Background
Ureaplasma species are among the most common isolates from the human urogenital tract [
1,
2]. Although ureaplasmas can be isolated from healthy individuals, epidemiologic studies have shown a strong association between Ureaplasmas and various diseases including non-gonococcal urethritis (NGU), bacterial vaginosis, infertility, prostatitis, epididymitis, urinary tract infection (UTI), nephrolithiasis, postpartum endometritis, chorioamnionitis, spontaneous abortion, premature birth, stillbirth and neonatal pneumonia [
1‐
4]. Animal studies have demonstrated the ability of ureaplasmas to induce pneumonia, pyelonephritis, and struvite uroliths (urinary tract stones primarily composed of magnesium, ammonium, and phosphate) [
5‐
10]. The most common host immune response to
Ureaplasma during disease states involves elevated pro-inflammatory cytokines, most notably IL-1α, IL-1β, IL-6, IL-8, MCP-1 and TNF-α, accompanied by infiltration of neutrophils and macrophages at sites of infection [
1,
10‐
12]. However, little work has been done to characterize the immune response during uncomplicated infections. Therefore, the complex interactions between
Ureaplasma and the host that lead to simple colonization versus inflammation and disease are largely unknown. In a recent study, we showed that the inbred rat strain Fischer 344 (F344) is susceptible to UTI induced by a rat adapted strain of
Ureaplasma parvum isolated from the urine of a patient with recurrent UTI [
13]. As part of that study, we found that 60% of infected F344 rats developed struvite uroliths, which were associated with an exaggerated inflammatory response that is similar to what has been reported in other disease states caused by
Ureaplasma infection [
1,
5,
6,
11,
12]. Interestingly, the other 40% of F344 rats developed uncomplicated UTI that was characterized by low concentrations of pro-inflammatory cytokines in urine as well as mild to moderate lesions in the lower urinary tract. Since F344 rats are an inbred strain, this particular infection model would be useful for identifying the host/
Ureaplasma interactions that confer disease or asymptomatic infection without confounding variables that would be introduced by genetic variability.
In the study reported here, we examined the innate immune response to UTI induced with varying microbial concentrations of U. parvum in the F344 rat. By applying an integrated approach that combines histopathology with cytokine profiling, we were able to identify innate immune response profiles that were significantly different between an uncomplicated UTI and a UTI accompanied by struvite formation. Our findings provide insights into innate immune responses that are likely involved in the development of complicated disease with Ureaplasma.
Methods
Ureaplasmapreparation and culture
A host-adapted strain of
U. parvum, designated strain 257-48 was used for the entire study [
13]. Fifty mls of
U. parvum in logarithmic growth phase was aliquoted into 1 ml volumes and stored at -80°C. This stock was used for all experiments.
For infection studies, one ml of the working stock was grown in 45 ml of 10B broth for 12 to 16 hours at 37°C. The Ureaplasma culture was pelleted by centrifugation at 10,000 × g, at 4°C, for 50 minutes. Due to the delicate nature of Ureaplasma, the pellet was resuspended in 15 to 20 ml of fresh 10B broth instead of saline, to give a final concentration of 109 CFU per ml then serially diluted to produce various inocula that contained 107, 105, 103, and 101 CFU per ml. The CFU of all inocula (including all serial dilutions) were confirmed by culture on A8 agar. For each infection experiment, at least two animals were included in each U. parvum dose group and experiments were replicated a minimum of 5 times.
Inocula and animal tissues were serially diluted 10-fold in 10B broth to 10-10 and 10-5, respectively. For CFU determination, 20 μl from each sample and its corresponding dilutions were plated on A8 agar. Agar plates were incubated at 37°C in 5% CO2; broth cultures were incubated at 37°C in ambient air. Agar cultures were incubated for at least 5 days before colonies were counted to determine CFU.
Animals
Specific pathogen free F344 virgin female rats were purchased from a commercial vendor (Charles River, Indianapolis, IN). All animals ranged in weight from 178–200 grams. Animal colonies were monitored and found free of the following pathogens: Sendai virus, H-1 virus, rat corona virus, sialodacroadenitis virus, reovirus type 3, Kilham rat virus, Hantaan virus, M. pulmonis, respiratory and enteric bacterial pathogens, endoparasites and ectoparasites. All animals were handled in accordance with procedures approved by the University of Florida Institutional Animal Care and Use Committee.
All animals were handled within a biosafety laminar flow hood. Rats were housed in Microisolator® (Lab Products, Inc., Maywood, NJ) cages in the same room under the same temperature and light conditions. Control animals were always handled before infected and housed in separate microisolator cages in order to prevent contamination with Ureaplasma. All food, water, bedding, and caging were autoclaved before use.
Rats were anesthetized and inoculated with sterile broth or
U. parvum inoculum into the bladder as previously described [
13]. For each infection experiment, a minimum of two rats per inoculum dose were infected, so that each dose was represented in each experiment.
Necropsy
Rats were necropsied at two weeks post-infection as previously described [
13]. Prior to euthanasia, free catch urine was collected for cytokine analysis. The bladder was processed for histopathologic evaluation. Each kidney was transected sagittally so that a portion of the renal pelvis was present in each section. One half of each kidney was processed for histopathologic evaluation. The remaining halves of the right and left kidneys for an individual animal were combined, minced in sterile 10B broth, and the medium was aseptically removed and cultured for
U. parvum.
Stone analysis
Bladder calculi were submitted to a commercial laboratory (Louis C. Herring and Co., Orlando, FL) and analyzed by integrated crystallography.
Histopathology
Bladder and kidney tissues were fixed in a paraformaldehyde-lysine-periodate [
14] solution for 24 hours, then washed 3 times in sterile saline and transferred to 70% ethanol prior to processing. Tissues were processed routinely and stained with hematoxylin and eosin (H&E).
Bladder lesions were scored by a system developed in a previous study [
13]. Epithelial changes in bladder tissues were scored as: 0 for none, 1 for minimal hyperplasia, ulceration or effacement of epithelium by inflammation; 2 for mild hyperplasia and rare dense inflammatory infiltrates, and 3 for the same changes noted in a score of 2 but accompanied with marked erosion and/or ulceration of the epithelial surface. Scoring for cell types that comprised the inflammatory infiltrate was: 1 for primarily mononuclear cells (lymphocytes, plasma cells and macrophages), 2 for mononuclear cells and neutrophils, and 3 for mononuclear cells, neutrophils and fibrous infiltrates. Kidney tissues were scored on the basis of total area affected, which was: 1 for less than 10%, 2 for 10 to 50%, and 3 for greater than 50%.
Detection of urinary cytokines
Urine from control and infected rats was analyzed for the presence of cytokines with a multiplex antibody-immobilized bead immunoassay (Lincoplex KIT, Linco Research, Inc., St. Charles, MO). The manufacturer's protocol was followed for the simultaneous detection of the following cytokines and chemokines: GM-CSF, IL-1α, IL-1β, IL-6, IL-10, IL-12p70, IFN-γ, IL-18, GRO/KC (growth related oncogene/keratinocyte chemoattractant- the rat analog for human IL-8), and TNF-α within the same aliquot of urine. Briefly, a standard cocktail was serially diluted in order to develop a standard curve for each analyte that ranged from 3.2 to 2000 pg/ml. Urine samples were diluted in assay buffer to obtain a total volume of 60 μl per well, and run in duplicate as previously described [
13].
Data analysis
Data from multiple experiments were grouped together in order to make statistical analysis possible. Wherever possible, data were analyzed by one-way ANOVA when more than two groups were included in the analysis. Fisher's PLSD test was used when the ANOVA indicated a significant difference among group means. An unpaired student's T test was used for comparisons between two groups. Contingency table analysis was used for comparisons of groups involving nominal data (positive vs. negative). Cytokine pattern recognition analysis was performed using JMP Genomics 3.0 (SAS Institute, Cary, NC). Datasets were initially evaluated by distribution analysis and normalized prior to analysis by one-way ANOVA using row by row modeling and Fischer C correction for multiple comparisons. For all analyses a probability of P < 0.05 was considered significant.
