Gluten-Sensitive Disease With Mild Enteropathy

GASTROENTEROLOGY 1996;111:608 616 Gluten-Sensitive Disease With Mild Enteropathy ANTONIO PICARELLI,* LUIGI MAIURI,, MARIA CRISTINA MAZZILLI, x SALVATORE COLETTA, PAOLA FERRANTE, x FABIO DI GIOVAMBATTISTA,* MARCO GRECO,* ALDO TORSOLI,* and SALVATORE AURICCHIO *Cattedra di Gastroenterologia I, Clinica Medica II, University La Sapienza, Rome; Children s Hospital Pausilipon, Naples; Department of Pediatrics, University Federico II, Naples; and x Department of Experimental Medicine, University La Sapienza, Rome, Italy lesions, together with other markers, has been proposed to identify patients with gluten sensitivity as potential or latent celiac patients. 3,4 Latent celiac patients are de- fined as patients who have apparently normal jejunal histology on a normal diet but, at some other time before or since, have a gluten-sensitive enteropathy. A number of markers have been proposed to identify these patients, such as increased number of intraepithelial lymphocytes (IELs); increased number of IELs bearing the g/d receptor 5,6 ; presence of serum antiendomysium antibodies 7 ; and high levels of intestinal immunoglobulin (Ig) M antigliadin antibodies, other IgM class antibodies, and IgA antigliadin antibodies in samples of jejunal fluid and whole-gut lavage fluid 8 (the so-called celiac-like intestine antibody pattern). Latent celiac disease is likely to be more represented in first-degree relatives of celiac patients; about half of them were shown to present immunologic indices of potential celiac disease, 9 and a dose effect of celiac HLA- DQA1 and HLA-DQB1 predisposing alleles was found in this population. Recently, Arranz and Ferguson 10 tried to use the ce- liac-like intestine antibody pattern to identify latent celiac disease in a series of patients attending a gastrointes- tinal unit to undergo diagnostic small bowel biopsy. In a few patients with idiopathic diarrhea or iron-deficient anemia, the celiac-like intestine antibody pattern was found in the presence of histologically normal small intestinal mucosa. Five celiac-like intestine antibody posi- tive patients had complete resolution of chronic high- volume diarrhea on a gluten-free diet (3 of them had a high IEL count on a gluten-containing diet that de- creased to normal on a gluten-free diet). No HLA typing was reported for these patients. In this report, we describe 10 patients with suggestive symptoms of celiac disease and with serum antiendomy- Background & Aims: Celiac disease is a permanent gluten intolerance strongly associated with HLA class II antigens and possibly showing milder changes of mucosal architecture. Ten patients with symptoms suggesting celiac disease and serum antiendomysium antibodies with normal mucosal architecture were studied. Methods: Immunohistochemical detection of mucosal immune activation and HLA typings were performed. Results: Mucosal immune activation, with normal mucosal architecture and normal g/d / intraepithelial lym- phocytes counts, was found on a gluten-containing diet. In 3 of 6 patients, multiple biopsy specimens showed one sample with severe villous atrophy. Clinical and immunomorphologic features were strictly gluten dependent. The mucosal immune activation was elicited in vitro by gliadin. Only 4 patients had the typical HLA typing of celiac disease. Conclusions: Gluten-sensitive celiac-like symptoms may occur in patients with serum antiendomysium antibodies, apparently normal intestinal mucosa, and HLA typing not commonly associated with celiac disease. These patients should undergo multiple biopsies, and signs of immunologic acti- vation should be sought accurately; in the presence of mucosal immune activation, a trial with a gluten-free diet should be encouraged to detect gluten dependency. In vitro immunologic response of small intestinal mucosa to gliadin may support the diagnosis of glutensensitive enteropathy. C eliac disease is defined as a permanent intolerance of the small intestine to gluten. 1 It may present with a wide spectrum of clinical symptoms and is also strongly associated to HLA class II antigens. 2 The diagnosis is based on the presence of total or partial villus atrophy with crypt hyperplasia on a gluten-containing diet with improvement in histological appearance after treatment with a gluten-free diet. An abnormal mucosal cell mediated immune response to gliadin is considered the central event in the pathogenesis of celiac disease (reviewed by Sollid and Thorsby 2 ). However, it is not present exclusively in patients with a flat mucosa on gluten-containing diet. Recognition of minor intestinal Abbreviations used in this paper: ICAM-1, intercellular adhesion molecule 1; IEL, intraepithelial lymphocyte; LPMNC, lamina propria mononuclear cell. 1996 by the American Gastroenterological Association 0016-5085/96/$3.00

