SEARCH

SEARCH BY CITATION

Keywords:

  • Kit immunohistochemistry;
  • zebrafish;
  • interstitial cells of Cajal (ICC);
  • gastrointestinal tract motility

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Gastrointestinal (GI) motility results from the coordinated actions of enteric neurons, interstitial cells of Cajal (ICC), and smooth muscle cells. The GI tract of the zebrafish has a cellular anatomy that is essentially similar to humans. Although enteric nerves and smooth muscle cells have been described, it is unknown if ICC are present in the zebrafish. Immunohistochemistry and PCR were used determine expression for the zebrafish Kit orthologue in the zebrafish gastrointestinal tract. Cells displaying Kit-like immunoreactivity were identified in the muscular layers of the adult zebrafish gastrointestinal tract. Two layers of Kit-positive cells were identified, one with multipolar cells located between the longitudinal and circular smooth muscle layers and one with simple bipolar cells located deep in the circular muscle layer. Primers specifically designed to amplify mRNA coding for two zebrafish kit genes, kita and kitb, and two kit ligands, kitla and kitlb, amplified the expected transcript from total RNA isolated from zebrafish GI tissues. The Sparse mutant, a kita null mutant, showed reduced contraction frequency and increased size of the GI tract indicating a functional role for kita. These data establish the presence of a cellular network with Kit-like immunoreactivity in the myenteric plexus region of the zebrafish GI tract, adjacent to enteric neurons. Expression of kita and kitb, and the ligands kitla and kitlb, were verified in the adult GI tract. The anatomical arrangement of the Kit-positive cells strongly suggests that they are ICC. Developmental Dynamics 236:903–911, 2007. © 2007 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Gastrointestinal (GI) motility is primarily mediated by complex interactions between enteric neurons, interstitial cells of Cajal (ICC), and smooth muscle cells. These cell types have been identified and characterized in many vertebrate model systems including the mouse, guinea pig, rat, dog, chick, and human (Faussone-Pellegrini and Thuneberg,1999; Komuro,1999; Sanders et al.,1999; Young,1999). It is now widely understood that ICC play a vital role in regulating GI motility. ICC generate a pacing signal that drives smooth muscle, mediates neuronal input to smooth muscle, and establishes a smooth muscle membrane potential gradient across the thickness of the circular smooth muscle layer (Huizinga et al.,1995; Farrugia et al.,2003; Strege et al.,2003; Ward et al.,2004; Sanders et al.,2006). Disturbances in ICC distribution have been correlated with GI dysmotilities in animal model systems, and in humans (He et al.,2000,2001; Huizinga et al.,2001; Lyford et al.,2002; Sanders et al.,2006). The zebrafish is a well-established model system for studies on development, and is an important model system for human disease (Dodd et al.,2000; Goldsmith,2004). Zebrafish larvae are transparent, which allows direct observation of organ function in the intact organism. The GI tract is functional and is apparently fully formed at 5 days post fertilization (dpf). Muscular contractions can be directly observed in unperturbed intact larvae. Therefore, the functional role played by enteric neurons, ICC, and smooth muscle can be examined within an intact, living network unlike other model organisms where direct observation of GI organogenesis and GI function is not possible.

Establishing a zebrafish-based model system for GI motility requires identification of enteric neurons, ICC, and smooth muscle cells in the GI tract, as well as the characterization of function at developmental time points. Smooth muscle cells are evident at 50 hr post fertilization (hpf) and the longitudinal and circular muscle layers are well established by 96 hpf (Wallace et al.,2005a). Enteric neurons are observed by 72 hpf, completely populate the GI tract by 96 hpf, and increase in number to 120 hpf (Kelsh and Eisen,2000; Shepherd et al.,2004). Therefore, both enteric neurons and smooth muscle cells appear well established by 5 dpf, coinciding with observations of spontaneous GI contractions and feeding (Holmberg et al.,2004). For comparison, ICC in the mouse model begin to develop at embryonic day 15, irregular and spontaneous electrical rhythmicity develops by embryonic day 18, and adult-like slow waves were recorded 9 days after birth (Ward et al.,1997; Beckett et al.,2006) ICC play a vital role in developing coordinated GI contractions that function to mix and propel the luminal contents of the gut (Sanders et al.,1999). However, the presence of ICC in zebrafish fish has not been reported previously.

Identification of ICC in the GI tract may be accomplished using Kit as a selective and specific marker (Huizinga et al.,1995; Burns et al.,1997; Kluppel et al.,1998; Faussone-Pellegrini and Thuneberg,1999; Komuro,1999; He et al.,2000,2001; Lyford et al.,2002) The proto-oncogene c-kit is expressed by ICC located within the tunica muscularis of the GI tract of mice, guinea pigs, rats, dogs, and humans. Antibodies specific for this protein label ICC within GI tissues in several species, including humans, and have been widely utilized to both determine ICC cellular morphology, and to characterize sub-populations of ICC located in distinct layers of the muscular wall of the GI tract (Ward and Sanders,1992; Ward et al.,1994; Burns et al.,1997; Ozaki et al.,2004; Komuro,1999). Several classes of ICC have been identified. Each class is distinguished according to cellular morphology, anatomical location, and function (Faussone-Pellegrini and Thuneberg,1999; Komuro,1999; Sanders et al.,1999). The myenteric plexus region, located between the circular and longitudinal muscle layers, exhibits the highest ICC density and the myenteric ICC network is continuous throughout the mammalian GI tract (Burns et al.,1997; Hirst and Edwards,2004). A second network of ICC, termed deep muscular plexus ICC, is located between the thin innermost and thicker outer layer of circular muscle of the mammalian small intestine (Burns et al.,1997; Komuro,1999). Deep muscular plexus ICC are typically bipolar and are oriented in parallel with circular smooth muscle, and receive excitatory and inhibitory synaptic input from the enteric nervous system (Wang et al.,1999; Wang et al.,2003; Iino et al.,2004).

Two orthologs of the mammalian c-Kit receptor tyrosine kinase have been identified in the zebrafish, kita and kitb (Parichy et al.,1999; Mellgren and Johnson,2005). The zebrafish Sparse mutant is characterized by a deficit in stripe melanocytes resulting from a null allele of kita (Parichy et al.,1999). Zebrafish kita plays an essential role for the migration and survival of embryonic melanocytes, but the role and/or expression of kita and kitb in the GI tract have not been reported. Because c-Kit signaling is essential for the normal development of ICC and rhythmic activity in the mouse GI tract, identification of the kita and kitb protein in the zebrafish GI tract as well as functional motility differences in the Sparse mutant would support the presence of zebrafish ICC (Ward et al.,1994; Huizinga et al.,1995; Kluppel et al.,1998). The objective of this study was to determine the presence or absence of Kit-like immunoreactivity within the muscular outer layers of the zebrafish GI tract, consistent with the presence of ICC.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Kit expression in the zebrafish GI tract was identified using a rabbit polyclonal antibody specifically designed to recognize C terminal amino acids 961–976 of human c-Kit. A continuous and extensive network of cells displaying Kit-like immunoreactivity was observed within the muscular layers of paraformaldehyde-fixed adult zebrafish GI tract (Fig. 1). Two distinct populations of cells were observed in separate layers of whole mount tissue. One layer displayed elongated cell bodies with multiple branching processes forming a loose but regular network pattern (Fig. 1B), and a second layer was comprised of bipolar or simple bifurcating cells located deeper within the circular muscle layer (Fig. 1C). Images shown are taken from mid-intestinal segments, as classified by Wallace et al. (2005a). Substantial differences in distribution of Kit-positive cells in each segment were not observed.

thumbnail image

Figure 1. Anti-Kit rabbit polyclonal antibody identifies putative ICC networks in the adult zebrafish GI tract. Full thickness confocal stack reconstruction of an adult zebrafish GI whole mount preparation immunostained using rabbit anti-Kit polyclonal antibody (A). Total thickness of reconstructed stack ≈20 μm. Reconstruction of a partial thickness stack (≈8-μm-thick) shows a kit-positive network in the myenteric plexus region with stellate-shaped, highly branching cells (B). A second distinct network of slender bipolar cells oriented parallel to the circular smooth muscle cells was located closer to the submucosal border and is shown in a partial stack reconstruction of a 6-μm-thick region (C). Kit-like immunoreactivity was observed in the GI tract of zebrafish larvae beginning at 7 dpf. Single confocal sections approximately 0.6-μm-thick of mid-saggital sections from the posterior end of the larvae GI tract show branching cells with slender processes in the tunica muscularis (D, arrows). Background staining of the mucosal layer prevented full-thickness stack reconstruction. The density of Kit-positive cells was increased at 11 dpf (E), and a network of Kit-positive cells was observed at 20 dpf (F). Scale bars = 20 μm.

