Vaginal microbiota and its role in HIV transmission and infection

Authors


Correspondence: Sarah Lebeer, KU Leuven, Centre of Microbial and Plant Genetics, Kasteelpark Arenberg 20, Box 2460, B-3001 Leuven, Belgium. Tel.: +32 16 321631; fax: +32 16 321966; e-mails: sarah.lebeer@biw.kuleuven.be; sarah.lebeer@ua.ac.be

Abstract

The urogenital tract appears to be the only niche of the human body that shows clear differences in microbiota between men and women. The female reproductive tract has special features in terms of immunological organization, an epithelial barrier, microbiota, and influence by sex hormones such as estrogen. While the upper genital tract is regarded as free of microorganisms, the vagina is colonized by bacteria dominated by Lactobacillus species, although their numbers vary considerably during life. Bacterial vaginosis is a common pathology characterized by dysbiosis, which increases the susceptibility for HIV infection and transmission. On the other hand, HIV infections are often characterized by a disturbed vaginal microbiota. The endogenous vaginal microbiota may protect against HIV by direct production of antiviral compounds, through blocking of adhesion and transmission by ligands such as lectins, and/or by stimulation of immune responses. The potential role of probiotics in the prevention of HIV infections and associated symptoms, by introducing them to the vaginal and gastrointestinal tract (GIT), is also discussed. Of note, the GIT is a site of considerable HIV replication and CD4+ T-cell destruction, resulting in both local and systemic inflammation. Finally, genetically engineered lactobacilli show promise as new microbicidal agents against HIV.

Introduction

The surfaces of the human body can be divided into four fundamental microbial niches, which differ in composition, but each play an important role in the entire life of an individual. These niches include the skin, the oronasopharyngeal cavity, the genital tract, and the gastrointestinal tract (GIT). The microbiota present in each of the niches provides a vast number of health effects to the host (Reid et al., 2011). For each of these niches, the symbiotic microbiota protects the host from pathogenic colonization by niche-specific metabolic exclusion of pathogens, by competition for adhesion sites and by stimulation of immune responses that activate production of antimicrobial components (Duerkop et al., 2009). The impact of the mucosal microbiota on the mucosal immune system is even more broad than the mere stimulation of antipathogenic responses, because more and more studies, such as with germfree mice, indicate a crucial role for immune regulation by symbiotic bacteria (Hooper et al., 2012). In addition, the microbiota has specific functions at each mucosal niche, of which the role of the GIT microbiota in digesting food components and energy conversions are probably the best known.

During the last decades, the genital and GIT microbiota have been increasingly recognized as important factors in the host defense against viral pathogens such as the human immunodeficiency virus (HIV). HIV is a member of the Lentivirus family based on its gene sequence homology, morphology, and life cycle. Two types of HIV are known, designated as HIV-1 and HIV-2, which are transmitted in the same way and manifest similar clinical syndromes. However, they differ in genetic structure, antigenicity, and pathogenicity: that is, HIV-2 is less pathogenic compared with HIV-1. Furthermore, HIV-2 transmits less efficiently and is almost exclusively found in West Africa (Reeves & Doms, 2002). This review aims to describe the role of the human microbiota and exogenously applied probiotics in relation to HIV infection, with emphasis on the molecular mechanisms involved. In addition to the current state of the literature, future applications, such as recombinant probiotics as potential new microbicides, are discussed.

Key immunological and microbiological aspects of the female reproductive tract

Defense system of the urogenital tract

The mucosal surface of the urogenital tract forms the first line of defense against microorganisms and viruses in this niche and separates the external environment from the internal sterile environment. It is generally believed that the upper genital tract is free of microorganisms (Heinonen et al., 1985). The mucosal epithelium in this upper niche should thus be extremely efficient in recognizing and subsequently responding to microorganisms, while at the same time avoiding chronic inflammation. In contrast to the upper genital tract, the cervical and vaginal epithelia are permanently colonized by microorganisms, mainly endogenous microbiota represented by lactobacilli in a healthy stage and a broad variety of potential pathogens during infections (see below). Therefore, this mucosa has adapted to a nonsterile, dynamic environment, continuously challenged by changes in hormone levels and a variety of inflammatory stimuli associated with sexual intercourse. For example, in the genital tract of premenopausal women, hormones control the expression of large numbers of genes responsible for the secretion of cytokines and chemokines, regulating cellular composition, immunoglobulin (Ig) secretion, and antigen presentation (Wira & Rossoll, 2003). Such experimental studies clearly show that the vaginal and cervical epithelium and immune cells provide a natural barrier against pathogens.

The upper female reproductive tract (FRT), which consists of the endometrium, endocervix, and Fallopian tubes, is characterized by the presence of Type I mucosae, a simple columnar epithelium formed by a single layer of ciliated columnar cells connected via tight junctions (Fig. 1). The ciliated columnar cells have between 200 and 300 finger-like hairs, called cilia, involved in movement of mucus as well as egg cells toward the Fallopian tubes and the uterus. Type I mucosae are characterized by the presence of polymeric Ig receptors, microfold cells (M-cells), presence of mucosa-associated lymphoid tissue (MALT), and secretory IgA antibodies (Iwasaki, 2010). A Type I mucosa is also present in the GIT. However, the GIT mucosa also shows significant differences with the upper FRT, such as the presence of Paneth cells and gut-associated lymphoid tissue (GALT), which includes Peyer's patches in the small intestine and isolated lymphoid follicles in the colon.

Figure 1.

General overview of the Type I and Type II mucosae present in the GIT and FRT. GIT and upper FRT (uterus and endocervix) are characterized by the presence of Type I mucosae. Both GIT and upper FRT are covered by a single layer of columnar epithelium in which the cells are connected by tight junctions. This epithelium is covered with a thick layer of mucus. IgA is the main antibody isotype detected in Type I mucosae. A typical characteristic of GIT is the presence of Peyer's patches, which represent mucosa-associated lymphoid nodes. In comparison, a unique structure of the upper FRT is the presence of LA. A striking difference between the GIT mucosae and upper FRT mucosae is the presence of an endogenous microbiota in GIT, while the upper FRT is believed to be sterile. The lower FRT is represented by Type II mucosae, characterized by a stratified squamous (nonkeratinized) epithelium in which low numbers of subepithelial immune cells are detected. The major antibody isotypes are IgG and a low concentration of IgA. A unique set of epithelial DCs, called LCs, is also present within this Type II mucosa. In addition, the figure shows the ectocervix–endocervix transformation zone, where stratified squamous epithelium ‘transforms’ to single columnar epithelium. This zone is highly enriched with immune cells, mainly CD4+ T cells. The commensal microbiota for each of these niches is also highlighted. The GIT microbiota includes a diverse number of microorganisms with three main enterotypes, dominated by resp. Bacteroides, Prevotella, and Ruminococcus. On the other hand, the vaginal microbiota is generally Lactobacillus dominated with commonly isolated species including Lactobacillus iners, Lactobacillus crispatus, Lactobacillus gasseri, and Lactobacillus jensenii. The vaginal microbiota is characterized by the production of lactic acid, which lowers the pH in the vaginal epithelium to the range of 4–5; production of H2O2 and bacteriocins, as well as bacteriophage release which affects community dynamics. In addition, some lactobacilli can be isolated from the ectocervix, in comparison with the endocervix and the upper FRT, which are considered to be sterile niches.

The lower FRT, that is, the vaginal canal and the ectocervix, contains Type II mucosae. They are characterized by multiple layers of nonkeratinized stratified squamous epithelium in which the surface cells are flattened, and the deeper cells are columnar and attached to a basal membrane (Robboy & Bentley, 2004). Such a Type II mucosa is lacking MALT, M- cells, and polymeric Ig receptors. Nevertheless, Type II mucosae are characterized by the presence of Langerhans cells (LCs) and high production of IgG antibodies (Iwasaki, 2010). In addition, cells of the squamous epithelium have no tight junctions between each other (Fig. 1). This permits the transport of small molecules between the cells within the epithelial space, including small viruses and toxic compounds from pathogens (Hickey et al., 2011).

The uterine and vaginal epithelial cells are not only a physical barrier against pathogens, but they are also able to actively recognize conserved microbe-associated molecular patterns of microorganisms via the expression of cognate pattern recognition receptors (PRRs). PRRs include Toll-like receptors (TLRs) and NOD-like receptors, which mediate the secretion of cytokines, chemokines, and antimicrobial peptides (Schaefer et al., 2004, 2005).

The epithelial cells from Type I and Type II mucosae in the FRT and in the GIT are also covered by a layer of mucus, consisting of mucins, which are complex high-molecular-mass O-glycoproteins (Andersch-Bjorkman et al., 2007). The mucus sources in Type I mucosae are specialized goblet cells, similar to the ones in the GIT, and mucus glands present in the crypts of the cervix, while the mucus source in Type II mucosae is local mucus-secreting epithelial cells (Andersch-Bjorkman et al., 2007). Different types of mucin genes (MUCs) are expressed in the FRT. For example, the endocervical epithelium expresses MUCs 1, 4, 5AC, 5B, and 6, while two types of cervico-vaginal mucus can be isolated from the FRT, which strongly depend on the stage of the menstrual cycle (Gipson et al., 1997). The estrogenic mucus, present at the proliferative stage and ovulation, is thin and watery with a low viscosity that permits sperm movement. In contrast, progestational mucus, present at high concentrations after ovulation and during the secretory phase, is thick, sticky, and blocks the passage of spermatozoa (Hickey et al., 2011). An important characteristic of the cervico-vaginal mucus is the low pH between 4 and 5 that affect the transmission of pathogenic bacteria and viruses including HIV.

The concentration and distribution of innate immune cells in the FRT depend on the hormone levels and menstrual cycle, and varies in the different parts of the FRT. In a steady state, low numbers of neutrophils, dendritic cells (DCs) or LCs, macrophages, and natural killer (NK) cells are present in the lower FRT. In contrast, the same cells are present in high concentration in the upper FRT, possibly related to the fact that this niche needs to be kept sterile (Pudney et al., 2005). FRT DCs are localized in the subepithelial stroma of the endometrium, while they are present within the epithelial layers in the vaginal epithelium and known as LCs (Fig. 1; Ginhoux et al., 2006). DCs are essential mediators in capturing the HIV virions by the DC-specific intercellular adhesion molecule 3 (ICAM3)-grabbing nonintegrin (DC-SIGN) receptor, which mediates subsequent transmission to T cells (Cameron et al., 1992; Geijtenbeek et al., 2000). Macrophages represent around 10% of the immune cells present in the upper FRT, but because their concentration is estrogen- and progesterone dependent, the upper FRT exhibits relatively high levels prior to menstruation (Givan et al., 1997). An important characteristic of vaginally presented macrophages in comparison with the GIT macrophages is the higher expression of the CD4, CCR5, and CXCR4 receptors. CD4 is known to be the main receptor for HIV virus, while CCR5 and CXCR4 function as coreceptors, which is in agreement with the greatest affinity of HIV virions to vaginal macrophages (Shen et al., 2009). Other important innate immune cells in the FRT are NK cells, which are account for 10–30% of the leukocytes. The numbers of NK cells do not change, except for the ones in the endometrium which can reach up to 70% during the menstruation period (Givan et al., 1997; Wira et al., 2005). The FRT NK cells – similarly to blood NK cells – produce a variety of pro-inflammatory cytokines such as granulocyte–macrophage colony-stimulating factor (GM-CSF), interleukin-8 (IL-8) and interferon (IFN) and subsequently promote the inflammatory response, induce macrophage activation and cytotoxic T-cell generation (Hickey et al., 2011). However, the FRT NK cells are characterized by specific activities such as the production of angiogenic growth factors and leukemia inhibitory factors essential for blood vessel development (Hickey et al., 2011). The importance of NK cells in the female innate defense is also demonstrated by the increased susceptibility of patients with defects in NK cell function for Herpes simplex virus-2 (HSV-2; Bloomfield & Lopez, 1980). In addition, vaginal NK cells – but not blood NK cells – can inhibit infection of target cells by HIV X4 strains but not R5 strains via secretion of CXCL12 (Mselle et al., 2009).

