I. SUMMARY OF THE PROBLEM
The exocrine pancreas is pivotal to the digestion and absorption of nutrients. Pancreatic disorders manifest either with pain secondary to inflammation or with malabsorption secondary to acinar malfunction, paucity, or dropout. Children suffer from the full range of pancreatic disease seen in adults, and some pancreatic disorders first become apparent in childhood.
The spectrum of pancreatic disease is listed below.
- Congenital anatomic abnormalities
- Formation anomalies: annular pancreas, pancreas divisum, pancreatic hypoplasia and agenesis.
- Situation anomalies: heterotopic pancreas
- Inborn errors
- Cystic fibrosis
- Congenital secretory deficiencies: Shwachman-Diamond syndrome, Johanson-Blizzard syndrome, Pearson bone marrow pancreas syndrome, isolated enzyme deficiencies (lipase, colipase)
- Hereditary pancreatitis
- Pancreatitis; idiopathic, viral, drugs, traumatic, metabolic, collagen vascular diseases; autoimmune; fibrosing, nutritional (tropical)
- Neoplastic disease (rare in childhood)
- Pancreatic insufficiency secondary to other disorders (Alagille syndrome, Celiac disease)
Most of these disorders are either uncommon or rare. The exceptions are cystic fibrosis (CF), which is the commonest severe autosomal recessive disease of humans, and nutritional pancreatitis. The gene frequency for cystic fibrosis in western communities is 1:25 to 1:35, with a live birth incidence of 1:2,500 to 1:5,000. Between 30,000 and 40,000 CF patients live in the United States. Nutritional pancreatitis is endemic in some parts of the tropics, including Kerala in Southern India, and parts of Indonesia and Africa.
MAJOR ADVANCES IN PANCREATIC DISEASE
Before the mid 19th century, knowledge of the pancreas was descriptive. The concept of the pancreas as digestive organ began with the description by Claude Bernard of lipolytic activity in pancreatic juice in 1856. The mechanisms governing pancreatic secretions were advanced by the discovery in intestinal extracts of secretin that drives water and anion secretion by Bayliss and Starling in 1902 and pancreozymin (cholecystokinin) by Harper and Raper in 1943. These hormones were synthesized by Mutt and Jorpes in the 1960s. Many other neuroenteric hormones that act on the pancreas have since been discovered. The 1980s saw the advent of the concept of ion channels and the realization that the CF gene was most likely an apical chloride channel.
Pancreatic Function Tests
An understanding of pancreatic physiology allowed pancreatic function tests to be developed. This was particularly important because of the inaccessible nature of the exocrine pancreas. These tests were either direct, which required sampling of pancreatic secretions obtained from a stimulated exocrine pancreas, or indirect, which relied on measuring malassimilated nutrients either in blood, urine, or feces. More recently, pancreatic-specific proteins such as cationic trypsinogen or pancreatic polypeptide have been measured in blood. The first use of secretion/cholecystokinin to stimulate the pancreas and obtain secretions in children was undertaken by Shwachman in 1943. This test was further refined by using marker perfusion and bilumen tubes. The modern method for measuring fecal fat was developed by Van der Kamer in 1949. Crossley and Elliott first measured immunoreactive cationic trypsinogen in blood spots in 1979, and this test forms the basis for newborn screening for CF.
Pancreatic Function and Relationship to Malabsorption
The development of pancreatic function tests allowed the concept of pancreatic reserve to be developed. Di Magno and Go in 1973 estimated on the basis of secretin/cholecystokinin stimulation that 90% of exocrine pancreatic lipolytic activity had to be lost before fat malabsorption became manifest. This was subsequently revised to a loss of 98% to 99% of lipase/colipase activity by Gaskin in 1984. This allowed patients to be classified as either pancreatic insufficient (PI), less than 2% of normal activity, or pancreatic sufficient (PS), 2% to 100% of normal activity. This was particularly important in CF, in which it was shown by Gaskin that PS patients had a markedly improved prognosis.
Advances in molecular biology have allowed the genes for specific pancreatic disorders to be pinpointed and the molecular structure of gene products determined. Intracellular localisation of gene specific proteins has followed. The cellular and biochemical consequences of gene disruption have begun to be determined. The first gene responsible for a pancreatic disorder the CF gene on chromosome 7q was characterized by Rommens, Tsui and Collins in 1989 (1) and the product CFTR (cystic fibrosis transmembrane regulator) structurally determined to be an apical chloride channel by Riordan. Since then, more than 850 separate mutations have been described in this gene, although the pathology is attributed in over 99% of individuals to approximately six common defects, of which a defect resulting in a deletion of phenylalanine in the first nucleotide binding fold ΔF508 is by far the most common (approximately 70% to 80% of homozygotes and 90% of heterozygotes). The determination of genotype has also allowed a functional classification of gene defects depending on the intracellular location of CFTR and its degree of residual function.
The hereditary pancreatitis gene on chromosome 7q was discovered by Whitcomb in 1996 (2) and shown to cause unusual stability in intracellular cationic trypsin preventing autocatalysis of trypsin. Two defects, R117H and N21I, are by far the predominant mutations (3). A defect in mitochondrial DNA has been shown to be responsible for Pearson bone marrow pancreas syndrome, and a single patient with pancreatic aplasia has been shown to have a defect in the pancreatic duodenal homeobox gene. The gene locations for lipase and colipase are known to be situated on chromosomes 10q and 6p, respectively. The genes responsible for Shwachman syndrome and Johanson Blizzard syndrome remain to be determined.
