The tetraspanins are a broadly expressed superfamily of transmembrane glycoproteins with over 30 members found in humans and with homologues conserved through distantly related species, including insects, sponges, and fungi. Members of this family appear to form large integrated signaling complexes or tetraspanin-enriched microdomains (TEMs) by their association with a variety of transmembrane and intracellular signaling/cytoskeletal proteins (49). These interactions link tetraspanins to an array of physiological functions and, in consequence, to numerous endogenous pathologies, including cancer development and inherited disorders (Table 1).
Members of the tetraspanin superfamily with reported links to pathologiesa
Tetraspanins are also known to have roles in the pathology of infectious diseases such as diphtheria, malaria, and numerous viral infections (Table 1). The literature currently indicates that specific tetraspanin family members are selectively associated with specific viruses and affect multiple stages of infectivity, from initial cellular attachment to syncytium formation and viral particle release. Thus, the relationship of tetraspanins with viruses appears to be particularly complex.
Here, we will consider this data in the context of recent developments in tetraspanin biology, particularly in our understanding of the architecture and function of TEMs. With the benefit of recent insights into tetraspanin function in cell fusion events and intracellular trafficking, we discuss common features of tetraspanin/viral associations which indicate a fundamental role for TEMs in a number of viral infections. We will also consider the existing therapeutic strategies for human immunodeficiency virus (HIV), hepatitis C virus (HCV), and human T-cell lymphotropic virus type 1 (HTLV-1), focusing on the potential therapeutic value of targeting TEMs, using peptide reagents based on tetraspanin extracellular regions.
THE TETRASPANIN SUPERFAMILY
Structural features of the tetraspanin superfamily.Tetraspanins are type III membrane glycoproteins which span the plasma membrane four times, producing two extracellular loops and short intracellular regions (Fig. 1). A defining structural signature of the superfamily is the “tetraspanin fold” of the larger extracellular loop (LEL). In this region, disulfide bonding of four absolutely conserved cysteines forms a subloop structure containing a region which is hypervariable between family members and between species homologues of the same tetraspanin (Fig. 1) (81, 142). The cysteines are present in three variously conserved motifs (CysCysGly, ProXSerCys [where X = any amino acid], and GluGlyCys) (Fig. 1), wherein the flexibility and constraint imparted by the conserved Gly residue in CysCysGly and by Pro in ProXSerCys contribute to subloop formation (68). Many members of the tetraspanin family have one or two additional pairs of Cys residues in the tetraspanin fold region, potentially allowing the formation of complex subloop structures. The region of the LEL outside of this subloop shows greater structural conservation, forming three α-helices which not only form a structural platform to present the tetraspanin fold but may also contribute independently to tetraspanin function (see below) (68, 152).
The tetraspanin CD81. The primary sequence of CD81 (single-letter amino acid code) is shown schematically in the context of its transmembrane architecture. Key conserved residues in the tetraspanin superfamily include charged transmembrane residues and canonical cysteines which are disulfide bonded to form a functional subloop in the LEL. Sites of potential modification by N-glycans and palmitoylation are shown, and approximate locations of sites of molecular associations (homodimerization/heterodimerization of tetraspanins and primary interaction with nontetraspanin proteins) are highlighted. CD81 lacks classical NXS/T glycosylation sites commonly found in other tetraspanins and a broadly conserved E/Q residue in the fourth transmembrane domain. The four residues which differ between human CD81 LEL and African green monkey (T163, F186, E188, D196) CD81 LEL are shown. PKC, protein kinase C.
Conservation of the canonical LEL Cys, Gly, and other residues, including charged transmembrane amino-acids, distinguishes members of the tetraspanin superfamily proper from other “tetraspan” proteins (168). Modification of tetraspanins includes O- and N-linked glycosylation. N glycosylation sites are largely located in the LEL, and glycosylation has been shown to regulate specific tetraspanin functions (111). N glycosylation is generally assumed to occur at classical AsnX Ser/Thr sites, but the LELs of some tetraspanins, including CD9 and CD81, contain the rarer AsnXCys N-glycosylation motif (138) which, interestingly, incorporates canonical cysteines. Mutation of the only AsnXSer sites in CD9 does not prevent N glycosylation, indicating that the AsnXCys site may be utilized in some cases (G. W. Moseley, L. J. Partridge, and P. N. Monk, unpublished data). Palmitoylation of tetraspanins has also been shown to be functionally significant, as it appears to participate in the formation of heterotetraspanin associations (14) and to regulate association with lipid rafts (see below) (16).
