Spontaneous electrical and Ca²⁺ signals in typical and atypical smooth muscle cells and interstitial cell of Cajal-like cells of mouse renal pelvis

Abstract

Electrical rhythmicity in the renal pelvis provides the fundamental drive for the peristaltic contractions that propel urine from the kidney to bladder for storage until micturition. Although atypical smooth muscles (ASMCs) within the most proximal regions of the renal pelvis have long been implicated as the pacemaker cells, the presence of a sparsely distributed population of rhythmically active Kit-positive interstitial cells of Cajal-like cells (ICC-LCs) have confounded our understanding of pelviureteric peristalsis. We have recorded the electrical activity and separately visualized changes in intracellular Ca²⁺ concentration in typical smooth muscle cells (TSMCs), ASMCs and ICC-LCs using intracellular microelectrodes and a fluorescent Ca²⁺ indicator, fluo-4. Nifedipine (1–10 μm)-sensitive driven action potentials and Ca²⁺ waves (frequency 6–15 min⁻¹) propagated through the TSMC layer at a velocity of 1–2 mm s⁻¹. High frequency (10–40 min⁻¹) Ca²⁺ transients and spontaneous transient depolarizations (STDs) were recorded in ASMCs in the absence or presence of 1 μm nifedipine. ICC-LCs displayed low frequency (1–3 min⁻¹) Ca²⁺ transients which we speculated arose from cells that displayed action potentials with long plateaus (2–5 s). Neither electrical activity propagated over distances > 50 μm. In 1 μm nifedipine, ASMCs or ICC-LCs separated by < 30 μm displayed some synchronicity in their Ca²⁺ transient discharge suggesting that they may well be acting as ‘point sources’ of excitation to the TSMC layer. We speculate that ASMCs act as the primary pacemaker in the renal pelvis while ICC-LCs play a supportive role, but can take over pacemaking in the absence of the proximal pacemaker drive.

For the last 35 years it has been thought that pyeloureteric peristalsis consists of propagating muscle contractions triggered by electrically active atypical smooth muscle cells (ASMCs) within a diffuse layer of cells lying adjacent to the typical smooth muscle cell (TSMC) layer which forms the funnel-shaped pelviureteric complex. These ASMCs are located predominately within the proximal regions of the renal pelvis at the base of the papilla and not in the ureter (Dixon & Gosling, 1973, 1982; Gosling & Dixon, 1974; Klemm et al. 1999; Lang et al. 2001). Atypical SMCs display many of the morphological features of cardiac sino-atrial pacemaker cells having a small nucleus and long branching processes that contain relatively few parallel contractile filaments. Intracellular microelectrode recordings, particularly in the most proximal regions of the renal pelvis (pelvicalyceal junction; PCJ) of rat and guinea pig (Klemm et al. 1999; Lang et al. 2001), reveal that short spindle-shaped ASMCs display high frequency (10–30 min⁻¹) spontaneous transient depolarizations (STDs) of a simple waveform that are reduced but not blocked by nifedipine (Tsuchida & Suzuki, 1992; Lang et al. 1995; Lang et al. 2001; Lang et al. 2006b).

More recently, evidence has been presented that a distinct population of cells with many of the distinctive morphological features of interstitial cells of Cajal (ICC), the pacemakers of the gastrointestinal (GI) tract, are sparsely distributed within the lamina propria and muscle layers of the renal pelvis and proximal ureter of both uni-calyceal and mutli-calyceal mammals. In mouse (Pezzone et al. 2003; Lang & Klemm, 2005; Lang et al. 2006b; Lang et al. 2007), human (Metzger et al. 2005) and rat (Metzger et al. 2004) but not guinea pig (Klemm et al. 1999), these ICC-like cells (ICC-LCs) are immuno-reactive to antibodies raised against Kit, the proto-oncogene for the receptor-tyrosine kinase which binds stem cell factor essential for the development of mast cells and other haematopoietic cells in the urinary tact and ICC in the GI tract. Indeed Kit immuno-reactive ICC-LCs appear in mouse embryonic ureter in culture at the same time as coordinated unidirectional peristaltic contractions in a manner blocked by the Kit antibody, ACK45 (David et al. 2005). Single enzymatically isolated ICC-LCs of the mouse renal pelvis have also recently been demonstrated to display autorhythmicity in the form of spontaneously occurring large long-lasting inward currents that are cation selective (Lang et al. 2007). These spontaneous inward currents could well provide a pacemaker drive for ureteric peristalsis, particularly after pyeloplasty or ureteral obstruction, conditions that would disconnect the ureter from its proximal ASMC pacemaker drive.

In this report we have used electrophysiological and Ca²⁺ fluorescence imaging to ascertain the primary role of ASMCs and ICC-LCs in the initiation of pelviureteric peristaltic contractions in the mouse renal pelvis. We observed propagating Ca²⁺ waves in TSMCs within the muscle wall with frequencies, time courses and conduction velocities similar to those recorded for propagating action potentials and muscle contraction (Klemm et al. 1999). We have also visualized spindle shaped ASMCs and fusiform ICC-LCs which display their own autorhythmicity, firing Ca²⁺ transients with 10-fold differences in their frequency and duration, which matched the parameters of non-propagating STDs and low frequency long plateau action potentials, respectively, recorded with intracellular microelectrodes. Ca²⁺ transients in ASMCs and ICC-LCs did not propagate over distances > 50 μm.

It was concluded that muscle contraction arises from Ca²⁺ entry through L-type Ca²⁺ channels which are opened during the time course of TSMC action potentials that freely propagate the length of the renal pelvis and blocked by relatively high concentrations of nifedipine (1–10 μm). It seems likely that TSMC action potentials were triggered by STDs (2–40 mV) arising in ASMCs which are acting as ‘point sources’ of excitation to evoke driven action potential discharge in the TSMC layer via gap junctions. The temporal characteristics of Ca²⁺ transients in ICC-LCs were correlated with the long plateau action potentials which did not evoke muscle wall contraction. Thus the initiation and propagation of autorhythmicity in the upper urinary tract bears little resemblance to ICC dependent mechanisms well characterized in the gastrointestinal tract. As in the bladder (Hashitani et al. 2004b) or corpus cavernosum (Hashitani et al. 2005), ICC-LCs in the renal pelvis may well play a supporting rather than an initiating role in muscle wall contraction and pelviureteric peristalsis.

Methods

Conventional Swiss o/b male mice, 4–6 weeks in age, were killed by cervical dislocation and exsanguination and the kidneys and attached ureters removed through an abdominal incision, using procedures approved by the Physiological Department Animal Ethics Committee at Monash University and Nagoya City University. The kidney was bathed in a bicarbonate buffered physiological salt solution (PSS). The upper urinary tract, from its point of attachment to the papilla and calyx (PCJ) to the pelviureteric junction, was dissected free of the kidney, opened along its longitudinal axis and loosely pinned out in a dissecting dish with the urothelial layer uppermost.

Intracellular microelectrode and tension recordings

Strips (2 × 5 mm²) of transversely cut portions of proximal or mid renal pelvis, or longitudinal full length strips of the renal pelvis (containing a portion of PCJ) were dissected free and one end was firmly pinned, urothelial side uppermost, into a silicone resin (Sylgard, Dow Corning Corp., Midland, MI, USA) coated recording chamber while the other end was attached to a force transducer via a thread attachment. The bath was then mounted on an inverted microscope and superperfused with PSS at 3–5 ml min⁻¹ at 37°C. Electrophysiological recordings were made using glass microelectrodes with resistances of 80–120 mΩ when filled with 1 m KCl. Membrane potential changes was recorded with a high impedance Axoclamp-2 preamplifier (Axon Instruments/Molecular Devices, Union City, CA, USA), low pass filtered at 1 kHz and stored digitally with tension changes on a personal computer using a Digidata 1200 DMA analog-to-digital interface, Axotape 8 and pCLAMP 8 software (Axon Instruments) for later analysis. Microelectrode tips were often filled with the fluorescent dye Lucifer Yellow (Sigma Industries, Australia). Fluorescence-filled cells were examined at the end of each impalement using an excitation wavelength of 400–450 nm. Micrographs of emissions at wavelengths between 520 and 580 nm were captured onto computer for later examination.

