LncRNA GAS5 regulates osteogenic differentiation of human periodontal ligament stem cells through miR-23b-3p/STAT5B axis

Abstract

Periodontal ligament stem cells (PDLSCs) play a crucial role in alveolar bone regeneration due to their inherent osteogenic differentiation capacity. However, the molecular mechanisms regulating PDLSC osteogenic differentiation remain incompletely understood. In this study, human PDLSCs were isolated and characterized, revealing that the cytoplasmic long non-coding RNA GAS5 was significantly upregulated during osteogenic induction. Functional assays demonstrated that GAS5 overexpression markedly enhanced alkaline phosphatase (ALP) activity, mineralized nodule formation, and the expression of osteogenic marker genes (RUNX2, OCN, OPN). Conversely, GAS5 knockdown attenuated these osteogenic phenotypes. Mechanistically, GAS5 functioned as a competitive endogenous RNA (ceRNA) by directly binding to miR-23b-3p, thereby preventing miR-23b-3p-mediated post-transcriptional silencing of its target gene STAT5B. Rescue experiments confirmed that STAT5B silencing abolished the pro-osteogenic effects induced by GAS5 overexpression, while miR-23b-3p inhibition reversed the osteogenic impairment caused by GAS5 knockdown. Collectively, our findings reveal a novel regulatory axis (GAS5/miR-23b-3p/STAT5B) that modulates PDLSC osteogenic differentiation, offering a potential therapeutic target for periodontal tissue regeneration strategies.

Highlights

• •GAS5 sponges miR-23b-3p, promoting osteogenic differentiation in PDLSCs.
• •miR-23b-3p targets STAT5B; repressed by GAS5, upregulating STAT5B.
• •Overexpression of GAS5 or inhibition of miR-23b-3p enhances osteogenesis in PDLSCs.
• •Knockdown of GAS5 or overexpression of miR-23b-3p suppresses osteogenesis in PDLSCs.
• •Targeting the GAS5/miR-23b-3p/STAT5B axis offers a promising therapeutic strategy for regenerating periodontal tissue and alveolar bone.

1. Introduction

Periodontitis, a highly prevalent condition worldwide, presents significant public health challenges [1]. The repair and regeneration of alveolar bone remain central to its treatment [2]. Periodontal ligament stem cells (PDLSCs), with their self-renewal capacity and multilineage differentiation potential, are considered a promising source for bone regeneration [3]. However, their osteogenic potential is influenced by various factors [[4], [5], [6]], which can stimulate the secretion of growth factors and signaling molecules that promote osteogenesis [7]. Therefore, elucidating the molecular mechanisms underlying the osteogenic differentiation of PDLSCs is essential.

Non-coding RNAs (ncRNAs), although not encoding proteins, play crucial roles in a wide range of biological processes [8]. They have been identified as key regulators of cellular signaling pathways and multiple diseases, including periodontitis, osteoarthritis, and osteoporosis [9,10]. Long non-coding RNA growth arrest specific 5 (GAS5), a member of the ncRNA family, plays a complex and enssential role in numerous cell biological processes, particularly in osteogenic differentiation and bone-related diseases [11].

Studies have demonstrated a close association between GAS5 and osteogenic differentiation. GAS5 expression is closely linked with the osteogenic capacity of bone marrow mesenchymal stem cells (BMSCs) [[11], [12], [13]]. In patients with osteonecrosis of the femoral head (ONFH), GAS5 expression was significantly downregulated in subchondral bone tissue. Conversely, during the osteogenic induction of BMSCs, GAS5 expression was upregulated, which enhanced alkaline phosphatase (ALP) activity [11]. These findings underscore the regulatory function of GAS5 in osteogenic differentiation. Moreover, GAS5 promoted osteogenic differentiation in BMSCs via the miR-135a-5p/FOXO1 pathway [12]. In osteoporosis models, GAS5 modulates osteogenic balance by participating in the ROS/SIRT1 feedback regulatory circuit [13]. GAS5 has also been applied to enhance bone healing. For instance, in rotator cuff injury models, GAS5 delivered through liposome nanoparticles improved bone quality and promoted tendon-bone healing [14]. Notably, compared with PDLSCs under normal conditions, 1220 and 1564 lncRNAs are significantly upregulated and downregulated, respectively, in inflamed PDLSCs, among which GAS5 shows a pronounced alteration in expression [15]. Given the prominent role of GAS5 in osteogenic differentiation and bone-related diseases, elucidating its regulatory mechanism in PDLSCs may provide new insight in therapeutic strategies for periodontitis and alveolar bone regeneration.

