Urothelial bladder afferent neurons in the rat are anatomically and neurochemically distinct from non-urothelial afferents

Abstract

There is mounting evidence underscoring a role for the urothelium in urinary bladder sensation. Previous functional studies have identified bladder primary with mechanosensitive properties suggesting urothelial innervation and/or communication. The current study identifies a group of urothelium-innervating afferent neurons in rat, and characterizes and compares the properties of these and non-urothelial afferent neuron populations. Lumbosacral (LS) primary afferent neurons were retrogradely labeled using intraparenchymal (IPar) microinjection or intravesical (IVes) infusion of tracer into the bladder. Using these techniques, separate populations of neurons were differentiated by dorsal root ganglion (DRG) somata labeling and dye distribution within the bladder. IPar- and IVes-labeled neurons accounted for 85.0% and 14.4% of labeled L6-S1 neurons (P<0.001), respectively, with only 0.6% of neurons labeled by both techniques. Following IVes labeling, dye was contained only within the periurothelial bladder region in contrast to non-urothelial distribution of dye after IPar labeling. Electrophysiological characterization by in situ patch-clamp recordings from whole-mount DRG preparations indicated no significant difference in passive or active membrane properties of IPar and IVes DRG neurons. However, calcium imaging of isolated neurons indicates that a greater proportion of IPar-than IVes-labeled neurons express functional TRPA1 (45.7% versus 25.6%, respectively; P<0.05). This study demonstrates that two anatomically distinct groups of LS bladder afferents can be identified in rat. Further studies of urothelial afferents and the phenotypic differences between non-/urothelial afferents may have important implications for normal and pathophysiological bladder sensory processing.

1. Introduction

The urinary bladder stores and evacuates urine through local, spinal, and supraspinal reflexes that involve coordination of efferent and afferent nervous system activity. Intravesical pressure during bladder filling is encoded by in-series, low threshold mechanosensitive afferents arising from lumbosacral (LS) dorsal root ganglia (DRG) that have receptive fields in the bladder and convey information along the pelvic nerve (PN) [1]. Changes in the activation threshold and encoding properties of these afferents, along with recruitment of activity and/or acquisition of mechanosensitivity by other types of afferents, is thought to contribute to bladder hypersensitivity associated with various pathological conditions (e.g., overactive bladder, interstitial cystitis/painful bladder syndrome; see [2]).

The urothelium, which was once thought to act simply as a protective barrier to sequester urine contents within the bladder lumen, plays an important role in bladder sensation via its transduction and signaling properties and communication with afferent nerves [3, 4]. Techniques that utilize retrograde labeling to study bladder-innervating primary afferent neurons have relied on the assumption that tracers injected into the intraparynchemal space label somata in the DRG that are broadly representative of all bladder-innervating afferents. Recently, a previously undefined subpopulation of bladder afferents that innervates the peri-urothelial bladder tissue (referred to here forward as “urothelial” afferents) was identified and characterized in the mouse [5]. The electrophysiological properties of urothelial afferents differed from those innervating non-urothelial bladder tissues, and were suggestive of increased excitability. The present study sought to determine whether a similar group of anatomically and functionally distinct urothelial bladder afferents could be identified in the rat. Bladder innervating DRG neurons were labeled either by intraparynchemal (IPar) injection or intravesical (IVes) infusion of fluorescent dye into the bladder, and were first differentiated on the basis of local dye distribution and somata labeling in L6-S1 DRG. Then, in situ patch-clamp recordings from IPar and IVes afferent neurons were performed using a whole-mount DRG preparation to examine electrophysiological properties. Finally, changes in intracellular calcium concentration evoked by application of an agonist to the transient receptor potential cation channel, subfamily A, member 1 (TRPA1) were measured. TRPA1 has been shown to play a role in persistent bladder hyperalgesia [6, 7] and mediates changes in sensory transduction in response to reactive metabolites and environmental chemicals [8, 9] often encountered in luminal urine contents.

