INTRODUCTION

Recent evidence suggests that the evolutionary emergence of vertebrates was accompanied by a sudden change in the structure and function of the ancient metazoan immune system (15, 18, 25, 40, 41, 52, 83, 109). To highly sophisticated but stereotyped innate defenses, vertebrates added adaptive immunity, which bestowed them with antigen-specific memory, leading to more rapid and efficient secondary immune responses. Adaptive immunity emerged twice in the early phases of vertebrate evolution; this remarkable example of convergent evolution not only attests to the unique selective advantage of adaptive immunity for vertebrates, but also provides immunologists with the unprecedented opportunity to identify the general principles governing the structure and function of adaptive immunity and its coevolution with innate defenses.

In anticipation of the imminent completion of whole-genome projects for many lower vertebrates, this review summarizes recent findings on the immune systems of jawless and jawed fishes to facilitate the discovery of common principles of innate and adaptive vertebrate immune systems and the concomitant recognition of species- or group-specific idiosyncrasies. It is anticipated that this type of genomic immunology will have an immediate impact on the understanding of mammalian immune systems and might provide new avenues for intervening in faulty immune functions in humans.

VERTEBRATE PHYLOGENY

Extant vertebrates comprise two groups: jawless vertebrates and jawed vertebrates. The jawless fishes (Cyclostomata) are a monophyletic group (51) of approximately 100 species of lampreys and hagfish (63); the extant jawed vertebrates (Gnathostomata) comprise approximately 6 × 10⁴ species (88). The available evidence from genome analyses of selected species suggests that the evolutionary origin of vertebrates was associated with whole-genome duplication events (103), although their precise timing is controversial (74). One hypothesis posits that jawless vertebrates diverged from the chordate lineage after a whole-genome duplication event and that the ancestor of all jawed vertebrates had experienced two genome duplications (66); in the stem lineage of ray-finned fishes, an additional genome duplication event took place, giving rise to up to eight copies per gene of the metazoan core genome complement (62) (Figure 1). Not unexpectedly, however, over the long period that has elapsed since the emergence of the different branches of vertebrates, many genes have experienced alteration of the 1:2, 1:4, and 1:8 relationships, either through loss of some or all paralogs or through further gene-specific amplification events. In a recent spectacular example, cartilaginous fishes were shown to have lost one of the four Hox clusters, HoxC, previously thought to be a primitive and constitutive feature of jawed vertebrates (70).

Nonetheless, the characteristic long-range syntenic relationships found in vertebrate genomes still often reflect the structure of the ancestral vertebrate genome; these genomic signatures represent important landmarks for understanding the origin of novel genes characteristic of vertebrate adaptive immune systems. For instance, the major histocompatibility complex (MHC) is embedded in a chromosomal context that occurs in four paralogous regions in the typical vertebrate genome (65). Yet the genes encoding key components of the MHC-dependent pathway of antigen presentation [particularly the MHC class I (MHCI) and class II (MHCII) genes] are found in only one paralogous segment, suggesting that they represent singular evolutionary innovations that emerged in the ancestor common to all jawed vertebrates.

THE MORPHOLOGICAL AND MOLECULAR BASIS OF VERTEBRATE IMMUNE SYSTEMS

Vertebrate-specific genetic innovations were accompanied by the emergence of novel cell types, such as lymphocytes, dendritic cells, and immune-related tissues such as thymus and spleen (Figure 2). As a result of over 500 million years of coevolution, a key aspect of vertebrate immune systems is the functional cooperation of innate and adaptive cell types to orchestrate effective immune responses (61). Not surprisingly, many of the molecules that are required for the crosstalk between innate and adaptive immunity are already present in the most basal vertebrates. For instance, lampreys possess a homolog of the gene encoding the macrophage migration inhibitory factor, which in mammals is known to be inducibly expressed by T cells and myelomonocytic cells during an inflammatory response (122); the same is true for interleukin 17 (IL-17) (141), a key proinflammatory cytokine, and IL-8 (96), a classical chemoattractant for leukocytes. The inventory of cytokines, interleukins, chemokines, and their receptors in fishes has been deduced from genome comparisons and, to a lesser extent, transcriptome data (2, 3, 44, 80, 82, 85, 99, 125, 147). Although a detailed appraisal of these studies is beyond the scope of this review, the available information suggests the universal presence of factors known to have important roles in the mammalian immune system; however, the extent of each gene family may vary according to evolutionary position and species.

