Heterogeneity of Celiac Disease: Clinical, Pathological, Immunological, and Genetic

PDFlib PLOP: PDF Linearization, Optimization, Protection

Page inserted by evaluation versionwww.pdflib.com – [email protected]

112

Heterogeneity of Celiac Disease:Clinical, Pathological, Immunological,and Genetic

ANNE FERGUSON,a HELEN GILLETT, KENNETH

HUMPHREYS, AND KATHLEEN KINGSTONE

Gastro-intestinal Unit, University of Edinburgh Department ofMedicine, Western General Hospital, Edinburgh EH4 2XU,United Kingdom

ABSTRACT: In this paper we consider recent new data on the pathologicalfeatures of gluten sensitivity and on the disease-associated antigens, in thecontext of a multistage hypothesis that we have been developing for the lastfive years. This incorporates concepts of oral tolerance induction, mucosalT-cell and antibody-mediated injury, and genetic contributions. Until now,there has been complete agreement that the diagnosis of celiac disease mustbe based on small bowel histology. There are patients with low-gradegluten-sensitive enteropathy, in whom the only morphological abnormalityis a high count of intraepithelial lymphocytes (IEL). Some, but not all, alsohave positive serum IgA anti-endomysium antibody (AEA). With good tech-niques, in a properly accredited laboratory, in a patient suspected on clini-cal grounds to have celiac disease, a positive serum IgA AEA test (perhaps,alternatively, high-titer anti-transglutaminase by ELISA), is virtually diag-nostic of the condition. Our hypothesis of a stepwise pathogenesis of severegluten-sensitive enteropathy is re-examined in the light of these new data. Itis evident that there are at least five different levels at which genetic influ-ences may operate.

In 1992, at the Coeliac Symposium in Dublin,1 one of us predicted that the fol-lowing advances in knowledge could well be achieved by the turn of the century:

• Precise definition—to include descriptions of what the disease-associatedgene(s) encode, the precise molecular conformation of the provoking sub-stance, and the trigger for sensitization.

• Firm diagnostic criteria—diagnosis to be based on the demonstration thatsensitization has occurred.

aAdditional correspondence information: Telephone: 44-131-537-1731; Fax: 44-131-537-1007; e-mail [email protected]

• Pathogenesis—will have been shown to involve at least two components,intestinal T-cell-mediated immunity to gluten and one or more other fac-tors; there will be a degree of heterogeneity with major (classical) andminor subsets, related to genetic and mechanistic differences.

In this paper we consider recent new data on the pathological features ofgluten sensitivity, and on the nature of disease-associated antigens, in the contextof these predictions from five years ago.

CLINICAL AND PATHOLOGICAL SPECTRUM OF GLUTEN-SENSITIVE ENTEROPATHY

In view of the enormous clinical heterogeneity observed in celiac patients, ithas been generally accepted that the criteria for definition and diagnosis shouldbe based on small bowel pathology.2,3 In other words, classical celiac disease isdefined as a permanent gluten-sensitive enteropathy. In patients with active celiacdisease, there are nutritional deficiencies and/or diarrhea; malabsorption syn-drome is now relatively rare. Clinically silent celiac disease is now being increas-ingly recognized, for example in asymptomatic people in whom jejunal biopsyhas been performed as part of a family study or in screening of blood donors. Verylittle research has been done on the different factors that lead to silent disease insome patients and to malabsorption and diarrhea in others.

The “enteropathy” of the above definition involves a group of pathologicalfeatures—villus and crypt sizes, crypt mitotic activity, epithelial cell damage, andintra-epithelial and lamina propria lymphoid cell infiltrates.

When quantitative histology and computerized image analysis are applied ingluten challenge and gluten withdrawal protocols, and in certain groups of gluten-sensitive individuals such as those with dermatitis herpetiformis, it becomes clearthat features such as villus/crypt architecture changes, lamina propria cell counts,and intra-epithelial lymphocyte (IEL) counts form a continuum, with the classi-cal flat lesion at one end of the spectrum and at the other end a mucosa with nor-mal villus and crypt architecture; but the only measurable abnormality is a highdensity or count of villus intra-epithelial lymphocytes, IEL.3 A change in IELcount or density in the surface epithelium of a duodenal or jejunal biopsy is thefirst and most sensitive index of the effects of gluten reintroduction into the dietof a celiac patient.4 Furthermore, there is now a substantial body of evidence thatgluten-sensitive enteropathy may be manifested only by IEL count. In otherwords, there are patients who have a high count of IEL in an otherwise normaljejunal biopsy while taking a normal diet, and that their IEL count falls withgluten exclusion and rises on gluten reintroduction. Some have no gastrointesti-nal symptoms, for example patients with dermatitis herpetiformis. But there areothers who have gluten-sensitive diarrhea as well as gluten-related IEL counts.5