The number of leukocytes in the FRT approximately accounts for between 6% and 20% of the total number of immune cells with higher numbers in the upper FRT (Givan et al., 1997; Hickey et al., 2011), while the number of leukocytes in the vaginal epithelium is remarkably low. By comparison, higher numbers of leukocytes, predominantly CD4+ T and CD8+ T cells, can be detected in the vaginal lamina propria, but still in lower numbers compared with the other parts of the FRT (Pudney et al., 2005). T cells in both the vaginal epithelium and the vaginal lamina propria appear to be predominantly of the memory phenotype. Furthermore, Pudney et al. (2005) reported that the ectocervical mucosa contained higher concentrations of CD4+ intraepithelial lymphocytes. These researchers have also shown that T cells and antigen presenting cells are most prevalent in the cervical transformation zone where the ectocervix transforms into endocervix and where the stratified epithelium ends and columnar monolayer begins. This suggests that this site functions as an immunological barrier to bacterial and viral pathogens and is the major inductive and effector site for cell-mediated immunity in the lower FRT (Fig. 1).

Unique for the upper FRT mucosae is the presence of several lymphoid aggregates (LA) consisting of an inner core of B cells surrounded by mainly CD8+ T cells and an outer halo of macrophages (Fig. 1). The size of the aggregates was found to vary throughout the stages of the menstrual cycle, with larger sizes during the secretory stage as compared with the proliferative stage (Yeaman et al., 1997, 2001). The absence of these aggregates in postmenopausal women provided evidence that these aggregates are under hormonal control. The functions of these LAs are not well documented yet. However, a potential role has been suggested in the suppression of cell-mediated immunity in the uterus during the secretory phase of the cycle, when ovulation and implantation are most likely to happen (Yeaman et al., 1997, 2001). This would imply that the presence of aggregates of immune cells is not dependent on the presence of infection.

In conclusion, the FRT is – in comparison with GIT mucosal systems – under strong hormonal control and is characterized by a unique distribution and phenotypes of DCs, LCs, NK cells, macrophages, neutrophils, B and T cells as well as unique set of Ig isotype. This indicates the presence of multiple levels of protection to minimize the risk of infection by potential bacterial and viral pathogens.

Healthy vaginal microbiota

Prior to the development of molecular methods, vaginal lactobacilli were identified by culture-dependent methods. The first reports on the presence of vaginal microbiota date from 1892 when the German scientist Albert Döderlein, reported on the presence of ‘Gram-positive, nonspore-forming rods, sometimes quite long and slender, with square or very tapering ends, occurring single or in chain, producing lactic acid that could inhibit the growth of pathogens’ (Döderlein, 1892). Till 1980, it was believed that the vaginal microbiota was dominated by Lactobacillus acidophilus, as determined by culture-dependent and microscopic methods, since in 1928, Stanley Thomas identified the Döderlein's bacillus as L. acidophilus (Thomas, 1928). However, after 1980, the group of microorganisms known as L. acidophilus was shown to be highly diverse. Based on different molecular methods, the group was separated into DNA-homology groups, which could not be distinguished biochemically and which form nine separate species of the L. acidophilus complex, that is, L. acidophilus, Lactobacillus amylolyticus, Lactobacillus amylovorus, Lactobacillus crispatus, Lactobacillus gallinarium, Lactobacillus gasseri, Lactobacillus iners, Lactobacillus jensenii, and Lactobacillus johnsonii (Du Plessis & Dicks, 1995; Falsen et al., 1999). As mentioned above, the culture-dependent methods excluded various Lactobacillus species. For example, most culture-dependent methods fail to identify L. iners, because it grows only on blood agar, but not in the normal Rogosa or de Man, Rogosa and Sharpe medium, typical for the growth of Lactobacillus (de Man et al., 1960). Lactobacillus iners was first isolated in a woman with a healthy vaginal microbiota and described based on culture-independent methods in 2002 (Burton & Reid, 2002). Recently, it was reported that L. iners has the smallest Lactobacillus genome consisting of a 1.3-Mbp single chromosome (Macklaim et al., 2011). Yet, the genome of L. iners contains a number of genes putatively involved in adaptation to the fluctuating vaginal environment. For example, L. iners has an iron-sulfur cluster (Fe-S) mainly detected in vaginal Lactobacillus isolates. For example, an Fe-S cluster was also recently reported in the genome of Lactobacillus pentosus KCA1 (Anukam et al., 2013) and Lactobacillus rhamnosus GR-1 (J. Macklaim, M.I. Petrova, A.S. Rodriguez, J. Vanderleyden, K. Marchal, G. Gloor, S. Lebeer and G. Reid, unpublished data). The Fe-S cluster from L. iners shares similarities with similar clusters from L. crispatus and L. johnsonii, common vaginal isolates, and is suggested to be involved in resistance to oxidative stress, especially in the vaginal niche where high levels of H2O2 are produced by other lactobacilli (Macklaim et al., 2011). In addition, several unique σ-factors were detected in the genome of L. iners. The σ-factors were reported to regulate gene transcription of various stress response genes and suggested to further contribute to survival in the presence of H2O2 (Macklaim et al., 2011). Up till now, L. iners is one of the most frequently isolated strains from vaginal mucosa of healthy premenopausal women (Lamont et al., 2011).

The composition of the vaginal microbiota changes drastically over time, based on changing estrogen levels during the maturation of women (Cribby et al., 2008). Soon after birth, the vaginal epithelium is colonized by a vast number of microorganisms. The majority of the vaginal bacteria originate from the GIT microbiota through a natural ascension independent of hygiene or from the surrounding skin epithelium. During this early stage of infancy, maternal estrogen induces thickening of the vaginal epithelium and, in this process, regulates the deposition of glycogen in the epithelial cells. During exfoliation of the epithelial cells, glycogen is released. Glycogen is a source of glucose, thereby favoring glucose-fermenting microorganisms (Boskey et al., 1999). The degradation of extracellular glycogen by lactobacilli is not yet well documented, but it was recently reported that L. iners differentially expresses genes responsible for breakdown of glycogen under vaginal conditions (Macklaim et al., 2013). This suggests that some lactobacilli are able to use the released glycogen as carbohydrate source. Postnatally, the maternal estrogen is metabolized, resulting in a thinning of the mucosa, and reduction of glycogen. A subsequent reduction in glucose-fermenting microorganisms, including lactobacilli, facilitates an increase in the vaginal pH, encouraging the proliferation of a wide range of aerobes and facultative anaerobes (Fig. 2). The vaginal microbiota is therefore during childhood mostly dominated by Gram-negative anaerobe bacteria including Veillonella, Bacteroides, Fusobacteria, and some Gram-negative cocci as well as Gram-positive anaerobe bacteria including Actinomyces, Bifidobacteria, Peptococcus, Peptostreptococcus, and Propionibacterium (Dei et al., 2010; Randelovic et al., 2012). The vaginal microbiota can include also some aerobic bacteria such as Staphylococcus aureus, Staphylococccus epidermidis, Streptococcus viridans, Enterococcus faecalis, Corynebacterium, and Diphteroides. Typical for the vaginal microbiota of prepuberal girls is the low frequency of lactobacilli, Gardnerella vaginalis, Prevotella bivia, Mycoplasma hominis, and yeast (Hill et al., 1995; Randelovic et al., 2012). With the beginning of puberty, the vaginal epithelium, under estrogenic control, once again thickens. As mentioned before, this glycogen-rich environment selects for glucose-fermenting microorganisms, provided glycogen can be broken down (Mac Bride et al., 2010). The microbiota present in this period of life has also been studied and is predominantly colonized by Lactobacillus species. Yamamoto et al. (2009) reported that the microbiota of adolescent girls is similar to the vaginal microbiota of adult women and dominated by L. iners, L. crispatus, L. jensenii, and L. gasseri. It remains to be investigated how this estrogen-induced increase of glycogen and its subsequent metabolism by vaginal lactobacilli is regulated at the molecular level.

Figure 2.

Changes in the vaginal mucosae during different stages of a woman's life. During prepuberty, the low levels of estrogen result in thin mucosae and low levels of glycogen, which stimulate the growth of diverse microbial species. In adult stage of life, the glycogen levels increase, due to the increase in the levels of estrogen. The degradation of glycogen to glucose selects for glucose-fermenting microorganisms such as Lactobacillus sp., which lower the vaginal pH and prevent growth of pathogenic bacteria. Postmenopause, the levels of estrogen once again decline, reducing the deposition of glycogen thereby selecting for a high diversity of bacterial species.

The vaginal microbiota in healthy adult women is mostly dominated by Lactobacillus species as mentioned above. In the last decade, the interest in the vaginal microbiota has strongly increased. The most frequently occurring species are L. crispatus, L. gasseri, L. iners, and L. jensenii (Table 1; Fig. 3). In a recent study, Ravel et al. (2011) showed that the vaginal microbiota can be divided in five major microbial communities based on samples from four ethnic groups (white, black, Hispanic, and Asian in North America; Fig. 3), as also nicely reviewed by Ma et al. (2012). According to Ravel et al. (2011), microbial communities belonging to group I (26, 2%), II (6, 3%), III (34, 1%), and V (5, 3%) were dominated by L. crispatus, L. gasseri, L. iners, and L. jensenii, respectively, and were isolated mainly from white and Asian women. The remaining 27% of the women forming community group IV, mainly black and Hispanic, were found to be part of a heterogeneous group. Group IV was characterized by strictly anaerobic bacteria, including Prevotella, Dialister, Atopobium, Gardnerella, Megasphaera, and Peptoniphilus (Ravel et al., 2011; Table 1). Other studies also reported that some women's vaginal ecosystems can be healthy without a Lactobacillus-dominant vaginal microbiota. Studies identified women with Atopobium vaginae, Megasphaera, and/or Leptotrichia species as dominant vaginal phylotype (Zhou et al., 2004, 2007; Srinivasan et al., 2012). The same species, which belong to the lactic acid producers, were also detected by Ravel et al. (2011) in group IV, suggesting that lactic acid production is crucial for an adult healthy vaginal ecosystem (Witkin et al., 2007; Ravel et al., 2011). In addition, diversity in the vaginal microbiota of different geographic area was also observed. For example, the vaginal microbiota of Nigerian, Belgian, and Brazilian women appears to be dominated mainly by L. iners (Anukam et al., 2006; Vitali et al., 2007; Martinez et al., 2008), whereas in Swedish, German, and Turkish women, L. crispatus is most commonly isolated (Kilic et al., 2001; Vasquez et al., 2002; Thies et al., 2007). In addition, the vaginal microbiota of Indian and Bulgarian women is dominated by Lactobacillus reuteri, L. gasseri, and Lactobacillus fermentum (Dimitonova et al., 2008; Garg et al., 2009; Table 1). Nevertheless, the exact number of vaginal community types is under discussion (Ma et al., 2012), in agreement with the discussion on the importance of GIT enterotypes (Arumugam et al., 2011). Recently, Koren et al. (2013) showed with data from the Human Microbiome Project that the differences in number of enterotypes across the human body depend on the sensitivity of enterotyping to the taxonomic depths used in constructing operational taxonomic units (Koren et al., 2013).

Table 1. Composition of vaginal microbiota during healthy and diseasea
Vaginal microbiota in healthVaginal microbiota associated with BV
  1. BV, bacterial vaginosis.

  2. a

    Based on Wertz et al. (2008), Ravel et al. (2011), Lamont et al. (2011), Gajer et al. (2012), and Ma et al. (2012).

Dominant vaginal Lactobacillus strainsCommon species isolated during BV
L. iners, L. crispatus, L. gasseri, L. jenseniiMycoplasma hominis, Gardnerella vaginalis, Prevotella spp., Mobiluncus spp., Staphylococcus epidermidis, Staphylococcus aureus, Streptococcus spp., E. coli
Less common Lactobacillus speciesOther species detected during BV
L. vaginalis, L. reuteri, L. rhamnosus, L. salivarius, L. plantarumAtopobium vaginae, Megasphaera spp., Leptotrichia spp., Clostridiales order, Dialister spp., Streptobacillus spp., Chloroflexi spp., Olsenella spp., Porphyromonas asaccharolytica, Shuttleworthia spp., Eggerthella hongkongensis
Rarely isolated vaginal Lactobacillus species
L. fermentum, L. brevis, L. agilis, L. coleohominis, L. suntoryeus
Other groups isolated in some women
Prevotella, Dialister, Atopobium, Gardnerella, Megasphaera, Peptoniphilus, Sneathia, Eggethella, Aerococcus, Finegoldia, Mobiluncus, Cryptobacterium, Gemella, Ruminococcus, Anaeroglobus, Leptotrichia, Parvimonas, Gardnerella
Figure 3.

