Immunopathogenesis of genital Chlamydia infection: insights from mouse models

ABSTRACT

Chlamydiae are pathogenic intracellular bacteria that cause a wide variety of diseases throughout the globe, affecting the eye, lung, coronary arteries and female genital tract. Rather than by direct cellular toxicity, Chlamydia infection generally causes pathology by inducing fibrosis and scarring that is largely mediated by host inflammation. While a robust immune response is required for clearance of the infection, certain elements of that immune response may also damage infected tissue, leading to, in the case of female genital infection, disease sequelae such as pelvic inflammatory disease, infertility and ectopic pregnancy. It has become increasingly clear that the components of the immune system that destroy bacteria and those that cause pathology only partially overlap. In the ongoing quest for a vaccine that prevents Chlamydia-induced disease, it is important to target mechanisms that can achieve protective immunity while preventing mechanisms that damage tissue. This review focuses on mouse models of genital Chlamydia infection and synthesizes recent studies to generate a comprehensive model for immunity in the murine female genital tract, clarifying the respective contributions of various branches of innate and adaptive immunity to both host protection and pathogenic genital scarring.

This review summarizes mouse model studies that significantly contributed to our current understanding of the protective and pathogenic immune responses taking place during genital Chlamydia infections.

INTRODUCTION

Chlamydia species cause a plethora of clinical syndromes comprising a significant public health burden worldwide. Different species and serovars of Chlamydia display tropism for different tissues underlying their respective disease phenotypes. Chlamydia pneumoniae causes an extremely common atypical pneumonia and is linked with the development of atherosclerosis and coronary artery disease (reviewed in Campbell and Kuo 2004; Porritt and Crother 2016). Certain serovars of Chlamydia trachomatis are tropic for ocular tissues, comprising the leading cause of infectious blindness worldwide (reviewed in Taylor et al. 2014). Chlamydia trachomatis serovars that display tropism for genital tissues cause the most common sexually-transmitted bacterial infections in the US (Braxton et al. 2018). Of the sexually-transmitted serovars, L serovars cause lymphogranuloma venereum (LGV), an infection of lymph nodes as well as genital and rectal tissue (reviewed in Mabey and Peeling 2002). Chlamydia trachomatis serovars D–K establish infection inside genital epithelial cells, cause pathology mostly in women, and can be asymptomatic or cause pelvic pain and mucopurulent cervical discharge. Repeated or chronic infection can lead to scarring of the upper genital tract, increasing the risk of ectopic pregnancy and infertility (reviewed in Brunham and Paavonen 2020). All these species of Chlamydia infect epithelial cells, inside which they replicate and persist (reviewed in Elwell, Mirrashidi and Engel 2016). The infected epithelium activates mechanisms of cell-autonomous immunity, undergoes cellular changes and produces proinflammatory cytokines, which recruit and activate innate immune cells such as neutrophils, macrophages, dendritic cells and potentially tissue-resident lymphoid cells (reviewed in Finethy and Coers 2016). As the infection proceeds, antigen-presenting cells activate adaptive immunity, leading to production of anti-Chlamydia antibodies by B cells and influx of Chlamydia-specific CD4 and CD8 T cells into infected tissue (reviewed in Poston and Darville 2018). The result is an inflammatory environment that contains and frequently clears the infection but also damages the infected tissue.

There has been significant controversy regarding which of these responses to Chlamydia infection mediate pathology. It was initially thought that the antigen-specific responses—mediated by T cells, Chlamydia-specific antibodies or even autoreactive antibodies resulting from molecular mimicry to Chlamydia Hsp60—resulted in tissue scarring and pathology (reviewed in Brunham and Peeling 1994). The contributions of antibodies to pathology have largely been discredited (Ness et al. 2008), with more focus placed on inflammatory signaling derived from infected epithelial cells and subsequent recruitment of innate immune cells, such as neutrophils, and induction of various T cell responses (reviewed in Darville and Hiltke 2010).

The studies regarding chlamydial pathogenesis in humans have been extremely informative in correlating certain immune responses with pathology. Mechanistic studies, however, require animal models, and while non-human primate models have occasionally been used (Bell et al. 2011), chlamydial disease has been studied mainly in mice and guinea pigs. Amongst these two models the immunological response in the guinea pig may better recapitulate that observed in humans (Rank and Sanders 1992; Rank, Bowlin and Kelly 2000), but genetic tractability and reagent availability make the mouse the most useful model to examine specific immune mechanisms on a granular level. After treatment with medroxyprogesterone to halt uterine epithelial cell shedding in estrus, Chlamydia infection can successfully be established in the murine genital tract via the intravaginal, intrauterine, or transcervical routes, depending on the Chlamydia strain (Tuffrey and Taylor-Robinson 1981; Gondek et al. 2012).

Different Chlamydia species display tropism for different host species, demonstrating evolutionary adaptation to varying host immune mechanisms. Chlamydia trachomatis mouse pneumonitis (MoPn) strain, now commonly referred to as Chlamydia muridarum, is the mouse-adapted Chlamydia species (Barron et al. 1981), subverts murine cell-autonomous immune mechanisms (Nelson et al. 2005; Coers et al. 2008), and establishes a self-resolving 4–8-week-long genital infection in mouse hosts. Human-pathogenic C. trachomatis serovars, in contrast, are rapidly cleared from the mouse female genital tract (Fig. 1) and fail to establish productive infections or induce pathogenic immune responses (Perry, Feilzer and Caldwell 1997). Likewise, C. muridarum is highly susceptible to the human cell-autonomous immunity and is therefore nonpathogenic in human hosts (Haldar et al. 2016). A primary C. muridarum genital infection is able to elicit robust genital pathology in mice. Significant inflammatory infiltrate predominated by neutrophils in the oviduct lumen (pyosalpinx) is readily detected in C. muridarum-infected mice at 14dpi and 21dpi, and resolves by 28dpi. Beginning at 28dpi and at later timepoints, glandular duct blockade causes uterine horn dilation; and tissue fibrosis and occlusion of the oviduct lumen cause oviduct dilation (hydrosalpinx) (Shah et al. 2005; Sun et al. 2015). These representations of genital scarring are associated with infertility in primary murine C. muridarum infection. Humans, on the other hand, seem to require secondary C. trachomatis infections to drive significant pathology (reviewed in Brunham and Paavonen 2020). The specific immune mechanisms that mediate human pathology, however, have not been fully elucidated. Mouse models using C. muridarum can therefore be used to implicate specific pathogenic immune mechanisms which can then be investigated further for their roles in human disease.