Two microelectrode studies

Transversely or longitudinally cut strips (2 × 5 mm²) of the renal pelvis were pinned tightly to the bottom of the recording chamber and impaled with two independently positioned intracellular microelectrodes (V₁ and V₂). Light micrographs were taken of the positions of the two electrodes allowing later analysis of their separation and orientation within the preparation. Conductivity between the two electrodes was examined using one electrode as a current passing electrode, while the other electrode was used to record the resulting electronic potentials at various points (< 200 μm) in both the transverse and longitudinal direction. The conduction velocity of spontaneous regenerative events that clearly propagated between the two recording sites were calculated from the distance between electrodes (10–3000 μm) and the interval between action potentials or from the inverse of the slope of the plot of pooled intervals against separation distance.

Measurements of Ca²⁺

Two (‘normal’ and ‘light’) loading protocols were used to specifically visualize the changes in the concentration of intracellular calcium ([Ca²⁺]_i) (Ca²⁺ transients) in TSMCs versus ASMCs and ICC-LCs, respectively. To visualize Ca²⁺ transients in the TSMC layer, preparations were pinned with the urothelium surface uppermost onto a block of Sylgard which had a window of 1.5 mm × 3 mm in the centre. To minimize Ca²⁺ transient distortion due to smooth muscle contractility, preparations were stretched using 15–20 tungsten wires (20 μm in diameter). After 30 min incubation in warmed (36°C) PSS and commencement of spontaneous muscle contractions, preparations were incubated in low Ca²⁺ PSS ([Ca²⁺]_o= 0.5 mm) containing 3–5 μm fluo-4 AM (FluoroPure, Molecular Probes, OR, USA) and cremophor EL (0.01%, Sigma) for 45–60 min at room temperature (‘normal’ loading). Following incubation, preparations were washed with dye-free, warmed PSS at a constant flow (2 ml min⁻¹) for 30 min.

To preferentially visualize Ca²⁺ signals in ASMCs and ICC-LCs in situ, preparations were pinned onto the sylgard square and incubated in low Ca²⁺ PSS ([Ca²⁺]_o= 0.5 mm) containing fluo-4 AM (0.1–1 μm) and cremphor EL (0.01%) for 15–30 min at 36°C (‘light’ loading). Following incubation, the preparations were washed with dye-free PSS (36°C) for 30 min. After either loading protocol the recording chamber was mounted on the stage of an inverted fluorescence microscope (IX70, Olympus) equipped with an electron multiplier CCD camera (C9100, Hamamatsu Photonics) and a high speed scanning polychromatic light source (C7773, Hamamatsu Photonics). Preparations were viewed with either a water immersion ×60 objective (UPlanApo 60, Olympus) or an air ×20 objective (UPlanApo 20, Olympus) and illuminated at 495 nm. The fluorescence emissions in desired areas as selected by the fluorescence capture software were measured through a barrier filter above 515 nm (sampling interval 23–200 ms), using a micro photoluminescence measurement system (Aquacosmos, Hamamatsu Photonics). Relative changes in [Ca²⁺]_i were expressed as the ratio (F_t/F₀) of the fluorescence generated by an event at time t (F_t) and the baseline fluorescence at t = 0 (F₀).

Immunohistochemistry of whole mounts of renal pelvis

Immuno-staining of unfixed whole mounts of proximal and mid renal pelvis for Kit positive cells, macrophages and cell nuclei was carried out as recently described for the ureteropelvic junction and proximal ureter (Lang et al. 2006, 2007) using Kit antibody (1 : 200 rabbit anti-human H-300; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA), FITC-dextran FD-70S (1.3 mg ml⁻¹ Sigma Industries, Australia) and Hoechst Blue. Cells were examined under bright field and with fluorescence microscopy; micrographs of immuno-reactive cells were taken separately for each fluorophore and later superimposed. The number and orientation of Hoechst Blue stained urothelial and SMC nuclei was often used to indicate the likely position of Kit-positive cells within the UPJ wall. Contractile filament-rich TSMC and ASMC with sparsely distributed contractile proteins were visualized by immuno-staining paraformaldehyde (4% in 0.1 m phosphate buffer) fixed whole mounts of renal pelvis for α-smooth muscle actin as previously described (Klemm et al. 1999).

Solutions and drugs used

The bicarbonate buffered PSS (PSS) for the electrophysiological experiments had the following composition (mm): NaCl 120, KCl 5, NaHCO₃ 25; CaCl₂ 2.5, MgSO₄ 1–2, KH₂PO₄ 1, NaHCO₃ 25 and glucose 11, bubbled with a 95% O₂–5% CO₂ gas mixture to establish a pH of 7.3–7.4. The concentration of all stock solutions ranged between 0.1 mm and 10 mm. Most drugs were dissolved in filtered distilled water and diluted with PSS to their final concentrations as indicated. Nifedipine (Sigma, St Louis, MO, USA) was dissolved in absolute ethanol or DMSO. Stock solutions were generally added 1 : 1000 dilution; 0.1% ethanol and DMSO had no effect on the recorded electrical activity.

Data analysis

Regenerative action potentials, STDs, contractions and Ca²⁺ transients were identified for analysis by the creation of unique templates using pCLAMP 9 software as previously described (Lang et al. 2006a). Various parameters of these captured events (n = 2–90) were measured and averaged under each experimental condition: baseline fluorescence or resting membrane potential, amplitude, time to peak amplitude, half-amplitude duration (1/2 width), maximal rate of rise of the initial phase, and interval, which was often converted to frequency (min⁻¹). The synchronicity of Ca²⁺ signals in ASMCs or ICC-LCs was analysed using the cross-correlation function of Clampfit 9 software (Axon Instruments). Synchronicity was estimated from the peak amplitude of the cross-correlation factor (CCF) plotted against lag period (between −4000 and +4000 ms) of Ca²⁺ signals for pairs of cells separated by a known distance.

Data from a number of similar experiments were averaged as indicated to provide a mean ± standard deviation of the mean. N denotes the number of tissues, n denotes the number of observations within a single experiment. In most experiments, paired or unpaired Student's t tests were used for tests of significance; P < 0.05 was accepted as statistically significant.

Results

Relationship between electrical and mechanical activity

Transversely and longitudinally cut strips of renal pelvis both developed spontaneous contractions within 5–15 min of being placed in the recording chamber and washed with PSS at 37°C. As previously described (Constantinou et al. 1978; Hannappel et al. 1982; Lang et al. 1998) the frequency of contraction of transversely cut strips decreased with distance from the PCJ. In contrast, longitudinally cut strips that contained a portion of PCJ contracted at the same high frequency as transversely orientated strips cut from the most proximal renal pelvis. In six longitudinal strips, spontaneous contractions had a mean amplitude, half-amplitude duration, maximal rate of rise and frequency of 0.07 ± 0.02 mN, 113.9 ± 107.7 ms, 0.032 ± 0.004 mN ms⁻¹ and 13.07 ± 3.09 min⁻¹, respectively (mean ±s.d. of the mean, N = 6). These parameters were compared with the contractions recorded in nine transverse strips cut from the mid renal pelvis which had a mean amplitude, half-amplitude duration, maximal rate of rise, and frequency of 0.24 ± 0.2 mN, 399 ± 163 ms (unpaired t test P < 0.05), 0.027 ± 0.019 mN ms⁻¹ and 6.05 ± 2.1 min⁻¹ (P < 0.05), respectively.

