Pancreatic Islet Abnormalities in Sudden Infant Death Syndrome

Transcription

1 The Japanese Society of Pathology Pancreatic Islet Abnormalities in Sudden Infant Death Syndrome Qualitative and Quantitative Analyses of 15 Cases Masa no ri H isao ka I, Joji Ha ra ta ke', Yasu hiro Na ka mura2, and Yohko Itoh3 The pancreata of 15 autopsy cases of sudden infant death syndrome (SIDS) and those of 14 age-matched controls were examined qualitatively and quantitatively to re-evaluate the relationship between pancreatic islet abnormalities and sudden death in infancy. Histopathologically, a diffuse or focal form of nesidioblastosis and septal islets were frequently observed in the pancreata of both groups. Endocrine cell dysplasia was found only in 2 infants who had died of SIDS. Quantitatively, there was little difference of islet cell composition between the SlDS cases and the controls. A relatively high proportion of islet cell area to total pancreatic tissue area was demonstrated in the SlDS group ( % in the pancreatic head ; % in the pancreatic body to tail) in comparison with the controls (5.32k 1.77% ; 5.63?1.6%). Although nesidioblastosis and septal islets were considered to be within the limits of normal variation during pancreatic development, endocrine cell dysplasia and quantitatively unusual proliferation of the pancreatic endocrine tissue suggest the possibility that abnormalities in the endocrine pancreas may be causally related to sudden death in infancy. Acta Pathol Jpn 42: , Key words : Sudden infant death, Pancreatic islet, Nesidioblastosis, Quantitative analysis thorough necropsy examination fails to reveal a cause of death(1). Although the incidence of SlDS in Japan, 1.2/1, live births, is lower than the figure of about 3/ 1, in Europe (2, 3), thorough investigation and appropriate measures for prevention of SlDS are urgently required. Although previous extensive studies have considered various plausible or possible causes of death in SIDS, including asphyxia, infection, abnormal respiratory physiology, and metabolic disorder (4-7), 6-7% of deaths are still unexplained (8, 9). Polak et a/. and some other investigators have found islet cell hyperplasia or nesidioblastosis in the pancreata of some SlDS cases, and suggested that hyperinsulinemic hypoglycemia due to abnormal 6-cell (insulinsecreting cell) proliferation could be related to the cause of sudden death (1-13). However, these sporadic case studies seem to be insufficient for establishing the cause of sudden death in infancy. In order to confirm the relationship between such islet cell abnormalities and SIDS, we examined qualitatively and quantitatively the pancreatic endocrine tissue of 15 infants who had died of SlDS and 14 age-matched co nt ro Is. Sudden infant death syndrome (SIDS) has been generally defined as the death of an infant or a young child, which is unexpected on the basis of history and where a - Received August 3, Accepted for publication October 12, 'Department of Pathology and Oncology, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, *Department of Pathology, St. Mary's Hospital, Kurume, and 3Department of Legal Medicine, Kurume University School of Medicine, Kurume. Mailing address : Masanori Hisaoka, M.D., Department of Pathology and Oncology, School of Medicine, University of Occupational and Environmental Health, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu 87, Japan. SlDS cases MATERIALS AND METHODS Fifteen infants who had died of SlDS at 11 days to 7 months after birth were carefully selected from among infant autopsy cases seen at our hospitals. All of these cases satisfied the criteria of SIDS; the deaths were sudden and unexpected, without a history of significant abnormalities, i.e. major congenital abnormalities, extreme prematurity, severe infection or serious trauma, and thorough autopsy had failed to reveal any pathological finding considered to be related to sudden death. During life, clinical laboratory tests had not been per-

