Cystitis outcomes are bacterial strain-dependent

First we assessed differences in host outcomes 28 days after infection with 10⁸ CFU of either UTI89 or CFT073 (Fig 1C and 1D). We found that CFT073 caused chronic cystitis lasting four weeks, but at a significantly lower rate than UTI89 did (30% incidence for CFT073 vs. approximately 50% incidence with UTI89; Fig 2A), consistent with a previous study [21]. Whether infected with UTI89 or CFT073, mice with chronic cystitis at 28 dpi had significantly higher bladder bacterial burdens (Fig 2B) and bladder edema (Fig 2C) than mice that spontaneously resolved their infection, in accordance with our definition of chronic cystitis. Kidney titers were also higher in mice with chronic cystitis than those that resolved their infection (Fig 2D). In a blinded analysis of scanning electron micrographs of chronically infected bladders at 28 dpi, there were no obvious differences between chronic UTI89 and chronic CFT073 cystitis, with both infections causing epithelial disruption, immune cell infiltration, and abundant extracellular rod-shaped bacteria (Fig 2E). However, compared to chronic UTI89 cystitis, chronic CFT073 cystitis at 28 dpi trended toward reduced: bladder titers (Fig 2B), bladder edema (Fig 2C), and kidney titers (Fig 2D), suggesting that chronic CFT073 cystitis at this time point may be less robust than chronic UTI89 cystitis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. CFT073 infection is less robust than UTI89 at 28 dpi and clears over time.

Juvenile Naive (7–8 week old) female C3H/HeN mice were infected with 10⁸ CFU UTI89, CFT073, or CFT073ΔfimH::FimH^UTI89 (“CFT073-FimH”) and monitored for 28 days. (A) Percentage of mice (infected as indicated) that developed chronic cystitis, which was defined as persistent, high-titer bacteriuria (>10⁴ CFU/ ml urine), bladder bacterial burden >10⁴ CFU/bladder and bladder inflammation at time of sacrifice [14]. * P < 0.05, ** P < 0.01, Fisher’s exact test. (B-D) Bacterial burdens (B) and edema (C) and kidney burdens (D) at 28 dpi in mice with chronic cystitis caused by UTI89 or CFT073 (“Chronic”) compared to mice that had spontaneously resolved these infections (“Resolved”). (E) Scanning electron microscopy (SEM) of bladders harvested from mice with chronic UTI89 or chronic CFT073 cystitis at 28 dpi. Representative images from N = 1 experiment with n = 3 mice per group are shown; scale bars, 50 μm. (F) UTI89 and CFT073 urine titers over time in mice meeting the criteria for chronic cystitis (persistent, high-titer bacteriuria >10⁴ CFU/ ml urine). (G) Mice infected with UTI89 (n = 12), CFT073 (n = 18), or CFT073ΔfimH::FimH^UTI89 (n = 15) were followed for 14 weeks. Dashed lines indicate our previously established cutoff for high-titer persistent bacteriuria, 10⁴ CFU/ml urine. Note that 4 out of 18 CFT073-infected mice resolved the infection between weeks 3 and 4 (indicated by the arrow); this “late-resolving” phenotype contributes to the lower incidence of CFT073 chronic cystitis at 4wpi, seen in 2A. (H) SEM at 28 dpi of a mouse that resolved a CFT073 infection between days 21 and 28 (a so-called “late resolver”); scale bar, 50 μm. Data are combined from two to three independent experiments; data points represent actual values for each individual mouse, zeros are plotted at the limit of detection, and bars indicate median values. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001, Kruskal-Wallis test with Dunn’s multiple test correction.

https://doi.org/10.1371/journal.ppat.1007457.g002

During chronic cystitis, CFT073 urine bacterial titers dropped significantly between 1 and 28 days post infection (dpi), though they remained above our predefined cutoff of 10⁴ CFU (Fig 2F). As well, in the course of this study we repeatedly observed that a subset of mice resolved their CFT073 infection between 21 and 28 days (i.e. persistent bacteriuria >10⁴ CFU until 21 dpi, but urine and/or bladder titers <10⁴ at 28 dpi), which we dubbed “late resolution.” Late resolution was not observed with UTI89 infection. Extending our studies past 28 days of infection revealed that every CFT073-infected mouse cleared bacteriuria within eight weeks (Fig 2G; note that four of 18 mice had “late resolution” between weeks three and four). In contrast, urine titers remained persistently high for 14 weeks post infection in the UTI89-inoculated mice that had developed chronic cystitis (Fig 2G). Analysis of a late-resolving CFT073-inoculated bladder by SEM revealed a calm surface at 28 dpi, with an intact epithelium having no immune cells or visible bacteria (Fig 2H).