TSMC action potentials

Intracellular microelectrode impalements of both transverse and longitudinally orientated strips of renal pelvis revealed three major patterns of electrical activity which were often recorded in the same preparation. The most common pattern of activity was characterized by the firing of regenerative action potentials that preceded muscle contraction on every occasion; the membrane potential between action potentials was mostly quiescent (Fig. 1A and B). The time course of these action potentials consisted of an initial spike followed by a plateau 200–800 ms in duration. In five typical impalements of transversely orientated strips of mid renal pelvis the averaged membrane potential, peak amplitude, half-amplitude duration and frequency of these action potentials were −42.8 ± 3.1 mV, 35.3 ± 4.3 mV, 442 ± 86 ms and 5.8 ± 1.1 min⁻¹, respectively.

Figure 1.

General electrical mechanical and morphological characteristics of the mouse renal pelvis A, intracellular microelectrode recordings of TSMC action potentials (Ai top panel) that were directly associated with muscle contraction (lower panel); dashed lines represent 0 mV. Aii, high frequency spontaneous transient depolarizations (STDs) not associated with muscle contractions were recorded in approximately 50% of impalements. Aiii, on eight occasions the recorded electrical activity was characterized by the presence of long plateau action potentials (at frequency of 1–3 min⁻¹) which did not trigger muscle contraction. B, greyscale fluorescence micrographs of fixed whole mount preparations of proximal (Bi) and mid (Bii) renal pelvis immuno-stained for α-smooth muscle actin. Ci, fluorescence micrographs of unfixed whole mount preparations of mouse proximal renal pelvis exposed to FITC-dextran FD-70S to label macrophages (Cii) and then the extracellularly binding Kit antibody H-300 (Ci). Ciii, superposition of Ci+Cii revealed that only < 50% of Kit-positive cells also displayed the FITC-dextran fluorescence (arrows). Calibration bars represent 50 μm.

When viewed under fluorescence illumination after filling with Lucifer Yellow, the cells firing these action potentials were spindle shaped in appearance, some 200–500 μm in length and 5–10 μm wide at the central nuclear region (data not shown). Fluorescence intensity decreased with distance from the point of impalement so that the ends of the cells were invariably not visible, which precluded any accurate measurement of cell lengths. The cells were similar in shape to the spindle cells in whole preparations that were immuno-positive for α-smooth muscle actin (Fig. 1B). As has been previously reported in the guinea pig and rat renal pelvis (Klemm et al. 1999; Lang et al. 2001), these regenerative action potentials were recorded more frequently with distance from the PCJ. α-Actin-immuno-staining of whole mount preparations also revealed an increased fluorescence with distance from the PCJ. α-Actin-positive cells formed a loosely packed basket weave network of short spindle shaped cells within the proximal regions of the renal pelvis (Fig. 1Bi). In the mid (Fig. 1Bii) and distal regions of the renal pelvis, α-actin-positive cells formed a layer of increasingly more tightly packed longer spindle-shaped cells. The shape and position of these cells allowed us, by analogy with previous microelectrode and histological examinations of the rat and guinea pig renal pelvis, to identify them as ‘typical’ smooth muscle cells (TSMCs) (Klemm et al. 1999; Lang et al. 2001).

STDs

Impalement of the most proximal regions of longitudinal strips also revealed cells that discharged STDs which were characterized by their high frequency (5–30 min⁻¹) and variable amplitudes. These STDs, although often summing to amplitudes > 20 mV, never evoked a driven action potential that triggered muscle contraction (Fig. 1B). The STDs illustrated in Fig. 1B that were not associated in time with muscle contraction had a mean peak amplitude, half-amplitude duration and maximal rate of rise of 17.6 ± 8.4 mV, 570 ± 141 ms and 0.85 ± 0.15 mV ms⁻¹, respectively (n = 30). The time course and parameters of the contractions and driven action potentials recorded in cells displaying STDs were similar to those in cells that were essentially quiescent between action potential discharge. In the transversely orientated strip cut from the mid renal pelvis illustrated in Fig. 1B the frequency of contraction and regenerative action potential discharge was 6 min⁻¹ while the frequency of STD discharge was 33 min⁻¹. The peak amplitude, half-amplitude duration and maximal rate of rise of these action potentials was 37.9 ± 0.9 mV, 352 ± 16 ms and 1.8 ± 3.8 mV ms⁻¹, respectively (n = 10); they triggered contractions with a half-amplitude duration of 463 ± 17 ms.

When filled with Lucifer Yellow and viewed with a fluoresecence microscope, the cells displaying STDS were spindle shaped, 20–50 μm long and 10–20 μm wide at the nucleus (data not shown). We have previously designated these short spindle shaped cells as ‘atypical’ smooth muscle cells (ASMCs; Klemm et al. 1999; Lang et al. 2001). However, in some recordings the impaled cell was clearly a longer TSMC. We have concluded that these impaled cells must have been electrically close to one or more active ASMCs.

ICC-LCs

The third population of electrical events recorded in the mouse renal pelvis was characterized by their very low frequency of discharge (< 4 min⁻¹), their very long plateaus (> 2 s) and the very low frequency at which we could record them. Figure 1Aiii is one representation of eight recordings of these long lasting action potentials and illustrates that, in contrast to driven action potentials, these plateaus were not associated with contraction of the renal pelvis smooth muscle. These plateau potentials were evident for the initial 10–30 min after impalement and tended to decline in duration with time until they were indistinguishable from the ongoing TSMC action potentials. Four cells which maintained a relatively constant half-amplitude duration over this time had a peak amplitude, half-amplitude duration and frequency of 37.9 ± 0.3 mV, 3.53 ± 1.28 s and 2.4 ± 1.3 min⁻¹, respectively – all significantly different from the same parameters measured from the driven action potentials recorded in the same impalement over the same time period (35.0 ± 4.6 mV, 0.53 ± 0.11 s and 11.2 ± 3.2 min⁻¹, respectively) (Fig. 9B).

Figure 9.

Summary of characteristics of electrical and Ca²⁺ signalling in TSMCs, ASMCs and ICC-LCs A, plots of mean interevent intervals against half-amplitude durations (1/2 width) of contractions (•) and TSMC action potentials (□) recorded in 5 preparations of the mouse renal pelvis. ○, a plot of the same parameters for Ca²⁺ waves recorded in 7 views from 4 preparations. B, plots of mean interevent intervals against 1/2 widths of TSMC action potentials (•) and long plateau action potentials (○) recorded simultaneously in 7 preparations. C, plot of mean interevent intervals against 1/2 widths of Ca²⁺ transients in ASMCs (•) and ICC-LCs (○) bathed in 1 μm nifedipine PSS in 7 preparations. D, comparison of mean interevent intervals of ASMCs and ICC-LCs (○) in 7 preparations. •, plot of mean interevent intervals of TSMC action potentials and contractions in 4 preparations for comparison. E, plots of peak cross-correlation factor (CCF) against separation distance (μm) between 32 pairs of ASMCs (Ei, N = 8) and ICC-LCs (Eii, N = 7). The straight line y = 0.284 + 0.000616x in Ei was fitted to pooled data by least squares regression.

