Characterization of Cyclooxygenase-2 and its induction pathways in response to high lipid diet-induced inflammation in Larmichthys crocea

The present study was conducted to investigate the effects of a high-lipid diet (HLD) on cyclooxygenase (Cox)-2 expression and the signalling pathways related to low-grade inflammation in the large yellow croaker (Larmichthys crocea). An isolated 2508 bp cDNA clone of cox-2 contained an open reading frame spanning 1827 bp encoding a protein with 608 amino acid residues. The over-expression of cox-2 was consistent with the activation of c-Jun N-terminal kinases (JNKs) and p38 mitogen-activated protein kinase (MAPK) in HLD-fed fish. The activation of the activator protein-1 (AP-1) and the nuclear transcription factor kappa-B (NF-κB) signalling pathways in HLD-fed fish and the significant increase of cox-2 promoter-luciferase activity in vitro indicated that AP-1 and NF-κB could combine cox-2 promoter to promote its transcription, respectively. Together, HLD-induced inflammation up-regulates cox-2 expression via JNKs and p38 MAPK-dependent NF-κB and AP-1 pathways. The present study provides important insight into the signal transduction pathways involved in HLD-induced inflammation, which is detrimental to the health and production of fish as well as to the health of fish consumers.

Scientific RepoRts | 6:19921 | DOI: 10.1038/srep19921 kinases 1&2 (ERK1/2) that are related to cell proliferation and survival 17,18 , JNKs and p38 MAPK that are related to inflammation and programmed cell death 19,20 . Several studies in mammals have reported that MAPKs could be activated by a HLD, and transduce the HLD stimulus to the nucleus by mediating downstream signalling cascades, such as NF-κ B and AP-1 21 . However, little information is available on the activation of MAPKs and their effect on downstream transcription factors in response to a HLD in fish.
Large yellow croaker (Larmichthys crocea), with its high level of production and critical role in human food or health 22 , has been widely cultured in southeast China. The use of a HLD has become more common in large yellow croaker commercial diets, because of its protein sparing effect. The long-term use of a HLD abnormally increases lipid accumulation in the liver 23 and the level of inflammation characterized by increased immune-related gene expression 8 and mitochondrial function disorder 3 . Furthermore, previous studies have indicated that lipid deposition and inflammatory signals of the large yellow croaker are strikingly similar to those of other fish species and mammals 23 . Thus, the large yellow croaker is an appropriate fish species to investigate the mechanisms of HLD-induced inflammation.
Studies on nutrition of this fish have been intensively conducted in recent years [24][25][26][27][28][29] , but no information is available on the molecular mechanism of HLD-induced inflammation. Thus, in this study, cox-2 from large yellow croaker was cloned and characterized, and the effect of a HLD on cox-2 expression and the signalling transduction pathway involved were investigated.

Results
Characterization of full-length cox-2 from large yellow croaker. The full-length cDNA of cox-2 from large yellow croaker was 2508 bp (GenBank Accession No. KP259877), including a 5′ untranslated terminal region (UTR) of 106 bp, a 3′ UTR of 575 bp, and an open reading frame (ORF) of 1827 bp encoding a polypeptide of 608 amino acid residues with a predicted molecular weight of 69.03 KDa and a theoretical isoelectric point of 6.74 (Fig. 1).
Tissue distribution of cox-2 in the large yellow croaker. The tissue-specific expression of cox-2 was examined in multi-tissues, including gill, intestine, kidney, liver, adipose, muscle, stomach and eye. The cox-2 transcripts were broadly expressed in all of the detected tissues. The highest expression was found in gill, followed by intestine, kidney, liver, adipose, muscle and stomach, whereas the lowest expression level was observed in eye ( Fig. 4).

Cox-2 and pro-inflammatory cytokine expression in response to dietary fatty levels. Compared
with that in the control group, cox-2 transcript levels were up-regulated in the liver of HLD fish by approximately 2.78-fold (P < 0.05) (Fig. 5). The levels of cox-2 translation paralleled its transcript levels (Fig. 6). The transcript expression levels of the pro-inflammatory cytokine TNFα and IL-1β were also up-regulated in HLD fish compared with the control group (P < 0.05) (Fig. 5). These data indicate that cox-2 might be induced by an HLD and may play a critical role in inflammation.
