Bacterial Infections | Spotlight

Urine Cytokine and Chemokine Levels Predict Urinary Tract Infection Severity Independent of Uropathogen, Urine Bacterial Burden, Host Genetics, and Host Age

Chelsie E. Armbruster, Sara N. Smith, Lona Mody, Harry L. T. Mobley
Shelley M. Payne, Editor
DOI: 10.1128/IAI.00327-18

ABSTRACT

Urinary tract infections (UTIs) are among the most common infections worldwide. Diagnosing UTIs in older adults poses a significant challenge as asymptomatic colonization is common. Identification of a noninvasive profile that predicts likelihood of progressing from urine colonization to severe disease would provide a significant advantage in clinical practice. We monitored colonization susceptibility, disease severity, and immune response to two uropathogens in two mouse strains across three age groups to identify predictors of infection outcome. Proteus mirabilis caused more severe disease than Escherichia coli, regardless of mouse strain or age, and was associated with differences in interleukin-1β (IL-1β), beta interferon (IFN-β), CXCL5 (LIX), CCL5 (RANTES), and CCL2 (MCP-1). In a comparison of responses to infection across age groups, mature adult mice were better able to control colonization and prevent progression to kidney colonization and bacteremia than young or aged mice, regardless of mouse strain or bacterial species, and this was associated with differences in IL-23, CXCL1, and CCL5. A bimodal distribution was noted for urine colonization, which was strongly associated with bladder CFU counts and the magnitude of the immune response but independent of age or disease severity. To determine the value of urine cytokine and chemokine levels for predicting severe disease, all infection data sets were combined and subjected to a series of logistic regressions. A multivariate model incorporating IL-1β, CXCL1, and CCL2 had strong predictive value for identifying mice that did not develop kidney colonization or bacteremia, regardless of mouse genetic background, age, infecting bacterial species, or urine bacterial burden. In conclusion, urine cytokine profiles could potentially serve as a noninvasive decision support tool in clinical practice and contribute to antimicrobial stewardship.

INTRODUCTION

Urinary tract infections (UTIs) and catheter-associated UTIs (CAUTIs) are among the most common health care-associated infections worldwide (1). If not appropriately managed, UTIs can progress to severe cystitis, pyelonephritis, bacteremia, and sepsis. These infections are especially prevalent in nursing facilities, where they are responsible for a substantial proportion of antibiotic prescriptions (2).

Older adults have an increased risk of developing UTIs, which is thought to be due to predisposing factors including comorbid illnesses, disability, physiologic changes related to aging, and a higher rate of asymptomatic bacteriuria (ABU) (3–5). Indeed, the prevalence of ABU in adults ranges from 1 to 2% in young adults to up to 50% of adults >75 years of age (5–8). The prevalence of UTI also exhibits age-dependent trends, particularly in women, with a decrease between youth and middle age followed by an increase in those >65 years of age (9). The increased incidence of ABU with age poses a significant challenge for diagnosing UTI in older adults. However, differentiating between ABU and UTI and providing appropriate antimicrobial treatment are essential as these infections are common sources of bacteremia, which is associated with a 1-year mortality rate of up to 66% (1, 10, 11). Thus, development of a noninvasive method to predict likelihood of progressing from ABU to UTI and severe disease across age groups would provide a significant advantage in clinical practice, particularly in this vulnerable population.

There is mounting evidence that the microbiology of ABU and UTI changes with age. For instance, Staphylococcus saprophyticus and Escherichia coli are more frequently identified in younger individuals than in older adults, while Proteus mirabilis, Enterococcus species, and Streptococcus agalactiae are more common in older adults, especially those in nursing facilities (5, 12–16). We recently determined that P. mirabilis was the most common organism in catheter-associated UTIs experienced by a cohort of nursing home residents (17), which is consistent with findings of previous studies (18–20). Notably, P. mirabilis was approximately twice as common in individuals aged 75 or greater than in those <75 years of age, while E. coli was more common in those <75 years of age. Thus, the microbiology of urine colonization may provide one noninvasive factor for differentiation of ABU and UTI in different age groups.

One explanation for age-dependent differences in the microbiology of ABU and UTI is a difference in the innate immune response to infection with age. Monocytes, macrophages, and neutrophils are all critical components of the host innate immune response to UTI (21–24). During UTI, macrophages and uroepithelial cells produce proinflammatory cytokines and chemokines that attract neutrophils to the site of infection and regulate antibacterial defenses, including interleukin-8 (IL-8), CCL2 (MCP-1), CCL5 (RANTES), tumor necrosis factor alpha (TNF-α), gamma interferon (IFN-γ), IFN-β, IL-1β, IL-6, IL-10, and IL-17 (22, 25–28). However, many of these defenses become impaired with age, including neutrophil migration, reduced ability of neutrophils to form extracellular traps, reduced phagocytosis and killing capabilities of innate immune cells, reduced cytokine production in response to infection, and slower wound healing (29, 30). Thus, age-dependent differences in the innate immune response to infection could lead to differences in uropathogen colonization susceptibility. This was recently demonstrated for experimental S. agalactiae infection, in which aged mice exhibited increased IL-1α, IL-12 (p40), granulocyte colony-stimulating factor (G-CSF), and CCL5 (RANTES) and decreased IL-3, IL-4, IL-9, and IL-10 compared to levels in young mice (31).

Specific uropathogens also differ in their abilities to stimulate or suppress the innate immune response (32–34), which may further contribute to colonization susceptibility with age. For example, inoculation of young CBA/J mice with P. mirabilis elicits higher levels of IL-1α, IL-1β, IL-6, and granulocyte colony-stimulated factor (G-CSF) than inoculation with E. coli (32) and higher levels of CCL2 (MCP-1), CXCL1 (keratinocyte-derived cytokine, KC), IL-6, IL-10, and IFN-γ than inoculation with Providencia stuartii (33). Similarly, inoculation of young C57BL/6 mice with S. agalactiae induces IL-1α, IL-1β, IL-6, CCL5 (RANTES), and CXCL5 (LIX), but not TNF-α (34), while infection with E. coli induces IL-10, IL-17, IL-23, TNF-α, and CXCL1 (KC) in addition to IL-1α, IL-1β, IL-6, CCL5 (RANTES), and CXCL5 (LIX) (28, 35). It is also notable that parity (multiple gestations) influences S. agalactiae and E. coli colonization susceptibility, and this is at least partly mediated through Toll-like receptor 4 (TLR4)-stimulated inflammation (31, 36). Host genetics also appear to influence age-dependent differences in uropathogen susceptibility as differences in colonization and disease severity have been observed between mouse strains (31, 36).

In this study, we monitored colonization susceptibility, disease severity, and immune response to two uropathogens (Proteus mirabilis and Escherichia coli) in two mouse strains (C57BL/6 and CB6F1 hybrids) across three age groups (young, mature adult, and aged) to identify common predictors of infection outcome. This study utilized virgin mice to exclude potential confounding from parity and two related mouse strains to account for potential differences in host genetics. Overall, our data confirm that colonization susceptibility and disease progression are influenced by host genetics, host age, and uropathogen, but the cytokines and chemokines produced in response to infection may serve as noninvasive predictors of disease severity independent of host genetics, age, or the infecting bacterial species.

RESULTS

Age influences susceptibility to uropathogen colonization.CB6F1 hybrid mice are a cross between female BALB/c and male C57BL/6 mice that are long lived and have fewer strain-specific abnormalities than either parental strain (37). We therefore chose to investigate the impact of age on uropathogen susceptibility first in CB6F1 mice and then in C57BL/6 mice. Virgin CB6F1 mice aged 6 to 8 weeks (young), 4 months (mature adult), or 20 to 24 months (aged) were transurethrally inoculated with P. mirabilis HI4320 or uropathogenic E. coli CFT073. Bacterial burden was enumerated from urine and bladder samples at 96 h postinoculation (hpi) to assess colonization of the urinary tract (Fig. 1). In the ascending model of UTI, kidney colonization occurs only following ascension of the ureters, and during severe infection the bacteria may reach the bloodstream by crossing the epithelial cells of the kidney proximal tubule and the endothelial cells of the capillary, thereby causing bacteremia. Kidney colonization was therefore assessed as a measure of disease severity, and spleen colonization was assessed as a surrogate for bacteremia, as previously described (38).

FIG 1
FIG 1

Age influences P. mirabilis disease severity and E. coli bladder colonization in CB6F1 hybrid mice. CB6F1 mice aged 6 to 8 weeks (young mice), 4 months (mature adult mice), or 20 to 24 months (aged mice) were inoculated transurethrally with 50 μl (1 × 108 CFU/mouse) of P. mirabilis (Pm) (A) or E. coli (Ec) (B). CFU were enumerated from urine, bladder, kidneys, and spleen at 96 hpi. Each symbol represents the value for an individual mouse; gray bars indicate median CFU counts, and dashed lines indicate the limits of detection. For P. mirabilis infections, the numbers of mice used were as follows: young, n = 14; mature adult, n = 13; aged, n = 12. For E. coli infections, the numbers of mice used were as follows: young, n = 6; mature adult, n = 6; aged, n = 5. Significant differences in bacterial burden between age groups were determined by two-way ANOVA with a multiple-comparison posttest, and individual differences by organ were determined by a nonparametric Mann-Whitney test. *, P < 0.05; **, P < 0.01.

