Role of CXCR4 in HIV Infection and its Potential as a Therapeutic Target

Tsutomu Murakami; Naoki Yamamoto


Future Microbiol. 2010;5(7):1025-1039. 

In This Article

Coreceptor-targeted Anti-HIV Therapy

Along with CCR5, CXCR4 is a major coreceptor in HIV infection. R5 HIV-1 is isolated predominantly during the acute and asymptomatic stages,[38] whereas X4 HIV-1 strains emerge in approximately 50% of infected individuals, and their emergence is associated with rapid CD4+ T cell loss and disease progression.[39,40] CXCR4, therefore, is a novel and attractive target for the development of new anti-HIV drugs. Significant efforts have been made to explore and identify small-molecule compounds that interfere with the interaction between gp120 and CXCR4. The structures of some of the well-studied CXCR4 inhibitors discussed below are shown in Table 1 & Figure 2.

Figure 2.

CXCR4 inhibitors.

AMD3100 & AMD070

Bicyclam AMD3100 (Table 1 & Figure 2) is a small molecule inhibitor that strongly restricts X4 HIV-1 infection.[41] The compound exhibits no antiviral activity against CCR5-utilizing HIV-1 strains. Correlation was observed between the inhibitory activity of AMD3100 on X4 HIV-1 replication, CXCR4 mAb binding and SDF-1α-induced signal transduction. Thus, AMD3100 is a specific antagonist of CXCR4. AMD3100 has proven effective not only in a severe combined immunodeficiency (SCID)-hu/Thy/Liv mouse model[42] but also in a proof-of-concept study on a patient infected with X4 HIV-1. Development of the compound as an anti-HIV drug was suspended primarily due to its cardiotoxicity.[43] An effort to overcome the poor oral bioavailability and side effects of AMD3100 led to the generation of AMD070 (Table 1 & Figure 2), which has tetrahydroquinolineamine as its pharmacophore and which is a potent and specific X4 HIV-1 inhibitor with high oral bioavailability. Although AMD070 also showed potential activity against X4 HIV-1 in a clinical Phase Ib/IIa proof-of-concept study, the FDA halted development of this compound due to liver histology changes in long-term preclinical toxicity studies.[44] Further studies are needed to determine in vivo toxicity of AMD3100 and AMD070, is due to blocking CXCR4 functions.

T22, T134 & T140

Self-defense peptides isolated from the hemocytes of the Japanese and American horseshoe crab tachyplesins and polyphemusins have antibacterial and antiviral activities.[45–47] Several years prior to the discovery of the HIV-1 coreceptors, Yamamoto and colleagues were able to show that T22 ([Tyr5, 12, Lys7]-polyphemusin) (Table 1 & Figure 2), a synthetic polyphemusin analog with 18 amino acid residues and two disulfide bonds, inhibits HIV-1 replication.[47,48] The two disulfide bonds form an antiparallel β-sheet structure, which is important for the antiviral activity of the peptide. The peptide inhibits replication of T-tropic but not M-tropic HIV-1. The determinant of this specific antiviral activity was mapped to the V3 region of the HIV-1 Env protein,[49] and the 50% effective concentration (EC50) was 290 nM in an anti-HIV assay using MT-4 cells. Subsequent to the discovery of the CXCR4 and CCR5 coreceptors, we demonstrated that the T22 peptide specifically blocks virus–cell and cell–cell infection mediated through HIV-1 Env interaction with CXCR4 and CD4,[50] as also reported in other CXCR4 inhibitors such as ALX40-4C and AMD3100.[41,51] It was also found that T22 suppresses Ca2+ mobilization induced by SDF-1α. Thus, T22 is a small CXCR4 antagonist that inhibits X4 HIV-1 infection via specific binding to the CXCR4 molecule.

