According to UNAIDS, women represent more than 50 percent of people living with HIV/AIDS Worldwide [1]. Despite the low probability of Human Immunodeficiency Virus (HIV) transmission during a single episode of unprotected vaginal sex (from 1/200 to 1/12000) [2], heterosexual transmission is the fastest growing phase of the HIV pandemic. Even if the transmission risk per exposure is lower, vaginal intercourse contributes more to the epidemic than anal intercourse or parenteral inoculation [2]. The mucosa-associated lymphoid tissue (MALT) is an important element in the interaction between pathogens and the host immune system. Since mucosal tissues of the female genital (FGT) and intestinal (gut-associated lymphoid tissue, GALT) tract are the main sites of HIV infection, HIV is considered a mucosal pathogen. The impact of HIV on the mucosal immune system is critical for disease outcome, and the state of the peripheral immune system does not necessarily reflect these profound mucosal dysregulations. Here, we will review alterations of the mucosal immune system caused by HIV infection and we will discuss how the impact of constant stimulation on mucosal immunity at the FGT affects HIV infection in commercial sex workers.

The mucosal immune system plays a critical role in protecting the host against many pathogens and environmental antigens while inducing tolerance to non-pathogenic commensal flora (reviewed by [3]). Mucosa associated lymphoid tissues (MALT) are specialized organized lymphoid systems that contain the majority of the lymphocytes of the body. The MALT is a non-sterile environment where lymphocytes and epithelial cells work together to maintain the balance between activation and tolerance. Mucosal immunology is increasingly gaining attention because of its important potential in the development of vaccines against mucosal pathogens such as influenza and HIV. The mucosal immune system is, in many ways, more complex than its systemic counterpart. Studies in macaques infected with simian immunodeficiency virus (SIV) and humans infected with HIV have shown that the SIV/HIV infection rapidly kills most CD4⁺ T cells at mucosal surfaces [4-6]. The mucosal surface of the FGT is the main route of entry of HIV, and the mucosal immune system is rapidly perturbed by HIV infection more than any other immune compartment. The fundamental role played by mucosal surfaces in the pathogenesis of HIV infection has recently been a major focus of HIV research.

Irrespective of the route of entry, transmitted virus promptly crosses the epithelial barrier to reach target cells, resulting in an extensive, and possibly permanent, depletion of the resident mucosal memory CD4⁺ T cells. This phenomenon is observed in the intestine as well as in other mucosal tissues including the FGT [5,7,8] (reviewed by [9]). The massive mucosal CD4⁺ T cell depletion and the breakdown of essential immunoregulatory mechanisms leads to a severe impairment of mucosal defences, which results in a high rate of secondary opportunistic infections and contributes to the morbidity and mortality of HIV infected individuals [10]. Non-human primate (NHP) studies have demonstrated that during the first day following infection, many critical events occur at the mucosal level which affects the rate of disease progression. In humans, these events occur early in the course of the infection, prior to systemic alterations and sometimes months to years before the individual realizes his/her HIV status (reviewed by [11]).

The female genital tract (FGT)

The majority of female HIV infection occurs through the FGT. With a low pH, the presence of a natural microflora and many antiviral proteins in the vaginal secretions/cervical mucus, the FGT environment is naturally considered to be a hostile environment for HIV transmission and replication (reviewed by [12]). Multiple factors influence this natural barrier against infection and an imbalance in mucosal regulation may provide an opportunity for HIV to penetrate the barrier and initiate a productive infection.

Hormones are important factors that can modify the mucosal immune compartment of the FGT. Throughout the menstrual cycle, hormone expression leads to changes in the thickness of the epithelial layer, frequencies and activation status of immune cell populations, antimicrobial peptide expression and cytokine/chemokine expression [13]. High levels of progesterone seem to be associated with an increased risk of HIV acquisition [14]. Recently, the use of hormonal contraception has also been associated with increased risk of HIV acquisition, showing the importance of hormone regulation on the mucosal immune response [15]. However, hormones are not the only factor that can influence the mucosal environment of the FGT; sexual activities and sexually transmitted infections (STI) are also influencing the mucosal milieu (reviewed by [16]), as discussed later in this review.

Although the mucosal tissues of the FGT are severely affected by the virus, the biological impact of HIV infection on the genital tissues and its implications for HIV replication and disease progression have still not been well characterized. Due to the challenges associated with obtaining fresh samples and the limited number of cells available with cytobrush sampling, the impact of HIV on the FGT immune system still remains unclear. The first steps of infection have been characterized in non-human primate models. Intravaginal inoculation of SIV in macaques showed that within 14 days post infection, there was an increase of CCL20 (MIP-3a), which was associated with an increase of CD123⁺ plasmacytoid dendritic cells (pDC). Those pDC, in turn, produced CCL3 and CCL4, which attracted CD4⁺CCR5⁺ T cells to the site of infection and promoted the dissemination of the infection [17]. By inhibiting immune activation and cytokine/chemokine production, it could be possible to block systemic infection. Supporting this idea, Li et al. showed that applying glycerol monolaurate, a widely use antimicrobial agent that blocks T cell activation, at the cervico-vaginal mucosa can block the secretion of CCL20 and protects against SIV infection [17].