Myelomonocytic Cells

Descendants of the myelomonocytic lineage have many functions, some related to morphogenesis and tissue repair, others more relevant for immune defense (45, 113, 131). A key facility is their capacity for phagocytosing pathogens and infected or damaged cells. In the context of adaptive immunity, such innate effector cells provide information about tissue damage and infection to the adaptive effector arm, comprising B and T cells (61). This is achieved by expression of cytokines, interleukins, and chemokines in response to their activation by damage- or pathogen-associated molecular patterns. In this way, they provide critical temporospatial contextual information and establish an effective reciprocal intercellular communication system that is required for the initiation and termination of effective immune responses.

A second role of myelomonocytic cells is the presentation of antigens to lymphocytes (45). Dendritic cells—the paradigmatic antigen-presenting cells in mammals—are also present in lower vertebrates; cells with the morphological, phagocytic, and gene expression characteristics of dendritic cells were identified in two teleost species (3, 84), and in zebrafish were additionally shown to activate T lymphocytes in an antigen-dependent manner (84). Hence, it appears that macrophages and dendritic cells represent an evolutionarily conserved feature of jawed vertebrates, a notion supported by morphological evidence for dendritic cells in lymphoid tissues of cartilaginous fishes (120). Given the similarities between the immune systems of jawless and jawed vertebrates, it is anticipated that lamprey and hagfish also possess antigen-presenting cells. Indeed, in mixed leukocyte reactions, the adherent myeloid fraction of hagfish blood cells drives the alloresponse, whereas lymphocytes are the responder cells (114).

Antigen-specific immune responses are initiated by providing native or processed foreign structures to lymphocytes. In jawed vertebrates, antigen presentation occurs via two functionally interconnected pathways culminating in the formation of MHC-peptide complexes on the cell surface (97). As noted above, the chromosomal region encoding the MHC is unique to the genomes of jawed vertebrates and does not exist in jawless vertebrates; it appears therefore that if a functionally equivalent antigen presentation system exists in lamprey, its molecular components are probably different from the paradigmatic MHC system. In jawed vertebrates, comparative genomic studies are beginning to delineate the ancestral structure of the extended MHC region (67, 104, 105, 106) and to define group-specific differences, such as the separation of MHCI and MHCII genes in teleosts (75).

Lymphocytes

Current evidence suggests that lymphocytes evolved as a distinct cell type of the immune system concomitantly with the emergence of vertebrates. Remarkably, all vertebrates possess two major types of lymphocytes—B cells and T cells (Figure 3)—which raises questions about their possible invertebrate origin (15). The dichotomy of B and T cell lineages was discovered in chicken (26), subsequently confirmed for all jawed vertebrates, and recently demonstrated for jawless fishes (47). B cells express antigen receptors on their cell surface as B cell receptors (BCRs) and upon antigen encounter can secrete them as antibodies [immunoglobulins (Igs)]. At least two different B cell lineages have been described in jawed vertebrates. In mammals, complementary functions are attributed to B1 and B2 cells (87). Recent evidence supports the existence of at least two functionally and genetically distinct B cell lineages in fish; in teleosts, B cells expressing two different Ig isotypes, termed IgM and IgZ (or IgT), have been described (31, 41, 49). IgZ/T-expressing lymphocytes are thought to constitute the B cell type associated with the intestinal immune system (153). It appears that the transcription factor IKAROS is required for the expression of IgZ/T and perhaps also for the development of the entire lineage (124). Cartilaginous fishes also possess several Ig isotypes, including IgM (41); whether the new antigen receptor (NAR) isotype of Igs in cartilaginous fishes (118) is expressed in a separate lineage is not known. Lamprey B-like lymphocytes express the variable lymphocyte receptor B (VLRB) antigen receptor on the cell surface and also secrete it upon antigen encounter (4, 5). Functional similarities between Ig⁺ and VLRB⁺ B lineage cells are indicated by the lineage-specific activity of homologous genes encoding cell-surface and signaling molecules (47). In jawless fish, there is no evidence to date for a second B-like lymphocyte lineage.

In contrast to B cells, the antigen receptors of vertebrate T cells are always cell-surface bound, even after antigenic stimulation. Two main functionally distinct lineages of T cells exist in jawed vertebrates, one expressing an αβ T cell receptor (TCR) and the other expressing a γδ TCR (72). Cartilaginous fishes are an exception to this rule, as they have been found to additionally express several unique types of antigen receptors [NAR-TCR, a variant δ chain (28), and chimeric Ig/TCR forms (27)] whose exact roles in adaptive immunity remain to be established. The presence of more than one genetically separable lineage of T cells is not unique to jawed vertebrates, but also holds true for lamprey (47) and possibly hagfish (71); the structurally related VLRA (47) and VLRC (69) antigen receptors appear to be expressed in a mutually exclusive fashion (69, 71) by distinct T-like cells, although their functional differences, if any, remain to be defined.

