Effect of Hydrogel Matrix and Fluid Shear Stress on the Behavioral Regulation of Mesenchymal Stem Cells

Development of a micromodel that recapitulates multiple mechanical properties to mimic the complex mechanical microenvironment is crucial for cell‐based research. Herein, a microsystem combining structure of hydrogel matrix and acoustic streaming (AS) to mimic the cellular microenvironment is proposed, which can realize multiplex cellular mechanical cues, including matrix stiffness, fluid shear stress (FSS) generated by AS, and matrix roughness. The results demonstrate that the cell spreading area enlarges with the increase of matrix stiffness, and cell spreading is encouraged by integrin β1 cluster to polymerize actin when combines the hydrogel matrix with FSS. In addition, FSS has the influence on the roughness of the hydrogel, which further affects the cell morphology and mechanical properties, inducing mesenchymal stem cells (MSCs) differentiation into astrocytes rapidly. Meanwhile, cell migration is also enhanced by FSS stimulation, particularly, undifferentiated cells at 22 kPa hydrogel have the fastest migration speed, and change the movement model from contact inhibition to contact stimulation migration. Especially, matrix roughness and stiffness dominantly control of cell spreading and differentiation, whereas FSS affects cell migration. In conclusion, the microsystem in this work shows superior performance in regulating the spreading, differentiation, and migration of MSCs, which provides a new tool for cell‐microenvironment study.


Introduction
Mesenchymal stem cells (MSCs) are a type of cells with ease of isolation, self-renewal, unlimited proliferation, and multidirectional differentiation potential, [1] which have attracted considerable attention in tissue engineering and regenerative medicine.In recent years, the regulation of mechanical factors on the behavior of MSCs is receiving more and more attention. [2]Mechanical cues in the growth microenvironment of MSCs are important for behavior of cell spreading, differentiation, and migration, including fluid force, [3,4] matrix stiffness, [5][6][7][8][9] topography, [10][11][12] roughness, [13,14] and so on.For example, by applying different mechanical stimulation, such as fluid force and magnetomechanical stimulation to regulate cell behavior and guide cell fate. [15,16]The recent studies also show the design of the adjustable viscoelastic hydrogels to mimic soft-hard tissue interfaces has facilitated in vitro studies of cellmatrix interactions and boosted potential biomedical applications. [17,18]hese mechanical cues can initiate signal perception based on integrin transmembrane proteins, induce focal adhesion formation, cytoskeleton assembly, and transmit signals from the cytoplasm to the nucleus by the linker of nucleoskeleton and cytoskeleton complex, thereby affecting the expression of genes related to cell behaviors. [19,20]For example, the nonlinear viscoelastic mechanical signal perception and response of myoblasts to extracellular matrix (ECM) during myogenesis are related to integrin-focal adhesion kinase signaling, actin polymerizationmediated myocardin-related transcription factor nuclear localization, and nuclear mechanotransduction. [21] Hydrogel matrix is widely used to research the interaction between cells and the substrate due to its ability to mimic the characteristics of ECM.[24] In general, MSCs prefer to form stable focal adhesions on the stiff matrix, showing that the cell spreading area and cytoskeletal organization increased with the matrix stiffness. [25]In addition, MSCs grown on the matrix of different stiffness tend to differentiate into tissue cells with corresponding stiffness, such as adipocytes on softer substrates and osteoblasts on stiffer substrates. [9,26]However, the osteogenic induction ability of MSCs was weakened when the cells were transferred to stiff substrate after long-term preconditioning of soft matrix. [27]oreover, matrix roughness also plays an extremely significant role in the manipulation of MSCs behaviors.Enhancement of osteogenesis on the hydrogels with high surface roughness than on smooth stiff substrates has been reported. [13]It has been found that increasing surface roughness slightly decreased cell spreading areas and inhibited osteogenic differentiation. [14]owever, most current researches only pay attention to one mechanical property of the matrix, while the cells are regulated by multiple mechanical cues.Therefore, it is necessary to combine the matrix properties with other mechanical factors to investigate the synergistic effect on cell behaviors.
MSCs in vivo are often in a complex dynamic microenvironment stimulated by various fluid forces due to body fluid flow.The previous researches have shown that fluid force had many effects on MSCs.For example, MSCs could respond to perfusion flow-induced shear stress and differentiate into vascular cell or osteoblast. [28]Under the appropriate rate of fluid force, the fate of rat bone marrow MSCs could be controlled toward osteogenic or chondrogenic differentiation. [29]Thus, the combination of hydrogel matrix and fluid force provides an alternative pathway to mimic the microenvironment of cell growth.
Currently, the generation of fluid force is mainly achieved through the microchannel powered by external pump, but the structure of the microchannel limits the space for cell growth and has poor adjustability.Recently, the acoustic streaming (AS) generated by the microscale gigahertz (GHz) bulk acoustic wave resonator has applied in cell manipulation, including controllable cell deformation, [30] intracellular drug delivery, [31] drug penetration, [32] cell mechanical motion, [33] and neurite outgrowth, [34] due to its advantages in wider mechanical control and controllable flow velocity distribution, and it is easier to control the spatial range of the applied area.
In this work, a small structured system combining hydrogel matrix with GHz AS to mimic the microenvironment was proposed to study the regulation of human bone marrow mesenchymal stem cells (hBMSCs) behaviors.When acting GHz AS on the hydrogel matrix, the roughness of matrix altered, thus forming three kinds of mechanical factors on the adherent cells: matrix stiffness, fluid shear stress (FSS), and matrix roughness.Under the synergistic effect of above three mechanical cues, the spreading area of the cells increased.Moreover, hBMSCs grown on the 22 kPa hydrogel were induced to differentiation into astrocytes, and cell migration was promoted after AS stimulation.

