Evolution of the Immune System

L Du Pasquier, University of Basel, Basel, Switzerland

2014 Elsevier Inc. All rights reserved.

Glossary
Activation-induced deaminase (AID) Enzyme of the
APOBEC family that removes, in the DNA, an amino group
from a cytidine, turning it into a uracil that has to be
replaced. Plays roles in somatic mutation, class switch, and
somatic gene conversion.
Adaptive immunity Immunity acquired within the
individual following an internal Darwinian selection of
randomly generated receptor repertoire.
Agnathan A vertebrate without jaws, for example the
lamprey.
Alternative spicing A mechanism by which different
isoforms of messenger RNAs are generated from the same
gene.
C-type lectin A protein domain, 110115 amino acids
long, binding carbohydrate in a calcium-dependent
manner. Consists of ve short beta strands, two short
helices, and the rest (50%) is an extended structure.
CD The CD nomenclature (cluster of differentiation) of
cell surface molecules is widely used in Immunology.
The following link shows a chart of the human CDs:
http://www.bdbiosciences.com/documents/Human_CD_
Chart.pdf.
Complement An ensemble of proteins that, once activated
in solution (blood), generates various effectors of innate
immunity (e.g., membrane attack complex and opsonins)
via successive protease actions.
Cytokine A small protein secreted by many cell types and
important for intercellular communication; an
immunomodulating agent.
Deuterostome Triploblastic Metazoa where the rst
natural opening to appear during development is the anus.
Dual oxidase Enzyme of the NAPDH-oxidase (peroxidase)
family involved in the production of reactive oxygen
species.
Gnathostome Jawed vertebrate.

Introduction

At some stage of evolution one or more species must have made the
momentous discovery that a very convenient source of food consists
of the tissues of other organisms. And whatever the rst form of life
on earth may have been we nd at the present time that all organisms have provided themselves with some membrane or shield to
protect themselves against invasion from the outside.
(William C. Boyd, Fundamentals of Immunology, 1956)

Because immunity is of such importance, it is understandable that it is not provided to our body by a single system.

Reference Module in Biomedical Research, 3rd edition

Immunoglobulin domain A protein domain roughly 100

amino acids long consisting of a two-layer beta-pleated
sheet sandwich.
Innate immunity Immunity depending entirely on the
expression of germline genes without somatic adaptations.
Interferon A family of cytokines very important for ghting
viral infection.
Interleukin Cytokines essentially produced by leukocytes.
Leucine-rich repeat A protein domain consisting of 2030
amino acid repeats that are rich in hydrophobic residues,
mainly leucines; the assembled domain has a horseshoe
shape.
Metazoan Multicellular organisms.
NLR NOD-like receptors; cytoplasmic proteins playing
a role in apoptosis and inammation, some of which can
function as intracellular pattern recognition receptors
(PRRs).
Nucleotide-binding oligomerization domain (NOD) A
leucine-rich repeat moiety and then the functionally
differentiating domain.
Pattern recognition receptor (PRR) Proteins that
recognize conserved epitope on microbial pathogens.
Prophenoloxidase (PPO) system A system generating via
enzymatic cascades the enzyme phenoloxidase, resulting in
the melanization of pathogens or damaged tissues.
Protostome Triploblastic organisms (i.e., composed of the
three fundamental layers: ectoderm, mesoderm, and
endoderm) where the rst natural opening to appear
during development is the mouth.
Reactive oxygen species (ROS) Chemically reactive
molecules containing oxygen; have a destructive effect on
invading pathogens.
Toll-like receptors LRR membrane surface receptors
recognizing structurally conserved epitopes present in
microbes.

Nowadays, one often reads that the human immune system

(Figure 1) consists of an innate arm inherited from primitive
species that provides a rst line of defense, and an adaptive arm
that is restricted to vertebrates and responsible for specicity
and memory. In fact, innate immunity is provided by
a constellation of different systems. These are often not related
to each other; they may operate at quite different topographical levels in the body, intracellularly or extracellularly; and
they were probably selected independently and coevolved in
a complex way with the adaptive arm of the system.
Therefore, a comparative study of our immune system with
an attempt at reconstituting its buildup during the evolution of

http://dx.doi.org/10.1016/B978-0-12-801238-3.00110-0

Evolution of the Immune System

Figure 1 Schematic representation of immunity in humans. Abs: Antibodies, AID: Activation-induced deaminase, AMP: Antimicrobial peptides,
APC: Antigen-presenting cell, B: B lymphocyte, BM: Bone marrow, C0 : Complement, CTL: Cytotoxic T lymphocyte, Fragment crystallizable (Fc)
receptor: Receptor binding the crystallizable fragment of the Ig molecule, IFN: Interferon, Il1-R: Interleukin 1 receptor, Igsf: Immunoglobulin superfamily, ITIM: Immunotyrosine inhibitory motif, LRR: Leucine-rich repeat, MHC: Major histocompatibility complex, NLR: NOD-like receptors, PDC:
Plasmacytoid dendritic cell, PKR: Protein kinase RNA activated, RAG1 and RAG2: Recombination-activating genes, RIG-1: Retinoic acidinducible
gene, RNAi: RNA interference, T: T lymphocyte, TLR: Toll-like receptor, TRIM: Tripartite motif-containing receptor.

Metazoa helps distinguish the essential from the accessory in

the complex assembly of its elements.

Early Evolution of Metazoa: Recruitment of Sensors

and Receptors, and Establishment of Signal
Transduction toward an Effector Mechanism
Some of the selection pressures and constraints that shaped the
evolution of immune systems of Metazoa (Figure 2) are presented in Table 1: One can see (in italics) that some of the
pressures will be met better by innate or adaptive characteristics
of the immune system.
Oceans, a major source of microorganisms, were plausibly
the environment in which the rst Metazoa appeared. Therefore, several hundred million years ago, they must have been
exposed to multiple and simultaneous threats to their individualities: bacteria, viruses, Protozoa, Fungi, Archaea, and
competing individuals of the same species. In addition, they
were exposed to putatively dangerous variants of their own self
due to mutations. Such a situation must have favored the early
development of a diversied immune system.
If avoiding contact with pathogens is impossible, immunity
can be achieved inside our body in the humoral environment
(e.g., soluble mediators with a complement or complementlike cascade (Table 2), or other enzymatic cascades such as
the prophenoloxidase pathway), ending in the generation of
a product directly or indirectly detrimental for the pathogen
(Cerenius et al., 2010).
The immune response can also be local: Each cell of the
organism has the possibility to launch cell-autonomous effector

mechanisms (MacMicking, 2012). In simple Metazoa like

Hydra, this takes place at the epithelium level (Augustin
et al., 2010).
In complex Metazoa, immunity can be the result of several
additional mechanisms involving specialized tissues and
mobile cells. Those mobile cells, the hemocytes or lymphocytes, originate from the mesoderm-derived hemopoietic
compartment. They are often able to phagocytose pathogens.
They patrol the body, and therefore they exert surveillance
and, later in the response, help launch a systemic reaction.
The rst event in which these cells participate is recognition
of the environment by receptors on their membrane, followed by a signal transduction and an effector phase in the
nucleus leading to the activation of a specic immunityrelated gene.

Recruitment of Receptors
Modern Protozoa that have a metazoan phase, like Dictyostelium, may suggest how some elements of the immune systems
were recruited. Dictyostelium exhibits a sorting capacity that is
indicative of self-recognition mechanisms governing phagocytosis. During the sexual cycle of this species, zygote giant cells
chemoattract and phagocytose hundreds of local amoebae.
This sexual phagocytosis is species and developmental stage
specic, and the process is mediated by receptors able to
distinguish self from nonself. Metazoa may have proted from
this distinction capacity and adapted it to protection of the
individual. Indeed, Protozoa and invertebrates share related
phagocyte receptors of the type found in Dictyostelium (e.g.,
pHG1-like) (Cornillon et al., 2000), but sharing is more

Evolution of the Immune System

Petromyzon

Agnathans
Cephalochordates

Branchiostoma
Ciona

Urochordates

Anus

Vertebrates

Gnathostomes

Hemichordates

Deuterostome
Echinoderms

Strongylocentrotus

Plathyhelminths
Annelids

Mouth
Protostome

Invertebrates

Mollusks

Biomphalaria

Nemerteans
Sipunculids
Nematodes
Arthropods

Anopheles

Drosophila

Penaeus

Cnidaria

Protozoa

Porifera
Placozoa

Figure 2

A simplied representation of Metazoa.

Table 1
Selection pressures exerted on the immune systems. An
ideal immune system
Should be rapid and efcient at eliminating the danger
Innate, constitutive elements
Should confer to each individual within a species the best chances of
survival under changing environment conditions
Somatic adaptations
Should have diversity and exibility
Many solutions: gene families, many molecular categories, and somatic
adaptations
Should be regulated (up- and downregulation)
Regulatory networks
Should be specic
Somatic adaptations and large repertoires
Should be economical (i.e., should not use too much genetic material or
too many cells)
Somatic adaptations and combinatorial usage of elements
Memory and transfer of protection to the progeny could be useful
Clonal selection, inheritable specificities, and maternal transfer of antibodies
Should not react with self; should allow symbiotic organisms
Selection
Central and peripheral tolerance
Modulations of the above are likely to be observed under the inuence of
many traits of the life histories of the organisms: genome size, level of
genomic instability, metabolic rates, progeny size, age at reproduction,
lifespan, acquisition or not of symbionts, and so on.

