Interstitial cells of Cajal: a new perspective on smooth muscle function

Abstract

Interstitial cells of Cajal (ICC) were described more than 100 years ago by Ramon y Cajal. For many years these cells were identified only by non-specific histological stains and later, more reliably, by electron microscopy. Ultrastructural features and the anatomical locations of ICC suggested important physiological roles for these cells. A breakthrough occurred in our ability to study ICC when it was recognized that antibodies for Kit could be used to identify ICC, even in living tissues. Signalling via Kit, a receptor tyrosine kinase, is also necessary for ICC development and maintenance of phenotype. Thus, blocking Kit, by a variety of techniques, caused loss of ICC in experimental animals and demonstrated the critical physiological functions of these cells in gastrointestinal motility. Loss of ICC in human gastrointestinal diseases may contribute to the motor pathologies observed. Unrestrained Kit signalling leads to the transformation of ICC and the development of gastrointestinal stromal tumours. Now ICC-like cells have been identified in a variety of smooth muscle tissues, and the race is on to discover whether these cells have equivalent or even novel functions in organs outside the gastrointestinal tract. This perspectives article gives a short overview of the history of ICC research and directions for future investigation.

A hypothesis emerges

Ramon y Cajal described nerve-like cells at the ends of motor neurons in organs innervated by peripheral nerves (cf. Cajal, 1911). These cells, which have been best described in the gastrointestinal (GI) tract, now provide promising new explanations for the motor physiology and pathophysiology of the hollow organs.

ICC were first identified by standard histochemical techniques that had been developed by classical histologists to identify neurons (e.g. silver staining, zinc iodide/osmic acid, and methylene blue). The co-staining of neural processes made it difficult to describe the organization of ICC and cellular associations definitively. With the advent of electron microscopy, it became possible to describe the ultrastructure of the cells observed by light microscopy in greater detail (cf. Rogers & Burnstock, 1966; Imaizumi & Hama, 1969; Gabella, 1972; Cook & Burnstock, 1976). Morphological features of ICC, such as gap junctions with smooth muscle cells and very close apposition with nerve terminals were consistent with functional roles for these cells in GI organs. Ultrastructural criteria became the gold standard for identification of ICC (see Christensen, 1992).

Close association of ICC with nerve terminals reinforced the view of Cajal that these cells may be involved in neurotransmission (Yamamoto, 1977). Gap junctions with smooth muscle cells, and collections of these cells into specific layers within GI muscles suggested that ICC might be pacemakers and provide a ‘conductive function’ (Faussone Pellegrini et al. 1977; Thuneberg, 1982). The role of ICC as pacemakers was an old idea that had originated in the early 1900s (e.g. Keith, 1915). Other structures and connectivity with surrounding cells also suggested that ICC could be stretch receptors (Thuneberg, 1989). All of these ideas, conceived from morphological examination, were awaiting appropriate tools to unravel the functional roles of these cells.

Electron microscopy allows rigorous examination of cellular structures and cell-to-cell interactions, but this method of visualization is limited, without great time and determination, in obtaining a widespread image of cellular distribution and function. It is also a technique that is too labour intensive to utilize routinely for diagnostic purposes. Evidence for the anticipated roles of ICC in function languished because practical and specific research methods to study these cells were not available. Methylene blue was purported to specifically damage ICC (see Thuneberg, 1989), but there are simply too many non-specific effects of this substance for it to be reliable (e.g. Sanders et al. 1989). The observant eyes of Philip Langton identified cells of unusual structure in dispersed populations of colonic smooth muscles. Electron microscopy verified that these cells were ICC, and we showed for the first time that ICC were electrically rhythmic (Langton et al. 1989). In other words ICC could be pacemakers as morphologists had suggested! Picking ICC from mixed populations of dispersed cells on the basis of ultrastructure is slow and prone to error, however, and better techniques were needed.

