REVIEW ARTICLE: Mucosal Innate Immunity as a Determinant of HIV Susceptibility
Shehzad Iqbal, Medical Science Building #6356, 1 King’s College Circle, University of Toronto, Toronto, ON, Canada M5S1A8.
Most human immunodeficiency virus (HIV) is acquired during sex, across a mucosal membrane. Despite many advances in our understanding of HIV pathogenesis, the initial events during mucosal transmission have been poorly characterized, and a better understanding of these events will probably be a key to the development of successful microbicide(s) and/or a preventative HIV vaccine. While a vast majority of mucosal HIV exposures do not result in productive infection, implying that innate mucosal immune defenses are highly protective, failure of these mucosal defenses resulted in over 3 million new HIV infections in 2006. We review recent findings regarding HIV mucosal immunopathogenesis, emphasizing the importance of innate immunity in natural protection from infection, and examine how natural or induced perturbations in the mucosal innate system may underpin HIV transmission. Given the great challenges to the development of HIV microbicides and vaccines, identification and enhancement of ‘natural’ innate immune defenses present attractive avenues for development of safe, non-toxic microbicides.
HIV currently infects approximately 40 million people globally, most of which are in sub-Saharan Africa.1 While the slow roll out of antiretroviral therapy (ART) in this region has started to reduce mortality from HIV/AIDS, therapy remains available for a minority of those in need, and the long-term sustainability of ART programs is unclear. Education and prevention programs can reduce HIV sexual risk taking and infection rates2, but these alone cannot reverse the pandemic. Education programs can be effective only in receptive populations, and may be difficult to implement and sustain for the marginalized populations that are in the greatest need. Such populations, particularly female sex workers, may act as a ‘core transmission group’ to the general population, and serve as a reservoir of infected individuals who continue to fuel the epidemic. Therefore, additional prevention strategies, such as an effective and safe HIV microbicide or vaccine, are urgently needed, and particularly within these high-risk subgroups.
Much attention has been focused on elucidating the mechanisms of natural HIV protection in highly exposed persistently seronegative (HEPS) populations, such as HIV ‘resistant’ sex workers, and it is easy to forget that to a large extent we are all ‘HIV resistant’. Estimates of the risk of HIV transmission per unprotected heterosexual sex act range from 3 to 50 per 10,000 unprotected exposures3, such that the probability of transmission per coital act is approximately 0.1–1%.4 This means that the protective efficacy of the genital mucosal barrier is 99–99.9%, which is much higher than that for preventative vaccines available for almost any infection, much less for HIV. Nonetheless, HIV infects an enormous number of people globally: how then has HIV spread so efficiently? Perhaps the most plausible explanation is that viral transmission is a highly variable process whereby transmission probabilities may vary tremendously from individual to individual and from one time to another within the same individual. We hypothesize that this variability might relate to natural or induced perturbations in innate immune mucosal defenses.
Mucosal immune defenses against potential microbial pathogens are poorly understood. Although this lack of knowledge is particularly acute for the male genital tract, it is in men that the promise of mucosal strategies to prevent HIV acquisition has been most dramatically demonstrated. A simple, cheap, one-time mucosal intervention – namely, male circumcision – has unequivocally demonstrated a protective effect of 50–60% in three carefully conducted randomized clinical trials, and efforts are now focused on how male circumcision programs can be scaled up safely.5–8 While the biological basis for protection has not been fully elucidated, it is likely to relate to a reduction in the number of mucosal HIV target cells in the male external genitalia. The foreskin contains a large number of CD4+ T cells and Langerhans cells [a subset of dendritic cells (DCs)], which are both felt to be key initial cellular targets for HIV.9
In this review, we will highlight several innate immune factors that are active at a mucosal level, that may be important determinants of HIV susceptibility. Given that microbicide development has been hampered by mucosal toxicity and unexpected increases in HIV susceptibility, we hypothesize that a better understanding of such ‘natural’ protective factors might be an important step in the development of safe and effective microbicides.
How does HIV cross the mucosal barrier?
