ARHGAP17 inhibits pathological cyclic strain-induced apoptosis in human Periodontal Ligament Fibroblasts via Rac1/Cdc42

Li Wanga, b, Xiaojie Yangb, Leilei Wanb, Shiwei Wangb, Yuehua Liua,*, Jinsong Panb,*

aDepartment of Orthodontic, Shanghai Stomatological Hospital, Fudan University, No.356 East Beijing Road, Shanghai, China
bDental Department, Shanghai 1st People’s Hospital Affiliated to Shanghai Jiao Tong University, No. 100 Haining Road, Hongkou District, Shanghai, China

*Corresponding author:
Yuehua Liu. Department of Orthodontic, Shanghai Stomatological Hospital, Fudan University, No.356 East Beijing Road, Shanghai, China. E-mail address: [email protected].
Jinsong Pan. Dental Department, Shanghai 1st People’s Hospital Affiliated to Shanghai Jiao Tong University, No. 100 Haining Road, Hongkou District, Shanghai, China. E-mail address: [email protected].

Short title: The role of ARHGAP17 in periodontal ligament cells

List of abbreviations:
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/1440-1681.13336

PDL: periodontal ligament
Rho-GAP: Rho GTPase activating protein Rho-GDI: Rho GDP dissociation inhibitor

Summary
Rho GTPase activating protein (Rho-GAP) and Rho GDP dissociation inhibitor (Rho- GDI) are two mainly negative regulators of Rho GTPase. Our previous work has found that Rho-GDI and Rho GTPase are involved in the response of human periodontal ligament (PDL) cells to mechanical stress. However, whether Rho-GAP also has a role in this process remains unknown. Here, we attempted to find the Rho-GAP gene that may be involved in pathological stretch-induced apoptosis of PDL cells. Human PDL fibroblasts were exposed to 20% cyclic strain for 6h or 24h, after which the expression levels of ARHGAP10, ARHGAP17, ARHGAP21, ARHGAP24 and ARHGAP28 were determined. Results showed that ARHGAP17 expression decreased the most obviously after treatment of stretch. In addition, ARHGAP17 overexpression abolished 20% cyclic strain-induced apoptosis. Therefore, ARHGAP17 has an important role in pathological stretch-induced apoptosis of human PDL fibroblasts. Moreover, we found that ARHGAP17 overexpression also alleviated cyclic strain-induced activation of Rac1/Cdc42, a major downstream target of ARHGAP17. Furthermore, two Rac1 inhibitors, NSC23766 and EHT 1864, both attenuated ARHGAP17 knockdown-mediated apoptosis in human PDL fibroblasts. Collectively, our data demonstrate that ARHGAP17 inhibits pathological cyclic strain-induced apoptosis in human PDL fibroblasts through inactivating Rac1/Cdc42. This study highlights the importance of Rho signaling in the response of human PDL fibroblasts to mechanical stress.

Keywords: ARHGAP17, Rac1/Cdc42, apoptosis, pathological cyclic strain, human Periodontal Ligament Fibroblasts

1. Introduction
Periodontal ligament (PDL) is a soft tissue that connect tooth with alveolar bone, and it is directly subjected to mechanical stress caused by occlusion of teeth. During orthodontic tooth movement, appropriate mechanical stimulation is essential for the remodeling of PDL tissue 1,2, while excessive mechanical force will cause cell death in PDL tissue 3.
Both mechanical tension and pressure have been found to promote apoptosis in human PDL cells 4,5. Moreover, many apoptosis-related genes, such as Caspase genes, express abnormally in human PDL fibroblasts exposed to mechanical stretch 6,7. Recent studies further reveal that Caspase3, Caspase5, Caspase8 and Caspase9 are implicated in cyclic stretch-induced apoptosis in human PDL fibroblasts8-11. In our previous work, we further found that Rho-GDIα, a Rho GDP dissociation inhibitor (Rho-GDI), is involved in the pro-apoptotic mechanism of pathological cyclic strain in human PDL fibroblasts by regulating Caspase3 and poly-ADP-ribose polymerase (PARP)8.
Rho GTPase, which belongs to Ras superfamily, is a family of small signaling G proteins. Like other G proteins, Rho GTPases are inactivated in GDP-bound state, while in GTP-bound state they are activated and interact with downstream effectors 12. Rho GTPase activating protein (Rho-GAP) is a main negative regulator of Rho GTPase activity, which stimulates the hydrolysis of GTP that is bound to Rho GTPase 13,14. Previous studies have focus on the role of Rho-GAP in cancer. For example, ARHGAP10 inhibits the proliferation, migration and invasion of ovarian cancer cells15. ARHGAP17 suppresses tumor progression in cervical cancer 16. Recently, mounting Rho-GAPs have been found to be involved in cellular mechanical stress response

