z*t
发帖数: 863 | 1 Beutler得奖，Medzhitov会不会有意见？
http://www.sciencedirect.com/science/article/pii/S1074761309002
Ruslan Medzhitov
This year marks the 20th anniversary of a publication (Janeway, 1989) that
in many ways revolutionized our understanding of the immune system. As part
of the proceedings of the Cold Spring Harbor Symposium on Immune Recognition
, the paper authored by the late Charles A. Janeway, Jr. was not a standard
peer-reviewed publication. In fact, Charlie used to refer to it as “the
best paper I've never published.” According to the Francis Crick's
quotation above, Charlie was definitely a “professional” who had many
ideas and who hit more than one jackpot. The pattern recognition theory was
the biggest one. In addition to suggesting a general principle of innate
immune recognition, it provided a conceptual framework for the integration
of the two types of immunity—innate and acquired. It also suggested a new
set of ideas that explained the initiation of the immune responses, the
mechanism of adjuvant effects, and the evolution of different forms of
immune recognition. Collectively, these ideas provided guiding principles
for our current understanding of the functioning of the innate immune system
and its connection with adaptive immunity.
Early Studies on the Initiation of Innate and Adaptive Immunity
In the late 80s and early 90s, immunologists focused much of their attention
on the mechanisms of adaptive immune recognition, including the structure
and function of the antigen receptors, the mechanisms of MHC restriction,
and lymphocyte development and activation. However, little was known about
the requirements for the activation of adaptive immune responses. Although
the question of self-nonself discrimination has preoccupied the minds of
immunologists for decades, the solution was thought to be largely provided
by the pioneering works of Lederberg, Burnet, and Medawar ( [Billingham et
al., 1953] , [Burnet, 1959] and [Lederberg, 1959] ). The central idea at the
time was that the removal or inactivation of autoreactive clones during
lymphocyte development allowed the remaining lymphocytes to respond to any
antigens they may encounter because these by definition would be nonself.
The concept was later modified by the two-signal model of Bretscher and Cohn
(Bretscher and Cohn, 1970) and by the costimulatory model suggested by
Lafferty and Cunningham (Lafferty and Cunningham, 1975). The first
experimental evidence for the two-signal requirement for T cell activation
was later provided by Marc Jenkins and Ronald Schwartz (Jenkins and Schwartz
, 1987), although the nature of the second signal remained unknown for a few
more years. Importantly, however, in an unrelated line of studies, Antonio
Coutinho and Göran Möller discovered the distinct roles of antigen
-specific and polyclonal signals in B cell activation (Coutinho and Moller,
1974). The polyclonal (or mitogenic) signal in their experiments was
provided by lipopolysaccharide (LPS), and thus their studies made one of the
first critical connections between antigen-specific receptor and nonclonal
microbial receptors in activation of B cell responses. Other studies
established adjuvant properties of LPS ( [Condie et al., 1955] and [Skidmore
et al., 1976] ). In the meantime, Ralph Steinman, working in Zanvil Cohn's
group, discovered dendritic cells and characterized their unique
immunostimulatory activity ( [Steinman and Cohn, 1973] and [Steinman and
Cohn, 1974] ). A critical connection between the initiation of T and B cell
responses was provided by the discovery of antigen-specific T and B cell
interactions by Antonio Lanzavecchia (Lanzavecchia, 1985). Together, these
and many other pioneering discoveries, including analyses of the specificity
of antibody responses by Karl Landsteiner (Landsteiner, 1945), provided the
essential background for future analyses of how the adaptive immune
response is initiated.
In parallel, fundamental advances were being made in studies of the innate
host defense mechanisms. During the 1970s, analyses of antimicrobial
mechanisms of neutrophils and macrophages, particularly by Zanvil Cohn and
his many pupils and collaborators, including James Hirsch, Seymour Klebanoff
, Carl Nathan, Siamon Gordon, Ralph Steinman, Alan Aderem, Ira Mellman, Bill
Muller, Gilla Kaplan, and many others, had established many basic
principles of macrophage and neutrophil functions, including phagocytic
process, degranulation, secretory function, arachidonic acid metabolism, and
oxygen-dependent and -independent bacteriocidal functions (reviewed in
Steinman and Moberg, 1994). Many of these functions were found to be
inducible in macrophages by microbial stimuli, including LPS, Bacillus
Calmette-Guérin (BCG), and zymosan. Thus, the idea of microbial stimulation
of innate effector functions was well appreciated by that time. In addition
, the discovery of antimicrobial peptides by Hans Boman revealed a powerful
microbicidal system, which is now known to be used by most and perhaps all
organisms for antimicrobial defense. In the field of antiviral immunity, the
crucial advances were the discovery of type I interferon (IFN) activity ( [
Isaacs and Lindenmann, 1957] , [Isaacs et al., 1957] and [Nagano and Kojima,
1958] ), cloning of the first type I IFN gene (IFN-β) by Tadatsugu
Taniguchi (Taniguchi et al., 1980), and the isolation of the IFN-α gene by
Shigekazu Nagata and Charles Weismann (Nagata et al., 1980). Importantly,
type I IFNs were known to be inducible by poly IC, presumably because it
mimics viral dsRNA—another important target of innate immune recognition.
