LectinocytochemicaL study of rat stomach mucosa under the conditions of cycLooxygenase-1/-2 bLockage and pretreatment with h-gLu-asp-gLy-oh

assessment of glycoconjugate expression on cell membranes using the lectin histochemistry technique may be a feasible approach for evaluating the functional state of the cell. The aim of this study was to evaluate carbohydrate determinants of rat stomach mucosa cell membranes under the conditions of CoX-1/2 blockage with indomethacin and pretreatment with the tripeptide h-glu-asp-gly-oh. Male Wistar rats were divided into 3 groups (n = 6 per group): 1st group (control) received vehicle; 2nd – indomethacin (35 mg/kg); 3rd – H-Glu-Asp-Gly-OH (10 μg) 30 min before indomethacin. Rats were sacrificed 24 hours later. gastric mucosa (gM) carbohydrate determinants were studied by lectin-peroxidase technique. The lectins panel included α-fucose(LABA), syalo(WGA, SNA), mannose(Con A, LCA) and galactose-specific (HPA, PNA, SBA) lectins. Intensity of lectin-receptor reaction was scored: 0 – no reaction; 1 – weak; 2 – mild; and 3 – strong reaction. CoX-1/2 blockage caused gM lesions, attenuated by h-glu-asp-gly-oh. Wga and SNa showed the highest affinity to GM. Indomethacin decreased SNA-labeling of epitheliocytes and mucocytes and laBa-labeling of chief cells. h-glu-asp-gly-oh reversed the glycosylation changes, caused by CoX1/CoX-2 blockage only in regards to labeling of chief cells with laBa, epitheliocytes and mucocytes with LCA, mucocytes with SNA. Predominantly H-Glu-Asp-Gly-OH under COX-1/COX-2 blockage had an effect opposite to indomethacin alone but glycosylation changes under these conditions differed significantly also from the control. CoX-1/CoX-2 blockage causes alteration of glycosylation processes in rat gM, mainly reduction of NeuNAc(α2-6)DGal and α-Fuc content. H-Glu-Asp-Gly-OH under the conditions of COX-1/COX-2 blockage leads to more profound changes in GM lectin-binding pattern compared to the independent effect of indomethacin and to control.


introduction
Nonsteroidal antiinflammatory drugs (NSAIDs), whose therapeutic effect is based on the inhibition of cyclooxygenase (COX) activity, are the most commonly prescribed medications worldwide for the management of pain and inflammation, however adverse effects of NSAIDs, in particular stomach mucosa toxicity, present one of the major health care challenges [1,2,[4][5][6]. Research data from the last decade provide scientific evidence that various short chain peptides may be worth attention as promising gastroprotective agents, especially for chronic NSAIDs users [2][3][4][5][6]. Short derivatives of regulatory peptides are known to be involved in the regulation of acid/base secretion in the gastrointestinal tract, motility, enzymatic activity, production of prostaglandins and nitric oxide and interaction with the neuroendocrine system [2,[4][5][6]. Important characteristics of oligopeptides, in contrast to conventionally used proton pump inhibitors or blockers of histamine-2 receptors, are lack of side effects and extremely low effective doses [2,4,5]. Oligopeptides were also reported to be resistant to the influence of gastric and enteric proteinases, which allows their oral administration to the patient [4].
Our previous studies demonstrated the gastroprotective effects of hexapeptide Arg-α-Asp-Lys-Val-Tyr-Arg and tripeptides H-Glu-Asp-Gly-OH and H-Lys-Glu-Asp-OH, that are mediated by the decrease of nitrosooxidative stress in the stomach, although the molecular mechanisms of biological effects of oligopeptides need to be elucidated [2,3].
Glycans decorate the surface of most living cells and organisms, creating a complex landscape of recognition sites and barriers that generally represent the first point of contact at the interface of a cell's biotic and abiotic environments [16]. Diseaseassociated glycosylation changes can derive from a huge number of molecular causes as glycan biosynthesis is, unlike the synthesis of DNA, RNA and proteins, a non-template-driven process involving numerous different participants [14]. Glycosylation as a common post-translational modification, occurring in most human proteins, plays crucial role in molecular recognition, cell-cell adhesion, molecular trafficking, receptor activation and signal transduction [23] Changes in glycosylation pattern was reported in many inflammatory and malignant gastrointestinal tract disorders [13,15,20,23]. In experimental studies, loss of intestinal mucin-type O-glycans was reported to promote the development of spontaneous duodenal tumors and increase susceptibility to colitis [13,15].
In recent years, the study of glycosylation processes is progressing remarkably by the development of glycan analysis systems, including histochemical methods, based on the use of lectins microarrays [7,12,24]. Lectins are carbohydrate binding proteins, other than enzymes or antibodies, which may be extracted from plants, animals and microorganisms but compared to antibodies against glycans, lectins are less expensive and have useful specificities for complex glycans [12]. Systematic analysis of the change in expression pattern of glycogens and lectins can bring about a comprehensive understanding of the genetic basis of glycobiological changes occurring in pathological condition [15].
Thus, the aim of this study was to evaluate carbohydrate determinants of rat stomach mucosa cell membranes under the conditions of COX-1/-2 blockage with indomethacin and pretreatment with the tripeptide H-Glu-Asp-Gly-OH.

material and methods
animals and experimental design. The studies were conducted on 18 male Wistar rats, aged 8 weeks, weighting 180-220 g, obtained from the Experimental Medical Laboratory of Danylo Halytsky Lviv National Medical University (Lviv, Ukraine). The studies were approved by the Institutional Ethic Committee of Danylo Halytsky Lviv National Medi cal University (Protocol No 2 of February 16 th , 2015). The experimental procedures were carried out in accordance with the international guidelines for the use and care of laboratory animals. The rats were kept in self-ventilating cages (6 animals per cage) at room temperature with light exposure following a 12/12 hour light/dark cycle. The animals were fed with concentrate feed K 120/2 for mice, rats and hamsters ("Dutch Concentrate Feed", Ukraine). All rats underwent a 24 hour fast prior to drug exposure, with water ad libitum.
To assess changes of the gastric mucosa glycoprofile under the conditions of COX-1/-2 blockage with indomethacin and pretreatment with tripeptide H-Glu-Asp-Gly-OH, the rats were randomly divided into 3 groups (n = 6 in each group). The control group received the vehicle, infused intragastrically using a probe (isotonic solution of sodium chloride). The second group was intragastrically administered indomethacin at a dose of 35 mg/kg (Sopharma, Bulgaria), dissolved in slightly alkaline distilled water [21]. The third group was pretreated with H-Glu-Asp-Gly-OH (10 µg), dissolved in saline 30 min before indomethacin introduction.
determination of gross mucosal damage. Twenty-four hours after the experimental procedures the animals, anesthetized by urethane at a dose of 1.1 mg/kg, were sacrificed by decapitation. The stomachs were rapidly excised, cut along the lesser curvature, washed in isotonic sodium chloride solution and pinned open for macroscopic examination. After gross inspection, hemorrhagic and ulcerative lesions of the gastric mucosa were counted and their lengths (mm 2 ) were measured with the method of planiometry by a researcher blind to the experimen-35 tal grouping. The lesions were scored according to macroscopic degree of injury: hyperemia (2); pellicle and small hemorrhages (4); pellicle and superficial erosion (6); several superficial erosions or one deep erosion (8); ulcer (10); and ulcer, erosions, hemorrhages (12) [2].
histologic evaluation of gastric mucosa and lectin histochemistry analysis. The pieces of stomach wall were excised, fixed in 4% neutral formalin, and embedded in paraffin wax according to the standard protocol. For general morphology studies, 5-to 7-μm-thick sections were stained with haematoxylin and eosin.
Lectin histochemistry investigations were conducted using the peroxidase-diaminobenzidine visualization protocol with diaminobenzidine tetrahydrochloride (Sigma, USA) [10]. The lectin panel is presented in Table 1. All lectins used and their peroxidase conjugates were prepared by Dr. V. Antonyuk ("Lectinotest", Ukraine). Microscopy was conducted and pictures taken using Granum R6053 photomicroscope (China), equipped with Echo-Imager 502000 (China), and the ToupView 3.7 computer program (USA). Peroxidase activity and relevantly the sites of lectin binding with glycoconjugates were determined according to the brown staining of oxidative polymerization of 3.3-diaminobenzidine product. Semi-quantitative evaluation of the staining intensity in stomach mucosa cells was performed by three independent observers, as described in the lite rature [22]. Staining intensity was classified into 4 categories: no reaction (0); weak (1); mild (2) and strong reaction (3).
Statistical analysis. Statistical analysis was performed using Origin Pro, GraphPad Prism and Microsoft Excel software. Since the data distribution in our study was normal, the mean values, standard errors and standard deviations, and upper and lower 95% confidential intervals for the means were calculated. Student's t-test was used, with differences of P < 0.05 considered to be significant. Results are expressed as means (M) ± standard error (m).

results
gross and histomorphologic assessment of the gastric mucosa under the conditions of CoX-1/-2 inhibition and pretreatment with h-glu-asp-gly-oh. In our studies the mucous membrane of control group rats was relatively intact ( Fig. 1, a), whereas COX-1/-2 inhibition with indomethacin caused the development of hyperemia, edema and abundant erosive and ulcerative lesions of the gastric mucosa with the mean area = 11.20 ± 0.65 mm 2 , and score of damage = 9.60 ± 0.61 ( Fig. 1, B, Fig. 2, a, B). Pretreatment with H-Glu-Asp-Gly-OH did not reverse the ulcerative effect of COX-1/COX-2 inhibition, but resulted in the decrease of the gastric lesion area ( Fig. 1, C, Fig. 2, a, B) and a minor trend of decreased damage score (P ˃ 0.05).
Histologic studies did not reveal pathological changes in the stomachs of the control group of rats, whereas indomethacin caused the development of  edema, loss of the surface epithelial cells, leukocyte infiltration in the submucosal layer, and decrease of the periulcerative mucosa thickness (Fig. 1, B). In the H-Glu-Asp-Gly-OH-pretreated group, better preservation of mucosal structural architecture (crypts and gastric glands) was noted as well as increased content of mucocytes ( Fig. 1, C).
Assessment of α-fucose-specific carbohydrate determinants in rat stomach mucosa. Staining with laburnum anagyroides bark agglutinin (LABA) revealed the presence of fucose-specific receptors in almost all cell components of the mucous-epithelial layer of the gastric mucosa, besides endocrinocytes (Fig. 3). The reactivity of gastric mucosa cells to LABA in control rats was as follows, mucocytes ˃ chief cells ˃ epitheliocytes ˃ parietal cells ( Table 2). Cyclooxygenase blockage with indomethacin caused a tendency to increase the α-Fuc content in mucocyes and epitheliocytes (P ˃ 0.05), whereas the number of fucose receptors decreased by 48% (P < 0.05) in chief cells and more than 2 times in parietal cells compared to control animals. Pretreatment with H-Glu-Asp-Gly-OH had a tendency to reverse the changes of α-Fuc content in chief cells (P ˃ 0.05) caused by indomethacin and showed a tendency to increase LABA-receptors in parietal cells (P ˃ 0.05), compared to the effect of cyclooxygenase (Fig. 3). But intriguingly H-Glu-Asp-Gly-OH introduction enhanced the decrease of LABA-reactivity of epitheliocytes (50%, P < 0.05) and mucocytes (50%, P < 0.05) caused by indomethacin, and in the tripeptide-pretreated group α-Fuc content in these cells was also lower compared to the control group (P < 0.05). Assessment of syalo-specific carbohydrate determinants in rat stomach mucosa. Staining with both syalo-specific lectins wheat germ agglutinin (WGA) and Sambucus nigra agglutinin (SNA) was shown to be feasible for identifying the epithelio-Gastric mucosa structural components: EP -epitheliocytes; CC -chief cells; PC -parietal cells; M -mucocytes. 1 Values are expressed as means (M) ± standard error (m). Student's t-test was used, differences of P < 0.05 were considered statistically significant. #P < 0.05, compared to control; *P < 0.05, compared to indomethacin effect

T a b l e 2. Intensity of lectin-receptor reactions in gastric mucosa cells under the conditions of cyclooxygenase-1/2 blockage and pretreatment with h-glu-asp-gly-oh in Wistar rats
Lectins and their carbohydrate specificity   cytes, chief and parietal cells as well as mucocytes of the gastric mucosa (Fig. 4, Fig. 5). Mucocytes of the control animals showed the highest reactivity to WGA among the visualized cell components of the gastric mucosa, however, other cells, apart from endocrinocytes, were also well stained (mucocytes ˃ chief cells = parietal cells ˃ epitheliocytes (Table 2).
No statistically significant changes of WGA-receptor number were revealed under the conditions of indomethacin-induced gastric lesions compared to control rats, but there was a tendency for increased WGA-reactivity in epitheliocytes (P ˃ 0.05) and decreased WGA-reactivity in chief, parietal cells and mucocytes (P ˃ 0.05, Fig. 4). Pretreatment with H-Glu-Asp-Gly-OH resulted in increased numbers of WGA-receptors in mucocytes, chief cells and epitheliocytes, that were decreased by cyclooxygenase blockage. However, the effect of the tripeptide enhanced the decreased WGA-staining of parietal cells, caused by indomethacin, so that the reactivity of these cells in the H-Glu-Asp-Gly-OH-pretreated group was 33% lower (P < 0.05) compared to control rats according to semi-quantitative evaluation.
Concerning SNA-labeling of the mucousepithelial barrier of the stomach, the reactivity of the cells was epitheliocytes ˃ mucocytes ˃ parietal cells ˃ chief cells (Fig. 5). Under the conditions of cyclooxygenase blockage with indomethacin, NeuNAc(α2-6)DGal content decreased by 22.2% (P < 0.05) in epitheliocytes and by 23.5% (P < 0.05) in mucocytes compared to control. In chief cells, SNA-labeling slightly increased (P ˃ 0.05) but did not change in parietal cells, in contrast to control animals. Pretreatment with H-Glu-Asp-Gly-OH reversed indomethacin-induced alterations of mu-cocyte carbohydrate determinants (P < 0.05) and showed a tendency to increase syalo-specific receptors (labeled by SNA) in epitheliocytes, compared to the effect of cyclooxygenase-1/2 blockage (P ˃ 0.05, Fig. 5). Administration of H-Glu-Asp-Gly-OH did not affect the staining profile of chief cells compared to indomethacin, but decreased SNA-reactivity of parietal cells almost 2 times compared to indomethacin-treated and control rats (P ˃ 0.05). As mentioned before, NeuNAc(α2-6)DGal content in parietal cells of control rats and animals with gastric lesions was considered to be the same, based on the evaluation of the lectin-receptor reaction.
assessment of mannose-specific carbohydrate determinants in rat stomach mucosa. Among mannose-specific lectins in our studies, the gastric mucosa of experimental rats showed stronger reactivity to lens culinaris agglutinin (LCA) than to Canavalia ensiformis agglutinin (Con A). Con A labeling in all groups of animals was restricted to only epitheliocytes, and a very slight reactivity to this lectin was marked in 2 animals under the conditions of ulceration, that anyway was supposed to be statistically insignificant for calculations ( Table 2). Cyclooxygenase blockage with indomethacin caused 25% reduction of αDMan content in epitheliocytes compared to control and the loss of ConA-receptors was enhanced by the effect of H-Glu-Asp-Gly-OH on the background of indomethacin.
The reactivity of gastric mucosa cells to LCA was epitheliocytes ˃ mucocytes ˃ chief cells ˃ parietal cells. Blockage of cyclooxygenase-1/2 with indomethacin tended to increase LCA-labeling in epitheliocytes (P ˃ 0.05) and to decrease labeling in chief cells, parietal cells, and mucocytes (P ˃ 0.05).

A B C A B C
In the H-Glu-Asp-Gly-OH-pretreated group the number of LCA-receptors decreased by 28% (P < 0.05) in epitheliocytes and increased by 64.3% in mucocytes (P < 0.05) compared to indomethacin alone, approaching the glycome of control rats. However, the noted loss of parietal cell and chief cell reactivity to LCA under the conditions of cyclooxygenase-1/-2 blockage was enhanced by pretreatment with H-Glu-Asp-Gly-OH. LCA-labeling of chief cells under the conditions of H-Glu-Asp-Gly-OH administration on the background of indomethacin was 6 times lower compared to control rats (P < 0.05). Assessment of galactose-specific carbohydrate determinants in rat stomach mucosa. Three galactose-specific lectins were used to assess the changes of the gastric mucosa glycopropfile under the conditions of cyclooxygenase inhibition and pretreatment with the tripeptide H-Glu-Asp-Gly-OH. The reactivi ty of gastric mucosa cell constituents to helix pomatia agglutinin (HPA) in control rats was mucocytes ˃ parietal cells ˃ chief cells ˃ epitheliocytes ( Table 2). Development of indomethacin-induced gastric lesions was accompanied by decreased HPAlabeling of epitheliocytes, but increased reactivity of parietal cells and mucocytes to this lectin (Fig. 6). Pretreatment with H-Glu-Asp-Gly-OH caused changes in the lectin-binding pattern of the gastric mucosa compared to the effect of indomethacin, in particular the statistically significant increase of HPA-receptors in mucocytes (P < 0.05) was noted and this difference was also detected versus the control group of animals (P < 0.05). Introduction of H-Glu-Asp-Gly-OH simultaneously with cyclooxygenase inhibition with indomethacin also resulted in a 79% increase of HPA-labeling of chief cells compared to indomethacin alone and the control group (Fig. 6). Pretreatment with H-Glu-Asp-Gly-OH did not affect the lectin-receptor reactions in epitheliocytes compared to the effect of indomethacin.
Reactivity of gastric mucosa cells to peanut agglutinin (PNA) was parietal cells = mucocytes ˃ chief cells ˃ epitheliocytes. Blockage of cyclooxygenase with indomethacin did not affect the glycoprofile of epitheliocytes but led to a tendency for decreased PNA-labeling of chief cells (P ˃ 0.05) and mucocytes (P ˃ 0.05) and a tendency toward increased parietal cell reactivity (P ˃ 0.05). The effect of H-Glu-Asp-Gly-OH on PNA-labeling of all cells, apart from chief cells, caused significant changes in glycosylation processes compared to the effect of indomethacin alone but this dynamic was not consistent with the glycoprofile of control rats. Thus, pretreatment with H-Glu-Asp-Gly-OH led to an almost 3-times decrease in PNA-labeling of parietal cells versus control (P < 0.05), statistically significant enhancement (P < 0.05) of mucocyte staining compared to control, and an almost 4-times increase in epitheliocyte reactivity to this lectin.
The sensitivity of gastric mucosa cells to soybean agglutinin (SBA) was epitheliocytes = mucocytes ˃ chief cells = parietal cells. Gastric ulceration, caused by cyclooxygenase activity blockage with indomethacin, was accompanied by 50% increased SBA-labeling of parietal cells (P < 0.05) and by 62.5% of mucocytes (P < 0.05). The tendency for a decrease in numbers of SBA-receptors in epitheliocytes and chief cells was noted in rats exposed to the ulcerative effect of indomethacin. Pretreatment with the tripeptide H-Glu-Asp-Gly-OH did not affect the glycoprofile of mucocytes compared to the effect of indomethacin alone, but decreased SBA-labeling of parietal cells, chief cells and epitheliocytes approached the glycome of control rats.

discussion
This study provides the first description of the rat stomach mucosa glycoprofile in health and under the conditions of COX-1/COX-2 blockage with indomethacin according to the staining with a comprehensive set of lectins with 4 different carbohydrate specificities (α-fucose-, syalo-, mannose-and galac-tose). The changes of glycosylation processes in rat stomach mucosa under the conditions of tripeptide H-Glu-Asp-Gly-OH administration on the background of COX-1/COX-2 blockage are also reported.
