Inhibition of the Extracellular Signal-regulated Kinase Pathway Restores the Discogenic Phenotype of Inammatory Intervertebral Disc Cells

Background: Intervertebral disc (IVD) degeneration is a spinal disease caused by trauma and/or repetitive mechanical overloading of the spine which triggers inammatory response pathways. Long-term disc inammation may lead to development of spinal pseudoarthrosis. The aim of the present study was to elucidate the role of the extracellular signal-regulated kinase (ERK) pathway in inammation-induced IVD cells. Methods: Inammatory human nucleus pulposus cells (NPC) were stimulated using tumor necrosis factor alpha (TNFα) and the ERK pathway was blocked using a selective molecule-based inhibitor U0126. Gene expression of catabolic and anabolic events, pro-inammatory, and NPC markers were investigated. The enzymatic activity of matrix metalloproteinases (MMP)2/9 were determined by gelatin zymography. The cytoxicity of U0126 concentrations on NPC was quantied using resazurin assay, and the specicity of U0126 on ERK1/2 signaling was determined. Results: The pro-inammatory cytokines like MMP3/13 and interleukin 6 in nucleus pulposus (NP) inammatory conditions were down-regulated by U0126 and a trend towards an increase of the NP-specic collagen type 2, aggrecan and keratin 19 was observed suggesting a recovery of the NP phenotype. U0126 does not seem to have effect on prostaglandin production, aggrecanases and some anabolic genes. We conrmed that U0126 selectively blocks the ERK phosphorylation and U0126 affects the cells metabolic activity only for high concentrations. Conclusions: Inhibition of ERK signaling down-regulates important metalloproteinase, pro-inammatory cytokines, and up-regulates NP markers in order to restore the discogenic phenotype of inammatory NPC.


Background
Low back pain affects millions of people in industrialized societies every year and represents, therefore, a leading cause of disability with signi cant economic and social burdens [1][2][3]. Chronic low back pain is strongly associated with intervertebral disc (IVD) degeneration and consists mainly in increasing pain during the execution of daily spinal movements. The IVD consists of an inner nucleus pulposus (NP) surrounded by the annulus brosus (AF) tissue, and a hyaline articular cartilage is located at the endplates between the IVD and two adjacent vertebral bodies. The NP is a gelatinous-like and avascular connective tissue containing a highly organized extracellular matrix rich in proteoglycans and collagens with few numbers of cells. The NP cells (NPC) actively regulate the homeostasis in the extracellular matrix by several growth factors and cytokines acting in an autocrine and paracrine fashion. The early stage of disc degeneration is often asymptomatic [4] and tends to develop a chronic in ammation within the spine. This produces high stress in the disc tissue, which begins with synthesis of catabolic enzymes [5], resulting in degradation and loss of the extracellular matrix (ECM) structures present in the IVD. The loss of IVD structures leads to leakage of the connective tissue from the NP through the AF, which causes compression of peripheral nerves, resulting in pain and inability of spinal movements.
Non invasive interventions including medications, steroid injection and physical therapy are most commonly recommended, and often of limited e cacy [6-7] since they do not maintain or restore the native tissue structure in degenerative discs. Currents surgical treatments are aimed to remove discogenic pain by replacement of the injured tissue with a functional biological substitute or prosthesis, however, surgery for disc degeneration also yielded to mixed clinical outcomes and often results in incomplete interbody fusion [8].
Interleukin 1 beta (IL1β) and tumor necrosis factor alpha (TNFα) are pro-in ammatory cytokines know as key mediators of the development/progression of disc degeneration and low back pain [5]. It has been shown, for instance, in many studies that TNFα is highly expressed in degenerative IVD tissues [9][10].
Surgical samples obtained from patients with history of low back pain revealed higher levels of TNFαpositive cells than autopsy from healthy controls [11]. It causes an up-regulation of ECM-degrading enzymes and decreases the expression of matrix-speci c matrix proteins [12]. Thus identifying and targeting signaling pathways responsible for the TNFα-mediated in ammation could be considered as a promising therapeutic option to regenerate the IVD tissues.
Extracellular signal-regulated kinase (ERK), a downstream of the mitogen-activated protein kinase (MAPK) signaling cascade, is an important in ammatory pathway that plays a critical role in the production of in ammatory cytokines and the activation of procatabolic responses induced by TNFα in chondrocytes-lineages cells [13]. Inhibition of MAPK signaling attenuates the decrease of collagen type II and aggrecan without inducing apoptosis in primary rat and immortalized chondrocytes [14]. In agreement with this, inhibition of MAPK/ERK activity enhances chondrogenesis of mesenchymes [15], and TNFα-induced NF-κB DNA binding in chondrocytes is reduced by inhibition of MAPK signaling [14].
Recent studies suggested a catabolic pathway of ERK in the degeneration of intervertebral disc. In AF cells, ERK was shown to mediate IL1-induced upregulation of cyclooxygenase [16], important aggrecanses and metalloproteinases [17], while ERK inhibition decreased the IL1β-induced apoptosis [18].
The role of the ERK pathway in effecting the action of TNFα in NPC, however, has not been fully elucidated. The aim of the present study is to clarify the role of ERK in in ammation-induced human NP cell culture model.

