Dr Rupert Kaul, University of Toronto, Medical Sciences Building #6356, 1 King’s College Circle, Toronto, ON, Canada M5S 1A8. E-mail: Rupert.email@example.com
Citation Kaul R, Cohen CR, Chege D, Yi TJ, Tharao W, McKinnon LR, Remis R, Anzala O, Kimani J. Biological factors that may contribute to regional and racial disparities in HIV prevalence. Am J Reprod Immunol 2011; 65: 317–324
Despite tremendous regional and subregional disparities in HIV prevalence around the world, epidemiology consistently demonstrates that black communities have been disproportionately affected by the pandemic. There are many reasons for this, and a narrow focus on socio-behavioural causes may be seen as laying blame on affected communities or individuals. HIV sexual transmission is very inefficient, and a number of biological factors are critical in determining whether an unprotected sexual exposure to HIV results in productive infection. This review will focus on ways in which biology, rather than behaviour, may contribute to regional and racial differences in HIV epidemic spread. Specific areas of focus are viral factors, host genetics, and the impact of co-infections and host immunology. Considering biological causes for these racial disparities may help to destigmatize the issue and lead to new and more effective strategies for prevention.
It was famously said by Kofi Annan that ‘in Africa, AIDS has a woman’s face’,1 but gender is by no means the most marked imbalance when it comes to the effects of HIV. While women now bear over half of the global HIV burden,2 it is only in the continent of Africa that women constitute the majority of infected persons. In contrast, there is a tremendous disparity in the effects of HIV along racial and ethnic lines that is apparent throughout the world. This imbalance is most marked at a continental level, given that approximately two-thirds of all HIV-infected persons are in Africa, but is also apparent within most regional subepidemics. The reasons underlying the racial and geographical imbalances in HIV prevalence are complex and have led to myths, stereotypes, stigma and discrimination that may impede the development of better HIV prevention tools and programs.
As is the case for all sexually transmitted infections (STIs), socio-economic and cultural factors have been hypothesized to be critical contributors to HIV transmission and increased HIV prevalence in Africa.3,4 Many of these sociocultural factors are potentially stigmatizing and include higher per-capita rates of commercial sex,5 increased partner exchange/concurrency,6,7 intimate partner violence,8–10 and traditions such as wife inheritance.11 There are data supporting the causal association of HIV with at least some of these factors, but it is unfortunate that a focus on the cultural and behavioural aspects of HIV transmission tends to implicitly lay blame for infection on affected communities or individuals.12 While a discussion of the sociocultural associations of HIV is beyond the scope of this review, our goal is to emphasize that there may be other causes for the geographical and racial imbalances in HIV prevalence that are equally important. Specifically, our goal is to explore possible biological cofactors that may enhance vulnerability and contribute to the substantial global racial disparities in HIV prevalence. Our hope is that a better understanding of such cofactors may allow the development of new HIV prevention tools while reducing stigma.
The epidemiology of HIV and race
There are major racial and geographical disparities in HIV prevalence. Not only do over two-thirds of HIV-infected individuals reside in Africa, a continent that holds under 12% of the global population,13 but there are major racial disparities within most regions of the world. In the USA, AIDS rates are ten times higher in African Americans than in white Americans.14 Specifically, the HIV prevalence in black men is six times that in white men, and in black women the rates are nearly eighteen times higher.15 Likewise, it is estimated that in Ontario, Canada approximately 22.5% of HIV-infected individuals and 3.9% of the provincial population are black, so that the HIV prevalence is increased six-fold in black men and 24-fold in black women16 (and R. Remis, personal communication). There can also be dramatic differences in the degree to which HIV affects districts and ethnic groups within individual African countries. For instance, the HIV prevalence in Nyanza province, Kenya is more than double that of the rest of the country (13.9% versus 6.3%), and those of Luo ethnicity (who predominate in this district) have an HIV prevalence over three times the national average (20.2% versus 6.3%).17
As sexual partnerships are generally formed within the same geographical region and/or community, it would not be surprising to find that this increased HIV prevalence would be associated with a higher HIV incidence. However, in many situations, the ‘per exposure’ rate of HIV acquisition seems to be disproportionately high. For instance, the annual HIV incidence within the control arm of the recent CAPRISA trial of tenofovir gel in KwaZulu-Natal was an astounding 9.1%, despite a low reported number of prior/new sexual partners. Likewise, HIV rates were 2.5–6 times higher in women than men aged 15–19 years from Kisumu (in Nyanza province, Kenya) without apparent gender differences in prior HIV exposure.18,19 These data strongly suggest regional differences in HIV susceptibility and additional susceptibility differences by gender.
