Main

Chronic pancreatitis is a chronic inflammatory process of the pancreas characterized by pain and progressive fibrosis leading to exocrine and/or endocrine insufficiency. Alcoholic and idiopathic chronic pancreatitis account for the majority of cases. Patients with chronic pancreatitis are at increased risk for developing pancreatic adenocarcinoma.

Several genetic factors have been identified to predispose persons to chronic pancreatitis. Genetic alterations in the cationic trypsinogen gene (PRSS1) cause hereditary pancreatitis (Online Mendelian Inheritance in Man [OMIM] 167800), and mutations in the serine protease inhibitor, Kazal type 1 (SPINK1) (OMIM 167790), and the cystic fibrosis transmembrane conductance regulator (CFTR) (OMIM 602421) genes have been associated with chronic pancreatitis.1–5 Because these alterations are present only in a subset of patients with chronic pancreatitis, other genetic determinants are likely important in predisposing humans to the disease. Homozygous and compound heterozygous alterations in the HFE gene have been reported in a family with hereditary pancreatitis and segregated with the disease.6

Hereditary hemochromatosis (HH) (OMIM 235200) is a metabolic disease characterized by accumulation of excess iron in a variety of organs, leading to their failure and the development of serious illnesses, such as cirrhosis, diabetes mellitus, cardiomyopathy, and arthritis. Homozygous p.C282Y alteration in the HFE gene is associated with iron overload (serum ferritin levels and transferrin saturation) and was reported to account for the majority of cases of HH.7 However, p.C282Y homozygosity is relatively common in the general population and by itself is insufficient to cause HH.8 A second alteration (p.H63D) in a compound heterozygous (pC282Y years/pH63D) or homozygous state (pH63D/pH63D) is associated with mild to moderate iron overload and found in a minority of patients with HH.7 Large population-based studies recently reported slightly higher transferrin saturation in heterozygous carriers of both variants, whereas no differences were seen in serum ferritin levels.8,9 Patients with HH are at high risk for developing hepatocellular carcinoma, and carriers of HFE variants have been reported to bear an increased cancer risk.10

Iron has been proposed to play a significant role in the development of alcoholic liver disease by exacerbating oxidative stress-induced tissue damage. In response to iron, hepatic stellate cells produce excess collagen and other matrix proteins leading to fibrosis.11 Although similar mechanisms are crucial for the development of pancreatic fibrosis, the effect of iron on the pancreatic stellate cell has not been well studied. In a mouse model of hemochromatosis, iron overload in the pancreas resulted in fibrosis of the gland with massive macrophage infiltration and loss of secretory epithelium.12 In patients with hemochromatosis, the development of diabetes is an irrevocable step in disease progression that leads to premature death, even if excess iron is subsequently removed.

It is known that chronic excessive alcohol consumption accentuates the clinical expression of hemochromatosis. Cirrhosis is up to nine times more likely in subjects with hemochromatosis who consume more than 60 g of ethanol per day.13 However, no effect of HFE heterozygosity on the development and severity of alcoholic cirrhosis has been observed in most studies.14–16 Nevertheless, HFE alterations have been shown to contribute to hepatic iron accumulation17 and to severity of fibrosis in nonhereditary liver disease, such as chronic hepatitis C.18 Some investigators have shown an association of p.C282Y and nonalcoholic fatty liver disease19 and porphyria cutanea tarda.20

The aim of our investigation was to determine whether heterozygous p.C282Y and p.H63D alterations are associated with chronic pancreatitis or pancreatic adenocarcinoma.

MATERIALS AND METHODS

Patients

After ethics committee approval, informed consent was obtained and whole blood was collected from 958 consecutive patients with chronic pancreatitis, acute pancreatitis, and pancreatic adenocarcinoma. The patients were recruited at all of the participating institutions between 1997 and 2003 and originated from Germany (n = 884), Austria (n = 11), and Finland (n = 63). The diagnosis of chronic pancreatitis was based on previously published criteria.2,21 Patients were classified according to their diagnosis:2,22,23 alcoholic pancreatitis (349 patients: 290 male, 59 female; mean age 50 years); idiopathic pancreatitis (231 adult patients: 134 male, 97 female; mean age 49 years; and 112 children [50 male, 62 female; mean age 8 years]); and familial pancreatitis (64 patients, 25 male, 39 female; mean age 8 years). In addition, 34 patients with acute pancreatitis (18 male, 16 female; mean age 52 years) and 168 patients with histologically confirmed pancreatic adenocarcinoma (101 male, 67 female; average age 62 years) were included. A subset of our patients (n = 157) were children. For 36 children, blood for genetic analysis was available from both parents; for nine children, blood was available from one parent. Each participating institution also collected blood from randomly selected blood donors to ensure that the controls originated from the same population (588 of German and 93 of Finnish origin, collected between 1997 and 2003). Blood was also collected from 100 persons with chronic alcohol abuse of German origin without a history or symptoms of acute or chronic pancreatitis (recruited at Central Institute of Mental Health, Mannheim for genetic association studies). In a subset of patients (n = 135), hospital charts were reviewed and data regarding presence of diabetes mellitus were recorded.

Mutational analysis

Genomic DNA was isolated from anticoagulated blood by standard procedures and subjected to polymerase chain reaction (PCR) amplification using a Primus 95 Plus cycler (MWG Biotech, Ebersberg, Germany) as previously described.24,25 PCR products were subjected to enzymatic digestion using RsaI and MboI because p.C282Y creates an additional restriction site for RsaI and p.H63D destroys an MboI restriction site. Restriction products were separated on a 3% NuSieve (Teknova, Hollister, CA) agarose gel (RsaI) and 3% agarose gel (MboI), and visualized by ethidium bromide staining. A subset of samples were analyzed by PCR amplification followed by a melting curve analysis using fluorescence resonance energy transfer probes and the LightCycler (Roche Diagnostics, Basel, Switzerland), as published previously with slight modifications.26

Statistical analysis

Differences between groups were compared with the χ2 and Fisher exact tests. Correction for multiple testing according to Bonferroni was applied where appropriate. Disease-allele association in parent–offspring trios was tested by the transmission disequilibrium test.27 A P value less than 0.05 was considered significant. Power analysis was performed (Statistical Package for the Social Sciences software; SPSS Inc., Chicago, IL) using the previously reported p.C282Y and p.H63D genotype frequencies from German and Finnish populations.28,29 The difference in frequency corresponding to odds ratios of at least 1.6 (p.C282Y) and 1.4 (p.H63D) could be detected at a power of 0.8 for the given sample sizes in patients with chronic pancreatitis and 2 and 1.7 in patients with pancreatic adenocarcinoma.

RESULTS

The demographic characteristics of the study population are summarized in Table 1. The genotype and allele distribution of the studied HFE alterations in patients and controls of German and Finnish origin are shown in Table 2. In the study population of German origin, no statistically significant difference was observed in genotype and allele frequencies between patients with alcoholic pancreatitis, idiopathic pancreatitis, familial chronic pancreatitis, acute pancreatitis, and pancreatic ductal carcinoma when compared with healthy and alcoholic controls. In the study population of Finnish origin, the distribution of genotypes and alleles was similar between patients with chronic alcoholic pancreatitis and healthy controls. Furthermore, the genotype and allele frequencies were similar and not statistically different between patients and controls of German and Finnish origin. No significant differences were detected between patients with chronic pancreatitis of different cause. Analysis of 19 informative parent–offspring trios did not reveal an increased transmission of either allele. p.H63D was transmitted in 7 of 15 families, p.C282Y in 2 of 4 trios (P > 0.05). The presence of secondary diabetes mellitus in patients with chronic pancreatitis was not associated with any specific genotype or allele (P > 0.05, Table 3).

Table 1 Demographic characteristics of the study population
Full size table
Table 2 Genotype and allele frequencies of the HFE gene p.C282Y and p.H63D variants
Full size table
Table 3 Genotype frequencies in patients with diabetes mellitus
Full size table

DISCUSSION

Most cases of chronic pancreatitis in industrialized countries can be attributed to alcohol abuse. However, only a subset of patients with heavy alcohol consumption develop pancreatitis. The cause of idiopathic chronic pancreatitis remains largely unknown. Some patients with a family history of pancreatitis have mutations in PRSS1 or SPINK1, while others do not. Therefore, it is hypothesized that new, yet unidentified risk factors are important in the disease development in these individuals. The discovery of genetic predispositions to chronic pancreatitis suggests that more may remain to be discovered.

The report of the p.C282Y variant segregating with the disease in a family with hereditary pancreatitis6 prompted us to investigate its role in a spectrum of sporadic and familial pancreatic diseases. To our knowledge this is the first study to focus on determining the frequency of the p.C282Y and p.H63D in patients with pancreatic disease. In the present study HFE alterations were not associated with alcoholic, idiopathic, or familial chronic pancreatitis. Furthermore, no association was seen in patients with pancreatic adenocarcinoma. In addition to healthy controls, we used patients with chronic severe alcohol abuse without chronic pancreatitis as a control group for patients with alcoholic pancreatitis. Nevertheless, comparison of patients with alcoholic pancreatitis, alcoholic patients without pancreatitis, and healthy controls did not reveal any significant differences. Our results agree with those of Frenzer et al.,15 who, focusing on alcoholic cirrhosis, used a small number of patients with chronic pancreatitis as controls. In our study the observed genotypes were comparable to the previously reported data from population studies in Germany and Finland.28,29

We hypothesized that HFE variants acting through altered iron metabolism may predispose patients to fibrosis or carcinogenesis. Several possible factors may account for the lack of association between HFE genotypes and various pancreatic diseases. First, iron-related genes other than HFE may be important. Recently, other genes that play significant roles in maintaining iron homeostasis were identified.30 Mutations in hepcidin (HAMP), iron-importer protein transferring receptor 2 (TFR2), and hemojuvelin (HJV) have been identified in patients with altered iron metabolism.7,30 Genetic alterations in these genes were not explored in our study. Second, because the HFE variants alone did not present a predisposing factor in our patients, a relationship between HFE and other genes that play a role in iron metabolism or cellular antioxidant defense may be important. The HFE wild-type gene product complexes with the transferrin receptor (TFRC). The two investigated HFE variants have been shown to increase the affinity of the receptor for transferrin and increase cellular iron uptake. Beckman et al.,31 when studying multiple neoplastic disorders, found no association for HFE and TFRC genotypes separately, but an interaction between p.C282Y and p.S142G variants was found in multiple myeloma, breast, and colorectal cancer. Third, because we did not determine markers of iron metabolism in blood or tissue, we cannot completely exclude the role of iron in predisposing patients to pancreatitis or pancreatic cancer. However, even an obvious biochemical phenotype of iron overload rarely results in significant organ damage, suggesting that a simple association is unlikely.32

Because the frequency of diabetes mellitus in patients with HH is high, data regarding presence of secondary diabetes mellitus in our patients were recorded. However, neither alteration was a risk factor for development of diabetes mellitus in our study population.

In summary, our study provides no evidence that carriers of p.C282Y, p.H63D, or both have an increased risk for developing chronic alcoholic, idiopathic, and familial pancreatitis or pancreatic adenocarcinoma.