2. Results

2.1. IVes- and IPar-labeled afferent neurons are distinct anatomical populations

In one group of rats, the number of bladder afferent neurons retrogradely labeled by IPar and/or IVes dye administration was quantified in histological DRG sections (Fig. 1). The route of dye administration had a profound effect on the number of labeled neurons (P < 0.0001 ). When Dil was delivered IVes and FB was delivered IPar (Fig. 1A-C), DiI-positive neurons accounted for 16.6% (27.0 ± 2.6 neurons per rat) of the total labeled neurons (811) in L6 DRG and 12.1% (12.4 ± 3.0 per rat) of the total labeled neurons (513) in S1 DRG, whereas FB-positive neurons accounted for 83.4% (135.2 ± 20.5 per rat) of all labeled neurons in L6 and 87.9% (90.2 ± 20.9 per rat) in S1 (n = 5 rats; P < 0.001; Fig. 1D-F). The overlap between IVes- and IPar-labeled neurons was extremely small, with ≤1% of labeled LS neurons expressing both FB and DiI (0.6 ± 0.2 per rat in L6 and 1.0 ± 0.3 per rat in S1). These data are based on a 2 hour IVes dwell time of DiI, following a parametric analysis in independent groups of rats in which labeled L6 DRG neurons were counted after 1, 2, or 4 hour dwell times. A dwell time of 1 hour labeled significantly fewer neurons than 2 hour (10.7 ± 0.7 versus 29.7 ± 3.8, respectively; P < 0.01 ), and 2 hour did not differ from 4 hour (31.0 ± 3.2) (n = 3 rats/group).

Fig 1. IPar- and IVes-labeled neurons are anatomically distinct subsets of bladder afferent neurons.

IPar injection of FB and IVes infusion of Dil retrogradely labeled distinct populations of L6 and S1 bladder DRG neurons (A-C). FB-labeled IPar neurons are indicated by arrowheads and a DiI-labeled IVes neuron is indicated by asterisks. One week following the labeling procedure, the total number of L6 and S1 neurons labeled by IPar FB was significantly higher than the number labeled by IVes DiI or by both FB and DiI (D). Of a total of (811) labeled L6 DRG neurons, 83.4% were positive for FB, 16.6% were positive for DiI, and 0.4% were double labeled (E). Of a total of (513) labeled S1 DRG neurons, 87.9% were positive for FB, 12.1% were positive for DiI and 1.0% were double labeled (F). **** P < 0.001 versus IVes and Double-labeled; two way ANOVA with Sidak’s multiple comparison tests, n=5 rats. Scale bar represents 25 μm.

Cell size was quantified in histological DRG sections from DiI and FB doublelabeled rats. Cell area was measured with ImageJ and used to calculate diameter. Neurons were binned into two groups that were ≤30 and >30 μm in diameter, based on an a priori size determination suggested by data indicating a correlation between neuronal soma size and peripheral conduction velocity [10]. 32.9% (49 out of 149) of IVes neurons and 67.1% (100 out of 149) of IPar neurons were ≤30 μm, with the remaining proportion in each group being >30 μm (P < 0.0001 ; Chi-square test). The median diameter of IVes neurons was 34.2 μm (with 28.1 μm and 38.7 μm as 25^th and 75^th percentiles, respectively) while the median diameter of IPar neurons was significantly smaller at 28.3 μm (with 25.7 μm and 32.6 μm as 25^th and 75^th percentiles, respectively) (P < 0.0001 ; Mann-Whitney U test).

To determine local distribution of dye after labeling, bladder tissue was harvested at 1 or 7 days following IPar or IVes DiI labeling. Integrated density was quantified in urothelial and non-urothelial tissues, revealing technique-dependent patterns of Dil distribution. After IPar administration, there was uniform distribution of DiI throughout the bladder smooth muscle, with no urothelial labeling, at 1 and 7 days post-labeling n = 4; Fig. 2A,B). In contrast, DiI was restricted to the urothelium following IVes infusion and did not redistribute into deeper tissues over time (n = 4; Fig. 2C,D). Two-way ANOVAs indicated a significant interaction between the labeling technique and distribution of dye at each time point (1 day: F(1,12) = 33.48, P < 0.001; 7 day: F(1,12) = 22.38, P < 0.005), with urothelial > non-urothelial labeling in IVes tissue and non-urothelial > urothelial labeling in IPar tissue (all P values < 0.05; Fig. 2E,F). There were no significant main effects of labeling method or tissue region on integrated density measurements.

Fig 2. IVes and IPar labeling techniques produce differential distributions of fluorescent dye in bladder tissue.

Technique-dependent patterns of DiI distribution were observed in bladder tissue following IVes and IPar labeling. Brightfield (BF) photomicrographs taken at 10× were used to differentiate urothelium (Uro) and non-urothelial (Non-Uro) bladder tissue as seen in (A) and (C) in a blinded fashion. Then, the integrated density of DiI distribution in the same Uro and Non-Uro regions was quantified (B, D). In IVes-labeled bladders, integrated density was significantly greater in urothelial than non-urothelial tissue, and the opposite was true in IPar-labeled bladder tissue. This distribution pattern was observed when tissue was harvested at 1 d (E) or 7 d (F) post infusion. * P < 0.05, ** P < 0.01, **** P < 0.0001 ; two-way ANOVA with Sidak’s multiple comparison tests; n=4 rats/group. Scale bars represent 50 μm.

2.2. Differences in sensitivity to allyl isothiocyanate in vitro

Ratiometric calcium imaging was performed on dissociated, retrogradely labeled DiI-positive IVes neurons (n = 43 from 6 animals) and IPar neurons (n = 46 from 7 animals) from L6-S1 DRG to determine their sensitivity to the TRPA1 agonist, allyl isothiocyanate (AIC). Increases in intracellular calcium concentration, measured as an increase in fluorescence from baseline (ΔF), were detected in a subset of both IPar and IVes DiI labeled neurons following application of 100 μΜ AIC (Fig. 3A). Neurons were considered responsive if they exhibited a F ΔF>1.0 and returned to baseline F following EGTA application. AIC evoked responses in 25.6% of IVes-labeled neurons compared to 45.7% of IPar-labeled neurons (Fig. 3E; P < 0.05), with a similar peak change in ΔF (2.42 ± 0.36 and 2.01 ± 0.19, respectively).

Fig 3. AIC sensitivity in IVes and IPar neurons.

Functional TRPA1 expressi determined by AIC-evoked calcium influx recorded as a change in fluorescence (peak F - F₀) as seen in representative images prior to drug application (A), post 100 μΜ AIC (B), and post 100 μΜ EGTA (C). Example traces show either an increase in F or no change for a responder (red line) and non-responder (black line), respectively. Following AIC application, a significantly higher percentage of IPar neurons (45.7%) exhibited a ΔF>1 compared to IVes neurons (25.6%). * P < 0.05; Chi-square; n=6–7 rats/group.

2.3. Electrophysiological properties of IVes- and IPar-labeled bladder afferents

Retrogradely labeled, DiI-positive Ives neurons (n = 19 from 3 animals) and IP neurons (n = 12 from 3 animals) were identified using differential interference contra and epifluorescence illumination in whole-mount DRG (Fig. 4A). Soma diameter of the neurons studied were 26.1 ± 0.9 μm and 24.0 ± 0.7 μm for IVes and IPar groups, respectively (P > 0.05). To maintain accuracy of electrophysiological parameters after advancing the electrode through layers of DRG tissues in situ, data was collected from neurons with baseline whole-cell currents of ≤50 pA (Vh = −73 mV). Neither IPar nor IVes neurons displayed spontaneous activity and did not differ in resting membrane potential, membrane capacitance, rheobase, or input resistance (Table 1 ; see Fig. 4B for typical responses to current injection). Both types of neurons exhibited action potentials of similar amplitude and width, with no difference in AP threshold (Fig. 4C, Table 1). The majority of neurons in each group fired single APs in response to suprathreshold current injection up to 10x rheobase, with only 16.7% IPar and 5.3% IVes neurons firing multiple action potentials (Fig. 4D).

Fig 4. Action potential properties did not differ between IPar and IVes neurons.

Differential interference contrast (DIC) (A) and epifluorescence (B) were used to identify DiI-positive bladder neurons retrogradely labeled by IPar or IVes methods. As shown by DiI-positive bladder neurons retrogradely labeled by IPar or IVes methods. As shown by representative traces in (C) and (D), there were no distinguishable differences in action potential or hyperpolarization characteristic. The majority (83.3% and 94.7%, respectively) of IPar and IVes neurons fired single action potentials (E). Scale bar represents 10 μm.

Table 1. Passive and active membrane characteristics of IPar and IVes neurons exhibit no significant differences.

Resting membrane potential (RMP), input resistance, cell capacitance, rheobase, action potential (AP) threshold, and AP width did not differ.

	RMP (mV)	Input R (MΩ)	Capacitance (pF)	Rheobase (pA)	AP threshold (mV)	AP width (ms)
IPar (n=12)	−68.6 ±1.4	250.4 ±27.1	33.3 ± 1.1	195.8 ±29.7	−34.5 ±2.0	2.4 ±0.3
IVes (n=19)	−68.8 ±1.3	219.0 ±23.6	31.7 ±1.1	235.8 ±25.8	−34.0 ± 1.7	2.2 ±1.2

Voltage-activated currents were elicited by voltage steps ranging from −103 to +57 mV in increments of 10 mV (250 ms). With no attempt made to isolate specific currents, two main currents were observed in response to the voltage steps: a transient inward current reflecting activation of voltage-gated Na⁺ channels, and an outward current reflecting activation of both inactivating and non-inactivating voltage-gated K⁺ channels. When outward currents were compared by quantifying the peak amplitude of the inactivating current and the amplitude of non-inactivating current at the end of each voltage step, no differences were seen between IPar and IVes neurons (Fig. 5B,C). In addition to the currents observed during voltage steps, both IPar and IVes neurons exhibited M-type voltage-gated K⁺ currents (M-currents). Representative traces from IPar and IVes neurons during deactivation from −20 to −60 mV show overlapping magnitude of M-currents (Fig. 5D) with mean peak amplitudes of 143.4 ± 16.1 pA and 139.2 ± 7.6 pA, respectively (Fig. 5E).

Fig 5. Amplitude of outward currents, including M-current, did not differ between urothelial and non-urothelial PN bladder neurons.

Representative traces of total currents (A) evoked by voltage steps (shown below the traces) from a holding potential of −73 mV revealed an inactivating inward current, an inactivating outward current (white triangles), and a non-inactivating outward current (black triangles). The amplitude of inactivating (B) and non-inactivating (C) outward currents were not different between IPar and IVes neurons. Representative traces of M-currents (D) in response to a −73 to +7 mV voltage step show similar amplitude between IPar and IVes neurons, and a quantitative comparison of peak M-currents showed no difference between groups (E).

3. Discussion

The present report describes basic anatomical, neurochemical, and electrophysiological properties of bladder primary afferent neurons in rat that were identified using intravesical and intraparenchymal retrograde labeling techniques. Recently, Kanda et al. [5] demonstrated that afferents innervating the periurothelial region of the mouse bladder were distinguishable and distinct from those innervating other layers of tissues. Further, this newly identified group comprises a similar proportion of mouse pelvic nerve bladder afferents as those that have previously been functionally identified as responsive to mechanical stimulation of the urothelium [11]. This study was a logical follow-up to determine whether a similar group of sensory afferents could be identified in the rat using intravesical infusion of dye into the bladder lumen (i.e., IVes) versus microinjection of dye into the bladder parenchyma (i.e., IPar). As in mouse, IVes and IPar techniques labeled two anatomically distinct groups of DRG neurons with little overlap (8 out of 1324 L6-S1 neurons from 5 rats). These two groups of neurons differ in somata diameter and in their AIC-sensitive proportions; however, IPar- and IVes-labeled neurons in the rat do not exhibit unique electrophysiological characteristics. There are examples in the literature of differences in sensory systems between mouse and rat, and this is perhaps unsurprising given the evolutionary divergence of the species and demonstrated transcriptomic variation that likely underlies molecular and behavioral differences between them. However, without knowing for example the functional phenotypes of subclasses of human bladder-innervating sensory afferent neurons, it is difficult to know which species exhibits the most translationally relevant phenotypic properties in this regard. Some comparative studies of urinary bladder or lower urinary tract phenotype have shown convergence of data in mouse and rat that differs from human (e.g., [12, 13]), while others have shown inter-species differences between mouse and rat where the physiology of one or the other is more similar to human (e.g., [14, 15]). A recent study by Ray et al. (2017; bioRxiv) describes comparative transcriptome profiling of human and mouse dorsal root ganglion with a focus on sensory-related genes and, although not organ or tissue-specific, such studies will aid in the advancement of translational sensory research.

Very little overlap was observed in FB-labeled IPar and DiI-labeled IVes neurons (0.8% of L6 and 1.0% of S1 neurons). In fitting with previous histological studies that have shown peripheral sensory terminal endings in proximity to the urothelium [16–18], quantification of fluorescent integrated density in IVes-labeled bladder tissue suggests that 12–16% of IVes-labeled neurons in rat may give rise to those fibers. Teased fiber recordings in mouse revealed that 9% of pelvic nerve afferents had “urothelial” type mechanical response profiles and 63% had “muscular” response profiles [11]. This is proportionally similar to IVes- and IPar-labeled neurons, respectively, in the present study, with potential overlap in each group in those previously functionally identified as “muscular/urothelial” (14%) [11].

Afferent neurons with terminal endings in proximity to the urothelium express ligand-gated receptors for a variety of urothelial-derived mediators that alter afferent signaling (see [19]). Although the precise role of IVes-labeled, urothelial afferent neurons in bladder sensation and function is unknown at this time, it is highly likely that they play an important role in chronic voiding dysfunction and pain associated with interstitial cystitis/painful bladder syndrome (IC/PBS). The etiology of IC/PBS appears to be multifaceted, including neural, immune, and/or psychological factors, however, a consistent phenotypic observation in IC/PBS patients is that of increased permeability of the urothelium [20], a stratified transitional epithelium lining the urinary bladder. Increased urothelial permeability permits urine contents such as ions and metabolic byproducts to interact with sensory afferent endings proximal to the urothelium, i.e., those we have identified using IVes labeling techniques. This notion is strongly supported by a preclinical study in which increased urothelial permeability induced by upregulation of claudin 2, a disrupter of tight junctions which has been shown to be upregulated in bladder pain syndrome biopsies [21], results in the development of spontaneous action potential firing in afferent neurons identified using intravesical delivery of adenovirus coding for GFP [22].

Although not tested in the present study, it was previously reported in mouse that specificity of urothelial afferent labeling by IVes procedures was obtained with DiI, but not with FB [5]. As dye-dependent labeling specificity is likely due to differences in uptake and transport mechanisms, we assume a similar discrepancy in dye specificity in rat. Dil is a lipophilic dye that acts primarily via passive diffusion along membrane surfaces, while FB is more readily taken up by intact terminals [23], suggesting active transport. In the previous and current studies, passage of DiI across the urothelial barrier was likely facilitated by the use of DMSO (10%) in the dye solution, as DMSO can increase lateral distance between lipid heads and facilitate the entrance of water [24]. Other compounds with high lipophilicity, such as paclitaxel, have been shown to penetrate the urothelium to a greater degree than drugs with low lipophilicity when used intravesically to treat bladder cancer [25]. Moreover, that little to no DiI was observed below the urothelium when delivered intravesically may be due to its drainage into capillaries as it diffuses toward the basal urothelium, as has been indicated for other lipophilic compounds [26].

Bladder afferent neurons have small myelinated (Αδ-fiber) or unmyelinated (C-fiber) axons and consist of mechanosensitive, thermosensitive, and chemosensitive populations [27]. These neurons and their axons express various receptors important for sensory transduction, including transient receptor potential ankyrin 1 (TRPA1). TRPA1 is a ligand-gated ion channel that is localized predominantly in small diameter (≤35 μm) neurons of sensory ganglia [28], including in bladder sensory neurons [6, 7, 29–31]. In the present study, it was histologically determined that IPar neurons had a significantly smaller overall mean diameter than IVes neurons, and also included a significantly larger proportion of neurons ≤30 μm. Notably, when an identical contingency analysis was performed using ≤35 μm as the categorical cutoff for binning, comparative results were still profoundly significant (P < 0.0001). This is in contrast to in situ measurement of IVes and IPar neuron size which indicated no difference between the two populations. This discrepancy could be due to a sampling bias, in which only labeled neurons in the outer two to three cell layers were used for electrophysiological assessment due to the in situ preparation. Previous conduction velocity studies and supportive immunohistochemical data have demonstrated that approximately two thirds of pelvic nerve (lumbosacral) bladder afferents are C-fiber neurons and one third are Aδ-fiber neurons [32–34]. Based on a limited sampling of rat DRG neurons, a soma diameter of ≤30 μm has been used as a correlate to identify C-fiber, and some Aδ-fiber, neurons [10]; thus one might extrapolate that the IPar group of neurons is comprised of a larger proportion of unmyelinated, C-fiber afferents. Moreover, a greater percentage of AIC-responsive afferents was observed in IPar (45%) than in IVes (25%) neurons, in fitting with the notion that TRPA1 is expressed predominately in small diameter neurons. This is in contrast to what might have been anticipated given what is known about detection and transduction of chemosensory information by TRPA1 and the (albeit indirectly) demonstrated proximity of IVes afferent terminals to the bladder lumen. Alternatively, reduced chemosensitivity in urothelial afferents may be an important evolutionary adaptation to food derived pungent molecules that are cleared from the body by renal excretion, and that otherwise may lead to bladder hyperexcitability. In either case, TRPA1-expressing urothelial afferents likely play a role in both infectious and non-infectious cystitis. TRPA1 has been shown to directly mediate cutaneous neurogenic inflammation and nociceptive behavioral responses produced by bacterial endotoxins, such as lipopolysaccharide, in a toll-like receptor 4-independent fashion [35]. As such, alterations in the sensitivity of urothelial afferents by gramnegative bacteria introduced to the lower urinary tract may increase bladder sensation and function associated with bladder infection. In patients with non-infectious cystitis (e.g., classic IC), mucosal biopsies exhibit increased expression of TRPA1 [36] which could be driven by upregulation in either urothelial afferents or urothelial cells.

The use of IVes and IPar labeling techniques together can lead to new insights for understanding pathological changes in bladder sensation and function following inflammation or organ injury. More specifically, the ability to identify and study urothelial bladder afferent neurons will aid in deciphering mechanisms related to urothelial-afferent signaling, and may be of particular use in understanding sensory changes in model systems in which urothelial dysfunction appears to be a pathophysiological component of symptom development (e.g., interstitial cystitis/painful bladder syndrome (IC/PBS), diabetic bladder dysfunction).

4. Experimental Procedures

4.1. Animals

Experiments were performed on female Sprague Dawley rats (Invigo), 200–250 g in weight that were randomly assigned to groups. Males were excluded because of difficulty in transurethral bladder cannulation without surgical incision. Rats were housed with ad libitum food and water access in solid-bottom plastic cages with wire bar and micro-isolator lids with commercially available wood chip bedding in the University of Alabama at Birmingham Animal Resources Program. All procedures were approved by the UAB Institutional Animal Care and Use Committee (IACUC) and met the standards for humane animal care and use set forth by the Animal Welfare Act and the NIH Guide for the Care and Use of Laboratory Animals. Data were collected and analyzed in a blinded fashion.

4.2. Reagents

Fast blue (FB; Polysciences Inc.) was used as a 1% w/v solution in sterile water. 1,1’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine perchlorate (DiI ; Molecular Probes) was prepared as a 10 mg/ml solution in DMSO and diluted 1:10 in sterile saline. A stock solution of allyl isothiocyanate (AIC, 100 mM; Pestanal®; Sigma-Aldrich) was prepared in 1-methyl-2-pyrrolidinone and diluted in HBSS. Final AIC concentration after addition to the recording dish during Ca²⁺ imaging was 100 μΜ.

4.3. Retrograde labeling

Bladder primary afferent neurons in were retrogradely labeled with fluorescent dyes by intraparynchemal (IPar) injection into the bladder wall or intravesical (IVes) infusion in to the bladder lumen as previously described [5]. Rats were anesthetized with isoflurane in oxygen (5% induction, 1.5–2% maintenance). For IPar labeling, a laparotomy was performed to expose the bladder and multiple injections of DiI or FB were made into the bladder wall using a microsyringe (total volume 15 μl). The abdominal muscle and skin layers were sutured separately. For IVes labeling, bladders were filled with 500 μl of saline for 5 minutes via a transurethrally placed 22-gauge angiocatheter. Bladders were then fully expressed and 500 μl of DiI solution was infused into the bladder and left to dwell for 2 hours (except where otherwise specified). For double-labeled animals, IPar FB labeling was always performed 24 hours prior to IVes DiI labeling. All experiments were carried out one week following labeling except where noted for bladder histology.

4.4. Histology

Rats were deeply anesthetized with isoflurane (5% in oxygen) and a cardiac exsanguination was performed. Bladders and L6-S1 DRG were dissected and drop fixed in 4% paraformaldehyde overnight, then cryoprotected in 30% sucrose for at least 24 hours. Both right and left ganglia were embedded in one cryoblock, then sectioned at 10 μm using a cryostat. Sections were slide mounted and coverslipped with 50% glycerol in PBS. Tissue sections for all analyses were viewed on a Nikon Ti80E microscope (Nikon Instruments, Inc.) and imaging was performed using a 10× dry objective. For parametric analysis of 1, 2, and 4 hour IVes DiI labeling, tissue sections were stained for 5 minutes with a 1 μl/ml bisbenzimide solution prior to coverslipping. Only neurons with a visibly stained nucleus were counted. Cells with a nucleus in the plane of the photomicrograph were analyzed using the freehand tool to trace around the perimeter, then the Analyze plug-in was used to determine area for that cell in μm². Area was converted to diameter for analysis. To determine localization of dye within bladder tissue, bladders were harvested from mice at 1 or 7 days following IVes labeling and 7 days following IPar labeling. A grid of 7000 pixels² area per point was overlaid on the brightfield images of four non-sequential histological bladder sections per rat using the Analyze plug-in. Five grid boxes of only urothelial tissue and five of non-urothelial tissue were identified and marked by an experimenter blinded to labeling method. The same ten boxes were then identified on the fluorescence images of the same tissue sections, and contrast was adjusted (low end 34, high end 220) to remove background. Integrated density was measured and normalized to pixel area. The five urothelial values and the five non-urothelial and values were averaged to obtain a single urothelial and a single non-urothelial measure for each rat. Data are shown as group mean integrated density/pixel area with the minimum and maximum value range.

4.5. Cell dissociation

Bladder afferent neurons were retrogradely labeled with DiI using IPar or IVes techniques as described above. One week later, rats were deeply anesthestized with inhaled isoflurane (5% in oxygen) and transcardially perfused with cold Ca²⁺/Mg²⁺-free Hanks’ Balanced Salt Solution (HBSS; Gibco) over 3 minutes to remove blood and improve cell survival. A laminectomy was performed and bilateral L6-S1 DRG were quickly removed and transferred into cold HBSS for dissociation as previously described [37]. Briefly, DRG were enzymatically treated with cysteine, papain, collagenase type II, and dispase type II to facilitate mechanical dissociation. Isolated neurons were plated on poly-d-lysine/laminin-coated microwells of a glass bottom culture dish in Dulbecco’s modified Eagle medium F12 (DMEM/F12; Gibco) containing 10% fetal bovine serum and penicillin/streptomycin (50 U/ml). After 2 hours at 37°C, cells were flooded with 2 ml of medium and used within 6 hours.

4.6. Ca²⁺ imaging

Ca²⁺ imaging was performed on retrogradely labeled IPar and IVes bladder afferent neurons to examine changes in intracellular calcium concentration in response to application of AIC, a TRPA1 receptor agonist. Neurons were incubated for 20 minutes at 37°C in the Ca²⁺ indicator Fura-2 AM ester (3 μM; TEFLabs) with 0.05% Pluronic F-127 (Invitrogen) in HBSS containing 5 mg/ml bovine serum albumin (Fisher Scientific). The Fura-2 AM solution was replaced with 2 ml of fresh HBBS without Fura-2 AM for an additional 20 minutes at 37°C. DiI-labeled cells were identified as regions of interest using a Nikon Ti80e microscope and 40× immersion objective. A ratio of fluorescence emission at 510 nm in response to excitations at 340 and 380 nm was acquired at 1 Hz (Lambda LS and 10-B Smart Shutter, Sutter Instruments) via CoolSnap HQ camera (Photometrics) and saved to a computer using Nikon Elements software. Baseline fluorescence (F₀) was measured for one minute prior to addition of AIC (final concentration 100 μM) followed by EGTA (final concentration 100 μM) to the chamber. Evoked increases in intracellular Ca²⁺ concentration were measured by calculating ΔF (peak fluorescence minus F₀). Cells that exhibited a ΔF>1.0 within 3 minutes of AIC application were considered responders. A percentage of responders was calculated for each cell preparation, and data are presented as the mean of these percentages.

