Metabolic dormancy in Chlamydia trachomatis treated with different antibiotics

ABSTRACT

Diseases caused by Chlamydia spp. are often associated with persistent infections. Chlamydial persistence is commonly associated with a unique non-infectious intracellular developmental form, termed an aberrant form. Although infectious chlamydiae can be cultured consistently in cells stressed to aberrancy, their role in persistence is not clear. Recovery from antibiotic stress was explored as a model to determine how survival of non-aberrant chlamydiae, in the presence of fully inhibitory drug concentrations, may participate in persistence. Assays included incubation in quinolones, tetracyclines, or chloramphenicol for differing lengths of time, followed by an extended recovery period in antibiotic-free media. Culturable elementary bodies were not detected during treatment with each antibiotic, but viable and culturable Chlamydia trachomatis emerged after the drug was removed. Time-lapse imaging of live, antibiotic-treated infected cells identified metabolically dormant developmental forms within cells that emerged to form typical productive inclusions. The effects of the increasing concentration of most tested antibiotics led to predictable inhibitory activity, in which the survival rate decreased with increasing drug concentration. In contrast, in fluoroquinolone-treated cells, there was a paradoxical increase in productive development that was directly correlated with drug concentration and inversely associated with aberrant form production. This model system uncovers a unique chlamydial persistence pathway that does not involve the chlamydial aberrant form. The association between productive latency and metabolic dormancy is consistent with models for many bacterial species and may lead to a different interpretation of mechanisms of chlamydial persistence in patients.

IMPORTANCE

The life history of most pathogens within the genus Chlamydia relies on lengthy persistence in the host. The most generally accepted model for Chlamydia spp. persistence involves an unusual developmental stage, termed the aberrant form, which arises during conditions that mimic a stressful host environment. In this work, we provide an alternate model for chlamydial persistence in the face of antibiotic stress. This model may be relevant to antibiotic treatment failures in patients infected with C. trachomatis.

INTRODUCTION

The obligate intracellular bacterium Chlamydia trachomatis is among the most prevalent of human pathogens and is associated with a disease of the genital tract and eye (1, 2). Other Chlamydia spp. cause diseases in humans and a wide variety of different animal species. Some of these infections are zoonotic (3, 4). Vaccination against chlamydial disease in humans remains elusive, and vaccines against diseases of veterinary significance are less than adequate (5–7). While antibiotic therapy is effective against most infections, treatment failures occur and the reasons for this are unclear.

A major aspect of chlamydial natural history is persistence within patients, regardless of the host species or specific chlamydial pathogen (8, 9). Chlamydial persistence is important in human chlamydial disease, as immune processes that target persistently infected cells lead to deleterious responses to the host tissue (5). Persistence may also be a factor in antichlamydial treatment failures (10).

All chlamydiae undergo a developmental cycle with both dormant and active cell forms (11). Elementary bodies (EBs) are minimally metabolically active cells that function in chlamydial attachment and infection, disseminating the infection. In contrast, reticulate bodies (RBs) are metabolically active and replicative, amplifying the progeny, but RBs cannot infect cells. Chlamydial development also includes an alternate non-infectious developmental stage, the enlarged, polyploid aberrant form (AF). A wide variety of different physiologic insults including beta-lactam treatment (12), interferon-gamma treatment (13), iron starvation (14), viral infection (15), and nutrient stress (16) lead to the formation of AFs in cultured cells. Upon removal of stress, chromosomal foci form within AFs and, through uncharacterized processes, AFs can re-enter the classic developmental cycle and generate infectious chlamydiae. This unusual response to stress is not simply an artifact of culture in vitro: AFs have been observed in tissue sections from human endocervical tissue (17) and in murine tissues following experimental chlamydial infection (18). The molecular processes leading to persistence in tissues and the actual role of AFs in clinical persistence remain elusive.

While AFs are a unique persistent developmental form in the chlamydial system, low levels of infectious forms are found in vitro during treatment with almost every stressor that generates AFs (14, 16, 19). This level of infectivity is, by definition, a function of EBs remaining quiescent in the persistent culture. The abundance and significance of these apparently dormant Chlamydia found in persistently infected cells remains unclear. The present study uses an antibiotic treatment and removal model system to explore dormancy-based C. trachomatis persistence in cell culture, in which minimally active, but non-infectious, developmental forms lead to productive inclusion formation after removal of antibiotic stress.

