Oral ferrous sulfate supplements increase the free radicalgenerating capacity of feces from healthy volunteers13

Elizabeth K Lund, S Gabrielle Wharf, Susan J Fairweather-Tait, and Ian T Johnson

ABSTRACT Background: Most dietary iron remains unabsorbed and hence may be available to participate in Fenton-driven free radical generation in conjunction with the colonic microflora, leading to the production of carcinogens or direct damage to colonocytes. Objective: Our aims were to measure the proportion of fecal iron available to participate in free radical generation and to determine the effect of an oral supplement of ferrous sulfate on free radical generation. Design: Eighteen healthy volunteers recorded their food intake and collected fecal samples before, during, and after 2 wk of supplementation (19 mg elemental Fe/d). Total, free, and weakly chelated fecal iron were measured and free radical production was determined by using an in vitro assay with dimethyl sulfoxide as a free radical trap. Results: Fecal iron increased significantly during the period of supplementation and returned to baseline within 2 wk. The concentration of weakly bound iron in feces ( 1.3% of total fecal iron) increased from 60 mol/L before to 300 mol/L during supplementation, and the production of free radicals increased significantly ( 40%). Higher-carbohydrate diets were associated with reduced free radical generation. Conclusion: Unabsorbed dietary iron may increase free radical production in the colon to a level that could cause mucosal cell damage or increased production of carcinogens. Am J Clin Nutr 1999;69:2505. KEY WORDS Dietary iron, free radicals, colon cancer, feces, ferrous sulfate, Fenton chemistry, humans Colorectal carcinoma develops progressively via the adenoma-carcinoma sequence that is associated with the induction of mutations, first in proliferating crypt cells and subsequently in abnormal clones derived from them. Such mutations may be mediated by fecal mutagens, free radicals, or both. Free radical production in fecal incubates is known to depend on diet (3). In many systems the production of superoxide and hydrogen peroxide by aerobic metabolism is favored by high concentrations of suitably chelated iron. Oxygen radicals are known to damage protein, lipids, and DNA under in vivo conditions, and this damage has been implicated in the induction of somatic mutations that may favor the development of several forms of cancer (4, 5). Unabsorbed dietary iron enters the colon and in conjunction with intraluminal bacteria may become available for participation in a combination of Haber-Weiss and Fenton-type reactions that generate hydrogen peroxide and hydroxyl radicals at the mucosal surface (68). Hydrogen peroxide or iron may also enter colonocytes and increase the risk of DNA damage in a manner similar to that described for immune cells (9), thus increasing the risk of a mutation, either as an initiating event or later in the adenoma-carcinoma sequence. Alternatively, ironmediated reactions may be involved in the conversion of procarcinogens to carcinogens within the lumen of the colon (6). Iron has also been shown to increase crypt cell proliferation in rats treated with a chemical carcinogen (10) and in rats fed high-fat diets (11). Even moderate oral iron supplementation has a significant effect on luminal iron concentrations and increases mucosal cell proliferation in rat colons (12). Intraluminal iron might therefore act in the initiation of carcinogenesis by causing DNA damage or at the promotion stage by stimulating polyp growth. The aim of the present study was to test the hypothesis that oral iron supplementation can modify the iron content of human feces so as to increase the formation of free radicals.

INTRODUCTION The epidemiologic evidence supporting a link between colorectal cancer and diet is strong (1). The stem cells of the colorectal mucosa lie near the complex, potentially mutagenic environment of the fecal stream; an understanding of the effects of diet on the chemistry of fecal material is therefore essential in the development of strategies to prevent colorectal cancer. A prospective study showed that high iron consumption is associated with an increased risk of colorectal cancer (2). Higher concentrations of plasma iron were associated with distal tumors, but a high dietary iron intake, independent of red meat consumption, was specifically associated with an increased risk of cancer of the proximal colon. 250

1 From the Institute of Food Research, Norwich Research Park, Norwich, United Kingdom. 2 Supported by the Ministry of Agriculture, Fisheries, and Food of England and Wales. 3 Address reprint requests to EK Lund, Institute of Food Research, Norwich Research Park, Colney, Norwich, NR4 7UA, United Kingdom. E-mail: liz.lund@bbsrc.ac.uk. Received April 16, 1998. Accepted for publication July 2, 1998.

