Modeling neurovascular dysfunction in Alzheimer’s disease using an isogenic brain-chip model

The hiPSC line MC0117 (APOE ε3/ε3, apparently healthy individual) was obtained from the Mayo Clinic’s Neuroregeneration Lab within the Center for Regenerative Medicine. The hiPSC line AG27609 (APOE ε4/ε4, late-onset AD patient) was obtained from the Coriell Institute for Medical Research/National Institutes of Aging’s Aging Cell Repository. Colonies of hiPSCs were maintained in 6-well plates coated with Matrigel (Corning) in complete mTeSR Plus media (StemCell Technologies), with media changes performed every other day. Cells were passaged at a 1:6 ratio, when 80% confluency was reached, using ReLeSR passing reagent (StemCell Technologies).

Differentiation of hiPSCs into iBMVECs was performed following a previously published protocol [25]. Briefly, hiPSCs colonies were singularized with Accutase (Thermo Scientific) and then plated into Matrigel-coated T75 flasks at a density of 25,000 cells/cm² in complete mTeSR Plus media, in presence of 10 µM of the ROCK inhibitor Y-27632 dihydrochloride (StemCell Technologies) for 24 h. Then, media was changed to DeSR1 media that consisted of DMEM/F12 media (Gibco), 1% non-essential amino acids (Corning), 1% glutamine (Corning), 0.1 mM β-mercaptoethanol (Gibco), and 6 µM of the GSK3β inhibitor CHIR99021 (StemCell Technologies). Cells were cultured on DeSR1 media for 24 h and then changed to DeSR2 media (DeSR1 media supplemented with 1X B27 (Gibco) and without CHIR99021). Cells were cultured in DeSR2 media for 5 days with daily media changes. Then, cells were cultured for 3 days in hECSR1 media that consisted of human endothelial cell serum free media (hECSFM) (Gibco), 1% B27, 20 ng/mL basic fibroblast growth factor (bFGF) (Millipore), 10 µM retinoic acid (Sigma), with one media change performed after 48 h. Media was then changed to hECSR2 media that consisted of hECSR1 media without bFGF and retinoic acid and supplemented with 1% penicillin/streptomycin (Gibco) for 24 h. Then, differentiated iBMVECs were singularized and plated into T75 flasks coated with 400 µg/mL human collagen IV (Sigma Aldrich) and 100 µg/mL fibronectin (Sigma Aldrich) in hECSR2 media supplemented with Y-27632 dihydrochloride for 2 hours, and then media was replaced with hECSR2 and maintained in culture for 24 to 48 h.

Pericytes were differentiated following a previously described protocol [36]. Briefly, hiPSCs were singularized with Accutase, plated into Matrigel-coated 100 mm dishes at a density of 25,000 cells/cm² in complete mTeSR Plus media, supplemented with 10 µM Y-27632 and maintained in culture for 24 h. Cells were then cultured for 3 days in N2B27 media, consisting of 1:1 Neurobasal (Gibco) and DMEM/F12 media, 1% glutamine, 1% penicillin/streptomycin, 1X B27 and 1X N2 (Gibco), and supplemented with 25 ng/mL bone morphogenic protein 4 (BMP4) and 8 µM CHIR99021. After that, cells were cultured for 48 h in N2/B27 media supplemented with 10 ng/mL platelet-derived growth factor BB (PDGF-BB) (Peprotech), and 2 ng/mL Activin A with daily media change. Differentiated pericytes were then maintained in N2/B27 media for 24 to 48 h.

