Kit signaling is essential for development and maintenance of interstitial cells of Cajal and electrical rhythmicity in the embryonic gastrointestinal tract

Authors


Abstract

Interstitial cells of Cajal (ICC) are specialized cells in smooth muscle organs that generate and propagate pacemaker activity, receive inputs from motor neurons, and serve as mechanosensors. In the gastrointestinal tract, development and maintenance of the ICC phenotype have been linked to intracellular signaling via Kit, but its role in development of ICC during embryogenesis is controversial. Here we have studied the development of functional ICC-MY during the late gestational period in mice. Blocking Kit with a neutralizing antibody before and after development of spontaneous electrical activity (E17 to P0) caused loss of ICC-MY networks and pacemaker activity. ICC-MY and pacemaker activity developed normally in W/+ and WV/+ heterozygotes, but failed to develop between E17 to P0 in W/WV embryos with compromised Kit function. Muscles treated with Kit neutralizing antibody or the tyrosine kinase inhibitor, imatinib mesylate (STI571), from E17-P0 for 3 days caused loss of functionally developed ICC-MY networks, but ICC-MY and pacemaker activity recovered within 9 days after discontinuing treatment with neutralizing antibody or imatinib mesylate. These data suggest that Kit signaling is an important factor in lineage decision and in the development of functional ICC in late gestation. ICC-MY demonstrate significant plasticity in gastrointestinal tissues. Manipulation of the ICC phenotype might provide useful therapies in gastrointestinal disease where the Kit-positive cell population is either lost or amplified. Developmental Dynamics, 2006. © 2006 Wiley-Liss, Inc.

INTRODUCTION

Slow wave pacemaker activity, which times the phasic contractions of the tunica muscularis in the gastrointestinal (GI) tract, is generated by a specialized population of cells known as interstitial cells of Cajal (ICC) (see Langton et al.,1989; Ward et al.,1994; Huizinga et al.,1995). In the small intestine, slow waves are initiated in an ICC network (ICC-MY) that lies within the region also occupied by the myenteric plexus between the circular and longitudinal muscle layers (Suzuki et al.,1986; Hara et al.,1986; Ward et al.,1994). Slow waves conduct into the adjacent, electrically coupled muscle layers (Kito et al.,2005), causing depolarization of smooth muscle cells and increasing the open probability of Ca2+ channels. Ca2+ entry activates the contractile apparatus and generates segmental and peristaltic contractions. A second population of ICC (ICC-DMP) lies in the deep muscular plexus near the submucosal aspect of the circular muscle layer (Rumessen et al.,1992; Torihashi et al.,1993; Zhou and Komuro,1992). ICC-DMP form close, synaptic contacts with nerve terminals of enteric motor neurons and are crucial for reception of motor neural inputs (Wang et al.,2003a, b; Iino et al.,2004, Ward et al.,2006).

ICC express the gene product of c-kit (Ward et al.,1994; Huizinga et al.,1995), a proto-oncogene that encodes the receptor tyrosine kinase, Kit (Yarden et al.,1987; Chabot et al.,1988). Blocking Kit function with neutralizing antibodies impairs the development of functional ICC (Maeda et al.,1992; Torihashi et al.,1995), and animals carrying non-lethal mutations in c-kit (Ward et al.,1994; Huizinga et al.,1995) or its natural ligand, stem cell factor (SCF; Ward et al.,1995), have defects in ICC networks, pacemaker activity, and enteric motor neurotransmission.

Mice express Kit protein by embryonic day E12.5, and functional pacemaker ICC-MY develop by E18, as shown by the occurrence of slow waves in embryonic muscles (Torihashi et al.,1997). We hypothesized that a lineage decision was made near E15 in which Kit-positive cells continued to express Kit and became functional ICC or lost Kit expression and became longitudinal smooth muscle cells (Torihashi et al.,1997), and this idea was supported by a study of Kluppel and coworkers (1998). We suggested that signaling via Kit was essential for the development of functional ICC from Kit-positive precursors. However, this concept was challenged by experiments on mice in which lacZ was knocked-in to the Kit locus, resulting in lacZ expression in ICC-like cells but loss of Kit function (Bernex et al.,1996). LacZ-positive cells were found in embryonic WlacZ/WlacZ mice in the anatomical positions normally occupied by ICC. A second study using Wbanded (Wbd) mutant mice also suggested that Kit expression is only necessary for the postnatal development and proliferation of ICCs but not for the initial cell lineage decision toward an ICC fate during embryogenesis (Kluppel et al.,1998). In homozygous animals with dual mutant alleles, Kit expression was abolished and ICC-like cells were present at birth, causing these authors to conclude that Kit signaling was not required for pre-natal development of ICC. Neither study included assays of ICC function, however, so it is unclear whether the cells identified by lacZ or methylene blue were physiological ICC or precursors that failed to develop into functional ICC. Here, using organotypic cultures, we investigated the role of Kit in the pre- and post-natal development of ICC by determining when Kit signaling becomes important for (1) lineage decision, (2) development of ICC networks, and (3) the onset of electrical rhythmicity. We have also studied the plasticity of ICC networks during development by investigating responses to and recovery from blocking Kit signaling with a Kit-neutralizing antibody (ACK2; Nishikawa et al.,1991) and the tyrosine kinase inhibitor, imatinib mesylate. We also examined W/WV mutant embryonic jejunal muscles for expression of ICC networks and electrical rhythmicity. These data suggest that Kit signaling is crucial from at least E15 onward for the development of functional ICC networks.

