The Physiology and Pathophysiology of Interstitial Cells of Cajal: Pacemaking, Innervation, and Stretch Sensation

13.1.1

The Early Discussions on Gut Pacemaker Activity

In the early 20th century, Walter Alvarez realized that electrical activity was underlying the movements of the gut organs and started to record “action currents.” He soon understood that there was a disconnect between the electrical activity and the contractile activity since often the electrical activity would continue even though the contractions waned. He also discovered that the frequency of contractions in the intestine was higher in the proximal compared to the distal end and so the concept of pacemaking in the gut and a key component of the pacemaker network, its frequency gradient, was born. Hence, rhythmic smooth muscle depolarizations (slow waves) were underlying the rhythmic contractions of the intestine. Nevertheless, Alvarez struggled with the physiology of peristalsis and he and Alexander Keith discussed a potential role for specialized pacemaker cells, which Keith had postulated based on anatomical observations and comparisons with the sinus node. In the end, Alvarez rejected Keith’s ideas based on the fact that he could cut the intestine into sections and every section would show peristalsis. Alvarez assumed that according to Keith’s ideas there would be a single-dominant pacemaker (a gut sinus node) that would govern the rest of the intestine, even though Keith suggested that each section of the intestine would have its own pacemaker. This discussion continues today, with theories assuming that gut pacemaking travels through the intestine similar to a cardiac action potential and theories assuming that a network of millions of continuously active pacemaker cells orchestrates pacemaking.

13.1.2

Electrical Activity Generates Gut Smooth Muscle Contraction in a Two-Step Process

Gut smooth muscle contracts on demand; hence, a trigger needs to initiate the contractions. In the heart, this trigger is the autonomous pacemaker action potential from the sinus node; in the gut, the trigger is a two-step process. Gut smooth muscle cells contract when calcium enters the cell cytoplasm, originating from the cell surroundings, potentially fortified by calcium from the sarcoplasmic reticulum. Calcium from the extracellular fluid enters the cell through calcium- and/or voltage-activated calcium channels. For the smooth muscle cells to reach the threshold for opening of the calcium channels, significant depolarization is needed. For gut smooth muscle that is dominated by pacemaker activity (slow waves) such as in the stomach and small intestine: the smooth muscle cells are constantly subjected to rhythmic depolarizations from the interstitial cells of Cajal (ICC) pacemaker cells, but this might not be enough to trigger contraction. Hence, another stimulus such as muscle stretch or neural activity is needed to further depolarize the cell and so bring the slow-wave plateau above the threshold for the generation of smooth muscle action potentials that bring calcium into the cells. The subsequent mechanisms to generate contraction are excellently reviewed by Bitar et al.

13.1.3

The Development of the ICC Pacemaker Hypothesis

In 1889, Cajal published on an interstitial cell type that came to be associated with his name although Cajal thought them to be interstitial neurons. In the following 90 years, the ICC remained a topic of investigation by anatomists, and suggestions on a potential role in gut motility control often surfaced. The first doctoral thesis on ICC was written by Jan Boeke at the University of Utrecht in 1949 who summarized his interpretation as follows: “Interstitial cells of Cajal are the neurohumoral region of the sympathetic nerve endings, the synaptic field, where humoral energy is produced necessary for the transmission of the nervous stimulus.” Nelemans and Nauta concluded in 1949 that “Since most organs containing interstitial cells show rhythmicity…. it seems to us most probable that we have to find the origin of this rhythmicity in the interstitial network” and their studies indicated communication between ICC and the extrinsic autonomic nervous system “The sympathetic and parasympathetic fibers likely make synaptic contact with the interstitial cells.”

In the late 1970s, physiologists started to deduce from their experiments that the slow-wave activity may not be a general property of smooth muscle cells. Edwin Daniel started to talk about hybrid cells that may interact with smooth muscle cells and nerves. Tadao Tomita surmized that the slow waves might originate from cells in between the circular and longitudinal muscle layer. Nelson Durdle scraped the submucosa off the circular muscle of the dog colon and observed the slow waves to disappear, suggesting that circular smooth muscle cells may not have intrinsic slow-wave activity.

