INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Infections due to Chlamydia trachomatis impose a significant medical and economical burden throughout the world (12, 38, 54). In areas with poor hygienic conditions, infections with this bacterium result in trachoma, the most common cause of preventable blindness in the world (12, 38). In addition, throughout the world this pathogen is one the leading causes of sexually transmitted diseases (12, 38, 42, 54). The acute genital infection generally subsides, but in some patients it may result in long-term sequelae. In females, among the long-term sequelae infertility, ectopic pregnancy, and chronic abdominal pain are relatively frequent (55, 56). Effective antimicrobial therapy is available to treat infected patients. This approach, however, has not been effective at controlling infection by this organism. There are several reasons for this disappointing outcome (25, 42). On one hand, a majority of the genital infections in women are asymptomatic, and thus patients do not seek treatment. In addition, in patients that are symptomatic, by the time the antimicrobial therapy is implemented permanent damage may have already occurred. Routine screening, followed by adequate therapy, in certain populations at high risk can decrease the overall prevalence of the disease, but this approach is temporary and can be practically applied only to limited groups. Thus, if we want to control the diseases produced by this pathogen the most effective approach will be to engineer a vaccine.

Attempts to implement a vaccine using whole organisms for C. trachomatis infections have a long history (8, 13, 41, 53). Four or five decades ago most efforts focused on developing a vaccine against trachoma using viable or inactivated whole organisms (13, 41, 53). More recently, adoptive immunization with dendritic cells pulsed ex vivo with inactivated whole chlamydial organisms has also been used as an experimental model to define the parameters of protection (19, 45). Several conclusions resulted from the earlier studies (13). The vaccine trials showed that an effective, although short-lived, protection could be achieved if the patients were immunized with sufficient antigen of the serovar that they were going to be subsequently exposed to. On the other hand, patients inadequately immunized suffered a hypersensitivity reaction when reexposed to C. trachomatis. This last observation has been interpreted as possibly an immune reaction to a chlamydial component, maybe the 60-kDa heat shock protein, that cross-reacts with some human antigen(s) (20). This possibility gave impetus for the development of a subunit chlamydial vaccine.

A total of 15 human serovars of C. trachomatis have been described (52). The trachoma biovar includes 12 serovars, A through K plus Ba, and the lymphogranuloma venereum group consists of the L1, L2, and L3 serovars. The A through C serovars have been isolated mostly from trachoma cases, while the D through K serovars mainly cause genital infections. The mouse pneumonitis (MoPn) biovar is the only isolate of C. trachomatis so far recovered from mice (24). The grouping of C. trachomatis isolates into distinct serovars appears to be based mainly on the differences in the amino acid sequence of the variable domains (VD) of the major outer membrane protein (MOMP) of this organism (10, 43). The MOMP is surface exposed, accounts for over 60% of the mass of the outer membrane, and has been shown to induce neutralizing antibodies (5, 41, 53). These characteristics have made the MOMP the preferred candidate for a subunit vaccine (10, 41, 53). In our laboratory we have utilized the C. trachomatis MoPn model, since this serovar, as a natural murine pathogen, can infect the genital tract of mice and produce a pathology that closely mimics that found in humans (27, 46). Here we show that vaccination with a highly purified preparation of the MOMP that was refolded in an attempt to reestablish the conformation of the native epitopes was able to protect mice against a genital challenge.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Organisms. The C. trachomatis MoPn biovar (strain Nigg II) was purchased from the American Type Culture Collection (Manassas, Va.) and grown in HeLa 229 cells (27). The elementary bodies (EB) were prepared as described by Caldwell et al. (5). The organisms were frozen at 70°C in SPG (0.2 M sucrose, 20 mM sodium phosphate [pH 7.4], and 5 mM glutamic acid).

Preparation of the C. trachomatis MoPn MOMP. C. trachomatis MoPn grown in HeLa 229 cells was washed with 10 mM phosphate-buffered saline (PBS), pH 7.4, and centrifuged and the pellet was treated with DNase (31). Following centrifugation the pellet was resuspended in 0.2 M phosphate buffer, pH 5.5, containing 0.001 M concentrations (each) of EDTA and phenylmethylsulfonyl fluoride (PMSF) and was extracted with 3-[(3-cholamidopropyl)-dimethylammonio]-1-propane sulfonate (CHAPS; Calbiochem-Novabiochem Corp., San Diego, Calif.) and Zwittergent 3-14 (Z3-14; Calbiochem-Novabiochem Corp.) as previously described (31). The MOMP was recovered in the supernatant and purified using a 1-cm by 35-cm hydroxylapatite column (5). Fractions containing the MOMP were pooled, run on a 5 to 20% gradient sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and stained with silver and Coomassie blue (Bio-Rad, Hercules, Calif.). In addition, an N-terminal amino acid analysis was performed on the purified MOMP by the core facilities at the University of California, Irvine.