Until the 1950s, pancreatic pathology was unable to be detected by imaging techniques unless pancreatitis had resulted in calcification. Two-dimensional gray scale ultrasound developed in 1958 allowed pancreatic swelling and pseudocysts to be detected. Direct visualization of pancreatic ductular structures became possible in 1968 with the use of the side-viewing endoscope and the direct cannulation and injection of dye into the pancreatic ducts (endoscopic retrograde cholangiopancreatography). Computerized tomographic and imaging of the pancreas became available in 1973 and allowed visualization of structures adjacent to the pancreas. Magnetic resonance cholangiopancreatography became available in 1994 and allowed noninvasive imaging of the pancreatic ducts. The last technique has not been widely used in children because of concerns regarding the safety of contrast agents and movement artefact.
MAJOR DISEASE STATES
Until 1938, when Andersen described the autopsy manifestations, CF was regarded as a subset of celiac disease. CF is the commonest severe autosomal recessive disease in humans. It results in a generalized exinocrinopathy in which ductular structures in organs with exocrine glands secreting protein, water, and anions become clogged with inspissated proteinaceous secretions. The secretory defect that results from a mutation in an apical chloride channel in the epithelial cells of exocrine glands resulting in reduction of C1−, HCO−3 and water secretion. Of CF patients, 85% are PI, 65% are PI at birth, and 20% become rapidly insufficient during the first few months and years of life (3). The 15% of CF patients who are PS have a markedly improved prognosis (4) and seldom develop significant disease in other organ systems.
Approximately 15% of CF present at birth with either meconium ileus or meconium peritonitis, and, at some stage of their disease course, 10% of patients develop distal intestinal obstruction syndrome. Five percent of CF patients develop significant liver disease secondary to focal biliary cirrhosis. This may result in death through either fulminant hepatic failure or variceal bleeding secondary to portal hypertension. CF liver disease is becoming an increasing problem as respiratory prognosis improves and most patients survive into adult life. The common underlying pathologic condition relates to obstruction of ductular structures with inspissated proteinaceous secretions. However, it is not understood why some patients with common identical severe mutations are PI at birth and some are not and develop PI later, and why some PI patients go on to develop meconium ileus or distal intestinal obstruction syndrome or liver disease and some do not.
The pathology of pancreatitis is poorly understood. Both bacterial infection and obstruction were recognized as important precipitants a century ago. Viral infections and drugs may result in direct cellular injury. Most cases of pancreatitis do not have an obvious precipitant.
Two advances are responsible for a more sophisticated understanding of pancreatitis. The first was the cloning of the genes for autosomal dominant hereditary pancreatitis and cystic fibrosis, the second and more recent advance is the appreciation that some chemokines are specifically expressed in acute pancreatitis, including interleukin 8, epithelial neutrophil activating peptide ENA78, and monocyte chemoattractant protein MCP-1 (5).
The diagnosis of pancreatitis is hindered by the lack of specific tests and is usually made on the basis of elevated serum pancreatic enzymes such as amylase and lipase in the presence of typical symptoms of pancreatic pain and a swollen pancreas on either ultrasound or computed tomography (CT) examinations. No consensus of opinion exits as to the gradation of childhood pancreatitis into mild, moderate, or severe disease.
Nutritional (tropical) pancreatitis was first described in 1955, and the cause is poorly understood. This disorder may result in chronic pain and ongoing malnutrition. Diabetes mellitus also may result from nutritional pancreatitis. The pancreas commonly displays calcification. Exocrine pancreatic function is compromised in malnutrition and may not completely recover. However, this disorder is prevalent in some tropical or third world countries where malnutrition is prevalent and not in other countries that are equally disadvantaged. It is likely that genetic and environment factors such as diet are just as contributory as malnutrition per se.
Pancreatic enzyme supplementation, usually with porcine enzyme extracts, is the only currently available therapy for PI. Pancreatic enzymes are rapidly degraded by stomach acid and pepsin and function best in an alkaline environment. Lipase and colipase require bile acids for optimal fat digestion and micelle formation. CF patients may have high gastric acid outputs and deficient alkalinisation of the duodenum secondary to deficient hepatic, pancreatic, and intestinal crypt
HCO3 production. These factors may limit the effectiveness of pancreatic enzyme supplements.
Enteric-coated microspheres have been shown to increase delivery of enzymes to the duodenum and distal small bowel. Although this results in improved utility, there have been problems with miscibility of microspheres with food and also suboptimal dissolution of the spheres in a relatively acid environment.
Other approaches to increasing enzyme utility have concentrated on improving the intestinal milieu and making it more enzyme friendly. H2 antagonists, protein pump inhibitors, and misoprostil have all been used to either reduce stomach acid or to render the duodenum more alkaline. These therapies, particularly protein pump inhibitors, do result in improved enzyme utility.
Lipase doses above 25,000 units of lipase/kg/d have been shown to be associated with fibrosing colonopathy, which may result in intestinal obstruction secondary to colonic stricture, failure to thrive, abdominal pain, and occasionally blood diarrhea and ascites (6). Marked ulceration and submucosal fibrosis is seen in the affected colon and occasionally regenerative nodular hyperplasia results. For this reason, it is recommended that lipase dose be kept below 10,000 units/kg/d.
