Multicellularity evolved well before 600 million years ago, and all multicellular animals have evolved since thenwith the need to protect against pathogens. There is no reason to expect their immune systems to be any lesssophisticated than ours. The vertebrate system, based on rearranging immunoglobulin-superfamily domains,appears to have evolved partly as a result of chance insertion of RAG genes by horizontal transfer. Remarkablysophisticated systems for expansion of immunological repertoire have evolved in parallel in many groups oforganisms. Vaccination of invertebrates against commercially important pathogens has been empirically suc-cessful, and suggests that the definition of an adaptive and innate immune system should no longer dependon the presence of memory and specificity, since these terms are hard to define in themselves. The evolutionof randomly-created immunological repertoire also carrieswith it the potential for generating autoreactive spec-ificities and consequent autoimmune damage.While invertebratesmay use systems analogous to ours to controlautoreactive specificities, they may have evolved alternative mechanisms which operate either at the level ofindividuals-within-populations rather than cells-within-individuals, by linking self-reactive specificities to reg-ulatory pathways and non-self-reactive to effector pathways.
Until recently, the prevailing paradigm for the evolution of theability to mount immune responses has been one in which a basic in-nate immune system evolved initially with limited repertoire. In thismodel the more sophisticated adaptive immune system evolved lateronly in the jawed vertebrates (Fig. 1) [1]. Corollaries of this paradigmare, firstly, that only the jawed vertebrates can be protected againstinfectious disease by a process of ‘vaccination’; and, secondly, thatonly the jawed vertebrates are likely to be susceptible to autoimmune
: +44 117 928 9505.).
sevier B.V.
ution of immune systems: Sp
disease. A simple examination of known phylogenetic trees makes itclear that this view is highly unlikely to be accurate, since mammalsdid not evolve from modern-day insects or sponges, but from com-mon ancestors more than 600 million years ago (mya) [2]. Since allthree of these groups have been subjected to natural selection by expo-sure to pathogens over the same period of time, it would be perfectlyreasonable to expect that immune systems might evolve to the sameextent in all three, but likely in different directions.
In fact, accumulating evidence suggests that this is exactly whathas happened: multiple mechanisms for expansion of repertoireseem to have evolved in many animal phyla. Although their evolutionhas been independent, the systems currently known have tended touse the same basic molecular building blocks. Presumably, also, they
ecificity and autoreactivity, Autoimmun Rev (2012), http://dx.doi.org/
Fig. 1. Common concepts describing the evolution of the immune system. Highly schematic phylogenetic trees demonstrating (A) increasing sophistication of the immune systemfrom ‘primitive’ invertebrates to the more sophisticated mammals. In this model, the immune systems of molluscs and arthropods are essentially a subset of the immune system ofmammals. (B) Increasing sophistication of the immune system with geological time. In this model, the immune systems of molluscs and insects are likely to be highly developed,but along very different lines to that of the mammals. (C) Major events in the evolution of the immunoglobulin-based immune system of the jawed vertebrates. Several indepen-dent events seem to have been involved, without any one of which our immune system would have been likely to develop very differently, as demonstrated by the differentsolutions evolved in the hagfish and lampreys (leucine-rich repeat molecules) and lower chordates.
2 M. Bailey et al. / Autoimmunity Reviews xxx (2012) xxx–xxx
have dealt with similar problems in terms of expression of repertoireacross populations of cells or of individuals, of recognition of self andnon-self, and of distinguishing between ‘dangerous’ and ‘harmless’non-self. An understanding of these independently-evolved solutionsto the same problem can help to identify necessary, common charac-teristics of an ‘adaptive’ immune system which our immune systemmust have.
2. The prototypic adaptive immune system
The prototypic adaptive immune system involving rearranging im-munoglobulin superfamily (IgSF) receptors appears to have evolved atthe same time as the appearance of the jaw in the vertebrates (Fig. 1)[3]. Immunoglobulin heavy and light chain loci, TCR α, β, γ and δ, areall present in the sharks, skates and rays, in the bony fish, and in allland tetrapods [4,5]. The strong probability is that these receptorsevolved from an earlier germ-line encoded IgSF pattern-recognition re-ceptor which acquired the ability to generate an expanded repertoire,firstly by undergoing somatic hypermutation and subsequentlyrearrangement. Such a receptor is present in themodern-day chordates
Please cite this article as: Bailey M, et al, Evolution of immune systems: Sp10.1016/j.autrev.2012.10.007
Branchiostoma and Ciona, often regarded as close to the ‘primitive’ancestors of the vertebrates [6,7]. The variable chitin-binding pro-teins (VCBPs) in these organisms include two variable-domain-typeIgSF domains which are highly polymorphic and also continuously ge-netically variable as a consequence of pointmutations, genetic exchangeand gene duplication [7].