RESULTS

Development of ICC Networks and Electrical Rhythmicity in the Jejunum

Kit immunohistochemistry has been used to label ICC in the gastrointestinal tract (e.g., Torihashi et al.,1995). We studied the appearance of ICC networks using Kit immunohistochemistry and correlated the development of ICC with the appearance of electrical slow wave activity using intracellular microelectrode recordings from jejunal muscles. By E17, Kit-immunopositive ICC had developed into a network within the myenteric region in the jejunum (ICC-MY; Fig. 1A). ICC-MY continued to develop as a distinct network through P0 (Fig. 1C,E,G). As previously reported, ICC-DMP develop largely after birth (Torihashi et al.,1997).

Figure 1.

Development of Kit-like immunoreactivity and electrical rhythmicity in the murine intestine. A,C,E: ICC-MY, using Kit immunofluorescence, in embryonic mice from E17–E19. G: ICC-MY in a P0 mouse. ICC-MY develop prior to birth and display an essentially mature appearance from E18. B,D,F,H: Associated with the development of ICC-MY is the onset of pacemaker activity. At E17, small irregular fluctuations in membrane potential (unitary potentials) were observed (B). At E18 slow waves were apparent (D). Slow waves increased in amplitude from E18 to birth (D,F,H), such that at P0 slow waves were about 20 mV in amplitude. Scale bar in G applies to A,C,E.

Associated with the development of ICC-MY was the development of electrical pacemaking activity (Fig. 1B,D,F,H). The rhythmic depolarizations were identified as slow waves by their resistance to nifedipine (1 μM; see Ward et al.,1994; Huizinga et al.,1995; Fig. 1D,F,H). At E17, the resting membrane potential (RMP) of jejunal circular muscle cells averaged −64.3 ± 0.7 mV (n = 6). At this point of development, most cells displayed irregular fluctuations in membrane potential, but no regular slow wave activity (Fig. 1B). At E18, RMP (most negative potentials between slow waves) averaged −61.3 ± 1.5 mV and irregular slow waves with an average amplitude of 10.2 ± 1.3 mV occurred at a frequency of 18.9 ± 2.1 cycles min−1 (n = 9; Fig. 1D). The amplitudes of slow waves increased in amplitude and frequency through the pre-natal period, such that by birth (P0), slow waves averaged 16.5 ± 1.5 mV in amplitude and occurred at a frequency of 19.4 ± 1.0 cycles min−1 (n = 18; Fig. 1H). Slow waves also continue to develop in amplitude and frequency after birth, such that adult levels of activity were noted by P10 (see Ward et al.,1994, 2006).

Dependence of ICC Networks and Electrical Rhythmicity on Kit Signaling

The dependence of the development of ICC-MY networks and electrical rhythmicity on the Kit signaling pathway prior to and after birth was examined using the Kit neutralizing antibody ACK2. Jejunal muscles were removed from E17–P0 animals and placed in organotypic culture with or without ACK2 (5 μg/ml; Fig. 2). Control cultured muscles (with or without non-immune serum) continued to display ICC networks and developed slow wave activity that was similar to the activity recorded from muscle strips removed from animals at equivalent time points. For example, E17 muscles contained a well-developed network of ICC (Fig. 2A) but were electrically quiescent at the time of the initiation of cultures (Fig. 2B). These muscles maintained the ICC-MY network and slow waves developed that averaged 30.7 ± 1.6 mV in amplitude and occurred at a frequency of 11.0 ± 1.7 cycles min−1 (n = 9) after 3 days in culture (Fig. 2C,D). Jejunal tissues removed from E18 animals also developed normal electrical rhythmicity in organotypic culture such that after 3 days in culture slow waves with an average amplitude of 24 ± 3.6 mV and frequency of 16 ± 0.9 cycles min−1 were observed (Fig. 2H,J,L). ICC networks and slow waves were absent from muscles cultured from E17 for 3 days in the presence of ACK2 (Fig. 2E,F).

Figure 2.

Embryonic development of ICC-MY and pacemaker activity is dependent upon Kit signaling. A,B: ICC-MY network and the lack of pacemaker activity in an E17 jejunum. C,D: ICC-MY network and electrical rhythmicity after an E17 muscle was cultured for 3 days. E.F: Disruption of the ICC-MY network and absence of pacemaker activity after an E17 muscle was cultured for 3 days in the presence of ACK2 (5 μg/ml). Inhibition of Kit signaling also disrupted ICC-MY and pacemaker activity in embryonic tissues after pacemaker activity developed. GL: ICC-MY and pacemaker activity in an E18 jejunum (G,H). I,J: ICC-MY network and electrical rhythmicity after an E18 muscle was cultured for 3 days. K,L: Disruption of the ICC-MY network and absence of pacemaker activity after an E18 muscle was cultured for 3 days in the presence of ACK2 (5 μg/ml). Scale bar in K applies to all immunohistochemical micrographs.

