Motion – Genetic testing is useful in the diagnosis of nonhereditary pancreatic conditions : Arguments for the motion

This article was originally presented at a symposium entitled, “Controversies in Gastroenterology”, sponsored by Axcam Pharma, Toronto, Ontario, April 8 to 10, 2002 Cell Biology & Physiology and Human Genetics; Division of Gastroenterology, Hepatology and Nutrition; and Center for Genomic Sciences, University of Pittsburgh, Pittsburgh, Pennsylvania, USA Correspondence: Dr David C Whitcomb, Division of Gastroenterology, Hepatology and Nutrition, UPMC Presbyterian, Mezzine level, C-wing, 200 Lothrop Street, Pittsburgh, Pennsylvania 15213, USA. Telephone 412-648-9604, fax 412-383-7236, e-mail whitcomb@pitt.edu DC Whitcomb. Motion – Genetic testing is useful in the diagnosis of nonhereditary pancreatic conditions: Arguments for the motion. Can J Gastroenterol 2003;17(1):47-52.

S ome people see genetic testing as a Pandora's Box: "Don't open it!"Others see genetic testing as an incredibly powerful diagnostic tool that will accurately reflect the patient's past, present and future, and that of his or her family.
The issues surrounding genetic testing in gastrointestinal and liver diseases emerged in 1996 with the discovery of the common genetic defects for hereditary pancreatitis (1).By 1998, patients with idiopathic chronic pancreatitis were being screened for cystic fibrosis transmembrane conductance regulator (CFTR) gene mutations (2,3).Finally, an association between idiopathic pancreatitis and mutations in the serine protease inhibitor, Kazal type 1 (SPINK1) gene was recognized in 2001 (4,5).Other pancreatitis mutations will soon be discovered.Genetic testing raises concerns among patients, gas-

CONTROVERSIES IN GASTROENTEROLOGY
troenterologists, genetic counsellors, insurance companies, and lawyers.These concerns are real and will be addressed.In the future, however, genetic testing should be offered to all patients with suspected pancreatitis, especially early in the course of the disease.This may prevent delays in diagnosis and improve treatment options.
The focus of this article will be on the three primary genes that are known to be associated with pancreatitis, and the challenges and opportunities for the future.

THREE MAJOR PANCREATITIS-ASSOCIATED GENES
The primary genes that have thus far been shown to be associated with pancreatitis are the cationic trypsinogen gene (PRSS1), the CFTR gene, the pancreatic secretory trypsin inhibitor (PSTI), or (SPINK1), gene.
Each of these primary pancreatitis-associated genes teaches us something different about genetics and issues surrounding genetic testing.

PRSS1 MUTATIONS
Cationic trypsinogen is among the most abundant molecules produced by pancreatic acinar cells (6).It plays a central role in hydrolyzing dietary proteins at lysine and arginine amino acid residues, and also in activating all other digestive proenzymes (6).Premature activation of trypsinogen within the pancreas leads to pancreatic autodigestion and is believed to be a key step in the pathogenesis of acute pancreatitis.Recurrent attacks of acute pancreatitis, as occur in patients with hereditary pancreatitis, eventually lead to chronic pancreatitis.

Clinical syndrome
Approximately 80% of persons with the PRSS1 N29I or R122H mutations have recurrent episodes of acute pancreatitis during childhood.The median age of onset is roughly 10 years; chronic pancreatitis develops during the subsequent decade of the disease in about half of patients with acute pancreatitis, and pancreatic cancer occurs by age 70 years in 40% of individuals with chronic pancreatitis.Smoking more than doubles the risk of developing pancreatic cancer and reduces the median age of onset by 20 years (20).