2 Acta Pathologica Japonica 42 (12): formed on most of the infants. Little abnormality was detected in the few examined, and all of the deaths of the infants selected were clinically "unexplained". In addition, these cases also conformed to the revised definition of SlDS drafted recently (14). Fourteen age-matched controls (aged 8 days to 8 months) were also selected from among infant autopsy cases seen at our hospitals. All of them were confirmed to have had a clear cause of death; sepsis or severe respiratory infection (5 cases), congenital cardiac anomalies (4 cases), hydrocephalus (2 cases), cystic dysplastic kidney (1 case), bronchopulmonary dysplasia (1 case) and acute monocytic leukemia (1 case). The duration of their clinical courses was generally short (mean, 37.4 days). Although exact clinical data on pancreatic hormones were unavailable, these controls showed no evidence of organic endocrine or metabolic diseases. No persistent abnormalities of glucose homeostasis were detected during their clinical courses. Corticosteroids were not administered except for two infants. Pancreatic sampling and immunohistochemistry Samples used in this study were obtained from formalin-fixed pancreatic tissues of subjects in both groups. Tissue specimens were cut from two discrete regions: the pancreatic head, and the body to tail region. Because of some sampling limitations, the total number of specimens obtained were 9 pancreatic head regions and 15 body to tail regions from SlDS cases (specimens from both regions were taken from each of 9 cases, and only one specimen of the body to tail region was taken from each of 6 cases), and 14 head regions and 11 body to tail regions from controls (specimens from both regions were taken from each of 11 controls, and only one specimen of the head region was taken from each of 3 controls). Each specimen was embedded in paraffin and routinely processed for histological examination. In addition to HE staining, immunoperoxidase staining for insulin, glucagon, somatostatin and pancreatic polypeptide (PP) using the avidin-biotin complex method was performed respectively on serial sections of each tissue specimen. Furthermore, immunostaining with combined antisera against these four pancreatic hormones was done on other sections. The antisera employed were rabbit polyclonal ones obtained commercially from DAKO JAPAN Co., Ltd. (Kyoto, Japan). The PP-cell-rich region was identified in the pancreatic heads of 7 SlDS cases (47%) and 12 controls (86%). Quantitative analysis The relative proportion of the endocrine tissue area to the total parenchymal area of the pancreas in each tissue section was calculated using an IBAS-2 automatic image-analysis system (KONTRON Bildanalyse GmbH, Eching, Germany). Optical views of the sections immunostained with combined antisera against the four pancreatic hormones were projected through a microscope onto square fields of a TV monitor at a magnification of x 185. Contours of the pancreatic islets (including endocrine cells clusters composed of 2 or more endocrine cells) in each field were traced on the monitor, and the total area of enclosed islets or cell clusters was summed automatically by the system. The relative proportion of the islet cell area was estimated by dividing the total summed area enclosed by the total parenchymal area, which was also calculated automatically. Twenty consecutive fields for each section were scanned, covering most of the parenchymal tissue area of the section. Although scanned total parenchymal areas were not markedly different among the subjects (about 17.6 mm2 per region), the number of pancreatic islets examined varied considerably among the subjects (about 3 to 15 islets per region). The area of isolated or single endocrine cells, observed sparsely in almost all of the specimens, was not evaluated in this analysis because of certain technical limitations in our method and the low contribution of such cell areas to the total measured endocrine tissue area. The populations of the four types of endocrine cells were also evaluated. Positively immunostained endocrine cells in 2 consecutive fields were counted microscopically at a magnification of x2 in each immunostained serial tissue