Figure 1.

Kinetics of bacterial burden in murine genital Chlamydia infection. Prior to activation of cell-autonomous immunity, C. trachomatis and C. muridarum grow unchecked inside genital epithelial cells. Innate IFNγ production induces expression of the IRG system, resulting in destruction of C. trachomatis but not C. muridarum. CD4 T cells ultimately clear genital C. muridarum infections after 4–7 weeks. Figure created with Biorender.com.

Chlamydia infection induces an array of immune responses, beginning with the infected epithelial cell, followed by activation and recruitment of innate immune cells such as neutrophils, macrophages and potentially tissue-resident lymphoid cells within the first week after infection, leading into subsequent cellular and humoral adaptive immunity (reviewed in Finethy and Coers 2016; Poston and Darville 2018). Not all of these responses are important for clearing the infection, nor does each of them cause tissue damage and subsequent scarring. An ideal vaccine would induce immune mechanisms that specifically destroy bacteria, resulting in protective immunity, without activating immune mechanisms that may cause tissue damage and disease sequelae. In this review, we describe the components of the anti-Chlamydia immune response during primary mouse female genital tract infection, parsing the respective contributions of branches of genital immunity to bacterial clearance and immunopathology.

Murine innate immunity controls C. trachomatis but not C. muridarum genital infection

Cell-autonomous immunity to Chlamydia infection in mouse cells

Once Chlamydia enter epithelial cells, they form a membrane-bound inclusion body inside which they replicate prior to lysing or extruding from infected cells (reviewed in Hybiske and Stephens 2008). The first approximately 24 h of murine genital Chlamydia infection is marked by rapid bacterial growth as the infection proceeds unchecked (Rank, Bowlin and Kelly 2000). During this time, bacteria are sensed by various cellular pattern-recognition receptors including Toll-like receptors 2 and 3 (TLR2 and TLR3) (Darville et al. 2003; Derbigny, Kerr and Johnson 2005; Derbigny et al. 2010, 2012; Yu et al. 2016), nucleotide-binding oligomerization domain-containing protein 1 (NOD1) (Welter-Stahl et al. 2006), as well as the nucleic acid sensors cyclic GMP–AMP synthase (cGAS), and stimulator of interferon genes (STING) (Prantner, Darville and Nagarajan 2010; Barker et al. 2013; Zhang et al. 2014a, 2017). In days 1–3 after infection, bacterial sensing leads to genital production of the cytokine gamma-interferon (IFNγ; Rasmussen et al. 1997), which induces cell-autonomous defense programs in both infected and non-infected cells (reviewed in Finethy and Coers 2016). Cell-autonomous immunity to intracellular bacterial pathogens such as Chlamydia is mediated in part by the immunity-related GTPase (IRG) family of proteins (reviewed in Pilla-Moffett et al. 2016), members of which localize to Chlamydia inclusions and trigger the lytic destruction of targeted inclusions (Coers et al. 2008; Haldar et al. 2013, 2015).

IFNγ treatment of murine cells is sufficient for robust restriction of C. trachomatis growth in non-immune cells such as mouse embryonic fibroblasts (MEFs) (Bernstein-Hanley et al. 2006; Coers et al. 2008), and indeed IFNγ production in the genital tract coincides with a decrease in C. trachomatis burden after the first day post-infection (dpi). However, the mouse-adapted pathogen C. muridarum is able to subvert the IRG system and evade IFNγ-mediated restriction in non-immune cells (Coers et al. 2008) corresponding with a further increase in bacterial burden in the female genital tract through 6 dpi (Perry, Feilzer and Caldwell 1997; Coers et al. 2011; Gondek et al. 2012). For this reason, mice are readily vaginally infected with C. muridarum as it can ascend to the uterus and oviducts, whereas C. trachomatis must be instilled transcervically into the uterus in order to establish infection in IRG-competent animals.

The importance and early sources of interferon-γ

As infection proceeds, macrophage and neutrophil populations begin to expand within infected tissue (Morrison and Morrison 2000; Rank, Bowlin and Kelly 2000; Lijek et al. 2018). Both cell types are able to phagocytose and destroy extracellular bacteria, but generally play different roles in immune responses to bacterial infection. Macrophages are the primary innate immune cell type in defense against intracellular bacterial pathogens, and once activated by IFNγ possess potent antimicrobial properties (reviewed in Mosser and Zhang 2008). Indeed, IFNγ priming further stimulates the ability of mouse bone marrow derived phagocytes to clear C. trachomatis or C. muridarum infections via complex mechanisms including perforin-2 (Fields et al. 2013), components of the autophagy pathway (Radomski et al. 2017), and phagolysosomal nitric oxide production (Rajaram and Nelson 2015; Nagarajan et al. 2018). Similar to epithelial cells, macrophage cytokine responses to Chlamydia are mediated by TLR2 (Darville et al. 2003) and STING (Prantner, Darville and Nagarajan 2010). Additionally, Chlamydia infection can activate the NLRP3 and AIM2 inflammasomes, as well as the noncanonical caspase11 inflammasome, leading to pyroptosis and production of pro-inflammatory cytokines (Prantner et al. 2009; Finethy et al. 2015; Chen et al. 2017; Webster et al. 2017; Allen et al. 2019).

Macrophages can display varying ontogenies and phenotypes (reviewed in Ginhoux and Guilliams 2016) and no genetic deletion or antibody depletion studies have effectively demonstrated the in vivo importance of macrophages to restriction of Chlamydia growth. Nonetheless, the aforementioned in vitro macrophage studies suggest that they do play important roles in vivo, both directly killing extracellular Chlamydia and potentiating immune responses through production of cytokines. However, the effectiveness of macrophage immunity varies with the Chlamydia species used. Whereas innate immunity, including macrophage-mediated processes, is sufficient to sterilize murine C. trachomatis infection in vivo, C. muridarum is able to overcome innate immune clearance in mouse models. Escape from destruction inside macrophages via inhibition of phagosome–lysosome fusion (Rajaram and Nelson 2015), combined with its ability to subvert epithelial cell-autonomous immunity (reviewed in Finethy and Coers 2016), explains how C. muridarum is able to establish a successful genital infection in mice.

