Leptin-a mediates transcription of genes that participate in central endocrine and phosphatidylinositol signaling pathways in 72-hour embryonic zebrafish (Danio rerio)

We analyzed microarray expression data to highlight biological pathways that respond to embryonic zebrafish Leptin-a (lepa) signaling. Microarray expression measures for 26,046 genes were evaluated from lepa morpholino oligonucleotide “knockdown”, recombinant Leptin-a “rescue”, and uninjected control zebrafish at 72-hours post fertilization. In addition to KEGG pathway enrichment for phosphatidylinositol signaling and neuroactive ligand-receptor interactions, Gene Ontology (GO) data from lepa rescue zebrafish include JAK/STAT cascade, sensory perception, nervous system processes, and synaptic signaling. In the zebrafish lepa rescue treatment, we found changes in the expression of homologous genes that align with mammalian leptin signaling cascades including AMPK (prkaa2), ACC (acacb), Ca2+/calmodulin-dependent kinase (camkk2), PI3K (pik3r1), Ser/Thr protein kinase B (akt3), neuropeptides (agrp2, cart1), mitogen-activated protein kinase (MAPK), and insulin receptor substrate (LOC794738, LOC100537326). Notch signaling pathway and ribosome biogenesis genes respond to knockdown of Leptin-a. Differentially expressed transcription factors in lepa knockdown zebrafish regulate neurogenesis, neural differentiation, and cell fate commitment. This study presents a role for zebrafish Leptin-a in influencing expression of genes that mediate phosphatidylinositol and central endocrine signaling.


INTRODUCTION
In humans and rodents, the absence of a functional leptin (Ob/Ob) or leptin receptor (Db/Db) causes hyperphagia, early-onset and morbid obesity, severe type II diabetes, hypogonadotropic hypogonadism, impaired thermogenesis, and dysregulation of bone growth and immune response (Friedman & Halaas, 1998). Leptin (LEP), a conserved 16 kDa pleiotropic peptide hormone, is primarily secreted from adipocytes in proportion to fat mass (Maffei et al., 1995;Zhang et al., 1994). Leptin receptor (LEPR) has 6 isoforms in mammals that vary in cytoplasmic domain length, tissue distribution, and signaling competence (Chen et al., 1996;Lee et al., 1996). The leptin receptor long isoform (OB-RB) is expressed in the hypothalamus and interacts with members of the Janus kinase (JAK) and signal transducers and activators of transcription (STAT) protein families. Leptin signaling in mammals mediates anorectic (CART, POMC) and orexigenic (AgRP, NPY) circuits in the arcuate nucleus as well as voltage-gated calcium channels in the lateral hypothalamus (Baskin et al., 1999;Jo et al., 2005;Park & Ahima, 2014;Schwartz et al., 1997). Ob/Ob, and Db/Db obese rodent mutants, and diet-induced obesity mammals have reduced STAT-3 signaling in the arcuate nucleus which is generally attributed to either an absence of the hormone, an inability to respond to the hormone, or central leptin resistance (Ghilardi et al., 1996;Münzberg, Flier & Bjørbaek, 2004;Vaisse et al., 1996). Tissue-specific deletion of OB-RB in rodent neurons, but not hepatocytes, results in obesity, indicating that the regulatory effects of leptin signaling on adipose mass are mediated centrally (Cohen et al., 2001). Leptin supplementation rescues morbid phenotypes associated with Ob/Ob but not with Db/Db rodents which agrees with parabiosis experiments (Coleman, 1973); these findings mimic the effects of recombinant leptin in LEP −/− and LEPR −/− human case studies (Farooqi et al., 1999;Farooqi et al., 2002;Farooqi et al., 2007;Gibson et al., 2004;Montague et al., 1997;Pelleymounter et al., 1995).