RESULTS

C. trachomatis recovery following treatment with quinolones, tetracyclines, and chloramphenicol

Initial experiments were conducted to determine the effect of increasing concentrations (1 and 10 µg/mL) of tetracycline (TET), chloramphenicol (CHL), or ofloxacin (OFX) on chlamydial development. McCoy cell monolayers infected with a C. trachomatis LGV-434 strain carrying a recombinant shuttle vector expressing mCherry [L2mCherry: Fig. S1 (20)] were examined during a 3–72-h incubation in differing concentrations of each antibiotic and then monitored for up to 72 h after drug removal. Treatment of cells from T = 0 with TET or CHL at 1 or 10 µg/mL led to complete elimination of chlamydial infectivity, as measured by lysis of cells and replating onto new cell monolayers. However, upon removal of the drug and further incubation of infected cells in antibiotic-free medium, bacterial growth and EB production were observed in all but one of the tested conditions (Fig. 1). Recovery was higher at 1 µg/mL antibiotic for TET and CHL, with no detectable recovery observed following treatment with 10 µg/mL TET within a 48-h recovery period following drug removal. These results demonstrate that apparently dormant, non-infectious chlamydiae persist in infected monolayers, and increasing concentrations of TET and CHL lead to lower recovery of infectious C. trachomatis.

Fig 1.

Quantification of EB production in a treatment/recovery model system for OFX, TET, and CHL. In each case, McCoy cells infected with C. trachomatis L2mCherry at a multiplicity of infection (MOI) of 0.3 were incubated for 24 h in medium containing the indicated antibiotic at either 1 or 10 µg/mL for 24 h. The cells were then washed and incubated an additional 48 h in antibiotic-free medium. Quantification of untreated (control) infections is shown for a 24-h incubation only. Data represent inclusion forming units, expressed as EB/mL, in a single infected well of a 24-well plate, with three technical replicates per treatment.

C. trachomatis recovery after quinolone treatment was different to that observed with TET and CHL. In all treatments, culturable EBs were undetectable while in the presence of drug, but EB recovery after removal of OFX was directly correlated with drug concentration (Fig. 1). There were also differences in chlamydial development in these treatments that was not observed with TET or CHL. Incubation of L2mCherry in 1 µg/mL OFX led to the accumulation of individual, enlarged fluorescent bacteria within infected cells, with the developing bacteria approaching the size of small AFs (Fig. 2A). These brightly fluorescent forms then ceased growth and remained generally unchanged throughout the entire duration of the analysis. Removal of OFX from cells treated with 1 µg/mL drug, followed by incubation for an additional 48 h in MEM-10 lacking OFX, led to limited productive development, with most of the inclusions remaining paused as single fluorescent developmental forms (Fig. 2D). In contrast, treatment of infected monolayers with 10 µg/mL OFX led to very low numbers of detectable fluorescent forms at any time during incubation in drug (Fig. 2B). Although the mCherry fluorescence within these cells was virtually undetectable, the presence of C. trachomatis in these monolayers was confirmed using antibody to the major outer membrane protein (MOMP), which is present at all stages of chlamydial development (Fig. 2C). Removal of the medium containing 10 µg/mL OFX and replacement with drug-free medium led to a markedly different phenotype compared with the cells treated with 1 µg/mL OFX: a higher number of inclusions developed in the cells treated with the higher concentration of OFX (Fig. 2E), leading to fully developed inclusions that are similar to those growing in the absence of drug (Fig. 2F). The concentration-dependent increase in productive C. trachomatis was also examined with an anti-MOMP antibody. Untreated infections show typical development (Fig. 2G), while infected cells recovered for 48 h after treatment with 1 µg/mL (Fig. 2H) or 10 µg/mL (Fig. 2I) OFX exhibited the concentration-dependent increase in productive development.

Fig 2.

Concentration-dependent chlamydial dormancy demonstrated by differential OFX treatment and recovery. Panels A, B, and D–F show McCoy cells infected with C. trachomatis strain L2mCherry infected under differing conditions and visualized for mCherry (red) and DNA (DAPI; blue). Panel F shows infected cells in normal medium, fixed with paraformaldehyde 24 h post-infection (hpi). Panels A, B, D, and E show similarly infected cells incubated in 1 µg/mL OFX (A and D) or 10 µg/mL OFX (B and E) for 24 h, followed by fixation (A and B) or incubation for 48 h in OFX-free medium (D and E). Panel C shows a higher magnification of infected cells treated with 10 µg/mL for 24 h, fixed with paraformaldehyde, and visualized for MOMP (green), mCherry (red), and DNA (blue). The space bar in B represents 20 microns for panels A, B, and D–F, while the bar in C represents 10 microns for that image. Panels G–I: MOMP (green) and mCherry (red) labeling of L2mCherry-infected cells either untreated for 24 h (panel G) or treated with OFX for 24 h and then recovered for 48 h in the absence of drug (panels H and I). The cells in panel H were treated with 1 µg/mL OFX, and the cells in panel I were treated with 10 µg/mL OFX prior to a 48-h incubation in OFX-free medium.

