Differential Susceptibilities to Azithromycin Treatment of Chlamydial Infection in the Gastrointestinal Tract and Cervix

Abstract

Evidence from animal studies suggests that chlamydiae may persist in the gastrointestinal tract (GI) and be a reservoir for reinfection of the genital tract. We hypothesize that there may be a differential susceptibility of organisms in the GI and genital tracts. To determine the effect of azithromycin on persistent chlamydial gut infection, C57BL/6 and BALB/c mice were infected orally and genitally and treated with azithromycin (Az) orally (20, 40, or 80 mg/kg of body weight), and the numbers of chlamydiae were determined from cervix and cecal tissues. The Az concentration in the cecum and cervix was measured by high-performance liquid chromatography with electrochemical detection (HPLC-ECD). Az treatment cleared genital infection in both C57BL/6 and BALB/c mice; however, GI infection was not cleared with the same doses. HPLC data showed the presence of Az at both sites of infection, and significant amounts of Az were measured in treatment groups. However, no significant difference in Az levels between the cecum and the cervix was observed, indicating similar levels of Az reaching both sites of infection. These data indicate that antibiotic levels that are sufficient to cure genital infection are ineffectual against GI infection. The results suggest a reevaluation of antibiotic therapy for chlamydial infection.

INTRODUCTION

Chlamydia trachomatis remains the most commonly reported infectious disease in the United States. More than 1.4 million cases of chlamydial infection were reported to the CDC in 2011, the largest number of cases ever reported to the CDC for any infection. The reported numbers of cases are highest for adolescent girls (15 to 19 years of age) and young women (20 to 24 years of age). A major problem with chlamydial infections is that in the majority of cases, the infections remain subclinical and “persist” for long periods of time if not treated. Of course, it is difficult to determine if the “persistent” infections are the original infection or the result of treatment failure or the individual has become reinfected from a partner.

Alternatively, it is possible that women become autoinfected from the gastrointestinal (GI) tract. It should be noted that in most animals, including birds, sheep, cattle, pigs, mice, and guinea pigs, the natural site of chlamydial infection is the GI tract, and the organisms are transmitted by the fecal-oral route. We and others have reported that mice remain infected in the GI tract with Chlamydia muridarum indefinitely in the absence of pathology (1–3) even though genital infections resolve within 3 to 4 weeks with an intense inflammatory response. Because chlamydiae persist in the GI tract of other mammalian and avian species, it is plausible that chlamydiae may also establish long-term colonization of the human GI tract. In fact, there is ample evidence that women become infected orally and that chlamydiae can be isolated from rectal swabs in the absence of anal intercourse (4). Jones and colleagues showed that women may be positive in the rectum but negative in the cervix (4). Thus, we have proposed that chlamydiae can persist in the gastrointestinal tract and that women can be reinfected in the genital tract by autoinfection from the rectum. That this is feasible, one need only consider that women commonly develop Escherichia coli urinary tract infections, while men do not.

However, if chlamydiae persist in the GI tract as a reservoir and this serves as a source for genital reinfection, then one must hypothesize that azithromycin (Az) treatment which clears the genital infection is not able to eliminate GI infection in all individuals. That this may be possible is supported by several studies which show treatment failure with respect to rectal chlamydial infections. It has been reported that women at no risk for reinfection had treatment failure rates with azithromycin of 8%, and in men, a higher rate of 23% was seen (5–8). A different report showed a 6% treatment failure rate for rectal chlamydial infection (8), and the most recent study showed an even higher rate, 21.4%, in both men and women (9) when patients were treated with Az. Thus, if chlamydiae persist in the gut, which then serves as a source for reinfection, we hypothesize that there may be a differential susceptibility of chlamydiae in the GI and genital tracts to Az treatment. To explore this hypothesis, we infected mice in the genital tract and GI tract with C. muridarum and evaluated the ability of Az treatment to cure the infection at both sites.

MATERIALS AND METHODS

Experimental animals and chlamydial infection.