Are the pathways stimulated by mammalian leptin signaling conserved across vertebrates? Genomes from all vertebrate classes have at least one copy of leptin (Londraville et al., 2017). Despite low sequence homology, fish leptin genes share a syntenic relationship with human lep (Gorissen et al., 2009;Howe et al., 2013;Kurokawa, Uji & Suzuki, 2005), and the zebrafish genome has orthologous members of other cytokine and receptor families including IFN-I (interferon type I), IFN-II (interferon type-II), IL (interleukins), chemokines, and TNF (tumor necrosis factors) (Savan & Sakai, 2006). Although all mammals express single LEP genes, many fishes express multiple leptin paralogs (likely due to an ancestral teleost whole-genome duplication event; Jaillon et al., 2004). As opposed to single LEP orthologues in humans and rodents, the impact of having multiple copies of leptin on fish physiology and gene regulation is not well-understood. Zebrafish lepa message is abundant in liver, and lepb expression is highest in ovary (Gorissen et al., 2009). The putative zebrafish lepb promoter contains a ∼1.3 kb enhancer element that responds to regeneration cues, and hepatic lepb expression is downregulated in response to fasting (Gorissen et al., 2009;Kang et al., 2016). Zebrafish lepa is induced by hypoxic cues in adult liver and embryonic muscle, presumably through hypoxia inducible transcription factor 1 (Chu, Li & Yu, 2010). The induction of lepa and lepb expression in response to hypoxia and tissue regeneration, respectively, suggests that zebrafish leptins may integrate different responses to stress. The zebrafish genome has one leptin receptor (lepr) which has broad tissue expression including gill, liver, ovary, spleen, gut, heart, and pituitary (Gorissen et al., 2009;Liu et al., 2010). In contrast to mammalian Db/Db analogues, adult zebrafish leptin receptor knockouts (lepr sa1508 ) do not have differences in adiposity, feeding, fecundity, or body size compared to controls (Michel et al., 2016), unlike the response to lepr knockout in medaka (Chisada et al., 2014). Knockout lepr sa1508 larvae have increased numbers of β-cells as well as upregulation of insulin-a and glucagon-a (which can be blocked by metformin) suggesting that peripheral leptin signaling accommodates glucose homeostasis, but not adipostasis in zebrafish (Michel et al., 2016). A more complete understanding of zebrafish Leptin-a and Leptin-b binding to the receptor, and their respective signaling cascades is needed to understand leptin function in fishes.
Stat-3 signaling stimulated by leptin has been documented for several fish species (Gong, Jönsson & Björnsson, 2016;Wu et al., 2016a;Wu et al., 2016b;Douros et al., 2018) and frogs (Cui et al., 2014). For Xenopus (Cui et al., 2014) and Tilapia (Douros et al., 2018), large-scale transcriptomic response to leptin stimulation is established. These innovative studies add much to our understanding of the evolution of leptin response, but they do not measure transcriptomic response with reduced leptin signaling.
Zebrafish embryos (<1 hpf) injected with lepa morpholinos (lepa knockdown) are verified to have reduced Leptin-a protein expression, developmental abnormalities (reduced eye size, reduced otic vesicle formation, bent notochord), and reduced metabolic rate, all of which are ameliorated by co-injection of recombinant zebrafish Leptin-a (lepa 'rescue';Liu et al., 2012;Dalman et al., 2013). In this study we used the established lepa knockdown model to investigate the role of Leptin-a in embryonic gene regulation. RNA prepared from lepa knockdown, lepa rescue, and control zebrafish embryos was processed on single-channel microarrays; digital expression estimates were generated for 26,046 unique gene symbol identifiers across 16 total samples. Differentially expressed genes (DEGs), KEGG pathways, and Gene Ontologies (GO) were evaluated for three pairwise comparisons: lepa knockdown compared to uninjected controls; lepa rescue compared to lepa knockdown; and lepa rescue compared to uninjected controls. We tested the null hypothesis that zebrafish Leptin-a mediates energy homeostasis through central endocrine mechanisms analogous to human and rodent leptins. We expected that the transcription of genes in adipocytokine and lipid signaling, endocrine pathways, JAK/STAT cascades, glucose homeostasis, and immune function respond to embryonic zebrafish lepa knockdown and recombinant Leptin-a rescue treatments.

