1. Introduction
In the last decades mucosal immunology of more evolved vertebrates is a strongly explored field of research. Although a well functioning mucosal immune system in fish would be very beneficial to live in a pathogen rich aquatic environment, not many details are known about this system in this group of animals. This review will summarize the data known about the intestinal immune system of teleosts and when appropriate will compare it with the mammalian intestinal immune system. In mammals two sites can be distinguished with respect to the gut mucosal immune system: the induction sites (GALT structures like Peyer’s patches) and the effector sites: lamina propria (LP) and the intraepithelial lymphocyte (IEL) compartment [1,2].
Very essential in the follicle-associated epithelium above the induction sites are the M cells that can actively transport exogenous antigens to the underlying lymphoid tissue and finally can result in IgA class switching [3] and the secretion of high amounts of dimeric IgA at the effector sites [1,2]. The secreted IgA is subsequently bound by the polymeric Ig receptor (pIgR) and transcytosed to the intestinal lumen or to the bile in the liver. The extracellular part of the receptor is then cleaved off and secreted as secretory component (SC) together with the IgA (and IgM, both having a joining J-chain for pIgR binding) at the mucus site. In the last decade it has been shown that dendritic cells (DC) can also take up antigen directly from the lumen and this new induction alternative is now well established [1,2]. Peyer’s patches, M cells, IgA and J-chain are not reported in teleost fish. In addition, lymph nodes are absent and there is hardly evidence for specific homing of mucosal lymphocytes.
This nearly excludes the existence of a common mucosal immune system, but does not exclude a local mucosal defense system. The first indications that fish could have local and/or mucosal responses came from the detection of specific antibodies in mucosal secretions after intestinal [4e8] or immersion [9e11] immunizations of a variety of fish species, while they were not or hardly detectable after systemic immunizations. This differential appearance of specific antibodies in combination with the detection of specific antibody-producing cells at mucosal sites after intestinal [12] or immersion vaccination [11,13] inspired many scientists to study mucosal structures in fish.
2. Food and antigen uptake in the gut of fish The digestive tract of fish starts as a straight tube without the presence of a clear stomach. During larval development the stomach develops in 85% of the bony fish species [14], while the other 15% of the fish species (i.e. Cyprinidae) will never develop a stomach and lack a region with low pH and predigestion. In these stomachless species the bile and pancreatic ducts immediately enter the gut just posterior of the esophagus, and the widened first straight part of the gut, the intestinal bulb, seems to have a storage function. In nearly all species investigated the intestine can be subdivided in three segments based on the microscopical anatomy of their mucosa, especially their enterocytes [15]:
1) The first segment (60e75% of the total gut length, dependent on the species), of which the enterocytes can be considered as absorptive cells. In functional studies on cyprinids [16e18] the uptake of food proteins appears to take place in the anterior 50% of the intestine.
2) The second segment (15e30% of the gut length), in which the enterocytes are characterized by large supranuclear vacuoles, irregular microvilli zone and high pinocytotic activity at the apical part. In this part a strong uptake of macromolecules, like horseradish peroxidase (HRP) and ferritin could be demonstrated in both stomachless [15,18e21] and stomach-containing species [22e25].
3) The third segment (5e15% of the gut length), is less well investigated but the enterocytes are suggested to have an osmoregulatory function and because of the minor microvilli they do not seem to have a nutritional function [23,26]. Many recent papers neglect the third segment and mention endgut or hindgut when they study the second gut segment.
It has to be mentioned that certain fish species (i.e. Gadidae) have a rectum in their hindgut which is mostly separated from the intestine by valves; a clear second segment is till now not found in cod [27]. The strong uptake capacity of the second segment is apparently not of nutritional value and drew the attention of immunologists since the eighties of the last century, when it was shown that antigens could be transported towards the local as well as systemic immune system [6,7,12,28]. Crucial observations with respect to the uptake and processing of antigens were done in carp [21] showing different transport routes for receptor mediated uptake (solid phase uptake of HRP) and fluid phase uptake (ferritin). HRP was immediately sorted out in the endolysosomal compartment and sent to the intercellular spaces, while ferritin followed the route to the large supranuclear vacuoles (SNV: endosomes) and never could be observed in the intercellular spaces. The same route is followed by soluble factors (mainly LPS) of a bacterin [6,7,12].
