Long non‐coding RNA AC018926 .2 regulates palmitic acid exposure‐compromised osteogenic potential of periodontal ligament stem cells via the ITGA2/FAK/AKT pathway

Abstract Although obesity has been proposed as a risk factor for periodontitis, the influence of excessive fat accumulation on the development of periodontitis and periodontal recovery from disease remains largely unknown. This study investigated the cellular response of periodontal ligament stem cells (PDLSCs) to elevated levels of a specific fatty acid, namely, palmitic acid (PA). The mechanism by which PA exposure compromises the osteogenic potential of cells was also explored. It was found that exposure of PDLSCs to abundant PA led to decreased cell osteogenic differentiation. Given that long non‐coding RNAs (lncRNAs) play a key role in the stem cell response to adverse environmental stimuli, we screened the lncRNAs that were differentially expressed in PDLSCs following PA exposure using lncRNA microarray analysis, and AC018926.2 was identified as the lncRNA that was most sensitive to PA. Next, gain/loss‐of‐function studies illustrated that AC018926.2 was an important regulator in PA‐mediated osteogenic differentiation of PDLSCs. Mechanistically, AC018926.2 upregulated integrin α2 (ITGA2) expression and therefore activated ITGA2/FAK/AKT signalling. Further functional studies revealed that inactivation of ITGA2/FAK/AKT signalling by silencing ITGA2 counteracted the pro‐osteogenic effect induced by AC018926.2 overexpression. Moreover, the results of bioinformatics analysis and RNA immunoprecipitation assay suggested that AC018926.2 might transcriptionally regulate ITGA2 expression by binding to PARP1 protein. Our data suggest that AC018926.2 may serve as a therapeutic target for the management of periodontitis in obese patients.

considered overweight in 2017, of which 650 million were obese, equating to 39% of the adult population, and the prevalence continues to rise. 1 Obesity is a significant risk factor for various diseases, such as diabetes, dyslipidaemia, cardiovascular disease, osteoarthritis and periodontitis. [2][3][4][5] Periodontitis, initiated by dental plaque, is a multifactorial chronic inflammatory disorder of periodontal tissues. It can cause irreversible damage to the periodontium and, if left untreated, eventually lead to tooth loosening and even tooth loss. 6,7 In 2010, severe periodontitis was listed as the sixth most prevalent chronic disease in the world. 8 Many studies have illustrated that patients with obesity were more prone to develop periodontitis and suffered from more severe periodontal destruction compared to the individuals without obesity. [9][10][11] A characteristic feature of obesity is the elevated systemic and local levels of fatty acids (FAs). Recent studies have indicated that a highfat diet can aggravate periodontal lesions by accelerating bone destruction and impairing bone healing. [12][13][14][15] Previous researches on bone marrow mesenchymal stem cells, osteoblasts and osteoclasts have shown that specific FAs can regulate bone metabolism through mechanisms such as inflammation, apoptosis, autophagy and oxidative stress, 16 resulting in impaired bone formation and enhanced bone resorption. 17,18 Periodontal ligament stem cells (PDLSCs) are a type of undifferentiated mesenchymal cell with stem cell properties isolated from PDL tissues. These cells can differentiate into cementum, alveolar bone and PDL-like tissue, which lays a theoretical foundation for the application of PDLSCs in the regeneration of periodontal tissue. 19 At the same time, the differentiation of PDLSCs is a complex process involving many factors. Some special conditions, such as aging, inflammation and high glucose levels, will damage the multi-lineage potential of PDLSCs. [20][21][22] Periodontal bone regeneration is currently a major challenge in periodontal therapy. However, the effect of a high-fat condition on the osteogenic differentiation of PDLSCs and its underlying mechanisms are still poorly studied.
Long non-coding RNAs (lncRNAs) are a class of RNA transcripts, lacking protein-coding potential, with more than 200 nucleotides in length. 23,24 LncRNAs are involved in a variety of pathological and physiological processes. 25,26 It is worth noting that lncRNAs can serve as both potential diagnostic/prognostic biomarkers and important therapeutic targets for bone-related diseases. [27][28][29][30] In recent years, scholars have made progress in uncovering the role of lncRNAs in regulating periodontal bone homeostasis. In the inflammatory microenvironment, lncRNA GACAT2 can regulate PDLSC cementoblastic differentiation by binding to PKM1/2 proteins 22 ; lncRNAs HIF1A-AS1 and HIF1A-AS2 regulate osteogenic differentiation of PDLSCs by affecting the activity of hypoxia-inducible factor (HIIF)-1α under hypoxia 31 ; lncRNAs may also mediate force-induced PDLSC osteogenic differentiation during orthodontic tooth movement in response to mechanical stress. 32 Based on the findings of these studies, we hypothesize that lncRNAs play a crucial role in the damaged PDLSC osteogenic potential induced by a high-fat environment. In this study, we established a high-fat condition with a specific FA, palmitic acid (PA), to monitor the PA-induced changes in the osteogenic differentiation of PDLSCs. A key lncRNA, AC018926.2, was screened out and identified to be downregulated in PDLSCs exposed to PA compared to those under a normal condition, and it was shown to be involved in regulating PDLSC osteogenic potential in an environment with PA. Mechanistically, the results revealed that AC018926.2 elicited its biological functions by increasing the expression level of integrin α2 (ITGA2) and therefore activating ITGA2/FAK/AKT signalling. Overall, our research suggested that AC018926.2 may serve as a novel therapeutic target for the recovery of periodontitis in obese individuals.

