Introduction

Maturation related alterations in intestinal structure are important for the development of small intestinal motility throughout the postnatal period (Thomson and Keelan 1986). The diameter, length and wall-thickness of the small intestine of adult animals have increased several folds when compared with measurements obtained from neonates. In addition, intestinal contractions and motility are greatly enhanced after birth (Lu et al. 2005).

It is widely accepted that interstitial cells of Cajal (ICCs) play an important role in the regulation of gastrointestinal motility (Thomsen et al. 1998). ICCs express c-kit gene product, Kit (a tyrosine kinase receptor) (Huizinga et al. 1995), which is essential for the determination of cell lineage, survival and proliferation during embryonic and neonatal periods (Torihashi et al. 1999; Beckett et al. 2007), and is a marker protein of ICCs. Kit positive ICCs or their precursors of ICCs first appear in the murine intestinal wall at E12.5 (Torihashi et al. 1997; Young et al. 1996) and subsequently, start to cluster around the myenteric nerve plexus. At the time of birth, one layer of ICCs, so called ICC-MY, is located between the longitudinal and circular smooth muscle layers (Ward and Sanders 2001). Studies have shown that ICC-MY fully differentiate and mature 48 h after birth and possess identical ultrastucture and electrophysiological activity (Liu et al. 1998). These cells act as the pacemaker cells that generate and propagate the slow waves (Lee et al. 1999; Ordog et al. 1999). ICC-DMP, another subtype of ICC in the small bowel, which are closely associated with nerve fibers (Ward et al. 2006; Zhou and Komuro 1992), and appear soon after birth (Ward and Sanders 2001). However, the postnatal development of ICCs, from neonatal to adult life, is poorly investigated, although the small intestine keeps lengthening and extending throughout this period, in spite of the fact that developmental disorders related to ICCs have been shown to be involved in a number of dysfunctions of the gastrointestinal tract in newborns and infants, such as intestinal pseudo-obstruction (Krishnamurthy et al. 1993), Hirschsprung’s disease (Vanderwinden et al. 1996b) and infantile hypertrophic pyloric stenosis (Vanderwinden et al. 1996a).

Moreover, the progenitor of ICCs in the postnatal gastrointestinal tract has, until now, not yet been unambiguously identified. A recent research has identified a type of cell, rare and resembling ICCs progenitor, which can be expanded by soluble Kit ligand, stem cell factor (SCF), and further differentiate into mature ICCs (Lorincz et al. 2008). Similarly, several in vitro experiments have shown that SCF can expand ICCs obtained from the embryonic and neonatal murine intestine (Wu et al. 2000), however, ICCs from the mice older than day-6 post-partum do not proliferate (Nakahara et al. 2002) and suggests that the proliferation of ICCs is time-limited and SCF-dependent under the culture conditions. In spite of the clear evidence for proliferation in vitro, it is not clear yet whether the proliferation of ICCs is involved in the postnatal development of ICCs.

To answer this question, we have used immunofluorescent staining to study alterations in morphology and number of ICCs in the murine small intestine over the period extending from neonatal (P0) to adult (P56) life.

Materials and methods

Animals

BALB/C mice were purchased from the Animal Center of the Third Military Medical University (Chongqing, China) and paired to produce offspring. Eighty-seven postnatal mice were used in our experiments. Animals were divided into 4 groups: 1) Twenty-four mice were killed at postnatal day 0 (P0), P2, P5, P8, P16, P24, P32 and P56; 2) twenty-one mice were injected with one dose of BrdU (0.1 mg/g) intraperitoneally 24 h before sacrifice at P2, P5, P8, P16, P24, P32 and P56; 3) twenty-one mice were injected daily with BrdU from P0 and were killed at P2, P5, P8, P16, P24, P32 and P56; 4) twenty-one mice were given one injection of BrdU (0.1 mg/g) on P10 and were killed at 1, 2, 8, 16, 24, 48 and 96 h later. All experiments were performed in accordance with our University Health Guide for the Care and Use of Laboratory Animals.

