Glycobiology Advance Access originally published online on April 5, 2006
Glycobiology 2006 16(7):103R-112R; doi:10.1093/glycob/cwj111
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
REVIEW |
T-cell recognition of glycolipids presented by CD1 proteins
Division of Rheumatology, Immunology and Allergy, Brigham and Women’s Hospital and Harvard Medical School, Smith Building Room 514, 1 Jimmy Fund Way, Boston, MA 02115
1 To whom correspondence should be addressed; e-mail: dyoung{at}rics.bwh.harvard.edu
2 To whom correspondence should be addressed; e-mail: bmoody{at}rics.bwh.harvard.edu
Received on March 3, 2006; revised on March 27, 2006; accepted on March 27, 2006
 |
Abstract |
|---|
Top
Abstract
T-cell activation by peptide...The most well-known molecular paradigm of antigen recognition by T cells involves partial digestion of proteins to generate small peptides, which bind to major histocompatibility complex (MHC) proteins. Recent studies of CD1, an MHC class I homolog encoded outside the MHC, have revealed that it presents diverse glycolipids to T cells. The molecular mechanism for lipid antigen recognition involves insertion of the lipid portion of antigens into a hydrophobic groove to form CD1–lipid complexes, which contact T-cell receptors (TCRs). Here, we examine the known antigen structures presented by CD1, the majority of which have sugar moieties that are capable of interacting with TCRs. Recognition of carbohydrate epitopes is precise, and lipid-reactive T cells alter systemic immune responses in models of infectious and autoimmune disease. These findings provide a previously unrecognized mechanism by which the cellular immune system can recognize alterations in many types of carbohydrate structures.
Key words:
glycolipid antigens
/
lipid antigens
/
NK T cells
/
T cells
/
-galactosyl ceramide
|
T-cell activation by peptide–major histocompatibility complex |
|---|
Based on studies initiated in the 1960s, the most well-known molecular mechanism of T-cell activation involves recognition of proteins encoded in the major histocompatibility complex (MHC), MHC class I and MHC class II (Mcdevitt and Benacerraf, 1969
; Zinkernagel and Doherty, 1974). MHC proteins predominantly present peptide antigens arising from partial digestion of intact proteins through the actions of endoproteases. Proteins are cleaved to peptides that have a length (8–22mer) that corresponds to the size of peptide antigen-binding grooves formed by the membrane-distal domains of MHC proteins. MHC–peptide complexes are generally formed within the endoplasmic reticulum or endosomal compartments of antigen-presenting cells (APCs) and are subsequently transported to the cell surface to allow peptide display to T cells, which are key initiators of cellular immune responses. The primary signal for T-cell activation involves direct contact of the T-cell receptor (TCR) with MHC–peptide complexes (Kappler et al., 1983; McIntyre and Allison, 1983).
On the basis of crystal structures of ternary MHC–peptide–TCR complexes and functional studies of T-cell activation, there is detailed information regarding how the variable regions of the TCR contact peptide antigens bound to MHC class I proteins, as shown in Figure 1 (Garboczi et al., 1996; Garcia et al., 1996). The components of the multimeric CD3 signaling complex, which directly contact antigens, are the TCR alpha and beta chains (TCR and TCRß). These two proteins result from the rearrangement of smaller gene segments (variable [V], diversity [D], and joining [J]) found in genomic DNA to produce highly diverse proteins, such that TCR and TCRß chains can be paired in millions of different combinations, which differ from clone to clone within the larger T-cell repertoire. The final products of these rearranged gene segments are transmembrane proteins that contain a short cytoplasmic tail and an extracellular domain composed of constant (C) and variable regions. The segments encoded by constant region genes are proximal to the transmembrane region and do not vary significantly for different ß TCRs. The more membrane-distal segments encoded by rearranged variable region genes, as well as N-region additions, differ considerably in the sequence and length of their outer loops, which form the margins of the beta sheet structure and directly contact peptide–MHC (Figure 1).
|
The trimolecular interaction among MHC–peptide–TCR is highly specific for both the MHC-encoded antigen-presenting molecule and the linear peptide antigen bound within its groove. T-cell responses are initiated through a transient adhesion of the TCR with peptide–MHC antigen complexes, forming interactions that are of much lower affinity than those of immunoglobulins with antigen. However, together with adhesion molecules, TCR–MHC interactions serve to initiate the formation of higher order aggregates that constitute the immunologic synapse between the APC and the T cell. T-cell activation is a complex phenomenon that is a function of the duration of immunological synapse formed at the interface between the T cell and the APC and an array of costimulatory receptors, adhesion molecules, and other auxiliary membrane-bound proteins (reviewed in Bromley et al., 2001). This process is, however, initiated and controlled by TCR contact with MHC–peptide and is highly specific for peptide antigens, such that a single amino acid substitution in the peptide of 9–22 amino acids can provide an all-or-nothing difference in the ability of an APC to result in T-cell activation (reviewed in Chien and Davis, 1993).