The macromolecules taken up by the fluid phase route are finally transferred to phagocytosing cells present in and under the intestinal epithelium, which in carp are resident and mobile macrophages. Although the mechanism of this transfer is never clarified well, the most logical way is the transfer of the SNV from epithelial cells to phagocytes, like described for transfer of melanin to human skin epithelial cells and apical parts of rodlets to pagocytosing chromatophores in the mammalian retina. The receptor mediated uptake of HRP results in a very fast peak of HRP in the blood of carp [29] and trout [30] 30 min after oral administration, while the appearance of ferritin or LPS inside macrophages is more a process of 4e8 h [7,21]. The HRP receptor is presently characterized, but the biological reason for this receptor mediated uptake of enzymes still not [31]. In contrast to mammalian M cells enterocytes of the second segment have functional endolysosomal organels and are not well able to phagocytose inactivated bacteria [32]. On the other hand phagocytosis cannot be completely excluded and probably needs certain inducing signals. For instance poly D,L-lactide-co-glycolic acid (PLGA) microparticles seem to pass the intestinal barrier, enter the body and finally can be found back in the systemic lymphoid system [33e35]. None of these papers showed the mechanism of penetration of the PLGA microparticles.
3. Immune competent cells in the intestinal mucosa
The presence of immune competent cells have been well studied in the gut of common carp, European sea bass and Atlantic salmon. It was already known for many years that leucocytes are abundantly present in lamina propria and intestinal epithelium [36e39]. However, the lack of suitable antibodies in fish hampered the distinction of subpopulations within the gut associated lymphoid system (GALT).
In teleosts the level of GALT organization is lower than in mammals, but they have a more diffusely organized immune system in their gut, containing many lymphoid cells, macrophages, eosinophilic and neutrophilic granulocytes. As far as we know dendritic cells (DC) are not well established in the mucosae of fish.
3.1. Ig-positive B cells
In sea bass [40,41], salmonids (Sunyer, this issue) and cyprinids [39] Ig+ cells are described in the intestinal mucosa, but the numbers described differ strongly between species. In carp the highest number of Ig + cells have been described, mainly present in the intestinal mucosa and consisting of B cells and plasma cells, the latter mostly smaller than found in systemic lymphoid organs [39].
The strong difference in the number of B cells found in this species may be due to the reactivity and/or affinity of the monoclonal antibodies used. As will be discussed later (Section 5) mucosal Ig may differ from systemic Ig and the antibodies made against serum IgM may not or hardly react with mucus Ig. However, the monoclonal antibody against carp serum IgM (WCI12) appears to crossreact with mucosal Ig/B cells [42] and hence it can be concluded from this species that abundant numbers of B cells are present in the lamina propria.
In isolated gut cell suspensions of carp the proportion of isolated B cells is low (5e10%), because cells could not be released easily from the connective tissue [39]. Another possibility for low numbers of B cells in the gut of certain species may be explained by a preference of mucosal plasma cells in the liver as demonstrated for an Antarctic species [43].
3.2. Ig-negative T cells
Monoclonal antibodies (mAb) specific for T cells or T cell subpopulations were used to detect these T cells in sea bass [40,41,44] and carp [45]. In both species T cells were abundantly present in the lamina propria and epithelium.
The intestinal epithelial lymphocytes (IEL) of a variety of teleost species were found to express T cell genes as CD3ε , ZAP70, TCR, CD4 and CD8 [46e49]. An anti-human CD3ε antibody raised against a well conserved recombinant peptide appeared to cross-react with a variety of fish species and demonstrated abundant T cells in epithelium and lamina propria of carp [50] and Atlantic salmon [51]. In sea bass, a significant number of DLT15+ cells does not express TCRβ and this may be the first indication for the presence of γδ T cells [52]. Preliminary results in sea bass and carp have shown the gene expression of γδTCR.
