Recurrent infections drive persistent bladder dysfunction and pain via sensory nerve sprouting and mast cell activity

Abstract

Urinary tract infections (UTIs) account for almost 25% of infections in women. Many are recurrent (rUTI), with patients frequently experiencing chronic pelvic pain and urinary frequency despite clearance of bacteriuria after antibiotics. To elucidate the basis for these bacteria-independent bladder symptoms, we examined the bladders of rUTI patients. We noticed a striking increase in neuropeptide content in the lamina propria and indications of enhanced nociceptive activity. In mice subjected to rUTI, we observed sensory nerve sprouting that was associated with nerve growth factor (NGF) produced by recruited monocytes and tissue resident mast cells. Treatment of rUTI mice with an NGF-neutralizing antibody prevented sprouting and alleviated pelvic sensitivity, whereas instillation of native NGF into naïve mice bladders mimicked nerve sprouting and pain behavior. Nerve activation, pain, and urinary frequency were each linked to presence of proximal mast cells, as mast cell deficiency or treatment with antagonists against receptors of several direct or indirect mast cell products were each effective therapeutically. Thus, our findings suggest NGF-driven sensory sprouting in the bladder coupled with chronic mast cell activation represent an underlying mechanism driving bacteria-independent pain and voiding defects experienced by rUTI patients.

One Sentence Summary:

NGF produced by monocytes and mast cells drives sensory nerve growth and sensitization following bacterial bladder infections.

Introduction

Urinary tract infections (UTIs) account for a high number of clinical visits each year due in part to the high susceptibility of women and high rates of recurrence (1–3). Most UTIs are bacterial in nature, with uropathogenic Escherichia coli (UPEC) strains accounting for most infections (4). Recurrence appears related to the persistence of UPEC in reservoirs within the urinary tract as the persisting bacteria are typically the same causal agents of the initial UTI (5). Indeed, studies in mice have revealed that UPEC can persist for months within bladder epithelial cells long after inflammation has subsided (6). Remarkably, rUTIs occur even in individuals without any preexisting conditions and whose immune systems are intact (7). The standard of care for UTIs is typically a 3–5 day course of antibiotic treatment. In patients with frequent recurrences, the off-label prophylactic antibiotic course could be in excess of 6 months (8–10). rUTI patients typically present with symptoms associated with altered bladder sensation such as pelvic pain, increased frequency, dysuria, and urgency, but they are routinely treated with antibiotics even when their urine cultures are negative (11, 12).

Increased nerve sensitivity has been linked to several chronic pain pathologies and is associated with nerve sprouting, or aberrant growth of nerve fibers, due to elevated neurotrophic factors including nerve growth factor (NGF) (13). Notably, local production of NGF has been implicated in the propagation of bladder nerves; furthermore, there are several reports of elevated urinary NGF in both first-time UTI and rUTI patients. NGF not only induces neuronal growth but also sensitizes sensory fibers by lowering the nociceptive activation threshold and increasing the relative expression of receptors(14–16), including transient receptor potential vanilloid (TRPV) receptors, which are suggested to mediate peripheral and visceral pain (17–21). Subsequent activation of these sensory nerves results in the release of neuropeptides such as substance P (SP) and calcitonin gene related peptide (CGRP), both of which are involved in pain transmission (22, 23).

Mast cells are a potential major source of bladder NGF and a pivotal modulator of inflammation during UTIs (24, 25), which are specialized secretory cells resident in the lamina propria (26). Included within their prestored secretory granules are histamine and heparin that aid in bradykinin production (27), as well as other bioactive molecules capable of sensitizing nerves (28, 29). Histamine has been shown to sensitize TRPV1⁺ nerves in irritable bowel syndrome patients, and in the bladder, it is able to sensitize nerves to distension during filling (30, 31). In addition, afferent nerves in the bladder express bradykinin receptors that can alter nerve excitability upon exposure to bradykinin (32). During acute infection, bladder mast cells degranulate, releasing these components and recruiting neutrophils to target uropathogens (33). Shortly afterwards, mast cells traffic across the basement membrane to the superficial epithelium to trigger exfoliation of epithelial cells, which are typically fully laden with infecting bacteria (34). Following exfoliation, mast cells promote quenching of local inflammation through secretion of IL-10, enabling rapid re-epithelization as urine can be toxic to the underlying tissue (26, 34, 35). Interestingly, mast cells in various tissue are found adjacent to peripheral nerves (36) and in view of their capacity to secrete NGF, histamine, and heparin, these cells have the potential to contribute to prolonged bladder sensitization during UTIs. Here, we elucidate the underlying basis for chronic pelvic pain and associated voiding aberrations following rUTIs.

Results

Pelvic sensitivity and urinary frequency after rUTI are associated with sensory sprouting

Many rUTI patients experience continuation of symptoms including pain (pelvic sensitivity) and void dysfunction even after the infection has subsided (37–40). Since both pain and micturition signals are transmitted via the peripheral nervous system, we focused our initial investigations on bladder sensory nerves. Bladder biopsies were collected from controls and rUTI patients currently experiencing pain with no culturable bacteria in their urine. Biopsies were analyzed for SP⁺, a neuropeptide and marker for nociceptive sensory nerves (41), and we found that rUTI patients displayed significantly increased SP staining (p<0.0001) in the lamina propria compared to control patients (Figure 1A–C and Figure S1A). To assay for the activation of the sensory nerves, which usually results in the release of neuropeptides, urine SP levels were assessed and found to also be elevated in rUTI patients (Figure 1D). These findings strongly suggest there are substantial and persistent perturbations to nociceptive activity in rUTI patients.

Figure 1: rUTI Induces Sensory Sprouting in the Bladder Evoking Pelvic Pain-like Behavior and Void Defects.

(A-B) Bladder biopsies were obtained from (A) control and (B) rUTI patients and immunostained for E-cadherin to identify intermediate epithelium and substance P (SP) to identify the neuropeptide. Scale bar, 50 μm (C) Quantification of SP immunostain from patient bladder biopsies, with each data point representing a single field of view (n=3 control and n=8 rUTI patients). (D) Urine from control and rUTI patients was collected and assayed for SP level (n=12 control and n=25 rUTI patients). (E) Schematic of rUTI model. Mice were infected with 10⁸ CFU UPEC once a week for three consecutive weeks. Functional assays were performed 14 days after the third infection. (F) rUTI or control (saline instilled) mice were mechanically probed to assess response frequency to pelvic stimulation(n=13–17 mice). (G) Frequency was assessed via cystometry analysis of control and rUTI mice (n=7–9 mice). Data is representative of 2–3 independent experiments. (H-K) Representative images (H,J) and quantification (I,K) of fluorescence staining in the bladder show increased presence of nerve fibers (Green, TUJ1⁺) in the lamina propria but not in the detrusor in mice following rUTI. Each plot point represents one image field. Data is representative of 2–3 independent experiments, with each containing at least two mice per treatment group. Scale bar, 100 μm. (L-N) (L) Representative (left) immunofluorescence and (right) 3D models and (M,N) quantification of SP⁺ innervation in control and rUTI lamina propria. Both neurite length and number of branch points increase in rUTI mice compared to the controls. Z-stacks were acquired per mouse bladder from at least two independent experiments, with at least two mice per treatment group. Scale bar, 30 μm. All comparisons were analyzed by unpaired Mann-Whitney U Test with p<0.05 used to define significance. Mean±SEM, ****p<0.0001, **p<0.01, *p<0.05, ns=not significant.

We next recapitulated these clinical findings in a murine model for mechanistic studies. Recently, we described an rUTI infection regimen in which 6–8 week old female C57BL/6 (WT) mice were subjected to three consecutive, naturally resolved (no bacteriuria) infections by uropathogenic Escherichia coli (UPEC) strain J96 (42) or sham challenge (control) 7 days apart (Figure 1E) (26). 14 days after the last challenge, mice were assessed for pelvic sensitivity and urinary frequency. Pelvic sensitivity was determined based on responses evoked by mechanical stimulation of the abdomen (43–45), and urinary frequency was determined based on cystometric analysis of bladder urodynamics (46). We found that compared to control mice, rUTI mice displayed increased pelvic sensitivity and urinary frequency (Figure 1F,G and S1B–D), confirming that these mice display similar functional features as clinical patients.

