Citalopram and Sertraline Exposure Compromises Embryonic Bone Development

ORIGINAL ARTICLE Citalopram and sertraline exposure compromises embryonic bone development

D Fraher1,2,6, JM Hodge2,3,6, FM Collier2, JS McMillan3, RL Kennedy3, M Ellis1,2, GC Nicholson3, K Walder1,2, S Dodd2, M Berk2,4, JA Pasco2,5, LJ Williams2,7 and Y Gibert1,2,7

Selective serotonin reuptake inhibitors (SSRIs) are the most commonly prescribed treatments for depression and, as a class of drugs, are among the most used medications in the world. Concern regarding possible effects of SSRI treatment on fetal development has arisen recently as studies have suggested a link between maternal SSRI use and an increase in birth defects such as persistent pulmonary hypertension, seizures and craniosynostosis. Furthermore, SSRI exposure in adults is associated with decreased bone mineral density and increased fracture risk, and serotonin receptors are expressed in human osteoblasts and osteoclasts. To determine possible effects of SSRI exposure on developing bone, we treated both zebrafish, during embryonic development, and human mesenchymal stem cells (MSCs), during differentiation into osteoblasts, with the two most prescribed SSRIs, citalopram and sertraline. SSRI treatment in zebrafish decreased bone mineralization, visualized by alizarin red staining and decreased the expression of mature osteoblast-specific markers during embryogenesis. Furthermore, we showed that this inhibition was not associated with increased apoptosis. In differentiating human MSCs, we observed a decrease in osteoblast activity that was associated with a decrease in expression of the osteoblast-specific genes Runx2, Sparc and Spp1, measured with quantitative real- time PCR (qRT-PCR). Similar to the developing zebrafish, no increase in expression of the apoptotic marker Caspase 3 was observed. Therefore, we propose that SSRIs inhibit bone development by affecting osteoblast maturation during embryonic development and MSC differentiation. These results highlight the need to further investigate the risks of SSRI use during pregnancy in exposing unborn babies to potential skeletal abnormalities.

Molecular Psychiatry (2016) 21, 656–664; doi:10.1038/mp.2015.135; published online 8 September 2015

INTRODUCTION women.14 To date, an understanding of the impact of SSRIs on The safety of maternal selective serotonin reuptake inhibitors fetal bone development is lacking. (SSRI) use during pregnancy is of increasing concern with recent The focus of this research, therefore, was to investigate effects studies suggesting associations between SSRIs and increased risk on bone formation during embryonic development following of persistent pulmonary hypertension, low birth weight, preterm exposure to SSRIs. Owing to the ethical and practical limitations of birth, seizures and even infant death.1–4 Decreased birth length studying maternal SSRI use on fetal bone development in humans, fi caused by SSRI use during pregnancy may suggest an effect on our group has used the zebra sh (Danio rerio) as a model to study fi skeletal development.5,6 In adults, SSRI exposure has been shown effects of SSRIs on embryonic bone formation. The zebra sh has been used as a vertebrate model organism for embryogenesis for to increase fracture risk, impair bone metabolism and decrease 15 – over 30 years. The external and rapid development of the bone mineral density.7 10 We recently showed that the human zebrafish makes it an excellent model for developmental biology, bone-forming cells, osteoblasts, and the bone-resorbing cells, – including bone development.16 20 Importantly, regulators of bone osteoclasts, express serotonin receptors and the serotonin formation, cellular organization and bone homeostasis are highly transporter, the target of SSRI action, suggesting a possible fi 21,22 11 conserved between zebra sh and humans. It is also known mechanism for the effect of SSRIs on bone homeostasis. that 84% of human disease-related genes are expressed in Furthermore, SSRIs can accumulate in the bone marrow for zebrafish,23 and these features have allowed us to visualize bone extended periods at concentrations much higher than those 12 formation in vivo following SSRI exposure. To complement our found in blood or the brain. Several SSRIs are currently embryonic in vivo studies, we utilized an in vitro model of human considered as options for use for the treatment of depression 13 osteoblast differentiation from early precursor cells that employs during pregnancy, including citalopram and sertraline. human adipose tissue-derived mesenchymal stem cells (AT-MSCs). Citalopram and sertraline account for ~ 70% of all SSRI prescrip- These multipotent precursor cells provide an excellent source of tions in the United States of America, including for pregnant osteoblast progenitor cells that can differentiate into mature