For mice inoculated with P. mirabilis, there was a clear difference in the incidence of kidney and spleen colonization between age groups (Fig. 1A). In young mice, 8/14 (57%) exhibited kidney colonization, and 4/14 (29%) had spleen colonization indicative of bacteremia. In contrast, only 4/13 mature adult mice (31%) had kidney colonization, and none showed signs of bacteremia. Aged mice also had a lower incidence of kidney colonization (3/12, 25%), but 5/12 (42%) exhibited spleen colonization. The incidence of bacteremia in aged mice was significantly greater than that in either young mice (P < 0.003 by chi-square) or mature adult mice (P < 0.001 by chi-square), indicating that age may be associated with a greater likelihood of developing secondary bacteremia.

For E. coli infection, all mice exhibited bladder colonization at 96 hpi, but aged mice exhibited a reduction in bacterial burden compared to that of young or mature adult mice (Fig. 1B). Kidney colonization and bacteremia were rare events, which precluded analysis of age-dependent differences in infection severity as measured by CFU count.

To determine if host genetics influence colonization susceptibility and disease severity with age, the infection study was repeated in C57BL/6 mice, a parental strain of the CB6F1 hybrids that is commonly used for experimental UTI studies (39). Virgin female mice were again used to exclude the potential influence of parity. Mice aged 6 to 9 weeks (young), 4 months (mature adult), or 12 to 24 months (aged) were transurethrally inoculated with either P. mirabilis or E. coli, and bacterial burden was enumerated 96 hpi (Fig. 2). For P. mirabilis infection, young mice exhibited the highest bacterial burdens and greatest incidence of kidney colonization (16/19, 84%) and bacteremia (9/19, 47%) (Fig. 2A). Mature adult mice had the lowest bacterial burdens overall, and none of these mice developed bacteremia. In contrast, aged mice exhibited reduced urine and kidney CFU counts compared to levels in young mice but high bacterial burden in the bladder and a similar incidence of bacteremia as young mice (8/24, 33%). Thus, despite differences in urine and bladder colonization levels, mature adult mice of both genetic backgrounds were significantly better at controlling P. mirabilis infection and limiting spread to the kidneys and spleen, while young and aged mice exhibited similar susceptibilities to severe disease. For E. coli infection, age did not significantly impact colonization susceptibility in C57BL/6 mice, and kidney colonization and bacteremia were again rare events (Fig. 2B).

FIG 2
FIG 2

Age influences colonization and disease severity for P. mirabilis but not E. coli in C57BL/6. C57BL/6 mice aged 6 to 9 weeks (young mice), 4 months (mature adult mice), or 12 to 24 months (aged mice) were inoculated transurethrally with 50 μl (1 × 108 CFU/mouse) of P. mirabilis (A) or E. coli (B). CFU were enumerated from the urine, bladder, kidneys, and spleen at 96 hpi. Each symbol represents the value for an individual mouse; gray bars indicate median CFU counts, and dashed lines indicate the limits of detection. For P. mirabilis infections, the numbers of mice used were as follows: young, n = 20; mature adult, n = 5; aged, n = 17. For E. coli infections, the numbers of mice used were as follows: young, n = 5; mature adult, n = 5; aged, n = 8. Significant differences in bacterial burden between age groups were determined by two-way ANOVA with a multiple-comparison posttest, and individual differences by organ were determined by a nonparametric Mann-Whitney test. *, P < 0.05; ***, P < 0.001.

Urine bacterial burden correlates with bladder bacterial burden but not disease severity.A bimodal trend in urine and bladder CFU counts was noted in both CB6F1 hybrid mice and C57BL/6 mice, particularly those inoculated with P. mirabilis. We therefore sought to determine if having a high urine CFU count correlated with a high bladder CFU count, incidence of kidney colonization, or incidence of bacteremia. To conduct this analysis, colonization data from the entire cohort of CB6F1 and C57BL/6 mice were combined (Fig. 3). In this combined cohort of 109 mice, urine CFU counts were correlated with bladder CFU counts (rs = 0.5719, P < 0.0001 by Spearman's correlation), and there was a clear demarcation between high and low urine and bladder colonizations that occurred between 105 and 106 CFU (Fig. 3A). Mice were therefore separated based on urine colonization below (n = 83) or above (n = 26) 2 × 105 CFU/ml. Of the mice exhibiting low urine colonization, only 18% exhibited high bladder colonization (Fig. 3B). Despite having low urine and bladder colonization, kidney colonization was present in 20% of these mice, and bacteremia was present in 18%. For mice with high urine colonization, 88% also had high bladder colonization, 23% had kidney colonization, and 27% had bacteremia (Fig. 3C). Interestingly, kidney colonization was significantly associated with high urine CFU counts and high bladder CFU counts (P < 0.0001 by chi-square for each), but bacteremia was not (P = 0.327 for high urine colonization and P = 0.483 for high bladder colonization by chi-square). Thus, other factors are likely to be better predictors of disease severity than urine CFU counts.

FIG 3
FIG 3

Urine CFU counts correlate with bladder CFU counts but not kidney colonization or bacteremia. The entire cohort of CB6F1 mice from the experiment shown in Fig. 1 and C57BL/6 (BL6) mice from the experiment shown in Fig. 2 were combined and color coded on the basis of bacterial species, mouse genetic background, age, and urine CFU count. Log10 CFU counts from urine, bladder, kidneys, and spleen for the entire cohort (A), for mice with urine colonization below 2 × 105 CFU/ml (B), and for mice with urine colonization at or above 2 × 105 CFU/ml (C) were plotted. Numerical values indicate the percentage and number of mice exhibiting urine colonization above 2 × 105 CFU/ml, bladder colonization above 2 × 105 CFU/ml, kidney colonization above the limit of detection, and spleen colonization above the limit of detection. Dashed lines the indicate limits of detection and 2 × 105. Each symbol represents the CFU count from a single mouse.

Disease severity is correlated with age and bacterial species.To further explore potential predictors of disease severity in the entire cohort of 109 mice, colonization data were separated by mouse age (Fig. 4) or bacterial species (Fig. 5). Bimodal urine and bladder colonization were present across all age groups of mice, with no significant associations (P = 0.430 by chi-square). However, there were striking differences in the incidence of kidney colonization and bacteremia between age groups. Young mice exhibited the highest incidence of kidney colonization (17/31, 55%), followed by mature adult (8/29, 28%) and aged mice (9/49, 18%), and the decline in kidney colonization with age was statistically significant (P = 0.002 by chi-square). Bacteremia also exhibited a significant age-associated trend (P = 0.003 by chi-square), with the highest incidence occurring in aged mice (16/49, 33%) followed by young mice (6/31, 19%), and it was completely absent in all 29 of the mature adult mice regardless of host genetics, bacterial species, or urine CFU count. Thus, mature adult mice may have a significantly different response to urinary tract infection than young or aged mice that allows them to better control the infection and prevent spread to the bloodstream.

FIG 4
FIG 4

Age correlates with incidence of bacteremia, independent of urine or bladder CFU count. The color-coded cohort of mice from the experiment shown in Fig. 3A was separated by age into groups of young (A), mature adult (B), and aged (C) mice. Numerical values indicate the percentage and number of mice exhibiting urine colonization above 2 × 105 CFU/ml, bladder colonization above 2 × 105 CFU/ml, kidney colonization above the limit of detection, and spleen colonization above the limit of detection. Dashed lines indicate limits of detection and 2 × 105. Each symbol represents the CFU count from a single mouse.

FIG 5
FIG 5

Bacterial species correlates with incidence of bacteremia but not bladder CFU count. The color-coded cohort of mice from the experiment shown in Fig. 3A was separated by the infecting bacterial species as follows: (A) P. mirabilis and (B) E. coli. Numerical values indicate the percentage and number of mice exhibiting urine colonization above 2 × 105 CFU/ml, bladder colonization above 2 × 105 CFU/ml, kidney colonization above the limit of detection, and spleen colonization above the limit of detection. Dashed lines indicate limits of detection and 2 × 105. Each symbol represents the CFU count from a single mouse.

