Electrical Characteristics of Suburothelial Cells Isolated From the Human Bladder

0022-5347/04/1712-0938/0 THE JOURNAL OF UROLOGY® Copyright © 2004 by AMERICAN UROLOGICAL ASSOCIATION

Vol. 171, 938 –943, February 2004 Printed in U.S.A.

DOI: 10.1097/01.ju.0000108120.28291.eb

ELECTRICAL CHARACTERISTICS OF SUBUROTHELIAL CELLS ISOLATED FROM THE HUMAN BLADDER G. P. SUI, C. WU

AND

C. H. FRY*

From the Institute of Urology and Nephrology, London, United Kingdom

ABSTRACT

Purpose: We measured the membrane electrical characteristics as well as the response to adenosine triphosphate of cells isolated from the suburothelial layer of the bladder. Materials and Methods: Suburothelial cells were isolated from biopsy samples of human bladder by collagenase disruption. Electrophysiological measurements were done under current and voltage clamp to record membrane potential and ionic currents using patch pipettes with a K⫹ based filling solution. Intracellular [Ca2⫹] was measured with Fura-2. Results: Cells were different from epithelial cells by their spindle-shaped appearance with projections at either end. The cells stained for vimentin but epithelial and smooth muscle cells did not. The cells had small membrane capacitance (27 ⫾ 16 pF) and a specific membrane resistance of 90 ⫾ 48 ⫻ 109 ⍀ cm2. Average membrane potential was ⫺63 ⫾ 14 mV but cells showed spontaneous spikes or random fluctuations of membrane potential. A small net inward current was superimposed by a larger outward current. Inward current was attenuated by the removal of extracellular Ca. Outward current showed large spontaneous fluctuations and was greatly decreased by 30 mM tetraethyl ammonium chloride. Adenosine triphosphate (30 to 100 ␮M) elicited an inward current of about 50 pA and large intracellular Ca2⫹ transients. Conclusions: These cells are electrically active which, in conjunction with the previous observation of connexin 43 labeling, suggests that they could act as an electrical network. A quantitative model of voltage distribution in such a network after the generation of inward current suggests that individual cells could not act as pacemakers, but rather a group of simultaneously activated cells could exert a peripheral excitatory effect that would amplify the magnitude of the original response. The implications of this in terms of bladder sensation are discussed. KEY WORDS: bladder, connexin 43, electric capacitance, urothelium, calcium

The role of the urothelium in bladder sensation has recently been studied, such that adenosine triphosphate (ATP) is released from the basolateral surface when the hydraulic pressure gradient across the bladder wall is altered.1 Furthermore, P2X3 purinoceptors are localised on suburothelial afferents2 and it has been proposed that ATP is a chemical mediator between changes to the transmural pressure gradient and afferent excitation. The hypothesis is strengthened in that P2X3 knockout mice show urinary retention,3 consistent with decreased afferent sensitivity to bladder filling. A suburothelial cell layer was recently described immediately below the basal lamina that has the microscopic characteristics of myofibroblasts.4, 5 These cells make close appositions with bare nerve endings and they are functionally connected via gap junctions composed on connexin 43. It is also significant that similar cells have been described as targets for efferent nitrergic fibers.6 Therefore, these cells could form an electrically and chemically connected functional syncitium in the suburothelial space and could act as an intermediary step in the signal transduction process between urothelial deformation and afferent excitation. Such a hypothesis is attractive for several reasons. It would have high gain (sensitivity) because local ATP release would have a more extensive effect through a functional syncitium of myofibroblasts. The ability to regulate the activity of these cells, possibly through a nitrergic pathway, would allow control of the gain of the

sensory pathway. Also, myofibroblasts are readily replaced, ensuring a robust sensory system in a tissue exposed to extremes of pressure and ionic composition. However, the functional cellular characteristics of these cells are unknown, although there have been functional studies in similar interstitial cells in the muscle mass of visceral smooth muscle organs, such as the gut and urethra.7, 8 In these initial experiments we isolated myofibroblast-like cells from the urothelium of human bladder biopsies, determined whether they show electrophysiological properties consistent with their ability to form an electrical functional syncitium and ascertained whether they respond to exogenously applied ATP. METHODS

