| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1
Dipartimento di Biochimica e Biotecnologie Mediche, Facoltà di Medicina e Chirurgia, and CEINGE-Biotecnologie Avanzate, Università "Federico II", via S. Pansini 5, 80131 Napoli, Italy;
2
Dipartimento di Pediatria, Facoltà di Medicina e Chirurgia, Università "Federico II", via S. Pansini 5, 80131 Napoli, Italy;
a author for
correspondence: fax 39-81-7463650, e-mail salvator{at}unina.it
Celiac disease (CD) is a genetically complex, multifactorial immune-mediated disease (1). The intestinal damage that characterizes the disorder is induced by dietary gluten ingestion in susceptible individuals (1). A specific HLA DQA10501/DQB10201 heterodimer, or in a few cases, the HLA DRB104 alleles, is associated with the disease (2). In fact, when presented by these HLA molecules, gluten-derived peptides cause T-cell activation in the intestinal mucosa, which is followed by cytokine production and mucosal intestinal damage (1).
The clinical manifestations of CD are very variable, and most of the symptoms are also present in other clinical conditions. Diagnosis of CD includes the presence of circulating gliadin and endomysium antibodies (EMAs) (3) and is traditionally confirmed by intestinal biopsy, according to European Society of Paediatric Gastroenterology and Nutrition (ESPGAN) criteria (4). The genetic diagnostic approach may be very useful in CD patients with IgA deficiency, a condition where the serological tools are less useful; in familial screening; or in cases of latent forms of CD. Regarding the genetic approach, despite numerous reports of HLA class II associations to CD in diverse countries (5)(6)(7)(8)(9), there has been no systematic study to establish the role of HLA typing in the diagnosis of CD, particularly in discriminating CD from other confounding diseases.
The aim of this study was to verify, using a PCR-based method that we recently established (10), the prevalence of the HLA heterodimer and of the HLA DRB104 alleles in healthy subjects, in CD-affected children, and in other age-matched subjects affected by confounding disease from South Italy. In a subgroup of CD patients in an active stage of disease, we also compared the diagnostic characteristics of genetic patterns and the presence of the EMAs and gliadin antibodies.
We examined two groups of patients from the Center for the Study of Gastrointestinal Disorders of the Department of Pediatrics of our University. One group (n = 74), affected by CD diagnosed according to the ESPGAN criteria (4), was examined retrospectively. The other group consisted of 80 patients (mean age, 6 years) at their first clinical examination whose history of clinical symptoms and/or results from laboratory tests were suggestive of CD. All 80 patients underwent jejunal biopsy. Celiac disease was diagnosed in 48 of 80 patients on the basis of the revised ESPGAN criteria (4); therefore, the total CD patients reached 122 cases in this study. In the other 32 of 80 patients, the small intestinal mucosa was histologically normal, and these patients were discharged with other diagnoses. These 32 patients served as gastrointestinal controls for our study. Blood samples from 116 healthy ethnically matched and unrelated adult controls, with no symptoms or family history of CD, obtained from the Blood Transfusion Service of our University, were also examined for HLA typing.
The study was approved by the Ethics Board on Human Subjects of our University, and informed consent was obtained from all individuals directly (adults) or from the parents (children).
Blood samples were immediately processed for genomic DNA extraction, using the proteinase K and phenol-chloroform procedure. The PCR-based methodology for the rapid detection of the HLA DQA10501, DQB10201, and DRB104 alleles is described elsewere (10). In the subjects lacking the heterodimer and DRB104 alleles, the other alleles at the HLA-DQB1 locus were typed by PCR amplification with sequence-specific primers, according to Olerup et al. (11).
Sera from the 80 patients at their first clinical examination were tested for IgA EMAs with an immunofluorescence method and for IgAIgG gliadin antibodies with an ELISA method.
The 
test and Fisher's exact test were used to
compare alleles and haplotype frequencies in patients and controls.
Diagnostic sensitivity, specificity, and the positive and negative
predictive values were calculated according to Galen and Gambino
(12) on the basis of the CD prevalence detected in our
Center, which is 52%.
