TRPV1 enhances the afferent response to P2X receptor activation in the mouse urinary bladder

Both TRPV1 and P2X receptors present on bladder sensory nerve fibres have been implicated in mechanosensation during bladder filling. The aim of this study was to determine possible interactions between these receptors in modulating afferent nerve activity. In wildtype (TRPV1+/+) and TRPV1 knockout (TRPV1−/−) mice, bladder afferent nerve activity, intravesical pressure, and luminal ATP and acetylcholine levels were determined and also intracellular calcium responses of dissociated pelvic DRG neurones and primary mouse urothelial cells (PMUCs). Bladder afferent nerve responses to the purinergic agonist αβMethylene-ATP were depressed in TRPV1−/− mice (p ≤ 0.001) and also in TRPV1+/+ mice treated with the TRPV1-antagonist capsazepine (10 µM; p ≤ 0.001). These effects were independent of changes in bladder compliance or contractility. Responses of DRG neuron to αβMethylene-ATP (30 µM) were unchanged in the TRPV1−/− mice, but the proportion of responsive neurones was reduced (p ≤ 0.01). Although the TRPV1 agonist capsaicin (1 µM) did not evoke intracellular responses in PMUCs from TRPV1+/+ mice, luminal ATP levels were reduced in the TRPV1−/− mice (p ≤ 0.001) compared to wildtype. TRPV1 modulates P2X mediated afferent responses and provides a mechanistic basis for the decrease in sensory symptoms observed following resiniferatoxin and capsaicin treatment for lower urinary tract symptoms.

play a role in human urinary bladder physiology and pathophysiology, as P2X receptors are expressed in bladder smooth muscle 7 and have been shown to mediate a non-adrenergic non-cholinergic component of contraction from patients with idiopathic detrusor instability 8 . P2X receptors have also been identified in myofibroblasts, urothelium, and sub-urothelial nerve fibres of human bladders [9][10][11] .
The transient receptor potential (TRP)V1 receptor is a member of the TRP superfamily, activated by noxious heat ≥ 43 °C, pH ≤ 6.0, and the vanilloids capsaicin and resiniferatoxin. TRPV1 is mostly confined to unmyelinated, small diameter primary afferent fibres and has been consistently implicated in nociception, and inflammatory pain 12,13 . In the urinary tract, TRPV1 is predominantly expressed on primary sensory afferent nerve fibres, whilst contradictory evidence both support and negate a functional role for TRPV1 in the urothelium [14][15][16][17][18][19][20] .
There is evidence to suggest that the TRPV1 receptor contributes to afferent nerve responses during bladder filling. TRPV1 knockout (TRPV1 −/− ) mice have reduced afferent responses to bladder distension, reduced reflex bladder activity and increased bladder capacity 21,22 . Moreover, application of capsaicin or RTX causes activation of bladder sensory nerves and increased reflex activity 21,23 , while systemic administration of a TRPV1 antagonist increases micturition threshold volume and decreases bladder contraction amplitude 24 . Previous studies have also shown that in experimental models of cystitis, genetic deletion or pharmacological blockade of TRPV1 prevents bladder hyperactivity [25][26][27] . However, TRPV1 is not considered to be a mechanically gated ion channel and thus the mechanism by which it is able to alter bladder sensitivity to filling remains unclear.
Previous clinical studies have shown that P2X3 immunoreactivity is significantly reduced in patients who respond positively to intravesical RTX therapy suggesting TRPV1 and P2X3 receptor expression may be coupled 28 . Moreover, in cultures of human urothelial cells TRPV1 was identified at both mRNA and protein level and stimulation led to ATP release 20 , consistent with earlier studies in rodents 22,29 . Such studies implicate the urothelium in bladder sensory signalling, which we expanded upon here, in order to better understand the role of TRPV1 in modulating purinergic evoked bladder afferent activity in the mouse.

