OUP user menu

HIV and the CCR5-Δ32 resistance allele

Eric de Silva, Michael P.H. Stumpf
DOI: http://dx.doi.org/10.1016/j.femsle.2004.09.040 1-12 First published online: 1 December 2004


The combination of molecular biology, epidemiology, virology, evolutionary and population genetics has enabled us to understand the delicate interplay between HIV and the CCR5-Δ32 HIV resistance allele. We here review and collect from the different approaches to show how they can be combined to elucidate the interaction between host and pathogen genetics in this system. We will present an overview of the normal role of CCR5, its involvement in HIV, the molecular biology of the CCR5-Δ32 allele and its probable origins. By focusing on this well-documented and important system we hope to demonstrate the power that such a “holistic” approach might offer in the study of infectious diseases.

  • Chemokines
  • Host–virus interactions
  • Population genetics

1 Introduction

Standard epidemiological models have often made very simplified assumptions about genetic heterogeneity in the susceptibility of the host population. While the mathematical formalism has been well developed for some time [1], the genetic contributions to host-heterogeneity are only beginning to be appreciated in the wake of the human genome project and the progress seen in the field of human population genetics. Understanding the mechanisms by which a host's genotype influences infectious disease dynamics is important for a number of reasons. First, heterogeneity in disease transmission or susceptibility can affect a number of crucial epidemiological parameters such as the potential number of hosts or the number of latent and non-clinical carriers of an infectious agent. Second, if some individuals are (partially) resistant to an infectious agent then their inclusion into clinical or epidemiological studies will skew the results. This in turn could potentially lead to misleading guidelines in preventative epidemiology. Third, understanding the molecular and genetic causes of differential disease susceptibility may point towards novel therapeutic approaches. Finally, study of the disease-causing polymorphisms can provide detailed insights into the evolutionary history of our own species. Here the situation may become rather complicated indeed. Following the pioneering work of Bartlett [2] we know that childhood diseases, which are followed by lifelong immunity after initial infection, require quite substantial population sizes for the pathogen to persist over time, unless a different species can act as a reservoir. Thus to some extent exposure of humans to pathogens may well have changed over say the last 100,000 years; diseases will no doubt have benefited from urbanization (e.g. for the first time in Babylon some 3500 years before present) and the apparent increase in emergent infections could just be a natural progression from such early beginnings. As such the study of the host contribution to infectious diseases may predominantly shed light on relatively recent phenomena in human history.

A host of genetic factors have been identified which contribute to differences in disease susceptibility, progression or pathology between individuals following challenge by a pathogen. Sickle-cell anaemia has become the canonical example and is discussed in great detail in the classic text by Bodmer and Cavalli-Sforza [3]. Other examples include the human Prion gene [4], and countless polymorphisms [5] in the HLA which for example show clear association with disease progression phenotypes, e.g. following infection by human T-lymphotropic virus type 1 (HTLV-1), patients with the promoter TNF-863A allele show faster progression to Human T lymphotropic virus type I associated myelopathy/tropical spastic paraparesis (HAM/TSP) [6].

One of the best documented genetic polymorphisms which regulates disease susceptibility is the CCR5-Δ32 HIV resistance allele [7]. This deletion polymorphism has been demonstrated to affect disease progression following challenge with HIV: individuals homozygous for the allele appear to be protected from HIV infection while progression to disease is delayed in heterozygous carriers of the allele. We believe that the study of CCR5-Δ32 is important both because of its clinical importance on HIV disease dynamics and as a case study in the co-evolution of a parasite–host system. Here we will review what is known about this allele and the corresponding wildtype, their respective roles during HIV infection and the evolutionary consequences of the CCR5-Δ32 allele for the virus and the host. We will conclude by discussing the potential importance of the insights gained from studying the CCR5-Δ32 HIV resistance allele has for the development of anti-retroviral drugs.

2 The human CCR5 gene

The chemokine CCR5 plays an important part in the immune system, where it is highly expressed on macrophages and CD4 T cells [8]. Chemokines or chemoattractant cytokines – generally recruit and activate leukocytes to and at sites of infection [9]. Motion along a gradient of chemokines allows cells to migrate in a particular direction. They are the only cytokines that interact with seven-transmembrane-domain-G-protein-coupled receptors (proteins that respond to an external signal and transmit it to an intracellular signal) [10]. Chemokines are labelled according to the structure at conserved sites of their cysteine residues such that CC chemokines (β-chemokines, attract mostly neutrophils) have neighbouring cysteines while CXC chemokines (α-chemokines, mostly attract monocytes and lymphocytes) have an additional amino acid separating cysteine residues. To date seven CC (of which CCR5 is one) and four CXC receptors are known (see the regularly updated http://csp.medic.kumamoto-u.ac.jp/). It appears that a primordial chemokine gene that underwent duplication and divergence is responsible for the different chemokine groups observed today [9].

