Selective serotonin reuptake inhibitors (SSRI) affect murine bone lineage cells

Abstract

Aims:

Data suggest pharmacological treatment of depression with selective serotonin reuptake inhibitors (SSRI) may impair bone health. Our group has previously modeled compromised craniofacial healing after treatment with sertraline, a commonly prescribed SSRI, and hypothesized potential culprits: alterations in bone cells, collagen, and/or inflammation. Here we interrogate bone lineage cell alterations due to sertraline treatment as a potential cause of the noted compromised bone healing.

Main methods:

Murine pre-osteoblast, pre-osteoclast, osteoblast, and osteoclast cells were treated with clinically relevant concentrations of the SSRI. Studies focused on serotonin pathway targets, cell viability, apoptosis, differentiation, and the osteoblast/osteoclast feedback loop.

Key Findings:

All cells studied express neurotransmitters (e.g. serotonin transporter, SLC6A4, SSRI target) and G-protein-coupled receptors associated with the serotonin pathway. Osteoclasts presented the greatest native expression of Slc6a4 with all cell types exhibiting decreases in Slc6a4 expression after SSRI treatment. Pre-osteoclasts exhibited alteration to their differentiation pathway after treatment. Pre-osteoblasts and osteoclasts showed reduced apoptosis after treatment but showed no significant differences in functional assays. RANKL:OPG mRNA and protein ratios were decreased in the osteoblast lineage. Osteoclast lineage cells treated with sertraline demonstrated diminished TRAP positive cells when pre-exposed to sertraline prior to RANKL-induced differentiation.

Significance:

These data suggest osteoclasts are a likely target of bone homeostasis disruption due to sertraline treatment, most potently through the osteoblast/clast feedback loop.

Introduction

Antidepressants, including selective serotonin reuptake inhibitors (SSRIs), are the 3^rd most commonly prescribed class of drugs in the United States with an estimated 1 in 10 Americans using them (Brinton et al., 2019, Joshi, 2018, Pratt et al., 2011). While the intended mechanism of action for SSRIs is to modulate the distribution of the central nervous system neurotransmitter serotonin, preclinical (Bab and Yirmiya, 2010, Battaglino et al., 2007, Bradaschia-Correa et al., 2017, Galli et al., 2013, Li et al., 2017, Ortuno et al., 2016, Warden et al., 2008, Weaver et al., 2018) and clinical data (Bab and Yirmiya, 2010, Brinton et al., 2019, Calarge et al., 2014, Hant and Bolster, 2016, Hodge et al., 2013, Joshi, 2018, Lanteigne et al., 2015, Rauma et al., 2016, Schweiger et al., 2018, Sheu et al., 2015, Vestergaard, 2009, Warden and Fuchs, 2016, Watts, 2017, Wu et al., 2014, Ziere et al., 2008) demonstrate systemic effects, including a negative impact on bone health. Specifically, data suggest a relationship between treatment with SSRIs and increased risk of fractures and reduced bone mineral densities (osteopenia and osteoporosis) (Richards et al., 2007, Ziere et al., 2008). Additionally, use of SSRIs have been linked to altered bone development, maintenance, remodeling, and healing (Bradaschia-Correa et al., 2017, Calarge et al., 2014, Cray et al., 2014, Durham et al., 2015, Fraher et al., 2016, Howie et al., 2018a, Karsenty and Yadav, 2011, Lanteigne et al., 2015, Sheu et al., 2015, Vestergaard, 2009, Wu et al., 2019, Wu et al., 2014). Although the preclinical and clinical observations are clear, research investigating the connection between SSRI use and bone health has not robustly extended to specific effects on the bone cells responsible for mineralized tissue homeostasis (Bradaschia-Correa et al., 2017, Hodge et al., 2013, Lanteigne et al., 2015, Sheu et al., 2015, Vestergaard, 2009, Wu et al., 2014).

