Infection, Genetics and Evolution 21 (2014) 492–496

Contents lists available at ScienceDirect

Infection, Genetics and Evolution

journal homepage: www.elsevier .com/locate /meegid

Discussion

Evolutionary responses of innate immunity to adaptive immunity

1567-1348/$ - see front matter � 2014 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.meegid.2013.12.021

⇑ Corresponding author.E-mail addresses: wardae@terpmail.umd.edu (A.E. Ward), benjamin.rosentha-

l@ars.usda.gov (B.M. Rosenthal).1 Tel.: +1 (301) 504 5408.

Amanda E. Ward a,⇑, Benjamin M. Rosenthal b,1

a University of Maryland-College Park, College Park, MD 20740, United Statesb United States Department of Agriculture, Beltsville Agricultural Research Center, Building 1040, Room 103, 10300 Baltimore Avenue, Beltsville, MD 20705-2350, United States

a r t i c l e i n f o

Article history:Received 25 August 2013Received in revised form 18 December 2013Accepted 20 December 2013Available online 8 January 2014

Keywords:Toll-like receptorsPattern recognition receptorsJAK/STAT signalingAgnathostome immunityInnate immunityVLR immunity

a b s t r a c t

Innate immunity is present in all metazoans, whereas the evolutionarily more novel adaptive immunityis limited to jawed fishes and their descendants (gnathostomes). We observe that the organisms that pos-sess adaptive immunity lack diversity in their innate pattern recognition receptors (PRRs), raising thequestion: did gnathostomes lose the diversity of their ancestors? Or might innate receptors have diver-sified in the lineage lacking adaptive immunity? We address this question by contextualizing PRRs intheir distinct functional roles in organisms possessing or lacking adaptive immunity. In particular, limitedPRR diversity in gnathostomes is accompanied by an expansion of the JAK/STAT signaling pathway, whichwould suggest that the development of adaptive immunity shifted the role of PRRs from the entirety ofpathogen recognition to regulators of subsequent immune responses. As PRRs became essential upstreamcomponents of the increasingly complex JAK/STAT signaling cascade in organisms possessing adaptiveimmunity, it may have limited their freedom to diversify. By contrast, PRR diversity continues to conferan advantage for organisms lacking the means to generate non-self recognition receptors via somaticmutation. Extensive deuterostome PRR diversity may have been driven by gnathostome adaptive immu-nity inducing diversification of shared pathogens, which exerted strong diversifying selection on deuter-ostome PRRs. Thus, the development of adaptive immunity changed the role of PRRs in immunity as wellas the selective forces on host receptors, deuterostomes, and pathogens.

� 2014 Elsevier B.V. All rights reserved.

1. Introduction

Immunity has evolved in all metazoans to be protective againstpathogens. The two major strategies employed by metazoans forthe first step of this process, recognition, are the use of a largeset of diverse, germ-line encoded pattern recognition receptors(PRRs) and a system for random generation and clonal selectionof antigen specific receptors. While oversimplified, the first strat-egy describes the typical innate immune system and the secondis characteristic of an adaptive immune system. Innate immunityis the evolutionarily more ancient system that is present, in someform, in almost all metazoans. In contrast, adaptive immunity hasevolved in complement to innate immunity in jawed vertebrates(McCurley et al., 2011).

Adaptive immunity, as elaborated in mammals and other jawedvertebrates (gnathostomes), is often viewed as the definitive solu-tion to pathogen recognition. Recent evidence has complicated thisclassical picture with the discovery of alternatives to the prototyp-ical model of adaptive immunity (Hedrick, 2004). The purpose of

this paper is to review several new lines of evidence concerningthe advent of distinct (but functionally analogous) systems ofadaptive immunity in vertebrates and the subsequent impact thatsuch innovation had on the diversity and function of innate path-ogen recognition receptors in organisms possessing, and in organ-isms lacking, the ability to synthesize nearly limitless repertoiresof molecules capable of recognizing pathogens.

1.1. The evolution of adaptive immunity

The cellular component of the adaptive immune system injawed vertebrates consists of T lymphocytes and B lymphocytes,which are activated by antigen presenting cells akin to those alsofound in animals lacking such specific effector cells. As they samplethe environment and phagocytose pathogens and damaged tissue,dendritic cells and other antigen presenting cells may be viewed as‘‘surveillance’’ and a link between the innate and adaptive immunesystems (Palucka and Banchereau, 1999). Until recently, no homo-logues for T and B lymphocytes were known in jawless fishes(agnathostomes). This apparent absence of adaptive immune cellssparked an immunological conundrum because sharks, hagfish,and lampreys exhibit graft rejection, a hallmark of a cytotoxic T-cell mediated immune response (Wang et al., 2008). More recently,it has been shown that hagfish and lampreys do have populations

A.E. Ward, B.M. Rosenthal / Infection, Genetics and Evolution 21 (2014) 492–496 493

of cells that have similar function to T and B cells but that lack thecharacteristic markers of T cell receptors and B cell receptors (TCRs,BCRs). In gnathostome immunity, immature T cells are shown self-antigens in the specialized lymphoid organ, the thymus. It is in thethymus that self-reactive T cells are negatively selected. A homol-ogous organ could not be found in agnathostomes until it was dis-covered that the tips of hagfish and lamprey gills served a similarfunction to the gnathostome thymus (Boehm et al., 2012).

In classical adaptive immunity, immature lymphocytes mustundergo V(D)J recombination to specify the binding characteristicsof each resulting TCR and BCR. This process of somatic recombina-tion is accomplished by the rearrangement of germline encodedgene fragments to produce structurally and functionally newreceptors. This process is engendered by two principal enzymes:recombination activating genes 1 and 2 (RAG1 and RAG2); whichare members of a family of enzymes known as ‘‘recombinases’’(Nishana and Raghavan, 2012). Although agnathostomes lack VDJgenes and RAG1/2, they nonetheless produce cell lines functionallyanalogous to gnathostome T and B cells (Boehm et al., 2012). Re-cent work by Pancer et al. (2004) showed that agnathostomes dohave functionally analogous receptors, called Variable LymphocyteReceptors (VLRs); these are produced by gene conversion of germline encoded fragments, in a similar but not identical process toTCR/BCR development. However, VLR structure is drastically differ-ent from that of TCR/BCRs; VLRs feature leucine rich repeats whichproduce a concave binding region, similar to the structure of thehighly evolutionarily conserved innate pattern recognition recep-tors, the Toll-like receptors (TLRs) (Pancer et al., 2004). Diversityin VLRs is produced through a process of gene rearrangement en-tirely distinct from the process by which RAG1 and RAG2 initiateV(D)J recombination. Instead, VLRs are produced by gene conver-sion, mediated by a different class of enzymes, the cytidine deam-inases. Despite differences in molecular mechanisms, VLR-basedimmunity follows a similar functional pattern as classical adaptiveimmunity, where VLRA is analogous to the T cell receptor and VLRBis analogous to the secreted B cell receptor (Boehm et al., 2012).The adaptive immune systems of gnathostomes and agnathosto-mes achieve strikingly similar outcomes by entirely distinct molec-ular mechanisms. This creates an interesting puzzle about how thetwo systems could have evolved.

The results of comparative genome analyses suggest that theancestor of gnathostomes and agnathostomes possessed the pre-cursors of both VLRs and TCR/BCRs, each of which was likely‘‘repurposed’’ from other functions. The gnathostome lineagedeveloped RAGs, which act on specific DNA sequences to producediverse TCRs and BCRs through the V(D)J system of somatic recom-bination. Agnathostomes instead relied on cytidine deaminases toproduce unique antigen recognition receptors. These enzymes eachform double stranded DNA breaks, but at different sequence spec-ificities. Thus, antigen recognition receptors with unique structuresbut similar function are mediated by these two systems of somaticdiversification (Boehm et al., 2012), a fascinating example bywhich independent evolutionary lineages have elaborated func-tionally analogous (convergent, but not evolutionarily homolo-gous) attributes.

However, these systems of adaptive immunity did not developin isolation from pathogens or existing innate immunity. Eventhough somatic recombination provides the adaptive immune sys-tem with the ability to generate receptors for the recognition ofover 1014 different antigens (McCurley et al., 2011), adaptiveimmunity still requires activation signals from the innate immunesystem. Considering this functional link, it is reasonable to hypoth-esize that innate immunity shaped the evolution of adaptiveimmunity in all its forms. Less evident though, is whether thedevelopment of adaptive immunity shaped the evolution of innateimmunity.

