Journal of General Virology (2002),
83, 1809-1829.
© 2002 Society for General Microbiology
Cell surface receptors, virus entry and tropism of primate lentiviruses
Paul R. Clapham1 and
Áine McKnight2
Center for AIDS Research, Program in Molecular Medicine, Department of Molecular Genetics and Microbiology, University of Massachusetts Medical School, Biotech II, 373, Plantation Street, Worcester. MA 01605, USA1
The Wohl Virion Center, Department of Immunology and Molecular Pathology, The Windeyer Institute for Medical Sciences, University College London, 46 Cleveland Street, London W1P 6DB, UK2
Author for correspondence: Paul Clapham. Fax +1 508 856 4075. e-mail paul.clapham{at}umassmed.edu
 |
Abstract
|
|---|
Human immunodeficiency virus (HIV) exploits cell surface receptors to attach to and gain entry into cells. The HIV envelope spike glycoprotein on the surface of virus particles binds both CD4 and a seven-transmembrane coreceptor. These interactions trigger conformational changes in the envelope spike that induce fusion of viral and cellular membranes and entry of the viral core into the cell cytoplasm. Other cell surface receptors also interact with gp120 and aid attachment of virus particles. This review describes these receptors, their roles in HIV entry and their influence on cell tropism.
|
Introduction
|
|---|
Human immunodeficiency virus (HIV) and the simian immunodeficiency virus (SIV) counterparts are retroviruses that belong to the family of lentiviruses that causes degenerative diseases. A hallmark of lentivirus disease is the failure of the host’s immunity to prevent continuous virus replication. Lentiviruses include maedi–visna virus of sheep, equine infectious anaemia virus and feline immunodeficiency virus (FIV) as well as HIV and SIV. All these lentiviruses establish persistent infections in cells of the monocyte/macrophage lineage that are associated with pathology in the brain (dementia) and joints (arthritis). FIV, HIV and SIV also infect T-lymphocytes and cause severe immune deficiency. For HIV, the cell surface receptors exploited for entry into cells are major determinants of cell tropism and pathology.
Lentiviruses are enveloped viruses that acquire a lipid membrane while budding from the membranes of an infected cell. After budding, the Gag proteins of the virus core are processed by the virion protease to form a mature infectious particle. The resulting cone-shaped core contains the viral genomic RNA that is delivered into a new cell to start a fresh cycle of replication. The first events that initiate infection are: (1) attachment of the virus particle to the cell surface and (2) fusion of the virus and cell membranes to deliver the virion core into the cell cytoplasm (Fig. 1A
). For HIV, attachment and fusion are mediated by the interaction of virion glycoprotein spikes with cell surface receptors.
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 3. The role of gp41 in the formation of the fusion pore. (A) Interaction of gp120 with CD4 and a coreceptor exposes the fusion domain (blue triangle). (B) Complexing of the leucine zipper-like region (red) and  -helix (blue) positions the virion and cell membranes close together. Several gp41 molecules are probably needed to form a fusion pore. (C) The outer membranes fuse together at the hemifusion stage. (D) The hemifusion stage is usually short-lived and fusion of the internal membranes follows rapidly. The core inside the virus particle is now exposed to the cell cytoplasm and further uncoating events are needed to allow disassociation from virion membrane and transport of the preintegration complex through the cytoplasm.
|
|
The glycoproteins of many enveloped viruses carry a hydrophobic fusion domain, which is held inside the native spike glycoprotein for protection from hydrophilic environments. Fusion is triggered by conformational changes in the envelope spikes directing the fusion domain to embed in the cell membrane. For example, following attachment to cell surfaces, influenza virus particles are internalized into endosomes. The low pH (<5·5) inside endosomes triggers reorientation of the influenza virus spike glycoproteins (haemagglutinins, HAs), exposing the fusion domain to the endosomal membrane and initiating fusion (Fig. 1B
) (reviewed by Skehel et al., 1995 ). HIV uses a different strategy to trigger the conformational changes needed for infection. This process is independent of low pH, being driven by interactions with specific receptors on the cell surface. Two receptors are usually essential for HIV entry: CD4 and a seven-transmembrane (7TM) coreceptor (Fig. 2). This review examines these and other cellular receptors that HIV exploits to attach to and gain entry into cells.
|
HIV envelope structure and attachment of HIV particles to cell surfaces
|
|---|
Attachment of HIV particles to cell surfaces is mainly attributed to the interaction of the spike glycoproteins with receptors. The envelope glycoproteins of HIV are made as a precursor molecule, gp160, which is cleaved in the Golgi apparatus by a cellular protease (for example, furin and related proteases) (Hallenberger et al., 1997 ) into a surface (SU) gp120 molecule, noncovalently attached to a transmembrane (TM) gp41. Each spike on a virus particle is made up of three gp120 and three gp41 molecules, held together as a trimer (Weiss et al., 1990 ) by determinants in gp41 (Earl et al., 1990 ). The SU glycoprotein gp120 is particularly exposed to host antibodies and contains five variable loops (V1 to V5) which may help replicating viruses to escape antibody-mediated neutralization. These variable loops are interspersed by more conserved regions. In comparison, gp41 is relatively conserved.
