SIRT4 Regulated by a Mechanosensor, PIEZO1 Shows a Protective Function to Suppress Ox-LDL Uptake in Endothelial Cells

Abstract

Background:

Endothelial cells (ECs) are key regulators of vascular function, adapting to mechanical forces, such as shear stress to maintain vascular homeostasis. Disruption of this adaptation, particularly in the regions of disturbed flow, contributes to endothelial dysfunction and the development of atherosclerosis later on.

Methods:

We prepared a custom-designed PDMS-based flow chamber to apply controlled shear stress (2 or 7 dynes/cm²) to human umbilical vein endothelial cells. ECs were cultured on gelatin-coated coverslips and exposed to different shear flows for up to 12 h. Cell alignment was confirmed by angle measurements using ImageJ. Gene expression of SIRT4, PIEZO1, NOTCH1, and LOX-1 was determined via qPCR, and protein levels were assessed by western blot. Specific gene knockdown was also conducted using siRNAs, targeting either PIEZO1 or SIRT4. Oxidized LDL uptake was evaluated using DiI-labeled Ox-LDL and quantified by fluorescence imaging. Immunofluorescence staining of ECs was performed to visualize VE-cadherin, F-actin, and nuclei. All quantitative data were subjected to statistical analysis.

Results:

We demonstrated that the mechanosensitive ion channel PIEZO1, regulates SIRT4 expression in response to shear stress. Under atheroprotective shear stress (7 dyne/cm²), PIEZO1-mediated upregulation of SIRT4 was observed, while atheroprone shear stress (2 dyne/cm²) led to reduced expression. Functional assays showed that SIRT4 protects endothelial cells from Ox-LDL uptake, a key factor in atherosclerosis. SIRT4 silencing increased Ox-LDL accumulation even under protective flow. This effect, and its link to LOX-1, was dependent on PIEZO1 signaling.

Conclusion:

Current findings suggest that the PIEZO1-SIRT4 axis may modulate endothelial responses to shear stress, offering a protective mechanism against Ox-LDL-induced dysfunction and pathology. Our study underscores the potential of SIRT4 as a therapeutic target to mitigate vascular disorders associated with oxidative stress and disturbed blood flow.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13770-025-00733-w.

Introduction

Endothelial cells (ECs), which make up the thin layer lining the inside of blood vessels, are vital to the health of human vascular system [1]. These cells are essential for controlling vascular functions, such as blood flow [2], inflammation [3], and barrier function that separates blood from the neighboring tissues [4]. Their cellular responses to mechanical forces, especially shear stress, which is a frictional force caused by continuous blood flow [5] pose significant implications on vascular homeostasis in regard to the capability to sense and react to varying mechanical cues [6]. In fact, shear stress has significant impact on the behavior of ECs and also contributes to the pathological development of vessels [7]. For instance, the regions of low shear stress or disrupted blood flow are more likely to develop cardiovascular diseases [8], such as plaque formation inside arteries and later atherosclerosis development [9]. In such regions, endothelial dysfunction is often accompanied by an accumulation of oxidized low-density lipoprotein (Ox-LDL), a critical factor in the initiation and progression of atherosclerosis [10]. Ox-LDL triggers pro-inflammatory responses [11], increases oxidative stress [12], and impairs endothelial function. Given the role of oxidative stress in endothelial dysfunction, understanding of key molecular regulators, such as PIEZO1 and SIRT4 in response to shear stresses or Ox-LDL could provide valuable insights into a new protective mechanism against atherosclerosis. PIEZO1 is one of the crucial elements in ECs deeply associated with the cellular sensing and reactions to external mechanical forces [13, 14].

As a mechanosensitive ion channel, PIEZO1 [15] passes calcium ions into the cell in response to the blood flow-driven shear stress [16]. Numerous cellular reactions are triggered by this calcium signaling, such as modifications in vascular tone [17], cytoskeleton reorganizations [18], and gene expression. However, the role of PIEZO1 in mechanotransduction-how cells convert mechanical signals into biological output-remains incompletely understood and a maladaptive signaling in disturbed flow conditions also warrants further investigation. On the other hand, the sirtuin family of proteins, which includes SIRT4, is well-known for its role in mitochondrial and metabolic regulation [19, 20]. Cellular stress responses, such as how cells respond to inflammation and energy depletion, have been associated with SIRT4 [21]. Furthermore, SIRT4 is implicated in modulating oxidative stress, lipid metabolism, and mitochondrial homeostasis, while highlighting its role in maintaining cellular energy balance [22]. Recent studies have suggested that SIRT4 exerts protective role in various pathological conditions, including metabolic disorders and cardiovascular diseases [23, 24], by regulating key signaling pathways in stress adaptation. However, its role in ECs has not been fully explored, especially in relation to the impact of mechanical cues on SIRT4. A new mechanism may disclose whether SIRT4 has a protective effect on ECs in response to various shear stresses, given the role of oxidative stress in atherosclerosis development.