Independent of mouse age, the infecting bacterial species had a significant impact on the likelihood of kidney colonization and bacteremia (Fig. 5). Significantly more mice inoculated with P. mirabilis (Fig. 5A) exhibited high urine CFU counts than those inoculated with E. coli (Fig. 5B) (P = 0.013 by chi-square), but a similar percentage of mice from each infection group had high bladder CFU counts (35% for P. mirabilis and 32% for E. coli). However, kidney colonization was more common in mice inoculated with P. mirabilis (P = 0.033 by chi-square), and only 2/34 mice inoculated with E. coli developed bacteremia (6%), whereas 20/75 mice inoculated with P. mirabilis developed bacteremia (27%, P = 0.011 by chi-square). Thus, the host response elicited by inoculation with E. coli may lead to less severe disease than the response elicited by P. mirabilis. Further exploration of how mature adult mice differ from young and aged mice and how E. coli infection differs from P. mirabilis infection may therefore provide predictors of infection severity.

Weight loss and CBC differentials differ with age but do not correlate with disease severity.To determine if mature adult mice experience fewer physiological signs of infection severity than mice of other ages, we analyzed weight loss and complete blood counts (CBCs) in a subset of CB6F1 hybrid mice that exhibited similar levels of bladder colonization (see Fig. S1 and S2 in the supplemental material). As expected, baseline weight was highest in aged mice and lowest in young mice (Fig. S1A). All mice experienced weight loss during the 96-h study, but only young mice exhibited recovery and weight gain. Aged mice exhibited the most dramatic weight loss overall (averaging a decrease of 6 to 7% of baseline weight by 72 hpi), which was significantly more pronounced than losses of mature adult (P < 0.0436) or young mice (P < 0.0048). These findings were independent of bacterial species; there were no significant associations between percent weight loss and incidence of kidney colonization or bacteremia, and no trends that would explain reduced infection severity in mature adult mice were uncovered.

Complete blood counts (CBCs) with automated differentials were also performed on this subset of CB6F1 mice prior to inoculation, at 48 hpi, and at 96 hpi to monitor monocytosis (greater than 10.7 × 103 monocytes/μl white blood cells), neutrophilia (greater than 38.9% of cells being neutrophils), and lymphocytopenia (less than 55.8% of cells being lymphocytes) in response to infection (Fig. S2). The only subtle difference uncovered in the differentials was an increase in neutrophilia and lymphocytopenia in aged mice compared to levels in young or mature adult mice. These findings were again independent of bacterial species; there were no significant associations between CBC differentials and the incidence of kidney colonization or bacteremia and no trends that would explain reduced infection severity in mature adult mice. Thus, weight loss and complete blood counts are not useful predictors of infection severity in our model.

Urine cytokine and chemokine profiles are predictive of disease severity.To determine if the innate immune response to infection correlates with disease severity, a multiplexed bead-based flow cytometry assay was used to quantify 13 cytokines and chemokines in urine samples collected from a subset of mice (n = 81) prior to inoculation and at 6, 24, and 96 hpi. All data from CB6F1 mice are presented in Fig. S3, and data from C57BL/6 mice are presented in Fig. S4.

In CB6F1 mice, TNF-α, IFN-γ, and IFN-β levels differed between age groups inoculated with P. mirabilis during the course of infection, as measured by two-way analysis of variance (ANOVA). When all 13 cytokines and chemokines for a given time point were analyzed, the multiple-comparison posttests indicated that the primary difference at 6 hpi was increased TNF-α for aged mice compared to the level in young or mature adult mice (Fig. 6A). By 24 hpi, aged mice had higher levels of TNF-α and IFN-β (Fig. 6A and C), and by 96 hpi aged mice had higher levels of TNF-α and IFN-γ (Fig. 6A and D). CB6F1 mice inoculated with E. coli similarly exhibited differences in TNF-α, IFN-β, and IFN-γ as well as IL-1β. The main differences at 6 hpi were higher levels of TNF-α and IL-1β in mature adult mice than in mice of the other age groups (Fig. 6A and B). At 24 hpi, mature adult mice exhibited differences in IFN-β, IFN-γ, and IL-1β levels compared to mice of the other age groups (Fig. 6B, C, and D). By 96 hpi, age was no longer a significant source of variation, and the multiple-comparison posttests identified only a slight difference in IFN-γ for aged mice compared to levels for mice of the other age groups (Fig. 6D). Overall, TNF-α and IL-1β levels differed the most between CB6F1 mice inoculated with P. mirabilis and those inoculated with E. coli (Fig. 6A and B).

FIG 6
FIG 6

Age and uropathogen have subtle effects on the innate immune response to infection in CB6F1 mice. Urine collected from mice inoculated with P. mirabilis and E. coli was used to quantify the magnitude of the innate immune response in terms of TNF-α (A), IL-1β (B), IFN-β (C), and IFN-γ (D) levels. For P. mirabilis infection, the numbers of mice used were as follows: young, n = 9; mature adult, n = 9; aged, n = 8. For E. coli infection, the numbers of mice used were as follows: young, n = 6; mature adult, n = 6; aged, n = 5. Error bars represent means ± standard deviations, and dashed lines indicate limits of detection. Significance was determined by two-way ANOVA with a multiple-comparison posttest. Significant differences between the results for different age groups for mice inoculated with P. mirabilis (*) and E. coli (#) are indicated; significant differences between the results for mice of a given age group for inoculation with P. mirabilis versus E. coli are also indicated (∧). One symbol indicates P < 0.05, two symbols indicate P < 0.01, and three symbols indicate P < 0.001.

In C57BL/6 mice, TNF-α was increased at 6 hpi in aged mice inoculated with P. mirabilis compared to the levels in mice of the other age groups (Fig. 7A), but there were no other significant differences. For mice inoculated with E. coli, the only differences were higher levels of LIX (LPS-induced CXC chemokine, CXCL5) in aged mice than in mice of the other age groups at 6 and 24 hpi (Fig. 7B) and an increase in IFN-γ at 96 hpi (Fig. 7C). In comparing the responses to each bacterial species, there were no significant differences in young or mature adult mice at any time point; but aged mice inoculated with P. mirabilis exhibited increased TNF-α at 6 hpi, and aged mice inoculated with E. coli had increased CLXC5 (LIX) at 24 hpi.

FIG 7
FIG 7

Age and uropathogen have subtle effects on the innate immune response to infection in C57BL/6 mice. Urine collected from mice inoculated with P. mirabilis and E. coli was used to quantify the magnitude of the innate immune in terms of TNF-α (A), CXCL5 (LIX) (B), and IFN-γ (C) levels. For P. mirabilis infection, the numbers of mice used were as follows: young, n = 5; mature n =adult, n = 6; aged, n = 9. For E. coli infection, the numbers of mice used were as follows: young, n = 5; mature adult, n = 5; aged, n = 8. Error bars represent means ± standard deviations, and dashed lines indicate limits of detection. Significance was determined by two-way ANOVA with a multiple-comparison posttest. Significant differences between the results for different age groups for mice inoculated with P. mirabilis (*) and E. coli (#) are indicated; significant differences between the results for mice of a given age group for inoculation with P. mirabilis versus E. coli are also indicated (∧). One symbol indicates P < 0.05, two symbols indicate P < 0.01, and three symbols indicate P < 0.001.

In analyzing the combined data set from all 81 mice, urine CFU counts were positively correlated with levels of the following cytokines and chemokines: CCL2 (MCP-1; at 24 hpi, rs = 0.2375, P = 0.0402; 96 hpi, rs = 0.2332, P = 0.0412), IL-6 (96 hpi, rs = 0.2767, P = 0.0149), IL-10 (6 hpi, rs = 0.2402, P = 0.0379; 24 hpi, rs = 0.2848, P = 0.0133; 96 hpi, rs = 0.2337, P = 0.0408), IL-17A (6 hpi, rs = 0.4184, P = 0.0002; 24 hpi, rs = 0.2523, P = 0.0290; 96 hpi, rs = 0.2799, P = 0.0137), TNF-α (24 hpi, rs = 0.3572, P = 0.0017; 96 hpi, rs = 0.3124, P = 0.0044), and IFN-β (6 hpi, rs = 0.2728, P = 0.0179; 24 hpi, rs = 0.2878, P = 0.0123). Urine CFU counts were therefore incorporated into all initial logistic regression models. For binary analyses, urine cytokine and chemokine responses were coded as high or low based on whether they surpassed the following cutoffs for at least one time point postinoculation: IL-1β cutoff of ≥10 pg/ml; IL-6, ≥10 pg/ml; IL-10, ≥3 pg/ml; IL-17A, ≥2.5 pg/ml; IL-23, ≥5 pg/ml; TNF-α, ≥30 pg/ml; IFN-β, ≥10 pg/ml; IFN-γ, ≥10 pg/ml; CXCL1 (KC), ≥3 pg/ml; CXCL5 (LIX), ≥5 pg/ml; CCL3 (MIP-1α), ≥10 pg/ml; CCL5 (RANTES), ≥3 pg/ml; and CCL2 (MCP-1), ≥3 pg/ml.