Observational studies of HIV transmission, often performed in the context of HIV serodiscordant couples, have not generally examined race as a cofactor in HIV transmission. However, a recent meta-analysis of observational studies examining the risk of transmission during heterosexual sex found that, in the absence of commercial sex, the per-exposure risk of male-to-female transmission was almost four times higher in low-income countries compared to high-income countries (0.30% versus 0.08%), and the risk of female-to-male transmission was increased ninefold (0.38% versus 0.042%).20 This does not prove that race itself is associated with biological differences in HIV susceptibility, but it clearly demonstrates that the increased HIV transmission in low-income countries is about more than partner selection or commercial sex.
The biology of HIV transmission
As already described, HIV transmission is much less efficient than one would expect from the size of the HIV pandemic. The per-exposure transmission rate for both penile-vaginal and vaginal-penile sex is roughly 0.05% in high-income and 0.3% in low-income countries,20 meaning that HIV transmission is the rare exception rather than the rule, whatever part of the world is under discussion. While there is a great deal that we do not understand about the biology of HIV transmission, we do know biological factors are critical determinants of exposure outcome.21,22 The most important determinants of transmission are (i) the HIV level in the blood and genital/rectal secretions of the HIV-infected partner and (ii) the number and density of HIV-susceptible target cells to which the virus can gain access at the site of exposure (usually the mucosal lining of the penis, rectum or female genital tract) in the HIV-uninfected partner.23,24 As will be discussed, these two critical determinants are affected by numerous, overlapping biological factors: we hypothesize that this biology, in addition to any sociocultural and economic factors, has played and continues to play an important role in the racial imbalance that characterizes the global HIV pandemic. Important biological factors and the potential interactions of these factors with race and geography are now reviewed under the broad headings of viral factors, host genetic factors, co-infections and host immunology.
HIV-1 group M viruses are subdivided into several subtypes or clades based on genetic heterogeneity: these clades have strong geographical associations,25 and considerable research has examined the potential associations of clade with HIV transmission. Clade C predominates globally and is responsible for most HIV infections in southern Africa and India, while clade B predominates in North America, Europe and Australia. East Africa is dominated by clade A and to a lesser extent D, while the recombinant virus designated CRF01_AE (previously clade E) is most common in Thailand. Early studies suggested that clade C and CRF01_AE were more easily transmitted through heterosexual sex,26 potentially because they bound preferentially to Langerhans cells in the vaginal mucosa and penis.27 Research continues in this area, and more recent work has found that HIV clade C shows enhanced replication (compared to clade A) in a dual virus culture system, as well as in an ex vivo cervical explant model;28 in addition, observational studies demonstrate that clade A may be transmitted more easily than clade D.29 While these data are interesting, it is probably fair to say that it remains unclear what effect, if any, virus clade has on patterns of HIV epidemic spread in the real world. Certainly, virus subtype cannot explain racial differences in HIV prevalence that are apparent in multiple regions and across different virus clades.
The clearest association of viral factors with HIV transmission is the plasma HIV RNA viral load, with higher plasma levels being associated with stepwise increases in the probability of transmission30 and in virus levels within genital secretions.21 There are a few data to suggest that plasma viral load varies substantially with viral clade, geography or race per se. However, co-infections such as tuberculosis, malaria and herpes simplex type 2 (HSV-2) have strong regional associations, can clearly increase the viral load and may enhance transmission;31 this is discussed later in more detail.
