Impact of a Low-Oxygen Environment on the Efficacy of Antimicrobials against Intracellular Chlamydia trachomatis

Abstract

Emergence of chronic inflammation in the urogenital tract induced by Chlamydia trachomatis infection in females is a long-standing concern. To avoid the severe sequelae of C. trachomatis infection, such as pelvic inflammatory diseases (PID), ectopic pregnancies, and tubal infertility, antibiotic strategies aim to eradicate the pathogen even in asymptomatic and uncomplicated infections. Although first-line antimicrobials have proven successful for the treatment of C. trachomatis infection, treatment failures have been observed in a notable number of cases. Due to the obligate intracellular growth of C. trachomatis, reliable antimicrobial susceptibility assays have to consider environmental conditions and host cell-specific factors. Oxygen concentrations in the female urogenital tract are physiologically low and decrease further during an inflammatory process. We compared MIC testing and time-kill curves (TKC) for doxycycline, azithromycin, rifampin, and moxifloxacin under hypoxia (2% O₂) and normoxia (20% O₂). While low oxygen availability only moderately decreased the antichlamydial activity of azithromycin in conventional MIC testing (0.08 μg/ml versus 0.04 μg/ml; P < 0.05), TKC analyses revealed profound divergences for antibiotic efficacies between the two conditions. Thus, C. trachomatis was significantly less rapidly killed by doxycycline and azithromycin under hypoxia, whereas the efficacies of moxifloxacin and rifampin remained unaffected using concentrations at therapeutic serum levels. Chemical inhibition of multidrug resistance protein 1 (MDR-1), but not multidrug resistance-associated protein 1 (MRP-1), restored doxycycline activity against intracellular C. trachomatis under hypoxia. We suggest careful consideration of tissue-specific characteristics, including oxygen availability, when testing antimicrobial activities of antibiotics against intracellular bacteria.

INTRODUCTION

Genital tract infections with Chlamydia trachomatis are the most common sexually transmitted disease (STD) in the United States (5), affecting mostly adolescents and adults under 25 years of age (5). Ninety-two million people are estimated to be infected worldwide (36). Severe clinical sequelae, such as pelvic inflammatory disease (PID), ectopic pregnancy, and tubal infertility, develop through chronic inflammatory processes of persistently infected tissues (11, 23).

First-line drugs, such as doxycycline and macrolides, have been successfully proven for the treatment of asymptomatic and uncomplicated C. trachomatis infections (5). However, treatment failures resulting in persistent or recurrent C. trachomatis infections with subsequent chronic tissue damage have been frequently observed (1, 21). Defining accurate treatment strategies for intracellular chlamydial infections is difficult, as standardized assays for in vitro testing of chlamydial isolates are hardly established and are limited by the fact that specific environmental conditions are not considered. The MIC test determines antimicrobial activity against freshly inoculated cells but does not reflect the potency of the antibiotic against pathogens at different intracellular developmental stages (31). In contrast, determination of time-kill curves (TKC) for antimicrobials against C. trachomatis allows simulation of clinical treatment conditions that are more closely related to the in vivo situation (31).

Hypoxia is known to modulate central host cell signaling pathways involved in metabolism and survival, and it regulates the expression of multidrug resistance transporters via hypoxia-inducible factor 1α (HIF-1α) (29). We and others showed that chlamydiae adapted well to a low-oxygen environment and directly interfered with the stabilization of HIF-1α, the central mammalian oxygen sensor, to replicate (17, 26). In addition, gamma interferon (IFN-γ) could not control intracellular growth of C. trachomatis in human fallopian tube cells in a low-oxygen environment, which can be found in the urogenital tract in women (25).

Reduced effectiveness of chemotherapeutical agents within oxygen-restricted areas is a well-known phenomenon (10). While diminished efficacy of antimicrobials against intracellular bacteria in cells overexpressing multidrug resistance protein 1 (MDR-1) has been reported, no direct correlation between oxygen availability and intracellular effectiveness of antimicrobials has been observed so far (20, 22, 30). Therefore, we investigated whether intracellular activities of recommended and alternative antimicrobials against C. trachomatis are maintained in a low-oxygen environment (16, 34).

MATERIALS AND METHODS

Bacterial strains and epithelial cell culture.

