|Home | About | Journals | Submit | Contact Us | Français|
Micturition, or urination, occurs involuntarily in infants and young children until the age of 3 to 5 years, after which it is regulated voluntarily. The neural circuitry that controls this process is complex and highly distributed: it involves pathways at many levels of the brain, the spinal cord and the peripheral nervous system and is mediated by multiple neurotransmitters. Diseases or injuries of the nervous system in adults can cause the re-emergence of involuntary or reflex micturition, leading to urinary incontinence. This is a major health problem, especially in those with neurological impairment. Here we review the neural control of micturition and how disruption of this control leads to abnormal storage and release of urine.
The storage and periodic elimination of urine depend on the coordinated activity of smooth and striated muscles in the two functional units of the lower urinary tract, namely a reservoir (the urinary bladder) and an outlet consisting of the bladder neck, the urethra and the urethral sphincter1,2. The coordination between these organs is mediated by a complex neural control system that is located in the brain, the spinal cord and the peripheral ganglia.
The lower urinary tract differs from other visceral structures in several ways. First, its dependence on CNS control distinguishes it from structures that maintain a level of function even after the extrinsic neural input has been eliminated. It is also unusual in its pattern of activity and in the organization of its neural control mechanisms. For example, the bladder has only two modes of operation: storage and elimination. Thus, many of the neural circuits that are involved in bladder control have switch-like or phasic patterns of activity, unlike the tonic patterns that are characteristic of the autonomic pathways that regulate cardiovascular organs. In addition, micturition is under voluntary control and depends on learned behaviour that develops during maturation of the nervous system, whereas many other visceral functions are regulated involuntarily.
Owing to the complexity of the neural mechanisms that regulate bladder control, the process is sensitive to various injuries and diseases. This Review summarizes the results of recent studies in animals and humans that have provided new insights into the sensory and motor mechanisms that underlie voluntary and reflex micturition, the changes in neural pathways that occur following disease or injury that alters lower-urinary-tract function, and new therapies for the treatment of neurogenic bladder dysfunction.
The requirement for voluntary control over the lower urinary tract necessitates complex interactions between autonomic (mediated by sympathetic and parasympathetic nerves) and somatic (mediated by pudendal nerves) efferent pathways1,2 (FIG. 1a). The sympathetic innervation arises in the thoracolumbar outflow of the spinal cord, whereas the parasympathetic and somatic innervation originates in the sacral segments of the spinal cord. Afferent axons from the lower urinary tract also travel in these nerves.
Sympathetic postganglionic nerves — for example, the hypogastric nerve — release noradrenaline, which activates β-adrenergic inhibitory receptors in the detrusor muscle to relax the bladder, α-adrenergic excitatory receptors in the urethra and the bladder neck, and α- and β-adrenergic receptors in bladder ganglia3,4 (FIG. 1b).
Parasympathetic postganglionic nerves release both cholinergic (acetylcholine, ACh) and non-adrenergic, non-cholinergic transmitters. Cholinergic transmission is the major excitatory mechanism in the human bladder4 (FIG. 1b). It results in detrusor contraction and consequent urinary flow and is mediated principally by the M3 muscarinic receptor, although bladder smooth muscle also expresses M2 receptors5. Muscarinic receptors are also present on parasympathetic nerve terminals at the neuromuscular junction and in the parasympathetic ganglia3,6. Activation of these receptors on the nerve terminals can enhance (through M1 receptors) or suppress (through M4 receptors) transmitter release, depending on the intensity of the neural firing. Non-cholinergic excitatory transmission is mediated by ATP actions on P2X purinergic receptors in the detrusor muscle7. Inhibitory input to the urethral smooth muscle is mediated by nitric oxide (NO) that is released by parasympathetic nerves4.
Somatic cholinergic motor nerves that supply the striated muscles of the external urethral sphincter arise in S2–S4 motor neurons in Onuf's nucleus and reach the periphery through the pudendal nerves1,2 (FIG. 1a). A medially placed motor nucleus at the same spinal level supplies axons that innervate the pelvic floor musculature.
Sensations of bladder fullness are conveyed to the spinal cord by the pelvic and hypogastric nerves1,8, whereas sensory input from the bladder neck and the urethra is carried in the pudendal and hypogastric nerves. The afferent components of these nerves consist of myelinated (Aδ) and unmyelinated (C) axons. The Aδ-fibres respond to passive distension and active contraction8 and thus convey information about bladder filling. The C-fibres are insensitive to bladder filling under physiological conditions (they are therefore termed ‘silent’ C-fibres) and respond primarily to noxious stimuli such as chemical irritation9 or cooling10. The cell bodies of Aδ-fibres and C-fibres are located in the dorsal root ganglia (DRG) at the level of S2–S4 and T11–L2 spinal segments. The axons synapse with interneurons that are involved in spinal reflexes and with spinal-tract neurons that project to higher brain centres that are involved in bladder control (see below).
A dense nexus of sensory nerves has been identified in the suburothelial layer of the urinary bladder in both humans and animals11,12, with some terminal fibres projecting into the urothelium13,14. This suburothelial plexus is particularly prominent at the bladder neck but is relatively sparse at the dome of the bladder15 and is thought to be critical in the sensory function of the urothelium.
Animal studies have shown that non-neuronal cells also play a part in bladder sensory mechanisms. Traditionally viewed as merely a passive barrier, the urothelium is now known to have specialized sensory and signalling properties that allow it to respond to chemical and mechanical stimuli and to engage in reciprocal chemical communication with nerves in the bladder wall1,16,17. These properties include the expression of nicotinic, muscarinic, tachykinin, adrenergic, bradykinin and transient-receptor-potential vanilloid receptors (such as TRPV1)18; responsiveness to transmitters released from afferent nerves; a close physical association with afferent nerves; and the ability to release chemical mediators such as ATP19, ACh and NO, which can regulate the activity of adjacent nerves and thereby trigger local vascular changes and/or reflex bladder contractions (FIG. 2).
The presence of muscarinic and nicotinic receptors in the urothelium20–24 has focused attention on the role of ACh as a chemical mediator of neural–urothelial interactions. ACh is released from the urothelium in response to chemical or mechanical stimuli16,25–27. Application of a muscarinic receptor agonist to strips of rat bladder tissue induces membrane-potential transients and Ca2+ transients that begin near the urothelial–suburothelial interface and then spread to the detrusor smooth muscle, raising the possibility that the urothelium could regulate the generation of spontaneous, non-voiding contractions in the bladder28. Indeed, in vivo intravesical administration of muscarinic or nicotinic receptor agonists or anticholinesterase agents that increase the levels of endogenous ACh facilitates reflex bladder activity in rats and cats26. Thus, the clinical effect of antimuscarinic agents in overactive bladder (a decrease in sensory symptoms) might be related to the blocking of muscarinic receptors in the urothelium or in afferent nerves16,26,29.
Stimulation of cholinergic receptors in the urothelium induces the release of ATP and an as yet unidentified smooth-muscle relaxant factor20,23,26. The function of ATP release from the urothelium has attracted considerable attention because intravesical administration of ATP induces detrusor overactivity by stimulating the purinergic receptor P2X ligand-gated ion channel 3 (P2X3) or P2X2/3 on afferent nerves1,30. Mice that lack P2X3 receptors exhibit reduced bladder activity and inefficient voiding, suggesting that activation of P2X3 receptors on bladder afferent nerves by ATP released from the urothelium is essential for normal bladder function31.
The discovery of a suburothelial layer of myofibroblasts (also referred to as interstitial cells32) that lie in close proximity to nerves and are extensively linked by gap junctions33 led to the proposal that these cells, together with afferent nerves, the urothelium and smooth muscle, might collectively have the properties to act as a stretch-receptor organ33. Furthermore, the demonstration that they express ATP-gated purinergic P2Y receptors34 raises the possibility that they might respond to urothelial ATP release.
The regulation of micturition requires connections between many areas in the brain and extensive tracts in the spinal cord that involve sympathetic, parasympathetic and somatic systems. Parasympathetic and sympathetic preganglionic neurons (PGNs) are located in the intermediate grey matter (laminae V–VII; FIG. 3) of spinal cord sacral and lumbar segments, respectively. Parasympathetic PGNs send dendrites into the dorsal commissure and into the lateral funiculus and lateral dorsal horn of the spinal cord35 and exhibit an extensive axon collateral system that is distributed bilaterally in the cord36. A similar axon collateral system has not been identified in sympathetic preganglionic neurons. The somatic motor neurons that innervate the external urethral sphincter are located in the ventral horn (lamina IX) in Onuf's nucleus, have a similar arrangement of transverse dendrites and have an extensive system of longitudinal dendrites that travel within Onuf's nucleus1,37.
Interneurons in the lumbosacral spinal cord of the rat that are involved in lower-urinary-tract function are located in the dorsal commissure, the superficial dorsal horn and the parasympathetic nucleus38–41 (FIG. 3c). Some of these interneurons send long projections to the brain41, whereas others make local connections in the spinal cord and participate in segmental spinal reflexes42.