COX-1 is a constitutive enzyme, catalyzing the reaction of prostaglandin production [2]. Prostaglandins are known to play a crucial role in the processes of gastric mucosa integrity maintenance, including contributing to the decrease of hydrogen chloride production in stomach in response to stimuli, the increase of bicarbonate secretion, production of the protective mucus in the gastrointestinal tract, phospholipid biosynthesis, and enhanced blood flow [1,2]. Gastrointestinal complications are probably the most critical adverse effects of NSAIDs and the indomethacin model of experimental gastric lesion induction is well described in the literature [1,2,5]. In our studies nonselective cyclooxygenase-1/2 blockage with indomethacin also resulted in extensive gastric mucosal damage. Pretreatment with a tripeptide, containing lysine, glutamate and asparaginate provided a mild cytoprotective effect in idomethacin-induced gastric lesions in rats. According to data reported in the literature, the tripeptide H-Glu-Asp-Gly-OH stimulates cell proliferation and protein biosynthesis, and significantly inhibits apoptosis in fibrobasts, increasing resistance of their mitochondria to damaging agents, as revealed in cultured human stomach epitheliocytes [11]. Short peptides were reported to affect the processes of gene expression and H-Glu-Asp-Gly-OH was shown to decrease the synthesis of the proinflammatory enzyme COX-2 along with TNF-α [11,17]. Taking into account that COX-2 is an inducible enzyme, upregulated in inflammation, that leads to overproduction of prostaglandin E2, a well established mediator of pain and inflammation [1,2,5], it may be presumed that COX-2 inhibition with H-Glu-Asp-Gly-OH is one of the explanations for the mild gastroprotective effect of this tripeptide. Attenuation of experimental gastric mucosa damage in rats under the conditions of selective COX-2 inhibition was reported in our earlier studies and research of other authors [2].
In the model of cistamine-HCl-induced gastric ulceration on the background of helicobacter pylori infection it was shown that subcutaneous administration of H-Glu-Asp-Gly-OH during 5 days led to the decrease of gastric lesions and inflammatory changes around the ulcerative defect as well as enhanced ulcer healing [15]. Both our previous studies and data in the literature showed the decrease of expression and activity of NO-synthase, lipid peroxidation processes, level of heat shock protein (HSP) 70 and p65 in experimental gastric lesions in animals under the effect of H-Glu-Asp-Gly-OH [3,11]. Stem cells, capable of differentiation to other cells, and intracellular signal cascades have also been reported to be some of the targets for the effects of the short peptides [11,17].
In our studies all 8 used lectins labeled the mucous-epithelial barrier of the gastric mucosa, although other scientists have reported the lack of LABA-, PNA-and SNA-labeling of healthy rat stomach mucosa [18]. However, in our research the cell reactivity to Con A (αDMan>αDGlс) was restricted to epitheliocyte labeling that is concordant with the data of other researchers [19]. Syalo-specific lectins WGA (DGlcNAc>NeuNAc) and SNA (NeuNAc(α2-6)DGal) showed the highest affinity to gastric mucosa in 3 experimental groups of rats, compared to other lectins (α-fucose-, mannose-and galactose-specific) and reactivity to WGA was higher in contrast to SNA. In a study by Falalyeyeva et al (2011) WGA, specific to NAcDGlc→NAcNeu, was also shown to exert the highest affinity to stomach mucosa in rats [18]. Development of indomethacininduced ulceration of gastric mucosa caused the statistically significant decrease of SNA-labeling of epitheliocytes and mucocytes that is indicative of the decrease of NeuNAc(α2-6)DGal content in tissue and LABA-labeling of chief cells, mediated by the diminished α-Fuc level. Roy et al. (2014) reported the increased reactivity of human gastric mucosa to WGA in gastritis [20] but our studies in rats showed only increased WGA-reactivity in stomach mucosa epitheliocytes.
Interestingly, introduction of the tripeptide H-Glu-Asp-Gly-OH on the background of COX-1/-2 blockage had a more profound effect on glycosylation processes in gastric mucosa compared to indomethacin alone. Pretreatment with H-Glu-Asp-Gly-OH in only a few cases (LABA-labeling of chief cells, LCA-labeling of epitheliocytes and mucocytes, SNA-labeling of mucocytes) reversed the glycosylation changes, caused by COX blockage with indomethacin. In most cases this tripeptide on the background of indomethacin administration had an effect opposite to indomethacin alone but glycosylation changes under these conditions differed significantly also from the control group. These changes may also be explained by masking of relevant carbohydrate determinants of stomach mucosa by other glycans.
A limitation of this study is the lack of data on the independent effect of the tripeptide H-Glu-Asp-Gly-OH on glycosylation processes in gastric mucosa.