Clomipramine causes osteoporosis by promoting osteoclastogenesis via E3 ligase Itch, which is prevented by Zoledronic acid

Patients taking antidepressants, including Clomipramine (CLP), have an increased risk of osteoporotic fracture. However, the effects of CLP on bone metabolism are unknown. Here, we demonstrate that WT mice treated with CLP for 2 weeks had significantly reduced trabecular bone volume and cortical bone thickness, associated with increased osteoclast (OC) numbers, but had no change in osteoblast numbers or bone formation rate. Bone marrow cells from CLP-treated mice had normal OC precursor frequency, but formed significantly more OCs when they were cultured with RANKL and M-CSF. CLP promoted OC formation and bone resorption and expression of OC-associated genes. CLP-induced bone loss was prevented by Zoledronic acid. At the molecular level, CLP inhibited the activity of the ubiquitin E3 ligase Itch. CLP did not promote OC formation from bone marrow cells of Itch−/− mice in vitro nor induce bone loss in Itch−/− mice. Our findings indicate that CLP causes bone loss by enhancing Itch-mediated osteoclastogenesis, which was prevented by Zoledronic acid. Thus, anti-resorptive therapy could be used to prevent bone loss in patients taking antidepressants, such as CLP.

OCs or OBs to cause bone loss. This is an important question because if CLP induces bone loss by promoting OC function, one can treat patients who take CLP with anti-resportive drugs.
Bisphosphonates are clinically used anti-resorptive drugs for patients with osteoporosis mainly by promoting apoptosis of mature OCs and inhibiting RANK expression of OC precursors 21 . Zoledronic acid (ZA), a bisphosphonate, has been considered as the most effective drug on treating bone tumor metastasis 22,23 and osteoporosis 24 . We used ZA to prevent bone loss in a mouse model of leukemia 25 .
A recent drug screening study reported that CLP inhibits the activity of ubiquitin E3 ligase Itch 26 . Itch negatively regulates multiple inflammatory signaling pathways, including JNK 27 and NF-κ B 28 in variety of cell types. We reported previously that OC precursors from Itch−/− mice have prolonged RANKL-induced NF-κ B activation and they formed more mature OCs in vitro and in vivo 29 , indicating that Itch is a negative regulator of osteoclastogenesis. Finding of CLP inhibiting Itch raises an interesting possibility that CLP may involve in osteoclastogenesis via the regulation of Itch. In this study, we examined the requirement of Itch for CLP-induced bone loss in Itch−/− mice, and if this is via OCs or OBs and can be prevented by ZA treatment.

Results
Mice-treated with Clomipramine have decreased bone volume. Clomipramine (CLP) is prescribed to patients with depression or other mental disorders in a wide range of doses (typically 25-250 mg/day. e.g. 12.5-125 mg/m 2 for a person with 2 m 2 of body surface), starting at the lowest dose and gradually increasing until an effective dose is reached 30 . To determine the effects of CLP on bone mass, we treated 2-month-old WT C57BL/6 J male mice with CLP at 10 mg/kg daily or with saline Control (Ctrl) for 2 weeks. This dose was selected because it is approximately equivalent to the 31 mg/m 2 dose administered to patients (for a patient with 2 m 2 of body surface, the dose is 62 mg/day, which is a lower effective dose used in the clinic) 31 . CLP reduced bone volume in both long bones and vertebrae. Tibial μ CTs showed that CLP decreased trabecular bone volume (BV/TV: 19.4% ± 3.3% vs. 24.5% ± 3.1% in Ctrl mice), trabecular thickness (0.048 mm ± 0.002 mm vs. 0.060 mm ± 0.004 mm in Ctrl mice) and cortical bone thickness (0.202 mm ± 0.006 mm vs. 0.224 mm ± 0.010 mm in Ctrl mice) (Fig. 1a). Decreased bone volume (BV/TV: 11.02% ± 1.12% vs. 13.46% ± 0.99% in Ctrl mice) was confirmed in Von Kossa-stained sections of vertebrae (Fig. 1b). CLP did not affect body weight or induce any pathologic changes in sections of internal organs, including liver, spleen and intestine (data not shown).
