MicroRNA-338-3p as a novel therapeutic target for intervertebral disc degeneration

Recent studies have demonstrated the pivotal role played by microRNAs (miRNAs) in the etiopathogenesis of intervertebral disc degeneration (IDD). The study of miRNA intervention in IDD models may promote the advancement of miRNA-based therapeutic strategies. The aim of the current study was to investigate whether intradiscal delivery of miRNA can attenuate IDD development. Our results showed that miR-338-3p expression was significantly increased in the nucleus pulposus (NP) of patients with IDD. Moreover, there was a statistically significant positive correlation between the expression level of miR-338-3p and the severity of IDD. Our functional studies showed that miR-338-3p significantly influenced the expression of extracellular matrix synthesis genes, as well as the proliferation and apoptosis of NP cells. Mechanistically, miR-338-3p aggravated IDD progression by directly targeting SIRT6, a negative regulator of the MAPK/ERK pathway. Intradiscal injection of antagomir-338-3p significantly decelerated IDD development in mouse models. Our study is the first to identify miR-338-3p as a mediator of IDD and thus may be a promising target for rescuing IDD.


INTRODUCTION
Intervertebral disc degeneration (IDD, OMIM 603932), one of most common spine disorders, is characterized by the progressive loss of disc space, endplate sclerosis, and low back pain (LBP) 1 . It is among the main factors affecting quality of life and exacts a huge burden on the global healthcare system 2,3 . An intervertebral disc is composed of the central nucleus pulposus (NP) and the outer annulus fibrosus (AF). The NP is involved in maintaining the homeostasis of proteoglycan and type-II collagen (Col II) in the extracellular matrix (ECM) 4 . A hallmark of IDD is the degradation of proteoglycans in the NP, leading to decreased disc height and reduced biomechanical function 5,6 . Clinically, IDD causes disc herniation and sciatica, which are associated with back pain and chronic disability. With regard to therapeutics, IDD presents with a complex of degenerative diseases. Conservative and surgical treatments have limited long-term effects in many IDD cases, as these methods only target the clinical symptoms but do not focus on the etiology of IDD 7 . Recent advancements in gene therapy have opened an avenue for the possible prevention of IDD development at the transcriptional or post-transcriptional levels [8][9][10][11] . There is therefore the need for a more in-depth understanding of IDD to facilitate the development of new therapeutic strategies 12 .
MicroRNAs (miRNAs), a group of endogenous noncoding RNAs, have emerged as key regulators in the development of skeletal pathophysiology 13 . Studies have reported that they bind to the 3′ untranslated region (3′ UTR) of target mRNAs, resulting in the modulation of gene expression 14 . Accumulating evidence has demonstrated that miRNAs act as central modulators in essential cellular processes, including cellular proliferation and apoptosis 15,16 .
Notably, dysregulation of miRNAs has been reported to be related to various degenerative diseases in humans [17][18][19][20] . Differential expression patterns of miRNAs have been noted in the pathogenesis of IDD. Several miRNAs, such as miR-155, miR-93, and miR-141, were significantly downregulated or upregulated in patients with IDD and thus were identified as targets of IDD-associated gene expression ranging from matrix-degrading enzymes to proinflammatory cytokines [21][22][23][24] . In the current study, we performed a comprehensive screening of 2000 miRNAs in human NP tissue. miRNA expression profiling led to the identification of miR-338-3p as substantially increased in degenerative NP tissue compared with control NP tissue. Moreover, miR-338-3p expression was positively correlated with the Pfirrmann grade of IDD. Our functional studies revealed that miR-338-3p induced IDD by promoting the catabolism of NP cells and NP cell apoptosis via the sirtuin 6 (SIRT6)/ mitogen-activated protein kinase (MAPK)/extracellular signalregulated protein kinase (ERK) pathway. Our in vivo study investigating the possible effects of intradiscal injection of an miR-338-3p inhibitor in mouse models revealed that this microRNA exerted an NP-protective effect. Thus, our findings provide, for the first time, crucial evidence that miR-338-3p holds promise as a possible target for the prevention of IDD.