CD4 is the major receptor for HIV and SIV (Sattentau et al., 1988 ). Each monomer of gp120 contains a binding site for CD4. Engagement of one CD4 molecule by a single gp120 in the trimeric spike is sufficient to induce conformational changes in all three glycoprotein monomers of the trimer (Salzwedel & Berger, 2000 ). HIV type 1 (HIV-1) strains adapted for replication in CD4+ T-cell lines (T-cell line-adapted, TCLA) have an affinity for CD4 up to 50 times higher than envelopes of primary isolates (Moore et al., 1992 ). Despite CD4 affinities that are often low, primary viruses are still dependent on CD4 for fusion; however, the importance of CD4 for attachment to cells is questionable. Furthermore, while some cell types targeted by HIV in vivo express high levels of CD4 (for example, T-cells), others, including macrophages and dendritic cells (DCs), express barely detectable amounts. In these situations, HIV may attach to cells by CD4-independent interactions involving sugar groups on the envelope glycoprotein with other sugars or lectin-like domains on cell surface receptors, such as the mannose-specific macrophage endocytosis receptor (Larkin et al., 1989 ). Table 1 lists cell surface molecules identified to interact with gp120. A cell surface protein (DC-SIGN) identified by its capacity to bind gp120 with high affinity (Curtis et al., 1992 ) is expressed on certain DC populations (Geijtenbeek et al., 2000 ). A closely related receptor (DC-SIGNR) expressed on endothelial cells binds HIV in a similar manner (Pohlmann et al., 2001 ). Gp120 also binds the glycolipid galactocerebroside (Gal-C) and its sulphated derivative, sulphatide (Fantini et al., 1993 ; Harouse et al., 1991 ). These molecules are expressed on neurons and glia in the brain (Harouse et al., 1991 ), colon epithelial cell lines (Fantini et al., 1993 ) and, importantly, on macrophages (Seddiki et al., 1994 ). Gal-C binds gp120 with a high affinity, similar to the binding affinity of monomeric gp120 for CD4. Gal-C supports suboptimal entry of particular HIV-1 strains without CD4, although infection requires a coreceptor (Delezay et al., 1997 ). Mondor et al. (1998) have shown that HIV virions attach to HeLa cells via an interaction between gp120 and the glycosaminoglycan heparan sulphate. This interaction can be demonstrated for X4 and R5X4 viruses but is less efficient for R5 virus envelopes, since it is mediated mainly by positively charged V3 loops interacting with negative sulphate groups on glycosaminoglycans (Moulard et al., 2000 ).
Besides direct interactions of the envelope glycoprotein with cell surface receptors, interactions also occur between cell-derived molecules incorporated onto virions and their ligands. Such interactions enhance the overall efficiency of virus entry. Examples include the integrin ICAM-1 (intercellular adhesion molecule-1), which assembles onto HIV particles (Paquette et al., 1998 ) and enhances attachment to cells expressing its ligand, LFA-1 (lymphocyte function-associated antigen-1) (Fortin et al., 1999 ).
Although HIV may attach to cells via a number of distinct interactions, fusion will not occur until sufficient CD4 and coreceptor molecules are recruited to trigger formation of a fusion pore. Thus, direct and early interactions with CD4 are likely to be the most efficient infection process with the fastest kinetics.
|
Fusion mechanism
|
|---|
HIV particles fuse with cell surface membranes. Envelope/coreceptor engagement triggers gp41 rearrangement and exposure of the fusion domain leading to fusion. The crystal structures of the extracellular gp41 domains have been determined (Chan et al., 1997 ; Weissenhorn et al., 1997 ). These gp41 structures represent the triggered fusion-activated form and have a rod-like structure with a bundle of six helices at their core. Similar TM structures have been reported for unrelated enveloped viruses, including influenza virus and Ebola virus, consistent with conserved mechanisms for fusion (Dutch et al., 2000 ). Three leucine zipper domains (each from one oligomer of the trimer) form a coiled-coil structure that extends up towards the N-terminal fusion domain. Three -helices proximal (but external) to the virion membrane interact with outer grooves of the coiled coil in an anti-parallel manner to form the six-helix bundle. The conformational events and intermediate structures that lead to the formation of the six-helix bundle are not clear. An extended coiled coil may form first, protruding the fusion domain towards the cell membrane. Complexing with the membrane proximal -helices would then position the two membranes close together (Fig. 3).
Several glycoprotein spikes form a ring and cooperate in the induction of a fusion pore. For influenza virus, three to six HA trimers are required (Danieli et al., 1996 ). Fusion proceeds from initial curvature of target and virion membranes, following insertion of the fusion peptide, to a short-lived hemifusion stage where only the outer lipid bilayers are fused and, finally, form a flickering pore that stabilizes and expands. Low pH-treated influenza virions usually carry HA1s that appear to be completely disorganized. This disordered state may prevent HA1 from physically hindering the two membranes approaching each other. The fate of gp120, CD4 and coreceptors following gp41 activation is not known. Neither is it clear if gp120 or the receptors play a role in establishing the ring of envelope spikes around the fusion pore or have roles in uncoating events immediately after fusion, as suggested by Chackerian et al. (1997) .
|
The major receptor, CD4
|
|---|
Early evidence that CD4 was the receptor for HIV-1 included the following observations: (1) infection was inhibited by monoclonal antibodies (MAbs) to CD4; (2) expression of CD4 on resistant cells conferred sensitivity to infection; and (3) CD4 could be coprecipitated with the HIV envelope.
CD4 is a ligand for MHC class II molecules interacting with the 
2 subunit. It is expressed predominantly on T-helper cells acting as an accessory receptor in the cellular immune response. Its role is to increase the avidity between helper T-cells and MHC class II+ antigen-presenting cells, forming part of a ternary complex with the T-cell receptor (TCR) in antigen recognition. CD4–MHC class II interactions also have roles in cell adhesion, enabling other receptor/ligands to contact.
CD4 is a member of the immunoglobulin superfamily and has four extracellular immunoglobulin-like domains (D1 to membrane proximal D4), a TM region and the cytoplasmic tail that associates with the kinase p56lck. The extracellular domain of CD4 extends 120 
, the distance required to span the length of the TCR and interact with MHC class II molecules (Janeway, 1992 ). CD4 is also expressed on cell types that do not express a TCR, such as monocytes, macrophages and DCs, all of which can be infected by HIV. The role of CD4 on these cells is not well understood; however, CD4 is a receptor for IL-16, a cytokine with chemoattractant activity for CD4+ lymphocytes, monocytes and eosinophils (Cruikshank et al., 1998 ).