Rearrangement of higher vertebrate IgSF loci is completely depen-dent on the presence of the rearrangement-associated genes RAG1and RAG2, which have been highly conserved through the vertebrates(approximately 50% similarity between shark and human sequences).The situation is complicated by RAG-like genes in invertebrates, butnothing similar has been found in the jawless fish, and current evidencesuggests that the RAG genes were acquired by horizontal transmissionofmicrobial integrase genes through the action of a transposon, retrovi-rus or herpesvirus, 400–450 mya [8,9]. Insertion of recombination sig-nal sequences into a pre-existing, somatically hypermutating pattern-recognition receptor would increase the potential repertoire generatedby allowing non-germline-encoded junctional diversity. The continuingpresence of multicluster IgH loci in sharks [(V–D–D–J–C)n] as well asthe more familiar translocon arrangements [Vn–(J–C)n, Vn–Jn–C and
ecificity and autoreactivity, Autoimmun Rev (2012), http://dx.doi.org/
3M. Bailey et al. / Autoimmunity Reviews xxx (2012) xxx–xxx
Vn–Dn–Jn–C] strongly supports the idea of an initial V–J–C structurewhich further expanded repertoire by duplicating either individualsegments or complete loci [5].
3. Through the looking glass: the way it might have been
However, it now appears that the dependency of our adaptive im-mune system on IgSF domains was only one of at least two possibili-ties present in the early vertebrates. Recent studies have identified anextremely sophisticated, rearranging adaptive immune system in thejawless fish (the lampreys and hagfish) based on leucine-rich repeat(LRR) molecules [10,11]. As with the IgSF adaptive immune system,this may have evolved from an earlier use of LRR molecules as pattern-recognition receptors with fixed repertoire, following a similar pathwayof limited expansion of repertoire before the appearance of full re-arrangement. Despite the apparent similarity amongst the LRR re-ceptors, the prototypic Toll in insects and the vertebrate Toll-likereceptors (TLRs), it now appears that these domains were co-optedinto immunological function independently in the two groups, presum-ably because of an underlying suitability for use in molecular recogni-tion (as in the IgSF domains) [12]. In the sea urchins, closely related tothe chordates, massive expansion of pattern-recognition receptor locihas occurred: the purple sea urchin genome contains approximately200 TLR loci [13]. This appears not to be an isolated phenomenon: inaddition to its IgSF-based VCBPs, the protochordate Branchiostoma hasalso expanded its TLR repertoire. Interestingly, Branchiostoma has alsoextensively duplicated the loci encoding TLR signalling molecules, rais-ing the possibility that the type of response initiated by ligation of thedifferent TLRs may also vary more than ours [14]. Thus, this appar-ently ‘primitive’ organism actually has at least two quite sophisticatedmechanisms for generating a relatively large repertoire of recognitionmolecules.
The pathways by which an earlier LRR pattern-recognition recep-tor evolved into a fully rearranging locus are not clear, but the systempresent in lampreys and hagfish has now been well characterised andincludes at least three loci, VLR-A, -B, and -C, in which upstream anddownstream LRR cassettes are inserted randomly into an expressionsite by a process of gene conversion, producing an extremely large,non-germ-line encoded repertoire [15]. Thus, somewhere aroundthe point where the jawed and jawless fish diverged, there appearto have been at least two alternative pathways by which a fully adap-tive immune system could have developed.
4. Unconventional mechanisms for expansion of immunologicalrepertoire
In fact, it seems likely that the many evolutionary pathways leadto diversification of immunological repertoire. At least two other,unrelated mechanisms have been described, in the insects and crus-taceans (indicating that it is likely to be present in all arthropods)[16,17] and one originally described in snails (molluscs, related tothe octopi and squids) but which may be much more widespread[18–20].
