NActa

. europathologlca

Acta Neuropathol (Berl) (1986) 71:243-250

9 Springer-Verlag1986

Aqueductal lesions in 6-aminonicotinamide-treated suckling mice* H. Aikawa 1,2, S. Kobayashi 2, and K. Suzuki 2 Division of Ultrastructural Research, National Center for Nervous, Mental and Muscular Disorders, 4-1-4, Ogawa-Higashi-Machi, Kodaira, Tokyo 187, Japan 2 Departments of Pathology (Neuropathology) and Neuroscience, the Rose F. Kennedy Center for Research in Mental Retardation and Human Development, Albert Einstein College of Medicine, Bronx, New York

Summary. Suckling mice which received a single intraperitoneal injection of 6-aminonicotinamide on the 5th postnatal day, consistently developed hydrocephalus. During the early stages of hydrocephalus (7-9 days after injection), aqueductal lesions were characterized by edematous ependymal and subependymal cells, and spongy changes in the periaqueductal area, which resulted in aqueduct stenosis. Later stages (after 20 days post-injection) showed that these edematous changes totally subsided, leaving an obliterated aqueduct which was similar to that of human congenital hydrocephalus. At the completely obliterated area, ultrastructural investigation disclosed a normallooking neuropil but no aqueductal lumen. In the remaining ependymal cell, increased intermediate filaments and lipid droplets occurred. These data suggest that acute ependymal cell degeneration during the perinatal period may result in the profile of aqueduct "agenesis" in human congenital hydrocephalus. Key words: 6-Aminonicotinamide - Aqueduct "agenesis" - Ependymal c e l l - Hydrocephalus Suckling mice

Human congenital hydrocephalus is often associated with an aqueductal lesion termed "aqueduct agenesis". It is morphologically characterized by the obliteration of the aqueductal lumen without ependymal cell or significant glial proliferation in the periaqueductal gray matter. This pathological condition is a unique feature in human congenital hydrocephalus and has been considered to be due to aqueduct malformation during the embryonic period [8]. Offprint requests to: H. Aikawa (adress see above) * Supported in part by research grants NS-03356, NS-J0803 from NINCDS, USPHS

In suckling mice receiving a single injection of 6-aminonicotinamide (6-AN) at day 5 postnatal and which developed hydrocephalus, we found the aqueductal lesions after day 20 post-injection (PI) to be identical to those of human congenital hydrocephalus. A preliminary report of this model of hydrocephalus has been published elsewhere [2]. In this communication, we will present the morphological details of aqueductal lesions in 6-AN-induced hydrocephalus in context of the pathogenesis of so-called "aqueduct agenesis".

Materials and methods Ultrastruetural study A total of 72 suckling mice (54 experimental and 18 controls) of the ICR strain (Institute of Cancer Research, Philadelphia, PA, USA) was used. The experimental group received a single intraperitoneal injection of 6-AN, 25 mg/kg of body weight, in 0.05 ml physiological saline solution on the 5th postnatal day. Control mice received saline only. Animals were killed on days 1, 3, 5, 7, 10, 20, 25 and 30 PI by cardiac perfusion with 2.5% glutaraldehyde in 0.01 M phosphate buffer (pH 7.4) for 10 min. Various parts of the brain includingseveral slices of the midbrain tissue and spinal cord were removed and post-fixed in 2% osmium tetroxide. After dehydration in a series of graded ethanol, specimens wre cleared with propylene oxide and embedded in Epon. Semithin sections were stained wit toluidine blue for the light microscope study and selected areas were trimmed for electron microscopy. Ultrathin sections were stained with uranyl acetate and lead citrate, and examined with a Hitachi H-600 electron microscope.