Binding site on CD4 for gp120
The interaction between CD4 and gp120 is conserved among all primate lentiviruses. The site on CD4 that interacts with gp120 has been mapped by mutagenesis and the structure of a gp120–CD4 complex reported (Kwong et al., 1998 ). The CD4 site that contacts gp120 forms a charged ridge on the N- terminal domain furthest from the cell membrane. This site is part of the CDR2-like region that corresponds to the second of three complementarity-determining regions (equivalent to the antigen-binding site on antibody molecules). F43 and the positive R59 residues in this region make multiple contacts with gp120 residues, including negatively charged D368, E370 and hydrophobic W427. The F43 side chain penetrates a hole on gp120 (Fig. 4A). About 63% of gp120–CD4 contacts are made by CD4 residues 40–48 (Kwong et al., 1998 ). The contact gp120 residues are derived from several discontinuous sequences and include conserved amino acids where the backbone of the polypeptide chain, rather than amino acid side chains, contacts CD4. Binding of gp120 to CD4 causes rearrangement of the gp120 core (Myszka et al., 2000 ) and movement of variable loops resulting in formation and/or exposure of a site that binds a coreceptor. The crystals determined by Kwong et al. (1998) reveal the structure of gp120 and CD4 complexed together; however, the native gp120 structure prior to CD4 binding remains less clear.
View larger version (56K):
[in this window]
[in a new window]
|
Fig. 4. Receptor-binding sites on gp120. (A) Critical amino acids involved in the interaction of gp120 and CD4. Amino acids from five different domains of gp120 make up the conformational binding site that interacts with CD4. Red bars mark critical contacts between F43 on CD4 and W427, E370 on gp120 as well as between R59 on CD4 and D368 on gp120. (B) Structure of gp120 showing bridging sheet situated between V1/V2 and V3 and ‘below’ the CD4-binding site. The coreceptor binding site is thought to consist mainly of sites on the bridging sheet and V3 loop.
|
|
CD4 itself undergoes conformational changes on binding gp120. The crystal structure of gp120–CD4 complexes do not reveal any rearrangements in domain D1D2 compared to uncomplexed CD4 (Kwong et al., 1998 ). However, CD4 has flexible regions between domains D2 and D3, as well as between domain D4 and the membrane. Such flexibility may be required for CD4 to ‘approach’ gp120 laterally and to orientate the coreceptor-binding site towards the cell surface (Fig. 2). Yachou & Sekaly (1999) showed that gp120 binding resulted in the loss of epitopes on domains D3 and D4 of CD4, consistent with conformational alterations distant from the gp120-binding site. There is evidence that both flexible domains are important for HIV infection. First, MAbs to the D2–D3 hinge block HIV infection but not gp120 binding to CD4 (Healey et al., 1990 ) and, second, deletions in the D4 membrane-flexible region delay infection and reduce V3 loop exposure, suggesting that conformational changes in gp120 are inefficient without CD4 flexibility (Moir et al., 1996 ).
Binding site on CD4 for MHC class II molecules
The sites on CD4 that interact with MHC class II molecules are complex and encompass a larger surface area compared to the gp120-binding site (reviewed by Ravichandran et al., 1996 ). Amino acids clustered along one side of CD4 domain D1 in CDR1 and CDR3 as well as domain D2 residues are involved in interaction with MHC class II molecules (Moebius et al., 1993 ). Evidence indicates that the CDR2 region on the opposite face of CD4 also binds MHC class II (Huang et al., 1997 ; Moebius et al., 1993 ) and may be involved in interactions that allow hetero-oligomers to form for augmentation of T-cell activation signals (Huang et al., 1997 ). Additional evidence suggests a direct association between CD4 and the TCR via the membrane proximal D3 and D4 domains of CD4 (Vignali et al., 1996 ).
IL-16
IL-16 was reported to form homodimers that then interact with the membrane-proximal D4 domain on dimeric CD4 (Liu et al., 1999b ). The ability of peptides derived from the D4 domain of CD4 to block IL-16-induced activation is consistent with domain D4 as the IL-16-binding site (Liu et al., 1999b ).
Human herpes virus-7 (HHV-7) exploits CD4 as a receptor
HHV-7 uses CD4 as a receptor for entry into cells and is inhibited by soluble gp120 as well as by CD4 MAbs (Lusso et al., 1994 ). However, CD4 transfection did not confer HHV-7 sensitivity for all cell lines (Yasukawa et al., 1997 ), perhaps indicating that HHV-7 requires a coreceptor for infection. If a coreceptor is required, it is not one of the major HIV coreceptors, CCR5 or CXCR4. Neither is needed for HHV-7 infection (Yasukawa et al., 1999 ).
|
Coreceptors
|
|---|
Soon after CD4 was shown to be the main receptor for HIV and SIV, evidence started to accumulate indicating that CD4 alone was not sufficient for HIV to fuse with the cells. Maddon et al. (1986) showed that CD4 expressed on mouse cells allowed virus to bind but did not confer virus entry. Data from this study (Maddon et al., 1986 ) suggested that mouse cells lacked a cofactor or coreceptor needed in addition to CD4 to trigger HIV entry. Of particular importance were the observations that HIV-1 isolates fell into two distinct groups depending on their biological properties. Asjo et al. (1986) described the two groups as slow-low and rapid-high depending on their replication rates in PBMCs. Related reports described the two virus groups as nonsyncytium-inducing (NSI) or syncytium-inducing (SI) (Tersmette et al., 1988 ) as well as macrophage-tropic (or M-tropic) and T-cell tropic (or T-cell line tropic, T-tropic) (Gartner et al., 1986 ). The differences between the two types of isolate were determined by gp120 env sequences and were located predominantly in the V3 loop (Hwang et al., 1991 ). This evidence was consistent with the two virus groups requiring different cell surface coreceptors for entry.
