Professor Kristina Broliden, Unit of Infectious Diseases, B2:00, Karolinska University Hospital, Solna, 171 76 Stockholm, Sweden. (fax: +46 8 7178501; e-mail: firstname.lastname@example.org).
The development of HIV-1 vaccines and microbicides remains hindered by our limited understanding of correlates of immune protection to infection. Evidence indicating that resistance to HIV-1 infection is indeed possible comes from HIV-1-exposed yet uninfected individuals, including cohorts of commercial sex workers and discordant couples. Despite their uninfected status some of these individuals have mucosal and systemic HIV-1-specific humoral and cellular immune responses in addition to their innate immune response. The combined contribution of innate and adaptive immunity as well as genetic factors is most likely of great importance for this protection against infection. Here we review data on the antibody responses and secreted immune molecules of the innate immune system in the female genital tract with emphasis on individuals who seem to resist HIV-1-infection despite repeated exposure to the virus.
The global HIV-1/AIDS epidemic is without a doubt one of this century's worst tragedies and its consequences are readily felt on social, demographic and economic fields. The viral infection has been accompanied by denial, stigma and discrimination from the moment HIV-1 was identified. The 2006 UNAIDS report on the epidemic provides us with numbers that are hard to grasp; approximately 40 million individuals are currently infected with HIV-1 of which 4 million were newly infected in 2005. Moreover, in 2005 nearly 3 million individuals lost their lives to AIDS . Worldwide, women make up the fastest-growing group of people living with HIV-1/AIDS, especially in the epicentre of the epidemic; the sub-Saharan region of Africa, being home to almost 64% (∼25 millions) of all people living with HIV-1. According to the 2006 UNAIDS report , women account for approximately 60% of all HIV-1 infected individuals in the region and amongst young people aged 15–24 years, the percentage reaches 75%. More than 90% of all adolescent and adult HIV-1 infections have resulted from heterosexual intercourse and due to social and biological factors women are particularly vulnerable to this route of transmission. Because of social inequalities women are indeed victims of a high prevalence of nonconsensual sex and they can rarely control condom use or the eventual high-risk behaviours of their partners. The main biological factor contributing to female vulnerability to heterosexual transmission of HIV-1 is the substantial mucosal exposure of the female genital tract to seminal fluids. However, we still lack knowledge about the relative contribution of different genetic and immunologic factors that may contribute to resistance against infection.
Anatomy of the female genital tract mucosa relevant to HIV-1 transmission
In order to better understand the biology of HIV-1 transmission upon sexual exposure the anatomy of the female genital tract is summarized. The endocervix is thus covered by mucin-secreting, simple columnar epithelium and the ectocervix is lined by nonkeratinizing stratified squamous epithelium. The area where the tough squamous epithelium gives away to the fragile single layer of columnar cells is called the transformation zone. Throughout the reproductive life the transformation zone lies at variable distances below the external cervical os. In a healthy woman of childbearing age, the cervix is well protected from abrasions during intercourse by the tougher squamous epithelium which covers the outside of the cervix. However, during a condition called cervical ectopy, the external cervical os is instead lined by a single layer of columnar cells. This condition mostly occurs in adolescence, during pregnancy or postpartum and is associated with susceptibility to infections that can more easily penetrate this type of epithelia. When intact, the epithelial layers and the covering mucus of the vagina and cervix constitute natural barriers to HIV-1 infection. However, if the vaginal mucosa is disrupted due to high-risk sexual behaviours, sexually transmitted diseases or cervical ectopy, micro ulcerations and breaks in the epithelium may occur. It is important to mention that micro ulcerations may occur during ‘normal’ coital acts as well. When studying the female genital tract mucosa several factors of relevance for HIV-1 transmission have to be taken into consideration. The status of the epithelia and mucosal secretions is influenced by differences in hormonal levels due to stage of menstrual cycle, pregnancy, age or use of oral contraceptives or intrauterine devices. Also, during ovulation and menstruation the cervical mucus plug which functions as a physical barrier to invading pathogens is absent. Studies have shown that being seropositive for herpes simplex type 2 (HSV-2) may increase the risk of HIV-1 acquisition amongst high-risk HIV-1-negative people exposed to HIV-1. Likewise, the infectiousness of individuals co-infected with HIV-1 and HSV-2 may increase during periods of HSV-2 reactivation. In fact, it has been stated that prevalent HSV-2 infection is associated with a threefold increased risk of HIV-1 acquisition amongst the general population, thus, in areas of high HSV-2 prevalence a high proportion of HIV-1 may be attributable to HSV-2 .