Here, relationship between PIEZO1 and SIRT4 in ECs is examined in this work, with particular attention to how shear stress affects SIRT4 expression via PIEZO1-mediated signaling. We hypothesize that in addition to the role in detecting mechanical forces and converting them into biological cues, PIEZO1 may function as a crucial regulator of SIRT4 in response to shear stress in ECs. Using a custom-made flow chamber that simulates physiologically relevant shear stress conditions, we seek to find evidences about whether SIRT4 protects against Ox-LDL uptake in ECs and how PIEZO1 affects SIRT4 expression. Our study may offer a fresh perspective of PIEZO1-SIRT4 axis regarding their role in maintaining endothelial function under shear stresses.

Materials and methods

Cell culture

Human umbilical vein endothelial cells (HUVECs, hereafter ECs) were purchased from Lonza (Basel, Switzerland). Cells were cultured in Endothelial Cell Growth Medium 2 (EGM-2; PromoCell GmbH, Germany) with the supplement mix, C-39216, 100 µg/ml streptomycin and penicillin combination (Gibco, Thermo Fisher Scientific, USA). HUVECs were grown on the tissue culture plastic (TCP) with the seeding density of 75,000 cells/cm² cultivated until full confluence in EGM2 medium. For the experiments, ECs were the seeded on top of gelatin (0.5% solution)-coated coverslips. Cell culture was performed in the incubator under the condition of 5% CO₂ at 37 °C. ECs with cell passage 3–7 were used throughout the whole experiments.

Flow chamber design and shear stress application

The master mold for the flow chamber was designed using AutoCAD software and manufactured from stainless steel. Once the master mold was ready, we prepared polydimethylsiloxane (PDMS) (SYLGARD™ 184, Dow Corning, USA) at the ratio of 1:10 (silicone elastomer base: curing agent) in accordance with the manufacture’s protocol. The PDMS solution was then degasified in vacuum chamber for 15 min and cured for 15 min at 150 °C. After then, we assembled both parts (top and bottom) into a complete set of flow chamber as shown in Fig. 1A, B. The inner flow chamber parameters is as follows: The distance (h) between the lid and the base was 1 mm and the width was 18 mm to fit the gelatin-coated coverslips seeded with ECs. The chamber design included 3 parallel flow channels where they could accommodate 9 samples in total at a time. Shear flow was applied for up to 12 h. The flow rate was determined by measuring the volume of liquid delivered by the peristaltic pump over a one-minute interval. Shear stress was calculated via the formula of $\tau =\frac{6\mu Q}{w{h}^2}$, where τ-is shear stress [Pa], μ-viscosity of cell culture media [Pa s], Q-flow rate [m²/s], w and h, width and height of the chamber, respectively [25]. Our system can adjust shear stress levels, simulating a low (2 dynes/cm²), and high (7 dynes/cm²) shear stress, respectively. The peristaltic pump allows a precise control of flow rate, enabling the adjustment of shear stress levels if needed. Once the shear flow was stopped at predetermined time point, we removed the chamber lid, collected samples carefully, and proceeded with further experiments.

Fig. 1.

Flow chamber setup. A Schematic illustration of our flow device setup, which is composed of top and bottom PDMS chambers and a peristaltic pump (not shown here); B Image of assembled flow chamber; C Schematic representation of shear flow-induced shear stress on the ECs; D Two different shear stresses (high, 7 dyne and low, 2 dyne) applied to the ECs in this study

Examination of cell alignment

To observe and quantify ECs alignment in response to the shear stress applied in the flow chamber, phase-contrast images were taken and analyzed using ImageJ (NIH). The images were first converted to grayscale, and background subtraction was then applied to minimize noise. Cell boundaries were drawn manually following the VE-cad signal to minimize discrepancies. The "Analyze Particles" function was used to identify and select individual cells, and "Fit Ellipse" option was applied to measure the orientation of the major axis of each cell relative to the flow direction. The extent of cell alignment under different shear stress conditions (2 and 7 dynes/cm²) was quantitatively determined by angle measurements.

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)

Gene expression of target genes, such as SIRT4, PIEZO1, NOTCH1, and LOX-1 was detected by RT-qPCR. RNeasy Mini Kit (QIAGEN, Germany) was employed to extract total RNA from the ECs with or without shear stress. Concentration and purity of the samples were determined using NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, USA). SuperScript VILO (Invitrogen, Thermo Fisher Scientific, USA) was used to synthesize cDNA according to the manufacturer’s protocol. Acquired cDNA was used as a template to amplify dsDNA using TB Green Premix Ex Taq (Takara Bio Inc., Japan). The reaction conditions were: 95 °C for 10 min, 40 cycles at 95 °C for 10 s, 57 °C for 30 s and 72 °C for 30 s. Target genes and primer sequences are listed in the Table 1.

Table 1.

List of primers and sequences

Target genes	Sequence (5′ → 3′)
VCAM1	GATACAACCGTCTTGGTCAGCCC (Forward)
	CGCATCCTTCAACTGGGCCTT (Reverse)
NOTCH1	GCAAGAACGCCGGGACAT (Forward)
	CGGCACTTGTACTCCGTCAG (Reverse)
GAPDH	CCATGGGGAAGGTGAAGGTC (Forward)
	TGATGACCCTTTTGGCTCCC (Reverse)
PIEZO1	TCCTCAACCACATGGTCACG (Forward)
	GCGATCTCGGTGAAGACGAT (Reverse)
SIRT4	TGCAATCAGACGGTCCCAC (Forward)
	CCTACGAAGTTTCTCGCCCA (Reverse)
LOX	CGTGACTGCTTCACTCTCTCA (Forward)
	ATCCAGTCTTGCGGACAAGG (Reverse)

Transfection of HUVECs with silencing RNA

Both siRNA PIEZO1 and siRNA SIRT4 were synthesized by BioNEER Corp. (Daejeon, Republic of Korea) using pre-setup configuration from the manufacturer. Before cell transfection using siRNAs, ECs remained under normal condition for 3 days. After then, cell culture medium was replaced with antibiotics-free media for 24 h before transfection procedure. The transfection was lasted for 5 to 6 h in 50 nM concentration using Lipofectamine RNAiMAX Reagent (Invitrogen, Thermo Fisher Scientific, USA) as a carrier in serum-free medium. The cells were then placed in fresh culture medium for 24 h to allow a complete transfection. As a negative control, AccuTarget™ Negative Control siRNA by BioNEER Corp. (Daejeon, Republic of Korea) was used throughout all the experiments.

DiI Ox-LDL treatment

To investigate Ox-LDL uptake by ECs with or without shear stress, the ECs were subjected to shear stress 12 h while the control group remained in static condition. Each experimental group consisted of three samples of HUVECs seeded on 18 mm gelatin-coated coverslips. After 12 h of shear stress exposure, those samples were transferred to fresh EGM-2 medium under static condition. The medium was supplemented with 10 μg/mL of DiI Ox-LDL (L34358; Invitrogen, USA), and the cells were incubated for an additional 12 h at 37 °C. Those samples were then washed with PBS 3 times before immunofluorescent staining. The fluorescence intensity of the DiI Ox-LDL inside the ECs was quantitatively analyzed using RGB profile plot plugin via ImageJ.

Western blot

Protein extraction out of the ECs was performed using RIPA Buffer (BYLABS, Republic of Korea) for cell lysis, supplemented with 1X protease inhibitor cocktail (ab271306; Abcam, UK). Protein concentration was measured using Pierce BCA assay kit (Thermo Fisher Scientific, USA). Total protein amount per loading was 20 μg for each sample. The protein samples were separated using 10–15% SDS-PAGE (4561023; BioRad, USA) under 100 V for 1 h in a constant voltage mode. After the running was over, the gels were transferred to polyvinylidene fluoride (PVDF) membrane under the conditions of 110 V for 80 min. Subsequently, the membrane was blocked with 5% skim milk for 1 h on orbital shake, followed by the incubation with following primary antibodies: anti-SIRT4 (ab124521; Abcam), anti-PIEZO1 (ab128245; Abcam, UK) and β-actin (C4; Life science, India), one at a time at 4 °C overnight. Next, membranes were washed 5 times in tris buffered saline with Tween (TBST) 1X and incubated with the HRP-conjugated secondary antibody anti-rabbit (6721; Abcam, UK) for 1 h at RT. After the membranes were washed again 5 times with TBST 1X, they were then subjected to the signal development via SuperSignal™ West Pico PLUS Chemiluminescent Substrate (34578; Thermo Fisher Scientific, USA) for 3 min. The protein band was detected using iBright CL1500 (Thermo Fisher Scientific, USA). The band intensity of target proteins was normalized to the intensity of β-actin. The results were quantitatively analyzed using ImageJ software (ImageJ 1.53q) and plotted in Graphpad Prism software (Prism 9).

