Controllable blood–brain barrier (BBB) regulation based on gigahertz acoustic streaming

The blood–brain barrier (BBB) constitutes the boundary of the cerebral blood vessels¹ and is a structural, functional, and physiological dynamic barrier composed of brain endothelial cells, pericytes, astrocytes, and basement membrane.^2–6 The BBB regulates the transport of ions, nutrients, and cells between the blood and the brain,^2,7,8 and it thus plays a critical role in the maintenance of homeostasis in the central nervous system (CNS).^8–10 At the same time, owing to the sizes of drug molecules and the effects of hydrophilicity, lipid solubility, and transduction pathways, the BBB prevents more than 98% of small-molecule drugs and 100% of macromolecular drugs from entering the CNS,^11,12 thus severely restricting CNS drug interventions and treatment of CNS-related diseases.^13–17 In recent years, techniques using nanoparticles,^2,18–20 ultrasonic cavitation,^16,21,22 and hypertonic solutions have been developed to help drug molecules bypass this barrier for delivery to the required sites for targeted therapy.^23,24 Among these techniques, nanoparticles increase circulation time by means of specifically designed composition, size, shape, and surface ligands, and they then cross the BBB through different surface strategies such as endocytosis.²⁵ Ultrasonic cavitation, on the other hand, uses the energy generated by oscillations of microbubbles under local ultrasonic irradiation to destroy the connections between endothelial cells and thereby allow trans-BBB delivery of drug molecules.²⁶ The principle of action of hypertonic solutions such as mannitol is the destruction of tight junction proteins (TJs) and adhesion proteins (AJs), with the consequent formation of pores in the BBB through which drug molecules can pass.²⁷ 

In recent years, the regulation of BBB function and structure by the shear force generated by blood flow has attracted much attention. Fluid shear stress is considered to be a major stimulator of BBB differentiation and has a profound impact on maintaining the structural and functional integrity of the BBB.²⁸ Previous studies have found that when endothelial cells are subjected to fluid shear stress, cell differentiation, proliferation, and cell cycle regulation are enhanced.²⁹ For example, Cucullo et al.³⁰ found that capillary shear stress (6.2 dyn/cm²) significantly upregulated the expression of TJs, AJs, drug transporters, and integrin-related genes, and significantly increased the content of cytoskeletal proteins, suggesting that fluid shear forces can promote the development and maturation of the BBB. By contrast, abnormal levels of fluid shear force can degrade the functional properties of the BBB. For example, Garcia-Polite et al.⁷ found that a high level of shear stress (40 dyn/cm²) decreased the expression of TJs such as zonula occludens-1 (ZO-1) and claudin-5 and changed their junction morphology. This indicates that the BBB will be damaged to a certain extent when the fluid shear force is higher than the physiological shear force.

To facilitate the study of fluid shear force regulation of the BBB and overcome the BBB impediment to molecular delivery, we use a solidly mounted thin-film bulk acoustic wave resonator (SMR) that can generate acoustic streaming (AS) to construct a BBB mechanical regulation system with continuously adjustable shear force. Our experimental results show that the gigahertz AS generated by the SMR in liquid can achieve rapid regulation of the BBB and enhance molecular delivery across the BBB by interrupting the TJs between endothelial cells.

A human cerebral microvascular endothelial cell line (hCMEC/D3) was obtained from Millipore Sigma. Endothelial cell medium (ECM) was obtained from ScienCell. Calcein-AM, Alexa Flour 594 ZO-1 monoclonal antibody (ZO1-1A12), 4% paraformaldehyde, and Blocker bovine serum albumin (BSA) (10%) in PBS were purchased from Thermo Fisher Scientific. Fluorescein isothiocyanate (FITC)-conjugated dextran (20 kDa, 40 kDa, 70 kDa, and 150 kDa) was obtained from Sigma-Aldrich. Phosphate-buffered saline (PBS, pH = 7.4) and 4',6-diamidino-2-phenylindole (DAPI) were obtained from Solarbio Life Sciences. Matrigel Basement Membrane Matrix Phenol Red Free and a 24-well Transwell were purchased from Corning Costar.