Finally, characterization of two mouse strains deficient in LPS
responsiveness, C3H/HeJ (Heppner and Weiss, 1965) and C57BL/10Cr (Coutinho
et al., 1977), and demonstration of their susceptibility to a bacterial
infection (O'Brien et al., 1980) demonstrated that LPS recognition plays a
critical role in host defense against at least some Gram-negative pathogens.
The two lines of study briefly outlined above were developing largely
independently of each other, as reflected in the way the innate and adaptive
immune systems were covered in textbooks. The critical role of innate
immune recognition in the control of adaptive immunity was not well
appreciated and a new conceptual framework was clearly needed to integrate
innate and adaptive immune recognition to explain the fundamental properties
of mammalian immunity.
Janeway's Hypothesis
Charles Janeway's unique contribution was in developing a new synthesis that
placed many immunological phenomena in a clear biological context. First,
he proposed that the costimulatory signal required for lymphocyte activation
was inducible (on antigen-presenting cells [APCs]). Whether and how the
costimulatory signals are regulated was not well understood at the time,
although expression of interleukin-1 (IL-1) by macrophages was found by Emil
Unanue and colleagues to be one of them (Weaver et al., 1988). Second, the
costimulatory signal was suggested to be inducible by conserved microbial
products, thus placing the activation of adaptive immunity under the control
of pathogen-sensing mechanisms. This also explained the adjuvant properties
of certain microbial stimuli. The fact that many microbial molecules, such
as LPS, had immunostimulatory properties was known before, but it was not
clear why some microbial molecules have these properties, including adjuvant
activity, whereas others do not. Neither was it known how these microbial
structures exerted their adjuvant effects. Janeway suggested that the actual
detection of infection was mediated by the receptors of the innate immune
system, rather than the antigen receptors. Specifically, he proposed that
innate immune system determined the origin of antigens recognized by T and B
cells and instructed the latter to initiate the response if antigen was of
microbial origin. Pathogen sensing, in turn, was proposed to be mediated by
a set of germline-encoded pattern recognition receptors that detect
conserved products of microbial biosynthetic pathways (known as pathogen-
associated molecular patterns [PAMPs]). Janeway further pointed out that
this form of immune recognition must be evolutionarily related to the immune
systems of invertebrates, which lack adaptive immunity. Finally, most
adjuvants were suggested to work in part by mimicking microbial infections,
by triggering the receptors of the innate immune system and inducing
costimulatory signals, thus “tricking” the adaptive immune system into
action. These and other ideas provided an elegant explanation of many
fundamental aspects of the functioning of the immune system. Remarkably, all
of them turned out to be fundamentally correct. Nevertheless, the new
theory did not seem to attract a lot of attention at first. The situation
started to change about a decade later, when new studies began elucidating
the pathways of innate immune recognition.
Toll-like Receptors
By the mid-90s, several receptors involved in innate immune recognition were
already known. These included CD14, a component of the LPS receptor complex
(Wright et al., 1990); Mannan-binding lectin (MBL), an activator of the
lectin pathway of complement (Takahashi et al., 2006); macrophage mannose
receptor (Stahl and Ezekowitz, 1998); and MARCO, a member of the scavenger
receptor family (Elomaa et al., 1995). MBL was particularly well
characterized in terms of its specificity and function. Its role in
complement activation and microbial opsonization was reminiscent of
immunoglobulins, suggesting that MBL functioned as a nonclonal version of
antibodies. To emphasize this point, Alan Ezekowitz referred to MBL as “
ante-antibody,” highlighting the evolutionary ancestry of the former (
Ezekowitz, 1991). However MBL, as well as pentraxins, although known at the
time to function as pattern recognition molecules, were not the type of
receptors that would be expected to control the expression of costimulatory
signals on APCs. Therefore, Janeway and his colleagues were looking for cell
-surface receptors, expressed on macrophages and dendritic cells (DCs), that
would be able to induce signaling pathways resulting in expression of B7.1
and B7.2 (now known as CD80 and CD86, respectively). Because many PAMPs are
either carbohydrates or glycolipids, the new pattern recognition receptors
were thought to probably contain a C-type lectin domain, similar to MBL and
macrophage mannose receptor. In addition, the intracellular portion was
expected to be linked to the activation of the NF-κB signaling pathway.
