On the pathology of experimental hydrocephalus

Brain Research, 95 (1975) 343-350 © ElsevierScientificPublishingCompany,Amsterdam- Printed in The Netherlands

343

ON THE PATHOLOGY OF EXPERIMENTAL HYDROCEPHALUS

R. N Y B E R G - H A N S E N , A. TORVIK AND R. B H A T I A

Laboratory of Neuropathology, Ullevdl Hospital, Oslo (Norway)

SUMMARY

Acute hydrocephalus was produced in newborn rabbits by injection of kaolin into the cisterna magna. The light microscopic changes which occurred in the ependyma and periventricular brain tissue were studied. Some animals also received intraventricular injection of Evans blue albumin (EBA) at various times after the kaolin injection to study the permeability of the ependyma. There was a progressive dilatation of the lateral ventricles from the second day after the kaolin injection. Marked hydrocephalus was seen after 2 weeks. The white matter of the cerebral hemispheres showed increasing reduction in volume with the degree of hydrocephalus. Neither destruction of brain tissue nor macrophage response or inflammation were seen. The ependyma adapted remarkably well to the increased intraventricular pressure by extensive flattening and stretching. No convincing breaks or ruptures were seen. There was a patchy spongy zone beneath the ependyma, probably indicating oedema of the periventricular white matter due to transventricular absorption of the cerebrospinal fluid (CSF). Denudement of the ependymal lining is not necessary for the concept of transventricular flow of CSF. No difference was seen in the penetration of EBA into the periventricular tissue between hydrocephalic and control animals. The reduction of the cerebral mantle thickness was probably caused by simple pressure atrophy. This indicates that the morphological changes may to a large extent be reversible if the hydrocephalus is properly treated within reasonable time. The role of the morphological changes in the pathophysiology of hydrocephalus if briefly commented upon in relation to certain aspects of human hydrocephalus.

INTRODUCTION

Controversy still exists about the exact mechanisms of the production and

344 absorption of the cerebrospinal fluid (CSF) in normal as well as in hydrocephalic brains. Although the choroid plexus of the ventricular system normally is regarded as the major site of CSF production, minor subsidiary sources cannot be ruled out. The CSF is normally produced by the choroid plexus by a secretion process, mainly as a bulk fluid at a rate of 0.35 ml/min in children and 0.40 ml/min in adults la. The formation seems to a large extent to be independent of the intraventricular pressureS, 11. The arachnoid villi along the superior sagittal and transverse sinus appear to be the major sites of CSF absorption, which is linearly related to the intracranial pressure. The CSF normally circulates from the ventricles through the subarachnoid space to the parasagittal region of the cerebral convexity, as is reflected by the radiopharmaceutical movement pattern encountered with cisternography. During hydrocephalus when the normal circulation pathways are blocked, alternative routes of absorption become important, especially as only a minor decrease in CSF production has been observed in such cases. In hydrocephalic children the rate of formation has been estimated to 0.30 ml/min, against 0.35 ml/min in normals 11. The CSF production thus seems to occur at a rate which varies little with the intraventricular pressure both in normal and hydrocephalic brainsS, 11. With regard to alternative routes of CSF absorption during hydrocephalus, several studies have indicated that a transventricular absorption occurs across the ventricular ependymal lining2, s,9,1v, and it has been assumed that the CSF escapes into the periventricular brain tissue through tears in the ependyma 9,~2,~3,~5 caused by the increased intraventricular pressure. The present communication directs attention to the morphological changes which occur during hydrocephalus and discusses their role in the pathophysiology of this condition. The experiments were designed to study the light microscopic changes in the ependyma and periventricular brain tissue after acute obstruction of the CSF pathways in newborn rabbits. METHODS

Acute hydrocephalus was produced in 10-day-old rabbits by injection of 0.1-0.2 ml of a thick suspension of kaolin into the cisterna magna. The animals were killed from 2 to 30 days after the injections and the brains fixed by perfusion with the modified SUSA fixative4, embedded in paraffin and cut in serial sections at 5/~m. Every 20th section was mounted and stained with haematoxylin and eosin, additional sections with various other methods for myelin sheaths, axons and astrocyte fibres. Control animals of the same age were used. (For details, see ref. 20.) RESULTS A N D DISCUSSION

The kaolin filled the entire subarachnoid space around the brain stem in all cases. An acute inflammatory reaction was seen in the meninges during the first days and then gradually subsided. The subarachnoid space appeared to be completely occluded in all animals. There were no ischaemic lesions of the brain stem.

345

Fig. 1. Photograph showing the dilatation of the ventricular system in a hydrocephalic animal killed 3 weeks after the injection of kaolin into the cisterna magna. The ventricles are normally slit-like.