Understanding the mechanics of chemokine action is complicated by the fact that chemokine receptors can bind to any number of chemokines, while chemokines in turn can bind to any number of chemokine receptors. Chemokine receptors are expressed on a range of cells of the immune system such as lymphocytes, monocytes, macrophages, eosinophils, basophils and platelets [10]. For instance CXCR4 (formerly referred to as fusin) receptors are expressed on T cells and CXCR5 receptors on B cells [10]. The role of CC chemokines is accordingly complex and diverse, for example monoclonal antibodies such as MIP-1α (see below) have been used in inflammatory models highlighting the function of chemokines in inflammation. Also, the use of chemokine receptors in locating antagonists (chemical inhibitors) is being actively pursued in HIV therapeutics. A good description of CC polymorphisms is given by Dean et al. [11].

2.1 Chemokine receptors and their role in infectious diseases

Chemokine receptors often take a central role in infectious disease progression as they are involved in enabling or facilitating viral entry into cells; overexpression of receptors can for example exacerbate their effect on disease progression in some cases [12]. The best documented involvement of chemokine receptors in disease is during the course of HIV infection which is discussed below in some detail. However, other instances of their importance for disease dynamics are also well known. Two of the parasites involved in Malaria enter erythrocytes via binding to the Duffy antigen receptor for chemokines (DARC) [13, 14]. There is evidence that HTLV-1 Tax protein is involved in the process of transactivating chemokine receptor gene promoters and so upregulating the cell-surface expression of CXCR4 and CCR5 [15]. Interestingly, cytomegalovirus encodes a chemokine (U238) that is related to chemokine receptors in humans [13].

Inflammatory diseases such as rheumatoid arthritis, asthma, and inflammatory bowel disease (IBD) seem most likely to involve CCRs and indeed CCR5 has been implicated in many such diseases as detailed in Section 3.2. The role of chemokine receptors in central nervous system infectious diseases (CNS), such as HIV-1 encephalitis and bacterial meningitis, has recently been discussed in detail [16]. Chemokine receptors have also been implicated in various other disease states such as psoriasis [17] and atherosclerosis [18]. For a breakdown of individual chemokine receptors and their properties as well as their role in inflammation see Murdoch and Finn [14].

2.2 The molecular genetics of CCR5 in health and disease

The gene for CCR5 is found on chromosome 3 in the p21.3–p24 area. In the coding region of CCR5 there are known to be at least 23, mostly rare, alleles. The majority of mutations on CCR5 are nonsynonymous (amino acid altering) implicating past selection pressures [19]. A number of SNPs (five common) have been reported in the cis-regulatory region of CCR5 resulting in six common haplotypes [12]. The five most common CCR5 promoter and/or cis-regulatory region polymorphisms have been shown to be in strong linkage disequilibrium with one another as well as with the CCR5-Δ32 deletion allele [20]. Further, a SNP in CCR5's promoter region has been shown to be in strong linkage disequilibrium with the 64I allele in CCR2 [21]. (The CCR2 gene has two forms – CCR2A and CCR2B – due to alternative splicing, with identical ligand binding specificities. Its most important ligand is monocyte chemoattractant protein-1 (MCP-1) which mediates monocyte chemotaxis responding to diseases such as rheumatoid arthritis. The CCR2-64I allele, which is involved in delayed progression to AIDS in HIV patients is reviewed by Berger et al. [8].) In total, nine CCR5 haplogroups (referred to as A, B, C, D, E, F*1, F*2, G*1, and G*2) have been identified (G*2 is the CCR5-Δ32 allele) [22, 23].

In repairing damaged tissue, macrophage inflammatory protein 1α (MIP-1α), macrophage inflammatory protein 1β (MIP-1β) and regulated on activation normal T expressed and secreted protein (RANTES) all latch onto CCR5 leading it to the correct site. (CD4 memory T cells are selectively recruited by RANTES). For an illustration of CCR5 structure (which does not concern us here) see figure 1 in McNicholl et al. [24]. It shows the three extracellular loops associated with chemokine binding, the seven transmembrane domains and the intracellular loops associated with cell signalling. The G-protein coupled interactions control the interaction between the receptor and its ligands.

