CD4-independent infection of HIV and SIV

Introduction The primate immunodeficiency viruses (HIV and SIV) are enveloped viruses that normally require successive interactions with CD4 and a co-receptor to trigger the fusion of viral and cellular membranes and entry into cells. The chemokine receptor CCR5 is the major co-receptor for HIV and SIV; however, CXCR4-using HIV-1 variants evolve in up to 50% of AIDS patients. The emergence of CXCR4-using viruses correlates with a more rapid depletion of CD4 cells and disease progression. Nevetheless, CD4 cell depletion and AIDS occur in patients from whom only CCR5-using viruses can be isolated. CD4-independent HIV variants that interact directly with co-receptors [1,2] can be isolated in vitro [2–4]. These viruses are sensitive to neutralizing antibodies, and particularly vulnerable if target cells do not express CD4 [5,6]. It is thus likely that CD4-independent viruses are selected against in vivo, and CD4-independent routes of infection are easily curtailed by antibody surveillance. These observations indicate that the two-receptor strategy adopted by HIV for entering cells may protect critical envelope sites from neutralizing antibodies. The result is that relatively few CD4- negative cell types are infected in vivo. CD4-independent SIVs with reduced affinity for CD4 have been isolated from rapid progressor rhesus macaques that failed to produce high titer neutralizing antibodies [7,8]. Moreover, macrophage-tropic SIVs are generally more CD4-independent than T-cell tropic SIVs [9], whereas a highly macrophage-tropic HIV-1 isolate from the brain could infect cells that express minimal levels of CD4 [10]. Together, these observations indicate that viruses with a reduced requirement for CD4 may evolve when neutralizing antibody concentrations are low. Such conditions may occur in rapid progressors (rhesus macaques or humans), immunoprivileged organs such as the brain, and late in disease when the immune response is waning. The emergence of variants that require less CD4 for infection may confer a more extensive colonization of macrophages in the brain and other non-lymphoid tissues, as well as T-cell subsets that express low levels of CD4. An expanded tropism among CD4+ T-cell populations could explain why CD4 T cell depletion and AIDS occur in individuals from which T-cell tropic CXCR4-using viruses cannot be isolated. Similarly, an enhanced macrophage tropism and infection of truly CD4-negative astrocytes in the brain may contribute to the HIV associated dementia that occurs in some AIDS patients. The viral envelope glycoprotein on virus particles consists of a surface gp120 non-covalently attached to a transmembrane gp41. These gp120/gp41 glycoproteins are arranged as trimers on the virus particle. The binding of gp120 to CD4 is usually required to induce a structural alteration in gp120 that exposes a binding site for the co-receptor. CD4 binding may also position the exposed co-receptor binding site close to the cell surface enabling contact with the co-receptors. This arrangement may enable envelope–co-receptor interactions to occur even if their affinities are not high. Co-receptor–envelope interactions then trigger rearrangements in the transmembrane gp41 that induce fusion. HIV receptors and fusion are reviewed in Clapham and McKnight [11]. HIV-1, HIV-2 and SIV variants have been isolated that can efficiently infect co-receptor-expressing cells in the absence of CD4. These CD4-independent viruses carry gp120 in a native conformation in which the co-receptor binding site is exposed and can interact directly with co-receptors [1,2] (Fig. 1). Such CD4-independent viruses might be expected to have an advantage in vivo and be able to colonize a wider range of cell types that express co-receptors but not CD4. This article reviews recent advances in understanding the envelope conformation of CD4-independent viruses, the prevalence of such HIV and SIV variants in vivo, and their impact on cell tropism.Fig. 1.: Conformations of gp120 and implications for CD4-dependent and CD4-independent infection. (a) Putative conformations of native gp120 monomers. The structure of native gp120 is not clearly understood. Myszka et al. [12] proposed that CD4 engagement induces the final folding of the bridging sheet and V1/V2 loop movement to expose a mature co-receptor binding site. Therefore, on CD4-bound gp120, the exposed co-receptor binding site comprises sites on the bridging sheet and in the V3 loop. The conformations shown speculate that part of the bridging sheet folds from an α-helix (left) as predicted [13,14]. It is not clear whether some HIV/SIV strains carry preformed bridging sheets but still require