Construction of an In Vitro Blood–Brain Barrier Micro-Organoid Model Using Decellularized Squid Mantle Scaffold Film

TL;DR

Researchers created a new in vitro model of the blood-brain barrier using a squid mantle scaffold to better study neurological diseases and drug permeability.

Contribution

A novel BBB micro-organoid model using decellularized squid mantle scaffold films for a more authentic NVU microenvironment.

Findings

The DSMS film supported the growth of endothelial and astrocyte cells, forming a BBB-like structure.

The model showed functional barrier properties with high TEER and low permeability to macromolecules.

LRP1 expression was used to assess model integrity and response to thrombotic conditions.

Abstract

Although blood–brain barrier (BBB) models are of great value in investigating neurological diseases, the structural complexity and intricate function based on cell–cell interactions of the BBB bring various limitations to the applications of existing models. In this study, a novel BBB micro-organoid model was established by culturing neurovascular unit (NVU) cells on a decellularized squid mantle scaffold (DSMS) film to reconstitute a more authentic and reliable NVU microenvironment for in vitro research. The DSMS applied was obtained from squid mantle scaffolds via decellularization, followed by defatting, and showed good biocompatibility with no cytotoxicity. The DSMS film was finally prepared by lyophilization. The lyophilized film exhibited a void ratio and pore size suitable for the adhesion and growth of endothelial cells (hCMEC/D3) and astrocytes (hACs), which led to the formation of a BBB-like spatial structure. The BBB micro-organoid model exhibited functional barrier properties, including an effective transendothelial electrical resistance (TEER) of approximately 230 Ω/cm2, restricted permeability to macromolecules—with apparent permeability coefficients (Papp) of 6.3 × 10−7 cm/s for 10 kDa and 2.7 × 10−7 cm/s for 70 kDa FITC–dextran—and expression of tight junctional complex (TJC) proteins such as vascular endothelial cadherin (VE-cad) and Zonula Occludens-1 (ZO-1). Furthermore, low-density lipoprotein receptor-related protein 1 (LRP1), a key receptor stably expressed in these two NVU cell types, was utilized as a critical indicator to assess the integrity of the BBB micro-organ model and its responsiveness to pathophysiological stimuli, particularly under thrombotic conditions. This study not only validates the feasibility of constructing a functionally competent BBB micro-organ model using DSMS films integrated with NVU cells but also provides a promising in vitro platform for subsequent studies on the BBB-related pathological mechanisms and the evaluation of drug permeability across the BBB.

Figures

Full text

1. Introduction

The blood–brain barrier (BBB) is the interface between the blood and the brain parenchyma. Its unique cellular structure comprises endothelial cells lining vessels, pericytes embedded within the basement membrane, and astrocytes whose endfeet ensheathe the abluminal vessel surface [1]. As a key functional component of the NVU, the BBB depends on the synergy of multiple cell populations within the NVU to maintain structural integrity and functional specificity [2]. The extensive capillary network in the human brain constitutes approximately 85% of the length of the brain’s vasculature and cover an endothelial surface area of 12 m^2^, which is pivotal for molecular exchanges [3]. Concurrently, the brain harbors intricate neural networks that engage in highly sophisticated interactions with cerebral vessels. In the BBB, wedge-shaped endothelial cells line inner blood vessels, forming extensive tight-junction proteins [4]. These proteins, together with receptors, transporters, efflux pumps and other cellular components, control molecular entry into and exit from the brain. The neurovascular unit refers to a multicellular unit in the brain, including vascular cells [e.g., endothelium and mural cells including pericytes and smooth muscle cells (SMCs)], glia (e.g., astrocytes, microglia), and neurons [5]. The BBB, constructed and regulated by the NVU cells (such as microvascular endothelial cells, astrocytes and pericytes), is the core structure maintaining central nervous system (CNS) homeostasis, with cell–cell interaction forming an essential part of its barrier integrity and function [6]. The NVU’s homeostatic balance is characterized by the BBB’s selective permeability [7]. Together, the NVU and BBB constitute a sophisticated defense system that protects the CNS and ensures its physiological functions [8]. However, this unique structure also limits the passage of therapeutic drugs [9]. Additionally, the BBB’s “intelligent” selective permeability hinders the development of highly realistic, credible in vitro BBB models [10].

To date, the main in vitro BBB models in laboratories are primarily divided into three types: Transwell models based on chambers, chip models with peristaltic pumps providing flow shear stress, and organoid models derived from pluripotent stem cells [11]. However, each of these three models has its own advantages and disadvantages [12]. The Transwell model has proven stable and reliable for in vitro BBB research over the past decade. It is mainly constructed with chambers based on PET or PC membranes: endothelial cells are seeded onto the semipermeable membrane to mimic the vascular side, while neural cells are cultured at the bottom of the lower chamber to mimic the brain side [13,14]. Despite its reproducibility and ease of use, the traditional Transwell system has limited ability to replicate microenvironment complexity [15]. The chemically composed semipermeable membrane cannot fully mimic the BBB basement membrane’s composition. PDMS-based microfluidic (chip) models have advanced significantly in recent years [16,17]. They reconstruct the BBB microenvironment using hydrogel systems, multicellular structures, and physiological mechanical forces (e.g., shear stress) [18], accurately mimicking the in vivo BBB environment [19]. However, their complex operation, high cost, steep learning curve, and inability to perform high-throughput tests hinder widespread adoption. The third type of model utilizes pluripotent stem cells, e.g., induced pluripotent stem cells (iPSCs) and embryonic stem cells ESCs, which can differentiate into key BBB cell types such as brain microvascular endothelial cells, astrocytes, and pericytes. These components self-assemble into the BBB spheroid models [20,21]. Compared to the standard Transwell models, this model better mimics in vivo physiological conditions. It also has self-assembly ability: after pluripotent stem cell differentiation, the BBB components assemble spontaneously. However, as a relatively new technology, it is not yet mature. The BBB spheroids differ in size and cellular composition between batches, reducing experimental reproducibility [22]. Notably, most pluripotent-stem-cell-derived BBB spheroid models currently lack in vivo-like BBB basement membrane components. This deficiency makes the model unable to mimic the mechanical support and signal transduction functions of the native BBB basement membrane [23]. To address these limitations, this study aimed to develop a novel and practical in vitro BBB model capable of reconstructing a tightly connected, in vivo-like microenvironment while incorporating a functional basement membrane structure.