3.3. Macrophages
Not many studies are dedicated to the intestinal macrophages of fish, probably again because of the lack of suitable markers. Although gut macrophages are morphologically described in a variety of teleost species [6,27,28,37] there is only functional evidence in common carp. Cyprinid fish have many macrophagelike cells in and under the intestinal epithelium and their number increases twofold after anal intubation with antigens [21,36,53]. This invasion is mainly due to small mobile macrophages, that remigrate and can be detected 1e2 days later loaded with intestinal antigen in other lymphoid organs. In the second gut segment of most cyprinids studied, many large resident intraepithelial macrophages can be found. These cells are strongly Ig+, because of their Ig-binding capacity [39,54]. These macrophages can be loaded with antigens from the lumen; after degradation antigenic determinants can subsequently be presented at their surface [6,53]. All these characteristics and the abundant presence of B and T cells again strongly indicate that local immune responses can be evoked. It has to be stated here that not each teleost species shows similar numbers and size of intestinal macrophages or have the same Igbinding capacity.
3.4. Granulocytes
In general granulocytes in fish can be considered as innate cells. Although intermediate cells are described, sometimes as separate subpopulations, two main types can be considered: neutrophils with the smallest granules having a typical rod-like electron dense structure (EM) and eosinophils with larger and electron dense granules, strongly resembling mammalian mast cells [38,55].
These eosinophils are mentioned basophils [38] in common carp, because they did not stain with eosin after standard fixation techniques. Reite [56] for the first time decided to call them mast cells/eosinophilic granule cells but it must be stated that fish do not have IgE and with the exception of one perciform fish [57] histamine is not demonstrated in these cells. On the other hand they can release tryptase [58], antimicrobial peptides as lysozyme [55,59], piscidin [60] and pleurocidin [61], they are abundant in mucosae, and they react strongly upon inflammation via migration and granule release [55].
The inflammation can be induced by a variety of factors: feed [50,55], infectious agents and chronic stress [62]. In fish as well as mammals neutrophils are considered as innate phagocytozing leucocytes abundantly present in the circulation and quickly recruited from the blood to sites of inflammation [63e66].
“Healthy” gut of fish in general contains less neutrophils than eosinophils/basophils, but they easier penetrate the epithelium [38] and their number strongly increase under danger or stress conditions [49]. For salmonids [66] and carp [10,67] neutrophil-specific mAbs are available to study the behaviour of neutrophils, but unfortunately it is hardly or not applied for teleost gut with the exception of the recent thesis of Sundh [62]; salmon neutrophils as well as eosinophils showed a strong invasive and activation reaction upon stress and infection and neutrophils are frequently penetrating the intestinal epithelium.
4. Mucosal immune responses and oral tolerance
As mentioned earlier mucosal immunizations result in detectable specific antibodies in mucosal secretions while these antibodies were hardy or not detectable after systemic immunizations [4e10]. These data were supported by the detection of specific antibody secreting plasma cells (ELISPOT) in gut and/or gills after mucosal immunization of carp [12] and sea bass [13] and their absence or scarcity in head kidney, spleen or blood. These results strongly indicate the induction of local responses, but also that differences have to exist between serum and mucosal Ig. In some of the mentioned studies [8,10] it is obvious that detectable antibodies in skin mucus appear rather late compared to the appearance of serum antibodies after parenteral immunization.
This might be due to a serious lag phase caused by the transport of bound specific antibodies in the multi-layered skin epithelium, because the studies on plasma cell detection did show comparable peak days after mucosal and systemic immunizations [12,13]. In carp, oral immunization with (alginate) encapsulated Vibrio anguillarum antigens resulted in high numbers of specific plasma cells in gills [12], which may be an indication for a common mucosal immune reaction. However, still unpublished results in carp and recent data in trout [49] did not reveal mucosa specific homing of IEL. Because IEL predominantly consist of putative T cells, tissue specific homing cannot be excluded yet for teleost B cells. Recently it has been suggested that “B-cells need a proper house, whereas T cells are happy in a cave” [68], indicating a higher dependency of B cells on secondary lymphoid tissue”. Whether the teleost gut is a “proper house” for B cells still has to be proven. Another phenomenon characteristic for intestinal immunology is oral tolerance or suppression of immune responses after (repeated) oral immunization.
