Ultrastructural observations in experimental hydrocephalus in the rabbit

Fingerprint changes in CSF composition associated with different aetiologies in human neonatal hydrocephalus: inflammatory cytokines

PURPOSE

Hydrocephalus (HC) has a multifactorial and complex picture of pathophysiology due to aetiology, age at and duration since onset. We have previously identified distinctions in markers of cell death associated with different aetiologies. Here, we examined cerebrospinal fluid (CSF) from human HC neonates for cytokines to identify further distinguishing features of different aetiologies.

METHODS

CSF was collected during routine lumbar puncture or ventricular tap from neonates with hydrocephalus, or with no neurological condition (normal controls). Total protein, Fas receptor, Fas ligand, stem cell factor (SCF), hepatocyte growth factor (HGF), vascular endothelial growth factor (VEGF), insulin growth factor-1 (IGF-1), tumour necrosis factor alpha (TNF-α) and interleukin 6 (IL-6) were measured and compared between 8 unaffected and 28 HC neonatal CSF samples.

RESULTS

Total protein was significantly (P < 0.05) raised in late-onset hydrocephalus (LOH). Fas receptor was raised (P < 0.05) in post-haemorrhagic hydrocephalus (PHH) and spina bifida with hydrocephalus (SB/HC), but no difference in Fas ligand was found. SCF was raised (P < 0.05) in SB/HC. HGF was found in all HC and was increased (P < 0.01) in PHH. Increased VEGF was found in PHH (P < 0.01) and SB/HC (P < 0.05). Variable levels of IL-6, TNF-α and IGF-1 were found in all HC groups compared with none in normal.

CONCLUSIONS

LOH was unusual with significantly raised total protein indicating an inflammatory state. Increased Fas receptor, VEGF, IGF-1 and HGF suggest anti-apoptotic and repair mechanism activation. By contrast, elevated TNF-α and IL-6 indicate inflammatory processes in these neonatal brains. Taken with our previous study, these data indicate that different pathophysiology, inflammation and repair are occurring in HC of different aetiologies and that additional treatment strategies may benefit these infants in addition to fluid diversion.

Soluble Fas (CD95/Apo-1), soluble Fas ligand, and activated caspase 3 in the cerebrospinal fluid of infants with posthemorrhagic and nonhemorrhagic hydrocephalus

Hydrocephalus may result in loss of tissue associated with neuronal degeneration, axonal damage, and reactive gliosis. The soluble form of the anti-apoptotic regulator Fas (sFas) and the pro-apoptotic factors soluble FasL (sFasL) and activated caspase 3 were studied in the cerebrospinal fluid of infants with hydrocephalus. Fifteen preterm infants with posthemorrhagic hydrocephalus undergoing serial reservoir puncture and seven term or near-term infants with nonhemorrhagic hydrocephalus and shunt surgery were included in the study. Twenty-four age-matched patients with lumbar puncture for the exclusion of meningitis served as controls. Elevated levels of sFas were observed in infants with posthemorrhagic hydrocephalus [median (range), 131 ng/mL (51-279 ng/mL)] and in nonhemorrhagic hydrocephalus [127 ng/mL (35-165 ng/mL)]. sFas concentrations were highest in a subgroup of eight patients with posthemorrhagic hydrocephalus developing periventricular leukomalacia [164 ng/mL (76-227 ng/mL)]. In contrast, in 24 control infants, sFas was low, in 15 cases below detection limit (0.5 ng/mL) and in nine cases, 24 ng/mL (20-43 ng/mL). sFasL and activated caspase 3 did not differ from control infants in all groups of patients. Increased intrathecal release of sFas in the cerebrospinal fluid of infants with hydrocephalus may serve as an indicator of brain injury from progressive ventricular dilatation.

Association of learning and memory impairments with changes in the septohippocampal cholinergic system in rats with kaolin-induced hydrocephalus

OBJECTIVE

The septohippocampal cholinergic (SHC) system plays an important role in the maintenance of normal memory and learning. However, the fact that memory and learning impairments under hydrocephalic conditions are directly related to the SHC system is less well known. We investigated the relationships between pathological changes in SHC neurons and impairments in memory and learning among hydrocephalic rats.

