| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Institut National de la Santé et de la Recherche Médicale U654 and Department of Genetics (I.G.), Hôpital Henri Mondor, 94010 Creteil, France; Department of Pathology (C.S., J.R.), Cliniques Universitaires Saint-Luc, Université Catholique de Louvain, B-1348 Brussels, Belgium; Department of Biology (C.B.-C.), Hôpital Saint-Antoine, 75012 Paris, France; Service Hospitalier Frédéric Joliot (M.R.), Commissariat a lEnergie Atomique, Direction des Sciences du Vivant, Departement de Recherche Médicale, 75270 Orsay, France; and Departments of Pediatrics, Radiology, and Surgery (L.H., N.B., J.-M.S., J.-J.R., F.B., F.J., C.N.-F., P.d.L.), Hôpital Necker-Enfants Malades, 75270 Paris, France
Address all correspondence and requests for reprints to: Dr. Irina Giurgea, Department of Genetics, Hôpital Henri Mondor, Créteil, 51, Avenue du Mal de Lattre de Tassigny, 94010 Creteil Cedex, France. E-mail: irina.giurgea{at}im3.inserm.fr.
| Abstract |
|---|
|
|
|---|
Patients: We here report two patients with bifocal CHI as evidenced by relapsing hypoglycemia after removal of the first focal lesion and the detection of a second, distinct, focal adenomatous hyperplasia during later surgery (patients 1 and 2) and a patient with a giant focal lesion involving the major part of the pancreas (patient 3).
Results: In the three patients, a germline, paternally inherited, mutation of SUR1 was found. In patients 1 and 2, haploinsufficiency for the maternal 11p15.5 region resulted from a somatic deletion specific for each of the focal lesions, as shown by the diversity of deletion break points. In patient 3, an identical somatic maternal 11p15 deletion demonstrated by similar break points was shown in two independent lesion samples, suggesting a very early event during pancreas embryogenesis.
Conclusion: Individual patients with focal hyperinsulinism may have more than one focal pancreatic lesion due to separate somatic maternal deletion of the 11p15 region. These patients and those with solitary focal lesions may follow the two-hit model described by Knudson. The stage of embryogenesis at which the somatic event occurs may account for the observed histological diversity (early event giant lesion, later event small lesion).
| Introduction |
|---|
|
|
|---|
The most common pathophysiological mechanism underlying both forms is a dysfunction of the pancreatic ATP-sensitive potassium channel composed of two tetramers encoded by the sulfonylurea receptor gene (SUR1 or ABCC8) and the inward-rectifying potassium channel gene (KIR6.2. or KCNJ11) (7, 8, 9, 10, 11). However, focal and diffuse CHI results from different molecular mechanisms.
Diffuse hyperinsulinism appears as a genetically heterogeneous endocrine disease related to mutations in one of the genes encoding the proteins involved in insulin secretion (7, 9, 10, 11, 12, 13). In the autosomal recessive form, the most frequently involved genes are SUR1 (7) or KIR6.2 (8, 14) and, to a lesser degree, the short-chain L-3-hydroxyacyl-CoA dehydrogenase gene (11). Autosomal dominant forms are less common and have been related to mutation of SUR1, the glucokinase (9), and the insulin receptor genes (15) and, in cases associated with hyperammonemia, the glutamate dehydrogenase (GLUD1) gene (10). In the neonatal period, diffuse CHI is mostly related to dysfunction of the pancreatic ATP-sensitive potassium channel, which is histologically characterized by hyperfunctional ß-cells described as containing abundant cytoplasm and large nuclei throughout the whole of pancreatic islets (1, 16, 17).