Results
Afferent nerve activity. Addition of αβMe-ATP (300 nM-100 µM) produced concentration-related increases in afferent nerve activity and intravesical pressure, indicating detrusor contraction ( Fig. 1a,b,c). The afferent response was characterised by an initial rapid component followed by a period of more sustained firing before returning to baseline over 2-3 minutes. Application of αβMe-ATP (30 µM) initiated a pronounced detrusor contraction (Fig. 1d). The maximal firing and time course profiles in response to repeated application of αβMe-ATP were reproducible (Fig. 1e,f) (p ≥ 0.05, n = 9).
The afferent nerve response to αβMe-ATP occurred concurrent with a moderate contraction of the detrusor smooth muscle. It is possible therefore that the changes in afferent nerve firing in the TRPV1 −/− mice was secondary to altered detrusor muscle activity. However, this would appear not to be the case since there was no significant difference in αβMe-ATP induced bladder contraction between preparations from TRPV1 +/+ and TRPV1 −/− mice despite markedly different afferent firing profiles. Furthermore capsazepine (10 µM) failed to alter contractile responses to αβMe-ATP (p ≥ 0.05, n = 4) (Fig. 3b). Bladder compliance was assessed by the pressure/ volume relationship during distention. Compliance was also unaffected by the loss of functional TRPV1 receptors (Fig. 3c). To further examine the potential link between afferent firing and contractile activity the response to αβMe-ATP was compared under calcium free conditions. Removal of calcium completely abolished αβMe-ATP induced contraction (Fig. 3f). However, a robust afferent response to αβMe-ATP was maintained in calcium free conditions (Fig. 3d,e,f). Moreover, application of capsazepine still attenuated the afferent and peak afferent response to αβMe-ATP in calcium-free buffer (Fig. 3d,e,f; 86.6 ± 16.3imp −s Vs 53.6 ± 10.8imp −s , (p < 0.05, n = 7).
Responses in the DRG. A proportion of DRG neurons responded to αβMe-ATP with an increase in intracellular calcium (Fig. 4a). A subset of DRG neurons from WT mice responded to capsaicin (around 60% Fig. 4b), however, as expected, DRG neurons from TRPV1 −/− failed to exhibit responses to capsaicin. There was no significant difference in maximal αβMe-ATP (30 µM) induced calcium rise in DRG neurons (24.4 ± 2.4% Vs 28.1 ± 3.2%) from TRPV1 +/+ and TRPV1 −/− mice respectively (p ≥ 0.05, N = 12, n = 80) (Fig. 4c). However, the percentage of DRG neurons responding to αβMe-ATP was significantly lower in the TRPV1 −/− mice compared to TRPV1 +/+ ( fig. 4d)  Primary mouse urothelial cells. In order to examine how the urothelium might contribute to purinergic-TRPV1 signalling, experiments were performed with PMUCs and compared with isolated pelvic DRG neurons. mRNA for the TRPV1 receptor was expressed in both PMUCs and DRGs (Fig. 5a). However, the levels of expression were significantly greater in DRG neurons than PMUCs (p ≤ 0.01, n = 3). PMUCs responded to the purinergic agonist ATP with an increase in intracellular calcium in a concentration-dependent manner Mediator release. ATP was released from the urothelium both at rest and in response to distension (Fig. 5d).

Discussion
A number of previous studies have shown that P2X3 sensitive afferent nerves are essential for normal bladder function and that purinergic responses in sensory neurons are largely mediated by P2X2, P2X3 and P2X2/3 receptors 5,6 . Moreover, P2X receptors have been shown to co-localise on afferents expressing TRPV1 30,31 , and co-immunoprecipitation studies have demonstrated a physical association of TRPV1 and P2X3 receptors in the DRG 32 . P2X3 immunoreactivity on bladder projecting neurons is significantly increased in patients with neurogenic detrusor overactivity 33 and is significantly reduced in those patients who responded positively to intravesical application of the TRPV1 agonist, resiniferatoxin, which causes receptor desensitisation, further suggesting there may also be a functional association between the TRPV1 and P2X3 receptors on bladder afferent neurons innervating the bladder 33 . The results presented in this study provide the first direct evidence that the TRPV1 receptor is able to influence bladder afferent responses to P2X stimulation, a process considered essential to mechanosensation during bladder filling.
αβMe-ATP is a non-selective agonist which activates P2X1 and p2X3 receptor subtypes. Previous studies have provided compelling evidence that P2X1 receptors located on the smooth muscle mediate bladder contraction 34 and P2X3 receptors located on the afferent nerves play a role in mechanosensation 4,6 . Bladders from wildtype mice responded to the application of αβMe-ATP with a transient bladder contraction and a profound afferent response. In calcium free Krebs, bladder contractions were lost but nerve firing persisted, albeit diminished. This suggests that the receptors responsible for these two components are distinct, and, given the body of evidence in the literature is seems likely that P2X1 receptors on the muscle mediate the contractile response while P2X3 receptors 34 at the nerve terminal drive the direct nerve response. Membrane bound ionotropic P2X1 and P2X3 receptors, when bound by ATP, mediate non-selective cation influx of Na + , and Ca 2+ , and the efflux of K + across the cell membrane under physiological ionic distribution to initiate cell depolarisation 35 . The degree of afferent firing that is maintained in Ca 2+ free buffer in our experiments is likely a consequence of direct Na + induced membrane depolarisation. Whilst P2X1 receptor mediated influx of both Na + and Ca 2+ influx in smooth muscle cells creates an excitatory junction potential, it is the subsequent opening of voltage gated calcium channels and calcium induced calcium release from intracellular stores which create the concentrations of intracellular calcium necessary for smooth muscle contraction 36 . Afferent responses to αβMe-ATP were also tested in the presence of the L-type calcium channel blocker nifedipine. Nifedipine significantly attenuated bladder muscle contraction with no effect on afferent firing (unpublished data), further supporting differential dependence of voltage gated calcium channels for muscle contraction and afferent firing in response to P2X receptor activation. Loss of TRPV1 reduced the bladder afferent response to αβMe-ATP application. It is possible that P2XR expression on bladder afferent endings is altered in the TRPV1 −/− , thereby affecting the sensitivity to αβMe-ATP. However, the effect of TRPV1 deletion was replicated by the acute treatment with capsazepine, indicating that the attenuated response to αβMe-ATP is unlikely due to a reduction in P2X receptor expression. The similar attenuation observed in the TRPV1 −/− and capsazepine treated wildtype mice is also consistent with capsazepine's pharmacological antagonism of the TRPV1 receptor rather than as a consequence of any non-specific or smooth muscle actions. Interestingly, knockout or pharmacological blockade of the TRPV1 receptor had no effect on detrusor muscle contraction stimulated by αβMe-ATP suggesting that the interaction was at the level of the afferent nerve or urothelium rather than secondary to changes in muscle tone. The effect of capsazepine on afferent firing persists after contractile responses are attenuated in calcium free conditions, again ruling out any effect on bladder contractility. These results are supported by a previous study from our laboratory showing that mechanosensitive responses to ramp distension are significantly attenuated by removal or block of TRPV1 with no concurrent effect on bladder compliance during ramp distension 21 .
TRP channels have been recently shown to modulate the sensitivity of other ion channels 37 through extensive phosphorylation of intracellular signalling molecules, which in the case of TRPV1 and P2X3/P2X2/3 have been shown to be regulated by the phosphoinositides PiP 2 and PiP 3 in peripheral neurons 38,39 . Thus it is possible that removal or blockade of the TRPV1 receptor reduces the sensitivity of P2X3 receptors by alterations to its intracellular signalling cascade.
A previous study into the role of TRPV1 in intestinal mechanosensitivity suggested that loss of TRPV1 function leads to an overall reduction in the pre-tuning of the afferents, thereby reducing excitability 40 . In order to test whether this would explain the phenomenon observed in our study we examined the effect of TRPV1 knockout on the bladder afferent responses to a nicotinic agonist, DMPP, as nicotinic agonists have previously been shown to stimulate capsaicin sensitive bladder afferents 41 . The rationale behind this was to stimulate the bladder by a non-purinergic mechanism in order to assess whether the loss of TRPV1 caused a reduction in nerve sensitivity to both agonists or if the response was specific for αβMeATP. We found no significant difference in the response to DMPP between the wild-type and knock-out mice suggesting that loss of TRPV1 did not result in an overall loss of bladder sensitivity, but that inhibited P2X receptor responses in the TRPV1 knockout is due to a specific interaction between the TRPV1 and P2X3 receptors.

Dorsal root ganglia.
To investigate this relationship in the absence of other potentially confounding factor in the bladder wall, responses to αβMeATP were determined in primary cultures of DRG neurons from lumbosacral (L6/S2) and thoracolumbar (T11/L2) regions from TRPV1 +/+ and TRPV1 −/− mice. DRG neurons from TRPV1 +/+ responded to αβMe-ATP with an increase in intracellular calcium with similar proportions to those reported by others 42 . We found that the magnitude of the calcium signal was not different between neurons from TRPV1 +/+ and TRPV1 −/− mice. The opening of P2X receptors following αβMe-ATP binding permits the influx of Na + , as well as Ca 2+ ions. Membrane depolarization of sensory neurons is dependent on Na + influx, and results in the opening of voltage-gated calcium channels to act as a second messenger of electric signalling 36 . Sensory neuronal signalling through calcium influx is further amplified by calcium-induced calcium release from intracellular stores which is much larger than the initial calcium influx through P2X receptor ion channels. Thus, it is likely that any difference in P2X-receptor excitability between TRPV1 −/− and TRPV1 +/+ mice is masked by the much larger calcium signals associated with the opening of calcium channels. The proportion of DRG neurons responding to αβMe-ATP was significantly reduced in the TRPV1 −/− neurons compared to the wildtypes. In a previous study, Kiyatkin et al. found that 30% of DRG neurons exhibit functional responses to both αβMe-ATP and capsaicin confirming that a subpopulation of neurons in the DRG co-express functional TRPV1 and P2X3 receptors 43 . We found the number of responsive DRG neurons in the TRPV1 −/− mice was reduced by about 25%, which likely represents the proportion of co-expressing cells in the wild type animals whose absence in the knockout animals would account for the attenuated ex-vivo afferent firing in response to αβMe-ATP. Since a similar attenuation was seen following the acute (30 min) treatment with capsazepine it would imply that P2X receptor expression is unchanged in the knockout but sensitivity to ligand is blunted. Together this data suggests that TRPV1 interacts with the P2X3 receptor to modify its sensitivity to agonist stimulation rather than the loss of TRPV1 alters P2X3 receptor expression and further experiments are planned to determine if this is indeed the case. The exact mechanisms involved in this interaction and down-stream signalling pathway still remain to be determined. The urothelium. The bladder urothelium plays a key role in modulating the excitability of bladder afferent nerves. This is achieved by the non-neuronal release of inhibitory (such as NO) and excitatory (such as ATP) mediators. In this study, we examined whether the absence of TRPV1 also influenced the ability of the urothelium to respond to purinergic receptor stimulation or alter transmitter release. In concordance with a number of previous studies, we found no evidence that capsaicin caused a calcium response from urothelial cells 14,15,18 suggesting that TRPV1 may not be functionally coupled to calcium entry or release from intracellular stores in the mouse urothelium, at least under baseline conditions. Paradoxically, however, ATP release from the bladder was attenuated in the TRPV1 −/− mouse, an effect which has previously been reported using isolated rat urothelial cells 22 .
As such, the role of TRPV1 in the urothelium remains controversial, and the detection of a functional TRPV1 receptor from a range of species, including human, appears to be dependent on the specific experimental procedures used, and perhaps an over reliance on immunocytochemistry in the absence of confirmatory assays 14-20 . In the present study TRPV1 mRNA was identified in the urothelium, albeit in low levels, suggesting that TRPV1 might be expressed in urothelial cells despite the failure of capsaicin to evoke a response. This suggests that the levels of TRPV1 mRNA translation may be insufficient to elicit a calcium response, yet high enough to affect ATP release. Immunohistochemical localisation of TRPV1 in the urothelium has been questioned, citing a lack of knockout studies and relevant controls 44 . Genetic deletion of TRPV1 can lead to compensatory changes that attenuate ATP release. Birder et al. (2002) showed that capacitance of the urothelium was blunted in the TRPV1 −/− mouse suggesting that urothelial trafficking mechanisms might be disrupted following deletion of TRPV1 22 . Importantly, however there was no difference in the sensitivity of PMUCs to ATP as indicated by similar pEC50s suggesting that the interaction between TRPV1 and purinergic receptors observed in the afferent recordings are unlikely to be due to a mechanism involving the urothelium.
In contrast to ATP, acetylcholine release was unaffected by the loss of TRPV1. This was surprising since several previous studies have shown that ATP mediates a positive feedback mechanism on ACh outflow from the urothelium 45,46 . The fact that ATP release was changed by the loss of TRPV1 but that ACh release was not suggests that this feedback mechanism was uncoupled. This may be as a direct result of the differential release mechanism for the two mediators (ie ATP is released vesicularly in a graded manner, whereas ACh is released independent of vesical trafficking in an all-or-nothing manner) 47 or a developmental compensation aimed at maintaining control over micturition. However, until further studies are conducted to understand how these two transmitters are coupled interpreting this data can only be speculative.
Concluding remarks. The TRPV1 receptor may have two synergistic roles in mediating bladder mechanosensitivity. A modulatory role in regulating the release of ATP from the urothelium, coupled with a decrease in afferent nerve sensitivity to ATP could combine to result in the deficit in mechanosensitivity observed during bladder distension 21 . Importantly it has been shown that the relative density of suburothelial P2X3 and TRPV1 immunoreactive nerves increases in disease 28,33 and is significantly reduced following successful treatment with intravesical instillation of resiniferatoxin or capsaicin to cause receptor desensitization. This pathway, therefore, may contribute a larger proportion of the mechanosensory response to distension during pathophysiological conditions. The present results provide a mechanistic basis for the decrease in sensory symptoms seen following resiniferatoxin and capsaicin treatment for lower urinary tract symptoms 48,49 , although the relative contribution of the urothelium and the sub-urothelial afferent nerves remains unknown.

Animals. All experiments were performed in accordance with the University of Sheffield's Animal Care
Committee under a project license issued by the UK Animals (Scientific Procedures) Act 1986. TRPV1 wildytpe (TRPV1 +/+ ) and TRPV1 Knockout (TRPV1 −/− ) mice with a genetic background of C57/BL6 were obtained from GlaxoSmithKline (Harlow, UK). Trans-membrane domains 2-4 of the mouse VR1 gene (i.e. DNA encoding amino acids 460-555) were replaced by the neo gene 50 . Mating pairs of TRPV1 −/− and TRPV1 +/+ N1F1 littermates were used to generate separate colonies of TRPV1 +/+ and TRPV1 −/− mice at the University of Sheffield. In the course of the study, genotyping was periodically performed to confirm the absence of the TRPV1 gene in the TRPV1 −/− mice. There were no overt differences in feeding behaviour, litter size, growth rate and body weight between TRPV1 −/− and TRPV1 +/+ groups. [16][17][18] week-old male TRPV1 −/− and TRPV1 +/+ animals were used in this study. Maintenance and killing of the animals followed principles of good laboratory practice in compliance with UK national laws and regulations. The mice were killed humanely by cervical dislocation. Afferent nerve recording. Nerve recording was conducted using an in vitro model previously described 12 .

Isolation of mouse urothelial cells.
Bladders from TRPV1 −/− and TRPV1 +/+ mice were dissected, pinned urothelial side up and incubated with MEM media containing 2.5 mg/ml dispase (2-3 hours at 21 °C). Cells were collected by gentle scraping and dissociated in 0.025% trypsin EDTA at 37 °C (10 minutes) and re-suspended in keratinocyte serum free media (KSFM) before plating overnight onto collagen (IV) coated coverslips.
