Glycine reduces tissue lipid peroxidation in hypoxia-reoxygenation-induced necrotizing enterocolitis in rats 1

Purpose:To assess the protective effect of glycine in an experimental model of Neonatal Necrotizing Enterocolitis (NEC). Methods: Fifty (50) neonatal Wistar rats, from a litter of six female rats and weighing 4 to 6 grams, were used. Five animals were cannibalized and the 45 remaining were distributed into three groups: the G1 normal control group (n=12); the G2 Group (n=16), of animals that underwent hypoxia-reoxygenation (HR); the G3 Group of animals (n=17) that underwent HR following a 5% intraperitoneal glycine infusion. The animals underwent hypoxia in a CO2 chamber receiving an air flow of 100% CO2 for 5 minutes and reoxygenation receiving an O2 flow at 100% for 5 minutes. One centimeter long small bowel and colon segments were prepared for histological analysis. The rest of the bowel was removed in a block and frozen at minus 80°C for homogenization and determination of tissue malondialdehyde (MDA). Tissue lesions were classified as Grade 0 to Grade 5, according to the level of damaged mucosa. Results: The animals in Group G1 had levels of small bowel and colon lesion significantly smaller as compared to the animals in Groups G2 and G3. The G2 group had mean MDA values significantly higher than the animals in the G1 (p = .015) and G3 (p=0.021) groups. MDA values did not differ significantly (p = 0.992) for the animals in groups G1 and G3. Conclusion: Glycine reduces tissue MDA levels (a measurement of lipid peroxidation) following HR in neonatal rats.


Introduction
Despite significant advances in the care provided to high risk newborns, necrotizing enterocolitis (NEC) continues to be the most important cause of mortality and morbidity in low-birth-weight infants.NEC is the most common gastrointestinal emergency in neonatal intensive care units; approximately 5% of low-birth-weight infants end up by developing NEC 1 .Most newborns that develop NEC receive parenteral nutrition and extended spectrum antibiotics; surgical indication is restricted to quite frequent complications, aiming in particular to act before intestinal perforations, responsible for the terrible evolution in those children.In such cases, surgical mortality can come to 45% 2 .
The children that survive major resections of large portions of the bowel develop the well known short-bowel syndrome, whose morbidity in the end definitively compromises their weight and height development, with implications also in their neurological, psychic and motor development 2 .Different hypotheses have been formulated to explain the origin of NEC.However it remains as a disease of unknown etiology and uncertain pathogenesis.Management is still based on empirical observations and no prevention method had shown to be fully successful 1 .The combination of gastrointestinal ischemia, which is the outcome of a redistribution of the splanchnic blood flow to vital organs such as the heart and the brain, enteral nutrition that possibly predisposes to lesion of the mucosal layer, and the translocation of pathogenic bacteria, associated with immature immunitary mechanisms of the bowel, would be involved in the development of the disease 1 .
Many animal models have been used in the assessment of hypoxia effects on the origin of NEC and shown that the reduced intestinal perfusion as well as tissue lesion would be caused by hypoxia-reoxygenation (HR) mechanism [3][4][5][6][7][8][9][10][11][12] .Thus, several agents were used in an attempt to prevent intestinal lesions in NEC such as: mother's milk 3 , vitamin E 9 , supplementation with bifidobacteria 11 and recombinant human interleukin-10 12 .Other agents were used seeking to attenuate the injuries caused by ischemia-reperfusion without, however, using NEC as an experimental model.Such was the case of: glucagon 13 , 21-amino-steroids 14 , pentoxifylline 15 , somatostatin 16 and perfluorocarbons 17 .More recently, by using a model of HR in rats, glycine was put to the test as a protective agent of intestinal injuries stemming there from [18][19][20] .Glycine is a non-essential simple amino acid formed by a carbon molecule bonded to an amino group and to a carboxyl group.On the plasmatic membrane, it brings into action a chlorine channel that either stabilizes or hyperpolarizes the membrane's potential.As a result, it blocks the additional intracellular calcium triggering the cytokines formation cascade 21 .Thus glycine has an anti-inflammatory, immune-modulator and cytoprotective effect.It protects against the shock caused by both hemorrhage and endotoxins, it prevents ischemiareperfusion-associated injury in several organs and tissues such as the liver, kidney, heart, bowel and skeletal muscle and it reduces renal and liver injury caused by drugs 21 .Considering the anti-inflammatory, immune-modulator and cytoprotective effects of glycine, we sought to assess its capacity to prevent intestinal injuries caused by hypoxia-reoxygenation in a previously described experimental model of NEC 9 .