A series of logistic regressions were used first to attempt to develop a best-fit model for predicting bacteremia. However, the majority of the cytokines and chemokines were dropped from the model for perfectly predicting the outcome or for collinearity. As models for predicting bacteremia could not be directly developed in the full data set, we took advantage of the fact that bacteremia did not occur in mature adult mice or during infection with E. coli, and best-fit models were developed to distinguish mature adult mice from those of the other age groups (Table 1) and mice inoculated with E. coli from those inoculated with P. mirabilis (Table 2). Mature adult mice could be distinguished from young and aged mice by a combination of high IL-23, high CCL5 (RANTES), and low CXCL1 (KC), as shown in Table 1 and Fig. 8. Unfortunately, the model failed to predict bacteremia in the full data set (data not shown). In a different best-fit model, mice inoculated with E. coli were associated with low IL-1β, IFN-β, and CCL5 (RANTES) and high CXCL5 (LIX) and CCL2 (MCP-1) (Table 2). The three cytokines that exhibited the greatest differences between bacterial species are shown in Fig. 9. Again, despite the finding that only two mice inoculated with E. coli developed bacteremia, this profile failed to predict bacteremia in the full data set (data not shown).

View this table:
TABLE 1

IL-23, CXCL1 (KC), and CCL5 (RANTES) levels can distinguish mature adult mice from mice of other age groups

View this table:
TABLE 2

IL-1β, IFN-β, CXCL5 (LIX), CCL5 (RANTES), and MCP levels can distinguish mice inoculated with E. coli from mice inoculated with P. mirabilis

FIG 8
FIG 8

Urine IL-23, CXCL1 (KC), and CCL5 (RANTES) levels distinguish mature adult mice from mice of other age groups. The entire cohort of CB6F1 mice from the experiments shown in Fig. S3 in the supplemental material and C57BL/6 mice from the experiment shown in Fig. S4 were combined and color coded in the same manner as described in the legend of Fig. 3 (on the basis of bacterial species, mouse genetic background, age, and urine CFU count). The color-coded cohort of mice was separated by age as follows: young and aged combined (A, C, and E) and mature adult (B, D, and F). Numerical values indicate the percentages and numbers of mice exhibiting high levels of IL-23 (A and B), CCL5 (RANTES) (C and D), or CXCL1 (KC) (E and F). Dashed lines indicate limits of detection and the cutoff values used to distinguish high from low. Each symbol represents the value for a single mouse.

FIG 9
FIG 9

Urine IL-1β, IFN-β, CCL5 (RANTES), CXCL5 (LIX), and CCL2 (MCP-1) levels distinguish mice inoculated with E. coli from those inoculated with P. mirabilis. The color-coded cohort of mice from the experiment shown in Fig. 8 was separated by bacterial species, as indicated. Numerical values indicate the percentages and numbers of mice exhibiting high levels of IFN-β (A and B), CCL5 (RANTES) (C and D), or CCL2 (MCP-1) (E and F). Dashed lines indicate limits of detection and the cutoff values used to distinguish high from low. Each symbol represents the value from a single mouse.

Given the association between bimodal urine colonization and the levels of certain cytokines and chemokines, an alternative best-fit model was developed to distinguish mice with low urine CFU counts that progressed to bacteremia from those with low urine CFU counts that did not experience severe disease (Table 3). Importantly, kidney colonization provided the strongest predictor of bacteremia in mice with low urine CFU counts, and this remained true when the model was applied to the full data set, including mice with high urine CFU counts. A final best-fit model was therefore developed to predict kidney colonization in the full data set (Table 4 and Fig. 10). The likelihood of kidney colonization was highest in mice with high IL-1β, high CXCL1 (KC), and low CCL2 (MCP-1). When the model was applied to predicting severe disease in the full data set (kidney colonization and/or bacteremia), high CXCL1 (KC) and low CCL2 (MCP-1) were the strongest predictors. The ability of this model to predict disease severity is in agreement with mature adult mice being associated with low CXCL1 (KC) (Table 1) and mice inoculated with E. coli being associated with high CCL2 (MCP-1) and low IL-1β (Table 2). This profile had a high sensitivity (93%) for detecting kidney colonization and severe disease but very low specificity (25% for kidney colonization and 23% for severe disease). The positive predictive value of the profile was only 56% in both cases, but the negative predictive value was 79% for kidney colonization and 76% for severe disease. Thus, this model would not accurately identify individuals who are likely to develop bacteremia, but it would be beneficial for identifying individuals who are not likely to progress from urine colonization to kidney infection or bacteremia and therefore do not necessarily need antimicrobial treatment for urine colonization.

View this table:
TABLE 3

Kidney colonization is the strongest predictor of bacteremia in mice with low urine CFU counts

View this table:
TABLE 4

IL-1β, CXCL1 (KC), and MCP are predictive of kidney colonization

FIG 10
FIG 10

Urine IL-1β, CXCL1 (KC), and CCL2 (MCP-1) levels are predictive of disease severity. The color-coded cohort of mice from the experiment shown in Fig. 8 was separated by those that did not develop kidney colonization above the limit of detection (A, C, and E) and those that did (B, D, and F). Numerical values indicate the percentages and numbers of mice exhibiting high levels of IL-1β (A and B), CXCL1 (KC) (C and D), or CCL2 (MCP-1) (E and F). Dashed lines indicate limits of detection and the cutoff values used to distinguish high from low. Each symbol represents the value from a single mouse.

To determine the impact of host genetics on the predictive value of the final model, a best-fit model was built for distinguishing C57BL/6 mice from CB6F1 hybrid mice based on their cytokine and chemokine responses during infection (Table 5). The only significant factors were high IL-17A and low CCL5 (RANTES) for C57BL/6 mice compared to levels for CB6F1 hybrid mice. Neither of these factors was part of the final best-fit model for kidney colonization in the full data set; including mouse strain in the model had no impact on the fit of the model, and the model still held predictive value when tested in C57BL/6 and CB6F1 mice separately (data not shown). Thus, a combination of high IL-1β, high CXCL1 (KC), and low CCL2 (MCP-1) is capable of predicting the likelihood of developing severe disease, regardless of differences in age, infecting species, or host genetics, in a cohort of 81 mice.

View this table:
TABLE 5

IL-17A, TNF-α, and CCL5 (RANTES) distinguish C57BL/6 from CB6F1 mice

DISCUSSION

Urinary tract infection (UTI) is one of the most common health care-associated infections worldwide (1), with approximately 80% of UTIs due to indwelling urinary catheters. Older adults are at even higher risk for UTI and catheter-associated UTI, and this is especially true of residents in nursing facilities (3, 5–8). These infections are responsible for the majority of antibiotic prescriptions in nursing facilities (2) although roughly one-third of cases are actually misdiagnosed asymptomatic bacteriuria (ABU), which should not be treated with antibiotics (6, 40). On the other hand, appropriate management of UTI and CAUTI is essential as these infections are common sources of bacteremia and subsequent mortality (1, 10, 11). Thus, it is imperative to better understand factors involved in the progression from urine colonization to severe disease and to develop a predictive tool for guiding treatment strategies in this vulnerable population.

Diagnosing UTI and CAUTI in residents in nursing facilities is challenging as the majority of these individuals are older adults for whom many signs and symptoms of infection are infrequent, entirely absent, or subtle and nonspecific. There is also mounting evidence suggesting differences in the microbiology of urine colonization and UTI across the life span, which may influence differences in the signs and symptoms of infection (5, 12–16, 41, 42). However, the underlying mechanisms that influence susceptibility to distinct uropathogens, infection presentation, and the likelihood of progressing to severe disease have remained largely unexplored, particularly with respect to differences across the life span. Here, we monitored colonization susceptibility, disease severity, and the innate immune response to two distinct uropathogens in two different but genetically related mouse strains across three major age groups (young, mature adult, and aged), with an emphasis on identifying common determinants of infection outcome to assess their potential as indicators of disease progression.

To summarize the combined results from both mouse strains, mature adult mice were better able to control bacterial urine colonization than young or aged mice, resulting in reduced bladder and kidney colonization and a reduced incidence of bacteremia. These findings provide support for the hypothesis that the microbiology of urine colonization differs with age, and they are in agreement with the microbiological differences that have been observed in human patients, particularly those with catheter-associated UTIs (5, 12–17). While parameters such as weight loss and white blood cell counts differed with age, these differences were not useful for predicting disease severity. In contrast, a combination of age-dependent and bacterial strain-dependent differences in the cytokine and chemokine profiles elicited during infection were capable of identifying mice that developed kidney colonization and bacteremia, regardless of differences in age, bacterial species, or host genetics. High urine levels of CXCL1 (KC) and low levels of CCL2 (MCP-1) were the strongest predictors of disease severity and exhibited a high sensitivity (93%) but low specificity (25% for kidney colonization and 23% for severe disease). The negative predictive value of the model was 79% for kidney colonization and 76% for severe disease, indicating that it may be useful for identifying individuals who are not likely to progress from urine colonization to kidney infection or bacteremia and therefore do not necessarily need antimicrobial treatment for urine colonization. Considering that the prevalence of severe disease was high in the full cohort of mice (57%), this model would have even greater negative predictive value in a population where the incidence of severe disease is relatively low. If so, this model would have the potential to significantly contribute to antimicrobial stewardship in health care settings, such as nursing facilities.