SSRIs target Solute Carrier Family 6 Member 4 (SLC6A4) to disrupt reuptake of serotonin resulting in increased extracellular serotonin and sustained activation of pre- and post-synaptic serotonin receptors (Joshi, 2018). Previously it was thought that serotonin only acted as a neurotransmitter (serotonergic) important to autonomic and motor function, hormone secretion, and mood; in the gut for regulation of intestinal movements; and as a vasoconstrictor. However, there are now preclinical and clinical data pointing to systemic effects of serotonin and SSRI use (de Vernejoul et al., 2012), specifically the impairment of bone mass and bone maintenance over time(Bab and Yirmiya, 2010, Battaglino et al., 2007, Bradaschia-Correa et al., 2017, Brinton et al., 2019, Galli et al., 2013, Hant and Bolster, 2016, Hodge et al., 2013, Lanteigne et al., 2015, Ortuno et al., 2016, Rauma et al., 2016, Schweiger et al., 2018, Sheu et al., 2015, Warden and Fuchs, 2016, Warden et al., 2008, Watts, 2017, Weaver et al., 2018, Wu et al., 2014). In preclinical work it appears that mice lacking brain serotonin develop bone loss akin to osteopenia (Ducy and Karsenty, 2010, Inose et al., 2011, Karsenty and Yadav, 2011). Now clinical data suggest a similar relationship (Bab and Yirmiya, 2010, Brinton et al., 2019, Hant and Bolster, 2016, Rauma et al., 2016, Schweiger et al., 2018, Warden and Fuchs, 2016). Several serotonin receptors, serotonin, and serotonin transport protein have been identified in bone cells (Chabbi-Achengli et al., 2013, Gustafsson et al., 2006, Hodge et al., 2013, Lavoie et al., 2017),(Bliziotes et al., 2006, Cray et al., 2014),(Battaglino et al., 2004, Ortuno et al., 2016).

In a recent manuscript, our group reported on a repeatable model of SSRI (sertraline, trade name Zoloft) exposure in conjunction with simulated bone removal/injury reflective of clinical craniofacial surgical intervention (Howie et al., 2018a). These data suggest sertraline diminishes bone regeneration reflecting a drug specific effect not related to the clinical symptomology of depression. Interestingly in this model there was an observed direct effect on the bone lineage cells; specifically, maintenance of osteoblast activity with diminished osteoclast activity suggesting the critical homeostasis between these cell types may be disrupted. Other previous work has established that bone lineage cells express Slc6A4, the intended target of SSRIs including sertraline, as well as other related G-Protein Coupled Receptors (GPCRs) with binding affinities for serotonin (Apostu et al., 2017, Bab and Yirmiya, 2010, Battaglino et al., 2007, Bliziotes et al., 2006, Chabbi-Achengli et al., 2013, Cray et al., 2014, Elefteriou, 2008, Galli et al., 2013, Gustafsson et al., 2006, Hodge et al., 2013, Lavoie et al., 2017, Lychkov, 2011, Tanaka et al., 2015, Zhang and Drake, 2012). To date no data has robustly addressed whether SSRIs target these cells to alter local serotonin related gene expression, bone cell activity, or the relationship between osteoblast and osteoclast lineage cells.

As we are now observing high use of SSRIs, patient populations are at undue risk of negative effects on bone that are yet to be understood. Specifically, what is not clear is whether SSRIs have direct reproducible effects on bone cells. Here we utilized precursor and mature osteoblast and osteoclast lineage cells with exposure to a common SSRI, sertraline, to address these questions. Our central hypothesis is that SSRIs affect bone cells directly, indicating unintended effects of these drugs.

Materials and Methods

Cell lines and Growth Conditions

MC3T3-E1 cells (Subclone 4 CRL-2593) were obtained from ATCC (Manassas, VA) and maintained as recommended to produce pre-osteoblast cells for studies. Briefly, cells were maintained in a 75cm² flask using Alpha Eagle Minimum Essential Medium (Alpha-MEM, Lonza, Walkersville Inc., MD, BE02–002F) with 10% Fetal Bovine Serum (FBS, Atlanta Biologicals, Atlanta, GA, S11150H) and 1% Penicillin/Streptomycin (Lonza, 10k/10k 17–602E) until 85% confluent when they were moved to a 175cm² flask after trypsin-EDTA (0.1%, Gibco/Fisher Scientific, Hampton, NH, 15400–054) dissociation. Cells (Pre-osteoblasts) were subculture as necessary throughout the experimentation. To induce osteoblast differentiation 0.25mM ascorbic acid (Fisher Scientific, A61–25), 0.1 μm dexamethasone (Fisher Scientific, AC230300010), and 10mM β-glycerophosphate (Fisher Scientific, L03425) were added to standard Alpha-MEM media to produce osteogenic media (OM, Osteoblasts).

