Downregulated microRNA-199a-3p differentiation of bone marrow mesenchymal stem cells by targeting Kdm3a in ovariectomized rats

Osteoporosis is a prevalent systemic skeletal disorder entailing bone fragility and increased fracture risk, often emerging in post-menopausal life. Emerging evidence implicates the dysregulation of microRNAs (miRNAs or miRs) in the progression of osteoporosis. This study investigated the effect of miR-199a-3p on osteoporosis and its underlying mechanism. We first examplished an ovariectomized (OVX)-induced rat osteoporosis model, and then isolated mesenchymal stem cells (MSCs) from bone marrow of the model rats. The overexpression and knock down of miR-199a-3p were conducted in OVX rats and MSCs to verify the role of miR-199a-3p on MSC differentiation. Calcium nodules were measured using alizarin red S (ARS) staining. RT-qPCR and Western blot assay were performed to measure the expression of miR-199a-3p, Kdm3a and osteogenic differentiation-related markers in rat tissues and cells. The correlation between miR-199a-3p and Kdm3a was confirmed using dual-luciferase reporter assay. The enrichment of Kdm3a at the Erk2 and Klf2 promoter was assessed using chromatin immunoprecipitation (ChIP) assay. Isolated MSCs were positive for CD29, CD44, CD90, and CD45, suggesting successful isolation of MSCs. There was increased expression of miR-199a-3p and inhibited osteogenic differentiation in OVX rats. Kdm3a was negatively targeted by miR-199a-3p. Our results also demonstrated that Kdm3a elevated the expression of Erk2 and Erk2 by promoting Erk2 and Klf2 demethylation, which further contributed to osteogenic differentiation. Overall, our results revealed a regulatory network of miR-199a-3p in osteogenic differentiation, highlighting miR-199a-3p as a potential target for therapeutic interventions in osteoporosis.


Introduction
Characterized by low bone mass, oteoporosis most frequently results from an elevation in bone resorption that is not sufficiently compensated by a corresponding enhancement in bone formation, resulting in a progressive net decline of bone mineral density (BMD) and disruption of bone microarchitecture [1,2]. Generally, osteoporosis and osteopenia have been diagnosed via measurement of BMD in the hip and spine using X-ray absorptiometry. However, this diagnostic procedure is usually performed after the onset of diseases symptoms, which delays the initiation of measures for prevention or treatment [3]. Yet, unlike the majority of chronic diseases, the approved osteoporosis treatment consists of a very limited repertoire of medicatinoa and dosage regimens [4]. Recently, targeting bone marrow mesenchymal stems (MSCs) as multipotent progenitor cells has emerged as a potential treatment to stimulate osteogenesis [5].
MicroRNAs (miRNAs or miRs), noncoding RNA molecules of only 20-22 bases in length, are recognized as important mediators of epigenetic modification, and there have been investigation of the their roles in bone remodeling, therapeutic implications for the treatment of osteoporosis [6,7]. A prior study revealed that mice upregulating miR-34a exhibit reduced bone resorption and enhanced bone mass, and that miR-34a could alleviate osteoporosis and bone metastasis by depressing osteoclastogenesis and Tgif2 expression [8]. Furthermore, miR-16-5p might function as a poteintial target for postmenopausal osteoporosis treatment, based on its suppression of osteogenesis by inhibiting the expression of vascular endothelial growth factor A (VEGFA) [9]. Besides, miR-199a-3p was identified as a biomarker for osteoporosis in the mandible of ovariectomized (OVX) mice [10]. MSCs isolated from mice treated with deficient lysine (K)-specific demethylase 3A (Kdm3a) have defective chromosome organization and an exacerbated DNA damage response, which manifests accelerated bone aging [11]. Kdm3a, a member of the KDM3 subfamily, has specificity to remove mono and dimethyl markers from H3K9 and is implicated in tumor/metastasis promotion, chemoresistance and other phenotypes, thus constituting a promosing therapeutic target in numerous cancers [12]. Interestingly, a bioinformatics analysis in the present study revealed that Kdm3a was a target for negative regulation by miR-199a-3p. Furthermore, in the cotext of myeloma, Kdm3a has been found to activate Krüppel-like factor 2 (Klf2), which is an important transcription factor in the regulation of endothelial cell function [13,14]. Klf2 has also been reported to expedite the self-renewal of human bone marrow stromal cells, which is conducive to tissue and organ regeneration [15].
Downregulation of Klf2 can increase osteoclast differentiation and inhibit osteoblast differentiation [16]. Klf2 is phosphorylated by extracellular-signal-regulated protein kinase 2 (Erk2), and the interaction or balance of effects of these competing processes may mediate ground state pluripotency of osteoblasts [17]. Erk2 is a form of ERK and thus a member of the mitogen-activated protein kinase (MAPK) family of signaling molecules, which has been Downloaded from http://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200314/901300/bcj-2020-0314.pdf by guest on 14 January 2021 Biochemical Journal. This is an Accepted Manuscript. You are encouraged to use the Version of Record that, when published, will replace this version. The most up-to-date-version is available at https://doi.org/10.1042/BCJ20200314 4 implicated in numerous cellular events, like cell proliferation, growth, differentiation, migration, survival, metabolism and transcription [18]. However, little is known about the effects of Erk2 on osteoporosis. Our study examined the possible regulatory network involving miR-199a-3p, Kdm3a, Klf2 and Erk2 in osteoporosis progression in a rodent model.

Ethics statement
The study protocol was ratified by the Institutional Animal Care and Use Committee of The Affiliated Hospital of Nantong University. Animal experiments took place in the Laboratory of The Affiliated Hospital of Nantong University and were performed in strict accordance to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health. Extensive efforts were made to minimize discomfort and numbers of the animals used in the study.

OVX-induced rat model of osteoporosis
Forty female Sprague Dawley rats (aged 12 weeks, weighing 230 ± 15g) were purchased from the Laboratory Animal Research Center, Fourth Military Medical University (Xi'an, China). The rats were fed with 15 g of a rodent diet per day containing 0.3% calcium to prevent weight gain and had ad libitum access to deionized water. After acclimation for one week, the rats were randomly assigned to experimental OVX and sham groups (n = 20 per group). In brief, rats were anesthetized by an intraperitoneal injection of 3% sodium pentobarbital (P3761, Sigma-Aldrich Chemical Company, St Louis MO, USA). Bilateral ovariectomy was conducted in the OVX group under sterile conditions, as previously described [19]. The remaining sham-operated 20 rats had their ovaries surgically exteriorized and the replaced. .
For the establishment of postmenopausal osteoporosis (PMO) animal model, the femoral and lumbar spine samples were retrieved from the rats in the OVX and sham groups and scanned with a micro-CT scanner (SCANCO Medical AG, Brüttisellen, Switzerland). SCANCO Medical microCT systems software were adopted to calculate the trabecular bone scanning parameters, and for 3D analysis, visualization, image management and data import/export (SCANCO Medical AG). Morphometric indices of the trabecular bone region were determined from microtomographic data sets using direct three-dimensional morphometry. The microarchitecture parameters in the evaluation of BMD, trabecular thickness (Tb.Th), bone volume/total volume ratio (BV/TV), bone cortical thickness, and cortical porosity were determined. Femoral and lumbar spine samples (16 cases with 8 in each group) were fixed by immersion in 4% paraformaldehyde for 48 h and the decalcificated with 10% ethylenediamine Downloaded from http://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200314/901300/bcj-2020-0314.pdf by guest on 14 January 2021 Biochemical Journal. This is an Accepted Manuscript. You are encouraged to use the Version of Record that, when published, will replace this version. The most up-to-date-version is available at https://doi.org/10.1042/BCJ20200314 5 tetraacetic acid (EDTA) for 6 weeks, whereupon they were prepared into 4-μm paraffinembedded sections before hematoxylin and eosin (HE) staining.

Alizarin red S (ARS) staining
MSCs at passage 3 were incubated into 6-well plates at 3 × 10 4 cells/well. Following 24 h of culture in the growth medium, the osteogenic induction medium was added and changed every 3 days. After 14-day osteogenesis induction, MSCs were fixed in 4% paraformaldehyde for 20 min, stained for 10 min with 40 mM ARS staining solution and finally photographed under a microscope (TE2000, Nikon, Japan).

RNA isolation and quantitation
Total RNA was isolated from hippocampus tissues using TRIzol reagents (16096020, Thermo Fisher Scientific, New Your, USA), which was then reverse-transcribed into cDNA using TaqMan™ MicroRNA Reverse Transcription Kit (4366596) and High-Capacity cDNA Reverse Transcription Kit (4368813; all from Thermo Fisher Scientific, New York, USA). miR-199a-3p primer sequences were provided from an miRNA cDNA synthesis Kit and an miRNA reverse transcription quantitative polymerase chain reaction (RT-qPCR) SYBR kit (TaKaRa, Tokyo, Japan). The primers are depicted in Table 1. Relative quantification (2 -ΔΔCt method) was applied to calculate the fold changes which were standardized by U6 (for miR-199a-3p) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH, for remaning genes).

Dual-luciferase reporter assay
The entire 3'-untranslated region (3'-UTR) gene fragment of Kdm3a was synthesized and introduced into the pGL3-reporter (Promega Inc., Madison, WI, USA). The mutant type (MUT) sequence was designed in the wild type (WT) of Kdm3a. Following restriction endonuclease treatment, WT and MUT were inserted into the pGL3-reporter. The correctly sequenced WT and MUT plasmids were then co-transfected with mimic-NC, miR-199a-3p mimic, inhibitor-NC, or miR-199a-3p inhibitor into HEK293T cells respectively. After 48 h of culture, the cells were lysed and the luciferase activity was measured using the Dual-Luciferase ® Reporter Assay System (Promega Inc.) and TD-20/20 Luminometer (E5311, Promega Inc.).

Statistical analysis
Statistical analysis was operated using SPSS 21.0 statistical software (IBM Corp., Armonk, NY, USA) with p < 0.05 as statistically significant. Measurement data were expressed as mean ± standard deviation. Data obeying normal distribution and homogeneity of variance between two Downloaded from http://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200314/901300/bcj-2020-0314.pdf by guest on 14 January 2021 Biochemical Journal. This is an Accepted Manuscript. You are encouraged to use the Version of Record that, when published, will replace this version. The most up-to-date-version is available at https://doi.org/10.1042/BCJ20200314 8 groups were compared using unpaired t-test, while one-way analysis of variance (ANOVA) or repeated measures ANOVA was utilized to compare data among multiple groups, followed by Tukey's test.

Successful establishment of rat osteoporosis model and isolation of bone marrowderived MSCs
All rats remained healthy 3 months after surgery. From mirco-CT images, it was observed that the characteristics of normal skeleton in sham-operated rats changed significantly in OVX rats (80.6%), CD90 (94.6%) and CD45 (0.046%) in sham-operated rats and corresponding expression in OVX rats (82.3%, 79.2%, 94.2% and 0.151%, respectively) were consistent with MSC characteristics (Fig. 1H). The RT-qPCR results revealed reduced expression of Alp, Bglap (Ocn) and Runx2 was reduced in femoral tissues of OVX rats (p < 0.05; Fig. 1I). ARS staining revealed that the bone marrow-derived MSCs exhibited impaired osteogenic differentiation in OVX rats, in contrast to those with sham treatment (Fig. S1A).

miR-199a-3p is overexpressed in OVX rats and inhibits osteogenic differentiation in bone marrow-derived MSCs
The correlation of miRNA expression with the progression of osteoporosis has been extensively reported. For instance, miR-365a-3p inhibits the differentiation of osteoblasts by targeting RUNX2 and thus promotes the development of osteoporosis [21]. miR-34a prevents osteoporosis and bone metastasis by inhibiting osteoclast production and Tgif2 [8]. In addition, miR-7b-5p attenuates the progression of osteoporosis by inhibiting adipose differentiation of human bone marrow-derived MSCs via regulating IRS2 [22]. Collectively, miRNAs can thus affect the development of osteoporosis in many different ways. miR-199a-3p is highly expressed in rats with OVX-induced osteoporosis, and promotes the adipogenic differentiation of MSCs [10,23]. The present study was designed to study the mechanism of how miR-199a-Downloaded from http://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200314/901300/bcj-2020-0314.pdf by guest on 14 January 2021 9 3p intervenes in osteoporosis. First, RT-qPCR revealed high expression of miR-199a-3p in tissues and cells of OVX rats compared to tissues and cells of sham-operated rats ( Fig. 2A, p < 0.05), suggesting the inhibitory role of overexpressed miR-199a-3p in osteogenic differentiation.

Kdm3a is downregulated in OVX rats and its overexpression promotes osteogenic differentiation in bone marrow-derived MSCs
Next, we aimed to elucidate the downstream regulatory mechanism underlying miR-199a-3p.

Kdm3a expedites osteogenic differentiation by upregulating Erk2 and Klf2 expressions in vitro
To explore the downstream factors of Kdm3a, ChIP assay manifested that the enrichment of Kdm3a was decreased in MSCs treated with sh-Kdm3a, while that of H3K9me1, H3K9me2, and H3K9me2 was enhanced (Fig. 5A
As revealed by Western blot assay, treatment with miR-199a-3p antagomir or BIX01294 resulted in an enhancement in the extents of Erk2 phosphorylation and Klf2 expression in rat femoral tissues (p < 0.05, Fig. 6E). As shown in Fig. 6F, G, the expression of Alp, Sp7(Osx), Bglap and Runx2 was increased in femoral tissues of miR-199a-3p antagomir-treated rats, and a similar trend was noted follwoing BIX01294 treatment. Coherently, these results suggested that the inhibition of miR-199a-3p could facilitate osteogenic differentiation to relieve osteoporosis in vivo.

Discussion
This dissertation unraveled that miR-199a-3p was upregulated in OVX rats and that miR-199a-3p upregulation inhibited osteogenic differentiation in vitro. Furthermore, downregulation of Kdm3a was observed in OVX rats, while the elevation of Kdm3a expression expedited osteogenic differentiation by promoting the expression of Erk2 and Klf2 in vitro. Meanwhile, Kdm3a was found to be negatively targeted by miR-199a-3p. These findings indicate the participation of miR-199a-3p/Kdm3a in osteogenic differentiation.
In previous work, miRs have been associated with the adipogenic commitment of MSCs [23]. Notably, the influence of miR-199a-3p on adipocyte differentiation has been previously highlighted, such that our findings concurred in showing decreased miR-199a-3p expression during differentiation in OVX rats [24]. Accordingly, we considered the possibility that miR-199a-3p may assume a significant role in osteogenic differentiation. In this work, results of the in vitro experiments highlight that miR-199a-3p overexpression can indeed suppress osteogenic differentiation of MSCs. Consistently, miR-199b-5p overexpression has been reported previously to assume a depressive role in osteogenic differentiation in vertebral ligamentum flavum cells [25]. ALP and Runx2 are important markers for osteoblastic differentiation, and their enhanced expression has been shown to be a factor in the promotion of osteogenic Downloaded from http://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200314/901300/bcj-2020-0314.pdf by guest on 14 January 2021 12 differentiation [26]. Besides, the upregulation of Alp and Sp7(Osx) has also been reported to act as a promoter of osteoblasts differentiation [27]. Consistent with these studies, our results documented the lower expression of ALP, Sp7(Osx), Bglap and Runx2 in response to miR-199a-3p overexpression. Osteoporosis has the feature of diminished BMD and microarchitectural deterioration of bone tissues owing to the augmented osteoclastic bone resorption [28]. In our experiments in vivo, we saw increased BMD after repressing miR-199a-3p in OVX rats, which indicated a notable increase in osteogenic differentiation.
To elucidate better the mechanism underlying the suppression of osteogenic differentiation of MSCs by miR-199a-3p, we investigated the correlation between miR-199a-3p and Kdm3a expression, which was further confirmed by dual-luciferase reporter assay, suggesting that Kdm3a was negatively targeted by miR-199a-3p. Kdm3a possesses diverse functions in the regulation of associated genes, which often are involved in metabolism and cell differentiation [29]. In addition, the histone demethylase Kdm3a has also been found to regulate the transcriptional program of the androgen receptor in prostate cancer cells [30]. Another study highlighted the contributory role of Kdm3a during cell differentiation, suggesting that Kdm3a assumes a pivotal role in the differentiation into parietal endoderm-like cells in carcinoma [31].

Conclusions
The data of our study provides novel insights into the mechanism underlying osteogenic differentiation. We uncovered that miR-199a-3p was upregulated in bone marrow-derived MSCs and OVX rat tissues, and that its downregulation could induce osteogenic differentiation by recruiting DNA demethylase Kdm3a at the Erk2 and Klf2 promoters (Fig. 7). Given this mechanism, we suppose that miR-199a-3p may serve as a therapeutic target for the treatment of osteoporosis in the future, and see a need to test this approach in prospective studies in osteoporosis patients.

Funding
None.

Competing Interests
The authors have no competing interests to declare.

Data Availability Statement
The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.  sham-operated rats. Measurement data were expressed as mean ± standard deviation and unpaired t-test was adopted to analyze the differences between two groups. Measurement data were expressed as mean ± standard deviation and unpaired t-test was adopted to analyze the differences between two groups. Cell experiments were conducted three times independently. Measurement data were expressed as mean ± standard deviation and unpaired t-test was adopted to analyze the differences between two experimental groups. Cell experiments were conducted three times independently. MSCs transfected with miR-199a-3p mimic + oe-NC or miR-199a-3p mimic + sh-NC.
Measurement data were expressed as mean ± standard deviation and unpaired t-test was adopted to analyze the differences between two groups. Cell experiments were conducted three times independently. 20 data were expressed as mean ± standard deviation and unpaired t-test was adopted to analyze the differences between two groups, while ANOVA was utilized to compare the data among multiple groups.         portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200314/901300/bcj-2020-0314.pdf by gue