Discussion
Ureaplasmas are an underappreciated pathogen of the urogenital tract. Despite strong epidemiological evidence and even experimental infections in humans that fulfilled Koch's postulates [
15], the etiologic role of
Ureaplasmas is confounded by the isolation of the microbe from the lower urogenital tract of normal, asymptomatic individuals. In addition, the severity of disease for most mycoplasmal infections depends on the host immune response. Therefore, experimental infections in genetically defined animal models will be critical to unraveling the key interactions in the host/parasite relationship that contribute to disease severity. By using a combined approach involving histopathology and cytokine profiling, we were able to further characterize the immune response associated with asymptomatic UTI and UTI complicated with struvite formation.
The F344 rat strain is highly susceptible to development of complicated UTI following experimental inoculation with
U. parvum [
13]. In this study, we showed that varying the inoculum dose of
U. parvum significantly affected the frequency of animals that remained colonized two weeks after inoculation. Therefore, the initial microbial load is important in establishing infection. However, once infected, the proportion of animals that developed complicated UTI in response to varying inocula of
U. parvum did not show a definitive dose response effect. More importantly, the immune response to infection in this group of animals with complicated UTI was consistent, regardless of initial inoculating dose. For example, the cytokine profile and urinary tract pathology of a struvite positive animal that was inoculated with 10
1 CFU was indistinguishable from a struvite positive animal that was inoculated with 10
9 CFU. This suggests that the initial microbial load of
U. parvum is not a critical factor in the development of complicated UTI in this rat strain. Further, it supports the concept that, once infection is established, the host inflammatory response is a key determinant of lesion severity in the urinary tract.
As previously reported [
13], animals with asymptomatic UTI had significantly less pro-inflammatory urine cytokines and tissue damage when compared to rats with struvites. By profiling the entire cytokine milieu, we were able to identify a significant predominance of cytokines such as IFN-γ, IL-18, and MCP-1 in the UTI group that work synergistically to regulate monocyte/macrophage activation [
16‐
18]. An intriguing finding was the significant emphasis of IFN-γ in the urine of animals with asymptomatic UTI, since this cytokine is a potent priming agent for macrophages [
19]. This cytokine profile also coincides with the cellular immune response in these animals that consisted of macrophages, lymphocytes, and plasma cells, which resembles a profile that may be seen during the healing or resolution phase of infection. We cannot rule out that these animals could be displaying a pre-resolution phase to infection, but there are indicators suggesting that these animals have compromised immune defense. For example, the immune profile of these animals was obtained while they were actively colonized with
U. parvum, and 65% exhibited an ascending infection into the kidneys. Further, the microbial load of
U. parvum in animals with asymptomatic UTI was equivalent to animals in the Struvite group. Another intriguing feature in animals with asymptomatic UTI was the overall lack of uroepithelial proliferation that was present in varying degrees in the Negative group as well as the Struvite group. A primary defense mechanism of uroepithelium exposed to bacteria involves desquamation, necrosis or apoptosis followed by proliferation [
20]. Therefore, the overall lack of this response in animals with asymptomatic UTI also implies that uroepithelial defense mechanisms may be perturbed by
U. parvum.
F344 rats with struvite uroliths had a similar clinical profile to what we have previously described [
13]. Specifically, these animals had the greatest concentration of pro-inflammatory cytokines in their urine (GRO/KC – the rat analog of human IL-8, IL-1α, IL-1β, IL-6 and TNF-α) and the most extensive inflammatory changes in bladder tissue. Since IL-1β is a known inducer of IL-8 and GRO chemokines in human and murine epithelial cells [
21‐
23], it is not an unexpected finding that these cytokines are closely linked in their expression. Cytokine pattern analysis showed this cytokine cluster is unique to animals with struvites. Moreover, there is a significant positive correlation between IL-1β, GRO/KC and the degree of histopathologic change, which suggests that IL-1β and GRO/KC are critical elements in a pro-inflammatory loop that leads to chronic active inflammation, epithelial hyperplasia and struvite formation as seen in struvite positive F344 rats. Most of the animals within the struvite group had uroepithelial hyperplasia or erosion with hemorrhage and inflammation within the kidneys, yet none of these animals had uroliths in the renal pelvis that could account for these lesions. Therefore, although mechanical irritation by the urolith itself may partially contribute to epithelial erosion or hyperplasia in the bladder, it cannot entirely account for the lesions that were present in the urinary tract of these animals.