1Metabolic Genetic Diseases Laboratory, Metabolic Research Unit, School of Medicine, Deakin University, Geelong, VIC, Australia; 2IMPACT and MMR Strategic Research Centres, School of Medicine, Deakin University, Geelong, VIC, Australia; 3Barwon Biomedical Research, University Hospital, Geelong, VIC, Australia; 4Orygen, The National Centre of Excellence in Youth Mental Health and the Centre for Youth Mental Health, Department of Psychiatry, Florey Institute of Neuroscience and Mental Health, University of Melbourne, Parkville, VIC, Australia and 5Department of Medicine, Northwest Academic Centre, The University of Melbourne, St Albans, VIC, Australia. Correspondence: Dr Y Gibert, IMPACT and MMR Strategic Research Centres, School of Medicine, Deakin University, 75 Pidgons Road, Waurn Ponds, Geelong, VIC 3217, Australia. E-mail: [email protected] 6These authors contributed equally to this work. 7These authors are co-senior authors. Received 18 September 2014; revised 15 May 2015; accepted 14 July 2015; published online 8 September 2015 Sertraline and citalopram in bone formation D Fraher et al 657 osteoblasts. To address the knowledge gap regarding the effects of resuspended and filtered through a 100-μm cell strainer to remove SSRI exposure on early bone formation, we have investigated their remaining tissue debris. Cells were pelleted by centrifugation and seeded actions in both an in vivo model of embryogenesis, as well as an at 1 × 106 in tissue culture flasks in basal medium and incubated at 37 °C in fi in vitro human osteoblast model. a humidi ed atmosphere with 5% CO2. Cells were passaged by treatment with 0.025% trypsin/EDTA in phosphate-buffered saline and diluted 1:10 in DMEM/FBS. MSCs were employed in assays after five passages. MATERIALS AND METHODS Ethics statement Differentiation of MSC in medium containing osteogenic factors All zebrafish studies were approved by the Deakin University Animal MSC (104 cells per well) were seeded in 6-mm diameter culture wells in Welfare Committee (81-2011). Human adipose tissue samples were DMEM/FBS and cultured overnight. For MSC differentiation, cells were then obtained with informed, written consent from healthy donors and all cultured in osteogenic medium (DMEM/FBS containing 100 nM dexa- protocols were approved by the Barwon Health Human Research Ethics methasone, 10 mM β-glycerophosphate and 100 mM ascorbate-2-phos- Committee (05–69). phate) in the absence or presence of citalopram or sertraline, and assessed for alkaline phosphatase (ALP) activity. Animal husbandry Zebrafish were reared and staged at 28.5 °C according to Kimmel et al.24 ALP activity assay To determine cellular ALP activity, cells were lysed in 0.1% Triton X-100 for 30 min at room temperature. A pre-warmed solution containing Pharmacological treatment − 10 mg ml 1 p-nitrophenylphosphate in 10% v/v diethanolamine buffer Citalopram (Sigma-Aldrich, St Louis, MO, USA) was stored at − 20 °C in a containing 0.5 mM MgCl2 (pH 9.8) was then added to the lysates and stock concentration of 24.7 mM dissolved in ethanol. Sertraline (Sigma- optical density of samples was assessed using a Tecan Genios Pro − Aldrich) was stored at 20 °C in a stock concentration of 29.2 mM dissolved photospectrometer, at 410 nm at 2.5 min intervals for 30 min. Results were in dimethyl sulphoxide. For treatment, embryos were placed in 50-ml converted to standard international units, equivalent to the conversion by tubes containing 25 ml of E3 embryo medium, supplemented with 0.003% 25 ALP of 1 mM of pNPP to p-nitrophenyl (pNP) per minute. A standard curve phenyl-thiourea to prevent pigmentation. SSRIs were added directly to was generated by serially diluting 1 mM pNP in diethanolamine buffer and the embryonic medium and embryos were allowed to develop in an data presented as relative standard international unit. incubator at 28.5 °C. Measurement of cell viability/apoptosis Bone, cartilage and apoptosis staining Cell viability and apoptosis were assessed by propidium iodide (PI) Alizarin red skeletal staining was performed as previously described by 26 staining and Annexin V (ANV) measurement, respectively. Previously Walker et al. Von Kossa staining was performed as described by Felber untreated MSCs (4 × 105) were freshly prepared and then treated with et al.27 Live embryos were stained for apoptotic cells with the vital dye 28 citalopram or sertraline for 48 h. Cells were dissociated by trypsin digest acridine orange (Sigma-Aldrich) as described by Sassi-Messai et al. (0.025%), recovered for 30 min and then washed in binding buffer Cleaved caspase 3 labeling was performed with BD Pharmigen caspase 3, (10 mM Hepes/NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl ) and stained with Active Form antibody (BD Biosciences, North Ryde, Sydney, Australia) and 2 a combination of ANV–FITC (5 μM per tube) and PI (10 μM) for 15 min in the Donkey anti-Rabbit IgG (H+L) secondary antibody, Alexa Fluor 488 (Life dark. A further 400 μl of binding buffer was added to each sample and the Technologies, Mulgrave, Melbourne, Australia) as described by Sorrells percentage of viable (ANV − /PI − ) and apoptotic cells (ANV+/PI − ) et al.29 Alcian blue stain was adapted from Schilling et al.30 in which a 0.1% was determined using flow cytometry analysis (FACSCalibur, CELLQuest alcian blue solution was prepared in 0.37% HCl/70% ethanol. software, BD Biosciences, Franklin Lakes, NJ, USA).

Whole-mount in situ hybridization Real-time RT-PCR analysis Embryos were fixed in 4% paraformaldehyde in phosphate-buffered saline Total RNA was isolated by lysing cells in Trizol and using the illustra overnight at 4 °C and then transferred and stored in 100% methanol. RNAspin Mini Kit (GE Healthcare, Melbourne, Victoria, Australia) and Whole-mount in situ hybridization using digoxigenin-labeled riboprobes reverse-transcribed using the Superscript III First Strand Synthesis was performed as described in Thisse and Thisse.31 The following probes SuperMix system (Life Technologies) as per the manufacturer’s instructions. were used: runx2a,32 runx2b,32 osterix,33 col10a1,34 sparc,35 spp1,35 sox9a,32 For each SSRI dose/time point, mRNA was extracted separately from three slc6a4a36 and caspase 3 (primers sequences: forward, 5ʹ-GCACTGACGT wells of a 12-well plate and the real-time PCR was performed in duplicate. AGATGCAGGA-3ʹ; reverse, 5ʹ-GCCAATTGCTACCAGATTCC-3ʹ). Images were To quantify the expression of human osteoblastic genes (Runx2, ALP, Sparc taken using the Axioskop 2 Imager (Zeiss, Carl Zeiss AG, Oberkochen, and Spp1), human apoptotic genes (Bax and Caspase 3) and the zebrafish Germany) and processed with the Adobe Photoshop software (San Jose, apoptotic gene (caspase 3), we employed quantitative real-time PCR CA, USA). (qRT-PCR) analysis of the cDNA in a 7500 Fast Real-Time PCR System (Applied Biosystems), using TaqMan Gene Expression Assays (Applied Cell media and reagents Biosystems, Hs00231692 (Runx2), Hs01029144 (ALP), Hs00234160 (Sparc), Dulbecco’s Modified Eagle’s Medium (DMEM), penicillin/streptomycin Hs00959010 (Spp1); Hs01016552 (Bax), Hs00234387 (Caspase 3) and solutions, paraformaldehyde, collagenase type-1, p-nitrophenylphosphate, zebrafish caspase 3.37 Relative gene expression units were determined − ΔCt p-nitrophenyl, diethanolamine, dexamethasone and dimethyl sulphoxide using the formula 2 x1000, where ΔCt values represent the difference were purchased from Sigma-Aldrich. β-glycerophosphate disodium salt between the Ct of the gene of interest and Beta-actin (amplified using was purchased from Merck Millipore (Kilsyth, Victoria, Australia). L-Ascorbic Taqman chemistry with forward primer (5′-GACAGGATGCAGAAGGA acid phosphate was purchased from NovaChem Pty (Melbourne, Victoria, GATTACT-3’), reverse primer (5′-TGATCCACATCTGCTGGAAGGT-3′) and Australia). Non-essential amino acids (100 × ) and fetal bovine serum (FBS) probe (5′(FAM)-ATCATTGCTCCTCCTGAGCGCAAGTACTC-(TAMRA)-3′). were purchased from Bovogen (Melbourne, Victoria, Australia). All other reagents were of analytical grade. Area measurement and statistical analysis Pictures of stained zebrafish embryos were mounted on a slide in a ventral Isolation and culture of adipose tissue-derived MSC orientation using an Axioskop 2 Imager (Zeiss). The area of stain was Human adipose tissue was collected from elective abdominoplasty measured using Cell Sens Dimension (Olympus, Tokyo, Japan). Statistical surgery. To isolate MSC, tissue was teased from blood vessels, minced analysis was performed using IBM SPSS Statistics version 21 (Armonk, NY, with a scalpel blade and digested for 30–45 min with 0.075% collagenase USA). Comparisons were made with an independent samples t-test. For at 37 °C with gentle agitation. Enzyme activity was neutralized with basal human samples, data are expressed as the mean ± s.e.m., where applicable. medium (DMEM containing 10% FBS, 50 U ml − 1 penicillin and 50 mg ml − 1 Differences between groups were determined by one-way analysis of streptomycin) and the cells were centrifuged at 1200 g for 10 min, variance, followed by Fisher’s multiple comparison test.