September 1996 MILD GLUTEN SENSITIVE ENTEROPATHY 609 Serial cryostat sections cut at 3 mm were fixed in ace- tone for 10 minutes and then tested with monoclonal antibodies to different activation markers and adhesion molecule (intercellular adhesion molecule 1 [ICAM-1] [1:400; Dako, Copenhagen, Denmark], interleukin 2 receptor [CD25] [1:30; Dako], B7/BB1 [CD80] [1:40; Becton Dickinson, Mountain View, CA], CD3 [1:200; Dako], HLA-DR [1:20; Becton Dick- Patients and Methods inson]) that have been found strongly expressed in atrophic Patients celiac disease mucosa. 11 14 Immunostaining was performed after the alkaline phosphatase antialkaline phosphatase Ten patients attended a gastrointestinal unit because method as described previously by Maiuri et al. 15 of symptoms suggesting celiac disease: 6 patients presented At least five slides for each sample were evaluated blindly. sium antibodies; histologically normal small intestinal biopsy specimens (with a normal count of IEL [with four exceptions] and a normal number of IELs bearing the g/ d receptor) showed signs of immunologic activation that disappeared after gluten withdrawal and was again evident in 3 patients after a 4-month gluten challenge. Surprisingly, only 4 of these patients had HLA class II antigens compatible with latent celiac disease, whereas the other 6 patients showed various HLA class II antigens not commonly associated with celiac disease. a first-degree relative of patients with celiac disease; all of the patients were negative for serum antiendomysium antibodies. Immunohistochemical Staining with chronic diarrhea and steatorrhea (in 2 patients with iron- The number of the cells bearing ICAM-1, CD25, and CD80 deficient anemia), and the other 4 patients showed relapsing antigens was evaluated in a standard area of 1 mm 2 of lamina hypocalcemic tetany, chronic hepatitis, recurrent aphthous sto- propria. The number of CD3 / cells was counted for 500 enteromatitis, and osteoporosis, respectively. Only 1 of them (patient cytes in each sample and calculated as the percentage of entero- 3) was a first-degree relative of patients with celiac disease. cytes. The number of IELs bearing the g/d receptor was calcu- The patients were investigated several times for the presence lated per millimeter of epithelium as reported previously. 6 of serum antigliadin and antiendomysium antibodies because of the persistence of clinical symptoms; 5 of the patients under- Small intestinal biopsy specimens, taken from patients treated with either normal diet or gluten-free diet, were cul- tured for 24 hours in the presence or absence of a peptictryptic digest of gliadin (1 mg/ml) or of maize prolamin peptides (3 mg/ml) before and after gluten-free diet. The cultures took place as described previously. 16 The immune response to gliadin was assessed by detecting in lamina propria mononuclear cells (LPMNCs), activation markers, and adhesion molecules (ICAM-1, CD25, and CD80); we have found that all these markers increased in celiac disease in remission on in vitro gliadin challenge, 15 and CD3 / cells were numbered in the mucosal intraepithelial compartment. The expression and distribution of the immunologic markers were assessed by immunohistochemistry as it was for the biopsy specimens before in vitro manipulation. went small intestinal biopsies at different times of the clinical history. The biopsy specimens were embedded in paraffin and showed a normal mucosal architecture. All patients at a certain stage of their illness developed serum antiendomysium anti- bodies; consequently, a further small intestinal biopsy at duo- denal-jejunal flexure was performed. Samples were embedded in OCT compound (Tissue Tek; Miles Laboratories, Elkhart, IN), frozen in liquid nitrogen, and kept at 070 C until cut. Moreover, 6 of these patients underwent multiple small intesti- nal biopsies around the Treitz region during endoscopic exami- nation. In all patients, small intestinal biopsies were also per- formed after 6 12 months of gluten-free diet; 3 of the patients were rechallenged with gluten for 4 months, and intestinal biopsies were performed as soon as clinical symptoms and serum antiendomysium antibodies reappeared. The other pa- tients did not give permission for gliadin challenge because they were concerned about the danger of worsening their rather good clinical condition that had been reached after several years of severe illness. Controls In Vitro Challenge With a Gliadin Digest Determination of Fecal Fat, Nitrogen, and Water Surgical biopsy specimens were taken from the proxi- mal jejunum of 7 consenting adult controls who underwent intestinal resection for biliary tract disease, gastric cancer (3 of the 7 patients), gastric ulcer (2 patients), or pancreatic disease, respectively. None of them was administered cytostatic or local antibiotic drugs. All samples were obtained far from the intestinal lesion and appeared histologically normal. Peroral biopsy specimens were obtained from the duodenal-jejunal flexure of 10 consenting adult controls who attended the gastrointestinal unit because of esophagitis (6 patients) or chronic nonspecific diarrhea (4 patients). None of the 10 patients was Quantitative determination of fecal fat, nitrogen, and water was performed on a 24-hour stool collection by a near infrared reflectance analysis. 17 Urinary Xylose Excretion Urinary xylose excretion was performed by means of a commercial xylose test kit (Roche, Milan, Italy) with bromo- aniline method. Patients were administered 25 g of D-xylose dissolved in 500 ml of water. Urinary samples were collected for 5 hours, and the total volume was measured. Pentoses reacted with 4-bromoaniline to produce a red-brown dye stuff. The color intensity was measured photometrically.