Immunofluorescent staining

Samples were fixed with 4% paraformaldehyde for 15 min, washed with PBS 2 times and permeabilized with 0.2% Triton X-100 for 10 min and washed with PBS 3 times more. After then, those samples were blocked with 3% bovine serum albumin (BSA) in PBS for 1 h. Subsequently cells were incubated with primary antibodies of VE-cad (ab96587, Abcam, UK) diluted in 1% BSA at the ratio 1:500 (antibody to 1% BSA) at RT for 1 h. Subsequently, the samples were subjected to PBS washing 3 times and stained with secondary antibodies in dilution of 1:500 (goat anti-rabbit Alexa Fluor 488 (Dylight 488)), along with 4′,6-diamidino-2-phenylindole (DAPI; R37605, Invitrogen, USA) and rhodamine-phalloidin (R415, Thermo Fisher Scientific, USA) staining for 1 h at RT.

Statistical analysis

To determine statistical significance, the experimental data were analyzed using one-way ANOVA for multiple comparisons and t-test for comparing two means via Graphpad Prism software (Prism 9). For all the experiments, each group were tested in triplicate unless different sample number was mentioned. Statistical significance is marked as *p < 0.05, ** p < 0.01, *** p < 0.001 or **** p < 0.0001.

Results

The PDMS chamber was structured to ensure uniform flow distribution, minimizing turbulence and thus delivering consistent mechanical stimulation to the ECs. Our flow chamber system can generate physiologically relevant shear stresses, specifically 2 and 7 dynes/cm², respectively (Fig. 1C). While 2 dynes/cm² represents a low shear stress condition, typically observed in disturbed or atheroprone regions of the vasculature [26], the shear stress of 7 dynes/cm² simulates a physiological flow, characteristic of arterial regions with a laminar flow [27] (Fig. 1D). This setup allows us to investigate the mechanotransduction pathways in ECs influenced by different shear stresses, particularly the role of PIEZO1 and SIRT4.

Once we successfully prepared shear flow chamber, we first validated ECs responses to different shear stress levels by assessing some parameters, such as cellular alignment and morphology. The fluorescent images of F-actin (red) and DAPI (blue) staining visualized the cytoskeletal organization and nuclear morphology of ECs, respectively (Fig. 2A). Under static condition, the normal cobblestone morphology and random orientation of the ECs were indicative of the absence of directional mechanical force, i.e., shear stress. Nuclear morphology remained largely unchanged because the cells were unable to align in response to low shear stress (2 dyne/cm²) and cytoskeletal reorganization was not observed. On the contrary, under higher shear stress (7 dyne/cm²), ECs exhibited a notable cell alignment along the flow direction, retaining an elongated, spindle-like morphology, which is typical of ECs under physiological shear stress. This result suggested a laminar flow adaptation, where cells would align with the shear force to enhance blood flow efficiency and cellular junction stability. Quantitative analysis also supported these morphological changes as appeared in the fluorescent images (Fig. 2B).

Fig. 2.

Effect of different shear stresses on ECs morphology. A Representative fluorescent images exhibit the morphological changes of ECs upon different levels of shear stress (static-0 dyne, atheroprone-2 dyne, and atheroprotective-7 dyne). Cells were stained with F-actin (red) to visualize cytoskeletal organization and with DAPI (blue) to assess nuclei morphology. The flow direction is indicated with the white arrow; B The graph shows the aspect ratio of ECs nuclei under different shear stress conditions. Aspect ratio was calculated from fluorescent images using ImageJ Fiji software; C Fluorescent images illustrate VE-cadherin localization in ECs to assess cell–cell junction integrity under different shear stress conditions; D A rose graph displays the distribution of ECs alignment angles under shear stresses. It demonstrates the predominant orientation of cells in response to higher shear stress, with angles measured relative to the direction of the flow. Data are presented as mean ± standard deviation. For statistics, one-way ANOVA test was used to determine the differences among the test groups. Statistically significant difference was marked as **** (p < 0.001). The scale bar is 50 μm (A) and 20 μm (C)