Spontaneous slow waves were recorded from muscles taken from E18 embryos and well-developed ICC networks were present (Fig. 2G,H). ICC were maintained and slow wave activity was further developed in muscles cultured for 3 days (Fig. 2I,J). However, ICC networks and slow waves were absent in muscles cultured from E18 for 3 days in the presence of ACK2 (Fig. 2K,L). Culturing P0 muscles in the presence of ACK2 for 3 days also resulted in loss of ICC-MY and electrical rhythmicity. Jejunal tissues obtained from 5 P0 neonates had a mean RMP of −65.2 ± 1.8 mV and displayed spontaneous slow waves that occurred at a frequency of 19.9 cycles min−1 and had an average amplitude of 12.1 ± 1.4 mV (Fig. 3A,B). Control P0 muscles that were cultured in the absence of ACK2 had well-developed ICC-MY and displayed rhythmic slow waves with an average amplitude of 29.2 ± 3.6 mV and frequency of 12.5 cycles min−1. The average RMP of control cultures was −68.6 ± 2.0 mV (Fig. 3C,D). After 3 days in culture in the presence of ACK2, ICC networks were severely disrupted, the average RMP was unchanged (−66.5 ± 1.9 mV), and the muscles were electrically quiescent (Fig. 3E,F).

Figure 3.

Neonatal development of ICC-MY and pacemaker activity is dependent upon Kit signaling. A,B: ICC-MY in a P0 neonate animal. At birth, ICC-MY were similar to adult tissues. However, slow waves were not fully developed to adult amplitudes or frequencies. C,D: ICC-MY network and electrical rhythmicity after a P0 muscle was cultured for 3 days. ICC-MY were maintained in cultured neonatal muscles and the amplitude of slow waves increased during 3 days in culture. E,F: Disruption of the ICC-MY and absence of pacemaker activity after a P0 muscle was cultured for 3 days in the presence of ACK2 (5 μg/ml). Scale bar in E applies to all immunohistochemical micrographs.

The importance of Kit signaling for the development of ICC-MY and the generation of electrical rhythmicity prior to birth was further examined using the white spotting mutant mouse W/WV. Adult W/WV mutant animals have a severely disrupted ICC-MY network and lack slow wave activity (Ward et al.,1994; Huizinga et al.,1995). Previous studies of intestinal muscles of W/WV were focused on adult muscles, and it is not known whether the defects in ICC-MY and rhythmicity are manifest during embryogenesis. We crossed WB/ReJ-W/+ (W/+) females) and C57BL/6J-WV/+ (WV/+) males and pregnant mothers and offspring were sacrificed at E18. All offspring were genotyped and jejunal tissues were collected for Kit-immunohistochemistry, electrophysiology, and organotypic cultures. Genotyping showed that an average of 25% of the offspring were W/WV while others were W/+ and WV/+ and +/+. For example, one WB/ReJ-W/+ mother produced nine embryonic offspring (F1-1 to F1-9). The first set of primers (Kit-W1 and Kit-W1r; see Table 1 for details) amplified a 437-bp product from a BALB/c control and all nine F1 embryos. The 437-bp product corresponds to a region spanning between exon 10 and exon 13. Four siblings (F1-4, F1-5, F1-6, and F1-8) showed a 203-bp product along with a faint 437-bp product, confirming they contained the W product (deletion of exon 11 and exon 12 [234 bp] in the W locus). The second set of primers (Kit-WV1 and Kit-WV1r) amplified a 239-bp fragment from samples obtained from the nine F1 embryos and the BALB/c control (Fig. 4). Nsi I digested the 239-bp products from four of the embryos (F1-1, F1-4, F1-5, and F1-9) into 152- and 87-bp fragments, revealing that these four embryos contained the WV point mutation. Nsi I failed to digest the 240-bp products obtained from the other five embryos or the BALB/c control (Fig. 4) Thus, the PCR and Nsi I digestion analyses revealed F1-2, F1-3, and F1-7 were +/+; F1-6, and F1-8 were W/+; F1-1 and F1-9 were WV/+; F1-4, and F1-5 were W/WV. Two additional sets of nested primers (1) Kit-W2 and Kit-W2r and (2) Kit-WV2 and Kit-WV2r (see Table 1 for details) were used to confirm the genotyping results (data not shown). The 437-, 203-, and 240-bp PCR products were confirmed by sequencing (Fig. 4D–I).

Table 1. Oligonucleotide Primers Used in This Studya
No.PrimersSequence (5′–3′)TmLocation (bp)
  • a

    All primers used for genotyping were designed and synthesized based on the cDNA sequence of the mouse Kit oncogene (c-Kit) (accession number: NM_021099).

1Kit-W1GAGTTCCATAGACTCCAGCGTCTTC591462–1486
2Kit-W1rCAACTGTCATGGCAGCATCCGAC591864–1886
3Kit-W2CAACGGCACGGTGGAGTGTAAG591492–1513
4Kit-W2rCCATATGCAGTGGCCTCAACGAC591832–1854
5Kit-WV1CAGAAAGAGAGGCCCTAATGTCGG591920–1943
6Kit-WV1rGCTCCGTTGAGTGCAGAAGGTTC592137–2159
7Kit-WV2CTGAAGGTCCTGAGCTACCTGG591946–1967
8Kit-WV2rCTTCTGCCTGCTCTTCTTGCTTTGAG602098–2123
Figure 4.