This phenomenon has been shown as decreased specific antibody production against bacterial and protein antigens in rainbow trout [69], salmon [70,71] and common carp [7,72]. Only in common carp it has been shown that oral administration of proteins can result in a genetic dependent suppression of antibody responses after systemic administration of the same antigen [72], which was comparable with oral tolerance studies in mammals. More recently, tolerance induction was also described for cytotoxic responses when animals were repeatedly immunized (anally or orally) with allogenic target cells (EPC) [73,74]. It is interesting that this suppression of cytotoxicity caused by repeated intestinal administration recovered when the same antigen was administered systemically [74]. None of the studies on oral tolerance paid attention to the mechanism behind this phenomenon. It is also not clear how oral or anal intubated whole allogeneic target cells can pass the intestinal barrier and result in specific cytotoxic cells in blood.
5. Mucosal immunoglobulin
The distinct antibody secretion in mucus can only be explained by differences in mucosal and systemic immunoglobulin (Ig). But in contrast to higher vertebrates IgA is not detectable in fish. Biochemical analysis of cutaneous mucus of a variety of teleost species revealed only IgM-like molecules, merely in tetrameric but also in dimeric and monomeric forms [75e80]. At present minor data are available on the presence of Ig in gut mucus, probably due to the more difficult access but also because the degradation of IgM in the strong proteolytic environment [79]. Next to IgM, IgD is a commonly found immunoglobulin in teleost [81], but till now this isotype is never correlated to mucosal immunity. Although, in sheepshead [82,83] and carp [42] differences were observed between serum and mucus Ig, molecular evidence was never found for different IgM subtypes. In carp, a mAb could be selected reacting with the H chain of mucus and not of serum IgM. This mAb could be used to detect specific plasma cells in gut and gills after mucosal immunization and was immunoreactive with bile ducts and capillaries in liver and with skin epithelial cells, just the sites were mucosal Ig can be expected [42].
In sheepshead (Archosargus probatocephalus), a noncovalent dimer of covalent dimeric subunits has been described associated to a 95 kDa secretory componentlike protein [82,83], but this was not found in bile of the same species and also never observed in other species. In teleost fish a J-chain is never found while this IgA or IgM-joining protein is present in all vertebrates, including cartilaginous fish [83]. Nevertheless, it can be concluded that teleosts have a mucosal immune system because intravenous administration of radiolabeled Ig never reached the mucosal secretions [84] and therefore the Ig in mucosal secretions must be the result of local synthesis [42,76,84]. Till now liver and bile are poorly studied, but the presence of Ig in bile and the accumulation of plasma cells within the liver portal tracts of an Antarctic teleost fish, Trematomus bernacchii [43], indicate that liver, like in mammals, may be an important secretory side for mucosal Ig. In 2005 a new previously unknown Ig isotype was described, called IgZ in zebrafish [85,86] and carp [87], IgT in trout [88] and IgH in fugu [89]. Recently IgT in trout is claimed to be specialized in mucosal immunity and IgT responses to a gut parasite are restricted to the intestine [90]. In common carp two IgZ subclasses are described (IgZ1 and IgZ2) of which IgZ2 is a two constant domain chimera with m1 and z4 domains. IgZ1 is found more abundant in systemic organs while IgZ2 is preferentially expressed in gills and gut [91]. Indeed, IgT or IgZ may be an important ‘last flag unfurled” [86], but on the other hand not every teleost species seems to have the IgT- or IgZ-like sequences, like channel catfish [92], while this species showed mucosal responses [76].