No specific therapy exists for acute pancreatitis. Most therapy is palliative and relies on bowel rest, adequate hydration, and pain relief through patient-controlled analgesia. Intravenous broad-spectrum antibiotics should be administered in severe acute pancreatitis. Octreotide has not been shown to be useful for acute pancreatitis but may be of benefit in treating pancreatic pseudocysts and pancreatic fistulae. Surgical debridement of the pancreas is seldom if ever needed in childhood pancreatitis.
No convincing therapy exists for recurrent pancreatitis. Antioxidants, particularly vitamins A, E, and C, have been used, as has selenium, but the evidence is far from compelling. Surgical intervention might be possible if a defined defect is demonstrated on endoscopic retrograde cholangiopancreatography (ERCP), but this is rarely the case.
Trypsin has been used to reduce pancreatic pain. Trypsin cleaves a molecule called monitor peptide, which releases cholecystokinin, rendering this molecule ineffective. Although this theoretically should reduce pancreatic drive, this therapy has not been shown to be useful in children. Most chronic pancreatitis become less severe with time, and if possible surgery should be avoided. Pancreatic pseudocysts can be drained into the stomach under CT guidance and stents placed, allowing resolution of these lesions.
Liver and Intestinal Disease in Cystic Fibrosis
Ursodeoxycholic acid has not been demonstrated on Cochrane meta-analysis to be beneficial in cystic fibrosis. This hydrophilic bile acid does, however, result in increased choleresis and is membrane protective. It has also been shown in most studies in CF patients to result in improved serum transaminase levels and improved biliary excretion on nuclear medicine scanning. One small study has suggested histologic improvement over a 2-year period (7).
Liver transplantation is an option for end-stage liver disease, as is vascular shunt surgery for portal hypertension. The management of distal intestinal obstruction syndrome has improved with the use of nonabsorbed balanced intestinal lavage solutions such as Golytely, which rely principally on polyethylene glycol and sodium sulfate to provide an iso-osmotic fluid load.
IMPACT ON CHILD HEALTH
The global impact of pancreatic disease in childhood is difficult to quantify. We do not know how many children in the developing world are malnourished, nor do we know how many of these will develop long-term PI or diabetes mellitus. Most of these children have no access to pancreatic enzyme therapy or insulin. Some may go through childhood racked with pain, and life expectancy is reduced. The financial costs of pancreatic disorders other than CF are not known.
CF is estimated to cost approximately $40,000 a year per patient in the United States. This places an intolerable burden on families in societies in which a socialized health system is not available and where lifetime capitation limits are placed either by health insurance firms or health maintenance organisations. The annual cost of CF to the health system in the United States is approximately 1.2 × 109 U.S. dollars. In Australia, pancreatic enzyme supplementation alone may cost $3,000 to $5,000 per annum.
Families may have considerable difficulty adjusting to the realization that the newborn baby they had previously thought to be normal has a chronic disabling condition that requires daily therapy and may well have a reduced life expectancy with reduced or absent fertility. Very few quality of life studies have been undertaken in CF, but it is expected that therapeutic demand causes considerable problems for individual patients from school age on and may impact on employment and marriage.
II. AREAS IN NEED OF INVESTIGATION OR IMPLEMENTATION
Further Clarify the Molecular Events Governing Pancreatic Development
Defects in the Xenopus laevulus–derived homeobox genes Pdx and HIXb9 result in pancreatic aplasia (8). The genes that are responsible for the development of the dorsal and ventral pancreas and regulate the fusion of the two glands and their ducts are unknown. Determination of these genes would help us to understand anatomic pancreatic anomalies such as annular pancreas and pancreas divisum.
Differentiation into Specific Cell Types
The genes that induce pancreatic stem cells to differentiate various cell types, including acinar cells, islet cells, somatostatin secretory cells, and the mensenchymally derived pancreatic stellate cells, are largely unknown. The homeobox genes Isl-1, pax4, and pax6 govern murine islet cell formation and cell differentiation into insulin and glucagon secreting cells, respectively. Further delineation of genes specific for cell differentiation may allow the development of strategies aimed at preserving pancreatic function through regeneration or promotion of specific cell types.
Expression of Cell-Specific Proteins
The genes for some specific pancreatic proteins such as cationic trypsinogen, lipase, and colipase have been cloned, and the genomic location of other proteins are known. The factors governing expression are not known. This is true both for an isolated point and also for different ontogenic stages for development. For example, at birth lipase secretion is suboptimal and increases over the first year of life. The events responsible for this maturational delay are unknown. The events involved in signal transduction and the nuclear transcription factors involved in gene expression need to be determined. Modification of nuclear transcription may provide opportunities to augment secretion of potentially useful products such as pancreatic enzymes.
Factors Governing Growth, Regeneration, and Programmed Cell Death
The exocrine pancreas is subject to trophic growth factors. These growth factors may be derived from the small intestine, secreted into the splanchnic circulation, and act in an endocrine fashion, or they may be derived locally from the pancreatic parenchyma or acini and act as paracrine factors on adjacent cells. Some of the endocrine trophic hormones such as cholecystokinin are well characterized and may result in pancreatic hyperplasia when infused in the mouse. Most endocrine trophic factors active on the exocrine pancreas, however, have not been characterized. The pancreatic specific effects, receptors, signal transduction, and nuclear transcription factors of most intestinally derived peptide trophic hormones are not known.