Tetraspanin-enriched microdomains.Tetraspanins affect multiple events in vitro, including cellular signaling, migration, adhesion, fusion, cytoskeletal reorganization, and proliferation, which appear relevant to altered expression patterns seen during cellular activation, differentiation, proliferation, and malignant transformation in vivo (reviewed in references 87 and 168). Tetraspanins are expressed by all mammalian tissues, although the complement of family members expressed is tissue specific (Table 1). It is unsurprising, therefore, that tetraspanin functions have been identified in an array of cell types, including platelets, epithelial/endothelial cells, muscle cells, and photoreceptor cells and cells of the immune, central nervous, and reproductive systems (87, 99). The basis for this broad functionality appears to be the capacity of tetraspanins to form multiple intermolecular interactions with a restricted but varied complement of transmembrane and intracellular molecules (Fig. 2).
Possible architecture of the tetraspanin web. The principal interactions of TEMs are robust primary interactions which involve direct extracellular tetraspanin-protein interactions. Secondary interactions are less robust and include heterotetraspanin interactions which facilitate indirect interaction of primary tetraspanin complex proteins. Tertiary interactions are the least-robust complexes and are thought to reflect large-scale coalescence of the higher order interactions into large signaling domains. Based on the known structure of the tetraspanin-rich urothelial plaque, this order of interactions is shown as interlinked hexagons (98). Interactions within TEMs include homotypic and heterotypic tetraspanin associations and association with adhesion molecules (integrins [9]), immune cell receptors (e.g., Fc receptors and T-cell receptor (TCR), CD4/CD8, and MHC proteins [103, 167]), growth factors and their receptors (pro-HB-EGF [136], pro-transforming growth factor α [144], and epidermal growth factor receptor [110]), intracellular signaling molecules (protein kinases hck, lyn, and PKC [167, 176] and phosphatidylinositol 4-kinase [172]), and cytoskeletal components (27), as well as members of a novel transmembrane immunoglobulin superfamily (EWI-2, EWI-F) (152).
Tetraspanins have no intrinsic enzymatic activity or typical signaling motifs (with the exception of the recently identified 14-3-3 binding motif in CD81 [20]), so it has been predicted that they act primarily as novel adapter proteins, facilitating the interaction of associated molecules in tetraspanin signaling networks (tetraspanin web [135] or TEMs [49]). The evidence that physical associations of tetraspanins are functionally relevant is compelling. For example, in vitro modulation of tetraspanins on cell adhesion/migration or growth have been shown to be dependent on tetraspanin-associated integrins (76, 93, 143) and progrowth factors (75, 144). Similar observations have been made in vivo in tetraspanin-null animals (167).
Analogies have been drawn between TEMs and lipid rafts. In the latter, protein-protein interactions are facilitated by association with membrane regions enriched for cholesterol and glycosphingolipids (149). In either lipid rafts or TEMs the associated proteins are believed to form large, integrated signaling platforms. Although TEMs and lipid rafts exist as separate entities, they have been shown to interact physically and functionally (17, 27).
Architecture of TEMs.Recent studies assessing the strength/proximity of tetraspanin interactions have indicated that TEMs are constructed from a hierarchy of interactions (Fig. 2) (49, 167). Primary (or high-order) tetraspanin interactions generally appear to be mediated by the LEL region, particularly the subloop and proximal regions, a notable exception being homotypic tetraspanin associations, which are suggested to be mediated by regions of the LEL away from the subloop (Fig. 1) (152). These are the most robust of TEM interactions, being stable in relatively hydrophobic detergents such as Triton X-100 and incorporating significant proportions of the cellular complement of the associated proteins. The fact that they can be stabilized by short covalent chemical cross-linkers indicates that they involve direct extracellular interactions.
Secondary interactions involve the transmembrane and/or intracellular regions (Fig. 1) (10, 14). These interactions are less stable in high-stringency detergents but are maintained in milder conditions (Brij-96, Brij-97) and involve tetraspanin-tetraspanin associations that show some dependency on palmitoylation (14). These interactions are thought to indirectly link primary complexes and thereby facilitate their cross talk.
Tertiary, low-order interactions are stable only in low-stringency detergents (CHAPS, Brij-99). While higher order interactions are isolated as discrete entities, tertiary interactions are large complexes containing multiple tetraspanins and associated molecules. Thus, tertiary interactions are hypothesized to result from the large-scale coalescence of higher order complexes. The study of tertiary interactions is complicated by their poor stability and extensive nature, meaning that the composition of the complexes is less well defined than for higher order associations. The use of low-stringency detergents to isolate these complexes suggests the possibility of nonspecific capture of molecules in insoluble micelles. However, several tertiary interactions appear to be functionally relevant or have been confirmed by different methods, including fluorescence resonance energy transfer (154). Importantly, Claas et al. (18) have demonstrated that tertiary interactions can exist independently of lipid rafts/large detergent-insoluble fractions.