Raw 264.7 cells (ATCC, TIB-71) were obtained as a gift from Dr. Beth Lee, The Ohio State University, and used to produce pre-osteoclast cells. These cells were maintained in a 75cm² flask using Dulbecco’s Modified Eagle’s Medium (DMEM, Lonza, 12604F) with 10% FBS and 1% Penicillin/Streptomycin until 85% confluent, when they were moved to a 175cm² flask after cell scraper dissociation. Cells (Pre-osteoclasts) were sub-cultured as necessary throughout the experimentation. To induce osteoclast differentiation murine RANK Ligand 50ng/ml (Peprotech, Secaucus, NJ, 315–11) was added to the standard DMEM media (RANKL, Osteoclasts).

To induce Selective Serotonin Re-uptake Inhibitor (SSRI) exposure, Sertraline HCL 20MG/ML (Rising, Saddle Brook, NJ, NDC:59762-0067-01) was purchased in liquid form from The Ohio State University Pharmacy. A low dose (34.2 ng/ml) or a high dose (342 ng/ml) was suspended in media mimicking clinical circulating drug levels (Howie et al., 2018b, NIH, 2002).

PCR for Neurotransmitter and G-Protein Coupled Receptors Identification

Pre-osteoblasts, osteoblasts (OM), Pre-osteoclasts, and osteoclast (RANKL), were seeded in triplicate at a density of 100,000 cells per well in triplicate in 6 well culture plates and treated with control un-supplemented media, and appropriate differentiation media as defined above supplemented with low (34.2 ng/ml) and high (342 ng/ml) dose sertraline. After 7 days (Pre-osteoblasts, Pre-osteoclasts, and osteoclasts) or 14 days (osteoblasts) RNA was isolated using the OMEGA bio-tek E.Z.N.A. Total RNA kit 1 (Omega Bio-tek, Norcross, GA, R6834-02) according to manufacturer’s protocol. Quality and quantity of RNA was assessed using a Synergy Hi Microplate reader and a Take3 Microvolume Plate (BioTek, Winooski, VT). Complimentary DNA synthesis was performed using Quanta qScript cDNA Synthesis reagents following manufacturers protocol (Quanta Biosciences, Beverly, MA, 95047-025). Presence of neurotransmitters and GPCRs was determined via PCR using cDNA, designed primers from Integrated DNA Technologies (Coralville, IA) (Table 1), Platinum Taq DNA Polymerase (Fisher Scientific, 100021273), and separation on 1.5% agarose gels employing beta actin (Forward GCAGGAGTACGATGAGTCCG / Reverse ACGCAGCTCAGTAACAGTCC) as a control. Each assay was repeated three independent times.

Table 1.

Expression of Serotonin Pathway Targets

Target	Pre-osteoblasts	Osteoblasts	Pre-osteoclasts	Osteoclasts	Primer Forward / Reverse
Dat					AAGATCTGCCCTGTCCTGAAAG / CATCGATCCACACAGATGCCTC
Sert	✕	✕	✕	✕	GACTCCTCCCCTCTAAGCCA / AGTTTGTAATGGGCCCGGAG
Net	✕	✕		✕	ACTCAGGCGCTCCTTTTCTT / GCACCTGCGGATTCATTC
Vmat1				✕	CGGAGTGCCTCACATCAAGA / CCACCACAGTGAGCAGCATA
Vamt2					TGTGACCAACACGACTGTCC / CAAGAGGAGCCGATTCCCTG
Pmat		✕			TTCTCGCTGCTAATGGGCAT / GTGGCTGTTTGAAAGCAGCT
5ht1a	✕	✕	✕	✕	TACGTGAACAAGAGGACGCC / CACTCGATGCACCTCGATCA
5ht1b	✕	✕	✕	✕	ACCCTTCTTCTGGCGTCAAG / AGGGCAGCCAACACACAATA
5ht1d					TGGGTGTTTCCGGATCTCCCTAT / AGGACAGACCGGAGGTAGGA
5ht1f	✕	✕	✕	✕	TGGAGCTTTCTACATCCCGC / GTAGTGGCTGCTTTGCGTTC
5ht2a					GAACCAACCTCTCCTGCGAA / ATGGTCCACACCGCAATGAT
5ht2b	✕	✕		✕	CTCGGGGGTGAATCCTCTGA / CCTGCTCATCACCCTCTCTCA
5ht2c					TGGCCTATTGGTTTGGCAGT / TTGCTACATACCGGTCCAGC
5ht3	✕	✕			ACACCGAGGAGGAACTGGCTAATAT / GACCACATTAGAGGGGTTGACTGGC
5ht4	✕	✕			ATGGTCAACAAGCCCTATGC / AGGAAGGCACGTCTGAAAGA
5ht5					CAGGACCTACAGGGTTGTGG / GCTGAGCATAGGCAGTAAAGG
5ht6	✕	✕	✕	✕	GTGGACCTCTCACAGTGGTG / GACCTGGTCAGTTCATGGGG