Studies of these molecular events have explained in detail how mammalian T-cell repertoires can read out the intracellular protein content of target cells by directly contacting peptide fragments displayed at the surface. T cells play a critical role in immunity to microbial infections, control of the rejection of transplanted organs, and tissue destruction during autoimmune diseases such as diabetes and multiple sclerosis. Therefore, most approaches to development of new vaccines, adjuvants (agents that improve antigen presentation), and transplantation of organs have focused on proteinaceous targets of T-cell responses and MHC-encoded proteins.
However, the glycolipid, glycoprotein, and glycan structures on mammalian cells and microbial pathogens also differ systematically, so that immune discrimination of self- and foreign structures can also be based on carbohydrate structure. It is well known that antibodies and complement proteins involved in humoral immunity specifically bind to a wide variety of molecular structures, including carbohydrates. However, until recently, it was thought that T-cell mediators of cellular immunity largely ignored carbohydrates and were targeted solely to recognizing foreign peptides. There was no a priori reason to assume that the structurally diverse TCRs that comprise mammalian T cell repertoires could not contact carbohydrates and initiate an immune response. Instead, the key question was whether cellular antigen presentation molecules could bind carbohydrate-containing antigens for display to T cells. In fact, MHC class I and II can present glycopeptides that result from post-translational glycosylation of proteins, resulting in specific activation of T cells (reviewed in Carbone and Gleeson, 1997; Haurum et al., 1999; Hudrisier et al., 1999; Lisowska, 2002). Also, MHC class II has been recently shown to bind and present zwitterionic bacterial polysaccharides to T cells (Cobb et al., 2004). However, until recently, cellular systems that predominantly focus on binding and presenting structurally diverse carbohydrates to T cells were unknown. The discovery of CD1-mediated antigen presentation has shown that human and mouse T cells can precisely recognize and respond to a variety of glycolipids found in microbial pathogens and mammalian cells.
|
CD1 antigen-presenting proteins |
|---|
In humans, CD1 genes are expressed as five protein isoforms, CD1a, CD1b, CD1c, CD1d, and CD1e. These genes were discovered and classified into distinct homology groups by Calabi and Milstein (1986
, 2000). CD1a, CD1b, and CD1c constitute the group 1 CD1 proteins, and CD1d is classified as group 2. CD1e was originally assigned to group 1 based on its amino acid sequence. It is currently, however, usually classified as the group 3 CD1 protein. This separate classification of CD1e is based on the degree of differences in amino acid homology and new studies showing that CD1e does not directly participate in display of antigens at the cell surface, but instead functions within endosomes as a lipid-transfer protein that promotes glycosidase-mediated alterations in antigen structure (Angenieux et al., 2000, 2003, 2005; de la Salle et al., 2005).
CD1 proteins are homologous to MHC class I, and like MHC class I, they consist of a heavy chain with three extracellular domains (1, 2, and 3), which bind to ß-2 microglobulin (Figure 1). The region of the antigen-presenting protein that holds and presents the antigen is formed by the 1 and 2 domains and is a hollow cavity, known as the antigen-binding groove. This structure is made up of the two alpha helices, which might be compared to the jaws of a vise, and a beta sheet floor. The solution of the three dimensional structures of CD1a, CD1b, and CD1d proteins by X-ray crystallography shows that the CD1 antigen-binding grooves are deeper than the grooves found in MHC class I or II (Zeng et al., 1997; Gadola et al., 2002; Zajonc et al., 2003). In addition, unlike MHC proteins that can utilize hydrogen bonding and charge–charge interactions to capture antigens, the inner surfaces of CD1 grooves are predominantly lined with non-polar amino acids, so that they are well adapted for binding lipid antigens through hydrophobic interactions.
|
CD1–lipid complexes |
|---|
The first evidence of CD1-mediated presentation of exogenous antigens came from the study of a human
ß T-cell lines generated by stimulation with extracts of Mycobacterium tuberculosis cell walls. One of these T-cell lines could only be activated by cells that expressed CD1b proteins and also required that the target cells be exposed to live mycobacteria or mixtures of cell wall products (Porcelli et al., 1992). Biochemical analysis of the antigens demonstrated that they were resistant to destruction by proteases and also soluble in organic solvents, suggesting that they might be lipidic in nature. Ultimately, mycolic acid (Figure 2), a large (
C80) fatty acid found in mycobacteria and related Actinomycetales species, was identified as the antigenic compound, providing the first direct evidence for T-cell recognition of lipids (Beckman et al., 1994). Subsequently, many lipid antigens have been isolated from microbial pathogens and mammalian cells, and the crystal structures of CD1–lipid complexes provide detailed insights into the molecular mechanisms of lipid antigen presentation (reviewed in Moody et al., 2005).
