Scanning electron microscopy of vascular architecture in the gastric mucosa of the golden hamster

Summary The three-dimensional architectures and the regional differences of the vascular system in the mucosa of the hamster stomach were revealed by scanning electron microscopy of corrosion casts. In the forestomach, the vascular network spreads two-dimensionally in a thin lamina propria. In the corpus and the antrum, the capillaries in the thick lamina propria are well developed, extending three-dimensionally along the gastric pits and glands. In the corpus, the submucosal arteries enter the lamina propria to become ascending capillaries, which project toward the top of the lamina propria and anastomose to create a capillary network beneath the mucosal epithelium. A subepithelial capillary is much wider in diameter than an ascending capillary and is, therefore, a sinusoid capillary. Subepithelial capillaries join descending venules, which are less numerous than the ascending capillaries. Near the gastric lumen, the capillaries in the corpus can be classified into two types: arched type in the cephalic (upper) region and honeycomb type in the caudal (lower) region. In the antrum, the submucosal arterial plexus is less well developed than that in the corpus. The mucosal aspect of the corrosion cast shows many clumps, formed by a unit of capillary network. Functional significances of different vascular architectures in the gastric mucosa of the forestomach, corpus, and antrum are discussed.This study was supported in part by grants from the Research Fund of the Ministry of Education, Science, and Culture, Japan     

Microangioarchitecture of gastric mucosa in man: antrum ventriculi

The blood vessels of the antrum area of human gastric tunica mucosa have been studied by scanning electron microscopy and compared with vertical and horizontal serial sections of perfusion-fixed antrum mucosa. In contrast to previous studies on the oxyntic corpus mucosa, where two superimposed capillary layers could be seen, the antrum mucosa presents only one capillary layer. It extends from the lymphatic follicles above the lamina muscularis mucosae up to the luminal surface. No further arterioles exist in the glandular part of the antrum mucosa. The luminal aspect of the capillaries show thin, honeycomb-like capillaries, differing from our previous findings in the corpus area. There the capillaries are wide and circulate in a dense, convoluted way. Arteriovenous anastomoses again could not be seen in the examined tissue. The differences in gastric microangioarchitecture might be explained by the extent of the glandular part in both regions. The findings in the antrum area very much resemble the basal capillary layer in the corpus mucosa. The critical height for one capillary layer in the glandular tissue of human gastric mucosa is estimated to be approximately 1 mm. Larger diameters, as in the oxyntic corpus mucosa, might require a second capillary layer.     

Microvasculature of the feline stomach

The feline gastric microvasculature was studied by corrosion of the injected vascular bed, which allowed evaluation of the vascular pattern of the different tunics. The serosal pattern consisted of numerous interconnected capillary vessels, forming a delicate network. Submucosal arteries supplied the muscular tunic through numerous anastomosing vessels that followed the direction of the smooth muscle fibers. The entire mucosal tunic was supplied by arterioles derived from the submucosal plexus; these gave rise to capillaries that surrounded the gastric glands and terminated in a diffuse, anastomosing subepithelial capillary network. Venules coursed through the mucosa in a perpendicular manner, forming submucosal veins.     

Studies on the microvascularization of the digestive tract by scanning electron microscopy of vascular corrosion casts. 1. Large intestine in rats and guinea pigs.

The colonic microvascularization of 10 adult Sprague-Dawley rats and guinea pigs (Cavia porcellus was studied by scanning electron microscopy of microvascular corrosion casts. The tunica muscularis is supplied by branches of the submucosal arteriolar plexus, according to the arrangement of muscle layers, longitudinally and transversally arranged capillaries are distinguished. The mucosal capillaries show a honeycomb pattern and mimic the openings of the mucosal glands. Parallel to the luminal aspect of the large intestine, the mucosal capillary loops often arise from the submucosal arterioles at the most abluminal aspect of the mucosa; however, some arterioles end just subjacent to these capillaries. The submucosal veins are located just subjacent to the capillaries of the lamina propria. The rat and guinea pig colonic vascular architecture revealed no differences, neither did the capillary density in different parts of the large intestine.     