Table 2
Conservation of complement components and function
among Metazoa
Thioester

Function

Homologies

C3, from diploblastic

C4, in gnathostomes
Alpha2M, all Metazoa
Thioesterbond-containing protein (TEP),
in protostomes
Inammation
In gnathostomes
Opsonization
Via C3b from diploblastic
Via TEP in protostomes
Cell lysis
Via the membrane attack complex (MAC)
C5bC9 in gnathostomes
Via MAC perforin (MACPF) Metazoa
C2, deuterostomes
C3, Bf from diploblastic
C1r, C1s, membrane attack perforin (MASP)
C1q, mannan-binding lectin-associated serine
protease (MBL)
Deuterostomes

Evolution of the Immune System

conspicuous when comparing phyla within Metazoa. During

Drosophila development, apoptotic cells are phagocytosed after
recognition by evolutionarily conserved scavenger receptors of
the family Draper, Eater, and Nimrod (a homolog of Ced1 of
Caenorhabditis elegans) family with multiple epithelial growth
factor (MEGF) domains. This family has homologs in the
vertebrate MEGF 10 molecules. The scavenger receptor B,
Croquemort, of the vertebrate cluster of differentiation 36
(CD36) family is expressed on y and other invertebrate
macrophages (Ferrandon et al., 2007).
Finally, the adaptation of phagocytosis to the Metazoa
immune system ended up equipping the surface of phagocytic
cells with a diversity of receptors, allowing the multiple
recognition events imposed by the challenges discussed here.
This diversity also offers the possibility of combinatorial usage
of elements, which multiplies the number of recognition
possibilities.
The molecules recruited as receptors have to obey certain
criteria: to be resistant to proteolysis and to pH changes, to
have certain binding capacities, to be able to stand on the
outside of a cell, to transduce a signal, to be part of an enzymatic cascade in solution, and so on.

Under such constraints, the actual number of protein categories recruited is restricted compared to the available number
of protein folds (Figure 3). Lectins, leucine-rich repeats, the
immunoglobulin superfamily (Igsf), epithelial growth factor
(EGF) receptor, and members of the complement family, all
with different structures and properties, were among the most
successful receptor and sensor families. In line with the
working hypothesis proposed here according to which, from
the onset of Metazoa, the immune system could be already
complex Porifera (Placozoa have not yet been investigated)
do have an available diversity of elements that are similar or
even homologous to those of vertebrates, although they may
not use them in an analogous way.
In term of specicity, innate immunity uses predominantly
a restricted set of germline-encoded receptors, generally known
as pattern recognition receptors (PRRs), which have been
selected during evolution because they recognize a relatively
small number of epitopes but are distributed on essential
molecules of the pathogens that cannot vary too much without
threatening their survival. These are called pathogen-associated
molecular patterns (PAMPs) (Janeway and Medzhitov, 2002).
In adaptive immune systems, with the introduction of somatic

LRR

IgS
F

Lectin

SRCR

Fn3

EGF

PGRP

TIR

Gnathostomata

Agnatha

Urochordata

+*

Echinodermata

Lophotrochozoa

Ecdysozoa

+**

Cnidaria

Porifera

Protozoa

Plants

rdata
Cephalochordata

GRP

Figure 3 Basic diversity of protein domains that can be used in immunoreceptors. This gure shows the availability of domains across Metazoa,
not their actual usage in immunoreceptors. In green: intracellular element. In beige: mostly involved in extracellular elements. C0 : Complement,
EGF: Epithelial growth factor, Fn3: Fibronectin 3, GRP: Beta 1-3 glucan receptor, Igsf: Immunoglobulin superfamily; can also be found intracellularly,
LRR: Leucine-rich repeat; can also be found intracellularly (e.g., in NLRs), PGRP: Peptidoglycan-recognizing protein, SRCR: Scavenger receptor
cysteine-rich, TIR: Tollinterleukin 1 receptor. ** Absent in C. elegans, * Absent in Ciona.

Evolution of the Immune System

adaptations and the possibility to create large repertoires of

receptors, this restriction will disappear.

Signal Transduction
In order to be of any use, a receptor with antigen-binding
capacity has to lead to an effector phase.
But many metazoan immune effector phases depend on
direct cell activation. The binding of an epitope to a membranebound receptor is followed by phagocytosis, or proliferation, or
secretion of a mediator. Again, because phagocytosis is
ancient, one can imagine that at least some elements of its
signaling pathways could be shared between Metazoa and
Protozoa. Indeed, some common mediators are used to
modulate phagocytosis in Dictyostelium and in macrophages
such as cyclic adenosine monophosphate and adenosine
(ODay and Keszei, 2012).
Other pathways had some of their elements present in
plants, like TIR domains and leucine-rich repeats (Figure 3)
(Ronald and Beutler, 2010).
Signal transduction involves several intermediates, the
acquisition of which through evolution remains a mystery. One
can imagine that if one successful pathway is generated, it will
be preserved and used under different conditions in different
organisms. This is why the elements of those cascades are often
shared not only among all Metazoa but also across kingdoms.
Many intracellular protein kinases, such as protein kinase A,
protein kinase C, focal adhesion kinase, and mitogen-activated
protein kinase; small G proteins such as Ras; and second
messengers like intracellular free calcium, all known to play
roles in vertebrates, regulate many hemocyte immune functions in invertebrates. This is also why these signaling cascades
are shared among different physiological systems: the NF
kappa B (NF-kB)-related cascades are shared by developmental
and immune pathways (Toll) (Gilmore and Wolenski, 2012;
Lemaitre and Hoffmann, 2007).
This conservation might be the Achilles heel of the hosts
immune system when confronted by a parasite. Indeed, from
the parasites viewpoint, the signal transduction could represent an ideal target to eliminate responses from the host:
jamming any intermediary in the pathway would ruin the
cascade. However, those mechanisms, which one considers
intuitively as stable, are made of elements that do precisely
evolve fast and probably escape parasite inuence in this way
(Schmid-Hempel, 2011).
The conservation of intracellular cascades across the animal
kingdom (Figure 4) does not mean automatically a conservation of function. For instance, the Toll pathway used in insects
shares homologous elements at the structural and signaling
levels with the Toll-like receptor (TLR) pathway, yet the principle of recognition is quite different in the two cases. Toll
recognizes a cytokine (spaetzle), whereas TLRs recognize
pathogens directly (Leulier and Lemaitre, 2008). Once more,
the properties of a single structure (the leucine-rich repeat that
forms the extracellular part of the receptor) can be used in
different, nonanalogous directions. Multiple effector phases
can follow gene activation: upregulation of transcription of
genes coding for diverse antipathogen molecules (in the
case of antimicrobial peptide production); overproduction of
the receptor itself (in the case of antibody synthesis); the

induction of migration and, in some organisms, of proliferation; and so on.

Later in Evolution: Multiplication of Receptor

and Effector Mechanisms
Multigene Families
The pressure for diversity must have affected differently the
evolution of receptor families in the function of which ligands
they recognize. To oversimplify, if they recognize constant
PAMPs, PRRs can also remain constant. Conversely, a rapidly
changing or highly polymorphic pathogen epitope may be
better met by a rapidly expanding family of host receptors and
effectors. Diversication within a population can be due to
polymorphism, but this mode of diversication does not
match well the selection pressure requiring individual protection (see Table 1).
Therefore, another way to increase diversity has been the
creation of multigene families by gene duplication. Duplication
can also be achieved via tandem duplication of one gene
involved in the recognition or effector phase. Duplication can
also be the result of a whole-genome duplication, without any
particular pressure exerted over a specic immunity gene, which
creates conditions different from the latter one. Such duplications, of which signatures are seen in the various paralogs of
immunity genes in vertebrates, may have played an essential role
in shaping the gnathostome immune system (Kasahara, 2007).
Examples of multigene families are presented in Figure 5.
Diversity is also seen at the level of effectors, with different
principles being used to damage pathogens: examples include
the production of reactive oxygen, and diverse antimicrobial
peptide families exploited and amplied differently according
to the organisms.

Cell Types Involved in Immunity

As Metazoa bauplans grew in complexity from Placozoa to
vertebrates, immunity has been progressively devoted to more
mechanisms. Division of the work became important. From the
surface and cytoplasm of each cell, from epithelia, immunity
has been delegated to some specialized cells and organs of the
hemopoietic system: hemocytes or coelomocytes in invertebrates, and various leukocytes and the uncommitted lymphocytes in vertebrates. These cells became the patrolling units of
the immune system. Among those, the granulocytes, monocytes, and macrophages are mainly involved in rapid innate
immune responses and tissue repair, and cells with similar
morphological and functional characteristics (e.g., granulocytes
secreting antimicrobial peptides) can be found in most Metazoa. Among the vertebrate lymphocytes, some exert innate
immunity function, like the newly characterized and evolutionarily conserved innate lymphocyte compartment; and
others (the T and B cells or their equivalents) are in charge of the
adaptive responses. By expressing only one receptor specicity
that is able to proliferate upon stimulation by an antigen, and
dividing with a shorter generation time than that of the organisms to which they belong (a few hours versus month or years),
the adaptive lymphocytes offer the possibility to ll in the
generation time gap with the fast-dividing invading pathogens.

