Chlamydia trachomatis Persistence in Vitro – An Overview

Abstract

Chlamydiae growing in target mucosal human epithelial cells in vitro can transition from their normal developmental cycle progression -- alternating between infectious but metabolically-inactive elementary bodies (EB) to metabolically-active but non-infectious reticulate bodies (RB) and back to EB -- into a state of “persistence”. Persistence in vitro is defined as “viable but non-cultivable chlamydiae” involving “morphologically enlarged, aberrant, non-dividing RB”. The condition is “reversible” to yield infectious EB on removal of the inducers, including penicillin, interferon-gamma, iron/nutrient starvation, concomitant herpes infection, or maturation of the host cell into its physiologically-differentiated state. All aberrant RB phenotypes are not the same due to differing up/down regulated chlamydial gene sets and subsequent host responses. While all persistence-inducing conditions exist in vivo, key questions include whether or not (i) aberrant chlamydial RB occur in vivo during the alternating acute-silent chronic-acute chlamydial infection scenario that exists in infected patients and animals and (ii) if such aberrant RB can contribute to prolonged, chronic inflammation, fibrosis and scarring .

Introduction – The Chlamydial Developmental Cycle

Genital strains of C. trachomatis serovariants (serovars) D–K target superficial mucosal epithelia of the urethra or the endocervix for initiating their infectious process. This process requires two distinct morphological forms of the bacterium – the elementary body or EB and the reticulate body or RB. The small (ca. 0.2 – 0.3µm), dense infectious EB make contact with the epithelial cell surface and, following ligand-receptor interaction, the EB are endocytosed (Fig. 1A). Rapid modification of the EB-containing endocytic vesicle by the EB allows escape of the EB endosome from the endocytic-lysosomal pathway and trafficking on cytoskeletal intermediate filaments to the nuclear hof/endoplasmic reticulum/Golgi activity center. After arrival at this destination, transformation of the essentially non-metabolically active EB into the larger (ca. 0.8µm), metabolically-active RB is triggered. The compacted EB DNA is relaxed, signals for DNA, RNA and protein synthesis are activated and RB cell division ensues (Fig. 1B). The RB-containing endosome intercepts trans-Golgi vesicles for endosome membrane expansion to accommodate the increasing number of RB progeny. This now microscopically visible intracellular membrane-bound chlamydial microcolony is termed an inclusion. After several hours of continued logarithmic growth and expansion of the inclusion (Fig.1C), depletion of nutrients and ATP scavenged by the growing RB from the infected host cell signal the maturation of the non-infectious RB back into the infectious EB (Fig. 1D). EB released from the extruded inclusion and/or lysis of the host cell can then infect neighboring epithelial cells to perpetuate the infectious process [1,2].

Figure 1.

Illustrated stages of the genital C. trachomatis developmental cycle in infected human endometrial epithelial cells. A. Attachment and entry of the small, infectious EB at 1 hpi. B. Early inclusion (8–12 hpi) containing a few larger, metabolically-active RB. C. Larger inclusion at mid-developmental cycle containing many RB and a few intermediate bodies (IB) – transitional forms for maturation into infectious EB. D. Maturing late developmental cycle inclusion containing more infectious EB and some RB. Epithelial cells infected with genital C. trachomatis serovars form only 1 large inclusion. E. An example of multiple (4) inclusions of C. suis (or C. psittaci or C. pneumoniae) per single infected cell. A,B -- Bar = 2 µm at 7,000 × magnification; C,D – Bar = 2 µm at 7,000 × mag., E – Bar = 2 µm at 4,200 × mag.

This in vitro characterized developmental cycle is fundamental to all Chlamydia and varies only in timing, from 48 – 72 – 96 hours, depending on the species, and the number of inclusions per host cell – from one in a C. trachomatis-infected cell to several inclusions per C. pneumoniae – or C. psittaci-infected cell (Fig. 1E). Stages of this developmental cycle have also been visualized in vivo by fluorescence microscopy and electron microscopy in tissue samples collected from patients and experimentally infected animals [3], respectively.