Animal immunization and challenge protocols. Seven- to 8-week-old BALB/c female mice (H-2^d) were obtained from Simonsen Laboratories (Gilroy, Calif.). All animal protocols were approved by the University of California, Irvine, Animal Care and Use Committee.

Immunoassays. Vaginal secretions and blood were collected before each immunization, and antibody titers were determined using an enzyme-linked immunosorbent assay (ELISA) as follows (27). Multiwell plates (96 flat-bottom wells; Corning Glass Works, Corning, N.Y.) were coated overnight with 1 µg of purified C. trachomatis MoPn EB per well. Serial dilutions of the serum samples were added to each well, and the plates were incubated for 2 h at 37°C. The antigen-antibody reaction was detected by adding horseradish peroxidase (HRP)-conjugated goat anti-mouse antibodies. The following class or subclass-specific antibodies were used: immunoglobulin G (IgG), IgG1, IgG2a, IgG2b, IgG3, IgA, and IgM (Southern Biotechnology Associates, Inc., Birmingham, Ala.). The reaction was measured at 405 nm using an ELISA reader (Bio-Rad Corp.). The substrate, 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonic) acid (ABTS) was used for color development.

Fertility studies. Six weeks after the intrabursal challenge, groups of four female mice were housed with a proven breeder male mouse for a maximum of 18 days and the pregnancies were assessed by measuring the weight of each mouse (27). Animals that had gained 5 to 10 g of weight by or before 18 days postmating were considered pregnant and were euthanized, and the number of embryos in each uterine horn was counted. After the first mating the female mice that did not gain weight were mated a second time with male mice that had successfully mated with another group of female mice and were monitored as described previously (27). All the animals that had not gained weight were euthanatized 25 days from the start of the second mating. The number of embryos in each uterine horn was counted when the mice were killed.

Statistics. Statistical analyses were performed with the Statview software package on a Macintosh computer. The two-tailed unpaired Student's t test, the Mann-Whitney's U test, and the Fisher's exact test were employed to determine the significance of the differences between groups.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of the vaccine candidate preparations. The purity of the MOMP preparation, following the hydroxylapatite column, was first assessed by gel electrophoresis and amino acid sequencing. As shown in Fig. 1, only one band, corresponding to the C. trachomatis MoPn MOMP, was detected on a gradient gel stained with silver. By N-terminal amino acid analysis only the expected sequence, L-P-V-G-N-P, of the mature MoPn MOMP was obtained, indicating a greater than 99% purity of the protein. Figure 2A shows a Coomassie blue-stained SDS-PAGE of the folded C. trachomatis MOMP that was vortexed (lane 2) or sonicated (lane 5) and run under nonreducing conditions. The folded MOMP was also boiled for 10 min (lane 6), boiled for 10 min in the presence of 30 mM dithiothreitol (DTT) (lane 3), or treated with 30 mM DTT but not boiled (lane 4). Sonication and boiling of the folded MOMP under reducing or nonreducing conditions resulted in a predominant single band with an apparent molecular size of approximately 42 kDa. Under nonreducing conditions or reducing conditions but not boiling there is a prominent band of 62 kDa and faint bands with approximate molecular sizes of 42 and 132 kDa. The 62- and 132-kDa bands most likely represent homopolymers of the MOMP not linked by disulfide bonds, since boiling, without a reducing agent, resulted in a single band of 42 kDa. Figure 2B shows a Western blot of the C. trachomatis MoPn MOMP under boiled-reduced and nonboiled-nonreduced conditions probed with serum from mice immunized with the vortexed and sonicated MOMP preparations. Under reduced conditions only one band is detected with a molecular size of 42 kDa, while in addition to the 42-kDa band there is a band at 62 kDa, with faint bands above and below it under nonreduced conditions. The lack of an identifiable band at 132 kDa may be due to the poor transfer of large-molecular-size components during blotting. Alternatively, a paucity of antibodies was elicited by this antigen or epitopes to which antibodies were formed were partially blocked by the conformation of the 132-kDa molecules. Western blots probed with monoclonal antibody MoPn-40 to a linear epitope on VD1 appeared similar to those reacted with serum from mice immunized with vortexed or sonicated MOMP. A faint band corresponding to the molecular size of the lipopolysaccharide (LPS) could be observed when the Western blots of the MOMP preparation were performed using low dilutions of the immune serum (data not shown).