Evolution of the Immune System

Drosophila
Fungi

Yeast

Vertebrate

Drosophila

Pathogens

Gram

Gram +

(Lys-type
peptidoglycan)
Proteolysis

PGRP-SD
PGRP-SA

PGN

TNF family

SPE

Fly hemolymph

Persephone

products
(PAMPs)

Spaetzle

Vertebrate extracellular

Pathogen
GRP

GRP

Receptors

Transcription
Factors

TNFR

PGRP-LC

dMyD88

MyD88

RIP/FADD

IMD/dFADD

Pelle, Tube

IRAK/TAKI

MEKK/IKK

PREDD

Cactus

IkB

IkB
Relish
(combined Ikb/NFkb)

Dorsal/DIF

Induced
Anti-microbials

Vertebrate intracellular

Adaptors

TLR

Fly fat-body cell

TIR
Toll

Drosomycin (fungi)
Defensin (Gram +)

NFkB

NFkB

Increase MHC levels

Upregulate co-stimulation
Pro-inflammatory cytokines

Diptericin
Drosocin

} (Gram)

Figure 4 Conservation of signaling cascade in innate immunity. d: Drosophila, FADD: Fas-associated death domain, IKK: Ikb kinase, IRAK: Interleukin 1 receptor associated kinase, IMD: Immunodecient, MEK: Mitogen-activated protein kinase/extracellular signal-related kinase kinase, MEKK:
MEK kinase, Myd88: Myeloid differentiation factor 88, NFKB: Nuclear factor kappa B, PGN: Peptidoglycan, PGRP: Peptidoglycan-recognizing protein,
Sa or Sd: short form of a or d, Lc: long form of c, R: Receptor, RIP: Receptor interacting protein kinase, SPE: Spaetzle processing enzyme, TAK1:
Transforming growth factor beta-activated kinase 1, TLR: Toll-like receptor, TNF: Tumor necrosis factor. Redrawn and modied from Flajnik, M.F.,
Du Pasquier, L., 2008. Evolution of the immune system. In Paul, W.E. (Ed.), Fundamental Immunology, sixth ed. Wolters Kluwer Lippincott Williams
and Wilkins, Philadelphia, pp. 56124.

These cells share transcription factors with each other, across

lineages and across species, indicating at some point an ancient
common origin (Figure 6) (Hartenstein, 2006).

Toward More Products than Genes: Combinatorial

Associations and Somatic Adaptations
Because they are under continuous pressure to diversify their
receptor repertoires, organisms have selected a simple solution
that increases the diversity without unduly increasing the number

of genes devoted to immunity: the combinatorial associations

of elements. This contributes to using the genome at its best.
At the protein level, multimeric receptors can be made by
assembly of different polypeptide chains (e.g., peptidoglycanrecognizing proteins (PGRPs), TLRs, variable light (VL) and
variable heavy (VH), and T-cell receptor (TCR) chains). The
nal number of receptors is equal to the product of the number
of each constituent.
At the nucleic acid level, combinatorial associations of
exons save genetic material, fullling the economy constraint

Evolution of the Immune System

Drosophila
melanogaster

Strongylocentrotus
purpuratus

Oikopleura
dioica

Ciona
Intestinalis

Branchiostoma

Lampetra

Homo sapiens

Protein

Pathogen sensors
TLR

222

48

21

10 +1ps

NLR

203

20

92

140220

20

SRCR

14

218

81

270

287 domains

81

PGRP

15

>20

ND

RIG-I-like helicases

12

ND

ND

C-type lectins

32

104

31

120

1215

ND

81

IgSF-ITIM

>3

ND

>6

>5

>3

>50

MyD88-like
(DEATH-TIR)

ND

SARM1-like
TIRAP-like
TICAM2-like

15

>2

12

ND

65

128

>7

ND

11

Key adaptors

Potential effector
PLA2

Figure 5 Multigene families across invertebrates immune systems compared to human. ITIM: Immunoreceptor tyrosine-based inhibition motif,
NLR: NOD-like receptor, PLA2: Phospholipase 2, RIG-1: Retinoid-inducible gene 1, SARM-1: Sterile a motif (SAM) and an Armadillo repeat motif TIR
motif-containing one protein, TICAM2: TIR domain-containing adapter molecule 2, TIRAP: TIR adaptor protein, TLR: Toll-like receptor.

(Table 1). They permit the economical building of a large

repertoire from a small number of genetic elements. Occurring
somatically and at random, they create during ones life an
individual repertoire that is the basis for an adaptive response
(Du Pasquier, 2006).
This evolutionary strategy has been selected several times in
evolution (in agnathans, gnathostomes, arthropods, mollusks,
and perhaps echinoderms) with variations on the theme:
mutually exclusive alternative splicing, DNA rearrangements
with or without random somatic mutation, and somatic gene
conversion (Du Pasquier, 2006). Let us stress the considerable
differences between events due to splicing that are not inheritable and those due to DNA rearrangement or mutation that
are transmittable to cellular progenies.
Exploiting the genome at its best also implies usage of
epigenetic regulation. In the immune system, it plays a role that
is currently more and more analyzed, but this is a topic for
which we have not yet phylogenetically relevant observations.

RNA Alternative Splicing

Alternative splicing is used widely across Metazoa and modies
the signaling intracytoplasmic moiety of a receptor. But it can
also be used to modify its extracellular recognition capacity.
The best example of such a case is Down syndrome cell adhesion molecule in crustaceans and insects. There, mutually

exclusive alternative RNA splicing, involving large numbers of

Igsf-duplicated exons, generates thousands of isotypes of
the molecule, the role of which in immunity seems important
but is not yet fully characterized. A role in phagocytosis is
conrmed (in addition to a role in the genesis of the nervous
system).
In tunicates, alternative splicing also leads to a large
repertoire of Fester products, the putative receptors of a
highly polymorphic tunicate FuHc locus presumably
involved in allorecognition. Other not understood somatic
events, probably at the RNA and posttranslational levels (?),
diversify the family 185/333 of glycine- and histidine-rich
effectors, which are strongly upregulated after lipopolysaccharide treatment in the sea urchin (reviewed in Flajnik and
Du Pasquier, in press).

A Typical Innate Immune System

With the mechanisms described thus far in this article, the
responses remain essentially innate, with limited and not
always well-understood somatic adaptations. An example of
a sophisticated response using such mechanisms is that of
Drosophila, which is presented in Figure 7 (Lemaitre and
Hoffmann, 2007).
A nagging question is whether immunity based only on
innate mechanisms can generate specic memory. Some

Evolution of the Immune System

Hematopoietic lineages of mesodermic origin

2 m

2 m

Mammalian lymphocyte

Insect plasmatocyte

m
gm

e
hsc

PU.1
g

mye

PU.1

GATA 1
FOG
ph

AML 1

lym

GATA / Srp
Gcm

lm

EBF / Col

C / EBP

GATA 2
b

pl

EBF

GATA 3

GATA / Srp
FOG / Ush

cr

GATA / Srp
AML 1/Lz

Figure 6 Sharing of transcription factor families in hematopoietic tissues from invertebrates and vertebrates: the different hemocyte and leukocyte
lineages, with their transcription factors. cr: Crystal cell, e: Erythrocyte, g: Granulocyte, gm: Granulocyte and macrophage precursor, hsc: Hematopoietic stem cells, lm: Lamellocyte, lym: Lymphoid precursor, m: Macrophage, mye: Myeloid precursor, ph: Hematopoietic precursor, pl: Plasmatocyte
precursor, t: T lymphocyte. Redrawn and modied from Hartenstein, V., 2006. Blood cells and blood cell development in the animal kingdom.
Annu. Rev. Cell Dev. Biol. 22, 677712.

immune responses of long-lived insects like the bumblebee

phenotypically have the features of a memory response
with an apparent increased specicity (Sadd and SchmidHempel, 2006). Yet whether it is a second induction,
a long-lasting response, or a reinitiation is unclear. Normally, the immune responses in invertebrates studied so far
have not shown any specic proliferation except the renewal
of used populations of hemolytic by feedback. No clonal
selection has been demonstrated in any invertebrate so far,
and specic memory investigations in the case in allograft
rejection have always led to ambiguous results. One key
question, curiously rarely addressed directly, is: Does any
invertebrate response involve cell proliferation in response
to a challenge (in opposition to a simple upregulation of
gene expression)?

