Histochemical localization of IGF-I and IGF-II mRNA in the rat between birth and adulthood

We describe the postnatal ontogeny and localization of insulin-like growth factors I and II (IGF-I and -II) in the rat. We have used oligodeoxyribonucleotide probes for in situ hybridization (hybridization histochemistry) and for Northern blotting. IGF-II mRNA is strongly expressed in liver, skeletal muscle, perichondrium, leptomeninges and choroid plexus of the newborn. Demonstrable levels fall dramatically in the liver at 18-20 days postnatally but persist for longer periods in muscle and remain undiminished throughout life in the pia/choroid plexus, indicating that different control mechanisms operate in these tissues. IGF-I mRNA is predominantly found in the liver. Its level in this organ rises well before levels of IGF-II fall. This suggests that distinct factors govern the expression of IGF-I and -II genes.     

Increased IGF-I and IGF-II mRNA and IGF-I peptide in fusing rat cranial sutures suggest evidence for a paracrine role of insulin-like growth factors in suture fusion

Premature cranial suture fusion, or craniosynostosis, can result in gross aberrations of craniofacial growth. The biology underlying cranial suture fusion remains poorly understood. Previous studies of the Sprague-Dawley rat posterior frontal suture, which fuses at between 12 and 20 days, have suggested that the regional dura mater beneath the cranial suture directs the overlying suture's fusion. To address the dura-suture paracrine signaling that results in osteogenic differentiation and suture fusion, the authors investigated the possible role of insulin-like growth factors (IGF) I and II. The authors studied the temporal and spatial patterns of the expression of IGF-I and IGF-II mRNA and IGF-I peptide and osteocalcin (bone morphogenetic protein-4) protein in fusing posterior frontal rat sutures, and they compared them with patent coronal (control) sutures. Ten Sprague-Dawley rats were studied at the following time points: 16, 18, and 20 days of gestation and 2, 5, 10, 15, 20, 30, 50, and 80 days after birth (n = 110). Posterior frontal and coronal (patent, control) sutures were analyzed for IGF-I and IGF-II mRNA expression by in situ hybridization by using 35S-labeled IGF-I and IGF-II antisense riboprobes. Levels of IGF-I and IGF-II mRNA were quantified by counting the number of autoradiograph signals per cell. IGF-I and osteocalcin immunoreactivity were identified by avidin-biotin peroxidase immunohistochemistry. IGF-I and IGF-II mRNA were expressed in dural cells beneath fusing sutures, and the relative mRNA abundance increased between 2 and 10 days before initiation of fusion. Subsequently, IGF-I and IGF-II mRNA were detected in the suture connective tissue cells at 15 and 20 days during the time of active fusion. In contrast, within large osteoblasts of the osteogenic front, the expression of IGF-I and IGF-II mRNA was minimal. However, IGF-I peptide and osteocalcin protein were intensely immunoreactive within these osteoblasts at 15 days (during the period of suture fusion). These data suggest that the dura-suture interaction may be signaled in a paracrine fashion by dura-derived growth factors, such as IGF-I and IGF-II. These peptides, in turn, stimulate nearby osteoblasts to produce bone-promoting growth factors, such as osteocalcin.     

Insulin-like growth factor I (IGF-I) in a growth-enhanced transgenic (GH-overexpressing) bony fish,the tilapia (<Emphasis Type="Italic">Oreochromis niloticus</Emphasis>): indication for a higher impact of autocrine/paracrine than of endocrine IGF-I