Early cellular targets of HIV-1 in the female genital tract
Primary target cells for HIV-1 infection
It is not yet understood where in the female genital tract HIV-1 most often traverse the epithelium upon primary infection. However, the vaginal and ectocervical stratified squamous epithelia are the dominant surfaces exposed to HIV-1 upon sexual intercourse. The cervix, and specifically the transition zone between endocervix and ectocervix, is a likely location for both viral transmission and induction of cell-mediated immunity to HIV-1 . Although further studies are needed, several routes for HIV-1 transmission have been suggested based on in vitro assays, ex vivo explant models and animal models (Fig. 1). These suggested routes of HIV-1 transmission are thoroughly addressed in references [4–6]. Physical breaches obviously allow access for the virus to the submucosal compartment but moreover, the intact thick epithelium of vagina and ectocervix is interspersed with CD4 and CCR5 expressing, immature Langerhans cells . If HIV-1-infected, these Langerhans cells may subsequently transmit virus to proliferating CD4+ memory T cells allowing further dissemination to regional lymphnodes . The virus can also bind to Langerin, a C-type lectin receptor, which is expressed on intraepithelial Langerhans cells [9–11]. Furthermore, HIV-1 can interact concomitantly with nonLangerhan dendritic cells in the submucosa by a CD4/CCR5 dependent pathway or by C-type lectin receptors (including DC-SIGN), adhesion molecules, complement receptors, Fc-receptors and heparin sulphate proteoglycans [9, 12–14]. The CD4/CCR5-dependent pathway of virus entry probably plays a critical role which is consistent with the observations that: (i) individuals with the homozygous CCR5delta32 deletion are protected from sexually acquiring HIV-1 ; (ii) preferentially CCR5-using virus isolates account for the sexually transmitted viral phenotypes [7, 16]; (iii) the CCR5 blocking molecule PSC-RANTES could prevent vaginal SHIV transmission in rhesus macaques . Langerhans cells are not present in the single cell epithelium of endocervix and the upper female genital tract and thus this is not a mechanism for viral entrance at these sites. HIV-1 entry by transcytosis through single columnar epithelial cell layer is a mechanism of virus entry in the intestinal and rectal mucosa but has not yet been proved in the female genital tract [18, 19].
Dissemination of the HIV-1 infection
Once viral particles or virus-infected cells have crossed the mucosal barrier via binding to Langerhans cells or via alternative mechanisms, replication is initiated in target cells that include CD4+ T cells and to a limited extent in macrophages in the lamina propria (reviewed in ). Using a monkey model, Miller et al.  demonstrated small founder populations of infected cells at the point of entry following intravaginal inoculation with simian immunodeficiency virus (SIV). They also showed the requirement for local proliferation and for continued seeding of distal lymphoid tissues to establish systemic infection. Although the monkey model of virus transmission in the genital tract is relevant as originally proposed  we need to clarify the mechanisms that subsequently occur during HIV-1 exposure to the human female genital tract. Further studies are also needed to clarify the relative importance of: (i) infection with free virus or virus-infected cells; (ii) infection across the single cell epithelia or multistratified epithelia; (iii) virus capture by Langerhans cells or other primary target cells; (iv) virus infections by CD4+CCR5+ dependent pathways or by C-type lectin or other receptors and (v) virus transport to regional lymphnodes by Langerhans cells, nonLangerhan dendritic cells or CD4+ T cells. It is likely that different mechanisms are involved depending on the status of the mucosa (i.e. due to hormonal influences and/or inflammatory processes) and the concentration of the virus inoculum (i.e. viral load in semen).
Anatomy and early cellular targets for HIV-1 transmission in the female genital tract in relation to HIV-1 prevention strategies
Although the probability of transmission for each sexual encounter with HIV-1 is believed to be quite small (approximately 1 : 200–1 : 2000, male-to-female) [23, 24], new data convincingly indicate that these rates vary greatly depending on the phase of the infection. In a study on rates of transmission in an Ugandan HIV-1-discordant couple's cohort (i.e. one partner is HIV-1-infected and the other not), the HIV-1-positive partner's stage of infection was highly associated with the rate of transmission to the HIV-1-negative partner . Wawer et al. conclude that the transmission of HIV-1 is the highest during the early-stage of infection and that the cut off for this direct link between viral load and transmission seems to be at 1500 HIV-1 RNA copies mL−1 of blood.