Modications at the DNA Level

This step is essential for the development of an adaptive
immune system because the changes introduced become
inheritable and allow a Darwinian variation or selection inside
one individual during its ontogeny. The two known modalities
of such DNA modication used in vertebrates are represented
in Figure 8.
One enzyme, a cytidine-deaminase (activation-induced
deaminase (AID) in gnathostomes), achieves several somatic
modications of the genes of the vertebrate lymphocyte
receptors. It causes the deamination of a cytidine and its
transformation into a nonacceptable uridine in the DNA
segment affected. This is followed by repairs, breaks, all
sources of mutation, heavy-chain class switch gene conversion, or copy choice variations. When applied to cassettes of

Evolution of the Immune System

Fat-body cell

Gut
Cytokine
Pathogens

Domeless

JAK
STAT

PGN

Imd

PGRP-Lx
Toll

ROS

AMPs

Spaetzle

Toll

Imd

Duox
Effectors
Left-over pathogens

Large
parasites
Gram

Phagocytosis

Gram +

Fungi

Coagulation
Melanization

Encapsulation

PPO
Eater
Draper
DSCAM Nimrod

SR-CI PGRP

Plasmatocyte

Crystal cell

Lamellocyte

Lymph gland

Figure 7 The innate immune system of an insect, Drosophila. Hematopoiesis takes place only in larvae in the lymph gland. AMP: Antimicrobial
peptides, DUOX: Dual oxidase, JAK/STAT: Janus kinase and signal transducer and activator of transcription, PPO: Prophenoloxidase, ROS: Reactive
oxygen species. For other abbreviations, see Figure 4. Redrawn and modied from Lemaitre, B., Hoffmann, J., 2007. The host defense of Drosophila
melanogaster. Annu. Rev. Immunol. 25, 697743.

leucine-rich repeat subunits of the variable lymphocyte

receptor of agnathans, it can create a potential repertoire of
immunoreceptors just as vast as that of antibody or TCR in
the Igsf of gnathostomes. The same enzyme is responsible for

the hypermutation and the heavy-chain class switch in gnathostomes. When applied to the avian immunoglobulin (Ig)
locus, it generates the B-cell repertoire by gene conversion
(Schatz, 2007).

10

Evolution of the Immune System

Assembly and diversification of genes encoding antigen receptors

Jawed vertebrates
V

Jawless vertebrates
J

LRR modules

LRRNT

LRRCT

1 2 3 4 5

LRR modules

6 7 8 9 n

RAG1-RAG2 V(D)J recombination

V

D J

Somatic hypermutation

Gene conversion

AID

AID

D J

U U

D J

Class-switch recombination

D J

D J

VLR gene assembly

AID

Copy from yV

Error-prone repair

D J

AID
S

(U

LRR modules

D J

LRR modules

1 2 3 4 5

Point mutation

Gene conversion tract

NT

LRRCT

CT

U U

LRR modules

6 7 8 9 n

Homology search
Extension by DNA polymerase

DNA deletion

LRRNT

1 2 3 4 5

Replaced exon

LRRNT
NT

LRRCT

2 5 7 9

CT

LRR modules

6 7 8 9 n

Copied LRR modules

Figure 8 Somatic generation of diversied repertoires of receptors in jawless (agnathan) and jawed vertebrates (gnathostomes). AID: Activationinduced deaminase, D: Diversity, J: Junction, LRR: Leucine-rich repeat, RAG: Recombination-activating gene, V: Variable, VLR: Lamprey variable
lymphocyte receptor. Redrawn and modied from Schatz, D.G., 2007. DNA deaminases converge on adaptive immunity. Nat. Immunol. 8, 551553.

Apparently, mollusks too can achieve somatic mutation of

their brinogen-related protein genes (made of two Igs plus
one brinogen domain and involved in anti-Schistosoma
immunity) (Loker et al., 2004). It will be interesting to investigate whether this is due to a homolog of AID. In gnathostomes, an extra enzyme set, RAG 1 and 2, plays a role in
generating somatically a repertoire of receptors. These enzymes
contribute to the somatic rearrangement, via combinatorial
usage, of scattered Variable, Diversity, and Joining (V, D, J)
segments that build up a variable domain of a B-cell receptor
(BCR) or a TCR. This enzyme pair is conserved, being detectable in sea urchin (where its function is not elucidated) and
apparently inherited from transposons of invertebrates (Transib and the molluskan N-RAG-TP would be candidates) (Panchin and Moroz, 2008). RAG 1 and 2 are not present in
lamprey, nor in amphioxus or Ciona. The presence of RAG 1
and 2 in an invertebrate was a surprise and might actually look
suspicious. However, the gene structure with the RAG2
homolog adjacent to the urchin RAG1 is similar to what is
found in gnathostomes. However, the presence of RAG1 core
sequences in cephalochordates, mollusks, and diploblastic
invertebrates suggests that the sea urchin observation is not an
artifact. It is not impossible that the same transposon has
invaded some genomes and not others, therefore creating
a heterogeneity in evolutionary histories. It must be admitted
that so far it is unclear how this new result ts into the puzzle of
the origins of adaptive immunity.
Another enzyme, the terminal deoxynucleotidyl transferase,
generates further somatic receptor diversity in all gnathostomes. This DNA polymerase diversies the third complementary determining region during Ig and TCR gene
rearrangement by adding nucleotides in a nontemplated
fashion, hence the name of N diversity that ensues.

Necessity of Selection: Emergence

of the Gnathostome Adaptive Immune System
Some phylogenetically distant organisms, under selection
pressure favoring diversication, have incorporated somatic
adaptations in their immune system (Flajnik and Kasahara,
2010). These randomly create a giant repertoire of immunoreceptors, some of which might interact with the self. This
means that such specicities could be used to monitor one
selfs integrity (the immune surveillance hypothesis), but it
means also that there will be a risk of autoimmunity. A selection process is necessary. Therefore, for them to be useful,
somatic adaptations had to occur in a context providing
a convenient unit of selection. In vertebrates, this unit turned
out to be the uncommitted lymphocyte (i.e., a cell expressing
only one antigen-specic receptor and exposed to selection by
self antigens in a special topographical location, the lymphoid
organs stromas and epithelia). In this way, if the cell expresses
an autoreactive dangerous receptor, only that cell has to be
eliminated; it will not inuence the fate of other, useful
lymphocytes. This requirement is strong, and it was invented at
least twice in evolution: in the agnathans and in the gnathostomes, independently (Hirano et al., 2011).

The Origin of Lymphocytes

Some invertebrate hemocyte transcription factors are homologous to those of vertebrate leukocytes in general, and
lymphocytes in particular (Figure 6). We can oversimplify by
considering the myeloid lineage homologous to the various
hemocyte lineages of invertebrates. But what about the T- and
B-lymphocyte lineages? The discovery of T- and B-cell markers
in agnathans suggests that these two lineages were ancient and

Evolution of the Immune System

predated the split between agnathans and gnathostomes. The

genesis of these two sets of lymphocytes could have therefore
preceded the acquisition of the specic antigen receptors
associated with them: the TCR and BCR of gnathostomes and
the variable lymphocyte receptor A and B (VLRA and VLRB) of
agnathans (Guo et al., 2009).

The Selection of Repertoires in Agnathans and Gnathostomes

For the agnathans, the presence of a homolog of the thymus,
the thymoid of the lamprey, rich in sterile transcripts of the
VLRA genes could suggest that selection happens there (Bajoghli
et al., 2011). No one knows whether it is due to a presented
peptide or to the recognition of native epitopes. So far, neither
class I or II nor immunoproteasome or TAP (transporter associated with antigen processing) have been isolated from either
hagsh or lampreys, but an MHC analog is not excluded.
In gnathostomes, the specialized cells are selected following
interaction with MHC class I and II molecules plus peptide
complexes in the thymus (T cells) or native antigen epitopes in
lymphoid organ stroma (B cells). Some T cells, the gamma delta,
are selected in a similar way. The presentation of peptides rather
than native epitopes allows the monitoring of one selfs constituents in a way almost as precise as checking ones DNA sequence.
It might have been another selective advantage of the process.

History of the Major Histocompatibility Complex

The term major histocompatibility complex (MHC) has led to
some confusion among comparative immunologists. Recognition of allogenic determinants leading to histoincompatibility is a property of most Metazoa; often, the rejection
reactions are under the control of a major region (e.g., Rosa
et al., 2010), which does not mean that it is a homolog of the
gnathostome MHC that we are going to envisage now. The
MHC backbone must be ancient because four MHC paralogs
have been detected in vertebrates (Flajnik and Kasahara, 2001).
Because of the two genome duplications that presumably
occurred in the ancestors of vertebrates, this means that this
region stemmed from a single ancient linkage group. Indeed,
such a region has been identied in amphioxus and in
Drosophila, but it does not contain class I or II (Danchin and
Pontarotti, 2004). It looks as if this region acquired the class I
and II genes late. And, indeed, Igsf domains used in those
molecules are of the rare C1 type, which is encountered only in
the receptors generated by somatic rearrangement in gnathostomes, with perhaps some homologous domains in the B7
family of co-receptors, of which distant homologs exist in
invertebrates. Within the MHC, class I genes, found on two or
three of the paralogs, may have predated class II genes (found
on only one chromosome). The MHC contains class I processing genes, like the immune proteasomes lmp2 and lmp7
and the TAP genes. These genes are functionally part of the
class I presentation pathway, and it makes sense to imagine
a co-segregation of genes belonging to the same pathway.
In that sense, the breaking of linkage seen in mammals is
a unique accident. Even in teleost sh, where the MHC has
been split into many fragments, the class I pathway gene
linkage is maintained. The other genes forming the class III
region, like complement, HSP, and so on, have homologs that
are detectable in invertebrates. It has been proposed that,

11

immunologically speaking, the original function of MHC

before presentation was to process invaders intracellularly. In
this context, the presence in the MHC or in the MHC paralogs
of many antiviral tripartite motif-containing molecule (TRIM)
genes, whose products target viruses toward the proteasome,
could suggest what was the ancient MHC linkage group before
the existence of classes I and II: an intracellular system for
antiviral defense. The adaptation of the proteasome to peptide
presentation would come later.
Another way of considering MHC evolution could be through
the MHC-linked marker heat shock protein (HSP) 70. HSPs
derived from human tumors have been shown to mediate the
re-presentation of their chaperoned peptide complexes and lead
to the in vitro activation of tumor-specic T cells. This ability
extends to other classes and perhaps phyla. It may have been at
some point an alternative presenting pathway placing this
genetic region under selection (Robert, 2003). It would be
interesting to examine this issue in agnathans. In addition, if we
have no clues as to where the peptide-binding region of classes I
and II comes from, the old proposal that it might be related to
HSP remains intellectually one of the most attractive hypotheses.