FERGUSON et al.: HETEROGENEITY OF CELIAC DISEASE 113

114 ANNALS NEW YORK ACADEMY OF SCIENCES

PATHOLOGICAL FEATURES THAT DO AND DO NOT CORRELATEWITH T-CELL ACTIVATION

Most, but not all, of the features of the celiac lesion can be reproduced inmodel systems of T-cell activation, such as graft-versus-host reaction in themouse—crypt hyperplasia, shortening of villi, high density of IEL infiltrate in thevillus epithelium, IEL mitosis, and HLA class II expression by crypt cells.Different cell activation and cytokine signals may underlie these components—γ-interferon and IL-2-influencing crypt cells, TNF associated with villus flatten-ing, and cytotoxic cells or other cytokines involved in the rare, severe hypoplas-tic lesion. However, it is also probable that some immune features that occur inceliac disease have a different pathogenesis.

The striking damage to surface enterocytes that occurs in celiac disease is nota feature in models of delayed-type hypersensitivity and cannot merely be attrib-uted to direct gluten toxicity. The cells at the surface of a “flat” biopsy takenafter an overnight fast will have moved out of the crypts onto the surface duringthe preceding 8–12 hours and therefore have not been exposed to gluten. Thereis convincing evidence for a role of complement in the celiac lesion6 and it islikely that IgM/IgG immune complex/complement-mediated damage probablyco-exists with delayed-type hypersensitivity (DTH) in untreated patients andmay be the major factor responsible for the changes in basal lamina, reducedenterocyte height, and derangement of enterocyte brush border. The antigen orantigens involved in these immune complexes are likely to include gluten, otherfoods, and autoantigens (see below).

ORAL TOLERANCE

Much work has been performed in rodents, showing that feeding of proteinantigens, including gliadin, induces the specific immune response of oral toler-ance. This is a powerful and prolonged suppression and downregulation of thecapacity of the animal to mount active antibody- and T cell-mediated immuneresponses to the antigen concerned.

Are nonceliacs immunologically tolerant to gluten, or merely not sensitized?There are no data for humans, but in mice, our comparative studies of immuneresponsiveness of genetically identical animals from gluten-free and gluten-con-taining diet colonies have clearly shown that mice eating a normal diet are tolerantfor both cellular and humoral limbs of the systemic immune response to gluten.7

It is important to note that human infants are born with a degree of alteredimmunity to gluten and other foods. They will have circulating, maternallyderived IgG anti-gliadin antibody in variable titer and if breast-fed will alsoreceive maternal IgA anti-gliadin antibody in milk. Small amounts of foodseaten by the mother, including gluten, are present in human breast milk.Furthermore, there may also be passive transmission of IgG and IgA anti-idiotype antibodies by these routes—antibodies to the antigen-combining (Fab)

site of specific antibodies, which may have the same conformation as the origi-nal antigen and can, themselves, act as immunogens. Whether or not there is alsosome passive transfer of cellular immunity, in milk or transplacentally, remainsto be established

Thus, although further work on active immunity and tolerance to foods ofweaning diets is urgently needed, research on influences of the otherimmunomodulatory influences such as passively transferred immunity, is alsorequired.

WORKING HYPOTHESIS, 1994: SEVERAL INDEPENDENT STEPS INTHE PATHOGENESIS OF CELIAC DISEASE

Our studies with mice showed that immunological sensitization to gliadin doesnot trigger the development of a T-cell-mediated lesion of the intestine when thediet contains gluten.8 Additional cofactors were required, which could act viaenhanced antigen presentation, recruitment of specific T cells in the mucosa,upregulation of the expression of class II antigens, failure of suppression, or agene that modulates the expression of mucosal cellular immunity.

Integration of knowledge of oral tolerance with recent changes in our per-spectives of gluten sensitivity generated the following hypothetical sequence9 asthe pathogenesis of gluten-sensitive enteropathy:

1. Failure of the development of oral tolerance for T-cell mediated immu-nity, either globally (many antigens) or confined to gluten.

2. The nontolerant individual is thus vulnerable to being actively immunizedto the food antigen (gluten) if a particular combination of diet, gut perme-ability, mucosal, and systemic immunomodulatory signals coincide. Thefrequency with which this occurs in nontolerant individuals, and at whatage, will depend on many intrinsic and environmental factors.

3. When active T-cell sensitization has occurred, and there is gluten in thediet, relatively subtle effects on the gut mucosa (low-grade pathology, e.g.,high IEL count with normal villi) occur in a proportion of those at risk.

4a. It is then only a matter of time until a critical combination of antigen doseand activated mucosal T effector cells occurs, and this then precipitatesthe evolution of severe enteropathy and malabsorption.

