ABSTRACT
In order to study the interaction of variants in in vivo infection, we employed an azithromycin-resistant mutant (AZ2) and its wild-type parent (SP6) in the guinea pig model of Chlamydia caviae conjunctival infection. When each strain was inoculated individually into conjunctiva, both attained the same level of growth, but AZ2 elicited less pathology. However, when equal numbers of the two strains were inoculated together into the guinea pig conjunctiva, SP6 produced a significantly greater number of inclusion-forming units than AZ2, and the pathology reflected that of a SP6 monoinfection. The goal of this study was to further characterize the dynamics of concomitant infection of these two distinct variants, with particular emphasis on the impact of the host response on the in vivo growth of each organism and the development of pathology. Animals infected with AZ2 had reduced conjunctival infiltration with CD45+ cells and neutrophils as well as a reduced interleukin-8 (IL-8) response. Gene expression of gamma interferon (IFN-γ), tumor necrosis factor alpha (TNF-α), CCL2, and CCL5 was also significantly lower in AZ2-infected animals. The lower inflammatory response induced by AZ2 was associated with its decreased ability to activate NF-κB via Toll-like receptor 2 (TLR2). In general, the inflammatory response in animals infected with both variants was greater than in infection with AZ2 alone, resulting in lower numbers of AZ2 than those of SP6 in the mixed infection. Our results suggest that the ability to elicit an inflammatory response is an important factor in the dynamics of mixed infection with strains that display different pathological phenotypes.
INTRODUCTION
One area in the study of chlamydial infections which has received virtually no attention is how the distribution of variants in an inoculum may influence the outcome of the infection and production of disease in the host. Typically, chlamydial infections have been viewed as consisting of a single serovar, but it has become clear that even within a population of a given serovar, there are variants that can differ dramatically in their physiology (1, 4, 9, 14). The serovar designation is strictly based on the presence of certain epitopes making up the major outer membrane protein, and there is no definitive evidence that the antigenic structure of the major outer membrane protein is directly associated with disease severity (6). It is entirely possible that the composition of the infecting population with respect to the nature of the individual variants present and the number and/or proportion of those variants within the population may be important factors in the development and/or severity of disease. A population may initially consist of multiple variants, or variants may arise during the infection through mutation. In a recent study by our laboratory, a plaque assay was performed with a suspension of Chlamydia caviae obtained from a conjunctival swab of an infected guinea pig (1). The initial inoculum of the guinea pig was from a laboratory stock that had never been plaque purified. There was a large variation in the sizes of the plaques observed, and plaque size was generally related to the growth rate of the organisms in the plaques, as assessed in HeLa cell culture. Moreover, when 10 of these plaques were cultured in the presence of azithromycin (AZM), azithromycin-resistant mutants were obtained at a rate of 3 × 10−8 to 8 × 10−10. This study clearly demonstrated that multiple variants exist within an inoculum and that mutations routinely occur in chlamydiae just as in other bacteria.
That a given inoculum is indeed a mixture of variants and that these variants may differ in their virulence were demonstrated in an elegant study by Sturdevant and coworkers (14). They infected 56 mice with a stock of Chlamydia trachomatis serovar D and then monitored the course of the infection. Not surprisingly, they noted that there was a great variability in the course of infection, with some mice having infections for as short as 10 days and a few mice with infections for as long as 77 days. They then plaque-purified isolates from a single mouse at 10 days and 49 days after infection and, upon reinoculation into naïve mice, observed that the isolates from each time were consistent with regard to the length of infection, i.e., the isolate purified at 10 days consistently produced infections of approximately 10 days. Interestingly, the isolate producing a long infection also elicited more inflammation and more severe pathology; however, there was no difference in the growth rate of either organism in vitro. Thus, these data demonstrate very clearly that a given stock population may have variants with differing virulence phenotypes and that the outcome of the infection and the severity of disease may depend upon the competition of the variants in vivo. This concept of multiple variants within an infecting bacterial population raises several questions. Is there a shifting of the proportional representation of individual variants within a population as a result of in vivo selective pressures? Do faster-growing variants, or variants that are more “fit,” have a competitive advantage within the overall population? Does the ability of one variant to activate a strong host response affect the growth or survival of other variants within the population? Consequently, do some individuals develop more severe pathology because they have a dominant variant that is more virulent than the other variants within the population?