E. A. Berger and colleagues cloned the first HIV coreceptor, CXCR4 (termed fusin) (Feng et al., 1996 ). Coexpression of CXCR4 with CD4 on mouse cells conferred fusion by SI or T-tropic (but not NSI/M-tropic) HIV-1 strains. Several groups reported CCR5 as the coreceptor for NSI viruses (Alkhatib et al., 1996 ; Deng et al., 1996 ; Dragic et al., 1996 ). CCR5 and CXCR4 are the major HIV-1 coreceptors and all strains can use one (R5 and X4 viruses) or both (R5X4 viruses) to enter CD4+ cells. R5 viruses are predominantly transmitted and persist throughout infection. Viruses that exploit CXCR4 emerge late in disease and can be isolated from up to 50% of AIDS cases. Both CCR5 and CXCR4 are members of the 7TM chemokine receptor family. More than a dozen other 7TM receptors have been shown to act as coreceptors on CD4+ cell lines for particular HIV-1 strains. These coreceptors are also chemokine receptors or are closely related orphan receptors. Currently, there is little evidence to suggest that coreceptors other than CCR5 and CXCR4 are used significantly in vivo.
The pattern of coreceptors used by SIV and HIV-2 is different from HIV-1, as expected (Clapham et al., 1991 ). SIV uses CCR5 but CXCR4 is rarely used (Meister et al., 2001 ). SIV strains predominantly exploited in the rhesus macaque (MAC) or cynomolgus animal models, such as African green monkeys (AGM), all use CCR5, as do primary isolates from sooty mangabey monkeys (SMM) (Chen et al., 1998a ). Several SIV strains have been shown to use CXCR4 on cell lines in vitro. In most cases, infection was insubstantial and its significance unclear (Owen et al., 2000 ; Schols & De Clercq, 1998 ). A switch from CCR5- to CXCR4-use associated with disease progression has not been reported in SIV animal models. T-tropic and M-tropic SIVMAC strains, however, are apparent. These variants differ in the way that CCR5 is exploited as a coreceptor (Edinger et al., 1997 ) and in their capacity to exploit low levels of surface CD4 (Bannert et al., 2000 ). SIVMAC, SIVSMM and SIVAGM strains often use other coreceptors in addition to CCR5, including GPR15/Bob, CXCR6/Bonzo and GPR1 (reviewed by Clapham & Weiss, 1997 ). Furthermore, the majority of red-capped mangabeys in Gabon are homozygous for a 24 bp deletion in their CCR5 gene and harbour an SIV strain that uses CCR2b and STRL-33 but not CCR5 (Chen et al., 1998b ).
GPR15 and CXCR6/Bonzo are used less frequently by HIV-1, while GPR1 is rarely exploited. Unlike the closely related SIVMAC/SIVSMM group viruses, HIV-2 variants that use CXCR4 do evolve in vivo, probably (as for HIV-1) during the later stages of disease. A minority of HIV-2 strains appear to use CXCR4 exclusively (Guillon et al., 1998 ; Reeves et al., 1999 ); however, most primary HIV-2 isolates use a much broader range of coreceptors compared to HIV-1 (Bron et al., 1997 ; McKnight et al., 1998 ; Morner et al., 1999 ). Coreceptors used by HIV-2 in the asymptomatic stage of disease are less clear, since virus loads are often very low and isolates are difficult to culture. R5 HIV-2 strains have been reported (Morner et al., 1999 ; Reeves et al., 1999 ); however, it is not certain whether such strains predominate during the asymptomatic phase and are preferentially transmitted over broadly tropic viruses. Table 2 lists 7TM receptors identified as coreceptors for HIV and SIV strains in vitro.
The interaction between gp120 and coreceptors
The gp120 sites that interact with coreceptors and determine tropism include the variable V1/V2 and V3 loops as well as a conserved region of -strands known as the bridging-sheet, which is situated between the V1/V2 and V3 loops (Fig. 4B). Variation in the variable loops may enable HIV to adjust the interaction with coreceptors so that different virus strains require subtly distinct sites on coreceptors to trigger fusion. Such variability may aid immune escape but will impact on virus phenotypes determined by coreceptor interactions. Such phenotypes include tropism, sensitivity to inhibition by coreceptor ligands and possibly the use of alternative coreceptors. This section describes the crucial gp120 and coreceptor regions involved in their interaction.
Chemokine receptors form rods in the membrane with a central pore surrounded by the 7TM regions. Bacterial rhodopsin is the only high resolution 7TM structure that has been resolved (Palczewski et al., 2000 ). Such proteins have four domains exposed on the cell surface: the N terminus and three extracellular loops (E1, E2 and E3). Coreceptors take up different conformations on cell surfaces and on different cell types (Baribaud et al., 2001 ; Lee et al., 1999 ), influencing their ability to support HIV infection. Such conformations may result from the formation of dimers, as reported for CCR5 (Lapham et al., 1999 ) or with heterologous chemokine receptors (Mellado et al., 1999 ). Associations may also occur with other cell surface molecules, as reported for CCR5 and CD4 (Wu et al., 1996 ; Xiao et al., 1999 ). Coreceptor sites involved in HIV entry are centred on the N terminus and E2. Mutagenesis studies showed the N terminus of CCR5 is important for coreceptor activity for HIV-1 R5 viruses (Hill et al., 1998 ). R5 strains, however, differ in their use of CCR5, as highlighted by the variation in their capacity to infect cells expressing different chimeric human/mouse CCR5 receptors (Picard et al., 1997a ). MAbs that bind the N terminus of CCR5 are most efficient at inhibiting gp120 binding, while E2-specific MAbs are potent inhibitors of fusion and infection (Lee et al., 1999 ; Olson et al., 1999 ; Wu et al., 1997a ). For SIVMAC, both M-tropic and T-tropic strains use CCR5; however, the former require the N terminus of CCR5, while E2 is crucial for T-tropic SIV strains (Edinger et al., 1997 ). It is unclear if there are CCR5-using HIV-1 strains with the properties of T-tropic SIV strains.