610 PICARELLI ET AL. GASTROENTEROLOGY Vol. 111, No. 3 Serum Antigliadin and Antiendomysium in 3 of the patients, all of the biopsy specimens were mor- Antibodies phologically normal, whereas in patients 2, 3, and 8, three of Antigliadin antibodies, IgA, IgG, and IgE classes, were four fragments showed normal histology, and one fragment determined by the in-house enzyme-linked immunosorbent showed severe villous atrophy and crypt hyperplasia as assay technique as described previously. 18,19 found in the classical, untreated patients with celiac disease. Antiendomysium antibodies detection was sought in sera In all of the patients on the gluten-containing diet diluted to 1:5 using an immunofluorescence commercial kit and in all biopsy specimens, the expression in LPMNCs (Eurospital Pharma, Trieste, Italy). Further dilutions were also of ICAM-1 (Figures 1 and 2A), CD25 (Figure 1), and performed, and the results were compared with those obtained CD80 (Figure 1) was significantly greater than in controls with a pool of sera of classical, untreated patients with celiac (Figure 1 and 2B) (Põ0.001); in 4 of 10 patients disease. (patients 3 5 and 8), HLA-DR was expressed in crypt HLA Typing epithelial cells, but there was no observed increase in the number of IELs bearing the g/d receptor (median, 1.78 The patients were typed for HLA-DRB1, -DQA1, and per millimeter of epithelium; range, 0.8 2.8 per milli- -DQB1 genes by polymerase chain reaction using sequence- meter of epithelium; vs. controls: median, 2.1 per millispecific primers. 20,21 DNA was extracted from fresh or frozen meter of epithelium; range, 1.1 3.2 per millimeter of peripheral blood using a salt-chloroform method described epithelium; P Å NS) (Table 2). After gluten withdrawal, previously. 22 The typing was performed using a DR and DQ the intestinal mucosa from all patients showed the disapsequence-specific primers kit (Dynal A.S., Oslo, Norway) and pearance of any feature of immunologic activation; in following the recommended procedures. The polymerase chain fact, the number of LPMNCs expressing ICAM-1 molereaction products were loaded on a 2% agarose gel and electrocules was reduced significantly with respect to that obphoresed in the presence of ethidium bromide. After visualizaserved during a gluten-containing diet. Moreover, CD25 tion on a UV transilluminator, a photograph was taken as a record. HLA typing came out according to the presence or and CD80 expression was almost absent, and the number absence of specific amplification products in each lane. of IELs was õ30 of 100 enterocytes in all patients (Table 2). The features of immunologic activation were again Statistical Analysis detectable after 4 months of in vivo gluten challenge in Student s two-tailed t test was used for comparing the the 3 patients for whom gluten challenge was accepted. specimens from patients with those from controls before culand The number of LPMNCs that expressed ICAM-1, CD25, ture, as well as for comparing the biopsy specimens exposed CD80 was similar to that found before gluten-free to gliadin with those exposed to media alone. The application diet (Figure 1). By that time, the number of IELs was of nonparametric tests (Wilcoxon two-tailed test) provided increased only in patient 8, who showed high IEL counts results concordant with those obtained by means of parametric before gluten withdrawal. tests. Results In Vitro Challenge With Gliadin Peptides Clinical Features and Laboratory Findings In all patients on gluten-free diet, 24 hours after in vitro gliadin challenge, the intestinal samples showed The details of clinical history, symptoms, and labo- remarkable modifications that mirrored the pattern obratory findings of the patients are shown in Tables 1 and served before the gluten withdrawal, i.e., expression of 2. None of the patients reported any history of allergy CD25 and CD80 in LPMNCs and increased expression and did not present with skin symptoms; none of the of ICAM-1 molecules in LPMNCs, as well as in the patients had high levels of serum IgE or antigliadin IgE extracellular matrix (Figure 3A). Such in vitro immune antibodies. response is very similar to that observed in patients with classical celiac disease in remission, as described pre- Histological and Immunohistochemical viously by Maiuri et al. 15 ; however, differently from those Findings patients, in none of the patients reported here, gliadin Histological and immunohistochemical results are challenge induced infiltration of the surface epithelium shown in Table 2 and in Figures 1 3. In all 10 patients, by CD3 / cells (median, 22.5 per 100 enterocytes; range, both before and after gluten withdrawal, the intestinal mordigest; 19 31 per 100 enterocytes after culture with gliadin phology was normal with a villous height/crypt depth ratio vs. median, 24.5 per 100 enterocytes; range, 18 of ú3. In 4 patients (patients 3 5 and 8), the number of 33 per 100 enterocytes after culture with medium alone; IELs was ú40 of 100 enterocytes. In 6 of these patients, P Å NS) even in the patients showing high IEL counts three more biopsy specimens were taken from around the while on the gluten-containing diet. Treitz region during endoscopy on a gluten-containing diet; A similar behavior on in vitro gliadin challenge was