Quantitative Real-Time PCR for Slc6a4, Osteoblast, and Osteoclast Targets

To quantify expression of specific targets, cDNA was subjected to quantitative PCR using Applied Biosystems TaqMan Gene Expression Master mix and targeted TaqMan gene expression assays for serotonin specific receptor Slc6A4 (Mm00439391_m1), osteoblast specific targets Runx2, Alp, Msx2, Dlx5, Osx, Bglap, Col1a1, Col1a2, and osteoclast specific targets Mcp-1, Rantes, Rank, Nfatc1, Ctsk, Mmp9, Calcr, Trap (Table 2). Data were normalized to 18S (Mm03928990_g1) ribosomal RNA expression by ΔCT. Quantitative data were compared for gene expression changes due to treatment with sertraline (low 34.2 ng/ml, high 342 ng/ml) by ΔΔCT methodology. Previously published statistical analysis methodology was used to determine differences for gene expression after sertraline treatment for osteoblast and osteoclast related targets of interest (Yuan et al., 2006). Differences were considered significant if p≤0.05. Data are presented as inverse of the ΔCT to allow for direct visual comparison between cell types for serotonin receptor (Fig 1A) and by fold change compared to control (no exposure) to highlight changes due to exposure (Fig 1B–D). Error bars represent the variability about Fold Change for visual representation.

Table 2:

Bone Cell Lineage Targets

	Target	TaqMan Gene Expression Assay
Osteoblast	Runx2	Mm00501584_m1
	Alp	Mm00475834_m1
	Msx2	Mm00442992_m1
	Dlx5	Mm00438430_m1
	Osx	Mm04209856_m1
	Bglap	Mm01741771_g1
	Col1a1	Mm00801666_g1
	Col1a2	Mm00483888_m1
Osteoclast	Mcp-1	Mm00442991_m1
	Rantes	Mm01302427_m1
	Rank	Mm00437132_m1
	Nfatc1	Mm01265944_m1
	Ctsk	Mm00484039_m1
	Mmp9	Mm00442991_m1
	Calcr	Mm00432282_m1
	Trap	Mm00475698_m1

Fig 1. Sertraline Exposure Modulates Serotonin Receptor and Bone Cell Lineage Related Gene Expression.

A. Relative baseline expression of serotonin transport receptor Slc6A4 in bone related cell types indicates the most expression in mature osteoclasts. B. Expression of serotonin transport receptor Slc6A4 can be modulated as compared to control in osteoblast and pre-osteoclast cells with sertraline. C. Expression of genes associated with osteogenesis were modulated with sertraline treatment in pre-osteoblasts and less so in mature osteoblasts. D. Osteoclast related gene expression was modulated in both pre-osteoclasts and osteoclasts with sertraline treatment. Low Sertraline = 34.2 ng/ml High Sertraline = 342 ng/ml n=3 assays/cell type *p≤0.05 **p≤0.01 ***p≤0.001 as compared to control

Osteoblast and Osteoclast Functional Assays

To determine the specific effect of sertraline on bone lineage cells, pre-osteoblasts, osteoblasts, pre-osteoclasts and osteoclasts were seeded in triplicate into 96 well plates at a density of 4,000 cells per well. Cells were treated with standard, osteoblast differentiation (OM), or osteoclast differentiation (RANKL) media alone, or supplemented with low (34.2 ng/ml) or high (342 ng/ml) dose sertraline for 7 days. Cell viability (proliferation) was assessed with the colorimetric MTS (Promega, Madison, WI, G3581) assay and apoptosis was assessed using the Apo-ONE Caspase-3/7 Assay (Promega, G7791) per manufacturer’s protocol using a Gen5 plate reader (BioTek). Each assay was repeated three times.