Size analysis of PCR products from cDNAs of a litter of nine E18 F1 offspring produced by mating a WB/ReJ-W/+ (female) and C57BL/6J-WV/+ (male). cDNAs for PCR templates were prepared from total RNA isolated from jejunum muscularis of the nine embryos. PCR products were generated through the use of two sets of primers. A: Kit-W1 and Kit-W1r. B: Kit-WV1 and Kit-WV1r. Kit-W1 and Kit-W1r amplified a 437-bp and/or a 203-bp fragment corresponding to a region spanning between exon 10 and exon 13, within which exon 11 and exon 12 (234 bp) were deleted in the W locus. Kit-WV1 and Kit-WV1r amplified a 240-bp fragment corresponding to a region spanning exon 14 and exon 15, in which a point mutation (C-T) occurred in the WV locus. This mutation introduces an Nsi I site. C: All the 240-bp products were digested with Nsi I, which generated a 153-bp fragment and an 87-bp fragment. The PCR products and the digested products were electrophoresed on 1.5% agarose gels. Brain cDNA from BALB/c was also used for positive control and NTC was a non-template control. A 100-bp marker indicates the size of the fragments. The genotypes of the nine embryonic mice from this litter generated using this procedure are indicated below C. DI: The sequence analysis of c-kit cDNAs of the + (D,E,H), W (F,G), and WV (I). The 437-, 203-, and 240-bp PCR products shown in A–C were gel-extracted and subcloned into the TA cloning vector pcDNA3.1. The plasmids were isolated and sequenced. A partial DNA sequence of the 437-bp PCR products between exon 10 and exon 11 is shown (D). Half of the 437-bp products sequenced contained a 12-bp insertion between exon 10 and exon 11 shown in the box in E. The 203-bp products showed a deletion of exon 11 and exon 12 (F). Half of the 203-bp products also contained a 12-bp insertion after exon 10 (G). A partial sequence of the 240-bp products from the + on exon 14 is shown (H). The 240-bp product from the WV contained a point mutation (C to T) indicated by arrowhead, which creates an Nsi I site (I).

Intracellular electrical recordings and immunohistochemical analysis were performed on jejunal muscles isolated from the F1-1 to F1-9 offspring. These experiments showed that +/+ embryos (F1-2, F1-3, and F1-7), W/+ embryos (F1-6, and F1-8) and WV/+ embryos (F1-1, and F1-9) genotypes displayed ICC-MY and electrical slow waves were recorded from these muscles (Fig. 5A–F). ICC-MY and electrical slow waves were absent in W/WV muscles at E18 (F1-4; F1-5; see Fig. 5G,H). To ensure that the development of ICC-MY and electrical rhythmicity was not just delayed in W/WV tissues, we cultured jejunal muscles from F1-1 to F1-9 E18 embryos for 3 days. After culturing, ICC-MY and slow waves were maintained in, W/+, and WV/+ muscles but absent in W/WV muscles (Fig. 6).

Figure 5.

Kit immunohistochemistry and intracellular recordings from genotyped E18 jejunal muscles (same muscles shown in Fig. 4). A,B: ICC-MY and electrical activity from a +/+ control. C,D: ICC-MY and electrical activity from the jejunum of a W/+ embryo. E,F: ICC-MY and slow wave activity from a WV/+ embryo. G,H: The absence of ICC-MY and pacemaker activity in a W/WV embryo. The few immunopositive cells in G are ICC-DMP that persist in W/WV animals, as previously reported (Torihashi et al.,1997). Scale bar in G applies to A,C,E.

Figure 6.

Organ culture of muscles from W/WV mutant embryos did not restore ICC-MY or pacemaker activity. ICC-MY and electrical recordings are shown for offspring of a WB/ReJ-W/+ (female) and C57BL/6J-WV/+ (male) mating after 3 days in organ culture. Muscles were put into culture on E18. A,B: ICC-MY and pacemaker activity from a +/+ embryo after 3 days in culture. C,D: ICC-MY and pacemaker activity from the jejunum of a W/+ embryo. E,F: ICC-MY and slow wave activity from a WV/+ embryo. G,H: The absence of ICC-MY and pacemaker activity in a W/WV embryo after 3 days in culture. The number of each genotype from this litter is noted in each panel and refers to the same animals as in Figure 4. Scale bar in G applies to A,C,E.

From a total of 22 embryos isolated from 3 WB/ReJ-W/+ females, 5 were identified as W/WV. Resting membrane potentials of +/+, W/+, and WV/+ tissues averaged −62.1 ± 0.9 mV and displayed electrical slow waves averaging 11.1 ± 0.9 mV in amplitude that occurred at a frequency of 20.1 ± 1.1 cycles min−1 (n = 17). In contrast, W/WV mutant embryonic tissues that had resting membrane potentials averaging −60.2 ± 2.0 mV (n = 5) were electrically quiescent and Kit-immunopositive ICC-MY were absent.

Kit Signaling and the Plasticity of ICC Networks

We further examined the importance of Kit signaling on the maintenance of ICC-MY and electrical rhythmicity by incubating neonatal tissues for 3–6 days with the tyrosine kinase inhibitor, imatinib mesylate (STI571 or Gleevec™). Imatinib mesylate (1–5 μM) caused a dose-dependent reduction in slow wave activity and disrupted ICC networks (Fig. 7). This inhibition was due to the inhibition of Kit rather than non-specific side effects of imatinib mesylate as the drug had no effect on slow wave activity when it was applied acutely for up to 1 hr (Fig. 7).

Figure 7.