Kit staining of unfixed whole mount preparations of proximal renal pelvis revealed punctate immuno-staining for Kit on both spindle- and stellate-shaped cells within the lamina propria and between the SMC bundles within the serosal surface of the proximal (Fig. 1Ci, N = 5) and mid (N = 7) renal pelvis. Kit immuno-positive spindle-shaped cells were often positioned in parallel to one another and to neighbouring SMC bundles. The stellate shaped cells had numerous short, often branching projections which did not appear to form close associations with neighbouring immuno-positive cells (Fig. 1Ci). As in the ureteropelvic junction and proximal ureter (Lang et al. 2006b; 2007), relatively few (1–5 in a field view of 10 cells) of the Kit-positive cells in the proximal renal pelvis also displayed the uptake of FITC-dextran FD-70S (Fig. 1Cii and iii). We speculate the infrequently recorded long plateau action potentials which do not trigger muscle contraction are in fact being recorded in either one of these sparsely distributed Kit-positive, FITC dextran-negative cells or a TSMC electrically close to one of these cells (Fig. 1Ciii).

Origin and propagation of electrical activity

When paired microelectrode (V₁ and V₂) recordings were made at various positions along the axial or transverse axes within longitudinal strips (separations 20–2000 μm), the most common recorded activity was spontaneous TSMC action potentials of similar amplitude recorded at each electrode with little decrement in their amplitude or time course (Figs 2Ab and c, and 3A). In five paired impalements (separation 500–2000 μm) in longitudinal strips of renal pelvis, the averaged membrane potential 500 ms before each action potential and the peak amplitude, half-amplitude duration, maximum rate of rise and discharge frequency of these propagating TSMC action potentials at microelectrode at V₁ were not significantly different (paired t tests all P > 0.05) from the equivalent parameters recorded at V₂ (Table 1). These data suggest that TSMC action potentials are regenerative and can propagate over long distances in the renal pelvis.

Figure 2.

Typical variation in recorded electrical activity in longitudinal preparations of mouse renal pelvis using single or pairs of intracellular microelectrodes A, the electrical activity at site b recorded with microelectrode V₁ is compared with the electrical recordings at sites a, c and d recorded with a second microelectrode, V₂. B, spontaneous action potentials could still be recorded in the distal portions of longitudinal preparations trisected into 3 equal sections. The frequency of driven action potential discharge in the most proximal section (Bia and iia) was little altered by the trisection, while the frequency of discharge in the most distal section was always less than the most proximal section (Biia and b).

Figure 3.

Propagation of driven action potentials in longitudinally cut strip of renal pelvis Ai–iii, TSMC action potentials were recorded with two intracellular microelectrodes; V₁ was held constant while V₂ was moved to various positions along the longitudinal axis. B, plot of pooled intervals between action potential discharge recorded at V₁ and V₂ orientated along the longitudinal (•) and transverse (○) axes against the separation between V₁ and V₂ (10–2500 μm). The conduction velocity of 1.65 mm s⁻¹ was obtained from the inverse of the slope of the straight line fitted to the pooled data obtained in the longitudinal axis. ▵, intervals versus separation plot for Ca²⁺ waves recorded in the TSMC layer for comparison.

Table 1.

Comparison of parameters of propagating TSMC action potentials recorded simultaneously with two intracellular microelectrodes (V₁ and V₂)

	Membrane potential (mV)	Amplitude (mV)	Half-amplitude duration (ms)	Max. rate of rise (mV s−1)	Frequency (min−1)
V1	−46.6 ± 4.7	32.8 ± 5.3	354 ± 70	1.2 ± 0.1	13.9 ± 6.3
V2	−57.4 ± 12.7	35.6 ± 8.8	329 ± 73	1.1 ± 0.2	14.7 ± 3.5

In transversely cut proximal strips of proximal renal pelvis (containing a portion of PCJ) it was evident that more than one region displayed high frequency STDs and that only one of these active regions drove the more distal regions of the muscle strip at any one time. Indeed paired microelectrode recordings in the proximal renal pelvis revealed that these high frequency STDs did not propagate over distances > 100 μm (Fig. 2Aa and b) (see below).

On most, but not every, occasion the action potentials recorded at the proximal electrode preceded the action potentials recorded more distally, indicating that the dominant pacemaker lay in the proximal end of the muscle strip (Fig. 2ivb and d). When longitudinal preparations were trisected into three equal portions, spontaneous electrical discharge continued in the most proximal section of the renal pelvic strip (Fig. 2Bia and iia). In 3 of 4 preparations, TSMC action potentials were also recorded in the two distal portions albeit at a reduced frequency of that recorded in the same regions when the tissue was intact (Fig. 2Biib). In the fourth preparation, low frequency action potential discharge was recorded in the middle portion of the trisected strip, while the distal third of the preparation was electrically quiescent. Thus in most preparations proximal regions provided the predominant pacemaker drive. However dislocation of these proximal pacemaker regions can unmask the presence of other pacemaker regions in more distal portions of longitudinal strips (Zhang & Lang, 1994).

The conduction velocity of TSMC action potentials propagating in the longitudinal direction was 1.4 ± 0.3 mm s⁻¹ (n = 12, range 0.5–4.5 mm s⁻¹), not significantly different from the conduction velocity of action potentials propagating in the transverse direction, 2.0 ± 1.2 mm s⁻¹ (n = 5, range 0.3–6.7 mm s⁻¹). A conduction velocity of 1.65 mm s⁻¹ was also obtained from the inverse of the slope (0.61 s mm⁻¹) of the straight line fitted to pooled action potential intervals (N = 7, n = 12) recorded in the longitudinal direction plotted against the distance between electrodes (filled circles Fig. 3B).

Effects of nifedipine

Blockade of smooth muscle L-type Ca²⁺ channels with nifedipine (1 μm) has allowed the study of the properties of the pacemaker cells responsible for autorhythmicity in many gastrointestinal and urogenital smooth muscle organs. Such studies have relied on the near complete blockade of Ca²⁺ entry into the smooth muscle cells and an appreciable electrical conduction within a ‘short cable’ syncytium so that pacemaker potentials can be recorded, whether impalements were in a pacemaker ICC or in a neighbouring smooth muscle bundle (Dickens et al. 1999; Hirst & Ward, 2003). We have made an extensive study of the effects of nifedipine on the electrical and contractile properties of the mouse renal pelvis and have found that the effects of nifedipine are far more variable than in other smooth muscle preparations.

In 10 preparations, spontaneously occurring TSMC action potentials and their associated contractions were only reduced by 1 μm nifedipine (for > 5–60 min)(Fig. 4Ai and ii). In five of these experiments, the peak amplitude and half-amplitude duration of these propagating action potentials (N = 5, n = 6–52) and the contractions they evoked were significantly (paired t test P < 0.05) reduced after > 10 min exposure to 1 μm nifedipine. However their frequency of discharge was not significantly affected (Table 2). It is clear that nifedipine significantly altered the shape of the action potential, reducing the right hand shoulder, which resulted in a smaller after-hyperpolarization component to reveal a more triangular shaped action potential.

Figure 4.

Effects of ‘L-type’ Ca²⁺ channel blockade with nifedipine on TSMC action potentials, STDs and long plateau action potentials A, TSMC action potentials were either reduced by 1–10 μm nifedipine in a concentration dependent manner (Ai–iv) or completely blocked by 1 μm nifedipine (Ci and ii V₁). B, STDs were little affected by nifedipine (1–10 μmBi–iii) but were blocked upon replacing the Ca²⁺ concentration in the bathing medium with an equimolar concentration of Mg²⁺ (Biv). C, long plateau action potentials recorded with V₂ did not propagate over any appreciate distance (V₁) and were little affected by 1 μm nifedipine, a concentration that readily blocked propagating TSMC action potentials (V₁). D, in the presence of 1 μm nifedipine, residual action potentials triggered during the repolarizing phase of a large hyperpolarizing electrotonic potentials evoked at V₂ readily propagated to V₁.

Table 2.