Locally produced trophic hormones in the exocrine pancreas are even less well understood. Recently, a number of specific growth factor genes such as β cellulin (a member of the EGF superfamily) have been shown to be expressed in pancreatic tissue (9). An understanding of pancreatic growth is important in understanding developmental anomalies as well as fetal and postnatal development but also may aid in understanding recovery of pancreatic acini from insults such as malnutrition or alcohol. It also may assist in our understanding of pancreatic neoplasia.
Programmed cell death or apoptosis may balance growth and remove dysfunctional cells. The genes responsible for pancreatic exocrine cell apoptosis remain to be characterized.
Develop New Tests and Therapies for Pancreatic Insufficiency and Pancreatitis
Tests for Pancreatic Insufficiency and Pancreatitis
The ideal pancreatic function test does not exist. Such a test should be inexpensive, easily performed, repeatable, and reliable. The ideal pancreatic function test should be specific for pancreatic disease and able to exclude patients with other digestive disorders such as small bowel mucosal disease, inherited defects of fat transport, or cholestasis. It should be able to define the exact level of pancreatic function in subjects with PI in whom partial impairment of exocrine function is present but nutrient assimilation is unaffected. There should be no interference from pancreatic enzyme supplements.
Most current tests have major limitations. Pancreatic stimulation tests are the only tests capable of measuring functional reserve. They are invasive uncomfortable, time consuming, and expensive and require each laboratory to establish its own normal range. The widely used 3- to 5-day fecal fat test is unpleasant for patients, parents, and laboratory staff. It is a test for malabsorption but does not separate out pancreatogenous malabsorption from other causes. This test and most other tests are subject to interference from pancreatic enzyme supplements with the exception of fecal elastase. Elastase is not found in enzyme supplements, and low levels of elastase indicate pancreatic disease but do not indicate the extent of pancreatic reserve.
Immunoreactive cationic trypsinogen with subsequent gene analysis for cystic fibrosis should be offered as a newborn screen in countries in which CF is prevalent. Nutritional prognosis is improved with early identification, and newborn screening may allow the timely introduction of novel therapies.
Acute pancreatitis poses a separate problem. In most instances, pancreatic enzymes efflux into the circulation, and elevations of circulating pancreatic isomylase, cationic trypsinogen, and lipase can be found. This is useful for diagnosis but does not give any indication of the severity of disease. The holy grail for pancreatitis would be the development of a test that allows early prediction of severity. Currently serum and urine trypsin activation peptides correlate with severity of illness in adults, but these tests require validation in children. A test that indicates severity may help predict outcomes and may guide therapeutic decisions such as when to re-feed patients.
New Therapies for Pancreatic Insufficiency and Pancreatitis
Pancreatic Enzyme Supplementation
Pancreatic enzyme supplements and adjuvant therapies even in compliant patients often fail to correct malabsorption. There is a need to develop alternative agents that maintain lipolytic activity in an acid environment. These agents would not require colipase as a cofactor and consequently would be less dependent on bile salts. Bacterial fungal, plant-derived, human, and animal gastric lipases are all potentially useful. Biologically derived material should be free of harmful contaminants such as plant lectins. Some of these agents could be produced by recombinant genetic engineering.
Three approaches are theoretically useful for developing specific therapies for pancreatitis. These are 1) inhibition of the catalytic effects of trypsin, 2) dampening down the inflammatory cascade responsible for the release of trypsin into the circulation, 3) reducing pancreatic calcification and fibrosis after the inflammatory insult. The first approach has proved largely ineffective. This may be due to damage occurring before treatment with trypsin inhibitors but also may relate to damage secondary to enzymes other than trypsin. Early recognition of disease is essential for this therapy to be effective, and this also holds true for therapy directed at the inflammatory cascade. The identification of specific chemokines in acute pancreatitis raises the possibility that monoclonal antibodies directed at inflammatory cytokines or receptor antagonists usually fusion proteins aimed at blocking binding to receptors could theoretically block acute pancreatitis.
Recurrent episodes of acute pancreatitis may result in fibrosis, calcification, and pancreatic insufficiency as well as chronic pain. Therapy aimed at reducing scarring could be potentially beneficial. Scarring results from collagen formation from primed pancreatic stellate cells. These cells produce collagen in response to transforming growth factor beta. Therapies aimed at downregulating TGFβ secretion or reducing pancreatic stellate cell responsiveness to TGFβ may be of benefit in chronic pancreatitis.
Modulation of Progression from PI to PS in CF
Therapies are needed to halt or reverse the progression to PI in CF. The factors governing progression are incompletely understood, especially because progression may vary between patients with identical genotypes. Two possible approaches involve a) increasing fluid secretion and b) treatment with docosahexanoic acid (DHA), which reverses pancreatic changes in the CF mouse model.