This pathway was already well established to play a pivotal role in innate
immunity (Kopp and Ghosh, 1995), and studies by Stephanie Vogel, Doug
Golenbock, Richard Ulevitch, and many others have demonstrated the role of
NF-κB in LPS signaling. The best-characterized receptors known at the time
to activate NF-κB were TNF and IL-1 receptors. The IL-1 receptor was
particularly interesting in this regard because it contained a cytoplasmic
domain, which was known to occur in Drosophila Toll, identified in 1988 by
Carl Hashimoto and Kathryn Anderson (Hashimoto et al., 1988), and in a
resistance protein from tobacco known as N protein, which gave it the name
TIR (for Toll, IL-1R, and Resistance protein) domain ( [Gay and Keith, 1991]
and [Whitham et al., 1994] ). The idea therefore was to find a new
transmembrane protein that would contain a C-type lectin ectodomain and a
cytoplasmic TIR domain.
In early 1996, a piece of an expressed sequence tag sequence was identified
in the NCBI database as containing similarity to a portion of a TIR domain
and used to screen a splenic cDNA library. To initial disappointment, the
full-length clone did not contain a C-type lectin domain. Instead, it was a
homolog of Drosophila Toll, which was known to have an endogenous protein
ligand Spatzle (Schneider et al., 1994). Moreover, one sequence with
similarity to Drosophila Toll was deposited in the database as “randomly
sequenced cDNA 786” (rsc786) (Nomura et al., 1994), and it was reported in
early 1996 that the protein it encoded did not activate NF-κB (Mitcham et
al., 1996). The reason for this negative result was that rsc786 is what is
now known as TLR1, which was subsequently shown by Alan Aderem and
colleagues to signal only as a heterodimer with TLR2. However, the clone
isolated in the Janeway lab encoded what is now known as TLR4, which
functions as a homodimer, and accordingly, its constitutively active form
was able to induce NF-κB. Furthermore, a few months later, in the summer of
1996, we learned from Jules Hoffmann about their stunning discovery of the
essential role of Drosophila Toll in antifungal defense. This finding (
Lemaitre et al., 1996) further reinforced the possibility that a mammalian
Toll homolog may also have an immune function, and indeed it turned out to
induce expression of genes encoding B7 and several cytokines (Medzhitov et
al., 1997), suggesting that TLR may couple innate immune recognition with
activation of adaptive immunity (Fearon, 1997). On the basis of this, and
given the immune function of Drosophila Toll discovered by Bruno Lemaitre
and Jules Hoffmann, human Toll was suggested to be involved in recognition
of microbial PAMPs (Medzhitov and Janeway, 1997). The assumption at the time
was that similar to its fly counterpart it would function downstream of a
proteolytic cascade triggered by microbial recognition. However, the
identity of the TLR ligands remained unknown for another year, despite
numerous attempts by multiple groups to test whether human TLR4 was a
receptor for LPS. All these attempts failed because the other key component
of the LPS receptor complex, MD2, was not known at the time and was
discovered by Kensuke Miyake and colleagues only two years later (Shimazu et
al., 1999). In the meantime, several additional members of the Toll-like
receptor family were identified ( [Chaudhary et al., 1998] and [Rock et al.,
1998] ).
The first evidence that TLRs may indeed be involved in microbial recognition
was provided by Paul Godowsky and colleagues (Yang et al., 1998). This
study indicated that TLRs could directly respond to bacterial products.
Specifically, it reported that TLR2 conferred responsiveness to crude
preparations of LPS, which also contained bacterial lipoproteins. The same
result was also reported a few months later by a group from Tularik (
Kirschning et al., 1998). We now know that TLR2 in these experiments
responded to lipoproteins rather than to LPS (Hirschfeld et al., 2000).