Effects of 1 μm nifedipine on the parameters of driven action potentials and contractions in mouse renal pelvis

Parameters	Control	1 μm Nifedipine
Driven action potentials (N = 5)
Amplitude (mV)	35.2 ± 4.3	26.8 ± 9.8*
Half-amplitude duration (ms)	360 ± 70	201 ± 39*
Frequency (min−1)	12.0 ± 3.9	14.6 ± 6.3
Contractions (N = 4)
Amplitude (mN)	0.15 ± 0.01	0.05 ± 0.02*
Half-amplitude duration (ms)	464 ± 24	131 ± 40*
Frequency (min−1)	6.9 ± 1.7	6.05 ± 0.67

^*P < 0.05.

Thus nifedipine-resistant (1 μm for > 5–20 min) action potentials still evoked smaller, shorter muscle contractions suggesting that blockade of L-type Ca²⁺ channel in the smooth muscle cells was not complete. In other experiments in which tension was not measured, regenerative action potentials recorded in 1 μm nifedipine were always associated with muscle contraction as viewed via a video camera attached to the microscope. Raising the nifedipine concentration to 3 or 10 μm (for > 2–60 min N = 6) blocked the discharge of all driven action potentials (Fig. 4Aiii–iv) and their associated contractions.

Cells firing TSMC action potentials that were less sensitive to nifedipine were also characterized by the presence of numerous small amplitude (< 10 mV) STDs which presumably propagated from neighbouring ASMCs and which either grew in amplitude or were only apparent in 1 μm nifedipine (Fig. 4Ai and ii). This increase in STD amplitude appeared to arise from a decrease in a membrane conductance of TSMC that was contributing to the after-hyperpolarization after each action potential (Fig. 4Bii and iv), although this decrease did not lead to a significant increase in the frequency of action potential discharge (see above). Raising the concentration of nifedpine to 3–10 μm reduced the amplitude but never blocked the discharge of these small STDs recorded in TSMCs (Fig. 4Aiii and iv).

In 13 preparations, nifedipine (1 μm for > 2 min) completely blocked all action potential discharge and their associated contractions. These recordings were invariably characterized by either the absence of any or the presence of only small amplitude STDs (< 2 mV) within the period between the regenerative action potentials. Action potential blockade by 1 μm nifedipine was usually associated with membrane depolarization of 8.6 ± 1.8 mV (range 1.6–18 mV, N = 8) (Fig. 4A).

This blockade of action potential discharge in nifedipine was not due to a loss of cell-to-cell conductivity. When two electrodes (V₁ and V₂) placed in either the axial or transverse orientation were separated by distances < 50 μm, the hyperpolarizing current injected into one electrode (V₂, Fig. 4Di) evoked an electrotonic potential at the second electrode (V₁, Fig. 4Di) that decreased with distance from the current passing electrode. The decay phase of electrotonic potentials evoked by the current passing electrode often evoked a ‘rebound’ regenerative response in the presence of 1 μm nifedipine, presumably due to voltage-dependent unblock of L-type Ca²⁺ channels, that readily propagated to the second recording electrode (Fig. 4Dii).

Large amplitude (> 10 mV) STDs recorded in 1 μm nifedipine, presumably on the direct impalement of an ASMC or an electrically close TMSC, were little affected when the concentration of nifedipine was raised (3–10 μm nifedipine, N = 3). In the cell illustrated in Fig. 4B, non-regenerative STDs in 1 μm nifedipine had an amplitude, half-amplitude duration, maximal rate of rise and interevent interval of 31.3 ± 3.9 mV, 228 ± 50 ms, 1.87 ± 0.3 mV ms⁻¹ and 1524 ± 1077 ms, respectively (frequency 39 min⁻¹) (n = 65), compared with 30.9 ± 4.0 mV, 385 ± 197 ms, 1.97 ± 0.24 mV ms⁻¹ and 1365 ± 807 ms, respectively (frequency 44 min⁻¹), in 10 μm nifedipine (n = 68) (all P > 0.05 unpaired t test). However, the dependence of STD generation on Ca²⁺ entry was established when their discharge was blocked upon replacing the Ca²⁺ concentration (1.5 mm) in the PSS with an equimolar concentration of Mg²⁺ (Fig. 4Biv, N = 3).

The third pattern of electrical activity recorded with intracellular microelectrodes was also little affected by 1 μm nifedipine. Figure 4C is representative of three experiments in which non-propagating long plateau action potentials recorded with V₂ were still present after 10 min exposure to 1 μm nifedipine, a concentration that was clearly sufficient to block the TSMC action potentials recorded propagating between V₁ and V₂.

Ca²⁺ waves in TSMC layer

In both the gastrointestinal (Torihashi et al. 2002; Yamazawa & Iino, 2002; van Helden & Imtiaz, 2003; Hennig et al. 2004; Park et al. 2006) and urogenital tracts (Hashitani et al. 2004a), Ca²⁺ fluorescence imaging is increasingly being used to examine the mechanisms of pacemaker generation and coupling/propagation to neighbouring pacemaker cells and smooth muscle bundles. When viewed under ×20 magnification (Fig. 5Ai top panel), renal pelvis preparations generated transients in [Ca²⁺]_i (Ca²⁺ waves) which swept across the field of view and which were always associated with propagating contractions. The gross morphology of the cells generating the Ca²⁺ waves when viewed at higher magnification (Fig. 5Aii bottom panel) suggested that they were TSMCs forming a loosely packed ‘basket weave’ network within a single layer of the preparation. In seven experiments these Ca²⁺ waves had a mean amplitude, half-amplitude duration and interevent interval of 0.17 ± 0.12 F_t/F₀, 490 ± 194 ms and 3174 ± 1838 ms, respectively (frequency 23.6 ± 10 min⁻¹) (N = 4, n = 9–20). Cross-correlation analysis of these Ca²⁺ waves at any two regions within the field of view revealed a prominent peak cross-correlation factor (CCF) near lag time 0 ms (Fig. 5Bii) suggesting that a 1 : 1 correlation existed between the activity of individual TSMCs within the muscle layer.

Figure 5.

Ca²⁺ waves in the typical smooth muscle (TSMC) layer of the renal pelvis A, sequential Ca²⁺ fluorescence intensity micrographs of the fluo-4 loaded TSMC layer with time intervals of 66 ms observed at ×20 (Ai) or ×60 (Aii) magnification. The Ca²⁺ wave was clearly seen as a transient increase in Ca²⁺ intensity propagating across the field of view; the arrow indicates a single TSMC. Bi, superimposed fluorescence intensities of the 3 regions (a–c) in Ai top panel plotted against time. Bii, cross-correlogram of the Ca²⁺ wave recorded at a and c (separation 110 μm) show a high degree of 1 : 1 synchronicity. C, Ca²⁺ waves recorded at 2 positions in a field of view (Ci) were reduced but not blocked upon exposure to 1 μm nifedipine (Cii). In other fields of view in the same preparation, Ca²⁺ waves in the TSMC layer were completely blocked by 1 μm nifedipine.

By sampling the Ca²⁺ fluorescence of two to three regions, separated by 50–200 μm, within a field of view (Fig. 5Ai) it was possible to accurately measure the conduction velocity of these Ca²⁺ waves as 0.97 ± 0.83 mm s⁻¹ (range 0.13–2.7 mm s⁻¹; N = 4, n = 8). These conduction velocities are similar to the conduction velocities recorded for driven action potentials using paired intracellular microelectrodes. In Fig. 3B individual determinations of the intervals of the Ca²⁺ waves passing two regions of fluorescence measurement within a field of view have been plotted on the interval versus microelectrode separation plots obtained previously in both the axial and transverse directions.

Ca²⁺ waves within the same preparation displayed a similar variable sensitivity to 1 μm nifedipine as driven action potentials. In many fields of view, propagating Ca²⁺ waves were completely blocked in 1 μm nifedipine (for > 10 min). However, if another field within the sample preparation was viewed small Ca²⁺ transients firing at a reduced frequency were often recorded. Figure 5C is a typical example of one of these sites in which the Ca²⁺ transients within two regions (Fig. 5Cia and b) within the field of view were not completely blocked in 1 μm nifedipine (for > 10–60 min) (Fig. 5Ciia and b).

Ca²⁺ transients in ASMCs

In the absence of nifedipine it was clear that other cells in preparations of mouse renal pelvis also displayed Ca²⁺ transients which were not directly related in time to the Ca²⁺ waves observed in the TSMC layer. These cells were best observed under high (×60) magnification (Fig. 6A). Figure 6B illustrates a typical experiment in which four short (10–50 μm in length) spindle shaped cells displayed high frequency (> 10 min⁻¹) Ca²⁺ transients in control PSS which were independent of the Ca²⁺ waves in the TSMC layer. It was not possible to perform a meaningful cross-correlation analysis of this experiment due to our inability to separate the Ca²⁺ transients in the ASMCs from the fluorescence increases associated with the Ca²⁺ waves that swept over these ASMCs in the overlying TSMC layer. Spindle shaped cells discharging Ca²⁺ transients formed randomly orientated bundles, which extended across the field of view (135 × 135 μm²) and not in the same plane of focus as the TMSC layer (Fig. 6A).

Figure 6.

Ca²⁺ transients in atypical smooth muscle cells (ASMCs) of the renal pelvis A, greyscale micrograph of fluo-4 loaded ASMCs; arrows indicate single ASMCs. Bi–iv sequential fluorescence intensity micrographs of 4 ASMCs (a–d) and a Ca²⁺ wave in the TMSC layer (e) taken at intervals of 1 s (arrow indicates direction of Ca²⁺ wave). C, Ca²⁺ fluorescence intensities for the 4 ASMCs (a–d) illustrated in B plotted against time reveals that the Ca²⁺ transients in these cells had no temporal relationship with the Ca²⁺ wave. D, Ca²⁺ transients recorded in 2 ASMCs (Di upper panels) were little affected by 1 μm nifedipine, which reduced Ca²⁺ waves in the TSMC layer (Dii lower panel).

The addition of 1 μm nifedipine markedly reduced (Figs 5C and 6D) or completely blocked (data not shown) the Ca²⁺ rises associated with the Ca²⁺ waves leaving an uncontaminated Ca²⁺ transient in these short spindle shaped cells. Thus most experiments investigating the properties of these Ca²⁺ transients were performed in 1 μm nifedipine unless otherwise stated. Table 3 summarizes the properties of the Ca²⁺ transients recorded in spindle shaped cells within seven preparations bathed in 1 μm nifedipine (N = 7, n = 12–90).

Table 3.

Comparison of parameters of Ca²⁺ transients in ASMCs and ICC-LCs in preparations of renal pelvis bathed in 1 μm nifedipine

Cell (N = 7)	Amplitude (Ft/F0)	Half-amplitude duration (ms)	Frequency (min−1)
ASMCs	0.22 ± 0.12	512 ± 118	11.5 ± 3.9
ICC-LCs	0.33 ± 0.17	1271 ± 440*	3.5 ± 2.4*

^*P < 0.05.

In eight preparations, Ca²⁺ transients in individual spindle shaped cells bathed in 1 μm nifedipine did not propagate more than one to two cell lengths. In Fig. 7A and B, cross-correlation analyses of the four cells displaying Ca²⁺ transients simultaneously within the field of view (Fig. 7Aa–d and Ba–d) revealed that a prominent peak CCF near lag time 0 ms was only observed in cells that were separated by < 30 μm (Fig. 7B lower panel). Cells that were separated by distances > 30 μm did not display a prominent peak CCF near lag time 0 ms and were therefore displaying little synchronicity. Figure 7C is representative of four experiments demonstrating that STDs (5–20 mV in amplitude) recorded at one electrode (V₁) also invariably failed to propagate to a second electrode (V₂) placed 50–60 μm in either the axial or transverse direction.

Figure 7.

STDs and Ca²⁺ transients in ASMCs do not propagate Ai–viii, sequential Ca²⁺ fluorescence micrographs of 4 ASMCs (a–d) taken at intervals of 300 ms. Bi, Ca²⁺ fluorescence intensity of the 4 cells (a–d) illustrated in A plotted against time. Bii, cross-correlagrams of the four cells (a–d) plotted in Bi reveals little 1 : 1 synchronicity except when cells (a–b and b–c) were separated by < 30 μm. C, in 1 μm nifedipine, STDs recorded with one microelectrode (V₁) were not recorded by a second microelectrode (V₂) situated 50–60 μm along the longitudinal (Ci) or transverse (Cii) axis.

We have plotted this relationship of peak CCF and separation distance (10–105 μm) in 32 pairs of cells from eight preparations (bathed in 1 μm nifedipine containing PSS) in Fig. 9Ei. The pooled data were fitted by least squares to the straight line y = 0.284 + 0.000616x. These data suggest that only neighbouring spindle shaped cells were synchronizing their discharge of Ca²⁺ transients.

The similarity in frequency of discharge, half-amplitude duration and failure to propagate over any appreciable distance as well as their general morphology and position within the renal pelvis wall (Dixon & Gosling, 1973, 1982; Gosling & Dixon, 1974; Klemm et al. 1999; Lang et al. 2001) lead us to suggest that the spindle shaped cells generating these short duration Ca²⁺ transients are in fact ASMCs.

Ca²⁺ transients in ICC-LCs

High magnification Ca²⁺ imaging of renal pelvic preparations also revealed another population of cells displaying Ca²⁺ transients which occurred at lower frequencies (< 4 min⁻¹) than either TSMC Ca²⁺ waves or ASMC Ca²⁺ transients (Fig. 8Bi). The morphology of these cells was variable, being triangular, fusiform or stellate shaped with oval shaped nuclear regions, 10–20 μm long (Fig. 8A). The density of these cells displaying low frequency Ca²⁺ transients was one to five per field of view compared with the density of ASMCs firing high frequency transients (> 20 per field of view). However, like ASMCs the Ca²⁺ transients recorded in these cells were little affected by 1 μm nifedipine (Fig. 8Bi and ii) so that all subsequent studies were made in nifedipine.

Figure 8.

Ca²⁺ transients in ICC-LCs A, sequential Ca²⁺ fluorescence micrographs of fusiform shaped ICC-LC (arrow indicates cell in inset) taken at intervals of 1 s. B, Ca²⁺ fluorescence intensity of TSMC Ca²⁺ wave and ICC-LCs plotted against time in the absence (Bi) and presence (Bii) of 1 μm nifedipine. Ci, Ca²⁺ fluorescence intensity of 4 ICC-LCs (a–d) in a field of view plotted against time. Cii, cross-correlagrams of the four cells (a–d) plotted in Ci reveals that ICC-LCs were not firing synchronicity except when separated by < 30 μm (a–d and b–c).

Table 3 illustrates that the amplitude and half-amplitude duration of the Ca²⁺ transients recorded in these fusiform cells were significantly larger than the equivalent parameters for ASMC Ca²⁺ transients recorded in the same seven preparations (P < 0.05 unpaired t tests)(Fig. 9C). In contrast, the discharge frequency of the Ca²⁺ transients recorded in these fusiform cells was significantly smaller than in ASMCs.

As the discharge frequency and half-amplitude duration of the long plateau action potentials (Fig. 9B) were similar to the equivalent parameters measured from Ca²⁺ transients recorded in the fusiform shaped cells (Fig. 9C), it seems likely that these low-frequency electrical discharges and Ca²⁺ signals are arising from the same population of cells. The sparse distribution of the cells displaying these low frequency Ca²⁺ transients also matches the sparse distribution of Kit-positive ICC-LCs in the proximal renal pelvis (Fig. 1Ciii), ureteropelvic junction and proximal ureter (Pezzone et al. 2003; Metzger et al. 2005; Lang et al. 2006b; Lang et al. 2007) suggesting that these low frequency signals are arising from ICC-LCs.