section, and the relative proportion of each cell type was estimated in each pancreatic region. In total, about 1,2 to 5,5 endocrine cells per region were identified and counted. In the cases showing a PP-rich region in the pancreatic head, the cell composition was evaluated respectively in PP-rich and PP-poor regions in the same section. Differences between means were assessed by Welch's t test. Pat ho log ica I findings RESULTS Microscopic examination of the pancreata from SlDS cases and controls revealed a smaller-sized lobular arrangement of acinar tissue and scattered pancreatic islets or small packets of endocrine cells within or outside the pancreatic lobules. Pancreatic islets varied in size and shape not only among the SlDS cases but also among the selected areas, even in a single case. A focal or diffuse form of unusual islet cell proliferation (focal or diffuse nesidioblastosis) was a common finding

3 8 72 Pancreatic Islet Abnormalities in SlDS (Hisaoka et a/.) Figure 1. Photomicrographs of various islet abnormalities observed in the pancreata of SlDS cases and controls. a : Focal nesidioblastosis (SIDS, pancreatic tail, immunoperoxidase staining with combined antisera). b: A septal islet (control, pancreatic head, HE). c: Endocrine cell dysplasia (SIDS, pancreatic tail, immunoperoxidase staining with combined antisera). d : Islet cell hypertrophy (arrows) (SIDS, pancreatic head, HE). in both groups. Each focal lesion was composed of large islets and partly confluent endocrine cell clusters in restricted areas, being evident in 6 SlDS cases (4%) and 5 controls (36%) (Fig. la). In each diffuse lesion, a large number of various-sized islets and small packets of endocrine cells were distributed evenly throughout the pancreatic tissue, being evident in 3 SlDS cases (2%) and 2 controls (14%). Septal islets, i.e. islets lying entirely within the septal connective tissue stroma of the pancreas, were constantly observed in both groups, and the sizes of the islets varied considerably (Fig. lb). This was evident in 13 SlDS cases (87%) and 13 controls (93%). lmmunohistochemically, the islets observed were composed of an admixture of four types of endocrine cells (B, A (glucagon-secreting), D (somatostatin-secreting) and Table 1. Pancreatic Islet Abnormalities in SlDS and Control Infants Islet abnormalities SlDS Controls (n=15) (n=14) Nesidio blastosis focal 6 (4%) 5 (36%) diffuse 3 ~2%) 2 (14%) Septal islets 13 (87%) 13 (93%) Endocrine cell dvsalasia 2 (13%) PP (PP-secreting) cells) with a relative predominance of B cells. The distribution of these cells within the islets showed no distinct abnormalities. Scattered endocrine cells or cell clusters within the acinar tissue were also immunostained positively for various types of hormones. Although a small number of immunostained endocrine cells were identified in the ductal epithelium in most of

4 Acta Pathologica Japonica 42 (12): Table 2. Relative Proportions (mean%) of Pancreatic Endocrine Cells and Areas in SlDS Cases and Controls B cell % Endocrine cells A cell D cell PP cell Total endocrine cells counted % Endocrine cell area SlDS (n=15) head region (n=9) 8.46 k4.9 PP-rich (n=7) f6.25 k2.1 f4.66 f8.5 f164 P P- poor (n=9) f3.65 f4.68 i5.13 f.61 k295 Body to tail region 8.66k4.23 a (n=15) k8.23 k3.65 f f1192 Controls (n=14) head region (n=14) PP-rich (n=12) k7.62 t7. k3.9 k13.57 t147 PP-poor (n=14) f f7.91 f.65 f1484 body to tail region (n=ll) k11.71 f8.77 f9.17 k.58 k1418 Each value indicates mean+sd; a Difference statistically significant, P< k k1.6 a 3 h g m 2? m - 2 al u C al.