Obviously male and female condoms that hinder virus or virus infected cells to reach their target cells in the female genital tract remain the most reliable method for HIV-1 prevention. On the other hand, if the transmission occurs solely through the cervical mucosa or other parts of the upper female genital tract, diaphragms (devices that cover the cervix and prevent access to the upper genital tract) would be sufficient to stop the virus from reaching their target cells in the case of an intact epithelial layer in the lower genital tract. Indeed, large trials are ongoing assessing whether latex diaphragms used during intercourse can protect women from contracting HIV-1 . The rationale for conducting these studies prior to knowing the specific biologic basis is the urge to have female-initiated methods in light of the fact that young, married women are at high risk of contracting the infection in certain geographical regions and often have difficulty negotiating condom use. Nevertheless, for future design of preventive measures in the absence of a protective vaccine, the biology of virus transmission and protective immunity at the local level is urgent to uncover. The use of a physical barrier like a diaphragm could indeed be a complement to a microbicide.
HIV-1-exposed uninfected individuals
A minority of individuals has been repeatedly exposed to HIV-1 for years without the occurrence of productive infection and such individuals are considered to be resistant to infection (as defined by HIV-1 IgG and HIV-1 RNA/DNA negativity in blood). These so called ‘exposed uninfected individuals’ have been identified in a variety of cohorts including: discordant couples, commercial sex workers (CSW), exposed healthcare workers and infants of HIV-1 infected mothers exposed in utero (reviewed in ). Multiple reports investigating the cause of this protection have been published and several protective mechanisms have been suggested including adaptive cellular and humoral immune responses, innate immune responses and genetic variations. It must be noted though that the ‘resistant’ status in many studies probably reflects the poor efficiency of the virus to infect the host upon sexual exposure rather than absolute resistance to infection. Measuring environmental exposures can be difficult and imprecise and the determinants for host susceptibility are multifactorial in mechanisms and biology, making it hard to determine the relative importance of each factor. Perhaps the strongest evidence of resistance comes from cohorts of sex-workers whom are repeatedly highly exposed to different HIV-1 subtypes during several years from multiple partners, and yet the virus lacks to productively infect the female. Knowledge of the possible immune mechanisms of such resistance would allow the targeted development of vaccines, other biological methods of preventing HIV-1 transmission as well as therapeutics.
Can adaptive immunity occur without invasive infection?
Surprisingly, some exposed uninfected individuals (preferentially those that are highly exposed) demonstrate HIV-1-specific CD4+ and CD8+ T-cell responses as well as HIV-1-neutralizing IgA antibodies both systemically and in the female genital tract mucosa (reviewed in ). Several explanations have been presented for the induction of adaptive immune responses, including; (i) a continuous virus replication at an extraordinarily low level in the submucosal tissue at the site of viral exposure (but without further dissemination of the infection to regional lymphnodes); (ii) single rounds of HIV-1 replication at one or a few occasions at local sites (genital submucosa) followed by viral clearance; (iii) immune processing of viral fragments presented to the immune system, without productive infection, in a way that allow stimulation of an adaptive immune response. Although it is not possible to provide a definitive scientific demonstration of resistance in humans (for obvious reasons), cohort studies of natural resistance in high-risk populations together with studies in animal models has thought us a great deal about the mechanisms of protection. In fact, in vivo experiments, using macaques to mimic the natural phenomena which is believed to take place in these cohorts of exposed uninfected individuals, show that when animals were exposed to subinfectious doses of SIV, the ‘exposed uninfected’ animals were protected from subsequent challenge with infectious virus inocula .
Humoral immune responses in the female genital tract of exposed uninfected individuals
The female genital mucosal immune system is responsible for two life essential tasks; to protect the reproductive tract from pathogenic mechanisms yet to be tolerant of sperm and embryo antigens so that reproduction may occur. Considering these tasks, a coordinated expenditure of immune molecules, cells and antibodies to combat invading pathogens are required and these will also challenge any HIV-1 exposure in the genital tract mucosa.
Antibody responses in HIV-1-infected individuals
The protective role of virus-neutralizing or other virus-inhibiting antibodies (including those that bind complement or mediate antibody-dependent cellular cytotoxicity) is most likely critical when the different responses are present at the mucosal site at time of viral exposure. Although the role of systemic and mucosal antibody responses during established HIV-1 infection is controversial, higher levels of systemic HIV-1-envelope specific neutralizing IgG antibodies have been described in subjects with long-term nonprogressive infection [30–32] and mothers who do not transmit HIV-1 to their children [33–35] as compared with HIV-1 infected individuals with progressive infection.