In order to study the interaction of variants in the context of in vivo infection, we employed an azithromycin-resistant mutant (AZ2) and its wild-type parent (SP6), both plaque purified, in the guinea pig model of C. caviae conjunctival infection (1). We previously observed that when each strain was inoculated individually into guinea pig eyes, both strains attained the same level of growth, but the AZ2 strain elicited significantly less pathology. Interestingly, when equal numbers of the two strains were inoculated together into the guinea pig conjunctiva, SP6 produced a significantly greater number of inclusion-forming units (IFU) than AZ2, and the pathology reflected that of infection with SP6 alone. Thus, the data indicated that one strain could have a dominant effect over another with respect to the intensity of disease in vivo. Therefore, the goal of this study was to further characterize the dynamics of concomitant infection with two distinct variants, one with a pathological phenotype and the other with a less pathological phenotype. Our results suggest that the ability to elicit an inflammatory response is an important factor in the dynamics of mixed infection with strains that display different pathological phenotypes.
MATERIALS AND METHODS
Experimental animals.Female Hartley strain guinea pigs, each weighing 450 to 500 g, were obtained from Charles River Laboratories (Boston, MA). All animals were housed individually in cages covered with fiberglass filter tops, given food and water ad libitum, and maintained on a 12/12 light/dark cycle. Each experimental group routinely consisted of five animals. All animal experiments and protocols were approved by the Animal Care and Use Committee of the Arkansas Children's Hospital Research Institute.
Conjunctival infection of guinea pigs.Stocks of the C. caviae variants were made according to standard methodology and were frozen at −80°C in sucrose-phosphate-glutamate buffer (SPG) until needed (10). Guinea pigs were infected in the conjunctiva of both eyes by instilling 20 μl of SPG containing various numbers of IFU of C. caviae SP6 and/or AZ2 directly into the conjunctival sac. When both strains were inoculated concomitantly, a suspension was made of equal numbers of IFU of each prior to inoculation into the eye. SP6 and AZ2 are plaque-purified isolates derived from the conjunctiva of a guinea pig infected with passage 55 of C. caviae, which has been continually passaged in this laboratory since 1974 (1). Pathological changes on each eye were assessed daily using a 0 to 4+ scale while evaluating palpebral and bulbar conjunctivae for erythema, edema, and exudation (10). The scores are defined as follows: slight erythema or edema of either the palpebral or bulbar conjunctiva, 1+; definite erythema or edema of either the palpebral or bulbar conjunctiva, 2+; definite erythema or edema of both the palpebral and bulbar conjunctivae, 3+; definite erythema or edema of both the palpebral and bulbar conjunctivae and the presence of exudate, 4+. Only one individual evaluated the pathology in order to maintain consistency. Conjunctival material for the isolation and quantification of chlamydiae was collected from the conjunctiva using a Dacron swab that was placed in sucrose-phosphate transport medium (13). There was no effect of the swab collection on the conjunctival pathology. The numbers of IFU were determined by culture in McCoy cells. In order to quantify the AZ2 mutants, swab material was cultured in the presence of azithromycin (AZM) (Sigma, St. Louis, MO) at 500 ng/ml. The total number of wild-type IFU was then determined by subtracting the number of IFU obtained in the AZM culture from that obtained in the culture without AZM.