pathway, such as ARHGAP21, ARHGAP24 and ARHGAP28 17-19. Our previous studies have found that Rho GTPase and Rho-GDI are involved in pathological cyclic strain-induced cytoskeleton rearrangement and apoptosis in human PDL fibroblasts 8,20,21. However, little is known about the role of Rho-GAP in the response of human PDL fibroblasts to pathological stretch.
In this study, we first investigated the effects of pathological cyclic strain on the expression levels of ARHGAP10, ARHGAP17, ARHGAP21, ARHGAP24 and ARHGAP28 in human PDL
fibroblasts. We found that ARHGAP17 decreased the most obviously after 6h or 24h of stretch treatment. Furthermore, we demonstrate that ARHGAP17 inhibits pathological cyclic strain-induced apoptosis in human PDL fibroblasts by inactivating its downstream target Rac1/Cdc42.

2. Results
2.1 Identification of human PDL fibroblasts
To identify human PDL fibroblasts, fibroblast marker vimentin and epithelial marker CK-19 were detected by immunocytochemical staining. Results showed that cells at the fourth passage stained positively for vimentin and negatively for CK-19 (Figure 1A, S1A). Moreover, Mesenchymal cell-associated surface markers (CD90 and CD105) and hematopoietic surface markers (CD34 and CD45) were measured by flow cytometer. Results showed that these cells were positive for CD90 and CD105 while negative for CD34 and CD45 (Figure 1B, S1B). Together, these data indicated that these cells were pure PDL fibroblasts. Previous studies have found that PDL fibroblasts maintain their fibroblast characteristics until 15th passage22,23. Here, passages 4-8 were used for subsequent experiments.
2.2 Pathological cyclic strain significantly inhibited ARHGAP17 expression in human PDL fibroblasts
Next, we investigated the effects of pathological stretch on the expression of Rho-GAP genes in PDL fibroblasts. Cells were exposed to 20% cyclic strain for 6h or 24h, and then the expression levels of ARHGAP10, ARHGAP17, ARHGAP21, ARHGAP24 and ARHGAP28 were detected

by RT-qPCR and western blotting. Results showed that after treatment with 20% cyclic strain, the mRNA levels of ARHGAP10, ARHGAP17, ARHGAP24 and ARHGAP28 were decreased, while ARHGAP21 mRNA remained at a constant level (Figure 2A, S2A). The protein levels of these Rho-GAP genes exhibited a similar trend (Figure 2B, S2B). As shown in Table 1, ARHGAP17 mRNA decreased the most obviously in PDL fibroblasts exposed to 20% cyclic strain. Therefore, we focused on the role of ARHGAP17 in pathological stretch-mediated apoptosis of PDL fibroblasts in subsequent experiments.
2.3 oeARHGAP17 overexpressed ARHGAP17 while siARHGAP17 knocked down ARHGAP17 expression in human PDL fibroblasts
Lentiviruses oeARHGAP17 and siARHGAP17 were constructed to regulate ARHGAP17 expression in human PDL fibroblasts. To examine their effectiveness, we determined the mRNA and protein levels of ARHGAP17 in PDL fibroblasts after 24h of lentivirus transduction. RT-qPCR analysis showed that ARHGAP17 mRNA was significantly up-regulated by oeARHGAP17 and down-regulated by siARHGAP17; and siARHGAP17-3 had the highest inhibition on ARHGAP17 mRNA expression (Figure 3A, S3A). Western blotting showed similar results in protein level (Figure 3B, S3B). Thus, oeARHGAP17 and siARHGAP17-3 were selected for subsequent analyses.
2.4 ARHGAP17 overexpression repressed pathological cyclic strain-induced apoptosis and activation of Rac1/Cdc42 in human PDL fibroblasts
To investigate if ARHGAP17 is involved in pathological stretch-mediated apoptosis of human PDL fibroblasts, cells were transduced with oeARHGAP17 and then were exposed to 20% cyclic strain for 24h. Flow cytometric analysis and Tunel assay showed that 20% cyclic strain promoted apoptosis in human PDL fibroblasts while ARHGAP17 overexpression attenuated cyclic strain-induced apoptosis (Figure 4A, 4B, S4A, S4B). Western blotting showed that 20% stretch increased the pro-apoptotic protein cleaved-Caspase3, while ARHGAP17 overexpression abolished stretch-mediated elevation of cleaved-Caspase3 (Figure 4C, S4C). Therefore, ARHGAP17 has an important role in pathological cyclic strain-induced apoptosis in human PDL fibroblasts.