6. The polymeric Ig receptor (pIgR)
The pIgR is an essential component for mucosal immunity in mammals. It is expressed by mucosal epithelia and hepatocytes, can bind IgA and IgM, transcytose these Igs to the luminal sides (or bile) and finally the extracellular part of the receptor is cleaved off and co-secreted with the IgA or IgM as a protecting secretory component [1,2]. In mammals it contains five Ig-like domains (ILDs), but in birds [93] and amphibians [94] only four ILDs are found. Recently, the pIgR amino acid sequence of four teleosts is published (fugu [95]; carp [96]; orange-spotted grouper [97]; rainbow trout [90]) and of two others (zebrafish and Atlantic salmon) it is reported on GenBank. The teleost pIgR only consists of two ILDs, which corresponds to the ILD1 and ILD5 of mammals [95e97].
It is obvious that the CDRs on ILD1 necessary for IgA binding are all three absent in teleosts. However, IgM binding studies showed that this small molecular weight pIgR can bind teleost IgM and cells strongly expressing the pIgR: skin epithelial cells, enterocytes and hepatocytes. The binding of pIgR to IgM is in mammals associated with the presence of a J-chain. Therefore it is postulated [95,97] that the J-chain has to be present in teleost as it is in cartilaginous fish. However, it is also suggested that the mammalian J-chain does not interact directly with the pIgR, but merely creates the necessary conformational site for binding with the pIgR [83]. On the other hand the J-chain of mammals, birds and amphibians are all able to polymerize human IgA while the J-chain of nurse shark is not, probably because the highly conserved J-chain region for pIgR interaction in tetrapods is lacking in cartilaginous fish [94]. Therefore, it is still questionable how efficient pIgR and IgM interaction occurs in teleosts without the presence of a J-chain. Recently, the interaction of pIgR and IgT has been evidenced in trout [90]. Another point of debate is whether the small molecular weight pIgR (35 kDa) will be cleaved off after transcytosis as a secretory component (SC) and whether it will be different between polarized epithelial cells or hepatocytes and non-polarized skin epithelial cells. With the exception of the sheepshead [82] an SC is never described in isolated mucosal IgM of other species. Moreover, the molecular weight (95 kDa) of this SC is much too large to be related to the teleost pIgR. The skin epithelia can in theory be well protected when Ig is bound to the surface of these cells. The transport of this bound Ig could well explain the earlier mentioned lag phase observed in the cutaneous antibody responses after mucosal immunization.
7. Intestinal T cells and their function
Mucosal T lymphocytes represent the major leucocyte population within the teleost gut but only limited data are available on their functional relevance [40,45,49]. Rainbow trout αβ IELs show clear changes in the TCRβ repertoire upon a systemic rhabdovirus infection comparable to spleen and pronephros T cells [49]. Recent data obtained in sea bass (Dicentrarchus labrax) provided further insight into their local defense roles [46]. Gene transcription studies evidenced the wide occurrence of T cells throughout the intestine and TCR- β and CD8- α transcripts exceeding CD4 ones that indicated the predominant occurrence of CD8- α expressing T cells. In particular, CD8- α transcripts significantly increased in the posterior intestinal segment. These data agree with previous findings about a significant increase of mucosal mAb DLT15+ T cells along the sea bass intestine towards the anus [40,41,52] that largely depends on the rise of DLT15+ IEL in the posterior segment [98]. This reflects a regional immune specialization as previously reported in other teleost species [7,99]
9. Oral vaccination strategies
Fish vaccination has a long history [144], but only in the last 20 years vaccination against bacterial and more recently viral diseases are established as successful prophylactic treatments [145]. Vaccination of fish by injection or immersion is now routinely used and some oral vaccines are recently commercialized. This contribution will be limited to oral vaccination as the most appealing method, because it is easy to administer without stress induction and suitable for mass vaccination at all ages [146]. It is worthwhile to state that the first oral vaccination paper was published by Duff in 1942, showing that trout could be orally vaccinated against Bacterium salmonicida [147]; prolonged feeding (64e70 days) with a B. salmonicida containing feed resulted in 25% mortality after an immersion challenge against 75% in the control animals. In the eighties many oral vaccination studies are reported, but most of them did not reach these protection values, if they reached protection at all. Johnson and Amend (1983) showed very good protection after anal intubation with V. anguillarum and Yersinia ruckeri bacterins compared to oral intubation or immersion [148].