During infection of the urinary tract, the cascade of events that leads to microbial control and clearance begins with Toll-like receptor (TLR) activation (for a review, see reference 43). Activation of TLR4 by bacterial lipopolysaccharide (LPS) induces a potent proinflammatory cytokine response that includes IL-8 (or the murine functional homologs CXCL1 [KC] and CXCL5 [LIX]) to attract neutrophils to the site of infection and IL-6 to stimulate mucosal IgA production and further inflammatory pathways. Experimentally, the magnitude of IL-6 induction by E. coli isolates has been shown to be associated with the O antigen of LPS, and the induction of a potent innate immune response at early time points postinoculation correlates with reduced bacterial burden (44). Additional factors produced during UTI that recruit and activate neutrophils and regulate antibacterial defenses include CCL2 (MCP-1), CCL5 (RANTES), TNF-α, and IFN-γ (22, 25–28). Thus, our predictive model for disease severity predominately relates to cytokines and chemokines involved in neutrophil and monocyte recruitment to the site of infection. Prior studies in young mice have shown that inoculation with E. coli induces TNF-α, IL-1β, IL-6, IL-10, IL-17, KC (CXCL1), RANTES (CCL5), LIX (CXCL5), and IFN-γ during the course of infection (28, 44–46), and inoculation with P. mirabilis induces CCL2 (MCP-1), RANTES, KC, IL-6, IL-10, IL-17A, IL-1β, TNF-α, and IFN-γ (33). Our cytokine and chemokine data are in agreement with these studies and uncovered differences in urine levels of IL-1β, IFN-β, CCL5 (RANTES), CXCL5 (LIX), and CCL2 (MCP-1) upon direct comparison of E. coli UTI to P. mirabilis UTI across the entire cohort of mice. However, the commonalities between mice, regardless of the inoculating species, will likely provide the most utility for translation to a clinical setting.

Notably, the impact of host genetic background on P. mirabilis colonization has never been addressed experimentally. Our results indicate that the level of urine and bladder colonization by P. mirabilis differs with age in C57BL/6 mice but not in CB6F1 hybrid mice. However, both genetic backgrounds exhibit age-dependent differences in the incidence of kidney colonization and bacteremia. With regard to E. coli infection, aged CB6F1 mice exhibited ∼20-fold reduction in bladder CFU counts compared to levels in young mice, which is consistent with prior work conducted in virgin C3H/HeN mice (36), but no reduction in bladder colonization was observed for C57BL/6 mice, which again suggests that genetic background may contribute to colonization susceptibility with age.

Prior studies have shown similar trends in bladder colonization between BABL/c, C57BL/6, and C3H/HeN mice at 24 and 72 h postinoculation with 1 × 108 CFU of E. coli (47), but there remains a possible role for genetic differences in colonization susceptibilities between C57BL/6 mice and the CB6F1 hybrids that result from the cross between C57BL/6 mice with BALB/c mice. One of the most striking immunological differences between these strains pertains to their T helper type biases. C57BL/6 mice have a Th1-type bias, which is associated with proinflammatory M-1 macrophage responses, including TNF-α, IL-1β, IL-6, IL-17A, IL-23, and production of nitric oxide, while BALB/c mice have a Th2-type bias and a more anti-inflammatory M-2 macrophage response, including induction of IL-10 and pathways that promote tissue repair (48). The association between IL-17A and host genetics in the present study suggests that differences in macrophage responses between the parental C57BL/6 mice and the CB6F1 hybrids may be the primary underlying difference in the context of susceptibility to UTI. BALB/c mice also carry the Nlrp1rs/s allele, which is associated with inflammasome activation, induction of caspase-1, and macrophage lysis, while C57BL/6 mice carry the Nlrp1rr/r allele (46). Considering the importance of IL-1β in our final predictive model and the link between this cytokine and inflammasome activation (49), the contribution of macrophage types and of the inflammasome response to infection during aging warrants further exploration.

Most importantly, several of the cytokines and chemokines produced in response to experimental UTI in mice have also been reported as induced during both human UTI and in vitro stimulation of human bladder urothelial cells, including IL-1β, IL-6, IL-8 (or the murine functional homologs CXCL1 [KC] and CXCL5 [LIX]), IL-10, and CCL5 (RANTES) (34, 39, 44, 50–54). The timing and magnitude of production of the immunomodulatory protein IL-10 in response to E. coli infection appear to contribute to colonization of the urinary tract in a murine infection model and during human UTI (28), indicating that age-related changes in IL-10 production during urine colonization could have a dramatic impact on colonization susceptibility and disease severity. It is also notable that increased urine IL-8 does not appear to be uropathogen specific in humans, which is consistent with our findings in mice, although the exact levels observed in one study were higher from E. coli-infected urine samples than from P. mirabilis-infected urine samples (52). Increased IL-8 has also been observed for patients with pyelonephritis compared to the level in those with lower UTI only, supporting the possible value of this cytokine as a predictor of disease severity (55).

In summary, our data indicate that the urine levels of IL-1β, CXCL1 (KC), and CCL2 (MCP-1) may provide an indication of the likelihood of progressing to severe disease, regardless of the infecting uropathogen, urine bacterial burden, or mouse genetic background. Identification of a urine cytokine profile that predicts likelihood of progression to severe disease would provide a significant advantage in clinical practice. However, additional studies are needed to confirm and further refine the model for murine UTI and to test its utility for predicting likelihood of severe disease in humans.

MATERIALS AND METHODS

Ethics statement.Animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Michigan Medical School (PRO00007111), in accordance with the Office of Laboratory Animal Welfare (OLAW), the U.S. Department of Agriculture (USDA), and the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). Mice were anesthetized with a weight-appropriate dose (0.1 ml for a mouse weighing 20 gm) of ketamine-xylazine (80 to 120 mg/kg ketamine and 5 to 10 mg/kg xylazine) by intraperitoneal injection and euthanized by inhalant anesthetic overdose with vital organ removal.

Bacterial strains and growth conditions.Proteus mirabilis HI4320 (isolated from the urine of a catheterized patient in a chronic care facility [19]) and uropathogenic Escherichia coli CFT073 (isolated from the blood and urine of a patient hospitalized with acute pyelonephritis [56]) were routinely cultured at 37°C with aeration in 5 ml of LB broth (10 g/liter tryptone, 5 g/liter yeast extract, 0.5 g/liter NaCl) or on LB solidified with 1.5% agar.

Mouse model of ascending UTI.Infection studies were carried out as previously described (38, 57). Female, virgin, C57BL/6 and CB6F1 mice aged 6 to 9 weeks (young) or 4 months (mature adult) were obtained from Jackson Laboratories, and 12- to 24-month-old mice (aged) were obtained from the National Institute on Aging specific-pathogen-free rodent colonies. Bacteria were cultured overnight in LB broth, diluted to 2 × 109 CFU/ml (optical density at 600 nm [OD600] of ∼2.0 for P. mirabilis or ∼4.0 for E. coli), and mice were inoculated transurethrally with 50 μl (1 × 108 CFU/mouse). Prior to infection and at 6, 24, 48, and 96 h postinoculation (hpi), mice were weighed using a Navigator NV212 scale (Ohaus), and urine was collected into a sterile microcentrifuge tube. Fifteen microliters of the collected urine was frozen at −80°C for a LEGENDplex assay, and the rest was used for verification of bacterial burden by plating on LB agar (data not shown). Blood was collected into capillary blood collection tubes (containing heparin and EDTA) by retroorbital bleed preinfection and at 48 hpi and by cardiac puncture at 96 hpi for complete blood count (CBC) with five-part differential, performed by the In-Vivo Animal Core at the University of Michigan. Mice were euthanized at 96 hpi, and bladders, kidneys, and spleens were harvested into phosphate-buffered saline (0.128 M NaCl, 0.0027 M KCl, pH 7.4). Tissues were homogenized using an Omni TH homogenizer (Omni International) and plated using an Autoplate 4000 spiral plater (Spiral Biotech). Colonies were enumerated with a QCount automated plate counter (Spiral Biotech), and data are expressed as the number of CFU per milliliter of voided urine and the number of CFU per gram of tissue to account for the larger organ size in older mice.