Immunofluorescence

Ten to five minutes before the abdomen was opened to expose the small intestine, a dose of papaverine (0.12 mg/g), a calcium antagonist, was injected intraperitoneally in order to abolish contractile activity in the small intestine. The entire small intestine from the pylorus to ileo-caecal junction was removed tenderly and placed into the PBS containing papaverine (0.5 mg/ml), and then the length and diameter of the small intestine were measured quickly. To obtain whole-mount preparations, the small intestine was inflated with acetone to the size identical with the one when the parameters were measured formerly. After 30 min fixation, and the longitudinal smooth muscle layer containing the ICC-MY was prepared with the aid of a dissection microscope. The immunostaining procedures have been previously described (Komuro and Zhou 1996). Briefly, after the specimens were incubated with primary antibody (Table 1) for 8 h at 4°C, the immunoreactivity was detected by using a Cy3-conjugated secondary antibody (anti-rat IgG, 1:100; Zymed), or FITC-conjugated secondary antibody (anti-mouse IgG, 1:100; DAKO), or Cy5-conjugated secondary antibody (anti-rabbit IgG, 1:100; Zymed). To identify the proliferation of ICCs, the whole-mount preparations were stained for ACK2 (Kit) as described above, and then double labeled for BrdU as follows: the specimens were additionally fixed in 4% paraformaldehyde in 0.1 M PB at pH 7.2 for 10 min. After rinsing in PBS, the specimens were treated with 2N HCl for 30 min at 37°C. To reveal BrdU labeling, the specimens were incubated with a mouse monoclonal antibody raised against BrdU (3 μg ml−1; DAKO) overnight at 4°C and then incubated with a FITC-conjugated secondary antibody (anti-mouse IgG, 1:100; DAKO). The stained results were examined with a BX51 fluorescence microscope (OLYMPUS, Japan) or a TCS SP5 confocal laser scanning microscope (Leica, Germany) with an excitation wavelength appropriate for Cy3 (552 nm), FITC (488 nm) and Cy5(625 nm). The Z stacking of confocal images at 3–5 μm intervals contained all the levels of positively stained cells and processes.

Table 1 The primary antibodies
Full size table

Measurement and statistical analysis

Six intestinal segments from each experimental animal were sampled in a random manner and intestinal diameter measured and immunofluorescent staining assessed. Photomicrographs of both Kit positive cells and Kit/BrdU double labeled cells were taken in 10 random fields (×200 magnification, 0.2607 mm2) per whole-mount preparation with a digital camera (SPOT, Diagnostic Instruments, Inc) mounted on a BX51 fluorescence microscope (OLYMPUS, Japan). The numbers of either Kit positive cells or Kit/BrdU double labeled cells were counted with Image-Pro Plus 5.0 (Media Cybernetics). The estimated surface area of ICC-MY in the small intestinal wall = (mean diameter − 2 × mean thickness of longitudinal smooth muscle layer) × π × length. Multiplying the estimated surface area of ICC-MY and mean density together made the estimated total cell number of ICC-MY. Data are expressed as means ± S.E.M. The n value reported in the text refers to the number of animal used. Differences in the data were evaluated by student t test, and P < 0.05 was taken as a statistically significant difference.

Results

Morphologic alterations of ICCs

The Kit positive ICC-MY were observed between the longitudinal and circular muscular layers around the myenteric nerve plexus. In the duodenum and jejunum at birth (P0)(Group 1), these cells were arranged into a loose cellular network that connected with each other via 2–3 short and slim processes with several branches (Fig. 1a, b). In comparison, ICC-MY in the ileum seemed to have fewer and shorter processes, or lacked definitive processes entirely, thus making the network seem incomplete (Fig. 1c). By P2 and P5, the processes of ICC-MY in the duodenum and jejunum were more numerous and longer but had thinner processes with branches than compared with that seen at P0 (Fig. 1d, e). Similar changes were seen in the ileum, albeit the processes were shorter and lower in number (Fig. 1f). Subsequently, at P8 and P12, the Kit-postive cells increased in density and possessed thicker and more abundant processes, and an compact cellular network was seen (Fig 1g–i). These cells were similar to the morphologically mature cells observed in P16 animals, which had a large number of winding branching processes. Well-developed ICCs cellular networks were observed in duodenum, jejunum and ileum by P32 (Fig. 1j–l) and P56.