Nevertheless, the demonstration that a TLR recognized a microbial stimulator
had its impact because it immediately suggested that all mammalian TLRs may
recognize different microbial ligands and that unlike Drosophila Toll, they
may recognize them directly. In a commentary piece accompanying Godowsky's
paper, Craig Gerard pointed out that the gene encoding human TLR4 was
located in a chromosomal region (9q32-33) syntenic to mouse chromosome 4, in
which LPS unresponsiveness in the C3H/HeJ strain was previously mapped,
suggesting that TLR4 may be involved in LPS recognition (Gerard, 1998).
Positional cloning of the lps locus by Bruce Beutler and colleagues (
Poltorak et al., 1998) demonstrated that TLR4 was required for LPS
responsiveness, thus providing the first genetic evidence for TLR4 function
in microbial recognition. The same result was later reported by Danielle
Malo and colleagues (Qureshi et al., 1999). Subsequent studies by Shizuo
Akira and many others have elucidated the specificities of other TLRs for
various microbial ligands (Takeda et al., 2003). Thanks to these
contributions, we now know that TLRs detect a variety of bacterial, viral,
fungal, and protozoal ligands. TLR signaling pathways have also been
elucidated in some detail, again in large part because of the numerous gene
“knockout” mice generated by Shizuo Akira and colleagues. The latest
breakthrough in the TLR field was the elucidation of the structures of
several TLRs, in some cases in complex with their ligands (Jin and Lee, 2008
). These structures allowed us to see in great detail the structural
principles of pattern recognition.
Other Pattern Recognition Receptors and Their Functions
Elucidation of the functions of TLRs in mammalian immunity has provided
experimental support for the concept proposed by Charlie Janeway two decades
ago. However, TLRs are not the only sensors of microbial infection. In May
of 1999, two independent studies reported identification and initial
characterization of NOD1, a founding member of a family of cytosolic
receptors ( [Bertin et al., 1999] and [Inohara et al., 1999] ). NOD1 and a
related protein, NOD2, were subsequently found to sense fragments of
bacterial peptidoglycans ( [Chamaillard et al., 2003] , [Girardin et al.,
2003a] , [Girardin et al., 2003b] and [Inohara et al., 2003] ). The role of
NOD proteins in host defense is still being elucidated, but genetic studies
have revealed their role in intestinal immunity (Kobayashi et al., 2005),
inflammation (Maeda et al., 2005), and the pathogenesis of Crohn's disease (
[Hugot et al., 2001] and [Ogura et al., 2001] ). In 2001, Gordon Brown and
Siamon Gordon identified Dectin-1 as a receptor for β-glucans (Brown and
Gordon, 2001). In subsequent studies, Dectin-1 was found to play important
roles in antifungal immunity ( [Gross et al., 2006] , [Leibundgut-Landmann
et al., 2007] , [Saijo et al., 2007] and [Taylor et al., 2007] ). In 2004,
Takashi Fujita and colleagues identified RIG-I and MDA-5 as intracellular
sensors for double-stranded RNA (Yoneyama et al., 2004). RIG-I was also
found in a screen for mutations permissive for hepatitis C virus replication
(Sumpter et al., 2005). RIG-I and MDA-5 were later shown to play essential
roles in sensing RNA viruses and initiating antiviral immunity ( [Gitlin et
al., 2006] , [Kato et al., 2005] and [Kato et al., 2006] ). Definition of
the RNA ligands recognized by RIG-I revealed the molecular pattern sensed by
this receptor ( [Hornung et al., 2006] and [Pichlmair et al., 2006] ).
Finally, in 2002, Jürg Tschopp and colleagues described the inflammasomes-
multiprotein complexes that activate caspase-1 and other inflammatory
caspases to generate and secrete mature forms of IL-1 family members (
Martinon et al., 2002). Although it is not yet clear how all the different
inflammasomes are activated, at least some of them may function as
components of a pattern recognition system (Meylan et al., 2006).
Together with pentraxins, collectins, and ficolins ( [Bottazzi et al., 2006]
and [Holmskov et al., 2003] ), these newly identified pattern recognition
receptors account for most of the pathogen-sensing capabilities of mammalian
hosts (Figure 1). However, there are clearly additional pathogen
recognition systems that remain to be identified. One example is the
intracellular receptors involved in recognition of viral and bacterial DNA
that trigger the type I interferon response ( [Ishii et al., 2006] , [Okabe
et al., 2005] and [Stetson and Medzhitov, 2006] ). It is likely that there
is more than one receptor involved in intracellular sensing of microbial DNA
(Takaoka et al., 2007). Another example is the receptor(s) involved in the
recognition of chitin. Chitin is an abundant polysaccharide produced by
arthropods and parasitic worms, and it was recently demonstrated that chitin
can trigger a type 2 inflammatory response (Reese et al., 2007). At least
some of the activities of chitin are mediated by TLR2 (Da Silva et al., 2008
), although additional receptors are likely to exist. Other receptors
involved in recognition of parasitic worms remain to be identified. Innate
immune recognition of helminthes is poorly understood, and it is not even
clear to what extent pattern recognition plays a role here. Indeed,
multicellular parasites may not be evolutionarily distant enough from
mammalian hosts to produce reliable targets for pattern recognition
receptors. With the exception of chitin, there are no other known molecular
products that are conserved among parasites and distinct from the mammalian
hosts. Therefore, pattern recognition may not play a major role in the
detection of multicellular parasites. Instead, other forms of innate immune
recognition may play a predominant role (see below).
