MHC and adaptive immunity in teleost fishes
The adaptive immune system has long been considered a key evolutionary innovation of the vertebrates, the product of two rounds of genome duplication that gave rise to the raw material necessary for the evolution of a highly specific immune response and immune memory. While comparative studies of a small number of model organisms have led to the commonly held view that the adaptive immune system has remained relatively static since its origin, recent studies of non-model organisms are challenging this notion, highlighting the fact that we have only begun to scratch the surface in terms of our understanding of immune system diversity. Some of the most exciting recent results have come from the comparative analysis of teleost fishes, a group that includes more than 40% of vertebrates, and shows remarkable diversity in immune system structure and function. Despite the repeated loss of key components of the adaptive immune machinery in this group, affected species are capable of mounting a robust response to immune challenge, suggesting that they have evolved alternative mechanisms of immune protection. Such deviations from the canonical model of vertebrate immunity create opportunities to explore common paradigms of immune function, and may contribute to new experimental approaches and methods of treatment.
KeywordsComparative genomics Evolutionary immunology Immune function Major histocompatibility complex Osteichthyes
Key components of innate and adaptive immunity in mammals and bony fish (after Sunyer 2013)
Osteichthyes (bony fish)
Secondary lymphoid organs
Spleen, thymus, lymph nodes
Macrophages, dendrites, neutrophils
Toll-like receptors (TLR) (Palti 2011)
TLR1–3, 5, 8–9, 14, 18–23
Complement system (Sunyer et al. 2003)
Interleukins (Secombes et al. 2011)
Chemokines (Alejo and Tafalla 2011)
HLA-E/-F/-G, CD1, MICA/B
U, Z, L, S, P
TCR, CD4, CD8
IgM, IgG, IgA, IgD, IgE
IgM, IgD, IgT, IgZ
Recent studies of immune system structure and function in non-model organisms have been increasingly challenging the consensus view that the adaptive immune system has remained relatively static since its origin in the common ancestor of vertebrates. While the field of evolutionary immunology is still in its infancy, examples of the loss of key components of immune memory in multiple lineages of teleost fish (Star and Jentoft 2012; Haase et al. 2013), novel antibody structure in sharks and camels (Flajnik et al. 2011), cross-talk between the innate and adaptive immune systems via natural killer cells (Rölle et al. 2013), and the discovery of alternative forms of adaptive immunity in more distant evolutionary lineages (Kurtz and Armitage 2006; Herrin and Cooper 2010; Barrangou and Marraffini 2014) emphasize that we have only begun to scratch the surface in terms of our understanding of immune system diversity. Exceptions to the canonical model of vertebrate immunity offer a unique window into the evolutionary process, and can contribute to a more nuanced understanding of immune system structure and function.
The genes of the MHC are recognized as an essential component of the vertebrate adaptive immune system, and are responsible for the recognition and presentation of foreign antigens. MHC loci have traditionally been subdivided into classical loci, which have a well-characterized function involving the presentation of antigen epitopes to T cell receptors, and non-classical loci, many of which have not yet been fully characterized (Kaufman et al. 1994). Classical MHC loci show high polymorphism and broad expression domains, while non-classical loci typically show low variability and tissue-specific expression. MHC loci experience high turnover rates (Nei et al. 1997), and phylogenetic reconstructions of the mammalian MHC suggest that non-classical loci have evolved multiple times independently following the degradation of duplicated classical genes (Hughes and Nei 1989).
Classical MHC class I molecules (MHC I) are located on all nucleated cells and are activated following the binding of antigens synthesized within the host (e.g., viruses). These proteins are then presented on the cell surface to cytotoxic CD8 T cells (CTLs), which become activated and destroy the infected cell, a process known as cell-mediated specific immunity (Barber and Parham 1993).
Classical MHC class II molecules (MHC II), in contrast, are restricted to professional antigen-presenting cells (APCs). The presentation of pathogen-derived antigens by MHC II on the surface of APC B cells and phagocytes is essential to the elicitation of CD4 T cell (TH) binding (i.e., humoral immunity), which activates B cell differentiation into plasma cells, producing antibodies specific to the invading pathogen, and memory cells, preserving a record of past infection. TH cells themselves then differentiate into effector cells, which activate B cells to produce cytokines and memory cells. Memory B and TH cells allow the body to respond more rapidly to secondary infection, promoting a higher affinity and accelerated immune response, and providing enhanced immunoprotection (Ahmed and Gray 1996). MHC II-dependent immune memory is considered a hallmark of the adaptive immune response (Flajnik and Kasahara 2009).