Cell preparation. Human bladder wall samples were obtained with patient consent from patients undergoing cystectomy or bladder augmentation. The urothelium was dissected from the underlying detrusor layer under ⫻10 to ⫻20 magnification. Urothelial samples were digested at 37C with a collagenase based medium, which has been described previously for the preparation of isolated human detrusor myocytes,9 and with constant stirring for 30 minutes. After treatment the tissue was partially disrupted. Two main cell types could be seen under a light microscope; namely round urothelial cells and a layer of ovoid or spindle-shaped cells with or without 1 or more dendrite-like structures (fig. 1). Solutions. Biopsy samples were placed immediately in Cafree HEPES buffered medium and used immediately for experiments. The solution was composed of 105.4 mM NaCl, 22.3 mM NaHCO3, 3.6 mM KCl, 0.9 mM MgCl2, 0.4 mM

Accepted for publication August 8, 2003. Study received local ethical committee approval. Supported by St. Peter’s Trust and Wellcome Trust. * Correspondence: The Institute of Urology and Nephrology, 48 Riding House St., London W1W 7EY, United Kingdom (telephone: ⫹44 20 7679 9376; FAX: ⫹44 20 7679 9584; e-mail: [email protected]). 938

939

ELECTROPHYSIOLOGY OF SUBUROTHELIAL CELLS RESULTS

FIG. 1. Micrographs of cells isolated from bladder wall. A, light micrograph shows suburothelial cell on dark field background. B, light micrograph demonstrates detrusor smooth muscle on dark field background. Note difference in size compared with A and B.

NaH2PO4, 19.5 mM HEPES, 5.4 mM glucose and 4.5 mM Na pyruvate. Cells were superfused at 37C with a solution composed of 118 mM NaCl, 4.0 mM KCl, 24 mM NaHCO3, 0.4 mM NaH2PO4, 1.0 mM MgCl2, 1.8 mM CaCl2, 6.1 mM glucose and 5.0 mM Na pyruvate gassed with 5% CO2-95% O2, pH 7.4. ATP was added to the superfusate from a 100 mM aqueous stock solution of Na2ATP. Immunohistochemical staining. Samples of cells that would have been used for experimental recording were stained for vimentin using mouse mAb 3B41112 457 (Boehringer Mannheim, Indianapolis, Indianapolis) (1:100) and AP 192C donkey antimouse IgG (Chemicon International, Temecula, California) as a secondary antibody conjugated to Cy3 (1:250). Further details have been described previously.4 Electrophysiological recordings. They were made using patch-type electrodes filled with a solution composed of 20 mM KCl, 110 mM aspartic acid, 5.45 mM MgCl2, 5.0 mM Na2ATP, 0.1 mM Na4GTP, 0.1 mM egtazic acid and 5.0 mM HEPES, adjusted to pH to 7.1 with KOH. Membrane potentials were recorded in current clamp mode and resting potentials were recorded with no current (I) passed into the cells (Ih ⫽ 0). Ionic currents were recorded under voltage clamp using 2-second depolarizing steps generally from a holding potential (Vh) of ⫺100 mV. To investigate the effect of ATP on net ion current a Vh of ⫺60 mV was used because it is close to the average resting membrane potential. Timeaveraged outward current magnitude was obtained by integrating the area under the trace using I ⫽ 0 as the lower limit. Measurement of intracellular [Ca2⫹] ([Ca2⫹]i). [Ca2⫹]i was measured using epifluorescence microscopy with the fluorochrome Fura-2. A suspension of 1 ml cells was loaded with 5 ␮l Fura-2 (1 mM stock) and left for about 30 minutes at room temperature to load. Cells were excited at 340/380 nm and 50 times per second, and fluorescence intensity was recorded at between 410 and 480 nm. The ratio of emission intensity when excited at 340 and 380 nm was used as an index of [Ca2⫹]i. The Fura-2 signal was calibrated using solutions of varying [Ca2⫹] in the absence of cells, as described previously.9 Data presentation and calculations. Data groups are presented as the mean ⫾ SD, except for values of ATP generated current when a median value (⫾ 25% IQs) is quoted because data were not normally distributed. Significance between data sets was tested using Student’s t test with null hypothesis rejected at p ⬍0.05. Specific membrane resistance (Rm) in ⍀ cm2 was calculated from the relationship, ␶m ⫽ Rm. Cm, where ␶m is the time constant of membrane potential (Em) change after injection of a small current into the cell and Cm is specific membrane capacitance, equal to 1 ␮F cm⫺2.10 Cell surface area (a) was calculated from cell capacitance (cm) by the relationship, a ⫽ cm/Cm. Rm was also calculated from the slope of the steady-state current-voltage relationship using the formulas, (rm ⫽ Em/I) and Rm ⫽ rm ⫻ a.