For differentiated osteoblasts, alkaline phosphatase activity was assessed using the Sigma-Aldrich Alkaline Phosphatase Activity kit (Sigma-Aldrich, St. Louis, MO, 86C-1KT) according to manufacturer protocol. Images were captured using an Olympus inverted scope and CellSens imaging software (Olympus, Miami, FL, TH4–100). Quantitative Alkaline Phosphatase Activity was assessed using SigmaFast p-nitrophenyl phosphate substrate on lysed cells (Sigma-Aldrich, N2765,T8790). Kinetic absorbance was read after 30 minutes incubation using a Gen5 plate reader (BioTek). Each assay was repeated three times.

For differentiated osteoclasts, Tartrate-resistant acid phosphatase (TRAP) was used to identify mature, multi-nucleated osteoclasts. Fixed cells were stained with TRAP stain containing 0.3 mg/ml Fast Red (Sigma, F3381) in TRAP Buffer pH5.0 (Acetate Buffer (0.2M sodium acetate, 0.2M acetic acid) with 0.3M sodium tartrate, 10mg/ml Napthol AS, and 0.01% Triton-X). Wells were imaged using an Olympus inverted scope and CellSens imaging software (Olympus, Miami, FL, TH4–100). Positive (red, multinucleated) osteoclasts were counted in each 10x image by two independent observers and averaged. Triplicate wells were also averaged, and treatments were compared using Analysis of Variance (ANOVA). Each assay was repeated three times.

RANKL:OPG Studies

The ratio of RANKL to osteoprotegerin (OPG) mRNA expression was compared for pre-osteoblasts and osteoblasts using quantitative PCR on synthesized cDNA (Section 4.3), Applied Biosystems TaqMan Gene Expression Master mix, and targeted TaqMan gene expression assays for RANKL (Mm00441906_m1) and OPG (Mm00435454_m1).

Protein was isolated from pre-osteoblasts, osteoblasts, pre-osteoclasts, and osteoclasts treated for 7 days with control or sertraline low (34.2 ng/ml) or high (342 ng/ml) dose supplemented media. Cells were seeded in triplicate at a density of 300,000 cells per well and supernatant was collected before cells were washed with cold Phosphate Buffered Saline (PBS, HyClone, Fisher Scientific, SH30256.01) and protein was extracted using radioimmunoprecipitation buffer (RIPA, G Biosciences, St. Louis, MO, 786–490) for 30 mins at 4°C with agitation. Supernatant and cell lysate was collected and used for RANKL (AbCam, Cambridge, MA, Ab100749) and OPG (AbCam, Ab100733) ELISA according to manufacturer protocol. The ratio of RANKL and OPG was calculated for each cell type for each treatment in both cell lysates and supernatant. Each assay was repeated three times.

Pre-Exposed Differentiation Experiments

Pre-Osteoblasts and pre-osteoclasts were pretreated for 7 days with control media or media supplemented with low (34.2 ng/ml) or high (342 ng/ml) dose sertraline. After pretreatment, cells were seeded in triplicate, and treatment continued for another 7 days when assays for differentiation were performed as above. Each assay was repeated three times.

For co-culture studies, pre-osteoblasts were pretreated with sertraline low (34.2 ng/ml) or high (342 ng/ml) dose and osteogenic media (OM) for 7 days, then were seeded in triplicate with differentiated osteoclasts at a ratio of 4 osteoblasts to 1 osteoclast and treated with sertraline low (34.2 ng/ml) or high (342 ng/ml) dose for another 7 days. The difference in osteoclast to osteoblast seeding density was based on proliferative potential of the different cell types. TRAP assay followed using the same protocol as outlined above and was repeated three times.