Afferent nerves from the bladder project to regions of the spinal cord that contain interneurons and parasympathetic PGN dendrites43,44 (FIG. 3). Pudendal afferent pathways from the urethra and the urethral sphincter exhibit a similar pattern of termination45,46. The overlap between bladder and urethral afferents in the lateral dorsal horn and the dorsal commissure indicates that these regions are probably important sites of viscerosomatic integration that might be involved in coordinating bladder and sphincter activity.
In the brain, many neuron populations are involved in the control of the bladder, the urethra and the urethral sphincter. Some, such as the serotonergic neurons of the medullary raphe nuclei, the noradrenergic neurons of the locus coeruleus and the noradrenergic A5 cell group in the brain stem, are non-specific ‘level-setting’ mechanisms with diffuse spinal projections47. Others are specific for micturition: these include the neurons of Barrington's nucleus (also called the pontine micturition centre (PMC)) and those of the periaqueductal grey (PAG), cell groups in the caudal and preoptic hypothalamus, and the neurons of several parts of the cerebral cortex, in particular the medial frontal cortex39,48 (FIG. 4).
There are interconnections between some of these brain areas and also between the brain and the lumbosacral spinal cord. Interneurons in the spinal cord project to the PAG49 and (in rats) to the PMC (FIG. 4). Neurons in the PMC receive input from the PAG and from the anterior and caudal hypothalamus, but few or no other inputs; in turn, they send descending axons back to the parasympathetic nucleus in the spinal cord. The paraventricular nucleus of the hypothalamus projects non-specifically to all autonomic preganglionic motor neurons in the spinal cord, including the sacral parasympathetic and sphincter motor nuclei50, but neurons in the lateral pons project rather selectively to the sphincter motor nucleus50,51. Thus, the supraspinal regulation of lower-urinary-tract function probably depends on multiple pathways carrying information between the brain and the spinal cord. The pathways that are specific for micturition47 are the topic of this Review.
The neural pathways that control lower-urinary-tract function are organized as simple on–off switching circuits that maintain a reciprocal relationship between the urinary bladder and the urethral outlet. Storage reflexes are activated during bladder filling and are organized primarily in the spinal cord, whereas voiding is mediated by reflex mechanisms that are organized in the brain (FIG. 5).
Throughout bladder filling, the parasympathetic innervation of the detrusor is inhibited and the smooth and striated parts of the urethral sphincter are activated, preventing involuntary bladder emptying. This process is organized by urethral reflexes known collectively as the ‘guarding reflex’. They are activated by bladder afferent activity that is conveyed through the pelvic nerves, and are organized by interneuronal circuitry in the spinal cord52,53 (FIG. 5a). Some input from the lateral pons, which is also known as the ‘L-region’ or the ‘pontine storage centre’, might facilitate sphincter reflexes or have a role in involuntary sphincter control51,54. In animals, reflex activation of the lumbar sympathetic pathway is involved in the inhibition of bladder smooth muscle21, contraction of the bladder outlet and inhibition of parasympathetic activity at the level of the autonomic ganglia31. The importance of the sympathetic control in humans is less obvious, as sympatholytic drugs or resection of the sympathetic chain have little effect on bladder storage.
In cats, afferent input from the bladder 41,50,55,56 ascends, through spinal interneurons in the lateral funiculus, to a relay station in the central PAG that in turn, through the lateral PAG, provides an excitatory input to the PMC50 (FIG. 6). The presumed interconnection between the central and lateral PAG implies that signal processing occurs in the PAG, probably enabling higher centres to control the excitatory input to the PMC. In rats, ascending projections from the spinal cord also terminate directly in the PMC57. Additional bladder sensory input to the brain is carried by the spinothalamic tract, the spinohypothalamic tract and dorsal-column pathways to the nucleus gracilis.
Pharmacological and electrophysiological studies have indicated that the circuitry that feeds bladder afferent activity to midbrain and pontine centres and transmits efferent signals from the pons to the sacral cord (FIG. 5b) allows the spinobulbospinal voiding-reflex pathway to function as a switch that is either in a completely ‘off’ mode (storage) or a maximally ‘on’ mode (voiding)52. During bladder filling the parasympathetic efferent pathway to the bladder, including a population of PMC neurons, is turned off 58,59, but at a critical level of bladder distension the afferent activity arising from tension receptors in the bladder switches the pathway to maximal activity58,60,61. Drugs injected into the PMC can change the set-point (that is, the bladder volume) for activation of the switch without altering the magnitude of the voiding reflex62. Similarly, lesions in brain areas that lie rostral to the pons (for example, the diencephalon and the cerebral cortex) can alter the set-point of the voiding reflex, indicating that the switching circuit is tonically modulated by inhibitory and excitatory influences from the forebrain.
