PIEZO1-Mediated Mechanotransduction Regulates Collagen Synthesis on Nanostructured 2D and 3D Models of Fibrosis

Abstract

Progressive fibrosis can lead to tissue malfunction and organ failure due to the pathologic accumulation of a collagen-rich extracellular matrix. In vitro models provide useful tools for deconstructing the roles of specific biomechanical or biological mechanisms, such as substrate micro- and nanoscale architecture, in these processes for identifying potential therapeutic targets. Here, we investigated how the mechanosensitive ion channel PIEZO1 influences fibrotic gene and protein expression in adipose-derived stem cells (hASCs). Specifically, we examined the role of PIEZO1 and the mechanosensitive transcription factors YAP/TAZ in sensing aligned or non-aligned substrate architecture to regulate collagen formation. We utilized both 2D microphotopatterned substrates and 3D electrospun polycaprolactone (PCL) substrates to study the role of culture dimensionality. We found that PIEZO1 regulates collagen synthesis in hASCs in a manner that is sensitive to substrate architecture. Activation of PIEZO1 induced significant morphological changes in hASCs, particularly when cultured on aligned substrates, leading to a 30–40% reduction in cell spreading area and increased cell elongation, in 3D-aligned cultures. Picrosirius Red staining and immunoblotting revealed that PIEZO1 activation reduced collagen accumulation in 3D culture. While YAP translocated to the cytoplasm following PIEZO1 activation, depleting YAP and TAZ did not change collagen expression significantly downstream of PIEZO1 activation, implying that YAP/TAZ translocation from the nucleus and decreased collagen synthesis may be independent consequences of PIEZO1 activation. Our studies demonstrate a role for PIEZO1 in cellular mechanosensing of substrate architecture and provide targetable pathways for treating fibrosis and for enhancing tissue-engineered and regenerative approaches for fibrous tissue repair.

1. Introduction

Tissue fibrosis is a pathologic process characterized by the formation of a fibrous, collagen-rich extracellular matrix (ECM), increased tissue stiffness, and a loss of normal capacity to repair tissue damage and restore homeostasis. Particularly, due to increasing rates of obesity worldwide, there has been growing interest in obesity-related pathologies such as adipose tissue fibrosis [1]. Alteration in ECM mechanical properties is believed to alter crucial signals that regulate cell function, leading to continual progression of this condition [2, 3]. Despite its prevalence across various soft tissues (including adipose tissue, heart, liver, skin, kidney, and eye), effective anti-fibrotic treatments are lacking [4]. Consequently, fibrosis often necessitates organ transplantation, and is a primary cause of morbidity and mortality in a variety of chronic inflammatory disease [5]. As the precise cellular and molecular mechanisms underlying its development, maintenance, and resolution remain incompletely understood, more study is needed [6].

Conventional methods have failed in identifying inhibitors of soluble factor signaling pathways in fibrosis therapy [7], prompting the exploration of mechanobiological pathways [2]. Recognizing the significant role of mechanical forces and tissue stiffness in fibrosis development, recent investigations have delved into mechanotransduction [8–13], informing the development of strategies for modulation or reversal of fibrotic remodeling. In recent research, engineered biomimetic in vitro models have emerged as viable alternative to animal models, seeking to replicate the physical characteristics of in vivo niche [14], aiding in understanding mechanobiological pathways in fibrosis progression [7, 15]. Consideration of dimension, substrate architecture, and stiffness is crucial in disease model construction using in vitro models.

PIEZO1, a mechanosensitive ion channel, is activated when membrane tension increases [16]. It plays crucial roles in human physiology such as touch sensation, and cell mechanosignaling [16–23]. Studies have also shown that PIEZOs are involved in pathology [24–26], including fibrosis [10, 13, 27]. For example, GsMTx4 (a small molecule PIEZO1 inhibitor) can alleviate kidney fibrosis, while Yoda1 (a small molecule PIEZO1 activator) promotes profibrotic responses in renal HK2 cells [9]. Overall, prior studies have identified PIEZO1 as a potential therapeutic target for fibrotic conditions in various organs, including the renal [9, 10, 13] and cardiac fibrosis [28]. In all these cases, PIEZO1 had pro-fibrotic effects. In adipose tissue, PIEZO1 is implicated in adipocyte plasticity, with increased expression in obese mice leading to fibrotic adipose tissue development [27, 29]. While in another study on adipose tissue, PIEZO1 knockdown showed no significant change in fibrotic gene expression, such as collagens, despite excessive collagen synthesis and deposition being hallmark of fibrosis [30]. A recent study also suggested PIEZO1 inhibition as a promising strategy for minimizing fibrosis, especially in organs where fat contributes to fibrotic processes [31]. Thus, further investigation is needed to understand if and how PIEZO1 regulates mechanosensation during the progression of fibrosis in adipose tissue.

YAP (Yes-associated protein) and TAZ (Transcriptional co-activator with PDZ-binding motif), crucial nuclear transducers of mechanical signals [32], have been shown to contribute significantly to the progression of fibrosis [11, 33, 34]. Their regulation and subsequent nuclear translocation are dependent on RhoGTPase activity and actomyosin cytoskeleton tension [35, 36]. In the context of fibrotic diseases, increased expression and nuclear localization of YAP/TAZ have been observed in remodeled fibrotic lungs [11, 12] suggesting their involvement in the fibrotic process. Additionally, ECM stiffening ex vivo and liver damage in vivo have been demonstrated to activate YAP in hepatic stellate cells during liver fibrosis [37]. Notably, targeted inhibition of YAP through knockdown or pharmacological approaches prevents the activation of hepatic stellated cells and subsequent fibrogenesis in mice models [37]. While YAP and TAZ are primarily associated with fibrosis, they also play a role in adipose tissue biology, where TAZ regulates adipogenic activity in mature adipocytes [38], and the knockout of both YAP and TAZ negatively impacts adipose tissue expansion in mice [39]. Together, these findings highlight the role of YAP and TAZ in fibrotic diseases as well as their involvement in adipose tissue remodeling, potentially through their mechanosensitive regulation. However, the precise role of PIEZO1 or YAP/TAZ pathways in tissue fibrosis remains to be determined, and a better understanding could provide important opportunities for identifying new therapeutic targets for this condition.

