Mucosal immunoglobulins of teleost fish: a decade of advances

Abstract

Immunoglobulins (Igs) are complex glycoproteins that play critical functions in innate and adaptive immunity of all jawed vertebrates. Given the unique characteristics of mucosal barriers, secretory Igs (sIgs) have specialized to maintain homeostasis and keep pathogens at bay at mucosal tissues from fish to mammals. In teleost fish, the three main IgH isotypes, IgM, IgD and IgT/Z can be found in different proportions at the mucosal secretions of the skin, gills, gut, nasal, buccal, and pharyngeal mucosae. Similar to the role of mammalian IgA, IgT plays a predominant role in fish mucosal immunity. Recent studies in IgT have illuminated the primordial role of sIgs in both microbiota homeostasis and pathogen control at mucosal sites. Ten years ago, IgT was discovered to be an immunoglobulin class specialized in mucosal immunity. Aiming at this 10-year anniversary, the goal of this review is to summarize the current status of the field of fish Igs since that discovery, while identifying knowledge gaps and future avenues that will move the field forward in both basic and applied science areas.

1. Introduction: the importance of understanding the biology of mucosal Igs in fish biology

Fish are covered by mucosal secretions that protect the living cells of all epithelial barriers from external aggressions. Many studies have highlighted the importance of mucus in fish physiology and immunological defense (Dash et al., 2018; Peatman et al., 2015). Early studies suggested that fish mucus contains immunoglobulins (Igs) that are produced locally rather than representing transudates from plasma (Rombout et al., 1993; Zhang et al., 2010). Further, these pioneer studies revealed that the composition of fish mucus changes in response to infection or vaccination (Salinas et al., 2011). In early studies, the presence of specific mucosal Igs were already detected in the fish skin mucus following vaccination with bacterins, thus already predicting that specific mucosal Igs were an important aspect of the fish immune response. Studies performed over the past decade have revealed that the mucosal immune system (MIS) of fish operates under the guidance of primordially conserved principles shared also by the MIS of tetrapod species. Importantly, it has become apparent that this arm of the immune system both in tetrapods and fish are clearly different from their systemic counterparts in a number of key points: First, mucosal barriers are not exclusively devoted to immunological functions and therefore they comprise tissues with high complexity and cellular heterogeneity. Second, mucosal barriers are colonized by microbial communities that establish unique relationships with the host mucosal immune system. Third, the mucosal immune system, despite of the high exposure to an antigen-rich environment, is able to remain in a relatively tolerogenic state, avoiding excessive immune responses and maintaining tissue homeostasis. The differences between the mucosal and the systemic immune systems are also reflected in the unique adaptations found in secretory Igs (sIgs). Once secreted by plasma cells found in barrier tissues, sIgs have to perform their functions within mixtures of mucins, antimicrobial molecules, lectins, proteases, and commensals and pathogens, among others. These structural and functional differences between mucosal and systemic Igs continue to be unveiled. These adaptations confer mucosal sIgs with unique properties that allow them to function in the external environment. Perhaps the most important adaptation stems from the presence of the microbiome, which at least in mammals, educates mucosal B-cells and shapes the sIg repertoire from birth (Li et al., 2020; Zhao and Elson, 2018).

Given that mucosal barriers are at the interface between the environment and the fish host, the MIS is the target of vaccines and other immune interventions in aquaculture. Understanding how sIgs function under homeostatic conditions and how they respond to antigenic and environmental perturbations is critical not only for our understanding of fish health in the wild and in captivity, but also for building up the necessary knowledge required for the rationale design of mucosal vaccines, the development of probiotics, immunostimulants and other immune interventions. Such interventions should enhance both specific and polyreactive sIg production while not damaging homeostasis between sIgs and microbiota and tissue integrity.

Ten years ago, IgT was discovered to be an immunoglobulin specialized in mucosal immunity. This represented an important discovery in the field as the role of that antibody class was unclear at that point, and more importantly, specialized immunoglobulins had only been described in tetrapod species. Aiming at this 10-year anniversary, the goal of this review is to provide a current view of the field of mucosal sIgs in teleosts since that discovery. With that goal in mind, here we present our perspective on knowledge gaps and future avenues that will move the field forward in both basic and applied science.

2. Where are sIgs found in fish?

Teleost fish are excellent models for the study of mucosal immunity because every barrier that separate the host from the environment is a living mucosal barrier. Thus, sIgs in fish are found in every mucosal barrier that has been investigated thus far. Traditionally, teleost mucosal immune responses have been mostly investigated in the skin, gut and gills. Thus, the skin associated lymphoid tissue (SALT), gut-associated lymphoid tissue (GALT) and gill-associated lymphoid tissue (GIALT) are the best described MALTs in teleosts (Salinas et al, 2011; Salinas 2015). Vertebrate MALTs include organized MALT (O-MALT) and diffuse MALT (D-MALT) (reviewed by Salinas, 2011, Salinas and Miller, 2020). D-MALT are found in all vertebrates as scattered lymphoid and myeloid cells that reside within all mucosal epithelia. The evolutionary origins of bona fide D-MALT are still a matter of debate, but recent studies suggest that lymphocytic organization at mucosal barriers may have evolved independently in different vertebrate lineages (Heimroth et al, 2020). While higher vertebrates contain O-MALT where sIgA responses are induced (i.e, Peyer’s patches, mesenteric lymph nodes), such lymphoid structures appear to be missing in teleosts, and thus, the question remains as to where and how are adaptive immune responses induced and mature in the different teleost MALTs.

Recently, the list of fish MALTs has been considerably expanded, with new studies reporting the presence of a nasopharynx-associated lymphoid tissue (NALT) as well as a buccal- and pharyngeal-associated lymphoid tissues (Yu et al., 2019). As a result, sIgs can be detected in the skin, gut, gills, nasal, buccal and pharyngeal mucus, at least in rainbow trout (Oncorhynchus mykiss) (Kong et al., 2019; Tacchi et al., 2014; Xu et al., 2013b; Xu et al., 2016; Yu et al., 2019; Zhang et al., 2010). Further, a recent study identified a bursa of Fabricius-like structure in cloaca of Atlantic salmon (Salmon salar) which contains IgM⁺ and IgT⁺ B-cells and therefore, the mucus surrounding this structure should also contain sIgs, and thus it might represent an additional teleost MALT (Loken et al., 2020). Whether this structure is found in other teleost species remains to be investigated. In summary, MALTs have been described in substantial detail in salmonids thus illuminating a large and complex network of lymphoid tissues at all mucosal barriers. However, many aspects of the anatomy and mechanisms of action remain to be described in these MALTs, including the mechanisms of antigen uptake, the localization and mechanisms of activation of dendritic cells and other antigen presenting cells (APCs), the distribution of CD4⁺ T lymphocytes and their subsets and the interactions between APCs, CD4⁺ T-cells with antigen and B-cells. Moreover, the anatomy of MALTs in other teleost species is not as well characterized as that of salmonids and therefore further studies are needed to ascertain whether this network is conserved in all teleosts.

Fish mucosal secretions contain the three major IgH classes found in teleost, IgM, IgD and IgT/Z (Danilova et al., 2005; Hansen et al., 2005; Wilson et al., 1997). The absolute protein concentrations of each of these three Igs appear to vary within a species (depending on the body site) as well as from species to species. It is clear however, that the Ig composition of fish mucus is very different to that of serum and that in several species, IgT/IgZ is enriched in mucosal secretions compared to serum. In this regard, the initial discovery of IgT as a specialized mucosal Ig has been corroborated in sea bass (Dicentrarchus labrax) (Buonocore et al., 2017), and zebrafish (Danio rerio) (Ji et al., 2020) and studies at the transcript level in other teleosts appear to also support such a role. Importantly, as discussed in detail later, while most teleosts possess IgT/IgZ, some do not, raising important questions about the diversity of mucosal immune systems that can be found within Actinopterygii.

3. Genomic, molecular and structural aspects of sIgs

3.1. Overview

Immunoglobulins (Igs) or antibodies (Abs) are expressed on the surface of B-cells (membrane-bound form) as B-cell receptors (BCR) or in soluble form in body fluids (i.e., plasma, mucus, peritoneal cavity fluid) (Salinas et al., 2011). Igs typically consist of two identical heavy (H) chains and two identical light (L) chains. Both H and L chains contain an amino-terminal variable (V_H or V_L) domain for antigen recognition, and one or more carboxyl-terminal constant (C) domains (C_H or C_L). The C_H domain defines the Ig isotype (i.e., antibody class) and mediates effector functions of the antibody molecule, including pathogen opsonization and neutralization, antibody-dependent cellular cytotoxicity (ADCC), and complement activation as reviewed in (Schroeder and Cavacini, 2010). So far, three major Ig isotypes have been identified in teleost fish, IgM, IgD and IgT/IgZ, which are designated by their heavy chains μ, δ and τ/ζ respectively (Danilova et al., 2005; Hansen et al., 2005; Wilson et al., 1997). For the most up to date nomenclature on salmonid IGHC genes please refer to Magadan et al., 2019a. The IgH locus of fish is extremely heterogeneous and diverse as evidenced by several comparative genomic studies (reviewed by (Fillatreau et al., 2013)). This astounding diversity has been acknowledged for decades, but with the advent of next-generation sequencing platforms and genomic tools, our current knowledge on fish IgH is greater than before, and such studies have confirmed the extreme volatility of IgH evolution (Fillatreau et al., 2013).

Remarkably, some deep-sea anglerfish have lost entire arms of their adaptive immune system, such as a functional RAG system, B-cells, T-cells and show loss of mhc1 gene diversity (Swann et al, 2020). These findings raise the question of how deep-sea anglerfishes respond to mucosal pathogens and microbiota in the absence of functional adaptive immune molecules such as sIgs. Combined, the current body of knowledge underscores the tremendous diversity of IgH classes and adaptive immune systems in teleosts (Fillatreau et al., 2013; Swann et al., 2020). This diversity should translate into equally diverse fish mucosal immune systems, most of which are essentially unknown.