Interaction between Hydrogel Matrix and GHz AS
The mechanical effect is very important for the cell behavior regulation.In vivo, tissue stiffness and the fluid force generated by the body fluid are the main mechanical forces on the cells.In order to mimic the mechanical state of cells in vivo, a microsystem based on the combination of hydrogel matrix and GHz AS (hydrogel-AS) was established (Figure 1a), including a 1.58 GHz solid mounted thin-film piezoelectric resonator (SMR), a polydimethylsiloxane (PDMS) chamber, and hydrogel matrices with different stiffness.When SMR is placed into the liquid environment, by which the acoustic waves energy generated will decay rapidly to drive the liquid flow and generate high-speed streaming (microvortices).Due to the focused acoustic wave at high resonant frequency (> GHz), a localized AS zone can be achieved according to the size of the device.By further placing the lid (here is hydrogel seeded with cells) opposite to the SMR, the strength of the AS, flow speed, and FSS applied on the hydrogel can be readily controlled.This provides a dedicate tool for applying mechanical stimulation on cell behavior regulations.In addition, four kinds of hydrogels with different stiffness were prepared by varying the concentration of monomer and cross-linker (Table S1, Supporting Information).The Young's modulus of them was measured by atomic force microscopy (AFM), which were 2, 22, 50, and 118 kPa, respectively (Figure S1, Supporting Information).To investigate the effect of hydrogel on GHz AS, a 2D finite element simulation model was established under the combined stimulation of GHz AS and hydrogel matrix or coverslip (Figure 1b), which demonstrated the microvortices were generated, and the shape of vortices did not change significantly with hydrogel or coverslip.In addition, the maximum flow velocity was concentrated near the SMR, and the flow velocity near the hydrogel or coverslip was about 0.01 ≈ 0.02 m s À1 .The result confirmed that the hydrogel had almost no effect on the AS distribution.
To explore the effect of GHz AS on the hydrogel matrix, the mechanical properties of four hydrogels with different stiffness and coverslip were measured by AFM before, during, and after the AS stimulation (Figure 1c and Figure S2, Supporting Information).Coverslip was used as a control group to exclude the effect of AS on the AFM probe.Figure 1d gives the roughness of the hydrogel before, during, and after the AS stimulation.The result suggested that when the AS acted on the hydrogel, its roughness would rise rapidly.The softer the matrix, the greater the roughness changed.The roughness of the hydrogel decreased when withdrawal of the AS stimulation, but it was still greater than that of the hydrogel before the AS stimulation, demonstrating that the effect of AS on the hydrogel would change the roughness of the matrix.The Young's modulus of the hydrogel before and after the AS stimulation is shown in Figure 1e, which indicated that the Young's modulus of the hydrogels with different stiffness decreased after the AS stimulation.Notably, as a stiff substrate, the roughness and Young's modulus of coverslip did not change prominently before and after AS stimulation (Figure S2b,c, Supporting Information), confirming that AS had no significant effect on the AFM probe.These results confirmed that GHz AS could change the roughness and Young's modulus of the hydrogel matrix.
In order to better compare the mechanical changes of hydrogels with and without AS stimulation, the relative changes of roughness and Young's modulus of hydrogels were statistically analyzed (Figure 1f ).The relative change is defined as the ratio of the absolute change before and after the AS stimulation to the initial value before the AS stimulation (|Variate After AS À Variate Before AS |/Variate Before AS ).It was observed that as matrix stiffness increased, the relative variation of Young's modulus of hydrogels decreased gradually, and the roughness decreased accordingly, but the relative change of roughness increased at 22 kPa.All the results indicated that FSS generated by GHz AS had a great influence on the soft matrix and a small influence on the stiff matrix.Meanwhile, compared with the Young's modulus, the effect on the roughness of hydrogels by FSS was more significant, and a more pronounced effect was observed as matrix stiffness decreased.All the above results indicated that FSS caused by GHz AS could increase the roughness and reduce the Young's modulus of the matrix, thus the combination of hydrogel matrix and GHz AS could produce three mechanical factors, including matrix stiffness, FSS, and matrix roughness, which provided synergistic effect on the cell behavior regulation.

