Aging‐associated decline in vascular smooth muscle cell mechanosensation is mediated by Piezo1 channel

Abstract Aging of the vasculature is associated with detrimental changes in vascular smooth muscle cell (VSMC) mechanosensitivity to extrinsic forces in their surrounding microenvironment. However, how chronological aging alters VSMCs' ability to sense and adapt to mechanical perturbations remains unexplored. Here, we show defective VSMC mechanosensation in aging measured with ultrasound tweezers‐based micromechanical system, force instantaneous frequency spectrum, and transcriptome analyses. The study reveals that aged VSMCs adapt to a relatively inert mechanobiological state with altered actin cytoskeletal integrity, resulting in an impairment in their mechanosensitivity and dynamic mechanoresponse to mechanical perturbations. The aging‐associated decline in mechanosensation behaviors is mediated by hyperactivity of Piezo1‐dependent calcium signaling. Inhibition of Piezo1 alleviates vascular aging and partially restores the loss in dynamic contractile properties in aged cells. Altogether, our study reveals the signaling pathway underlying aging‐associated aberrant mechanosensation in VSMC and identifies Piezo1 as a potential therapeutic mechanobiological target to alleviate vascular aging.

Silanization of silicon molds was carried out with tridecafluoro-1,1,2,2, -tetrahydrooctyl)-1trichlorosilane (Sigma-Aldrich) overnight in vacuum.Negative PDMS molds were made from the silicon master mold by adding Sylgard 184 silicone elastomer base and curing agent (Dow Corning) mixture by 10: 1 mass ratio to the silicon master mold and baking at 110°C for 30 mins.
After curing, negative PDMS molds were peeled off from the silicon mold and silanizied in vacuum overnight.PDMS micropillar arrays were generated by casting a layer of Sylgard 184 silicone elastomer base and curing agent mixture by 10: 1 mass ratio on the surface of the silanized negative molds, and the negative molds were then covered with oxygen plasma (350 W, PlasmaEtch) treated glass coverslips (22mm x 22mm, Electron Microscopy).After curing in an oven heated at 110 °C for 48 hours, the glass coverslips containing PDMS micropillar arrays were peeled off from the negative molds and cut into small pieces.Fabricated PDMS micropillar array substrates were immersed in 100% ethanol and sonicated to restore collapsed micropillars, then dried with a critical point dryer (Samdri®-PVT-3D), then mounted on a 60 mm petri dish with a 15 mm hole in the center.Finally, PDMS micropillar arrays were functionalized using microcontact printing with fibronectin (50 μg/ml; Sigma-Aldrich) and Alexa-Fluor 647conjugated fibrinogen (25 μg/ml; Life Technologies).Schematic illustration of micropillar fabrication and functionalization is presented in Fig. S2.

Microbubble attachment to VSMCs
Biotinylated VesselVue microbubbles (Sonovol) of diameter between 4-5 µm were used in this study.First, the microbubble solution was mixed with streptavidin (10 mg/ml, ThermoFisher) at a volume ratio of 20:1 for 30 min and kept at room temperature to form streptavidin-conjugated microbubbles via streptavidin-biotin binding.The streptavidin-conjugated microbubbles were washed twice with phosphate buffered saline (PBS, Gibco) to remove the unbound streptavidin.
Next, biotinylated RGD peptides (2 mg/ml, Vividtide) were added to the streptavidin-coated microbubble solution in the ratio of 2:21 for 20 min at room temperature, then 1.2 μL of the microbubble solution was collected from the top layer and mixed with 48.8 μL of cell culture medium.Cell culture medium was aspirated from the petri dish containing VSMCs on micropillars.The prepared microbubble solution was added on top of the micropillar arrays with adherent cells.The petri dish was then flipped for 10 min to allow microbubbles to bind to the adherent cells via floatation.Lastly, the petri dish was flipped back and gently washed to remove any unbound microbubbles.VSMCs with single microbubbles attached were selected for experimentation with ultrasound tweezers.

Image recognition-based quantification of partial actin deviation
The fluorescent images of F-actin were first skeletonized using a Canny edge detection (MATLAB 2020b).The actin CSK was detected by applying Hough transform (a line extraction technique for digital image processing (Ballard, 1981)) .Next, the Hough-Peaks associated with the Hough transform matrix were found.Finally, the lines associated with the Hough-Peaks were determined.
The angle associated with the lines was used to calculate the orientation of the actin fiber associated with the detected line.The partial actin deviation with the detected lines were then calculated based on formula established in a previous study (Liu, Mollaeian, & Ren, 2018).Detailed pipeline is presented in Fig. S8.Schematic illustration of the PDMS micropillar array fabrication procedure which includes (i) silicon master mold manufacturing using photolithography, (ii) creating PDMS negative mold, and (iii-iv) making PDMS micropillar array on a glass substrate from the negative molds though a soft lithography fabrication process.B) Schematic illustration of the cell seeding process, including (i) coating the PDMS stamp with protein solution containing fibronectin and fluorescence-conjugated fibrinogen adhesive proteins, (ii-iii) transferring the adhesive proteins onto the tips of the micropillars using a microcontact printing technique, and (iv) loading of cells onto the functionalized micropillar array substrate.

Supporting Tables.
Table S1.Top 20 differentially expressed genes in each cell clusters.
Figure S1.Schematic illustration of the ultrasound tweezers-based micromechanical system for single-cell mechanosensation study.A) Experimental apparatus including a function generator, a power amplifier, and an ultrasound transducer that applies ultrasound pulses to generate acoustic radiation force on the microbubble attaching on cell membrane causing its displacement.The cell is loaded on a PDMS micropillar array substrate and the cellular force response is observed with an inverted microscope.B) Schematic illustrations showing attachment of microbubble onto cell membrane via RGD-integrin binding (top panel), and cell traction force measurement based on the deflections of micropillars underneath the cell (bottom panel).Figure is created with BioRender.com.

Figure S2 :
Figure S2: Micropillar array fabrication, microcontact printing, and cell seeding.A)Schematic illustration of the PDMS micropillar array fabrication procedure which includes (i) silicon master mold manufacturing using photolithography, (ii) creating PDMS negative mold, and (iii-iv) making PDMS micropillar array on a glass substrate from the negative molds though a soft lithography fabrication process.B) Schematic illustration of the cell seeding process, including (i) coating the PDMS stamp with protein solution containing fibronectin and fluorescence-conjugated fibrinogen adhesive proteins, (ii-iii) transferring the adhesive proteins onto the tips of the micropillars using a microcontact printing technique, and (iv) loading of cells onto the functionalized micropillar array substrate.

Figure S3 :
Figure S3: Traction force measurement using the PDMS micropillar array.A) The pipeline stages showing calculation of traction force using Cellogram software (open source).The software first detects the coordinate of single micropillar and presents them as a circle dot (i).Next, a mesh is generated to connect all the detected micropillars (ii).Energy minimization is performed to detect the reference position of the micropillars (iii).Displacement vector is obtained by subtracting the displaced coordinates with the reference coordinates.Finally, an elastic beam theory based on displacement and spring constant of the micropillars is applied to calculate the traction forces exerted by the cells.B) Example of the workflow showing traction force mapping from the fluorescent images (red channel) of the micropillars.The v panel shows a representative traction force heatmap (left) and a traction force vector map (right) of VSMC.

Figure S4 :
Figure S4: Microbubble attachment to cell.A) Single cells are seeded on a fibronectin-coated micropillar array substrate.B) Cell culture media is removed, and microbubble solution was added to the cells.C) The culture dish is flipped for 20 minutes to facilitate microbubble attachment via floatation.D) Unbounded bubbles are washed away with cell culture media twice, and cells bound with microbubbles are then suspended in fresh culture media for following ultrasound stimulation experimentation.

Figure S5 :
Figure S5: Force dynamic response of VSMCs without ultrasound stimulation and treated with Blebbistatin with inhibited myosin-mediated contractility.A) Temporal evolution of normalized traction force of VSMCs on PDMS micropillar array without ultrasound stimulation.Gray lines represent temporal force response of individual cells, and black line represents average force response of all measured cells (n=6).B) Dynamic force response of Blebbistatin treated young cells (n=5).C) Instantaneous amplitude and D) Instantaneous frequency of young versus young + Blebbistatin cells (n=5).In C, D, data are presented as mean values ± SEM, P-values were calculated using Student's t-test.* denotes p <0.05.

Figure S6 :
Figure S6: Instantaneous frequency spectrum analysis.A) Pipeline of Hilbert-Huang transform methodology, including tracking individual micropillar's velocity, performing empirical mode decomposition (EMD) of the velocity signal, and then applying Hilbert transform to the decomposed signal to get the instantaneous spectrum.B) Representative images showing displacement of individual pillars (white dots) and velocity (green arrows).C) Application of EMD on the pillar velocity.(i) Pillar velocity variation over time serving as the input signal.(ii)

Figure S7 .
Figure S7.Pathway analysis of differentially expressed genes in young versus old VSMCs.A) GO enrichment analysis demonstrating alterations in VSMC biological processes with aging, based on the Gene Ontology Resource knowledgebase.The enriched pathways were categorized into aging-associated pathways, vascular functions, and mechanosensitive functions.B) KEGG pathway analysis performing on differentially expressed genes in young and old VSMCs (n=3).

Figure S8 .
Figure S8.Subcellular F-actin analysis.A) Outline of image recognition-based actin CSK quantification method to calculate partial actin deviation.B) An example demonstrating subcellular actin analysis of young and old cells, including raw F-actin fluorescent image (i), cellular selection using boxes of 128x128 pixel (ii), F-actin fiber detection (iii), and partial actin deviation calculation.C) Polar histogram quantifying the subcellular distribution of the PADs in peripheral, periproximal and medial section of young and old cells as indicated.

Figure S9 .
Figure S9.Comparison of VSMC morphology and actin arrangement on micropillar and flat substrates.A) Representative immunofluorescent staining images and B) quantification of cell morphology features (cell area and circularity), F-actin fluorescent intensity, and F-actin anisotropy of young and old VSMCs after 24-hour culture on micropillar and flat glass substrates (n=40).In B, data are presented as mean values ± SEM, P-values were calculated using Student's t-test.* denotes p <0.05, ** p<0.005, **** p<0.00005, and ns denotes no significant difference.

Figure S10 .
Figure S10.Compassion of Piezo1 gene expression in young and old VSCMs isolated from 5 different mice.qRT-PCR analyses were performed on four independent cultures (Batch 1-4) from the same vendor (Creative Bioarrays) and one culture (Batch 5) from a different vendor (Cell Biologics) to confirm Piezo1 expression in young and old VSMCs (n=3).Data are presented as mean values ± SEM, P-values were calculated using Student's t-test.* denotes p<0.05, ** p<0.005, **** p<0.00005.

Figure S11 .
Figure S11.Single-cell RNA sequencing mapping of young, old, AngII-induced premature, and Piezo1-reversed aging VSMCs.A) Heatmap representation of top 20 differentially expressed genes across four cell clusters.B) UMAP projection of each VSMC subset labeled by obligotagged antibodies.C) Dot plot showing average and percent expression of genes encoded for aging-associated markers and mechanosensitive channels in the four cell clusters.D) Feature plots showing expression of housekeeping genes across the VSMC subsets.E) UMAP projection highlighting the expression of VSMC markers in cluster of young cells.

Figure S12 .
Figure S12.Expression of aging-associated markers and vascular cells markers in young, old, AngII-induced premature, and Piezo1-reversed aging VSMCs.A) Violin plots for markers of Spp1, Cd44, and Piezo1 among the four clusters.B) Violin plots quantifying expression of cytoskeletal genes in young versus young+AngII cluster.

Figure S13 .
Figure S13.Inhibition of Piezo1 in VSMCs with Silencer Pre-designed Piezo1 siRNA.A) qRT-PCR analysis (n=4) of Piezo1, VSMC marker ACTA2 and TAGLN, and B) Immunofluorescent analysis (n=40 cells per group) of Piezo1 expression in old cells (control) and old cells transfected with Silencer Pre-designed Piezo1 siRNA (AM16708, Life Technologies).C) Time course measurement of Ca 2+ response in VSMCs treated as indicated in response to Yoda1 stimulation (n=15).D) Normalized global traction force dynamics.Peak force response and response time (t) are indicated in the figure.(n=5).All statistical analysis was performed by Student's t-test and data are presented as mean values ± SEM. **** denotes p<0.00005.

Figure S14 .
Figure S14.Schematic of Piezo1-dependent signaling pathway that mediates VSMC mechanosensation.The figure is created with BioRender.com.