View larger version (74K): [in this window] [in a new window]   FIG. 1.   Silver-stained SDS-PAGE (5 to 20%) of the purified C. trachomatis MoPn MOMP. Lane, 1 molecular size standards (in kilodaltons). Lanes 2 through 6, preparation of reduced MOMP. Lane 2, 8 µg; lane 3, 4 µg; lane 4, 2 µg; lane 5, 1 µg; lane 6, 0.5 µg.

View larger version (67K): [in this window] [in a new window]   FIG. 2.   (A) SDS-PAGE of the purified C. trachomatis MoPn MOMP. Lane 1, molecular size standards. Lane 2, preparation of folded MOMP electrophoresed under vortexed-nonreducing conditions; lane 5, preparation of folded MOMP electrophoresed under sonicated-nonreducing conditions. Lanes 3, 4, and 6, folded MOMP electrophoresed after it was boiled for 10 min in the presence of 30 mM DTT (lane 3), treated with 30 mM DTT but not boiled (lane 4), or boiled for 10 min (lane 6). (B)Lane 1, molecular size standards. Lanes 2 through 7, C. trachomatis MoPn folded MOMP boiled in the presence of 30 mM DTT (lanes 2, 4, and 6) or not boiled and not reduced (lanes 3, 5, and 7). The blots were probed with serum from mice immunized with the vortexed MOMP preparation (lanes 2 and 3), the sonicated MOMP (lanes 4 and 5), or with the monoclonal antibody MoPn-40 that recognizes the VD1 of the MOMP (lanes 6 and 7).

Immune response following vaccination with the MOMP. An antibody response was detected in serum and vaginal samples 2 weeks following the initial immunization with MOMP. Antibody titers continued to rise until the day before the intrabursal challenge (Table 1). At that point the total IgG and IgA antibody titers in serum of the mice immunized with the vortexed MOMP were 12,800 and 1,600, respectively, while in the vaginal wash the IgG titer was 100 and the IgA titer was 50. Animals immunized with the sonicated MOMP had IgG and IgA titers in serum of 218,000 and 4,050, respectively, while in the vaginal wash they were 128 and 8, respectively. Control mice immunized intranasally (i.n.) with EB had IgG titers in serum and the vaginal wash of 48,600 and 80, respectively, while the IgA titers were 5,400 and 320, respectively.

View larger version (40K): [in this window] [in a new window]   FIG. 3.   Reactivity to octameric synthetic peptides of sera collected from mice immunized with the C. trachomatis MoPn vortexed and sonicated MOMP on the day before the genital challenge. The VDs of the MOMP are marked at the bottom of the figure.

Vaginal shedding of C. trachomatis following the intrabursal challenge. As shown in Table 3, 25 of the 28 (89.9%) control animals inoculated with ovalbumin shed C. trachomatis over the 6 weeks of the experiment. In contrast, only 7 of the 29 (24.1%) mice immunized with the vortexed MOMP had positive cultures during that period of time (P < 0.05). The duration and intensity of the shedding were also significantly different between these two sets of mice (Table 3). On average, the number of IFU recovered was 1 log₁₀ lower in the animals immunized with vortexed MOMP than in those inoculated with ovalbumin. A more limited protection was observed in the group of mice immunized with the sonicated MOMP. Of the 25 mice, 16 (64%) had positive cultures, and only during the first week was there a statistically significant difference in the number of IFU recovered from the mice immunized with the sonicated MOMP versus that recovered from ovalbumin. As expected, almost complete protection was obtained in the group inoculated i.n. with EB. Of the 18 mice in this group, only 1 (5.6%) had a positive culture on the second week, with a very low level of C. trachomatis MoPn IFU detected.