Human NP isolation and culture
Human IVD were collected from trauma patients undergoing spinal fusion surgery. The procedure was performed with patients' written consent and was approved by the local ethical committee of the canton of Bern, Switzeland. The disc materials were collected from patients aged between 17 and 51 years old (34 ± 16.9 [mean ± SD]). The NP tissue was separated from the outer AF of the disc by an experienced surgeon and subsequently processed within 24 h after the surgery. Data were anonymized and solely the sex and age were recorded from each donor. The NPC were isolated by sequential digestion of the NP tissue fragments with 1.9 mg/mL pronase (Roche Basel, Switzerland) for 1 h followed by collagenase II (Worthington, London, UK) at 37°C overnight on a plate shaker. The remaining undigested NP tissue debris was removed by ltration through a 100 µm cell strainer (Falcon, Becton Dickinson, Allschwil, Switzerland) and cell viability was determined by Trypan blue exclusion.
The NPC were expanded in proliferation medium (Dulbecco's Modi ed Eagle Medium [DMEM, Sigma-Aldrich, Basel, Switzerland] containing 10% Fetal Bovine Serum [FBS, Sigma-Aldrich] and penicillin/streptomycin [P/S, 100 U/mL and 100 µg/mL, respectively, Merck, Darmstadt, Germany]) until con uency. In order to amplify the low number of NPC obtained after digestion of the biopsy, the NPC were expanded in the proliferation medium until reaching several passages and a cell stock was cryopreserved for further analysis. Low-passage (less than 3 passages) NPC were used in this study.
Induction of pro-in ammatory environment and ERK inhibition NPC were seeded at a density of 5 × 10 4 cells/well in 24-well plates in the proliferation medium and left overnight for cell adherence. Induction of a pro-in ammatory environment in NPC was performed by addition of 10 ng/mL of human recombinant TNFα (Peprotech, London, UK) to the wells. Control cultures represented NPC with (positive control) or without (negative control) TNFα.
Alternatively, the cells were pretreated with a selective molecule-based inhibitor of ERK, U0126 (Selleck Chemicals, Houston, USA) at 0.5 µM and 5 µM for 1 hour to inhibit ERK pathway activation. The cultures were maintained for 3 days and collected thereafter for downstream applications.