Acellular scaffolds are widely used in tissue engineering. Their core advantage is simulating the structure and function of natural extracellular matrix (ECM) [24]. They also have good biocompatibility and controllability [25]. Most of these scaffolds are derived from mammals, including decellularized tissues from humans, pigs and cattle. However, these mammalian-derived acellular matrices have limitations [26], such as limited donor sources and ethical and religious concerns. In contrast, acellular marine collagen scaffolds can overcome these drawbacks [27]. Cytotoxicity testing of material extracts was performed following the ISO 10993-5 standard [28,29]. Leachates from the film could release residual components. These components may affect the proliferation, migration, and survival of endothelial cells and ACs [26,30]. These cellular behaviors are crucial for establishing a BBB model [6]. Therefore, the effects of the DSMS film extracts on hCMEC/D3 cells and hACs were examined. The DSMS film serves as a biomimetic basement membrane for the BBB modeling due to its compositional homology with the native structure. As a collagen-rich (types I/IV) natural connective tissue, it mimics the core ECM environment essential for barrier function [31]. Due to these properties, sheet-like DSMS has attracted widespread attention in recent years [32].

Based on this, we propose utilizing the DSMS film to mimic the BBB basement membrane to support the NVU cells in constructing a highly simulated in vitro organ model of the BBB [33]. The most adaptable cell lines including hCMEC/D3 and SVG p12 were selected to establish the in vitro BBB micro-organ model in this study, as they highly recapitulate the two most critical cell types of NVU cells—brain microvascular endothelial cells and astrocytes. After the model construction, efficacy evaluation was performed using the standard validation system for in vitro BBB models: TEER assay and immunofluorescence staining were combined. TEER measurements confirmed that the model’s barrier function met core criteria for in vitro BBB models, while immunofluorescence clearly showed specific expression of tight-junction proteins [e.g., claudin-5 (CLDN-5), ZO-1], verifying the model’s reliability from both functional and structural perspectives. LRP1 is not only a key functional receptor on BBB endothelial cells, but is also stably expressed in hACs and involved in BBB regulation [34]. Its overall expression and distribution can directly reflect the physiological state of the barrier. Based on this, we propose a new BBB detection method: detecting BBB effectiveness by targeting LRP1. In a thrombotic environment, LRP1 serves as a key protein regulating BBB function and linking the fibrinolytic system [35]. Given that tissue-type plasminogen activator (tPA) is expressed by capillary endothelial cells [36] and that LRP1 mediates its internalization in the circulatory system [37], the interaction between these two molecules may be crucial for BBB physiology. The practicality of the model can be evaluated by detecting the expression of LRP1 protein. Meanwhile, monitoring the changes in LRP1 expression and distribution can reflect both the structural and functional status of the BBB and the activity of the fibrinolytic system, providing a new perspective for the accurate evaluation of the BBB in thrombus-related brain diseases.

This study developed a novel BBB micro-organoid model by employing a DSMS film to mimic the BBB basement membrane and constructed a self-assembled spatial structure [18]. The model restricted the passage of exogenous substances by regulating the paracellular permeability via TJCs, thereby maintaining homeostasis of the BBB microenvironment. This is the first BBB micro-organoid constructed using sheet-like DSMS film [38], which could further help expand the application of acellular matrices in micro-organoid engineering. This study presents a novel approach to evaluating the BBB function through targeted detection of LRP1. The validity of the micro-organoid model was confirmed by assessing LRP1-mediated BBB dysfunction and alterations in the fibrinolytic system under thrombotic conditions. Demonstrating intact basal barrier properties, the model enables responsive feedback to BBB-disrupting agents and exhibits selective transport, thereby offering a reliable platform for elucidating BBB transport mechanisms and conducting in vitro preclinical studies.

2. Materials and Methods

2.1. Chemicals and Reagents

Fresh squids were purchased from Zhoushan Marine Bioengineering Co., Ltd. (Zhoushan, China). Sodium dodecyl sulfate (Shanghai, China), ethanol (Shanghai, China), acetic acid (Shanghai, China), and Triton X-100 (Shanghai, China) were used in the decellularization process. All the chemicals were of analytical grade and used without further purification. Astrocyte medium (AM 1801; Sciencell, Carlsbad, CA, USA) and endothelial cell medium (ECM 1001; Sciencell, Carlsbad, CA, USA) were used for the cell culture. Rabbit polyclonal anti-ZO-1 and fluorescein isothiocyanate isomer (FITC) goat anti-rabbit IgG were purchased from Proteintech (Wuhan, China). Rabbit polyclonal anti-VE-cadherin was purchased from Abbkine (Wuhan, China). Horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG was purchased from Affinity (Bioscience, Cincinnati, OH, USA).

2.2. Preparation of Decellularization Mantle Scaffold

Squid mantles were harvested from freshly killed squids (body weight: 30 ± 5 kg). The mantles were rinsed with ddH_2_O (liquid-to-solid ratio ≥ 5:1 to ensure complete submersion) to thoroughly remove residual blood and mucus. After being cut into pieces, the samples were subjected to five freeze–thaw cycles between −80 °C and 37 °C for auxiliary physical decellularization (Figure 1). Subsequently, the samples were treated sequentially with 1% sodium dodecyl sulfate (SDS) at 4 °C for 2 h [39] and a mixed solution of 0.05% trypsin and 0.5% Triton X-100 at 4 °C for 24 h [40]. In order to provide favorable cell growth conditions and to improve cell adhesion, the samples were further decolorized and decellularized with an acetic acid/ethanol/water mixture (8:25:67, v/v/v) at 4 °C for 4 h [36,41,42]. Before each treatment step, the mantle was rinsed with 1% antibiotic-containing ddH_2_O and phosphate-buffered saline (PBS) (liquid-to-solid ratio ≥ 5:1) for 2 h to eliminate residual reagents and bacteria. After the above procedures, the samples were washed on a magnetic stirrer (liquid–solid ratio ≥ 5:1) for 12 h, followed by ten cycles of ultrasonic cleaning (30 s per cycle, ultrasonic power: 100 W) to thoroughly remove remaining reagents. The samples were then lyophilized overnight in a freeze dryer to obtain the DSMS film. After lyophilization, the final product was collected and sterilized by ^60^Co irradiation. Fresh undecellularized mantle was used as the control group, which was also lyophilized for subsequent analysis.