Full-size image
High-quality image (615K)
Figure 1.
Pattern Recognition Receptors
Major classes of mammalian pattern recognition receptors and some of their
representative microbial ligands are shown. Secreted pattern recognition
molecules activate complement and function as opsonins. Transmembrane and
intracellular receptors trigger signaling pathways for innate host defense
and activate adaptive immune responses. Mechanisms of inflammasome
activation are still being defined. NLRs, NOD-like receptors; CRD,
carbohydrate recognition domain; ITAM, immunoreceptor tyrosine-based
activation motif.
View Within Article
Pattern recognition also does not explain immune responses to transplants
and to most allergens. However, there is no other theory (and there may
never be) that can explain all immune responses, simply because they do not
all follow the same rules. Rather, different mechanisms of immune
recognition may sometimes work in specific combinations to trigger optimal
immune responses. Therefore, different theories need not be competing, so
long as they explain distinct phenomena. Thus, in addition to pattern
recognition theory, three other concepts have been proposed to explain the
mechanisms of activation of immune responses. These are the missing self
theory, the guard theory, and the danger theory.
Missing Self Recognition
The missing self concept was proposed by Klas Kärre to explain the
specificity of tumor cell recognition by NK cells (Karre et al., 1986). The
molecular basis for missing self recognition by NK cells is the engagement
of a class of inhibitory receptors (initially discovered by Wayne Yokoyama
and colleagues), typically containing immunoreceptor tyrosine-based
inhibitory motifs (ITIMs) (Karlhofer et al., 1992). They recruit SHP-1
tyrosine phosphatase to block Syk-72 dependent NK activating pathways,
including cytotoxicity (Raulet and Vance, 2006). The missing self concept
also applies to many other forms of innate immune recognition. Activation of
the alternative complement pathway (Gotze and Muller-Eberhard, 1971), for
example, is negatively regulated, among other things, by Factor H, which can
distinguish between microbial and host cells by binding to host-specific
sialic acids and preventing the alternative pathway of complement activation
on host cells (Meri and Pangburn, 1990). There is growing evidence that
phagocytosis is also negatively regulated by a variety of ITIMs containing
inhibitory receptors, including SIRPα and probably Siglec proteins. Thus,
recognition of CD47 on target cells by SIRPα prevents their phagocytosis by
macrophages (Oldenborg et al., 2000). Siglecs may play a similar role by
recognizing host-specific sialic acids, analogously to Factor H (Crocker and
Varki, 2001). Sialic acid recognition on antigens by CD22 may play a
similar role in controlling B cell activation (Doody et al., 1995).
It should be noted that missing self recognition is essentially a negative
control of some activating signal and thus always works in conjunction with
an activating pathway. The activating signal itself can be triggered by a
variety of mechanisms. Thus, the alternative complement pathway is activated
by properdin, which was discovered in the early 1950s by Louis Pillemer (
Pillemer et al., 1954). Properdin itself has features of a pattern
recognition protein (Kemper and Hourcade, 2008). Phagocytosis is positively
regulated by a variety of phagocytic receptors (Underhill and Ozinsky, 2002)
, some of which are pattern recognition receptors that directly recognize
microbial cells. Finally, NK cell cytotoxic activity is triggered by
activating receptors, which recognize ligands that are induced on target
cells upon viral infection and by some forms of stress (Lanier, 2005). How
these ligands are induced in virally infected cells is not yet well defined,
but it is possible that the intracellular sensors of viral infections, such
as RIG-I and MDA-5, may play a role. Thus it appears that in some (but
certainly not all) cases, the positive signals that are opposed by the “
missing self” pathways may be inducible by the pattern recognition
receptors. In other cases, they may be triggered by mechanisms that are
explained by the guard theory.
