Portal de Periódicos da CAPES

Study of Two Ectopeptidases in the Susceptibility to Celiac Disease: Two Newly Identified Polymorphisms of Dipeptidylpeptidase IV

2000; Lippincott Williams & Wilkins; Volume: 30; Issue: 4 Linguagem: Inglês

10.1097/00005176-200004000-00024

1536-4801

Fabienne Clot, Marie‐Claude Babron, Selvaggia Percopo, Mara Giordano, F. Bouguerra, Françoise Clerget‐Darpoux, Luigi Greco, Jean‐Louis Serre, Marie‐Claude Fulchignoni‐Lataud,

Gastrointestinal disorders and treatments

Celiac disease (CD) is a T-cell–mediated enteropathy induced by ingestion of wheat gluten and related proteins from other cereals in genetically susceptible persons (1). The damage of the small intestine is characterized by villous atrophy, crypt cell hyperplasia and infiltration of lymphoid cells. The disease is strongly associated with HLA class II molecules, but other non-HLA genetic risk factors may contribute to the development of CD, as suggested by the difference in concordance rates between HLA-identical siblings (30%) and monozygotic twins (70%) (2). A theory was suggested to explain the pathogenesis of CD, consisting of an alteration in peptidase enzymatic activities in the brush border of the small intestinal epithelial cells and responsible for an incomplete degradation of the gluten, resulting in the production of toxic peptides. Among the enzymes involved in the digestive process are the dipeptidylpeptidase IV (DPPIV; EC 3.4.14.5; CD26) and aminopeptidase N (APN; EC 3.4.11.2; CD13). Dipeptidylpeptidase IV, located on chromosome 2, is a serine exopeptidase that cleaves X-proline or X-alanine dipeptides from the N terminus of polypeptides (3). Expressed in the brush border membranes of the small intestine and in other tissues, its activity has been shown cytochemically to decrease significantly in enterocytes of children with CD (4) and to be mostly affected in the ileum of rats and piglets in animal models for human CD (5). Moreover, in the rat, a functional role of DPPIV in the intestinal hydrolysis of prolyl peptides and assimilation of proline-rich proteins, such as gliadin, was described (6). For these reasons, the DPPIV gene can be considered a good candidate gene for causing susceptibility to CD. Furthermore, it plays a significant role in the process of activation and proliferation of human thymus-derived lymphocytes (7) and could participate in the immune reactivity present in patients with CD. The APN enzyme is also a good candidate gene for the susceptibility to CD, as studied by Giordano et al. (8). Located on chromosome 15, it is a membrane-bound metallopeptidase of the epithelial cells from renal proximal tubules and the intestinal brush border (9). It plays a role in the final digestion of peptides. It can trim amino-terminal ends of peptides at the surface of the antigen-presenting cells that protrude out of major histocompatibility complex (MHC) class II, and has significant effects on T-cell stimulation and specificity (10). To test the possible implication of DPPIV and APN genes as risk factors for CD, we took the opportunity of having nuclear families with affected sibling pairs to use two different statistical approaches: maximum likelihood score (MLS) and the transmission disequilibrium test (TDT). A sample of 210 Italian families (152 with two affected siblings and 58 with one affected chid) and one of 42 Tunisian families (2 with two affected siblings and 40 with one affected child) were collected. Diagnosis of patients with CD was established according to ESPGAN criteria (11). Because a polymorphism within or flanking the gene is crucial for TDT testing, a first step was to find such polymorphisms within the DPPIV gene. No such polymorphism was reported in the literature. The 10 longest exons including large intronic sequences were sequenced within 12 unrelated persons using the procedure described in Table 1.TABLE 1: PCR primers for genomic amplification of 10 exons of DPPIV geneThe sequencing of 10 exons and large intronic sequences of DPPIV, allowed the identification of two bi-allelic polymorphisms in intron 8, at position +47 after exon 8, resulting in the nucleotide change T→C, and at position −14 before exon 9, resulting in the nucleotide change A→G. Typing of these newly identified polymorphisms within