Quantitation of proinflammatory response.A LEGENDplex (BioLegend) assay was used to quantify 13 cytokines and chemokines from frozen urine samples: RANTES (CCL5), TNF-α (tumor necrosis factor alpha), KC (keratinocyte-derived cytokine, CXCL1), IL-23, IL-10, MIP-1α (macrophage inflammatory protein 1 alpha), IFN-β, LIX (LPS-induced CXC chemokine, CXCL5), IL-6, MCP-1 (monocyte chemoattractant protein 1, CCL2), IL-17A, IFN-γ, and IL-1β according to the manufacturer's protocol. Streptavidin-phycoerythrin (SA-PE) intensity was analyzed by a FACSCanto instrument (Becton Dickinson), and each analyte was quantified relative to the kit standard curve using LEGENDplex software, version 7.0.

Statistical analysis.Significance was assessed by two-way analysis of variance (ANOVA) with a multiple-comparison posttest or nonparametric Mann-Whitney test using GraphPad Prism, version 7 (GraphPad Software, San Diego, CA). All P values are two-tailed at a 95% confidence interval (CI). Pearson's chi-square tests, Spearman correlations, and logistic regression models were performed using Stata/MP, version 13 (StataCorp LP, College Station, TX).

ACKNOWLEDGMENTS

We thank members of the University of Michigan and University at Buffalo Departments of Microbiology and Immunology for helpful comments and critiques. We also thank members of the University of Michigan Geriatrics Center and Institute of Gerontology, the University of Michigan Infection Prevention in Aging research group, the Claude D. Pepper Older American Independence Centers, and the National Institute of Aging rodent core for their guidance and support of this project.

This work was supported by a University of Michigan Geriatrics Center and Claude D. Pepper Older Adults Independence Center Pilot and Exploratory Studies Core grant (C.E.A.), and the Pepper Center is supported by the National Institute of Aging (P30 AG024824). This work was also supported by the National Institute of Diabetes Digestive and Kidney Disorders (K99/R00 DK105205 to C.E.A.), the National Institute of Allergy and Infectious Diseases (R01 AI059722 to H.L.T.M.), and the National Institute of Aging (R01 AG032298, R01 AG041780, and K24 AG050685 to L.M.).

The sponsors were not involved in the study design, methods, subject recruitment, data collections, analysis, or preparation of the paper. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders.

FOOTNOTES

    • Received 1 May 2018.
    • Returned for modification 22 May 2018.
    • Accepted 6 June 2018.
    • Accepted manuscript posted online 11 June 2018.

  • Supplemental material for this article may be found at https://doi.org/10.1128/IAI.00327-18.