Fig. 1
figure 1

Confocal images of murine small intestine whole-mount preparations. ICC-MY were labeled with Kit and illustrate alterations of ICC morphology and density in the duodenum (a, d, g, J), jejunum (b, e, h, k), and ileum (c, f, i, l) at P0 (a–c), P5 (d–f), P12 (g–i) and P32 (j–l). a–c: ICCs formed a loose cellular network interconnected by 2–3 slim, short, primary processes with very few branch at P0; ICCs in the duodenum (a) and jejunum (b) seemed to be better developed compared with cells in the ileum (c). d–f: ICCs had more elongated and slim processes with several branches and had a higher density at P5 in the duodenum (d) and jejunum (e), while ICCs of the ileum had comparably fewer processes and a lower density (f). g–i At P12, ICCs projected their thick cytoplasmic processes to form an intact cellular network, which was not so apparent in the ileum (i). j–l: ICCs developed further into a complicated network with gross processes at P32, which were almost identical to that of adult. Scale bar a–l 75 μm

Full size image

Density and estimated total cell number of of ICCs

The quantitative data showed that the mean density of Kit-positive ICC-MY in the small intestine was about ~750 mm−2 at P0, increasing to a maximum of 1,010 mm−2 at P12 (Table 2). Subsequently, the cell number decreased to about 720 mm−2 at P32, a value which was similar to that seen in adult mice (P56) (Table 2). The length and the average diameter of small intesine were 71.3 and 0.98 mm respectively at P0, and these values increased to 297.2 and 3.93 mm by P32 (Table 2). The estimated area of the small intestinal wall increased about 17-fold from 200 mm2 at P0 to 3,520 mm2 at P32 (Table 2). Therefore, given the density of ICC-MY and the estimated area of small intestinal wall, the total cell number of ICC-MY was estimated at about 1.52 × 105 at P0 and at 25.29 × 105 at P32, suggesting a more than 15-fold increase in total cell number (Table 2).

Table 2 The mean density and estimated total cell number of Kit positive ICC-MY, and the mean diameter, full length and estimated total area of the small intestinal wall of mice with different ages from P0 to P56
Full size table

Detection of proliferative ICCs

In the preceding section we showed that the total number of ICC-MY increased by more than 15-fold. We wished to know the underlying mechanisms for this increase and therefore BrdU was used to detect the presence of proliferative events of ICCs. BrdU was incorporated into Kit positive cells 1–2 h after injection into the P10 mice (Group 4), most cells had 2–3 thick processes without branch and were not connected with the cellular network (Fig. 2a–c), The number of such cells amounted to 32.4 and 48.2 mm−2 at 1 and 2 h respectively (Table 3). Then paired Kit/BrdU double labeled cells began to show up 8 h post-BrdU-injection, and the double positive cells always distributed in pairs at 24 h (Table 3). Each paired double labeled cells was characterized by closely located cell bodies and similar features (Fig. 2d–f). Quantitation showed that the cell numbers had nearly doubled within 24 h (Table 3). These cells further developed into the cells with prolonged processes and branches and had features similar to those of mature ICC-MY 48 h later (Fig. 2g–i), although the density decreased slightly at 48 and 96 h, possibly due to the expansion intestinal wall area (Table 3).