For X4 strains, E2 is critical. Deletion of the N terminus of CXCR4 affects some but not all strains (Picard et al., 1997b ), although, when present, participates in binding gp120 (Doranz et al., 1999 ). Chimeric coreceptors support X4 virus entry providing E2 is present (Lu et al., 1997 ); however, Brelot et al. (1999) showed that X4 strains vary in their use of CXCR4 E2 with different isolates dependent on distinct E2 residues for activity.
Electrostatic charge interactions are also likely to enhance gp120–coreceptor interactions. The N terminus of CCR5 (and often other coreceptors) is negatively charged due to three acidic amino acids and four (potentially) sulphated tyrosines, which are important for coreceptor function (Farzan et al., 1998 ). These negative residues may aid interactions with positive amino acids in and around the bridging sheet on gp120 (Kwong et al., 1998 ). Moreover, the V3 loops of X4 strains are positively charged, while E2 of CXCR4 contains five negative residues and these oppositely charged faces may interact, as suggested by Platt et al. (2001) . Mutagenesis of all five acidic residues does not eliminate HIV infection (Wang et al., 1998 ). Thus negatively charged residues at the N terminus of CCR5 and in E2 of CXCR4 may enhance their use by electrostatic interactions with R5 and X4 strains, respectively; however, they may not determine the specificity of the interaction.
The coreceptor-binding site on gp120 involves the conserved bridging-sheet that lies between the protruding V1/V2 and V3 loops, as well as some residues in V3 itself (Kwong et al., 1998 ). Antibodies to both regions block gp120–coreceptor interactions (Trkola et al., 1996 ; Wu et al., 1996 ). The V3 loop has long been known as a determinant of tropism and now coreceptor usage. Positive residues in V3 that confer an SI phenotype correlate with the use of CXCR4. The role of the V1/V2 loops in the coreceptor interaction is less clear, since an HIV-1 mutant with deleted V1/V2 loops was infectious (Cao et al., 1997 ), while recombinant gp120 deleted for V1/V2 bound coreceptors (Wu et al., 1996 ). However, when present, V1/V2 loops influence tropism (Koito et al., 1994 ; Westervelt et al., 199 1) and coreceptors used (Cho et al., 1998 ; Ross & Cullen, 1998 ). Sites in the V1/V2 loops, the bridging-sheet and the V3 loop may contribute to at least two specific interactions with coreceptors centred on the N terminus and E2. A ‘high affinity’ interaction at both sites may not be needed to trigger infection and may explain why the specificity of the coreceptor interaction can be predominantly mapped to either the N terminus or E2. In summary, diverse virus strains vary in the sites and specific amino acids of coreceptors that they exploit for recognition and triggering fusion. The capacity of HIV to vary the Env and coreceptor residues involved in their interaction will be a major mechanism of immune evasion.
Chemokine binding to chemokine receptors
Chemokines are chemoattractant proteins with roles in immune development, inflammation, immunity, embryogenesis and development. Chemokines are 70–120 residue polypeptides with a common folding pattern that forms an -helix underlying three anti-parallel -strands and a less-structured N terminus. The first N-terminal loop carries receptor-binding specificities and, once bound, the N terminus itself is thought to interact with a second receptor site to induce receptor activation and signalling. As for gp120, both the N terminus and the E2 of chemokine receptors are implicated in chemokine binding. The gp120- and chemokine-binding sites overlap but can be separated by mutagenesis. For detailed descriptions of chemokine and chemokine receptor structure and function, see the review by Rojo et al. (1999) .
|
Does HIV signal through CD4 or coreceptors?
|
|---|
Signals induced by HIV interacting with CD4 and/or coreceptors may prepare the intracellular environment for early replication steps during virus entry or modify cell conditions for virion production when newly made envelopes are trafficking onto the cell surface. Inactivated virions or cross-linked gp120 signal via CD4 and p56lck, inducing nuclear translocation of NF-
B on primary lymphocytes to promote cell cycle progression and commitment for virus production (Briant et al., 1996 ). Corbeil & Richman (1995) reported that apoptosis of CD4+ T-cells triggered during late stages of replication (Corbeil & Richman, 1995 ) does not occur in cells expressing CD4 receptors lacking the signalling capacity of their cytoplasmic tail (Corbeil et al., 1996 ). Soluble forms of both R5 and X4 envelope glycoproteins signal via CCR5 or CXCR4 in vitro (Davis et al., 1997 ; Hesselgesser et al., 1997 ; Weissman et al., 1997 ) and focal adhesion kinase was shown to associate with CCR5 following gp120 binding (Cicala et al., 1999 ). Whether a single virion ligates sufficient receptors to induce a signal during entry remains controversial. However, Arthos et al. (2000) reported that the capacity of viral envelopes to signal via CCR5 on macrophages correlated with the ability of viruses to undergo early post-fusion events.
