|
Now I Know My CpGs |
|
[编者的话] The vertebrate innate immune system detects the presence of infection by the recognition of molecular patterns that are present in many pathogens but absent from host tissues. For example, lipoproteins, peptidoglycans and lipopolysaccharides are all detected by one or more of the Toll-like receptors (TLRs), which have emerged as key receptors in this innate immune function. A recent study reports that a CpG motif, an immune stimulatory sequence characteristic of bacterial but not vertebrate DNA, is also detected by the TLR system.
The vertebrate immune system is
frequently faced with new antigens. Some antigens are associated with
pathogens and necessitate a rapid response if the organism is to
survive. Other antigens might initially be expressed as the result of
physiological changes (e.g. puberty) or can be ingested in food, in
which case immune responses should be avoided. One of the central
problems for the immune system is in deciding which antigens to respond
to - how does it tell when it is really being infected? The answer to
this question has much greater implications than merely satisfying
scientific curiosity. If it were possible to 'trick' the immune system
at will into reacting as if a particular antigen was from an infectious
pathogen, then it should be possible to design safer and more effective
vaccines and cancer immunotherapies. Activation of Innate Immunity The proinflammatory effects of
bacterial extracts have been recognized for centuries. In fact, crude
extracts known as 'Coley's toxins' were used in treating >1000
patients with advanced malignancy, with some remarkable successes
(reviewed in reference
1). Mycobacterium bovis bacillus Calmette-Guérin (BCG) is highly
active in many mouse tumor models and continues to be useful in the
immunotherapy of bladder cell cancer in humans. Studies to determine the
active component within BCG revealed the surprising finding that the DNA
itself was the component capable of inducing anti-tumor immune responses
in mice [2,3].
Subsequently, we demonstrated that bacterial DNA activates several cell
types (figure
1) and that the immune stimulatory effects resulted from the
presence of CpG dinucleotides in particular base contexts. These CpG motifs distinguish
bacterial DNA from vertebrate DNA in three ways. First, vertebrate
genomes uniformly show an unexplained feature termed CpG suppression -
the frequency of CpG dinucleotides is only ~25% of the frequency that
would be expected if base usage was random. Second, vertebrate genomes
show CpG skewing, that is, the CpG dinucleotides that are present more
commonly occur in base contexts that are not immune stimulatory than in
base contexts that are [4].
Third, the CpG dinucleotides in vertebrate genomes are usually
methylated, which abolishes their immune stimulatory effect [5]. That the unmethylated CpG
motifs in bacterial DNA are responsible for its immune stimulatory
activity is demonstrated by the fact that this activity is lost when
these motifs are methylated5. However, demethylation of the CpG motifs
in vertebrate DNA does not restore immune stimulatory activity [6],
probably owing to the skewing in the CpG motifs and the presence of
inhibitory sequences [4]. Toll-like Receptors In recent years, a new paradigm
has arisen to explain the mechanisms used by the immune system to detect
infections rapidly. This function is fulfilled by the innate immune
system, through cells such as dendritic cells (DCs), macrophages,
monocytes and neutrophils. These innate immune cells lack the
sophisticated and exquisitely antigen-specific receptors of T and B
cells, but have instead evolved a set of pattern recognition receptors
(PRRs) that enable them partially to sense the world around them.
Different classes of pathogens, be they viruses, Gram-negative or
Gram-positive bacteria, protozoans or multicellular parasites, possess
certain molecular signatures that are shared across many genera and
families of microorganisms, but are absent from vertebrate cells. Termed
pathogen-associated molecular patterns (PAMPs) by Janeway and colleagues
[7],
these motifs bind to the PRRs on innate immune cells. Some PRRs appear
to be expressed only on particular cell subsets, whereas others are
expressed more widely. Depending on the cell types that are activated
and the molecular pathways that are triggered by particular PRRs or PRR
combinations, the immune system as a whole appears to integrate the
resulting signals, which presumably allows it to then trigger the type
of immune response that has been evolutionarily selected for this
pattern or combination of molecular patterns [8]. A major recent discovery in
immunology is the identity of a family of key PRRs, the toll-like
receptors (TLRs) [9,10].
Of the ten reported mammalian TLRs, until recently only two had
identified ligands (Table
1). Immune activation by bacterial endotoxins was known to require
TLR-4 [11],
whereas recognition of bacterial lipoproteins and peptidoglycans
required TLR-2 [12,13].
In a recent issue of Nature, Hemmi et al. reported that TLR-9 is
required for immune activation by a CpG motif in a synthetic
oligodeoxynucleotide (CpG ODN) [14].
This exciting finding reinforces the role of the TLRs in innate immune
recognition of PAMPs and provides novel insights into the molecular
mechanism of action of CpG DNA. To interpret this finding fully, it is
necessary to review the prior state of knowledge concerning the
mechanism of action of CpG DNA. Cell Uptake and Endosomal
Acidification and/or Maturation The previously identified TLR
ligands are thought to induce innate immune cell signaling by
interacting with their microbial ligands (and specific host-produced
cofactors) at the cell membrane. By contrast, CpG DNA requires cell
uptake for activation to occur, suggesting that if it binds the CpG
motifs directly, TLR-9 might not work at the cell membrane [5].