Genetics and HIV susceptibility
Variation in host genetics would perhaps be the most intuitive mechanism for geographical and racial differences in HIV prevalence. Indeed, the best-described association of genetic resistance to HIV infection is homozygosity for CCR5Δ32, which is phenotypically characterized by an absence of the HIV co-receptor CCR5 on the cell surface.32–34 This genotype is associated with near-complete resistance to sexual HIV acquisition, and stem cell transplantation from a CCR5Δ32 homozygous donor has resulted in the functional cure of HIV.35 While this gene is present at a frequency of approximately 10% in people of European descent, it is much less common in non-Europeans.36
However, not all genetic associations of HIV resistance are increased in non-black populations. A reduced number of gene duplications encoding CCL3L1, which encodes the CCR5 ligand MIP1α, may be associated with increased HIV susceptibility,37 although there are conflicting data in this area.38 African populations have higher copy numbers of this gene duplication,37 and other genetic associations of relative HIV resistance have also been mapped in Africa.39–41 Overall, while there is clear racial variation in several genes associated with differential HIV susceptibility, the degree of variation in the genetic determinants mapped to date is insufficient to explain the global associations of HIV and race.
Co-infections, geography and HIV transmission
Dramatic regional and racial variation in the prevalence of co-infections that may enhance HIV transmission means that this is likely to be an important contributor to global disparities in the HIV pandemic.31 Clinical trials have shown that the blood HIV RNA viral load was reduced to varying degrees by therapy of each of tuberculosis (a drop as high as >3.0 log10 copies/mL), malaria (approximately 0.3 log10 copies/mL), geohelminths (approximately 0.2 log10 copies/mL), schistosomiasis (approximately 0.4 log10 copies/mL) and filiariasis (approximately 0.8 log10 copies/mL).31 No clinical trials have assessed the impact of therapy for these co-infections on HIV transmission, but models suggest that a 0.3 log10 increment in the plasma viral load would be associated with a 20% increase in HIV transmission, while a 1.0 log10 increment would increase transmission by 100%.42 On this basis, it has been estimated that malaria has caused an excess 8500 HIV infections in a Kenyan community of 200,000 with high malaria rates.43 Clearly, co-infections that are endemic in sub-Saharan Africa can impact HIV transmission and may in part explain the disproportionate spread of HIV in this region.
The HIV RNA blood viral load in the blood correlates with that in the genital tract, albeit incompletely, and this is probably the reason for the association between blood viral load and transmission probability. Therefore, one might expect that genital co-infections would have a particularly dramatic effect on HIV levels in genital secretions, and this is the case: the semen HIV viral load was increased by approximately 0.7 log10 copies/mL in men with gonorrhea44 and 1.0 log10 copies/mL during semen CMV reactivation.45 Both genital infections and bacterial vaginosis (BV), an imbalance in the normal vaginal flora, have a similar effect in the female genital tract.22 HSV-2 merits individual mention, because suppressive therapy in HIV/HSV-2 co-infected individuals with acyclovir-based medications has been consistently associated with a reduction in both the blood and genital tract HIV viral load,31 although a recent clinical trial of HSV-2 suppression in HIV co-infected individuals did not reduce HIV transmission to their sex partners.46 Furthermore, genital infections do not only increase HIV transmission from a co-infected individual, but they have been consistently linked with increased HIV susceptibility in an HIV-uninfected person,47 likely due to immune alterations outlined in the next section. HSV-2 infection, even if asymptomatic (as most cases are) increases HIV susceptibility approximately threefold in both men and women,48 and BV increases a woman’s susceptibility by 60%.49
Genital co-infections may play a key role in HIV transmission, but for them to play a role in racial and geographical imbalances in HIV prevalence, a similar imbalance must exist in their own prevalence. Studies have shown that this is the case. For instance, while the HSV-2 seroprevalence is around 15–20% in white women from the USA, it is over 50% in black women from the USA50 and African/Caribbean women from Canada,51 and it may exceed 80% in adult women from sub-Saharan Africa.52 Rates of BV in women from sub-Saharan Africa are approximately double those in the rest of the world,53 and within North America BV preferentially affects African-American women for reasons that are poorly understood.54,55 Given that both HSV-2 and BV each predispose to the other and to the acquisition of a range of other STIs,56 it is clear that genital co-infections may be an important mechanism driving the association of black race and HIV prevalence.
