Editor—Hereditary pancreatitis (HP) is an autosomal dominant disease with about 80% penetrance that mainly afflicts white families.1 Although pancreatitis was hypothesised to result from inappropriate activation of pancreatic zymogens by Chiara2 in 1896, and the genetic nature of HP was identified by Comfort et al 3 in 1952, the precise mechanism underlying the pathogenesis of HP has remained a mystery until recently. By familial linkage analysis, a genetic defect was mapped to chromosome 7q35 by Le Bodic et al,4 and independently confirmed by two other groups in 1996.5 6 Soon after, a single G to A mutation resulting in an arginine (R) to histidine (H) substitution (R117H) in the third exon of the cationic trypsinogen gene was identified as being associated with HP by Whitcombet al.7

Trypsinogen is synthesised in the acinar cells of the pancreas and is activated into trypsin upon cleavage of the activation peptide by enterokinase. Trypsin plays a central role in pancreatic exocrine physiology by acting as the trigger enzyme which leads to the activation of all the pancreatic digestive proenzymes as well as trypsinogen itself. When the R117H mutation was identified, Whitcombet al 7 concluded that this mutation did not affect the tertiary structure of trypsin, nor alter its catalytic activity or interfere with trypsin inhibitor binding, since the three dimensional position of R117 was located on the opposite surface of the trypsin molecule to the catalytic and trypsin inhibitor binding sites. They hypothesised instead that the R117H mutation eliminated a “fail safe” mechanism for the inactivation of trypsin by abolishing an important autolytic site. Thus, the stabilised mutant enzyme would disrupt the trypsin activation/inhibition balance and trigger the pancreatic autodigestion process which results in pancreatitis under certain conditions. This model coincided with Chiara's pancreatitis hypothesis2 and has been supported by in vitro mutagenesis data. When the R117 of rat trypsin was replaced by other amino acids, the rate of autolysis of certain mutant enzymes was significantly slower than that of the wild type protein.8 9 While the R117H mutation has been shown to be a common mutation in HP by several laboratories world wide,10-13 further mutations in the cationic trypsinogen gene have been reported recently.13-15 These mutations are also presumed to facilitate the trypsin autodigestion process by altering either the tertiary structure of the protein or the binding of the pancreatic secretory trypsin inhibitor (PSTI).

However, mutations in the cationic trypsinogen gene do not appear to be the whole story. When 14 HP families from different regions of France were scanned for mutations in the cationic trypsinogen gene by denaturing gradient gel electrophoresis (DGGE) analysis, no mutations were detected in either part of the promoter region, in the intron/exon junctions, or in the gene coding sequence of six families. Furthermore, segregation analysis of one family with microsatellite markers (D7S640, D7S495, D7S684, D7S661, D7S676, D7S688) showed that the affected subjects had inherited two different haplotypes.13 Locus heterogeneity in HP was also suggested by the negative linkage and absence of the R117H mutation in two out of eight families studied by Dasouki et al.12 These findings, along with the incomplete penetrance of HP, indicated that another gene, or maybe even more than one, is involved in the pathogenesis of HP.

Human PSTI, a single chain polypeptide consisting of 56 amino acids, is also synthesised in the acinar cells of the pancreas. Its main physiological function is believed to be the prevention of the trypsin driven digestive enzyme activation cascade and of pancreatic autodigestion.16 17 Because of this central role of PSTI as a negative regulator of trypsin activity, it has been speculated that mutations in this gene may contribute to the development of pancreatitis.13 To date, no mutations have been reported in the human PSTI gene, which is located on chromosome 5.17 We therefore sought to investigate the possibility of mutations in the PSTI gene in a cohort of hereditary and sporadic chronic pancreatitis patients, as part of a continuing effort to gain further insight into the molecular basis of this disorder.

The human PSTI gene is approximately 7.5 kb long and is separated into 4 exons.17 By designing three exonic primer pairs (sequence not shown), we first successfully amplified the three introns of the PSTI gene from genomic DNA samples. The sizes of them were ∼1.7 kb, 1.5 kb, and 3.5 kb respectively, with a total length of ∼6.7 kb, which is within the range of 7.5 kb. The three PCR fragments were then cloned into the pGEM^®-T vector (Promega) and the inserts partially sequenced using T7 and SP6 promoter primers. Their identity was confirmed by comparing the resulting sequence with the published corresponding exon/intron boundary sequence.17 With the availability of the intronic sequence of ∼100 bp immediately flanking each exon, combined with the published sequence of the 5′ regulatory region and the 3′ untranslated region of thePSTI gene, five DGGE primer pairs were designed to allow for a complete scanning of the 334 bp DNA sequence upstream from the translation start point, as well as of all four exons and corresponding exon/intron junctions of the gene. Detailed information about gel preparation, buffer system, and electrophoresis apparatus for DGGE analysis has been described in our previous paper.18 Specifically, the primer sequence, annealing temperature, optimal linear gradient range, and migration time for each amplicon are set out in table 1. The presence of a DNA variant, which was indicated by an altered pattern in the DGGE analysis, was first confirmed by independent PCR/DGGE analysis. Then a second PCR was performed under the same conditions as for DGGE analysis and the resulting PCR product was cloned and sequenced on an ABI 310. For identifying heterozygous mutations, at least three colonies were sequenced using the T7 and SP6 promoter primers. Each DNA variant has been confirmed by reamplifying and resequencing from both strands in order to avoid artefacts introduced by PCR or sequencing errors.

Among the 14 French HP families we previously studied, the R117H mutation was detected in four families, a K23R mutation in one family, a N29I in two families, and a –28delTCC in one family. None of these cationic trypsinogen mutations was detected in the remaining six families.13 Considering the fact that a certain fraction of HP families do not carry the trypsinogen mutations, and ∼20% of the subjects carrying these mutations are non-symptomatic, we decided to analyse all the 14 families in order to screen for a possible disease causing mutation and also for a possible second mutation which may have an effect on phenotype in the PSTIgene. DGGE analysis and subsequent sequencing showed three DNA variants. They were –253T>C, IVS1-37T>C, and a missense mutation c.101A>G (resulting in N11S) respectively, named according to the recommendations for a nomenclature system for human gene mutations.19 The –253T>C variant was detected in two families with the R117H mutation, one family with the K23R mutation, and two families without trypsinogen mutations, both in affected patients and unaffected family members. Also, homozygosity was observed in one patient and one unrelated, disease free subject. Futhermore, this variant was not present in some patients in these families and its frequency evaluated in control chromosomes was ∼20%. Thus, the –253T>C variant is clearly a natural polymorphism. The IVS1-37T>C and c.101A>G (N11S) variants occurred together in one family without trypsinogen mutations and were present in the same haplotype. They have been classified as neutral polymorphisms based primarily on the fact that they did not segregate with the disease and that they were present in control chromosomes. Moreover, the IVS1-37T>C variant did not appear to affect the splice recognition sites and the c.101A>G variant did not replace the asparagine (N) at position 11 of the protein with an amino acid of different physical characteristics, although N11 is conserved in the human and two types of rat PSTI proteins.20

Owing to the similar clinical, laboratory, and pathological features of hereditary and sporadic chronic pancreatitis, we also undertook DGGE analysis of sporadic chronic pancreatitis. An additional three heterozygous DNA variants were separately identified in one out of 30 patients with sporadic chronic pancreatitis. The first was a C to T transition at position 163 of the PSTI cDNA, resulting in a proline (P) to serine (S) change at position 32 of the protein (c.163C>T (P32S)). This variant did not change a conserved amino acid and it was present in control chromosomes, indicating a harmless effect on phenotype. The second was a G to A substitution at position 41 upstream from the translation start site (−41G>A). This –41G is not conserved in the human and the two types of rat PSTI genes and is located ∼20 bp downstream from the main transcription start site. This suggests that the −41G>A substitution could not have a significant effect on the transcriptional or translational activity of the PSTI gene. We believe it to be a rare polymorphism as it was not detected in 400 control chromosomes. The third variant was a C to T transition at position 174 of the cDNA resulting in a silent mutation at position 35 of the protein (c.174C>T (C35C)). In addition, a heterozygous C to T substitution at position 22 upstream from the translation start site (−22C>T) was detected in two out of 200 control subjects (representing 400 chromosomes). All of the DNA variants detected in this study as well as their frequency evaluated in control chromosomes are described in table 1.

DGGE analysis is one of the most sensitive and efficient established mutation scanning techniques to date. It can allow for the discrimination of DNA molecules differing by as little as only one base change. Using this technique, we identified nearly 100% of mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene in a certain population.21 In this study, we detected up to seven different DNA variants. Given this high sensitivity of DGGE analysis, although we cannot exclude the possibility of mutations in the more upstream 5′ regulatory region or in the remaining intronic sequences, our results strongly suggested that the PSTIgene could be neither a cause nor a cofactor in the development of HP. Future research into this disease may be directed towards other pancreatic digestive proenzyme genes such as the anionic trypsinogen and mesotrypsinogen genes.22

When mutations in the cationic trypsinogen gene were identified as the molecular basis of HP, it was questioned whether they could predispose patients to develop sporadic pancreatitis. Until now, these cationic trypsinogen mutations have not been detected in sporadic chronic pancreatitis.15 23 In this study, although seven different DNA variants in the human PSTIgene were identified in sporadic chronic pancreatitis, none of them seems to have a functional effect on phenotype. Recently, mutations in the CFTR gene have been reported to be closely associated with this disorder24 25 and it would be interesting to look at whether CFTR also plays a role in the hereditary form of pancreatitis.

In conclusion, this study is the first comprehensive search for possible mutations in the human PSTI gene that may be linked to pancreatitis, and represents the first identification of seven DNA variants of the gene. Furthermore,PSTI has been excluded from involvement in the pathogenesis of hereditary and sporadic chronic pancreatitis.