Mucosal microvascular organization of the rat colon

The mucosal microvascular architecture of the rat colon is described from vascular casts and intravital microscopy. Arterial break-up into the mucosal capillary bed invariably occurs at the submucosal/mucosal interface. The mucosal capillaries drain into venules only at the opposing, luminal aspect, i.e., mucosal venules transverse the mucosa without receiving further capillary tributaries. Intravital microscopy of luminal flow confirmed cast predictions. Further, capillary flow was unusually unidirectional, i.e., rarely static or reversible. The possible functional importance of this particular microvascular architecture to water absorption in the colon is discussed.     

Blood vascular network of the rat lymph node: tridimensional studies by light and scanning electron microscopy

Many aspects of the blood vascular network of the lymph node are unknown, and others need confirmation. We have studied the blood vasculature of rat peripheral lymph nodes by means of carbon perfusion and vascular cast corrosion techniques. At the hilus of the node, an artery gives off arterioles running in medullary cords towards the cortex. Some reach the peripheral cortex directly, branching there into slender cortical vessels. Other arterioles enter the periphery of the deep cortex units, and then head towards the peripheral cortex. Upon reaching it, they curve part way above the center of the deep cortex units and provide slender branches to the overlying peripheral cortex. Dense plexuses of capillaries arise from arterioles in the medullary cords, in the periphery of the deep cortex units, and in the outermost stratum of the extrafollicular zone of the peripheral cortex. In the cortex, the draining high endothelial venules are restricted to the extrafollicular zone and to the periphery of the deep cortex units. At the cortico-medullary junction, these peculiar venules transform into regular medullary venules which form the hilar veins. In contrast, the folliculo-nodules and center of the deep cortex units are little vascularized by a loose capillary network, while no vessels occur in the subsinus layer. These features of the node vascular network are of interest in relation to the node architecture.     

Histological and ultrastructural specialization of the digestive tract of the intestinal air breather hoplosternum thoracatum (teleost)

The digestive tract of Hoplosternum thoracatum consists of an esophagus, gastric area, anterior digestive intestine with elaborate folds, digestive intestine with decreasing folds and thin, smooth-surfaced respiratory intestine. The upper tract has a mucoid columnar lining which is gently folded, whereas the gastric area has numerous pits opening into the tubular secretory glands. Striated muscle comprises the anterior muscularis but is replaced by inner circular and outer longitudinal smooth muscle layers in the gastric region. The digestive intestinal mucosa is elaborately folded, consisting of columnar cells with prominent brush borders. Mucosa, submucosa, circular and longitudinal muscularis and serosa layers are present throughout the tract. Goblet cells occur in both the digestive and respiratory intestine. Major changes that appear in the respiratory intestine are a drastic reduction in mucosa epithelial thickness and the penetration of an elaborate capillary bed into the epithelium. The other basic layers are not significantly reduced in thickness. The air-blood barrier consists of the thin epithelium, basement lamina and very thin capillary endothelium. Regional cellular composition and ultrastructural features are correlated with respective digestive and respiratory functions.     

Endometrial vascularity during pseudopregnancy in the rabbit

The vascular architecture in the rabbit uterus was studied during pseudopregnancy. Uteri at 2, 3, 4, 6, 7, 8, 9, 11, 13, 16, 17, 18, 20, 24 and 28 days after sterile mating were subjected to one of four techniques: neoprene latex casts, transparent sections, frozen sections, and histological sections. Measurements were made microscopically of the thickness of myometrium and of the subepithelial capillary plexus in the different mucosal folds. In the estrous rabbit, the circular arteries in the uterine muscular layer give off arterioles which pass upwards, with a few branchings through the endometrium to the uterine lumen. These arterioles reach the surface of the mucosal folds and break up into the subepithelial capillary plexus. This plexus is connected to the tips of the venules which run down through the endometrium to the endometrial vascular plexus at the base of the endometrium: some of the venules connect with the circular venous vessels in the muscular layer. With advanced stages of pseudopregnancy, the capillaries among the glands become stretched and elongated. Maximal branching of the folds occurs at 4 to 9 days of pseudopregnancy. The “branching activity” was consistently higher in the placental than in the periplacental or in the obplacental folds. Such changes reached a maximum at 6 to 7 days p.c., after which the capillaries became gradually shorter and tortuous. The development of arterioles in the mucosa was marked at 3 to 6 days p.c. The thickness of the plexus in the periplacental fold and in the obplacental fold as a percentage of the thickness in the placental fold was highly correlated with the stage of pseudopregnancy. At 9 to 11 days p.c., these ratios reached a minimum of 70 to 80%.     