Discovery of a C. trachomatis Persistence Mode In Vitro

Historically, C. trachomatis was associated with post-gonococcal urethritis or cervicitis (PGU/PGC) or non-gonococcal urethritis/cervicitis (NGU/NGC). In the early 1960 –1970s, patients with gonococcal urethritis or cervicitis were treated with penicillin, prior to N. gonorrhoeae acquiring penicillin resistance. However, approximately 3 months later, some of these patients would present with PGU/PGC, from whom C. trachomatis was isolated [4,5].

Manire and Matsumoto [6] exposed C. psittaci Cal-10 –infected L929 cells in vitro to 200µ/ml penicillin 1 hr post-infection (hpi). EB to RB transformation proceeded but the developmental cycle was apparently stalled as no new infectious EB progeny could be recovered. Transmission electron photomicrographs revealed the normal RB continued to grow but could not undergo binary fission; hence, enlarged, morphologically aberrant RB were visible. Once the penicillin was removed and after a lag period of RB internal and external budding, the RB returned to normal size, underwent cell division and matured into infectious EB. These morphological changes are illustrated in Figure 2 with C. trachomatis-infected endometrial (HEC-1B) cells exposed to 10u/ml penicillin.

Figure 2.

Penicillin-induced aberrant RB in C. trachomatis serovar E-infected human endometrial epithelial cells cultured in vitro. A. Standard inclusion containing RB and EB in a C. trachomatis-infected HEC-1B cell control at 24 hpi, not exposed to penicillin. B,C. C. trachomatis-infected HEC-1B cells at 36 hpi, previously exposed to Penicillin G (20 µ/ml) at 1 hpi. The RB are swollen, abnormal and non-dividing (RB*). The altered metabolic disconnect between transcription/translation in abnormally enlarged, non-dividing RB and cell envelope component biosynthesis results in excessive shedding of RB outer envelope blebs into the inclusion lumen; these antigen-containing (MOMP, LPS, etc.) blebbed vesicles somehow fuse with or cross the inclusion membrane and emerge into the cytoplasm of the infected host cell (arrowheads). D. The C. trachomatis-infected HEC-1B cells pre-exposed to Penicillin G (20 µ/ml) for 35 hrs, followed by removal of the antibiotic and continued cultivation of the infected cells for an additional 12 hrs. The appearance of miniature bodies is one morphological signal of recovering RB and their reprogrammed attempt to undergo cell division to produce progeny. A,B,C – Bar = 10 µm at 2,900 × mag., D. – Bar is 2 µm at 10,000 × mag.

Studies of these events utilizing transcriptomics and proteomics have confirmed there is continued genome replication and mRNA synthesis, in the aberrantly enlarging RBs, though cell division is altered [7]. Further, Galasso and Manire [8] demonstrated that C. psittaci cultured in HeLa cells in vitro in the presence of penicillin and antiserum for 6–9 months maintained a viable, persistent infection which, on removal of the antibiotics and antibodies, was reversible and resulted in an acute infection in the tissue cultured cells with recoverable infectious EB.

Concomitant with or soon thereafter, a flurry of additional experimental conditions were shown to reproduce the viable but non-dividing altered RB phenotype. The inducers included interferon gamma (IFNγ), nutrient/amino acid starvation, iron depravation, the host cell differentiation state and herpes virus infection. In the final analysis, the findings by many different researchers using differing experimental conditions and different chlamydial species were the same as described by Manire and Matsumoto [6]. Thus, it is now generally accepted by investigators in the field that there exists in vitro a state of “chlamydial persistence” defined by “viable but non-cultivable chlamydiae” involving “morphologically enlarged, non-dividing, aberrant RB” which is “reversible” to yield infectious EB.

Since there are two elegant and detailed reviews on chlamydial persistence by Beatty et al. [9] and Hogan et al. [10] covering the extensive literature on the subject over a 40 year period, here there will follow only a brief summary of the subject of chlamydial persistence and the aberrant RB phenotype in vitro in order to focus on the key questions of the Chlamydia Immunology and Control Expert Advisory Meeting, i.e. what definitive evidence is needed to confirm whether or not C. trachomatis aberrant RB persistence can or does occur in vivo and, if so, what might be the consequences? My apologies to colleagues for studies not acknowledged due to page limitations and the fact that the focus of this meeting was human C. trachomatis infections.