C. trachomatis serovar L2 (ATCC VR-902B) and HEp-2 cells (ATCC CCL-23) were used in the present study. A total of 5 ×10⁴ cells per well in 24-well plates (Greiner bio-one, Frickenhausen, Germany) or 2.5 × 10⁵ cells per well in 6-well plates (Greiner bio-one) were grown in RPMI 1640 medium (PAA Laboratories, Cölbe, Germany) supplemented with 10% fetal bovine serum (Gibco/Invitrogen, Karlsruhe, Germany), nonessential amino acids (PAA Laboratories), 2 mM glutamine (PAA Laboratories), 2 μg/ml amphotericin B (PAA Laboratories), and 100 μg/ml gentamicin (PAA Laboratories) under normoxic (20% O₂) and hypoxic (2% O₂) conditions (Toepffer Laboratory Systems, Göppingen, Germany).

Chemicals.

Doxycycline, azithromycin, rifampin, cyclosporine A (CsA), and probenecid were purchased from Sigma Aldrich (Deisenhofen, Germany), and moxifloxacin was purchased from Bayer Vital GmbH (Leverkusen, Germany).

Determination of the MIC for C. trachomatis.

A total of 5 × 10⁴ HEp-2 cells per well were seeded in 24-well plates and cultured for 24 h under normoxic and hypoxic conditions. The culture medium was subsequently changed to RPMI 1640 medium supplemented with 5% fetal bovine serum, and the cells were infected with C. trachomatis serovar L2 at 3 × 10⁵ inclusion-forming units (IFU)/ml with or without the respective antibiotics in different concentrations. MICs were determined after 48 h of incubation by immunofluorescence staining with a mouse anti-chlamydial lipopolysaccharide (LPS) antibody (kindly provided by Helmut Brade, Borstel, Germany), and a polyclonal rabbit fluorescein isothiocyanate (FITC)-labeled anti-mouse IgG antibody (Dako, Hamburg, Germany) was used for detection of C. trachomatis.

TKC.

TKC was performed as described previously with minor modifications (31). Briefly, 5 ×10⁴ HEp-2 cells per well were seeded in 24-well plates and cultured for 24 h under normoxic and hypoxic conditions. The culture medium was subsequently changed to RPMI 1640 medium supplemented with 5% fetal bovine serum, and the cells were infected with C. trachomatis serovar L2 at 3 × 10⁵ IFU/ml with or without the respective antibiotics in different concentrations. The reported serum levels (non-serum protein-bound antibiotics) were selected as working concentrations for doxycycline (2 μg/ml), azithromycin (0.5 μg/ml), moxifloxacin (3 μg/ml), and rifampin (8 μg/ml) (9, 31, 35). After the indicated time points, the cells were washed twice with medium to remove the remaining antibiotic and subsequently cultured for 48 h for determination of the recoverable chlamydiae as described previously (31). the numbers of recoverable chlamydiae at the indicated time points were normalized to those of the untreated controls (0 h postinfection [p.i.]) under normoxia and hypoxia, respectively, and displayed as a percentage of the controls. The means of duplicate performances of four independent experiments were used for calculation of statistical significance.

Real-Time PCR.

A total of 2.5 × 10⁵ HEp-2 cells/well were seeded in 6-well plates overnight and cultured in RPMI 1640 medium with 5% fetal bovine serum for up to 24 h under normoxic and hypoxic conditions. Extraction of total RNA, reverse transcription to cDNA, and PCR amplification using the LightCycler Detection System (Roche Molecular Biochemicals, Mannheim, Germany) were performed at the indicated time points as described previously (27). Relative quantification of MDR-1 or multidrug resistance-associated protein 1 (MRP-1) mRNA expression levels (37) was performed against the endogenous control β-actin gene (3). The primer sequences were as follows: MDR1 forward, 5′-CCCATCATTGCAATAGCAGG-3′; MDR1 reverse, 5′-GTTCAAACTTCTGCTCCTGA-3′; MRP1 forward, 5′-GGACCTGGACTTCGTTCTCA-3′; MRP1 reverse, 5′-CGTCCAGACTTCCTTCATCCG-3′; β-actin forward, 5′-CCTGGCACCCAGCACAAT-3′; β-actin reverse, 5′-GGGCCGGACTCGTCATAC-3′.

MDR-1 and MRP-1 inhibition assays.