4.7. Patch-clamp electrophysiology

One week after IPar injection or IVes infusion of DiI, mice were deeply anesthetized with isoflurane (5% in oxygen) and cardiac exsanguination was performed. Bilateral L6 DRG were rapidly dissected into cold Leibovitz-15 medium (Mediatech, Inc.) and surface connective tissues were carefully removed. Whole DRG were anchored in a recording chamber mounted on the stage of an Olympus IX50 microscope and submerged in a Krebs solution containing (in mM): 117 NaCl, 3.5 KCl, 2.5 CaCl₂, 1.2 MgCl₂, 1.2 NaH₂PO₄, 25 NaHCO₃, and 11 glucose, with pH 7.3 and 325 mOsm and saturated with 95% O₂ and 5% CO₂. DRG were exposed to Krebs containing 0.05% dispase II and 0.05% collagenase type II for 3–5 minutes, washed in enzyme-free Krebs, then continuously perfused (2 ml/minute) with normal Krebs solution at room temperature. Retrogradely labeled neurons were first identified under epifluorescence illumination and then viewed using an infrared differential intensity contrast optical system. Whole-cell patch-clamp recordings were performed on randomly selected labeled neurons by an experimenter blinded to labeling condition. Electrodes for patch-clamp recordings were fabricated from thin-wall capillaries using a Flaming/Brown P-97 puller (Sutter Instrument Co.). Recording electrodes had resistances of 6 MΩ when filled with internal solution containing (in mM) 105 K-Gluconate, 35 KCl, 2.4 MgCl₂, 0.5 CaCl2, 5 EGTA, 10.0 HEPES, 5.0 Na₂ATP, 0.33 GTP-Tris salt; pH 7.35 (adjusted with Tris-base) and 320 mOsm (adjusted with sucrose). Signals were amplified using an Axopatch 200B amplifier (Axon Instruments) with a low-pass filter set at 2 kHz and digitized at 10 kHz. Data were acquired using pCLAMP10 software (Axon Instruments). Junction potential between bath and electrode solution was calculated to be −13 mV and was corrected for in the data analysis.

After gaining whole-cell access and in voltage-clamp mode, membrane capacitance and input resistance were measured at the beginning of the recording using a seal test with voltage pulses of 5 mV. Unless otherwise indicated, neurons were held at −73 mV (voltage command at −60 mV). To record voltage-activated currents, whole-cell currents were evoked by a series of voltage steps ranging from −103 to +57 mV in 10 mV increments of 250 ms. Isolation of Na⁺ or K⁺ currents was not performed in this study. To examine M-currents, deactivating tail currents were measured following the voltage step of −20 mV as described previously [38].

Resting membrane potential and spontaneous activity were determined upon switching to current-clamp mode. To determine membrane excitability, step current pulses from −100 to +900 pA (20 pA per step, 250 ms duration) were injected into cells through patch-clamp electrodes. Neuronal excitability was characterized by action potential (AP) threshold, rheobase, number of APs evoked in response to depolarizing current injection, and spontaneous activity [39]. Rheobase was defined as the least amount of current required to evoke an AP and AP threshold was defined as the greatest depolarization reached before AP generation. Other passive and active electrophysiological properties were assessed including cell capacitance, input resistance, and AP amplitude, width, and shape.

4.8. Data analysis

All data are represented as mean ± SEM unless noted. Graphpad (Prism) was used to compare all non-electrophysiological data. Two-way ANOVA followed by Sidak’s multiple comparisons were used to analyze and compare the number of IPar- and IVes -labeled neurons in double-labeling experiments, and the integrated density of fluorescence in bladder tissue. The number of IVes neurons labeled by dye infusions of varying duration was analyzed by one-way ANOVA followed by Student’s t-tests. Calcium imaging data were analyzed using a Chi-square test. Electrophysiological data from IPar- and IVes-labeled neurons were analyzed using pCLAMP software and Clampfit 10 module (Molecular Devices) and were compared using Student’s t-tests. For all analyses, P values <0.05 were considered significant.

Highlights.

• Urothelium-innervating sensory afferent neurons were identified in the rat.

• Urothelial afferents comprised 14.4% of urinary bladder L6-S1 DRG neurons.

• Non-urothelial afferents comprised 85% of bladder-innervating L6-S1 neurons.

• Afferent populations did not differ in their electrophysiological parameters.

• Fewer urothelial than non-urothelial afferents expressed functional TRPA1.