Nembutal

The computational mesh used for the simulations in this paper was constructed from the “Waxholm Space Atlas of the Sprague Dawley Rat Brain v4” (RRID: SCR_017124) [73 (link)–75 ], available under the licence CC-BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/) at https://www.nitrc.org/projects/whs-sd-atlas. The atlas provides a detailed segmentation of different regions within the rat brain.
In the original study behind the atlas [73 (link)], the animal was anaesthetized by intraperitoneal injection of a mixture of Nembutal (Ovation Pharmaceuticals, Inc., Lake Forest, IL) and butorphanol, and transcardially perfused with 0.9% saline and ProHance (10:1 v:v) for 4 minutes followed by a flush of ProHance in 10% phosphate-buffered formalin (1:10 v:v). All procedures and experiments in their work were approved by the Duke University Institutional Animal Care and Use Committee [73 (link)].
Since the models in this paper do not separate between tissue from different regions of the brain, the segmentation is mainly of interest for removing unwanted sections. Most importantly, we wanted to remove the segments representing various parts of the ventricles. Moreover, we removed some external artefacts such as the spinal trigeminal tract, the optic nerves, and parts of the auditory system [74 (link)].
The various segments in the raw data file have a few irregularities. For example, in regions where the lateral ventricles are very thin, small groups of unlabeled voxels create holes in the 3D reconstruction of the ventricles. To repair these irregularities, we have made use of 3D Slicer (https://www.slicer.org/), an open-source software application for visualization and analysis of medical images [76 (link)]. 3D Slicer provides a segment editor with tools for manual labelling of voxels, hole filling and surface smoothing. After refining the segmentation of the ventricular system, it may be removed from the original volume to create a realistic representation of the brain surface. The surface is exported as an stl-file to be used in the meshing algorithm.
The creation of the computational mesh is performed by SVMTK (https://github.com/SVMTK/SVMTK), which provides a python API for 3D mesh generation methods from the CGAL library. The mesh generation algorithm consists of a Delaunay refinement process followed by an optimization phase [77 ]. Following the procedures described in [78 ], we created the mesh illustrated in Fig 2a.
To solve the Eqs (1) and (3), we use the finite element method for the discretization in space and an implicit Euler method to integrate the resulting ordinary differential systems in time.
In this paper, we choose a resolution for the spatial mesh of h = 1/32. The temporal domain is [0, T] with T = 360min with a time step of Δt = 1min. Details of the mesh and time resolutions can be found in Appendix C.2 in S3 Appendix. The numerical scheme has been implemented using the FEniCS Library [79 , 80 ], and the linear system was solved using the generalized minimal residual method (GMRES) and the incomplete LU (ILU) preconditioner. Our code is publicly available on GitHub at the following link: https://github.com/jorgenriseth/multicompartment-solute-transport.

Immunohistochemistry was performed using a standard protocol [20 (link),21 (link)]. In brief, mice were anesthetized with an i.p. injection of Nembutal (pentobarbital sodium, Dainippon Pharmaceutical Co., Ltd., Osaka, Japan) at 0.5 mg/kg and transcardially perfused with phosphate-buffered saline (FUJIFILM Wako), followed by 4% paraformaldehyde phosphate buffer solution (FUJIFILM Wako). The brains were immersed in 4% paraformaldehyde for 24 h before being transferred to a 30% sucrose solution (FUJIFILM Wako) for 24 h. Coronal brain sections of 40 μm thickness were cut using a cryostat (Carl Zeiss MicroImaging GmbH, Jena, Germany) after the brains were rapidly frozen in OCT compound (Sakura Finetek, Torrance, CA, USA). Microglia were detected in the prefrontal cortex (PFC) and hippocampus slices dissected from frozen brains using the rabbit polyclonal anti-mouse ionized calcium-binding adapter molecule 1 (Iba-1) antibody (FUJIFILM Wako, 019-019741). Microglial TNF-α expression was confirmed using the goat polyclonal anti-mouse TNF-α antibody (R&D Systems, Minneapolis, MN, USA. AB-410-NA). The secondary antibodies used were Alexa Fluor™ 488-conjugated anti-rabbit IgG (1:300; Invitrogen, Carlsbad, CA, USA) and Alexa Fluor™ 594 anti-goat IgG (1:300; Invitrogen). The nuclei in the slices were stained with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen). Cell images were acquired using a fluorescence microscope (Axio Scope.A1; Carl Zeiss, Oberkochen, Germany). The levels of Iba-1 signals were obtained using ImageJ 1.53K software (NIH Image, Bethesda, MD, USA).

In this study, we used tissue from 10 adult normal monkeys (Table 2) that had been kept in cryoprotection in Dr. Rausell′s tissue bank in the Department of Anatomy, Histology, and Neuroscience, Medical School, Autonomous University of Madrid (Madrid, Spain).
The tissue preparation has been described in detail in previous studies [49 (link)]. Briefly, monkeys were anesthetized with intramuscular ketamine and given an overdose of intravenous Nembutal. They were then perfused through the ascending aorta with normal saline followed by a solution of 4% paraformaldehyde and 1% glutaraldehyde in phosphate buffer (PB, 0.1 M, pH 7.4). The brain was removed and blocked. All the blocks were subsequently postfixed in 4% paraformaldehyde for 4 h, infiltrated with 30% sucrose in 0.1 M PB at 4 °C with gentle agitation, frozen in dry ice, and stored at −80 °C. The frozen blocks of the flattened cortex were cut tangential to the pia mater into 25–30 μm thick sections in a freezing sliding microtome. Alternate series of sections were collected in a sterile cryoprotectant solution. These series would later be processed for Nissl staining, immunohistochemistry and double immunofluorescence for MCT8/OATP1C1 and a number of cell/vessel markers.