MAPK activity, AP-1 and NF-κB signal pathways involved in dietary lipid level-induced cox-2 expression. The HLD increased the phosphorylation of JNK and p38 MAPK in HLD fish compared to the control group, whereas no significant effect on the phosphorylation status of ERK1/2 was observed among the different lipid levels (Fig. 7). No significant difference in ERK1/2, JNK and p38 total protein was observed among the different dietary lipid levels. Therefore, it is likely that JNK and p38 MAPK but not ERK 1/2 up-regulated cox-2 expression. The phosphorylation of c-Jun was increased by the HLD, indicating the nuclear translocation level of c-Jun. Similarly, the HLD increased IKKα /β and Iκ Bα protein phosphorylation, which induced the degradation of cytoplasmic Iκ Bα and subsequently increased the nuclear level of NF-κ B subunits p65. These data indicate that the HLD stimulated the expression of cox-2 depending on the AP-1 and NF-κ B pathways via the activation of the JNK and p38 MAPK pathways (Fig. 8).
Dual-luciferase reporter assays. Compared with the empty vector pCS2+ , the expression of recombinant pCS-NF-κ B and pCS-AP-1 resulted in 2.91-and 2.22-fold (P < 0.05) increases in pGL3-Cox-2 expression, respectively, and the combination of pCS-NF-κ B and pCS-AP-1 resulted in a 4.15-fold (P < 0.05) increase in pGL3-Cox-2 expression. Compared to the combination of pCS-NF-κ B and pCS-AP-1 without an inhibitor, the combination of pCS-NF-κ B and pCS-AP-1 with a NF-κ B or AP-1 inhibitor resulted in a significant decrease in pGL3-Cox-2 expression. Compared to pCS-AP-1 or pCS-NF-κ B independent treatments, no significant differences were observed in pGL3-Cox-2 expression of the combination of pCS-NF-κ B and pCS-AP-1 with a NF-κ B or AP-1 inhibitor treatment. These results indicate that both NF-κ B and AP-1 activate the expression of luciferase reporter genes, suggesting that they could promote the transcription of cox-2 and that their regulation could be regarded as independent actions (Fig. 9).

Discussion
The ORF of cox-2 encodes a 608 amino acid polypeptide, which showed a high similarity to Cox-2 from other fish and mammals. The evolutionary relationship of Cox-2 is consistent with traditional taxonomy. Additionally, the Scientific RepoRts | 6:19921 | DOI: 10.1038/srep19921 highest cox-2 expression was found in gills, a primary defence against external pathogens, thus indicating that cox-2 may play a significant role in innate immunity.
Previous reports on fish species have indicated that cox-2 is expressed at low levels in resting cells and is significantly induced by various fatty acids, such as saturated fatty acids 8 and n-6 polyunsaturated fatty acids (PUFAs) 9 . Saturated fatty acids are positively associated with inflammation, mainly due to the significant effects of palmitic and stearic acids. Many studies have demonstrated that saturated fatty acids increase the saturation of membrane phospholipids 30 , initiate the unfolded protein response 31 , lead to endoplasmic reticulum stress 32 , affect mitochondrial metabolism 33 and promote ROS accumulation 34 . In general, n-6 PUFAs have pro-inflammatory activity and can play important roles in immune function. Human inflammatory cells typically contain high proportions of the n-6 PUFA arachidonic acid (20:4n-6) and low proportions of n-3 PUFA, mainly because arachidonic acid is the precursor of two-series prostaglandins and four-series leukotrienes, which are highly-active mediators of inflammation 35 . Furthermore, some negative effects of obesity are associated with excessive levels of n-6 PUFAs 36-39 . In the present study, HLD-induced inflammation may be due to the high levels of saturated fatty acid and n-6 PUFAs in the HLD. Although many studies involving fatty acids induced inflammation have been reported in mammals, limited information is available in fish.