Immunohistoehemistry with glialfibrillary aeidic protein (GFAP) Four mice (two experimental and two controls) were killed on days 24 and 28 PI by intracardiac perfusion with cold physiological saline under anesthesia. The whole brain was immediately dissected and frozen in a dry ice/acetone bath. Ten-micrometer coronal secions of the brain were cut and mounted on glass slides, air-dried for 2 0 - 6 0 min, and stained immunohistochemically. The avidin-biotin-peroxidase method was used as the staining procedure [13]. Sections were fixed in 4%

244

H. Aikawaet al.: 6-Aminonicotinamide-inducedhydrocephalus weakness of the limbs usually developed by day 5 PI. After day 7 PI, enlargement of the head, lower positioning of the ear and postive transillumination of the head became apparent, suggesting that hydrocephalus was developing (Fig. 1) Hydrocephalic mice became inactive and in some, subdural hematoma beneath the skull was observed after day 20 PI. Affected animals survived for about a month. No abnormalities were observed in control mice. Morphologic studies

Fig. 1. A 6-aminonicotinamide(6-AN)-treated mouse on day 9 post-injection (PI). Enlargement of the head and lower placement of the ear are characteristicin this hydrocephalicmouse

Fig. 2. Coronal section of the cerebrum of the 6-AN-treated mouse on day 9 PI. Markedenlargementof the lateral ventricles is apparent, x 8 paraformaldehydefor 10 min and treated with 0.3% hydrogen in methanolfor 30 min. The slideswere then incubated serially with 3% normal goat serum for 20 min, 1:1000 anti-GFAP antiserumfor 60 rain, 1:200 biotinylatedanti-rabbitIgG (Vector Lab, CA) for 45 rain, avidin-biotin-peroxidasecomplex(Vector Lab) for 60minm, and 0.05% diaminobenzidine solution containing 0.01% hydrogen peroxide for 8 rain. As a control, instead of anti-GFAP serum, normal rabbit serum was used at the same dilution. Sliceswere counterstainedwith hematoxylin.

Results

Clinical observations 6-AN-treated mice consistently showed weight loss, diarrhea and weakness of the hindlimbs about day 3 PI. Neurological manifestations such as ataxia and

In all mice treated with 6-AN, enlargement of the lateral ventricles was consistently observed after day 7 PI, while the 4th ventricle was not dilated (Fig. 2). Early lesions in 6-AN-treated mice were characterized by edematous ependymal and subependymal cells and spongy changes of glial cells in the periaqueductal gray matter, resulting in aqueduct stenosis. Electon microscope (EM) examination revealed that edematous ependymal and glial cells had dilatation. of the perinuclear and rough endoplasmic recticulum cisternae [t]. Intracytoplasmic edema of ependymal cells was noted as early as 24 h PI and these changes were followed by spongy changes of glial cells. By day 5 PI, almost all ependymal cells lining the ventricles were swollen except fore tanycytes in the 3rd ventricle and the higher columnar cells in the subcommissural organ. Spongy changes of glial cells in the gray matter were also prominent in the brain stem, cerebellum and spinal cord. Choroid plexuses and the blood vessels were normal. Neurons showed no edematous changes. However, chromatolytic changes were observed in the brain stem nuclei and in the anterior horn cells in the lumbar spinal cord [3]. By day 7 PI, the lumen of the aqueduct was obliterated by edematous ependymal cells and spongy changes were most prominent around the periaqueductal area (Fig. 3A). EM disclosed tightly apposed edematous ependymal cells with nuclear blebs at the site of obstruction and associated cilia and microvilli were still observed (Fig. 4). In subependymal regions, the extracellular space was dilated. The junctional complexes of the ependymal cells remained intact. After day 15 PI, edema of ependymal and subependymal cells gradually subsided in the lateral and 4th ventricles, and were scarcely observed by 20 PI. The ependymal cells on the lateral ventricles were stretched and flattened without any edematous changes in the cytoplasm on day25 PI. However, in the aqueductal lesion and central canal of the spinal cord, vacuolar changes in the ependymal cells were still observed from day 15 PI (Fig. 3 B) to day 25 PI (Fig. 3 C), while spongy changes in the periaqueductal

