Introduction

In 2003, bisphosphonate-associated osteonecrosis of the jaw (BPONJ) was described for the first time [1]. Since then, the incidence of this complication has dramatically increased [2,3,4]. Bisphosphonates cause reduced bone formation and wound healing, which are further risk factors for developing an osteonecrosis. This mainly occurs after intake of highly potent nitrogen-containing bisphosphonates (BP) such as pamidronate or zoledronate [5, 6]. The pathogenesis of this disease is not yet completely known and seems to be a multifactorial disease.

Various studies have shown that nitrogen-containing BP not only affect the metabolism of osteoclasts to reduce bone resorption but also have a negative impact on the metabolism of different cell lines, such as osteoblasts, fibroblasts, keratinocytes, and endothelial cells. In particular, nitrogen-containing BP such as zoledronate have shown a toxic effect on different cell types dependent on the administered dose [7,8,9,10].

The periodontal ligament and the alveolar bone are exposed to a cyclic mechanical loading due to the process of mastication. The occlusal forces of the teeth are transmitted to the alveolar bone by the cells of the periodontal ligament, resulting in permanent turnover of the alveolar bone. Especially periodontal fibroblasts play an important role in the remodeling of the periodontal ligament and interaction with the surrounding cells of the alveolar bone. This mechanical loading is missing after extraction of a tooth and might reduce the bone turnover, especially in patients taking BP [11]. Patullo et al. described the hypofunction due to missing mastication as one of the risk factors for alveolar bone loss [12]. Mechanical loading caused by mastication or orthodontic tooth movement affects cell differentiation, synthesis of the extracellular matrix, and the synthesis of inflammatory cytokines of periodontal fibroblasts [13, 14].

Interleukin (IL)-1β plays a major role. It stimulates osteoclasts via the IL-1-receptor and enhances the synthesis of prostaglandin (PG) E2 and matrix-metallo-proteinases (MMP) [15,16,17]. Bisphosphonates also affect the production of inflammatory cytokines. Suzuki et al. observed an inhibiting effect of etidronate on the synthesis of cyclooxygenase (COX)-2 and PGE2 in different cell lines [18]. Zoledronate, however, showed a pro-inflammatory effect on neutrophil granulocytes and fibroblasts [19, 20]. Liu et al. showed that clodronate reduced the PGE2 synthesis caused by mechanical loading [21].

The present study analyzes the interaction between the bisphosphonates clodronate (non-nitrogen) and zoledronate (nitrogen-containing), the pro-inflammatory cytokine IL-1β, and mechanical loading regarding cell viability, COX-2 gene expression, PGE2 synthesis, and MMP-8/TIMP-1 synthesis of human periodontal ligament fibroblasts (HPdLF).

Material and methods

Cell culture

Cell cultures were prepared and maintained according to standard cell culture procedures. Commercially available HPdLF (Lonza, Basel, Switzerland) were maintained in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, Carlsbad, USA) containing 10% FCS, 1% l-glutamine, and 1% penicillin/streptomycin/neomycin. Cell morphology was observed under light microscope for typical fibroblast morphology. Additionally, alkaline phosphatase was analyzed, which is characteristic for ligament fibroblasts, but not for gingival fibroblasts [22]. The cells were cultured in an incubator with 5% CO2 and 95% air at 37 °C. Cells were passaged at regular intervals depending on their growth characteristics using 0.05% Trypsin-EDTA solution (PAA, Pasching, Austria). Cells were used for experiments at passages four to six.

Mechanical strain devices and incubation with IL-1β and bisphosphonates

As an established experimental model of tensile strain, HPdLF were seeded at a density of 1.0 × 105 cells/well on flexible-bottomed six-well plates (Bioflex® Plates, Flexcell® International Corporation, Hillsborough, USA) and cultured in supplemented DMEM. After 2 days for cell attachment, spread, and growth to subconfluency, the medium was replaced by medium with 1% fetal calf serum (starvation medium) and with 10 nM IL-1β and or two different bisphosphonates (clodronate or zoledronate) of 5 μM for 48 h. A Flexercell Strain Unit (Modell FX 3000, Dunn Labortechnik Gmbh, Asbach, Germany), which is capable of controlling the magnitude, type, and frequency of cell deformation, was used to generate cyclic tensile strain (CTS) in HPdLF. Cells were subjected to CTS of 3% elongation (2 cN/mm2) in an incubator for 12 h. Experiments were repeated three times to confirm reproducibility, and three wells were used for each group.

MTT assay

Cell number of HPdLF was evaluated directly after application of CTS in the six-well plates with the MTT (3-(4,5-dimethylthiazol2-yl)-2,5-diphenyltetrazolium bromide) colorimetric assay (Sigma, München, Germany). Viable cells ferment tetrazolium bromide to formazan, which can be measured after cell lysis photometrically at 550 nm. The experiments were performed in triplicate.

