Intrathecal antibody distribution in the rat brain: surface diffusion, perivascular transport and osmotic enhancement of delivery.

Abstract

It is unclear precisely how macromolecules (e.g. endogenous proteins and exogenous immunotherapeutics) access brain tissue from the cerebrospinal fluid (CSF). We show that transport at the brain-CSF interface involves a balance between Fickian diffusion in the extracellular spaces at the brain surface and convective transport in perivascular spaces of cerebral blood vessels. Intrathecally-infused antibodies exhibited size-dependent access to the perivascular spaces and tunica media basement membranes of leptomeningeal arteries. Perivascular access and distribution of full-length IgG could be enhanced by intrathecal co-infusion of hyperosmolar mannitol. Pores or stomata present on CSF-facing leptomeningeal cells ensheathing blood vessels in the subarachnoid space may provide unique entry sites into the perivascular spaces from the CSF. These results illuminate new mechanisms likely to govern antibody trafficking at the brain-CSF interface with relevance for immune surveillance in the healthy brain and insights into the distribution of therapeutic antibodies. The precise mechanisms governing the central distribution of macromolecules from the cerebrospinal fluid (CSF) to the brain and spinal cord remain poorly understood, despite their importance for physiological processes such as antibody trafficking for central immune surveillance, as well as several ongoing intrathecal clinical trials. In the present study, we clarify how IgG and smaller single-domain antibodies (sdAb) distribute throughout the whole brain in a size-dependent manner after intrathecal infusion in rats using exvivo fluorescence and invivo three-dimensional magnetic resonance imaging. Antibody distribution was characterized by diffusion at the brain surface and widespread distribution to deep brain regions along the perivascular spaces of all vessel types, with sdAb accessing a four- to seven-fold greater brain area than IgG. Perivascular transport involved blood vessels of all caliber and putative smooth muscle and astroglial basement membrane compartments. Perivascular access to smooth muscle basement membrane compartments also exhibited size-dependence. Electron microscopy was used to show stomata on leptomeningeal coverings of blood vessels in the subarachnoid space as potential access points allowing substances in the CSF to enter the perivascular space. Osmolyte co-infusion significantly enhanced perivascular access of the larger antibody from the CSF, with intrathecal 0.75m mannitol increasing the number of perivascular profiles per slice area accessed by IgG by ∼50%. The results of the present study reveal potential distribution mechanisms for endogenous IgG, which is one of the most abundant proteins in the CSF, as well as provide new insights with respect to understanding and improving the drug delivery of macromolecules to the central nervous system via the intrathecal route.