Abstract

In the past few years, enormous insights have been gained into the mechanisms of Human Immunodeficiency Virus Type One (HIV-1) transmission and primary infection. The discovery of coreceptors as necessary for viral entry has revealed an intricate relationship among virus envelop protein, coreceptor, target cell, and disease state. As early as the late 1980's, cohorts designed for the study of HIV and Acquired Immune Deficiency Syndrome (AIDS) had brought to light individuals who remained seronegative despite multiple exposures to the virus. Now researchers were able to use the plausible idea of mutations in coreceptors to adequately explain this phenomenon. The further discovery of the mutant allele Dccr5 - which encodes the HIV coreceptor CCR5 - and subsequent work supported this idea. Cohort studies revealed that the allele appeared to prevent HIV-1 transmission and primary infection as well as possibly delaying the progression to AIDS. The following review article is thus an attempt to consolidate and summarize the most important findings that came out of the work done on the Dccr5 allele.

Background

The replication kinetics of HIV-1 can be broadly divided into two distinct stages. The first stage encompasses initial transmission of the virus and the asymptotic phase of viral replication. In almost all cases, only M-tropic, non-syncytium inducing (NSI) strains, which infect primary macrophages and lymphocytes, are present in this first stage (Connor, 1994; Roos, 1992; Schuitemaker, 1992). The second and more virulent stage witnesses a decrease in the rate of viral replication. T-tropic, syncytium-inducing (SI) viruses, which infect CD4+ transformed cells and primary lymphocytes but not macrophages, are predominant (Doranz, 1996; Feng, 1996; Schuitemaker, 1992). A majority of the time (50%-60%) thestandard delineation between HIV-1 and AIDS coincides with the delineation between the two stages in HIV-1 replication and tropism in infected individuals (Connor, 1994; Karlsson, 1994; Schuitemaker, 1992). and tropism in infected individuals (Connor, 1994; Karlsson, 1994; Schuitemaker, 1992).

Figure One. The CCR5 Receptor (Wild-Type Protein).
The CCR5 Receptor and Gene
http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?uid1262810&form=6&db=n&Dopt=g

The CCR5 protein is encoded by the CMKBR5 gene, which is located in region p21.3 of human chromosome 3 (Dean, 1996; Samson, 1996c). Like most other chemokine receptor genes, the 1.055 kb CMKBR5 gene contains no introns (Combadiere, 1996; Samson, 1996a). To date, twenty different mutations have been documented in the CMKBR5 gene, sixteen of which (80%) are nonsynonymous or "codon-altering". Seventeen of the twenty mutations are point mutations, with eleven being transversions and six being transitions. Among the other three mutant alleles are a single base pair deletion, a trinucleotide deletion, and a 32-base pair deletion (Dccr5) (Ansari-Lari, 1997; Carrington, 1997). Three of the mutations - the point mutation at position 303, the single-base pair deletion, and the 32-base pair deletion - result in premature truncation of the protein (Ansari-Lari, 1997; Carrington, 1997; Quillent, 1998).

The high frequency of nonsynonymous mutations suggests an adaptive role for the polymorphisms, perhaps a role that emerged as a result of historic selective pressures. The discovery that all missense mutations occurring at highly conserved positions within the CMKBR5 gene were of Caucasian origin further bolsters this possibility. In other ethnic populations, there are dramatically fewer mutations in the CMKBR5 gene, and these mutations demonstrate a more natural distribution throughout the coding region (Carrington, 1997).

The mysterious selective pressure was determined when the frequency of one of the mutant alleles, Dccr5, was found in a north-south gradient in Europe, with the highest frequencies in the countries the farthest north (Libert, 1998; Martinson, 1997; Yudin, 1998). Analysis of flanking microsatellite loci, which were discovered to be in linkage disequilibrium, has narrowed the origin of this mutation to between 2,000 to 4,000 years ago (Carrington, 1997; Libert, 1998). Further mathematical analysis has revealed that the historical pressure that has brought Dccr5 to its high present-day allele frequency occurred in the 1300's - the century of the Black Plague. Researchers believe that the mutant allele Dccr5 conferred an advantage to carriers when the epidemic struck Europe in the mid-1300's. In support of this theory, the severity of the Black Plague in each area of Europe is proportional to that of the present-day frequencies of the Dccr5 allele; the Black Plague also exhibits a north-south gradient (Stephen O'Brien, July 1998, presentation (NIH)). This theory adequately explains the variations in the frequencies of the various polymorphisms in the CMKBR gene among different ethnicities and why the Dccr5 allele constitutes almost all 20% of variations seen in the CMKBR gene (Ansari-Lari, 1997; Carrington, 1997; Libert, 1998).