Download figure to PowerPoint

Spontaneous contractions of the zebrafish GI tract begin at 4 dpf, coinciding with the development of enteric neurons, but the possibility for ICC contributing to the development of rhythmic contractions has not been explored (Kelsh and Eisen,2000; Shepherd et al.,2004; Wallace et al.,2005b). Kit expression was identified in the GI tract of paraformaldehyde-fixed zebrafish larvae. The GI tract was carefully dissected from larvae prior to immunostaining to maximize antibody penetration of intact tissues. Single confocal sections taken midway through the digestive tube are shown for 7, 11, and 20 dpf larvae (Fig. 1D–F, respectively). Cells displaying Kit-like immunoreactivity were observed in the outer layer of the tunica muscularis, indicated by arrowheads. Kit expression appeared to increase from 7 to 11 dpf, and it was possible to identify an apparent network of Kit-positive cells at 20 dpf. Kit-like immunoreactivity was not observed prior to 7 dpf (data not shown). It was not possible to separate the mucosa from the tunica muscularis in adult or larvae GI tissues, which contributed to high background staining and prevented complete confocal stack reconstructions of the full-thickness digestive tract in larvae. The ACK2 rat monoclonal antibody that has been widely used to identify mouse ICC also specifically identified cellular networks in acetone-fixed adult and larvae zebrafish GI tissues (data not shown).

Two types of Kit-positive cells are observed at high magnification: one branching cell with prominent nuclei (black arrow) and thinner bipolar cells (Fig. 2A). The anatomical position of these cells was determined in transverse sections of adult zebrafish GI tissues. Hematoxilyn and eosin stained transverse sections show that the outer longitudinal and inner circular smooth muscle layers are approximately 2 and 4–5 cell layers thick, respectively (Fig. 2B). Kit-like immunoreactivity was consistently observed in two distinct layers of cells in transverse sections of adult GI tissues (Fig. 2C). The outer layer of cells with Kit-like immunoreactivity appeared more dense and continuous, and a second thinner layer of Kit-positive cells was observed closer to the lumen. The inner layer was discontinuous and oriented in parallel with the circular smooth muscle cells. Both layers of Kit-positive cells were observed in anterior, mid, and posterior intestinal segments. The anatomical position of Kit-positive cells was also determined using transverse sections of zebrafish larvae. A single continuous layer of Kit-positive cells within the tunica muscularis of the GI tract was observed in 13 dpf larvae (Fig. 2D).

thumbnail image

Figure 2. Cells with Kit-like immunoreactivity form a dense network in the myenteric plexus region, and a second network close to the circular smooth muscle-submucosal border. High magnification of Kit-positive cells reveals individual stellate-shaped cells (white) in a full thickness stack (A). Circular and longitudinal muscle layers of the adult GI tract are clearly observed in H&E stained transverse sections (B). Kit-positive cells are found in the myenteric plexus region, and a thin layer of Kit-positive cells is located near the circular smooth muscle-submucosal border (C). A single layer of Kit-positive cells was observed in the tunica muscularis of transverse sections of zebrafish larvae (white arrow) (D). Scale bars = 20 μm.

Download figure to PowerPoint

The relative anatomical positions of Kit-positive cells, enteric neurons, and smooth muscle cells in the zebrafish GI tract was examined further for comparison with established mammalian model systems. A pan-neuronal antibody, anti HuC/D, was used to identify cell bodies, and anti-acetylated α-tubulin antibody was used to identify neural processes. Composite images resulting from dual labeling experiments with whole mounted tissue show that cells displaying Kit-like immunoreactivity are in the same area as neuronal cell processes, labeled with acetylated α-tubulin antibody (Fig. 3A). Composite images of transverse sections show Kit-positive cells and neural cell processes are located in the myenteric plexus region (Fig. 3B). Similarly, dual labeling with anti Hu C/D and anti-Kit antibody shows neural cell bodies located near Kit-positive cells in whole mounted tissues, and in the myenteric plexus region in transverse sections (Fig. 3C and D). Enteric nerves were only occasionally observed to extend deep into the circular muscle layer near the thin inner layer of Kit-positive cells (data not shown). Dual labeling with anti-SM22 antibody to detect smooth muscle cells and anti-c-Kit antibody identified longitudinal and circular smooth muscle layers, and Kit-positive cells located between the layers in the whole mounted tissue (Fig. 3E). Smooth muscle cells were identified using an anti-Desmin antibody in transverse sections. Desmin is an intermediate cytoskeleton filament protein expressed by GI smooth muscles. Composite images of transverse sections dual labeled with anti-Desmin and anti-Kit antibody show one dense layer of Kit-positive cells positioned between the longitudinal and circular muscle layers and a second thin layer of Kit-positive cells near the innermost circular smooth muscle cells (arrowhead, Fig. 3F).