|
Before these studies, there was little precedent that immunoglobulins, B-cell receptors (BCRs), or TCRs could specifically bind to unsubstituted alkyl chains. Therefore, it was difficult to understand how alkyl chains, which are insoluble in biological solutions and conformationally flexible, could specifically interact with BCRs or TCRs, which usually contact antigens based on hydrogen bonding or charge–charge interactions. Crystal structures now confirm that the aliphatic hydrocarbon chains present in glycolipids and lipopeptides are generally contained within the antigen-binding groove, so that they are sequestered from aqueous solvent and form hydrophobic interactions within the globular 1–2 domain of CD1 (Figure 1) (Zeng et al., 1997; Gadola et al., 2002; Zajonc et al., 2003; Batuwangala et al., 2004; Giabbai et al., 2005; Koch et al., 2005; Zajonc, Cantu et al., 2005; Zajonc, Crispin et al., 2005; Zajonc, Maricic et al., 2005). Thus, the lipid moieties serve to anchor the antigens within the 1–2 superdomain of CD1 proteins and mainly contact CD1 proteins rather than TCRs. All mammalian and avian CD1 proteins studied to date, whether or not they have been crystallized, have non-polar amino acids in positions that are homologous to known groove-forming residues; so it is reasonable to assume that this orientation holds for most, and probably all, CD1 proteins.
The display of lipids to T cells involves a molecular mechanism whereby the polar head group, which for many antigens, includes a simple or complex carbohydrate moiety, protrudes from the groove, so that it is exposed at the surface of CD1 and is available for direct contact with the TCR. The polar head groups incorporate a variety of functional groups allowing the complete range of fundamental intermolecular interactions including dipole–dipole, hydrogen bonding, ionic, and Van der Waals interactions to come into play during interaction with the TCR, as is possible with conventional peptides bound to MHC class I and II. The orientation of functional groups such as hydroxyl, carboxylic acids, and amines in a head group that lie outside the groove would be expected to lead to a high degree of molecular specificity in the interaction with the TCR. Consistent with this idea, structure–function studies have demonstrated that the responses to glycolipid are highly specific for the glycosyl or peptidic moieties of glucose monomycolate (GMM), -linked ceramides, lipopeptides, and other antigens (Kawano et al., 1997; Moody et al., 1997, 2000, 2004).
|
Invariant natural killer T cells |
|---|
Natural Killer (NK) T cells originally derived their name from the fact that some of them express receptors such as CD161, which are encoded in the NK complex. However, unlike NK cells, which are primarily controlled by NK locus-encoded proteins and lack TCRs, the central mechanism of NK T-cell activation involves CD1d-mediated presentation of glycolipids to TCRs (Bendelac et al., 1995
; Kawano et al., 1997). The TCRs found on NK T cells are composed of nearly invariant TCR chains, V14J18 in mice and V24J18 in humans. NK T cells are a particularly abundant population of specialized T cells, which compromise between 0.1 and 10% of all T cells in humans and mice (Benlagha et al., 2000; Matsuda et al., 2000; Gumperz et al., 2002). Unlike conventional MHC-restricted T cells, which generally require more than a week to expand to large numbers and become fully activated, NK T cells can be rapidly activated within minutes to hours to produce cytokines that influence the functions of many other immune cells. In human and murine models of systemic diseases, NK T cells are activated or influence the outcomes of in vivo models of autoimmunity, infection, allergy, and infectious disease (reviewed in Kronenberg, 2005).
The first known and most potent antigens for NK T cells are -galactosyl ceramide (GalCer) and -glucosyl ceramide, which were discovered through a high-throughput screen of synthetically produced glycolipids in assays of immune-mediated regression of experimental tumors (Figure 3) (Cui et al., 1997; Kawano et al., 1997). The anomeric linkage is critical to antigenicity, as ß-galactosyl ceramide shows no ability to activate NK T cells (Figure 3). Most mammalian monosacharides have a ß-linkage to the sphingosine base, which leads to the speculation that the NK T-cell specificity for the -linkage might have physiologic significance in preventing autoreactivity to common self-glycosyl ceramides, while recognizing -linked ceramides and their natural homologs as foreign. Although -galactosyl ceramides are only known to be made by marine sponges and synthetic chemists, naturally occurring bacterial -linked galacturonide and glucuronide antigens for NK T cells have recently been identified in Sphingomonas and related bacteria (Kinjo et al., 2005; Mattner et al., 2005).
|
NK T cells can also be activated by natural lipids that have structures unrelated to -linked sphingolipids, including phosphatidylinositol-based compounds (Gumperz et al., 2002), phosphatidylethanolamine (PE) (Rauch et al., 2003), gangliosides (Wu et al., 2003), isoglobosides (Zhou et al., 2004), and leishmania lipophosphoglycans (Amprey et al., 2004). Although each of these naturally occurring compounds can activate NK T cells, they are generally less potent or activate a smaller percentage of NK T cells than seen with -galactosyl ceramides, leading to the idea that -galactosyl ceramides are a kind of lipid "superantigen" for NK T cells. Representative lipid stimulants of NK T cells are shown in Figure 3, and inspection of these structures shows that among these various lipids, only Sphingomonas-derived galacturonyl ceramides show substantial homology with -galactosyl ceramide. The conserved -anomeric linkage and the sphingolipid base structure suggest that these two antigens activate NK T cell using substantially similar molecular mechanisms.
