The roles of mechanosensitive ion channels and associated downstream MAPK signaling pathways in PDLC mechanotransduction.

The present study aimed to investigate whether the cytoskeleton, the Piezo1 ion channel and the transient receptor potential cation channel subfamily V member 4 (TRPV4) ion channel are equally functional in the mechanotransduction of periodontal ligament cells (PDLCs) and to reveal the interplay of these mechanically sensitive ion channels (MSCs). Human PDLCs (hPDLCs) were pretreated with cytochalasin D (the inhibitor of actin polymerization), GsMTx4 (the antagonist of Piezo1) and GSK205 (the antagonist of TRPV4), and then subjected to periodic mechanical loading. The expression levels of macrophage colony stimulating factor (M‑CSF), receptor activator of NF‑κB ligand (RANKL) and cyclooxygenase‑2 (COX2) in hPDLCs were detected via western blotting. Osteoblast mineralization induction capacity of the hPDLCs was also studied and the mitogen‑activated protein kinase (MAPK) expression profile was determined via protein microarray. The expression of Piezo1 and TRPV4 in the PDLCs was significantly increased at 8 h after loading. These differences in expression were accompanied by increased expression of M‑CSF, RANKL and COX2. Compared with the control group, key PDLC biomarkers were suppressed after mechanical loading following treatment with the inhibitors of Piezo1 (GsMTx4) and TRPV4 (GSK205). The phosphorylated‑MAPK protein array showed differential biomarker profiles among all groups. The present study suggested that both MSCs and the cytoskeleton participated as mechanical sensors, and did so independently in hPDLC mechanotransduction. Furthermore, the Piezo1 ion channel may transmit mechanical signals via the ERK signaling pathway; however, the TRPV4 channel may function via alternative signaling pathways.

Introduction

Dental implants and ankylosed teeth cannot be moved by orthodontic force, which highlights the importance of periodontal ligament cells (PDLCs) in orthodontic tooth movement (1).

PDLCs are mechanically sensitive cells that may be induced by mechanical force, and they express a series of cytokines, including macrophage colony stimulating factor (M-CSF), receptor activator of NF-κB ligand (RANKL), osteoprotegerin (OPG) and inflammatory factors (2,3). The combined functional interventions that result from these factors serve to control the development of alveolar bone (4,5). However, the processes underlying the transduction of the associated mechanical signals to elicited biological signals, which is mediated by PDLCs, remain to be elucidated.

To date, 2 mechanotransduction mechanisms in the cells have captured considerable attention. First, attention has focused on the cytoskeletal-integrin-focal adhesion pathway, which receives and delivers the signals following transformation of microfilaments and microtubules (6,7). Second, attention has also focused on the mechanically sensitive ion channels (MSCs) on the surface of the cytomembrane, which transduce signals by increasing membrane permeability of the ion channels and triggering the influx of extracellular calcium (8). It has been proposed that the activation of calcium ion channels and subsequent calcium influx is associated with cytoskeletal transformation on induction by stress stimuli (9). However, this process was contradicted by Cox et al (10), which states that the integrity of the cytoskeleton is irrelevant in the context of Piezo1 ion channel function. The functional roles played by MSCs in orthodontic force-induced PDLC activation and the relationship between these two types of mechanotransduction have been poorly studied.

Piezo 1 and transient receptor potential cation channel subfamily V member 4 (TRPV4) are two typical MSCs that have received widespread attention from the research community. Piezo1 was first identified in a mouse neuroblastoma cell line; it was determined to respond to mechanical stimuli in as little as 5 msec and trigger calcium influx into the cells (11). A distinctive feature of the microscopic structure of Piezo1 is the flexible blades region, which is proposed to rotate and expose the central ion-conducting pore under mechanical stimulus (12).

Distinct from Piezo1, TRPV4 was initially recognized as an osmotically activated channel (13). Further studies identified that TRPV4 could be activated by fluid shear stress and phorbol ester (14,15). However, the gating mechanisms of TRPV4 remain to be elucidated. Although Piezo1 and TRPV4 are found in several mechanically sensitive cells (16–18), the downstream signal transduction pathways remain unknown.