We assessed bladder innervation in our rUTI mice to further compare to clinical observations. When bladders were harvested from mice subjected to this experimental regimen, sectioned, and subjected to immune microscopy, we noticed a striking increase in the density of nerve fibers using the pan-neuronal marker TUJ1 (Figure 1H–K and S1E–G). This overt increase was specific to the lamina propria (Figure 1H,I), as we did not see a noticeable difference when we analyzed the detrusor (Figure 1J,K). We determined whether there was an increase in innervation of sensory nerves specifically. Employing probes for SP, we were able to confirm that rUTI mice display sensory nerve sprouting in the lamina propria based on increased immunofluorescence signal observed in 3D models used for quantification, verifying our previous findings (Figure 1L–N, Figure S1H, and Supplemental Movie 1,2; movie 1 and 2 show an overview of a representative 3D model of control (movie 1) or rUTI (movie 2) bladder SP⁺ nerves, highlighting the increased density and number of branching of nerves in rUTI compared to control). As lamina propria nerves are responsible for reception of sensory input, this data also places sensory sprouting as a potential cause of an abnormal sensory response, namely pain-evoked behavior and urinary frequency. To confirm that sensory sprouting was specific to rUTI, mice receiving only a single infection were assessed for sprouting at the equivalent time point of rUTI mice (28 days after first infection). Contrary to rUTI mice, single infected mice did not display increased sensory innervation compared to control mice (Figure S1I,J). To rule out persistent inflammation or altered detrusor function as alternative underlying causes of pain-like behavior and urinary frequency (47–50), bladders were harvested from rUTI mice and assessed for immune infiltration by histology and flow cytometry (Figure S1K–R) or detrusor function by muscle contractility assays (Figure S1S–V). rUTI mice did not display persistent infiltration of immune cells nor noticeable differences in detrusor function. Lastly, to rule out sprouting of sympathetic nerves, which could also contribute to altered bladder function, we probed control and rUTI mice for tyrosine hydroxylase, a marker of sympathetic nerves. rUTI mice did not demonstrate increased TH⁺ nerve presence, supporting the observed sprouting phenotype to be specific to the sensory nerve population in the bladder (Figure S1W–Y). Taken together, these data indicate profound sensory sprouting manifests in the bladder lamina propria specifically after rUTI, suggesting that increased pelvic sensitivity and urinary frequency are specifically associated with the sensory sprouting that occurs after rUTI.

Sensory sprouting and sensitization are mediated by elevated NGF during rUTI

After observing the profound increase in sensory innervation in the mouse bladder after multiple infections, we next characterized the underlying mechanisms involved. Sensory nerves throughout the body are maintained by neurotrophic factors produced and secreted by surrounding non-neuronal cells (51–55). Among these are nerve growth factor (NGF) and brain derived neurotrophic factor (BDNF), both of which have well-defined roles both in fostering the growth of nerve fibers during wound repair after tissue injury through axonal outgrowth and increasing their sensitivity through upregulation of nociceptive receptors and mediators (19, 22, 23, 56, 57). Based on this, we determined whether there were changes in NGF or BDNF expression in the bladder after rUTI by measuring protein levels in the bladder by ELISA. We found that while bladder lysates obtained from rUTI mice 7 days after third infection did not display altered BDNF compared to saline instilled mice, they did display significantly elevated NGF (p<0.05) compared to saline controls (Fig 2A,B), suggesting NGF but not BDNF as a potential mediator of sprouting and sensitization.

Figure 2: NGF Mediates Sensory Nerve Growth and Development of Pelvic Sensitivity during rUTI.

(A,B) Bladders were harvested from control and rUTI mice 7 days after the third instillation, homogenized, and analyzed for (A) NGF and (B) BDNF protein by ELISA (n=3–6 mice). (C-E) (C) Representative (left) immunofluorescence and (right) 3D models and quantification of SP⁺ (D) neurite length and (E) number of branch points in WT bladders after daily instillation of NGF into the bladder by transurethral catheterization (n=3 mice). Scale bar, 80 μm. (F) Mice instilled with NGF daily for 7 days were assessed for pelvic sensitivity 2 days after the final instillation (n=3 mice). (G-H) Neurons from L6 and S1 DRGs were harvested from mice and treated with either vehicle, NGF (25 ng/ml) or NGF and GW-441756 (7.25 μM) overnight. Scale bar, 50 μm. (H) The percent of neurons with neurites were quantified per group (n=5–6 images, approximately 150 neurons were counted per group). (I-L) rUTI mice were treated with vehicle or GW-441756 daily (IP and intravesicular instillation) starting at the second infection. Bladders were harvested 14 days after final infection and probed for SP to assess innervation. (I-K) (I) Representative (left) immunofluorescence and (right) 3D models and quantification of SP⁺ (J) neurite length and (K) number of branch points in rUTI mice treated with either vehicle or GW-441756 (n=5–6 mice per group (L) A separate group of vehicle or GW-441756 treated rUTI mice were assessed for pelvic sensitivity 14 days after final infection (n=4–6 mice). Z-stacks were acquired per mouse bladder from at least two independent experiments, with at least two mice per treatment group. Scale bar, 70 μm. (M-O) (M) Representative (top) immunofluorescence and (bottom) 3D models and quantification of SP⁺ (N) neurite length and (O) number of branch points in mice treated weekly (intraperitoneal injection; IP) with a NGF neutralizing antibody during rUTI (n=5 mice per group). Scale bar, 80 μm. (P) Mice undergoing NGF neutralizing were also assessed for pelvic sensitivity 7 days after the final anti-NGF treatment (n=5 mice per group). (A,B,D-F,J-L) Comparisons were analyzed by unpaired Mann-Whitney U Test with p<0.05 used to define significance. (H,N-P) Comparisons were analyzed by one-way Anova with Tukey post hoc test. p<0.05 used to define significance. Mean±SEM, ****p<0.0001, **p<0.01, *p<0.05, ns=not significant.

To address whether NGF mediates sprouting and sensitization, we undertook two approaches. First, we assessed whether instillation of native mouse NGF into the bladder of naïve mice was able to generate a similar phenotype as rUTI. Daily instillation of NGF for 1 week recapitulated in naïve mice both the sensory sprouting and increased pelvic sensitivity observed after rUTI (Figure 2C–F) suggesting NGF alone can induce the rUTI phenotype. Second, we inhibited NGF mediated signaling. NGF binds to the high affinity nerve growth factor receptor TrkA, leading to immediate signaling events such as sensitization of nociceptive receptors, in addition to delayed effects such as nerve growth after internalization and trafficking of the NGF:TrkA complex to the neuron cell body in the dorsal root ganglia (DRG) (19, 32, 58). Initially, we assessed whether the TrkA antagonist GW-441756 is capable of blocking nerve growth in vitro. Neurons were cultured from naïve mouse L6 and S1 DRGs, which innervated the bladder (59), and treated with either vehicle, native mouse NGF, or native mouse NGF in combination with GW-441756. While we observed extensive neurite growth in the neurons treated with NGF alone compared with vehicle treated cells, addition of GW-441756 abrogated this effect (Fig 2G,H), demonstrating the capacity of the antagonist to block NGF:TrkA mediated nerve growth. To further demonstrate that bladder innervating neurons specifically can sprout in response to NGF, WT mice were injected in the bladder wall with a fluorescent tracer to identify bladder specific neurons in the L6 and S1 DRGs, and these DRGs were harvested, cultured, and treated with NGF and GW-441756. As before, neurite growth was observed in the traced neurons after exposure to NGF, and this response was prevented with coadministration of GW-441756, revealing that bladder innervating neurons sprout in response to NGF in a TrkA dependent manner (Figure S2). After confirming its activity in vitro, we evaluated if the TrkA antagonist would prevent nerve growth during rUTI in mice. As before, all mice received three total instillations of E.coli, with each instillation occurring 7 days apart. Starting 2 hours before the second bacterial instillation, mice received once daily vehicle or GW-441756 treatment until 14 days after the third infection. Strikingly, compared with rUTI mice treated with vehicle, GW-441756-treated rUTI mice displayed significantly less bladder lamina propria innervation (p<0.01), highlighted by comparatively less nerve length and number of branch points (Fig 2I–K), revealing the ability of this antagonist to block NGF mediated nerve growth in our rUTI mouse model. We then evaluated whether blocking TrkA would prevent the development of pelvic sensitivity that we observe during rUTI. Compared with vehicle-treated rUTI mice, GW-441756-treated rUTI mice did not display increased pain-like behavior, suggesting that TrkA antagonism also prevents development of pelvic sensitivity (Fig 2L). Lastly, to confirm the impact of inhibiting NGF, we utilized a neutralizing antibody (anti-NGF) targeted at the neurotrophic factor. Weekly administration of anti-NGF in mice undergoing rUTI prevented both sensory sprouting as well as increased pelvic sensitivity, compared to rUTI mice treated with an isotype control antibody (Figure 2M–P), further confirming the role of NGF during rUTI sprouting and pain-like behavior. Taken together, these data reveal NGF to be the primary neurotrophic factor involved in both sensory nerve growth and pelvic sensitization during rUTI.