The system in snails involves fibrinogen-related proteins (FREPs)of which the best-studied, FREP-3 includes two IgSF domains whichare highly polymorphic, somatically diversified, and appear to engagein high levels of crossing over at meiosis [18]. As a consequence, therepertoire of FREPs expressed varies markedly between individualsnails within a population. Knock-down experiments have confirmedthat FREP-3 is important in resistance of snails to the intermediatestages of the human pathogen Schistosoma mansoni [21]. However,the mechanism by which repertoire is generated suggests that thefull repertoire of FREPs may be expressed at the level of the population,rather than of individuals. That is, each snail may express a differentsubset of the total, possible repertoire, and selection by pathogen resultsin survival to reproduce only of those individuals with appropriate
Please cite this article as: Bailey M, et al, Evolution of immune systems: Sp10.1016/j.autrev.2012.10.007
repertoires (Fig. 2). This may be an appropriate strategy for specieswhich produce large numbers of offspring with a high mortalityrate. However, extreme selection pressure applied by one set of patho-gens could very quickly result in vulnerability to other, rapidly-evolvingpathogens capable of escaping the selected repertoire. It seems likelythat the value of crossing-over at meiosis is in ‘re-shuffling’ the reper-toire in each generation to prevent this over-selection.
In the arthropods, it has become apparent that a previously de-scribed locus involved in neuronal development, encoding the Downsyndrome cell adhesionmolecule (Dscam) is also involved inmediatingimmunity. The Dscam locus encodes multiple copies of several exons,which can be variably incorporated into the final translated mRNA byalternative splicing. In Drosophila, the potential combinatorial diversityof this system is 38,016 concentrated, interestingly, into two IgSF do-mains [16]. These molecules directly bind bacteria and have beenimplicated in resistance to bacterial infections [22]. They may also ex-plain the considerable recent literature describing the surprising, empir-ical success of vaccination in at least one group of invertebrates,commercial shrimp, against a particularly pathogenic virus, althoughthe role of Dscam in clearance of viral infectionshas not been established[23,22].
5. The implications of multiple, independently-evolved ‘adaptive’immune systems
The finding that there are at least two, if not more, independently-evolved ‘adaptive’ immune systems allows us to re-evaluate a num-ber of questions to which we previously believed we had workableanswers:
• is some form of expansion of repertoire, whether innate or adap-tive, the norm across animal phyla rather than exclusive to verte-brates?
• what characteristics define an ‘adaptive’ immune system and dis-tinguish it from an ‘innate’ immune system?
• to what extent does expansion of pathogen-specific repertoire re-sult in expansion of self-specific repertoire, and is autoimmunityan inevitable consequence?
• is ‘vaccination’ a strategy worth re-examining in invertebrates toprotect them against their own pathogens or to interrupt transmis-sion of human pathogens?
5.1. What constitutes an adaptive or innate immune system?
Briefly, ‘memory’ and ‘specificity’ do not qualitatively distinguishthe adaptive immune system of jawed fish from other, recently iden-tified innate immune systems. Increased resistance to pathogens onsecondary challenge is seen in invertebrates, as a consequence of ac-tivation of effector mechanisms with some degree of specificity dur-ing primary infection [24]. Similarly, the repertoire of many ‘innate’immune systems is much larger than previously thought. The mostuseful definition is likely to be based on the ability to generate expan-sion of specific cells after recognition of ‘antigen’. By this definition, thesystems present in jawed and jawless vertebrates are clearly adaptive,the somatically hypermutating VCBPs in basal chordates and FREPsin molluscs may or may not be, while the Dscam of arthropods is theleast likely.
5.2. Autoimmunity
Clearly,whenever the repertoire of possible immunological receptorsis randomly generated (by rearrangement or somatic hypermutation),there is a potential for generating autoreactive specificities. In the proto-typic adaptive immune systemof the jawed vertebrates,mechanisms fornegative selection in the thymus, bone marrow, bursa of Fabricius andgerminal centres are relatively well established [25,26]. A comparable
ecificity and autoreactivity, Autoimmun Rev (2012), http://dx.doi.org/
Re-arrangement of IgSF Recombination ofRe ofgenes in somatic lymphocyte precursors
germ-line genes during reproduction
Removal of autoreactive specificities
Expansion of pathogen-reactive specificities
Removal of non-pathogen-reactive
specificities
Fig. 2. Proposed model for generation of immunological repertoire within individuals and at the population level. Open symbols, neutral repertoires; grey symbols, pathogen-reactive repertoires; black symbols, self-reactive repertoires. (A) Immunoglobulin repertoire is generated by random rearrangement of lymphocyte precursors in a developing,individual human: self-reactive specificities are eliminated in the bone marrow, and clonal expansion of pathogen-reactive cells results in skewed representation in the population.(B) Repertoire is generated within a population of snails by recombination of FREP genes during reproduction: snails with repertoires including self-reactive specificities are selec-tively disadvantaged, as are those without appropriate pathogen-reactive specificities.
