|
|
Mutations of the Cystic Fibrosis Gene in Patients with Chronic Pancreatitis
Click here for a personal experience with the Cystic Fibrosis Gene in Patients with Pancreatitis
Nicholas Sharer, M.R.C.P., Martin Schwarz, Ph.D., Geraldine Malone, B.Sc., Andrea Howarth, M.Sc., John Painter, M.R.C.P., Maurice Super, F.R.C.P., and Joan Braganza, D.Sc. Data gathered at http://content.nejm.org/cgi/content/full/339/10/645 ABSTRACT Background The pancreatic lesions of cystic fibrosis develop in utero and closely resemble those of chronic pancreatitis. Therefore, we hypothesized that mutations of the cystic fibrosis transmembrane conductance regulator (CF TR ) gene may be more common than expected among patients with chronic pancreatitis. Methods We studied 134 consecutive patients with chronic pancreatitis (alcohol-related disease in 71, hyperparathyroidism in 2, hypertriglyceridemia in 1, and idiopathic disease in 60). We examined DNA for 22 mutations of the CF TR gene that together account for 95 percent of all mutations in patients with cystic fibrosis in the northwest of England. We also determined the length of the noncoding sequence of thymidines in intron 8, since the shorter the sequence, the lower the proportion of normal CFTR messenger RNA. Results The 94 male and 40 female patients ranged in age from 16 to 86 years. None had a mutation on both copies of the CF TR gene. Eighteen patients (13.4 percent), including 12 without alcoholism, had a CF TR mutation on one chromosome, as compared with a frequency of 5.3 percent among 600 local unrelated partners of persons with a family history of cystic fibrosis (P<0.001). A total of 10.4 percent of the patients had the 5T allele in intron 8 (14 of 134), which is twice the expected frequency (P=0.008). Four patients were heterozygous for both a CF TR mutation and the 5T allele. Patients with a CF TR mutation were younger than those with no mutations (P=0.03). None had the combination of sinopulmonary disease, high sweat electrolyte concentrations, and low nasal potential-difference values that are diagnostic of cystic fibrosis. Conclusions Mutations of the CF TR gene and the 5T genotype are associated with chronic pancreatitis.
In 1969, Paris et al.1 described two siblings who had the autosomal recessive disease cystic fibrosis and whose father and paternal uncle and grandfather had chronic calcific pancreatitis. Not only did this report hint at a shared molecular basis for pancreatic damage in these two conditions,2 but it also, since the children's mother was apparently unaffected, underlined the importance of naturenurture interactions in the pathogenesis of the sporadic form of chronic pancreatitis.3 The presentation of chronic pancreatitis typically resembles that of acute pancreatitis; subsequent attacks can be anticipated until all secretory parenchyma is destroyed. Alcoholism is a major etiologic factor, exposure to cigarette smoke and occupational exposure to volatile hydrocarbons independently increase the risk,4,5 duct-obstructing lesions initiate a few cases, and there may also be an underlying metabolic or autoimmune disorder. The cause of the rare hereditary form has been identified as a mutation in the cationic trypsinogen gene at locus 7q35.6 The disease is idiopathic in up to 40 percent of affected patients in developed countries. The exocrine pancreas is invariably affected in cystic fibrosis.7 The lesion has been described as "basically a diffuse form of chronic pancreatitis."8 The damage begins in utero7 and can be identified in neonates on the basis of elevated blood concentrations of pancreatic enzymes, classically trypsinogen. Pancreatic biopsy in the first year of life reveals interstitial inflammation,9 but this is not found on postmortem examination after the failure of pancreatic exocrine function. The progression of the disease is usually rapid and painless; however, a few patients may have an attack of pancreatitis or pancreatic calculi. In 1989, the cystic fibrosis transmembrane conductance regulator (CF TR) gene was identified at locus 7q31.10 This discovery led to the suggestion that insufficiency of the CFTR protein may underlie the overlapping clinicopathological facets of chronic pancreatitis and cystic fibrosis.2 As a first step toward testing this hypothesis, we examined the frequency of CF TR mutations in a cohort of patients with chronic pancreatitis. Methods Study Design In a study of 600 unrelated partners of persons with a family history of cystic fibrosis in the northwest of England, the rate of carriage of CF TR mutations was 5.3 percent (95 percent confidence interval, 3.5 to 7.1 percent).11 Therefore, we calculated that a minimum of 106 patients with chronic pancreatitis would have to be examined for the study to be able to detect, at a power of 90 percent, a doubling of this frequency. Study of the attendance register at the weekly pancreatobiliary clinic indicated that this target could be met within six months. We decided at the outset that patients who were found to have a CF TR mutation would undergo supplementary tests for atypical cystic fibrosis12 and that their first-degree relatives would be offered the opportunity to undergo screening for CF TR mutations.11 Patients The study was approved by the hospital's ethics committee. Between January 1993 and June 1993, consecutive white patients with chronic pancreatitis were enrolled after they gave informed consent. The diagnosis of chronic pancreatitis was based on standard criteria: abnormalities on histologic analysis of biopsy specimens, visible calculi on x-ray films, unequivocally abnormal findings on endoscopic pancreatography,13 or impaired exocrine secretory capacity, determined by the secretinpancreozymin test (bicarbonate or enzyme output more than 2 SD below the mean in normal subjects)13 or the p-aminobenzoic acid excretion index (results more than 3 SD below the mean in normal subjects in our version of this tubeless test, which uses bentiromide).14,15 Patients with a periampullary lesion that obstructed duct drainage were excluded. For the purposes of the study, alcoholism was defined as the daily intake of at least 80 g of ethanol by men and at least 60 g of ethanol by women for two years before the first symptom of pancreatitis, and cigarette smokers were defined as those who smoked 10 or more cigarettes per day.5 Those who drank lower amounts of alcohol were classified as nonalcoholics, and those who smoked zero to nine cigarettes per day were classified as nonsmokers. Job histories were also available, but we did not analyze these data because of the difficulty of quantifying occupational exposure to hydrocarbon.6 All patients were taking antioxidant supplements to control pain, and most had been taking them for about five years. The rationale for antioxidant therapy has been discussed previously.3,16,17,18 A frequent starting regimen consisted of six tablets containing organic selenium, beta carotene, and vitamins C and E (Wassen, Leatherhead, United Kingdom), and eight tablets of methionine (Evans Medical, Horsham, United Kingdom) per day in divided doses, for total daily supplements of 600 µg of organic selenium, 9000 IU of beta carotene, 0.54 g of vitamin C, 270 IU of vitamin E, and 2 g of methionine. Doses were adjusted after periodic measurement of blood vitamin C, selenium, and glutathione. DNA Studies We extracted
DNA from buccal cells obtained by having the patients rinse
their mouths with 10 ml of 4 percent sucrose.19
The CF TR locus was examined for the 22 mutations
that together account for 95 percent of all such mutations
in patients with cystic fibrosis in the northwest of England.20
The amplification-refractory mutation system Elucigene CF(4)m
kit (Zeneca Diagnostics, Macclesfield, United Kingdom) was
used to detect the four most common mutations: Assessment for Atypical Cystic Fibrosis Pulmonary spirometry and sinopulmonary radiography were performed whenever possible. Studies of the transepithelial nasal potential difference24 were undertaken in a subgroup of patients with a CF TR mutation and in comparable groups of patients with chronic pancreatitis and normal subjects with no CF TR mutations. The test involves the perfusion of a nasal mucosal electrode with standard buffer, then with 100 µM amiloride in standard buffer to block the epithelial sodium channel, then with 100 µM amiloride in low-chloride buffer to stimulate chloride movement, and finally with 100 µM amiloride and 10 µM isoproterenol in low-chloride buffer to increase intracellular cyclic AMP. Pilocarpine iontophoresis25 was used to obtain sweat samples from the same subgroups; a minimum of 100 mg of sweat was analyzed to determine sodium and chloride concentrations. Several of the younger patients had already undergone sweat testing at their initial presentation. Statistical Analysis A chi-square statistic with Yates' correction, a binomial distribution, the MannWhitney U test, and Fisher's exact test were used as appropriate. All P values were two-tailed. A P value of less than 0.05 was considered to indicate statistical significance.26 Results Characteristics of the Patients The cohort of 134 patients included 71 with alcohol-related pancreatitis, 3 with a metabolic problem (2 with hyperparathyroidism and 1 with severe hypertriglyceridemia), and 60 with idiopathic pancreatitis. There were 94 male and 40 female patients. The age at onset of symptoms varied widely, from 5 to 81 years, as did age at the time of the study (16 to 86 years). Ninety-nine patients smoked 10 or more cigarettes a day. Large-duct disease was identified on the basis of an abnormal pancreatogram or radiographic evidence of calculi in 80 percent of the patients, and small-duct disease was identified on the basis of histologic findings or impaired exocrine function in the others. An attack of pancreatitis was the usual presenting symptom, and increasing pain was the usual reason for referral. DNA Studies No patient
had a mutation on both copies of the CF TR gene. Eighteen
patients (13.4 percent; 95 percent confidence interval, 8.2
to 20.4 percent) had a CF TR mutation on one chromosome,
as compared with 32 of the 600 unrelated partners of persons
with a family history of cystic fibrosis (5.3 percent; 95 percent
confidence interval, 3.5 to 7.1 percent; P<0.001). The most
common mutation was
Analysis of the polyT sequence identified the 5T allele in 14 of 134 patients, or 10.4 percent (95 percent confidence interval, 5.8 to 16.9 percent); the frequency is 5.0 percent in the general population22 (P=0.008). It was present in 4 of the 18 patients with a CF TR mutation and in 10 (all males) of the other 116 patients (22.2 percent vs. 8.6 percent, P=0.10). The clinicopathological features of the patients with chronic pancreatitis classified according to whether they had a 5T allele or a CF TR mutation alone or in combination are summarized in Table 2. On the basis of the work of Chillón et al.,27 it is likely that the CF TR mutation and the 5T allele were on opposite chromosomes.
Assessment for Atypical Cystic Fibrosis None of the 18 patients with a CF TR mutation alone or in combination with a 5T allele met the diagnostic criteria for cystic fibrosis when all the evidence was considered.12 A review of family histories revealed cystic fibrosis in close relatives of two unrelated patients. Subsequent screening identified the disease in the outwardly healthy infant son of another patient. There
were no sinopulmonary symptoms or signs or radiologic abnormalities
in 133 patients. The one exception was a 76-year-old woman
(Patient 17 in Table 1) who was a former smoker with
the Nasal potential-difference tests (Figure 1) were interpreted with reference to published studies,24 after we confirmed that the pattern was abnormal in patients with classic cystic fibrosis by testing two such patients (data not shown). As compared with the mean (±SE) value in 12 normal subjects with no CF TR mutations (7.2±0.7) , the base-line value was significantly lower in the subgroup of patients with a CF TR mutation (10.5±1.2, P=0.02) but not in the subgroup with no CF TR mutations (8.1±0.5). However, no patient had a value that was diagnostic of cystic fibrosis (approximately 50 mV)12 and the results in our two patients with cystic fibrosis were close to this value (48 and 56 mV). Moreover, the patterns and gradients of responses to the various perfusates were similar in the three subgroups. Only one of the four patients who were heterozygous for both a CF TR mutation and the 5T allele agreed to be tested (Patient 10 in Table 1). His base-line value was the lowest recorded (17.6 mV), and it changed to 11.8 mV after exposure to amiloride representing a change of 32.9 percent, as compared with approximately 70 percent in patients with cystic fibrosis12 and changed to 20.7 mV with a low-chloride perfusate and to 22.5 mV after exposure to isoproterenol. Sweat tests showed a stepwise increase in electrolyte concentrations, with the normal subjects having the lowest concentrations and the patients with a CF TR mutation the highest concentrations (Figure 2). There was no significant change in electrolyte concentrations in five patients who were examined twice, at presentation and again for the study, while receiving antioxidant therapy. A sweat chloride concentration of at least 60 mmol per liter, which is suggestive of cystic fibrosis,12 was found in three patients, one with a normal CF TR genotype and two with a CF TR mutation but normal results on spirometry and nasal potential-difference tests (Patients 8 and 15 in Table 1). Alcoholism, but not cigarette smoking, as defined in this study, may have contributed to this outcome (mean chloride concentration, 51 mmol per liter in eight patients with alcoholism and 37 mmol per liter in 18 patients without alcoholism, irrespective of the CF TR genotype; P=0.02).
Male patients with cystic fibrosis frequently have azoospermia,12 as was found in a patient (Patient 1 in Table 1) who has been described previously in another context.28 Of the other nine male patients, four are fathers, one had normal results on semen analysis, and four declined to undergo semen analysis. CF TR Genotype and the Pancreatic Phenotype Among
patients with cystic fibrosis, mutations have been described that
result in pancreatic insufficiency (e.g., Discussion Our investigation had three linked objectives: to assess whether, because certain features of pancreatic involvement overlap in chronic pancreatitis and cystic fibrosis,2 CF TR mutations are more common than would be expected among patients with chronic pancreatitis; to determine whether patients with a mutation had an atypical form of cystic fibrosis12; and to seek clues to the possible role of CF TR mutations in the development of chronic pancreatitis.3 In a cohort of 134 patients, we found that the frequency of a CF TR mutation was nearly 2.5 times as high as expected and that the frequency of the 5T allele was twice as high as expected. CF TR mutations were associated with idiopathic rather than alcohol-related disease, as was the case in preliminary reports by others,30,31,32 and there was a high rate of mutations among nonsmokers or those who smoked fewer than 10 cigarettes a day. A diagnosis of atypical cystic fibrosis was not justified in these patients,12 but a few patients had a partial pattern of extrapancreatic involvement compatible with this diagnosis, and a more extensive molecular genetic analysis might have identified other such patients. However, on the basis of our findings, we conclude that chronic pancreatitis should be added to the list of conditions in which a mutant CF TR gene has pathogenetic importance. This list includes congenital absence of the vas deferens,27 nasal polyposis,33 diffuse bronchiectasis,34 and bronchopulmonary allergic aspergillosis in adults.35 There is a fundamental difference between our findings and those reported in patients with congenital absence of the vas deferens.27 Among patients with congenital absence of the vas deferens, 19 percent had a CF TR mutation in both copies of the gene without a 5T allele on either chromosome (a genotype that is calculated to reduce the percentage of normal CFTR mRNA to less than 3 percent of normal), and 34 percent had a 5T allele with a CF TR mutation on the opposite chromosome or were 5T homozygotes (genotypes that would result in a reduction of functional CFTR mRNA to between 8 and 12 percent of normal).23,27 In contrast, none of our patients with chronic pancreatitis had two CF TR mutations, 11 percent had a CF TR mutation and no 5T allele, and 3 percent had both a CF TR mutation and a 5T allele. These data suggest that CF TR mutations are a sufficient explanation for the problem with the development of the vas deferens in at least 50 percent of affected persons but that the relation between CF TR mutations and the development of chronic pancreatitis is more subtle. Ductal obstruction is generally regarded as the initiating event in both chronic pancreatitis and cystic fibrosis. However, this theory is undermined by several observations,2,3,13,36,37 as well as by histologic evidence to the contrary.7,8,9,28 We favor an alternative explanation wherein the acinar cell is a direct target2,3,38 and the damage is amplified when bicarbonate-producing epithelium is affected in a manner that reduces the pH within the intra-acinar space and lumen of ductules.39 The involvement of the CFTR protein in intracellular-vesicle targeting, movement of macromolecules, and membrane recycling,37,40 over and above its role in ion transport, is central to this concept. The sweat studies in our patients were not meant to be definitive but rather to document the findings in a newly characterized subgroup with a CF TR mutation, since increased sweat electrolyte concentrations have been reported in patients with alcohol-related chronic pancreatitis.41,42 We found that the electrolyte concentrations were independent of CF TR mutations, although concentrations were higher in patients with a mutation (Figure 2). It is impossible to assess the meaning of the lower base-line values for nasal potential difference in the patients with a CF TR mutation, because several ion channels contribute to the result.43 In conclusion, our study identifies mutations of the CF TR gene as a risk factor for chronic pancreatitis. Further studies are needed to explain why chronic pancreatitis does not develop in the majority of persons with a CF TR mutation and to examine the relation between a CF TR mutation and a mutation in the cationic trypsinogen gene.44
Abstracts of this work were presented at a meeting of the British Society of Gastroenterology, Manchester, United Kingdom, September 1820, 1996, and the International Conference on Cystic Fibrosis, Jerusalem, Israel, June 1721, 1996. We are indebted to Zeneca Diagnostics for the amplification-refractory mutation system kits, to Dr. L.P. Hunt for statistical advice, to Dr. K. Southern for instruction on nasal potential-difference methods, to Mr. D. Marland for sweat electrolyte analyses, to Ms. S. Turner and Ms. B. Bosch for respiratory spirometry, to Ms. T. Roberts for CFTR screening, to Mrs. S. Postill for nursing assistance, and to Ms. T. Knott for secretarial help. From the Pancreato-Biliary Unit (N.S., J.P.) and University Department of Medicine (J.B.), Manchester Royal Infirmary; and the Regional Genetic Service, Royal Manchester Children's Hospital (M. Schwarz, G.M., A.H., M. Super) both in Manchester, United Kingdom. Address reprint requests to Dr. Braganza at Manchester Royal Infirmary, Oxford Rd., Manchester M13 9WL, United Kingdom. References
Related Letters:
Correction:
Mutations of the Cystic Fibrosis Gene in Patients with Chronic Pancreatitis
Mutations of the Cystic Fibrosis Gene and Pancreatitis This article has been cited by other articles:
|