Knowledge on the interaction between HIV and human FGT is however limited. Even among healthy women, knowledge about mucosal immunity at the FGT is limited. Studies on vaginal and cervical tissue have gathered some pieces of information about the distribution of various immune subsets that could interact with HIV in these tissues. Immune cells from non-inflamed vaginal and cervical tissue derived from women undergoing hysterectomies showed a distinct lymphocyte distribution throughout the FGT. The lower vaginal mucosa contains few leucocytes with a majority of memory CD8⁺ T lymphocytes, while ectocervix tissue contains similar lymphocyte subsets with a higher proportion of CD4⁺ and CD8⁺ lymphocytes [18]. The transformation zone represents the transition between the multilayer epithelium of the ectocervix and the single-layer columnar epithelium of the endocervix. The immune cell-rich transformation zone was shown to be the main point of weakness for SIV entry but infection is not restricted to this area (reviewed by [19]). It is the location with the highest proportion of immune cells (CD8⁺ and CD4⁺ T cells, granulocytes and macrophages) including a small accumulation of CD56⁺ natural killer cells (NK) [20]. This innate immune subset combats HIV replication by non-cytolytic mechanisms including the production of anti-viral cytokines or by direct lysis of infected cells (reviewed by [21]). While the predominant NK cell population in the lower reproductive tract display the cytolytic CD56dimCD16⁺ phenotype, NK cells from the upper reproductive tract are mainly the cytokine-secreting CD56bright subset [22].

Distribution of antigen-presenting cell (APC) subsets localised in the FGT has been identified by immunostaining of ectocervical, endocervical and endometrial biopsies. Distinct populations expressing APC markers were present throughout these tissues, which harboured differing expression patterns of C-type lectins, important attachment factors for HIV (reviewed in [23]). Langherans cells (LC) expressing the langerin receptor were found within the epithelial layer while DCSIGN+ and/or mannose receptor (MR)+ DCs and macrophages were localised in the submucosal compartment along with CD123⁺ pDC and adjacent to CD4⁺CCR5⁺ T cell populations [24-26]. The role of LCs and submucosal DCs in HIV infection remains controversial. These cells are targets for HIV-infection and have the potential to disseminate the virus through C-type lectin-dependant and replication-dependant mechanisms. However, their function during in vivo infection remains to be clearly defined (reviewed by [2]). While well studied in the nonhuman primate model, the cascades of innate immune events that follow exposure to HIV and contribute to the spreading of the infection remain poorly described in the human FGT. Nonetheless, when viral dissemination occurs, it leaves profound immune alterations at the FGT.

Chronic HIV infection promotes a pro-inflammatory response at the genital tract resulting in higher mucosal levels of cytokine and chemokine expression that will, in turn, contribute to cell recruitment. MCP-1, MIP-1b, RANTES, MCP-3, MIG, TNF-a and IFN-g are overexpressed in HIV infected commercial sex workers (CSW) compared to HIV uninfected CSW and low risk women. In contrast, MIP-1a is underexpressed compared to HIV uninfected CSW [27,28]. These immunoregulatory molecules are involved in lymphocyte chemotaxis and the inflammatory response. Secreted by various cell types at the FGT such as epithelial, stromal or immune cells, MCP-1, MIP-1a, MIP-1b, RANTES and MCP-3 attract neutrophils, monocytes, macrophages, immature DCs, T cells and NK cells to the inflamed tissue, while MIG attracts activated T cells and NK cells (reviewed by [29]).

However, beta chemokines released by activated T cells such as RANTES, MIP-1a and MIP-1b may play a dual role during HIV infection. In addition to attract activated T cells, which increases the number of target cells for the viruses, they also inhibit cell infection. By competing with HIV for CCR5 receptor binding, these chemokines inhibit HIV infection in vitro [30-32] and their spontaneous or induced expression by PBMCs has shown to have a protective role against HIV acquisition among discordant couples (reviewed by [33]). Effect of betachemokines might depend on the timing of their secretion. Secreted prior to the infection, they could be protective by blocking the access to CCR5. However, their secretion following infection may increase the potential number of target cells and contribute to the establishment infection foci. Triggered during inflammatory responses, TNF-a drives HIV-1 transcription through the activation of cellular transcription factors (review by [34]). HIV-induced secretion of TNF-a by epithelial cells contributes to epithelial barrier impairment by disrupting tight junctions, suggesting an important contribution of TNF-a in viral transcytosis and dissemination across the epithelial barrier [35]. Interestingly, cytokine concentrations in CVL during the acute phase of HIV infection did not differ from the concentrations observed prior to infection [36]. This suggests that cervical cytokine concentrations may increase over time in the course of HIV-infection. In fact, elevated levels of cytokines in women with early HIV infection were correlated with the presence of STIs [36].

Infections with other STIs such as Herpes simplex virus type 2 (HSV-2), human cytomegalovirus (CMV), Neisseria gonorrheae, Chalmydia trachomatis and many others also enhance cytokine and chemokine expression at the FGT (reviewed by [37]). STIs influence HIV susceptibility either by breaching the epithelial tissue, recruiting HIV target cells into the site of infection, or by generating a proinflammatory local immune milieu (reviewed by [37]). Elevated levels of pro-inflammatory cytokines are observed in CVL from symptomatic and asymptomatic women infected with a broad array of STIs [38,39] and higher concentration of inflammatory cytokines (IL-1β, IL-6, IL-8, and sCD40L) in CVL specimens correlate with an increased HIV acquisition among South African women, especially among Nesseria gonorrhea infected women [38]. HSV-2 infection negatively synergized with HIV in the mucosal compartment to favor viral replication. Indeed, HSV-2 infection increases the number of potential HIV target cells at the FGT. In turn, HIV infection impairs the HSV-2 mucosal immune control and local HSV-2 reactivation increases the risk of transmission of both viruses [40]. Women with cervico-vaginal inflammation due to bacterial vaginosis or cervicitis had higher level of pro-inflammatory cytokines such as IL-1b and IL-6 and this proinflammatory environment is likely to be important in susceptibility to HIV infection [41].

In addition to profound alterations of the cytokine/chemokine environment, the FGT of HIV-positive women exhibits fundamental changes in cells populations such as macrophage and lymphocyte populations and in immunoglobulin (Ig), even in the absence of symptoms [42]. Immunohistological studies of the ectocervix have shown an important redistribution of macrophage subsets during HIV-infection. The epithelium of HIV-infected women showed an increase in both the number of CD68⁺ macrophages and the proportion of activated (HLA-DR+) subsets, with an increased proportion of macrophages harbouring a suppressive phenotype in the stroma [42]. CD1a+ LC subsets acquired a better capacity to present antigens and were depleted in the stroma of HIV-positive women, as observed during cervicitis [20]. This decline may be explained by a rapid LC emigration from the epithelium, similar to what is seen during infection of ex vivo organ culture (reviewed by [2]). These observations suggest the presence of a local mechanism for T cell activation, however the suppressive phenotype displayed by the local macrophages may impair the emergence of a local immune response [42].

An important gap in the knowledge is the impact of HIV-infection on innate cell populations of the FGT. Well characterized in the uterine compartment where they are abundant and play an important role in reproductive biology (reviewed by [43]), NK cells seem to be scarce in the lower FGT, i.e. the ectocervix compartment, of HIV-infected and uninfected women [42] and are poorly described in the FGT. While observed in the higher part of the FGT (endocervix and transformation zone) of uninfected women [20] and in vaginal biopsies of monkeys [44], no studies have addressed the impact of HIV-infection on NK phenotype and distribution in the human FGT. We have observed in cervico-vaginal mononuclear cell (CMC) samples that NK cells comprise about 10% of the CD3 negative lymphocytes population of the FGT, harbouring a more activated phenotype (CD69⁺) among HIV-infected CSWs (Lajoie and Kirwan, Unpublished data).

Gamma-delta T cells are another important mediator of innate immunity that plays an important role bridging innate and acquired immunity. Activated by stress-induced molecules without the need for antigen processing and presentation, gamma-delta T cells also recognize target cells by a NK-like mechanism of activating and inhibiting receptors and participate to the local inflammatory response by secreting cytokines and chemokines (reviewed by [45]). During systemic HIV infection, gamma-delta T cell subsets are perturbed. Two subsets of gamma-delta T cells are commonly found in humans. The Vdelta1 subset expresses tissue regulation genes (IL-10 and IL-11) and is the predominant subset found in mucosal compartment [46]. The Vdelta2 subset is more prevalent in the circulation and expresses pro-inflammatory genes upon mitogen stimulation [46]. Inversion in the ratio of Vdelta2 to Vdelta1 is observed in HIV-infected subjects. Contraction of the circulating Vdelta2 subsets occurs concurrently with expansion of the Vdelta1 subset in the gut of HIV infected patients [47] (reviewed by [45]). Circulating Vdelta2 T cells become anergic in HIV-infected patients and are depleted by an apoptosisdependant mechanism (reviewed by [45]). These alterations persist despite ARV therapy [47,48]. Despite early reports of gamma-delta T cells derived from the human cervix [49] recent studies have failed to identify a gamma-delta T cell population in the lower FGT during non-inflammatory conditions [42]. Monkey studies, however, have demonstrated an important homing of gamma-delta T cells to the cervico-vaginal mucosa of monkeys immunized by a gp120/p27/alum vaccine targeting the iliac lymph node [50]. Further analyses have demonstrated that gamma-delta T cells protected macaques against SIV infection by secreting beta-chemokines (RANTES, MIP1-a and MIP-1b) that block CCR5 binding and prevent infection of new target cells [50].

In addition to gamma-delta T cells, CD1d-restricted invariant NKT (iNKT) cells represent another rare lymphocyte subset that link innate and adaptive immunity. At this time, no reports of the presence or absence of iNKT populations at the human female genital tract have been published, despite the presence of CD1d expression in vaginal tissues [51,52]. In mice, however, iNKT activation is protective against genital tract Chlamydia muridarum infection [53], demonstrating a role for iNKT activation in mucosal immunity that has yet to be confirmed in humans. Data on innate cell subsets at the FGT during HIV-infection are scarce and we are far away to fully understand their contribution to HIV acquisition and pathogenesis. A better comprehension of the role of these subsets at the FGT is needed in order to harness them for vaccine and microbicide development.