Systemic and mucosal immune associations of HIV transmission
As stated above, a critical determinant of HIV susceptibility is the number and density of HIV-susceptible target cells to which the virus can gain access at the site of exposure. Perhaps the clearest demonstration of this is the fact that male circumcision reduces HIV acquisition by approximately 60%.57,58 This is presumably because of the direct removal of the HIV target cells that are present in the foreskin,59,60 although the pathophysiology and immune correlates of HIV acquisition in the foreskin remain poorly defined. In keeping with this, 88% of Kenyan men aged 15–49 years are circumcised, and the HIV prevalence is 3% in circumcised men as opposed to 13% in those uncircumcised:17 men of Luo ethnicity do not traditionally practice circumcision, leading to a lower circumcision rate of 21.5% in these men and of 44.8% in Nyanza province.17 Therefore, we would argue that biology is likely to be the major contributor to the disproportionate impact of HIV within this community and area that was described earlier.
The level of immune activation in an individual may also be an important predictor of their susceptibility (if HIV uninfected) or infectiousness (if HIV infected). In keeping with this, immune activation is substantially dampened in rare individuals who are relatively resistant to HIV infection.61 HIV replicates more efficiently within activated CD4+ T cells,62 in part because cell activation leads to increased surface expression of the HIV co-receptor CCR5.63 As immune activation and inflammation are key host responses to an invading pathogen, endemic infections such as malaria in an HIV-infected individual would be expected to indirectly increase virus replication and blood levels, and thereby causing the enhanced HIV transmission that was described in the previous section. BV and genital co-infections such as HSV-2 lead to dramatic increases in activated CD4+ T cells directly within the genital mucosa of an HIV-infected individual,64–66 and therefore may have an even greater effect on the genital HIV viral load and subsequent HIV transmission.
Systemic immune activation may be increased in healthy African individuals67–69 and might be hypothesized to increase HIV susceptibility, although at least part of this phenomenon is probably driven by higher rates of co-infections such as HSV-270 and geohelminths67–69 that were often not screened in these studies. Whether this systemic immune activation increases the per-exposure HIV acquisition risk may hinge on whether it is associated with a corresponding increase in HIV-susceptible target cells at the site of virus exposure (i.e., the mucosal lining of the cervix, vagina, rectum or penis). While HSV-2 is clearly associated with increases in activated T cells within both the genital tract and blood,65,70 it is not known whether non-sexually transmitted infections (malaria and so on) have the same effect.
Interestingly, recent work from our group demonstrated higher numbers of activated CD4+ target cells in the female genital tract of young women from Kisumu (Nyanza province, Kenya) compared to San Francisco (USA), independent of genital co-infections or other behavioural practices.71 These data are summarized in Fig. 1, which shows that the total number of activated CD4+ T cells collected on a cytobrush was increased in Kisumu women (Fig. 1a); this was not attributed to higher overall CD4+ T cell numbers, but rather to substantial increases in the percentage of CD4+ T cells that were activated (Fig. 1b). In addition, vaginal levels of secretory leucocyte protease inhibitor (SLPI), an innate mucosal immune protein with anti-HIV properties in vitro,72 were substantially lower in Kisumu participants (Fig. 1a). Whether these differences will be confirmed in future studies, and if so whether they were attributable to systemic infections or other factors such as genetics or differences in the genital microbiome are important questions that may lead to the development of novel prevention strategies.
Racial disparities in HIV prevalence are profound, both between regions and within regions. These disparities are not often discussed, perhaps because it is assumed that they are driven by stigmatizing socio-behavioural factors such as sexual concurrency or promiscuity, partner violence and so on. While such factors may be important in some contexts, the purpose of this review has been to emphasize that biological factors such as endemic co-infections and immunology also play a key role. To develop better prevention tools, it is critical that communities, researchers and policy makers come together to discuss and investigate these tremendous disparities in an open and non-judgmental fashion.
This work was supported by grants from the Canadian Institutes of Health Research (RK, HET-85518; LRM and DC, salary support). Study sponsors played no role in the writing of the manuscript or decision to submit for publication. No author has any financial or personal relationship posing a conflict of interest in relation to this study.
Study concept and initial draft: RK; manuscript revisions: CRC, TJY, DC, WT, LRM, OA, JK, RR.