HIV molecular epidemiology: transmission and adaptation to human populations

Abstract

Purpose of review

To provide an update on the origin of the HIV epidemic and insights into how the immune response is shaping virus evolution.

Recent findings

Characterization of archival samples showed that by the 1960s, HIV had already diverged within humans. It is now estimated that HIV has been in humans since at least the early 1900s. However, despite the potential for different divergent viruses to spread, surprisingly few viruses successfully expanded to cause the global epidemic. In approximately 80% of cases, productive infection is the result of infection with only a single virus or single virus-infected cell. After transmission, HIV evolves at a rapid rate driven by the immune pressure until the virus reaches a delicate survival balance: on one hand avoiding elimination through the development of cytotoxic T-cell immune escape mutations, and on the other sacrificing replication fitness as these mutations may come with a severe fitness cost to the virus. People infected with these ‘attenuated’ cytotoxic T-cell escape viruses can have a survival advantage. Cytotoxic T-cell responses are molding HIV diversity at a population level resulting in a loss of some of the common immune epitopes.

Summary

Insights into the origin of HIV and its evolution between populations and within individuals is essential to understanding HIV pathogenesis and imperative for the design of effective biomedical interventions such as vaccines.

Introduction

The rapid evolution of HIV is a hallmark of this virus and high diversity is one of the key reasons for the success of this pathogen. In order to meaningfully intervene through vaccines, microbicides or antiretroviral therapy, scientists need to develop a holistic and precise understanding of virus evolution in the context of host responses.

HIV origin

HIV-1 is thought to have crossed into humans three times to result in three distinct phylogenetic lineages (M, N and O) with group M responsible for the HIV pandemic [1]. A few years ago an elegant study was published whereby investigators analyzed primate feces from the jungles of central Africa. This study not only definitively identified the chimpanzee (Pan troglodyte troglodyte) as the natural host of the group M and N viruses but also determined that the likely geographical location of the epidemic was in west-central Africa [1]. Until recently, however, the natural host of the group O viruses remained inconclusive. These group O viruses are most closely related to viruses infecting gorillas, and new evidence suggests that chimpanzees are also the origin of infection of these gorillas [2]. This suggests that chimpanzees could be the common source of all three HIV lineages. However, one cannot rule out that the group O viruses may have been introduced into humans via a gorilla intermediate [2,3].

Using estimated rates of viral diversification it is possible to date the common ancestor for the HIV M group, and to estimate the latest time at which the transfer from non-human primates occurred. The identification of a second HIV-positive specimen collected in 1960 from the Democratic Republic of the Congo (DRC, previously Zaire), together with the information on the earliest HIV sample characterized from 1959 [4], showed that even at this time the group M viruses had diversified significantly (~12%) suggesting that it had been evolving in humans for longer than previously thought [5^••]. It is now estimated that HIV has been in humans since at least 1902–1921. Interestingly, this timing coincides with urbanization when the first towns were emerging in this region of Africa.

Global spread

Since the introduction of HIV into humans, the virus has evolved into distinct phylogenetic groups or subtypes labeled A1, A2, B, C, D, F1, F2, G, H, J and K (Fig. 1) [6]. In addition, there are now at least 43 mosaic viruses circulating (called circulating recombinant forms) which comprise of components of more than one subtype [7]. The highest diversity of HIV remains in the region where HIV originated in west-central Africa and despite the potential for divergent viruses to spread, surprisingly few viruses have successfully expanded with 90% of the epidemic comprising of four subtypes (A, B, C and D) and two circulating recombinant forms (CRFs) (CRF01_AE and CRF02_AG). Historically, many more viruses undoubtedly emerged from Africa; however, these strains did not establish themselves within transmission networks, or were of lower fitness which limited their dispersal. Founder effects can probably account for most of the current dominant epidemics whereby a single chance introduction of an ancestral virus resulted in major spread. This is clearly illustrated in a study by Gilbert et al. [8] who investigated the spread of subtype B to Haiti and the US. They demonstrated that the epidemic originated in Haiti around 1966 and within 5 years a single transmission event occurred that culminated in the subtype B epidemic in the US. HIV was evidently smoldering in North America for approximately 10 years prior to its discovery in 1981. This subtype has become massively successful accounting for most infections in the industrialized world (Fig. 1).