September 1996 MILD GLUTEN SENSITIVE ENTEROPATHY 611 Table 1. Clinical History and Symptoms Patient 1 Patient 2 Patient 3 Patient 4 Patient 5 Patient 6 Patient 7 Patient 8 Patient 9 Patient 10 20/M 32/F 28/F 37/M 45/M 28/M 28/F 53/F 60/F 71/F Normal diet Antiendomysium antibodies Present Present Present Present Present Present Present Present Present Present Symptoms Delayed growth, Since 2 yr: Since 7 yr: Since 2 yr: Since 10 yr: Since 11 yr: Since 2 yr: Since 5 yr: chronic Since 10 yr: Since 2 yr: hyposideremia relapsing recurrent noninfectious chronic chronic recurrent diarrhea, chronic diarrhea, osteoporosis, as a child; vomiting, aphthous chronic hepatitis, diarrhea, severe diarrhea, hypoglycemia steatorrhea, weight loss, fatigue, and since 5 yr: abdominal stomatitis, elevated diarrhea (up to dyspepsia, (every 10 20 weight loss, and dyspepsia, and weight loss chronic pain, weakness, transaminase 15 stools/day weight loss days), anemia anemia of 10 kg diarrhea, headache, and values (ALT, since 1 yr), and of 10 kg fatigue, steatorrhea, hypocalcemic dyspepsia seven times steatorrhea and stomatitis, and recurrent tetany (once normal; AST, four recurrent paresthesiae, abdominal monthly), times normal), aphthous recurrent pain stomatitis, diarrhea, and stomatitis; abdominal ecchymosis, steatorrhea; liver since 1 yr: pain, paraesthesiae, biopsy: mild recurrent diarrhea, and bone pain, chronic hepatitis abdominal cheilitis infertility, and pain weight loss of 3kg Gluten-free diet Antiendomysium antibodies Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Symptoms No clinical No clinical No clinical Normal Disappearance of No clinical No clinical No clinical No clinical Weight increase symptoms symptoms symptoms transaminase diarrhea symptoms symptoms symptoms and symptoms and of 7 kg and weight and weight values and and weight and weight weight increase weight increase increase of 3 increase disappearance of increase of increase of 5 of 5 kg of 3 kg kg of 5 kg diarrhea 5 kg kg Gluten challenge Antiendomysium antibodies Present Present Present Symptoms Fatigue, Hypocalcemic Diarrhea, weight diarrhea, tetany (two loss of 3 kg, steatorrhea, monthly in the and anemia and weight last 2 mo); loss of 1.5 kg and weight loss of 2 kg, diarrhea, and steatorrhea AST, aspartate aminotransferase.

612 PICARELLI ET AL. GASTROENTEROLOGY Vol. 111, No. 3 Table 2. HLA Typing and Histological and Laboratory Findings Morphology g/d / cells IEL (CD3 / cells/ (per millimeter 100 enterocytes) epithelium) Antigliadin Xylose HLA type antibodies Stools (g/24 h) (% of ingested xylose) Normal Guten-free Normal Gluten- Patient DR DQ diet diet diet free diet IgA a IgG a IgE a Fat b Nitrogen c Water d Urinary excretion e 1 5,7 A1*0201/0501 27 21 1.8 1.3 1.18 0.93 0.57 5.68 7.09 248 17.05 B1*0201/0301 2 1,2 A1*0101/0201 28 24 0.9 1.4 0.84 0.71 0.6 5.36 7.86 252.7 22.9 B1*0501/0502 3 5,6 A1*0103/0501 52 27 2.6 3.1 0.91 0.4 0.4 2.21 7.87 280 15.7 B1*0301/0603 4 4 A1*0302 46 23 2.8 3.1 0.6 0.54 0.6 5.56 9.78 335 51.58 B1*0301/0402 5 2,6 A1*0102/0102 49 23 2.8 3.1 1.13 0.7 0.66 11.84 12.68 604 10.1 B1*0502/0604 6 1,6 A1*0101/0102 32 25 1.9 2.3 1.0 0.73 0.4 1.61 2.25 79.9 32.98 B1*0501/0604 7 6,7 A1*0102/0201 26 29 1.7 1.5 1.94 0.78 0.49 2.28 7.11 316 16.5 B1*0201/0604 8 5,7 A1*0201/0501 47 22 1.4 1.3 1.24 0.92 0.23 8.9 10.9 584 ND B1*0201/0301 9 5,6 A1*0103/0501 23 21 0.8 1.1 0.71 0.56 0.25 8.5 11.3 602 ND B1*0601/0301 10 5 A1*0501 26 20 1.1 1.3 0.63 0.47 0.31 4.1 7.1 80.3 5.24 B1*0301 NOTE. In patients 2, 3, and 8, multiple biopsy specimens showed one fragment with villous atrophy and the others with normal morphology; in the other samples, all the mucosal fragments were morphologically normal. a Optical density: the values express the ratio between patient s optical density and control s optical density { 3SD (normal values: IgG, õ1; IgA, õ0.8; and IgE, õ0.7). b Normal values, 2.3; SD, 1.23. c Normal values, 3.3; SD, 1.67. d Normal values, 125.8; SD, 70.1. e Urinary xylose excretion. Normal values, ú25%. shown by biopsy specimens from all 10 patients before whereas in patients with classical atrophic celiac disease, gluten withdrawal. The number of LPMNCs that ex- antiendomysium antibodies were detected until 1:80 pressed ICAM-1, CD25, and CD80 significantly increased working dilution. Serum antigliadin antibody levels are with respect to the values observed before and after culture shown in Table 2. with medium alone (P õ 0.01) (Figure 3B); also at this time point, a significant increase of the number of CD3 / Fecal Fat, Nitrogen, and Water: Urinary cells infiltrating the surface epithelium was not observed Xylose Excretion (median, 31 per 100 enterocytes; range, 19 50 per 100 enterocytes after culture with gliadin digest; vs. median, Results are shown in Table 2. 