Therefore, it was concluded that oral vaccines have to be protected against degradation and when possible in a cost-effective way. The first approaches in this direction were based on bio-encapsulation of V. anguillarum bacterins in juvenile ayu [149] gilthead seabream and carp [150], using plankton as water flea, rotifers or brine shrimp nauplii as antigen-delivery vehicle. The results also showed that this approach has to be applied on the right age, because too young animals still developed a kind of “neonatal” tolerance (immune-suppression). Other attempts to protect antigens for degradation in the gut are the use of (Eudragit) coated bacteria [151], microbial biofilms [152,153] and recently bacterial ghosts [154]. A variety of microparticles have been tested and most studies report a better delivery in the endgut, elevated antibody production and/or higher protection using alginate microparticles [12,155,156], liposomes [157,158], poly D,L-lactide-co-glycolic acid (PLGA) microparticles [33e35,159,160] and chitosan microparticles [161,162]. Alginate particles adapted in their formulation to the characteristics of the digestive tract appeared to be promising in carp and trout [12] and evoked mucosal as well as systemic immune responses.
PLGA and chitosan particles are claimed to be biodegradable mucoadhesive microspheres, both very suitable for encapsulation of pDNA vaccines [35,162]. Although microencapsulation has shown its potency for oral vaccination of fish there are only a few oral vaccines on the market, which may be due to the stability of the vaccine in the feed and the limited protection obtained [163] but probably more to the cost effectiveness of these vaccines. From this view the development of “edible vaccines” may be promising for aquaculture. For instance viral trimeric G proteins can be produced in potato tubers (or other consumption plants) and subsequently used for the preparation of fish feed. When LTB (the pentameric non-toxic variant of cholera toxin) is combined with the G protein a very stable LTB-G protein construct is produced by the plants and it finally result in a much better uptake in the second segment and immunological memory formation [164,165]. At present the main problem to be solved in this approach is the elevation of the amount of transgenic proteins produced by the plants, to avoid an extra concentration step. In fact LTB can be considered as a mucosal adjuvant. In general, intestinal vaccination is hampered by the lack of suitable mucosal adjuvants and also in fish more research is needed to obtain optimal effects with minor amounts of antigens.
10. Gut microbiota and probiotic immune-stimulation
The intestinal micro-ecology in fish is a relatively less-studied aspect, but over the last years an impressive amount of data has accumulated on the application of probiotics [166]. The microbial community in the gut of fish is strongly influenced by the environment and within this complex ecosystem the microbes compete for space and nutrients for their survival. Bacteria (107 e108 per gm) are the dominant microorganisms in fish intestine of which Proteobacteria are the main phylogenetic type [167e170], while in mouse and humans 99% of the intestinal bacteria are of the Firmicutes and Bacteroidetes type [171]. Resident microbes, occasional pathogens, as well as those artificially delivered microbes (probiotics) can have a significant role in intestinal immune responses. They co-exist in a dynamic equilibrium within the gut through continued and active signaling [172].
This symbiotic consortium of microbes starts to develop with the first feeding in the larval phase and possibly reaches a stable composition at the juvenile stage following autogenic and allogenic succession [173,174]. The changes in the composition of the microbiota are very influential during the early stages of fish when intestinal colonization and development of the immune system is taking place. For instance the absence of microbiota during the developmental stages of zebrafish led to arrested differentiation and altered function [175]. As in mammals, the interactions between host and commensal bacteria and the benefits derived from these processes are still poorly understood. Anyhow, the innate immune system protects the host by maintaining the integrity of the intestinal barrier, using pathogen recognition receptors (like TLR). For instance, the generic microbial molecule LPS could serve as ligand for TLR [176,177] and regulate alkaline phosphatase, a marker of enterocyte maturation and lipid absorption [175]. The beneficial microbial consortium uses multiple independent signals to interact with the host and co-exist with pathogenic bacteria. It has been suggested that an imbalance of the microbiota, favouring the less-traditional varieties (termed “dysbiosis”) could potentially lead to immune-related disadvantages to the host, including the lack of balance between regulatory and effector T cells and the increased differentiation of inflammatory effector immune cells [174].