| MATERIALS AND METHODS
Detailed materials and methods are described in the Supplementary Information.

| PA exposure compromised the osteogenic differentiation of PDLSCs
Primary human PDLSCs were successfully isolated from the PDL tissues of 6 fresh teeth ( Figure S1A). The results of the colony-forming unit (CFU) ( Figure S1B) and CCK-8 ( Figure S1C) assays confirmed that the PDLSCs have the ability to proliferate and multiply. The cells could also differentiate into osteogenic, adipogenic and chondrogenic lineages, as verified by Alizarin red S staining, Oil red O staining and Alcian blue staining, respectively ( Figure S1D). Flow cytometry analysis ( Figure S1E) showed that these PDLSCs positively expressed CD105, CD44, CD90 and CD146 and negatively expressed CD45 and CD34. Then, these cells were used for the following in vitro experiments.
In order to build a high-fat condition, PA was added to the osteogenic medium (PA group) to adjust its final concentration to 200 μM, and the control group was treated with the same amount of solvent control (FA-free bovine serum albumin, BSA). The use of PA at a concentration of 200 μM was based on previously published literature. [33][34][35][36] Then, we systematically monitored the effects of PA exposure on the osteogenic potential of PDLSCs through a series of related experiments. We found that ALP activity was significantly decreased in the PA group compared with the control group, as indicated by ALP staining and quantification ( Figure 1A). Alizarin red S staining and quantification on Day 21 showed that extracellular mineralization was also reduced in the PA group ( Figure 1B). Meanwhile, when the cells were exposed to PA for 14 days, the expression levels of osteogenesis-related genes COL1, RUNX2, ALP, BMP2 and OCN were dramatically impaired, as demonstrated by a quantitative realtime polymerase chain reaction (qRT-PCR) assay ( Figure 1C). Moreover, Western blot analysis revealed that the levels of osteogenesisrelated proteins COL1, RUNX2 and BMP2 were significantly decreased in the PA group on Day 14 ( Figure 1D). These results indicate that PA exposure could impair the osteogenic differentiation of PDLSCs.

| lncRNA microarray analysis and qRT-PCR validation
As the osteogenic differentiation of PDLSCs was significantly inhibited after 14 days of PA exposure, we performed a microarray analysis to identify the expression profiles of lncRNAs in PDLSCs 14 days after osteogenic induction when cells were exposed to PA (PA group) or BSA (Ctrl group). A total of 760 lncRNAs (p < 0.05 and fold change >1.5) were differentially expressed between the two groups, of which  (C) Expression levels of the osteogenesis-related genes COL1, RUNX2, ALP, BMP2 and OCN in PDLSCs after 14 days of exposure to PA measured by qRT-PCR. (D) Expression levels of the osteogenesis-related proteins COL1, RUNX2, ALP and BMP2 in PDLSCs after 14 days of exposure to PA determined by Western blot analysis. All experiments were performed with 3 biological replicates. Data are presented as the mean ± SD (n = 3). *p < 0.05, **p < 0.01, and ***p < 0.001 represent significant differences between the indicated columns, while NS represents no significant difference.
, by qRT-PCR assays. All experiments were performed with 3 biological replicates. Data are presented as the mean ± SD (n = 3). *p < 0.05, **p < 0.01 and ***p < 0.001 represent significant differences between the indicated columns, while NS represents no significant difference. confirmed by the quantitative evaluation of osteogenesis-related genes and proteins. As shown in Figure 3D

