Abstract
In this study, the localization and appearance of neuronal nitric oxide synthase-immunoreactive (nNOS-IR) nerve cells and their relationships with the developing gastric layers were studied by immunocytochemistry techniques and light microscopy in embryonic rat stomach. The stomachs of Wistar rat embryos aged 13–21 days were used. The first nerve cells containing nNOS-IR were seen on embryonic Day 14. The occurrence of mesenchymal cell condensation near nNOS-IR neuroblasts on embryonic Day 15 may reflect an active nerve element-specific mesenchymal cell induction causing the morphogenesis of muscle cells. Similarly, the appearance of glandular structures after nNOS-IR neuroblasts, on embryonic Day 18, suggests that the epithelial differentiation may depend on inputs coming from nNOS-IR neuroblasts, as well as other factors. Observation of nNOS-IR nerve fibers on embryonic Day 21 demonstrates that at this stage they contribute to nonadrenergic noncholinergic relaxation. In conclusion, depending on this study's results, it can be said that cells and tissues might be affected by NO secreted by nNOS-IR nerve cells during the development and differentiation of embryonic rat stomach.
N
Normal NADPH-diaphorase (NADPHd) in the brain was recently shown to be NOS and was found in neurons and their processes in rat digestive tract (Aimi et al. 1993). NADPHd-positive neurons have previously been described in the myenteric plexus of the embryonic mouse gut and have also been noted in the myenteric plexus and submucous plexus of the adult mouse (Epstein et al. 1983).
NO is an important paracrine signaling molecule in a variety of cell types (Synder and Bredt 1991). This kind of signaling plays a critical role in regulating cell-to-cell interactions, including many interactions among different types of cells that take place during embryonic development (Synder and Bredt 1991). Therefore, some investigators suggested that NO may be effective in cell and tissue differentiation (Faussone-Pellegrini et al. 1996).
The cell interactions within the neural crest-derived aggregates and the enteric microenvironment may be important in the establishment of the differentiated phenotypes and neural circuitry within and between ganglia (Gershon et al. 1993). Microenvironmental changes might also influence cell differentiation and, more specifically, the myoblast-differentiated nerve cell interactions are important for smooth muscle cell differentiation and mucosal differentiation (Gershon et al. 1993). Recent genetic studies demonstrated the importance of distinct receptor-ligand pathways in modulating the critical cell-to-cell interactions during gastrointestinal development, and highlight the potential clinical significance of alterations in these pathways in certain gastrointestinal diseases and disorders (McHugh 1996). Harasawa et al. (1979) reported that gastric motility may be one of the important factors for epithelial growth or differentiation because they observed a further decrease in the gastric emptying rate in gastric ulcer with a history of recurrence.
Up to now, few reports have described the mechanism of normal development and regeneration of gastric mucosa. The ultrastructure of differentiated cells and their maturation process are well known. Thus far, however, no reports have dealt with the relationship between nNOS-IR neuronal elements and other tissues in gastric development, i.e., mutual interactions during the development of muscle layers and nNOS-IR neuronal elements.
We have therefore investigated the occurrence, localization, and morphological features of nNOS-IR neuronal elements in embryonic rat stomach. In the course of this study we have also looked at the development of nNOS-IR neurons and its relationship to the developing smooth muscle, epithelium, and gland formation in rat stomach.
Material and Methods
Animals and Tissue Collection
Seventy-two adult female and 36 adult male Swiss albino rats weighing 190–230 g (Akdeniz University Medical Research Center; Antalya, Turkey) were used in this study. A vaginal smear from each female was examined daily and rats entering estrus were mated with adult males of the same strain. Vaginal smears were examined on the following morning for the presence of sperm. If sperm was observed, this day was designated as Day 1 of pregnancy and the females were housed in plastic cages under light-dark cycles of 12 hr. All rats were fed standard rat food pellets and were given water ad libitum throughout the experimental period.