Several lines of growth hormone (GH)-overexpressing fish have been produced and analysed for growth and fertility parameters. However, only few data are available on the growth-promoting hormone insulin-like growth factor I (IGF-I) that mediates most effects of GH, and these are contradictory. Using quantitative real-time RT-PCR, radioimmunoassay, in situ hybridization, immunohistochemistry, and radiochromatography we investigated IGF-I and IGF binding proteins (IGFBPs) in an adult (17 months old) transgenic (GH-overexpressing) tilapia (Oreochromis niloticus). The transgenics showed an around 1.5-fold increase in length and an approximately 2.3-fold higher weight than the non-transgenics. Using radioimmunoassay, the serum IGF-I levels were lower (6.22 ± 0.75 ng/ml) in transgenic than in wild-type (15.01 ± 1.49 ng/ml) individuals (P = 0.0012). Radioimmunoassayable IGF-I in transgenic liver was 4.2-times higher than in wild-type (16.0 ± 2.21 vs. 3.83 ± 0.71 ng/g, P = 0.0017). No hepatocytes in wild-type but numerous hepatocytes in transgenic liver contained IGF-I-immunoreactivity. RT-PCR revealed a 1.4-times higher IGF-I mRNA expression in the liver of the transgenics (10.51 ± 0.82 vs. 7.3 ± 0.49 pg/μg total RNA, P = 0.0032). In correspondence, in situ hybridization showed more IGF-I mRNA containing hepatocytes in the transgenics. A twofold elevated IGF-I mRNA expression was determined in the skeletal muscle of transgenics (0.33 ± 0.02 vs. 0.16 ± 0.01 pg/μg total RNA, P < 0.0001). Both liver and serum of transgenics showed increased IGF-I binding. The increased IGFBP content in the liver may lead to retention of IGF-I, and/or the release of IGF-I into the circulation may be slower resulting in accumulation of IGF-I in the hepatocytes. Our results indicate that the enhanced growth of the transgenics likely is due to enhanced autocrine/paracrine action of IGF-I in extrahepatic sites, as shown here for skeletal muscle.     

IGF-I and IGF-II mRNA expression in slow and fast growing families of USDA103 channel catfish (Ictalurus punctatus)

The objective of this study was to examine insulin-like growth factor (IGF)-I and IGF-II mRNA levels in fast and slow growing families of catfish. Relative levels of IGF-I and IGF-II mRNA were determined by real-time PCR. Family A exhibited a specific growth rate (SGR) of 3.6 and was designated as fast growing, while family H exhibited a SGR of 3.1 and was designated as slow growing (P=0.017). Levels of IGF-II mRNA were 3.3-fold greater (P=0.006) in muscle for the fast growing family compared to the slow growing family. Levels of IGF-II mRNA were 1.8-fold greater (P=0.049) in liver for the fast growing family compared to the slow growing family. Levels of IGF-II mRNA from both fast and slow families were 12.2-fold greater (P<0.001) in muscle and 5.8-fold greater (P=0.021) in liver, respectively, compared to levels of IGF-I mRNA. Muscle and liver levels of IGF-I mRNA were similar between families. Elevated levels of IGF-II mRNA in muscle and liver compared to IGF-I mRNA, as well as differences in levels of IGF-II mRNA between fast and slow growing families of fish suggests a role of IGF-II in growth of channel catfish.     

Distinct methylation patterns in histone H3 at Lys-4 and Lys-9 correlate with up- & down-regulation of genes by ethanol in hepatocytes

Ethanol induced liver injury is associated with a global change in gene expression but its mechanisms are not known. We studied whether alcohol-induced gene expression is associated with post-translational methylations of histone H3. Primary culture of rat hepatocytes was treated with ethanol (50 or 100 mM) for 24 h and the status of methylation of H3 at lys 4 (H3dimeK4) or lys 9 (H3dimeK9) was monitored by Western blotting using antibodies to dimethylated histone H3 at lys 4 or lys 9. The cells exposed to ethanol showed strikingly opposing behaviors in methylation patterns; H3dimeK9 methylation was decreased whereas H3dimeK4 increased. Similar results were obtained in the interphase nuclei. Their binding on the metaphase chromosomes exhibits distinct site specific pattern of accumulation. Next, chromatin immunoprecipitation of the ethanol treated samples with antibodies for methylated lys 4 or lys 9 histone H3 followed by amplification of the immunoprecipitated DNA, was used to determine their association with the promoters of genes up- or downregulated by ethanol. Lys4 methylation was associated with ethanol upregulated genes (Adh, GST-yc2) whereas lys 9 methylation with downregulated genes (Lsdh, cytP4502c11) demonstrating a difference between these two methylations. These results suggest that exposure of hepatocytes to ethanol changes the expression of several susceptible genes which are associated with site specific modification of dimethylated forms of histone H3 amino termini at their regulatory regions.     

