Neurovascular relationships in hippocampal slices: physiological and anatomical studies of mechanisms underlying flow-metabolism coupling in intraparenchymal microvessels.

Experiments were carried out to investigate the functional and anatomical relationships between neuronal elements and cerebral microvessels in 300-350-microm thick coronal hippocampal slices maintained at 33-35 degrees C, obtained from 150-200 g male Wistar rats. Cerebral arterioles (9-22 microm in diameter) were visualized in situ and pre-constricted by 22.0+/-6.6% by the addition of the thromboxane A2 agonist U46619 (75 nM), to the bathing medium. The glutamate agonist N-methyl-D-aspartate (0.01-1 mM) produced a dose-related increase in luminal diameter of pre-constricted vessels. In the presence of 4 microM haemoglobin to scavenge nitric oxide from the extravascular environment of the slice, the increase in diameter evoked by 0.1 mM N-methyl-D-aspartate was significantly reduced from 17.5+/-4.6% to 4.8+/-1.7% indicating that N-methyl-D-aspartate-induced vasodilatation of cerebral microvessels is mediated via a mechanism which involves neuronally-derived nitric oxide. In a parallel anatomical study, beta-nicotinamide adenine dinucleotide phosphate-dependent diaphorase staining was used to reveal the enzyme nitric oxide synthase in vascular endothelium and neurons in slices. A small subpopulation (< 11 cells per slice) of darkly-stained multipolar neurons, 21-32 microm in diameter was observed to give rise to a dense network of fine diaphorase-reactive nerve fibres that ramified throughout the whole of the hippocampus and appeared to come into close apposition with arterioles. Morphometric analysis of the relationship between cerebral microvessels, beta-nicotinamide adenine dinucleotide phosphate, reduced form-dependent diaphorase-reactive neuronal elements and individual pyramidal layer neurons, identified by filling with biocytin, revealed that for a given point on a pyramidal layer neuron, the proximity of the nearest diaphorase-reactive nerve fibre was less than 10 microm, whilst the distance to the nearest arteriole (the smallest functional unit for controlling blood flow) was in excess of 70 microm. Such a distance would probably preclude diffusion of vasoactive metabolites in effective concentrations from the area of increased neuronal activity. We therefore propose that the diaphorase-reactive nerve network constitutes the functional link. It is possible that during periods of increased neuronal activity, spillover of glutamate from synapses may activate the diaphorase-reactive network. Release of nitric oxide from the network in the vicinity of local cerebral arterioles may then produce relaxation of the vascular smooth muscle, enabling increased blood flow into the capillary network supplying the region of increased metabolic activity. This study has shown that the process whereby increases in neuronal activity elicit a local change in cerebral blood flow remains functionally intact in hippocampal slice preparations. Nitric oxide of neuronal origin appears to be involved in mediating the coupling between neurons and cerebral arterioles. Stereological analysis of the relationship between neuronal and vascular elements within hippocampal slices suggested that a small subpopulation of nitric oxide synthase-containing neurons which give rise to a diffuse network of fine nitric oxide synthase-containing nerve fibres that lie in close apposition to cerebral arterioles may provide the anatomical substrate for coupling of blood flow to metabolism.