Bone marrow cells from mice-treated with Clomipramine have increased osteoclast forming potential ex vivo. To determine if CLP affects OC precursor frequency in bone marrow (BM), we stained BM cells from CLP-treated mice with cell surface markers for OC precursors and mesenchymal progenitors and subjected them to flow cytometric analysis. We defined OC precursors as CD11b + Gr1 −/low cells and mesenchymal progenitors as CD45 − CD105 + cells 32,33 . We did not detect any difference in the percentage of CD11b + Gr1 −/low and CD45 − CD105 + cell populations between CLP-treated and Ctrl mice (CD11b + Gr1 −/low cells: 7.1% ± 0.2% vs. 6.1% ± 0.3% in Ctrl mice; CD45 − CD105 + cells: 0.81% ± 0.03% vs. 0.83% ± 0.02% in Ctrl mice) (Fig. 2a). When we cultured these BM cells in our OC formation assays, we found that cells from CLP-treated mice formed many more OCs and OC area than Ctrl mice (0.81 mm 2 ± 0.12 mm 2 vs. 0.31 mm 2 ± 0.03 mm 2 ) (Fig. 2b). In contrast, BM cells from CLP-treated mice had no significant changes in OB differentiation as indicated by CFU-ALP colony formation assays (5.1 ± 0.7 vs. 5.3 ± 1.0 in Ctrl mice) (Fig. 2c).

Clomipramine increases osteoclast formation and bone resorption in vitro.
To determine if CLP directly affects OC formation and function, we cultured BM cells from WT mice with M-CSF and RANKL for 7 days in the presence of different concentrations of CLP (0-10 μ M). We found that CLP significantly increased OC numbers (Fig. 3a) and area, at a dose of 0.1 μ M and 0.3 μ M (13.3 mm 2 ± 0.8 mm 2 vs. 6.06 mm 2 ± 1.3 mm 2 in Ctrl). The expression levels of OC-related genes, including nfatc1, acp5 and ctsk, were significantly increased in RNA extracted from CLP-treated cells (Fig. 3b). In addition, CLP increased total OC numbers and resorption area per bone slice when they were cultured on the bone slices, but not pit area/OC (Fig. 3c).
Zoledronic acid treatment prevents Clomipramine-induced bone loss. Bisphosphonates are used to inhibit bone resorption in patients with osteoporosis 34 . Because our data suggest that bone loss caused by CLP treatment is due mainly to increased OC-mediated bone resorption, we next tested if zoledronic acid (ZA), a commonly used bisphosphonate, could prevent the reduction in bone mass caused by CLP. We treated WT mice with ZA (0.25 mg/kg, i.p. twice/week) or sterile water for 2 weeks 25 and followed with CLP treatment for an additional 2 weeks. Compared to mice that were treated with sterile water (vehicle for ZA) followed by saline (vehicle for CLP), CLP treated mice developed bone loss and increased OC numbers as showed in Fig. 1. ZA restored CLP-reduced bone volume (Fig. 4a). Analysis of H&E-and TRAP-stained sections revealed that ZA restored CLP-induced bone loss and OC formation (Fig. 4b). Mice deficient in the ubiquitin E3 ligase, Itch, do not develop Clomipramine-induced bone loss. The molecular mechanisms whereby CLP induces its antidepressant effects have not been well investigated. Some early reports indicate that CLP may affect serotonin uptake 35 , a Rho-kinase pathway and Ca(2+ ) -channels 36 . A more recent study using high throughput screening for inhibitors of the ubiquitin E3 ligase, Itch, indicates that CLP can specifically inhibit Itch ligase activity by binding irreversibly to HECT domains or catalytic cysteines 26 . Interestingly, we reported that Itch negatively regulates OC formation 29 . Thus, we hypothesized that Itch−/− mice would not develop CLP-induced osteoclastogenesis and bone loss. We treated BM cells from WT and Itch−/− mice with CLP in our OC formation assays and found that Itch−/− cells formed more OCs than WT cells (7.9 ± 1.35 mm 2 vs. 3.0 ± 0.65 mm 2 ) 29 . However, CLP treatment increased OC formation from WT cells (6.4 ± 0.75mm 2 vs. 3.0 ± 0.65mm 2 ), but not in Itch−/− cells ( Fig. 5a and b). At basal levels, Itch−/− mice have low bone mass and high OC number compared to WT mice because itch is a negative regulator of OC formation as we previously reported 29 . We found that CLP administration did not cause bone loss in Itch−/− mice, but it caused bone loss in WT littermates (Fig. 5c). BM cells from CLP-treated WT mice formed more OCs when they were cultured with RANKL and M-CSF than cells from vehicle-treated WT mice, but this effect was not observed in cells from CLP-treated Itch−/− mice (Fig. 5d). Consistently, CLP increased OC numbers and surface area in WT mice, but not in Itch−/− mice (Fig. 5e).

Discussion
Antidepressant medications are primarily prescribed for depression and anxiety 37,38 and for some other non-psychiatric conditions 39 . The effects of antidepressants on bone have not been studied in depth, and the findings of some reports are controversial. For example, the antidepressant, lithium, accelerates proliferation of human bone marrow-derived mesenchymal progenitor cells in vitro, associated with inhibition of GSK-3β and accumulation of β -catenin 40 . Local application of lithium promotes bone regeneration by affecting both osteoblasts and osteoclasts (OCs) 41 . In contrast, 5-hydroxytryptophan, the precursor of serotonin, causes bone loss by stimulating osteoclastogenesis via RANKL 42 . In the current study, we investigated the bone's effects of Clomipramine (CLP), a drug used to treat patients with obsessive-compulsive disorder, major depression, panic attacks, anxiety disorders and premature ejaculation. We found that CLP directly stimulates OC formation and causes bone loss in mice via the ubiquitin E3 ligase, Itch, and that Zoledronic acid (ZA) prevents CLP-induced bone loss. These findings suggest that ZA or other anti-resorption drugs may prevent bone loss and reduce fracture risk in patients who take CLP and other antidepressants.
OC formation requires the RANKL/RANK signaling. RANKL and RANK interactions recruits TNF receptor-associated factor 6 (TRAF6), leading to activation of essential OC transcription factors including NF-κ B 43,44 . Itch is a negative regulator of NF-κ B signaling in lymphocytes and macrophages 28,45,46 and it inhibits OC formation by promoting deubiquitination of TRAF6 and suppressing NF-κ B signaling 29 . Rossi et al. used an ELISA-based high throughput screening assay and screened more than 20,000 compounds (1040 compounds from the NINDS library and ~20,000 compounds from various commercial suppliers) for Itch inhibitors. CLP was the only compound with high affinity that inhibited Itch ligase activity. We found that Itch deficiency abolished CLP-promoted osteoclastogenesis and bone loss, suggesting that Itch inhibition is one of the mechanisms by which CLP induces bone loss.
CLP inhibits the reuptake of serotonin in the central nervous system through interaction with serotonin 5-HT2 and 5-HT3 receptor subtypes 47,48 . Thus, it is possible that CLP treatment leads to elevated circulating serotonin levels to inhibit osteoblast function because serotonin produced in the gut negatively regulates osteoblasts 20 . Our study cannot rule out the potential contribution of the serotonin pathway to CLP-induced bone loss. However, we did not detect significant influence of CLP in osteoblast parameters both in vivo and ex vivo, suggesting that CLP unlikely inhibits osteoblast function. In contrast, CLP increased OC-mediated bone resorption, strongly indicating that CLP affects mainly OCs. CLP promoted OC formation on plastic dishes peaked at 0.3 μ M and OC stimulatory effect declined thereafter. CLP stimulated OC formation and bone resorption on bone slices at 1 μ M. Our explanation for this difference is that bone slices provide survival factors, making OCs resistant to potential toxic effect of high dose CLP on cells.