Conceptually, TEMs may form extended structures of primary tetraspanin complexes indirectly connected in large contiguous networks (Fig. 2). However, it is unlikely that TEMs are static structures, and the differing properties of primary interactions compared with those of lower order interactions may relate, at least in part, to a more transitory nature of the lower order interactions, not only to their spatial arrangement. The localization of tetraspanin complexes, subcellularly and in the plane of the membrane, is known to change in response to certain stimuli, as does the molecular composition of isolated complexes (59, 118). The dynamic subcellular localization of tetraspanins has been suggested to relate to roles in trafficking of web-associated molecules or as novel chaperone proteins in integrin maturation (9, 33, 60, 89, 134, 173). Moreover, in B cells the CD81/CD21/CD19 coreceptor complex interacts dynamically with lipid rafts, modulated by palmitoylation of CD81 (16), and regulated association of CD82, lipid rafts, and the cytoskeleton has been reported (27). Cytoskeletal association of TEMs is likely to be involved in their subcellular/membrane distribution. Thus, TEMs, like lipid rafts, are likely to be highly dynamic and regulable structures.
Tetraspanin function in vivo.A major feature of the tetraspanin web hypothesis (135) was that it provided an explanation for the similar functional effects of different members of the superfamily observed in vitro. This hypothesis has received support from studies of tetraspanin-null mice, where similar effects of knockouts of CD151, CD81, CD37, Tssc-6 have been seen in T cells (167) and in the CD9-knockout mouse, where CD81 expression in vitro can compensate for deficiencies in sperm-egg fusion (65). Moreover, tetraspanin-knockout mice are largely healthy, generally displaying only mild phenotypic abnormalities. This is surprising, as quite dramatic in vivo phenotypes result from mutation of the primary partners of some tetraspanins (69, 166). Thus, it appears that there is a certain degree of functional redundancy between tetraspanins.
While single-tetraspanin knockouts display mild phenotypes, phenotypic abnormalities are widespread and are found in T cells, platelets, and keratinocytes of CD151-null mice (77, 166), in the central nervous system, B and T cells, and retinal pigment epithelial cells of CD81-null mice (39, 100, 151), and in gametes, smooth muscle cells, and the peripheral nervous system of CD9-null mice (58, 99, 140). It would appear likely, therefore, that TEMs per se have important roles in the physiology of most, if not all, cell types. The removal of a single tetraspanin shows some redundancy, owing to the extensive nature of TEMs and the expression of close homologues. However, the clear phenotypic abnormalities observed indicate that not all functions of a single tetraspanin can be replicated by their counterparts (as is suggested by the evolutionary conservation of the members of this large family). The production of knockout mice lacking multiple tetraspanins may lead to a delineation of the overlapping and unique functions of evolutionarily close members of the tetraspanin family.
As described above, studies of the structure of TEMs have relied heavily on detergent solubility studies, which are clearly limited and relatively insensitive. Similarly, with the data currently available, the functional/physical relationship of tetraspanin associations remains inferred, as it has not formally been correlated through mutagenesis of tetraspanins. The re-expression of mutated tetraspanins in null mice may therefore permit more intricate and physiologically relevant investigations of the structure and functional architecture of TEMs.
TETRASPANINS IN VIRAL INFECTIONS
Given the broad physiological importance of tetraspanins, it is not surprising that tetraspanins have been commonly identified in cellular disorders (Table 1). There is increasing evidence for the involvement of tetraspanins in infections by various microbes, but they have been most commonly identified as playing a role in viral infections. The remainder of this review will focus on these associations.
FLAVIVIRUSES: HCV AND CD81
Interaction of CD81 and HCV-E2. The most exhaustively documented involvement of a tetraspanin in viral infection is of CD81 in HCV infection. HCV is a small, enveloped RNA virus whose genome encodes a single polyprotein that is processed by viral/cellular proteases into structural and nonstructural proteins. Structural proteins include the envelope glycoproteins E1 and E2, which mediate viral binding and entry into host cells (78).
CD81 was the first protein ligand identified for HCV, specifically for E2 protein. The interaction was identified using a cDNA library from a subclone of a human T-cell lymphoma line with high E2 binding capacity and was confirmed by the demonstration that monoclonal antibodies to CD81-LEL can inhibit the binding of recombinant E2 to Epstein-Barr virus-transformed B cells (122) and that human CD81-LEL specifically binds E2 from HCV-infected plasma. The chimpanzee is the only species other than humans that shows susceptibility to HCV (132), and vaccination with E2 glycoproteins can protect chimps from the virus. The physiological significance of CD81/E2 association was implied by the demonstration that this protection correlated with the ability of sera to block their interaction (122).
