Osteogenic differentiation of gingival and periodontal ligament fibroblasts inhibits osteoclast formation in 2D and 3D fibrin cultures

Abstract

Bone regeneration requires coordinated interactions between multiple cell types responding to biochemical and biomechanical cues from the extracellular matrix (ECM). Gingival fibroblasts (GFs) and periodontal ligament fibroblasts (PDLFs) contribute to the regeneration of lost alveolar bone, as they can differentiate into osteoblast-like cells and regulate osteoclastogenesis. However, it remains unclear whether these two fibroblast populations contribute similarly to bone regeneration and how ECM properties modulate their osteogenic differentiation and subsequent osteoclast-inducing capacity. Here, we compared the osteogenic potential of GFs and PDLFs from 12 donors, cultured in 2D monolayers or 3D fibrin hydrogels under normal or osteogenic conditions, and assessed how fibroblast-conditioned media regulates osteoclast formation from peripheral blood mononuclear cells (PBMCs). Both fibroblast populations underwent osteogenic differentiation, as evidenced by increased alkaline phosphatase (ALP) activity, calcium deposition, and robust mineral nodule formation in 3D fibrin hydrogels. Osteogenic stimulation reduced macrophage colony-stimulating factor (M-CSF) secretion and increased osteoprotegerin (OPG) release. Functionally, conditioned media from non-osteogenic cultures promoted formation of multinucleated TRAcP-positive osteoclasts and upregulated osteoclast-associated genes, including TRACP, RANK, Cathepsin K, and the fusion marker DC-STAMP. In contrast, osteogenic differentiation attenuated the osteoclast-inducing capacity of both fibroblast populations. Across all parameters, GFs and PDLFs displayed largely comparable osteogenic and osteoclast-modulating phenotypes, irrespective of culture dimensionality. These findings indicate that osteogenic differentiation functions as a regulatory switch that diminishes the osteoclast-inductive potential of oral fibroblasts and support the interchangeable use of GFs and PDLFs in periodontal regenerative strategies.

1. Introduction

The alveolar bone, together with cementum and periodontal ligament (PDL), forms the supporting apparatus of the teeth (Bartold et al., 2019). As a key component of this complex, alveolar bone anchors teeth and absorbs masticatory forces (Cho and Garant, 2000). In periodontitis, progressive destruction of these supporting tissues ultimately leads to tooth loss if left untreated (Grzesik and Narayanan, 2002). Moreover, the absence of functional mechanical loading following tooth loss disrupts the balance between bone formation and resorption (Zhan et al., 2025). Consequently, regenerative and tissue engineering strategies have gained interest as approaches to restore both tissue architecture and physiological function (Parisi et al., 2024).

Alveolar bone regeneration involves the coordinated activity of multiple cell types, including gingival fibroblasts (GFs) and periodontal ligament fibroblasts (PDLFs) (Alfonso Garcia et al., 2020). These fibroblasts actively interact with surrounding cells and respond dynamically to biochemical and biomechanical cues from the extracellular matrix (ECM) (Parisi et al., 2024). GFs are located within the gingival connective tissue and contribute to tissue maintenance, wound healing, and immune modulation, whereas, PDLFs reside in the PDL, the highly specialized connective tissue between the cementum and alveolar bone, and are essential for PDL and alveolar bone stability and function (Naruishi, 2022).

Periodontium-derived fibroblasts exhibit remarkable plasticity and share several characteristics with mesenchymal stem cells (MSCs), including spindle-shaped morphology, expression of markers such as CD73, CD90, CD105, and fibronectin, and multilineage differentiation potential (osteogenic, adipogenic, and chondrogenic) under inductive conditions (Parisi et al., 2024; Denu et al., 2016; Claeys et al., 2020). These shared features highlight their potential as alternative cell sources for regenerative applications. For example, Wu et al. reported successful periodontal regeneration in a canine model using autologous human GFs that were induced toward a mineralizing phenotype through stimulation with an osteogenic medium containing β-glycerophosphate, ascorbic acid, and dexamethasone (Wu et al., 2018). Consistent with these findings, we have previously observed similar time-dependent mineral deposition and osteogenic differentiation capacity in both GFs and PDLFs following osteogenic induction (Ceylan et al., 2024; Prins et al., 2024).

Besides cellular components, the ECM plays a crucial role in bone regeneration. Early healing involves formation of a provisional fibrin-rich matrix, generated through thrombin-mediated polymerization of fibrinogen, which provides structural support and biochemical cues guiding cell adhesion, migration, and differentiation (Bartold and Ivanovski, 2025; Bujoli et al., 2019). Compared with collagen, fibrin adsorbs substantially more serum fibronectin and enhances osteoblast proliferation and differentiation, as shown by increased RUNX2 expression, ALP activity, and calcium deposition (Oh et al., 2014). Moreover, fibrin is frequently applied as an osteoconductive carrier for cells and bioactive molecules in bone tissue engineering (Noori et al., 2017). Sasaki et al. generated bone-like tissues by culturing bone marrow MSCs within three-dimensional (3D) fibrin hydrogels prior to subcutaneous implantation in immunodeficient mice (Sasaki et al., 2015). These constructs showed increasing mechanical strength and progressive accumulation of calcium and osteopontin during in vitro culture, indicating active matrix mineralization. Following implantation, the engineered constructs developed into organized cortical bone-like tissue, demonstrating the capacity of 3D fibrin hydrogels to direct structured bone regeneration in vivo.

While regenerative approaches often focus on osteogenesis, successful bone regeneration also requires controlled osteoclast-mediated remodeling, and continuous communication between bone cells (Shariati et al., 2025). This requires tight coupling between osteoclast and osteoblast activity to ensure proper bone architecture and long-term stability (Shariati et al., 2025; Schindeler et al., 2008). Importantly, both GFs and PDLFs express osteoprotegerin (OPG) and the receptor activator of nuclear factor κB (RANK), the central regulators of osteoclastogenesis (de Vries et al., 2006; Sokos et al., 2015). Consistent with this, our previous findings demonstrated that osteogenic differentiation of these fibroblasts is associated with a reduced capacity to stimulate osteoclast formation (Ceylan et al., 2024; Prins et al., 2024).

