Dysregulated Bone Metabolism Is Related to High Expression of miR-151a-3p in Severe Adolescent Idiopathic Scoliosis

Adolescent idiopathic scoliosis (AIS) is a common complex disease, and bone homeostasis plays an important role in its pathogenesis. Recent advances in epigenetic research show that dysregulated miRNAs may participate in the development of orthopedic diseases and AIS. The aim of this study was to detect differentially expressed miRNAs in severe AIS and elucidate the mechanism of miRNA deregulation in the pathogenesis of AIS. In the present study, miRNA expression profiles were detected in severe and mild AIS patients as well as healthy controls by miRNA sequencing. Candidate miRNAs were validated in a larger cohort. Primary osteoblasts from severe AIS patients were extracted and isolated to determine the effect of the candidate miRNAs on bone metabolism. Finally, we determined the methylation level in primary osteoblasts from severe AIS patients. The result showed that miR-151a-3p was overexpressed in severe AIS patients. Reduced GREM1 expression was observed in primary osteoblasts from severe AIS patients. miR-151a-3p directly inhibited GREM1 in primary osteoblasts. Relatively lower methylation levels were detected in primary osteoblasts from severe AIS patients. In conclusion, our study revealed that plasma miR-151a-3p levels may serve as a biomarker for severe AIS. Overexpression of miR-151a-3p may interrupt bone homeostasis via inhibiting GREM1 expression. Our result may provide a new biomarker for the early detection of AIS and increase our understanding of the pathogenesis of AIS.


Introduction
Adolescent idiopathic scoliosis (AIS) is a common complex disease affecting 0.4-5.0% children worldwide [1,2]. The spinal curve continues to deteriorate in approximately 10% AIS patients [3]. Severe AIS patients suffer from physical and mental issues, such as severe back pain and lower selfesteem [4]. Currently, the only treatment for severe AIS is surgery, which is costly and associated with long-term implications [5]. Therefore, it is important to perform early screening and preventive intervention for early-stage AIS patients. Previous studies suggested that several factors, such as age, Risser sign, and premenarche status, serve as predictive factors for curve progression [6,7]. However, the actual predictive power of these factors is insignificant. The reason is partly attributed to our limited understanding of the pathogenesis of AIS. Recent studies suggested that genetic factors play an important role in the pathogenesis of AIS [8]. It is important to identify several genetic and epigenetic signatures of curve progression.
miRNA is a class of small noncoding RNAs (21-25 nt) that are essential for numerous key cellular processes, such as cell differentiation and proliferation [9]. It is widely reported that miRNAs are deregulated in cancers as well as metabolic and degenerative diseases [10][11][12]. Recent advances in epigenetic research of AIS reveal noncoding RNAs (ncRNAs), including long ncRNAs (lncRNAs) and microRNAs (miRNAs), as promising biomarkers for early screening and prognostic predictions of AIS [13]. For instance, a recent study reported that a lowly expressed miRNA MIR4300 is related to severe AIS [14]. In another study, severely deregulated miRNAs were identified and suggested for use as biomarkers for AIS diagnosis [15]. In addition, some studies explored the mechanism underlying differentially expressed miRNAs between AIS and healthy control. A downregulated lncRNA in mesenchymal stem cells (MSCs) was associated with the suppression of osteogenic differentiation and may be involved in the pathogenesis of AIS [16]. Impaired osteocyte function in AIS was correlated with deregulated miR-145-5p/β-catenin signaling [17]. However, the detailed mechanism between miRNAs and severe scoliosis requires further investigation.
In the present study, miRNA expression profiles were detected in severe and mild AIS patients and healthy controls by miRNA sequencing. After bioinformatics analysis, candidate miRNAs were validated in a larger cohort. Primary osteoblasts from severe AIS patients were extracted and isolated to determine the effects of candidate miRNAs on bone metabolism. Finally, we determined the methylation level in primary osteoblasts from severe AIS patients. The aim of this study was to detect differentially expressed miRNAs in severe AIS and elucidate the mechanism of miRNA deregulation in the pathogenesis of AIS.

Ethics Statements. The Ethics Committee of Xiangya
Hospital of Central South University approved the protocol of this study. All participants involved in this project signed a written consent form, and guardians provided a signature for subjects under the age 18 of years. All data are used for scientific research.