As reviewed by Parra et al, in contrast to mammals, the teleost igh locus encodes all three Ig genes in a translocon organization (Parra et al., 2016). The typical igh locus encoding Dτ-Jτ-Cτ clusters (for IgT) are generally embedded between V_H gene segments and the Dμ/δ-Jμ/δ-Cμ-Cδ locus (for IgM and IgD), while the V_H gene segments are shared by all igh genes (Parra et al., 2016). With the absence of class switch recombination in fish, the recombination of a V_H segment to Dμ/δ-Jμ/δ-Cμ-Cδ regions would therefore delete Cτ regions in most teleosts (Mashoof and Criscitiello, 2016), resulting in the development of mutually exclusive B-cell lineages expressing either IgT or IgM/D. As determined by the three Ig isotypes, four B-cell subsets have been identified in teleost species: three subsets uniquely expressing either surface IgM (IgM⁺ B-cells) (Edholm et al., 2011; Edholm et al., 2010; Granja and Tafalla, 2019), IgD (IgD⁺ B-cells) (Castro et al., 2014; Chen et al., 2009; Edholm et al., 2011; Edholm et al., 2010; Granja and Tafalla, 2019) or IgT/IgZ (IgT⁺ B-cells) (Schorpp et al., 2006; Zhang et al., 2010), and another subset that co-expresses surface IgM and IgD (IgM⁺IgD⁺ B-cells) (Edholm et al., 2010; Xu et al., 2016; Zhang et al., 2010). There are however conflictive reports on the organs where the single positive IgM⁺ and IgD⁺ B-cells populations can be found as well as in their percentages in these tissues. While some reports find these two B-cell populations in very high or significant proportions in gut and gills (IgD⁺ B-cells) and spleen (IgM⁺ B-cells) respectively (Castro et al., 2014; Perdiguero et al., 2019), other studies find them in very low or insignificant numbers in the same organs (Ramirez-Gomez et al., 2012; Xu et al., 2016)

3.2. IgM

IgM is the predominant Ig class in teleost body fluids and is considered the most prevalent isotype in systemic immune responses, which are typically detected in plasma or serum (Parra et al., 2013). The ighμ gene consists of four μ constant (Cμ) and two transmembrane (TM) exons, and is highly conserved in all teleost species examined to date. The secretory form of ighμ transcript (sIgM) possesses all four Cμ domains and a secretory tail, while the membrane-bound form of ighμ (mIgM) transcript generally splices the TM exons to Cμ3 exon, therefore mIgM only utilizes Cμ1-Cμ3 and a TM domain (as reviewed by (Parra et al., 2016)). This structure is also conserved with the ighμ gene of more recently examined species, including rohu, (Labeo rohita), ballan wrasse (Labrus bergylta), turquoise killifish (Nothobranchius furzeri) and the southern platyfish (Xiphophorus maculatus) (Bilal et al., 2019; Bradshaw and Valenzano, 2020; Piazzon et al., 2016; Saravanan et al., 2020). However, exceptions are present in medaka and Antarctic fish, in which mIgM contains only Cμ1 and Cμ2 (Magadán-Mompó et al., 2011; Quiniou et al., 2011), while an additional mIgM form found in zebrafish only contains one Cμ (Hu et al., 2011).

As opposed to the pentameric mammalian IgM, teleost IgM in serum and mucus is chiefly tetrameric (~660–800 kilodaltons (kDa). In contrast to the subunits of mammalian IgM which are all covalently bound through disulfide bonds, under denaturing conditions, the teleost fish IgM subunits are present in various redox forms that vary in their degree of inter-heavy chain disulfide polymerization (Bromage et al., 2004; Kaattari et al., 1998). It is worth mentioning that a correlation between an increased disulfide polymerization and a greater affinity of trout IgM to antigen has been found (Ye et al., 2010). Unlike mammalian IgM, the association of the different IgM subunits in teleosts occurs in the absence of the joining (J) chain (Kaattari et al., 1998; Rombout et al., 1993; Xu et al., 2013b; Xu et al., 2016; Zhang et al., 2010). Serum IgM concentration in teleost fish varies from 0.6-21 mg/ml among different species and fluctuates with various factors (i.e. temperature, water quality, fish size, stress, infection, vaccination, etc.) (Olesen and Jorgensen, 1986; Parra et al., 2016; Solem and Stenvik, 2006). However, IgM is present at a very low concentration (~4.5-280 μg/ml) in mucosal secretions when compared to that of serum (Table 1). Increased IgM titers are detected in serum after immunization or infection and provide a protective function against a variety of pathogens (as discussed below, and Table 2). In turn, the induction of antigen-specific IgM titers in mucosal surfaces is much lower than those found in serum, and typically lower than those observed for IgT (Kong et al., 2019; Xu et al., 2013b; Xu et al., 2016; Yu et al., 2019; Zhang et al., 2010). It is worth pointing that affinity maturation of the IgM response has been demonstrated in teleosts although affinity increases in these species are much lower than those observed in mammals (Kaattari et al., 2002). Moreover, somatic hypermutation of IgM and IgD genes has also been described in teleosts, although its contribution to the degree of Ab affinity maturation remains to be investigated (Jiang et al., 2011; Magor, 2015; Yang et al., 2006; Abos et al., 2018; Perdiguero et al., 2019). In the apparent absence of germinal centers or organized lymphoid tissues in teleosts, a key question that remains to be solved is how and where are antigen-specific IgM responses induced in teleost lymphoid tissues.

Table 1.

Concentrations of Immunoglobulins in rainbow trout body fluids

	Concentration (μg/ml)				Reference
	IgM	IgD	IgT	IgT/IgM ratio in mucus / IgT/IgM ratio in serum
Serum	2520–5534	18.59–35.6	3.7–11
Gut mucus	~74.9		~7.1	~63	(Zhang et al., 2010)
Skin mucus

Table 2.

Concentrations of IgT and IgM in body fluids of rainbow trout that survived parasitic infection

	IgT		IgM
	Conc. (μg/ml)	Fold increase versus control	Conc. (μg/ml)	Fold increase versus control
Gut	~200	~51 fold	~60	no change	(Zhang et al., 2010)
Skin	~2	~10 fold	~5.5	no change	(Xu et al., 2013b)
Gill	~14	~10 fold	~22	no change	(Xu et al., 2016)
Buccal	~50	~8 fold	~800	no change	(Yu et al., 2019)
Pharyngeal	~15	~5 fold	~100	no change	(Kong et al., 2019)
Serum	~24-41	~3-5 fold	~6200-25000	~3-5 fold

Historically, the ability to detect teleost IgM has been restricted to a few species, although recently, anti-IgM monoclonal Abs have been developed in several teleost species (Bilal et al., 2016; Huang et al., 2019; Li et al., 2018; Ronneseth et al., 2015; Yang et al., 2017b, 2018), Such antibodies have been instrumental to identify IgM⁺ B-cells as well as their phagocytic activities, a capacity that was first described in (Li et al., 2006). Similar to their mammalian counterparts, fish IgM has been demonstrated to exert neutralizing and agglutinating activities against viral and bacterial pathogens respectively, and to activate the classical pathway of complement upon pathogen recognition (Boshra et al., 2004; Magnadottir et al., 1997; Ye et al., 2013). IgM has been shown to coat a significant portion of the microbiota ( ~12–50%) on different mucosal surfaces, albeit to a lesser degree than that of IgT which is the prevalent Ig in that capacity (Kong et al., 2019; Xu et al., 2013b; Xu et al., 2016; Yu et al., 2019; Zhang et al., 2010). Thus, it is plausible that IgM plays a relevant role in the maintenance of microbiota homeostasis at mucosal sites. In that regard, a recent study performed on fish transiently depleted of IgT showed that the portion of IgM-coated gill microbiota increased in the absence of IgT, in contrast, no change in IgD-coating levels was observed (Xu et al., 2020b), thus suggesting a greater role of IgM than that of IgD in microbiota homeostasis.

3.3. IgD

IgD has been found in all teleost species examined so far (Banerjee et al., 2017; Xu et al., 2019). Due to the multiple rounds of gene duplication and deletion, the ighδ gene encodes 2-16 δ constant (Cδ) exons, which display a remarkable structural plasticity among species (Parra et al., 2016). Unlike mammalian IgD, teleost ighδ transcripts are chimeric consisting of a Cμ1 domain infused by several Cδ regions. A recent study has reported an IgD/IgT chimera in the European sea bass (Buonocore et al., 2020). The IgD/IgT chimeric transcript cloned from a gill cDNA is composed of Cδ1-6 and Cτ3-4, which maybe a result of genomic rearrangement, rather than alternative splicing from the same igh locus. The highest expression of IgD/IgT chimera is found in head kidney, followed by gills and gut, while sea bass IgT is highly expressed in gut and gills (Buonocore et al., 2020; Buonocore et al., 2017). Moreover, the expression level of this chimeric Ig in gills and gut is higher than that of IgD, but lower than that of IgT (Buonocore et al., 2020). This chimeric IgD/IgT is also found in the transcriptomes of Morone saxatilis and M. chrysops (Buonocore et al., 2020). Salmonids have two ighδ genes named IGHD and IGHDD (for duplication) according to the latest nomenclature (Magadan et al., 2019a). In rainbow trout and zebrafish, a single ighδ gene is utilized for both secreted and membrane-bound forms (Ramirez-Gomez et al., 2012; Zimmerman et al., 2011). Interestingly, catfish possesses two distinct ighδ genes encoding the secretory and membrane-bound transcripts separately (Bengtén et al., 2006), while only a membrane-bound form of ighδ transcript has been found in fugu, medaka and Nile Tilapia (Magadán-Mompó et al., 2011; Saha et al., 2004; Wang et al., 2016). IgD protein has been examined in only two fish species, catfish and rainbow trout. Two variants of the secreted form of IgD (~130 kDa and ~180 kDa) are present in catfish, with a total concentration of ~40 μg/ml in serum (Edholm et al., 2011), whereas two monomeric variants with long (~370 and ~400 kDa ) and short (~240 kDa) forms of secreted IgD are present in rainbow trout, and their concentration is ~2–80 μg/ml in serum (Parra et al., 2016). The concentration of IgD in the mucus of several mucosal sites has been reported to be lower than in serum (Table 1) (Kong et al., 2019; Xu et al., 2016; Yu et al., 2019). Although IgD has been shown in several studies to be transcriptionally induced during viral, bacterial and parasitic infection (Basu et al., 2016; Makesh et al., 2015; Wang et al., 2016; Xu et al., 2019; Zhang et al., 2018), recent studies have revealed the absence of pathogen-specific IgD titers in the trout serum and mucus in response to parasite and bacterial infection (Kong et al., 2019; Xu et al., 2016; Yu et al., 2019; Zhang et al., 2020). Therefore, the role of IgD in host defense against pathogens remains unclear, although a potential role as a pattern recognition molecule has been hypothesized for secreted catfish IgD (an Ig lacking the VH region) (Edholm et al., 2010; Parra et al., 2016). With regards to a potential role of IgD in fish immunity, in 2016, secreted IgD was first reported to coat a significant portion of the fish gill microbiota (Xu et al., 2016) albeit lower than that of IgM and IgT. Subsequent studies showed similar results in other mucosal sites, including the gut (Perdiguero et al., 2019), buccal cavity (Yu et al., 2019), and pharyngeal cavity (Kong et al., 2019). Thus, it is possible that IgD plays a role in the regulation and homeostasis of the fish microbiota, as suggested in (Perdiguero et al., 2019).

3.4. IgT/IgZ

Until recently it was thought that most teleosts contained IgT/IgZ genes. However, a recent study found absence of IgT/IgZ in 25 species out of 73 Actinopterygii species analyzed including Siluriforms, Beloniforms, Gobiforms and others (Mirete-Bachiller et al., 2020), highlighting that these clades must possess mucosal immune systems that operate independently of IgT. Further analyses of Holostean genomes identified an IgT molecule with four CH domains in all Holostei and an additional IgT molecule with three CH domains in spotted gar (Lepisosteus oculatus) (Mirete-Bachiller et al., 2020). The complexity of the teleost IGHT locus is further compounded by the independent rounds of duplication and deletion of IGHZ found in cyprinodontiforms (Bradshaw and Valenzano, 2020) and by reports of Ig chimeric molecules including the carp (C. carpio) IgM/IgT (Savan et al., 2005a) and the chimeric IgD/IgT recently described in Mediterranean seabass (D. labrax) and striped bass (Morone sp.) (Buonocore et al., 2020).