Quantitative measurements of nuclei morphology from the DAPI-stained images revealed a significant hike in the aspect ratio of cell nuclei under atheroprotective shear stress (7 dyne/cm²). The difference was statistically significant between 7 and 2 or static group (0). This elongation in nuclear shape is explained by a mechanoadaptive response, where ECs stretch and align themselves in the direction of blood flow, reflecting the cells’ ability to sense and respond to mechanical stress. This adaptation is critical for minimizing the frictional forces exerted by the flow, allowing cells to better withstand mechanical stress and maintain vascular health. To further investigate the impact of shear stress on the structural integrity of cell–cell junctions, the expression and localization of VE-cadherin, a key component of adherens junctions, were examined. VE-cadherin plays an essential role in maintaining endothelial barrier function and facilitating strong intercellular connections. Under 7 dyne/cm², VE-cadherin showed robust localization at the cell borders (Fig. 2C), supporting the formation of stable and cohesive junctions. These well-defined cell–cell junctions enhance the mechanical strength of the endothelial layer, providing a protective mechanism against mechanical disruptions and inflammatory insults. In contrast, those cells exposed to atheroprone shear stress (2 dyne/cm²) exhibited weaker VE-cadherin expression and irregular localization at the cell boundary. This disrupted VE-cadherin would contribute to the instability of endothelial junctions, making these regions more prone to inflammatory cell infiltration and endothelial dysfunction. It is notable that VE-cadherin expression was also present under static condition, which is natural in the culture of ECs. In fact, we considered static condition as a positive control. Our particular interest was to confirm the difference of cell–cell junction between 2 and 7 dyne condition.

In addition, a rose plot analysis was performed to better understand the cells’ alignment relative to the given flow direction (Fig. 2D). Under static condition or atheroprone shear stress, cells displayed a more randomized orientation, indicating the absence of directional mechanical cues. However, when exposed to the atheroprotective cue, the cells disclosed a pronounced alignment parallel to the flow direction. This alignment reflects a coordinated cytoskeletal rearrangement, enabling cells to adapt to the mechanical cue and thus reduce mechanical stress on their structures. Interestingly, when PIEZO1 was silenced and HUVECs were subjected to 7 dynes shear stress, the cells failed to align properly, indicating that PIEZO1 is essential for the proper cytoskeletal response to shear stress (Supplementary Fig. 1A). In fact, in the areas of high, laminar shear stress, such as straight arterial segments, such adaptation foster a stable, anti-inflammatory environment that protects against plaque formation. Conversely, in the arterial branches or bifurcations, which is the regions with low or disturbed shear stress, the lack of proper mechanoadaptive responses may contribute to increased vulnerability and ECs dysfunction [28].

To further investigate endothelial responses to mechanical and metabolic stress, we examined PIEZO1, SIRT4, and NOTCH1 [29, 30], as these genes play essential roles in vascular adaptation. PIEZO1, a mechanosensitive ion channel, directly detects shear stress and activates downstream signaling pathways. SIRT4, a mitochondrial protein, regulates metabolic and cellular stress responses [31, 32]. NOTCH1 is crucial for vascular homeostasis, mediating flow-dependent endothelial alignment and function [33]. Among them, PIEZO1 and SIRT4 are of particular interest in this study due to their roles in shear stress detection and stress adaptation, respectively. When their expression levels were determined under static and flow condition, respectively, both PIEZO1 and SIRT4 expressions were significantly upregulated upon the exposure to atheroprotective shear stress (7 dyne/cm²) (Fig. 3A). This result suggested that shear stress-induced PIEZO1 activation on the cell surface may affect SIRT4 expression and function in mitochondria, thereby maintaining cellular resilience under physiologic laminar flow. However, NOTCH1 expression was downregulated, suggesting an inverse relationship with SIRT4 and PIEZO1. Interestingly, the ECs exposed to atheroprone shear stress (2 dyne/cm²) disclosed an increased NOTCH1 and decreased SIRT4 and PIEZO1 expression. In atherosclerosis-prone areas, ECs are susceptible to oxidative stress and inflammation, which is consistent with the current results: significantly decreased PIEZO1 and SIRT4 expression under low or disturbed flow conditions. The situation may become worse by upregulated NOTCH1 expression, since NOTCH1 has been connected to endothelial inflammation nd the formation of atherosclerotic plaque in disturbed flow area [34].

Fig. 3.

Effect of PIEZO1 on Sirt4 under different shear stresses in ECs. A Gene expression patterns of PIEZO1, NOTCH1, and SIRT4 in ECs under varying shear flow conditions. Data are presented as fold changes relative to that of static condition, suggesting differential regulatory responses of these genes to flow dynamics; B Gene expression levels of SIRT4 in ECs under different shear flows and PIEZO1 silenced; (C) Gene expression levels of PIEZO1 in ECs under different shear flow conditions and SIRT4 silenced; D Protein level of PIEZO1 and SIRT4, respectively as assessed via western blot when HUVECs were exposed to different levels of shear stresses for 12 h; Based on the western blot data, quantitative analysis of PIEZO1 (E) and SIRT4 (F) as presented in fold changes relative to that of beta-actin. Data are presented as mean ± standard deviation. For statistics, Welch and Brown-Forsythe ANOVA test was used to assess the difference among the test groups. Statistically significant difference was marked as * (p < 0.05), ** (p < 0.01) or *** (p < 0.001)

To better understand the relationship between PIEZO1 and SIRT4, we conducted gene silencing experiments to determine whether PIEZO1 is somehow interconnected with SIRT4 under shear stress conditions. The results showed that silencing PIEZO1 did not alter SIRT4 expression at all in either atheroprotective or atheroprone condition. This hinted to a connection that SIRT4 might be regulated by PIEZO1 under shear stress. We demonstrated that silencing PIEZO1 prevents the upregulation of SIRT4 under shear stress, indicating that PIEZO1 is required for shear stress-induced SIRT4 expression and may function as an upstream of this mechanotransduction pathway (Fig. 3B). Our idea that PIEZO1 may function independently within the mechanotransductive network, possibly acting as an upstream regulator of SIRT4 under flow condition, was further supported by the fact that silencing SIRT4 had no effect on PIEZO1 expression under any shear condition (Fig. 3C). These gene expression patterns were validated by western blot, which showed that both SIRT4 and PIEZO1 protein levels were significantly lower under the atheroprone flow (2 dyne/cm²) but significantly higher under the atheroprotective flow (7 dyne/cm²) (Fig. 3D). Quantatitative analysis further supported the same trend: the difference was statistically signficiant between 2 and 7 dyne (Fig. 3E, F). Taken together, concurrent upregulation of SIRT4 under atheroprotective flow implies its involvement in the endothelial protective response, along with the elevated PIEZO1 level as a sensor of laminar flow condition.

Under normal blood flow condition, ECs would retain an atheroprotective phenotype, characterized by elongated and aligned morphology and decreased lipid accumulation. This protective effect helps prevent endothelial dysfunction and slows the progression of atherosclerosis. Since the accumulation of Ox-LDL in ECs is a crucial step in the early stages of atherosclerosis, SIRT4's role on this process is extremely important for vascular health [21]. When the effect of SIRT4 on Ox-LDL uptake was examined between static and shear stress condition, the ECs subjected to atheroprotective shear stress (7 dyne/cm²) exhibited a significantly decreased Ox-LDL uptake when compared to those under static condition (Fig. 4A). Quantitative analysis also supported this result, demonstrating statistically significant difference between them (Fig. 4B). Further study with or without SIRT4 silencing under static condition revealed that silenced SIRT4 led to noticeably higher Ox-LDL uptake of ECs than the control group (Fig. 4C). The difference was statistically significant (Fig. 4D). Our findings imply that SIRT4 is a crucial component in regulating endothelial lipid homeostasis and that it may have a protective effect by reducing Ox-LDL accumulation. As such, upregulated SIRT4 level under normal shear stress may contribute to lowering Ox-LDL uptake and stop lipid buildup in the vascular endothelium. Interestingly, SIRT4 silenced ECs also showed increased levels of VCAM-1 which may indicate a pathologic change into a inflammatory phenotype (Supplementary Figure 2).

Fig. 4.

Regulatory role of either atheroprotective flow or SIRT4 on the Ox-LDL uptake in ECs. A Fluorescent images show Ox-LDL uptake by ECs under either static or atheroprotective shear flow condition (7 dyne); B Quantitative analysis disclosed that the ECs subjected to a normal shear flow showed a significant decline of Ox-LDL uptake compared to those under static condition, highlighting the protective effect of atheroprotective flow against lipid accumulation; C Representative fluorescent images illustrate Ox-LDL uptake by ECs under either control or SIRT4 silenced group; D In quantitative analysis, SIRT4 silenced ECs exhibited significantly increased Ox-LDL uptake compared to that of control, an indicative of the regulatory role of SIRT4 in lipid homeostasis. Data are presented as mean ± standard deviation. For statistics, Student t-test was used to assess the difference between the two groups. Statistically significant difference was marked as **** (p < 0.0001). The scale bar is 20 μm

In order to examine the combined effects of SIRT4 and shear stress, we next evaluated Ox-LDL uptake under atheroprotective flow with targeted silencing of either SIRT4 or PIEZO1, respectively. First of all, silenced SIRT4 caused a significant increase of Ox-LDL uptake, even under atheroprotective flow when compared to that of control siRNA group with the flow (Fig. 5A). We also observed a significantly reduced Ox-LDL uptake in ECs under the atheroprotective shear stress, compared to that of static condition. Interestingly, however, combined effect of SIRT4 silencing and shear flow resulted in significantly incresed Ox-LDL uptake when compared to that of atheroprotective flow alone (Fig. 5B).

Fig. 5.

Influence of either SIRT4 or PIEZO1 silencing on endothelial Ox-LDL uptake under atheroprotective shear flow condition. A Fluorescent images display endothelial uptake of Ox-LDL, with or without SIRT4 silencing under atheroprotective shear flow (7 dyne). The control group is the ECs in the static condition; B Quantitative analysis discloses that under the shear flow condition, SIRT4 silenced ECs led to increased Ox-LDL uptake compared to that of non-silenced group; C In addition, when the effect of silenced PIEZO1 was examined, the amount of Ox-LDL uptake remained unchanged between control and atheroprotective group with silenced PIEZO1, indicating that silencing obstructed mechanosensitive abilities of ECs; D When gene expression levels of LOX-1 was examined under different shear stress levels, its expression was significantly higher at atheroprone shear flow (2 dyne); E Effect of either SIRT4 or PIEZO1 silencing on the LOX-1 expression under atheroprotective shear flow. The results suggested both genes are crucial in suppressing LOX-1 expression. Data are presented as mean ± standard deviation. For statistics, One-way ANOVA test was used to determine the difference among the test groups. Statistically significant difference was marked as * (p < 0.05) ** (p < 0.01), *** (p < 0.001) or **** (p < 0.0001). The scale bar is 20 μm

Our result stronlgy indicates that SIRT4 is a crucial component for lowering lipid accumulation in ECs, highlighting the protective role of SIRT4, because its absence negates the beneficial effects of atheroprotective flow on lipid uptake. Instead, silencing PIEZO1 had little notable impact on Ox-LDL uptake, even under atheroprotective flow (Fig. 5C). It is notable, however that while PIEZO1 itself has little effect in suppressing Ox-LDL uptake, PIEZO1-mediated cellular pathways do have impact. This suggests that since PIEZO1 is deeply involved in mechanosensitive responses of ECs, its function on Ox-LDL uptake may be indirect or dependent on the activity of SIRT4. Additionally, when we looked at the expression levels of LOX-1, an Ox-LDL receptor that mediated Ox-LDL uptake in ECs, LOX-1 was highly upregulated under the atheroprone flow (2 dyne) compared to that of static and 7 dyne condition (Fig. 5D). In addition, when either PIEZO1 or SIRT4 were silenced, LOX-1 expression was significantly upregulated under shear flow conditions (Fig. 5E). Although LOX-1 expression was elevated upon PIEZO1 knockdown, this did not result in increased Ox-LDL uptake (Fig. 5C), suggesting that additional regulatory mechanisms are involved in Ox-LDL internalization. It is also possible that PIEZO1 might influence other steps in the uptake process, such as cytoskeletal dynamics, vesicle trafficking, or receptor localization, independent of LOX-1 expression. This suggests that both PIEZO1 and SIRT4 might contribute to the shear-dependent suppression of LOX-1, which is associated with an atheroprotective endothelial phenotype. The loss of either genes finction may disrupt this regulatory mechanism, leading to increased LOX-1 expression despite the presence of shear flow. Notably, there was no significant difference in LOX-1 level between SIRT4 and PIEZO1 silencing groups, indicating that both genes might share the same pathway to mediate this effect. Elevated LOX-1 level in ECs for prolonged time leads to lipid accumulation, oxidative stress, and inflammation. This contributes to endothelial dysfunction by impairing nitric oxide availability and promoting a pro-atherogenic phenotype [35]. Persistent LOX-1 activation can potentially accelerate vascular inflammation and plaque formation.

Discussion

This work provides a new insight into the mechanosensitive control of ECs function, with an emphasis on PIEZO1's role in controlling SIRT4 expression and its implications for endothelial defence against Ox-LDL accumulation. Here, we report our findings, where mechanical forces affect specific gene expression and lipid uptake in ECs by simulating either physiological or pathological shear stress conditions using a custom-made flow chamber system. With possible ramifications for vascular health and diseases, our results propose a new regulatory connection between PIEZO1 and SIRT4 as a key mechanism in endothelial adaptation to shear stress. One of the primary findings is that under shear stress, SIRT4 expression depends on PIEZO1 activation. While SIRT4 has been linked to stress responses and metabolic regulation, little is known about how it functions in ECs, especially in relation to intracellular pathway of mechanotransduction. Current results show that PIEZO1 upregulation is closely interconnected to a concurrent rise of SIRT4 expression under atheroprotective shear stress (7 dyne/cm²). This implies that SIRT4 expression is stimulated by shear stress-induced PIEZO1 activation, suggesting a plausible signal transduction link between mechanical cues and the mitochondrial defence mechanism of ECs. Given the mechanosensitive nature of PIEZO1, it is likely that calcium influx acts as a key downstream mediator in its regulatory effect on SIRT4. Calcium-dependent pathways, such as those involving CaMKII or calcineurin-NFAT axis, may be contributing ones. Further investigation into these signaling cascades and potential transcriptional regulators will help clarify the molecular mechanisms underlying the PIEZO1-SIRT4 relationship.

It is interesting to note that silencing PIEZO1 did not increase SIRT4 expression levels, even under atheroprotective shear stress, suggesting that PIEZO1 acts as an upstream regulator of SIRT4. Our study also highlights SIRT4's role in controlling Ox-LDL uptake and maintaining lipid homeostasis. ECs with intact SIRT4 expression demonstrated significantly lower Ox-LDL uptake, particularly under atheroprotective shear stress, in comparison to static or atheroprone condition. On the other hand, regardless of the shear stress applied, silencing SIRT4 led to increased Ox-LDL uptake, empathizing the significance of SIRT4 in controlling lipid accumulation in ECs. This effect was amplified under atheroprotective flow, where SIRT4 expression further decreased Ox-LDL uptake. This suggests that SIRT4 plays a crucial role in endothelial defence against inflammation and lipid-induced oxidative stress upon the aid of laminar blood flow. By suggesting the cooperative action of PIEZO1 and SIRT4 to maintain vascular health, current findings contribute to expanding our knowledge regarding the implicatons of endothelial mechanotransduction that is pivotal for the vascular health.

The disturbed, non-laminar blood flow areas commonly seen at arterial branching points [36] or curvatures are closely associated with the early onset of atherosclerosis and endothelial dysfunction. Our custo-made flow chamber offers several technical advantages. For example, it allows precise control of different level of shear stresses, which is essential for simulating the mechanical cues ECs experience in the bloodstream in vivo. Such feature is crucial for studying how cells respond to particular mechanical cues under the pathological condition in vitro. Our framework to study the mechanoadaptive responses of ECs under different shear flow milieu, 2 and 7 dyne/cm² allows the replication of either pathological or protective haemodynamic environment [37]. While our study examined the endothelial response to acute shear stress, the observed PIEZO1-SIRT4 regulatory axis may also have relevance in the context of chronic vascular disease. Persistent alterations in shear stress, as seen in disturbed or low shear flow regions of the vasculature, are associated with endothelial dysfunction and metabolic reprogramming. Given the roles of mechanotransduction and mitochondrial regulation in these processes, our findings may provide insight into how prolonged mechanical stimuli contribute to vascular pathologies such as atherosclerosis. The base material, PDMS was chosen due to its biocompatibility and flexibility to fabricate [38]. Additionally, easy handling and user-friendly setup minimize technical complexity, ensuring consistency across the experiments while reducing time and effort required for operation. Furthermore, the system's capacity to process a large number of samples enhances experimental throughput and reproducibility, making it suitable for large-scale or comparative studies. Collectively, these features make our system a powerful tool for investigating the issue of mechanobiology underpinning vascular health and disease.

In this study, while harnessing our custom-made flow chamber design, we have provided an in vitro model that closely mimics physiological blood flow, allowing for a deeper understanding of the dynamic interactions between mechanical cues and cellular signaling. We demonstrated the critical role of shear stress in regulating ECs behavior through, particularly involving PIEZO1 and SIRT4. The SIRT4 expression was dependent on PIEZO1 activation, suggesting a probable link between them in regard to mechanotransduction pathway. Furthermore, our study emphasizes the protective role of SIRT4 in ECs against Ox-LDL uptake, offering a new insight into its implication as a therapeutic target in vascular diseases. Taken together, this work not only advances our knowledge of ECs behavior under atheroprotective or atheroprone shear stress but highlights PIEZO1-SIRT4 axis as a potential target in the initiation and treatment of atherosclerosis.

Supplementary Information

Below is the link to the electronic supplementary material.

Data availability

The data that supports the findings in this study are available from the corresponding authors upon reasonable request.

Declarations

Conflict of interest

The authors declare no conflicts of interest.

Ethical approval

There are no animal experiments carried out for this article.