Concurrent to the increased levels of cytokine and chemokine expression, elevated numbers of activated CD8⁺ T cells can be found at the FGT of HIV-infected women, indicating active recruitment of nonresident CD8⁺ lymphocytes [42]. Very few studies have characterized the HIV-specific T cell response in the FGT of HIV infected women. HIV-specific CTL responses are much less frequent in the cervical compartment than in the periphery [54], yet, following suppression of viremia by ART, CTL responses are preserved in the cervix while lost in the blood [55]. Gumbi et al. analysed the link between mucosal inflammation and HIV-specific CD8⁺ T cell responses in the cervix of HIV infected women. They found that genital HIV shedding was associated with an increased level of pro-inflammatory cytokines such as TNF-a, IL-1b, IL-6 and IL-8, although there was no association between HIV shedding and the cervical HIV-specific CD8⁺ T cell response [56]. These results support the hypothesis that, during infection, release of viral particles at the genital tract is exacerbated by inflammatory cytokines and that the CTL response is not sufficient to contain HIV replication. CD4⁺ T cells, preferentially IL-17 and IL- 22 expressing T cell populations, are depleted in the cervix following the establishment of HIV infection [7,54]. While well described in the gut, the relevance of Th17 and Th22 cell depletion in the FGT is not understood. It has been suggested that depletion of these populations may contribute to the increased frequencies of candidiasis and others genital tract infections in HIV-infected women [7].

Deregulation of the mucosal immune system also influences systemic immunity. High genital levels of pro-inflammatory IL-1b, IL-6 and IL-8 were associated with lower systemic CD4⁺ T cell counts while MCP-1 positively correlated with plasma viral load [57,58]. In early infection, elevated cervicovaginal lavage (CVL) pro-inflammatory cytokines were associated with lower blood CD4⁺ count and higher blood viral load and genital HIV RNA concentrations [36]. These observations suggest that alterations of the mucosal compartment affect the systemic compartment. ARV therapy may restore CD4+ count at the FGT since higher cervical CD4⁺ T cell percentage was observed in treated women [55]. However, ART does not seem to entirely suppress viral production at the FGT, as low levels of virus can still be detected even if plasma viral load is undetectable [55]. To our knowledge, no studies have addressed the impact systemic cytokines/ chemokines expression on the FGT immune system. Further studies need to be done in order to understand the relationship between both compartments.

Even if there is still an important lack of knowledge about the impact of HIV on the mucosal immune system of the FGT, it is quite certain that the impact of HIV on the mucosal environment is more important than previously realized. A growing body of studies have already begun to unravel the impact of HIV infection on the gastrointestinal tract (GALT), the most studied compartment of the mucosal immune system.

The gastro-intestinal lymphoid tissue (GALT)

The gastro-intestinal tract contains the majority of the CD4⁺ T cells in the body [59] and is an important site for early viral replication and severe CD4⁺ T cell depletion during HIV infection. Compared to the FGT, we have a better comprehension of the impact of HIV/SIV infection on the GALT environment. The magnitude of T cells lost in the GALT of an HIV infected patient is not reflected in the peripheral blood. Studies have shown that early in infection there is a drastic loss of the memory T helper cell population. Data from multiple groups showed that, in the first days post infection, nearly 60% of the CD4⁺ T cells are depleted [4-6], and that this depletion is largely irreversible. Indeed, HIV-infected individuals on anti-retroviral therapy show a slow and incomplete restoration of CD4⁺ T cells in the GALT compared with the systemic compartment [60]. HIV-specific CTL response may, in part, protect against this rapid decline by limiting mucosal replication during chronic infection. The level of Gag-specific CD8⁺ T cells is more elevated at the rectal mucosa than in the systemic compartment and is positively associated with CD4 counts and inversely with the plasma viral load [61].

Various T cell populations are differentially affected by the presence of the virus. Th17 and Th22 populations are involved in the generation of immune responses to microbial pathogens at mucosal surfaces and play a critical role in normal barrier homeostasis [18,62] (reviewed by [63]. In HIV infection, Th17 and Th22 cells are depleted more severely than the bulk CD4⁺ T cell population in the intestinal compartment [64,65]. The strong depletion of these cells is associated with the development of malabsorption and nutriment complications in rhesus macaques [66] and with a decreased expression of markers regulating barrier integrity and intestinal function [67,68]. This change also coincides with increased inflammation and expression of immune activation genes. Damage at the mucosal barrier seems to increase absorption of microbial products at the systemic level. The lost of barrier integrity is associated with elevated plasma levels of LPS (and sCD14), which in turn contributes to chronic immune activation and the eventual loss of CD4⁺ T cell-mediated immunoregulation [67].

In the gut, gamma-delta1 T cells are important players in maintaining mucosal homeostasis, and represent about 50% of the intrepeithelial lymphocytes and 10% of the lamina propria lymphocytes (reviewed by [69]). In HIV-infected patients, high levels of immune activation, as measured by elevated levels of B2-microglobulins and neoprepin, inversely correlated with duodenal gamma-delta intraepithelial lymphocytes, suggesting a role for these cells in limiting the immune activation in patients [70]. Underscoring this hypothesis, a decrease in the mucosal gamma-delta T cell count was observed toward the end of life in AIDS patients [70].