The first approach has shown promise in human small intestinal explants and cultured CF small intestinal epithelial cells. The phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX) has been shown to partially correct the defect in CFTR (Cystic Fibrosis transmembrane conductance regulator)-mediated apical chloride secretion. No evidence exists as to the effect of this agent on pancreatic ductular epithelial cells. For this agent and similar drugs to progress to therapeutic trials, they would need a wide therapeutic index and not produce toxicity. Docosahexanoic acid an omega 6 fatty acid, reverses the pathologic changes seen in the exocrine pancreas of CFTR-deficient mice (10). The mechanism of action is unknown. Fatty acids can act as second messengers, and it is possible that in the mouse DHA may activate an alternative chloride channel. This therapy needs to be explored in a variety of CFTR defects in animal models and trialed in humans. Both pancreatic secretagogues and DHA would have to be started early in life and preferably in utero to have a reasonable prospect of reversing pancreatic pathology. DHA supplementation in breast-feeding mothers and addition of DHA to infant formula have resulted in disappointing increases in serum DHA levels. If the mechanism of action were known, it might be possible to engineer a more potent and specific synthetic fatty acid.
Gene therapy for CF has been frustratingly elusive. After the discovery of the CFTR gene, pervading sentiment held that it would only be a matter of time before this became a reality. Most current techniques do not result in long-term incorporation of CFTR into the genome. Repeated administration is problematic and in the case of the most widely used vectors, adenoviruses, and adeno-associated viruses, the human immune response has resulted in respiratory compromise (11). It is also difficult to deliver the gene vector to the target tissue, and, in the case of the pancreas ERCP, probably provides the only realistic method of delivering gene vectors to the pancreatic ducts. This is technically difficult in very small infants, who constitute the group most likely to benefit from gene therapy.
Retroviral or Lentiviral vectors that have been mutated, preferably in two widely separated sites to render them nonreplicating, represent the best chance for genomic incorporation of CFTR. The remote chance of reconstitution of wild-type virus, particularly if human immunodeficiency virus (HIV) is used to construct the gene vectors, renders this approach ethically fraught.
Characterize the Clinical Parameters of Childhood Pancreatitis
Published studies of pancreatitis in childhood are retrospective, the severity of pancreatitis is not assessed, and clinical detail is often lacking. In developed societies, we do not know the incidence or prevalence of childhood pancreatitis, and because of this we are unable to assess whether this disorder is increasing in prevalence and if so why? We are also unable to determine the economic cost to society. Prospective studies are required to assess the incidence, risk factors for pancreatitis, the best diagonistic criteria, the natural history of the disorder, and ultimately the long-term prognosis.
Similar prospective studies are needed in the developing world to look at nutritional pancreatitis. Such studies would examine incidence and prevalence and look at predisposing risk factors such as diet, micronutrient deficiencies, and infective load. A simple screening test for pancreatic damage in nutritional pancreatitis would greatly aid case definition and identification.
Prospective studies would allow more specific investigation of idiopathic and other forms of pancreatitis. Patients with acute pancreatitis should be screened for predisposing allelic mutations such as the 5T polythymide in defect in intron 8 in the CFTR gene and the R117H and N21I mutations in the cationic trypsinogen gene. Prospective studies would also allow established and novel therapies to be assessed and the utility of new imaging modalities such as MRCP to be determined.
Prospective studies in nutritional pancreatitis would allow further investigation of deficiency of trace metals such as zinc and selenium and free radical scavengers such as vitamins A, C, and E and whether supplementation may favorably alter the course of the disease.
Define the Pathogenesis of Acute and Chronic Pancreatitis
The molecular and cellular events resulting in acute and chronic pancreatitis are largely unknown. Normal and pancreatitic exocrine pancreatic tissue is seldom available for genetic analysis but if available gene microarray technology could conceivably identify genes that may in themselves be causative for pancreatitis or contribute to the perpetuation of pancreatitis. Candidate genes could include genes that encode cytokines specific for pancreatitis, nuclear transcription factors, genes that govern zymogen release, cellular adhesion molecules, metallothionases and apoptotic factors. Nanotechnology with the development of injectable computerized microrobots capable of sampling tissue may well progress in this area in the foreseeable future.
A wide range of disparate clinical disorders result in acute pancreatitis. A study of the molecular consequences of these disorders may provide insight on the causation of pancreatitis. For example, the antiepileptic drug sodium valproate and Pearson Bone Marrow Pancreas Syndrome both result in mitochondrial damage or deficiency with resulting impaired generation of intracellular energy. Similarly, homocystinuria may result in pancreatitis. Defects in another gene involved in homocystine metabolism, methylene tetrahydrofolate reductase, result in raised homocystine levels and are a significant contributor to cardiovascular mortality through endothelial damage and activation of thrombosis. The prevalence of predisposing polymorphic alleles for this gene should be surveyed in patients with idiopathic pancreatitis. Patients with autoimmune pancreatitis should be investigated and the target antigens in the pancreas determined.
The identification of specific genes such as 5T defect in CFTR and R117H in cationic trypsinogen should allow animal models to be created in which firstly predisposing events can be charted and secondly tissue can be obtained to plot the intracellular and molecullar consequences.
The best example of this is found with the R117H mutation. It is believed that the major consequence of this defect is that an inherently stable activated trypsin is no longer subject to autocatalysis. This finding does not explain why patients with this disorder have periodic bouts of pancreatitis rather than constant pancreatitis, nor does it explain what the trigger events are for this form of pancreatitis.
Define the Molecular Details of Pancreatic Disease in Cystic Fibrosis
The function of the normal product of the CF gene CFTR is not completely understood in normal pancreatic ductular epithelium. An understanding of normal function is required to dissect the effects of various mutational abnormalities or to begin to understand how we could intervene therapeutically through signal transduction or upregulation of alternative ion channels.