However, it is not yet clear how the other structures, which have much larger glycans or diacylglycerol units, can activate NK T cells expressing TCRs similar or identical to those which recognize -linked ceramides. One possibility is that subsets of NK T cells have differing specificities for antigen based on subtle and as yet poorly understood differences in the Vß chains present in the "nearly invariant" TCRs (Behar and Cardell, 2000). A second possibility is that, despite the obvious differences in the overall structures among each of these antigens, some conserved feature, such as a proximal saccharide in a larger glycan unit, provides a minimal epitope to facilitate TCR binding to CD1d. A third possibility is that certain of these antigens may function to promote NK T-cell activation by a mechanism that is independent of CD1d contact with TCRs. Although it is certain that CD1d--galactosyl ceramide complexes bind to the TCR (Sidobre et al., 2002), gram-negative lipopolysaccharide can stimulate NK T cells by a mechanism that does not involve the presentation of the carbohydrate to the TCR but instead involves a TCR-independent mechanism whereby receptors on the APC trigger IL-12 release (Brigl et al., 2003). Determining whether each of these antigens activates via TCR-dependent or TCR-independent mechanisms is an active area of inquiry.
|
Diverse CD1-restricted T cells |
|---|
Although invariant NK T cells utilize TCRs with highly conserved TCR
and TCRß chains, CD1d can also present lipids to the more general class of T cells that do not express the invariant NK TCR (Cardell et al., 1995; Behar and Cardell, 2000). In fact, viewed broadly, the overall repertoire of T cells that recognize CD1a, CD1b, CD1c, and CD1d is characterized by T cells expressing diverse TCRs with no apparent conservation of TCR or TCRß chain usage (Grant et al., 1999). These diverse CD1-restricted T cells can recognize a wide range of structures including non-lipidic small molecules, unglycosylated lipids, lipopeptides, and glycolipids. In general, the responses of T-cell clones to each antigen structure are highly specific, and T-cell clones responding to one type of glycolipid antigen do not cross-react with structurally distinct antigens. This gives rise to the hypothesis that the repertoire of diverse CD1-restricted T cells is involved in a broader immunosurveillance of many classes of self and foreign lipids within target cells. Each class of antigen is considered in turn based on the presence and complexity of its glycan moieties. |
Lipid antigens |
|---|
Unglycosylated lipids may be presented by CD1 and recognized by T cells. For example, the first known antigen discovered in the CD1 system was isolated from the cell walls of M. tuberculosis (Porcelli et al., 1992
) and was subsequently identified as mycolic acid (Beckman et al., 1994). Mycolic acids are long-chain carboxylic acids with a hydroxyl group at ß-carbon and an alkyl branch at the -carbon, such that their overall lipid length can range from C30 to C80 (Figure 2). Mycolic acids are produced only by mycobacteria and related Actinomycetales species and lack any structural homologs in mammalian cells. Thus, mycolates can be considered foreign lipids, which are efficiently detected based on their long chain length (Moody et al., 2002), and may represent a biomarker for detecting the presence of mycobacterial infection (Ulrichs et al., 2003). Mycobacteria are chronic intracellular pathogens that have mechanisms for inhibiting peptide antigen display by MHC class I and II through sequestration of protein antigens within phagosomes, as well as partial deacidification of the phagosomes, which would normally digest them and produce the small peptides required for MHC class II presentation. However, mycobacterial lipids are known to escape from the phagosomes of infected cells (Russell, 2003), leading to the hypothesis that dispersion of lipids to non-infected subcellular compartments or bystander cells might represent an alternate route for activation of ß T cells.
A second class of lipids presented by the CD1 system are lipopeptides. CD1a proteins present dideoxymycobactins (Figure 2), which have a peptide-like structure consisting of both amino and organic acids. These lipopeptide antigens are likely precursors to mycobactins, a class of iron-binding compounds that are produced by infectious mycobacteria and have been shown to be required for persistent infections within cells, where iron is scarce (De Voss et al., 1999, 2000; Wagner et al., 2005). The discovery of this antigen raises interesting questions concerning the possibility that commonly occurring peptides that are post-translationally modified by prenylation or acylation might also be antigens and that CD1-restricted T cells might discriminate peptide structure (reviewed in Van Rhijn et al., 2005).