Signalling per se is not needed for coreceptors to function for virus entry, since the pertussis toxin blocks signal induction but not HIV infection on CD4+ cell lines (Aramori et al., 1997 ). Truncation of the CCR5 cytoplasmic region or mutation of the DRY motif both block signal transduction but do not effect the capacity of CCR5 to act as a coreceptor (Gosling et al., 1997 ). Despite the observations of J. Corbeil and colleagues, signalling by newly synthesized envelopes will be minimized by several mechanisms (Vpu- and Nef-induced) that downregulate CD4. Loss of CD4 will prevent efficient Env–coreceptor interactions and signalling during late stages of replication. Finally, shed gp120 may interact with cell surface receptors on uninfected cells and induce signals; however, it is not known whether sufficient concentrations of shed gp120 are present in vivo.
|
Cell tropism of HIV in immune and nonimmune tissues
|
|---|
In vivo, HIV replication is restricted to haematopoietic cells that express CD4 and either CCR5 or CXCR4. HIV therefore contrasts with other highly lymphotropic retroviruses, such as human T-lymophotropic virus type I, where virus receptors are expressed on diverse cell types (Weiss et al., 1985 ). For other retroviral receptors see the review by Sommerfelt (1999) . The main cells infected by HIV are T-helper lymphocytes, macrophages and DCs. In vitro, R5 viruses infect primary cultures of both lymphocytes and macrophages (Berger et al., 1998 ), while X4 isolates also infect T-cell lines. The capacity of X4 strains to infect macrophages is controversial. However, we and others have shown that primary X4 isolates infect at least some populations of macrophages (Simmons et al., 1996 ; Valentin et al., 1994 ). In the blood of individuals that carry R5 viruses, the CD4+CD45RO+ memory T-cells carry most of the provirus load, although CD45RA+ naive cells are also infected. When CXCR4-using strains emerge, their tropism is broader and new cell populations are targeted. On T-cells, CCR5 expression is mainly restricted to memory cells, while CXCR4 expression is widespread and predominates on naive T-cells (Bleul et al., 1997 ). Symptomatic, X4 virus-carrying individuals have an increased provirus load in naive T-cells consistent with an expanded T-cell tropism (Blaak et al., 2000 ; Ostrowski et al., 1999 ). Early studies suggested that monocytes were infrequently colonized in vivo (Schnittman et al., 1989 ). However, recent reports indicate that monocytes harbour replication-competent viruses in patients treated with highly active anti-retroviral therapy (HAART) (Lambotte et al., 2000 ).
Whether DCs are infected has been controversial. Blood DCs form two distinct populations: CD11c+ myeloid and CD11c- plasmacytoid. Both populations express CD4, CCR5 and CXCR4 and support at least some level of HIV replication in vitro (Patterson et al., 2001 ). Their sensitivity to infection and extent of replication depends on their stage of maturation and phenotype (Bakri et al., 2001 ; Granelli-Piperno et al., 1998 ; Patterson et al., 1999 ). Blood plasmacytoid DCs are more sensitive to both HIV R5 and X4 viruses than myeloid DCs (Patterson et al., 2001 ). Immature myeloid dendritic cells were reported to selectively support replication by R5 viruses (Granelli-Piperno et al., 1998 ; Reece et al., 1998 ; Zaitseva et al., 1997 ). More mature cells are permissive to R5 and X4 virus entry; however, replication blocks prior to (Granelli-Piperno et al., 1998 ) and after (Bakri et al., 2001 ) provirus integration are described. Immature DCs, such as Langerhans’ cells at mucosal membranes, may be the first cells encountered by transmitting HIV. Such maturing cells potentially carry HIV either as DC-SIGN-trapped virus (Geijtenbeek et al., 2000 ; Masurier et al., 1998 ) or as infected cells to lymph nodes, where association with T-cells provides a potent medium for the rapid amplification of progeny virus.
Chemokines also influence the types of cells that become infected (Cocchi et al., 1995 ). Several CD4+CCR5+ T-cell clones from uninfected and nonprogressing HIV-1+ individuals were resistant to infection due (at least in part) to endogenously produced -chemokines (Saha et al., 1998 ; Vyakarnam et al., 2001 ). T-cell clones from AIDS patients were substantially more sensitive to infection by R5 viruses, consistent with an increasing colonization of CD4+CCR5+ T-cells as disease progresses. Along mucosal membranes, there is extensive stromal cell-derived factor 1 (SDF-1) expression and downregulation of CXCR4 on T-lymphocytes (Agace et al., 2000 ). Langerhans’ cells taken from under the skin express little surface CXCR4, whereas, on culture, high concentrations of CXCR4 held internally in vesicles are rapidly expressed (Zaitseva et al., 1997 ). These observations may explain the restricted transmission of X4 viruses across mucosal membranes and why DCs in vitro and away from the SDF-1-rich environment of mucosa support at least the early entry stages of X4 virus replication. Another explanation is needed to explain selective transmission of R5 viruses directly into the blood (Wilkinson et al., 1998 ). Thus, soluble factors such as chemokines in the tissue milieu or produced endogenously by target cells have a major influence on tropism.
In nonimmune tissues and organs, resident-specialized macrophages carry the virus load; for example, in the liver, HIV antigens are detected in Küppfer cells. The brain is colonized by HIV-1 early in infection and eventually results in dementia or related pathology in up to 30% of AIDS cases. The brain is physically isolated from the blood by the blood–brain barrier, a system of tight, gap junctions between endothelial cells in blood capillaries. HIV is probably carried into the brain by infected monocytes, macrophages or activated T-cells. The main brain cell types infected are perivascular macrophages and microglia (reviewed by Gabuzda & Wang, 2000 ). Astrocytes do not express CD4 but may be occasionally infected in neonates (Saito et al., 1994 ). Whether HIV-1 adapts to use brain-expressed coreceptors for replication in brain cells is unclear. Neurotropic and neurovirulent SIV variants have been isolated from infected macaques (Zink et al., 1998 ). However, it is not known if equivalent variants are involved in HIV-1 brain infection. It is also controversial whether CXCR4-using viruses colonize the brain when they emerge late in disease and the vast majority of virus isolates and envelope sequences from brain tissue indicate that R5 viruses predominate. Recently, Gorry et al. (2001) reported isolation of M-tropic R5X4 and X4 strains from the brain tissue of dementia patients. Whether such CXCR4-using strains are implicated in brain pathogenesis is not known and controversial. However, shed gp120 from X4 viruses has been shown to induce apoptosis of neurons (reviewed by Gabuzda & Wang, 2000 ). The majority of HIV-1 isolates, including R5 and X4 strains, from blood infect both primary microglial cells and macrophages in vitro (Ghorpade et al., 1998 ; He et al., 1997 ; Hibbitts et al., 1999 ; Shieh et al., 1998 ). R5 isolates that replicate more efficiently in microglial cultures have been selected in vitro (Shieh et al., 2000 ). Enhanced replication in cultured microglia was conferred by mutations in V1, an envelope region associated with coreceptor use. Specific amino acids at particular sites in the V3 loop (or motifs) have also been associated with envelopes in the brain (Power et al., 1994 ). The significance of such motifs is highly controversial but could be associated with the use of alternative brain-encoded coreceptors or with the differential use of CCR5.