Figure Two. The Dccr5 Receptor (Mutant Protein).

The approach taken by Dean et al. (1996) that demonstrated the importance of the CCR5 receptor in HIV-1 transmission was the most objective. After discovering the Dccr5 allele, Dean's research team examined 170 loci for genotypic associations between HIV-1-infected versus HIV-1-uninfected individuals. At only one of the 170 loci was there a significant distortion of genotypic frequencies between the two groups: the CMKBR gene (Dean, 1996).

Prompted to confirm these findings, other research teams genotyped individuals in their cohorts and found a significantly elevated frequency of Dccr5 homozygotes among high risk-uninfected or exposed-uninfected individuals (Aarons, 1997; Huang, 1996; Paxton, 1996), in percentages similar to what Liu et al. (1996) and Dean et al. (1996) reported. In the cohort studied by Liu et. al. (1996), which comprised 25 exposed-uninfected individuals, 3 (12%) were homozygous for the Dccr5 allele in comparison to a genotype frequency of 1.4% in the unexposed population (Huang, 1996). In another study, 3.4% (15 of 446 individuals) were homozygous for Dccr5, which was statistically different from the frequency found in the normal population. Furthermore, in the same study, when the genotypes of the 30 participants with the greatest number of homosexual contacts in the previous 6 months were assessed, the frequency of the homozygous Dccr5 genotype rose to 16.7%. The percentage climbed even higher to 25% among high risk individuals who remained uninfected for eight or more years (Paxton, 1996). Another study found a slightly higher frequency, 33.3%, of Dccr5 homozygotes among those at highest risk for seroconversion for the longest period (Huang, 1996). Zimmerman et al. (1996), who reported a statistically significant six-fold increase in the frequency of Dccr5 homozygotes (a rise to 4.5%) in a group of highly-exposed seronegative individuals, concluded that such a distortion of genotype frequencies suggested complete penetrance of the allele in the homozygous condition. Yet, even if penetrance is complete, many exposed-uninfected individuals display the CCR5 homozygous wild-type genotype, suggesting that other unknown resistance factors exist (Zimmerman, 1997).

For many months after, not a single individual homozygous for Dccr5 was found among the tens of thousands of genotyped HIV-1 seropositive individuals (Biti, 1997; Dean, 1996; Huang, 1996; Liu, 1996; Michael, 1997; Paxton, 1996; de Roda Husman, 1997; Samson, 1996b; Zimmerman, 1997); although Zimmerman et al. (1997) found that two individuals, reported to be transiently viremic with HIV-1, were homozygous for Dccr5.

To date, six different seropositive individuals homozygous for Dccr5 have been independently reported (Balotta, 1997; Biti, 1997; Meyer, 1997; Michael, 1998; O'Brien, 1997; Theodorou, 1997). In all individuals in which the virus has been genotyped, only SI variants, which use the CXCR4 receptor, have been found (Balotta, 1997; Michael, 1998; O'Brien, 1997; Theodorou, 1997). Even during the first few years of infection, during which only M-tropic strains have been found in an overwhelming majority of infected individuals (van't Wout, 1994; Zhu, 1993), only SI variants have been discovered in these individuals (Balotta, 1997; Michael, 1998; O'Brien, 1997; Theodorou, 1997). The most accepted explanation for these exceptions is that primary infection occurred by means of SI variants which use the CXCR4 receptor (Bratt, 1998; Carrington, 1997; Michael, 1998). Consistent with this theory (Cornelissen, 1995), two of the Dccr5 homozygotes have experienced rapid progression to AIDS, a characteristic of SI variants (Biti, 1997; Michael, 1997). Despite these rare cases, the Dccr5 homozygous condition may be said to confer protection against transmission.