The concentration-dependent phenotype was also examined in a urogenital strain, C. trachomatis D/UW-3. Incubation of D/UW3 in 1 or 10 µg/mL OFX led to a similar pattern of higher productive inclusion formation following incubation in the higher concentration of drug. Three different antigens were examined in these analyses: GroEL (Fig. 3), Mip (Fig. S2), and MOMP (not shown). In cells incubated in drug for 24 h, antigen is readily detectable in single bacteria incubated in 1 µg/mL OFX, but all labeling is below the limits of detection with the 10 µg/mL treatment (Fig. 3A and B). However, removal of the antibiotic led to higher numbers of inclusions formed in the cells treated at 10 µg/mL (Fig. 3C and D). This phenotype was also observed in parallel experiments conducted in the human HeLa cell line (Fig. S3), and therefore, the source of host cells did not affect the described concentration-dependent recovery from OFX treatment.

Fig 3.

C. trachomatis D/UW-3 exhibits concentration-dependent dormancy during treatment with OFX. Infected cells were treated with either 1 µg/mL (panels A and C) or 10 µg/mL (panels B and D) for 24 h and then either fixed with methanol (panels A and B) or incubated in antibiotic-free medium for an additional 48 h prior to methanol-fixation (panels C and D). Each monolayer was then labeled with Mab B9, with specificity to chlamydial GroEL (red). DNA is labeled with DAPI (blue). The scale bar in A represents 20 microns for each image.

EB output differences were examined quantitatively by assessing the numbers of infectious units produced at each stage of the drug treatment for C. trachomatis L2-mCherry. An initial growth curve demonstrated that L2mCherry had typical growth kinetics in the absence of OFX (Fig. 4A), whereas there was no observed EB production in wells incubated in either 1 or 10 µg/mL OFX. Productive growth of L2mCherry EBs was demonstrated after removal of OFX and washing of the cells (Fig. 4B). The EB output was similar for each drug concentration in wells incubated in drug for 24 h and then in drug-free medium for 24 additional hours, but the nature of the EB accumulation changed as the drug-free culture period was extended. In repeated experiments where OFX was added for the first 24 h of culture followed by 48 h of culture in drug-free medium, cells treated with 10 µg/mL OFX always generated significantly higher numbers of infectious particles than those produced following treatment with OFX at 1 µg/mL. The relationship between increased OFX concentration and increased numbers of inclusion forming units (IFU) remained consistent when cells were incubated during the washout phase for longer periods of time: a 72-h drug-free culture period led to higher numbers of EB in both groups, but incubation in the higher concentration of OFX (10 µg/mL) always led to higher levels of recovery than those treated with OFX at 1 µg/mL (Fig. 4B).

Fig 4.

Quantification of infectious EB output in McCoy cells treated with differing concentrations of OFX. (A) Infectious EB output in cells infected with C. trachomatis L2mCherry over the course of 24 h of growth in McCoy cells, with or without OFX in the medium. Cells were infected at an MOI of 0.3. Lysates for each treatment group (no treatment, 1, or 10 µg/mL OFX) were titrated for infectious EBs at different points in a 30-h growth cycle. Panels B–F show infectious EB output for various treatment regimens with either low (1 µg/mL) or high (10 µg/mL) concentrations of OFX. Where indicated, numbers separated by a slash show the hours cultured in OFX/hours cultured in the absence of OFX. (B) A 48-h washout period is required to facilitate higher output in the 10 µg/mL-treated cells. (C) Extended incubation in OFX leads to lower total IFU output in cells treated with both 1 and 10 µg/mL OFX, but the relationship between EB output after washout remains consistent. (D) C. trachomatis L2mCherry elementary body recovery from cells treated with either 1 or 10 µg/mL OFX for varying amounts of time, followed by a 48-h recovery period. Cells were untreated with OFX (black bars) or treated with 1 or 10 µg/mL OFX (light or dark-gray bars, respectively). (E) MIC determination in C. trachomatis L2mCherry-infected cells subjected to treatment and removal of OFX. In each group, cells were treated as indicated for 24 h and then allowed to recover for 48 h in the absence of antibiotic. At the end of the experiment, infected cells were lysed, and the progeny was assessed for growth on increasing concentrations of OFX. (F) C. trachomatis strain D/UW-3 differentially treated with OFX at 0, 1, and 10 µg/mL and then recovered for 48 h in antibiotic-free medium. The left graph shows that strain D/UW-3 is sensitive to any tested concentration of OFX. The right graph shows EB output in cells treated with OFX at 1 and 10 µg/mL and then recovered for 48 h.

We then examined the effects of increasing lengths of time on OFX prior to rinsing and a subsequent 48-h culture in drug-free medium. Identical triplicates of L2mCherry-infected cells were incubated in 1 or 10 µg/mL OFX for between 3 and 24 h and then incubated in drug-free medium for 48 additional hours. The resulting data show that the observed difference in EB production following treatment with either 1 or 10 µg/mL OFX was non-existent following a 3-h treatment with OFX, but the difference increased in magnitude and acquired statistical significance as the length of drug treatment was extended (Fig. 4D). The effect of increasing the length of time in OFX to 48 h, prior to a 48-h culture period in the absence of OFX, was then examined (Fig. 4C). Results from these experiments demonstrate that survival in the presence of either 1 or 10 µg/mL OFX decreases with increasing lengths of time in drug but that the relative differences between treatment groups, after a 48-h washout, remains consistent.