2.3. Decellularized Scaffold Characteristics

2.3.1. Microstructure Observation

The microstructure of the DSMS film was observed by scanning electron microscopy (SEM). Briefly, the DSMS film was cut into small pieces of 1 mm × 1 mm and fixed on the sample stage with conductive adhesive tape. The microstructure of the DSMS film surface after gold spraying was observed using a Smur4800 cold field-emission scanning electron microscope (FE-SEM, Hitachi SU5000, Tokyo, Japan).

2.3.2. Circular Dichroism (CD) Spectroscopy Analysis

A manual grinder was used to grind the lyophilized DSMS into fine powder. An appropriate amount of the powder was accurately weighed and dissolved in acetic acid to prepare a homogeneous sample solution at a concentration of 0.1–0.5 mg/mL. After baseline scanning, the solvent in the cuvette was replaced with the prepared DSMS film solution, ensuring no air bubbles were present. The cuvette was placed back into the sample compartment, and circular dichroism scanning was performed under preset conditions. (Circular dichroism spectrometer, Chirascan VX, Leatherhead, UK).

2.3.3. UV–Vis Analysis

The DSMS film was dissolved in 0.5 M acetic acid solution at a concentration of 0.1 mg/mL. The mixture was treated with concurrent heating (45 °C) and ultrasonication (3 cycles) to facilitate dissolution, followed by clarification via centrifugation [43]. Baseline correction was first performed using the pure acetic acid solvent as a blank control, with a background scan across 200–600 nm on a UV–Vis spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) to eliminate solvent-derived interference. The centrifuged DSMS solution was then loaded into a cuvette, and an absorption spectrum was acquired over the same 200–600 nm range.

2.3.4. Fourier-Transform Infrared Spectrometer (FTIR) Analysis

The FTIR was switched to attenuated total reflection (ATR) mode and preheated to a stable state. A background scan was performed over the wavenumber range of 4000–800 cm^−1^ (resolution: 4 cm^−1^, 32 cumulative scans) using an empty ATR probe as the reference to achieve baseline correction [44,45]. Subsequently, the front and back membrane samples of the DSMS film were respectively placed in tight, flat contact with the crystal surface of the ATR probe, with uniform pressure applied to ensure gap-free adhesion. Spectral scans were conducted for both sample types over the same wavenumber range to collect infrared absorption spectra [46]. Finally, the acquired spectra were subjected to baseline correction and automatic smoothing, and the positions and intensities of absorption peaks corresponding to characteristic functional groups were compared and analyzed.

2.3.5. Quantitative Determination of Residual DNA and RNA Content

Genomic double-stranded DNA (dsDNA) was extracted from each sample using a Genomic DNA Extraction Kit (DP304, TIANGEN, Beijing, China). RNA was isolated from each group with a SteadyPure Universal RNA Extraction Kit (AG21017, Accurate Biotechnology, Nanjing, China). The concentrations of dsDNA and RNA were quantified using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). DSMS film before and after decellularization was stained with DAPI cell staining solution (Beyotime, Beijing, China). Images were captured using a fluorescence microscope (ICX41, Sunny, Ningbo, China) and the fluorescence intensity was quantitatively analyzed with analysis software.

2.3.6. Atomic Force Microscopy (AFM)

Atomic force microscopy (AFM; Dimension Icon, Bruker Corp., Bremen, Germany) with a tip (φ = 5 μm) was used to measure the surface roughness and surface height and to generate three-dimensional (3D) models of both the front and back sides. Data were processed using Nanoscope software (v.1.8).

2.4. Construction of Biomimetic BBB Micro-Organoid Model

2.4.1. Cell Cultures

The human cerebral microvascular endothelial cell line (hCMEC/D3, Meisen, Zhejiang, China) and human astrocytes (SVG p12, Meisen, Zhejiang, China) were cultured in endothelial cell medium (ECM 1001; Sciencell, Carlsbad, CA, USA) and astrocyte cell medium (AM 1801; Sciencell, Carlsbad, CA, USA), respectively, in a 5% CO_2_ incubator at 37 °C. The hCMEC/D3 cells and human astrocytes were reseeded at a density of 1 × 10^4^ cells/well into 96-well plates (Corning, Corning, NY, USA) and the DSMS film was coated on the plates, respectively, for CCK-8 assay and live/dead-cell staining. To construct an in vitro BBB model for TEER measurement, hCMEC/D3 cells and human astrocytes were seeded at densities of 7.0 × 10^5^ cells/cm^2^ and 3.5 × 10^5^ cells/cm^2^, respectively, onto DSMS films mounted in CellCrown inserts (Millipore, Santa Clara, MA, USA). For immunofluorescence (IF) staining and quantitative real-time polymerase chain reaction (RT-qPCR) analysis, the complexes that had formed the BBB micro-organoid model after completion of culture were removed from the CellCrown inserts. The culture medium was changed every other day until confluence was reached.

2.4.2. The BBB Micro-Organoid Constructed by NVU Cells

First, DSMS films were sterilized by double-sided γ-irradiation (2 h per side) [47] (Figure 2). Immediately after sterilization, the scaffold pieces were transferred into sterile inert glass bottles, and complete medium was added. After sealing the glass bottles, they were placed in a 37 °C, 5% CO_2_ incubator for constant-temperature immersion for 24 h. Subsequently, in a sterile biosafety cabinet, the scaffold pieces were gently rinsed twice with sterile PBS to remove excess medium. A sterile CellCrown (cell culture insert) was placed in a multi-well plate. The pretreated DSMS film piece was picked up with sterile forceps and accurately placed in the middle of the circular groove of the CellCrown, ensuring tight contact between the film edge and the inner wall of the insert to prevent leakage. Meanwhile, a leak test was performed to confirm no liquid leakage. Astrocytes were first seeded onto the abluminal (bottom) side of the DSMS film within the insert assembly at a density of 3.5 × 10^4^ cells/cm^2^. The plate was placed in a 37 °C, 5% CO_2_ incubator for 24 h of culture until the cells were fully attached to the surface of the DSMS film and spread [48]. Thereafter, rat tail collagen coating was conducted: the DSMS film coated with rat tail collagen was evenly laid on the surface of the culture plate and left to stand for 12 h to ensure stable bonding between rat tail collagen and the DSMS film. Subsequently, hCMEC/D3 cells were seeded: the medium in the insert was aspirated. The hCMEC/D3 cell suspension was then added to the luminal side at a density of 7 × 10^4^ cells/cm^2^. The plate was returned to the incubator for continued culture, constructing a “hCMEC/D3 –scaffold–hACs” double-layer cellular structure. After seeding, 1/2 of the complete medium volume was replaced daily (to avoid nutrient deficiency affecting cell function) [11], supporting the self-assembly of the two cell types into a BBB micro-organoid model complex.

2.5. CCK-8 Assay

Cell viability was assessed using the Cell Counting Kit-8 (CCK-8) assay in two different formats:

1. Direct seeding on DSMS: Cells (hCMEC/D3 or astrocytes) were seeded at 1 × 10^4^ cells/well in 100 μL medium into 96-well plates, either uncoated or pre-coated with DSMS film, and cultured for 24 h. After 1, 3, 5, 7,10 and 12 days of culture, 10% CCK-8 solution (AbMole BioScience, Houston, TX, USA) was added, and the optical density at 450 nm was measured using a microplate reader (Agilent, Santa Clara, CA, USA).
2. Treatment with DSMS leachates: Cells (100 μL, 1 × 10^4^ cells/well) were seeded as in Format 1 on uncoated plates and cultured for 24 h. Subsequently, the medium was replaced with DSMS film leachates of different concentrations. The DSMS film leachates were prepared by immersing sterile DSMS film in serum-free medium at a solid–liquid ratio of 10 mg/mL (w/v) under constant-temperature shaking (37 °C, 100 rpm) for 24 h, followed by filtration through a 0.22 μm sterile filter to obtain the stock leachate (100%, v/v). After 24 h of culture, 10% CCK-8 solution (AbMole BioScience, Houston, TX, USA) was added, and the optical density at 450 nm was measured using a microplate reader (Agilent, Santa Clara, CA, USA).

2.6. LIVE/DEAD Cell Staining

Viability of cells in the co-culture model on days 1, 3, and 7 was assessed using a LIVE/DEAD viability/cytotoxicity kit. The specific procedures are as follows: First, the samples were washed three times with phosphate-buffered saline (PBS). Subsequently, the samples were stained using a LIVE/DEAD Cell Staining Kit (Beyotime, Beijing, China) in accordance with the manufacturer’s instructions. An appropriate volume of Calcein AM/PI working solution was added under dark conditions. For 96-well plates, 100 μL of the working solution was added to each well. After incubation for 30 min, the samples were washed three times with PBS again to remove residual reagents. Imaging was performed using a fluorescence microscope (ICX41, Sunny, Ningbo, China). Live and dead cells were indicated by green (Calcein AM) and red (PI) fluorescence, respectively.

2.7. Migration Assay

To investigate the effect of the DSMS film leachate on cell proliferation and migration, the in vitro scratch assay was performed. The DSMS film leachates were prepared by immersing sterile DSMS film in serum-free medium at a solid–liquid ratio of 10 mg/mL (w/v) under constant-temperature shaking (37 °C, 100 rpm) for 24 h, followed by filtration through a 0.22 μm sterile filter to obtain the stock leachate (100%, v/v). hCMEC/D3 cells and astrocytes in the logarithmic growth phase were seeded into 6-well plates at a density of 2 × 10^6^ cells/well and cultured until 100% confluence. Subsequently, sterile pipette tips (200 μL) were used to create a scratch on the underlying cell layer in each well [49]. After washing with phosphate-buffered saline (PBS), the leachate was added for cell culture. Cell migration was monitored every 12 h using an inverted microscope. ImageJ software (Version: 2.14.0/1.54f) was employed for quantitative analysis of the scratch area.

2.8. Ultrastructure of the Blood- Brain Barrier Micro- Organoid Model

For ultrastructural analysis, cell-seeded constructs were rinsed twice with PBS and fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.0) for 4 h at 4 °C. Samples were gradually dehydrated with gradient-diluted ethanol (30%, 50%, 70%, 80%, 90%, 95% and 100%) for 10 min per step, with two changes of 100% ethanol for 20 min each [23]. Afterwards, the samples were dehydrated in a freeze dryer. Finally, samples were sputter-coated with a gold/palladium layer prior to SEM observation.

2.9. TEER Assay

Barrier function was indicated by the temperature-corrected TEER (tcTEER) values, measured by a Millicell ERS-2 system (Millipore, Santa Clara, MA, USA) in the in vitro micro-organoid model after temperature calibration. Cells were seeded onto the DSMS film coated with rat tail collagen, which was placed in CellCrown inserts. As a control, cells were seeded on Transwell inserts coated with type I. The initial tcTEER measurement (day 1) was taken 24 h post seeding. Measurements were then recorded daily for up to two weeks [50]. Before each measurement, the culture medium was equilibrated to the appropriate temperature. During measurement, the TEER probes (silver/silver chloride electrode probes) were completely immersed in the apical and basolateral chambers, respectively.