3 CCR5-Δ32

This allele has received much attention (see for example [25] and references therein) due to the fact that individuals homozygous for this mutation are resistant to infection by the HIV-1 virus while those heterozygous for this allele who are HIV-positive have a delayed onset to AIDS of 2–3 years [26]. Homozygosity of CCR5-Δ32 in Caucasians was measured to be about 1%, with heterozygosity being anything up to 20%[25].

With the removal of 32 nucleotides from the CCR5 gene the translation machinery encounters a stop codon too soon, resulting in a truncated version of the peptide being manufactured [23]. It no longer has the final three transmembrane domains and lacks both extracellular and intracellular loops. This mutated version of the protein remains in the cytoplasm, unable to migrate to the cell surface. Attention to this mutation first increased with its discovery in individuals resistant to HIV-1 infection [27].

3.1 The CCR5-Δ32 allele and HIV

Infection by the HIV-1 virus can be roughly divided into two stages. In the first, the so-called M-tropic phase, the transmembrane glycoprotein gp120 on the surface of the HIV-1 virus binds to the CD4 receptor and CCR5 coreceptor of macrophage cells [28] thereby gaining access to the inside of these cells (see Fig. 1). Inside the macrophages the virus typically produces billions of virions per day [12] which can remain hidden in these cells for several years without apparently harming them [28]. Regions of the gp120 protein then undergo mutation until they are able to bind more efficiently to the CXCR4 coreceptor of CD4 T lymphocytes (as opposed to the CCR5 coreceptor of macrophages) and thus infiltrate these cell types as well [28]. This latter stage is thus referred to as the T-tropic phase which results in the increasing destruction of the T-lymphocyte pool. As the number of CD4 T cells drops – from about 1000 – to 200 mm3 of blood [27], and since they are responsible for the body's immune response, it is this phase that leads to immunodeficiency and the ensuing onset of opportunistic infections: AIDS. It has been estimated that ninety percent of all HIV-1 infections are due to the M-tropic strain [12]. In vitro, T-tropic viruses (sometimes called X4 strains as they bind to CXCR4 [28]) induce the production of large multinucleated cells; these viral strains are therefore also known as syncytium-inducing virus [28]. Likewise M-tropic viruses (sometimes called R5 strains due to CCR5 binding) are referred to as non-synctium inducing. 30–50% of infected (M-tropic) persons have HIV-1 isolates which then use CXCR4 as a coreceptor (X4 strain) or use both CCR5 and CXCR4 coreceptors (known as R5X4 strains). Although an individual can progress to AIDS without these variants [29] they are related to an accelerated progression to AIDS [29] and even without them the R5 strain remains so CCR5 is always involved [29].

Figure 1

The primary components involved in the infection of target cell by HIV-1. Gp120 binds to the CD4 receptor and heparan sulphate proteoglycan (which aids binding and stabilises it) resulting in conformational changes allowing now exposed regions of gp120 to bind with a coreceptor (such as CCR5 in the M-tropic case, not shown here). The V3 region of gp120 binds to the receptor (CXCR4 in the T-tropic case above) and gp120 docks opening the way for gp41 to complete fusion.

Individuals homozygous for CCR5-Δ32 cannot produce complete CCR5 and their cell surfaces are therefore devoid of this receptor; no other side-effects have been reported [19]. Thus the HIV-1 virus cannot attach to these cells and subsequently enter the cytosome [30]. This in turn makes the CCR5-Δ32 homozygous individuals resistant to infection by HIV-1 [31]. Heterozygous carriers of the CCR5-Δ32 express less functional receptors on the cell surface which appears to slow down the progression to AIDS by 2–3 years [26]. While CCR5 is to be found on the cell surface in homozygous carriers of the wildtype, in CCR5-Δ32 homozygous individuals the truncated protein product is held in the endoplasmic reticulum [30]. Before arriving at the cell surface CCR5 is known to pass through the endoplasmic reticulum and while CCR5 can be post-translationally modified by phosphorylation, CCR5-Δ32 cannot [30]. In heterozygous individuals both alleles are expressed, but in addition to there being only half as much of the normal allele being produced CCR5-Δ32 also seems to reduce cell surface expression of the normal allele by holding the CCR5 in the endoplasmic reticulum [30]. In this way normal CCR5 can no longer mediate HIV-1 infection which explains the delayed progression of the disease seen in such individuals [26]. Thus, the mutant CCR5-Δ32 is lacking certain post-translational abilities resulting in it being confined to the endoplasmic reticulum and causing it to form heteroplexes with the normal CCR5 product, holding it back as well. All this has been observed by Benkirane et al. [30] who used confocal immunofluorescence to visualise potential intracellular influences of CCR5-Δ32 on CCR5. They further find that phosphorylation (but not expression) of the CCR5 protein is influenced by MIP-1β. If CCR5 is stimulated by MIP-1β after it has been translated phosphorylation takes place. It is interesting that of all the CCR5 ligands (MIP-1α, MIP-1β and RANTES) MIP-1β is the only one that is not used by any other chemokines: it is unique to CCR5 [9]. MIP-1α and RANTES for instance, also bind to CCR1 [10]. Very recently Agrawal et al. [32] have argued that CCR5-Δ32 confers resistance not just because CCR5 does not reach the surface of the cell, but also because the CCR5-Δ32 truncated protein is able to downregulate the expression of any CCR5 and CXCR4 on the cell surface by scavenging them.