Recruited monocytes and bladder resident mast cells are the primary immune sources of NGF during rUTI

After identifying NGF as the primary neurotrophic factor driving sensory nerve growth and sensitization, we next determined its cellular source in the bladder. When we assessed hematopoietic (CD45⁺) and non-hematopoietic (CD45^-) cell populations in control and rUTI bladders 7 days after third infection, we found that there was a significant increase in the CD45⁺NGF⁺ cell population (p<0.01) in rUTI bladders (Figure 3A–B), likely corresponding to the elevated NGF in the bladder tissue at this same time point (see Figure 2A). Based on this and our data demonstrating lack of sprouting in single UTI mice (Figure S1I,J), we reasoned that the changes in NGF expression that initiate sprouting likely occur in CD45⁺ cells at some time after the second infection. To elucidate this, we first investigated the distribution of various innate immune cells in the bladder after one or two infections (Figure 3C). Compared with naïve mice, single-infected mice displayed a shift toward a higher frequency of CD11b⁺Ly6G⁺ neutrophil and CD11b⁺Ly6C⁺ monocyte populations, as well as fewer CD11b⁺F4/80⁺ macrophage and dendritic cell (both CD11b⁺CD11c⁺ and CD11b^-CD11c⁺) populations, consistent with an acute response to infection (Figure 3C and S3A–F). By Day 7 after infection, these shifts returned close to that of the naïve mouse (Figure 3C), consistent with a resolved bladder that has cleared urinary bacteria (26). Noticeably, a second infection elicited a strong monocyte and neutrophil response, demonstrating an ability of the bladder to respond to subsequent infections (Figure S3C,D). We did not observe any changes in the FcεRI⁺cKit⁺ MC population (Figure S3G) between controls and infected mice throughout the course of infections. Among these cell type, macrophages, monocytes, and mast cells have each been reported to express several neurotrophic factors including NGF (53, 60–62). Based on this, we then assessed the NGF expression within these three cell types and found that while macrophages displayed an overall decrease in NGF median fluorescence intensity (MFI) (Figure 3D), both monocytes (p<0.0001) and mast cells (p<0.05) displayed significant increases in NGF MFI 24 hours after the second infection (Figure 3E,F). It is likely that sensory sprouting does not occur after only one infection in mice due to (1) the lack of subsequent influx of NGF⁺ monocytes that does occur with second and third infection, and (2) the lack of increase in NGF production in both monocytes and mast cells that occurs with a subsequent infection. Altogether, these data reveal monocytes and mast cells as two of the primary cells responsible for enhanced NGF production in the bladder during rUTI.

Figure 3: Infiltrating monocytes and resident mast cells mediate sensory sprouting via NGF production after rUTI.

(A-B). Flow cytometry analysis of single-cell suspensions of bladder cells 7 days third UTI or saline to assess NGF. Corresponding gating strategy is provided in Supplemental Figure S3. Live cells were gated on NGF expression and quantifications of (A) CD45⁺ and (B) CD45^- cells were determined (n=5 mice). (C) Donut charts generated from flow cytometry analysis (see Figure S3) of CD45⁺ bladder cells at baseline, 24 hours after first UTI, 7 days after first UTI, and 24 hours after second UTI (given 7 days after first UTI). (D-F) Analysis of NGF MFI expression in (D) macrophages, (E) monocytes, and (F) mast cells (n=5 mice). (G,H) Neurons from L6 and S1 DRGs were harvested from mice and treated with either vehicle, native NGF (25 ng/ml), or cultured with bone marrow derived (G) monocytes or (H) mast cells in absence or presence of TrkA antagonist, GW-441756 (7.25 uM) overnight. The percent of neurons with neurites were quantified per group (n=6–10 images, approximately 180 neurons were counted per group). (I,J) Quantification of SP⁺ (I) neurite length and (J) number of branch points in Ccr2^−/− bladders after rUTI. (n=5 mice). (K-M) (K) Representative (left) immunofluorescence and (right) 3D models and quantification of SP⁺ (L) neurite length and (M) number of branch points in mice treated twice weekly (intraperitoneal injection; IP) with an antibody targeting Ly6C during rUTI (n=5 mice per group). Scale bar, 80 μm. (N,O) Quantification of SP⁺ (N) neurite length and (O) number of branch points in Kit^w-sh/w-sh bladders after rUTI (n=5 mice). (P-S) (P,R) Void frequency and (Q,S) pelvic sensitivity were assessed in (P) Ccr2^−/−, (Q) anti-Ly6C treated, and (R,S) Kit^w-sh/w-sh mice following rUTI (n=4–5 mice). All data representative of 2–3 independent experiments. (A-B,I,J,N-P,R,S) Comparisons were analyzed by unpaired Mann-Whitney U Test with p<0.05 used to define significance. (D-H,L,M,Q) Comparisons were analyzed by one-way Anova with Tukey post hoc test. p<0.05 used to define significance. Mean±SEM, ****p<0.0001, ***p<0.001, **p<0.01, *p<0.05, ns=not significant.

Next, we investigated if either monocytes or mast cells are consequential in promoting bladder nerve sprouting. Using in vitro culture (63, 64), we confirmed that both monocytes and mast cells induce nerve growth when cultured with neurons isolated from L6 and S1 DRGs (Figure 3G,H). Addition of the TrkA antagonist to the co-culture of either monocytes and neurons or mast cells and neurons prevented nerve growth, supporting a specific role of NGF from either immune cell in inducing this phenomenon (Figure 3G,H). We then confirmed the role of monocytes and mast cells in our in vivo model by utilizing genetic models that either display reduced monocyte migration (Ccr2^−/−) or mast cell deficiency (Kit^W-sh/W-sh). Inflammatory monocytes express the CCR2 chemokine receptor, allowing them to migrate in response to the chemokine CCL2 (65). Conceivably, the absence of their migration to the bladder should prevent sensory sprouting by limiting NGF content in the bladder. Likewise, mice lacking mast cells should also be protected from sensory sprouting if mast cells are also a vital source of NGF in the bladder during rUTI. After confirming that both Ccr2^−/− and Kit^W-sh/W-sh mice each lack monocytes or mast cells in the bladder after infection (Figure S4A–E) (26, 33, 66–68), we first examined neurite length and number of branch points of bladder nerve fibers in sham challenged and rUTI Ccr2^−/− mice. Here, we found no differences between the two groups of mice in either parameter after rUTI, indicating that Ccr2^−/− mice were protected from the sensory sprouting seen in WT mice, which implicates the role of inflammatory monocytes (Figure 3I,J and S4F) in sensory sprouting. Furthermore, utilization of a Ly6C blocking antibody to deplete monocytes during rUTI in WT mice also prevented sensory sprouting, compared to isotype antibody treated rUTI mice (Figure 3K–M). Likewise, when we assessed neurite growth and number of branch points in Kit^W-sh/W-sh mice, we similarly found no difference in either parameter after rUTI (Figure 3N,O and S4G). These findings confirm that both monocytes and mast cells are necessary for NGF mediated hyperinnervation in the bladder after rUTI. In addition, we questioned whether monocytes or mast cells contributed to the development of increased urinary frequency and pelvic sensitivity, since we previously demonstrated a connection between these phenomena (Figure 1F–G). When bladder function and pelvic sensitivity in control and rUTI monocyte-targeted or mast cell-targeted mice were compared, no differences in frequency or pain-like behavior were observed (Figure 3P–S) suggesting that in rUTI mice, the accumulation and/or presence of both NGF⁺ Ly6C⁺ monocytes and NGF⁺ mast cells were necessary for both sensory sprouting and subsequent pelvic sensitivity and urinary frequency.

Sustained release of NGF sensitizes nociceptors and promotes pain-like responses and urinary frequency

After determining the cause of sensory sprouting in the bladder after rUTI, we next determined what drives persisting sensitization and associated pathology after resolution of infections. We hypothesized that NGF may be involved as it has an established role in sensitizing peripheral nerves to subsequent stimulation and we found it to be elevated through 7 days after a third infection (Fig 2A). Interestingly, when we treated rUTI mice with the TrkA antagonist GW-441756 on days 13 and 14 after the third infection, we found that it decreased pelvic sensitivity, confirming the persisting role of NGF and supporting the notion of it sensitizing the bladder (Fig 4A). To corroborate this, we assessed the bladder nerves for signs of nociceptive sensitization. We first probed for pERK⁺ nerves in the rUTI bladder since phosphorylation of ERK has been used as a marker of nociception in neurons (69–71). We observed increased pERK expression in SP⁺ sensory nerves adjacent to bladder mast cells in rUTI mice compared to saline control bladders (Figure 4B,C and S5A), suggesting that these nerve fibers are sensitized. In addition, to confirm sensitization of bladder nerves specifically, we utilized conjugated wheat germ agglutinin (WGA) to trace bladder nerves to their neurons in the DRGs of saline and rUTI mice. We found that rUTI mice had a higher number of bladder-traced neurons that were also pERK⁺ (Figure 4D,E), supporting the conclusion that bladder nerves are sensitized during rUTI. Interestingly, when we examined cross-sections of rUTI mast cell-deficient (Kit^W-sh/W-sh) mice for the presence of sensitized sensory nerves, we found no difference in the numbers of sensitized nerves between controls and rUTI mice (Figure 4F and S5B), indicating that mast cells are likely responsible for their sensitization after rUTI. When we assessed pelvic sensitivity in rUTI mice that were depleted of mast cells specifically after sprouting had occurred (Mcpt5-iDTR) (34, 72) (Figure S5C), we found that these mice were protected from increased responses after rUTI (Figure 4G), further confirming the role of mast cells in bladder nerve sensitization after rUTI.

Figure 4: Mast Cells Mediate Pain-like Behavior and Void Defects via Mediator Release.

(A) rUTI mice were treated with either vehicle or GW-441756 on days 13 and 14 after final infection and assessed for pelvic sensitivity (n=5 mice). (B,C) Bladders were harvested from control and rUTI mice and immunofluorescence was used to identify lamina propria mast cells and activated (pERK⁺) SP⁺ nerves in frozen sections. (C) Mast cells (identified by avidin staining; pointed by the arrow) were noted in proximity of sensory nerves. Scale bar, 50 μm. (B) Increased presence of activated nerves (pERK⁺SP⁺) in rUTI bladders compared to saline control (n=6 mice). (D,E) (D) Control and rUTI mouse bladders were injected with fluorescent WGA (WGA-647) and L6 and S1 DRGs were harvested 5 days later to assess bladder innervating neurons. DRGs were subsequently stained with (green) Beta III Tubulin and (red) pERK as shown in the merged image. Double positive neurons are demarcated by a star (n=3 mice). Scale bar, 50 μm. (E) Quantification of pERK⁺ neurons among the traced neurons. (F) Bladders were harvested from control and rUTI Kit^w-sh/w-sh mice and activated SP⁺ nerves were quantified (n=6 mice). (G) rUTI was performed in MCPT5-iDTR mice or littermate controls prior to DT injection to deplete mast cells. Pelvic sensitivity was assessed in both groups of mice after DT injection (n=3 mice). (H) Mast cells were cultured (64) and treated with 200 μM SP for 24 hours. Supernatant was collected from untreated and stimulated mast cells and assessed for NGF protein content. (I) Urine was collected from WT control and rUTI mice 7 days after the first and second infection and 14 days after the third infection and assess for histamine levels by ELISA (n=4–11 mice). (J,K) Quantification of SP⁺ (J) neurite length and (K) number of branch points in Trpv1^−/− mice during rUTI (n=4 mice per group). (L) Trpv1^−/− mice were also assessed for pelvic sensitivity after rUTI (n=4 mice per group). (M,N) WT mice instilled with saline, bradykinin, or histamine for 15 min were assessed for (M) pelvic sensitivity and (N) frequency 1 hour after instillation (n=5 mice). (O) Trpv1^−/− mice instilled with histamine for 15 min were assessed for pelvic sensitivity 1 hour after instillation (n=5 mice). (P,Q) WT mice after control or rUTI treated with (P) pyrilamine maleate or (Q) icatibant were assessed for pelvic sensitivity (n=5 mice). (A,C,E-H,O-Q) Comparisons were analyzed by Mann-Whitney U Test with p<0.05 used to define significance. (I,M,N) Comparisons were analyzed by unpaired Student T-Test with p<0.05 used to define significance. (J-L) Comparisons were analyzed by one-way Anova with Tukey post hoc test. p<0.05 used to define significance. Mean±SEM, ***p<0.001, **p<0.01, *p<0.05, ns=not significant.

As the monocyte population has returned to baseline level (Figure S1M), it is likely that mast cells are a prominent source of NGF after rUTI resolution (14 days after last infection). Since mast cells respond to SP, which is present in the sprouting sensory nerves after rUTI, via the MRGPRB2 receptor (73), we assessed whether treatment of cultured mast cells with SP could induce NGF secretion. We found that SP was able to induce NGF secretion from cultured mast cells, implying the potential formation of a positive feedback loop between the sprouted SP⁺ sensory nerves and mast cells (Figure 4H). This implication was also supported by the finding of partially degranulating mast cells in the bladder of rUTI mice compared to saline controls (Figure S5A).

In addition to NGF, mast cells release several other mediators that can stimulate surrounding nociceptors. These include bioactive compounds such as histamine and heparin. Histamine is linked to neuronal activation and detectable in the urine of rUTI patients (74, 75). When we assessed control and rUTI urine for histamine content, we found rUTI urine to contain higher levels of histamine (Figure 4I). Mast cell heparin contributes to bradykinin production, which is known to mediate nociceptive activation of nerves. Both histamine and bradykinin signaling involves activity of histamine or bradykinin receptors, respectively, leading to enhancement of TRPV1 channel activity and nociceptor activation (30, 76). To ascertain the involvement of TRPV1 during rUTI, Trpv1^−/− mice underwent saline or rUTI and were assayed for sensory nerve sprouting and pelvic sensitivity. While Trpv1^−/− mice did display sensory nerve sprouting, they were protected from developing pelvic sensitivity, suggesting a role of TRPV1 specifically in pain-like behavior after rUTI in mice (Figure 4J–L). Furthermore, when resiniferatoxin (RTX) was used to deplete sensory nerves (77, 78) in mice after rUTI (Figure S5D) we found that mice no longer displayed urinary frequency (Figure S5E), implicating a role in TRPV1 in altering bladder function after rUTI. After confirming both histamine and bradykinin can activate TRPV1⁺ nerves in vitro by stimulating calcium mobilization (Figure S5F), we next determined whether these products specifically are involved in disease symptoms after rUTI. First, we investigated if introducing each mediator directly into the naïve mouse bladder was sufficient to increase pelvic sensitivity and urinary frequency. We found that a single administration of either histamine (0.0825 mg/kg in 30μL) or bradykinin (0.00795 mg/kg in 30μL) into the bladders of naïve mice each increased pelvic sensitivity (Figure 4M), which was comparable in magnitude to that previously observed in the rUTI mouse bladders (Figure 1F). Interestingly, only histamine instillation in the naïve mouse bladder led to urinary frequency (Figure 4N), comparable to that seen previously in the bladders of rUTI mice (Figure 1G), suggesting possible differences in the respective roles of these mediators. When histamine was instilled in Trpv1^−/− mice, it failed to elicit increased pelvic sensitivity, confirming the role of TRPV1 in histamine mediated nociception (Fig 4O) (79). Lastly, to demonstrate the specific roles of histamine and bradykinin in our rUTI model, we treated rUTI mice with the histamine receptor (H₁R) antagonist pyrilamine maleate or the bradykinin 2 receptor (B₂R) antagonist icatibant (80, 81). rUTI mice treated with two IP injections of the H₁R antagonist or B₂R antagonist reduced pelvic sensitivity, demonstrating the role of histamine and bradykinin in mediating pathology after rUTI (Figure 4P,Q). These data support the notion that MC products such as histamine and heparin-generated bradykinin are sufficient to provoke pelvic sensitivity and/or urinary frequency via TRPV1 and their respective receptors and, when considered with the observation of NGF-mediated nerve sprouting and sensitization in the rUTI bladder, may play a role in mediating these symptoms after rUTI (Figure S6).