The sequence in the LEL essential for binding E2 is contained within residues 164 to 201, a region that forms the epitope for antibody 1D6, which inhibits E2 binding to CD81 (Fig. 1) (50). Inhibition of E2 binding by a small peptide analogue of CD81 LEL has highlighted the role of residues 176 to 189 (30). African green monkey (AGM; Cercopithecus aethiops) CD81 differs from human CD81 LEL by only four residues (Fig. 1), but AGM-CD81 is unable to bind to E2 and AGMs are not susceptible to HCV. Reciprocal mutations of human and AGM CD81 have shown that the key residue in this interaction is CD81 Phe186 (50). The importance of Phe186 was confirmed when tamarin monkey (Saguinus imperator) CD81 was shown to be able to bind to E2 with a higher affinity than human CD81 (1). Phe186 is conserved in tamarin CD81, and mutation of this residue can inhibit E2 binding (97). Mutational analysis has determined that the E2 binding site on CD81 also comprises residues Ile182, Asn184, and Leu162, forming a hydrophobic E2 binding surface of approximately 806 Å2 (31). It has also been suggested that the SEL domain is important during HCV infection, although this appears to relate primarily to its role in promoting surface expression of CD81 by masking an intracellular retention signal (91).
E2 binds to cells of human origin but not to cell lines derived from AGM, rat, and rabbit (36). Thus, the interaction appears to be dependent on human (h)CD81 and not to require other molecules in human cells, as rat cell lines can bind E2 protein when transfected with hCD81. Binding to this particular tetraspanin is also specific, as E2 does not bind to hCD9-, hCD63- or hCD151-transfected rat cells (36).
Based on this evidence, it has been suggested that CD81 may be a cellular receptor for HCV. However, recent data indicates that CD81 is not sufficient for HCV entry, and the relevance of the CD81-HCV E2 association for viral entry remains unclear. One study has reported that binding of HCV to three hepatocyte-derived cell lines was not blocked by anti-CD81 antibody (137). It has also been shown that hepatocarcinoma cells that have lost the expression of CD81 mRNA still show a small but detectable level of HCV binding (120). Tamarins are not susceptible to HCV infection, but tamarin CD81 binds to E2 with great affinity (1), while hCD81 transgenic mice cannot be infected with HCV (91) and certain human liver cell lines are nonpermissive to infection by HCV-HIV pseudotypes despite CD81 expression. Thus, it appears that formation of the CD81-E2 complex does not necessarily confer susceptibility to the virus. It has also been observed that different HCV strains utilize CD81 to different extents (70, 94, 131); E2 subtype 1b, for instance, shows a lower binding affinity to CD81-LEL than E2 subtype 1a. The CD81-LEL Phe186Leu mutation has been shown to abolish the binding of CD81 LEL to soluble E2 strain H but not to strain Con 1, even though this mutation has no effect on the ability of transfected HepG2 cells to support infection by pseudotype viruses expressing either Con 1 or strain H E2 (175). Another study using a more diverse range of HCV glycoproteins revealed that the CD81-LEL Asp196Glu mutation reduced the infectivity of HCV pseudotypes expressing CH35, HJC4, and C6a1 glycoproteins by 70 to 96% (94). These data suggest that all HCV pseudotypes require CD81 for infection of HepG2 cells but that the importance of the interactions between CD81 LEL and HCV glycoproteins differs between viral strains. The available evidence suggests that CD81 represents an important component in HCV cellular receptor activity but that this process is complex and likely to involve multiple receptor complexes. Further research will be required to establish the in vivo relevance of these in vitro observations.
CD81 in viral entry and release.There are other candidate receptors for HCV, namely the scavenger receptor class B type 1 (SR-B1) (139), dendritic cell-specific intracellular adhesion molecule 3-grabbing nonintegrin (DC-SIGN) (84, 126), and low-density lipoprotein receptor (LDLR) (170). One study has established a requirement for both CD81 and SR-B1 for binding and infectivity of HCV pseudoviruses, indicating a role for CD81 as a coreceptor in pseudotype infections (8). Similarly, pseudotype infections can be inhibited or enhanced by CD81 silencing or expression, respectively (175), indicating a role for CD81 in viral entry. However, CD81 alone may be insufficient for HCV entry, as the antibody-mediated internalization rate of CD81 is extremely low compared with that of other surface proteins such as CD4 and CD71 (120). Chimeras consisting of CD81 and the cytoplasmic domains of two receptors that frequently undergo endocytosis (LDLR and the transferrin receptor [TfR]) have been constructed (156); both chimeras significantly increase the endocytosis of recombinant HCV-E2 protein, anti-CD81 antibody, and E2-enveloped viral particles from the serum of HCV-infected patients. Expression of the chimeras in the Huh7 hepatoma cell line also resulted in increased viral replication, suggesting that additional factors are required for efficient endocytosis of wild-type CD81/virus complexes. Thus the mechanism of receptor-HCV complex internalization is still poorly understood. Some lines of in vitro evidence indicate the involvement of receptor complexes incorporating CD81, but further research will be required to identify additional receptor components which contribute to efficient endocytosis
As discussed above, TEMs may have a general role in regulating the trafficking of proteins associated with tetraspanins. There is precedent of a pathogenic exploitation of these processes: CD81 has been linked to entry of plasmodium into hepatocytes in the context of a parasitophorus vacuole (147). Thus, it is possible that viral proteins which interact with tetraspanins become incorporated into TEMs, facilitating their transport into or out of the cell. In the absence of human CD81, mammalian cells transfected with HCV E1-E2 cDNA almost completely retain these proteins in the endoplasmic reticulum. However, when cotransfected with hCD81, a small fraction of these proteins pass through the Golgi and are found associated with extracellular exosomes (92). The CD81 Asp196Glu mutation decreases the susceptibility to infection (see above), despite having only a minimal effect on binding of E2. It is possible that this mutation decreases infectivity by altering the intracellular trafficking characteristics of CD81 (50), lending further support to the notion that CD81-E2 interactions are involved in facilitating trafficking of HCV particles within the cell. Thus, CD81 may play a role in facilitating viral glycoprotein translocation into exosomes during viral exit and not exclusively in cell entry.