PE has been shown to be an important self-antigen for CD1d presentation to both variant ß T cells (Figure 2) and NK T cells (Figure 3) (Rauch et al., 2003; Agea et al., 2005). Because PE and certain other self-antigens in the CD1 system are always present within cells, this raises a question as to whether such lipids generate autoreactive responses. One possibility is that the cellular immune system can discriminate among individual molecular species within a larger lipid class based on subtle differences in their lipid size saturation, state, or other factors. Related to this, antigens with fine structural modifications might serve as a biomarker of some physiological state that is regulated by T cells. For example, one study of a PE-reactive T-cell clone showed that it could discriminate among various forms of PE and that a cis double bond in one of the alkyl chains confers enhanced antigenicity (Rauch et al., 2003). Furthermore, a study of polyclonal T-cell responses in human patients with seasonal allergies showed that CD1-restricted T cells demonstrated a preferential response to diacylglycerols with a diunsaturated acyl group (Agea et al., 2005). This fine specificity for the length and saturation state of the alkyl groups in phospholipids may be of physiological significance, because cypress pollen, a potent allergen, produces diunsaturated PE and phosphatidylcholine (PC), as compared with PE and PC normally found in mammalian cells (Bedinger et al., 1994; Wolters-Arts et al., 1998; Zinkl et al., 1999).
Phenyl 2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonate (PPBF) (Figure 2) is a man-made compound presented by CD1d, which represents a comparatively small antigen presented by CD1d to ß T cells (Van Rhijn et al., 2004). This compound is somewhat enigmatic as a CD1-presented antigen, because it lacks an alkyl chain, which is common to most other types of CD1-presented antigens. Among the various cellular lipids presented by the CD1 system, there is no direct evidence for the presentation of sterols, but the ability of this polyaromatic hydrocarbon to stimulate T cells raises the possibility that ringed structures, in addition to alkyl chains, can be inserted into the grooves of certain CD1 proteins. Also, PPBF shows structural similarity to sulfa antibiotics and other sulfur-containing drugs, suggesting the possibility that T cells might be involved in hypersensitivity reactions.
|
Monosaccharide glycolipid antigens |
|---|
Many antigens presented by the CD1 system contain a simple carbohydrate or monosaccharide as the polar head group. These include glucose monomycolates (GMM), phosphatidylinositols, mannosyl phosphomycoketides, galactosyl ceramides, and sulfogalactosyl ceramides (Figure 4). GMM is a monosaccharide glycolipid antigen presented by CD1b, which has a mycolic acid in ester linkage to the 6-hydroxy position on the sugar. CD1b-restricted T cells that recognize free mycolic acid do not cross-react with GMM, and GMM-specific CD1b-restricted T cells likewise fail to be activated by the free mycolic acid (Moody et al., 1997
). In addition, GMM-reactive T cells do not respond to galactose- or mannose-monomycolates, even though these compounds differ from glucose only in the stereochemistry of a single hydroxy group. This ability to precisely discriminate of carbohydrate structure is typical of CD1-restricted T cells and can be compared with MHC-restricted T cells, which can discriminate between two peptides that differ only in one amino acid side chain. On the basis of the crystal structures of CD1b-GMM complex (Gadola et al., 2002), CD1a-sulfatide complexes (Zajonc et al., 2003), and CD1d--galactosyl ceramide complexes (Koch et al., 2005; Zajonc, Cantu et al., 2005), the carbohydrate moieties, are known to protrude from the grooves of CD1a, CD1b, and CD1d proteins. Therefore, this specificity is probably mediated by differential TCR interactions with defined and recognizable substitutions on the hexose sugar.
|
Sulfatides are produced by mammalian cells and are found at particularly high levels in the central nervous system in mammals (Figure 4). Polyclonal T cells that are activated by sulfatide have been isolated from normal subjects, multiple sclerosis patients, and mice with experimental allergic encephalomyelitis, a model for multiple sclerosis in humans (Shamshiev et al., 2002; Jahng et al., 2004). In addition, treatment of mice with sulfatide was able to prevent the development of experimental autoimmune encephalomyelitis in wild-type mice but not in CD1d-deficient mice (Jahng et al., 2004). Whereas most models of CD1 autoreactivity emphasize an indirect role of CD1-restricted T cells in activating other effector cells, the ability of CD1-restricted T cells to recognize an antigen that accumulates in target tissues has raised the hypothesis that CD1-restricted T cells participate directly in antigen recognition and tissue destruction in autoimmune disease.