The anti-Chlamydia activities of both macrophages and infected epithelial cells are potentiated by IFNγ (Johansson et al. 1997). Whereas wildtype mice demonstrate decreased C. trachomatis burden from 3 to 6 dpi, C. trachomatis-infected mice lacking IFNγ or the IRG family members Irgm1 and Irgm3, or wildtype mice infected with C. muridarum, demonstrate a comparable increase in burden through 6 dpi (Gondek et al. 2012). These findings beautifully illustrate that while IFNγ in the first week after infection is sufficient and necessary to restrict C. trachomatis growth in an IRG-dependent manner, it is insufficient to clear C. muridarum infection within that timeframe.

IFNγ-producing CD4 T cells (Th1 cells) are the major producers of IFNγ in the latter stages of infection (Nagarajan et al. 2011), but more controversy surrounds the source(s) of IFNγ in the first several days after infection. There are few tissue resident immune cell types known to secrete IFNγ that have the potential to respond quickly to genital Chlamydia infection. The most obvious is the natural killer (NK) cell, the most prominent subpopulation of innate lymphoid cells (ILCs) whose functions include the induction of cytotoxicity in infected cells and the production of cytokines such as IFNγ (reviewed in Zhang and Huang 2017). NK cells are present in the genital tract early after infection, and administration of anti-NK1.1 or anti-asialo-GM1 depleting antibodies eliminate IFNγ-producing cells in the iliac lymph nodes of C. muridarum-infected mice (Tseng and Rank 1998). However, these depletion schemes also eliminate other ILC populations that may produce significant amounts of IFNγ. Notably, IFNγ-producing ILC1s are important sources of IFNγ in other mucosal tissues such as the gut, as are ‘ex-ILC3s,’ i.e. IL-17-producing non-cytotoxic ILCs that convert into IFNγ-producing ILCs (reviewed in Artis and Spits 2015; Vivier et al. 2018). While the physical presence of these alternative ILC subpopulations in the murine genital tract has not been demonstrated, mice deficient for ILCs, as achieved by genetic deletion of the IL-7 receptor or deletion of both Rag2 and the IL-2 receptor common γ-chain (γc), were unable to control early C. trachomatis growth in an IFNγ-dependent manner (Xu et al. 2020). These findings suggest that ILC1 and/or ex-ILC3 populations in the genital tract may be responsible for the production of innate IFNγ prior to Th1 cell influx, and that this innate production of IFNγ is sufficient to control early C. trachomatis growth. It is possible that other innate-like cell types, such as γδT cells, may provide a redundant early source of genital IFNγ, and additional studies are warranted to evaluate these genital ILC populations and the mechanisms of their activation during C. trachomatis infection.

Although C. muridarum is able to evade IFNγ-mediated restriction in epithelial cells and macrophages, innate IFNγ does nonetheless appear to play a pivotal role in preventing C. muridarum dissemination from the female genital tract into other tissues. Earlier studies using polyclonal stocks of the C. muridarum Nigg strain demonstrated prolonged vaginal shedding of bacteria, increased bacterial dissemination throughout the course of the infection and rare instances of lethality in vaginally infected IFNγ-deficient mice. Wildtype mice on the other hand survived infection, displayed minimal bacterial dissemination and cleared the vaginal infection within approximately 4 weeks (Cotter et al. 1997; Perry, Feilzer and Caldwell 1997). These findings were essentially confirmed in subsequent studies, with the notable distinction that the use of the more virulent C. muridarum Weiss strain (Sturdevant and Caldwell 2014) or hypervirulent plaque-purified clonal isolates of the Nigg strain (Poston et al. 2018; Mercado et al. 2020) resulted in more pervasive lethality in IFNγ-deficient mice. The systemic and sometimes lethal infection of vaginally inoculated IFNγ-deficient mice is not seen in mice lacking conventional αβ T cells (Mercado et al. 2020), suggesting that ILC-derived innate IFNγ plays an important role in preventing C. muridarum dissemination and host survival. In support of this conclusion the total elimination of ILCs in Rag2-/γc-deficient mice led to greater C. muridarum dissemination, accelerated weight loss and an earlier onset of death compared to infected Rag1-deficient mice (Mercado et al. 2020). Altogether, these findings suggest that innate IFNγ rather than Th1-derived IFNγ is responsible for limiting early C. muridarum dissemination and subsequent lethality, although Th1 cells are clearly important for ultimate clearance of the genital infection (Mercado et al. 2020). The mechanisms by which innate IFNγ prevents C. muridarum dissemination remain unclear. Additional studies are needed to parse the distinct roles of innate and Th1-derived IFNγ in defense against C. muridarum infection.

Neutrophils play a minor role in defense against Chlamydia

Apart from IFNγ-responsive macrophages, neutrophils are the other major innate immune cell type that responds early in genital Chlamydia infection. Neutrophils are present at low levels in uninfected tissues and are recruited to infected tissues from the bloodstream (reviewed in Ley et al. 2018; Petri and Sanz 2018). Similar to macrophages, neutrophils are voracious phagocytes and are able to destroy engulfed bacteria using reactive nitrogen and oxygen species. Neutrophils possess additional functions that are effective against extracellular pathogens: in a specialized form of programmed cell death, neutrophils release cytoplasmic and nuclear contents such as DNA and oxygen radicals, forming a neutrophil extracellular trap (NET) that is toxic to microbes and host tissue alike. Neutrophils therefore comprise the primary effector immune cell type in infection with extracellular pathogens (reviewed in Ley et al. 2018), with their roles in infection with an intracellular pathogen such as Chlamydia less established.

Neutrophils are present in Chlamydia-infected genital tissue as early as 24 hpi (Rank et al. 2011) and have been shown to destroy extracellular Chlamydia in vitro (Naglak, Morrison and Morrison 2017). Protective roles for neutrophils in vivo vary with the Chlamydia species used. Antibody depletion experiments have shown that neutrophils may play a role in defense against C. muridarum within the first 2 days of infection (Rank et al. 2011) but not throughout the course of infection (Lee et al. 2010; Frazer et al. 2011). For defense against C. trachomatis mouse infection, on the other hand, neutrophils appear to be dispensable (Lijek et al. 2018). Interestingly, electron micrographs of infected genital epithelia have shown neutrophils entering epithelial cells to engulf whole inclusions or extruding infected epithelial cells into the uterine lumen (Rank et al. 2011). While these observations suggest novel mechanisms for protective functions of neutrophils in defense against intracellular pathogens, their contribution to in vivo immunity has not been shown. Because C. muridarum is able to survive inside murine epithelial cells, it is quite possible that neutrophils function predominantly in the destruction of extracellular C. muridarum, thus limiting cell-to-cell spread. Neutrophils would therefore be largely dispensable in defense against C. trachomatis, which is readily destroyed by immune-primed murine epithelial cells. Alternatively, Chlamydia may have evolved mechanisms to interfere with neutrophil function or recruitment (Dolat and Valdivia 2020). Altogether, neutrophils appear to play only a minor protective role in genital Chlamydia infection.

