What is the blood-brain barrier (not)?

In 1900, summarizing his experiments with toxins and Ehrlich's earlier observations with intravital dyes, the Berlin physician Lewandowski concluded that 'brain capillaries must hold back certain molecules'. Illustrating this phenomenon with persuasive beauty, the subsequently evolving metaphor of a 'Bluthirnschranke' (blood-brain barrier, BBB) gained wide acceptance, but the extension of its meaning into the context of inhibiting leukocyte recruitment into the brain is imprecise. On the basis of the original work by Ehrlich, Lewandowski and Goldmann we re-define the BBB as a capillary barrier for solutes, and clarify that leukocyte recruitment requires two differentially regulated steps: (i) passage across postcapillary venules into Virchow-Robin spaces, and (ii) subsequent progression across the glia limitans into the neuropil. We propose that the second step frequently involves perivascular antigen-recognition and the induction of ectoenzymes, for example matrix metalloproteinasies (MMPs).