32 per 100 enterocytes; range, 20 45 per 100 enterocytes Genetic Markers after culture with medium alone; P Å NS). All the features of the immunologic activation described, both in patients The results are shown in Table 2. Only patients on the normal diet or after gluten withdrawal, were gliadin 1 and 8 carried HLA alleles (DR 5 and 7; DQA1*0501 specific because they did not occur after challenge with and B1*0201) commonly present in patients with celiac maize prolamin peptides (Figure 3A and B) and also disease, whereas the HLA typings of the remaining 8 showed specificity for celiac small intestine because they patients were not typical of celiac disease. Patient 4 was did not occur in control patients without celiac disease DR4 positive, but the patient did not show the who were challenged with gliadin digest (Figure 3C). DQB1*0302 allele usually found in the few DR4-posi- Serum Antiendomysium and Antigliadin tive patients with celiac disease. 23 Antibodies Discussion In all the patients, serum antiendomysium antibodies This report deals with two aspects of gluten sensigas- were detected until 1:32 working dilution, tivity present in a subgroup of patients attending a

September 1996 MILD GLUTEN SENSITIVE ENTEROPATHY 613 Figure 1. Expression of CD25, CD80, and ICAM-1 in LPMNCs in small intestinal mucosa from controls ( ; nå17) and from patients on normal diet ( ;nå10) or gluten-free diet ( ; nå10) and after gluten challenge ( ; nå3). CD25, CD80, and ICAM-1 expression is significantly increased in patients on normal diet (as well as after gluten challenge) with respect to control values (P õ 0.01). No signifi- cant differences were observed between the values in controls and in patients on gluten-free diet. *P õ 0.01 vs. control values and vs. patient values after gluten-free diet; P Å NS vs. control values. biopsy specimens. In patients who underwent multiple biopsy specimens, only one randomly selected biopsy fragment was analyzed at this time point. Moreover, immunocytochemical analysis of these samples showed signs of immunologic activation, such as enhanced expression of ICAM-1, presence of LPMNCs bearing CD25 or CD80, and DR expression in crypts (only in some cases). All of these features have been found in intestinal samples of patients with overt celiac disease with villous atrophy. However, in the intestinal biopsy specimens of the patients described in this report, these features of immunologic activation were less severe. In fact, the number of LPMNCs that expressed the tested immunologic markers were in between those observed in atrophic celiac disease samples (unpublished data) and those observed in non celiac disease controls. In all of these patients, the number of IELs bearing the g/d T- cell receptor was within normal range. All the immunologic features observed in the intestinal biopsy specimens of these patients, as well as the clinical symptoms and serum antiendomysium antibodies, were shown to be strictly gluten dependent. In fact, they disappeared after gluten-free diet and were again detectable after 4 months of gluten challenge in the 3 patients in whom the challenge was accepted. The gluten sensitivity of these patients was further supported by the in vitro gliadin challenge of the intestinal explants, both on a gluten-containing diet and on a gluten-free diet. We have characterized previously the early pathogenetic immunologic events leading to celiac disease in small intestinal biopsy specimens of patients with celiac disease in remission. 15 In fact, we have shown 15 that, in this tissue, gliadin triggers a sequential cascade of immunologic modifications in vitro that lead to a specific immune reaction after 24 hours, reproducing the following various immunologic aspects of the celiac lesion without mucosal damage: increased expression of ICAM-1, CD80 and/or CD25 in LPMNCs, DR expression in crypts, and migration of CD3 / cells in the upper mucosal layers with infiltration of epithelial compartment by not-activated T lymphocytes. In all of the patients described in this report, both on gluten-containing diet and on gluten-free diet, gliadin challenge induced similar immunologic modifications. The only difference observed with respect to the patients with CD in remission who were studied previously was the absence in all of the patients described here of a significant increase of the number of T lymphocytes within the surface epithelium on in vitro gliadin challenge. So far, we do not know whether the absence troenterological unit for symptoms suggesting celiac disease: first, the absence of histological abnormalities in the presence of intestinal and serological immunoreaction to gluten and, second, the lack of association to the HLA dimer DQ (a1*0501 and b1*0201) present in approximately 90% of the celiac disease population. These two aspects could be, but are not necessarily, intercorrelated. In fact, we have identified 10 patients presenting with symptoms that suggest celiac disease who showed normal intestinal morphology at routine histological examination, similarly to those described by Arranz and Ferguson 10 and by Cooper et al. 24 All of these patients were investigated over a long period because of the persistence of clinical symptoms; all developed serum antiendomysium antibodies at a certain stage of their illness. Small intestinal biopsy specimens were then analyzed carefully by morphometry and typed according to Marsh. 25 In fact, it has been shown 25 that at least three distinctive, interrelated patterns of mucosal change define the spectrum of gluten sensitivity and are arbitrarily termed infiltrative, hyperplastic, and destructive. The infiltrative (type 1) lesion comprises normal mucosal architecture in which the villous epithelium is markedly infiltrated by T lymphocytes. In 4 of the patients described in the present report intestinal biopsy specimens showed this morphological pattern because they showed an increased number of IELs with normal mucosal architecture. Neither hyperplastic (type 2) nor destructive (type of this phenomenon is related to the absence of villous 3) mucosal lesions were observed in any of the patients atrophy characteristic of these patients. Therefore, in studied at the microscopic examination of the intestinal these patients, a gluten-dependent (in vivo and in vitro)

614 PICARELLI ET AL. GASTROENTEROLOGY Vol. 111, No. 3 Figure 2. ICAM-1 expression in small intestinal mucosa from (A) patients and (B) controls. (A) ICAM-1 is detected in many LPMNCs (arrow), as well as in the extracellular matrix. (B) ICAM-1 is detected only on endothelial cells and in a few mononuclear cells. Immunohistochemistry (alkaline phosphatase staining technique) is used. The strong staining of the brush border is caused by endogenous alkaline phosphatase (original magnification 1501). Therefore, we conclude that the patients described in this study are affected by mild enteropathy and that the immunohistochemical analysis is essential for the diagnosis of the ongoing immune reaction to gluten. Surprisingly, the HLA class II antigens of these patients were different from those usually found in patients with celiac disease. In fact, celiac disease is primarily associated with a particular DQ a/b heterodimer encoded by DQA1*0501 and DQB1*0201 alleles inherited in cis with DR3 or in trans with DR5/7 haplotypes; in fact, the DQ dimer is present in approximately 90% of patients with celiac disease from all the populations exam- ined. In our population, 13 of 13 full-genotyped patients with celiac disease described previously 15 carried such a heterodimer. In the patients described in the present study, only 2 patients showed HLA typical for celiac disease (both typed as 5 and 7) and may then be considered patients with classical celiac disease with normal mucosal morphology or with patchy mucosal lesion. No immunologic activation of the intestinal mucosa is a constant phenomenon in the absence of morphological abnormalities. We have also analyzed the other mucosal samples in those 6 patients in whom multiple biopsy specimens were available. The features of immunologic activation were present in all of the biopsy specimens. However, in 3 of these 6 patients, an interesting feature was observed; three of four biopsy fragments were morphologically normal (with signs of immunologic activation) as described earlier, and, on the contrary, one fragment showed severe villous atrophy. A patchy lesion is then observed. In our experience, in patients with classical celiac disease, the mucosal lesion may be patchy; however, this event is quite rare in celiac disease. We have observed previously that, in 5 of 25 patients with severe villous atrophy in which multiple biopsy specimens were performed, only one of five biopsy fragments showed normal histology with signs of immunologic activation (unpublished data).