REFERENCES

  1. 1.↵
    1. Hooton TM,
    2. Bradley SF,
    3. Cardenas DD,
    4. Colgan R,
    5. Geerlings SE,
    6. Rice JC,
    7. Saint S,
    8. Schaeffer AJ,
    9. Tambayh PA,
    10. Tenke P,
    11. Nicolle LE
    . 2010. Diagnosis, prevention, and treatment of catheter-associated urinary tract infection in adults: 2009 International Clinical Practice Guidelines from the Infectious Diseases Society of America. Clin Infect Dis 50:625–663. doi:10.1086/650482.
    OpenUrlCrossRefPubMedWeb of Science
  2. 2.↵
    1. Loeb M,
    2. Brazil K,
    3. Lohfeld L,
    4. McGeer A,
    5. Simor A,
    6. Stevenson K,
    7. Walter S,
    8. Zoutman D
    . 2002. Optimizing antibiotics in residents of nursing homes: protocol of a randomized trial. BMC Health Serv Res 2:17. doi:10.1186/1472-6963-2-17.
    OpenUrlCrossRefPubMed
  3. 3.↵
    1. Hazelett SE,
    2. Tsai M,
    3. Gareri M,
    4. Allen K
    . 2006. The association between indwelling urinary catheter use in the elderly and urinary tract infection in acute care. BMC Geriatrics 6:15–15. doi:10.1186/1471-2318-6-15.
    OpenUrlCrossRefPubMed
  4. 4.↵
    1. Juthani-Mehta M
    . 2007. Asymptomatic bacteriuria and urinary tract infection in older adults. Clin Geriatr Med 23:585–594. doi:10.1016/j.cger.2007.03.001.
    OpenUrlCrossRefPubMed
  5. 5.↵
    1. Foxman B
    . 2010. The epidemiology of urinary tract infection. Nat Rev Urol 7:653–660. doi:10.1038/nrurol.2010.190.
    OpenUrlCrossRefPubMed
  6. 6.↵
    1. Nicolle LE,
    2. Bradley S,
    3. Colgan R,
    4. Rice JC,
    5. Schaeffer A,
    6. Hooton TM
    , Infectious Diseases Society of America, American Society of Nephrology, American Geriatric Society. 2005. Infectious Diseases Society of America guidelines for the diagnosis and treatment of asymptomatic bacteriuria in adults. Clin Infect Dis 40:643–654. doi:10.1086/427507.
    OpenUrlCrossRefPubMedWeb of Science
  7. 7.↵
    1. Nelson JM,
    2. Good E
    . 2015. Urinary tract infections and asymptomatic bacteriuria in older adults. Nurse Pract 40:43–48. doi:10.1097/01.NPR.0000460855.44987.c1.
    OpenUrlCrossRef
  8. 8.↵
    1. Evans JK,
    2. McOwan A,
    3. Hillman RJ,
    4. Forster GE
    . 1995. Incidence of symptomatic urinary tract infections in HIV seropositive patients and the use of cotrimoxazole as prophylaxis against Pneumocystis carinii pneumonia. Genitourin Med 71:120–122.
    OpenUrlPubMed
  9. 9.↵
    1. Griebling TL
    . 2005. Urologic diseases in America Project: trends in resource use for urinary tract infections in women. J Urol 173:1281–1287.
    OpenUrlCrossRefPubMedWeb of Science
  10. 10.↵
    1. Babich T,
    2. Zusman O,
    3. Elbaz M,
    4. Ben-Zvi H,
    5. Paul M,
    6. Leibovici L,
    7. Avni T
    . 2017. Empirical antibiotic treatment does not improve outcomes in catheter-associated urinary tract infection: prospective cohort study. Clin Infect Dis 65:1799–1805. doi:10.1093/cid/cix680.
    OpenUrlCrossRef
  11. 11.↵
    1. Daniels KR,
    2. Lee GC,
    3. Frei CR
    . 2014. Trends in catheter-associated urinary tract infections among a national cohort of hospitalized adults, 2001–2010. Am J Infect Control 42:17–22. doi:10.1016/j.ajic.2013.06.026.
    OpenUrlCrossRef
  12. 12.↵
    1. Kahlmeter G
    . 2000. The ECO-SENS Project: a prospective, multinational, multicentre epidemiological survey of the prevalence and antimicrobial susceptibility of urinary tract pathogens—interim report. J Antimicrob Chemother 46(Suppl 1):15–22. doi:10.1093/jac/46.suppl_1.15.
    OpenUrlCrossRefPubMedWeb of Science
  13. 13.↵
    1. Medina-Polo J,
    2. Guerrero-Ramos F,
    3. Pérez-Cadavid S,
    4. Arrébola-Pajares A,
    5. Sopeña-Sutil R,
    6. Benítez-Sala R,
    7. Jiménez-Alcaide E,
    8. García-González L,
    9. Alonso-Isa M,
    10. Lara-Isla A,
    11. Passas-Martínez JB,
    12. Tejido-Sánchez Á
    . 2015. Infecciones urinarias adquiridas en la comunidad que requieren hospitalización: factores de riesgo, características microbiológicas y resistencia a antibióticos. Actas Urol Esp 39:104–111. doi:10.1016/j.acuro.2014.08.001.
    OpenUrlCrossRef
  14. 14.↵
    1. Laupland KB,
    2. Ross T,
    3. Pitout JD,
    4. Church DL,
    5. Gregson DB
    . 2007. Community-onset urinary tract infections: a population-based assessment. Infection 35:150–153. doi:10.1007/s15010-007-6180-2.
    OpenUrlCrossRefPubMed
  15. 15.↵
    1. Gravey F,
    2. Loggia G,
    3. de La Blanchardière A,
    4. Cattoir V
    . 2017. Bacterial epidemiology and antimicrobial resistance profiles of urinary specimens of the elderly. Med Malad Infect 47:271–278. doi:10.1016/j.medmal.2017.03.002.
    OpenUrlCrossRef
  16. 16.↵
    1. Ipe DS,
    2. Sundac L,
    3. Benjamin JWH,
    4. Moore KH,
    5. Ulett GC
    . 2013. Asymptomatic bacteriuria: prevalence rates of causal microorganisms, etiology of infection in different patient populations, and recent advances in molecular detection. FEMS Microbiol Lett 346:1–10. doi:10.1111/1574-6968.12204.
    OpenUrlCrossRefPubMed
  17. 17.↵
    1. Armbruster CE,
    2. Prenovost K,
    3. Mobley HLT,
    4. Mody L
    . 2017. How often do clinically diagnosed catheter-associated urinary tract infections in nursing home residents meet standardized criteria? J Am Geriatr Soc 65:395–401. doi:10.1111/jgs.14533.
    OpenUrlCrossRef
  18. 18.↵
    1. Warren JW,
    2. Tenney JH,
    3. Hoopes JM,
    4. Muncie HL,
    5. Anthony WC
    . 1982. A prospective microbiologic study of bacteriuria in patients with chronic indwelling urethral catheters. J Infect Dis 146:719–723. doi:10.1093/infdis/146.6.719.
    OpenUrlCrossRefPubMedWeb of Science
  19. 19.↵
    1. Mobley HLT,
    2. Warren JW
    . 1987. Urease-positive bacteriuria and obstruction of long-term urinary catheters. J Clin Microbiol 25:2216–2217.
    OpenUrlAbstract/FREE Full Text
  20. 20.↵
    1. Breitenbucher RB
    . 1984. Bacterial changes in the urine samples of patients with long-term indwelling catheters. Arch Intern Med 144:1585–1588. doi:10.1001/archinte.1984.00350200077013.
    OpenUrlCrossRefPubMedWeb of Science
  21. 21.↵
    1. Guiton PS,
    2. Hung CS,
    3. Hancock LE,
    4. Caparon MG,
    5. Hultgren SJ
    . 2010. Enterococcal biofilm formation and virulence in an optimized murine model of foreign body-associated urinary tract infections. Infect Immun 78:4166–4175. doi:10.1128/IAI.00711-10.
    OpenUrlAbstract/FREE Full Text
  22. 22.↵
    1. Schiwon M,
    2. Weisheit C,
    3. Franken L,
    4. Gutweiler S,
    5. Dixit A,
    6. Meyer-Schwesinger C,
    7. Pohl JM,
    8. Maurice NJ,
    9. Thiebes S,
    10. Lorenz K,
    11. Quast T,
    12. Fuhrmann M,
    13. Baumgarten G,
    14. Lohse MJ,
    15. Opdenakker G,
    16. Bernhagen J,
    17. Bucala R,
    18. Panzer U,
    19. Kolanus W,
    20. Grone HJ,
    21. Garbi N,
    22. Kastenmuller W,
    23. Knolle PA,
    24. Kurts C,
    25. Engel DR
    . 2014. Crosstalk between sentinel and helper macrophages permits neutrophil migration into infected uroepithelium. Cell 156:456–468. doi:10.1016/j.cell.2014.01.006.
    OpenUrlCrossRefPubMed
  23. 23.↵
    1. Frendeus B,
    2. Godaly G,
    3. Hang L,
    4. Karpman D,
    5. Lundstedt AC,
    6. Svanborg C
    . 2000. Interleukin 8 receptor deficiency confers susceptibility to acute experimental pyelonephritis and may have a human counterpart. J Exp Med 192:881–890. doi:10.1084/jem.192.6.881.
    OpenUrlAbstract/FREE Full Text
  24. 24.↵
    1. Haraoka M,
    2. Hang L,
    3. Frendéus B,
    4. Godaly G,
    5. Burdick M,
    6. Strieter R,
    7. Svanborg C
    . 1999. Neutrophil recruitment and resistance to urinary tract infection. J Infect Dis 180:1220–1229. doi:10.1086/315006.
    OpenUrlCrossRefPubMedWeb of Science
  25. 25.↵
    1. Svensson M,
    2. Irjala H,
    3. Svanborg C,
    4. Godaly G
    . 2008. Effects of epithelial and neutrophil CXCR2 on innate immunity and resistance to kidney infection. Kidney Int 74:81–90. doi:10.1038/ki.2008.105.
    OpenUrlCrossRefPubMedWeb of Science
  26. 26.↵
    1. Olszyna Dariusz P,
    2. Florquin S,
    3. Sewnath M,
    4. Branger J,
    5. Speelman P,
    6. van Deventer Sander JH,
    7. Strieter Robert M,
    8. van der Poll T
    . 2001. CXC chemokine receptor 2 contributes to host defense in murine urinary tract infection. J Infect Dis 184:301–307. doi:10.1086/322030.
    OpenUrlCrossRefPubMedWeb of Science
  27. 27.↵
    1. Godaly G,
    2. Bergsten G,
    3. Hang L,
    4. Fischer H,
    5. Frendéus B,
    6. Lundstedt A-C,
    7. Samuelsson M,
    8. Samuelsson P,
    9. Svanborg C
    . 2001. Neutrophil recruitment, chemokine receptors, and resistance to mucosal infection. J Leukoc Biol 69:899–906.
    OpenUrlPubMedWeb of Science
  28. 28.↵
    1. Duell BL,
    2. Carey AJ,
    3. Tan CK,
    4. Cui X,
    5. Webb RI,
    6. Totsika M,
    7. Schembri MA,
    8. Derrington P,
    9. Irving-Rodgers H,
    10. Brooks AJ,
    11. Cripps AW,
    12. Crowley M,
    13. Ulett GC
    . 2012. Innate transcriptional networks activated in bladder in response to uropathogenic Escherichia coli drive diverse biological pathways and rapid synthesis of IL-10 for defense against bacterial urinary tract infection. J Immunol 188:781–792. doi:10.4049/jimmunol.1101231.
    OpenUrlAbstract/FREE Full Text
  29. 29.↵
    1. Hazeldine J,
    2. Harris P,
    3. Chapple IL,
    4. Grant M,
    5. Greenwood H,
    6. Livesey A,
    7. Sapey E,
    8. Lord JM