Disruption of ICC-MY networks and electrical rhythmicity exposed to imatinib mesylate (STI571 or Gleevec™). Blockade of Kit with imatinib mesylate led to dose-dependent (1–5 μM) loss of ICC-MY and pacemaker activity. A,D: ICC-MY and pacemaker activity of a P0 muscle cultured for 3 days under control conditions. B,C: The loss of ICC-MY in a P0 muscle cultured for 3 days in the presence of imatinib mesylate (3 and 5 μM). E,F: Associated with the loss of ICC-MY was a dose-dependent reduction in electrical pacemaker activity. Acute administration of imatinib mesylate for up to 1 hr had no effect on pacemaker activity in jejunal tissues (G,H). Scale bar in C applies to A,B.

The plasticity of ICC-MY and electrical rhythmicity was examined by treating muscles cultured from E18-P0 for 3 days with either ACK2 or imatinib mesylate and then removing these substances after ICC and electrical activity were lost. The muscles were allowed to recover for up to 9 days following the removal of ACK2 or imatinib mesylate. Removal of ACK2 from the culture media resulted in restoration of ICC-MY (Fig. 8) and a gradual return in electrical slow wave amplitude (Fig. 9). Prior to ACK2 application, the average resting membrane potential of jejunal tissues from 5 P0 neonates was −64.7 ± 0.9 and slow wave activity averaged 16.7 ± 3.3 mV in amplitude and occurred at a frequency of 22.7 ± 2.2 cycles min−1. ICC-MY and electrical slow waves were absent after 3 days exposure to ACK2. Three days after removal of ACK2 from the media, ICC-MY were partially recovered and slow waves averaged 7.0 ± 3.0 mV in amplitude and occurred at a frequency of 15.3 ± 2.4 cycles min−1 (n = 5). Six days after ACK2 was removed, slow waves averaged 21.8 ± 3.7 mV in amplitude and occurred at a frequency of 19.3 ± 1.5 cycles min−1. At 9 days, slow waves were similar in amplitude and frequency to control tissues cultured in the absence of ACK2 (i.e., 9 days recovery after ACK2 removal slow waves were 27.5 ± 2.5 mV in amplitude and occurred at 20.5 ± 1.5 cycles min−1; control tissues that were cultured for the same time period in media only displayed slow waves with an average amplitude of 35.0 ± 2.9 mV and frequency of 21.0 ± 1.0 cycles min−1; P > 0.05 for both slow wave amplitude and frequency). Removal of imatinib mesylate also resulted in recovery of ICC-MY and electrical slow waves over a 6–9-day period in 3 muscles (Fig. 10).

Figure 8.

Plasticity of ICC-MY after blocking Kit with ACK2 for 3 days in culture. AC: ICC-MY in P0 muscles that were cultured for 3–12 days, as denoted. ICC-MY were absent from muscles after culturing in the presence of ACK2 for 3 days (as in Fig. 2). Removal of ACK2 from the culture media for 3–9 days resulted in recovery of ICC-MY (DF, 3–9 days of recovery, respectively). Scale bar in F applies to all panels.

Figure 9.

Pacemaker activity is restored in muscles after ICC-MY recover. AD: The post-natal development of slow waves in P0 muscles cultured for 3–12 days under control conditions. After exposure to ACK2 for 3 days, slow wave activity was abolished. Removal of ACK2 from the culture media led to a gradual return in electrical slow wave activity over 3–9 days (FH, respectively).

Figure 10.

Plasticity of ICC-MY and pacemaker activity after inhibition with imatinib mesylate for 3 days in culture. A,D: ICC-MY and pacemaker activity after 6 days in culture under control conditions. B,C: Recovery of ICC-MY 3 days after the removal of imatinib mesylate (3 and 5 μM; 3 days exposure). E,F: Recovery of pacemaker activity 3 days after removal of imatinib mesylate (3 and 5 μM; following 3 days exposure). Scale bar in C applies to A,B.

DISCUSSION

In the present study, we have shown that ICC of the murine small intestine depend upon Kit signaling through the period of late gestation to develop from Kit-positive precursors into functional pacemaker cells. This conclusion is based on observations using two experimental approaches: (1) Treating muscles from E17 with a neutralizing antibody for Kit caused loss of ICC networks in small intestinal muscles and blocked the development of electrical rhythmicity. Even after electrical rhythmicity had developed (i.e., by E18 or at P0), treating muscles with the neutralizing antibody or with the tyrosine kinase inhibitor, imatinib mesylate, caused disappearance of ICC and loss of pacemaker activity. (2) Embryonic W/WV mice with compromised, but not complete, loss of Kit function (Nocka et al.,1990) failed to develop ICC-MY and pacemaker activity during the late gestational period as occurred in wildtype mice and W heterozygotes. Blocking Kit signaling during the late gestational or early neonatal periods does not close a developmental window on the ability of intestinal muscles to development a functional population of pacemaker cells. We found that terminating the blockade of Kit signaling caused development of a functional ICC-MY network within about 9 days. These data demonstrate the extensive plasticity of ICC in the gastrointestinal tract. We have previously shown that Kit neutralization can inhibit pacemaker activity in jejunal tissues from P10 animals, but the time required to block slow waves (35 days) was much longer than that required to block activity in P0 muscles (the present study), suggesting the sensitivity of ICC to signaling via the Kit receptor varies as a function of the developmental period of the animal (Ward et al,1999).