Eight-week-old C57BL/6 and BALB/c mice were purchased from Harlan Laboratories (Indianapolis, IN). The C. muridarum Nigg strain, which has been passaged in our laboratory since 1977, when it was obtained from the American Type Culture Collection as a yolk sac preparation, was used for all the infections. C57BL/6 and BALB/c mice were given a 2.5-mg subcutaneous injection of progesterone (Depo-Provera, Upjohn, Kalamazoo, MI) 7 days prior to infection to place the mice in a state of anestrus. All mice were infected both orally and intravaginally. For oral infection, mice received 3 × 10⁶ inclusion-forming units (IFU) by oral gavage using a sterile standard stainless steel feeding needle with a ball tip and without anesthesia to prevent aspiration of chlamydiae into the lungs and death from respiratory infection. Mice were infected intravaginally with a sterile pipette tip at 3 × 10⁵ IFU as described previously (10). All protocols were approved by the Institutional Animal Care and Use Committee.

Chlamydial culture.

Chlamydiae were quantified from the genital tract using a cervicovaginal swab with a Dacroswab and then cultured in McCoy cells according to standard procedure (11). In order to quantify chlamydiae in GI tissue, the cecum was dissected longitudinally and the contents were removed by washing with phosphate-buffered saline (PBS). The epithelium was gently scraped with a scalpel blade and deposited in a sterile Eppendorf tube containing two 4-mm glass beads and 1 ml of 2-sucrose-phosphate buffer transport medium with 0.1 mg/ml gentamicin, 0.2 mg/ml vancomycin, and 2.5 μg/ml of Fungizone. The tubes were vortexed for 1 min and then sonicated for 1 min. After centrifugation to pellet the large cell debris, the supernatants were diluted with Eagle's minimal essential medium supplemented with 10% fetal bovine serum (FBS) and 5% glucose, pH 7.4, and inoculated onto cell monolayers for the determination of IFU. The number of IFU was determined by culturing in HeLa cells according to standard protocol (11). We found that removal of food for 12 h prior to euthanasia and culture of cecal scrapings in HeLa cells increased the sensitivity of detection of chlamydiae in the cecum. Chlamydial inclusions were stained with C. muridarum-specific immune serum and with a fluorescent secondary anti-mouse IgG antibody (Alexa Fluor 488).

Experimental design.

Mice were infected both orally (3 × 10⁶ IFU) and genitally (3 × 10⁵ IFU), and Az (APP Pharmaceuticals, Lake Zurich, IL) treatment was given as a single dose orally after 10 days of infection at three different doses (20, 40, and 80 mg/kg of body weight). Another group remained untreated as a control. Mice treated with doxycycline (APP Pharmaceuticals, Lake Zurich, IL) received 10 mg/kg by intraperitoneal injection for 7 days. Genital swabs were collected on days 3, 6, and 9 before antibiotic treatment and on days 4 and 7 after antibiotic treatment. Antibiotic treatments were based on regimens given clinically for chlamydial infection (12). Mice were euthanized after 11 days after initiation of antibiotic treatment, and cecal epithelial cells were collected for chlamydial isolation on a HeLa cell monolayer. Each experiment was performed twice with 5 animals per group.

Assessment of antibody levels.

Cecal contents were collected at the time of euthanization and placed in a sterile Eppendorf tube containing 150 μl PBS–1% bovine serum albumin (BSA). Tubes were weighed before and after collection of cecal contents to determine the weight of sample in each tube. Samples were diluted with additional PBS with 1% BSA to reach a uniform concentration within each group, normalized to the lowest weight cecal contents. Samples were vortexed for 1 min and centrifuged at 16,000 rpm for 15 min. Supernatants were collected and stored at −80°C. Mice were bled to collect sera, which were stored at −80°C. Specific anti-C. muridarum IgA and IgG levels were measured in cecal and serum samples, respectively, by enzyme-linked immunosorbent assay (ELISA). C. muridarum antigen was purified as previously described (11), and serum or cecal content extract was added to the plate with an initial dilution of 1:10. Goat horseradish peroxidase-conjugated anti-mouse IgG antibody was obtained from Southern Biotechnology Associates (Birmingham, AL), and goat horseradish peroxidase-conjugated anti-mouse IgA was obtained from Serotec (Raleigh, NC). The color reaction was developed with ABTS (2,2′-azinobis(3-ethylbenzthiazolinesulfonic acid) (catalogue no. A1888; Sigma, St. Louis, MO) (3).

Azithromycin extraction from tissues and quantitative determination of azithromycin.

