The inflammatory bowel diseases (IBD),⁵ Crohn's disease and ulcerative colitis (UC), are multifactorial disorders whose etiology remains unknown. Evidence supports a pathogenic role of arachidonic acid (AA)-derived inflammatory mediators within the gastrointestinal tract. Two groups of AA metabolites exist that may play a role as mediators of colonic inflammation: cyclooxygenase products (primarily prostaglandins) and lipoxygenase products (Brasitus 1983). Vilaseca et al. (1990) demonstrated participation of prostaglandin E₂ (PGE₂), 6-keto-prostaglandin F₁ (6-keto-PGF₁), leukotriene B₄ (LTB₄) and thromboxane B₂ (TXB₂) in the development of chronic lesions in an animal model of inflammatory colitis in rats. One mechanism of inflammatory modulation is through the competition of polyunsaturated fatty acids (PUFA) in fish oil and AA [20:4(n-6)] as precursors of eicosanoid synthesis. The PUFA present in fish oil are eicosapentaenoic acid [20:5(n-3); EPA] and docosahexaenoic acid [22:6(n-3); DHA]. Eicosapentaenoic acid and DHA compete with AA for incorporation into membrane phospholipids, resulting in increased proportions of EPA and DHA in plasma phospholipids and red blood cell (RBC) membrane phospholipids (von Schacky et al. 1985). The increased incorporation of EPA and DHA leads to increased production of 3-series prostaglandins and thromboxanes (PGE₃ and TXB₃) and 5-series leukotrienes (LTB₅) which have reduced inflammatory potential. Thus, the inflammatory response modulated via prostaglandins, leukotrienes and thromboxanes may be diminished by the action of dietary (n-3) fatty acids on eicosanoid biosynthesis.

Short-chain fatty acids (SCFA), produced from bacterial fermentation, also may influence the overall health of colonic mucosa in various colonic disorders. It is unlikely that SCFA added directly to nutritional formulas will reach the large bowel. Thus, it may be advantageous to provide fermentable fiber or oligosaccharides to supply SCFA to the large bowel. Potential substrates include fructooligosaccharides (FOS), xylooligosaccharides (XOS) and gum arabic. Fructooligosaccharides and XOS have been implicated in increasing the densities of bifidobacteria (Hidaka et al. 1991, Okazaki et al. 1990) in the gastrointestinal tract. Bifidobacterium species (spp.) produce acetic and lactic acids during fermentation of FOS and XOS, resulting in a lower colonic pH which prevents enteric colonization of potentially pathogenic microorganisms (Gibson and Roberfroid 1995). Moreover, oligosaccharides have been implicated in increasing beneficial bacteria (i.e., Bifidobacterium spp.) that alter luminal pH leading to inhibition of potential pathogens, stimulation of immune response and restoration of normal intestinal flora during antibiotic therapy (Gibson and Roberfroid 1995, Mitsuoka 1990). Thus, addition of fish oil and oligosaccharides to enteral formulas may be advantageous to gastrointestinal health via modulation of the inflammatory response at the mucosal level while maintaining integrity of the intestinal tract.

As a result of greater interest in providing nutritional management to patients with IBD, it is important to evaluate nutritional formulas and their potential impact on lipid profiles of phospholipids and eicosanoid formation. Many difficulties exist in conducting these experiments in human subjects; thus, utilization of an animal model is warranted. Use of swine in biomedical research has been growing steadily. Miller and Ullrey (1987) described many advantages of using the pig as a model for the human. Pigs were used in this study to determine the effects of fish oil, FOS, XOS, gum arabic and antioxidants as supplemental ingredients in an enteral formula. The objective was to evaluate and assess changes in plasma, RBC, and colonic phospholipid fatty acid and prostaglandin profiles in pigs fed an ulcerative colitis nutritional formula (UCNF) compared with a similar formula minus these ingredients.

MATERIAL AND METHODS

Table 1. Composition of enteral formulas fed to growing pigs [View Table]

Table 2. Fatty acid composition of enteral formulas fed to growing pigs [View Table]

Fecal samples were collected from the killed pigs immediately following blood sampling on d 0, 7, 14 and 21. After making a midline abdominal incision, the rectum was incised to expose the distal rectum, and fresh fecal samples were collected into preweighed screw-cap plastic tubes containing 15 mL Carey-Blair transport media (Meridian Diagnostics, Cincinnati, OH). Samples then were weighed and stored in liquid nitrogen until analyzed.

Total urine (24-h collection) was collected on d 0, 7, 14 and 21 into buckets placed beneath the metabolic crates containing 5 mL of 6 mol/L HCl. Urinary output was recorded and an aliquot stored at 20°C until analyzed for nitrogenous constituents.

