Minute amounts of dietary proteins permeate the gut of healthy infants (1,2) and, less frequently, of adults (3,4). Increased permeability to proteins has been shown in celiac disease (5), in infectious diarrhea (6,7), in food allergy (8-10), immunoglobulin A (IgA) deficiency (11), and infantile colic (12).
Patients with celiac disease show decreased absorption of monosaccharides (13-15) and polyethylene glycols (15,16) and increased permeability of disaccharides (13-15) and 51Cr-EDTA (17-19). Dual sugar tests, especially those using a disaccharide/monosaccharide ratio, have been suggested as suitable screening tests for celiac disease (20,21). However, the results with different probes are contradictory, and permeability to sugar molecules does not necessarily reflect protein permeability (22). Although the exact pathogenesis of celiac disease is unknown, there is an inappropriate immunologic reaction against gliadin fractions of cereals in genetically susceptible individuals. A prerequisite for this immunologic response is that the antigenic fraction of gluten will permeate the gut wall to come into contact with immunocompetent cells, and uptake of protein antigens may have a pathogenetic role in celiac disease (23).
Our aim was to assess gut permeability to human α-lactalbumin (ALA), bovine β-lactoglobulin (BLG), mannitol, and lactulose in celiac disease. We specifically wanted to investigate the relationship between protein and sugar permeability.
MATERIALS AND METHODS
Forty-six permeability tests were performed on 33 patients with celiac disease. These were consecutive patients diagnosed at the Children's Hospital, University of Helsinki, or at Jorvi County Hospital, Espoo, Finland. The mean (range) age of the patients was 10.2 (1.4-19.4) years. A peroral small bowel biopsy was performed on all 22 newly diagnosed child patients because of suspected celiac disease, which was confirmed by this biopsy, on 17 patients on a gluten-free diet to show normalization of the gut mucosa, and on seven patients after a gluten challenge to verify the diagnosis of celiac disease according to current criteria (24). On each occasion a gastrointestinal permeability test was performed the following day in a research ward at either hospital as follows: after an overnight fast, patients provided a urine sample and received a solution by mouth of 4 g mannitol and 8 g lactulose in 100 ml of purified water (osmolality 680 mOsm); the amount given was 2 ml/kg body weight, not exceeding 100 ml. All urine during the next 5 h was collected. Two hours after drinking the sugar solution the children received by mouth a measured dose of milk containing equal amounts of human milk and cow's milk (CM); a venous blood sample was drawn 1 h later. We performed the same permeability test for the sugar probes on 18 healthy adults with a mean age of 26 years.
Concentrations of ALA and BLG were measured in serum by means of a sensitive solid-phase double-sandwich fluoroimmunoassay (2), with microplates coated with polyclonal anti-ALA or BLG. Levels are expressed in relation to the amount of milk protein received per meal as micrograms of ALA or BLG/liters of serum/grams of ALA or BLG given/kilograms of body weight. The assay is similar to that used by Paganelli and Levinsky (3), where antigens can be detected both when free and when in immune complexes. Samples were analyzed in duplicate. The sensitivity of the assay was 0.25 μg/l for ALA and 1.0 μg/l for BLG. The coefficient of variation (CV) within assays was 3-10% and between assays 3-24%. Mannitol and lactulose permeabilities were assessed by measuring their recovery in urine by subtracting the starting level from the 5 h collection level; the result is given as percentage of the oral dose administered. For measurement of mannitol by gas-liquid chromatography (25), urine samples were desalted by Duolite mixed resin (BDH Ltd., Dorset, U.K.) and centrifuged at 2,000 g and supernatants were vacuum dried. Drying with nitric gas and derivatization with bistrifluoroacetamide-trimethylchlorosilane-pyridine was done before analysis; CV between assays was 15%. Lactulose was measured enzymatically (26); CV between assays was 6%. Gliadin (27) and CM antibodies (CMAb) (28) were measured with an enzyme-linked immunosorbent assay (ELISA) coated with wheat gliadin (BDH) and with CM formula (Valio, Finland). Results were compared with a positive reference serum and expressed as percentages. CV was 10% for gliadin antibodies, 5% for IgG, and 13% for IgA CMAbs. Reticulin antibodies (IgA) were determined by the indirect immunofluorescence technique (29) using rat kidney as the antigenic substrate. Definition of the jejunal biopsy finding was according to Kuitunen et al. (30); subtotal villous atrophy (SVA), villous height < 150 μm; partial villous atrophy (PVA), villous height 150-200 μm; and normal, villous height > 300 μm.