Collectively, there are three phases of the early (pre-adaptive) immune response to genital Chlamydia infection (Fig. 1). In the first day of infection Chlamydia growth proceeds unchecked prior to activation of cell-autonomous immune defenses elicited by IFNγ priming. Unprimed infected epithelial cells sense intracellular Chlamydia via TLR2, TLR3, NOD1 and STING, secrete proinflammatory cytokines and activate an unknown cellular source—likely to be ILCs and/or γδT cells—to produce IFNγ. In response to IFNγ, the IRG system is induced in mouse epithelial cells, leading to an initial restriction of the growth of C. trachomatis but not C. muridarum. Macrophages and neutrophils then respond, killing extracellular and possibly intracellular Chlamydia and further secreting cytokines via activation of the same pattern-recognition receptors as in epithelial cells as well as inflammasomes. These functions are sufficient to clear C. trachomatis infection, but because C. muridarum is able to survive inside cytokine-primed epithelial cells as well as macrophages, C. muridarum infection persists beyond innate immunity and requires an adaptive immune response for clearance.

Adaptive immunity provides sustained IFNγ production to clear an established Chlamydia infection

Th1 cells comprise the crucial element of adaptive immune responses

While the innate response restricts Chlamydia growth in the genital tract in the first week of infection and limits dissemination to other tissues, dendritic cells in genital tissue ingest Chlamydia antigens, migrate to the draining iliac lymph nodes and present these antigens via the major histocompatibility complexes (MHC) I and II to naïve T cells (Su et al. 1998; Marks et al. 2007). The corresponding lymphocyte responses consist predominantly of Th1- and Th17-polarized CD4 T cells as well as CD8 T cells (Ramsey and Rank 1991; Igietseme et al. 1993, et al. 1994; Roan and Starnbach 2006; Roan et al. 2006; Coers et al. 2011; Gondek et al. 2012; Li and McSorley 2013; Vicetti Miguel et al. 2016), which migrate to the genital tract in a CXCR3- and CCR5-dependent manner (Olive, Gondek and Starnbach 2011; Lijek et al. 2018) to execute different effector functions. Genital CD4 and CD8 lymphocyte numbers peak at 14 dpi and remain elevated for weeks after the resolution of infection (Morrison and Morrison 2000; Gondek et al. 2012). While B cells do respond to Chlamydia infection, they are likely dispensable for clearance of primary infection from the genital tract (Ramsey, Soderberg and Rank 1988; Johansson et al. 1997), and instead produce antibodies which aid in preventing bacterial dissemination to distant tissues (Li and McSorley 2013; Poston et al. 2018; Malaviarachchi et al. 2020). Additionally, B cells have a protective role in reinfection (Su et al. 1997; reviewed in Li and McSorley 2015), possibly by production of Chlamydia-specific IgA in genital secretions (Shillova et al. 2020). However, the preponderance of evidence suggests that the primary adaptive cell type responsible for defense against primary Chlamydia infection is the Th1 CD4 T cell.

Th1 cells differentiate in the draining lymph nodes in response to the cytokine IL-12 and develop into IFNγ-secreting effector cells (reviewed in Zhu and Paul 2010). Mouse models deficient for all T cells (Rank, Soderberg and Barron 1985), αβT cells (Perry, Feilzer and Caldwell 1997), CD4, MHC class II (Morrison, Feilzer and Tumas 1995), IL-12 (Perry, Feilzer and Caldwell 1997), or IFNγ (Rank et al. 1992; Cotter et al. 1997; Johansson et al. 1997; Perry, Feilzer and Caldwell 1997; Ito and Lyons 1999; Perry et al. 1999a) all demonstrate increased susceptibility to primary C. muridarum infection. Furthermore, adoptive transfer of Chlamydia-specific Th1 cells confers protection against C. muridarum in nude recipient mice (Igietseme et al. 1993), and against C. trachomatis infection in IFNγ-deficient recipient mice (Gondek, Roan and Starnbach 2009; Gondek et al. 2012). With respect to C. trachomatis, the IFNγ produced by Th1 cells is not necessary for clearance, as mice lacking all T and B cells are able to clear C. trachomatis infection spontaneously (Sturdevant and Caldwell 2014), demonstrating that the innate sources of IFNγ are sufficient to induce cell-autonomous destruction of C. trachomatis in epithelial cells and macrophages. Interestingly, mice deficient for IRG-mediated epithelial cell-autonomous immunity are able to spontaneously clear C. trachomatis infection after a 2-week delay, correlated with an increase in Th1 responses (Coers et al. 2011) indicating that sustained IFNγ production is able to drive clearance of Chlamydia infection via macrophages, neutrophils and lymphocytes. Likewise, as C. muridarum is able to evade the IRG system and persist beyond the early innate phases of the immune response, strong Th1 responses are required to successfully clear the infection by 28–45 dpi. Indeed, in mice deficient for TCRβ C. muridarum does not disseminate to other tissues but bacterial growth and shedding in the genital tract continues unchecked at 60 dpi (Mercado et al. 2020), confirming that T cell responses are required for clearance of C. muridarum infection in the genital tract even if they are dispensable for controlling C. muridarum dissemination.