1.2. The effect of adaptive immunity on the evolution of innateimmunity

We can observe the effect of the adaptive immune system onthe innate immune system by comparing the diversity of innateimmune components in organisms that possess or lack an adaptiveimmune system. The most easily studied are the pattern recogni-tion receptors, which are present in some form in all metazoans.A comparison of PRRs and innate immune signaling adaptor mole-cules in humans (Homo sapiens), lampreys (Petromyzon marinus),sea urchins (Stronglyocentrotus purpuratus), amphioxus (Branchios-toma floridae), and fruit fly (Drosophila melanogaster), demonstratesfar more PRRs in the species that lack adaptive immune systems(Huang et al., 2008). The Toll-like receptor family, a group of PRRsthat are highly conserved across all metazoans and play integralroles in the innate recognition of bacterial and viral motifs (Jane-way and Medzhitov, 2002) provide an instructive contrast. 10Toll-like receptors have been identified in humans, 21 in lamprey,72 in amphioxus, 222 in sea urchins, and 9 in Drosophila (Fig. 1)(Huang et al., 2008; Satake and Sekiguchi, 2012). Even though lam-prey have VLR-based adaptive immunity, they still have twice thediversity in Toll-like receptors that humans have. Interestingly, thegreatest diversity is found in amphioxus and sea urchins, which donot possess an adaptive immune system. Clearly, organisms pos-sessing adaptive immunity harbor fewer innate immune receptors.

Either of two hypotheses could conceivably account for this dif-ference, however, the two hypotheses are not mutually exclusiveand could conceivably occur simultaneously. The ‘‘gnathostomeloss’’ hypothesis would posit that abundant ancestral innate im-mune receptor diversity was lost in the evolutionary lineage thatinherited adaptive immunity. Reduced need for such heritable var-iation would, according to this view, have resulted in waning selec-tive pressure to maintain and further diversify germline-encodedpathogen recognition. Adaptive immunity, powered by somaticrecombination, would have ended such organisms’ reliance on a di-verse set of innate immune receptors. In its simplest formulation,this hypothesis would predict selection against a number of path-ogen recognition receptors or the lack of maintenance of thosegenes.

The second explanation, which may be termed ‘‘deuterostomegain,’’ posits that ancestral innate immune receptors underwentmarked diversification in those organisms lacking adaptive immu-nity. Why such organisms, and only such organisms, would haveundergone diversification in their heritable pathogen recognitionreceptors remains a matter of some interest; could adaptive immu-nity in vertebrates have engendered PRR diversity in deutero-stomes, indirectly, by promoting the antigenic diversification ofshared pathogens? This possibility focuses attention on the differ-ence in magnitude of the selective pressures faced by vertebratesand deuterostomes.

An examination of the phylogeny and evolution of Toll-likereceptors, cytokine receptors, and the JAK/STAT signaling pathwaycomponents, provides a means to evaluate the relative merits of‘‘gnathosome loss’’ versus ‘‘deuterostome gain.’’

1.3. Selective pressures on Toll-like receptors

Toll-like receptors (TLRs) are a class of Type I transmembranereceptors featuring leucine-rich repeats which produce a concaveligand-binding site. They are highly evolutionarily conservedacross metazoans (Wiens et al., 2007) however, recent evidenceprovided by Maximum Likelihood analysis by Areal et al. (2011)has shown that all mammalian TLRs are still undergoing positiveselection and a similar analysis of primate TLRs by Wlasiuk andNachman (2010) showed six of ten TLRs have recently experiencedpositive selection. It should be noted that human TLRs may have

2

STAT Family Members

Toll-like Receptors

Mammalia

Reptilia

Amphibia

Osteichythes

Chondrichthyes

Petromyzondia

Myxini

Urochordata

Cephalochordata

Echinodermata

Molluska

Annelida

Nematoda

Arthropoda

Cnidaria

Porifera

10

21

43

222

9 1

2

8

7

AdaptiveImmunity

Humans

Bony Fishes

Sharks

Sea Urchin

Sea Squirt

Lamprey

Deuterostomia

Fruit Fly

Sponges

Lancelet

1

Gna

thos

tom

a

Cytokine Receptor

Chains

1

36

Fig. 1. Toll-like receptor (TLR) diversity is greatest in echinoderms but smaller in mammals. The pattern of TLR diversity is mirrored by an expansion of cytokine receptorsand the STAT family of adaptor signaling molecules (the JAK family experienced a similar expansion but for simplicity only the number of STAT family members are shown).Taken together, these patterns support the notion that with the advent of adaptive immunity, the role of pattern recognition receptors shifted to initiators of downstreamimmune signaling and appropriate responses.