|
Polysaccharides presented by CD1 |
|---|
Lipoarabinomannan (LAM) and structurally related glycolipids such as phosphatidylinositol mannoside with 2, 4, or 6 mannose residues activate T cells that are restricted by CD1b or CD1d (Figure 5) (Sieling et al., 1995
; Fischer et al., 2004; de la Salle et al., 2005). These antigens, along with the mycobacterial acylated sulfotrehalose (Gilleron et al., 2004), GM1 gangliosides (Shamshiev et al., 1999), glycosyl phosphatidylinositols (Schofield et al., 1999), and lipophosphoglycans (Amprey et al., 2004), demonstrate that glycolipids containing carbohydrate structures with a wide range of size and complexity can activate T cells. One general question that arises from the discovery of antigens with polysaccharide moieties having five or more sugar units is how such large carbohydrates could be positioned between the CD1 and the TCR without sterically blocking the contact of the TCR with isoform-specific determinants on the outer surface of CD1 and thereby losing the CD1 restriction. This issue is highlighted by docking exercises in which models of CD1 TCRs can be brought in apposition to CD1 proteins bound to lipids with no carbohydrate or with a small lipopeptide (Grant et al., 2002; Zajonc, Crispin et al., 2005). These models show little excess room for the much larger carbohydrates discussed here to fit between the TCR and the CD1; yet, these molecules are able to function as CD1-restricted antigens that are recognized by T cells.
|
Recognition of large carbohydrate structures might be explained by any of several distinct mechanisms. One possibility is that the large carbohydrate moieties are cleaved enzymatically after their uptake into APCs and before their contact with TCRs. There is evidence that -mannosidase, ß-galactosidase, and CD1e function to alter the structures of ceramide and isogloboside antigens within the endosomes of APCs (Prigozy et al., 2001; Zhou et al., 2004; de la Salle et al., 2005). However, the cellular cleavage hypothesis does not explain the presentation of all large glycans. CD1b-presented ganglioside GM1 contains a pentasaccharide head group (Figure 5), and, judging from the crystal structure of CD1b bound to the structurally related glycolipid, GM2 ganglioside, it appears that nearly the complete carbohydrate moiety extends above the surface of the CD1 protein (Gadola et al., 2002). Furthermore, structure–function studies have shown that the terminal galactose and the sialic acid are necessary for T-cell activation. Last, GM1 ganglioside can be presented by recombinant CD1b proteins bound to plastic plates, thereby ruling out cellular cleavage of the antigen before TCR contact (Shamshiev et al., 1999).
A second possibility is that large, branched carbohydrates might extend laterally along the plane of CD1–TCR contact, so that only one or two of the carbohydrate units are positioned directly between the TCR and the CD1. This model predicts that the bulk of the larger carbohydrate reside is simply shifted to the side of the CD1–TCR contact site and would be largely irrelevant for recognition or might conceivably bind to the lateral surface of CD1 or TCR. Related to this, it is also possible that multiple TCR-docking orientations could play a role, as was recently proposed for unusually large peptides presented by MHC1 (Miles et al., 2005; Tynan, Borg et al., 2005; Tynan, Burrows et al., 2005). Large peptides, having a loop that extended from the MHC I groove and which might be expected to interfere with MHC presentation, were shown to interact with the TCR via two distinct orientations. In this case, different docking orientations appeared to increase the number of allowed molecular interactions while sharing a minimal required footprint and maintaining the MHC restriction via a slightly limited interaction of the TCR with the antigen-presenting protein. In addition, one of the TCR loops was also observed to move slightly aside and provide a "pocket" for the protruding bulged peptide. These mechanisms might be expected to work in the case of large glycans presented to TCRs as well.
The last possibility is that certain glycolipid stimulants of CD1-restricted T cells act indirectly to alter the production of other lipids or the synthesis of CD1 proteins themselves, so that the large lipids do not themselves bind in the groove but rather indirectly stimulate the process leading to CD1-dependent T-cell activation (Brigl et al., 2003).
T-cell activation by LAM is particularly puzzling, because its recognition might be accounted for by almost any of the mechanisms outlined above. LAM is a very large complex carbohydrate often containing in excess of 100 monosaccharide units and having a molecular weight often in excess of 15 kD, compared with most CD1-presented antigens that have a glycan unit much less than 1 kD. Although LAM might be trimmed before interacting with CD1 or TCRs, one study suggested that reducing the size of its carbohydrate moiety reduced its ability to activate T cells (Sieling et al., 1995). The major portion of the carbohydrate head might protrude laterally from the TCR–CD1 complex, but this group is so unusually large that then one might expect interference with the other proteins involved in forming the immunologic synapse. Last, a recent study has shown that some forms of LAM can trigger translation of new CD1 proteins in myeloid APCs via its ability to activate Toll-like receptor 2 (Roura-Mir et al., 2005). This observation raises the possibility that LAM activates T cells indirectly by triggering increases in the expression of CD1 proteins at the cell surface (but is not actually bound in the groove), in which case, its large size would not be relevant to potentially blocking CD1–TCR interactions.