Matrix Stiffness Regulated Cell Spreading
The cell-matrix interface has been shown to exert a considerable influence on stem cell shape and function. [10,35]In order to study the effect of matrix stiffness on hBMSCs, the polyacrylamide hydrogels, one of the most widely used hydrogels in regenerative medicine and stem cell research, with Young's modulus of ≈ 2, 22, 50, and 118 kPa were utilized.The hBMSCs with concentration of 0.2 Â 10 5 /mL were seeded on the hydrogels with different stiffness and cultured in growth medium for 7 days, and the bright field images are shown in Figure 2a.The hBMSCs exhibited distinct cell morphology on hydrogels with different stiffness.Highly spreading morphology was observed in most cells on the 118 kPa hydrogel, while cells on the 2 kPa hydrogel were smaller, illustrating that the cell spreading was more notable with the increase of matrix stiffness.The statistics of cell area verified that the average cell area increased with hydrogel stiffness (from 2 to 118 kPa), same as observed in previous studies. [36]Oppositely, the aspect ratio of the cells decreased with the increase of the matrix stiffness, which demonstrated that stiff matrix promoted cell spreading.It can be inferred that due to the increase of matrix stiffness, the cells gradually spread and changed from irregular shape to polygon, resulting in the decrease in the aspect ratio of the cells.
The integrins act as the cell surface receptors that connect the ECM to the intracellular actin cytoskeleton, thereby mechanically integrating extracellular and intracellular compartments. [37]n particular, integrin β1, as the most important integrin protein, plays an important role in the adhesion between cells and ECM.In order to explore the response mechanism of matrix stiffness to cell morphological changes, the F-actin was stained with phalloidin (red color) and integrin β1 was stained with antiintegrin β1 and Alexa Fluor 488 (green color) after 7 days of culture (Figure 2a and Figure S3, Supporting Information).With the increase of matrix stiffness, aligned F-actin was more neatly.Highly organized cytoskeleton was observed in most cells on the 118 kPa hydrogel, while cells on the 2 kPa hydrogel showed monomeric or spot-like structure.Meanwhile, more punctuated integrin β1 was stained on the membrane of cells which grew on the stiffer hydrogel, demonstrating promoted ECM-integrin interaction.In addition, a close correlation between cell spreading and actin assembly was found that cells with highly spreading also tended to have aligned actin assembly, which is aligned with the relationship between actin assembly and cell spreading shown in previous researches. [15]Above results demonstrated that the spreading behavior of cells was pronouncedly altered by matrix stiffness.
The formation of cytoskeleton is directly related to cell adhesion. [38,39]In order to analyze the effect of matrix stiffness on cell mechanical properties, AFM was used to obtain the height image and force curve of hBMSCs after 1 day of growth on four kinds of hydrogel matrixes (Figure 2b and Figure S4, Supporting Information).The Young's modulus and adhesion force of the cells were analyzed based on the force curve.It can be seen that the Young's modulus of the cells and adhesion force between the cells and AFM probe gradually increased with stiffness of matrix.As discussed above, the degree of actin polymerization increased with the increase of matrix stiffness.Most recent studies have concluded that the cytoskeleton is the most important determinant of cell stiffness, as its dissolution fluidizes the cells, while its stability or increased assembly makes the cells stiffer, [40] which is consistent with our findings.With the increase of matrix stiffness, the cytoskeleton assembly effect was more obvious and the Young's modulus of cells also increased, thus improving the cell stiffness.In particular, the Young's modulus of the cytoplasm and nucleus and the adhesion force with the AFM probe change with the substrate stiffness similar to that of the cell, but the adhesion force between the cytoplasm and the AFM probe is higher than that of the nucleus.As cell adhesion is specifically mediated by integrin receptors, [41] it is speculated that high adhesion in the cytoplasmic region was related to the distribution of integrin β1.Additionally, it can be hypothesized that the cells with strong adhesion were highly spreading and the actin cytoskeleton was well organized.The above results confirmed that the stiff matrix promoted the spreading and adhesion of cells.

