Interstitial Cells of Cajal: Update on Basic and Clinical Science
- First Online:
- Cite this article as:
- Huizinga, J.D. & Chen, J. Curr Gastroenterol Rep (2014) 16: 363. doi:10.1007/s11894-013-0363-z
- 751 Views
The basic science and clinical interest in the networks of interstitial cells of Cajal (ICC) keep growing, and here, research from 2010 to mid-2013 is highlighted. High-resolution gastrointestinal manometry and spatiotemporal mapping are bringing exciting new insights into motor patterns, their function and their myogenic and neurogenic origins, as well as the role of ICC. Critically important knowledge is emerging on the partaking of PDGFRα+ cells in ICC pacemaker networks. Evidence is emerging that ICC and PDGFRα+ cells have unique direct roles in muscle innervation. Chronic constipation is associated with loss and injury to ICC, which is stimulating extensive research into maintenance and repair of ICC after injury. In gastroparesis, high-resolution electrical and mechanical studies are beginning to elucidate the pathophysiological role of ICC and the pacemaker system in this condition. Receptors and ion channels that play a role in ICC function are being discovered and characterized, which paves the way for pharmacological interventions in gut motility disorders through ICC.
KeywordsInterstitial cells of Cajal (ICC)PDGFRα+Enteric nervous system (ENS)Chronic constipationGastroparesisInflammationColon motilityPacemaker cellsNitric oxideGuanylate cyclaseGastroparesisChronic constipationGut transitGastrointestinal transitIon channelsReceptors
interstitial cells of Cajal
ICC associated with the myenteric plexus (also called ICC-MY and ICC-AP)
ICC associated with the deep muscular plexus (small intestine)
ICC associated with the submuscular plexus (colon)
tyrosine-protein kinase Kit or CD117
Enteric nervous system
- PDGFRα+ cells
Platelet-derived growth factor receptor-alpha positive cells (specialized fibroblast-like cells)
Long Distance Contraction (colon)
Rhythmic Propulsive Motor Complex (colon)
High Amplitude Propulsive Contraction (an RPMC identified in human colon with amplitude > 100 mm Hg)
Recent studies on relationships between interstitial cells of Cajal (ICC) injury and gastrointestinal (GI) disease
Injury to ICC
Relationship Between ICC Injury and GI Dysfunction Proposed or Shown / Notes
Decrease in c-Kita in full thickness archival biopsies
ICC-MP in terminal ileum show ultrastructural injury
Communication with mast cells appears to provide recovery
ICC-MP showed no degeneration or cytological changes
ICC-DMP shows ultrastructural injury
Injury selective to ICC
Slow transit constipation
Reduction in Ano1 immuno-reactivity
Not specifically studied
Loss of c-Kita, ultrastructural injury but no loss of cells
Injury related to degree of inflammation, full recovery when inflammation resolved
Idiopathetic and diabetic gastroparesis
Ultrastructural injury to ICC, differences between types
Loss of c-Kita
Delayed gastric emptying
Inverse correlation between ICC count and 4-h gastric retention in DG but not IG
Loss of c-Kita
Loss of c-Kita
High resolution mapping with 256 electrodes shows slow wave irregularities
Ultrastructural injury to ICC, abnormal slow waves in vitro
Loss of c-Kita
C-kit positivity related to extent of dilation
Loss of c-Kita
With and without megacolon
C-Kit was significantly more reduced in patients with megacolon
Loss of c-Kita
Loss of c-Kit in proximal but not distal part
Loss of c-Kita
No significant difference of loss in proximal and distal part
Streptozotocin-induced diabetes in rats
Reduction in c-kit protein levels
Delayed gastric emptying
Curcumin (reducing oxidative stress) improves gastric emptying and c-kit protein levels
Reduction in c-Kita in gastric antrum
Impaired up-regulation of HO-1 expression associated with ICC injury
More than 50 % loss of c-Kita in small intestine
Slow wave mapping with 121 electrodes showed no changes
Diabetic enteropathy in mice
No change in ICC in small intestine and colon (immuno and electron microscopy)
Increased fecal output, increased intestinal transit
Change in enteric nerves likely cause of dysmotility
Diabetic gastroparesis in mice
Loss of c-Kita
Delayed gastric emptying
Gastric emptying and ICC restored by upregulation of heme oxygenase I (reducing oxidative stress)
Structure–Function Relationships of ICC
Control of Colonic Motor Patterns
We have come a long way from inferring gut motor function from agonist-induced contractions on muscle strips or from measuring a “motility index” of in vivo pressure recordings that grouped all motor activities together into a number that said nothing about the type of contractions or motor patterns. In particular, spatiotemporal mapping techniques of video recordings of motor activities in vivo or in whole organs in vitro have given us unprecedented insight into details of motor patterns, far exceeding knowledge gained from classical manometry or in vitro muscle strip studies. Spatiotemporal mapping and high-resolution manometry will dominate searches for physiological and pharmacological control of motility for some time to come.
The key to many controversies related to myogenic versus neurogenic origins of motor patterns is that abolishment of a motor pattern by tetrodotoxin (TTX) does not show exclusive control by the ENS. First, a nonneural stimulus might create the same pattern under different conditions, and second, neurally dependent ICC pacemaker activities might underlie a motor pattern that would be inhibited by TTX. Most motor patterns are orchestrated by an intricate combination of neurogenic and myogenic control systems [38•, 45•].
In studies on muscle strips of the human colon, slow rhythmic contractions at about 1/min were observed [43•, 49], possibly related to cyclic motor complexes recorded in whole human colons in vitro at 0.25/min . Spencer et al. proposed that the pacemaker is intrinsic to the ENS, although a polarized intrinsic neural reflex was not demonstrated. In the study by Carbone et al., the strong rhythmic contractions were shown not to be due to burst firing by motor neurons, but to rely on intrinsic properties of the muscular apparatus. The rhythm was speculated to occur through motor neurons exciting but not driving the pacemaker system, likely originating in ICC-MP [43•].
Although good studies on the human colon using traditional manometry are emerging, such as a recent 24-h colonic manometry study , such studies focus on the most marked colonic activities, such as the so-called high-amplitude (>100 mm Hg) propulsive contractions (HAPCs) that might occur only a few times in 24 h . The study of Singh et al.  had as one of the objectives to differentiate myopathies from neuropathies. Myopathies were deduced from the presence of most motor activities, albeit of low amplitude, and neuropathies from the absence of responses to a meal or the absence of occurrence of HAPCs. Future studies using high-resolution manometry will better define motor patterns, and much more basic research will reveal the myogenic and neurogenic parts of the control mechanism of human colonic motility.
Interactions Between ENS and ICC
Innervation of ICC serves to modulate pacemaker activity, serves to facilitate the sensory function of ICC networks, and serves to mediate nitrergic innervation. ICC are heavily innervated; every ICC is innervated by multiple varicosities of enteric sensory and/or motor nerves via synapse-like structures  (Fig. 2). This is different from smooth muscle cells that are primarily innervated through varicosities in the extracellular space. Although nerve varicosities can be as close as 20 nm to smooth muscle cells, in general they are innervated by nerve varicosities that are further away from the cell membrane, thereby creating the possibility of innervation of multiple cells through single varicosities or a string of varicosities along an axon . When excitatory nerves increase the pacemaker frequency of the stomach , the frequency of peristaltic contractions will increase (if the general muscle excitation surpasses the mechanical threshold). Enteric sensory AH neurons have been shown to communicate with ICC, likely affecting intestinal pacemaker activity .