There are several feasible routes by which HIV transmission might occur across the genital mucosal epithelium, and which one predominates in ‘the real world’ is a matter of ongoing debate. Most studies have focused on HIV transmission across the female genital mucosa, probably related to practical issues of mucosal sampling. Our understanding of the mucosal mechanisms of HIV transmission after penile HIV exposure is lacking, particularly for circumcised men.
In the female genital tract (FGT), HIV can be transmitted ex vivo by either cell-associated or cell-free virus, although there is a suggestion that transmission after exposure to cell-associated virus may be more efficient; both are present in the semen of an HIV-infected male partner.10–12 Whichever mode is responsible for transmission, the virus must make its way through several non-specific mucosal barriers, including the cervico-vaginal mucus and a low intravaginal pH, to interact with susceptible host target cells.13,14 It is likely that the initial cellular targets for HIV after mucosal inoculation are either activated CD4+ T cells or DCs/macrophages.10 Genital DCs lie within the subepithelial lining and are potent antigen presenting cells that may express a surface adhesion molecule known as Dendritic Cell-Specific Intracellular adhesion molecule-3 Grabbing Non-integrin (DC-SIGN).15,16 This adhesion molecule binds and internalizes HIV virions, generally in the absence of productive DC infection, a process that may induce DC activation, maturation and migration to the submucosa and/or regional lymph nodes, where DC-associated antigens are presented to CD4+ T cells and productive infection may occur.15,17,18 In addition, mucosal T cells are relatively plentiful, and while CD8+ T cells predominate intraepithelially, the CD4/CD8+ T-cell ratio in the submucosa is similar to that in blood. In an ex vivo model of the vaginal epithelium, the initial HIV cellular targets were activated CD4+ T cells and intraepithelial Langerhans cells, through independent mechanisms, with the former appearing to be the most efficient.19 Whether the same holds for transmission across the cervical epithelium is not clear, and it may be that transmission is more common across the cervical epithelial monolayer than the more robust pluristratified squamous vaginal epithelium. Regardless, it is likely that any natural or external factors that increase the number of genital mucosal DCs or activated CD4+ T cells will increase the probability of HIV acquisition after exposure.
The uniquely structured cervical epithelium is composed of columnar epithelial cells, is relatively enriched for immunologically active cells, and is subject to weakening or ulceration through hormonal changes, aging or concurrent infections.10 It is likely that disruption of the integrity of the epithelial barrier arising out of any cause, either in cervix or the vagina, represents another portal of entry for the virus. In addition to the differences in the physical barrier presented by the cervical and vaginal epithelia, HIV is able to cross the cervical monolayer by transcytosis, although the significance of this process to HIV transmission in vivo is not known.20 Whether epithelial cells themselves can be productively HIV infected remains controversial but a possible mechanism of viral acquisition.21,22
Potential innate defenses against HIV in the genital tract
The genital mucosa is not a sterile environment. The mucosal barrier of the genital tract is the first line of defense against invasion by potential microbial pathogens, including components of the host’s own microflora, and numerous innate antimicobial factors are active at the level of the genital mucosa. The microflora of the FGT, as well as serving as a reservoir for potential pathogens, also provides important protection against microbial invaders including HIV. In particular, the low vaginal pH that is maintained through hydrogen peroxide production by Lactobacillus species has antiviral activity and provides a less conducive environment for colonization by potential gram-negative bacterial pathogens.23
Several innate immune proteins with anti-HIV activity are present at mucosal surfaces, including the genital tract. The anti-leukoprotease Secretory Leukocyte Protease Inhibitor (SLPI) was first isolated from saliva, and this protein can potently protect activated CD4+ T cells from HIV infection at concentrations found in semen and saliva.24,25 SLPI is primarily active against HIV isolates using CCR5 as a co-receptor (R5 isolates), which are responsible for most HIV sexual transmission.24 Although the mechanism of action for this selective inhibition is not well defined, SLPI inhibited HIV infection through binding to the annexin II receptor on the macrophage surface, inhibiting infection at the post-fusion but the pre-reverse transcription phase of the viral life cycle.26 Another potentially important mucosal factor is lactoferrin, a component of breast milk and the genital tract that has been shown to inhibit HIV at the early stages of viral infection in vitro, most likely through preventing viral uptake.27–32
