Chlamydia trachomatis is the most common bacterial pathogen associated with sexually transmitted genital tract infections both in the United States [1] and worldwide [2]. It is generally accepted that most female genital tract infections with C. trachomatis are both asymptomatic and without severe sequelae [3]; and, that despite improved screening programs and the availability of highly effective antibiotics [4–6], there has been a significant increase in the incidence of C. trachomatis genital tract infection within the last half decade [2, 7]. Although epidemiologic studies suggest that prior infection with C. trachomatis confers some short term protection against reinfection [8, 9], the exact nature of this acquired immunity remains undetermined as do issues relating to the serovar specificity and possible involvement of this immunity in the more severe sequelae associated with multiple and/or chronic infection [10–12].

It has been advanced that serovar specific immune responses, particularly those made to the major outer membrane protein (MOMP), contribute to protection, whereas responses to broadly shared antigens, particularly those induced by chlamydial heat shock protein 60 (Hsp60), are associated with the immunopathologic injury that contributes to ectopic pregnancy and tubal infertility [13–19]. There is also increasing evidence from both human epidemiologic studies and animal model experiments to support the hypothesis that a protective immune response to reinfection is a complex, site of infection variable interaction between C. trachomatis and specific Th1 dominant cellular responses and interferon-gamma mediated Th1 augmenting humoral responses [20, 21].

In attempts to understand the salient features and specific components of this interplay, a great deal of work has been performed using slightly modified versions of a murine model of female genital tract infection first described by Tuffrey and Taylor-Robinson [22]. In our laboratory, we routinely use strains belonging to the human oculogenital biovar of C. trachomatis and have previously reported significant differences in the course of infection among strains belonging to 7 oculogenital serovars, which loosely correlated with the prevalence of the serovars among human clinical isolates, especially for the most and least prevalent serovars, i.e. D and E versus H and I [23].

The purpose of the present study was to expand on these observations by assessing the degree of homotypic and heterotypic protection against reinfection that follows resolution of infection with human isolates of C. trachomatis. Two strains were selected from the previously studied collection of strains: the strain of serovar D which was shown to establish the longest duration of infection (median duration of 38 days) and greatest humoral response; and the serovar H strain which had the shortest duration (median duration of 7 days) and lowest humoral response. In addition, lower genital tract infection with serovar D was shown to both ascend into the uterine horns with a greater frequency than serovar H, as well as shed more infectious units during the acute phase of infection [23]. In a separate study we demonstrated a link between certain in vitro growth characteristics and differences in the level of cytotoxic activity associated with elementary bodies between these strains [24]. Thus, strain selection was made with the intention of reflecting the greatest diversity observed among the strains studied in the murine model, as well as to represent serovars that have significantly different prevalence rates among human isolates in the hope that the results of the study would provide some insight into the possible causes of these differences. Duration of genital tract infection was used to determine the extent of protection, and humoral and cellular immune response data were evaluated to identify factors that associated with any observed protection.

Using a murine model of C. trachomatis female genital tract infection, we have demonstrated that homotypic immunity against reinfection was induced following initial infection with either serovar D or H, but that a more significant level of protection was observed following infection with serovar D. However, heterotypic protection against reinfection was strain dependant and was seen only when the initial infection was with the more virulent and immunogenic strain of serovar D. These results are the first demonstration of heterotypic immunity between two oculogenital serovars following female genital tract infection in the mouse model, as well as being the first comparative study that suggests a possible strain dependent restriction on the process. The findings are consistent with other studies that have observed heterotypic immunity in the context of MoPn and serovar E. Also consistent with prior reports is the correlation between the level of protection observed and the virulence of the strains used to infect the mouse genital tract, with the more virulent MoPn inducing a more solid level of heterotypic immunity against serovar E than was observed in the reverse situation [25, 26]. Finally, the observed differences in heterotypic and homotypic responses between serovars D and H could explain in part the relatively stable differences in the frequency of these two serovars among human isolates from different geographic locations [27–33], i.e. a prior infection with a more virulent/immunogenic serovar would confer a greater degree of protection against infection with a less virulent/immunogenic serovar, thus reducing the incidence of the latter by reducing the efficiency of serial transmission.