2.10. FITC–Dextran Penetration Test

The FITC–Dextran Penetration Test was used to quantitatively evaluate the integrity of the in vitro BBB micro-organoid model. First, 70 kDa and 10 kDa FITC–dextran were serially diluted, and their absorbance was measured using a multi-functional microplate reader (Agilent, Santa Clara, CA, USA) to generate fluorescence standard curves. Then, 70 kDa and 10 kDa FITC–dextran were added to the upper side of the BBB micro– organoid model. Subsequently, 10 μL of liquid was aspirated from the lower chamber every hour and measured using a multi-functional microplate reader (Agilent, Santa Clara, CA, USA). The permeability was calculated based on the fluorescence standard curves and the permeability calculation formula.

2.11. Immunofluorescent Staining Microscopy

For immunofluorescence (IF) analysis, after the tcTEER values stabilized (approximately 10 days later), the BBB complexes were removed from the CellCrown inserts for immunofluorescence staining. The complexes were fixed in 4% paraformaldehyde at 4 °C for 30 min, then permeabilized with 0.5% Triton X-100 for 20 min, wash three times with PBS, and blocked with 5% (v/v) bovine serum albumin (BSA) in TBST for 1 h at room temperature. Anti-ZO-1 and anti-VE-cadherin primary antibodies were diluted 1:1000 in 3% BSA and incubated with the complexes overnight at 4 °C. After three washes with PBS, secondary antibodies (FITC-conjugated goat anti-rabbit IgG antibodies) diluted 1:500 in 3% BSA were added and incubated at room temperature in the dark for 1 h. Finally, nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) diluted 1:500 in 3% BSA for 10 min at room temperature, followed by three additional PBS washes. Fluorescence microscopy was used to observe three random fields of view per well.

2.12. RNA Extraction from hCMEC/D3 Cells and Quantitative RT-qPCR

Five treatment groups were set up as follows: blank control (no intervention); TNF-α alone (20 ng/mL, diluted in 0.1% BSA/PBS, incubation for 6 h); TNF-α + LRPAP1 (20 nM LRPAP1 pretreatment for 4 h, followed by incubation with 20 ng/mL TNF-α for 6 h); TNF-α + thrombin + fibrinogen (20 ng/mL TNF-α incubation for 6 h, then treatment with 2 U/mL thrombin + 2 mg/mL fibrinogen for 8 h); and quadruple treatment (20 nM LRPAP1 pretreatment for 4 h → 20 ng/mL TNF-α incubation for 6 h → 2 U/mL thrombin + 2 mg/mL fibrinogen treatment for 8 h, with reagent concentrations consistent with the above groups). Total RNA was extracted from these cells using the SteadyPure Quick RNA Extraction Kit (Accurate, Hunan, China) according to the manufacturer’s protocol. The quantity and purity of total RNA were determined using a Nanodrop (Thermo Fisher, Waltham, MA, USA). cDNA was synthesized from 1 µg of total RNA using the Evo M-MLV RT Mix Kit with gDNA Clean for qPCR (Ver.2, Accurate, Hunan, China). Quantitative RT-qPCR was conducted in a 20 µL reaction system using the SYBR Green Premix Pro Taq HS qPCR Kit and gene-specific primers. The cycle threshold (Ct) values were collected and normalized to the internal reference gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The relative mRNA expression levels of each target gene were calculated using the 2^(−ΔΔCt)^ method. Primers for LRP1, PLAT (tPA), claudin-5 (CLDN-5), Occludin (OCLN), and GAPDH were designed and synthesized by Sangon Biotech (Shanghai, China; https://www.sangon.com/; accessed on 22 February 2026). Primers for LRP1, PLAT (tPA), claudin-5 (CLDN-5), Occludin (OCLN), and GAPDH were designed and synthesized by Sangon Biotech (Shanghai, China; https://www.sangon.com/; accessed on 24 September 2025), and the primer sequences are listed in Table 1.

2.13. Statistical Analysis

For statistical analysis, GraphPad Prism 9 software was employed. Data were presented as mean ± standard deviation (SD) from three independent experiments. Two-group comparisons were performed using Student’s t-test. Single-factor multi-group comparisons were analyzed by one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test. Two-factor multi-group comparisons were assessed by two-way ANOVA with appropriate post hoc tests. A p-value < 0.05 was considered statistically significant, with * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. All tcTEER values were expressed as relative values, with the initial value set to 100%. Quantitative analysis of positive-stained cells for various markers was conducted using ImageJ in a blinded manner.

3. Results

3.1. Characterization of the DSMS Film

Squid mantle was processed into acellular membranes to recapitulate the human BBB basement membrane and emulate its native structural complexity. The DSMS film was successfully fabricated through a sequential process involving detergent treatment (SDS and Triton X-100), extensive rinsing, and lyophilization, resulting in an off-white, membranous scaffold (Figure 3a). Macroscopic observation revealed a grayish-white appearance with a textured outer surface and a smoother inner surface.

These scaffolds presented a grayish-white appearance with a relatively rough outer surface and a smoother inner surface. Light microscopic observations revealed a region-specific porous architecture: the front side (designated for endothelial cell seeding) exhibited uniformly distributed, relatively large pores, while the back side (for astrocyte seeding) possessed smaller, denser pores (Figure 3a). The SEM further complemented these findings, demonstrating that the front surface was relatively smooth (in contrast to the rough back surface); the front epidermal layer was flat with abundant pores, whereas the opposing side displayed a dense cotton-like fibrous structure (Figure 3b). Cross-sectional SEM analysis additionally unveiled distinct interconnected pores among fibers, forming a multi-layered porous configuration with a well-organized fibrous network and prominent topological characteristics—features favorable for mimicking the native BBB basement membrane microenvironment [51].

