MICROSCOPY RESEARCH AND TECHNIQUE 47:325–335 (1999) Receptors in Interstitial Cells of Cajal: Identification and Possible Physiological Roles MARIA-GIULIANA VANNUCCHI* Department of Anatomy, Histology and Forensic Medicine, Section of Histology ‘‘E. Allara,’’ University of Florence, Florence, Italy KEY WORDS neural transmission; Kit-receptor; NK1 receptor; nitric oxide; somatostatin ABSTRACT Interstitial cells of Cajal (ICCs) are specialized cells of the gastrointestinal tract forming distinct populations depending on their location in the gut wall. Morphological observations and functional data have led to the hypothesis of two functions for the ICCs: (1) as pacemakers of the rhythmic activity; (2) as intermediaries in neural inputs to the muscle. The identification of specific receptors on the ICCs has represented an important step in the knowledge of these cells. Immunohistochemical labeling of these receptors provided information on both ICC morphology and contacts (particularly those with nerve endings) and on the functions of these cells. All ICC possess the Kit receptor, which represents the best tool to identify these cells under the light microscope. It has been demonstrated that this receptor is essential for ICC differentiation, and, by using mutant mice lacking the Kit-related gene, it has been possible to discriminate among all the ICC those with a primary role as pacemakers. The ileal ICC, in particular those at the deep muscular plexus, express the tachykinin receptor NK1 and a subtype of somatostatin receptors and contain nitric oxide synthase. All these data support a primary role of these ICC in neural transmission. Microsc. Res. Tech. 47:325–335, 1999. r 1999 Wiley-Liss, Inc. INTRODUCTION One century ago Ramon y Cajal first described thin, elongated and branched cells within the gastrointestinal wall that, on the basis of their shape and location, he interpreted as peculiar neuronal cells. Although Cajal identified these cells using a light microscope (LM) after staining with methylene blue, the definitive technique to unequivocally identify the ‘‘interstitial cells of Cajal’’ (ICC) is the transmission electron microscopy (TEM). TEM allows one to see the peculiar cytological features and the ‘‘interstitial’’ position of the ICC between nerve endings and smooth muscle cells. These ultrastructural aspects led several authors to suggest that the ICC could be involved in intestinal pacemaker activity and neural transmission. In the last decade, giant steps have been made in the knowledge of the ontogeny, structural characteristics, and related functions of ICC. It has been shown that these cells possess receptors responsible for (1) their development and differentiation (Kit receptor); (2) their putative role in excitatory (NK1 receptor) and inhibitory (NO) neural transmission; and (3) their putative ability to respond to hormones (somatostatin receptor). KIT RECEPTOR The kit receptor (kit-r) is encoded by the protooncogene c-kit located on the W (white spot) locus of chromosome 5 of mouse; it is a tyrosine kinase similar to other receptors for growth factors whose interactions with the ligands initiate signaling via kinase cascade to the cell nucleus. The discovery that ICC express the kit-r has represented a fundamental advance in the study of these cells. r 1999 WILEY-LISS, INC. Kit-r and ICC Identification As previously stressed, ICC are identifiable unequivocally only under the TEM. The fact that the kit-r in the gut is expressed by ICC provides a way to identify these cells under the LM. Although other cells such as the mast-cells are kit-r positive, their shape and localization should avoid confusing them with ICC. Therefore, it is generally accepted that the kit-r immunolabeled cells are ICC (Lavin et al., 1998). In this regard however, it is necessary to consider some concern that the kit-r study alone could lead to erroneous conclusions. First, owing to the great number of kit-r antibodies commercially available, the possibility of a low affinity or specificity of some of them has to be kept in mind when ICC are studied using the LM only. To avoid wrong interpretations, it is necessary to use different kit-r antibodies whenever possible. Secondly, absence of kit-r labeling might be associated with the presence of otherwise normal ICC (Klüppel et al., 1998). Finally, kit-r positive labeling may not necessarily imply a structural and functional integrity of ICC. In conclusion, although the kit-r immunohistochemistry maintains its importance in LM investigations, the study of ICC, especially in pathological conditions, still requires the contribution of the ultrastructural examination. Abbreviations used: ICC-DMP ⫽ ICCs within the deep muscular plexus region of the small intestine; ICC-MP ⫽ ICCs within the myenteric plexus region of stomach, small intestine and colon; ICC-IM ⫽ ICCs intramuscular of the lower esophagus and pyloric sphincter regions, stomach, and colon; ICC-SM ⫽ ICCs along the submucosal surface of the circular muscle bundles of the colon. *Correspondence to: Maria Giuliana Vannucchi, Department of Anatomy, Histology and Forensic Medicine, Section of Histology ‘‘Enrico Allara,’’ Viale G. Pieraccini, 6 I-50139 Florence, Italy. E-mail: email@example.com Received 22 February 1999; accepted in revised form 1 August 1999 326 M.