HIV-infection also alters mucosal NK cell subsets. Mounting evidence suggests an important role of NK cells in the control of HIV replication. Control of HIV replication has been associated with certain HLA-Bw4 alleles, including HLA-B57 and HLA-B27, and their inhibitory and putative activating NK receptors, killer immunoglobulinlike (KIR) 3DL1 and KIR3DS1 respectively. Combinations of KIR and HLA alleles have been associated with HIV disease outcome, with enhanced NK cell function delaying AIDS progression (reviewed by [71]). A recent study has observed the presence of two distinct NK cell subsets residing in the gut with altered frequencies during HIV-infection [72]. Using immunochemistry of gut biopsies, Sip et al. [72] demonstrated that intraepithelial and lamina propria NK cells express two different isoforms of Nkp46, a natural cytotoxicity receptor expressed on all NK cells. In addition to harbouring a more terminally differentiated phenotype (CD57⁺), intraepithelial NK subset were observed more frequently among individuals with the protective KIR/HLA genotype associated with spontaneous control of HIV replication, implying a potential implication for intraepithelial NK in the control of viremia. Moreover, the fact that both subsets expanded in the gastro-intestinal tract of HAART treated HIV-positive subsets with incomplete CD4+ reconstitution has led to the suggestion that these cells accumulate in absence of viral replication to potentially compensate for the GALT compromised immunity [72].

Over the past few years, our understanding of the role of the mucosal immune system in HIV pathogenesis has improved. The challenge now is to understand the full implications of the deregulation of the mucosal environment in HIV pathogenesis as well as the impact of perturbations in mucosal immunity on the systemic immune system. The elucidation of mechanisms to prevent mucosal T cell depletion and the impairment of the mucosal immune system is critical to move forward in the fight against HIV infection.

Important insights on the pathogenesis of HIV infection come from the studies of monkey models. Infection of the African natural hosts, like African Green Monkeys (AGM) and Sooty Mangabeys (SM), with Simian Immunodeficiency Virus (SIV) leads to a non-pathogenic SIV infection. In marked contrast, infection of rhesus macaques results in a rapid progression to AIDS that mimics human HIV infection. Unlike pathogenic models, natural primate hosts preserve a healthy level of systemic CD4⁺ T cells and do not progress to AIDS despite high viremia [73] (reviewed by [74]). One of the main characteristics of this model is the limited immune activation associated with the presence of SIV [75]. Like HIV infection in humans, SIV infection of natural hosts is associated with a rapid and profound depletion of the mucosal memory CD4+ T cell subset [76]. However, in natural hosts, the mucosal immune activation that characterizes pathogenic primate models and HIV infected individuals is not observed [76]. In the natural host model, the mucosal T cell depletion observed during the acute phase of infection does not progress and reaches a stable plateau (in SM) or is followed by a significant recovery of these cells (in AGM) during the chronic phase of the infection [77] . In the monkey pathogenic model, as well as in HIV infected rapid/normal progressors, an important loss of Th17 cells in the MALT occurs. This depletion of the Th17 population does not happen in the non-pathogenic model and might explain why the mucosal integrity remains intact [78]. Another rather important factor that distinguishes between the pathogenic and non-pathogenic monkey models is the rapid resolution of the innate immune response observed in non-pathogenic infection. Rhesus macaques and humans show a persistent type I interferon response throughout the chronic phase of infection while natural host resolve it within 4 to 8 weeks. Hence, it has been proposed that natural host of SIV actively downregulate the innate and adaptive immune responses to the virus to avoid progression to AIDS (reviewed by [74]). Rapid resolution of innate response might contribute to the maintenance of the mucosal compartment and barrier integrity preventing uncontrolled immune activation and disease progression.

Among those who are HIV infected, some individuals have the incredible capacity to naturally control HIV replication or maintain normal CD4⁺ T cell counts; these individuals are termed Elite Controllers (EC) and long term non progressors (LTNP) respectively. Like in the SIV natural host capable of maintaining a normal CD4⁺ T cell count despite the presence of active SIV replication, long term non progressors (LTNP) present a normal CD4⁺ count and lower expression of immune activation and mucosal inflammation genes than HIVinfected individuals (as measured in microarray analysis) [79,80]. Low levels of systemic immune activation are not, however, the only factor contributing to a lack of disease progression in EC/LTNP populations. Despite low viral loads and normal CD4⁺ T cell counts, HIV infected ECs have higher generalized immune activation than HIV negative individuals [81,82] suggesting an important role in maintaining low mucosal immune activation for delayed disease progression. Interestingly, EC patients do not suffer from the mucosal CD4⁺ T cell depletion observed among natural host NHPs or HIV progressors [4] and elevated mucosal CD4⁺ T cell and Gag-specific CD8⁺ T cell responses have also been observed among LTNP patients [80]. These data seem to indicate that despite elevated immune activation, EC and LTNP individuals are able to maintain improved mucosal immune responses that might explain their slow progression to AIDS.

These findings are quite intriguing. How natural hosts, ECs and LTNPs succeed in maintaining normal mucosal immune function despite T cell depletion is an unanswered question. What drives mucosal immune activation and chronic inflammation in normal HIV infected individuals? Is mucosal immune activation responsible for disease progression? The capacity to regulate mucosal immunity seems to be important for protection against HIV progression, but can the same be said for HIV acquisition?