NP cells mitochondrial activity
The NPC were seeded at a density of 2 × 10 3 cells/well in 96-well plates in the proliferation medium and left overnight for cell adherence. To address whether blocking the ERK pathway and a combination of TNFα and ERK inhibition may induce a cytotoxic effect on IVD cells, NPC were stimulated with 10 ng/mL TNFα and increasing concentration of a U0126 ranging from 0.1 µM to 10 µM. The cultures were collected after 3 days and the cells viability was determined with a resazurin red solution (Sigma-Aldrich) as previously described [19]. Brie y, the cultures were incubated with 50µM resazurin red solution in a humidi ed atmosphere (5% CO 2 , 37°C) for 2 hours. The absorbance was measured at 580 nm on a microplate reader (SpectraMax M5, Bucher Biotec, Basel, Switzerland).

Analysis of gene expression
Total RNA was extracted as previously described [20] from the NPC after stimulation with TNFα and inhibition of the ERK signaling. NP hall markers, including aggrecan (ACAN), collagen type II (COL2A1) and cytokeratin 19 (KRT19), anabolic markers including insulin-like growth factor 1 (IGF1) and transforming growth factor beta 1 (TGFβ1), catabolic markers including matrix metalloproteinase 3 (MMP3), MMP13 and a disintegrin and metalloproteinase with thrombospondin motifs 5 (ADAMTS5), pro-in ammatory markers including IL6 and cyclooxygenase-2 (COX2), and the ribosomal 18S RNA as reference gene, were determined. Human-speci c oligonucleotide primers (Table 1) (Microsynth, Balgach, Switzerland) were designed with Beacon DesignerTM (Premier Biosoft, California, Palo Alto, USA) based on the sequences of the nucleotides from GenBank. The e ciency and melting curves of the amplicons were tested to determine speci c ampli cation. Table 1 Custom-designed DNA primers used in real-time qPCR study. Amplicons were generated using a 2-step ampli cation cycling (95° for 15s and 57°C for 30s for 45 cycles) and SYBR-green mastermix.

Gelatin zymography
A gel zymography was performed in cell lysates after treatment of NPC with TNFα and/or inhibition of the ERK signaling pathway to analyze the pattern of MMP2/9 in different treatment groups. Brie y, the cell layers were washed with PBS and homogenized with 0.5% Triton-X (Sigma-Aldrich) containing a protease inhibitor (Sigma-Aldrich). The extracts were applied to a gelatin-containing 10% SDS-PAGE and the gel was incubated in a developing buffer composed of 50mM Trizma Base and 5mM CaCl 2 (pH 8.0) overnight at 37°C. Subsequently, the polyacrilamide gel was stained with 0.1% Coomassie blue R250 solution (Sigma-Aldrich) for 2 hours.

Western Blot analysis
Primary human NP cells (250'000 cells/well) were seeded into 6-well culture plates and grown for one week in the proliferation medium without additional factors to allow for cell adherence. The medium was changed and treatment of the cells with TNFα, TNFα together with U0126 (0.1 µM, 0.5 µM, 1 µM and 5 µM) and DMSO (control) was initiated. The cells were pre-treated with various concentrations of U0126 for 1 hour and subsequently TNFα (10 ng/mL) was applied to the cultures in order to simulate an in ammatory environment. After 30 min of treatment, adherent cells were rinsed with PBS and immediately lysed in the cell lysis buffer (CelLytic, Sigma-Aldrich) containing a protease inhibitor cocktail (25µg/ml, Sigma-Aldrich) and phosphatase inhibitor (1 mM NaF, 1 mM sodium orthovanadate, 4 mM sodium tartrate, 1.15 mM sodium molybdate and 2 mM imidazole, Biotool, USA). The total protein content in the cell layers was determined by Bradford protein assay. The samples (10 µg) were subjected to 12% SDS-PAGE and transferred to PVDF membrane. Non-speci c binding sites were blocked using 5% skim milk in Tris-buffered saline with 0.05% Tween (TBST) for one hour at room temperature. The membrane was washed several times with TBST followed by incubation with a rabbit anti-phospho-ERK1/2 and/or ERK1/2 primary antibodies (Cell Signaling Technology, USA) at 4°C overnight. Detection was achieved using a goat anti-rabbit antibody (LI-COR Biosciences GmbH, Germany), and the band intensity was quanti ed with Odyssey Software (LI-COR).