© 2016 Macmillan Publishers Limited Molecular Psychiatry (2016), 656 – 664 Sertraline and citalopram in bone formation D Fraher et al 658 RESULTS of slc6a4a was not detected in these areas at the early time points Sertraline and citalopram inhibit bone development 50 and 62 hpf (data not shown). To assess the effects of citalopram and sertraline on bone in vivo, To further investigate the effects of citalopram and sertraline on zebrafish embryos were treated from 36 h post fertilization (hpf), a bone development, we examined the expression of genes developmental stage during embryogenesis when bone formation involved in early (runx2a and runx2b), intermediate (osterix, also begins,21 through to 130 hpf. At 130 hpf, zebrafish embryos still called sp7) and mature (collagen 10a1, osteonectin (sparc) and osteopontin (spp1)) stages of osteoblast differentiation in fish. rely on nutrition from the oocyte-provided yolk, preventing fi possible differential nutritional effects on development, whereas Zebra sh embryos were treated with SSRIs from 36 hpf, when pre-osteoblast cells begin expressing runx2b in forming skeletal bone mineralization progressed enough to observe any effects 21 due to chemical exposure. To determine the dosage and toxicity condensations. We allowed the embryos to develop until fi 50 hpf and performed whole-mount in situ hybridization with pro le of each drug, embryos were exposed to increasing 4 incremental doses until toxic effects (smaller or misshapen bodies) runx2a and runx2b probes (n 10). Citalopram and sertraline had no effect on runx2a or runx2b expression compared with were observed. Embryos were treated with 10, 20 and 50 μM control embryos, notably in the area of the forming opercle citalopram and sertraline (data not shown). Both citalopram and – – sertraline treatments caused growth retardation in the embryos at (Figures 2a d, e h arrowheads). The expression of osterix (osx), a μ μ later marker of osteoblast differentiation, was also unaffected in 50 M, whereas treatment at 20 M was toxic for citalopram- – treated embryos, but not sertraline-treated embryos (data not embryos treated from 36 to 62 hpf (Figures 2i l, arrowheads, n410). Expression of collagen 10a1 (col10a1) which, in zebrafish, is shown). Therefore, we chose a dose range of 5–15 μM for first expressed from 63 hpf onwards in osteoblasts,21 was citalopram and 10–30 μM for sertraline (Figures 1c–h, k–p). Both citalopram and sertraline treatments decreased tissue decreased in the area of the developing opercle (10/12 citalopram treated, 14/18 sertraline treated) in embryos exposed to SSRI mineralization in developing bones, observed by staining with the – matrix calcification dye alizarin red, in a dose-dependent manner treatment from 36 to 72 hpf (Figures 2m p, arrowheads). The – – expression of two bone matrix proteins, sparc and spp1, was (Figures 1a h, i p). This inhibition of bone formation was notably 4 present in the vertebral column (arrows) and, in the opercle, the reduced in the opercle (n 10, all treated embryos had reduc- fi tions) in citalopram- and sertraline-treated embryos (Figures 2q–t, posterior part of the operculum, a speci ed bone that covers and – protects the gills in fish (arrowheads). Magnification of the opercle u x, arrowheads). Therefore, citalopram and sertraline treatments bone shows decreased calcification upon exposure to citalopram appear only to affect the expression of genes present later in osteoblast differentiation. and sertraline at 15 and 30 μM, respectively, compared with control embryos (Figures 1q–t). Quantification of the amount of To determine whether SSRI treatment had a transient effect on calcified tissue in the whole embryo was consistent with the bone development, only inhibiting late bone development, we effects observed in the opercle, with reductions in the total area of treated embryos at an early time range, from 36 to 56 hpf. Embryos alizarin red staining at higher SSRI concentrations (Figures 1u, v). treated with citalopram and sertraline from 36 to 56 hpf showed To assess bone mineralization under SSRI exposure, embryos no differences in alizarin red staining at 130 hpf compared with controls (Supplementary Figure 2A and D). In addition, embryos were treated with 15 μM citalopram and 30 μM sertraline from 36 to 150 hpf (a time point more suitable for observing von Kossa treated from 36 to 56 hpf showed no differences in the expression of col10a1, sparc and spp1 in the opercle at 72 hpf (Supplementary staining) and stained with von Kossa stain to label matrix Figure 2 E–H,I-L,M-P, arrowheads). Therefore, the effects of phosphate (Figures 1w, z). Both citalopram- and sertraline-treated citalopram and sertraline on bone development appear to result embryos showed reduced mineralization compared with controls, from the action of SSRIs during late osteoblast development. notably in the opercle (Figures 1w–z, arrowheads and insets). Owing to the reduction of bone calcification and mineralization that we observed from citalopram and sertraline exposure, we The effects of citalopram and sertraline on bone development are monitored cartilage development in embryos exposed to the SSRIs. independent of apoptosis in zebrafish embryos We were unable to detect changes in the expression of sox9a,a As cell death could cause reduced mineralization, calcification and chondrogenic marker, or differences in cartilage morphology, gene expression, we investigated necrosis and apoptosis in labeled with alcian blue, a cartilage stain (Supplementary Figure treated embryos. Bright-field images showed no evidence of 1A–P). Therefore, the effects of citalopram and sertraline are tissue necrosis (opaque cells) from 15 μM citalopram and 30 μM specific to bone development and are not chondrogenesis. sertraline treatments when embryos were exposed from 36 to As the mode of action for SSRIs is to inhibit proper functioning 50 hpf (data not shown), 36 to 72 hpf (data not shown) and 36 to of the serotonin transporter, we performed whole-mount in situ 96 hpf (Figures 3a–d). To assess apoptosis, zebrafish embryos were hybridization to identify expression of the transporter in stained with acridine orange, a fluorescent dye marking apoptotic embryonic zebrafish bone. In zebrafish, there are two identified cells, following citalopram and sertraline exposure from 36 to serotonin transporters, slc6a4a and slc6a4b.36 At 72 hpf, we 96 hpf (Figures 3e–h, arrowheads). There was no evidence of detected expression of slc6a4a in the opercle (Figure 1a', b' apoptosis occurring in areas of developing bone and no arrowheads) and cleithrum (Figure 1c' bracket) bones. Expression detectable increase in overall apoptosis. Acridine orange staining