Guard Theory
Another theory of immune recognition, known as “guard theory,” was
proposed by Dangl and Jones to explain the genetics of host resistance in
plants (Dangl and Jones, 2001). Individual resistance genes in plants are
often required to confer protection from a specific strain of pathogen
carrying a particular virulence factor. (In keeping with the overly
confusing terminology in animal immunity, “virulence factors” in plant
immunity are referred to as “avirulence factors.”) The idea of the guard
theory is that rather than detecting pathogens directly, the products of at
least some individual resistance genes monitor, or “guard,” certain
cellular processes that are often targeted by the virulence factors of
pathogens. Disturbances in these processes are sensed by resistance gene
products, which become activated and trigger a host defense response, for
example, the hypersensitivity response. This form of microbial detection
thus does not require direct molecular recognition of the pathogen, but
rather is based on sensing the results of virulence factor activity by
monitoring their effects on the host. Because plants also use pattern
recognition for pathogen sensing, it is likely that the two mechanisms
normally work together (Jones and Dangl, 2006).
Do the principles of guard theory apply to animal immunity? Two lines of
recent evidence suggest that they do. First, at least some inflammasomes (
specifically the NALP3 inflammasome) appear to function by monitoring
cellular membrane integrity (Ogura et al., 2006). As it happens, cellular
membrane integrity is often targeted by pathogens. Many Gram-positive
pathogens, for example, produce pore-forming exotoxins that can be sensed by
the NALP3 inflammasome (Mariathasan et al., 2006). The NALP3 inflammasome
can also be activated by noninfectious stimuli that affect cellular
membranes, although the mechanisms involved remain poorly defined (Martinon
et al., 2009). Interestingly, NALP3 and other NOD-like receptors are
structurally similar to the products of many plant resistance genes,
although it is too early to say whether this similarity extends to their
mechanisms of action.
The second example of sensing virulence activity is based on the detection
of a proteolytic activity. In Drosophila, the proteolytic cascade that leads
to processing of Spatzle and activation of Toll is normally triggered by
pattern recognition molecules (Hoffmann, 2003), but can also be triggered by
a virulence factor (a protease) secreted by a fungal pathogen. This
protease, in turn, can trigger (and therefore is sensed by) the Spatzle-
activating proteolytic cascade resulting in the induction of antifungal
immunity (Gottar et al., 2006). In mammals, proteolytic activity of protease
allergens is detected by unknown sensors expressed on basophils, resulting
in the initiation of a T helper 2 cell response (Sokol et al., 2007). This
mechanism of allergen sensing presumably evolved to detect proteases
secreted by parasitic worms, whereas allergens happen to trigger the same
pathway inadvertently. As mentioned above, the pattern recognition strategy
may not be sufficient or even practical for the detection of multicellular
parasites. The host may have evolved instead a set of sensors that detect
abnormal biochemical activities associated with helminth infections, much
like guard proteins in plants detect activities of virulence factors. Cell-
autonomous sensing of cellular perturbations associated with viral
infections may also be coupled with the induction of positive signals for NK
cell activation.
Interestingly, in some contexts, inflammasome activation and IL-1 production
are critical for the induction of T cell responses ( [Ben-Sasson et al.,
2009] and [Ichinohe et al., 2009] ), just as sensing of allergen protease
activity by basophils is critical for Th2 cell responses (Sokol et al., 2007
). This suggests that indirect sensing of “virulence activity” can indeed
be linked with the activation of adaptive immunity.
Danger Theory
The danger theory was proposed by Polly Matzinger in 1994 as an alternative
to self-nonself discrimination theories (Matzinger, 1994). The main point of
the danger model is that the immune system is not concerned with the origin
of the antigen. Rather, the immune response was suggested to be initiated
by “danger signals” released from damaged tissues. Thus, anything that can
cause tissue damage, be it a pathogen or a surgical procedure, was proposed
to induce an immune response. Benign stimuli not associated with tissue
damage, whether microbial or not, were proposed to be ignored by the immune
system. Thus, according to the danger model, initiation of the immune
response is not dependent on microbial recognition, but rather on the
ability of pathogens (or a transplantation procedure) to cause tissue damage
. Moreover, according to the danger model, pattern recognition receptors (
for example, TLRs) did not evolve to recognize their microbial ligands (for
example, LPS); rather, the microbial ligands evolved to bind to and activate
TLRs (Matzinger, 2002).