Mitogen-activated protein kinase (MAPK) refers to a group of protein kinases that are associated with Piezo1 and the TRPV4 channel (19,20). It has been identified that an ERK1/2 inhibitor decreased the expression of Piezo1 in neonatal rat ventricular myocytes, whereas this effect was not observed when p38 and JNK inhibitors were applied (21). Additionally, the p38 inhibitor SB203580 enhanced the expression of TRPV4 in the dorsal root ganglion (22). Collectively, these observations suggest that MAPKs may participate in signal transduction pathways downstream of MSCs under conditions of mechanical loading.

In the present study, human primary PDLCs were subjected to stretch using a Flexcell device, leading to a model of stress-induced transformation. The roles played by MSCs in PDLC mechanotransduction were functionally analyzed by deconstructing the cytoskeleton using cytochalasin D (cytoD), or by blocking the Piezo1 channel using GsMTx4 or the TRPV4 channel using GSK205. The expression profiles of the MAPK signaling pathway in PDLCs when both of the MSCs were specifically blocked by targeted inhibition was also investigated.

Materials and methods

Cell culture

Human PDLCs were obtained from premolars that were extracted from 4 young donors for orthodontic consultation and treatment at the Jiangsu Stomatological Hospital. All donors were healthy ethnic Han Chinese females between 12 and 14 years old. The donors and their legal guardians were fully informed of the purpose of this study and provided written informed consent. All human experimental protocols were approved by the Ethics Committee of Shanghai Tenth People's Hospital [policy no. 2008 (20)]. The periodontal ligament was scraped from the root surfaces of the teeth and digested with collagenase type I (Sigma-Aldrich; Merck KGaA) for 30 min at 37°C. Cells were collected and resuspended in low-glucose DMEM (HyClone; GE Healthcare Life Sciences) that was supplemented with 15% FBS (ScienCell Research Laboratories, Inc.), 100 U/ml penicillin-G and 100 µg/ml streptomycin sulfate (HyClone; GE Healthcare Life Sciences). Cells were passaged when they reached ~90% confluence, and those from passages 3–5 were used in subsequent experiments.

Primary mouse osteoblasts were isolated from 20 2-3-day-old BALB/c neonatal female mice (Beijing Vital River Laboratory Animal Technology); animals were sacrificed on arrival. All animal experimental protocols were approved by the Ethics Committee of Shanghai Tenth People's Hospital (policy no. SHDSYY-2017-2473). The calvarial bones of the mice were cut into fractions and digested using 0.25% trypsin for 30 min and 1 mg/ml collagenase type II for 10 min (Sigma-Aldrich; Merck KGaA) at 37°C. Following digestion, the fractions were resuspended in DMEM supplemented with 10% FBS and incubated at 37°C in a humidified atmosphere of 95% air and 5% CO2. Cells were passaged when they reached 90% confluence, and those from passages 3–5 were used in subsequent experiments.

Cell loading

Human periodontal ligament cells (hPDLCs) were plated in BioFlex culture plates (type I collagen-coated; Flexcell International Corporation) and divided into four groups. The first was a control group, in which cells were loaded onto a Flexcell tension system (FX-5000T, Flexcell International Corporation) without any treatment. The second were cells in the cytoD group, which were pretreated with cytoD (Sigma-Aldrich; Merck KGaA) at a concentration of 5 µg/ml to depolymerize F-actin in the cells and incubated at 37°C for 20 min, cells were refreshed by addition of fresh culture medium without cytoD. For the third (GSK205) and fourth (GsMTx4) groups, fresh medium respectively containing 30 µM GSK205 (cat. no. 616522; Merck KGaA) and 500 nM GsMTx4 (cat. no. ab141871; Abcam) were added separately before cell loading to suppress the TRPV4 and Piezo1 channels, and cells were subjected to these compounds during loading Next, the plates were set on the FX-5000T. Then, 15% stretch was applied onto the cells for 4, 8 and 12 h. According to the study of Lu et al (9) on three different cell types, a stretch time duration of <3 sec fails to initiate calcium influx. Consequently, a 0.25-Hz square waveform of periodic loading, which persisted for 3 sec with a 1-sec release, was selected for the present study.

Calcium measurements

hPDLCs in all groups were washed three times in PBS following the stretch procedure. Then, the cells were treated with 1 µM Fluo-4AM (Beyotime Institute of Biotechnology) and incubated for 20 min in the dark at 37°C. Afterwards, the cells were washed three times in PBS and observed by fluorescence microscopy (DMI3000B; Leica Microsystems GmbH). The fluorescence intensity was measured using ImageJ 2X 2.1.4.6 software (National Institutes of Health). To do the measurements, the image was first converted to 8-bit grayscale and then inverted. The threshold was then set to ensure that only the cell regions were selected, and measurements were performed to obtain the area of the cells, the integrated intensity of the region and the average gray value of the selected region. The average gray value of each image was used for statistical analysis.

Cytoskeletal fluorescence staining

hPDLCs were washed three times in PBS following stretching and fixed in 4% paraformaldehyde (Beijing Leagene Biotech Co Ltd.) for 15 min at room temperature. Thereafter, the cells were permeabilized in 0.5% Triton X-100 for 15 min at room temperature and then rinsed three times in PBS. ActinGreen™ 488 ReadyProbes® Reagent (Thermo Fisher Scientific, Inc.) was added to all plates, after which they were incubated at room temperature for 30 min in the dark. Finally, cells were washed three times in PBS and observed by fluorescence microscopy at ×200 magnification (DMI3000B; Leica Microsystems GmbH). The average gray value of the cells were also measured by the method described above.

Reverse transcription-quantitative (RT-q)PCR

After 0, 4, 8 or 12 h loading, total RNA was extracted from the PDLCs using the Tiangen RNAprep pure kit (Tiangen Biotech Co., Ltd.). cDNA was obtained with the PrimeScript™ RT Master Mix reagent (Takara Bio, Inc.) following the manufacturer's instructions. The cDNA was added to a 20-µl system for qPCR using the SYBR Green Reaction kit (Roche Diagnostics) using the ABI Prism 7300 Real-Time PCR system (Applied Biosystems; Thermo Fisher Scientific, Inc.). The reaction was carried out under the following conditions: 94°C for 2 min, then 40 cycles of 94°C for 30 sec and 55°C for 1 min. The comparative 2−∆∆Cq method was used to calculate the fold-change of mRNA (23). Primers used are listed in Table I.

Table I.

Sequences of primers used in the present study.

mRNA	Primer sequence
RANKL	Forward: 5′-ACCGACATCCCATCTGGTT-3′
	Reverse: 5′-GCCATCCTGATTAACTATTAGTT-3′
OPG	Forward: 5′-AAGCCTTCTCTAACCTCTCC-3′
	Reverse: 5′-GCCCTCGCTTATGATCTGTC-3′
COX2	Forward: 5′-TTGAAATGGCAGTTGATTCCTTT-3′
	Reverse: 5′-TATCCTCTTTCTCAGGGTGCTTG-3′
Piezo1	Forward: 5′-GGCAACATGAGGGAGTTCATTAACTC-3′
	Reverse: 5′-TTCTCCGTCAGGTAGTTGACAATGTG-3′
TRPV4	Forward: 5′-GGCAACTTCCTCACCAAGA-3′
	Reverse: 5′-GGGTATTTCTTCTCTGTCTCT-3′
GAPDH	Forward: 5′-GGCACAGTCAAGGCTGAGAATG-3′
	Reverse: 5′-ATGGTGGTGAAGACGCCAGTA-3′

COX2, cyclooxygenase-2; RANKL, receptor activator of NF-κB ligand; OPG, osteoprotegerin; TRPV4, transient receptor potential cation channel subfamily V member 4.

Osteoblast mineralization induction

Murine osteoblasts that were derived from passages 3–5 were seeded into 24-well plates. Culture media from all four PDLC treatment groups, which were obtained following loading, were added to the plates when the cells reached ~90% confluence. The media was changed every three days. After 10 days of culture, cells were fixed in 95% ethanol at room temperature for 15 min, and then stained with 0.2% Alizarin red at room temperature for 5 min.

Western blot analysis

Cells obtained from the control group, the cytoD group and the GSK205 and GsMTx4 groups were washed in PBS following loading, following which they were lysed and their total protein was extracted using a whole-cell lysis kit (Wanleibio Co., Ltd.). Protein concentration was determined with an enhanced bicinchoninic acid protein assay reagent kit (cat. no. P0009; Beyotime Institute of Biotechnology). Proteins were subjected to electrophoresis in a 12% SDS polyacrylamide gel (40 µg protein per lane) and then transferred on to a PVDF immunoblotting membrane (EMD Millipore). Then, 5 µl molecular weight standards (cat. no. 26616; Fermentas; Thermo Fisher Scientific, Inc.) was loaded onto the gels as a reference. The membrane was blocked in 5% skimmed milk for 2 h at room temperature and incubated with cyclooxygenase-2 (COX2; 1:500; cat. no. ab179800; Abcam), RANKL (1:500; cat. no. ab65024; Abcam) and OPG (1:500; cat. no. ab183910; Abcam) primary antibodies at 4°C overnight. The membrane was washed six times in Tris-buffered saline containing 0.05% Tween 20 (TBST) for 5 min per wash. The membrane was then incubated in secondary antibodies (1:1,000; cat. no. WLA023; Wanleibio, Co., Ltd.) for 45 min at 37°C and washed six times in TBST. The membrane-immobilized protein bands were detected with chemiluminescent detection reagents (cat. no. WBKLS0100; EMD Millipore) in the WD-9413 fluorescence imaging system (Beijing Liuyi Biotechnology Co., Ltd.). The gray values for the visible bands were analyzed with the Gel-Pro Analyzer software 4.0 (Meyer Instruments, Inc.).

Expression profiling using protein array

Whole-cell proteins extracted from PDLCs in all four groups after the loading procedure were screened for their protein expression profiles using a Human Phospho-MAPK array kit (cat. no. ARY002B; R&D Systems, Inc.) according to the manufacturer's instructions, variations of >20% are considered potential candidates in the mechanotransduction of PDLCs.

Statistical analysis

Each experiment was repeated three times independently and data are presented as the mean ± SD. Statistical analyses were performed by one-way analysis of variance using GraphPad Prism 6 software (GraphPad Software, Inc.). Tukey's analysis was performed for multiple comparison tests. P<0.05 was considered to indicate a statistically significant difference.

Results

Increased mRNA expression of COX2, RANKL, OPG, Piezo1 and TRPV4 in PDLCs following periodic mechanical loading

The mRNA levels of key biomarkers, including COX2, RANKL and OPG, were increased after periodic loading at 0.25 Hz, indicating that the PDLCs were activated after stretch (Fig. 1). Additionally, the increased expression of TRPV4 and Piezo1 suggested that both ion channels were related to PDLC activation following loading (Fig. 1). However, it was noted that while most of the expression levels peaked at 8 h, additional stretch loading reduced PDLC activation, which could be attributed to the plate membrane losing its elasticity. Another reason for this observation may have been due to adaptation of the mechanosensitive cells and attenuation of the mechanotransduction process following loading, as was discussed in our previous study (24).

Figure 1.

Stretch loading induces the mRNA expression of COX2, RANKL, OPG, TRPV4 and Piezo1 in PDLCs. mRNA expression patterns of COX2, RANKL, OPG, TRPV4 and Piezo1 in PDLCs following loading. (A) OPG, (B) RANKL, (C) COX2, (D) TRPV4 and (E) for Piezo1. *P<0.05, **P<0.01 vs. 0 h. PDLC, periodontal ligament cell; COX2, cyclooxygenase-2; RANKL, receptor activator of NF-κB ligand; OPG, osteoprotegerin; TRPV4, transient receptor potential cation channel subfamily V member 4.

GsMTx4 and GSK205 do not interfere with cytoskeletal reconstruction in PDLCs, while CytoD does not affect calcium influx following loading

Fluorescent images showed that the intracellular calcium concentrations of PDLCs treated with GsMTx4 and GSK205 decreased compared with the control group following loading; however, cells that were treated with cytoD exhibited a fluorescence intensity similar to that of the control group (Fig. 2). This result suggested that the roles played by MSCs were independent of cytoskeletal integrity. The fluorescent images of the cytoskeleton showed that in cells that were pretreated with cytoD, a wrinkled appearance was noted and the intensity of the fluorescence decreased. By contrast, neither GsMTx4- nor GSK205-treated cells exhibited any differences in fluorescence intensity or cell morphology (Fig. 3).

Figure 2.

Calcium influx is inhibited by GsMTx4 and GSK205 after 8 h of loading. (A) Control group, (B) cytoD-pretreated group, (C) GsMTx4-treated group and (D) GSK205-treated group. The histogram shows quantified fluorescence intensity in each group. Compared with the group loaded normally (as the control), equivalent fluorescence intensity was emitted by cells that were pretreated with cytoD and comparatively less fluorescence was observed in the GsMTx4- and GSK205-treated groups. *P<0.05 vs. Control. cytoD, cytochalasin D.