Cells utilize mechanosensitive ion channels as a means of perceiving their physical environment. Research indicated that substrate properties and architecture influence the activity of these ion channels. For instance, PIEZO1 has been indirectly implicated in detecting the nanoroughness of substrates in interactions between neurons and astrocytes [40], and it has been observed to influence the differentiation of neural progenitor cells based on substrate stiffness [41]. Additionally, PIEZO1 has been identified as a sensor for confinement during cell migration [42]. Despite these advances, the impact of substrate architecture on PIEZO1-mediated mechanotransduction in collagen synthesis is not fully understood.

Adipose-derived stem cells (ASCs) have gained increasing interest as a valuable cell source to study fibrosis and test potential therapeutic interventions, due to their potentail antifibroitic effects. ASCs are multipotent adult stem cells with self-renewal and multilineage potential, and are easily isolated, making them a suitable cell source for in vitro studies. Three methods by which ASCs are thought to counteract fibrosis have been proposed: regenerative activity, direct cell-to-cell contact, and paracrine signaling. They interact with a variety of cell types-including immune cells, endothelial cells, and fibroblasts- all significant contributors to fibrosis. Their ability to counteract fibrosis is thought to arise from releasing growth factors that affect the different cell types they interact with [43–47]. For example, due to their paracrine factors, ASCs demonstrated therapeutic promise for liver fibrosis [48–50]. Moreover, ASCs are hypothesized to play a crucial role in the anti-fibrotic effects of autologous fat grafting, a technique showing promise in treating skin fibrosis, particularly through their secreted factors [51]. Our rationale for using ASCs in these fibrosis models serves both mechanistic and therapeutic purposes. ASCs are particularly well-suited for studying fibrosis progression mechanisms because they reside in adipose tissue, which can undergo fibrotic changes during obesity-related conditions. Their multipotent nature allows differentiation into fibroblast-like cells under specific conditions, potentially contributing to fibrosis development. Additionally, the high sensitivity of ASCs to mechanical cues provide a robust system for studying mechanotransduction in fibrosis. While our primary focus is understanding fibrosis progression, ASCs also offer therapeutic potential through their demonstrated anti-fibrotic effects via paracrine signaling and direct cell-to-cell contact. Understanding how ASCs respond to different mechanical environments can therefore inform strategies for their therapeutic application in fibrotic conditions

The goal of this study was to elucidate PIEZO1 mechanosensing mechanisms in hASCs within a stiff microenvironment, and to understand how substrate architecture influences profibrotic signaling. Both 2D model (microphotopatterning) and 3D models (3D fibrous scaffold) were utilized to investigate role of culture dimensionality in-vitro. We hypothesized that PIEZO1 signaling is regulated by the alignment or lack thereof of substrate architecture, modulating Ca²⁺ signaling and collagen synthesis, and that this process is regulated downstream via YAP/TAZ signaling. Our findings indicate a critical role for PIEZO1 in regulating collagen expression, influenced by both 2D and 3D matrix architecture. Our findings shed new light on the mechanosensitive processes involved in both physiologic and pathologic fibrous tissue formation.

2. Materials and Methods

2.1. Cell culture.

Immortalized human adipose stem cells (ASC52telo, ATCC SCRC4000^™) were expanded in monolayer in Mesenchymal Stem Cell (MSC) Basal Medium (ATCC, PCS-500–030). 10 mL of MSC supplement and 6 mL of LAlanyl- L-Glutamine (ATCC, PCS-500–040) and 0.2 mg/mL G418 (Sigma, G8168) were included as medium supplements. hASCs were seeded (30,000 cells/cm²) onto aligned patterned or unpatterned (flat) surfaces in the 2D model, or for cell culture on PCL mats with aligned or random fibers. Cells were cultivated for 7 days; cultured media (mesenchymal stem cell basal medium with ascorbic acid (50 μg/ml) was replaced every three days.

2.2. Preparation of micropatterned substrates.

On large (75.2 × 50.4 × 1 mm) microscope coverslips (Ted Pella, Redding CA), cell-adhesive patterns were produced by photoablating thin (150 nm) polyvinyl alcohol hydrogel films (PVA) (Sigma-Aldrich, St. Louis, MO) utilizing lithography techniques. Briefly, large cover glasses were activated as described in [52], and their surface was then spin-coated with PVA and S1805 photoresist. Coverslips were baked for one minute at 100°C. Either a laser writer or a mask aligner wrote the patterns. The photoresist was subsequently developed with a developer for 1 minute. After 5 minutes of drying at 70°C, cover glass surfaces were etched using reactive ion etching (RIE) device. Photoresist was removed with remover PG (Fisher Scientific, Waltham, MA) for 4 minutes. Cover glasses were then washed with deionized water. UV light was utilized to disinfect cover glasses for 1 hr. Micropatterns comprised of two micro-scale architectures: 1) aligned, with straight parallel lines (2 μm wide, 5 μm center-to-center distance), and 2) unpatterned (flat), with fully ablated areas (2 cm by 0.2 cm), as depicted in (Fig. S1). Patterns were stored at 4°C in 1X PBS (Gibco, Grand Island, NY). This technique produces patterns that are stable for one month prior to cell seeding and two weeks in culture. Pattern area was coated with 20 μg/ml fibronectin (Sigma-Aldrich, St. Louis, MO) to functionalize the ablated cell-adhesive regions prior to cell culture. Then, cover glasses were incubated at 37°C for 1 hr, washed three times with 1X PBS, and seeded with hASCs. The selective adsorption of fibronectin to patterned surfaces was visualized using fluorescently labeled (green) fibronectin (Cytoskeleton, Inc, Denver, CO). Cells began to align in the pattern direction 2 hrs after seeding (Fig. S1).