CD4 binding for V1/V2 movement (middle). On envelopes capable of CD4-independent infection, the co-receptor binding site is (at least partly) open, implying that the bridging sheet is formed and exposed, whereas V1/V2 positioning does not impede access (right). (b) CD4-dependent and independent interactions with co-receptors. (Left) CD4–gp120 interactions expose the binding site for the co-receptor. (Right) The co-receptor binding site is pre-exposed on a CD4-independent envelope and co-receptor interactions occur independently of CD4.Mechanisms of CD4 independence Consideration of gp120 structure Several reports have described HIV-1 and HIV-2 variants selected in vitro for their capacity to infect co-receptor+ cells that lack CD4 [2,4,15,16]. Insights into the mechanisms of CD4-independence have come from rapid progress in understanding the structure of gp120. It is clear that such variants simply bypass CD4 and interact directly with co-receptors. A high resolution structure of gp120 complexed with CD4 and a gp120-specific antibody was reported by Kwong et al. [17]. Although this structure lacked the envelope variable loops, it provided an understanding of the sites on gp120 that bind CD4 and co-receptors. This view has been refined by antibody mapping and mutagenesis studies. The co-receptor binding site consists of a relatively conserved sheet of β-strands known as the ‘bridging sheet’ [18,19] and residues in the V3 loop, the major determinant of co-receptor-use (Fig. 1) [18–21]. It has been proposed that gp120 residues in the base of the V3 loop and in conserved region 4 (C4, immediately downstream from the β21 strand; see Kwong et al. [17]) bind the N-terminus of CCR5 [22,23], whereas the V3 loop crown and other residues in conserved regions of gp120 (C1, C2 and C3) are involved in interactions with the CCR5 extracellular loops [23]. The V1/V2 loops also influence co-receptor use [24,25]; however, it is less clear whether V1/V2 residues contact the co-receptor. These loops are not essential for infection because replication-competent viruses deleted for V1/V2 can be constructed [26–28]. The precise location of the V1/V2 and V3 loops is not known because these regions were absent from the gp120 structure determined by Kwong et al. [17]. The stems of V1/V2 and V3 were present, and these sit on either side of the bridging sheet [17] (Fig. 1). Moreover, antibody-binding studies demonstrated that the co-receptor binding site is shielded by the extended V1/V2 loop and is exposed after CD4 binding [29–31]. The precise structure of native gp120 before it binds CD4 is not known. However, calorimetric studies indicated that extensive conformational changes occur in the core and periphery of gp120 when it binds to CD4. These changes are thought to involve: (i) the folding of the bridging sheet to form the conserved part of the co-receptor binding site; (ii) the pulling together of inner and outer gp120 domains; and (iii) movement of the V1/V2 loops to uncover the co-receptor binding site [12]. CD4 binding thus locks gp120 into an optimal gp120 structure for co-receptor binding and the induction of gp41 conformational changes required for fusion and virus entry into cells. CD4-independent variants isolated in vitro CD4-independent HIV-1 and HIV-2 variants have been isolated in vitro by passage through cells lacking CD4. CD4-independent HIV-2 variants (ROD/B and VCP) were described first and were shown to bind directly with CXCR4 [3,4,15] in a high affinity interaction [32]. La Branche et al. [33] cloned a variant of HIV-1 HXBc2 (8x) after serial passage through CD4-negative BC7 cells (a derivative of the Sup T1 T-cell line) that replicated in CD4-negative CXCR4+ cells, and carried a stably exposed co-receptor binding site [1]. Both HIV-1 8x and the HIV-2 ROD/B variants were shown to retain the capacity to bind CD4 [1,7], but did not require CD4 for the infection of CXCR4+ cells. Multiple amino acid substitutions were detected in the envelope region of these viruses; however, sequences in V3, around the base of V4 (i.e. close to, or part of the co-receptor binding site) were critical for the CD4-independent phenotype [33,34]. CD4-independence of both ROD/B and 8x also required changes in gp41 that are likely to enable fusion to be triggered after weaker interactions between gp120 and co-receptors in the absence of CD4. Together, these observations indicate that these HIV isolates have undergone a succession of amino acid changes in envelope to achieve CD4-independence. These changes in the native gp120 exposed the co-receptor binding site, increase envelope affinity for the co-receptor and enhanced fusigenicity. In contrast, Kolchinsky et