Unidentified cell surface proteins bind DNA in a sequence-independent
manner, internalizing it into acidified endosomal vesicles [15].
In contrast to cellular activation by lipopolysaccharide (LPS),
endosomal acidification and/or maturation is required for CpG-induced
activation [16-18].
This internalization is accompanied by the rapid generation of reactive
oxygen species and the induction of mitogen-activated protein kinase
signaling pathways, culminating in the activation of transcription
factors such as nuclear factor (NF)-kB and activator protein 1 (AP-1) [15].
TLR-9 has a transmembrane domain; however, these findings suggest that
its role in mediating CpG-induced signaling could be fulfilled through
expression in an intracellular compartment such as the endosome, or
perhaps through recruitment of additional cofactors, which might only
occur within the endosomes or other intracellular sites. It is worth
noting that direct associations have generally not been reported between
other TLRs and their ligands. Instead, cofactors might be required, such
as CD14 and LPS-binding protein in the case of TLR-4. Potential
cofactors for TLR-9 have not yet been identified although, like most
other TLRs, signaling requires the presence of the adapter protein MyD88
[19,20]. Other Mechanisms Involved in
Recognition of CpG Motifs It is worth pointing out that
the requirement for TLR-9 in CpG-induced signaling has only been
demonstrated using a single type of CpG motif in a synthetic ODN [14].
Because bacterial DNA has a wide range of CpG motifs, and these are in a
native phosphodiester backbone instead of the phosphorothioate backbone
that is commonly used with CpG ODN, it remains possible that TLR-9 could
be only part of the story. It is therefore intriguing that an
independent group of researchers has recently suggested a role for the
DNA-dependent protein kinase, DNA-PK, in mediating immune activation by
both bacterial DNA and CpG ODN [21].
In this study, Chu et al. reported that CpG DNA sequences induce the
catalytic activity of DNA-PK in vitro and that mice genetically lacking
the catalytic subunit of DNA-PK fail to respond to CpG stimulation.
Although the activity of DNA-PK had previously been shown to be induced
in a sequence-independent fashion by the ends of double-stranded DNA
(dsDNA) and by other DNA structures, these investigators reported that
the optimal activation of DNA-PK activity required the presence of
unmethylated CpG motifs, and was reduced by DNA methylation. DNA-PK
activation had been previously associated with the cell activation
response to damaged DNA, but is now suggested by these investigators to
be more generally linked to the activation of host defenses in response
to foreign DNA. These results do not clarify
whether the sequence-specific immune activation observed is caused by
preferential binding of DNA-PK to CpG DNA, or whether a multi-protein
complex, possibly including TLR-9 or other factors, is involved.
Although the findings of Chu et al. are provocative and stimulating, the
hypothesis that CpG-induced DNA-PK catalytic activity is required for
the resulting immune stimulation is inconsistent with previous reports
that CpG DNA is a highly effective immune stimulator in severe combined
immunodeficient (SCID) mice, as these mice have no DNA-PK catalytic
function [22].
Previous studies have reported that dsDNA sequences without CpG motifs
can trigger cellular activation through pathways that appear to be
independent of those activated by CpG motifs [23].
Given the known ability of dsDNA sequences to activate DNA-PK, this
latter observation can be more easily reconciled with a role for DNA-PK
than can CpG-specific regulation. Resolution of the questions raised by
these disparate studies will require substantial additional
investigations. Different CpG DNA Molecules
Trigger Distinct Immune Effects Multiple mechanisms are likely
to be involved in the immune stimulatory effects of CpG ODN, as ODN with
different backbones and different sequence motifs can induce
dramatically different profiles and kinetics of immune activation [24-27].
For example, ODN containing phosphodiester backbones are particularly
effective at activating NK cells, whereas CpG motifs in
nuclease-resistant phosphorothioate backbones have dramatically enhanced
B-cell stimulatory properties but reduced NK-cell stimulation [5,24].
Based on these properties, we have suggested that several distinct
families of CpG ODN can be recognized (Table
2) [28].
These observations suggest that distinct CpG ODN activate multiple
independent as well as shared signaling pathways in parallel. A possible
explanation for this is that the ODN might serve as a kind of scaffold
upon which various CpG-specific or non-specific DNA-binding proteins can
form different complexes, depending on the DNA backbone, the CpG motifs
and the other DNA motifs present. In considering these various models,
it must be kept in mind that long DNA sequences are not required for
immune activation; sequences as short as six bases show some activity,
although perhaps not the full range of effects that can be observed with
longer ODN [29]. Conclusion The ability of the innate
immune system to distinguish microbial from self DNA based on
differences in the content of unmethylated CpG sequences is an elegant
demonstration of the PRR principle. To the reductionist, it is pleasing
that the immune system should have solved the problem of self-non-self
discrimination by using a single family of proteins, the TLRs, to detect
such a diverse group of microbial molecules (Table
1). At this time, the identity of the protein(s) that specifically
bind to CpG motifs remains unclear, as well as whether there are
different CpG-binding proteins for different motifs or different
cofactors that are recruited into a multi-protein complex. Regardless of
the ultimate resolution of these outstanding questions, the recent
reports of proteins required for CpG-induced immune activation provide
important new directions for research in this exciting field.
|
|
|
|
1999-2005 中国科学院上海生命科学研究院生物信息中心 |