Iron supplementation may increase disease activity in ulcerative colitis, possibly through the production of reactive oxygen species from the Fenton reaction.
Iron supplementation may increase disease activity in ulcerative colitis, possibly through the production of reactive oxygen species from the Fenton reaction.
To assess the effects of two doses of oral iron on intestinal inflammation and oxidative stress in experimental colitis.
Colitis was induced in rats by giving 5% dextran sulphate sodium in drinking water for 7 days. First, using a 2 × 2 factorial design, rats with or without dextran sulphate sodium received the regular diet or a diet containing iron 3%/kg diet. Second, rats with dextran sulphate sodium-induced colitis were supplemented with iron 0.3%/kg diet and compared with rats on dextran sulphate sodium and regular diet. The body weight change, histological scores, colon length, rectal bleeding, plasma and colonic lipid peroxides, colonic glutathione peroxidase and plasma vitamin E and C were measured. Faecal analysis for haem and total, free and ethylenediaminetetra-acetic acid-chelatable iron was also performed.
Iron 3% and iron 0.3% increased the activity of dextran sulphate sodium-induced colitis, as demonstrated by higher histological scores, heavier rectal bleeding and further shortening of the colon. This was associated with increased lipid peroxidation and decreased antioxidant vitamins. Faecal iron available to the Fenton reaction was increased in a dose-dependent manner.
Iron supplementation taken orally enhanced the activity of dextran sulphate sodium-induced colitis and is associated with an increase in oxidative stress.
Inflammatory bowel disease often leads to iron deficiency anaemia, requiring iron therapy. However, oral iron supplementation has been observed clinically to produce disease exacerbation.1–3 This may be related to increased oxidative stress due to the Fenton reaction.
Oral iron is poorly absorbed in the upper gastrointestinal tract and most of it ends up in the inflamed colon. There, via the Fenton reaction, it can react with superoxide (O2•–) and hydrogen peroxide (H2O2), produced by activated neutrophils in the inflamed colonic mucosa. This will lead to the production of the highly reactive and toxic free hydroxyl radical (OH•).
The Fenton reaction can be described as:
The pro-oxidative imbalance created by this overproduction of free radicals, collectively called reactive oxygen species, produces an oxidative stress.4 This oxidative stress can lead to further tissue damage and amplify inflammation by increasing mucosal and vascular permeability, recruiting neutrophils and activating transcription factors, such as nuclear factor kappa- B (NF-κB), that up-regulate the transcription of adhesion molecules, cytokines and enzymes, all involved in the inflammatory response.5, 6
The antioxidant defence system can become weaker in situations in which the oxidative stress is sustained, such as during long episodes of colitis. This may further contribute to tissue damage. Indeed, antioxidant vitamins, such as vitamins E and C, and carotenoids, as well as the antioxidant enzymatic systems such as superoxide dismutase, catalase and glutathione peroxidase, have been reported to be deficient in patients suffering from inflammatory bowel disease.7–9
The purpose of this study was to investigate the effect of various doses of oral iron supplementation on intestinal inflammation and the imbalance between the production of reactive oxygen species and the antioxidant defence system, using a rat model with dextran sulphate sodium (DSS)-induced colitis.
Male Wistar rats (Charles River) weighing 100–120 g were housed individually under standard conditions. All animals were handled according to the guidelines of the Canadian Council on Animal Care, and the protocol was approved by the University of Toronto Animal Care Committee. Three diets were used to test the effect of iron supplementation on the colon. The regular diet was the Purina Rat Chow #5001 (Purina, Minneapolis, MN, USA) containing 270 mg of iron per kilogram (0.027%). The iron supplemented groups received exactly the same diet, but containing either 3000 mg of iron per kilogram (iron 0.3%) or 30 000 mg of iron per kilogram (iron 3%) (Purina Test Diets). Rat chow was supplemented using pentacarbonyl iron, a 99% (w/w) pure form of elemental iron, as microscopic spheres of 4.5–5.2 μm in diameter (Sigma). This is an iron source of choice for food fortification in the USA.10 The iron 3% diet is approximately 100-fold the regular diet and is commonly used to induce iron load in rats.11, 12 The 10-fold iron diet (iron 0.3%) is comparable to the amount of iron therapy clinically used (for example, iron sulphate, 300 mg twice daily, for a total of 120 mg of elemental iron, being 12-fold the average 10 mg of dietary iron consumed daily). Acute colitis was induced by giving 5% DSS (MW 8000; Sigma, Canada) in drinking water for the study duration. Rats without colitis received regular drinking water.
