|
|
||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Physiology3
Department of Veterinary Preclinical Sciences,4 The University of Liverpool, Liverpool L69 3BX, United Kingdom
| ABSTRACT |
|---|
|
|
|---|
calcium, interstitial cells, parturition, pregnancy, uterus
| INTRODUCTION |
|---|
|
|
|---|
An earlier study [12] had briefly mentioned the presence of cells with complex geometry and terminal arms in freshly dissociated preparations of uterine myocytes, but no study of them was made. The light microscopic appearance of these cells closely resembled that of ILCs and ICCs in the smooth muscles mentioned above, suggesting that ILCs may be present in myometrium.
The aims of this work were therefore to investigate if ILCs were present in uterine tissue of two species, rat and human, to characterize them, and to determine if they played a role in pacemaker activity.
| MATERIALS AND METHODS |
|---|
|
|
|---|
Cell Isolation
Rats were humanely killed by cervical dislocation under CO2 anesthesia. The uteri were removed and small strips of longitudinal smooth muscle were dissected. Human uterine biopsies were collected in cooled physiological salt solution (HBSS, Sigma) and small strips were dissected along the bundles of smooth muscle with the aid of a dissection microscope. After dissection, the strips were incubated in low-Ca (50 µM) HBSS for 60 min at 36°C. Single cells were produced by enzymatic digestion using 0.1% collagenase Type 1A and 0.03% elastase Type IIA (both from Sigma) dissolved in low-calcium HBSS. Digestion time was 1 h for rat myometrium and 2 h for human myometrium. After digestion, fluffy strips were washed 23 times with low-Ca HBSS and transferred into KB medium (see below [13]). After 1015 min of incubation in KB medium, strips were triturated using a fire-polished glass pipette to release single cells. The suspension obtained consisted of smooth-muscle cells (
90%) and ILCs (
10%). Cells were kept in KB medium at 4°C and used within 4 h after isolation.
Strips
Thin (0.30.5 mm) strips of human myometrium were treated with collagenase (Type IA, 0.1%) for 2030 min at 36°C in low-Ca HBSS. After the enzyme treatment, cells in the strips were more easily visualized. The strips were washed with the low-Ca HBSS and placed into a perfusion chamber mounted on the stage of the Olympus IX-51 inverted microscope. Wide-field images of partially digested myometrial tissue were taken via a water immersion 40 x 0.75 numerical aperture objective lens, using an UltraPix cooled CCD camera controlled by UltraView Software (Perkin Elmer, Cambridge, UK).
Immunohistochemistry
Dispersed cells (adhering to glass slides coated with polylysine) and tissue samples were fixed in 4% (w/v) paraformaldehyde in 0.1 M phosphate-buffered saline. For light microscopy, frozen section and cells were incubated for 16 h in mouse monoclonal antibodies against vimentin (V9 prediluted; DakoCytomation, Cambridge, UK) or c-kit (NCL-cKIT, 1:10; Vector Laboratories, Peterborough, UK), followed by a 1-h incubation in goat anti-mouse IgG conjugated to FITC (1:500, Sigma, UK) or donkey anti-mouse IgG conjugated to Cy3 (1:5000, Jackson ImmunoResearch, Stratech Scientific Ltd, Luton, UK). FITC staining was usually combined with staining of smooth-muscle cells, using phalloidin labeled with TRITC (Sigma, UK). Tissue sections were also double stained for vimentin and a pan-neuronal marker PGP9.5 (rabbit anti-PGP9.5, 1:5000; Ultraclone, Isle of Wight, UK), using the FITC-labeled anti-mouse IgG and donkey anti-rabbit IgG conjugated to Cy3 (Jackson ImmunoResearch). Coverslips were mounted with Vectashield containing 4',6'-diamidino-2-phenylindole (DAPI; Vector Laboratories).
Electron Microscopy
Small tissue samples and dispersed cells were fixed in 4% paraformaldehyde and 2% glutaraldehyde, followed by 1% osmium tetraoxide, and processed routinely into resin. For immunohistochemical labeling, small tissue samples were fixed in paraformaldehyde and stained for vimentin, using monoclonal V9 and the immunoperoxidase method, before fixing in glutaraldehyde and osmium tetraoxide and embedding in resin.