Microarray analysis of gnotobiotic zebrafish [135] receiving first intestinal bacteria showed transcriptional changes similar to mice and hence pointing to an evolutionarily conserved role of microbiota in vertebrate development. The innate immune genes involved were serum amyloid A1 (Saa1), C-reactive protein (Crp), complement component 3 (C3), angiogenin 4 (Ang4, a microbiota-regulated antibiotic protein) suppressor of cytokine signaling 3 (Socs3), myeloperoxidase (Mpo; a granulocyte-specific marker) and the oxidative stress response gene glutathione peroxidase 2 (Gpx2). In zebrafish only the microbiota of fish (dominated by Aeromonas and Pseudomonas spp., besides Vibrio and Lactococcus) and not that transplanted from mouse displayed an increased expression of innate immune response biomarkers e Saa, complement factor b (Cf) and Mpo [178]. Further Saa and Cf were responsive to a large number of bacterial strains tested while the granulocyte marker Mpo was specifically expressed with Pseudomonas aeruginosa. Although not all reports on this subject in fish are mentioned here, it is clear that host-microbial commensalism as well as the etiology of pathogenic states in the intestine of fish needs much more attention, especially as the application of probiotics seem to be a powerful tool for disease resistance in fish. Probiotics, currently defined as “live microorganisms which when administered in adequate amounts confer a health benefit on the host” [179], are regarded as an alternative viable therapy and have become an integral part of the aquaculture practices for improving growth and disease resistance. Effects were documented in protection against a variety of bacterial diseases but recently even against protozoan and viral infections: Ich in Oncorhynchus mykiss [180] and iridovirus in Epinephelus coioides [181].
Protection against pathogens has been claimed through competitive exclusion for adhesion sites, production of organic acids, hydrogen peroxide and other compounds such as antimicrobials, siderophores and lysozyme and modulation of fish physiology and immunity [166,182,183]. However, the mechanisms through which probiotics effectively modulate the immune system are not well elucidated and the effects found on local gut immunity are even more sparse [183e185]. Viable probiotics (Lactobacillus rhamnosus) seem to be better immune stimulators than killed bacteria in rainbow trout [186] probably because of their ability to colonize the host gut. Recent studies in gilthead seabream [187,188] and European sea bass [189] on early administration (during gut metamorphosis) of autochthonous bacterial strains showed profound effects on intestinal microflora, gut physiology and immunity of fish larvae. In seabream, a multispecies probiotic formulation (autochthonous Lactobacillus fructivorans þ Lactobacillus plantarum from human feces) raised the number of intestinal IgMþ cells and acidophilic granulocytes [190]; this formulation was significantly more effective (in terms of fry survival and improved stress response) when administered early. However, juvenile seabream fed diets supplemented with inactivated Lactobacillus delbreckii ssp. lactis and Bacillus subtilis (individually or combined) showed also higher total serum IgM and numbers of gut IgMþ cells and acidophilic granulocytes than untreated controls [191]. In sea bass, the autochthonous bacterium Lactobacillus delbrueckii delbrueckii raised the number of intestinal T cells (with increased total body TcR-b transcripts) and acidophilic granulocytes, concomitant to lower transcription of key inflammatory genes (IL-1b, Cox-2, IL-10, TGF-b). Despite this down-regulation of cytokine expression increased survival, increased weight gain, decreased cortisol levels and improved stress responses were monitored [192]. In Atlantic cod autochthonous intestinal bacteria also appeared to be potential probionts [193] and proteomic analysis of probiotic stimulated cod indicated a lower level of immune stimulation but an up-regulation of proteins involved in growth and development [194]. It will be clear that the data given on probiotic stimulation are fragmentary and probably bacterial strain or fish species dependent and hence much more research is necessary to explain the protection reported after feeding probiotics.