Longitudinal data for intrauterine levels of fetal IGF-I and IGF-II

OBJECTIVE: An increasing body of evidence supports a major role for the insulin-like growth factors (IGFs) in the control of human fetal growth. Individual data at various times of pregnancy suggest that IGF-I and IGF-II levels remain stable up to the 33rd week of pregnancy. Thereafter, both increase to reach values 2-3 times higher at term. In order to provide an accurate reflection of fetal IGFs in utero, we sampled fetal blood from the umbilical cord by cordocentesis. METHODS: We measured IGF-I and IGF-II in 12 fetuses longitudinally for up to 5 times between the 21st week of gestation and delivery. RESULTS: All patients showed a progressive increase in IGF-I and IGF-II levels. Data determined during different time intervals (before 29th, 29th to 32nd, after 32nd week) were compared and the main increase was found after the 32nd week. The median for IGF-I before the 29th week was 33.5 ng/ml (range 19-40.5) and increased to 41 ng/ml (32-59) between the 29th to 32nd and further to 54.1 ng/ml (range 17-70) thereafter. During the same time interval, the median for IGF-II increased from 217 ng/ml (86-326) to 349 ng/ml (227-467). In 7 patients, cord blood after delivery was available. For IGF-II a further increase was consistently found after birth (from 282 ng/ml (175-511) to 393 ng/ml (297-513)), whereas only 2 fetuses showed an increase in IGF-I. CONCLUSION: We conclude that in human fetuses, IGF-I and IGF-II levels increase longitudinally throughout pregnancy. Therefore, they may become important markers of healthy fetal development.     

Progestins induce down-regulation of insulin-like growth factor-I (IGF-I) receptors in human breast cancer cells: potential autocrine role of IGF-II.

Insulin-like growth factor-I (IGF-I) receptors are present in breast cancer cells and may play a role in breast cancer cell growth. We have studied the effect of progestins on IGF-I receptors in T47D human breast cancer cells. T47D cells constitutively express high levels of progesterone receptors and are a model for studying the regulation of cellular functions by progestins. Treatment of T47D cells with either progesterone or the synthetic progestin promegestone (R5020) decreased IGF-I receptor content by approximately 50%, as measured by Scatchard analysis and receptor biosynthesis studies. In contrast to progestins, estradiol, dexamethasone, and dihydrotestosterone did not influence IGF-I receptor content. No effect of R5020 was seen after 12 h of incubation, a near-maximal effect was seen after 24 h, and greatest effects were seen after 72 h. R5020 decreased IGF-I receptor mRNA abundance, indicating that progestins acted at the level of gene expression. However, progestins also increased the secretion of IGF-II, a ligand for the IGF-I receptor. In contrast to IGF-II, T47D cells did not express IGF-I. The addition of exogenous IGF-II to T47D cells down-regulated both IGF-I receptor binding and IGF-I receptor mRNA abundance. This study indicates, therefore, that progestins regulate IGF-I receptors in breast cancer cells and suggests that this regulation occurs via an autocrine pathway involving enhanced IGF-II secretion.     

Sequence of IGF-I, IGF-II, and HGF expression in regenerating skeletal muscle

Various cytokines are thought to play a role in muscle regeneration, however, the interaction and mechanisms of action of these cytokines remains largely unknown. In this study, we investigated the role of HGF, IGF-I, and IGF-II during myogenesis using the regeneration model of skeletal muscle as well as myoblast culture. RT-PCR analysis revealed that HGF and IGF-I expressions were markedly upregulated, in regenerating muscle. In contrast, there was no significant difference in IGF-II expression between normal and regenerating muscle. Immunohistochemical analysis demonstrated that HGF was expressed mostly by myocytes during the early stages of muscle regeneration. Additionally, HGF inhibited the formation of myotubes by myoblasts, but promoted cellular proliferation. Otherwise, IGF-I and IGF-II were expressed by myocytes through the early to middle stages of muscle regeneration. The addition of HGF to myoblast growing in vitro significantly increased the number of cells. These findings indicate that these three cytokines have pleiotropic effects in regenerating skeletal muscle.     