Composition of vaginal microbiota in healthy adult women. The vaginal microbiota in healthy adult women can be divided into five community groups. Group I, II, III, and V are dominated by Lactobacillus species belongs to Lactobacillus crispatus, Lactobacillus gasseri, Lactobacillus iners, and Lactobacillus jensenii respectively, while group IV is diverse in composition. In addition, community group IV can be separated into two subgroups – IV-A and IV-B. Community group IVA is characterized by the presence of L. crispatus or L. iners and strict anaerobes, while IVB has a higher proportion of members of the genus Atopobium, Prevotella, Parvimonas, Sneathia, Gardnerella, and Mobiluncus. Figure mainly based on Ravel et al. (2011) and Gajer et al. (2012).

Until now, most studies are based on the collection of vaginal samples at a single time point. However, a recent study highlights the dynamics of the vaginal microbiota over a short period of time (16 weeks; Gajer et al., 2012). The authors were able to isolate five community groups in agreement with the Ravel et al. (2011). Groups or community states I–III were dominated by L. crispatus, L. gasseri, or L. iners, respectively, while community states IV-A and IV-B were heterogeneous in composition (Fig. 3). Community state IV-A was characterized by a modest proportion of either L. crispatus, L. iners or other Lactobacillus species as well as low numbers of strict anaerobic bacteria. Group IV-B was dominated by diverse number of bacteria belongs mainly to genus Atopobium, Prevotella, Parvimonas, Sneathia, Gardnerella, or Mobiluncus (Gajer et al., 2012). The authors were able to show that some of the vaginal bacterial communities markedly change over time switching from one to other class, whereas other stays relatively stable. For example, the vaginal communities dominated by L. crispatus often transform to a community state III dominated by L. iners, or to IV-A. In addition, community group III dominated by L. iners shifts more often to community type IV-B, but in rare cases to IV-A. In comparison, the community group dominated by L. gasseri rarely transits to other types and stays stable over time. Furthermore, the authors reported that the fluctuation of the vaginal communities and their constancy are affected by time in the menstruation cycle and to a certain extent by sexual activity. Nevertheless, the vaginal community function was maintained despite the changes in bacterial composition as indicated from the metabolite profiles (Gajer et al., 2012).

Of particular interest, these vaginal microbiota were recently shown to play a key role in the colonization of newly born infants (Dominguez-Bello et al., 2010). The authors showed that vaginally delivered infants acquire bacterial communities that resemble their own mother's vaginal microbiota, dominated mainly by Lactobacillus. In contrast, C-section infants were shown to harbor bacteria similar to those found on skin surface, which is dominated by Staphylococcus, Corynebacterium, and Propionibacterium. It has been suggested that this might explain the susceptibility of C-section-delivered infants to certain pathogens, but this requires further study (Dominguez-Bello et al., 2010). Furthermore, the vaginal microbiota has been shown to change drastically during pregnancy, characterized by a reduction in overall diversity and richness, but an increase of Lactobacillus species, Clostridiales, Bacteroidales, and Actinomycetales (Aagaard et al., 2012). This might be related to the vertical transmission of these microorganisms during vaginal delivery. The importance of vaginal transmission of beneficial lactobacilli from mother to child needs to be taken into account when the application of intrapartum antibiotics is considered upon maternal Group B streptococcal colonization (Ohlsson & Shah, 2013), which is currently a common practice that could have unforeseen side effects.

In postmenopausal women with a decrease in estrogen levels, the vaginal microbiota changes (Fig. 2). The microbiota in this stage was reported to be dominated by L. iners and G. vaginalis and was characterized by a lower abundance of Candida, Mobiluncus, Staphylococcus, Sneathia, Bifidobacterium, and Gemella. Also, the presence of other lactobacilli is strongly reduced, and consequently, growth of potential pathogenic bacteria is increased (Gupta et al., 2006; Hummelen et al., 2011b).

Overall, the microbiota of the human vaginal tract is characterized by a much lower diversity and number of microbial species in comparison with the GIT microbiota. The reason for the lower diversity of the vaginal microbiota is still unclear, but may be linked to differences in the nutrient availability, reduced competition with indigenous organisms, but also different immune activity and functioning, as described above (Cribby et al., 2008).

Vaginal microbiota during bacterial vaginosis

In comparison with the healthy vaginal microbiota dominated by lactobacilli, an abnormal vaginal microbiota is characterized by an increased diversity of species, mostly pathogens. These pathogens are able to infect the vaginal epithelium of the host, which subsequently results in severe infections such as bacterial vaginosis (BV), yeast vaginitis, or urogenital tract infections, when pathogens from the genital tract migrate to the bladder epithelium. For example, yeast vaginitis is characterized by an overgrowth of fungal pathogens mainly consisting of Candida albicans, but also Candida glabrata, Candida krusei, and Candida tropicalis. BV is associated with an abnormal growth of vaginal bacterial pathogens and a reduced number of vaginal Lactobacillus species. Symptomatic BV is mostly detected using the Amsel clinical criteria (Amsel et al., 1983) or the Nugent score (Nugent et al., 1991). The Amsel criteria are based on the presence of three or four symptoms such as vaginal pH higher than 4.5, a milky homogeneous vaginal discharge, detection of fishy odor and significant presence of clue cells (Amsel et al., 1983). The Nugent score is based on Gram staining of the vaginal fluid and quantification of the number of lactobacilli. The score ranges between normal (0–3) through intermediate (4–6) to BV (7–10) (Nugent et al., 1991). Common species detected by culture-dependent techniques from women with BV are anaerobes such as M. hominis, G. vaginalis, Prevotella spp., and Mobiluncus spp. as well as a variety of Gram-positive (S. epidermidis, S. aureus, Streptococcus spp.) and Gram-negative bacteria (Escherichia coli; Table 1). Recent molecular methods demonstrated that specific microorganisms are detected in particular stages of BV. Many species present during BV were prior to the use of molecular-based techniques not detected. For example, species belonging to the Clostridiales order, Megasphaera spp., Leptotrichia spp., Dialister spp., Chloroflexi spp. Olsenella spp., and Streptobacillus spp. can now also be detected (Wertz et al., 2008; Table 1). A typical example is A. vaginae, which was isolated for the first time in 1999 based on 16S rRNA gene sequencing (Rodriguez et al., 1999). Since then, it has been often isolated and detected in women with BV, much more than in those with a healthy vaginal microbiota (Lamont et al., 2011). Furthermore, some species previously designated as Lactobacillus minutus and Lactobacillus rimae within the lactic-acid-producing bacteria have now been reclassified to the genus Atopobium.

A typical characteristic for the species associated with BV is the formation of thick adherent biofilms on the vaginal epithelium, which are resistant to treatment with a broad range of antibiotics. Some studies suggest a role for H2O2-producing Lactobacillus in the prevention of BV, although this is not yet clear. Most studies report that the number of H2O2-producing Lactobacillus species, such as L. crispatus and L. jensenii, decrease during BV (Antonio et al., 1999; Song et al., 1999; Balkus et al., 2012). On the other hand, the less H2O2-producing L. iners is often detected during BV. The presence of L. iners during BV might be explained by better adaptation of this strain to the changes in the environment during BV, for example, better adaptation to the increased pH (Wertz et al., 2008). Furthermore, unknown factors during BV might be toxic for the other lactobacilli but not for L. iners.

In conclusion, the cause of BV and the associated reduction of lactobacilli are still unknown. One study suggests that lysogenic phages might be involved in the depletion of vaginal Lactobacillus (Kilic et al., 2001). The authors reported that four morphotypes of phage isolates from vaginal lactobacilli were able to infect a broad host range of Lactobacillus species of other women. Most lactobacilli containing lysogenic phage DNA were able to release phage particles into the environment, suggesting that these lactobacilli could be a source of infective phages and could drastically reduce the Lactobacillus population and by extension could lead to BV (Kilic et al., 2001), but the exact trigger is unknown. Martin et al. (2009) investigated whether chemotherapeutic and antibiotic treatment inducing SOS responses could be an important trigger, by the induction of resident prophages from their lysogenic hosts. Their results provide a complicated picture on the role of lysogenic phages. The authors found that lysogeny appears to be widespread among the tested vaginal Lactobacillus strains, although half of the strains harbored prophage sequences that were not responsive to SOS activation. In addition, in some cases when prophage induction could be achieved by treatment with mitomycin C, viable phages were not generated. Furthermore, most of the tested lactobacilli were able to produce H2O2, which is also an inducer of the SOS response and induction of lytic cycle. Taken together, their results suggest that H2O2 production selects for strains that harbor SOS-insensitive, defective prophages in vaginal lactobacilli, which are unable to promote phage lysis. They suggest that resident phages in H2O2-producing lactobacilli might have coevolved with their hosts, resulting in a more stable population (Martin et al., 2009). Nevertheless, this needs to be further documented. Damelin et al. (2011) showed higher levels of lysogeny in L. crispatus in comparison with L. jensenii, the former being also more frequently isolated from patients with BV. The authors propose that the high levels of lysogeny might be related to the better survival of L. crispatus compared with L. jensenii under the changing conditions in the vaginal environment during BV, by the presence of survival and/or adherence promoting genes by the lysogens (Damelin et al., 2011), but this remains to be further documented. On the other hand, in a more recent study, Macklaim et al. (2013) reported that CRISPRs genes, responsible for antibacteriophage defense, are highly expressed in L. iners under BV conditions, suggesting that L. iners is able to response to the changes in the phage load during BV. In addition, CRISPRs genes were recently also detected in the genome of L. pentosus KCA1 together with abortive infection (Abi) systems that can target different phases of phage development (Anukam et al., 2013). These mechanisms were also recently supported by Reyes et al. (2010), who propose that changes in the phage population in the human fecal microbiota are related with changes in the bacterial community and therefore can be used as markers for disease states. The obtained results are also supported by the metagenomic data from the Human Microbiome Project, which show that the CRISPRs genes are most abundant in Lactobacillus-targeting phages sequences in the vagina (Rho et al., 2012). Clearly, the exact role of bacteriophages in the vaginal environment and BV pathogenesis is still not well understood, and future studies are needed.

In addition, it is also of interest to mention that the urogenital tract appears to be the only microbial niche in the human body that shows clear differences in microbiota between women and men (Mändar, 2013). In male, the microbiota is present in the lower genital tract, mostly in urethra and coronal sulcus. Recently, 16S rRNA gene sequencing was used to characterize the microbiota of the coronal sulcus and urine collected from 18 adolescent men over three consecutive months. The coronal sulcus microbiota of most participants was more stable than their urine microbiota, and the composition was strongly influenced by circumcision. BV-associated taxa, including Mycoplasma, Ureaplasma and Sneathia, were detected in the corona sulcus specimens from sexually experienced and inexperienced participants (Nelson et al., 2012). Nevertheless, whether this is related to the pathogenesis of BV remains to be further investigated.

Key immunological and microbiological aspects of HIV infection

The process of HIV infection

The female genital tract mucosa is a portal of entry for several clinically relevant sexually transmitted viruses including HIV. At present, 34 million people are estimated to be infected with HIV, and c. 2.7 million new infections occurred worldwide in 2010 (http://www.unaids.org/globalreport/global_report.htm). Women appear to be more easily infected with HIV than men. Differences in social rank, behavior, sex hormone regulation, and especially organization of the mucosal surface appear to be involved (Iwasaki, 2010). The FRT mucosa – as described above – has indeed a unique structure and function, under the control of a variety of sex hormones (Wira & Rossoll, 2003).