Through an effort to downsize T22 by structure–activity relationship studies, it was found that T134[52] and T140[53] (Table 1 & Figure 2), 14-mer peptides with a single disulfide bond, had stronger anti-HIV-1 activity than T22. Although T140 lacks four amino acids and one disulfide bond, it maintains an antiparallel β-sheet structure. Through an Ala-substitution scanning study, it was demonstrated that Arg2, Nal3, Tyr5 and Arg14 form the pharmacophore of T140, which is useful information for the rational design of peptide derivatives with higher anti-HIV activity.[54] Downsizing T140 resulted in the development of a cyclic pentapeptide, FC131, which has strong CXCR4-antagonistic activity and could serve as a template for further modification.[55] Systemic toxicity of the T compounds was evaluated with TN14003, a derivative of T140.[56] CB-17 SCID mice were injected with TN14003 at 100 ng/g body weight twice weekly for 45 days. Although no damage in the liver and kidney was observed by hematoxylin and eosin staining, further careful in vivo study is required for assessing the safety of the T compounds.

KRH-1636 & KRH-3955

In an attempt to discover a novel, small nonpeptide CXCR4 antagonist Yamamoto and colleagues screened more than 1000 compounds from the Kureha Corporation (Tokyo, Japan) chemical library. This screen led to the identification of KRH-1636 (Table 1 & Figure 2), which strongly and specifically inhibits X4 HIV-1 replication, both in vitro and in vivo.[57] KRH-1636 blocks the replication of various X4 HIV-1 strains in a nanomolar range and has low cytotoxicity (CC50: 400 µM in MT-4 cells). KRH-1636 also strongly inhibits both SDF-1α binding to CXCR4 and CXCR4-mediated Ca2+ signaling and blocks binding of monoclonal antibodies to CXCR4 without down-modulating the coreceptor. Importantly, KRH-1636 also inhibits X4 HIV-1 replication in a human peripheral blood lymphocyte (PBL)/ SCID mouse model. Furthermore, KRH-1636 was absorbed into the blood after intraduodenal administration in rats. KRH-1636 did not show severe toxicity in rats that received the compound (1.5 mg/kg per day) by intravenous administration for 4 days.[57]

Vigorous efforts to search for more potent and orally bioavailable CXCR4 antagonists were undertaken through a combination of chemical modification of KRH-1636 and biological assays, leading to the identification of KRH-3955 (Table 1 & Figure 2).[58] KRH-3955 very potently inhibits replication of X4 HIV-1 in activated peripheral blood mononuclear cells (PBMCs) from different donors and effectively restricts clinical HIV-1 isolates. The EC50 of KRH-3955 is approximately 1 nM, and it also blocks replication of recombinant X4 HIV-1 containing resistance mutations in reverse transcriptase and protease, as well as isolates with T-20-resistant mutations in the envelope protein. KRH-3955 inhibits both SDF-1α binding to CXCR4 and Ca2+ signaling through the coreceptor. Moreover, KRH-3955 does not induce CXCR4 internalization but inhibits the binding of anti-CXCR4 monoclonal antibodies that recognize the second or third extracellular loops of the receptor. The compound shows an oral bioavailability of 26% in rats and oral administration blocks X4 HIV-1 replication in the human PBL/SCID model. Thus, KRH-3955 is a strong CXCR4 antagonist and seems to be a new promising therapeutic agent against HIV-1 infection and AIDS although further studies are definitely needed on in vivo safety of the compound.