A total of 5 × 10⁴ HEp-2 cells per well were seeded in 24-well plates and cultured for 24 h under hypoxic conditions. The culture medium was subsequently changed to RPMI 1640 medium supplemented with 5% fetal bovine serum, and the cells were infected with C. trachomatis serovar L2 with 3 × 10⁵ IFU/ml with or without the inhibitor for 8 h. Doxycycline efficacy with and without inhibitory treatment was analyzed 2 h and 4 h p.i. and normalized to the inhibitory effect of CsA or probenecid 8 h p.i.

Statistics.

Data are indicated as mean ± standard error of the mean (SEM). The data were evaluated using unpaired Student t tests. P values of ≤0.05 were considered statistically significant.

RESULTS

MIC testing of antimicrobials against C. trachomatis under normoxia and hypoxia.

The MICs for doxycycline, azithromycin, moxifloxacin, and rifampin against C. trachomatis were calculated under normoxic (20% O₂) and hypoxic (2% O₂) conditions (Table 1). No significant differences were observed for doxycycline, moxifloxacin, and rifampin. Interestingly, azithromycin showed significantly higher MIC values (0.08 μg/ml versus 0.04 μg/ml; n = 3; P < 0.05) for C. trachomatis when the cells were cultured under hypoxic conditions in comparison to normoxic conditions (Table 1 and Fig. 1).

Table 1.

MICs of antibiotics against C. trachomatis under normoxia and hypoxia

Antibiotic	MIC (μg/ml)
	Normoxia (20% O2)	Hypoxia (2% O2)
Doxycycline	0.05	0.05
Azithromycin	0.04a	0.08
Moxifloxacin	0.05	0.05
Rifampin	0.005	0.005

^aP < 0.05.

Fig. 1.

Comparison of the MIC values for azithromycin (AZM) under normoxia (20% O₂) and hypoxia (2% O₂) 48 h p.i. Immunofluorescence staining with mouse anti-chlamydial LPS antibody visualized chlamydial inclusions (green). Evans blue counterstaining of host cells (red) was used for better characterization of intracellular inclusions. Under normoxia, no visible chlamydial growth was observed at a concentration of 0.04 μg/ml, whereas complete eradication of chlamydiae was obtained at a concentration of 0.08 μg/ml under hypoxia. Nox, normoxia; Hox, hypoxia. Representative images from three independent experiments are shown.

TKC of antimicrobials against C. trachomatis under normoxia and hypoxia.

To uncover differences in the activities of antimicrobials against intracellular chlamydial growth under normoxic and hypoxic conditions, we performed TKC. C. trachomatis growth without antibiotic treatment was not significantly different under normoxia and hypoxia (5 × 10⁸ ± 4 × 10⁷ IFU/ml versus 5 × 10⁸ ± 6 × 10⁷ IFU/ml; n = 7; P = 0.4) (Fig. 2A), confirming what was reported by Juul et al. (17). In contrast to MIC testing, TKC revealed significant differences in the antichlamydial activities of all tested antibiotics between normoxic and hypoxic conditions. Thus, doxycycline (2 μg/ml) and azithromycin (0.5 μg/ml) treatments were significantly less effective against early intracellular chlamydial growth when the cells were grown in hypoxia (Fig. 2 and 3). When the doxycycline concentration was reduced (0.5 μg/ml), significantly fewer chlamydiae were killed under hypoxia than under normoxia, whereas an increase to 8 μg/ml resulted in complete eradication of chlamydiae under both conditions (Fig. 2B). The same trend was observed for azithromycin (Fig. 3B). In contrast, moxifloxacin and rifampin were equally efficient in eradicating intracellular C. trachomatis at concentrations of 3 μg/ml and 8 μg/ml, respectively (data not shown). However, when the concentrations of moxifloxacin and rifampin were reduced to 0.5 μg/ml, C. trachomatis was significantly more effectively eradicated under normoxia than under hypoxia (Fig. 4A and B).

Fig. 2.

TKC of doxycycline (DOX) against C. trachomatis under normoxic and hypoxic conditions. (A) Chlamydial growth was not significantly different between normoxic and hypoxic conditions (n = 7). (B) Immunofluorescence staining showed reduced efficacy of DOX (2 μg/ml) against C. trachomatis under hypoxic conditions. (C) Significant differences in the antichlamydial activities of doxycycline between normoxia and hypoxia were also observed when the concentration of the antimicrobial was reduced (0.5 μg/ml), but not when the concentration was increased (8 μg/ml). The numbers of recoverable chlamydiae at the indicated time points were calculated as a percentage of the untreated control under normoxia and hypoxia (n = 4; mean ± SEM; *, P < 0.05).