Hematopoietic and Lymphoid Tissues

All immune effector cells (including B lymphocytes) develop and differentiate in general hematopoietic tissues; T cells are an exception, as they develop in a dedicated and anatomically distinct tissue environment, the thymus (recently reviewed in 17). Hematopoietic tissues in fishes occur in many forms (50). Sometimes they are associated with the gut tube, as exemplified by the typhlosole of lamprey larvae; in cartilaginous fishes they can be found in the so-called Leydig's organ, which is appended to the esophagus, and the spiral valve, a specialized region of the lower intestine. Another site of significant hematopoietic activity in fishes is the kidney. B cell development in hematopoietic organs is believed to occur in specialized domains or niches that are spatially separated from environments supporting development of other hematopoietic cell types, such as erythroid and myeloid cells. As indicated above, T cell development is always confined to distinct lymphoepithelial structures in the pharynx, although in that location they can assume many forms. In lamprey, thymopoietic tissue is distributed throughout the entire gill basket in small structural units termed thymoids (10) located at the tips of gill filaments; in jawed vertebrates, thymopoietic tissue is present in larger aggregations associated with fewer pharyngeal arches, culminating in the appearance of two laterally symmetric lobes in teleosts. The morphological differentiation of the thymus into cortical and medullary zones is remarkably conserved in all jawed vertebrates; in mammals, it underlies the sequential differentiation steps of thymocytes from uncommitted precursors to mature T cells expressing a diverse repertoire of MHC-restricted and self-tolerant TCRs. Whether the thymus equivalent in lamprey also exhibits functionally distinct areas is not yet known.

With regard to secondary lymphoid tissues (Figure 2), fish are notable for the presence of spleen- and gut-associated lymphoid tissue and the lack of lymph nodes, which made their evolutionary debut in birds and are also found in mammals. Secondary lymphoid tissues are thought to provide an environment conducive to the efficient initiation and regulation of immune responses. Indeed, the histological structure of these tissues in cartilaginous and bony fishes suggests the close apposition of T, B, and dendritic cells (120). The presence of similar structures in lamprey has not yet been ascertained.

THE COMMON GENETIC TOOL KIT OF VERTEBRATE IMMUNE SYSTEMS

Arguably, the most dramatic changes in vertebrate immune systems concern the genes and genetic networks that are involved in the somatic diversification of antigen receptors.

Clonal Expression of Somatically Diversified Antigen Receptors

Jawless and jawed vertebrates share fundamental strategies for the elaboration and expression of the antigen receptor repertoires in lymphocytes (15, 18, 25, 40, 41, 52, 83, 109). Functional receptor genes are assembled in combinatorial fashion from incomplete, germline-encoded segments in lymphocytes, one allele at a time; successful assembly of a functional antigen receptor gene on one allele prevents assembly of the second. Only if the first instance is unsuccessful is a further attempt made to assemble the second allele. As a result of this allelic exclusion mechanism, each lymphocyte expresses only a single antigen receptor. Although the precise molecular mechanisms of this somatic diversification process differ between jawless and jawed vertebrates (see below), it has the potential to generate a vast repertoire of different receptors from limited genetic material because, to a first approximation, the assembly process yields structurally different receptors in each lymphocyte. Clonal expression of the anticipatory primary repertoire of antigen receptors is essential for antigen-specific memory formation in the immune system, because exposure to one particular antigen activates only a subset of lymphocytes, namely those that bear the cognate receptor.

Antigen Receptor Assembly and Structure

Commonalities in design aside, jawless and jawed vertebrates use different assembly mechanisms and protein structures to generate functional antigen receptors. Structurally, the antigen receptors of jawed vertebrates belong to the Ig superfamily. BCRs and TCRs are obligate hetero-oligomers composed of two membrane-spanning heavy chains and two associated light chains in the case of Igs, and two membrane-spanning molecules (either α and β or γ and δ) in the case of TCRs; in lymphocytes, antigen receptors are associated with cell-type-specific coreceptor components, transmitting antigen binding to downstream signaling pathways. Functional antigen receptor genes are formed by means of V(D)J recombination by recombination activating gene (RAG) family proteins; variable (V)-, diversity (D)-, and joining (J)-type segments are assembled to form functional antigen receptor genes. In addition to the combinatorial use of these sequence elements, non-germline-encoded sequence variability at the junctions (typically encoding the antigen-binding surface of the antigen receptors) makes a significant contribution to the diversity of functional BCRs and TCRs. With the advent of high-throughput sequencing methods, the size of the antigen repertoires can now be studied not only in mice and humans but also in lower vertebrates. For instance, a recent study estimated that the zebrafish antibody repertoire comprises approximately 5 × 10³ different Ig heavy-chain sequences per fish (149). Assuming that the light-chain repertoire is of the same magnitude, approximately 2.5 × 10⁷ different combinations of heavy and light chains (that is, potential specificities) could be formed; however, the zebrafish does not possess a large enough number of lymphocytes to express this large repertoire.