Measurement of mucosal HIV-1-specific IgA antibodies
Measurement of systemic and mucosal HIV-1-specific IgG antibodies are standardized and well validated with both commercial and in-house assays . There are however no optimal methods available for quantification or measurement of the epitope-specificity of HIV-1-specific IgA antibodies in mucosal fluids. The concentration of HIV-1-specific IgA antibodies in cervicovaginal lavage fluids of HIV-1 infected patients is low as measured by antigen-binding assays (enzyme-immuno assays or Western blotting assays) . Due to these difficulties it is helpful, and probably more biologically relevant, to measure the functional activity (such as antibody-mediated inhibition of HIV-1 infection) of IgA purified cervicovaginal fluids. HIV-1-specific IgA antibodies to conformational epitopes can also be missed in enzyme-immuno assays or Western blotting-based methods whereas they may show functionality in other type of assays. IgA-mediated virus neutralization has thus readily been demonstrated in cervicovaginal fluid samples of HIV-1-infected women. The epitope-specificity of the antibodies may also be measured in the functional assays by blocking the activity with preincubation of the antibody sample and relevant peptides or by peptide-purification of the antibody sample . The functional assay does not obviously discriminate the innate from the adaptive IgA response and can thus measure many specificities including allo-reactivity to cellular antigens.
Mucosal HIV-1 neutralizing IgA in exposed uninfected individuals
By definition, exposed uninfected individuals do not have HIV-1-specific IgG antibodies neither in blood nor in mucosal fluids. The humoral antiviral activity mediated by these individuals is thus mediated by non-HIV-1-specific IgG antibodies (such as allo-reactive antibodies) or by HIV-1-specific or non-HIV-1-specific IgA antibodies. Although the presence of HIV-1-specific IgA antibodies in mucosal fluids of exposed uninfected individuals is controversial  this presence is probably dependent on the frequency of HIV-1-exposure (number of HIV-1-infected sexual partners, type of sexual activity and viral load in semen) . The methodological difficulties in measuring the antibodies has been outlined above and thus measurement of functionality is more relevant. Our group and others have previously detected two types of functional/biologic activity: HIV-1-neutralizing IgA antibodies in a peripheral blood mononuclear cell-based assay as well as IgA mediated inhibition of HIV-1 transcytosis across epithelial cells in exposed uninfected individuals [41, 42]. Regarding specificity, the neutralizing IgA derived from exposed uninfected commercial sex worker cohorts have a reportedly broad cross-clade activity, i.e. able to neutralize several HIV-1 clades. In contrast, IgA derived from exposed uninfected discordant partner cohorts seem to lack this characteristic . This may be a reflection of the broad spectrum of viruses encountered by the commercial sex workers whilst the discordant partners are only exposed to the HIV-1-positive partner's virus, further implying the role of virus exposure as a ground for the corresponding immune response and that the response is indeed partly HIV-1-specific.
The IgA purification method we and others have used allows IgA1 antibodies to be specifically purified by the use of jacalin-beads. The IgA2 subclass has generally been equivalent to anti-carbohydrate antibodies targeting bacterial infections and these carbohydrate–protein interactions have been considered to be weaker than protein-protein interactions [44–46]. Nevertheless, viruses (such as the HIV-1 virus) shield their antigenic protein epitopes with glycan structures . These glycan structures could theoretically become immunogenic and thus accessible to the pre-existing anti-carbohydrate antibodies. Such pre-existing antibodies reacting with HIV-1 may also be present due to previous encounter with the virus in the mucosa. In fact, one of the few broadly HIV-1 neutralizing monoclonal antibodies, 2G12, that can protect against viral challenge in animal models  and delay viral rebound in HIV-1 infected patients  targets the carbohydrate-masked ‘silent’ face of gp120. Our own preliminary data indicate that both mucosal IgA1 and IgA2 antibodies from exposed uninfected individuals indeed have HIV-1 neutralizing capacity (Hirbod T and Broliden K, unpublished).
The presence of HIV-1 inhibiting IgA, but not IgG, antibodies in exposed uninfected individuals is puzzling. We lack basic immunologic knowledge how the virus can yield an IgA response and thereby by-pass the production of HIV-1-specific IgG antibodies. Although the mucosal tissue contains antibody-producing cells that are susceptible to TGF-β (a well-known factor for IgA switching) this does not explain the absence of HIV-1-specific IgG. It seems clear though that the number of viral exposure is critical for the induction of an IgA-mediated HIV-1-neutralizing response as stated above .
Are HIV-1 neutralizing IgA antibodies associated with HIV-1 protection?
Although HIV-1-neutralizing IgA responses have been described in the blood and genital tract of exposed uninfected individuals there is a lack of well controlled prospective studies performed. Such studies could potentially reveal the significance of IgA responses with regards to protection against HIV-1 acquisition. We have performed such a study in a cohort of HIV-1-uninfected Kenyan commercial sex workers participating in an HIV-1 prevention trial from 1998 to 2002 . After trial completion, the commercial sex workers who acquired HIV-1 were matched with persistently uninfected controls based on study arm, sexual risk taking and time of cohort enrolment. Our preliminary findings indicate that the presence of HIV-1-neutralizing genital tract IgA is indeed associated with subsequent protection against sexual HIV-1 acquisition (Hirbod T, unpublished data). Within the original HIV-1 prevention trial  there was an association between HSV-2 infection and HIV-1 acquisition identified.