Cell characteristics. Figure 1 shows examples of suburothelial cells used for experiments. The spindle-shaped cell was typical of those used for experiments (fig. 1, A) and samples of these cells always stained for vimentin. Figure 1, B shows for comparison a cell isolated from the smooth muscle layer at the same magnification. Such cells never stained for vimentin. Membrane electrical properties were recorded cell capacitance, cm (27 ⫾ 16 pF, n ⫽ 53) was measured after membrane rupture with patch electrodes, and is small compared to smooth muscle cells. It yielded a cell surface area of 2.7 ⫻ 10⫺5 cm2. Membrane resistance was estimated by the passage of small depolarizing or hyperpolarizing currents (ie ⌬Em ⬍ ⫾5 mV). 1) From the time constant (␶) of the Em change ␶ was 90 ⫾ 48 milliseconds in 15 preparations and at Cm ⫽ 1 ␮F.cm⫺2 mean Rm was 9.0 ⫻ 104 ⍀ cm2. 2) It was also estimated from the slope of the steady-state current-voltage relationship with the value of rm ⫽ 3.1 ⫾ 1.1 ⫻ 109 ⍀ in 15 preparations, from which mean Rm ⫽ 3.1 ⫻ 109 ⫻ 2.7 ⫻ 10⫺5 ⫽ 8.5 ⫻ 104 ⍀ cm2. Resting membrane potential. In 16 cells Em was recorded in current clamp (I ⫽ 0 mode). Figure 2 shows that Em was not quiescent. The average Em per cell was obtained from a section of recording over several tens of seconds immediately after the patch electrode had gained access to the intracellular space. In 16 cells the average value was ⫺63 ⫾ 13 mV. Figure 2, A shows an example of a fairly stable resting potential with occasional spikes of depolarization. Such a pattern was observed more frequently in cells with a more negative Em value. Spikes rarely achieved an over shoot

FIG. 2. Membrane potential records of 2 suburothelial cells recorded under current clamp (Ih ⫽ 0) using KCl filled patch electrodes. A, quiescent cell with occasional spikes of activity. B, cell with more random fluctuations in membrane potential.

940

ELECTROPHYSIOLOGY OF SUBUROTHELIAL CELLS

FIG. 3. Ionic currents measured in suburothelial cells. A, inward current evoked by depolarization from ⫺100 mV to ⫺20 or 0 mV. Inset, early period of similar voltage clamp steps from ⫺40 to 0 mV. B, outward currents in another cell evoked by depolarization from ⫺100 mV to ⫺40, ⫺20 and 0 mV. C, effect of 30 mM TEA on outward currents in 3 superimposed traces on depolarization from ⫺100 to 0 mV before, during and after TEA application. D, Mean current-voltage relationship ⫾ SD of steady-state outward current. Outward current was magnitude of total time averaged integral throughout clamp step.

potential and they more generally peaked between about ⫺45 and ⫺25 mV. Figure 2, B shows that with cells at less negative potentials spike activity was much less apparent and a more randomly fluctuating pattern was observed. Ionic currents. Figure 3, A shows an example of ionic currents recorded at depolarization from a holding potential of ⫺100 mV. In this example small transient inward currents were followed by sustained outward currents. The inward current peaked at about ⫺10 mV and it was smaller at 0 mV as the larger outward current overlaid more the inward component. A detailed characterization of the current was not done. However, the current was reversibly attenuated when extracellular Ca was omitted from the superfusate. The outward current in the example described was a smooth trace. However, figure 3, B shows that in most cells additional large transient outward currents were also seen. They obscured any inward current since they tended to be larger at the beginning of the depolarizing step and, thus, they explained why net inward current was only recorded in a minority of cells. The magnitude of these outward transients was greater with the larger depolarizing steps in the range examined from ⫺50 to 0 mV. Figure 3, C shows that outward current transients were reversibly attenuated by 30 mM tetraethylammonium chloride, while they affected the steady-state current much less. Figure 3, D shows a current-voltage relationship for the time averaged outward current. The current showed outward going rectification, ie it was larger as the depolarizing step increased, and it showed a reversal potential at about ⫺80 mV, which further suggests that it was a K⫹ current. Actions of ATP on [Ca2⫹]i and membrane current. The effect of ATP on [Ca2⫹]i and membrane current was measured because the purine is proposed to be released from the urothelium when the bladder wall is stretched and, therefore, it could influence the activity of the myofibroblast layer. Figure 4, A shows a recording of [Ca2⫹]i during exposure to 30 ␮M ATP. A large transient increase in [Ca2⫹]i was recorded, which returned to baseline in the continued presence of ATP. The Ca2⫹ transient did not show an under shoot at

return to baseline. Similar transients were observed in all 6 cells tested, originating from a baseline [Ca2⫹]i of 95 ⫾ 13 nM to a peak of 708 ⫾ 154 nM. ATP was also applied to several epithelial cells from the urothelium/suburothelium biopsy but no Ca2⫹ transients were observed. Figure 4, B shows the effect of 100 ␮M ATP on net mem-

FIG. 4. Effect of ATP. A, intracellular Ca2⫹ transient at application of 30 ␮M ATP in isolated cell. Cell was held under voltage clamp at ⫺60 mV throughout experiment. B, inward current generated during superfusion with 100 ␮m ATP.

ELECTROPHYSIOLOGY OF SUBUROTHELIAL CELLS

brane current. The cell was voltage clamped at a potential of ⫺60mV, similar to the average resting membrane potential of these cells. After a delay ATP generated an inward current accompanied by a number of smaller transients. Similar results were seen in 18 of 26 cells that were exposed to 100 ␮M ATP. The median maximum peak inward current was 41 pA (23,79 pA IQs). Of the remainder no response was observed in 3 cells and in the remainder a small outward current was measured. All cells were voltage clamped at ⫺60 mV during ATP exposure. DISCUSSION

We isolated cells from the urothelial layer of the bladder that were distinct from epithelial cells. They were spindleshaped but distinct from smooth muscle cells due to their small size and shape (fig. 1). Moreover, they stained for the intermediate filament vimentin, which was not so with detrusor smooth muscle cells and, thus, they were not contaminants from the underlying detrusor. However, the cell population may have some heterogeneity because not all responded to the addition of ATP, for example. It may represent different stages in differentiation. Our previous study described a suburothelial cell layer in the bladder wall, identified as myofibroblasts, which was the exclusive site for vimentin staining as well as connexin 43.4 In this study desmin was confined to the base of the epithelial layer and always absent from the suburothelial layer. The fact that the cell population used for these experiments but not flat epithelial cells or larger smooth detrusor cells also stained for vimentin suggests that they are from this suburothelial population, although more extensive characterization is still necessary. Electrical characteristics of the cells. The cells had many characteristics found in excitable cells. Resting membrane potential was about ⫺60 mV. Many cells had spikes which, if not all or none action potentials because of the inconsistency of the amplitude, at least represented regenerative re-

941

sponses. Inward current preceded outward current under voltage clamp. Inward current was generated at application of an agonist. These characteristics and the abundance of the gap junctions protein connexin 43, which was identified in a previous study,4 mean that it is not unreasonable to propose that these cells could form an electrical network that would distribute an electrical signal over a reasonable area at focal depolarization. The distribution of intracellular electrical signals via gap junctions from cell to cell over a significant area would be facilitated by the relatively high membrane resistance. The value of the specific membrane resistance was similar to that of detrusor smooth muscle.11 The quantification of the spread of electrical signals is described. Ionic currents and membrane potential. Using KCl filled electrodes many cells generated transient net inward current in the range of ⫺40 to 0 mV before larger outward current developed. The current depended partially on extracellular Ca2⫹ and with its voltage dependence11, 12 it suggests that it is at least in part a Ca2⫹ current. Outward current showed outward going rectification at positive potentials and in many examples demonstrated large superimposed transients. These transients in particular were susceptible to extracellular tetraethyl ammonium ions, suggesting that a significant fraction of outward current is carried by a Ca2⫹ dependent fraction, as in for example detrusor smooth muscle.11, 12 Together these data suggest that these cells can generate transient electrical responses, which was confirmed under current clamp (fig. 2). The membrane potential of about ⫺60 mV is sufficient to allow L-type Ca2⫹ channels to recover from inactivation and, thus, they would be available to support the depolarizing phases of such spikes. Effects of ATP. Interest in the action of ATP stems from the hypothesis that stretch of urothelium releases purine, resulting finally in sensory nerve activation.13 The suburothelial network is ideally placed immediately below the urothelium to act as an intermediate cell layer that could contract and exert stretch dependent activation of the sensory nerves.

FIG. 5. Two-dimensional voltage distribution in network of electrically coupled cells. A, cell network is modeled as disk with current spreading radially from central point. B, voltage distribution as function of radial distance (r) from current source supplied by 1, 10 or 250 cells (50 pA per cell, Ri ⫽ 1,000 ⍀ cm and 10 ␮m cell layer). Vertical dotted line indicates distance from current source when change in membrane potential equals 5 mV. C, effect of Ri at 250, 1,000, 2,000 and 5,000 ⍀ cm on radial distribution of voltage from current source of 250 cells in 10 ␮m layer. D, influence of cell layer thickness (10, 50 and 100 ␮m) on voltage distribution from current source of 250 cells (Ri ⫽ 2,000 ⍀ cm).

942

ELECTROPHYSIOLOGY OF SUBUROTHELIAL CELLS

This observation also suggests a role for nitrergic activation6 because it should raise intracellular cyclic nucleotides and in turn modulate contractile sensitivity. Thus, nitrergic nerves could act as gain controllers for this system. Electrical responses would amplify the signal prior to afferent activation by transmission through a functional syncitium of these cells via connexin 43 gap junctions. We emphasize that it is not clear whether the electrical responses are causal or follow the increase in [Ca2⫹]i. If causal, it would imply that purinoceptors are of the P2X ionotropic family.14 If they follow the increase in [Ca2⫹]i, it implies that they may be generated by Ca2⫹ activated ion channels. The latter possibility is suggested by the somewhat variable nature of the membrane current responses to ATP, ie inward current in most cells but outward current in others because Ca2⫹ activated ion channels that carry inward and outward current have been described. Evaluation of this question is an important objective. The role of suburothelial cells in the integration of electrical signals. The extent to which an electrical signal spreads after focal activation of a region and the number of cells contributing to the initiation of such a signal may be calculated from the electrical characteristics of these cells and the magnitude or current evoked on application of agonists such as ATP. The analysis assumes that the cell layer is a finitely thick disk, so that current spreads in 2 dimensions from a depolarized central point (fig. 5, A). The Appendix shows an exact solution. Figure 5, B to D shows plots of depolarization magnitude (⌬Ed) at various distances from the central point after generation by each cell of 50 pA, as with the application of ATP (fig. 4). The dotted lines show the distance at which ⌬Ed was 5 mV, a lower limit at which inward current might be evoked and a regenerative response might be evoked. Figure 5, B shows the voltage variation for initiation of current in 1, 10 or 250 cells at an intercellular coupling resistance (Ri) of 1,000 ⍀ cm and a cell layer of 10 ␮m. Only simultaneous activation of many cells generates ⌬Ed ⬍5 mV, and then it would only propagate about 200 ␮m. Figure 5, C shows that ⌬Ed depends on the value of Ri. In these examples increasing Ri from 1,000 to 2,000 and 5,000 ⍀ cm increased ⌬Ed ⫽ 5 mV to 500 and 820 ␮m, respectively, while a decrease to 250 ⍀ cm drastically decreased depolarization. Figure 5, D shows the effect of the functional thickness of the layer as it increases from 10 to 50 and 100 ␮m radial depolarization is successively attenuated. While they are only estimates, several valid conclusions may be drawn that aid in the interpretation of the functional properties of this cell layer. 1) Depolarization of a single cell or a small group of cells is insufficient to generate even local significant depolarization in a cellular network because intracellular electrical coupling between adjacent cells is sufficient to prevent local accumulation of charge and, hence, depolarization. Thus, these cells are unlikely to act individually as focal pacemakers, but rather require a larger number to be depolarized simultaneously. 2) The Ri value between cells is critical to the extent of depolarization spread. A value of 1,000 ⍀ cm is almost that of detrusor smooth muscle.15 As Ri increases, the spread of depolarization also increases or a smaller number of cells would generate significant depolarization. Thus, determining the exact value of Ri remains a critical experimental objective. However, even at a large Ri value a single cell is still unlikely to evoke significant depolarization, so that the cells must be regarded as working in a network. 3) The thickness of the layer is important. A thicker layer provides a more 3-dimensional structure and greater local decrement of depolarization. It is crucial to quantify depolarization in the depth of the layer that may be achieved only using a network of these cells.

CONCLUSIONS

Overall these data and associated studies show that a layer of suburothelial, nonepithelial bladder cells have electrical and structural characteristics that allow them to act as an electrical network. Although many details require elucidation, to our knowledge we present the first characterization of these cells, which we propose have a controllable step in the process of bladder sensory function. Chemicals were obtained from Sigma, Poole, United Kingdom. APPENDIX: 2-DIMENSIONAL CURRENT FLOW IN A DISK-LIKE LAYER OF CELLS

The model assumes that depolarizing current is generated at a central location and spreads radially and homogeneously to other cells in the discoid network (thickness, b). The solution will calculate the change in voltage (membrane potential) as a function of radial distance (r). Current is assumed to flow in the intracellular space that offers resistance (Ri in ⍀ cm) and it is lost to the extracellular space through cellular membranes that have a specific resistance (Rm) and capacitance (Cm). The general time independent differential equation for voltage distribution as a function of distance in a 2Ri 1 dV d2 V ⫺ ⫹ V ⫽ 0,16 which 2-dimensional space is dr2 r dr bRm i 0 Ri has a solution V共r兲 ⫽ 䡠 K 共r/␭兲 with the boundary condibRm 0 tion V(r) 3 0, r 3 ⬁. K0 is a zero-order modified Bessel function of the second kind,17 i0 is the current generated at the origin and ␭2 ⫽ (bRm/2Ri). Voltage distribution, V(r), may be calculated since all variables may be experimentally determined (i0, Rm, Ri and b). Bessel function values may be looked up in standard tables. In this case they were calculated using the Excel (Microsoft, Redmond, Oregon) function BESSELK(r/␭,0). Plots of V(r) as a function of r were plotted in figure 5 with variation of the described parameters to investigate those that most influence voltage distribution. REFERENCES