Other studies have concluded that Kit signaling is not important for the development of ICC during embryogenesis, but Kit must be functional shortly after birth to maintain and expand the ICC population. These results came from two studies using transgenic animals with mutations expected to completely abolish expression of Kit protein. In the first study, lacZ was introduced into the first exon of c-kit at the W/Kit locus in mouse embryonic stem cells (Bernex et al,1996). This created a c-kit null allele referred to as WlacZ. LacZ expression in animals with the transgene overlapped the expression of Kit in WlacZ/+ embryos, and the pattern of ICC-like distribution was the same in WlacZ/+ and WlacZ/WlacZ embryos. These authors concluded that ICC do not depend upon Kit expression during embryonic development. In the complete absence of Kit, cell migration, proliferation, and survival of lacZ-positive cells in the GI tract was apparently not impaired. In the second study, another mutation resulting in c-kit null mice (W banded mutation; Wbd/Wbd) were found to have cells at birth that were labeled with methylene blue and distributed in a pattern equivalent to ICC. These cells, however, failed to develop after P5 (Kluppel et al.,1998). The authors concluded that Kit signaling is not necessary for initial cell lineage decision for mesenchymal precursors to develop toward the ICC phenotype during embryogenesis, but Kit signaling is required for postnatal development and proliferation of ICC. Unfortunately, neither of these studies included assays of function, so it is unclear whether the development of ICC was completed (i.e., from Kit-positive precursors to functional ICC) in the absence of Kit. Our study argues that while development of ICC precursors (i.e., mesenchymal Kit-positive cells; see Torihashi et al.,1997) may not require Kit signaling, the next stage of development, true lineage decision toward a functional ICC phenotype, absolutely depends upon the Kit signaling pathway. ICC appear to receive stimulation of Kit via stem cell factor expressed by adjacent smooth muscle cells (Horvath et al.,2005). We have previously shown that precursors of ICC remain in an undifferentiated state if Kit is blocked (Torihashi et al.,1999). Thus, it is possible that mesenchymal precursor cells with null c-kit alleles express lacZ (as in Bernex et al.,1996), occupy the same approximate anatomical locations as mature ICC, and label with methylene blue (as in Kluppel et al.,1998) but these cells are not actually ICC (i.e., they lack the mature physiological functions of ICC).

It is also possible that in the complete absence of functional Kit protein, some form of biological compensation occurs, such that possibly another growth factor receptor facilitates ICC development. It should be noted that animals with W/W mutations, rescued from anemia by placental transplantation of wildtype hematopoietic stem cells, were viable (Fleischman and Mintz,1979), and this is hard to explain in the complete absence of functional ICC or other cells to replace the normally critical functions of ICC in the gastrointestinal tract (Sanders et al.,2006). The W mutation results in a non-functional Kit protein that is not present in the plasmalemma, analogous to the consequences of WlacZ/WlacZ and Wbd/Wbd mutations. The WV mutation encodes a protein that has reduced kinase activity, but some functional Kit is retained in compound heterozygous animals carrying W and WV alleles (Nocka et al.,1990). In the present study, we show that the severe loss-of-function of Kit in W/WV mutants was sufficient to block the development of functional ICC. Thus, the dependence upon Kit function for development of pacemaker ICC was clearly manifest in W/WV mutants.

At the present time, we are unclear about the fate of Kit-positive cells that are diverted from development of the ICC phenotype. Previous studies have investigated ICC in neonatal tissues treated with Kit-neutralizing antibody (Torihashi et al.,1999). As in the present study, blocking Kit caused loss of ICC in the small intestine. However, there was no evidence of cell death, cell necrosis, or apoptosis in ICC networks after chronic block of Kit. During the process of loss of ICC, remaining Kit-positive cells began to express smooth muscle-like antigens, such as desmin and smooth muscle myosin. Thus, it was suggested that blockade of Kit signaling causes redifferentiation of functionally mature ICC toward a smooth muscle cell phenotype. In the present study, we have presented data suggesting that this process may be reversible if Kit blockade is terminated or the tyrosine kinase function of Kit is restored. Thus, the ICC phenotype may possess a rather robust degree of plasticity that is regulated via Kit signaling.

The precise cellular changes that occur in response to the blockade of the Kit signaling pathway and the restorative changes that occur during redevelopment of functional ICC networks is an extremely interesting direction for future investigation, since a variety of human gastrointestinal motility disorders are accompanied by loss of ICC (Sanders et al.,2006; Vanderwinden and Rumessen,1999). In the other direction, proliferation of Kit-positive cells leads to formation of life-threatening gastrointestinal stromal tumors (Hirota et al.,1998; Nishida et al.,1998). Thus, developing the means to manipulate the ICC phenotype may have profound therapeutic benefits for these patients.

In summary, we have shown that blocking Kit signaling during late gestation either by pharmacological techniques or as a result of genetic defects results in failure of ICC networks and pacemaker function to develop in the murine small intestine. Late in gestation, sometime between E15 and E18, is a critical period for ICC development in which a lineage decision occurs and Kit-positive precursors begin to develop toward a functional ICC phenotype. Our data, in contrast to previous reports, indicate that Kit signaling during this period is critical for the development of ICC. After the development of ICC function, Kit signaling remains highly critical for the continued development and maintenance of the ICC phenotype. However, if the development of ICC is disrupted and Kit immunoreactivity in ICC falls to levels that are not resolvable by immunofluorescence techniques, functional ICC can regenerate if the blockade of Kit signaling is removed. These data suggest the exciting possibility that it may become possible to use the inherent plasticity of ICC for therapeutic purposes.