HESN and protection against HIV

Over the last 30 years, many groups have studied the correlates of susceptibility to HIV, but some of the most interesting findings are provided by studying HIV-exposed seronegative (HESN) individuals. Despite repeated exposure to HIV, those individuals exhibited a natural resistance to HIV infection. Many cohorts of HESN have been identified, including commercial sex workers (CSW), discordant couples, hemophiliac individuals and children born from HIV infected mothers. Studying HESN has been an important step in understanding determinants of natural protection against HIV acquisition.

Many observations arising from these cohorts have led to the identification of different factors involved in natural protection. Numerous studies have examined the immunological and genetic correlates of protection among HESNs (for more details see [83,84]). One of the most studied correlate of protection is a deletion of 32 base pairs in the CCR5 gene. This polymorphism, found in 10% of the European population, has not been observed in the majority of HESN cohorts and cannot explain all cases of natural protection [85]. The presence of systemic HIV-specific T cell responses [33,86], polymorphisms in the Interferon Regulatory Factor (IRF) -1 gene [87-89], carrying homozygous KIR3DS1 genotype or combined KIR3DL1 high expressor alleles and HLA-B*57 genotype [90,91] and increased mucosal levels of anti-proteases are among others factors that have been described to explain this natural protection against HIV infection [92]. Compared to the GALT, the female genital compartment harbours IgG as the dominant antibody isotype. In the FGT, IgG and IgA originate from blood circulation and local production (reviewed by [93]) and functional capacity of IgA to neutralize HIV-1 and block trancytosis across the epithelium has been demonstrated [94,95]. However, the protective role of cervical IgA in HIV acquisition remains controversial. While some studies have reported a higher proportion of cervical IgA [95-97] among HESN CSWs and discordant couples, others reported its absence [98,99] or correlated its frequency at FGT with higher daily exposure to the virus [95,100]. Hence, it is still unclear to what extent cervical IgA protects against HIV.

Protection against vaginal HIV acquisition has also been correlated with stronger interactions between NK and immature DCs. Immature DCs (iDC) from HESNs induced higher IFN-g production from NK cells that become more potent in eliminating autologous iDCs upon in vitro isolation or differentiation [101]. In addition to altered NK-DC cross-talk, several studies looking at KIR genotypes or NK activities have linked increased NK activity to protection against sexually and intravenously acquired HIV infection [102-104] (reviewed by [33]). Recently, we’ve developed a model of protection termed Immune Quiescence (IQ) to explain HIV resistance [105] among the Pumwani cohort of CSWs in Nairobi, Kenya. Decreased levels of T cell activation and host proteins required for HIV-1 replication were observed in HESN sex workers that limit available target cells for viral replication [105]. Factors that affect lymphocytes trafficking to the FGT and that regulate the inflammatory state of the genital mucosa could be key contributors to the induction of immune quiescence and, therefore, protection to HIV [106,107].

In addition to hormonal cycles, sexual activities influence the immune environment of the FGT. Contact between seminal fluid and cells of the FGT triggers a cascade of events involved in clearance of seminal debris, selection of fertilizing sperm, and induction and maintenance of immunological tolerance toward paternal antigens (review by [16]). Following coitus, seminal fluid induces a proinflammatory response illustrated by the secretion of GM-CSF, IL-6, IL-8, IL-1A, MCP-1 and CCL20 by epithelial cells [108-111] (reviewed by [112]). This local increase in cytokines and chemokines correlates with the recruitment of neutrophils, macrophages, DCs and lymphocytes in the epithelial and stromal layers of the cervix [110,113]. The remodeling of cervical immune cell populations following coitus might, therefore, affect HIV acquisition at the mucosa.

Fluctuation in seminal components between individuals causes variations in the pattern of cytokine synthesis and leukocyte recruitment during this early inflammatory response [114]. The HIV status of the male partner also influences cytokine concentrations in seminal plasma, and the increased concentration of seminal TGF-b observed during the chronic phase of HIV infection correlates with lower secretion of pro-inflammatory mediators by semen-exposed genital epithelial cells [115]. In addition to TGF-b, sperm contains prostaglandins, IL-8 and many other regulatory factors [110,111,114-116]. TGF-b contributes to the early inflammatory response but also to the induction of the Treg population involved in tolerance to paternally derived antigens [117]. In addition, exposure to semen during unprotected sexual intercourse results in peripheral T cell proliferation in response to the partner’s cells [118,119]. Nonetheless, this alloimmune response to a partner’s semen appears to be subject to important inter-individual variations. While the majority of individuals develop an immune response against their partner’s cells, a few become tolerized and respond to allostimulation with a pattern of immunoregulation that cannot be explained by the matching of HLA class II alleles [118,119].

Sex work results in constant immune stimulation as a consequence of exposure to seminal fluid from multiple partners rather than only one partner. Some studies have looked at the impact of commercial sex work on systemic immune responses and found that sex work results in suppression of the alloimmune response, decreased of CXCR4 expression on naïve CD4⁺ T cells, and an elevated proportion of activated CD8⁺ T lymphocytes [120,121]. Previous studies have compared cytokine expression patterns from CVL and blood to conclude that immune activation of the systemic compartment does not reflect the level of activation at the mucosal compartment [27,28]. But what happens in the vaginal compartment upon frequent sexual activities? So far, few studies have looked at the impact of sex work on immune activation at the FGT. Shedding light onto this question, our group has observed that the duration of sex work positively correlates with higher proportion of NK cells in CMCs and a reduction of cytokine/chemokine expression in CVL of CSWs from Nairobi, Kenya (Lajoie et al. In preparation). After two years of commercial sex work, HIV-uninfected women harbored lower levels of MIP-3a, ITAC, MIG, IL-1a, IL-1b, IL-Ra, IL-8, IL-10, IP-10, MCP-1, MDC, MIP1-a, MIP- 1b and TNF-a in CVL compared to sexually active, HIV-negative, non-CSW women. This difference was even more pronounced among individuals with a history of >10 years of sex work. In a recent study, cervical and peripheral cellular responses to mitogen stimulation were compared between low-risk women and CSWs active in the field for 1 to 38 years. Following mitogen stimulation, PBMC and CMC from CSWs demonstrated a reduced production of Th17/Th22-related cytokines in both compartments, but no difference in the magnitude of the immune response was observed according to duration of sex work [107]. Immunostaining of endocervical biopsies demonstrated that active sex work alters DC phenotypes at the FGT. Despite no observed DC accumulation, DCs subsets from CSWs expressed higher levels of CLR receptors compared to women from the low-risk population [24]. This distinct pattern of CLR expression in women practicing active commercial sex work might be induced by the inflammatory response triggered upon frequent exposure to sex-derived antigens and could increase the HIV susceptibility of CSWs [24].

However, we have to keep in mind that most of these studies have compared HESN CSWs to sexually active low-risk women from the general population. Therefore, it is impossible to discriminate the effect of frequent exposure to sexual products of multiple partners to the pressure exerted by HIV upon the accumulating years of commercial sexual activities. These years of exposure may result in selection of women having low levels of baseline immune activation in the genital mucosa. Both factors can play a significant role in protection, and so far, none of them have been excluded. Any factors that are capable of influencing immune activation at FGT are likely to play an important role in HIV susceptibility. How sex work affects susceptibility to HIV infection is a question that remains to be understood.

Intriguing findings regarding this question comes from an observation from the Pumwani cohort in Nairobi. CSWs from this cohort experience many unprotected sexual exposures, more than 500 during their years of commercial sex work, and despite that, they remain uninfected [122,123]. However, this protection is not absolute and late seroconversions of previously resistant CSWs have occurred and were associated with a reduction of daily clients or interruption in commercial sexual activities. Retirement from sex work results in the loss of peripheral HIV-specific immune responses, but some late seroconversions cannot be entirely explained by a lack of preexisting HIV-specific CTL responses or by viral CTL escape [123]. Why the protection against HIV infection wanes during the arrest of sexual activities and what drives this protection has remained unanswered.

We therefore performed a preliminary study to assess the levels of mucosal and systemic immune activation in CSWs already planning a break from commercial sex work. We observed that an interruption of at least 3 weeks had a major impact on the immune system compared to CSWs that maintained their commercial sexual activities. Sex work is clearly a source of ongoing immune stimulation at the FGT, and an arrest in commercial sexual activities led to a decrease of the activation markers on peripheral and cervical lymphocytes, particularly among HIV-infected CSWs, as illustrated by CCR5 expression (Figure 1A). Surprisingly, even in the context of HIV-induced immune activation, interruption of sex work significantly reduces immune activation of HIV-positive CSWs, demonstrating the strong impact of allostimulation on peripheral and mucosal immune activation. Importantly, HIV-negative individuals experienced a dramatic increase in mucosal activation upon resumption of sex work (Figure 1B), and it is known that higher levels of cellular activation correlate with higher cellular susceptibility to HIV infection [124,125]. By increasing the number of potential HIV targets at the FGT, interruption of sex work may render HIV-negative women more susceptible to HIV acquisition. These observations may partly explain the previously observed late seroconversions of resistant women [123].

Figure 1: Impact of an interruption in commercial sex work among HIV infected and uninfected CSWs. (A) Proportion of CCR5 expressing peripheral and cervical CD4⁺ and CD8⁺ T cells among HIV-infected CSWs that interrupted their commercial sexual activities for 4 to 8 weeks (HIV+ Sex Interruption, blue line). Proportion of CCR5 expressing cells from women that remained active in sex work (HIV+ control, black line) is also represented. HIV-positive Controls (n=8), HIV-positive SWI (n=10). B) Frequency of CD69 and HLA-DR expression of cervical mononuclear cells in HIV-negative that were newly enrolled in the Pumwani Sex Worker Cohort (New Negative) during commercial sex work (SW), following an interruption in SW and a resumption to commercial SW. New Negative (n=5). Difference between time points was calculated by one-way Repeated Measures ANOVA, if significant, P-values were calculated by Paired t tests.