Although the danger theory gained popularity in the past decade, there
appears to be a lot of confusion as to what exactly is considered to be “
danger signals.” The term “danger signal” is now often used to refer to
microbial TLR ligands, which is ironic because the implication is the exact
opposite of the main postulate of the danger theory. In other cases, PAMPs
are being distinguished from DAMPs (damage-associated molecular patterns).
There are certainly signals released from damaged tissues that trigger
inflammatory and tissue repair responses. Tissue damage of course initiates
the inflammatory response, but in the absence of infection, this response
leads to tissue repair, rather than adaptive immunity, which in this case
should be autoreactive. It is clear that microbial ligands of pattern
recognition receptors can trigger the immune response even in isolation from
the pathogens. Similarly, nonpathogenic microbes that do not cause tissue
damage can also induce immune responses, as evidenced, for example, by the
generation of IgA antibodies specific to commensal bacteria (Macpherson and
Slack, 2007). Thus, microbial recognition appears to be sufficient to
initiate the immune response, even in the absence of tissue damage.
Microbial recognition, however, does not explain all adaptive immune
responses. Some types of cell death are thought to be immunogenic in some
contexts ( [Green et al., 2009] and [Rock and Kono, 2008] ), and receptors
that can sense immunogenic forms of cell death are beginning to be
identified ( [Sancho et al., 2009] and [Yamasaki et al., 2008] ). Certain
aspects of the danger model may indeed be valid; tissue damage is likely to
play a role in the regulation of immune responses, although it is unlikely
to be necessary or sufficient to induce an adaptive immune response. But
then again, this depends on the definition of adaptive immune responses. If
there are adaptive immune responses that play roles other than protection
from microbial infections, than we will all have to reconsider many of the
arguments on which we currently rely.
The Janeway model and the Matzinger model started off as alternative views
on the mechanisms of initiation of immunity. The two models have now
coexisted for 15 years. As Francis Crick once noted: “The basic trouble is
that nature is so complex that many quite different theories can go some way
to explaining the results.” Eventually, the differences in our views will
undoubtedly degenerate into purely semantic arguments, at which point they
will become irrelevant.
Innate Control of Adaptive Immunity
The principles of innate control of adaptive immunity have been discussed
before ( [Fearon and Locksley, 1996] and [Iwasaki and Medzhitov, 2004] ) and
will not be covered here in detail. However, some aspects of this important
subject have recently become controversial, thus generating a great deal of
confusion.
As discussed above, several families of pattern recognition receptors are
now known and a few more remain to be identified. In addition, other
strategies of innate immune recognition, in particular sensing of virulence
activities and missing self recognition, are likely to play an important
role in the control of adaptive immunity. It is now well established that
several pattern recognition receptors are coupled with the induction of
adaptive immunity, including TLRs, NLRs, Dectin-1, RIG-I, and MDA-5 (Palm
and Medzhitov, 2009b). Additional, currently unknown receptors that can
induce adaptive immune responses must exist as well because some immune
responses (for example, some Th2 cell responses) cannot be accounted for by
any of the known innate immune-sensing mechanisms. It should also be self-
evident that if more than one pathway is sufficient to induce an immune
response, then none of the individual pathways is required in general,
except in the absence of stimuli for any of the alternative pathways.
Numerous studies have now demonstrated that some adaptive immune responses
are dependent on the TLR-MyD88 pathway, whereas others are not (Palm and
Medzhitov, 2009b). However, a recent report (Gavin et al., 2006) claimed
that TLRs are not involved in the control of adaptive immunity and that
common adjuvants do not work through TLRs. This caused a great deal of
confusion, given that the reasons for these discrepant results were unclear.
That main reason for the discrepancy, as it turns out, has to do with the
type of antigen used in immunizations. Unlike the previous publications ( [
Pasare and Medzhitov, 2004] and Pasare and Medzhitov, 2005 C. Pasare and R.
Medzhitov, Control of B-cell responses by Toll-like receptors. Nature, 438
(2005), pp. 364–368. | View Record in Scopus | | Full Text via CrossRef |
Cited By in Scopus (290) [Pasare and Medzhitov, 2005] ), in which native
protein antigens were used for immunizations, Gavin et al. exclusively
examined immune responses to haptenated antigens. However, haptenated
antigens, unlike native protein antigens, have intrinsic immunogenic
activity (Palm and Medzhitov, 2009a). Thus, the results reported by Gavin et
al. are not directly comparable to the previous reports. Second, adjuvants
have at least two types of activities: First, they trigger innate immune
recognition pathways, commonly by engaging one or more pattern recognition
receptors, and second, mineral oil (a component of Freund's adjuvants) and
alum have a depot effect that is critical when soluble antigens are used (
but probably dispensable for particulate antigens). Thus the TLR-independent
, adjuvant-mediated enhancement of the antibody response to haptenated
antigens reported by Gavin et al. most likely reflects the depot effect, as
previously reported for MyD88-independent Th2 cell responses (Schnare et al.