There is however some evidence that reduced CCR5 dosage rather than CCR5 hijacking in CCR5-Δ32 heterozygotes is responsible for measured lower levels of surface CCR5 [33]. Other work has shown that depending on the mRNA ratio of CCR5-Δ32 to the wild-type in heterozygotes, protection from the HIV-1 virus differs between peripheral blood lymphocytes and monocyte-derived macrophages [34]. HIV-1 is able to infect the central nervous system and this fits with the expression of CCR5 in microglial cells [35].

Another chemokine receptor implicated in delayed progression to AIDs is CCR2B. A single nucleotide mutation (valine replacing isoleucine) leads to the V64I polymorphism in the corresponding protein product. This non-synonymous polymorphism has an allelic population frequency of 10–15% among Caucasians and African Americans, and appears to delay progression to AIDS by 2–4 years [36]. This delay is stronger in Africans probably due to the lack of CCR5-Δ32 in African populations. Furthermore, CCR5 has also been shown to be linked to the other chemokine receptors CCR1, CCR2, CCR3 and CCR4 [25]. Two SNPs in the RANTES promoter region have been examined and found to be associated with slower rates of CD4 fall-off as well as reduced risk of HIV infection [37]. Further, the distribution of CCR5 and its ligand MIP-1β and RANTES haplotype pairs are population dependent and complex in their protection against HIV-1 infection and progression to AIDs [23]. It appears that for envelope-mediated membrane fusion of M-tropic HIV-1 and SIV (Simian immunodeficiency virus) to CCR5, aspartic acids at locations 2 and 11 and glutamic acid at location 18 of the CCR5 amino acid terminus are necessary [28], see Fig. 2. O'Brien and Moore [12] discuss possible epistasis between CCR5, CCR2 and SDF (the ligand for CXCR4) and their possible joint role in delayed AIDS progression. It is also worth noting at this point that five to ten percent of HIV-infected people do not progress to AIDs for 15 years or more (see http://www.unaids.org). It has been suggested that polymorphisms in the CCR5 promoter region may account for this variability via differences in the amount of CCR5 transcription [22, 21].

Figure 2

Illustrates the 32 bp deletion from CCR5.

3.2 The CCR5-Δ32 allele and other diseases

Following the observation of the role of CCR5-Δ32 in HIV infection there has been considerable interest in the potential involvement of CCR5 in other diseases [11]. Given that blocking the CCR5 receptor could offer an extremely attractive therapeutic approach to control HIV there have been concerns about the overall quality of life of people without functional CCR5 on their cell surface [12]. Since homozygous individuals not appear to suffer adverse effects such concerns seem to have been largely unfounded. The fact that CCR5 absence is not detrimental suggests considerable levels of redundancy among chemokine receptor function [12]. As far as the virus is concerned in fact HIV is able to utilise CCR2B and CCR3 as coreceptors but they are not as effective as CCR5.

In light of the prominent role which CCR5 plays in the immune system it is not surprising that a large number of investigators have examined its role in other diseases and illnesses, in mice as well as humans [11] and in vitro as well as in vivo [28]. CCR5 knockout in mice has in some cases resulted in animals more susceptible to certain infections or higher mortality as a result of certain infections [12]. Table 1 summarises much of the work which has focused on humans. It is important to remember, however, that some of the results are contradictory or ambiguous.