Messenger RNA extraction and reverse transcriptase polymerase chain reaction

Cells were detached with 0.05% Trypsin-EDTA solution directly after stretching and were then individually harvested. Messenger RNA (mRNA) was isolated using the peqGOLD Total RNA KIT (peqLab Biotechnologie Gmbh, Erlangen, Germany). This included a DNAse digestion step. Both the quantity (260 nm) and quality (ratio 260/280 nm) of the RNA were determined by using a NanoDrop Spectrophotometer ND-100 (peqLab Biotechnologie Gmbh, Erlangen, Germany). Reverse transcription (RT) of RNA (100 ng) was performed by standard protocols with Gene Amp PCR System 2400 (Perkin Elmer, Massachusetts, USA) and iScript cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, USA) in a total volume of 20 μl.

Primers (Eurofins MWG Operon, Ebersberg, Germany) were designed using the NCBI-nucleotide library and Primer3-design to detect the mRNA levels (Table 1). All primers had been matched to the mRNA sequences of the target genes (NCBI Blast software). As housekeeping genes, actin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were evaluated. We were able to show the most stable expression for actin and GAPDH using specialized freeware called GeNorm.

Table 1 Oligonucleotide primer sequences used for real-time PCR
Full size table

Quantitative real-time PCR was performed with the IQ5-I-Cycler and IQ5 Optical System software version 2.0 (Bio-Rad Laboratories, Hercules, USA) according to the manufacturer’s instructions, which included an initial denaturation at 95 °C, annealing temperature of 56 °C, and an elongation temperature of 71 °C over 40 cycles. q-PCR amplification was conducted with a reaction mixture containing SYBR Green Supermix (Bio-Rad Laboratories, Hercules, USA), an appropriate amount of paired primers, and 2 μl template cDNA; the reason for this was to determine the threshold of the SYBR green fluorescence curve at the exponential part. This method was applied to calculate the cycle number and CT-value for quantitation.

Furthermore, the CT-values of the actin and GAPDH housekeeping genes and the individual primer efficacy were taken into account. Single product formation was confirmed by melting point analysis. Data were obtained from three individual experiments and normalized to the CT of actin and GAPDH. CDNA from individual cell experiments was analyzed in triplicate PCR. The relative expression levels of each mRNA were evaluated by using a modification of the ΔΔCT method [23].

Enzyme-linked immunosorbent assay

PGE2, MMP-8, and TIMP-1 were measured in supernatants by Quantikine Human Immunoassay (R&D Systems, Inc. Minneapolis, USA) according to the manufacturer’s instructions using a microplate reader (Metertech, Inc., Taipei, Taiwan). The assays were performed in triplicate.

Statistical analysis

SPSS 21.0 (IBM-SPSS, Ehningen, Germany) was used for statistical analyses. To detect the difference between the groups, one-way ANOVA was used with the post hoc Tukey test. A p value <0.05 was considered statistically significant. Box and whisker diagrams: The black bar in the middle of each box represents the median. The box includes all values between the 25th and 75th percentiles. Whiskers indicate values still within the 1.5 interquartile range (IQR). An asterisk (*) represents a significant difference compared to control cells.

Results

MTT assay

Zoledronate reduced the cell number of HPdLF compared to control cells (60.3 vs. 100%). Clodronate, IL-1β, and CTS had no influence on cell number (Fig. 1a). The combination of clodronate and IL-1β did not influence the cell number of HPdLF, whereas zoledronate combined with IL-1β reduced the cell number significantly under 60% compared to control cells (p < 0.05). The addition of CTS abolished this reduction and led to a significant increase of the cell number in all groups (Fig. 1b).

Fig. 1
figure 1

MTT assay. a Cell number test for human periodontal ligament fibroblasts (HPdLF) exposed to 5 μM clodronate, 5 μM zoledronate, and 10 nM IL-1β for 48 h or to 12 h of cyclic tensile strain at strengths (CTS) of 3% compared to control cells set to 100%. b Cell number test for HPdLF for different combinations of the single conditions. (X-axis = different strengths of SMS, Y-axis = cell viability compared to control group in percent). The assays were performed in triplicate, *p < 0.05. The whiskers at each value represent the standard deviation

Full size image

COX-2 gene expression and PGE2 protein level

IL-1β led to a 10-fold increase of COX-2 gene expression. Clodronate and zoledronate did not increase COX-2 gene expression. Clodronate slightly reduced the increase induced by IL-1β, whereas zoledronate enhanced this increase. IL-1β in combination with CTS led to a 25-fold increase of COX-2 gene expression. Clodronate or zoledronate combined with mechanical loading did not increase COX-2. The combination of mechanical loading, IL-1β, and zoledronate increased COX-2 70-fold (Fig. 2).

Fig. 2
figure 2

Quantitative RT-PCR-results for cyclooxygenase (COX)-2 of human periodontal ligament fibroblasts (HPdLF) exposed to 10 nM IL-1β, 5 μM clodronate (clo.), 5 μM zoledronate (zole.), or combination of both for 48 and 12 h of cyclic tensile strain (CTS) at strengths of 3% compared to control cells (contr.). Data were obtained from three individual experiments (means ± SD, *p < 0.05)

Full size image

The ELISA of PGE2-synthesis was done as an indirect proof of COX-2 at the protein level. Corresponding to the results at gene level, we observed significantly increased values of PGE2 after incubation of HPdLF with IL-1β. Zoledronate enhanced this increase nearly 4-fold. CTS alone did not increase PGE2 synthesis. Values for CTS combined with IL-1β were similar to IL-1β alone. The highest values of PGE2 were found after incubation with IL-1β and zoledronate combined with mechanical loading (Fig. 3).

Fig. 3
figure 3

Prostaglandin (PG) E2 secretion in supernatant liquor of human periodontal ligament ibroblasts (HPdLF) exposed to 10 nM IL-1β, 5 μM clodronate (clo.), 5 μM zoledronate (zole.), or combination of both for 48 and 12 h of cyclic tensile strain at strengths of 3%. Data given in picograms per milliliter relative to control cells

Full size image

MMP-8 and TIMP-1 protein synthesis

IL-1β increased MMP-8 synthesis significantly. Clodronate and zoledronate did not alter MMP-8 synthesis, but reduced the IL-1β-induced MMP-8 synthesis slightly. Mechanical loading increased MMP-8, and when combined with IL-1β, this increase was enhanced. Both bisphosphonates erased the MMP-8 increase induced by mechanical loading (Fig. 4).

Fig. 4
figure 4

Matrix metalloproteinase (MMP)-8 secretion in supernatant liquor of human periodontal ligament fibroblasts (HPdLF) exposed to 10 nM IL-1β, 5 μM clodronate (clo.), 5 μM zoledronate (zole.), or combination of both for 48 and 12 h of cyclic tensile strain at strengths of 3%

Full size image

TIMP-1 values were not altered by IL-1β, clodronate, or zoledronate. The combination of IL-1β and zoledronate showed a slight reduction of TIMP-1. TIMP-1 synthesis was increased by mechanical loading in all groups especially in combination with IL-1β (Fig. 5).

Fig. 5
figure 5

Tissue inhibitor of matrix metalloproteinase (TIMP)-1 secretion in supernatant liquor of human periodontal ligament fibroblasts (HPdLF) exposed to 10 nM IL-1β, 5 μM clodronate (clo.), 5 μM zoledronate (zole.), or combination of both for 48 and 12 h of cyclic tensile strain at strengths of 3%

Full size image

Discussion

The present study analyzed the metabolism of HPdLF under mechanical loading, IL-1ß, and bisphosphonates. For the first time, we observed a positive interaction between mechanical loading and bisphosphonates regarding the cell number. The study also shows for the first time an additive pro-inflammatory effect between IL-1ß, mechanical loading, and zoledronate regarding COX-2 gene expression and PGE2 synthesis.

Zoledronate reduced the cell number of HPdLF, and IL-1ß enhanced this effect. Mechanical loading had a positive effect on the cell number and negated the decreased caused by zoledronate. Pabst et al. and Walter et al. observed this negative effect of zoledronate regarding the cell viability of different cell lines at the same concentration of 5 μM. Higher concentrations, such as 50 μM, enhanced the reduction of the viability [8, 24]. Agis et al. described a reduced cell viability and higher apoptosis rate of HPdLF by zoledronate at concentrations from 30 to 100 μM [25]; these results correspond to the present study, which additionally showed that IL-1ß enhanced this effect.

In a previous study, we combined zoledronate and clodronate with high levels of mechanical loading (10% elongation). We observed an additive negative effect of static mechanical loading regarding the reduction of cell viability [26]. In this study, we used a different protocol for mechanical loading entailing low forces and cyclic loading, which simulates mastication forces. The deviating results indicate that the type of mechanical loading might be a crucial effect modifier regarding the influence of bisphosphonates. Further studies are desirable to analyze this interaction more properly and maybe discover new insights into the pathomechanism of BPONJ.

IL-1ß alone and in combination with mechanical loading increased COX-2 gene expression and PGE2 synthesis. The combination IL-1ß, mechanical loading, and zoledronate led to the highest COX-2 gene expression and PGE2 synthesis. Diercke et al. could previously show that IL-1ß led to an increased COX-2 gene expression of human cementoblasts, which was enhanced by mechanical loading in the form of compression [27].