MicroRNAs (miRNAs), another major class of non-coding RNAs, have been shown to promote the progression of periodontitis by negatively regulating target genes [16]. During stem cells differentiation into osteoblasts, miRNAs acted as post-transcriptional regulators, influencing lineage commitment by modulating the osteogenic gene expression [17]. Among them, miR-23b-3p was a multifunctional regulatory factor involved in biological development and is recognized as an important biomarker for the diagnosing and treatment of osteoporosis [18,19]. It is well documented that miR-23b-3p directly regulated osteogenesis in stem cells. Under inflammatory conditions, it modulated the differentiation of PDLSCs by targeting the transcription factor RUNX2 [20], while also enhancing the osteogenic potential of MC3T3-E1 cells to support bone formation [21]. Additionally, miR-23b-3p molecule targeted the MRC2 gene and its upregulation can inhibit the Wnt signaling pathway, thereby suppressing osteogenic differentiation and promoting postmenopausal osteoporosis [22]. Despite its well-established importance in stem cell function, the precise mechanisms by which miR-23b-3p functions remain unclear. Based on a combination of literature review and bioinformatics predictions, we hypothesize that miR-23b-3p may interact with lncRNA GAS5. However, whether GAS5 regulates miR-23b-3p through a sponge mechanism to influencing osteogenic differentiation remains unknown.

Signal transducer and activator of transcription 5B (STAT5B), a pivotal component of the JAK2/STAT5B signaling pathway, played essential regulatory roles in osteogenic differentiation, cell proliferation, apoptosis, and immune modulation [23]. STAT5B was particularly critical for regulating the expression of insulin-like growth factor I (IGF-I), a major mediator of intracellular development and extracellular matrix expansion, and served as a central hub within the growth hormone (GH)-STAT5B-IGF-I transcriptional axis [24]. The regulatory function of STAT5B was significantly influenced by interactions with microRNAs, such as miR-134, which suppressed neoplastic growth and stem cell renewal by simultaneously targeting the KRAS and STAT5B signaling pathways [25]. However, the role of miRNA-mediated regulation of STAT5B in directing stem cell osteogenic differentiation remains insufficiently explored.

This study investigated the role of LncRNA GAS5 in regulating STAT5B expression through its interaction with miR-23b-3p and examines how this regulation impacts the osteogenic differentiation of PDLSCs. Furthermore, by elucidating the relationship between this gene regulatory network and alveolar bone regeneration, this study aimed to establish a theoretical foundation for developing targeted therapeutic strategies for periodontitis based on non-coding RNA regulation.

2. Materials and methods

2.1. Cell culture

This study protocol was approved by the Ethics Committee of Xiamen Medical College Stomatological Hospital and conducted in accordance with the Declaration of Helsinki (approval number: HS20211208002). Informed consent was obtained prior to tooth collection. PDLSCs were isolated from the periodontal ligament tissues of healthy premolars extracted for orthodontic reasons.

One-third of the root surface of each extracted tooth was scraped, and the collected tissue was washed 6-8 times with phosphate-buffered saline (PBS) containing with 100 U/mL penicillin and 100 mg/mL streptomycin. The tissue was then minced into 1 mm² pieces using tissue shears and transferred to a 1.5 mL centrifuge tube. 2 mg/mL type I collagenase (2344440, Gibco, USA) was added, and mixture was digested at 37 °C in a 5% CO₂ for 30 min. The digestion was terminated by adding three volumes of complete medium. The digested suspension was centrifuged at 1000 rpm for 3 min, and the supernatant was discarded. The resulting pellet was resuspended in α-MEM supplemented with 20% fetal bovine serum (FBS) (A5670701, Gibco, USA), 100 U/mL of penicillin and 100 mg/mL of streptomycin, and cultured in a 25 cm² flask at 37 °C with 5% CO₂. When the cells reached 80% to 90% confluence, they were passaged using 0.25% trypsin. Cells at passages 3-5 (P3–P5) were used for subsequent experiments.

Osteogenic induction was initiated when the cells reached 70-80% confluence using osteogenic differentiation media (OM) containing 10% FBS (Gibco, USA), 10 mmol/L β-glycerophosphate (Sigma, USA), 50 μg/mL ascorbic acid (Sigma, USA), and 100 nmol/L dexamethasone (Sigma, USA).

2.2. Flow cytometry analysis

Flow cytometry was performed to characterize PDLSCs surface markers using the QB Human MSC Analysis Kit (QuantoBio, Beijing, China). According to the manufacturer's protocol, cells were incubated with antibodies against CD73, CD105, CD45, and HLA at 4 °C in the dark for 30 min. After rinsing with PBS, the samples were analyzed using a FACSCalibur flow cytometer (BD Biosciences, USA).

2.3. Bioinformatics analysis

ENCORI (rnasysu.com/encori/), DIANA-LncBase (microrna.gr/LncBase), and Rnahybrid (bibiserv.cebitec.uni-bielefeld.de/rnahybrid) were used to predict miRNAs potentially interacting with GAS5. Venn diagrams were constructed to identify intersecting miRNAs. To predict miR-23b-3p target genes, ENCORI (rnasysu.com/encori/), miRDB (mirdb.org/index.html), and GeneCard databases (genecards.org) were used. The intersecting target genes were further analyzed using the STRING database to construct the protein–protein interaction (PPI) network.