There is little evidence supporting protective roles for other populations of T cells in primary Chlamydia infection in mouse models. Th2 cells normally function in immunity to helminth infections and play important roles in allergic responses (reviewed in Zhu and Paul 2010). Th17 cells are important for defense against extracellular pathogens, as the primary Th17 cytokine IL-17 largely functions to promote neutrophil recruitment and activity in infected tissue (Li, Casanova and Puel 2018). Adoptive transfer of Chlamydia-specific Th2- or Th17-polarized cells fails to confer any level of protection against C. trachomatis (Hawkins, Rank and Kelly 2002; Gondek, Roan and Starnbach 2009). Furthermore, transfer of Th2 cells actually worsens Chlamydia infection, likely by inhibiting protective Th1 responses (Andrew et al. 2013), and increased Th2 responses in IL-12-deficient mice are associated with increased rates of hydrosalpinx (Zhang et al. 2014b). Interestingly, IL17^−/− mice infected with C. muridarum demonstrate decreased and shortened bacterial shedding, correlated with decreased neutrophil and macrophage influx into the oviducts (Andrew et al. 2013). Another study showed that IL17ra^−/− mice demonstrate partially impaired Th1 immunity, but wildtype levels of bacterial burden in C. muridarum infection (Scurlock et al. 2011), demonstrating an indirect role for Th17 cells in supporting protective Th1 responses. On the whole, however, the evidence demonstrates that Th17 cells, similar to neutrophils, play a minor role in protection against primary female genital Chlamydia infection.

CD8 T cells support protective immunity via production of IFNγ

CD8 T cells are important in defense against intracellular bacterial and viral infections and function to produce cytokines such as IFNγ and tumor necrosis factor α (TNFα) as well as to identify and destroy infected host cells (reviewed in Zhang and Bevan 2011). CD8 T cells are indeed induced in genital Chlamydia infection and present in the genital tract (Roan and Starnbach 2006). It is therefore surprising that they appear to be largely dispensable for defense against primary Chlamydia infection. Adoptive transfer of Chlamydia-specific CD8 T cells has shown protection against dissemination in C. trachomatis infection (Starnbach, Bevan and Lampe 1994), and either mild protection when transferred to athymic nude mice (Igietseme et al. 1994) or no protection when transferred to naive wildtype mice (Su and Caldwell 1995) in genital C. muridarum infection. A more recent study demonstrated no difference in the course of C. muridarum genital infection or dissemination to gut tissues in CD8-deficient mice compared to wildtype mice (Xie et al. 2020). While CD8 T cells are able to lyse infected cells in vitro (Beatty and Stephens 1994), their protective function in vivo appears to be associated with or dependent upon the secretion of IFNγ (Igietseme et al. 1994; Starnbach, Bevan and Lampe 1994; Lampe et al. 1998) rather than cytotoxic activity (Perry et al. 1999b). They are therefore mostly redundant to Th1 cells and play only a minor supporting role in defense against primary Chlamydia infection.

It is clear that protection against female genital Chlamydia infection in mouse models depends primarily upon IFNγ. Innate IFNγ, likely produced by ILCs and/or γδT cells, provides sufficient stimulation to infected cells to clear C. trachomatis infection and assists in preventing dissemination in C. muridarum infection. After the first week post-infection, adaptive immunity takes over the role of producing IFNγ. Th1 cells produce the majority of IFNγ with minor support from CD8 T cells. Lymphocytes persist in the genital tract for weeks after inoculation and are able to spontaneously clear C. muridarum infection in wildtype mice after 28–45 days (Fig. 1).

Determinants and mechanisms of genital pathology in Chlamydia infection

Factors underlying the differences in pathogenicity of Chlamydia strains

Elimination of the bacteria is achieved via a multifaceted immune response, yet at the same time it is also the components of the immune system that largely mediate tissue damage and pathologic sequelae. Understanding this double-edged sword of the immune response to genital Chlamydia infection is therefore crucial to design treatments and vaccines that prevent pathology. A plethora of mouse studies targeting various molecular and cellular mechanisms of disease have shed light on this issue, although much remains to be understood. Chlamydia trachomatis serovar L2 causes LGV in human hosts, and while it comprises an excellent infection model in murine hosts to study the immune mechanisms that are important in limiting bacterial growth, it does not cause pronounced genital pathology in immunocompetent mice (Lijek et al. 2018). Mouse studies examining immunopathology have therefore generally employed C. muridarum or C. trachomatis serovar D, which mirror the disease pathology observed in the human female genital tract, with genital scarring leading to adhesions, hydrosalpinx, chronic pelvic inflammation and infertility (reviewed in Darville and Hiltke 2010).

In total, two main factors underlie these strain-specific differences in pathogenicity. First, C. muridarum undergoes sustained replication in the mouse host, resulting in a productive and persistent infection, while C. trachomatis is cleared within 14 dpi. The short-lived presence of proinflammatory bacterial ligands in C. trachomatis infection may simply be insufficient to induce a robustly pathogenic immune response; indeed, in mice treated with IFNγ-depleting antibodies, C. trachomatis serovar L2 establishes a prolonged high-burden infection and pathology ensues (Zhong et al. 1989). Furthermore, mice deficient for Chlamydia sensor TLR2 (Darville et al. 2003) or LPS sensor and inflammasome component caspase11 (Allen et al.2019) both demonstrate equivalent C. muridarum growth compared to wildtype mice but develop significantly less genital pathology. These findings suggest that the presence of bacterial ligands, regardless of the strain, drives pathogenic inflammation.

Second, there may be bacterial strain-specific factors that directly mediate tissue pathology independent of bacterial growth. While the genomes of human and rodent Chlamydia species share approximately 99% synteny, stark genomic diversity is found in a 50 kb ‘plasticity zone’ within the Chlamydia chromosome (Read et al. 2000) that has been the focus of much study. Within the plasticity zone, C. muridarum and C. trachomatis serovar D harbor genes encoding putative cytotoxins that bear resemblance to Clostridium toxin B. In contrast to C. muridarum and C. trachomatis serovar D, C. trachomatis serovar L2 does not express a functional cytotoxin (Belland et al. 2001; Carlson et al. 2004; Thalmann et al. 2010). Analysis of C. muridarum plasticity zone mutants demonstrated wildtype bacterial growth with decreased genital pathology (Rajaram et al. 2015), suggesting that the putative cytotoxin may underlie strain-specific differences in pathogenicity. Additionally, C. muridarum and many strains of C. trachomatis harbor a small plasmid that is associated with virulence in primate ocular models (Kari et al. 2011) and murine genital models (O'Connell et al. 2007; Frazer et al. 2011; Lei et al. 2014; Sigar et al. 2014). The plasmid contains the virulence factor Pgp3 that inhibits bacterial killing by cationic antimicrobial peptides (Hou et al. 2015, 2019), is essential for the establishment of persistence (Yang et al. 2020), comprises an immunodominant antigen in infected women (Chen et al. 2010a), and is required for C. muridarum-induced genital pathology in mice (Lei et al. 2014; Liu et al. 2014; Yang et al. 2020). Altogether, the capacity of a Chlamydia strain to induce genital pathology appears to depend upon its ability to persist in genital tissue and its intrinsic ability to induce tissue damage based on the presence of one or more putative toxins.