EXPERIMENTAL PROCEDURES

Animals

Pregnant BALB/c mice at precisely known stages of gestation were obtained from Jackson Laboratory (Bar Harbor, ME) or Harlan Sprague Dawley (Indianapolis, IN). WB/ReJ-W/+ (W/+; female) and C57BL/6J-WV/+ (WV/+; male) mice were obtained from Jackson Laboratory and the two mice mated to produce WBB6F1/J-W/WV (W/WV) along with WBB6F1/J-W/+ (W/+), WBB6F1/J-WV/+ (WV/+), and WBB6F1/J-+/+ (+/+). The gestational age of embryos was taken from the middle of the preceding night after discovery of the presence of a vaginal copulation plug. Mice were sacrificed at embryonic days E17–19 for molecular, morphological, and electrophysiological investigations. Neonate BALB/c, W/WV mutants and their siblings (W/+, WV/+, +/+) were also used on the first day following birth (P0). Animals were anesthetized by isoflurane (Baxter, Deerfield, IL) inhalation and exsanguinated following cervical dislocation. The small intestines were removed from fetuses or neonates and opened by sharp dissection along the mesenteric border and contents washed with Kreb's Ringer buffer (KRB). The mucosa was subsequently removed, revealing the underlying circular muscle layer. Muscles above the central point of the small intestine were used for this study, and this region was defined as the jejunal region. Previous studies have demonstrated pre-natal development of electrical rhythmicity in these muscles (Ward et al., 1997). The use and treatment of animals was approved by the Institutional Animal Use and Care Committee at the University of Nevada.

Molecular Studies

RNA isolation and RT-PCR.

Following mating of WB/ReJ-W/+ (W/+; female) and C57BL/6J-WV/+ (WV/+; male) mice, total RNA was extracted from the jejunums of the resulting F1 progeny at either E18 or P0. The tissues were stripped free of mucosa and TriZol™ (Invitrogen, La Jolla, CA) was used to isolate total RNA according to the manufacturer's instructions. Poly(A)+ RNA was isolated from the total RNA using Oligotex mRNA Mini kit (Qiagen, Chatsworth, CA). First-strand cDNA was synthesized using 200 U SuperScript II™ (Invitrogen) and 500 μg/μl of oligo dT primers at 42°C for 50 min in the presence of 100 ng poly(A)+ RNA in a 40 μl-reaction volume. Gene-specific oligonucleotide primers were designed and synthesized for RT-PCR (Table 1). The PCR reaction was performed using the GeneAmp PCR system 2700 (Applied Biosystems, Foster City, CA) and by using 12.5 μl of 2 × AmpliTaq Gold PCR Master Mix (Applied Biosystems), 1 μl of the synthesized cDNA and 10 pM of the primers in a 25-μl reaction volume. A Two-Step PCR method (95°C for 10 min, then 30 cycles of 95°C for 15 s and 60°C for 1 min) was used. After PCR, 2 μl of the RT-PCR product was analyzed on a 1.5% agarose gel. The 240-bp PCR products amplified by RT-PCR were gel-eluted using a PCR purification kit (MinElute kit, Qiagen) and analyzed by Nsi I digestion.

DNA sequencing and analysis.

The 437-, 203-, and the 240-bp PCR products were gel-extracted, subcloned into the TA cloning vector pcDNA3.1 (Invitrogen), and transformed into One Shot Top10 electro-competent cells (Invitrogen). Six individual plasmid clones from each F1 mouse were isolated using the QIAPrep Miniprep kit (Qiagen) for sequencing. DNA sequencing was performed using T7 and BGH primers at Nevada Genomic Center (University of Nevada, Reno, NV). The DNA sequences were analyzed by Vector NTI Suite v.6.0 (InforMax, Inc., MD) and by the BLAST program (Altschul et al.,1990).

Genotyping.

For genotyping, cDNAs from PCR templates were prepared from total RNA isolated from the jejunums as described above. Segments of jejunum isolated from the same animals were used for immunohistological and functional analyses. PCR products were generated through the use of two sets of primers Kit-W1 and Kit-W1r and Kit-WV1 and Kit-WV1r. Kit-W1 and Kit-W1r amplified a 437-bp product from all of the F1 generation from the WB/ReJ-W/+ (W/+; female) and C57BL/6J-WV/+ (WV/+; male) cross. The 437-bp product corresponds to a region spanning between exon 10 and exon 13 of the c-kit gene. Approximately 50% of the F1 generation showed a 203-bp product along with a faint 437-bp product, confirming they contained the W product. This 203-bp product is from a point mutation (GT to AT) at the 5′-splice donor site of intron 11, which leads to a deletion of exon 11 and exon 12 (234 bp) in the W locus (Hayashi et al.,1991). Kit-WV1 and Kit-WV1r amplified a 240-bp fragment from all samples. The 240-bp product corresponds to a region spanning between exon 14 and exon 15, within which a point mutation (C to T) can occur at the WV locus (Nocka et al.,1990). This mutation introduces an Nsi I site. In order to identify offspring expressing the WV point mutation, 240-bp products were digested with Nsi I, which generates a 153-bp fragment and an 87-bp fragment if the WV point mutation is present. Approximately 50% of the F1 generation was found to contain the WV product. Thus, the PCR and Nsi I digestion analyses revealed which individuals of the F1 generation were +/+; W/+, WV/+, or W/WV. We used two additional sets of the nested primers Kit-W2 and Kit-W2r and Kit-WV2 and Kit-WV2r (see Table 1) to confirm genotyping results (data not shown). The 437-, 203-, and 240-bp PCR products were confirmed by sequencing. The sequence of all the 437-bp products matched 100% with the reported c-kit cDNA (MN_021099). In addition, about 50% (2 or 3 of the 6 sequenced clones per each F1 mouse) of the 437-bp products contained a 12-bp insertion between exon 10 and exon 11 (Fig. 4). This insertion is due to an alternative splicing at the 5′ end of the exon 11 and introduces four additional amino acid residues in Kit. The sequences of the 203-bp products confirmed a deletion of exon 11 and exon 12 (234 bp) in the W locus. The 12-bp cryptic exon was also found in all 203-bp products. Sequencing of the 240-bp products confirmed the presence of the point mutation (C to T; Fig. 4).