2.4. Fluorescence in situ hybridization (FISH)

The GAS5 probe was synthesized by Zolgene Biotechnology (China). FISH was performed using the Zolgene in situ hybridization detection kit. Briefly, PDLSCs slides were fixed with 4% paraformaldehyde, washed with distilled water, digested with 3% citrate-pepsin mixture at 37 °C for 3-10 min, permeabilized with 0.1% Triton X-100, and prehybridized at 42 °C for 2 h. A 0.5 μM probe hybridization solution was then added, and hybridization was carried out overnight at 45 °C. Slides were washed, counterstained with DAPI, and mounted. Images were aquired using a pathological slide scanning system.

2.5. Cell transfection

To examine the regulatory function of lncRNA GAS5 in PDLSCs, a GAS5-overexpressing lentiviral vector (LV5-GAS5) and its negative control (LV5-NC) were were constructed (GenePharma, Shanghai, China). Transduction was performed using polybrene (GenePharma) and efficiency was verified by quantitative real-time polymerase chain reaction (qRT-PCR) and green fluorescent protein (GFP) fluorescence microscopy.

For microRNA (miRNA) manipulation, miR-23b-3p mimic, mimic negative control (mimic-NC), miR-23b-3p inhibitor, and inhibitor negative control (inhibitor-NC) were purchased from GenePharma (Shanghai, China). Transfection were performed with Lipofectamine 3000 (Invitrogen, USA) following the standard instructions.

To silence STAT5B, small interfering RNA (siRNA) sequences targeting STAT5B (si-STAT5B) and non-targeting control siRNA (si-NC) were transfected using Lipofectamine 3000. Silencing efficiency was calidarted by qRT-PCR.

2.6. Dual-luciferase reporter assays

Luciferase plasmids containing wild-type (WT) or mutant (MUT) GAS5 and STAT5B were constructed. PDLSCs were co-transfected with plasmid and miR-23b-3p mimic or mimics NC. Luciferase activity was measured 48 h post-transfection using the Dual-Glo Luciferase Assay System (E1910, Promega, Inc.).

2.7. qRT-PCR

Total RNA was extracted using TRIzol reagents (Invitrogen, USA). cDNA synthesis for mRNA was performed using the PrimeScript RT Reagent Kit (TaKaRa BIO INC.) and qPCR was carried out using the TB Green® Premix Ex Taq™ II kit (TaKaRa BIO INC.). miRNA reverse transcription and qPCR were performed using Vazyme kits and the miRNA Universal SYBR qPCR Master Mix (Vazyme Biotech Co., Ltd.), respectively. GAPDH and U6 served as the internal references. Relative expression was calculated using the 2^−ΔΔCt method. The experiments were repeated three times. Primer sequences are presented in Table 1.

Table 1.

Sequencing of primers, mimics and inhibitors.

Gene	Forward primer sequences5′-3′	Reverse primer sequences5′-3′
ALP	CTGGGAAATCTGTGGGCAT	GGGGCATCTCGTTGTCTG
RUNX2	ACGGGGACCTACAGACAGC	AGACACCAAACTCCACAGCC
OPN	ACAGACCCTTCCAAGTAAGTCC	GTCATCTACATCATCAGAGTCGTT
OCN	AGCCCTCACACTCCTCGC	TCACTACCTCGCTGCCCTC
STAT5B	GCCACCATCACGGACATTAT	TTCGGGAGGAGCTGGGAC
GAPDH	CAAAAGGGTCATCATCTCTGC	TGATCTTGAGGCTGTTGTCATAC
miR-23b-3p	CGATCACATTGCCAGGGAT	AGTGCAGGGTCCGAGGTATT
U6	CTCGCTTCGGCAGCACA	AACGCTTCACGAATTTGCGT
si-GAS5	GAUGAGAAUAGCUACUGAATT	UUCAGUAGCUAUUCUCAUCTT
23b mimics	AUCACAUUGCCAGGGAUUACCAC	GGUAAUCCCUGGCAAUGUGAUUU
23b inhibitor	GUGGUAAUCCCUGGCAAUGUGAU

2.8. Western blot

Total protein was extracted using RIPA buffer containing 1% phenylmethylsulfonyl fluoride (Epizyme Biomedical). After centrifugation, proteins concentration was determined by a BCA protein assay kit (Beyotime, Haimen, China). Proteins were separated by SDS–PAGE, and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA), blocked for 20 min and incubated overnight at 4 °C with the with the following antibodies: anti-RUNX2 (Abcam, ab76956, 1:1000), anti-osteocalcin (OCN) (Abcam, ab133612, 1:1000), anti-osteopontin (OPN) (Proteintech, Cat.#22952-1-AP, 1:1000), and anti-STAT5B antibodies (Proteintech, Cat.#66427, 1:5000). Following washing with TBST, the membranes were incubated with the species-matched secondary antibody at 25 °C for 1 h. The visualization of proteins was achieved using an ECL luminescence reagent (Millipore, Billerica, MA, USA). Subsequently, the gray values corresponding to the bands were quantified using ImageJ software.