To comprehensively elucidate the structural adaptability of the DSMS film in simulating the BBB model, this study combined AFM for surface topography analysis and pore structure characterization to reveal its unique structural advantages in BBB model construction. The average surface roughness was calculated as Ra = 2.45 nm and Rq = 3.76 nm, with height variation ranges of −11.8 nm to 11.5 nm, −12.3 nm to 12.1 nm, and −12.0 nm to 11.8 nm across the regions, respectively (Figure 3d). This smooth topography minimizes mechanical hindrance to endothelial cell spreading and provides a foundation for the formation of tight junctions. Meanwhile, Image J analysis confirmed that this side had an average porosity of 51.938% and a mean surface pore size of 7.32 μm. The relatively large and uniform pores not only ensure efficient diffusion of nutrients and oxygen but also provide sufficient space for endothelial cell adhesion and proliferation. These two characteristics synergistically meet the physiological requirements for the continuity and barrier function of the endothelial layer in the BBB. In contrast, the structure of the back side of the matrix (astrocyte seeding side) was specifically tailored to the characteristics of astrocytes, with an average surface roughness of Ra = 4.0 nm and Rq = 5.78 nm, and height fluctuation ranges of −23.5 nm to 23.3 nm, −24.0 nm to 23.7 nm, and −23.8 nm to 23.6 nm, respectively. The multiple adhesion sites and mechanical anchorage provided by the rough surface can promote the extension of astrocyte processes and morphological maturation. Additionally, the back side had an average porosity of 68.372% and a mean surface pore size of 8.821 μm (smaller and denser), which further enhances support for the construction of functional astrocyte networks and facilitates the formation of a synergistic regulatory relationship with endothelial cells. This region-specific structural combination of “smooth surface + large pores” on the front side and “rough surface + small pores” on the back side precisely matches the growth needs and functional division of labor of the two cell types [52]. It provides an ideal matrix foundation for recapitulating the intercellular interactions and natural microenvironmental characteristics of the BBB in vitro.

The CD spectrum of the DSMS film displayed the characteristic peaks of native collagen triple helices (~195 nm positive; ~208 nm and ~222 nm negative), confirming the preservation of its functional secondary structure. UV–Vis spectroscopy showed negligible absorption at 234 nm (nucleic acid marker) and across the 260–280 nm range, verifying efficient decellularization and high scaffold purity. Attenuated total reflection–Fourier-transform infrared (ATR-FTIR) analysis revealed consistent amide I (1650 cm^−1^), amide II (1540 cm^−1^), and carbohydrate-related (1030 cm^−1^) peaks on both film sides, indicating uniform chemical composition (Figure 4a–c, Table 2). The consistent peak positions and intensities between the front (smooth, large-pored) and back (rough, small-pored) sides confirmed uniform chemical composition across the scaffold. Efficient decellularization was confirmed by fluorescence imaging of DAPI-labeled nuclei, showing over 95% cell removal. Decellularization was further validated by >95% nuclear removal in DAPI imaging and quantitative nucleic acid levels below 15 ng/mg (dsDNA) and 10 ng/mg (RNA), corresponding to >90% removal (Figure 4d,e). As a core component of the BBB basement membrane, preserved collagen integrity is essential for supporting endothelial and astrocyte adhesion/proliferation, providing a solid structural basis for the BBB model construction.

3.2. The Good Biocompatibility Exhibited by the DSMS Film

Human cerebral microvascular endothelial cells and astrocytes were cultured in well plates coated with the DSMS film. Subsequently, their proliferation, migration, survival rate, and death rate were evaluated.

First, CCK-8 cell viability assays were performed for two distinct cell culture formats: direct-contact cell culture and leachate-based cell culture. In the first experiment, the DSMS film leachates were tested at concentrations of 20%, 40%, 60%, 80%, and 100% (v/v). Compared with the control group, these leachates did not reduce the viability of hCMEC/D3 cells or astrocytes except for the 100% leachate group. In the 100% leachate group, cell viability decreased to 75% of the control level. In contrast, the 20% leachate significantly enhanced cell viability. The OD values measured at 450 nm for hCMEC/D3 cells showed an increase of 23.2% compared with the control group. For astrocytes, the OD value increased by 17.5% (both p < 0.05). The reason for reduced cell viability in the 100% leachate group is as follows.

In the second experiment, the optical density (OD) at 450 nm of hCMEC/D3 cells and human astrocytes was slightly lower in 96-well plates coated with the DSMS film than in uncoated (control) plates on day 1. However, at all other time points (days 3, 5, 7, 10, and 12), the OD values in the DSMS-film-coated plates were significantly higher than those in uncoated plates (Figure 5a). Specifically, compared with the uncoated group, the OD values of hCMEC/D3 cells in the DSMS-film-coated group increased by 23.6%, 38.2%, 51.7%, 67.4%, 72.9%, and 78.3% at each time point (days 1 to 12). For astrocytes, the OD values increased by 19.8%, 31.5%, 44.2%, 59.6%, 65.1%, and 70.5% respectively (all p < 0.05).The slight OD reduction in the DSMS-film-coated group on day 1 can be explained by the difference in surface topological features between the DSMS film and conventional plastic culture plates. Both hCMEC/D3 cells and astrocytes rely on ECM signals for adhesion and spreading [53]. On day 1, cells were in the adaptive phase of “matrix recognition and adhesion site integration”, leading to slightly lower metabolic activity (reflected by OD values) compared with cells on plastic plates. This is consistent with the “time-dependent adaptation of cell–matrix interaction” characteristic.

Scratch assay showed distinct migratory responses of hCMEC/D3 and ACs to the DSMS film extract (Figure 5c,d). The extract significantly promoted hCMEC/D3 migration in a time-dependent manner: 24 h migration rate was 50.85% (vs. 14.80% in blank control, * p < 0.05) and increased to 75.00% at 48 h (vs. 36.37% in control, * p < 0.05). In contrast, the extract had a negligible effect on ACs migration: even at 48 h, the migration rate was only slightly elevated to 22.00% compared with 16.37% in the control group. Though ACs showed marginal significance (* p < 0.05), the migratory improvement was far weaker than that of hCMEC/D3. These data demonstrate that the extract selectively enhances hCMEC/D3 migration, which may facilitate functional endothelial monolayer formation during the BBB micro-organoid construction.