Sexual transmission of HIV is mediated by exposure to infectious virions and/or infected lymphocytes and monocytes present in the semen. The ratio of transmissibility of cell-free vs. cell-associated viruses is still uncertain, but both are sources of infections and should be targeted by intervention strategies. Under normal circumstances, the incidence of HIV transmission from males to females is very low, within the range of one productive infection for every 200–2000 exposures (Schellenberg & Plummer, 2012). This can probably be explained by the presence of the protective Type II mucosa of the female vaginal epithelium cells (Schellenberg & Plummer, 2012). In contrast, the endocervix mucosa, as described above, is a single layer of columnar epithelium, which can easily support HIV transmission (Lederman et al., 2006). However, the endocervix is protected by a thick mucus layer that provides a physical barrier that efficiently can trap HIV virions or infected donor cells (Maher et al., 2005). Yet, free HIV virions can eventually cause infection in many different ways. For example, it has been reported that HIV-1 can bind to and enter epithelial cells from the lower FRT by transcytosis, or endocytosis and subsequent exocytosis, thereby causing productive infection (Wu et al., 2003; Stoddard et al., 2009; Fig. 4). Transcytosis of HIV virions has been demonstrated in vitro using genital tract–derived cell lines and primary human endocervical tissue (Stoddard et al., 2009). Once released from the epithelial cells, the virions can infect underlying leukocytes (Fig. 4). As mentioned above, the squamous epithelial cells also lack tight junctions, so that particles such as HIV-1 can be transported through the narrow gaps between the cells to the draining lymphatics as single particles or by infected donor lymphocytes and macrophages (Maher et al., 2005; Fig. 4). Moreover, disruption of the integrity of the vaginal mucosa, which can occur during sexual intercourse, BV, or other inflammatory conditions of the vagina, could allow direct contact of the virus particles with intraepithelial LCs and γδ CD4+ T cells, or might allow HIV-1 to reach suprabasal or basal epithelial cells that are more susceptible to viral transcytosis. Furthermore, viral entry could occur close to the transformation zone of the endocervix, which is highly enriched with CD4+ T cells (Fig. 1), thereby causing productive infection (Pudney et al., 2005).

Figure 4.

Vaginal invasion by HIV virions. The vaginal mucosa belongs to Type II and is therefore covered by a multicellular layer of nonkeratinized squamous cells. Mucus secreted by cervical glands and local epithelial cells provides a barrier to invading pathogens. In and underneath the epithelial cell layer, many innate immune cells are located, including LCs, γδT cells, DCs, and macrophages. Free HIV virions can eventually cause infection in many different ways. (1) Infection by free HIV virions: Free HIV virions trapped by the mucus layer can penetrate the epithelium through gaps between the epithelial cells and therefore reach the stroma and the submucosa. Furthermore, infected donor cells could also be trapped in the mucus layer and in this way release new virions. The virions can also be captured by LCs located in the epithelium layer and subsequently transferred to the submucosa. (2) HIV virions can also infect γδ CD4+ T cells on the epithelium. (3) Transcytosis: free virions or those released from the donor cells can be transferred through the basal epithelium cells by transcytosis, productively infecting these cells or penetrating between epithelial cells. (4) Migration of HIV virions to the submucosa: free virions or infected donor cells can migrate through a disrupted vaginal mucosa and reach the vaginal submucosa. (5) Infection of immune cells and local immune response: once in the stroma, free virions can infect the underlying CD4+ T cells or macrophages. Furthermore, the virions could also make contact with DCs and subsequently get transported to CD4+ T cells. (6) Spreading the infection: infected DCs, CD4+ T cells, macrophages, and LCs transport the virions subsequently to the submucosa and to the local lymph nodes. Figure mainly based on (Lederman et al. (2006), Hladik & McElrath (2008) and Iwasaki (2010).

Another possible way of HIV invasion occurs through LCs. Vaginal LCs were shown to efficiently internalize HIV-1 into their cytoplasmic compartment. After the LCs exit, the epithelium at the basal site, they are able to transport the HIV virions and spread the infection (Hladik et al., 2007). However, it is not well documented whether LCs can produce and release new HIV-1 virions. LCs express HIV receptors including CD4, CCR5, and the C-type lectin receptor langerin, but they do not express CXCR4 and DC-SIGN. Several studies suggest that HIV virions can efficiently bind to the LCs in different ways resulting in endocytosis of the virus and subsequent transport to the lymph nodes (Hussain & Lehner, 1995; Hladik et al., 2007). However, it has also been shown that LCs may act as a natural barrier to HIV-1 infection, by internalizing HIV-1 particles through langerin and degrading them intracellularly (de Witte et al., 2007). Another possible and more likely mechanism of infection with HIV is through DCs, which express both DC-SIGN and CCR5, unlike LCs, and therefore can support HIV infection. However, their exact role in mucosal transmission is not clear. In situ studies in a human explant model did not succeed to identify infected DCs in the cervico-vaginal stroma (Collins et al., 2000; Cummins et al., 2007). On the other hand, HIV-infected DCs were identified in tissue biopsies of the vaginal stroma of HIV-1-infected women. Although it is not clear whether HIV can efficiently infect DCs, the importance of DCs in capturing (trapping) the virus and transferring it by an ‘infectious synapse’ to CD4+ T cells and thus transmitting the viral infection is well documented (Cameron et al., 1992; Geijtenbeek et al., 2000). CD4+ T cells are the main target cells for HIV. They are present within the vaginal and the endocervical mucosa as memory T cells expressing high levels of CCR5 coreceptor (Hladik et al., 2007). In addition, macrophages and/or NKs cells present in the vaginal mucosa can also be a target for HIV (Cummins et al., 2007; Harada et al., 2007).

Although several types of innate immune cells can be targeted by HIV, the human body and particularly the innate immune system provide a response to the HIV infection by induction of several different cytokine and activation of several chemokine pathways. The innate immune response toward HIV is a double-edged sword, of which some responses efficiently restrict the virus, while other responses actually promote virus replication and/or transmission to target cells. The main recognition of HIV nucleic acid is through TLRs. TLR7 and TLR8 are able to recognize single-stranded viral RNA (ssRNA; Heil et al., 2004), while TLR3 recognizes double-stranded viral RNA (dsRNA; Alexopoulou et al., 2001). TLR3, TLR7, and TLR8 are localized in the endosome of macrophages, monocytes, and DCs (Heil et al., 2004). Hence, TLR3, TLR7, and TLR8 play important roles in antiviral innate immune response such as against HIV. Recognition of HIV RNA by one of the TLRs activates the NF-κB pathway and the production of TNF as well as activation of IFN-inducible genes, followed by production of Type I IFN (IFN-α/β). Increased levels of type I IFN are followed by a decline of HIV load by induction of a large number of signaling pathways, some of which can contribute to HIV-1 restriction (Florey et al., 2011). In addition, the expression of the viral accessory protein Nef by HIV stimulates production of the pro-inflammatory cytokines TNF and IL-6, which could stimulate proliferation of naive T cells (Swingler et al., 1999; Olivetta et al., 2003). Furthermore, TLR2 receptors present on the surface of innate immune cells and epithelial cells are able to specifically recognize gp120 from the viral envelope and subsequently to increase the production of TNF (Olivetta et al., 2003). Additionally, gp120 binds to the CCR5 coreceptor, which results in the release of chemokines such as macrophage inflammatory protein 1α/β (MIP-1α/β), MCP-1, and RANTES from macrophages. These chemokines recruit T cells and monocytes to the site of virus infection (Fantuzzi et al., 2001). Some of the cytokines produced by innate immune cells, CD4+ T and CD8+ T cells, can also have a significant impact on HIV replication and progression of the disease. For example, TNF can – depending on the receptor used – increase apoptosis of HIV-infected cells or initiate NF-κB signaling, resulting in increased production of virus (Herbein & Khan, 2008). IL-2 also plays a crucial role in HIV infection by initiating a Th1 CD4+ T-cell response and activating expression of IFN-γ and TNF (Duh et al., 1989; Reuter et al., 2012). Yet, Akinsiku et al. (2011) demonstrated that IL-2-expressing CD8+ T cells are related to high levels of protection during HIV infection (Akinsiku et al., 2011). Furthermore, IL-7 and IL-15, similarly to IL-2, can drive a CD8+ T-cell response, which might benefit the host during an HIV infection. Nevertheless, IL-2 as well as IL-7 and IL-15 can also induce HIV replication in T cells (Moran et al., 1993; Ferrari et al., 1995; Mueller et al., 2008). One of the important cytokines with well-known antiviral activity is IFN-γ. However, during HIV infections, IFN-γ has a dual role similar to the other cytokines described above and can both enhance HIV infection and inhibit viral replication (Reuter et al., 2012). In addition, some chemokines such as RANTES and MIP-1β can be considered as a potential inhibitors of HIV by recognizing CCR5, a major coreceptor used by HIV and predominantly responsible for viral transmission (Handen & Rosenberg, 1997).

BV associated with increased susceptibility to HIV

BV is associated with increased cases of spontaneous abortion, premature birth, and endometritis. Furthermore, BV is associated with increased susceptibility to sexually transmitted infections (STIs) such as HSV-2, gonorrhea, Trichomonas vaginalis, and Chlamydia trachomatis as well as HIV. For example, a study in Kenya showed that the presence of BV and absence of lactobacilli are significantly associated with acquisition of HIV (Martin et al., 1999). Furthermore, the presence of H2O2-producing lactobacilli showed to be protective, not only against HIV, but also against Neisseria gonorrhoeae, C. trachomatis, and T. vaginalis (Martin et al., 1999). However, three studies performed respectively in Thailand (Cohen et al., 1995), Uganda (Sewankambo et al., 1997), and Malawi (Taha et al., 1998, 1999) could not provide a clear relationship between BV and HIV infection. Different mechanisms are proposed for the increased susceptibility to HIV in the case of BV, including: (1) activation of the immune response and subsequent inflammation during BV; (2) disruption of the vaginal epithelium and subsequent transmission of HIV to the subepithelium where immune cells are present; and (3) reduced number of Lactobacillus species resulting in an increased pH and reduced H2O2 concentration compromising the protection of the vaginal epithelium. Indeed, an activated immune response appears to occur during yeast vaginitis and BV (Witkin et al., 2007). For example, levels of pro-inflammatory cytokines such as TNF, IL-1, IL-6, and IL-8 were shown to be significantly increased in BV (Spear et al., 2007). Furthermore, higher levels of IL-8 recruit immune cells and therefore may increase the risk of HIV infection (Narimatsu et al., 2005). Some in vitro studies using vaginal fluids from women with a normal vaginal microbiota and women with BV also highlight the potential role of BV in HIV infectivity. For example, the vaginal fluid from women with BV was shown to induce HIV expression and virus production in latently HIV-1-infected monocyte U1 cells, as compared with the vaginal fluid from women with a normal vaginal microbiota. In addition, it was shown that the levels of HIV-1 RNA in vaginal fluid of HIV-positive women were significantly increased in the presence of BV and Candida vaginitis, as compared with the levels in women with a normal vaginal microbiota (Cu-Uvin et al., 2001; Sha et al., 2005). Specific studies have focussed on the exact role of vaginal pathogens on their ability to stimulate HIV expression. For example, it was shown that G. vaginalis and M. hominis can significantly induce HIV expression in a monocyte cell lines. Other pathogens such as Peptostreptococcus asaccharolyticus, P. bivia, Streptococcus agalactie, and Streptococcus constellatus also have the ability to induce HIV expression in vitro (Al-Harthi et al., 1999; Hashemi et al., 1999, 2000; Simoes et al., 2001). In contrast, L. acidophilus appears not to stimulate HIV expression (Hashemi et al., 2000). Furthermore, it was shown that vaginal fluids from women with BV, but not from healthy women, induced IL-12 and IL-23 production in monocyte-derived DCs as well as CD40 and CD83 dendritic maturation markers (St. John et al., 2007). BV fluids also caused a decrease in the endocytotic ability of monocyte-derived DCs and an increased proliferation of T cells as compared with fluids from women with a normal vaginal microbiota (St. John et al., 2007). In agreement with St. John, Rebbapragada et al. (2008) also reported increased levels of pro-inflammatory cytokines and CD4+ T cells in HIV-infected women with BV, which was related with increased levels of HIV-1 RNA. Taken together, these results might be important for understanding the increased susceptibility of HIV infections during BV, because DCs are important cells in the recognition of the virus and subsequent transmission to the CD4+ T cells. Although St. John et al. (2009) found that BV fluids do not appear to increase transinfection from DCs to T cells, indirect effects of DC maturation on HIV transmission during BV cannot be ruled out.