Neutrophils, Th17 cells and CD8 T cells are pivotal drivers of genital pathology

The combination of bacterial ligands, acting through immune recognition by TLR2, inflammasomes and possibly other sensors such as STING, as well as toxin-mediated tissue damage induce inflammatory responses that act as the main drivers of genital pathology in Chlamydia infection. A study comparing C. muridarum infection in 11 mouse strains revealed varying susceptibilities to hydrosalpinx that did not correlate strain-specific variation in bacterial burden (Chen et al. 2014), demonstrating that host-mediated responses to infection drive pathology. The various components of the immune response to genital Chlamydia infection execute their antimicrobial functions with distinct mechanisms that each have a different potential for collateral damage to host tissue. Genetic deletion or antibody-blocking and -depleting experiments have examined the relative contribution to pathology of many immune signaling molecules and cell types (Table 1).

Table 1.

Host factors implicated in pathogenesis in murine Chlamydia infection. Comprehensive summary of in vivo Chlamydia infection models that examined inflammation and pathology. Listed are the target host factors; a brief description of the model design; the Chlamydia strain and substrain used (when available); the relevant phenotypes of the listed model and the reference for the experiments.

Host factor	Model design	Chlamydia strain	Phenotype (compared to wt)	Citation
Cytokines
IL-1	IL1R −/−	C. muridarum (Nigg)	Increased bacterial burden, decreased pathology	Nagarajan et al. (2012)
	IL1Ra −/− (IL-1R antagonist)	C. muridarum (Nigg)	Decreased burden, enhanced pathology	Nagarajan et al. (2012)
	IL1A −/−	C. muridarum (Nigg)	Decreased pathology	Gyorke et al. (2020)
	IL1B −/−	C. muridarum (Nigg)	Decreased pathology	Prantner et al. (2009); Gyorke et al. (2020)
IL-6	IL6 −/−	C. muridarum (Nigg clone CMG13.32.1)	Increased burden, decreased pathology	Sun et al. (2017)
IL-17	IL17 −/−	C. muridarum (Weiss)	Decreased burden, decreased pathology	Andrew et al. (2013)
IL-12	IL12p35 −/−, IL12p40−/−	C. muridarum (Nigg)	Increased pathology	Chen et al. (2013)
Type I IFN	IFNAR −/−	C. muridarum (Nigg)	Decreased burden, decreased pathology	Nagarajan et al. (2008)
C5	C5 −/−	C. muridarum (Nigg)	Decreased pathology	Yang et al. (2014)
TNFα	TNFA −/-	C. muridarum (Nigg)	Decreased pathology	Murthy et al. (2011)
	TNFAR1 −/-	C. muridarum (Nigg)	Decreased pathology	Dong et al. (2014)
IFNγ	IFNG −/−	C. muridarum (Nigg 1942)	Increased pathology	Murthy et al. (2007); Scurlock et al. (2011)
Signaling molecules
MyD88	MyD88 −/−	C. muridarum (Nigg)	Increased uterine inflammation and hydrosalpinx	Chen et al. (2010b); Nagarajan et al. (2011)
STING	STING gt/gt	C. muridarum (Nigg)	Increased incidence of bilateral hydrosalpinx	Zhang et al. (2017)
TLR2	TLR2 −/−	C. muridarum (Nigg)	Decreased pathology	Darville et al. (2003)
TLR3	TLR3 −/−	C. muridarum (Nigg)	Increased burden, increased pathology	Carrasco et al. (2018)
CD28	CD28 −/−, CD80/86−/−	C. muridarum (Nigg)	Decreased cytokine production and pathology	Chen et al. (2009)
CD40	CD40 −/−, CD40L−/−	C. muridarum (Nigg)	Decreased IFNγ production, increased burden
Inflammasomes
Caspase 11	Casp11 −/−	C. muridarum (Nigg)	Increased burden, decreased pathology	Allen et al. (2019)
	Casp1/Casp11−/−	C. muridarum	Decreased pathology	Cheng et al. (2008)
	Casp1/Casp11−/−	C. trachomatis L2	Increased fertility in infected mice	Igietseme et al. (2013)
	Pan-Caspase inhibitor	C. trachomatis L2	Increased fertility in infected mice	Igietseme et al. (2013)
Cell effector molecules
MMPs	MMP chemical inhibitor	C. muridarum	Decreased ascension to upper genital tract, decreased pathology	Imtiaz et al. (2006)
MMP9	MMP9 −/−	C. muridarum (Weiss)	Decreased ascension to upper genital tract, decreased pathology	Imtiaz et al. (2007)
iNOS (RNS)	NOS2 −/−	C. muridarum (Weiss)	Increased pathology	Ramsey et al. (2001a)
Phagocyte oxidase	P47phox -/−	C. muridarum (Weiss—Ramsey et al.; Nigg3 clone CMG0.1.1–Dai et al.)	Decreased pathology	Ramsey et al. (2001b); Dai et al. (2016a)
Perforin	Perforin −/−	C. muridarum (Nigg)	Decreased pathology	Murthy et al. (2011)
TAP1	TAP1 −/−	C. muridarum (Nigg)	Decreased pathology	Murthy et al. (2011)
Chemokine receptors
CXCR2	CXCR2 −/−	C. muridarum (Weiss)	Decreased pathology	Lee et al. (2010)
CXCR3	CXCR3 −/−	C. muridarum (Nigg)	Delayed bacterial clearance, decreased pathology	Jiang et al. (2017)
	Anti-CXCR3 antibody depletion	C. trachomatis D	Decreased T cell infiltrate, decreased pathology	Xie et al. (2003)
CCR7	CCR7 −/−	C. muridarum (Weiss)	Decreased burden, decreased pathology	Li et al. (2017)
Cell types
Neutrophils	Anti-GR1 antibody depletion	C. trachomatis D	Decreased neutrophil and monocyte infiltrate, decreased pathology	Lijek et al. (2018)
	Anti-Ly6G antibody depletion	C. trachomatis D	Decreased neutrophil infiltrate, decreased pathology	Lijek et al. (2018)
	Mcl1-LysM-Cre (absence of mature neutrophils)	C. muridarum (03DC39)	Increased burden, decreased pathology	Zortel et al. (2018)
CD8+ T cells	CD8 −/−	C. muridarum (Nigg)	Decreased pathology	Murthy et al. (2011)
	Anti-CD8 antibody depletion	C. muridarum (Nigg—Murthy et al.) (Nigg3 clone G13.32.1–Yu et al.)	Decreased pathology	Murthy et al. (2011); Yu et al. (2019)
	OT-I (ovalbumin-specific CD8 T cells)	C. muridarum (Nigg)	Decreased pathology	Manam, Nicholson and Murthy (2013); Vlcek et al. (2016)
CD8-derived TNFα	Adoptive transfer of TNFA−/− CD8 T cells	C. muridarum (Nigg)	Decreased pathology compared to transfer of wt CD8+ T cells	Murthy et al. (2011)
CD4+ T cells	Anti-CD4 antibody depletion in CD8−/− mice	C. muridarum (Nigg3 clone G13.32.1)	Decreased pathology	Yu et al. (2019)
Regulatory T cells	Anti-CD25 antibody depletion	C. muridarum	Decreased pathology	Moore-Connors et al. (2013)