In most teleost species, the ighτ/ζ genes encode four τ/ζ constant (Cτ/ζ) exons and two TM exons. Unlike ighμ, both secretory and membrane-bound forms of ighτ/ζ use all four exons of Cτ/ζ domains (Parra et al., 2016). However, different numbers of Cτ/ζ have been reported in several species: three Cτ/ζ domains in stickleback, ayu, and Antarctic fish (Gambón-Deza et al., 2010; Giacomelli et al., 2015; Kato et al., 2015); two domains in fugu (Savan et al., 2005b). Up to date, three IgT subclasses (IgT1, IgT2 and IgT3) have been identified in rainbow trout. All IgT subclasses contain 4 CH domains and are detectable in trout serum (Zhang et al, 2017). IgT1 and IgT3 include both membrane and secretory forms whereas IgT2 is only secretory (Danilova et al., 2005). Importantly, a previously reported monoclonal antibody (Zhang et al., 2010) against trout IgT1 recognizes both IgT2 and IgT3. Interestingly, two subclasses of IgZ have been identified from zebrafish, in which IgZ is present in serum and gill/skin mucus, while IgZ2 is exclusively observed in the mucus (Ji et al., 2020).

Structural and functional analyses of IgT/IgZ protein have thus far only been analyzed in rainbow trout and zebrafish (Ji et al., 2020; Parra et al., 2016). The IgT protein is expressed as a monomer (~180kDa only identified in trout) in serum, while it is found mostly as a polymer associated by non-covalent interactions in mucus (~4–5 monomers), although small amounts of monomeric IgT are also present in this body fluid (Kong et al., 2019; Xu et al., 2013b; Xu et al., 2016; Zhang et al., 2010). Compared to IgM, IgT concentration is much lower in mucus (see Table 1). Nevertheless, it is important to note that the ratio of IgT/IgM is much higher in the mucus when compared to that of serum (~20–124 fold) (Table 1). Upon infection, pathogen-specific titers of IgT are highly prevalent in the mucus from different mucosal sites, in contrast to those of IgM which are overwhelmingly dominant in the serum (Kong et al., 2019; Tacchi et al., 2014; Xu et al., 2013b; Xu et al., 2016; Yu et al., 2019; Zhang et al., 2010). It is worth pointing that pathogen-specific IgT titers can also be detected systemically (i.e., serum) but at a much lower levels than those found in mucus from several MALTs (Kong et al., 2019; Xu et al., 2013b; Xu et al., 2016; Yu et al., 2019). Supporting the key role of IgT in mucosal surfaces, it has recently been shown that IgT-depleted fish become highly susceptible to a mucosal parasite (Ich) while no compensatory IgM or IgD responses were observed (Xu et al., 2020b). Similar to the role of IgA in mammals, IgT has also been shown to be the prevalent teleost Ig in coating the microbiota of several fish MALTs including the gut, skin, gill, nose, buccal cavity, and pharyngeal cavity (Kong et al., 2019; Tacchi et al., 2014; Xu et al., 2013b; Xu et al., 2016; Yu et al., 2019; Zhang et al., 2010). A key role of IgT in teleost microbiota homeostasis was recently demonstrated in fish that were transiently depleted of IgT, in which upon depletion, fish developed a marked dysbiosis. Moreover, the microbiota of these fish was able to significantly translocate across the epithelium and reach systemic circulation as demonstrated by the presence of high levels of LPS in the plasma of IgT-depleted fish (Xu et al., 2020b), thus indicating a key role of IgT in containing the microbiota within mucosal sites.

With regards to which cells secrete IgT, it has been shown that upon stimulation of trout head kidney leukocytes with LPS and vibrio bacterin, the main IgT-secreting cells displayed surface IgT as well as high forward scatter (FSC) and low side scatter (SSC) properties (Zhang et al., 2010). Whether these large cells represent plasmablast- or plasma-like cells, remains to be determined.

3.5. Transport of sIgs into fish mucus

Polymeric Ig receptor (pIgR) mediates the transport of sIgs across epithelia and into the mucus of luminal areas through the mechanism of transcytosis (Gurevich et al., 2003; Hamuro et al., 2007; Kaetzel, 2005). This receptor is expressed by epithelial cells and has already been identified in several teleost species (Feng et al., 2009; Hamuro et al., 2007; Xu et al., 2013a; Zhang et al., 2010). Several structural and functional differences have been identified between teleost and mammalian pIgR (reviewed in (Kong et al., 2018). pIgR has been detected by IHC in rainbow trout gut, gills, skin, buccal and pharyngeal cavity, and olfactory epithelium. Further, trout secretory component (SC) has been found associated with IgT and IgM in the mucus of all these mucosal sites (Yu et al., 2018; Yu et al., 2019). In addition, the free form of SC has been detected in mucus but not the serum of several teleosts, (Hamuro et al., 2007; Kelly et al., 2017; Xu et al., 2013a) and it can directly bind to both microbiota and bacterial pathogens, thus suggesting a role for SC in innate immunity and/or the homeostasis of microbial communities at fish mucosal surfaces (Kelly et al., 2017). Interestingly, pIgR is also highly expressed in the liver of teleosts where it may be responsible for the transport systemic Igs into the fish gut, as is the case in mammals (Brandl et al., 2017; Giffroy et al., 1998; Salinas and Parra, 2015). Teleost pIgR binds IgM (Hamuro et al., 2007; Xu et al., 2013a; Zhang et al., 2010) and IgT (Zhang et al., 2010). Whether or not pIgR transports IgD in teleosts remains to be elucidated.

Teleost pIgR expression is regulated in response to infection. For instance, grass carp (Ctenopharyngodon idella) pIgR has been found to be highly up regulated in spleen, liver, kidney, skin, gill and gut, 4, 8, 12 and 24 h after Flavobacterium columnare infection. Similarly, olive flounder (Paralichthys olivaceus) infected with Vibrio anguillarum was also found to up regulate pIgR transcripts in skin, gills, gut, liver, spleen and head kidney 4, 8, 12 , 24 and 48 h post-infection (Sheng et al., 2019). In many of these studies, pIgR transcript increases correlated with up regulation of sIgs expression, suggesting that fish require more pIgR in order to be able to transport increased amounts of sIgs in response to pathogenic challenge. Other interesting immune associations have been recently discovered with pIgR in Asian sea bass (Lates calcarifer), where vitellogenin (Vg) was found to interact with pIgR and it was suggested that Vg may play a role as a pattern recognition receptor in this species (Yang et al., 2017a).

Interestingly, it has been shown that turbot (Scophthalmus maximus) egg IgM binds to pIgR, and the expression and visualization of pIgR by in situ hybridization in the yolk sac and the embryo could be related with the maternal transfer of sIgs into the embryo (Qin et al., 2019). Altogether, the above studies strongly support the role of teleost pIgR in the transport of sIgs to mucosal tissues and highlight additional immune roles of pIgR in mucosal homeostasis.

4. Interactions between fish sIgs and microbiota

4.1. Fish microbiomes

Teleost fish, like all metazoans, live in symbiosis with diverse microorganisms that provide them with vast physiological and metabolic benefits. The field of fish microbiomes has blossomed in the past 8 years, with many studies characterizing which bacterial species colonize each mucosal site in many diverse teleost species. Moreover, studies have revealed what environmental and genetic factors are among the major determinants of fish microbial communities. In general, it appears that environment is a stronger determining factor of fish microbial assemblies than host genetics (Brown et al., 2019; Stagaman et al., 2017; Vasemagi et al., 2017) . Further, body site, diet, and life cycle stage are all drivers of fish microbial community assembly (Llewellyn et al., 2016; Lowrey, 2014; Stagaman et al., 2020) . Interestingly, comparisons between wild caught and laboratory zebrafish microbiomes indicate that the microbiome of laboratory animals can model that of wild counterparts (Stagaman et al., 2020). Whether this is the case for other fish species is unknown but, for instance, the gut microbiome of wild Atlantic salmon contained a high abundance of Mycoplasmataceae (Llewellyn et al., 2016), and similar findings have been reported in farmed and laboratory reared salmonids (Lowrey et al., 2015; Zarkasi et al., 2014). For further reading on fish microbiomes refer to the following reviews (Kelly and Salinas, 2017; Llewellyn et al., 2014; Stagaman et al., 2020).

4.2. sIg Coating

Igs contain Fc and V regions that can interact with antigens in a non-specific or specific fashion, respectively. As mentioned earlier, mucosal secretions include heterogenous mixtures of different sIgs. Thus, bacterial coating by IgA, IgG, and IgM has been reported in mammals (Chen et al., 2020). However, sIgA is the chief mucosal Ig in that capacity (Donaldson et al., 2018; Wilmore et al., 2018), at least in laboratory mice. Nevertheless, it has recently been shown that in contrast to these mice, human IgM coats a similar percentage of gut microbiota than that of IgA. In fact, microbiota coated by human sIgM were also coated by SIgA, and thus, it is apparent that at least in humans, IgM may also play a key role in the control of microbiota homeostasis (Magri et al., 2017). It will be pivotal to assess whether gut IgM have such capacity in non-laboratory mice, as the lack of such IgM coating in laboratory mice maybe partly due to the very clean environment conditions where they live. Interestingly, as reviewed above, a significant portion of the gut microbiota in rainbow trout are coated by IgM and thus, this species might be instrumental in understanding the role of vertebrate IgM in microbiome homeostasis.

The ability of sIgA to coat fecal bacteria in humans was first discovered in 1996 (Van der Waaij et al., 1996). Mucosal IgA is generated via T-dependent (TD) and T-independent (TI) pathways and both mechanisms have been shown to generate microbiota-coating sIgA. Several studies have revealed that sIgA coats both beneficial as well as pathogenic taxa (Bunker et al., 2015; Palm et al., 2014). Interestingly, bacteria can be found either highly or lowly coated by IgA, and some of the highest coated taxa are colitogenic bacterial species that are involved in the pathology of inflammatory bowel disease (Palm et al., 2014). Based on these initial studies on sIgA coating, it was proposed that TI-induced IgA broadly targets non-invasive commensals, while TD-induced IgA coats penetrant commensals and invasive pathogens (Bunker et al., 2015; Chen et al., 2020; Pabst and Slack, 2020; Palm et al., 2014).