Am J Clin Nutr 1999;69:2505. Printed in USA. 1999 American Society for Clinical Nutrition

FECAL FREE RADICALS AND ORAL IRON SUBJECTS AND METHODS Subject protocol Ethical approval was obtained from the Research Ethics Committee of the Norfolk and Norwich Health Care Trust before the start of the study. Twenty healthy members of the staff of the Norwich Research Park (10 male and 10 female) gave their writ ten, informed consent to participate. Eight men aged 2248 y (x : 30 y) with a body mass index (in kg/m 2) of 1834 (x : 23) and 10 women aged 2154 y (x : 37 y) with a body mass index of 2027 : 23) completed the study. Of the 10 female subjects, 3 were (x postmenopausal and 2 of these were receiving hormone replacement therapy. No subjects were taking vitamin or mineral supplements during, or had taken them for 6 wk before, the study. An initial blood sample was taken from all subjects before final recruitment to the study and sent for routine biochemical screening at the Norfolk and Norwich Hospital. Hemoglobin, red blood cell count, total hematocrit, mean red cell volume, and zinc protoporphyrin were measured by standard methods to ensure that no volunteers were anemic. Each volunteer was asked to weigh and record his or her normal food intake during the first 7 d of the study. During this period subjects were also asked to collect 3 fecal samples according to normal bowel habit. Subjects then took a 100-mg supplement of FeSO47H2O (prepared as capsules by a hospital pharmacy) every day for 14 d (subsequent analysis showed that the capsules contained 19 mg elemental Fe). During the second week of supplementation, subjects were asked to collect an additional 3 fecal samples. Food intake was again recorded, in this case by portion size with reference to a set of photographs of standard portions. In the second week after the end of supplementation the subjects were again asked to collect 3 fecal samples. Iron content of fecal samples Fecal samples were weighed and half of the sample was homogenized in a blender (Stomacher 400; Seward, London) with an accurately measured volume of deionized water of approximately equal weight to the feces. The homogenized material was then subsampled for the subsequent measurement of free radical production and concentrations of water-soluble iron, EDTA-chelatable iron, and heme iron. The total iron concentration of the feces was measured after the feces collections were freeze-dried and then ashed in silica crucibles at 480 C for 48 h. The ash was dissolved in warm, concentrated hydrochloric acid and the solution diluted to an appropriate volume with distilled water. The iron content was then measured with an atomic absorption spectrophotometer (PU 9100X; Philips, Cambridge, United Kingdom). Values were calculated from a standard curve and analytic accuracy was confirmed by using National Bureau of Standards standard reference material 8431 mixed diet (Office of Standard Reference Materials, Washington, DC). Free iron in feces was assessed by using an adaptation of the method described by Simpson et al (13). A preweighed sample of fecal homogenate containing 5 g feces (wet weight) was mixed with a measured volume of water to give a final volume of water plus feces of 15 mL. The sample was centrifuged for 30 min at room temperature at 6000 g and the supernate collected. The pellet was washed with an additional 10 mL water and centrifuged and the 2 supernates were combined and the volume recorded. Readily exchangeable iron was then