Calcium imaging of cultured urothelial and DRG neurones. Cultured DRG neurones and urothelial cells (20-24 hours) were loaded with 2 μM fura-2-acetoxymethyl ester (fura-2AM) for 30 minutes at 37 °C and washed (Hepes buffer: Hepes 10 mM, NaCl 135 mM, KCl 5 mM, glucose 10 mM, CaCl 2 2 mM and MgCl 2 1 mM, pH 7.4) for 1 hr prior to imaging. Cells were stimulated with either ATP, αβMethylene ATP, or capsaicin for 3-5 minutes and changes in intracellular calcium [Ca 2+ ] i, were monitored in real-time. Ionomycin (Sigma) (5 µM) was applied as a positive control in DRG experiments. Results are expressed either as relative fluorescence (RF) or normalised to percentage of maximum ionomycin-induced calcium influx to overcome intra-experimental variability. [Ca 2+ ] i is represented as the ratio between the fluorescence signal at 340/380 nm. Numbers are quoted as: N = number of mice and n = number of cells.
Experimental protocols. In nerve recording experiments, the preparation was allowed an equilibration period of 30 min before any experimental protocol commenced. Control bladder distensions were always carried out using isotonic saline (NaCl, 0.9%) at a rate of 100 µl min −1 to a maximum pressure of 40 mmHg; at this pressure the infusion pump was stopped and rapid evacuation of the fluid occurred by opening the two-way catheter in the dome. This was repeated several times at intervals of 10 min to assess the viability of the preparation and reproducibility of the intravesical pressure and neuronal responses to distension. Following equilibration, bladders were partially filled (12 mmHg) with the outflow tap closed and left to equilibrate until bladder pressure and baseline nerve activity were stable (approx. 45 mins). αβMethylene-ATP (30 µM) or dimethyl-4-phenylpiperazinium (DMPP) (100 µM) were applied to the bath by bolus dose and washed out with Krebs buffer after 60 seconds with the perfusion system. Krebs buffer was continually perfused for one hour to minimize receptor desensitisation before a second administration of experimental compound. In some experiments, capsazepine (CPZ) (10 µM), or vehicle control (0.01% DMSO) was perfused in the Krebs solution for the period prior to the second stimulus. Bladder contraction was measured as a change in intravesical pressure (mmHg) and latency was calculated as the time (in seconds) from drug application to maximal nerve firing.
Mediator release. Sequential filling (ramp distension) of a catheterised bladder (as described above) was performed (0.9% NaCl, to 40 mmHg). Intraluminal fluids were collected and the amount of ATP and Acetylcholine (ACh) in the samples was determined using the luciferin-luciferase ATP bioluminescent assay and the Amplex red ACh/ACh esterase assay kit (Molecular Probes).
SCientiFiC REPORts | (2018) 8:197 | DOI:10.1038/s41598-017-18136-w qPCR. mRNA from cultured urothelial cells and DRG neurones (20-24 hrs) was isolated (RNAeasy minikit, Qaigen) and cDNA was synthesised by reverse transcription using superscript III (Invitrogen) from mRNA following the manufacturers protocol. cDNA was amplified by PCR for 35 cycles with the following primers and Sybr green fluorescence (Biorad). Primer sequence for the TRPV1 receptor was forward 5′ CAAGGCTCTATGATCGCAGG 3′ and reverse 5′GAGCAATGGTGTCGTTCTGC 3′. All samples were assayed in triplicate in the same plate. The relative amount of target mRNA was calculated by the 2 −ΔΔCt method using β-actin as a housekeeping gene.

Data analysis and Statistics.
Whole nerve multi-fibre afferent nerve activity was quantified using a Spike2 software which counted number of action potentials crossing a pre-set threshold (Cambridge Electronic Design, Cambridge, UK).
For afferent nerve recording experiments, baseline activity was subtracted from stimulus induced response. Data are presented as means ± S.E.M. Statistical analysis was carried out using either a 1 or 2-way ANOVA and appropriate post-hoc test or a Student's T-test, where appropriate. Tukey post-hoc test was used for 2-way ANOVA when n values for groups was the same, and Bonferroni for all other ANOVA, with significance confirmed at P < 0.05.