Fig. 2
figure 2

Confocal microphotographs showing presumptive proliferations of ICCs on whole mount preparations of ileum in the mice at P10. a–c A number of double labeled Kit+ (red) and BrdU+ (green) cells (arrowheads) were observed 1 h after one dose of BrdU injection. These cells had 2–3 thick processes without branches, and did not appear mature. d–f Kit+/BrdU+ were often distributed in pairs at 24 h, each pair of these cells had adjacent cell bodies (arrows) and identical profiles. g–i These paired double labeled cells (arrowheads) had developed features similar to those of mature cells at 48 h. Scale bar: a–i 30 μm

Full size image
Table 3 The density of Kit/BrdU double labeled cells in ICC-MY of the mice10-day post-partum from 1 to 96 h after one dose of BrdU injection
Full size table

Proliferative ability of ICCs

To investigate the proliferative ability of mice with different ages, BrdU/Kit double labeled ICC-like cells were detected 24 h after one dose of BrdU injection (Group 2). The analysis showed that the BrdU labeled ICCs were seen frequently on whole mount preparations (Fig. 3a–c), and about 178.4 such cells (mm−2) were observed at P2 (Table 4). The value gradually decreased to 102.2 and 68.4 respectively by P8 (Fig. 3d–f) and P16 (Fig. 3g–i) (Table 4). The BrdU positive ICC-MY were rare in P24 samples (Fig. 3j–l) and were not present by P32 (Fig. 3m–o) or P56 (Table 4). These results suggest that ICC proliferation is age-dependent and may cease just beyond P24.

Fig. 3
figure 3

Confocal images of ICC-MY labeled with Kit (red) and BrdU (green) on whole-mount preparations of the murine ileum at P2 (a–c), P8 (d–f) and P16 (g–i) showing the proliferation of ICCs 24 h after a single dose of BrdU. a–c At P2, a large amount of Kit/BrdU double labeled cells (arrowheads) were observed with slim and short processes. d–f At P8, ICC-MY had longer and thicker processes, some of which were both Kit and BrdU positive (arrows); the number of these cells was less than that seen at P2. g–i At P16, a number of Kit+/BrdU+ (arrowheads) also could be observed, which indicated that the proliferation of ICCs persist until early adulthood. The BrdU positive ICC-MY were rare to see (arrow) in P24 samples (j–l) and were not present by P32 (m–o). Scale bar: a–o 50 μm

Full size image
Table 4 Mean density and percentage of Kit/BrdU double labeled cells in the ICC-MY of the mice with different ages after one dose of BrdU injection for 24 h or daily injection of BrdU from birth
Full size table

To determine the fraction of ICC-MY which resulted from proliferation, daily BrdU injection was performed from birth until the animals were killed (Group 3). At P2, about 28% ICC-MY were both Kit and BrdU positive. The fraction increased to 35% at P5 (Fig. 4a–c; Table 4), reaching a peak value of 64% by P24 (Fig. 4d–f) and P32 (Table 4), which implied that ICC-MY expanded their volume mainly by the way of proliferation.

Fig. 4
figure 4

Confocal microphotographs showing the ICCs which emerged by proliferation on whole mount preparations of ileum with daily injection of BrdU from P0. a–c A large number of double labeling of Kit (red) and BrdU (green) cells (arrowheads) were observed at P5. More double positive cells (arrows) were seen at P24 (d–f). Scale bar: a–f 60 μm

Full size image

Origin of the proliferative ICCs

Since vivid proliferation was seen in the postnatal development of ICCs, then we wanted to investigate the origin of these proliferating ICCs. A recent study has identified a very small group of Kit+/CD44+/CD34+/insulin receptor+/insulin-like growth factor I receptor (IGF-IR)+ cells in murine stomach as the putative progenitors of ICCs (Lorincz et al. 2008). Interestingly, our results showed that most Kit/BrdU double positive cells were also labeled with CD44 (arrows Fig. 5a–d) and a number of Kit+/BrdU+ cells were also labeled with IGF-IR (arrow Fig. 5e–h) in the mice after one dose BrdU injection for 1 h at P10 (Group 4). In addition, CD34, a cellular marker of hematopoeitic cells and endothelial cells of blood vessels, was also found expressed by a number of BrdU labeled ICC-MY (arrowhead Fig. 5i–l). Besides that, vimentin+/IGF-I+ fibroblast-like cells are regarded as the progenitors of ICCs (Pucilowska et al. 2000). Our finding indicated that such cells were very rare around myenteric plexus in the small intestine (arrow Fig. 5m–o) which suggested these cell might the progenitors of ICCs in a certain situation.