Figure 1. Citalopram and sertraline exposure decreased bone calcification and mineralization during zebrafish embryogenesis. Embryos were treated with citalopram (c–h) or sertraline (k–p) from 36 to 130 h post fertilization (hpf) and stained with alizarin red after fixation at 130 hpf. Embryos were treated with 5 μM (c and d), 10 μM (e and f)or15μM (g and h) citalopram and 10 μM (k and l), 20 μM (m and n)or30μM (o and p) sertraline. Citalopram- and sertraline-treated embryos showed a dose-dependent decrease in bone calcification, specifically in the opercle (arrowheads), compared with control siblings (a, b, i and j). The area of staining in (pixel)2 of the opercle was measured in control fish (q, close up of the opercle in r), citalopram-treated (s) and sertraline-treated embryos (t). Both citalopram and sertraline treatments showed a decrease in alizarin red staining compared with control fish. Total area of alizarin red staining was measured in citalopram- and sertraline-treated fish (u and v). Both treatments showed a dose-dependent decrease in area of alizarin red staining with increasing concentration. Values are shown as relative to control, *significantly different to control (Po0.05). Von Kossa staining, to measure bone mineralization, of 15 μM citalopram (x)or 30 μM sertraline (z)-treated fish from 36 to 150 hpf showed reduced mineralization of the opercle (arrowheads, insets) compared with controls (w and y). Expression of the serotonin transporter, slc6a4a, is shown in wild-type embryos at 72 hpf (a'–c') located in the opercle (arrowheads) and cleithrum (bracket).

Molecular Psychiatry (2016), 656 – 664 © 2016 Macmillan Publishers Limited Sertraline and citalopram in bone formation D Fraher et al 659 of embryos treated from 36 to 50 hpf and 36 to 72 hpf also not change in embryos treated with citalopram or sertraline from showed no increase in apoptosis (Supplementary Figure 3A–D, 36 to 52 hpf or from 36 to 72 hpf (Figures 3i–p, blue arrowheads). E–H). In addition to no change in acridine orange staining, qRT-PCR analyses from whole-embryo mRNA conﬁrmed that expression of caspase 3 (a key caspase involved in apoptosis) did caspase 3 mRNA levels were not increased in citalopram- or

Figure 2. Citalopram and sertraline exposure affected the expression of only late zebraﬁsh osteoblast markers. Whole-mount in situ hybridization was performed on embryos treated with 15 μM citalopram and 30 μM sertraline. Embryos treated from 36 to 56 h post fertilization (hpf) (c, d, g and h) did not show a difference in runx2b or runx2a expression in the developing opercle (arrowhead) compared with controls (a, b, e and f). Embryos treated from 36 to 62 hpf (k and l) did not show a difference in osx expression compared with controls (i and j). col10a1 expression in embryos treated from 36 to 72 hpf (o and p) was reduced, prominently in the opercle (arrowheads), compared with controls (m and n). The expression of sparc (q–t) and spp1 (u–x) was reduced in embryos treated from 36 to 72 hpf.