Defensins are small cationic proteins with significant antimicrobial effects. They are divided into two groups, both exhibiting anti-HIV properties; the α-defensins (produced by neutrophils, macrophages and γδ T cells) and the β-defensins 1-3 (produced by epithelial cells) primarily produced by epithelial cells of the mucosa.33 Interestingly, β defensins have been shown to have differential expression, where HBD-1 (human β defensin-1) is expressed constitutively while HBD-2 and HBD-3 are inducible and have been shown to inhibit HIV, particularly X4 type infections.34–36 Defensins may serve an important role as a ‘bridge’ between the innate and adaptive immune response, by inducing DC maturation through TLR-4 receptors.37 In addition to the defensins, other cationic proteins may play a role in defense against HIV infection. A proteomics approach to the identification of cervical innate factors with HIV inhibitory activity found that cationic proteins found in cervical lavage samples were collectively able to inhibit viral infection in vitro.38 The authors of this study postulated that the sum of all cationic proteins present in the vaginal fluid, which included defensins, SLPI, lactoferrin and several previously undescribed proteins, may serve as a critical defense against infection in the genital mucosa.
The type I (IFN α/β) and type II (IFN γ) interferons are also potent inhibitors of HIV infection, and may be produced by a variety of immune cells as a component of both the innate and adaptive immune responses.10,39 Type I interferons interfere with the HIV replication cycle through multiply described mechanisms, one of which is through its ability to induce the anti-viral activity of another intracellular innate defense molecule APOBEC3G (apolipoprotein B mRNA-editing enzyme-catalytic polypeptide-like 3G).40,41 IFN α can also stimulate antigen cross presentation between cytotoxic T lymphocytes (CTLs) and DCs.42 While several of these cytokines play a pivotal role in HIV-specific T-cell-mediated immunity, IFN α is expressed early in response to HIV infection, and may therefore be viewed as an innate defense against infection.43
In addition to the type I interferons, genital secretions contain significant levels of several other cytokines/chemokines with HIV inhibitory activity that may alter HIV susceptibility in vivo (see below). While CD4 is the primary HIV receptor, a co-receptor is also required for efficient cell entry. This is generally either CCR5 for the so-called R5 viruses that are responsible for most sexual transmission, or CXCR4 for the X4 viruses that may emerge as HIV disease progresses.44 Cytokines/chemokines present in genital secretions have selective ability to inhibit different strains of HIV based on their co-receptor usage: the chemokine SDF-1 (CXCL12), which is found within the subepithelial layer of the cervix, is able to inhibit X4 viruses competitively.45 Similarly, the β chemokines MIP-1α, MIP-1β, and regulated on activation, normal, T-cell expressed, and secreted (RANTES) are all ligands for the CCR5 receptor, and can block cell infection by R5 viruses.46 Nonetheless, as the latter chemokines are generally produced by activated immune cells (including CD4+ T cells), so there may be a fine balance between their anti-HIV activity and the increase in potential HIV target cells that their presence may herald.
In vivo innate immune correlates of altered HIV susceptibility
As noted above, the presence of anti-HIV activity in vitro does not mean that the presence and/or levels of a given innate immune protein will correlate with protection against HIV acquisition in vivo. This may be particularly true, if an inflammatory process leads to production of the innate immune protein, since the inflammatory process itself may ‘trump’ any increase in anti-HIV factors, either by directly breaching the mucosal barrier, recruiting potential HIV target cells to the genital mucosa, or by enhancing the production of other immune proteins that may enhance HIV replication. While numerous mucosal innate immune factors demonstrate in vitro anti-viral activity, studies correlating these factors with ‘real world’ HIV susceptibility are less common. However, several groups have examined mucosal levels of innate immune proteins in populations demonstrating increased or decreased susceptibility to HIV infection, and these results are summarized in Table I. Ideally, levels of an innate immune factor associated with protection against HIV should be increased in individuals with reduced HIV susceptibility, and reduced in individuals with increased susceptibility. However, immune factors meeting these criteria are rare, with the possible exception of SLPI, and this makes it difficult to prove that such factors are actually mediating HIV protection.