-G. VANNUCCHI Kit-r and the Various ICC Populations Combined morphological and functional studies of ICC have demonstrated the existence of distinct populations of ICC in each gut region and muscle layers, and distinct roles in intestinal motility have been attributed to each population (Faussone-Pellegrini, 1992; Thuneberg, 1989). The results obtained in studies on kit-r seem to confirm such a hypothesis. In the ileum, for example, where two types of ICC (ICC-MP and ICCDMP) have been described, the investigations done in W mutant mice and in anti-kit-r treated animals have shown significant differences between the two ICC populations: (1) in sensitivity to kit-r expression to reach complete differentiation during the early stages of postnatal life (Torihashi et al., 1995); (2) in the persistence of the ICC-DMP and the loss of the ICC-MP in adult mice with the W mutation (Ward et al., 1994; Torihashi et al., 1997). Consistent with these observations, we report here a different kit-r labeling distribution and intensity between the two ileal ICC populations of three different animal species (see Experimental Procedures). Kit-r and ICC Ontogeny and Differentiation The use of kit-r molecular biology and immunohistochemistry in grafts of intestinal segments (Young et al., 1996) or in quail-chick chimeras for vagal neural crest (Lecoin et al., 1996) have demonstrated that ICC: (1) originate from gut mesenchyma and not from the neural crest cells that colonize the gut; (2) develop in the absence of neurons that are considered the source of the stem cell factor (Steel factor), the ligand for the kit-r. Various attempts to evaluate the importance of kit-r in ICC differentiation have demonstrated that, although kit-r is already expressed by these cells in the embryo, its presence is not necessary for their migration, proliferation, survival (Bernex et al., 1996), and early differentiation (Klüppel et al., 1998). On the contrary, the presence of kit-r becomes necessary after birth (2–5 days) to guarantee cell division and ICC network formation (Klüppel et al., 1998). At this time, the cells are close to myenteric neurons and nerve endings that express the Steel factor (Torihashi et al., 1997). The binding of the ligand to the ICC receptor might be important for cell proliferation after birth. Kit-r and Pacemaker Function The observation that treating mice with kit-r antibodies led to the loss of the rhythmic activity in the mouse intestine (Maeda et al., 1992) suggested that cells located in the intestine normally contain the kit-r and are responsible for the slow-wave activity. Two groups showed by immunohistochemistry that the ICC express kit-r and that these cells disappear in the anti-kit-r treated (Torihashi et al., 1995) or W mutant mice (Huizinga et al., 1995; Ward et al., 1994). Besides, electrophysiological recordings confirmed that the lack of ICC was associated with the loss of the rhythmic activity (Huizinga et al., 1995; Ward et al., 1994) and demonstrated that neural inputs were also reduced, although not abolished, suggesting a role for ICC in neural transmission also (Torihashi et al., 1995) (see below). ICCS AND NEURAL TRANSMISSION The consistent relationship of ICC with nerve endings on one side and with the smooth muscle cells on the other side suggested that these cells are potential intermediaries in autonomic neurotransmission (Daniel and Posey-Daniel, 1984). Indeed, electrophysiological studies (Berezin et al., 1990; Huizinga et al., 1990; Smith et al., 1989) have demonstrated that ICCs are innervated and also possibly intercalated between nerve endings and smooth muscle cells. Furthermore, a combined morphological and functional study of the ileum of Steel mutant mice revealed a lack of the ICC-MP network and of the electrical rhythmicity even though ICC-DMP were maintained, as were the typical excitatory and inhibitory inputs (Ward et al., 1995). These findings indicate that (1) ICC-MP are not essential for mediating inputs from the enteric nervous system and (2) the ICC-DMP are the primary mediators in neural inputs. In another important study in animals treated with kit-r antibody after birth, Torihashi et al. (1995) found that several populations of ICC, such as all ICC-MP of the gut, ICC-DMP of the ileum, ICC-SM and ICC-IM of the colon, were reduced in number, and, in parallel, the slow-wave activity was abolished and the neural responses were reduced, although not abolished. Altogether these findings confirm that ICC can mediate neural inputs and that some of the ICC populations are more important than others in neural transmission. TACHYKININ RECEPTORS Three types of tachykinin receptors (NKr) have been described that bind with different affinities a series of neuropeptides called tachykinins including substance P (SP), neurokinin A (NKA), and neurokinin B (NKB). With the exception of the last, the tachykinins are commonly synthesized and stored in enteric neurons (Sternini et al., 1989). NKr are three distinct G-proteincoupled receptors (NK1r, NK2r, and NK3r with the highest affinity to SP, NKA, and NKB, respectively) (Maggi et al.,, 1993). Immunohistochemical studies have shown that all the three receptors are detectable in the gut (Grady et al., 1996) with a region-specific distribution (Vannucchi and Faussone-Pellegrini, unpublished data). However, only NK1r has been observed on ICC. NK1r and ICC-DMP All ICC are intimately associated with nerve endings. However, differences among the ICC types in the frequency of these contacts have been reported, i.e., between the ICC-MP and ICC-DMP in the small intestine or between the ICC-MP and ICC-IM of the stomach. These contacts are both more numerous and closer with ICC-DMP and ICC-IM than with ICC-MP, both in laboratory mammals and humans (Faussone-Pellegrini, 1992; Thuneberg, 1989). The presence of nerve endings close to ICC-MP suggested a neural control of slow-wave activity (Huizinga et al., 1997), in particular, of the premature or delayed slow wave (Cayabyab et al., 1997). Conversely, ICC-DMP and ICC-IM, which show a larger number of contacts with nerve ending, could be the intermediate step between enteric neurons and smooth muscle cells in the control of gut motility. In this regard, however, it has been reported that, in the dog RECEPTORS IN INTERSTITIAL CELLS OF CAJAL ileum, ICC-DMP produced slow waves of a different character from those produced by the ICC-MP (Jimenez et al., 1996). Indeed, immunohistochemical studies have shown that ICC-DMP are closely apposed to SP positive nerve endings and express its preferred receptor, the NK1r (Portbury et al., 1996; Sternini et al., 1995; Vannucchi et al., 1997; Vannucchi and FaussonePellegrini, unpublished data). Recently, a combined study of immunohistochemistry and pharmacology in the guinea pig ileum, has