Figure 1. Regional HIV-1 subtype and circulating recombinant form (CRF) distribution.

Modified and updated from Hemelaar et al. [6]. 03_AB = CRF03_AB; 12_BF = CRF12_BF; 01_AE = CRF01_AE; 07_BC = CRF07_BC; 08_BC = CRF08_BC; RF = unique recombinant form.

Population bottlenecks result in a drastic reduction of population size with diminished genetic variability and, in RNA viruses, are usually associated with reduction in viral fitness [9]. The question has been asked as to whether HIV replication fitness has changed over the course of the epidemic. Ex-vivo replication fitness assays have been developed and early cross-sectional studies comparing fitness in samples collected in the late 1980s compared to the early 2000 suggested that HIV was becoming less pathogenic over time [10]. However, viral fitness can increase with disease progression and a similar study which controlled for time postinfection found the opposite results in that there was a trend for viruses to gain in fitness over the course of the epidemic in Amsterdam [11]. Similarly, conflicting results were obtained in cohort studies. A comparison of early disease progression markers in large cohorts spanning Europe, Canada and Australia showed trends of decreasing CD4 cell counts and increasing viral loads from 1985 to 2002 [12], whereas no change in disease progression was observed in the US from 1984 to 2005 [13]. One of the major drawbacks of comparing historical to more recent data is that advances in diagnostic techniques may change the readout and thus bias results. These studies are also difficult to interpret within the context of host genotype, subtype, nutritional status, access to healthcare – all of which might impact on disease progression. Whether HIV replicative fitness has changed since the start of the pandemic, remains inconclusive.

Transmission

As vaccines and microbicides target transmitted viruses there is intense interest in understanding the genetic characteristics of these variants, which may differ from those of chronic infection [14]. It has been shown that despite a swarm of closely related viruses present in the donor, there is a genetic bottleneck associated with transmission so that only a limited number of variants get transmitted to the recipient [14–17]. Recent studies have significantly advanced our understanding of transmission: first, by investigating a large number of individuals in very early stage of infection, many of whom were HIV RNA-positive antibody-negative; and second, by applying more quantitative methods using single-genome amplification (SGA) which has enabled investigators to infer the sequence of the infecting virus and thus enumerate the number of infectious units [18^••,19^••,20,21^••]. This more standardized approach has also enabled us to directly compare results between cohorts.

The largest of these studies have been done in 102 subtype B-infected individuals infected via men who have sex with men (MSM) route or heterosexually [19^••], and in 69 subtype C heterosexually infected men and women [21^••]. Interestingly, both studies estimated that approximately 80% of productive infections were the result of a single virus or a single virus-infected cell. The transmission of a single infectious unit provides a window of opportunity for vaccines which, at least in the majority of cases, would need to protect against a very small incoming viral dose of limited diversity. It is still an unresolved question whether these viruses have distinctive features associated with sexual transmission. However, a recent study of vertical transmission in seven mother–infant pairs showed that replicationally fit viruses were selectively transmitted to the infant in the presence of less fit variants in the mother [22^•].

Approximately 20% of people were infected with multiple viruses (between two and six variants) [19^••,21^••] and there is interest in understanding factors that disrupt this bottleneck not only because it has vaccine implications but also because high diversity following transmission has been associated with more rapid disease progression [23–25]. Studies in macaques have shown that there is a relationship between dose and frequency of multiple variant transmission suggesting that factors that increase HIV transmission rates will also increase frequency of multiple variant transmission [26]. Similarly these macaques studies showed that whereas intravenous inoculation resulted in multiple variant transmission, mucosal routes of inoculation recapitulated human studies, implying that it is the mucosal barrier that is responsible for the population bottleneck during transmission. This is further supported by recent studies confirming that sexually transmitted infections, which disrupt the mucosal barrier, are associated with increased frequency of multiple variant transmission [18^••]. Lastly, mode of transmission may also be important with increased frequency of multiple variant transmissions reported in MSM [27].