Table I. Innate Mucosal Immune Associations of Altered HIV Susceptibility
|SLPI||↓ Replication24||↓ in BV50|
↓ by nonoxynol-965
|↑ (in infant saliva) reduced MTCT48|
↑ (in maternal vagina) reduced MTCT47
↔ in HEPS49,50
|Trappin-2||↓ Replication51||No data||↑ in HEPS FGT**51|
|Lactoferrin||↓ HIV replication; ↓ epithelial adsorption30||↑ by BV, N. gonorrhoeae, C. trachomatis86–88||↑ in HEPS FGT50|
|CCR5 expression on genital CD4+ T cells||↑ Replication (R5 HIV strains)||↑ by HSV2, H. ducreyi89,90↑ by OC use63||↔ in HEPS FGT53|
|DC-SIGN expression on dendritic cells||↑ CD4+ T-cell infection in trans15||↑ by HSV289||No data|
|Defensins (α/β)||↓ Replication||↑ by T. vaginalis91|
↑ by cervicitis87
|↑ (in maternal breast milk) reduced MTCT 91|
↑/↔ in HEPS cervix49,54
|IFNa (type I interferons)||↓ Replication41,42||No data||↑ in HEPS cervix49|
|β chemokines (RANTES, MIP1α, MIP1β)||↓ Replication||↑ in BV50|
↑ by nonoxynol-965
|↑ in HEPS FGT53|
↑/↔ in HEPS cervix 49,50
|Inflammatory cytokines (IL-1β, TNFα, IL6, IL10)||↑ Replication||↑ in BV, STIs 67,68,86|
↑ by nonoxynol-969
↔ by HSV2 (in absence of shedding) 89
|↔ in HEPS FGT53|
|Epithelial cell integrity||↓ HIV infectivity||↓ by nonoxynol-993,94↓ by STIs95||No data|
Mucosal Innate Immunity in Populations with Reduced HIV Susceptibility
Elucidation of factors providing natural HIV protection is hampered by the fact that HIV infection is a relatively rare event after sexual exposure. Practically, this means that to demonstrate meaningful associations, studies must either recruit a very large number of participants with rare HIV exposures, or focus on smaller groups with very high HIV exposure levels: the latter has been a more common strategy. Groups that have been described as having reduced HIV mucosal susceptibility include (i) sexually exposed HEPS female sex workers, men who have sex with men, and individuals within an HIV serodiscordant relationship; and (ii) perinatally exposed (during vaginal delivery) or postnatally exposed (breast milk), uninfected children of HIV-infected mothers (see Table I). These groups are quite heterogeneous, with clear differences in the frequency of HIV exposure and the diversity of viruses to which individuals are exposed; in addition, factors related to the partners’ viral strain can only be examined in some of these groups. Such factors must be kept in mind before drawing conclusions from studies of such individuals.
Higher vaginal levels of SLPI in HIV-infected mothers have been associated with reduced HIV transmission during childbirth, and higher saliva levels in infants with reduced HIV acquisition through breast-feeding.47,48 However, the association between genital tract SLPI levels and reduced sexual HIV acquisition is less clear.49,50 When a proteomics approach was used to characterize novel mucosal proteins potentially associated with HIV resistance in over 500 Nairobi sex workers and lower risk women, the anti-protease Trappin-2 was elevated in HEPS individuals, and was shown to have anti-HIV inhibitory properties in vitro.51 While these preliminary findings require confirmation and further study, proteomics may represent a useful scientific approach to the elucidation of novel mucosal factors associated with HIV protection.