    . 2014. Impaired neutrophil extracellular trap formation: a novel defect in the innate immune system of aged individuals. Aging Cell 13:690–698. doi:10.1111/acel.12222.
    OpenUrlCrossRefPubMed
  30. 30.↵
    1. Verschoor CP,
    2. Johnstone J,
    3. Loeb M,
    4. Bramson JL,
    5. Bowdish DM
    . 2014. Anti-pneumococcal deficits of monocyte-derived macrophages from the advanced-age, frail elderly and related impairments in PI3K-AKT signaling. Hum Immunol 75:1192–1196. doi:10.1016/j.humimm.2014.10.004.
    OpenUrlCrossRef
  31. 31.↵
    1. Sullivan MJ,
    2. Carey AJ,
    3. Leclercq SY,
    4. Tan CK,
    5. Ulett GC
    . 2016. Increased age, but not parity predisposes to higher bacteriuria burdens due to Streptococcus urinary tract infection and influences bladder cytokine responses, which develop independent of tissue bacterial loads. PLoS One 11:e0167732. doi:10.1371/journal.pone.0167732.
    OpenUrlCrossRef
  32. 32.↵
    1. Alteri CJ,
    2. Himpsl SD,
    3. Mobley HL
    . 2015. Preferential use of central metabolism in vivo reveals a nutritional basis for polymicrobial infection. PLoS Pathog 11:e1004601. doi:10.1371/journal.ppat.1004601.
    OpenUrlCrossRefPubMed
  33. 33.↵
    1. Armbruster CE,
    2. Smith SN,
    3. Johnson AO,
    4. DeOrnellas V,
    5. Eaton KA,
    6. Yep A,
    7. Mody L,
    8. Wu W,
    9. Mobley HLT
    . 2017. The pathogenic potential of Proteus mirabilis is enhanced by other uropathogens during polymicrobial urinary tract infection. Infect Immun 85:e00808-16. doi:10.1128/IAI.00808-16.
    OpenUrlCrossRef
  34. 34.↵
    1. Tan CK,
    2. Carey AJ,
    3. Cui X,
    4. Webb RI,
    5. Ipe D,
    6. Crowley M,
    7. Cripps AW,
    8. Benjamin WH,
    9. Ulett KB,
    10. Schembri MA,
    11. Ulett GC
    . 2012. Genome-wide mapping of cystitis due to Streptococcus agalactiae and Escherichia coli in mice identifies a unique bladder transcriptome that signifies pathogen-specific antimicrobial defense against urinary tract infection. Infect Immun 80:3145–3160. doi:10.1128/IAI.00023-12.
    OpenUrlAbstract/FREE Full Text
  35. 35.↵
    1. Sivick KE,
    2. Schaller MA,
    3. Smith SN,
    4. Mobley HL
    . 2010. The innate immune response to uropathogenic Escherichia coli involves IL-17A in a murine model of urinary tract infection. J Immunol 184:2065–2075. doi:10.4049/jimmunol.0902386.
    OpenUrlAbstract/FREE Full Text
  36. 36.↵
    1. Kline KA,
    2. Schwartz DJ,
    3. Gilbert NM,
    4. Lewis AL
    . 2014. Impact of host age and parity on susceptibility to severe urinary tract infection in a murine model. PLoS One 9:e97798. doi:10.1371/journal.pone.0097798.
    OpenUrlCrossRef
  37. 37.↵
    1. Miller RA,
    2. Nadon NL
    . 2000. Principles of animal use for gerontological research. J Gerontol A Biol Sci Med Sci 55:B117–B123. doi:10.1093/gerona/55.3.B117.
    OpenUrlCrossRefPubMedWeb of Science
  38. 38.↵
    1. Armbruster CE,
    2. Smith SN,
    3. Yep A,
    4. Mobley HLT
    . 2014. Increased incidence of urolithiasis and bacteremia during Proteus mirabilis and Providencia stuartii coinfection due to synergistic induction of urease activity. J Infect Dis 209:1524–1532. doi:10.1093/infdis/jit663.
    OpenUrlCrossRefPubMed
  39. 39.↵
    1. Carey AJ,
    2. Tan CK,
    3. Ipe DS,
    4. Sullivan MJ,
    5. Cripps AW,
    6. Schembri MA,
    7. Ulett GC
    . 2016. Urinary tract infection of mice to model human disease: practicalities, implications and limitations. Crit Rev Microbiol 42:780–799. doi:10.3109/1040841X.2015.1028885.
    OpenUrlCrossRef
  40. 40.↵
    1. Loeb M,
    2. Simor AE,
    3. Landry L,
    4. Walter S,
    5. McArthur M,
    6. Duffy J,
    7. Kwan D,
    8. McGeer A
    . 2001. Antibiotic use in Ontario facilities that provide chronic care. J Gen Intern Med 16:376–383. doi:10.1046/j.1525-1497.2001.016006376.x.
    OpenUrlCrossRefPubMedWeb of Science
  41. 41.↵
    1. Arinzon Z,
    2. Shabat S,
    3. Peisakh A,
    4. Berner Y
    . 2012. Clinical presentation of urinary tract infection (UTI) differs with aging in women. Arch Gerontol Geriatr 55:145–147. doi:10.1016/j.archger.2011.07.012.
    OpenUrlCrossRefPubMed
  42. 42.↵
    1. Mouton CP,
    2. Bazaldua OV,
    3. Pierce B,
    4. Espino DV
    . 2001. Common infections in older adults. Am Fam Physician 63:257–268.
    OpenUrlPubMedWeb of Science
  43. 43.↵
    1. Spencer JD,
    2. Schwaderer AL,
    3. Becknell B,
    4. Watson J,
    5. Hains DS
    . 2014. The innate immune response during urinary tract infection and pyelonephritis. Pediatr Nephrol 29:1139–1149. doi:10.1007/s00467-013-2513-9.
    OpenUrlCrossRefPubMed
  44. 44.↵
    1. Horvath DJ,
    2. Patel AS,
    3. Mohamed A,
    4. Storm DW,
    5. Singh C,
    6. Li B,
    7. Zhang J,
    8. Koff SA,
    9. Jayanthi VR,
    10. Mason KM,
    11. Justice SS
    . 2016. Association of O-antigen serotype with the magnitude of initial systemic cytokine responses and persistence in the urinary tract. J Bacteriol 198:964–972. doi:10.1128/JB.00664-15.
    OpenUrlAbstract/FREE Full Text
  45. 45.↵
    1. Ingersoll MA,
    2. Kline KA,
    3. Nielsen HV,
    4. Hultgren SJ
    . 2008. G-CSF induction early in uropathogenic Escherichia coli infection of the urinary tract modulates host immunity. Cell Microbiol 10:2568–2578. doi:10.1111/j.1462-5822.2008.01230.x.
    OpenUrlCrossRefPubMedWeb of Science
  46. 46.↵
    1. Garcia TA,
    2. Ventura CL,
    3. Smith MA,
    4. Merrell DS,
    5. O'Brien AD
    . 2013. Cytotoxic necrotizing factor 1 and hemolysin from uropathogenic Escherichia coli elicit different host responses in the murine bladder. Infect Immun 81:99–109. doi:10.1128/IAI.00605-12.
    OpenUrlAbstract/FREE Full Text
  47. 47.↵
    1. Hopkins WJ,
    2. Gendron-Fitzpatrick A,
    3. Balish E,
    4. Uehling DT
    . 1998. Time course and host responses to Escherichia coli urinary tract infection in genetically distinct mouse strains. Infect Immun 66:2798–2802.
    OpenUrlAbstract/FREE Full Text
  48. 48.↵
    1. Sellers RS,
    2. Clifford CB,
    3. Treuting PM,
    4. Brayton C
    . 2012. Immunological variation between inbred laboratory mouse strains: points to consider in phenotyping genetically immunomodified mice. Vet Pathol 49:32–43. doi:10.1177/0300985811429314.
    OpenUrlCrossRefPubMed
  49. 49.↵
    1. Tschopp J,
    2. Schroder K
    . 2010. NLRP3 inflammasome activation: the convergence of multiple signalling pathways on ROS production? Nat Rev Immunol 10:210–215. doi:10.1038/nri2725.
    OpenUrlCrossRefPubMedWeb of Science
  50. 50.↵
    1. Ko YC,
    2. Mukaida N,
    3. Ishiyama S,
    4. Tokue A,
    5. Kawai T,
    6. Matsushima K,
    7. Kasahara T
    . 1993. Elevated interleukin-8 levels in the urine of patients with urinary tract infections. Infect Immun 61:1307–1314.
    OpenUrlAbstract/FREE Full Text
  51. 51.↵
    1. Candela JV,
    2. Park E,
    3. Gerspach JM,
    4. Davidoff R,
    5. Stout L,
    6. Levy SM,
    7. Leach GE,
    8. Bellman GC,
    9. Lad PM
    . 1998. Evaluation of urinary IL-1α and IL-1βin gravid females and patients with bacterial cystitis and microscopic hematuria. Urol Res 26:175–180. doi:10.1007/s002400050043.
    OpenUrlCrossRefPubMedWeb of Science
  52. 52.↵
    1. Oregioni O,
    2. Delaunay P,
    3. Bruna P,
    4. Gaudart A,
    5. Lemichez E,
    6. Boquet P,
    7. Landraud L
    . 2005. Urinary interleukin-8 is elevated in urinary tract infections independently of the causative germs. Cytokine 31:415–418. doi:10.1016/j.cyto.2005.06.009.
    OpenUrlCrossRefPubMedWeb of Science
  53. 53.↵
    1. Otto G,
    2. Braconier J,
    3. Andreasson A,
    4. Svanborg C
    . 1999. Interleukin-6 and disease severity in patients with bacteremic and nonbacteremic febrile urinary tract infection. J Infect Dis 179:172–179. doi:10.1086/314534.
    OpenUrlCrossRefPubMedWeb of Science
  54. 54.↵
    1. Storm DW,
    2. Patel AS,
    3. Horvath DJ,
    4. Li B,
    5. Koff SA,
    6. Justice SS
    . 2012. Relationship among bacterial virulence, bladder dysfunction, vesicoureteral reflux and patterns of urinary tract infection in children. J Urol 188:236–241. doi:10.1016/j.juro.2012.03.025.
    OpenUrlCrossRefPubMed
  55. 55.↵
    1. Sheu J-N,
    2. Chen M-C,
    3. Lue K-H,
    4. Cheng S-L,
    5. Lee I-C,
    6. Chen S-M,
    7. Tsay GJ
    . 2006. Serum and urine levels of interleukin-6 and interleukin-8 in children with acute pyelonephritis. Cytokine 36:276–282. doi:10.1016/j.cyto.2007.02.006.
    OpenUrlCrossRefPubMedWeb of Science
  56. 56.↵
    1. Mobley HL,
    2. Green DM,
    3. Trifillis AL,
    4. Johnson DE,
    5. Chippendale GR,
    6. Lockatell CV,
    7. Jones BD,
    8. Warren JW
    . 1990. Pyelonephritogenic Escherichia coli and killing of cultured human renal proximal tubular epithelial cells: role of hemolysin in some strains. Infect Immun 58:1281–1289.
    OpenUrlAbstract/FREE Full Text
  57. 57.↵
    1. Johnson DE,
    2. Lockatell CV,
    3. Hall-Craigs M,
    4. Mobley HL,
    5. Warren JW
    . 1987. Uropathogenicity in rats and mice of Providencia stuartii from long-term catheterized patients. J Urol 138:632–635. doi:10.1016/S0022-5347(17)43287-3.
    OpenUrlCrossRefPubMedWeb of Science
View Abstract
Back to top
Download PDF
Citation Tools
Print

Alerts
Email

KEYWORDS

CAUTI
Escherichia coli
Proteus mirabilis
UTI
aging
bacteremia
chemokine
cytokine
pyelonephritis
urinary tract infection

Related Articles

Cited By...