Electrophysiology
Electrophysiological properties of isolated cells were studied using conventional whole-cell patch clamp technique in both current-clamp and voltage-clamp modes. An aliquot of the cell suspension was placed in a perfusion chamber mounted on the stage of an inverted microscope (Olympus IX 50, Japan), allowed to settle for 5 min, and perfused with prewarmed (35°C) Krebs solution (see below). An EPC-9 patch-clamp amplifier controlled by an IBM-compatible PC running the Pulse 8.53 data acquisition software (HEKA, Germany), was used in the experiments. Patch pipettes of 35 m
were fabricated from 1.5-mm glass capillaries.
Pipettes were filled with solution containing (in mM) NaCl, 5; KCl, 40; K-Glutamate, 90; MgCl2, 1; HEPES, 10; EGTA, 0.01; Mg-ATP, 5; (pH 7.2 adjusted with NaOH). Krebs solution contained (in mM) NaCl, 140; KCl, 5.4; CaCl2, 2; MgCl2, 1.2; Glucose, 10; HEPES, 10 (pH 7.4 adjusted with NaOH). KB medium used for cell storage was composed of (in mM) KCl, 40; KH2PO4, 10; KOH 105, taurine, 10; glucose, 11; EGTA, 0.1; N-tris[hydroxymethyl1]methyl-2-aminoethanesufonic acid, 10. Glutamic acid was used to adjust the pH of this solution to 7.2.
| RESULTS |
|---|
|
|
|---|
Enzymatic dissociation of the pregnant rat and human uterus consistently revealed cells of strikingly different appearance from smooth-muscle cells, obtained in the same isolation (Figs. 1, 3, 4). These cells were variable in their morphology but will be referred to as ILCs in the subsequent discussion. The most obvious characteristics of these non-smooth-muscle cells were that they had a bi- or multipolar stellate appearance (Fig. 1). Some cells had many thin spines projecting from them, as seen in Figure 1C. There was considerable diversity in their appearance, consistent with the in situ data described below. Some of the cells looked similar to smooth-muscle cells but with distended centers and forked ends. The relaxed smooth-muscle cells (Fig. 1D) had a mean length of 212 ± 15 µm and mean diameter of 24 ± 3 µm and were the majority of cells (
90%) in every isolation. The non-smooth-muscle cells could also be distinguished by their noncontractile behavior. Thus, only 2/11 cells identified by their morphology as being ILC contracted when depolarized or damaged by the patch pipette at the end of the experiment. In contrast, all smooth-muscle cells tested in the same experiments contracted.
|
|
|
In Situ
In strips of rat myometrium exposed briefly to collagenase to increase cell visualization (2030 min), the same cells were seen as in the dispersed cell preparations. Figure 2 shows pieces of rat uterine tissue prepared in this way. In Figure 2A, the clubbed bipolar cell type is evident, and Figure 2B shows the thinner type with several projections. The close apposition of these cells with smooth-muscle cells and their relative sizes can also be discerned (Fig. 2C).
|
Immunohistochemistry
Tissue sections were examined for c-kit immunoreactivity with a c-kit antiserum that we had used previously to label ICCs in human gut fixed with paraformaldehyde [14], as were the uterine samples in the present study. Neither ILCs nor myocytes within rat and human uterus were labeled: immunoreactivity was detected only in small, round cells lacking projections that resembled mast cells (not shown). Thus, c-kit, an accepted marker for enteric ICCs and mast cells, did not label the ILCs in uterus. In contrast, vimentin, a marker for ICCs and fibroblast-like cells, showed intense immunoreactivity in both rat and human uterine samples. Long, slender cells with fine processes were distributed throughout the muscle layers (Fig. 3, a and b), as well as in the connective tissue of the endometrium. Many of the vimentin-positive cells were closely associated with axons (Fig. 3d).
Vimentin-staining of dispersed cells revealed a population of cells with multiple fine processes, morphologically distinct from smooth-muscle cells (Fig. 3c). Figure 4 shows three-dimensional reconstructions of the ILCs from human myometrium stained with antibodies against vimentin.
Electron Microscopy
Cells resembling ICCs were identified in both the dispersed cell preparations (Fig. 5) and intact tissue (Fig. 6). These ILCs were characterized by large numbers of caveolae, many mitochondria, and a dense mass of intermediate filaments in their cytoplasm. In dispersed cells, the smooth endoplasmic reticulum was dilated, but this was less evident in ILCs in intact tissue: rough endoplasmic reticulum was either absent or restricted to a few cisternae close to the nucleus. The cells extended numerous fine processes and formed extensive gap junctions with neighboring ILCs (Figs. 5 and 6). They also contacted smooth-muscle cells, but no membrane specialization was apparent at the points of contact (Fig. 5C). Vimentin immunoreactivity was detected in cells showing the ultrastructural features of ICCs (Fig. 6b) as well as in fibroblasts and some endothelial cells. ILCs were morphologically distinct from fibroblasts, the latter contained abundant rough endoplasmic reticulum and had few caveoli associated with their cell membrane.
|
|
Electrophysiology
The electrophysiological properties of ILCs were studied using current-clamp and voltage-clamp modes of the patch-clamp technique and compared with those of smooth-muscle cells. Table 1 compares the electrophysiological characteristics of ILCs and smooth-muscle cells (SMC) obtained in the same isolation. ILCs had large input resistance, ranging between 1.2 and 12 G
. For current-clamp experiments, only cells with input resistance below 5 G
were selected, to avoid possible artefactual changes in transmembrane voltage due to the bias input current of the amplifier headstage. Resting membrane potential was recorded for 210 min in every cell to see if spontaneous slow waves of depolarization occurred. The ILCs failed to produce regular slow waves of depolarization described in classical ICCs, although some irregular excursions of membrane potential ranging from 10 to 35 mV were observed (see Fig. 7A). Furthermore, ILCs did not generate action potentials in response to depolarizing current. Only passive electrotonic potentials were recorded when current pulses were applied (Fig. 7B). In voltage-clamp experiments, rapidly inactivating outward current was recorded in response to depolarizing voltage pulses above 40 mV (Fig. 7C). This current disappeared when K+ was substituted for Cs+ in the pipette solution, indicating that the outward current was carried by K+ flowing through K+ channels. No inward current was recorded in ILCs dialyzed with either K+ or Cs+-containing pipette solution (not illustrated). In contrast, all SMC exhibited inward current and were capable of generating action potentials (Fig. 7D).
|
|
| DISCUSSION |
|---|
|
|
|---|
Differences from Other Cell Types
As shown in our figures, the ILCs could easily be distinguished from smooth-muscle cells, axons, and fibroblasts, the other main cell types present in myometrium. This identification was based on morphology, ultrastructure, and electrophysiological properties, which were consistent from single cells to in situ preparations.
The single-cell studies revealed non-smooth-muscle uterine cells that were similar to those shown in the uterus by Yoshino et al. [12], and in the recent reports from other smooth muscles. Yoshino et al. [12] estimated that these nonsmooth cells comprised around 10% of the population of cells in a digestion, a figure that we also found and that is comparable with the proportion of these cells (6%) in a digest of mesenteric artery [7]. Although not systematically examined, the ILCs were much more populous than any remaining cell types, such as fibroblasts or endothelial cells. A consistent finding in all these studies and our own is the pleomorphism. Thus, these cells had, in common vimentin staining, thin projections and lack of contractility, but each had its own unique shape, as shown in our figures and those of the other studies quoted.
In situ, there were many ILCs between smooth-muscle fiber bundles and spreading spider-like between smooth-muscle cells with their extended cytoplasmic processes. The cells were present throughout the uterus of rat and women. Horiguchi et al. [15] suggested that thick smooth-muscle layers, in their case, gastric antrum, have septal ICCs, infiltrating between muscle bundles and layers, providing a pathway for the transmission of depolarization. The distribution of the ILCs in uterus appears similar to septal ILCs.
Our ILCs were negative for c-kit, a marker for enteric ICCs, but were vimentin positive. In a previous study, no c-kit expression was found in human uterus, although the c-kit receptor ligand was detected in myometrium [16]. More recent reports have also failed to detect any myometrial c-kit expression, and evidence for its expression in uterine tumors is mixed [17, 18]. Thus, our data suggest that uterine ILCs are of mesenchymal origin but are distinct from smooth-muscle cells and are not classical ICCs.
The lack of c-kit staining does not preclude a pacemaker role of these cells. In the urethra, Sergeant et al. [4] described vimentin-positive cells with pacemaker activity, but which were negative to a panel of c-kit antibodies. In contrast, Povstyan et al. [6] described c-kit-positive cells in portal vein but were unable to identify any pacemaker role for the cells. It is also the case that not all ICCs are involved in pacemaking and initiating the propagation of slow waves. Recently, roles for ICCs in transmitting or modifying signals between nerves and smooth muscle have been reported [11, 15, 19, 20]. As Horiguchi et al. [15] expressed it, a picture is emerging regarding the functional significance of ICC, suggesting that a division of labor exists, with ICCs involved in pacemaking and/or neurotransmission. In the uterus, ILCs were found closely associated with axons and smooth-muscle cells and, thus, they may be involved in modulation of neurotransmission.