Eight or 10 embryos were removed from each rat uterus after sacrifice by ether overdose from Day 13 to Day 25 of pregnancy. Six embryos were used at each developmental stage. The abdomens of the fetal rats were opened under sterile conditions and stomachs were detached at the esophageal and pyloric junctions. Between embryonic Days 18 and 21 (E18–21) of gestation, each stomach was divided into fore-stomach (proximal), corpus (middle), and pylorus (distal).
Whole stomachs for E13–17 and forestomach, corpus, and pylorus samples for E18–21 were fixed in 10% neutral formaldehyde (pH 7.2) for 6 hr and were dehydrated in ethanol, cleared in xylene, and embedded in paraffin. All the samples were serially sectioned at 6 μm perpendicular to the longitudinal axis of the stomach.
Conventional Staining
Some of the paraffin sections were hydrated and stained with hematoxylin (Merck; Darmstadt, Germany) and eosin (Merck) (HE) for routine morphological studies (Gurr 1973) and viewed through an Axioplan Microscope (Zeiss; Oberkochen, Germany).
nNOS Immunocytochemistry
To identify nNOS nerve cells, a polyclonal antibody to nNOS (rabbit anti-nNOS; Transduction Laboratories, Lexington, KY) was applied to the rest of the paraffin sections. The paraffin sections were dewaxed with xylene, dehydrated through graded alcohols, and placed in PBS (pH 7.2). The paraffin sections were placed in citric acid (Sigma, St Louis, MO; pH 6) and were exposed to microwaving (750 W) twice for 5 min. Endogenous peroxidase was blocked by treatment with 3% H2O2 in methanol for 15 min and then sections were placed in PBS (pH 7.2). Nonspecific binding sites were blocked with 5% normal swine serum (Signet Laboratories; Dedham, MA) in PBS. Tissue sections were incubated overnight at 4C with the nNOS primary antibody (diluted 1:1000). The immunocytochemical procedure was controlled by replacing the primary antibody with IgG (Santa Cruz Biotechnology; Santa Cruz, CA) of the appropriate non-immunized species (Taguchi et al. 2000). After that, sections were incubated for 30 min with biotinylated swine anti-rabbit immunoglobulins (DAKO; Santa Cruz, CA), followed by a third incubation with streptavidin-peroxidase complexes (DAKO) for 30 min. After each incubation sections were rinsed in PBS. The peroxidase activity was developed with 0.03% 3,3′-diaminobenzidine tetrahydro-chloride (DAB; DAKO). The sections were counterstained with hematoxylin, dehydrated, and mounted in mounting medium (BioGenex; San Ramon, CA) (Wehby and Frank 1999).
In these sections, nNOS-IR staining was evaluated in a semi-quantitative fashion under light microscopy and selected ganglia were photographed (Kodak; Rochester, NY). The evaluations were denoted as — (no staining), —/+ (very weak staining), + (weak staining), ++ (distinct), +++ (intense), ++++ (most intense).
The light microscopic morphometric studies of the stomachs from each embryonic age were performed using a Zeiss eyepiece attached to the ocular of an Axioplan Microscope (Zeiss) under ×400 magnification. Previously selected areas observed in sections were screened to identify nNOS-IR nerve cell bodies from forestomach to pylorus of the embryonic stomach, a few microscopic areas from each slide. nNOS-IR nerve cells in myenteric ganglia were counted in each microscopic area and the diameters of cell bodies with nuclei were measured. The half of the sum of short and longitudinal diameters was accepted as the diameter of a cell body. Similarly, between E17–21 the thickness of the muscle layer of the stomach from each embryonic age was measured under ×400 magnification using the same standardized light microscopy.
The primers used for nNOS were 5′-ACCCCGT-CCTTTGAATACCAG-3′, sense; 5′-GACGCTGTTGAATC-GGACCTT-3′, antisense.
Statistical Analysis
Student's t-test was used for statistical analysis. The results were expressed as mean ± SD.
Results
Morphological Observations
At 13–15 days of gestation the gastric epithelium was thick and composed of multiple cell layers. The border between the epithelium and the underlying mesenchyme was smooth. There were many blood vessels in the mucosal connective tissue. At E15, some mesenchymal cells were observed to be dense in the future circular muscle layer area. The serosa was composed of monolayered cuboidal epithelium (Figures 1A–1C). At E13, in the gastric wall, there was no nNOS immunoreactivity in the cells in mesenchyme. At E14, solitary nNOS-IR-positive cells were first seen faintly labeled, with oval or rounded shapes, in mesenchymal tissue, and contained a voluminous nucleus surrounded by a thin ring of cytoplasm (Figure 2A). At E15, some nNOS-IR neuronal cells were situated at the outer boundary of the mesenchyme from which the circular muscle layer originates. nNOS-IR reaction product appeared to be concentrated at one pole of the cell, as at the previous stage, because most nuclei were eccentrically placed. Weak nNOS-IR staining was cytoplasmic and nuclei were unstained (Table 1).
At E16, mesenchymal cell condensation was seen to develop partly into the circular muscle. The primordium of the outer longitudinal muscular layer was evident between the presumptive circular smooth muscle layer and the serosa. Neuroblasts were first seen as small clusters of cells at this gestational period. The neuroblast groups became arranged towards the periphery of the presumptive circular muscle layer to form Auerbach's plexus. The initial feature distinguishing these groups from the surrounding mesenchyme was an increasing density of packing. The intensity of nNOS-IR staining was similar to that at E15 (Figure 1D; Table 1).
At E17, the gastric epithelium was composed of two or three layers of tall and columnar cells. At this stage the basement membrane became mildly irregular, and protruded into the mesenchyme in some mucosal regions. The height of the epithelium became irregular at E17 and onward. Cavities, which apparently corresponded to future gastric pits, developed at the epithelial surface. The ganglia forming the myenteric plexus were readily distinguished between the circular and longitudinal muscle layers (Figure 1E). Widespread nNOS-IR neuronal cells were first observed between the circular and longitudinal muscle layers. Generally, one or two nNOS-IR neurons were detected in some ganglia of myenteric plexus, and each of these cells was found in close correspondence with of groups of smooth muscle cells of the circular muscle layer. Some nNOS-IR nerve cells appeared to be monopolar and their shapes were rounded and oval. The intensity of nNOS-IR staining was similar to that observed at previous stages (Figures 2B; Table 1).
At E18 and E19, the epithelium was composed of one or two cell layers in corpus and pylorus but of three or four cell layers in forestomach. The formation of the gastric glands was first observable in the epithelium at E18. Epithelial cells were arranged radially at the base of the primitive pits, beneath which the basement membrane protruded toward the mesenchyme (Figures 1F and 1G). nNOS-IR neuronal cells were very widespread at the myenteric ganglia of corpus and pyloric regions but not in the forestomach (Figures 2C–2E). At E18–19, most nNOS-IR nerve cells were monopolar. There was some variation in the intensity of nNOS-IR staining, which depended on embryonic age and location and was higher than that at previous stages. The most intense nNOS-IR staining was first seen in the corpus at E18 and then in the forestomach at E19 (Table 1).
At E20, as the gastric glands continued to grow, the surface epithelium became progressively simple columnar. Parietal cells were first recognized in the mucosal epithelium. Neuronal cells spread rapidly as clusters of neuroblasts between the circular and longitudinal muscle layers of the stomach (Figure 1H). nNOS-IR nerve cells were bipolar in morphology. In different gastric regions, nNOS-IR neurons were oval or rounded. Most myenteric ganglia contained densely stained nNOS-IR nerve cells. In corpus and forestomach there was no variation in the intensity of nNOS-IR staining at E20 compared with the previous stage (Table 1).
