The Role of Altered Neuron Connections on Outcome
The cerebral cortex from a hydrocephalic brain viewed at very high magnification with the electron microscope. Damage can be seen in the form of watery, swollen dendrites (d2) and axons (a), as well as dark degenerating dendrites (d1). Although the number of synapses is reduced, some are still present (arrows), suggesting that the brain is still maturing, although to a lesser degree than normal. Scale bar = 5 microns.

Synaptic Connections

Neurons interact with each other by highly complicated, but selective, "wiring diagrams" , formed by an axon from one neuron connecting to the cell body or dendrites of another neuron via a synapse. Synapses comprise the narrow clefts between axon endings and dendrites. Electrical impulses travel down the axon and cause the release of chemicals (neurotransmitters), which cross the synaptic cleft and stimulate the next neuron. We and others have shown, using a variety of techniques, that axons degenerate, dendrites deteriorate, the number of synapses is reduced, and the levels of neurotransmitters decrease during untreated Hydrocephalus. Shunting can reverse these alterations, but even early treatment cannot bring all of these changes back to normal. Thus, "biological" treatments that could supplement shunting and either protect neurons from damage or promote their recovery are being pursued.

Axon Connections

Connectivity of the cerebral cortex may be irreversibly altered by hydrocephalus. Previous studies have support ed this hypothesis by demonstrating reductions in cortical synapses, decreased monoamine levels, pyknotic neurons in layers V and VI, dendritic atrophy, demyelination and axonal degeneration in periventricular white matter. To examine this possibility directly, a study using an axonal tracer was initiated on kittens in which hydrocephalus was induced at 9-11 days of age by intracisternal injection of kaolin. At 10-14 days post-kaolin, 5 hydrocephalic animals received low pressure ventriculoperitoneal (VP) shunts. Normal age and weight matched animals served as controls. Hydrocephalic and shunted animals were monitored by ultrasound to document progression of hydrocephalus. Injections of horseradish peroxidase conjugated with wheat-germ agglutinin (WGA-HRP) were made unilaterally in cortical areas 3,4 and 6 in the following sequence: hydrocephalic animals at 9-15 days post-kaolin; shunted animals at 1,2 and 4 weeks post-shunt; control animals at times corresponding to the ages of the hydrocephalic and shunted animals. Tissue from the entire brain was processed by routine HRP histochemistry and analyzed using light microscopy.

Camera lucida drawings representative of coronal sections, each at the same magnification showing injection sites in the control (A), Hydrocephalic (B) and VP shunted brains (C). Contralateral cortex was devoid of both retrograde and anterograde labelling in untreated hydrocephalus. With shunting, contralateral cortex exhibited retrograde and anterograde labeling which approximated the control animal.
Retrograde labeling of neuronal cell bodies in the contralateral somatosensory cortex was absent in untreated hydrocephalic animals, but returned to normal after VP shunting. Ipsilateral ventral lateral (VL), centromedial (CM) and centrolateral (CL) nuclei of the rostral thalamus and the ventral tegmental area (VTA) exhibited a decrease in retrograde labeling in hydrocephalus compared to normal control animals. VP shunting returned retrograde labeling in these areas to normal. Retrograde labeling in the ipsilateral ventrobasal nucleus (VB) remained relatively unaffected in both hydrocephalic and shunted animals.

Camera lucida drawings at rostral thalamic levels in the moderately hydrocephalic condition (C,D) as compared to a normal control (A,B) and the VP shunted animal (E,F). In general, retrograde labelling was reduced in hydrocephalus and approximated the control after VP shunting. Anterograde labelling in thalamic nuclei was reduced in hydrocephalus and was not completely restored by VP shunting. Anterograde labelling in the internal capsule was unaffected. Each dot represents 3 retrogradely labelled neurons.

Anterograde labeling in the contralateral somatosensory cortex was present in shunted animals but lacked the same distribution as normal controls. Anterograde labeling within the internal capsule remained virtually unchanged, suggesting that corticobulbar, corticopontine and corticospinal pathways were unaffected in hydrocephalus. The preservation of the anterograde labeling of the internal capsule continued throughout the crus cerebri in the brainstem. Anterograde label in VL decreased in hydrocephalus compared to normal controls. VP shunting increased anterograde labeling in VL but not to normal levels. Ipsilateral anterograde label was absent in CL nuclei in the hydrocephalic and VP shunted groups. Likewise, anterograde labeling of VB nuclei remained relatively unaffected.

Photomicrographs from thalamus showing HRP labelling in control (A-C), hydrocephalic (D-F), and VP shunted (G-I) animals. In the centrolateral nucleus (left column)and the VL nucleus viewed with bright field (middle column) and darkfield (right column) optics, both retrograde and anterograde labelling was reduced in hydrocephalus. VP shunting restored retrograde labelling but anterograde labelling was still reduced. Scale bar same for all panels.

The results suggest that (1) cortical connectivity was impaired in untreated hydrocephalus and involves both afferent and efferent pathways, (2) shunting improved both cortical afferent and efferent connectivity and, (3) complete reestablishment of the cortical efferent pathways did not occur. Such changes in cortical pathways, if permanent, could be responsible for many of the motor and cognitive deficits seen clinically in afflicted hydrocephalic children.