2.3. Preparation of 3D electrospun polycaprolactone (PCL) substrates.

Electrospun polycaprolactone (PCL) substrates with aligned and random fibers (Nanofiber Solutions, Dublin, OH) were used as our 3D fibrous scaffold (Fig. S3). The approximate diameter of the nanofiber polymers is 700 nm. The thickness of fiber layers on the bottom of the plate is ~ 20 μm. The fiber scaffolds are greater than 90% porous. 24 well plates or plate inserts were used for all RT-qPCR, immunolabeling and Western blot. They were rinsed 2–3 times with 1x PBS and fibers allowed to air-dry. Mats were coated with 20 μg/ml fibronectin (Sigma-Aldrich, St. Louis, MO) for 2 hrs at 37°C to functionalize fibers prior to cell culture. After 3 washes of 1x PBS, mats were pre-incubated in media for at least 2 hrs at 37°C. The media was removed by aspiration and cells were seeded on substrates with aligned or random fibers.

2.4. Immunostaining.

The media was aspirated, and the cells were washed twice with 1x PBS. Cells were fixed with 4% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA) in 1x PBS at room temperature (RT) for 15 minutes, followed by 3 washes with 1x PBS. hASCs were then permeabilized with 0.1% Triton X-100 (Sigma-Aldrich, St. Louis, MO) in 1x PBS for 5 minutes at room temperature. Samples were then rinsed twice with 1x PBS followed by blocking with goat serum (Vector^® laboratories, Newark, CA) for 30 minutes in RT. Primary antibodies for YAP (Cell signaling, Danvers, MA, 1:200), PIEZO1 (Novus Biologicals, Cat #: NBP1–78537AF67, Centennial, CO, 1:25), Collagen I (Cell Signaling, Danvers, MA, 1:1000) were used. Samples were covered by primary antibodies and incubated for 1 hr at RT, followed by three washes with 1x PBS tween for 5 minutes each. The secondary antibody was diluted according to manufacturer protocol in 1x PBS tween and incubated at RT for 1 hr. The samples were then washed three times for 5 minutes each with 1x PBS tween and stained for 20 minutes with TRITC-coupled Phalloidin (Invitrogen, Carlsbad, CA). After three washes with 1X PBS, cells were mounted using mounting media with DAPI (Vectashield, Newark, CA) on a microscope slide (Fisher Scientific, Waltham, MA).

2.5. Picrosirius Red Staining.

hASCs were fixed for 15 minutes at RT with 4% paraformaldehyde in 1x PBS. The cells were rinsed three times with 1x PBS and treated for 1 hour at RT with picrosirius red staining solution (0.1% Direct Red 80, 0.1% Fast Green FCF, dissolved in a saturated aqueous solution of picric acid, which contains 1.2% picric acid in water) followed by incubation with 1% acidified water (acetic acid). After 5 minutes, cells were washed twice with 1x PBS and afterwards kept at 4°C until imaging.

2.6. Real-time PCR.

Following the manufacturer’s instructions, RNA was extracted using a total RNA purification kit (NORGEN, Auburn, WA) followed by cDNA synthesis utilizing iScript cDNA kit (BIORAD, Hercules, CA). Quantitative real time-PCR using iTaq Universal SYBR GREEN Supermix (BIORAD, Hercules, CA) was used to assess the mRNA expression level of the target genes.

2.7. Primer Design.

Using the NCBI Primer-BLAST and Primer3 websites, primers for the target genes were designed and purchased from idtDNA. Primers are listed in Table S1. Using the 2(-ΔΔCT) approach, the relative fold change was assessed for gene expression analysis. The expression of the target gene was normalized relative to the housekeeping gene (GAPDH) and the gene expression of the control (aligned) samples.

2.8. siRNA knockdown.

Small interfering RNA (siRNA) transfections were performed utilizing Dharmafect1 (Horizon Discovery) according to the manufacturer’s guidelines. On-TARGET plus siRNA pools were used to transfect target genes. The sequences of siRNA pool for each gene are listed in Table S2. The total concentration of all siRNAs employed for transfection was 25 nM. 48 hours after transfection, RNA samples were isolated, and 96 hours after transfection, protein samples were isolated. siRNA method was used to knock down PIEZO1, YAP and TAZ genes on our hASCs culture both on micropatterns and 3D fibrous scaffold.

2.9. Western blotting.

Cells were lysed in RIPA buffer (Cell Signaling, Danvers, MA) with 2.5% CHAPS (Sigma-Aldrich, St. Louis, MO) and protease inhibitor (ThermoFisher, St. Louis, MO). Then, protein content was assessed using the BCA Assay (ThermoFisher St. Louis, MO). 25 μg of proteins were separated in each well using 6% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel with pre-stained molecular weight markers (ThermoFisher St. Louis, MO) and transferred to PVDF (polyvinylidene) membranes. The PVDF membranes were incubated overnight at 4°C with anti-PIEZO1 (Protein tech, Rosemont, IL) and anti-GAPDH (Protein tech, Rosemont, IL) antibodies. The relevant HRP-conjugated secondary antibodies (Cell Signaling, Danvers, MA) were utilized for detection. Using the iBright FL1000 Imaging System, immunoblots were imaged and evaluated (Thermo Fisher, St. Louis, MO). After normalization with the signal intensity of GAPDH, histograms reflect the signal intensity and area of protein bands in arbitrary units.

2.10. Imaging with Confocal microscopy.

hASCs cultured on fibronectin coated micropatterns or PLC mats were imaged using a confocal microscope (LSM 880, Zeiss, Dublin, CA). Z-stack images were acquired with a 40x objective. Images were taken from at least four field of views from each sample. Alexafluor 647 was used for secondary antibody YAP, Alexafluor 488 was used for phalloidin, 405 fluorophore was used for DAPI.

2.11. Ca²⁺ Imaging.

Cells were cultivated in MatTek Petri dishes coated with fibronectin. Before imaging, hASCs were stained for 1 hr with Fura red-AM (ThermoFisher Scientific, St. Louis, MO) and Fluo-4-AM (ThermoFisher Scientific, St. Louis, MO). Before imaging, the staining medium was removed and replaced with imaging media (HBSS media). The images were captured with a confocal microscope (LSM 880, Zeiss, Dublin, CA) fitted with a 10x objective. During imaging, the temperature was maintained at 37°C using an incubation stage. One minute of baseline time series photos were captured before the DMSO (control) solution was added to the well. The imaging continued for an additional 2 minutes. The PIEZO1 agonist Yoda1 was added to each well at concentrations of 1 μM, 2.5 μM, 5 μM, or 10 μM for an additional 2 minutes, and the Ca²⁺ signaling response of hASCs was recorded.