2.9. ALP staining

ALP staining was performed using the BCIP/NBT Alkaline Phosphatase Color Development Kit (Biotechnology, Haimen, China). Cells were fixed in 4% paraformaldehyde (PFA), washed, incubated in staining buffer for 1 h at 25 °C, rinsed, and observed microscopically.

2.10. Alizarin red staining

After 21 days of osteogenic induction, mineralized nodules were stained using the Alizarin Red S Staining Kit (Biotechnology, Haimen, China). To quantify mineralization, nodules were eluted with 1% cetylpyridinium chloride for 1 h, and absorbance was measured at 562 nm using a microplate reader.

2.11. RNA-pull down

Biotin-labeled miR-23b-3p probes and negative control probes were incubated with streptavidin beads for 2 h. PDLSCs (2 × 10⁷) were lysed, and the lysates were incubated with bead-probe complexes overnight. Beads were washed, and bound RNA was extracted using TRIzol. GAS5 enrichment was measured by qRT-PCR.

2.12. Statistical analysis

All experiments were independently repeated at least three times (n = 3), with technical triplicates for each biological sample to ensure data reproducibility. Statistical analyses were performed using GraphPad Prism 8.0 software. Between-group comparisons were conducted using Student's unpaired t-test and multiple group comparisons used one-way analysis of variance (ANOVA) with Tukey's post-hoc test. P < 0.05 was considered statistically significant.

3. Results

3.1. Isolation and identification of PDLSCs

In vitro culture observations revealed that after 7 days of incubation, spindle-shaped cells migrated from the edges of the periodontal tissue explants and adhered to the bottom of the culture flask (Fig. 1A). Subculture was performed when the cell density reached approximately 80-90% confluence. Immunophenotypic characterization via flow cytometry confirmed that the isolated cells exhibited typical mesenchymal stem cell (MSC) surface marker profiles, with high expression of CD73 (95.58%) and CD105 (94.07%), and minimal expression of hematopoietic lineage markers CD45 (0.74%) and HLA (5.61%) (Fig. 1C).

Fig. 1.

Isolation, culture, and characterization identification of PDLSCs. (A) Primary PDLSCs were extracted from the periodontal ligament tissue and cultured in standard culture medium. Scale bar, 200 μm. (B) Osteogenic differentiation potential of PDLSCs was assessed by alkaline phosphatase (ALP) staining (7 days post-induction) and Alizarin Red S staining (21 days post-induction) to evaluate early and late osteogenic phenotypes, respectively. Scale bar, 250 μm. (C) Flow cytometry analysis demonstrated that PDLSCs expressed positive markers CD73 and CD105, while CD45 and HLA were negative, indicating that the cultured PDLSCs exhibited mesenchymal stem cell-like characteristics. (D) After 7 and 14 days of osteogenic induction, mRNA levels of ALP, OPN, OCN, and RUNX2 were significantly increased, as determined by qRT-PCR analysis, indicating that PDLSCs possess the ability to undergo osteogenic differentiation. (E) The protein levels of RUNX2 and OCN in PDLSCs were analyzed using Western blotting after 7-days of osteogenic induction with an osteogenic differentiation medium. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

For osteogenic differentiation assays, ALP activity was significantly upregulated after 7 days of osteogenic induction compared with the uninduced control group. At 21 days post-induction, Alizarin Red S staining revealed substantially increased mineralized matrix deposition relative to the control group (Fig. 1A). Molecular analysis further demonstrated osteogenesis-related gene expression changes: qRT-PCR results showed that mRNA levels of ALP, OPN, OCN, and RUNX2 were significantly increased in induced cells (P < 0.05, Fig. 1D). Consistent with the transcriptional data, Western blotting analysis further confirmed that protein expression levels of OCN and RUNX2 were also significantly elevated in the osteogenically induced group (Fig. 1E).

Collectively, these findings confirm the successful isolation of MSCs from periodontal tissues and validates their osteogenic differentiation potential.

3.2. Subcellular localization and osteogenic regulatory role of lncRNA GAS5 in PDLSCs

RNA-FISH was performed to determine the subcellular localization of lncRNA GAS5 in PDLSCs. 18S ribosomal RNA (rRNA) served as a cytoplasmic reference. The results showed that GAS5 was predominantly localized in the cytoplasm (Fig. 2A).

Fig. 2.