4 M. Bailey et al. / Autoimmunity Reviews xxx (2012) xxx–xxx
mechanism for deletion of autoreactive VLR-A lymphocytes in lampreyhas been proposed in a structure at least partially homologous to thethymus of jawed fish [27], and this may indicate an earlier evolutionaryhistory in common with the IgSF-based system of jawed vertebrates.However, it also seems likely that some sort of mechanism must alsoexist for regulating autoreactive VCBPs in chordates and FREPs in mol-luscs. Since these systems appear to have evolved independently, theymay also have evolved unrecognisable mechanisms for immunoregula-tion, and there are several obvious possibilities:
• Limitations on structure may prevent the generation of autoreactivespecificities. This would depend on the ‘set’ of auto-epitopes beingnon-overlapping with the ‘set’ of pathogen epitopes. Such a mecha-nism might function effectively if specificity remained restrictedto classes of pathogen-associated molecular patterns not representedin hostmolecules (classical PRR–PAMP interactions). Variationwouldneed to be restricted to prevent acquisition of broader reactivity, lim-iting the value of such a mechanism of repertoire expansion.
• Negative selection might operate at the level of the whole organismrather than individual cells within an organism. If the repertoire re-mains fairly limited and the probability of generating autoreactivespecificities is low, any individual who does so may simply notsurvive— effectively dying from autoimmune disease. Again, a mech-anism such as this might work perfectly well in those groups whichproduce large numbers of offspring: both pathogen repertoire andautoreactive repertoire might be controlled at the level of the popu-lation rather than the individual.
• A more sophisticated possibility would be the selective linking ofautoreactive specificities to negative immunoregulatorymechanisms.The self-specific, inhibitory NK receptors of the land tetrapods func-tion in this way by linkage to an intracellular negative signalling path-way. The mechanisms by which marine sponges reject transplantsseem to involve a similar kind of self-recognition system allowingrejection of ‘missing self’ [28,29]. In addition, it seems likely that atleast a subset of thymically-derived regulatory T-cells are autoreactive,but are pushed into a differentiation program linking TCR ligation
Please cite this article as: Bailey M, et al, Evolution of immune systems: Sp10.1016/j.autrev.2012.10.007
to pathways mediating intercellular immunoregulation, ratherthan into the apoptosis associated with conventional negative se-lection [30,31]. A mechanism in which the default pathway is todifferentiate into an immunoeffector cell, but autoreactive specific-ities are pushed into immunoregulatory pathways,might avoid the sig-nificant cost of negative selection. Interestingly, recent studies havedemonstrated pathogen specific receptors within loci encoding NK re-ceptors, together with some form of NK ‘memory’ on re-challenge [32].The speedwithwhichNK receptors appear to be evolving suggests thattheymay be quite analogous to the FREPs or VCBPs [33]. Thus, receptorsystems which encode both self-recognition and non-self-recognitionmolecules are not uncommon.
5.3. Vaccination
Several studies have described empirical phenomena which parallelthat of ‘memory’ in the prototypic, IgSF-based, adaptive immunesystem [24]. The simplest mechanism which might account for thesephenomena is probably transient upregulation of effector mechanismswith some level of specificity. For example, many of the antimicrobialpeptides produced by insects have partially restricted specificity, e.g.,for Gram positive bacteria, conferring a level of protection against otherGram positive bacteria during the recovery phase, but not againstGram negative bacteria or fungi [34]. However, the identification ofmechanisms for expanding repertoire in several groups of invertebrates,together with an expanding literature on their ‘vaccination’ [35,36],raises the possibility that future research onmodulating the immune re-sponses of invertebrates of commercial importance (e.g. pollinators orintermediate hosts of human or animal pathogens) may be worthexploring.