At E21, all the layers of the embryonic gastric wall, which included myenteric and submucous plexuses and smooth muscle layers, developed progressively similar to those in the adult. Many parietal cells were observed in the developing gastric glands. The primitive plexus appeared to have progressively developed as the ganglia became more widely dispersed (Figure 1J). Throughout the stomach, all neurons were recognizable and many of them were nNOS-IR-positive. In some myenteric ganglia, nNOS-IR nerve cells formed large groups of three or four cells in the corpus and forestomach (Figures 2F and 2G). Most of the nNOS-IR-positive nerve cell bodies had smooth-contoured outlines. Generally, nNOS-IR nerve cells were strongly labeled with nNOS antibody in all gastric regions (Table 1).
At this stage the nerve fibers of nNOS-IR positive neurons were often found in close topographical relationship to nNOS-IR-positive and -negative neurons in the myenteric ganglion. In forestomach and corpus, nNOS-IR nerve fibers were clearly seen stretching from the myenteric ganglia to reach the submucosa and they were numerous in the circular muscle layer. nNOS-IR nerve fibers reached the luminal side of a circularly arranged muscle layer and extended to below the epithelium. nNOS-IR-stained fibers were very few in the submucosal and subepithelial areas, but numerous in the muscle layer (Figures 2F and 2G). In corpus and pylorus, nNOS-IR neuron bodies were first clearly identified at the submucosal surface of circular muscle layer, where they appeared in lesser numbers than in Auerbach's plexus, and their shapes were fusiform (Figures 2G and 2H). The longitudinal muscle contained no nNOS-IR nerve fibers in all gastric sections. Except for the myenteric plexus, there were no nNOS-IR nerve fibers in the pylorus, but the other nNOS-IR characteristics were similar to those in fore-stomach and corpus (Figure 2H).
Morphometric Measurements
In myenteric plexus of embryonic stomach, mean nNOS-IR nerve cell number did not change significantly at the early stages. At E17 the mean number of nNOS-IR nerve cells was 0.98 ± 0.15, which was a significant increase compared to the E16 age group (p<0.05). In myenteric plexus of the developing stomach, nNOS-IR nerve cell numbers increased continuously from E17 to E21. Between E17 and E21, in the embryonic stomach there were more nNOS-IR stained neurons at each microscopic area (×400). In the embryonic stages studied, the largest increase in mean number of nNOS-IR nerve cells was seen as 3.46 ± 0.23 in myenteric plexus of stomach at E21, resulting in a statistically significant increase in mean number of nNOS-IR neurons compared to the E20 age groups (p<0.05) (Table 2).
In the same way, similar increases were seen in the diameter of nNOS-IR nerve cell bodies between E14 and E21. The diameters of nNOS-IR neuron bodies varied with the embryonic ages and average soma diameters ranged from 4.58 ± 0.42 μm at E14 to 12.08 ± 1 μm at E21 (Table 2).
The tunica muscularis increased significantly in thickness from 30.4 ± 0.45 μm at E16 to 45.50 ± 0.70 μm at E18 (p<0.05), and then the thickness decreased to 34 ± 0.96 μm at E21 (p< 0.05) (Figure 3; Table 2).
Discussion
There have been few descriptions of the first appearance of nerves in the developing stomach. NADPHd-containing nerve cells could be detected on Day 12 in the embryonic mouse stomach (Branchek and Gershon 1989). In contrast to this finding, we observed the first nNOS-IR nerve cells at E14 in rat stomach. nNOS-IR-positive neurons were weakly stained at E14-E17 and more heavily stained at E18-E21. The fact that for the same incubation time as with the nNOS-IR reaction neurons in younger embryonic stages displayed a lighter intensity of nNOS-IR labeling may be explained by the presence of less enzyme, less enzyme activity, or both. It is also possible that the form of the enzyme differs at different embryonic stages (Balaskas et al. 1995). The intensity in nNOS-IR staining may vary depending on embryonic age and localization. nNOS-IR staining is usually more intense in the older embryos. This might reflect a varying degree of nNOS-IR neuronal cell maturation (Schafer et al. 1999). As summarized in Tables 1 and 2, if nNOS-IR labeling were light the counts for nNOS-IR-positive neurons would be low.
