Proceedings of The Physiological Society

University of York (2002) J Physiol 539P, S020


Quantitative evidence for production and convective flow of brain interstitial fluid

I.R. Mihaylov, P. Batchelor* and N.J. Abbott

BBB Group, Centre for Neuroscience Research and *Department of Radiological Sciences, King's College London, London SE1 1UL, UK

The classical concept of brain interstitial fluid (ISF) turnover proposed the presence of convective (bulk) flow of ISF from perineuronal spaces to the cerebrospinal fluid (CSF) compartment via perivascular spaces (Weed, 1914), acting as an effective ‘lymphatic' system for the CNS (Weller et al. 1992). The pattern of clearance of intracerebral tracers gave evidence for ISF flow (Cserr et al. 1981). This study aimed to quantify ISF flow at the microscopic level.

Experiments accorded with UK legislation. Four adult female Wistar rats were anaesthetised (250 mg ketamine I.P.). 20 µl of artificial CSF containing tracer Evans Blue-albumin (EBA, 65 mg ml-1) was stereotactically infused in the grey-white matter junction of the frontal lobe of over a period of 1 h. Animals were humanely killed 5 h later and the brains removed, frozen and serially sectioned. Coronal sections were immerse-fixed in cold 10 % formaldehyde and counter-stained with 1 % neutral red, photographed and digitised. Tracer distribution was similar in the four brains. The digital images of one rat brain were used to construct a 3-dimensional model of tracer spread using in-house registration/alignment software. The brain interstitial distribution of EBA was quantified from regional optic density measurements using Analyze 3.1 software (Biomedical Imaging Resources, Mayo Clinic, USA). The relative optic density was resolved as an exponential function of distance by regression analysis. Comparison with the theoretical distribution for ‘high flow microinfusion' (Morrison et al. 1994) was made by dividing the experimental by the theoretical exponential function to give a ratio R; the effect of post-infusion diffusion was also considered.

R values for tracer movement in the medio-lateral and anterior-posterior directions were >> 1, evidence for the influence of production and bulk flow of ISF. For movement towards the ventricles and subarachnoid space, R was 522 and 320, respectively. Although tracer was seen tracking along white matter tracts, R for cortical grey matter was ~3.6 times greater than for white matter (792 and 222, respectively).

The results are compatible with endogenous secretion of ISF by brain capillaries, with flow towards both ventricles and subarachnoid space, and greater ISF secretion in grey compared with white matter, although white matter provides routes of low resistance for fluid flow. This study has implications for CNS tissue clearance pathways, routes of malignant cell dissemination, and drug delivery strategies.

I.R.M. was supported by a Physiological Society Vacation Studentship.

Where applicable, experiments conform with Society ethical requirements