Experimental design
The animal experiment was approved by the Bioethical Committee in Research from Federal University of São Paulo (UNIFESP) under number 0560/04.Fifty (50) neonatal OUT B EPM-1 Wistar rats (Rattus norvegicus albinus, Rodentia mamalia), from a litter of six female rats and weighing 4 to 6 grams, were used.Five animals were cannibalized and, the forty-five remaining ones were randomized into three groups: the normal control G1 Group (n=12); the G2 Group (n=16), with animals undergoing hypoxia-reoxygenation (HR); the G3 Group (n=17), with animals undergoing HR following a glycine intraperitoneal infusion.The animals underwent hypoxia in a rodents CO 2 kill chamber where they received an air flow containing 100% CO 2, for 5 minutes.Following hypoxia the animals were resuscitated with an air flow containing O 2 at 100% for 5 minutes and then kept close to their respective mothers in a normothermic environment [9].All animals were given breast milk before and after the procedure..The glycine injection was administered 30 minutes prior to hypoxia-reoxygenation and it was maintained once a day until the animals' euthanasia.All animals underwent euthanasia by decapitation on day four of life.

Intestinal fixation and tissue processing
One centimeter-long small bowel (ileum) and colon samples were prepared for histological analysis.The rest of the bowel was then removed in a block and immediately frozen at minus 80°C for subsequent homogenization and measurement of tissue Malondialdehyde (MDA).

Histopathological examination
The one centimeter specimens were fixed in 10% formalin, dehydrated, embedded in paraffin and stained with hematoxylin and eosin.Six micrometer sections were made in each specimen, analyzed under optical microscopy and graded according to the extent of the tissue lesion, according to Chiu et al 22 , as: Grade 0: mucosa with no alterations (Figure 1).Grade 1: well formed villi, with no cell lyses or inflammatory process, however with formation of Grunhagen's subepithelial space (Figure 2).Grade 2: presence of cell lyses, formation of Grunhagen subepithelial space and increased spacing between villi (Figure 3).Grade 3: destruction of the villi free portion, presence of dilated capillary and inflammatory cells (Figure 4).Grade 4: structural destruction of villi, which in some cases appeared only roughly and formed by inflammatory cells and necrotic material, with bleeding and basal glandular ulceration (Figure 5).Grade 5: destruction of the entire mucosal tunic, with no glandular structure found, but only amorphous material deposited on the submucosal layer (Figure 6).

Determination of malondialdehyde
Malondihaldeyde (MDA) is a final product of lipid peroxidation and a well established parameter to determine the increase of free radicals in the bowel tissue 12 .In order to determine MDA levels the thiobarbituric acid method (TBA), proposed by Kohn and Liversedge 23 was used and the levels were expressed in nmol MDA/mg of protein.The protein content of the homogenate was determined by the "coomassie brilliant blue" (CBB) procedure.Tissue samples were defrosted, duly weighed, and a volume equivalent to 5 times the weight of TRIS 0,01M/pH 7.4 buffer solution was then added.Tissue samples were homogenized in ice bath 4 times, for 30 seconds each and, subsequently, centrifuged for 5 minutes at 10.000 rpm at 4°C.

Protein measurement
The CBB reactant interacts with protein enabling its quantification by using a standard albumin curve with known concentrations.
Preparation of the CBB reactant CBB 250G, 100 mg, was dissolved in 50 ml of 95% ethanol.100 ml of 85% phosphoric acid was added and constantly stirred.Distilled water for a final 1 liter volume was added.The reactant was let to rest for 24 hours and then it was filtered and kept in a dark vial.By using a bovine serum albumin (BSA) 10 mg/ml stock solution, we prepared 200 ìl of the following solutions: 0.1 mg/ml, 0.2 mg/ml, 0.4 mg/ml, 0.8 mg/ml and 1 mg/ml (bovine serum albumin/ml of water) to have the bovine serum albumin standard curve (standard curve for protein measurement).Homogenate 25 ìl and 50 ìl was collected, and then diluted in 5 times a TRIS 0,01M/pH 7.4 buffer; 2.5 ml of the CBB was added and readings were taken at 595 nm, 10 min after the CBB reactant was added.

Malondialdehyde measurement
Four hundred (400 ìl) of the centrifuged homogenate supernatant was collected and the following was added to it: 1 ml of 20% trichloroacetic acid and 400 ìl of 1.6% thiobarbituric acid, mixture that was incubated for 30 minutes at 95°C.Lipids were extracted by adding n-butanol (1.6 ml) and stirring vigorously.The sample was again centrifuged for 10 minutes at 3000 rpm.Absorbance of the organic layer was determined through reading at 510, 532, and 560 nm.The following equation proposed to minimize the interference of both heme pigments and hemoglobin in the measurement of MDA 25 was used : The calibration curve was drawn with 1,3,3 tetramethoxypropane malondialdehyde bis (acetyl-dimethyl).

Statistical analysis
The quantitative variables were represented by mean, standard error (SE), minimum and maximum and the qualitative variables were represented by absolute (n) and relative (%) frequencies.The Kolmogorov-Smirnov test was applied to test the normal distribution of the parameters.One Way Analysis of Variance (ANOVA) was performed to compare the study groups as for nmol/mg of protein and nmol/mg of tissue.The differences were found through the Tukey multiple comparison test.The Kruskal-Wallis test was used in analyzing the extent of tissue injury in the three sites assessed.The differences were found under the Dunn multiple comparison test.We set p < 0.05 as the significance level for all the tests, represented by *.

Histopathology
The animals in the G1 Group showed ileum lesion scores significantly lower as compared to the animals in the G2 and G3 Groups that showed no significant differences among one another as can be seen in Table 1.The G1 Group animals had colon lesion scores significantly lower as compared to the G2 and G3 Groups that did not differ significantly among one another as can be seen in Table 1.
neonatal dogs, promoted intestinal ischemic necrosis by producing hypoxia with 10% O 2 for 2 hours.Cohen et al 7 exposed neonatal pigs to a partial 50% O 2 pressure, achieving microscopic and macroscopic evidence of NEC.Caplan et al 8 described a NEC model by feeding the animals on an artificial formula and asphyxiating them for 60 seconds with 100% Nitrogen twice a day followed by cold exposure.Other investigators were also able to show ischemic lesion of the intestinal mucosa in a model similar to the one used in this paper 9,10,12 .
One of the reasons accepted to explain hypoxiaassociated lesions would be that neonatal asphyxia would lead to a redistribution of the blood flow by triggering splanchnic vasoconstriction, diverting the flow to vital organs such as the heart and the brain and causing intestinal ischemia 5 .Several mechanisms are involved in the onset and progression of this ischemia-associated lesion, such as: increased production of hyper reactant peroxides, increased synthesis of adhesion molecules with neutrophils infiltration, increased lipid peroxidation and increased production of inflammatory mediators such as cytokines 25 .Several techniques have been used in an attempt to use NEC attenuating or preventive agents.Mother's milk, unlike artificial milk formulas, could prevent the development of NEC 3 .Vitamin E, in the 30 IU/Kg/day dose 9 and recombinant human interleukin, in the 75 µg/Kg dose 12 could reduce lipid peroxidation and histopathological lesion in the intestinal tissue of neonate rats.Supplementation with bifidobacteria could reduce the risk of NEC by reducing the level of plasma endotoxins and intestinal phospholipase A 2

11
. The use of nitric oxide has already been tested in NEC models also, resulting in reduced intestinal inflammatory reaction 26 .Agents such as glutamine and dexamethasone also showed some effect by reducing intestinal inflammatory reaction 27 .
Glycine is a non-essential amino acid that protects the gut against lesions caused by the ischemia-reperfusion phenomenon [18][19][20]28 . It s considered an anti-inflammatory, immune modulator agent of direct cytoprotective function 21 .Lee et al 19 , in a model of intestinal ischemia and reperfusion, showed that local 20% glycine mesenteric intravenous infusion increased mucosal protein and DNA content, reduced the intestinal myeloperoxidase activity and maintained glutaminase activity in the mucosa.Another two studies 19,20 , also in an intestinal HR model in rats, showed the protective effect of glycine used in systemic intravenous infusion, by reducing the apoptosis cascade 19 and preserving the integrity and contractility of the intestinal wall 20 .In the present study, the rats received 1,7 a 2,2 mg glycine per gram body weight.There are reports that recognized the dose to prevent intestinal lesions due ischemia-reperfusion is 0,5-3 mg glycine per gram body weight [18][19][20] .
Lipid peroxidation is a complex process that can occur in biological membranes that are made up of molecular oxygen-reactant polyunsaturated fatty acids, leading to the production of lipid hydroperoxides and their metabolites.Most cases involving lipid peroxidation start from a chain reaction that spreads out, mediated by the presence of free

Malondialdehyde
The groups' MDA levels are shown in Table 2.
A statistically significant difference was found among the groups studied in regard to Malondialdehyde mean values (p = .002) The G2 Group had malondialdehyde mean values significantly higher in relation to the G1 Group (p = 0.015) and also higher than the animals in the G3 Group (p=0.021).
The animals in the G1 and G3 Groups showed levels of no significant difference (p = 0.992)