CLP is available to patients as capsules for oral administration. We used the daily intraperitoneal injection in our study. This treatment design raises two questions. One is if the serum concentration of the drug in our mice is comparable exposures to humans. The plasma half-life of CLP in mice after intraperitoneal injection is about 2 hours, which is much shorter than in man (about 22 hours) after oral administration 49 . However, even with much shorter half-live, we still observed significantly increased OCs and reduced bone volume in CLP-treated mice, suggesting that it is a potential stimulator of osteoclastogenesis in vivo. Another question is the rationale for 2 weeks of CLP dosing. CLP has been used in many mouse behavior studies 13,35 . In C57BL/6 J mice, 2 weeks treatment of CLP exerts anti-depression effect 50 , suggesting 2-week dosing may represent a reasonable regimen to study the in vivo effect of CLP. Because anti-depressants are often given chronically, we suspect that longer duration of CLP treatment will cause more severe bone loss.
Few studies have addressed the management of fracture risk in patients who take anti-depressants. We found that ZA could reduce bone loss in CLP-treated mice. ZA is used to treat a number of diseases associated with elevated bone resorption, including osteoporosis, osteolytic bone metastases, and multiple myeloma. ZA or other anti-resorption drugs could therefore be useful for prevention of bone loss in antidepressant users. The clinical significance of our study is twofold. It first suggests that monitoring bone mineral density and serum bone resorption markers should be included as components of standard care for patients who take antidepressant medication. It also suggests that anti-resorptive medication may have beneficial effects in these patients.
In summary, we found that CLP, one of the tricyclic antidepressants used for treating patients with depression, anxiety, and obsessive-compulsive disorders, caused decreased bone mass in mice. We found that CLP, an inhibitor of the E3 ligase, Itch, increased OC formation in vivo and in vitro, and slightly increased osteoblast formation in vitro, but not in vivo. ZA prevented CLP-induced bone loss in mice. Hence, anti-resorptive drugs may be beneficial for patients on antidepressant drugs who are at risk for osteoporotic fractures.

Materials and Methods
Ethics statement. All experiments with animals were performed in strict accordance with the Animal Experimentation Guidelines of the National Institute of Infectious Diseases, and the protocol was approved by the University of Committee on Animal Resources in the University of Rochester.
Mice and treatment. Two-month-old wild-type C57BL/6 J male mice were purchased from Jax laboratory.
Itch−/− mice on a C57BL/6 J background were generated and genotyped by PCR 51 . Although both male and female patients take anti-depressants, we used male mice in the study because we wanted to avoid the potential influence of estrogen on bone in female mice. Three sets of experiments were performed. 1) To study the role of Clomipramine (CLP, SIGMA, C-7291) on bone mass, mice (n = 5-7/group) were treated with water as control (Ctrl) or CLP (10 mg/kg body weight) by intraperitoneal (i.p.) injection daily for two weeks. 2) To study the effect of Zoledronic acid on CLP-induced bone loss, mice were pre-treated with saline or Zoledronic acid 25 (ZA, SIGMA, SML0223; 0.25 mg/kg body weight) by i.p. injection twice/week for two weeks, then followed by CLP for another 2 weeks (n = 4-6/group). 3) To study if itch is required for CLP-induced bone loss, age-(3-month-old) and gender-(male) matched Itch−/− mice (n = 4-5/group) were treated with CLP (10 mg/kg body weight) by i.p. injection daily for two weeks. Mice were sacrificed 4 hours after the last treatment and samples were harvested for analyses.