Fig. 3.

TKC of azithromycin against C. trachomatis under normoxic and hypoxic conditions. (A) Immunofluorescence staining showed reduced efficacy of AZM (0.5 μg/ml) against C. trachomatis under hypoxic conditions. (B) Differences in the efficacy of AZM in early intracellular C. trachomatis eradication with respect to the oxygen availability were maintained when the concentration was decreased to 0.04 μg/ml and were even more pronounced when the concentration was enhanced to 0.5 μg/ml. The numbers of recoverable chlamydiae at the indicated time points were calculated as a percentage of the untreated control under normoxia and hypoxia (n = 4; mean ± SEM; *, P < 0.05).

Fig. 4.

TKC of rifampin (RIF) and moxifloxacin (MXF) against C. trachomatis under normoxic and hypoxic conditions. RIF (0.5 μg/ml) (A) and MXF (0.5 μg/ml) (B) showed reduced antichlamydial activity under hypoxia. The numbers of recoverable chlamydiae at the indicated time points were calculated as a percentage of the untreated control under normoxia and hypoxia (n = 4; mean ± SEM; *, P < 0.05).

MDR-1 and MRP-1 mRNA expression levels under normoxic and hypoxic conditions.

We could demonstrate by TKC testing that antibiotic efficacy was reduced under hypoxia compared to normoxia. It is known that the intracellular activities of antimicrobials strongly depend on cellular accumulation and are tightly regulated by multidrug resistance transporters. We therefore investigated the transcriptional activities of the MDR-1 and MRP-1 genes under normoxia and hypoxia. Within 24 h, both MDR-1 (2.0-fold ± 0.2-fold; P < 0.05) and MRP-1 (1.4-fold ± 0.1-fold; P < 0.05) were significantly upregulated under hypoxia compared to normoxia (Fig. 5). Stabilization of HIF-1α was observed by Western blot analysis in cells that were incubated under hypoxia, but not under normoxia (data not shown).

Fig. 5.

Quantification of MDR-1 and MRP-1 mRNA expression levels under normoxic and hypoxic conditions. The data were normalized to normoxic values at 0 h and 24 h. A significant increase in MDR-1 and MRP-1 mRNA expression under hypoxia was observed within 24 h (n = 6; mean ± SEM; *, P < 0.05).

Functional relevance of MDR-1 and MRP-1 expression to reduced antichlamydial activity of doxycycline under hypoxia.

To test whether hypoxia-induced upregulation of MDR-1 and MRP-1 expression is responsible for reduced activity of doxycycline under hypoxia, we blocked MDR-1 with CsA and blocked MRP-1 with probenecid. Control studies were also conducted to verify that each inhibitor, at the concentrations employed, did not affect epithelial cell viability over the assay period (data not shown). C. trachomatis growth was slightly inhibited by 10 μM CsA and 2.5 mM probenecid. To analyze the impacts of the inhibitors on doxycycline efficacy in hypoxia, we normalized the respective values against the rates of chlamydia recovery with and without CsA or probenecid treatment in the absence of doxycycline. Inhibition of MDR-1 by CsA significantly increased the antichlamydial activity of doxycycline (2 μg/ml) under hypoxic conditions compared to the untreated control cells (Fig. 6A). In contrast, inhibition of MRP-1 using probenecid did not restore the antichlamydial activity of doxycycline but rather attenuated the efficacy of doxycycline against chlamydial growth under hypoxic conditions (Fig. 6B).

Fig. 6.

MDR-1, but not MRP-1, inhibition restores antichlamydial activity of doxycycline under hypoxia. Incubation for 8 h with 10 μM CsA (A), but not with 2.5 mM probenecid (B), resulted in increased intracellular eradication of C. trachomatis when the cells were treated with doxycycline under hypoxic conditions. Doxycycline efficacy with and without inhibitory treatment was analyzed 2 h and 4 h p.i. and normalized to the inhibitory effect of CsA or probenecid at 8 h p.i. (n = 4 for CsA; n = 3 for probenecid; mean ± SEM; *, P < 0.05).