The lateral ventricles showed a progressive dilatation from the second day after the kaolin injection. Most animals were markedly hydrocephalic after 2 or 3 weeks (Fig. 1). Enlargement of the third and fourth ventricle was only seen in cases kept alive for 2 weeks or longer. There was a considerable thinning of the cerebral hemisphere, increasing progressively with the degree of hydrocephalus (Fig. 1). The reduction o f the mantle thickness affected mainly the white matter, and only to a minor extent the neuropil of the cerebral cortex. Neither inflammatory response nor loss of cells or astrocytic proliferation were seen. The volume reduction thus seems to be the result of a simple pressure atrophy rather than destruction of neural tissue. The ependymal lining of the ventricles was flattened and stretched in all hydrocephalic animals in contrast to the relatively tall and heavily ciliated normal ependyma (Figs. 2 and 3). These changes were most prominent along the lateral wall of the lateral ventricles where small gaps (Fig. 3) between the intact individual cells could be observed. The ependymal flattening was much less pronounced over the grey matter along the medial wall of the lateral ventricle. Using serial sections which are indispensable for this purpose, we did not

347 observe true breaks or ruptures of the ependyma. This is in contrast to other reports on the pathology of experimental hydrocephalus~,6,9,15, al. However, artificial tears of the ependyma which invariably occur even in well fixed tissue were also observed in our material. Ependymal tears due to the increased intraventricular pressure have been assumed to allow the escape of CSF into the periventricular tissueg, 12,14,15. Transventricular bulk flow of CSF undoubtedly occurs during hydrocephalusZ,S,9,17,19. However, the fact that protein molecules may pass freely between intact ependymal cells even in normal brains 3,19 indicates that ependymal defects are not necessary for the passage of CSF into the periventricular brain tissue in hydrocephalic brains. The present experiments are in accordance with this line of reasoning, and provided no support for the concept of CSF movement through simple breaks or ruptures of the ependyma during hydrocephalus. However, the small gaps which were observed between the intact individual ependymal cells (Fig. 3) probably facilitate the transventricular absorption of the CSF. Thus, in an autoradiographic study with radiopharmaceutical 13qodine serum albumin (RISA) in chronic communicating hydrocephalus in the dog, Strecker et al. 19 observed migration of the labelled albumin between the individual ependymal cells. However, these observations need electron microscopic confirmation. It should be added that the penetration of protein molecules between ependymal cells in normal brains does not appear to represent bulk flow of CSF into the brain parenchyma. Beneath the ependyma a consistent patchy sponginess of the subependymal tissue was observed (Fig. 4), most marked in the white matter along the lateral wall of the lateral ventricle, especially in the area of the dorsolateral ventricular angle at the corpus callosum. The significance of these changes is somewhat uncertain. They are probably an expression of increased fluid content or oedema in the subependymal tissue. The volume of the brain intra- and extracellular space cannot be determined by light microscopy. However, the sucrose, sodium and mannitol space of the periventricular white matter have been shown to be larger in hydrocephalic than in normal brains 10, indicating enlargement of the extracellular space. Furthermore, the hydrocephalic white matter has been reported to have increased content of water 7,s, sodium and chloride and a significant loss of lipids and proteinsL The water and electrolyte composition of the cerebral cortex is normal 7. The sponginess of the periventricular white matter (Fig. 4) supports the concept of transependymal passage as an alternative pathway of CSF absorption during

Fig. 2. Microphotograph showing the normal structure of the ventricular wall in the parietal region in a 10-day-old control animal. Note the relatively tall and ciliated ependymal cells and the normal structure of the subependymal white matter. × 400. Fig. 3. Microphotograph showing the ependymal lining in the parietal region of the lateral ventricle from a hydrocephalic animal killed 3 weeks after the injection of kaolin. Note the flattened and stretched ependymal cells with small gaps (arrow) between the individual cells (cf. Fig. 2). × 400. Fig. 4. Microphotograph of the lateral ventricular wall from a hydrocephalic animal killed l week after the injection of kaolin. Note the sponginess of the subependymal tissue, probably indicating oedema of the periventricular white matter (cf. Fig. 2). × 200.

348 hydrocephalus. Thus, in chronic communicating hydrocephalus in dogs, Strecker et al. x9 demonstrated transependymal migration of labelled albumin to such an extent as to suggest bulk flow. The most prominent transfer of the radiopharmaceutical was found in the angle between the corpus callosum and the caudate nucleus 19. The possibility that the sponginess of the white matter in hydrocephalic brains represents fluid derived from the intracerebral blood vessels seems less likely, but cannot be completely ruled out. The fate of the CSF, once within the periventricular tissue, is not known. The experiments of Strecker and James 18 who determined the transfer of radiopharmaceutical a3qodine labelled albumin from the ventricular system to the general circulation, suggest that some of the fluid is absorbed into the vessels of the periventricular brain tissue. However, as these vessels represent a true barrier for absorption, the exact mechanism of absorption is essentially unknown. The spongy changes of the subependymal tissue seen in the present investigation (Fig. 4) were maximal after one week and then gradually subsided, possibly indicating that an approximation to a new equilibrium between the production and absorption of the CSF was taking place. Experimental evidence in favour of this has been reported by Weller et al. ~1 who made 2-week-old dogs hydrocephalic by infusion of silicone oil into the cisterna magna. These authors reported that the ensuing oedema of the periventricular white matter subsided as the intraventricular pressure returned to normal levels in the 80-day-old animals. The progressive dilatation of the ventricles caused by the elevation of the intraventricular pressure during hydrocephalus produces an increasing surface area for transventricular absorption of the CSF (Fig. 1). Furthermore, decrease of the resistance to transventricular absorption seems likely in view of the flattened and stretched ependyma observed in the present study (Fig. 3). These features may after some time allow absorption to keep pace with the production, and the intraventricular pressure may at this stage subside to normal levels. It may thus be suggested that 'arrest' of hydrocephalus may occur when a new balance is established between the extent of the ventricular wall area available for absorption and the amount of CSF secreted by the choroid plexus. This line of reasoning may bear some relationship to certain variants of human hydrocephalus presenting dementia, gait disturbances and urinary incontinence (occult, low pressure, normal pressure hydrocephalus) ~. In these cases there is little or no evidence for increased intracranial pressure when the distention of the ventricles has taken place 1, probably because a new equilibrium is established between the formation and absorption of the CSF. In the present investigation attempts were also made to study the penetration of protein across the ventricular ependymal lining. For these permeability studies some hydrocephalic animals and controls of the same age received intraventricular injections of 0.2 ml of 1 ~ Evans blue dissolved in 5 o/ bovine albumin (EBA) from 3 to 30 days after the kaolin injections. The animals were killed from 15 rain to 3 h after the dye injections and the brains fixed with 10~,i formalin. The penetration of the dye into the periventricular tissue was studied on the gross specimen as well as under the light and fluorescent microscope. A slow penetration of the dye into the periventricular tissue was seen after the

349 injections both in normal and hydrocephalic animals. There was apparently no convincing difference between hydrocephalic and control animals. In accordance with the observations of Ogata e t al. 15 who used horseradish peroxidase as a tracer, we found no evidence of increased passage of CSF in any particular areas of the ventricular wall (see, however, ref. 13). Altogether the present experiments showed that the ependyma adapted remarkably well without tearing under the increased intraventricular pressure when the normal anatomical pathways of the CSF circulation were blocked in newborn rabbits. The stretched and flattened, but essentially intact, ependymal cells observed (Fig. 3) may well be consistent with increased permeability of CSF across the ependyma, which apparently exerts little restraint on the migration of CSF in such cases. Breaks and ruptures of this lining are not necessary for the concept of transventricular bulk flow of CSF. Since the present experiments unveiled no frank destruction of periventricular tissue, the progressive attenuation of the cerebral mantle was probably caused by a simple pressure atrophy, mainly of the white matter. This suggests that the morphological changes may to a large extent be reversible if the hydrocephalus is properly treated within reasonable time. This may also have some bearings to certain aspects of human hydrocephalus, in which relative good preservation of intellectual functions may be found despite large dilatation of the lateral ventricles, adding further support to the view that interference with cerebral function in hydrocephalus is not necessarily due to neuronal destruction. Furthermore, in humans the alterations of the white matter long precede changes of the cerebral cortex and the deep nuclear grey matter (the basal ganglia and thalamus), possibly indicating greater vulnerability of the white matter to the increased pressure. The relative preservation of the grey matter is doubtless a factor to be taken into account in the surprising degree of intellectual development which often occurs in children in whom hydrocephalus has been successfully relieved. This seems also to be true in some properly treated cases of so-called 'normal' pressure hydrocephalus in adults in whom a considerable intellectual content may be compatible with large hydrocephalus. However, since there may be severe loss of neural tissue and extensive gliosis, especially around the lateral ventricles in cases of long-standing human hydrocephalus16, it should be emphasized that such patients ought to be treated as early as possible.

REFERENCES l ADAMS,R. D., FISHER, C. M., HAKIM, S., OJEMANN,R. G., AND SWEET, W. H., Symptomatic occult hydrocephalus with 'normal' cerebrospinal-fluid pressure. A treatable syndrome, New Engl. J. Med., 273 (1965) 117-126. 2 BERING,E. A., AND SATO, O., Hydrocephalus: changes in formation and absorption of cerebrospinal fluid within the cerebral ventricles, J. Neurosurg., 20 (1963) 1050-1063. 3 BRIGHTMAN,M. W., The intracerebral movement of proteins injected into blood and cerebrospinal fluid in mice. In A. LAJTHAAND D. H. FORD (Eds.), Brain Barrier Systems, Progress in Brain Research, Vol. 29, Elsevier, Amsterdam, 1968, pp. 19-40. 4 CAMMERMEYER,J., Differences in shape and size of neuroglial nuclei in the spinal cord due to individual, regional and technical variations, Acta anat. (Basel), 40 (1960) 139-172.

350 5 CLARK, R. D., AND MILHORAT, T. H., Experimental hydrocephalus. Part 3. Light microscopic

findings in acute and subacute obstructive hydrocephalus in the monkey, J. Neurosurg., 32 (1970) 400-413. 6 DE, S. N., A study of the changes in the brain in experimental internal hydrocephalus, J. Path. Bact., 62 (1950) 197-208. 7 FISHMAN,R. A., ANDGREER, M., Experimental obstructive hydrocephalus, Arch. Neurok (Chic.), 8 (1963) 156-161. 8 HOCHWALO,G. M., Lux, W. E., SAHAR,A., AND RANSOHOFF,J., Experimental hydrocephalus, changes in cerebrospinal fluid dynamics as a function of time, Arch. NeuroL (Chic.), 26 (1972) 120-129.

9 HOCHWALD,G. M., SAHAR,A., SADIK,A. R., AND RANSOHOFF,J., Cerebrospinal fluid production and histological observations in animals with experimental obstructive hydrocephalus, Exp. NeuroL, 25 (1969) 190-199. 10 LEVIN, V. A., MILHORAT,T. H., FENSTERMACHER,J. D., HAMMOCK,M. K., AND RALL, D. P., Physiological studies on the development of obstructive hydrocephalus in the monkey, Neurology (Minneap.), 21 (1971) 238-246, 11 LORENZO,A. V., PAGE, L. K., AND WATTERS,G. V., Relationship between cerebrospinal fluid formation, absorption and pressure in human hydrocephalus, Brain, 93 (1970) 679-692. 12 Lux, W. E., HOCHWALD,G. M., SAHAR,A., AND RANSOHOFF,J., Periventricular water content. Effect of pressure in experimental chronic hydrocephalus, Arch. Neurol. (Chic.), 23 (1970) 475479. 13 MILHORAT,T. H., Hydrocephalus and the Cerebrospinal Fluid, Williams and Wilkins, Baltimore, Md., 1972. 14 MILHORAT,T. H., CLARK,R. G., HAMMOCK,M. K., AND McGRATH, P. P., Structural, ultrastructural and permeability changes in the ependyma and surrounding brain favoring equilibration in progressive hydrocephalus, Arch. Neurol. (Chic.), 22 0970) 397-407. 15 OGATA,J., HOCHWALD,G. M., CRAVIOTO,H., AND RANSOHOFF,J., Light and electron microscopic studies of experimental hydrocephalus, Acta neuropath. (Berl.), 2l (1972) 213-223. 16 RUSSEL,D. S., Observations on the Pathology of Hydrocephalus, Her Majesty's Stationery Office (Special Report Series Medical Research Council, No. 265), London, 1949. 17 SAHAR,A., HOCHWALD,G. M., AND RANSOHOFF,J., Alternate pathway for cerebrospinal fluid absorption in animals with experimental obstructive hydrocephalus, Exp. Neurol., 25 (1969) 200-206. 18 STRECKER, E. P., AND JAMES,A. E., The evaluation of cerebrospinal fluid flow and absorption: clinical and experimental studies, Neuroradiology, 6 (1973) 200-205. 19 STRECKER,E. P., JAMES,A. E., KONIGSMARK,B., AND MERZ, T., Autoradiographic observations in experimental communicating hydrocephalus, Neurology (Minneap.), 24 (1974) 192-197. 20 TORVTK,A., BHATIA,R., AND NYBERG-HANSEN,R., The pathology of experimental obstructive hydrocephalus, Neuropath. appl. Neurobiol., (1975) in press. 21 WELLER,R. O., WlgN~EWSKI,H., SI-tULMAN,K., AND TERRY, R. D., Experimental hydrocephalus in young dogs: histological and ultrastructural study of the brain tissue damage, J. Neuropath. exp. Neurol., 30 (1971) 613-626.