- Email: [email protected]
Tropical calcific pancreatitis: Strong association with SPINK1 trypsin inhibitor mutations
GASTROENTEROLOGY 2002;123:1020 –1025
Tropical Calcific Pancreatitis: Strong Association With SPINK1 Trypsin Inhibitor Mutations EESH BHATIA,* GOURDAS CHOUDHURI,‡ SADIQ S. SIKORA,§ OLFERT LANDT,¶ ANDREAS KAGE,㛳 MICHAEL BECKER,# and HEIKO WITT# *Departments of Endocrinology, ‡Gastroenterology, and §Surgical Gastroenterology, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India; 㛳Institute of Laboratory Medicine and Pathobiochemistry, Charite´, ¶TIB MOLBIOL, Berlin, Germany; and #Kinderklinik, Charite ´ , Humboldt-University, Berlin, Germany
Background & Aims: Tropical calcific pancreatitis (TCP) is a chronic pancreatitis unique to developing countries in tropical regions. The cause of TCP is obscure. Whereas environmental factors, such as protein energy malnutrition and ingestion of cassava, have been implicated, a genetic predisposition to the disease also may be important. In the present study we report on mutations in the serine protease inhibitor, Kazal type 1 (SPINK1) gene in north Indian patients with TCP. Methods: We studied 66 unrelated TCP patients (44 men, 49 with diabetes, and 6 with family history of TCP), 25 relatives, and 92 healthy control subjects. Samples were analyzed for SPINK1 variants (ⴚ53C>T, L14P, N34S, P55S, and 272T>C) and cationic trypsinogen (PRSS1) variants (A16V, K23R, N29I, and R122H) by melting curve analysis. Results: Twenty-nine patients (44%) carried the N34S missense mutation, of whom 9 (14%) were homozygotes. In contrast, only 2 (2.2%) control subjects were N34S heterozygotes (prevalence ratio 20.2; 95% confidence interval 5.0 – 81.8; P < 0.0001 vs. TCP). The severity of pancreatitis did not differ between TCP patients with or without N34S, or among those heterozygous or homozygous for N34S. Among TCP patients with or without diabetes, the frequency of N34S carriers (43% vs. 47%) and N34S homozygotes (14% vs. 12%) was similar. Conclusions: TCP is highly associated with the SPINK1 N34S mutation. The high prevalence of N34S in TCP patients with and without diabetes suggests that these 2 subtypes have a similar genetic predisposition. The genetic predisposition to TCP resembles, at least in part, the idiopathic chronic pancreatitis found in industrialized countries.
ropical calcific pancreatitis (TCP) is an idiopathic chronic pancreatitis (ICP) unique to nonindustrialized countries located in the tropics.1– 8 Patients present at a young age with recurrent abdominal pain and nutritional deficiencies. The chronic pancreatitis leads to progressive beta-cell deficiency, and patients often develop insulin-requiring diabetes (fibrocalculous pancreatic diabetes, [FCPD]), before age 30.1– 4,8 In older reports
T
the prognosis was described as dismal, with many patients succumbing to complications related to malnutrition and infection.1,2 The origin of TCP is obscure.1,2 Specific predisposing factors, such as alcohol abuse, biliary tract stones, and metabolic disease, are absent. It has been proposed that environmental factors, such as protein energy malnutrition6 or the consumption of cassava (a source of cyanogenic glycosides),7 may play a pathogenic role. However, TCP is frequently found in regions where cassava is not consumed4,8 and is also observed in patients from higher economic levels, in whom malnutrition is unlikely.1 The familial clustering of TCP suggests that genetic defects may predispose to the disease.9 Environmental factors may possibly be operative on the background of a genetic predisposition to pancreatitis. However, no association has been found among the different candidate genes studied, including the reg1A (pancreatic stone protein) gene10 or the cationic trypsinogen (PRSS1) gene.11,12 In a preliminary study, 2 of 18 north Indian TCP patients had heterozygous mutations of the cystic fibrosis transmembrane regulator (CFTR) gene.13 However, this frequency was similar to that in the healthy white population. Chronic pancreatitis is thought to be the result of inappropriate trypsin activity within the pancreatic parenchyma.14 Gain-of-function mutations of the PRSS1 gene have been found to be associated with chronic pancreatitis.15–18 More recently, mutations in the serine protease inhibitor, Kazal type 1 (SPINK1), an inhibitor of intrapancreatic trypsin activity, have been reported to Abbreviations used in this paper: CFTR, cystic fibrosis transmembrane regulator; FCPD, fibrocalculous pancreatic diabetes; FRET, fluorescence resonance energy transfer; ICP, idiopathic chronic pancreatitis; PRSS1, cationic trypsinogen; SPINK1, serine protease inhibitor, Kazal type 1; TCP, tropical calcific pancreatitis. © 2002 by the American Gastroenterological Association 0016-5085/02/$35.00 doi:10.1053/gast.2002.36028
October 2002
be associated with chronic pancreatitis in Europe and the United States.19 –24 These mutations are thought to result in a decreased inhibitory capacity within the pancreas.19 A substitution of asparagine by serine (N34S) in exon 3 is the most common SPINK1 mutation in patients with ICP.19 –24 Recently, in a preliminary study, Rossi et al.25 reported that SPINK1 mutations are associated with FCPD, but not with TCP without diabetes. In the current study, we report on SPINK1 gene mutations in a large cohort of TCP patients, families, and control subjects from north India. Our findings suggest that SPINK1 mutations are strongly associated with TCP, both with and without diabetes.
Materials and Methods Subjects TCP was diagnosed by a history of recurrent abdominal pain and radiologic evidence of pancreatic intraductal calculi. Patients with a history of alcohol intake were excluded, as were those with other causes of chronic pancreatitis, including metabolic disorders (hypercalcemia, hypertriglyceridemia), biliary duct stones, and anatomic anomalies (pancreas divisum). Fasting plasma glucose was measured in all subjects, and diabetes was diagnosed using the criteria of the American Diabetes Association (1997).26 Based on the fasting plasma glucose level, TCP patients were classified into TCP without diabetes and FCPD. Patients receiving treatment with insulin or oral hypoglycemic drugs were classified as FCPD, irrespective of their current plasma glucose level. We studied 66 unrelated patients (44 men) with TCP who attended the Endocrinology and Gastroenterology services at the Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, between January 2000 and September 2001. All patients belonged to the north Indian state of Uttar Pradesh or to adjoining regions. None consumed cassava or gave a history suggestive of malnutrition in childhood. The age (mean ⫾ standard deviation) of the subjects was 31.5 ⫾ 10.9 years, and age at onset of pain was 19.7 ⫾ 9.9 years. Fifty-eight percent of patients had onset of pain ⬍20 years. Twenty-seven patients (41%) had undergone surgery or an endoscopic procedure for pain relief. The patients’ body mass index (BMI) was 18.1 ⫾ 3.4 kg/m2; a low BMI (⬍18 kg/m2) was found in 53% of patients. The patients had severely diminished exocrine function (fecal chymotrypsin 2.2 ⫾ 2.2 U/g stool (n ⫽ 39); normal ⬎8.4 U/g stool). Forty-nine patients (74%) had FCPD; of these, 40 (82%) required insulin therapy. The clinical characteristics of FCPD and TCP subjects without diabetes are shown in Table 1. A family history of TCP was present in 6 (9%) index patients. We studied 17 relatives (8 additional subjects with TCP and 9 unaffected) from these 6 families. In addition, we studied 8 unaffected relatives from 3 different families, where only the proband had TCP. Ninety-two healthy north Indian subjects (physicians and hospital staff), belonging to the same
TCP AND SPINK1 TRYPSIN INHIBITOR MUTATIONS
1021
Table 1. Clinical Characteristics of Patients With Tropical Calcific Pancreatitis
Age (yr) M:F Age at onset of pain (yr)a Body mass index (kg/m2) Age at onset of diabetes (yr) Fecal chymotrypsin (U/g) (normal ⬎8.4 U/g) Surgery for pain relief Fasting plasma glucose (mg %) Plasma glucose 2 hours postprandial (mg %)
TCP with diabetes (FCPD) (n ⫽ 49)
TCP without diabetes (n ⫽ 17)
32.6 ⫾ 10.1 33:16 19.3 ⫾ 9.6 18.0 ⫾ 3.4 28.1 ⫾ 7.8 2.4 ⫾ 2.3 (n ⫽ 29) 12 (24%) 213 ⫾ 78 (70–427) 351 ⫾ 105 (n ⫽ 44)
28.0 ⫾ 13.0 11:6 20.7 ⫾ 13.0 18.5 ⫾ 3.5 – 1.8 ⫾ 1.9 (n ⫽ 10) 15 (88%)b 77 ⫾ 16c (60–109) –
NOTE. Mean ⫾ standard deviation (range). TCP, tropical calcific pancreatitis; FCPD, fibrocalculous pancreatic diabetes. a The age at onset of pain was accurately given by 60 patients. b P ⬍0.0001; c P ⬍0.0001 Student t test.
region, served as controls. Informed written consent was obtained from all subjects, and the study was approved by the institutional ethics committee.