Figure 3.

Structural integrity of the cytoskeleton is compromised by cytoD after 8 h of loading. (A) Control group, (B) cytoD-pretreated group, (C) GsMTx4-treated group and (D) GSK205-treated group. The histogram shows quantified fluorescence intensity in each group. CytoD destroyed F-actin expression in PDLCs; it was found that the fluorescence was reduced and the fibers were irregularly arranged. In the GsMTx4- and GSK205-treated groups, no significant difference was observed. *P<0.05 vs. Control. cytoD, cytochalasin D.

Destruction of the cytoskeleton and inhibition of mechanosensitive ion channels may compromise PDLC activation induced by periodic stretching

The biomarkers COX2, RANKL and OPG were all secreted by PDLCs after mechanical loading. COX2 is the rate-limiting enzyme that converts arachidonic acid to prostaglandins (PGs), which can regulate local inflammation and play a key role in the reconstruction of alveolar bone under orthodontic forces (25). The ratio of RANKL/OPG directly controls the balance between osteogenesis and osteoclastogenesis (26). It has been established that increased expression of these factors in PDLCs is induced by mechanical loading (27,28).

In the present study, the observed increases in biomarker expression were compromised by cytoD, GsMTx4 and GSK205 treatment (Fig. 4), which indicated that the efficiency of mechanotransduction was reduced. The results suggested that the cytoskeleton, Piezo1 and TRPV4 may affect signal transduction in PDLCs after stretch loading. The culture of osteoblasts with loaded PDLC supernatants also demonstrated that the supernatant of the cells treated with the selected inhibitors exhibited a diminished capacity to promote osteogenesis (Fig. 5).

Figure 4.

Expression levels of COX2, RANKL and OPG are downregulated by the inhibitors after 8 h of loading. Following treatment of cells with the inhibitors, the expression of COX2, RANKL and OPG was decreased compared with that found in the normally loaded group. *P<0.05, **P<0.01 vs. Control. COX2, cyclooxygenase-2; RANKL, receptor activator of NF-κB ligand; OPG, osteoprotegerin; cytoD, cytochalasin D.

Figure 5.

Culture supernatants collected from inhibitor-treated PDLCs exhibit reduced osteogenesis-promoting capacity after 8 h of loading. Osteoblasts that were cultured in the culture supernatant of PDLCs following loading exhibited differential capacities for mineralization. (A) Osteoblasts on plates: (a) Control group, (b) cytoD-pretreated group, (c) GsMTx4-treated group and (d) GSK205-treated group. (B) Osteoblasts under a light microscope: (a) Control group, (b) cytoD-pretreated group, (c) GsMTx4-treated group and (d) GSK205-treated group. Results suggested that the induction of osteoblast mineralization by loaded PDLCs could be suppressed by specific inhibitors. PDLC, periodontal ligament cell; cytoD, cytochalasin D.

Roles played by cytoskeletal and mechanosensitive ion channels in PDLC mechanotransduction may be mediated by MAPK pathways

Although all three inhibitors decreased PDLC activation that was otherwise induced by stretch, the profiles of MAPK-associated proteins in each group were observed to differ. Following a microarray analysis of phosphorylated MAPKs, the three groups of cells that were treated by the inhibitors were compared with the control group. Fold changes in the phosphorylation of proteins in each group are shown in Fig. 6.

Figure 6.

Fold change of phosphorylated-MAPK proteins in each group was compared with the control group. *P<0.05 and variations of >20% vs. Control. MAPK, mitogen-activated protein kinase; RSK, ribosomal S6 kinase; MSK, mitogen- and stress-activated kinase; MKK, MAPK kinase; HSP, heat shock protein; CREB, cAMP response element-binding protein.

Discussion

It has been established that PDLCs are mechanosensitive cells that secrete multiple cytokines to regulate the reconstruction of local alveolar bone (28,29). This mechanism could be considered the foundation of their role in orthodontic tooth movement (29,30). PDLCs at the pressure side are capable of continuous RANKL expression (28). Additionally, increased expression of PGE2, tumor necrosis factor-α, interleukin (IL)-1β, IL-2, IL-6, IL-8 and M-CSF has been detected in periodontal ligament loaded by orthodontic force (31,32). Tension loading can also activate PDLCs (32).

In the present study, the expression of COX2, RANKL and OPG in PDLCs was elevated following tension loading, which corresponds with results obtained from previously published studies (27). ELISA was not performed in the present study, as the aim of employing RT-qPCR was to find the time point of peak molecular expression of the genes of interest. In addition, changes in mRNA levels were considered to be more sensitive than the estimation of protein levels by assays such as ELISA.

Among the biomarkers, COX2 regulates local inflammation by regulating the synthesis of PGs (25). Increased COX2 expression has been detected in multiple mechanosensitive cells, and is considered to be an indication of extraneous stimulation intensity (24). RANKL and OPG are factors that play opposing roles and both participate in bone metabolism (26,33). Increases in these factors suggested that PDLCs were activated by periodic stretch loading. Additionally, increased TRPV4 and Piezo1 expression implied that calcium ion channels may be involved in mechanotransduction of stretch loading. When PDLCs are subjected to stress, increased levels of TRPV4 and Piezo1 enable greater efficiency of the cells in the mechanotransduction process and regulate their biological adaptation to the environment (34). Similar feedback regulation can be detected in cytoskeletal reconstruction following loading, as reported in our previous study (24). The strengthening of mechanical sensors in the cells following loading may be attributable to positive feedback mechanisms, which requires further investigation.

The roles played by the cytoskeleton in the process of stress-strain-cell signal transduction have been established (35–37). MSCs, particularly TRPV4 and Piezo1, have also been considered key factors in mechanotransduction (8,38). It has been suggested that mechanical loading can induce calcium influx and calcium-dependent cytoskeletal reorganization (39). However, it remains to be determined as to whether cytoskeletal transformation induced by stress can activate the calcium ion channels on the cytomembrane, or whether the calcium influx that is initiated by calcium ion channels can regulate cytoskeletal reorganization through downstream signaling pathways.

A previous study employing the patch-clamp assay revealed that the integrity of the cytoskeleton is not essential for the functioning of the Piezo1 channel (10). Conversely, certain investigators do not trust the patch-clamp database, as the surfaces of the cytomembrane are subjected to continuous tension forces during an experimental study, which may lead to results that differ from those found in normal cells (34). By contrast, another study found decreased calcium influx when induced pluripotent stem cells were pretreated with cytoD prior to application of 20% stretch force; such effects were not observed when 10–15% stretch was used (9).

In the present study, 15% stretch was loaded onto the PDLCs, and no effect on calcium influx was found by treatment with cytoD, consistent with the abovementioned studies. Additionally, such contradictions as are described in previous studies may imply that the Piezo1 channel, unlike TRPV4, is not involved in cytoskeletal reconstruction. In support of this, Matthews et al (40) reported that the force applied to β-1 integrins could activate the TRPV4 ion channel. In summary, no meaningful interactions between the cytoskeleton and these two ion channels were detected by the present study.

Fluxes in calcium ion concentrations play a critical role in cell-mediated biological activities that are regulated by calcium ion channels at the cytoplasmic membrane (41,42), thus enabling MSCs to regulate the cell cycle, proliferation, differentiation and apoptosis through calcium-dependent pathways upon application of a mechanical force (43). In the present study, administration of Piezo1 and TRPV4 antagonists decreased intracellular calcium concentrations following loading. The antagonists also attenuated the activation of three important biomarkers, RANKL, OPG and COX2, in PDLCs. The results suggested that both Piezo1 and TRPV4 are key factors in PDLC mechanotransduction. Jin et al (8) reported the effect of Piezo1 in this process; however, the mechanical loading in that study was performed under conditions of static pressure. This, combined with the results of the present study, suggests that both tension and pressure forces are recognized by the Piezo1 channel, which could initiate the expression of key cellular signals. By contrast, TRPV4 is a special ion channel that possesses similar permeability for the divalent cations Ca2+, Sr2+, Mg2+ and Ba2+; however, under physiologic conditions, Ca2+ is the predominant ion that migrates through the TRPV4 channel (13,44).

Son et al (38) studied the role of TRPV4 in PDLCs and found that when TRPV4 was activated by osmotic pressure, calcium influx was otherwise evoked. Meanwhile, although the expression of RANKL was increased, it was found that OPG expression remained unchanged (38). In the present study, the expression levels of both RANKL and OPG were decreased. This could be formally attributed to the fact that the mechanical force applied in the present study was a periodic stretch force. These conditions differed from those reported in the previous study (38) and could have resulted in diverse effects on TRPV4. Analysis of existing data has revealed that, although both Piezo1 and TRPV4 participate in mechanotransduction, they differ in terms of permeability, an ability to activate conditions and in their downstream signaling pathways.

That being considered, the MAPK family of signal transducing pathways consists of three major cascades, ERK, JNK and p38, all of which participate in cell activation by mechanical stress (45–47). However, it remains to be elucidated whether mechanical signals that are transduced by MSCs are indeed delivered by MAPK in PDLCs. It has been reported that the opening of Piezo1 induced by mechanical force can initiate chondrocyte apoptosis through ERK1 and ERK2 signaling pathways (20).

In the present study, no significant difference was found in the phosphorylation of ERK1/2 when the Piezo1 channel was inhibited. By contrast, a significant increase in glycogen synthase kinase (GSK)3α/β phosphorylation was observed. As GSK3 is a negative regulator of ERK1/2 (48), this might imply that GsMTx4 downregulates the ERK1/2 signaling pathway via the activation of GSK. Additionally, GSK3β contributes to β-catenin activation via Ras suppression (49). Furthermore, GSK3β can also directly inhibit the transcription factors c-Fos and c-Jun (50). The results of microarray analysis also showed increased phosphorylation of JNK3, JNK Pan and MAPK kinase (MKK)3 and 6 after Piezo1 channel attenuation. Among these, MKK3 and 6 are members of the p38-mediated signaling pathway, which catalyze p38 phosphorylation (51). Further, the JNK signaling pathway can activate c-Jun (52). It might be hypothesized that when the Piezo1 channel is blocked, the ERK-related signaling pathway is also inhibited. Increased functional activation of both JNK and p38 may be attributable to cellular compensation mechanisms under conditions where the ERK signaling pathway was compromised.

In contrast to the Piezo1 channel, when the TRPV4 ion channel was blocked by GSK205, increased phosphorylation of ERK, JNK, AKT and p70S6 kinase was observed. In addition, initiation of the AKT pathway is known to result from mechanical stimuli (53). The absence of calcium in the extracellular matrix has been noted to attenuate AKT phosphorylation (54). Until now, evidence supporting an interaction between TRPV4 and the MAPK signaling pathway in mechanotransduction, especially in PDLCs, has been lacking.

It has been reported that blocking of the ERK1/2 signal transduction pathway may prevent increases in TRPV4 channel activity induced by hypoosmotic stress in chondrocytes (55). Conversely, the effects of TRPV4 inhibition on MAPK sub-members have rarely been reported. As, in the present study, no significant suppressive effect was observed in the context of MAPK signaling molecules, it might be concluded that the signals received and transduced by TRPV4 were not delivered via the MAPK-mediated signaling pathways.

Microarray analysis also revealed that phosphorylation of ERK and p70S6 was enhanced, whereas that of p38α was diminished following loading and under conditions when the stress fibers were disrupted by cytoD. This observation suggests that p38 may participate in signal delivery via the cytoskeletal network. As reported by Qi et al (56), interference of stress fiber assembly by Rac1 knockdown may dampen p38 phosphorylation in vascular smooth muscle cells loaded by cyclic strain. Another study revealed the roles of p38 signaling in PDLC mechanotransduction loaded with mechanical stress (57).

There were certain limitations in the present study. Protein assays are an efficient way to screen out potential candidates in the mechanotransduction of PDLCs; the roles of the proteins need to be verified via western blotting or in vivo experiments. Furthermore, it has been established that several signaling molecules, including MAPK, Wnt and Notch, are involved in the process of mechanotransduction (58,59); however, the present study only focused on the MAPK pathway. Due to insufficient time and financial constraints, verification experiments could not be performed in the present study. Further investigations into these pathways will be conducted in the future.

In summary, the present study revealed the roles of Piezo1 and TRPV4 in the mechanotransduction of hPDLCs. In addition, it was suggested that both Piezo1 and TRPV4 may impart their functional activities via different signaling pathways distinct from actin cytoskeletal rearrangement. Furthermore, the present study indicated that the ERK and p38 signaling pathways might participate in mechanotransduction mediated by the Piezo1 channel and actin cytoskeleton, respectively. However, a similar role was not observed for the TRPV4 channel. Other signal transduction pathways might be engaged as a compensatory mechanism under conditions when an individual pathway might be compromised.