Introduction

Chronic pancreatitis (CP) is a continuing or relapsing inflammatory disease of the pancreas. CP follows recurrent attacks of acute pancreatitis with all of the common complications of CP, including unrelenting pain, parenchymal and ductal calcifications, duct distortion, fibrosis, maldigestion, and diabetes mellitus. The etiological factor of about one-third of chronic and acute recurrent pancreatitis patients is unknown and this group of patients is classified as suffering from idiopathic pancreatitis (IP). Mutations of the cystic fibrosis transmembrane conductance regulator (CFTR) gene have been associated with IP.1,2,3,4 Active trypsin is inhibited by trypsin inhibitors as serine protease inhibitor Kazal type 1 (SPINK1), and when trypsin activity exceeds the inhibitor effect of SPINK1 it will continue to activate trypsinogen and other zymogens, leading to autodigestion of the pancreas and pancreatitis. Mutations in the SPINK1 gene have been reported in hereditary and idiopathic acute and chronic pancreatitis.4,5,6,7 Witt et al5 reported that 22 out of 96 CP patients carried a SPINK1 mutation and that 18 of them carried a missense mutation in codon 34 (N34S). Pfützer et al6 found that 26 out of 57 IP patients carried a SPINK1 mutation and that 23 of them carried the N34S, confirming the association of this mutation with CP and postulating that alterations in the SPINK1 can act as a disease modifier gene in this pathology. Audrézet et al7 demonstrated that four out of 39 IP cases carried the N34S mutation in the SPINK1 gene.

The aim of this study was to search for mutations in the SPINK1 gene in association with susceptibility to IP in Italian patients and to study the possibility of a synergistic effect between SPINK1 and CFTR gene mutations in IP.

Patients and methods

Cases and controls

The population under study consisted of 32 IP Italian subjects (18 males, 14 females, mean age 34.6 years, range 7–59 years), suffering from either idiopathic chronic pancreatitis (17) or recurrent acute pancreatitis (15). The criteria for diagnosis were as follows: for idiopathic recurrent pancreatitis, multiple episodes of usually clinical mild pancreatitis in the absence of the etiological factors most commonly associated with pancreatitis such as colelithiassis, alcohol abuse, pancreas divisum, hypercalcemia, dyslipidemia and viral infections; and for idiopathic chronic pancreatitis: progressive pancreatic inflammatory disease characterized by irreversible morphological changes and typically causing pain and/or permanent loss of function, where none of the possible causes previously mentioned in the other definitions had been detected. A total of 50 control individuals from the same Italian population were analyzed in this study. Mutation N34S was tested in a total of 100 control individuals. The patients had been previously analyzed for mutations in the CFTR gene.8,9

DNA extraction and analysis

DNA was extracted from peripheral blood leukocytes by standard procedures.

Mutation N34S of the SPINK1 gene was detected by PCR with modified primers that produce a restriction site during in vitro amplification (RG-PCR)10 using forward primer 5′CAATCACAGTTATTCCCCAG3′, and modified reverse primer 5′TGGTGCATCCATTAAGTGCA3′ which creates a restriction site for endonuclease Bsp1286 I when allele 34S is present. Digestion products were analyzed by 8% acrylamide gel electrophoresis.

The 5′UTR region, the four exons and the exon–intron boundaries of the SPINK1 gene were amplified as described11 and denaturing gradient gel electrophoresis (DGGE) analysis of the amplified fragments was performed. Amplified fragments showing an altered migration pattern were directly sequenced using the BigDye terminator Cycle Sequencing kit (Perkin–Elmer) on a 377 ABI PRISM sequencer.

Results and discussion

Mutation N34S was tested in 32 IP, and 100 control individuals. The mutation was present in three IP patients, one of whom was homozygous, and in one control individual (Fisher's exact test, P=0.044) (Table 1).