Since Rac1/Cdc42 is a major downstream target of ARHGAP17 24, which has been found to promote apoptosis in many cells 25-27, we also investigated the effect of pathological cyclic strain on Rac1/Cdc42 activity. Results showed that 20% stretch enhanced the activity of Rac1/Cdc42 while ARHGAP17 overexpression abrogated stretch-promoting activation of Rac1/Cdc42 (Figure 4D, S4D), indicating that pathological cyclic strain promotes the activation of Rac1/Cdc42 via regulating ARHGAP17.
2.5 ARHGAP17 knock-down promoted the apoptosis of human PDL fibroblasts via enhancing the activity of Rac1/Cdc42
To investigate if ARHGAP17 affects the apoptosis of human PDL fibroblasts by regulating Rac1/Cdc42, the cells were transduced with siARHGAP17, and two Rac1 inhibitors, NSC23766 and EHT 1864, were also used. Western blotting showed that ARHGAP17 knock-down increased the activity of Rac1/Cdc42, while both two Rac1 inhibitors abolished siARHGAP17-induced activation of Rac1/Cdc42 (Figure 5A, S5A). Flow cytometric analysis and Tunel assay showed that ARHGAP17 knock-down promoted the apoptosis of PDL fibroblasts while both two Rac1 inhibitors attenuated siARHGAP17-induced apoptosis (Figure 5B, 5C, S5B, S5C). Taken together, these findings indicated that ARHGAP17 knock-down promoted the apoptosis of human PDL fibroblasts via enhancing the activity of Rac1/Cdc42.

3. Discussion
Mechanical stress is critical for chewing and orthodontic treatment, and acts directly on PDL tissue. Physiological mechanical force is essential for the homeostasis of PDL tissue 1, while pathological mechanical stimulation will lead to PDL cell apoptosis 8. Studies have reported that Caspase proteins are involved in cyclic strain-induced the apoptosis of human PDL fibroblasts, including Caspase3, Caspase5, Caspase8 and Caspase9 9-11,28. Our previous works found that Rho GTPase and its negative regulator Rho-GDIα, are involved in cyclic strain-induced cytoskeletal rearrangement and apoptosis in PDL cells 8,20,21. Here, we further investigated the role of Rho-GAP, another important regulator of Rho GTPases, in cyclic strain-induced apoptosis in human PDL fibroblasts.

To screen Rho-GAP genes that may be involved in the response of human PDL fibroblasts to pathological stretch, we measured the expression levels of ARHGAP10, ARHGAP17, ARHGAP21, ARHGAP24 and ARHGAP28. We found that ARHGAP17 expression level decreased the most obviously after treatment of 20% cyclic strain. Furthermore, we found that ARHGAP17 overexpression attenuated the apoptosis of human PDL fibroblasts induced by 20% cyclic strain. Collectively, ARHGAP17 has an important role in pathological stretch-induced apoptosis. Rac1/Cdc42 is a major downstream target of ARHGAP17, and is inactivated due to ARHGAP17-mediated GTP hydrolysis on Rac1/Cdc42.24 Overexpression of Rac1/Cdc42 has been found to promote apoptosis in many cells, such as macrophages and chondrocytes.25,26 Therefore, we further explored if ARHGAP17 modulated pathological stretch-induced apoptosis by regulating Rac1/Cdc42. Our study found that ARHGAP17 overexpression partly abolished 20% cyclic strain-mediated activation of Rac1/Cdc42. Moreover, both two Rac1 inhibitors, NSC23766 and EHT 1864, attenuated the apoptosis of human PDL fibroblasts induced by ARHGAP17 knock-down. Taken these together, we conclude that ARHGAP17 inhibits pathological stretch-induced apoptosis in human PDL fibroblasts by inactivating Rac1/Cdc42 (Figure 6).
Previous studies of ARHGAP17 mainly focus on its role in cancer. ARHGAP17 has been
found to be a cancer suppressor in colon, cervix and breast cancer 16,29,30. To our knowledge, no study has investigated the role of ARHGAP17 in human PDL fibroblasts. Moreover, whether ARHGAP17 is involved in the process of cell apoptosis remains unknown. Here, we found that ARHGAP17 knock-down promotes the apoptosis of human PDL fibroblasts in the physiological context, while ARHGAP17 overexpression inhibits pathological stretch-induced apoptosis. Therefore, our study reveals that ARHGAP17 has an anti-apoptosis function and is involved in the response of human PDL fibroblasts to pathological mechanical stimulation.
Rho signals have been found to participate in stretch-promoting cytoskeletal rearrangement 20,21. Our previous study also observed that Rho-GDIα is involved in pathological stretch-promoting apoptosis 8. In this study, we further found that ARHGAP17 and Rac1/Cdc42 mediate pathological stretch-induced human PDL fibroblasts apoptosis, underlining the importance of Rho signaling in the response of human PDL fibroblasts to mechanical stress.