Subjects.
For the miRNA sequencing cohort, all serum samples were obtained from 10 AIS patients, including five severe and five mild patients, and five healthy controls. For the qPCR validation cohort, a total 40 AIS patients and 40 healthy controls were recruited. In addition, facet joints and bone tissues of posterior column vertebrae were harvested from 21 severe AIS patients with a Cobb angle greater than 50 degree and 20 nonscoliosis patients with thoracolumbar spinal injuries during surgery. The clinical characteristics of each cohort are presented Table 1. Two spinal surgeons confirmed the diagnosis of each AIS patient according to physical examination and the standard anteroposterior radiography method. In this study, severe AIS patients exhibited a maximum Cobb angle of greater than 50°, and mild AIS patients exhibited a maximum Cobb angle less than 40°. All patients with familial inherited diseases, congenital vertebral malformations, neuromuscular diseases, skeletal dysplasia, and connective tissue abnormalities were excluded from the study.
2.3. RNA Extraction and Small RNA Sequencing. A peripheral venous blood sample was aspirated from each subject in the morning, and the plasma was isolated and stored in an -80°C freezer for future analysis. Total cell-free RNA was extracted from plasma using the miRNeasy Serum/Plasma kit (Qiagen, CA, USA) per the protocol's instructions. A small RNA library was established by using TruSeq the Small RNA Sample Prep Kit (Illumina, San Diego, CA, USA). Then, the qualified library was loaded on an Illumina HiSeq 2500 platform for high-throughput sequencing (Lianchuan Biotechnology, Hangzhou, China). The raw data of each sample were filtered by ACGT101-miR (LC Sciences, Houston, Texas, USA) to remove junk sequences. Next, all of the 18-to 26-nucleotide sequences were mapped to the miRBase21.0 database (http://www.mirbase.org/) to identify known miRNAs and novel 5p or 3p miRNAs [18]. Differentially expressed miRNAs were assessed using an analysis of variance (ANOVA) test. To predict the target genes of each miRNAs, TargetScan7.2 (http://www.targetscan.org/vert_72/) and miRDB (http://mirdb.org/) databases were used to identify the potential binding site [19,20].

Real-Time
Quantitative PCR (RT-PCR). Total RNA was extracted from the bone tissue of the facet joint, posterior column vertebrae, or primary osteoblasts using TRIzol reagent (Invitrogen, Thermo Fisher Scientific) according to the protocol's instruction. A miRNA cDNA Synthesis Kit (Thermo Fisher Scientific) was used for miRNA reverse transcription. In brief, an adaptor sequence, i.e., 3 ′ poly-A tailing and 5 ′ ligation, was added to mature miRNA. Then, a standard reverse transcription reaction was performed. For mRNA reverse transcription, cDNA was synthesized using the Taq-Man cDNA synthesis kit (Applied Biosystems, CA, USA) following the manufacturer's protocol. All primers used in this study are provided in supplemental table S2. Real-time PCR was conducted by using the SYBR Green Master Mix (Takara, Kusatsu, Shiga, Japan) according to the manufacturer's protocol. GAPDH and U6 were used as internal controls.
2.5. Isolation and Culture of Primary Osteoblasts. For human primary osteoblast isolation, chopped cancellous bone tissues from the facet joint or posterior column vertebrae were isolated as described before [21]. In brief, samples were washed with phosphate-buffered solution (PBS) thrice and then incubated with 0.25% trypsin (Gibco, CA, USA) for 30 min at 37°C. Samples were centrifuged after complete medium was added. The supernatant was discarded, and the pellet was resuspended in 0.1% collagenase type I (Sigma-Aldrich, St. Louis, MO, USA). The sample was incubated in a 37°C incubator for 4 h and then cultured with F-12 medium (Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS) in an incubator (37°C and 5% CO 2 ).
2.6. Western Blot. Total cellular proteins from primary osteoblasts were collected in lysis buffer with protease inhibitors. Protein samples were then sonicated and boiled for 5 min. After quantification using the BCA protein assay (Thermo Fisher Scientific, MA, USA), twenty micrograms of each protein sample were loaded into a 10% SDS gel for separation and then transferred to a PVDF membrane (Bio-Rad, CA, USA). The membranes were blocked with skim milk for 1 h at room temperature and incubated with antibodies against GREM1 (1 : 2000, HPA007526, Sigma-Aldrich) overnight at 4°C on a shaker. After three washes in PBST, the membrane was incubated with goat anti-rabbit IgG antibody (1 : 10000, ab150077, Abcam) for 1 h at room temperature. The blot was detected using the Chemiluminescence Protein Detection Module (Millipore, CA, USA). GAPDH (1 : 5000, 2 BioMed Research International EPR16891, Abcam) was used as an internal control. Band intensity was measured using the ImageJ (v 1.52p; National Institutes of Health) program.
2.9. Staining. Alizarin Red staining was conducted after primary osteoblasts were transfected with miR-151a-3p mimic and incubated with 50 ng/ml BMP-2 for two weeks. Briefly, cells were cultured in six-well plates, incubated with BMP-2 and transfected with miR-151a-3p mimic. After two weeks, the cells were fixed with 4% paraformaldehyde (PFA) for 30 min. After three washes in PBS, cells were incubated with 0.4% Alizarin Red S (Sigma-Aldrich, MO, USA) for 10 min. Staining was assessed by microscopy. The calcium concentration was determined using 10% cetylpyridinium chloride (CPC) (Sigma-Aldrich, MO, USA). Cells were incubated with 10% CPC for 15 min at room temperature. Then, the diluted sample was transferred into a 96-well plate. Absorbance at 562 nm was detected using a plate reader (BD Biosciences, CA, USA).  Values are shown as the means ± SD. AIS: adolescent idiopathic scoliosis; BMI: body mass index. † P < 0:05 and ‡ P < 0:01 between the severe AIS group and mild AIS group; * P < 0:05 and * * P < 0:01 between the severe AIS group and control group.
3 BioMed Research International test was performed to compare the anthropometric data from the AIS and control groups. Receiver operating characteristic (ROC) curve analysis was conducted to determine the efficiency of each miRNA in predicting progression of the Cobb angle. Two-sided Student's t-tests were performed to compare experimental results from the AIS and control groups. P values < 0.05 were considered statistically significant.