MATERIALS AND METHODS Patient samples
Nucleus pulpous (NP) tissues were collected from 110 patients with IDD who underwent lumbar microdiscectomy (mean age 56.3 ± 9.1 years). All IDD patients were clinically diagnosed by two qualified spine surgeons via physical examination and MRI. On the basis of the Pfirrmann classification 25 , all patients were classified as having mild/moderate degeneration (grades 1 or 2) or serious degeneration (grades 3 or 4). Normal NP tissues were collected from 103 patients with lumbar vertebral fracture who underwent anterior spine surgery (mean age 53.7 ± 4.6 years). These trauma patients had no prior history of preoperative low back pain or IDD.

The injury-induced IDD model and therapeutic experiment
The injury-induced IDD mouse model was established via needle puncture as previously described 23,26 . Briefly, following general anesthesia of 12week-old C57BL/6 mice, the mouse coccygeal discs Co6/Co7 were exposed and then punctured with a 31-G syringe needle through the AF to the NP. The needle was then left in the intervertebral disc at a depth of 1.5 mm for a period of 10 s. The adjacent Co7/Co8 disc levels were not punctured, as they were used as contrast segments.
For recombinant adenovirus construction, the sequence for sirt6 from human genomic DNA, which was cloned by PCR, was inserted into the pDC311-U6-MCMV-EGFP vector (Ad-sirt6) (Hanbio Co. Ltd, Shanghai, China), and sirt6 siRNA was purchased from Life Technologies. Adenoviral vectors encoding green fluorescent protein (Ad-GFP) were used as controls. A total of 15 injury-induced IDD mice were included in the study and were randomly divided into three groups (n = 5). Three, seven and fourteen days after puncture intervention, the intervertebral discs were injected with Ad-sirt6 (1 × 10 9 viral particles), Ad-GFP (1 × 10 9 viral particles), and sirt6 siRNA, respectively. The mice were sacrificed by pentobarbital overdose eight weeks after adenovirus injection.
The ethics committee of The First Affiliated Hospital of Guangxi Medical University approved the animal experimentation protocol (2018-KY-NSFC-025), and all animal experimentations were conducted in accordance with approved guidelines.

Cell cultures
Human NP tissues were sliced into 1-mm 3 pieces and washed twice with 10% fetal bovine serum (PBS) (Gibco, NY, USA), followed by incubation in DMEM (Gibco, NY, USA). Subsequently, a resuspension of NP cells was cultured in DMEM, 1% penicillin-streptomycin, and 10% FBS. The cells were then cultured in an incubator at 37°C and 5% CO 2 , dissociated from the substrate with trypsin, and split at a ratio of one to three.

Quantitative real-time polymerase chain reaction (qRT-PCR)
TRIzol (Invitrogen Life Technologies, CA, USA) was used for total RNA extraction. The RNA templates were then synthesized into cDNA using an iScripts cDNA Synthesis kit (Quanta BioSciences, MD, USA), and GAPDH was used as the control for normalization. The isolated miRNA was then quantified using a qScript microRNA cDNA synthesis kit (Quanta BioSciences, MD, USA), with U6 snRNA used as the internal control. Then a SYBR Green real-time PCR kit (Quanta Biosciences) was used to perform qRT-PCR, with a comparative threshold cycle (ΔΔCt) adopted for the calculation of gene expression. The primer sequences are shown in Supplementary Table 1.

miRNA microarrays and miRNA target prediction
The miRNA microarray study was conducted using an Affymetrix 4.0 miRNA Array (Thermo Fisher Scientific, MA, USA). Total RNA was obtained from six individual NP samples with or without IDD. Based on a volcano plot and fold-change filtering, we identified the differentially expressed miRNAs in the IDD samples. Candidate miRNAs were included in the analysis they met the following criteria: (1) exhibited more than a 5-fold change or less than a 0.2-fold change, and (2) differences had p-values less than 0.05. Gene Cluster software (Stanford University) and DAVID software were used for hierarchical cluster analysis and functional group analysis, respectively. TargetScanHuman (www.targetscan.org/) and microRNA.org were used to forecast the miRNA target genes and to analyze mRNAbinding sites.
Luciferase constructs and reporter assay SIRT6 mRNA 3′ UTR fragments, including the binding sites of wild-type (wt) and mutant (mut) miR-338-3p, were inserted into a psiCHECK-2 luciferase reporter vector (Promega, WI, USA). An miR-338-3p expression plasmid or a control plasmid was then cotransfected with SIRT6 wt-or mut-3′ UTR and their respective control plasmids into human NP cells. All luciferase assays were performed using a Dual-Glo Luciferase Assay System (Promega, WI, USA), and luciferase activity was measured by a fluorescence microplate reader (BioTek, USA).