Other Inducers of C. trachomatis Persistent Aberrant RB In Vitro

(i) IFNγ and Amino Acid Starvation

At the same time the early penicillin studies were being carried out, other Chlamydia researchers discovered the growth-inhibiting effects of immunological mediators on chlamydiae-infected cells, the most notable being the cytokine IFNγ [11,12]. Clever studies by Byrne et al. [13] revealed the cytokine effect to be on the host cell which, in turn, affected chlamydial growth modulation. In human epithelial cells, IFNγ induces indoelamine-2,3-dioxygenase, an enzyme that catalyzes the degradation of tryptophan – an essential amino acid for both eukaryotic cells and chlamydiae. However, addition of exogenous indole, in the presence of IFNγ, results in a return to the normal chlamydial developmental cycle; this rescue event is evident for genital strains only [14]. Interestingly, genital serovars possess functional tryptophan synthease (trpBA) genes to convert indole, secreted by local vaginal flora, into tryptophan, which allows these chlamydiae to escape killing by IFNγ and establish a persistent infection; ocular serovars have mutations in the Trp operon that inactivate the synthease enzyme, making them incapable of using endogenous indole to synthesize tryptophan. Subsequently, Beatty et al. [15] demonstrated that exposure of C. trachomatis-infected cells to physiological concentrations of IFNγ resulted in enlarged, aberrant RB. Moreover, though these altered RB were deficient in the cell envelope structural components, major outer membrane protein (MOMP) and lipopolysaccharide (LPS), they continued to synthesize and secrete HSP-60, an immuno-destructive antigen associated with the chronic inflammatory fallopian tube pathology associated with pelvic inflammatory disease, ectopic pregnancy, and infertility [16]. The latter finding also raised the possibility that the persistent aberrant RB state was induced as a stress response to nutrient starvation.

Complete transcriptome analysis by Belland et al. [2] of C. trachomatis serovar D growth in HeLa cells exposed to IFNγ demonstrated the up-regulation of many genes involved in active metabolic processes in the aberrant RB, including those involved in DNA repair and recombination, protein translation, and phospholipid utilization. Additionally, stress response gene up-regulation was observed as was down-regulation in the genes involved in proteolysis, peptide transport, cell division, RB to EB redifferentiation and compaction of EB DNA (histone genes). The authors’ interpretation was that the chlamydial response to IFNγ-tryptophan nutrient limitation had evolved as a coordinated transcriptional response to control the transition between the classic EB-RB-EB active growth and the aberrant RB persistent growth states, and that the latter represented an alternative life style used by chlamydiae to avoid the host immune response. This premise has been challenged on the basis of in vitro experiments showing that cytotoxic T lymphocytes harvested from the spleens of mice previously immunized with C. trachomatis were able to recognize and kill L-cells infected with aberrant RB following exposure of the cells to IFNγ or penicillin (17).

Since chlamydiae are obligate intracellular bacteria, an available supply of soluble nutrients from the host cell cytoplasm is essential for productive chlamydial growth. Hence, by addition of the eukaryotic protein synthesis inhibitor cycloheximide to chlamydiae-infected coverslip cultures, there is less competition by the host cell for nutrient pools and chlamydial inclusions are considerably larger and easier to detect and count. Early studies by Moulder [18] growing C. psittaci 6BC in fibroblasts in alternating nutrient rich medium 199 versus Eagle’s Minimal Essential Medium led to overt or covert “cryptic body” chronic infections, respectively. Extensive studies by the Pearce laboratory [19] examined the growth of C. trachomatis serovar L2 in McCoy cells in Eagles Minimal Essential Medium containing 0–100% amino acid levels. When the medium amino acid pools were reduced to 10%, chlamydial inclusions at 48 hpi were filled with swollen RB and irregular budding RB. A return to productive infection and maturation of RB to EB could be accomplished by addition of cysteine or L-isoleucine.