Fig. 5
figure 5

Confocal microphotographs showing the Kit+/BrdU+ cells expressing CD44 (a–d), IGF-IR (e–h) and CD34 (i–l) on whole mount preparations of small intestine in the mice at P10 after one dose BrdU injection. a–d Most Kit (red)/BrdU (green) double labeled cells were also CD44 positive (blue) (arrows). A number of them were stained with IGF-IR (blue) (arrow) (e–h), or CD44 (blue) (arrowhead) (i–l). m–o Very rare vimentin+ (green)/IGF-I+ (red) fibroblast-like cells were observed on whole mount preparations (arrow) counterstained with DAPI (blue) at P8. Scale bar: a–l 20 μm, m–o 30 μm

Full size image

Discussion

The present study demonstrates that: (1) The estimated total cell number of Kit positive ICCs increases about 15-fold from neonatal to adult life in mice along with the extension and elongation of the small intestine; (2) ICCs expand their volume mainly by an age-dependent proliferation which persists from day-0 to day-24 post-partum; (3) Some of the newly emerged Kit/BrdU double positive cells were also labeled with CD34, CD44 and IGF-IR which implies ICCs progenitors may be the origin of proliferation.

Morhphogic and quantitative alterations of ICCs

ICCs begin to appear during the embryonic period (Torihashi et al. 1995; Faussone-Pellegrini et al. 1996; Torihashi et al. 1997; Ward et al. 1997), and the ICC-DMP first appear at the time of birth (Ward and Sanders 2001). ICC-MY are immature at birth both in ultrastructure and physiologic functions (Liu et al. 1998). About 2 days after birth, ICC-MY develop ultrastructural features and electrophysiological features that are essentially identical to mature ones (Faussone-Pellegrini et al. 1996; Liu et al. 1998). Our results show that Kit-positive ICC-MY are loosely interconnected with each other by several slim and very short processes, and a loose cellular network is observed at birth. A complete cellular network is seen at P2, with increasing thick, branching processes appearing from P5 to P32. The processes of ICCs are closely connected with SMCs and enteric nerves which can propagate the slow waves and receive nervous integration (Rumessen and Thuneberg 1991). The gradual enrichment and growth of processes might contribute to the postnatal development of the intestinal motility, in other words these morphological alterations would be adapted for the large increase in smooth muscle cells (Lu et al. 2005) and enteric neurons. In addition, it seems that ICC-MY of the duodenum and jejunum develop sooner and are morphologically better than those of the ileum, which is probably due to a better developed motility in these regions (Lu et al. 2005).

The small intestine increases in diameter and prolongs in length for multi-fold from postnatal period to adulthood in rats (Lu et al. 2005), and our results reveal similar alterations in mice. Accordingly the estimated wall area of small intestine increases more than 17-fold during this period. The mean density of ICC-MY does not alter dramatically with an increase of approx 40% from P0 to P12 and then decreased by approx 30% at P32. However, it is noteworthy that the estimated total number of ICCs in the small intestine increases by more than 15-fold over the neonatal to adulthood period in mice. This result indicates that ICCs would keep increasing in cell number during postnatal development rather than remaining constant.