An important moment in the history of ICC physiology was the 9th International Symposium on Gastrointestinal Motility held in 1983 in Aix-en-Provence where Lars Thuneberg showed data on methylene blue mediated destruction of ICC-MP (see Section 13.2.1 ) in the intestine and the loss of slow waves as a consequence, providing the first indirect physiological evidence of a pacemaker function of ICC. Thuneberg also showed videos of contracting ICC in culture. For many physiologists present, it was the first time that ICC came on their radar screen, and for some it was the start of incorporating ICC into their physiological research from then on. Thuneberg published his ideas about ICC as pacemaker cells, focusing on the mouse small intestine, in his seminal thesis “Interstitial cells of Cajal: intestinal pacemaker cells?” A similar hypothesis based on electron microscopic observations on the human esophagus was put forward by Faussone Pellegrini who published her original ideas in Italian in 1976, later translated into English.

Dissection experiments provided evidence for a role of ICC in pacemaking in the colon. The original observations from Durdle et al. that scraping the submucosa from the circular muscle layer caused loss of slow waves in the dog colon were confirmed by Smith et al. in 1987. The junction of smooth muscle and submucosa was shown to contain ICC. Conklin and coworkers showed that, also in the cat colon, slow waves depended on an intact interface with the submucosa. Many studies followed to characterize the ICC-SMP as the pacemaker cells of the canine colon. From the ICC-SMP (see Section 13.2.4 ), slow waves propagate actively into the circular muscle layer where they rhythmically depolarize the network of muscle cells. The slow-wave plateau remains at a relatively constant level of depolarization; from the submucosa to the myenteric plexus, the slow-wave amplitude decreases, this is not due to electrotonic decay but to a gradient in membrane potential. When ICC-SMP are removed from the circular muscle layer, the smooth muscle cells at the ICC-SMP border depolarize by about 18 mV. Hence, ICC-SMP not only provide rhythmic depolarization but they also hyperpolarize adjacent smooth muscle cells.

13.1.4

The Definitive Experiments on ICC Pacemaking in the Mouse Small Intestine

Thuneberg’s hypotheses were based on the ICC-MP in the mouse small intestine, and physiologists turned again to this model to provide definitive proof of ICC as pacemaker cells. Maeda et al. described how Kit antibodies eliminated Kit-positive cells from the intestine and affected the rhythmicity of the intestinal contractions. This was followed by studies on mutant mice that have a malfunctioning kit receptor, which causes inability to develop ICC-MP after birth ; it was shown that these mice lost the ability to generate slow-wave activity and slow-wave-driven peristalsis in the proximal small intestine. The small intestine still functions, the musculature generates normal action potentials because the loss of ICC results in marked depolarization of smooth muscle cells allowing action potential generation at the resting membrane potential instead of at the plateau of slow waves. In 1998, the definitive piece of evidence was reported : ICC isolated from the myenteric plexus area of the mouse small intestine, held in short term culture, produced spontaneous rhythmic inward currents with a reversal potential of + 10 mV or + 17 mV whereas smooth muscle cells did not. It was thought that the channel involved was a nonselective cation channel.

13.1.5

Role of ICC in Setting Smooth Muscle Cell Membrane Potential

The circular muscle layer of the stomach and small intestine in dogs, cats, mice, and humans have a resting membrane potential gradient with the myenteric plexus region being up to 23 mV hyperpolarized compared to the submucosal border. The advantage may be that the musculature can generate weak and strong contractions dependent on the region depolarized enough to generate action potentials. An important contributor to the gradient is carbon monoxide, which is generated by ICC-MP. Carbon monoxide is generated by heme oxygenase 2 (HO2) and in HO2-KO mice and ICC-MP depleted tissue, the gradient is abolished.

In the dog and mouse colon, a membrane potential gradient exists but the most hyperpolarized region is adjacent to the submucosa. In the mouse colon, the gradient was shown to depend on carbon monoxide generated by HO2, but the enzyme was found in submucosal neurons and not in ICC-SMP. Hydrogen sulfide also plays a role in chronic hyperpolarization of smooth muscle by blocking nitric oxide synthesis and as such contributes to the maintenance of the membrane potential gradient. The hydrogen sulfide generating enzymes are found in enteric neurons.