Mutations that result in PS should also be studied, as should the 1% to 2% of patients homozygous for the common allele ΔF508/Δ F508 who maintain pancreatic sufficiency for protracted periods. Current mouse models of cystic fibrosis have concentrated on severe alleles such as Δ F508/Δ F508 and G551D/G551D and it would be beneficial to have a variety of models that encompass all classes of CFTR mutations.
The factors governing the variable progression of PS to PI are not understood. CF patients have increased phosopholipid-bound arachidonic acid and decreased DHA pancreatic tissue levels. It is possible that increasing tissue fatty acid levels may retard progression to PI. The major protein contributant to pancreatic ductular plugs in pancreatitis is the predominant protein found in zymogen granules, glycoprotein 2. This 78-kd protein is released by exocytosis into the acini. Differential release of glycoprotein 2 could result in varying rates of progression in cystic fibrosis. Other components of ductular plugs also may be problematic, for example, the high Na+ concentrations may inactivate β defensins, which protect against colonisation and invasion by pathogens (12).
CFTR dysfunction results in increased expression of some cytokines and decreased expression of others in both pancreas and lung. The mechanisms responsible for this are unknown.
Characterize the Pathophysiology of Liver and Intestinal Disease in Cystic Fibrosis
The diagnosis of cystic fibrosis liver disease is problematic. CF liver disease is very patchy on liver biopsy, and consequently most studies have not used histologic appearances to confirm disease. Other studies have used clinical hepatomegaly, raised transaminases, and ultrasound appearances. Because of the variable criteria used, the reported prevalence of liver disease varies widely. CF liver disease is attributed to deficient cholangiolar secretion of water and anions.
Biliary epithelial cells express CFTR. The regulation of CFTR function in the biliary tract is not completely understood. The anion exporting activities of normal and mutated CFTR and the effects of CFTR activity on the structure, function, and expression of biliary mucin should be studied. Differences in CFTR function are unlikely to explain why some patients with identical gene defects develop liver disease and some patients do not. Two possible explanations exist for these differences. The first includes gene modifiers and the second involves an acquired structural predisposition to obstructive liver disease through the development of common bile duct strictures.
The effects of the gene defect may be amplified if other potentially hepatotoxic factors exist. CF patients are known to provide toxic hydrophobic glycoconjugates of bile acids and differential excretion could predispose to damage. Similarly, biliary mucin secretion could vary between patients producing greater obstruction in some patients compared with others. CF patients have increased intestinal permeability (13), and it is conceivable this could result in differential antigen sieving by the liver with deposition of immune complexes. Liver fibrosis is controlled by hepatic stellate cells, which are also responsible for vitamin A storage. Deficiency of vitamin A at tissue level and differential expression of TGFβ with differential activation of hepatic stellate cells also may result in differences in inherent susceptibility to CF liver disease. Most of these contentions remain to be explored.
Gaskin et al (14) in 1989 identified by percutaneous transhepatic cholangiogram a subset of CF patients with liver disease who had demonstrable stenosis of the common bile duct as it passed through the head of the pancreas. This was thought to be caused by fibrosis in the head of the pancreas and has been found in both PI and PS CF patients. Symptoms of right upper quadrant pain and raised transaminases improved markedly after cholecystojejunostomy. There is not universal agreement about this phenomenon, and larger-scale studies with an agreed imaging modality are required. Further larger long-term case-controlled studies of the effect of ursodeoxycholic acid are required in CF liver disease.
The intestinal manifestations of CF—meconium ileus and distal intestinal obstruction syndrome—are also variably expressed even within families in which, for example, one baby with CF may have meconium ileus and another will not. The reasons for this are not understood, and gene modifiers may come into play. Differential expression of intestinal mucins should be investigated, and variable intestinal transit also should be considered. Variable intestinal transit also may partially explain the increased intestinal permeability seen in CF, but this phenomenon also needs further investigation.
The mechanisms underpinning fibrosing colonopathy should be determined. The enteric coating Eudralgit was initially thought to be implicated, but this has been discarded. Other possibilities include 1) a direct injurious effect of enzymes; 2) high-dose enzymes resulting in the accumulation of a toxic metabolite; 3) an immune response to the enzymes resulting in local injury (colonic stricture) and distant injury (nodular hyperplasia of the liver). The first possibility is unlikely, because we would expect more fibrosing colonopathy, although interindividual variations in intestinal transit may predispose to injury. Pancreatic enzyme supplements do not mimic life; they are high in lipase but do not contain phospholipase A. This imbalance could result in the generation of toxic metabolites such as lysolecithin, which has been shown to be directly injurious to intestinal epithelia. Couper et al (15) have shown a brisk systemic and mucosal immune response to pancreatic enzymes and recent evidence from Toronto shows this peaks at the time of peak prevalence of fibrosing colonopathy after enzyme initiation.
The risk of neoplasia in CF needs to be determined. Adenocarcinoma of the small intestine is increased in adult CF patients. This problem may become increasingly apparent as more patients survive well into adult life. It is possible that patients with long-standing liver disease may develop hepatocellular carcinoma. Cirrhosis per se is known to increase the risk of hepatocellular carcinoma in normal individuals.