The antigen receptors of jawless fishes are composed of leucine-rich-repeat-containing proteins (108). Like IgM (1, 92), the secreted form of VLRB is thought to form multimers (5), attesting to the remarkable functional similarity between the antibody-like molecules of jawless and jawed vertebrates. The assembly of functional VLR genes from incomplete genomic elements in lamprey lymphocytes is thought to be achieved by cytidine deaminases (CDAs) (5, 95, 117) that are expressed in a lineage-specific fashion—CDA1 for T lineage cells and CDA2 for their B lineage counterparts (47). The estimated potential diversity of VLRs is on the order of 10¹⁴ (5)—rivaling that of Igs and TCRs and far exceeding the total number of lymphocytes in jawless vertebrates.

Evolutionary Origin of Antigen Receptor Genes

It has been noted that the genes eventually subjected to somatic diversification in vertebrate lymphocytes probably already existed in the vertebrate common ancestor (15, 18, 25, 40, 41, 52, 83, 109) (Figure 4). Phylogenetic analyses indicate that VLR-like genes are derived from the gene encoding the vertebrate-specific glycoprotein Ib α (GPIbα) membrane protein. Moreover, chordate genomes contain genes encoding ancestral Ig-like and TCR-like proteins; the same appears to be true for the activation-induced cytidine deaminase (Aid)–Apobec genes implicated in the diversification of VLR genes in jawless fish. Indeed, CDAs might have been co-opted from an ancient defense mechanism that targeted foreign genetic material (48); they continue to also be present in the genomes of jawed vertebrates, in which they carry out auxiliary immune functions such as somatic hypermutation of Ig-encoding genes (35). Intriguingly, in some species, such as birds, they are even required for the generation of the primary repertoire of Igs (116, 136), representing one of many striking functional parallels between jawless and jawed vertebrates. However, whereas the diversification of VLR genes appears to rely solely on the products of the genes present in the genome of the vertebrate common ancestor, somatic diversification of Ig and Tcr genes depends on the activity of Rag recombinases; these are thought to have been incorporated into the genomes of gnathostomes via lateral gene transfer by transposases after the divergence from cyclostomes (123). This apparent evolutionary discontinuity supports the notion of the independent evolutionary origins of diversification mechanisms in the two sister groups of vertebrates.

Studies addressing the evolutionary histories of germline-encoded receptor families with important functions in the mammalian immune system have revealed many group- and species-specific differences; however, it appears that representatives of the complement system (100), the natural killer receptors (152), and the major families of extra- and intracellular pathogen receptors—such as lectins (144), Toll-like receptors (TLRs) (68, 107), peptidoglycan recognition proteins (22), NOD-like receptors (NLRs) (77, 126, 134), and retinoic acid-inducible gene I (RIG-I)–like receptors (121, 135, 154)—are all present in lower vertebrates. Moreover, at least some components of the intracellular signaling cascades that are functionally required downstream of these receptors appear to be shared among all vertebrates (19, 47, 55, 89, 90, 112, 142, 146).

Lineage-Specific Transcription Factors and Signaling Components

The functional similarities between the adaptive immune systems of the two sister groups of vertebrates extend even beyond antigen receptor formation and expression. Not only do they appear to employ a shared set of proinflammatory chemokines, cytokines, and their associated receptors to initiate and regulate immune responses (see above), but according to the results of recent genome comparisons and expression studies, they appear to make use of the same evolutionarily conserved key transcription factors regulating the development and differentiation of hematopoietic lineages, including lymphocytes (6, 12, 36, 64). The same applies to transcription factors critically required for the differentiation of distinct stromal cell types in primary lymphoid organs (9).

Collectively, a picture of fish immunity emerges that suggests adherence to a common design principle for all vertebrates, namely the cooperation of innate and adaptive immunity based on germline-encoded and somatically diversified antigen receptors; jawless and jawed vertebrates realized adaptive immunity using distinct molecular means and thus responded to the same demand in quite different ways (14).

PROBING THE FUNCTION OF FISH IMMUNE SYSTEMS

While the outlines of innate and adaptive immune systems in vertebrates are beginning to emerge, studies on the immune systems of jawless and jawed fish promise to yield additional details on functional commonalities as well as some species-specific adaptations that should prove particularly informative with respect to the malfunctioning of the human immune system. We briefly discuss some of the pertinent experimental strategies below.

Transient Interference with Gene Function

For the analysis of immunologically relevant developmental processes and effector functions that manifest themselves early in development, two versatile methods of genetic intervention are widely used. Gene-specific knockdown (13) is made possible by injection of antisense oligonucleotides, singly or in combination, at the one-cell stage. When target sequences encompass initiation codons, translation of both maternal and zygotic mRNAs is inhibited; alternatively, antisense oligonucleotides targeting splice donor or acceptor sites affect only zygotic mRNAs and result in the translation of either nonfunctional proteins or specific (alternatively spliced) isoforms. Hence, through appropriate target sequence design, the maternal and zygotic effects of a specific gene can be distinguished and specific isoforms selectively produced. The ease with which gene functions can be knocked down by antisense oligonucleotides offers the opportunity to address more than one gene at a time, for instance, to overcome redundant functions of paralogous genes or to interrogate cooperativity in genetic networks. For instance, thymus homing in zebrafish (53) and medaka (9) was found to require the cooperative activity of two chemokines, ccl25a and cxcl12a, to achieve robust colonization of the thymic rudiment during embryogenesis (Figure 5).

Gain-of-function analyses can be performed by injection of in vitro transcribed mRNAs or proteins of one or more genes in wild-type or mutant forms (54). This strategy can be applied to address many questions. For instance, during positional cloning projects following the identification of specific mutants, the lack of additional alleles frequently requires further measures to be able to firmly associate a sequence variant with a specific phenotype. In this case, rescue of the mutant phenotype can be employed to prove the functional identity of the implicated gene. Moreover, injection of mRNAs encoding mutant forms of the relevant gene enables the assessment of their functional capacity or lack thereof, including the possibility of identifying dominant-negative variants. An alternative method to achieve gain-of-function phenotypes is to inject either cDNA expression constructs (54) or large fragments of genomic DNA (such as those contained in bacterial artificial chromosomes) (151), which can be superior to RNA injection because of their often more sustained expression or even tissue-specific expression. However, injections of antisense oligonucleotides, mRNAs, and cloned genomic DNA all produce only transient effects; therefore, these procedures are applicable only to the investigation of innate and adaptive immunity features that are relevant in early stages of development.

Isolation of Gene-Specific Mutations

Examining the effects of genetic perturbations on immunological functions in later stages of life requires stable genetic modifications. Locus-specific genetic modification by homologous recombination has not yet been achieved in lower vertebrates. However, two alternative strategies approximate the versatility of the targeted gene disruption that is a routine procedure in mice. Gene-specific identification of mutations from a pool of mutagenized genomes—targeting induced local lesions in genome (TILLING)—has been successfully used to establish fish models with mutant versions of immunologically relevant genes, such as rag1 (150). Clearly, the practical value of this approach depends on both the number of mutagenized genomes available for screening and the mutation density; whereas the former parameter scales with the effort required for exhaustive screening, the number of undesired, possibly even epistatic, background mutations determines the extent of the subsequent genetic purging required.

Whereas the success of TILLING depends to a large extent on chance, the ability to introduce germline mutations via sequence-specific artificial nucleases (24) provides a more targeted procedure for the generation of gene-specific mutations. So far, however, it has not been used to generate mutations of immunological relevance.

Live Imaging

Arguably the most important advantage of small-vertebrate models for immunological research is the opportunities for live imaging. In mammals, this is technically demanding and often hampered by the complex physiology of the process. These difficulties notwithstanding, two-photon microscopy has been successfully used to observe cell migration and interaction in lymph nodes of mice and to derive novel information on migration routes and cellular interaction modes (8); however, other processes are difficult to study using this method as observations are limited by anatomical difficulties or the short observation periods. A case in point here is the process of thymus homing, for which direct long-term observation in mammals has not yet been possible. Recently, however, we have succeeded in doing this in transgenic zebrafish, the relevant cell types of which were marked by expression with different fluorescent proteins; we were able to demonstrate that the proliferation of epithelial cells is a cell-autonomous process and not dependent on the presence of hematopoietic cells, and that thymus colonization is driven by an evolutionarily conserved cooperative action of ccl25a and cxcl12a chemokines (53) (Figure 5). Another spectacular application of live imaging in lower vertebrates has been the direct observation of how immune effector cells interact with infectious agents (29).

Functional Analysis of the Immune Response

Most of the information regarding the structure, components, and functions of the immune system has been obtained from the traditional models of immunological research: humans, mice, and birds. In contrast to these well-established and extensively studied model systems, immunological research in lower vertebrates is less advanced, often simply because of the paucity of analytical tools, reagents, and robust in vitro assay procedures. Furthermore, many of the traditional models in fish immunology, such as trout and carp, although amenable to transgenic manipulation, are impractical genetic model systems owing to their long generation times. This problem also affects genetic studies on lamprey and hagfish. Nonetheless, these established models have been used extensively for the analysis of immune system function, particularly in the context of infection biology (37, 86, 143). To overcome this limitation, genetically tractable fish species such as zebrafish are being used as models for immunological research (81, 91). In addition to the well-established procedures for transgenesis in all varieties, recent reports have described the successful development of in vivo and in vitro assays suitable for the analysis of immune function in this species (130). For instance, transplantation of hematopoietic cells has been used to study the developmental potential of specific hematopoietic cell types and to probe certain aspects of histocompatibility (34, 138); in vitro T cell activation assays using purified antigen-presenting cells have also been successfully performed (84).

IMMUNODEFICIENCY SYNDROMES

Compared with the situation in mammals, our understanding of the genetic underpinnings of immune function in fish is still limited. However, a growing number of fish models are being generated that carry mutations in genes associated in humans (and mice) with immunological deficiency or dysregulation. These models can be used to examine which aspects of the often complex phenotypes are evolutionarily conserved and which features are species specific.

Forward Genetic Screens

The low cost of keeping small fish species of high fecundity has prompted several researchers to conduct large-scale forward genetic screens in the hope of identifying components of previously unknown pathways crucial for immune system development and function. For instance, to examine the genetic basis of blood and immune cell development, several groups have performed large-scale and tissue-restricted screens in zebrafish (115, 148) and medaka (42). Forward genetic screens concerned with the adaptive immune system of zebrafish (16, 140) and medaka (60) have focused mainly on T cell development. This was done primarily because T cell development in the thymus begins within a few days of fertilization (78) and thus much earlier than B cell development (32). Hence, whole-mount RNA in situ hybridization for rag1 expression can be used to detect aberrant T cell development in the thymus of fish embryos (16, 60, 78, 140).

Two types of such screens have been carried out, one requiring the treatment of fish with a mutagenic chemical (ENU) (16, 60, 140) and the other using insertional mutagenesis by retroviruses (43). Mutations with immunological phenotypes have so far been identified only from chemically mutagenized fish. These were conducted as either haploid or diploid screens. Haploid screens—also known as gynogenetic or early-pressure screens—are performed by stimulating the duplication of maternal genetic material subsequent to fertilization with inactivated sperm (79). Because all sequence variants in the maternal genome are thus rendered homozygous, the desired phenotypes are often confounded by the inevitable presence of many background mutations, so that phenotypes often change or disappear altogether when founder females are outcrossed to isolate the desired mutation. Nonetheless, this procedure has the advantage of being fast and has been successfully used in several laboratories (Table 1). The classical diploid screen is more time-consuming but also enables the identification of dominant and recessive mutations, depending on the generation in which the phenotypes are assessed. The initial concern that the polyploid nature of teleost genomes might also confer a high degree of robustness on the genetic pathways regulating immune development and function, therefore constituting a serious obstacle to the identification of mutants, was refuted by the discovery of several intriguing examples of immune defects (Table 1).

Table 1

Viable recessive mutations^a affecting definitive hematopoiesis in zebrafish (Danio rerio)

Encouragingly, these screens have led to the discovery of mutations in genes known in mammals to regulate essential aspects of hematopoiesis and lymphopoiesis, and have provided new candidate genes for human immunodeficiency syndromes (57, 58, 59, 93, 124, 128, 129, 139, 150) (Table 1). Overall, these results illustrate the remarkable functional similarity among vertebrate immune systems.

Another aspect of the study of immune function that promises to be transformed by the use of small-vertebrate models is the field of infection biology. The unique biology of the zebrafish has allowed for the development of high-throughput assays for susceptibility to infection by specific pathogens. For instance, a forward genetic screen for innate susceptibility to Mycobacterium marinum identified a hypersusceptible mutant; detailed analysis of the genetic basis uncovered a crucial role of leukotriene A(4) hydrolase in the regulation of inflammatory states (137) (see also Table 2).

Table 2

Examples of genes involved in human primary immunodeficiency with evolutionarily conserved functions in zebrafish (Danio rerio)^a

Epistatic Analysis

The availability of various lymphopoietic mutants has even made it possible to predict the existence of so far unidentified components of immunologically relevant pathways. For instance, the identification of mutants for the interleukin 7 receptor alpha chain gene (il7r) and the gene encoding the intracellular signal transducer jak3 in zebrafish and their distinct phenotypes in early T cell development strongly suggest that teleosts possess a functional homolog of IL-7 that has so far eluded identification, possibly owing to the low sequence conservation of interleukins; genetic interaction analysis has additionally demonstrated that the components of the presumed IL-7 signaling pathway are functionally equivalent to those of their mammalian counterparts (58).

