Gingival cell growth with antiresorptive treatment combined with corticosteroids or antiestrogen

Abstract Objectives Antiresorptive treatment has been shown to impair mucosal cell proliferation, migration, and viability. However, in the clinic, antiresorptives are often used in combination with other drugs. We studied the effect of antiresorptives combined with a corticosteroid or antiestrogen on oral mucosal keratinocytes and fibroblasts. Material and methods Human gingival keratinocyte and fibroblast cell lines were exposed to bisphosphonates (BPs) and denosumab in different concentrations and durations together with an antiestrogen or corticosteroid. Changes in cell viability, proliferation and migration after exposures were measured. Data were evaluated with hierarchical linear mixed model for repeated measurements. Results Bisphosphonate exposure suppressed keratinocyte and fibroblast cell viability, proliferation, and migration in a time‐dependent manner. Combining a corticosteroid or antiestrogen with BPs further increased this negative effect. Denosumab alone had a mild positive effect on keratinocyte and fibroblast growth. When denosumab was combined with a corticosteroid or antiestrogen, cell growth was suppressed. Conclusions Our results show that coexisting medications may increase the negative impact of BPs or denosumab on oral mucosal cells.

Medication-related osteonecrosis of the jaw (MRONJ) is a potential complication of BP/denosumab therapy and a growing problem (Ruggiero et al., 2014;Schiodt et al., 2015). However, MRONJ is a combination of osteonecrosis and ulceration of the mucosa. In order to cure osteonecrosis, keratinocytes and fibroblasts need to close the wound.
We evaluated the effect of antiresorptive treatment combined with corticosteroid or antiestrogen exposure on oral mucosal keratinocytes and fibroblasts. Based on previous epidemiological and in vitro studies, we hypothesized that BP/denosumab treatment of gingival mucosal cell lines will result in impaired growth and that antiestrogen or corticosteroid therapy will further affect cell growth. The purpose of this study was to show the additive causative effect of corticosteroids and antiestrogen on mucosal wound healing.
Growth media were changed every 3 days using Dulbecco's Modi-

| Assays of viability and proliferation
The cells were first exposed to either of the two BPs or denosumab for 24, 48, and 72 h with each concentration and for denosumab additionally for 144 h (day 6). These time points were determined from previous studies (Acil et al., 2012;McLeod et al., 2014;Otto et al., 2010;Ravosa et al., 2011;Reid, Bolland, & Grey, 2007;Soydan et al., 2015;Walter et al., 2011). The cells were then further exposed to antiestrogen or corticosteroid. The control conditions included Both assays were conducted in duplicate with four separate replicates.
The cells were monitored for 24 h. The migration process was automatically recorded with IncuCyte Zoom Software and relative wound density data for statistical analysis. The results were analyzed at the time points of 8, 16, and 24 h after "wounding" with relative wound density percentages, as recommended by the manufacturer. The assay was conducted in eight separate replicates.

| Statistics
All results were recorded in Microsoft Excel (Microsoft Corp., Washington, DC), with the exception of the wound healing results, which were analyzed with the appropriate IncuCyte software. Analyses were further performed using SAS System, version 9.4 for Windows (SAS Institute Inc., Cary, NC).
In the migration experiment, changes in absorbance and relative wound density were evaluated with hierarchical linear mixed model for repeated measurements. Drug exposure was handled as a between-factor and time point of measurement (8, 16, 24 h) as a within-factor. Interaction between drug exposure and time was also included in the model. The same modeling techniques were used to analyze viability and proliferation. Concentration, time, and drug exposure were included in the model. All statistical tests were performed as two-sided, with the significance level set at 0.05.

| Statement of ethical approval
This article does not contain any studies with human participants or animals performed by any of the authors, and therefore, no ethics approval was needed.

| Statement of informed consent
This article does not contain any studies with human participants or animals performed by any of the authors, and therefore, no consent was needed. Both BPs radically decreased the viability of fibroblasts in a timedependent manner (p < .0001) ( Table 1). Corticosteroid exposure alone suppressed fibroblast viability (all p < .0001). Decreased viability was detected in fibroblasts exposed to antiestrogen alone 24 h after exposure, but not at later time points. Combined antiestrogen and higher BP concentrations decreased fibroblast viability (all p < .0036) compared with either BP or antiestrogen exposure alone. The impact of combined BP and corticosteroid exposure on fibroblast viability was somewhat inconsistent, but always greater than that of BP or corticosteroid exposure alone (all p < .0156).

| Keratinocyte and fibroblast viability
Denosumab exposure resulted in inconsistently increased viability in fibroblasts (Table 1) (p < .05). Combined denosumab and corticosteroid exposure decreased fibroblast viability relative to denosumab exposure alone (all p < .01). Compared with corticosteroid alone, no significant differences emerged.

| Keratinocyte and fibroblast proliferation
Both BPs radically decreased keratinocyte proliferation in a timedependent manner (all p < .0001) and especially at higher concentrations (100 and 500 μg/mL) (Tables 1 and 2). Corticosteroid exposure alone also suppressed keratinocyte proliferation. Combined BP and corticosteroid exposure of keratinocytes further decreased proliferation after 24-48 h (p < .0006) ( Table 3). Keratinocytes exposed to Corticosteroid exposure alone also suppressed fibroblast proliferation. Antiestrogen-exposed fibroblasts had a lower proliferation than negative controls 24 h after exposure (p < .0001). The effect of combined corticosteroid and BP exposure on fibroblast proliferation was inconsistent, but always negative relative to BP or corticosteroid exposure alone. The effect of BPs with antiestrogen was inconsistent in fibroblast proliferation.
Denosumab exposure inconsistently increased fibroblast proliferation (p < .0336). There was no effect of combined denosumab and corticosteroid exposure on fibroblast proliferation. There was also no effect of combined exposure of denosumab and antiestrogen on fibroblast proliferation (all p > .44).

| Keratinocyte and fibroblast migration
Keratinocyte migration was decreased after BP exposure. It was 59% lower than in negative controls (p < .001) even 72 h after zolendronate (100 μmol/L) exposure. Migration was also decreased   Table 1).
Migration of only denosumab-exposed keratinocytes did not differ from negative controls. Combined denosumab-and corticosteroidexposed keratinocytes showed impaired migration after 8 h, but not at other time points (p < .04).
Migration of fibroblasts was impaired after BP exposure. Similarly, it was decreased after corticosteroid alone (all p < .001). Fibroblast migration was further impaired with combined BP and corticosteroid exposure (p < .0001).

| DISCUSSION
The main finding of this study is that combined treatment of corticosteroids or antiestrogens with BP/denosumab elicits an additive negative impact on oral mucosal cell viability, proliferation, and migration. This effect of combined exposures on oral mucosal cells has not been shown previously in the literature. The finding suggests that oral wound healing may be delayed in patients receiving these treatments.
Viability, proliferation, and migration of oral mucosal cells decreased after BP exposure. This study corroborates previous research, despite some differences in the experimental setup and analysis systems and the cell lines (Acil et al., 2012;Basso et al., 2013;Kim et al., 2011;Landesberg et al., 2008;Manzano-Moreno et al., 2019;McLeod et al., 2014;Pabst et al., 2012;Ravosa et al., 2011;Soydan et al., 2015;Taniguchi et al., 2019;Walter et al., 2011). In particular, BP concentrations have differed between studies (0.25-500 μmol/L). The high variance in the concentrations of zolendronic acid is due to the slow re-distribution in bone and the fact that its terminal half-life has not been adequately determined (Scheper et al., 2009). According to previous studies, the maximum serum concentration (C max ) of zolendronic acid is dose-dependent, ranging from 403 to 2252 ng/mL (Chen et al., 2002). Some theoretical models (Otto et al., 2010) and a small study of mandibular bone BP concentrations (Scheper et al., 2009) have reported a range of 0.4-126 μmol/L. How high the concentration is in mucosa after increased exposures due to tooth extraction or infection is debatable.
BP binds to the bone and during bone resorption it is released into the surrounding tissues in an uncontrolled fashion, resulting in a 100-fold increase in osseal BP concentrations (Baron et al., 2011;Chen et al., 2002). In our study, BPs had a strong negative effect on gingival epithelial cells in a time-and dose-dependent fashion.
Our results are consistent with those of Kuroshima et al. (2016) who reported that denosumab alone did not impair gingival fibroblast growth. Moreover, we observed that denosumab alone increased mucosal cell growth. Regarding dosages used to treat osteoporosis, denosumab maximum serum concentrations (C max ) of 6 μg/mL occur in 10 days, but in the case of malignancies the dosage is twofold higher and administered two to six times more often (Amgen Ltd, 2017). Based on this, we chose a wide variety of concentrations: 0, 5, 100, and 500 μmol/L for BP and 0, 6, 25, and 600 μg/mL for denosumab, and as the medium was changed the exposure did not continue.
Both corticosteroid and antiestrogen exposure alone impaired epithelial and/or fibroblast cell growth, and the changes were often additive with BP or denosumab. To our knowledge, the in vitro effects of BP/denosumab combined with antiestrogen or corticosteroids have not been previously investigated. It is well known that corticosteroids impair wound healing, suppressing immunological and inflammatory responses (Baxter & Forsham, 1972). Our results demonstrate that exposure of BP-affected cells to corticosteroids has a more severe negative impact on the cells. We propose that the corticosteroid effect on the RANKL/OPG system (Komori, 2016) together with BPinduced geranylgeraniol pyrophosphate suppression lead to further suppression of cell growth. This requires confirmation in future studies.
Our findings could offer a preliminary explanation for the previous epidemiological data of corticosteroid and antiestrogen therapy contributing to the risk of MRONJ (de Boissieu et al., 2016;McGowan et al., 2018;Ruggiero et al., 2014). In addition, the results may explain the observation of the prevalence of MRONJ being higher with prolonged BP therapies than with denosumab therapies (de Boissieu et al., 2016;McGowan et al., 2018;Ruggiero et al., 2014). Patients may have delayed wound healing after necrotic bone removal. Note: Comparison of Least Square Means (LSMeans) between bisphosphonates/denosumab exposed cells and bisphosphonate/denosumab with antiestrogen exposed cells. Abbreviations: D, denosumab; AE, antiestrogen; P, pamidronate; Z, zolendronate.
Denosumab alone did not affect cells negatively, but surprisingly increased cell growth. However, when administered in combination with corticosteroid, denosumab impaired epithelial cell growth. It is therefore possible for denosumab to cause alterations in oral soft tissues in the clinical setting. The frequency of MRONJ caused by zolendronate or denosumab has been reported to be 0.02-6.7% and 0.04-1.9%, respectively (Ruggiero et al., 2014). BP binds to the bone, exposing the oral mucosa to its negative effects over time, while denosumab does not negatively impact the mucosa, nor does the drug have long-term effects (Ruggiero et al., 2014).
The main strength of this study was that the evaluation of the effects of BP and denosumab on gingival cells was performed

CONFLICTS OF INTEREST
The authors declare no conflict of interest.

AUTHOR CONTRIBUTIONS
Authors TS and JR conceived the study concept. JR and HME developed the theory. HME conducted the experiments with laboratory assistance. EL performed the computations and verified the analytical methods, and HME reported the analytical findings.
TS and JR supervised the work. HME drafted the manuscript. All authors discussed the results and contributed to the final manuscript.

DATA AVAILABILITY STATEMENT
All data related to this manuscript is available upon request.