Statistical analysis
Differences in real time PCR and resazurin sodium salt assay were evaluated by a one-way ANOVA with Bonferroni's post-hoc test using GraphPad Prism (version 6.0 f for Mac OS; GraphPad Software Inc., La Jolla, CA USA). p < 0.05 was considered signi cant.

U0126 effect on the NP metabolic activity
The cellular metabolic activity was analyzed using the rezasurin assay to verify whether U0126 induces any cytotoxic effect on the cells. There was a dose-dependent decrease in the NPC viability with increasing concentrations of U0126 as compared to the control group (TNFα). The NPC metabolic activity was signi cantly reduced (p = 0.0016, Fig. 1) by U0126 at 0.5µM and beyond 1µM (p < 0.0001). These data suggested that U0126 is a toxic component at high concentrations (micro-molar range). In addition, TNFα (10ng/ml) alone signi cantly reduced (p = 0.001) the NPC metabolic activity as compared to untreated cells. The TNFα-stimulated NPC were morphologically different, spindle shaped with broblastlike appearance and less dense compared to untreated cells as observed microscopically.

Gelatinases expression
Gel zymoraphy was performed in cell layer lysates for the detection and assessment of the gelatinases (MMP2/9) expressed in the different treatment groups. It was able to detect both the MMP2 and MMP9 and their respective pro-MMP2/9 in the lysates (Fig. 3). However, the band intensities were almost the same in the lysates with insigni cant variations upon the treatment groups.

Activation of ERK signalling
Western blots were carried out to assess the effects of TNFα and U0126 on the ERK mitogenic dependent intracellular signaling. TNFα induced the ERK signaling pathway in human NPC as demonstrated by the phosphorylation of ERK1/2 (Fig. 4A). We found that the peak ERK phosphorylation was achieved after 30 minutes of TNFα treatment in NPC. The cytokine-mediated activation of the ERK pathway was typically transient with a maximal TNFα stimulation obtained followed by decline of p-ERK dynamics after the rst 30min of TNFα treatment (Fig. 4B). Further western blots investigating the effect of U0126 on ERK pathway in TNFα-stimulated NPC were carried out at 30min of cytokine exposure in order to bene t from the maximal TNFα stimulation effect. U0126 abolished the the TNFα-mediated ERK phosphorylation in a dose-dependet manner (Fig. 5A). U0126 treated cells at 1µM along with TNFα resulted in phosphorylated ERK results comparable to the untreated control group (Fig. 5B).