Subcellular localization of lncRNA GAS5 in PDLSCs and its regulatory role in osteogenic differentiation. (A) FISH assay was performed to determine the subcellular localization of lncRNA GAS5 in PDLSCs; representative images show that GAS5 is predominantly localized in the cytoplasm. Scale bar, 20 μm. (B) The relative expression levels of GAS5 in PDLSCs during osteogenic differentiation were measured at various induction times using qRT-PCR. (C) Fluorescence microscopy was used to observe the green fluorescent protein (GFP) signal in PDLSCs transfected with GAS5 overexpression lentivirus. Scale bar: 100 μm. (D) Semi-quantitative analysis of mineralization was performed via Alizarin Red S staining quantification (corresponding to panel F). (E) ALP staining was conducted to evaluate ALP activity in the transfected PDLSCs. Scale bar: 500 μm. (F) Alizarin Red S staining was used to assess mineralized nodule formation in transfected PDLSCs, with quantification shown in panel (D). Scale bar: 250 μm. (G) qRT-PCR validation of GAS5 expression levels in PDLSCs after lentivirus-mediated overexpression or knockdown. (H, I) qRT-PCR was employed to quantify the mRNA expression levels of osteogenic markers in the transfected PDLSCs. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

To elucidate the potential regulatory role of GAS5 in PDLSC osteogenesis, we first examined its expression profile during a 14-day osteogenic induction period. qRT-PCR analysis showed that GAS5 transcript levels gradually and significantly increased throughout the induction process (Fig. 2B). To further investigate its function, we established gain- and loss-of-function models: GAS5 overexpression was achieved via lentiviral transduction, while GAS5 knockdown was mediated by siRNA transfection. Successful transfection was verified by qRT-PCR and GFP fluorescence (Fig. 2C and G).

GAS5 overexpression significantly enhanced the transcription of key osteogenic markers, including ALP, OCN, OPN, and RUNX2, whereas GAS5 knockdown markedly reduced their expression (Fig. 2H and I). Morphologically, ALP staining revealed increased ALP enzymatic activity in GAS5-overexpressing PDLSCs, and quantitative Alizarin Red S (ARS) staining demonstrated significantly increased mineral deposition in the LV5-GAS5 group (P < 0.0001). Conversely, GAS5 knockdown substantially reduced matrix mineralization (Fig. 2D–F).

Together, these results confirm that GAS5 functions as a positive regulator of PDLSC osteogenic differentiation.

3.3. LncRNA GAS5 functions as a sponge for miR-23b-3p

To explore whether GAS5 regulates osteogenesis through a competitive endogenous RNA (ceRNA) mechanism, we used three bioinformatics tools - ENCORI, LncBase, and RNAhybrid - to predict potential miRNAs targeting GAS5. Venn diagram analysis identified five candidate miRNAs: hsa-miR-23b-3p, hsa-miR-449c-5p, hsa-miR-205-5p, hsa-miR-188-5p, and hsa-miR-485-5p (Fig. 3A). Preliminary qRT-PCR screening showed that only miR-23b-3p was significantly downregulated during osteogenic induction (Supplementary Fig. S1). Given its distinct expression pattern and reported role in osteogenesis [20], miR-23b-3p was selected for further investigation.

Fig. 3.

LncRNA GAS5 acts as an endogenous sponge by directly targeting miR-23b-3p in PDLSCs. (A) Venn diagram illustrating overlapping miRNAs predicted to bind the lncRNA GAS5 by three bioinformatics tools (ENCORI, LncBase, and RNAhybrid). (B) Temporal expression profile of miR-23b-3p during osteogenic induction of PDLSCs (0, 3, 7, and 14 days) detected by qRT-PCR. (C) Schematic diagram of the conserved binding motif between GAS5 and miR-23b-3p predicted by bioinformatics analysis; complementary nucleotide sequences are highlighted in red. (D) The effects of co-transfecting GAS5 wild type (WT) or GAS5 mutant (MUT) with miR-23b-3p mimics or a negative control into PDLSCs were analyzed using a luciferase reporter assay. (E) The expression levels of miR-23b-3p in PDLSCs overexpressing GAS5 were analyzed using qRT-PCR. (F) RNA pull-down assay validating the direct interaction between GAS5 and miR-23b-3p. Biotin-labeled miR-23b-3p (miR-23b-3p-Bio) or scrambled control (NC-Bio) was used to pull down RNA complexes; qRT-PCR showed significantly higher GAS5 enrichment in the miR-23b-3p-Bio group.∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

Bioinformatics analysis predicted a conserved miR-23b-3p binding site within the GAS5 sequence (Fig. 3C). To experimentally validate this interaction, dual-luciferase reporter assays were performed. Dual-luciferase assays demonstrated that miR-23b-3p overexpression significantly reduced luciferase activity in the GAS5-WT reporter but had no effect on the GAS5-MUT construct (Fig. 3D), confirming sequence-specific binding.

An RNA pull-down assay further validated the direct interaction: biotin-labeled miR-23b-3p significantly enriched endogenous GAS5 transcripts compared with the negative control probe (Fig. 3F). Additionally, miR-23b-3p expression progressively decreased during osteogenic induction (Fig. 3B) and was significantly reduced in GAS5-overexpressing cells (Fig. 3E).

Collectively, these data establish that GAS5 directly binds to miR-23b-3p and acts as a ceRNA in PDLSCs.