Determine the Mechanism of Pancreatic Dysfunction in Other Disorders
Determining the genes and gene products responsible for other pancreatic disorders may provide fresh insights in pancreatic pathology and allow new therapies to be developed. Shwachman-Diamond syndrome is unique amongst pancreatic disorders in that malabsorption improves with age. The reasons for this are unknown and are unlikely to be plumbed until the gene is identified. The best prospect for this lies in the identification of informative families with more than one member afflicted. Similarly, the reasons for pancreatic dysfunction in Alagille's syndrome are unknown. The recently identified genetic defect in the Notch Ligand Jagged 1 may provide a clue. Notch is a transmembrane receptor responsible for determining early cell fate in a number of cells, including epithelial cells. The utility of Jagged 1 and Notch are unknown for pancreatic epithelial cells.
Differential expression of CFTR also may play a role in other pancreatic disorders. It has been recently shown that differential expression of CFTR in the biliary tree may contribute to sclerosing cholangitis and if this held true for pancreatic ducts this could be responsible for variable expression of disease such as the secondary pancreatic insufficiency seen in celiac disease.
Determining Whether Pancreatic Disorders in Later Life Have Their Genes in Childhood
The Barker hypothesis holds that events in later life may result directly from intrauterine programming. For example, it has been shown that the age of onset of juvenile diabetes mellitus, a disorder of the endocrine pancreas is related to birth weight. This has been postulated to be due to reduced β cell mass. It is possible therefore that the progression of pancreatic disease in disorders such as CF may be partially determined by gestational factors such as placental function, which could conceivably reduce acinar cell mass. This also may be important in nutritional pancreatitis in which in utero malnutrition may already have reduced functional pancreatic reserve.
Disorders of the pancreas in later life also may partially governed by events more classically associated with childhood disorders. CFTR status and function has not been assessed in either alcoholic pancreatitis or pancreatic adenocarcinoma.
III. PROPOSED PLAN TO ACHIEVE GOALS
Critical mass is the major problem impeding progress in CF and other pancreatic diseases. Too few investigators currently work in this field. The following are required to redress this problem and are common to all the areas requiring investigation or implementation.
- The pancreatic component of pediatric gastroenterology training should be formalized, and the recent NASPGAN guidelines should serve as a template (16).
- Pediatric gastroenterologists interested in pancreatic disorder and cystic fibrosis should continue to meet at forums such as NASPGAN or the USCFA meeting to determine standard definitions of disease, diagnostic criteria, and investigative protocols.
- All CF clinics should have a designated pediatric gastroenterologist in attendance.
- Pediatric gastroenterologists should be involved in the transition of CF patients to adult clinics. This would allow long-term follow-up of therapeutic interventions and disease outcomes.
- All developed countries should have a CF disease register and database.
- A central international register of all basic science research into pancreatic disorders of childhood should be established. This could be undertaken by a group of individuals or by organisations such as NASPGN, ESPGN, or the USCFA.
- A central international register of all clinical research and therapeutic trials into pancreatic disorders of childhood should be established. It is essential that epidemiologic studies and therapeutic trials, particularly in rarer disorders, should be collaborative and multicenter.
- Redundant tissue from pancreas transplant programs should be separated into cell types and made available to basic science research programmes in pancreatic disease.
Factors Affecting Pancreatic Development
Pancreatic tissue and tissue obtained from normal children and adults should be screened by gene microarray for candidate homeobox genes, nuclear transcription factors, and genes controlling apoptosis. Large animal models such as the pig should be established to assess both overexpression and deletion of the gene products identified by the above process. These animals should be screened for anatomic abnormalities, ultrastructural abnormalities, and biochemical abnormalities. Transgenic animals should be assessed temporally to allow the ontogenic development of the pancreas to be charted. Transgenic or site deleted animal models expressing gut related regulatory peptides or trophic hormones should also be assessed for pancreatic abnormalities.
Mutated animals would allow ready access for both in vitro and in vivo studies of molecules involved in signal transduction. They also would allow improved techniques for isolation of pancreatic stem cells and the long-term culture of nontransformed pancreatic cells, particularly acinar and pancreatic stellate cells. Matrix culture techniques allowing cellular co-culture to assess the paracrine effect of trophic hormones should be developed for pancreatic acinar, stellate, and endocrine cells.
These studies would be best funded by national governmental funding organizations such as the NIH or the United Kingdom or Canadian MRC.
New Tests and Therapies for Pancreatic Insufficiency and Pancreatitis
Gene microarray technology will allow the identification of genes and gene products specific to the normal pancreas. These gene products should be screened for in serum, urine, and feces of normal individuals, CF patients with either pancreatic insufficiency or pancreatic sufficiency, and patients with acute and chronic pancreatitis. The identification of quantitative differences may allow the development of specific tests for pancreatic insufficiency and pancreatitis.
Alternative safe lipolytic acid stable enzymes need to undergo trials in humans. Trials should be undertaken with DHA supplementation in newborn CF patients. Such studies should also be considered in mothers who have undergone prenatal diagnosis and have elected to proceed with a CF affected pregnancy. Studies of engineered synthetic n-6 fatty acids should be undertaken in CF animal models and structural examination of the pancreas undertaken.
Newborn screening using immunoreactive cationic trypsinogen and genetic testing for common CF mutations should be lobbied for in developed countries where this is not already available. This will allow therapies that hich retard disease progression to be initiated age appropriately. Safe vectors for long-term incorporation of the normal CFTR gene into the gene need to be developed.
Identification of pancreatic-specific molecules and gene therapy vectors are best funded by governmental agencies and large nongovernmental agencies such as the USCFA. Trials of lipolytic agents, especially phase I and phase 2 trials, are best funded by industry. Formula companies and biotechnology companies should fund studies of DHA and engineered synthetic fatty acids.
Clinical Parameters of Childhood Pancreatitis
Multicenter epidemiologic studies would determine the clinical, biochemical, and genetic parameters of acute and chronic pancreatitis. Patients should be screened for CFTR mutations and mutations in the cationic trypsinogen gene. Additionally, a radiologic arm should allow the development and assessment of magnetic resonance pancreatography in children. A classification of disease severity should be developed. These studies should be funded by both governmental funding agencies and in-house sources.
Epidemiologic studies of nutritional (tropical) pancreatitis are needed and should concentrate on specific ethnocultural differences in dietary practices and deficiencies of trace metals and free radical scavengers. If specific deficiencies are identified, large-scale population supplementation studies should be undertaken. These studies should be funded by the WHO or FAO.
Pathogenesis of Acute and Chronic Pancreatitis
Large animal models for pancreatitis should be established. In the first instance these should be based on the gene defects known to predispose to pancreatitis. These models would facilitate the study of pancreatic tissue for upregulation or downregulation of chemokine genes. They would allow tissue to be obtained for the study of the molecular regulation of these genes and also facilitate the development of techniques to allow sampling in acute pancreatitis. They would facilitate the study of trigger events such as bacterial infection and also allow therapies directed at modulating the inflammatory chemokine cascade such as tumor necrosis factor antibody or receptor antagonist to be assessed. Funding should be obtained from governmental or in-house sources, although studies of specific therapies could be assisted by industry.
Molecular Details of Pancreatic Disease in Cystic Fibrosis
Studies should be undertaken in fetal tissue obtained from CF-affected pregnancies. Pancreatic tissue should be cultured from individuals with a range of CF defects. Ion fluxes should be determined either by patch clamping, ion-specific dyes, or by electron microscopic x-ray diffraction analysis. Signal transduction should be studied and molecules of interest such as DHA assessed. Cytokine expression should also be assessed. The range of animal models in CF should be expanded to include animals with PS CF. The ontogeny of disease could be assessed in these models and therapeutic trials undertaken. Funding should come from National Cystic Fibrosis Foundations or governmental funding agencies such as NIH.
Pathophysiology of Liver and Intestinal Disease in Cystic Fibrosis
Pediatric gastroenterologists should establish a consensus on the diagnosis of CF liver disease and a protocol for assessing progression of disease and the effect of therapy. Large-scale long-term population-based studies are required to establish whether ursodeoxycholic acid is efficacious in CF and whether common bile duct stricture contributes to liver disease.
CFTR regulation should be studied in biliary epithelium obtained from CF and normal fetal pancreas and from CF patients who undergo liver transplantation. CFTR regulation should be studied in CF and normal human fetal intestinal epithelium as well as intestinal biopsy specimens obtained endoscopically from older CF patients. Similar studies should also be undertaken in patients who are born with meconium ileus to determine quantitative difference. Intestinal mucin characteristics and mucin gene expression should be studied in a variety of human CF genotypes. Mucin expression should then be correlated with intestinal permeability. Mucosal immunity and immune exclusion warrants further study in patients with altered permeability and patients with fibrosing colonopathy.
Distal intestinal obstruction develops in small animal models of CF such as mice, and therapy can be assessed in these animals. However, liver disease develops over time, and there is a need to develop a large animal model to facilitate studies of CF liver disease. Such a model would also allow assessment of intestinal transit abnormalities and the mechanisms underpinning fibrosing colonopathy. Cancer risks should be able to be assessed through CF registries, national mortality data, and improved surveillance through to adult life. Funding of these studies should be obtained from governmental agencies and the cystic fibrosis foundations.
Mechanism of Pancreatic Dysfunction in Other Disorders
International collaboration will lead to the identification of the gene for Shwachman syndrome. Rare disease registers or paediatric surveillance units would allow informative families to be identified. Once the gene is identified, an animal model can be created and the disease ontogeny studied at a structural biochemical level. CFTR status should be assessed in other disorders associated with pancreatic dysfunction. The function of Notch and its ligand Jagged should be determined in normal pancreatic tissue. These studies are best funded from in-house sources, although agencies such as the March of Dimes Foundation may contribute to genetic research.
Disorders of the Pancreas in Late Life
The exocrine pancreas should be studied in intrauterine growth restriction models such as the carunculized lamb and such animals followed temporally. Epidemiologic studies of idiopathic pancreatitis later in life should look at parameters such as birth weight and gestational age and lifestyle factors known to be associated with intrauterine growth restriction such as insulin resistance, hypertension, obesity, and cardiovascular disease. Common disorders of the pancreas in adult life such as alcoholic pancreatitis and pancreatic adenocarcinoma should be assessed for CFTR heterozygosity.
Studies looking at intrauterine factors pertaining to pancreatic disorders should be funded by institutional funding bodies when animal studies are involved or by governmental funding. Long-term epidemiologic studies could be incorporated into studies similar to NHANES (National Health and Nutrition Examination Survey) in the United States. Studies looking at molecular predisposition to adenocarcinoma of the pancreas could be funded by bodies such as the National Cancer Institutes or Cancer Foundations.