The gigahertz-frequency SMR was fabricated by a microelectromechanical process. On the basis of past experience,^31–33 the acoustic resonator devices were fabricated on 100 nm undoped silicon (Si) wafers, starting with the deposition of aluminum nitride (AlN) and silicon dioxide (SiO₂) layers alternately on the Bragg reflector through physical vapor deposition (PVD) and chemical vapor deposition (CVD), respectively. Then, a sandwich structure comprising a bottom electrode (BE), a piezoelectric layer (AlN), and a top electrode (TE) was deposited and patterned layer by layer [Figs. 1(b) and 1(c)].

hCMEC/D3 cells were cultured in ECM at 37 °C in a 5% CO₂ incubator: 150 μl of hCMEC/D3 cell suspension at a density of 1 × 10⁵ was evenly seeded in Transwell chambers coated with Matrigel in advance, and the ECM was replaced every 24 h until the BBB model became mature. To monitor the growth of the BBB, its trans-endothelial electrical resistance (TEER) value was measured every 24 h by chopstick electrodes connected to a transmembrane resistance analyzer (EVOM2).

The microelectromechanically fabricated SMR was diced to a side length of 2 mm and then fixed on top of a slender electronic vacuum booster (EVB) board, with communication provided through gold wires. The SMR device thus constructed was vertically fixed on a three-axis displacement stage (with one end of the SMR device downward) and inserted into Transwell chambers cultured with the BBB model. The input power was controlled through a sinusoidal signal generated by a signal generator and amplified by a power amplifier. The distance between the SMR device and BBB model was controlled to be 1.5 mm by adjusting the Z axis of the displacement stage. Input signals of different powers (300 mW, 400 mW, and 500 mW) were then applied to the SMR through the signal generator, with the time of each action being controlled at 10 min.

The cell culture medium in the Transwell chamber was replaced with a solution of dextran conjugated with FITC, and the dextran molecules permeated across the BBB and accumulated in the lower chamber. After standing for 10 min, the solution in the lower chamber was collected, and its fluorescence intensity was measured with a microplate reader (Varioskan LUX) to characterize the molecular permeability. The emission spectrum of the collected solution under excitation light of 488 nm was measured with a microplate reader. To accurately obtain the fluorescence intensity of the solution, this was taken to be the average value of the fluorescence intensity corresponding to wavelengths in the range of 515–525 nm: the spectrum of each group was scanned twice with two wells and the average was then calculated. Three replications were performed for each molecular weight probe. To obtain the concentration of the solution from the measured fluorescence intensity, the original dextran solution was diluted by gradient and the fluorescence intensity was measured to obtain a calibration curve corresponding to concentration and fluorescence intensity. The fluorescence intensity of the collected solution was measured according to the calibration curve to obtain its molecular concentration.

Cells were fixed with 4% paraformaldehyde solution for 10 min, followed by treatment with 0.25% TritonX-100 for 25 min to permeabilize them. Then, after blocking with 5% BSA for 1 h, ZO-1 monoclonal antibody conjugated with a red fluorescent molecule (1:200 in 1% BSA) was added and incubated at room temperature for 3 h. The nuclei were finally counterstained with DAPI for 20 min. At the end of each step, the samples were washed three times with PBS for 3 min each. After staining, fluorescence micrographs were taken with a confocal microscope (Leica TCS SP8) for analysis and observation.

The experimental data are expressed in the form of mean ± standard deviation, and each group of experiments has four replicates. The fluorescence intensity of the images was obtained by ImageJ software.

The SMR consists of two electrodes with a piezoelectric layer (AlN) sandwiched between them. When an external electric field is applied through the electrodes, the piezoelectric material induces mechanical vibrations due to the inverse piezoelectric effect. The resonant frequency of the SMR designed for this experiment is 1.58 GHz, When the SMR is operating in a liquid, gigahertz AS is generated [Fig. 1(a)]. A two-dimensional finite element simulation (FEM) of AS was performed, as shown in Fig. 1(d), where the white arrows indicate the direction of the liquid flow. It can be seen that when the SMR is active in the liquid, dissipation of acoustic energy induces AS, which flows down from the middle and back from both sides, thus producing a vortex-like streaming in the liquid.

The different shear forces that can be generated by the gigahertz AS were also verified by FEM simulation. The flow rates of this streaming at different distances from the SMR (Fig. 2) (500 μm, 1000 μm, and 1500 μm) were simulated. The results show that the velocity of the AS is greater at locations closer to the SMR. The mechanical force imposed on the cells by the AS at the solid–liquid interface is caused by the sharp reduction of the velocity to zero.³¹ Therefore, the closer the cells are to the SMR, the greater is the mechanical force that can be applied. Thus, the distance between the SMR and the cell can be controlled by the Z axis to produce a specific magnitude of force imposed on the cell.

The effect of the gigahertz AS on BBB integrity was first investigated. To visualize the changes in the endothelial cells layer, hCMEC/D3 cells were stained with calcein-AM (green fluorescence) and observed with a confocal microscope. As shown in Fig. 3(a), the distribution of endothelial cells was relatively dispersed on the first day; then, after five days of culture, through continuous proliferation and confluence, there were no obvious gaps between cells, and a dense cell barrier was thus formed. This suggested that the BBB model had formed a tight barrier of cells on the sixth day under these culture conditions. To more accurately monitor cell growth, the EVOM2 instrument was used to measure the TEER of the BBB models every 24 h [Fig. 3(b)]. TEER is a classical method to determine the integrity of the BBB, with a continuously increasing TEER value usually indicating continuous maturation of a BBB model.³⁴ The measurement results showed that the TEER values of the BBB models increased continuously over time and reached their highest value of 20 Ω cm² on the sixth day, after which they stabilized over the following days [Fig. 3(c)]. This indicated that the integrity of the BBB improved over time and matured at day 6. A valid BBB model is thought to exhibit lower molecular penetration efficiency because of the restricted access to cellular pathways.³⁴ Therefore, to further follow the growth of the BBB models, the penetration of molecules across the barrier was also monitored using a solution of 20 kDa dextran conjugated with FITC, with the measurement results being shown in Fig. 3(d). The fluorescence intensity of the molecules that penetrated and accumulated in the lower chamber became weaker with time and stabilized around the minimum value at day 6. This indicated that the number of dextran molecules permeating across the barrier decreased with extended culture time, which indicates that connections between endothelial cells were gradually established in the BBB model, with a relatively dense and mature culture being achieved on the sixth day. It is known that the compactness of the BBB is due to the presence of TJs and AJs between endothelial cells, which significantly reduce paracellular permeability. ZO-1, one of the TJs, was selected to characterize the maturation of the BBB model on the sixth day of culture. We performed immunofluorescence staining of the BBB with the ZO-1 protein on the sixth day. The resulting immunofluorescence images in Fig. 3(e) show the expression of ZO-1 among endothelial cells. The above results all indicate that on the sixth day, a relatively mature BBB model had been established in the Transwell, enabling subsequent experiments to be carried out.

After the BBB models had been cultured to maturity (i.e., after six days of culture), AS stimulation was applied to them using the SMR. A series of experiments were carried out to observe the effect of AS on the BBB. Different input powers (300 mW, 400 mW, and 500 mW) were applied to the SMR to achieve different magnitudes of shear force on the BBB. To visualize the results, cells were fluorescently stained with calcein-AM and then imaged using both fluorescence and brightfield confocal microscopy. As can be seen from both the fluorescence and brightfield images in Fig. 4(a), at a SMR input power of 300 mW, the BBB integrity was not affected. However, when the power was increased to 400 mW, the fluorescence intensity of the cells in the area affected by the AS decreased [indicated by the red arrows in Fig. 4(a)], and the brightfield image shows that the cells did not detach. It is speculated that the decrease in fluorescence intensity may be caused by calcein-AM being pumped out of the cells under the action of the AS. With a further increase in input power to 500 mW, both the fluorescence and brightfield images show that the BBB was significantly damaged, some cells were peeled off from the substrate, and the integrity of the BBB was destroyed. As shown in Fig. 4(b), measurements of the TEER value after AS stimulation with different input powers also reflected this trend. The above results all indicate that the integrity of the BBB can be destroyed under higher-power AS but remains intact under lower-power AS.