demonstrated that the NK1r on ICC-DMP is a true receptor since in the presence of the agonist (SP), it aggregates on the plasma membrane and internalizes in the cytoplasm of ICC (Lavin et al., 1998). Moreover, these authors quantified the relationship between SP-IR nerve endings in the DMP and NK1r-IR related ICC and demonstrated that up to 90% of these nerve endings are apposed closely to ICC. It is noteworthy that neither NK2r nor NK3r antibodies label ICC-DMP (Grady et al., 1996; Vannucchi and Faussone-Pellegrini, unpublished data). A study of the time course of appearance of NK1r on ileal ICC during rat development has shown that ICC-DMP express NK1r very early after birth and reach the maximum intensity and distribution of IR at 1 week of postnatal life with no further changes in adult life. Interestingly, TEM studies have shown that during this period of life ICC-DMP acquire the adult features and that their differentiative steps are temporally related to those of the enteric plexuses (see Experimental Procedures). Unexpectedly, ICC-IM of the stomach, to which a role similar to that of the ICC-DMP has been attributed, never stained with the NK1r antibody. However, it has been recently reported that ICC-IM of the human stomach can be stained with a monoclonal antibody to NK1r (Smith et al., 1998). The possibility that the antibody used does not recognize a slightly modified NK1r in the stomach has to be considered. NK1r and ICC-MP In previous papers, it was reported that the NK1r-IR at the MP of the small intestine was detectable on ICC of the duodenum but not of the ileum (Portbury et al., 1996). In a more recent work, Lavin et al. (1998) showed that NK1r-IR was detectable in the cytoplasm of ICC-MP of the ileum, but only after the receptor internalization was induced by the agonist (SP), and suggested that the absence of immunoreactivity in resting conditions was due to a receptor conformation that masked the site for antibody interaction. Besides, they found that in the duodenum almost 40% of the SP-IR nerve endings detected in the tertiary plexus of the MP were closely apposed to ICC. In our laboratory, using the same antibody, we detected NK1r-IR on ICC-MP in laminae of rat and guinea pig ileum and showed that SP-IR nerve endings of the tertiary plexus of the MP were rarely close to the NK1r-IR ICC (see Experimental Procedures). This finding suggests that the difficulty in ICC-MP labeling by the available NKr1 antibodies might be due to the presence of few receptors that concentrate after internalization, rather than to changes in their conformation. A reduced number of NK1r on ICC-MP would better fit with the scarcity of closely apposed SP-IR nerve endings. 327 NK1r-IR was never observed in any ICCs population of the colon, in guinea pig (Portbury et al, 1996), rat (Grady et al., 1996), or mouse (personal observations). NITRIC OXIDE TRANSMISSION Nitric oxide (NO) is a gas that easily diffuses across biological membranes (Moncada et al., 1991). Many of its actions are mediated by activating soluble guanylate cyclase. The regulatory step of its production is a specific enzyme, nitric oxide synthase (NOS). Two types of NOS have been identified in biological tissues: a constitutive form and an inducible form. Both are di-oxygenases that synthesize NO in the presence of arginine, are cytosolic and require NADPH as a coenzyme. However, the constitutive NOS is calciumcalmodulin dependent and produces a picomolar amount of NO, while the inducible NOS is calcium-independent and produces larger amounts of NO when activated. Several isoforms of both enzymes have been found (Moncada and Higgs, 1993) but only the constitutive NOS has been identified in neurons (neuronal NOS), as well as in the endothelium, platelets, mesangial cells, and, membrane-bound, in skeletal muscle. The activity of constitutive NOS is normally inhibited at the resting calcium concentration present in cells. Its activation is dependent on events that increase the free calcium in the cytoplasm. NO and ICCs The presence of NOS-positive neurons in the gastrointestinal tract has been amply reported (for review see Lefebvre, 1995) and a loss of these neurons has been associated with pathological conditions such as achalasia, congenital esophageal stenosis, hypertrophic pyloric stenosis (Mearin et al., 1993; Vanderwinden et al., 1992), and an increase of them with idiopathic chronic constipation (Cortesini et al., 1995). Furthermore, numerous NOS-containing nerve endings have been described in the DMP of the ileum, apparently surrounding unstained cells identified under the TEM as ICCDMP (Matini and Faussone-Pellegrini, 1997). A significant increase in cGMP was reported in ICC of the small intestine and colon in response to exogenous NO (Shuttleworth et al., 1993; Young et al., 1993). The availability of mutant mice for the kit-r has allowed experiments to be carried out to demonstrate that ICC mediate NO transmission. In fact, in W mutants, the ICC-IM of the stomach fail to develop, and the NOdependent inhibitory neuroregulation is greatly reduced in spite of a normal presence and distribution of the NOS-positive neurons and nerve fibers. At variance with the results reported for the ileum, ICC-MP of the stomach are unaffected by the kit-r mutation (Burns et al., 1996). In a more recent study, the same group (Ward et al., 1998) has demonstrated the importance of ICC-IM in mediating nitrergic neurotransmission also at the level of the lower esophageal and pyloric sphincters. This group found that in W mutant mice the lack of ICC-IM was associated with a marked decrease in the response to nitrergic stimulation. In contrast to the findings of Burns et al. (1996) in the fundus, the loss of the ICC-IM in the sphincters did not completely abolish the inhibitory response, and it was still possible to slightly enhance muscle contractions adding NOS inhibitors. This finding suggested the existence of a 328 M.