METHODS

Rats with kaolin-induced hydrocephalus were prepared with injections of kaolin suspension into the cisterna magna. Learning and memory performance was assessed with the passive avoidance and Morris water maze tests. Ventricular sizes were measured for the lateral and third ventricles. Acetylcholinesterase and choline acetyltransferase immunostaining was performed to investigate degenerative changes in cholinergic neurons in the medial septum and hippocampus.

RESULTS

Hydrocephalic rats demonstrated significant learning and memory impairments in the passive avoidance and Morris water maze tests. Decreased hesitation times in the passive avoidance test and markedly increased acquisition times and decreased retention times in the Morris water maze test indicated learning and memory dysfunction among the hydrocephalic rats. The numbers of cholinergic neurons in the medial septum and hippocampus were decreased in the hydrocephalic rats. The decreases in choline acetyltransferase and acetylcholinesterase immunoreactivity were significantly correlated with enlargement of the ventricles.

CONCLUSION

Impairment of spatial memory and learning may be attributable to degeneration of SHC neurons. These results suggest that learning and memory impairments in rats with kaolin-induced hydrocephalus are associated with the dysfunction of the SHC system induced by ventricular dilation.

Effects of brain ventricular shape on periventricular biomechanics: a finite-element analysis

OBJECTIVE

A computer simulation based on the finite-element method was used to study the biomechanics of acute obstructive hydrocephalus and, in particular, to define why periventricular edema is most prominent in the anterior and posterior horns.

METHODS

Brain parenchyma was modeled as a two-phase material composed of a porous elastic matrix saturated by interstitial fluid. The effects of the cerebrovascular system were not included in this model. The change in the shape of the ventricles as they enlarged was described by two variables, i.e., the stretch of the ependyma and the concavity of the ventricular wall. The distribution of stresses and strains in the tissue was defined by two standard mechanical measures, i.e., the mean effective stress and the void ratio.

RESULTS

With obstruction to cerebrospinal fluid flow, the simulation revealed that the degree of ventricular expansion at equilibrium depended on the pressure gradient between the ventricles and the subarachnoid space. Periventricular edema was associated with the appearance of expansive (tensile) stresses in the tissues surrounding the frontal and occipital horns. In contrast, the concave shape in the region of the body of the ventricle created compressive stresses in the parenchyma. Both of these stresses seem to be direct consequences of the concave/convex geometry of the ventricular wall, which serves to selectively focus the forces (perpendicular to the ependyma) produced by the increased intraventricular fluid pressure in the periventricular tissues.

CONCLUSION

The distribution of periventricular edema in acute hydrocephalus is a result not only of increased intraventricular pressure but also of ventricular geometry.

Progressive changes in cortical water and electrolyte content at three stages of rat infantile hydrocephalus and the effect of shunt treatment

Infantile hydrocephalus causes injury to the developing brain and despite surgical treatment, neurological deficits persist. The H-Tx rat develops inherited hydrocephalus in late gestation. Rapid postnatal ventricular enlargement, results in severe hydrocephalus by 21 days after birth. This is accompanied by changes in cortical morphology and metabolite content that indicate possible changes in intracellular composition. This study has tested the hypothesis that tissue water and electrolyte content is altered in hydrocephalus. The objective was to gain further insight into the mechanisms leading to neuronal damage. Water and electrolyte content (Na+, Cl-, and K+) were measured in the cerebral cortex of control and hydrocephalic rats at 4, 11, and 21 days after birth, and at 21 days in rats that received alleviating shunt surgery at 4 or 11 days. At all ages, hydrocephalic tissue was significantly increased over control for cortical water, Na+, and Cl- content. Additionally, at the intermediate (11-day) and advanced (21-day) stages there were significant decreases in K+ content, consistent with previous observations of decreases in organic osmolytes and energy metabolites. This suggests that by 11 days there are intracellular changes, probably through impaired membrane homeostatic mechanisms. In shunt-treated rats, the extracellular constituents were almost normal, although a small increase over control values persisted. The decrease in intracellular K+ was not corrected in either group of shunt-treated rats. It is concluded that early hydrocephalus is characterized by extracellular edema that largely reverses with shunt treatment. Subsequently, as the hydrocephalus progresses, there is a breakdown of cell homeostasis and an irreversible loss of intracellular constituents.