H. Aikawa et al. : 6-Aminonicotinamide-inducedhydrocephalus

245

Fig. 3 A - D . Chronology of the aqueductal changes in the 6-AN-treated mice. One-micrometer Epon section. Toluidine blue stain. Fig. 3. A On day 7 PI, the aqueduct is obliterated with vacuolated ependymal and subependymal cells. Spongy changes in the periaqueductal gray matter are also prominent, x 140. B On day 15 PI, spongy changes in the periaqueductal gray matter subside. However, vacuolated ependymal cells still remain. Only the dorsal portion of the aqueductal lumen (arrow) is recognizable, x 140. C On day 25 PI, aqueductal lumen is sealed with spongy parenchymal tissue. Ependymal cells without intracytoplasmic edema are observed at the dorsal portion (arrow). x 160. D On day 30 PI, a cluster of surviving ependymal cells is noted at the dorsal and ventral portion (arrows). However, no lumen is seen in the middle of the aqueduct. Spongy changes and gliosis are not observed in the periaqueductal lesion, x 160. Inset. Higher magnification of the surviving ependymal cells indicated with upper arrow in D. They form a rosette at the light microscopic level, x 360

246

H. Aikawa et al.: 6-Aminonicotinamide-induced hydrocephalus

Fig. 3E, F. Immunohistochemistry of the aqueduct with GFAP. E Aqueduct of normal control mouse. The lumen is patent and reaction products are noted in the subependymal and mid-sagittal astrocytes, x t80. I~Aqueductal lesion of 6-AN-treated mouse on day 28 PI. The lumen is completelyobliterated and reaction products are prominent in the dorsal and ventral portion of the aqueduct where surviving ependymal cells are noted. In the obliterated portion, slight increase of reaction product is noted, x 180

gray matter gradually subsided. Edematous changes remained in some ependymal cells of the 3rd ventricle but the lining was well preserved. In some instances, subependymal vessels protruded into the 3rd ventricle through the ependyma after day 15 PI (Fig. 5). By day 30 PI, edematous changes of ependymal and glial cells totally subsided. Only a few clusters of ependymal cells were observed in the dorsal and ventral portion of the obliterated aqueduct (Fig. 3 D). Some remaining ependymal cells formed clusters, or rosettes (Fig. 3 D, inset). Immunohistochemistry with anti-GFAP serum revealed no significant glial proliferation in the periaqueductal gray matter, although a diffuse increase of reaction products was noted in the perivascular end-feet and perikarya and processes of astrocytes in the midbrain, in comparison to normal mice (Fig. 3 E). The remaining ependymal cells displayed strong reactivity with anti-GFAP serum, and a slight increase of reaction products was observed in the obliterated area in the 6-AN-treated mice on day 28 PI (Fig. 3F). The rostral-caudal extent of the luminal obstruction of the aqueduct ranged from 80 ot 200 gm. The central canals of the lumbar spinal cord were also totally obliterated and lacked

ependymal cells on day 30 PI (Fig. 6). Disappearance of the central canal of the cervical spinal cord was found in two mice, however, no obstruction was noted at the thoracic level. Ultrastructural study of surviving ependymal cells in the aqueduct on day 30 PI revealed abundant intermediate filaments and lipid droplets within an elongated perikaryon (Fig. 7). They possessed elongated electron-dense junctional complexes. At the obliterated portion of the aqueduct and centrl canal of the lumbar spinal cord, normal-appearing neuropils without any trace of the lumen were observed. Some dendrites showed swelling but ependymal cells were not recognizable.

Discussion

The agent 6-AN is a well-known gliotoxin and edematous changes of glial cells have been reported in many experimental animals [6, 9, 11, 12, 20, 26-29]. Dilatation of the rough endoplasmic reticulum and perinuclear cisternae is a consistent pathological change in the astrocytes and oligodendrocytes in af-

H. Aikawaet al. : 6-Aminonicotinamide-inducedhydrocephalus

247

Fig. 4. An electron micrograph of an aqueductallesion of 6-AN-treated mouse on day 7 PI. The lumen is obliteratedwith edematous ependymal cells. Those which have dilatation of perinuclear cisternae (PC) and nuclear blebs show more electron-dense chromatin pattern. In the narrowed lumen (asterisk), numerousmicrovilli and cilia are observed. Junctional complexesbetween the ependymalcells (arrows) are intact. x 8,750

fected animals. However, in previously reported studies, no pathological changes have been reported in ependymal cells. Chamberlain [5] produced congenital hydrocephalus in rats by prenatal treatment of the pregnant female with 6-AN and observed a vacuolation of neuroblasts and cellular rarefaction, but no ependymal changes were detected in his model of hydrocephalus. None of the adult animals which were treated with 6-AN developed hydrocephalus [6, 9, 11, 12, 20, 26-29]. In our previous studies of neonatal mice treated with 50 mg/kg body wt of 6-AN, we observed edematous ependymal changes and hydrocephalus with 100% incidence and pointed out that ependymal cells were one of the target cells by 6AN treatment during the neonatal period [1], One drawback of previous studies was the high mortality rate and thus, no chronic stage of hydrocephalus or