1. Ferguson, D. R., Kennedy, I. and Burton, T. J.: ATP is released from rabbit urinary bladder epithelial cells by hydrostatic pressure changes—a possible sensory mechanism? J Physiol, 505: 503, 1997 2. Lee, H. Y., Bardini, M. and Burnstock, G.: Distribution of P2X receptors in the urinary bladder and the ureter of the rat. J Urol, 163: 2002, 2000 3. Cockayne, D. A., Hamilton, S. G., Zhu, Q.-M., Dunn, P. M., Zhong, Y., Novakovic, S. et al: Urinary bladder hyporeflexia and reduced pain-related behaviour in P2X3-deficient mice. Nature, 407: 1011, 2000 4. Sui, G. P., Rothery, S., Dupont, E., Fry, C. H. and Severs, N. J.: Gap junctions and connexin expression in human suburothelial interstitial cells. BJU Int, 90: 118, 2002 5. Wiseman, O. J., Fowler, C. J. and Landon, D. N.: The role of the human bladder lamina propria myofibroblast. BJU Int, 91: 89, 2003 6. Smet, P. J., Jonavicius, J., Marshall, V. R. and de Vente, J.: Distribution of nitric oxide synthase-immunoreactive nerves and identification of the cellular targets of nitric oxide in guinea-pig and human urinary bladder by cGMP immunohistochemistry. Neuroscience, 71: 337, 1996 7. Sergeant, G. P., Hollywood, M. A., McCloskey, K. D., Thornbury, K. D. and McHale, N. G.: Specialised pacemaking cells in the rabbit urethra. J Physiol, 526: 359, 2000 8. Hirst, G. D. and Edwards, F. R.: Generation of slow waves in the antral region of guinea-pig stomach—a stochastic process. J Physiol, 535: 165, 2001 9. Wu, C., Sui, G. P. and Fry, C. H.: The role of the L-type Ca(2⫹) channel in refilling functional intracellular Ca(2⫹) stores in guinea-pig detrusor smooth muscle. J Physiol, 538: 357, 2002 10. Weidmann, S.: Electrical constants of trabecular muscle from mammalian heart. J Physiol, 210: 1041, 1970

ELECTROPHYSIOLOGY OF SUBUROTHELIAL CELLS 11. Sui, G.-P., Wu, C. and Fry, C. H.: The electrophysiological properties of cultured and freshly isolated detrusor smooth muscle cells. J Urol, 165: 627, 2001 12. Montgomery, B. S. I. and Fry, C. H.: The action potential and net membrane currents in isolated human detrusor smooth muscle cells. J Urol, 147: 176, 1992 13. Ferguson, D. R.: Urothelial function. BJU Int, 84: 235, 1999 14. Burnstock, G.: P2 purinoceptors: historical perspective and classification. Ciba Found Symp, 198: 1, 1996

943

15. Fry, C. H., Cooklin, M., Birns, J. and Mundy, A. R.: Measurement of intercellular electrical coupling in guinea-pig detrusor smooth muscle. J Urol, 161: 660, 1999 16. Jack, J. J. B., Noble, D. and Tsien, R. W.: Electric Current Flow in Excitable Cells. Oxford, United Kingdom: Clarendon Press, 1966 17. Abramovitz, M. and Stegun, I. A.: Handbook of Mathematical Functions with Formulas, Graphs and Mathematical Tables. Washington, D. C.: National Bureau of Standards, U. S. Government Printing Office, vol. 55, 1964