Flow cytometry analysis of lymphocytes in C. caviae-infected conjunctivae.For analysis of the numbers and types of lymphocytes present, infected conjunctivae were harvested and processed individually to produce a single-cell suspension as described previously (15). The entire upper conjunctiva from one eye was used to prepare the suspension. Approximately, 1 × 105 to 2 × 105 cells were stained for individual cell surface markers or isotype controls (5 μg/ml) for 20 min on ice. The cell suspensions were individually incubated first with purified mouse anti-guinea pig CD45 (clone IH-1) or mouse anti-pig CD11R1 (clone MIL4) (AbD Serotec, Raleigh, NC) that recognizes a subset of neutrophils, porcine eosinophils, and NK cells and cross-reacts with guinea pig species. While the antibody recognizes guinea pig neutrophils, it is possible that eosinophils and NK cells may also be counted. Nevertheless, histopathologic examination of infected conjunctivae has shown that the vast majority of inflammatory cells at the site of infection are neutrophils (5), so it is quite likely that the majority of cells counted are neutrophils. Eosinophils are very rarely observed in infected conjunctivae. All primary antibodies were purchased from AbD Serotec (Oxford, United Kingdom). After a wash step, R-phycoerythrin-conjugated goat anti-mouse Ig(H+L) sera (SouthernBiotech, Birmingham, AL) was added. Dead cells were excluded from analysis by the addition of DAPI (4′,6-diamidino-2-phenylindole; Invitrogen) into the final cell suspension. Flow cytometric analysis was performed using a FACSAria cell sorter (BD Biosciences, San Jose, CA), and data were analyzed using FCS Express software.
Quantitative analysis of cytokines and chemokines in C. caviae-infected conjunctivae.Conjunctivae were excised immediately after euthanasia of guinea pigs and stored in RNAlater (Qiagen Scientific, Germantown, MD) at −20°C until further use. Total RNA was extracted from homogenized conjunctivae with TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA). For each conjunctiva sample, 0.5 μg total RNA was treated with RNase-free DNase (Promega, Madison, WI) for 25 min at 37°C, followed by incubation at 70°C for 10 min to inactivate the DNase. RNA samples were then converted into cDNA using oligo(dT), random hexamer primers, and the SuperScript III reverse transcriptase kit from Invitrogen Life Technologies according to the manufacturer's recommendations. Real-time PCR was performed using iQ SYBR green supermix (Bio-Rad Laboratories, Inc., Hercules, CA) in a Bio-Rad iCycler. No-template controls and DNA melting curve analysis were used as controls to ensure the lack of contaminating DNA in the RNA preparations and to rule out primer-dimer formation, respectively. Fold induction of mRNA was determined from the threshold cycle (CT) values normalized for guinea pig β-actin expression and then normalized to the value derived from conjunctivae of healthy, uninfected guinea pigs. Primer sequences for gamma interferon (IFN-γ), interleukin-8 (IL-8), tumor necrosis factor alpha (TNF-α), CCL2 (MCP-1), and CCL5 (RANTES) were published by Wang et al. (15). All primers were synthesized by IDT, Inc. (Coralville, IA).
To quantify IL-8 protein levels in conjunctival secretions, guinea pigs were sedated with ketamine, and absorbent sponges (ear wicks; De Royal Industries, Powell, TN), approximately 2 × 10 mm, were placed under the conjunctiva for 15 to 20 min to absorb tears. The sponges were frozen at −70 C until needed. The sponges were then eluted with phosphate-buffered saline (pH 7.4). Interleukin-8 (IL-8) was measured by enzyme-linked immunosorbent assay (ELISA) using a kit for the quantification of human IL-8 (human CXCL8/IL-8 DuoSet catalog no. DY208; R&D Systems, Minneapolis, MN), which cross-reacts with guinea pig IL-8.
Assessment of NF-κB translocation via TLR2.In order to assess the ability of the C. caviae variants to induce translocation of NF-κB via stimulation of the Toll-like receptor 2 (TLR2), HEK293-hTLR2 cells [transfected with TLR2 (InvivoGen, San Diego, CA)] were used. HEK293 cells not transfected with TLR2 were used as controls for TLR2 stimulation. These cells are also stably transfected with the SEAP (secreted alkaline phosphatase) reporter gene, which is under the control of the NF-κB response element. Cells were grown in Dulbecco's modified Eagle medium (DMEM) (Mediatech, Inc., Manassas, VA), 4.5 g/liter glucose, 10% (vol/vol) fetal bovine serum (FBS) (Atlas Biologicals, Fort Collins, CO), 100 μg/ml Normocin (InvivoGen, San Diego, CA), 2 mM l-glutamine (HyClone, Waltham, MA), and 1× HEK-Blue selection (InvivoGen, San Diego, CA). Cells were seeded in a 96-well plate at 2 × 105 cells/ml in 0.2 ml growth medium (4 × 104 cells/well) and incubated at 37°C and 5% CO2 for 2 days, resulting in 70% confluence. The cells were then infected with either SP6 or AZ2 suspended in DMEM at multiplicities of infection (MOI) of 1, 2, and 5. Cells treated with 1 μg/ml of a synthetic TLR2 ligand, Pam3CSK4 (InvivoGen, San Diego, CA), served as positive controls. Supernatants from untreated cells were used as negative controls. The plates were centrifuged at 3,000 rpm (1,800 × g) and 34°C for 1 h and then incubated at 37°C in 5% CO2. Culture supernatants were collected after 6 and 24 h and stored at −80°C. TLR2 activation was measured using the SEAPorter assay kit (Imgenex, San Diego, CA), according to the manufacturer's instructions.