al. [26] described a mutant of the CCR5-using HIV-1 ADA strain, in which the removal of a single glycosylation site at the downstream base of the V2 loop conferred the capacity to infect cells via CCR5 without CD4. The structural models of gp120 [17] suggested that this glycosylation group is unlikely to occlude the co-receptor binding site directly, and its removal probably results in the repositioning of the V1/V2 loops [26]. This interpretation is supported by a variant of the HIV-1 NDK isolate in which CD4-independence corrrelated with the acquisition (rather than loss) of two glycosylation sites in V1/V2 [35]. An ADA mutant with V1/V2 deleted was also able to infect CD4− CCR5+ cells, and its gp120 was shown to bind CCR5 directly [26]. Similarly, an SIV lacking V1/V2 could infect cells expressing CCR5 in the absence of CD4 [28]. These observations confirmed that the V1/V2 domain is not essential for infection, and that its removal or repositioning exposes a site that can bind the co-receptor. Table 1 lists well-known and characterized CD4-independent HIV and SIV.Table 1: CD4-independent HIV and SIV: origins and characteristics.CD4-independent variants thus fall into two groups. First are those in which a deletion or change in the orientation of V1/V2 unmasks the co-receptor binding site. Second are those that contain mutations elsewhere in the envelope. Such mutations may also shift V1/V2 orientation; however, changes in V3 and the base of V4 are also likely to increase envelope affinity for co-receptor binding (or pre-folding of the bridging sheet). In addition, amino acid substitutions in gp41 may enable gp41 conformational rearrangements and fusion to occur after a weaker env–co-receptor interaction. Receptors that substitute for CD4 Alternative receptors that can bind gp120 and induce exposure of the co-receptor binding site have not been described. Many cell-surface receptors bind gp120 via carbohydrate domains, often to high affinity. These include galactosylceramide, DC-SIGN, and mannose-binding protein [11]. However, none of these receptors appears to participate in envelope activation and fusion. Galactosylceramide does confer low-level infection of neuronal cell lines [46], although relatively inefficient, CD4-independent infection of astrocytes may involve an astrocyte receptor [47,48]. Whether these receptors induce weak gp120 conformational changes is not known. Saha et al. [49] also reported the efficient infection of CD8+ cells that is blocked by the presence of CD8 antibodies. These observations are consistent with the use of CD8 as an alternative receptor; however, they have yet to be confirmed by others. Differences between HIV-1 and HIV-2/SIV? Many HIV-2 and SIV isolates can infect cells without CD4 at least to some extent, whereas HIV-1 isolates are predominantly CD4-dependent [39]. Only one HIV-1 primary isolate with an unusual V3 loop sequence has been shown to infect CD4-negative cells, and not efficiently [36]. These observations may mean that the native structure of HIV-2 and SIV envelopes is subtly distinct from HIV-1, and carries an exposed or partly exposed co-receptor binding domain. The HIV-1 ADA isolate also highlights possible distinctions in the envelopes of different HIV/SIV strains. For ADA, a single V2 substitution confers CD4-independence (see above) [26], whereas HIV-1 8x and HIV-2 strains ROD/B and VCP required the accumulation of successive changes in gp120 and gp41 [33,34]. The basis of such strain differences is not known, but could reflect variation in gp120/co-receptor affinities or differences in how readily gp41 conformational changes are induced. For example, viruses with a high affinity for CCR5 might not require further mutations to increase CCR5 affinity to a level sufficient to support binding and infection in the absence of CD4. Such viruses may therefore be rendered CD4-independent by relatively few mutations that simply confer exposure of the co-receptor binding site. Another possibility comes from Myszka et al. [12], who suggested that gp120/CD4 cell binding induces folding of the bridging sheet as well as movement of V1/V2 to expose the co-receptor binding site (Fig. 1). It is thus possible that the bridging sheet is already folded in the native gp120 of some viruses, e.g. HIV-2 and SIV, but is still partly or completely protected by the V1/V2 loop. Then relatively few gp120 changes that move V1/V2 could confer CD4-independence. Fig. 1 outlines some of these possibilities. CD4-independence and tropism CD4-independence of SIV isolates has been shown to correlate with macrophage tropism [9,50]. As rhesus macrophages express very low levels of CD4 [51,52], it follows that CD4-independent SIV will have an increased capacity to infect such cells. A similar correlation between the macrophage tropism and CD4-independence of HIV-1 has not been described. However, Gorry et al. [10] reported that an HIV-1 isolate made from the brain tissue of an individual with HIV-associated dementia infected cells that carry low levels of both CD4 and CCR5. Martin et al. [53] also reported that an HIV-1 variant adapted for replication in brain microglia in vitro also infected cells expressing low levels of CD4 and CCR5. These isolates are not CD4-independent, but carry an enhanced capacity to infect human macrophages, which (like rhesus macrophages) also express only modest CD4 levels [51,52,54]. Thus, highly macrophage-tropic SIV and HIV-1 strains evolve that require less cell surface CD4 for infection. CD4-independence of SIV is usually defined by testing the infection of CD4-negative cell lines that express high levels of CCR5. The extent to which such CD4-independent strains enter cells via a CD4-independent pathway and carry an expanded tropism for CD4-negative cells in vivo is less clear. Thus, the CD4-independent phenotype of SIV revealed in vitro, may simply reflect an adaptation in vivo for infection of cells that express minimal cell surface CD4. CD4-independent SIV in vivo CD4-independent SIV isolates from lung and brain CD4-independence of SIV correlates with macrophage tropism [9]. SIV envelopes or isolates derived from tissues, in which macrophages form the main target for infection, thus confer macrophage tropism and infection of CD4-negative CCR5+ cell lines in vitro [9]. These include envelopes from lung (SIVmac 316) [43] and brain tissue (SIVmac 17E-Fr [45] and SIVmac 1100 [9,55]) of diseased macaques. Although these viruses are CD4-independent, their infectivity is increased when CD4 is co-expressed [9], thus reflecting their enhanced tropism for macrophages. CD4-independent SIV from rapid progressor rhesus macaques Two independent groups have reported that rapid progression to AIDS in macaques was accompanied by the selection of CD4-independent, highly macrophage-tropic viruses [7,8]. Most SIV-infected macaques survive for 1–2 years; however, a small percentage progress rapidly to a wasting disease in 3–6 months of infection. Both Ryzhova et al. [7] and Dehghani et al. [8] demonstrated that SIV envelopes amplified by polymerase chain reaction from rapid progressors carried mutations in a gp120 motif (GDPE) that directly contacts CD4. These changes were associated with a 40-fold reduction in binding affinity for CD4 [7]. Other changes in the gp120 of these variants occurred in V3 as well as V1/V2, where the loss of an N-linked glycosylation site resulted. These latter mutations probably increased exposure of the co-receptor site and affinity for CCR5 [8]. Like the lung and brain derived, CD4-independent SIV described above, it is likely that SIV from rapid progressors will also have an enhanced tropism for CCR5+ macrophages. However, it is not yet clear whether the reduced CD4 affinity of these SIV correlates with an expanded tropism for truly CD4-negative cells in vivo. CD4-independent viruses are neutralization sensitive Several recent studies have shown that CD4- independent variants are more sensitive to neutralizing antibodies than CD4-dependent viruses [5,6,9,50]. Kolchinsky et al. [6] showed that the CD4-independent ADA strain (which lacked a V2 glycosylation site) was substantially more sensitive to neutralizing antibodies than the CD4-dependent virus it was derived from. Sensitivity was increased to a range of neutralizing monoclonal antibodies that recognized different envelope epitopes, including the CD4 binding site, CD4-induced epitopes (that overlap the co-receptor binding site), the V3 loop and gp41 determinants. Hoffman et al. [1] also showed that the HIV-1 HXBc2 8x variant was more sensitive to neutralization by HIV-positive human sera as well as the CD4-induced epitopes and by monoclonal antibodies to CD4 binding site. Similarly, the neutralization sensitivity of the CD4-independent HIV-2 RODB strain was increased for monoclonal antibodies that targeted different gp120 epitopes [5]. Several studies also demonstrated that for CD4-dependent viruses, pre-exposure of the co-receptor binding site (induced by soluble CD4) imparts ultrasensitivity to neutralizing antibodies [5,9,56]. The mechanisms that confer increased sensitivity to antibodies that target different sites on gp120 and gp41 are not completely clear. However, it is possible that gp120 structural alterations, which expose the co-receptor binding site, may also open up other sites that can be targeted by neutralizing antibodies. All studies to date agree that the sensitivity of CD4-independent HIV/SIV to neutralizing antibodies is more evident when infection is carried out on CD4-negative cells [5,9]. This observation indicates that entry via CD4 has evolved as a robust pathway, and is difficult for the humoral immune response of the host to curtail effectively. The CD4-independent infection route is easily abrogated by antibody surveillance, and is thus likely to be strongly selected against in vivo. However, several reports (described above) have indicated that CD4-independent viruses or strains that are less dependent on CD4 are present in vivo. How do such CD4-independent variants survive if they are sensitive to neutralizing antibodies? Rapid progressor rhesus macaques that yield CD4-independent viruses, develop weak immune responses and may have low concentrations of neutralizing antibodies at best [7,8]. Neutralization-sensitive, CD4-independent SIV may also flourish in immunoprivileged sites, e.g. in the brain, where protection by the blood–brain barrier keeps B cells and antibody levels low, or late in disease [57] when the immune system is compromised. The prevalence of such CD4-independent viruses (with an enhanced tropism for macrophages) in immunocompetent macaques during the asymptomatic phase is not known. In contrast to SIV, macrophage tropism of HIV-1 does not correlate with a capacity to infect cells in the absence of CD4. Nevertheless, HIV-1 variants with a reduced requirement for CD4 and CCR5 have been reported [10,53] (see above). These include a variant of the R5 HIV-1 BORI isolate that was adapted by passage through microglial cell cultures [53] as well as a CCR5-using brain isolate from a patient with HIV-associated dementia [10]. Both these variants have an enhanced capacity for replication in cultured microglia and macrophages. Like their SIV counterparts, these HIV-1 variants are also more sensitive to neutralization [10,53]. These observations are consistent with the hypothesis that the immune privileged of the brain the emergence of variants that can efficiently in brain microglia and macrophages that express only low levels of CD4. CD4-negative cell types infected by HIV and SIV The extent to which CD4-negative cells are infected in vivo is however, it is generally thought to be low Moreover, is clear yet that HIV-2 or SIV (that CD4-independence in infect more CD4-negative cells in vivo with HIV-1 [5]. The best of CD4-negative cell infection is in the brain where CCR5-using viruses Several groups reported the of HIV sequences in astrocytes particularly in AIDS Moreover, et al. were able to rhesus astrocytes vivo that were infected with Several studies both CCR5 and CXCR4 on astrocytes however, CCR5 is on studies indicated that infection of brain cells in is but could be sufficient to HIV to the blood–brain Moreover, et al. reported SIV infection of in a blood–brain barrier human support the infection of some not HIV isolates et al. described a primary from a single that CD4 but not CCR5 This supported infection by HIV and SIV strains that were able to use a range of alternative although the for HIV-1 replication in the brain is because CCR5-using strains In contrast, et al. described cells that express both CCR5 and CXCR4 but not CD4. Together, these observations indicate that in the brain are likely to be Other CD4-negative brain cell e.g. are infected with only some studies infected Two recent studies and detected HIV-1 sequences in isolated These are observations and could indicate infection, but are difficult to with the from studies only infection. In the of CD4-independent HIV/SIV has provided into different conformational of gp120. viruses that can infect cells in the absence of CD4 carry a pre-exposed co-receptor binding site. However, this gp120 structure confers to neutralizing antibodies, particularly for the infection of CD4-negative cells. via CD4 is thus a robust pathway that CD4-dependent viruses against neutralizing antibodies as well as some protection for CD4-independent These observations could explain the presence of CD4-independent SIV detected in vivo, where such viruses may have evolved to low levels of CD4 on macrophages other cell than for infection of truly CD4-negative cell Therefore, CD4-independent SIV can be detected in vivo even infection of CD4-negative cells in vivo is generally low and of for HIV Several have reported the infection of CD4-negative brain astrocytes in which replication is to the phase This that CD4-negative astrocytes are targeted by viruses in the brain, where neutralizing antibodies are by the blood–brain Although the impact of astrocyte infection on is such infected cells will contribute to HIV that are to on the