The possible involvement of CD81 in facilitating vacuole trafficking/fusion/release is consistent with hypothesized roles for other tetraspanins in physiological processes (167, 169) and similar to events described for the role of CD63 in HIV-1 infection (see below). In vitro studies showing that anti-CD81 antibody delayed fusion of a mouse myogenic cell line support a role for CD81 in membrane fusion events (155). Anti-CD9 antibody induced the same effect, suggesting that TEMs containing different tetraspanins may be involved in this event. Further evidence for TEM activity includes the demonstration that antibodies to CD9 or CD81 delayed syncytia formation in a rhabdomyosarcoma cell line (155). CD9 is also involved in sperm-egg fusion (15, 79, 99), and significantly, the impaired fusion seen in CD9-null animals can be compensated in vitro by CD81 expression (65). Also, recent data suggest that soluble LEL domains of both CD9 and CD63 can inhibit concanavalin A-induced fusion of human blood monocytes, suggesting that a range of tetraspanins may be involved in the mechanism of cell fusion (V. Parthasarathy, G. W. Moseley, A. Higginbottom, P. N. Monk, R. C. Read, and L. J. Partridge, unpublished data). Therefore, TEMs containing multiple tetraspanins are involved in membrane fusion events which are likely to relate to roles in viral infection.
Potential roles of CD81 and E2 in modulation of the immune response.Although HCV-infected patients show both cellular and humoral immune responses, over 85% are unable to eliminate the virus. In common with other leukocyte-expressed tetraspanins, CD81 has roles in modulating immune cell function (17, 28, 86). An intriguing possibility is that the interaction of CD81 with E2 could modify immune responses in HCV-infected individuals. It has been shown that cross-linking of natural killer (NK) CD81 with immobilized E2 inhibits NK cell response to interleukins-2, -12 and -15, which may affect early innate immune response to HCV infections (21, 159), while E2 cross-linking of CD81 on B- and T-cell lines alters proliferation and aggregation (36). E2 engagement of CD81 also causes the secretion of the chemokine RANTES by CD8+ T cells and the subsequent down-regulation of its receptor, CCR5, a mechanism by which HCV might interfere with lymphocyte migration (109). These events may provide strategies for immune evasion or, alternatively, may relate to the immune dysregulation effects of chronic HCV infections, including type 2 diabetes mellitus, glomerulonephritis, arthritis, and non-Hodgkin's lymphoma (157). As CD81 is widely expressed and involved in an array of functions, there are a number of potential biological effects of E2-CD81 interaction. Flint et al. (36) hypothesized that interactions between E2 on infected hepatocytes and CD81 on uninfected cells could prime the latter for infection. These are certainly intriguing possibilities, and the genuine physiological significance of the biological effects of E2 interactions with CD81 awaits further research.
RETROVIRUSES: HTLV-1, HIV, AND FIV
HTLV-1.HTLV-1 is a type C retrovirus that infects CD4+ lymphocytes in vivo (130), although CD8 T cells may also serve as a reservoir (107). This tropism contrasts with the in vitro ability of viral Env protein to bind and enter multiple cell types (13, 45, 67, 153, 158). The receptor for HTLV-1 is unknown, but a likely candidate is the human GLUT1 glucose transporter (88). Molecules involved in syncytium formation, virus adsorption, and infection include ICAM-1, ICAM-3, VCAM-1, and tetraspanins (22, 52, 123, 124). Cell-cell fusion via syncytium formation is thought to be one of the main transmission mechanisms of HTLV-1 (19). The tetraspanin CD82 was identified as an antigen recognized by several antibodies found to inhibit HTLV-1-mediated syncytium formation (38, 57). Coexpression of CD82 with HTLV-1 Env glycoprotein in COS-1 cells had an inhibitory effect on syncytium formation and HTLV-1 infectivity, although the incorporation of Env into nascent virions was unaffected (124).