Mild irritant prevents ethanol-induced gastric mucosal microcirculatory disturbances through actions of calcitonin gene-related peptide and PGI2 in rats

Pretreatment with a mild irritant such as 1 M NaCl prevented ethanol-induced mucosal injury, which was abolished by indomethacin, suggesting involvement of endogenous PGs. With the use of intravital microscopy, we investigated the mechanism in microcirculation whereby a mild irritant prevents ethanol-induced mucosal injury. Microcirculation of the basal part of gastric mucosa in anesthetized rats was observed through a window with transillumination. Diameters of arterioles, collecting venules, and venules were measured with an electric microscaler. One molar NaCl alone caused dilation of arterioles and constrictions of collecting venules and venules, which were inhibited by indomethacin. Ethanol (50%) applied to mucosa constricted collecting venules and venules but dilated arterioles. Constriction of collecting venules resulted in mucosal congestion. Pretreatment with 1 M NaCl inhibited ethanol-induced constrictions of collecting venules and venules, and administration of indomethacin or a calcitonin gene-related peptide (CGRP) antagonist, CGRP-(8-37), abolished elimination of constrictions. Topical application (1 nM-10 microM) of PGE2 or beraprost sodium (a PGI2 analog) to microvasculature markedly and dose-dependently dilated arterioles, whereas that of PGE2, but not beraprost, slightly constricted collecting venules. Pretreatment of microvasculature with a nonvasoactive concentration of PGE2 (100 nM) or beraprost (1 nM) completely inhibited ethanol-induced constriction of collecting venules. The inhibitory effect of beraprost but not of PGE2 was abolished by CGRP-(8-37). Present results suggest that the mechanism whereby 1 M NaCl prevents ethanol-induced injury is elimination of constrictions of collecting venules and venules by CGRP whose release may be enhanced by PGI2 but not by PGE2.     

Does Immunohistochemistry Allow Easy Detection of Lymphatics in the Optic Nerve Sheath?

We evaluated the validity of anti-D2-40 and anti-LYVE-1 (antibodies against lymphatic endothelium) for IHC diagnosis and semiquantification of lymphatic vessels in the dura mater of the intraorbital portion of the human optic nerve (ON). Fourteen specimens were analyzed using light microscopy within 12 hr postmortem. We found in all specimens that both D2-40 and LYVE-1 stained lymphatic vessels as well as venules and arterioles. Our findings show lymphatic vessels in the meninges of the intraorbital portion of the human ON. Anti-D2-40 and anti-LYVE-1 antibodies, however, are not found to be exclusively specific to the endothelial layer of lymphatics because they also stain the endothelial layer of venules and arterioles. For the unequivocal identification of lymphatics, additional morphological criteria are necessary. Nevertheless, D2-40 and LYVE-1 staining allows rapid identification of endothelial layers. (J Histochem Cytochem 56:1087–1092, 2008)     

Changing pattern of cytokeratin 7 and 20 expression from normal epithelium to intestinal metaplasia of the gastric mucosa and gastroesophageal junction