consequences of ependymal cell damage could be examined. In the current study with reduced amounts of 6-AN, we have demonstrated the aqueductal lesions which were closely similar to a congenital absence of aqueduct ("aqueduct agenesis") in mice which survived the acute stage. It is apparent that even a single administration of 6-AN causes long-lasting severe consequences. 6-AN is an antagonist of niacin and blocks the activities of NADP-dependent enzymes, such as 6phosphogluconate dehydrogenase in the pentose phosphate pathway, and subsequently greatly influences the glycolytic metabolism of the cell [9, 10, 16]. Almost selective damage of glial cells, not neurons, observed in 6-AN-treated adult animals suggest significant metabolic differences among these cell types. Also the degree of glial cell changes

248

H. Aikawa et al.: 6-Aminonicotinamide-induced hydrocephalus

Fig. 6. Lumber spinal cord of 6-AN-treated mouse on day 30 PI. No lumen or ependymal cells are seen in the central canal (arrows). One micrometer Epon section. Toluidine blue stain, x 45

Fig. 5. Ependymalceilsin the 3rd ventricleon day 15 PI. A blood vesselprotrudes into the ventricle through the edematousependymal cell layer. Macrophages are observed in the lumen and around the vessel. One micrometer Epon section. Toluidine blue stain, x 250

following 6-AN treatment was reported to be different in young-adult and aged rats [12]. Thus, the ependymal lesion by 6-AN in our neonatal mice is likely to reflect a specific metabolic state of this particular type of cells during different periods of development and maturation. Hydrocephalus could be produced by the treatment of pregnant females with various teratogenic agents [5, 15, 18, 21, 30, 31) or by postnatal treatment with kaolin [19] or silicone-oil [32]. Several mutant mice with congenital hydrocephalus have also been reported [4, 17, 22]. In these experimental hydrocephalus, diffusion of cerebro-spinal fluid into the subependymal space through the ependymal cell layer has been well recognized, but loss of ependymal cells in the aqueduct or spinal central canal have not been observed. In contrast, our model of hydrocephalus shows the most striking pathological findings in the aqueduct to be ependymal cell degeneration during the early stage and ependymal cell loss during the chronic stage. Surviving ependymal cells in the chronic aqueductal lesion appeared normal at the light microscopic level. However, at the ultrastructural level, abundant intermediate filaments and many lipid

droplets were noted within the cytoplasm. Positive immunostaining with anti-GFAP serum indicated these filaments to be related to astroglial filaments. GFAP positivity has been reported in the human ependymal cell during early period of brain development [24] and in neoplastic ependymal cells [7]. Since reactive or regenerative astrocytes in 6-AN-treated mice have been reported to have many gial filaments and lipid droplets [6], filament accumulation in surviving ependymal cells was probably a reactive or regenerative change of ependymal cells which had previously undergone edematous changes. Elongated juncitonal complexes may also characterize the regenerative ependymal cells, because these structures were similar to those of "ependymal-rosette" in ependymoma [25] or in culture [23], but rarely seen in normal ependymal cells. It is surprising, even at the ultrastructural level, that the lumen of the aqueduct and spinal central canal was totally absent in the chronic stages of 6AN-treated mice. Following exfoliation or erosion of damaged ependymal cells, cellular process might have transversed the lumen and interwoven with the aqueduct and/or central canal of the spinal cord. The pro-

H. Aikawaet al.: 6-Aminonieotinamide-inducedhydrocephalus

249

Fig. 7. An electron micrograph of the remainingependymal cells at the aqueductallesion on day 30 PI. The cells show elongated cytoplasmin which numerous intermediate filaments and lipid droplets (small arrows) are observed. No dilatation of the rough endoplasmic reticulum and perinuclear cisternae are seen. Elongated junctional complexesare prominent. The lumen (large arrows) is filledwith numerous cilia. x 5,050. Inset. Higher magnificationof the area indicated (asterisk). Numerous intermediate filaments are observed, x 35,000

trusion of subependymal blood vessels into the ventricle between the ependymal cells, as seen in Fig. 5, might suggest the disruption of junctional complexes of these cells and support the hypothesis of an extention of cellular processes across the lumen. The other example of an aqueductal lesion identical to that of 6-AN-treated mice is that described following experimental mumps virus infection in newborn rodents [14]. The presence of viral inclusions in the ependymal cells was demonstrated in the acute stage by fluorescence microscopy. Johnson and Johnson [14] emphasized the similarity of the virus-induced chronic aqueductal lesion to "aqueduct agenesis" in human congenital hydrocephalus. Active inflammatory changes or reactive gliosis in the surrounding midbrain structure were absent in these lesion. They suggest that perinatal infection by specific viruses may be a possible etiological factor for human congenital

hydrocephalus with "aqueduct agenesis". Our results with 6-AN suggest that damage of ependymal ceils by toxic agents may also be considered as another possible etiological factor for "aqueduct agenesis". There are two factors which are common to these experimental hydrocephalus associated with "aqueduct agenesis". One is the use of neonatal animals and the other is the use of the agents which selectively damage ependymal cells. Therefore, it is possible that any experiments fulfilling the above conditions may produce similar aqueductal lesions, or so-called "aqueduct agenesis". Our data suggest that "aqueduct agenesis", the main pathological change of human congenital hydrocephalus, may be a residuum of acute ependymal degeneration during the perinatal period rather than congenital or genetic aqueductal malformation during embryogenesis.

250

H. Aikawa et al.: 6-Aminonicotinamide-induced hydrocephalus

Acknowledgements. We thank Dr. Lawrence F. Eng for provid-

15. Kalter H (1963) Experimental mammalian teratogenesis. A study of galactoflavin-induced hydrocephalus in mice. J Morphol 112:303-317 16. Kauffman FC, Johnson EC (1974) Cerebral energy reserves and glycolysis in neural tissue of 6-aminonicotinamidetreated mice. J Neurobiol 5: 3 7 9 - 392 17. Kohn DF, Chinookoswong N, Chou SM (1984) Congenital hydrocephalus. Am J Pathol 114:184-185 18. Newberne PM, O'Dell DL (1958) Histopathology of hydrocephalus resulting from a deficiency of vit. B~2. Proc Soc Exp Biol Med 97: 6 2 - 65 19. Ogata J, Hochwald GM, Cravioto H, Ransohoff J (1972) Distribution of intraventricular horseradish peroxidase in normal and hydrocephalic cat brains. J Neuropathol Exp Neurol 31:454-463 20. O'Sullivan BM, Blakemore WF (1980) Acute nicotinamide deficiency in the pig induced by 6-aminonicotinamide. Vet Pathol 17: 748 - 758 21. Overholster MD, Whitley JR, O'Dell BL, Hogan AG (1954) The ventricular system in hydrocephalic rat brains produced by a deficiency of vitamin Blz or folic acid in the maternal diet. Anat Ree 120:917-933 22. Raimondi AJ, Clark SJ, McLone DG (1976) Pathogenesis of aqueductal occlusion in congenital routine hydrocephalus. J Neurosurg 45: 6 6 - 77 23. Raine CS, Bornstein MB (1974) Unusual profiles in organotypic cultures of central nervous tissue. J Neurocytol 3:313-325 24. Roessmann U, Velasco ME, Sindely SD, Gambetti P (1980) Glial fibrillary acidic protein (GEAP) in ependymal cells during development. An immunocytochemical study. Brain Res 200:13-21 25. Russell DS, Rubinstein LJ (1977) Ependymomas. In: Pathology of tumours of the nervous system, 2nd edn. Arnold, London, pp 204-212 26. Sasaki S (1982) Brain edema and gliopathy induced by 6amino-nicotinamide intoxication in the central nervous system of rats. Am J Vet Res 43 : 1691 - 1695 27. Schneider H, Cervos-Navarro J (1974) Acute gliopathy in spinal cord and brain stem induced by 6-aminonicotinamide. Acta Neuropathol (Berl) 27 : 11 - 23 28. Schochet SS Jr (1970) Pathogenesis of 6-aminonicotinamide neurotoxicity. VIth International Congress of Neuropathology Proceeding, Masson, Paris, pp 89 [abstr] 29. Sotelo C, Rio JP (1980) Cerebellar malformation obtained in rats by early postnatal treatment with 6-aminonicotinamide. Role of neuro-glia interactions in cerebellar development. Neuroscience 5:1737-1759 30. Stempak JG (1964) Etiology of trypan blue-induced antenatal hydrocephalus in the albino rat. Anat Rec 148:561 571 31. Stempak JG (1965) Etiology of antenatal hydrocephalus induced by folic acid deficiency in the albino rat. Anat Rec 151:287-296 32. Weller RO, Wisniewski H, Shulman K, Terry RD (1972) Experimental hydrocephalus in young dogs: histological and ultrastructural study of the brain tissue damage. J Neuropathol Exp Neurol 30: 613 -- 626