September 1996 MILD GLUTEN SENSITIVE ENTEROPATHY 615 to confer celiac disease susceptibility. Two of the patients described in this study carried such heterodimer; neither patients nor their ancestors came originally from Sar- dinia. Interestingly, 5 of 10 patients typed as DR6, and 3 of the 5 patients carried the DQ (a1*0102 and b1*0604). It appears that this group did not show the HLA typings more common in patients with celiac disease, but a larger sample will be necessary to exclude any association of HLA with the gluten-sensitive disease described in this report. Therefore, the patients we have studied are affected by gluten-sensitive disease because clinical and laboratory findings as well as the signs of immunologic activation of the intestinal mucosa both in vivo and in vitro are gluten dependent. Nevertheless, some features differentiate these patients from those affected by classical celiac disease because they do not show severe enteropathy (consequently, they may be considered latent similarly to the patients described by Arranz and Ferguson 10 ), HLA typical for celiac disease in 6 of 10 of them, or increased density of g/d IELs, with the latter phenomenon being possibly influenced by a celiac disease typical genetic background. 6 Therefore, classical celiac disease is only the top of a hypothetical iceberg; a low-grade or mild gluten-sensitive enteropathy may be not only present in latent patients with celiac disease with celiac HLA class II molecules but also in patients without such a genetic background. Gluten-dependent intestinal mucosal immune activation is common to all these forms of glutensensitive disease, whereas villous atrophy may occur only in some of them in a variable extension along the small intestine. This may be a consequence of some not yet identified phenomena, such as different immune mechanisms or different genetic factors. Troncone and Ferguson have already suggested that there may be separate (genetic) factors that regulate the capacity to gluten sensitization on one hand and a susceptibility to develop severe enteropathy as an expression of mucosal T-cell mediated immunity on the other hand. 27 In conclusion, we have observed that gluten-sensitive celiac-like symptoms may occur in patients who present with serum antiendomysium antibodies and small intestinal mucosa that is apparently normal on routine histological examination; some of these patients may not have HLA typical for celiac disease. In these patients, multiple biopsy specimens should be performed, and signs of immunologic activation should be sought accurately; in the presence of signs of immunologic activation, a trial with a gluten-free diet should be encouraged to verify their gluten dependency. In vitro immunologic response of small intestinal mucosa to gliadin peptides may further support the diagnosis of gluten-sensitive enteropathy. patients were typed as DR3, the most common DR allele in patients with celiac disease, and the only patient who was DR4 positive was DQB1*0302 negative. Recently, Congia et al. 26 reported that Sardinian patients, who are DQ (a1*0501 and b1*0201) negative, may carry an- other heterodimer, DQ (a1*0101 and b1*0501), able Figure 3. Effect of in vitro gluten challenge in control and patient tissues. (A) Patients on gluten-free diet. (B) Patients on normal diet. (C) Controls without celiac disease. Significant differences were observed in the expression of CD25, CD80, and ICAM-1 in patient tis- sues after 24 hours of in vitro gliadin challenge with respect to the values observed before culture or after culture with media alone or with maize prolamin peptides (P õ 0.01). No significant differences were observed in non celiac disease control tissues. *P õ 0.01 vs. values before culture and after culture with medium alone or with maize prolamin peptides. P Å NS vs. values before culture and after culture with medium alone.

616 PICARELLI ET AL. GASTROENTEROLOGY Vol. 111, No. 3 References intestine of celiac patients. Gastroenterology 1996;110:1368 1378. 1. Meuvisse GW. Diagnostic criteria in coeliac disease. Acta Ped 16. Fais S, Maiuri L, Pallone F, De Vincenzi M, De Ritis G, Troncone Scand 1970;59:461 464. R, Auricchio S. Gliadin induced changes in the expression of 2. Sollid LM, Thorsby E. HLA susceptibility genes in celiac disease: MHC class II antigens by human small intestinal epithelium. genetic mapping and role in pathogenesis. Gastroenterology Organ culture studies with coeliac disease mucosa. Gut 1992; 1993;105:910 922. 33:472 478. 