It is currently unclear whether intestinal metaplasia at the esophagogastric junction and in the distal esophagus represent a continuum of the same underlying disease process, i.e., gastroesophageal reflux, or constitute different entities with a different pathogenesis. Biopsies below the Z line might show specialized epithelium in some patients and the question is whether this is another form of short segment Barrett's esophagus or whether it is related to a generalized atrophic process of the stomach. Data from recent studies regarding the expression of cytokeratin CK7 and CK20 in intestinal metaplasia (IM) found at the gastroesophageal junction are conflicting. Prompted by these data we undertook the present study: a) to evaluate the expression of CK7 and CK20 in IM of the gastric cardia and to compare the findings with those in patients with Barrett's esophagus and IM of the gastric corpus and antrum mucosa; and b) to evaluate the immunophenotype of non-intestinalized cardiac mucosa and to compare it with that of normal gastric epithelium. We studied the expression of CK7 and CK20 on biopsy specimens from patients with long-segment Barrett's esophagus (n=17) and surgical resection and biopsy specimens of gastric cardia (n=15), corpus (n=14) and antrum (n=22) from patients with histological evidence of IM. Eighty-four biopsy specimens from 42 patients (antrum n=15, corpus n=20, cardia n=7) without evidence of IM were studied as a control group. We observed an immunophenotype characterised by diffuse moderate to strong CK7 staining on the surface and crypt epithelium combined with strong CK20 staining on the surface and superficial part of the crypts in 94.1% (16/17) of the cases with long-segment Barrett's esophagus, but in none of the 36 cases with IM in distal stomach (antrum and corpus). IM in the gastric cardia expressed the immunophenotype seen in IM of the gastric mucosa in 93.3% (14/15) of the cases. On the other hand, normal cardiac epithelium expressed patchy strong CK7 staining on the surface epithelium and on both, superficial and deep parts of the pits combined with patchy strong CK20 staining on the surface epithelium and superficial pits, a feature permitting distinction of the normal cardiac epithelium from those of the normal gastric antrum and corpus epithelium. We conclude that the expression of cytokeratins 7 and 20 can be used to distinguish the origin of IM of the gastroesophageal junction. The CK7/20 immunophenotype of IM in the gastric cardia closely resembles that of the IM in the gastric antrum and corpus and is different from IM in long-segment Barrett's esophagus. In contrast, the CK7/20 immunophenotype of the cardiac epithelium is different from that of the gastric antrum and corpus mucosa, suggesting that cardiac epithelium might not be a native normal gastric epithelium but one that is acquired as a consequence of longstanding inflammation. Changing pattern of CK7 and CK20 expression from normal to intestinalized epithelium suggests that IM arising from cardiac epithelium might have distinctive features.     

Differentiation of blood vessels in the adipose tissue of lean and obese fetal pigs, studied by differential enzyme histochemistry

Subcutaneous adipose tissue was obtained from fetuses removed from pregnant obese (Ossabaw) and lean (crossbred) sows at three stages of gestation (70, 90, and 110 days). Histochemical analysis for nucleo-side phosphatase (NPase), alkaline phosphatase (APase), and NADH tetrazoleum reductase (NADH-TR) was conducted on fresh-frozen cryostat sections. Age- associated changes in NPase and NADH-TR reactions in the arteriolar system were correlated with the morphological development of the medial layer of arterioles and arteries. For instance, a strong NPase reaction in small arterioles was associated temporally with the assumption of a normal smooth muscle cell morphology and arrangement in the medial layer. In the youngest fetuses, strong NADH-TR reactions were only evident in small and presumptive arterioles and venules (associated with fat cells). Little NADH-TR reactivity was evident in larger arterioles and venules in 70-day tissue. Arteries and large arterioles were distinguished from veins and venules (strong reactions vs. weak reactions) with NADH-TR and NPase reactions in the oldest fetuses. In the younger fetuses, the NPase distinction (arterioles vs. veinules) was obvious before NADH-TR distinction. Small adipocyte-associated vessels were APase positive in the youngest fetuses, but APase reactivity was limited to short segments of vessel between arterioles and capillaries in the oldest fetuses. With the following exceptions, all the above observations were independent of fetal strain. In obese fetuses (110 day) small venules and small arterioles were equally reactive for NPase activity. Capillaries in obese fetuses (110 day) were NADH-TR reactive, whereas no activity was evident in capillaries from lean fetuses (110 day).(ABSTRACT TRUNCATED AT 250 WORDS)     

Selective ablation of spinal afferent neurons containing CGRP attenuates gastric hyperemic response to acid.

The gastric mucosa, in particular submucosal blood vessels, are innervated by afferent neurons containing neuropeptides such as calcitonin gene-related peptide. Stimulation of sensory neurons innervating the gastric mucosa increases submucosal blood flow. Since sensory neurons supplying the stomach are of dual origin from nodose and dorsal root ganglia, we examined the effect of selective ablation of either the vagal or spinal sensory innervation to the upper gastrointestinal tract on the increase in gastric mucosal blood flow in response to acid back diffusion into the gastric mucosa. Perineural application of capsaicin to the celiac/superior mesenteric ganglia, but not to the vagus nerves, significantly inhibited by 53% the hyperemic response to acid back diffusion. Tissue levels of immunoreactive calcitonin gene-related peptide in the gastric corpus were significantly reduced (by 73%) by periceliac capsaicin treatment, but unaffected by perivagal capsaicin treatment. These data suggest that spinal capsaicin-sensitive afferents containing calcitonin gene-related peptide immunoreactivity are involved in mediating increases in gastric mucosal blood flow. This increase in gastric mucosal blood flow mediated by sensory neurons may act as a protective mechanism against mucosal injury, similar to responses seen in other tissues such as skin.     

Oxytocin is expressed by both intrinsic sensory and secretomotor neurons in the enteric nervous system of guinea pig

Single- and double-immunostaining techniques were used systematically to study the distribution pattern and neurochemical density of oxytocin-immunoreactive (-ir) neurons in the digestive tract of the guinea pig. Oxytocin immunoreactivity was distributed widely in the guinea pig gastrointestinal tract; 3%, 13%, 17%, 15%, and 10% of ganglion neurons were immunoreactive for oxytocin in the myenteric plexuses of the gastric corpus, jejunum, ileum, proximal colon, and distal colon, respectively, and 36%, 40%, 52%, and 56% of ganglion neurons were immunoreactive for oxytocin in the submucosal plexuses of the jejunum, ileum, proximal colon, and distal colon, respectively. In the myenteric plexus, oxytocin was expressed exclusively in the intrinsic enteric afferent neurons, as identified by calbindin 28 K. In the submucosal plexuses, oxytocin was expressed in non-cholinergic secretomotor neurons, as identified by vasoactive intestinal polypeptide. Oxytocin-ir nerve fibers in the inner circular muscle layer possibly arose from the myenteric oxytocin-ir neurons, and oxytocin-ir nerve fibers in the mucosa possibly arose from both the myenteric and submucosal oxytocin-ir neurons. Thus, oxytocin in the digestive tract might be involved in gastrointestinal tract motility mainly via the regulation of the inner circular muscle and the balance of the absorption and secretion of water and electrolytes.     

Histological and ultrastructural study of the digestive tract of rice field eel, Monopterus albus

The histological characteristics of the digestive tract and the ultrastructure of mucosal cells of the stomach and intestine of rice field eel, Monopterus albus, are described to provide a basis for future studies on its digestive physiology. The digestive tract of the rice field eel is a long and coiled tube composed of four layers: mucosa, lamina propria‐submucosa, muscularis and serosa. The pharynx and oesophagus mucosa is lined with a stratified epithelium. The stomach includes the cardiac and pyloric portions and the fundus. Many gastric pits are formed by invaginations of the mucosal layer and tubular gastric glands formed by the columnar cells in the fundus. The intestine is separated from the stomach by a loop valve and divided into a proximal portion and a distal portion. The proximal intestinal epithelium consists of columnar cells with microvilli towards the lumen and goblet cells. The enterocytes are joined at the apical surface by the junctional complex, including the evident desmosomas. Numerous lysosomes and some vesicles are evident in the upper cytoplasm of the cells, and a moderate amount of endoplasmic reticulum and lysosomes are scattered in the supranuclear cytoplasm. The epithelium becomes progressively thicker and the folds containing large numbers of goblet cells are fewer and shorter in the distal portion of the intestine. At the ultrastuctural level, the columnar cells of the tubular gastric glands have numerous clear vacuoles and channels. A moderate amount of pepsinogen granules are present in the stomach. The enterocytes of the intestinal mucosa display a moderate amount of endoplasmic reticulum and lysosomes, and long and regular microvilli.     

Stomach stress and strain depend on location, direction and the layered structure

The stomach is as other parts of the gastrointestinal tract functionally subjected to dimensional change. Hence, the biomechanical properties are of functional importance. Our group has previously demonstrated that the stress–strain properties of the rat and rabbit stomach wall were species-, location- and direction-dependent. We further wanted to study the anisotropic biomechanical properties of the stomach wall in pigs. Furthermore, we made an in-depth biomechanical test on the layered wall of the stomach in different regions. Two stomach strips were cut both in longitudinal direction (parallel with the greater curvature) and circumferential direction (perpendicular to the greater curvature) from the gastric fundus, corpus and antrum. One strip was used for the non-separated (intact) wall test and the other one was separated for the test on the mucosa–submucosa and muscle layers individually. The length, thickness and width of each strip were measured from digital images. The uni-axial stress and strain were computed from the force generation and the tissue strip deformation during stretching. The muscle layer was the thickest in the antrum whereas the mucosal–submucosal layer was the thickest in the corpus of the stomach (P<0.01). The strips from the corpus were stiffest among the three regions in both longitudinal and circumferential directions (P<0.001). The longitudinal strips was stiffer than the circumferential strips in all three regions (P<0.001) and the mucosa–submucosa strips was stiffer than the intact wall and the muscle layer in both directions for the fundus and the corpus (P<0.001). The constant a of the intact wall and mucosa–submucosa layer was in both directions linearly associated with the mucosa–submucosa thickness. In conclusion, the uni-axial stress–strain curves of pig stomach were location-, direction- and layer-dependent. The stiffer wall in the corpus is likely due to its thicker mucosa, i.e., the stiffness of the mucosa–submucosa layer seems can explain the intact wall stiffness. Since the structure and function of the pig stomach are similar to the human stomach, we believe that the data obtained from this study can be extended to humans. Detailed biomechanical mapping of the stomach will likely help us to understand physiological functions of the different parts of the human stomach, such as gastric accommodation and mechanosensation.     

A model of oxygen exchange between an arteriole or venule and the surrounding tissue

A mathematical model is developed to study the effect of capillary convection on oxygen transport around segments of arterioles and venules that are surrounded by capillaries. These capillaries carry unidirectional flow perpendicular to the vessel. The discrete capillary structure is distributed in a manner determined by the capillary blood flow and capillary density. A nonlinear oxyhemoglobin dissociation curve described by the Hill equation is used in the analysis. Oxygen flux from the vessel is expressed as a relationship between Sherwood and Peclet numbers, as well as other dimensionless combinations involving parameters of the capillary bed. A numerical solution is obtained with a finite difference method. The numerical results obtained within the physiological range of parameters allow the prediction of longitudinal gradients of hemoglobin-oxygen saturation along the arterioles and venules.     

Stimulant effect of nitric oxide generator and roxatidine on mucin biosynthesis of rat gastric oxyntic mucosa.