The roles of MHC I and II molecules have traditionally been thought to be clearly delineated to the presentation of endogenous, and exogenous, antigens, respectively. However, MHC I molecules can also recognize and present exogenous antigens to cytotoxic T cells through a process known as cross-presentation (Ackerman and Cresswell 2004). While this process was once thought to be restricted to phagocytes, evidence of cross-presentation by MHC I in B cells has attracted considerable research interest, due its potential as a target for a new generation of vaccines (Basta and Alatery 2007).
Gene and genome duplications have long been recognized as an important source of genetic variation (Ohno 1970; reviewed in Taylor and Raes 2004). The MHC region is thought to have emerged as a result of two rounds of whole-genome duplication early in vertebrate evolution, which provided the raw material necessary for the evolution of the adaptive immune system (the “big bang” theory of adaptive immunity; Kasahara 1997; Schluter et al. 1999; Flajnik and Kasahara 2009). Gene rearrangement and loss following these whole-genome duplications ultimately led to the tight clustering of major histocompatibility loci in modern tetrapods (Kelley et al. 2005).
MHC I and II genes are among the most variable loci in the vertebrate genome (Reche and Reinherz 2003; Kelley et al. 2005). The peptide-binding regions (PBR) of classical MHC loci interact directly with antigens, and the unique distribution of genetic variation at PBR sites is one of the textbook examples of balancing selection (Hughes 2007). MHC genes were initially identified as “immune response” loci due to their high specificity, with antibody responses to particular antigens dependent on MHC genotype (Snell 1948). Due to the highly targeted nature of MHC-mediated immunity, any given individual carries protection against only a small subset of potential pathogens. This high degree of immune specificity emphasizes the fact that innate immunity remains essential to the immune response of higher vertebrates, particularly when an individual’s adaptive immune system is unable to effectively recognize and present a specific pathogen.
While the structure and function of MHC loci have been well-explored in a small number of mammalian models, we still have a relatively poor understanding of immune function outside a restricted number of non-mammalian vertebrates that have been studied largely in isolation. As a consequence, efforts to reconstruct the evolution of adaptive immunity in vertebrates (e.g., Flajnik and Kasahara 2009; Cooper and Herrin 2010; Sunyer 2013) have been hampered by a lack of comparative data, which has contributed to the perception that the adaptive immune system has been essentially static since its origin 500 million years ago. Below, I highlight some of the unexpected recent insights gained from the study of immune function in teleost fish, and argue that deviations from the canonical immune system offer a unique opportunity to test long-held hypotheses concerning adaptive immune function.
Fish evolution and the MHC
Fish are the most diverse group of vertebrates, with more than 1000 species of cartilaginous fish (Chondrichtyes) and close to 32,000 bony fishes (Osteichthyes), including the lobe-finned fishes (Sarcoptergyii), the closest evolutionary relatives of modern tetrapods (mammals, amphibians, and reptiles) (Froese and Pauly 2010). More than 99% of bony fish diversity is found in the ray-finned fishes (Actinopterygii), a group that originated ca. 400 million years ago (Near et al. 2012), and gave rise to the more than 31,000 species of modern teleosts (Froese and Pauly 2010), following a lineage-specific genome duplication (Meyer and Van de Peer 2005). The exceptional biodiversity of modern teleosts has been attributed in part to the genetic raw material generated by this early genome duplication (Meyer and Van de Peer 2005; Santini et al. 2009).
The whole-genome duplication in the ancestor of modern teleosts had immediate effects on immune diversity in the group, doubling the number of loci associated with immune function, and creating the opportunity for both sub- and neo-functionalization. The unique complements of toll-like receptors (Palti 2011), immunoglobulins (Hsu et al. 2006), and major histocompatibility loci (Dijkstra et al. 2013) in modern teleosts are thought to reflect a complex evolutionary history of gene duplication, rearrangement, and loss in this group. Comparative genomic analyses suggest that this early whole-genome duplication was instrumental in breaking the genetic linkage between MHC class I and II loci in early teleosts (Palti et al. 2007; Dijkstra et al. 2013), allowing the independent segregation of these two key components of the adaptive immune system. This fundamental deviation from the tight physical linkage of the MHC in most vertebrates means these loci are not, strictly speaking, a complex, which has led to the alternative nomenclature of MH for this group (Stet et al. 2003).