Statistical analyses.Differences in the pathological responses and the numbers of IFU, CD45+ cells, and polymorphonuclear leukocytes (PMNs) among groups of animals were compared statistically using either a 2-factor analysis of variance (ANOVA) (group, days) with a Tukey test or 2-factor ANOVA with repeated measures followed by a Tukey test if measurements were taken from the same animals. All other assays were compared using a two-tailed t test. Differences were always considered to be statistically different if P was <0.05.
RESULTS
Dose response.We had previously determined that infection of the conjunctiva with 104 IFU of either SP6 or AZ2 resulted in comparable levels of bacterial growth, but when 104 IFU of each strain was inoculated concomitantly into the same conjunctiva, AZ2 did not compete well and had a peak number of IFU that was significantly lower than that for SP6 (1). In order to determine whether the inoculum size of each variant would have an effect on the competition between SP6 and AZ2 in the same tissue, groups of five animals each were inoculated with doses of 103, 104, or 105 IFU of either SP6 or AZ2, and additional groups were inoculated with 103, 104, or 105 IFU each of both SP6 and AZ2. Gross pathology was observed and graded daily. Conjunctival swabs were collected every 3 days to enumerate the number of chlamydial IFU.
At all doses, there was significantly less pathology in animals infected with AZ2 alone than in either the SP6 group alone or the mixed infection of SP6 and AZ2 at all doses (P < 0.001) (Fig. 1). While there were some differences in the kinetics of the infections, the animals infected with SP6 alone or SP6 plus AZ2 had similar peak levels of pathology at each dose tested.
Effect of various doses of SP6, AZ2, and mixed SP6 and AZ2 inocula on the gross pathological response in the conjunctiva. The curves represent repeated daily scores of five animals in each group.
When the numbers of IFU were compared, animals infected with either variant at the same dose generally had similar infection curves, particularly in the initial part of the infection (Fig. 2). However, it was interesting to note that the AZ2 group attained higher peak levels of infection than the SP6 group at 103 and 104 IFU doses and the combined group at 104 and 105 IFU doses even though it elicited less pathology than either the SP6 or combined group. There was also a trend for the animals infected with SP6 to resolve their infections more quickly than either the AZ2 group or the mixed-infection group. At a dose of 103 IFU, the course of the SP6 infection was significantly different than those of both the AZ2 and the mixed infection (P < 0.001). Similarly, at a dose of 104 IFU, the SP6 group was significantly different than the AZ2 group but not the mixed-infection group, and at 105 IFU, the SP6 group was only significantly different than the mixed-infection group. Guinea pigs infected with both variants also had total numbers of IFU similar to those of the animals with the individual variants. Moreover, these animals resolved their infections more slowly than the animals infected with SP6 alone.
Kinetics of bacterial growth following conjunctival infection with various doses of SP6, AZ2, and mixed SP6 and AZ2. The first column represents the growth curves of chlamydiae in animals infected with SP6 and AZ2 individually and the total number of IFU of the mixed infection. The second column represents the SP6 and AZ2 populations within the mixed infection. The curves represent repeated numbers of IFU for five animals in each group. The statistical analysis was performed using a 2-way ANOVA on repeated measures with a Tukey post hoc analysis. The mean competitive index (CI) for the mixed infection is shown at day 6, the time of the peak infection and maximum IFU difference between the two variants in the second column. The CI represents the ratio of the output AZ2/SP6 on day 6 compared to the input AZ2/SP6 ratio. The input dose was a ratio of 1:1.