- L V al - 1 c._ r Q? a NS Pc.3 m n i n.! control SlDS control SlDS head body to tail Figure 2. The proportion of endocrine cell area to total pan. creatic tissue area in SlDS cases and controls. the specimens, no well formed ductulo-insular complex was noted in either group. Endocrine cell dysplasia, i.e. an apparent loss of the normal centrilobular aggregation of large islets with increased numbers of small islet cell clusters arranged in a haphazard manner, irregularity of islet contours and prominent islet cell nuclear hypertrophy, was found only in 2 cases (13%) of SlDS (Figs. lc, d). Although some of the hypertrophic islet cells in endocrine cell dysplasia were positively immunostained for insulin, others were stained for glucagon or somatostatin. The pancreatic endocrine cell abnormalities observed in this study are listed in Table 1. Quantitative analysis In SlDS cases, the mean proportions of the islet cell area were % (mean2sd) in the pancreatic head and % in the body to tail. These values were relatively higher than those of controls ( % in the head, P>.5; s in the body to tail, P<.3) (Table 2). In addition, there were 6 SlDS cases with prominently large islet cell areas above 1%. The proportion of the islet cell area in the body to tail region was slightly higher than that in the head region in both groups. The individual results regarding the proportion of islet cell area are shown diagrammatically in Fig. 2. In the PP-poor region of the pancreatic head and in the body to tail, B cells were predominant with a proportion of about 56-57%, followed by D cells at about 23-24%. The PP-rich region in the pancreatic head showed a PPcell predominance at about 53%, followed by B cells at about 26%. There was no significant difference in the mean cell composition of the islets in the head or the body to tail, between the SlDS cases and the controls (Table 2). DISCUSSION Morphologically, unusual endocrine cell proliferation in the pancreas has been observed in some SlDS cases, and hypoglycemia due to excess insulin secretion by these proliferated endocrine cells has been implicated as a possible cause of sudden death in infancy(1o-13). A variety of morphologic abnormalities in the endocrine

5 8 74 Pancreatic Islet Abnormalities in SlDS (Hisaoka et a/.) pancreas, such as nesidioblastosis (synonyms ; islet cell hyperplasia, adenomatosis), islet cell hypertrophy, endocrine cell dysplasia, ductulo-insular complex and septal islets, have been considered to be common and essential pathologic changes observed in the pancreas of neonates with hyperinsulinemic hypoglycemia (15-2). In order to confirm a causal relationship between inappropriate insulin secretion and sudden death in infancy, it is necessary to evaluate the volume of pancreatic endocrine tissue and/or serum levels of pancreatic hormones as well as the histology of the endocrine pancreas in affected individuals. However, it is rather difficult to obtain clinical data in view of the fact that affected infants die suddenly and unexpectedly, and we have been unable to find any pancreatic hormonal assay studies in SIDS. In addition, morphometric analysis of pancreatic endocrine tissue from infants with SlDS has not been performed except for only two studies, one of which supported the hypothesis with the evidence of increased pancreatic endocrine tissue (endocrine tissue; 1.8% of mean total area) whereas the other obtained negative findings (data not shown) (13, 21). Details of the methodology used in those studies were not well documented and were probably different from ours. In the present study, diffuse or focal nesidioblastosis and septal islets were observed in SlDS cases as well as controls at almost the same frequency. Hence, these findings were presumed to be within the normal range of variation for the endocrine pancreas in infancy. Recently, some investigators have suggested that these morphologic findings are not specific, but rather normal or indicative of persistent development of the pancreas after birth (22-24). Jaffe et a/. proposed a related term endocrine cell dysplasia to describe the morphologic islet cell abnormalities observed in the pancreata of hypoglycemic patients with infantile hyperinsulinemia, providing further subtle definitions but no statement about the condition being causally related to hypoglycemia (24, 25). In our study, only two SlDS cases showed consistent features of endocrine cell dysplasia, including islet cell hypertrophy, which has been considered to be of diagnostic significance in hyperinsulinism (26-28). Although hyperinsulinism might have been related to the sudden deaths of the infants with