What contributes to the regulation of the levels of immune activation in the FGT and why sexual activity seems to influence this process are questions that require attention. As mentioned previously, it has been reported that HESNs present a quiescent phenotype at the FGT and periphery and this low level of activation is less suitable for the establishment of HIV infection and may contribute to protection [105-107,125,126]. The way the mucosal immune system responds to commercial sexual activities and HIV pressure could contribute to this quiescent phenotype. And how HESN CSWs regulate their immune response to that constant immune pressure might be the key to their protection. What could drive this state of quiescence in HESN CSWs during their regular days of work and how does this protection wane when they interrupt commercial sexual activities? These questions need to be explored.

Pivotal in the regulation of the immune response, Treg subsets, that constitute 5-10% of CD4⁺ T cells [127], may play an important role in shaping immune balance at the FGT and protecting against HIV acquisition. Frequencies of peripheral Treg among heterosexual couples practicing unprotected intercourse negatively correlate with in vitro HIV infectivity of CD4⁺ T cells [119]. Emerging population of Treg limits allogenic stimulation and contributes to reduced immune activation and HIV replication [119,128]. Moreover, HESN CSWs harbor higher frequencies of peripheral Treg, which correlates with a lower proportion of CD69⁺ activated T cells and lower cellular infectivity [125,129]. By actively regulating levels of immune activation at periphery but also at mucosal sites, Treg may contribute to the protection of HESNs.

But what drives the emergence of Treg? Sex itself might be a part of the answer. In a mouse model, seminal fluid regulates the accumulation of Treg at the FGT [130]. Semen contains high concentrations of TGF-b [114,115,131] and exposure to seminal plasma converts naïve T cells into functional FOXP3-negative Treg that secrete TGF-b [132]. Constant exposure to antigen also influences the emergence of Treg. Mucosal sites, including the vagina, are responsible for maintaining the delicate balance between induction of tolerance to commensal microbiota and the capacity to mount an effective immune response against pathogens [133] (reviewed by [134]). In the gut, repeated exposures to low-dose antigen will induce the development of regulatory T cell populations while high doses exposure induce anergy of T cells [135] (reviewed by [136]). In this way, it is reasonable to believe that frequent exposure to stimulants contained in semen (including HIV) might affect T cell responsiveness at the FGT of CSWs. Likewise, T cell receptor signaling pathways and antigen processing and presentation pathways are down-regulated in HESN CSWs compared to age-matched low-risk HIV negative control group [105,126]. Despite quiescent baseline cell activity, these women did not have a defect in their capacity to fight infection because they have the same ability to respond to recall antigen stimulation as controls [105]. Hence, constant exposure to semen could affect the threshold of T cell activation in HESNs, without impairing their capacity to mount an efficient immune response. To date, the frequency of Tregs in the FGT of HESN has poorly been explored.

Observations from a recent study that immunized rhesus macaques with a live-attenuated SHIV lentiviral vector producing continual low doses of virions support the idea that continuous exposure to low dose antigen might alter vaginal immune environment. Six to eight months following rhesus monkeys immunization, immunized monkeys experienced a decrease in immune activation levels, consistent with reduced levels of vaginal indoleamine 2,3-dioxygenase expressing DC, normally triggered by pro-inflammatory signals (reviewed by [137]), and a reduced number of blood pDCs with impaired capacity to produce IFN-a compared to non-immunized monkeys. Three days following SIV challenge, CD4⁺ FOXP3⁺ Treg populations expanded in the vaginal tract of immunized monkeys, and this expansion correlated with lower activation at the genital tract. These results suggest that SHIV-immunization had specific effects on innate immunity and immunoregulation that collectively increased the threshold for T cell activation resulting in avoidance or quick resolution of innate inflammatory responses in immunized monkeys after SIV challenge. The modest antiviral effector CD8⁺ T cell response elicited by SHIV immunization could then participate in the elimination of the few HIV-infected cells at mucosal sites. Thus, a modest CTL response, insufficient to control HIV replication in other immunization models, in the context of mucosal immunoregulatory T cell populations and an anti-inflammatory environment was able to maintain control of SIV replication [138].

After more than 30 years of HIV research, numerous pieces of the puzzle are still missing. In the past few years we have begun to take a fresh look at the different compartments of the immune system and realized that immune dysregulation caused by HIV is even more striking in mucosal compartments. HIV infection alters homeostasis of the mucosal immune system at the GALT but also at the FGT. Conversely, we also came to understand that immune regulation of the mucosal compartments affects susceptibility to HIV acquisition. The fine balance between target cells and effectors cells at the FGT is likely to be an important determinant of HIV resistance. Understanding mechanisms of protection is not an easy task and few longitudinal studies have looked at the variation of immune parameters over time and sexual exposure in HESNs. Protection might rest on the capacity of HESNs to rapidly regulate pro-inflammatory responses triggered by commercial sexual activities or HIV exposure while maintaining innate and adaptive defenses that clear the infection, observations that are missed in the setting of a cross-sectional study. Many environmental or genetic factors may contribute to the emergence of protective cell populations. The maintenance or generation of a regulatory environment as indicated by fewer activated T cells may be an important factor in limiting HIV target cells at the FGT.

Over the last few years, it became more obvious that mucosal immune system plays an important function in regulating the entire immune system activation, however, we are still facing an important lack of knowledge. There is an urgent need for future studies in order to have a better understanding of the role of mucosal immune regulation on the risk of HIV acquisition and/or progression to AIDS.

The authors declare no conflicts of interest.