Although the involvement of nitric oxide (NO) in an increasing gastric mucus metabolism has been reported, information on whether or not its activation is limited to the specific mucus-producing cells is lacking. In this paper, we report the effect of the exogenous NO-donor, isosorbide dinitrate (ISDN), and second-generation histamine H2 receptor antagonist roxatidine (2-acetoxy-N-(3-[m-(1-piperidinylmethyl)phenoxy]propyl)acetamide hydrochloride) which is demonstrated to accelerate the mucin metabolism mediated by endogenous NO, on the mucin biosynthesis in distinct sites and layers of the rat gastric mucosa using an organ culture technique. Radiolabeled mucin was obtained from the tissue of full-thickness and the deep corpus layer, and the antrum of the rat stomach incubated for 5 hr with [3H]glucosamine(GlcN) in vitro. With the addition of ISDN to the culture medium, 3H-labeled mucin in the full-thickness corpus mucosa increased to 124-145% of the control (p<0.05), but not in the antrum. This stimulation of the mucin synthesis disappeared by the removal treatment of the surface mucous cell layer which has immunoreactivity of neuronal NO synthase. Similarly, roxatidine stimulated the mucin biosynthesis in the full-thickness corpus mucosa, but not in the gland mucous cell layer. These results suggest that the stimulation of the mucin biosynthesis mediated by NO is restricted to the surface mucous cells of the rat gastric oxyntic mucosa.     

Ultrastructural evidence of remodelling in the microvasculature of the normal rabbit and human eye

Several investigators have recently presented ultrastructural evidence for remodelling in the mammalian, including human, choriocapillaris. This evidence consists of cytoplasmic processes off endothelium and pericytes that penetrate the basal lamina of the capillary and enter the pericapillary space and redundant layers of basal lamina interpreted as the result of secretory activity subsequent to cellular movement within the wall of the capillary. This report extends these observations to the remaining microvasculature of the choroid--its arterioles and venules--and to another part of the ocular microvasculature--the pars plana of the ciliary body--of the rabbit and human eye. Cytoplasmic processes and redundant basal laminae were observed in the microvasculature at both sites, most frequently associated with venules and capillaries. Cytoplasmic processes and redundant basal laminae are generally associated with cellular movement and vessel growth during ocular neovascularization. However, their presence in histologically normal eyes suggests that these phenomena occur in the absence of neovascularization, i.e. that the microvasculature is remodelled in the normal eye.     

Gastric mucosal protection against ethanol by EP2 and EP4 signaling through the inhibition of leukotriene C4 production

Prostaglandin (PG)E derivatives are widely used for treating gastric mucosal injury. PGE receptors are classified into four subtypes, EP(1), EP(2), EP(3), and EP(4). We have tested which EP receptor subtypes participate in gastric mucosal protection against ethanol-induced gastric mucosal injury and clarified the mechanisms of such protection. The gastric mucosa of anesthetized rats was perfused at 2 ml/min with physiological saline, agonists for EP(1), EP(2), EP(3), and EP(4), or 50% ethanol, using a constant-rate pump connected to a cannula placed in the esophagus. The gastric microcirculation of the mucosal base of anesthetized rats was observed by transillumination through a window made by removal of the adventitia and muscularis externa. PGE(2) and subtype-specific EP agonists were applied to the muscularis mucosae at the window. Application of 50% ethanol dilated the mucosal arterioles and constricted the collecting venules. Collecting venule constriction by ethanol was completely inhibited by PGE(2) and by EP(2) and EP(4) agonists (100 nM) but not by an EP(1) or an EP(3) agonist. Ethanol-induced mucosal injury was also inhibited by EP(2) and EP(4) agonists. When leukotriene (LT)C(4) levels in the perfusate of the gastric mucosa were determined by ELISA, intragastric ethanol administration elevated the LTC(4) levels sixfold from the basal levels. These elevated levels were significantly (60%) reduced by both EP(2) and EP(4) agonists but not by other EP agonists. Since LTC(4) application at the window constricted collecting venules strongly, and an LTC antagonist reduced ethanol-induced mucosal injury, reductions in LTC(4) generation in response to EP(2) and EP(4) receptor signaling may be relevant to the protective action of PGE(2). The present results indicate that EP(2) and EP(4) receptor signaling inhibits ethanol-induced gastric mucosal injury through cancellation of collecting venule constriction by reducing LTC(4) production.