494 A.E. Ward, B.M. Rosenthal / Infection, Genetics and Evolution 21 (2014) 492–496

experienced unusual selective pressures due to human migrationpatterns and more recently, health practices, but since both studiesconsider humans as part of larger groups, mammals and primates,the results do not merely reflect the selective pressures of humans.In both studies, the majority of positively selected codons were inthe leucine rich repeats which comprise the ligand binding do-main, so it is reasonable to conclude this evolution is driven byhost pathogen interactions.

Evidence for ongoing positive selection in gnathostome TLRsrenders the ‘‘gnathostome loss’’ hypothesis less plausible, at leastin its simplest formulation. If ancestral PRR diversity were selectedagainst (owing to its metabolic costs and redundancy with theantigen recognition provided by adaptive immunity), this wouldbe in contrast to extant data that substantiate the persistence ofselection maintaining PRR function. However, the ‘‘deuterostomegain’’ hypothesis does not exclude the possibility that PRRs in ver-tebrates with adaptive immune systems also underwent positiveselection, but rather that this selection was not as strong as theselective pressures experienced by deuterostome invertebrates. Ifthe ancestor had similar TLR diversity to current protostomes (dro-sophila have 9 members) then it is reasonable to assume that ver-tebrates, after acquiring adaptive immunity, underwent a modestdiversification in their TLR repertoire (mammals have between10 and 13 TLRs) (Satake and Sekiguchi, 2012). Meanwhile, deuter-ostomes underwent a much larger diversification.

It is interesting to consider how the advent of adaptive immu-nity may have created conditions necessitating such an expansionof PRRs in deuterostomes. This idea seems plausible (though by nomeans yet established) given the fact that certain deuterostomesand gnathostomes share a common pool of microbial and parasiticthreats. If the advent of vertebrate adaptive immunity engenderedantigenic diversifications among such pathogens, diversification of

PRRs in deuterostomes might have been a necessary response. Inan evolutionary sense, immunity does not require perfect defenseagainst all pathogens; to proliferate, an organism needs only com-paratively more effective defense (Hedrick, 2004). A sudden in-crease in infection would exert a strong selective pressure oninvertebrate deuterostomes, which might explain the extremediversity of TLRs in basal chordates and echinoderms (48 and222 members respectively).

Marine ecosystems, where lower chordates and vertebrates fre-quently interact, would seem to magnify the potential of competi-tion mediated by a shared pool of pathogens. For example, theparasitic worm Echinorhychnus coregoni Linkins is known to infectboth sea lamprey (McLain, 1952; Van Cleave, 1921), which lackclassical adaptive immunity and teleost ‘‘whitefish’’ (Van Cleave,1919), which have adaptive immunity. Also, more distantly, thenematode Anisakis simplex is known to affect both mollusks suchas squid (Pascual et al., 1995), which lack any form of adaptiveimmunity, and marine mammals such as dolphins (Aznar et al.,2003), which possess adaptive immunity. Both of these examplesare parasitic worms which induce an IgE/CD4 based immune re-sponse in vertebrates (MacDonald et al., 2002) and both examplesserve to illustrate the point that, despite large evolutionary dis-tances, host species can still share common pathogens.

If host–pathogen interactions of shared pathogens were suffi-cient to affect the evolution of immunity in gnathostomes and deu-terostomes, perhaps the reverse is also true. It is reasonable tohypothesize that as the types of immunity increased, there was amirrored increase in microbial diversity; more complex defensivesystems may have engendered more ways to circumvent such de-fenses. It is well documented that microbes modulate immune re-sponses in their favor by mimicking host signaling proteins(Epperson et al., 2012). If there were an increase in the complexity

A.E. Ward, B.M. Rosenthal / Infection, Genetics and Evolution 21 (2014) 492–496 495

of signaling pathways, pathogens would adapt to new niches.Additionally, perhaps if pathogen recognition pathways were ableto process signals with more subtlety, hosts would be able to dif-ferentiate between symbiotic and pathogenic microbes more easilyand therefore develop more diverse microbiomes. However, all ofthis relies on the expansion of host signaling capabilities.

1.4. Expansion of host signaling capabilities

Cytokines are a class of proteins that can be produced by allnucleated cells and generally have immuno-modulatory or hema-topoietic functions. Cytokines binding to specific cell surface cyto-kine receptors initiate downstream signaling events that regulatethe host immune system. Given their central role in shaping im-mune responses, especially adaptive immune responses, it is feasi-ble that cytokine receptors could have been influenced by thedevelopment of adaptive immunity. Analysis by Liongue and Ward(2007) of the sea squirt, mosquito, fruit fly, zebrafish, and humangenomes revealed only one cytokine receptor chain in mosquitos,two in sea squirts and fruit flies, and 36 cytokine receptor chainsin the zebrafish genome. They reason that the large expansion ofcytokine receptor chain diversity between urochordates and verte-brates occurred via a combination of whole genome, tandem, anden bloc duplication events.

Downstream of cytokine receptors, the components of the JAK/STAT pathway experienced a similar expansion. This is reasonableconsidering the pathway is used in many biological systems(including hematopoiesis, immune cell development, stem cellmaintenance, organismal growth, and mammary gland develop-ment), but is heavily involved in the signaling pathways of interfer-ons and cytokines (Rawlings et al., 2004). This expansion is evidentin that Drosophila possess only one member of the JAK family,while jawed fishes (specifically teleosts) possess five and humanspossess four. In a similar manner, Drosophila possess only onemember of the STAT family, sea squirts possess two members, tel-eosts possess eight members and humans possess seven members.Clearly, there has been an expansion of the number of componentsand complexity of this signaling pathway, mostly in telosts andtheir descendants. In-depth bioinformatics has revealed that therewas very little diversification of the JAK and STAT families beforethe divergence of urochordates. However, a major expansion coin-cided with the development of adaptive immunity in jawed fish, apattern that is parallel to that seen in cytokine receptor chains.There is also evidence that between the divergence of urochordatesand fish there were two whole genome duplication events (Lion-gue et al., 2012). Perhaps with the redundancy provided by a dupli-cated genome, changes in cytokine receptor chains and JAK/STATcomponents were more tolerated and could be positively selecteduntil they became new members of the signaling pathway. Then,even as many paralagous genes that were created by the wholegenome duplication were lost, the functionally distinct and posi-tively selected new cytokine receptor chains and JAK/STAT mem-bers were retained in the genome of vertebrates. These changestook place at the same time that key lymphoid cytokines wereevolving, so together they allowed more subtlety in regulation ofthe complex emerging adaptive immune system (Liongue et al.,2012).

This pattern of diversified cellular communication seems to beindicative of a larger pattern wherein vertebrates, rather thanevolving new PRRs, developed new ways to regulate the powerof their newfound adaptive immune systems. With more signalsand more ways to process and interpret those signals, adaptiveand innate immune cells could respond to infection with moresubtlety and nuance. Perhaps because the evolution of adaptiveimmunity required an expansion of cellular and intracellular sig-naling networks, there was reduced selective pressure on innate

immune PRRs. Rather than developing new innate immune recep-tors to initiate immunological signals, immune cells in organismsthat posses adaptive immune systems could process their currentset of inflammatory signals in more ways with increased signalingnetwork diversity.

The expansion of cytokine and JAK/STAT signaling, in conjunc-tion with the development of adaptive immunity, seems more con-sistent with the ‘‘deuterostome gain’’ hypothesis. If the evolutionof adaptive immunity caused the expansion of cellular communi-cation between the innate and adaptive immune systems, in-creased molecular cooperation would have ensued. The resultmay have been an increase in the efficacy of immune responses,which would thereby reduce the need for expansion of innatePRRs. Perhaps chordates, which did not experience the JAK/STATexpansion, had to compensate by diversifying their PRR repertoire.The expansion of the JAK/STAT signaling pathway would not ruleout the reduction in PRR effectors envisioned by the ‘‘gnathostomeloss’’ hypothesis, but rather a combination of the two mechanismsmay have created the current pattern of PRR and signaling networkdiversity. It is reasonable that in the wake of two whole-genomeduplication events, signaling network components faced positiveselection while PRRs faced strong purifying selection.

2. Conclusion

Through examination of both TLR and JAK/STAT diversity, it hasbecome apparent that the two mechanisms for the evolution of in-nate immunity, as influenced by adaptive immunity, are notequally plausible but still compatible. The ‘‘gnathostome loss’’explanation relies on the negative or purifying selection of innatePRRs, which is in contrast to extant evidence that PRRs maintainfunction in vertebrates. It is apparent that the major class of PRRs,the TLRs, had limited diversity in ancestral protostomes but haveretained structure and function. Moreover, the adaptive immunesystem requires activating signals from the innate immune systemto become effective at clearing infections. Those organisms thatdeveloped a costly and complex system for fighting infection main-tained the core PRRs with the ability to activate it effectively.

The ‘‘deuterostome gain’’ hypothesis posits, more plausibly, thatorganisms lacking adaptive immunity experienced increased pres-sure to diversify their PRRs, although notably in the absence ofwhole genome duplications making the massive expansion allthe more impressive. Since vertebrate adaptive immunity can gen-erate specific receptors for virtually any epitope, vertebrates relyless on their innate immune system to recognize all PAMPs whenneutralizing infectious agents. In vertebrates, positive selectionon the TLRs and the expansion of JAK/STAT signaling componentssubstantiate the notion that increased complexity of their cellularsignaling network reduced the pressure on their PRRs to furtherdiversify. Indeed, situating each PRR in a complex signaling net-work may have created conditions where further changes in PRRstructure would have disrupted their new signaling functions. Ver-tebrates still rely on innate immunity to help initiate adaptive im-mune responses, but the value of further diversification in PRRs hasbeen reduced or even reversed.

By contrast, organisms lacking adaptive immunity rely solely ontheir PRRs for immune surveillance; since their pathogens rapidlyevolve, the diversity of deuterostome PRRs must quickly evolve aswell. Situated as effectors of simpler signaling cascades, these PRRswould seem amenable to the continuous elaboration of newdiversity.

Gnathostome PRRs must have experienced negative selectionafter whole genome duplications (hence why mammals do nothave four times as many TLRs as invertebrates), although it seemsunlikely that purifying selection acting alone could cause the large

496 A.E. Ward, B.M. Rosenthal / Infection, Genetics and Evolution 21 (2014) 492–496

observed difference in the number of deuterostome TLRs and gna-thostome TLRs when already selecting against duplicated alleles.However, it does illustrate the point that the present pattern ofPRR diversity was likely caused by a combination of ‘‘deuterostomegain’’ and ‘‘gnathostome loss’’ in varying strengths. ‘‘Deuterostomegain’’ acting alone is still plausible but ‘‘gnathostome loss’’ wouldrequire extraordinary negative selection on gnathostome PRRs inthe wake of whole genome duplications; so if it played a role inshaping PRR diversity, it most likely acted in conjunction with‘‘deuterostome gain.’’

The precise reasons that deuterostomes posses more diversePRRs than do gnathostomes may never be known with certainty.Inferring the events impelling either loss or gain in extinct ances-tors is, admittedly, fraught with difficulty and is not amenable toeither direct observation or testing. Nonetheless, comparing vari-ous lines of evidence in extant descendants of each evolutionarylineage sheds light on the relative plausibility of possible hypothe-ses. Ultimately, the evidence to date favors the notion that deuter-ostomes underwent diversification in PRRs that was not alsoexperienced in those of our ancestors who passed on to us adaptiveimmunity. Even though gnathostomes evolved an entirely newsystem of host defense, we will still be plagued by pathogensand the red queen must keep running her eternal race.

Conflicts of interest

None.

Acknowledgement

The University Honors Program at The University of Maryland.

References

Aznar, J., Herreras, V., Balbuena, J., Raga, J., 2003. Population structure and habitatselection by Anisakis simplex in 4 odontocete species from northern Argentina.Comp. Parasitol. 70 (1), 66–71.