FSS Stimulation Promoted Cytoskeleton Assembly
FSS is of vital importance for the regulation of cell growth, and the cells under different scales of FSS display distinguishable behaviors. [42]Our previous study has found that the power and distance of the AS are positively correlated with the flow velocity, and the GHz AS had no effect on cell activity, which produced FSS at the μN level. [30]In order to find the optimal parameters of AS stimulation on hBMSCs, the stimulation power, duration time, and distance of SMR from cells were adjusted to achieve FSS of different scales.After the cells were seeded on the coverslip for 24 h, the FSS induced by AS was applied to the adherent cells.Subsequently, the cell state was recorded by inverted phase contrast microscope.On the 7th day, the cells were fixed, stained with phalloidin and anti-integrin β1, and then observed by confocal microscope (Figure 3a). Figure 3b,c shows the cell images of hBMSCs under AS stimulation at different power and time.It can be seen that the cells spread larger after 4 days of AS stimulation compared with that of day 0. In order to compare the effects of AS with different power and time on cell growth more clearly, the area and aspect ratio of cells were analyzed within 4 days after AS stimulation.When the power was less than 300 mW, the aspect ratio of cells decreased with the increase of power and culture time.When the power was 500 mW, the aspect ratio of the cells increased slightly.In addition, when the stimulation time was less than 10 min, the cell area increased; while the stimulation time was 15 min, the cell area decreased, illustrating the higher FSS generated by AS with ultralong stimulation time may inhibit cell spreading.In brief, the AS stimulation with power of 300 mW and time of 10 min was the most suitable FSS for cells, which could promote cell spreading under these parameters.
Finally, the distant of AS stimulation was explored (Figure 3d).In order to ensure the effect of GHz AS on cell activity, hBMSCs were stained with calcein-AM before AS stimulation (Figure S5, Supporting Information).It can be observed that AS had little effect on cell activity.Nevertheless, a little part of cells fell off with the stimulation distant of 1.0 mm, while the cells adhered well with the stimulation distant of 3.0 and 5.5 mm.After 7 days of culture, the F-actin and integrin β1 were visualized by confocal microscope (Figure S6, Supporting Information).First, the cell area and the aspect ratio were analyzed.As the distant of the stimulation decreased, the area of the cell increased and the aspect ratio decreased, indicating that the FSS was positively correlated with cell spreading.Compared with the no AS treatment group (control group), the cytoskeleton after AS stimulation was mostly filamentous, and the arrangement was more orderly with the decrease of the distant of the stimulation, which illustrated that the FSS generated by AS promoted cytoskeleton assembly.Furthermore, integrin β1 was primarily localized in the cell membrane and cytoplasm without AS stimulation, while it acquired a more cytoplasmic distribution with AS stimulation (Figure S6, Supporting Information).The statistical analysis of integrin β1 further showed that with the decrease of stimulation distance, the average fluorescence intensity of integrin β1 was higher and the adhesion length was larger.Previously, our study has confirmed that the FSS generated by AS induced neurite elongation by mediating the formation of actin stress fibers rather than through the mitogen-activated extracellular signalregulated kinase pathway. [34]Therefore, it is speculated that the integrin β1 could sense the mechanical signal of FSS and transmit the signal to the cytoskeleton, which affected the cell behaviors by regulating its arrangement and assembly.As mentioned above, there was a positive relationship between integrin β1 cluster, cytoskeleton assembly, and cell spreading: under the stimulation of AS, integrin β1 cluster and actin assembly were promoted and cell spreading was enhanced.In summary, when the stimulation power was 300 mW, the stimulation time was 10 min and the stimulation distant was 3.0 mm; the regulation of FSS based on AS on cells was the most suitable.Therefore, these parameters of AS were selected in the following experiments.

Synergistic Effect of Matrix and FSS on Cell Spreading
After finding the promoting effect of matrix stiffness and FSS on cell spreading, a multiple model that combined matrix stiffness and FSS applied on hBMSCs was explored by applying the AS to the hydrogel matrix with cells.The whole experimental process is shown in Figure 4a.Cells were seeded on the hydrogel for 1 day first; then the FSS generated AS was applied to the adherent cells.Cell images before and after AS stimulation are shown in Figure S7 (Supporting Information).
The morphology and mechanical properties of the cells were analyzed by confocal microscopy (Figure 4b and Figure S8, Supporting Information) and AFM (Figure 4c and Figure S9, Supporting Information).Compared with the cells on hydrogels alone, the cells on hydrogels combined with AS stimulation were more spread out and the cytoskeleton was arranged more orderly.Especially, actin arrangement in the cytoplasmic region was observed in cells on the 2 kPa hydrogel after AS stimulation, while the actin was mostly punctate on the 2 kPa hydrogel individually.The statistics of the average fluorescence intensity and adhesion length of integrin β1 with or without AS showed that the average fluorescence intensity and adhesion length increased significantly after AS stimulation.These results indicated that the FSS generated by AS promoted the expression of integrin β1 in hBMSCs grown on the hydrogel matrix, thereby activating the cytoskeleton assembly.Consistent with the effect of matrix stiffness individually, increasing hydrogel stiffness after AS stimulation led to an increase in the cell spreading area and a decrease in the cell aspect ratio, while the Young's modulus of the cells and adhesion force between the cells and AFM probe increased.In addition, the variations of cell area and aspect ratio, Young's modulus, and adhesion force with and without AS were analyzed (Figure 4d,e).Specifically, the cell area and aspect ratio increased, while the Young's modulus of cell and adhesion force between cells and AFM probe decreased after AS stimulation.For example, the cell area increased by 176 μm 2 and the cell aspect ratio increased by 0.32 at 22 kPa after AS stimulation, while the Young's modulus of cell decreased by 1.24 kPa and cell adhesion with AFM probe decreased by 0.03 nN at 22 kPa after AS stimulation.It should be noted that the change of cell area on 22 kPa was the most obvious, which may be due to the maximum change of roughness of 22 kPa matrix caused by FSS.
The variations of Young's modulus of cytoplasm and nucleus and adhesion force with AFM probe without or with AS are displayed in Figure S10 (Supporting Information).The Young's modulus of the nucleus changed more significantly than that of the cytoplasm, and showed a decreasing trend with the increase of matrix stiffness, whereas the nucleus showed an increasing trend.It was speculated that it could be due to the fact that the FSS based on AS preferentially worked on the nucleus compared with the cytoplasm, leading to have a greater effect on the nucleus.Although the influence of FSS on the adhesion force between the nucleus and AFM probe was greater than that of the cytoplasm, the adhesion force of them did not alter significantly with the stiffness of the matrix.These results indicated that FSS reduced the Young's modulus of hBMSCs and adhesion force between hBMSCs and AFM probe, and preferentially acted on the nucleus when compared to the cytoplasm, resulting in a greater change in the mechanical properties of the nucleus.
These observed findings confirmed cell spreading was promoted after long-term culture of cells with AS.Notably, due to the effect of AS on the matrix roughness, the spreading area of cells on 22 kPa hydrogel changed more obviously after AS treatment.In conclusion, compared with matrix stiffness and FSS, matrix roughness significantly affected cell spreading.

Synergistic Effect of Matrix and FSS on Cell Differentiation
Adhesion-mediated signal induces cell pseudopod formation and drives cell morphological change to explore the surrounding environment. [43,44]Our results illustrated that the morphology of hBMSCs cultured on the hydrogels was slightly different without and with the stimulation of AS (Figure 5a).The further quantification of pseudopods (Figure 5b,c) showed that cells grown on the 22 kPa hydrogel exhibited more and longer pseudopods than cells grown on hydrogel of other stiffness.In order to better compare the effects of matrix stiffness combined with FSS on cells, the relative changes of the number and length of pseudopods, spreading area and aspect ratio, Young's modulus of cells, and adhesion with AFM probe were statistically analyzed (Figure 5d,e,f ).Among them, the relative variation is defined as the absolute variation of the cells cultured on different stiffness hydrogels before and after AS stimulation and the initial value of the cells before AS stimulation (|Variate After AS À Variate Before AS |/Variate Before AS ).These results showed that the relative changes of the number and length of pseudopods at 22 kPa were the largest among four hydrogels with different stiffness, and its relative changes of adhesion and spreading area were also the largest, indicating that FSS had the greatest effect on the cells at 22 kPa.It may be speculated that FSS exerted the most significant effect on the roughness of the 22 kPa hydrogel, resulting in the most alterations with the number and length of cell pseudopods, cell area, and adhesion force at 22 kPa under the combined effect of three mechanical factors of FSS, matrix stiffness, and matrix roughness, which drove the change of cell morphology.
In particular, the cells grown at 22 kPa with and without GHz AS have shown the morphology of astrocytes.In order to determine the specification of the hBMSCs differentiation, the expression level of neuron and astrocytes markers was investigated.Microtubule-associated protein 2 (MAP2) is considered as a neuron maturation marker, and glial fibrillary acidic protein (GFAP) is the marker of astrocytes. [45]Therefore, immunofluorescence staining was carried out for assessing the level of MAP2 and GFAP in hBMSCs at 22 kPa without or with AS for 7 days (Figure 5g).Notably, the protein level of GFAP was higher with or without GHz AS than that of MAP2.To accurately reflect neural or glial differentiation and determine the role of FSS, the mean fluorescence intensity of MAP2 and GFAP was quantified (Figure 5h).The result showed the hydrogel matrix at 22 kPa was more favorable for the differentiation of hBMSCs into astrocytes.Meanwhile, the increased expression level of GFAP under AS stimulation was observed, suggesting that FSS promoted glial differentiation of hBMSC.Collectively, matrix stiffness dominated the glial differentiation of hBMSC, and FSS could further promote the efficiency of differentiation induction.

Synergistic Effect of Matrix and FSS on Cell Migration
In order to explore whether cell morphology and mechanical changes induced by combination of matrix and GHz AS mediate cell migration, the cell movement was tracked in real time with and without AS.After being stimulated by AS, the time-lapse images of the cells stained with calcein-AM within 10 h were recorded by confocal microscopy (Figure 6a and Movie S1-S6, Supporting Information).It has been shown that the migration of cell populations can be induced by two types of cell-cell interactions, contact inhibition of migration or contact stimulation of migration. [46][53] From Figure 6b, it was found that for the undifferentiated cells on the 22 kPa hydrogel, one cell encountering another cell would change the direction of migration (contact inhibition of migration).However, for undifferentiated cells at 22 kPa with AS, one cell contacting another cell could activate the cell migration (contact stimulation of migration).The result illustrates the FSS can induce movement of cells change from contact inhibition to contact activation migration, which may accelerate the migration of undifferentiated cells on 22 kPa.
Directional components of cell migration were also investigated by plotting individual migration traces in x and y directions (Figure 6c-h), which indicated random locomotion of cells on different matrices without or with AS.In addition, the final migration distance of each cell between the starting and ending positions was quantified (Figure 6i); it can be seen that the migration distance of cells grown on coverslips was similar to that of differentiated cells on 22 kPa, and the migration distance of undifferentiated cells on 22 kPa was the longest.Furthermore, the cell migration distance increased after AS, and the migration distance of undifferentiated cells on 22 kPa increased significantly, which was 2.2 times than that of cells without AS, indicating that FSS generated by AS promoted cell migration.The statistics of cell migration speed also supported this result (Figure 6j).With the previous studies from our group, the FSS generated by AS can promote the mechanical movement of cells by regulating cytoskeleton assembly, [33] which is consistent with our results.Moreover, it can be inferred that due to the weak adhesion of cells on 22 kPa, the migration speed was faster, and the migration speed of differentiated hBMSCs was slower, which was consistent with the fact that hBMSCs on 22 kPa differentiated into astrocytes because the migration ability of astrocytes is weaker than that of hBMSCs. [54]All findings showed that FSS could induce cells to switch from contact inhibition migration to contact stimulation migration and promote cell migration.Additionally, compared with cells grown on coverslips, cells cultured on hydrogel matrix migrated farther and more quickly.Overall, both matrix stiffness and FSS could influence cell migration, whereas FSS had a more dramatical effect on their migration.
The regulation of matrix stiffness combined with FSS on the behavior of hBMSCs is systematically summarized in Figure 7.In this work, the hydrogel was combined with GHz AS to construct the mechanical regulation of cell behavior by matrix stiffness, FSS, and matrix roughness.It was found that after AS stimulation, cell spreading on hydrogels was promoted.In addition, cells on 22 kPa hydrogel were induced to differentiate into astrocytes and their migration was promoted with AS.

Conclusion
In summary, a microsystem combining matrix and FSS to mimic microenvironment of cell growth was proposed.When the FSS   worked on the hydrogel matrix, the roughness of the matrix increased, which constituted a regulation system for the combined effect of three mechanical factors: matrix stiffness, FSS, and matrix roughness.According to the systematic analysis of cell morphology and mechanical properties, it was found that stiff matrix could facilitate the cell spreading by enhancing cell adhesion.Moreover, the combination of matrix and FSS could further enhance cell spreading by integrin β1 sensing mechanical signals to polymerize actin.Based on the regulation of cell morphology by matrix and FSS, the ability of this system to induce the directional differentiation of hBMSCs into astrocytes on the 22 kPa hydrogel was further confirmed.Finally, the timelapse images of hBMSCs on the coverslip and 22 kPa hydrogel were recorded without or with AS, showing that the migration ability of undifferentiated cells at 22 kPa was faster, and the FSS generated by AS could change the cell migration mechanism and improve cell migration.It can be concluded that when the hydrogel matrix was combined with GHz AS, the obvious change of matrix roughness led to a significant alteration in cell spreading area.In addition, the matrix stiffness was more dominant in the regulation of cell differentiation, while cell migration could be modulated by both matrix stiffness and FSS, but FSS had a more profound influence.The microsystem that combines three mechanical cues of matrix stiffness, FSS, and matrix roughness in this work provides new potential in regenerative medicine.