|
Conclusion |
|---|
In summary, it has been shown that
ß T cells can specifically discriminate among the structures of a wide variety of polar head groups, including ß-hydroxy carboxylic acids, monosaccharides linked in various positions, inositol, PE, substituted polyaromatic hydrocarbons, branched carbohydrates, and peptidic structures. Like conventional T cells that recognize MHC class I and MHC class II, CD1-restricted T cells can express diverse TCRs and can recognize individual antigen structures without cross-reactivity. Therefore, the general functions of ß T cells are much broader than previously recognized, because they are capable of responding to many types of structures other than simple peptides. It is also noteworthy that CD1 can also play an important role in determining the ultimate immunologic response by selectively binding only some types of lipid tails, based on their overall length, saturation state, and substitutions. This may help provide needed discrimination between foreign antigens and endogenous glycolipids having the same carbohydrate head group.
One major question in CD1 research currently is whether the handful of glycolipids depicted in Figures 2–
5 are the only types of carbohydrate ligands recognized by T cells or is it possible that nearly any form of glycolipid that fits within CD1 grooves might initiate T-cell responses. On the basis of the combinatorial possibilities generated by polysaccharides with differing linkages, the potential array of glycolipid antigens for T cells is extremely diverse. Clearly, the targeting of cellular immune response is more versatile than previously thought, and the ability of TCRs to recognize a very wide variety of glycolipid antigens provides a new way to develop glycolipids as immunomodulatory drugs and vaccines.
|
Conflicts of interest statement |
|---|
None declared.
|
Acknowlegdments |
|---|
This work is supported by grants from the Pew Foundation, the Cancer Research Institute, the Mizutani Foundation for Glycoscience, and the NIH (AI 49313, AR48632).
|
Abbreviations |
|---|
APC, antigen-presenting cell; GMM, glucose monomycolate; LAM, lipoarabinomannan; MHC, major histocompatibility complex; NK, natural killer; PE, phosphatidylethanolamine; TCR, T-cell receptor
|
References |
|---|
Agea, E., Russano, A., Bistoni, O., Mannucci, R., Nicoletti, I., Corazzi, L., Postle, A.D., De Libero, G., Porcelli, S.A., and Spinozzi, F. (2005) Human CD1-restricted T cell recognition of lipids from pollens. J. Exp. Med., 202, 295–308.
Amprey, J.L., Im, J.S., Turco, S.J., Murray, H.W., Illarionov, P.A., Besra, G.S., Porcelli, S.A., and Spath, G.F. (2004) A subset of liver NK T cells is activated during leishmania donovani infection by CD1d-bound lipophosphoglycan. J. Exp. Med., 200, 895–904.
Angenieux, C., Fraisier, V., Maitre, B., Racine, V., van der Wel, N., Fricker, D., Proamer, F., Sachse, M., Cazenave, J.P., Peters, P. and others. (2005) The cellular pathway of CD1e in immature and maturing dendritic cells. Traffic, 6, 286–302.[CrossRef][ISI][Medline]
Angenieux, C., Salamero, J., Fricker, D., Cazenave, J.P., Goud, B., Hanau, D., and de La Salle, H. (2000) Characterization of CD1e, a third type of CD1 molecule expressed in dendritic cells. J. Biol. Chem., 275, 37757–37764.
Angenieux, C., Salamero, J., Fricker, D., Wurtz, J.M., Maitre, B., Cazenave, J.P., Hanau, D., and de la Salle, H. (2003) Common characteristics of the human and rhesus macaque CD1e molecules: conservation of biochemical and biological properties during primate evolution. Immunogenetics, 54, 842–849.[ISI][Medline]
Batuwangala, T., Shepherd, D., Gadola, S.D., Gibson, K.J., Zaccai, N.R., Fersht, A.R., Besra, G.S., Cerundolo, V., and Jones, E.Y. (2004) The crystal structure of human CD1b with a bound bacterial glycolipid. J. Immunol., 172, 2382–2388.
Beckman, E.M., Porcelli, S.A., Morita, C.T., Behar, S.M., Furlong, S.T., and Brenner, M.B. (1994) Recognition of a lipid antigen by CD1-restricted alpha beta+ T cells. Nature, 372, 691–694.[CrossRef][Medline]
Bedinger, P.A., Hardeman, K.J., and Loukides, C.A. (1994) Travelling in style: the cell biology of pollen. Trends Cell Biol., 4, 132–138.[CrossRef][Medline]
Behar, S.M. and Cardell, S. (2000) Diverse CD1d-restricted T cells: diverse phenotypes, and diverse functions. Semin. Immunol., 12, 551–560.[CrossRef][ISI][Medline]
Bendelac, A., Lantz, O., Quimby, M.E., Yewdell, J.W., Bennink, J.R., and Brutkiewicz, R.R. (1995) CD1 recognition by mouse NK1+ T lymphocytes. Science, 268, 863–865.