In contrast, when the individual variants within the mixed-infection group were quantified, SP6 always had a higher peak level than AZ2, so the curves were significantly different at all doses (P < 0.001). Thus, it appears that regardless of the inoculation dose, when AZ2 is inoculated together with SP6, it is not able to compete as well even though in animal monoinfection, it actually can attain higher levels.
Parameters of the inflammatory response.While the gross pathological examination of infected animals is perhaps the most meaningful measure of the inflammatory response, we wanted to measure various parameters of inflammation to determine their relationship to inoculation with each of the variants as well as in a mixed infection. Therefore, groups of guinea pigs were inoculated in both conjunctivae with 104 IFU of either SP6 or AZ2 or both SP6 and AZ2. On days 2, 5, and 8 following inoculation, four animals from each group were euthanized, and tissue was collected for the determination of the numbers of CD45+ cells and PMNs and for quantification of chemokine and cytokine expression by real-time PCR. A portion of the conjunctival tissue was also removed and preserved in RNAlater for quantitative PCR analysis of important chemokines and cytokines. In addition, prior to euthanization, tears were collected for the assessment of protein levels of IL-8. Swabs were collected for IFU determination every 3 days, and the gross conjunctival pathology was scored daily. The course of the infection was as observed previously, with SP6, AZ2, and SP6 plus AZ2 having approximately the same level of infection (data not shown). However, as observed previously in the mixed infection, the course of infection for AZ2 was significantly less than for SP6.
As before, the SP6 group had significantly greater pathology than the AZ2 group (Fig. 3). While the level of pathology in the SP6 and mixed-infection groups was no different, the SP6 group attained the peak level of pathology more quickly, so that the course of infection was significantly different. When the numbers of CD45+ cells and PMNs were quantified in the conjunctival tissue, both cell types increased more rapidly in the SP6 group than in both the AZ2 and mixed-infection groups, paralleling the earlier increase in pathology in that group (Fig. 3). By 6 days after infection, the numbers of both cell types in the mixed-infection group were the same as the SP6 group; however, neither the CD45+ cells nor PMNs attained the same peak levels in the AZ2 group as in the other groups, reflecting the observed lower gross pathological response in the AZ2 group.
CD45+ and PMN levels in conjunctival tissue collected at various times after infection with SP6, AZ2, or SP6 and AZ2 mixed. Each point represents the mean of four animals. The cell numbers are the total number of each cell type in the entire upper conjunctiva of one eye. The mean pathology scores assessed in all animals on the days prior to and at euthanasia are included in the top panel for comparison to the cell numbers. The statistical analysis was performed using a 2-way ANOVA with a Tukey post hoc analysis.
A similar pattern was observed when levels of IL-8 were measured in ocular secretions (Fig. 4). IL-8 levels in the SP6 group were significantly higher than those in the AZ2 group throughout the infection. While over the course of the infection, the IL-8 levels in the mixed-infection group were significantly different than the SP6 group, post hoc analysis indicated that the levels at day 2 and day 5 were not statistically different. The mixed-infection group had somewhat higher levels of IL-8 than the AZ2 group on days 2 and 5, but overall, the curve was not significantly different. It was interesting in that the increase in IL-8 levels in the SP6 and combined groups at day 2 foreshadowed the onset of the pathological response on day 3 in those groups.
IL-8 levels in conjunctival secretions collected at the time of euthanasia. Each point represents the mean of four animals. The mean pathology scores assessed in all animals on the days prior to and at euthanasia are included in the top panel for comparison to the IL-8 levels. The statistical analysis was performed using a 2-way ANOVA with a Tukey post hoc analysis.
Conjunctival tissue from day 2 was processed to measure the expression of IL-1β, gamma interferon (IFN-γ), IL-8, TNF-α, CCL2 (MCP-1), and CCL5 (RANTES) by quantitative PCR (Table 1). Again, similar to the other inflammation parameters, the SP6 group had significantly greater expression of each of the genes than the AZ2 group. Only the level of TNF-α expression was significantly lower in the mixed-infection group than in the SP6 group. There were no significant differences in any of the chemokines/cytokines between the mixed-infection group and the AZ2 group. Thus, in general, the levels observed in the mixed-infection group were always at a midlevel between the SP6 alone and the AZ2.