Levels of β-chemokine expression were increased in the blood of the HIV-uninfected partner within a discordant relationship, and the cervico-vaginal levels of RANTES (but not other β-chemokines) were increased 10-fold in Nairobi HEPS sex workers.52,53 In addition, increased RANTES expression was observed in cervical biopsy tissue from HEPS women using immunofluorescence, along with increased levels of IFNα.49 Genital defensin levels have been measured in several HEPS cohorts, and only α defensins have been increased in such individuals.54 No changes in β defensins were apparent in the cervix of HEPS women, although their cervical levels were increased in HIV-infected individuals.49 No studies have comprehensively examined the integrity of the genital epithelium in HEPS individuals, although rates of sexually transmitted infections (STIs) (including ulcerative infections) were similar in HEPS and HIV-susceptible female sex workers.55
Mucosal Innate Immunity in Populations with Increased HIV Susceptibility
There is strong evidence that certain natural or external factors are associated with increased HIV acquisition, and it is also interesting to consider the innate immune correlates of increased HIV susceptibility. While not comprehensive, factors enhancing susceptibility can be broadly divided into: infectious conditions [STIs, particularly Herpes simplex virus type 2 (HSV2)]; hormonal factors (progesterone-based and perhaps estrogen-based contraceptives); external substances applied to the vagina (douching and detergent-based microbicides such as nonoxynol-9); and alterations in the normal vaginal microbial flora [bacterial vaginosis (BV)].3,56–60
An intact genital epithelium is likely to be a key innate defense, and disruption of this barrier is likely to be a key reason for the increased HIV susceptibility seen in association with ulcerative STIs (HSV2, syphilis, chancroid).61 In addition, epithelial degradation because of the application of mucosal chemicals/detergents such as nonoxynol-9 and certain douching agents, or to physical trauma, can increase susceptibility to HIV infection.10 Changes in the physical characteristics of the vaginal epithelium under hormonal influence, particularly thinning caused by exogenous or endogenous progesterone, increase susceptibility both in a macaque model and in some human cohorts.10,62 However, the impact of oral (estrogen-based) contraceptives (OCs) on HIV susceptibility remains controversial, with some studies demonstrating increased susceptibility but others showing no effect.58 Interestingly, those demonstrating increased HIV susceptibility have generally been nested within high-risk groups such as sex workers. In addition to alterations in the physical properties of the vaginal epithelium and the quality and quantity of vaginal mucus, CCR5 expression may be upregulated on CD4+ T cells from the cervix of women taking OCs, providing a potential mucosal immune basis for increased susceptibility.63,64
In addition to the direct effects of STIs and chemical agents on the genital epithelium, they induce the recruitment of activated immune cells, increasing the number of mucosal HIV target cells. Even in the absence of HSV2 reactivation, infected women have increased numbers of both CCR5+ CD4 T cells and DC-SIGN+ DCs in the cervix, and nonoxynol-9 application increases cervical macrophage numbers.65 Influx of activated immune cells also leads to the release of inflammatory cytokines, such as interleukin-1 (IL-1), which can upregulate HIV infection in vitro.65,66 Increased IL-1 in genital secretions has been seen in association with several genital infections, including BV,67,68 as well as after a single vaginal application of nonoxynol-9.69 BV is not only associated with local inflammation in vivo, but with increases in vaginal pH because of reduced genital lactobacilli, reduced levels of SLPI and defensins 50,70, and may also enhance HIV replication because of undefined soluble HIV-inducing factors.71 While nonoxynol-9 use was also associated with a dramatic decrease in vaginal SLPI levels, β-chemokine levels were actually increased.65 The latter observation emphasizes the need for studies to examine the broader mucosal immune milieu, rather than focus on individual potentially protective mucosal factors.
The adaptive immune system: too little too late?