Calcein staining showed green fluorescence of live cells, while PI staining revealed a small number of red fluorescent signals from dead cells; the proportion of live cells was relatively low in the merged images (Figure 6). On day 3, the green fluorescence of Calcein became significantly denser, the degree of cell spreading was enhanced, and the red fluorescent signals of PI further decreased. On day 7, the green fluorescence of Calcein almost filled the visual field, cells formed a dense monolayer, and only scattered red fluorescent signals of PI were observed. Quantitative analysis showed that the cell survival rate was approximately 70% on day 1, and exceeded 90% on both day 3 and day 7; there was no significant difference between the survival rates on day 3 and day 7 (p > 0.05). Cells in the complex exhibited a time-dependent trend of survival and proliferation during culture: on day 1, cells were in the initial adaptation stage with a relatively low survival rate; starting from day 3, cells proliferated rapidly and maintained a high survival rate; by day 7, a stable cell monolayer was formed. These results indicate that the system exhibits good biocompatibility, provides a suitable growth microenvironment for cells, and supports long-term cell survival and function maintenance [27,54]. These results indicate that the DSMS film provides a favorable environment for the culture of the hCMEC/D3 and SVG p12 and exhibits good biocompatibility [55,56].

3.3. Establishment of the Barrier in the Constructed Micro-Organoid Model

The SEM images demonstrate that the DSMS film effectively maintains the morphological integrity of both hCMEC/D3 and hACs, and also facilitates the formation of a dense monolayer by hCMEC/D3 cells (Figure 7a,b). The hCMEC/D3 cells displayed a typical “cobblestone-like” morphology and attached to the surface of the DSMS film with a robust cell–DSMS adhesion, which indicated that the DSMS film provided suitable mechanical support and a signaling-conductive microenvironment for the NVU cells. Meanwhile, the ACs cells presented a “stellate” or “multi-protrusive” morphology on the back side of the DSMS film, whose cellular protrusions extended deep into the matrix. The complex three-dimensional network structure formed by the ACs cells is highly consistent with the physiological functions of astrocytes including nutritional support and structural anchoring, which suggested the participation of the matrix provided by the DSMS film in maintaining the astrocytes’ functional morphology. The intact barrier formed by these NVU cells in the constructed micro-organoid model showed no obvious intercellular gaps as shown in SEM images, which was also supported by the stable tcTEER phase with relatively high values (Figure 7c).

Barrier integrity was successfully achieved in the constructed the BBB micro-organoid model by the eighth day post seeding, when the tcTEER value stabilized at approximately 200–250 Ω/cm^2^ (310.15 K) (Figure 7c). The tcTEER value exceeded 200 Ω/cm^2^ (310.15 K) at day 6, which could last for 5 days of cultivation. The barrier in the BBB micro-organoid model utilizing the DSMS film exhibited resistance tcTEER values comparable to those of the traditional model using PET membrane. All these results obtained by SEM and TEER assay indicated that the DSMS film support constructs an artificial BBB in the micro-organoid model.

3.4. Barrier Functions Mediated by TJCs of the BBB Micro-Organoid Model

According to previous studies, the BBB exhibited selective permeability of substances with low molecular weight (MW) below 0.5–1 kDa [57,58]. Therefore, FITC–dextrans with different molecular weights (10 and 70 kDa) were separately applied to verify the barrier function in the BBB micro-organoid model (Figure 8b,c). The apparent permeability coefficients of both 10 and 70 kDa FITC–dextrans through different membranes were all below 20 × 10^−6^ cm/s over 3 h, which indicated the low permeabilities of the barrier [14,59]. As suggested by the Pₐₚₚ values between the PET group (5 × 10^−6^–10 × 10^−6^ cm/s) and the PET–collagen group (1 × 10^−6^–3 × 10^−6^ cm/s), collagen supplement was necessary for the barrier construction and its normal function [60]. Although the similar essential role of collagen in the barrier construction did not show in the case of the DSMS film, the Pₐₚₚ values in the DSMS film (about 1.5–6 × 10^−6^ cm/s) and the DSMS–collagen (about 1–6 cm/s) group were at the order of 10^−7^ cm/s for both 10 kDa and 70 kDa FITC–dextran, which were comparable to those of Transwell-based BBB models and met the construction requirements for in vitro BBB models [59,61].

This expression pattern was highly consistent with that of tight-junction proteins in the physiological BBB, indicating that functional tight junctions had been successfully formed by endothelial cells in the model (Figure 9). Combined with the previously verified permeability and TEER data, these results collectively demonstrated that the constructed BBB micro-organoid model possessed intact structural and functional features of the blood–brain barrier, confirming the successful establishment of the micro-organoid model.

3.5. The Physiological Function of the BBB Micro-Organoid Model Mediated by Cell–Cell Interaction Based on the NVU Microenvironment

As shown, the TEER values of the blank group and the LRPAP1 alone group stayed around 220 Ω/cm^2^. This showed that LRPAP1 had no significant effect on the normal BBB. However, the TEER values of the TNF-α group and the TNF-α + thrombin + fibrinogen group decreased to 100–110 Ω/cm^2^. This highlighted the severe disruption of the BBB integrity by the combined effects of inflammation and coagulation. Notably, the TEER value of the LRPAP1-pretreated group (LRPAP1 + TNF-α + thrombin + fibrinogen) recovered significantly compared to the TNF-α + thrombin + fibrinogen group. This confirmed that LRPAP1 could partially reverse the BBB damage caused by combined inflammation and coagulation. Activating LRP1 effectively preserves the structural and functional integrity of the BBB.

To investigate the expression of thrombosis-related proteins and tight-junction proteins at the transcriptional level, we performed RT-qPCR assays [62,63]. LRPAP1 (pretreatment before TNF-α and thrombin + fibrinogen addition) exerted a partial reversal effect on the TNF-α and TNF-α + thrombin + fibrinogen-induced downregulation of LRP1 and PLAT mRNA expression [64,65] (Figure 10d–g). Mechanistically, this suggests that LRP1, in conjunction with its ligand PLAT (tPA), plays a critical role in mediating the regulation of OCLN expression by LRP1 itself [66,67]. This regulatory interplay is essential for maintaining the structural integrity and functional barrier properties of the BBB in thrombotic environments [68]. In contrast, the reversal effect of LRPAP1 on CLDN-5 was not obvious, while it showed a relatively significant reversal effect on OCLN [62,69]. These results collectively indicate that LRPAP1 differentially regulates the expression of thrombosis-related and tight-junction proteins in response to combined inflammatory and thrombotic stimuli, with more pronounced effects on OCLN than on CLDN-5 [64].