First, the effect of both DSS and iron 3% was evaluated in a 2 × 2 factorial design. This study was conducted to determine whether this amount of iron, usually used in iron load protocols, would increase intestinal inflammation and produce oxidative stress. After 7 days of acclimatization, the rats were weighed and randomly assigned to one of the four experimental groups: control (regular diet and water, n=6); control + iron 3% (iron 3% supplemented diet and water, n=22); DSS (regular diet and DSS in water, n=16); and DSS + iron 3% (iron 3% supplemented diet and DSS in water, n=26). Food and water intakes were monitored. On day 7, all rats were anaesthetized by intraperitoneal administration of sodium pentobarbital (50 mg/kg body weight) and sacrificed by cardiac puncture. Blood was then drawn into an ethylenediaminetetra-acetic acid (EDTA)-containing vacutainer, centrifuged at 3000 g for 10 min and plasma was stored at −80 °C for future analysis. Plasma for vitamin C was stabilized immediately with 100 g/L metaphosphoric acid prior to freezing. The colon from the colocaecal junction to the anal verge was removed and the length was recorded. It was then opened and cut longitudinally into two pieces for histological evaluation and tissue measurements, respectively. For histological examination, the colon was fixed in 10% formalin, embedded in paraffin and stained with haematoxylin and eosin. Staining was also performed with Prussian Blue for qualitative determination of iron. The other part of the colon was stored at – 80 °C until assayed for lipid peroxides and glutathione peroxidase. Fresh samples of faeces were taken daily for the measurement of haem and total, free and EDTA-chelatable iron. Half of the samples were used for the measurement of total iron and the other half were homogenized with an equal amount of water (w/w). Based on the effect of iron 3%, we also studied the effect of iron 0.3% in rats with DSS-induced colitis (DSS + 0.3% iron, n=14). The same protocol was followed.
Crypt and inflammatory scores were determined according to the scoring system of Murthy et al.13 by a pathologist who was unaware of the experimental protocol. The method of scoring crypts was based on the following: grade 0, intact crypt; grade 1, loss of bottom one-third of the crypt; grade 2, loss of bottom two-thirds of the crypt; grade 3, loss of entire crypt with the surface epithelium remaining intact; and grade 4, loss of the entire crypt and surface epithelium (erosion). These changes were quantified with regard to the percentage affected by the disease process: (1) 1–25%, (2) 26–50%, (3) 51–75% and (4) 76–100% of the surface area examined. The crypt score is the product of the grade and the percentage area involvement. The inflammation score was performed according to Onderdonk and Bartlett14 and graded as 0–3: grade 0, normal; grade 1, focal inflammatory cell infiltration, including polymorphonuclear leucocytes; grade 2, inflammation cell infiltration, gland dropout and crypt abscess; and grade 3, mucosal ulcers. The extent of involvement was estimated in the same way as for the crypt score and the final inflammation score was given by the product of the inflammation grade and the extent of involvement. The combined score is the summation of the crypt and inflammatory score.