Genetic testing
Genetic testing for hereditary pancreatitis has been available for research and clinical purposes for the past five years.The major issues surrounding genetic testing were recently reviewed (21)(22)(23).In general, the indications for clinical genetic testing vary widely according to the severity of the disease, the age of onset, the availability of surrogate markers and the possibility of an effective intervention.Patients are also likely to have their own reasons for pursuing genetic testing.Reasons for PRSS1 mutation testing generally include verification of a clinical suspicion, helping patients to understand or validate their condition, and assisting individuals at risk of pancreatitis (and ultimately pancreatic cancer [24]) in making lifestyle decisions (involving diet and smoking, for example) that minimize the risk of disease (21).Indeed, identification of an established pancreatitis-associated gene mutation can be valuable in expediting the evaluation of recurrent pancreatitis in children and preventing the unnecessary pursuit of elusive causes of pancreatitis in adults (eg, sphincter of Oddi dysfunction or occult alcohol abuse).A consensus statement on genetic testing for suspected hereditary pancreatitis was developed at the Third International Symposium on Inherited Diseases of the Pancreas (in Milan, Italy, in April 2001), and has been published in a special issue of the journal Pancreatology (25).Such investigations are meant for patients with a clinical and family history suggestive of hereditary pancreatitis.

CFTR mutations causing cystic fibrosis
The primary CFTR mutation-related disorder is cystic fibrosis (CF), which is one of the most common life-threatening autosomal recessive disorders in Caucasians (26).Major mutations in both alleles of the gene result in the commonly recognized clinical features of CF: abnormal sweat chloride concentrations, neonatal hypertrypsinogenemia, pancreatic pseudocysts, fibrosis (hence the term 'cystic fibrosis'), chronic pancreatitis, and progressive pulmonary disease.Among CF patients, 66% have a three base pair deletion of the phenylalanine-coding codon 508 (F508), although approximately 1000 other mutations have been reported (27,28).African-Americans may have their own set of 'common' CF mutations, including the 3120+1G>A mutation, which occurs at a frequency of 12.3% in a representative population (29).Most CFTR mutations can be placed into one of five severity categories based on the observed or presumed molecular consequences (30,31).Typical CF patients with pancreatic insufficiency tend to have two severe mutations (ie, Class I, II, or III), whereas CF patients with pancreatic sufficiency from birth have at least one CF 'mild allele' (ie, Class IV or V) (31).

CFTR mutations associated with idiopathic chronic pancreatitis
In 1998, two groups reported a significant association between idiopathic chronic pancreatitis and various CFTR mutations (2,3).Indeed, several mild, 'pancreas sufficient' mutations (eg, CFTR R117H and the intron 8 '5T allele' [IV8 5T], which results in an 80% reduction of exon 9 expression [32,33]) seem to be associated with idiopathic chronic pancreatitis (2,3) and other features of CF, including congenital bilateral absence of the vas deferens (32,34).Other mild CFTR mutations (eg, L997F [35]) may also be associated with neonatal hypertrypsinemia and/or idiopathic pancreatitis but not with lung disease or abnormal sweat chloride.Although initial reports suggested that idiopathic chronic pancreatitis was associated with a single allelic mutation of CFTR, more recent evidence suggests that patients with chronic pancreatitis may actually have compound heterozygous mutations of CFTR and mild CF.This argument is based on at least three observations: • Healthy parents of children with CF (obligate CFTR mutation carriers without CF) do not have an increased incidence of acute or chronic pancreatitis compared with the general population (36).
• The prevalence of functional classes IV and V CFTR mutations (eg, R117H, IV8 5T) among patients with idiopathic chronic pancreatitis without a coexisting severe mutation is only about 1.5-fold higher than the expected number (which is not statistically significant) (37).
• Severe and mild CFTR compound heterozygous mutations are strongly associated with idiopathic chronic pancreatitis, and the majority of these patients have abnormal nasal bioelectric responses, thus demonstrating CFTR dysfunction (37).
Therefore, a subset of patients with chronic pancreatitis appear to have atypical CF.Indeed, genetic screening of infants who were thought to be heterozygous F508 carriers (on the basis of a normal sweat chloride but elevated immunoreactive trypsinogen, which is associated with pancreatic injury) frequently yielded evidence of R117H mutation (9%) or IV8 5T allele (20%) (38).These data led to the following conclusions (2,3,37): • Individuals with a single mutant CFTR allele are asymptomatic carriers.
• Persons with two severely affected CFTR alleles have classic CF.
• Persons with compound heterozygous CFTR genotypes with a severe (classes I to III) and a mild (class IV or V) mutation may be at risk for pancreatitis, congenital bilateral absence of the vas deferens (39,40), sinusitis, or other CF-associated disorders.