2.12. Image Analysis.

To quantify the ratio of nuclear/cytoplasmic YAP fluorescence intensity values, we used the following formula: $\frac{\raisebox{1ex}{${\sum }_{nuc}I$}\!\left/ \!\raisebox{-1ex}{$A_{nuc}$}\right.}{\raisebox{1ex}{$\sum _{cyto}I$}\!\left/ \!\raisebox{-1ex}{$A_{cyto}$}\right.}$ , Where ${\sum }_{nuc}I$ and $\sum _{cyto}I$ indicate the cumulative intensity values of pixels in nuclear and cytoplasmic region respectively, and $A_{nuc}$ and $A_{cyto}$ are the area of the corresponding regions quantified using ImageJ. For the quantification of cell morphology, we employed ImageJ to calculate cell, nucleus area and aspect ratio. Cell aspect ratio was calculated by fitting an ellipse to the cell body or nucleus based on the fluorescent signals from the phalloidin and DAPI channels. The aspect ratio of cell or nucleus was determined as the ratio between the longest and shortest diameters of the fitted ellipse.

In Ca²⁺ signaling experiments, a ratiometric study using ImageJ was conducted. After thresholding and using the analysis particle tool, cell pixel intensities were quantified and normalized to control section values. The highest normalized intensity (Fmax/F) represented cell signal. Cells with intensity exceeding the control mean plus five standard deviations were considered to be responding cells. Calculations were performed using a custom MATLAB program.

2.13. Statistical Analysis.

For the examination of gene expression, YAP localization and morphological data, a two-way ANOVA followed by a post-hoc test was utilized.

3. Results

3.1. Human adipose stem cells (hASCs) express functional PIEZO1 channels.

We first examined the presence and functionality of PIEZO1 in hASCs. By utilizing Western blotting, we confirmed the presence of PIEZO1 protein expression in hASCs (Fig. 1A). When exposed to the PIEZO1-specific chemical agonist Yoda1, hASCs responded by increasing the intracellular Ca²⁺ levels as measured by peak Ca²⁺ signal intensity values (Fig. 1B). Yoda1 (1μM) induced a delayed, sustained Ca²⁺ elevation, while higher doses caused a rapid, pronounced Ca²⁺ increase with quicker return to baseline (Fig. 1B). This effect was dose-dependent (P<0.0001), 94.4% with 10 μM and 83.7% response with 2.5 μM Yoda1 (Fig. 1C). We note that approximately 25% of cells exhibited Ca²⁺ responses without Yoda1 exposure. This baseline activity may be attributed to spontaneous Ca²⁺ oscillations, minor mechanical stimulation during the experimental procedure, basal PIEZO1 activity, or the involvement of other Ca²⁺ signaling pathways. These data indicate that ASCs express functional PIEZO1 ion channels.

Fig. 1. hASCs express functional PIEZO1.

A) Protein expression of PIEZO1 in hASCs by western blotting. 286 kD, and 36 kD are expected sizes for PIEZO1, and GAPDH proteins respectively. B) Normalized intracellular Ca²⁺ fluorescence intensity ΔF_max/F in response to different doses of Yoda1. C) Percentage of cells with Ca²⁺ response for different Yoda1 concentrations. Data presented as mean ± SEM. For C and D, n=3 samples; for group comparison B, C, one-way ANOVA with Dunn post hoc test, ****P<0.0001.

To directly test the role of PIEZO1 in hASC responses, we used siRNA knockdown [53]. We verified knockdown efficiency in mRNA and protein levels using real time PCR, western blot analysis and immunostaining (Fig S2. A, B, C). While PIEZO1 primarily functions at the plasma membrane, the intracellular signal we observed (Fig S2.C) likely represents protein molecules trafficking through the secretory pathway during transport to the membrane [54, 55]. Gene expression analysis showed 92.5 % knockdown efficiency in mRNA level (P<0.05) and Western blot data showed 73% knockdown efficiency in protein level.

3.2. PIEZO1 modulates collagen synthesis by hASCs

We utilized our multicellular micropattern model system with two configurations: aligned, with straight parallel lines, and unpatterned (flat), with fully ablated areas on glass surface to assess hASCs’ collagen synthesis in response to PIEZO1 modulation on aligned versus non-aligned substrate architecture. As excessive collagen deposition is a hallmark of fibrosis, there has been increasing interest in studying the regulation of collagen expression to better understand the progression of fibrosis [2]. To investigate collagen synthesis, we measured gene expression, intracellular protein production, and initial deposition of collagens. We evaluated collagen synthesis by hASCs cultured on micropatterns using three different conditions: a control group (no treatment), a PIEZO1 antagonist group (treated with 2 μM GsMTx4, with media changes every three days), and a PIEZO1 agonist group (exposed to 2.5 μM Yoda1 for 30 minutes daily). After seven days, we performed Picrosirius red (PSR) staining of cells to visualize collagens. We found that GsMTx4 resulted in increased collagen synthesis, whereas Yoda1 reduced collagen synthesis (Fig. 2A, Fig. S6A). Moreover, PSR staining of hASCs cultured for seven days on both aligned and flat substrates showed enhanced collagen synthesis in PIEZO1 siRNA-treated groups compared to those treated with non-targeting siRNA (Fig. 2B, Fig. S6B).

Fig. 2. PIEZO1 modulates collagen synthesis in hASCs.

A) Picrosirius red staining of hASCs in response to PIEZO1 antagonist (2 μM, GsMTx-4), and PIEZO1 agonist (2.5 μM, Yoda1) in micropatterns. B) Picrosirius red staining of Nontargeting siRNA and PIEZO1 siRNA on micropatterns. C) Picrosirius red staining of hASCs in response to PIEZO1 antagonist (2 μM, GsMTx-4), and PIEZO1 agonist (2.5 μM, Yoda1) on 3D fibrous scaffold. D) Picrosirius red staining of Nontargeting siRNA and PIEZO1 siRNA on 3D fibrous scaffold. E) mRNA levels of COL1A1 and COL3A1 normalized to GAPDH in response to 2 μM GsMTx-4, and 2.5 μM Yoda on micropatterns. F) mRNA levels of COL1A1 and COL3A1 normalized to GAPDH in response to Nontargeting siRNA and PIEZO1 siRNA G) mRNA levels of COL1A1 and COL3A1 normalized to GAPDH in response to 2 μM GsMTx-4, and 2.5 μM Yoda on 3D fibrous scaffold. H) mRNA levels of COL1A1 and COL3A1 normalized to GAPDH in response to Nontargeting siRNA and PIEZO1 siRNA on 3D fibrous scaffold. Data presented as mean ± SEM. For E, n=3 samples; for F, G and H, n= 5 or 6 samples; for group comparison E, F, G, H, two-way ANOVA with Dunn post hoc test, *P<0.05. All data are normalized to the control aligned group. Our statistical analysis involved two types of comparisons: All groups were compared to the control aligned group. The flat/random groups were also compared to their respective control groups.

While 2D models provide reductionist systems to test specific architectural features on cell responses, 3D models can provide additional information and better replicate the complex architecture and cellular interactions found in living tissues compared to 2D models. Therefore, collagen synthesis was also investigated in a three-dimensional (3D) model. hASCs were cultured on 3D polycaprolactone (3D fibrous scaffold) substrates with aligned and random oriented fibers for a period of seven days. Picrosirius Red (PSR) staining (Fig. 2C, Fig. S6C) showed that the inhibition of PIEZO1 via GsMTx4 increased collagen synthesis, while activation of PIEZO1 through Yoda1 exposure either reduced collagen synthesis (on 3D-aligned substrates) or had no effect (on 3D-random substrates). In line with our findings in 2D, PIEZO1 knockdown increased collagen synthesis in 3D (Fig. 2D, Fig. S6D).

Inhibition of PIEZO1 also led to an upregulation of the collagen genes COL1A1 and COL3A1 compared to untreated controls on 2D-aligned surfaces (Fig. 2E, P<0.05). While overall COL1A1 expression was higher in cells cultured on unpatterned (flat) surfaces, compared to those grown on aligned surfaces, the increase in COL1A1 and COL3A1 gene expression due to PIEZO1 inhibition was not significant on flat surfaces. Treatment with Yoda1 did not have significant effect on collagen gene expressions. Additionally, PIEZO1 knockdown significantly increased COL3A1 expression in cells on flat substrates and increased both COL1A1 and COL3A1 expression in cells on aligned substrates (Fig. 2F, P<0.05). These findings suggest that collagen expression is regulated by both substrate architecture (2D-aligned vs. 2D-flat) and PIEZO1 activity, with more pronounced effects on COL1A1 than COL3A1. We recognize that cellular mechanotransduction pathways can be sensitive to experimental conditions, including transfection processes. Even non-targeting siRNA transfection can induce mild cellular stress, potentially altering baseline cellular responses. Despite these baseline variations, comparing PIEZO1 siRNA groups to their respective non-targeting siRNA controls allows us to isolate PIEZO1-specific effects on collagen production and cellular behavior.

COL1A1 and COL3A1 mRNA levels remained relatively unchanged under exposure to either GsMTx4 or Yoda1 in 3D (Fig. 2G). The only significant difference was in COL3A1 levels, between cells exposed to Yoda1 and grown on randomly oriented fibers, and the aligned controls (Fig. 2G, P<0.05). mRNA levels of both COL1A1 and COL3A1 showed a significant, 2.5–3-fold increase following PIEZO1 knockdown in cells grown on aligned 3D substrates (Fig. 2H, P<0.05). Cells on 3D-random substrates did not show a significant change in collagen gene expression. These results indicate that in 3D environments, while the effects of substrate architecture and PIEZO1 modulation on collagen expression are more subtle than in 2D conditions, PIEZO1 knockdown demonstrates a more pronounced effect on COL1A1 compared to COL3A1.

To specifically quantify collagen I protein expression in a 3D context, we used immunolabeling and Western blotting on hASCs cultured on 3D fibrous scaffold. hASCs were exposed to Yoda1 to activate PIEZO1, or PIEZO1 was knocked down and collagen protein were measured when in the presence or absence of PIEZO1. Western blots showed that PIEZO1 activation was associated with an approximately 2-fold reduction in collagen levels in cells cultured on both aligned and random fibers (Fig. 3A, C). Immunostaining results were consistent with Western blots. Conversely, the knockdown of PIEZO1 increased collagen at protein levels on both aligned and random fibers (Fig. 3B, D). Western blot data for COL1A1 in 3D is presented as a representative example. Future studies could extend this analysis to other 2D configurations, COL3A1, additional time points, and other relevant proteins. Such extensions would enhance our understanding of PIEZO1’s role in collagen regulation across different contexts.

Fig. 3. PIEZO1 modulates collagen synthesis in hASCs at the protein level.

A) Representative nuclear (blue), actin (green), and Collagen (pink) staining of hASCs cultured on 3D fibrous scaffold for Control and 2.5 μM Yoda1 groups. B) Representative nuclear (blue), actin (green), and Collagen (pink) staining of hASCs cultured on 3D fibrous scaffold for Non targeting siRNA and PIEZO1 siRNA groups. C) Protein expressions of Collagen I in control and 2.5 μM Yoda1 groups by western blotting (left). Normalized intensity values of protein expression quantified using WB (right). D) Protein expressions of Collagen I in Non targeting siRNA and PIEZO1 siRNA groups by western blotting (left). Normalized intensity values of protein expression quantified using WB (right).

3.3. PIEZO1 activation leads to increase cytoplasmic YAP

YAP localization serves as an indicator of cell responses to mechanical cue in the microenvironment; YAP is in its active form when it is localized in the Nucleus [35]. In order to test if the YAP/TAZ pathway was involved in the PIEZO-mediated changes in collagen in hASC cells, we quantified YAP localization following PIEZO1 activation with Yoda1. Nuclear accumulation of YAP or TAZ has been widely used as a primary indicator of the active YAP/TAZ-associated signaling pathways. Nuclear-to-cytoplasmic (N/C) ratio of YAP serves as a read out of cellular mechanosensing [35]. hASCs cultured on either aligned or flat micropatterned surfaces were treated with (or without) Yoda1 for 1 hour. Then, immunostaining and confocal imaging were used to measure changes in YAP localization and fluorescence intensity ratios. Image analysis revealed that the activation of PIEZO1 in hASCs induced a decrease in nuclear/cytoplasmic ratio of YAP (Fig. 4A, C) on either aligned or flat substrates (P<0.05). Our immunofluorescence analyses indicated that PIEZO1 activation on hASCs cultured on 3D fibrous scaffold also decrease in nuclear/cytoplasmic ratio of YAP (P<0.05) (Fig. 4B, E). These findings show that PIEZO1 modulation has a significant influence on YAP activity in 2D or 3D cellular environments.

Fig. 4. PIEZO1 activation modulates YAP activity and influences cell morphology.

Representative nuclear (blue), actin (green), and YAP (pink) staining of control and Yoda1 groups cultured on A) micropatterns (2D model, top) and B) on 3D fibrous scaffold (bottom). C) YAP nuclear localization of hASCs cultured on micropatterns (n=3). Quantification of the D) Cell morphology (spread area and aspect ratio) on micropatterns (n=3). E) YAP nuclear localization of hASCs cultured on 3D fibrous scaffold (n=3). Quantification of the F) Cell morphology (spread area and aspect ratio) on 3D fibrous scaffold (n=3). Data presented as mean ± SEM. for group comparison C, D, E, and F, two-way ANOVA with Sidak post hoc test, *P<0.05.

In examining the role of PIEZO1 modulation on cell and nucleus morphology, we noted a decrease in hASCs spread area (approximately 30–40 %) following PIEZO1 activation in both 2D and 3D cultures (P<0.05) (Fig. 4D, F). This reduction in cell spread area, however, was not accompanied by a significant change in the extent of cell elongation as cell aspect ratio changes with PIEZO1 activation were not significant, except for 3D-aligned cells that we observed cell elongation following PIEZO1 activation (Fig. 4D, F). Furthermore, we observed a statistically significant decrease in nuclear size in cells treated with Yoda1 and cultured on unpatterned, flat surfaces and 3D substrates (P<0.05) (Fig. S7A, B). Additionally, in cells grown on aligned substrates, elongation of nuclei was statistically significant both in 2D and 3D (P<0.05) (Fig. S7A, B).

3.4. PIEZO1 depletion leads to nuclear accumulation of YAP in hASCs

To validate our observation on the impact of PIEZO1 activation by Yoda1 on YAP localization, we performed PIEZO1 knock down in cells cultured on both 2D and 3D cell cultures, following the experimental conditions described in the previous section. We expected no significant alteration in YAP localization when PIEZO1 knocked down ASCs were exposed to Yoda1. Knockdown of PIEZO1 in hASCs cultured on micropatterns in both aligned and flat conditions resulted in an increased nuclear-to-cytoplasmic (N/C) ratio of YAP (P<0.05) (Fig. 5A, C). To discern if this effect was specifically attributable to a loss of PIEZO1 function, we exposed both non-targeting siRNA and PIEZO1 siRNA groups to Yoda1 and assessed the YAP N/C value as before. Contrary to our expectations, Yoda1 exposure led to a decreased nuclear-to-cytoplasmic YAP ratio in the PIEZO1 knockdown groups (P<0.05) (Fig. 5B, C). Therefore, we hypothesize that the residual PIEZO1 proteins, not eliminated post-siRNA transfection, may still respond to PIEZO1 activation by Yoda1.

Fig. 5. PIEZO1 knockdown modulates YAP activity and influences cell morphology in 2D.

Representative nuclear (blue), actin (green), and YAP (pink) staining of A) Non-targeting siRNA and PIEZO1 siRNA groups cultured on micropatterns. B) non-targeting siRNA+Yoda1(2.5 μM) and PIEZO1 siRNA+Yoda1 (2.5 μM) groups cultured on micropatterns. Quantification of the C) YAP nuclear localization of hASCs cultured on micropatterns (n=3). Quantification of the D) Cell morphology (spread area and aspect ratio) on micropatterns (n=3). Data presented as mean ± SEM. for group comparison C, and D, two-way ANOVA with Sidak post hoc test, *P<0.05.

Using our 3D model, PIEZO1 knockdown resulted in an elevated N/C value compared to the non-targeting control group (P<0.05) (Fig. 6A, B, C). In the presence of Yoda1, N/C value did not significantly change compared to the PIEZO1 siRNA groups. Moreover, no significant difference in YAP localization was observed between aligned and random fiber orientations, suggesting that YAP activity in this context is independent of substrate architecture (Fig. 6C).

Fig. 6. PIEZO1 knockdown modulates YAP activity and influences cell morphology in 3D.

Representative nuclear (blue), actin (green), and YAP (pink) staining of A) Non-targeting siRNA and PIEZO1 siRNA cultured on micropatterns. B) non-targeting siRNA+Yoda1 (2.5 μM) and PIEZO1 siRNA+Yoda1 (2.5 μM) groups cultured on micropatterns. Quantification of the C) YAP nuclear localization of hASCs cultured on micropatterns (n=3). Quantification of the D) Cell morphology (spread area and aspect ratio) on micropatterns (n=3). C, and D, two-way ANOVA with Sidak post hoc test, *P<0.05.

Our morphological analysis indicated that cultured ASCs only on unpatterned, flat surfaces reduced in size (spread area) following PIEZO1 knockdown was statistically significant (P<0.05) (Fig. 5D). Additionally, hASCs in PIEZO1 knockdown groups demonstrated reduced elongation compared to their non-targeting siRNA group when cultured on aligned micropatterns (Fig. 5D). The nuclei of hASCs cultured on aligned patterns exhibited reduced spreading when PIEZO1 was knocked down (P<0.05) (Fig. S8A). These results indicate the PIEZO1 contributes to cell spreading and it is involved in cells aligning their morphology in response to matrix alignment.

Expanding our morphological investigation to encompass a 3D model yielded findings that were not paralleled with those observed in the 2D model in the absence of PIEZO1. In the 3D environment, hASCs demonstrated increased spreading on both aligned and random 3D fibrous scaffold (P<0.05) (Fig. 6D). Furthermore, an elongation of the nucleus was observed in hASCs cultured on substrates with aligned fibers (P<0.05) (Fig. S8B). The specific morphological responses mediated by PIEZO1 appear to be context-dependent, varying between 2D and 3D matrix environments.

3.5. Differential regulation of collagen genes by PIEZO1 activation and YAP/TAZ knockdown in 2D vs 3D models.

To elucidate the regulatory role of YAP and TAZ in the context of collagen modulation via PIEZO1, we employed siRNA-mediated gene silencing of the YAP and TAZ genes. The efficacy of knockdown at both the mRNA and protein levels was evaluated through RT-qPCR (P<0.05) and Western blot analyses. The results demonstrate a successful reduction in YAP and TAZ expression (Fig. S9A, B).

To investigate the impact of YAP and TAZ depletion, individually and in combination, on collagen gene expression during PIEZO1 activation, we used RT-PCR to compare the collagen gene expression patterns of cells treated with or without Yoda1 exposure, in both 2D and 3D substrate contexts (Fig. S9C, D). As we previously observed, PIEZO1 activation with Yoda1 did not have significant effect on collagen gene expressions (Fig. 2E), we also found here that activation of PIEZO1 in the context of YAP or TAZ knockdown, as well as the concurrent knockdown of both YAP and TAZ, does not significantly alter collagen expression patterns on micropatterned substrates (P<0.05) (Fig. S9C). COL1A1 gene expression was significantly downregulated in both aligned and random substrate cultures in 3D culture when hASCs with YAP/TAZ knockdown and YAP knockdown cells were exposed to Yoda1 (P<0.05) (Fig. S9D).

In the 3D model, a reduction in COL3A1 expression was observed in both aligned and random substrate cultures within the TAZ and combined YAP/TAZ siRNA groups (P<0.05) (Fig. S9D). The effect of PIEZO1 activation on collagen gene expressions in the absence of YAP and/or TAZ was more pronounced in a 3D model as compared to a 2D model.

4. Discussion

In this study, we identified a role for PIEZO1 and the mechano-responsive transcriptional activators YAP/TAZ in sensing substrate architecture and regulating collagen synthesis in human ASCs (Fig. 7). Using both 2D and 3D nanostructured in-vitro models, we explored the interaction between PIEZO1 and the YAP/TAZ signaling pathway in this process.

Fig. 7. Regulation of collagen expression in hASCs by PIEZO1 modulation via YAP/TAZ dependent mechanotransduction.

Schematic illustrating Regulation of collagen expression in ASCs by PIEZO1-mediated mechanotransduction and PIEZO1 interaction with YAP.

Our findings demonstrated that PIEZO1 activity can influence collagen expression and morphological responses of hASCs in a manner dependent on substrate architecture. Interestingly, PIEZO1 knockdown led to increased collagen synthesis, suggesting a regulatory role in limiting collagen synthesis. This PIEZO1-mediated collagen modulation in mRNA level was sensitive to the substrate architecture, with a significant effect observed when hASCs were cultured on aligned substrates in both 2D and 3D contexts. Our 3D model revealed reduced collagen I protein synthesis upon PIEZO1 activation. These findings contrast with a recent study in adipocytes, which proposed PIEZO1 inhibition as a potential strategy for mitigating fibrosis in skin [31]. Instead, our data showing that PIEZO1 inhibition or knock-down increased collagen synthesis in both 2D and 3D microenvironments using human adipose-derived stem cells suggest PIEZO1 as a potential target for promoting fibrosis resolution in adipose tissue. The discrepancy may be explained by cell-type specific functions, varying microenvironmental cues, and/or different mechanical forces experienced by different tissues. The effect of PIEZO1 effect on hASCs, which contribute to adipose tissue homeostasis, may differ from its role in chronic disease states or injury-induced fibrotic processes in other tissues. Differences in experimental conditions could also contribute to these contrasting effects on collagen regulation. We also observed that PIEZO1 knockdown had a greater effect on collagen expression in hASCs grown on aligned as compared to those grown on random substrates. This finding is consistent with previous studies showing that substrate alignment affects various cellular processes [56–61]. For example, Kishore et al. [62] found that aligned collagen fibers more effectively promote tendon-specific markers, including collagen III.

Furthermore, PIEZO1 activation induced significant morphological changes by reducing cell spreading area and increasing cell elongation in hASCs, especially in cells cultured on aligned substrates. This is in line with existing literature highlighting the role of PIEZO1 in cellular morphology changes [55, 63–68]. Holt et al. [63] explored PIEZO1 localization in migrating keratinocytes during wound healing. Their results indicate that PIEZO1 activity regulates cell shape by promoting cell polarization in keratinocytes, potentially by modulating the subcellular localization of PIEZO1 channels in migrating cells. In another study on HEK cells cultured on stripe patterns, Jetta et al [55] showed that PIEZO1 regulates cell spreading. PIEZO1-expressing HEK cells exhibited elongated morphology, which was abolished by PIEZO1 knockout as well as inhibition with GsMTx4 [55]. These results are particularly relevant to understanding cell responses to their mechanical environment, which is crucial in tissue engineering.

In this study, we examined the response to Yoda1 exposure to 1 hour to study early mechanotransduction responses. The observed morphological changes represent immediate cellular reactions to mechanical stimuli rather than long-term phenotypic changes. This approach helps understand how cells initially process mechanical signals in their environment, particularly relevant for early mechanotransduction in fibrotic processes. We often observed changes in cell aspect ratio without corresponding changes in cell spreading area. This observation highlights the complex nature of cellular responses to mechanical and chemical stimuli, emphasizing the importance of considering multiple morphological parameters in cell mechanobiology studies. This differential response might be attributed to distinct regulatory mechanisms for the two characteristics. Aspect ratio is primarily influenced by cytoskeletal organization, particularly actin stress fiber alignment, while cell spreading depends more on focal adhesion formation and strength. PIEZO1 modulation might preferentially activate pathways affecting cytoskeletal organization over those regulating cell spreading. While the Yoda1-induced nuclear elongation is indeed subtle, it is consistent across 2D and 3D dimensions, suggesting a genuine biological response. PIEZO1 activation may influence nuclear elongation through multiple mechanisms, including cytoskeletal reorganization, enhanced nuclear-cytoskeletal coupling via LINC complex proteins, Ca²⁺-mediated signaling, and amplification of existing mechanical cues from aligned substrates. The Hippo pathway is a critical regulator of organ size control and stem cell renewal [69, 70], suggesting that PIEZO1 could influence cellular processes related to tissue homeostasis and regeneration through its effects on the Hippo/YAP signaling cascade. We found that PIEZO1 activation induces YAP accumulation in the cytoplasm, resulting in reduced YAP activity. Multiple studies have shown interactions between these pathways and implicated them in cellular processes in other tissues/cells [41, 71–75]. Xiong et al. [71] demonstrated that PIEZO1 activation promotes ovarian cancer metastasis via the Hippo/YAP signaling axis. The absence of PIEZO1 causes YAP to be excluded from the nucleus in human neural stem cells [41] and during zebrafish outflow tract valve development [73], implying that PIEZO1 functions upstream of YAP. PIEZO1 and PIEZO2 are upstream regulators of signal transduction pathways that activate YAP for stem cell differentiation [74] and in regulating self-renewal of vocal fold mucosal epithelia [75]. However, YAP signaling induces PIEZO1 to promote oral squamous cell carcinoma cell proliferation [72], placing PIEZO downstream of YAP. Interestingly, knocking down PIEZO1 did not influence the transcriptional targets of YAP or YAP/TAZ nuclear localization in that study. Consequently, the authors propose that while the interplay between YAP signaling and PIEZO1 may vary depending on the cellular environment, these molecules could serve as a shared mechanotransduction mechanism, bridging organ development in stem cells [72].

We found that YAP and TAZ do not significantly influence collagen gene expression when PIEZO1 is activated on micropatterned substrates. This indicates independence between the mechanotransduction pathway mediated by PIEZO1 and the signaling mediated by YAP/TAZ in regulation of collagens. These results contrast with some previous research on the interplay between YAP/TAZ and PIEZO1 pathways in controlling collagen expression in other cell types. For example, Wang et al. [76] studied osteoblast cells response to mechanical loads and mechanistically investigated expression of collagens and showed that PIEZO1 regulates the expression of type II and IX collagens through YAP. PIEZO1 activation triggers Ca²⁺ influx and YAP translocation to the nucleus of heart valve interstitial cells, further supporting the role of YAP in PIEZO1-mediated collagen regulation [66]. This discrepancy bears further study.

Additional molecular pathways need to be investigated to gain a better understanding of the downstream mechanisms controlling collagen regulation. The interplay between PIEZO1 activation and YAP signaling is indeed complex and involves multiple pathways. PIEZO1- mediated Ca²⁺ influx significantly influences YAP signaling cascades. Fu et al [10] identified PIEZO1 as a key sensor of matrix stiffness, activating YAP-mediated extracellular matrix (ECM) accumulation through the p38 MAPK signaling pathway. Niu et al [77] demonstrated that PIEZO1 activation by mechanical tractions leads to YAP signaling activation via focal adhesion kinase (FAK) in cardiac fibroblasts, illustrating another mechanism by which mechanical cues can modulate YAP localization and activity through PIEZO1-mediated calcium signaling. Building on previous findings [78], we propose a mechanical signaling pathway where the absence or inhibition of PIEZO1 results in elevated membrane tension. This increased tension potentially induces higher actomyosin contractility, leading to stretching and flattening of the nucleus. Such nuclear deformation may favor the import rather than export of YAP to the nucleus, resulting in enhanced YAP activity [78], and subsequent upregulation of fibrotic gene expression.

We acknowledge a limitation in not directly assessing PIEZO1’s potential impact on stem cell differentiation. While our short-term experiments and specific changes suggest mechanosensing rather than differentiation, the interplay between PIEZO1 activity, stemness, and differentiation in hASCs remains an important area for future investigation. Such studies could provide valuable insights into PIEZO1’s broader role in stem cell biology and tissue engineering.

In future studies, exploring the interplay between PIEZO1 and other mechanotransduction pathways, such as integrin signaling, will be crucial in understanding how cells respond to substrate topography and mechanical cues to regulate collagens in disease context. Previous research has highlighted the crucial role of the interaction between PIEZO1 and integrins in mechanotransduction, influencing various physiological and pathological processes [79–83]. Cross-talk between PIEZO1 and integrins creates a feedforward loop that enhances glioma cell aggressiveness. This cross-talk contributed to increased tumor tissue stiffness, subsequently promoting glioma invasion and proliferation [82]. Furthermore, recent studies [83] suggests that the loss of PIEZO1’s interaction with integrins in cancer cells maybe a contributing factor to the uncontrolled growth and altered mechanical properties characteristic of tumors. Although we did not investigate PIEZO1 and integrin interactions in the present study, future research should explore this promising avenue to better understand their role in fibrosis progression. Future studies may wish to investigate potential crosstalk and feedback mechanisms between PIEZO1 and YAP/TAZ signaling pathways across various cellular contexts and developmental stages. Previous research indicates a bidirectional relationship between PIEZO1 and YAP/TAZ signaling, emphasizing the importance of gaining deeper insights into their interplay. A deeper understanding of these pathways could provide insights into new therapeutic targets for diseases involving dysregulated mechanotransduction, such as fibrosis. It is worth noting that our PCL mat system, while providing 3D topography, differs from fully embedded 3D models like hydrogels or thick scaffolds. Cells in our system experience surface-based 3D cues rather than complete 3D encapsulation, which may influence cellular behavior and mechanosensing. Future comparisons with fully embedded models could offer insights into how dimensionality affects PIEZO1-mediated processes in hASCs. Additionally, while our combined approach of gene expression analysis, picrosirius staining, and immunostaining provides insights into different stages of collagen synthesis, these methods don’t directly quantify extracellular collagen deposition. Hydroxyproline assays or other quantitative methods may be used to directly measure extracellular collagen accumulation in future studies.

Data Availability Statement

The raw/processed data required to reproduce these findings are available upon request.