Although a variety of HIV-specific adaptive immune responses have been described in the genital mucosa of HEPS, including cellular immune and IgA responses, most global HIV exposures do not occur in such a context. Details of adaptive HEPS immune responses are beyond the scope of this review, and interested readers should refer to recent excellent reviews on the topic.72–74 When previously unexposed macaques were infected by simian immunodeficiency virus through a mucosal route, vigorous local, and systemic cellular immune responses were generated eventually, but lagged behind the dissemination of the virus, and appeared to represent ‘too little, too late’ on the behalf of the infected host.75 Therefore, it seems likely that the 99–99.9% efficacy of HIV protection seen in ‘non-HEPS’ people after sexual exposure is mediated by innate factors, as opposed to the later development of HIV-specific adaptive immunity. Of course, it is still possible that vaccine-induced HIV-specific immunity will further enhance this innate protection against HIV acquisition, and potentially have a dramatic impact on the global pandemic, although it seems unlikely that an effective vaccine will be available in the short term.
The genetics of HIV susceptibility
Reduced HIV susceptibility has been associated with a number of genetic polymorphisms: most strikingly, the CCR5-Δ32 mutation confers near complete resistance to sexual HIV acquisition.74 Natural HIV resistance has also been linked to polymorphisms of innate immune molecules, such as an SDF-1 3’ UTR mutation in a group of HEPS individuals, and with human leukocyte antigen (HLA) haplotypes including HLA-B18, HLA-A2/6802, HLA-DRB1*01, and HLA-A11.76–79 Although such observations have been very useful in elucidating HIV pathogenesis, and in the development of new antiretroviral drugs and candidate microbicides, rates of HIV sexual transmission remain consistently low even in the absence of known genetic polymorphisms, implying that they are not the main basis of ‘natural’ innate resistance to HIV acquisition.
Translating natural HIV protection into prevention
If innate protection against HIV is so effective, then how have research findings in this area translated into novel strategies to prevent HIV acquisition? Description of the CCR5-Δ32 mutation resulted in the new class of CCR5 inhibitor antiretroviral drugs, and the development of a PSC-RANTES analogue as a microbicide candidate.80,81 Similar advances have not yet been fuelled by descriptions of natural innate inhibitors of HIV, but there are areas of significant promise.
One such area is the design of microbicides that aim to enhance the pH barrier of the vaginal tract. This might be achieved through the (re-)establishment of lactobacillus species in the genital flora, and the FGT introduction of engineered lactobacilli that express additional protective factors, such as HIV fusion inhibitors, might provide even higher levels of protection.82,83 One major challenge to microbicides that modify the genital pH will be the alkaline nature of seminal plasma. In addition, a range of polyanionic microbicides that may interfere with HIV binding to mucosal target cells are in development.10 Despite the anti-HIV activity of innate mucosal cationic charged proteins, few related compounds are in development as candidate microbicides, with the exception of the theta-defensins, which are not normally expressed in humans.84 Population-based studies of STI prevention to reduce HIV transmission have had disappointing results, particularly in the setting of an established HIV epidemic, where most HIV transmission may occur in the context of stable couples without STIs.85 Large-scale trials are currently examining HSV2 suppression as strategy to prevent HIV transmission within stable, HIV serodiscordant couples.
Summary and future directions
As we continue to elucidate the innate factors that provide such efficient ‘natural’ protection against HIV acquisition after sexual exposure, it becomes clear that this protection is dependent on the synergistic activity of numerous mucosal factors. Levels of these factors may fluctuate with the hormonal cycle or during intercurrent infections, and measurement of just one or two of these factors may be inadequate to elucidate the effects of an intervention – such as a candidate microbicide – on HIV susceptibility. The identification of innate factors that are central to HIV protection will therefore both inform the development of new, safe microbicides, and also the monitoring of microbicide candidate safety/toxicity for those currently in clinical trials. Basing future microbicides on key innate proteins naturally found in the genital tract may have the advantage of reduced toxicity and demonstrated efficacy. If HIV protection rests not on one or two such key factors, but rather on the interplay of numerous factors, then techniques that are able to assess the entire genital mucosal proteome may be useful to assess microbicide safety.
This study was supported by grant from Canadian Institutes of Health Research (RK, grants #HOP-81735 and HOP-75350; SMI, salary support) and the Canadian Research Chair Program (RK, salary support).