DISCUSSION

Infection with C. trachomatis is the most common bacterial STD, with more than 2.8 million new cases estimated to occur each year in the United States (6, 7). Reinfections from untreated or inadequately treated sex partners and treatment failures perpetuated the high prevalence over the last years. Doxycycline (100 mg orally twice a day for 7 days) and azithromycin (1,000 mg orally in a single dose) are the preferred treatment regimens for uncomplicated urogenital tract infections with C. trachomatis in women (5, 6). However, treatment failures are frequently reported, suggesting inadequate eradication of chlamydiae from the sites of infection (21). C. trachomatis treatment failure has been observed in up to 14% of patients, whereas only 2% of patients seem to fail initial treatment in Mycobacterium tuberculosis infections (1, 4). In addition, the propensity of C. trachomatis to induce persistent infections when subinhibitory antibiotic concentrations are applied is of major concern. In the persistent state of chlamydiae, the pathogen remains viable in an atypical intracellular inclusion and resists high doses of otherwise effective antimicrobials (13).

The use of MIC testing to predict the therapeutic relevance of antimicrobial activity against intracellular pathogens is questionable because a MIC test does not take into account host cell-specific and environmental factors (21). TKC therefore evolved as a suitable tool to evaluate antimicrobial growth inhibition of bacteria at different intracellular developmental stages (15, 31). In this study, we investigated the impact of low oxygen, which can be observed in the female urogenital tract under physiological conditions, on the antichlamydial activities of doxycycline, azithromycin, rifampin, and moxifloxacin. While the MICs of the respective substances remained unchanged, except for a slight increase for azithromycin, when the tests were performed under hypoxia, significant changes between the antimicrobial activities in the time-dependent killing of intracellular chlamydiae were observed. Reduced effectiveness against intracellular chlamydial growth under hypoxia was observed for doxycycline and azithromycin at therapeutic serum concentrations, whereas moxifloxacin and rifampin lost antichlamydial activity exclusively at reduced concentrations. Therapeutic serum concentrations reflect the concentrations of unbound antibiotics, and it is important to take the binding of antibiotics by serum proteins in the experimental setting into account (12, 24, 32, 38). As protein binding is almost proportional to the percentage of serum in the medium, 5% serum, as used in this study, can be expected to have had almost no effect on the concentrations of unbound antibiotics in our experiments (38). Our findings are not only of interest for the treatment of chlamydial infections, but could also have an impact on treatment strategies for other obligate or facultative intracellular bacteria. With regard to recent findings in human and animal studies, changes in local oxygen levels are linked to the transition between active and latent M. tuberculosis infections (14, 18, 28). Furthermore, it has been shown in a guinea pig model of tuberculosis that M. tuberculosis isolates resistant to antibiotic treatment are found in hypoxic areas of the granuloma (18).

Intracellular activity of antimicrobials requires efficient drug penetration and accumulation within the cell and is constrained by excretion, metabolism, and inactivation, resulting in reduced bioavailability (33). Although various drug transporters are found in different cells (2), active drug efflux pumps, such as the ATP-binding cassette (ABC) transporter proteins MDR-1 and MRP-1, have an important role in the regulation of intracellular drug concentrations of antimicrobials (20, 22, 30). Localized at the plasma membranes of most cell types, MDR-1 and MRP-1 are distributed almost ubiquitously in numerous organs (19). In macrophages and epithelial cells, several groups have demonstrated that expression of MDR-1 and MRP-1 affected the intracellular concentrations of macrolides and quinolones (20, 30). Furthermore, higher concentrations of doxycycline, macrolides, rifampin, and quinolones were required to kill Listeria monocytogenes in MDR-1-overexpressing cells (22). To test whether MDR-1 and MRP-1 were responsible for the reduced effectiveness of doxycycline and azithromycin in cells that were cultured under hypoxia, we analyzed the hypoxia-induced expression and functional relevance of both proteins. Hypoxic cells that were characterized by increased stabilization of HIF-1α showed enhanced transcription of MDR-1 and MRP-1. It was shown previously that HIF-1α directly regulates MDR-1 expression via a hypoxia-responsive element (HRE) within the gene promoter region (8). Our data showing that incubation with CsA, but not with probenecid, restored the antichlamydial activity of doxycycline suggest a functional role of MDR-1 in the reduced effectiveness under hypoxia.

Taken together, our data indicate that intracellular activities of antimicrobials against C. trachomatis depend on the local oxygen availability. Reduced efficacies of first-line antimicrobials in hypoxic areas might account for the insufficient eradication and subsequent persistence of C. trachomatis within the diseased tissue. Improvement of the currently available in vitro models and establishment of human tissue models of C. trachomatis infection for antibiotic testing will help to predict the successful outcome of treatment strategies more precisely.