Table 1 Mutations and polymorphisms in the SPINK1 gene and distribution among IP, and controls individuals
Full size table

The four exons and exon–intron boundaries, and the 5'UTR region of the SPINK1 gene were screened for sequence variations by DGGE analysis in 32 cases of IP, and in 50 control individuals. In the 5′UTR region, polymorphism −253T>C and the −41G>A variant were found. The common polymorphism −253T>C was present in both groups with similar frequencies and the −41G>A variant was present in one IP patient only. In IVS1, variant −37T>C was observed in three IP patients, and it was in linkage disequilibrium with mutation N34S in exon 3. Beside the N34S mutation, we found the following in exon 3 and boundaries: variant 174C>T(C58C) in one IP patient, variants IVS2−23A>T in two controls, and a novel sequence variant IVS3+12T>G in one IP patient. In exons 1 and 4 and their exon–intron boundaries, no sequence alteration was observed. All the DNA variants detected in this study as well as their distribution in both study groups are summarized in Table 2.

Table 2 Analysis of mutations in CFTR and SPINK1 mutations in ICP patients
Full size table

Four sequence variations were found in IP patients, and of these only the N34S mutation, which was present in three patients (9.3%), has previously been considered as being associated with IP. This result is similar to that reported by Audrézet et al,7 who reported the presence of the N34S mutation in 10% of ICP patients, and by Ockenga et al,12 who reported a similar prevalence of SPINK1 mutations in a cohort of adult IP patients. A higher frequency of SPINK1 mutations was described by Witt et al5 and by Pfützer et al,6 where the frequency of N34S in CP patients was 18.75 and 40.3%, respectively.

In this study, no other sequence alteration was observed that could be associated with an altered function of the SPINK1 protein. The −41G>A variant, which was present in one IP and in no control individual, has been discussed by Audrézet et al7 as having a possible role in transcription efficiency of the protein. However, it was also present in 3/190 control individuals in the study by Pfützer et al,6 so its relevance in IP remains to be determined. Interestingly, the N34S mutation was present in two patients who carried the 5T variant and in one who carried the L997F variant in the CFTR gene (Table 2). The patients carrying one CFTR gene and one SPINK1 gene mutation had no family history of IP. No relatives were available for a segregation study of the mutations. The 5T variant of the CFTR gene is one of the three known alleles with a variant number of thymidines (5,7 or 9) in a polythymidyl tract (T-tract) of the acceptor splice site in intron 8. The variant 5T is associated with low amounts of normally processed mRNA13 and, although this allele does not cause CF, it can modify the phenotypic effect of other mutations.14 It can be hypothesized that the added action of a lower level of correctly processed CFTR gene product, together with an altered function of the SPINK1 gene product, could have a deleterious effect upon the pancreatic phenotype. A high frequency of the L997F mutation of the CFTR gene has been described in our cohort of IP patients8 and it has been postulated that this mutation, together with other CFTR gene mutations, can cause IP as an atypical, monosymptomatic form of CF. As with the 5T variant, it can be hypothesized that the L997F mutation combined with a SPINK1 mutation can have a synergistic effect upon the development of IP. IP patients carrying a SPINK1 mutation as well as a CFTR gene mutation have also been reported by Witt et al15 as unpublished data. Noone et al4 reported nine out of 24 CP patients carrying the N34S SPINK1 gene mutation. Three of these patients carried the SPINK1 mutation together with a CFTR mutation: two carrying the 5T variant in the CFTR gene, and one carrying two CFTR mutations (F508del and H117H). In three other patients carrying the N34S mutation, the authors could not exclude the presence of CFTR rare mutations as only the common CFTR gene mutations were screened. Ockenga et al,12 revealed double heterozygosity for CFTR Y1092X mutation and SPINK1 R65Q mutation, not only in a patient but also in his unaffected relatives, indicating that other genetic or environmental factors are needed to develop PI. However, the possible relevance of the novel SPINK1 gene mutation R65Q in IP is still unclear.

The frequency of double heterozygotes for SPINK1 and CFTR gene mutations among IP patients compared with control individuals is significantly higher (3/32 IP, 0/100 controls, Fisher's exact test P=0.013). The presence of patients who carry mutations at the two genes suggests that epistatic interactions in trans-heterozygotes may contribute to disease pathogenesis. The distribution of SPINK1 mutations in IP patients was not homogeneous, 0/10 patients carrying a CF causing mutation, 3/7 carrying a CFTR variant and 0/15 carrying no CFTR gene variant (χ2 after permutation P=0.037). Severe CFTR mutations were not identified in this study, possibly because of the low number of observations. Our observation suggests a possible interaction of CFTR and SPINK1 gene variants in determining IP.