Operating in this way, the reflex circuitry shown in FIG. 5 would lead to involuntary bladder emptying (that is, incontinence) whenever the bladder volume reached a critical level. However, in continent individuals the firing of the voiding reflex is under strict voluntary control, enabling one to plan to void at a socially acceptable time and place. The decision to void, which is a crucial aspect of human behaviour, is based on a combination of factors, including one's emotional state, an appreciation of the social environment and the sensory signals arising from the bladder. Knowledge of the extent to which one's bladder content is comfortable and ‘safe’ is central in this process. Thus, voluntary control of the bladder and the urethra has two important aspects, namely registration of bladder filling sensations and manipulation of the firing of the voiding reflex. The PAG has a pivotal role in both49. On the one hand it receives and passes ascending bladder signals to higher brain centres and into the realm of conscious sensation. On the other hand it receives projections from many higher brain centres and also controls the primary input to the PMC; during bladder filling, therefore, such higher brain centres (particularly the prefrontal cortex47) can suppress the excitatory signal to the PMC and thus prevent voiding or incontinence; when voiding is consciously desired, they can allow the PMC to be excited.
Our understanding of the processing of bladder sensations in humans has been greatly advanced in recent years by the advent of functional brain imaging. Imaging studies have shown activation of the PAG during bladder filling63,64 (FIG. 6); this is in keeping with its postulated role in receiving bladder afferents and relaying them (perhaps through the thalamus) to the insula, where normal visceral sensations, such as desire to void, are thought to be mapped65. Consistent with this postulate, the insula was active in most imaging studies of urine storage66 and its activation apparently increased with bladder filling67. By contrast, bladder cooling did not significantly activate the PAG68, suggesting that afferents related to noxious bladder stimulation might follow a different pathway to the thalamus and the cerebral cortex.
The PAG has multiple afferent and efferent connections, not only with the thalamus, the prefrontal cortex and the insula, but also with other higher brain centres. The influence of these centres — the anterior cingulate cortex particularly — probably determines how much attention one pays to signals coming from bladder afferents and how one reacts to them, whether by deciding to void or by recruiting mechanisms (for example, urethral sphincter contraction) that allow voiding to be postponed. The anterior cingulate cortex is an extensive region, and different parts were activated in different studies66. The dorsal part seems to be especially strongly activated by bladder distension in subjects with “poor bladder control” (REF. 69), suggesting a strong emotional reaction to a situation the subjects experienced as threatening. This pattern of abnormal activity might help to distinguish an abnormal sensation of urgency from a strong but normal desire to void.
Both clinical observations and observations from functional imaging strongly suggest that in humans the frontal lobes play an important part in determining the appropriateness of micturition. For example, the right inferior frontal gyrus, which is part of the prefrontal cortex, was active during storage in all the studies on which FIG. 6a is based63,64. The prefrontal cortex is thought to be the seat of planning complex cognitive behaviours and the expression of personality and appropriate social behaviour and has a role in attention and response-selection mechanisms70. It has strong and direct connections with the PAG47, suggesting that it might be responsible for tonic suppression of voiding that is relaxed only when voiding is both desired and socially appropriate. In keeping with this postulate, in subjects who cannot sustain such suppression (that is, in subjects who have poor bladder control) the prefrontal cortex responds abnormally weakly to bladder filling67. The prefrontal cortex has multiple connections with the anterior cingulate gyrus71 and both regions have direct or indirect connections with the PAG, the hypothalamus and other areas that are associated with autonomic control. The caudal hypothalamus has direct access to the PMC (FIG. 6b) and responded to changes in bladder volume or sensation in two imaging studies69,72. This might represent a further layer of control of the micturition reflex that permits voiding only if it is judged ‘safe’ to do so (REF. 49).
The existence of a pontine storage centre in humans (FIG. 5a) is not certain54, but several imaging studies have shown activation near the expected location (that is, ventrolateral to the PMC) during storage or withholding of urine66.
Animal studies have shown that reflex micturition is mediated by a spinobulbospinal pathway that passes through the PMC in the rostral brainstem49,55,73 (FIG. 5b). Excitation of the PMC activates descending pathways that cause urethral relaxation and, some seconds later, activation of the sacral parasympathetic outflow. This results in contraction of the bladder and an increase in intravesical pressure and the flow of urine. Relaxation of the urethral smooth muscle is mediated by activation of the parasympathetic pathway to the urethra, which triggers the release of NO, and by the removal of adrenergic and somatic cholinergic excitatory inputs. Secondary reflexes elicited by the flow of urine through the urethra facilitate bladder emptying1.