Microglia were differentiated from hiPSC adapting previously described methods [47, 48]. In short, hiPSCs were singularized with Accutase and plated at a density of 3 × 10⁶ cells per well on AggreWell 800 6-well plates (StemCell Technologies) in embryoid body media that consisted of mTeSR plus media supplemented with 10 µM Y-27632, 50 ng/mL BMP4, 20 ng/mL stem cell factor (Peprotech), and 50 ng/mL vascular endothelial growth factor 121 (Peprotech). Plates were spun down at 300 g for 3 min and maintained in culture for 2 days, before performing half media change. On day 4, embryoid bodies were collected and transferred to Matrigel-coated 6-well plates in 3 mL of hematopoietic media, consisting of X-VIVO 15 media (Lonza), supplemented with 1% glutamine, 1% penicillin/streptomycin, 55 µM β-mercaptoethanol, 100 ng/mL Macrophage Colony Stimulating Factor (Peprotech), and 25 ng/mL IL-3 (Peprotech). Cells were maintained in culture for 15 days, with half media changes performed every 5 days. After 15 days in culture, primitive macrophage precursor cells (PMPs) are produced and released to the media. On day 15, PMPs were harvested and plated in 6-well plates in RMPI 1640 media at a density of 180,000 cells/cm². Cells were allowed to attach for 1 h and then media was replaced with complete microglia media that consisted of RMPI 1640 media supplemented with 1% glutamine, 1% penicillin/streptomycin, 100 ng/mL IL-34 (Peprotech), and 10 ng/mL Granulocyte-Macrophage Colony-Stimulating Factor (Peprotech). PMPs were cultured for 10 days with media change performed every other day. Differentiated microglia were maintained in microglia media for up to 10 days.

Neural progenitor cells (NPCs) were differentiated from hiPSCs using the STEMdiff SMADi Neural Induction Kit (StemCell Technologies). Singularized hiPSCs were plated in Matrigel-coated 6-well plates at a density of 2.5 × 10⁵ cells/cm² in STEMdiff neural induction media (NIM) + SMADi supplement and 10 µM Y-27632. Daily media changes were performed for 6 days using STEMdiff NEM + SMADi. On day 6, cells were passaged to Matrigel-coated 6-well plates at a density of 2.0 × 10⁵ cells/cm² in STEMdiff NIM + SMADi supplement and 10 µM Y-27632. Daily media changes were performed for another 6 days using STEMdiff NEM + SMADi. Cells were passaged again on day 12 following the same procedure and cultured for an additional 6 days. On day 18, NPCs were passaged into Matrigel-coated 6-well plates at a density of 1.25 × 10⁵ cells/cm² in neural progenitor media (StemCell Technologies). NPCs were maintained in culture for 7 days with daily media change.

Neurons were differentiated from NPC using the STEMdiff forebrain neuron differentiation kit (StemCell Technologies). Briefly, NPCs were cultured in 6-well plates coated with 15 µg/mL poly-L-ornithine (PLO) (Sigma) and 5 µg/mL laminin (Sigma) in STEMdiff forebrain neuron differentiation media (StemCell Technologies) at a density of 125,000 cells/cm². NPCs were maintained in culture for 7 days with daily media Change to achieve neuronal differentiation. On day 7, differentiated neurons were passaged to PLO/laminin-coated 6-well plates at a density of 80,000 cells/cm² in STEMdiff forebrain neuron maturation media (StemCell Technologies). Neurons were cultured for 14 days with media changes performed every other day.

Astrocytes were differentiated from NPCs using the STEMdiff astrocyte differentiation kit (StemCell Technologies). Briefly, NPCs were cultured in Matrigel-coated 6-well plates at a density of 1.5 × 10⁵ cells/cm² in astrocyte differentiation media (StemCell Technologies). Cells were cultured for 7 days in STEMdiff astrocyte differentiation media with daily media change. Astrocyte precursor cells were passaged on days 7 and 14 and seeded in Matrigel-coated plates at a density of 1.5 × 10⁵ cells/cm² in astrocyte differentiation media. On day 21, differentiated astrocytes were passaged into Matrigel-coated plates at a density of 1.5 × 10⁵ cells/cm² in astrocyte maturation media (StemCell Technologies). Astrocytes were maintained in culture for 7 days with daily media change and then passaged twice at the same seeding density. Mature astrocytes are observed after 21 days in culture.