In teleost fish, the first report of sIgs coating bacterial commensals dates back to 2010, where rainbow trout gut bacteria were shown to be coated by IgT and to a lesser extent IgM (Zhang et al., 2010). Additionally, double coated microbial populations were also identified in trout gut. Since that first report, bacterial coating by sIgs has been reported in rainbow trout skin, gills, gut, olfactory organ, buccal cavity, pharyngeal cavity (Kong et al., 2019; Tacchi et al., 2014; Xu et al., 2013b; Xu et al., 2016; Yu et al., 2019; Zhang et al., 2010). In all these reports, IgT was always identified as the predominant isotype associated with bacteria, except in the olfactory organ where IgT and IgM coated similar percentages of the microbiota, and double coated populations were the most abundant. In addition, in several of these MALTs, microbiota could be found double coated with IgT and IgM, while a very tiny portion was found triple-coated with sIgT, sIgM and sIgD. In that regard, the unique bacterial community structure of each fish mucosal tissue (Lowrey et al., 2015) combined with unique physico-chemical properties (i.e oxygen levels, pH, oxygen peroxide levels) and immunological factors (i.e antimicrobial peptides) may all contribute to the differences in coating levels among different fish MALTs.

In order to understand the biological implications of sIg coating of commensals, it is critical to know what bacterial taxa are coated by each sIg class. To answer this question, IgA-Seq was recently developed in mammals. This technique consists of sorting bacteria coated by sIgA and then performing 16S rDNA sequencing to determine the diversity and proportions of the sIgA-coated microbial community. As previously stated, these studies revealed that sIgA coats specific subsets of bacteria which can be both beneficial but also pathogenic, such as colitogenic bacteria (Bunker et al., 2015; Palm et al., 2014). Recently, our laboratories performed the first IgT-Seq experiments in fish, using rainbow trout gill-associated bacteria from naive fish. Similar to the results from IgA-Seq studies in mammals, it was observed that sIgT coats a broad but well-defined range of bacteria accounting for approximately 25% of the entire gill bacterial community. Importantly, the trout gill IgT-coated community was enriched in Actinobacteria and Firmicutes, mostly belonging to anaerobic and facultative anaerobic taxa. IgT targeted potentially beneficial taxonomic groups such as Clostridiales and Propionibacteriales, which can produce short-chain fatty acids. Short-chain fatty acids are known to enhance barrier integrity, to induce mucin gene and AMP expression, and to dampen intestinal inflammation in mammals (reviewed by (Parada Venegas et al., 2019)). IgT also targeted potentially pathogenic bacteria such as members of the orders Enterobacteriales and Pseudomonadales. Interestingly, IgT coated disease-driving bacteria included Candidatus Branchiomonas, Acinetobacter sp., Serratia sp., Veillonela sp., Streptococcus sp., and Stenotrophomonas sp. Combined, these studies have revealed that sIgT is able to target both beneficial and pathogenic bacteria, suggesting several mechanisms by which sIg coating contributes to mucosal homeostasis in fish (Xu et al., 2020b) as discussed below (Figure 1).

Figure 1: Schematic representation of the different direct and indirect mechanisms by which sIg coating contributes to microbiota homeostasis at fish mucosal barriers.

Finally, it is important to remember that a large number of teleost taxa do not encode IgT/IgZ. In other taxa, such as rainbow trout, IgT has evolved three sub-classes with potentially unique functions. This raises exciting biological questions as to how microbiota-teleost mucosal immunity interactions have co-evolved in systems where IgT/IgZ is not present and how microbiota may have driven diversification of Ig subclasses at mucosal surfaces.

4.3. What are the functional consequences of sIg coating?

The relevance of bacteria sIg coating in teleosts has recently been unveiled by the use of a novel experimental model in which IgT is transiently depleted during a 4 week period using an antibody-based approach (Xu et al., 2020b). In this IgT depletion model, examination of the trout gill microbiome over the course of depletion provided insights into the biological functions of sIgT coating. First, IgT depletion led to a profound dysbiosis in the gill mucosa in which 39 operational taxonomic units (OTUs) were confirmed to have different bacterial abundances when comparing IgT-depleted versus control fish. Significantly, a ~10-fold increase in the ratio Bacteroidetes-to-Firmicutes was observed in IgT-depleted animals, thus confirming further a state of global dysbiosis in these fish. Second, absence of IgT resulted in losses of beneficial taxa belonging to the phylum Proteobacteria. Thus, sIgT coating is likely contributing to colonization of beneficial bacteria, at least in trout gills (Figure 1). Third, IgT depletion resulted in expansions of pathogenic taxa, indicating the IgT coating likely inhibits overgrowth of pathogens. Fourth, IgT depletion resulted in translocation of bacteria across the gill epithelium and into systemic circulation, indicating that IgT coating is necessary to contain bacteria within the external mucosal layer. Further, other indirect mechanisms of action cannot be ruled out (Figure 1). One of these mechanisms is the production of functionally relevant metabolites by coated microbes that successfully colonize the mucosal barrier. Indeed, bacterial-derived metabolites such as sphingolipids can stimulate sIg secretion (Sepahi et al., 2016) and therefore potentially promote coating of more bacterial cells. In that regard, a significant decrease in the abundance of Sphingomonadales (a bacterial order known to produce high amounts of sphingolipids) was observed in IgT-depleted fish (14.2% in control fish and 5% in IgT-depleted fish), which strongly suggests that IgT is critical for the colonization of these beneficial bacteria (Figure 1). Finally, in mammals, sIgA promotes bacteria-bacteria interactions in the gut therefore shaping the symbiotic community and their function (Nakajima et al., 2018). Overall, these above-listed sIgT-coating functions have also been reported in mammalian studies for sIgA (Bunker et al., 2015; Palm et al., 2014), thus highlighting that similar evolutionary pressures have shaped mucosal Igs for millions of years in the control of microbiota homeostasis. While much work needs to be performed to fully appreciate the biological roles of fish sIg coating of microbiota, Figure 1 illustrates the current knowledge as well as other potential roles of sIg coating in fish. It is worth-pointing that similar to what has been described in sIgA-deficient mammals, dysbiosis caused by absence of IgT was found to be associated with increased mucosal tissue damage, and inflammation. Critically, IgT recovery in the transiently IgT-depleted fish was sufficient to restore both tissue and microbiome homeostasis. In addition to corroborating a key role of sIgT in the homeostasis of the microbiota, these data strongly suggest that the roles of sIgT in microbiota coating and mucosal tissue homeostasis are interwoven.

5. sIg responses to infection

Mucosal tissues are natural, physical, biochemical, dynamic, and semipermeable barriers that facilitate the exchange of nutrients, water, ions and proteins with the external environment (Salinas et al., 2011). Mucosal tissues are also immunological barriers located at the interface between the host and the pathogen-rich environment. Since most infections in fish initially take place at mucosal sites, mucosal innate immune responses, including cytokine and chemokine responses, are critical for activating local immune cells as well as for recruiting leukocytes from other mucosal or systemic lymphoid organs to the site of infection. Further, innate immune responses at these surfaces also shape mucosal adaptive immune pathways (Salinas, 2015; Zou and Secombes, 2016). The fish mucosal adaptive immune response is presumably mediated by local T- and B-cells and/or lymphocytes recruited from systemic lymphoid organs, or both, although the specific areas in the MALT where mucosal antibody responses are induced are thus far unknown (Salinas et al., 2011). Here, we propose three different pathways of sIg induction at fish mucosal sites based on the current knowledge available. These three pathways are not mutually exclusive and therefore one or more may operate at the same time or may be turned on depending on the type of infection, pathogen load or route of immunization (Figure 2). Below we summarize what has been reported over the last 10 years regarding sIgs at fish mucosal barriers during early and late immune responses to pathogens.

Figure 2: Testing for the existence of a common mucosal immune system (CMIS) in teleosts.

Hypothetical models for B cell activation and antigen-specific sIg production in teleost fish in response to infection. In Model 1, infection of one MALT (the gut), results in local activation and proliferation of B cells as well as production of sIgs. Activated B cells would then migrate to other MALTs. Here the gut would be the inductor site, and only skin and gills are depicted as effector sites but all other MALTs would behave in a similar fashion. In Model 2, infection at one MALT, in this case the gut, only results in local activation and proliferation of B cells but no migration of activated B cells to other MALT. Thus, sIg production is restricted to the site of infection. In Model 3, infection may result in antigen transport to the spleen where B cell activation and proliferation would take place first. Next, B cells would migrate to all MALTs where they would differentiate into plasma cell and produce sIgs. All these models are only hypothetical and further experimental evidence is currently needed to support or refute them.

As already stated, IgT is an Ig specialized in fish mucosal immune responses (Zhang et al., 2010) with analogous roles to that of IgA (Rodríguez et al., 2005). Nevertheless, IgM and IgD can also be detected in fish mucosal secretions (Salinas et al., 2011; Xu et al., 2020b; Xu et al., 2016) and possibly contribute to some degree to the defense against mucosal pathogens, especially in species lacking IgT, although the specific functional contributions of these sIgs to protective mucosal responses remain unknown. It is important to highlight that up until recently, most studies that evaluated antibody titers have focused on IgM responses due to the lack of antibody reagents to measure IgT responses in several fish species (Parra et al., 2016). Consequently, the majority of the studies so far published rely on gene expression analyses of IgM, IgT and IgD upon pathogenic exposure, and therefore it is unclear whether such data may correlate with antigen-specific antibody responses to these pathogens (Dixon et al., 2018). Because of the large body of new reports since 2011 on sIg responses to pathogens, and due to space constraints, below we discuss only the most representative reports, while all other studies have been summarized in Supplementary Table 1.

5.1. sIgs responses in GALT

Similar to higher vertebrates, the fish gut mucosa plays an important role in the fight against pathogens (Lanning et al., 2004; Salinas, 2015; Shields, 2000; Sommer and Bäckhed, 2013). The first studies indicating that IgT was an sIg specialized in mucosal immunity were performed in the gut of rainbow trout, where significant increases of IgT⁺ B-cells could be detected after infection with the myxozoan endoparasite Ceratomyxa shasta, while IgM⁺ B-cells numbers remained unchanged when compared to those of control fish (Zhang et al., 2010). These studies showed for the first time compartmentalized Ig responses in mucosal versus systemic sites in a fish: while parasite-specific IgT titers where overwhelmingly detected in the trout gut mucus, IgM-specific responses were for the most part observed in serum. Subsequent studies have evaluated sIgT responses in the gut of other fish species, some of which are in agreement with the findings in rainbow trout. For instance, gilthead seabream (Sparus aurata) infected with the parasite Enteromyxum leei for over 100 days had significant IgT- but not IgM-specific titers in the gut mucus while IgM titers against the parasite were only detected in the serum (Piazzon et al., 2016). Interestingly, 64 days post-infection, both membrane IgT and IgM transcripts were up regulated in the gut, although the number of IgT⁺ and IgM⁺ B-cells were not quantified. Previous studies using the same model had shown increased IgM⁺ B-cell levels and IgM expression in gut after parasitation (Estensoro et al., 2012). Further, a recent study found that E. leei induced a polyclonal expansion of diverse IgT and IgM repertories in the posterior gut, implying a role of both isotypes in the seabream adaptive immune response against this parasite (Picard-Sánchez et al., 2020). Polyclonal expansions could however be the result of parasite-induced dysbiosis whereby translocated bacteria activate IgM and/or IgT B-cells polyclonally. Thus, the exact roles of specific IgM responses to E. leei in teleosts still remain to be fully elucidated.