Relative to other components of innate immunity, neutrophils are by far the most implicated in inducing immunopathology. Chlamydia muridarum strain-specific induction of neutrophils (Frazer et al. 2011) as well as mouse strain-specific duration of early pyosalpinx (Zhang et al. 2014a) are associated with long-lasting hydrosalpinx. Depletion of neutrophils results in decreased pathology in genital infection with C. trachomatis (Lijek et al. 2018) or C. muridarum (Lee et al. 2010; Zortel et al. 2018). Mice deficient for certain neutrophil effectors such as MMP9 (Imtiaz et al. 2007) and phagocyte oxidase (Ramsey et al. 2001a; Dai et al. 2016b), as well as signaling mediated by neutrophil-derived cytokines such as TNFα (Murthy et al. 2011; Dong et al. 2014) and IL-1 (Nagarajan et al. 2012), all demonstrated reduced pathology in C. muridarum infection. IL-1 is produced in two forms: IL-1α is produced by epithelial and hematopoietic cells in Chlamydia-infected genital tracts (Gyorke et al. 2020) and IL-1β is produced by hematopoietic cells such as macrophages and neutrophils upon inflammasome activation (Prantner et al. 2009). Both forms of IL-1 act on the IL-1 receptor and function to potentiate pro-inflammatory cytokine programs that result in further recruitment of neutrophils and tissue damage (reviewed in Dinarello 2018). Indeed, mice lacking IL-1α, IL-1β or the IL-1 receptor demonstrate decreased neutrophil recruitment and genital scarring upon genital C. muridarum infection, with minimal effects on bacterial burden (Prantner et al. 2009; Nagarajan et al. 2012; Gyorke et al. 2020).

Neutrophils are unlikely to be the sole contributor to genital pathology, as pathology was only partially reduced in all of these studies. Macrophages, which have a prominent role in host-protective immunity, produce MMPs, TNFα (Kol et al. 1998) and IL-1β (Finethy et al. 2015; Allen et al. 2019) in Chlamydia infection, and therefore also have the potential to cause tissue damage. Further studies are warranted to identify the protective and pathogenic contributions of specific macrophage populations in genital Chlamydia infection.

Just as the various subsets of CD4 T cells play different roles in clearance of Chlamydia infection, they also likely make different contributions to pathogenesis. Th1 cells, through their production of IFNγ, limit genital bacterial growth and are essential for clearance of C. muridarum infection. Th1 cells therefore play an almost exclusively host-protective role. Th17 cells, on the other hand, appear to play a more pathogenic role. This should not be surprising, as their signature cytokine, IL-17, functions mainly through neutrophils, which are largely pathogenic. Indeed, IL17RA^−/− mice displayed decreased neutrophil influx in C. muridarum-infected genital tracts (Scurlock et al. 2011), and Th17-mediated immunity was correlated with increased genital pathology in a reinfection model using C. trachomatis serovar D (Vicetti Miguel et al. 2016). However, IL17^−/− mice infected with C. muridarum demonstrated no difference in C. muridarum bacterial burden but increased genital pathology at 35 dpi (Andrew et al. 2013), suggesting a protective role for IL-17. It has been proffered that Th17 cells support Th1 responses (Scurlock et al. 2011), and further studies are warranted to better characterize the roles of Th17 cells in Chlamydia infection outside of their role in neutrophil recruitment.

Whereas CD8 T cells play a modest and largely redundant protective role in Chlamydia infection via their production of IFNγ, a plethora of studies have demonstrated their potent pathogenicity. Both genetic deficiency of CD8 or antibody-mediated depletion of CD8 T cells in wildtype mice reduced genital pathology in C. muridarum infection. Adoptive transfer of wildtype CD8 T cells into CD8^−/− mice restored pathology (Murthy et al. 2011). Chlamydia muridarum infection in OT-I transgenic mice, in which all CD8 T cells are specific for ovalbumin, induced no pathology (Manam, Nicholson and Murthy 2013), yet adoptive transfer of wildtype CD8 T cells again restored pathology in these mice (Vlcek et al. 2016), implying that Chlamydia-specific CD8 T cells are required for pathology. Additional studies indicated that CD8 T cells both produce and respond to TNFα during the immunopathogenesis of genital C. muridarum infections. Whereas TNFA^−/− mice displayed no pathology in C. muridarum infection, transfer of wildtype CD8 T cells restored pathology; further, TNFA^−/− CD8 T cells did not restore pathology in CD8^−/− recipient mice (Murthy et al. 2011), indicating that production of TNFα by CD8 T cell is important for pathogenesis. Notably, CD8 T cells were required to express TNFR2 (tumor necrosis factor receptor 2), in order to promote pathology in C. muridarum-infected mice. Because TNFα-producing, Chlamydia-specific wildtype CD8 T cells were unable to restore pathology in TNFR1^−/− mice (Manam et al. 2015), these data in aggregate suggest that TNFR2 signaling in CD8 T cells and TNFR1 signaling in other cell types are responsible for TNFα-mediated pathology associated with C. muridarum infections. Interestingly, depletion of CD8⁺ T cells from OT-I mice resulted in an increase in pathology in C. muridarum infection, and adoptive transfer of OT-I CD8⁺ T cells suppressed pathology in recipient wildtype or CD8^−/− mice (Xie et al. 2020). These findings suggest that, while Chlamydia-specific CD8+ T cells certainly promote pathogenesis, non-specific CD8⁺ T cells may paradoxically suppress pathology. Further studies are necessary to elucidate the underlying mechanisms.