Result
3.1. Differentially Expressed MicroRNA Profiles. To identify differentially expressed microRNAs among severe and mild AIS patients and nonscoliosis controls, we performed plasma miRNA deep sequencing on a cohort of five severe AIS patients, five mild AIS patients, and five healthy controls. The clinical characteristics of three groups were similar with the exception of the main curve Cobb angle ( Table 1). The results indicated that 32 miRNAs were differentially expressed in the three groups (P < 0:05) (Figure 1(a)). In total, 39 miRNAs were differentially expressed in the severe AIS group compared with the healthy control group, including 23 upregulated and 16 downregulated miRNAs (Figure 1(c), supplemental table S1). Comparing the severe AIS group with the mild AIS group, 11 miRNAs were differentially expressed, including three upregulated and eight downregulated miRNAs (Figure 1(d)). Eight miRNAs were differentially expressed, including seven upregulated and one downregulated miRNAs, in the mild AIS group compared with the healthy control group. Venn diagram analysis revealed that no shared differentially expressed miRNAs were noted between the mild AIS vs. control groups and severe vs. mild AIS groups, indicating that no representative miRNAs for the mild AIS group can be detect in our cohort (Figure 1(b)). Regarding the severe AIS group, seven miRNAs were shared by all the groups with the exception of mild AIS vs. the control group. Among all the seven miRNAs, three were upregulated in each profile: miR-941, miR-151a-3p, and miR-148b-5p.

Validation of Candidate miRNAs in a Larger Cohort.
To identify the expression pattern of candidate miRNAs among severe, mild AIS patients, and healthy controls, quantitative real-time PCR was performed to determine plasma expression levels of miR-941, miR-151a-3p, and miR-148b-5p in a larger verification cohort consisting of 40 severe AIS patients, 40 mild AIS patients, and 40 healthy controls. The clinical characteristics of the validation cohort are presented in Table 1. The results demonstrated that miR-941 and miR-151a-3p expression levels are significantly altered in the three groups (Figures 2(a) and 2(b)). Significantly increased expression of these two miRNAs is noted in the mild AIS group compared with the control group. However, the miR-148b-5p expression pattern was not consistent with miRNA sequence results (Figure 2(c)). The predictive power of these miRNAs for progressive scoliosis was calculated based on receiver operating characteristic (ROC) analysis. The area under the curve (AUC) values for each miRNA were 0.845, 0.885, and 0.762 (corresponding 95% confidence inter-vals are 0.767-0.904, 0.815-0.936, and 0.676-0.835) (Figures 2(d)-2(f)). In addition, miR-151a-3p exhibits the highest expression differences among the three groups and the most powerful diagnostic accuracy for severe AIS among all these three miRNAs. Thus, miR-151a-3p was selected as the candidate miRNA involved in the development of progressive AIS and used in subsequent studies.