While fish sIgs in the gut coat a large percentage of resident microbiota, reports regarding the role of sIgs in fish gut after bacterial infection are still scarce. Prior to the discovery of IgT as a mucosal immunoglobulin, all studies on gut Igs to bacterial infections had focused on IgM responses (Salinas et al., 2011). Subsequent studies have for the most part focused on the measurement of sIg transcript levels upon gut bacterial infections, except for a report in zebrafish which recently showed an increase in gut IgZ⁺ B-cells together with an increase of IgZ and IgZ2 specific titers against E. tarda 3, 7 and 14 days post-infection in gill and skin mucus (Ji et al., 2020). It is unclear how specific Ab titers against E. tarda can be elicited at such an early point (i.e., 3 days), which for fish is very unusual. Future studies will have to address this conundrum and prove the adaptive and specific nature of such a response. With regards to transcript studies, up regulated IgT mRNA levels were detected in the hindgut and stomach of olive flounder (P. olivaceus) after E. tarda bath infection, thus implying a role for IgT against bacterial pathogens in flounder gut immunity. However, in the same study, intraperitoneal (i.p.) injection of the pathogen also induced up regulation of IgT transcripts in spleen thus suggesting additional roles of IgT in the systemic compartment (Du et al., 2016). In a different study, both IgT and IgM transcript levels appeared to be elevated in rainbow trout midgut after exposure to Yersinia ruckeri (Evenhuis et al., 2012).

In contrast to sIg responses to other pathogen types, sIg responses to viral infection in the gut remain largely unexplored. A study conducted in rainbow trout showed an increase of IgM- and IgT⁺ B-cells in foregut and pyloric caeca 7 days post-infection with infectious pancreatic necrosis virus (IPNV), while those B-cell populations remained unchanged in hindgut and midgut (Ballesteros et al., 2014). In a different study also in rainbow trout, an increase of IgM, IgT and IgD transcripts was detected in gut 2 and 6 days post-infection with viral hemorrhagic septicemia virus (VHSV) (Leal et al., 2019). Another study that only focused on IgD showed a ~2.5 fold up regulation of IgD transcripts in L. rohita gut after 5 days of rhabdovirus infection (Basu et al., 2016). Thus, it is unclear at this point which sIgs can be induced to recognize viruses in teleost GALT.

5.2. sIgs responses in GIALT

Fish gills play critical immune roles supported by resident immune cells, including B and T-cells, monocytes, macrophages, neutrophils, thrombocytes, dendritic-like cells, natural killer (NK)-like cells, eosinophilic granule cells (EGCs), rodlet cells, and melanin-containing cells, that combined form the GIALT (Aas et al., 2017; Koppang et al., 2010; Koppang et al., 2015; Rességuier et al., 2020; Xu et al., 2016).

Gills are the target tissue of several fish ectoparasites. For example, several studies have focused on sIg responses against the parasite Ichthyophthirius multifiliis (Ich). Ich infection induced significant parasite-specific IgT responses, together with significant increases in the number of IgT⁺ B-cells in trout gills (Xu et al., 2016). Of note, increased IgT⁺ B-cells numbers could be explained to some degree by the observed local proliferation of these lymphocytes in the gill tissue, although recruitment of other IgT⁺ B-cells to the site of infection could not be ruled out. In contrast, the same study showed that upon infection, IgM⁺ B-cell numbers remained unchanged when compared to control fish, while parasite-specific IgM titers in the gill mucus were much lower than those of IgT. Conversely, high parasite-specific IgM titers were found in the serum, while those of IgT were almost negligible. In contrast, no parasite-specific IgD titers was detected neither in the gill mucus nor in the serum. Altogether, similar to what was found in GALT, these data pointed to IgT as the main sIg against Ich infection in rainbow trout GIALT (Buchmann, 2020; Gomez et al., 2013; Parra et al., 2015). Another well-studied parasite that affects salmonids is Neoparamoeba perurans, the causative agent of amoebic gill disease (AGD) (Marcos-López and Rodger, 2020). AGD induced a down regulation in IgT and IgM transcript levels together with a down regulation of molecules involved in antigen processing and B-cell activation in Atlantic salmon (Young et al., 2008). Such results may suggest an inhibition of sIg production by this pathogen. Similarly, a different study showed that Atlantic salmon infected with Entamoeba perurans had reduced total IgM levels in gill mucus after a long exposure period with high infection, suggesting again an inhibitory role played by the parasite, or a possible immune exhaustion of the IgM response (Marcos-López et al., 2017). Further studies have been conducted analyzing gene expression of sIgs in gill during AGD infection, obtaining dissimilar results depending on the species and infection time point (Pennacchi et al., 2016; Pennacchi et al., 2014; Valdenegro-Vega et al., 2015). More detailed information on AGD and the host response can be found in (Marcos-López and Rodger, 2020). Finally, a recent study showed the induction of specific IgM responses against the monogenean ectoparasite Heterobothrium okamotoi in gill mucus from grass puffer (Takifugu niphobles) (Matsui et al., 2020). Interestingly, specific IgM caused deciliation of the oncomiracidia (Matsui et al., 2020). Very few studies have addressed IgD responses to gill parasites. Specifically, IgD transcript levels were up regulated both in L. rohita and rainbow trout after infection with Argulus sp. and Ich, respectively (Basu et al., 2016). Similarly, infection of dojo loach with a fungus (Saprolegnia sp.) induced an up regulation of gill IgD transcripts. The question remains whether up regulated IgD transcripts translate into pathogen-specific IgD titers or not.

With regards to sIg responses to gill bacterial pathogens, some in depth work has been performed with F. columnare, a bacterial pathogen that causes severe outbreaks in salmonids and that induces significant gill pathology (Declercq et al., 2013). sIgT was found to be the predominant bacteria-specific sIg induced in the gill mucosa upon F. columnare infection whereas negligible or no IgM- or IgD-specific titers were detected in the gill mucus, respectively (Tongsri et al., 2020; Xu et al., 2016). In contrast, high titers of IgM against this bacterium were detected only in the serum, thus indicating a compartmentalization of pathogen-specific IgT and IgM responses in mucosal and systemic sites respectively, similar to that previously described in the gut (Xu et al., 2016; Zhang et al., 2020). A recent study (Tongsri et al., 2020) corroborated the results from the aforementioned work (Xu et al., 2016). Moreover, in this study, a 75 day-long longitudinal analysis of Ig transcripts in rainbow trout gill upon F. columnare infection showed that mIgT and mIgM transcripts were induced as early as 2 dpi (~2-4 fold), reaching maximum expression at 75 dpi, with a ~45-fold and ~54-fold increase, respectively. These findings may suggest a role for both Igs in the early and late immune responses against this pathogen. Interestingly, IgD transcripts were only found up regulated at 28 dpi. While both IgT and IgM transcripts were significantly up regulated, only IgT⁺ B-cell numbers significantly increased in the gill filaments at 28 dpi and 75 dpi, a finding that correlated with the significant bacteria-specific IgT titers detected in the gill mucus at these time points. In contrast, IgM⁺ B-cells were rarely found using IHC, and significant bacteria-specific IgM titers were only found in serum, while no IgD-specific titers could be detected (Tongsri et al., 2020). Overall, findings of this 2020 study indicate a lack of correlation between increased mIgM and mIgD transcripts and the presence of IgM-and IgD-specific titers, respectively, thus strongly suggesting that the study of Ig transcripts cannot be a substitute for the evaluation of pathogen-specific Ig responses. In another study, E. tarda-specific IgZ titers were detected mainly in zebrafish serum and to a lesser extent also in skin and gill mucus, while IgZ2 specific titers in these mucosal sites were lower than those of IgZ. In contrast to IgZ, the serum did not contain E. tarda-specific IgZ2 (Ji et al., 2020). These data suggested that while IgZ2 is specialized in mucosal responses, IgZ appears to play a role in both systemic and mucosal compartments. Accordingly, IgZ2⁺ B-cells were prevalent in skin and gill mucosal sites supporting a more dominant role of IgZ2 in mucosal immunity. Interestingly, zebrafish IgZ and IgZ2 exhibited different but complementary activities in antibacterial immunity: while IgZ was preferentially involved in a C1q-mediated bactericidal mechanism, IgZ2 appeared to participate in bacterial neutralization due to its coating of the bacteria (Ji et al., 2020). These data indicate different roles of the two subtypes of IgZ in zebrafish and thus, it will be of interest to analyze further whether IgT/Z subclasses of other fish species also play differential roles in systemic and mucosal immunity.

Changes in gill sIg upon bacterial infection have also been analyzed in other models but only at the transcriptional level. For example, in olive flounder (P. olivaceus), IgT transcript levels were increased in both gills and systemic immune tissues following an E. tarda challenge, although the induction levels depended on the route of infection. More specifically, bath-immersion induced higher up regulation of IgT transcripts in gill, skin, hindgut and stomach when compared to i.p. injection, whereas increases in spleen IgT transcripts were better induced by i.p. injection of the pathogen. These data suggest that IgT is prevalently induced in several mucosal surfaces when flounder are infected via a mucosal route (Du et al., 2016). Ayu fish (Plecoglossus altivelis) infected with V. anguillarum by immersion had increased IgT and IgM transcripts after 10 and 20 days in gill and trunk kidney, although the magnitude of the increase was greater in gills than in trunk kidney; IgD transcript levels, in turn, remained unchanged (Kato et al., 2015). Interestingly, other studies in L. rohita showed instead an up regulation of gill IgD transcripts 24 h after infection with A. hydrophila (Basu et al., 2016), thus suggesting a very early response of IgD in gills in this infection model. Combined this body of work highlights the need for further in-depth studies of sIg responses to gill bacterial pathogens.

Gill sIg responses to viral infection continue to be ill documented. Thus, very few studies have been carried out since our previous review on this subject in 2011 (Salinas et al., 2011). Infection studies in Atlantic salmon (Salmo salar) with infectious salmon anemia virus (ISAV) showed that sIgT transcripts were up regulated (~500 fold) in the gill ILT after 20-24 days post-infection (Austbø et al., 2014), thus suggesting a role for the ILT in the induction of IgT responses in the gill. In rainbow trout infected with viral hemorrhagic septicemia virus (VHSV), the expression of IgM and IgT transcripts was significantly induced in gills 1 dpi, while transcript levels returned to basal levels by 3 dpi (Aquilino et al., 2014). In another study conducted in European seabass (D. labrax), fish were infected by immersion with betanodavirus and it was observed that gene expression levels of IgT and IgM did not change significantly in the spleen or gill 24 h post-infection in comparison to those of control fish (Buonocore et al., 2017). Since betanodavirus is a neurotropic virus that causes encephalopathy in fish (Doan et al., 2017), the gill may not be a target tissue in this infection model and sIg responses may be induced at other body sites. Clearly, more studies are required to assess the role of the different sIgs in immune defense against viral pathogens in the teleost gill.

5.3. sIgs responses in SALT

The integument is the largest surface area in fish and therefore the target of multiple external aggressions as well as pathogen attack. Teleost skin is a living mucosal epithelium where both innate and adaptive immune responses against pathogens take place (Ángeles Esteban, 2012; Xu et al., 2013b). The first study indicating a key role for IgT against skin pathogens was performed in rainbow trout infected with Ich. This parasite prevalently induced in the skin mucus Ich-specific IgT responses while negligible IgM responses against the parasite were detected (Xu et al., 2013b). Accordingly, the parasite was overwhelmingly coated by sIgT in the skin of infected fish, while very little IgM coating was detected. Moreover, after infection, significant accumulations of IgT⁺ B-cells were detected in the skin epidermis, while numbers of IgM⁺ B-cells remained unchanged (Xu et al., 2013b). In a subsequent study, Ich-infected fish also showed a significant up regulation of IgT, IgD and IgM in skin after 24 h and 7 dpi, being the highest increase detected for IgT followed by IgD and IgM (Zhang et al., 2018), although it remains to be seen if immunoglobulin protein levels and specific titers are also up regulated at those early time points. Interestingly, zebrafish re-infected several times with Ich revealed unique immune responses at each infection cycle. The primary early response (1 h, 24 h and 48 h after infection) was characterized by onset of innate immune responses and neutrophil mobilization. In turn, upon the fourth infection cycle, up regulation of IgM and IgZ2 transcripts was noted in the skin within the first 48 h (Jørgensen et al., 2018). Another study in nile tilapia (Oreochromis niloticus) showed transcript up regulation of IgT and pIgR in skin after 5 days post-reinfection with the monogenean ectoparasite Gyrodactylus cichlidarum, while IgM transcripts remained unchanged (Zhi et al., 2020). Atlantic salmon infected with sea lice (Lepeophtheirus salmonis) had increased IgT transcript levels in the skin 24-26 dpi, while IgM transcript levels were unchanged, thus supporting a role for IgT in skin mucosal immune responses against this important pathogen (Gallardi et al., 2019). In contrast, another study conducted in Atlantic salmon infected with sea lice showed that spleen IgT transcripts increased higher than those of IgM at day 15 post-infection. However, IgM transcripts were more highly up regulated in skin 5 and 15 dpi, in comparison to those of IgT (Tadiso et al., 2011). In a different set of studies, skin IgT transcripts were found more up regulated than those of IgD in rohu (L. rohita) 3 days post-infection with the ectoparasite A. siamensis, while IgM transcripts were only detected in head kidney at day 30 post-infection. In that organ, IgM expression was more up regulated than that of IgT and IgD (Kar et al., 2015). Another study in Gibel carp (Carassius auratus gibelio) naturally infected in fish farms with the parasite Myxobolus honghuensis showed increased IgT, and to a lesser degree ,IgM transcripts in the pharynx (Zhao et al., 2019). Collectively, these studies suggest a key involvement of IgT in the mucosal immune response against skin parasites.

Compared to parasites, skin sIg responses to bacteria or viruses have been less investigated. Loach (Misgurnus anguillicaudatus) bath-infected with F. columnare had elevated IgT transcripts in skin 7 days post-infection, whereas a remarkable up regulation of IgM transcripts occurred in the spleen (Xu et al., 2019). High up regulation of IgT transcripts in the skin (i.e., ~510 fold increase) was also observed in Nile tilapia (O. niloticus) 24 h after E. tarda bath infection, whilst no changes were observed in head kidney. In contrast, infection with this pathogen had no effect on the expression of IgM transcripts in skin or head kidney tissues (Velazquez et al., 2018). Red mark syndrome (RMS) is a skin disease of rainbow trout that is potentially caused by Midichloria-like bacteria (Cafiso et al., 2016). RMS infection by cohabitation in rainbow trout revealed an increase in IgD⁺ B-cells 61 and 82 days post-cohabitation, while IgM⁺ B-cells increased 82 and 97 days post-cohabitation. In contrast, skin IgT⁺ B-cells were scarce and remained unchanged during the whole trial (von Gersdorff Jørgensen et al., 2019). In addition, RMS induced skin lesions at different levels, which were classified as severe and moderate. It was shown that IgD and IgM transcripts were up regulated in skin of fish with severe and moderate lesions, respectively, both 61 and 82 days post-cohabitation. On the other hand, IgT transcripts were up regulated in fish with severe and moderate lesions 61 days post-cohabitation, while after 82 days, such transcripts were only up regulated in fish with severe lesions.

Fish viruses are known to infect skin epithelial cells (Anders and Yoshimizu, 1994; Liu et al., 2020). A study conducted in gilthead seabream naturally infected in fish farms with lymphocystis disease virus showed a down regulation of skin IgM transcripts together with a significant reduction in total serum IgM, however whether this effect was caused by the virus or by other environmental factors remains unclear (Cordero et al., 2016). Overall, teleost skin sIg responses to viral infection are still not well understood at this point due to the lack of studies in this area.

5.4. sIgs responses in nasal, buccal and pharyngeal MALTs

Many fish pathogens infect the olfactory organ (reviewed by Das and Salinas, 2020). The nasopharynx-associated lymphoid tissue (NALT) of fish is a MALT that shares the main characteristics of other fish MALTs and responds to both pathogens and vaccines (Das et al., 2020; Tacchi et al., 2014). A unique aspect of trout NALT is that B-cells are located intraepithelially rather than in the lamina propria as it is the case in the gut (Tacchi et al., 2014). For a recent review on fish nasal immune responses see Das and Salinas 2020. Using the Ich trout infection model (Yu et al., 2018), IgT was found to coat Ich trophonts in rainbow trout olfactory organ after infection, which correlated with significant parasite-specific IgT titers in the nasal mucus as well as increased IgT⁺ B-cell numbers and their proliferation in NALT. As observed also in the skin (Xu et al., 2013b), gill (Xu et al., 2016) and gut mucosa (Zhang et al., 2010), Ich-specific IgM and IgD titers were low or negligible respectively in nasal mucus.

Recently, some studies have focused on the mucosal immune response of the buccal and pharyngeal cavity of rainbow trout against Ich (Kong et al., 2019; Yu et al., 2019). These studies have shown that these two cavities contain previously unrecognized MALTs (Kong et al., 2019; Yu et al., 2019). In these mucosa, Ich infection induced very similar responses to those induced in the skin (Xu et al., 2013b), gills (Xu et al., 2016), gut (Zhang et al., 2010) and NALT (Yu et al., 2018). More specifically, parasite-specific IgT titers were prevalent in both buccal and pharyngeal mucosal secretions, where upon infection, Ich was also found mostly coated by IgT. Moreover, in both mucosal sites, the authors could detect significant increases in the number of IgT⁺ B-cells as well as important local proliferative IgT⁺ B-cell responses upon infection. In contrast, IgM and IgD titers against the parasite were very low or absent respectively in both the buccal and pharyngeal cavities. In a different study, rainbow trout buccal mucosa was studied upon F. columnare infection. Up regulation of IgT transcripts was detected 2- and 28 days post-infection, while IgM expression also increased at 28 dpi albeit in moderation when compared to IgT. In contrast, IgD transcripts did not change significantly (Xu et al., 2020a). In addition, IgT⁺ B-cells increased in the buccal mucosa 28 dpi, showing also an important proliferative response. In addition, IgT specific titers were observed in buccal mucus, whereas those of IgM where mainly detected in serum 28 dpi. Combined, this recent body of work suggest the presence of an immunological continuum in the gastrointestinal tract of salmonids from the mouth to the anus that is chiefly dominated by sIgT responses. Further, all of the above studies appear to indicate that sIgT is the prevalent Ig responding to pathogens in all teleost MALTs, while the specific contribution of IgM and IgD require further validation.

6. sIgs responses to mucosal vaccination

At present, vaccination is the most effective method for disease prevention and pathogen control in aquaculture facilities (Rombout and Kiron, 2014). Vaccines that stimulate mucosal sites (i.e., mucosal vaccines) have long been proposed as an alternative to injected vaccines. Mucosal vaccines in fish first target mucosal tissues thereby stimulating local immune responses (Salinas et al., 2015; Wang et al., 2020; Xu et al., 2020b). Vaccine induction of specific sIgs at mucosal barriers should induce pathogen neutralization, thereby preventing infection (Munang’andu et al., 2015b; Rombout and Kiron, 2014; Wang et al., 2020; Wilson et al., 2020). Mucosal vaccination is also safe to fish and avoids deleterious side-effects inflicted by injection (Parra et al., 2015). In addition, mucosal vaccines may be more suitable for the aquaculture industry as they allow large-scale vaccination of fish. Despite all the clear advantages offered by mucosal vaccines, most of the current strategies used for mucosal vaccination, with the exception of nasal vaccines (which are only at experimental stages), generate in general poor or mediocre immune protective responses (Munang’andu et al., 2015a), As a result, only a handful of mucosal vaccines (mostly immersion vaccines) are licensed for its use in aquaculture (Ma et al., 2019).

Similar to the studies on sIg responses to pathogens, an important bottleneck in the development of mucosal vaccines has been the lack of specific antibody reagents to IgT in several species, which has precluded the analysis of antigen-specific IgT titers (Dixon et al., 2018). In that regard, a number of mucosal immunization routes have been shown to induce sIg transcripts (especially those of sIgT) upon vaccination, however, whether transcriptional increases translate into antigen-specific sIg titers, is a key point that remains to be determined. It is worth pointing that some of the mucosal immunization routes studied appear to induce different degrees of protective responses, although it is unclear at this point which immune mechanisms (i.e., cell-mediated versus humoral) are responsible for providing protection. An important consideration not taken into account during the development of a fish vaccine, is whether the pathogen is more susceptible to be eliminated by IgM or IgT responses. We propose as a potential solution to first evaluate whether in the absence of a specific Ig, fish become more susceptible to that pathogen, and if it does, then vaccines should be developed to specifically induce that particular Ig class. For example, in the absence of IgT, trout become much more susceptible to the Ich parasite (Xu et al., 2020b). This suggests that future vaccines against this parasite ought to probably induce IgT specific protective responses in order to eliminate the parasite. Future fish models devoid of IgM should also be developed to evaluate the specific contribution of IgM in eliciting protective responses to a given pathogen.

Thus, much work remains to be done to improve the efficacy of mucosal vaccines in fish. Most studies conducted with mucosal vaccines in fish since 2011 are summarized in Supplementary Table 2, while below we discuss the most representative reports in this area.

6.1. Oral vaccination

Oral vaccination represents an attractive delivery method in aquaculture due to the possibility of including the vaccine as part of the fish diet. However, this method of vaccination presents many challenges as shown by the fact that a very small number of oral vaccines are currently available for human or animals, including fish (Embregts et al., 2016). It is believed that fish oral vaccines have been unsuccessful in part because of the strong digestive enzymes present in the fish gastrointestinal tract which may overly degrade the antigen before it gets processed by antigen presenting cells. To solve this issue, several antigen carriers and encapsulation strategies are being developed to protect antigen from degradation (see (Moges et al., 2020; Mutoloki et al., 2015; Plant et al., 2011; Vinay et al., 2018) reviews). These carriers are designed with the goal to encapsulate recombinant antigens, DNA vaccines, bacterins, or attenuated pathogens (Adams, 2019; Dadar et al., 2017; Gomez-Casado et al., 2011; Heras, 2010; Tian et al., 2008; Xiao et al., 2011). For example, oral vaccination using Piscirickettsia salmonis previously inactivated and encapsulated into a proprietary biological matrix was effective in promoting antigen-specific IgM serum responses in rainbow trout, Atlantic salmon and coho salmon when used as a booster immunization after a first i.p. immunization with a commercial vaccine against P. salmonis (Tobar et al., 2015). In this study, IgM specific titers significantly increased after 400-1200 degree-days, while IgT titers were not evaluated. It is unclear whether this oral vaccine induced significant protective responses. Oral vaccination of olive flounder with alginate microparticles containing encapsulated formalin-inactivated Streptococcus parauberis induced IgM-specific titers in serum but not in gut mucus 3 weeks post-vaccination (Kim et al., 2020). Nevertheless, the immunization protocol resulted in an increase in the survival rate of 50% in small olive flounder fingerlings and 37.5% in bigger size fish (Kim et al., 2020). In a different study, an oral vaccine against IPNV using an alginate-encapsulated DNA vaccine was developed. In rainbow trout, this vaccine induced IgT and IgM⁺ B-cell recruitment into the digestive tract, especially in the pyloric caeca, 10 days post-vaccination (dpv), although vaccine-mediated protection was not evaluated (Ballesteros et al., 2013). Rainbow trout orally vaccinated with an IPNV DNA-VP2 gene encapsulated with sodium alginate induced increased IgT transcripts in the pyloric caeca and thymus 7 dpv, while IgM up regulation was only observed in the thymus (Ballesteros et al., 2012). In a different study with the same virus, oral vaccination of rainbow trout with a DNA vaccine encoding the VP2 gene encapsulated with CS-TPP (chitosan/tripolyphosphate) nanoparticles resulted in relative percentage survivals of 47-70%, depending on the vaccine dose used (Ahmadivand et al., 2017). Oral immunization of grouper larvae (Epinephelus coioides) with inactivated nervous necrosis virus (NNV) nanoencapsulated along with artemia nauplii, induced an increase in both IgM and IgT transcripts 1 and 4 weeks post-immunization in skin, gut and gill, while no protective studies were conducted (Kai et al., 2014). A different study, also with NNV, tested a nanostructured recombinant protein (NNV-C protein) vaccine in Senegalese sole (Solea senegalensis) administered i.p. or orally. This vaccine induced an increase of antigen-specific IgM titers in serum 14 and 30 dpv regardless of the route, however i.p. immunization induced ~1.5 fold more titers than oral immunization. In addition, both IgM and IgT transcripts were up regulated in the intestine at 3 dpv in the orally vaccinated group, while i.p vaccination induced an up regulation of IgM transcripts in head kidney. The previous report did not study potential protective responses induced by the vaccine, and thus, it is not possible to evaluate the functional relevance of those findings with regards to vaccine efficacy (Thwaite et al., 2020). In a different study, an oral non-encapsulated Photobacterium damselae bacterin vaccine was applied to gilthead seabream (S. aurata), and after 75 days fish were bath-challenge against P. damselae. Thus, the challenge induced an increase of serum IgM titers while high IgT titers were only detected in gut mucus. Protective responses were not evaluated (Piazzon et al., 2016).

Other strategies of oral vaccination include probiotic-based vaccines, as tested in common carp (Cyprinus carpio). In this type of vaccine delivery system, probiotics engineered to express G protein of spring viremia of carp (SVCV) induced high antigen-specific IgM titers in serum after 14, 28, 42 and 56 dpv. The vaccine reduced mortalities by 80% upon challenge with SVCV 42 dpv (Jia et al., 2020). Following a similar strategy, common carp larvae of 18 days post-hatch were orally immunized against cyprinid herpesvirus-3 (CyHV-3) using live Artemia previously fed with Saccharomyces cerevisiae expressing viral PORF65 tagged with GFP. This vaccine induced antigen-specific IgM titers in whole body homogenates as well as an up regulation of IgM transcripts in spleen and liver. Interestingly, IgT transcripts remained unchanged, while mortality was reduced by 30% at 6 weeks post-vaccination (Ma et al., 2020).

Additional strategies are being developed to potentially improve the effectiveness of oral vaccines. For example, in European seabass, recombinant TNF was used as adjuvant and orally delivered in combination with an inactivated V. anguillarum vaccine. This strategy resulted in the induction of IgT transcripts in the gut 28 and 171 dpv, the increase of IELs in hindgut 55 dpv and an increase of pathogen-specific IgM titers in serum 30, 45 and 155 dpv (Galindo-Villegas et al., 2013). Additionally, upon challenge with V. anguillarum 203, 300 and 376 dpv, 40-60% reduction in mortality rates was observed, thus indicating a protective effect of the vaccine. Prior to 2011, specific IgM titers had been reported to be induced in the gut in several vaccination studies using different oral delivery methods (reviewed by Salinas et al., 2011). Additional information on oral vaccination in fish has recently been reported in (Ma et al., 2019; Moges et al., 2020; Munang’andu et al., 2015b; Vallejos-Vidal et al., 2017).

6.2. Bath/immersion vaccination

Similar to oral vaccines, bath vaccines are especially attractive for aquaculture due to the easily-applicable, stress-free methodology used to deliver the vaccine. This type of immunization can potentially stimulate simultaneously all external MALTs, however it requires the use of high amounts of vaccine (Makesh et al., 2015). Most studies on the types of Ig responses induced by bath vaccination have only measured pathogen-specific IgM titers, thus probably missing the potential induction of specific IgT. For example, a study using a Y. ruckeri formalin inactivated and non-encapsulated vaccine delivered by immersion in rainbow trout induced serum specific IgM titers 4, 8 and 26 weeks post-vaccination. In addition, upon Y. ruckeri challenge, mortalities were shown to decrease by 40% and 28% after 8 and 26 weeks, respectively (Raida et al., 2011). In another study conducted in rainbow trout, a live non-attenuated F. psychrophilum bath immersion (2 × 10⁸ CFU ml⁻¹) strategy induced high pathogen-specific IgM titers in trout serum 26 and 47 dpv, reducing the associated mortality after i.p. challenge by 40% (Lorenzen et al., 2010). Similarly, in a different study also in rainbow trout, a live-attenuated F. psychrophilum immersion vaccine induced an increase of pathogen-specific IgM titers in serum after 14, 28, 42 and 56 dpv, while RPS was increased by 51-72% after challenging with wild-type F. psychrophilum (Ma et al., 2018). In a study conducted in E. coioides, fish were immunized by immersion, oral and intramuscular vaccination, using a recombinant capsid protein from nervous necrosis virus (NNV). After 14 and 28 dpv, specific IgM titers were induced in serum, being the highest titers those induced by immersion, followed by intramuscular and oral immunization routes. In addition, fish were challenged with NNV 30 dpv and RPSs of 40%, 30% and 20% were observed for the intramuscular, immersion and oral routes, respectively (Chien et al., 2018). In a different study, bath vaccination with formalin-killed Y. ruckeri in rainbow trout induced moderate specific IgM titers in serum after 3 months of vaccination, however a significant increase of IgM titers was observed 20 days after exposure to live bacteria. In addition, an increase in IgM⁺ B-cells in spleen was observed in fish that were bath-vaccinated three times during the three month trial period. No vaccine-mediated protective responses were evaluated in this study (Jaafar et al., 2018). In order to evaluate responses induced by bath vaccines, a number of studies have quantified IgM, IgD and IgT transcripts upon vaccination. For example, a study of bath-immunized rainbow trout with formalin-inactivated A. salmonicida was conducted with or without low frequency sonication. At 35 dpv, IgM, IgD and IgT transcripts were significantly up regulated (in a similar level for IgM and IgT and lower for IgD) in the gills of animals that were bath vaccinated using sonication, while no differences were detected in the non-sonicated treatment or i.p. vaccination. However, the opposite results were obtained in spleen, were bath and i.p. vaccination without sonication induced an increase in IgM, IgT and IgD transcripts, IgD responses being the largest, followed by IgM and IgT. These findings suggest that low frequency sonication can act as an adjuvant in fish mucosal tissues (Labarca et al., 2015), although it is unclear whether this strategy induces pathogen-specific titers. Furthermore, in a different study, bath immunization with inactivated betanodavirus in grouper (E. coioides) induced an up regulation of IgM and IgT transcripts in skin, gut and gill 1, 2 and 3 weeks post-immunization (Kai et al., 2014). In another study in turbot, IgM transcriptional responses were characterized after bath vaccination with live attenuated V. anguillarum, followed by challenge with wild type V. anguillarum. This regime induced the up regulation of both secreted and membrane IgM transcripts in spleen (3, 7, 28 dpv, and 1 day post-challenge (dpc)), gut (7, 21 and 28 dpv, and 1 dpc), kidney (3, 7 and 14 dpv and no up regulation after challenge), skin (3, 7, 14, 21, and 28 dpv and 1, 2 and 3 dpc) and gill (3, 7, 21 and 28 dpv and 2 and 3 dpc) (Gao et al., 2014). Ig titers were not evaluated. A further study also in turbot (S. maximus) has shown that formalin-inactivated V. anguillarum bath vaccination increased gene expression of IgT in gill, skin, spleen, head and trunk kidney, liver, foregut, midgut and hindgut at early (2-14) and late (21-28) dpv, while IgM transcripts were not studied (Tang et al., 2018).

Most of the immersion vaccination studies in fish deliver the antigen/bacterin without a carrier. However, the use of nanoparticles encapsulating the pathogen/antigen has also been tested in bath vaccines in an attempt to improve efficacy. For instance, a study in olive flounder (P. olivaceus) used a formalin inactivated vaccine against viral hemorrhagic septicemia virus (VHSV) encapsulated with PLGA nanoparticles and squalene as adjuvant. The vaccine was delivered twice by immersion or once by immersion followed by an oral booster, and fish were then challenged after 30 days with live VHSV. IgM titers increased in skin and gut mucus 48 h after the first immersion and booster vaccination. Such titers also increased 96 h post-challenge, although skin mucus titers were only quantified in the immersion/immersion treatment while gut mucus titers were evaluated in the immersion/oral treatment. In skin, both strategies triggered significant IgM transcriptional responses 48 h post-boost vaccination and 24, 48 and 96 h post-challenge. Gut IgM transcripts were up regulated 24, 48 and 96 h post-challenge in both immunization strategies whereas IgT expression was up regulated 48 h post-boost vaccination and 24, 48 and 96 h post-challenge. The immersion/immersion vaccine regime resulted in 60% protection whereas immersion/oral immunization led to 73.3% protection compared to 0% survival in control fish (Kole et al., 2019).