The concentration-dependent OFX phenotype was also quantified in cells infected with strain D/UW-3. As was seen in the immunofluorescence microscopy, D/UW-3 exhibited a similar concentration-dependent increase in quantifiable EB production (Fig. 4F).

One concern that was examined was whether the differential drug treatment followed by extended culture in the absence of drug led to differential selection of OFX-resistant mutant strains. Ofloxacin resistance in chlamydiae can be generated by culturing chlamydia-infected cells in sub-inhibitory concentrations of the drug, and this has been used by different investigators to generate differently resistant strains for genetic experiments (21, 22). Testing of progeny from infections incubated in either 1 or 10 µg/mL OFX demonstrated that there was no significant increase in OFX resistance in any of the treatment groups (Fig. 4E).

Live cell imaging of C. trachomatis recovery following antibiotic treatment

To definitively determine the cellular source of productive inclusion development observed during antibiotic recovery, we employed live cell imaging of a paired promoter reporter strain to follow the developmental cycle after drug removal. For these experiments, we used a dual-fluorescent reporter chlamydial strain (L2-BMEC) that expresses GFP from and early RB promoter (euoprom) and RFP from a late EB-associated promoter (hctBprom). Imaging the expression of GFP and RFP in single inclusion facilitates the determination of the kinetics of the developmental cycle (23).

We first determined the kinetics of recovery of C. trachomatis in infected cells treated with OFX. Cells were treated with OFX (0.5, 1, 5, 10, 15, and 20 µg/mL) at infection, and the kinetics of recovery and completion of the developmental cycle was assessed using live cell imaging. At 24 hpi, the number of euoprom+ inclusions were quantified for 8 fields of view for each treatment (Fig. 5A). At this time, the OFX was removed, the plates were washed, and medium lacking antibiotic was added. At 72 hpi (48 h after washout), the number of inclusions that successfully completed the developmental cycle (euoprom+ and hctBprom+) was quantified for eight fields of view for each treatment (Fig. 5A). After 24 h of treatment (before washout), just as we saw with fixed cells, there were observable GFP-expressing chlamydial inclusions. Again, the number of GFP-positive inclusions was dose dependent, with significantly more positive inclusions at 0.5 µg/mL and very few at 20 µg/mL (Fig. 5A). However, recovery from treatment showed an inverse relationship between successful completion of the developmental cycle with more inclusions successfully completing the cycle at higher OFX concentrations (Fig. 5A). The lowest treatment had the fewest inclusions that produced normal productive cycle kinetics, while the highest had the most (0.5 µg/mL = 0.5%, 1 µg/mL = 2.7%, 5 µg/mL = 8.2%, 10 µg/mL = 22.8%, 15 µg/mL = 21.2%, and 20 µg/mL = 34.2; Fig. 5B). Analysis of the live cell microscopy data for the OFX recovery experiments showed that the euoprom+ aberrant inclusions observable after 24 h OFX treatment did not recover to produce productive inclusions (euoprom increase and hctBprom+) (Fig. 5C, arrows). Instead, previously unrecognized, euoprom-negative, cryptic early developmental forms were the source of the productive inclusions and the number of these cryptic inclusions increased with OFX concentration (Fig. 5C, circles).

Fig 5.

Live cell imaging of chlamydial growth inside OFX-treated cells. Cells cultured in six-well trays were infected with L2-BMEC and treated with 0.5, 1, 5, 10, 15, and 20 µg/mL OFX for 24 h before washing and replacement with drug-free media. (A) Quantification of euoprom+ inclusions at 24 hpi (panel A, left) and inclusions that successfully completed the developmental cycle after OFX washout (panel A, right) per field of view (*P < 0.001 ANOVA). (B) Live cell expression kinetics from single inclusions. Each line in the graphs shows euoprom-clover expression (top row) and hctBprom-mKate2 expression (bottom row) for 75 total hours of culture, which represents a 24-h culture in drug followed by a 51-h washout period. For the untreated sample, cells were infected at the time of OFX washout. The number of inclusions whose fluorescence is driven by hctBprom and that therefore have completed the developmental cycle is indicated by n. The percentage is the number of productive inclusions (euoprom+ and hctBprom+) compared with the inoculum. (C) Single micrographs from live cell imaging showing that the euoprom+ inclusions at washout (red arrows) did not complete the cycle while non-visible Chlamydia eventually express euoprom and hctBprom, completing the cycle (purple circles).

