Introduction

Interstitial cells of Cajal (ICC) are considered the pacemaker cells in gastrointestinal (GI) muscles and mediate input from enteric motor nerves to the smooth muscle cells/layers. The ability to visualize ICC has arisen from the discovery that ICC express ckit [1]. Immunohistochemical techniques using cKit antibodies have expanded our knowledge about ICC structure, networks and the interactions with intestinal smooth muscle and the enteric nervous system (ENS).

Coordinated gastrointestinal motility is dependent on the integrated activity of both the ENS and the ICC networks. Our studies have confirmed the reports indicating that clinical dysmotility is related to the disruption of ICC network integrity in the gut wall in addition to ENS abnormalities [2–7]. As such, we are amongst a growing consensus proposing that knowledge of ICC pathology may be necessary in surgical decision making and patient prognosis [8–10].

To address this, we have derived a simple and rapid immunohistochemical technique applicable to clinical pathological use, to clearly identify the ICC, in addition to the ganglion cells, in resected human tissue at the time of surgery.

Materials and methods

At the time of pathological assessment of fresh resected intestine from patients less than 18 years of age, colonic tissue samples were obtained for immunohistochemistry.

Fixation

Frozen sections of unfixed tissue were cut on a cryostat at a thickness of 14 μm using tissue oriented in each of two planes (axial and cross section, see Fig. 1) and thaw-mounted onto chrome-alum-gelatin-coated slides. The sections were widely separated on the slide. Sections were air-dried and then treated with the following fixation protocols (Table 1): no fixation, formaldehyde/picric acid for durations of 5, 10, 20, and 30 min, cold acetone for 2 min, at −20°C or for 10 min, at −20°C, or formalin for 30 min, or cold ethanol for durations of 30 s, 2, 5, 10, and 15 min at 4°C and finally cold methanol/acetone (5 min, 4°C). Sections were ringed with a wide band of clear nail polish to form a hydrophobic boundary.

Fig. 1
figure 1

Diagrammatic representation of tissue sections grouped onto glass slide as axial (a) and cross-sectional (c) orientation

Full size image
Table 1 Fixation protocols
Full size table

Immunohistochemistry

Irrespective of the method of fixation, the following immunohistochemical protocol was applied: Rinse the tissue sections in PBS for 5 min and do not allow to dry from this point onward. Incubate in PBS containing 0.3% Triton X (PBST) mouse anti-cKit (Dakocytomation, 1:100) to section I and rabbit anti-NF 68 (Dakocytomation, 1:100) to section II and incubate with vibration for 1 h at room temperature in a humid atmosphere. Rinse in PBS and apply fluorescent-labeled sheep anti-mouse Cy3 to section I and goat anti-rabbit Cy3 to section II [1:400 in PBST containing Hoechst’s (1:10,000)] for 10 min with vibration [11]. Coverslip with anti-fade and observe or store at −40°C.

A similar result can be obtained without vibration using 3-h incubation in primary antibody and 15 min in secondary antibody. Double immunolabeling can be carried out in a single section using sheep anti-mouse Alexa Fluor 488 and goat anti-rabbit Alexa 594 (1:400) as secondary antibodies

Other immunohistochemical methods and markers which were evaluated and were not found useful for the purposes of this protocol are listed in Table 2.

Table 2 Neuronal markers
Full size table

Results were examined semiquantitatively using fluorescence microscopy. The resulting images were compared and photographed. The exposure, gain, and offset were held constant across experimental conditions. Optimal times of incubation in primary and secondary antibodies were studied, and we selected the most rapid condition that still gave a robust result.

Results

We found that the most reliable protocol providing optimal fast immunohistochemistry was the following (heretofore referred to as the Bettolli Protocol):

  1. 1.

    Frozen sections of unfixed tissue were cut at a thickness of 14 μm on a cryostat using tissue oriented in two planes (axial and cross section) as represented in the illustration, Fig. 1, and in the photomicrograph, Fig. 2.

  2. 2.

    These sections were thaw-mounted onto chrome-alum-gelatin-coated slides.

  3. 3.

    These were air-dried and then fixed 10 min in ice-cold absolute ethanol.

  4. 4.

    The sections were grouped and widely separated on the slide and circled with a wide band of clear nail polish (to form a hydrophobic boundary).

  5. 5.

    The sections were rinsed in PBS for 5 min and incubated in PBS containing 0.3% Triton X (PBST) mouse anti-cKit (Dakocytomation, 1:100) to section I and rabbit anti-NF 68 (Dakocytomation, 1:100) to section II. Incubation included vibration for 1 h at room temperature in a humid atmosphere.

  6. 6.

    Finally, the sections were rinsed in PBS and fluorescent-labeled sheep anti-mouse Cy3 to section I and goat anti-rabbit Cy3 to section II (1:400 in PBST containing Hoechst’s, 1:10,000) were applied for 10 min with vibration [11]. A similar result can be obtained without vibration using 3-h incubation in primary antibody and 15 min in secondary antibody, Fig. 3.

  7. 7.

    Coverslip with anti-fade and store at −40°C.

Although double immunolabeling can be carried out in a single section using sheep anti-mouse Alexa Fluor 488 and goat anti-rabbit Alexa 594 (1:400), it failed to provide adequate clarity and intensity of immunolabeling.

Fig. 2
figure 2

Photomicrograph of sections from normal pediatric colon showing the application of the Bettolli protocol. a and b are cross sections of the bowel. c and d are axial sections of the bowel. a and c sections were treated for cKit immunochemistry and show extensive labeling of ICC networks around the myenteric plexus (mp) and in the smooth muscle layers. b and d show neurofilament NF 68 immunostaining. Note the extensive labeling of ganglion cells and nerve fibers in the circular muscle (cm)

Full size image
Fig. 3
figure 3

Photomicrograph of sections from pediatric transverse colon comparing staining with the Bettolli protocol (a–d) with standard immunohistochemistry (e and f). The results are from the 1 h with vibration (a and b) and 3 h ‘no vibration’ Bettoli protocol where similar intensity and staining clarity were obtained within the longitudinal (lm) and circular (cm) muscle layers and at the myenteric plexus (mp). These were comparable to the results for sections treated with the standard 24-h immunohistochemical method for neuronal cells and ICC

Full size image

cKit worked well under all rapid fixation conditions, even at 15 min. Much of the work in optimizing the procedure involved identifying the most robust neural marker. Fixation protocols were eliminated when they decreased staining intensity or led to diffusion of marker protein within the tissue.

Using the Bettolli protocol which was designed for combined ICC and neural staining, the resultant staining was comparable with overnight standard immunohistochemical methodology for these cell types, Fig 3.

Employment of the Bettolli protocol in assessment of colonic HD clearly showed changes in the neuronal and ICC networks, Fig. 4.

Fig. 4
figure 4

Photomicrographs of sections from colonic Hirschsprung’s disease showing the application of the Bettolli protocol (1 h with vibration). a (cKit) and b (NF 68) are micrographs of sections from the proximal ganglionated colon. c (cKit) and d (NF 68) are sections from the distal aganglionic colon. The myenteric (mp) ganglia in the proximal colon (b) are absent in the diseased colon (d). The ICC network normally present around the myenteric plexus (mp) seen in (a) is reduced and disorganized in the aganglionic colon (c). cKit immunostaining in proximal ganglionic and distal aganglionic muscle layers are both reduced when compared to normal colon (see Fig. 3)

Full size image

Discussion

At present the management of all GI disorders is non-specific due to an incomplete knowledge of the control of enteric muscular function. Most of the GI disorders are associated with symptomatic smooth muscle dysfunction. There is now considerable data to show that in addition to the ENS, the ICC networks of the GI wall are also necessary for normal intestinal motility [12–14].

We and others have observed the correlation between ICC abnormality and intestine disease in the bowel [2–10].

Together these data point to the practical significance of determining the disruption of ICC networks in GI disorders where intestinal dysmotility cannot be explained satisfactorily using current pathology staining techniques. Such techniques focus universally upon general H&E assessment of enteric cellular organization and morphology. Some centers also employ cholinesterase or NADPH-dependant diaphorase staining in combination with H&E. However, these approaches do not allow for fully visualizing neural and ICC networks.

On this basis we set out to develop a protocol for fast, reliable characterization of neural and ICC networks in resected bowel at the time of surgery, which would allow the researcher and pathologist to undertake a more complete and more accurate histological assessment of the resected tissue. This would ultimately provide the clinician with more complete pathophysiological data for diagnostic and prognostic purposes.

The original studies of Cajal describing the existence of a specialized cell type in the tunica muscularis used silver impregnation (Golgi’s method) and methylene blue (Ehrlich’s method) [15, 16]. Taxi, however, using light and electron microscopy, distinguished ICC from neurons, macrophages, and Schwann cells [17]. Lecoin et al. showed that ICC came from the gut mesenchyme and were not derived from neural crest cells [18].

More recently, the characterization of ICCs has been possible following the identification of kit expression in ICC, which could be mapped by immunohistochemistry using anti-cKit [1, 19, 20]. This led to identification of a subgroup of ICC, which were found to be pacemaker cells [21–25]. Immunohistochemistry using ckit antibody also stains mast cells. However, minimal experience is required to differentiate the stained ICC from mast cells. Mast cells are round and show a varied and often diffuse pattern of localization compared to the distinct stellate-shaped ICCs with long thin processes, which are arranged in three-dimensional networks at different levels of the gut wall. A major ICC network (ICC-AP) is located against the primary neural network of the gut wall; the myenteric plexus and ICCs also occur closely apposed to the intrinsic innervation of the muscularis consistent with the proposed role for ICC as a mediator of specific neural inhibitory control of the smooth muscle. The localization and distribution of ICC and enteric neural networks appears to be critical for normal gut function. Therefore, pathological assessment of the gut wall would benefit from identification and examination of these networks simultaneously.

For the past 20 years, the characterization of the enteric nervous system and understanding of neural circuits underlying gut functions has benefited from the development of sensitive staining techniques and antibodies for neurochemical and neural cells [26]. For our purposes, we examined a variety of proven and specific neural anti-bodies for their usefulness when applied in conjunction with anti-cKit. Our results show that we have successfully developed a protocol for parallel ICC and neurite immunohistochemistry for the clinical pathology laboratory. The protocol incorporates cKit immunohistochemistry together with a specific neurofilament antibody NF 68, which allowed simultaneous assessment of neuronal cells and ICC in adjacent tissue sections. The individual cell types were easily identified and the distribution of stained cells and processes could be readily compared between healthy and pathological tissues. Our methodology fulfills all necessary pathological criteria. It is accurate, dependable, universally applicable, affordable, and can be applied before completion of the surgery. We found that cutting tissue sections in both axial and cross-sectional orientation allowed better visualization of the ICC networks and their localization relative to the neural and muscular layers. This is of particular importance when evaluating the ICC in longitudinal muscle, as these are difficult to see in cross section.

In conclusion, we propose that standard H&E stain used in combination with our rapid neuronal and ICC immunohistochemistry protocol enables a fast, comprehensive, and accurate assessment of the pathophysiology of signaling networks controlling gut motility while the patient is still in the operating room. This protocol does not require specialized laboratory equipment or expertise.

We propose that the addition of this simple and rapid immunohistochemical assessment in the pathologist evaluation of surgical specimens would result in a more complete characterization for diagnosis and prognosis of the pediatric patient.