Animal care and ethical procedures
All zebrafish and associated animal procedures were reviewed and approved by The University of Akron's Institutional Animal Care and Use Committee (IACUC, approval reference # 08-6B). Adult zebrafish were obtained from Aquatic Tropicals (Bonita Springs, FL), and maintained in aquatic fish housing systems at 28.5 • C on 13:11 light/dark cycles. Embryonic life staging, aquarium and animal maintenance procedures, diet, and husbandry approaches were completed as described in 'The Zebrafish Book' (Westerfield, 1995). Age of embryo was scored as hours post fertilization (hpf) or days post fertilization (dpf). Embryos were raised at 28.5 • C using tank water supplemented with 0.01% (w/v) methylene blue (Sigma Aldrich) from 0-2 dpf.

qPCR validation of microarray results
Relative expression for 96 transcripts via qPCR was estimated in the lepa knockdown, lepa rescue, and control groups (prepared as detailed in section 2.2) using RT 2 Signal Transduction Pathway Finder qPCR Arrays (Qiagen). Each qPCR array was prepared in duplicate (n = 2 for control, knockdown, and rescue treatments) with RNA derived from 5 embryos, and equal amounts of total RNA was reverse transcribed with a qScript Flex cDNA synthesis kit (Quanta Bio) with oligo dT primers (along with no template, no primer, and no reverse transcriptase controls). Primer sequences for the qPCR assays are available from the manufacturer (cat# PAZF-014Z) including 5 reference genes (acta 1b, b2m, hprt1, nono, rpl13a) to which the data were normalized. Assays were run with RT 2 SYBR Green Master Mix (Qiagen) on an Applied Biosciences 7300 cycler. Data were analyzed with the Qiagen Gene Globe data analysis web portal (Qiagen). Within the 96 transcripts, those identified as DEG by the microarray analyses were compared across assays (qPCR vs. microarray).

Differentially Expressed Genes (DEGs)
We previously validated lepa morpholino knockdown using an immunoblot and mismatch morpholinos (Liu et al., 2012). At 72 hpf, lepa knockdown have reduced metabolic rate (Dalman et al., 2013), physical malformations pertaining to neurosensory organs (small eyes, otoliths, hindbrain), enlarged pericardial cavity and yolk sac, irregular curvature of the notochord and tail, thinning of spinal nerves, and reduced pigmentation. Injection of both rxLeptin-a and lepa morpholinos alleviates the lepa knockdown phenotype where embryos injected with higher concentrations of rxLeptin-a more closely resemble uninjected control morphology (Liu et al., 2012). Here, microarray expression estimates from 75,212 probesets (∼1.2 million probes) were filtered to represent 26,046 unique gene symbol identifiers derived from the Zv9 reference assembly. In total, n = 16 microarrays were used to evaluate gene expression data derived from 40 uninjected control embryos, 20 embryos with lepa morpholino knockdown, and 20 embryos with lepa rescue; each microarray sample was derived from a pool of 5 embryos at 72 hpf.
Differentially expressed genes (DEGs) have log 2 fold change <−0.5 or >0.5 and adjusted p.value <0.01 ( Fig. 1). DEGs were rank-ordered by adjusted p.value. Genes that are differentially expressed between lepa knockdown and uninjected control microarray samples represent changes in gene expression that correspond to decreased Leptin-a signaling. Similarly, DEGs in the lepa rescue samples compared to uninjected controls present the combinatorial effect of reduced lepa translation (via morpholino) and induction by recombinant protein. Genes that are differentially expressed in lepa rescue compared to lepa knockdown samples respond to the induction of Leptin-a signaling (alone) because both conditions were treated with the same concentration of lepa morpholinos. While these microarray data alone cannot distinguish between first or second order gene targets of zebrafish Leptin-a signaling, this study is the first to measure the whole transcriptome response to reduced leptin signaling in a non-mammal.