Other CXCR4 Inhibitors

In addition to the compounds above, there are several other CXCR4 inhibitors. These include: ALX40-4C (Table 1 & Figure 2)[51,59] and Arg-mimetic conjugates,[60,61] POL3026 (β-hairpin mimetic)[62,63] and CGP64222.[64] ALX40–4C, a polypeptide of nine D-Arg residues stabilized by terminal protection, inhibits X4 HIV-1 infection as well as R5X4 (dual-tropic) HIV-1 infection in the context of CXCR4 use.[51] The peptide also blocks binding of SDF-1α and the anti-CXCR4 monoclonal antibody 12G5 to CXCR4. Although no significant reduction in viral load was observed, ALX40–4C was well tolerated by 39 out of 40 HIV-infected individuals for a 1-month treatment period in Phase I/II clinical trials.[59] The β-hairpin motif in the polyphemusin and T22, which are described in the previous section, was used to design a β-hairpin mimetic POL3026. The cyclic peptide specifically inhibits X4 HIV-1 infection.[62] Importantly, POL3026 showed excellent stability in human plasma and favorable pharmacokinetics when administered subcutaneously in dogs.[63]

Molecular Interactions between CXCR4 Inhibitors & the CXCR4 Receptor

The molecular interactions between CXCR4 inhibitors or antagonists with the CXCR4 receptor have been examined primarily using three methods. The first method involves examining the effect of CXCR4 inhibitors on the binding of several anti-CXCR4 monoclonal antibodies that recognize various regions of the receptor. In the second, the effect of mutations in CXCR4 on the inhibitory activity of CXCR4 antagonists is investigated together with binding of certain anti-CXCR4 monoclonal antibodies, such as 12G5, or the ligand SDF-1α to the receptor. Thirdly, radio or photo-labeled inhibitors are used to identify the binding sites of inhibitors directly.

Using various CXCR4 mutants, the determinants of AMD3100 sensitivity were identified as four Asp residues at positions 171, 182, 193 and 262, as well as Glu288, indicating the importance of the second extracellular loop (ECL)2 and ECL3 and their connected transmembrane domains (Figure 3).[65–68] Reduced binding of radiolabeled T140 was observed when CXCR4 was mutated at Asp171, Arg188, Tyr190, Gly207 and Asp262, again suggesting the contribution of ECL2 to inhibitor binding.[69] Furthermore, T140 photolabeling recently revealed that the peptide actually interacts with the fourth transmembrane domain of CXCR4.[70] To determine the binding site(s) of KRH-3955, its effect on the binding of four different anti-CXCR4 monoclonal antibodies was examined, and the data obtained suggested that KRH-3955 binds a region composed of all three CXCR4 ECLs. Further studies were performed using various CXCR4 point mutants to narrow down the binding site(s) of KRH-3955. While AMD3100 lost its activity when Asp171, Asp262 or Glu288/Leu290 was mutated, unexpectedly, His281 was the only residue whose mutation affected the inhibitory activity of KRH-3955 (Figure 2).[58] Together, these data indicate that several acidic amino acid residues in CXCR4, such as Asp171 and Asp262, are crucial for the binding of most of the CXCR4 inhibitors.

Figure 3.

Serpentine diagram of the CXCR4 receptor.
White letters in blue circles and black letters in red circles represent amino acid residue substitutions that severely reduce the inhibitory activities of AMD3100 and KRH-3955, respectively.
ECL: Extracellular loop.

CCR5-targeted Anti-HIV Therapy

Initially, natural CCR5 ligands, such as MIP-1α and RANTES, or their protein (or peptide)-based inhibitors, were studied as potential CCR5 inhibitors. However, these molecules were not developed further as HIV therapeutics due to their poor pharmacokinetics and bioavailability. Several small molecule CCR5 antagonists have been reported to date (Table 1): TAK-779,[71] TAK-652 (TBR-652)[72] and TAK-220[73] (Takeda; Osaka, Japan), AK602 (presently termed aplaviroc; Ono; Osaka, Japan),[74] SCH-C[75] and SCH-D (renamed vicriviroc)[76] (Schering-Plough/Merck), UK-427,857 (presently termed maraviroc),[77] PF-232798[78] (Pfizer; Sandwich, UK) and INCB9471 (Incyte, DE, USA).[79] Among them, the FDA approved maraviroc in 2007, for treatment-experienced adult patients, in combination with other antiretroviral drugs. Vicriviroc is being terminated in Phase III clinical trials.