Areal, Helena, Abrantes, Joana, Esteves, Pedro, 2011. Signatures of positive selectionin Toll-like receptor (TLR) genes in mammals. Evol. Biol. 11 (368), 1471–2148.

Boehm, Thomas, McCurley, Nathaniel, Sutoh, Yuichu, Schorpp, Michael, Kasahara,Masanori, Cooper, Max D., 2012. VLR-based adaptive immunity. Annu. Rev.Immunol. 30, 203–220.

Epperson, M.L., Lee, C.A., Fremont, D.H., 2012. Subversion of cytokine networks byvirally encoded decoy receptors. Immunol. Rev. 250, 199–215.

Hedrick, Stephan, 2004. The acquired immune system: a vantage from beneath.Immunity 21, 607–615.

Huang, Shengfeng, Yung, Shaochun, Guo, Lei, 2008. Genomic analysis of theimmune gene repertoire of amphioxus reveals extraordinary innate complexityand diversity. Genome Res. 18, 1112–1126.

Janeway, Charles A., Medzhitov, Ruslan, 2002. Innate immune recognition. Annu.Rev. Immunol. 20 (1), 197–216.

Liongue, Clifford, Ward, Alister, 2007. Evolution of class I cytokine receptors. BMCEvol. Biol. 7 (120).

Liongue, Clifford, O’Sullivan, Lydia, Trengrove, Monique, Ward, Alister C., 2012.Evolution of JAK–STAT pathway components: mechanisms and role in immunesystem development. PLoS One 7 (3), 1–16.

MacDonald, Andrew S., Araujo, Maria Ilma, Pearce, Edward J., 2002. Immunology ofparasitic helminth infections. Infect. Immun. 70, 427–433.

McCurley, Nathanael, Hirano, Masayuki, Das, Sabyasachi, Cooper, Max, 2011.Immune related genes underpin the evolution of adaptive immunity injawless vertebrates. Curr. Genomics 13, 86–94.

McLain, A., 1952. Diseases and parasites of the sea lamprey, Petromyzon marinus, inthe Lake Huron basin. Trans. Am. Fish. Soc. 81, 94–100.

Nishana, Mayilaadumveettil, Raghavan, Sathees C., 2012. Role of recombinationactivating genes in the generation of antigen receptor diversity and beyond.Immunology 137, 271–281.

Palucka, Karolina, Banchereau, Jacques, 1999. Dendritic cells: a link between innateand adaptive immunity. J. Clin. Immunol. 19, 12–25.

Pancer, Z., Amemiya, C., Ehrhardt, G., Ceitlan, J., Gartland, L., Cooper, M., 2004.Somatic diversification of variable lymphocyte receptors in the agnathan sealamprey. Nature 430, 174–180.

Pascual, S., Gonzalez, A., Arias, C., Guerra, A., 1995. Helminth infection in the short-finned squid Illex coindetii (Cephalopoda, Ommastrephidae) off NW Spain. Dis.Aquat. Organ. 23, 71–75.

Rawlings, J., Rosler, K., Harrison, D., 2004. The JAK/STAT signaling pathway. J. CellSci. 117, 1281–1283.

Satake, H., Sekiguchi, T., 2012. Toll-like receptors of deuterostome invertebrates.Front. Immunol. 3 (34), 1–7.

Van Cleave, H.J., 1921. Acanthocephala from the Eel. Trans. Am. Microsc. Soc. 40 (1),1–13.

Van Cleave, H.J., 1919. Notes on the life cycle of two species of acanthocephala fromfreshwater fishes. J. Parasitol. 56 (4), 167–172.

Wang, E., Worschech, A., Marincola, F., 2008. The immunologic constant of rejection.Trends Immunol. 29 (6), 256–262.

Wiens, Matthias, Korzhev, Michael, Perovic-Ottstadt, Sanja, 2007. Toll-like receptorsare part of the innate immune defense system of sponges. Mol. Biol. Evol. 24 (3),792–804.

Wlasiuk, G., Nachman, M.W., 2010. Adaptation and constraint at Toll-like receptorsin primates. Mol. Biol. Evol. 27 (9), 2172–2186.