CD82 is expressed on the surfaces of most types of peripheral blood mononuclear cells (PBMCs) and granulocytes, although CD16+ NK cells show very low levels of expression and CD4+ T cells express higher levels than CD8+ T cells (57). CD82 is heterogeneously glycosylated in T cells, with apparent molecular mass ranging from 35 to 50 kDa (57). Activation of T cells (e.g., using phytohemagglutinin) and/or infection with HTLV-1 results in up-regulation of CD82 expression and also increases the amount of glycosylation (38, 57). The glycosylation state of CD82 has been shown to be essential for its correct functioning (111, 112), possibly by altering the affinity of CD82 for membrane proteins such as integrins. Coimmunoprecipitation experiments show that HTLV-1 Env glycoproteins interact with highly glycosylated forms of CD82, although interactions with immature forms of the tetraspanin suggest that these interactions occur early in the secretory pathway (124). The tetraspanin CD151 has also been shown to be up-regulated in HTLV-1-infected cells (47, 48), and this has been linked to increased adhesion of infected cells to the extracellular matrix. CD82 also appears to be involved in cellular adhesion and motility events (112), and it is possible that CD151 and CD82 have common functions during HTLV-1 infection. CD82 has recently been proposed to act as an adaptor protein, linking lipid rafts and the actin cytoskeleton in T cells (26). In this role, it is likely to control the function of molecules such as ICAM-1 and β1 integrins that are directly involved in the cell-cell interactions that lead to syncytium formation (22). Thus, although CD82 does not act as a cellular receptor for HTLV-1 or affect expression levels of HTLV-1 proteins, it is likely to function in concert with CD151 and other tetraspanins in TEM-dependent cell fusion events.
HIV-1.Entry of the retrovirus HIV (genus, Lentivirus) into cells occurs in three distinct steps: the attachment of virions to cell surface CD4+ by envelope gp120 protein, the subsequent interplay of the conformationally altered gp120 with cellular CCR5 and CXR4 coreceptors, and, finally, envelope gp41 protein-mediated fusion to target cells (102). The HIV genome consists of two copies of RNA which are reverse transcribed to DNA to form the preintegration complex containing the viral matrix (MA), integrase, and Vpr proteins as well as several host proteins, which is imported into the nucleus and incorporated into the host chromosomes, where it can become latent for 3 to 10 years (44). The subtypes of HIV, HIV-1 and HIV-2, have the same modes of transmission and are associated with similar opportunistic infections, but immunodeficiency seems to develop more slowly with HIV-2 (90, 148).
A role for the tetraspanin CD63 in HIV-1 infection was first suggested when it was shown to be up-regulated from intracellular vesicles to the surfaces of HIV-1-infected cells and selectively incorporated into budding structures and newly synthesized virus particles (95, 96). Consistent with a role in viral release, vesicle fractions in infected H9 T cells contain increased levels of CD63 compared with noninfected cells (41).
Macrophages are sources of virus reservoirs during HIV-1 infection: previous work has described organelles within macrophages as preferential sites for virus accumulation in macrophages (114). Expression analysis revealed high levels of CD63 and major histocompatibility complex type II (MHC II) and low levels of Lamp 1 in these organelles, resembling the compartments where MHC II molecules undergo the final stages of maturation. CD63 is also enriched on the surface of multivesicular bodies, and it is presumed that HIV-1 particles mimic the mechanisms utilized in the production of natural internal vesicles (128). Thus it is possible that CD63, like CD81 and CD82 (see above), may be involved in viral trafficking and/or association with intracellular vesicles for fusion and release.
It has recently been shown that HIV-1 infection is inhibited by anti-CD63 antibodies, but not by antibodies to CD9 or CD81. This was specific for macrophages and HIV-1 strains which use the CCR5 coreceptor, as X4 virus and T-cell infection were unaffected by anti-CD63 antibodies (161). Preliminary data also shows that recombinant tetraspanin-LEL proteins can also specifically inhibit macrophage infection but, in contrast to antibody inhibitors, can also inhibit CXCR4-tropic virus; significantly, the effect is seen with a range of tetraspanin-LELs (including CD9 and CD81) (S. H. Ho, F. Martin, G. W. Moseley, L. J. Partridge, C. Cheng-Mayer, and P. N. Monk, unpublished data). These experiments indicate a role for multiple tetraspanins, probably organized in TEMs, in the infection of macrophages by HIV. The differences observed between the studies using tetraspanin LELs and antibodies may be due to a number of factors, as neither approach is physiological. However, tetraspanin LELs are small proteins and are not prone to the same artifacts as antibodies such as Fc receptor coengagement, known to suppress HIV replication in macrophages (119). Alternatively, the differing efficacy of antibodies may relate to epitope-specific events, as described for anti-CD82 in syncytium formation (57). It is also possible that anti-CD63 antibodies bind directly to the viral envelope, known to contain higher levels of CD63 than other tetraspanins, thus inhibiting infection by steric hindrance (117). However, the apparent involvement of CD9 and CD81 in HIV-1 infection implies that TEMs are involved in HIV-1 infection and that this role may involve vacuolar targeting and membrane fusion events in viral release.
FIV.Feline immunodeficiency virus (FIV) infection also appears to involve tetraspanins, in this case CD9 (164). In 104-C1 PBMC cells transfected to express feline (f)CD9, the anti-CD9 antibody vpg15 was shown to delay reversibly, but not completely block, reverse transcriptase activity following infection (29). As the inhibitory effect was enhanced upon addition of antibody postinfection, it appeared that CD9 can affect events after cellular entry. Treatment of chronically infected cells with vpg15 reduced the numbers of budding particles at the plasma membrane and inhibited FIV spread in cell culture, so it appears likely that viral assembly and/or release (29) is inhibited by anti-CD9 antibody, indicating a role for fCD9 similar to that of other tetraspanins in viral infections. FIV can infect CD9-negative 3201 cells (55), implying that CD9 is not essential for FIV infection or at least that other tetraspanins can take over this role. However, ectopic expression of CD9 in 3201 cells enhances FIV infectivity, suggesting that CD9 plays a direct role in infection rather than an indirect role via cell signaling pathways affected by anti-CD9 antibody ligation (163).
PARAMYXOVIRUS
Canine distemper virus (CDV) (a morbillivirus of the Paramyxoviridae family) is a negative-stranded RNA virus, similar to human measles virus, for which the host cell surface receptor(s) are currently not identified (46). As for FIV, an anti-CD9 antibody has been shown to inhibit viral infection (83) and transfection of CD9 into cell lines increases viral production, leading to a greater number of infectious centers and larger syncytia, consistent with the effects of CD9 expression on syncytium formation by a rhabdomyosarcoma cell line (155). CD9 does not appear to be a receptor for CDV (83), and anti-CD9 antibody does not block CDV attachment to cells (141). The involvement in subsequent events (83), including membrane fusion/syncytium formation, is comparable to that of other tetraspanins in human viral diseases and CD9 in FIV.
TETRASPANINS AS POTENTIAL THERAPEUTIC TARGETS FOR VIRUSES THAT INFECT HUMANS
HCV infects 170 million people worldwide, causing liver diseases that include hepatitis, cirrhosis, hepatocellular carcinoma, and disorders relating to immune dysregulation (125, 157). HIV/AIDS is responsible for more than 25 million deaths since 1984, with an estimated 37.8 million individuals currently infected (Joint United Nations Program on HIV/AIDS, UNAIDS, 2004). HTLV-1 presently infects 20 million individuals worldwide (32) and is associated with adult T-cell-leukemia/lymphoma and myelopathy/tropical spatic paraparesis. Treatments for these viruses include antiviral drugs such as interferons, viral enzyme inhibitors, and viral-cell receptor interaction inhibitors (24, 25, 35, 62, 71, 100, 157), which can substantially lower the viral load (115) but cannot eliminate the virus. Further problems associated with such treatments include toxicity and side effects, high cost, development of viral resistance, and low efficacy owing to viral genotype specificity/patient variables (11, 43, 61, 101, 108, 121, 145, 157, 171). Efforts to develop conventional vaccines for HIV have been hampered by potential pathogenicity and poor capacity to elicit protective immune responses (6, 54, 80, 162). Several new vaccination approaches for HIV (e.g., DNA and viral vectors) are in clinical trials (3, 4, 7, 23, 82, 105, 133, 146). However, in common with conventional vaccines, most of these strategies have been shown to be ineffective due either to their poor capacity to elicit immune protection (40, 85) or to the various mechanisms of HIV immune evasion, such as high rate of mutations and host genome insertion (34, 53, 72, 73, 106, 116, 127).
As described above, various groups have identified tetraspanins as the targets of antibodies which inhibit the infectivity of a range of viruses but have generally identified a specific tetraspanin in each of these studies. However, it appears that at least two tetraspanins may be involved in HTLV-1 infection and at least four in HIV-1 infection. Thus, we hypothesize that (i) several tetraspanins will be involved in some or all of the viral life cycles here mentioned, and (ii) that tetraspanins are likely to be involved in many other viral infections as yet undiscovered. While definition of the precise roles of TEMs in viral infection awaits further research, it has already been shown that peptides based on tetraspanin-LELs can inhibit tetraspanin functions such as sperm/oocyte fusion (51). A peptide inhibitor of only 14 residues, designed to mimic the region 176 to 189 of CD81 has been shown to inhibit the CD81-E2 interaction (30). With a greater understanding of the association of TEMs with viral proteins, particularly the primary ligand interactions, it should become possible to selectively target TEMs for therapy. The E2 binding site on CD81 has been mapped to specific amino acids in the LEL subloop (31, 160). Based on these findings, several bi-imidazole-based compounds that can also inhibit E2 binding to CD81 have been identified (160). These resemble the solvent-exposed face of the helix of CD81 LEL that contains Phe186, thus mimicking the E2 binding site on CD81. Two of the compounds have 50% inhibitory concentration values below 100 μM and represent an important proof of the concept that tetraspanin-based drugs can disrupt biological function. Thus, targeting TEMs may provide a novel strategy to inhibit critical processes of viral infection.
CONCLUDING REMARKS
The precise nature of the relationship of tetraspanins with viruses is presently poorly defined and is complicated by the apparently disparate functions of specific tetraspanins in infections by different viruses (summarized in Fig. 3). However, much of the data describing this relationship has been derived from in vitro studies. Thus, it is unclear whether modulation of immune cell line function by LEL-CD81 ligation, for example, has any genuine significance in vivo. Similarly, while CD81 is known to interact with HCV coat proteins in vitro, the stage of infection at which this interaction occurs in vivo is unknown and need not necessarily occur during cell entry. Indeed the interaction of CD82 with HTLV-1 proteins occurs postinfection, while CD9 and CD63 appear to affect FIV/CDV and HIV-1 late in infection. Of particular significance is the recent demonstration that soluble LEL proteins from several different tetraspanins can inhibit HIV-1 infection and the overlapping effects of certain antitetraspanin antibodies in syncytium formation by rhabdomyosarcoma cells. This suggests that there is functional overlap between different tetraspanins involved in viral infection processes, which implies the involvement of TEMs.
Model for the role of TEMs in viral infection. Current evidence (represented in inset) indicates that viral proteins may associate with TEMs by forming direct primary interactions with a tetraspanin (black arrows) which, via secondary/tertiary interactions of tetraspanins (grey arrow), incorporate the viral protein into the tetraspanin web, permitting indirect interaction with tetraspanin-associated molecules (dotted arrow). Various data indicate that, through these interactions, TEMs may be involved in initial viral attachment (no. 1) and entry (no. 2) into the cell. Some data also indicate that viral interaction with TEMs may couple to intracellular signaling pathways, priming cells for infection or altering immune cell functions (no. 3). TEMs may interact with new viral proteins after infection/transcription in the biosynthesis/secretion pathway (no. 4). Multiple lines of evidence also indicate that TEM involvement in membrane fusion events may be exploited for viral exit from cells whether via direct release by secretory vesicles or in exosomes (no. 5 and 6) or via cell-cell fusion to form syncytia (no. 7). In the latter, tetraspanin-associated cell adhesion molecules (e.g., ICAM, integrins) are likely to mediate intercellular interactions in the fusion process.
Thus, rather than being passive targets for opportunistic viral attachment to cells, tetraspanins are likely to play a more fundamental role in viral infectivity through the involvement of TEMs in viral protein maturation/trafficking, in cell-cell fusion events or in other membrane fusion events involved in viral release, processes which have clear parallels in normal TEM physiology. In this context, the selective primary interaction of viral protein with a specific tetraspanin (e.g., CD81 with HCV E2) to connect it via secondary/tertiary interactions to tetraspanin complexes involved in these processes would be typical of the incorporation of nontetraspanin proteins into the TEMs (Fig. 3). Targeting TEMs for viral therapies may therefore target a principal mechanism in the infectious cycles of HIV-1, HTLV-1, and HCV which viral mutation would not readily circumvent, providing new antiviral therapies so urgently required by an ever-increasing percentage of the world's population.
- Copyright © 2005 American Society for Microbiology
REFERENCES
- 1.↵
- 2.
- 3.↵
- 4.↵
- 5.
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.
- 64.
- 65.↵
- 66.
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.↵
- 122.↵
- 123.↵
- 124.↵
- 125.↵
- 126.↵
- 127.↵
- 128.↵
- 129.
- 130.↵
- 131.↵
- 132.↵
- 133.↵
- 134.↵
- 135.↵
- 136.↵
- 137.↵
- 138.↵
- 139.↵
- 140.↵
- 141.↵
- 142.↵
- 143.↵
- 144.↵
- 145.↵
- 146.↵
- 147.↵
- 148.↵
- 149.↵
- 150.
- 151.↵
- 152.↵
- 153.↵
- 154.↵
- 155.↵
- 156.↵
- 157.↵
- 158.↵
- 159.↵
- 160.↵
- 161.↵
- 162.↵
- 163.↵
- 164.↵
- 165.
- 166.↵
- 167.↵
- 168.↵
- 169.↵
- 170.↵
- 171.↵
- 172.↵
- 173.↵
- 174.
- 175.↵
- 176.↵