6. Conclusions
It now seems clear that selective pressure to diversify the repertoireof pathogen-associated molecular patterns which can be recognisedby an immune system has been continuously operating over hundreds
ecificity and autoreactivity, Autoimmun Rev (2012), http://dx.doi.org/
5M. Bailey et al. / Autoimmunity Reviews xxx (2012) xxx–xxx
of millions of years in all animal groups, rather than only in a smallgroup of vertebrates as they acquired a jaw. This pressure has drivenduplication of receptor loci and the independent evolution of severaldifferent mechanisms for generating non-germline-encoded receptors.Once this occurs, mechanisms for limiting the potential damage causedby autoreactive specificitiesmust also begin to evolve in parallel. Initial-ly, such mechanisms may be relatively crude, such as natural selectionagainst the individuals which create them, and this may be satisfactoryin species which overproduce offspring. However, a more efficient sys-tem might be to link autoreactive specificities to immunoregulatoryrather than immunoeffector pathways, and such a system might ac-count for mammalian natural Tregs. In future, studies of immune sys-tems outside the jawed vertebrates have the potential to impact onour understanding of and ability to control immune responses not onlyin humans and in our domesticated species, but also in a range of inver-tebrate species whose lives affect our own, either directly or indirectly.
Take home messages
• All current species have had the same time to evolve under pathogenchallenge, and we should expect comparable levels of sophisticationof their immune systems.
• Our adaptive immune system (the ‘prototypic’ system) based on im-munoglobulin superfamily molecules evolved in the jawed fish.
• Several alternative approaches to randomly expanding repertoirehave recently been described both in vertebrates and in invertebrates.
• These organisms must face similar problems with deletion ofautoreactive specificities, but have solved them in different ways.
• Vaccination of invertebrates may be an important area for futureresearch.
References
[1] Medzhitov R, Janeway CAJ. An ancient system of host defense. Curr Opin Immunol1998;10:12–5.
[2] Pick KS, Philippe H, Schreiber F, Erpenbeck D, Jackson DJ, Wrede P, et al. Improvedphylogenomic taxon sampling noticeably affects nonbilaterian relationships. MolBiol Evol 2010;27:1983–7.
[3] Rast JP, Litman GW. Towards understanding the evolutionary origins and earlydiversification of rearranging antigen receptors. Immunol Rev 1998;166:79–86.
[4] Criscitiello MF, Ohta Y, Saltis M, McKinney EC, Flajnik MF. Evolutionarily conservedTCR binding sites, identification of T cells in primary lymphoid tissues, and surprisingtrans-rearrangements in nurse shark. J Immunol 2010;184:6950–60.
[5] Lee V, Huang JL, Lui MF, Malecek K, Ohta Y, Mooers A, et al. The evolution of mul-tiple isotypic IgM heavy chain genes in the shark. J Immunol 2008;180:7461–70.
[6] Dishaw LJ, Giacomelli S, Melillo D, Zucchetti I, Haire RN, Natale L, et al. A role forvariable region-containing chitin-binding proteins (VCBPs) in host gut-bacteriainteractions. Proc Natl Acad Sci U S A 2011;108:16747–52.
[7] Dishaw LJ, Ota T, Mueller MG, Cannon JP, Haire RN, Gwatney NR, et al. The basisfor haplotype complexity in VCBPs, an immune-type receptor in amphioxus. Im-munogenetics 2010;62:623–31.
[8] Agrawal A, Eastman QM, Schatz DG. Transpositionmediated by RAG1 and RAG2 andits implications for the evolution of the immune system. Nature 1998;394:744–51.
[9] Litman GW, Rast JP, Fugmann SD. The origins of vertebrate adaptive immunity.Nat Rev Immunol 2010;10:543–53.
[10] Pancer Z, Amemiya CT, Ehrhardt GRA, Ceitlin J, Gartland GL, Cooper MD. Somaticdiversification of variable lymphocyte receptors in the agnathan sea lamprey.Nature 2004;430:174–80.
Please cite this article as: Bailey M, et al, Evolution of immune systems: Sp10.1016/j.autrev.2012.10.007
[11] Han BW, Herrin BR, Cooper MD, Wilson IA. Antigen recognition by variable lym-phocyte receptors. Science 2008;321:1834–7.
[12] Leulier F, Lemaitre B. Toll-like receptors — taking an evolutionary approach. NatRev Genet 2008;9:165–78.
[13] Hibino T, Loza-Coll M, Messier C, Majeske AJ, Cohen AH, Terwilliger DP, et al.The immune gene repertoire encoded in the purple sea urchin genome. Dev Biol2006;300:349–65.
[14] Huang S, Yuan S, Guo L, Yu Y, Li J, Wu T, et al. Genomic analysis of the immunegene repertoire of amphioxus reveals extraordinary innate complexity and diver-sity. Genome Res 2008;18:1112–26.