thumbnail image

Figure 3. Dual staining shows the relative position of Kit-positive cells and enteric neurons or smooth muscles. Cell morphology is shown in whole mounted adult tissues (left), and the position and orientation of each layer is shown in transverse sections (right). Composite images using anti-Kit and the anti-tubulin antibodies show Kit-positive cells (A, red) near neuronal cell processes (green, A). Composite images using the pan-neuronal antibody anti-Hu (green) show cells with Kit-like immunoreactivity (red) near neural cell bodies (C). Transverse sections show 2 separate layers of kit-positive cells. Kit-positive cells and neural processes (B) or neural cell bodies (D) were observed in the myenteric plexus region. Dual staining with anti-Kit and anti-SM22 or anti-Desmin antibodies show the relative positions of Kit-positive and smooth muscles. Dual staining in whole mount tissues shows outer longitudinal and inner circular smooth muscles (green, E) and cells with Kit-like immunoreactivity (red). Composite images of transverse sections clearly show the layers of longitudinal and smooth muscle separated by Kit-positive cells, and a second layer of kit-positive cells near the innermost border of circular smooth muscle. Sale bars = 20 μm.

Download figure to PowerPoint

Expression of the zebrafish orthologs of mammalian c-Kit was verified by determining mRNA expression for the kita and kitb genes within GI tissues. Reverse transcriptase PCR was performed on cDNA prepared from total RNA isolated from adult zebrafish GI tissues. The presence of mRNA for two known orthologs of the c-Kit receptor, kita and kitb, as well as orthologues of c-Kit ligand, kitla and kitlb, were determined using specific primer sets for each gene, and for β-actin as a positive control (see Table 1). Primers were intron-spanning to rule out the possibility of genomic contamination. Products of the expected size for kita, kitb, kitla, and kitlb were amplified (Fig. 4). The identity of the bands was confirmed by sequencing. These data show for the first time the presence of mRNA encoding kita, kitb, and the ligands for these receptors, kitla and kitlb, in the zebrafish GI tract.

Table 1. Oligonucleotide Primer Sequences
TargetSequenceExpected size (bp)
kita (AF 153446)F: GTT ATC CCA CTC CTC AGA TCA AGT564
 R: TCA CAG CTA CAG TCATCA CAG TGT 
kitb (DQ072166)F: GGG AGG AAT CAC CAT CAG AA234
 R: CTC AGG TGG AAA TCGTGG TT 
kital (AY929068)F: CACAGTTGCTGCCTATTCCA580
 R: GGTGAGGAGCCACCTGAGAT 
Kitbl (AY929069)F: GGC TGC ATT TGA ACC TGT ATC C542
 R: GTG TCT GCA CAC CCTAAA GAA TCC 
β actin (BC067566)F: GAT ACG GAT CCA GAC ATC AGG GTG TCA TGG TTG GTA580
 R: GAT ACA AGC TTA TAGCAG AGC TTC TCC TTG ATG 
thumbnail image

Figure 4. RT-PCR amplifies the zebrafish Kit receptor tyrosine kinase kita and kitb, and Steel factor, kitla and kitlb, from isolated adult GI tract. Amplification products are shown for primer sets designed for each gene (Table 1), and for β-actin. Products of correct size for each primer set were observed and confirmed by sequencing.

Download figure to PowerPoint

Development of ICC requires functional c-Kit signaling during embryogenesis in the mouse model, and also to maintain ICC in adults (Maeda et al.,1992; Torihashi et al.,1995; Beckett et al.,2006). We examined homozygous Sparse mutants spab5 (ZDB-FISH-980202-47) for functional differences in GI motility and for differences in appearance of the GI tract to determine a role for the kita gene. Contraction frequency was reduced in Sparse mutants compared to wild-type 7 dpf larvae. Contraction frequency averaged 0.59 ± 0.05 (mean ± standard error, n = 20) contractions per minute in wild-type larvae, and 0.33 ± 0.03 (mean ± standard error, n = 23, P ≤ 0.05) contractions per minute in Sparse mutants (Fig. 5). Inactivation of c-Kit by injection of the neutralizing antibody ACK2 in the mouse resulted in distension of the stomach, small intestine, and colon, and, therefore, the size of the GI tract in Sparse mutants was compared to wild0type larvae (Maeda et al.,1992; Torihashi et al.,1999). Sparse and wild-type larvae were incubated for 1 hr in media containing FITC-labeled dextran, anesthetized, and washed. The lumen of the GI tract was observed using fluorescence microscopy. The cross-sectional area of the lumen was outlined from fluorescent images and measured using Image Pro Plus software (Media Cybernetics, version 5.0). The size of the GI tract in 7 dpf Sparse mutants was larger compared to 7 dpf wild-type larvae, 12,331 ± 467 μm2 and 14,985 ± 483 μm2, respectively (mean ± standard error, n = 20 wild-type, 12 Sparse, P ≤ 0.05). Adult Sparse zebrafish consistently exhibited a distended GI tract, which was most apparent in the intestinal bulb (data not shown). Cells displaying Kit-like immunoreactivity were observed in Sparse mutant larvae and in adult GI tissues, with a similar density as wild-type zebrafish (data not shown).

thumbnail image

Figure 5. The pigment pattern mutation Sparse, a kita null mutant, has decreased spontaneous GI contractions and an increased size of the GI tract. Optically transparent zebrafish larvae (7 dpf) were videotaped and contraction frequency was 0.59 ± 0.05 contractions/minute (mean ± SE, n = 20) in wild type, and 0.33 ± 0.03 contractions/minute (P < 0.05, n = 23) in homozygous Sparse mutant larvae. Cross-sectional area of the GI tract was measured in single frames for each experiment. Cross-sectional area was less for wild-type, 12,331 ± 467, compared to the Sparse mutant larvae, 14,985 ± 483 (P < 0.05).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Results presented here show that anti-Kit antibodies identify a network of cells located between the circular and longitudinal muscle layers of the adult zebrafish GI tract and a separate, second group of cells in a region analogous to the boundary between the inner and outer circular muscle in the mouse. Expression of mRNA encoding for the two zebrafish orthologues of mammalian Kit, kita, and kitb, as well as two orthologues for mammalian Kit ligand, kitla and kitlb, was confirmed in adult zebrafish GI tissues. The Kit-positive cells were distributed in the myenteric plexus region near enteric neurons. A reduction in contraction frequency and an increase in size of the GI tract were observed in 7 dpf Sparse mutant larvae, which exhibit a null kita allele. These data are consistent with expression of zebrafish orthologs to the mammalian c-Kit receptor tyrosine kinase, kita and kitb, in the zebrafish GI tract. Furthermore, the data strongly indicate the presence of ICC in the zebrafish GI tract, which are distributed similarly to mammalian ICC, thus establishing the zebrafish as a suitable model system for mammalian GI motility.

Expression of kita and kitb was not detected previously in zebrafish larvae at 4 dpf using in situ hybridization techniques (Parichy et al.,1999; Mellgren and Johnson,2005). Our data confirm this result. Kit-like immunoreactivity was not detected in 5-dpf larvae, and was first detected at 7 dpf. It is likely that expression of kita and kitb occurs developmentally later then 4 dpf, or that mRNA levels in the 4-dpf GI tract occur at very low and not detectable levels by in situ hybridization techniques. Functional data show that spontaneous contractions of the GI tract that develop near 4 dpf are poorly organized (Holmberg et al.,2004). Coordinated contractions, which are better organized and contribute to propulsive motility movements, develop gradually between 5 and 7 dpf, consistent with the time course of appearance of Kit immunoreactivity (personal observation). The role of ICC on intestinal transit has been well characterized in the mouse model and mutant mice lacking ICC exhibit spontaneous but poorly organized contractions and delayed intestinal transit (Torihashi et al.,1995,1997; Ward et al.,1995). The poorly organized contractions in early zebrafish development are, therefore, similar to the ICC-deficient mutant mouse model. The morphological data from these studies are consistent with the presence of myenteric ICC and deep muscular plexus ICC, similar to the mouse small intestine. However, no obvious differences were observed in the distribution of Kit-positive cells in anterior, mid, or posterior regions of the zebrafish GI tract, whereas in the mouse model deep muscular plexus ICC distribution differs regionally along the GI tract. The functional role of Kit-positive cells in the zebrafish GI tract, as well as the relationship between Kit-positive cells and enteric neurons, remains to be explored.