Additionally, for all treatments (0.5, 1, 5, 10, 15, and 20 µg/mL OFX), euoprom activity was first detected at ~12 h post-OFX washout, while hctBprom activity was detected at ~24 h post-washout which is similar to untreated kinetics.

Together, these data suggest that OFX-mediated inhibition of early gene expression creates cryptic early inclusions that can re-enter the productive chlamydial developmental cycle when the drug pressure is removed. It also demonstrates that early gene expression in the presence of OFX leads to aberrant inclusions that cannot recover to produce a productive infection.

Next, we examined the recovery kinetics for CHL treatment. Cells were infected with L2-BMEC at an MOI of ~0.3 and treated with CHL (1 µg, 5 µg, and 10 µg) at infection for 24 h followed by washout and live cell imaging. Again, RB germination and growth were evaluated by GFP expression and expression of the RFP from the late promoter was evaluated for EB formation which together assess the developmental cycle (23). For all treatments (1, 5, and 10), at the start of imaging (24 hpi), no Chlamydia-expressing GFP from the euo promoter could be detected in the infected cells (Fig. S3). However, after washout, a subset of Chlamydia were able to recover and complete the developmental cycle, expressing both GFP and RFP (Fig. S3). The kinetics of recovery was slightly delayed and dose dependent. Recovery at the 1-µg/mL dose GFP expression was detected at ~38 hpi (14 hours post-drug removal) while at the 5-µg and 10-µg treatments, GFP expression was detected at ~40 hpi (16 h post-drug removal) (Fig. 5A). Additionally, the number of Chlamydia that recovered was also dose dependent. For the lowest dose (1 µg/mL), 33.3% ± 2.9% of the inoculum recovered while for 5 µg/mL and 10 µg/mL, 18.6% ± 2.6% and 23.8% ± 1.8% recovered, respectively (Fig. 6B). In all cases, the developmental cycle followed wild-type kinetics after initiation of GFP expression (Fig. 6A).

Fig 6.

Live cell imaging of chlamydial growth inside CHL-treated cells. Cos-7 cells cultured in six-well glass bottom plates were infected with L2-BMEC and treated with 1, 5, and 10 µg/mL CHL for 24 h before washing and replacement with drug-free media. The infections were monitored using live cell imaging for 75 h. (A) Expression intensities from the euo promoter and hctB promoter from >50 individual inclusions were monitored via automated live-cell fluorescence microscopy, and the mean intensities are shown. Cloud represents SEM. Y-axes are denoted in scientific notation. (B) Live cell expression kinetics from single inclusions. Each line in the graphs show euoprom-clover expression (top row) and hctBprom-mKate2 expression (bottom row) for 75 total hpi, which represents 24 hpi in drug followed by a 51-h washout period. The percentage of fully received inclusions (hctB+) was compared with untreated and reported as % recovered. All were significantly lower than untreated (P < 0.01; multiple comparison of means using the Tukey honest significant difference [HSD] test).

We also investigated the kinetics of recovery from TET treatment. Cells were again infected with L2-BMEC and treated with 1, 5, or 10 µg/mL TET at the time of infection. TET was washed out 24 h after infection of cells and were imaged to determine the kinetics of recovery. Again, for each treatment condition at the start of imaging (24 hpi), no C. trachomatis-expressing GFP from the euo promoter could be detected (Fig. S4). Recovery from TET treatment was dose dependent like CHL treatment, but the TET dose dependance was more pronounced than that observed with CHL. In chlamydial inclusions treated with 1 µg/mL TET, GFP expression was first detected at ~42 hpi (18 h post-washout), while for 5 µg/mL TET, GFP expression was first detected at ~50 hpi (26 h after washout) and at ~55 hpi (31 h after washout) for the 10 µg/mL treatment (Fig. 7A). Again, similar to CHL treatment, the percentage of Chlamydia that recovered was also dose dependent. Treatment with 1 µg/mL resulted in 45.9% ± 5.6% of the inoculum recovering while for the 5 µg/mL treatment resulted in 14.0% ± 1.8% of the inoculum recovered, and at 10 µg/mL, recovery was reduced to 8.4% ± 1.4% (Fig. 7B). Again, in all treatments, the C. trachomatis that recovered, as measured by GFP expression, went on to produce a productive cycle as measured by hctBprom activity (Fig. 7B). After the initial delay in recovery, the kinetics of the cycle were similar to the untreated for all three concentrations.

Fig 7.

Live cell imaging of chlamydial growth in TET-treated cells. Cos-7 cells cultured in six-well glass bottom plates were infected with L2-BMEC and treated with 1, 5, and 10 µg/mL TET for 24 h before washing and replacement with drug-free media. The infections were monitored using live cell imaging for 75 h. (A) Expression intensities from the euo promoter and hctB promoter from >50 individual inclusions were monitored via automated live cell fluorescence microscopy, and the mean intensities are shown. Recovery showed a pronounce dose response. Cloud represents SEM. Y-axes are denoted in scientific notation. (B) Live cell expression kinetics from single inclusions. Each line in the graphs show euoprom-clover expression (top row) and hctBprom-mKate2 expression (bottom row) for 75 total hpi, which represents a 24 hpi in drug followed by a 51-h washout period. The percentage of fully received inclusions (hctB+) was compared with untreated and reported as % recovered. All were significantly lower than untreated (P < 0.01) (multiple comparison of means—Tukey HSD). Additionally, each increasing dose had significantly less recovery (P < 0.01) (multiple comparison of means—Tukey HSD).

To characterize the germination state of the OFX-arrested EB-like forms, we asked if the high OFX-treated chlamydial cells were capable of any early gene expression-dependent processes during infection. We have previously shown that early after entry (0–4 hpi), the nascent C. trachomatis-containing inclusions are trafficked along the host microtubule network and, in multiply infected cells, coalesce at the host cell microtubule organizing center (MTOC). This process is dependent on both de novo transcription and translation (24). To determine how OFX treatment affected this process, we infected cells with L2-BMEC at an MOI of ~20 and treated the infected cells with either 1 or 20 µg/mL of OFX. The infected cells were fixed and stained for Chlamydia and microtubules at 24 hpi. For the 1 µg/mL OFX-treated cells, the Chlamydia coalesced at the MTOC of the cell, and as shown in previous figures, the chlamydial cells became larger and expressed GFP from the euo promoter (Fig. 8). Chlamydia also coalesced at the MTOC of the infected cells when treated with 20 µg/mL OFX for 24 h; however, the chlamydial cells remained small and did not express GFP from the euo promoter suggesting that these cells initiated some aspects of germination but did not fully germinate. Additionally, consistent with our previous study, the trafficking phenotype was dependent on de novo protein synthesis, as inclusion coalescence was inhibited when 10 µg/mL CHL was added at the time of infection (Fig. 8). These data suggest that the arrested persistent chlamydial cell forms induced by high OFX treatment were metabolically active and initiated a very early step in establishing the inclusion environment but were arrested in a state before full germination.

Fig 8.

The high OFX-induced persister state is capable of early germination-dependent processes. High OFX-treated Chlamydia microtubule trafficking comparisons between low (1 µg/mL OFX), high (20 µg/mL OFX), and OFX plus CHL (20 µg/mL, 10 µg/mL CHL). Cos-7 cells were infected with L2-BMEC and treated with antibiotics at infection. Cells were fixed at 24 hpi and stained with an anti-LPS antibody to visualize Chlamydia (red). Microtubules were stained with anti-beta tubulin monoclonal antibody (pink), and clover expression from the euo promoter was visualized in green. Images are z-projected confocal micrographs. Size bar = 10 µm.

DISCUSSION

Clinical persistence is important in many different chlamydial disease processes. Contemporary modeling of chlamydial persistence involves the development of aberrant forms—large, non-infectious, polyploid developmental forms that are maintained intracellularly in stressed, infected cells. Here, we present a model for a different type of chlamydial persistence—the presence of dormant, reduced-metabolically active chlamydiae in cells undergoing antibiotic treatment.

Classical bacterial persistence mechanisms are modeled in vitro by examining the fate of either growing or dormant bacteria in a population. The original association of bacterial persistence with metabolic dormancy was demonstrated over 70 years ago (25), but the molecular basis for these processes has been uncovered only recently. Excellent recent reviews delineate the clear differences between bacterial resistance and persistence and explore the subtle variations that define bacterial tolerance vs persistence (26). Bacterial tolerance and persistence are associated with quiescent—or dormant—forms that remain viable during periods of bactericidal stress. In contrast to many organisms that undergo stress-mediated dormancy in vitro or in vivo, Chlamydia spp. do not form biofilms, do not have toxin/antitoxin systems, and do not synthesize or respond to ppGpp alarmones (27, 28). Therefore, any molecular understanding of dormancy in this system is unexplored.

We investigated the effects of three antibiotics on the ability of Chlamydia to tolerate and persist after treatment. We determined the ability of chlamydial infections to recover from treatment with CHL (targets the 50-s ribosome subunit), TET (targets the 30-s ribosome subunit), and OFX (targets DNA gyrase). For all tested drugs, the C. trachomatis that recovered to produce a productive infection were quiescent with no detected gene expression.

In the CHL-treated experiments, there was only a slight dose-dependent decrease in Chlamydia that was able to recover from treatment and this decrease was associated with a slight delay in initiation of the developmental cycle for the highest two CHL treatments (5 and 10 µg/mL). For TET treatment, both the dose-dependent decrease in recovery and dose-dependent delay of recovery were more dramatic. This is likely due to differences in the reversible binding kinetics of these antibiotics to the chlamydial ribosomes.

Intriguingly, we observed a paradoxical increase in productive growth in the monolayers treated with higher concentrations of OFX. In non-chlamydial systems, quinolone antibiotics facilitate a similar concentration-dependent survival property that was first identified in 1975 (29) and has recently been attributed to reduced oxidative stress in bacteria treated with higher quinolone concentrations (30). The actual participation of oxidative stress in survival of bacteria during quinolone treatment remains an active area of investigation, and there are excellent reviews on the subject that present very different perspectives (31, 32). Virtually, all the published analyses have been conducted in extracellular bacteria, and any secondary effects of quinolone antibiotics for intracellular organisms are uninvestigated. We show here that C. trachomatis exhibits a concentration-dependent increase in survival in OFX-treated cells, and this survival is associated with developmental forms that exhibit dramatically reduced translational activity. Chlamydial development is initiated in cells treated with lower OFX concentrations, but this development is terminated at the single-cell stage in the large majority of treated cells. These active forms enter the midcycle expression profile (Fig. 3 and 5; Video S1 and S2) and make both early and midcycle inclusion membrane proteins but do not develop further upon removal of drugs. In contrast, infected cells treated with a higher concentration of OFX or ciprofloxacin (10 or 20 µg/mL) appeared to be quiescent by early gene expression yet proceeded to recover robust productive development after drug washout. Higher drug concentration led to higher IFU output in every treatment model, including extended incubation in drug and extended periods after washout. Intriguingly, the high concentration OFX-treated chlamydia did initiate very early gene expression leading to intracellular trafficking, one of the earliest events in establishing the chlamydial inclusion. This suggests that the persistent forms induced by high OFX reinitiated some gene expression but were inhibited from germinating. Furthermore, these data may indicate that the transition from the EB to the replication competent RB is a multistage process.

While quinolones have a strong inhibitory effect on C. trachomatis growth in vitro, their efficacy is not high in patient treatment (33). There are many possible reasons for this difference, one hypothesis might be that patient treatment leads to the host cells containing dormant forms identified in this report. This would be very different than the AF-mediated persistent form that is described for many other stressors.

The live cell imaging data summarized in Fig. 5 to 7 and shown in Video S1 and S2 provide evidence for the origins of productive development in the infected cells subjected to drug treatment followed by washout and incubation in drug-free medium. For all antibiotic treatments, translationally quiescent developmental forms are the source of recovery and productive inclusions. For OFX, this is more dramatic as translationally active developmental forms are seen that grow to an enlarged single-cell stage, but growth is arrested at that point and these forms do not continue even after an extended incubation in drug-free medium. In contrast, like those seen for CHL and TET, developmental forms that are invisible during treatment with either the low or high concentration of OFX become fluorescent and subsequently develop after removal of the antibiotic. While this productive development is observed in cells treated with both low and high amounts of antibiotic, cells in the high-drug treatment group generate quantitatively more productive inclusions and significantly higher numbers of infectious EBs. Conversely, the opposite was observed in the formation of translational competent (aberrant) cells, high OFX resulted in significantly fewer while low OFX treatment resulted in more of these forms. This is a concentration-dependent phenomena: quantitative live cell imaging analysis over a range of OFX concentrations generates a statistically significant increase in productive inclusion formation over a range of 0.5 to 20 µg/mL OFX.

This study shows that in addition to AF persistence that, despite not having a true stringent response, chlamydial cells can act as classical bacterial persister cells when subjected to antibiotic stress. Intriguingly, OFX treatment increases the number of bacteria that enter this state. For these experiments, we initiated treatment at T = 0, when medium was first added to infected monolayers. Therefore, the cell forms that acted as persisters were EBs after cell entry and were apparently rendered dormant before robust translation was detected. We hypothesize that treatment with high OFX concentrations is trapping the germinating EB in a persister-like state, perhaps by inhibiting the initiation of DNA replication. Treatment with low OFX allowed a higher number of early EBs to proceed into the early stages of germination, and these then became sensitive to the lethal effects of OFX and perhaps host cell killing responses. The idea that the early EBs after entry but before robust gene expression acts as a persister cell is supported by the TET and CHL data, which show that cells in this state are tolerant to reversible translational inhibition.

The work in this study demonstrates that chlamydiae may persist during infections using multiple strategies. This includes both AF and a novel, non-infectious dormant persistent form that is more consistent with established processes exhibited by many species. We are currently exploring the nature of the dormant forms identified in the OFX treatment model, with a goal of understanding the cellular processes leading to the dormancy-associated chlamydial persistence.