Possibly, the neuron-preferential staining is due to a specific permeability of the plasma membrane for nNOS-IR neuronal cells. In addition, the nNOS-IR-positive appearance in neuroblasts is strictly related to the onset of metabolic activities specific to neurons, such as the synaptic activities (Faussone-Pellegrini 1987). Therefore, the strictly or weakly nNOS-IR neurons may be related to the differentiating steps of the gastric plexuses. The morphologically immature nNOS-expressing immature neurons are capable of producing NO and of transmitting information, but their physiological significance can change in the course of development (Smits and Lefebvre 1996). It is quite likely that nNOS-containing neurons may have expressed NO at the earlier stages of the present study because it has been suggested by Branchek and Gershon (1989) that NADPHd-containing cells may be neuronal precursors migrating to the murine bowel.
Developing enteric neurons express a wide variety of neuroactive substances at or even before the early embryonic stage (Balaskas et al. 1995). Vasoactive intestinal polypeptide-immunoreactive neurons have been found in the chick proventriculus at E5.5–6.5 (Epstein and Poulsen 1991). Somatostatin has been detected in the chick proventriculus-gizzard junction at E4 and in the chick cecum at E6.5 (Epstein et al. 1992). Met-enkephalin- and substance P-immunoreactive neurons have been shown to be present in the chick foregut at E5 and E6, respectively (Epstein et al. 1983). The early expression of all these bioactive substances, including nNOS, suggests that there is a close time relation between the morphogenesis and the biochemical and functional maturation of gastric nerve plexus (Fekete et al. 1991).
(
In the present study, the first nNOS-IR neuroblasts normally appeared in mesenchyme at earlier stages than did the pits. Therefore, differentiation of the epithelium may depend on several inputs coming from the local mesenchyme. Some of these signals may be neuropeptides that are secreted early by neuroblasts (Ignarro 1991). Therefore, NO may be effective in cell and tissue differentiation. Once synthesized, it can migrate to lipid-soluble parts of cells, such as the membranes (Ignarro 1991). Payne and Kubes (1993) showed that NO produced by nNOS modulates epithelial permeability in the feline small intestine. Although the precise mechanism of the effects of NO is not known, some studies have suggested that the mucosal effects of NO are mediated via nerves, prostaglandin synthesis and/or via a direct effect involving the synthesis of cGMP (Tamai and Gaginella 1993).
Brown et al. (1992) reported that NO may regulate epithelial cell secretion in rat gastric mucosa. Because the cell bodies of secretomotor neurons are found in submucous ganglia (Ignarro 1991), our data suggest that NO may be a mediator in intrinsic secretomotor reflex pathways of the stomach. nNOS studies of rat myenteric plexus indicate that NOS-IR-positive nerve fibers make synaptic contacts with NOS-IR and NOS-non-IR myenteric neurons (Nichols et al. 1993). The present study reveals NOS-IR nerve cells in Henle's plexus of embryonic rat submucosa. Taken together, these findings suggest that, in addition to its neuroef-fector action on the circular smooth muscle, NO may regulate the activity of both myenteric and submucous neurons. In this regard, some investigators (Nichols et al. 1993) have reported functional evidence in support of the existence of NO interneurons in rat duodenum. Seelig et al. (1985) could not find a morphological relation between the nerve processes and glandular epithelial cells by use of electron microscopy.
There are no functional reports indicating that NO directly affects secretory processes or transmits sensory information in the gastrointestinal tract (Ginneken et al. 1998). We detected nNOS-IR nerve fibers in the gastric mucosa at E21. Therefore, they may simply follow the blood vessels and pass along the glands rather than innervating them (Ginneken et al. 1998). In the present study, nNOS polyclonal antibody stained neuronal elements of the plexus but did not give positive results in the gastric epithelium.
Distribution of nNOS immunostaining intensity due to different ages and regions in the developing rat stomach between days 13–21 of gestation a
a-, no staining; -/+, very weak staining; +, weak staining; ++, distinct; +++, intense; ++++, most intense.
Although we first observed nNOS-IR reaction product at E14, the mesenchymal condensation for future inner muscular layer was seen at E15 and for the circular muscle layer at E16. Therefore, there appears to be a considerable period of time between the formation of nNOS-IR reaction product and the development of the muscle layer, during which period nerve tissue-specific mesenchymal cell induction could take place (Seelig et al. 1985). Primarily, the establishment of cell-to-cell contacts facilitates the exchange of information and, secondly, when contact is established between a differentiated cell and a differentiating one, it is reasonable to conceive that trophic substances released by the former influence the differentiation of the target cell (Faussone-Pellegrini 1987). Therefore, it can be suggested that during the prenatal developmental stage neuronal functions might be different from those in adulthood and also that nNOS-IR reaction product can have an important role in inducing smooth muscle cells (Faussone-Pellegrini 1987).
Myectomies performed on guinea pig colon revealed myenteric nitrergic neurons to be either inter-neurons or motor neurons innervating the circular muscle layer (Barbiers et al. 1993, Jarvinen et al. 1999). In the present study, the presence of nNOS immunoreactivity in nerve fibers of the circular muscle suggests involvement of nitrergic neurons in nonadrenergic, noncholinergic relation in the developing stomach.
Seki et al. (1993) reported a circularly arranged muscle layer of rat stomach, which responded to acetylcholine stimulation at E17. In studies of the development of the murine enteric nervous system, synthesis of acetylcholine was detectable at E10-E12 and neurons were morphologically identified at E12 (Rothman and Gershon 1982). Takahashi and Ow-yang (1997) demonstrated that the accommodation reflex involves the vagal efferent pathway and nicotinic synapses, using NO released from the myenteric plexus. These data may indicate that gastric motility occurred at the embryonic developmental stage and that NO was a neurotransmitter to mediate gastric relaxation.
(
Means ± SD of number and diameter of nNOS-IR nerve cell bodies in the myenteric ganglia of developing rat stomach between days 14–21 of gestation.
∗ p<0.05.
In this study, primitive gastric pits started to form 1 day after the appearance of the circularly arranged muscle layer in mesenchyme. The gastric glands were observed on the mucosal surface 2 days after the formation of circular muscle layer, i.e., on gestational Day 18. Therefore, we consider that gastric motility is one of the important factors for epithelial growth and differentiation and that there are some relationships between the initiation of gastric motility and the formation of gastric glands.
Our measurements revealed significant increases in nNOS-IR nerve cell number from E17 onward, with no significant increase at E16, which is easily explained by the increase of the amount of the smooth muscle that must be innervated (Llewellyn-Smith et al. 1992). Prior studies indicated that smooth muscle development proceeds by the linear differentiation of the distinct smooth muscle cell phenotypes that result from the hierarchical expression of specific gene products (McHugh 1996).
Histograms of the mean thickness of tunica muscularis in developing rat stomach between Days 17 and 21 of gestation.
These observations of nNOS immunoreactivity at the embryonic stages studied demonstrate the dynamic development that occurs within the gastric nerve system at prenatal stages as chemical coding is established, reflexes develop, and neuromuscular contacts are determined.
In conclusion, it can be said that nNOS expression in the rat stomach occurs early during embryonic life. We believe that the development of the gastric wall, i.e., smooth muscle layer, epithelium, and blood vessels, may be closely associated with nNOS-IR neuronal elements. However, further studies are necessary to better understand the function of nNOS-IR neurons in the developing mammalian stomach.
Footnotes
Acknowledgements
Supported by the Research Fund of Akdeniz University, Antalya, Turkey, in partial support of the masters thesis of ZB.