-G. VANNUCCHI parallel nitrergic innervation of ICC-IM and of smooth muscle cells at the sphincter levels. These data, and the observation of loss of ICC in tissues from patients affected by pyloric stenosis and achalasia, suggest an association between the status of ICC-IM and sphincter disorders. NOS-Positive ICC NOS-positivity has been described in the following ICC types: ICC-SM of the canine colon (Xue et al., 1994); ICC-MP of the rat ileum (Matini and FaussonePellegrini, 1997); ICC-IM of the mouse gastric sphincters (Ward et al., 1998), and ICC-DMP of the guinea-pig ileum (see Experimental Procedures). The significance of the presence of a constitutive NOS isoform in ICC is a matter of debate and some hypotheses have been proposed. Neural inputs into ICC can induce NO production, which in turn can cause smooth muscle relaxation (Ward et al., 1998). A second possible role of NOS in ICC could be to amplify inhibitory signals mediated by nitrergic pathways since it has been observed that NO increases calcium concentration in ICC, which leads to NO production by ICC (Sanders, 1996; Ward et al., 1998). A third possible function of NOS in ICC could be to modulate neural transmission by NO action on neurotransmitter release from nerve terminals, as has been reported in the central nervous system (Garthwaite, 1991). ICC AND SOMATOSTATIN (SOM) RECEPTORS Som is a fourteen-amino-acid polypeptide widely distributed in the nervous system as well as in the endocrine tissues. It acts as a neuromodulator in the nervous system and as a neurohormone in other organs. The presence of Som in the digestive system had been well described (Costa et al., 1980; Ekblad et al., 1988), but contrasting results have been obtained regarding its actions on intestinal smooth muscle. In fact, it has been reported that through activation of inhibitory neurons and/or inhibition of excitatory neurons, Som caused muscle relaxation in guinea-pig intestine in vitro (Furness and Costa 1979), while it inhibited NO release and thereby excited intestine circular muscle in canine ileum in vivo (Jimenez et al., 1995). It has been shown that rat ICC-DMP express a particular subtype of Som receptors, sst2A, and are surrounded by Som-IR nerve endings (Sternini et al., 1997). Interestingly, although the entire gut was investigated, the positivity for sst2A was present only on these ICC (besides the neurons and some endocrine cells). Moreover, sst2A-IR on ICC-DMP was confined to the plasma membrane while it appeared to be located in the cytoplasm of neurons and endocrine cells. These findings have been interpreted as further support for the possibility that Som influences smooth muscle activity by directly activating sst2A-r-IR ICC-DMP. ICCS AND OTHER RECEPTORS Several immunohistochemical studies indicate that other neurotransmitters may act on ICC. In particular, it has to be mentioned that nerve endings positive for ChAT (Vannucchi and Faussone-Pellegrini, 1996) or VIP (Berezin et al., 1990; Matini et al., 1997) are present at the DMP closely apposed to unlabeled cells possibly identifiable as ICC. Indeed, the presence of VIP receptors on ICC has been hypothesized on the basis of electrophysiological data (Berezin et al., 1990). At the present, however, the lack of specific antibodies to identify these receptors does not allow confirmation of the existence of any other receptor on ICC. EXPERIMENTAL PROCEDURES Tissue Peparation for Immunohistochemistry Specimens of guinea-pig, rat, and mouse ileum were cut to yield transverse and tangential sections following the procedures previously reported (Vannucchi et al., 1997), and some specimens of rat and guinea pig ileum were prepared as laminae containing either the submucosa plus the circular muscle layer or the longitudinal muscle layer plus the myenteric plexus following the procedure described in Portbury et al. (1996). All the specimens were fixed in 4% paraformaldehyde in 0.1 M phosphate buffered saline (PBS) pH 7.4, for 4 hours at 4°C. Then, those to be frozen were placed in 30% sucrose in PBS, overnight at 4°C. The following day these specimens were embedded in OCT compound (Miles, Elkhart, IN) and frozen at ⫺80°C. Transverse and tangential sections (14 µm thick) were obtained from the frozen specimens with a cryostat. After fixation, the laminae were carefully washed in PBS. Both cryostat sections and laminae were pre-incubated in PBS containing 3% normal goat serum and 0.5% Triton X-100 and incubated with primary antisera in moist chambers. Antibodies To label the neurokinin receptor-1, NK1r polyclonal antibody (a generous gift of Dr. P. Vigna) was used, raised in rabbit against a synthetic fragment corresponding to the intracellular C-terminal portion of the rat receptors (Grady et al., 1996; Vigna et al., 1994), at a final dilution of 1:2,500 for 24 hours at 4°C. The kit transmembrane receptor protein was labeled with the kit-r (Ab-1) (Calbiochem, San Diego, CA) polyclonal antibody raised in rabbit and used at the concentration of 1 mg/ml for 48 hours at 4°C. Substance P was identified by using a mouse anti-SP monoclonal antibody (SP14, a generous gift of Dr. J.Y. Couraud) at 1:400 overnight at 4°C. NOS anti-mouse monoclonal antibody (nNOS) (Transduction Laboratories) was used at a final dilution of 1:1,000 for 24 hours at room temperature. Negative controls were performed by omitting the primary antibodies or substituting them with a nonimmune rabbit or mouse serum in order to check the specificity of the immunostaining. At the end of incubation, the polyclonal primary antisera in the sections and laminae were revealed by using fluorescein (DTAF)conjugated pure goat anti-rabbit IgG (H⫹L; Jackson Immuno-Research, West Grove, PA) or rhodamine antimouse IgG (Fab specific) TRITC conjugate (Sigma, St. Louis, MO) secondary antibodies. RESULTS AND DISCUSSION ICC and KIT Receptors Kit-r Distribution in ICC-MP and ICC-DMP of Adult Guinea Pig, Rat, and Mouse Ileum. In the guinea pig, kit-r-IR was detected in ICC-MP and in ICC-DMP (Fig.1A), with the ICC-MP showing a higher intensity of staining. ICC-MP had a round or triangular body and two, sometimes three, rarely four, main processes (Fig.1D). The surface of the body and processes was smooth and kit-r-IR was distributed over the RECEPTORS IN INTERSTITIAL CELLS OF CAJAL 329 Fig. 1. Kit-immunoreactivity (IR) in transverse sections (A–C) and in the lamina containing the longitudinal muscle layer (D). A: Guinea pig. IR is detected in ICC-MP and ICC-DMP; the intensity of the labeling is higher in ICC-MP. B: Rat. IR is detected only in ICC-DMP; C: Mouse. IR is detected in ICC-MP and in ICC-DMP. The intensity of labeling is higher in ICC-MP. D: Guinea-pig. IR is detected on the ICC-MP body and processes and has a patchy distribution. A,B,C, ⫻1,000, scale bar ⫽ 10 µm. D, ⫻450, scale bar ⫽ 22 µm. entire cell surface, although the intensity of labeling was higher at the level of wide patches present on the cell body and on the roots of the main processes. The ICC-DMP had an oval or round body and 2–4 main processes, two of which always originated at the opposite poles of the cell. The surface of both the body and processes was smooth and kit-r-IR was detected over the entire cell surface and showed a punctate aspect (data not shown). In the rat, kit-r-IR was detected only on the ICC-DMP (Fig.1B). The cells had an elongated body and 2–3 main processes that ran parallel to the muscle cells. The kit-r-IR was detected along the entire cell surface and had a punctate appearance. In the mouse, kit-r-IR was detected on ICC-MP as well as on the ICC-DMP (Fig.1C); however, the ICC-MP showed a higher intensity. These data represent the first comparison of the kit-r distribution in the ileal ICC of three of the animal species most used in research. As reported before, several kit-r antibodies are commercially available. We used a kit-r antibody against the intracellular portion of the receptor that, in our hands, demonstrated good selectivity since other cells such as mast cells were very rarely labeled. This antibody showed interesting intraand inter-species differences in kit-r expression. In particular, in each species the labeling intensity and/or distribution always differed between the two ileal ICC networks, with the ICC-DMP showing a similar labeling in all the animals. Moreover, kit-r-IR at the MP had a patchy organization while it was evenly distributed at the DMP. These data are in agreement with those reported for W mutant mice and anti-kit-r treated newborns that indicate a diversity between the two ileal ICC networks in kit-r sensitivity and dependence during maturation and in the maintenance of the differentiated state. ICC and NK1 Receptors Timing of Appearance and Distribution of NK1 Immunoreactivity (IR) and Tachykinergic-(SP/ TK)-IR Nerve Fibers in the ICC of Rat Ileum During Ontogenesis and in Adult Rats, Mice, and Guinea Pigs. In 18-day-old fetuses and in the newborn, NK1r-IR was detectable only at the MP region as 330 M.-G. VANNUCCHI Fig. 2. Neurokinin 1 receptor immunoreactivity (NK1r-IR) in the rat ileum during development. A: Five days of postnatal life. IR is detected at the deep muscular plexus (DMP) on cells (asterisks) which probably are interstitial cells of Cajal (ICCs). B: Seven days of postnatal life. The IR at the DMP region is particularly intense. C: Fourteen days of postnatal life. ICCs at the DMP have a spindle shape. D: Adult rat. ICCs have a spindle shape and the IR has a punctate aspect. mp ⫽ myenteric plexus; dmp ⫽ deep muscular plexus; ocl ⫽ outer circular muscle layer. A,B, ⫻400; C,D, ⫻1,000. Scale bar ⫽ 25 µm (A,B), 10 µm (C,D). Reproduced from Vannucchi et al. (1997) with permission of the publisher. faint patches on the neuronal contour. From 3 to 5 days of postnatal life, the immunoreactivity became detectable in the DMP on non-neuronal cells, which probably corresponded to ICC. At 3 days of postnatal life these cells were dispersed. At 5 days (Fig. 2A), they formed a continuous layer between the inner and outer portions of the circular muscle layer. All these ICC had an elongated shape and the IR was irregularly distributed on their plasma membrane. At the end of the first week of postnatal life, the IR at the DMP region was particularly intense and all ICC showed the same high degree of labeling (Fig. 2B). These ICC were oriented in parallel to the circular muscle cells, forming a one-cellthick layer. The ICC were spindle-shaped and most of them had a labeling distributed uniformly over the entire plasma membrane . By 14 postnatal days (Fig. 2C), most of the NK1r-IR ICCs in the DMP region were similarly intensely and uniformly labeled. In the adult, the ICC maintained their arrangement in rows parallel to the circular muscle cells and their spindle shape and showed similar intensity of labeling to those observed during the suckling and weaning periods. The NK1r-IR, however, was irregularly distributed along the cell surface, with a punctate appearance (Fig. 2D). From 7 days of postnatal life to adult, no change in the intensity of labeling was observed. Similar to the rat, adult guinea pig and mouse showed very intense NK1-IR at the DMP. The IR had a punctate distribution over the plasma membrane. In the rat (and mouse) the cell body was elongated or spindle-shaped (Fig. 3A) and, in the rat, thin and ramified branches were seen to protrude laterally from the body and processes. In the guinea pig, the body of these cells appeared oval- or roundshaped with two to four main processes; two of them always originated at the opposite poles of the cell, and the surface of both the body and processes was smooth (Fig. 3C). In laminae obtained from guinea pig and rat ileum, the NK1r-IR was detectable also on ICC-MP, although these cells showed a very faint labeling (Fig. 3B,D). In rat, from birth, SP/TK-IR varicose fibers could be seen to emerge from the MP and to gradually penetrate the circular muscle layer towards the DMP, reaching it by 5 days (Fig. 4A). At 7 days of postnatal life (Fig. 4B), RECEPTORS IN INTERSTITIAL CELLS OF CAJAL 331 Fig. 3. Neurokinin 1 receptor immunoreactivity (NK1r-IR) in laminae of adult rat and guinea-pig ileum containing the circular muscle layer or the longitudinal muscle layer plus the myenteric plexus. A: NK1r-IR ICCs at the DMP of rat ileum. The cells have a spindle or elongated body, the processes are thick and possess several transverse and short branches. The cells appear regularly organized in rows. C: NK1r-IR ICCs at the DMP of guinea-pig ileum. The cells have a round or oval body and long and thin processes. The IR in both animals has a punctate aspect. B,D: NK1r-IR ICCs at the MP of rat (C) and guinea-pig (D) ileum. Both cells have a triangular shape and three processes (asterisks) and the labeling is extremely faint. A,B,C,D, ⫻1,000. Scale bar ⫽ 10 µm. SP/TK-IR varicose fibers were detected in the DMP region. Between the first and third postnatal weeks, a progressive enrichment in SP/TK-IR at the DMP was observed (Fig. 4C). In the adult animal, the DMP showed the most intense reactivity for SP/TK, now appearing in transverse sections as a continuous line of varicose nerve fibers (Fig. 4D). A similar labeling distribution and intensity were observed in the ileum of adult mice and guinea pigs. These data demonstrate that the NK1r-IR on the ICC-DMP and SP/TK-IR in nerve varicosities of the DMP appeared very early after birth. Interestingly, the intensity of the labeling for the NK1r was at its maximum at 7 days of postnatal life, while that for SP/TK reached the highest intensity in the adult age. It has been reported that at 7 days of postnatal life the ICC-DMP have already acquired adult features (Faussone-Pellegrini et al., 1996) and the high expression of the NK1r with a peculiar distribution, different from that observed in the adult animal, could imply a trophic role of these receptors during a crucial period of the ICC differentiation. Furthermore, immunohistochemical studies in the rat ileum have demonstrated that at this age the DMP shows NOS-IR, but not VIP or ChAT labeling (Matini et al., 1997; Vannucchi and FaussonePellegrini, 1996). These data, together with the reported results on the presence of SP/TK-IR at this age, also indicate that, very early in postnatal life, the DMP contains both excitatory (SP) and inhibitory (NO) nerve endings whose final targets seem to be ICC. The difference between the earliest ages and the adulthood in the NK1r distribution on ICC-DMP probably depends on a progressive membrane re-organization of the same receptors in a cell that increases in size and changes in shape. Relationship Between the Tachykinergic Innervation and the Two Ileal ICC Populations. Simultaneous labeling with NK1r (Fig. 5A,B) or with the kit-r polyclonal antibodies (Fig. 5C) and the SP monoclonal antiserum revealed that, in the laminae obtained from guinea-pig and rat ileum, numerous SP-IR varicose nerve fibers were closely apposed to the ICC-DMP body and processes and some of these fibers surrounded the entire cell body, especially in the guinea pig (Fig. 5B). In 332 M.-G. VANNUCCHI Fig. 4. Substance P/tachykinins-immunoreactivity (SP/TK-IR) in the rat ileum during development. A: Five days of postnatal life. IR varicose fibers (asterisks) emerging from the myenteric plexus, reach the DMP region. B: Seven days of postnatal life. The varicose IR fibers participate in the formation of the DMP and surround unlabeled cells. C: Twenty-one days of postnatal life. IR varicose fibers become to be uniformly distributed in the DMP region (asterisks). D: Adult rat. The DMP shows an uniform and intense IR. mp ⫽ myenteric plexus; dmp ⫽ deep muscular plexus; ocl ⫽ outer circular muscle layer; icl ⫽ inner circular muscle layer. A,B, ⫻1,000. Scale bar ⫽ 10 µm, C,D, ⫻400. Scale bar ⫽ 25 µm. Reproduced from Vannucchi et al. (1997) with permission of the publisher. contrast, in MP, most of the SP-IR nerve fibers were far from the ICC, and SP-IR nerve varicosities were rarely seen near these cells (Fig. 5C). The observation that no other ICC population in the gut but those in the ileum express the NK1r-IR (Portbury et al., 1996; personal observation) suggests that the NK1r-IR labeling represents a immunohistochemical marker to identify the ileal ICC. The marked difference in the labeling intensity between the ICCDMP and ICC-MP confirms that they are two distinct ICC populations, and indicates that NK1r-IR is particularly useful for identifying the ICC-DMP. The different NK1r labeling intensity of these cells has an important functional significance and supports the hypothesis that the two ICC populations play different roles. In fact, using the simultaneous labeling, it was observed that the intense NK1r-IR ICC-DMP were close to many SP positive nerve varicosities in both rat and guinea pig ileum, while the faint NK1r-IR ICC-MP were close to very few SP-positive nerve varicosities. These findings indicate that the ICC-DMP are implicated in the neurotransmission, whereas ICC-MP, considered as pacemak- ers (Huizinga et al., 1995; Ward et al., 1994), are little involved in neurotransmission. ICC and NOS NOS-IR was detected in guinea-pig ileum, in the cytoplasm of non-neuronal cells located at the level of the DMP that were identified as ICC (Fig. 6). The presence of NOS in some ICC populations has also been reported in other species (see Introduction), and the possible significance of this enzyme in ICC is discussed both in the Introduction and in the Concluding Remarks. CONCLUDING REMARKS Kit-r immunohistochemistry is the best tool to identify the ICC under the LM. The kit-r is involved in the development and maintenance of the differentiated state of ICC after birth. However, among ICC, ICCDMP seem to depend on the presence of kit-r for their differentiation only partially and during a very short period of postnatal life; besides, in adult W mutant mice these cells are normally organized. The existence of RECEPTORS IN INTERSTITIAL CELLS OF CAJAL Fig. 5. Neurokinin 1 receptor (NK1r) (green) and Substance P (SP) (red-orange) in the laminae of adult rat and guinea-pig ileum containing the circular muscle layer or the longitudinal muscle layer plus the myenteric plexus. A: NK1r-IR ICCs at the DMP of the rat ileum with closely apposed SP-IR varicose fibers. B: NK1r-IR ICCs at the DMP of Fig. 6. NOS-immunoreactivity (IR) in transverse sections of guineapig ileum. The ICC-DMP (icc-dmp) appeared intensely labeled. x1000. Scale bar ⫽ 10 mm. other, unidentified, local factors whose presence during the earlier stages of postnatal life could compensate the kit-r impairment and guarantee the ICC differentiation, has been hypothesized to explain this diversity. 