For colonic measurements, tissues were thawed, washed in normal saline, blotted dry, weighed and homogenized with a 10-fold (v/w) amount of ice-cooled tris-HCl buffer, pH 7.4, containing 5 mM butylated hydroxytoluene and centrifuged at 3000 g for 10 min prior to analysis of lipid peroxidation products and glutathione peroxidase activity. Plasma and tissue lipid peroxides were measured using commercially available kits (LPO-586, R & D Systems, Minneapolis, MN, USA). This assay measures free malondialdehyde and 4-hydroxyalkenals. Lipid peroxides were expressed in nmol/mg protein in colonic tissue and nmol/mL in plasma. Glutathione peroxidase activity of colonic homogenate was determined at 25 °C with t-butyl hydroperoxide (0.3 mmol/L) as the substrate using a kit (Bioxytech GPx-340, OXIS International, Portland OR, USA). One unit of glutathione peroxidase is defined as one micromole of the reduced form of nicotinamide adenine dinucleotide phosphate oxidized per minute. Glutathione peroxidase was expressed in mU/mg protein. The protein concentration in the colonic homogenate was determined by the Biuret method.15 Plasma α-tocopherol was determined by the high performance liquid chromatography technique.16 This method uses an isocratic solvent (methanol–acetonitrile–tetrahydrofuran, 50:45:5, v/v/v), a reverse phase C18 column and fluorescence spectrophotometry at 294 nm following a hexane extraction using 200 μL of plasma sample. Plasma vitamin C was measured by spectrophotometry.17 In this method, total biologically active vitamin C levels were determined at 521 nm using 2,4-dinitrophenyl hydrazine as a chromagen.
The total iron concentration of the faeces was measured after ashing the faeces in silica crucibles at 525 °C for 48 h. The ash was dissolved in nitric acid and the solution diluted to an appropriate volume with distilled water. The iron content was then measured with an atomic absorption spectrophotometer (Hitachi Z8000).18 Free iron in the faeces was assessed using an adaptation of the method described by Simpson et al.18 Preweighed samples (0.5 g of wet faeces) of faecal homogenates were mixed with 3 mL of water. The samples were centrifuged for 30 min at room temperature at 6000 g and the supernatants were collected. The pellets were washed with an additional 2 mL of water and centrifuged, and the two supernatants were combined and the volume recorded. The EDTA-chelatable iron was then assessed in the same sample by washing it an additional two times in 2 mL TE buffer (10 mmol tris-HCl/L, 1 mmol EDTA/L) and then by measuring the iron content of the combined supernatants. The iron content of the resultant solutions was measured by atomic absorption spectrophotometry. Free and chelatable iron are both potentially available for the Fenton reaction.
Haem in the faeces was measured by the Hemoquant assay.19 Iron does not interfere with this assay.20 In brief, an average of 20 mg of faecal homogenates were weighed accurately and added to 4 mL of a solution containing, per litre, 2.5 M oxalic acid, 90 mmol FeSO4, 50 mmol uric acid and 50 mmol mannitol or 1.5 M citric acid at 80 °C and maintained at 80 °C for 90 min. The supernatants (0.5 mL) were mixed with 3 mL ethyl acetate–acetic acid and 1 mL potassium acetate (3 mol/L). The upper phase (1.25 mL) was then mixed with 0.5 mL 1-butanol and 3.8 mL potassium acetate (93 mol/L) in 1 mol KOH/L. The upper phase (0.5 mL) was added to phosphoric acid (2 mol/L)–acetic acid (9:1, v/v). The lower phase was then measured by fluorometry at an emission wavelength of 650 nm using an excitation wavelength of 400 nm. The measured fluorescence of the samples was compared with that of a standard 0–50 μg/L solution of coproporphyrin in 1.5 mol/L HCl (Sigma). Results were expressed per gram of wet faeces.
For the factorial design study, the two factors were defined as DSS and iron. The overall effects of iron and DSS were determined with a two-way analysis of variance (ANOVA). Fisher’s exact test was used to compare the mortality rate between the groups. Parameters from rats with DSS-induced colitis supplemented with iron 0.3% were also compared with those of the DSS group on a regular diet, using the unpaired t-test. When appropriate, one-way ANOVA was also used. The statistical package SAS (SAS Version 7.1, SAS Institute, Cary, NC, USA) was used. The level of significance was determined at P < 0.05. Results are reported as the mean ± S.E.M.