Vaginal microbiota in HIV-infected women

In a first detailed molecular study performed in the USA, the diversity of the vaginal microbiota in HIV-infected women was reported (Spear et al., 2008). The study compared the diversity of the vaginal microbiota in HIV-infected women with BV or without BV, as well as in HIV-negative women with and without BV. Higher microbial diversity was detected in HIV+BV+ women in comparison with the HIVBV+ women. Furthermore, in HIV+BV+ women, three additional taxa were detected belonging to Propionibacteriaceae, Anaerococcus, and Citrobacter. The higher diversity of vaginal microbiota in HIV+ women might be related to suppressed immunity and promoted growth of vaginal pathogens. In addition, the authors report no significant differences between the HIV+ BV and HIVBV women, where the microbiota is dominated by Lactobacillus species. However, the microbiota in HIV+BV showed also the presence of Bifidobacteriaceae, Catonella, Coriobacterineae, and Prevotella, although in low concentrations (Spear et al., 2008). Some reports have shown that Candida spp. vaginitis is also frequently found in HIV-infected women (Minkoff et al., 1999; Ohmit et al., 2003), suggesting indeed that immunity in the lower genital tract is severely compromised during HIV infections. A study of the vaginal microbiota of HIV+ Tanzanian women confirmed the results reported above (Hummelen et al., 2010a). Indeed, a BV microbiota was detected in several microbial profiles of HIV-infected women. The most prevalent isolated species belong to P. bivia or members of the order Clostridiales (probably originating from the gut) and the family Lachospiraceae. BV appears to be more prevalent in HIV+ women with CD4+ counts of ≤ 200 cells mm−3. The percentage of L. crispatus in women with CD4+ cell counts of ≤ 200 cells mm−3 is also significantly lower than in women with high CD4+ cell counts (Spear et al., 2011). Nevertheless, in some HIV+ women, a normal vaginal microbiota represented by H2O2-producing L. crispatus or L. jensenii, but also L. iners and L. gasseri was also shown (Hummelen et al., 2010a; Spear et al., 2011; Balkus et al., 2012). Of particular interest, the presence of normal vaginal microbiota dominated by Lactobacillus in HIV-positive women was not associated with increased levels of HIV RNA, suggesting that Lactobacillus could be involved in reducing HIV shedding and subsequent transmission of the virus. In addition, detection of L. crispatus was reported to be associated with a 35% lower risk of HIV-1 RNA shedding (Mitchell et al., 2012). In case of women using highly active antiviral therapy (HAART), L. jensenii was associated with decreased levels of HIV-1 RNA (Mitchell et al., 2012). These results form an important incentive to study the potential of exogenously applied Lactobacillus species as probiotics for the treatment of BV and prevention of HIV transmission.

Potential role of endogenous vaginal microbiota for the prevention of HIV infection

As yet indicated, the vaginal microbiota dominated by lactobacilli seems to play a key role in the prevention of a number of urogenital diseases such as BV, yeast infections, and STIs such as HSV-2 and HIV. Postulated mechanisms against HIV include a direct inhibitory effect of the vaginal microbiota by production of lactic acid, H2O2, bacteriocins, and lectin molecules. In addition, indirect mechanisms such as prevention of the growth of microorganisms associated with BV, stimulation of the immune system and/or by the epithelial barrier function can also be envisaged (Fig. 5). However, many details still need to be unraveled.

Figure 5.

Postulated strategies of vaginal probiotic Lactobacillus strains to prevent HIV infections in humans. (1) In the vaginal lumen, Lactobacillus strains can inhibit HIV directly by producing lactic acid, H2O2, bacteriocins, and other inhibitory agents; (2) Lactobacilli can also preserve the integrity of the vaginal epithelium and compete with BV pathogens for receptors on the vaginal epithelium; (3) By producing different inhibitory agents, lactobacilli can directly kill bacterial and other viral pathogens; (4) The vaginal lactobacilli could also capture HIV by lectin-mediated binding to the HIV glycoproteins and in this way prevent infection; (5) The vaginal lactobacilli might enhance the local immune system during health and disease and thereby inhibit HIV infection. Nevertheless, all these postulated mechanisms remain to be substantiated, especially in vivo. Figure based on Klebanoff & Coombs (1991), Olmsted et al. (2005), Pretzer et al. (2005), Corey (2007b) and Hummelen et al. (2010b).

Production of antiviral components by the vaginal microbiota

As mentioned before, the presence of lactic-acid–producing bacteria is a hallmark of a healthy vaginal ecosystem. The vaginal mucosa is characterized by a pH of 4–5 depending on the Lactobacillus strains present. For example, when the microbiota is dominated by L. crispatus, the vaginal mucosa reaches a pH of 4.0 ± 0.3, while the presence of L. gasseri, L. iners, and L. jensenii results, respectively, in pH values of 5.0, 4.4, and 4.7 of the mucosal vagina (Ravel et al., 2011). Generally, the low pH of the vaginal mucosa is believed to be the main strategy to prevent bacterial and viral infections in the vaginal niche. For example, cell-free HIV-1 is inactivated at an acid pH (Martin et al., 1985). The same results were also reported by Ongradi et al. (1990), showing that the infectivity of cell-free HIV particles depends of the pH value and the time of incubation. The authors observed that incubation at pH 5.7 for 2 h is enough to inhibit the infectivity of HIV, whereas at lower pH in the range of 5.4, only 20 min are sufficient to completely inactivate the virus. However, the cell-associated viral infectivity was not inhibited by low pH values (Ongradi et al., 1990). On the other hand, O'Connor et al. (1995) observed that HIV is more acid stabile with no substantial reduction in infectivity occurring at pH levels as low as 4.5. On the other hand, an acidic environment appears to result in a decreased activation of T lymphocytes, which may result in decreased lymphocyte susceptibility to HIV-1 infection (Fig. 5; Hill & Anderson, 1992; Olmsted et al., 2005). Monocytes, macrophages, and lymphocytes, which might act as vectors for sexual transmission of HIV, were found to completely lose motility at a pH slightly below 6.0 (Olmsted et al., 2005). Lai et al. (2009) also reported that acid human cervico-vaginal mucus, obtained from donors with normal lactobacilli-dominated vaginal microbiota, efficiently traps HIV, causing it to diffuse > 1000-fold more slowly than it does in water. At an acidic pH 4, lactic acid, but not HCl, could abolish the negative surface charge on HIV without lysing the viral envelope, which may alter HIV surface protein structures and/or possibly inactivate the virus by disrupting the envelope membrane and exposing the capsid (Lai et al., 2009). Taken together, these results support the idea that maintaining a low pH in the vaginal lumen by production of lactic acid is important to reduce HIV transmission. This is also relevant for other sexually transmitted viruses such as HSV-2. For example, a significant inhibition of the replication of HSV-2 by acidic (pH 4.5) supernatant of L. fermentum 301 isolated from healthy Bulgarian women was reported (Dimitonova et al., 2007).

Apart from the maintenance of a low vaginal pH by the production of lactic acid, H2O2 has been suggested as an important antipathogenic strain-specific characteristic of vaginal lactobacilli. Various strains of L. crispatus, L. jensenii, and L. gasseri have been shown to produce H2O2 (Antonio et al., 1999; Song et al., 1999; Balkus et al., 2012). The production of H2O2 by Lactobacillus species seems to be a nonspecific host defense mechanism against different pathogens. The effect of H2O2 on the survival of different sexually transmitted viruses including HIV is not well studied. Klebanoff & Coombs (1991) could demonstrate a virucidal effect of a natural L. acidophilus isolate on HIV, based on its ability to produce and release H2O2. The authors were able to show that their L. acidophilus isolate, producing H2O2, at a density of 107 CFU mL−1 was virucidal to HIV by measuring the decrease of the viral replication in T lymphoblast CEM cells. The anti-HIV activity was not observed when this L. acidophilus strain was treated for 15 min at 100 °C or when it was replaced by a strain unable to produce H2O2, suggesting the important role of H2O2 in eventual inhibition of the HIV infections. Furthermore, the virucidal effect was inhibited by catalase, but not by heat-inactivated catalase (Klebanoff & Coombs, 1991). Besides these results, no other reports on the effect of H2O2 against HIV have been published so far, to the best of our knowledge.

A third and also strain-specific characteristic of lactobacilli is the production of bacteriocins. Bacteriocins are defined as microbial small peptide molecules with activity against closely related microorganisms (Cotter et al., 2013). However, recently, a broader activity spectrum against a wide variety of microorganisms and viruses was documented for several bacteriocins. For example, vaginally isolated L. fermentum and Lactobacillus brevis strains were shown to produce bacteriocin-like molecules against HSV-2 (Dimitonova et al., 2007). Although the authors were not able to completely characterize or purify the bacteriocin-like molecules, they were able to show that supernatant isolated from L. fermentum and L. brevis at neutral pH and treated with catalase showed activity against HSV-2, thereby postulating that bacteriocin molecules could be responsible for this activity (Dimitonova et al., 2007). It has been shown that genital lesions caused by HSV-2 are an important cofactor to increase the rate of HIV transmission and infection (Corey, 2007a). Recently, HSV-2 infection of keratinocytes was even shown to enhance the susceptibility of LCs to R5-tropic virus (which require CCR5, but not X4 for cell entry) by an indirect mechanism that involves the antimicrobial peptide LL-37 (Ogawa et al., 2013). Therefore, products from vaginal Lactobacillus strains that inhibit HSV-2 would potentially have benefits in indirectly preventing HIV infection.

An interesting specific class of bacteriocins includes lantibiotics, ribosomally synthesized peptides, produced by Staphylococcus, Lactobacillus, and Actinomycetes. A typical characteristic of lantibiotics is the presence of the amino acid residues lanthionine or methyllanthionine (Willey & van der Donk, 2007). The best studied lantibiotic, nisin (belonging to the type I lantibiotics), is widely used as a food preservative (Cotter et al., 2005), but we recently showed that it has no activity against HIV (Ferir et al., 2013). Another interesting bacteriocin is LabyA1, which belongs to a novel class of type III lantibiotics containing labionin, known as labyrinthopeptins (Meindl et al., 2010). Although LabyA1 was isolated from the actinomycete Actinomadura namibiensis DSM 6313 (Wink et al., 2003), we could show that it exhibits a broad anti-HIV activity in different T-lymphocytic cell cultures as well as in peripheral blood mononuclear cells (PBMCs). Furthermore, it was also shown to inhibit viral cell–cell transmission between persistently HIV-infected T cells and uninfected CD4+ T cells. LabyA1 did not inhibit HIV binding to DC-SIGN; nevertheless, it subsequently inhibited the transmission of HIV captured by DC-SIGN+-cells to uninfected CD4+ T cells. In addition, the safety profile of LabyA1 revealed no effect on the growth of vaginal lactobacilli populations at concentrations up to 120 μM (Ferir et al., 2013).

The genome sequences of various vaginal lactobacilli show the presence of bacteriocin-related genes. For example, in the genome of L. iners, three genes were found to encode putative bacteriocin immunity proteins, together with bacitracin resistance protein and a putative enterocin A immunity protein (Macklaim et al., 2011). The fact that L. iners has various defense systems against other bacteriocin-producers suggest they are commonly produced in this vaginal ecosystem, but this remains to be further substantiated. Furthermore, in the genome of the vaginal isolate L. pentosus KCA1, a cluster responsible for the biosynthesis of a novel antibacterial bacteriocin designated pentocin KCA1 was detected (Anukam et al., 2013). In addition, the draft genome sequence of the vaginal strain L. rhamnosus GR-1 suggest the presence of a bacteriophage peptidoglycan hydrolase related to lysostaphin (J. Macklaim, M.I. Petrova, A.S. Rodriguez, J. Vanderleyden, K. Marchal, G. Gloor, S. Lebeer and G. Reid, unpublished data), a special class of bacteriocins first identified in Staphylococcus strains (Donovan, 2007). Interestingly, Liu et al. (2011) showed that lysostaphin-expressing Lactobacillus plantarum WCFS1 could inhibit S. aureus in a modified genital tract secretion medium (Liu et al., 2011). However, similar bacteriocin activity against HIV remains to be documented.