In conclusion, we elucidate that ARHGAP17 is involved in pathological stretch-induced apoptosis of human PDL fibroblasts by inactivating Rac1/Cdc42. This study highlights the crucial role of Rho signaling in the response of human PDL fibroblasts to mechanical stress.

4. Materials and Methods
4.1 Samples collection and cell culture
This study was approved by the Ethics Committee of Shanghai Stomatological Hospital, Fudan University. Human PDL cells were collected from healthy premolar tissues of three different donors (11-year-old boy; 12-year-old girl; 15-year-old girl) for orthodontic reasons. Cells were cultured in Dulbecco’s modified Eagle medium (Hyclone, USA), supplemented with 20% fetal bovine serum (Gibco, USA), 100 U/ml penicillin and 100 mg/ml streptomycin. The medium was placed in a humidified incubator of 5% CO2 at 37℃.
4.2 Identification of human PDL fibroblasts
Fibroblast marker vimentin and epithelial marker CK-19 were often detected to identify human PDL fibroblasts31,32. Here, cells at fourth passage were immunocytochemically stained with anti-vimentin and anti-CK-19 antibodies for characterization. Firstly, cells were fixed with 4% paraformaldehyde for 30min, and then cultured with 3% H2O2 for 10min. After being blocked with 1% bovine serum albumin (Thermo, USA), cells were incubated with primary antibodies against vimentin (ab45939, 1:500, Abcam, USA) or CK-19 (NBP2-29804, 1:500, Novus, USA) at room temperature for 1h, and then incubated with horseradish peroxidase (HRP)-conjugated secondary antibody (D-3004, 1:1000, Long-Island Biotec, China) for 30min at room temperature. For visualization of antibodies, cells were stained with diaminobenzidine and hematoxylin in turn.
Moreover, flow cytometric analysis was conducted for the characterization of mesenchymal cell-associated surface markers (CD90 and CD105) and hematopoietic surface markers (CD34 and CD45). Briefly, cells were digested with 0.25% ethylenediaminetetraacetic acid (EDTA)-trypsin and then washed using PBS. Afterwards, these cells were resuspended at a concentration of 1×106 cells/ml. 100μl cell suspension were incubated with CD34-PE (No. 551387, BD Biosciences, USA), CD45-PE (No. 561087, BD Biosciences), CD90-PE-CyTM5 (No. 561972, BD Biosciences)

and CD105- PerCP-CyTM5.5 (No. 560819, BD Biosciences) for 25min at 4℃ in the dark. Finally, the stained cells were washed by PBS and then detected using a flow cytometer (BD Biosciences, USA).
4.3 Cyclic strain treatment
Cyclic strain treatment was following our previous workflow 8. PDL cells were seeded into 6-well Bioflex plates at 3×105 cells/well density. When reaching 95% confluence, cells were subjected to 20% cyclic stretch at 0.1 Hz for 6 or 24h using Flexcell Tension Plus system FX-5000T (Flexcell International, USA). Cells cultured under the same condition except without cyclic strain loading were used as control. The morphologies of cells with or without stretch treatment were examined using an inverted phase-contrast microscope (Figure S6).
4.4 Real-time RT-PCR (RT-qPCR)
Total RNA was extracted using Trizol reagent (Invitrogen, USA), and then was digested with endonuclease DNase I to reduce residual DNA. The cDNA was synthesized using reverse transcription kit (Fermentas, USA) according to the manufacturer’s protocol. RT-qPCR was performed with SYBR Green Mix (Thermo, USA) in ABI 7300 Real-Time PCR System (ABI, USA). The mRNA levels of target genes were normalized to β-actin and GAPDH mRNA. PCR primers are listed in Table 2.
4.5 Western blotting
Total protein was extracted with RIPA lysis buffer and then quantified by BCA protein assay kit (Thermo, USA). Equivalent protein amounts were separated by 10% SDS-PAGE and transferred to PVDF membranes. After being blocked with 5% nonfat milk, these membranes were incubated using primary antibodies and HRP-conjugated secondary antibody (A0208, 1:1000, Beyotime, China) in turn. Primary antibodies included antibodies against ARHGAP10 (ab222805, 1:500, Abcam, USA), ARHGAP17 (ab74454, 1:500, Abcam, USA), ARHGAP21 (ab122350, 1:500, Abcam, USA), ARHGAP24 (18834-1-AP, 1:500, proteintech, USA), ARHGAP28 (26766-1-AP, 1:500, proteintech, USA) and β-actin (60008-1-Ig, 1:5000,
Proteintech, USA). Finally, the target proteins were determined using ECL-HRP detection kit (Millipore, USA).

4.6 Lentivirus construction and transduction
The lentiviruses for ARHGAP17 overexpression (oeARHGAP17) and knock-down (siARHGAP17-1, siARHGAP17-2 and siARHGAP17-3) were constructed commercially. Briefly, the target DNA fragments of oeARHGAP17 and siARHGAP17 were inserted into the plasmid pLVX-Puro (Clontech, USA) and PLKO.1 (Addgene, USA), respectively. Then the vector plasmids were co-transfected into 293T cells with packaging plasmids psPAX2 (Addgen, USA) and pMD2G (Addgen, USA), High-titer recombinant lentiviruses were collected after 72h of transfection. The sequences of ARHGAP17 siRNAs are as follows: siARHGAP17-1, GCAGAGGAAGCAGCTTGCAAGATTG; siARHGAP17-2,
GCGATTGTGTTAGGCCCTAACTTGT; siARHGAP17-3, TGTCTTGAAGAAGTTGGCTTTCTTT.
For lentivirus transduction, 5×105 cells were seeded into 6-well plates and 2ml medium was added into each well. When reaching 50-70% confluence, cells were transduced with lentiviruses at a multiplicity of infection (MOI) of 5 for 24h at 37℃ . Afterwards, the medium was replaced with fresh medium. Subsequent analyses were performed 48h after transduction.
4.7 Apoptosis detection
Flow cytometric analysis was carried out to detect apoptotic cells. Cells were digested using 0.25% trypsin-EDTA solution and then suspended in PBS to count. Next, 5×105 cells were centrifuged at 1000 rpm for 5min and resuspended using 195μL Annexin V-FITC (Beyotime, China). Afterwards, cells were incubated with extra 5μL Annexin V-FITC (Beyotime, China) for 15min, and then incubated with 5μL propidium iodide for 5min. Apoptotic cells were measured in flow cytometer (BD Biosciences, USA).
Apoptosis was also detected by Tunel assay using in situ cell death detection kit (No. 11684817910, Roche Diagnostics, Shanghai, China). Cells were fixed with 4% paraformaldehyde for 30min and then digested using trypsin for 40min. Afterwards, the cells were incubated with TUNEL reaction mixture in a humidified chamber at 37˚C for 1h. 4′, 6-diamidino-2-phenylindole (DAPI) was used to stain the nuclei. Tunel-positive cells were visualized by a fluorescence microscope (Nikon Eclipse Ni, Japan). The percentage of apoptotic cells was determined by counting the

TUNEL-positive cells and the total number of cells in five random high fields.
4.8 Rac1/Cdc42 activity assay
Rac1 activity was detected using Rac1 Activation Assay Biochem KitTM (Cytoskeleton, USA). Total protein was extracted using lysis buffer and then quantified using Precision Red Advanced Protein Kit (Cytoskeleton, USA). Extra lysis buffer was added to equalize the protein concentration in each sample. Equal volumes of samples were incubated with 15μg PAK-PBD beads for 1h at 4 ˚C. Afterwards, the beads were washed four times using 500μl wash buffer and then boiled with 20μl 2×Laemmli sample buffer for 2min. Active Rac1 was detected using western blotting with anti-Rac1 antibody (1:500; Cytoskeleton, Denver, CO, USA). For Cdc42 activity assay, the pull-down experiments were the same as described for Rac1, but active Cdc42 was detected with anti-Cdc42 antibody (1:250; Cytoskeleton).
4.9 Statistical analysis
Experimental data were shown as mean value ± standard deviation. Statistical analyses were conducted in prism ver. 8.0.2 (GraphPad, USA). One-way ANOVA and subsequent Tukey test were selected to compare the differences among groups. P-value<0.05 means statistically significant.