Cytokine/chemokine expression in conjunctival tissue at day 2 postinfection
TLR2 stimulation.Because there clearly was a defect in the ability of AZ2 to elicit an inflammatory response, especially at the level of expression of various critical chemokines and cytokines, we hypothesized that there was a defect in the ability of AZ2 to activate chemokine and cytokine pathways in the host cells. Since TLR2 has been demonstrated to be a key Toll-like receptor for the production of pathology in chlamydial infections (3), we infected HEK293-hTLR2 cells with various MOI of either SP6 or AZ2 and harvested the supernatants for analysis of NF-κB activation. At both 6 and 24 h postinfection, SP6 elicited significantly more NF-κB activation than AZ2 (Fig. 5), indicating that there is indeed a defect in AZ2 that prevents optimal activation of TLR2, although this defect did not completely inhibit NF-κB activation. The defect likely involves the initial interaction with TLR2, since the response is already lower at 6 h, at which time the organism is still in the very early stages of infection. That the defect is associated only with TLR2 was apparent when we infected HEK293 cells alone with either SP6 or AZ2. No response was seen with either of the C. caviae strains in these cells (data not shown).
Decreased NF-κB activity in vitro in HEK293-hTLR2 cells infected with AZ2 or SP6 at 6 and 24 h postinoculation. The supernatants from the 6-h cultures were undiluted, while those from the 24-h cultures were diluted 10-fold so that the values would be within the range of the standards. The bars represent the means and standard deviations of triplicate cultures. Pam3CSK4 (PAM) served as the positive control. The statistical analysis was performed using a 2-way ANOVA with a Tukey post hoc analysis. The P values represent the comparison of the AZ2 and SP6 variants at each MOI. OD405, optical density at 405 nm.
DISCUSSION
An important concept with regard to understanding the dynamics of chlamydial infection in the human host is that even though a single serovar may be present in the majority of cases, there are still multiple genetic variants within the infecting population (1, 4, 9). Consequently, the course of the infection and resultant pathology may be greatly influenced by the pathogenic potential and the numerical representation of individual variants within the population. In some ways, the ecology of the infection at the local site is much like a niche in nature in which some species are more competitive and can become dominant within the niche. This same dynamic appears to be occurring within a chlamydial population of variants in the ecological niche of the conjunctiva or genital tract. In our previous study, both SP6 and AZ2, when inoculated alone into the conjunctiva of guinea pigs, had similar growth rates and peak numbers of organisms, even though in vitro, SP6 has a faster replication rate and increased burst size compared to AZ2 (1). This suggested that there is a host factor(s) that is responsible for either increasing the growth rate of AZ2 or decreasing the growth rate of SP6. As demonstrated in the present study, it is apparent that SP6 is able to elicit a significantly greater inflammatory response than is AZ2, as evidenced by a more intense pathological response in the conjunctivae of SP6-infected guinea pigs. The in vivo pathology caused by SP6 was supported by demonstration of a significant increase in the number of CD45+ cells and PMNs at the local site as well as an increased level of IL-8 in ocular secretions and increased local expression of TNF-α, IL-1β, CCL2, and IFN-γ.
In contrast, AZ2 was severely deficient in its ability to induce proinflammatory cytokines and IL-8, resulting in a decreased cellular influx of CD45+ cells and PMNs compared to SP6. The inability to elicit a strong inflammatory response was apparently advantageous to AZ2 because the decreased number of PMNs and, consequently, diminished killing of chlamydiae would allow AZ2 to expand its numbers in vivo despite its lower growth rate and lower burst size. It was also of interest in our dose-response experiments that at two doses (103 and 104 IFU), AZ2 actually reached higher levels than SP6. That PMNs are an active participant in the conjunctival C. caviae infection was visualized by us in a recent transmission electron microscopy study (11). We reported that C. caviae was restricted to infection of the superficial epithelial cells and that PMNs were the only inflammatory cell in contact with chlamydia-infected cells. We also observed numerous examples of PMNs with phagocytized elementary bodies and reticulate bodies, indicating an active role for PMNs in controlling the infection. Moreover, we have also demonstrated in mice infected intracervically with Chlamydia muridarum that depletion of PMNs by treatment of mice with antibody to PMNs results in an increase in the number of chlamydiae in the tissue, indicating that PMNs are important to eliminate organisms and control the infection until the adaptive immune response becomes effective (12).