[15] Boehm T, McCurley N, Sutoh Y, Schorpp M, Kasahara M, Cooper MD. VLR-basedadaptive immunity. Annu Rev Immunol 2012;30:203–20.
[16] Watson FL, Püttmann-Holgado R, Thomas F, Lamar DL, Hughes M, Kondo M, et al.Extensive diversity of Ig-superfamily proteins in the immune system of insects.Science 2005;309:1874–8.
[17] Armitage SA, Freiburg RY, Kurtz J, Bravo IG. The evolution of Dscam genes acrossthe arthropods. BMC Evol Biol 2012;12:53.
[18] Zhang S, Adema CM, Kepler TB, Loker ES. Diversification of Ig superfamily genes inan invertebrate. Science 2004;305:251–4.
[19] Dong Y, Dimopoulos G. Anopheles fibrinogen-related proteins provide expandedpattern recognition capacity against bacteria and malaria parasites. J Biol Chem2009;284:9835–44.
[20] Fan C, Zhang S, Li L, Chao Y. Fibrinogen-related protein from amphioxusBranchiostoma belcheri is a multivalent pattern recognition receptor with a bacte-riolytic activity. Mol Immunol 2008;45:3338–46.
[21] Hanington PC, Forys MA, Loker ES. A somatically diversified defense factor, FREP3,is a determinant of snail resistance to schistosome infection. PLoS Negl Trop Dis2012;6:e1591.
[22] Watthanasurorot A, Jiravanichpaisal P, Liu H, Söderhäll I, Söderhäll K. Bacteria-induced Dscam isoforms of the crustacean, Pacifastacus leniusculus. PLoS Pathog2011;7:e1002062.
[23] Rowley AF, Pope EC. Vaccines and crustacean aquaculture — a mechanistic explo-ration. Aquaculture 2012;334:1–11.
[24] Kurtz J, Franz K. Innate defence: evidence for memory in invertebrate immunity.Nature 2003;425:37–8.
[25] Metzger TC, Anderson MS. Control of central and peripheral tolerance by Aire.Immunol Rev 2011;241:89–103.
[26] Pelanda R, Torres RM. Central B-cell tolerance: where selection begins. ColdSpring Harb Perspect Biol 2012;4:a007146.
[27] Bajoghli B, Guo P, Aghaallaei N, Hirano M, Strohmeier C, McCurley N, et al. A thy-mus candidate in lampreys. Nature 2011;470:90–4.
[28] Haseley SR, Vermeer HJ, Kamerling JP, Vliegenthart JF. Carbohydrate self-recognitionmediates marine sponge cellular adhesion. Proc Natl Acad Sci U S A 2001;98:9419–24.
[29] Sabella C, Faszewski E, Himic L, Colpitts KM, Kaltenbach J, Burger MM, et al. Cyclo-sporin A suspends transplantation reactions in the marine sponge Microcionaprolifera. J Immunol 2007;179:5927–35.
[30] Hsieh C, Lee H, Lio CJ. Selection of regulatory T cells in the thymus. Nat RevImmunol 2012;12:157–67.
[31] Picca CC, Simons DM, Oh S, Aitken M, Perng OA, Mergenthaler C, et al. CD4(+)CD25(+)Foxp3(+) regulatory T cell formation requires more specific recognitionof a self-peptide than thymocyte deletion. Proc Natl Acad Sci U S A 2011;108:14890–5.
[32] Sun JC, Madera S, Bezman NA, Beilke JN, Kaplan MH, Lanier LL. Proinflammatory cy-tokine signaling required for the generation of natural killer cell memory. J Exp Med2012;209:947–54.
[33] Parham P, Norman PJ, Abi-Rached L, Guethlein LA. Variable NK cell receptorsexemplified by human KIR3DL1/S1. J Immunol 2011;187:11–9.
[34] Uvell H, Engström Y. A multilayered defense against infection: combinatorialcontrol of insect immune genes. Trends Genet 2007;23:342–9.
[35] Ning D, Leng X, Li Q, Xu W. Surface-displayed VP28 on Bacillus subtilis spores in-duce protection against white spot syndrome virus in crayfish by oral administra-tion. J Appl Microbiol 2011;111:1327–36.
[36] Syed MS, Kwang J. Oral vaccination of baculovirus-expressed VP28 displays en-hanced protection against white spot syndrome virus in Penaeus monodon. PLoSOne 2011;6:e26428.
ecificity and autoreactivity, Autoimmun Rev (2012), http://dx.doi.org/