We compared differences between groups with the Mann-Whitney U test because of the skewed distribution of data; results are given as two-tailed p values. We sought associations between data with Spearman's rank correlation test; results are given as correlation coefficient Rho and two-tailed p value, corrected for ties; p < 0.05 was considered statistically significant.
The study was approved by the Ethics Committee of each hospital. A small bowel biopsy was performed on all celiac disease patients for initial diagnosis, to show normalization of the gut mucosa on a gluten-free diet or for final diagnosis of celiac disease when on a gluten-provocation diet according to current criteria (24). Informed, written consent was obtained from all parents. The peripheral venous blood sample was small, and the compounds tested are used widely as foods. A good noninvasive permeability test could serve as a follow-up method in patients with gastrointestinal disease and thus be of benefit to many children.
Group characteristics appear in Table 1. Of the celiac disease patients, one had selective IgA deficiency; three, diabetes; and one, nephrotic syndrome. Five patients were tested when newly diagnosed and after 8-20 months on a gluten-free diet, and seven patients with celiac disease were studied on a gluten-free diet and after 3 months on gluten provocation, or sooner if symptoms reappeared (median 2.6 months); one teenager was studied after being on gluten provocation for 9 months because of personal reasons. The remaining 21 patients were tested only once.
Gut Permeability to Human α-Lactalbumin and Bovine β-Lactoglobulin
ALA was found in 19 of 42 patients tested. Among celiac disease patients with detectable ALA, a greater proportion had SVA or PVA than did patients in whom ALA was not detected (Fig. 1): χ2 = 6.4, p = 0.01.
We compared the absorption of ALA in patients at different dietary phases who showed various degrees of jejunal damage (Fig. 1). Newly diagnosed patients did not differ significantly in their absorption of ALA from those on a gluten-free diet, but after gluten provocation they absorbed significantly more ALA than when on a gluten-free diet, p = 0.01. Absorption of ALA in newly diagnosed patients and those on gluten provocation was similar. Patients with SVA and those with either SVA or PVA absorbed more ALA than did those with a normal jejunal structure, p = 0.02 and 0.03, respectively.
Of 42 patients tested, only four had demonstrable absorption of BLG, three had SVA, and one showed normal jejunal structure.
Gut Permeability to Mannitol and Lactulose
Newly diagnosed patients absorbed less mannitol than those on a gluten-free diet, p = 0.006 (Table 2). Patients on gluten provocation absorbed less mannitol than did newly diagnosed patients, those on the gluten-free diet, or controls: p = 0.009, 0.001, and 0.005, respectively. Patients with either SVA or PVA absorbed less mannitol than those with normal histology:p = 0.001 and 0.006, respectively. Patients with normal jejunal biopsy absorbed more mannitol than did controls, p = 0.04, but those with villous atrophy did not differ significantly from controls. Newly diagnosed patients absorbed more lactulose than did other groups, but the difference was significant only when compared to controls, p = 0.007. Patients showing SVA on jejunal biopsy absorbed more lactulose than did patients with normal histology and controls; the latter was statistically significant, p = 0.029. The lactulose/mannitol (L/M) ratio was significantly higher in newly diagnosed patients than in controls and in patients on the gluten-free diet, p = 0.002 and 0.04, respectively. Patients on the gluten-provocation diet also had a higher L/M ratio than controls, p = 0.04, and than patients on the gluten-free diet, but this latter difference was not statistically significant. Patients with either SVA or PVA had a higher L/M ratio than controls (Fig. 2), and the L/M ratios in patients with normal jejunal biopsy and in controls were similar.
Relationship between Gut Permeability to ALA and Mannitol and Lactulose
We made a correlation analysis between the absorption of ALA and of mannitol (Fig. 3). In all patients, the correlation coefficient Rho was -0.36, p = 0.02, a finding caused by two patients with especially high permeability to ALA and low permeability to mannitol and thus not applicable to the majority of patients. When patients were divided into groups according to disease state or biopsy finding, no statistically significant correlation could be found. Nor was any statistically significant correlation found between absorption of ALA and of lactulose. In the correlation analysis between the ALA and L/M ratios in all patients, Rho was 0.31, p = 0.05. This figure also resulted from the two patients with high absorption of ALA and low absorption of mannitol and is not representative of the whole group. When patients were divided into groups according to disease state or biopsy finding, no statistically significant correlation was found.
Relationship between ALA Permeability and Gliadin, Reticulin, and CM Antibodies
Patients with detectable ALA after a milk feed had positive IgA or IgG gliadin antibodies more often than did those in whom no ALA was detected. The frequencies of reticulin antibodies observed in patients with detectable or undetectable ALA did not significantly differ from expected frequencies (Table 3). Median levels of CM antibodies (percentages of positive reference) in children with and without detectable permeability to ALA were 15.1 and 4.8 for IgA and 7.9 and 5.1 for IgG, respectively. The correlation between permeability to ALA and CMAbs was weak, Rho = 0.34, p = 0.03 for IgA; for IgG there was no correlation.
Permeability Results in Celiac Disease Patients Studied in Two Disease States
The L/M ratio became significantly lower (p = 0.03, Wilcoxon signed rank test) and the absorption of ALA decreased when newly diagnosed patients were on the gluten-free diet. During gluten provocation, five of seven patients showed an increase in L/M ratio and in four of five the absorption of ALA increased (Fig. 4).
This study shows that permeability to protein macromolecules occurs in celiac disease and that permeability is related both to disease state and to jejunal histologic findings. Patients with detectable ALA had villous atrophy more often, and a significantly higher level of ALA was found in those with villous atrophy than in those with normal jejunal histology. Our findings are in agreement with those of Husby et al. (5), but Pitcher-Wilmott et al. (31) found no abnormal protein permeability in celiac disease. Human ALA and bovine BLG were detected in 19 (45%) and four (10%) of our 42 subjects, respectively. We did not measure protein permeability in healthy controls, but in an earlier study we detected ALA in only one of 20 healthy infants older than 8 months (2,32).
We have here demonstrated abnormal uptake of mannitol and lactulose in celiac disease patients with villous atrophy; mannitol absorption was decreased and lactulose increased. However, the relationship between permeability to protein macromolecules and to sugar probes was weak. Lactulose is thought to be absorbed intercellularly and its absorption to be augmented in mucosal damage (13,20). Permeability to protein is transcellular and occurs by endocytosis (33). In this study the absorption of ALA was increased in mucosal damage, but because there was no correlation with lactulose absorption, we speculate that ALA is absorbed by a different route than is lactulose. Jalonen et al. (6) found, in rotavirus gastroenteritis, a correlation between permeability to BLG and L/M ratio and an inverse correlation between BLG and mannitol but no correlation for lactulose. They and others (34) suggest that mannitol is absorbed transcellularly because its uptake diminishes in villous atrophy when the absorption area diminishes. Villous atrophy could also have changed the proportion of immature cells in the intestinal epithelium, leading to increased endocytosis (35) and increased BLG absorption; both mannitol and proteins would then be absorbed transcellularly, only the mechanisms being different. Accordingly, sugar molecules do not reflect protein permeability, and they should be used cautiously to interpret pathogenetic phenomena since in celiac disease it is the protein antigen gliadin and fractions of gliadin that are involved in immunologic reactions leading to mucosal damage of the small intestine. It is unlikely that increased gut permeability is a primary factor in the development of celiac disease because permeability changes normalize on a gluten-free diet and upon gluten challenge increase again (21), which also was shown in this study. Moreover, the L/M ratios in celiac disease patients with normal jejunal biopsy and in controls were similar.