| Reduced expression of AC018926.2 impaired the osteogenic differentiation of PDLSCs under a normal condition
In addition to gain-of-function assays, we also performed loss-offunction assays. siRNA transfection technology was utilized to knock down the expression of AC018926.2 in PDLSCs under a normal condition, and Figure 4A shows the knockdown efficiency.  All experiments were performed with 3 biological replicates. Data are presented as the mean ± SD (n = 3). *p < 0.05, **p < 0.01 and ***p < 0.001 represent significant differences between the indicated columns, while NS represents no significant difference.
3.5 | AC018926.2 was positively correlated with the activity of the ITGA2/FAK/PI3K/AKT pathway Because the regulatory mechanisms of lncRNAs are associated with their subcellular localization, 37 we conducted subcellular fractionation assay to evaluate the distribution of AC018926.2. We found that AC018926.2 was primarily observed in the nuclear fraction and was less prevalent in the cytoplasm ( Figure 5A). This suggested that AC018926.2 may regulate the transcription of target genes. 38 To better understand how AC018926.2 affects PDLSC osteogenic differentiation, we performed RNA transcriptome sequencing following AC018926.2 knockdown in PDLSCs. As shown in the heatmap, 196 differentially expressed genes were identified (fold change >2 and p < 0.05) in the PDLSCs from the si-AC018926.2 group compared to those from the si-NC group, including 138 downregulated genes and 58 upregulated genes ( Figure 5B). Furthermore, the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was utilized to investigate the signalling pathways that may be impacted by AC018926.2. Figure 5C  2) or without (si-NC) AC018926.2 knockdown were assessed with Western blot analysis. The relative intensity of ITGA2 was normalized to β-actin, and the relative intensity of p-FAK and p-AKT was normalized to respective total FAK and total AKT (F, G). All experiments were performed with 3 biological replicates. Data are presented as the mean ± SD (n = 3). *p < 0.05, **p < 0.01 and ***p < 0.001 represent significant differences between the indicated columns. support vector machine (SVM) values were higher than 0.5 ( Figure 7A). Therefore, these four proteins were selected for further analysis. To determine the effects of DNMT3B, DNMT1, PARP1 and SUZ12 on ITGA2 transcription levels, we transfected PDLSCs with DNMT3B, DNMT1, PARP1 and SUZ12 siRNAs, and the knockdown efficiency is shown in Figure 7B. qRT-PCR assays revealed that DNMT3B and SUZ12 knockdown evidently enhanced the gene level of ITGA2, PARP1 knockdown notably suppressed the gene level of ITGA2, while DNMT1 knockdown did not influence the gene level of ITGA2 ( Figure 7C). Subsequently, DNMT1 was excluded, and RNA immunoprecipitation (RIP) assays were used to detect the interaction between the remaining three proteins, DNMT3B, SUZ12 and PARP1 and AC018926.2. The results showed that AC018926.2 was significantly enriched in PARP1 antibody when compared to the control F I G U R E 6 ITGA2 inhibition impaired the AC018926.2 overexpression-rescued osteogenic differentiation of PDLSCs under the PA condition. (A, B) ITGA2, p-FAK, total FAK, p-AKT and total AKT protein levels in the control and si-ITGA2 groups measured by Western blot analysis. The relative intensity of ITGA2 was normalized to β-actin, and the relative intensity of p-FAK and p-AKT was normalized to respective total FAK and total AKT. The cells were incubated in normal medium transfected with si-ITGA2. (C) Representative images of ALP staining and quantification of ALP activity in PDLSCs after osteogenic induction for 14 days (scale bar: 500 μm). (D) Representative images of Alizarin red S staining and quantitative analysis of the calcium mineral deposits formed by PDLSCs after osteogenic induction for 21 days (scale bar: 500 μm). (E) The effect of ITGA2 silence on the expression levels of the osteogenesis-related genes COL1, RUNX2, ALP, BMP2 and OCN in AC018926.2 overexpressed PDLSCs following 14 days of osteogenic induction measured by qRT-PCR. (F) The effect of ITGA2 silence on the osteogenesis-related proteins COL1, RUNX2, ALP and BMP2 in AC018926.2 overexpressed PDLSCs following 14 days of osteogenic induction determined by Western blot analysis. The cells were incubated in osteogenic medium with PA and co-transfected with LV-AC018926.2 and si-ITGA2 (C-F). All experiments were performed with 3 biological replicates. Data are presented as the mean ± SD (n = 3). *p < 0.05, **p < 0.01 and ***p < 0.001 represent significant differences between the indicated columns, while NS represents no significant difference.
immunoglobulin G (IgG) antibody (14.5-fold change compared to IgG by qRT-PCR from RIP assay) ( Figure 7D). Meanwhile, the gene expression level of PARP1 was not affected by AC018926.2 downregulation or upregulation ( Figure 7E). These data indicated that AC018926.2 specifically interacted with PARP1. To further verify that AC018926.2 influences the expression of ITGA2 by binding to PARP1, we overexpressed AC018926.2 and silenced PARP1 to examine the expression of ITGA2 measured by qRT-PCR. We found that PARP1 Data are presented as the mean ± SD (n = 3). *p < 0.05, **p < 0.01 and ***p < 0.001 represent significant differences between the indicated columns, while NS represents no significant difference. downregulation impaired the impact that AC018926.2 upregulation had on enhancing the mRNA level of ITGA2 ( Figure 7F).