A critical distinction between UTI89 and CFT073 is each strain’s allele of the mannose-binding fimH adhesin, which tips the type 1 pilus and mediates bladder colonization [23]. UTI89 has the FimH variant A62/V163, whereas CFT073 has the FimH variant S62/A163, which has a lower in vitro affinity for mannose and reduced virulence in the mouse bladder [21]. We found that a strain of CFT073 that expresses UTI89’s higher mannose affinity fimH allele, CFT073ΔfimH::FimH^UTI89, had a significantly greater incidence of chronic cystitis at 28 dpi than wild-type CFT073 did (Fig 2A), in line with previous findings [21]. Strikingly, this strain was able to persist longer in the bladder than wild-type CFT073, but nevertheless all but one mouse eventually cleared the infection over the course of 14 weeks (Fig 2G). These observations indicate that while the high mannose affinity variant FimH enhances and prolongs the colonization of CFT073ΔfimH::FimH^UTI89 in the mouse urinary tract relative to the wild type strain, other factors unique to CFT073 account for its propensity for being cleared from the host. As well, in the course of this work we became aware that the CFT073 strain used has a frameshift mutation in rpoS [24]. However, a CFT073 strain with a wildtype rpoS allele still had “late resolution” between days 21 and 28, and was cleared from all mice within eight weeks (S1 Fig), indicating that the rpoS frameshift mutation does not explain CFT073 clearance relative to UTI89.

An acute checkpoint predicts chronic CFT073 infection, which leaves an imprint on the bladder

We have previously shown that chronic UTI89 cystitis in C3H/HeN mice is predicted by an immune checkpoint at 24 hours post infection (Fig 1B), and that chronic bladder inflammation lasting four weeks results in a bladder imprint that persists after antibiotic therapy (Fig 1E). At 24 hpi, mice infected with CFT073 or UTI89 did not significantly differ in overall urine bacterial titers, pyuria, or pro-inflammatory serum cytokines. However, when mice were grouped by disease outcome, regardless of whether infected with UTI89 or CFT073 those that ultimately would develop chronic cystitis lasting four weeks had activation of the immune checkpoint, indicated by: higher urine bacterial burdens (Fig 3A), though here just a trend with UTI89; greater pyuria (Fig 3B); and increased serum levels of G-CSF and the IL-8 analogue KC, relative to the mice that would resolve their infection (Fig 3C and 3D). As well, there were trends toward higher serum IL-5 and IL-6 levels in mice that would develop chronic CFT073 cystitis (Fig 3E and 3F), but, unlike with UTI89-infected mice, the differences were not statistically significant. Thus, while we saw some minor differences between UPEC strains, checkpoint activation predicts the development of chronic cystitis lasting four weeks for both UTI89 and CFT073 in a similar manner.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. An acute host-pathogen checkpoint predicts the outcome of CFT073 bladder infection.

At 24 hpi, urine and serum was collected from mice infected with 10⁸ CFU UTI89 or CFT073, and urine bacterial burden (A), pyuria (B) and serum cytokines (C-F) were assessed. Infection outcomes (chronic cystitis vs. resolution) were determined over the course of four weeks, and data are plotted according to the four week infection outcome. There were no statistically significant differences in the overall data (i.e. all UTI89-infected mice vs. all CFT073-infected mice) at 24 hpi. Data are combined from two to three independent experiments. Data points represent actual values for each individual mouse, zeros are plotted at the limit of detection, and bars indicate median values. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001, Kruskal-Wallis test with Dunn’s multiple test correction. PMN, polymorphonuclear neutrophil.

https://doi.org/10.1371/journal.ppat.1007457.g003

Scanning electron microscopy (SEM) analysis of bladders 28 days after antibiotic therapy showed that mice with a history of chronic CFT073 infection, referred to herein as CFT073^HC mice, had significantly smaller urothelial cells (average cell surface area approximately one eighth the size of Adult Naive cells) (Fig 4A and 4B), similar to the previously observed reduced size of urothelial cells from mice with a history of chronic UTI89 (UTI89^HC) infection [15]. This change in bladder cell morphology is long-lasting in UTI89^HC mice: SEM analysis performed six months after antibiotics showed that UTI89^HC urothelial cells were still about one fourth the size of age-matched naive controls (S2A and S2B Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Chronic CFT073 cystitis causes bladder remodeling phenotypes that persist after antibiotic therapy.