Primary antibodies (all from Thermo Fisher): Zonula occludens 1 (ZO-1) (33-9100, 1:1000 dilution), P-glycoprotein (P-gp) (MA1-26528, 1:250 dilution), claudin-5 (35-2500, 1:500 dilution), glial fibrillary acidic protein (GFAP) (PA1-10004, 1:2000 dilution), microtubule-associated protein 2 (MAP2) (13-1500, 1:1000 dilution), beta 3 tubulin (50-4510-82, 1:1000 dilution), ionized calcium binding adaptor molecule 1 (IBA1) (10904-1-AP, 1:500 dilution), platelet-derived growth factor receptor beta (PDGFRβ) (MA5-15143, 1:500 dilution).

Secondary antibodies (all from Jackson Immuno Research): Alexa Fluor 488 goat anti-rabbit (111-545-003, 1:5000 dilution), Rhodamine (TRITC) goat anti-mouse (115-025-166, 1:2000 dilution), Alexa Fluor 647 donkey anti-chicken (703-605-155, 1:5000 dilution), Alexa Fluor 594 goat anti-rabbit (111-585-144, 1:2500 dilution), Alexa Fluor 488 donkey anti-mouse (715-454-150, 1:2500 dilution).

Differentiation of hiPSC-derived cells was confirmed by immunocytochemistry. Differentiated cells were cultured in 8-well µ-slides (Ibidi), fixed with 4% paraformaldehyde (PFA), permeabilized with 0.2% tween20 and blocked with 1% bovine serum albumin (BSA). Characterization of iBMVECs was performed using antibodies against claudin-5, ZO-1, and P-gp. Pericytes were characterized using antibodies against PDGFRβ. Microglial cells were characterized with anti-IBA1 antibodies. Neurons were characterized using anti-MAP2, while astrocytes were characterized using anti-GFAP antibodies. Cells were then incubated with fluorescent secondary antibodies. Micrographs were acquired using a 20x objective in an inverted confocal microscope (Nikon ECLIPSE Ti2).

Differentiated cells were cultured on the Chip-S1 from Emulate Inc (Boston, MA USA). The Chip-S1 consists of two parallel fluidically independent microchannels made from polydimethylsiloxane (PDMS). The dimensions of the top channels are 1000 μm wide by 1000 μm height, and bottom channel’s dimensions are 1000 μm wide by 200 μm height. The overlapping co-culture area of the channels is 17.1 mm². Channels are separated by a 50 μm porous PDMS membrane, with a 7 μm pore size and a 40 μm pore separation. Before culturing cells, the surface of the membranes was activated by introducing a mixture of the proprietary ER1 and ER2 reagents into the brain (top) and vascular (bottom) channels and exposing them to 2-cycles UV light irradiation for 10 min. The vascular channel of the chip was then coated with a mixture of 400 µg/mL human collagen IV and 100 µg/mL fibronectin in hECSFM. The brain channel was coated with 15 µg/mL PLO and 5 µg/mL laminin in BrainPhys media (StemCell Technologies). Chips were incubated overnight at 37 ℃ and 5% CO₂. The following day, the vascular channel was washed with hECSR1 media and seeded with iBMVECs at a density of 8.5 × 10⁵ cells/cm² in hECSR1 media supplemented with 10 µM Y-27632. Chips were inverted immediately after seeding and incubated at 37 ℃ and 5% CO₂. Then, 2 hours later chips were returned to the upright position and sterile filter pipette tips were placed in the outlets of the channels and 200 µL of hECSR1 media were dispensed into the channels, leaving filter pipette tips on the inlet ports. Chips were returned to the incubator at 37 ℃ and 5% CO₂ for overnight incubation. The following day, the vascular channel was washed with 200 µL of hECSR2 media, leaving filter tips in the inlet and outlet ports. Then, the brain channel was washed with 200 µL of proprietary microglia-forebrain neuron-astrocyte triculture media -referred to as brain media- (StemCell Technologies) that consisted of complete the STEMdiff forebrain neuron differentiation media supplemented with Microglia Supplement 2 (StemCell Technologies). After that, differentiated neurons (45,000 cells/cm²), astrocytes (20,000 cells/cm²), microglia (7,500 cells/cm²) and pericytes (12,500 cells/cm²) were pooled into a sterile microtube and seeded on the brain channel using brain media. Sterile filter tips were left in the inlet and outlet ports of the brain channel and chips were returned to the incubator at 37 ℃ and 5% CO₂ for overnight incubation. The following day, both channels were washed with their respective media and chips were connected to the POD modules (Emulate Inc.). The modules consist of a chip-housing unit with reservoirs for cell culture media dispensing and effluent collection. Brain media was dispensed in the inlet reservoir of the brain channel and hECSR2 media was dispensed in the inlet reservoir of the vascular channel. Then, chips and PODs were connected to the Zoë Culture Module (Emulate Inc.), which was set to a flow rate of 20 µL/h for brain channel media and 60 µL/h for vascular channel media. Chips were incubated under flow conditions for 5 days, with daily media replenishment.

Paracellular permeability across the endothelial barrier was assessed by measuring the passage of 0.5 kDa Lucifer Yellow (LY) (Invitrogen), 3 kDa Dextran Cascade Blue (CB) (Invitrogen), and 10 kDa FITC-Dextran (Invitrogen) from the vascular channel media into the brain channel media by crossing the endothelial barrier. Tracers were dissolved in hECSR2 media at 20 µg/mL for 0.5 kDa LY, 100 µg/mL for 3 kDa Dextran CB and 10 kDa FITC-Dextran. Samples from the inlet and outlet reservoirs from both channels were collected daily for 5 days. Fluorescence intensity was read using a microplate reader (BioTek) using the following excitation/emission parameters: 485/528 nm for 10 kDa FITC-Dextran, 375/420 nm for 3 kDa Dextran CB, and 425/536 nm for 0.5 kDa LY. Apparent permeability was calculated using the following equation. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \begin{aligned} Papp = & \frac{{Q_{R} *Q_{D} }}{{SA*\left( {Q_{R} + Q_{D} } \right)}} \\ & *ln\left[ {1 - \frac{{C_{{R,0}} *\left( {Q_{R} + Q_{D} } \right)}}{{\left( {Q_{R} *C_{{R,O}} + ~Q_{D} *C_{{D,O}} } \right)}}} \right] \\ \end{aligned} $$\end{document}

where P_app is the apparent permeability in units of cm/s, SA is the surface area of the co-culture channel (0.17cm²), Q_R & Q_D are the fluid flow rates in the receiving and dosing channels, respectively, in units of cm³/s, and C_R,0 and C_D,0 are the recovered concentrations in the receiving and dosing channels, respectively. Three cell-free chips were used to obtain Papp values to use as reference.

Detection of cell-specific markers on the brain chip was assessed via immunocytochemistry. After 5 days in culture under flow conditions, chips were washed 3 times with DPBS and cells within both channels were fixed with 4% PFA, permeabilized with 0.2% tween20, and blocked with 1% BSA. The brain channel was incubated overnight at 4 ˚C with primary antibodies against MAP2, GFAP, IBA1 and PDGFRβ, while the vascular channel was incubated with antibodies against ZO-1. Chips were then incubated with fluorescent secondary antibodies for 2 h, counterstained with DAPI, and imaged using an inverted confocal microscope with a 10X magnification objective (Nikon ECLIPSE Ti2).

Protein was extracted from the brain and vascular channels of the brain-chips. After 5 days in culture, both channels were washed 3 times with 200 µL of sterile DPBS, which was aspirated after the last wash. Outlet ports of the chip were blocked with sterile pipette tips, then 60 µL of sterile DPBS was dispensed into the brain channel via the inlet port and 60 µL of complete radioimmunoprecipitation assay (RIPA) buffer (Thermo Scientific), supplemented with 1x protease and phosphatase inhibitors (Sigma) and 1 mM phenylmethylsulphonyl fluoride (Cell Signaling), were dispensed into the vascular channel. Chips were incubated at 4 ℃ for 30 min and sample was collected from the vascular channel. After that, DPBS was aspirated from the brain channel and replaced with 60 µL of RIPA buffer, while the vascular channel was filled with 60 µL of DPBS. Chips were incubated at 4 ℃ for 30 min and sample was collected from the brain channel. Protein concentration was then quantified by the bicinchoninic acid (BCA) method.

Protein samples (2.5 µg) from the vascular channel of the brain chips were used to quantify levels of the TJ protein claudin-5 (ABIN6962352), the membrane-bound receptor low-density lipoprotein receptor-related protein 1 (LRP-1) (ABIN6968348), and transporter P-gp (ABIN6958484) via ELISA kits from Antibodies Online. Kits were used according to manufacturer’s specifications and absorbance was read at 450 nm using a microplate reader. Protein values were interpolated in a standard curve and normalized by mg of protein.

Activity of the membrane-bound transporter P-gp was evaluated using the multidrug efflux transporter P-glycoprotein (MDR1/P-gp) ligand screening kit (ab284553) from Abcam. Briefly, media was aspirated from the brain and vascular channels and channels were washed twice with 100 µL warmed efflux assay buffer. Half of the chips were incubated with a fluorogenic P-gp substrate for 30 min at 37 ℃ and 5% CO₂, while the other half was first incubated for 30 min with 100 µM of the P-gp inhibitor verapamil, before incubation with a mixture of 100 µM verapamil and 1x fluorogenic P-gp substrate at 37 ℃ and 5% CO₂ for 30 min. Chips were then imaged using a Nikon ECLIPSE Ti2 confocal inverted microscope, under a stage-top incubator (Tokai Hit) maintaining physiological conditions of 37 ℃ and 5% CO₂. The entire vascular channel of the brain-chips was imaged with a 10X objective and a FITC filter. Fluorescence intensity of the entire vascular channel was quantified using NIS Element software (Nikon).

The Milliplex Human Amyloid Beta and tau magnetic bead panel (Millipore Sigma, HNABTMAG-68k) was used to quantify levels of Aβ1–40, Aβ1–42, p-tau181, and total tau present within protein recovered from the vascular and brain channels (7.5 µg/protein lysate), as well as media effluent from both vascular and brain outlet reservoirs (pooled media from 5 days). The multiplex kits were used according to manufacturer’s specifications and plates were read using the BioPlex 200. The concentrations of Aβ1–40, Aβ1–42, p-tau181 and total tau were interpolated using a standard curve and normalized by mg of protein.

The BioPlex Pro Reagent Kit III (BioRad, 71304090 M) with a customized Human Cytokine screening panel was used to assess cytokines present in pooled media samples from vascular and brain effluent collected for 5 days. Magnetic beads targeting IL-6, TNF-α, IL-1β, MCP-1, IFN- γ, IL-12, IL-15 and IL-17a (BioRad, 17007269) were used. The multiplex kits were used according to manufacturer’s specifications and plates were immediately read using the BioPlex 200 using the same parameters as the MilliPlex kit.

Data were obtained from two independent experiments that utilized cells from two independent differentiations. Each experiment consisted of N = 24 chips. Sample size for each individual endpoint is included in the respective figure legends. Data were analyzed using GraphPad Prism (v.10.0 for Windows, Boston MA USA). Normal distribution of the data was tested using the Shapiro-Wilk test. Comparisons of means of two experimental groups was performed using two-tailed t-test. For experiments with more than two groups, a one-way analysis of variance (ANOVA) was used followed by post-hoc Bonferroni multiple comparison test. For experiments with two factors, a mixed effect model was used to analyze the data, with hiPSC line (control and AD) and time (days) as fixed effects and individual chips as random effects, followed by Tukey’s test for multiple comparisons.

To construct the brain-chip model, hiPSCs were direct differentiated into neurons, astrocytes, pericytes, microglia and iBMVECs. Differentiation was confirmed by the expression of cell-specific markers for each cell type (Fig. S1). Differentiated iBMVECs were cultured on the vascular channel of brain-chips, while differentiated neurons, astrocytes, microglia, and pericytes were cultured on the brain channel of the chip (Fig. 1A). Homogeneous distribution of the cultured cells was observed along both channels. Formation of a continuous monolayer of iBMVECs on the vascular channel was confirmed by immunostaining with the TJ protein ZO-1 (Fig. 1B). The presence of neurons, astrocytes, microglia, and pericytes along the brain channel was confirmed by immunostaining with the cell specific markers MAP2 (Fig. 1C), GFAP (Fig. 1D), IBA1 (Fig. 1E) and PDGFRβ (Fig. 1F), respectively. There were no obvious morphological differences between cells differentiated from the hiPSC lines MC0117 (control) and AG27609 (AD) (Fig. 2), as evidenced by the staining patterns for the proteins ZO-1 (Fig. 2A and B), P-gp (Fig. 2A and B), claudin-5 (Fig. 2C and D), MAP2 (Fig. 2E and F), GFAP (Fig. 2E and F), IBA1 (Fig. 2G and H), and PDGFRβ (Fig. 2I and J).

Paracellular permeability to different size molecules was assessed as a measure of BBB function during a 5-day period of culture under flow conditions. As expected, very high values of Papp were observed on cell-free chips, being 1.7 × 10⁻⁵, 1.85 × 10⁻⁵ and 2.09 × 10⁻⁵ for 10 kDa dextran, 3 kDa dextran and 0.5 kDa lucifer yellow, respectively. There was no interaction between hiPSC line (control and AD) and time for, 0.5 kDa lucifer yellow (Fig. 3A) [F (4, 127) = 1.066, p = 0.3763], 3 kDa dextran (Fig. 3B) [F (4, 127) = 1.280, p = 0.2814], or 10 kDa dextran (Fig. 3C) [F (4, 127) = 1.489, p = 0.2092]. Paracellular permeability increased overtime for 0.5 kDa lucifer yellow (Fig. 3A) [F (2.611, 82.89) = 37.84, p < 0.0001], 3 kDa dextran (Fig. 3B) [F (2.363, 75.03) = 26.96, p < 0.0001], and 10 kDa dextran (Fig. 3C) [F (2.240, 71.12) = 25.58, p = < 0.0001]. Notably, the increased permeability was only significant between days 1 and 2 for all tracers. There was a significant effect of hiPSC line (control vs. AD) on paracellular permeability for, 0.5 kDa lucifer yellow (Fig. 3A) [F (1, 32) = 10.85, p = 0.0024], 3 kDa dextran (Fig. 3B) [F (1, 32) = 35.49, p < 0.0001] and 10 kDa dextran (Fig. 3C) [F (1, 32) = 21.12, p = < 0.0001], with AD brain-chips presenting a significant increase in permeability for the 3 tracers at all time points, with the exception of day 4, for 0.5 kDa lucifer yellow (p = 0.2658) and 10 kDa dextran (p = 0.1487). Analysis within hiPSC line demonstrated size-dependent permeability for both, control brain-chips (Fig. 3D) [F (2, 45) = 67.75, p < 0.0001)] and AD brain-chips (Fig. 3E) [F (2, 51) = 39.40, p < 0.0001)], with higher permeability values observed for lower molecular weight tracers. Alterations in paracellular permeability are often related to decreased expression of TJ proteins. Therefore, we quantified levels of the TJ protein claudin-5 in protein lysates from the vascular channel of the brain-chips (Fig. 3F). We observed a decrease in claudin-5 expression in AD brain-chips compared to control brain-chips (p = 0.0230). Expression of ZO-1 was also observed in AD brain-chips compared to control brain-chips (p = 0.0103) (Supplementary Fig. 2). These data indicate that AD brain-chips present increased paracellular permeability to different size molecules that was related with a decrease in the expression of TJ proteins.

Because impaired Aβ clearance is implicated in AD, we quantified levels of Aβ transporters LRP1 (Fig. 4A) and P-gp (Fig. 4B) in protein lysates from the vascular channel. We found decreased expression of LRP1 (Fig. 4A) in AD brain-chips, compared to control brain chips (p = 0.0031), while the expression of P-gp remained unchanged (p = 0.1342). Then, we assessed the activity of P-gp on iBMVECs cultured on the vascular channel. To that end, we inhibited P-gp function by incubating the vascular channel of the brain-chips with 100 µM verapamil and then adding a fluorescent P-gp substrate that will diffuse into the iBMVECs. When functioning properly, P-gp will actively transport the substrate out of the cells, leading to a decrease in fluorescence signal (Fig. 4C and D). Activity of P-gp was decreased in AD brain-chips, compared to control (p = 0.0334). As expected, verapamil decreased P-gp activity, compared to untreated chips, in both control (p < 0.0001) and AD-brain chips (p = 0.0313), although the difference was smaller in AD brain-chips. These data suggest that mechanisms involved in Aβ clearance are altered in AD brain-chips from this particular donor.

Pathological hallmarks of AD include increased production of Aβ and tau. We quantified the amount of Aβ40, Aβ42, total tau and pTau 181 in protein lysates and cell culture media from the brain and vascular channels of the brain-chips. Aβ40 levels were below the limit of detection for all samples analyzed. Surprisingly, we observed decreased levels of Aβ42 (Fig. 5A, p < 0.0001), total tau (Fig. 5C, p < 0.0001) and pTau 181(Fig. 5E, p < 0.0001) in protein lysates obtained from the brain channel of AD brain-chips. Levels of total tau were also decreased in the media effluent of AD brain-chips compared to control (Fig. 5H, p = 0.0080), while the levels of Aβ42 (Fig. 5G) and pTau 181 (Fig. 5J) were not different between the two groups. On the other hand, a completely different pattern on the levels of these proteins was observed in samples obtained from the vascular channel of the brain-chips. We observed a significant increase in the levels of total tau (Fig. 5D, p = 0.02750) and pTau 181 (Fig. 5F, p = 0.0331) in protein lysates from the vascular channel of AD brain-chips. Levels of total tau were increased (Fig. 5I, p = 0.0476), while pTau 181 (Fig. 5K, p = 0.0009) were decreased in effluent media collected from the vascular channel of AD brain-chips. For Aβ42, there was no difference between the two groups on protein lysates from the vascular channel (Fig. 5B, p = 0.6745) and its values were below the limit of detection for effluent media from the vascular channel. These data indicate that brain-chips constructed with hiPSCs from this AD patient have lower levels of Aβ and tau in the brain channel, although vascular tau accumulation was detected.

Neuroinflammation is also recognized as a contributing factor to AD pathology. We quantified levels of inflammatory markers in effluent media from the brain and vascular channels of the brain-chips. Levels of TNF-α, IL-1β, IFN-γ, IL-12, IL-15 and IL-17a were below the limit of detection for all samples analyzed. We observed increased levels of MCP-1 on brain (Fig. 6A, p < 0.0001) and vascular (Fig. 6, p = 0.0365) effluents from AD brain-chips. Similarly, levels of IL-6 were also increased in brain (Fig. 6C, p < 0.0001) and vascular (Fig. 6D, p = 0.0009) effluents from AD brain-chips. These data suggest that increased production of some neuroinflammatory markers was observed in AD brain-chips for this specific hiPSC line.