Immunohistochemistry.

Jejunal tissues from E17 to P0 mice were pinned onto the base of a Sylgard dish mucosal side up. The tissues were opened along the mesenteric border and luminal contents washed with Krebs Ringer bicarbonate solution (KRB). The mucosa was removed by sharp dissection and the remaining muscularis tunica fixed in acetone (4°C; 10 min). Following fixation, preparations were washed for 30 min in phosphate buffered saline (PBS; 0.1M pH 7.4). Non-specific antibody binding was reduced by incubation in 10% rabbit serum for 1 hr at room temperature. Tissues were incubated overnight at 4°C with a rat monoclonal antibody raised against c-kit protein (ACK2; 5 μg/ml in PBS; Gibco BRL, Gaithersburg, MD). Immunoreactivity was detected using FITC-conjugated secondary antibody (FITC-anti-rat 1:100 in PBS, 1 hr, room temperature). Control tissues were prepared in a similar manner, omitting either ACK2 or the secondary antibody from the incubation solution. Whole mounts were examined using a Zeiss LSM 510 Meta laser scanning confocal microscope with an excitation wavelength appropriate for FITC (488 nm). Confocal micrographs are digital composites of Z-series scans of 10–15 optical sections through a depth of 20–35 μm. Final images were constructed using Zeiss LSM software.

Organ culture.

Jejunal segments were isolated from E17 to P0 animals and opened along the mesenteric border. Luminal contents were removed with KRB and the mucosa removed by sharp dissection. Muscle strips (5 × 2 mm) were cut and pinned to the base of a sterile tissue culture chamber slide lined with Sylgard, with the mucosal side of the circular muscle facing upward. Tissues were washed 4 times with KRB and placed in M199 media (Sigma) containing penicillin (200 U/ml), streptomycin (200 μg/ml), and amphotericin B (0.5 μg/ml), washed another 4 times, and incubated at 37°C (90% humidity and 95% O2–5% CO2 for up to 12 days with culture media being changed every second day. Some tissues were incubated in M199 media containing ACK2 antibody (5 μg/ml); ACK2 was not added to chamber slides containing control tissues. In a separate series of control experiments, bovine non-immune serum (5 μg/ml; Hyclone, Logan, UT) was added as a control for ACK2.

Electrophysiological studies.

After removing the mucosa, jejunal muscle strips (4 × 2 mm for fetal tissues and 8 × 4 mm for neonates) were cut and pinned to the Sylgard floor of a recording chamber with the mucosal side of the circular muscle facing upward. Alternatively, tissues were taken at different time periods in organ culture and placed in the recording chamber. Electrical recordings were made in the presence of nifedipine (10−6 M) to reduce muscle contractions and facilitate impalements of cells for extended periods.

Circular muscle cells were impaled with glass microelectrodes filled with 3M KCl and having resistances between 80–120 MΩ. Transmembrane potential was measured using a high input impedence amplifier (WPI S-7071) and outputs were displayed on an oscilloscope. Electrical signals were recorded onto a PC running AxoScope 9.0 data acquisition software (Axon Instruments, Union City, CA). Data analysis and construction of figures were performed using Clampfit analysis software (Axon Instruments). Data are expressed as means ± standard errors of the mean. Differences in the data were evaluated by Student's t-test, P < 0.05 were taken as a statistically significant difference.

Solutions and drugs.

Muscles were maintained in KRB (37.5 ± 0.5°C; pH 7.3–7.4) containing (in mM): Na+, 137.4; K+, 5.9; Ca2+, 2.5; Mg, 1.2; Cl, 134; HCO3, 15.5; H2PO4, 1.2; dextrose, 11.5 and bubbled with 97% O2–3% CO2. Imatinib mesylate was obtained as a gift from Novartis Pharma (Basel, Switzerland) and was dissolved in dH2O at a stock concentration of 10 mM before being added to the organ culture media or perfusion solution of electrophysiological experiments at a final concentration of 1–5 μM.

Nifedipine was obtained from Sigma-Adrich Co. (St Louis, MO) and dissolved in ethanol at a stock concentration of 10 mM before being added to the perfusion solution of electrophysiological experiments at a final concentration of 1 μM.

Ancillary