Electrophysiology
Our electrophysiological data does not support a role for the ILCs as excitatory pacemakers or generators of slow waves within the myometrium. Thus, these ILCs did not exhibit regular spontaneous depolarizations in current clamp. In voltage-clamp experiments, outward currents could be recorded, but no inward currents could be. In contrast, smooth-muscle cells did fire spontaneously and had inward, as well as outward, current. The myocytes were contractile in response to depolarization, but the ILCs were not. Thus, we suggest that the spontaneous electrical behavior exhibited by the myometrium is an inherent property of the smooth-muscle cells within the myometrium. There are relatively few direct studies of this proposal, although earlier electrophysiological studies [21] certainly show pacemaker potentials presumed to arise from impaled myocytes.
Physiological Implications
As reported by Faussone-Pellegrini and Thuneberg [22], general criteria for typical ICCs can be given, but there remains a gray zone. There now appears to be ample evidence for such a gray zone for ICC-like cells in smooth muscles (our data and [3, 4, 6]). Given the different properties of these ILCs, despite their morphological and other similarities, it seems apparent that smooth muscles contain a range of such cells, from pacemakers to unknown support cells. Povstyan et al. [6] demonstrated no difference in the electrophysiological characteristics between portal vein smooth-muscle cells and the ILCs and could find no evidence for the cells being pacemakers. A similar conclusion was reached for mesenteric ILCs [7]. In the uterus of both rat and human, there is clearly a distinct population of cells seen both in situ and upon cell dispersion: they express K+ channels and vimentin and make close contact with myocytes, but a clear functional role for them remains elusive. We suggest that the intrinsic activity in uterine muscle is the property of the smooth-muscle cells themselves. It also remains to be seen if the ICC-like cells change their characteristics once labor has been initiated.
The close contact seen at both light and electron microscopic levels between the ILCs and myocytes suggests some degree of communication or coupling between the two cell types. Could this be an inhibitory signaling pathway? ICCs have also been associated with relaxation in gut smooth muscle via influencing neurotransmission [20, 23]. In particular, it is the NO component of relaxation that seems to be targeted. In the bladder and proximal urethra, Smet et al. [8, 11] described interstitial cells, present in large numbers with extensive dendritic arborization and staining heavily for cGMP, as well as vimentin. The ILCs in the uterus had large outward currents. Perhaps they perform an inhibitory role on intrinsic uterine activity. We suggest that the intrinsic activity is the property of the smooth-muscle cells.
| FOOTNOTES |
|---|
2 Correspondence. FAX: 44 151 794 5321; s.wray{at}liv.ac.uk ![]()
Received: 24 June 2004.
First decision: 14 July 2004.
Accepted: 17 August 2004.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
S. Allix, E. Reyes-Gomez, G. Aubin-Houzelstein, D. Noel, L. Tiret, J.-J. Panthier, and F. Bernex Uterine Contractions Depend on KIT-Positive Interstitial Cells in the Mouse: Genetic and Pharmacological Evidence Biol Reprod, September 1, 2008; 79(3): 510 - 517. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. P Fischer, J. A Hutchinson, D. Farrar, P. J O'Donovan, D. F Woodward, and K. M Marshall Loss of prostaglandin F2{alpha}, but not thromboxane, responsiveness in pregnant human myometrium during labour J. Endocrinol., April 1, 2008; 197(1): 171 - 179. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Seda, F. M Pinto, S. Wray, C. G Cintado, P. Noheda, H. Buschmann, and L. Candenas Functional and Molecular Characterization of Voltage-Gated Sodium Channels in Uteri from Nonpregnant Rats Biol Reprod, November 1, 2007; 77(5): 855 - 863. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Shmygol, K. Noble, and S. Wray Depletion of membrane cholesterol eliminates the Ca2+-activated component of outward potassium current and decreases membrane capacitance in rat uterine myocytes J. Physiol., June 1, 2007; 581(2): 445 - 456. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Smith Parturition N. Engl. J. Med., January 18, 2007; 356(3): 271 - 283. [Full Text] [PDF] |
||||
![]() |
T. B. Bolton Calcium events in smooth muscles and their interstitial cells; physiological roles of sparks J. Physiol., January 1, 2006; 570(1): 5 - 11. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. McHale, M. Hollywood, G. Sergeant, and K. Thornbury Origin of spontaneous rhythmicity in smooth muscle J. Physiol., January 1, 2006; 570(1): 23 - 28. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |