Blood-brain-barrier organoids for investigating the permeability of CNS therapeutics

Abstract

In vitro models of the blood-brain barrier (BBB) are critical tools for the study of BBB transport and the development of drugs that can reach the CNS. Brain endothelial cells grown in culture are often used to model the BBB; however, it is challenging to maintain reproducible BBB properties and function. 'BBB organoids' are obtained following coculture of endothelial cells, pericytes and astrocytes under low-adhesion conditions. These organoids reproduce many features of the BBB, including the expression of tight junctions, molecular transporters and drug efflux pumps, and hence can be used to model drug transport across the BBB. This protocol provides a comprehensive description of the techniques required to culture and maintain BBB organoids. We also describe two separate detection approaches that can be used to analyze drug penetration into the organoids: confocal fluorescence microscopy and mass spectrometry imaging. Using our protocol, BBB organoids can be established within 2-3 d. An additional day is required to analyze drug permeability. The BBB organoid platform represents an accurate, versatile and cost-effective in vitro tool. It can easily be scaled to a high-throughput format, offering a tool for BBB modeling that could accelerate therapeutic discovery for the treatment of various neuropathologies.

Figure 1|. Characterization of each cell type used for establishing BBB organoids.

a) Fluorescence images showing the expression of CD31 (Abcam, dilution: 1:100) and VE-Cadherin (Cell Signaling Technology, dilution: 1:100) (endothelial cell markers) in primary human brain microvascular endothelial cells (HBMEC) (ScienCell). b) (Top) Bright field image showing the morphology of immortalized human cerebral microvascular endothelial cells (hCMEC/D3) (Cedarlane), and (middle) fluorescence image showing the expression of CD31. c) (Top) Bright field image of primary human brain vascular pericytes (ScienCell) and (middle) fluorescence image showing the expression of the pericyte marker NG2 (Millipore, dilution: 1:100). d) (Top) Bright field image of primary astrocytes (Lonza) and (middle) fluorescence image showing the expression of astrocyte marker GFAP (Sigma-Aldrich, dilution: 1:100). (b-d, bottom panels) Cells incubated with secondary IgG only (no primary IgG) were used as negative controls. Nuclei of cells were stained with the Hoechst dye (blue). Secondary antibody anti-rabbit Alexa Fluor 546 (red) or anti-mouse Alexa Fluor 488 (green) was used. Scale bar: 100 μm (Bright field: 20x objective; confocal: 40x objective).

Figure 2|. Culture, collection and assessment of BBB organoids.

a) Formation of multicellular BBB organoids in a 96-well plate by co-culturing hCMEC/D3 (Cedarlane), pericytes (ScienCell) and astrocytes (Lonza) in BBB working medium. Samples of each cell type were validated by immunofluorescence staining prior to co-culturing, as shown in figure 1. Wells are coated with 1% (w/v) agarose to form a concave low-attachment surface. The organoids assemble in the center of each well (blue arrows). b) Representative bright field images showing examples of acceptable and unacceptable organoids on the basis of their physical appearance. In our experiments, out of a total of 1,056 organoids, 950 were considered acceptable, yielding an average success rate of 90% (cell numbers are pooled from 11 independent organoid cultures). Scale bar: 100 μm (10x objective). c) Organoids (indicated by cyan arrows) are collected in a microcentrifuge (0.2-mL PCR) tube in BBB working medium. (d) Fluorescence image showing the expression of tight junction marker (ZO-1; red) (Life Technologies, dilution: 1:100) on the organoid surface. (e) Fluorescence image showing the expression of P-glycoprotein (P-gp; red) efflux pump (Abcam, dilution: 1:100) on the organoid surface. Nuclei of organoids are stained with Hoechst dye (blue). As a negative control, organoids are stained with secondary antibody anti-rabbit Alexa Fluor 546 (red) or anti-mouse Alexa Fluor 488 (green) only (no primary antibody). Scale bar: 50 μm (40x objective).

Figure 3|. Quantification of average fluorescence intensity in a BBB organoid.

a) Confocal fluorescence z-stack images of a BBB organoid (at a depth between 56–96 μm), showing the accumulation of a known BBB-permeable compound, angiopep-2^, (5 μM concentration, incubated for 5 h at 37°C) within the organoid core. The core area 50 μm from the organoid surface (blue dotted circle) is defined as the region of interest. Permeability of the compound is determined by measuring the total MFI within the core area of these optical sections. b) Orthogonal view of a organoid (from panel a), showing the core region to be quantified (blue dotted lines). Within the z-direction, quantification is performed at 56 μm depth from the surface up to 96 μm depth (yellow dotted lines and arrows). Scale bar: 100 μm (20x objective).

Figure 4|. Analysis of angiopep-2 permeability using the BBB organoid platform by confocal microscopy.

a) Confocal fluorescence z-stack images showing the transport of rhodamine-labeled angiopep-2 (2711.6 Da), rhodamine-scramble peptide (2711.6 Da), dextran-rhodamine (4400 Da) and unconjugated rhodamine dye (380.8 Da) into the BBB organoids. Organoids were established using hCMEC/D3, pericytes and astrocytes, and then, incubated with each compound (10 μM concentration) at 37°C for 5 hrs. Each organoid was imaged by confocal fluorescence microscopy (using a 20x objective), and the mean fluorescence intensity in the organoid core was quantified using ImageJ. b) Orthogonal view of the organoids showing the distribution of the compounds. White dotted lines indicate organoid border. Scale bar: 100 μm. c) Bar graph displaying the quantification of the total mean fluorescence intensity of the optical sections (from a)). Statistical analysis was performed using One-way ANOVA and Tukey’s multiple comparisons test. Graph depicts the sum of the mean fluorescence intensity of the core area from optical sections at a depth of 56 – 96 μm with standard deviation error bars (**** p< 0.0001; *** p<0.001; n_organoids = 5). d) Graph depicting the transport of Bip(1) in a concentration-dependent manner after 2 hrs of incubation (n_spheroid = 5–7). e) Time course analysis of the transport of Cy5-Bip(1) into the BBB organoid over 26 hours (at 10 μM concentration) (n_spheroid = 4–6). The graphs display mean Cy5.5 fluorescence intensity quantified at 88 μm depth from the surface of each spheroid with SD error bars.

Figure 5|. Setup and analysis of BKM120 and dabrafenib permeability by MALDI-MSI.

a) Setup of the cryostat for sectioning of BBB organoids. Organoids were collected at the bottom of a PCR tube and snap frozen in a dry ice/ethanol bath. The cap of the tube was then mounted onto the sample disc. b) Conductivity was measured on both sides of an indium tin oxide (ITO) coated glass slide with a voltage-meter to determine the correct ITO coated side for mounting of tissue sections. ITO coating resulted in a measurable electrical resistance on the surface of the slide, while the uncoated side was non-conductive. For MSI, cryo-sectioned tissues must be mounted onto the ITO-coated side. c) H&E stained images of the cryo-sections from each group confirmed the presence of BBB organoid tissues (n_organoid = 150). Serial sections were used for MSI. ‘Scan’ images were used to identify the position of organoid sections on the ITO slide. ‘Pixel’ images show the pixel distribution within the selected scanned region (spatial resolution of 30 μm). MALDI MSI ion images reveal the distribution/level of both drugs in the organoid sections, showing a high level of BKM120 accumulation within the organoids (in green m/z 411.1751 ± 0.001). Dabrafenib was not detected within the organoids (m/z 520.1083 ± 0.001). Scale bars for BKM image: 100 μm, Scale bar for Dabrafenib image: 200 μm. Images from panel (c) are adapted from Cho et al., Nature Communications (2017).