REVIEW ARTICLE: Antimicrobial Polypeptides are Key Anti-HIV-1 Effector Molecules of Cervicovaginal Host Defense
Alexander M. Cole, Department of Molecular Biology and Microbiology, Burnett School of Biomedical Sciences, 4000 Central Florida Blvd., Bldg 20, Rm 236, Orlando, FL 32816, USA.
Mucosal surfaces of the cervix and vagina are portals for heterosexual transmission of human immunodeficiency virus type 1 (HIV-1) and, therefore, play a fundamental role in the pathogenesis of primary infection. Cationic antimicrobial polypeptides including defensins are the principal effector molecules of mucosal innate immunity against microbes and viruses such as HIV. In cervicovaginal secretions, antimicrobial polypeptides constitute the majority of the intrinsic anti-HIV-1 activity, synergism between cationic polypeptides is complex, and full anti-HIV-1 activity involves the complete complement of cationic polypeptides. Periods in which cationic antimicrobial polypeptide expression is reduced are likely associated with increased susceptibility to HIV-1 infection. This review provides an overview of the role of cationic antimicrobial polypeptides in innate cervicovaginal anti-HIV-1 host defense, and discusses how hormones and bacterial infections can regulate their expression. Emphasis is placed on the θ-defensin (retrocyclin) class of anti-HIV-1 peptides and their potential for development as topical microbicides to prevent HIV-1 transmission.
Approximately, 40 million people are currently living with human immunodeficiency virus (HIV) worldwide according to the 2006 World Health Organization estimates.1 There has been a dramatic increase in the global spread of HIV type 1 (HIV-1) especially via the heterosexual mode of transmission. At present, women represent 50% of infected individuals worldwide and nearly 60% in Africa. The natural sexual transmission of HIV-1 occurs through mucosal surfaces such as the vaginal mucosa.2 Vaginal subepithelial stromal tissues are densely populated with dendritic cells (DC), macrophages, and T lymphocytes that express both CD4 and the HIV-1 coreceptors, CXCR4 and CCR5.2–4 The mechanisms by which HIV-1 journeys across the vaginal mucosa are not completely understood but likely involve the epithelial cells.5 Upon reaching the lamina propria, HIV-1 can either directly infect T lymphocytes or macrophages, or adhere to or infect DC, whose subsequent passage to lymph nodes initiates viral replication at these sites.6,7 While a majority of studies have focused on mucosal adaptive immunity of the female reproductive tract, only recently the attention has been focused on the innate immune factors. Mounting evidence suggests that the vaginal and cervical epithelia are more than simple physical barriers to protect against invading pathogens.8–10 On the contrary, these surfaces and their overlying fluids are replete with antimicrobial peptides and proteins that act as effectors of innate host defense.
Antimicrobial peptides and proteins are principal effector molecules of innate mucosal host defense
The antimicrobial properties of mucosal tissues are largely because of the contributions of bioactive polypeptides. In 1922, Alexander Fleming attributed the antimicrobial properties of human nasal secretions to an enzyme he named ‘lysozyme’.11 Since the discovery of lysozyme, many other antimicrobial components of mucosal secretions have been identified (reviewed in 12). Lysozyme and lactoferrin, two polypeptides expressed at appreciable concentrations by many mucosae, are stored in and secreted from serous cells in submucosal glands and are also released from neutrophils that migrate to inflamed tissues. The serine protease inhibitor secretory leukocyte protease inhibitor (SLPI) is also present at potentially antimicrobial concentrations.13 Other antimicrobial (poly)peptides that are present in most mucosal secretions include multiple members of the defensin family, the cathelicidin LL-37 and secretory phospholipase A2.14–16 Most antimicrobial peptides and proteins are broad-spectrum microbicides that target bacteria, fungi, and viruses.17–19 Some are enzymes that disrupt essential microbial structures, as exemplified by the muramidase activity of lysozyme and its ability to hydrolyze peptidoglycan, a key structural component of bacterial cell walls. Others bind essential nutrients, denying them to microbes, as is typified by the host defense strategy of lactoferrin. Still others, especially small antimicrobial proteins and peptides, act by disrupting microbial or viral membranes.20–25 Although antimicrobial peptides can be evolutionarily and structurally diverse, common properties typically include spatial separation of polar and non-polar residues (amphipathicity) and a net positive charge at physiological pH (cationicity). These attributes assist peptide binding and insertion into anionic microbial membranes, or prevent membrane–membrane interactions such as viral fusion.