The Thymus
Thanks to specic marker expression studies, the region
anatomically homologous to the thymus of gnathostomes, that
of pharyngeal pouches, has been shown to be involved in the
production of VLRA T-like cells like in agnathans, leading to
the hypothesis that perhaps selection of the repertoires was
occurring there. The presence of many sterile transcripts in the
lymphocytes present in this region reinforces the hypothesis
(Boehm et al., 2012).
The gnathostome thymus where central selection takes
place is the rst organ to become lymphoid during ontogeny. It
originates from pharyngeal pouches, but not always from the
same, unlike the thyroid. Not only the position but also the
number of pouches involved can vary (from ve in chondrichthyans to one in mammals). Several species retain more
than one thymus; for instance, the urodele axolotl has three.
Finally, selection based on the afnity of interactions
between the TCR and peptide MHC class I and II ligands
eliminates the dangerous (central deletion), keeps the useful
(positive selection), and ignores the mediocre, and it shapes
the T-cell repertoire of CD8 and CD4 T cells. This is achieved
centrally in the medulla, where the repertoire is probed against
the self. The thymus of gnathostomes has conserved its cortex
medulla architecture in species ranging from cartilaginous sh
to mammals. If central deletion is not complete, mechanisms
of peripheral tolerance (with suppression of lymphocyte
activity by the regulatory T cells) are also at work. An analogous
process, not mediated by MHC, takes place for B cells in
lymphoid organs. The other lymphoid organs of vertebrates are
represented in Figure 14.

Evolution of the Immune System within

Gnathostomes
Let us rst compare the setup of the TCR, BCR, and MHC genes;
the lymphoid organs; and the innate components, and then the
inuences of the variations noticed on the immune responses.

12

Evolution of the Immune System

also shows an interesting example of convergent evolution: the

independent acquisition of monovalent antibodies in the shark
and the camel, creating a parallel repertoire probably with very
different properties from the VH to VL pairs one. Much of the
VH germline repertoire has been conserved over extremely long
periods of vertebrate evolution. All vertebrate groups have one
to four L-chain isotypes, all of which can be categorized as k, l,
or s (and, in cartilaginous sh, s-car). The advantage of possessing multiple isotypes is poorly understood since experiments on L-chain-less antibodies show very little contribution
of the L chain to the antibody function. The organization of Ig
genes throughout gnathostomes is depicted in Figure 11
(Flajnik and Du Pasquier, 2008; Litman et al., 2010).
The generation of diversity of the BCR can be obtained by
multiple means, as discussed throughout the remainder of this
section.

T-Cell Receptor
Genes encoding the two types of TCR, a/b and g/d, have existed
since the earliest jawed vertebrates, in which their assembly and
diversication are due to somatic rearrangement of gene
segments mediated by RAG1 and RAG2 enzymes (Litman et al.,
2010). We do not have a complete picture of TCR genes in all
vertebrates, but an organization similar to the mouse and human
a/d locus, with an interspersion of alpha and delta elements in
the same chromosome, can be seen in teleostei (Figure 9).
So far, all alpha loci in vertebrates have many J segments,
and all beta loci have a large number of Vb. There can be
variation within classes: in humans and mice, g/d T cells
constitute a limited portion of the T-cell population, whereas
larger repertoires are seen in chicken, sheep, cattle, and rabbits,
the so-called GALT species (see explanation in the Section Gene
Conversion). The g/d TCR can be used either adaptively or for
innate recognition when oriented rearrangements due to
microhomologies of some gamma delta genes end up creating
the equivalent of innate receptors, as in the skin of mice or the
blood of humans.
Special isotypes are found in marsupials, and hybrid
molecules made up of new antigen receptor (NAR) and TCR in
shark remind us of the many possibilities offered by the Igsf
domain and of the independent evolutionary pathways chosen
(Flajnik and Du Pasquier, in press).

Somatic Rearrangement
Combinatorial association of DNA segments (rearrangement)
is mediated by RAG1 and RAG2 enzymes, with variations on
the theme in species ranging from cartilaginous sh to
mammals. Structured into many clusters, each chondrichthyan
IgH and L-chain gene consists of different V, D (no D in light
chains), J, and C elements where V, D, and J genes rearrange
only within one cluster. In the absence of rearrangement
between clusters, diversity of the primary repertoire is lower
than in other vertebrates. Bony sh, frogs, reptiles, and
mammals have the so-called translocon conguration where
the possibility of combinatorial rearrangement enables more
diversication than in cartilaginous sh (Hsu, 2009).

B-Cell Receptor
A summary of the gnathostome Ig heavy-chain isotypes
diversity and domain organization is given in Figure 10, and it
13V

12 J

Cd

27 J 1

C 1

Jd

TRA/D Danio

51 V

J 2

C 2

V 2

TRB

C 1

C 2

C 5

C 4

Cg3

C 6

TRG
Locus 2

Locus 1

100 V

61 J

TRA/D Homo

6467 V

J 1

C 1

J 2

C 2

TRB

1215 V

J 1

C 1

J 2

C 2

TRG

Figure 9 T-cell receptor gene organization in teleost sh and mammals. C: Constant, D: Diversity, J: Junction, TCR A,D, B,G: T-cell receptor alpha
delta and betagamma, V: Variable. Data assembled from Lefranc, M.P., Lefranc, G., 2001. The T Cell Receptor. Facts Book. Academic Press,
London; and Castro, R., Bernard, D., Lefranc, M.P., Six, A., Benmansour, A., Boudinot, P., 2011. T cell diversity and TcR repertoires in teleost sh.
Fish Shellsh Immunol. 31, 644654.

Evolution of the Immune System

13

Mammals
IgM

IgD

IgG

IgE

IgA
CAMELID

IgG
Birds
IgM

IgY

IgA

Reptiles
IgM

IgY

Amphibians

Teleost Fish

Cartilaginous
fish

IgM

IgD-like

IgM

IgD-like

IgM

Analogy

IgY

IgX

IgF

Ig Z,T

IgMg j

IgW

No analogy
IgNAR

Conservation
homology

Convergence

Figure 10 Immunoglobulin isotypes in gnathostome vertebrates. From Flajnik, M.F., Du Pasquier, L., 2008. Evolution of the immune system. In
Paul, W.E. (Ed.), Fundamental Immunology, sixth ed. Wolters Kluwer Lippincott Williams and Wilkins, Philadelphia, pp. 56124 with modifications.

Gene Conversion
In birds, the organization is similar, but all V genes except those
most 30 to the D elements are pseudogenes. Only one rearrangement is possible in light chains, and a very few are
possible in heavy chains, leaving the diversication to
a somatic, AID-mediated, gene conversion process that incorporates sequences from the pseudogenes into the functionally
rearranged gene during ontogeny. A similar if not identical
situation exists in some mammalian species that rely on gutassociated lymphoid tissue to generate Ig diversity (and
are therefore referred to as GALT species, a division that we
have seen in this article for the TCR gamma delta). These are
exceptions with a reduced germline-expressed repertoire, but
gene conversion and somatic hypermutation compensate for
this situation.

Somatic Hypermutation
The enzyme AID participates in both somatic hypermutation
and class switch.
Hypermutation is encountered throughout gnathostomes
and perhaps also in agnathans, with some variation on
frequency and modalities from shark to human.

Class Switch
Bona de heavy-chain class switch (i.e., with excision of DNA
containing the eliminated isotope constant region gene) occurs
from amphibians thanks to a relatively well-conserved switch
box repetitive region. However, a form of switch in chondrichthyans allows the V of one cluster to be associated with the
C of another, following still-unknown mechanisms but representing at least an analog of switch.
Isotypic and allelic exclusion are encountered throughout
all gnathostomes. Curiously, cartilaginous sh and birds show

co-evolution of Ig gene architecture for H and L loci, whereas

teleosts have shown that this is not a rule.