|
Significance of coreceptors for transmission, replication and pathogenesis in vivo
|
|---|
For HIV-1, current data support a model where R5 viruses predominate early in the asymptomatic phase, before strains able to use CXCR4 and often several other coreceptors (R5X4++) emerge (Scarlatti et al., 1997 ).
Coreceptor and chemokine polymorphisms
Individuals who are homozygous for a 32 bp deletion (
32) in the CCR5 gene are greatly protected from infection whether infection is via sex (Dean et al., 1996 ), blood contact (Wilkinson et al., 1998 ) or from mother-to-child transmission (Philpott et al., 1999 ). The 32 bp deletion results in a premature stop codon and a truncated CCR5 protein that fails to reach the cell surface (Benkirane et al., 1997 ). Homozygotes are therefore effectively CCR5-. The protection conferred by 32 CCR5 homozygosity indicates that SI/X4 strains are rarely transmitted, although a small number of HIV+ 32 CCR5 homozygotes has been reported. Where tested, these individuals carry viruses that use CXCR4 rather than alternative coreceptors (Michael et al., 1998 ). Individuals heterozygous for 32 CCR5 are not protected from infection (Huang et al., 1996 ) but survive longer (Dean et al., 1996 ). These individuals express lower levels of CCR5 (Wu et al., 1997b ), partly due to a halved CCR5 gene dosage but also because the 32 CCR5 protein interacts with full-length CCR5 in the secretory pathway and retains it there (Benkirane et al., 1997 ).
Other human CCR5 polymorphisms include a single change (m303) in the CCR5 gene reported in one family (Quillent et al., 1998 ). m303 causes a premature stop codon that prevents the expression of CCR5. PBMCs from one affected family member who also carried a 32 CCR5 gene (m303/32 CCR5) were resistant to infection by HIV-1 R5 strains. Several other CCR5 single nucleotide polymorphisms (SNPs) result in amino acid substitutions that interfere with -chemokine binding and/or coreceptor activity (Carrington et al., 1999 ; Howard et al., 1999 ). The influence of these rare SNPs on HIV in vivo is not known.
Several polymorphic alleles and SNPs in the CCR5 gene promoter region have been identified (Martin et al., 1998b ; McDermott et al., 1998 ). One allele (CCR5 P1, characterized by a pattern of 10 specific bases at particular sites) has been shown to accelerate disease progression in homozygous individuals (Martin et al., 1998b ).
A polymorphism in CCR2b that results in a V64I change in a TM domain slows disease progression. CCR2b V64I has no effect on sexual transmission (Smith et al., 1997 ), although a protective effect on mother-to-child transmission was reported (Mangano et al., 2000 ). The V64I change does affect the capacity of CCR2b to act as a coreceptor or signal in response to chemokines (Lee et al., 1998 ). Protection may be due to another CCR5 promoter polymorphism (-1835) linked to V64I in the adjacent CCR2b gene. The -1835 polymorphism unlinked to CCR2b V64I is rare and, to date, studies disagree on whether it protects or accelerates (Martin et al., 1998b ; Mummidi et al., 1998 ). Mellado et al. (1999) reported that CCR2b V64I, but not wild-type receptors, formed heterodimers with CXCR4 when costimulated by monocyte chemotactic protein 1 (MCP-1) or SDF-1. Such heterodimers may reduce CXCR4 available for HIV infection and thus could explain CCR2b V64I protection.
CXCR4 is indispensable to mammals. In mice, both CXCR4 (Ma et al., 1998 ; Zou et al., 1998 ) and SDF-1 (Ma et al., 1998 ) ‘knockouts’ are lethal. So far, only three rare CXCR4 polymorphisms have been reported that are not linked to pathogenesis (Cohen et al., 1998 ; Martin et al., 1998a ). A polymorphism in SDF-1 (the CXCR4 ligand) gene was reported as protective (Winkler et al., 1998 ). However, other studies showed a faster disease rate (Mummidi et al., 1998 ; van Rij et al., 1998 ) or a more rapid decline in CD4 cell numbers (Balotta et al., 1999 ). This G to A ‘mutation’ is located in the 3' noncoding region of SDF-1 mRNA and may influence mRNA stability. Two SNPs in the promoter of the RANTES gene (-471 and -96) also slow disease (Gonzalez et al., 2001 ; McDermott et al., 2000 ), while one study found an effect on transmission risk (McDermott et al., 2000 ). G at -96 led to an increase in RANTES expression providing an explanation for protection (Liu et al., 1999a ). Two SNPs in the first intron of macrophage inflammatory protein 1 (MIP-1) were also reported to influence disease progression (Gonzalez et al., 2001 ). Together, these observations provide evidence that -chemokines act protectively in vivo. Table 3 summarizes the known polymorphisms in HIV receptors and their ligands that influence the course of HIV infection (reviewed by Carrington et al., 1999 ).