In conclusion, bath vaccination can stimulate sIg responses at multiple fish MALTs and several modifications in the mode of delivery may boost bath vaccine efficacy. Whether the potency of the sIg responses induced by bath vaccines in each MALT are similar or different, is an important aspect that remains to be investigated. Further information on bath vaccination strategies can be obtained in (Bøgwald and Dalmo, 2019).

6.3. Anal vaccination

Anal immunizations are not as commonly used as other mucosal delivery routes, while anal cannulations are more commonly employed for bacterial challenges rather than for immunization (Torrecillas et al., 2017). Up regulation of sIgT and sIgD transcripts in rainbow trout gut were recorded 28 days post anal vaccination with an attenuated F. psychrophilum strain, although IgM specific titers in serum could not be detected and IgT responses were not studied. In this study, i.p. immunized fish showed an increase of total IgM in serum, gill mucus and gut mucus, however, specific titers were not evaluated (Makesh et al., 2015). With the exception of the aforementioned study, all anal vaccination studies have focused on IgM where the most common read-out is IgM specific titration in serum. In a different study, anal immunization of trout with a thymus independent antigen (TNP-LPS) induced an increase in serum specific IgM titers as well as a significant increase in IgM⁺ B-cell proliferation in the spleen, and to a lesser extent in the head-kidney, at 30 dpv. In contrast, no IgM⁺ B-cell proliferative responses could be detected in gut (Martín-Martín et al., 2020). Other studies in rainbow trout have shown that anal immunization with formalin-inactivated Y. ruckeri did not induce specific IgM titers in serum 6 months after immunization or after 7 dpc, while a 60% survival increase was observed in vaccinated animals. Thus, these findings suggested that other Igs or protective mechanisms are responsible for the observed protection (Villumsen et al., 2014). Nevertheless, in another study in rainbow trout, anal immunization with FITC-KLH induced specific serum and gut mucus IgM titers against KLH 4 weeks post-vaccination, being the response higher when FITC-KLH was combined with Freund’s complete adjuvant (Cain et al., 2000). In a model using genetically modified plants that expressed recombinant viral antigens, common carp were immunized through the anal route with the LTB-influenza-parvo fusion protein, a strategy that resulted in the induction of antigen-specific IgM titers in serum (Companjen et al., 2006). Combined, anal immunization studies have yielded variable results depending on the model system. Importantly, gut IgT responses have been overlooked and therefore it is still unknown whether anal immunization induces antigen-specific IgT responses in the gut when compared to other immunization routes.

6.4. Nasal vaccination

Nasal vaccination has been shown to be an effective vaccine delivery method to control infectious diseases in higher vertebrates (Tlaxca et al., 2015; Yusuf et al., 2017), where it has shown to typically induce protective IgA responses in mammals (Yanagita, Hiroi et al. 1999) and birds (Yanagita et al., 1999; Zhao et al., 2016). In addition, a lower antigen dosage is required to induce an immune response in the NALT when compared to other delivery routes (Bernocchi et al., 2017). Since NALT was discovered in teleost fish (Tacchi et al., 2014), this opened up the possibility of delivering vaccines intranasally in these species (Dong et al., 2020; LaPatra et al., 2015; Magadan et al., 2019b). In that regard, studies in rainbow trout have demonstrated the efficacy of nasal vaccination in reducing fish mortalities and/or in stimulating several immune parameters, including recruitment of CD8 T-cells, increased expression of chemokines, interleukins and AMPs (Dong et al., 2020; LaPatra et al., 2015; Salinas et al., 2015). Nasal vaccines do not circumvent the need to manipulate individual fish but offer many advantages including their high efficacy, immediate neuronal-mediated immune protection and the fact that no adjuvant is required (reviewed by (Das et al., 2020)). Thus far, it is not known if nasal vaccination induces antigen-specific IgM, IgT or IgD responses in the nasal mucosa. At the transcript level, intranasal delivery of live attenuated IHNV vaccine in rainbow trout resulted in increased IgM expression levels in NALT 14 dpv and moderate increases in IgD and IgT transcripts (Tacchi et al., 2014). Interestingly, nasal vaccination with Y. ruckeri bacterin in rainbow trout resulted in a significant modification of the IgM (IgH μ) repertoire in trout spleen, together with a lower diversification and higher relative IGHV2 usage of IgT repertoires. Importantly, perturbations of the IgT and IgM nasal B-cell repertoire were observed in trout i.p. vaccinated with Y. ruckeri bacterin. However, nasal vaccination resulted in weak IgM-specific titers in serum while no IgT specific titers could be detected despite the fact that at this time point (28 dpv) nasal vaccination of trout with this vaccine has been proven to be highly protective (Magadan et al., 2019b). Whether nasal vaccination results in specific sIg responses in other MALTs has not yet been investigated. However, mammalian studies have revealed that nasal vaccines can protect far distant mucosal sites (Bernocchi et al., 2017; Neutra and Kozlowski, 2006) and therefore this possibility cannot be rule out in teleosts. Further information on nasal immunity and vaccination in fish has recently been reviewed in (Das et al., 2020).

7. Concluding remarks and future avenues

The body of work here summarized highlights a breadth of mucosal immunoglobulin studies in aquacultured finfish species as well as in model species such as zebrafish. Since the discovery of the role of IgT ten years ago as an Ig specialized in mucosal immunity, studies on fish sIgs have blossomed. Since most pathogens need to trespass mucosal barriers in order to infect fish, understanding how sIg responses develop in these sites is crucial if we are to develop vaccines and other immunotherapeutics that stimulate protective immunity. Moreover, all of these large mucosal areas, including the skin, are populated by a vast and diverse population of microbiota. Thus, it is imperative to understand the mechanisms by which sIgs control and maintain microbiota homeostasis if we want to develop strategies (i.e., probiotics, immunobiotics) that will promote healthy mucosal barriers.

With regards to the interactions between fish sIgs and the microbiome, such studies have only begun to be explored in our field. We recently demonstrated the dual specialization of sIgT in protection of mucosal sites from pathogens and preservation of microbiota homeostasis. Since only the effects of IgT coating have been evaluated, the biological consequences of microbiota coating by other sIg are yet to be determined. Moreover, the reported presence of microbiota double or triple coated with IgT, IgM and/or IgD deserves further investigation.

How different environmental perturbations affect the equilibrium between fish sIgs and microbiota, two key elements of host mucosal health, is only beginning to be elucidated. Perturbations in sIgT result in dysbiosis and bacterial breaching of the mucosal barriers. Thus, fish pathogens that hijack or harm the sIg system and/or mucosal barriers are likely doubly harmful to the fish host, since they will also cause dysbiosis and potential opportunistic infection. Identifying environmental factors (ie. diet, water quality, chemical signals from conspecifics, toxicants) and interventions (i.e husbandry practices, dietary formulations, probiotics etc) that prevent sIg imbalances is therefore critical for fish health. This would require large field studies which have to yet be carried out. Moving the field forward clearly will entail novel fundamental immunological discoveries that will enable the development of translational interventions in the field of fish diseases.

From the wealth of studies reviewed herein, most of the reports analyzing titers and the binding capacity of sIgs to pathogens or microbiota support a key role of sIgT in mucosal immune responses to both types of microbes, while sIgM appears to play a less prevalent role. Another important point that arises from these studies is that pathogen-specific IgD responses are not induced upon infection. In contrast, a number of studies have confirmed that IgD coats a minor, but significant portion of the microbiota of several fish MALTs, and thus, IgD might contribute to the control and homeostasis of the microbiota. While most studies have been performed using parasitic or bacterial pathogens, very few have used viruses or fungi, and at this point it is not known whether these pathogens can induce specific mucosal IgT responses. Another aspect that remains poorly studied pertains the effector functions played by sIgT. Whether this is an Ig that can activate complement at fish MALTs is a critical point that remains to be assessed. In that regard, it is worth noting that mammalian sIgA is considered an anti-inflammatory Ig in part because it cannot activate complement at mucosal surfaces, a property that is critical to avoid harmful inflammation at these sites. Whether sIgs have the capacity to agglutinate or neutralize pathogens, and whether sIgT binding to the pathogen has opsonizing capacity remain unanswered.

With regards to the development of mucosal vaccines, it remains to be well studied which are the best immunization routes that induce antigen-specific mucosal sIg responses, although some recent strategies of antigen encapsulation and delivery to mucosal sites have shown promise in inducing protective responses. Future studies will have to develop more effective strategies of both antigen delivery as well as better antigen carriers that protect the antigen from being degraded before reaching the targeted mucosal site. At the same time, there is a need to develop effective mucosal adjuvants that potentiate both cell-mediated and humoral adaptive mucosal immune responses as such adjuvants are clearly missing for fish mucosal vaccines. Moreover, it is pivotal to evaluate the existence of a common mucosal immune system (CMIS, see figure 2), with the goal to examine which of the MALTs is most efficient to target for mucosal vaccination. This requires to expand our current understanding of the biology of induction and effector sites in teleost MALT and to elucidate whether and which of the MALTs contain some sort of organized or semi-organized lymphoid tissue functionally analogous to that of tetrapod MALTs (e.g., mesenteric lymph nodes, Peyer’s patches, tonsils). This will provide the necessary knowledge to improve fish vaccines and other immunotherapeutics.

A critical consideration in the study of teleost sIg responses to pathogens, vaccines and microbiota is the fact that most studies use changes in Ig transcript levels as a proxy for antigen- or pathogen-specific antibody responses. However, increases in Ig transcripts may not correlate with increases in antigen/pathogen-specific Ig titers as evidenced by a few recent studies. Further, many pathogens may induce non-specific polyclonal responses, thus activating B cells in a polyclonal fashion, akin to that shown for a number of PAMPs, such as LPS in mammals. In addition, increases in Ig transcripts may be the result of changes of the microbiome as result of the infection. Moreover, it is also well known that changes in the microbiome composition within mucosal surfaces can also induce the activation of mucosal B cells and production of sIgs. Finally, primers employed in many reports to amplify fish Ig transcripts do not distinguish between sterile and productive transcripts, an overlooked point in the field. Hence, further efforts need to be devoted to develop monoclonal antibody reagents to detect specific Igs from the most relevant aquaculture and model fish species.

As Comparative Immunologists, the diversity of adaptive immune systems in teleosts gives us the unique opportunity to unravel unique mucosal immune systems, for instance, mucosal immune systems that operate in the absence of IgT. This diversity cannot be found in other vertebrate groups like mammals and offers the possibility of unveiling unique immunological innovations to the problem of microbiota homeostasis and pathogen control. In summary, the field of teleost mucosal Igs can provide fundamental insights into the biology of vertebrate mucosal immune systems while improving fish health and welfare in aquaculture.

Highlights.

• Teleost sIgs are critical for mucosal homeostasis

• Seven different teleost mucosa-associated lymphoid tissues (MALTs) have been identified in fish

• Teleost sIgT promote microbiota colonization and their containment within the mucosa

• Upon infection or vaccination most studies indicate a predominant role for IgT in teleost MALTs

• Not all teleosts have IgT, thus suggesting alternative mucosal immune systems