Major Histocompatibility Complex: Antigen-Presenting Cells

The variations in MHC architectures (Figure 12) (Flajnik and
Du Pasquier, in press) turn out to be quite numerous across
vertebrates. They can inuence the presence or absence of genes
(class II, CD4 invariant chain), the number of class I genes
(from one classical in Xenopus and chicken to more than 10 in
axolotl and cod the number of complement-associated genes),
or their polymorphism. It corresponds to a preserved linkage
group except in teleost sh.
High levels of polymorphism and genetic diversity are
characteristic of most classical MHC genes. The retention of
ancient allelic lineages through selection accounts for the high
levels of diversity observed within species.
Although classes I and II are expressed either ubiquitously
or on a variety of tissues, presentation leading to immune
responses occurs predominantly on a specialized set of cells,
the antigen-presenting cell (dendritic cell); but, curiously,
numerous variations of the theme can be seen across vertebrates. Dendritic cells, for instance, exist in birds that also have
at least two other sets of phagocytic cells, heterophils and
thrombocytes, that may also present antigen. Similarly, B cells
can phagocytose and present in species from sh to mammals.
Analogs of those cells can be found in all vertebrates since
they resemble macrophages, histiocytes, and Langerhans cells.
Until recently, the costimulatory molecules expressed on the
surface of those cells have been the object of less attention
than the class I and II ones, but many have now been traced
in nonmammalian vertebrates; one example is the B7 family
(with their T-cell counterpart cytotoxic T lymphocyte A4

14

Evolution of the Immune System

(Horn shark,
nurse shark, skate)
(Catfish, trout,
zebrafish)

Bony fish

Cartilaginous
fish

Differences in Ig gene organization

H
IgNAR
IgW

(
(
(

C TMs

V D1D2 J

CW

Amphibians

(Xenopus)

Birds

(Chicken)
(Mouse, humans)

V I 1-n

V2

((

V>10 D

D
Vn

V II 1-n

n200

V DDJ C

Joined

( n10

( (
(

Vn

V2 J C

J
C

Vn 6

n5

VJC

n>(5-100) or

D J C

Vn J C

( (

Joined (2-100)

Ds 1-n

J4

V V V

Sw

C TMs
V

C 2
J

n60

V I 1-32

Mammals

Vn

V D1D2D3 J C NAR

C C 2

L(,)

V2 D

V II

Ds 15

V IIIXI

J 8-9

C Sw

C Sw C

H
Vn

L(,)
V VV

L()

( (

J
JJ V V

C
JJ C

60 kb

C J C

Ds

Sw

C TMs

Sw C

H
25 pseudogenes V

L ()
H
L ()
L ()

VI

((

VI
V2

J2 C2

V II

((

Vn

Ds

Vn

JC

V1 J3 C3

J 1-4

Sw C TMs C

C 3

C 1

C 1b

C 2a

C
J4

C1

Figure 11 Immunoglobulin gene organization in vertebrates. H: Heavy chain, L: Light chain, TM: Transmembrane. For other abbreviations,
see Figure 9. From Flajnik, M.F., Du Pasquier, L., 2008. Evolution of the immune system. In Paul, W.E. (Ed.), Fundamental Immunology, sixth ed.
Wolters Kluwer Lippincott Williams and Wilkins, Philadelphia, pp. 56124, with modications.

(CTLA4) and CD28) in amphibians and teleost sh, where its

complexity is greater than in mammals (Flajnik et al., 2012).
Other members homologs can even be detected in invertebrates in conserved linkage groups (Zucchetti et al., 2009).
Ancestor-like molecules with similar domain composition can
be identied in some invertebrates. So it is most likely that the
presenting cells of the dendritic cell type have been recruited
in the presenting pathway without losing some ancient
activity inherited from invertebrates (e.g., the use of the
conserved TLRs).

Links between Evolutionary Variations in Lymphoid Organs,

Architectures, and Immune Response
Lymphoid organs are perhaps the only elements of the adaptive
immune system that show some kind of progressive complexication across vertebrates (Figure 14). This complexication (acquisition of lymph nodes and germinal center sites
where somatic mutants are generated and bred) parallels
a progressively better exploitation of somatic adaptation,

resulting in a better factor of immunization, a better afnity

maturation of the antibody responses, and a better memory
(Figure 13).
The absence of germinal centers in cold-blooded vertebrates
could explain why afnity maturation remains moderate in
shark, teleost sh, and frogs, whereas it can generate 10 000
more afne antibodies in mammals at the peak of a secondary
response. The availability of mutants is sufcient but not the
selection, which requires the specic topography and chemistry
provided by the germinal centers.
These variations between species in exploiting somatic
adaptations may be the source of the variations in the interactions between innate and adaptive immune systems among
vertebrates. When adaptive immunity is not used at its best, like
in cold-blooded vertebrates perhaps, one sees innate strategies
coming back, for instance the use of germline-rearranged Ig in
shark. On the other hand, multiplication of innate pathways
can be seen in poor-responder sh that duplicated their C3
genes, or in the frogs where the number of AMP in the skin is
remarkably high.

15

Evolution of the Immune System

MHC comparative organization

Hypothetical jawed
vertebrate UR MHC

Class III

CL II
TAP 1
TAP 2
LMP 2*
LMP 7*
LMP 7/X*
RING3*
CL Ia
C4*
Bf/C2 (2 genes)*
LMP 7*
CL Ib

TAP 1
CL II
CL II

CTX (IgSFV)

CL IIa

FABGL*
TAPBP*
ZNF297*
DAXX*
KNSL2*
RxRB*
COLL11A2*

BG (IgV)
NKR
LECTIN
CL II
TAPBP**
CL II
RING3*
DM
DM
DMB
CL Ia
TAP 1
TAP 2
CL Ia
C4*

CL II (4 loci)
RING3*
DM
DM
CL Ia
LMP 2*
LMP 7*
TAP 1
TAP 2
NOTCH*
PBX2*
RAGE (IgV)
LPAAT
C4*
MECL1*
TNXB*
PPT2
C2*
Bf*
NEU
AIF1
BAT2*

Human

CL II(DO)
RING3*
CL II(DM)

CD1 (2 loci)
BG (>10 loci IgV)
CL Ib
CL/

BTL II (IgVC1)
NOTCH4*
PBX2*
RAGE (IgVo)
LPAAT
PPT2*
TNXB*

Bf*

C4 (2 loci C)*

IKBL
C5NK2B*

Class I

XMIV (IgSFVJ) (6 loci)

Bf*

Ly6 (2 loci)
C4*

LMP 2*
TAP 1
LMP 7*
TAP 2
CL II (DO)
CL II (DO)
CL II (DR)

TNFSF (3 loci)*

CL II

KNSL2*
DAXX*
ZNF297*
TAPBP*
RPS18
SACM2L
FABGL
HKE4
RXRB*
COL11A2*
CL II(DP)

Class II

PBX2*
SKII2W
CSK2B
BAT2*
DSP
PPT2*
HKE4
SACM2L
RPS18

Chicken

R-fpy*

CL Ib

KNSL2*
ZNF297*
DAXX
TAPBP*
CL Ia
LMP*
RXRB-L*
CL Ia
COLL11A2*
RING3*
FABGL*
TAP 2
LMP 2*
LMP 2/7*
MECL1*
LMP 7*
CL Ia

Xenopus

Adaptive MHC

Class I/II

CL II(DM)
CL Ia
TAP 1
TAP 2
LMP 2*
LMP 7*
MECL1*
RING3*
NOTCH4*
PBX2*
PPT2*
TNX*
C4*
Bf*
HSPA1*
BAT2*
TNFSF*

Teleost

Extended class II

ZNF297*
COLL11A2*
RXRB*
TAPBP (IgVC1)*
FABGL*
CL II /

Nurse shark

SKII2W
Bf(C)*
C2 (C)*
NEU (sialidase)
HSPAI (3 loci)
G6 (3 loci)

HSPA1 (2 loci)

Class Ib (XNC) (Many loci)

KBL

Inflamatory region

CSK28
BAT2*
AIF1
NKP30 (IgV)
LST1
TNFSF (3 loci)*
CL Ib (MIC)
Class II

CL Ia (A,B,C)
CL Ib (ABC-like)
MOG (IgV)
BTN (8 loci, IgVC1)

Figure 12 Major histocompatibility complex organization in gnathostome vertebrates. Names in red refer to genes involved in immunity. Vertebrates compared to a hypothetical ancestral MHC (left). Gene nomenclature is that used in GenBank and is available at http://www.ncbi.nlm.nih.gov/
genbank/.