3.4. GAS5 negatively modulates miR-23b-3p during the process of osteogenic induction

To elucidate the regulatory mechanism underlying the GAS5/miR-23b-3p interaction in PDLSC osteogenesis, we conducted a series of experimental validations. First, we characterized the functional role of miR-23b-3p through gain- and loss-of-function assays. PDLSCs were transfected with miR-23b-3p mimics to induce overexpression or with inhibitors to suppress its activity before initiating osteogenic induction, followed by assessment using qRT-PCR and alizarin red staining quantification. Inhibition of miR-23b-3p significantly promoted the formation of the mineralized matrix, whereas its overexpression had the opposite effect, markedly reducing mineralized matrix formation (Fig. 4A and B). Consistently, the expression levels of osteogenic differentiation markers, as determined by qRT-PCR, corresponded with the results from ARS staining (Fig. 4D).

Fig. 4.

The influence of miR-23b-3p on osteogenic differentiation. (A) Alizarin red staining of PDLSCs treated with mimics and inhibitors. Scale bar: 250 μm. (B) Quantitative analysis of calcified nodules after alizarin red staining. (C, D) qRT-PCR was employed to assess the expression of osteogenic induction-related genes in PDLSCs following various treatments. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

Next, to investigate the mechanistic interplay, we established a dual overexpression model by co-transfecting GAS5-overexpressing PDLSCs with miR-23b-3p mimics. Quantitative analysis of osteogenic markers revealed that the upregulation of miR-23b-3p significantly abrogated the GAS5-induced increase in ALP, OPN, OCN and RUNX2 expression (Fig. 4C) (P < 0.05). This reciprocal regulation supports a ceRNA mechanism, in which GAS5 functions as a molecular sponge for miR-23b-3p, thereby alleviating the repression of downstream targets essential for osteogenic differentiation.

3.5. STAT5B as a target of miR-23b-3p in PDLSC osteogenic differentiation

To identify downstream targets of miR-23b-3p, we screened ENCORI, miRDB, and GeneCards databases and identified 79 intersecting candidate genes (Fig. 5A). STRING-based PPI network analysis (Fig. 5B) revealed several hub genes, among which six were selected for expression validation. qRT-PCR indicated that STAT5B was the most significantly downregulated gene following miR-23b-3p overexpression (Fig. 5C). Notably, it has been reported that STAT5B is closely related to cell proliferation and that the STAT family can modulate the expression levels of target genes, leading to chondrocyte proliferation, which in turn stimulates bone development and promotes growth [[26], [27], [28]]. Collectively, These findings suggest that STAT5B may function as a downstream target gene of miR-23b-3p. Therefore, we further investigated the direct interaction and functional relationship between miR-23b-3p and the STAT5B gene.

Fig. 5.

Screening of downstream target genes of miR-23b-3p related to the osteogenic differentiation of PDLSCs. (A) Venn diagram illustrating the intersection of target genes for miR-23b-3p, as predicted and identified by the ENCORI, TargetScan, and GeneCards databases. (B) PPI network of the target genes. (C) qRT-PCR analysis of 6 candidate target genes of miR-23b-3p. (D) Effect of miR-23b-3p on STAT5B expression in PDLSCs. (E) Predicted binding sites of miR-23b-3p within the targeting STAT5B. (F) Dual-luciferase reporter assay results for STAT5B-WT and STAT5B-MUT constructs. (G) Western blot analysis of STAT5B expression following GAS5 overexpression. (H) Analysis of the correlation between STAT5B and GAS5 expression levels in periodontal tissue. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

To validate this putative targeting relationship, we first predicted the potential binding sequences of miR-23b-3p within the STAT5B transcripts using the ENCORI database. Both wild-type and mutant forms of STAT5B were subcloned into the pmirGLO plasmid, and PDLSCs were co-transfected with miR-23b-3p mimics alongside either the STAT5B-WT or STAT5B-MUT constructs. Dual-luciferase reporter assay revealed a significant reduction in luciferase activity in cells co-transfected with miR-23b-3p mimics with the wild-type STAT5B plasmid, whereas no discernible change was observed in the mutant group. These results confirmed that miR-23b-3p directly binds to the 3′-UTR of STAT5B (Fig. 5E and F). We further evaluated STAT5B expression patterns under conditions of miR-23b-3p inhibition or overexpression in PDLSCs. Our findings revealed that inhibiting miR-23b-3p expression led to an increase in STAT5B levels, while overexpressing miR-23b-3p resulted in a decrease in STAT5B levels (Fig. 5D). Additionally, investigated the effect of lncRNA GAS5 overexpression on STAT5B protein expression during PDLSC osteogenic differentiation and found that GAS5 overexpression significantly increased STAT5B protein abundance (Fig. 5G). Finally, correlation analysis of periodontitis tissue samples revealed a strong positive correlation between the expression levels of GAS5 and STAT5B (r = 0.8272, P < 0.0001; Fig. 5H).

3.6. The GAS5/miR-23b-3p/STAT5B axis regulates osteogenic differentiation of PDLSCs

Given the reciprocal regulation among GAS5, miR-23b-3p, and STAT5B, we hypothesized that GAS5 enhances PDLSC osteogenesis by regulating STAT5B through miR-23b-3p. To confirm STAT5B's functional role in PDLSCs, we transfected cells with si-STAT5B to knock down STAT5B; qPCR analysis demonstrated that STAT5B silencing downregulated the osteogenic marker genes ALP, OPN, OCN, and RUNX2 (Fig. 6F). Similarly, GAS5 knockdown led to a reduction in STAT5B expression (Fig. 6E).

Fig. 6.

The lncRNA GAS5/miR-23b-3p/STAT5B axis regulates osteogenic differentiation in PDLSCs. (A-D) Co-transfection of lncRNA GAS5 siRNA and miR-23b-3p inhibitor into human PDLSC cells reversed the upregulation of STAT5B, OPN, OCN and RUNX2 protein expression induced by the miR-23b-3p inhibitor alone, as assessed by Western blotting. (E) Knockdown of lncRNA GAS5 significantly reduced STAT5B mRNA expression, quantified by qPCR. (F) STAT5B silencing downregulated the expression of osteogenic marker genes (ALP, OPN, OCN, RUNX2), as determined by qPCR. (G) Semi-quantitative analysis of alizarin red staining results. (H) Alizarin red staining was performed to assess mineralized nodule formation across different treatment groups. Scale bar: 250 μm. (I) Western blot analysis of PDLSCs co-transfected with GAS5 overexpression constructs, si-STAT5B, and respective negative controls; STAT5B silencing attenuated the GAS5-mediated upregulation of osteogenic proteins (OCN, OPN, RUNX2). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

To delineate the functional role of the GAS5/miR-23b-3p/STAT5B axis in osteogenic differentiation, we conducted rescue experiments. Inhibition of miR-23b-3p alone significantly increased STAT5B, OPN, OCN, and RUNX2 protein levels, whereas co-transfection with si-GAS5 abolished these effects (Fig. 6A-D). Conversely, silencing STAT5B attenuated the GAS5-induced upregulation of osteogenic proteins in GAS5-overexpressing PDLSCs (Fig. 6I). Consistent with these findings, quantification of ARS staining further validated this regulatory axis (Fig. 6G and H).

Collectively, these data support our hypothesis that GAS5 acts as a ceRNA to sequester miR-23b-3p, thereby alleviating miR-23b-3p-mediated repression of STAT5B and promoting osteogenic differentiation in human PDLSCs.

4. Discussion

In this study, we elucidated a novel regulatory pathway involving lncRNA GAS5, miR-23b-3p, and STAT5B in the osteogenic differentiation of PDLSCs, which are essential for periodontal tissue regeneration. Our findings demonstrate that during the osteogenic differentiation of PDLSCs, the expression of miR-23b-3p is significantly reduced, whereas the expression of lncRNA GAS5 and STAT5B is upregulated. Importantly, we identified lncRNA GAS5 as a negative regulator of miR-23b-3p, which in turn negatively regulates STAT5B expression. This regulatory network indicates that GAS5 modulates STAT5B expression through its interaction with miR-23b-3p, thereby influencing the osteogenic differentiation potential of PDLSCs. These findings not only enhance our understanding of the molecular mechanisms underlying PDLSCs differentiation but also highlight potential therapeutic strategies for promoting periodontal tissue regeneration.

Accumulating evidence has established that lncRNAs and miRNAs play crucial regulatory roles in bone development and various bone-related pathological conditions [29,30]. Specifically, lncRNAs can function as ceRNAs by sequestering miRNAs through shared microRNA response elements (MREs), thus forming regulatory networks that modulate the osteogenic differentiation of MSCs and the expression of downstream targets [31]. This ceRNA-mediated crosstalk has been implicated in the pathogenesis of periodontitis and other bone metabolic diseases [32]. Consistent with these findings, previous studies have shown that GAS5 functions as a ceRNA in osteogenic regulation [33]. In line with this, our RNA-FISH results revealed that lncRNA GAS5 is predominantly localized in the cytoplasm of PDLSCs, - a distribution necessary for ceRNA activity, as miRNA-mRNA interactions occur primarily in the cytoplasm. Functional experiments further confirmed that GAS5 overexpression significantly enhances PDLSC osteogenic differentiation, whereas GAS5 knockdown suppresses this capacity.

To validate the direct interaction between GAS5 and miR-23b-3p, we performed dual-luciferase reporter assays and RNA pulldown assays. Co-transfection of miR-23b-3p mimic with a wild-type GAS5 reporter significantly reduced in luciferase activity, while mutations in the miR-23b-3p complementary seed region of GAS5 abolished this inhibitory effect. These results, consistent with bioinformatics predictions, confirm the specific sponge-like activity of GAS5 toward miR-23b-3p. According to the ceRNA hypothesis, RNAs containing the same microRNA response element (MRE) can regulate post-transcriptional gene expression by competing for miRNA binding [34]. In this study, GAS5 functions as a molecular sponge for miR-23b-3p, thereby alleviating its inhibitory effect on downstream targets genes and promoting osteogenic differentiation. Notably, miR-23b-3p has previously been shown to suppress key osteogenic transcription factors such as RUNX2 [35]. Consistent with this, our results demonstrated that GAS5 overexpression significantly increases RUNX2 protein expression, further supporting the ceRNA-mediated regulatory role of GAS5 in PDLSC osteogenesis.