sertraline-treated embryos (Figure 3q). We further evaluated at 13 days in AT-MSCs that were exposed to sertraline from 0 to 7, apoptosis by using antibody labeling for cleaved caspase 3, the 8 to 13 and 0 to 13 days (Figure 4 aviii). Therefore, in the human activated form of the caspase 3 protein. Embryos treated with model, sertraline shows a higher capacity to inhibit osteoblast citalopram and sertraline from 36 to 96 hpf showed no increase in differentiation than citalopram, although AT-MSCs exposed to cleaved caspase 3 labeling compared with controls (Figures 3r–u, citalopram did show modest decreases in ALP activity and arrows). Similarly, embryos treated from 36 to 50 hpf or 36 to 72 expression at 7 and 14 days, respectively. To determine the role hpf did not show increases in cleaved caspase 3 labeling of apoptosis in decreasing osteoblast differentiation, Annexin V (Supplementary Figure 3 I-L,M-P). Therefore, cell death from staining following 48 h of exposure was performed. Sertraline apoptosis or necrosis does not appear to be the cause of treatment induced a minor, dose-dependent increase in the decreased bone development. percentage of apoptotic cells, whereas citalopram had no effect (Figure 4 bi). Furthermore, neither sertraline nor citalopram treatment affected expression levels of apoptotic markers Bax Osteoblast differentiation from human AT-MSCs is inhibited by and Caspase 3 compared with control levels at 14 days (Figure 4 citalopram and sertraline Bii–iv), whereas a significant increase in Bax mRNA was observed Both citalopram and sertraline decreased ALP activity, a marker of after 7 days of sertraline exposure (Figure 4 bii). osteoblast lineage, in a dose-dependent manner in AT-MSC cultured in osteogenic media for 7 days (Figure 4 ai). Sertraline was more potent than citalopram, with decreases in ALP activity of DISCUSSION 55 and 16% at 10 μM, respectively (Po0.001). AT-MSCs were SSRI use during pregnancy – – exposed to citalopram and sertraline at intervals of 0 7, 8 14 and The central focus of these experiments was to study the effects of – 0 14 days and ALP activity was assessed at 14 days (Figure 4 aii). SSRIs specifically during embryonic bone formation. Depression is Sertraline-treated AT-MSCs showed decreased ALP activity when common during pregnancy, with a study showing the prevalence exposed from 8 to 14 and 0 to 14 days. Consistent with this of depressive symptoms to be 17%, with another reporting that finding, expression of ALP mRNA was reduced by 42% after 7 days 3.8% of women take an SSRI during pregnancy.38,39 There is and by 48% after 14 days of treatment with sertraline, and by 29% existing evidence suggesting that SSRI use by pregnant mothers is when treated with citalopram after 14 days (Figure 4 aiii, iv). There associated with birth defects. Chambers et al. showed an was a trend for decreased expression of Runx2 mRNA at 7 days of association between persistent pulmonary hypertension of the treatment; however, this only reached significance in the case of newborn and SSRI exposure after completion of the 20th week of sertraline at 14 days, with a 53% decrease (Figure 4 av, vi). In gestation.1 Wen et al. examined a retrospective cohort of 972 addition, Sparc expression was significantly decreased at 13 days pregnant women who had received at least one SSRI prescription when AT-MSCs were treated with sertraline from 0 to 7 and 0 to within the year before delivery compared with a cohort of 3878 13 days (Figure 4 avii). AT-MSCs treated from 8 to 13 days showed women who had not taken SSRIs.2 The researchers found that a decrease in Sparc expression approaching significance (P = 0.054; infants from mothers who had taken SSRIs were at a higher risk for Figure 4 avii). Furthermore, the expression of Spp1 was decreased low birth weight, preterm birth, seizures and even infant death.2

Figure 3. Citalopram and sertraline did not affect apoptosis in the developing zebrafish. Embryos were treated with 15 μM citalopram or 30 μM sertraline. Bright-field images of embryos treated from 36 to 96 h post fertilization (hpf) showed no increase in cell necrosis (a–d), which would be observed as opaque tissue. Embryos treated from 36 to 96 hpf were stained with acridine orange dye to label apoptotic cells (e–h, arrowheads) and showed comparable levels of labeled cells with controls (open arrowheads label neuromasts, not to be confused as apoptotic cells). Whole-mount in situ hybridization was performed with a caspase 3 probe on embryos treated from 36 to 52 hpf (k and l) and 36–72 hpf (o and p). Treated embryos showed no difference in expression (blue arrowheads) compared with controls (i, j, m and n). Quantitative real-time PCR shows that there was no significant difference in relative gene expression of caspase 3 in treated embryos compared with the controls (q). Cleaved caspase 3 antibody labeling of treated embryos (s and u, arrows) was consistent with levels seen in control embryos (r and t).

Furthermore, Alwan et al.40 showed an association between showed that SSRIs appear to affect the later stages of osteoblast maternal SSRI use and infant anencephaly, craniosynostosis and differentiation, as only expression of later zebrafish osteoblast omphalocele.40 These findings have raised concerns regarding the markers, col10a1, sparc and spp1, were reduced by SSRI's exposure use of SSRIs during pregnancy, reflected by these data. (Figure 4c). Interestingly, the onset of expression of these genes coincides with the initiation of the expression of the serotonin fi SSRIs and bone formation transporter slc6a4a in bone tissue during zebra sh embryogenesis – Although evident in the elderly, osteoporosis is more accurately a (Figures 1a' c'). It is possible that the presence of the transporter is disorder of the lifespan, beginning with variable rates of bone required for the inhibitory actions of the SSRIs to manifest and formation in youth, and driven by variable rates of bone loss in support the observation that early osteoblast progression is adulthood. Although there has been a rise in research examining the unaffected by SSRIs. Therefore, on the basis of these results we effects of SSRIs on the bone, few studies have looked at the effects on can hypothesize that maternal SSRI use may put the fetus at risk of bone during development. One study by Dubnov-Raz et al. analyzed bone malformations in the later stages of pregnancy, depending thebonedensityofinfantsbetween1and4daysoflifewhose on the timing of expression of the serotonin transporter in human mothers had been using SSRIs during gestation.6 Although a differ- bone development, when fetal osteoblasts are maturing. ence in the bone density of infants exposed to SSRIs was not evident, these infants were shorter and had smaller heads, indicating a Concentration of SSRIs and bone formation possible defect in skeletal growth. A limitation of that study was that The concentrations of citalopram and sertraline administered to ultrasound, rather than the gold standard, dual-energy X-ray absorpti- the zebrafish embryos in this study were higher than the ometry, was used to determine bone mineral density in order to recommended maximum dosage prescribed for use in humans, avoid potential negative health impacts of radiation on the children. which is up to 40 mg per day for citalopram and up to 200 mg Our study was able to observe the initial stages of bone per day for sertraline.42,43 At this dosage, the citalopram serum formation in a developing organism in vivo. Our data indicated that levels can be as high as 250 ng ml − 1 (equivalent to a concentration both citalopram and sertraline decreased bone mineralization of 0.8 μM), whereas sertraline serum levels can be as high as − 1 43,44 during zebrafish embryogenesis and osteoblast activity during 200 ng ml (equivalent to a concentration of 0.7 μM). Whereas differentiation. It is reasonable to speculate that these effects could both of these concentrations are still much lower than the hold true in human fetuses, which could present future health risks. concentrations used in our experiments, a previous study has A negative effect on bone maturation in the fetus could have shown that the SSRI, fluvoxamine, can accumulate in the bone lifelong implications as it has been shown that skeletal growth in marrow at much higher concentrations than are present in adulthood is influenced by intrauterine development.41 Our results circulating serum.12 Fluvoxamine levels were shown to be as high