Variations in the Innate Arm of the Immune System

Pattern Recognition Receptors
The addition of an adaptive compartment with obvious
advantages for an immune response has modied the role of
innate components and the fundamental role of innate
immunity in vertebrates. This new situation has resulted in the
specic evolution of certain PRR families and explains the
evolutionary divergence between vertebrate PRRs and their
invertebrate relatives. Divergences and variations can also be
seen within vertebrates. The innate immunity pathways have
acquired functions other than the direct production of antimicrobial products and are now often involved in generating

an inammatory response with effectors such as cytokines

(type I interferons, inammatory cytokines, and chemokines)
that pave the way to an adaptive response.
TLRs are instrumental in both launching innate immune
responses and inuencing adaptive immunity; they cooperate
with NOD-like receptors (NLRs) and retinoic acidinducible
gene protein 1 (RIG-1)-like receptors (RLRs) in innate immunity. Invertebrates appeared to have no orthologs of mammalian-type TLRs, whereas those are conserved in humans, sh,
and lampreys. Yet while some TLR members were commonly
retained in both mammals and birds, others were separately
lost or gained, or they diverged independently. Sea urchins

16

Evolution of the Immune System

Immune responses in jawed vertebrates

100

10

Primary + secondary
Nurse shark

Factor of immunization

100

10

Primary + secondary
Cod

Secondary

100

10

Primary
Xenopus

100
Secondary
10
Primary
Mouse
2

10

14

18

22

26

Weeks

Figure 13 Antibody responses in gnathostome vertebrates. Factor of immunization plotted in function of immunization schedules. Notice the differences in kinetics and intensities. The poor response of the cod could be due to the absence of MHC class II in this species. From Star, B.,
Nederbragt, A.J., Jentoft, S., Grimholt, U., Malmstrm, M., Gregers, T.F., Rounge, T.B., Paulsen, J., Solbakken, M.H., Sharma, A., Wetten, O.F.,
Lanzn, A., Winer, R., Knight, J., Vogel, J.H., Aken, B., Andersen, O., Lagesen, K., Tooming-Klunderud, A., Edvardsen, R.B., Tina, K.G., Espelund, M.,
Nepal, C., Previti, C., Karlsen, B.O., Moum, T., Skage, M., Berg, P.R., Gjen, T., Kuhl, H., Thorsen, J., Malde, K., Reinhardt, R., Du, L., Johansen,
S.D., Searle, S., Lien, S., Nilsen, F., Jonassen, I., Omholt, S.W., Stenseth, N.C., Jakobsen, K.S., 2011. The genome sequence of Atlantic cod reveals
a unique immune system. Nature 477, 207210.

with their large number of TLRs offer the best example of

independent diversication (Satake and Sekiguchi, 2012).
Many of the NLR genes discovered in mammals are
preserved in nonmammalian vertebrates, including sh,
although teleostei possess an additional unique group of NLRs.
Orthologs of RIG-I and melanoma differentiation-associated
gene 5 (MDA5), recognizing cytosolic, virus-derived RNA and
subsequently inducing the production of interferons, exist in
birds, amphibians, and some but not all of the teleost species
tested so far (Hansen et al., 2011).
Homologous molecules may be involved in rather different
pathways. For example, the vertebrate PGRP is structurally
related to the protostomian one, without suggesting direct
orthology. Mammals have four PGRPs, which are structurally
homologous to the insect PGRPs but are secreted. All are
involved as effectors of innate immunity via their proinammatory and/or bactericidal activities that are not shared
with insects (Dziarski and Gupta, 2005).
The TRIM receptors involved in ubiquitination proteins
(RING-Box-Coiled Coil motif followed by different C-terminal
domains (RBCC)) participate in many cellular processes,

including antiviral immunity. They are also involved in the

regulation of innate immune responses through the modulation of PRR signaling pathways. This family has been greatly
and specically diversied in vertebrates. In teleost sh, many
members belong to a specic category, the nTRIMs, the
expression of which is greatly amplied following virus infection. Most human trim genes of the other classes have only one
or two orthologs in sh. Finally, some trim genes encoding
RBCC-B30.2 proteins are preferentially located in the vicinity of
MHC or MHC paralogous genes in sh and mammals, which
suggests that they may have been part of the ancestral MHC
(Boudinot et al., 2011).
Globally, if one considers the families of innate receptors or
effectors across vertebrates NLRs, RLRs, TLRs, PGRPs, Aim2like receptors, TLRs, and TRIMs one notices independent
evolution patterns most of the time.

Natural Killer Cells

It has been proposed that NK-like recognition of an ancestral
class I or class I-like molecule may have been at the origin of the
adaptive immune system. NK-cell analogs are in fact numerous

Evolution of the Immune System

in invertebrates. In gnathostomes, the receptors of NK cells

recognize self MHC, but they can do it with lectin domains in
rodents or with Igsf domains in primates. NK-cell receptor
evolution is a nice example of an evolutionary story linked to
the pathogenic environment (Vilches and Parham, 2002).
They are the most rapidly evolving molecular component of
the gnathostome immune system. Most ligands for these
diverse NK receptors are MHC class I molecules, or molecules
of host or pathogenic origin related to MHC class I.
In addition, viruses are hypothesized to supply the evolutionary pressure on diversication of NK-cell receptors. This
receptor can be expressed on a category of cells with effector
function (such as perforin production, also encountered in CTLs)
that can be shared with invertebrate relatives like the cytotoxic
cells of sipunculid worms. Perforin domains are related to the
cholesterol-dependent cytolysins existing in bacteria!

Variations in Regulatory Networks: Cytokines

Because cytokines evolve fast, it is difcult to trace their history.
Most of the time, only analogy in function has been reported.
The majority of cytokine and chemokine genes are found in
all vertebrates. In fact, bony sh have more chemokines and

Thymoid
Galt
NK?

AID-L (2CDA)
Gen conv
LRR
VLR A , B, C
Allograft memory?

Il17,8
TNF
CxCR8

receptors than any other vertebrate! In mammals, the chemokine receptor genes are found on the four chromosomes containing the homeobox genes, which provided evidence for the
two rounds of duplication hypothesis. This could mean once
more that precursors of these genes should be found in invertebrates. Thus far, however, only two mammalian-type chemokines or receptors have been detected in agnathans. None of
the so-called adaptive cytokines seem to be present in the lower
deuterostomes.
Cytokines like TNF and interleukin 17 seem to be more
primordial than others, with homologs in some invertebrates
such as Ciona, amphioxus, and sea urchins (Flajnik and Du
Pasquier, in press).

Conclusion
As a conclusion, Figure 14 summarizes crucial aspects of the
immune system and immune responses of vertebrates. The
elements used to draw this scheme and most of the content of
this chapter can be found in recent textbooks such as the various
editions of Fundamental Immunology (Paul, 2008), which contain
specic chapters related to the evolution of the immune system.

Thymus
Spleen
Galt
Epigonal
DCs
NK?

Thymus
Spleen
Galt
Head kidney
DCs
NK

Thymus
Spleen
Galt
Bone marrow
DCs
NK

RAG AID
Ig no comb.

RAG AID
Ig comb.
TCR
MHCI/II

RAG AID
Ig comb.

Switch

TCR
MHCI/II

TB-cell memory
Poor Ab matur.

TB-cell memory
Poor Ab matur.

TH1 TH2
Type II IFN
Il20 TGF

Il2/15
Spleen
Epigonal organ
Thymus

Switch

TCR
CD3 /
MHC I/II
TB-cell memory
Poor Ab matur.

Il4
TGF
Treg

Thymus
Spleen
Galt
Bone marrow
Lymph nodes?
Germinal centers
FDCs
NK

Thymus
Spleen
Galt
Bone marrow
Lymph nodes
Germinal centers
FDCs
NK

RAG AID
Ig no comb. gen conv

RAG AID
Ig comb. gen conv

Switch

Switch

TCR
CD3 /
MHC I/II

TCR
CD3 with and
MHCI/II

TB-cell memory
Good Ab matur.?

TB-cell memory
Good Ab matur.

Il2/15

Treg
TH1,2,17
Lymph nodes

Thymoid

Thymus
patches

Spleen

Thymus
Rectal gland

17

Spleen

Fat column
(adult)
Head kidney

Galt
Bursa

Nodes?

Thymus

Galt
Bone marrow
Typhlosolis
(larva)

Teleostei
Agnatha

Aves

Anura

Spleen

Chondrichthyes

Bone marrow

Mammalia
Reptilia
Amphibia

Osteicthyes

Vertebrata

Gnathostomata

Figure 14 Recapitulating the immune system and responses of vertebrates. AID: Activation-induced deaminase, CDA: Cytidine deaminase, Comb:
Combinatorial, FDC: Follicular dendritic cells, GALT: Gut-associated lymphoid tissue, Gen. conv.: Gene conversion, Ig: Immunoglobulin, Il: Interleukin,
MHC: Major histocompatibility complex, NK: Natural killer cells, TCR: T-cell receptor, TGF: Transforming growth factor, TNF: Tumor necrosis factor,
TH1 and TH2: T helper cells 1 and 2, Treg: Regulatory T cells.

18

Evolution of the Immune System

Acknowledgments
I thank Drs D. Brites, H. Etlinger, and I. Lefkovits for critical
reading of this article. I thank Drs I. Chrtien and D. Visozo for
the artwork.

See also: Antigen Presentation; CD4 and CD8 Molecules:

Molecular Biology, Expression, and Function; Cellular and
Molecular Principles Underlying Allorecognition
and Allotransplant Rejection by T Lymphocytes; Development
of Human T Lymphocytes; Effector CD4 T Lymphocytes;
Evolution of the Immune System; Fc Receptors in Immune
Responses; Food Allergies; Genetically Engineered Antibody
Molecules; HLA, the Human Major Histocompatibility Complex;
Human Organ-Specic Autoimmune Disease; Human Primary
Immunodeciencies; Immunological Tolerance; Immunology
of Hematopoietic Stem Cell Transplantation; Inammasomes
and Danger Signals in the Immune System; Innate Immunity
and Allograft Rejection; Lymphocytes That Participate in Innate
Immune Responses; Mechanisms Underlying Transplantation
Tolerance; Natural Killer Cells; Receptors in Antiviral Immunity;
Role of IL-10 and the IL-10 Receptor in Immune Responses;
Structure and Function of Immunoglobulins; Systemic Lupus
Erythematosus and Systemic Autoimmunity; T-Cell Activation;
The gc Family of Cytokines: A Pleiotropic Family That Controls
Proliferation, Survival, and Differentiation; The Drosophila
melanogaster Model in Innate Immunity; The T-Cell Antigen
Receptor; Thymus; Toll-Like Receptor Function and Signaling;
Type I Interferons in Immunity; V(D)J Recombination, Somatic
Hypermutation, and Class Switch Recombination; Vaccines
against Cancer; White Blood Cells and Lymphoid Tissue.