The big break

So often in science paradigm-changing discoveries come from unsuspecting quarters. The ICC field is a prime example of this phenomenon. A team of Japanese scientists, working under the leadership of Shin-Ichi Nishikawa, developed reagents to test the role of the protooncogene product Kit. Kit had been implicated as a critical growth-factor-signalling pathway in haematopoesis, but other cells in the body also express Kit. So these investigators developed a neutralizing antibody to block Kit signalling to determine the consequences of reduced Kit function (Nishikawa et al. 1991). Administration of Kit-neutralizing antibodies to newborn animals resulted in the development of pathological gastric and intestinal distension and abnormal contractile behaviour within 10 days of birth (Maeda et al. 1992). Our group was fortunate to collaborate with Professor Nishikawa at this time, and we found that a majority of the Kit-positive cells in the GI tract were ICC. Blocking the Kit pathway greatly reduced the population of ICC in the GI tract and negatively impacted neural responses and abolished pacemaker activity (Torihashi et al. 1995). Kit mutant offspring, such as W/W^V mice, have compromised Kit function, lack certain classes of ICC and have profound pacemaker and motor neurotransmission defects (Ward et al. 1994; Huizinga et al. 1995). At last there were reagents to label ICC, methods to specifically manipulate ICC populations, and animal models to study the role of ICC (and the functions of these cells) in GI physiology, and progress defining the physiology of ICC proceeded rapidly after these discoveries.

There are at least two populations of ICC in the GI tract. Some cells lie between or at the edges of the muscle layers and generate and propagate electrical slow wave activity, and others lie within muscle bundles in close apposition with enteric neurons. W/W^V mice lack the ICC between the muscle layers in the small intestine and lose the ability to generate pacemaker activity (Ward et al. 1994; Huizinga et al. 1995). In contrast, these animals lose intramuscular ICC in the stomach and have greatly decreased responses to enteric motor neurotransmission (Burns et al. 1996; Ward et al. 2000). These studies suggested a ‘division of labour’ between the classes of ICC (Sanders, 1996), but this idea has been supplemented by the recent work of David Hirst and colleagues. Intramuscular ICC are not without rhythmic potential, however, and they generate fundamental rhythmic oscillations known as unitary potentials (Edwards et al. 1999). These events can summate in response to depolarization or neural inputs, and actually drive the pacemaker cells between the muscle layers that are normally dominant (Hirst et al. 2002). Intramuscular ICC are also stretch receptors that can influence the membrane potential (i.e. regulate the excitability of the entire ICC–smooth muscle syncytium) and regulate the frequency of pacemaker activity (Won et al. 2005).

It was not realized immediately after the discovery of the importance of Kit in ICC development that Kit mutants are important new models for intestinal pseudo-obstruction, gastroparesis, pyloric stenosis, lower oesophageal sphincter achalasia, and a variety of other GI motility disorders. However, Kit antibodies were soon used to label ICC in humans, and pathologists began to report reduced ICC populations in muscles of patients with a variety of motility disorders (see Vanderwinden & Rumessen, 1999; Sanders et al. 2006). It is interesting to note that some disorders thought for decades to be enteric neuropathies may be related to loss of ICC (cf. Ördög et al. 2000). Losing the subpopulation of ICC that mediates neurotransmission will have functional consequences rather similar to loss of motor neural innervation. Could loss of ICC explain other vague motor defects in smooth muscle organs observed in the human population?