A complex history of gene and genome duplications in teleosts has generated extraordinary variation in MH structure and diversity in this group (Grimholt 2016). While comparative analyses of MH variation are complicated by high levels of recombination and gene conversion among loci, the accumulation of whole-genome sequences for a variety of teleost species has provided unique insights into the evolutionary history of the MH in this group. High-quality genome assemblies based on deep sequencing, coupled with expression data from a broad range of tissue types, have been used to reconstruct the evolutionary history of the MH I (Grimholt et al. 2015) and MH II (Dijkstra et al. 2013) gene families, and to explore novel aspects of MH structure and function in teleosts.
MH I: ancient divergence and unresolved function
All classical loci fall into the U lineage, which also includes a variety of non-classical loci that share sequence homology, but lack the stereotypical peptide-anchoring residue motif found in human HLA-A2 (Grimholt et al. 2015). A phylogenetic reconstruction of the alpha 1 domain region of U lineage loci suggests that much of the variation in this lineage is old, reflecting evolutionary events dating from the teleost-specific genome duplication, a pattern which differs from the high turnover rates observed in mammals (Hughes and Nei 1989). Interestingly, a number of species deviate from this pattern, and show dramatic expansions of the U lineage (Fig. 1). U lineage loci from both the stickleback (Gasterosteus aculateus; 29 loci) and the Atlantic cod (Gadus morhua; >100 loci) form tight monophyletic groups, consistent with recent expansions in these species. The unique distribution of genetic variation at the U lineage loci of the cod has led to the suggestion that this species may employ a novel mechanism of MH I presentation (Star and Jentoft 2012; Grimholt et al. 2015), an idea that will be explored in greater detail below.
All teleosts also carry at least one gene from the Z lineage (Fig. 1), which consists of non-classical loci in which alpha 1 and 2 domains are conserved, while the alpha 3 domain is highly variable. Interestingly, despite the ubiquity of Z lineage genes in fish, these loci appear to be lacking in tetrapods (Grimholt et al. 2015). Here again, patterns of sequence divergence at Z lineage loci suggest that much of the genetic variability in this lineage reflects evolutionarily ancient events, and the remarkable conservation of amino acids making up the PBR of Z lineage loci suggests that the majority of Z lineage molecules bind a similar, as yet unidentified ligand. A smaller group of atypical Z loci lacks the anchoring residues characteristic of a functional peptide-binding region, and appear to have an as yet undermined function.
The remaining MH I lineages of teleost fish (S, L, and P) all appear to lack peptide-binding ability, and while phylogenetic analysis suggests that all three of these lineages are ancient in origin, their patchy distribution across the teleost phylogeny suggests that these non-classical loci have been frequently lost during the evolution of the group. While expression data indicate that these genes are expressed at low levels in vivo, the function of these loci remains unknown (Grimholt et al. 2015).
MH II: out and about without a chaperone?
In contrast to MH I, MH II molecules are heterodimers composed of separately coded IIα and IIβ loci that interact to form the peptide-binding groove involved in the binding of extracellular antigens. After assembly in the endoplasmic reticulum, MH II molecules must move through the cytoplasm to a specialized endosome known as the MH class II compartment, and are protected during transport by a specialized invariant chain protein, which binds to the MH II PBR and prevents the binding of intracellular ligands (Rocha and Neefjes 2008). Once in the endosome, the invariant chain is degraded, and a dedicated molecular chaperone removes the active component of the invariant chain from the PBR, stabilizes the molecule, and facilitates the binding of peptides generated by the degradation of extracellular proteins. This molecular chaperone (HLA-DM in humans, H2 in mice) is itself a non-classical MH II molecule, which is highly conserved in all tetrapods (Dijkstra et al. 2013), emphasizing its central role in MH II function.
The recent analysis of MH II loci in the genomic sequences of teleosts by Dijkstra et al. (2013) has uncovered a remarkable difference between the MH II system of teleosts and that all other vertebrates. While teleosts carry classical MH II and invariant chain genes, they lack the DM molecular chaperone, and classical loci show variability at key amino acid sites associated with chaperone interaction in tetrapods, indicating that teleosts must employ an alternative to the tetrapod DM molecular chaperone system (Dijkstra et al. 2013).
While teleosts have apparently lost the DM system, they maintain both classical and non-classical loci, and phylogenetic analysis indicates that the three major lineages of MH class II genes (DA, DB, and DE) are evolutionarily ancient, with the divergence of classical DA loci and non-classical DB loci dated to the time of the teleost-specific genome duplication. Interestingly, while teleost MH II loci form a monophyletic group, classical and non-classical loci share a variety of characteristics with their tetrapod counterparts, suggesting that despite high genic turnover in evolutionarily divergent lineages, similar functional constraints may be driving their evolution.