Focal CHI has been shown to result from a paternally inherited mutation in the SUR1 or KIR6.2 genes, both located in the 11p15.1 region, and loss of the maternal 11p15 allele [(loss of heterozygosity (LOH)]. LOH is a somatic event restricted to the pancreatic lesion (18). Histologically, focal adenomatous hyperplasia is a small poorly delineated lesion composed of normally structured hyperplastic islets (ß-cells surrounded by non-ß-cells), separated by few exocrine acini, thus maintaining a normal lobular pancreatic architecture (19). A high proliferation rate of ß-cells was shown inside the lesion, whereas in the normal adjacent pancreas, small resting islets, made of packed endocrine cells with scanty cytoplasm, exhibit no sign of proliferation (2). Furthermore, loss of the maternally expressed CDKN1C gene within the lesion is evidenced by the absence of immunohistochemical staining of the corresponding protein, in contrast to normal surrounding islets (19). LOH is common in tumor cells. Knudson (20) proposed a genetic model for retinoblastoma in which the inherited Rb gene mutation is described as the first hit and the tumor-restricted mutation as the second hit. This model has been enlarged to hereditary tumors. However, the two hits are necessary but may not be sufficient to trigger malignancy. In fact, the function of the gene involved in the mutational event determines the type of tumor. In CHI, LOH in 11p15 results in the loss of the wild-type SUR1 allele as well as in imbalanced expression of several imprinted genes controlling cell growth, i.e. H19, IGF-II, CDKN1C, and thereby permitting tumor growth (21). Indeed, IGF-II, a paternally expressed gene, has proliferative and antiapoptotic effects (22, 23).
To further investigate the mechanism underlying LOH in focal forms, we report three patients with severe neonatal hyperinsulinism. All patients were resistant to medical treatment and required pancreatectomy. In the first two patients, pancreatic venous sampling showed localized insulin secretion, suggesting focal CHI; whereas in the third patient, diffuse insulin secretion was suspected. The first two patients both had two distinct foci of islet cell hyperplasia, and the third patient had a very large area of islet cell hyperplasia involving the major portion of the pancreas.
| Patients and Methods |
|---|
|
|
|---|
We studied three children presenting with severe hyperinsulinemic hypoglycemia at birth. The diagnostic criteria for congenital hyperinsulinism included recurrent fasting and fed hypoglycemia (<3 mmol/liter) with inadequately elevated insulin plasma levels requiring high rates of iv glucose (>10 mg/kg·min) and increase of plasma glucose after iv glucagon injection. Furthermore, hypoglycemia responded to neither diazoxide at 15 mg/kg·d, given orally three times a day during 5 d (24), nor iv octreotide at 1050 µg/kg·d. Pancreatic venous sampling (PVS) and pancreatic arterial calcium stimulation were, until recently, the only preoperative procedures available for localizing the site of insulin secretion (25, 26). PVS allows collection of venous blood samples from the entire pancreas (head, isthmus, body, and tail) for measurements of plasma glucose, insulin, and C-peptide levels (25). Patients with a focal lesion have high plasma insulin and C-peptide levels in one or more contiguous samples and low plasma insulin and C-peptide levels in the remaining pancreatic samples. By contrast, patients with diffuse hyperinsulinism have high plasma insulin and C-peptide concentrations in all pancreatic samples (25, 26). Recently it has been shown that an [18F]fluoro-L-3,4-dihydroxyphenylalanine positron emission tomography scan permitted distinguishing between focal and diffuse forms (27).
Patient 1 was suspected of having focal insulin secretion located to the head of the pancreas after selective PVS (Fig. 1A
). An [18F]fluoro-L-3,4-dihydroxyphenylalanine positron emission tomography scan (27) further suggested that the focal lesion was situated in the head of the pancreas. Preoperative histology confirmed that the focal hyperplasia was located in the inferior one third part of the pancreatic head. A partial pancreatectomy limited to the inferior one third part of the pancreatic head was performed. Another distinct macroscopic fibrous lesion was seen on the anterior wall of the second part of the duodenum under the ampulla of Vater, measuring 10 x 10 mm. This duodenal lesion was separated 15 mm from the head of the pancreas. Histology from frozen sections of this fibrous lesion was performed during surgery and concluded to a second focus of islet cell hyperplasia. After surgery, the child still had a few hypoglycemias (3 mmol/liter). A second surgery was performed with the resection of the entire head and part of the isthmus of the pancreas. The histological analysis revealed a hyperplastic lesion in the inferior half part of the remaining pancreatic head, in continuity with the lesion resected during the first surgery. The child was cured by the second surgery, and no further pre- and postprandial hypoglycemic levels have been observed.
|
Patient 3 was suspected to have diffuse insulin secretion from the whole pancreas after PVS (Fig. 1C
). However, preoperative histology revealed a very large and ill-delimited focal lesion of adenomatous hyperplasia involving almost the whole pancreas (for histological findings, see below). A resection of a large portion of the pancreas, leaving the superior two thirds of the head, was performed. Despite this large partial pancreatectomy, the child remained severely hypoglycemic. The second surgery involved a near-total pancreatectomy after which the child was cured and no further hypoglycemic episodes were noted.