such islet abnormalities, we feel it is still difficult to determine such a relationship based only on histologic features. In the present quantitative analysis, there was little difference in the proportions of various islet cell types among SlDS cases, controls and other neonates or infants investigated previously (29). Our study failed to demonstrate any significant B-cell predominance or an apparent decrease in the number of D cells indicative of abnormal insulin secretion. Some other studies have also suggested that there was no apparent difference in islet cell composition between infants with hyperinsulinemic hypoglycemia and age-matched controls (22, 23). Nevertheless, the investigators mentioned that they could not rule out the possibility that unknown, poorly defined or histologically undetermined abnormalities (such as small insulin-secreting tumors) might have played a role in the pathogenesis of hyperinsulinism (22, 23). Our quantitative analysis revealed a relative increase of endocrine cell area in SlDS cases, especially in the body to tail region of the pancreas, in comparison with the controls. In addition, some cases showed a prominently large endocrine cell area exceeding 1% of the total pancreatic tissue area. The proportions of the islet cell areas in SlDS cases were also higher than those in control infants in another quantitative study of the endocrine pancreas in infants of diabetic mothers, also using an automatic image-analyzer (PP-rich region; %, PP-poor region; %) (29). The mean values of islet cell areas in our controls corresponded closely to those of the latter study. Hence, the present findings might indicate unusual proliferation of the endocrine pancreas, and it is suggested that the unusually proliferated pancreatic endocrine cells might play a role in abnormal regulation or modulation of the pancreatic endocrine system, which in turn might be related to the sudden deaths of some infants. However, the question of whether or not the causative mechanism of SlDS is a single entity is still open. In cases which show no abnormalities of the endocrine pancreas qualitatively or quantitatively, other mechanisms proposed previously, such as abnormal respiratory physiology, infections or metabolic disorders (4-7), should be considered. We also speculate that the cases of SIDS may be composed of the infants really lacking any detectable abnormalities and those having overlooked or underestimated abnormalities, which may have an undetermined causal relationship with sudden death in infancy. The selection of controls might be a delicate problem in the quantitation of islet cell composition, since medical treatments, such as corticosteroid or glucose infusion, may affect the cellular composition of the pancreatic islets(3). A number of untreated and/or normoglycemic infants should be collected as controls for this sort of analysis, although most of our controls were untreated with corticosteroids. In addition, the present SlDS cases lacked clinical information or symptoms suggestive of endocrine abnormalities such as hyperinsulinism. A clinical examination, including serum levels of insulin and glucose in the early neonatal period, is

6 Acta Pathologica Japonica 42 (12) : required to support the hypothesis and avoid any future critical condition if a relationship between SIDS and pancreatic endocrine tissue abnormalities does, in fact, exist. Acknowledgements : The authors are grateful to Prof. Minoru Morimatsu and Prof. Hiroshi Kimura of the Departments of Pathology and Legal Medicine, Kurume University School of Medicine, for supplying the pancreatic materials. REFERENCES 1. Beckwith JB. Observations on the pathological anatomy of the sudden infant death syndrome. In Bergman AB, Beckwith JB, and Ray CG, eds. Proceedings of the Second International Conference on Cause of Sudden Death in Infants. University of Washington Press, Seattle, Washington, 197: Berry PJ and Keeling JW. The investigation of sudden unexpected death in infancy. In Anthony PP and MacSween RNM, eds. Recent Advances in Histopathology. No. 14. Churchill Livingstone, New York, 1989 : Shiono H, Tabata N, Fujiwara M, et a/. Sudden infant death syndrome in Japan. Am J Forensic Med Pathol 9 : 5-8, Williams AL, Uren EC, and Bretherton L. Respiratory viruses and sudden infant death. Br Med J 288: , Southall DP. Role of apnea in the sudden infant death syndrome: A personal view. Pediatrics 81 : 73-84, Naeya RL. The sudden infant death syndrome: A view of recent advances. Arch Pathol Lab Med 11 : , Emery JL, Howat AJ, Variend S, and Vawter GF. Investigation of inborn errors of metabolism in unexpected infant deaths. Lancet 2: 29-31, Knowelden J, Keeling J, and Nicholl JP. A Multicentre Study of Post-Neonatal Mortality. Medical Care Research Unit, University of Scheffield, London, Norvenius SG. Sudden infant death syndrome in Sweden in and Acta Pediatr Scand [SUPPI] 333: 1-138, Aynsley-Green A, Polak JM, Keeling J, et a/. Averted sudden neonatal death due to pancreatic nesidioblastosis. Lancet 1 : , Cox JN, Guelpa G, and Terrapon M. Islet-cell hyperplasia and sudden infant death. Lancet 2 : , Polak JM, Aynsley-Green A, Bloom SR, and Wigglesworth JS. Nesidioblastosis as a cause of sudden neonatal death. Scand J Gastroenterol 13 [Suppl 491 : 143, Polak JM and Wigglesworth JS. Islet-cell hyperplasia and sudden infant death. Lancet 2: , Willinger M, James LS, and Catz C. Definig the sudden infant death syndrome (SIDS): Deliberations of an expert panel convened by the National Institute of Child Health and Human Development. Pediatr Pathol 11 : , Brown RE and Young RB. A possible role for the exocrine pancreas in the pathogenesis of neonatal leucine-sensitive hypoglycemia. Am J Dig Dis 15 : 65-72, Dahms BB, Landing BH, Blaskovics M, and Roe TF. Nesidioblastosis and other islet cell abnormalities in hyperinsulinemic hypoglycemia of childhood. Hum Pathol 11 : , Heitz PU, Kloppel G, Hacki WH, ef a/. Nesidioblastosis : The pathologic basis of persistent hyperin. sulinemic hypoglycemia in infants : Morphologic and quantitative analysis of seven cases based on specific immunostaining and electron microscopy. Diabetes 26 : , Yakovac WC, Baker L, and Hummeler K. Beta cell nesidioblastosis in idiopathic hypoglycemia of infancy. J Pediatr 79: , Grampa G, Garganitini L, Grigolato PG, and Chiumello G. Hypoglycemia in infancy caused by beta cell nesidioblastosis. Am J Dis Child 128: , Thomas CG Jr, Underwood LE, Carney CN, et a/. Neonatal and infantile hypoglycemia due to insulin excess. Ann Surg 185: , Loo SWH, Kozakewich HPW, WaId RA, and Vawter GF. Sudden infant death syndrome and nesidioblastosis : No evidence for pathologic relationship. Lab Invest 46 : 1, Gould VE, Memoli VA, Dardi LE, and Gould NS. Nesidiodysplasia and nesidioblastosis of infancy : Structural and functional correlation with the syndrome of hyperinsulinemic hypoglycemia. Pediatr Pathol 1 : 7-31, Goudswaard WB, Houthoff HJ, Koudstaal J, and Zwierstra RP. Nesidioblastosis and endocrine hyperplasia of the pancreas: A secondary phenomenon. Hum Pathol 17 : 46-53, Jaffe R, Hashida Y, and Yunis EJ. The endocrine pancreas of the neonate and infant. In Rosenberg HS Pnd Bernstein J, eds. Perspectives in Pediatric Pathology, Vol. 7. Masson, New York, 1982 : Jaffe R, Hashida Y, and Yunis EJ. Pancreatic pathology in hyperinsulinemic hypoglycemia of infancy. Lab Invest 42 : , Rahier J, Falt K, Muntefering H, et a/. The basic structural lesion of persistent neonatal hypoglycemia with hyperinsulinism : Deficiency of pancreatic D cells or hyperactivity of B cells? Diabetologia 26: , Rahier J, Wallon J, and Henquin J-C. Cell population in the endocrine pancreas of human neonates and infants. Diabetologia 2 : , Goossens A, Gepts W, Saudubray JM, et a/. Diffuse and focal nesidioblastosis : A clinicopathological study of 24 patients with persistent neonatal hyperinsulinemic hypoglycemia. Am J Surg Pathol 13 : , Milner RDG, Wirdnam PK, and Tsanakas J. Quantitative morphology of B, A, D, and PP cells in infants of diabetic mothers. Diabetes 3: , Barresi G, lnferrera C, Gemelli M, and Luca de F. Morphological aspects of the endocrine pancreas in a predonisolone-treated infant. Histopathology 1 : 21-28, 1977.