The plot expands

Based on observations from the gut, investigators of other smooth muscle tissues began to wonder whether ICC-like cells might be distributed and contribute to physiological regulation in other hollow organs. Reports began to appear, and this issue of The Journal of Physiology provides further evidence that functional populations of pacemaker cells and Kit-positive cells are present in a variety of smooth muscle tissues (e.g. Sergeant et al. 2000; Exintaris et al. 2002; McCloskey et al. 2002; Harhun et al. 2004; Lang & Klemm, 2005; Brading & McCloskey, 2005). At present, ICC-like cells have been reported in muscles throughout the urinary tract, reproductive tract, and vascular system (blood vessels and lymphatics). Kit is not easy to resolve in some muscles, and Kit may not be universal as an obligatory signalling pathway for ICC-like cells. In fact, there is even variability to Kit sensitivity within the GI tract, and different populations of ICC show substantial variability in their responses to blocking Kit signalling (see Ward et al. 1994). Thus, lack of resolution of Kit immunoreactivity does not mean that ICC-like cells are not present. Clearly, more and better immunoreactive markers are needed to understand the role of these cells in some organs. It should be noted that there have also been morphological descriptions of cells, referred to as interstitial cells of Cajal, in locations other than smooth muscle tissues (e.g. Hinescu & Popescu, 2005; Popescu et al. 2005). There is currently no information about the physiological role of ICC in these locations.

The cell biology of specific cell types is greatly advanced by the development of techniques to isolate and culture cells. Isolation of ICC has been difficult. When Kit was discovered as a marker for ICC, we were enthusiastic, because of the extracellular epitopes for some antibodies, that a simple means of cell identification was possible. Unfortunately, extracellular epitopes of Kit appear labile to enzymatic digestion, and it is difficult to resolve freshly dispersed ICC with this approach. ICC can be cultured, however, and the pacemaker function of these cells is retained for several days in culture conditions (Thomsen et al. 1998; Koh et al. 1998). This has made it possible to study the mechanism of spontaneous pacemaker activity, and mathematical models of ICC pacing are even beginning to emerge (Imtiaz et al. 2002; Youm et al. 2006; Edwards & Hirst, 2006). ICC can also be sorted by automated fluorescence cell sorting techniques (Ördög et al. 2004), and new techniques, using uptake of fluorescent neurotransmitters, have provided a means to isolate specific classes of ICC. Soon modern genomics techniques will make it possible to investigate important questions about what makes ICC unique and what makes ICC susceptible to disease and environmental factors such that these cells disappear in pathophysiological conditions.

The issue of development of ICC is extremely important as cases in humans have suggested that ICC can be affected during either development or during adulthood. In the small intestine of mice, ICC emerge from a population of Kit-positive precursor cells at about embryonic day 15 (Torihashi et al. 1997). At about this time a lineage decision is made where Kit-positive cells signalled via Kit retain Kit expression and develop into functional ICC. Kit-positive precursors that are not signalled via this pathway become smooth muscle cells. Kit signalling is critical during the late gestational period and after birth for the development and maintenance of ICC populations (Torihashi et al. 1995, 1999). Loss of Kit later in life might be explained by a breakdown in Kit signalling, and even upstream effects such as loss of Kit ligand (stem cell factor) might cause a defect in Kit signal. This appears to be the result in diabetic gastroparesis where the loss of insulin or IGF-1 signalling causes reduced stem cell factor expression and loss of ICC (Horvath et al. 2006). It is also possible that inflammatory mediators, such as nitric oxide, could have deleterious effects on ICC, as in post-surgical ileus (Yanagida et al. 2004).

ICC display a high degree of plasticity, and loss of these cells in pathological conditions does not necessarily mean permanent loss. We found, for example, that treatment of muscles with a neutralizing Kit antibody resulted in nearly quantitative loss of ICC, but the affected cells did not die (Torihashi et al. 1999). In fact, when Kit was blocked ICC appeared to transdifferentiate toward a smooth muscle-like phenotype. Abolition of the conditions that block Kit causes regeneration of ICC (S. M. Ward & K. M. Sanders, unpublished observations). Regeneration of ICC also occurred in an animal models of intestinal obstruction and intestinal resection (Chang et al. 2001; Yanagida et al. 2004). Removal of the obstruction or healing after surgery caused restoration of functional ICC. One wonders if the restoration of stem cell factor in diabetes or reduction in inflammatory factors in postsurgical ileus or inflammatory bowel disease might lead to restoration of functional ICC in patients. Will it soon be possible for patients with motor defects and ICC loss to recover if conditions contrary to the ICC phenotype can be improved? There is some suggestion that motor defects in non-GI muscles may reflect changes in ICC populations or function (cf. Brading & McCloskey, 2005). It is likely that associations between defects and ICC will continue to emerge during the next decade. Research into the factors affecting ICC development, ICC depletion in disease models, and regeneration of the ICC phenotype are exciting directions for future research in this field.