All DSS-teated rats developed signs of colitis, such as bloody diarrhoea, within the 7 days of the study. As seen in Table 1, water intake was significantly lower in all the DSS-treated rats compared to those without DSS. Rats with DSS ate significantly less than rats without DSS, and those on iron also ate less compared to rats without iron. Furthermore, rats on DSS lost more weight during the study compared to rats without DSS, and there was also a significant reduction in weight in rats taking iron when compared with no iron treatment. There was no mortality observed in the control and control + iron 3% groups. However, the mortality rate was significantly higher in the rats on DSS and iron compared with DSS alone (33.3% vs. 6.3%, P ≤ 0.05).
As previously described with this model of colitis,13 the colonic length of rats treated with DSS was significantly shorter than that of rats without DSS (control, 18.1 ± 0.4 cm; DSS, 15.8 ± 0.1 cm; control + iron 3%, 16.4 ± 0.1 cm; DSS + iron 3%, 13.6 ± 0.2 cm; P ≤ 0.05). Iron also reduced significantly the colon length in rats supplemented with iron when compared with that of rats on a regular diet (P ≤ 0.05). No interaction was seen.
Both DSS and control rats supplemented with iron 3% showed iron deposition in the non-inflamed proximal segments of the colonic mucosa (Figure 1A). In the distal part of the colon, where colitis was present, iron-containing debris was observed at the surface of some of the erosions and no superficial epithelial cells were left to demonstrate iron deposition (Figure 1B). No iron deposition was visible in rats on a regular diet. As previously reported,21 microscopic examination of the DSS-treated rats showed that the colitis involved mainly the distal colonic mucosa and submucosa, with areas of erosion, crypt distortion and inflammatory infiltration. Using the scoring system, inflammation and crypt scores were higher in rats treated with DSS + iron 3% compared with those treated with DSS alone (Figure 2). In control rats supplemented with iron 3%, histological examination did not demonstrate any inflammation or crypt distortion, and therefore the scoring system was not used.
Iron significantly increased lipid peroxidation, based on higher plasma and colonic lipid peroxides, when compared with rats on a regular diet (Table 2). On the other hand, the administration of DSS in rats did not significantly increase the lipid peroxide levels when compared with rats without DSS. The activity of colonic glutathione peroxidase was significantly higher in DSS-treated rats than in rats without DSS. However, the activity of this enzyme was not significantly affected by iron 3% supplementation. The plasma concentrations of α-tocopherol were significantly lower in iron 3% supplemented rats than in rats on a regular diet. There was no significant difference with DSS.
Based on the results of the previous study, another group of rats with DSS-induced colitis was studied with a lower dose of iron (iron 0.3%). This group was compared with the DSS group.
There were no significant differences between the DSS and DSS + iron 0.3% groups with respect to their body weight change and DSS intake. No mortality was observed in the DSS + iron 0.3% group.
The colonic length was significantly shorter in the DSS + iron 0.3% group compared with the DSS group (Figure 3). There were significantly higher inflammatory and combined scores of colitis in the iron 0.3% group compared with the DSS group (Figure 3). However, the difference in crypt scores was not statistically significant (P=0.1). In the DSS + iron 0.3% group, Prussian Blue staining of colonic tissue showed iron deposition qualitatively similar to that described in the iron 3% group (Figure 1A,B).
At a more clinically relevant dose of iron (10-fold the regular diet), there was still a significant increase in plasma and colonic lipid peroxides in the DSS + iron 0.3% group compared with the DSS group (Figure 4). Furthermore, the activity of colonic glutathione peroxidase was reduced in rats supplemented with iron 0.3% (Figure 5). In addition, plasma levels of α-tocopherol and ascorbic acid in rats supplemented with iron 0.3% were significantly lower than in rats on a regular diet (Figure 6).
As shown in Figure 7, there was an increase in the amount of faecal haem over time with DSS. Starting on day 6 of the study, the rats supplemented with iron 0.3% had significantly more rectal bleeding than the rats on DSS alone. Rats supplemented with iron 3% also bled significantly more than the DSS group. This started at day 2, earlier than the two other groups. Overall, the rats supplemented with iron bled significantly more than the rats on DSS alone in a time– and dose–response manner. Diet supplementation with iron 0.3% and 3% caused a highly significant increase in the total iron concentration in the faeces of rats compared with those on DSS alone (Table 3). The increase in faecal combined water-soluble and EDTA-chelatable iron, which is available for the Fenton reaction, after iron 0.3% and 3% supplementation was also highly significant (Table 4).
The results of the present study showed that oral iron supplementation, at either high or moderate doses, increased the disease activity in rats with DSS-induced colitis. This is based on increased faecal haem and histological scores, as well as reduced colonic length. In addition, iron supplementation induced oxidative stress in both healthy animals and rats with DSS-induced colitis, as reflected by reduced plasma α-tocopherol levels and increased products of lipid peroxidation in either the colon or the plasma. Furthermore, a high dose of oral iron given to rats with DSS-induced colitis increased mortality and induced a dramatic weight loss.
During intestinal inflammation, there is an intense flux of circulating neutrophils into the inflamed mucosa, which amplifies the inflammation by releasing large amounts of superoxide and hydrogen peroxide.22–24 These free radicals interact with iron (Fe) to yield the highly reactive hydroxyl radical (OH•) or OH•-like species via the Fenton reaction.25, 26 This increased production of hydroxyl radicals may enhance intestinal inflammation by increasing mucosal and vascular permeability as well as the recruitment and activation of neutrophils.5 Furthermore, hydroxyl radicals may attack the polyunsaturated fatty acids from the lipid membranes and induce lipid peroxidation.
The increased activity of colitis induced by iron supplementation is clinically relevant to ulcerative colitis because iron deficiency anaemia caused by mucosal bleeding is common in this patient population, and is frequently treated by oral iron.27 Because iron is poorly absorbed, most of it ends up in the colon where it is in contact with the inflamed colonic mucosa. In our study, incremental doses of iron supplementation increased, in a dose-dependent manner, faecal iron concentration, thereby increasing the available form of iron (free and chelatable iron) to levels known to drive the Fenton reaction maximally.28 Such concentrations of iron in faeces produce significant amounts of hydroxyl radicals, as previously demonstrated by Lund et al.29 In addition, iron may induce in situ membrane lipid peroxidation of colonocytes, as suggested by the presence of iron in the colonic epithelial cells and the increased colonic lipid peroxides in all rats supplemented with iron. Similar results were also reported by Soyars and Fischer30 who demonstrated an increase in mucosal iron concentration following iron supplementation with 10-fold the regular diet in the rat.
Oral iron can also be absorbed by up to 40%, and therefore systemic iron may have contributed to the oxidative stress and increased intestinal inflammation in the supplemented rats. We have previously shown that parenteral iron has this effect.31 However, with oral supplementation, iron deposition is seen only at the level of the mucosa, while parenteral iron leads to iron deposition throughout the intestinal wall. This suggests that the effect of oral iron is probably more from the lumen rather than systemic.
The effect of oral iron on intestinal inflammation has recently been reported by others. In using the iodoacetamine model of colitis in rats, Reifen et al. observed that oral iron supplementation can enhance mucosal damage.32 However, the microscopic assessment of the mucosa was qualitative and no data on myeloperoxidase measurements were given. Another group assessed the effect of dietary and topically administered iron in the interleukin-10–/– mouse model of colitis.33 No significant enhanced colonic inflammation was noted on histology after iron supplementation, possibly due to the relatively small number of mice per group (eight mice per group). Nevertheless, oral as well as rectal administration of iron resulted in increased colonic pro-inflammatory cytokine production. This suggests a role of iron in the modulation of inflammatory mediators in the colon.
The increase in lipid peroxidation in otherwise healthy rats supplemented with iron has also been observed by other investigators.34–36 However, no inflammation or crypt distortion was observed in the control + 3% iron group despite high colonic lipid peroxide. The fact that iron supplementation by itself did not produce intestinal inflammation supports the idea that the presence of activated neutrophils from pre-existing inflammation is necessary to produce this detrimental effect from iron. One mechanism may be through the activation of NF-kB by iron-generated free radicals, mainly in cells already primed by inflammatory mediators.6, 37
Our observations that iron supplementation leads to a decrease in plasma α-tocopherol and vitamin C levels agree with previous data in animals.12 This effect of iron has also been reported in humans, where erythrocytes and spleens from subjects with iron overload from repeated transfusions underwent excessive lipid peroxidation that was associated with decreased levels of vitamin E38, 39 and C.40 Vitamin E is a very potent liposoluble antioxidant and constitutes an important defence mechanism against lipid peroxidation. Several studies have found that, when membrane lipids are exposed to an oxidative stress, vitamin E in synergy with vitamin C prevents lipid peroxidation until it is significantly consumed.41, 42 Therefore, it is possible that the reduction in vitamin E and C found in our study was due to consumption from increased lipid peroxidation induced by iron, as weights and intake were not different. On the other hand, in rats with colitis supplemented with iron 3%, the low food intake could also have contributed to the decrease in antioxidant vitamin levels and hence to oxidative stress.
We also examined the effect of iron supplementation on colonic glutathione peroxidase as this key antioxidant enzyme is up-regulated in inflamed colonic tissue of ulcerative colitis patients.43 As expected, we observed a significant increase in glutathione peroxidase activity in rats with DSS-induced colitis. Intriguingly, the activity of colonic glutathione peroxidase was not modified by a high dose of iron supplementation, but was significantly reduced in rats supplemented with iron 0.3%. Although iron can clearly inactivate glutathione peroxidase in vitro,44 results from other animal studies are variable depending on the dose of iron supplementation and the tissue studied.34–36, 45
Rats with DSS-induced colitis constitute a good model to investigate the relationship between iron supplementation, oxidative stress and inflammation. Although the mechanism of DSS-induced colitis remains unclear, oxygen-derived radicals may be implicated in the pathogenesis of colitis, as suggested by higher urinary 8-hydroxydeoxyguanosine excretion,46 higher plasma markers of lipid peroxidation,47 significant depletion of several mucosal antioxidants48 and the protective effect of antioxidants including 5-aminosalicylic acid.49 This model was reproduced in this study and is characterized by a predominantly left-sided colitis manifested by bloody diarrhoea, weight loss, shortening of the colon, mucosal ulcers and an inflammatory infiltrate.13, 21, 50 Shortening of the colon is characteristic of this model of colitis,13 and could be the result of abnormality in the muscle layers. Due to the 2 × 2 factorial design used in the initial study and the surprisingly high plasma and colonic lipid peroxide levels in the control + iron 3% group, as opposed to the DSS + iron 3% group, we could not detect a statistically significant increase in oxidative stress with DSS. Significant reduction in food intake and thus oral iron in the DSS + iron 3% group may have contributed to this discrepancy.
In conclusion, we demonstrated that oral iron supplementation aggravated colonic injury caused by DSS-induced colitis, as shown by validated histological scores, quantitative evaluation of rectal bleeding and colon length. This increase in intestinal damage is associated with an increase in lipid peroxidation and a decrease in antioxidant vitamins. Because iron supplementation is used to treat anaemia in patients with active colitis, these findings are clinically relevant, and future studies should investigate efficient ways to minimize the colonic mucosal exposure to iron, especially in the presence of intestinal inflammation. Another alternative would be to add antioxidants to scavenge free radicals produced by the Fenton reaction, thereby preventing lipid peroxidation and reducing mucosal damage.
This study was supported by grants from the Canadian Digestive Disease Foundation and the Medical Research Council (grant # MT-15488). Dr J. Carrier was supported by the Armstrong Fellowship from the Crohn’s and Colitis Foundation of Canada. We acknowledge the contribution of the histotechnology staff at St Joseph’s Health Center.