|Home | About | Journals | Submit | Contact Us | Français|
The interstitial cells of Cajal (ICC) are mesenchymal derived "pacemaker cells" of the gastrointestinal (GI) tract that generate spontaneous slow waves required for peristalsis and mediate neuronal input from the enteric nervous system1. Different subtypes of ICC form distinct networks in the muscularis of the GI tract 2,3. Loss or injury to these networks is associated with a number of motility disorders4. ICC cells express the KIT receptor tyrosine kinase on the plasma membrane and KIT immunostaining has been used for the past 15 years to label the ICC network5,6. Importantly, normal KIT activity is required for ICC development5,6. Neoplastic transformation of ICC cells results in gastrointestinal stromal tumor (GIST), that frequently harbor gain-of-function KIT mutations7,8. We recently showed that ETV1 is a lineage-specific survival factor expressed in the ICC/GIST lineage and is a master transcriptional regulator required for both normal ICC network formation and for of GIST tumorigenesis9. We further demonstrate that it cooperates with activating KIT mutations in tumorigenesis. Here, we describe methods for visualization of ICC networks in mice, largely based on previously published protocols10,11. More recently, the chloride channel anoctamin 1 (ANO1) has also been characterized as a specific membrane marker of ICC11,12. Because of their plasma membrane localization, immunofluorescence of both proteins can be used to visualize the ICC networks. Here, we describe visualization of the ICC networks by fixed-frozen cyrosections and whole mount preparations.
Here, we use the large intestine as an example. H&E stain of a sucrose protected fixed frozen section used for immunofluorescence shows some artifacts compared to a reference paraffin section (Figure 2), where the yellow line demarks the myenteric plexus between the circular and longitudinal muscle layers and the green line demarks the border between the mucosa and muscularlis. Examining the cryosection, from the serosa to mucosa, you first encounter the thinner longitudinal muscle (LM) layer followed by the thicker circular muscle (CM) layer. The section should be in parallel with the longitudinal muscle so the nuclei of LM should be somewhat elongated while the nuclei of the CM should be small and round. Between the LM and CM are myenteric plexi made of neurons that stain with Pgp9.5. Pgp9.5 also stains neuronal processes throughout the CM. The myenteric ICC network (ICC-MY) surround the myenteric plexus and are made of multipolar cells. Within the longitudinal muscle layer, there are rare intramuscular ICCs (ICC-IMs) which are bipolar cells running in parallel with the muscle. Within the circular muscle layer, there are abundant ICC-IMs which are also bipolar cells that run parallel with the circular muscle and are cross-sectioned with this cut. At the junction of the muscularis and submucosa, the submucosal ICC network consisting of ICC-SMPs is a flat network of multipolar ICCs. Examining the whole mount Kit immunostaining of the large intestine, one should expect to encounter the ICC-MY network first, and then the ICC-IM of the CM, and ICC-SMP networks, as one focuses from the coverslip (Figure. 3). Volumetric measure of ICC networks using 3D reconstruction of whole-mount images have been described10. Using wide-field microscope, there is significant out-of-plane fluorescence in the raw images which can be removed after deconvolution (Figure 3). Incomplete stripping of mucosa will result in high background fluorescence. ANO1 is another marker of ICC’s and co-immunostaining of KIT and ANO1 shows complete overlap of ICC staining (Figure 4).
Figure 1. Representative dissected mouse gastrointestinal tract, reprinted with permission13.
Figure 2. A) Representative H&E staining of the mouse large intestine of cryosections prepared as described and reference paraffin embedded section showing some shrinkage artifacts of cryosections. The yellow line demarks the myenteric plexus and the blue line separates the mucosa from the muscularis. B) Representative Kit (red, ACK-2 ) and Pgp9.5 (green) double immunofluorescence with DAPI (blue) counterstaining of the mouse large intestine. LM: longitudinal muscle CM: circular muscle. Scale bar 20 μm.
Figure 3. Representative Kit (red, ACK-2) whole-mount immunofluorescence of the same field at three focal planes, the ICC-MY plane at the LM/CM border, the ICC-IM plane in the CM, and the ICC-SMP plane at the CM/submucosa border. Raw images and deconvoluted images are shown. Scale bar 20 μm.
Figure 4. Representative double-immunofluorescence of Ano1 (green) and Kit (red, ACK-2) in large intestine. Scale bar 20 μm.
Interstitial cells of Cajal were initially characterized by Santiago Ramóón y Cajal exactly one century ago using methylene blue and silver chromate stains of gastrointestinal muscularis. Cajal initially thought ICC’s were neurons based on their processes that are reminiscent of axons and dendrites. Over many ensuing years, study of ICC biology has been limited by lack of specific markers until the discovery of that KIT not only is expressed in ICC, but is also required for their development6. Since then, KIT immunofluorescence has been widely used in the study of ICC biology and has also led to the appreciation of ICC in other contractile organs such as the urinary bladder. Recently, ANO1 has been identified as a second reliable ICC marker.
There have been numerous publications over the past 15 years using immunofluorescence to identify ICC using various fixation and mounting techniques. In our hands "fixed-frozen" cryosections that have been prefixed with paraformaldehyde and sucrose protected works best compared to acetone or paraformaldehyde post-fixation after cryosectioning. For whole mounts, we have found that acetone fixation followed by mucosal dissection gives the most robust results.
Historically, the ACK-2 has been the Kit antibody of choice for mouse ICC identification. It is a rat monoclonal that recognizes the extracellular domain and it is also a blocking antibody that causes ileus when given to live mice. However, the ACK-2 epitope is slowly destroyed by paraformaldehyde fixation and is not rescued by standard antigen retrieval methods. We have found that the ACK-4 epitope is more resistant to paraformaldehyde fixation. Compared to fixed frozen blocks and sections, paraffin embedded blocks and sections preserves morphology better and is more amenable to long term archiving (Figure 4). Neither ACK-2 nor ACK-4 have been successfully used in paraffin sections. We have characterized that D13A2, a new rabbit monoclonal antibody, works exceptionally well in both fixed-frozen and paraffin sections of the GI tract.
Kit is expressed in several other cell types, including melanocytes, hematopoietic cells, germ cells, and in particular mast cells which are also found in the GI tract. Luckily, most mast cells are found is the mucosa and not in the muscularis. Double staining with rabbit anti-Ano1 and rat anti-Kit can eliminate non-specific staining of each antibody (Figure 4). It is also worth noting that the relative expression of Kit and Ano1 between subclasses of ICC seems different. For example, in the small intestine, Kit staining is much more intense in myenteric ICC compared to ICC of the deep muscle plexus whereas Ano1 staining is comparable.
No conflicts of interest declared.
This work was supported by the National Cancer Institute (K08CA140946, to Y.C.), (5F32CA130372, to P.C.), (R01CA102774 and R01HL055748 to P.B.), the Department of Defense (PC094302 to Y.C.), the Shuman Family Fund for GIST Research (to P.C.), and the Starr Cancer Consortium (to P.C., Y.C., C.L.S. and P.B.). We would like to thank Katia Manova, Ning Fan, and Mesurh Turkekul of the MSKCC Molecular Cytology core facility for help with cryosectioning and immunostaining.