View this table:
Table 1

Table summarising the results of research into the effects of CCR5-Δ32 on the resistance to contraction or post-contraction progression of various diseases

HIV-1[38, 39, 40]CCR5-Δ32 homozygosity slows replication kinetics of peripheral blood mononuclear cells
Breast cancer[42]
Childhood asthma[44, 45][46]
Hepatitis C[48]
Adult cytomkegalovirus[49]In HIV-infected patients with hemophilia
Homozygous sickle cell disease[50]CCR5-Δ32 allele reduces inflammation in patients with the disease
Liver transplantation outcome[51]
Multiple sclerosis[52, 53]In [52] a modulation of severity in MS by CCR5-Δ32
Pulmonary sarcoidosis[54, 55]In [55] evidence that CCR5-Δ32 allele increases susceptibility
Chagas’ disease[56]CCR5-Δ32 allele frequency too low in case/control populations
HIV-1 disease progression[5760][20]In [58] applies to HIV-1 infected children undergoing antiretroviral treatment
Coronary artery disease[61]
Crohn's disease[62][63]Homozygous CCR5-Δ32 patients have longer survival of renal transplants
Inflammatory bowl disease[64]
Placental trophoblast HIV-1 infection[65]Refers to CCR5-Δ32 state of infant
Insulin-dependent diabetes mellitus[66]Study involved children only
  • Yes: CCR5-Δ32 allele confers resistance, No: CCR5-Δ32 allele does not confer resistance, Ambiguous: current evidence is ambiguous as to whether CCR5-Δ32 allele confers resistance or not.

4 Population genetics of CCR5-Δ32

One aspect of the CCR5-Δ32 allele which has attracted much interest recently [25, 67, 68, 24, 11] is that the allele frequency varies among populations living in different geographical areas, in particular a north–south cline in the frequency of the deletion in Europe. It is of fundamental evolutionary interest to determine whether such a distribution could have arisen by chance (e.g. through genetic drift in structured populations) or if it has been caused by natural selection. If a purely neutral model of evolution [69, 70] can be ruled out then the next important and obvious question regards the cause of a potential selective cause. In this section we will review recent work regarding the human population genetics of the CCR5-Δ32 allele.

4.1 Geographic distribution of the CCR5-Δ32 allele

Fig. 3 shows the population frequencies of the CCR5-Δ32 allele across Europe. Allelic frequencies are found to be high in Northern Europe and gradually decrease as one moves further south until (off the map in Fig. 3) there are no observed CCR5-Δ32 alleles. In particular the allele has been absent in population samples from Lebanon, Georgia, Saudi Arabia, Korea, main land China, and among American Indian population samples where individuals belonging to the Cheyenne, Pueblo, and Pima ethnic groups have been studied. Furthermore the allele also appears to be absent from Venezuelan, Japanese and black West African populations [71, 72, 39]. This raises the interesting question as to what in human history might have caused such patterns of the allele distribution.

Figure 3

Gradient of CCR5-Δ32 mutation [27].

It has previously been suggested that this mutation originated within Nordic populations (where its frequency is highest today) and that it spread south-wards due to the movement of the Vikings in the eight to the tenth centuries [68] resulting in the geographic frequency distribution now observed. While this potentially explains the north–south gradient it does not explain the causative agent that forced this mutation to reach such high levels in the first place (if there was one). The ability to separate between a purely migratory model and a selective-pressure scenario has been discussed and it has been argued [73] that the geographical distribution of the allele and its age need be considered together if one wants to make progress. Use of all the CCR5 alleles jointly [74] from various latitudinal European populations in a principal component analysis could indicate if the observed pattern is due to migratory effects [75].

4.2 Allele age estimates

Stephens et al. [25] considered over 4000 individuals from 38 ethnic populations. Because of the absence of the CCR5-Δ32 allele in so many global populations we can almost certainly rule out recurrent mutations and can safely assume that it originated from a single deletion event (this is also supported by the analysis of the CCR5 haplotype network [22]). Stephens et al. [76] used the intra-allelic variability at two nearby microsatellite loci to determine the time back to the most recent common ancestor of all individuals carrying the CCR5-Δ32 allele in their sample. Their point estimate for the age of the most recent common ancestor is approximately 700 years before present (with a heuristic confidence interval ranging from 275 to 1875 years before present).

Libert et al. [35] obtained a different estimate using the haplotype of very polymormphic microsatellites upstream and downstream from CCR5. They estimated the time since the deletion which lead to CCR5-Δ32 in the European population using microsatatellite mutation rates to be 1400 years and crossing-over rates to be 3400 years. They further suggest that CCR5-Δ32 may have occurred more than once in Europe as there is a 10 bp repeat next to the CCR5-Δ32 bp deletion.