μCT, histology, and histomorphometric analysis. Soft tissues were removed from left tibiae and lumbar 1 (L1) vertebrae, fixed in 10% formalin at 4 °C for 24-48 hours on a rocker and transferred to 70% ethanol until μ CT scanning. Specimens were scanned at high resolution (10.5 μ m) on a VivaCT 40 μ CT scanner (Scanco Medical) using 300 ms integration time, 55 kVp energy and 145 μ A intensity. Then these tibiae were decalcified in 10% EDTA at 4 °C for 21 days after formalin fixation and embedded in paraffin. Sections (4 μ m thick) were stained with H&E for bone volume and osteoblast (OB) analysis and with TRAP-Fast Green for osteoclast (OC) analysis using OsteoMeasure system. OB and OC numbers and surfaces per millimeter bone surface were analyzed in tibial sections with point counting using an eyepiece grid and an Olympus TH4-100 microscope. Tibial and vertebral bone volumes were analyzed in digital slides generated using an Olympus VS120 image analysis system and automated analysis and algorithms that we developed using a Visiopharm Image Analysis System 52 .

Double calcein labeling.
Calcein was administered to mice (10 mg/kg body weight) by i.p. injection at 3 days and 1 day before they were sacrificed. Vertebrae (L4) were harvested, fixed in 10% formalin at 4 °C for 24-48 hours, and embedded in LR white acrylic resin. Bone sections of L4 (4 μ m thickness) were cut and observed under fluorescence microscopy. Bone formation rate (BFR) were analyzed using an OsteoMetrics image analysis software system (OsteoMetrics) 32,52 . Flow cytometry. Bone marrow (BM) cells were flushed from left femora with PBS. Red blood cells were lysed using lysis buffer (Thermo Fisher Scientific, A1049201). After being washed with PBS, cells were suspended in PBS with 2% FBS. Cells were stained with various fluorescein-labeled Abs and subjected to flow cytometric analysis using a Becton-Dickinson FACSCanto II Cytometer. FITC-anti-CD45, PE-anti-CD105, APC-anti-CD11b, and PECY5-anti-Gr1 antibodies were purchased from eBioscience. Results were analyzed by Flowjo7 data analysis software (Ashland, OR).
Cell culture and analysis. For OC formation assays, BM cells were harvested from right femora and cultured in 96-well plates (5 × 10 4 cells/well) in α -MEM with 10% FBS, 1:50 dilution of conditioned medium containing M-CSF (from an M-CSF producing cell line) and a sub-optimal concentration of RANKL (1 ng/ml) in the presence of different doses of CLP for 7 days. After multinucleated cells were observed under inverted microscopy, cells were fixed by 10% formalin and stained for TRAP activity. TRAP-positive cells with more than 3 nuclei were counted for TRAP + OC number using standard stereology methods 33 . For OC bone resorption assays, BM cells were cultured on sterile cow bone slices (size: 2 mm × 3 mm; thickness: 0.75 mm) under the same conditions for 10 days as described above. Cells were removed and bone slices were stained with 0.1% toluidine blue to visualize resorption pits. The area of pits was quantified and data were expressed as area of pits (mm 2 )/slice. For CFU-Alkaline phosphatase (ALP)-colony formation assays, BM cells from CLP treated mice or from Ctrl mice were cultured in 6-well plates (2 × 10 6 cells/well) in α -MEM containing 15% FBS with 50 μ g/ml ascorbic acid and 10 mM β -glycerophosphate for 21 days. Cells were fixed with 70% ethanol and stained for ALP activity. The number of ALP + colonies with more than 20 cells was counted, as we described previously 32 .

Statistical analysis.
All results are listed as mean ± SD. Statistical analysis was performed using GraphPad Prism 5 software (GraphPad Software Inc., San Diego, CA). Comparisons between 2 groups were analyzed using a 2-tailed unpaired Student's t test. Comparisons among 3 or more groups were carried out using one-way ANOVA followed by Dunnett's post-hoc multiple comparisons. p values < 0.05 were considered statistically significant.