In vitro studies of the mutant allele Dccr5 further bolster the evidence that the homozygous condition confers resistance against transmission. In fusion assays, cells from Dccr5 homozygotes fail to support cell-cell fusion mediated by M-tropic strains but not by T-tropic strains (Aarons, 1997; Connor, 1996; Dragic, 1996; Huang, 1996; Liu, 1996; Paxton, 1996; Picchio, 1997; Rana, 1997; Samson, 1996b). Paxton et al. (1996), however, found that a few cells from Dccr5 homozygotes could be infected by M-tropic isolates, but 1000-fold more virus was required to establish infection, and, once infected, these few cells failed to replicate further. Liu et al. (1996) reported similar findings. Perhaps this phenomenon could explain the transient viremia that Zimmerman et al. (1997) witnessed in two Dccr5 homozygotes.

In regard to dual-tropic strains, Samson et al. (1996b) found that cells homozygous for Dccr5 failed to support infection by dual-tropic strains, but Rana et al. (1997) and Zimmerman et al. (1997) found no differences in infectibility between M-tropic and dual-tropic strains.

Non-permissive cells transfected with CCR5 cDNAs from wild-type homozygotes strongly support viral entry of M-tropic and dual-tropic isolates, and while those transfected with CCR5 cDNAs from Dccr5 homozygotes do not support viral entry of M-tropic isolates and weakly support viral entry of dual-tropic isolates (Liu, 1996; Rana, 1997; Zimmerman, 1997). The fact that transfection renders non-permissive cells permissive to viral infection also confirms that the CCR5 receptor is the major coreceptor required by HIV-1 for entry. Lastly, RT-PCR analysis reveals that CCR5 mRNA is expressed at about the same level or greater in the Dccr5 homozygous condition as it is the CCR5 homozygous condition (Liu, 1996).

The block to HIV-1 infection therefore lies at the level of virus fusion (Connor, 1996). In conjunction with the fact that the Dccr5 protein does not fuse with the cell membrane (Liu, 1996; Rana, 1997; Samson, 1996b), these studies and a study reporting that neutralizing antibodies to the natural ligands of the CCR5 receptor do not increase infectibility (Dragic, 1996) rule out the possibility that the block to infection is due to increased ligand secretion of MIP-1a, MIP-1b, and RANTES. Increased ligand secretion has been shown to suppress HIV-1 infection by M-tropic strains (Alkhatib, 1996; Cocchi, 1995; Deng, 1996; Dragic, 1996; Paxton, 1996). In some highly exposed-uninfected individuals, up to a 10-fold greater secretion of the ligands has been observed, which is perhaps a result of the loss of negative inhibition (Dragic, 1996; Liu, 1996; Paxton, 1996).

An in vivo murine study has also confirmed the differential resistance of Dccr5 homozygotes to infection by M-tropic isolates but not to T-tropic isolates. In addition, infection with the dual-tropic isolate 89.6 showed a 10-fold-reduced plasma viremia seven days post-infection (Picchio, 1997).

Transmission and Primary Infection in Dccr5 Heterozygotes

Whether heterozygosity for Dccr5 confers protection against HIV-1 infection has generated controversy. Some studies have found evidence that heterozygosity confers partial protection against transmission (Husain, 1998; Liu, 1996; Malo, 1997/8; Michael, 1997; Samson, 1996b), while other studies have found no significant differences in infectibility between the homozygous wild-type genotype and the heterozygous genotype in both adults (Balotta, 1997; Dean, 1996; Eugen-Olsen, 1997; Huang, 1996; Malo, 1997/8; Paxton, 1998; Quillent, 1997; Zimmerman, 1997) and children (parental transmission) (Misrahi, 1998). Studies supporting partial protection have found a significantly reduced frequency of heterozygotes among seropositive individuals in comparison to the general population or control groups of unexposed individuals (Malo, 1997/8; Michael, 1997; Samson, 1996b). Samson et al. (1996b) was the first research group to find evidence for partial protection, reporting both a 35% decrease in Dccr5 heterozygous genotype among seropositive individuals and a reduced efficiency in supporting virus fusion in cells from Dccr5 heterozygotes. Another study reported that the frequency of the Dccr5 allele was significantly higher among seronegative individuals (12.5%) than among seropositive individuals (7.1%); in high risk seronegative individuals, the frequency of the mutant allele was 23.8% (Michael, 1997). Liu et al. (1996) found that the efficiency of virus replication was significantly reduced in cells from Dccr5 heterozygotes in comparison to virus replication in cells from homozygous wild-type individuals.

In looking at plausible explanations for the discrepancies, an in vitro examination of a proposed transdominant effect in heterozygotes has shown a moderate role for transdominance in fusion inhibition (Husain, 1998). Liu et al. (1996), however, added inactive CCR5 vectors into transfected cells and did not observe a transdominant effect. Perhaps variability in expression could resolve the issue since among CCR5 wild-type individuals receptor expression may vary up to 20-fold (cited in Bratt, 1998). In a study conducted by Picchio et al. (1997), virus infectivity by M-tropic strains in two out of three heterozygotes was comparable to that in wild-type homozygotes, but in the third heterozygous individual, poor virus replication was observed. Huang et al. (1996) and Liu et al. (1996) also found similar variability, although not quite as great in magnitude. Rana et al. (1997) did not find any difference in virus replication between cells from wild-type homozygotes and Dccr5 heterozygotes, but suggested that perhaps a difference could manifest itself at lower concentrations of virus.

Hoffman et al. (1997), though, may provide the most interesting explanation. Hoffman and his colleagues have discovered a significantly increased frequency of the heterozygotes genotype in exposed-uninfected individuals (28%) versus infected individuals (11%) in sexual relationships in which one partner was seropositive while the other remained seronegative (Hoffman, 1997). When they examined the genotypic distribution further by comparing acquisition solely through heterosexual intercourse and acquisition through homosexual intercourse, they found that heterozygosity conferred no significant protection against transmission through homosexual intercourse (28% uninfected versus 20% infected) but appeared to confer protection against transmission through heterosexual intercourse (27.6% uninfected versus 0% infected); unfortunately, this difference was not significant due to the small number of heterosexual individuals involved in the study (Hoffman, 1997). In many of the largest studies, cohorts comprise mostly or entirely homosexual individuals (Dean ,1996; Huang, 1996; Michael, 1997). The cohort examined by Samson et al. (1996b) who reported that Dccr5 heterozygosity confers partial protection against infection included individuals infected through mechanisms other than homosexual intercourse (cited in Hoffman, 1997). In support of this argument, Malo et al. (1997/8) found a significantly reduced frequency of Dccr5 heterozygotes among individuals infected through intercourse but not among individuals infected through contaminated blood products, in whom, frequently, progression to AIDS occurs more rapidly, thus suggesting the plausible rare transmission of SI variants. This, together with the fact that almost all studies on HIV-1-infected individuals show delayed progression to AIDS, higher CD4+ T cell counts, lower viral load, a lower mortality rate, or at least one of the above among Dccr5 heterozygotes (Bratt, 1998; Dean, 1996; Doranz, 1996; Eugen-Olsen, 1997; Huang, 1997; Liu, 1996; Meyer, 1997; Michael, 1997; Misrahi, 1998; Paxton, 1998; Rappaport, 1997; de Roda Husman, 1997; Samson, 1996b; Stewart, 1997; Zimmerman, 1997), supports the idea that heterozygosity confers partial protection either against viral infection or viral replication.

Applications

The findings on Dccr5 and its role in HIV-1 transmission and primary infection provide a target for HIV-1 therapies and vaccines. Among other venues, researchers have investigated ways to produce effective chemokine receptor antagonists and ways to reduce surface expression levels of the HIV-1 coreceptors. Although no one approach has yet led to a solution, this work appears promising since individuals with one or copy copies of the Dccr5 allele are healthy and exhibit no observed phenotype.