MATERIALS AND METHODS

Chlamydia strains and culture conditions

Chlamydia trachomatis strains used in this study were propagated using standard culture conditions and purified on a 30% renografin pad (34). Frozen aliquots of each strain were stored at −80°C prior to use in the experiments. Both the McCoy cell line (murine, fibroblast, ATCC CRL-1696) and HeLa cell line (human, epithelial, ATCC CCL-2) were cultured in DMEM + 10% fetal calf serum (FCS) at 37°C in 5% CO₂. Chlamydia were inoculated onto cells in PBS (multiplicity of 0.3, unless indicated) and centrifuged at 545 × g for 1 h at room temperature. After centrifugation, the inocula were removed and cells were incubated in DMEM + FCS + 1 µg/mL cycloheximide. Ofloxacin (Sigma O8757) was dissolved in 1 M NaOH and stored frozen at 50 mg/mL. This stock solution was thawed and diluted to concentrations indicated in the results. Other antibiotics were purchased from Sigma and prepared according to standard methods (35). Unless indicated differently, each antibiotic was removed from culture 24 hpi to observe the transition of C. trachomatis from dormancy to productive growth.

Three different C. trachomatis strains were examined in this work. C. trachomatis strain L2mCherry expresses the mCherry protein, driven by a constitutive GroEL promoter. The strain is brightly fluorescent in live infected cells or cells fixed with 4% paraformaldehyde PBS (Fig. S1). C. trachomatis D/UW-3 is a prototype strain originally cultured from an asymptomatic female (36, 37). C. trachomatis strain L2-BMEC expresses the mKate protein from the hctB promoter and the clover protein from the euo promoter (23).

Microscopy of fixed images

Immunofluorescence microscopy was conducted using described methodologies (34) following fixation with either 100% methanol or 4% paraformaldehyde in PBS, as indicated for each experiment. Antibodies used in this work included anti-chlamydial GroEL [CTL_0365: antibody B9: (38)], anti-C. trachomatis MOMP [L2-I10 (39)], anti-chlamydial MIP (clone 154, (40)], and anti-beta tubulin (Sigma). All immunofluorescent images were collected following culture of infected cells on glass coverslips for the indicated number of hours. Fixed cells were then labeled and inverted onto mounting medium consisting of Vectashield Antifade Mounting Media (Vector Laboratories, Newark CA) containing the DNA-specific dye 4′,6-diamidino-2-phenylindole. Images were collected on a Leica DMLB microscope with a QImaging camera and Qcapture Pro software (Teledyne Photometrics, Tucson, AZ). Additional confocal images were acquired using a Nikon CrestOptics X-Light coupled with Nikon Elements imaging software. Images were taken using a 100× oil-immersion objective.

Live cell microscopy

Cos-7 cells were seeded on a multi-well glass-bottom plate, and infections were synchronized by incubating monolayers with L2-BMEC EBs in Hank’s balanced salt solution (HBSS, Gibco, Grand Island, NY) for 15 minutes at 37°C while rocking. The inoculum was removed, and cells were washed with prewarmed (37°C) HBSS with 1 mg/mL heparin and 1 mM cycloheximide. The HBSS was replaced with fresh RPMI-1640 containing 10% FetalPlex (FP, Gemini Bio-Products, Sacramento CA), 10 µg/mL gentamicin, 1 mg/mL heparin, 1 mM cycloheximide, and the indicated concentrations of OFX. OFX was removed at 24 hpi, and the infected monolayers were washed 3× in HBSS and 1× in phenol red-free RPMI-1640 (Gibco). The final wash was replaced with phenol red-free RPMI-1640 containing 10% FP, 10 µg/mL gentamicin, and 1 mg/mL heparin. Washed plates were placed in an OKOtouch CO2/heated stage incubator immediately upon washout and imaged for a further 54 h at 30-min intervals. Fluorescence images were acquired via epifluorescence microscopy using a Nikon Eclipse TE300 inverted microscope with a ScopeLED lamp at 470 nm and 595 nm and BrightLine Bandpass filters at 514/30 nm and 590/20 nm. A 20×/0.4NA dry objective lens was used. DIC was used to auto-focus images. Image acquisition was performed using an Andor Zyla sCMOS camera in conjugation with μManager software (41). Multiple fields were imaged for each treatment and the fluorescent intensity of individual inclusions was monitored using the Trackmate plug-in in ImageJ (42). Inclusion fluorescent intensities were graphed in Python as previously described (23). Statistics were performed using ANOVA from statsmodels (https://www.statsmodels.org/stable/index.html).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/iai.00339-23.

Supplemental material. iai.00339-23-s0001.pdf.

Figures S1 to S5; legends for Videos S1 and S2.

Video S1. iai.00339-23-s0002.mov.

Live-cell time-lapse movie of Cos-7 cells infected with C. trachomatis L2-BMEC and treated with 1 µg/mL OFX at infection for 24 hours.

Video S2. iai.00339-23-s0003.mp4.

Live-cell time-lapse movie of Cos-7 cells infected with C. trachomatis L2-BMEC and treated with 20 µg/mL OFX at infection for 24 hours.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.