Benlagha, K., Weiss, A., Beavis, A., Teyton, L., and Bendelac, A. (2000) In vivo identification of glycolipid antigen-specific T cells using fluorescent CD1d tetramers. J. Exp. Med., 191, 1895–1903.[CrossRef][Medline]
Brigl, M., Bry, L., Kent, S.C., Gumperz, J.E., and Brenner, M.B. (2003) Mechanism of CD1d-restricted natural killer T cell activation during microbial infection. Nat. Immunol., 4, 1230–1237.[CrossRef][ISI][Medline]
Bromley, S.K., Burack, W.R., Johnson, K.G., Somersalo, K., Sims, T.N., Sumen, C., Davis, M.M., Shaw, A.S., Allen, P.M., and Dustin, M.L. (2001) The immunological synapse. Annu. Rev. Immunol., 19, 375–396.[CrossRef][ISI][Medline]
Calabi, F. and Milstein, C. (1986) A novel family of human major histocompatibility complex-related genes not mapping to chromosome 6. Nature, 323, 540–543.[CrossRef][Medline]
Calabi, F. and Milstein, C. (2000) The molecular biology of CD1. Semin. Immunol., 12, 503–509.[CrossRef][ISI][Medline]
Carbone, F.R. and Gleeson, P.A. (1997) Carbohydrates and antigen recognition by T cells. Glycobiology, 7, 725–730.
Cardell, S., Tangri, S., Chan, S., Kronenberg, M., Benoist, C., and Mathis, D. (1995) CD1-restricted CD4+ T cells in major histocompatibility complex class II-deficient mice. J. Exp. Med., 182, 993–1004.
Chien, Y.H. and Davis, M.M. (1993) How alpha beta T-cell receptors ‘see’ peptide/MHC complexes. Immunol. Today, 14, 597–602.[CrossRef][ISI][Medline]
Cobb, B.A., Wang, Q., Tzianabos, A.O., and Kasper, D.L. (2004) Polysaccharide processing and presentation by the MHCII pathway. Cell, 117, 677–687.[CrossRef][ISI][Medline]
Cui, J., Shin, T., Kawano, T., Sato, H., Kondo, E., Toura, I., Kaneko, Y., Koseki, H., Kanno, M., and Taniguchi, M. (1997) Requirement for Valpha14 NKT cells in IL-12-mediated rejection of tumors. Science, 278, 1623–1626.
de la Salle, H., Mariotti, S., Angenieux, C., Gilleron, M., Garcia-Alles, L.F., Malm, D., Berg, T., Paoletti, S., Maitre, B., Mourey, L. and others. (2005) Assistance of microbial glycolipid antigen processing by CD1e. Science, 310, 1321–1324.
De Voss, J.J., Rutter, K., Schroeder, B.G., and Barry, C.E., 3rd (1999) Iron acquisition and metabolism by mycobacteria. J. Bacteriol., 181, 4443–4451.
De Voss, J.J., Rutter, K., Schroeder, B.G., Su, H., Zhu, Y., and Barry, C.E., 3rd (2000) The salicylate-derived mycobactin siderophores of Mycobacterium tuberculosis are essential for growth in macrophages. Proc. Natl. Acad. Sci. U. S. A., 97, 1252–1257.
Fischer, K., Scotet, E., Niemeyer, M., Koebernick, H., Zerrahn, J., Maillet, S., Hurwitz, R., Kursar, M., Bonneville, M., Kaufmann, S.H., and Schaible, U.E. (2004) Mycobacterial phosphatidylinositol mannoside is a natural antigen for CD1d-restricted T cells. Proc. Natl. Acad. Sci. U. S. A., 101, 10685–10690.
Gadola, S.D., Zaccai, N.R., Harlos, K., Shepherd, D., Castro-Palomino, J.C., Ritter, G., Schmidt, R.R., Jones, E.Y., and Cerundolo, V. (2002) Structure of human CD1b with bound ligands at, 2.3 A, a maze for alkyl chains. Nat. Immunol., 3, 721–726. (PDB ID: 1GZP)[ISI][Medline]
Garboczi, D.N., Ghosh, P., Utz, U., Fan, Q.R., Biddison, W.E., and Wiley, D.C. (1996) Structure of the complex between human T-cell receptor, viral peptide and HLA-A2. Nature, 384, 134–141.[CrossRef][Medline]
Garcia, K.C., Degano, M., Stanfield, R.L., Brunmark, A., Jackson, M.R., Peterson, P.A., Teyton, L., and Wilson, I.A. (1996) An alphabeta T cell receptor structure at, 2.5 A and its orientation in the TCR-MHC complex. Science, 274, 209–219.