Blocking HIV by binding to lectins of the vaginal microbiota

Molecules on the cell surface of the vaginal microbiota that directly interact with pathogens or host cells are postulated to play a role in the exclusion of bacterial and/or viral pathogens. Such interactions could be established via carbohydrate-binding proteins known as lectins, which interact specifically with carbohydrates on surface of pathogens and are highly specific for the ligand of interaction. Therefore, lectins on the cell surface of vaginal microorganisms could play a role in pathogen exclusion by (1) competitively binding to the same glycans on the host surface, thereby blocking adhesion or (2) by binding glycans on the pathogenic surfaces, thereby blocking virulence mechanisms such as adhesion and invasion (Fig. 5). Recently, some lectins, especially the ones highly specific for recognition of mannose (e.g. actinohivin and griffithsin) and N-acetylglucosamine (GlcNAc) residues (e.g. Urtica dioica agglutinin, UDA), have been shown to possess activity against HIV by binding of the glycans on the viral envelope and thereby blocking the virus entry process (Balzarini, 2007). Some of these lectins can inhibit the infection of T cells by cell-free virions through binding to the HIV gp120 glycoprotein. They can also block the interaction between HIV and the macrophage mannose receptor, thereby preventing the infection of macrophages. Furthermore, some lectins inhibit syncytia formation between virus-infected T lymphocytes and uninfected T cells. Lectins can also prevent the capture of HIV-1 particles to DC-SIGN-expressing cells. Additionally, lectins can block the DC-directed transmission of the virus to uninfected T cells (Balzarini, 2007).

Nevertheless, the information on lectins encoded by Lactobacillus species and especially vaginal isolates is still limited. Only a few studies reported on the presence of lectin-like proteins on the cell surface of probiotics that are involved in their adhesion capacity. The best documented lectin from a Lactobacillus strain is the mannose-specific adhesin (Msa) from L. plantarum WCFS1 (Pretzer et al., 2005). In silico analyses of the Msa protein revealed that this protein shows similarity with a mucus-binding protein (Mub) from L. reuteri and with SasA, an LPxTG-containing cell surface protein from S. aureus, referred to as a Concanavalin A-like lectin (Pretzer et al., 2005). The mannose-specific recognition was determined by pronounced agglutination of Saccharomyces cerevisiae cells (having mannans on their surface) by L. plantarum WCFS1 (Pretzer et al., 2005). The fact that Msa is highly mannose specific highlights the potential of using this strain or the protein for prevention and/or treatment of infections caused by pathogens that contain mannose residues on their surface, such as HIV and C. albicans. Some other studies suggest the presence of lectin-like molecules on the surface of lactobacilli. For example, Sun et al. (2007) demonstrated that L. plantarum Lp6 adheres to mucus in a lectin-like manner and significantly agglutinates S. cerevisiae as reported for L. plantarum WCFS1. A more recent study suggests that L. acidophilus FN001 also adheres in a lectin-like manner by recognition of carbohydrate moieties (Sun et al., 2010). The adhesion capacity of L. acidophilus FN001 was shown to be strongly inhibited in the presence of d-mannose and methyl-α-d-mannoside, suggesting that L. acidophilus FN001 contains mannose-specific protein(s) on its surface that mediate its adhesion to the Peyer's patches. Furthermore, L. acidophilus FN001 showed agglutination activity toward rabbit red blood cells in a mannose-specific manner, which decreased after protease pretreatment, and was able to strongly inhibit the adhesion of E. coli ATCC25922 to Peyer's patches in vitro (Sun et al., 2010). However, the lectin-like protein of L. acidophilus FN001 remains to be identified.

These few examples of lectins from Lactobacillus species were only reported for GIT isolates. Recently, we were able to detect three putative lectin-like proteins in the draft genome of L. rhamnosus GR-1 (J. Macklaim, M.I. Petrova, A.S. Rodriguez, J. Vanderleyden, K. Marchal, G. Gloor, S. Lebeer and G. Reid, unpublished data). Functional characterization of one of these lectin-like proteins by mutant analysis showed that such proteins are involved in tissue-specific adhesion of L. rhamnosus GR-1 to vaginal epithelial cells (M.I. Petrova, S. Malik, T.L.A. Verhoeven, M. van den Broek, J. Macklaim, G. Gloor, G. Reid, J. Balzarini, J. Vanderleyden and S. Lebeer, unpublished data). Other molecules on the surface of vaginal lactobacilli might also show activity against HIV. For example, Su et al. (2013) recently reported a ‘CD4-like receptor’ localized on the cell surface of Lactobacillus casei ATCC393 that can bind HIV and described its potential role in the inhibition of HIV infection in vitro. In addition, the ability of some vaginal lactobacilli to autoaggregate through interaction of cell wall proteins or lectins might be important factor in inhibition of HIV infections. The autoaggregating strains could inhibit HIV by direct capture of the virus or by strong adhesion to the epithelial cells, thereby competiting for epithelial cell receptors. Recently, we were able to show that sortase-dependent proteins are involved in the high autoaggregation and adhesion capacity of the vaginal isolate L. plantarum CMPG5300 (Malik et al., 2013). However, future studies are required to investigate the potential of L. plantarum CMPG5300 and other vaginal lactobacilli and their surface proteins for the prevention of HIV and related infections.

Stimulation of the immune system by the vaginal microbiota

As described above, the human vaginal epithelium is less densely populated by immune cells as compared with the GIT. Based on comparison with the GIT (Lebeer et al., 2010) and skin (Naik et al., 2012), it is tempting to speculate that the vaginal microbiota could regulate and stimulate the immune system to efficiently prevent bacterial and fungal pathogens as well as viral infections such as HIV. However, little is known about the relationship between the vaginal microbiota and the immune system. As described above, HIV infections are characterized by an increase of pro-inflammatory cytokines and pro-inflammatory responses. This has been linked with a disruption of the integrity of the vaginal mucosa and consequently further activation of HIV in infected people. Therefore, vaginal microbiota that can reduce a pro-inflammatory response could contribute to a decreased activation of HIV. This postulated mechanism has only been fragmentarily documented. For example, in vaginal epithelial multilayers treated with TLR agonists, significant reduction of IL-6 and IL-8 expression after treatment with L. crispatus ATCC 33820 was observed (Rose et al., 2012). Also, L. crispatus ATCC 33820 and to lesser extend L. jensenii ATCC25258 could induce a significant reduction of TNF secretion as well as some pro-inflammatory chemokines (MIP-1β and RANTES; Rose et al., 2012).

In another recent study, the role of the vaginal isolates L. rhamnosus GR-1 and L. reuteri RC-14, in the prevention of C. albicans vulvaginitis by stimulation and modulation of the vaginal immune responses was reported (Wagner & Johnson, 2012). These lactobacilli were shown to suppress C. albicans-induced NF-κB inhibitor kinase alpha (Iκκα), TLR2, TLR6, IL-8, and TNF expression, suggesting that they inhibit NF-κB signaling. In addition, C17β-estradiol was also shown to suppress expression of IL-1α, IL-6, IL-8, and TNFα mRNA, which is of high interest given the crucial role of estrogen hormones in the regulation of the vaginal ecosystem. As mentioned before, it has been shown that initiation of NF-κB signaling results in increased production of HIV, which enhances progression of the virus. Therefore, vaginal lactobacilli might be involved in controlling HIV infections by suppressing NF-κB effects (Wagner & Johnson, 2012). On the other hand, the vaginal strain L. rhamnosus GR-1 was also shown to stimulate TLR4 at both mRNA and protein level in cells challenged with E. coli, with concomitant increased NF-κB activation and TNF release (Karlsson et al., 2012). The role of these findings in relation to HIV expression remains to be studied.

Of interest, a recent study showed an important role for a new type of cytokine, designated IFN-ɛ, with potential antipathogenic activity against HSV-2 and Chlamydia muridarum. INF-ɛ was reported to be exclusively expressed by epithelial cells in the FRT (uterus, cervix, vagina, and ovary). Interestingly, IFN-ɛ expression was not induced by typical PRR signaling such as stimulation with synthetic TLR ligands or in vivo infection with pathogens, but showed to be clearly under hormonal regulation (Fung et al., 2013). However, the interplay between vaginal microbiota and IFN-ɛ production remains to be studied.

Inhibition of BV pathogens by the vaginal microbiota

As outlined above, the presence of vaginal pathogens and progression to BV or yeast vaginitis is significantly associated with a higher risk of a subsequent HIV infection, progression of disease and increased transmission of the virus to noninfected individuals. Therefore, the role of a healthy vaginal microbiota in the prevention of HIV could also be indirect by inhibiting urogenital infections (Fig. 5). Inhibition of bacterial or yeast pathogens by vaginal microbiota can be due to production of lactic acid, H2O2, and bacteriocins, as described to be also important for protection against HIV, as well as competition for receptor sites on the vaginal epithelium or by modulation of the vaginal immune system. Indeed, several studies have shown the importance of vaginal lactobacilli in the prevention of BV. For example, it was shown that four vaginal isolates – L. acidophilus CRL 1259, L. crispatus CRL 1266, Lactobacillus paracasei ssp. paracasei CRL 1289, and Lactobacillus salivarius CRL 1328 – were able to inhibit the adhesion capacity of S. aureus to vaginal epithelial cells by exclusion and competition (Zarate & Nader-Macias, 2006). Lactobacillus acidophilus CRL 1259 and L. paracasei ssp. paracasei CRL 1289 were also able to inhibit the adhesion of Group B streptococci. In a more recent study, the potential role of vaginal lactobacilli against G. vaginalis, P. bivia, Mobiluncus spp., and Bacteroides fragillis was reported (Matu et al., 2010). The vaginal lactobacilli were isolated from 107 Caucasian women, with L. jensenii as the dominant species. None of the strains was able to inhibit the growth of B. fragillis, but several strains were able to inhibit the growth of P. bivia, G. vaginalis, and Mobiluncus spp. based mainly on the production of lactic acid and possibly bacteriocins (Matu et al., 2010). In another study, the effect of cervico-vaginal lavage (CVL) from healthy women in the prevention of E. coli infection was recently reported (Kalyoussef et al., 2012). The CVL inhibited clinically isolated E. coli strains, but showed no activity against Lactobacillus. The inhibition of E. coli was related to the concentration of two human and four lactobacilli proteins present in the CVL. The bacterial proteins corresponded to L. jensenii and L. crispatus proteins and showed similarity to Lactobacillus cripatus S-layer proteins (Kalyoussef et al., 2012). S-layers proteins of lactobacilli have been previously shown to be involved in pathogen exclusion such as of enterohemorrhagic E. coli O157:H7 (Johnson-Henry et al., 2008) and Junin virus (JUNV; Martinez et al., 2012), as well as immune stimulation via DC–SIGN interaction (Konstantinov et al., 2008).

Finally, lactobacilli could also inhibit other STIs such as HSV-2 and in this way may prevent infections of HIV. For instance, Conti et al. (2009) have reported that the capacity of lactobacilli to adhere to vaginal epithelial cells is related to the degree of inhibition of HSV-2 infection. Lactobacillus brevis CD2 was reported to have a higher inhibition capacity related to a high adhesion to epithelial cells, whereas L. plantarum FV9 and L. salivarius FV2 showed intermediate and low adhesion and inhibition of HSV-2, respectively. Recently, the antimicrobial peptide LL-37 produced by HSV-2-infected keratinocytes was shown to be involved in the enhancement of HIV infection of LCs by strongly upregulating the expression of HIV receptors in LCs (Ogawa et al., 2013). Whether vaginal lactobacilli have an inhibitory effect on this LL-37 induction remains to be studied.

Importance of the GIT for HIV replication and progression of disease

HIV infections have also a profound impact on the GIT. They are characterized by increased GIT inflammation, increased permeability and malabsorption, inflammatory infiltrates of lymphocytes and damage to the GIT epithelial layer, including villous atrophy, crypt hyperplasia, loss of tight junctions, and villous blunting (Brenchley & Douek, 2008). These events have a detrimental effect on the function of the intestine and increase the risk of GIT infections and disorders (Kelly et al., 2009). Furthermore, the hallmarks of HIV infection include chronic activation of the immune system and a depletion of CD4+ T cells, which is more prominent in the GIT than in peripheral blood and lymph nodes (Mehandru et al., 2004). The GIT is thus considered as a preferable site for progression of HIV infection. Furthermore, other leukocyte subsets in the GIT are altered because the majority of the body's lymphocytes are localized in the GIT (MacDonald, 2008). Indeed, increased turnover, cell cycle perturbations, apoptosis, and altered functionality among CD8+ T cells, B cells, and innate immune cells are observed (Fig. 6). Of note, DCs are significantly decreased, including abnormal levels of plasmacytoid and myeloid DCs, and loss of mucosal CD103+ DCs (Klatt et al., 2012; Fig. 6). Also, macrophages, which are potential targets for HIV infection, have been shown to have a reduced capacity to phagocyte bacterial products in mucosal tissues of HIV patients (Pugliese et al., 2005).

Figure 6.

GIT mucosa during chronic infection with HIV. Typical characteristics of HIV infections are the morphological changes in the GIT mucosa, such as blended villi and crypt hyperplasia, as well as a compromised barrier function due to damaged tight junctions between epithelial cells. Apoptosis of epithelial cells induces an increased microbial translocation and permeability and causes chronic inflammation. Furthermore, decreased numbers of immune cells, especially CD4+ T cells, are observed. Th17 cells playing an important role in maintaining the integrity of the mucosa and defense against pathogens are easy targets for the HIV virions and are especially depleted. In addition, skewing of the dominant Th2/Th1 in favor of Th2 T cells is also observed with subsequent increased levels of IgE antibodies. Increased numbers of Th2 T cells activate further HIV replication and cause bacterial translocation. Figure based on Brenchley et al. (2006), Brenchley & Douek (2008), Hummelen et al. (2010b), and Lackner et al. (2012).

During HIV infection, the virus appears also to switch the immune system by activation of Th2 responses over Th1 (Clerici & Shearer, 1993). Indeed, HIV patients with progressive disease have symptoms that closely resemble allergies (Rancinan et al., 1997). Th2 skewing induces an increase of the production of IL-4, IL-5, and IL-13 with deleterious effects on the humoral and cellular immunity, similar to the immune deficiency of allergic patients (Clerici & Shearer, 1993). The increased levels of IL-4 subsequently hyperactivate B cells and increase the production of IgE (Fig. 6). Furthermore, it has been suggested that CD4+ Th2 cells efficiently support HIV replication, while Th1 dominant do not (Maggi et al., 1994). As expected, Th1 cytokines are drastically suppressed in HIV patients. It seems that the early antiviral cytotoxic T-lymphocyte response against HIV fails, while IgE synthesis prevents the production of effective antiviral Ig. This immune imbalance may induce an increased inflammation and barrier dysfunction in the GIT, as the increased IL-4 levels can compromise the antimicrobial function of Th17 (Harrington et al., 2005). Although the Th17 response could have a protective role against HIV infection, Th17 cells appear to be an easy target for HIV replication, which results in a depletion of CD4+ IL-17-producing cells in humans in vivo (Brenchley et al., 2008). Furthermore, depletion of Th17 cells from the gut in HIV-1 infection is associated with microbial translocation, chronic immune activation, and disease progression (Chege et al., 2011). Taken together, HIV infection and progression could be considered as a disease of the GIT tract.

The source of systemic inflammation during HIV infection remained unidentified until 2006 when it was demonstrated for the first time that depletion of intestinal mucosal immune cells in the GIT results in systemic immune activation through the increased translocation of microorganisms and bacterial products from the intestinal tract (Brenchley et al., 2006). Indeed, Brenchley et al. reported that HIV-infected humans with progressive disease show increased levels of plasma lipopolysaccharide (LPS) in the bloodstreams, as an indicator of increased microbial translocation. Furthermore, chronic in vivo stimulation of monocytes by the increased levels of LPS associated with activation of innate and adaptive immune was reported. The pro-inflammatory environment may then cause further damage to the gut barrier function, increasing bacterial translocation and subsequently systemic inflammation (Fig. 6), although this remains to be further documented. Interestingly, a decrease in plasma levels of LPS and reconstruction of CD4+ T cells was observed after HAART (Brenchley et al., 2006). The authors also reported that in nonprogression patients, plasma LPS was shown to be lower than in those patients with progressive HIV infection (Brenchley et al., 2006).

The intestinal microbiota of HIV patients is also significantly altered and appears to contain higher numbers of pathogens and lower numbers of beneficial microorganisms. For example, Gori et al. (2008) reported that the relative amount of bifidobacteria and especially lactobacilli in the HIV-positive population were significantly lower in comparison with healthy people. In addition, higher levels of Pseudomonas aeruginosa in HIV-positive individuals were reported compared with the levels of healthy individuals. Similarly, C. albicans was detected in 100% of the feces samples of HIV-positive individuals (Gori et al., 2008). The observation that HIV patients are especially susceptible to C. albicans infection could be explained by the diminished amount of Th17 cells, which are easy targets for HIV and of crucial importance for anti-Candida host defenses. An abnormal GIT microbiota during HIV infections was also reported by Merlini et al. (2011). More than 90% of HIV-infected patients harbored a bacterial population enriched with Enterobacteriales, while < 60% harbored Lactobacilllales (Merlini et al., 2011). In a more recent study, higher levels of Enterobacteriales were also reported in HIV-positive subjects compared with controls. Furthermore, the proportions of Enterobacteriales and Bacteroidales were significantly correlated with duodenal CD4+ T-cell depletion and peripheral CD8+ T-cell activation, respectively (Ellis et al., 2011).

Future studies need to focus on the role of GIT microbiota in the possible prevention of HIV infections similar to the vaginal microbiota. It was reported that transmission of HIV through GIT varied within the range of one productive infection for every 20–200 exposures for rectal transmission, one over 2500 for the upper GIT transmission, and one over 5–10 for upper GIT transmission with breastmilk (Hladik & McElrath, 2008). However, recent studies show that mouse mammary tumor virus, also a retrovirus, interacts with the GIT microbiota, which induces immune evasion pathways (Kane et al., 2011). Therefore, the successful transmission of the virus probably depends on the commensal microbiota. Furthermore, the GIT microbiota can have a large impact on viral infections at other sites in the human body. Nevertheless, this requires further study.

Probiotics and their potential role in HIV infection

During the last decades, an increased number of scientific studies highlight the potential of exogenously applied probiotics to promote human health, such as for the prevention and treatment of viral infections. Probiotics are defined as ‘live microorganisms which, when administrated in adequate amounts, confer health benefits on the host’ (WHO/FAO, 2001). Only rather recently, intervention trials have been performed with exogenously applied probiotic Lactobacillus bacteria in which their potential for improving the lives of HIV-infected patients or even for preventing HIV infections is investigated. The rationale behind this is that probiotics can have a dual role depending on the site of the body they are active. On the one hand, they can be used to increase the Lactobacillus numbers in the vaginal mucosa, which is the entry point for HIV infection as well as for transmission of the virus. Furthermore, exogenously applied lactobacilli can be used for the treatment and prevention of BV, often associated with HIV infections. For example, it was shown that daily oral intake of L. rhamnosus GR-1 and L. reuteri RC-14 resulted in colonization of the vagina, with a concomitant reduction in bacterial and yeast pathogens in this niche (Reid et al., 2003; Reid, 2008).

On the other hand, exogenously applied probiotics may also exert health benefits by enhancing the GIT mucosa, because the GIT is identified as a site of considerable early HIV replication, CD4+ T-cell destruction, and systemic inflammation, as mentioned above. The question then is whether blocking the first step of the cascade – the leaky gut – could prevent HIV from proceeding forward and eventually keep infected individuals healthier (Wenner, 2009). Various probiotic organisms applied in the GIT, mainly lactobacilli and bifidobacteria, have been shown to enhance intestinal epithelial barrier function, reduce inflammation, and support effective Th-1 responses. For example, probiotics can enhance the gut barrier function and reduce bacterial translocation (Luyer et al., 2005; Forsyth et al., 2009) by improving the beneficial interactions between the commensal enteric microbiota and the host during health and disease. Moreover, probiotics can restore GALT homeostasis by, for example, inducing regulatory mechanisms to down-regulate inflammation (Pessi et al., 2000; Braat et al., 2004). In vitro studies showed evidence that probiotics can skew away the immune system from a Th-2 dominant state (Iwabuchi et al., 2009; Hougee et al., 2010) and influence DCs to skew T cells toward Th-1 polarization (Mohamadzadeh et al., 2005), thereby causing a recovery in intestinal tolerance. Furthermore, probiotics are able to make the intestinal environment less hospitable for pathogens by producing antimicrobial compounds, lowering the pH, and reducing the adhesion and invasion of pathogens. These beneficial characteristics could prevent GIT infections and improve the quality of life of HIV patients (Hummelen et al., 2010b).

Intervention studies with probiotics in HIV patients

In recent studies, several intervention studies with HIV/AIDS patients were reported. The best studied probiotic organisms for these applications are L. rhamnosus GR-1 and L. reuteri RC-14, originally isolated from a healthy female urogenital tract with the capacity to adhere to urogenital epithelial and vaginal cells (Reid et al., 1987). These strains have also a documented capacity to inhibit the growth and adhesion of urogenital and intestinal pathogens, such as inhibition of the growth, adhesion, and biofilm formation of C. albicans by L. rhamnosus GR-1 (McMillan et al., 2011; Kohler et al., 2010). Moreover, L. rhamnosus GR-1 and L. reuteri RC-14 produce lactic acid that kills bacteria and viruses, including HIV (G. Reid, pers. commun.), H2O2, and bacteriocin-like compounds. Lactobacillus rhamnosus GR-1 can colonize the vagina (Reid et al., 1994; Gardiner et al., 2002) and the intestine (Reid et al., 2001) upon exogenous application. In addition, they show a good survival capacity in milk and hence have the ability to be delivered in a yogurt form without deterioration of taste or structure (Hekmat et al., 2008). These characteristics define L. rhamnosus GR-1 as a potential vaginal probiotic that is active in the GIT, the vagina, and the distal urethra and could restore health in these areas.

An initial study conducted in Nigeria administrated yogurt supplemented with L. rhamnosus GR-1 and L. reuteri RC-14 to 24 female HIV/AIDS patients for 15 and 30 days, respectively, which was associated with complete resolution of diarrhea, nausea, and flatulence in all tested subjects and a slight increase in CD4+ counts for selected patients (Anukam et al., 2008). These positive results have led to the development of specialized kitchens in Mwanza (Tanzania), and the production of Fiti probiotic yogurt supplemented with L. rhamnosus GR-1. Thirty days after the consumption of Fiti probiotic yogurt, almost all participants showed improvement in their weight and higher levels of vitamins and micronutrients than the controls. Furthermore, consumers of the probiotic yoghurt had fewer fungal infections, less episodes of diarrhea, and a lower degree of tiredness (Reid, 2010).

These promising results led to a second study using the same L. rhamnosus GR-1 Fiti probiotic from the Mwanza community kitchen (Irvine et al., 2010). Probiotic consumption for 70 days by 68 participants, compared with a control group of 82 participants, showed to increase the CD4+ count while most treated patients could also work 2 h longer a day, had less frequency of fever, and no diarrheal symptoms. A more recent study investigated the effect of long-term intake of capsulated L. rhamnosus GR-1 and L. reuteri RC-14 strains on the health of HIV/AIDS patients (Hummelen et al., 2011a). From baseline to 10 and 25 weeks of probiotic intake, the CD4+ counts increased in the probiotic group, while no differences in the immune markers IFNγ, IL-10, IgG, and IgE, diarrhea incidence or adverse events were observed. In addition, in a study performed in Canada, consumption of L. rhamnosus GR-1 yogurt supplemented with micronutrients was shown to support the immune system of HIV-infected individuals and improve their ability to perform daily tasks (Hemsworth et al., 2012). Further studies will be important to investigate the mechanisms by which L. rhamnosus GR-1 and L. reuteri RC-14 support health benefits.