MH II-mediated immune memory is considered an essential component of the vertebrate adaptive immune system, but the recent discovery of species that lack the MH II/CD4 pathway associated with specific immunity and immune memory suggests that vertebrate immunity is far more dynamic than was once thought. Both gadiform (cod and its allies) and syngnathiform fishes (seahorses, pipefish, and seadragons) have independently lost classical MH II loci (Fig. 1), yet these species nonetheless exhibit a robust response to immune challenge, suggesting that they may have evolved alternative mechanisms of immune protection. I discuss our current understanding of immune system structure and function in these species below.
Atlantic cod: evolutionary novelty of immune system structure and function
The Atlantic cod (G. morhua) was the first vertebrate species found to lack MH II activity (Pilström et al. 2005), and the recent completion of its full genome sequence indicates that the cod has lost MH II, CD4, and invariant chain genes while experiencing dramatic expansions of MH I and Toll-like receptor (TLR) families (Star et al. 2011). Transcriptome screening of related gadoid fishes indicate that immune architecture has remained stable in this group, suggesting that the loss of MH II activity in this group was an ancient event that predated its diversification (see below).
While the majority of teleost fishes show a robust response to bacterial infection, with pronounced activation of humoral and cell-mediated specific immunity as well as a full complement of phagocytic cells, antimicrobial peptides, and natural antibodies, produced independent of specific infection (Uribe et al. 2011), MH II-deficient G. morhua fail to show a humoral immune response following bacterial infection (Pilström et al. 2005), despite normal levels of B cell activity (Rønneseth et al. 2007) and concentrations of natural antibodies that are among the highest detected in a marine fish species (Magnadóttir 1998). Phagocyte titers in the head kidney and spleen of cod exceed those detected in fish with a fully intact immune system (Rønneseth et al. 2007).
The recent discovery that MH I molecules may play a role in cross-presentation (Basta and Alatery 2007) suggests a possible alternative mechanism of immune protection in the cod. MH I recognition of extracellular antigens could offer a degree of immune protection in species such as the cod that have lost the ability to produce MH II. Intriguingly, TLR, an important component of the innate immune system involved in the detection of microbial peptides, have also been implicated in cross-presentation (Datta et al. 2003), suggesting a possible connection between innate and adaptive immunity, and a mechanism through which MH I/CTL activity could compensate for the loss of the MH II/TH pathway. The expansion of both TLR and MH I families (Star et al. 2011) and the evolution of MH II-like functionality in MH I genes of the cod (Malmstrøm et al. 2013) are consistent with this hypothesis. Genes of the Tlr21 (12 loci) and Tlr22 (2 loci) families are the only cell-surface TLRs in the cod (Sundaram et al. 2012), and are thus ideal candidates for potential cross-presentation activity.
G. morhua fails to produce a specific antibody response upon primary infection, despite high levels of Ig (Pilström et al. 2005), and lacks antibody-mediated immune memory upon secondary infection (Mikkelsen et al. 2011). Interestingly, while vaccination against Vibrio anguillarum in G. morhua had no impact on antibody production, interferon and interleukin expression were enhanced, even 50 days after treatment (Mikkelsen et al. 2011). It remains unclear whether this extended period of immune activation reflects a protracted and persistent response to primary infection, or whether the cod relies on a CD4-independent source of immune memory, a result which would have important implications for our understanding of adaptive immune function.
Syngnathid fishes have independently lost MH II
The recent transcriptome profiling of the pipefish (Syngnathus typhle) has revealed an independent example of loss of the MH II pathway (MH II, invariant chain, CIITA, and CD4/8b) (Haase et al. 2013), but in contrast to the pattern found in the cod, these genes are present and functional in the closely related seahorse (Hippocampus abdominalis) (Bahr and Wilson 2012; Bahr et al. 2012), suggesting an independently derived and evolutionarily recent genetic event (Fig. 1). The MH II gene of H. abdominalis exhibits high levels of PBR polymorphism and broad expression (Wilson et al. 2014), consistent with its function as a classical MH II locus.
The recent whole-genome sequencing of the tiger tail seahorse (Hippocampus comes; Lin et al. 2016) and the gulf pipefish (Syngnathus scovelli; Small et al. 2016) is consistent with the results of earlier transcriptome work and suggests that the loss of MH II function is a trait likely shared by all Syngnathus pipefish. This clear difference between Hippocampus seahorses, which show a minimal functional classical MH II system, and Syngnathus pipefish, which have completely lost the MH II pathway, raises questions concerning the timing of the loss of MH II in this group, and whether the loss of this pathway is associated with the expansions of other components of adaptive and innate immunity as in the cod. Pipefish and seahorses are both members of the family Syngnathidae, a diverse group of close to 300 species, offering an opportunity to explore the genetic architecture of adaptive immunity at a family-wide level.