(ii) Iron Deprivation

Virtually all prokaryotes have an absolute requirement for iron or related translational metals for survival; intracellular bacterial pathogens require iron for entry and replication in mammalian cells, and chlamydiae are no exception. Addition of the iron-chelating chemical desferal to the medium bathing C. trachomatis serovar E-infected polarized endometrial epithelial HEC-1B cells led to a distinctive form of persistent infection with small-sized inclusions containing enlarged, aberrant, non-dividing RB, with loose, wavy outer envelope membranes. Reversal to productive infection and recovery of infectious EB could be achieved with the removal of desferal and supplementation of the culture with iron-saturated transferrin [20]. Given that C. trachomatis infections in the genital tract are most prevalent in young females during their peak reproductive years [21] when iron concentrations fluctuate considerably from iron-sufficient to iron-deficient during the menstrual cycle, modulated in particular by estrogen, it is possible to envision chronic chlamydial infections in these women due to alternating aberrant RB persistent/productive infection episodes.

(iii) Host-cell Differentiation State

In the early 1980s, infection of human monocytes with C. psittaci resulted in inclusions containing enlarged abnormal RB; only after the monocytes matured into macrophages did productive chlamydial growth of normal RB and EB resume [22]. Reproducible results of ultrastructurally abnormal RB have also been reported by one group of investigators comparing C. trachomatis serovar K growth in human monocytes and synovial tissues from patients with chlamydial arthritis with laryngeal Hep-2 cells as controls [23,24]. All 11 gene groups, representing glycolysis, pentose phosphate pathway, TCA cycle, electron transport, sigma factors and DNA replication were up-regulated in the active stages of growth; in aberrant RB phenotype infection, chromosome replication continued but cytokinesis was severely down-regulated.

In 1985, Shirley Richmond proposed that chlamydiae survive in a persistent aberrant RB phenotypic form in the undifferentiated epithelial cells at the base of endometrial glandular cells [25]. As the epithelial cells migrate up the glands, differentiation into the productive form of infection is achieved as the epithelial cells reach and repopulate the mucosal surface. Partial support for this proposal has been provided by Guseva et.al. [26] using pig reproductive tissues cultured ex vivo. C. suis produces aberrant RB/persistent-like inclusions and few infectious chlamydiae within infected, primary swine glandular epithelial cells but follows a more normal developmental sequence in more differentiated, swine luminal epithelial cells. Productive replication in luminal cells damages the epithelium and serves as a source of released infectious EB, while aberrant RB phenotype forms in glandular tissue serve as a “silent” reservoir for continued infection. Another example of this phenomenon is illustrated in Figure 3, which reveals by transmission electron microscopy the asynchronous inclusion development as well as aberrant RB in inclusions in C. trachomatis-infected glandular epithelial cells derived from the ductal breast carcinoma cell line MCF-7. Host cell differentiation-dependent replication of obligate intracellular parasites is not without precedence; human papilloma virus is highly dependent upon basal cell differentiation into keratinized squamous epithelial cells for productive replication while persistent host infection is maintained by chronic viral episome infection of the basal cell layer.

Figure 3.

Asynchronous inclusion development and aberrant RB in C. trachomatis E-infected glandular epithelial cells derived from ductal breast carcinoma MCF-7 cells cultured in vitro. N denotes normal RB and EB in a typical inclusion; P denotes morphologically enlarged, aberrant RB in other inclusions in juxtaposed epithelial cells. Bar = 2 µm at 7,000 × mag.

(iv) Herpes Simplex Infection

A somewhat surprising culture condition was recently reported to induce typical chlamydial persistence. Infection of HeLa cells 24 hours after inoculation with C. trachomatis serovar E with herpes simplex virus type 2 (HSV2) resulted in swollen, aberrantly-shaped, electron-translucent RB 20 hours later [27]. As previously reported with IFNγ exposure, chlamydial MOMP was decreased while HSP60-1 production was increased. However, new data suggest that co-infection of the same cell by chlamydiae and HSV2 is not necessary to trigger a chlamydial persistent aberrant RB phenotype. Mere contact of chlamydiae-infected epithelial cells with UV-inactivated virions alone or with surface antigens on fixed HSV-singly-infected cells activates a host cellular response resulting in chlamydial persistence, again with increased HSP60 production. Eventual degradation of the defective HSV virions by the host epithelial cell results in reversal of chlamydial persistence and a return of productive chlamydial infection [28]. Since there is considerable data on HSV2 infection of endocervical epithelial cells, especially in PAP smears, there are a number of ways that C. trachomatis-infected cells could come in contact with the appropriate viral ligands in the absence of cellular co-infection in vivo. First, because HSV infection produces 50–200 defective viral particles/plaque forming unit (PFU), cellular co-infection with defective virions is likely more frequent than that with replication-competent HSV. Second, viral glycoproteins released from infected cells could induce persistence by interacting with receptors on chlamydiae-infected epithelial cells. Third, contact between C. trachomatis-infected cells and viral glycoproteins on the surface of neighboring, HSV-infected cells or on infiltrating, HSV-abortively-infected monocytes or T cell could produce this effect as well.