251

assessed in the same sample by washing it an additional 2 times in 10 mL TE buffer (10 mol tris HCl/L, 1 mmol EDTA/L)/g sample and then measuring the iron content of the combined supernates. The iron content of the resultant solutions was measured by atomic absorption spectrophotometry as described above. The water content of the fecal samples was calculated from a subsample by weighing before and after freeze-drying. Heme iron was measured by the HemoQuant assay (14). In brief, 20 mg fecal homogenate was weighed accurately before the addition of a solution containing 2.5 mol oxalic acid/L and 90 mmol FeSO4/L at 80 C or 1.5 mol citric acid/L at 80 C and maintained at 80 C for 90 min. The supernate (0.5 mL) was mixed with 3 mL ethyl acetateacetic 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 (3 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, by vol). The lower phase was then measured by fluorometry at an emission wavelength of 653 nm by using an excitation wavelength of 402 nm. Standards were prepared from cyanomethemoglobin up to a concentration of 8.87 mmol/L. Oxalic acid removes iron from the porphyrin ring structures whereas citric acid treatment does not. Porphyrin rings do not emit light when iron is bound to them; under the conditions used in this assay, therefore, heme iron could be calculated as the difference between the 2 values. Free radical production in fecal samples The effect of dietary iron on free radical production in human feces was explored by using an assay developed by the method of Babbs and Gale (7, 15) and based on the following reaction: Dimethyl sulfoxide + OH methanesulfinic acid + CH3 (1) Each fecal subsample (12 g) was incubated overnight in trisbuffered saline (pH 7.0) containing 5% dimethyl sulfoxide (0.7 mol/L), glucose (0.1%), and Na2EDTA (50 mmol/L) at 37 C. The sample was then centrifuged at 900 g for 10 min at room temperature, the supernate removed, and the protein removed as a precipitate by lowering the pH to 1.0 for 10 min by adding 12 mol HCl/L. The pH was then returned to 7.4, the sample was centrifuged at 900 g for 10 min at room temperature, and the supernate was stored at 20 C before batch analysis of the methanesulfinic acid content. Standards were prepared freshly before each assay, with 075 mmol methanesulfinic acid/L in the incubation medium, and both samples and standards were processed identically. A 2-mL aliquot was mixed well with 0.2 mL H2SO4 (1 mol/L) and then with 1-butanol (4 mL). The upper phase was mixed with 2 mL sodium acetate buffer (0.5 mol/L, pH 5.0) and then centrifuged at 500 g for 3 min at room temperature before 1.8 mL of the lower aqueous phase was removed. The lower aqueous phase was then adjusted to a pH of 2.5 by adding HCl (1 mol/L) before the addition of Fast Blue BB salt (0.03 mol/L; Sigma, Dorset, United Kingdom) to form the colored product diazosulfone acid. Once the color reaction reached a plateau, after 10 min in the dark, 1.5 mL toluene:1-butanol (3:1, by vol) was added and the sample was mixed for 120 s before separation of the phases by centrifugation at 500 g for 3 min at room temperature. The upper phase was then removed, washed with 1-butanolsaturated water, and measured by scanning spectrophotometry at a peak absorbance between 340 and 520 nm. Peak absorbance was invariably between 410 and 420 nm.

252

LUND ET AL

TABLE 1 Hemoglobin, hematocrit, mean red cell volume, and zinc protoporphyrin concentrations in subjects before the start of supplementation1 Hemoglobin g/L Women (n = 10) 133.8 2.0 (126147) Men (n = 8) 151.8 2.9 (138164) Both (n = 18) 142.3 11.8 (126164) 1 x SEM; range in parentheses. 0.3815 0.01 (0.3490.431) 0.4280 0.01 (0.3750.459) 0.4035 0.034 (0.3490.431) Hematocrit Mean red cell volume fL 89.86 0.80 (87.195.4) 92.51 0.79 (90.597.9) 91.12 2.74 (87.197.9) Zinc protoporphyrin g/g hemoglobin 1.85 0.13 (1.32.5) 1.74 0.07 (1.42.0) 1.80 0.34 (1.32.5)

Analysis of dietary intake Diaries from the initial 7-d weighed-intake measurements were analyzed with the commercial software package FOOD BASE (Institute of Brain Chemistry and Human Nutrition, Queen Elizabeth Hospital for Children, London). A second set of diaries was used as reminders to the volunteers to consume similar menus during supplementation as during the period when fecal samples were collected before supplementation began and to check for significant variations from the control diet. Statistical analysis The data were analyzed by using the MINITAB statistical package (release 8, Macintosh version; Minitab Inc, State College, PA). The general linear models technique was used to analyze the repeated measures of iron concentration and free radical production, taking into account between-subject variation. Linear regression analysis was used to assess the effect of diet on the measured variables and the correlation between measured variables.

between the concentrations of available iron and total iron in feces (r = 0.596, P < 0.001): Available iron = (0.0135 total iron) 1.1 (2)

The increase in both water-soluble and EDTA-chelatable iron after supplementation with ferrous sulfate was highly significant (P < 0.001) in both men and women (Figure 1). Free radical production After supplementation there was a signicant increase in the free radicalgenerating capacity of the fecal samples measured as mol methanesulfuric acid/kg wet wt feces (P < 0.001; Figure 2). Analysis of all data collected throughout the experiment showed signicant correlations between free radical generation and water-soluble iron (r = 0.673, P < 0.001) and between free radical generation and EDTA-chelatable iron (r = 0.559, P < 0.001). Before supplementation there was no correlation between heme iron concentrations in feces and free radical generation, despite some evidence of a correlation between heme iron concentrations and water-soluble iron (r = 0.470, P < 0.05). These effects were not sex dependent and there was no signicant effect of pre- or postmenopausal state in female subjects. Analysis of the dietary intake data from the women revealed a possible inverse correlation between baseline free radicalgenerating capacity of feces and the total carbohydrate content of the diet (r = 0.627, P = 0.05 ). This relation was not signicant in men. However, in men, the estimated intake of dietary ber was inversely associated with the increase in free radicalgenerating capacity caused by the iron supplement (r = 0.711, P < 0.05; Figure 3).

RESULTS Initial screening Mean values for initial hemoglobin concentration, hematocrit, mean red cell volume, and zinc protoporphyrin for all subjects are summarized in Table 1. Potentially relevant aspects of the subjects estimated habitual dietary intakes are given in Table 2. Iron intakes ranged from 9.0 to 24.7 mg Fe/d in men and from 9.1 to 47.2 mg Fe/d in women. (Two women had exceptionally high intakes of iron as a result of consuming game.) No significant correlation was found between estimated dietary iron intake and any of the indexes measured. Fecal iron concentrations Neither the total fecal iron concentration nor heme iron in feces correlated with dietary iron intake before supplementation. The only dietary factor significantly associated with fecal iron was dietary starch, which was inversely correlated with the fecal iron concentration (r = 0.454, P < 0.05). There was a tendency for fat intake to be positively associated with an increased fecal iron concentration (r = 0.419, P = 0.08). Supplementation with ferrous sulfate caused a highly significant increase in the total iron concentration in feces in both men and women. Concentrations returned to baseline within 2 wk after supplementation ended (Table 3). Just > 1% (1.35%) of the total iron concentration of feces was in a form that was likely to be available for participation in free radical generation or for mucosal uptake, and this was independent of the total amount of iron present. Thus, there was a linear relation

DISCUSSION Most iron within tissues is bound to protein, but the small pool of free iron, which is essential for several cellular
TABLE 2 Subjects daily dietary intakes as calculated from 7-d weighed intake diaries1 Men (n = 8) Total energy (kJ/d) Protein (% of energy) Fat (% of energy) Carbohydrate (% of energy) Starch (% of energy) Sugar (% of energy) Fiber (g) Dietary iron intake (mg Fe/d) Phytic acid (g/d) 1 x SEM. 10 217 470 15.5 0.9 35.8 1.9 44.2 2.3 26.8 2.0 17.2 1.2 15.4 1.7 16.4 1.9 0.08 0.03 Women (n = 10) 10 878 689 18.7 1.3 31.9 1.9 44.6 2.2 23.2 1.5 21.0 1.8 13.7 1.4 26.6 4.4 0.08 0.03

FECAL FREE RADICALS AND ORAL IRON

TABLE 3 Iron concentration of feces before, during, and after supplementation with ferrous sulfate (100 mg/d) 1 Baseline During supplementation

253

After supplementation

Iron concentration ( g/g wet wt) Men (n = 8) 108.5 6.8a 277.6 30.0b a Women (n = 10) 90.6 10.6 284.2 44.6b ( g/g dry wt) Men (n = 8) 369.1 21.7a 969.5 96.5b Women (n = 10) 352.4 21.7a 985.3 96.7b 1 x SEM. Values in the same row with different superscript letters are signicantly different, P < 0.05.

123.2 25.2a 108.9 17.9b 465.7 103.6a 391.5 37.8a

processes, is known to catalyze free radical production, particularly in diseased states. Iron was first proposed as a potential carcinogen > 35 y ago (16) and the role of iron in the etiology

of various cancers has been discussed extensively in 2 review papers (17, 18). The overall iron status of the body is tightly controlled by regulation of iron uptake in the small intestine. Once iron stores are replete, dietary iron remains almost completely unabsorbed and high concentrations of iron can develop in the distal gut lumen after the reabsorption of water. The hypothesis that the presence of iron in the fecal stream may lead to the generation of free radicals, and that this effect may be procarcinogenic for colorectal cancer, was advanced by both Blakeborough et al (6) and Babbs (7). However, the effect of oral iron intake on the quantity of iron available for free radical generation in the human fecal stream has not been reported previously. To participate in reactions leading to free radical production, iron must be either freely water soluble or readily bound to small organic ions such as ascorbic acid and citrate. In his original study, Babbs (7) wanted to establish whether reactive oxygen species could be generated by fecal bacteria in the presence of suitably chelated iron. Iron EDTA was used as the source of iron in the assay system because EDTA is considerably more stable than naturally occurring small organic ions. In the pres-

FIGURE 1. Mean ( SEM) concentrations of water-soluble iron in feces and iron that was soluble in the presence of EDTA at pH 7.5 (readily chelatable iron) before (initial), during (supplement), and after (post) supplementation with 100 mg FeSO4/d. The increase in both forms of iron during supplementation was significant (P < 0.001). Sex had no significant effect.

FIGURE 2. Effect of iron supplementation (100 mg FeSO4/d) on mean ( SEM) free radical production in feces determined by using methanesulfinic acid (MSA) as an end product in an in vitro assay. The increase during iron supplementation was significant (P < 0.001). Sex had no significant effect.

254

LUND ET AL

FIGURE 3. Inverse correlation between the presence of fiber in the diet and the increase in the free radicalgenerating capacity of feces after iron supplementation in men (P < 0.05).

ent study, we wished to determine whether an increase in the concentration of intraluminal iron derived from the oral route would lead to any quantifiable variation in free radical production in feces; we therefore used sodium EDTA as the chelating agent, to avoid adding iron to the system from an external source. The present results clearly indicate that oral iron supplementation increases the concentration of fecal iron in a form potentially available for the generation of free radicals. Using rat feces, Babbs (7) showed an increase in free radical generation over iron concentrations ranging from 0 to 100 mol/L. In the present study, we showed that the concentration of water-soluble iron is normally equivalent to 25 mol/L in human feces and that with oral supplementation concentrations rose to > 100 mol/L. If the pool of easily chelatable iron is also available for free radical production, then the total concentration of iron in the active intraluminal pool rose from 60 to 350 mol/L, which far exceeds the concentration required for maximal free radical production in Babbss system (7). It was also clearly established in the present study that consumption of 100 mg FeSO 4/d was associated with a marked increase in the production of free radicals in feces. The mean increase in free radical production was less than the rise in the putative active pool of fecal iron, but this is explicable if, as the previous results of Babbs (7) suggest, free radical production in feces is maximal at 100 mol/L. The inverse associations between free radical production and carbohydrate intake in women and fiber intake in men provide some limited evidence that habitual intake of dietary fiber may suppress the production of reactive oxygen species. In a recent study, Erhardt et al (3), using a similar approach to measure free radical production, reported a significant reduction in the free radicalgenerating capacity of feces from volunteers consuming high-fiber, low-fat diets compared with low-fiber, high-fat diets. This reduction was associated with a 42% reduction in total fecal iron. In the present study, we found no effect of fiber on baseline total fecal iron concentrations, but starch intake was negatively correlated with fecal iron concentrations. The range of fiber intakes in our study was low compared with those in the study by Erhardt et al (3) and was nearer to the intake of the low-fiber group. The fat intake of our subjects was intermediate to that in

the high- and low-fat diets studied by Erhardt et al (3), as was carbohydrate intake. Clarification of such interactions between major dietary components, iron concentrations in the feces, and free radical production will require more detailed dietary analysis with a larger number of subjects. Graf and Eaton (19) suggested that the putative protective effects of ber-rich foods against colorectal cancer are not necessarily due to the carbohydrate constituents of cell walls, but to the associated phytate content of plant cells. We attempted to explore this hypothesis indirectly by using the information on phytate content available in the computerized food database used to analyze the food diaries. No relation was found between estimated habitual phytate intake and free radicalgenerating capacity, but the program indicated that the reliability of the data on the phytate content of foods was low; thus, the ndings cannot be regarded as denitive. A recent in vitro study by Lu et al (20) showed that lignin can act as a free radical scavenger, yet we found no reduction in free radical generation with increasing concentrations of dietary lignin. The present results clearly establish that orally administered iron increases the rate at which reactive oxygen species are generated within fecal material in an artificial in vitro system. To what extent do these results provide information relevant to the intraluminal environment in vivo? The first obvious difference between the 2 systems is that the colonic lumen is predominantly anaerobic whereas atmospheric oxygen is available in the in vitro assay system. However, the existence of a microclimate near the intestinal mucosal surface is widely recognized, and it seems probable that, as Babbs (7) proposed, sufficient oxygen tension is present at the mucosal surface to enable the production of reactive oxygen species. In rats, intracolonic oxygen tension was reported to be 11.1 mm Hg as determined by mass spectrometry (21), and in a more recent study the mucosal oxygen tension measured by a surface probe was 9 mm Hg compared with 65 mm Hg in serosal tissue (22). The pH of the colonic contents may also affect free radical production. In the proximal human colon, the pH of the bulk phase varies between 5.5 and 7 depending on the rate of fermentation of carbohydrate. In animal models a microclimate has been shown to exist in the colon (23) such that the pH at the mucosal surface is buffered close to 7.0, the pH used in the in vitro incubations in the present study.

FECAL FREE RADICALS AND ORAL IRON The reactions observed in vitro are also dependent on the presence of small organic molecules that are not present in vivo. In our preliminary experiments, no detectable free radical generation was found unless EDTA was added to the incubation medium (data not shown). However, it has been shown that heme breakdown products such as bilirubin and biliverdin, which are known to be present in the lumen of the large intestine, can promote iron-induced free radical generation in vitro (7). Thus, the conditions necessary for the generation of oxygen free radicals may well exist in the proximal colon, which is where epidemiologic evidence for an association between iron intake and an increased risk of bowel cancer has been observed (2). In view of the widespread use of iron supplements and the fortification of foods as a prophylactic measure against iron deficiency disorders, further studies on the significance of free radical production in the human fecal stream seem warranted. REFERENCES
1. Willett WC, Stampfer JM, Colditz GA, Rosner BA, Speizer FE. Relation of meat, fat, and ber intake to the risk of colon cancer in a prospective study among women. N Engl J Med 1990;323: 166472. 2. Wurzemann JI, Silver A, Schreinemachers DM, Sandler RS, Everson RB. Iron intake and risk of colon cancer. Cancer Epidemiol Biomarkers Prev 1996;5:5037. 3. Erhardt JG, Lim SS, Bode JC, Bode C. A diet rich in fat and poor in dietary fiber increases the in-vitro formation of reactive oxygen species in human feces. J Nutr 1997;127:7069. 4. Knekt P, Reunanen A, Takkunen H, Aromaa A, Heli-vaara M, Hakulinen T. Body iron stores and risk of cancer. Int J Cancer 1994;56: 37982. 5. Nelson RL, Davis F, Sutter E, Sobin LH, Kikendall JW, Bowen P. Body iron stores and risk of colonic neoplasia. J Natl Cancer Inst 1994;86:45560. 6. Blakeborough MH, Owen RW, Bilton RF. Free radical generating mechanisms in the colon: their role in the induction and promotion of colorectal cancer. Free Radic Res Commun 1989;6:35967. 7. Babbs CF. Free radicals and the etiology of colon cancer. Free Radic Biol Med 1990;8:191200. 8. Graf E, Eaton JW. Dietary suppression of colonic cancerfiber or phytate? Cancer 1985;56:7178. 9. Buttke TM, Sandstrom PA. Oxidative stress as a mediator of apo-

255

ptosis. Immunol Today 1994;15:710. 10. Nelson RL, Yoo JC, Tanure JC, Andrianopoulos G, Misumi A. Effect of iron on experimental colorectal carcinogenesis. Anticancer Res 1989;9:147782. 11. Thompson LU, Zhang L. Phytic acid and minerals: effect on early markers of risk for mammary and colon carcinogenesis. Carcinogenesis 1991;12:20415. 12. Lund EK, Wharf SG, Fairweather-Tait SJ, Johnson IT. Increases in the concentrations of available iron in response to dietary iron supplementation are associated with changes in crypt cell proliferation in rat large intestine. J Nutr 1998;128:1759. 13. Simpson RS, Sidhar S, Peters TJ. Application of selective extraction to the study of iron species present in diet and rat gastrointestinal tract contents. Br J Nutr 1992;67:43744. 14. Schwartz S, Dahl J, Ellefson M, Ahlquist D. The HemoQuant test: a specific and quantitative determination of heme (hemoglobin) in feces and other materials. Clin Chem 1983;29:20617. 15. Babbs CF, Gale MJ. Colorimetric assay for methanesulfinic acid in biological samples. Anal Biochem 1987;163:6773. 16. Richmond HG. Induction of sarcoma in the rat by iron dextran complex. Br Med J 1959;1:9479. 17. Sahu SC. Dietary iron and cancer: a review. Environ Carcinog Ecotox Rev 1992;10:20537. 18. Weinberg ED. The role of iron in cancer. Eur J Cancer Prev 1996; 5:1936. 19. Graf E, Eaton JW. Suppression of colonic cancer by dietary phytic acid. Nutr Cancer 1993;19:119. 20. Lu F-J, Chu L-H, Gau R-J. Free radical-scavenging properties of lignin. Nutr Cancer 1998;30:318. 21. Bornside GH, Donovan WE, Meyers MD. Intracolonic tensions of oxygen and carbon dioxide in germfree, conventional, and gnotobiotic rats. Proc Soc Exp Biol Med 1976;151:43741. 22. Zabel DD, Hopf HW, Hunt TK. The role of nitric oxide in subcutaneous and transmural gut tissue oxygenation. Shock 1996;5:3413. 23. Rechkemmer G, Wahl M, Kuschinsky W, von Engelhardt W. pHmicroclimate at the luminal surface of the intestinal mucosa of guinea pig and rat. Pflugers Arch 1986;401:3340.