Methods Genomic DNA was extracted from peripheral blood leukocytes by standard techniques. Coded DNA samples of the patients and the control subjects were analyzed for PRSS1 variants (A16V, K23R, N29I, and R122H), and SPINK1 variants (⫺53C⬎T, L14P, N34S, P55S, and 272T⬎C) by melting curve analysis.
Polymerase Chain Reaction Primers flanking the coding region of the PRSS1 and SPINK1 genes were designed according to the nucleotide sequences published by Rowen et al. (PRSS1; GenBank #U66061) and Witt et al. (SPINK1; GenBank #AF286028). With the exception of the forward primer for exon 3 of SPINK1 (5⬘-TAT GAC CCT GTC TGT GGG AC-3⬘), all primers used have been described previously.16,19 Polymerase chain reaction (PCR) was performed using 0.5 U AmpliTaq Gold polymerase (Perkin Elmer, Braunschweig, Germany), 400 mol/L dNTPs, and 0.1 mol/L of each primer in a total volume of 25 L. Cycle conditions were an initial denaturation for 12 minutes at 95°C, followed by 40 cycles of 15 seconds denaturing at 95°C; 30 seconds of annealing at 54, 56, 58 or 64°C; 40 seconds of primer extension at 72°C; and a final extension for 2 minutes at 72°C in an automated thermal cycler (Biometra, Go¨ttingen, Germany).16,19
Melting Curve Analysis Melting curve analysis for the N34S mutation was performed using a pair of fluorescent resonance energy transfer (FRET) probes and a LightCycler (Roche Diagnostics, Mannheim, Germany), as previously described.19 FRET probes were
1022
BHATIA ET AL.
designed and synthesized by TIB MOLBIOL (Berlin, Germany). For the other variations, the following FRET probes were used: SPINK1: ⫺53C⬎T, 5⬘-TTC TGC GTC CAG AGG TCA GTT GAA AAC T-FL and 5⬘-LC Red 640-ACC GCA CTT ACC ACA TCT CTT CAG; L14P, 5⬘-GAA GGT AAC AGG CAT CTT TCT TCT CAG TG-FL and 5⬘-LC Red 705-CTT GGC CCC GTT GAG TC; P55S, 5⬘-CAC GCA TTC ATT GGA ATA AGT AT-FL and 5⬘-LC Red 705-CCA TCA GTC CCA CAG ACA GG; 272C⬎T, 5⬘-CCT CGC GGT GAC CTG ATG GGA-FL and 5⬘-LC Red 640-TTC AAA ACC TTG GTT CTC AGC AAG GCC. PRSS1: A16V, 5⬘-GAC TTG CCT TCT CCC TTC CCA TCT CCA-FL and 5⬘-LC Red 640-CCA GTT GCT GCC CCC TTT GAT; K23R and N29I, 5⬘-AGG GAC ACC TGG TAG GGG ACA GAA TTC TCC-FL and 5⬘-LC Red 705-TGT AGC CCC CAA CGA TCC TGT CAT CAT (K23R) and 5⬘-LC Red 640-CAC AGT TGT AGC CCC CAA (N29I); R122H, 5⬘-CAA CGC CCA CGT GTC CAC CA-FL and 5⬘-LC Red 705- TCT CTG CCC ACC GCC CCT CCA GCC.
DNA Sequencing A novel variation detected by melting curve analysis [272C⬎T; 274C⬎T] was analyzed by direct DNA sequencing. We purified PCR products with spin columns and performed cycle sequencing using 1.5 L purified template, 3 L BigDye terminator mix (Applied Biosystems), and 3 L forward or reverse primer (0.3 mol/L). The reaction products were purified with spin columns and loaded onto an ABI 373A fluorescence sequencer (Applied Biosystems). DNA sequences were analyzed by sequencing both strands.
Statistical Analysis Continuous data were compared using the Student t test, and categorical data were compared using 2 test or the Fisher exact test. A 2-tailed P value ⬍0.05 was considered significant.
Results SPINK1 Mutations The frequency of SPINK1 mutations in the cohort of 66 unrelated index patients (FCPD, 49 patients; TCP without diabetes, 17 patients) and in 92 control subjects is shown in Table 2. Twenty-nine patients (44%) carried the N34S mutation, in contrast to only 2 (2.2 %) control subjects (prevalence ratio, 20.2; 95% confidence intervals, 5.0 – 81.8; P ⬍ 0.0001). Nine patients (14%) were homozygous for the N34S mutation, whereas both control subjects were heterozygous carriers. Comparing TCP patients with an earlier (⬍20 years) or later (⬎20 years) age at onset of pain, the frequency of N34S carriers (54% vs. 36%) or N34S homozygotes (14% vs. 16%) was not significantly different. The severity of pancreatitis (based
GASTROENTEROLOGY Vol. 123, No. 4
Table 2. SPINK1 Variants in Subjects With Tropical Calcific Pancreatitis and Controls
SPINK1 variants
Controls (n ⫽ 92)
All TCP patients (n ⫽ 66)
Ratio (all TCP patients: controls)
N34S (homozygotes ⫹ heterozygotes) N34S homozygotes N34S heterozygotes
2 (2.2) 0 2 (2.2)
29 (43.9) 9 (13.6) 20 (30.3)
20.2 [5.0–81.8]a Not calculatedb 13.9 [3.4–57.6]a
NOTE. Figures in parentheses are the percentage of patients and controls carrying the mutation. Prevalence ratios and 95% confidence intervals compare the frequency of individuals with different abnormal SPINK1 genotypes in the TCP vs. control groups. No ⫺ 53 C ⬎ T or L14P variations were observed; there were no compound N34S/P55S heterozygotes; 1 patient was compound heterozygous for N34S and [272C ⬎ T; 274C ⬎ T]. The frequency of the 272C ⬎ T variant was similar in control subjects (24 of 92, 26.1%) and TCP patients (28 of 66, 42.4%). a P ⬍ 0.0001, Fisher exact test. b P ⫽ 0.0003, Fisher exact test.
on age at onset of pain, frequency of diabetes, and surgical intervention for pain relief) did not differ between TCP patients with or without the N34S mutation or between subjects heterozygous or homozygous for N34S. Among the 49 tested patients who had FCPD, 21 (43%) carried an N34S mutation. Of these, 7 (14%) were N34S/N34S and 14 (29%) were N34S/WT. Of the 17 patients who had nondiabetic TCP, 8 (47%) carried an N34S mutation; 2 (12%) were N34S/N34S and 6 (35%) were N34S/WT. The frequency of N34S heterozygotes or homozygotes did not differ between FCPD and nondiabetic TCP patients. Of the 9 patients homozygous for N34S, 7 (78%) had diabetes; this frequency of diabetes was similar to that in subjects not carrying the N34S homozygous mutation (42 of 47, 74%). The frequency of the P55S and 272C⬎T variations were similar in patients and control subjects, and no ⫺53C⬎T or L14P mutations were detected. In the assay for the 272T⬎C polymorphism, 3 patients and 2 control subjects showed an unknown melting curve. DNA sequencing revealed a C–T transition at nucleotide 274, together with the frequent 272C⬎T polymorphism on the same allele [272C⬎T; 274C⬎T]. Of the 3 families studied with a single affected TCP patient, 2 of 8 unaffected family members carried the N34S mutation (1 heterozygous and 1 homozygous). In one of these families, the proband (onset of pain at age 17 years, insulin requiring diabetes at 22 years) carried a heterozygous N34S mutation, inherited from the unaffected mother. The proband’s wife, though asymptomatic, was an N34S carrier. A history of consanguinity was denied. The patient’s unaffected son (age 8 years, no
October 2002
TCP AND SPINK1 TRYPSIN INHIBITOR MUTATIONS
1023
Discussion
Figure 1. Families with heterozygous or homozygous N34S carriers. Dark squares or circles denote affected family members. N34S/ N34S, homozygotes; N34S/WT, heterozygotes; WT/WT, wild-type.
abdominal pain, normal pancreatic ultrasonography and plasma glucose level) was homozygous for N34S. PRSS1 Mutations Mutations in PRSS1 were absent in all TCP patients, including those with a family history of pancreatitis, and in all control individuals. SPINK1 Mutations in Familial TCP Of the 6 families in which at least 2 family members had TCP, the N34S SPINK1 mutation was detected in 4. Of the 10 individuals with TCP in these 4 families, 7 carried the N34S mutation. In 1 of the families, both of the affected siblings with FCPD were homozygous for N34S and inherited the mutation from asymptomatic heterozygous parents (Figure 1A). In the 3 other families, the TCP patients were heterozygous for N34S. In 1 of these families, the index patient (nondiabetic TCP) and the mother and an aunt (both FCPD) carried the heterozygous N34S mutation (Figure 1B). In another family, the index FCPD patient carried a heterozygous N34S mutation, but no DNA was available for genetic testing from the affected sister. In the fourth family, the index patient (FCPD) was heterozygous for N34S, but 3 other affected family members (also FCPD) carried the wild-type allele. In the fifth family, the proband and his affected nephew were heterozygous for the P55S variation, and the father and brother of the proband were heterozygous carriers and were asymptomatic. Finally, in the last family, neither the proband nor her affected father had any SPINK1 mutation.
Although TCP has certain clinical features similar to idiopathic chronic pancreatitis, distinctive features include an early age of onset, large intraductal calculi, and a progressive course with high frequency of diabetes.1 Our study provides evidence that TCP is strongly associated with the N34S mutation in the SPINK1 gene. Approximately 50% of the TCP patients were carriers for N34S, whereas 14% were homozygous for this mutation. Thus, in terms of this genetic predisposition, TCP is similar to idiopathic chronic pancreatitis found in industrialized countries.19 –24 The association of the same SPINK1 mutation with chronic pancreatitis in 2 different racial groups strengthens the hypothesis that this mutation, or another mutation closely linked to it, predisposes for disease. Some authors have tried to distinguish between FCPD and TCP on the basis of clinical findings.27 Nonetheless, TCP without diabetes and FCPD are indistinguishable on the basis of their pathologic findings.1,4,28,29 Within the clinical spectrum of TCP, the features of pain, diabetes, and exocrine deficiency occur in varying combinations, and onset of diabetes is usually associated with an advanced stage of TCP.4 In TCP, islet beta-cell function correlates with exocrine function,30 diminishing with increasing duration of disease.31 Thus FCPD is likely to result from a progressive impairment of beta-cell function in patients with tropical pancreatitis. Recently, Rossi et al.25 reported that TCP and FCPD have a separate genetic background. This was based on their findings that 4 of 6 subjects with FCPD, but 0 of 3 TCP patients without diabetes, carried a SPINK1 mutation. However, in our study, we found a similar high prevalence of SPINK1 mutations in both of these subgroups. This suggests that the genetic predisposition to pancreatitis is similar in FCPD and TCP without diabetes. Whether, as proposed by Kambo et al.,32 other genetic loci (e.g., HLA class II genes, insulin gene) confer susceptibility to the development of diabetes requires further study. In the families studied, TCP was associated with different patterns of N34S inheritance. In 1 family, 2 affected children were homozygous for N34S, whereas the unaffected parents were heterozygous. This finding could indicate a recessive inheritance pattern. However, in another family, the 3 affected members were heterozygous for N34S. In another family, only the index patient was N34S heterozygous, whereas 3 other affected family members carried a wild-type allele. Finally, in a simplex family, the proband was N34S/WT; his N34S homozygous offspring was found to have no evidence of TCP. In
1024
BHATIA ET AL.
summary, our data suggest that despite the striking association between N34S and TCP, this genetic alteration acts in a manner that is presently unclear. The prevalence of TCP in India is likely to be higher than that of ICP in industrialized countries.1,4 While the prevalence of TCP in the general population in north India has not been studied, a field survey from the Quilon district in Kerala, south India, reported a prevalence of 0.12%.5 Because the aforementioned survey was conducted in a district “endemic” for the disease, the prevalence of TCP in other regions of India may be lower than this figure. Clinic based estimates of FCPD are variable, but range from 5% to 30% of the patients with diabetes below age 30 years.2,4 In addition, it has also been reported that a subclinical “pancreatopathy” (as suggested by reports of elevated serum trypsin in asymptomatic subjects) is frequent in healthy Indian subjects.33 This high rate of clinical and sub-clinical pancreatic disease may be related in part to the N34S mutation, but other (hitherto unknown) genetic factors or unique environmental factors are also likely to play a role. We found a prevalence of N34S mutations in healthy subjects of 2.2%, though this estimate is unstable since our control population was small (n ⫽ 92), and data from other studies are not available for the north Indian population. Nevertheless, this prevalence of N34S carriers is far higher than that of TCP. In addition, there was no specific pattern for inheritance of the N34S mutation in families. Finally, N34S homozygotes did not reveal greater severity in their clinical features, when compared with heterozygotes. These data are analogous to that found in studies on ICP.19 –24 They would suggest that while a heterozygous N34S mutation is strongly associated with TCP, it may not be adequate to cause the disease without other genetic and/or environmental factors. In a study by Noone et al., CFTR mutations and the N34S mutation in the SPINK1 gene independently predispose to ICP.24 In a previous report on a small cohort of TCP patients, the frequency of CFTR mutations (0.08) was lower than that reported in ICP in Caucasian subjects (0.20-0.24).13 Of the 18 TCP subjects previously studied for CFTR mutations, 16 were analyzed in the current study. DNA from 1 of the 2 patients with CFTR mutations was not available. One patient, homozygous for the 5T allele of the CFTR gene, was heterozygous for N34S. Thus, while both SPINK1 and CFTR genetic loci predispose to TCP and ICP, they are likely to contribute in different proportions to the susceptibility. The P55S variation is not believed to be diseaseassociated with chronic pancreatitis.19,20,34 Similarly, in
GASTROENTEROLOGY Vol. 123, No. 4
the current study, the frequency of P55S did not differ in patients and control subjects. The [272C⬎T; 274C⬎T] alteration is also unlikely to be disease-predisposing, since the frequency of this variation was similar in TCP and control subjects. Mutations in the cationic trypsinogen (PRSS1) gene were not detected in the current study. This confirms earlier reports of the absence of PRSS1 mutations in TCP subjects.11,12 The absence of trypsinogen mutations is in contrast to ICP, where these are found in approximately 6% of children and adolescents.35 Recently, Noone et al. have also reported a PRSS1 mutation in four ICP patients without a family history of pancreatitis.24 In conclusion, there is a strong association between the N34S mutation of the SPINK1 gene and TCP, both with and without diabetes. The high prevalence of early-onset chronic pancreatitis in tropical countries may be the consequence of alterations in the SPINK1 gene, mutations in other genes, and/or environmental factors (e.g., protein energy malnutrition) which are unique to developing countries. The genetic predisposition to TCP is similar, in part, to idiopathic chronic pancreatitis found in industrialized countries.
References 1. Mohan V, Pitchumoni CS. Tropical calcific pancreatitis. In: Berger HG, Warshaw AL, Bu¨chler MW, eds. The pancreas. 1st ed. London: Blackwell Science, 1998:688 – 697. 2. Abu-Bakare A, Taylor R, Gill GV, Alberti KGMM. Tropical or malnutrition-related diabetes: a real syndrome? Lancet 1986;1:1135– 1138. 3. Viswanathan M. Pancreatic diabetes in India: an overview. In: Podolsky S, Viswanathan M, eds. Secondary diabetes: the spectrum of diabetic syndromes. New York: Raven, 1980:105–116. 4. Mohan V, Nagolimath SJ, Yajnik CS, Tripathy BB. Fibrocalculous pancreatic diabetes. Diabetes Metab Rev 1998;14:153–170. 5. Balaji LN, Tandon BN, Tandon RK, Banks PA. Prevalence and clinical features of chronic pancreatitis in Southern India. Int J Pancreatol 1994;15:29 –34. 6. Shaper AG. Chronic pancreatic disease and protein malnutrition. Lancet 1960;1:1223–1224. 7. McMillan DE, Geevarghese PJ. Dietary cyanide and tropical malnutrition diabetes. Diabetes Care 1979;2:202–208. 8. Bhatia E, Baijal SS, Kumar R, Choudhuri G. Exocrine pancreatic and -cell function in malnutrition related diabetes among North Indians. Diabetes Care 1995;18:1174 –1178. 9. Mohan V, Chari ST, Hitman GA, Suresh S, Madanagopalan N, Ramachandran A, Viswanathan M. Familial aggregation in tropical fibrocalculous pancreatic diabetes. Pancreas 1989;4:690 – 693. 10. Hawrami K, Mohan V, Bone A, Hitman GA. Analysis of islet regenerating (reg) gene polymorphisms in fibrocalculous pancreatic diabetes. Pancreas 1997;14:122–125. 11. Rossi L, Whitcomb DC, Ehrlich GD, Gorry MC, Parvin S, Sattar S, Ali L, Azad Khan AK, Gyr N. Lack of R117 mutation in the cationic trypsinogen gene in patients with tropical pancreatitis from Bangladesh. Pancreas 1999;17:278 –280. 12. Hassan Z, Mohan V, McDermott MF, Ali L, Ogunkolade WB, Aganna E, Cassell PG, Deepa R, Khan AK, Hitman GA. Pancreatitis in fibrocalculous pancreatic diabetes is not associated with
October 2002
13.
14. 15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
common mutations in the trypsinogen gene. Diabetes Metab Res Rev 2000;16:454 – 457. Bhatia E, Durie P, Zielenski J, Lam D, Lam D, Sikora SS, Choudhuri G, Tsui LC. Mutations in the cystic fibrosis transmembrane regulator gene in patients with tropical calcific pancreatitis. Am J Gastroenterol 2000;95:3658 –3659. Steer ML, Meldolesi J. The cell biology of experimental pancreatitis. N Engl J Med 1987;316:144 –150. Whitcomb DC, Gorry MC, Preston RA, Furey W, Sossenheimer MJ, Ulrich CD, Martin SP, Gates LK Jr, Amann ST, Toskes PP, Liddle R, McGrath K, Uomo G, Post JC, Ehrlich GD. l. Hereditary pancreatitis is caused by a mutation in the cationic trypsinogen gene. Nat Genet 1996;14:141–145. Witt H, Luck W, Becker M. A signal peptide cleavage site mutation in the cationic trypsinogen gene is strongly associated with chronic pancreatitis. Gastroenterology 1999;117:7–10. Creighton J, Lyall R, Wilson DI, Curtis A, Charnley R. Mutations in the cationic trypsinogen gene in patients with chronic pancreatitis. Lancet 1999;354;42– 43. Chen JM, Montier T, Ferec C. Molecular pathology and evolutionary and physiological implications of pancreatitis-associated cationic trypsinogen mutations. Hum Genet 2001;109:245–252 Witt H, Luck W, Hennies HC, Claen M, Kage A, Lass U, Landt O, Becker M. Mutations in the gene encoding the serine protease inhibitor, Kazal type 1 are associated with chronic pancreatitis. Nat Genet 2000;25:213–216. Pfu¨tzer RH, Barmada MM, Brunskill APJ, Finch R, Hart PS, Neoptolemos J, Furey WF, Whitcomb DC. SPINK1/PST1 polymorphisms act as disease modifiers in familial and idiopathic chronic pancreatitis. Gastroenterology 2000;119:615– 623. Chen J-M, Mercier B, Audrezet M-P, Raguenes O, Quere I, Ferec C. Mutations of the pancreatic secretory trypsin inhibitor (PST1) gene idiopathic chronic pancreatitis. Gastroenterology 2001; 120:1061–1062. Plendl H, Siebert R, Steinemann D, Grote W. High frequency of the N34S mutation in the SPINK1 gene in chronic pancreatitis detected by a new PCR-RFLP assay. Am J Med Genet 2001;100: 252–253. Ockenga J, Do¨rk T, Stuhrmann M. Low prevalence of SPINK1 gene mutations in adult patients with chronic idiopathic pancreatitis. J Med Genet 2001;38:243–244. Noone PG, Zhou Z, Silverman LM, Jowell PS, Knowles MR, Cohn JA. Cystic fibrosis gene mutations and pancreatitis risk: relation of epithelial ion transport and trypsin inhibitor gene mutations. Gastroenterology 2001;121:1310 –1319. Rossi L, Pfu¨tzer RH, Parvin S, Ali L, Sattar S, Azad Kahn AK, Gyr
TCP AND SPINK1 TRYPSIN INHIBITOR MUTATIONS
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
1025
N, Whitcomb DC. SPINK1/PST1 mutations are associated with tropical pancreatitis in Bangladesh. Pancreatology 2001;1:242– 245. American Diabetes Association. Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 1997;20:1183–1197. Azad Khan AK, Ali L. Tropical calcific pancreatitis and fibrocalculous pancreatic diabetes in Bangladesh. J Gastroenterol Hepatol 1997;12(suppl):S48 –S52. Balaraman NM, Latha P. Pancreas in chronic calcific pancreatitis. In: Balakrishnan V, ed. Chronic pancreatitis in India. Trivandrum, India: St. Joseph’s Press, 1987:115–119. Govindrajan M, Mohan V, Deepa R, Ashok S, Pitchumoni CS. Histopathology and immunohistochemistry of pancreatic islets in fibrocalculous pancreatic diabetes. Diabetes Res Clin Pract 2001;51:29 –38. Yajnik CS, Shelgikar KM, Sahasrabudhe RA, Naik SS, Pai VR, Alberti KGMM, Hockaday TDR, Katrak A, Dandona P. The spectrum of pancreatic exocrine and endocrine (beta-cell) function in tropical calcific pancreatitis. Diabetologia 1990;33:417– 421. Mehrotra RN, Bhatia E, Choudhuri G. Beta-cell function and insulin sensitivity in tropical calcific pancreatitis from North India. Metabolism 1997;46:441– 444. Kambo PK, Hitman GA, Mohan V, Ramachandran A, Snehlatha C, Suresh S, Metcalfe K, Ryait BK, Viswanathan M. The genetic predispostion to fibrocalcific pancreatic disease. Diabetologia 1989;32:45–51. Yajnik CS, Sahasrabudhe RA, Naik SS, Katrak A, Shelgikar KM, Kanitkar SV, Narayanan VA, Dandona P. Exocrine pancreatic function (serum immunoreactive trypsin, fecal chymotrypsin, and pancreatic isoamylase) in Indian diabetics. Pancreas 1990;5: 631– 638. Chen J-M, Mercier B, Audrezet M-P, Ferec C. Mutational analysis of the human pancreatic secretory trypsin inhibitor (PST1) gene in hereditary and sporadic chronic pancreatitis. J Med Genet 2000; 37:67– 69. Witt H. Gene mutations in children with chronic pancreatitis. Pancreatology 2001;1:432– 438.
Received October 31, 2001. Accepted May 16, 2002. Address requests for reprints to: Heiko Witt, M.D., Kinderklinik, Charite ´, Humboldt University, D-13353 Berlin, Germany. e-mail: [email protected]; fax: (49) 30-450-566-917. Supported by a grant from the Sonnenfeld Stiftung, Berlin, Germany.
Tropical Calcific Pancreatitis: Strong Association With SPINK1 Trypsin Inhibitor Mutations EESH BHATIA,* GOURDAS CHOUDHURI,‡ SADIQ S. SIKORA,§ OLFERT LANDT,¶ ANDREAS KAGE,㛳 MICHAEL BECKER,# and HEIKO WITT# *Departments of Endocrinology, ‡Gastroenterology, and §Surgical Gastroenterology, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India; 㛳Institute of Laboratory Medicine and Pathobiochemistry, Charite´, ¶TIB MOLBIOL, Berlin, Germany; and #Kinderklinik, Charite ´ , Humboldt-University, Berlin, Germany
Background & Aims: Tropical calcific pancreatitis (TCP) is a chronic pancreatitis unique to developing countries in tropical regions. The cause of TCP is obscure. Whereas environmental factors, such as protein energy malnutrition and ingestion of cassava, have been implicated, a genetic predisposition to the disease also may be important. In the present study we report on mutations in the serine protease inhibitor, Kazal type 1 (SPINK1) gene in north Indian patients with TCP. Methods: We studied 66 unrelated TCP patients (44 men, 49 with diabetes, and 6 with family history of TCP), 25 relatives, and 92 healthy control subjects. Samples were analyzed for SPINK1 variants (ⴚ53C>T, L14P, N34S, P55S, and 272T>C) and cationic trypsinogen (PRSS1) variants (A16V, K23R, N29I, and R122H) by melting curve analysis. Results: Twenty-nine patients (44%) carried the N34S missense mutation, of whom 9 (14%) were homozygotes. In contrast, only 2 (2.2%) control subjects were N34S heterozygotes (prevalence ratio 20.2; 95% confidence interval 5.0 – 81.8; P < 0.0001 vs. TCP). The severity of pancreatitis did not differ between TCP patients with or without N34S, or among those heterozygous or homozygous for N34S. Among TCP patients with or without diabetes, the frequency of N34S carriers (43% vs. 47%) and N34S homozygotes (14% vs. 12%) was similar. Conclusions: TCP is highly associated with the SPINK1 N34S mutation. The high prevalence of N34S in TCP patients with and without diabetes suggests that these 2 subtypes have a similar genetic predisposition. The genetic predisposition to TCP resembles, at least in part, the idiopathic chronic pancreatitis found in industrialized countries.
ropical calcific pancreatitis (TCP) is an idiopathic chronic pancreatitis (ICP) unique to nonindustrialized countries located in the tropics.1– 8 Patients present at a young age with recurrent abdominal pain and nutritional deficiencies. The chronic pancreatitis leads to progressive beta-cell deficiency, and patients often develop insulin-requiring diabetes (fibrocalculous pancreatic diabetes, [FCPD]), before age 30.1– 4,8 In older reports
T
the prognosis was described as dismal, with many patients succumbing to complications related to malnutrition and infection.1,2 The origin of TCP is obscure.1,2 Specific predisposing factors, such as alcohol abuse, biliary tract stones, and metabolic disease, are absent. It has been proposed that environmental factors, such as protein energy malnutrition6 or the consumption of cassava (a source of cyanogenic glycosides),7 may play a pathogenic role. However, TCP is frequently found in regions where cassava is not consumed4,8 and is also observed in patients from higher economic levels, in whom malnutrition is unlikely.1 The familial clustering of TCP suggests that genetic defects may predispose to the disease.9 Environmental factors may possibly be operative on the background of a genetic predisposition to pancreatitis. However, no association has been found among the different candidate genes studied, including the reg1A (pancreatic stone protein) gene10 or the cationic trypsinogen (PRSS1) gene.11,12 In a preliminary study, 2 of 18 north Indian TCP patients had heterozygous mutations of the cystic fibrosis transmembrane regulator (CFTR) gene.13 However, this frequency was similar to that in the healthy white population. Chronic pancreatitis is thought to be the result of inappropriate trypsin activity within the pancreatic parenchyma.14 Gain-of-function mutations of the PRSS1 gene have been found to be associated with chronic pancreatitis.15–18 More recently, mutations in the serine protease inhibitor, Kazal type 1 (SPINK1), an inhibitor of intrapancreatic trypsin activity, have been reported to Abbreviations used in this paper: CFTR, cystic fibrosis transmembrane regulator; FCPD, fibrocalculous pancreatic diabetes; FRET, fluorescence resonance energy transfer; ICP, idiopathic chronic pancreatitis; PRSS1, cationic trypsinogen; SPINK1, serine protease inhibitor, Kazal type 1; TCP, tropical calcific pancreatitis. © 2002 by the American Gastroenterological Association 0016-5085/02/$35.00 doi:10.1053/gast.2002.36028
October 2002
be associated with chronic pancreatitis in Europe and the United States.19 –24 These mutations are thought to result in a decreased inhibitory capacity within the pancreas.19 A substitution of asparagine by serine (N34S) in exon 3 is the most common SPINK1 mutation in patients with ICP.19 –24 Recently, in a preliminary study, Rossi et al.25 reported that SPINK1 mutations are associated with FCPD, but not with TCP without diabetes. In the current study, we report on SPINK1 gene mutations in a large cohort of TCP patients, families, and control subjects from north India. Our findings suggest that SPINK1 mutations are strongly associated with TCP, both with and without diabetes.
Materials and Methods Subjects TCP was diagnosed by a history of recurrent abdominal pain and radiologic evidence of pancreatic intraductal calculi. Patients with a history of alcohol intake were excluded, as were those with other causes of chronic pancreatitis, including metabolic disorders (hypercalcemia, hypertriglyceridemia), biliary duct stones, and anatomic anomalies (pancreas divisum). Fasting plasma glucose was measured in all subjects, and diabetes was diagnosed using the criteria of the American Diabetes Association (1997).26 Based on the fasting plasma glucose level, TCP patients were classified into TCP without diabetes and FCPD. Patients receiving treatment with insulin or oral hypoglycemic drugs were classified as FCPD, irrespective of their current plasma glucose level. We studied 66 unrelated patients (44 men) with TCP who attended the Endocrinology and Gastroenterology services at the Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, between January 2000 and September 2001. All patients belonged to the north Indian state of Uttar Pradesh or to adjoining regions. None consumed cassava or gave a history suggestive of malnutrition in childhood. The age (mean ⫾ standard deviation) of the subjects was 31.5 ⫾ 10.9 years, and age at onset of pain was 19.7 ⫾ 9.9 years. Fifty-eight percent of patients had onset of pain ⬍20 years. Twenty-seven patients (41%) had undergone surgery or an endoscopic procedure for pain relief. The patients’ body mass index (BMI) was 18.1 ⫾ 3.4 kg/m2; a low BMI (⬍18 kg/m2) was found in 53% of patients. The patients had severely diminished exocrine function (fecal chymotrypsin 2.2 ⫾ 2.2 U/g stool (n ⫽ 39); normal ⬎8.4 U/g stool). Forty-nine patients (74%) had FCPD; of these, 40 (82%) required insulin therapy. The clinical characteristics of FCPD and TCP subjects without diabetes are shown in Table 1. A family history of TCP was present in 6 (9%) index patients. We studied 17 relatives (8 additional subjects with TCP and 9 unaffected) from these 6 families. In addition, we studied 8 unaffected relatives from 3 different families, where only the proband had TCP. Ninety-two healthy north Indian subjects (physicians and hospital staff), belonging to the same
TCP AND SPINK1 TRYPSIN INHIBITOR MUTATIONS
1021
Table 1. Clinical Characteristics of Patients With Tropical Calcific Pancreatitis
Age (yr) M:F Age at onset of pain (yr)a Body mass index (kg/m2) Age at onset of diabetes (yr) Fecal chymotrypsin (U/g) (normal ⬎8.4 U/g) Surgery for pain relief Fasting plasma glucose (mg %) Plasma glucose 2 hours postprandial (mg %)
TCP with diabetes (FCPD) (n ⫽ 49)
TCP without diabetes (n ⫽ 17)
32.6 ⫾ 10.1 33:16 19.3 ⫾ 9.6 18.0 ⫾ 3.4 28.1 ⫾ 7.8 2.4 ⫾ 2.3 (n ⫽ 29) 12 (24%) 213 ⫾ 78 (70–427) 351 ⫾ 105 (n ⫽ 44)
28.0 ⫾ 13.0 11:6 20.7 ⫾ 13.0 18.5 ⫾ 3.5 – 1.8 ⫾ 1.9 (n ⫽ 10) 15 (88%)b 77 ⫾ 16c (60–109) –
NOTE. Mean ⫾ standard deviation (range). TCP, tropical calcific pancreatitis; FCPD, fibrocalculous pancreatic diabetes. a The age at onset of pain was accurately given by 60 patients. b P ⬍0.0001; c P ⬍0.0001 Student t test.
region, served as controls. Informed written consent was obtained from all subjects, and the study was approved by the institutional ethics committee.
Methods Genomic DNA was extracted from peripheral blood leukocytes by standard techniques. Coded DNA samples of the patients and the control subjects were analyzed for PRSS1 variants (A16V, K23R, N29I, and R122H), and SPINK1 variants (⫺53C⬎T, L14P, N34S, P55S, and 272T⬎C) by melting curve analysis.
Polymerase Chain Reaction Primers flanking the coding region of the PRSS1 and SPINK1 genes were designed according to the nucleotide sequences published by Rowen et al. (PRSS1; GenBank #U66061) and Witt et al. (SPINK1; GenBank #AF286028). With the exception of the forward primer for exon 3 of SPINK1 (5⬘-TAT GAC CCT GTC TGT GGG AC-3⬘), all primers used have been described previously.16,19 Polymerase chain reaction (PCR) was performed using 0.5 U AmpliTaq Gold polymerase (Perkin Elmer, Braunschweig, Germany), 400 mol/L dNTPs, and 0.1 mol/L of each primer in a total volume of 25 L. Cycle conditions were an initial denaturation for 12 minutes at 95°C, followed by 40 cycles of 15 seconds denaturing at 95°C; 30 seconds of annealing at 54, 56, 58 or 64°C; 40 seconds of primer extension at 72°C; and a final extension for 2 minutes at 72°C in an automated thermal cycler (Biometra, Go¨ttingen, Germany).16,19
Melting Curve Analysis Melting curve analysis for the N34S mutation was performed using a pair of fluorescent resonance energy transfer (FRET) probes and a LightCycler (Roche Diagnostics, Mannheim, Germany), as previously described.19 FRET probes were
1022
BHATIA ET AL.
designed and synthesized by TIB MOLBIOL (Berlin, Germany). For the other variations, the following FRET probes were used: SPINK1: ⫺53C⬎T, 5⬘-TTC TGC GTC CAG AGG TCA GTT GAA AAC T-FL and 5⬘-LC Red 640-ACC GCA CTT ACC ACA TCT CTT CAG; L14P, 5⬘-GAA GGT AAC AGG CAT CTT TCT TCT CAG TG-FL and 5⬘-LC Red 705-CTT GGC CCC GTT GAG TC; P55S, 5⬘-CAC GCA TTC ATT GGA ATA AGT AT-FL and 5⬘-LC Red 705-CCA TCA GTC CCA CAG ACA GG; 272C⬎T, 5⬘-CCT CGC GGT GAC CTG ATG GGA-FL and 5⬘-LC Red 640-TTC AAA ACC TTG GTT CTC AGC AAG GCC. PRSS1: A16V, 5⬘-GAC TTG CCT TCT CCC TTC CCA TCT CCA-FL and 5⬘-LC Red 640-CCA GTT GCT GCC CCC TTT GAT; K23R and N29I, 5⬘-AGG GAC ACC TGG TAG GGG ACA GAA TTC TCC-FL and 5⬘-LC Red 705-TGT AGC CCC CAA CGA TCC TGT CAT CAT (K23R) and 5⬘-LC Red 640-CAC AGT TGT AGC CCC CAA (N29I); R122H, 5⬘-CAA CGC CCA CGT GTC CAC CA-FL and 5⬘-LC Red 705- TCT CTG CCC ACC GCC CCT CCA GCC.
DNA Sequencing A novel variation detected by melting curve analysis [272C⬎T; 274C⬎T] was analyzed by direct DNA sequencing. We purified PCR products with spin columns and performed cycle sequencing using 1.5 L purified template, 3 L BigDye terminator mix (Applied Biosystems), and 3 L forward or reverse primer (0.3 mol/L). The reaction products were purified with spin columns and loaded onto an ABI 373A fluorescence sequencer (Applied Biosystems). DNA sequences were analyzed by sequencing both strands.
Statistical Analysis Continuous data were compared using the Student t test, and categorical data were compared using 2 test or the Fisher exact test. A 2-tailed P value ⬍0.05 was considered significant.
Results SPINK1 Mutations The frequency of SPINK1 mutations in the cohort of 66 unrelated index patients (FCPD, 49 patients; TCP without diabetes, 17 patients) and in 92 control subjects is shown in Table 2. Twenty-nine patients (44%) carried the N34S mutation, in contrast to only 2 (2.2 %) control subjects (prevalence ratio, 20.2; 95% confidence intervals, 5.0 – 81.8; P ⬍ 0.0001). Nine patients (14%) were homozygous for the N34S mutation, whereas both control subjects were heterozygous carriers. Comparing TCP patients with an earlier (⬍20 years) or later (⬎20 years) age at onset of pain, the frequency of N34S carriers (54% vs. 36%) or N34S homozygotes (14% vs. 16%) was not significantly different. The severity of pancreatitis (based
GASTROENTEROLOGY Vol. 123, No. 4
Table 2. SPINK1 Variants in Subjects With Tropical Calcific Pancreatitis and Controls
SPINK1 variants
Controls (n ⫽ 92)
All TCP patients (n ⫽ 66)
Ratio (all TCP patients: controls)
N34S (homozygotes ⫹ heterozygotes) N34S homozygotes N34S heterozygotes
2 (2.2) 0 2 (2.2)
29 (43.9) 9 (13.6) 20 (30.3)
20.2 [5.0–81.8]a Not calculatedb 13.9 [3.4–57.6]a
NOTE. Figures in parentheses are the percentage of patients and controls carrying the mutation. Prevalence ratios and 95% confidence intervals compare the frequency of individuals with different abnormal SPINK1 genotypes in the TCP vs. control groups. No ⫺ 53 C ⬎ T or L14P variations were observed; there were no compound N34S/P55S heterozygotes; 1 patient was compound heterozygous for N34S and [272C ⬎ T; 274C ⬎ T]. The frequency of the 272C ⬎ T variant was similar in control subjects (24 of 92, 26.1%) and TCP patients (28 of 66, 42.4%). a P ⬍ 0.0001, Fisher exact test. b P ⫽ 0.0003, Fisher exact test.
on age at onset of pain, frequency of diabetes, and surgical intervention for pain relief) did not differ between TCP patients with or without the N34S mutation or between subjects heterozygous or homozygous for N34S. Among the 49 tested patients who had FCPD, 21 (43%) carried an N34S mutation. Of these, 7 (14%) were N34S/N34S and 14 (29%) were N34S/WT. Of the 17 patients who had nondiabetic TCP, 8 (47%) carried an N34S mutation; 2 (12%) were N34S/N34S and 6 (35%) were N34S/WT. The frequency of N34S heterozygotes or homozygotes did not differ between FCPD and nondiabetic TCP patients. Of the 9 patients homozygous for N34S, 7 (78%) had diabetes; this frequency of diabetes was similar to that in subjects not carrying the N34S homozygous mutation (42 of 47, 74%). The frequency of the P55S and 272C⬎T variations were similar in patients and control subjects, and no ⫺53C⬎T or L14P mutations were detected. In the assay for the 272T⬎C polymorphism, 3 patients and 2 control subjects showed an unknown melting curve. DNA sequencing revealed a C–T transition at nucleotide 274, together with the frequent 272C⬎T polymorphism on the same allele [272C⬎T; 274C⬎T]. Of the 3 families studied with a single affected TCP patient, 2 of 8 unaffected family members carried the N34S mutation (1 heterozygous and 1 homozygous). In one of these families, the proband (onset of pain at age 17 years, insulin requiring diabetes at 22 years) carried a heterozygous N34S mutation, inherited from the unaffected mother. The proband’s wife, though asymptomatic, was an N34S carrier. A history of consanguinity was denied. The patient’s unaffected son (age 8 years, no
October 2002
TCP AND SPINK1 TRYPSIN INHIBITOR MUTATIONS
1023
Discussion
Figure 1. Families with heterozygous or homozygous N34S carriers. Dark squares or circles denote affected family members. N34S/ N34S, homozygotes; N34S/WT, heterozygotes; WT/WT, wild-type.
abdominal pain, normal pancreatic ultrasonography and plasma glucose level) was homozygous for N34S. PRSS1 Mutations Mutations in PRSS1 were absent in all TCP patients, including those with a family history of pancreatitis, and in all control individuals. SPINK1 Mutations in Familial TCP Of the 6 families in which at least 2 family members had TCP, the N34S SPINK1 mutation was detected in 4. Of the 10 individuals with TCP in these 4 families, 7 carried the N34S mutation. In 1 of the families, both of the affected siblings with FCPD were homozygous for N34S and inherited the mutation from asymptomatic heterozygous parents (Figure 1A). In the 3 other families, the TCP patients were heterozygous for N34S. In 1 of these families, the index patient (nondiabetic TCP) and the mother and an aunt (both FCPD) carried the heterozygous N34S mutation (Figure 1B). In another family, the index FCPD patient carried a heterozygous N34S mutation, but no DNA was available for genetic testing from the affected sister. In the fourth family, the index patient (FCPD) was heterozygous for N34S, but 3 other affected family members (also FCPD) carried the wild-type allele. In the fifth family, the proband and his affected nephew were heterozygous for the P55S variation, and the father and brother of the proband were heterozygous carriers and were asymptomatic. Finally, in the last family, neither the proband nor her affected father had any SPINK1 mutation.
Although TCP has certain clinical features similar to idiopathic chronic pancreatitis, distinctive features include an early age of onset, large intraductal calculi, and a progressive course with high frequency of diabetes.1 Our study provides evidence that TCP is strongly associated with the N34S mutation in the SPINK1 gene. Approximately 50% of the TCP patients were carriers for N34S, whereas 14% were homozygous for this mutation. Thus, in terms of this genetic predisposition, TCP is similar to idiopathic chronic pancreatitis found in industrialized countries.19 –24 The association of the same SPINK1 mutation with chronic pancreatitis in 2 different racial groups strengthens the hypothesis that this mutation, or another mutation closely linked to it, predisposes for disease. Some authors have tried to distinguish between FCPD and TCP on the basis of clinical findings.27 Nonetheless, TCP without diabetes and FCPD are indistinguishable on the basis of their pathologic findings.1,4,28,29 Within the clinical spectrum of TCP, the features of pain, diabetes, and exocrine deficiency occur in varying combinations, and onset of diabetes is usually associated with an advanced stage of TCP.4 In TCP, islet beta-cell function correlates with exocrine function,30 diminishing with increasing duration of disease.31 Thus FCPD is likely to result from a progressive impairment of beta-cell function in patients with tropical pancreatitis. Recently, Rossi et al.25 reported that TCP and FCPD have a separate genetic background. This was based on their findings that 4 of 6 subjects with FCPD, but 0 of 3 TCP patients without diabetes, carried a SPINK1 mutation. However, in our study, we found a similar high prevalence of SPINK1 mutations in both of these subgroups. This suggests that the genetic predisposition to pancreatitis is similar in FCPD and TCP without diabetes. Whether, as proposed by Kambo et al.,32 other genetic loci (e.g., HLA class II genes, insulin gene) confer susceptibility to the development of diabetes requires further study. In the families studied, TCP was associated with different patterns of N34S inheritance. In 1 family, 2 affected children were homozygous for N34S, whereas the unaffected parents were heterozygous. This finding could indicate a recessive inheritance pattern. However, in another family, the 3 affected members were heterozygous for N34S. In another family, only the index patient was N34S heterozygous, whereas 3 other affected family members carried a wild-type allele. Finally, in a simplex family, the proband was N34S/WT; his N34S homozygous offspring was found to have no evidence of TCP. In
1024
BHATIA ET AL.
summary, our data suggest that despite the striking association between N34S and TCP, this genetic alteration acts in a manner that is presently unclear. The prevalence of TCP in India is likely to be higher than that of ICP in industrialized countries.1,4 While the prevalence of TCP in the general population in north India has not been studied, a field survey from the Quilon district in Kerala, south India, reported a prevalence of 0.12%.5 Because the aforementioned survey was conducted in a district “endemic” for the disease, the prevalence of TCP in other regions of India may be lower than this figure. Clinic based estimates of FCPD are variable, but range from 5% to 30% of the patients with diabetes below age 30 years.2,4 In addition, it has also been reported that a subclinical “pancreatopathy” (as suggested by reports of elevated serum trypsin in asymptomatic subjects) is frequent in healthy Indian subjects.33 This high rate of clinical and sub-clinical pancreatic disease may be related in part to the N34S mutation, but other (hitherto unknown) genetic factors or unique environmental factors are also likely to play a role. We found a prevalence of N34S mutations in healthy subjects of 2.2%, though this estimate is unstable since our control population was small (n ⫽ 92), and data from other studies are not available for the north Indian population. Nevertheless, this prevalence of N34S carriers is far higher than that of TCP. In addition, there was no specific pattern for inheritance of the N34S mutation in families. Finally, N34S homozygotes did not reveal greater severity in their clinical features, when compared with heterozygotes. These data are analogous to that found in studies on ICP.19 –24 They would suggest that while a heterozygous N34S mutation is strongly associated with TCP, it may not be adequate to cause the disease without other genetic and/or environmental factors. In a study by Noone et al., CFTR mutations and the N34S mutation in the SPINK1 gene independently predispose to ICP.24 In a previous report on a small cohort of TCP patients, the frequency of CFTR mutations (0.08) was lower than that reported in ICP in Caucasian subjects (0.20-0.24).13 Of the 18 TCP subjects previously studied for CFTR mutations, 16 were analyzed in the current study. DNA from 1 of the 2 patients with CFTR mutations was not available. One patient, homozygous for the 5T allele of the CFTR gene, was heterozygous for N34S. Thus, while both SPINK1 and CFTR genetic loci predispose to TCP and ICP, they are likely to contribute in different proportions to the susceptibility. The P55S variation is not believed to be diseaseassociated with chronic pancreatitis.19,20,34 Similarly, in
GASTROENTEROLOGY Vol. 123, No. 4
the current study, the frequency of P55S did not differ in patients and control subjects. The [272C⬎T; 274C⬎T] alteration is also unlikely to be disease-predisposing, since the frequency of this variation was similar in TCP and control subjects. Mutations in the cationic trypsinogen (PRSS1) gene were not detected in the current study. This confirms earlier reports of the absence of PRSS1 mutations in TCP subjects.11,12 The absence of trypsinogen mutations is in contrast to ICP, where these are found in approximately 6% of children and adolescents.35 Recently, Noone et al. have also reported a PRSS1 mutation in four ICP patients without a family history of pancreatitis.24 In conclusion, there is a strong association between the N34S mutation of the SPINK1 gene and TCP, both with and without diabetes. The high prevalence of early-onset chronic pancreatitis in tropical countries may be the consequence of alterations in the SPINK1 gene, mutations in other genes, and/or environmental factors (e.g., protein energy malnutrition) which are unique to developing countries. The genetic predisposition to TCP is similar, in part, to idiopathic chronic pancreatitis found in industrialized countries.
References 1. Mohan V, Pitchumoni CS. Tropical calcific pancreatitis. In: Berger HG, Warshaw AL, Bu¨chler MW, eds. The pancreas. 1st ed. London: Blackwell Science, 1998:688 – 697. 2. Abu-Bakare A, Taylor R, Gill GV, Alberti KGMM. Tropical or malnutrition-related diabetes: a real syndrome? Lancet 1986;1:1135– 1138. 3. Viswanathan M. Pancreatic diabetes in India: an overview. In: Podolsky S, Viswanathan M, eds. Secondary diabetes: the spectrum of diabetic syndromes. New York: Raven, 1980:105–116. 4. Mohan V, Nagolimath SJ, Yajnik CS, Tripathy BB. Fibrocalculous pancreatic diabetes. Diabetes Metab Rev 1998;14:153–170. 5. Balaji LN, Tandon BN, Tandon RK, Banks PA. Prevalence and clinical features of chronic pancreatitis in Southern India. Int J Pancreatol 1994;15:29 –34. 6. Shaper AG. Chronic pancreatic disease and protein malnutrition. Lancet 1960;1:1223–1224. 7. McMillan DE, Geevarghese PJ. Dietary cyanide and tropical malnutrition diabetes. Diabetes Care 1979;2:202–208. 8. Bhatia E, Baijal SS, Kumar R, Choudhuri G. Exocrine pancreatic and -cell function in malnutrition related diabetes among North Indians. Diabetes Care 1995;18:1174 –1178. 9. Mohan V, Chari ST, Hitman GA, Suresh S, Madanagopalan N, Ramachandran A, Viswanathan M. Familial aggregation in tropical fibrocalculous pancreatic diabetes. Pancreas 1989;4:690 – 693. 10. Hawrami K, Mohan V, Bone A, Hitman GA. Analysis of islet regenerating (reg) gene polymorphisms in fibrocalculous pancreatic diabetes. Pancreas 1997;14:122–125. 11. Rossi L, Whitcomb DC, Ehrlich GD, Gorry MC, Parvin S, Sattar S, Ali L, Azad Khan AK, Gyr N. Lack of R117 mutation in the cationic trypsinogen gene in patients with tropical pancreatitis from Bangladesh. Pancreas 1999;17:278 –280. 12. Hassan Z, Mohan V, McDermott MF, Ali L, Ogunkolade WB, Aganna E, Cassell PG, Deepa R, Khan AK, Hitman GA. Pancreatitis in fibrocalculous pancreatic diabetes is not associated with
October 2002
13.
14. 15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
common mutations in the trypsinogen gene. Diabetes Metab Res Rev 2000;16:454 – 457. Bhatia E, Durie P, Zielenski J, Lam D, Lam D, Sikora SS, Choudhuri G, Tsui LC. Mutations in the cystic fibrosis transmembrane regulator gene in patients with tropical calcific pancreatitis. Am J Gastroenterol 2000;95:3658 –3659. Steer ML, Meldolesi J. The cell biology of experimental pancreatitis. N Engl J Med 1987;316:144 –150. Whitcomb DC, Gorry MC, Preston RA, Furey W, Sossenheimer MJ, Ulrich CD, Martin SP, Gates LK Jr, Amann ST, Toskes PP, Liddle R, McGrath K, Uomo G, Post JC, Ehrlich GD. l. Hereditary pancreatitis is caused by a mutation in the cationic trypsinogen gene. Nat Genet 1996;14:141–145. Witt H, Luck W, Becker M. A signal peptide cleavage site mutation in the cationic trypsinogen gene is strongly associated with chronic pancreatitis. Gastroenterology 1999;117:7–10. Creighton J, Lyall R, Wilson DI, Curtis A, Charnley R. Mutations in the cationic trypsinogen gene in patients with chronic pancreatitis. Lancet 1999;354;42– 43. Chen JM, Montier T, Ferec C. Molecular pathology and evolutionary and physiological implications of pancreatitis-associated cationic trypsinogen mutations. Hum Genet 2001;109:245–252 Witt H, Luck W, Hennies HC, Claen M, Kage A, Lass U, Landt O, Becker M. Mutations in the gene encoding the serine protease inhibitor, Kazal type 1 are associated with chronic pancreatitis. Nat Genet 2000;25:213–216. Pfu¨tzer RH, Barmada MM, Brunskill APJ, Finch R, Hart PS, Neoptolemos J, Furey WF, Whitcomb DC. SPINK1/PST1 polymorphisms act as disease modifiers in familial and idiopathic chronic pancreatitis. Gastroenterology 2000;119:615– 623. Chen J-M, Mercier B, Audrezet M-P, Raguenes O, Quere I, Ferec C. Mutations of the pancreatic secretory trypsin inhibitor (PST1) gene idiopathic chronic pancreatitis. Gastroenterology 2001; 120:1061–1062. Plendl H, Siebert R, Steinemann D, Grote W. High frequency of the N34S mutation in the SPINK1 gene in chronic pancreatitis detected by a new PCR-RFLP assay. Am J Med Genet 2001;100: 252–253. Ockenga J, Do¨rk T, Stuhrmann M. Low prevalence of SPINK1 gene mutations in adult patients with chronic idiopathic pancreatitis. J Med Genet 2001;38:243–244. Noone PG, Zhou Z, Silverman LM, Jowell PS, Knowles MR, Cohn JA. Cystic fibrosis gene mutations and pancreatitis risk: relation of epithelial ion transport and trypsin inhibitor gene mutations. Gastroenterology 2001;121:1310 –1319. Rossi L, Pfu¨tzer RH, Parvin S, Ali L, Sattar S, Azad Kahn AK, Gyr
TCP AND SPINK1 TRYPSIN INHIBITOR MUTATIONS
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
1025
N, Whitcomb DC. SPINK1/PST1 mutations are associated with tropical pancreatitis in Bangladesh. Pancreatology 2001;1:242– 245. American Diabetes Association. Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 1997;20:1183–1197. Azad Khan AK, Ali L. Tropical calcific pancreatitis and fibrocalculous pancreatic diabetes in Bangladesh. J Gastroenterol Hepatol 1997;12(suppl):S48 –S52. Balaraman NM, Latha P. Pancreas in chronic calcific pancreatitis. In: Balakrishnan V, ed. Chronic pancreatitis in India. Trivandrum, India: St. Joseph’s Press, 1987:115–119. Govindrajan M, Mohan V, Deepa R, Ashok S, Pitchumoni CS. Histopathology and immunohistochemistry of pancreatic islets in fibrocalculous pancreatic diabetes. Diabetes Res Clin Pract 2001;51:29 –38. Yajnik CS, Shelgikar KM, Sahasrabudhe RA, Naik SS, Pai VR, Alberti KGMM, Hockaday TDR, Katrak A, Dandona P. The spectrum of pancreatic exocrine and endocrine (beta-cell) function in tropical calcific pancreatitis. Diabetologia 1990;33:417– 421. Mehrotra RN, Bhatia E, Choudhuri G. Beta-cell function and insulin sensitivity in tropical calcific pancreatitis from North India. Metabolism 1997;46:441– 444. Kambo PK, Hitman GA, Mohan V, Ramachandran A, Snehlatha C, Suresh S, Metcalfe K, Ryait BK, Viswanathan M. The genetic predispostion to fibrocalcific pancreatic disease. Diabetologia 1989;32:45–51. Yajnik CS, Sahasrabudhe RA, Naik SS, Katrak A, Shelgikar KM, Kanitkar SV, Narayanan VA, Dandona P. Exocrine pancreatic function (serum immunoreactive trypsin, fecal chymotrypsin, and pancreatic isoamylase) in Indian diabetics. Pancreas 1990;5: 631– 638. Chen J-M, Mercier B, Audrezet M-P, Ferec C. Mutational analysis of the human pancreatic secretory trypsin inhibitor (PST1) gene in hereditary and sporadic chronic pancreatitis. J Med Genet 2000; 37:67– 69. Witt H. Gene mutations in children with chronic pancreatitis. Pancreatology 2001;1:432– 438.
Received October 31, 2001. Accepted May 16, 2002. Address requests for reprints to: Heiko Witt, M.D., Kinderklinik, Charite ´, Humboldt University, D-13353 Berlin, Germany. e-mail: [email protected]; fax: (49) 30-450-566-917. Supported by a grant from the Sonnenfeld Stiftung, Berlin, Germany.