Mechanical stress alters protein O‐GlcNAc in human periodontal ligament cells

Abstract Protein O‐linked N‐acetylglucosamine (O‐GlcNAc) is a post‐translational modification of intracellular proteins that regulates several physiological and pathophysiological process, including response to various stressors. However, O‐GlcNAc's response to mechanical stress has not been investigated yet. As human periodontal ligament (PDL) cells are stimulated by compression force during orthodontic tooth movement that results in structural remodelling, in this study we investigated whether mechanical stress induces any alteration in protein O‐GlcNAc in PDL cells. In this study, PDL cells isolated from premolars extracted for orthodontic indications were exposed to 0, 1.5, 3, 7 and 14 g/cm2 compression forces for 12 hours. Cell viability was measured by flow cytometry, and protein O‐GlcNAc was analysed by Western blot. Cellular structure and intracellular distribution of O‐GlcNAc was studied by immunofluorescence microscopy. We found that between 1.5 and 3 g/cm2 mechanical compression, O‐GlcNAc significantly elevated; however, at higher forces O‐GlcNAc level was not increased. We also found that intracellular localization of O‐GlcNAc proteins became more centralized under 2 g/cm2 compression force. Our results suggest that structural changes stimulated by compression forces have a significant effect on the regulation of O‐GlcNAc; thus, it might play a role in the mechanical stress adaptation of PDL cells.

single N-acetylglucosamine saccharide is attached to serine and threonine hydroxyl groups of nuclear and cytoplasmic proteins. 3 Hundreds of proteins have been identified as being subjected to protein O-GlcNAc including transcription factors, cytoskeletal and signalling components and metabolic enzymes. 4 This modification is highly dynamic when responding to various stimuli. Protein O-GlcNAc shares some features when compared to serine/threonine phosphorylation, such as the existence of cyclic enzymes or the occupation of the same residues on proteins. 5 There is compelling evidence that O-GlcNAc is involved in the regulation of several physiological (transcription, 6 nutrient sensing 7 and cell cycle regulation 8 ) and pathological processes (diabetes, 9 cancer, 10 Alzheimer disease 11 ). For example, in patients with diabetes, chronic hyperglycaemia will lead to elevated levels of protein O-GlcNAc modification 12 which could have several deleterious consequences, such as altered transcriptional factor activity or interference with phosphorylation. 13 Importantly, recent researches have also found altered O-GlcNAc levels associated with increased stress tolerance.
In conditions such as oxidative stress, hypoxia or heat shock, the reg-

| Cells and culture conditions
Primary human PDL cells were isolated from healthy, non-carious first premolars undergoing tooth extraction for orthodontic indi- large cytoplasm-to-nucleus ratio, single cells capable to migrate and forming large bundles, without any apparent contact inhibition (Appendix Figure A1). The cells were cultured in a 1:1 mixture of EMEM and Ham's F12 medium (Lonza) supplemented with 10% foetal bovine serum (Thermo Fischer Scientific), 1% non-essential amino acids, penicillin (100 U/mL, Sigma-Aldrich ™ ) and streptomycin (100 µg/mL, Sigma-Aldrich ™ ). The cells were incubated at 37°C, 5% CO2 in a humidified incubator. Subculturing was performed after reaching confluency. The medium was refreshed 12-24 hours prior to each experiment. At least 5 independent human PDL primary cell lines were obtained, and experiments were carried out between the third and fifth passages. Prior to sample collection, written informed consents were obtained from all patients. The procedures were approved by the Regional Committee for the Research Ethics

| Application of mechanical stress
To simulate continuous mechanical stress, a previously described compression method was used. 18,19 Briefly, PDL cells were plated

| Immunofluorescence microscopy
Cells were grown on coverslips in 6-well plates until approx. 15% confluency in complete media. Subsequently, the coverslips were turned over; thus, the cellular layers were put between the coverslips and the surface of the plates. The coverslips were either subjected to 0 (just coverslip) or 2 g/cm 2 compressive force for 12 hours. Control cells were kept on the coverslips in upside position. At the end of the treatments, all coverslips were turned back upside carefully, the media was removed, and the cells were  Figure A2).

| Data analysis
Data are presented as mean ± standard deviations (SD) throughout.
Comparisons were performed by one-way ANOVA plus Dunnett's post hoc test using SPSS software. Statistically significant differences between groups were defined as P Values < 0.05 and are indicated in the figure legends.

| Cell viability
We used compression force of up to 14 g/cm 2 to test the mechanical resistance of cultured PDL cells. Control samples were incubated for the same time period but did not receive any mechanical load. Cells in the 0 g/cm 2 group were covered with only a coverslip. The average mechanical load because of the coverslips on the cells is ~26 ± 0.4 mg, that is <2% of the next, smallest load in our experimental setup. We assessed the cell viability after 12 hours of mechanical compression by PI staining. As shown in Figure 1A, dead cells could be clearly separated from live cells by their increased PI uptake. The viability of control cells was 97.1 ± 2.88%, whereas the viability of the 0 g/cm 2 group was similar, 97.5 ± 1.1%. The percentage of living cells in the compressed groups were not significantly different from either the control or the 0 g/cm 2 group (P = 0.390); however, we found a decreasing tendency towards the increased mechanical compression. Viability was 95.6 ± 3.29% in the 1.5 g/cm 2 group, which decreased to 88.8 ± 8.6% in the 14 g/cm 2 group ( Figure 1B).

| Mechanical stress altered protein O-GlcNAc
The effect of mechanical stress on O-GlcNAc levels was inves-

| Immunofluorescence detection shows altered O-GlcNAc distribution in compressed PDL cells
We have also investigated the relationship between the cytoskel- with coverslip only. In cells exposed to 2 g/cm 2 compression force, these membrane protrusions seemed to be retracted and narrowed.
Correspondingly, O-GlcNAc staining also tended to be more central-

| D ISCUSS I ON
In this study, we demonstrated for the first time that protein O-GlcNAc is significantly altered in response to mechanical compression in human PDL cells in vitro. PDL cells enduring 1.5-3 g/cm 2 pressure for 12 hours developed a significant increase in O-GlcNAc levels compared with control. The intracellular distribution of O-GlcNAc-rich proteins also changed upon compression force. Our results suggest that the regulation of O-GlcNAc has an "optimal" range in relation to mechanical challenges.
The orthodontic force is an extrinsic mechanical stimulus that evokes cellular responses and therefore allows orthodontic tooth movement. 20,21 The biologic effect also depends on the PDL area over which the force is distributed; therefore, the net effect should be considered as force per unit area or pressure. Ren et al have concluded that very small pressure (<2 g/cm 2 ) can stimulate biologic responses. However, neither the exact threshold nor the optimal pressure magnitude could be defined. 22 Other studies found that light force is preferable to avoid potential overloading that can hinder tooth movement. 23 In most in vitro studies, compression force was used in the range of 0-4 g/cm 2 and 2-3 g/cm 2 was found to induce the most significant changes. 18,19,24 Our results correlate with these studies as O-GlcNAc changes were most prominent at 1.5-3 g/cm 2 compression force. In contrast to in vitro studies, large individual differences have been shown in response to orthodontic forces in vivo.
Early animal studies have suggested that a continuous forces of not more than 15-20 g/cm 2 should be used for optimum biological tooth movement, as higher forces may lead to adverse effects. 25 More recently, forces between 1.2 and 10 g for up to 14 days were used to move rat molars. 26,27 This discrepancy from in vitro findings could be explained by different bone/mineral density, variation in individual anatomic structures, or different structure of the collagen fibres and cellular activity. One of the limitations of our study is that we have used beta-actin as an internal control for the Western blot analyses.
Actin is widely used as a general loading control in stress-related Western blot studies 28,29 ; on the other hand, its dynamic nature is F I G U R E 1 Periodontal ligament cells survive mechanical compression for 12 h. Cells were exposed to 0-14 g/cm 2 of mechanical compression for 12 h. The ratio of living cells was measured using propidium iodide (PI) viability assay via flow cytometry. (A) Representative dot plot chart of PDL cells compressed by 7 g/cm 2 force for 12 h. PI staining (y-axis, FL3 Log) was plotted against forward scatter (x-axis, FS Lin) which is proportional to cell size. Live (bottom region) and dead (upper region) cells were separated based on FL3 signal intensity (PI staining). (B) Average ratio of live cells compared with the total number of cells. Bars are representing mean values ± SD from at least 3 independent experiments. *P < 0.05 vs control a key player in the cellular motility and morphology. Nevertheless, a few studies suggest that the overall level of actin is not changing significantly after mechanical stress. 19,30 The success of orthodontic tooth movement depends largely on the remodelling capability of alveolar bone. Impaired glucose metabolism in diabetes has been demonstrated to influence bone metabolism and bone formation. 31,32 Several studies have reported significant differences in bone response to orthodontic stimulus under diabetic conditions. 31,33 Furthermore, metformin administration was found to diminish the adverse effects of diabetes on tooth movement. 31 There is an increasing number of evidences demonstrating a strong linkage between increased O-GlcNAc and diabetic complications. 34  Pathophysiologic processes such as tumour growth and inflammation in closed compartments will also introduce mechanical challenges to cells. Thus, we think that our data may contribute to the better understanding of these processes.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest. Note: Data are means ± SD, n = 12 cells/ group # P = 0.763 vs control ‡ P = 0.016 vs control.

AUTH O R CO
TA B L E 1 Average distances of O-GlcNAc-positive regions from the nuclei F I G U R E 3 Protein O-GlcNAc distribution pattern compared with actin, tubulin and vimentin staining after compression stress in human PDL cells. PDL cells were grown on coverslips (control), or the coverslips were turned upside-down for 12 h (coverslip) or turned upsidedown and compressed with 2 g/cm 2 force for 12 h (2 g/cm 2 ). Immunofluorescence labelling was performed on PDL cells fixed on coverslips. Representative epi-fluorescence images of cells stained with anti-cytoskeletal filament antibodies (actin, tubulin or vimentin-green, first column), anti-O-GlcNAc antibody CTD110.6 (red, second column) and Hoechst nuclear staining (blue, 3rd column) are shown. Corresponding merged images are shown in the 4th column, whereas higher magnification images (outlined by squares in the merged images) displaying membrane protrusions are shown in the 5th column