A total of 19,987 genes were not differentially expressed in any part of the microarray dataset (76.73%). 1,461 genes were differentially expressed between lepa knockdown and control treatments; 425 of the 1,461 probesets were unique to this comparison. Similarly, there were 5,105 DEGs in the lepa rescue treatment compared to the controls; 3,125 of the 5,105 probesets were unique to this comparison. Additionally, 1,716 genes were differentially expressed in lepa rescue samples compared to lepa knockdown where 329 of the 1,716 probesets were unique to this comparison. Ultimately, 43 genes were differentially expressed among all three pairwise comparisons (Fig. 1). DEGs in each comparison are appended in the Supplementary Gene List. Microinjections, library preparation, and microarray scanning for n = 16 samples were executed across 5 independent trials. The sample median probeset intensities are consistent across the series of microarrays regardless of scan date or treatment (Fig. S1). Principle component analysis illustrates that the cumulative proportion of variance for principle components 1, 2, and 3 is 65.97%, or 38.77%, 18.58%, and 8.62%, respectively. lepa knockdown, lepa rescue, and uninjected control samples form segregated clusters that do not contain members from different Volcano plots (1A-1C) represent the adjusted p.value versus fold change for each gene across three pairwise comparisons (listed below each plot). Each point represents one gene. Differentially expressed genes are teal; red points denote genes that do not meet differentially expressed selection criteria. Venn diagram (1D) represents the number of differentially expressed genes (DEGs) in each comparison at adjusted p.value < 0.01 and log 2 fold change ≤ 0.5 or > 0.5. ''morphant-control'' refers to genes that are differentially expressed in lepa knockdown zebrafish compared to uninjected controls. ''rescue-morphant'' refers to genes that are differentially expressed in lepa rescue zebrafish compared to lepa knockdown. ''rescue-control'' refers to genes that are differentially expressed in lepa rescue zebrafish compared to uninjected controls.
Full-size DOI: 10.7717/peerj.6848/ fig-1 treatments, demonstrating that the most variable features in the dataset are more similar among microarrays with a common treatment as opposed to common scan date (Fig. S1).
The 50 top-ranked DEGs from each pairwise comparison are listed in Tables 1-3. KEGG and GO enrichment analyses were performed on each differentially expressed gene set (separately) to generate inferences on the regulation of the transcriptome by reduced versus induced Leptin-a signaling using knockdown and rescue treatments, respectively. Opsin 1 medium-wave-sensitive 1 (opn1mw1) has the largest reduction in expression of all genes in both lepa knockdown (−6.10 log 2 fold change) and lepa rescue (−3.82 log 2 fold change) treatments compared to control. In the lepa knockdown and control contrast, the top-5 ranked genes include immunoglobulin superfamily DCC subclass member 3-like (LOC100537029, 3.27 log 2 fold change) and immunoglobin superfamily member 21b (igsf21b, −2.25 log 2 fold change). Von Willebrand factor (vwf; 3.55 log 2 fold change) and SI:DKEY-24I24.3 (3.52 log 2 fold change) are the two highest-ranked genes in the lepa rescue treatment compared to the controls. In lepa rescue zebrafish, leptin-b (lepb) is upregulated (1.76 log 2 fold change) compared to controls. In lepa knockdown zebrafish, lepa and lepr are differentially expressed (1.01 and 0.77 log 2 fold changes, respectively) compared to controls which may be compensatory to lepa knockdown.

Gene ontology enrichment
The lepa knockdown compared to control GO data were generated from 1,461 DEGs (970 Entrez gene ID's); lepa knockdown have 54 enriched biological process terms (P < 0.01) including visual perception, ncRNA metabolic process, rRNA metabolic process, ribosomal large/small subunit biosynthesis, neurological system process, protein-chromophore linkage, spinal cord development, nervous system development and neuron differentiation, regulation of neurotransmitter levels and secretion, synaptic signaling, and regulation of neurogenesis (Fig. 5). The GO data for the lepa rescue and knockdown contrast were generated from 1,716 DEGs (999 Entrez gene ID's) that correspond to 22 enriched terms including sensory perception, Notch signaling, brain/head/gland/ventral spinal cord development, regionalization, olfactory and semaphorin receptor activity, and protein tyrosine kinase activity (Fig. 6). Biological process Gene Ontologies from the lepa rescue and control comparison were produced using 5,105 DEGs (3,036 Entrez gene IDs); lepa rescue has 39 enriched biological process terms including JAK-STAT cascade, nervous system process, synaptic signaling, synapse organization, sensory perception, regulation of voltage-gated calcium channels, neurotransmitter secretion, and phototransduction (Fig. 7). Genes that map to each GO category in Figs. 5-7 are appended in the Supplementary Gene List. From 1,461 DEGs in lepa knockdown compared to controls, there were 112 transcription factors (7.67%) (Fig. S2). There are 5 differentially expressed transcription factors with a log 2 fold change greater than 2 (foxn4, fosl1a, cbx7a, and atf5a). Two transcription factors (pou2f2a, insm1a) ranked in the top-10 of all 1,461 DEGs between lepa knockdown and controls (Table 1). Additionally, the Gene Ontology enrichment analysis for transcription factor biological process has 65 terms that envelop many aspects of development (spinal cord, endocrine system, pancreas, brain, head, neural retina, nervous system, epithelium), as well as regulation of neurogenesis, neuron differentiation, cell fate commitment, and hindbrain morphogenesis (Fig. S3).