Are All Chlamydia Aberrant RB the Same?

Extensive comparative genomic, transcriptomic and proteomic analyses have been applied recently to C. trachomatis [29], C. pneumoniae [30] and C. psittaci [31] growing under normal versus aberrant RB persistent conditions induced by penicillin, IFNγ or iron-deprivation. The bottom line is that while the phenotypic criteria of morphologically enlarged, aberrant, non-dividing RB, which are viable but non-cultivable, remain similar, emerging transcriptional profiles do show significant differences in up-regulated or down-regulated gene profiles between the different persistence models. Due to the complexity of the data on this subject, the reader is referred to the Tables and Discussion in the Hogan et al. [10] review. However, one example will be given. Beatty et al. [14] found in IFNγ-induced aberrant C. trachomatis RB a reduced production of MOMP and an increased production of HSP60, whereas Nicholson and Stephens [7] found no change in the ompA/groEL gene transcripts in penicillin-induced aberrant C. trachomatis RB. Conversely, Matthews et al. [32] detected up-regulation of the ompA gene encoding MOMP in IFNγ-induced aberrant C. pneumoniae RB. Further, Gerard et al. [33] reported that HSP60-1/groEL was expressed predominantly in acute phase growth of C. trachomatis serovar K and that it was the HSP60-copy 2/CT604 gene transcript/protein that was increased in iron-induced persistent cultures. The latter finding was confirmed by LaRue and Raulston et al. [34] in iron-deprived cultures of C. trachomatis serovar E aberrant RB. Thus, the current consensus is that the transcriptional response of the chlamydiae differs according to the persistence inducing stimuli and furthermore, that each of these likely generates a different host cell response.

The rationale for such comparisons is to find genes that are consistently specific for the aberrant RB phenotype in chlamydial infection in vitro in order to extend these studies successfully to patient tissues. The ultimate goal is to determine whether or not the aberrant RB phenotype exists in chlamydiae-infected tissues in vivo and, if possible, determine the inducing mechanism(s). Leading candidate gene sets for in vivo investigations may be up-regulation of the early gene euo with down-regulation of the hctB gene. The latter encodes the HC-2 histone protein that modulates/compacts chlamydial DNA for RB-to-EB re-differentiation for production of infectious EB.

Does the Chlamydia Aberrant RB Phenotype Exist In Vivo and, If So, What May Be the Consequences?

Historically, latent chronic chlamydial infections in birds and animals were well known to early chlamydia investigators. In a few cases, presumably resolved chlamydial infections could be reactivated by immunosuppressive agents, such as cyclophosphomide. Clinical conditions that could be ascribed to persistent in vivo infections with very low levels or even nonexistent elementary body production could include: (i) asymptomatic urethritis in males and cervicitis in females; (ii) females with silent pelvic inflammatory disease; (iii) culture-positive individuals treated with antibiotics, who seemed to resolve the infection, but in whom reactivation or re-isolation occurred weeks or months later with the same genotype despite denials of re-exposure; and (iv) subjects with culture-negative but strong serological titers and epidemiological associations. Again, the readers are referred to the erudite reviews of these issues by Beatty et al. [9] and Hogan et al. [10].

It should be noted that the use of the term “persistence” to describe the in vitro phenomenon has created great confusion in the literature as the same term is used to describe clinically silent, chronic infection in humans and animals. In the context of in vivo chlamydial infection, “persistence” does not necessarily connote the aberrant RB phenotype identified in in vitro studies as described above. In fact, little is known of the biologic mechanisms used by chlamydiae to persist in vivo. To avoid confusion in this paper, the term “aberrant RB phenotype” rather than “persistence” has been used to refer to the in vitro defined phenomenon.