Target Prediction of miR-151a-3p and Expression Level
Validation. To further verify the relationship between miR-151a-3p and progressive AIS, we obtained facet joint and other bone structures from postcolumn vertebrae from severe AIS patients during corrective surgery and extracted the total RNA and protein from the cancellous bone of the facet joint and vertebral plate. We determined miR-151a-3p expression levels by quantitative real-time PCR. The miR-151a-3p levels of AIS patients are significantly increased compared with nonscoliosis controls (Figure 3(a)). To identify the downstream target gene of miR-151a-3p in AIS, we employed two miRNA target prediction databases, TargetS-can7.2 and miRDB, to predict the potential targets of miR-151a-3p (supplemental table S3, 4). Among the top 100 predicted targets of each database, 25 genes were the same as demonstrated by the Venn diagram (Figure 3(b)). We screened these 25 genes based on gene function and expression pattern since most of them are not involved in skeletal development or expressed in bone tissue. Finally, five predicted target genes were identified as potential genes that interact with miR-151a-3p, including GREM1, ME1, PFN2, YTHDF3, and NAMPT. Then, we determined mRNA expression levels of these genes in posterior column vertebrae from severe AIS patients. In AIS patients, only GREM1 is significantly downregulated compared with the control group (Figure 3(c)). To further confirm this result, we isolated primary osteoblasts from bone tissues of the posterior column vertebrae and determined GREM1 protein expression levels in primary cells. Reduced GREM1 protein levels were noted in AIS patients (Figures 3(d) and 3(e)).

GREM1 Is
Directly Regulated by miR-151a-3p. To further elucidate the interaction between miR-151a-3p and GREM1, luciferase reporter assays were performed by cotransfecting a wild-type or mutant GREM1 luciferase reporter vector with miR-151a-3p mimic into HEK-293T cells. The potential binding site and mutant GREM1 reporter vector sequence is reported (Figure 4(a)). The level of fluorescence was significantly decreased in cells cotransfected with the wild-type GREM1 3 ′ UTR plasmid and the miR-151a-3p mimic (Figure 4(b)). The results indicate that GREM1 is directly regulated by miR-151a-3p.

Overexpression of miR-151a-3p Inhibits Mineralization of
Primary Osteoblasts by Suppressing GREM1 Expression. GREM1 expression can be stimulated by BMP2. Thus, GREM1 exhibits a BMP2 antagonist effect. The balance between these two factors is necessary for skeletal homeostasis. To elucidate the detailed role of elevated miR-151a-3p expression in AIS progression, we first evaluated the effect of miR-151a-3p on GREM1 mRNA expression in primary    BioMed Research International osteoblast incubated with 50 ng/ml BMP2. BMP2 incubation increased GREM1 mRNA expression levels, which is consistent with previous studies. Overexpression or inhibition of miR-151a-3p along with BMP2 incubation significantly decreased or increased GREM1 expression (Figure 5(a)). Then, we performed Alizarin Red staining to evaluate the mineralization level of primary osteoblasts incubated with 50 ng/ml BMP2 with or without 50 nM miR-151a-3p mimic transfection for two weeks. The results showed that miR-151a-3p overexpression combined with BMP2 incubation significantly increased the mineralization of AIS primary osteoblasts (Figures 5(b) and 5(c)).

Discussion
AIS is a complex disease, and genetic factors play an important role in its pathogenesis. Although previous studies have reported many causative genes and related underlying mechanisms [22], it is still difficult to predict scoliosis development and progression in adolescents. In the current study, we performed microRNA sequence analysis in AIS patients with severe and mild spine curvature, which is typically referred to as the Cobb angle and identified miRNAs associated with progressive AIS. We next identified miR-151a-3p as a biomarker for severe AIS using bioinformatics analysis combined with expression validation. Here, miR-151a-3p may also contribute to the development of a greater Cobb angle via inhibition of GREM1 expression in osteoblasts to interrupt bone homeostasis, which is an influential factor in spine development. Finally, we observed relatively lower methylation levels in primary osteoblasts from severe AIS patients, which may cause high miR-151a-3p expression levels. This study linked the imbalance in epigenetic regulation with severe AIS, which might partially explain why mutations have not been identified in a large proportion of AIS patients.
The relationship between miRNA and AIS has been described in a few recent papers. In a study based on RNA sequencing data from plasma samples, circulating miR-122-5p, miR-27a-5p, and miR-223-5p levels were upregulated in AIS patients [15]. In addition, another study identified seven highly expressed miRNAs in bone marrow mesenchymal   Figure 4: Validation of miR-151a-3p targeting GREM1 mRNA. (a) Schematic figure demonstrating alignment of miR-151a-3p, wild-type, and mutant GREM1 3′UTR sequences. (b) Luciferase reporter assay was performed after HEK-293T cells were cotransfected with a random sequence. miR-151a-3p or control mimic as well as the plasmid vector with wild-type or mutant GREM1 3′UTR. * P < 0:05 versus the control group (n = 3).