qPCR validation of the microarray
Several genes within the qPCR dataset were also identified as DEG in the microarray experiments (10 for knockdown vs. control and 5 for rescue vs. control; Fig. S4). Fold differences >1 are positive changes in expression, and <1 are negative. The direction of expression change is the same between microarray and qPCR experiments for 8 out of 10 genes for the comparison of knockdown to control. For the rescue vs. control comparison, 1 out of 5 genes has the same direction of change when expression is measured by microarray vs. qPCR.

DISCUSSION
Null mutations in leptin (LEP) or leptin receptor (LEPR) impair lipid metabolism, endocrine signaling, and energy homeostasis in humans and rodents, but there are few non-mammal model organisms with reduced leptin signaling. Mammalian leptin is widely characterized in adults, and to a lesser extent in juveniles or embryos (Antczak & Van Blerkom, 1997). Embryonic leptin expression precludes terminal adipocyte differentiation in mammals suggesting that leptin modulates embryogenesis apart from its role as an adipocytokine (Antczak & Van Blerkom, 1997). Given the intrinsic limitations of in vivo embryonic leptin manipulation using mammals, we tested transcriptomic response to leptin manipulation in developing zebrafish using lepa morpholino knockdown. Our  lepb, lepr, LOC794738, LOC100537326, akt3, pik3r1, prkaa2, socs2, socs5b, socs9, npy8br, npy8ar, mc5ra, foxo1b, jak2a, jak2b, camkk2, map3k12, agrp2 and cart1. Genes displaying large fold changes (vwf, opn1lw1, igsf21b, LOC100537029), and others that regulate lipid (acacb, dagla, daglb, lpl, crot, plin1, cpt1b, pank2) or RNA metabolism (smg5, dicer1, LOC796505, LOC570775) are also included. Each gene (row) is color coded to represent high (red) and low (green) levels of expression. Columns are color coded by treatment-blue (lepa rescue), red (lepa knockdown), green (uninjected controls). Zebrafish lepa and lepb are co-expressed in zebrafish pituitary which may mediate endocrine and local (paracrine, autocrine) axes, respectively (Gorissen et al., 2009). Affirming a role for Leptin-a signaling in the central nervous system, neuroactive ligand-receptor interactions are enriched KEGG pathways in lepa knockdown and rescue zebrafish compared to controls, but fewer genes respond to knockdown as opposed to rescue suggesting there are differential responses to reduced versus induced Leptin-a signaling (Fig. 2). Additionally, lepa rescue DEGs are analogous to neuroendocrine factors associated with the human leptin pathway (and classical mediators of changes in appetite and metabolism), including agouti-related peptide-2 (agrp2), cocaine-andamphetamine-regulated transcript 1 (cart1), neuropeptide Y receptor Y8 (npy8ar, npy8br), and melanocortin receptor 5a (mc5ra) (Fig. 4B). Zebrafish homologs of the human lep/pi3k/akt, and lep/shp2/mapk cascades are differentially regulated in lepa rescue (but not lepa knockdown) treatments including protein kinase B (akt3), phosphatidylinositol 3-kinase (pik3r1), insulin receptor substrate (LOC794738, LOC100537326 ), Janus kinase 2 (jak2a, jak2b), forkhead box O1 (foxo1b), AMP-activated protein kinase (prkaa2), acetyl-CoA carboxylase beta (acacb), calcium/calmodulin kinase (camkk2), mitogen-activated protein kinase kinase kinase 12 (map3k12), and suppressors of cytokine signaling (socs2, socs5b, socs9) (Fig. 4B). Consequently, KEGG phosphatidylinositol signaling (Fig. 3) and GO biological process JAK/STAT cascade (Fig. 7) are enriched in lepa rescue zebrafish. In addition to prkaa2 and acacb, lipolytic factors also respond to lepa rescue (but not knockdown) including pantothenate kinase 2 (pank2) which mediates a