4.3 Has the allele been under selection?

As outlined above and also indicated by its restricted geographic distribution the CCR5-Δ32 allele appears to be of relatively recent origin. Although we know only relatively little about the precise demographic processes in Europe over the last 1000 years or so (such as founder effects and patterns of migration/gene-flow), the relatively large frequency of the allele in Northern Europe could suggest that selection may have favoured CCR5-Δ32 in the past.

Before considering this scenario there are however other possibilities that could have given rise to the observed north–south frequency cline in the geographic CCR5-Δ32 in Europe. If the allele originated in northern Europe then it may have increased in frequency by drift and dissipated south by migration/gene flow. It has been argued that the case of a north–south selection differential and gene-flow from north to south by migration cannot easily be distinguished from the genetic data alone. There is moreover a third plausible scenario as demonstrated by the recent work of [77]. These authors show that during a migration even (such as the Paleolithic settlement of Europe) mutations that occur close to the population wavefront are quite likely to increase in frequency as the wave progresses. This could leads to the situation where the frequency is highest at the location where the population wave came to a halt but low at the location where the mutation occurred. Further population genetic modelling in this area is needed to determine the evolutionary history of the CCR5-Δ32 allele. For the remainder of this section we will however concentrate on recent work into possible selection pressures on CCR5-Δ32.

HIV-1 has not been around for long enough in humans to have exerted selective pressure on CCR5-Δ32 populations. Additionally, in places in Africa where the incidence of HIV is very high natural selection will result in more individuals with advantageous haplotypes that increase the time between HIV-contraction and AIDs-onset. Schliekelman et al. [78] estimated the frequency of such an AIDs-delaying haplotype in a population over 100 years finding an increase from 0.4 to 0.53 delaying the average onset of AIDS in infected individuals by a year. They find this selective intensity to be comparable to that of CCR5-Δ32 over the past 700 years and argue that given the short period of time that the plague epidemics lasted (see below), the partial resistance gained by heterozygous CCR5-Δ32 individuals offered a selective advantage that does not account for the current high disease frequency.

4.4 Smallpox versus plague

This still leaves the question as to possible causes for the selection pressure on CCR5-Δ32. The age estimate of Stephens et al. [25] falls into the period in European history, which is commonly referred to as the black death. During this period bubonic plague swept across Europe killing approximately 30% of the population within a mere 6 years [79]. This has lead to the early hypothesis that the causative parasite of bubonic plague, the bacterium Yersinia pestis, has exerted the selection pressure on the CCR5-Δ32 allele, necessary to drive it up in frequency.

In the absence of experimental verification for such a theory Galvani and Slatkin [80] have used a population genetic model to study the selective impact plague and smallpox would have had on a resistance allele. They used an age-structured model (parameterized using historical data such as burial and census records) together with a generic disease model to study the expected dynamics resulting from the historical plague and smallpox epidemics in Europe. They observe that based on known disease characteristics (such as the age profile of disease mortality) smallpox is a much more likely cause for selection pressure potentially experienced by CCR5-Δ32. In contrast to the instances of Plague epidemics (including two pandemics: the Black Death, 1346–1352 and the Great Plague, 1665–1666) smallpox has apparently been endemic from at least 200 AD. Moreover, smallpox is a disease that primarily strikes the young and thereby exerts greater selection pressure than bubonic plague which tends to affect older individuals. Smallpox as a means of forcing CCR5-Δ32 to high frequencies was first suggested by [67]. While the bubonic plague took the lives of some 23 million people between 1346 and 1352 it has been estimated that smallpox killed over 300 million people prior to its eradication in 1979.

In the model of Galvani and Slatkin the allelic frequency of the resistance gene (equivalent to CCR5-Δ32) never reaches todays levels and does not exceed 0.8% if bubonic plague is considered (it is 16% in some populations today). However in the smallpox model, a putative resistance gene reaches an allelic frequency of 10% in 680 years.