nuclear families was performed using polymerase chain reaction (PCR) amplification with the primers of exon 8 and 9 described in Table 1, followed by allele-specific oligoprobe hybridization. Hybridizations were performed at 42°C overnight with oligoprobes labeled with digoxigenin-dideoxyuridine triphosphate. Then the membranes were rinsed with the following buffers: 5 minutes with 2× SSPE and 0.1% sodium dodecyl sulfate (SDS); once for 10 minutes with 50 mmol/l Tris (pH 8), 3 mol/l tri methyl ammonium chloride, 2 mmol/l ethylenediaminetetraacetic acid (EDTA) (59°C), and 0.1% SDS; and 5 minutes with 2× SSPE, and 0.1% SDS. The nonradioactive detection was performed using an antidigoxigenin antibody conjugated to alkaline phosphatase and CSPD (Boehringer-Mannheim, Mannheim, Germany). Linkage between CD and microsatellite markers flanking the DPPIV locus (D2S2370, D2S2190, D2S306, D2S156, and D2S2380) and the APN locus (D15S1038, D15S158, D15S963, and D15S996) was evaluated in 80 affected Italian sibling pairs by MLS with Mapmaker/Sibs program (12). The results showed MLS values near zero. Thus, there is no evidence for linkage between CD and the DPPIV or APN loci, although a linkage between markers flanking the APN region and CD was reported in Irish and British families affected by CD (13,14). However, in our situation, a gene may have had too weak an effect to be detected with the MLS approach. We therefore used the two newly identified DPPIV intragenic polymorphisms and two APN bi-allelic polymorphisms described by Giordano et al. (8) to perform a TDT test. In fact, the TDT may be a more powerful linkage test than the MLS test in the presence of a strong gametic disequilibrium (15). It compares the alleles transmitted by heterozygous parents to their affected child with the nontransmitted alleles by use of a χ2 test of homogeneity with 1 df(16). Because affected sibling pairs were used in the TDT analysis, the TDT is considered to be a test for linkage in the presence of association, as noted by Martin et al. (17) From the results reported in Table 2, the values of TDT were not significant for any of the studied polymorphisms in the Italian and Tunisian populations.TABLE 2: Transmission of DPPIV and APN bi-allelic polymorphisms from heterozygous parents to Italian and Tunisian childrenThe Tunisian sample for the whole study is small and could detect a predisposing factor only in the presence of a strong linkage disequilibrium or if the factor has a strong effect. Under these conditions, even small samples used in a TDT analysis can be performant. As few as 30 informative meioses made it possible to demonstrate the implication of the D2S1471 locus in diabetes (P = 0.028) (18). The Italian sample is very large for both candidate genes. The two DPPIV polymorphisms have been analyzed in many Italian families, leading to a large number of informative meioses, 250 informative meioses for E8+47 and 304 for −14E9. Similar conditions are obtained for the two APN polymorphisms by pooling our results with those of a previous Italian sample (8) : 321 informative meioses for A281G and 360 for G2957C. Therefore, these negative results may be due more to the absence of a linkage disequilibrium or more probably to the nonimplication of DPPIV and APN as genetic risk factors in the Italian population than to a lack of power. If the statistical data do not support that the DPPIV or APN regions could be risk factors for CD in the Italian population, replication studies in Tunisian and other populations should be investigated for both genes, to settle whether they are factors that predispose to CD. Despite the absence of association between CD and the two identified DPPIV polymorphisms, our study allows an estimation of their frequencies in the Italian population: the E8+47 T-allele frequency is 0.70, and the −14E9 A-allele frequency is 0.64. These new polymorphisms could be useful as markers for human genomic studies on chromosome 2 and could be used in a candidate gene approach for other diseases. Acknowledgments: Supported by a fellowship from Association Française contre les Myopathies (F.C.), by the Italian Coeliac Society and Consiglio Nazionale delle Ricerche, by the Société Nationale Française de Gastroentérologie, and by a Franco-Italian Galilee grant.