References
Augustin, R., Fraune, S., Bosch, T.C., 2010. How Hydra senses and destroys
microbes. Semin. Immunol. 22 (1), 5458.
Bajoghli, B., Guo, P., Aghaallaei, N., Hirano, M., Strohmeier, C., McCurley, N.,
Bockman, D.E., Schorpp, M., Cooper, M.D., Boehm, T., 2011. A thymus candidate
in lampreys. Nature 470, 9094.
Boehm, T., McCurley, N., Sutoh, Y., Schorpp, M., Kasahara, M., Cooper, M.D., 2012.
VLR-based adaptive immunity. Annu. Rev. Immunol. 30, 203220.
Boudinot, P., van der Aa, L.M., Jouneau, L., Du Pasquier, L., Pontarotti, P., Briolat, V.,
Benmansour, A., Levraud, J.P., 2011. Origin and evolution of TRIM proteins: new
insights from the complete TRIM repertoire of zebrash and puffersh. PLoS One 6,
e22022.
Castro, R., Bernard, D., Lefranc, M.P., Six, A., Benmansour, A., Boudinot, P., 2011. T cell
diversity and TcR repertoires in teleost sh. Fish Shellsh Immunol. 31, 644654.
Cerenius, L., Kawabata, S., Lee, B.L., Nonaka, M., Sderhll, K., 2010. Proteolytic
cascades and their involvement in invertebrate immunity. Trends Biochem. Sci. 35,
575583.
Cornillon, S., Pech, E., Benghezal, M., Ravanel, K., Gaynor, E., Letourneur, F.,
Brckert, F., Cosson, P., 2000. Phg1p is a nine-transmembrane protein superfamily member involved in dictyostelium adhesion and phagocytosis. J. Biol. Chem.
275, 3428734292.
Danchin, E.G., Pontarotti, P., 2004. Towards the reconstruction of the bilaterian
ancestral pre-MHC region. Trends Genet. 20, 587591.
Du Pasquier, L., 2006. Germline and somatic diversication of immune recognition
elements in Metazoa. Immunol. Lett. 104, 217.
Dziarski, R., Gupta, D., 2005. Peptidoglycan recognition in innate immunity.
J. Endotoxin Res. 11 (5), 304310 (Review).
Ferrandon, D., Imler, J.L., Hetru, C., Hoffmann, J.A., 2007. Nat. Rev. Immunol. 7, 862874.

Flajnik, M.F., Du Pasquier, L., 2008. Evolution of the immune system. In:
Paul, W.E. (Ed.), Fundamental Immunology, sixth ed. Wolters Kluwer Lippincott
Williams and Wilkins, Philadelphia, pp. 56124.
Flajnik, M.F., Du Pasquier, L. Evolution of the immune system. In Paul, W.E. (Ed.),
Fundamental Immunology, seventh ed. Wolters Kluwer Lippincott Williams and
Wilkins, Philadelphia, in press.
Flajnik, M.F., Kasahara, M., 2001. Comparative genomics of the MHC: glimpses into
the evolution of the adaptive immune system. Immunity 15, 351362.
Flajnik, M.F., Kasahara, M., 2010. Origin and evolution of the adaptive immune
system: genetic events and selective pressures. Nat. Rev. Genet. 11, 4759.
Flajnik, M.F., Tlapakova, T., Criscitiello, M.F., Krylov, V., Ohta, Y., 11 Apr 2012.
Evolution of the B7 family: co-evolution of B7H6 and NKp30, identication of a new
B7 family member, B7H7, and of B7s historical relationship with the MHC.
Immunogenetics (Epub ahead of print).
Gilmore, T.D., Wolenski, F.S., 2012. NF-kB: where did it come from and why?
Immunol. Rev. 246, 1435.
Guo, P., Hirano, M., Herrin, B.R., Li, J., Yu, C., Sadlonova, A., Cooper, M.D.,
2009. Dual nature of the adaptive immune system in lampreys. Nature 459,
796801.
Hansen, J.D., Vojtech, L.N., Laing, K.J., 2011. Sensing disease and danger: a survey
of vertebrate PRRs and their origins. Dev. Comp. Immunol. 35, 886897.
Hartenstein, V., 2006. Blood cells and blood cell development in the animal kingdom.
Annu. Rev. Cell Dev. Biol. 22, 677712.
Hirano, M., Das, S., Guo, P., Cooper, M.D., 2011. The evolution of adaptive immunity
in vertebrates. Adv. Immunol. 109, 125157.
Hsu, E., 2009. V(D)J recombination: of mice and sharks. Adv. Exp. Med. Biol. 650,
166179.
Janeway Jr., C.A., Medzhitov, R., 2002. Innate immune recognition. Annu. Rev.
Immunol. 20, 197216.
Kasahara, M., 2007. The 2R hypothesis: an update. Curr. Opin. Immunol. 19,
547552.
Lefranc, M.P., Lefranc, G., 2001. The T Cell Receptor. Facts Book. Academic Press,
London.
Lemaitre, B., Hoffmann, J., 2007. The host defense of Drosophila melanogaster.
Annu. Rev. Immunol. 25, 697743.
Leulier, F., Lemaitre, B., 2008. Toll-like receptorsdtaking an evolutionary approach.
Nat. Rev. Genet. 9, 165178.
Litman, G.W., Rast, J.P., Fugmann, S.D., 2010. The origins of vertebrate adaptive
immunity. Nat. Rev. Immunol. 10, 543553.
Loker, E.S., Adema, C.M., Zhang, S.M., Kepler, T.B., 2004. Invertebrate immune
systemsdnot homogeneous, not simple, not well understood. Immunol. Rev. 198,
1024.
MacMicking, J.D., 2012. Interferon-inducible effector mechanisms in cell-autonomous
immunity. Nat. Rev. Immunol. 12, 367382.
ODay, D.H., Keszei, A., 2012. Signalling and sex in the social amoebozoans. Biol. Rev.
Camb. Philos. Soc. 87, 313329.
Panchin, Y., Moroz, L.L., 2008. Molluscan mobile elements similar to the vertebrate recombination-activating genes. Biochem. Biophys. Res. Commun. 369,
818823.
Paul, W.E. (Ed.), 2008. Fundamental Immunology, sixth ed. Wolters Kluwer Lippincott
Williams and Wilkins, Philadelphia.
Robert, J., 2003. Evolution of heat shock protein and immunity. Dev. Comp. Immunol.
27, 449464.
Ronald, P.C., Beutler, B., 2010. Plant and animal sensors of conserved microbial
signatures. Science 330, 10611064.
Rosa, S.F., Powell, A.E., Rosengarten, R.D., Nicotra, M.L., Moreno, M.A.,
Grimwood, J., Lakkis, F.G., Dellaporta, S.L., Buss, L.W., 2010. Hydractinia allodeterminant alr1 resides in an immunoglobulin superfamily-like gene complex.
Curr. Biol. 20, 11221127.
Sadd, B.M., Schmid-Hempel, P., 2006. Insect immunity shows specicity in protection
upon secondary pathogen exposure. Curr. Biol. 16, 12061210.
Satake, H., Sekiguchi, T., 2012. Toll-like receptors of deuterostome invertebrates.
Front. Immunol. 3, 34.
Schatz, D.G., 2007. DNA deaminases converge on adaptive immunity. Nat. Immunol.
8, 551553.
Schmid-Hempel, P., 2011. Evolutionary Parasitology. Oxford University Press, Oxford.
Star, B., Nederbragt, A.J., Jentoft, S., Grimholt, U., Malmstrm, M., Gregers, T.F.,
Rounge, T.B., Paulsen, J., Solbakken, M.H., Sharma, A., Wetten, O.F., Lanzn, A.,
Winer, R., Knight, J., Vogel, J.H., Aken, B., Andersen, O., Lagesen, K., ToomingKlunderud, A., Edvardsen, R.B., Tina, K.G., Espelund, M., Nepal, C., Previti, C.,
Karlsen, B.O., Moum, T., Skage, M., Berg, P.R., Gjen, T., Kuhl, H., Thorsen, J.,

Evolution of the Immune System

Malde, K., Reinhardt, R., Du, L., Johansen, S.D., Searle, S., Lien, S., Nilsen, F.,
Jonassen, I., Omholt, S.W., Stenseth, N.C., Jakobsen, K.S., 2011. The genome
sequence of Atlantic cod reveals a unique immune system. Nature 477, 207210.
Vilches, C., Parham, P., 2002. KIR: diverse, rapidly evolving receptors of innate and
adaptive immunity. Annu. Rev. Immunol. 20, 217251.

19

Zucchetti, I., De Santis, R., Grusea, S., Pontarotti, P., Du Pasquier, L., 2009. Origin
and evolution of the vertebrate leukocyte receptors: the lesson from tunicates.
Immunogenetics 61, 463481.