13.2.1

ICC-MP (ICC Associated With the Myenteric Plexus)

ICC-MP provide primary pacemaking to the stomach and small intestine and secondary pacemaker activity to the colon ( Fig. 13.1 ). They are found around the myenteric ganglia and in between the ganglia, they form a continuous network of pacemaker cells around the circumference and along the length of the organ. ICC, in general, have features of smooth muscle cells and fibroblasts and their electron microscopic identification and differentiation needs expert evaluation; their identity is in part dependent on their connection to associated cells. Principle pacemaker ICC appear to have a stellate-shaped cell body with a large nucleus and relatively sparse cytoplasm and processes extending in all direction, connected to other ICC-MP by gap junctions. Pacemaker ICC are rich in mitochondria, likely related to the need for ongoing pacemaker activity and rich in caveolae related to calcium handling. ICC have intermediate filaments but no thick filaments and are not contractile. Thuneberg showed dramatic contractile activity in isolated ICC after a few days in culture, but recent studies using calcium imaging do not provide evidence for contractile activity in situ. An example of ICC density: there are about 750 cells per mm ² in ICC-MP in the mouse small intestine or a total of 2.5 million. ICC-MP are Kit positive, the Kit protein is essential for ICC development, and recovery after injury, and has as natural ligand, steel factor. Hence, mutations in both the Kit protein and steel factor (the Sld/Sld mice) lead to absence of ICC-MP and exposure to Kit antibodies will destroy the ICC. Enteric nerves regulate ICC-MP activity.

13.2.2

ICC-DMP (ICC Associated With the Deep Muscular Plexus of the Small Intestine)

ICC-DMP are positioned at the outer edge of the circular muscle but separated from the submucosa by a thin layer of circular muscle cells, hence the identification of an inner and outer circular muscle layer with the ICC-DMP in between. The ICC-DMP are associated with a plexus of nerve varicosities along the axons, but no neuronal cell bodies, hence no ganglia. ICC-DMP are usually bi-polar in shape and run along the neighboring circular muscle cells. They are exceptionally well connected to each other and to the outer layer of smooth muscle cells via gap junctions. ICC-DMP have close contacts with nerve varicosities, containing accumulations of synaptic vesicles, via synapse-like junctions. Hence ICC-DMP are intercalated between nerves and smooth muscle cells and can act as an accessory route for neuromuscular transmission, as originally suggested by Cajal. ICC-DMP are Kit positive but they are completely normal in W/Wv or Sld/Sld-mutant mice, hence in contrast to ICC-MP, ICC-DMP maturation is not dependent on the Kit protein. Immunohistochemistry using Kit or Ano1 antibodies shows ICC-DMP processes to be aligned with the long axis of circular muscle cells. In addition to neurotransmission, ICC-DMP are thought to play a role in secondary pacemaking, that is, stimulus-induced pacemaking, where a neural stimulus or distention may evoke rhythmic depolarizations that will interact with the primary pacemaker ICC-MP to generate specific motor patterns such as segmentation. The intimate association of the ICC-DMP with enteric nerves facilitates neurotransmission and neural stimulus-induced pacemaking.

13.2.3

ICC-IM (Intramuscular ICC)

Intramuscular ICC are found throughout the gastrointestinal tract and form a three-dimensional network within the longitudinal and circular muscle layers. Only in the small intestine of small animals are they scarce or absent. They are sometimes subdivided into ICC-CM and ICC-LM. ICC-IM are connected to ICC-MP and as such facilitate the passage of pacemaker activity to the circular muscle layer ( Fig. 13.3 ). ICC-IM are usually bipolar with their long axes along the long axes of smooth muscle cells. They connect with neighboring smooth muscle cells via many large gap junctions and often show close contacts with nerve terminals via synapse-like junctions.

ICC-IM in the esophagus and stomach are connected to vagal afferents; intramuscular arrays (IMAs), vagal stretch receptors, are all intimately connected to ICC-IM. Mutant mice with diminished ICC-IM also show a marked reduction in IMAs suggesting mutual dependence. Destruction of the nodose gangion results in loss of IMAs in the esophagus as well as loss of ICC-IM. Vagal stimulation of ICC-IM in the stomach may induced pacemaker activity in these cells which then becomes dominant over the primary pacemaker activity of the ICC-MP. ICC-IM are therefore monitoring smooth muscle activity and modifying it through their interactions with enteric and extrinsic nerves.

In larger animal and humans, the circular muscle is divided into lamellae with septa in between that extend from the submucosa to the myenteric plexus area. ICC line the septal borders and therefore run in a different direction compared to regular ICC-IM. The septal ICC can be regarded as part of the ICC-IM network although they are sometimes categorized as a separate subtype.

13.2.4

ICC-SMP (ICC Associated With the Submuscular Plexus of the Colon and Antrum)

Colonic ICC-SMP are observed at the interface between the submucosa and the circular muscle layer. They form a network connected by gap junctions and this network is loosely attached to the smooth muscle cells, that is, some sections are gap junction-coupled to smooth muscle cells and other sections are found disconnected and wander into the submucosa. The ICC-SMP not only connect to each other but also connect to ICC-IM in the septa, so that a continuing network of pacemaker cells is present throughout the colon ( Fig. 13.1 ).

A specialized pacemaker function has been proposed for ICC-SMP in the colon, they are primarily responsible for generating the slow waves, and ICC-MP acting as secondary pacemaker cells providing stimulus-dependent pacemaking. ICC-SMP are connected to enteric nerves and have neurotransmitter receptors such as NK1. Consistent with ICC-MP being stimulus-dependent pacemakers in the colon, ICC-SMP-associated contraction patterns are much less influenced by neural blockade compared to activity orchestrated by ICC-MP, and electrical nerve stimulation in vitro also affects ICC-MP-related activity much more compared to the ICC-SMP-related activity. Inhibitory innervation has little effect on ICC-SMP generated slow waves.

In the rat antrum, ICC-SMP were of very similar structure as the ICC-MP. They may functionally be connected to ICC-IM. ICC-SMP have been described in the human antrum, they have a close association with nerve networks but a specific function has not been elucidated yet.

The nonganglionated submuscular plexus is distinct from the ganglionated submucosal plexus; ICC are also present in association with the submucosal plexus and may be involved in orchestration of contractions by the muscularis mucosa.

13.2.5

ICC-SS (ICC of the Subserosa of the Colon)

ICC-SS were described in 1982 by Thuneberg in the mouse. Komuro studied their structure extensively in the guinea pig colon and found them to be distributed within a thin layer of connective tissue lining the mesothelium, appearing to constitute part of the serosa, but commonly referred to as being part of the subserosa. Several branches extend from ICC-SS cell bodies, diverging into secondary and tertiary processes to form an extensive two-dimensional network. In the guinea pig colon, they are connected by gap junctions to each other and by peg and socket junctions and close appositions to smooth muscle cells. Komuro observed that the ICC-SS connect to ICC-IM in the longitudinal muscle and suggested them to be stretch receptors that might trigger contractions of the longitudinal muscle. Connections with ICC-IM in the longitudinal muscle were also seen using 3D imaging. In the guinea pig colon, few nerve fibers were present in this connective tissue layer and a relationship with ICC was difficult to ascertain. ICC-SS may also help in stretch-induced nutrient absorption in the microcirculation and fluid drainage into lymph vessels similar to the fibroblast network underlying epithelial cells in the mucosa. A study by Vanderwinden on the mouse colon showed that ICC-SS were rare at birth but fully developed into a two-dimensional network 7 days after birth, well connected by gap junctions, aligned with smooth muscle cells and in contact with them through peg and socket junctions, and in close apposition to nerve fibers.

Administration of the Kit antibody imatinib for 4 days beginning 8 days after birth markedly reduced ICC-SS together with ICC-MP and ICC-IM ; ICC recovered within 4 days after withdrawal of the drug. Imatinib did not have this effect 24 days after birth suggesting a specific role of Kit in neonatal development.

Rumessen studied ICC-SS by electron microscopy in the human colon and found them to have myoid features, also rich in caveolae and secretory organelles. They did not find gap junctions and although small nerve bundles appeared abundant in the subserosal layer, close contacts with ICC appeared inconspicuous. ICC-SS appeared connected by peg and socket contacts; connections with smooth muscle cells were close appositions as well as peg and socket structures.

13.3.1