Significantly, there is evidence that HIV is a lot more similar in etiology to poxvirus than to Yersinia pestis [81]; for example infection is facilitated through the use of chemokine receptors. Although, receptors for poxviruses have not been identified except for Myxoma virus [81], the latter utilizes CCR5 and CXCR4 amongst others. From the crystal structure of the cowpox virus, which is ∼95% similar to smallpox, a site of conserved residues has been implicated in the binding of CC-chemokines [82]. Of the 20 complete poxvirus genomic sequences currently in sequence databases Table 1 in Seet et al. [83] lists known poxvirus modulation of chemokines. They also discuss molluscum contagiosum virus and fowlpox virus which are the only two poxviruses observed to possess chemokine-mimics. There is a viral CC chemokine inhibitor (vCCI) expressed by many poxviruses that binds CC chemokines with high affinity and can compete with CC chemokine receptors in the host [84]. For Yersinia pestis such evidence is lacking and very recently in fact it was shown that the CCR5-Δ32 allele does not prevent infection of mice by Yersinia pestis [85]. Here, normal mice and mice with the CCR5-Δ32 allele were infected with Yersinia pestis and there was no observed difference between the two especially with respect to growth in the bacteria or time until death. This supports the smallpox population genetic model assuming infection mechanisms between mice and men are not significantly different. It is perhaps interesting to note at this point that CCR5-Δ32 homozygous mice displayed atypical immunological responses.

5 Evolutionary constraints of and potential therapies for HIV-1

There are currently around 20 anti-HIV drugs, most successfully used in combination, such as the so-called highly active antiretroviral therapy (HAART) [86]. Understanding the structure of the HIV-1 virus and its method of cellular entry will almost certainly benefit the development of novel treatments and vaccines [87]; here, as we have seen, the molecular genetics of the host can provide valuable clues. Here we will discuss the implications host genetics can have for the evolution, and hence treatment of HIV.

The HIV-1 genome is less than 10 kb long and has nine genes [28]. The molecules responsible for entry are the cell-surface attachment glycoprotein gp120 and the membrane spanning protein gp41, both of which make up the spikes which extend from the viral envelope (Fig. 1) [28]. Half of the weight of gp120 is made of its sugar coating which seems to act as protection against antibodies and protein-degrading enzymes. Gp120 has five constant regions (C1–C5) and five variable regions (V1–V5) and it is the charge and amino acid content of the V3 region which determine its target cell [88]. It appears that the CCR5 binding site is one of the most highly conserved surfaces on the gp120 core, more so than the CD4 site to which it is physically close [89]. CD4 binding (the first stage in HIV-1 entry) and binding to heparan sulfate proteoglycan (HSPG) seems to stimulate conformational changes in gp120 structure resulting in the motion of the V1–V2 stem away from, and the V3 loop towards, the coreceptor binding site (CCR5) [88]. Actual fusion begins once gp120 has docked to CCR5 and gp41 inserts its now exposed hydrophobic fusion peptide into the target cell. Poignard et al. [88] discuss the use of structural information of gp120 in the development of vaccines. Further, the T-20 peptide, which inhibits the fusion of gp41 has been approved for clinical use.

A nice example of the adaptability that HIV-1 is capable of comes from SIV in mangabey monkeys [12]. In other monkey species SIV uses CCR5 as a coreceptor (in much the same was as HIV-1 does in humans). However in the red-capped mangabey subspecies SIV uses CCR2 as a coreceptor. Further, there is a CCR5-Δ24 mutation in red capped mangabeys with an allelic frequency of 87%. This suggests that this mutation was under strong selection in the past because of an older version of SIV but that the virus was able to adapt and switch to using the CCR2 coreceptor instead. Therapeutic agents to inhibit HIV-1 entry could speed up HIV-1 evolution. The development of CCR5 blockers may increase the chances of a mutated strain that uses CXCR4 instead as a coreceptor arising. HIV-1 could evolve to use other coreceptors. There are around a dozen other seven-transmembrane-domain-G-protein-coupled receptors that HIV-1 could untilize in vitro as coreceptors [29], however inefficiently. Continued therapeutic blocking of CCR5 and CXCR4 could result in selective pressures which increase the efficiency with which HIV-1 uses alternative coreceptors.

The use of HAART since 1996 has resulted in recovered immunity and decline in morbidity and mortality by 80%[87]. The extent of HIV-1 infection is reflected by the CD4 count, which denotes the degree of immunodeficiency, and viral load (number of HIV RNA copies per ml of plasma), which forecasts progression towards AIDS. Viral replication is suppressed by antiretroviral therapy which currently work against two replication enzymes of the virus: reverse transcriptase (converts RNA to DNA post-cell entry) and viral protease (processes viral proteins during construction and budding of virions). Failure of HAART therapy may be a result of its toxicity, gradual drug resistance in patients and/or non-compliance of treatment (long-term drug use is a necessity) [86].

