CHEMOKINE RECEPTORS AND HUMAN IMMUNODEFICIENCY VIRUS INFECTION

Paul D. Bieniasz and Bryan R. Cullen

Department of Genetics and Howard Hughes Medical Institute, Duke University Medical Center, Durham, NC 27710

Received 12/15/97 Accepted 12/19/97

6. PLASTICITY OF HIV CORECEPTOR USAGE - IMPLICATIONS FOR PATHOGENESIS AND TREATMENT

6.1 Coreceptor targeted inhibitors of HIV entry

The identification of coreceptors that mediate primate lentivirus entry might well have major significance for attempts to treat HIV infection. The chemokines themselves are inhibitors of HIV entry (21-23), and modified forms of RANTES, which bind CCR-5 yet fail to induce signal transduction, have similar properties (83, 84). Recently, beta chemokine homologues encoded by Kaposi's sarcoma associated herpesvirus have also been shown to block HIV-1 infection (85, 86). The lack of an overt phenotype associated with CCR-5 delta32, even among homozygotes (40, 42), suggests that antagonists targeted specifically to this coreceptor should be relatively free of undesirable side effects. However, this is less likely to be the case for CXCR-4 targeted antagonists; mice in which the SDF-1 has been ablated have a lethal defect in B cell lymphopoesis (87). Nevertheless, it might be possible to develop small molecule inhibitors that block viral interactions with CXCR-4 without compromising SDF-1 signal transduction, Peptide and bicyclam compounds targeted to CXCR-4 that prevent T-tropic HIV infection have been described (88-90). However, all of these compounds also block signalling through CXCR-4. A potentially even more serious obstacle in the development of coreceptor targeted threrapeutics is the plasticity of envelope coreceptor interactions. It is quite conceivable that coreceptor targeted compounds would simply select for strains that use an alternative coreceptor, or different regions of the same coreceptor. Furthermore, since AIDS has been documented in CCR-5 delta32 homozygotes, it is unlikely that even total ablation of the CCR-5 entry pathway will achieve much more than retardation of disease progression, once infection has been established. It is even possible that CCR-5 targeted therapy might accelerate the course of disease by selection of viral strains that use alternative coreceptors (most notably CXCR-4, but also including CCR-2b and CCR-3) whose occurrence is associated with disease progression (91). It remains possible, however, that combinations of agents that ablate the coreceptor function of both CCR-5 and CXCR-4 would have a significant impact on the ability of HIV to propagate in vivo.

6.2 Changing patterns of coreceptor usage during HIV-1 disease progression - cause or consequence of immune suppression?

It is interesting to speculate as to why primate lentiviruses have evolved plasticity in their interactions with coreceptors. The ability to use utilize multiple functionally redundant contacts with coreceptors could conceivable facilitate immunological escape. Thus, in the face of a neutralizing antibody response (which, in significant part, is directed against V3 sequences (92), a major modulator of coreceptor interaction), the selection of variants with altered envelope sequences would be permitted without compromising the ability of the virus to use a given coreceptor. Furthermore, changes in envelope sequence that enable the virus to use additional coreceptors, while retaining the ability to interact with CCR-5, could be tolerated. It appears that dua-tropic strains that use both CCR-5 and CXCR-4 are less tolerant of perturbations in CCR-5 sequence than are M-tropic strains (71, 72), suggesting that acquisition of the ability to utilize CXCR-4 might involve the sacrifice of a degree of functional redundancy and/or affinity in the envelope/CCR-5 interaction.

During the early, asymptomatic phase of HIV-1 infection M-tropic strains predominate. It is quite possible that this is the simply consequence of a 'founder effect': It is likely that the major route of transmission of HIV-1 is via infection using CCR-5 as a coreceptor, given the observation that CCR5 delta32 homozygous individuals are much less likely to be HIV infected than other genotypes. However, why do strains that also use CXCR-4 tend to arise only late in the course of infection, when in many cases they retain the ability to use CCR-5? While it has often been noted that SI viruses are associated with poor prognosis it is important to emphasise that the occurrence of SI viruses that can use CXCR-4 (and usually CCR-5) is only a correlate of disease progression (91, 93). We can as yet, draw no conclusions about causality. Although it is possible that a presumptive expansion in tropism might accelerate the course of HIV-1 disease, particularly since CXCR-4 is expressed on the naive T-cells subset from which CCR-5 is largely absent (94), it is equally possible that CXCR-4 utilizing viruses arise as a consequence of, rather than being the cause of immunosuppression. It appears that in addition to the selective transmission of M-tropic viruses, that there may, in fact, be an advantage to the virus in using CCR-5, at least early in infection. Indeed, viruses isolated from a laboratory worker who was accidentally infected with the exclusively T-tropic, IIIB strain became M-tropic with time (95). Although the coreceptor utilization profile of these strains have not yet been documented, a selective advantage for CCR-5 utilizing viruses early in infection is suggested. In addition, a chimeric virus (containing HIV-1 derived envelope sequences in an otherwise SIV genome) replicates to higher titers in macaques if the envelope is derived from the dua-tropic 89.6 strain than does a chimera containing the envelope of the T-tropic strain IIIB (96). Since precisely the opposite observation was made upon cultivation of these viruses in cultured macaque PBL, it is clear that the selective pressures on coreceptor utilization are different in vivo as opposed to in vitro. At least two scenarios could be envisaged that provide CCR-5 utilizing virues with a selective advantage: Firstly, it is conceivable that a major selective influences are the result of envelope directed neutralizing immune responses, which could impose a restriction on the nature of coreceptor utilization. Indeed, neutralization phenotype and viral tropism (and therefore, presumably, coreceptor selection) are not independent properties of the viral envelope (97). A. general observation is that primary HIV-1 isolates (which tend to use CCR-5 but not CXCR-4) are somewhat more resistant to neutralization by naturally occurring HIV-1 antibodies than are laboratory adapted strains (98). Thus, it is possible that immunological constraints are placed on envelope sequence and conformation that could negatively influence the ability of the virus to use CXCR-4 as a coreceptor. Therefore, the occurrence of CXCR-4 utilising variants late in the course of disease might be a reflection of the infected individuals compromised ability to mount immune responses to new variants that are capable of utilising CXCR-4 that would have been readily neutralized earlier during the infection.

Conversely, a selective advantage for viruses which use CCR-5 may be immune response-independent. An equally plausible hypothesis is that CXCR-4 utilizing strains only become prevalent late in disease as a result of the selective depletion of CD4+ memory T-cells (the subset that preferentially expresses CCR-5). These are replaced at a high rate by naive counterparts which, (at least in uninfected individuals) express higher levels of CXCR-4 than of CCR-5 (94), thus providing T- or dual-tropic strains with a selective advantage.

Regardless of the selective pressures on coreceptor usage, the occurrence of SI viruses in only about 50% of patients with advanced disease indicates that the ability of the virus to use CXCR-4 is by no means necessary for the onset of immune suppression. Conversely, the fact that infected CCR-5 delta32 homozygotes also have progressive disease (43-46) indicates that disease causation cannot be universally ascribed to viruses that use this coreceptor either.