Cross-correlation analyses of the Ca²⁺ transients of the four ICC-LCs displayed in Fig. 8Ci revealed little 1 : 1 synchronicity in their activity, except when they are separated by < 30 μm (Fig. 8Cii cells c–d). An analysis of 32 pairs of ICC-LCs in seven preparations (Fig. 9Eii) revealed that the average peak CCF (0.38 ± 0.26, range 0.04–0.94) was significantly larger than the same estimate for 32 pairs of ASMCs (0.25 ± 0.1, range 0.08–0.44, N = 8) (Fig. 9Ei; P < 0.05 unpaired t test). There was also little reduction of peak CCF with distance between ICC-LCs (Fig. 9Eii) suggesting that electrotonic spread plays no role in the degree of synchronicity shown in 1 μm nifedipine.

As it was not technically possible to make recordings of electrical activity simultaneously with changes in Ca²⁺ levels in our preparations we compared the interevent intervals and half-amplitude durations of the three types of Ca²⁺ signals recorded with the equivalent variables for the three types of electrical activity recorded with intracellular microelectrodes. In Fig. 9A it is clear that muscle contractions, driven action potentials and Ca²⁺ waves all occur at a frequency of about 10 min⁻¹ and have a half-amplitude duration of about 500 ms supporting our notion and direct demonstration that driven action potentials propagate and trigger a rise in internal Ca²⁺ and muscle contraction.

Figure 9B compares the properties of driven action potentials and long plateau action potentials recorded simultaneously in four preparations, while Fig. 9C compares the intervals and the half-amplitude durations of the Ca²⁺ transients arising in high frequency ASMCs and the low frequency ICC-LCs recorded simultaneously in seven preparations. It can be seen that the mean interevent intervals of the long plateau action potentials are similar to the low frequency Ca²⁺ transients recorded in the fusiform ICC-LCs. In Fig. 9D Ca²⁺ transients in ASMCs and ICC-LCs (open circles) show little 1 : 1 correlation in their activity confirming they were not arising from the same population of cells, nor were they synchronized. In contrast, there was a 1 : 1 correlation in the temporal relationship between driven action potentials and contraction (Fig. 9D filled circles) in four preparations of mid renal pelvis.

Effects of gap junction blockers

Gap junctions are likely to provide electrical coupling between cells within the TSMC layer (Santicioli & Maggi, 2000) as well as low resistance pathways between ASMCs, ICC-LCs and TSMCs (Klemm et al. 1999). We have used the uncoupler of gap junctions, 18β-glycyrrhetinic acid (18β-GA, 50 μm), to study the propagation of electrical and Ca²⁺ signals in the TSMC layer and ASMC bundles. In four experiments, propagating driven action potentials recorded with a pair of intracellular microelectrodes (V₁ and V₂ < 50 μm apart) were rapidly blocked (within 5 min) in the presence of 50 μm 18β-GA (Fig. 10A). In one experiment, driven action potential and STD discharge was still evident at V₁ and absent at V₂. In all cases, 18β-GA blockade of propagation was associated with a reduction in the amplitude of the electrotonic potentials recorded at V₁ in response to a large hyperpolarizing current applied to V₂ (Fig. 11Bi and ii). These effects were reversible after 10 min wash (data not shown). Similar results were obtained when another putative gap junction blocker, carbenoxolone (CBX 50 μm, N = 3), was used.

Figure 10.

Effects of a gap junction uncoupler, 18β-glycyrrhetinic acid (18β-GA) A, propagating action potentials recorded with a pair of intracellular microelectrodes (V₁ and V₂ < 50 μm apart) were blocked by 50 μm 18β-GA in a manner associated with a reduced amplitude of the electrotonic potentials recorded at V₁ in response to a large hyperpolarizing current applied to V₂ (Bi and ii). C, 18β-GA (50 μm) rapidly blocked propagating Ca²⁺ waves in the TSMC layer (Ci and ii) as well as the Ca²⁺ transients recorded in ASMCs (Ciii and iv).

Figure 11.

Schematic representation of the mouse pelviureteric system and relationships between ASMCs, TSMCs and ICC-LCs A, diagram illustrating the gross morphology of the pelviureteric system and the regions of the pelvi-calyceal junction (PCJ), proximal renal pelvis (Prox. RP), and the ureteropelvic junction consisting of the distal renal pelvis (Dist. RP) and ureter. Shaded areas represent typical positions of the longitudinally (L) and transversely (T) cut strips of renal pelvis dissected for experimentation. B, schematic cross sectional diagram of the proximal renal pelvis illustrating the distribution and interconnectivity of TSMCs, ASMCs and ICC-LCs, blood vessels (BV), nerve bundles (N), lamina propria (LP) and urothelium. C, relative distribution of ASMCs, TSMCs and ICC-LCs along the mammalian pelviureteric system.

We also examined the effects of 18β-GA (50 μm) on the Ca²⁺ waves recorded in the TSMC layer (N = 3) and the Ca²⁺ transients recorded in a number of ASMCs within a field of view (N = 3, n = 8). Figure 10Ci and ii illustrates that Ca²⁺ waves recorded at two points within a field of view were rapidly blocked by 50 μm 18β-GA. In contrast, Ca²⁺ transients in ASMCs were slowly reduced by 18β-GA (50 μm) (Fig. 10Ciii and iv). Ca²⁺ waves and propagating driven action potentials also rapidly (within 1 min) recovered upon washout (Fig. 10Ci and ii), while ASMC Ca²⁺ transients were slower to recover over 10–30 min (data not shown).

Discussion

The data presented here represent the first visualization of three distinct populations of cells responsible for the spontaneous electrical activity which drive peristalsis in the pelviureteric system. Rhythmic Ca²⁺ waves were recorded in a distinct layer of long (> 100 μm) TSMCs, which when impaled by an intracellular microelectrode displayed action potentials that were directly demonstrated to trigger muscle contraction (Figs 1 and 5). Attempts to block smooth muscle L-type voltage-dependent Ca²⁺ channels with low concentrations (1 μm) of nifedipine were only partially successful as tissues displayed a considerable regional variability in their sensitivity to nifedipine. In any individual preparation, Ca²⁺ waves, contractions and TSMC action potentials recorded in one region could be completely blocked while neighbouring regions were only partially reduced by 1 μm nifedipine. However, these reductions were not associated with any significant changes in discharge frequency. Contractions, Ca²⁺ waves and driven action potentials were invariably blocked when nifedipine concentration was raised (3–10 μm). In the guinea pig, both contractions propagating the length of the ureter (Meini et al. 1995; Santicioli & Maggi, 1998) and the whole cell Ca²⁺ channel current recorded in freshly isolated single myocytes (Lang, 1990) have a similar resistance to nifedipine blockade.

This heterogeneous sensitivity of TSMC L-type Ca²⁺ channels to nifedipine has complicated our interpretations of the electrical activity recorded in the absence or presence of nifedipine. The randomly orientated ‘basket weave’ structure of both TSMC and ASMC layers has also precluded the dissection of thin, ‘short cable’ preparations that can be obtained in other gastrointestinal (Cousins et al. 2003) and urogenital (Hashitani et al. 2001) smooth muscles. Thus we have not been able to assume that the electrical events recorded in 1 μm nifedipine directly reflect the behaviour of any neighbouring pacemaker cells present. We believe that cells displaying large STDs (Fig. 4B) are generated in ASMCs electrically close to the recording electrode, while the smaller variable amplitude STDs are generated in ASMCs that are electrically distant from the recording electrode to a varying degree (Fig. 4A).