While many researchers strongly suspect that the insidious nature of some chlamydial infections and the chronic inflammatory consequences of genital reproductive disease sequelae are due, in part, to phenotypically aberrant chlamydiae, the collective body of data is still circumstantial and the topic is controversial [35]. Barriors to research into this subject include the following: infections due to the aberrant RB phenotype are not detectable by culture; detection of the presence of chlamydial DNA and/or antigen in diseased tissues of patients may result from delayed nucleic acid clearance following successful eradication of viable organisms; ultrastructural evidence of abnormal pleomorphic RB in macrophages from chlamydiae-infected patient tissues is suggestive but, in isolation, not proof; and high titers of antibodies to the chlamydia immuno-destructive HSP60 antigen in a group of high risk individuals is not necessarily synonymous with current chronic infection.

How to detect the existence of aberrant RB phenotype infections in vivo was one of the key questions posed to the workshop participants and will be a challenge to clinicians, investigators using experimental animal models and researchers involved in translational studies. One possibility might be to collect chlamydiae-positive only endocervical and endometrial biopsy specimens, survey tissues for morphologically aberrant or normal-appearing inclusions and undertake extensive, painstaking analyses of chlamydial inclusions by methods such as electron microscopy, confocal microscopy and molecular studies of isolated inclusions. All such data could be placed in the context of clinical findings and data collected from patients, such as demographics, sexual risk factors, past STIs and antibiotic treatment history. Non-invasive approaches such as use of cytobrushes for sampling organisms and cells from the genital tract, as pioneered by the Quayle laboratory to track changes in the T cell repertoire during chlamydial infection, might also provide valuable information on chlamydial growth characteristics and evidence for in vivo persistence [36]. However, it should be remembered that chlamydia-specific stimulation of adaptive immune response effectors can be indicative of either viable organisms or the presence of chlamydial antigen.

All of the aforementioned inducers of the chlamydial aberrant RB phenotype in vitro are likely to function in vivo at physiological concentrations. Penicillin, amoxicillin and ampicillin, all three of which can induce the chlamydial aberrant RB phenotype in vitro (unpublished data), are widely used in the United States with 68.7 million prescriptions dispensed in 2005 [www.imshealth.com/ims/portal/front/articleC/0,2777,6599_73914140_73916014,00.html]. In fact, there are two reports that claim that penicillin-induced aberrant RB are more refractory to killing via azithromycin [37,38] due to their slowed metabolism. These observations support the case for the aberrant RB phenotype as one of the mechanisms, if not the most important mechanism, utilized by chlamydiae to persist in vivo.

Finally, another key question related to the potential consequences of the chlamydia aberrant RB phenotype in vivo is the nature of the immune response triggered by aberrant RB-infected versus acutely-infected epithelial cells. Can the aberrant RB chlamydiae state (i.e. swollen, undividing but viable RBs) actually result in a protracted, chronic infection that leads to immune-mediated damaging sequelae? At the research bench, experiments can be designed to ask if different chemokine signals are generated in acutely- and aberrant RB-infected endocervical and endometrial cells exposed in vitro to penicillin versus IFNγ versus iron-deprivation and whether the signals are anti-inflammatory (IL-11) or pro-inflammatory (IL-8). Actually, aberrant RB induced by penicillin may be the better model in this case since both IFNγ and iron-deprivation have a definite effect on epithelial cells that may confound the data. Our fundamental understanding of the immunological consequences of acute and damaging chlamydial infection has been generated in experimental animal models and these contributions have been invaluable [39–41]. In this current supplemental issue, Drs. Darville and Hiltke will review the pathogenic immune responses to acute chronic infection and Drs. Rank and Whittum-Hudson will review the development of protective immunity to Chlamydia infection in animal models and the underlying effector immune mechanisms.

In summary, the Chlamydia Immunology and Control Expert Advisory Meeting, sponsored by the Centers for Disease Control and Prevention, provided an appropriate venue for an in depth look at how far we have come in understanding chlamydial infections, with an emphasis on genital C. trachomatis infections, and importantly, where we need to focus our immediate future research directions. The timing is right for a shift to more translational studies on tissues from chlamydiae-infected patients. While the complexities and limitations of human studies are not trivial, well-designed prospective protocols and analyses will yield valuable information. Support for such efforts is a must.