The physiological role of kita and kitb in the zebrafish GI tract is unknown. Parichy et al. did not observe motility defects in kita null mutants (Sparse), and homozygous adults appeared normal (Parichy et al.,1999). However, motility patterns are complex, and quantification of spontaneous contractions in zebrafish larvae is notoriously difficult. Further, erratic spontaneous contractions of GI smooth muscles do not require ICC in the mouse model. Data reported here suggest that early spontaneous rhythmic contractions observed in zebrafish larvae before 7 dpf also do not require ICC, because Kit-positive cells were not observed prior to 7 dpf. Therefore, overt GI motility deficits are not predicted for the Sparse mutant during early development and would not be observed. A role for kita in ICC development is indicated by data from Sparse mutants showing a reduced frequency of spontaneous GI contractions in 7-dpf larvae. Further, the increased surface area of the GI tract in Sparse mutants is highly similar to observations of an expanded small intestine in the W/Wv mutant mouse (Maeda et al.,1992; Torihashi et al.,1995). Complete c-Kit inactivation mutants or the absence of Kit ligand is lethal for the mouse model but homozygous Sparse mutants survive to adulthood. This raises the possibility for a different physiological function for mammalian Kit and zebrafish kita. The recent discovery of a second Kit orthologue in the zebrafish, kitb, provides the possibility that kitb may be sufficient to support partial or complete ICC development and survival in the zebrafish GI tract (Mellgren and Johnson,2005). Survival of homozygous kita null mutants suggests that kitb at least partially supports ICC development. Kit-like immunoreactivity was also observed in Sparse larvae and adult GI tissues, indicating that kitb contributes to ICC development and maintenance. Elucidating the specific roles for the zebrafish receptor tyrosine kinase kita and kitb during ICC development in the zebrafish, and examining whether kita and kitb are co-orthologous for human c-Kit will contribute to the understanding of ICC development during pathophysiological conditions that contribute to GI dysmotility in the clinical setting.

ICC are the primary cell type that expresses c-Kit in the GI tract but mast cells are also labeled by anti-Kit antibody. The two cell types are easily distinguished because mast cells are round and granular in appearance but typical ICC display long, slender processes and an oval body (Faussone-Pellegrini and Thuneberg,1999; Komuro,1999). In the present study of the zebrafish GI tract, the cellular morphology of cells displaying Kit-like immunoreactivity showed multiple branching processes, or were slender and bipolar, and round cells were never observed. Definitive confirmation of these cells as ICC may require electron microscopy to examine their ultra structure to determine if they posses features associated with mammalian ICC, such as a high density of mitochondria and intermediate filaments, an absence of thick filaments, gap junction contacts with smooth muscle cells, and surface caveolae (Komuro,1999). Characterization of the association between enteric neurons and ICC will contribute to determining the functional role and classification of ICC in the zebrafish.

The GI tract of the zebrafish lacks a stomach, and presents as a folded tube in the adult. The anterior segment, or intestinal bulb, exhibits the largest lumen and may function as a reservoir and mixing chamber, similar to a stomach. The anterior, mid, and posterior portions roughly approximate the small and large intestine with respect to absorptive function (Wallace et al.,2005a). Direct observation of muscular contractions in intact larvae show that contractions in the anterior segment are more frequent and more random when compared to mid and posterior segments where less frequent but more ordered anterograde propagations appear. It will be interesting to determine the role for enteric neurons and ICC in these distinctly different motility patterns. Although the development of contraction frequency has been reported previously, functional studies, such as assessing transit through the GI tract, are lacking (Holmberg et al.,2004).

In summary, the results in this study show Kit-like immunoreactivity in the tunica muscularis of the zebrafish GI tract and provide functional evidence for the role of kita in GI motility. Kit-positive cells in the zebrafish appear to correspond to myenteric ICC and deep muscular plexus ICC, based on cellular morphology and distribution. Expression of mRNA encoding kita, kitb, kitla, and kitlb was demonstrated in the zebrafish GI tract using PCR analysis. Mutants lacking functional kita showed a decreased contraction frequency compared to wild-type zebrafish. Taken together, these data are consistent with the presence of ICC in the zebrafish GI tract, and suggest that the zebrafish may be a suitable model system for human GI physiology. More work will be required to determine if Kit-positive cells in the zebrafish GI tract initiate spontaneous, rhythmic contractions, transmit electrical activity along the long axis of the GI tract, and coordinate enteric neurotransmission similar to mouse ICC.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Aquaculture

Wild-type and Sparse mutant zebrafish (ZFIN ID 960809-7 and 980202-7) were obtained from the Zebrafish international resource center and maintained according to standard guidelines in accordance with IACUC guidelines (Westerfield,1993). Wild-type long-finned golds (Scientific Hatcheries, Huntington Beach, CA) were also used for some preliminary experiments and no differences in immunohistochemical staining were observed when comparing strains (data not shown). Fish were maintained at 28°C in system water comprised of deionized water containing 240 mg/L Instant Ocean salts and 75 mg/L NaHCO3 with 20% system water change each day (pH was adjusted to ≈ 7.2, conductivity ≈450 PPM). Zebrafish were fed 3 times daily, alternating Cyclopeeze (Argent, Redmond, WA) with live brine shrimp, and maintained on a 14-hr/10-hr light/dark cycle. Crosses were performed in the morning, and embryos were maintained in embryo medium in 400-ml beakers kept in a water bath set to 28°C. Larvae were fed hatchfry encapsulation, grade 0, beginning at 7 dpf, and live brine shrimp after 11 dpf (Argent, Redmond, WA).

Immunohistochemistry

Adult and larvae zebrafish were anesthetized in system water containing MS222 (3-aminobenzioc acid ethyl ester, Sigma Chemical Co., St Louis, MO) and sacrificed for immunohistochemistry. Intact larvae and freshly dissected adult GI tissues were fixed in freshly prepared 4% paraformaldehyde in phosphate buffered saline (Fisher) with pH adjusted between 7.3 and 7.4 for a minimum of 2 hr, but not longer than overnight. Immunostaining using the ACK2 antibody requires acetone fixation, and for these experiments larvae and tissues were fixed for 15 min in ice-cold acetone. Fixed tissues were washed 4 times in PBS containing 0.02% sodium azide. Larvae were carefully dissected such that the GI tract was separated from the body so that the head remained attached to the GI tract. Although time-consuming, removing the GI tract in this manner allowed direct access for antibody-antigen binding, producing consistent immunostaining. Nonspecific binding of primary antibody to tissues was minimized by incubation for at least 1 hr at 4°C in blocking solution comprised of 10% normal donkey serum (NDS, Chemicon) and phosphate buffered saline containing 0.02% sodium azide, 0.1% Triton-X-100, and 0.05% -Tween 20 (PBS-TT). Primary antibodies were diluted in PBS-TT containing 5% normal donkey serum and were applied for 24–48 hr at 4°C on an orbital platform. After washing 4 times in PBS-TT, tissues were incubated with appropriate secondary antibody conjugated to a fluorescent marker and diluted in PBS-TT containing 2.5% normal donkey serum for 24 hr at 4°C on an orbital platform. Nonspecific immunoreactivity was assessed by immunostaining tissues or larvae in an identical manner but with the primary antibody omitted. The optimal concentration for each primary and secondary antibody was determined using serial dilutions. Antibodies were applied simultaneously during dual labeling experiments, followed by a PBS-TT wash and the simultaneous application of two appropriate secondary antibodies. Tissues were washed with PBS-TT and mounted on glass slides using Slow Fade medium (Invitrogen).

Paraffin-embedded sectioned tissues.

Adult fish were anesthetized in MS 222, decapitated posterior to the gills, tails were removed to aid fixative penetration and were immersed in freshly prepared 4% paraformaldehyde overnight at 4°C on an orbital platform. The GI tract was subsequently dissected and washed with PBS followed by 70% ethanol. Tissues were dehydrated and paraffin embedded using a Tissue-Tek VIP E150 processor with 1-hr infiltration steps at 70% (1 step), 95% (2 steps), 100% (2 steps) ethanol, followed by xylene (3 steps) and paraffin infiltration (30 min, 4 steps) comprising a 12-hr schedule. GI tissues were sectioned into anterior, mid, and posterior portions and embedded in a paraffin block oriented for transverse sections. One piece of paraformaldehyde-fixed mouse GI tract (small intestine) was included in each block to serve as a positive control for immunostaining. Paraffin blocks were stored at 4°C until sectioning. Specimen blocks were trimmed, soaked in ice water for 30 min, and sectioned at either 4 or 8 μm using a Reichart-Jung microtome. Sections were placed on slides (Superfrost plus, Fischer), baked at 56°C for 1 hr, and stored at −20°C until de-paraffinization.

Slides with sectioned tissues were placed on a 56°C warming tray for 5 min, and then a xylene bath for de-paraffinization (2 steps, 5 min). Tissue sections for hematoxylin and eosin staining were rehydrated using a graded alcohol series; 100% (2 steps, 3 min), 95% (2 steps, 3 min), followed by dIH2O (1 step, 10 min). Tissue sections used for immunohistochemistry were similarly rehydrated followed by PBS-TT (1 step, 10 min). Preliminary immunostaining experiments using anti-Kit antibody resulted in a weak signal, and therefore antigen retrieval techniques were utilized to enhance specific immunoreactivity. Slides were immersed in antigen retrieval solution (Abcam) at 100°C for 20 min., allowed to cool to room temperature in the same solution, and rinsed in PBS-TT. Immunostaining was performed at room temperature. Tissue sections were incubated in blocking solution (1 hr), primary antibody solution (1 hr), and secondary antibody solution (1 hr). Optimal dilutions were determined for each primary and secondary antibody. Wash steps using PBS-TT were performed after the primary and secondary antibody incubations. Sections were cover slipped after application of Slowfade (Invitrogen).

Fluorescence and transmitted light imaging.

Tissues were examined with conventional light and fluorescence microscopy using an Olympus BX51 microscope equipped with an Optiscan z-axis controller (Prior Scientific) and supported on an anti-vibration table (Technical Manufacturing Corporation, Peabody, MA). Images were captured using a Spot RT digital camera (Diagnostic Instruments) or a Retiga EXi digital camera (QImaging) using Image Pro Plus software version 5.0 (Media Cybernetics). High-resolution images were collected using an Opti grid structured light imaging system (QIOPTIC, Rochester, NY). A 100-W mercury lamp was used for epi-fluorescence illumination with appropriate excitation-emission filter sets for each fluorophore. A laser-scanning confocal microscope (model LSM 510, Zeiss) was used for images in Figure 1. Images were reconstructed from confocal stacks of z-series scans as indicated.

Functional GI motility measurements.

GI motility was assayed using live larvae that were incubated in embryo medium containing blue food dye to enhance contrast of the GI tract lumen. The dye did not effect larvae survival, heart rate, or GI tract contraction frequency. Larvae were fully anesthetized and mounted laterally in 1.2% agar to permit an optimal viewing of the GI tract, and to prevent drift during filming. A drop of anesthetic was placed on the agar to keep it from drying out and to keep larvae anesthetized during filming. Spontaneous GI contractions were recorded continuously for 10 min using a Cannon Optura Xi digital video camera and converted to digital format (Pinnacle Studio AV/DV). Contractions were counted manually at a single position in the mid-intestine during replay of digitized video at an increased rate.

Reverse Transcriptase PCR.

Experiments were done to determine if the zebrafish Kit receptors kita and kitb, and the Kit ligand (Steel factor) kitla and kitlb, are expressed in GI tissues. Total RNA was prepared from freshly dissected GI tissues of adult wild-type zebrafish, and from whole zebrafish (used as a positive control), using the RNeasy kit (Qiagen, Chatsworth, CA). First-strand synthesis was performed using random decamer primers. Gene-specific PCR was performed using Taq Pfx (Invitrogen) with 1 μl of the reaction mixture from the first-strand synthesis and primers specifically designed for each gene, and run for 35 cycles. Optimal magnesium concentration and primer annealing temperature were independently determined for each primer set. Amplification products were resolved on a 2.5% agarose gel.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The authors thank Dr. David Brannigan and Mary Georger for expert technical assistance with tissue preparation and histological procedures, Kyle Leonard for H&E staining, and Jodi Davis for zebrafish care.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES