Post‐menopausal oestrogen deficiency induces osteoblast apoptosis via regulating HOTAIR/miRNA‐138 signalling and suppressing TIMP1 expression

Abstract In this study, we aimed to explore the molecular mechanisms underlying the development of osteoporosis in post‐menopausal females. Real‐time PCR was conducted to measure the expression of potential lncRNAs involved in the osteoporosis of post‐menopausal females. In addition, Western blot and IHC assays were used to study the possible correlation among HOTAIR, miR‐138 and TIMP1, while a computational analysis was carried out to predict the ‘seed sequence’ responsible for the binding between miR‐138 and HOTAIR/TIMP1. Furthermore, luciferase reporter assays were conducted to validate the negative regulatory relationship between miR‐138 and TIMP1/HOTAIR. To evaluate the effect of oestrogen on the function of HOATIR and its downstream effectors, luciferase activity was measured in cells cotransfected with different vectors or treated with different doses of oestrogen. The results of the luciferase assay were further validated by real‐time PCR, Western blot, MTT assay and flow cytometry. Among the candidate lncRNAs, HOTAIR was the only lncRNA down‐regulated in post‐menopausal females. HOTAIR bound to miR‐138 and negatively regulated its expression. Meanwhile, miR‐138 could also bind to TIMP1 mRNA and reduce its expression. Furthermore, a dose‐dependent up‐regulation of HOTAIR was observed in cells treated with oestrogen, and the elevated HOTAIR increased the level of TIMP1 by targeting miR‐138. In addition, oestrogen promoted cell viability and suppressed cell apoptosis, and effects of oestrogen were blocked by the silencing of HOTAIR. Therefore, it can be concluded that oestrogen deficiency could induce the apoptosis of osteoblasts and lead to osteoporosis in post‐menopausal females via modulation of the HOTAIR/miR‐138/TIMP1 signalling axis.


| INTRODUC TI ON
As a disease caused by the loss of bone mineral content and featured by bone strength reduction, osteoporosis affects the life quality of millions of people worldwide, especially those suffering from pathological fractures. 1,2 Bone re-modelling tends to maintain a balance between the formation of new bones via osteoblast activation and the resorption of existing bones. Increasing evidence has demonstrated that the apoptosis of osteoblasts is a major mechanism underlying the pathogenesis of osteoporosis. 3 Therefore, the induction of osteoclast apoptosis and the prevention of osteoblast apoptosis may be used as the potential strategy in the treatment of osteoporosis. 3 Long non-coding RNAs (lncRNAs) are RNA transcripts of longer than 200 nucleotides that do not encode proteins. 4 Although ln-cRNAs were once deemed as transcriptional 'noises', they have been recently shown to play critical roles in many biological processes, such as cell proliferation, transcriptional regulation and tumorigenesis. [5][6][7] It has been reported that many lncRNAs can 'sponge' the activity of microRNAs (miRNAs) by sharing a common miRNA response element (MRE) and inhibiting the expression of their 'target' miR-NAs. 8,9 Furthermore, lncRNAs such as HOTAIR have been implicated in osteoblast differentiation. 10 In addition, the effect of SMAD7 on osteoblast differentiation was also investigated by down-regulating its expression using miR-17-5p. The results of many previous studies collectively suggested that HOTAIR participates in the control of osteoblast differentiation by regulating the expression of miR-17-5p and its target gene, SMAD7. So far, several microRNAs (miRNAs), including miR-214, miR-148a, miR-10 and miR-9, have been implicated in the development of osteoporosis. [11][12][13] For example, miR-138 was demonstrated to suppress osteoblast differentiation in mesenchymal stem cells (MSCs), indicating that this miRNA plays a critical role in the progression of osteoporosis. It has also been shown that miR-138 acts as a negative regulator during the differentiation of human MSCs into osteoblasts, while the inhibition of miR-138 expression can promote such differentiation.
Previous studies have demonstrated that the overexpression of HOTAIR suppressed the expression of miR-138 during lipopolysaccharide (LPS)-stimulation in chondrocytes and rats with rheumatoid arthritis (RA). In addition, the regulatory relationship between miR-138 and HOTAIR has been further clarified by luciferase assays. In summary, the above results suggested that miR-138 acts as a target gene of HOTAIR. 14 Past studies have shown that the decline in the level of ovarian oestrogen during menopause could lead to a rapid loss of trabecular micro-architecture, thus increasing the rate of bone resorption and the severity of cortical porosity, which in turn increase the risk of osteoporosis and fracture. 15 Oestrogen deficiency can affect skeletal homeostasis by inducing bone resorption, and the severity of oestrogen deficiency becomes greater over time. 16 Interestingly, oestrogen deficiency was shown to alter the local distribution of mineral density, thus modifying the mechanical properties of the bones. [17][18][19][20][21] Furthermore, it has been demonstrated that the onset of oestrogen deficiency can lead to bone loss in the mandible. [22][23][24] LncRNAs MALAT1 and HOTAIR not only serve as transcriptional targets of miRNAs but are also involved in the regulation of hormone-sensitive genes, such as PSA, hTERT and pS2, that are targeted by oestrogen in prostate cancer cells. 25 Previous studies have shown the correlation between the significant reduction in the expression of tissue inhibitor of metalloproteinase-1 (TIMP1) and increased apoptosis of osteoblasts under endoplasmic reticulum (ER) stress. In summary, TIMP1 exerts a critical effect on osteoblast apoptosis and the development of osteoporosis. 26

| Cell culture and transfection
HFOB and MG63 cells were cultured in a DMEM medium containing 10% foetal bovine serum (charcoal-stripped medium was used for oestrogen treatment assay and regular medium for others). When the cells reached approximately 80% confluence, they were adjusted to a concentration of 2 × 10 6 /mL and seeded into 96-well plates for transfection. The transfection was performed using Lipofectamine

| Cell proliferation assay
Transfected HFOB and MG63 cells were collected and adjusted to a density of 20,000 cells/ml. Subsequently, the cells were seeded into a 96-well plate and cultured at 37℃ and 5% CO 2 . Each treatment group was assayed in triplicate. In the next step, 20 μL of MTT reagent (5 mg/mL, Sigma-Aldrich) were added into each well and incubated in the dark for 4 hours at 37°C. Finally, after the addition of 150 μL of DMSO followed by 10 minutes of incubation on a shaker to dissolve the precipitates, the absorbance value in each well was measured on a TFC instrument (Thermo Fisher Scientific) at a wavelength of 490 nm. A growth curve of the cells was then plotted using absorbance values as the ordinate and the culture time as the abscissa.

| Luciferase assay
In order to investigate the regulatory relationship of HOTAIR/miR- were treated with 5 nmol/L and 10 nmol/L of oestrogen, respectively, for 24 hours before the luciferase activity in the cells was measured to evaluate the effect of oestrogen on the transcription of HOTAIR promoter. In particular, the luciferase activity of transfected cells was measured on a Glomax20/20 luminometer (Promega) using a Dual-luciferase assay kit (Promega). Each experiment was repeated three times.

| Western blot analysis
After the tissue and cell samples were collected, they were lysed using a cell lysis buffer before the protein concentration in the lysate was measured by a BCA kit (Thermo Fisher Scientific). Subsequently, Biotechnology). Subsequently, the membrane was developed using an ECL solution (Amersham Bioscience) and visualized using a Scion Image analysis system (Scion Corporation). The relative expression of TIMP1 protein was calculated based on the ratio between the optical density of TIMP1 protein and that of the internal control β-actin.

| Immunohistochemistry (IHC) assay
The samples of formalin-fixed demineralized bone tissues and treated cells were fixed in a 10% formaldehyde solution, soaked in paraffin, dewaxed with xylene, hydrated with gradient alcohol,  SPSS 20.0 software was used for data analysis. Measurement data were presented as mean ± standard deviations. Student's t tests were used for the comparison between two groups, while one-way analysis of variance (ANOVA) was used for the comparison among multiple groups, followed by Bonferroni's test.

| Statistical analysis
All tests were two-sided, with P < .05 considered as statistically significant.

| HOTAIR was down-regulated in postmenopausal females
In this study, bone tissue samples were collected from 34 osteoporosis patients, including 18 post-menopausal females and 16 pre-menopausal females. The demographic parameters, risk factors and BMD of the subjects were presented in Table 1

| The negative correlation between the expression of miR-138 and TIMP1/HOTAIR
Subsequently, the expression of miR-138 and TIMP1/HOTAIR was compared between the post-menopausal and pre-menopausal groups. As shown in Figure 1F, the relative expression of miR-138 was significantly elevated in the post-menopausal group compared with that in the pre-menopausal group. However, the relative expression of both TIMP1/HOTAIR mRNA ( Figure 1G) and protein ( Figure 1H) was significantly reduced in the post-menopausal group.
The above results were also confirmed by the results of IHC assays ( Figure 1I), indicating the presence of a negative correlation between the expression of miR-138 and TIMP1/HOTAIR.

| MiR-138 could bind to HOTAIR
To explore the regulatory relationship between miR-138 and HOTAIR, a computational analysis was conducted using an online miRNA database. As shown in Figure 2A, there was a matching 'seed sequence' on miR-138-5p allowing it to bind to HOTAIR. To verify the role of this 'seed sequence', we cotransfected cells with miR-138-5p mimics and wild-type or mutant HOTAIR, and then measured the relative luciferase activity in transfected cells. As shown in Figure 2B, the relative luciferase activity of HFOB cells cotransfected with wild-type HOTAIR and miR-138 mimics was significantly reduced compared with that in the negative control cells, while the relative luciferase activity of HFOB cells cotransfected with mutant HOTAIR and miR-138 mimics showed no major change. Similar results were also obtained in MG63 cells ( Figure 2C), confirming the fact that HOTAIR is a direct target of miR-138.

| TIMP1 mRNA was a target gene of miR-138
Furthermore, to explore the regulatory relationship between miR-138 and TIMP1, we carried out a computational analysis using the online target gene predicting tool. As shown in Figure 2D, TIMP1 was identified as a potential target gene of miR-138 with a miR-138 binding site located in the 3'UTR of TIMP1. Subsequently, we cotransfected HFOB and MG63 cells with miR-138 mimics and wildtype or mutant 3'UTR of TIMP1, and then measured the relative F I G U R E 2 MiR-138 could bind to HOTAIR while TIMP1 mRNA was a target gene of miR-138 (Replicate number N = 3). A, A potential 'seed sequence' responsible for the binding of miR-138-5p to HOTAIR. B, Relative luciferase activity of HFOB cells cotransfected with wild-type HOTAIR and miR-138 mimics was evidently higher than other HFOB cell groups (*P value of < .05 compared with the negative controls). C, Relative luciferase activity of MG63 cells cotransfected with wild-type HOTAIR and miR-138 mimics was evidently higher than other MG63 cell groups (*P value of < .05 compared with the negative controls). D, A potential 'seed sequence' of miR-138 was located in the 3'UTR of TIMP1. E, Relative luciferase activity of HFOB cells cotransfected with wild-type TIMP1 mRNA and miR-138 mimics was evidently higher than other HFOB cell groups (*P value of < .05 compared with the negative controls). F, Relative luciferase activity of MG63 cells cotransfected with wild-type TIMP1 mRNA and miR-138 mimics was evidently higher than other MG63 cell groups (*P value of < .05 compared with the negative controls) luciferase activity in transfected cells. As shown in Figure 2E, the HFOB cells cotransfected with miR-138 mimics and the 3'UTR of wild-type TIMP1 showed apparently decreased luciferase activity compared with the cells transfected with negative controls, while the HFOB cells cotransfected with miR-138 mimics and the 3'UTR of mutant TIMP1 showed no change in luciferase activity. Similar results were also obtained in MG63 cells ( Figure 2F), confirming the fact that TIMP1 acts as a virtual target gene of miR-138.

| Oestrogen exhibited its effect via regulating the level of HOTAIR
According to the above results, HOTAIR acted as an upstream regulator in the molecular pathway of HOTAIR, miR-138 and TIMP1. Since HOTAIR is differentially expressed between post-menopausal and pre-menopausal females, the effect of oestrogen on the expression of HOTAIR was also evaluated in this study by treating HFOB and MG63 cells with 5 nmol/L or 10 nmol/L of oestrogen, respectively.
As shown in Figure 3A, the relative luciferase activity of HFOB cells was the lowest in the negative control group and the highest in the 10 nmol/L oestrogen treatment group, indicating that oestrogen acted as an enhancer of HOTAIR expression.
The relative expression of HOTAIR and its downstream targets, miR-138 and TIMP1, was also measured in cells treated with 5 nmol/L or 10 nmol/L of oestrogen, respectively. Accordingly, the relative expression of HOTAIR ( Figure 3B), as well as the relative expression of TIMP1 mRNA and protein ( Figure 3D), in the oestrogen-treated HFOB cells was gradually increased in a dose-dependent manner.
On the contrary, the relative expression of miR-138 ( Figure 3C

| D ISCUSS I ON
As a disease of bone degeneration frequently observed at an older age, osteoporosis is characterized by changes in bone density and bone micro-structures. 28 During osteoporosis, the dynamic equilibrium between bone formation and bone absorption is impaired. 29 During bone formation, osteoblasts play an important role in bone growth, bone metabolism and bone repair. 30 Therefore, the dys- TIMP1 has been implicated in the metabolism of bones. In addition, the increased expression of TIMP1 in osteoblasts has been shown to increase mineral density and bone mass. 37 Xie et al also demonstrated that the inhibition of TIMP1 expression could lead to the apoptosis of osteoblasts. 38 Furthermore, TIMP1 was found to reduce the number of neuronal cells in patients with neurodegenerative diseases who were exposed to ER stress. 39

| CON CLUS ION
In conclusion, our findings demonstrated a novel role of oestrogen in the development of osteoporosis by regulating the expression of HOTAIR. In addition, HOTAIR acted as a competing endogenous RNA for miR-138 and its target gene TIMP1, which in turn plays an important role in the apoptosis of osteoblasts. Finally, we confirmed that oestrogen could at least partially affect the progress of osteoporosis by controlling the apoptosis of osteoblasts and by upregulating the expression of HOTAIR, an lncRNA playing a protective role against osteoporosis.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest.

F I G U R E 3
Oestrogen could increase the luciferase activity of HOTAIR while influencing the relative expression of HOTAIR and its downstream targets, miR-138 and TIMP1. The silencing of HOTAIR could also block the effect of oestrogen on HOTAIR and its downstream targets. Oestrogen could also increase cell viability and suppress cell apoptosis via mediating the HOTAIR/ mir-138/ TIMP-1 signalling axis (Replicate number N = 3). A, Relative luciferase activity of HOTAIR was elevated in HFOB cells treated with 5 nmol/L and 10 nmol/L of oestrogen (*P value of < .05 compared with the negative controls). B, Relative expression of HOTAIR was increased in HFOB cells treated with 5 nmol/L and 10 nmol/L of oestrogen, and the transfection of HOTAIR siRNA silenced the expression of HOTAIR (*P value of < .05 compared with the negative controls; ** P value of < .05 compared with the E2 10 nmol/L + NC siRNA group). C, Relative expression of miR-138 was inhibited in HFOB cells treated with 5 nmol/L and 10 nmol/L of oestrogen, and HOTAIR siRNA increased the miR-138 expression (*P value of < .05 compared with the negative controls; **P value of < .05 compared with the E2 10 nmol/L + NC siRNA group). D, TIMP1 mRNA expression was up-regulated by 5 nmol/L and 10 nmol/L of oestrogen, and HOTAIR siRNA partly blocked the effect of oestrogen in HFOB cells (*P value of < .05 compared with the negative controls; **P value of < .05 compared with the E2 10 nmol/L + NC siRNA group). E, The viability (% of control) of oestrogen-treated HFOB cells was dose-dependently increased, while HOTAIR siRNA reduced cell viability (*P value of < .05 compared with the negative controls; ** P value of < .05 compared with the E2 10 nmol/L + NC siRNA group). F, Apoptosis index of HFOB cells was dose-dependently reduced by oestrogen (5 nmol/L and 10 nmol/L), while silencing of HOTAIR blocked the effect of oestrogen (*P value of < .05 compared with the negative controls; ** P value of < .05 compared with the E2 10 nmol/L + NC siRNA group). G, Relative luciferase activity of HOTAIR was elevated in MG63 cells treated with 5 nmol/L and 10 nmol/L of oestrogen (*P value of < .05 compared with the negative controls). H, Relative expression of HOTAIR was increased in HFOB cells treated with 5 nmol/L and 10 nmol/L of oestrogen, and the transfection of HOTAIR siRNA silenced the expression of HOTAIR (*P value of < .05 compared with the negative controls; ** P value of < .05 compared with the E2 10 nmol/L + NC siRNA group). I, Relative expression of miR-138 was inhibited in HFOB cells treated with 5 nmol/L and 10 nmol/L of oestrogen, and HOTAIR siRNA increased miR-138 expression (*P value of < .05 compared with the negative controls; ** P value of < .05 compared with the E2 10 nmol/L + NC siRNA group). J, TIMP1 mRNA and protein expression was up-regulated by 5 nmol/L and 10 nmol/L of oestrogen, and HOTAIR siRNA partly blocked the effect of oestrogen in HFOB cells (*P value of < .05 compared with the negative controls; ** P value of < .05 compared with the E2 10 nmol/L + NC siRNA group). K, The viability (% of control) of oestrogen-treated HFOB cells was dose-dependently increased, while HOTAIR siRNA reduced cell viability (*P value of < .05 compared with the negative controls; ** P value of < .05 compared with the E2 10 nmol/L + NC siRNA group). L, Apoptosis index of HFOB cells was dose-dependently inhibited by 5 nmol/L and 10 nmol/L of oestrogen, while HOTAIR silencing promoted cell apoptosis (*P value of < .05 compared with the negative controls; ** P value of < .05 compared with the E2 10 nmol/L + NC siRNA group) Shi: Investigation (equal); Methodology (equal); Software (equal); Visualization (equal). Rui-ming Zhou: Investigation (supporting); Project administration (lead); Supervision (lead); Writing-review & editing (lead).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.