Fertility studies. Table 4 shows the results of the fertility studies. In the ovalbumin control group, only 3 of the 28 (10.7%) mice had embryos in both uterine horns. In contrast, of the mice immunized with vortexed MOMP, 16 of the 29 (55.2%) had embryos in both uterine horns (P < 0.05). However, animals immunized with the sonicated MOMP did not have a statistically significant increase in fertility. The group inoculated i.n. with EB and the fertility control group also had higher fertility rates than the ovalbumin control (P < 0.05). Furthermore, mice inoculated with ovalbumin or sonicated MOMP had significantly fewer embryos in the challenged left uterine horn than the fertility control animals (P < 0.05). In contrast, no significant differences were observed between the number of embryos in the right and left uterine horn in the animals immunized with vortexed MOMP, or in those inoculated i.n. with EB, compared with the number of embryos in the fertility control group.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this report we have shown that immunization with a preparation of the C. trachomatis MOMP, extracted directly from the organism, can induce significant protection against a genital challenge. This is, to our knowledge, the first time that a highly purified antigen of C. trachomatis has provided protection against a genital infection and infertility. This parallels the encouraging results recently reported using an anti-idiotypic antibody to the exoglycolipid antigen of C. trachomatis to protect mice against an ocular challenge (57) and the positive data obtained with DNA plasmids coding for the MOMP against an intranasal challenge (58).

Of the several antigens present in the chlamydial outer membrane, the MOMP appears to be the most obvious candidate for a vaccine against chlamydial genital infections (10, 41, 53). Based on that premise, various groups have used preparations of the MOMP, peptides corresponding to the VD, or DNA plasmids coding for the MOMP as an antigen in an attempt to induce protection (26, 31, 36, 44, 48, 49). Unfortunately, the results have been disappointing. Several explanations could account for these negative results, including the route of delivery, dose of antigen, and adjuvant preparations. On the other hand, a factor that has been considered by several investigators as potentially critical in inducing a protective immune response is the structural conformation of the epitopes of the MOMP (26, 36, 44, 49). In this respect, we have shown that a purified preparation of the chlamydial outer membrane complex, consisting mainly of the MOMP, the 60-kDa cysteine-rich protein, and LPS, can induce protective immunity against a genital challenge (31). Similar results have also been reported using a preparation of the outer membrane to protect sheep and guinea pigs against Chlamydia psittaci (2, 47). It was, in part, on the basis of these results that we considered the need to refold the MOMP before vaccination in an attempt to reconstitute the protective epitopes present in the native outer membrane of C. trachomatis. Here, our data suggest that epitopes that are formed by refolding the MOMP may be critical for inducing a protective immune response. Refolding may reconstitute some of the nonlinear protective epitopes present in a single MOMP molecule in the native EB and also those formed by homopolymers of MOMP (23). The fact that sonication of the folded MOMP resulted in a decrease in its protective ability supports this interpretation, although an alteration of linear epitopes by sonication cannot completely be excluded. This conclusion is also upheld by the finding that a MOMP preparation extracted and purified in the same manner as that reported here but not refolded failed to induce protection (31). Obviously, until the crystal structure of the MOMP and its conformation in the EB are elucidated, no definitive conclusions can be reached.

That a specific conformation of the epitopes appears to be important for inducing a protective response suggests that the humoral immunity plays a substantial role in this animal model (6, 22). This supports the observation originally described with vaccines against trachoma in humans and monkeys, indicating that the protection was serovar specific (13). It is important, however, to consider that the development of protective CD4⁺ T-cell responses may also depend on the proper conformation of the inducing epitopes. For example, Sjolander et al. (40) have shown that the native surface antigen 2 from Leishmania major can induce a protective response, while the same antigen produced in Escherichia coli failed to elicit protection. Since protection in mice against L. major is dependent on T helper cells, the authors concluded that correct protein folding and/or posttranslation modification maybe needed for eliciting a protective CD4⁺ T-cell response.

Other than the innate immune factors and the natural anatomical barriers present in the genital tract, protection against infection, called sterilizing immunity, will have to be mediated by antibodies on the mucosal surface (41, 53). A protective role for local antibodies in humans was first suggested by Brunham et al. (3), who showed an inverse correlation between the titer of IgA in genital secretions and the quantity of infectious units of C. trachomatis isolated from the cervix. Experimental proof that a monoclonal IgA antibody to a conformational epitope of MOMP could neutralize in vitro the infectivity of C. trachomatis MoPn has been reported (30). Subsequently it was demonstrated that passive immunization with this IgA antibody resulted in a significant decrease in the number of mice infected and in the intensity and duration of vaginal shedding following an intravaginal challenge (32).