Of note, a recent study also reported a possible role for a probiotic yogurt containing L. casei Shirota in the clearance of human papillomavirus (HPV)-related cervical lesions (Verhoeven et al., 2013). Whether such protection against HPV-related lesions of the vaginal epithelium will indirectly results in decreased susceptibility for HIV infection remains to be investigated. The results indirectly can be implemented for the prevention of HIV infections by protecting the vaginal epithelium from other pathogens such as HPV.

Recombinant lactic acid bacteria as microbicidal agents against HIV

Over the last 10 years, there is an increased interest to enlarge the anti-HIV potential of Lactobacillus species and other lactic acid bacteria by genetic engineering. Potential anti-HIV molecules for expression by probiotic bacteria are proteins targeting the cell binding and fusion processes of HIV, thereby blocking infection and transmission. Hereby, the glycoproteins on the cell surface of HIV, that is, the transmembrane gp41 and the external gp120, which are important for the virus entry into the cells, are important targets (Botos & Wlodawer, 2005). In the process of infection, gp120 binds to the CD4 receptor and CXCR4 or CCR5 coreceptor on the host T cells and macrophages, which triggers a conformational change in gp41 and initiates fusion of the viral envelope with the host cell. Soluble CD4 receptors, chemokines, neutralizing antibodies, and other active molecules such as lectins or carbohydrate-binding agents have yet been reported to inhibit the initial binding of gp120 to the host cell, while peptides that bind the amino-terminal domain of gp41 are reported to block the fusion and entry into the host cell (Nikolic & Piguet, 2010). These are all interesting ‘anti-HIV’ molecules for genetic engineering in lactobacilli.

Certain studies have delivered proof of concept that bioengineered probiotic strains can be used for delivery of different active molecules against HIV. The first study that succeeded in the expression of active anti-HIV molecules was carried out in the vaginal L. jensenii strain Xna-651 (Chang et al., 2003). Lactobacillus jensenii was genetically modified to secrete the first two extracellular domains of the human CD4 receptor (2D CD4) responsible for high-affinity binding to HIV gp120. Lactobacillus-derived 2D CD4 molecules were able to inhibit HIV-1 infections of cultured cells in a dose-dependent manner in vitro (Chang et al., 2003). Furthermore, coincubation of the engineered bacteria with HIV-1 led to a significant decrease in virus infectivity of HeLa cells expressing CD4_CXCR4_CCR5. Anchoring of the 2D CD4 on the Lactobacillus surface was later successfully achieved using a natural vaginally isolated L. jensenii strain 1153 (Liu et al., 2008). Other authors investigated the role of different natural immune inhibitors, which are ligands for HIV receptors such as CCR5 and ICAM-1. For example, Chancey et al. (2006) designed a L. casei 393 strain expressing a single-chain variable fragment (scFV) specific for the intracellular adhesion molecule 1 (ICAM-1). ICAM-1 molecules are involved in cell–cell interactions and presentation of HIV antigens of infected monocytes and CD4+ T lymphocytes to uninfected cells. The recombinant anti-ICAM-1 scFV antibody was indeed shown to block cell-associated HIV-1 transmission across an in vitro culture model of the cervical epithelium (Chancey et al., 2006). Furthermore, Liu et al. (2007) were also able to express a range of HIV inhibitors by stable chromosomal integration in the genome of L. reuteri RC-14. Four different HIV inhibitors were included in the study: MIP-1β (inhibiting CCR5 coreceptor binding), T-1249, a next-generation T-20 peptide inhibitor (inhibiting hairpin formation and membrane fusion between the virus and host cell), CD4D1D2-IgKLC, and CD4D1D2-IgG2HC (the light- or heavy-chain variable domain of human IgG2 replaced by the D1D2 domain of human CD4). The authors achieved two different cellular locations for the recombinant HIV inhibitors – cell wall anchored and secreted in the surrounding medium. All recombinant inhibitors used by the authors proved to inhibit HIV and SHIV infections in human PBMCs (Liu et al., 2007). More recently, Vangelista et al. (2010) were able to construct a recombinant L. jensenii 1153 strain, which secretes the anti-HIV-1 chemokine RANTES and C1C5 RANTES, a mutated analogue that acts as a CCR5 antagonist and is devoid of pro-inflammatory activity. The recombinant wild-type RANTES and C1C5 RANTES showed inhibitory activity against HIV-1 infection in CD4+ T cells and macrophages in vitro. The authors were able to show a similar antiviral activity against six different R5 HIV strains, suggesting in vitro cross-clade protection (Vangelista et al., 2010).

Other authors have used fusion inhibitor peptides derived from the gp41 transmembrane envelope glycoprotein, which exhibit virus-blocking properties upon overexpression. Rao et al. (2005) were able to construct a recombinant probiotic strain of E. coli Nissle 1917 that secretes an HIV–gp41–hemolysin A hybrid peptide, which blocks the fusion and entry into target cells. Remarkably, the recombinant E. coli Nissle 1917 was able to colonize the murine colon and cecum, and in low concentration also rectum, vagina, and small intestine and to persist for a period of weeks to months. Furthermore, the recombinant strain was able to grow and secrete HIV–gp41–hemolysin A hybrid peptide in situ (Rao et al., 2005). In addition, Pusch et al. (2006) were also able to engineer L. plantarum ATCC 14917 and L. gasseri ATCC 9857 in a way that they secrete various HIV-1 C-peptide fusion inhibitors (FI-1, FI-2, and FI-3) targeting the highly conserved heptad repeat-1 region in the transmembrane ectodomain of gp41 and therefore neutralizing HIV infections.

A different approach to target the HIV envelope and thereby inhibit HIV infection is through the use of lectins. The first recombinant lectin secreted by probiotic bacteria was cyanovirin-N (CV-N). Giomarelli et al. (2002) studied the anti-HIV-1 activity of the microbicidal compound CV-N, secreted by recombinant Streptococcus gordonii. CV-N, derived from the cyanobacterium Nostoc ellipsosporum, was shown to have antiviral activity by binding to high mannose residues on the HIV-1 viral envelope, thereby blocking the fusion of HIV-1 with the cell membrane and the transmission of its infection. The authors showed that soluble recombinant CV-N was able to specifically bind gp120 in a concentration-dependent manner, and recombinant strains expressing CV-N on their surface were able to capture HIV virion (Giomarelli et al., 2002). In further studies, Pusch et al. (2005) used two different strains, Lactococcus lactis MG1363 and L. plantarum NCIMB8826, for expression of CV-N and two different cell locations – intracellular and extracellular. The low concentration of extracellular recombinant CV-N at first was successfully increased after codon optimization and recharging of the N-terminal domain of the fused CV-N (Pusch et al., 2005). Liu et al. (2006) were able to express CV-N in the natural vaginal isolate L. jensenii 1153. Recombinant CV-N was shown to decrease the infectivity of both CCR5-tropic HIV and CXCR4-tropic HIV in vitro. The recombinant L. jensenii strain was capable of colonizing the vagina and producing full-length CV-N in situ when administered intravaginally to mice during the estrus phase (Liu et al., 2006). Furthermore, the recombinant L. jensenii was able to colonize the vaginal mucosa of Chinese rhesus macaques and to reduce by 63% the transmission of simian/human immunodeficiency virus. Colonization and prolonged antiviral protein secretion by the genetically engineered lactobacilli did not cause any increase of pro-inflammatory markers (Lagenaur et al., 2011). Nevertheless, more studies are required to conclude the safety of the recombinant Lactobacillus strain. For example, other studies have shown that CV-N promotes secretion of pro-inflammatory cytokines and chemokines from human PBMCs (Huskens et al., 2008). Moreover, CV-N has a mitogenic activity that results in morphological changes in subpopulations of the treated cells. This mitogenic characteristic of CV-N can cause the development of cells that are more susceptible to HIV infection (Huskens et al., 2008). Other studies have reported that it also activates quiescent CD4+ T cells and promotes T-cell proliferation (Balzarini et al., 2006; Huskens et al., 2008). These side effects of CV-N as a microbicidal drug suggest that it is not an ideal agent for the prevention and treatment of HIV infection.

Recombinant Lactobacillus strains can be also used to induce specific immune responses against HIV. For example, Kajikawa et al. (2012) were able to design L. acidophilus NCFM expressing HIV-1 Gag protein on the bacterial surface with coexpression of Salmonella flagellin used as adjuvant. The authors were able to show that the recombinant L. acidophilus strain induces a specific immune response in vitro and in vivo and therefore can serve as a vaccine vector to promote an immune response against HIV (Kajikawa et al., 2012).

Thus, the use of bioengineered probiotic strains in the prevention of HIV infections has already shown promising results. However, most of the studies performed so far were only carried out in vitro. Future studies need to investigate the potential role of the recombinant probiotics in vivo in order to proof the safety profile and efficacy of the bioengineered probiotic strains as mucosal delivery system and as potential new microbicides. In addition, the recombinant strains or the recombinant proteins produced by the lactobacilli should not interfere with the normal vaginal microbiota. For example, some abiotic microbicides were reported to modify cervico-vaginal innate immunity and actually reduce the number of Lactobacillus species in vitro (Fichorova et al., 2011). Furthermore, some abiotic microbicides were shown to affect the vaginal microbiota and shift the colonization of the vaginal mucosa to the presence of strict anaerobes and a significantly reduction in lactobacilli (Ravel et al., 2012). These results suggest that culture-dependent and independent evaluation of candidate microbicides on the vaginal microbiota should be considered. For example, we recently were able to show that different lectins with activity against HIV do not inhibit the growth of a wide variety of vaginal lactobacilli, do not affect the viability of these bacterial isolates, and have no significant impact on their adhesion properties to human epithelial HeLa cell monolayers (Petrova et al., 2013). In addition, ethical and moral considerations should be also taken into account, because the bioengineered probiotics are genetically modified organisms. Nevertheless, the future of recombinant lactobacilli is still open, because new technics of gene expression will be developed and optimized.

Conclusion and perspectives

The human microbiota plays an important role during the entire life of humans of which protection against pathogenic infections is well known. Although various studies have already addressed the importance of the GIT microbiota and its relation with diseases as well as HIV, studies on the urogenital microbiota are lagging behind. Nevertheless, an emerging number of studies have documented the role of the vaginal microbiota in the prevention of BV and related diseases. Because the vaginal epithelium is an important entry point for HIV, a better understanding of the vaginal microbiota and its relationship with the immune system is needed. Such detailed understanding includes focused genetic studies on the members of the beneficial vaginal microbiota in search for probiotic factors. Ultimately, molecular studies should allow tailored application of specific probiotic strains, either orally via the GIT or vaginally, based on well-documented modes of action. Future studies should also exploit the promise of genetic engineering of vaginal microbiota with specific anti-HIV molecules. Other strategies can also be envisaged, which are perhaps more plausible. Of note, applications such as replenishment of lactic acid – in comparison with short-chain fatty acids in the GIT or application of glycogen like sugars – in comparison with the prebiotics for the GIT, can be considered, given the apparent important role of the latter in the establishment of a healthy vaginal ecosystem.

It can be concluded that the reported studies that focus on the reconstitution of a healthy vaginal microbiota to improve the life of HIV patients have shown promising results so far in combination with standard medical treatments. Evidently, these studies need to be carried out on a larger scale and with more parameters analyzed in order to make the conclusions more firm. The very convincing studies addressing the important role of the GIT microbiota on bowel diseases and functioning of the gut epithelium are a good driver to further invest in these types of studies.

Acknowledgements

Part of the research of the authors is supported by the KU Leuven (PF 10/18). We acknowledge the valuable help of Shweta Malik, Hanne Tytgat, Marijke Segers, and Elke Lievens for carefully reading the manuscript.

Ancillary