Discussion
Current treatments to alleviate chronic low back pain are principally surgical with very variable outcomes.
For example, a 10-40% failure rate was observed in lumbar surgeries with or without spinal fusion, and similar results were obtained in patients whom underwent a discectomy of herniated disc with recurrence after two years of surgery [21][22][23][24]. Innovative strategies for IVD with regenerative medicine and/or pharmacological inhibitors of disc degeneration are urgently wanted.
The present study investigated the contribution of ERK signaling in in vitro in ammation-induced human NPC by TNFα and whether ERK inhibition using selective mitogen-activated protein kinase kinase enzyme (MEK) inhibitor can reverse the in ammatory phenotype of these cells. We found that TNFα stimulation in NPC induced a pro-in ammatory microenvironment of human IVD cells, in particular NPC, characterized by increased expression of important catabolic events and pro-in ammatory mediators. Herein, the expression of matrixproteinases including MMP3 and MMP13 was upregulated in TNFα-stimulated NPC. MMP are believed to be the major proteolytic enzymes responsible for ECM degradation in the IVD leading to disc degeneration [25]. The expression of numerous metalloproteinases at the transcript and protein levels has been studied in several human IVD as well in experimental animal models revealing the catabolic changes and their mediation in the progression of IVD degeneration [26-28].
This study further supports that the TNFα cytokine is a potent mediator of in ammatory response in IVD and in particular in NPC leading to ECM degradation through increased catabolism and therefore to disc degeneration. In ammatory processes exacerbated by TNFα and/or IL1β are believed to trigger disc degeneration and in later stages low back pain. For instance, surgical samples obtained from patients with history of low back pain revealed higher levels of TNFα-positive cells than autopsy from healthy controls [11]. In addition, these intradiscal pro-in ammatory cytokines are implicated in the onset and progression of IVD degeneration and discogenic pain, and are produced by native IVD cells including NPC and AFC as well to in ltrating macrophages [5,[9][10]. Taken together, TNFα is an essential initiator of a pro-in ammatory environment in IVD tissue and cells which leads to the tissue ECM degradation and disorganization, and therefore to disc degeneration and painful spine. The importance of TNFα cytokine in discogenic pain led to multiple clinical trials using TNFα inhibitors, which resulted in mixed results [29][30][31] and highlighting, therefore, for further research studies [12]. In particular, monoclonal antibodies against TNFα have shown promise for mitigating disc degeneration and relieving low back pain. Anti-TNFα treatment signi cantly decreases the concentration and activity of MMP1 and MMP3 in ex vivo IVD tissues isolated from patients with herniated discs [32]. Despite obvious bene t of TNFα mAbs, some patients do not respond to them and/or many will develop recurrent disease despite continuing dosing which hampers the clinical use of these antibodies. [33]. In addition, In iximab which is a TNFα blocker and a chimeric IgG1 antibody did not appear to interfere with spontaneous resorption of disc herniation over a prolonged period based on MRI diagnosis in a randomized controlled study [34]. Therefore, further research to elucidate the mechanism by which in ammatory cascade is initiated through TNFα is required for targeted pharmacological treatment of IVD degeneration.
Within this study, we aimed to block the TNFα downstream signaling pathway by targeting the ERK from the MAPK family. First, we reproduced an in vitro in ammatory environment in NPC and subsequently the ERK pathway was blocked through U0126. The current study sought to elucidate the role of ERK1/2 signaling pathway in TNFα-mediated catabolic effect in NPC.
We used the resazurin red assay to examine the possibility of any cytotoxic effect of the ERK inhibitor on NPC. It was found that U0126 was cytotoxic in NPC at large concentrations translated by a dosedependent reduction of cells' viability with increasing U0126 molarity. This nding further supports the critical role of MAPK, in particular ERK1/2 pathway, in the regulation of mammalian cell proliferation as previously documented [35].
Blocking the ERK pathway in in ammation-induced human NPC resulted in a down-regulation of MMP3 and MMP13 to similar levels observed in control NPC, at least for the mRNA levels. It was previously demonstrated that CCAAT/enhancer binding protein beta (C/EBPβ) in the TNFα promoter region was suppressed in the presence of an ERK inhibitor PD98059 and the p38-MAPK inhibitor SB202190, but not the JNK inhibitor SP600125 in rat NPC [36]. In addition, the C/EBPβ and MMP13 expression was colocalized in chondrocytes in in ammatory arthritic patients [37]and that pro-in ammatory cytokines such as IL1β and TNFα binds to MMP3 and MMP13 promoter regions and stimulates their expression [37][38].