Discussion

The underlying basis for why rUTI patients suffer chronic pelvic pain despite their urine cultures being negative has until now remained a mystery. Our studies reveal that this condition is, at least in part, attributable to excessive sensory nerve sprouting in the bladder. Our studies of bladder biopsies and urine of rUTI patients revealed not only enhanced SP staining in the lamina propria, but also elevated levels of urinary SP, suggesting that sensory nerves in the bladder of rUTI patients are in an activated state. This finding of sensory nerve modulation after UTI was recapitulated in our mouse model of rUTI, where mice subjected to three consecutive UTIs were found to display significant enhancement of neurite length and branch points specifically in the lamina propria 14 days after the third infection. This layer of the bladder is increasingly thought to play an integrative role in signal transduction to the central nervous system involving both nociception and mechanosensation (82). We have attributed this rUTI-induced sensory nerve sprouting to enhanced production of NGF, a common bladder neurotrophic factor. Evidence for the role of NGF comes from our findings that NGF specifically was elevated in bladder tissue of rUTI mice. Furthermore, antagonists to the NGF receptor TrkA and neutralizing antibodies targeting NGF were found to block neuronal sprouting in vitro and in vivo when administered during rUTI, and administration of native NGF into the bladders of naïve mice was able to cause sensory sprouting. Whereas several cell types produce NGF, we found that the hematopoietic cell population in the bladder displayed a positive shift in their proportion of total cells that express NGF during rUTI. Among these NGF⁺ cells were a combination of both bladder-resident and recruited cells, namely recruited Ly6C⁺ inflammatory monocytes and bladder-resident mast cells. Inflammatory monocytes migrate as a large wave into the bladder after each bout of UTI only to wane to base-line levels 7 days later, presumably when most of the bacteria had been cleared. A recent report has also shown that large numbers of monocytes arrive in the bladder following UTIs to facilitate bacterial uptake (83). Interestingly, although there is robust recruitment of monocytes to the bladder during the first infection, the relative expression of intracellular NGF within monocytes did not noticeably increase until after the second infection, as was the case for mast cells. Presumably, the dramatic increase in NGF production during the second infection may be related to a general shift in the bladder following UTI toward a repair and reepithelization mode, as NGF also promotes tissue repair (35, 84–86). The importance of inflammatory monocytes and mast cells to nerve sprouting in the bladder was deduced from the findings that targeting of either cell types using genetic knockout models (Ccr2^−/− or Kit^W-sh/W-sh) or depletion studies using Ly6C neutralizing antibody prevented sprouting after infection. It is noteworthy that both cell types appear to be necessary for sprouting to occur during rUTI. Quite possibly, monocytes may be necessary for an early increase in NGF by virtue of their migration to the bladder during acute infection, while mast cells play a more critical role in sustaining elevated NGF after inflammation has subsided. Since the monocyte population during acute infection is noticeably larger than that of the mast cell population, their recruitment would theoretically result in a much faster increase in total NGF in the bladder, compared with that achieved by mast cells alone. Without mast cells present, NGF within bladder may closely mimic the population shift of monocytes from acute infection to resolution and effectively be unable to sustain nerve growth. Regardless, rUTI leads to dramatic remodeling of bladder tissue including enhancement in innervation.

The pathological consequences of bladder nerve sprouting were revealed to be pelvic pain-like behavior and voiding defects, as (1) WT mice that have experienced rUTIs exhibited these pathologies in contrast to uninfected mice, and (2) Ccr2^−/− and Kit^W-sh/W-sh mice exhibited no voiding defects even after rUTIs. We have identified TRPV1 on the burgeoning bladder nociceptors as the critical pain sensor responsible for mediating rUTI symptoms by demonstrating protection from pelvic pain behavior after rUTI in Trpv1^−/− mice and protection from bladder voiding defects after specifically depleting this cation channel with resiniferatoxin (RTX) treatment in the bladder in rUTI WT mice. That a connection exists between bladder nociceptors and the circuitry controlling bladder pain-like behavior and voiding comes from our observations that administering compounds known to modulate TRPV1 on neurons such as histamine and bradykinin (76, 87) were sufficient to trigger both pain-like behavior and urinary frequency even in naïve mice. It has been reported that bladder-innervating nociceptors can mediate “reflex-driven” bladder voiding but in pathologic states the same circuitry can contribute to debilitating bladder pain and/or frequency (88). NGF is known to sensitize nerves by lowering their activation threshold and promoting increased production of pain receptors, which is likely occurring during rUTI based on our data revealing persistent elevation of NGF after rUTI resolution and effective targeting of NGF through either NGF neutralization or TrkA receptor blockade resulting in reduced pain. In addition to the increased physical presence of sensory nerves in the bladder, the nerves themselves may have increased expression of TRPV1, H₁R, B₂R, and other receptors, leading to increased activation. Whether increased receptor production occurs during rUTI and how long pain behavior persists after rUTI remains to be seen; however, the increased presence of nociceptive nerves appears to be the critical factor driving the bladder into a pathological state during rUTIs.

Mast cells and their potential roles in pain have been the focus of several studies in the bladder, with many predominantly focusing on sterile bladder pathologies (89–94). For example, in sterile pathology interstitial cystitis/painful bladder syndrome (IC/PBS), mast cells were found to play a pivotal role in contributing to pain through increased numbers in the bladder and through histamine release (90, 91). In addition, nerve sprouting and spatial proximity of mast cells to nerves in the bladder is observed clinically in IC/PBS (13, 89). Here, we depict the relationship between mast cells and nerves in orchestrating pain transmission after rUTI. Mast cell-deficient mice exposed to multiple UTIs failed to exhibit either pain-like behavior or aberrant bladder voiding. Strikingly, we observed that depleting mast cells after exposing these mice to rUTIs prevented both pelvic pain-like behavior and aberrant voiding responses in the hyperinnervated bladder, suggesting that mast cells were necessary for triggering these bladder pathologies. In this particular model, mucosal mast cells are not specifically depleted, suggesting that these mast cell-mediated effects we observe are due to connective tissue mast cells; this is further supported by previous work revealing the majority of bladder mast cells are depleted after diphtheria toxin injection in this model (34). We noted mast cell activation morphologically consistent with partial degranulation in the rUTI bladder compared to saline control mice, as opposed to the fully degranulated phenotype commonly observed with classical IgE mediated degranulation. The partial degranulation of mast cells is often described as one undergoing “piece meal degranulation”, which allows selective and sustained release of specific mast cell products including histamine (95, 96). This form of degranulation has been shown previously to be associated with visceral hypersensitivity induced by NGF released by mast cells in the gut (97), which is comparable to our observation in our rUTI mouse model, as well as neurogenic inflammation in the skin resulting in pain and itch (98). The work presented here supports SP released from stimulated sensory nerves as a likely candidate causing mast cell activation, as SP⁺ nerves sprout during rUTI and mast cells respond to SP. This would correlate with the observation of increased urinary SP in patients after rUTI and it is known that both human and mouse mast cells express receptors for SP (73, 99–101). Furthermore, since mast cell products can stimulate and sensitize TPRV1⁺ nerves to bladder distension (102), natural filling of the bladder after rUTI may be sufficient to initiate noxious stimulation of nerves and subsequent neuropeptide release and mast cell activation. Alternatively, UPEC which are known to persist in small numbers in the bladder for extended periods after each bout of UTI (6, 103, 104) could be a candidate, although it is unknown if these residual, quiescent bacteria are able to stimulate nerves in the urothelium. Regardless, this crosstalk between mast cells and nerves and the possible formation of a “continuous signaling loop” can potentially persist for protracted periods and contribute to chronic pain and defective voiding.

Many rUTI patients continue to receive antibiotic treatments even though there is no significant culturable bacteria in their urine. However, this effect may not be because they are eliminating residual bladder bacteria but rather, because antibiotic treatment may somehow impact brain signaling pathways underpinning pain regulation (105–107). One of the limitations of our work is we cannot completely rule out the contribution of other immune cells, even though our data does suggest that several other innate immune cell populations appear to decrease throughout the course of subsequent infections. Although we found CCR2 expressing monocytes to be important for both nerve sprouting and bladder function, the cellular source of its major chemokine CCL2 is currently unknown. Several resident bladder cell types have the potential of producing this chemokine during UTI (108–111). Whichever cells produce CCL2 during UTI, it is clear that it is only for a limited period of time as monocyte recruitment subsides by the third day of each infection. It is also important to note that in our model, both voiding defects and pelvic sensitivity were observed in rUTI mice up to 14 days after a third infection. At this time it is not known whether these symptoms persist for even longer and if the hyperinnervation phenotype represents a permanent change in bladder tissue following rUTIs. In view of our findings that neutralization of NGF signaling and mast cell mediator receptor blockade can markedly reduce pain-like behavior in rUTI, it may be appropriate to introduce therapy targeted at preventing sensory nerve sprouting and activation which should offer better and early relief from the most pressing symptoms experienced by the rUTI patient population.

Materials and Methods

Study Design

We utilized a rUTI mouse model to recapitulate clinically relevant symptoms and phenotypes. Using this infection model, we specifically assessed voiding behavior, pain-like behavior, and nerve anatomy with the aid of several assays, techniques, and analyses including cystometry, histology, and 3D model reconstruction. ELISA and flow cytometry were used to determine relevant proteins and cell types connected to the rUTI pathology and multiple pairs of targeting mechanisms were used in tandem to confirm these observations, including usage of receptor antagonists and neutralizing antibodies for NGF, CCR2^−/− mice and Ly6C blocking antibodies for monocytes, and Kit^W-sh/W-sh and MCPT5-iDTR mice for mast cells. To further confirm these studies, native NGF, histamine, and bradykinin were administered to naïve mice to recapitulate the primary observations in the rUTI model. Samples were collected from human patients to corroborate our mouse data and published clinical observations. Informed consent was obtained from all subjects involved in the study. Human samples were acquired according to Ewha Womans University Medical Center Institutional Review Board Protocol number 2014–12-050.

Human Samples

Tissue biopsy and urine collection

Bladder biopsies and urine were obtained from control and rUTI patients who were experiencing pain. Recurrent UTI of these patients were diagnosed based on medical records indicating the presence of pyuria and bacteriuria (urine culture positive) on two occasions within the last six months or at least three times in a year. These patients exhibited the following symptoms: dysuria and frequent urination, or urgency, or suprapubic discomfort. The following are exclusion criteria. Pregnant individuals and those with prolonged indwelling catheters, kidney stones, bladder cancer, urological surgery within 6 months or anatomical abnormalities. Patient cohorts were all female with a median age of 62 (range 29–79). Control urine was obtained from patients who have no history of rUTI. Urine collection was performed via urethral catheterization. The collected urine from these patients was cultured to determine the presence of uropathogens. The rest of the urine was analyzed by ELISA (Human Substance P ELISA kit; Cusabio, Cat: CSB-E08357h).

Tissue processing and immunostaining

Bladder biopsies were processed after fixation in formaldehyde. Paraffin-embedded tissues were cut at 5 um and deparaffinized. To remove formalin-induced antigen masking, heat-mediated antigen retrieval was performed in sodium citrate buffer. After blocking the tissue sections with a solution containing 10% donkey serum, 0.3% Triton X, and 1% bovine serum albumin in PBS, samples were incubated with anti-substance P antibody (Santa Cruz, Cat: sc-21715), anti-E-cadherin (BD, Cat: 610182), and fluorochrome-conjugated wheat germ agglutinin. To visualize the target, a fluorochrome-conjugated secondary antibody was used.

Animal Studies

All mice used in studies were female mice purchased at 6–8 weeks old and were 7–8 weeks old at the initiation of each experiment. Female C57BL/6J (Stock No: 000664), MC-deficient Wsh mice (Kit^W-sh/W-sh), Ccr2^−/− (Stock No: 004999), iDTRf^+/+ (Stock No: 007900), and Trpv1^−/− (Stock No: 003770) mice were each purchased from Jackson Laboratories. MCPT5-iDTR mice were generated by crossing MCPT5-cre mice with DTR mice. Advillin-GCaMP6 mice were generated by crossing GCaMP6f mice with Advillin-Cre mice (Jackson Laboratory). Mcpt5-Cre strain was a gift from Dr. Axel Roers, University of Technology, Dresden, Germany. Advillin-Cre strain was a gift from Dr. Fan Wang at Duke University. Mice were housed in Duke University approved facilities where the animals received water and food ad libitum. Experiments were performed in accordance with a Duke University Institutional Animal Care and Use Committee approved animal protocol.

Bacterial Culture and Bladder Infection

Clinical isolate E.coli strain J96 (UPEC) was used for all infection experiments. UPEC was grown stationary for 18–22 hours at 37°C in Luria-Bertani broth (Sigma) before utilization in infection procedures. Mice were anesthetized with ketamine (100 mg/kg; intraperitoneal injection) prior to bladder infections. A 2 cm long PE-10 catheter was inserted into the bladder via the urethra until the bladder was reached. Urine was expelled from the bladder by applying gentle pressure to the abdomen. 10⁸ UPEC in 30 μL was then instilled in the bladder slowly via a needle inserted in the open end of the catheter and allowed to remain for 30 minutes. After, the bacterial suspension was removed from the bladder. A second and third identical infection was performed at 7 and 14 days after the first infection, respectively, for the rUTI model.

Pelvic Sensitivity Assay

Pelvic sensitivity was determined by von Frey filament application. Tests were performed blindly. Mice were allowed to acclimate to the testing chamber for at least 1 hour before performing the test. A 0.04g filament (Ugo Basile) was applied to the pelvic area a total of 10 times and responses (licking, jumping, or other sudden movements) were recorded. For 50% paw withdrawal threshold, a range of filaments was applied in an ascending manner to the hind-paw until a positive response was recorded, according to the Up-Down method. Scores were determined based on a conversion chart.

Cystometry

Surgery was performed on mice 7 days before cystometry analysis. A PE 10 catheter was inserted into the dome of the bladder and routed from the abdominal cavity to the back of the mouse and attached to the skin until the day of testing. Cystometry was performed in fully conscious, restrained mice. Mice were contained in a small cylindrical device (Natsume Seisakusho) that was situated inside a small animal cystometry lab station (Med Associates). To measure void volume, an analytical balance was placed below the animal, with volume defined as the amount increase on the balance. Saline was infused through the catheter at a fixed rate of 15 μL per min for 2–3 hours via a remote-controlled syringe pump. The pump was equipped with an in-line pressure transducer to measure pressure changes. Measurements were recorded with the Med-CMG software (Med Associates) and analyzed with CMG Analysis software (version 1.06 Med Associates). 5 consecutive voiding cycles collected at least 30 minutes after the start of the experiment were used for data analysis. Frequency was defined as the number of voids per hour.

Microscopy

Cross-section

Bladders and DRGs were harvested from mice after euthanasia fixed in 4% paraformaldehyde for 1.5 hours. Tissues were then washed three times in PBS before being placed in a 30% sucrose solution for 3 days at 4°C. Samples were then placed in OCT compound (Sakura Finetek) and frozen at −80°C before cryosection. 15 μm sections were collected on glass slides and placed in acetone for 20 minutes at 4°C. Slides were rehydrated and blocked with a 2% BSA/0.1% saponin PBS solution. Primary and secondary antibodies were applied in this solution as well. Primary incubation was performed overnight at 4°C using individual antibodies against beta 3 tubulin (Clone Tuj1; R&DSystems Cat: MAB1195), substance P (Santa Cruz Cat: sc-21715), pERK (Cell Signaling Cat: 9101). Each primary was used at 1:500 dilution. Three washes in the blocking buffer was performed before secondary antibody application. Secondary incubation was performed for 1 hour at ambient temperature using a combination of secondaries (Jackson ImmunoResearch) or conjugated probes including avidin-FITC (Biolegend Cat: 405101) and wheat germ agglutinin (ThermoFisher Cat: W32466). After 3 washes, slides were mounted in Fluoromount containing DAPI.

Whole Mount

After collecting the bladders, the samples were fixed using 4% paraformaldehyde at 4°C overnight. Each bladder was then embedded in low melting-point agarose for vibratome sectioning, to achieve 500 μm thick sections. Sections were then permeabilized in methanol gradients, followed by blocking in phosphate buffered saline (PBS) with 5% donkey serum, 5% DMSO, 0.3% TritonX-100 and 0.2% bovine serum albumin (BSA) at ambient temperature overnight. For immunostaining, samples were incubated with primary and secondary antibodies for 3 days at 4°C, respectively, followed by washing in wash buffer (0.3%Triton X-100 + 3%NaCl in PBS) to wash the samples after each incubation. Primary Abs: Rabbit anti-TUJ1 (Abcam Cat: ab18207); secondary Abs: Avidin-FITC, Wheat Germ Agglutinin-AF555 (ThermoFisher Cat: W32464), and Goat Anti-Rat AF647 (Thermofisher) with 4′,6-diamidino-2-phenylindole (DAPI). Stained samples were then optically cleared by dehydration in methanol gradients, and clearing in BABB (benzyl alcohol and benzyl benzoate V:V = 1:2) before imaging.

For confocal imaging, the detrusor and lamina propria were captured respectively on the Olympus FV1000 using the 20X LUCPLFLN20X objective for each captured area with 250 μm Z-stack. Both urinary tract infected and untreated groups were acquired with the same imaging conditions.

Images were analysed using the Imaris software (Bitplane, Ver. 9.5, 9.6, and 10.0). The surface function of Imaris was used to reduce the non-specific signal and render 3D modeling surfaces for both the TUJ1 channel and DAPI channel based on the mean fluorescence intensity, to achieve the morphology of nerve fibers and the whole tissue. In addition, Innervation Index was defined as follows:

$$\text{Innervation}\text{Index}=(\text{nerve}\text{fiber}\text{volume})/(\text{tissue}\text{volume})\times 1000$$

Innervation Index was calculated for each capture area and carried out statistical comparison of both the UTI and control groups. Software used: Imaris (Bitplane) for imaging processing and 3D modeling, Prism (Graphpad, Ver. 10) for statistics and plotting.

Samples used for nerve length and branch point quantification was processed for whole mount in a modified manner. Samples were fixed using 4% paraformaldehyde at ambient temperature for 1.5 hours and subsequently washed thrice with PBS before blocking in 2% BSA, 3% goat serum, and 0.3% Triton X-100 PBS solution for 1.5 hours at ambient temperature. Primary Abs: Rat anti-SP (1:500), Sheep anti-TH (Millipore Sigma, Cat AB1542; 1:250), Mouse anti-Tuj1 (Rndsystems; 1:250) Rat isotype (Santa Cruz Cat: sc-2026). Staining protocol was identical to a protocol used for cross-section; only, secondary incubation was performed for 2 hours. Samples were mounted in aqua polymount and imaged. All images were acquired from the lower third of the bladder, in the trigone. 20 μm Z-stacks taken just below the basement membrane (lamina propria) were used for Imaris processing. 3D models were constructed in Imaris and nerve length and the number of branch points were automatically quantified with the software.

For the image quantifications of phospho-ERK, the mean pixel intensity of phospho-ERK staining was measured using NIH ImageJ. The images from a minimum of two bladder sections representing n=3–4 mice in each group were used in the quantification.

Histology

Mast cell staining was performed using Carnoy’s fixative and 0.1% toluidine blue. The degranulating mast cells were visually counted from the serial bladder sections representing n=5 mice in each group and presented as an average of degranulating mast cells per mouse bladder. Standard hematoxylin and eosin staining was performed to visualize tissue cellularity.

Bladder Contraction Assay

Bladders were isolated from mice and incubated in cold Krebs solution, pH 7.4. Dome and trigone were removed, and mucosa was removed to expose the detrusor region. Two circumferential muscle strips (2mm × 6mm) were used in myographs (DMT 820M, Aarhus, Denmark). Tissue was maintained in Krebs solution (37°C, 5% CO₂) and stretched to a resting tension of 2 mN for 45 min equilibration. Next, 120 mM potassium chloride, carbachol, or electrical field stimulation was used to evoke detrusor contractions. Dose curves were used for carbachol and electrical field stimulation. Atropine (10 μM) or PPADs (100 μM) were added to electrical field stimulation to assess noncholinergic and purinergic mediated contraction. Responses were recorded using Lab Chart 8 software and Power Lab acquisition hardware (AD Instruments, Colorado Springs, CO). Data were measured as a percent of the maximum response to potassium chloride or relaxation from carbachol.

Neurotrophin and Histamine ELISA

NGF, BDNF, and Histamine ELISA were each performed according to manufacturer instructions (NGF, Raybiotech ELM-bNGF-1; BDNF, Biosensis BEK-2225-SET; Histamine, Enzo ENZ-KIT140A). For NGF and BDNF, bladders were isolated and homogenized before collecting lysate. Protein concentrations were calculated and 50 μg of lysate were used per sample. For Histamine, urine was collected from anesthetized mice and spun down to remove particulates. 100 μL of urine was used per sample.

DRG Culture and Neurite Growth Assay

L6 and S1 dorsal root ganglia were collected from in WT mice and incubated with enzymes (2 mg/mL collagenase and 2.5 mg/mL dispase-II) to dissociate cells for 1 hour at 37C. Flame-polished Pasteur pipettes were used to mechanically agitate the ganglia prior to plating on poly-d-lysine coated glass coverslips. Cells were allowed to adhere for 30 min before addition of native NGF (25 ng/ml, Sigma Cat: N6009) or GW-441756 (7.25 μM; Tocris, Cat: 2238). Cells were allowed to incubate for 14 hours before processing. Coverslips were fixed using 4% paraformaldehyde at ambient temperature for 20 min and subsequently washed twice with PBS before blocking in 2% BSA/0.1% saponin in PBS solution for 1 hour. Primary incubation was performed overnight at 4°C using an antibody against beta 3 tubulin at 1:200 dilution. Three washes in the blocking buffer were performed before secondary antibody application. Secondary incubation was performed for 1 hour at ambient temperature using a FITC conjugated secondary (Jackson ImmunoResearch). After 3 washes, slides were mounted in Aqua Polymount. Neurite⁺ neurons were defined as neurons with processes that were equal to or greater in length than the diameter of the cell body. At least 150 neurons were quantified per treatment group for analysis.

NGF Instillation in vivo

Naïve WT mice were administered via transurethral instillation into the bladder (identical to the instillation technique used in “Bladder Infection” above) native NGF (2 ug in 40uL 10% DMSO/PBS) or vehicle (40 uL 10% DMSO/PBS) daily while under isoflurane anesthesia for 30 min for 7 days. 2 days after the final instillation, mice were assessed for pelvic sensitivity as outlined above. Afterwards, bladders were harvested to assess innervation neurite length and number of branch points via whole mount microscopy and 3D reconstruction.

NGF Neutralization in vivo

Mice were administered once-weekly (intraperitoneal injection; IP) injections of anti-NGF (Bio X Cell, Cat: SIM0017; 10 mg/kg) or isotype antibody (Bio X Cell, Cat: BP0301; 10 mg/kg) starting 1 day before the second of three infections. Each mouse received a total of 3 injections, with the final injection occurring on day 6 after the third infection. Mice were assessed for pelvic sensitivity on day 14 after the third infection and bladders were subsequently harvested to assess innervation neurite length and number of branch points via whole mount microscopy and 3D reconstruction.

TrkA Antagonism in vivo

rUTI mice were treated with GW-441756 (1 mg/kg) or vehicle (10% Tween 80 in PBS) daily starting on day 7 after the first infection (total of 21 treatments). Each treatment solution was administered via IP injection (total volume 100 μL) and bladder instillation (total volume 50μL). These mice then underwent pelvic sensitivity tested at 14 days after third infection as outlined above. For mice receiving TrkA antagonism after rUTI, GW-441756 was administered at the same dose and volume as above on day 13 and 14 after the third infection. Pelvic sensitivity testing was then performed as outlined above.

Bone marrow derived cell culture (monocytes and mast cells) and DRG Co-Culture

Bone marrow derived monocytes and mast cells were generated from female C57BL/6J according to Francke et al and Malaviya et al, respectively. Monocytes or mast cells were added to cultured L6-S1 neurons 30 min after plating at an immune cell to neuron ratio of 5:1. GW-441756 (7.25 μM) was added to relevant wells 5 min before addition of either monocytes or mast cells. 24 hours after addition of immune cells to neuron culture, coverslips were fixed and processed according to Neurite Growth Assay section above. At least 180 neurons were quantified per treatment group for analysis.

Ly6C Depletion

Mice were administered twice-weekly (intraperitoneal injection; IP) injections of anti-Ly6C (Bio X Cell, Cat: BE0203; 10 mg/kg) or isotype antibody (Bio X Cell, Cat: BE0089; 10 mg/kg) starting 2 days and 30 min before the second of three infections. Each mouse received a total of 6 injections, with the final 2 injections occurring on days 5 and 7 after the third infection. Mice were assessed for pelvic sensitivity and sensory innervation in a manner identical to the description in “NGF Neutralization in vivo” above.

Retrograde Tracing of Bladder Neurons

Naïve, saline instilled, and rUTI mice were anesthetized and a midline incision of the abdomen was performed to expose the bladder. Wheat germ agglutinin conjugated to Alexa Fluor 647 (1 mg/kg) was injected via 30-gauge needle and Hamilton syringe into 3–4 areas of the bladder, injecting a total volume of 10 uL. After suture and recovery of the mice, euthanasia was performed 5 days later and L6 and S1 DRGs were harvested and either processed according to “DRG Culture and Neurite Growth Assay” or processed according to “Microscopy Cross-Section” described above. For traced DRGs used in “DRG Culture and Neurite Growth Assay”, only traced neurons (displaying signal for WGA-647) were used for quantification.

Sensory Nerve Depletion

Resiniferatoxin (RTX; 1 μM in 30μL, Sigma Cat: R8756) was instilled in the bladders of isoflurane-anesthetized mice for 20 minutes. Catheterization was identical to that used for UPEC infection. Cystometry was performed 24 hours after RTX depletion. Depletion of neuropeptide was confirmed in a separate group of mice by probing bladders harvested 24 hours after instillation with antibodies for SP.

Mast Cell Depletion

To deplete MCs, MCPT5-iDTR and littermate controls (DTR) mice were injected (i.p.) with 400 ng diphtheria toxin (DT) once every other day for a total of 5 injections. Depletion was confirmed via whole-mount microscopy. Mice received the first DT injection 1 week following the third bladder infection.

Flow Cytometry

Bladders were harvested from mice 24 hours after first or second infection and single-cell suspensions were prepared. Briefly, bladders were placed in an HBSS solution containing collagenase (Sigma, 1 mg per mL and DNase I (Sigma, 200 μg per mL) for 40 min. The tissue was then passed through a 70-μm cell strainer and transferred to a PBS buffer solution containing 3% fetal bovine serum and 5 mM EDTA. 1% anti-mouse CD16/CD32 (BD Biosciences Cat: 553141) and 5% each of normal rat and normal mouse serum in buffer solution was used to block at 4°C for 15 min. Staining was performed for 30 min at 4°C protected from light using the following antibodies (all anti-mouse BioLegend unless stated otherwise): FITC-FcεR1 (Cat: 134305), PE Cy7-cKit (Cat: 135112), BV510-CD11b (Cat: 101263), Pacific Blue-CD45 (Cat: 103126), BV785-Ly6C (Cat: 128041), BV650-Ly6G (Cat: 127641), APC Cy7-CD11c (Cat: 117324). For intracellular staining, samples were processed according to the BD Cytofix/Cytoperm kit (BD Biosciences) and subsequently stained with PE-NGF (Santa Cruz, Cat: sc-365944) for 30 min. 7-AAD was used as a cell viability marker. Gating strategy for all flow cytometry experiments are depicted in Figure S3 and S4. Data was collected with LSRFortessa X-20 Cell Analyzer (BD Biosciences) and analyzed with FlowJo software (TreeStar).

Calcium Imaging

Dorsal root ganglia were collected from Advillin-GCaMP6 mice and incubated with enzymes (2 mg/mL collagenase and 2.5 mg/mL dispase-II) to dissociate cells for 1 hour at 37C. Flame-polished Pasteur pipettes were used to mechanically agitate the ganglia prior to plating on poly-d-lysine coated glass coverslips. Cells were grown in a neurobasal + 2% B27 supplement medium for 24 hours before testing.

Prior to testing cells were placed in imaging buffer containing 140 mM NaCl, 10 mM D-(+)-Glucose, 1 mM MgCl₂, 2 mM CaCl₂, 5 mM KCl, 10 mM HEPES, pH = 7.4, osmolarity = 320 mOsm/L. Ca²⁺ signals were measured using green emitted light in a 3 s interval. Histamine (Tocris, Cat: 3545; 300μM in imaging buffer), Bradykinin (Tocris, Cat: 3004; 5μM in imaging buffer), or imaging buffer alone was added at interval 40 until interval 80. Capsaicin (Tocris, Cat: 0462; 300nM in imaging buffer; positive control) was added at interval 240 to activate TRPV1⁺ neurons. Baseline average fluorescence intensity was calculated as F₀. Fold change of signal amplitudes at each time point are presented as $\mathrm{\Delta }\mathrm{F}/{\mathrm{F}}_0=({\mathrm{F}}_{\mathrm{t}}-{\mathrm{F}}_0)/{\mathrm{F}}_0$; the ratio of fluorescence difference (${\mathrm{F}}_{\mathrm{t}}-{\mathrm{F}}_0)$ to basal value (${\mathrm{F}}_0$). ${\mathrm{F}}_{\mathrm{t}}$ represents the fluorescence intensity at each time point.

Mediator Instillation and Receptor Inhibition

Histamine (0.0825 mg/kg in 30μL) or Bradykinin (0.00795 mg/kg in 30μL) was instilled in the bladder of isoflurane anesthetized mice for 20 minutes. Catheterization was identical to that used for UPEC infection. Cystometry was performed 1 hour after instillation. Pelvic sensitivity assay was performed 30 min after instillation. For H₁R, mice received an intraperitoneal injection of pyrilamine maleate (Sigma, Cat: P5514; 5 mg/kg) 24 hours and 2 hours before the pelvic sensitivity test. For B₂R, mice received an intraperitoneal injection of icatibant (Tocris, Cat: 3014; 150 nmole/kg) 24 hours and 2 hours before the pelvic sensitivity test.

Statistical Analysis

Pelvic sensitivity assays were performed by an individual blinded to experimental conditions. Results were analyzed using Graphpad Prism ver 10 by either the Mann-Whitney U test, Student’s T-test, or Kruskal-Wallis test as indicated in each figure, each with p<0.05 used to define significance. Number of independent experimental repeats are indicated in each figure. Data depicted as Mean ± SEM, ****p<0.0001, ***p<0.001, **p<0.01, *p<0.05, ns=not significant.

Inclusion and Diversity

One or more of the authors of this paper self-identifies as an underrepresented ethnic minority in science. One or more of the authors of this paper received support from a program designed to increase minority representation in science.

Data and materials availability

The data for this study is available in Data file S1 and upon reasonable request from the corresponding author.