Genetic testing
CFTR is a large molecule with 1480 amino acids coded for by over 4400 nucleotides in 24 exons (26).Idiopathic chronic pancreatitis appears to be associated with some loss of CFTR function, which can be caused by many combinations of CFTR mutations.However, screening of the entire CFTR gene for mutations is difficult and very expensive, thereby limiting this approach to specialized research laboratories.Some commercial laboratories do offer clinical testing for a panel of mutations that are commonly associated with CF, but these panels may not include many of the 'mild' CFTR mutations that are associated with pancreatitis (37).Furthermore, since CFTR mutations are prevalent in the population, the identification of a polymorphism does not necessarily mean that it is the cause of pancreatitis.Nor does the presence of a CFTR polymorphism confer a high risk of pancreatitis in asymptomatic individuals.For example, the incidence of acute pancreatitis in the general population is 1 in 16,000 (41), and the presence of the CFTR R117H genotype increases the risk 2.6-fold (37); thus, the overall risk of pancreatitis with this mutation is only 2.6 in 16,000, or 0.16%.A major dilemma regarding genetic testing of patients with recurrent acute or chronic pancreatitis has already arisen.Cohn et al (37) suggest that, at the present time, CFTR testing may be considered for individuals in whom pancreatitis appears to be the earliest manifestation of classic CF, or for young patients with pancreatitis and borderline sweat chloride values, so that individuals can be referred to CF centres or for genetic counselling.The situation is becoming more complicated, however, because the World Health Organization (WHO) is about to publish the Classification of Cystic Fibrosis and Related Disorders in the upcoming International Classification of Diseases (ICD-11).Dodge reported on the deliberations of several major European working groups during the Third International Symposium on Inherited Diseases of the Pancreas in Milan, Italy, on March 4 to 6, 2001; the conference proceedings have also been published (42).According to these guidelines, all patients with 'chronic pancreatitis' require testing for CFTR mutations, and the presence of even a single CFTR mutation identifies the patient as having a CF-associated disorder.This formulation raises a number of serious questions for which there are no answers, including: Which mutations must be considered?How can a single common CFTR mutation be diagnostic of CF-associated chronic pancreatitis?Is testing for CFTR mutations always warranted in patients with chronic pancreatitis?
The author believes that CFTR mutation analysis may play an important role in the future, but most of the work has been on severe CF-causing mutations in Caucasians, and large national studies have not been conducted for other ethnic populations.

SPINK1/PSTI GENE MUTATIONS
SPINK1 function PSTI, which is also known by its UniGene name, SPINK1, is a 56-amino acid peptide that specifically inhibits trypsin by physically blocking the active site.SPINK1 is synthesized by pancreatic acinar cells together with trypsinogen and is located in the same zymogen granules.In the mechanistic models of pancreatic acinar cell protection, SPINK1 acts as the first line of defence against prematurely activated trypsinogen (1,(4)(5)(6)43).Because only a limited amount of SPINK1 is produced, however, it is capable of inhibiting only approximately 20% of the trypsin that could potentially be formed.

SPINK1 mutations
Because gain-of-function mutations of the trypsin molecule are associated with acute and chronic pancreatitis, it was hypothesized that loss of trypsin inhibitor capacity might have similar effects.By 2000, it was appreciated that SPINK1 mutations indeed predispose to chronic pancreatitis (4,5,44).This extremely important discovery was made by screening a large, well-defined population for mutations in candidate genes.SPINK1 N34S and P55S mutations are relatively prevalent, being present in roughly 2% of the general population (5,44).In addition, a number of intron mutations, which have not been fully characterized, occur in patients with pancreatitis (unpublished observations).SPINK1 mutations occur in families of pancreatitis patients who do not have trypsinogen mutations, but the mutations do not consistently segregate with the disease (5,44).Thus, SPINK1 mutations do not cause hereditary pancreatitis in an autosomal dominant pattern.Proof of the role of SPINK1 mutations, however, is the finding that 23% to 25% of patients with idiopathic chronic pancreatitis harbour them (4,5).

Gene-gene interactions
We have argued that SPINK1 mutations act as disease promoters in that they lower the threshold for initiating pancreatitis or possibly worsen the severity of pancreatitis that is caused by other genetic or environmental factors (5).It has also been shown that the presence of PRSS1 and SPINK1 mutations in the same person influences the phenotype (45).The author's group and the Midwest Multicenter Pancreatic Study Group have collected a large database from kindreds of patients with hereditary pancreatitis; it currently includes 717 individuals, of whom 368 (51%) have confirmed pancreatitis.Unaffected individuals include relatives and spouses, who are essential for conducting genetic linkage studies, as well as 46 relatives who were found to be silent carriers for a PRSS1 mutation.From this data set, 39 persons were identified with PRSS1 mutations but without SPINK1 mutations, and 22 with both PRSS1 and SPINK1 mutations.The median age of disease onset was older for patients with mutations of PRSS1 alone than for patients with mutations in both genes.The possible interaction between SPINK1 mutations and CFTR mutations may also be important.Because the SPINK1 mutation is not sufficient to cause disease in most individuals (as already mentioned), the significantly earlier age of onset seen in persons with both PRSS1 and SPINK1 mutations supports the concept that SPINK1 acts as a disease modifier.

Genetic testing for SPINK1 mutations
The discovery of this disease-associated mutation raises the question of when to undertake genetic testing.Identification of SPINK1 mutations in patients with early chronic pancreatitis may provide important information about the cause of the disorder.On the other hand, because fewer than 1% of persons with heterozygous SPINK1 mutation alone are at risk of developing pancreatitis, it is not worthwhile to test asymptomatic subjects.SPINK1 mutations are associated with early onset of symptoms; therefore, testing is unlikely to be fruitful in patients who develop pancreatitis after the age of 20 years.

Fears about genetic testing
The decision about whether to proceed with genetic testing is often complex.Currently, the utility of this technique is limited by our incomplete understanding of the etiology of pancreatic disease and our inability to use genetic information to alter the clinical outcomes.Even though these factors will change in the future, the patients' fears of genetic testing result in significant hesitation.
The author's group recently studied the motivations and concerns of patients who were offered genetic testing for hereditary pancreatitis (22).By far, the major concern was insurance discrimination; this is the case even with hemochromatosis, a disease in which genetic testing can identify patients who could be effectively treated with phlebotomy.In a recent study, more than 25% of 124 asymptomatic subjects who were identified as hereditary hemochromatosis carriers reported that they were either required to pay higher premiums or denied coverage altogether (46).
Unfortunately, as of 2001, fewer than half of the states in the United States (Arizona, California, Colorado, Connecticut, Florida, Georgia, Hawaii, Illinois, Maryland, Minnesota, Nevada, New Hampshire, New Jersey, New York, North Carolina, Ohio, Oregon, Tennessee, Texas, Virginia and Wisconsin) have passed legislation prohibiting genetic discrimination in health insurance (46).Participation in a research study, rather than attendance at a clinical laboratory, is an attractive option for individuals who do not want their test results to be included in their medical records.In this setting, the results are disclosed directly to the patient, who can then decide to whom the results will be disclosed.

Accuracy of testing
Modern and properly applied genetic testing techniques yield positive and negative predictive values of almost 100% for identifying specific mutations.The pretest probability of a test for a PRSS1 mutation depends on several factors.The strongest predictors of a genetic basis for a case of pancreatitis include a typical family history (especially an autosomal dominant inheritance pattern) and an early age of onset of symptoms.
A delayed onset of pancreatitis (ie, over age 20 years) does not preclude a genetic etiology.Although 93% of patients in the Midwest Multicenter Pancreatic Study Group-Pittsburgh study developed symptoms by the age of 30 years (45), only 40% of affected persons enrolled in the European Genetic Register of Hereditary Pancreatitis and Familial Pancreatic Cancer study became ill by that age (21).These conflicting results suggest the possibility of selection or recruitment bias in studies that were not designed to explore the age of onset of symptoms.The latter study (21) might have overestimated the age of onset of disease because of a failure to fully elucidate the patients' histories.
Impediments to investigation include the facts that many patients are unaware of their family histories, have few relatives, have unaffected first-degree relatives or have ancestors in whom the diagnosis of pancreatitis was not established (unpublished observations).Moreover, the underlying frequency of mutations varies across populations.For example, the prevalence of PRSS1 mutations ranges from 0% to 19% among various patient populations with presumed idiopathic chronic pancreatitis (18,(47)(48)(49), perhaps because of differences in regional settlement patterns.The likelihood of positive test results is much higher in endemic regions.

Interpretation of test results
The results of genetic testing have implications for both the subject and his or her extended family.Roughly 80% of individuals with either the R122H or the N29I trypsinogen mutation will experience at least one episode of acute pancreatitis (ie, the disease penetrance is 80%) (50)(51)(52)(53).About one-half of clinically affected persons with either mutation will develop symptomatic chronic pancreatitis, and there is an increased risk of pancreatic cancer (23,54).Furthermore, there is a 50% chance of passing on the mutation to each child.
In asymptomatic individuals, a positive test result signifies an increased risk of pancreatitis, which possibly diminishes with age.A negative test result in a patient whose family is known to harbour a pancreatitis-associated gene mutation essentially eliminates the risk of developing that particular genetic form of pancreatitis.On the other hand, if such a mutation has not been previously identified in the family, then a negative test result in an asymptomatic person is considered 'noninformative' because the subject might have inherited a different pancreatitis-causing mutation.
The situation with CFTR and SPINK1 mutations is more complex because pancreatitis-associated mutations occur in more than 1% of North American Caucasians, even though the prevalence of chronic pancreatitis is small (1 of 16,000 persons).Identification of one of the common pancreatitis-associated mutations of these genes in a patient with pancreatitis suggests that the mutation contributed to the etiology of pancreatitis, whereas the same mutation in an unaffected individual confers only a slight risk (approximately 5%) of ever experiencing pancreatitis.

Future developments
The major questions regarding genetic testing in the future will be who to test and when to test them.The technique is appropriate when the following conditions are met: 1.The patient will derive clinical benefit from the test results.
2. The test addresses the clinically relevant mutations.
3. The test is affordable, reliable, and easy to perform.
4. Insurance discrimination and other disincentives are minimal.
5. The patient can give informed consent.
6.The results are properly interpreted and communicated.
It is clear that testing should be undertaken for markers of a very high risk for pancreatitis, such as the PRSS1 R122H and N21I mutations (21)(22)(23).
Greater opportunities for genetic testing will come with better understanding of the genetic and environmental risk factors for pancreatitis.This goal can be accomplished by means of national studies involving thousands of patients and appropriate controls.We must also develop sensitive and specific methods for the early detection of chronic pancreatitis, and find ways to limit disease progression.
The development of new molecular diagnostic methods that allow for the simultaneous screening of hundreds of disease-associated mutations would make genetic testing extremely useful.Once these conditions are met, patients with typical features of pancreatitis should be surveyed for environmental factors and offered genetic screening to determine their risk for pancreatitis.Patients who are shown to be at increased risk would then undergo a sensitive and specific confirmatory test (for example, endoscopic ultrasound with biopsy [23]).Effective therapy could be initiated for patients with the diagnosis of early pancreatitis.In this way, a disease like chronic pancreatitis, which cannot be cured, can be readily prevented.
Indeed, we must risk opening the mystical box; is it Pandora's, or panacea's, or somewhere in between?