In a study on an HIV-1-discordant couple cohort, Clerici et al.  detected that the epitope specificity of IgA was different to that of HIV-1-infected individuals, indicating that this region may be important for immune protection. The IgA antibodies recognized a restricted epitope on the gp41 protein (QARILAV) situated within the coiled coil pocket that is involved in virus interaction with the target cell. It remains to be seen whether IgA reactivity against this epitope is specific for exposed uninfected individuals representing other risk groups, such as commercial sex workers. If so, it would be an interesting region to use as a vaccine target.
Are samples from exposed uninfected individuals able to neutralize virus beyond the glycoprotein rich mucous of the female genital tract and thus prevent further dissemination of the infection?
As discussed, if HIV-1 passes through the female genital tract mucosa the virus will most likely be rapidly associated to intraepithelial Langerhans cells, submucosal dendritic cells and CD4+ memory T cells. As a consequence the probability of a manifest infection may increase. Studies using co-cultures of HIV-1-exposed dendritic cells and T cells in vitro (representing viral transmission from dendritic cells to T cells in the female genital mucosa) have resulted in a massive viral replication (reviewed in ). Therefore, we sought to explore if samples from exposed uninfected individuals would be able to inhibit this virus propagation. The neutralizing ability of the same study subjects were also tested in traditional peripheral blood mononuclear cell-neutralization assays, rather mimicking the events at the luminal side of the genital tract (i.e. before dendritic cell-HIV-1 interaction) (Fig. 2). Consistent with previous data published by our group and others, the majority of cervicovaginal lavage and plasma samples from exposed uninfected women were able to neutralize virus in the peripheral blood mononuclear cell-based neutralization assay. That is, samples from exposed uninfected individuals have the ability to neutralize HIV-1 and hinder the infection of target cells. However, if the virus is allowed to bind (either in cis [infection of dendritic cells]; or trans [binding without infection of dendritic cells] reviewed in ) to dendritic cells, neither systemic nor mucosal samples from exposed uninfected individuals seem to have the ability to prevent further dissemination of the virus. On the contrary, HIV-1 infected individuals indeed comprise the ability to neutralize virus that has already infected dendritic cells . Based on these findings, we hypothesize that the noted neutralizing ability is due to the presence of HIV-1 specific IgG in HIV-1-positive individuals (the presence of HIV-1 binding IgG antibodies was confirmed in all samples from HIV-1 infected patients) and that the HIV-1 neutralizing IgA activity of exposed uninfected individuals rather limits the spread of HIV-1 prior to dendritic cell–T cell HIV-1 transfer.
Innate host defence of human cervico-vaginal mucosa
The innate immune responses are defined as rapid and inherent and are activated by conserved structures on invading pathogens. The responses utilize a wide spectrum of cells (such as macrophages and NK cells) and molecules to confer protection against potential invading pathogens and assists in regulating subsequent adaptive immune responses. This host defence is mediated by both physical and chemical defence barriers (such as the cervical mucus plug) as well as antimicrobial peptides and proteins (such as calprotectin, SLPI, lysozyme, lactoferrin, defensins, cathelicidin and histones) which are released into the overlying cervicovaginal mucosal fluid. Alterations during the menstrual cycle have been observed in cervical mucus for several innate molecules such as LIF , SLPI  and β-defensins , and the use of oral contraceptives has been observed to decrease the expression of β-defensins  in the cervix mucosa. The expression of innate molecules is obviously also altered due to sexually transmitted diseases (STDs) causing inflammation and lesions in the epithelium that normally protects the submucosal compartment. Although knowledge of endogenous factors that may mediate resistance to HIV-1 transmission and early dissemination in the female genital tract are still largely unclear, several studies indeed describe such protective responses. In the following section factors of relevance for the defence against HIV-1 and specifically those that have been studied in exposed uninfected individuals will be described.
Chemokines are secreted by various types of immune system cells and have specific receptors of relevance for their antiviral activity, including HIV-1 (i.e. CCR5). The antiviral activity of the corresponding chemokines (Regulated upon Activation Normal T-cell Expressed and Secreted (RANTES), MIP-1α, MIP-1β) is exerted by competition with HIV-1 for receptor binding capacity and the attraction of immune cells with antiviral activity. Elevated expression levels of RANTES, MIP-1α and MIP-1β due to genetic polymorphism have indeed been associated with resistance to infection [57, 58]. The antiviral capacity of RANTES has been further supported by Iqbal et al.  who observed significantly higher levels of the chemokine in cervical washings of exposed uninfected commercial sex workers. Furthermore, protection against Simian HIV chimaera (SHIV) infection mediated by a vaginally applied RANTES analogue has been suggested in a monkey model .
Secretory leukocyte protease inhibitor (SLPI) is a serine protease inhibitor secreted by nonciliated epithelial cells lining mucosal surfaces . When McNeely et al. evaluated a panel of proteins expressed in mucosal tissues, for protection of macrophages against HIV-1 the researchers concluded that even though several of the investigated proteins displayed anti-HIV-1 activity, SLPI was the only factor that significantly blocked infection at physiological concentrations (1–10 μg mL−1) . Although a putative SLPI receptor has not yet been described, studies indicate that the mechanism by which SLPI inhibits HIV infection appears to involve interactions with host cell target rather than direct binding to the virus . Furthermore, studies have implicated that this activity occurs at an early stage of cellular infection, prior to viral reverse transcription . At the 2006 XVI International AIDS conference, Iqbal et al.  reported on the elevation of a 6 kDa biomarker in the genital mucosa of exposed uninfected individuals when compared with HIV-1-uninfected and HIV-1-infected controls. The biomarker was subsequently recognized as Trappin-2, which is a SLPI-like molecule, and using in vitro assays identified as a potent inhibitor of HIV-1 infection.
Endogenous antimicrobial peptides
One group of molecules that have been studied extensively in their role as anti-viral factors are the defensins, secreted by epithelial cells. The defensin group is divided into α- and β-defensins, and it is β-defensins that are particularly produced by epithelial cells . The α-defensins are suggested to function more systematically, allowing immune cells to access vascularized tissues, whereas β-defensins primary have a protective role of defending the mucosal surfaces from microbes . A strong, constitutive production of α-defensin has been observed in exposed uninfected individuals suggesting a role in the protective immune response that characterizes these individuals . The human β-defensin 1 is suggested to be constitutively produced by the epithelial cells of the female genital tract, whilst β-defensin 2 seems to be induced by infection . Furthermore, β-defensin 2 and 3 have been shown to effectively block HIV-1 infection by modulating the CXCR4 co-receptor and interacting directly with virions . In a recent study, Feng et al.  showed that β-defensin 3 actually competes with SDF-1 (the α-chemokine natural ligand for CXCR-4) for cellular binding to HIV-1 target cells. The human catheliodin LL-37 was also recently shown to inhibit HIV-1 in vitro .
The fact that Interferon alpha (IFNα) is expressed as one of the earliest responses to infection has attributed the cytokine a dominant role in directing the forthcoming innate and adaptive immune responses to pathogens . Not only does IFNα play a major role in directing and linking the innate and adaptive immune responses against an invading pathogen, but the cytokine also plays a major role in the host viral defence by directly inhibiting the intracellular-viral lifecycle of a wide variety of viruses. In an extensive review, Stark et al.  describe how IFNα can block entry/uncoating, viral RNA transcription, viral protein translation, and many other mechanisms for several viruses. In 2001, Soumelis et al.  demonstrated a negative correlation between the number of circulating pDCs (the main producer of IFNα) and HIV-1 viral load in the blood of HIV-1-positive patients. It has since been suggested that IFNα may be used as an antiviral and immunotherapeutic drug against HIV-1 by boosting the host's immune system in response to HIV-1 infection. In addition, another Interferon molecule, IFNγ, has been associated to natural resistance to HIV-1 infection .
Intracellular host factors with anti HIV-1 activity
Host resistance factors that are able to interfere with the early steps of HIV-1 infection include TRIM-5alpha and APOBEC3G [75–78]. At the 2006 XVI International AIDS conference, Piacentini et al.  presented data on upregulation of APOBEC 3G and 3F in cervical biopsies of exposed uninfected individuals as compared with HIV-1 infected and low-risk uninfected individuals. In addition, the detection of Murr1, a genetic restriction factor that seem to inhibit HIV-1 replication in resting CD4+ T cells, is believed to contribute to the regulation of asymptomatic HIV-1 infection and the progression of AIDS  but has not been examined in exposed uninfected individuals to our knowledge.
What happens to the female genital innate immune milieu when women engage in high HIV-1 encountering risk behaviours?