Figure 4. Effect of citalopram and sertraline on osteoblastic maturation and induction of apoptosis in human MSC. Human MSCs were cultured in osteogenic media in the presence of citalopram and sertraline. (ai, ii) alkaline phosphatase (ALP) activity at 7 and 14 days; (iii, iv) ALP gene expression at 7 and 14 days; (v, vi) Runx2 gene expression at 7 and 14 days; (vii) Sparc gene expression at 13 days; (viii) Spp1 gene expression at 13 days. (bi) Percentage of apoptotic cells (Annexin V+ve) after 48 h of treatment and assessment using flow cytometry; (ii, iii) Bax gene expression at 7 and 14 days; (iv, v) Caspase 3 gene expression at 7 and 14 days. Data are expressed as mean ± s.e.m. where applicable. RT-PCR values are shown relative to control values. Statistical significance is indicated thus: citalopram different to control, ★; sertraline different to control, #; citalopram different to sertraline, μ; Po0.05. (c) Model of SSRI action on bone formation during zebrafish development. MSC, mesenchymal stem cell; RT-PCR, real-time PCR; SSRI, selective serotonin reuptake inhibitor.

11 as 80 μM in the bone marrow of human patients, which is more citalopram did not. These findings are in the context of previous than five times the concentration of citalopram and more than 2.5 reports of citalopram acting as an anti-apoptotic agent in vivo in times the concentration of sertraline used to treat zebrafish brain structures and that sertraline was a known inducer of – embryos in this study. Furthermore, fluvoxamine remained at an apoptosis.45 47 On the basis of these results, it was expected that elevated level, above 60 μM, in the bone marrow even as the citalopram would not induce apoptosis; however, it was anticipated circulating level fell below detectable levels following cessation of that sertraline might, thus preventing proper bone mineralization. treatment. These data demonstrate that SSRIs have the capacity to However, this was not observed in our in vivo model, as both bioaccumulate in the bone marrow. Therefore, SSRI use, even sertraline and citalopram exposure did not induce apoptosis during transiently during pregnancy, might have deleterious effects on embryogenesis. In comparison, during bone differentiation from bone formation, as SSRIs remain at high levels in the bone marrow human MSCs, a high level of sertraline exposure led to a small even after their levels have dropped in blood. As the bone marrow increase in apoptosis after 48-h exposure and an increase in is essential for bone formation via osteoblast activity, it is tempting expression of the apoptosis regulator Bax after 7-day exposure, to hypothesize that SSRI use by a pregnant mother during the time whereas after 14-day exposure Bax expression was unchanged from that the bone marrow has developed in the fetus may affect control. However, the apoptotic changes seen at the earlier time points were not accompanied by alterations to expression of osteoblast activity and bone mineralization. Caspase 3 and may have been transient. In the human model of osteoblast differentiation from MSCs, we observed a strong SSRI and apoptosis reduction in ALP expression after 14-day treatment of both We recently reported that most SSRIs (sertraline, fluvoxamine, sertraline and citalopram. Furthermore, we observed a reduction fluoxetine and paroxetine) induced apoptosis in bone cells, whereas in Runx2 expression at 14 days of exposure in sertraline-treated cells

Molecular Psychiatry (2016), 656 – 664 © 2016 Macmillan Publishers Limited Sertraline and citalopram in bone formation D Fraher et al 663 and a trend of reduction in citalopram-treated cells. These 5 Davidson S, Prokonov D, Taler M, Maayan R, Harell D, Gil-Ad I et al. Effect of reductions at 14 days of treatment cannot be caused by apoptosis exposure to selective serotonin reuptake inhibitors in utero on fetal growth: as neither citalopram- nor sertraline-treated cells exhibited a higher potential role for the IGF-I and HPA axes. Pediatr Res 2009; 65:236–241. level of Bax or Caspase 3 after 14 days of exposure. Furthermore, 6 Dubnov-Raz G, Hemilä H, Vurembrand Y, Kuint J, Maayan-Metzger A. Maternal use citalopram-treated cells did not exhibit an increase in Bax expression of selective serotonin reuptake inhibitors during pregnancy and neonatal bone 88 – but nevertheless displayed a strong reduction in ALP mRNA after density. Early Hum Dev 2012; :191 194. 7 Chau K, Atkinson SA, Taylor VH. Are selective serotonin reuptake inhibitors a 14 days of differentiation. We therefore conclude from these data 2012 fi secondary cause of low bone density? J Osteop 2012; : 323061. that during zebra sh embryogenesis and human osteoblast 8 Haney EM, Chan BK, Diem SJ, Ensrud KE, Cauley JA, Barrett-Connor E et al. differentiation, apoptosis is not the biological process by which Association of low bone mineral density with selective serotonin reuptake inhib- decreases in bone mineralization and in ALP activity occur. itor use by older men. Arch Int Med 2007; 167: 1246–1251. During zebrafish embryogenesis, osteoblasts are first detected by 9 Richards JB, Papaioannou A, Adachi JD, Joseph L, Whitson HE, Prior JC et al. Effect gene expression at around 36 hpf,21 whereas bone mineralization is of selective serotonin reuptake inhibitors on the risk of fracture. Arch Int Med first detected at around 96 hpf.48 During this time, osteoblasts 2007; 167: 188–194. rapidly mineralize bone to produce the mature skeleton. The goal 10 Williams LJ, Henry MJ, Berk M, Dodd S, Jacka FN, Kotowicz MA et al. Selective of this rapid production is to form a supportive skeleton fast serotonin reuptake inhibitor use and bone mineral density in women with a history of depression. Int Clin Psychopharmacol 2008; 23:84–87. enough to provide rigidity for the developing organism. Apoptosis 11 Hodge JM, Wang Y, Berk M, Collier FM, Fernandes TJ, Constable MJ et al. Selective at these developmental stages would be counterproductive to the serotonin reuptake inhibitors inhibit human osteoclast and osteoblast formation skeletal building process. Therefore, it is possible that, although and function. Biol Psychiatry 2013; 74:32–39. 11 SSRIs can induce apoptosis in adult osteoblasts, the developing 12 Bolo N, Hode Y, Macher J-P. Long-term sequestration of fluorinated compounds in bone cells are more resistant to apoptosis – a feature previously tissues after fluvoxamine or fluoxetine treatment: a fluorine magnetic resonance reported in other stem/progenitor cells this apoptotic stimulus is spectroscopy study in vivo. Magn Reson Mater Phys Biol Med 2004; 16:268–276. overcome in developing bone cells.49 This could have potentially 13 MayoClinicStaff. Antidepressants: Safe During Pregnancy?: Mayo Clinic; 2012 prevented the SSRIs from inducing apoptosis in our models of bone (cited 4 August 2014). Available from http://www.mayoclinic.org/healthy-living/ formation during early embryogenesis. pregnancy-week-by-week/in-depth/antidepressants/art-20046420. 14 Grohol JM. Top 25 Psychiatric Medication Prescriptions for 2013: Psych Central; 2014 (cited 2014 24 July 2014). Available from http://psychcentral.com/lib/top-25- CONCLUSION psychiatric-medication-prescriptions-for-2013/00019543. 15 Lieschke GJ, Currie PD. Animal models of human disease: zebrafish swim To our knowledge, this is the first study to demonstrate that SSRIs into view. Nat Rev Genet 2007; 8:353–367. directly influence bone formation during vertebrate embryonic 16 Zon LI. Zebrafish: a new model for human disease. Genome Res 1999; 9:99–100. development and in the differentiation of MSCs into osteoblasts. 17 Spoorendonk KM, Peterson-Maduro J, Renn J, Trowe T, Kranenbarg S, Winkler C Our results confirm previous studies suggesting that sertraline et al. Retinoic acid and Cyp26b1 are critical regulators of osteogenesis in the axial may have higher bone toxicity than citalopram, especially during skeleton. Development 2008; 135: 3765–3774. human MSC differentiation into osteoblasts. Given the high 18 Knopf F, Hammond C, Chekuru A, Kurth T, Hans S, Weber CW et al. Bone fi fi prevalence of bone-related disease,50 the use of SSRIs to treat regenerates via dedifferentiation of osteoblasts in the zebra sh n. Dev Cell 2011; 20:713–724. psychiatric disorders during pregnancy, and the knowledge that 19 Fisher S, Halpern ME. Patterning the zebrafish axial skeleton requires early chordin SSRI agents affect bone formation, our study has overt clinical function. Nat Genet 1999; 23: 442–446. implications. If supported by human clinical and epidemiological 20 Barrett R, Chappell C, Quick M, Fleming A. A rapid, high content, in vivo model of data, it may be time to re-evaluate the current guidelines for SSRI glucocorticoid‐induced osteoporosis. Biotechnol J 2006; 1: 651–655. prescriptions, incorporating recommendations for SSRI use during 21 Li N, Felber K, Elks P, Croucher P, Roehl HH. Tracking gene expression during pregnancy in order to prevent potential bone defects in babies zebrafish osteoblast differentiation. Dev Dyn 2009; 238:459–466. born to mothers experiencing mood disorders. 22 Flores MV, Tsang VWK, Hu W, Kalev-Zylinska M, Postlethwait J, Crosier P et al. Duplicate zebrafish runx2 orthologues are expressed in developing skeletal ele- ments. Gene Expr Patterns 2004; 4:573–581. CONFLICT OF INTEREST 23 Howe K, Clark MD, Torroja CF, Torrance J, Berthelot C, Muffato M et al. The zebrafish reference genome sequence and its relationship to the human genome. The authors declare no conflict of interest. Nature 2013; 496:498–503. 24 Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF. Stages of embryonic ACKNOWLEDGMENTS development of the zebrafish. Dev Dyn 1995; 203: 253–310. 25 Geisler R. Zebrafish: A Practical Approach (The Practical Approach Series, 261). We thank the staff members of the Deakin University zebrafish facility for providing Oxford Univ Press: New York, NY, USA, 2002, pp 175–212. excellent husbandry care. YG is supported by funding from the Molecular and 26 Walker M, Kimmel C. A two-color acid-free cartilage and bone stain for Medical Research Strategic Centre at Deakin University. MB is supported by a NHMRC zebrafish larvae. Biotech Histochem 2007; 82:23–28. Senior Principal Research Fellowship (1059660). LJW is supported by a NHMRC Career 27 Felber K, Croucher P, Roehl HH. Hedgehog signalling is required for perichondral development Fellowship (1064272). osteoblast differentiation in zebrafish. Mech Dev 2011; 128:141–152. 28 Sassi-Messai S, Gibert Y, Bernard L, Nishio S-I, Lagneau KFF, Molina J et al. The phytoestrogen genistein affects zebrafish development through two different REFERENCES pathways. PLoS One 2009; 4: e4935. 1 Chambers CD, Hernandez-Diaz S, Van Marter LJ, Werler MM, Louik C, Jones KL 29 Sorrells S, Toruno C, Stewart RA, Jette C. Analysis of apoptosis in zebrafish et al. Selective serotonin-reuptake inhibitors and risk of persistent pulmonary embryos by whole-mount immunofluorescence to detect activated Caspase 3. J hypertension of the newborn. N Engl J Med 2006; 354:579–587. Vis Exp 2013; 82: e51060. 2 Wen SW, Yang Q, Garner P, Fraser W, Olatunbosun O, Nimrod C et al. Selective 30 Schilling TF, Piotrowski T, Grandel H, Brand M, Heisenberg C-P, Jiang Y-J et al. Jaw serotonin reuptake inhibitors and adverse pregnancy outcomes. Am J Obstet and branchial arch mutants in zebrafish I: branchial arches. Development 1996; Gynecol 2006; 194: 961–966. 123:329–344. 3 Andrade SE, Raebel MA, Brown J, Lane K, Livingston J, Boudreau D et al. Use of 31 Thisse C, Thisse B. High-resolution in situ hybridization to whole-mount zebrafish antidepressant medications during pregnancy: a multisite study. Am J Obstet embryos. Nat Protoc 2007; 3:59–69. Gynecol 2008; 198: 194 e1–194 e5. 32 Eames BF, Amores A, Yan Y-L, Postlethwait JH. Evolution of the osteoblast: ske- 4 Laine K, Heikkinen T, Ekblad U, Kero P. Effects of exposure to selective serotonin letogenesis in gar and zebrafish. BMC. Evol Biol 2012; 12:27. reuptake inhibitors during pregnancy on serotonergic symptoms in newborns 33 Hammond CL, Schulte-Merker S. Two populations of endochondral osteoblasts and cord blood monoamine and prolactin concentrations. Arch Gen Psychiatry with differential sensitivity to Hedgehog signalling. Development 2009; 136: 2003; 60: 720–726. 3991–4000.