Experimental Section
Cell Culture: Primary hBMSCs were purchased from Oricell and cultured in hBMSC growth medium (Oricell) containing L-glutamine, supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin in a humid incubator of 37 °C and 5% CO 2 .To ensure the pluripotency of cells, the primary cells were passaged every 4-5 days within ten passages.The culture medium was changed every 2-3 days.Before experiments, cells were seeded on poly-L-lysine-modified coverslips or fibronectin-modified hydrogels with concentration of 0.2 Â 10 5 /mL.
Hydrogel Preparation: The polyacrylamide hydrogel was prepared as previously reported. [55]In brief, the coverslip was treated with 0.5% 3-(trimethoxysilyl)propylmethacrylate (Aladdin) and 0.5% glutaraldehyde (Aladdin) for 30 min, respectively.Subsequently, 40% acrylamide (Sigma-Aldrich) and 2% bis-acrylamide (Sigma-Aldrich) were mixed in different proportions, as shown in Table S1 (Supporting Information).The polymerization of hydrogel was initiated by the addition of N,N,N 0 , N 0 -tetramethylethylenediamine (final 0.2% by volume, Aladdin) and 10% ammonium persulfate (final 0.1% by volume, Aladdin) to the above mixture.Immediately, 20 μL solution was pipetted onto the glass slide and the aminosilanized coverslip was placed on the top of the droplet.After polymerization for 1 h at room temperature, 0.2 mg mL À1 of Sulfo-SANPAH (Invitrogen) diluted in 10 mM 4-(2-hydroxyethyl)-1-piperazineethan esulfonic acid (Aladdin, pH 8.5) was coated on each hydrogel and incubated for 30 min at room temperature.The hydrogel and solution were then exposed to UV (365 nm) and photoactivated in a 12-well plate for 10 min.Finally, the hydrogel was washed with PBS to remove any excess reagents and incubated with 10 μg mL À1 fibronectin (Invitrogen) solution at 4 °C overnight.The polyacrylamide hydrogel was soaked in PBS and could be stored at 4 °C for 1-2 weeks.Before seeding the cells, the polyacrylamide hydrogel was subjected to UV for 1 h.
Resonator Fabrication: The SMR, including the Bragg reflector, bottom electrode, piezoelectric layer, and top electrode, was fabricated with a micro electro mechanical systems standard process. [56,57]Briefly, the Bragg reflector consisting of three alternating layers of aluminum nitride (AlN) and silicon dioxide was mounted on a silicon substrate with plasmaenhanced chemical vapor deposition and reactive sputtering.After being deposited onto the Bragg reflector with radio frequency, the molybdenum was patterned as the bottom electrode through plasma etching.Au was then evaporated and patterned as the top electrode by a lift-off process, and AlN was deposited as a piezoelectric layer between the bottom and top electrodes.The SMR was controlled by a sinusoidal signal (1.58 GHz), which was generated by a signal generator (Agilent, N5171B) and amplified by a power amplifier (Mini-Circuits, ZHL-5W-422þ).
System Setup for Cell Regulation: The system used for cell regulation mainly included SMR, PDMS chamber, coverslip, or hydrogel with cells.PDMS was prepared according to the conventional method. [58]Then the cured PDMS was cut into 8 mm Â 8 mm blocks, and punched out a small hole with a diameter of 5 mm.After UV sterilization of PDMS chamber and SMR for 20 min, the PDMS chamber was adhered to the SMR and filled with PBS.Afterward, the coverslip or hydrogel with cells was aligned to the center of the SMR and placed above the PDMS.
When the SMR began to work, the GHz acoustic wave was coupled with the liquid, and the high-energy acoustic wave attenuated to drive the liquid flow.When the flow encountered the coverslip or hydrogel with cells, the fluid would rapidly decelerate and flow along the fluid-solid interface to form a high-speed vortex, thereby generating shear stress on the hydrogel or cell surface.
Finite Element Simulation: The simulation model of AS field and hydrogel matrix was established based on fluid-solid coupling.The attenuation of acoustic bulk force induced the generation of GHz AS.In the simulation, bulk force area was set to 1000 μm (width) Â 1000 μm (height).Liquid material was set as water, and the velocity field was described by the incompressible Navier-Stokes equation.The hydrogel was simulated using a rectangular linear elastic material with a length of 100 μm and a width of 1000 μm.
AFM: Height imaging and force curves of hydrogels and hBMSCs were obtained by quantitative imaging (QI) mode of Nano Wizard4 AFM (Bruker).The hydrogel was placed in PBS medium, while the cells were placed in hBMSC culture medium without FBS.When performing AS and AFM concurrently, the SMR was laterally placed and fixed in a petri dish containing PBS to enable it to generate the AS effect, and the distance between SMR and hydrogel was 2.5 mm.The AFM scanning was started during SMR working.The SNL-10 probe (spring constant: 0.3 N m À1 ) was chosen for hydrogel imaging, while the PFQNM-LC-A-CAL probe (spring constant: 0.1 N m À1 ) was selected for cell imaging.When scanning the hydrogel, the setpoint was set to 0.1 V and the scanning speed was set to 50 μm s À1 .When scanning the cell, the setpoint was set to 0.4 V and the scanning speed was set to 30 μm s À1 .The mechanical properties of hydrogels as well as cells were obtained with Hertz/Sneddon contact mode with JPK Data Processing software.
Time-Lapse Imaging: The growth state of the cells was observed in real time by time-lapse imaging with confocal microscope.In brief, cells were incubated with calcein-AM (Invitrogen, diluted with culture medium at 1:1000) at 37 °C for 15 min and imaged using confocal microscope over a period of 10 h.In order to maintain the normal growth of cells, cells were cultured in a small cell incubator with controlled temperature (37 °C), humidity (95%), and CO 2 (5%).The incubator was installed on the confocal microscope to achieve long-term continuous observation.During the 10 h observation, the cell images of each group were captured from five positions every 10 min.The migration trajectory and corresponding displacement of the cells within 10 h were obtained by the Trackmate plugin in Image J software.

Figure 1 .
Figure 1.Construction and mechanical characterization of hydrogel-AS system.a) The schematic diagram of hydrogel-AS system.b) 2D finite element simulation of hydrogel-AS system.AS was generated by SMR with input power of 300 mW and distant of 1 mm.c) Height images of hydrogels with different stiffness scanned by AFM before, during, and after AS stimulation.d) Roughness and e) Young 0 s modulus of hydrogels with different stiffness before, during, and after AS stimulation (n = 3).f ) The relative variation of roughness and Young's modulus of hydrogels with different stiffness (n = 3).The data are represented as mean AE standard deviation (SD).