Since less than half of our patients had ALA in their sera, the ALA test could not be used to screen for celiac disease; due to overlapping permeability results in different stages of celiac disease, the sensitivity of a detectable ALA for predicting villous atrophy was only 55% and its specificity was 77%. It was thus not a good test for predicting the timing of a new biopsy. Mannitol or lactulose values did not distinguish strongly enough patients from controls nor celiac disease patients with normal histology from those with villous atrophy. The L/M recovery ratio was significantly higher in newly diagnosed patients than in controls, but overlapping recovery values in the two groups were many. Our opinion, shared by others (36), is that sugar absorption tests using mannitol and lactulose cannot be used as screening tests for celiac disease or in the follow-up of the diagnosed patient for assessment of diet adherence or for timing of a new jejunal biopsy, as has been suggested (20,21). However, according to new diagnostic criteria (24), a gluten challenge is needed today only when there is doubt about the initial diagnosis or when the biopsy was inadequate or uncharacteristic.
An antibody response requires permeation of the protein through the gut epithelium and contact with the immune system of the gut. Patients with detectable permeability to ALA more frequently had gliadin antibodies, implying mucosal damage or a relation between increased permeability and level of gliadin antibodies. However, we have shown earlier (37) that there is no direct association between increased permeability to protein and level of serum antibodies to food protein. Variability in immune response is the probable reason for differences in serum antibody responses between individuals.
We conclude that permeability to proteins and sugar molecules is abnormal in celiac disease patients with mucosal damage but that the mechanisms of penetration may be different.
Acknowledgment: We thank Pekka Kuitunen, M.D., for performing the biopsies and coordinating the permeability tests in Jorvi Hospital as well as for the evaluation of the jejunal biopsy specimens and Annikki Sarnesto, Ph.D., nurses Maija-Liisa Nisula and Taina Keituri, and medical laboratory technologists Terttu Louhio and Sirkku Kristiansen for their assistance. Support was by Finska Läkaresällskapet, the Sigrid Jusélius Foundation, and the University of Helsinki, Finland.
1. Jakobsson I, Lindberg T, Lothe L, Axelsson I, Benediktsson B. Human α-lactalbumin as a marker of macromolecular absorption. Gut
2. Kuitunen M, Savilahti E, Sarnesto A. Human α-lactalbumin and bovine β-lactoglobulin absorption in infants. Allergy
3. Paganelli R, Levinsky RJ. Solid phase radioimmunoassay for detection of circulating food protein antigens in human serum. J Immunol Meth
4. Husby S, Jensenius JC, Svehag S-E. Passage of dietary antigen into the blood of healthy adults. Quantification, estimation of size distribution, and relation of uptake to levels of specific antibodies. Scand J Immunol
5. Husby S, Foged N, Höst A, Svehag S-E. Passage of dietary antigens into the blood of children with coeliac disease. Quantification and size distribution of absorbed antigens. Gut
6. Jalonen T, Isolauri E, Heyman M, Craindenoyelle AM, Sillanaukee P, Koivula T. Increased beta-lactoglobulin absorption during rotavirus enteritis in infant—relationship to sugar permeability. Pediatr Res
7. Holm S, Andersson Y, Gothefors L, Lindberg T. Increased protein absorption after acute gastroenteritis in children. Acta Paediatr
8. Dannaeus A, Inganäs M, Johansson SGO, Foucard T. Intestinal uptake of ovalbumin in malabsorption and food allergy in relation to serum IgG antibody and orally administered sodium cromoglycate. Clin Allergy
9. Heyman M, Grasset E, Ducroc R, Desjeux J-F. Antigen absorption by the jejunal epithelium of children with cow's milk allergy. Pediatr Res
10. Juvonen P, Jakobsson I, Lindberg T. Macromolecular absorption and cows' milk allergy. Arch Dis Child
11. Cunningham-Rundles C, Brandeis WE, Good RA. Bovine antigens and the formation of circulating immune complexes in selective immunoglobulin A deficiency. J Clin Invest
12. Lothe L, Lindberg T, Jakobsson I. Macromolecular absorption in infants with infantile colic. Acta Paediatr Scand
13. Menzies IS, Laker MF, Pounder R, et al. Abnormal intestinal permeability to sugars in villous atrophy. Lancet
14. Hamilton I, Cobden I, Rothwell J, Axon ATR. Intestinal permeability in celiac disease: the response to gluten withdrawal and single-dose gluten challenge. Gut
15. Ukabam SO, Cooper BT. Small intestinal permeability to mannitol, lactulose, and polyethylene glycol 400 in celiac disease. Dig Dis Sci
16. Stenhammar L, Fälth-Magnusson K, Jansson G, Magnusson K-E, Sundqvist T. Intestinal permeability to inert sugars and different-sized polyethyleneglycols in children with celiac disease. J Pediatr Gastroenterol Nutr
17. Bjarnason I, Peters TJ, Veall N. A persistent defect in intestinal permeability in coeliac disease demonstrated by a 51
Cr-labelled EDTA absorption test. Lancet
18. Turck D, Ythier H, Maquet E, et al. Intestinal permeability to [51Cr]EDTA in children with Crohn's disease and celiac disease. J Pediatr Gastroenterol Nutr
19. Bjarnason I, Maxton D, Reynolds AP, Catt C, Peters TJ, Menzies IS. Comparison of four markers of intestinal permeability in control subjects and patients with coeliac disease. Scand J Gastroenterol
20. Juby LD, Rothwell J, Axon ATR. Lactulose/mannitol test: an ideal screen for celiac disease. Gastroenterology
21. van Elburg RM, Uil JJ, DeMonchy GR, Heymans HSA. Intestinal permeability in pediatric gastroenterology. Scand J Gastroenterol
22. Weaver LT, Coombs RRA. Does “sugar” permeability reflect macromolecular absorption? A comparison of the gastro-intestinal uptake of lactulose and beta-lactoglobulin in the neonatal guinea pig. Int Arch Allergy Appl Immunol
23. Walker WA, Isselbacher KJ. Uptake and transport of macromolecules by the intestine. Possible role in clinical disorders. Gastroenterology
24. Walker-Smith JA, Guandalini S, Schmitz J, Shmerling DH, Visakorpi JK. Revised criteria for diagnosis of coeliac disease. Working Group of European Society for Paediatric Gastroenterology and Nutrition. Arch Dis Child
25. Laker MF. Estimation of disaccharides in plasma and urine by gas-liquid chromatography. J Chromatogr
26. Geier H, Klostermeyer H. Enzymatische bestimmung von lactulose. Z Lebensm Unters Forsch
27. Savilahti E, Perkkiö M, Kalimo K, Viander M, Vainio E, Reunala T. IgA antigliadin antibodies: a marker of mucosal damage in childhood coeliac disease. Lancet
28. Tainio V-M, Savilahti E, Arjomaa P, Salmenperä L, Perheentupa J, Siimes MA. Plasma antibodies to cow's milk are increased by early weaning and consumption of unmodified milk, but production of plasma IgA and IgM cow's milk antibodies is stimulated even during exclusive breast-feeding. Acta Paediatr Scand
29. Mäki M, Hällström O, Vesikari T, Visakorpi JK. Evaluation of a serum IgA-class reticulin antibody test for the detection of childhood coeliac disease. J Pediatr
30. Kuitunen P, Kosnai I, Savilahti E. Morphometric study of the jejunal mucosa in various childhood enteropathies with special reference to intraepithelial lymphocytes. J Pediatr Gastroenterol Nutr
31. Pitcher-Wilmott RW, Booth I, Harries J, Levinsky RJ. Intestinal absorption of food antigens in coeliac disease. Arch Dis Child
32. Kuitunen M, Savilahti E, Sarnesto A. Human α-lactalbumin and bovine β-lactoglobulin absorption in premature infants. Pediatr Res
33. Heyman M, Desjeux JF. Significance of intestinal food protein transport. J Pediatr Gastroenterol Nutr
34. Anonymous. Intestinal permeability. Lancet
35. Heyman M, Crain-Denoyelle AM, Desjeux JF. Endocytosis and processing of protein by isolated villus and crypt cells of the mouse small intestine. J Pediatr Gastroenterol Nutr
36. Catassi C, Rossini M, Ratsch IM, et al. Dose dependent effects of protracted ingestion of small amounts of gliadin in coeliac disease children: a clinical and jejunal morphometric study. Gut
37. Kuitunen M, Savilahti E. Mucosal IgA, mucosal cow's milk antibodies, serum cow's milk antibodies and gastrointestinal permeability in infants. Pediatr Allergy Immunol
Keywords:© Lippincott-Raven Publishers
Gut permeability; Human α-lactalbumin; Mannitol; Lactulose