| DISCUSSION
In the present study, we observed that PA exposure impaired the osteogenic potential of PDLSCs. Moreover, we found that AC018926.2 was significantly downregulated in PDLSCs exposed to PA condition, which was associated with the compromised osteogenic potential of PDLSCs. Mechanistically, as an important regulator in the osteogenesis process, AC018926.2, mainly distributed in the nucleus, Increasing evidence has confirmed that obesity and obesityrelated complications are associated with the onset, progression and recovery of periodontitis. 2,3,49,50 However, the specific mechanism is still under investigation. In obese patients, plasma-free FAs, derived from adipose tissues and diet, are elevated and have proinflammatory effects. 51 Therefore, increased levels of free FAs in obese patients might contribute to the association between obesity and periodontal destruction. 52 PA is the most abundant saturated FA in daily diet and adipose tissues, 53 and elevated PA levels in the serum can be converted into increased PA levels in the bone tissues. 33 Based on this, we focused on how PA will affect PDLSC osteogenic differentiation.
Nevertheless, there remains controversy in the previous studies regarding the effect of PA on bone metabolism. Some studies suggest that PA exhibits high lipotoxicity and can negatively affect the osteogenic differentiation of osteoblasts, 18,33,54 while other studies suggest that PA affects bone homeostasis mainly by influencing osteoclasts rather than osteoblasts. [48][49][50] According to our present study supported by systematical observation, when exposed to PA, the osteogenic potential of PDLSCs was notably impaired (Figure 1). The discrepancies between previous studies may be partially related to different PA exposure duration. Considering obesity is a chronic state, to better simulate in vivo situation, we adopted the prolonged PA exposure (at least 14 days as was used in our study) rather than the 6-8 days incubation protocol in most previous studies. [34][35][36] LncRNAs are important components of non-coding RNAs. It has been found that lncRNAs can regulate gene expression at the transcriptional, posttranscriptional and epigenetic levels through various mechanisms, thereby positively or negatively influencing the osteogenic differentiation of PDLSCs under both physiological and pathological conditions. 30,55,56 However, whether lncRNAs are involved in the regulation of the damaged osteogenic potential of PDLSCs under a PA condition remains to be determined. We explored the lncRNAs that are differentially expressed in PA-exposed PDLSCs and their controls by microarray analysis as they may be potential targets for peri-  (Figures 3 and 4). Based on these observations, AC018926.2 was identified as an active regulator in the osteogenic differentiation of PDLSCs.
From the results of RNA sequencing and KEGG analysis, we found that the FAK and PI3K/AKT pathway were significantly downregulated following AC018926.2 knockdown ( Figure 5C). Previous studies have shown that activation of the FAK/PI3K/AKT pathway is positively associated with cell osteogenic differentiation. 39,40 As an important member of the integrin family, ITGA2 has been reported to enhance the osteogenic capacity of cells by activating the FAK signalling pathway. 43,44 Furthermore, the RNA sequencing results showed that ITGA2 expression was decreased after AC018926.2 knockdown ( Figure 5B). In addition, qRT-PCR assays confirmed that the gene level of ITGA2 was affected by overexpression/knockdown of AC018926.2 ( Figure 5D, E). Therefore, we speculated that AC018926.2 promoted the osteogenic differentiation of PDLSCs by activating the ITGA2/ FAK/PI3K/AKT pathway through regulating the expression of ITGA2.
As expected, the ITGA2/FAK/PI3K/AKT pathway was activated following AC018926.2 upregulation and inactivated following AC018926.2 downregulation (Figure 5F, G). When the pathway was inhibited by specific siRNAs for ITGA2, the ability of overexpressed AC018926.2 to enhance osteogenic differentiation of PDLSCs under PA exposure was blocked ( Figure 6).
The subcellular localization of lncRNAs is closely related to their functional mode and regulatory mechanism. 37 It has been proven that lncRNAs located in the nucleus can transcriptionally regulate their downstream target genes through interactions with specific proteins. 57 To further investigate the mechanism by which AC018926.2 influences the expression of ITGA2, we used bioinformatics analysis to predict potential proteins that might interact with AC018926.2 and potential transcription factors of ITGA2, and unfortunately there was no intersection of the both above mentioned. Among the predicted proteins that might interact with AC018926.2, DNMT3B, DNMT1, PARP1 and SUZ12 have been reported to be associated with the transcription of ITGA2. 46-48 Therefore, they were selected for subsequent experiments, and qRT-PCR assays demonstrated that the knockdown of DNMT3B, PARP1 and SUZ12 significantly affected the level of ITGA2 in PDLSCs ( Figure 7C). RIP assays confirmed the specific binding of AC018926.2 to PARP1 but not to DNMT3B or SUZ12 ( Figure 7D). The rescue experiment verified that AC018926.2 transcriptionally promoted ITGA2 expression in PDLSCs by recruiting PARP1 ( Figure 7F). Nevertheless, this regulation may not be direct, and more mechanistic investigations are required to further confirm this hypothesis. PARP1 has been reported to regulate gene expression by interacting with chromatin, chromatin modifiers, or transcription factors. 58 Meanwhile, recent studies have indicated that PARP1 is also an RNA-binding protein. Man et al. 59  hust.edu.cn/AnimalTFDB4/#/). PARP1 is located in the nucleus, which is consistent with the nuclear localization of AC018926.2. We assumed that AC018926.2 that is enriched in the nucleus might independently recruit PARP1 to target promoters or facilitate its interaction with the transcription factor of ITGA2. The specific mechanisms will be further explored in our future work.
It should be noted that this is a relatively preliminary exploration on the effect of a high-fat condition on the osteogenic differentiation potential of PDLSCs. Although we identified for the first time that AC018926.2 served as an active regulator in the process of PA exposure-impaired osteogenic differentiation of PDLSCs, and AC018926.2 was positively correlated with the activity of ITGA2/ FAK/PI3K/AKT pathway, it is not clear whether AC018920.2 changed in abundance due to altered transcriptional regulation of AC018920.2 or simply due to differing content of cell types following 14 days of treatment with PA at present. Nevertheless, our results clarified that the forced expression of AC018926.2 was able to reverse PA-compromised PDLSC osteogenic differentiation, thus indicating that AC018926.2 is an important target for improving osteogenic differentiation potential of cells exposed to a PA condition. To go deeply into the research, we will try to use methods such as single-cell sequencing to further explore whether the reduced expression of AC018926.2 is due to the transcriptional downregulation in all cells or due to the altering abundance of individual cell types present in these cultures.
In conclusion, our study identified and validated that the lncRNA

CONFLICT OF INTEREST STATEMENT
The authors declare no potential conflicts of interest to declare.

DATA AVAILABILITY STATEMENT
The lncRNA microarray data are available in the GEO databases (accession number GSE218434).
F I G U R E 8 A proposed model illustrating the function and mechanism of AC018926.2 in PDLSC osteogenic differentiation with PA exposure.