Although IgA antibodies can block an infection at the site of entry, the challenge will be to develop a vaccination protocol that induces and maintains a high level of antibodies in the genital mucosa. Thus, in addition to inducing protective antibodies a vaccine most likely will also have to induce cell-mediated immunity. In this respect it appears that, at least based on experimental data with mice, a predominant Th1 response will be more efficient at eradicating a chlamydial infection than a dominant Th2 response (15, 21, 37). In general, immunization with viable organisms, including C. trachomatis, induces a predominant Th1 response, while nonviable antigens tend to elicit a Th2 dominant response (4, 18, 27, 31, 58). The vortexed MOMP preparation used here elicited a predominant Th1 response, while the sonicated MOMP induced mainly a Th2 response, as shown by the IgG2a/IgG1 ratio and the levels of IFN- and IL-4 produced by the splenocytes. These findings, in addition to the stronger T-cell proliferative response elicited by EB in mice immunized with the vortexed MOMP, suggest that the epitopes of this preparation more closely resemble those present in the native EB than those in the immunogenic sites of the sonicated MOMP.

There are presently several mouse models to test vaccine protocols and analyze the pathogenesis of a genital chlamydial infection (1, 9, 51). Each one of them has unique advantages, but not without shortcomings. The model of intrabursal inoculation used here was originally described by Barron et al. (1) and offers the advantage that all strains of mice so far tested, independent of age, develop infertility following a relatively low challenge with the C. trachomatis MoPn serovar (18). The obvious shortcoming of this model is the unnatural route of inoculation. The model in which the mice are challenged intravaginally requires a large dose of C. trachomatis MoPn, and certain strains of mice appear to be completely resistant to the development of infertility if they are inoculated past a certain age (9, 29). To overcome this limitation, an alternative model was developed in which the mice are treated with progesterone before they are challenged with Chlamydia (51). Treatment with progesterone, however, not only makes the genital epithelium more susceptible to a chlamydial infection but also induces a significant alteration of the immune system (14, 17, 35). This, in our opinion, may be a significant shortcoming when testing a vaccine candidate, since treatment with progesterone may abrogate the protective immune response induced by the antigen. Thus, presently we do not have an ideal animal model to test vaccine candidates against chlamydial genital infections. We think, however, that if different groups of investigators test their vaccine candidates in several murine models, it is likely that, based on their accumulated experiences, conclusions could be reached that are applicable to a large segment of the human population by the time that a vaccine candidate is ready for clinical trials.

Several steps will have to be explored before we can implement a vaccination protocol in humans with a MOMP preparation. The most obvious is the need to obtain quantities of MOMP sufficient for vaccination. Although it is possible to increase the production of MOMP using tissue culture techniques, this issue can probably be better addressed by exploring some of the recombinant techniques now available. However, this will not be an easy task, since it will require not only the production of the MOMP but also the purification and refolding of the protein into the proper conformation. Structural characterization of the MOMP preparation that we have described, including its crystallization, could be very helpful to start addressing this problem. The use of MOMP will still pose another question, and that is the still-unresolved issue of whether or not this protein from one serovar can protect against all the other C. trachomatis serovars (34, 49). This will have to be addressed in the murine and monkey models before clinical trials are performed with humans (1, 9, 33, 51).

In addition, the issue as to whether or not the trace amounts of LPS present in this MOMP preparation may directly or indirectly play a role in protection will probably not be answered until a recombinant vaccine is tested. The fact that a nonrefolded preparation of MOMP that was extracted and purified using the same procedure utilized here did not induce protection suggests that if the LPS plays a role, it most likely will do so indirectly by modifying the MOMP (31). Alternatives to the use of Freund's adjuvant will also have to be explored. Various adjuvants are under intense investigation for other vaccines, so knowledge gained with other antigens may be applicable to the MOMP (7). In conclusion, we have shown that a purified and refolded preparation of the MOMP can induce protection against a genital challenge with C. trachomatis in a mouse model. A significant amount of work, however, is required before we can consider applying this approach to humans.