Because ICC are directly innervated and because they have gap junction contact with smooth muscle cells, the idea that ICC are actual conduits for innervation of smooth muscle cells, consistent with Ramon y Cajal’s original hypothesis , has gained considerable attention. But are ICC “mediators of smooth muscle innervation”? This term is used when the hypothesis is discussed that a smooth muscle function is affected by innervation via ICC, where ICC are deemed to be a conduit without necessarily a specific function other than transmitting the signal from nerve to muscle. For example, using WWv mice that lack intramuscular ICC (ICC-IM), electrical stimulation of nitrergic nerves  did not cause significant muscle relaxation, and stimulation of cholinergic nerves  did not lead to excitatory junction potentials in smooth muscle leading to the conclusion that innervation did not occur via direct communication between nerves and smooth muscle but that ICC was an essential intermediary. Subsequent experiments showed that the Ws rat fundus, also lacking ICC-IM, readily relaxed to the same type of stimulation of nitrergic nerves  or contracted normally via cholinergic neurotransmission , indicating that direct innervation of smooth muscle was the major pathway. Other studies showed mixed results . The discrepancies between these studies was solved elegantly by Friebe and his coworkers, who first showed that all nitrergic inhibition was mediated by cGMP generated by guanylate cyclase . Then they showed that mice that lack guanylate cyclase specifically in smooth muscle cells had normal nitrergic innervation. It was suggested that relaxation might be mediated by ICC or PDFGRα+ cells or that NO might release VIP from enteric nerves . A subsequent study showed that mice that lacked guanylate cyclase in both smooth muscle and ICC did not show any nitrergic innervation [2••] or, more correctly, that an NO donor (Fig. 1) or electrical field stimulation of enteric nerves did not relax smooth muscle cells. Hence, direct innervation to smooth muscle leads to normal relaxation, but, stunningly, when this pathway is blocked through deletion of the smooth muscle guanylate cyclase, the ICC pathway appears to be able to give complete and near normal smooth muscle relaxation. This occurs via NO-induced cGMP in ICC, which then transmits this signal to smooth muscle cells, by cGMP passing through gap junction and/or by hyperpolarizing the smooth muscle cells. There were differences noted in ICC or smooth-muscle-mediated relaxation. NO-induced relaxation was normal in muscle that was precontracted with 10 μM carbachol and also was normal in the presence of nifedipine, whereas the ICC-mediated pathway was absent under these conditions. This needs further investigations. The conclusion is that two or more parallel ways of neurally mediated nitrergic smooth muscle relaxation exist. It is possible that other cells, such as the PDGRF α+ cells, which contain guanylate cyclase cells, are also involved. Klein and coworkers used a c-Kit CreERT2 knock-in allele to target ICC [61•]. This achieved a 50 % reduction in ICC that resulted in loss of slow wave activity and severely reduced transit, confirming the critical role of ICC in normal motor function. In the small intestine, electrical field stimulation to induce neurally mediated inhibitory and excitatory junction potentials failed, but in the colon, completely normal fast and slow inhibitory junction potentials remained. This provides evidence for a role of ICC in neurotransmission, but it also shows that normal inhibitory innervation to the muscle is present in the colon of these mice. Therefore, their conclusion that “ICC . . . are essential for transmission of signals from enteric neurons to GI smooth muscle cells” may be a bit of an overinterpretation in light of their data on the colon and in light of other studies indicating parallel innervation.
There is little doubt that nitrergic innervation plays an important role in normal gut motility, yet when colonic motility was studied in nNOS knock-out mice, all normal motility features were present, although the frequency of propulsive activity and transit were moderately affected [42•]. Even in the study of Groneberg et al. [2••], transit in almost all mice with guanylate cyclase knocked out in both smooth muscle cells and ICC showed total gut transit times in the normal range. This again shows that alternative pathways are present for many gut functions.
What do these new insights into ICC nerve interactions contribute to our understanding of the pathophysiology of GI motor dysfunction? For example, in chronic constipation, it is well established that ICC are diminished. This means that innervation may be compromised, although to what extent is still difficult to assess—first, because of the existence of parallel pathways [2••] and because the direct muscle innervation appears to be more powerful [2••]. Moreover, we do not know whether the ICC pathway and the direct muscle pathway are completely overlapping. It is possible that certain neuronal programs preferentially use one pathway over the other ; this is an interesting topic for further investigations.
ICC and Inflammation
An inflammatory process can cause significant damage to the neuromuscular apparatus, but the gut organs also have a dramatic capability to recover from the most serious injuries. ICC can be very sensitive to an inflammatory process, but the injury can be quite selective to subpopulations of ICC. In an animal model of inflammation using Nippostrongylus Brasiliensis, ICC-DMP were seen to be virtually wiped out at day 30 postinfection, whereas ICC-MP showed little structural and functional injury . In mice infected by Trichinella Spiralis, ICC-DMP showed significant injury at day 10 postinfection but fully recovered at day 30 . In the same model, ultrastructural studies showed that the ICC-MP had undergone significant damage but c-Kit immunohistochemistry did not reveal injury; recovery was almost complete at day 40 postinfection . Human appendicitis showed a relationship between degree of inflammation, ultrastructural injury to ICC, and loss of the c-Kit receptor, although no marked loss of ICC was observed . An interesting phenomenon shown in many studies, including this appendicitis study, is that one of the first injuries to ICC is the retraction of processes and subsequent loss of contact with nerves. Importantly, interval appendicitis (surgery performed weeks after inflammation has receded through antibiotic treatment) showed marked recovery of ICC injury . In Crohn’s disease, ICC in the terminal ileum can be significantly diminished and injured, but there is no linear relationship between years of disease and diminished presence of ICC . In fact, in long-standing Crohn’s disease, ICC networks can be quite normal, and it was hypothesized that mast cells might play a positive role in the maintenance of a healthy ICC population . Also, in the colon of Crohn’s disease, although ICC-MP appeared less dense, they were ultrastructurally normal ; ICC associated with the submuscular plexus (ICC-SMP), however, showed ultrastructural signs of injury, whereas neighboring glial cells and fibroblast-like cells were undisturbed, with mast cells connecting to the fibroblast-like cells . In ulcerative colitis, a recent immunohistochemical study found that all ICC types were reduced (loss of c-Kit) by more than 50 % in the left colon, together with significant neuronal damage . Mast cells were markedly increased in the musculature. In a postinfectious IBS rat model using wild type and cytolethal distending toxin (CDT) negative strains of C. jejuni, it was demonstrated that Campylobacter CDT is an important factor in the development of chronic altered bowel patterns, rectal inflammation, and reduced ICC-DMP .
In conclusion, inflammation induces, most often, ultrastructural changes to ICC and, most often, loss of the c-Kit receptor, but it does not always lead to loss of ICC. Hence, reduction of c-Kit density is not equivalent to loss of ICC; on the other hand, normal c-kit immunohistochemistry might still be associated with ICC structural injury. Importantly, injury to ICC can be restricted to a subset of ICC, and, often, associated cells are undisturbed, except nerves that are usually affected. Gut tissue shows a dramatic capability of recovery to injury. Even with chronic inflammation, the fate of ICC is not doomed; it appears possible that partial recovery can occur, possibly mediated by mast cells. Of great interest are the factors that are responsible for regeneration, and many factors have been identified , including recently anoctamin-1 [68, 69]. Important issues for future studies are the injury to communication between ICC and neighboring cells—in particular, nerves, smooth muscle cells, PDGFRα+ cells, and immune cells—as well as the correlation between structural injury and loss of specific ICC functions. The study of impaired recovery of ICC or under- or overexpression of growth factors may contribute to the role of ICC in chronic neuromuscular disorders, and the study of epigenetics might reveal critical factors involved .
ICC and Pacemaker Dysfunction in Gastroparesis
Gastroparesis is a term used for patients with real or perceived gastric retention accompanied by nausea and/or vomiting and/or bloating. It is no doubt a heterogeneous disorder, but despite that, it appears that the most common intrinsic defect is being recognized in the ICC . Recent studies using full-thickness biopsies have confirmed this, together with the finding of CD45 and CD68 immune cell infiltration in the musculature . Interestingly, ICC counts inversely correlated with 4-h gastric retention in diabetic gastroparesis, but not in idiopathic gastroparesis. There was also a significant correlation between loss of ICC and enteric nerves in diabetic, but not in idiopathic, gastroparesis [16••]. No distinguishing features were found between ultrastructural injuries in both types of gastroparesis . Injury was not observed in ICC-associated PDGFRα+ cells [72•]. There was a similar immune infiltrate in diabetic and idiopathic gastroparesis, but ICC injury was not correlated with this [16••]. Heme oxygenase-1 might be important in ICC maintenance during gastroparesis, as deduced from a study on streptozotocin-induced diabetes in rats , confirming an earlier study . In the same model, curcumin, which has antioxidant and antiinflammatory properties, upregulated stem cell factor and c-Kit and was protective for ICC injury .
High-resolution mapping of electrical activities using arrays of hundreds of extracellular electrodes has given us much insight into the two-dimensional organization of pacemaker activity . This technique has now established that abnormalities in pacemaker activity accompany gastroparesis, such as changes in pacemaker frequency, but also the occurrence of ectopic pacemakers and conduction block [18•]. The pacemaker region in the stomach was shown to be associated with high-velocity and high-amplitude localized circumferential propagation of slow waves. Interestingly, rapid circumferential propagation also emerges during a range of gastric dysrhythmias, elevating extracellular amplitudes and organizing transverse wavefronts, likely disturbing motility. Bidirectional coupling between ICC-MP and circular ICC-IM networks was suggested to be associated with these electrical activity patterns . Also, in vitro, irregular, or loss of slow wave activity was observed in a model of gastroparesis .
In a detailed study on intestinal dysmotility in the RIP-I ⁄ hIFNb transgenic mice after 3.5 months of type 1 diabetes induced with streptozotocin, in contrast to most other models of diabetes, no changes in ileal and colonic ICC were observed, only selective changes in neurotransmission that appeared to underlie increased transit in the small intestine .
Continuing Discovery of Ion Channels and Receptors on ICC as Potential Targets for ICC Modulation
Our knowledge about mechanisms by which neurotransmitters and other regulatory substances affect ICC function is steadily increasing. Ion channels in the ICC membrane are discovered or further elucidated. Receptors are identified, and their agonists and functional roles explored. This accumulating knowledge will lead to our understanding of normal ICC function and how this can be modulated. The ideal would be to be able to selectively affect ICC function as pacemakers or mediators of innervation. But a lot of work is still needed to understand the intracellular workings of the ICC. To that end, the rhythmic ongoing transient calcium oscillations in ICC may play a key role (Movie 3) , and it is not surprising that many regulatory substances can affect intracellular calcium. Both neurotensin  and prostaglandin F2α  increase these calcium oscillations, and subsequently, nonselective cation channels are activated, leading to increased ICC excitation. Also, 5HT enhances spontaneous calcium oscillation in ICC-MP of the mouse small intestine via 5HT3 receptors and might, therefore, affect ICC excitability . Intracellular calcium can also mediate the regulation of excitability by CCK via the CCK-1 receptor, activating the HCN channel in the gastric antrum [78–80] initiating rhythmic inward currents. Pacemaker currents can also be activated by sphingolipids via sphingosine 1-phosphate . Both histamine via H1 receptors  and Substance P can excite ICC via membrane depolarization . The maxi chloride channel in ICC is also activated by intracellular calcium . The maxi chloride channel is one of the candidates for the pacemaker channel in the ICC-MP of the small intestine, together with the chloride channel Ano1 [85, 86]. The maxi chloride channel in ICC has complicated kinetics—in particular, when patch clamping of in situ ICC is attempted and single-channel analysis has been hampered by the difficulty of assessing channel opening and closing transitions. Phase portrait analysis has given us exciting new ways of understanding channel gating in ICC [87, 88].
Many studies have been performed in recent years that bring us closer to our understanding of the physiological role of ICC in controlling GI motility and their role in pathophysiology of motility disorders. This review has highlighted some of these studies dealing with several but not all aspects of gut motor function. Of critical importance are the studies that highlight how ICC collaborate with nerves and other cells in and outside the pacemaker networks to perform functions as pacemakers and mediators of innervation. Detailed knowledge is accumulating on receptors and ion channels in the ICC cell membrane and intracellularly that will bring us in the near future to pharmacological experiments to alter or restore ICC function to affect motor abnormalities.
The study was financially supported by Grant 81170249 from the National Natural Science Foundation of China (NSFC) and by Grant MOP12874 from the Canadian Institutes of Health Research (CIHR). As always, we appreciate the discussions with Dr. Xuan-Yu Wang.
Compliance with Ethics Guidelines
Conflict of Interest
Dr. Huizinga states that this work was possible due to ongoing research support from the Canadian Institutes of Health Research (CIHR). Dr. Chen declares that this work was possible due to ongoing research support from the National Natural Science Foundation of China (NSFC).
Human and Animal Rights and Informed Consent
This article does not contain any studies with human or animal subjects performed by the authors.