Genetic screens potentially allow not only the identification of novel pathways relevant to immune function but also the isolation of genetic modifiers. To assess the role of candidate modifier genes, gene-specific knockdown or gain-of-function approaches (see above) can be employed. For instance, many immunologically relevant mutations, such as those affecting early stages of lymphocyte development, cause activation of p53-dependent checkpoints; hence, inhibition of p53 function by injection of specific antisense oligonucleotides suppresses the ensuing apoptosis, thus alleviating the phenotype and enabling further examination (for a recent example, see 93). However, it is also possible to identify modifiers and enhancers in an unbiased manner by conducting secondary genetic screens. Although this strategy has not yet been applied to an immunological phenotype, a recent study on the genetic basis of erythropoiesis illustrates this point. A suppressor screen for secondary mutations alleviating the severe defect in erythropoiesis observed in fish lacking wild-type trim33 identified the cdc73 gene as a functional modifier; remarkably, fish simultaneously mutant for trim33 and cdc73 exhibited normal red blood cell formation (7). Hence, the use of enhancer and suppressor screens routinely carried out in invertebrates, such as worms and flies, is within reach for small vertebrates, even though the full potential of first-generation screens has not yet been fully exploited. As an alternative to secondary genetic screens, it is possible to perform chemical screens to examine the ability of small-molecule effectors to modify the phenotype of specific genetic alterations (56).

THE PROMISE OF COMPARATIVE IMMUNOGENOMICS

For comparative immunology to achieve its goals, analyses will have to be directed to comparisons both within and between the two groups of vertebrates. Although the number of species of extant jawless fish is relatively small, they nonetheless occupy very different ecological niches; the far more numerous cartilaginous and bony fishes encompass species with extremely divergent life histories, for instance, with respect to the mode of fertilization (internal or external) and life span (from only several weeks to more than 100 years). Hence, although most immunological studies still focus on mice and humans, fish immunologists have at their disposal an extensive list of diverse life forms to establish the core structure of adaptive immune systems and assess the changes related to their different habitats. The imminent completion of hundreds of vertebrate genome projects under the Genome 10K proposal (46) promises to revolutionize our concept of the structure and evolutionary trajectory of immunologically relevant gene families in lower vertebrates. Below, we briefly discuss several strategies that are available to arrive at a comprehensive appraisal of immune system structure and function from genomics alone.

All metazoans possess an innate immune system that is based on a large variety of germline-encoded cell-surface and intracellular pathogen receptors; these receptors are activated by molecules specific to viruses (e.g., double-stranded RNA), bacteria (e.g., lipopolysaccharides), and fungi (e.g., proteoglycans), and receptor-ligand interactions then initiate multilayered immune defenses. Because these pathogen recognition receptors possess evolutionarily conserved modular domain structures, the diversity of these gene families can be easily deduced from genome and transcriptome data. For instance, fish possess a unique set of TLRs, distinct in number and structural diversity from those found in mammals (68, 107). Hence, molecules acting as sensors of pathogens and danger and their downstream signaling pathways provide key information about the relative importance of hard-wired and rapidly deployed innate responses compared with slower adaptive responses. Indeed, several human disease phenotypes related to innate immunity have been reproduced, at least to a limited extent, in zebrafish (Table 2).

Interpretation of these data sets will initially be guided by the conceptual framework developed for experimentally well-studied species, such as mammals, but in the long term will be complemented by consideration of species- and/or group-specific peculiarities. For instance, the recently reported analysis of the cod genome revealed a loss of the MHCII pathway (132), which in mammals is essential for the function of several helper and regulatory T cell subsets. Nonetheless, the cod enjoys a long life span (approximately 20 years) and hence must have adapted in unique ways to ensure the functionality of adaptive immunity. The discovery that it lacks the MHCII pathway was not unexpected, because previous work had shown that this species lacks robust antibody production upon infection and immunization (127). Genome analysis not only confirmed the absence of MHCII genes and genes encoding other components of this pathway, such as CD4, but also suggested possible compensatory adaptations in the form of an increased number of MHCI genes, an expanded repertoire of TLR genes, and possibly others (132). Because other species are not adapted to such a constriction of adaptive immune function, lack of MHCII expression causes a severe immunodeficiency syndrome; in humans, this is known as bare lymphocyte syndrome and is caused by mutations in CIITA, a central transcriptional regulator of MHCII expression (133). This example illustrates a second potential strategy for analyzing vertebrate genomes for immunologically relevant features: In addition to major gene families, the search for genes homologous to nonredundant effector molecules of the human immune system should prove exceptionally informative. At present, mutations in over 200 genes are known to cause distinct immune disorders, the major classes being categorized as severe combined immunodeficiency (SCID), combined immunodeficiency (CID), and common variable immunodeficiency (CVID) (38, 101). Mutations in and/or knockdown of some of these genes in fish have been generated and found to cause strikingly similar phenotypes (Table 2). Given the amenability of the zebrafish to genetic and chemical perturbations, these models should prove uniquely useful for human immunodeficiency research, particularly with the aim of developing novel treatment strategies transcending current options (39, 101).

The identification of specific nonredundant regulators of immune functions in mammals provides a useful starting point for comparative genomic analysis of vertebrate genomes. These index genes can be used to derive evolutionary trajectories of presumed core functions—for instance, the development of lymphoid organs (such as the thymus and lymph nodes) and/or specific immunological functions (such as affinity maturation and class switch recombination of antibodies). This strategy is exemplified by an analysis of the evolutionary history of the transcription factor Foxn1, which is essential for the differentiation of thymic epithelial cells and hence thymopoiesis in mammals (98). Based on phylogenetic analysis of Foxn1 orthologs and paralogs, the pharyngeal endoderm was defined as a common expression domain of this gene family in all chordates (9). This information ultimately guided the identification of presumptive thymopoietic tissue in lamprey larvae, which is a morphologically inconspicuous tissue at the tips of their gill filaments that had eluded discovery for more than 100 years (10). Likewise, because the emergence and expression patterns of specific genes of the tumor necrosis factor (TNF) family of ligands and receptors are directly associated with the presence of lymph nodes in mammals (119), it should be possible to delineate the evolutionary origins of this major innovation of vertebrate immune systems and to examine whether relevant genetic pathways already exist in cryptic form in fishes but have not yet been drawn into the realization of specific developmental or morphogenetic programs. An analysis of this kind was conducted for the zebrafish homolog of AID, which in mammals is required for somatic hypermutation and Ig class switching (111). Ig class switching has not yet been demonstrated in fish; nonetheless, fish Aid perfectly complemented this particular aspect of Aid deficiency in mice (11), suggesting that the functional capacity of Aid for class switching preempted its deployment in evolutionarily more recent species.

FROM COMPARATIVE TO SYNTHETIC IMMUNOLOGY

A major impediment to experimental comparative immunology is the fact that the long generation times and complex life histories of many evolutionarily, ecologically, or economically important species make them genetically intractable. In some instances, this roadblock could potentially be circumvented by use of xenotransplantation of specific cell types or tissues. For instance, zebrafish carrying a loss-of-function mutation in the gene encoding the c-myb transcription factor were found to lack definitive hematopoiesis (129); the severely impaired innate immune system and lack of an adaptive immune system in this mutant could provide favorable conditions for transplantation of allogeneic and xenogeneic grafts. Moreover, because in this case the host species is genetically tractable, recipients could be transgenically modified to accommodate any special requirement the donor cells might have.

The same strategy could in principle also be applied to the induction of novel immune functions. For instance, from work in mice, the genetic requirements for the formation of lymph nodes are well known (119). Given that certain members of the TNF receptor family—most of which do not exist in the genomes of lower vertebrates lacking lymph nodes—play essential roles in this process, it would be interesting to examine whether ectopic expression of such genes in fish would initiate lymph node development. The same logic could be applied to the replacement of antigen receptor genes. For instance, would replacement of the cognate Ig loci in mouse B cells by the lamprey VLRB locus generate a functional alternative antibody system?

Comparative information addressing core aspects of the immune system might also guide attempts to generate synthetic immunological functionalities in vivo. This strategy has recently been applied to the partial reconstruction of thymopoiesis in a dysfunctional thymic rudiment of mice. Remarkably, expression of only two evolutionarily conserved factors—Cxcl12, a dual-function protein (chemoattractant and cytokine), and Dll4, a nonredundant Notch ligand for early T cell progenitors—sufficed to enable the generation of CD4⁺CD8⁺ TCR-expressing immature thymocytes in the refunctionalized epithelium in vivo (21). These experiments suggest a strategy for investigating how expression of crucial thymopoietic factors in ectopic sites or noncognate cell types could lead to organoids that might augment or replace failing immune functions.

CONCLUSIONS

The discovery of an alternative adaptive immune system in jawless vertebrates provides an unprecedented opportunity to carry out comparative immunological studies at all levels, from genome structure, gene content, and transcriptional complexity to functional analyses of the different effector cell types and lymphoid tissues. Collectively, these studies promise to reveal the principles governing the function of innate and adaptive immunity and their functional cooperation during immune homeostasis and immune responses. Thus, comparative immunology will not only inspire studies aimed at reconstructing primordial functionalities in genetically tractable species but also impact the fledgling field of synthetic immunology, which aims to rebuild immune functions from minimal sets of core components. Ultimately, the information derived from these studies will contribute to the development of new treatment modalities for failing human immune systems.

SUMMARY POINTS

disclosure statement

The authors are not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review.