Defensins are the most well-studied family of cationic antimicrobial peptides, and are expressed by the epithelial cells and leukocytes of mammals and birds.17,26–28 Three defensin subfamilies exist in vertebrates: α-defensins, β-defensins, and θ-defensins. All contain six cysteines, have principally β-sheet structures that are stabilized by a tri-disulfide motif unique to each subfamily of defensin,29 and are derived from an ancestral gene that existed before the divergence of reptiles and birds.30 Human β-defensin-1 (HBD-1) is an antimicrobial peptide constitutively expressed at low levels at many mucosal sites, whereas a structurally similar peptide, HBD-2, is predominantly induced at sites of inflammation.31,32 The α-defensins human neutrophil peptides 1–3 (HNP 1–3) are generally degranulation byproducts of neutrophils that extravasate into mucosal tissues during infection.33θ-defensins were first isolated from the leukocytes and bone marrow of the rhesus monkey, Macacca mulatta.34–36 These unique, arginine-rich, circular 18-residue peptides arose from two precursor peptides, each of which contributes nine-residues and three cysteines to the mature θ-defensin.34–36 They too have a β-sheet structure37 and represent the first truly macrocyclic peptides of vertebrate origin. While the mechanism of fusing the nine-residue peptides is not understood, it is believed that their intracellular cyclization occurs via two post-translational head-to-tail ligations. These splices join the backbones of the two nine-residue precursors to make the resulting peptide circular, and oxidation of the six cysteines form an internal tri-disulfide ladder. Whereas α- and β-defensin peptides are produced by both human and non-human primates,34 humans are not believed to produce mature θ-defensin peptides because of a premature termination codon that would inhibit its complete translation.38
Human cervicovaginal fluid contains a number of anti-HIV-1 peptides and proteins
A layer of mucosal fluid covers the cervical and vaginal epithelia, and is composed of secretions from the cervical vestibular glands, endometrial and oviductal fluids, and plasma transudate.39 The vagina serves as a host for numerous commensal microorganisms, which release organic acids and their own bioactive peptides to kill pathogenic invaders. The vaginal epithelial cells, cervical glands, and neutrophils contribute the majority of antimicrobial peptides to the milieu of the vaginal fluid, including calprotectin, lysozyme, lactoferrin, SLPI, HNP-1, -2, and -3, and HBDs.8 Until recently,40 evidence for the biologic role of antimicrobial polypeptides in vaginal anti-HIV-1 host defense had been largely circumstantial. Lysozyme and lactoferrin have been shown to inhibit the HIV-1 infection in vitro by preventing the adsorption and penetration of the virus.41–43 Rabbit, rat, and guinea pig defensins can inhibit HIV-1-induced cytopathogenicity of a CD4+ human T-cell line.44 Several studies extended these findings to confirm that human α-defensins (HD) are active against HIV-1,45 and that the mechanism of action involved both the inhibition of HIV-1 replication through intracellular interference with protein kinase C (PKC) activity and direct inactivation of HIV-1 virions.45,46β-defensins were also shown to inhibit HIV-1 infection. Quinones-Mateu and colleagues47 revealed that HIV-1 induced β-defensin expression in human oral epithelial cells. Interestingly, expression of β-defensins down-modulated the expression of CXCR4, a chemokine coreceptor important for entry of X4 strains of HIV-1. Several laboratories have reported an inhibitory activity of SLPI against HIV-1 (reviewed in 48). SLPI (>100 ng/mL) was shown to block HIV-1 internalization by inhibition of viral entry or capsid uncoating.49,50 Conversely, studies from another group revealed that while saliva inhibited HIV-1 infection, recombinant SLPI was not active against HIV-1 in vitro,51 leaving the activity of SLPI against HIV-1 in question. However, as several antimicrobial peptides and proteins are known to synergize, the anti-HIV-1 activity of SLPI might be realized in concert with the other antimicrobial peptides and proteins of cervicovaginal secretions. Indeed, the sum total of these and other antimicrobial peptides and proteins synergize to contribute to the innate host defense of the vagina.40