C57BL/6 and BALB/c mice were infected orally (3 × 10⁶ IFU) and genitally (3 × 10⁵ IFU), and Az treatment was given after 10 days of infection at three different doses (20, 40, and 80 mg/kg of body weight). Mice were euthanized after 24 h of Az treatment, and both the cecum and cervix were collected for Az extraction as described previously (13). In brief, ceca and cervices were individually weighed and then homogenized in sterile 1× PBS on ice until a uniform suspension was obtained. Two milliliters of acetonitrile was added, and the samples were homogenized again. Samples were centrifuged for 5 min at 3,300 rpm, and the supernatants were transferred to a 15-ml glass centrifuge tube and evaporated to dryness. To dried samples, 500 μl of 0.06 M potassium carbonate was added, and the residue was vortexed for 30 s. Extraction was carried out with 3 ml of methlyl-tert-butyl ether, and centrifugation was done for 5 min at 3,300 rpm. Az quantitation was carried out from the aqueous phase of the samples using high-performance liquid chromatography (HPLC) with electrochemical detection (ECD). The stock of Az used to treat animals was used as a standard.

RESULTS

C. muridarum in the GI tract is not susceptible to azithromycin treatment.

In order to determine if Az treatment is able to eliminate persistent chlamydial infection in the gastrointestinal tract, mice were infected orally and genitally. Genital swabs collected at days 3, 6, and 9 showed that all animals were positive for chlamydiae; however, no statistically significant difference between C57BL/6 and BALB/c mice was observed in the course of the early infection (data not shown). On day 10, mice received various doses of Az (20, 40, and 80 mg/kg body weight of mice) orally, and cervicovaginal swabs were collected on days 4 and 7 after treatment to determine if the infection had been cleared in the genital tract.

Mice treated with Az cleared genital infection by 4 days after Az treatment at all three doses, compared to findings for control mice that did not receive antibiotic (data not shown). To determine if GI infection had been eliminated, mice were euthanized 11 days after antibiotic treatment, and cecal epithelial cells were processed for isolation of chlamydiae. Of the mice receiving 20 or 40 mg/kg of Az, 95% of each strain remained infected in the cecum (Fig. 1). Of the mice receiving the highest dose (80 mg/kg), 40% of C57BL/6 and 60% of BALB/c mice were still positive for GI infection. Mice of the two strains receiving 80 mg/kg that were positive for chlamydiae in the GI tract after antibiotic treatment had similar IFU levels in the cecum compared to control untreated mice, demonstrating that antibiotic treatment did not have a partial effect on the number of chlamydiae in the cecum (Fig. 2). In contrast, all mice receiving all doses of Az were negative in the genital tract for chlamydial isolation (data not shown). These data clearly indicate that there is a differential susceptibility of chlamydiae to Az, dependent upon the site of infection. In addition, to determine if the 10-fold difference between the oral and genital infection doses had an impact on Az efficacy in the GI tract, C57BL/6 mice were infected both genitally and orally with 3 × 10⁶ IFU, followed by treatment with Az at 20- or 40-mg/kg doses beginning on day 10. Mice cleared genital infection 4 days after azithromycin treatment at both doses of Az; however, mice remained infected in the GI tract even 10 days after antibiotic treatment at both doses (data not shown). This confirms that a higher inoculating dose in the genital tract does not alter the ability of Az to resolve the genital infection.

Fig 1.

Mice do not clear GI infection after azithromycin treatment even at a higher dose (80 mg/kg body weight). The graph represents the percentage of mice infected in the cecum, where the number of mice positive for GI infection was determined and the percentage was calculated by dividing it by the total number of mice infected. Data are averages for 10 animals per strain; “∗∗” indicates that there is a statistically significant difference between control and treatment groups at an 80-mg/kg dose in both strains as per a one-tailed t test (P < 0.01).

Fig 2.

Mice that are positive have IFU levels in the cecum similar to those of control mice. Data are presented as IFU/g of cecum. Each bar is an average for 10 animals per strain. There are no statistically significant differences between control and treatment groups or between strains.

Quantitation of azithromycin.

It is known that neutrophils transport Az to the site of inflammation (14); however, our previous data and those of others showed no inflammation associated with chlamydial infection in the GI tract (1–3). Since there is no inflammation, it is possible that Az was not being transported as readily to the chlamydiae in the cecum; therefore, it was important to determine if Az was actually accumulating at the site of infection, both in the cecum and in the cervix. We measured the levels of Az in both cecal and cervical tissues by HPLC with electrochemical detection (Fig. 3). Significant amounts of Az were detected both in the cervix and cecum of animals treated with all doses compared to findings for animals not treated with antibiotic (one-tailed t test, P < 0.005). The differences between doses were significant for both cervix and cecum samples compared to control samples; however, there was no difference observed between cecal and cervical levels of Az, indicating that similar amounts of antibiotic reached both sites of infection. Thus, the data suggest that some other mechanism exists which prevents Az from having an effect on chlamydiae in the gastrointestinal tract, even though it is effective in the same animal at the cervical site.

Fig 3.

Azithromycin levels in the cecum and cervix were measured by HPLC-ECD. (Upper panel) Results show levels (μg/mg of tissue) of Az in the cervix. (Lower panel) Results show levels of azithromycin in the cecum. Data are averages for 5 animals per strain. A One-tailed t test showed significant differences between control and treatment groups in both cervix and cecal tissues (∗, P < 0.05 according to a one-tailed t test).

Antibody response.

We wanted to determine if Az treatment had an effect on the host antibody response to chlamydial infection. Therefore, cecal IgA and serum IgG levels were measured after antibiotic treatment. No significant differences between C57BL/6 and BALB/c mice in either cecal IgA or serum IgG levels were observed when results were evaluated by a one-tailed t test (Fig. 4). Moreover, there were no significant differences found between animals treated with antibiotic and control animals, suggesting that antibiotic treatment 10 days after infection does not affect the local and systemic antibody response.

Fig 4.

Serum and cecal antibody levels were measured by ELISA. (Upper panel) Cecal IgA levels show the local antibody response in both C57BL/6 and BALB/c mice treated with Az. (Lower panel) These data represent serum antibody levels (IgG) in both C57BL/6 and BALB/c mice treated with Az. Each bar represents an average of 10 animals per strain.

Treatment of chlamydial GI infection with doxycycline.

According to CDC guidelines, people who are infected with Chlamydia are given 1 g of Az as single-dose therapy or 100 mg of doxycycline at multiple doses (2 doses/day) for 1 week. A recent study showed a lower treatment failure rate with a group given doxycycline (5.2%) than with a group treated with azithromycin (23%) in men (15). In order to determine if doxycycline clears GI infection in the mouse, BALB/c mice were infected orally and genitally. Mice were given 0 or 10 mg/kg body weight of doxycycline beginning 10 days after infection for 7 consecutive days to maintain an antibiotic regimen similar to that given to patients.

Mice cleared genital infection with doxycycline treatment by day 4 of antibiotic treatment (data not shown). In contrast to results with Az, chlamydiae were completely cleared from the GI tract in the doxycycline-treated group. Similar to the case with the Az treatment, doxycycline treatment did not alter the specific IgA response in cecal secretions or IgG in serum compared to results for the untreated group (data not shown).

DISCUSSION

A major concern with chlamydial infections in women is that the infections are often long lasting. Ineffective immune response, reinfection from an external source, and treatment failure are all plausible explanations, although they are all difficult to verify with a given patient. It has also been proposed that chlamydiae may be present in a nonreplicating form in the genital tract; however, this concept is based strictly on in vitro studies with penicillin or gamma interferon-treated cultures, and there is no in vivo evidence for such a phenomenon. Women who are diagnosed with chlamydial genital infection are routinely treated with a single 1-g dose of azithromycin, and based on follow-up cultures, the breakthrough rate is relatively low (16–18). If treatment is indeed largely effective, then the primary mechanism for continued infection would be reinfection from an infected partner. Indeed, Whittington and colleagues found that of 792 treated women, 6.3% were culture positive upon follow-up, but of that number, 3.7% indicated in their interview that they had not had sexual activity following treatment (19).

An alternative mechanism for reinfection is that individuals harbor the organism at another anatomical site and become reinfected from that site. It has been well known for decades that chlamydiae infect and persist in the GI tract of virtually all animals, including birds, cattle, sheep, pigs, mice, and guinea pigs, and are transmitted via the fecal-oral route. Perry, Igietseme, and colleagues previously demonstrated that C. muridarum could persist in the GI tract of mice for as long as 260 days in the complete absence of a pathological response (1, 2). We recently characterized this model, showing the target organs to be the cecum and large intestine and having an initial peak infection at 3 to 4 weeks but then decreasing to a constant level of infection for up to 100 days (3). Interestingly, while there was a strong mucosal immune response to the infection, the cell-mediated immune response decreased to baseline even though the infection was still present. We have hypothesized that if chlamydiae inhabit the GI tract of most other hosts, there is no reason that this could not occur in humans as well, since there are many scenarios by which women and men can be infected orally. Jones and colleagues reported that 11% of women with positive genital cultures had positive rectal cultures as well, compared to only 2.7% with negative genital cultures (4). Overall, 5.2% of women had positive rectal cultures, and no association was seen between positive rectal cultures and a history of rectal intercourse. Therefore, it is entirely conceivable that women could become reinfected in the genital tract through contamination from rectal secretions, much as they would acquire an E. coli urinary tract infection. Nevertheless, the study by Jones did not demonstrate long-term carriage of chlamydiae in the GI tract. It is important that new studies be performed to confirm the GI tract as a site of persistent infection in humans. Based on the data presented in this study, clear information regarding GI carriage in humans would suggest that a reevaluation of management strategies is important.

If this scenario is indeed realistic for humans, then it would follow that there may be a differential susceptibility of chlamydiae to Az in the genital tract versus the GI tract. In this study, we have demonstrated using the mouse model that Az treatment is able to clear genital tract infection but is unable to eliminate chlamydial infection in the cecum within the same animal. Three different doses were effective in clearing the genital infection, and only at the highest dose were some mice cured of the infection in the cecum; nevertheless, at that dose, 40% of C57BL/6 mice and 60% of BALB/c mice were still positive. The 80-mg/kg dose is roughly equivalent to the dose given humans, although it is obviously difficult to extrapolate directly from the mouse to the human. There was no major difference in the response to treatment in either strain of mice, even though BALB/c mice are generally more susceptible to infection than C57BL/6 mice. Interestingly, in the mice receiving the 80-mg/kg dose, there was no difference in the number of IFU in the cecum from that of untreated mice. Therefore, these data suggest the possibility that treatment of women with Az may be effective against genital infection but is less effective in eliminating GI infection, so that following treatment, they may become autoinfected from rectal secretions. That rectal infections are more refractive to Az treatment has been shown in several clinical studies. A study in men at no risk for reinfection on rectal chlamydial infection showed a 6% treatment failure rate (8). A more recent report of both men and women for rectal chlamydial infection showed a higher treatment failure rate of 21.4% (9).

The mechanism for increased resistance to Az treatment in the GI tract is still unclear. When we measured the levels of Az in the GI and genital tract by HPLC, results indicated that Az was absorbed into the cecum and cervix at all three doses at comparable levels in both C57BL/6 and BALB/c mice. Thus, it would appear that the presence of antibiotic in the cecum is not the limiting factor in GI infection clearance. It is possible that GI tract microbiota may be diluting the Az impact on chlamydiae in cecal epithelial cells. In addition to this, it has been suggested by Horner (7) that heterotypic resistance could develop, and this phenomenon happens with chlamydiae at high infectious loads (20, 21). Furthermore, the lack of an inflammatory response would remove the possibility of polymorphonuclear leukocytes (PMNs) delivering the Az directly to the site of infection (14). Another possibility is that treatment failures could be a consequence of the single-dose therapy for Az. It has been shown that treatment with multiple doses of doxycycline has a lower failure rate for rectal infections than a single-dose-of-azithromycin treatment of chlamydial infections (8, 9). In our study, we also demonstrated that a 7-day regimen of doxycycline was able to eliminate the organism at both sites. It is possible that had we used multiple doses of Az, we may have also been able to eliminate the infection. These data may suggest that perhaps multiple doses of Az should be given to patients as well.

The concept that the reservoir for persistent infection is in the GI tract has substantial biological validity, and there is much documentation to demonstrate that women are indeed infected in the GI tract. If, as this study suggests, Az is not as effective in eliminating GI chlamydial infection, then the treatment protocols for chlamydial genital infections may have to be reexamined.