Mice were initially mock-infected (“Adult Naive”), or infected with 10⁸ CFU CFT073, and followed for four weeks, at which time ten days of antibiotics were initiated. Four weeks after the start of antibiotic therapy, convalescent bladders were harvested and assessed by microscopy. (A and B) Scanning electron microscopy (SEM) was used to assess bladder epithelial (urothelial) superficial cell size. (A) Superficial cells from N = 2 replicates with a total of n = 4 mice per group were measured in ImageJ and the average surface area (pixels²) was calculated. Data points represent the average of all measurements for a given mouse. * P < 0.05, Mann-Whitney U test. (B) Representative images from the analysis in (A). Scale bars, 25 μm. (C and D) Cell morphology and differentiation was assessed via immunofluorescence of paraffin-embedded bladder sections from N = 2 staining experiments with bladder sections from n = 3 Adult Naive mice and N = 5 CFT073^HC mice; representative images are shown. In (C) uroplakin IIIa is in green, Trp63 in red, keratin 5 in white and nuclei in blue. Scale bars, 50 μm. In (D), keratin 20 is shown in white, E-cadherin in green, Trp63 in red and nuclei in blue. Scale bars, 100 μm.

https://doi.org/10.1371/journal.ppat.1007457.g004

Immunofluorescence microscopy of bladder sections revealed that while CFT073^HC mice expressed the terminal differentiation marker uroplakin IIIa (Fig 4C), expression of the terminal differentiation marker keratin 20 was weak and patchy, in line with previous observations for UTI89^HC mice [15] (Fig 4D). Weak and patchy keratin 20 staining was also found six months after initial infection in UTI89^HC mice (S2C Fig), suggesting that these infection-induced changes to the host may be permanent. Finally, histological examination of bladder sections from CFT073^HC mice 28 days after antibiotic therapy revealed the presence of lymphoid follicles (S3 Fig), which were previously found to develop between two and four weeks post infection in UTI89^HC mice [14]. Thus, chronic CFT073 infection results in a constellation of bladder remodeling phenotypes similar to those seen after chronic UTI89 infection, suggesting that bladder remodeling is a broad consequence of chronic E. coli infection.

The ability of UPEC to cause recurrent UTI is bacterial strain-dependent

C3H/HeN mice become resistant to UPEC UTI with age [14,15,25], but UTI89^HC mice have enhanced susceptibility to UTI relative to age-matched (so-called “Adult”) Naive mice, a phenomenon we have termed “sensitization” to rUTI [14] (Fig 1G). To assess whether CFT073^HC mice were also sensitized to rUTI, mice with an initial chronic or mock infection were given antibiotics at four wpi and then challenged four weeks later with 10⁷ CFU of an isogenic strain with a different antibiotic resistance marker. Acute and chronic rUTI outcomes were then assessed. Previously, we found that UTI89^HC mice displayed a colonization resistance phenotype within the first 12 hours after challenge infection: UTI89 was less able to invade and replicate within remodeled bladder epithelial cells relative to naive ones, possibly due to their smaller cell size, resulting in significantly lower CFUs in UTI89^HC mice relative to age-matched Adult Naive mice [15]. We found that CFT073^HC mice also displayed a colonization resistance phenotype, with significantly reduced CFUs six hours after CFT073 infection in both UTI89^HC and CFT073^HC mice compared to age-matched Adult Naive mice (Fig 5A). Kidney titers were not statistically significantly different at this time point (Fig 5A). Despite early colonization resistance, UTI89 is nevertheless able to cause severe UTI at 24 hpi in about half of UTI89^HC mice, as indicated by high urine titers (median titer ~10⁶ CFU/ml) (Fig 5B) that precede the development of recurrent chronic cystitis over the 28 day challenge (Fig 5C). In contrast, CFT073 is not able to cause severe UTI in CFT073^HC mice, with median urine titer ~10⁴ CFU/ml urine at 24 hpi (Fig 5B) and zero incidence of chronic cystitis during the 28 day challenge (Fig 5C). Notably, CFT073^HC mice were only protected from same-strain challenge: UTI89 titers were high in the urine at 24 hpi (median titer ~10⁶ CFU/ml) (Fig 5B) and caused recurrent chronic cystitis in 50% of CFT073^HC mice (Fig 5C and S4 Fig). Likewise, UTI89^HC mice were susceptible to recurrent CFT073 cystitis (Fig 5B and 5C and S4 Fig), in line with previous observations [15], and remained susceptible to CFT073 for at least six months after antibiotic therapy (Fig 5D). Thus, whereas a history of UTI89 chronic cystitis predisposes to rUTI caused by either UTI89 or CFT073, a history of chronic CFT073 cystitis predisposes to rUTI caused by UTI89, but not caused by CFT073.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. CFT073^HC mice are protected against CFT073 rUTI, but not against UTI89 rUTI.