It is interesting that HIV-1 patients treated with HAART show different changes in the levels of CCR5 and CXCR4 expression [90]. MIP-1α, MIP-1β and RANTES as the agonists for CCR5 may have uses in antiviral treatments by blocking the M-tropic HIV-1 virus. However they are not alone; monocyte chemoattractant protein (MCP)-2 seems to do the same thing [91]. Antagonistic CC chemokines [28] that target CCR5 receptors should not have any toxic side-effects as individuals without these receptors appear normal (such antagonists are also being considered as anti-inflammatory therapies). Also, the amount of blocking agent administered needs to be considered as in CCR5 homozygous individuals CCR5 expression can vary by 20-fold [12]. CCR5 antagonists include small-molecule agents, monoclonal antibodies and modified chemokines. Small molecule antagonists [29] include TAK-770 which blocks binding by arresting CCR5 function at nanomolar concentrations; SCH-C/D which may halt R5 HIV-1 entry by deforming the structure of the CCR5 extracellular domain; UK-427,857 a CCR5 antagonist and E913 which blocks the binding of MIP-1α to CCR5. It has been demonstrated that MIP-1β can no longer bind to CCR5 once gp120 has bound to CD4 in a free M-tropic isolate [92]. There is also much research into CXCR4 antagonists. More recently, there has been a flurry of research into using RNA interference to effectively switch off the CCR5 gene [93] thus inhibiting the expression of surface CCR5. There are also Intrakine coreceptors. These antagonists block the surface expression of CXCR4 by attaching chemokines to the endoplasmic reticulum [94]. It has been mentioned that CCR5-Δ32 heterozygotes may respond better to HARRT than CCR5 homozygotes [12].

How likely is it that the virus will evolve to be able to overcome the constraint imposed by the CCR5-Δ32 allele? Quite unlikely we believe. The reason for this is that the allele prevents the virus from establishing a huge presence inside a host. The adaptive evolution of viruses depends in a complex way on the extent of selection pressure and the evolvability of the virus [95]. The latter in turn depends on the number of viral genomes and on the mutation rate. A small viral population is much less likely to give rise to escape mutants than a larger population. Moreover, as discussed above, the dependence on a fully functional CCR5 receptor is very strong. For this reason it is believed that a potential HIV strain that does not depend on CCR5 must differ from the normal wildtype at several nucleotides. For this reason the evolution of an escape mutant seems highly unlikely (even if the intermediate states were fully viable). Finally, as the viral population size in carriers of the CCR5-Δ32 allele appears to be reduced such individuals are less likely to infect others.

6 Conclusions

We have summarized recent results about the molecular biology, virology and host genetics related to the role of the CCR5-Δ32 HIV-resistance allele. At present this system probably offers one of the most detailed insights into the complex interplay between viral and host genetics available. No doubt the situation for the interplay between the virus and the CCR5-receptor is somewhat simpler than we would expect to be the case in general as the HIV-resistance conferred by CCR5-Δ32 segregates in an essentially Mendelian fashion. This may of course not always be the case for host genetic influence on infectious disease dynamics. We believe, however, that the combined analysis of the different factors will prove to give important insights in other cases, too.

A particular attraction of this system for us is that the molecular biology and evolutionary genetics (of the virus as well as the human host) fit together so nicely. We do understand in exquisite detail how the virus employs the CCR5 wildtype to establish itself inside the host. Furthermore we now know why the CCR5-Δ32 allele effectively stops infection in homozygous carriers and delays it in heterozygous carriers. Evolutionary (viral) genetics then allows us to conclude that it is highly unlikely that HIV will evolve CCR5-Δ32 escape mutants. Finally we can use a combination of epidemiology and population genetics to study why and how the allele has reached such high frequencies in Europe in the first place.


EdS and MPHS thank the Wellcome Trust for generous financial support through a Career Development Fellowship to MPHS.


  • Editor: G.M. Ihler


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
  49. [49].
  50. [50].
  51. [51].
  52. [52].
  53. [53].
  54. [54].
  55. [55].
  56. [56].
  57. [57].
  58. [58].
  59. [59].
  60. [60].
  61. [61].
  62. [62].
  63. [63].
  64. [64].
  65. [65].
  66. [66].
  67. [67].
  68. [68].
  69. [69].
  70. [70].
  71. [71].
  72. [72].
  73. [73].
  74. [74].
  75. [75].
  76. [76].
  77. [77].
  78. [78].
  79. [79].
  80. [80].
  81. [81].
  82. [82].
  83. [83].
  84. [84].
  85. [85].
  86. [86].
  87. [87].
  88. [88].
  89. [89].
  90. [90].
  91. [91].
  92. [92].
  93. [93].
  94. [94].
  95. [95].
View Abstract