5. Conflict of interest
The authors declare no conflicting interests.

6. Funding
This study was supported by National Natural Science Foundation of China Youth Fund (No. 11602140).

7. References
1. Meeran NA. Biological response at the cellular level within the periodontal ligament on application of orthodontic force–An update. J Orthod Sci 2012;1:2.
2. Krishnan V, Davidovitch Z. Cellular, molecular, and tissue-level reactions to orthodontic

force. Am J Orthod Dentofac 2006;129:469. e1-. e32.
3. Hatai T, Yokozeki M, Funato N, et al. Apoptosis of periodontal ligament cells induced by mechanical stress during tooth movement. Oral Dis 2001;7:287-90.
4. Zhong W, Xu C, Zhang F, Jiang X, Zhang X, Ye D. Cyclic stretching force‐induced early
apoptosis in human periodontal ligament cells. Oral Dis 2008;14:270-6.
5. Goga Y, Chiba M, Shimizu Y, Mitani H. Compressive force induces osteoblast apoptosis via caspase-8. J Dent Res 2006;85:240-4.
6. Ritter N, Mussig E, Steinberg T, Kohl A, Komposch G, Tomakidi P. Elevated expression of genes assigned to NF-κB and apoptotic pathways in human periodontal ligament fibroblasts following mechanical stretch. Cell Tissue Res 2007;328:537-48.
7. Xu C, Hao Y, Wei B, et al. Apoptotic gene expression by human periodontal ligament cells following cyclic stretch. J Periodont Res 2011;46:742-8.
8. Wang, Pan J, Wang T, Song M, Chen W. Pathological cyclic strain-induced apoptosis in human periodontal ligament cells through the RhoGDIα/caspase-3/PARP pathway. PLoS One 2013;8:e75973.
9. Zhao D, Wu Y, Xu C, Zhang F. Cyclic-stretch induces apoptosis in human periodontal ligament cells by activation of caspase-5. Arch Oral Biol 2017;73:129-35.
10. Hao Y, Xu C, Sun S, Zhang F. Cyclic stretching force induces apoptosis in human periodontal ligament cells via caspase-9. Arch Oral Biol 2009;54:864-70.
11. Wu Y, Zhuang J, Zhao D, Xu C. Interaction between caspase ‐ 3 and caspase ‐ 5 in the
stretch‐induced programmed cell death in the human periodontal ligament cells. J Cell Physiol 2019;234:13571-81.
12. Mackay DJ, Hall A. Rho GTPases. J Biol Chem 1998;273:20685-8.
13. DerMardirossian C, Bokoch GM. GDIs: central regulatory molecules in Rho GTPase activation. Trends Cell Biol 2005;15:356-63.
14. Schwartz M. Rho signalling at a glance. J Cell Sci 2004;117:5457-8.
15. Luo N, Guo J, Chen L, Yang W, Qu X, Cheng Z. ARHGAP10, downregulated in ovarian cancer, suppresses tumorigenicity of ovarian cancer cells. Cell Death Dis 2016;7:e2157-e.

16. Guo Q, Xiong Y, Song Y, Hua K, Gao S. ARHGAP17 suppresses tumor progression and up-regulates P21 and P27 expression via inhibiting PI3K/AKT signaling pathway in cervical cancer. Gene 2019;692:9-16.
17. Borges L, Bigarella CL, Baratti MO, et al. ARHGAP21 associates with FAK and PKCζ and is redistributed after cardiac pressure overload. Biochemical and biophysical research communications 2008;374:641-6.
18. Duval D, Rimbert A, Kyndt F, et al. 0030: ARHGAP 24/(FilGAP); a new gene responsible for mitral valve prolapse. Archives of Cardiovascular Diseases Supplements 2014;6:72-3.
19. Yeung C-YC, Taylor SH, Garva R, et al. Arhgap28 is a RhoGAP that inactivates RhoA and downregulates stress fibers. PLoS One 2014;9.
20. Pan J, Wang T, Wang L, Chen W, Song M. Cyclic strain-induced cytoskeletal rearrangement of human periodontal ligament cells via the Rho signaling pathway. PLoS One 2014;9:e91580.
21. Wu J, Song M, Li T, Zhu Z, Pan J. The Rho-mDia1 signaling pathway is required for cyclic strain-induced cytoskeletal rearrangement of human periodontal ligament cells. Exp Cell Res 2015;337:28-36.
22. Costa FPd, Puty B, Nogueira LS, et al. Piceatannol Increases Antioxidant Defense and Reduces Cell Death in Human Periodontal Ligament Fibroblast under Oxidative Stress. Antioxidants 2020;9:16.
23. Nogueira LS, Vasconcelos CP, Mitre GP, et al. Oxidative damage in human periodontal ligament fibroblast (hPLF) after methylmercury exposure. Oxidative Medicine and Cellular Longevity 2019;2019.
24. Richnau N, Aspenström P. Rich, a rho GTPase-activating protein domain-containing protein involved in signaling by Cdc42 and Rac1. J Biol Chem 2001;276:35060-70.
25. Liang S, Liu H, Zhou A. Lovastatin-induced apoptosis in macrophages through the Rac1/Cdc42/JNK pathway. J Immunol 2006;177:651-6.
26. Wang G, Beier F. Rac1/Cdc42 and RhoA GTPases antagonistically regulate chondrocyte proliferation, hypertrophy, and apoptosis. J Bone Miner Res 2005;20:1022-31.
27. Fukuda A, Hikita A, Wakeyama H, et al. Regulation of osteoclast apoptosis and motility by

small GTPase binding protein Rac1. J Bone Miner Res 2005;20:2245-53.
28. Wu Y, Zhao D, Zhuang J, Zhang F, Xu C. Caspase-8 and Caspase-9 functioned differently at different stages of the cyclic stretch-induced apoptosis in human periodontal ligament cells. PLoS One 2016;11:e0168268.
29. Pan S, Deng Y, Fu J, et al. Tumor suppressive role of ARHGAP17 in colon cancer through
Wnt/β-catenin signaling. Cellular Physiology and Biochemistry 2018;46:2138-48.
30. Kiso M, Tanaka S, Saji S, Toi M, Sato F. Long isoform of VEGF stimulates cell migration of breast cancer by filopodia formation via NRP1/ARHGAP17/Cdc42 regulatory network. International journal of cancer 2018;143:2905-18.
31. Zhang LN, Wang XX, Wang Z, Li KY, Xu BH, Zhang J. Berberine improves advanced glycation end products ‑ induced osteogenic differentiation responses in human periodontal ligament stem cells through the canonical Wnt/β ‑ catenin pathway. Molecular medicine reports 2019;19:5440-52.
32. Ying Y, Luo J. Salidroside promotes human periodontal ligament cell proliferation and osteocalcin secretion via ERK1/2 and PI3K/Akt signaling pathways. Experimental and therapeutic medicine 2018;15:5041-5.

Tables
Table 1. The quantified effects of 20% cyclic strain on the mRNA levels of ARHGAP10, ARHGAP17, ARHGAP21, ARHGAP24 and ARHGAP28 in human PDL fibroblasts

The mRNA levels of five Rho-GAP genes in human PDL fibroblasts at 6h or 24h after 20% cyclic strain treatment were normalized to their expression levels at 0h after 20% cyclic strain treatment. These ratios were shown below corresponding gene name as ‘A/B’. A: the ratio is calculated using GAPDH as reference gene. B: the ratio is calculated using β-actin as reference gene.

Table 2. Primers for RT-qPCR

Genes Forward primers (5’-3’) Reverse primers (5’-3’)
ARHGAP10 AGAAGGGACTCTGAACGG CTCGGAGGAGGAAGACAG
ARHGAP17 GCTAACCAGACCGTGGGCAGAG CTGGAAACATGCCACCAA
ARHGAP21 AACCTTCACGACCCATTC TCGCCTTCTTCTTCCATT
ARHGAP24 CCCTGAGAACTCCAACTC CTGGTGGCACGACTACTT
ARHGAP28 GATACTGGGCTTCTACAT TTCCCTCTAACACTTTCT
β-actin ACACTGTGCCCATCTACG TGTCACGCACGATTTCC
GAPDH AATCCCATCACCATCTTC AGGCTGTTGTCATACTTC

Figure legends:

Figure 1. Identification of human PDL fibroblasts. (A) Immunocytochemical staining showed that

cells at the fourth passage stained positively for vimentin and negatively for CK-19. Nuclei were stained blue by hematoxylin (×200). (B) Mesenchymal cell-associated surface markers (CD90 and CD105) and hematopoietic surface markers (CD34 and CD45) were measured by flow cytometer.

Figure 2. The expression levels of ARHGAP10, ARHGAP17, ARHGAP21, ARHGAP24 and ARHGAP28 in human PDL fibroblasts at 0h, 6h or 24h after 20% cyclic strain treatment. (A) The mRNA levels were detected by RT-qPCR. GAPDH and β-actin were selected as reference genes.
(B) The protein levels were detected by western blotting. *p<0.05, **p<0.01 and ***p<0.001 vs 0h.

Figure 3. The effects of lentivirus oeARHGAP17 and siARHGAP17 on the expression level of ARHGAP17 in human PDL fibroblasts. (A) The mRNA level was detected by RT-qPCR. (B) The protein level was detected by western blotting. ****p<0.0001 vs vector; ####p<0.001 vs siNC.

Figure 4. oeARHGAP17 attenuated pathological cyclic strain-induced apoptosis and activation of Rac1/Cdc42 in human PDL fibroblasts. Cells were treated with oeARHGAP17 and 20% cyclic strain for 24h. (A and B) Apoptotic cells were detected by flow cytometer (A) and Tunel assay (B). (C) Protein levels of ARHGAP17 and cleaved-Caspase3 were measured by western blotting.
(D) Rac1/Cdc42 activity was determined using Rac1/Cdc42 activation assay. ****p<0.0001 vs control; ##p<0.01 and ####p<0.0001 vs vector+20% 24h.

Figure 5. ARHGAP17 knock-down promoted the apoptosis of human PDL fibroblasts via enhancing the activity of Rac1/Cdc42. Cells were pre-incubated with Rac1 inhibitor, NSC23766 or EHT 1864, and then transduced with siARHGAP17. (A) The activity of Rac1/Cdc42 was detected by Rac1/Cdc42 activation assay. (B and C) Apoptotic cells were determined by flow cytometer (B) and Tunel assay (C). ****p<0.0001 vs siNC; ####p<0.0001 vs siNC+NSC23766.

Figure 6. Schematic figure outlining the involvement of ARHGAP17 and Rac1/Cdc42 in the

apoptosis of human PDL fibroblasts exposed to pathological cyclic strain.

Figure S1. Identification of human PDL fibroblasts. (A) Immunocytochemical staining showed that cells at the fourth passage stained positively for vimentin and negatively for CK-19. Nuclei were stained blue by hematoxylin (×200). (B) Mesenchymal cell-associated surface markers (CD90 and CD105) and hematopoietic surface markers (CD34 and CD45) were measured by flow cytometer.

Figure S2. The expression levels of ARHGAP10, ARHGAP17, ARHGAP21, ARHGAP24 and ARHGAP28 in human PDL fibroblasts at 0h, 6h or 24h after 20% cyclic strain treatment. (A) The mRNA levels were detected by RT-qPCR. GAPDH and β-actin were selected as reference genes.
(B) The protein levels were detected by western blotting. *p<0.05, **p<0.01, ***p<0.001 and
****p<0.0001vs 0h.

Figure S3. The effects of lentivirus oeARHGAP17 and siARHGAP17 on the expression level of ARHGAP17 in human PDL fibroblasts. (A) The mRNA level was detected by RT-qPCR. (B) The protein level was detected by western blotting. **p<0.01 and ***p<0.001vs vector; ##p<0.01, ###p<0.001 and ####p<0.001 vs siNC.

Figure S4. oeARHGAP17 attenuated pathological cyclic strain-induced apoptosis and activation of Rac1/Cdc42 in human PDL fibroblasts. Cells were treated with oeARHGAP17 and 20% cyclic strain for 24h. (A and B) Apoptotic cells were detected by flow cytometer (A) and Tunel assay (B). (C) Protein levels of ARHGAP17 and cleaved-Caspase3 were measured by western blotting.
(D) Rac1/Cdc42 activity was determined using Rac1/Cdc42 activation assay. ***p<0.001 and
****p<0.0001 vs control; #p<0.01 and ####p<0.0001 vs vector+20% 24h.

Figure S5. ARHGAP17 knock-down promoted the apoptosis of human PDL fibroblasts via

enhancing the activity of Rac1/Cdc42. Cells were pre-incubated with Rac1 inhibitor, NSC23766 or EHT 1864, and then transduced with siARHGAP17. (A) The activity of Rac1/Cdc42 was detected by Rac1/Cdc42 activation assay. (B and C) Apoptotic cells were determined by flow cytometer (B) and Tunel assay. ****p<0.0001 vs siNC; ###p<0.001 and ####p<0.0001 vs siNC+NSC23766.

Figure S6. Pathological cyclic strain changed the morphology of human PDL fibroblasts. (A) Human PDL fibroblasts without cyclic strain treatment are aligned multi-directionally. (B) Human PDL fibroblasts exposed to 20% cyclic strain for 24h are aligned in one direction, with the long axis perpendicular to the cyclic strain force vector. The black arrow represents the direction of the cyclic strain. (×200).