3. Ferguson A, Arranz E, O Mahony S. Clinical and pathological spec- 17. Picarelli A, Greco M, Di Giovambattista F, Ramazzotti A, Cedrone trum of coeliac disease active, silent, latent, potential. Gut 1993; C, Corazziari E, Torsoli A. Quantitative determination of foecal fat, 34:150 151. nitrogen and water by reflectance analysis (NIRA). Assessment of 4. Marsh MN. Gluten sensitivity and latency: can patterns of intestiods. Clin Chim Acta 1995;234:146 156. its accuracy and reproducibility compared with chemical methnal antibody secretion define the great silent majority? Gastroenterology 1993;104:1550 1562. 18. Volta U, Lenzi M, Cassani F, Lazzari R, Bianchi FB, Pisi E. Gliadin 5. Maki M, Holm K, Collin P, Savilahti E. Increase in g/d T cell antibodies in coeliac disease. Lancet 1983;1:1285. receptor bearing lymphocytes in normal small bowel mucosa in 19. Bonamico M, Culasso F, Pitzalis G, Mariani P. B2 microglobuline latent coeliac disease. Gut 1991;32:1412 1414. levels in coeliac disease. J Pediatr Gastroenterol Nutr 1990;11: 330 336. 6. Holm K, Maki M, Savilahti E, Lipsanen V, Laippala P, Koskimies 20. Olerup O, Zetterquist H. HLA-DR typing by PCR amplification with S. Intraepithelial g/d T cell receptor lymphocytes and genetic sequence-specific primers (PCR-SSP) in 2 hours: an alternative to susceptibility to coeliac disease. Lancet 1992;339:1500 serological DR typing in clinical practice including donor-recipient 1503. matching in cadaveric transplantation. Tissue Antigens 1992; 7. Collin P, Helin H, Maki M, Hallstrom O, Karvonen Al. Follow-up 39:225 235. of patients positive in reticulin and gliadin antibody tests with 21. Olerup O, Aldener A, Fogdell A. HLA-DQB1 and -DQA1 typing by normal small-bowel biopsy findings. Scand J Gastroenterol 1993; PCR amplification with sequence-specific primers (PCR-SSP) in 2 28:595 598. hours. Tissue Antigens 1993;41:119 134. 8. O Mahony S, Arranz E, Barton JR, Ferguson A. Dissociation be- 22. Mullembach R, Lagoda PJL, Welter C. An efficient salt-chloroform tween systemic and mucosal humoral immune responses in coeextraction of DNA from blood and tissues. Trends Genet 1989; liac disease. Gut 1991;32:29 35. 5:391 397. 9. Holm K, Savilahti E, Koskimies S, Lipsanen V, Maki M. Immuno- 23. Mantovani V, Corazza GR, Bragliani M, Frisoni M, Zaniboni MG, histochemical changes in the jejunum in first-degree relatives of Gasbarrini G. Asp57-negative HLA DQb chain and DQA1*0501 patients with celiac disease and the celiac disease marker DQ allele are essential for the onset of DQw2-positive and DQw2- genes. HLA class II antigen expression, interleukin-2 receptor negative coeliac disease. Clin Exp Immunol 1993;91:153 156. positive cells and dividing crypt cells. Gut 1994;35:55 60. 24. Cooper BT, Holmes GKT, Ferguson R, Thompson RA, Allan RN, 10. Arranz E, Ferguson A. Intestinal antibody pattern of celiac dis- Cooke WT. Gluten-sensitive diarrhea without evidence of celiac ease: occurrence in patients with normal jejunal biopsy histology. disease. Gastroenterology 1980;79:801 806. Gastroenterology 1993;104:1263 1272. 25. Marsh MN. Gluten, major histocompatibility complex and the 11. Brandtzaeg P, Helstensen TS, Hvatum M, Kvale D, Scott H. The small intestine. A molecular and immunobiologic approach to the serological and mucosal immunologic basis of coeliac disease. spectrum of gluten sensitivity (celiac sprue). Gastroenterology In: Walker WA, Harmatz PR, Wershil BK, eds. Immunophysiology 1992;102:330 354. of the gut. Volume 11. London: Academic, 1993;295 333. 26. Congia M, Cucca F, Frau F, De Virgiliis S. A gene dosage effect 12. MacDonald TT. The role of activated T lymphocytes in gastrointes- of the DQA1*0501/DQB1*0201 allelic combination influences tinal disease. Clin Exp Allergy 1990;20:247 252. the clinical heterogeneity of coeliac disease. Hum Immunol 13. Arnaud-Battendier F, Cerf-Bensussan N, Amsellem R, Schmitz J. 1994;40:138 142. Increased HLA-DR expression by enterocytes in children with ceathy 27. Troncone R, Ferguson A. Animal model of gluten induced enteropliac disease. Gastroenterology 1986;91:1206 1212. in mice. Gut 1991;32:871 875. 14. Sturgess RP, Macartney JC, Makgoba MW, Hung CH, Haskard DO, Ciclitira PJ. Differential upregulation of intercellular adhesion Received November 3, 1995. Accepted May 6, 1996. molecule 1 in coeliac disease. Clin Exp Immunol 1990;82:489 Address requests for reprints to: Salvatore Auricchio, M.D., Depart- 492. ment of Pediatrics, University Federico II, via S. Pansini 5, 80131 15. Maiuri L, Picarelli A, Boirivant M, Coletta S, Mazzilli MC, De Naples, Italy. Fax: (39) 81-546-9811. Vincenzi M, Londei M, Auricchio S. Definition of the initial immu- Presented at the 1995 International Meeting of Mucosal Immunolnologic modifications upon in vitro gliadin challenge in the small ogy in San Diego, California.