View this table:
[in this window]
[in a new window]
|
Table 3. Human polymorphisms in chemokine and coreceptor receptor genes that influence HIV infection and disease progression
|
|
X4 virus transmission
The protection from HIV-1 infection conferred by the 32/32 CCR5 genotype indicates that viruses using CXCR4 or other coreceptors rarely transmit. X4 strains have less opportunity for transmission since R5 strains predominate in most infected individuals. Extensive SDF-1 expression by mucosal epithelia may act as a barrier to X4 viruses (Agace et al., 2000 ). Harouse et al. (1999) showed that R5 but not X4 SHIV (Chimeric SIV/HIV-1) strains replicated extensively in rhesus macaque gut lymphoid tissue. Thus, X4 viruses that penetrate the SDF-1 barrier at the mucosa may arrive in associated lymphoid tissue containing T-cells and macrophages with CXCR4 downmodulated and inaccessible. Dual-tropic SI strains that use CCR5 and CXCR4 are very sensitive to inhibition by CCR5 chemokines when CXCR4 is absent (Kledal et al., 1997 ). Such sensitivity may be sufficient for -chemokines to prevent R5X4 transmitting via a CCR5 route.
32 CCR5 homozygotes may also be protected from blood transfer that bypasses the mucosa (Wilkinson et al., 1998 ). This observation suggests that there are restrictions to X4 viruses beyond the mucosa and that an incoming virus may need to establish infection at particular site(s) for transmission to be successful. Perhaps a critical tissue or site is permissive for R5 strains but not X4 viruses. Again, observations by Harouse et al. (1999) that gut lymphoid tissue in macaques supports extensive replication by R5 but not X4 SIV/HIV viruses provide a precedent.
Suppression of X4 strains?
R5 virus replication and variation would be expected to generate variants that can use CXCR4 in a short period of time. Yet, SI/X4 viruses apparently do not always evolve in vivo, can be isolated from only about 50% of AIDS patients and infrequently from HIV-1 subgroup C-infected individuals (Tscherning et al., 1998 ). X4 viruses are not isolated from SIVMAC-infected rhesus macaques, even though some SIV strains can use CXCR4 in vitro (Owen et al., 2000 ). It is not known if undetectable levels of CXCR4-using viruses are always present. Regardless, the mechanisms that prevent X4 viruses from predominating in vivo are not understood. Whatever the nature of the restriction, it breaks down and/or is breached during the later stages of disease when CXCR4-using viruses emerge in HIV-1-infected individuals. Valentin et al. (1998) described how IL-4 downregulates CCR5 while upregulating CXCR4 and enhancing HIV expression. Thus, IL-4 may select for X4 viruses and against R5 strains, a possibility supported by the observation that HIV+ individuals carrying a polymorphism in the IL-4 gene promoter that increases expression were more likely to harbour X4 viruses (Nakayama et al., 2000 ).
Two early studies suggested that SI variants present in the acute phase were later suppressed in favour of NSI viruses at seroconversion (Cornelissen et al., 1995 ; Lathey et al., 1997 ) and speculated that SI suppression was due to an immune-mediated mechanism (Lathey et al., 1997 ). However, primary X4 viruses are as resistant to neutralizing antibodies as R5 strains, while a role for T-cell immunity is difficult to envisage, since T-cell epitopes on the envelope glycoprotein are few and unlikely to distinguish between R5 and X4 strains (reviewed by Michael & Moore, 1999 ). The current consensus strongly favours infrequent transmission of CXCR4-using strains and their emergence only late in disease at the peak of virus diversity (Shankarappa et al., 1999 ). If X4 strains are frequently present at low levels in infected individuals, new therapies aimed at CCR5 may provide X4 viruses with a selective advantage.
The role of other coreceptors
The extent to which HIV-1 exploits coreceptors other than CCR5 or CXCR4 in vivo is thought to be minimal (Zhang & Moore, 1999 ). The growing number of different 7TM receptors that support HIV and SIV infection of cell lines in vitro therefore does not accurately predict coreceptor usage in vivo. High-level expression of alternative coreceptors ‘out of context’ on cell lines seems to deliver them to the cell surface in an active form that can confer virus entry. Factors in vivo that may prevent many of the same alternative coreceptors from functioning (as envisaged for CXCR4) are not known. Recent evidence implicates CXCR6/Bonzo for HIV-1 infection of some T-cells (Sharron et al., 2000 ) and CCR8 for thymocytes (Lee et al., 2000 ). Furthermore, one or more unidentified coreceptors frequently support HIV-2 and SIV infection of primary T-cells and macrophages in vitro (Chen et al., 1998a ; Simmons et al., 2000 ; Sol et al., 1997 ; Zhang et al., 2000 ). Despite these observations, there is no evidence yet to indicate that coreceptors other than CCR5 or CXCR4 significantly influence HIV or SIV replication in vivo.
|
CD4-independent infection
|
|---|
Many reports describe CD4-independent infection of various cell types by HIV-1 in vitro (reviewed by Clapham et al., 1996 ). Such infection is generally inefficient, although HIV-1 and HIV-2 variants selected in vitro are substantially more proficient (Clapham et al., 1996 ; Hoffman et al., 1999 ). These variants interact directly with coreceptors. Primary HIV-2 isolates generally infect CD4- coreceptor+ cells more efficiently than HIV-1 (Reeves et al., 1999 ). CD4-independent infection, however, is ultrasensitive to inhibition by neutralizing antibodies as well as coreceptor ligands (Puffer et al., 2002 ) and evidence that CD4- cells are infected with any frequency in vivo is limited.
|
Therapies targeted at HIV entry
|
|---|
TOP
Abstract
Introduction