Intestinal colonization promotes immunopathogenesis of genital Chlamydia infection

While components of both innate and adaptive immunity can contribute to tissue damage and subsequent scarring, recent studies have implicated immune reactions in distant tissues in genital pathogenesis. Chlamydia muridarum and C. trachomatis serovars D–K—the serovars most responsible for genital pathology in their respective host species—are not known to cause gastrointestinal illness in immunocompetent mice or humans, but can colonize the gut in both species (Yeruva et al. 2013; Gratrix et al. 2014, 2015, 2016; Peters et al. 2014; Craig et al. 2015; Musil et al. 2016). Indeed, C. muridarum inoculated via the genital tract can spread hematogenously to the mouse gut (Zhang et al. 2015; Dai et al. 2016a; Wang et al. 2016). Recent studies have demonstrated that genital infection with an attenuated C. muridarum mutant deficient for gut colonization is unable to induce hydrosalpinx unless wildtype C. muridarum is inoculated intragastrically (Tian et al. 2020). Since intragastric inoculation of wildtype C. muridarum does not result in dissemination to the genital tract or genital pathology on its own (Wang et al. 2016), these findings suggest that some component of the immune system is educated in the gut to generate a pathogenic immune response in the Chlamydia-infected genital tract, giving rise to a ‘two-hit’ model for Chlamydial pathogenesis (Zhong 2018). The nature of this mysterious gut-genital immune axis and its relevance to human disease remains to be fully elucidated.

Relevance to human infection

Genetic variants in specific immune pathways identified in humans, as well as observed immune responses, have been shown to correlate with either protection from infection or susceptibility to pathology, and the conclusions from these human cohort studies largely agree with the findings from more tractable mouse models. For example, polymorphisms in bacteria-sensing pattern-recognition receptors correlate with increased rates of tubal factor infertility in women with prior C. trachomatis exposure (den Hartog et al. 2006), cervical cell production of IFNγ and IL-12 positively correlates with fertility in C. trachomatis seropositive women, and production of IL-1β, IL-6 and IL-8 correlates with infertility (Agrawal et al. 2009). Moreover, cervical mucus neutrophils (Peipert et al. 1996) and markers of neutrophil activation correlate with endometritis (Wiesenfeld et al. 2002) in C. trachomatis-infected women. Unsurprisingly, robust T cell responses to Chlamydia antigens correlate with tubal factor infertility in women with prior C. trachomatis infection (Kinnunen et al. 2003; Tiitinen et al. 2006; Srivastava et al. 2008). However, strong IFNγ responses in T cells stimulated with Chlamydia antigens correlate with protection against C. trachomatis infection (Cohen et al. 2005), and conversely weak IFNγ T cell responses correlate with increased susceptibility (Debattista et al. 2002). Polymorphisms in TNFA are correlated with increased rates of tubal factor infertility in women who had been infected with C. trachomatis (Ohman et al. 2009). Finally, polymorphisms in HLA-A31, an HLA allele corresponding to CD8 T cell activation, are also correlated with C. trachomatis pelvic inflammatory disease (Kimani et al. 1996). These correlative studies have demonstrated that, for the most part, the different mechanisms of the human immune response to genital Chlamydia infection play similarly protective or pathogenic roles as in the mouse system.

CONCLUSION

The inflammatory response to genital Chlamydia infection is truly a double-edged sword: it is required for clearance of the infection but also induces tissue pathology that can cause infertility. The wealth of studies targeting individual components of the immune response have demonstrated that IFNγ-mediated responses, including cell-autonomous immunity and Th1 cells, are protective, whereas neutrophils, Th17 cells and CD8 T cells are largely dispensable for immunity and instead contribute to tissue pathology. After disruption of the epithelial barrier due to neutrophil-derived MMPs or reactive oxygen species, or CD8 T cell-derived TNFα, wound-healing fibrogenic responses are required to repair the damage, resulting in scarred genital tissue and disease sequelae (Fig. 2). While the murine host demonstrates slightly different immune characteristics compared to humans, and while human Chlamydia species have undoubtedly evolved adaptions specifically tailored for colonization of their natural host, mouse models of infection are nonetheless highly informative, shaping our understanding of the mammalian immune response to Chlamydia infections and providing insights for the development of human vaccines.

Figure 2.

Course of immune responses during murine genital Chlamydia infection. (A) Infectious elementary bodies (EBs) ascend through the cervix into the uterine horns and potentially to the oviducts and ovaries. EBs enter epithelial cells, establish membrane-bound inclusion bodies, and differentiate into replicative reticular bodies. (B) Bacteria are sensed by epithelial cell pattern recognition receptors, and epithelial cells produce cytokines and chemokines, including IL-1α, IL-6 and IL-8, to activate nearby immune cells. (C) Chlamydia extrude inclusions from infected cells or lyse infected cells, as bacteria spread to neighboring cells. (D) Tissue-resident immune cells (such as ILCs or γδT cells) produce IFNγ, which activates epithelial cell-autonomous immunity. C. trachomatis is effectively cleared by cell-autonomous immunity, but C. muridarum infection persists. (E) Innate immune cells such as macrophages and neutrophils accumulate in the genital tract, engulfing and destroying extracellular bacteria. Neutrophils comprise the most prominent source of matrix metalloproteases (MMPs) and reactive oxygen species, which contribute to tissue damage. (F) Chlamydia-specific lymphocytes traffic to the genital mucosa. Th1 cells sustain production of IFNγ, driving clearance of Chlamydia infection. Th17 cells produce IL-17, strengthening pathogenic neutrophil responses. CD8⁺ T cells produce TNFα, further driving tissue damage. (G) Wound-healing responses repair damaged tissue, resulting in pathogenic fibrosis and genital scarring. Figure created with Biorender.com.

Conflicts of Interest

None declared.