Neuroepithelial and ependymal changes in HTX rats with congenital hydrocephalus: an ultrastructural and immunohistochemical study

To investigate the pathogenesis of congenital hydrocephalus the brains of HTX rats aged between 16 days and 4 weeks and the brains of normal Wistar rats of the same ages were examined. In the fetal HTX rat brains, the lateral ventricles were symmetrically dilated from 20 days of gestation. The neuroepithelium bordering the ventricles showed thinning with cellular disarrangement and deformity. Similar neuroepithelial abnormalities were also found in the lateral ventricles of the HTX rat brain with no macroscopic signs of hydrocephalus at 20 days of gestation. The neuroepithelium showed flattening of the cells, widening of the intercellular spaces, formation of microvilli on the detached lateral cell surfaces, and frequent macrophage infiltration. On the other hand, the neuroepithelial cells of the third ventricle and the aqueduct were affected less severely or showed no significant abnormalities. Immunohistochemically, most of the neuroepithelium and ependyma of the lateral ventricles were positive for vimentin in both prenatal and postnatal hydrocephalic HTX rats, while a small number or none of those in normal control rats were positive. These morphological changes suggested that preferential involvement of the lateral ventricular neuroepithelium might be closely associated with the pathogenesis of congenital hydrocephalus in HTX rats.

Ultrastructural changes in the deep cortical pyramidal cells of infant rats with inherited hydrocephalus and the effect of shunt treatment

Pathological changes in the cortical gray matter in infantile hydrocephalus vary with the age at onset and may not be reversible with shunt treatment. We have used electron microscopy to investigate the sequence of pathological change and the effect of shunt treatment on layer VI pyramidal cells from infant H-Tx rats with inherited early-onset hydrocephalus. Tissue was prepared from the frontal and visual cortex of control and hydrocephalic rats at 4, 11, and 21 days after birth, together with 21-day rats previously treated with ventriculosubcutaneous shunts at 4-5 or 10-11 days after birth. Both cortical regions gave similar results but the effects were more severe in the visual cortex. In the early stages of hydrocephalus, the pyramidal cells were in clusters with fewer mature dendrites and less cytoplasmic organization than those in control rats, and some neuronal processes were vacuolated. In intermediate hydrocephalus the changes were more severe, with vacuolated cytoplasm, fewer cytoplasmic organelles, frequent swollen processes, and infrequent synapses. In advanced hydrocephalus at 21 days, many neurons showed degenerative changes, with edematous Golgi and dilated endoplasmic reticulum, distorted mitochondria, and single ribosomes. The neuropil contained many spongy areas with distended profiles. Shunt treatment prevented most of the changes if carried out at 4 days. Shunt treatment at 11 days also gave a dramatic recovery at the cellular level, but there were more immature pyramidal cells and edematous processes in the neuropil than in the 4-day-treated rats. The changes in hydrocephalus are consistent with progressive neuronal damage, which is largely prevented by early shunt treatment.

Neuropathological changes caused by hydrocephalus

The medical literature concerning neuropathological changes caused by hydrocephalus is reviewed. In both humans and experimental animals the ependyma suffers focal destruction, cerebral blood vessels are distorted and capillaries collapse, there is damage to axons and myelin in the periventricular white matter, and occasionally neurons suffer injury. The damage appears to result from mechanical distortion of the brain combined with impaired cerebral blood flow. If ventriculomegaly develops very early, foci of cortical dysgenesis may be the result. The character and distribution of pathological changes are dependent on the age at which hydrocephalus develops, the rate and magnitude of ventricular enlargement, and the duration of hydrocephalus. Diversionary shunting of cerebrospinal fluid can only incompletely reverse the damage and the potential for reversal diminishes as the duration of hydrocephalus increases.

Reconstitution of shunted mantle in experimental hydrocephalus

The morphological mechanism of the reconstitution of shunted mantle was studied histopathologically in 22 kaolin-treated hydrocephalic puppies. A remarkable attenuation of cerebral mantle to less than 1 cm in thickness was seen on computerized tomography (CT) scans of four animals sacrificed 1 to 2 months after kaolin treatment (preshunt group). Ventricular shunting resulted in successful recovery of the mantle on repeated CT scans obtained 1 to 2 months after shunting in seven animals (postshunt group). In the remaining 11 animals the cerebral mantle, which had been reduced to 4 mm in thickness prior to shunting, failed to recover even 2 months after the procedure (shunt-refractory group). On gross inspection, the preshunt specimens showed marked thinning of the white matter, with the cortical ribbon well preserved, while the postshunt specimens consisted predominantly of thickened white matter. Histopathological examination of the attenuated white matter of the preshunt specimens showed decreased nerve-fiber density, myelin destruction with myelin regeneration and/or repair of myelin sheaths, and reactive astrocytosis, which were prominent especially in the periventricular white matter. The main findings in the reconstituted white matter of the postshunt specimens were extensive myelin regeneration of residual axons and remarkable astroglial proliferation with mesenchymal reaction, particularly at capillaries. No clear evidence of increased numbers of nerve fibers or axonal regeneration was observed. The shunt-refractory specimens showed remarkable attenuation of cortex, in which reduced numbers of neurons and loss of cortical lamination were noted, with vestigial white matter. The results indicate that astroglial proliferation with mesenchymal reaction and myelin regeneration contribute to the reconstitution of the cerebral mantle volume following ventricular shunting in this model. It is suggested that the critical factor for mantle reconstitution in chronic hydrocephalus is whether cortex is preserved.

Silicone oil-induced hydrocephalus in the rabbit

Hydrocephalus was induced in adult rabbits by injecting silicone oil into the cisterna magna. Mean intracranial pressure was significantly elevated for approximately 36 h post-injection, during which time maximal ventricular dilatation was attained. Stretching and compression of periventricular tissue and capillaries accompanied dilation of the lateral ventricles. Ventricular dilation promoted mitotic activity among the periventricular astroglia. Ventriculomegaly altered the metabolism of the monoamine neurotransmitters in the cortex, hippocampus, diencephalon, hypothalamus, and brain-stem. Ischemic injury to neurons of the hippocampal formation, particularly the dentate gyrus, was observed when hydrocephalus had persisted for more than 4 weeks. Cerebrospinal fluid shunting effectively reversed the neuropathologic changes only when done in the early stages of hydrocephalus. When hydrocephalus persisted for 8 weeks, rapid reversal of changes in the ependyma and periventricular capillaries was prevented largely by periventricular gliosis.

Neuronal effects of experimentally induced hydrocephalus in newborn rats

To determine the effects of increased cerebrospinal fluid (CSF) pressure on neuronal morphology, obstructive hydrocephalus was induced by injecting kaolin into the fourth ventricle and cisterna magna of 1-day-old rats. The animals were sacrificed 10 to 12 days later, at which time severe ventriculomegaly and cortical thinning were apparent in the parieto-occipital region. Tissue from this area was processed by rapid Golgi methods. Well impregnated pyramidal neurons were examined by light microscopy, and their somatic and dendritic features compared to those of age-matched littermate controls. The somata of medium pyramidal neurons were unaffected, but their basilar dendrites had fewer branches and those that remained were shorter. A variable reduction in dendritic spines occurred, such that some branches were totally denuded while others exhibited spine densities similar to those seen in control animals. The most striking alteration was the occurrence of frequent dendritic varicosities. These enlargements of the dendritic shaft separated by extremely thin constrictions gave the affected segment a beaded appearance. Both dendritic spine loss and varicosity formation were most notable on distal portions of individual branches and within regions of the dendritic tree closest to the ventricular and meningeal surfaces. These alterations are consistent with other reports of dendritic changes associated with aging, mental retardation, and alcohol exposure. These observations suggest that hydrocephalus causes dendritic deterioration or retardation of dendritic maturation. The fact that neuronal morphology was not more severely affected may indicate that these effects are reversible.

Tracer study on a paracellular route in experimental hydrocephalus

Considering the possibility of a paracellular pathway for edema resolution, we studied the intracerebral movement of proteins and ionic lanthanum in rats with experimental hydrocephalus. Hydrocephalus was induced by injection of kaolin suspension into the cisterna magna. After induction of hydrocephalus, horseradish peroxidase (HRP), microperoxidase (MP), or lanthanum chloride (LaCl3) were perfused into the ventricle system. HRP and MP were localized mainly in the intercellular spaces between ependymal cells, glial cells, and in perivascular spaces and were restricted by endothelial tight junctions. Ionic lanthanum (La3+), however, penetrated these tight junctions and moved between the blood and CSF cavities by paracellular pathways. These findings indicate that in obstructive hydrocephalus, the tight junctions may constitute part of a paracellular pathway for the transendothelial movement of small solutes, although they prevent the movement of larger molecules.

A possible paracellular route for the resolution of hydrocephalic edema

Considering the possibility of a paracellular route for edema resolution we studied the microvasculature of the subependymal and subcortical white matter in hydrocephalic rats. Normal adult rats were used as controls. After injection of kaolin suspension into the cisterna magna, the animals were killed at intervals of 1, 2, 4, and 8 weeks. In hydrocephalic rats at 1 week after kaolin injection, widening of the interendothelical cleft between the tight junction (dehiscence) was seen in 27 of 76 (35%) vessels. At 2 weeks after kaolin injection, the number of the dehiscences had increased (39/7:56%) and some were enlarged, forming interendothelial blisters. At 4 weeks in hydrocephalic rats, both dehiscences and blisters were still prominent (45/73:63%) and at 8 weeks the dehiscences were still prominent, but the number of the blisters had decreased (25/81:31%). The blisters and dehiscences were most pronounced in the corpus callosum and occipital regions. Following i.v. injection of horseradish peroxidase, the interendothelial dehiscences and blisters were completely devoid of the marker substance. These findings indicate that in obstructive hydrocephalus the tight junctions may constitute part of a paracellular pathway for the resorption of interstitial edema fluid.

Cerebrospinal fluid outflow resistance in rabbits with experimental meningitis. Alterations with penicillin and methylprednisolone

Acute bacterial meningitis may be associated with increased intracranial pressure, neurological sequelae such as communicating hydrocephalus, and a slow response to antibiotic therapy. Alterations in cerebrospinal hydrodynamics are at least partially responsible for these complications. Constant, low-flow short-duration manometric infusion studies through a hollow-bore pressure monitoring device in direct continuity with the supracortical subarachnoid space were performed in rabbits with experimental meningitis. Maximal resistance to cerebrospinal fluid (CSF) outflow from the subarachnoid to vascular space was markedly increaed in acute pneumococcal meningitis when compared to control, uninfected animals (6.77 +/- 3.52 vs. 0.26 +/- 0.04 mm Hg/microliter per min, P less than 0.001). Similar elevations (8.93 +/- 4.15 mm Hg/microliter per min were found in experimental Escherichia coli meningitis. Despite eradication of viable bacteria from the CSF by penicillin therapy during the acute stage of pneumococcal meningitis, resistance remained elevated (6.07 +/- 4.68 mm Hg/microliter per min) and had not returned to normal up to 15 d later. Administration of methylprednisolone during the early stages of acute pneumococcal meningitis reduced mean peak outflow resistance towards control values (0.59 mm Hg/microliter per min) and no "rebound" effect was apparent 24 h later. These hydrodynamic alterations in experimental meningitis prevent normal CSF absorption and decrease the ability of the bran to compensate for changes in intracranial volume and pressure.