Despite extensive knowledge of fibroblast biology, their relative contribution to bone regeneration remains incompletely understood. In particular, GFs represent an attractive cell source because of their abundance and ease of harvesting compared with PDLFs. However, it remains unclear whether GFs can function comparably to PDLFs in bone regeneration and how ECM dimensionality influences their osteogenic differentiation and subsequent osteoclast-inducing capacity. Therefore, the present study aimed to directly compare the osteogenic differentiation of GFs and PDLFs cultured in two-dimensional (2D) monolayers and 3D fibrin hydrogels and to assess their ability to induce osteoclast formation. In the first part of the study, osteogenic differentiation was assessed by ALP activity, calcium (Ca²⁺) deposition, matrix- and morphology-related changes using scanning electron microscopy (SEM) and confocal microscopy, gene expression profiling, and ELISA-based quantification of secreted factors. In the second part, osteoclast-inducing capacity was analyzed using quantitative PCR (qPCR), tartrate-resistant acid phosphatase (TRAcP) activity assay, and TRAcP staining. We hypothesized that culturing fibroblasts in 3D fibrin hydrogels, as a more physiologically relevant environment, promotes osteogenic differentiation of GFs to levels comparable to PDLFs and enhances the osteoclast-modulating capacity of both cell types. Notably, this study includes a relatively large donor cohort of GF and PDLF samples, which has not been previously reported in a single study.

2. Materials & methods

2.1. Cell preparation

GFs and PDLFs were isolated from twelve systemically healthy individuals who underwent third molar extraction at the Department of Oral and Maxillofacial Surgery, Amsterdam UMC, Academic Centre for Dentistry Amsterdam, University of Amsterdam and Vrije University Amsterdam, Amsterdam, the Netherlands. Written informed consent was obtained from all participants prior to participation, and the study protocol was approved by the Medical Ethics Committee of Amsterdam UMC (protocol 2016.105).

Gingival tissue was collected from coronally attached gingival sites exhibiting no clinical signs of inflammation (probing depth ≤ 3 mm, absence of bleeding on probing, and no clinical attachment loss). PDL tissue was retrieved from the middle third-root region of the extracted teeth. The collected tissue samples were rinsed twice with Gibco BRL, Paisley, Scotland) supplemented with 10% fetal calf serum (FCS; HyClone, Logan, UT, USA) and 1% PSF (100 U/mL penicillin, 100 μg/mL streptomycin, 250 ng/mL amphotericin B; Antibiotic–Antimycotic Solution, Sigma-Aldrich, St. Louis, MO, USA). Samples were cultured at 37 °C in a humidified atmosphere containing 5% CO₂. Cells at passage 4 were used for all experiments.

2.2. Fibrin gel preparation

For 2D cultures, GFs and PDLFs were seeded in 48-well tissue culture plates at a density of 3 × 10⁴ cells/well.

For 3D cultures, fibrin hydrogels were prepared following the procedure as described by de Jong et al. (de Jong et al., 2016). Human fibrinogen depleted of plasminogen, von Willebrand factor, and fibronectin (Enzyme Research Laboratories, South Bend, IN, USA) was dissolved in DMEM to obtain a 4 mg/mL stock solution. This fibrinogen solution was mixed at a 1:1 ratio with a cell suspension, yielding final concentrations of 2 mg/mL fibrinogen and 2 × 10⁵ cells/mL. Fibrin polymerization was triggered by the addition of human α-thrombin (1 IU/mL final concentration; Prolytix, Essex Junction, VT, USA).

The selected cell density for the 3D fibrin constructs was based on matching the cell number per matrix surface area to that used in 2D. A 48-well plate provides approximately 1.1 cm² of growth surface, onto which 3 × 10⁴ cells were seeded, resulting in a cell density of ~2.7 × 10⁴ cells/cm². Our fibrin hydrogels have an estimated total surface area of ~2.5 cm². Thus, approximately 6 × 10⁴ cells per gel would be the equivalent to the 2D cultures, corresponding to a final concentration of 2 × 10⁵ cells/mL in a 300 μL construct. This concentration also aligns with typical ranges (10⁵–10⁶ cells/mL) reported in fibroblast-mediated fibrin or collagen remodeling assays (Duong et al., 2009; Jansen et al., 2013). For hydrogel preparation, 300 μL of the fibrin-cell mixture was dispensed into each well and allowed to polymerize for 1 h at room temperature, followed by an additional hour in a humidified incubator at 37 °C with 5% CO₂. Medium was not added during polymerization to avoid disturbance of fibrin fibril assembly.

After polymerization, 300 μL of culture medium was added to each well for both 2D and 3D cultures. At day 0, cultures were maintained in DMEM supplemented with 10% FCS and 1% PSF (Just Medium; JM). For all subsequent time points, the medium was additionally supplemented with 50 μg/mL ascorbic acid (Sigma-Aldrich) and is referred to as normal medium (NM). All culture media were refreshed twice a week throughout the experiment.

2.3. Osteogenesis

To induce osteogenic differentiation, GFs and PDLFs cultured under both 2D and 3D conditions were maintained for 3 weeks in osteogenic medium (OM) consisting of DMEM supplemented with 10 nM β-glycerophosphate (Sigma-Aldrich, St. Louis, MO, USA) and 50 μg/mL ascorbic acid (Sigma-Aldrich).

After the 3-week culture period, differentiation was evaluated using ALP activity, DNA content, and Ca²⁺ quantification. In addition, morphological and matrix-related changes were examined by SEM and confocal microscopy. Gene expression associated with osteogenic differentiation was analyzed by qPCR, and secreted factors from GFs and PDLFs were analyzed by ELISA to evaluate their osteoclast-inducing capacity.

2.3.1. ALP activity

ALP activity was evaluated exclusively in 2D cultures. GFs and PDLFs were collected at days 1 and 14, lysed in 200 μL of water, and stored at −20 °C until analysis. ALP activity was quantified using 4-nitrophenyl phosphate disodium salt (Merck, Darmstadt, Germany) as the ALP substrate at pH 10.3, following the method described by Lowry et al. (Lowry, 1957). Absorbance was measured at 405 nm using a Synergy HT microplate reader (BioTek Instruments Inc., Winooski, VT, USA).

Total DNA content (ng/mL) was determined using the CyQUANT Cell Proliferation Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA), with fluorescence recorded at 485 nm (excitation) and 528 nm (emission) on the same microplate reader. ALP activity was normalized to DNA content and expressed as the ALP/DNA ratio (μmol ALP/ng DNA).

2.3.2. Calcium quantification

Ca²⁺ deposition was quantified in 2D cultures only. GFs and PDLFs were collected at days 1, 7, and 21 by adding 0.5 mL of 0.5 N acetic acid to each well, after which the lysates were stored at −20 °C until analysis.

For measurement, 10 μL of each sample was transferred in duplicate into a 96-well plate and mixed with 300 μL of working reagent consisting of ethanolamine/boric acid buffer (14.8 M, pH 11), orthocresolphtalein complexone (OCPC), and hydroxyquinoline (all from Sigma-Aldrich) in Milli-Q water. Plates were incubated for 10 min at room temperature, and absorbance was subsequently recorded at 570 nm using a Synergy HT microplate spectrophotometer.

A calcium standard curve was (0–100 mg/mL) prepared using a Ca²⁺ stock solution prepared by dissolving calcium chloride dihydrate in Milli-Q water, serially diluted in acetic acid. Final Ca²⁺ concentrations are reported in mg/mL.

2.3.3. Scanning electron microscopy (SEM)

To assess mineralization at day 21, SEM was performed on 3D cultures. Fibrin hydrogels were first rinsed with PBS and fixed in 300 μL of a solution containing 3% paraformaldehyde (PFA) and 3% glutaraldehyde. Samples were stored at 4 °C until further processing.

Following fixation, the hydrogels were digested with trypsin to expose the embedded cells and were then dehydrated through a graded ethanol series, followed by acetone. Dehydrated specimens were dried using a critical point dryer (Leica EM CPD300) with liquid carbon dioxide, sputter-coated with a platinum–palladium alloy using a Leica EM ACE600, and mounted onto aluminum stubs with carbon adhesive tabs.

Imaging was conducted using a Zeiss Gemini Sigma 300 field emission gun scanning electron microscope (FEG-SEM; Carl Zeiss Microscopy GmbH, Jena, Germany) at 2.00 K magnification (EHT = 3.00 kV, Signal A = SE2).

2.3.4. Confocal imaging

Cell alignment and matrix organization in the 3D fibrin hydrogels were assessed by confocal microscopy at day 21. Hydrogels were washed with PBS, fixed for 30 min in formalin, and subsequently stored in PBS containing 1% PSF and 1% formalin at 4 °C.

After blocking in PBS/BSA, samples were incubated with a primary antibody against human collagen type I (mouse IgG3, 1:100; Abcam, Cambridge, UK) for 2–5 days, followed by washing and incubation with Alexa Fluor 488-conjugated goat anti-mouse IgG (1:500; Molecular Probes, Thermo Fisher Scientific) together with Phalloidin–Alexa Fluor 647 (1:500; Molecular Probes, Thermo Fisher Scientific) for at least 2 days. Nuclei were counterstained with DAPI (1:500; Molecular Probes, Thermo Fisher Scientific) for 5 min. The gels were mounted inverted on a glass-bottom dish (Cellview by Greiner) in Vectashield mounting medium (Vector Laboratories, Newark, CA, USA) for imaging.

Imaging was performed using a Leica Stellaris 8 confocal microscope equipped with a 20×/0.75 dry objective. Excitation was provided by laser lines at 405 nm (DAPI), 479 nm (Alexa Fluor 488), and 633 nm (Alexa Fluor 647), with emission collected on hybrid detectors. Images were acquired in xyz scan mode at 600 Hz with a 1 AU pinhole and a pixel dwell time of 1.2 μs. Overview and high-resolution scans used voxel sizes of 1.14 μm and 0.28 μm, respectively. This setup enabled visualization of cell orientation and cytoskeletal structure within the 3D fibrin matrix.

2.3.5. ELISA

Supernatants were collected from GFs and PDLFs at day 14 from both 2D and 3D cultures maintained in either NM or OM. Concentrations of human M-CSF and OPG were quantified using commercially available ELISA kits (R&D Systems, Minneapolis, MN, USA), according to the manufacturer's instructions.

2.4. Osteoclastogenesis

PBMCs were isolated from a buffy coat of a healthy donor (Sanquin, Amsterdam, the Netherlands) using density gradient centrifugation with Ficoll-Paque. Permission to use buffy coats for osteoclastogenesis experiments was obtained from Sanquin (approval number NVT230.0). The buffy coat was diluted 1:1 in PBS supplemented with 1% sodium citrate (pH 7.4). Subsequently, 25 mL of the diluted blood was carefully layered onto 15 mL of Lymphoprep (Axis-Shield PoC AS, Oslo, Norway) and centrifuged for 30 min at 800 × G without brake. The mononuclear cell layer at the interphase was collected, washed three times in PBS-citrate, and resuspended in culture medium.

PBMCs were seeded at 1 × 10⁶ cells per well in 96-well plates and cultured for 3 weeks in conditioned medium derived from GF and PDLF cultures grown in 2D and 3D conditions. As a control, PBMCs were also cultured in DMEM supplemented with serum and antibiotics, designated as just medium (JM).

Osteoclast differentiation and activity were evaluated at multiple time points through qPCR, TRAcP activity assays, and cytochemical TRAcP staining.

2.4.1. TRAcP activity and staining

Osteoclast formation was evaluated by quantifying TRAcP activity and by cytochemical TRAcP staining. For the osteoclast quantification, cells were fixed at day 21 in 4% PBS-buffered formaldehyde for 10 min and rinsed with distilled water. TRAcP staining was performed using the Acid Phosphatase, Leukocyte Kit (Sigma-Aldrich) according to the manufacturer's protocol. Nuclei were counterstained with DAPI (Thermo Fisher Scientific) for 5 min. The formation of the purple precipitate characteristic of the TRAcP assay was monitored microscopically and the reaction was stopped before excessive color development occurred, allowing clear visualization of DAPI-stained nuclei. Only cells containing three or more DAPI-positive nuclei were classified as multinucleated osteoclasts. Multinucleated TRAcP+ cells were imaged using a Leica DMIL microscope equipped with a DFC7000T camera (Leica Microsystems), and six standardized fields per well were analyzed at 100× magnification.

For TRAcP activity, cells were harvested at day 21, lysed in 100 μL of sodium acetate buffer (0.1 M NaAc, pH 5.8, containing 0.1% Triton X-100), and stored at −20 °C until analysis. TRAcP activity was measured using 4-nitrophenyl phosphate bis (cyclohexylammonium) salt (Sigma-Aldrich) at pH 5.8, following the method described by Ljusberg et al. (Ljusberg et al., 1999). Absorbance was recorded at 405 nm using a Synergy HT microplate reader (BioTek Instruments). Total protein concentration (μg/mL) was determined with the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) and measured at 540 nm. TRAcP activity was expressed as arbitrary units per microgram of protein (AU/μg).

2.4.2. RNA isolation and qPCR

For the osteogenic differentiation experiments, total RNA was isolated from 2D and 3D cultures of GFs and PDLFs at days 1 and 14. For 3D samples, fibrin hydrogels were snap-frozen in liquid nitrogen and mechanically disrupted using sterile RNase-free disposable pestles (Thermo Fisher Scientific). TRIzol Reagent (Thermo Fisher Scientific) was subsequently added, and RNA extraction proceeded according to the manufacturer's instructions. For 2D cultures, RNA was isolated using the RNeasy Mini Kit (Qiagen, Hilden, Germany).

For the osteoclastogenesis experiments, RNA was isolated from PBMC-derived cultures at day 21 using the RNeasy Mini Kit. RNA concentration and purity were determined with a Synergy microplate reader (BioTek, Santa Clara, CA, USA). cDNA was synthesized using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific) with both Oligo(dT)₁₈ and random hexamer (D(N)₆) primers. Gene-specific primers were designed using Primer Express software v2.0 (Applied Biosystems, Foster City, CA, USA), ensuring that each amplicon spanned at least one intron to avoid genomic DNA amplification. Primer sequences are listed in Table 1.

Table 1.

Primer sequences used for quantitative PCR.

Gene	Sequence 5′-3′	Amplicon length (bp)	Ensembl gene ID
HMBS	TCCAAGCGGAGCCATGTCTG	192	ENSG00000256269
	CCTGTGGTGGACATAGCAAT
COL1A1	TCCGGCTCCTGCTCCTCTTA	336	ENSG00000108821
	GGCCAGTGTCTCCCTTG
SPARC	CTGTCCAGGTGGAAGTAGG	233	ENSG00000113140
	GTGGCAGGAAGAGTCGAAG
MMP2	CCGTGGTGAGATCTTCTTCTTC	232	ENSG00000087245
	GCTGGTCAGTGGCTTGGGGTA
MMP9	TGACAGCGACAAGAAGTG	219	ENSG00000100985
	CGTGGCTCAGGTTCAGG
TGFB1	CTACTACGCCAAGGAGGTCA	199	ENSG00000105329
	CACGTGCTGCTCCACTTT
CSF1	CACCATGCGCTTCAGAGAT	208	ENSG00000184371
	CCAGTCCTTGTCAAGGAGAT
ACP5	CACAATCTGCAGTACCTGCAAGAT	128	ENSG00000102575
	CCCATAGTGGAAGCGCAGATA
CTSK	CCATATGTGGGACAGGAAGAGAGTT	149	ENSG00000143387
	TGCATCAATGGCCACAGAGA
TNFRSF11A	CCTGGACCAACTGTACCTTCCT	67	ENSG00000141655
	ACCGCATCGGATTTCTCTGT
DCSTAMP	ATTTTCTCAGTGAGCAAGCAGTTTC	101	ENSG00000164935
	AGAATCATGGATAATATCTTGAGTTCCTT

HMBS, hydroxymethylbilane synthase (encoding for porphobilinogen deaminase; PBGD); COL1A1, collagen type I alpha 1 chain; SPARC, secreted protein acidic and rich in cysteine (osteonectin); MMP2, matrix metalloproteinase-2; MMP9, matrix metalloproteinase-9; TGFB1, transforming growth factor beta-1 (TGF-β1); CSF1, colony-stimulating factor 1 (M-CSF); ACP5, tartrate-resistant acid phosphatase (TRAcP); CTSK, cathepsin K; TNFRSF11A, tumor necrosis factor receptor superfamily member 11a (encoding for receptor activator of nuclear factor-κB (RANK)); DCSTAMP, dendritic cell-specific transmembrane protein.

For each gene, the first sequence represents the forward primer, the second sequence the reverse primer.

qPCR was performed on a LightCycler 480 system (Roche, Basel, Switzerland). The cycling protocol consisted of an initial activation step at 94 °C for 10 min, followed by 40 cycles of denaturation at 95 °C for 30 s and combined annealing/extension at 60 °C for 1 min. A melting curve analysis was included at the end of each run to verify amplification specificity and exclude nonspecific products.

Gene expression levels were normalized to the housekeeping gene HMBS (encoding PBGD), which showed stable expression across all experimental conditions. Relative expression was calculated using the ΔCt method (ΔCt = Ct_gene of interest − Ct_HMBS), and data were presented as 2^(−ΔCt).

2.5. Statistical analyses

Statistical analyses were performed using SPSS software (version 30.0; IBM, New York, USA), and graphs were generated with GraphPad Prism (version 10.2.3; GraphPad Software, La Jolla, CA, USA). Data distribution was assessed using the Shapiro-Wilk test, which confirmed non-normal distribution for all datasets, including ALP activity, DNA content, Ca²⁺ quantification, qPCR, ELISA, TRAcP activity, and osteoclast quantification.

For ALP activity and DNA content, within-group changes between days 1 and 14 for each cell type were evaluated using the Wilcoxon signed-rank test, whereas comparisons between GFs and PDLFs at each time point were performed using the Mann-Whitney U test. For Ca²⁺ quantification, temporal changes across days 1, 7, and 21 within each cell type were analyzed using Friedman's test with Dunn's post hoc correction, and pairwise comparisons between GFs and PDLFs at individual time points were conducted using the Mann-Whitney U test.

For qPCR data from osteogenic differentiation experiments, changes in gene expression over time within each cell type were assessed using Friedman's test with Dunn's post hoc correction. Pairwise comparisons between GFs and PDLFs at matched time points were performed using the Mann-Whitney U test. For osteoclastogenesis experiments, changes across conditioned medium types were analyzed using Friedman's test with Dunn's post hoc correction, while direct comparisons between GFs- and PDLFs-derived conditions at each time point were evaluated using the Wilcoxon signed-rank test. The same statistical approach was applied to TRAcP activity and osteoclast quantification.

For ELISA, differences in protein secretion across experimental conditions within each cell type were analyzed using Friedman's test with Dunn's post hoc correction. Pairwise comparisons between GFs and PDLFs, as well as between 2D and 3D conditions at corresponding time points, were assessed using the Wilcoxon signed-rank test.

A p-value ≤0.05 was considered statistically significant for all analyses.

3. Results

3.1. Osteogenesis

3.1.1. Increased ALP activity and Ca²⁺ deposition in osteogenic medium

To evaluate the osteogenic differentiation of GFs and PDLFs, DNA content, ALP activity, and Ca²⁺ deposition were assessed in 2D cultures over time (Fig. 1). Because ALP activity was normalized to cell number, DNA content was first quantified at days 1 and 14.

Fig. 1.

Osteogenic differentiation of gingival fibroblasts (GFs) and periodontal ligament fibroblasts (PDLFs). A) DNA content of GFs and PDLFs cultured in 2D monolayers at days 1 and 14. B) ALP activity normalized to DNA content in GFs and PDLFs at days 1 and 14. C) Calcium deposition in GFs and PDLFs at days 1, 7, and 21. Both cell types showed significant increases in ALP activity and calcium deposition over time, with no significant differences between GFs and PDLFs at corresponding time points. Data are presented as median and range. *p < 0.05, **p < 0.01, ***p < 0.001 N = 12 donors.

Both cell types showed a significant increase in DNA content from day 1 to day 14 (GFs: p < 0.01; PDLFs: p < 0.001), indicating cell proliferation over time. No significant differences in DNA content were observed between GFs and PDLFs at either time point (p > 0.05; Fig. 1A).

ALP activity, an early marker of osteogenic differentiation, increased significantly in both GFs and PDLFs at day 14 compared to day 1 (GFs: p < 0.01; PDLFs: p < 0.001), showing progression toward an osteogenic phenotype. No significant differences in ALP activity were detected between the two cell types at either time point (p > 0.05; Fig. 1B).

To assess matrix mineralization, Ca²⁺ deposition was quantified at days 1, 7, and 21 (Fig. 1C). Both GFs and PDLFs exhibited a progressive increase in Ca²⁺ deposition over time, with significantly higher levels at day 21 compared with days 1 and 7 (p < 0.05 and p < 0.001, respectively). No significant differences in Ca²⁺ deposition were observed between GFs and PDLFs at any time point (p > 0.05; Fig. 1C).

Collectively, these findings demonstrate that both GFs and PDLFs undergo osteogenic differentiation and matrix mineralization under osteogenic culture conditions, with no apparent differences between the two cell populations.

3.1.2. Osteogenic medium induces mineralized matrix formation in 3D fibrin hydrogels

SEM was performed at day 21 to evaluate the ECM organization and mineral deposition in 3D fibrin hydrogels. Both GFs and PDLFs were shown to be dispersed through the fibrin hydrogels (Fig. 2 A–D). In contrast to cultures maintained in NM, OM induced clear mineralization in both cell types, as evidenced by the presence of mineral-like nodules within the fibrin network (Fig. 2 C and D, arrows). These nodules were mostly deposited on the ECM, which likely consists of newly deposited collagen fibers with cells embedded within this network. NM cultures displayed a predominantly fibrillar architecture without apparent mineralized deposits (Fig. 2 A and B).

Fig. 2.

SEM analysis of mineralization in 3D fibrin hydrogels. Representative SEM images of GFs and PDLFs cultured in 3D fibrin hydrogels for 21 days under normal medium (NM) or osteogenic medium (OM) conditions. Cultures maintained in OM showed extensive mineral nodule formation embedded within a dense extracellular matrix, whereas NM cultures display a predominantly fibrillar structure with no mineral deposition. Scale bar: 10 μm.

3.1.3. Intracellular collagen localization in 3D fibrin hydrogels

Confocal microscopy was performed at day 21 to assess cell morphology and collagen organization within 3D fibrin hydrogels. Representative images of GFs and PDLFs cultured under NM conditions are shown in Fig. 3. Both fibroblast populations displayed an elongated morphology with branched cytoplasmic extensions extending throughout the fibrin matrix, indicating active interaction with the surrounding scaffold.

Fig. 3.

Confocal analysis of collagen localization in 3D fibrin hydrogels. Representative confocal images of GFs and PDLFs cultured in 3D fibrin hydrogels under normal medium (NM) conditions at 21 days. Cells display elongated morphology with branched protrusions extending within the hydrogel matrix. Collagen signal is predominantly localized within the cytoplasm rather than forming organized extracellular fibrillar structures. A) Nuclei stained with DAPI (blue), B) Collagen type I (green), C) F-actin cytoskeleton (red), D) Merged image showing intracellular collagen localization. Scale bar: 10 μm.

Collagen type I staining was detected in both GFs and PDLFs and was predominantly localized within the cell body and cellular extensions rather than as an organized extracellular fibrillar network. The staining pattern appeared largely intracellular and pericellular, suggesting ongoing collagen synthesis and processing at this time point. No clear extracellular collagen fiber formation was observed within the fibrin hydrogels.

3.1.4. Gene expression analysis of fibroblasts

To evaluate the effects of culture dimensionality and osteogenic stimulation on fibroblast phenotype, the expression of genes related to extracellular matrix production, matrix remodeling, and osteoclast regulation was assessed by qPCR at days 1 and 14 (Fig. 4).

Fig. 4.

Gene expression profiles of GFs and PDLFs. Relative gene expression levels were measured by qPCR at days 1 and 14 in GFs and PDLFs cultured in 2D monolayers or 3D fibrin hydrogels under NM or OM conditions. A) COL1A1, B) SPARC (Osteonectin), C) MMP2, D) MMP9, E) TGFB1 (TGF-β1), F) CSF1 (M-CSF). Data are presented as median and interquartile range. *p < 0.05, **p < 0.01, ***p < 0.001 N = 12 donors.

Collagen type I (COL1A1) and osteonectin (SPARC), two genes associated with matrix formation and mineralization, were expressed by both GFs and PDLFs under all culture conditions (Fig. 4A–B). COL1A1 expression was generally maintained throughout the culture period, with the highest levels observed in GFs cultured under 3D NM conditions. SPARC expression increased over time in both cell types and was generally higher in PDLFs than in GFs, particularly under 3D culture conditions (p < 0.001), indicating active matrix maturation and mineralization-associated activity.

Matrix metalloproteinases (MMPs) play a major role in ECM degradation and regeneration, including bone regeneration (Khoswanto, 2023). Matrix remodeling was assessed by analyzing the gelatinases MMP2 and MMP9 (Fig. 4C–D). Both fibroblast populations readily expressed MMP2, which showed a tendency toward higher expression under 2D culture conditions than under 3D conditions. In contrast, MMP9 expression remained low across all experimental groups, with only modest increases observed under selected conditions, suggesting a limited contribution of MMP9 to matrix remodeling in this model.

TGF-β1 (TGFB1; Fig. 4E), a key regulator of ECM production and bone remodeling (Wu et al., 2016), remained largely stable across culture conditions in both fibroblast populations. This finding indicates that osteogenic stimulation and fibrin culture had only a limited effect on its transcriptional regulation.

Given its role in osteoclast precursor survival, M-CSF (CSF1; Fig. 4F) expression was also evaluated. While several condition-specific differences were observed, M-CSF expression remained within a relatively narrow range and no consistent differences between GFs and PDLFs were detected (p > 0.05).

Overall, the gene expression data demonstrated that both fibroblast populations responded similarly to osteogenic stimulation and 3D culture, with only modest differences observed between GFs and PDLFs.

3.1.5. Decreased M-CSF and increased OPG secretion in osteogenic conditions

Because of their key roles in osteoclastogenesis, the secretion of M-CSF and OPG was quantified in conditioned media collected from GF and PDLF cultures at day 14 (Fig. 5). M-CSF promotes the survival and proliferation of osteoclast precursors, whereas OPG inhibits osteoclast formation by preventing RANKL-RANK signaling (Florencio-Silva et al., 2015; Fuller et al., 1993; Boyce and Xing, 2007; Roodman, 1999). Since both factors can be produced by periodontium-derived fibroblasts, their levels were measured in conditioned media from 2D and 3D cultures maintained under NM or OM conditions.

Fig. 5.

Secretion of osteoclast-regulating factors by GFs and PDLFs. A) Macrophage colony-stimulating factor (M-CSF) and B) Osteoprotegerin (OPG) levels measured by ELISA in conditioned media collected at day 14 from GFs and PDLFs cultured in 2D monolayers or 3D fibrin hydrogels under NM or OM conditions. In both cell types, OPG levels were significantly higher in 2D OM than in NM, whereas M-CSF levels decreased under both 2D and 3D OM conditions. Data are presented as median and range. *p < 0.05, **p < 0.01, ***p < 0.001 N = 12 donors.

M-CSF secretion was reduced following osteogenic stimulation in both fibroblast populations (Fig. 5A). In GFs, M-CSF levels were significantly lower under 2D OM than 2D NM conditions (p < 0.01), also in PDLFs, a significant reduction was observed under 3D OM compared to 3D NM conditions (p < 0.05). Apart from a higher M-CSF level in PDLFs than GFs under 2D OM conditions (p < 0.05), no substantial differences between the two fibroblast populations were observed.

In contrast, OPG secretion increased following osteogenic stimulation (Fig. 5B). Both GFs and PDLFs exhibited significantly higher OPG levels under 2D OM than 2D NM conditions (p < 0.001 and p < 0.01, respectively), whereas no significant effects of osteogenic stimulation were detected in 3D cultures. No differences in OPG secretion were observed between GFs and PDLFs under any culture condition (p > 0.05; Fig. 5B).

Collectively, osteogenic stimulation shifted the secretory profile of both fibroblast populations toward a less osteoclast-supportive phenotype, characterized by reduced M-CSF and increased OPG secretion.

3.2. Osteoclastogenesis

3.2.1. Osteoclast formation is reduced by osteogenic preconditioning of fibroblasts

Given the reduced M-CSF and increased OPG secretion observed under OM conditions, we next investigated whether osteogenic preconditioning alters the osteoclastogenic capacity of fibroblasts. To evaluate the effects of fibroblast-derived factors without the influence of exogenous osteoclastogenic supplements, PBMCs were cultured in conditioned media collected from fibroblasts maintained under either NM or OM conditions in the absence of additional M-CSF and RANKL.

TRAcP, an enzyme expressed by osteoclast precursors and multinucleated osteoclasts that is involved in bone resorption (Hayman, 2008), was assessed at day 21 by both enzymatic activity and staining. Osteoclast formation was quantified by counting TRAcP-positive multinucleated cells containing three or more nuclei (Fig. 6A–B). Representative images demonstrated abundant multinucleated osteoclasts in cultures exposed to conditioned media derived from fibroblasts maintained under NM conditions, whereas substantially fewer osteoclasts were observed in cultures exposed to conditioned media derived from OM cultures.

Fig. 6.

Osteoclast-inducing capacity of GFs and PDLFs. A) Representative micrographs of TRAcP-stained osteoclast cultures. TRAcP-positive multinucleated cells with three or more nuclei (white arrows) were counted as osteoclasts. Nuclei are counterstained with DAPI (blue); TRAcP staining appears purple. B) Quantification of osteoclast numbers at day 21. Osteoclast formation was higher in NM-derived conditioned media than in OM-derived conditioned media for both cell types. C) TRAcP activity measured at day 21 in osteoclast cultures treated with conditioned media derived from GFs and PDLFs cultured in 2D or 3D under NM or OM conditions. Data are presented as median and range. JM: Just Medium. *p < 0.05, **p < 0.01, ***p < 0.001 N = 12 donors. Scale bar: 50 μm.

Quantification confirmed that conditioned media derived from NM-cultured fibroblasts induced significantly more osteoclast formation than conditioned media derived from OM-cultured fibroblasts (Fig. 6B). In GF-derived conditioned media, osteoclast numbers were significantly higher under 2D NM than 2D OM conditions (p < 0.01). Similarly, in PDLF-derived conditioned media, osteoclast numbers were significantly higher under both 2D NM (p < 0.01) and 3D NM (p < 0.001) conditions compared with their corresponding OM groups. No significant differences in osteoclast numbers were observed between 2D and 3D cultures within either fibroblast population (p > 0.05; Fig. 6B). The baseline reference condition group, in which PBMCs were cultured in JM alone, consistently showed the lowest osteoclast numbers, confirming that fibroblast-derived factors were required to support osteoclast differentiation under these conditions (Fig. 6B).

TRAcP activity was additionally measured as an indicator of osteoclast differentiation (Fig. 6C). Although TRAcP activity was generally higher in conditioned-media groups than in the JM group, no significant differences were observed between GF- and PDLF-derived conditioned media or between NM and OM conditions (p > 0.05; Fig. 6C). Together, these findings indicate that osteogenic preconditioning of fibroblasts reduces their capacity to support osteoclast formation, despite only modest effects on overall TRAcP activity.

3.2.2. Gene expression analysis of osteoclasts

To further characterize osteoclast differentiation, expression of genes associated with osteoclastogenesis, cell fusion, and resorptive activity was analyzed in PBMC cultures exposed to conditioned media derived from GFs and PDLFs (Fig. 7).

Fig. 7.

Osteoclast-related gene expression in response to fibroblast-derived media. Relative gene expression levels were measured by qPCR in osteoclast cultures at day 21 following stimulation with conditioned media derived from GFs and PDLFs. A) TNFRF11A (RANK), B) DCSTAMP, C) ACP5 (TRAcP), D) CTSK (Cathepsin K). Data are presented as median and interquartile range. JM: Just Medium. *p < 0.05, **p < 0.01, ***p < 0.001 N = 12 donors.

Expression of RANK (TNFRSF11A; Fig. 7A), the receptor responsible for mediating RANKL-induced osteoclastogenesis (Wada et al., 2006), showed considerable variability across experimental conditions. Although several condition-specific differences were observed, no consistent pattern distinguishing GF- from PDLF-derived conditioned media emerged.

Downstream of RANK/RANKL signaling, which induces the expression of other osteoclastogenic genes involved in osteoclast precursor fusion (DCSTAMP) and resorptive function (ACP5, CTSK) were evaluated (Fig. 7B–D) (Yavropoulou and Yovos, 2008).

Overall, expression of these genes was elevated in cultures exposed to fibroblast-conditioned media compared with the control condition, indicating successful induction of osteoclast differentiation. For both GF- and PDLF-derived conditioned media, expression of DCSTAMP and ACP5 generally tended to be higher under NM than OM conditions, consistent with the greater number of osteoclasts observed in the NM groups. Similarly, Cathepsin K (CTSK) expression remained elevated across most experimental conditions, reflecting the acquisition of an osteoclast-like phenotype.

No significant differences between GF- and PDLF-derived conditioned media were observed for DCSTAMP, TRAcP (ACP5), or cathepsin K (CTSK) expression (p > 0.05). Collectively, these findings indicate that conditioned media derived from both fibroblast populations supported osteoclast differentiation, whereas osteogenic conditioning of fibroblasts generally reduced the expression of osteoclast-associated genes.

4. Discussion

Tissue engineering strategies in periodontal regeneration aim to restore lost periodontal structures while ensuring integration and long-term stability of regenerated tissue. In alveolar bone regeneration, this requires coordinated coupling between bone formation and bone resorption (Schindeler et al., 2008). The present study compared the osteogenic differentiation capacity and osteoclast-modulating properties of GFs and PDLFs cultured in conventional 2D monolayers and 3D fibrin hydrogels. The results demonstrate that both fibroblast populations acquire an osteoblast-like phenotype under osteogenic stimulation and, importantly, that osteogenic differentiation reduces their capacity to support osteoclastogenesis. Overall, GFs and PDLFs exhibited similar functional behavior irrespective of culture dimensionality.

This finding is particularly relevant for GFs, which are abundant, easily accessible, and can be harvested with minimal donor-site morbidity compared with PDLFs (Alfonso Garcia et al., 2020). More broadly, our results support the concept that fibroblasts residing in soft tissues can actively participate in hard tissue regeneration. This challenges the traditional view of periodontal regeneration, which has largely focused on excluding soft tissues from the regenerative site through the use of occlusive barriers, and highlights the potential of selective use of soft tissue-derived cells with osteogenic potential in periodontal and alveolar bone regeneration.

Both fibroblast populations successfully underwent osteogenic differentiation, as evidenced by increased ALP activity, calcium deposition, and mineral nodule formation. Although some differences in the expression of individual genes were observed between GFs and PDLFs, these were not reflected in the functional outcomes. ALP is a key regulator of mineralization (Golub et al., 1992), and its progressive increase in both cell populations, together with increased calcium deposition, confirms acquisition of an osteoblast-like phenotype. These findings are consistent with previous studies demonstrating the osteogenic potential of both GFs and PDLFs and further support the concept that these fibroblast populations exhibit largely comparable osteogenic behavior under appropriate stimulatory conditions, making them useful models for studying periodontal bone homeostasis (Ceylan et al., 2024; Prins et al., 2024; Cayami et al., 2022; de Vries et al., 2026).

The 3D fibrin hydrogels were included to mimic the provisional matrix that forms during early wound healing. While fibrin influenced the expression of several matrix-related genes, it did not substantially alter either osteogenic differentiation or the subsequent regulation of osteoclastogenesis. Matrix-associated genes, including COL1A1, SPARC, and MMP2 collectively indicated active ECM remodeling during differentiation. Collagen type I provides the structural framework for mineral deposition (Clarke, 2008), whereas osteonectin facilitates interactions between collagen fibrils and mineral crystals during matrix maturation (de Wildt et al., 2019; Rosset and Bradshaw, 2016). Previous studies have similarly reported dynamic regulation of these markers during osteogenic differentiation in 3D environments (Alves et al., 2015; Berendsen et al., 2010; Inanc et al., 2006). For example, Berendsen et al. observed reduced COL1A1 expression in PDLFs cultured in 3D collagen matrices during mineral induction (Berendsen et al., 2010), whereas Inanc et al. reported increased expression of osteogenic matrix markers, including osteonectin, in PDLFs cultured within 3D matrices (Inanc et al., 2006). In line with these observations, our findings suggest that the transcriptional differences observed between culture conditions primarily reflect matrix remodeling and maturation rather than differences in osteogenic capacity. Despite variations in gene expression, both fibroblast populations exhibited comparable ALP activity, calcium deposition, and mineralized nodule formation. Although calcium deposition was quantified only in 2D cultures, SEM analysis confirmed robust mineral nodule formation in 3D fibrin hydrogels under osteogenic conditions, further supporting successful osteogenic differentiation in both cell types.

Confocal microscopy further demonstrated predominantly intracellular collagen localization within the fibrin hydrogels. This likely reflects active procollagen synthesis and intracellular processing within the secretory pathway (Wu et al., 2025). Because efficient extracellular fibril formation requires procollagen cleavage and organized pericellular assembly mediated by fibronectin-integrin interactions (Kadler et al., 2008), the predominantly intracellular signal observed at day 21 likely represents ongoing matrix production and maturation rather than impaired collagen synthesis. Together, these findings indicate that both fibroblast populations actively interact with and remodel the fibrin microenvironment.

A major finding of the present study was that osteogenic differentiation altered the paracrine profile of both fibroblast populations. During bone remodeling, osteoclast recruitment and activation are orchestrated by several coupled regulators, including the RANK/RANKL/OPG axis and M-CSF signaling (Boyce and Xing, 2008; Rolph and Das, 2020). Osteogenic stimulation consistently increased OPG secretion while reducing M-CSF production. Since OPG functions as a soluble decoy receptor that prevents RANKL-RANK interactions and M-CSF is required for osteoclast precursor survival, proliferation, and migration (Rolph and Das, 2020), these changes collectively indicate a shift toward a less osteoclast-supportive microenvironment. Interestingly, these effects were more evident at the protein than transcript level, which is consistent with the extensive post-transcriptional and post-translational regulation of cytokine secretion (Chambers et al., 1993).

The functional relevance of the altered secretome became evident in the osteoclastogenesis experiments. Conditioned media derived from fibroblasts maintained under normal culture conditions consistently promoted osteoclast formation, whereas conditioned media derived from osteogenically differentiated fibroblasts significantly reduced osteoclast numbers. This reduction was accompanied by lower expression of osteoclast-associated markers, including DC-STAMP, ACP5, and CTSK, collectively indicating reduced osteoclast differentiation, fusion, and maturation. DC-STAMP is essential for fusion of osteoclast precursors into multinucleated cells (Omi and Mishina, 2022; Yagi et al., 2005), whereas ACP5 and CTSK are hallmark markers of mature osteoclasts and bone resorption (Florencio-Silva et al., 2015; Dai et al., 2020). Accordingly, the reduced formation of multinucleated TRAcP+ osteoclasts, together with the lower expression of osteoclast-associated genes, supports the concept that osteogenic maturation of fibroblasts shifts the microenvironment toward a less osteoclastogenic state. Interestingly, despite these changes, overall TRAcP enzymatic activity remained largely unchanged. This is consistent with findings by Chiba-Ohkuma et al., who showed that TRAcP activity does not necessarily correlate with osteoclast size or multinucleation status (Chiba-Ohkuma et al., 2025). Therefore, osteoclast number and maturation state may provide a more sensitive reflection of osteoclastogenic potential than TRAcP activity alone.

Interestingly, RANK expression was increased under certain osteogenic conditions despite reduced osteoclast formation. However, RANKL remained undetectable in our system, and elevated OPG levels suggest that receptor upregulation likely reflects compensatory regulation rather than enhanced osteoclastogenesis. Similar observations have been reported in other systems where osteoclast precursors increase receptor expression in response to reduced ligand availability (Takai et al., 1998). Collectively, the changes in M-CSF, OPG, osteoclast-associated genes, and osteoclast numbers indicate that osteogenic differentiation induces a functional switch toward an osteoclast-inhibitory phenotype.

Because the conditioned media used for osteoclastogenesis contained either ascorbic acid alone (NM) or in combination with β-glycerophosphate (OM), potential direct effects of these supplements on osteoclast formation should be considered. However, culturing PBMCs in NM, OM, or JM alone did not affect osteoclast numbers (data not shown), suggesting that the observed effects primarily reflect fibroblast-derived paracrine regulation. Although ascorbic acid and β-glycerophosphate have been reported to influence osteoclastogenesis under certain conditions (Le Nihouannen et al., 2010; Noh and Yim, 2011; Ragab et al., 1998), the effects are highly context dependent and often require exogenous RANKL stimulation. In our model, osteoclastogenesis depended entirely on fibroblast-derived factors in the absence of exogenous M-CSF and RANKL, making a direct effect of these supplements unlikely.

This interpretation is further supported by our previous studies in which fibroblasts were first cultured under osteogenic conditions for three weeks to induce mineralization, followed by washing and co-culture in regular medium without ascorbic acid or β-glycerophosphate (Ceylan et al., 2024; Prins et al., 2024). In those studies, osteogenic differentiation similarly reduced osteoclast formation, showing that mineralization and altered paracrine signaling, rather than the continuous presence of osteogenic supplements, were responsible for the inhibitory effect. Together, these findings indicate that the reduced osteoclastogenesis observed under OM conditions is primarily driven by osteogenic changes in the fibroblast secretome, characterized by decreased M-CSF and increased OPG secretion, rather than by direct effects of osteogenic supplements on osteoclast precursors.

This study has several limitations. First, 2D cultures were not fibrin-coated, limiting direct assessment of fibrin-specific effects independent of dimensionality. Second, although osteogenic differentiation was assessed using multiple complementary approaches, inclusion of a MSC positive control would have provided additional validation. Third, osteoclast function was not evaluated using resorption assays. Fourth, PBMCs were derived from a single donor, and donor-specific variation in osteoclast precursor populations cannot be excluded. Fifth, the absence of condition-specific early reference time points prevented evaluation of the immediate effects of culture dimensionality and osteogenic stimulation. Finally, osteoclastogenesis was assessed using conditioned media rather than direct co-culture systems. Nevertheless, previous studies have shown that GFs can stimulate osteoclastogenesis in indirect co-culture systems and increase the formation of TRAcP+ multinucleated osteoclasts together with expression of osteoclast differentiation markers such as TRAcP and DC-STAMP (Novello et al., 2024). These observations support the biological relevance of paracrine regulation and strengthen the interpretation of the present findings.

5. Conclusion

In summary, both GFs and PDLFs can differentiate toward an osteoblast-like phenotype and deposit mineralized matrix, irrespective of culture dimensionality. When cultured with NM, they stimulate osteoclast formation, as assessed with conditioned medium. Osteogenic differentiation of these fibroblasts, however, reduces their osteoclast-inducing capacity, although their secreted factors remain sufficient to support osteoclastogenesis. These results indicate that mineralized matrix formation critically alters the secretome of the fibroblasts, thereby preventing osteoclast differentiation. Two key factors were identified: M-CSF that was higher in NM and OPG that was higher in OM. Overall, GFs and PDLFs display comparable osteogenic and osteoclast-modulatory properties, supporting their interchangeable use in periodontal regenerative strategies when appropriate biochemical cues and matrix environments are provided.

One-sentence summary

Osteogenic differentiation of gingival and periodontal ligament fibroblasts inhibits their ability to induce osteoclast formation, irrespective of culture dimensionality or fibroblast origin.

CRediT authorship contribution statement

Merve Ceylan: Writing – original draft, Visualization, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Teun J. de Vries: Writing – review & editing, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Data curation, Conceptualization. Wendy J.A.M. Runderkamp: Writing – review & editing, Data curation. Ineke D.C. Jansen: Writing – review & editing, Data curation. Irene M. Schimmel: Writing – review & editing, Data curation. Daisy I. Picavet-Havik: Writing – review & editing, Data curation. Marja L. Laine: Writing – review & editing, Supervision, Project administration, Methodology, Conceptualization. Ton Schoenmaker: Writing – review & editing, Supervision, Project administration, Methodology, Investigation, Data curation, Conceptualization.

Source of funding

Institutional funding from Academic Centre for Dentistry Amsterdam (ACTA), University of Amsterdam and Vrije University Amsterdam, Amsterdam, the Netherlands.

Declaration of competing interest

The authors declare no competing interests.

Data availability

Data will be made available on request.