, 2001). Immunogenicity, on the other hand, was provided as a result of the
use of haptenated antigens. The fact that TLRs are not required for all
immune responses have been well documented before ( [Rivera et al., 2006] ,
[Schnare et al., 2001] and [Way et al., 2003] ), as is the fact that TLRs
and several other pattern recognition receptors are sufficient to induce
adaptive immunity, thereby making the requirement for any particular
receptor dependent on the pathogen or antigen and adjuvant used in
immunization.
TLRs in Human Immunity
Recent studies have uncovered an important role of TLRs in human immunity.
Thus, individuals carrying mutations in TLR5 were found to be susceptible to
Legionnaires' disease (Hawn et al., 2003). Surprisingly, TLR3 deficiency in
humans resulted in susceptibility to herpes simplex encephalitis (Zhang et
al., 2007). In addition, patients with null mutations in IRAK4 and MyD88 are
highly susceptible to several bacterial infections, which often result in
lethality early in life ( [Ku et al., 2007] , [Medvedev et al., 2003] and [
von Bernuth et al., 2008] ). If the patients passed the critical period in
the first few years of life, however, compensatory host defense mechanisms,
presumably in large part antibody-mediated, afford protection from some
pathogens, although the patients still suffer from recurrent pyogenic
bacterial infections. Although these findings were interpreted as indicating
a narrowly restricted and redundant role of TLRs in human immunity (Ku et
al., 2007), this interpretation is incongruent with the fact that almost
half of the patients with IRAK4 deficiency die from infections, even though
a majority of the patients are presumably on antibiotic medications. In a “
natural” environment, without modern medical interference, presumably all
of them would be dead from infections early in life. Clearly, however, in
some of the patients examined so far, the adaptive immune response can
partially compensate for TLR signaling defects. It would be interesting to
learn in the future the rules of immunological compensation, which
presumably operate in most types of nonlethal immunodeficiencies.
Conclusions
The 20 years that passed since Janeway's publication have seen a tremendous
progress in our understanding of the innate immune system and its role in
mammalian host defense. Still, to be sure, we will continue “approaching
the asymptote” for many more years. No theory is ever complete, and pattern
recognition concept will continue to evolve and eventually integrate with
other concepts into a more general theory. We soon will know the entire
repertoire of pattern recognition receptors and will learn more about other
innate immune-sensing mechanisms. We will then be able to learn all possible
pathways for the activation of adaptive immunity. Perhaps at that point we
will be able to address far more daunting questions of the requirements for
the generation of protective immunity. Here we know very little, and this in
part explains our continuous failures in developing new effective vaccines.
Once the principles of protective immunity are understood, we may finally
be able to deliver on the promise of defeating major infectious diseases
with vaccination. But then again, pathogens may have a different opinion on
this matter.
Acknowledgments
This review is not meant to be a comprehensive account of the history of
innate immunity, and therefore many important contributions are not
discussed or cited here. Some of the views expressed here reflect my
personal opinion and should not be taken as representing a consensus in the
field. I would like to thank T. Ichinohe for helping with the figure and my
friends and colleagues for reading the manuscript. This article is dedicated
to the memory of C. Janeway. | s*********y
发帖数: 387 | 2 THANKS FOR SHARING.
I THINK THERE IS A BIG INCONSISTENT BETWEEN THESES TWO GROUPS IN SEVERAL
EARLY S AND N PAPER. RUSLAN IS THE ONE PROVED TO BE WRONG?
NOT SURE.
WAITING MORE INFORM
part
Recognition
standard
【在 z*t 的大作中提到】
: Beutler得奖，Medzhitov会不会有意见？
: http://www.sciencedirect.com/science/article/pii/S1074761309002
: Ruslan Medzhitov
: This year marks the 20th anniversary of a publication (Janeway, 1989) that
: in many ways revolutionized our understanding of the immune system. As part
: of the proceedings of the Cold Spring Harbor Symposium on Immune Recognition
: , the paper authored by the late Charles A. Janeway, Jr. was not a standard
: peer-reviewed publication. In fact, Charlie used to refer to it as “the
: best paper I've never published.” According to the Francis Crick's
: quotation above, Charlie was definitely a “professional” who had many
|
|