Flow cytometry, cell counting kit-8 (CCK-8), and EdU assay
Cellular apoptosis was measured using an apoptosis kit with Annexin V-FITC/PI (Cat. No: V13242, Invitrogen), and analysis was performed with Beckman Coulter EPICS Altra (USA). The detection of cellular proliferation was performed using CCK-8 (Dojindo Laboratories, Kumamoto, Japan). Human NP cells that were transfected with different sequences (miR-338-3p mimics, control mimics, miR-338-3p inhibitor, control inhibitor, SIRT6, SIRT6 siRNA, or control siRNA) were then added into 96-well plates and cultured for 24, 48, and 72 h. Next, CCK-8 was added to the cells, which were incubated for 3 h, and the absorption was then evaluated at 450 nm. The degree of cellular proliferation was assessed by EdU assay. Human NP cells were incubated with EdU medium (Sigma-Aldrich, MO, USA), and the cells were collected and stained with Hoechst 33258.

Fluorescence in situ hybridization (FISH)
For the analysis of human NP using FISH, digoxin-labeled miR-338-3p probes with locked nucleic acid modifications were designed and synthesized using Exiqon QIAGEN (Hilden, Germany). The fluorescence signals were detected using a FISH kit (Exiqon-QIAGEN) with a Nikon A1Si laser-scanning confocal microscope (Nikon, Tokyo, Japan) used for image analysis.

Immunofluorescence and TUNEL staining
Intervertebral discs were fixed, decalcified, and embedded in 10% formalin, 10% EDTA, and paraffin. Sagittal sections of disc tissue were then cut every 3 μm from the midsagittal plane. The sections were then stained with Alcian blue for the histological analysis. Masuda's method was then used to grade the histological staining score 27 . The cultured NP cells were then fixed, permeabilized, and blocked with 4% paraformaldehyde, 0.3% Triton X-100, and 1% BSA. For the immunofluorescence analysis, the following primary antibodies were incubated with sections of cells: anti-MMP9 and anti-Col II antibodies (ab73734 and ab34712, respectively; Abcam, Cambridge, UK) at 4°C overnight. The sections and cells were then treated with a secondary antibody for 20 min at room temperature, followed by the capture of immunofluorescence images using a Zeiss LSM780 confocal microscope (Carl Zeiss, Oberkochen, Germany). For the detection of apoptotic activity, a TUNEL assay was performed using a DeadEnd Colorimetric TUNEL System in accordance with the manufacturer's protocol (Cat. No: G7360, Promega).

Statistical analysis
All statistical analyses were performed using SPSS version 17.0 software (SPSS Inc., Chicago, IL). Differences between two or more groups were analyzed by two-tailed unpaired Student's test or one-way ANOVA with Tukey's post hoc test. A p-value less than 0.05 was considered statistically significant.

Identification of differentially expressed miRNAs between IDD patients and normal controls
To investigate dysregulated miRNAs, we scanned the expression profile of miRNAs based on microarrays of human NP tissues (3 IDD patients vs. 3 normal controls). The microarray revealed approximately 600 differentially expressed miRNAs between the IDD and control patients (Fig. 1a). The results of the unsupervised clustering analysis of miRNAs are shown in a volcano plot (Fig. 1b) and heat map (Fig. 1c). We identified a total of 14 upregulated and nine downregulated miRNAs that exhibited a mean fold change of more than 5-fold or less than 0.2-fold as well as a p-value less than 0.05. Specifically, miR-338-3p, miR-198, and miR-874-3p were found to be significantly dysregulated in the NP of IDD samples but not in that of the control samples (Supplementary Table 2). We performed qRT-PCR to identify the three most highly expressed miRNAs based on a cohort of 110 patients with IDD and 103 normal controls. Among the most highly expressed miRNAs, miR-338-3p was observed to be the most significantly upregulated in the NP tissues obtained from IDD patients (Fig. 1d). Therefore, it was selected as the candidate miRNA for further study. FISH confirmed that miR-338-3p was significantly increased in the NP of IDD patients compared to that of the control (Fig. 1e). Our results also revealed that the expression levels of miR-338-3p were positively correlated with the Pfirrmann grade of IDD (n = 56, r = 0.855, P < 0.01, Fig. 1f). In human NP cells, the expression levels of miR-338-3p were also found to be upregulated in IDD patients but not in normal controls (Fig. 1g).
The effect of miR-338-3p on the proliferation and apoptosis of human NP cells To determine the influence of miR-338-3p on the etiopathogenesis of IDD, we found that transfection with miR-338-3p mimics or inhibitor affected the cellular proliferation and apoptosis of human NP cells. The high transfection efficiency of miR-338-3p was confirmed by Cy3-labeled fluorescence and qRT-PCR (Fig. 2a,  b). At three different time periods (24 h, 48 h, and 72 h), the miR-338-3p inhibitor significantly activated the cellular proliferation of NP cells, whereas the miR-338-3p mimics significantly repressed the cellular proliferation of NP cells (Fig. 2c). Thus, to determine the effects of miR-338-3p on the proliferation of NP cells, an EdU assay was performed (Fig. 2d). Correspondingly, the transfection of miR-338-3p mimics significantly promoted the apoptosis of NP cells (Fig. 2e). However, the silencing of miR-338-3p notably upregulated the expression of ECM-related proteins (Col II and Aggrecan), whereas overexpression of miR-338-3p significantly enhanced the expression of matrix-degrading enzymes (MMP9 and ADAMTS4) (Fig. 2f). These observations were further supported by immunofluorescence assay (Fig. 2g). Thus, our findings illustrated that the silencing of miR-338-3p can promote the proliferation of NP cells and enhance matrix synthesis.

SIRT6 is a target gene for miR-338-3p
A heat map was used to identify the dysregulated mRNAs in IDD (Fig. 3a). Gene ontology analysis predicted the functional characterization of downregulated genes, including those involved in disc development, ECM structural constituency, and extracellular region (Fig. 3b). Based on the miRNA-mRNA network and Venn analysis, SIRT6 was identified as a target gene for miR-338-3p (Fig. 3c-e). Furthermore, the computational alignment scores indicated a high conservation of miR-338-3p among various animal species (Fig. 3f). The luciferase reporter assay confirmed that miR-338-3p mimics can inhibit the luciferase activity of the wild-type SIRT6-3′ UTR. In contrast, the miR-338-3p mimics failed to inhibit the luciferase activity of the mutant SIRT6-3′ UTR (Fig. 3g). Our results thus indicated that miR-338-3p can directly bind to the SIRT6 3′ UTR, which suppressed the mRNA and protein expression levels of SIRT6 in human NP cells (Fig. 3h, i).

DISCUSSION
To the best of our knowledge, our study (both in vitro and in vivo) is the first to report on the possible functional role that can be played by miR-338-3p in the search for a viable therapeutic target for IDD treatment. In accordance with miRNA expression profiling, we found that the expression levels of miR-338-3p were remarkably increased in degenerated human discs, which was further confirmed via qRT-PCR in an independent cohort. Notably, we found a noteworthy positive correlation between miR-338-3p expression and the severity of IDD. Thus, our results indicated that miR-338-3p can serve as a possible novel biomarker for the prognosis of IDD. Based on several computational tools, SIRT6 was determined to be the direct target of miR-338-3p. Furthermore, our results revealed that the expression levels of SIRT6 were downregulated in degenerated NPs and that miR-338-3p can directly interact with SIRT6. Our findings also showed an inverse relation between miR-338-3p and SIRT6 and demonstrated that both miR-338-3p and SIRT6 may be important mediators in the pathogenesis of IDD.  Fig. 3 SIRT6 as a direct target gene of miR-338-3p. a Microarray analysis showing the differential gene expression between patients with IDD and normal controls. b Gene ontology analysis showing the highest enrichment scores of the upregulated GO terms. c, d Identification of SIRT6 as a potential regulatory target of miR-338-3p. e Sequential alignment showing the binding sites between miR-338-3p and SIRT6. f Computational alignment scores indicating high conservation of miR-338-3p. g Relative luciferase activity detection in human NP cells cotransfected with miR-338-3p mimic/inhibitor and SIRT6 3′UTR-wild/mutant-type reporter plasmids. h, i Analysis of the expression levels of SIRT6 mRNA and protein via qRT-PCR and western blotting following transfection with miR-338-3p mimics/inhibitor in human NP cells. ***P < 0.01.
SIRT6 is a member of the sirtuin family of deacetylases that are involved in diverse pathologies, including neurodegeneration, metabolic homeostasis, and aging [30][31][32] . The link between SIRT6 and MAPK signaling has been well reported in multiple systems. For instance, SIRT6 has been reported to protect against hepatic ischemia/reperfusion injury by regulating the transcriptional status of MAPK/ERK signaling-related genes 33 . Recent studies have provided evidence that increased activation of the MAPK signaling pathway promotes disc degeneration [34][35][36] . However, the possible functional role of SIRT6 deacetylase activity in the MAPK signaling pathway is yet to be fully understood. We hypothesize that SIRT6 may be involved in influencing the development of IDD through the MAPK signaling pathway and that its deacetylase activity is essential in engaging this process. Our study is possibly the first to identify that SIRT6 can modulate MAPK/ERK signaling via transcriptional control of GRB2, which is an upstream activator of this signaling pathway. From the results of our coimmunoprecipitation study, we observed that the enhanced expression of GRB2 was likely a direct consequence of the deacetylating activity of SIRT6. Furthermore, our findings revealed that the upregulation of GRB2 expression in NP cells led to a significant increase in the expression of target proteins in MAPK/ERK signaling (p-ERK1/2), as well as those involved in cellular senescence or apoptosis (P16, Bax, and Caspase 3), ultimately resulting in the upregulation of the expression levels of MMP9 and ADAMTS4 and downregulation of the expression of Col II and Aggrecan. The modulation of the MAPK signaling pathway induced by SIRT6 siRNA or the miR-338-3p mimics is possible via the regulation of GRB2 expression. Nevertheless, our findings do not exclude the possibility that other genes are involved in the control of MAPK/ERK signaling. Deeper exploration into the understanding of the mechanism by which miR-338-3p regulates the expression of SIRT6 and GRB2 is needed in future studies.
Our study provides a novel perspective for the clinical translation of miR-338-3p as a diagnostic and treatable option for IDD. The increased expression of miR-338-3p observed in IDD patients and animal models signified that miR-338-3p can be used as a stable biomarker for this common musculoskeletal disease. Our in vitro study revealed that in the absence of miR-338-3p, the expression of SIRT6 was upregulated and the MAPK/ ERK signaling complex was suppressed, resulting in a reduced apoptosis rate of NP cells. Our results therefore indicated the viability of using miR-338-3p as a possible direct therapeutic target for reversing the development of IDD. In our preclinical study, we found that antagomir-338-3p (an miR-338-3p inhibitor) protected intervertebral discs against the typical pathological pattern of IDD by inhibiting the apoptosis of NP cells, enhancing the expression of extracellular matrix proteins (Col II and Aggrecan) and inhibiting the expression of extracellular matrixdegrading enzymes (MMP9 and ADAMTS4). These findings suggest that the possible underlying therapeutic mechanism of miR-338-3p inhibition is reversing the imbalance between catabolic and anabolic factors involved in IDD development. Collectively, our study introduces a new option for miR-338-3pbased therapeutics in IDD prevention.
In summary, our study revealed that miR-338-3p was significantly increased in human degenerative NPs and involved in the direct targeting of SIRT6, leading to the activation of the MAPK/ERK signaling pathway in NP cells. Our results showed that antagomir-338-3p, an miR-338-3p