Hormonal regulation of antimicrobial peptides and other mediators of vaginal innate immunity
Levels of sex hormones fluctuate during the menstrual cycle, influencing functional changes that prepare the genital tract for conception and pregnancy. Estradiol is the predominant sex hormone in the proliferative (follicular) phase of the menstrual cycle, whereas progesterone peaks and estradiol remains elevated during the secretory (luteal) phase. Progesterone thins the vaginal epithelium, increases vaginal pH, and reduces the amount of cervical mucus, whereas estrogen has the opposite effect. Many protective functions of the female genital tract, including antimicrobial polypeptide expression, appear to be under hormonal control. For example, the concentration of HD-5 produced by endometrial epithelial cells is highest during the secretory and post-ovulatory stage of the menstrual cycle.52 SLPI levels also vary throughout the menstrual cycle in the cervical mucus but not in serum,53 and the concentrations of α-defensins and β-defensins in cervicovaginal lavage (CVL) from normal patients can vary by up to 50-fold, suggesting hormonal regulation of antimicrobial peptides.8 In the endometrium, levels of HBDs and SLPI are under cycle control, and levels of SLPI are highest during the secretory phase.54–56 Moreover, the expression of SLPI from endometrial cells has been shown to be upregulated by progesterone in the rhesus macaque57 and in humans.55 Intravaginal estrogen treatments given to post-menopausal women increased the level of SLPI in their vaginal secretions.58
Other innate immune processes are influenced by female sex hormones, many of which have an indirect effect on the expression of antimicrobial polypeptides. High doses of 17β-estradiol enhanced LPS-induced IL-1β expression from uterine cells through the binding of estrogen receptor, and resulted in increased secretion of HBD-2.59 17β-Estradiol also increased the expression of the chemokine CCL20/MIP3α with a concomitant inhibition of tumor necrosis factor-α (TNF-α) production.60 It is notable that CCL20/MIP3α and HBD-2 are structurally similar and both act as ligands for CCR6, serving as a link between innate and adaptive immunity by chemoattracting immature DC and lymphocytes to sites of infection.61,62 However, to date, there have been a paucity of studies linking hormone levels with innate immune function. Hormonal changes associated with the phase of a woman’s menstrual cycle impact viral dynamics and shedding of HIV from the genital tract. It would therefore be important to understand how hormones such as 17β-estradiol and progesterone regulate the expression of anti-HIV-1 polypeptides in the female genital tract, and how levels of anti-HIV-1 polypeptides correlate with HIV-1 susceptibility.
Bacterial vaginosis causes dysfunction in innate immunity of the vaginal mucosa
Bacterial vaginosis (BV) is the most prevalent polymicrobial condition of women of reproductive age, and is often associated with increased risk of sexually transmitted diseases, including HIV-1.63–66 BV is caused by overgrowth of either non-resident bacteria in the vagina, or resident bacteria that under normal conditions are underrepresented members of the endogenous vaginal microbiota. Most women with BV have either a markedly reduced concentration of vaginal lactobacilli or are completely devoid of lactobacilli,67 both of which result in an elevation in vaginal fluid pH. It is unclear whether vaginal pH contributes to the initiation of BV or becomes elevated as BV becomes established. Additionally, while nearly a quarter of women with BV are populated with H2O2-producing lactobacilli,67 the number of these lactobacilli or amount of H2O2 liberated are presumably insufficient to prevent colonization by non-lactobacilli microbes. These and other findings underscore the importance of maintaining concentrations of vaginal lactobacilli in equilibrium. Likewise, if homeostasis of the cervicovaginal mucosa is altered, for example as a result of BV, perturbation of the resident vaginal flora leads to an increase in pH, an increase in non-resident or underrepresented microbes, and the ascension of pathogens into the sterile upper genital tract.
Several recent studies have concluded that BV greatly increases a woman’s susceptibility to HIV,63,64,68–71 yet only recently have studies emerged which implicate innate host defense dysregulation during BV. A paper by Valore and colleagues72 provided direct evidence that CVL fluid from women with BV was deficient in antimicrobial polypeptides and antibacterial activity as compared with healthy women or women with another infectious condition, vulvovaginal candidiasis (VVC). Effective treatment of BV normalized the levels of antimicrobial peptides, suggesting that low levels of antimicrobial polypeptides were a direct result of BV. Similarly, Draper and colleagues reported that SLPI was decreased in women with BV and sexually transmitted infections (STI), suggesting that decreased SLPI levels may increase susceptibility to HIV-1 infection.73 These collective studies are in stark contrast to VVC, in which the levels of antimicrobial polypeptides remained normal,72 and the evidence implicating VVC in increasing HIV susceptibility was less compelling.74 The relationship between BV and susceptibility to HIV-1 infection is likely complex, and in the future it will be important to determine how BV influences the normal anti-HIV polypeptide profile in the vaginal mucosa, and how reduced expression of anti-HIV polypeptides might be correlated with a woman’s susceptibility to HIV.
Despite the expression of cationic anti-HIV-1 polypeptides in the cervicovaginal mucosa, women are still infected by HIV-1. Very little is understood about the presence or absence of antimicrobial and antiviral polypeptides in the vagina and other mucosal sites of viral transmission, and perhaps the proteins and peptides are not present in the ideal place or at optimal concentrations. With regards to innate immune genes, copy polymorphism is unique to defensins,75 whose copy number can vary widely between individuals.76β-defensin genes can range from 2 to 8 per diploid genome, while α-defensins can range from 5 to 14 copies per genome, and the level of expressed peptide (and conceivably the level of anti-HIV-1 activity) is directly correlated to gene copy number. Antimicrobial polypeptides might be incomplete inhibitors of HIV-1 infection, able only to reduce but not eliminate the chance of infection and progression of disease. It is also plausible that endogenous human antimicrobial peptides in the vagina were once effective against lentiviruses such as HIV-1, but that the virus’ high mutation rate77,78 simply outpaced our innate host defense. A speculative yet appealing theory is considered in the next section, namely, that most humans may have evolutionarily ‘lost’ a host defense factor that once provided protection against HIV-1.
Retrocyclin, an ancestral human θ-defensin, can potently prevent HIV-1 infection
Human bone marrow expresses two different mRNA transcripts homologous to rhesus monkey circular θ-defensins.38 However, the genes for human θ-defensins code for an in-frame termination (stop) codon in the signal peptide sequence, theoretically preventing its translation.38 Humans can express θ-defensin mRNA in human bone marrow, spleen, thymus, testis, and skeletal muscle,79 and in cervicovaginal epithelia (unpublished results). However, to date, we have not been successful in isolating θ-defensin peptides from humans. Instead, we recreated these circular 18-residue peptides by solid phase synthesis and named them retrocyclins (‘retro-’ from the Latin for back or backward, and ‘-cyclin’ meaning circular). Importantly, retrocyclins could potently protect primary T cells from in vitro infection by both X4 and R5 strains of HIV-138 and were much more active in vitro than human α- and β-defensins. We and others have also revealed that retrocyclin analogs can remarkably protect CD4+ T cells against infection by clinical isolates of HIV-1 from a number of different clades80 and induce very little resistance in HIV-1.81 Retrocyclins are highly effective in the presence of human vaginal fluid,82 are not proinflammatory,38,82 and can prevent organotypic cervicovaginal tissues from HIV-1 BaL infection.82 Moreover, retrocyclins exhibited little to no hemolytic activity nor cytotoxicity against H9 cells and ME-180 cervical carcinoma cells at up to 500 μg/mL, a concentration that far exceeds the amount required for complete protection against most strains of HIV-1 (2–10 μg/mL).
Aside from the single base pair change that contributed the premature stop codon, the genes for retrocyclin otherwise remain extraordinarily intact,79 suggesting a contemporary role for these genes. It is possible that individuals, who are inherently more resistant to HIV-1 transmission, might also express retrocyclin peptides. Whether certain human populations have a corrected premature stop codon, or whose translation machinery can read through the premature stop codon, is not yet known. It would therefore be imperative to undertake studies to genetically test individuals who are naturally resistant to HIV-1 infection for the presence of intact retrocyclin genes and expression of retrocyclin peptides. While we may never realize whether human susceptibility to HIV-1 infection is a direct result of the evolutionary loss of retrocyclin peptides, retrocyclin-based microbicides that can be applied topically to augment the natural antiretroviral activity of human cervicovaginal fluid would be important in reducing HIV-1 transmission.
The present work has been supported by grants from the National Institutes of Health: AI052017, AI065430, and AI060753 (to A.M.C.).