Histochemistry and molecular analyses
Samples from the normal and lesional pancreas were fixed in Bouins solution for conventional microscopy and investigation of endocrine tissue by immunohistochemical detection of proinsulin, insulin, glucagon, and somatostatin and in 4% formalin to study CDKN1C peptide (p57KIP2) expression were performed as previously described (2, 19, 28).
After informed consent was obtained for all patients, DNA studies were performed. To analyze the allele loss in the lesional pancreas, we used DNA from the hyperplastic lesion, the normal pancreas, and leukocytes from the patients and their parents. Genotyping was performed by simple fluorescent PCR assays using the following markers: D11S4046, D11S902, D11S904, D11S4152, D11S4154, D11S914, D11S1776, D11S1751, D11S935, D11S4185, D11S4083, D11S905, D11S4109, D11S1313, D11S4191, D11S987, D11S1314, and D11S901.
For all patients, genomic sequence analysis was performed on exons and intron-exons boundaries of the ABCC8 or SUR1 and KCNJ11 or KIR6.2 genes as previously defined (7, 8).
| Results |
|---|
|
|
|---|
The histological examination of pancreatic and duodenal lesions of patient 1, the pancreatic and portal lesions of patient 2, and the large focal lesion of patient 3 all revealed the characteristic histological features of focal adenomatous hyperplasia with hyperplastic islets replacing the exocrine pancreas (Fig. 2A
, patient 2). Both lesions of patients 1 and 2 were surrounded by normal pancreatic tissue. The hyperplastic islets were normally organized with ß-cells in the center and non-ß-cells at the periphery (data not shown). They were easily recognized at lower magnification by insulin immunodetection (Fig. 2B
, patient 1). According to the well-established molecular bases of this form of hyperinsulinism, the hyperplastic islets were hyperfunctional with strong proinsulin immunolabeling (data not shown) and had lost CDKN1C protein expression, whereas this protein was detected in the islets located outside the hyperplastic area (Fig. 2
, C and D, patient 1). For patient 3, histological findings were more particular with a compact adenomatous lesion involving 95% of the tail of the pancreas (Fig. 2E
) and 7080% of the body (Fig. 2F
). No hyperfunctional islets were found in the inferior half of the head and uncinate process of the pancreas (Fig. 2G
). In this case also, no CDKN1C immunostaining was found in the hyperplastic areas, but normal staining was detected in islet cell nuclei outside the lesion (data not shown).
|
Genotyping of pancreatic DNA revealed a somatic deletion of the maternal allele, involving the 11p15 region, in the focal lesions of all three patients (Fig. 3
). For patients 1 and 2, presenting with two pancreatic lesions, a diversity of centromeric deletion break points was observed, suggesting that deletions were specific for each of the focal lesions in the same patient. For patient 3, presenting with a giant pancreatic lesion, the centromeric break point of the somatic deletion was identical between two independent lesion samples, suggesting an identical somatic maternal deletion.
|
| Discussion |
|---|
|
|
|---|
We report here three patients with clinical criteria for hyperinsulinemic hypoglycemia (3) and a paternally inherited SUR1 gene mutation. These patients have unusual focal lesions: two cases of bifocal lesions (patients 1 and 2) and a giant lesion involving the major portion of the pancreas (patient 3).
In the lesions from patients 1 and 2, haplo-insufficiency for the maternal 11p15.5 region resulted from two somatic deletions. As shown by a diversity of deletion break points, these deletions occurred independently and were specific to each of the focal lesions. Therefore, LOH may be regarded as the second hit according to the model of Knudson (20) based on the paradigm of retinoblastoma, whereas the germline, inherited mutation may be regarded as the first hit. Unlike mutations of genes involved in cell growth or DNA repair, the germline SUR1 mutation is not expected to trigger the second hit more likely than expected by chance. Thus, here the rate of somatic mutation (loss of 11p15 maternal allele) is probably lower than that of the retinoblastoma gene.
In our two bifocal patients, hyperinsulinism occurred at birth, similar to four other reported children with multifocal lesions (32, 36). In addition, recurrence of hypoglycemia occurred immediately after the initial surgery, suggesting that the second lesion was already present. No postnatal recurrent lesions have been described in patients or their relatives, particularly in the paternal families of children with focal CHI. These observations support the rationale that the somatic events are of antenatal onset. Furthermore, focal forms are always associated with maternal 11p15 LOH, which probably results in overexpression of the paternally expressed IGF-II gene. Moreover, IGF-II appears to be active only at early stages of development, in particular during embryonic growth, and its expression decreases after birth (37).
In patient 3, an identical somatic maternal deletion was shown in two independent lesion samples, as shown by similar break points. In this patient with a giant form, the second hit must have occurred early during pancreas embryogenesis. The human pancreas develops from the foregut endoderm as dorsal (d 26 after conception) and ventral (next few days) outgrowths (38, 39). The dorsal pancreatic bud gives rise to the superior half of the head, body, and tail of the pancreas, whereas the ventral pancreatic bud gives rise to the inferior half of the head and the uncinate process. During embryogenesis, these buds extend into the surrounding mesenchyma, the ventral portion rotating in the clockwise direction behind the developing duodenal loop. In this location, it abuts on the dorsal primordium, with which it fuses at the end of the embryonic period (d 56 after conception). The endocrine cells in the Langerhans islets derive from gut endoderm and expression data have shown relatively mature ß-cells containing insulin-positive cells at d 52 after conception (40). Patient 3 had extended adenomatous hyperplasia in the superior half of the head, body, and tail of the pancreas, the whole of which derives from the dorsal embryonic bud. Thus, the somatic event probably occurred in the dorsal pancreatic bud. In fact, if onset of the lesion is secondary to a somatic deletion in one of the cells of the dorsal pancreatic bud, there should be more nondeleted cells in the proximal pancreas than deleted ones. During pancreas growth, the proportion of deleted cells increases because of their high proliferation rate. Indeed, in the postnatal period, ß-cells from the focal lesion were shown to have a high proliferation rate (19). As a consequence, the density of deleted cells increases in the distal part of the pancreas (the tail).
In conclusion, individual patients with focal hyperinsulinism may have more than one focal pancreatic lesion due to separate somatic maternal deletion of the 11p15 region. Furthermore, patients with plurifocal/giant lesions and those with focal CHI with small solitary lesions may follow the same two-hit model as described by Knudson. In addition, the stage of embryogenesis at which the somatic event occurs may account for the observed histological diversity, an early event resulting in a giant lesion and a somewhat later event being responsible for a small circumscribed lesion.
| Acknowledgments |
|---|
| Footnotes |
|---|
First Published Online August 1, 2006
Abbreviations: CHI, Congenital hyperinsulinism; LOH, loss of heterozygosity; PVS, pancreatic venous sampling.
Received February 23, 2006.
Accepted July 25, 2006.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
M. Eszlinger, K. Krohn, S. Hauptmann, H. Dralle, T. J. Giordano, and R. Paschke Perspectives for Improved and More Accurate Classification of Thyroid Epithelial Tumors J. Clin. Endocrinol. Metab., September 1, 2008; 93(9): 3286 - 3294. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. Barthlen, O. Blankenstein, H. Mau, M. Koch, C. Hohne, W. Mohnike, T. Eberhard, F. Fuechtner, B. Lorenz-Depiereux, and K. Mohnike Evaluation of [18F]Fluoro-L-DOPA Positron Emission Tomography-Computed Tomography for Surgery in Focal Congenital Hyperinsulinism J. Clin. Endocrinol. Metab., March 1, 2008; 93(3): 869 - 875. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |