CHEMICAL ENVIRONMENT OF THE CENTRAL NERVOUS SYSTEM 



1875 



fig. 6. P 32 concentration of various parts of the human brain 

 following intravenous injection of 2 mc of the tracer per 70 kg 

 of body weight. [From Bakay (9).] 



account for the appearance of electrical spike activity 

 in fiber tracts joining the area postrema and adjacent 

 medulla following intravenous 3 m sodium chloride. 



It may be concluded that for substances to reach 

 the functional cellular elements of most of the central 

 nervous system from the blood stream, they must 

 pass through one or more 'barriers' which in some 

 fashion regulate this intra-extravascular traffic. The 

 biochemical machinery of these barriers is -still ob- 

 scure, but is clearly of major significance in under- 

 standing the control of the chemical environment of 

 neurons. Krogh (98) has recently reintroduced the 

 concept of lipoid solubility as a predominant factor 

 in determining barrier permeability. This hypothesis, 

 while quite satisfactory for some compounds, is, for 

 example, unable to account for the rapid penetration 

 of barbiturates (87) or for the relative penetration of 

 a series of sulfonamides (56). The blood-brain pene- 

 tration of these latter compounds is not related to 

 lipoid solubility nor to molecular weight, but rather 

 to their dissociation constants. 



In general, however, there is considerable similarity 

 between blood-brain barrier permeability and cell 

 membrane permeability. It appears that for solutes 

 to pass from the plasma to the intercellular fluid of 

 the central nervous system (with the exceptions 

 noted) they must, for the most part, pass through, 

 not between cells. This barrier does not represent a 

 simple sievelike impediment to otherwise free diffu- 

 sion, but rather a distinct polarized secretory mecha- 



nism which governs the concentration of certain 

 solutes in the intercellular fluid. 



Despite the mass of data which has accumulated, 

 the only generalizations that can be made remain 

 phenomenological: "... that the blood-brain barrier 

 (including the blood-cerebrospinal fluid barrier) 

 is a mechanism for producing a peculiarity in the 

 exchange of most substances between the plasma and 

 the intercellular fluid of the central nervous system; 

 that this peculiarity manifests itself as a decreased 

 rather than an increased rate of exchange; and that 

 this phenomenon is not localized to the choroid 

 plexus, but is associated with the entire cerebral 

 vasculature" (146). 



Anatomy of the Blood-Brain Barriet 



The anatomy of this panvascular barrier has re- 

 in, lined surprisingly elusive, but the claims and 

 counterclaims have developed a fervor exceeding 

 almost any other aspect of this subject. Following the 

 early unsatisfactory choroid plexus hypothesis of 

 Goldmann referred to previously, various anatomical 

 sites were proposed \% 1 1 iih involved, among other 

 structures, the meninges and the reticuloendothelial 

 system (170). However, the polemic dispute resolved 

 into two major camps, the perivascular glial mem- 

 brane theory and the capillary endothelium theory. 

 Hauptmann & Gartner (66) and Hoff (82) clearly 

 formulated the hypothesis di.it the barrier is located 

 in the perivascular pia-glia membrane. This mem- 

 brane was thought to lie composed of the invaginating 

 pia which follows die vessels from the surface into the 

 depth of the brain, supported by a closelv adherent 

 layer of astrocytes (fig. 4). Standard histological 

 techniques and light microscopy failed to reveal 

 whether or not this double membrane accompanied 

 the vasculature throughout its terminal ramifications, 

 enveloping the capillaries. However, in electron- 

 micrographs (figs. 2, 4) of cerebral capillaries, astro- 

 cytic feet can be found in association with most of 

 the capillary surface (114). Maynard ft al. (114) 

 claim that the expanded astrocytic processes form 

 but a single incomplete layer around capillaries and 

 do not overlap one another to am extent. They 

 estimate that this 'sheath' covers about 85 per cent 

 of the total capillary surface and state that the pia 

 does not continue down to surround the capillaries, 

 providing a mesodermal barrier between the vessels 

 and the ectodermal elements of the central nervous 

 system, as had been proposed. At that level, the 

 morphological blood-brain barrier consists only of the 