Paracellular transport of molecules across the BBB is highly restricted mainly due to the presence of TJs and AJs between adjacent endothelial cells, which are critical for the maintenance of BBB structure and function.¹⁶ To further explore the mechanism of the regulation of BBB by AS when the power of the SMR is 400 mW, a representative TJs, ZO-1, was selected to study its changes under AS stimulation. The ZO-1 protein was stained with a ZO-1 monoclonal antibody conjugated with a red fluorescent molecule, and three-dimensional layer scanning was performed with a confocal microscope to obtain three-dimensional fluorescence images. As can be seen from Fig. 5(a), when there was no AS stimulation, ZO-1 protein in the BBB filled the gaps between cells to form a tight barrier. However, Fig. 5(b) shows that when 400 mW AS was applied to the BBB, the ZO-1 protein between cells was disrupted and pores appeared between cells. From the perspective of TJs, this experiment proves that AS can increase the penetration of molecules by reducing the integrity of the BBB.

The ability of drug molecules to cross the BBB is crucial for therapy of CNS diseases. However, existing methods for transporting drugs across the BBB are often complex, with severe side effects in some cases.¹³ The discovery in the above experiments that AS can create pores between dense endothelial cells by interrupting TJs suggests that this effect could be used to enhance the penetration of molecules across cellular barriers. Dextran molecules with different molecular weights (20 kDa, 40 kDa, 70 kDa, and 150 kDa) conjugated with FITC were selected for trans-BBB penetration experiments in which the concentration of dextran that penetrated through the BBB from the upper chamber and accumulated in the lower chamber within 10 min was measured. As can be seen from Fig. 6, when there was no AS stimulation, the higher the molecular weight of the dextran, the fewer molecules penetrated into the lower chamber across the BBB, which is consistent with the previous understanding that the BBB is selective to molecular size of the molecule: the larger the molecular weight of a molecule, the less likely it will be to penetrate. At a molecular weight of 150 kDa, almost no molecules penetrated through the BBB into the lower chamber in the control group without stimulation. By contrast, in the experimental group, under stimulation by 400 mW AS, about 1 nM of 150 kDa dextran molecules penetrated into the lower chamber within 10 min. For dextran with other molecular weights, the penetration in the experimental groups was also significantly better than in the control groups. This indicates that AS can promote the passage of molecules across the BBB. The promotion of molecular penetration across the BBB by AS provides a new way to solve the problem of how to deliver drugs into the brain to treat diseases.

Overall, this work has demonstrated that controllable AS generated by a gigahertz bulk acoustic wave resonator device is an effective tool to regulate the BBB and facilitate the delivery of molecules across it in a controllable manner. The experimental results confirm that the integrity of the BBB can be regulated by adjusting the input power of the resonator to generate controllable AS. This AS can facilitate molecular penetration across the BBB by interrupting TJs and thereby creating pores in the BBB. Since the resonator can generate AS of controllable magnitude and also facilitates miniaturization, this approach provides a new platform for studying the regulation of the BBB by shear forces, as well as a new method for effectively promoting the delivery of drug molecules across the BBB.

The authors acknowledge financial support from the National Natural Science Foundation of China (Grant No. 61971302).

The authors have no conflicts to disclose.

Y.W. conceived the idea and designed the experiments. H.Q. performed the experiments and analyzed the results. S.Z. performed the simulation. S.H. and J.L. contributed to the editing of the paper. H.Q. and Y.W. wrote the paper.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

 Hang Qi received the B.S. degree from Tianjin University, Tianjin, China, in 2020. She is currently working toward a M.S. degree at Tianjin University, Tianjin, China. Her research interests include blood–brain barrier models, organs on chips, and acoustic actuators.

 Shuaihua Zhang received the B.S. degree from Northeastern University, Qinhuangdao, China, in 2019. He is currently working toward a Ph.D. degree at Tianjin University, Tianjin, China. His research interests include biosensors, microfluidics fluorescence and impedance flow cytometry, droplet technology, and acoustic actuators.

 Jiaxue Liang received a B.S. degree from Jiangsu University, Jiangsu, China. She is currently pursuing a Master’s degree at Tianjin University, Tianjin, China. Her research focuses on in vitro blood–brain barrier sensing.

 Shan He received a B.S. degree from Chongqing University, Chongqing, China, in 2017, and she is studying for a Ph.D. at Tianjin University. Her research interest focuses on intracellular delivery and cell mechanics studies.

 Yanyan Wang is an associate professor at the School of Precision Instrument and Optoelectronics Engineering, Tianjin University. Her research interests include micro/nano biomedical sensors, microfluidic bioanalytical systems, and micro/nano electromechanical systems.