Although studies of exposed uninfected individuals have pointed out a number of innate and adaptive immune responses that may confer the resistant status, little has been done to describe the expression of innate immune molecules in the cervical tissue of women at risk of infection. In fact, when we reviewed the literature on genital tract mucosa and innate immunity, we could not identify any articles that sought to investigate the variation in expression of innate molecules which have been associated with anti-HIV-1 activity (such as RANTES, SLPI or IFNα) at the single cell level. Therefore we set up a study seeking to describe and quantify the innate molecule expression in the female genital tract. Three distinct risk groups were recruited: HIV-1-positive women, low-risk healthy HIV-1-negative women and commercial sex-workers. None of the recruited commercial sex workers were HIV-1-positive, despite the fact that they had engaged in sex-work from 3–16 years. Interestingly, a higher expression of IFNα and RANTES was detected in exposed uninfected female sex workers and HIV-1 infected individuals as compared with low-risk HIV-1 uninfected controls . We hypothesize that a constitutive induction of IFNα and RANTES expression in cervical mucosa may contribute to protection of sexual HIV-1 transmission in subjects with a higher risk behaviour. It must however be pointed out that these molecules may be protective by their immunomodulatory and HIV-1 receptor (CCR5) blocking capacity, respectively, if they are present prior viral inoculum. On the other hand if they are not ‘in the right place at the right time’ they may have an opposite effect by promoting inflammation including up-regulation of CCR5 and thereby recruiting more target cells for HIV-1 replication and causing a mucosal environment that favours viral replication and viral dissemination. Indeed, a massive beta-chemokine expression is present already prior to peak viremia in individuals undergoing acute HIV-1 infection . Recent studies also implicate a role for IFNα in the continued loss of CD4+ T cells that is a hallmark of progressive HIV-1 infection . Thus, the effect of different molecules in a systemic immune response during a disseminated infection must be interpreted in another context than the effects of the same molecules at the mucosal levels acting prior to infection.
Although many innate factors have anti-viral activity the relative effect they have on HIV-1 transmission remains to be established. There are probably also several hitherto undiscovered innate immune factors including CAF [84, 85] that have an important role in the natural resistance against HIV-1 infection at the level of the female genital tract mucosa as the transmission rate of HIV-1 following sexual intercourse indeed is relatively low [23, 24].
Cellular immunity in the female genital tract of exposed uninfected individuals
Although HIV-1-specific cytotoxic T lymphocytes are important for controlling the viremia in HIV-1-infected individuals they are unable to prevent eventual immunosuppression (AIDS) and death . However, HIV-1-specific cytotoxic T lymphocytes have also been described in some (but not all) exposed uninfected populations, both derived from blood [87–90] and the female genital mucosa [87, 91]. In a report from 2001, Kaul et al.  were able to detect cytotoxic T-lymphocyte responses directed against predefined HIV-1 epitopes in the blood and genital tract of more than 50% of investigated exposed uninfected women in a sex-worker cohort. Furthermore, the investigators observed that the specificity of these epitopes were different from the cytotoxic T-lymphocyte epitopes detected in HIV-1-positive women. In fact, the occurrence of ‘late’ seroconversion (i.e. seroconversion due to a break from sex-work and thus lacking to maintain the cytotoxic T-lymphocyte responses), which has been reported by the same investigators , is suggested to be associated with a loss and a switch in the epitope specificity of these cytotoxic T-lymphocyte responses . These data remain to be confirmed in larger prospective studies. It would also be interesting to study the qualitative differences in the cytotoxic T-lymphocyte functional response including perforin/granzyme expression as well as antigen-specific polyfunctional cytokine responses in protected versus those individuals that failed to contain sterilizing immunity .
Genetic resistance to HIV-1 infection in exposed uninfected individuals
In 1996, it was described how a mutant allele of the HIV-1 co-receptor CCR5 was present in individuals resistant to infection by HIV-1 [15, 96, 97]. A 32-base-pair deletion within the coding region of CCR5 resulted in the generation of a nonfunctional receptor that did not allow the virus to fuse to the cell. The finding of this co-receptor mutation is one of the most solid evidence of resistance in exposed yet uninfected individuals. However, it is important to note that this co-receptor mutation is absent in sub-Saharan African populations, thus, it cannot be denoted as the reason for resistance found in these populations. Other genetic factors have also been linked to viral resistance including other HIV-1 co-receptor mutations and certain HLA types (reviewed in ). The role of host genetics in resistance to HIV-1 infection provides important insights into the HIV-1 epidemic but it must be noted that the studies cannot be translated for clinical use to determine HIV-1 susceptibility on an individual basis. Determinants of resistance are most likely dependent on a complex association of multiple genes affecting the immune response and viral susceptibility of host cells .