333 the guinea-pig ileum are closely apposed to SP-IR varicose fibers that, in some instances, surround the entire cell body. C: NK1r-IR ICCs at the MP of guinea-pig ileum and SP-IR varicose fibers of the tertiary plexus. A–C: ⫻950, scale bar ⫽ 10 mm. The observation that very early after birth ICC-DMP express as much NK1r as in the adult, has suggested that this receptor could also have a trophic function during the earlier stages of life affecting the differentiation and network formation of this ICC population. Still not clear is the significance of the kit-r presence on ICC precursors from the first stages of the embryonic life up to birth. During this period of prenatal life it has been demonstrated that (1) the absence of kit-r does not apparently affect the normal development of ICC; and (2) ICCs develop normally during embryogenesis even in the absence of neurons that are the source of the kit-r specific ligand, the Steel factor. The presence of sst2A receptor on ICC confirms the complexity of the roles played by these cells that can also be intermediaries in Som effects on the gut muscles. The interesting observations that some receptors such as kit-r and NK1r, which belong to two different classes of receptors, are expressed by ICC very early during ontogenesis, before the corresponding ligand becomes necessary for cell differentiation (Kitr and Steel factor) or is available in the synaptic vesicle (NK1r and SP), need special attention. They suggest that the commitment to synthesize the appropriate molecules for the differentiation or for the functional acquisition by the cell is independent of the presence of the specific ligand and could be genetically determined, or is due to local agents that act at less differentiated stages. TEM investigations have shown that ICC are 334 M.-G. VANNUCCHI consistently close to nerve endings from the very first stages of development and could be the source of agents able to induce the receptor expression. Indeed, these contacts between the ICC-blast and the nerve endings have been interpreted as the means by which some of the blasts present become ICC instead of smooth muscle cells or fibroblasts. On the other hand, it could also be hypothesized that the ability of the differentiating ICC to express these receptors is, in turn, the way to influence the expression of the chemical codes in the closely apposed nerve endings. ACKNOWLEDGMENTS The author thanks Dr. L. Corsani in performing immunohistochemistry and P. Guasti and L. Calosi for assistance in preparing the photographs. REFERENCES Berezin I, Huizinga JD, Farraway L, Daniel EE. 1990. Innervation of interstitial cells of Cajal by vasoactive intestinal polypeptide containing nerves in canine colon. Can J Physiol Pharmacol 68:922–932. Bernex F, De Sepulveda P, Kress C, Elbaz C, Delouis C. 1996. Spatial and temporal patterns of c-kit-expressing cells in WlacZ/⫹ and WlacZ/WlacZ mouse embryos. Development 122:3023–3033. Burns AJ, Lomax AE.J, Torihashi S, Sanders KM, Ward SM. 1996. Interstitial cells of Cajal mediate inhibitory neurotransmission in the stomach. Proc Natl Acad Sci USA 93:12008–12013. Cayabyab FS, Jimenez M, Vergara P, de Bruin H, Daniel EE. 1997. Influence of nitric oxide and vasoactive intestinal peptide on the spontaneous and triggered electrical and mechanical activities of the canine ileum. Can J Physiol Pharmacol 75:383–397. Cortesini C, Cianchi F, Infantino A, Lise M. 1995. Nitric oxide synthase and VIP distribution in enteric nervous system in idiopathic chronic constipation. Dig Dis Sci 40:2450–2455. Costa M, Furness JB, Llewellyn-Smith IJ, Davies B, Oliver J. 1980. An immunohistochemical study of the projections of somatostatincontaining neurons in the guinea-pig intestine. Neuroscience 5:841– 852. Daniel EE, Posey-Daniel V. 1984. Neuromuscular structures in opossum esophagus: role of interstitial cells of Cajal. Am J Physiol 246:G305–G315. Ekblad E, Ekman R, Hakanson R, Sundler F. 1988. Projections of peptide-containing neurons in rat colon. Neuroscience 27:655–674. Faussone-Pellegrini MS. 1992. Histogenesis, structure and relationship of interstitial cells of Cajal (ICC): from morphology to functional interpretation. Eur J Morphol 30:137–148. Faussone-Pellegrini MS, Matini P, Stach W. 1996. Differentiation of enteric plexuses and interstitial cells of Cajal in the rat gut during pre- and postnatal life. Acta Anat 155:113–125. Furness JB, Costa M. 1979. Actions of somatostatin on excitatory and inhibitory nerves in the intestine. Eur J Pharmacol 56:69–74. Garthwaite J. 1991. Glutamate, nitric oxide and cell-cell signalling in the nervous system. Trends Neurosci 14:60–67. Grady EF, Baluk P, Bohm S, Gamp PD, Wong H, Payan DG, Ansel. J, Portbury. A.L, Furness JB, McDonald DM, Bunnett NW. 1996. Characterization of antisera specific to NK1 NK2, and NK3 neurokinin receptors and their utilization to localize receptors in the rat gastrointestinal tract. J Neurosci 16:6975–6986. Huizinga JD, Berezin I, Daniel EE, Chow E. 1990. Inhibitory innervation of colonic smooth muscle cells and interstitial cells of Cajal. Can J Physiol Pharmacol 68:447–454. Huizinga JD, Thuneberg L, Kluppel M, Malysz J, Mikkelsen HB, Bernstein A. 1995. W/Kit gene required for interstitial cells of Cajal and for intestinal pacemaker activity. Nature 373:347–349. Huizinga JD, Thuneberg L, Vanderwinden JM, Rumessen JJ. 1997. Interstitial cells of Cajal as targets for pharmacological intervention in gastrointestinal motor disorders. Trends Pharmacol Sci 18:393– 403. Jimenez M, Vergara P, Christinck F, Daniel EE. 1995. Mechanism of action of somatostatin on the canine ileal circular muscle. Am J Physiol 269:G22–G28. Jimenez M, Cayabyab FS, Vergara P and Daniel