4b. Simultaneously with 4a, a wide range of immune effector cells arerecruited, and other dietary antigens, and possibly auto-antigens, becomeinvolved, thus perpetuating and worsening the enteropathy and explainingwhy tissue damage persists for weeks or months after strict dietary exclu-sion of gluten.

We suggested that differences in the proportion of those “at risk” whoprogress to the next step 1–>2, 2–>3, 3–>4, will explain striking differences indisease frequency in groups that share the same genetic make-up, for example,


Swedes and Danes, British infants in the late 1960s, and now healthy andaffected relatives of celiacs.

As discussed below, emerging differences in the genetic, clinical, and immuno-logical features between patients with mild and severe enteropathy now raiseother possibilities.

ANTI-ENDOMYSIUM ANTIBODY—WHERE DOES THIS FIT WITHINTHE PATHOLOGICAL AND CLINICAL SPECTRUM?

In clinical practice, the presence of IgA anti-endomysium antibody (AEA) inserum—as detected by immunofluorescence techniques—is highly predictive ofclassical celiac disease (TABLE 1) and correlates better with diagnosis than do any ofthe antibody tests with gliadin. Now that the autoantigen has been identified as tis-sue transglutaminase (as described in the previous paper), it is possible to use themore sensitive immunoassay techniques to detect lower titers of AEA than is possi-ble using immunofluorescence; and also to establish whether anti-transglutaminaseantibodies develop in patients with conditions other than celiac disease. As shown inFIGURE 1, serum IgA antibodies to transglutaminase, measured by ELISA, correlatevery well with the presence of celiac disease and with diet status; this sensitive,quantitative test can reveal the presence of antibody in amounts that fall below thethreshold for detection of AEA, but are still abnormally high. This was the case forall five patients with nonceliac small bowel disease, as shown in FIGURE 1. How this


TABLE 1. Correlation of IgA AEA test results and biopsy findings for patientsfrom Edinburgh whose serum was submitted to the gastrointestinal laboratoryfor serological testing

A. 838 samples submitted for celiac serology during 1995

• IgA AEA examined by immunofluorescence method with umbilical vein; titer1 in 5: 776 cases negative, 62 positive.

B. Small bowel biopsy data for 44 patients with negative AEAa

• Biopsy normal, (34)• High IEL only, (7)• Partial villus atrophy (inflammatory bowel disease), (1)• IgA deficiency; subtotal villus atrophy; celiac disease, (2)

C. 62 patients with positive IgA AEA: further details

• Biopsy normal, (2)• High IEL only, (6)• Severe enteropathy—celiac disease confirmed, (27)• Known celiac, (16)• No biopsy, empirical GFD, (4)• No biopsy, (2)• No information, (5)

aNumber of patients in parentheses.

phenomenon relates to the mechanisms of tissue injury, for example in small bowelCrohn’s disease, remains to be elucidated.

It is becoming clear that many (but not all) patients with gluten-related symp-toms, normal villus/crypt architecture but high counts of IEL, also have positiveAEA (by immunofluorescence) in serum. We have studied eight patients withgluten-related diarrhea, morphologically normal biopsy but high counts of IEL,and found that three also had postive serum IgA AEA. On a trial of gluten-freediet, diarrhea resolved and IEL count dropped, both in those patients withdetectable and with undetectable AEA (unpublished observations). There is arecent report of 10 patients with absolutely normal small bowel histology,(including normal IEL count in six of the 10), in whom positive AEA paralleledgluten-sensitive symptoms such as diarrhea.10

NEED FOR FURTHER REVISION OF CRITERIA FOR CELIACDISEASE DIAGNOSIS IN CLINICAL PRACTICE

We submit the following criteria for debate; in our own clinical practice, thisis the consensus we have reached and are now using:


FIGURE 1. Titers of IgA antibody to guineau pig tissue transglutaminase, in serum.Arbitrary units; interim reference range <1,200 units/ml. Results are shown for patients withclassical celiac disease, untreated; gluten-free diet treated celiac patients; immunologicallynormal patients; and five patients with nonceliac intestinal disease—two with Crohn’s dis-ease, one giardiasis, one scleroderma with intestinal involvement, one intestinal ischemia.


• With good techniques, in a properly accredited laboratory, a positive serumIgA AEA test in a patient, suspected on clinical grounds to have celiac dis-ease, is virtually diagnostic of the condition.

• Before starting gluten-free diet treatment, it is important to “stage” thepatient, by a combination of small bowel biopsy pathology and a functionaljejunal test, for example, sugar permeability (so that treatment response canbe monitored).

• A high IEL count in an otherwise morphologically normal biopsy plus pos-itive AEA is celiac disease (low-grade pathology type).