Similarly, treatment of rat NPC with ERK1/2 inhibitors (PD98059 and U0126) abolished the antagonistic effect of TGF-β1 on TNFα mediated MMP3 catabolic response [39], which further supports our nding on the implication of ERK pathway in in ammatory human NPC. Taken together, TNFα induces an in ammatory cascade in mammalian cells, in particular NPC, by up-regulation and modulation of MMP family members such as MMP3/MMP13 through ERK1/2 pathway, and inhibition of ERK signaling can reverse this catabolic effect. ERK1/2 is, therefore, considered a downstream signaling pathway of TNFα and this MAPK might be a target for the increased MMP enzymatic activity. Although, the enzymatic activities mediated by the gelatinases MMP2/9 were visible on gel zymography in NPC cell layer lysates, we could not detect signi cant variations of MMP2/9 following treatment of NPC with TNFα. This could be explained by an inappropriate model for the detection of MMP2/9, the intracellular gelatinases MMPs are normally secreted in the extracellular compartment for a variety of cell lineages [40][41][42]. Therefore, it is more relevant to assess the incorporated gelatinases in culture supernatant rather than within the cells lysates [43] as also detected previously for IVD cells [44].
The activation of an in ammatory microenvironment through stimulation of NPC with TNFα resulted in an increased ADAMTS5 expression as compared to controls. Blocking the ERK1/2 pathway with U0126, however, did not alter the expression of ADAMTS5. This observation might be explained by a differentiational regulation of aggrecanase-mediated proteoglycan degradation including ADAMTS4-5 which is mediated through NF-B activation and not ERK1/2 in bovine NPC [45]. Similarly, it was suggested that ADAMTS4 expression and promoter activity increased in NPC following TNFα and IL-1β treatments [46] and treatment of the cells with NF-B inhibitor abolished this inductive effect of the cytokines on ADAMTS4 mRNA and protein expression. This further supports that modulation of ADAMTS5 in NPC is mediated through NF-B which might explain our observation on the role of ERK1/2 in ADAMTS5 expression.
The results of this study revealed a trend towards an increase in transcripts levels of the NP-speci c markers including COL2 and ACAN, but also a signi cant increase of KRT19 in in ammatory NPC that were treated with U0126. In addition, increased expression of anabolic genes like IGF1 was observed suggesting a restoration of the NPC phenotype following the inhibition of the ERK pathway in in ammatory cells. Wei et al. [17] observed an imbalance between anabolic and catabolic events in rat AFC activated with IL1, and ERK inhibition signi cantly blocked the catabolic and in ammatory effects of IL1 in AFC.
Within this study, we provided evidence that TNFα-mediated in ammation in human NPC is triggered through the ERK1/2 pathway which in terms increased pro-in ammatory mediators like MMPs and decreased anabolic genes characteristic of a degenerated IVD cells. The ERK1/2 pathway was modulated through TNFα as observed by Western Blot assay and simultaneous treatment of NPC with TNFα and U0126 abolished this effect. We also showed that U0126 is a speci c inhibitor of ERK1/2 pathway.
There are some limitations of the current study. First, due to the di culty and availability in obtaining healthy and non-degenerated human disc tissues samples, the results were limited to gene expression analysis of some anabolic and catabolic genes, and we did not investigate their protein levels more in details. Second, we did not address the contribution of other MAPK like p38-MAPK and JNK in in ammation, although some pilot data were generated (data not shown). In addition, we used in vitro monolayer cultures of NPC, which might differ from 3D cultures like cells-seeded scaffolds or ex vivo IVD samples and the combined effects of in ammatory mediators with biomechanical stimuli was not addressed which is the aim for further investigations.

Conclusions
In conclusion, the current study provides evidence that ERK pathway is implicated in NPC in ammatory process and ERK1/2 inhibition could provide some protection against the adverse effect of TNFα.

Consent for publication
Written informed consent was obtained from the patients/participants for publication of the manuscript and gures.

Availability of data and materials
Data of this study are available from the corresponding author upon reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This study was supported by the Swiss National Science Foundation (project # 310030_153411).     Western blot of phosphorylated ERK1/2 (p-ERK) and their respective total ERK1/2 in NPC after TNFα stimulation. The NPC were stimulated with TNFα (10 ng/ml) and the cells lysates were collected at different time points (minutes until 2h) with control being time 0 min before stimulation (A). p-ERK and total ERK were determined using western blot, the signal intensities were quanti ed, and the p-ERK were normalized to total ERK (B).