Mice were initially infected with 10⁸ CFU UTI89 or CFT073 and monitored over four weeks, at which time antibiotic therapy was initiated. Those mice that developed chronic cystitis during the initial infection were then challenged four weeks after antibiotics (A-C) or six months after antibiotics (D) with 10⁷ CFU UTI89 or CFT073 as indicated. Initially mock-infected mice were included as a control (“Adult Naive”). (A) Mice were sacrificed six hours after infection with CFT073 and bladder and kidney titers were assessed. * P < 0.05, Kruskal-Wallis test with Dunn’s multiple test correction. (B and C). Mice were challenged with UTI89 or CFT073 and followed over four weeks. (B) Urine titers at 24 hours post challenge, representing acute cystitis. (C) Incidence of chronic cystitis over the four week challenge. * P < 0.05, Fisher’s exact test. (D) Incidence of chronic cystitis over the four week challenge, six months after the initiation of antibiotics, compared to age-matched (“Naive”) mice and mice that had resolved their initial infection (“UTI89^HR”). * P < 0.05, Fisher’s exact test. # denotes the number of mice per group. Data are from two to three independent experiments. Data points represent actual values for each individual mouse, zeros are plotted at the limit of detection, and bars indicate median values.

https://doi.org/10.1371/journal.ppat.1007457.g005

T cell depletion abrogates protection from rUTI in a bacterial strain-dependent manner

The observations that chronic CFT073 infection resolves over time (Fig 2) and confers protection against CFT073 challenge (Fig 5) suggest that chronic CFT073 cystitis may elicit an adaptive immune response that clears the infection and protects against challenge. Thus we questioned whether components of the adaptive immune system played a role in this model. We assessed bacterial-specific IgG levels in the serum and found that both CFT073 and UTI89 elicited antibody responses, though anti-CFT073 IgG levels were somewhat higher than anti-UTI89 IgG levels (S5 Fig). We also tested whether T cell depletion would affect chronic and/or recurrent cystitis outcomes. For other chronic Gram-negative bacterial infections, depletion of both CD4⁺ and CD8⁺ T cells resulted in significantly worse infection phenotypes than single depletions did [26–29]. Hence, we tested the effects of both single subset depletions and combined depletions, compared to an equivalent dose of isotype (0.5 or 1 mg, respectively), during an initial UPEC infection. Depletion of lymphoid subsets and effects on myeloid subsets was determined four weeks after the initial infection, and at the time of the challenge infection, by flow cytometry of spleen single cell suspensions (S6 Fig).

Depletion of CD4⁺ and CD8⁺ T cell subsets did not affect the incidence of chronic cystitis caused by UTI89 (Fig 6A). For CFT073 infection, the incidence of chronic cystitis at 28 dpi was higher in the single depletion groups (67% for CD4-depleted and 62% for CD8-depleted, compared to 46% in isotype-treated mice, S7A and S7B Fig), but the difference was not statistically significant, and similar proportions of mice cleared the infection after seven days (S7C Fig). Furthermore, mice depleted of single T cell subsets were still protected from same-strain rUTI (S7D and S7E Fig). In the combined CD4⁺ and CD8⁺ T cell depleted group, the rate of chronic CFT073 cystitis was 61%, vs. 40% in the isotype-treated group (Fig 6A), and CFT073 urine titers at 28 dpi were significantly elevated in the combined T cell-depleted mice relative to isotype-treated controls (Fig 6B). Notably, significantly fewer mice depleted of both CD4⁺ and CD8⁺ T cell subsets resolved the CFT073 infection after 7 dpi, relative to isotype-treated mice (Fig 6C). Accordingly, combined CD4⁺ and CD8⁺ T cell depletion also prevented the so-called “late resolver” phenotype (wherein some mice resolve their CFT073 infection between days 21 and 28 post infection; see Fig 2G). In the T cell-depleted group, n = 12 mice resolved the CFT073 infection: nine resolved within the first week and none between days 21 and 28. In contrast, of the n = 21 isotype-treated mice that resolved the CFT073 infection, ten resolved between days 21 and 28 (Fig 6D and 6E).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Depletion of CD4⁺ and CD8⁺ T cell subsets increases chronic and recurrent CFT073 cystitis.

Mice infected with 10⁸ CFU UTI89 or CFT073 were given weekly doses of anti-CD4 and anti-CD8 antibodies, or isotype control, and followed for four weeks. (A) The incidence of chronic cystitis was determined at 28 dpi; # indicates the number of mice per group. (B) Urine bacterial burdens at 28 dpi. (C) CFT073-infected mice from (A) were binned according to the length of their infection. ** P < 0.01, Fisher’s exact test for resolving 8–28 dpi. (D) CFT073 urine titers over the course of infection from isotype-treated (left panel) and anti-CD4/8-treated (right panel) mice. Dashed lines indicate our previously established cutoff for high-titer persistent bacteriuria, 10⁴ CFU/ml urine. (E) Shown are the percentages of mice with persistent high-titer CFT073 bacteriuria (>10⁴ CFU/ml urine) over time. (F-J) At 28 dpi, antibiotic therapy was initiated. Those mice that developed chronic cystitis during the initial infection were then challenged four weeks after antibiotics with 10⁷ CFU of the same strain as their initial infection. Depletion antibodies were not administered during the challenge infection. (F) The incidence of recurrent chronic cystitis was determined at four weeks post challenge; # indicates the number of mice per group. (G) CFT073 urine titers over the course of the challenge infection in CFT073^HC mice, from mice initially isotype-treated (left panel) and initially anti-CD4/8-treated (right panel). Some samples were unavailable at three days post challenge. (H) Shown are the percentages of CFT073^HC mice with persistent high-titer recurrent CFT073 bacteriuria (>10⁴ CFU/ml urine) during the challenge infection. (I and J) Shown are bladder and kidney bacterial burdens (I) and bladder edema (J) at time of sacrifice. Data are combined from two to four independent experiments; data points represent actual values for each individual mouse, zeros are plotted at the limit of detection, and bars indicate median values. *, P < 0.05, Mann-Whitney U test. rUTI, recurrent urinary tract infection.

https://doi.org/10.1371/journal.ppat.1007457.g006

Based on the above results, we tested whether the abrogation of CD4⁺ and CD8⁺ T cell responses during the initial CFT073 infection would render mice susceptible to recurrent CFT073 cystitis. Four weeks after the initial infection and concomitant T cell depletion, mice were given antibiotics and then challenged four weeks later with the same strain as their initial infection, without additional T cell depletions. UTI89^HC mice were susceptible to developing recurrent UTI89 cystitis, whether they were initially treated with anti-CD4/8 or isotype (Fig 6F). In contrast, T cell depletion had a striking effect on CFT073 challenge of CFT073^HC mice. No mice (out of n = 8) in the isotype-treated group developed recurrent chronic CFT073 cystitis, consistent with our observations in untreated mice (see Fig 5C); one mouse had persistent bacteriuria but had a bladder titer <10⁴ CFU. However, four of 11 mice that received anti-CD4/8 during their initial chronic infection developed recurrent chronic CFT073 cystitis lasting four weeks (Fig 6F–6H) and one mouse was a “late resolver,” similar to what is seen during CFT073 infection of juvenile naive mice (see Figs 2G and 6D). Bladder and kidney bacterial burdens did not differ between isotype vs. anti-CD4/8 treatments in UTI89^HC mice after UTI89 challenge, but in CFT073^HC mice challenged with CFT073, titers at sacrifice were higher in those that received anti-CD4/8 than those that received isotype control (Fig 6I). As well, bladder edema four weeks post challenge with CFT073 was significantly greater in CFT073^HC mice that received anti-CD4/8 than in those that received isotype control (Fig 6J). In sum, these experiments show that depletion of both CD4⁺ and CD8⁺ T cell subsets during an initial CFT073 infection renders C3H/HeN mice susceptible to chronic and same-strain recurrent cystitis, whereas T cell depletion during a UTI89 infection does not impact chronic or same-strain recurrent cystitis.

Ethics statement

All mouse experiments were conducted according to the National Institutes of Health guidelines for the housing and care of laboratory animals. All experiments were performed in accordance with institutional regulations after review and approval by the Institutional Animal Care and Use Committee (IACUC) at Washington University School of Medicine in St. Louis, MO (animal protocol number 20150226).

Bacterial strains

The uropathogenic E. coli isolates used in this study were the human cystitis isolate UTI89 [18] and the antibiotic-resistant derivatives UTI89 attHK022::COMGFP (kanamycin-resistant) and UTI89 attl::PSSH10-1 (spectinomycin-resistant) [47]; the human urosepsis isolate CFT073, wild-type and ΔrpoS [19], and the antibiotic-resistant derivatives CFT073ΔrpoS HK::Cm (chloramphenicol-resistant), CFT073ΔrpoS HK::Kan (kanamycin-resistant), and CFT073ΔrpoS ΔfimH::FimH^UTI89 (kanamycin-resistant) [21]. Strains were cultured statically in lysogeny broth (LB) at 37°C for two to three overnight passages to induce type 1 pilus expression. In experiments where mice had an initial infection and subsequent challenge infection, the challenge strain was marked with different antibiotic resistance from the initial strain.

Mouse infections

C3H/HeN mice were obtained from Envigo Research Model Services, Inc. (formerly Harlan). All mice were female and 7–8 weeks old (“Juvenile”) at the time of the first infection (or mock-infection in the case of the Adult Naive group). Bacterial inocula were prepared as previously described [14] and 10⁷ or 10⁸ CFU of bacteria were inoculated into the bladders of C3H/HeN mice by transurethral catheterization under isoflurane anesthesia as previously described [48]. For infections in Juvenile Naive mice, 10⁸ CFU was always used, since C3H/HeN mice develop chronic cystitis in an infectious dose-dependent manner. For challenge infections, 10⁷ CFU (of a differently antibiotic-marked strain than the initial infection) was used, to facilitate comparisons with prior studies [14,15,17]. To monitor infection outcomes, urine, bladder and kidney bacterial burdens were determined as previously described [14]. Mice were humanely euthanized by isoflurane inhalation followed by cervical dislocation. Chronic cystitis was defined as persistent high-titer bacteriuria (>10⁴ CFU/ml urine) at every urine collection time point (1, 3, 7, 10, 14, 21, and 28 days post infection, and weekly thereafter if experiments continued). A cutoff of 10⁴ CFU/ml persistent bacteriuria is highly specific and sensitive for the detection of chronic cystitis [14]. In addition, chronic cystitis was also indicated by a high bladder bacterial burden (>10⁴ CFU/bladder) and a visibly enlarged, inflamed bladder (indicating inflammation and edema) at time of sacrifice (≥28dpi). Resolution of cystitis was defined as urine bacterial titer dropping below 10⁴ CFU/ml urine during at least one time point, and/or bladder bacterial burden <10⁴ CFU at time of sacrifice. For long-term “challenge” experiments, beginning four weeks after the initial infection, all mice (including mock-infected controls, and regardless of infection outcome) were treated with trimethoprim and sulfamethoxazole in the drinking water (54 and 270 μg/ml water, respectively) for ten days. Urine samples were collected weekly after the initiation of antibiotics to confirm sterile urines and any mice with treatment failure were excluded from subsequent analysis. Four weeks or six months after the initiation of antibiotics to clear initial infection, mice were challenged with 10⁷ CFU of bacteria. The same criteria were used to define chronic cystitis during the challenge infection as in the initial infection (persistent bacteriuria >10⁴ CFU/ml urine; bladder burden >10⁴ CFU; enlarged bladder), with the exception that urine titers could be <10⁴ CFU/ml on the first day post challenge, as we have found that checkpoint activation in mice with a history of chronic cystitis may take longer (up to 72 hours) than in Juvenile Naive mice (24 hours).

Histopathology and immunofluorescence

Bladders were aseptically harvested and fixed overnight in methacarn (60% methanol, 30% chloroform, 10% glacial acetic acid), bisected to give two halves per bladder, paraffin-embedded and sectioned. For histopathological examination, slides were stained with hematoxylin and eosin and imaged with a Zeiss Axio Scan Z.1 brightfield slide scanner. Immunofluorescence experiments were conducted as previously described [15]. Primary antibodies used were uroplakin IIIa (mouse monoclonal, 10R-U103a, Fitzgerald), Trp63 (rabbit polyclonal, GTX102425, GeneTex), E-cadherin (goat polyclonal IgG, AF748, R&D Systems), cytokeratin 5 (chicken polyclonal, 905901, BioLegend) and cytokeratin 20 (mouse monoclonal, M7019, DAKO). Samples were mounted in ProLong Gold Antifade Mountant with DAPI (ThermoFisher Scientific) and fluorescence was visualized on a ZEISS Axioskop Observer.Z1 microscope.

Scanning electron microscopy

Two microscopy sample preparation protocols were used during the course of this work. For urothelial cell size determination six months after antibiotics, bladders were aseptically harvested, bisected, splayed in dissecting trays, and fixed in EM fixative as previously described [15]. The fixative comprised 2% paraformaldehyde and 2% glutaraldehyde in 0.1M sodium phosphate buffer, pH 7.4. Samples were prepared by critical point drying. Briefly, samples were post-fixed in 1% osmium tetroxide, dehydrated in increasing concentrations of ethanol, then dehydrated at 31.1°C and 1072 PSI for 16 minutes in a critical point dryer. Samples were mounted on carbon tape-coated stubs and sputter-coated with gold/palladium under argon. After these experiments, subsequent SEM samples were processed with a different protocol, which we verified gave the same results as prior studies. Thus, for all other experiments, the EM fix comprised 2% paraformaldehyde, 2.5% glutaraldehyde in 0.15M sodium cacodylate buffer with 2 mM CaCl₂ at pH 7.4, and after post-fixing in 1% osmium tetroxide and critical point drying, samples were sputter-coated with iridium. For urothelial cell size determination after UTI89 vs. CFT073 infection, bladders were aseptically harvested, bisected, splayed in dissecting trays, and fixed in EM fixative before further processing. For the bladders obtained during chronic UTI89 or CFT073 cystitis at 28 dpi, the “balloon method” was used, wherein bladders were instilled with fixative prior to harvest, as previously described [49]. All bladders were imaged on a Zeiss Crossbeam 540 FIB-SEM or a Zeiss MERLIN FE-SEM. For urothelial cell size determinations, ImageJ 1.47v (National Institutes of Health, USA) was used to calculate epithelial cell surface area in five 500x magnification fields per bladder half (ten per mouse).

Assessment of acute inflammation and checkpoint activation

Bacteria were enumerated in urine or bladder samples by serial dilution as previously described [14]. Urine sedimentation to determine neutrophil influx was performed as previously described [14]. Briefly, 80 μl of a 1:10 dilution of urine was centrifuged onto poly-L-lysine-coated glass slides and stained with a Hema 3 kit (Fisher Scientific). Slides were examined by light microscopy and the average number of polymorphonuclear leukocytes (PMN) per high-powered field (hpf; 400x magnification) was calculated from counting five fields. A semi-quantitative scoring system was used: 0, less than 1 PMN/hpf; 1, 1–5 PMN/hpf; 2, 6–10 PMN/hpf; 3, 11–20 PMN/ hpf; and 4, >20 PMN/hpf. Bladder edema was assessed by pre-weighing empty bladder homogenization tubes, weighing the bladder-containing tubes, and calculating the difference. Serum cytokines were assessed using the Luminex-based multiplex cytometric bead array platform (Bioplex, Bio-Rad, Hercules, CA) using beads individually purchased for IL-5, IL-6, KC, and G-CSF. The assay was performed according to the manufacturer’s instructions, except using 10-fold less standard and half the amount of coupled beads and detection antibodies indicated in the protocol, as described in [50]. Individual samples were run in duplicate and the mean value was used for plotting and statistical analysis.

