Water influx into cerebrospinal fluid is primarily controlled by aquaporin-4, not by aquaporin-1: 17O JJVCPE MRI study in knockout mice

Abstract

Recent studies on cerebrospinal fluid (CSF) homeostasis emphasize the importance of water flux through the pericapillary (Virchow-Robin) space for both CSF production and reabsorption (Oreskovic and Klarica hypothesis), and challenge the classic CSF circulation theory, which proposes that CSF is primarily produced by the choroid plexus and reabsorbed by the arachnoid villi. Active suppression of aquaporin-1 (AQP-1) expression within brain capillaries and preservation of AQP-1 within the choroid plexus together with pericapillary water regulation by AQP-4 provide a unique opportunity for testing this recent hypothesis. We investigated water flux into three representative regions of the brain, namely, the cortex, basal ganglia, and third ventricle using a newly developed water molecular MRI technique based on JJ vicinal coupling between O and adjacent protons and water molecule proton exchanges (JJVCPE imaging) in AQP-1 and AQP-4 knockout mice in vivo. The results clearly indicate that water influx into the CSF is regulated by AQP-4, and not by AQP-1, strongly supporting the Oreskovic and Klarica hypothesis.

Fig. 1

Phenotypic appearance. Wild-type (WT) and knockout (KO) mice were virtually indistinguishable phenotypically.

Fig. 2

Schematic diagram of pulse sequences. Pulse sequence consists of double spin-echo with hyperbolic secant pulse (HSP) followed by 31 consecutive refocusing pulses using B1 insensitive refocusing (BIR4) pulses. AHP, adiabatic half passage pulse.

Fig. 3

Region of interest (ROI). (a) Sagittal scout image: Imaging slab was set to 6 mm caudal from the top of the cerebrum. (b) Fast spin-echo image: ROIs were selected semi-automatically using image processing software. BG, basal ganglia; CSF, cerebrospinal fluid.

Fig. 4

Decay curve fitting. Intensities at the steady state of each area, expressed as % against the averaged intensity of identical pixel before administration of H₂¹⁷O, were determined by fitting their time course by the function: I=I₀+ae^−bt. I₀ denotes the normalized signal intensity (SI) at infinite time (t=∞) calculated from the fitted curve. i.v., intravenously.

Fig. 5

Representative time course. Representative time course of signal intensities (SIs) within pixels of each region of interest (ROI) shown in Fig. 3 following intravenous (i.v.) H₂¹⁷O administration in WT (a), AQP-1-KO (b), and AQP-4-KO (c) mice. AQP-1 (−/−), AQP-1-KO mouse; AQP-4 (−/−), AQP-4-KO mouse; KO, knockout; WT, wild-type. Blue, cortex; red, basal ganglia (BG); green, cerebrospinal fluid (CSF) within the third ventricle. Each dot represents the intensity of each pixel within the ROI.

Fig. 6

I₀ of three regions of interest. Values of I₀ in the cortex and basal ganglia (BG) are virtually identical among the three groups. In contrast, I₀ of cerebrospinal fluid (CSF) within the third ventricle is significantly higher in AQP-4-KO mice compared with AQP-1-KO and WT mice. I₀ of CSF within the third ventricle in AQP-1-KO mice is virtually identical to WT mice. *P<0.01 vs. WT, ^§P<0.01 vs. AQP-1 (−/−). AQP-1 (−/−), AQP-1-KO; AQP-4 (−/−), AQP-4-KO mice; KO, knockout; WT, wild-type.