Sustainability of Nile tilapia net-cage culture in a reservoir in a semi-arid region

Among the tools used to measure sustainability in aquaculture, sets of indicators allow a holistic view of a system in its social, environmental, and economic dimensions. Approaches that align indicators with models such as the Drivers-Pressure-State-Impact-Response (DPSIR) framework can improve understanding of this sustainability. This study evaluated the sustainability of cage production systems for Nile tilapia in the Santa Cruz Reservoir, to determine whether a set of indicators used with the DPSIR conceptual model was effective to study the sustainability of the system. The 49 indicators used were calculated from information obtained from questionnaires and from monitoring the production system. Sustainability was also modeled and compared with hypothetical scenarios, with different fish stocking densities. The results indicated that the production system is economically feasible, generating profit and distributing income. However, the income generated benefits few people and is not fixed in the community. Environmentally speaking, the system is highly dependent on inputs, especially the nutrients nitrogen and phosphorus, and energy, as well as increasing sedimentation of nutrients in the reservoir. In the social dimension, the venture employs few workers. The modeling showed that the system is potentially sustainable, and that changes in stocking density decreased this sustainability. In summary, the system showed many sustainable features, whereas some others need to be modified to improve the general sustainability.     

Sequence of IGF-I, IGF-II, and HGF expression in regenerating skeletal muscle

Various cytokines are thought to play a role in muscle regeneration, however, the interaction and mechanisms of action of these cytokines remains largely unknown. In this study, we investigated the role of HGF, IGF-I, and IGF-II during myogenesis using the regeneration model of skeletal muscle as well as myoblast culture. RT-PCR analysis revealed that HGF and IGF-I expressions were markedly upregulated, in regenerating muscle. In contrast, there was no significant difference in IGF-II expression between normal and regenerating muscle. Immunohistochemical analysis demonstrated that HGF was expressed mostly by myocytes during the early stages of muscle regeneration. Additionally, HGF inhibited the formation of myotubes by myoblasts, but promoted cellular proliferation. Otherwise, IGF-I and IGF-II were expressed by myocytes through the early to middle stages of muscle regeneration. The addition of HGF to myoblast growing in vitro significantly increased the number of cells. These findings indicate that these three cytokines have pleiotropic effects in regenerating skeletal muscle.     

Growth and IGF-I response of juvenile Nile tilapia (Oreochromis niloticus) to changes in water temperature and dietary protein level

Water temperature and dietary protein level play an important role in influencing the growth and insulin-like growth factor I (IGF-I) in Nile tilapia juveniles. The combined effect of temperature (20–34 °C) and dietary protein level (25–50%) on the specific growth rate (SGR), feed efficiency (FE), serum IGF-I level and hepatic IGF-I mRNA level was examined under laboratory conditions by employing central composite design and response surface method. Results showed that the linear effects of temperature and dietary protein level on the SGR, FE, serum IGF-I and hepatic IGF-I mRNA level were significant (P<0.05); the quadratic effects of temperature and dietary protein level on the FE and serum IGF-I were significant (P<0.05). The interaction of temperature and dietary protein level on the FE, serum IGF-I and hepatic IGF-I mRNA level all proved significant (P<0.05). The optimal temperature/dietary protein level combination was determined, i.e., 29.9 °C/40.3%, at which the greatest SGR (2.748%/d) and FE (0.775) were simultaneously arrived. Both SGR and FE were linearly correlated with serum IGF-I or hepatic IGF-I mRNA level. These results suggested that optimum combination of temperature and dietary protein level would enhance tilapia growth efficiency and IGF-I would regulate growth and FE.     

Distribution of IGF-I mRNA and IGF-I binding sites in the rat kidney

Summary In the present study we have investigated the distribution of IGF-I mRNA and IGF-I binding sites in the rat kidney. The distribution of IGF-I mRNA was investigated using a simple and sensitive non-radioactive in situ hybridisation technique based on probe labelling with digoxigenin labelled-UTP followed by detection with conventional immunocytochemical techniques. IGF-I mRNA was found predominantly in medullary collecting ducts and sparsely in cortical collecting duct cells. In addition IGF-I mRNA was expressed in scattered proximal tubular cells in the cortex and in cells confined to the glomerular tuft. IGF-I binding sites were studied using radiolabelled IGF-I and conventional autoradiographical techniques on tissue sections. It was found that IGF-I binding sites were widely distributed throughout the entire kidney and that the specific binding was highest in the inner medulla. These findings add further complexity to the understanding of IGF-I production and action on renal structures.     

Toxic effects of chlorpyrifos on the growth,hematology, and different organs histopathology of Nile tilapia,Oreochromis niloticus

Chlorpyrifos is a widely applied insecticide that permeates on most waterways and affects aquatic organisms. The growth performances, hematological and histological impacts on Nile tilapia, Oreochromis niloticus following a 60 day of exposure to varying concentrations of chlorpyrifos 20 EC (T₁ 08 µgL^?1, T₂ 16 µgL^?1, and T₃ 32 µgL^?1) were compared to a control T_c 0 µgL^?1. The 96-hour LC50 of chlorpyrifos 20 EC was calculated as 46.80 μgL^?1. The water quality parameters were recorded regularly. The value of dissolved O₂ and NH₃ stayed rather steady, although temperature varied considerably. It was revealed that as chlorpyrifos levels go up, the percentage of weight gain (WG %), specific growth rates (SGR), and survival rate decreases. The control group T_c had the highest percentages of SGR weight (1.16 ± 0.58) and the T₃ group had the lowest percentages of SGR weight (0.25 ± 0.77). The hematological assessment showed significant differences of hemoglobin concentration, white blood cell counts and red blood cell numbers between chlorpyrifos treatment and control group (P < 0.05). Histological alterations in the liver, gills, and muscle tissues reported to be worse for T₃ as compared to others. There were no statistical differences in GSI, HSI values between control and treatment groups. The chlorpyrifos 20 EC was shown to be highly toxic to O. niloticus at sub-lethal dosages.     

Distribution of IGF-I mRNA and IGF-I binding sites in the rat kidney.

In the present study we have investigated the distribution of IGF-I mRNA and IGF-I binding sites in the rat kidney. The distribution of IGF-I mRNA was investigated using a simple and sensitive non-radioactive in situ hybridisation technique based on probe labelling with digoxigenin labelled-UTP followed by detection with conventional immunocytochemical techniques. IGF-I mRNA was found predominantly in medullary collecting ducts and sparsely in cortical collecting duct cells. In addition IGF-I mRNA was expressed in scattered proximal tubular cells in the cortex and in cells confined to the glomerular tuft. IGF-I binding sites were studied using radiolabelled IGF-I and conventional autoradiographical techniques on tissue sections. It was found that IGF-I binding sites were widely distributed throughout the entire kidney and that the specific binding was highest in the inner medulla. These findings add further complexity to the understanding of IGF-I production and action on renal structures.     

Effects of IGF-I, IGF-II, bFGF and PDGF on the initiation of mRNA translation in C2C12 myoblasts and differentiating myoblasts.

In order to study the mechanisms by which growth factors stimulate protein synthesis, C2C12 myogenic cells were treated with a variety of growth factors and the recruitment of free ribosomes to polysomes was quantified. All experiments were conducted on C2C12 myoblasts (24 h prior to induction of fusion) and differentiating myoblasts (24 h after induction of fusion). After the 2 h incubation, cells were rinsed with phosphate buffered saline and quickly frozen at -80 degrees C. Cell lysates were fractionated on 15-60% sucrose gradients by centrifugation at 200,000 x g for 1 h. Absorbance at 254 nm was recorded continuously across the gradient. The response to each of the four growth factors, IGF-I and-II, basic fibroblast growth factor (FGF), and platelet-derived growth factor was a decrease (P < 0.05) in monosome peak height and a increase (P < 0.05) in polysome percentage (P < 0.05). All responses were linear, except IGF-I, and the monosome peak height response to FGF which were quadratic (P < 0.05). None of the growth factors had a significant effect (P > 0.05) on RNA concentrations over the 2-h incubation. Protein content did not vary due to growth factor or level of treatment. This corroborates the hypothesis that the acute increase of protein synthesis exhibited by growth factor treated cells is due to an increase in the activity of existing ribosomes rather than an increase in ribosome synthesis. These results suggest that we can study the mechanisms regulating protein synthesis in muscle cells effectively by studying shifts in ribosomal activity. This method gave more consistent results than the H3-tyrosine incorporation and has the added benefit of not requiring the use of radioactivity. The strong correlation between monosome peak heights and percentage polysomes will allow researchers to measure total protein synthetic activity in a culture from the free or cytoplasmic fraction and to reserve the polysomes for other uses. The similarity of response among the various growth factors may indicate a common mechanism for increasing the initiation of protein synthesis.     

Identification and partial characterisation of a chitinase from Nile tilapia, Oreochromis niloticus

Measurement of chitinase activity in extracts from stomach, intestine, and serum of Nile tilapia with the artificial substrates 4-methylumbelliferil beta-D-N,N'-diacetylchitobioside and 4-methylumbelliferil beta-D-N,N'N"-triacetylchitotrioside (4MU[GlcNAc](2,3)) showed that an endochitinase was involved in the liberation of the fluorophore 4-methylumbelliferone (MU). Enzymes were isolated from tilapia serum by a combination of gel filtration, ion exchange, and reverse-phase chromatography. The molecular mass of the enzyme was estimated to be 75 kDa by SDS-PAGE, suggesting that the enzyme occurs as a monomer. The partially purified enzyme showed maximal activity at pH 7.0 when assayed with 4MU[GlcNAc](2) and lost its activity below pH 5.0 and above pH 8.0. The optimal pH of the purified enzyme toward the substrate 4MU[GlcNAc](3) was pH 9.0 and activity was lost below pH 8.0 and above pH 9.0. Our study has revealed the presence of a chitinolytic enzyme in the gastrointestinal tract and serum that may play a role in digestion and/or defense.     

Differential up- and down-regulation of type I and type II receptors for adrenocorticosteroid hormones in mouse brain

Adult female mice were adrenalectomized and ovariectomized and the concentration of Type I and Type II receptors in whole brain, kidney, and liver cytosol determined at various time thereafter by incubation with [3H]aldosterone (+ RU 26988 to prevent binding to Type II receptors) or [3H]dexamethasone, respectively. Type I receptor binding in brain was found to undergo a dramatic biphasic up-regulation, with levels six times that of intact levels by 24 h post-surgery and a doubling again by 4-8 days post-surgery. By 16 days, however, Type I specific binding had returned to intact levels. Similar, but less dramatic fluctuations were seen in kidney and liver, whereas much smaller fluctuations were seen for Type II receptors in all three tissues. In a follow-up study with Scatchard analyses we observed a similar transient up- and down-regulation in maximal binding for Type I, and to a lesser extent Type II receptors in all three tissues. As expected, the apparent binding affinity for both receptors increased after surgical removal of competing endogenous steroids. Radioimmunoassays revealed that plasma concentrations of corticosterone were reduced to near undetectable levels by 24 h post-surgery. A direct comparison of male and female mice revealed no sex-related differences in Type I receptor binding capacity fluctuations in brain cytosol after adrenalectomy-gonadectomy. Lastly, treatment with exogenous aldosterone or corticosterone was found to prevent adrenalectomy-gonadectomy-induced up-regulation of Type I and, to a lesser extent, Type II receptors in brain. Somewhat surprisingly, the potency of these two adrenocorticosteroids appeared to be very similar for both receptor types.     

Molecular characterization and immunohistochemical localization of tilapia piscidin 3 in response to Aeromonas hydrophila infection in Nile tilapia

The antimicrobial activity of tilapia piscidin 3 (TP3) was determined in vitro against a locally isolated Aeromonas hydrophila. A 388 bp fragment was amplified from the TP3 cDNA and sequenced. The coding sequence (CDS) of TP3 was estimated to be 231 bp codes for 76 amino acids long and stop codon. In silico analysis was performed to detect both the signal peptide and the prodomain cleavage sites to follow the amino acids number 22 and 70, respectively. Based on this, a peptide 23 amino acids long with a remarkably high computed antimicrobial probability was synthesized and used in the subsequent experiments. The antimicrobial activity of TP3 was determined with minimum inhibitory concentration (MIC) and minim um bactericidal concentration (MBC) methods. TP3 exhibited relatively weak antimicrobial activities against the tested bacteria. A challenge experiment was then performed in Nile tilapia with low and high doses of A. hydrophila, followed by timely recognition; after 3, 6, 24 h, and 7 days of the specific TP3 gene expression, immunohistochemical localization was also performed. Histopathological examination revealed provoked inflammatory responses and congestion in the same organs of TP3 expression. Immunohistochemical localization showed that A. hydrophila induced tilapia fish to express TP3 after 24 h within the gills, intestine, hepatopancreas, spleen, and posterior kidney. In quantitative real time (RT)‐polymerase chain reaction analysis, the high dose showed higher mRNA expression levels than the low dose, and its expression levels increased in the A. hydrophila‐infected fish. It was therefore concluded that TP3 plays an essential role in fish immunity.