Figure 2 .
Figure 2. The morphological and mechanical regulation of hBMSCs by matrix stiffness.a) Effect of matrix stiffness on cell morphology.(I) The bright field and fluorescent images of F-actin and integrin β1 of hBMSCs after 7 days of culture on hydrogels with different stiffness.(II) The cell area and (III) aspect ratio on hydrogels of different stiffness on day 7 (n = 30 cells).b) Effect of matrix stiffness on cell mechanics.(I) The AFM images of hBMSCs after 1 day of culture on hydrogels with different stiffness.(II) The Young 0 s modulus of the cells and (III) adhesion force between the cells and AFM probe on hydrogels with different stiffness on day 1 (n = 30 cells).(IV) The Young 0 s modulus and (V) adhesion force of cytoplasm and nucleus on hydrogels with different stiffness on day 1 (n = 30 cells).The data are represented as mean AE SD.

Figure 3 .
Figure 3.The morphological regulation of hBMSCs under GHz AS stimulation.a) The experimental diagram of hBMSCs stimulated by GHz AS. b) The effect of different power of AS on hBMSCs.(I) The bright field images of cells on days 0 and 4 under different power of AS stimulation.(II) The cell area and (III) aspect ratio of days 0, 1, 2, and 4 under different power of AS stimulation (n =s 30 cells).c) The effect of different duration time of AS on hBMSCs.(I) The bright field images of cells on days 0 and 4 under different time of AS stimulation.(II) The cell area and (III) aspect ratio of days 0, 1, 2, and 4 under different time of AS stimulation (n = 30 cells).d) The effcet of different distant of AS on hBMSCs.(I) The fluorescent images of F-actin and integrin β1 of hBMSCs on day 7 under different distant of AS stimulation.(II) The cell area and (III) aspect ratio of day 7 under different distant of AS stimulation (n = 20 cells).(IV) The mean fluorescence intensity (n = 50 cells) and (V) maximum adhesion length of integrin β1 on day 7 under different distant of AS stimulation (n = 30 cells).The data are represented as mean AE SD.

Figure 4 .
Figure 4.The morphological and mechanical regulation of hBMSCs by combination of hydrogel matrix and GHz AS. a) The experimental diagram of hBMSCs stimulated by combination of hydrogel matrix and GHz AS. b) Synergistic effect of matrix stiffness and GHz AS on cell morphology.(I) The fluorescent images of F-actin and integrin β1 of hBMSCs cultured on hydrogels of different stiffness on day 7 under AS stimulation.(II) The mean fluorescence intensity (n = 50 cells) and (III) maximum adhesion length of integrin β1 on day 7 without or with AS stimulation (n = 20 cells).(IV) The cell area and (V) aspect ratio on hydrogels of different stiffness on day 7 under AS stimulation (n = 30 cells).c) Synergistic effect of matrix stiffness and GHz AS on cell mechanics.(I) The AFM images of hBMSCs cultured on hydrogels of different stiffness on day 1 after AS stimulation.(II) The Young 0 s modulus of cells and (III) adhesion force between cells and probe on hydrogels of different stiffness on day 1 after AS stimulation (n = 30 cells).d) The variation of area and aspect ratio and e) Young's modulus of cells and adhesion force between cells and AFM probe with different stiffness without or with AS stimulation (n = 30 cells).The data are represented as mean AE SD.

Figure 5 .
Figure 5.The regulation of cell differentiation by the combination of hydrogel matrix and GHz AS. a) The representative images of hBMSCs cytoskeleton on the hydrogels with different stiffness without or with AS. b) Quantification of pseudopodia number per cell and c) maximum pseudopodia length per cell with different treatment (n = 20 cells).d) The relative variation of pseudopodia number and maximum pseudopodia length, e) spreading area and aspect ratio, and f ) Young's modulus and adhesion of cells with different stiffness without or with AS stimulation.g) Representative images of MAP2 and GFAP immunofluorescent staining on the 22 kPa hydrogel without and with AS (day 7).h) Quantification of the mean fluorescence intensity of MAP2 and GFAP (n = 100 cells).The data are represented as mean AE SD.

Figure 6 .
Figure 6.The regulation of cell migration by the combination of hydrogel matrix and GHz AS. a) Time delay images of cell moving with coverslip, 22 kPa (undifferentiated cell) and 22 kPa (differentiated cell) without or with AS. White dashed lines indicate the position of the cell before and after the movement.b) The change of cell migration mechanism under AS stimulation on 22 kPa (undifferentiated cell).c) The time-dependent trajectory map of hBMSCs with coverslip, e) 22 kPa (undifferentiated cells) and g) 22 kPa (differentiated cells) without AS (n = 30 cells).d) The time-dependent trajectory map of hBMSCs with coverslip, f ) 22 kPa (undifferentiated cells) and h) 22 kPa (differentiated cells) with AS (n = 30 cells).i) Statistics of cell migration distance and j) speed with different treatment (n = 30 cells).The data are represented as mean AE SD.

Figure 7 .
Figure 7. Summary of the regulation of hBMSCs behaviors by combination of hydrogel matrix and GHz AS.