Using a C3H murine model of UTI, we have unveiled an inflammation-dependent checkpoint occurring at the host-pathogen interface within the first 24 hours of infection that has significant ramifications upon the long term fate of disease. Furthermore, subsequent development of chronic cystitis alters the host susceptibility to further bacterial cystitis after antibiotic therapy, resulting in increased susceptibilities to both severe acute symptomatology and chronic infections. We identified elevated IL-5, IL-6, G-CSF, and KC as a serum biomarker signature prognostic of the development of chronic cystitis in C3H mice. This signature was accompanied by severe bladder inflammation at 24 hpi as indicated by marked bladder edema and pyuria. While the severity of bladder damage and systemic signs diminish after the early acute stage of infection, these mice remain persistently bacteriuric and develop chronic cystitis. This improvement in clinical signs despite persistent bacteriuria is similar to the natural course of UTI reported in a majority of women from placebo-controlled studies 
. The alternate host response was characterized by mild to moderate bladder inflammation in the acute stage with minimal or no systemic signs. These mice typically achieved resolution of acute bacteriuria within 1–2 weeks, though recurrent bacteriuria after this time was common, either because of transient periurethral colonization or rUTI. The fate of disease outcome appears to be determined within the first 24 hpi and is infectious dose-dependent, indicating that this response is at least partially stochastic. Understanding the molecular and cellular mechanisms of this acute checkpoint may have far-reaching implications for how UTI and other chronic and recurrent infections are treated and evaluated.
Our findings indicate that a glucocorticoid-sensitive, pro-inflammatory response to early acute infection is required for the efficient development of chronic cystitis. Glucocorticoids are made in the adrenal gland and, among other functions, act as potent anti-inflammatory agents. Specifically, they block the production of acute mediators of inflammation such as TNFα, IL-1, IL-6, and arachidonic acid derivatives such as prostaglandins and leukotrienes, in part through blockade of NF-κB activity 
. For example, dexamethasone is commonly used as an adjunct therapy to treat acute bacterial meningitis, as its anti-inflammatory activity appears to improve survival 
. The immunosuppressive effects of a single dexamethasone treatment are transient, providing evidence for a mucosal immune checkpoint early in UPEC infection that, when triggered, leads to urinary tract immunopathology and facilitates chronic cystitis.
The resistance of C3Hscid
mice to chronic cystitis implicates lymphocytes as immune cells necessary for the development of chronic cystitis in C3H mice. Specifically, our data suggest that one or more lymphocyte populations with innate properties may play a necessary role in mediating a number of the early severe inflammatory responses to UPEC that are associated with the development of chronic cystitis. Candidate lymphocytes include γδ T cells, natural killer T (NKT) cells, and B1 cells, which make natural antibodies. γδ T cells have previously been implicated in host resistance to UTI in C57BL/6J mouse infection models 
, and they are normally found in the bladder lamina propria of naïve mice 
. Acute host responses that were muted or absent in C3Hscid
mice included severe pyuria, weight loss, and elevation of serum IL-6, G-CSF, and KC. Induction of both IL-6, an acute phase pro-inflammatory cytokine that causes fever and malaise if systemic, and IL-8, a chemotactic cytokine for granulocytes, are strongly associated with UTI in both humans and mice 
. The combination of elevated IL-6, KC, an IL-8 analog, and G-CSF, a growth factor that promotes granulocyte development and release from the bone marrow, in both the serum and urine would explain the weight loss and severe acute PMN response associated with the development of chronic cystitis in mice 
. Yet, these acute molecular and cellular responses are inadequate for resolving infection in C3H mice. Furthermore, a recent study suggests that such a response may contribute to disease pathogenesis, as depletion of G-CSF surprisingly resulted in lower UPEC titers in the bladders of C57Bl/6J mice at 48 hpi, despite reduced neutrophil infiltration into the bladder 
. Similarly, treatment with G-CSF exacerbates disease in a murine model of Klebsiella pneumoniae
respiratory tract infection 
. Further studies are needed to identify the immune cell populations required for the development of chronic cystitis, and to clarify the role of these early innate responses in disease pathogenesis.
We have found a strong association between elevated serum IL-5 at 24 hpi and the development of chronic cystitis. IL-5 promotes eosinophil and B1 lymphocyte development and IgA class switching and has been reported to enhance production of IL-6 in kidney cell lines 
. However, the role of IL-5 in UTI pathogenesis is unclear. While “low level” IL-5 mRNA expression has been reported in the mouse bladder early in UPEC infection of C3H/HeN mice, IL-5 has not otherwise been implicated in UTI pathogenesis 
. T helper cells polarized to produce type 2 cytokines (TH
2 cells) have long been considered the primary source of IL-5, but the acute serum IL-5 response to UPEC infection in this study was still present in C3Hscid
mice. This finding is consistent with a previous study in mice, which found that non-lymphoid cells, such as mast cells or basophils, are the major source of IL-5 in the peripheral organs 
. Activation of either of these cell types could mediate the bladder edema seen in C3Hscid
mice. However, if IL-5 is originating from the bladder, it does not appear to be able to access the bladder lumen as IL-5 was not elevated in the urines of mice at 24 hpi, regardless of disease outcome.