© 2016 Macmillan Publishers Limited Molecular Psychiatry (2016), 656 – 664 Sertraline and citalopram in bone formation D Fraher et al 664 34 Albertson RC, Yan Y-L, Titus TA, Pisano E, Vacchi M, Yelick PC et al. Molecular 43 UniversityofIowa. Sertraline (Zoloft) Drug Level The University of Iowa Depart- pedomorphism underlies craniofacial skeletal evolution in Antarctic notothenioid ment of Pathology Laboratory Services Handbook: The University of Iowa Health fishes. BMC. Evol Biol 2010; 10:4. Care; 2013 (cited 3 August 2014). Available fromhttps://www.healthcare.uiowa. 35 Laue K, Pogoda H-M, Daniel PB, van Haeringen A, Alanay Y, von Ameln S et al. edu/path_handbook/handbook/test2628.html. Craniosynostosis and multiple skeletal anomalies in humans and zebrafish result 44 MayoClinic. Test Catalog: Citalopram, Serum: Mayo Medical Laboratories; 2014 from a defect in the localized degradation of retinoic acid. Am J Hum Genet 2011; (cited 1 August 2014). Available from http://www.mayomedicallaboratories.com/ 89 – : 595 606. test-catalog/Clinical+and+Interpretive/83730. ‐ 36 Norton WH, Folchert A, Bally Cuif L. Comparative analysis of serotonin receptor 45 Kosten TA, Galloway MP, Duman RS, Russell DS, D'Sa C. Repeated unpredictable (HTR1A/HTR1B families) and transporter (slc6a4a/b) gene expression in the stress and antidepressants differentially regulate expression of the bcl-2 family of zebrafish brain. J Comp Neurol 2008; 511: 521–542. apoptotic genes in rat cortical, hippocampal, and limbic brain structures. 37 Mei J, Zhang Q-Y, Li Z, Lin S, Gui J-F. C1q-like inhibits p53-mediated apoptosis and Neuropsychopharmacology 2007; 33: 1545–1558. controls normal hematopoiesis during zebrafish embryogenesis. Dev Biol 2008; 46 Murray F, Hutson PH. Hippocampal Bcl-2 expression is selectively increased 319: 273–284. following chronic but not acute treatment with antidepressants, 5-HT or 38 Domar A, Moragianni V, Ryley D, Urato A. The risks of selective serotonin reuptake 1A 5-HT receptor antagonists. Eur J Pharmacol 2007; 569:41–47. inhibitor use in infertile women: a review of the impact on fertility, pregnancy, 2C/2B 47 Chen S, Xuan J, Wan L, Lin H, Couch L, Mei N et al. Sertraline, an antidepressant, neonatal health and beyond. Hum Reprod 2013; 28: 160–171. induces apoptosis in hepatic cells through the mitogen-activated protein kinase 39 Josefsson A, Berg G, Nordin C, Sydsjö G. Prevalence of depressive symptoms in 137 – late pregnancy and postpartum. Acta Obstetr Gynecol Scand 2001; 80:251–255. pathway. Toxicol Sci 2014; : 404 415. 40 Alwan S, Reefhuis J, Rasmussen SA, Olney RS, Friedman JM. Use of selective 48 Seritrakul P, Samarut E, Lama TT, Gibert Y, Laudet V, Jackman WR. Retinoic acid fi 26 serotonin-reuptake inhibitors in pregnancy and the risk of birth defects. N Engl J expands the evolutionarily reduced dentition of zebra sh. FASEB J 2012; : Med 2007; 356: 2684–2692. 5014–5024. 41 Cooper C, Fall C, Egger P, Hobbs R, Eastell R, Barker D. Growth in infancy and bone 49 Szegezdi E, O’Reilly A, Davy Y, Vawda R, Taylor DL, Murphy M et al. Stem cells mass in later life. Ann Rheum Dis 1997; 56:17–21. are resistant to TRAIL receptor‐mediated apoptosis. J Cell Mol Med 2009; 13: 42 FDA. FDA Drug Safety Communication: Revised Recommendations for Celexa 4409–4414. (Citalopram Hydrobromide) Related to a Potential Risk of Abnormal Heart Rhythms 50 Health UDo, Services H. Bone Health and Osteoporosis: a Report of the Surgeon with High Doses: U.S. Food and Drug Administration; 2012 (cited 3 August 2014). General. US Department of Health and Human Services, Office of the Surgeon Available fromhttp://www.fda.gov/drugs/drugsafety/ucm297391.htm. General: Rockville, MD, USA, 2004.

Supplementary Information accompanies the paper on the Molecular Psychiatry website (http://www.nature.com/mp)