Following electrocution on d 0, 7, 14 and 21, the large bowel of each pig was excised, dissected free of mesenteric attachment, cleaned in cold PBS solution and measured for length and diameter. The whole tissue was placed on a smooth surface and formed into six equal turns of equal length. Therefore, regardless of animal weight or large bowel length, the tissues were collected from the same relative sites. Two samples from each sampling site (i.e., cecum, proximal colon, middle colon and distal colon) were collected for histopathology. Immediately after the whole tissue samples were obtained, they were washed in PBS solution and placed in 250 mL Nalgene^TM bottles containing Bouin's solution. The remaining tissue was scraped with a surgical blade to remove the mucosa. The entire tissue was scraped, the mucosa was mixed, and an aliquot was obtained. The aliquot was stored in liquid nitrogen until analyzed for prostaglandins, fatty acids, and total protein.

5

Vitamins E and A were analyzed via HPLC (Rudy et al. 1989) at the University of Illinois Toxicology Laboratory in Urbana, IL. All analyses were completed within 24 h of obtaining the sample. Vitamin C (ascorbic acid) analysis was completed via HPLC at the Michigan Animal Health Diagnostic Laboratory. The vitamin C method consisted of using a reversed-phase C18 column (3.9 cm × 150 mm; 60 A; 4 µm) with a mobile phase consisting of 0.05 mol/L sodium phosphate, 0.05 mol/L sodium acetate, 300 mg/L dodecyltrimethylammonium bromide, 40 µmol/L EDTA (disodium salt), and 5% (v/v) methanol in water at pH 4.8 coupled to an electrochemical detector.

Lipids were extracted from plasma, RBC, and colonic and cecal mucosa samples by a methanol/chloroform procedure (AOAC 1984). Briefly, the plasma or RBC were added dropwise to a known amount of cold methanol. The mucosal samples were first homogenized in methanol using a tissue homogenizer. Equal volumes of chloroform were added and the mixture vortexed (<5 min) or homogenized as for the mucosa. The remainder of the chloroform was added for a final ratio of 2:1 chloroform/methanol with a ratio of 30:1 of solvent/sample for plasma. Red blood cells were extracted similarly; however, the final ratio was 2:1 chloroform/methanol with a ratio of 50:1 of solvent/sample. The mucosa had a final ratio similar to that of RBC. The mixtures were filtered through Whatman® #1 filter paper to remove the precipitated proteins. The resulting supernatant was dried under N₂. The remaining lipid residue was resuspended in 2 mL chloroform. The phospholipid fraction was isolated by silicic acid chromatography. Thus, the lipid residue was chromatographed on activated silica gel PrepSep^TM (Fisher Scientific, Fairlawn, NJ) preconditioned with 10 mL chloroform. Neutral lipids were eluted with 10 volumes of chloroform while the phospholipids were eluted with 10 volumes of methanol. The isolated phospholipids were dried under N₂. The resulting residue was hydrolyzed, methylated, and the fatty acid methyl esters analyzed by gas chromatography (Sukhija and Palmquist 1988). A free fatty acid [nervonic acid; 24:1(n-9)] was added to the samples prior to extraction to verify that no free fatty acids were contaminating the procedure of pure isolation of phospholipid components. Fatty acids were identified by comparing retention times to known standard mixtures.

Prostaglandins (PGE, TXB₂, and 6-keto-PGF₁) were isolated according to Powell (1982) from plasma, cecal mucosa and colonic mucosa thawed at room temperature. Briefly, to 500 µL of plasma, 75 µL of 1 mol/L HCl was added to acidify to pH 3. Mucosal samples (100-300 mg) were homogenized in 15% ethanol, centrifuged (4000 × g) for 10 min, and the supernatant was removed. The supernatant was acidified to pH 3 using 34 µL of 1 mol/L HCl. The mixtures were passed across PrepSep^TM-C₁₈ cartridges (octadecyl) (Fisher Scientific) preconditioned with 20 mL 95% ethanol followed by 20 mL water. The cartridges were washed with 20 mL 15% ethanol and 20 mL petroleum ether which was discarded. The prostaglandins were eluted with 20 mL ethyl acetate containing 1% methanol (HPLC grade). The eluant was evaporated to dryness under N₂ and resuspended in 1 mL enzyme immunoassay (EIA) buffer. Reconstituted samples were stored in liquid nitrogen until analyzed for prostaglandins. Prostaglandin E, TXB₂, and 6-keto-PGF₁ were assayed by enzyme-linked immunosorbent assay using EIA Kits 514016, 519031 and 515211, respectively (Cayman Chemical, Ann Arbor, MI).

Total nitrogen was quantified using the Kjeldahl method (AOAC 1984). Urea nitrogen was determined using Sigma Kit 640 (Sigma Diagnostics, St. Louis, MO). Total protein was quantified using Sigma Kit 610. 

Colonic and cecal mucosal samples were thawed at room temperature and analyzed for total protein using a Bio-Rad DC Protein Assay Kit (Bio-Rad Laboratories, Hercules, CA).

The thoracic and abdominal organs were evaluated grossly by a pathologist at the time of electrocution. Samples of liver, cecum, and proximal, middle and distal colon fixed in Bouin's were embedded in paraffin and stained with hematoxylin and eosin (H & E). Crypt depth was measured using a microscope at 10X power using Image-1 (release 4.1) software (Universal Imaging, West Chester, PA). The H & E histology sections were examined by a pathologist without knowledge of the treatment group of the pigs. Time on the enteral formulas (0, 7, 14 and 21 d) was known in advance. A section of liver and two sections of cecum and proximal, middle and distal colon were examined.

RESULTS

Table 3. Plasma phospholipid fatty acid profiles in growing pigs fed enteral formulas containing fish oil and oligosaccharides1 [View Table]

Table 4. Red blood cell phospholipid fatty acid profiles in growing pigs fed enteral formulas containing fish oil and oligosaccharides1 [View Table]

Table 5. Colonic mucosa phospholipid fatty acid profiles in growing pigs fed enteral formulas containing fish oil and oligosaccharides1 [View Table]

The proportion of red blood cell phospholipid fatty acid data are presented in Table 4. Total SAT, MONO and PUFA were unaffected by treatment. However, total (n-3) PUFA were increased 220% (P < 0.05) whereas total (n-6) PUFA were decreased 32% (P < 0.05) within 7 d for UCNF pigs. The control pigs had a 29% increase (P < 0.05) in total (n-6) PUFA whereas total (n-3) PUFA were unaffected. The effect noted for total (n-6) PUFA was attributed to 18:2(n-6), with minor contributions from 20:2(n-6) and 22:4(n-6). Red blood cell phospholipid AA was unaffected by treatment. Similar changes in RBC phospholipids compared with plasma phospholipids for the individual (n-3) PUFA [20:5(n-3), 22:6(n-3) and 18:3(n-3)] were observed. A dramatic increase was exhibited with EPA, while DHA was increased ~75% by d 7 and continued to increase over time because of ingestion of UCNF.

The phospholipids of colonic mucosa had fatty acid profiles similar to those in plasma (Table 5). Again, the total SAT were unaffected by treatment, whereas total PUFA and (n-6) PUFA were lowered ~40% (P < 0.05) with a concomitant elevation of 530% (P < 0.05) in total (n-3) PUFA because of ingestion of UCNF. Individual (n-9) fatty acids [18:1(n-9), 20:1(n-9) and 22:1(n-9)] were lowered (P < 0.001) in the colonic mucosa as a result of UCNF consumption compared with the control treatment. Colonic mucosa alterations similar to those seen in plasma also occurred within 7 d. Thus, the change in phospholipid fatty acids occurred rapidly at the plasma level and continued to be observed at a comparable rate in the colonic mucosa. Phospholipid fatty acid profiles of cecal mucosa mirrored those of colonic mucosa in response to the UCNF diet (data not shown).

1

1

Table 6. Plasma, colonic, and cecal prostaglandin profiles in growing pigs fed enteral formulas containing fish oil and oligosaccharides1 [View Table]

Table 7. Vitamin profiles in growing pigs fed enteral formulas containing fish oil and oligosaccharides1 [View Table]

Table 8. Urinary constituents in growing pigs fed enteral formulas containing fish oil and oligosaccharides1 [View Table]

DISCUSSION

The present study also demonstrated that 0.8 g EPA/(kg body weight·d) and 0.4 g DHA/(kg body weight·d) were able to significantly change the fatty acid composition of the tissues that led to a significant reduction in proinflammatory prostaglandin synthesis in the plasma and colonic and cecal mucosa. Prostaglandins, which have a broad spectrum of inflammatory activities and have been implicated in the pathogenesis of inflammatory processes (Lauritsen et al. 1988, Sharon and Stenson 1984, Sharon et al. 1978), are altered because of (n-3) PUFA through metabolic competition between the (n-6) and (n-3) PUFA. The effect of (n-3) PUFA on prostaglandin synthesis results from the ability of (n-3) PUFA to replace the more common (n-6) PUFA at the sn-2 position of glycerophospholipids. Fish oil intake increases availability of EPA and DHA to compete with AA for membrane phospholipid incorporation.