To determine the transcriptional profiles of thrombosis-associated proteins and tight-junction proteins, we conducted RT-qPCR assays. LRPAP1 pretreatment partially reversed the downregulation of LRP1 and PLAT mRNA levels induced by TNF-α, thrombin, and fibrinogen (Figure 10d–g). Based on these findings, we hypothesize that LRPAP1 intervention may activate LRP1 and enhance its expression in the BBB organoid. This upregulation could potentially facilitate the binding of LRP1 to its ligand, PLAT (tPA). Mechanistically, this LRP1-PLAT interaction is essential for mediating OCLN expression [70], which in turn preserves the structural and functional integrity of the BBB under thrombotic conditions. In contrast, LRPAP1 exerted a negligible reversal effect on CLDN-5 expression, whereas it significantly restored OCLN expression. Collectively, these results demonstrate that LRPAP1 differentially regulates the expression of thrombosis-associated and tight-junction proteins in response to combined inflammatory and thrombotic stimuli, with more prominent effects on OCLN than on CLDN-5.

In conclusion, LRP 1 was successfully expressed in the established BBB organoid and effectively regulated the barrier function under thrombotic conditions, which implies the successful construction of a BBB organoid capable of mediating cell–cell interactions within the NVU microenvironment.

4. Discussion

Since developing a practical BBB in vitro model capable of simulating the genuine NVU microenvironment is one of the keys to advancing research on cardiovascular and cerebrovascular diseases, this study provided an idea of a novel BBB micro-organoid model utilizing a marine-derived decellularized ECM scaffold.

The DSMS film provided the architecture for the critical morphological characteristics of the BBB in the micro-organoid model, with its apical side offering a smooth, large-pored surface for cerebral microvascular endothelial cells (hCMEC/D3) and its basolateral side providing a rough, small-pored surface for hACs. Through the specific spatial structure of the DSMS film and via contact co-culture of these two cell lines [71], a confluent endothelial monolayer with a deeply integrated astrocytic network was rapidly formed, as shown by SEM images. Furthermore, the preservation of native collagens, especially collagen IV, was confirmed by CD spectroscopy. Collagen IV is the fundamental component of the basement membrane of the BBB and supports integrin-mediated signaling for the maintenance of barrier function [72]. Collectively, the DSMS film exhibits region-specific structural, topographical, and compositional features that mimic the native BBB microenvironment, highlighting its potential for the BBB-related research and tissue engineering applications.

The functional data strongly validate the model. The achieved TEER plateau (~220 Ω/cm^2^) is within the range reported for other complex co-culture models and signifies the development of robust tight junctions [73]. The barrier functions have been further evaluated by examining the expression and localization of key regulating factors of TJCs, mainly including ZO-1 and VE-cad. Although the DSMS film was not completely transparent, which may influence the image clarity under fluorescence microscopy, both the TJCs’ (ZO-1 and VE-cad) expressions (in green) and cell nuclei (in blue) were successfully observed. The TJC protein ZO-1 exhibited a continuous net-like structure between hCMEC cells of the barrier in the BBB micro-organoid model, which was uniformly distributed along the cell boundaries without obvious discontinuities or fragmentation in merged images. The TJC protein VE-cad displayed a serrated expression pattern, which was well localized at cell borders in the merged images. All these expressions and localizations of ZO-1 and VE-cad indicated the formation of the TJC-mediating barrier function in the BBB micro-organoid model. The size-selective permeability to dextrans and the clear visualization of continuous ZO-1 and VE-cad junctions provided complementary evidence of a functional paracellular barrier. Interestingly, coating the DSMS film with rat tail collagen I did not enhance barrier function, and in some cases, even impaired it compared to the DSMS film alone. This suggests that the native, complex ECM composition of the DSMS film may provide more balanced cues for BBB maturation than a single, exogenous collagen type, potentially avoiding disruption of endogenous integrin signaling pathways.

LRP1, as a regulatory protein, serves as a crucial target for thrombosis therapy, and the regulatory effect of LRPAP1 on LRP1 also plays a vital role in thrombosis. LRP1 is expressed by hACs [74,75,76], and it is involved in maintaining the integrity of TJCs and regulating the barrier function of the BBB [62]. The tPA is expressed in capillary endothelial cells [77]. Additionally, LRP1 mediates the internalization of circulating tPA in the circulatory system [77]. Therefore, LRP1 has been applied in this study as an appropriate medium to testify the NVU cells’ physiological activities on the DSMS film and also investigate the cell–cell interaction in the BBB and its mediating barrier function.

A key strength of this model is its demonstrated responsiveness to pathophysiological stimuli. The reversible barrier disruption induced by TNF-α and thrombotic factors [78], coupled with the partial protection afforded by LRPAP1, illustrates its utility for mechanistic studies. The differential effect of LRPAP1 on OCLN versus CLDN5 mRNA expression is particularly intriguing and warrants further investigation, as it suggests distinct regulatory pathways for different tight-junction components under stress [79]. This aligns with growing evidence of the multifaceted role of LRP1 in the BBB regulation, neuroinflammation, and thrombosis [80].

Compared to chip-based or stem-cell-derived models, the DSMS film platform offers significant advantages in terms of simplicity, cost, and potential for standardization and scaling. The use of a readily available marine resource addresses ethical and supply concerns associated with mammalian tissues [81]. However, a recognized limitation of any biologically derived scaffold is batch-to-batch variability [82]. Unavoidable factors such as film thickness and roughness may exert a certain influence on the cell availability, cell adhesion and the barrier functionality. While our process yielded consistent results, future work could focus on further standardizing the decellularization protocol and establishing quality-control metrics (e.g., precise porosity, glycosaminoglycan content) to minimize variability.