It is likely that SP6 was able to attain comparable in vivo peak levels in spite of the presence of PMNs because its increased replication rate and burst size would compensate for the killing activity of the PMNs. It is interesting to note that, in general, the infection in SP6-infected animals resolved more quickly than infection in the AZ2 group. This could be the result of the greater number of PMNs at the tissue site and consequently a greater microbicidal effect, and/or the adaptive immune response may have been induced more quickly, as suggested by the increased levels of cytokine and chemokine expression in SP6-infected animals. In particular, increased expression of key chemokines for activated T cells (CCL2 and CCL5) was observed in the SP6 group compared to the AZ2 group. We previously reported that depletion of PMNs in guinea pigs during chlamydial conjunctival infection decreased expression of CCL5 and recruitment of CD4 and CD8 T cells by day 7 following infection, indicating the importance of PMNs in the development of the adaptive immune response (5). These experiments signify that in order to understand the dynamics of in vivo infection, it not only is important to consider bacterial physiological parameters such as replication rates but also is critical to determine the capability of a given variant to activate the host response.
The interplay of host response parameters and bacterial physiological parameters becomes even more critical when one considers the dynamics of infection with two different variants in the same infecting population. The availability of two strains of C. caviae with different degrees of virulence and differing sensitivities to azithromycin allowed us to create an artificial population of two variants in order to study the dynamics of mixed infection. Moreover, the guinea pig conjunctival model is ideal because gross pathology can be quantified and swabs can be collected for enumeration of both variants in the same animal as the infection progresses. When SP6 and AZ2 were coinoculated at different total doses in a 1:1 ratio, we observed that at each dose, SP6 attained a peak number of organisms at approximately 1 log greater than AZ2. In addition, in contrast to monoinfection, in which SP6 declined more rapidly than AZ2, SP6 resolved at the same rate as AZ2 in the mixed infection. This experiment repeated the observation we made previously but was nonetheless of interest because in monoinfection, AZ2 actually reaches somewhat higher levels of infection than SP6. In contrast to infection with AZ2 alone in which there are few PMNs elicited, the presence of the more pathogenic SP6 in the mixed infection results in the recruitment of more PMNs. Consequently, because the less robust AZ2 cannot replicate as quickly as SP6, it appears that the rate of killing by PMNs has a greater impact on the number of AZ2, leading to a greater proportion of SP6 in the mixed infection. These results indicate that competition among strains in vivo is multifactorial, including not only factors associated with the organisms themselves such as reproductive capability but also the impact of the host response and the ability of the strains to elicit that response or be affected by it.
With respect to the overall nature of the disease in animals inoculated with both of these variants, even though AZ2 was outcompeted numerically by SP6, it still influenced the intensity of the host response. IL-8 levels were lower in the mixed-infection group than in the SP6 group, as was the transcription of all of the chemokines and cytokines measured, although only the expression of TNF-α was significantly lower. Consequently, both CD45+ cells and PMNs increased more slowly in the mixed-infection group than in the SP6 group. These results likely reflect the influence of the AZ2 within the mixed population, as the IL-8 and chemokine/cytokine responses elicited by AZ2 alone were dramatically lower in all instances. Nevertheless, the gross pathological response produced by the mixed-infection group was not profoundly different from the SP6 group. Thus, these data suggest that even though in a mixed infection, the pathological outcome is dictated to a great extent by the dominant variant, the less pathogenic variant still influences the pathological outcome.