The frequencies of the DQA10501, DQB10201, and
DRB104 alleles in the celiac patients, in patients with
other gastrointestinal diseases, and in the healthy subjects are
reported in Table 1
A. The DQA10501/DQB10201 heterodimer was
present in a significantly (P <0.001) higher percentage of
cases (87%, 106 of 122) with respect to the two control groups. Eight
of the 16 celiac patients who lacked the heterodimer had
DRB104 alleles (50%), compared with 12% and 20% of the
disease and healthy controls, respectively (P <0.05). Our
data on CD in southern Italy confirm the primary association of the
disease with the heterodimer and a weaker association with the
DRB104 alleles (5)(6)(7)(8)(9). The percentage frequency
of the heterodimer in our CD population is in accordance with those
obtained in other geographical areas 96% (the former Czechoslovakia),
91.3% (United Kingdom), 93% (Spain), and 75.5% (Switzerland)
(5)(9)(13).
|
In our CD patients, the DRB104 alleles, in absence of heterodimer, were present in a higher percentage of cases (50%) than in controls (20%). A similar finding was reported for CD patients (75%) and in healthy controls (13%) in a Spanish population (7). On the other hand, the DRB104 allele frequency was the same (6%) in CD patients and in healthy controls in central Italy (8). Consequently, our data support the heterogeneity of the genetic expression of CD in the various geographical areas. The heterodimer was present in 27% of the healthy subjects; this percentage is comparable with those reported for other healthy Caucasians of southern and northern European countries such as Spain (25%) (7) and Norway (26%) (6), but higher than those reported for Switzerland (17%) (9) and central Italy (Rome, 18%) (8).
Eight of the 122 CD patients lacked both the heterodimer and the DRB104 alleles; therefore, we further screened the DQB1 locus to investigate if other alleles were present in association to CD (2)(14). Two of these eight CD patients showed the DQ2 molecule (DQB10201 allele in heterozygosis); but one of them carried also the DQB10501 allele, also found in Sardinian patients to be associated to CD (15); four showed the DQ7 molecule (DQB10304 or DQB10301 alleles). Another patient carried the DQ8 molecule encoded by the DQB10302 allele, and the remaining patient had the haplotype DQB10501 in homozygosis (15). The DQ8 molecule seems to be an alternative to DQ2 in influencing susceptibility towards CD, being present in up to 20% of celiac patients not bearing of DQ2 in the Mediterranean area (2)(14)(16). Our data do not support an earlier finding that HLA-DQ7 is a nonsusceptible molecule (2). In fact, DQ7 was present in 50% of our CD patients in the absence of the heterodimer and of the DRB104 alleles. Because the DQ7 molecule is very similar to the DQ8 molecule, it could alternatively present similar gluten-derived peptides to restricted T cells (17). DQ7 has been detected in a few CD cases in the absence of DQ2 but in association with the DRB104 alleles (18).
The diagnostic power of HLA typing and of the immunological approach in
CD was calculated in 80 patients whose symptoms suggested CD and is
reported in Table
1B. At the CD prevalence of 52% in our Center, the
positive predictive value of the heterodimer was 88%, whereas the
negative predictive value was 80%. EMAs were present in all CD
patients except two; one was 18 months old and had the heterodimer, the
other presented the DQ7 molecule.
Three EMA apparently false-positive results were obtained among the gastrointestinal controls, one of whom was a patient affected by Down syndrome, a condition known to often be associated to CD; he was negative for heterodimer and DRB104 alleles. Of the other two, the first had the heterodimer and the second had DRB104 alleles. These latter two patients, notwithstanding the apparently healthy morphology of the intestinal mucosa, were monitored as potential celiac patients. To our knowledge, this is the first combined study of HLA typing of gastrointestinal disease controls and CD patients. Given the positive predictive value of 88% obtained in our CD patients, the heterodimer typing should be used in addition to an immunological approach and, when required, to discriminate affected patients from patients with other confounding diseases among patients with symptoms and signs suggestive of CD.
Although a non-HLA locus has been recently found to be associated to CD (19), the HLA heterodimer is a strong genetic factor for the disease susceptibility (1)(2). Our rapid HLA typing of the DQA10501/DQB10201 and DRB104 alleles provides a useful adjunctive tool in the diagnosis of CD and could represent another laboratory indication of the disease in suspected cases, particularly (a) when CD diagnosis is complicated by ambiguous histological and immunological patterns; (b) in celiac disease latency, where the presence of EMAs occurs in symptom-free patients with apparently healthy mucosal morphology; (c) to lend support to the CD diagnosis when a small intestinal biopsy is not available; and (d) for family screening.
Acknowledgments
This study was supported by grants from Ricerca Sanitaria Finalizzata (Regione Campania, Italy), CNR, and Progetto Finalizzato Biotecnologie (CNR, Roma).
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