Development of microbicides and mucosal vaccines for protection against sexual transmission of HIV-1
A long neglected goal of HIV-1/AIDS research is the development of products designed to protect women against sexually transmitted diseases, including HIV-1, which would provide an alternative to male condoms. One such general method, as mentioned above, is the use of microbicides (vaginal or rectal application of a compound designed to prevent the sexual transmission of HIV-1) (reviewed in ). The female genital tract mucosa thus contains several cell types expressing different viral receptors for HIV-1 entry. An effective microbicide could work by preventing the virus from reaching its target cells by (i) targeting virus binding, fusion and replication; (ii) targeting viral receptors on the target cells and/or (iii) serving as a physical or chemical barrier in the mucosa. However, given the many routes of viral transmission from the mucosal surface to the submucosa (Fig. 1), a truly protective candidate may also have to block virus–cell interactions to prevent an establishment of local infection and further dissemination of the virus. Phase III clinical trials are indeed ongoing to test the efficacy of microbicide candidates and many new products have been developed for future trials.
Mucosal vaccines have been successfully developed against a few pathogens, including; oral vaccines against poliovirus, Salmonella typhi, V. cholerae and rotavirus and a nasal vaccine against influenza virus. Although the mucosal-associated lymphoid system (MALT) is well-defined in the gastro-intestinal, respiratory and nasal tracts, a corresponding genital associated lymphoid system has not been clearly defined. However, studies in nonhuman primates aiming to trace the migration of T- and B-cells have revealed that the internal and external iliac lymph nodes may function as inductive sites from which cells migrate to effector sites in the lamina propria of the cervico-vaginal mucosa (and rectum) . Studies in both humans and monkeys have shown that whether transmitted mucosally or parenterally, HIV-1 and SIV replicate preferentially in mucosal tissues, such as the intestinal mucosa, that are rich in CD4+ CCR5+ memory T cells [101–103]. Therefore, the ultimate goal of an anti-HIV-1 vaccine should be to first interrupt mucosal transmission at its earliest stages, before the virus has crossed the epithelial barrier and infected its first target cell, in order to prevent the establishment of viral reservoirs in mucosal tissues. It is also becoming clear that a successful vaccine must induce both the cellular and humoral immune response. As we have tried to point out, mucosal antibodies and particularly secretory IgA play an important role in protection against HIV-1 infection (reviewed in ) and stimulation of the interplay between innate and adaptive immunity seems to be crucial to elicit a mucosal and systemic response. A few mucosal vaccination studies in mouse models which provide proof-of-principle that mucosal immunization can induce IgA responses with neutralizing activity have been performed to date. [105–107]. These studies show that immunological induction at one mucosal site often results in immune responses at distal mucosal sites, probably due to specific ‘homing’ receptors on mucosal lymphoid cells. However, immunizations aimed to stimulate a mucosal HIV-1-specific protective immune response in the female genital tract in humans still need to be optimized and may also include immunization strategies with the aim of enhancing the primary allogeneic and CCR5-directed responses which are involved in the natural protection against HIV-1 infection .
We have here described how the mucosal transmission is the initial step towards systemic HIV-1 infection, and thus inhibition of viral mucosal transmission seems to be the most efficient method to prevent infection. The virus concentrations are relatively low upon entry and the viral particles are still susceptible to neutralization by antibodies and other molecules of the immune system. However, as soon as the virus is transmitted to T cells, virus replication and dissemination will greatly accelerate the infection and severely reduce the chance of any efficient clearance of the virus. Several studies have shown the contribution of the mucosal innate and humoral immune factors in HIV-1-exposed uninfected individuals to their resistant status. By taking these observations a step further, new interesting approaches using active mucosal immunization to produce a local immune response in the female genital tract have been described in experimental animal models. If we manage to understand the mechanisms of protection and induce these artificially, an AIDS vaccine or other biological methods of preventing HIV-1 transmission like microbicides would become a reality.
However, there still exist rather large gaps in our knowledge about protective mechanisms against HIV-1 acquisition. From our point of view, this is due to a severe lack of prospective studies conducted on mucosal sites in the human body, where the majority of HIV-1 transmissions occur. As Dr Sharon Hillier, a pioneer in the field of women's health and sexual transmission of pathogens, eloquently has said: ‘Infection happens in human environments and not in a test tube’. Although the field of HIV-1 research has indeed shifted during the past years and today a search on the internet for ‘Mucosa + HIV-1’ provides the seeker with a rather great range of publications within this field, the ongoing quest to find an effective vaccine is still hindered by our limited number of studies focusing on basic immunology and microenvironment of mucosal tissues.
Conflict of interest statement
The authors do not have any conflicts of interest to declare
The authors thank Sam Hirbod for the illustrations. Financial support for our own unpublished findings was received from the Swedish Research Council and SIDA/SAREC.