Statistics

Standard two-way Analysis of Variance (ANOVA) with post-hoc Bonferroni analyses were conducted for all comparisons based on means where appropriate. Violations of homogeneity of variance and normality were corrected through transformations of the data where needed. Differences were considered significant when p≤0.05. All discrete data are presented as means ± standard error of the mean (SEM). All statistical analyses were completed using SPSS 23.0 (IBM, Armonk, NY, USA).

Results

Neurotransmitters and G-protein Coupled Receptors on Bone Lineage Cells

In order to determine the sensitivity of bone related cell types to serotonin modulation, Polymerase Chain Reaction (PCR) was used to identify the presence of neurotransmitters and GPCRs associated with the serotonin pathway (Table 1). Both osteoblast and osteoclast lineage cells demonstrated presence of serotonin pathways targets. Specifically, Sert, 5ht1a, 5ht1b, 5ht1f, and 5ht6 were expressed in all bone lineage cell types (pre-osteoblasts, osteoblasts, pre-osteoclasts, and osteoclasts). Vmat1 was only present in mature osteoclasts, and Pmat was only present in mature osteoblasts. Net and 5ht2b were present in all cell types expect pre-osteoclasts. 5ht3 and 5ht4 were only identified within osteoblast lineage cells (pre-osteoblasts and osteoblasts).

Sertraline Modulates Serotonin Receptor Slc6A4 and Bone Cell Gene Expression

Serotonin transport receptor Slc6A4 was most robustly expressed in osteoclasts as compared to all other bone cell types (p<0.001) (Fig 1A). Expression of this receptor was modulated with administration of sertraline in both mature osteoblasts (p=0.020) and pre-osteoclasts (p=0.043) (Fig 1B).

Treatment with sertraline decreased expression of Msx2 (p=0.05) and Bglap (p=0.003 (low, 34.2 ng/ml), p=0.018 (high, 342 ng/ml)) in pre-osteoblast cells while increasing expression of Osx (p=0.019) and Col1a2 (p=.009) in those cells as well. A lesser effect of sertraline treatment was noted for mature osteoblasts with high dose (342 ng/ml) sertraline only increasing expression of Alp (p=0.05) (Fig 1C).

Pre-osteoclasts demonstrated reduced expression of Mcp-1 (p=0.004), and Rank (p=0.01) with high dose (342 ng/ml) sertraline and an increase in Ctsk (p=0.05) expression with low dose (34.2 ng/ml) sertraline. Expression of Calcr was undetectable in pre-osteoclasts. Administration of high dose (342 ng/ml) sertraline increased expression of Mmp9 (p=.003) and Trap (p=.002) in mature osteoclasts (Fig 1D).

Sertraline Affects Bone Lineage Cell Function

Sertraline had cell type specific effects on bone lineage cells. Specifically, treatment with low (34.2 ng/ml) and high (342 ng/ml) dose reduced apoptosis in pre-osteoblast cells as compared to control (p=0.032, p=0.001 respectively) (Fig 2B), while having no effect on proliferation or differentiation of osteoblast lineage cells (Fig 2A,C). Mature osteoclast cells were sensitive to sertraline, displaying a reduced level of apoptosis with high dose (342 ng/ml) treatment as compared to control (p=0.006) and low dose (34.2 ng/ml) (p=0.016) (Fig 2E). Sertraline had no effect on proliferation or differentiation of osteoclast lineage cells (Fig 2D,F).

Fig 2. Sertraline Exposure Affects Bone Lineage Cell Function.

A. Proliferation (MTS) of osteoblast lineage cells with sertraline treatment. B. Apoptosis is reduced in pre-osteoblast cells with low and high dose sertraline. C. Representative cells (4x Magnification) stained for alkaline phosphatase and stain quantification indicate that differentiation of osteoblasts is not affected by sertraline exposure. D. Osteoclast lineage cell proliferation is not affected by sertraline. E. High dose sertraline reduced apoptosis in mature osteoclasts. F. Representative cells (4x Magnification) stained for tartrate-resistant acid phosphatase (TRAP) and stain quantification confirm that sertraline does not affect osteoclast differentiation. OD = Optical Density RFU = Relative Fluorescence Units Low Sertraline = 34.2 ng/ml High Sertraline = 342 ng/ml n=3 assays/cell type *p≤0.05 **p≤0.01 as compared to control, $p≤0.05 as compared to low sertraline scale=0.5 mm

Sertraline Alters Maintenance of the RANKL:OPG Axis

A balance of RANKL and OPG is necessary for maintenance of the feedback loop connecting bone formation cells (osteoblast lineage) and bone resorption cells (osteoclast lineage). Increasing concentrations of sertraline treatment reduced the expression of the RANKL:OPG ratio in pre-osteoblasts compared to control (p=0.037 low 32.4 ng/ml, p=0.002 high 342 ng/ml). Low dose (34.2 ng/ml) sertraline also reduced the RANKL:OPG ratio in mature osteoblasts (p=0.03) (Fig 3A). This relationship was not similarly identified at the protein level in either cell lysates (Fig 3B) or the supernatant collected from pre-osteoblasts or osteoblasts (Fig 3C).

Fig 3. Sertraline Exposure Affects the Symbiosis of Bone Resorption and Bone Formation.

A. Expression of RANKL:OPG is altered in pre-osteoblasts and osteoblasts at the level of mRNA. B-C. The ratio of RANKL:OPG is not altered at the protein level in either cell lysate (B) or supernatant (C) from osteoblast lineage cells. Low Sertraline = 34.2 ng/ml High Sertraline = 342 ng/ml n=3 assays/cell type *p≤0.05 **p≤0.01 as compared to control

Pre-Exposure with Sertraline Primes Bone Lineage Cells for Altered Differentiation

Pre-treating osteoblast lineage cells with high dose (342 ng/ml) sertraline resulted in an increase in osteogenic differentiation as measured by alkaline phosphatase activity in osteoblast cells as compared to control (p=0.003) and low dose (34.2 ng/ml) sertraline (p=0.004) (Fig 4A). Alternatively, pretreatment of osteoclast lineage cells showed so such effect on differentiation (Fig 4B). Co-culture of mature osteoblasts with osteoclasts demonstrated the propensity of sertraline pre-treatment to reduce osteoclast differentiation at both low (34.2 ng/ml, p<0.001) and high (342 ng/ml, p<0.001) doses demonstrating a stepwise reduction between low (34.2 ng/ml) and high (342 ng/ml) dose sertraline (p=0.027) (Fig 4C).

Fig 4. Sertraline Pre-treatment Affects Bone Lineage Cell Differentiation.

A. Quantification and representative images of alkaline phosphatase activity in osteoblast lineage cells pre-treated for 7 days with sertraline. B. Quantification and representative images (4x Magnification) of tartrate-resistant acid phosphatase positive osteoclasts pre-treated for 7 days with sertraline. The reduction in osteoclast with sertraline treatment is not statistically significant. C. Quantification and representative images of tartrate-resistant acid phosphatase positive differentiated osteoclasts co-cultured with mature osteoblasts pre-treated for 7 days with sertraline. Low Sertraline = 34.2 ng/ml High Sertraline = 342 ng/ml n=3 assays/cell type **p≤0.01 ***p≤0.001 as compared to control $p≤0.05 as compared to low sertraline scale=0.5 mm

Discussion

Here we utilized murine bone cell lines to determine if delivery of a commonly prescribed SSRI, sertraline, altered bone cell activity. Our hypothesis was supported by data our group collected demonstrating decreased bone wound healing and osteoclast activity in a murine model treated with this antidepressant drug (Howie et al., 2018a). Data garnered here suggest firstly that osteoclast cells have much greater expression of the pharmacological target of SSRIs, Slc6a4, compared to their precursor cells or osteoblast lineage cells. Macrophages, which are closely related to the RAW 264.7 cell line used here as osteoclast precursors and to create differentiated osteoclasts, are also known to express serotonin transporters (Durairaj et al., 2015). Here the significantly greater expression of SSRI target receptor in differentiated osteoclast highlights their potential as a cell population susceptible to modification by this drug treatment. Further, Slc6a4 appears to be at least partially modulated at the message level showing decreased expression after sertraline treatment. Although osteoblast and osteoclast gene targets did show some alterations after sertraline treatment it was the pre-osteoclasts that demonstrated decreases in targets important in recruitment and fusion (Mcp-1) and differentiation and activation (Rank) suggesting SSRIs including sertraline may decrease osteoclastogenesis (Hodge et al., 2013, Khan et al., 2014). Note targets for osteoblasts and osteoclasts were chosen based on the known temporal expression patterns related to the function of these cell types.

We further interrogated cell activity of the bone cell lines after sertraline treatment and except for decreases in apoptosis in pre-osteoblasts and mature osteoclasts, no differences were observed for differentiation or proliferation of these cells. Although at first perplexing, the modality of treatment (concurrent) was the likely culprit in lack of downstream alterations. Additionally, though we did note a decrease in apoptotic activity there was not a concomitant increase in proliferation or cell number indicating that perhaps cells were dying via some other process.

Support for our assertion that concurrent treatment and differentiation may be driving the lack of downstream alterations came from our assessment of the gene expression ratios for Rankl:Opg and the relative observed differences for the RANKL:OPG protein ratios. Thus, the critical experiment necessitated pre-treatment of the cells with the SSRI sertraline and subsequent downstream interrogation of differentiation. These data support our previous in vivo work that suggested similar or greater levels of osteoblast activity after SSRI treatment, but an absolute decrease in osteoclast activity. This is interesting in that it is well documented that osteoblasts regulate osteoclast activity. Thus we would have expected that an increase in osteoblast activity would drive a similar increase in osteoclast activity, however we observed what is potentially an uncoupling of these cell types as a result of sertraline treatment(Boyce and Xing, 2008). In these experiments this decoupling was accentuated by the co-culture of osteoblast and osteoclast cell lines allowing for the interpretation that SSRIs including sertraline can specifically target osteoclast precursors to disrupt differentiation, but also may negatively impact the osteoblast/osteoclast feedback loop causing a further observable decrease in osteoclastogenesis (Hodge et al., 2013, Martin and Sims, 2015).

In the context of the literature, these data are concerning. Although no large-scale study has dissected the mechanism of direct effects of SSRIs on bone, research has partially addressed the effects of serotonin on bone health (Brinton et al., 2019, Hant and Bolster, 2016, Rauma et al., 2016, Warden and Fuchs, 2016, Watts, 2017, Weaver et al., 2018). As described above, studies targeting brain serotonin suggested a decrease in bone densities in a murine model (Ducy and Karsenty, 2010, Inose et al., 2011, Karsenty and Yadav, 2011). Further, due to the self-reported nature of clinical data, longitudinal clinical investigations are unable to adequately control for the obvious confounders present in large datasets of patients who have poor bone health or a fracture and are on these drugs; confounders include poor overall health and clinical depression (Brinton et al., 2019, Richards et al., 2007, Williams et al., 2015, Ziere et al., 2008). Further as serotonin is still not widely considered a hormone, (strictly defined as a protein produced by an organ that has effects away from this organ) but rather a neurotransmitter (Fouquet et al., 2019, Wu et al., 2019), research has been slow to identify and characterize what the systemic effects of serotonin modulation may be and how these drugs may have a system-wide impact. Data included here suggest that what is generally identified as a side effect of these drugs, may rather be an on-target effect in an area away from the intended location of effect (synaptic clefts of the central nervous system). Overall, these data support the interpretation that circulating SSRIs including sertraline may directly target osteoblast and osteoclast lineage cells to disrupt cellular activity and homeostasis of these cell types (Hodge et al., 2013).

Conclusions

Considering these data, future analyses will focus on the mechanistic pathway of effect by which osteoclasts directly and the osteoblast/osteoclast feedback loop are altered due to sertraline and other SSRI treatment. As precursors for both cell types also contribute to bone maintenance, injury repair, and remodeling (Fierro et al., 2017) attempts should be made to understand if these cells are diminished in response or in activity at remodeling and injury sites in vivo. Although much data has focused on the serotonin/SSRI/osteoblast axis very little work has incorporated osteoclasts in this paradigm (de Vernejoul et al., 2012, Fraher et al., 2016). These data suggest osteoclasts may be a more vulnerable cell type in these relationships.