EE. 1996. Heterogeneity in electrical activity of the canine ileal circular muscle: interaction of two pacemakers. Neurogastroenterol Motil 8:339– 350. Klüppel M, Huizinga JD, Malysz J, Bernstein A. 1998. Developmental origin and kit-dependent development of the interstitial cells of Cajal in the mammalian small intestine. Dev Dyn 211:60–71. Lavin ST, Southwell BR, Murphy R. 1998. Activation of neurokinin 1 receptors on interstitial cells of Cajal of the guinea-pig small intestine by substance P. Histochem Cell Biol 110:263–271. Lecoin L, Gabella G, Le Douarin N. 1996. Origin of the c-kit-positive interstitial cells in the avian bowel. Development 122:725–733. Lefebvre RA. 1995. Nitric Oxide in the peripheral nervous system. Ann Med 27:379–388. Maeda H, Yamagata A, Nishikawa S, Yoshinaga K, Kobayashi S, Nishi K, Nishikawa S. 1992. Requirement of c-Kit for development of intestinal pacemaker system. Development 116:369–375. Maggi CA, Patacchini R, Rovero P, Giachetti A. 1993. Tachykinin receptors and tachykinin receptor antagonists. J Auton Pharmacol 13:23–93. Matini P, Faussone-Pellegrini MS. 1997. Ultrastructural localization of neuronal nitric oxide synthase-immunoreactivity in the rat ileum. Neurosci Lett 229:45–48. Matini P, Mayer B, Faussone-Pellegrini MS. 1997. Neurochemical differentation of rat enteric neurons during pre- and post-natal life. Cell Tissue Res 288:11–23. Matini P, Mayer B, Faussone-Pellegrini MS. 1997. Neurochemical differentiation of rat enteric neurons during pre- and post-natal life. Cell Tissue Res 288:11–23. Mearin F, Mourelle M, Guarner F, Salas A, Riveros-Moreno V, Moncada S, Malagelada JR. 1993. Patients with achalasia lack nitric oxide synthase in the gastro-oesophageal junction. Eur J Clin Invest 23:724–728. Moncada S, Higgs A. 1993. The L-arginine-nitric oxide pathway. N Engl J Med 329:2002–2012. Moncada S, Palmer RMJ, Higgs AE. 1991. Nitric Oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev 43:109–142. Portbury AL, Furness JB, Young HM, Southwell BR, Vigna SR. 1996. Localisation of NK1 receptor immunoreactivity to neurons and interstitial cells of the guinea-pig gastrointestinal tract. J Comp Neurol 367:342–351. Sanders KM. 1996. A case for interstitial cells of Cajal as pacemakers and mediators of neurotransmission in the gastrointestinal tract. Gastroenterology 111:492–515. Shuttleworth CW, Xue C, Ward SM, de Vente J, Sanders KM. 1993. Immunohistochemical localization of 3’,5’-cyclic guanosine monophosphate in the canine proximal colon: responses to nitric oxide and electrical stimulation of enteric inhibitory neurons. Neuroscience 56:513–522. Smith TK, Reed JB, Sanders KM. 1989. Electrical pacemakers of canine proximal colon are functionally innervated by inhibitory motor neurons. Am J Physiol 256:C466-C477. Smith VC, Sagot JY, Couraud AMJ. 1998. Localization of the neurokinin 1 (NK1. receptor in the human antrum and duodenum. Neurosci Lett 253:49–52. Sternini C, Anderson G, Frantz G, Krause JE, Brecha N. 1989. Expression of substance P/neurokinin A-encoding preprotachikinin messenger ribonucleic acids in the rat enteric nervous system. Gastroenterology 97:348–356. Sternini C, Su D, Gamp PD, Bunnett NW. 1995. Cellular sites of expression of the neurokinin1 receptor in the rat gastrointestinal tract. J Comp Neurol 358:531–540. Sternini C, Wong H, Wu SV, De Giorgio R, Yang M, Reeve J Jr, Brecha NC, Walsh JH. 1997. Somatostatin 2A receptor is expressed by enteric neurons, and by interstitial cells of Cajal and enterochromaffin-like cells of the gastrointestinal tract. J Comp Neurol 386:396– 408. Thuneberg L. 1983. Interstitial cells of Cajal. In Wood JD, editor. Handbook of physiology. The gastrointestinal system, Vol. 1. Bethesda: Am Physiol Soc; p 349–386. Torihashi S, Ward SM, Nishikawa SI, Nishi K, Kobayashi S, Sanders KM. 1995. kit-r dependent development of interstitial cells and electrical activity in the murine gastrointestinal tract. Cell. Tissue Res 280:97–111. Torihashi S, Ward SM, Sanders KM. 1997. Development of c-kitpositive cells and the onset of electrical rhythmicity in murine small intestine. Gastroenterology 112:144–155. Vanderwinden JM, Mailleux P, Schiffmann SN, Vanderhaeghen JJ, De Laet M.H. 1992. Nitric oxide synthase activity in infantile hypertrophic pyloric stenosis. N Engl J Med 327:511–515. Vannucchi MG, Faussone-Pellegrini MS. 1996. Differentation of cholinergic cells in the rat gut during pre and post-natal life. Neurosci Lett 206:105–108. Vannucchi MG, De Giorgio R, Faussone-Pellegrini MS. 1997. NK1 receptor expression in the interstitial cells of Cajal and neurons and RECEPTORS IN INTERSTITIAL CELLS OF CAJAL tachykinins distribution in rat ileum during development. J Comp Neurol 383:153–162. Vigna SR, Bowden JJ, McDonald DM, Fisher J, Okamoto A, McVey DC, Payan DG, Bunnett NW. 1994. Characterization of antibodies to the rat substrate P (NK1) receptor and to a chimeric substance receptor expressed in mammalian cells. J Neurosci 14:834–845. Ward SM, Burns AJ, Torihashi S, Sanders KM. 1994. Mutation of the proto-oncogene kit-r blocks development of interstitial cells and electrical rhythmicity in murine intestine. J Physiol (Lond) 480: 91–97. Ward SM, Burns AJ, Torihashi S, Harney SC, Sanders KM. 1995. Impaired development of interstitial cells and intestinal electrical rythmicity in steel mutants. Am J Physiol 269:C1577–C1582. 335 Ward SM, Morris G, Reese L, Wang XY, Sanders KM. 1998. Interstitial cells of Cajal mediate enteric inhibitory neurotransmission in the lower esophageal and pyloric sphincters. Gastroenterology 115:314– 329. Xue C, Pollock J, Schmidt HH.H.W, Ward SM, Sanders KM. 1994. Expression of nitric oxide synthase by interstitial cells of the canine proximal colon. J Auton Nerv Syst 49:1–14. Young HM, McConalogue K, Furness JB, de Vente J. 1993. Nitric oxide targets in the guinea-pig intestine identified by induction of cyclic GMP immunoreactivity. Neuroscience 55:583–596. Young HM, Ciampolini D, Southwell BR, Newgreen DF. 1996. Origin of interstitial cells of Cajal in the mouse intestine. Dev Biol 180:97–107.