C3H background mice are known to have differing susceptibilities to various bacterial infections 
. Currently, the specific genetic basis for susceptibility to chronic cystitis and chronic pyelonephritis are unknown. However, a recent mouse genetics study by Hopkins and colleagues has made significant progress in identifying genetic loci associated with susceptibility to cystitis and pyelonephritis at 10 dpi in second generation crosses of C3H/HeJ (reported as susceptible to each) and Balb/c (resistant) mice 
. In those studies, the authors utilized quantitative trait loci (QTL) analyses to discover that the genetic basis for chronic cystitis and pyelonephritis differed. This is consistent with our data comparing experimental UPEC infection in the various C3H sub-strains through 4 wpi, which demonstrate that chronic cystitis and chronic pyelonephritis can be independent disease outcomes. In C3H/HeN, C3H/HeSnJ, and C3HeB/FeJ mice with chronic cystitis, concurrent kidney infection was an inconsistent finding and, when present, was largely limited to the renal pelvis (pyelitis). Equally, the majority of C3H/HeJ mice with chronic pyelonephritis did not have cystitis at sacrifice, as defined by the presence of bladder inflammation and bacterial titers greater than 104
cfu. Only in C3H/HeOuJ mice was there a strong association between chronic cystitis and pyelonephritis.
Since previous work had demonstrated that UPEC infection of both C3H/HeJ and C3H/HeOuJ mice resulted in similarly high bladder titers at 14 dpi, it had been hypothesized that chronic cystitis occurred independently of TLR4 signaling 
. However, our studies have now demonstrated that these two strains differ in their susceptibility to chronic cystitis. We suggest that differences in experimental methodology, including the use of ten-fold higher inocula and study durations limited to 2 weeks, likely prevented resolution of the differing susceptibilities of C3H/HeOuJ and C3H/HeJ mice to chronic cystitis in previous studies 
. In support of the hypothesis that enhanced TLR4 signaling contributes to the development of chronic cystitis, which in turn enhances host susceptibility to recurrent symptomatic infection, a TLR4 polymorphism that results in diminished responses to LPS has recently been associated with reduced susceptibility to rUTI in humans 
. We also found that a subset of C3H/HeJ mice developed a persistent biofilm-like bladder colonization in the absence of robust inflammation. Bacterial biofilms and reduced TLR4 expression on neutrophils have each been associated with asymptomatic bacteriuria in humans 
. Thus, the innate host responses in the urinary tract must be fine-tuned to successfully eliminate UPEC infection while also maintaining tissue integrity.
Intracellular bacterial communities (IBC) had not been previously observed during the chronic stage of UTI and thus their role in bacterial persistence was unknown. However, IBC formation is known to be required during the acute stages of infection 
which is a prerequisite for subsequent chronic infection. For example, we recently identified residues in FimH, the mannose-binding tip adhesin subunit of type 1 pili, that are under positive selection in UPEC strains isolated from human patients with UTI and that function in IBC formation 
. A double mutation in two of these residues abolished IBC formation despite retaining the ability to bind to mannose, bind to the urothelium, and invade the urothelium. This mutant was severely defective in a mouse model of UTI, and behaved similarly to a fimH
knockout or a FimH receptor binding mutant. Therefore, this mutant separated the phenotypes of (i) mannose-inhibitable binding and invasion of the urothelium and (ii) IBC formation 
. Its phenotype in a mouse model of UTI indicates that both of these phenotypes are critical for acute stages of UTI, which we have now shown are a prerequisite for persistence. However, the lack of development of chronic cystitis in C3Hscid
mice and in C3H mice treated with dexamethasone was not due to the lack of IBC formation in the acute stages of infection since IBCs developed normally in the acute stages of infection in these mice. Furthermore, we found evidence of IBC formation in the chronic stage of infection in C3H/HeJ mice, which have muted inflammatory responses to UPEC infection. This finding implicates mucosal inflammation, including urothelial reactivity and exfoliation, in restricting continuous IBC formation in immunocompetent mice. It also raises the hypothesis that one mechanism of asymptomatic bacteriuria, for which UPEC infection of C3H/HeJ mice has been proposed as a model 
, may be continuous IBC formation in hosts with muted urothelial inflammatory responses to UPEC.
In summary, we have discovered a new basis for understanding UTI that provides a possible mechanism for both chronic and recurrent infection. We propose that, in females that are genetically predisposed to enhanced mucosal TLR4 signaling, initial episodes of UTI due to UPEC or Klebsiella pneumoniae
may be particularly severe in both degree of symptomatology and duration. If allowed to progress past the early acute stage before initiation of antibiotic therapy, these individuals could then develop altered bladder mucosal responses to gram-negative uropathogens. Upon repeated exposure to gram-negative uropathogens, these individuals would then be at increased risk for developing severe, symptomatic rUTI. It is also possible that a quiescent intracellular reservoir (QIR) state may exist in humans, similar to what has been described in mice 
, not only serving as seeds for recurrent UTI, but also potentially modulating the mucosal response to UPEC due to the chronic persistence of bacteria within mucosal cells. This hypothetical model not only provides a logical basis for further, intensive murine and human UTI studies, but also extends our general understanding of chronic and recurrent gram-negative infections of mucosae.