STAT5B is a critical component of the Janus kinase (JAK)-STAT signaling pathway. In the present study, we identified STAT5B as a direct target gene of miR-23b-3p. Binding sites between miR-23b-3p and the 3′-UTR of STAT5B were initially predicted using bioinformatics tools such as TargetScan [36] and miRDB [37], and subsequently validated by dual-luciferase reporter assay. Overexpression of miR-23b-3p significantly reduced luciferase activity in cells transfected with the wild-type STAT5B 3′-UTR construct but not in the mutant construct, confirming direct targeting. Together, these findings establish a mechanistic link among GAS5, miR-23b-3p, and STAT5B in regulating PDLSC osteogenic differentiation. Functionally, STAT5B regulates osteogenesis through its role in the JAK-STAT pathway. Activation of STAT5B enhances the JAK2/STAT5B cascade downstream of growth hormone (GH), thereby promoting the expression of osteogenic markers - including ALP, OCN, and RUNX2 [38]. Conversely, STAT5B knockdown increases adipogenic markers such as PPARγ and C/EBPα, indicating a critical role in guiding progenitor cells toward osteoblast rather than adipocyte lineages [39]. These observations align with previous studies demonstrating the involvement of STAT5B in bone homeostasis and osteoblast differentiation [40].

Functional validation of the GAS5/miR-23b-3p/STAT5B axis further confirmed its essential role in osteogenesis. Rescue experiments showed that GAS5 knockdown reversed the STAT5B upregulation induced by miR-23b-3p inhibiton. Similarly, osteogenic markers such as OPN, OCN, and RUNX2 exhibited parallel trends, with significant increases following miR-23b-3p inhibiton. Mechanistically, GAS5 acts as a ceRNA that sequesters miR-23b-3p, thereby relieving the translational repression of STAT5B. This regulatory model is consistent with the established ceRNA paradigm, in which lncRNAs fine-tune miRNA-target interactions to modulate signaling outputs [41]. Furthermore, potential feedback regulation within this axis warrants further investigation. For example, STAT5B may regulate GAS5 transcription, while miR-23b-3p may target STAT5B under specific conditions, forming a dynamic feedback loop. Similar regulatory loops have been reported in STAT3-mediated regulatory networks [42], supporting the viability of the feasibility of this hypothesis.

Although our study identifies the GAS5/miR-23b-3p/STAT5B axis as a key regulator of PDLSCs osteogenic differentiation, several mechanistic aspects remain to be clarified. Notably, STAT5B-dependent osteogenesis require further investigation. Additionally, our mechanistic insights are limited to in vitro systems, which cannot fully recapitulate the complex in vivo periodontal microenvironment. To address these limitations, we plan to establish a mouse model of periodontal defects to validate the regulatory role of GAS5 in vivo and to evaluate the therapeutic potential of targeting this ceRNA axis for periodontal regeneration. Future studies will also explore the involvement of this axis in other skeletal disorders-such as osteoporosis and fracture healing-and determine its tissue-specific therapeutic relevance using preclinical models, including rat calvarial defect models. These efforts will deepen our understanding of ceRNA-mediated osteogenic regulation and support the development of targeted therapies for skeletal diseases.

In summary, our study demonstrates that GAS5 competitively binds to miR-23b-3p, acting as a ceRNA. This interaction alleviates the translational repression of STAT5B by miR-23b-3p, thereby enhancing STAT5B expression. The upregulation of STAT5B subsequently promotes the expression of key osteogenic genes, including RUNX2, ALP, OPN, and OCN, ultimately facilitating the osteogenic differentiation of PDLSCs (Fig. 7).

Fig. 7.

Schematic diagram illustrating the potential mechanism by which the GAS5/miR-23b-3p/STAT5B axis regulates the osteogenic differentiation of PDLSCs.

Ethical approval

Periodontal ligament tissue and periodontitis samples were collected from the Affiliated Stomatological Hospital of Xiamen Medical College. The research protocol received approval from the Medical Ethics Committee of the same institution (approval number: HS20211208002).

Funding

This work was supported by the Xiamen Medical and Health Guidance Project (No. 3502Z20214ZD1268), the Xiamen Medical and Health Guidance Project (No. 3502720224ZD1333to Y.H.), and the National Natural Science Foundation of China (No. 82001021).

CRediT authorship contribution statement

Qiufang Zhang: Data curation, Formal analysis, Investigation, Methodology, Writing – original draft. Yan Huang: Formal analysis, Methodology, Project administration. Liping Huang: Data curation, Investigation, Software, Validation. Zhisheng Zhang: Project administration, Resources, Supervision. Qiliang Zuo: Funding acquisition, Supervision, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Data availability

No datasets were generated or analyzed during the current study.