Consistent with human epidemiologic data [8, 9], the protection observed in this model against reinfection with oculogenital serovars of C. trachomatis was not complete, but rather acted in a way that reduced the duration of infection and level of bacterial shedding (data not shown) during infection. With the notable exception that serovar H infection resulted in a lesser quantitative and qualitative humoral response, a thorough analysis of the individual and collective data was unable to identify any specific element(s) that correlated with protection. This finding is consistent with a report in which a similarly extensive analysis of cellular and humoral responses was performed in a comparison of the acquired immunity induced by infection with MoPn and serovar E [25]. Although not proof, this does support the current working hypothesis that acquired immunity to C. trachomatis female genital tract infection is a complex and integrated phenomena that relies on both Th1 and Th2 type responses made during the course of infection, which in turn enhance innate immune responses upon reinfection [20]. This complexity, which likely arises out of the ever changing physiologic and immunologic milieu within and between anatomically distinct but connected regions of the female genital tract, may account for the difficulty identifying specific components of what may be a flexible pattern of responses that lead to a given individual's level of protection against or risk of severe upper genital tract pathology [21].

Women with recurrent C. trachomatis infection are at increased risk of reproductive sequelae, including pelvic inflammatory disease, ectopic pregnancy and tubal infertility [10–12], which have been linked to both cellular and humoral immune responses induced during infection [13–19]. How the nature and level of homotypic and/or extent of heterotypic immunity in the murine model extrapolate to the risk of upper genital tract pathology and infertility was not assessed in this study, and is an area of investigation that has yet to be systematically addressed. Most of the experimental data relating to the severe sequelae associated with C. trachomatis female genital tract infection has been obtained in studies using C. muridarum, MoPn [34, 35]. As a result, it has not been possible to clearly assess the immunologic features that are thought to contribute to severe sequelae within the human female genital tract, because the damage that occurs within the murine genital tract following infection with MoPn is a consequence of acute and not chronic infectious processes and/or recurrent infection [36, 37]. Typically in the mouse and in most women, human oculogenital serovars are limited in their ability to ascend with any major pathologic consequence from the initial site of infection within the lower genital tract. However, infection of female C3H/HeN with a strain of serovar E has been shown to ascend and cause infertility without gross pathology at a low incidence following a single infection and with increased incidence upon reinfection [26]. Of particular interest is that none of these mice developed hydrosalpinx, which is the hallmark of upper tract infection with MoPn, indicating a possible alternative mechanism for the induction of infertility, one different from the tubal dilation, scarring and associated hydosalpinx that occurs as a sequelae to MoPn infection. Although no specific immunologic mechanism was identified as contributing to the development of infertility, it was noted that the C3H/HeN mouse is an Hsp60 responding strain and that Hsp60 responding strains of mice are more susceptible to MoPn induced infertility while non-responders can be infected but do not experience severe upper tract pathology. It will be worthwhile to determine if the two C. trachomatis strains used in the present study show a similar relationship to each other in C3H/HeN mice and whether homotypic and/or heterotypic elements of immunity play a role in the progression of events that lead to severe upper tract pathology, which is essentially the reason why C. trachomatis genital tract infections are significant and why intervention and prevention strategies are needed.

In conclusion, we demonstrate in the first study comparing phenotypically different strains representing two human oculogenital serovars that both the level of homotypic protection against reinfection and the ability to confer heterotypic protection correlated with the virulence/immunogenicity of the strain. Extrapolating the results to human epidemiologic data could explain in part the relatively stable differences in the frequency between the most and least prevalent serovars based on a serovars ability to induce or not induce heterotypic immunity. Although no specific cellular or humoral factor(s) could be identified that associated with the observed protection, it is clear that heterotypic immunity can be induced and that the systematic study of human oculogenital serovars in the mouse model of female genital tract infection could provide information that leads to an understanding of what distinguishes a protective from an immunopathic response to infection.