While experimental investigations of immune function in syngnathid fishes are in their infancy, studies of wild-caught animals have yielded intriguing results concerning immune activation in Syngnathus pipefish (e.g., Roth et al. 2012a; Birrer et al. 2012; Beemelmanns and Roth 2017). Plasma isolated from wild-caught animals collected from across the species range of S. typhle showed stronger antimicrobial activity against sympatric bacteria, consistent with a specific immune response to local pathogen communities, but individuals failed to show an enhanced immune response following repeated exposures, consistent with an absence of immune memory in this species (Roth et al. 2012a). This research has been extended to demonstrate that the previous immune experience of both parents influences offspring immune function (Roth et al. 2012b), an effect that has recently been shown to extend to grand-offspring (Beemelmanns and Roth 2017).
Comparative genomics sheds light on immune structure and function
In an example of the potentially transformative power of high-throughput genomics, Malmstrøm et al. (2016) recently examined macroevolutionary patterns of immune gene diversity through the de novo sequencing of 66 teleost species, a substantial addition to the ten high-quality genomes that were previously available for this group. The authors specifically set out to test the compensatory hypothesis of Star et al. (2011), which proposed that expansions of MH I in the cod could represent a potential compensatory response to the loss of MH II in this species. Genome-level comparisons allowed Malmstrøm et al. (2016) to determine that the loss of MH II, CD4, and invariant chain genes occurred ca. 100 million years ago in the common ancestor of gadiform fishes, a family of >600 species inhabiting a wide variety of aquatic environments. While the majority of gadiforms showed evidence of classical MH I gene expansions, gene expansions were also found in a variety of fish species with an intact MH II system.
Remarkably, Bregmaceros contori, the most basal gadiform included in the study, was inferred to have lost not only MH II, but also lacked classical MH I loci, an observation that led the authors to conclude that expansions of MH I in other gadiform lineages occurred after the loss of MH II, consistent with the hypothesis of immune compensation (Star et al. 2011; Star and Jentoft 2012). While the identification of a free-living vertebrate lacking both MH I and II functionality is a potentially revolutionary discovery, and the expansion of sequencing resources for >60 teleost species opens the door for a wide range of comparative genomic analyses, it is important to note that the Malmstrøm et al. (2016) dataset was based on low coverage sequencing, and the genome assemblies are consequently rather incomplete. Deep sequencing and functional analyses of a cross-section of species included in this analysis, especially the enigmatic B. contori, will be necessary to confirm these intriguing results.
The increasing accessibility of genomic data for a broad range of vertebrates is at the root of the nascent field of evolutionary immunology, which focuses on understanding the ecological and evolutionary context of the immune response via the comparative analysis of individuals, populations, and species (Cooper and Herrin 2010; Boehm 2011).
While recent comparative analyses of immune system structure and function in teleosts have drawn attention to unique aspects of immune structure in this group, careful experimental studies will be needed to clarify the functional implications of these deviations from the canonical model of vertebrate immunity. The importance of MH I and II to teleost adaptive immunity has been clearly demonstrated through research identifying associations between individual alleles and pathogen resistance (Grimholt et al. 2003; Wedekind et al. 2004), but we still have no idea how MH II molecules move from the endoplasmic reticulum to the endosome, or how the exchange of invariant chain proteins for antigenic peptides takes place in the absence of a DM analog. Equally intriguing, while species lacking a functional MH II locus are capable of mounting a robust response to immune challenge, the mechanisms underlying this functionality remain almost entirely unknown. These observations are not only biologically interesting, but have the potential to lead to novel methods of treatment through the identification of alternative forms of immune protection. Paying attention to the myriad ways in which evolution has solved the problem of immune protection is certain to lead to unique perspectives, and a more nuanced appreciation of immunological diversity.
I am grateful to Martin Flajnik, Unni Grimholt, and Oriol Sunyer for illuminating discussions on the incredible diversity of vertebrate immunity.
- Grimholt U (2016) MHC and evolution in teleosts. Biology (Basel). doi: 10.3390/biology5010006
- Magnadóttir B (1998) Comparison of immunoglobulin (IgM) from four fish species. Icel Agr Sci 12:47–59Google Scholar
- Uribe C, Folch H, Enriquez R, Moran G (2011) Innate and adaptive immunity in teleost fish: a review. Vet Med 56:486–503Google Scholar