Previous studies have shown that mechanical loading might have pro-inflammatory effects, especially in an inflamed environment [28,29,30]. Nokhbehsaim et al. [30] could show that the effect of mechanical loading on the pro-inflammatory effect of IL-1ß is dependent on the duration of mechanical loading. They concluded that the pro-inflammatory effect of mechanical loading decreases under long-lasting mechanical loading. Long et al. [31] described an anti-inflammatory effect of mechanical loading. They observed reduced COX-2 gene expression and PGE2-synthesis after cyclic mechanical loading of 3–6% elongation. In contrast to our study, they used heterogenic cells of the periodontal ligament (PDL). PDL cells are composed of different cell types such as fibroblasts, osteoblasts, and cementoblasts. In a previous study, we could show that, for example, fibroblasts and osteoblast react differently to mechanical loading [32]. This might be an explanation for the deviating results.

Roemer et al. analyzed the effect of mechanical loading and the presence of bacteria on PDL cells. Mechanical loading and bacteria separately led to a 5-fold increase of COX-2 gene expression. The combination of both increased the COX-2 expression in the cells about 56-fold [29]. Similar observations were made by Nogueira et al. 2014 [28]. They could demonstrate that mechanical loading and the presence of Fusobacterium nucleatum led to a higher gene expression of COX-2 than both factors alone. Cyclic mechanical strain increased the pro-inflammatory effect of F. nucleatum. Both studies used different methods to simulate inflammatory conditions, but the results are in agreement with our study that there might be a synergistic effect between a pro-inflammatory stimulus and mechanical loading.

The combination IL-1ß, mechanical loading and zoledronate led to the highest COX-2 gene expression and PGE2 synthesis. This shows a pro-inflammatory effect of zoledronate in the presence of IL-1ß and mechanical loading for the first time. According to current knowledge, there is only one previous in vitro study analyzing the inflammatory effect of mechanical loading and a bisphosphonate. Liu et al. observed an anti-inflammatory effect of clodronate regarding increased COX-2 and PGE2 values after mechanical compression [21]. They concluded that clodronate might also decrease the COX-2-dependent RANKL synthesis and thus reduce bone resorption as well as orthodontic tooth movement. In our study, clodronate showed neither pro-inflammatory nor anti-inflammatory effects. Further studies analyzing the interaction of different types of BP with mechanical loading are urgently needed due to the important clinical context regarding inflammation and necrosis of the alveolar bone.

IL-1ß increased MMP-8 synthesis compared to the control cells. This is in concordance with the results of previous studies. Abe et al. showed a concentration-dependent increase of MMP-8 gene expression through stimulation with IL-1ß [15]. Mechanical loading increased both MMP-8 and TIMP-1 synthesis. The ratio between these two factors is important for the remodeling of the extracellular matrix in the PDL. In a previous study, we could show that the strength of mechanical loading is relevant for the changes of the MMP-8/TIMP-1 ratio [14]. This ratio is crucial for orthodontic tooth movement and mastication forces to avoid PDL destruction which might be caused by an isolated MMP-8 increase.

Both clodronate and zoledronate inhibited the MMP-8 synthesis, but not the TIMP-1 synthesis. This is in concordance with the results of Buduneli et al. They observed in an in vivo study an inhibiting effect of alendronate regarding the MMP-8 gene expression in rats. After injection of LPS to induce periodontitis, the animals showed an increase of MMP-8. After the injection of alendronate, a nitrogen-containing bisphosphonate, the MMP-8 gene expression was reduced [33]. Teronen et al. also described an inhibiting effect of clodronate and zoledronate on MMP-8 in vivo. Patients with clodronate intake showed a 50% reduction in MMP-8 in the gingival sulcus fluid around implants [33,34,35]. They concluded that bisphosphonates might be helpful to prevent implant loss caused by increased MMP-8 values due to periodontitis. Thus, both factors, mechanical loading and the bisphosphonate, might have a positive impact on the MMP-8/TIMP-1 ratio to prevent loss of the extracellular matrix.

Conclusion

For the first time, we were able to demonstrate that cyclic mechanical loading abolishes the negative effects of zoledronate and IL-1ß in terms of the viability of HPdLF in vitro. Mechanical loading might stimulate and reactivate the cells. However, the combination of mechanical loading, inflammation, and zoledronate increased the production of COX-2-dependent PGE2, and the combination of these factors should therefore be avoided for clinical reasons. Zoledronate and clodronate inhibited MMP-8 synthesis caused by IL-1ß or mechanical loading, which might prevent loss of the extracellular matrix.

Clinical relevance

Periodontal inflammation should be treated initially before BP intake to prevent decreased cell viability of the periodontium and increased inflammation, which might be enhanced by the addition of mastication forces.