ing us with anti-serum against GFAP and Dr. Cedric S. Raine for his critical reading of this manuscript.

References 1. Aikawa H, Suzuki K, Ito N, Iwasaki Y, Nonaka I (1984) 6-amino-nicotinamide-induced hydrocephalus in suckling mice. J Neuropathol Exp Neuro143 : 511 - 521 2. Aikawa H, Suzuki K (1985) Aqueduct stenosis induced by a single injection of antivitamin. Dev Brain Res 22:284-287 3. Aikawa H, Suzuki K (1986) Lesions in the skin, intestine and central nervous system induced by an antimetabolite of niacin. Am J Pathol 122: 335 - 342 4. Borit A, Sidman RL (1972) New mutant mouse with communicating hydrocephalus and secondary aqueductal stenosis. Acta Neuropathol (Berl) 21 : 3 1 6 - 331 5. Chamberlain JG (1970) Early neurovascular abnormalities underlying 6-aminonicotinamide (6-AN)-induced congenital hydrocephalus in rats. Teratology 3:377-387 6. Chui E, Garcia JH (1979) Pathogenesis of 6-aminonicotinamide neurotoxicity: new structural analysis. In: Zimmermann HM (ed) Progress in neuropathology, vol 4. Raven Press, New York, pp 3 4 1 - 359 7. Duffy PE, Graf L, Huang Y-Y, Rapport MM (1979) Glial fibrillary acidic protein in ependymomas other brain tumors. Distribution, diagnostic criteria, and relation to formation of processes. J Neurol Sci 40:133 - 146 8. Friede RL (1975) Hydrocephalus-special pathology. In: Developmental neuropathology. Springer, New York, pp 214-229 9. Griffiths IR, Kelly PAT, Grome JJ (1981) Glucose utilization in the central nervous system in the acute gliopathy due to 6-aminonicotinamide. Lab Invest 44:547-552 10. Herken H, Lange K, Kolbe H, Keller K (1974) Antimetabolic action on the pentose phosphate pathway in the central nervous system induced by 6-aminonicotinamide. In: Ganazzani E, Herken H (eds) Central nervous system. Studies on metabolic regulation and function. Springer, New York, pp 41 - 5 4 11. Horita N, Oyanagi S, Ishii T, Izumiyama Y (1978) Ultrastructure of 6-aminonicotinamide (6-AN)-induced lesions in the central nervous system of the rats. I. Chromatolysis and other lesions in the cervical cord. Acta Neuropathol (Berl) 44:111-119 12. Horita N, Ishii T, Izumiyama Y (1981) Ultrastructure of 6amino-nicotinamide (6-AN)-induced lesions in the central nervous system of the rats. III. Alterations of the spinal gray matter lesion with aging. Acta Neuropathol (Berl) 53 : 2 2 7 235 13. Hsu SM, Raine L, Fanger H (1981) Use of avidin-biotinperoxidase complex (ABC) in immunoperoxidase technique: a comparison between ABC and unlabeled antibody (PAP) procedures. J Histochem Cytochem 29: 577- 580 14. Johnson RT, Johnson KP (1968) Hydrocephalus following viral infection: the pathology of aqueductal stenosis developing after experimental mumps virus infection. J Neuropathol Exp Neurol 27:591-606

Received June 27, 1986/Accepted July 14, 1986