Eicosapentaenoic acid is a poor substrate for cyclooxygenase enzymes compared with AA; thus, its conversion to prostaglandins is decreased. When the circulating fatty acid profiles are altered by increased EPA, cellular fatty acid profiles will be altered and the capacity to synthesize prostaglandins and thromboxanes will be reduced as in the current study. Additionally, alterations in phospholipase activity are likely to be important; however, increased substrate may be a more important determinant of the eicosanoid-synthesizing capacity of the tissue. This observation was demonstrated by Pacheco et al. (1987) in human subjects with IBD. It has been reported by others that increased (n-3) PUFA reduces proinflammatory eicosanoid production (i.e., LTB₄, PGE₂ and TXB₂) (Hawthorne et al. 1992, Stenson et al. 1992, Vilaseca et al. 1990). Therefore, the decrease in AA with a concomitant increase in EPA and DHA probably accounts for most of the alterations in eicosanoid production observed in our study.

Fermentable fiber is known to enhance epithelial proliferation (Howard et al. 1995). The effects may be due to providing more butyrate, the principal metabolic substrate for the colonic epithelium (Roediger 1980). Additionally, Harig et al. (1989) inferred that diversion colitis may represent an inflammatory state resulting from a nutritional deficiency in the colonic epithelium, which may be effectively treated with local application of SCFA. They speculated that SCFA were the missing nutrients. Gum arabic and FOS are soluble fibers known to increase SCFA production both in vitro and in vivo. Previous research in our laboratory has indicated increased SCFA production in vitro using human fecal inoculum with gum arabic and FOS (Bourquin et al. 1993, Titgemeyer et al. 1991). Gum arabic exhibited extensive fermentation, resulting in >6.5 mmol total SCFA/g of substrate after 24 h. Gum arabic also resulted in greater proportions of both propionate and butyrate compared with other fiber sources. Wolf et al. (1994) evaluated FOS, oligofructose and XOS in vitro using human fecal inoculum and reported an increase in the total SCFA by 6 h. Total SCFA was greatest for XOS (8.72 mmol/g) compared with FOS and oligofructose (6.86 and 6.95 mmol/g, respectively) at 12 h (Wolf et al. 1994). Results of an in vivo study using rats (Campbell et al. 1996) fed FOS and XOS at 6% of the diet indicated increased total SCFA compared with no fiber or 5% cellulose in the rat cecum. Furthermore, Younes et al. (1995) investigated diets containing 7.5% fiber as oat fiber, gum arabic, FOS, or XOS and a control (no fiber) diet. They reported a significant increase in SCFA concentration for gum arabic, FOS and XOS compared with the control and oat fiber diets. Thus, in the current study, supplemental fermentable fiber providing SCFA directly to the luminal surface of the colonic epithelium may maintain gastrointestinal tract health, gut integrity and epithelial cell proliferation.

By maintaining epithelial cell proliferation, the cecum could have a greater capacity for absorption and greater blood flow. According to Younes et al. (1995), increasing fermentable fiber to the cecum will increase fecal nitrogen excretion with a concomitant decrease in urinary nitrogen excretion. The effect may be due to an increase in SCFA production leading to an increase in intestinal microflora that require a source of nitrogen for protein synthesis. In the present study, urinary nitrogen excretion increased in both groups because of an increase in nitrogen intake; however, the increase was significantly lower in the UCNF group compared with the control group by d 14. This was probably due to the fermentable substrates in the UCNF group increasing SCFA production that may increase microflora, cecum absorption, and cecal blood flow. By increasing the blood flow to the cecum, blood urea diffusion into the lumen may be enhanced, and this could provide urea for bacterial protein synthesis, thus trapping nitrogen for increased fecal nitrogen excretion and decreasing urinary nitrogen excretion (Younes et al. 1995).

Oligosaccharides also have been implicated in increasing beneficial bacteria (i.e., Bifidobacterium spp.) that alter luminal pH, leading to inhibition of potential pathogens, stimulation of immune response, and restoration of normal intestinal flora during antibiotic therapy (Gibson and Roberfroid 1995, Mitsuoka 1990). However, in the present study, bifidobacteria were unaffected by treatment. This result may be due to the pigs being healthy and having normal levels of bifidobacteria; thus, a significant increase above normal levels may have been difficult to achieve. Additionally, adult pigs may not be as susceptible to changes in bifidobacteria as neonatal pigs.

These data indicate that UCNF elevates (n-3) PUFA levels in plasma, RBC, and colonic and cecal mucosal phospholipids and decreases the synthesis of proinflammatory prostaglandins. The reduction of PGE and TXB₂ may be of biological significance, because these prostaglandins have been demonstrated to be elevated in IBD patients and implicated in the pathogenesis of the disease. Furthermore, the UCNF may be of practical importance because the effects of the diet were evident within 7 d of sole source feeding. In conclusion, the data demonstrate that clinical studies utilizing human subjects are warranted with the UCNF to further examine its impact on inflammatory responses in IBD patients.