That the pathology caused by a mixed population is affected by numerical representation of each variant within the population and also by the ability of a given variant to elicit an inflammatory response has important implications for human disease. The severity of genital tract or conjunctival disease in humans may depend to some extent on the makeup of the infecting population. If the dominant organism within the population is able to elicit a strong inflammatory response, then the disease may be more severe, but if the dominant organism elicits only a weak response, then the disease could be milder or subclinical. Clearly, our experiments are somewhat simplistic in that we evaluated the impact of only two organisms in a population, one distinctly pathogenic and the other distinctly poorly pathogenic. It is likely that in clinical infections, the populations are far more complex. Nevertheless, our study demonstrates that the composition of the infecting population may be a key factor in the intensity and quality of disease. It is important to note that there may even be fluctuations in the composition of the population as the infection progresses, moves to a different tissue site (e.g., uterus or fallopian tubes), or is transmitted to a new host. The report by Sturdevant and coworkers supports the later concept (14). In their study in which variants characterized by a short infection course and variants characterized by a long infection course were isolated, it was apparent that the inoculating population had both variants, but as the short-term variant was cleared, only the long-term variant remained. If transmission to a new host occurred early in the infection, the recipient would receive both variants, while late in the infection, the recipient would receive only the long-term variants. Therefore, the composition of a transmitted bacterial population may change or remain relatively consistent depending on when in the course of a given infection transmission to a new host occurs. This would not only have an impact on the course of disease in the newly infected recipient but would also impact the representation of a given variant within the host population.
In addition to the use of SP6 and AZ2 in competition experiments, the observation that AZ2 elicits significantly less pathology is of interest. The AZ2 strain was derived by culturing SP6 in the presence of azithromycin and is resistant to azithromycin because of a single base change in the 23S rRNA. The single base difference apparently causes sufficient conformational change in the 23S rRNA so that azithromycin cannot bind to it. The data presented in this study indicate that AZ2 does not elicit a chemokine and cytokine response to the same degree as the parent SP6. That the decreased response was extended to several different chemokines and cytokines suggested that AZ2 may have a defect in its ability to activate signaling pathways. This was confirmed when AZ2 showed a significantly lower activation of NF-κB than that of the parent SP6 when cultured in HEK293-hTLR2 cells. The decreased response was seen as early as 6 h after inoculation of the culture, suggesting that the signaling defect was an early event. The data suggest that AZ2 is unable to activate TLR2 to the same extent as its parent. Whether this defect relates to a downstream event in protein synthesis in AZ2 related to the base change in 23S rRNA or some other difference in the host-parasite interaction in epithelial cells has not been determined. In preliminary experiments, we have observed that other azithromycin-resistant strains with either the same or a different point mutation in the 23S RNA also had decreased NF-κB activity in the HEK293-hTLR2 cell assay (our unpublished data). Nevertheless, it is also possible that there is another gene mutation or polymorphism completely unrelated to the 23S rRNA that is responsible for the decreased TLR2 response. Since AZ2 still maintains its plasmid, it is unlikely that the defect is similar to that seen by O'Connell and colleagues (8). In their study, a plasmidless mutant of C. muridarum was unable to elicit a TLR2 response and, consequently, did not produce the same pathological response as the wild-type strain. Moreover, Carlson et al. (2) reported that a plasmidless C. trachomatis (L2) strain is not as virulent as the plasmid-containing wild type. Miyairi and colleagues also reported that a significant difference in virulence between two C. psittaci variants was unrelated to the presence or absence of the plasmid (7). Together, these studies suggest that there are multiple mechanisms by which chlamydiae may become less or more virulent, and one cannot designate a single polymorphism or mutation as being universally responsible for virulence in chlamydial species.
Regardless of the mechanism, the low pathogenic potential of AZ2 resulting from the decreased ability to activate TLR2 supports the previous observation by Darville and coworkers that activation of TLR2 is a critical event in the production of a pathological response in chlamydial infection (3). The results in our study extend their observations to a different chlamydial species and animal model as well as a different site of infection, indicating that activation of TLR2 is indeed a key factor in the induction of an inflammatory response in chlamydial infection.
ACKNOWLEDGMENTS
This study was supported by grant U19 AI084044 from the NIAID, NIH.
We acknowledge Rachel Binet for her helpful discussions.
The opinions or assertions contained herein are the private ones of the author(s) and are not to be construed as official or reflecting the views of the DoD or the USUHS.
FOOTNOTES
- Received 12 October 2011.
- Returned for modification 6 November 2011.
- Accepted 22 November 2011.
- Accepted manuscript posted online 5 December 2011.
- Copyright © 2012, American Society for Microbiology. All Rights Reserved.