Resistance to Coreceptor Inhibitors

Theoretically, HIV-1 acquires resistance to coreceptor inhibitors, including coreceptor ligands or anti-coreceptor antibodies, either by changing the way it uses coreceptors or switching coreceptor usage, such as shifting from CCR5 to CXCR4. Experiments to select viruses resistant to coreceptor inhibitors have been mainly performed in two ways. In one approach, a selection experiment is performed using cell lines that express only CXCR4 or CCR5.[80–84] In this case, resistant viruses did not show coreceptor switching, but rather started to use the same coreceptor in a drug-bound form. Many mutations in resistant HIV-1 were found in the gp120 region of HIV-1 Env, especially in the V2 and V3 regions. It is generally considered that it is much more difficult for the virus to gain resistance to coreceptor inhibitors than to HIV-1 inhibitors that target viral proteins such as reverse transcriptase. Indeed, 145 passages over 1.5 years were required to obtain T134-resistant NL4–3, which shows modest resistance to the T134 peptide (15-fold increase).[84] The T134-resistant HIV-1 is also resistant to AMD3100 as well (15-fold increase). Interestingly, effective concentration of T134 against AMD3100-resistant HIV-1 is almost the same as that against the wild-type strain. In the second resistant virus generation strategy, selection experiments are performed using a cell that expresses both CXCR4 and CCR5, such as human PBMC or PBL.[85–95] Sequence analysis of resistant variants revealed different patterns of amino acid changes in the envelope regions. In some cases, resistant mutations were mainly located in the V3 region of the HIV-1 envelope. Amino acid changes also occurred throughout gp160 in the absence of changes in the V3 loop.[89] Furthermore, detailed analysis revealed that three physically proximal mutations in the gp41 fusion peptide region are responsible for viral resistance to vicriviroc.[93] It is of note that most HIV-1 variants with resistance to CCR5 inhibitors remain CCR5-tropic, although emergence of CXCR4-using variants from SF162 was also reported in cell culture passage in the presence of maraviroc.[94]

Another important issue in viral drug resistance is whether resistance mutations affect virus fitness and/or sensitivity to neutralizing antibodies. Reduced fitness was reported for HIV-1 isolates resistant to CXCR4 inhibitors such as AMD3100.[96] On the other hand, resistance to the CCR5 inhibitor AD101 (SCH-30851, a precursor of vicriviroc) is not associated with a fitness loss.[87] Escape mutants from two different CCR5 inhibitors, AD101 and vicriviroc, were examined for their sensitivity to well-known neutralizing monoclonal antibodies, such as b12, 2G12, 2F5 and 4E10, as well as to sera from HIV-1-infected individuals.[92] The rationale for this study is that at least some of the escape mutants against CCR5 inhibitors change amino acids in their envelope region(s), which can be targets for neutralizing antibodies. Interestingly, the escape mutants were more sensitive than the parental isolates to a subset of the neutralizing antibodies and some sera from HIV-1-infected individuals, indicating that the humoral immune response could exert selection pressure in vivo. Further studies will be required to more completely examine how drug resistance alters virus fitness and/or susceptibility to the immune system.

Several studies have been reported on clinical resistance to CCR5 inhibitors. Coreceptor tropism was examined of 64 HIV-1-infected patients who were given short-term monotherapy of maraviroc. Phylogenic analysis of CXCR4-using variants selected only in two patients revealed that those variants were most likely derived from a CXCR4-using reservoir.[97] Clinical resistance to vicriviroc was examined using both subtype C and subtype D of HIV-1. In the case of an HIV-1 subtype C-infected subject, amino acid changes within the V3 loop were sufficient to confer vicriviroc resistance.[98] On the other hand, amino acid changes in the V3 loop as well as the C4 domain are fully responsible for vicriviroc resistance in HIV-1 subtype D.[99]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: