|Home | About | Journals | Submit | Contact Us | Français|
Gonorrhea is the second most commonly reported infectious disease in the United States, and incidence has been increasing in recent years. Antibiotic resistance among clinical isolates has reached a critical point at which the CDC currently recommends only a single class of antibiotic for treatment. These developments have hastened the search for a vaccine to protect against gonococcal infections. Vaccine efforts have been thwarted by the ability of the gonococcus to antigenically vary most surface structures. The transferrin-iron transport system is not subject to high frequency phase or antigenic variation and is expressed by all pathogenic Neisseria. Vaccine formulations comprised of epitopes of the transferrin-binding proteins complexed with inactivated cholera toxin generated antibodies with potentially protective characteristics. These antigens, and others predicted from genome sequence data, could be developed into a vaccine that protects against neisserial infections.
Gonorrhea, caused by Neisseria gonorrhoeae, is the second most commonly reported infectious disease in the United States. In 2006, there were 358,366 reported cases according to the CDC , which represents an increase of 5.5% from the previous year. Rates of disease have been increasing at an alarming rate in the Western and Southeastern United States; in these regions, incidence increased 2.9 and 12.3%, respectively, between 2005 and 2006. Gonorrhea is a world-wide problem, impacting both industrialized and developing countries. In 1999, the WHO estimated the global prevalence of gonorrhea at 62.4 million cases . Even this extraordinary number likely represents an underestimate as the disease is under-reported and often asymptomatic.
Localized gonococcal infections impact both males and females . In males, the infection is usually symptomatic, with hallmark manifestations including a purulent urethral discharge and dysuria. If untreated, the infection can ascend resulting in epididymitis. In females, up to 50% of infections may be asymptomatic [4–6], thus prolonging the time to diagnosis and treatment. This population also serves as an important reservoir for the disease. If symptoms are present, the most common is a purulent cervical discharge; however this is often minimal and/or goes unnoticed. As a result, ascending infection, and the consequent morbidity, is disproportionately suffered by the female population; these infections include pelvic inflammatory disease, salpingitis and ectopic pregnancy. Ultimately, untreated infections in both women and men can lead to infertility. Disseminated gonococcal infections can also be a downstream consequence of asymptomatic infection; as one might expect, this manifestation is most commonly experienced by females. Symptoms include fever, rash and septic arthritis. Finally, perinatal infections also occur, which can result in conjunctivitis in newborns. If untreated this can lead to blindness; however this rarely occurs in developed countries due to the mandatory administration of ophthalmic antibiotics following a vaginal childbirth.
While gonococcal infections are treatable with antibiotics, a dramatic increase in the incidence of antibiotic resistance has occurred in the recent past [7, 8]. Penicillin resistance emerged in the 1970’s and resistance to tetracycline emerged approximately 10 years later. This led to the abandonment of these antibiotics as recommended treatment regimens. While a decrease in the incidence of these resistances has occurred in the absence of selection, in 2004 16% of strains tested were resistant to one or both of these antimicrobial agents. Until recently, the fluoroquinolones were recommended therapy for treatment of gonococcal infections. However, in 2006 resistance to this class of antibiotic had risen to 13.8%, compared to 9.4% the year before. This increase led the CDC to announce in 2007 that fluoroquinolones were no longer the recommended therapy . This development is of great concern as it restricts the treatment of gonococcal infections to a single class of antimicrobial agents: the third generation cephalosporins. Clearly, new approaches are urgently needed to treat this common sexually transmitted infection (STI), which has profound health consequences. One promising new development in the treatment of gonococcal infections is the identification of novel topical agents with microbicidal activity against N. gonorrhoeae .
In addition to emerging antibiotic resistance, another concern is the role that STIs play in the transmissibility of HIV. HIV genomic RNA levels are higher in the seminal plasma and vaginal secretions of men and women with concurrent gonococcal infections [10, 11]. When the bacterial STI is treated with antibiotics, HIV RNA and virus titers decrease. Therefore, gonococcal infection coincident with HIV infection exposes contacts to higher viral doses, increasing the risk of acquiring the infection. This observation, along with the emergence of antibiotic resistances, has hastened the search for an effective, cross-protective and long-lasting gonococcal vaccine.
Gonococcal infections do not elicit protective immunity and there is no vaccine to prevent the disease. Many challenges have been identified in the decades-long search for a protective gonococcal vaccine. First, the gonococcal cell surface is extremely variable, being composed of protein and polysaccharide antigens that rapidly change in antigenic character. The list of antigens that are subject to high-frequency phase and antigenic variation includes pilin, opacity proteins (Opa), lipooligosaccharides (LOS), and several outer membrane iron transporters. Two basic mechanisms are employed by the Neisseriae to generate antigenic diversity. The pilin protein, which polymerizes to form the macromolecular pilus structure critical for adherence, is antigenically variable as a result of homologous recombination between an expression locus and any one of several storage or silent pil loci. Variant pilin proteins arise at a frequency of approximately 1/1000 per cell per generation, demonstrating the degree of pilin heterogeneity found within gonococcal populations . The second basic mechanism employed by the Neisseriae to generate antigenic diversity is slipped-strand mispairing, which occurs at the site of polymeric sequence tracts . Increases or decreases in the number of repeats result in changes in expression. In most loci, the repeats are located in the structural gene, in which case alteration in the number of repeats can lead to frame-shift mutations. For example, the N. gonorrhoeae genome contains up to 11 different Opa loci, each of which consists of a promoter and a set of repeats within the coding region. At any one time, 0–11 of these Opa genes may be expressed into full-length protein; however, in vitro the tendency is to express fewer and in vivo up to 5 Opa proteins are expressed simultaneously . The slipped-strand mispairing mechanism is also responsible for variable expression of some iron transport proteins (Fig. 1). The loci encoding LbpAB , HpuAB , and FetA  all contain polymeric repeat regions, which result in rapid, on-off switching of these gene products .
Another impediment to vaccine development efforts has been the capability of the gonococcus to block deposition of functional antibody by decoration of LOS with sialic acid. Specific epitopes of LOS can be modified by exogenously-supplied CMP-NANA, resulting in an LOS structure with unique physiochemical and size characteristics. This modification prevents the deposition of bactericidal antibodies directed at gonococcal porin . In addition, antibodies generated against a conserved antigen, RmpM, similarly prevent deposition of functional anti-porin antibodies . Thus, even antigens that are not susceptible to high-frequency variation can be rendered unavailable to immune factors by virtue of blocking antibody or sialylation of the cell surface.
Development of a protective anti-gonococcal vaccine also requires elicitation of mucosal immunity in the genital tract, where the infection first establishes a foothold. This is complicated by the absence of specialized inductive sites, which are present in other mucosal tissues (for example, Peyer's patches in the gut) and by the fact that this niche must be relatively tolerant of foreign antigens [20–22]. Studies in which antigens are delivered locally to the genital tract have, in large part, been less successful than those in which antigens are delivered to other mucosal sites, such as the nasal mucosa [23, 24]. Intranasal immunization has shown promise in the development of genital mucosal and systemic immune responses potentially effective against a variety of STI agents [25–27].
In search of a cross-protective gonococcal vaccine, we have focused on receptors that are necessary for nutrient acquisition. One such nutrient is iron, which is consistently identified as an essential factor for bacterial survival in vivo and in tissue culture systems [28–30]. Figure 1 shows the receptor systems with which N. gonorrhoeae is known to acquire iron, along with a group of incompletely-characterized, potential transporters (Tdfs, Fig. 1). The TbpAB protein system is employed to bind transferrin, and relieve it of iron, which is then transported through the outer membrane. Subsequently, iron is bound by FbpA in the periplasm, which relays the nutrient to a membrane permease (FbpBC) for entry into the cytoplasm. The FbpABC protein system is also required to shuttle iron passed into the periplasm from lactoferrin, which is received and stripped of iron by the surface exposed LbpAB proteins. As indicated above, the LbpAB system is phase variable due to a repeat element within the structural genes. Moreover, the LbpAB locus is largely deleted in approximately 50% of gonococcal strains tested . Thus, we have focused on the TbpAB system as a potential vaccine target due to its ubiquitous expression among isolates. In addition, this system is not subject to phase or antigenic variation like so many other outer membrane antigens. However, this system is repressed under high iron conditions, a situation likely experienced in vivo, as is true of other iron acquisition systems. We determined that expression of the TbpAB system was required to initiate signs and symptoms of urethritis in a male model of experimental infection . The strain employed in these studies was a naturally occurring isolate that lacked a functional LbpAB locus. When a functional LbpAB locus was reconstituted into a TbpAB mutant, the resulting strain was capable of eliciting a urethral infection, suggesting that the Lbp system can functionally compensate for the absence of the Tbp system . However, the fact remains that half of gonococcal isolates cannot express LbpAB, making these antigens poor vaccine targets.
Other possible iron transport systems to target for vaccine development include HpuAB and FetA (Fig. 1). The HpuAB system is necessary for gonococcal growth on hemoglobin as a sole iron source . Presumably there is a dedicated periplasmic binding protein dependent system for transporting heme into the cell; however, to date such a system has not been identified. The HpuAB system, like the LbpAB, is subject to phase variation and was selectively expressed by gonococcal variants isolated from women during the first half of the menstrual cycle . The final iron transport system characterized in N. gonorrhoeae consists of FetA and an associated ABC transport system (Fig. 1). FetA has been shown to bind to ferric enterobactin with weak affinity and to provide the ability to grow on ferric-enterobactin as a sole iron source . There are predicted members of a periplasmic binding protein dependent systems encoded downstream of FetA; however, these gene products have not been well characterized. Likewise, the benefits that FetA or the downstream genes provide in vivo are unknown.
While the Tbp proteins are expressed in vivo  and are necessary to initiate infection in males , native infections result in limited amounts of Tbp-specific antibody . This result concurs with those reported by others, cumulatively suggesting that gonococcal infections elicit only modest humoral and mucosal antibody responses [39–41]. Perhaps this is related to the Opa-mediated T-cell suppression mentioned above, and perhaps other gonococcal factors contribute as well . Nonetheless, native infections generate modest immune responses, a concept that seems to correlate well with the lack of protective immunity elicited by infection. Thus, our approach to vaccine development has been to identify Tbp-specific epitopes that are well conserved among strains and which, when combined with an appropriate adjuvant and immunization route, will generate a protective immune response where native infection does not.
We developed a topology model of TbpA (Fig. 2; [42, 43]) based upon its homology with other outer membrane TonB-dependent transporters such as FepA and FhuA, expressed by E. coli. This class of transporters requires TonB-derived energy to accumulate ligands in the periplasm of Gram-negative bacteria. Several TonB-dependent transporters have been crystallized [44–50], and their conserved structures include a β-barrel, which forms a pore through the outer membrane, and an amino-terminal plug or hatch region, which occludes the pore until ligand and TonB-derived energy are provided. We have focused our vaccine development efforts on surface-exposed regions of TbpA, which are likely to be involved in ligand binding and accessible to immune factors. These regions (Fig. 2) have been identified by a combination of computer analysis, epitope insertion, loop deletion, and antibody binding experiments [42, 43, 51]. We found that portions of loops 2, 3 5, 7 and 10 are indeed surface exposed in N. gonorrhoeae. Of these, loop 5 appears to be critical for transferrin binding as this region retains transferrin binding characteristics when expressed in recombinant form in E. coli. In addition, deletion of this region of the protein abrogated transferrin binding and transferrin-iron internalization capabilities by N. gonorrhoeae.
The second member of the transferrin-iron uptake system is TbpB, which is lipid modified and entirely surface exposed (Fig. 1 and Fig. 2). The presence of TbpB is not required for growth on transferrin, but TbpB makes iron internalization more efficient in short-term assays [52, 53]. No crystallized homologs of TbpB are available for homology modeling, thus hypothetical models lack specificity and detail (Fig. 2). TbpB is comprised of two similar lobes, each of which contains transferrin binding motifs . The amino-terminal lobe is responsible for the unique ability of TbpB to retain transferrin binding capabilities following SDS-PAGE . TbpB, unlike TbpA, preferentially binds holo-transferrin . The ability to selectively bind the iron-laden form of transferrin likely contributes to the role that TbpB plays in increasing the association rate with and dissociation rate from the gonococcal cell surface. In addition to its discriminatory function, TbpB could play a more active role in iron removal from transferrin.
We took advantage of the adjuvanticity of detoxified forms of cholera toxin in generating TbpAB antigens for vaccine studies. The B subunit of cholera toxin has shown promise as a mucosal adjuvant in a number of studies, particularly when delivered intranasally [23, 55]. We chemically conjugated recombinant, full-length TbpA or TbpB with cholera toxin B subunits. These antigens were administered to female BALB/c mice intranasally, after which we analyzed the resulting mucosal and systemic immune responses. Both conjugated proteins elicited vaginal and serum antibodies, although TbpB was much more immunogenic than TbpA. The serum antibodies elicited in animals immunized with either protein conjugate were bactericidal in the presence of human complement. However, sera from the TbpA-immunized animals were more cross-reactive with heterologous strains than sera from TbpB-immunized animals . This result is in agreement with those of Thomas et al. , who also found that recombinant TbpB alone generated a vigorous immune response, but the resulting antibodies only recognized the surface of the homologous strain. Thus, we conclude that the most effective and cross-reactive antibodies are generated by immunization with a combination of TbpA and TbpB conjugates and that the resulting antibodies generated against both proteins could have synergistic effects. A similar phenomenon has been described for outer membrane protein antigens from N. meningitidis 
In a second study , we produced hybrid proteins between the A2 subunit of cholera toxin and Tbp-specific peptides. The A2 subunit non-covalently interacts with the cholera toxin B subunits to generate GM1-ganglioside binding-competent pentamers. We generated one chimera between the A2 domain of cholera toxin and the amino terminal half of TbpB (Fig. 2). With the idea that TbpB would be more immunogenic and TbpA would generate cross-reactive antibodies, we created a second chimera which fused the amino terminus of TbpB with loop 2 of TbpA (Fig. 2) and the A2 cholera subunit. These chimeras were used as immunogens to vaccinate BALB/c mice. While intranasal immunization alone generated low levels of Tbp-specific antibodies, following an intraperitoneal boost, both mucosal and systemic antibodies were elicited against TbpB. Likewise detectable TbpA-specific antibodies were detectable in the serum after immunization with the L2-containing chimera. Interestingly, sera from the mice immunized with the double chimera were the most cross-reactive and were bactericidal against heterologous gonococcal strains. Furthermore, vaginal secretions from mice immunized with the double chimera were capable of interfering with transferrin-dependent growth of both the homologous strain, and a heterologous strain. Therefore, as with the previous study, the double chimera between L2 of TbpA and the amino terminus of TbpB generated a more biologically-functional antibody response, in which the TbpA- and TbpB-specific antibodies appeared to act in synergy. We are currently testing these and other Tbp-specific antigens to identify those that are protective in a mouse model of female genital tract infection .
While the transferrin receptor system is required to initiate infection in human males, expression of the Tbps is not necessary for gonococcal survival in the female mouse model . Other iron sources appear to be available in the female genital tract besides transferrin, lactoferrin, and hemoglobin. Thus, we sought to identify other potential iron transport systems encoded within the gonococcal genome. Three putative iron transporters were identified by Turner et al.  and we identified a fourth when searching the genome sequence of gonococcal strain FA1090. These four TonB-dependent transporters were named TdfF, TdfG, TdfH and TdfJ [62, 63]. While the ligands for these putative transporters (cumulatively labeled Tdfs in Fig. 1) have not yet been identified, they all share homology with outer membrane transporters for ferric-siderophores or heme. We insertionally inactivated each of these transporters and found that one, TdfF, was expressed and required for growth within human cervical epithelial cells . The inability of the tdfF mutant to grow within epithelial cells was overcome by addition of excess iron to the cell growth medium, suggesting that TdfF is responsible for iron acquisition within epithelial cells during iron restriction. Consistent with this hypothesis, we determined that TdfF expression is repressed under high iron conditions and induced under low iron conditions, as is the case with the other neisserial iron transport systems. Cumulatively, these results suggest that TdfF is important for in vivo survival of the gonococcus and therefore might prove to be a valuable vaccine antigen for future study. The role that the other Tdfs play in gonococcal growth and/or nutrient acquisition has not yet been defined; however, these outer membrane proteins could also serve as useful vaccine targets, even if their in vivo functions have not yet been elucidated.
Several antigens have been evaluated as possible vaccine antigens to protect against gonococcal disease, with varying degrees of success. Pili were early targets for vaccine development as they function in multiple aspects of gonococcal pathogenesis including attachment, host cell signaling and natural transformation. However, the high-frequency phase and antigenic variation deployed by this antigen limits the utility of such a vaccine. Recently, Hansen et al.  demonstrated that a surface exposed and variable epitope of pilin was critical for immunogenicity in mice. Without this epitope, the conserved regions of pilin alone were immunosilent but were not capable of actively suppressing immunity to other antigens . Thus, given the diversity of gonococcal pili and the fact that the level of diversity correlates with immunogenicity of this antigen, the goal of a cross-protective pilin-based vaccine seems unrealistic.
Another potentially tempting outer membrane protein for vaccine development is Opa. While variable expression of Opa proteins perhaps limits their utility in this regard, one could envision employing a cocktail of Opa proteins as vaccine immunogens. In vivo growth of the gonococcus selects for Opa-positivity; however, no specific Opa protein has been identified as crucial for survival in a human male challenge model of gonococcal disease . Moreover, recent evidence indicates that Opa protein binding to host cell receptors results in suppressed activation and proliferation of T-cells, which would have the undesirable effect of diminishing any acquired immune response . Lee et al.  have recently extended this observation to membrane vesicles containing Opa proteins; thus a vaccine approach based upon vesicles derived from any of the Neisseriae should be devoid of Opa proteins in order to generate the most effective immune response.
Porin proteins are abundant in the outer membrane and necessary to allow diffusion of small, hydrophilic molecules into the periplasm. N. gonorrhoeae expresses one of two porin alleles: PorB1A (previously known as P.1A) or PorB1B (previously known as P.1B). The former allele is associated with gonococcal isolates capable of causing disseminated infections; the latter allele is more frequently identified among localized isolates. Expression of either allele can result in serum resistance, although the mechanisms employed differ between porin alleles . The PorB protein is not subject to high frequency phase or antigenic variation and is essential for growth. Thus, this antigen would seem to be an ideal vaccine target. As indicated above, functional anti-PorB antibody deposition is blocked by LOS sialylation and by RmpM-specific antibodies. However, vaccination with preparations lacking RmpM could alleviate the latter problem. Sparling and co-workers tested a variety of immunization strategies to discern which presentation of PorB would result in functional PorB-specific antibodies . A DNA-based delivery system generated the lowest level of antibody, but a booster with recombinant protein increased this level modestly. Immunization with recombinant PorB generated significantly more antibody, but the response was strain specific, generating antibodies that reacted only with those strains expressing the same allele as was present in the vaccine. Intranasal administration of PorB-containing outer membrane vesicles generated the most functional and potentially protective responses. Specific and cross-reactive antibodies were detected in the serum and mucosal secretions; serum antibodies were bactericidal in the presence of human complement . These results suggest that vesicles, which retain outer membrane proteins in a native conformation, could be the preferred means of antigen delivery, particularly via the intranasal route. However, this enthusiasm is tempered to some degree by the knowledge that these vesicles also contain LOS, which can be immunostimulatory but also quite toxic. As indicated above, vesicles are also likely to contain one or more Opa proteins which could suppress the adaptive immune response. However, this effect would be masked in mouse models and mouse immunization studies as the Opa-receptor interaction is host restricted .
The term "reverse vaccinology", used by Rappuoli and co-workers [69, 70] refers to the process of identifying promising vaccine antigens by application of bioinformatic approaches to genomic sequence data. Employing this approach, several vaccine candidates have been identified in Neisseria meningitidis, which cumulatively have the potential to serve as a universal vaccine to prevent meningococcal disease caused by serogroup B . Similarly, Comanducci et al. have identified a novel surface adhesin of N. meningitidis, which also has vaccine potential . In the gonococcus, similar genomic searches have identified a peptidyl-prolyl isomerase that is involved in persistence in macrophages , and more recently a potential vaccine antigen with similarity to the E. coli OmpA protein . This N. gonorrhoeae OmpA-like protein was found to facilitate interaction with human cervical epithelial cells and to be important for survival in the mouse female genital tract . Presumably this sampling of antigens represents the tip of the iceberg, and as more genome sequences become available from other gonococcal strains, the opportunities for vaccine development will expand dramatically.
N. gonorrhoeae causes the very common sexually transmitted infection, gonorrhea, which has become increasingly difficult to treat. A single class of antimicrobial agent, the third generation cephalosporins, is now the sole recommended course of treatment for this infection due to increased resistance to other agents. The impact of the disease is disproportionately experienced by women, given the incidence of asymptomatic and ascending infections among this group. Women also serve as an important reservoir for the disease. In addition, HIV transmission is facilitated by concomitant N. gonorrhoeae infection in both men and women. Thus, identification of preventative measures to combat gonococcal infections is important and the subject of ongoing research. Protective vaccines to prevent gonorrhea have been sought for decades; however, antigenic variation by and poor immunogenicity of many potential targets has hampered these efforts. While there is a protective, capsule-based vaccine against the very similar Neisseria meningitidis, this option is not available for vaccine development against N. gonorrhoeae as the latter lacks a polysaccharide capsule. Moreover, the meningococcal vaccine is not protective against all serogroups. Most notably, the serogroup B capsule is not immunogenic but is possessed by N. meningitidis strains endemic in many industrialized nations. Thus, ongoing efforts are aimed at improving the meningococcal vaccine and developing a gonococcal vaccine.
Iron acquisition is critical for the survival of most bacterial pathogens; as such the surface exposed components that contribute to this process could represent conserved, vulnerable targets for vaccine development. We have focused our efforts on the transferrin-iron acquisition system since it is ubiquitously expressed by the pathogenic Neisseriae, required for infection, and not subject to high-frequency variation. Gonococcal and meningococcal transferrin binding proteins are similar in sequence [52, 75], suggesting that a vaccine comprised of these antigens could potentially protect against both gonococcal and meningococcal infections. Antibodies generated against meningococcal transferrin binding proteins have been shown to be protective in animal models of infection [76, 77]. Subunit vaccines comprised of the gonococcal transferrin binding proteins conjugated to or fused with the B subunit of cholera toxin elicited specific antibodies, both systemically and mucosally. These antibodies had biologically-important properties, including complement-dependent bactericidal activities and growth inhibitory functions. Utilizing both TbpA and TbpB in antigen preparations resulted in increased immunogenicity and cross-reactivity. Genomic sequencing and bioinformatic approaches have identified other, as yet uncharacterized, potential vaccine antigens. Some of these new gene products are important for various aspects of gonococcal infection or pathogenesis, interference with which could result in amelioration or prevention of disease.
Gonorrhea rates in the United States have declined from historic highs, but incidence has risen again in recent years. Estimates of unreported or unrecognized disease rival the reported incidence numbers. Added to the dramatic rise in antibiotic resistance among gonococcal isolates, this represents a public health crisis. Efforts to develop a vaccine to prevent this common infection have so far been unsuccessful. However, several candidate antigens have shown promise with regard to immunogenicity and elicitation of cross-reactive, biologically-relevant responses. A successful gonococcal vaccine is expected to be composed of one or several surface-exposed antigens. Immunization by the mucosal route may be the preferred method as this would elicit local genital tract antibodies, in addition to systemic responses that could prevent dissemination. Given that the gonococcus is capable of colonizing both males and females in addition to multiple niches within a single individual, it seems likely that distinct surface antigens are employed in these various environments. A successful vaccine would therefore be expected to consist of factors that are expressed and vulnerable to attack in each of these niches, which could necessitate the incorporation of antigenically and functionally diverse antigens into a cocktail, subunit-based vaccine. Appropriate adjuvants will have to be identified or developed, which increase immunogenicity of conserved epitopes and preserve protein conformation in the context of a mucosally-delivered antigen.
C.N. Cornelissen's research is supported by Public Health Service grants AI47141 and AI065555 from the National Institute of Allergy and Infectious Diseases at NIH. The author also acknowledges the current and past members of her laboratory and Ann Jerse for critical reading of this manuscript.