• A high IEL count in an otherwise normal biopsy, with negative AEAor normal biopsy with normal IEL count, but positive AEA:These patients, if symptomatic, should have a trial of gluten-free diet.

• We still need to agree on how to monitor the effects of treatment in theseatypical patients.

TABLE 2. Hypothetical staging (1997 version) of the pathogenesis, pathology,and immunology of celiac disease and examples of likely genetic influences

Stage in Evolutionof Celiac Disease

Small BowelHistology (when

eating gluten)

Likely Relevance ofGenetic Factors

Contributing to theEvolution of Celiac

Disease

ImmunologicalFeatures

A—global tolerance-capacity

B—tolerance togluten or related epitopes

Not tolerant Histology normal Antibody negative

C—gut epitheliumand subepithelialtissues many dif-ferent properties

Sensitization event Transient damage Release of transgluta-minase in smallbowel, risk ofimmunization

Not celiac haplotypeSensitized, goodmucosalimmunoregulation;resistance toenteropathy

Normal or high IEL Gut immunization togliadin complexesand autoantigens(antibodies only)

D—celiac haplotypeE—? γδT-cells

Sensitised, poormucosalimmunoregulation;predisposition toenteropathy

Flat (severe enteropathy)

Gut immunisation togliadin complexesand autoantigens(antibodies and Tcells)

ROLES OF GENETICS AND ENVIRONMENT IN CELIAC DISEASEHETEROGENEITY—STATUS OF THE SEVERAL-STAGE

HYPOTHESIS IN 1997

Clearly, the discriminant potential of genetic research will be greatly enhancedby ensuring that the pathologic and clinical features of patients being used forgenetic research work are very clearly defined, so that appropriate subsets can beanalyzed separately.

A recent study from Italy10 showed that most patients with positive AEA andgluten-sensitive symptoms but normal villi, did not have high counts of γ,δT-cellsand did not possess the typical celiac haplotype. Thus we suggest that the influ-ence of the celiac-associated DQ heterodimer may be at the level of expressionof severe rather than mild enteropathy and may not actually be crucial to the ear-lier stages of gluten nontolerance or sensitization.

ACKNOWLEDGMENTS

We thank Mr. John Bode for his critical and excellent technical work, includ-ing the development of our in-house AEA test. This work has been supported bygrants from the Scottish Office Acute Healthcare Committee and by donations tothe Edinburgh Intestinal Immunology Research Fund.

REFERENCES

1. FERGUSON, A. 1994. Pathogenesis of coeliac disease: The way ahead. InGastrointestinal Immunology and Gluten-Sensitive Disease. C. Feighery & C.O’Farrelly, Eds.: 269–273. Oak Tree Press. Dublin.

2. FERGUSON, A., E. ARRANZ & S. O’MAHONY. 1993. Clinical and pathological spectrumof coeliac disease —active, silent, latent, potential. Gut 34: 150–151.

3. FERGUSON, A., E. ARRANZ & S. O’MAHONY. 1992. Definitions and diagnostic criteriaof latent and potential Coeliac Disease. In Common Food Intolerances 1:Epidemiology of coeliac disease. S. Auricchio & J. K. Visakorpi, Eds.: 119–127.Karger Basel.

4. ZIEGLER, K. & A. FERGUSON. 1984. Coeliac disease. In Function and Dysfunction ofthe small intestine. R. M. Batt & T. L. J. Lawrence, Eds.: 149–166. LiverpoolUniversity Press. Liverpool, United Kingdom.

5. ARRANZ, E. & A. FERGUSON. 1993. Intestinal antibody pattern of coeliac disease:Occurrence in patients with normal jejunal biopsy histology. Gastroenterology 104:1263–1272.

6. HALTENSEN, T. S. et al. 1992. Association of subepithelial deposition of activatedcomplement and immunoglobulin G and M response to gluten in celiac disease.Gastroenterology 102: 751–759.

7. TRONCONE, R. & A. FERGUSON. 1988. Gliadin presented via the gut induces oral tol-erance in mice. Clin. Exp. Immunol. 72: 284–287.

8. TRONCONE R. & A. FERGUSON. 1991. Animal model of gluten induced enteropathy inmice. Gut 32: 871–875.


9. FERGUSON, A. 1995. Coeliac disease research and clinical practice—maintainingmomentum into the twenty-first century. In Clinical Gastroenterology, CoeliacDisease. P. D. Howdle, Ed. 9(2): 395–412. Bailliere Tindall, London.

10. PICARELLI, A. et al. 1996. Gluten-sensitive disease with mild enteropathy.Gastroenterology 111: 608–616.


Documents

Heterogeneity of Celiac Disease: Clinical, Pathological, Immunological, and Genetic