Serum ELISAs

Blood was obtained from convalescent mice (four weeks after the initiation of antibiotics) via submandibular puncture using a 5 mm lancet (Goldenrod) and collected in a BD Microtainer serum separator tube (BD Biosciences). UTI89 and CFT073 were grown as described above and 1 ml of OD₆₀₀ = 1 bacteria was lysed by boiling at 100°C for 5 minutes. Lysed bacteria were diluted tenfold in PBS and applied to high-affinity Immulon 2 HB 96 well plates (Thermo Fisher) overnight at 4°C. Plates were washed with PBS containing 0.05% Tween 20 and a dilution series of serum was applied for 90 minutes at room temperature. After washing, a 1:10,000 dilution of HRP-conjugated goat anti-mouse IgG (Invitrogen) was applied for 90 minutes at room temperature. After a final wash step, the TMB substrate reagent set (BD Biosciences) was used to develop the reaction. Absorbance was measured on a VersaMax microplate reader (Molecular Devices).

T cell depletions

Depletion antibodies were obtained from BioXCell: α-mouse CD4 (GK1.5, BP0003-1), α-mouse CD8a (2.43, BP0061), and rat IgG2b isotype control (LTF-2, BP0090). Mice were given intraperitoneal injections one day prior to infection and at days 7, 14, and 21 post infection. For depletions of both CD4⁺ and CD8⁺ T cell subsets, mice were treated with 500 μg α-CD4 and 500 μg α-CD8, or 1 mg isotype control, in a 500 μl volume of pH 7.0 dilution buffer (BioXCell). For single subset depletions, mice received 500 μg α-CD4 or 500 μg α-CD8, or 500 μg isotype control, in a 250 μl volume of pH 7.0 dilution buffer (BioXCell). T cell depletions were performed only during the initial infection, not the challenge infection.

Flow cytometry

Spleens were dissected and placed in RPMI (Gibco) on ice until spleens were removed from all animals. Each sample was dissociated by placing the spleen between two frosted glass slides and rubbing the slides together. The spleens were placed in ACK buffer (Gibco) for 5 minutes at room temperature to lyse red blood cells. Ten mls of RPMI was added to each tube and tubes were inverted to mix, then centrifuged at 1300 rpm at 4°C for 5 minutes. The supernatant was discarded and the pellet was resuspended in 3 mls RPMI and pushed through a 70 μm cell strainer (Fisher). The cells were washed twice in 1X dPBS and centrifuged at 500 X g. The pelleted cells were resuspended in Fc block (BD Biosciences) and incubated in 1 μl of live/dead blue fluorescent reactive dye (Thermo Fisher/Invitrogen, L34962) per 10⁶ cells for 20 minutes at 4°C, then washed in 1X dPBS. Cells were stained with surface markers, all from Biolegend, in two separate panels: Panel 1: CD45 (30-F11), CD8a (53–6.7), CD3 (17A2), CD19 (GD5), and CD4 (GK1.5); Panel 2: CD45 (30-F11), CD11b (M1/70), Ly6G (1A8), F4/80 (BM8), and Ly6C (HK1.4). Sample data was acquired on the LSR II flow cytometer (BD Biosciences) recording 10,000 events. The data was analyzed using FlowJo software (version 10.0). The relative proportion of cellular infiltrates was expressed as a percentage of live cells.

Statistical analysis

Our experience with the C3H/HeN chronic and recurrent cystitis model shows that when assessing bimodal responses, five mice per replicate (two replicates minimum) is the minimum number necessary to overcome any biological and/or technical variability and provide reliable and interpretable results within an experiment; a statistical analysis was not used to determine the number of mice per experiment. Investigators were not blinded to the infection history of the animals, except for Fig 2E. Statistics were performed in GraphPad Prism v7.01. A two-tailed Fisher’s exact test was used to test for significant differences in the incidence of chronic cystitis and kinetics of infection resolution. For bacterial titers, pyuria, cytokine expression levels, and surface area differences in superficial cells, comparisons of three or more groups were performed with the Kruskal-Wallis test followed by Dunn’s multiple test correction. Pairwise comparisons were performed with a two-tailed Mann-Whitney U test. P < 0.05 was considered statistically significant.