Discussion
Neonatal necrotizing enterocolitis is the most frequent and lethal disease affecting the gastrointestinal tract of preterm infants.Although the etiology of NEC has not been well defined yet, hypoxia definitely plays an important role in its pathogenesis [3][4][5][6][7][8][9][10][11][12] .Several animal models have been proposed to show the relevance of hypoxia in the development of NEC-associated lesions [3][4][5][6][7][8][9][10][11][12] .Barlow and Santulli 3,5 promoted intestinal lesion in neonatal rats fed on artificial milk, by putting the animals in a plastic bag for 3 to 5 minutes a day.Harrison et al 4 showed early alterations in the intestinal mucosa and in the endothelial capillary cells under electronic microscopy in neonatal dogs exposed to 7% O 2 hypoxia for two hours.Hansbrough et al 6 , using radicals.Lipid hydroperoxides accumulate in the membrane inactivating its receptors and enzymes, affecting its functions, leading it to be unstable and making it permeable to ions.A simple method of high sensitivity, very much used as a lipid peroxidation marker, involves thiobarbituric acid-reactive substances, such as lipid hydroperoxide derivatives.Hence, Malondihaldeyde (MDA) is a quite adequate indicator of lipid peroxidation caused by free radicals 10 .
By using the NEC model proposed by Okur et al 9 , we sought to assess the protective effect of glycine used through the intraperitoneal route.The model used in the experiment proved to be useful, easily reproduced, requiring the administration of no enteral formulae.An important limitation that had already been previously shown 10 , however, was the slight histological alteration found in the different groups.According to Kallakuri et al 20 involvement of the lesion beyond the mucosal layer would be necessary for the protective effect of glycine to be possibly assessed.Özkan et al 29 proposed, by using the same model, to have the animals undergoing repeated periods of HR, which would cause more marked intestinal lesions.In our experiment the animals undergoing HR, not previously taking glycine infusion, tended to show histological lesions more important than the animals in the other two groups: no injury to very mild injury (grade 0-1) was 12.6% in the HR group but was 41% in glycine-treated group.Moderate injury (grade 2-3) was ~63% in the HR group but was 47% in glycine-treated group.Severe injury (grade 4-5) was 25% in the HR group but was decreased to 12% in glycine-treated group.Similar phenomenon or even more obvious tendency is observed in the colon (severe injury was decreased from 25% to 6%), even though no statistical difference was shown among the groups.Our findings were similar to the ones reported by Lee et al 18 , where glycine did not change histological lesions but diminished intestinal myeloperoxidase activity and maintained mucosal glutaminase activity.
Kallakuri et al 20 showed that glycine did not prevent the mucosal cover decrease, but kept the full thickness of the wall and the circumference radius of the villi.MDA levels, however, showed that glycine has been able to prevent lipid peroxidation.The group undergoing HR had mean MDA values significantly higher than those in the group undergoing HR despite previously protected by the use of glycine (p = .021).The absence of difference between the control group and the group that used glycine (p= 0.992) showed that the level of protection provided by glycine was so important that it provided a peroxidation level similar to that of normal control rats.Intestinal lesion in NEC not only causes damage to the organ affected, but also, by triggering the release of inflammatory mediators into the blood stream, but it can also lead to the dysfunction and failure of multiple organs 30 .
It remains to be known to what extent such findings can actually benefit infants with NEC, and finding the dose capable of changing the history of this disease that still claims for so many lives among low birth weight infants.

Conclusion
Glycine reduces tissue malondialdehyde (MDA) levels (a measurement of lipid peroxidation) following hypoxiareoxygenation (HR) in neonatal rats.

FIGURE 4 -
FIGURE 4 -Microscopic appearance of the ileum from a rat showing destruction of the villi free portion, presence of dilated capillary and inflammatory cells

FIGURE 1 -FIGURE 2 -
FIGURE 1 -Microscopic appearance of the ileum from a rat showing mucosa with no alterations

FIGURE 5 -FIGURE 6 -
FIGURE 5 -Microscopic appearance of the ileum from a rat showing structural destruction of villi, which in some cases appeared only roughly and formed by inflammatory cells and necrotic material, with bleeding and basal glandular ulceration FIGURE 6 -Microscopic appearance of the ileum from a rat showing destruction of the entire mucosal tunic, with no glandular structure found, but only amorphous material deposited on the submucosal layer.

TABLE 1 -
Scores of tissue lesion of the Ileum and colon