Outcome of urogenital infection with Chlamydia muridarum in CD-14 gene knockout mice

A mouse model of urogenital chlamydial infection has been established for the study of immunity [1, 2] and pathogenesis [3‐7] to chlamydial infections. While in-roads have been made into understanding protective immunity in this model, much less is known about the factors associated with pathogenesis leading to the chronic sequelae of infection. One of the sequelae is tubal scarring and infertility, and a surrogate marker of infertility – hydrosalpinx formation [7, 8]. Several recent reports suggest that acute inflammatory responses may cause an initial insult leading to these outcomes [7, 9] but persistent infections that drive chronic inflammation are also possible causes [6].

Toll-like receptors (TLR) are a primary means of initiating innate immune system responses through the recognition of pathogen-associated molecular patterns (PAMPs) [10, 11]. We read with interest the recent work of Darville, et al., that assessed the role of TLR-2 and TLR-4 in the mouse model chlamydial infection [4]. In this study, they found that although TLR-2 knockout (-/-) mice had a similar infection course as wild type (wt) controls, they showed a significant reduction in oviduct pathology and the accompanying proinflammatory cytokine production. A similar effect was not observed in TLR-4 mice. These results indicated that TLR2 is a major means of signaling to induce innate immune responses in this model.

At the time of the Darville report, our laboratory was involved in assessing the role of CD14 in the murine model. CD14 has been proved to be an accessory protein for TLR4 recognition of lipopolysaccharide (LPS) and subsequent transduction of intracellular signaling via nuclear factor-kappaB [12]. While the LPS of chlamydia has been shown to have relatively weak endotoxic activity [13], we thought it was possible that CD14 could also function as a critical accessory protein for other PAMPs such as heat shock proteins (HSP) [14] or for the removal of apoptotic cells by macrophages – an event which could regulate inflammatory responses [5, 15‐17]. In humans, reports have shown that a functional polymorphism in the CD14 gene is not involved in susceptibility to C. trachomatis infection or fallopian tube pathology [18]. Nonetheless, similar functional polymorphisms have been variously associated with susceptibility to C. pneumoniae infection and possibly pathological immune responses [19‐22]. Based on these observations, we hypothesized that CD14 would be involved in chlamydial pathogenesis in the mouse model through one of these means or by heretofore unrecognized pathways.

Along with LPS-binding protein (LBP) and MD-2, CD14 has been proved to be an accessory protein for TLR4 recognition of lipopolysaccharide (LPS) and subsequent transduction of intracellular signaling via nuclear factor-kappaB for proinflammatory cytokine production [12]. While the LPS of chlamydia has been shown to have relatively weak endotoxic activity [13], we thought it possible that a high concentration of chlamydial LPS at the foci of an infection, could serve to synergize with other PAMPs to initiate or influence inflammatory responses. Other possible chlamydial PAMP that could engage CD14 exist. For example, it has been shown that CD14 binds heat shock proteins (hsp) [14] and the chlamydial hsp has been implicated as a hypersensitizing antigen at least partly responsible for the pathogenesis of C. trachomatis infections [27, 28]. Also, regulation of apoptosis in the murine model has been proved to have profound effects on immunopathology [5, 15] and CD14 has been shown to serve as receptor on macrophages for the recognition of cells undergoing apoptosis [16, 17]. In other models of Gram-negative bacterial infection, CD14 gene knockout mice had been proved to have altered outcomes of infection. Examples of these include Acinetobacter bauminii [29], Legionella pnuemophila [30], Borrelia burdorferi [31] and Salmonella typhimurium [32].

Several recent reports indicate that the Chlamydiae can engage TLR through LPS and non-LPS-dependent pathways [13, 20, 28]. However, TLR2, and not TLR4 appears to be important for the initial signaling pathway for proinflammatory cytokine production during chlamydial infection [20, 33]. Most relevant to the present findings, Darville, et al., have recently shown a lack of involvement of the TLR4 but a prominent role for TLR2 in the MoPn model of female urogenital chlamydial infection and in murine fibroblasts and macrophages [4]. TLR2-/- mice and primary cell cultures produced lower levels of proinflammatory chemokines and cytokines, displayed significantly lower pathology scores and a lower rate of hydrosalpinx formation when compared to wt control mice. TLR4-/- mice showed no significant change in infection course or pathology but some increases in cytokine production were observed compared to wt control mice and primary cell lines derived from these [34].

Our present results showing a lack of involvement of CD14 in pathogenesis largely agree with the work of the Darville report regarding TLR4. We make this correlation with Darville's findings because CD-14 is an integral component TLR4 signaling pathway. In this regard, it functions as a co-receptor for LPS [12], HSP [14], and apoptotic cells for phagocytic clearance by macrophages [17]. Some divergence with the findings of Darville, et al. was found in the observation that there was a statistically significant abbreviation of infection course in CD14-/- mice, particularly during the time frame of onset of the adaptive immune response and infection resolution (days 10–21 post-infection) [35]. Though the reason for this observation remains unclear, it could be that signaling through CD14-TLR4 pathway induces a proinflammatory cytokine profile that protracts resolution of the infection by impeding the appropriate adaptive immune response. A complete and exhaustive study of cellular or antibody-mediated immune responses would be required to substantiate this hypothesis.

It should also be noted that the MoPn agent is a natural pathogen for the mouse and sustains a significantly more vigorous infection, a far higher incidence of ascending upper genital tract infection, and higher acute inflammatory responses than human C. trachomatis serovars in the mouse [3]. Perhaps, a strain difference in chlamydial PAMP expression or in host expression of TLRs could account for this. In this respect, the application of our results to human infections, as with all animal models, should be carefully interpreted.

Another caveat for interpreting our results relates to our use of progesterone pretreatment. This is used to enhance infection by halting the rapid estrus cycle of the mouse and providing a lush epithelial environment for growth of this pathogen that prefers epithelial cells for replication. Without such treatment, we have found that a one hundred percent infection rate cannot be obtained unless mice are inoculated on 2 or more consecutive days with much higher doses of the organism. While we can find no reports in the literature related to progesterone modulating CD14 expression, we cannot completely rule out that pretreatment with progesterone could have blurred our results.

Due to the required intracellular replication of the Chlamydiae, it is more likely that intracellular pattern recognition receptors are involved in initiating host responses. To this end, the aforementioned work by Darville, et al. [4] and more recent findings of Welter-Stahl, et al., suggest that this may be the case [36]. These findings indicate that intracellular pattern recognition receptors such as Nod1 and TLR2, rather than those found extracellularly, are likely candidates. While PAMP such as a rudimentary peptidoglycan expressed by the Chlamydiae may contribute to the initiation of the host response, certainly other chlamydial ligands are involved. These findings complicate vaccine development that targets neutralization of chlamydial PAMPs in order to prevent harmful inflammatory responses because restriction of expression of TLR agonists to the intracellular niche protects these antigens from neutralizing antibodies.

The results herein imply that engagement of CD14 in this model does not affect chronic disease progression (e.g. hydrosalpinx formation and by extension, infertility) but does impede resolution of the infection.