critical step of coenzyme-A biosynthesis, lipoprotein lipase (lpl), diacylglycerol lipase (dagla, daglb), perilipin-1 (plin1), carnitine O-octanoyltransferase (crot ), and muscle carnitine palmitoyltransferase 1B (cpt1b) (Fig. 4B). Although lists of differentially expressed genes are a useful outcome of these studies, perhaps the most relevant results come from GO (gene ontology) and pathway analyses (e.g., KEGG). The goal of these analyses is to uncover functions/pathways that are statistically over/underrepresented in the treatment. Once a pathway is identified, all genes within the pathway are legitimate targets for downstream hypothesis testing, regardless of whether each gene in the pathway is differentially expressed (Khatri, Sirota & Butte, 2012). Notch signaling is differentially regulated between lepa knockdown and rescue treatments ( Fig. 4A; Table 4). Notch signaling is a conserved juxtacrine signaling pathway that, in addition to many other roles, directs neuron cell fate (Wakeham et al., 1997). Notch signaling, neurogenesis, and neural differentiation respond to lepa knockdown ( Fig. 4A; Fig. S3). GO enrichment pertaining to synaptic signaling, sensory perception, and neurotransmitter secretion is enriched in both lepa knockdown and rescue treatments compared to controls Figs. 5 and 7. Ob/Ob and Db/Db rodents have impaired nerve fiber extensions in arcuate nucleus (Bouret, Draper & Simerly, 2004). Similarly, lepa and LEPR zebrafish knockdown have thinning of spinal nerves along with malformations of sensory organs (Liu et al., 2012) indicating that there may be conserved roles between the embryonic zebrafish and human leptin signaling pathways tied to neurogenesis and synapse organization. This pathway analysis may also be less susceptible to gene:gene variation among replicates. Validation of expression data for 10 genes between the microarray and control groups is generally good, with 8/10 genes matching in the direction of change (e.g., up or down both in microarray and qPCR data, Fig. S4). However, agreement between the two methods for the rescue vs. control comparison is poor (1 of 5). We speculate that this treatment is inherently more complex (both knockdown with MO and rescue with RX leptin), and thus the response more variable. Other transcriptomics studies also report relatively low rates of agreement between transcriptomic and qPCR data (Cui et al., 2014).
Other laboratories investigated transcriptomic response to leptin manipulation in non-mammals. Denver's laboratory measured individual gene response and transcriptome response to leptin incubation in Xenopus brain preparations (Cui et al., 2014). They noted strong response of socs genes, similar to this study (Fig. 4B). Recently, Borski's laboratory published results of an ex vivo transcriptomic study on adult tilapia pituitary tissue treated with tilapia Leptin-a in culture (Douros et al., 2018). The overwhelming effect was increased carbohydrate metabolism, consistent with a leptin knockout in zebrafish (Michel et al., 2016). Several aspects of our study are consistent with Douros et al. (2018), such as protein kinase B (akt3), Janus kinase 2 (jak2a, jak2b), insulin receptor substrate (LOC794738, LOC100537326 ), AMP-activated protein kinase (prkaa2), acetyl-CoA carboxylase beta (acacb) (Fig. 4B), and the general stimulatory effects on ribosome assembly and function (Table 4). Differences between the two studies likely reflect developmental window (embryos vs. adults), and tissue (whole embryo vs pituitary and liver); it is not surprising that leptin may have different effects on the transcriptome of developing whole embryos versus specific tissues of adults, even in the same species.
Phototransduction genes elicit the same response to lepa knockdown and rescue treatments compared to controls suggesting this effect is not directly correlated with Leptin-a signaling ( Fig. 4A; Table 4). Correlations between the lepa knockdown dataset and unrelated morpholinos suggest that there may be a phototransduction artifact of morpholinos which is often manifested as reduced eye size and opsin expression. Having acknowledged these possibilities, it is also important to note that these artifacts are typically seen at much higher morpholino dose/embryo than used in this study. Although we assert that our results generally reflect the effects of lepa knockdown in zebrafish embryos, it is also likely that the general effects of morpholino oligonucleotides are reflected in the data. Eukaryotic gene expression is regulated at multiple levels including transcriptional machinery and epigenetic factors, mRNA processing and transport, and post-translational modifications. The transcriptome represents the complete spectrum of RNA species that are present at one reference point in time. Expression microarrays are molecular technologies adapted to qualitatively and semi-quantitatively describe transcribed regions of the genome using a single experiment. Structurally, antisense morpholino oligonucleotides (MO) are short ∼25 mer, nucleic acid analogues with alternative morpholine rings in place of the native deoxyribose and ribose moieties present in DNA and RNA, respectively (Summerton & Weller, 1997). Uncharged phosphodiamorate linkages between morpholino nucleotides juxtapose the negatively charged phosphodiester backbones of native DNA/RNA. These structural rearrangements confer resistance to RNase H nucleolytic cleavage while the uncharged MO backbone reduces off-target electrostatic interactions during delivery and diffusion (Summerton, 1999). MOs are generally directed against the translation initiation site of a sense strand mRNA as described here. Alternatively, mRNA transport, maturation, and processing can be manipulated by directing MOs against splice boundaries of precursor mRNAs (Kloosterman et al., 2007). Gene knockdown is catalyzed by complementary basepairing between antisense MO and target mRNA. Translation-blocking MO:mRNA hybrids sterically exclude the ribosome from executing translation which leads to reduced target protein synthesis and/or impaired processing of target RNA (Bill et al., 2009;Summerton, 2007). However, MOs generally have a temporal range of effectiveness. Cytological concentrations of MOs are reduced with every subsequent cell division which restricts most MO applications to early life stages (Nasevicius & Ekker, 2000).
Finally, we understand that the interaction among the morpholino and recombinant protein is/are likely complex. The degree to which the knockdown have reduced expression, and the degree to which the recombinant leptin replaces what is reduced by morpholinos is unlikely a perfect match. This undoubtedly contributes to different pathways identified in each of the comparisons.

CONCLUSIONS
These microarray data describe cell signaling pathways and gene targets regulated by embryonic zebrafish Leptin-a. Differentially expressed genes from the lepa rescue zebrafish embryos influence expression of genes that participate in central endocrine and phosphatidylinositol signaling pathways which agrees with the function of leptin signaling in mammals. Regulators of metabolism respond to lepa rescue that agree with LEP/AMPK/ACC and LEP/PI3K/AKT cascades in human (Minokoshi et al., 2002) ( Fig. 4B). There are also markers consistent with increased central leptin signaling in the CNS among lepa rescue zebrafish which includes cocaine-and-amphetamine regulated transcript like-1 (cart1; 0.99 log2 fold change) and agouti-related peptide-2 (agrp2; -1.21 log2 fold change). Although we did not identify differentially expressed members of the STAT transcription factor family, Janus kinase 2 (jak2a, jak2b), insulin receptor substrate (LOC794738, LOC100537326 ), insulin receptor a (insra), Ser/Thr protein kinase B (akt3), PI3K (pik3r1), and forkhead box O1 (foxo1b) are differentially expressed in lepa rescue zebrafish which aligns with coregulation of the central PI3K/Akt/FoxO1 cascade by leptin and insulin receptor signaling in mammals (Xu et al., 2005). Taken together, our results propose that cell signaling pathways regulated by the A-type zebrafish leptin paralog align with human and rodent leptins including LEP/AMPK/ACC and LEP/PI3K/AKT but functional assays using fish leptin signaling knockouts are needed to validate these gene expression data.

ADDITIONAL INFORMATION AND DECLARATIONS Funding
The experiments in this article were supported by NIH 2R15DK079282-02 to Richard Londraville and Qin Liu and The Choose Ohio First Bioinformatics Scholarship Program provided by the Ohio Board of Reagents to Matthew Tuttle and Mark Dalman. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.