ASMCs

Two other patterns of electrical and Ca²⁺ transient activities were observed in both longitudinally and transversely cut strips of the mouse renal pelvis (Fig. 11A). High frequency STDs and Ca²⁺ transients were recorded, although not simultaneously, in short (< 50 μm) spindle shaped ASMCs whether viewed with Lucifer Yellow or fluo-4. ASMCs formed small bundles that were also randomly orientated and not in the same plane of focus as the TSMC layer confirming the presence of distinct ASMC and TSMC muscle layers reported in previous light and electron microscopic studies of the pelviureteric system (Fig. 11B) (Gosling & Dixon, 1974, 1978; Dixon & Gosling, 1982; Klemm et al. 1999). The discharge frequency and time course of small (< 10 mV) or large STDs were similar (Fig. 9B and C). However large STDs were recorded only occasionally in the distal renal pelvis and more often when impalements were made in the most proximal regions of the renal pelvis strips. This suggests that impalement of ASMC was more likely in the proximal renal pelvis in agreement with previous electron microscopic demonstrations that ASMCs were preferentially distributed near the base of the papilla (Fig. 11C) (Gosling & Dixon, 1974; Dixon & Gosling, 1982; Klemm et al. 1999).

Paired microelectrode recordings revealed that, in contrast to driven action potentials, STDs did not propagate over distances > 50 μm (Fig. 7). Ca²⁺ transients in ASMC also did not travel more than one to two cell lengths (Fig. 4). The cross-correlation analyses of 32 pairs of ASMCs separated by < 100 μm revealed that there was little synchronicity in their spontaneous activity except when separations were < 30 μm (Fig. 9Ei). However there must be some form of cross talk between ASMCs in close apposition (Klemm et al. 1999) because the gap junction blocker, 18β-GA, blocked both Ca²⁺ waves in the TSMC layer and the Ca²⁺ transients in ASMCs. If ASMCs were completely asynchronous and communicating only with TSMCs, it would be expected that the Ca²⁺ transients in ASMCs would be unaffected by 18β-GA.

If ASMCs were triggering Ca²⁺ waves in the TSMC layer then one or a few ASMCs must be acting as a point source of excitation and on most occasions this site of excitation must have been situated outside the field of view. On only 3 occasions were we able to observe the origin of the Ca²⁺ waves such that spontaneous rises of Ca²⁺ fluorescence were observed to radiate from a single region within the preparation. It was only possible on one occasion to observe (at ×60 magnification) the ASMC that appeared to be driving a neighbouring ASMC which then triggered Ca²⁺ waves in the adjacent TSMC layer that propagated across the preparation. As ASMCs do not form any extensive network nor securely drive TSMCs like ICCs in the GI tract, we propose that they may be constantly providing an excitable drive to neighbouring TSMCs so that the frequency of driven action potentials and their associated Ca²⁺ waves may be primarily set by the refractory period of TSMCs (Lang et al. 1998; Klemm et al. 1999).

ICC-LCs

A number of Kit-positive cell types have recently been identified in the upper urinary tract using fluorescence and light microscopy. Kit-positive cells, lying within the lamina propria, the septa between the inner and outer muscle layers and within muscle bundles, have been described as being spindle shaped, fusiform with two distinct dendrites and stellate in appearance. Many Kit-positive cells also stain for FITC–avidin (Pezzone et al. 2003; David et al. 2005; Lang et al. 2006b), CD34 (Metzger et al. 2004), AA1 or CC1 (Metzger et al. 2005) indicating that these cells are other haemopoietic cells such as mast cells or macrophages. In our experience these dual labelled cells represent < 50% of the Kit-positive cells (Fig. 1C) (Lang et al. 2006b; Lang et al. 2007). Most researchers agree that Kit-positive cells form a diffuse network and that their number decreases with distance from the papilla base (Fig. 11C) (Lang & Klemm, 2005). However, many of the preparations studied to date have yet to be examined under the electron microscope to confirm that these Kit-positive cells display the necessary morphological criteria that distinguishes ICC in the intestine from myofibroblasts (Klemm et al. 1999). In the present experiments low frequency Ca²⁺ transients were recorded in fusiform cells that were sparsely distributed in each field of view, while long plateau action potentials were recorded only infrequently. However, the cells firing long lasting action potentials are presumably well coupled to the TSMC layer as they were recorded in the absence of driven action potential discharge in only 2 of 15 recordings (Figs 1C and 4Ci). This is perhaps in good agreement with previous electron microscopic observations in the guinea pig renal pelvis of the high degree of close appositions between neighbouring ICC-LCs (82% of cells), between ICC-LCs and TSMCs (4%), and the 18% of ICC-LCs which did not form any close appositions within the plane of section (Klemm et al. 1999).

Recently, single stellate and spindle shaped ICC-LCs have been enzymatically isolated from the mouse ureteropelvic junction and their electrical properties examined using patch clamp techniques (Lang et al. 2006b; Lang et al. 2007). Isolated ICC-LCs were distinguished by their lack of expression of either 4-aminopyridine-sensitive, voltage-activated K⁺ channels or large-conductance Ca²⁺-activated K⁺ channels which were present in freshly isolated TSMCs. ICC-LCs also uniquely display high frequency spontaneous transient inward currents which sum to form large inward currents (LICs) which appeared at a frequency of 1–3 min⁻¹ in cells bathed in 1 μm nifedipine PSS (Lang et al. 2006b; Lang et al. 2007). LICs were calculated to involve a 10–1000-fold increase in the cell membrane conductance for cations (Lang et al. 2006b; Lang et al. 2007). Although this has yet to be directly demonstrated, it seems likely that the low frequency long plateau potentials recorded occasionally with intracellular microelectrodes in the absence or presence of 1 μm nifedipine are being generated by Kit-positve ICC-LCs firing LICs.

Ca²⁺ transients in both ASMCs and ICC-LCs and all spontaneous electrical activity were blocked upon removal of Ca²⁺ from the bathing solution or upon blockade of Ca²⁺-ATPase with cyclopiazonic acid (CPA, 10 μm for 10–60 min) (H. Hashitani & R. J. Lang, unpublished data). We also have evidence that Ca²⁺ transients in ASMCs and ICC-LCs display differing sensitivities to blockers of Ca²⁺ release from ryanodine-dependent Ca²⁺ stores. We speculate that STDs and plateau potentials are being triggered by the spontaneous rises in [Ca²⁺]_i observed in fluo-4 loaded ASMCs and ICC-LCs, respectively. These Ca²⁺ transients also arise from the release of Ca²⁺ from internal stores in a manner that is sustained by influx of Ca²⁺ from the extracellular space. However the exact mechanisms of autorhythmicity and any differences in their pharmacological profiles in both AMSCs and ICC-LCs have yet to be fully elucidated.

In conclusion, the ability of the ureter to maintain rudimentary peristaltic waves and movement of urine from the papilla to the bladder after pyeloplasty, uretero-ureteral anastomosis or ureteral obstruction in the absence of a proximal pacemaker drive was previously thought to arise ‘autonomously’, by which stretch of the ureter upon urine collection at a site distal to the site of injury causes action potential discharge and muscle contraction (Santicioli & Maggi, 1998). However the presence of ICC-LCs in regions distant from the papilla may well be responsible for this intrinsic rhythmicity, particularly in human ureter (Schwentner et al. 2005). As such, pyeloureteric ICC-LCs with their unique pacemaker mechanisms may well provide a selective pharmacological target when considering non-surgical interventions to alleviate hydronephrosis arising from UPJ remodelling during and after ureteric blockade or pyeloplasty.