HIV adaptation to humans

As HIV is passaged through populations, the virus is continually adapting to avoid immune responses. There is evidence to support that mutations associated with immune evasion are accumulating at a population level [28,29^•,30^••]. This imprinting on the viral genome appears to be due to cytotoxic T-cell (CTL) responses which recognize infected cells through the presentation of short peptides by the human leukocyte antigen (HLA). The virus escapes these responses through mutations which interfere with epitope processing, binding or recognition. A study of 2800 patients spanning five continents showed that HLA frequency is correlated with the persistence of certain polymorphisms in the viral genome [30^••]. HLA alleles which occur at high frequency in certain populations, such as HLA-B^*51, HLA-B^*27 and HLA-B^*57, appear to be driving the fixation of certain CTL escape mutations within viruses in a given population. We do not yet know if infection of individuals bearing these HLAs with these ‘less immunogenic viruses’ will affect their disease progression; however, presumably they would elicit CTL responses to other subdominant epitopes.

These escape mutations are not necessarily ‘bad news’ to the host as some of them come with a fitness cost to the virus [31,32]. One group of HIV-1-infected individuals, the elite controllers, is able to control viremia to below 50 copies/ml and many different mechanisms have been suggested to be associated with this control including host genetics and immune responses [33–36]. Viruses infecting elite controllers are enriched for HLA-associated polymorphisms and one virological mechanism which is thought to contribute to this viremic control is HLA-mediated attenuation of viruses [37,38^•,39]. Mutations that carry a fitness cost are also known to revert to wild type in the absence of immune pressure and the speed at which this happens is related to the fitness cost of the mutation [29^•]. However, the rate or timing of reversion is affected by the presence of compensatory mutations that partially restore replication fitness [40]. In fact the selection of one CTL escape mutation above others that are more effective at disrupting epitope binding is influenced by the ability of some CTL escape mutations to occur in the presence of compensatory mutations [41]. Therefore, the persistence of variants carrying escape mutations depends on whether the fitness cost of the escape mutation outweighs the selective advantage of immune evasion [31].

Certain HLAs such as B^*57/5801 appear to drive HIV to a more attenuated form and two studies have shown that individuals who do not have these protective HLAs can also benefit. We have shown that transmission of viral variants carrying HLA-B^*57/5801-associated Gag escape mutations to HLA-mismatched recipients was associated with lower viral loads and higher CD4 cell counts during early infection [42^••]. A similar study of HIV-infected donor and recipient couples demonstrated that transmission of variants carrying a high number of CTL escape mutations within Gag was significantly associated with reduced viremia in the recipients 6 months after infection [43^••]. Together these results suggest that infection with viruses containing CTL mutations that come with high fitness costs may ameliorate infection in these individuals. Vaccine immunogens that elicit CTL responses that drive the virus into an attenuated form could provide a survival advantage in vaccine recipients who become infected despite vaccination. However, the potential loss of epitopes in circulating viruses [30^••] does imply that vaccine immunogens would need to be reviewed over time to ensure they reflect the evolving diversity.

Conclusion

HIV has been evolving in humans since at least the early 1900s. Despite high diversity in the west-central African geographical origin of the current pandemic, a limited number of viruses have spread with only four subtypes and two circulating recombinant forms responsible for 90% of infections. Even though a number of studies previously described genetic complexity of transmitted variants, it is only in the past year when it was definitively shown that 80% of infections were the result of single virus or single virus-infected cell in both subtypes B and C, the major contributors of the global pandemic. Similar figures have been shown for other subtypes and cohorts [18^••]. There is accumulating evidence that HIV is adapting to human populations resulting in an increased number of viruses carrying mutations which make them less recognizable by individuals bearing certain common HLAs [30^••,44]. Some of these CTL escape mutations come with a fitness cost to the virus and infection with these attenuated CTL escape viruses have been associated with lower viral loads. Transmission of these CTL escape viruses to HLA-mismatched individuals has also been associated with lower viral loads and high CD4+ cell counts. However, how the circulation of these attenuated viruses will impact on pathogenesis and the epidemic remains to be defined. Understanding viral evolution will provide vital information to understanding HIV transmission and pathogenesis, which are key research areas needed to design more effective HIV vaccines – the greatest challenges facing HIV research.