The immune response of animals within the negative group was highly variable and most likely comprises a mixed population of rats, including animals that never became colonized as well as animals that cleared the infection at various time points post-inoculation. Therefore, interpretation of data from this group of animals is difficult and is done with caution. In spite of this limitation, profiling urine cytokine data and bladder lesion scores by inoculum dose was informative. The threshold dose for successful colonization appears to be between log 5 and log 7 CFU, since 64% and 43% of rats respectively were culture negative 2 weeks post-inoculation. The animals within the log 5 and log 7 CFU inoculation groups also had the greatest flux in both pro-inflammatory (TNF-α, IFN-γ and GRO/KC) and anti-inflammatory (IL-4, and IL-10) cytokines. Interestingly, pattern analysis of urine cytokine data showed two distinct clusters. The first cluster identified in the negative group included IL-2, IL-4, IL-10, and TNF-α; these cytokines were notable as they were not part of the cytokine cluster groupings of U. parvum infected animals. Therefore, these cytokines may be critical in directing a more efficient immune response that leads to bacterial clearance with minimal tissue damage. The second cytokine cluster in the negative group included IFN-γ and GRO/KC, which are significant cytokines in the UTI and Struvite groups, respectively. Except for animals that were inoculated with log 9 CFU of U. parvum, the expression pattern of IFN-γ and GRO/KC was not inversely related as they were in culture positive animals. This may be reflecting a more balanced immune response than what is seen in U. parvum positive animals, and we suggest that this balanced response may be critical to resolution of infection and prevention of severe disease.
The variable clinical outcome to experimental inoculation with U. parvum in the F344 rat is an interesting phenomenon since this is an inbred strain. In this study, both genetic and environmental influences on disease were minimized to the extent possible. All of the animals in this study originated from the same colony. Further, rats were housed under the same barrier maintained conditions in order to minimize environmental variability. Despite our efforts, it was common to find that a rat that developed asymptomatic UTI had co-habited the same cage with a rat that developed struvites or was culture negative at time of necropsy. Therefore, external environment could not account for the varying clinical outcome in our study. However, our experimental inoculation procedure may be a critical source of variability. Although attempts were made to reduce mechanical trauma caused by catherization, it is possible that the trauma may have been sufficient to shift the immune response towards a pro-inflammatory profile in a subset of animals. Once this occurred, the pro-inflammatory cycle progressed until infection was resolved (Negative group) or the study was terminated (Struvite group). Another possible explanation for our findings may involve the actual placement of ureaplasmas within the urinary tract at time of inoculation. For example, if the catheter disrupted the uroepithelial barrier so that a sufficient number of ureaplasmas were deposited into the submucosa instead of the mucosal surface, this could elicit a different inflammatory response cascade than what would normally occur if the microorganisms were only present on the mucosal surface of the bladder. The results of this study were similar to what we have previously reported, thus showing the consistency and reproducibility of this model of infection. Moreover, the clinical outcome to ureaplasmal infection in the F344 rat is similar to what occurs in humans. The complex interactions between most mucosal pathogens and the host that lead to uncomplicated colonization versus inflammation and disease are largely unknown. Therefore, this model may be particularly useful for identifying the molecular events that confer asymptomatic infection, complicated infection as well as resolution of infection with an opportunistic pathogen of the urogenital tract.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
LR designed and executed animal infection studies, data analysis and manuscript preparation. MR performed histopathologic evaluation of tissues and developed the lesion scoring system implemented in this study. MBB participated in the design and coordination of the study, and helped draft the manuscript. All authors read and approved the final manuscript.