Single-unit recordings in cats and rats have revealed that several populations of PMC neurons exhibit firing that is correlated with reflex bladder activity58,59,61,74–76; these include neurons that are silent in the absence of bladder activity but fire prior to and during reflex bladder contractions (direct neurons); neurons that are active during the period between bladder contractions and are inhibited during contractions (inverse neurons); and neurons that fire transiently at the beginning and end of bladder contractions (on–off neurons). A large percentage of direct neurons project to the lumbosacral spinal cord, whereas only a small percentage of inverse neurons do. Thus, it has been speculated that inverse neurons function as local inhibitory neurons in the PMC. Both direct and inverse neurons exhibit excitatory synaptic responses to electrical stimulation of afferent axons in the pelvic nerve58.
Functional imaging studies into the process of voiding in humans showed that the cortical and brainstem areas involved were comparable to those that are involved in cats: voiding was associated with activation in the prefrontal cortex, the insula, the hypothalamus, the PAG and in a region that is comparable to that of the PMC in cats77,78. In study participants who were unable to void, slightly different prefrontal activation was seen, together with a more ventral region of pontine activation that supported the existence of a pontine storage centre (see FIG. 5a). Another PET study confirmed the involvement of the insula, the PAG and the PMC in voiding79. These observations seem to be compatible with the concept outlined above — that voluntary voiding implies interruption of the tonic suppression (by the prefrontal cortex) of PAG input to the PMC.
Various neurotransmitters have been implicated in the central control of the lower urinary tract. Putative excitatory transmitters include glutamic acid, tachykinins, pituitary-adenylate-cyclase-activating polypeptide, NO and ATP61,80. Glutamic acid, acting on NMDA (N-methyl-d-aspartate) and non-NMDA receptors, seems to be the essential transmitter in spinal and supraspinal reflex pathways that control the bladder and the external urethral sphincter81–83. Inhibitory amino acids (GABA (γ-aminobutyric acid) and glycine) and opioid peptides (enkephalins) exert a tonic inhibitory control in the PMC and regulate bladder capacity1,62,84. These substances also have inhibitory actions in the spinal cord. Some transmitters (dopamine, serotonin (5-hydroxytryptamine, 5-HT), noradrenaline, ACh and non-opioid peptides, including vasoactive intestinal polypeptide and corticotropin-releasing factor) have either inhibitory or excitatory effects, depending on the type of receptor that is activated, the receptors' location in the CNS and the species83.84. For example, dopamine elicits inhibitory effects on micturition through D1-like receptors and facilitatory effects through D2-like receptors. On the other hand, activation of 5-HT1A receptors inhibits micturition in the cat but facilitates it in the rat80.
The mechanisms that are involved in the storage and periodic elimination of urine undergo marked changes during prenatal and postnatal development83,85. In the fetus, before the nervous system has matured, urine is presumably eliminated from the bladder by non-neural mechanisms; however, at later stages of development voiding is regulated by primitive reflex pathways that are organized in the spinal cord. As the human CNS matures postnatally, reflex voiding is eventually brought under the modulating influence of higher brain centres (FIG. 7). In adults, injury or disease of the nervous system can lead to the re-emergence of primitive reflexes86,87.
In many species (for example, rats and cats), voiding in neonates depends on an exteroceptive somato-bladder reflex mechanism that is triggered when the mother licks the genital or perineal region of the young animal. This reflex is organized in the sacral spinal cord and has an afferent limb in the pudendal nerve and an efferent limb in the pelvic nerve85. Similar reflexes have been identified in human infants. During the postnatal period this reflex becomes progressively weaker and is eventually replaced by an inhibitory perineal-to-bladder reflex and the adult form of reflex voiding85, but spinal-cord injury in adult animals can cause the re-emergence of the excitatory somato-bladder reflex. The developmental and spinal-cord-injury-induced plasticity in sacral parasympathetic reflex pathways is due in part to alterations in glutamatergic excitatory transmission between interneurons and parasympathetic PGNs42,58. It has been proposed that these synaptic changes are due to competition between segmental and supraspinal inputs to the PGNs. Thus, synaptic remodelling in the sacral parasympathetic nucleus and in the brain is likely to be an important factor in the postnatal maturation of voiding reflexes.
In humans, disruption of the suprapontine circuitry as a result of anterior cerebral lesions or the degeneration of dopaminergic neurons in Parkinson's disease removes tonic inhibitory control over the PMC, resulting in decreased bladder capacity and detrusor overactivity. Pharmacological studies in animals following decerebration or cerebral infarction have revealed that suprapontine lesions lead to upregulation of NMDA-glutamatergic and D2-dopaminergic excitatory mechanisms and downregulation of NMDA-glutamatergic and M1-muscarinic inhibitory mechanisms in the brain88,89. Experimental cerebral infarction in rats also increases expression in the PMC of Cox2 and the immediate early genes c-fos and Zif268 (REF. 89), and these effects were blocked by an NMDA-glutamatergic antagonist. Injection of an RNA-synthesis inhibitor into the pons blocked the upregulation of c-fos and Zif268 and suppressed the detrusor overactivity that was induced by cerebral infarction, suggesting that the effect of suprapontine lesions on bladder activity is mediated in part by changes in synaptic transmission in the PMC89.
SCI rostral to the lumbosacral level eliminates voluntary and supraspinal control of voiding, leading initially to an areflexic bladder and complete urinary retention, followed by a slow development of automatic micturition and neurogenic detrusor overactivity (NDO) that is mediated by spinal reflex pathways90. However, voiding is commonly inefficient owing to simultaneous contractions of the bladder and the urethral sphincter (detrusor–sphincter dyssynergia) (FIG. 7c).
In cats with spinal lesions and disconnection of the sacral cord from brainstem centres, a segmental sacral spinal reflex emerges that drives reflex bladder contractions and is mediated by capsaicin-sensitive C-fibre afferents. This is the fundamental cause of reflex detrusor contractions in response to low-volume filling and of the underlying pathophysiology of NDO in experimental models91 (FIG. 8). Sufficient clinical evidence exists to support the view that a comparable process occurs in humans following spinal-cord lesions, so there have been extensive efforts to therapeutically reduce the C-fibre afferent input. The emergence of C-fibre bladder reflexes seems to be mediated by several mechanisms, including changes in central synaptic connections and alterations in the properties of the peripheral afferent receptors that lead to sensitization of the ‘silent’ C-fibres and the unmasking of responses to mechanical stimuli91,92.
In rats, bladder afferent neurons undergo both morphological93 and physiological changes following SCI94; it has been speculated that this neuroplasticity is mediated by the actions of neurotrophic factors, such as nerve growth factor (NGF), that are released in the spinal cord or the urinary bladder95–97. Indeed, the production of neurotrophic factors increases in the bladder after SCI95,96, and in rats chronic administration of NGF into the bladder induces bladder hyperactivity and increases the excitability of dissociated bladder afferent neurons. By contrast, intrathecal application of NGF antibodies suppressed neurogenic detrusor overactivity96 and detrusor–sphincter dyssynergia97 in SCI rats.
Animal and human studies also support a role for the suburothelial expression of TRPV1 (REF. 98), P2X3 (REF. 98) and/or the sensory neuropeptides substance P (SP) and calcitonin-gene-related peptide (CGRP)99 in the pathophysiology of human NDO, and patients with NDO have increased TRPV1- and P2X3-immunoreactive suburothelial innervation compared with controls100.
Urgency incontinence with underlying IDO is a common condition, but its fundamental cause remains to be discovered. IDO probably has a number of pathogeneses depending on the patient's sex and age101. The hypothesis that IDO is myogenic was based on the finding that detrusor myocytes from people with IDO showed increased excitability and increased tendency for activity to spread between cells, resulting in coordinated myogenic contractions of the whole detrusor102. An alternative hypothesis proposed that the problem was neurogenic103. Subsequently a ‘modular model’ was suggested in which units of the bladder wall consisting of an area of smooth muscle supplied by either an intramural ganglia or possibly interstitial cells becomes hyperexcitable, so that physiological ‘micromotions’ become synchronous and cause involuntary detrusor contraction104. Most recently, Griffiths et al.67 demonstrated, using functional MRI, that “poor bladder control” was associated with abnormally weak activation of the orbitofrontal cortex, suggesting a central cause. These hypotheses might not be mutually exclusive and the cause of IDO might vary according to the patient group and the type of IDO.
There is accumulating evidence that aberrant afferent activity plays an important part in IDO: compared with controls, women with IDO have an increased density of suburothelial nerve fibres that are immunoreactive for SP and CGRP99, whereas men with detrusor overactivity that is associated with urethral obstruction have increased incidence of a positive ice-water test, which is thought to be related to activation of a C-fibre-afferent-evoked bladder reflex105. Moreover, intravesical administration of the TRPV1 agonist resiniferatoxin (RTX) delayed or suppressed involuntary detrusor contractions in patients with IDO106, and biopsies from patients with IDO had increased expression of TRPV1 and P2X3 that ‘normalized’ in response to botulinum A neurotoxin (BoNTX/A) treatment (see below)100. However, a recent study found that neurokinin-A-induced responses were impaired in detrusor muscle from patients with IDO but not from patients with neurogenic detrusor overactivity (NDO)107. It was concluded that IDO and NDO might have different pathophysiologies, although it remains unknown whether the findings reflect a consequence of other disorder-related factors or whether tachykinins (such as neurokinin A) are directly involved in the generation of some types of DO108.
Bladder pain and overactivity induced by cystitis have been attributed in part to changes in the electrical properties of capsaicin-sensitive C-fibre bladder afferent neurons. These neurons normally show phasic firing in response to prolonged depolarizing current pulses, whereas neurons isolated from rats with chronic chemically induced bladder inflammation exhibit tonic firing and a reduced threshold for the initiation of action potentials109. These changes have been linked to a reduction in a low-threshold A-type K+ current (IA) that controls neuronal excitability. Blockade of this K+ current with 4-aminopyridine, heteropodatoxin or SP unmasks tonic firing in normal cells, whereas treatment with KW-7158, a drug that opens A-type K+ channels, reverses this effect110. KW-7158 also reduces detrusor overactivity in rats with cystitis. NGF has been implicated in the mechanism of afferent sensitization because cystitis, like SCI, increases NGF expression in the bladder111. Chronic administration of NGF to the bladder or the spinal cord mimics the effect of chemically induced cystitis on IA currents and firing in bladder afferent neurons112.
Patch-clamp recordings in capsaicin-sensitive afferent neurons from cats with feline interstitial cystitis (FIC), a chronic naturally occurring painful bladder condition, revealed changes in IA currents and firing that were similar to those described in rats with cystitis or rats that had been treated with NGF113. In addition, the currents that were induced by capsaicin were increased in amplitude and exhibited a slower desensitization114. The effects of capsaicin were normalized by treating the neurons with an inhibitor of protein kinase C that presumably leads to dephosphorylation of the capsaicin-sensitive channel TRPV1. These findings suggest that the bladder symptoms in cats with FIC are related in part to changes in K+ and TRPV1 channels in sensory neurons.
Changes in the chemical communication between the urothelium and C-fibre afferent nerves might also be involved in afferent sensitization in cystitis. ATP excites sensory nerves by activating P2X2/3 receptors, and bladder distension releases ATP from the urothelium (FIG. 2). In FIC cats115 and patients116 with interstitial cystitis, ATP release from urothelial cells is increased, raising the possibility that enhanced signalling between the urothelium and afferent nerves might trigger painful bladder sensations. Prostaglandins are also released from the urothelium and trigger hyperalgesia by acting through G-protein-coupled receptors and second-messenger pathways. This leads to changes in afferent excitability through phosphorylation of the Nav1.8 Na+ channel that generates tetrodotoxin-resistant Na+ currents117. Down-regulation of Nav1.8 expression in bladder afferent pathways reduced the bladder hyperactivity and immediate-early-gene expression in the spinal cord that was induced by chemical irritation of the bladder118.
The rationale for using intravesical vanilloids to treat human detrusor overactivity was the demonstration in animal experiments that capsaicin-sensitive C-fibre afferents were functionally important in the pathogenesis of volume-determined reflex contractions91,94,119. The action of capsaicin and other vanilloids is mediated by TRPV1: activation of TRPV1results in membrane depolarization and massive release of sensory neuropeptides, which can lead to desensitization and cell death. Capsaicin initially excites C-fibres and then causes a period of prolonged desensitization120. The use of intravesical vanilloids in human NDO was therefore aimed at desensitizing bladder afferents121; prior instillation of a local anaesthetic reduced the capsaicin-induced irritation without blocking its effect122. A meta-analysis of the literature reported symptomatic and urodynamic improvement in 84% of NDO patients treated with intravesical capsaicin123 (although it seems likely that there was ‘double counting’, and negative experiences had not been published). The authors subsequently showed that diluting capsaicin with glucidic acid improved tolerability of the treatment and that enduring symptomatic and urodynamic benefits were possible with this preparation124.
RTX is a thousand times as neurotoxic as capsaicin but has equal pungency and can therefore produce the same effect at much greater dilutions. Uncontrolled studies showed that RTX was successful in treating NDO, but a large multi-centre placebo-controlled study using escalating doses failed to show clinical response in most of the treatment groups (although this was probably due to accidental inactivation of RTX during the administration of the solution). The effect of intravesical administration of vanilloids in patients with interstitial cystitis has also been disappointing. A multi-centre, randomized placebo-controlled trial using RTX showed no improvement over placebo in lower-urinary-tract functioning, pain levels and quality-of-life scores125.
At the time of writing, despite the many positive reports of the clinical efficacy of these agents and the sound theoretical basis for their use, no licensed intravesical vanilloid therapy is currently available. However, it is thought that several pharmaceutical companies have development programmes in progress for agents that block the TRPV1 mechanism.
Injection of BoNT/A into the detrusor was introduced as a new treatment of intractable NDO on the theoretical basis that it would temporarily block the pre-synaptic release of ACh from the parasympathetic innervation and produce a paralysis of the detrusor smooth muscle126. Subsequent clinical results have exceeded all expectations, but the evidence so far suggests that it might have a more complex mechanism of action, through the inhibition of excitatory-neurotransmitter release from bladder afferents and urothelial cells127 (FIG. 2).
In cultured rat DRG, BoNT/A induced a delayed, long-lasting inhibition of SP release and blocked neuronal glutamate release in an animal model of inflammatory pain128, leading to a reduction in neurally mediated inflammation. In a rat model of spinal NDO, BoNT/A significantly reduced the abnormal distension-evoked urothelial release of ATP129 and reduced the DO that was evoked by application of ATP or capsaicin in rat bladders in vivo130.
Biopsies from a mixed population of NDO and IDO patients treated with intradetrusor BoNT/A showed no change in the overall suburothelial afferent-nerve density during the clinical response but a progressive decrease and, finally, normalization of P2X3 and TRPV1 suburothelial nerve immunoreactivity, suggesting that BoNT/A affects the expression of sensory receptors in suburothelial nerve fibres100. P2X3 immunoreactivity showed the fastest change, which correlated with improvements in patients' sensation of urgency.
By inhibiting urothelial ACh release, BoNT/A might block ACh's proposed excitatory effect on suburothelial afferent nerves, myofibroblasts or detrusor parasympathetic nerve endings during urine storage. Inhibition of the increased urothelial ATP release would reduce its proposed excitatory effect on suburothelial and urothelial P2X3 receptors and P2Y receptors in the myofibroblast network. In addition, there might be an attenuation of central sensitization, leading to further peripheral desensitization through decreased central SP release.
As anticipated, BoNT/A's effect on efferent pathways is also significant. Both animal131 and clinical studies show post-BoNT/A decreases in detrusor pressures during both filling and voiding, in conjunction with post-treatment increases in post-void residual urine volumes126,132–134. It seems likely that the exceptional efficacy of BoNT/A for the treatment of urgency incontinence is due to its combined effect on those intrinsic afferent and efferent systems that are pathologically up-regulated in detrusor overactivity.
Several studies have shown that the benefits of BoNT/A are maintained for 9–11 months on average126,132 and that subsequent injections have comparable efficacy to the first, both in terms of the size of the effect and the duration of its action134. Furthermore, BoNT/A is as effective in patients with IDO135–137 as it is in those with intractable NDO with spinal aetiology. Minimally invasive methods of delivering the injections in an outpatient setting138 mean that this treatment will probably have far-reaching beneficial consequences for the treatment of incontinence.
Although therapies that improve bladder emptying in the context of prostatic hypertrophy are available, less attention has been given to the development of agents that are effective at treating the reduced urinary retention or incomplete emptying that are associated with neurological or functional disorders. This is because these conditions are not as prevalent as urinary incontinence, and because clean, intermittent self-catheterization already provides symptomatic relief; however, medication that could enhance efficient voiding would nevertheless be valuable.
Dense neuronal staining for NO synthase (NOS) has been identified in the urethral sphincters of animals and humans139,140, and substantial animal experimental data show that NO is an important inhibitory neurotransmitter: its release results in sphincter relaxation during voiding. The hypothesis that augmentation of NO might be effective in reducing detrusor–sphincter dyssynergia following SCI141 was tested using the NO donor isosorbide dinitrate. This produced a significant reduction in striated sphincter pressure at rest and during dyssynergic contraction, although there was no reduction in residual urine volume142. The use of a target-specific phosphodiesterase inhibitor might allow more subtle manipulation of the NO signalling cascade by reducing the breakdown of NO. However, sildenafil citrate, a selective phosphodiesterase type 5 (PDE-5) inhibitor, did not improve sphincter relaxation in women with a primary disorder of sphincter relaxation143, possibly because the PDE-5 isoform that is targeted by this drug is not present in the female urethral sphincter.
During the past two decades, knowledge of the complex neural control of the apparently simple process of micturition has grown significantly. This has given us a better insight into the control of bladder storage and voiding in health and disease and a more complete understanding of the mode of action of existing therapies. The development of future treatments to reduce involuntary detrusor contraction and improve bladder emptying will doubtless be soundly based on this scientific understanding and be taken forward knowing that improvements in bladder control have a positive impact on patients' well-being.
Derek Griffiths' homepage: http://ourworld.compuserve.com/homepages/nacs
ALL LINKS ARE ACTIVE IN THE ONLINE PDF
Supported by US National Institutes of Health grants DK049430 and DK077783 to W.D. C.F.'s contribution to this work was undertaken at University College London Hospital/University College London, who received a proportion funding from the Department of Health's National Institute for Health Research Biomedical Research Centres funding scheme.
Competing interests statement: The authors declare competing financial interests: see web version for details.