The over-expression of cox-2 plays a crucial role in the development and progression of inflammation 40,41 . In the present study, high levels of cox-2 expression were found in the liver of HLD-fed fish, and the increased cox-2 expression was associated with increased pro-inflammatory cytokine TNFα and IL-1β transcripts (Fig. 5). In mammals, the up-regulated cox-2 expression in liver is associated with pathological conditions, such as acute liver failure, hepatic fibrosis and cirrhosis, and hepatocarcinogenesis 41,42 . Given the conservation of cox-2 function, cox-2 over-expression in the present study supports its important role in HLD-induced hepatic inflammation. In fish, MAPKs are associated with cox-2 expression through several stimuli, including metals 43,44 and parasites 45 , but the effects of an HLD on MAPKs remain unknown. A number of reports in mammals have demonstrated that several signal transduction pathways are simultaneously stimulated by an HLD, and these signals appear to converge at the common MAPK pathway 46,47 . In the present study, the HLD caused a marked increase   Relative cox-2 mRNA expression was determined by quantitative realtime PCR (qRT-PCR) and expressed relative to β -actin levels. Results are expressed as means ± S.E.M. (n = 3). Different letters above the bars denote significant differences among tissues at the P < 0.05 level (P = 0.000) as determined by one-way ANOVA followed by Tukey's test (SPSS). in the level of phosphorylated JNK and p38 MAPK, whereas the phosphorylation of ERK was not affected (Fig. 7). These findings suggest that two types of MAPKs, JNK and p38 MAPK, but not ERK, are involved in the HLD-induced up-regulation of cox-2.
Previous studies have suggested that MAPKs activate separate effector pathways, demonstrating crosstalk between receptor systems that converged further downstream in the cell nucleus in mammals. The AP-1 and NF-κ B activation can be selectively triggered by all MAPKs in a stress-dependent manner 48,49 . The qRT-PCR results indicated that an HLD up-regulated cox-2 expression at the transcriptional level (Fig. 5). The activation of transcription factor AP-1 and NF-κ B pathways was thoroughly studied to investigate the downstream molecular mechanism of HLD action during cox-2 induction. In the present study, a high level of c-Jun phosphorylation was observed in HLD-fed fish, which revealed increased c-Jun nuclear translocation and AP-1 activation (Fig. 8). The HLD also activated NF-κ B signalling pathways by the rapid phosphorylation of IKKα /β and the ubiquitination and proteolytic degradation of Iκ Bα , which then resulted in the translocation of NF-κ B to the nucleus (Fig. 8). In support of the findings in the present study, Dembinska et al. have demonstrated that the human obesity-induced inflammatory process depends on the AP-1 and NF-κ B signalling pathways 50 . Therefore, the HLD-induced over-expression of cox-2 is regulated by JNK and p38 MAPK through the AP-1 and NF-κ B pathways.
The human cox-2 promoter contains multiple potential cis-activating elements, including NF-κ B and AP-1 binding sites 51 . The transcription factors AP-1 and NF-κ B act as environmental sensors, detecting changes in the extracellular milieu through multiple signalling cascades 52 . Previous studies have demonstrated that NF-κ B is not required for lipopolysaccharide (LPS)-induced cox-2 expression in murine macrophages by the dominant negative inhibition of NF-κ B and cox-2 reporter gene activity 53 , whereas AP-1 binding activity is increased by LPS in macrophages 54 . Kim et al. have demonstrated that the application of a JNK inhibitor has no significant effect on papillomavirus E5-induced cox-2 expression compared with NF-κ B inhibitor and have suggested that NF-κ B is a major modulator of papillomavirus E5-induced cox-2 expression with AP-1 playing a reduced role 55 . These studies indicate that the regulation of AP-1 and NF-κ B in cox-2 transcription is not necessary through different stimuli. In the present study, dual-luciferase reporter assays showed that AP-1 and NF-κ B significantly increased cox-2 promoter-luciferase activity. As confirmed by pathway inhibitors, the regulation of cox-2 expression by AP-1 and NF-κ B occurred independently, in agreement with previous studies (Fig. 9). These findings indicate that AP-1 and NF-κ B interact with cognate cis-acting elements within the cox-2 promoter to promote cox-2 transcription independently.
In conclusion, the full-length cDNA encoding Cox-2 from the large yellow croaker was cloned and characterized. The transcription factor AP-1 and NF-κ B specifically bind to the cox-2 promoter and independently promote its expression. A HLD increases cox-2 transcription levels, which is mediated by JNKs and p38 MAPK-dependent NF-κ B and AP-1 activation. The present study provides important insight into the signal transduction pathways, which may potentially be applied to the discovery of disease mechanisms, thereby offering great benefits to the aquaculture industry and to human health.

Materials and Method
Ethics statement. The present study was performed in strict accordance with the Standard Operation Procedures (SOPs) of the Guide for the Use of Experimental Animals of Ocean University of China. All animal care and use procedures were approved by the Institutional Animal Care and Use Committee of Ocean University of China. Fish were anesthetized with eugenol (1:10,000) (Shanghai Reagent Corp., Shanghai, China) to minimize suffering before being assigned to cages and sampling. Figure 7. HLD activates the phosphorylation of JNK1/2 and p38 MAPK but not ERK1/2. Quantitative results of p-ERK1/2, p-JNK1/2 and p-p38 protein levels, which were adjusted with the total ERK1/2, p38, and JNK1/2 protein levels in the liver of experimental fish were analysed using Western blot. Data are expressed as A.U. of the Western blots and are depicted as a ratio of p-ERK1/2 (pThr202/Tyr204) to total ERK1/2, p-JNK1/2 (pThr183/Thr185) to total JNK1/2, and p-p38 (pThr180/Thr182) to total p38 (n = 3 in each group). All data are presented as mean ± S.E.M. *P < 0.05, **P < 0.01.
Scientific RepoRts | 6:19921 | DOI: 10.1038/srep19921 Experimental diets and feeding process. The experimental diets design and the feeding process were given in Yan et al. 56 . The ingredient and nutrient composition of the experimental diets are showed in Tables 1 and 2. In brief, whitefish meal and soybean meal were chosen as the main protein sources. Fish oil and soybean lecithin were chosen as the main lipid sources, which provided essential fatty acids and phosphatides, respectively. Three isoproteic (43% crude protein) diets were formulated to contain graded levels of lipid (12% and 18% on a Quantitative results of p-c-Jun, p-IKKα /β , p-Iκ Bα protein levels, which were adjusted with the total c-Jun, IKKα /β and Iκ Bα protein levels, were analysed using Western blot. Nuclear NF-κ B p65 protein levels which were adjusted with the total NF-κ B p65 protein levels, were also analysed using Western blot in the liver of experimental fish. Data are expressed as A.U. of the Western blots and are depicted as a ratio of p-c-Jun (pSer73) to total c-Jun, p-IKKα /β (pSer176/Ser180) to total IKKα /β , p-Iκ Bα (pSer32/Ser36) to total Iκ Bα and nuclear-NF-κ B p65 to total NF-κ B (n = 3 in each group). All data are presented as mean ± S.E.M. *P < 0.05, **P < 0.01.  Table 1). The diet with 12% crude lipid was used as the control because this dietary lipid level is optimal for the growth of large yellow croaker 57 .
At the start of the experiment, the fish were fasted for 24 h. The fish (average body weight 150.0 g) were randomly distributed into 6 cages (1.5 × 1.5 × 2.0 m) with 40 fish in each cage. Each diet was randomly allocated to triplicate cages, and the fish were hand-fed twice daily for 10 weeks.
Sample collection. At the termination of the experiment, fish were fasted for 24 h and anesthetized with eugenol (1:10,000) (purity 99%, Shanghai Reagent, China) before sampling. Tissues including kidney, intestine, spleen, heart, liver, brain and muscle from experimental fish were collected, frozen in liquid nitrogen and then stored at − 80 °C for the analysis of immune related gene expression.
RNA extraction and cDNA synthesis. RNA extraction and cDNA synthesis has been previously described in Yan et al. 56 with slight modification. In brief, total RNA was extracted from kidney, intestine, spleen, heart, stomach, liver, brain and muscle samples with Trizol Reagent (Invitrogen, USA) according to the manufacturer's instructions, and electrophoresed on a 1.2% denaturing agarose gel to test the integrity. The quantity and quality of the total RNA were assessed using the Nano Drop ® ND-1000 spectrophotometer (Nano-Drop Technologies, Wilmington, DE, USA). The 260/280 nm absorbance ratios of all samples ranged from 1.90 to 2.07, indicating a satisfactory purity of the RNA samples. First-strand cDNA was reverse transcribed from the DNase-treated RNA using PrimeScript TM RT reagent Kit (Takara, Japan).
The cloning and characterization of full-length cox-2 cDNA. The cloning has been previously described in Dong et al. with slight modification 58 . Degenerate primers were designed based on highly conserved regions from the genes of other fish to amplify internal fragments, and gene-specific primers were designed based on the known sequences of the internal fragments cDNA to clone the 3′ -and 5′ -end by rapid amplification of cDNA ends (RACE) through a two-round PCR using the SMARTer TM RACE cDNA Amplification Kit (Clontech, California, USA) ( Table 3). PCR amplifications using the primers (Table 3) and Taq DNA Polymerase (Takara, alian, China) were performed with an initial denaturation at 95 °C for 3 min and 35 cycles of "95 °C for 30 s, 60 °C for 30 s, and 72 °C for 1 min", followed by a final extension at 72 °C for 10 min. All PCR products were run on a 1.5% agarose gel, and then purified by SanPrep PCR urification Kit (Sangon Biotech, Shanghai, China). PCR products were cloned into pEASY-T1 simple cloning vector (TransGen, Beijing, China) and sequenced in BioSune (Shanghai, China). The nucleotide and deduced amino acid sequence of cox-2 from large yellow croaker were analyzed using BioEdit 7.0.1 and Expasysearch program (http://au.expasy.org/tools/). The sequences of cox-2 from different species were compared by the NCBI BLAST search program. A multiple sequence alignment was performed using ClustalW (http://www.ebi.ac.uk/clustalw/) and a phylogenetic tree of Cox-2 was made by MEGA 4.0 (http://www. megasoftware.net).
Quantitative realtime PCR (qRT-PCR) analysis. The first-strand cDNA synthesis was the same as described above. First-strand cDNA was diluted by 4 times using sterilized double-distilled water. Quantitative realtime PCR (qRT-PCR) was carried out in a quantitative thermal cycler (Mastercyclerep realplex, Eppendorf, Germany). The amplification was performed in a total volume of 25 μ l, containing 12.5 μ l of 2× SYBR ® Premix Ex TaqTMII (Takara, Japan), 9.5 μ l of sterilized double-distilled water, 1 μ l of each primer (10 μ M) ( Table 3) and 1 μ l of the diluted first strand cDNA product. The real-time qPCR amplification began with 2 min at 95 °C, followed by 40 cycles of 10 s at 95 °C, 10 s at 60 °C, and 20 s at 72 °C. Melting curve (1.85 °C increment/min from 58 °C to 95 °C) was performed after the amplification phase for confirmation. Each sample was run in triplicate. Reference Beta-actin gene was used as internal control 59 . A four-fold serial dilution of the cDNA samples quantified 5 concentrations in triplicate was used to assess PCR efficiencies for each assay. The primer amplification efficiency was analyzed according to the following equation E = 10 (−1/Slope)−1 . To calculate the expression of cox-2, the comparative CT method (2 −ΔΔct method) was used as described by Livak et al. 60 .

Western blot. Total protein was extracted from the liver of experimental fish with Total Protein Extraction
Kit (Applygen Technologies Inc, Beijing, China), and nuclear and cytosolic fractions were collected using a nuclear protein extraction kit (Pierce, Nashville, TN, USA), according to the instructions of the manufacturers. Protein concentration was measured using BCA kit (Pierce). Fifty micrograms of protein was loaded onto 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride membranes (Pall Corporation, Port Washington, NY, USA). The membranes were incubated with 5% skimmed milk for 1 h at room temperature. Immunoblots were obtained using antibodies against the following proteins: ERK1/2, ERK1/2 (pThr202/Tyr204), JNK1/2, JNK1/2(pThr183/Thr185), p38, p38 (pThr180/ Thr182), IKKα /β , IKKα /β (pSer176/Ser180), Iκ Bα , Iκ Bα (pSer32/Ser36), c-Jun, c-Jun (pSer73), NF-κ B p65 and beta-actin (Cell Signaling Technology, Danvers, MA, USA). Western blots were exposed using SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific, Waltham, MA, USA) and film images were scanned by Epson Perfection V33 (China).  and c-Jun ORF (GenBank Accession No. XM_010739773.1) fragments were amplified with the primers (Table 3) Statistical analysis. Software SPSS 17.0 (SPSS Inc.) was used for all statistical evaluations. All data were subjected to a one-way analysis of variance (ANOVA) and followed by Tukey's multiple-range test. The level of significance was chosen at P < 0.05 and the results were presented as means ± standard error of the mean.  Table 3. Sequences of the primers used in this study.