Too much of a good thing

As proven regularly at pubs during meetings of The Physiological Society, too much of a good thing can be deleterious. While critical for the development and maintenance of the ICC phenotype, abnormal enhancement in Kit function is dangerous to the GI tract and the individual. Kit-positive cells can transform and form gastrointestinal stromal tumours (GIST). A subset of GISTs are due to a variety of polymorphisms in c-kit genes that encode Kit receptors that are constitutively active (cf. Hirota et al. 1998; Candelaria et al. 2005). Without the need for stem cell factor signalling, it is thought that ICC are transformed and form tumours that can become life threatening. Inhibition of the receptor tyrosine kinase function of Kit by imatinib mesylate has profound therapeutic effects on GISTs. These are the first solid tumours that are effectively treated by a drug with a specific molecular target.

The long and winding road ahead

Having had long-term interest in studies of ICC, we are invigorated by the range of biological function these cells appear to affect. We now know that: (i) ICC are pacemakers and actively propagate electrical slow waves in GI muscles; (ii) ICC, to a large extent, mediate both inhibitory and excitatory motor neurotransmission; (iii) ICC serve as non-neural stretch receptors in GI muscles, affecting both smooth muscle excitability and slow wave frequency; and (iv) ICC form intimate associations with the intramuscular terminals of vagal afferents, thus they may also have a role in afferent signalling. At present, most of the functions of ICC have been deduced from studies of mice. Thus, it is important to determine whether the physiology of ICC is species dependent and whether ICC express similar mechanisms and functions in other animals and humans. It should be noted that the similarities in ICC morphology and interactions with surrounding cells suggest similar functions. It will also be important to determine whether ICC-like cells function in non-GI muscles in an equivalent manner to those in the gut, or whether these cells provide even more diverse functions in other organs.

Controversy remains about the basic mechanisms that are responsible for the physiological role of ICC. For example, most of the steps in pacemaking have been described, but specific ionic conductances, molecular entities, and the exact sequence of events to produce pacemaker rhythmicity are still debatable. What are the postjunctional effects of neurotransmitters in ICCs; what ionic mechanisms explain inhibitory and excitatory junction potentials? What is the role of non-electrical mechanisms in neurotransmission; are ICC involved in this aspect of neural signalling? What proteins participate in mechano-sensitive responses of ICC? Do ICC also have secretory or paracrine functions? Are the mechanisms of rhythmicity and neurotransmission common to non-GI ICC?

Finally, in terms of translational biology, the issue of plasticity is the most important question in the ICC field. The factors and circumstances responsible for loss and recovery of ICC are of intense interest to biomedical scientists hoping to translate the observation that ICC decrease in motility disorders to information that is useful to physicians treating patients with dysmotilities. This work will require more animal models and research tools for manipulating the microenvironments of ICC in vivo and in vitro. These tools may be engines for the next important discoveries in the ICC field. There are likely to be new revelations about the role of ICC in human motor disorders in organs other than the GI tract. It might be prudent to investigate the state of ICC networks in a variety of disorders associated with autonomic and enteric neuropathies, electrical and mechanical dysfunction, and inappropriate responses to neurotransmitters and stretch. The next clinically important breakthroughs in this area are likely to come from observant investigators with tissues, animals or patients with difficult to explain smooth muscle pathologies.