Proliferation of ICCs

Considering great increase of the ICCs cell number, the question arises what mechanism is involved in the dramatic increase in cell number? BrdU incorporation has been applied to detect the proliferative ICCs. Our result shows that, after daily injection of BrdU from birth, about 28% of Kit labeled ICCs are also BrdU positive at P2, and this ratio increases to 65% by P32, thus indicates the proliferation is a major method for ICCs to expand their volumes. Besides that, to measure the proliferative ability of ICCs of the small intestine at different ages, the incorporation of BrdU is detected at 24 h after one dose injection of BrdU to mice with different ages. It is noteworthy that the BrdU/Kit double labeled ICC-like cells are most abundant on day-2 and are still added until day-24 post-partum albeit in declining numbers, after which ICCs addition ceases. These results strongly suggest that the proliferation of ICCs is age-dependent and can persist from neonatal to almost adult life. In contrast, under in vitro culture conditions, only ICCs obtained from mice before P6 can proliferate in response to Kit ligand which suggests a time-limited and SCF-dependent proliferation of ICCs (Nakahara et al. 2002). The difference between in vitro and in vivo conditions might be explained by the fact that in vivo conditions are more appropriate for proliferation of ICCs than that in vitro conditions, i.e. proliferation of ICCs is dependent on suitable microenvironments. Addition of the Kit blocker (Imatinib) to organ cultures of the small intestine of juvenil mice leads to loss of ICCs, however, ICCs do reappear after removal of the agent (Beckett et al. 2007). Similarly ICCs in the small intestine of adult animals can regenerate after a reduction in cell numbers following gut transection and anastomsis (Yanagida et al. 2004; Mei et al. 2006), or obstruction (Chang et al. 2001). The recovery of ICCs in these cases might be the result of a re-expression of Kit protein of ICCs. Proliferation of ICCs or ICCs precursors is evident in the postnatal development, thus it is probable that proliferative events are also involved in the recovery of ICCs numbers as well as the resumption of Kit receptor.

Origin of the proliferative ICCs

Several investigations have proven that ICCs originate from mesenchymal cells (Kluppel et al. 1998; Lecoin et al. 1996; Wu et al. 2000; Young et al. 1996; Young 1999). The precursors of ICCs have the ability to differentiate towards ICCs under the stimulation of Kit signaling (Ward and Sanders 2001). A recent study has identified a very small group of Kit+/CD44+/CD34+/insulin receptor+/insulin-like growth factor I receptor (IGF-IR)+ cells in murine stomach as the putative progenitors of ICCs which could be expanded under the stimulation of SCF (Lorincz et al. 2008). The present study shows that BrdU is initially only incorporated (within an hour after injection) into a type of Kit positive cell with bi- or tripolar thick processes without branches, which does not seem to be mature ICC-MY. Then paired Kit+/BrdU labeled cells are often encountered 24 h post-BrdU-injection, concomitant with a cell number doubling. These paired cells further change into the features similar with mature ones. Besides that, some of Kit+/BrdU+ cells were also labeled with CD44, IGF-IR and CD34. Therefore, it is possible that these Kit positive cells with thick bi- and tripolar might be ICCs precursors at a certain stage of commitment (Huizinga and White 2008; Lorincz et al. 2008) that can proliferate and further differentiate into mature ICCs during the postnatal development. However, further investigations into the ultrastructural, physiologic and biologic characteristics of these cells are necessary to further substantiate this hypothesis.

In addition, vimentin+/IGF-I+ fibroblast-like cells are regarded as the progenitors of ICCs (Pucilowska et al. 2000), and ICC progenitor cells may switch their Kit expression (Horiguchi and Komuro 2000). And it may be possible that the Kit phenotype may switch between mature ICC and fibroblast-like progenitor cells. Our finding indicates that these vimentin+/IGF-I+ fibroblast-like cells are very small in amount (<2%) in the small intestine comparing with the Kit+ cells from P0 to P32 (data not shown) which suggested these cells might the progenitors of ICCs at certain stage and the switch of the phenotype between mature ICCs and the progenitor seemed to contribute slightly to the proliferation of ICCs and increase of ICCs cell numbers.

It is known that a number of motility disorders in children are associated with the reduced ICCs numbers, such as intestinal pseudo-obstruction (IPO) (Isozaki et al. 1997; Kenny et al. 1998; Wang et al. 2000) and Hirschsprung’s disease (HD) (Yamataka et al. 1995; Horisawa et al. 1998). This investigation has shown clear age-dependent changes in the morphology and cell number (by proliferation) of ICCs during the postnatal developmental period. In this regard, the precursors of ICCs might be a potential target for the future disease treatment, either through in vivo and/or in vitro expansion of these cells and might be a feasible therapeutic strategy.