NOTOCHORDAL CELL-DERIVED MATRIX INHIBITS MAPK SIGNALING IN THE DEGENERATIVE DISC ENVIRONMENT

Chronic low back pain is often caused by intervertebral disc (IVD) degeneration. Preceding this degenerative process, the main cellular phenotype in the nucleus pulposus shifts from notochordal cells (NCs) to nucleus pulposus cells (NPCs). In previous studies, porcine NC-derived matrix (NCM), containing NC-secreted factors, induced matrix anabolic effects and inhibited pro-inflammatory mediators in NPCs in vitro and in degenerated


Introduction
Low back pain (LBP) is a major health and socioeconomic problem throughout the world, and its incidence is rising due to ageing of the human population (Clark and Horton, 2018).Degeneration of the intervertebral disc (IVD) is a common cause for chronic LBP (Freemont, 2009).Current treatments for chronic LBP are only symptomatic, and include physiotherapy, analgesic medication (e.g.Non-Steroidal Anti-Inflammatory Drugs, opioids), and eventually surgery.Therefore, there is an urgent need for minimally invasive (regenerative) strategies that promote biological IVD repair (Bach et al., 2022).
During IVD degeneration, the activity of catabolic enzymes such as matrix metalloproteinases (MMPs) and disintegrins and metalloproteinases with thrombospondin motifs (ADAMTS) increases under the influence of locally produced inflammatory cytokines, e.g.interleukin-1β (IL-1β) (Le Maitre et al., 2005) and Tumor Necrosis Factor (TNF) (Wang et al., 2017a).Pathways that are known to play a role in this catabolic process in the degenerated IVD include inflammatory signaling pathways, e.g.Mitogenactivated protein kinase (MAPK) and its downstream Nuclear factor-κB (NF-κB) signaling (Daniels et al., 2017;Mi et al., 2018;Ni et al., 2019).Altogether, these changes lead to a loss of healthy extracellular matrix (ECM) and decreased structural integrity of the IVD.Because of abovementioned events, the normal function of the IVD deteriorates, which leads to a vicious circle, since the IVD becomes more vulnerable to damage under loading (Vergroesen et al., 2015).
In the young and healthy IVD, large vacuolated notochordal cells (NCs) are present in the nucleus pulposus (NP) (Hunter et al., 2004).These cells disappear, while smaller chondrocytelike nucleus pulposus cells (NPCs) appear during maturation and ageing in several species that spontaneously develop chronic LBP due to IVD degeneration, including humans and dogs (Bach et al., 2022).Previous work has already demonstrated that NC-secreted biomolecular factors evoke anabolic responses in various cell/tissue types, such as mesenchymal stromal cells (MSCs), annulus fibrosus (AF) cells, endplate (EP) chondrocytes, NPCs, NP explants and in vivo rat IVDs (for all studies, see (Bach et al., 2022)).Thus, NC-based treatment strategies for IVD degeneration leading to chronic LBP, appear to be promising.
In an attempt to exploit the instructive capacity of the specialized NC matrix as a cell-free treatment approach, porcine NC-derived matrix (NCM) from healthy NC-rich NP tissue, as a first step towards clinical translation was previously used (Bach et al., 2018).NCM, containing ECM and biologic factors secreted by NCs, may act rather comparable to demineralized bone matrix, which contains ECM and growth factor components (e.g.Bone Morphogenetic Proteins (BMPs) (Pietrzak et al., 2012)) native to bone and is employed in clinical practice for bone healing (Gruskin et al., 2012).Porcine NCM induced an anabolic response in bovine NPCs and inhibited TNFα and IL-1β expression in vitro (Schmitz et al., 2022;de Vries et al., 2019).These effects seemed to be tissue dependent, as NCM induced more potent anabolic effects compared to bovinederived NP matrix (i.e.NP tissue devoid of NCs), while the latter stimulated deposition of collagen type I typically present in degenerative fibrotic processes (de Vries et al., 2018b).Moreover, NCM also induced in vitro and in vivo ECM anabolic effects on degenerated canine IVDs and inhibited the gene expression of inflammatory cytokines (IL-1β and TNF), production of Prostaglandin E2 (PGE2), and the accumulation of advanced glycation end products (AGEs) evidenced by the brown discoloration of tissue (Bach et al., 2018).
Despite the abovementioned promising observations, the underlying mechanism behind the anabolic effects of NCM on IVD cells remains elusive.Therefore, the aim of this study was to determine the mode of action of porcine NCM in the degenerative IVD environment.Insights into the mechanism will allow for a better understanding of the underlying biology and contribute to the development of NC-based therapeutic approaches.

Experimental setup
To explore the underlying mechanisms of NCM, this study examined its effects, thereby focusing on cell phenotype and inflammatory signaling pathways that could potentially be involved in NCM's catabolic and anabolic mechanisms.For this purpose, canine and human NPCs (both species that spontaneously develop IVD degeneration and LBP (Bach et al., 2022)) were cultured in monolayer with and without NCM.After 6, 24, and 72 hours, reverse transcriptionquantitative polymerase chain reaction (RT-qPCR) was performed on inflammatory markers.To confirm short term inflammatory-like responses, an NF-ĸB reporter chondrosarcoma cell line (Neefjes et al., 2021) was incubated with and without NCM for 6 hours.After 72 hours, protein was collected from cultured NPCs and targeted proteomics was performed using DigiWest technology (Treindl et al., 2016) to determine which pathways were activated by NCM.Furthermore, results were confirmed at tissue level in mildly and moderately degenerated in vivo canine IVDs treated with NCM (immunohistochemistry for specific proteins identified with DigiWest; Dualspecificity phosphatase 5 (DUSP5), (phosphorylated) extracellular signal-regulated kinase ((p-) ERK), and cytokeratin 19 (KRT-19)) and specific proteins of interest (interleukin-1β (IL-1β), interleukin-1 receptor I (IL-1R), Forkhead Box F1 (FOXF1), and Paired box protein 1 (PAX1)) (Fig. 1).

Fig. 1. Experimental setup.
Human and canine nucleus pulposus cells (NPCs) from degenerated IVDs were cultured for 6, 24, and 72 hours in monolayers with and without porcine-derived NCM (10 mg/mL).At each timepoint, samples were collected for RT-qPCR and after 72 hours also for targeted proteomics.Additionally, a chondrosarcoma cell-derived NK-κB reporter cell line was cultured with and without 10 mg/mL NCM for 6 hours.Results were confirmed with immunohistochemistry of mildly and moderately degenerated dog IVDs 6 months after intradiscal injection with either no, 1× or 2× 10 mg/mL NCM (Bach et al., 2018).
Table 1.Details of the canine and human donors used in the cell culture experiments.Canine nucleus pulposus cells (NPCs) were obtained from all IVDs, whereas the human NPCs were obtained from the lumbar part of the spine.

Cell culture and protein isolation
Complete spines were collected from Beagles (chondrodystrophic dog breed) that had been euthanized in unrelated research studies (approved by the Utrecht University Animal Ethics Committee; Centrale Commissie Dierproeven, Table 1).IVDs from human donors (Table 1) were obtained during a standard postmortem diagnostic procedure in which part of the lumbar spine was collected within 48 hours after death, as approved the Local Medical Ethical Committee number (12-364).The material was used in line with the code 'Proper Secondary Use of Human Tissue' as installed by the Federation of Biomedical Scientific Societies.IVDs were opened under sterile conditions and NP tissue was collected by precise separation from AF and EPs (Bach et al., 2015).
For targeted proteomic analysis, NPCs were washed after 72 hours, twice with ice cold Phosphate-Buffered Saline (PBS) and protein was collected by scraping.Protein was collected in 10-30 μL PBS in Eppendorf Protein LoBind tubes (Z666505, Sigma-Aldrich, St. Louis, MO, USA) until further analysis.Provided that NCM stimulated deposition of its ECM components within 72 hours, interfering components were removed with a 2D clean-up kit (80648451, GE Healthcare, Boston, MA, USA) according to the manufacturer's instructions.

DigiWest technology and data analysis
Targeted proteomics was performed on 4.5 ± 0.7 µ g (canine) and 9 µ g (human) protein per sample with DigiWest® (Reutlingen, Germany), a proprietary immunoassay technology which transfers Western Blot to a high-throughput bead-based microarray platform (Treindl et al., 2016).In short, the proteins in each sample were separated by gel electrophoresis (NuPAGE® Novex® 4-12% Bis-Trisprotein gel, 1.0 mm, 12 well, 150 V, 90 min), blotted onto a polyvinylidene difluoride membrane (30 V, 75 min) and biotinylated (one hour), followed by a Ponceau staining.After one wash with Phosphate-Buffered Saline + Tween (PBST) and drying overnight, the lanes were then cut into 96 strips to generate different molecular weight fractions, and the proteins on the strips were eluted in 96-well plates (Greiner Bio-One, Frickenhausen, Germany).After protein elution in 10 µ L 8 M Urea in 100 mM Tris-HCl pH 9.5 including 1 % Triton-X100, neutravidin coated, color-coded MagPlex beads (Luminex, Austin, TX, USA) were added to the proteins.After overnight-coupling, leftover coupling sites were blocked with the use of deactivated NHS-PEG12biotin (500 µ M, 1 h).Beads were pooled and the original Western Blot lanes reconstructed by reassigning the color IDs of the MagPlex beads to the molecular weight fraction.DigiWest beads were blocked in assay buffer (ELISA blocking reagent, Roche, Rotkreuz, Switzerland) supplemented with 0.2 % milk powder, 0.05 % Tween-20, 0.02 % sodium azide) in a 96-well plate (Corning, NY, USA).Beads were incubated in 30 µ L primary antibody (Supplementary File 1) at 15 °C overnight.After washing twice with PBST, R-Phycoerythrinconjugated secondary antibody was added for 1 h at 23 °C.Beads were then washed twice with PBST and readout was performed on a Luminex PlexMAP 3D instrument (Treindl et al., 2016).
For peak identification and quantification of the antibody specific signals, the DigiWest analysis tool was employed.This tool uses the 96 values for each initial lane obtained from the Luminex measurements on the 96 molecular weight fractions, this identifies the peaks at the appropriate molecular weight, calculates a baseline using the local background, and integrates the peaks.The reported values present the peak specific fluorescence intensity (AFI, accumulated fluorescence intensity) (Treindl et al., 2016).
For the canine samples, a total of 237 antibodies for NP phenotypic markers and proteins involved in inflammatory signaling pathways plus species-specific blanks and protein loading controls were employed (Supplementary File 1).For some antibodies, no protein was detected resulting in 71 (canine) and 49 (human) remaining antibodies that were used for further analysis.For these antibodies, the signals were filtered to exclude proteins that were undetectable in the majority of the samples by employing.The cut-off value of > 400 average fluorescence intensity (AFI) was calculated as follow: nSamples*baseline + ngroups*baseline (to enable transformation, all signals below detection limit were arbitrarily assigned a baseline value of 33 AFI).Applying this filter, at the cut-off value of 400 AFI, removed 14 of the 71 antibodies leaving 57 antibodies (canine) and removed 7 of the 49 antibodies leaving 42 antibodies (human) for further analysis.
Analysis of the DigiWest data was conducted in R-Studio (v1.1),R (v3.4.4),Bioconductor 2.3.Scaling and normalization was conducted groupwise based on the streptavidin signal and visualized with multidimensional scaling.Differential protein expression patterns (log2 fold-change; log2FC) were obtained using pairwise exactTests using edgeR (v3.20.9)followed by posthoc False Discovery Rate (FDR) correction according to the Benjamini and Hochberg's method to calculate the adjusted pvalues.Thereafter, the significantly differentially expressed proteins were further analysed using "toppgene" and subsequently, a Panther overrepresentation test using the Reactome, pathways as the annotation data set was performed.For the antibodies producing multiple bands, the summed values over the bands were used and the individual single band values removed.

Immunohistochemical staining of (NCMtreated) canine IVDs
In a previous study, the effect of 10 mg/mL NCM was determined on degenerated canine IVDs in vivo (n = 5 Beagles) (Bach et al., 2018).Briefly, 50 μL of NCM was intradiscally injected in mildly (spontaneously; no NX-IVDs) and moderately (induced by partial NP removal; NX-IVDs) degenerated canine IVDs.After three months, NCM was reinjected in a moderately and mildly degenerated IVD (2× NCM) to determine whether multiple injections would exert a more beneficial effect than a single injection.Six months after the first injections, IVDs were collected for histological purposes.The spinal units (½ vertebra -EP -IVD -EP -½ vertebra) were sagittally transected and the samples were fixed in 4 % buffered formaldehyde for 14 days, decalcified in PBS with 0.5 M Ethylenediaminetetraacetic acid for two months and embedded in paraffin (Bach et al., 2018).
The histological samples (control, 1× NCM, 2× NCM, all from mildly and moderately degenerated IVDs) were included in the current study to determine the effect of NCM on protein immunopositivity in vivo (Table 3).For this purpose, 5-µ m sections were mounted on Microscope KP+ slides (KP-3056, Klinipath, Duiven, The Netherlands) and deparaffinized through xylene (two times 5 minutes) and graded ethanol (100 %, 96 %, 70 %; two times 5 minutes each), followed by one PBS rinse.The sections were blocked for 10 minutes with 0.3 % H2O2 in PBS and washed two times for 5 minutes with PBS + 0.1 % Tween (PBST 0.1 %).If needed, antigen retrieval was performed (Table 3).Thereafter, sections were blocked for 30 minutes with 5 % bovine serum albumin (BSA; A3095, Sigma-Aldrich, St. Louis, MO, USA) in PBS and incubated overnight at 4 °C with the primary antibody in 5 % BSA in PBS and negative control IgGs (Table 3).The negative controls (Table 3) showed no specific staining.The next day, the sections were washed with PBST 0.1 % before the secondary antibody (in similar concentration as primary antibody) was applied for 60 minutes at room temperature (Table 3).After washing with PBS, the sections were incubated with the liquid 3,3'-diaminobenzidine substrate chromogen system (K3468, Dako, Denmark) for 2 minutes and counterstained with Mayers hematoxylin (1.09249.0500,Merck, Darmstadt, Germany) for one minute.Thereafter, they were washed with tap water for 10 minutes and dehydrated with graded ethanol (70 %, 96 %, 100 %) and xylene (two times 5 minutes each) and mounted with Pertex (00811-EX, Histolab, Sweden).Raw images of the NP region (two per NP) were captured with a Leica DFC420C digital camera (Leica Microsystems, Amsterdam, The Netherlands) mounted to a BX60 microscope (Olympus, Leiderdorp, The Netherlands) and Leica Application Suite (V4.2) software package.All positively and negatively stained cells were manually counted (nuclear and cytoplasmatic staining).Adobe Photoshop CC 2017 18.1.0(San Jose, California, USA) was used to manually count (positively stained) cell numbers in two randomly selected NP areas per IVD section as described previously (Bach et al., 2016b).The mean percentage of cells that stained positive over the total number of cells present (ratio) was determined for every target protein.
Immunopositivity thus indicates ratio of positive cells, not intensity of the staining.

Statistical analysis
Statistical analyses on the RT-qPCR and immunohistochemical data were performed using IBM SPSS Statistics 24 (IBM Corp., Armonk, NY, USA).Normal distribution was tested with the Shapiro-Wilk test.If data were normally distributed, a one-way analysis of variance (ANOVA) was used, while if data were not normally distributed, a Mann-Whitney U test was performed.Because multiple samples didn't show any immunopositivity, for the KRT19 immunohistochemical data of the in vivo study a Chi-Square test was performed.The one-way ANOVA and Mann-Whitney U tests were followed by a Benjamini Hochberg post-hoc test to correct for multiple comparisons.p-values < 0.05 were considered as statistically significant.

NCM induces an initial inflammatory response
Gene expression analysis was performed on canine (cNPCs) and human NPCs (hNPCs) from degenerated IVDs cultured for 6, 24, and 72 hours in monolayers with or without 10 mg/mL NCM.IL-1β and TNF mRNAs were not detected at any time point.NCM induced IL-8 expression in cNPCs and IL-6 expression in hNPCs after 6 hours of treatment, but not at later time points (Fig. 2b,d).IL-6 expression was not detected in cNPCs (Fig. 2a) and IL-8 expression was not significantly affected by NCM treatment in hNPCs (Fig. 2e).COX2 expression was increased by NCM in cNPCs after 6 and 24 hours and in hNPCs after 6, 24 and 72 hours of treatment (Fig. 2c,f).Furthermore, increased NF-κB signaling activity was detected in the reporter cell line by its exposure to NCM derived from 3 individual porcine donors (Fig. 2g).

NCM affects multiple signaling pathways
Targeted proteomic analysis showed that several pathways were affected by NCM treatment both in vitro and in vivo, which results are discussed in more detail below.Additional pathways that were affected in canine and human NPCs are included in Table 4 and 5 respectively.

NCM inhibits other pathways associated with proinflammatory mediators
NF-κB (a transcription factor involved in various biological processes, i.e. inflammation, immunity, differentiation, cell growth, and apoptosis) is a homo-or heterodimeric complex formed by the Rel-like domain-containing proteins RELA, RELB, NF-κB1 and NF-κB2 (Bren et al., 2001).Only RELB was lower expressed in NCM-treated cNPCs versus controls, but not in hNPCs (Fig. 8c,f).Affirmatively, pathway analysis indicated that NF-κB and Toll-like receptor cascades were inhibited by NCM (Activation of AP-1 family of transcription factors; Table 4 and 5).Also the Phospho-Phospholipase A2 (PLA2) pathway was inhibited in cNPCs treated with NCM (Table 4).The PLA2 enzymes convert phospholipids into arachidonic acid (a precursor of prostaglandins, associated with inflammation), and facilitate membrane repair and production of inflammatory lipid mediators (Leslie, 1997).

AGE receptor and RUNX2 signaling is negatively influenced by NCM
In cNPCs, but not hNPCs, advanced glycosylation end product (AGE) receptor (RAGE) signaling was negatively affected by NCM (Table 4 and 5).AGEs have been shown to increase levels of reactive-oxygen-species and promote inflammation (Illien-Junger et al., 2013), but can also induce osteogenic differentiation and calcification in the IVD (Illien-Jünger et al., 2016).In line with this, RUNX2-mediated signaling was also negatively regulated by NCM in both hNPCs and cNPCs (Table 4 and 5).

NCM may inhibit angio-and neurogenesis in cNPCs
Reactome pathway analysis indicated that NCM treatment reduced vascular endothelial growth factor (VEGF) and neurite outgrowth signaling in cNPCs (Table 4).

Cell cycle regulation is stimulated by NCM
In cNPCs, expression of Glyceraldehyde 3phosphate dehydrogenase (GAPDH) and cell division cycle protein 2 homolog (CDC2), both involved in cell proliferation/cell cycle regulation, were increased in NCM-treated NPCs versus controls (p < 0.001, Fig. 8a,b).In hNPCs, GAPDH expression was increased, while CDC2 expression was decreased after NCM treatment (p < 0.001, Fig. 8d,e).Furthermore, Reactome analysis of cNPCs showed that cell cycle regulation (mitosis) pathways were stimulated (Table 4).In contrast, senescence and apoptosis pathways were mainly suppressed by NCM in cNPCs, and not affected in hNPCs (Table 4 and  5).
According to Mann-Whitney U statistical analysis, there was a non-significant trend towards increased KRT19 expression in NC-2× NCM IVDs compared with NX-control IVDs (p = 0.1).In accordance with the DigiWest results, additional Chi-Square statistical analysis indicated that the number of dogs in which KRT19 was expressed in the NP was significantly higher in 2× NCM NX-IVDs compared with the control NX-IVDs (p < 0.05, Fig. 10a).In contrast to the DigiWest data, however, PAX1 and FOXF1 immunopositivity was increased in Beagle NX-IVDs treated with 2× NCM compared with NXcontrol IVDs (p < 0.05, Fig. 10b,c).FOXF1 immunopositivity was also increased in 2× NCMtreated noNX-IVDs when compared with controland 1× NCM-treated noNX-IVDs (p < 0.001).The fold enrichment of the proteins in the pathways is represented compared with expected values from the reference list.The pathways in the table indicates over-representation of this category in the analyzed list: more genes are observed than expected on the basis of the reference list (for this category, the number of genes in the list is greater than the expected value) (Mi et al., 2019).Pos: positively influenced pathway, Neg: negatively influenced pathway, Pos/neg: both positively and negatively regulated proteins in this pathway.

Discussion
The aim of this study was to determine the mode of action of porcine NCM in the degenerative IVD environment.Our findings demonstrate that porcine NCM induces an initial inflammatory response, but thereafter exerts a prolonged antiinflammatory effect mainly by influencing the MAPK pathway.This was confirmed at the IVD level in a dog model where disc degeneration was induced by partial nuclectomy (Bach et al., 2018).Furthermore, based on the pathway analysis of differential protein expression upon NCM treatment in canine and human NPCs, new hypotheses involved in the instructive capacity of NCM are indicated.

NCM initially induces pro-inflammatory mediators in native NPCs
In previous work, prolonged anti-inflammatory effects of porcine NCM were shown in vitro in bovine NPCs cultured for 28 days (Schmitz et al., 2022;De Vries et al., 2015) and in vivo in dog IVDs treated with NCM for 6 months (Bach et al., 2018).Interestingly, the current study shows that NCM induced an initial inflammatory response in canine and human NPCs and that this response was mediated at least in part by activated NF-κB signaling.This could be due to fragmented cellfree (porcine) DNA remnants in NCM, which are known to cause inflammation (Motwani et al., 2019;Poli et al., 2017), e.g. in chronic diseases like osteoarthritis (Nagata et al., 2010)

NCM inhibits pro-inflammatory mediators by affecting the MAPK pathway
After 72 hours, the initial NCM-mediated inflammatory response gradually dissolved.At this timepoint, NCM mainly affected MAPK signaling in canine and human NPCs, by decreasing p-ERK/p-JNK/p-PKC expression via increased DUSP1/5/6 levels.The latter was confirmed in vivo in 6-months-treated canine IVDs.p-ERK1/2 immunopositivity was absent in all in vivo samples, which most probably is related to extensive period of fixation followed by months of decalcification not allowing for appropriate immunohistochemical analysis of phosphorylated proteins (Wolf et al., 2014).This study additionally indicated that the inflammation-related PLA2 and NF-κB cascades (downstream of the MAPK pathway) (Bren et al., 2001;Tian et al., 2018;Wang et al., 2017b) were inhibited by NCM.Since these pathways are known to be involved in production and release of pro-inflammatory mediators in degenerated IVDs (Daniels et al., 2017;Ni et al., 2019), these findings could well explain the antiinflammatory effects of NCM.Furthermore, others also showed that inhibiting MAPK signaling could reduce the catabolic effects of inflammatory mediators in the IVD (Park et al., 2016;Sun et al., 2020).

NCM inhibits also other pathways associated with pro-inflammatory mediators
The current study indicates that NCM also inhibited RAGE signaling, which is in line with our previous study with indications that NCM prevented AGE accumulation in vivo (Bach et al., 2018).AGEs have also been shown to induce osteogenic differentiation and calcification in the IVD (Illien-Jünger et al., 2016;Svenja et al., 2015).Interestingly, together with RAGE signaling, the RUNX2 signaling was also inhibited by NCM, indicating that it might prevent disc hypertrophy/calcification, processes that can occur during advanced stages of IVD degeneration (Rutges et al., 2010).
Angio Interestingly, NC-conditioned medium (NCCM) has previously been shown to exert both anti- (Cornejo et al., 2015) and proangiogenic (de Vries et al., 2018a) effects, whereas NCM did not affect vessel growth (de Vries et al., 2018a).Furthermore, NCCM inhibited neurite growth in one study (Purmessur et al., 2015), whereas another study showed that NCCM and NCM both increased neurite expressing cell numbers (de Vries et al., 2018a).In our previous in vivo study, no indications for increased VEGF expression, nerve or vessel ingrowth were present after six months of NCM treatment (Bach et al., 2018).Nonetheless, future studies that test NCM-based regenerative strategies in vivo should carefully assess the risk of angio-and neurogenesis during a prolonged time period.

NCM affects the expression of proteins involved in cell cycle regulation and proliferation
In previous work, it was established that NCM increased cell proliferation in vitro (Bach et al., 2018;de Vries et al., 2019).DigiWest analysis shed light on the possible mechanisms behind this effect of NCM.The ERK/JNK-MAPK signaling pathway is, besides involved in inflammation, also involved in cell cycle regulation, by promoting and/or inhibiting proliferation and apoptosis, dependent on the specific context (cell type, stimulus, pathologic status) (Cagnol and Chambard, 2010;Guma and Firestein, 2012;Mi et al., 2018).Although expression of key proteins (p-ERK/p-JNK/p-PKC) of the MAPK pathway was inhibited by NCM, p-S6 levels were increased.p-S6 is involved in the regulation of amongst others cell size and proliferation (Ruvinsky et al., 2009;Ruvinsky and Meyuhas, 2006), but it also stimulates cartilage repair (Zhang et al., 2020).P-S6 can, besides by ERK/JNK-MAPK, also be activated by other pathways, e.g.mammalian target of rapamycin (mTOR), which is in turn activated by amongst others certain growth factors (Hay and Sonenberg, 2004).Therefore, the increased p-S6 levels in NCM-treated NPCs, possibly caused by the growth factors present in NCM (Bach et al., 2022), might facilitate cell proliferation possibly facilitating an anabolic response.
Regulation of cJun, which is involved in cell cycle regulation and proliferation, is complex (Wisdom et al., 1999) and depends on the activation period (Guma and Firestein, 2012).Our study showed decreased p-cJun expression in hNPCs, but increased p-cJun levels in cNPCs treated with NCM.This may be explained by cJun being reported to induce cell cycle progression and inhibit apoptosis by distinct mechanisms; cJun-mediated G1 progression is independent of serine 63/73 phosphorylation, whereas inhibition of apoptosis requires serines 63/73 (Wisdom et al., 1999).Thus, in the canine NPCs cJun could have been increased because cell proliferation was stimulated by NCM independent of specific JNK activation.Additionally, the Fibroblast growth factor-4 (FGF4) retrogene insertion is responsible for chondrodystrophy and IVD degeneration in dog breeds like Beagles (Brown et al., 2017).There are indications that FGF4 influences proliferation via c-Jun activation (Kook et al., 2013), and therefore, this retrogene in combination with NCM treatment could have contributed to the increased cJun expression in Beagle NPCs compared with hNPCs.
Furthermore, NCM-treated NPCs demonstrated an increased GAPDH and CDC2 expression, the latter only in cNPCs.GAPDH was initially identified as a glycolytic enzyme and considered as a housekeeping gene, but emerging evidence indicates that GAPDH is involved in diverse functions other than its role in energy metabolism, e.g.cell cycle regulation, proliferation and apoptosis (Carujo et al., 2006;Kosova et al., 2017;Zhang et al., 2015).Interestingly, GAPDH enables the activation of CDC2 (Carujo et al., 2006), which is also a key player in these processes (Dorée and Hunt, 2002;Haneke et al., 2020;Liu et al., 2008).
Altogether, NCM affects cell proliferation and apoptosis via different players.
NCM improves the NPC phenotype in vivo at the long-term follow up NCM treatment improved the cNPC phenotype in vivo after 6 months, indicated by increased expression of well-known phenotypic NP markers KRT19, FOXF1, and PAX1 (Akker et al., 2017;Richardson et al., 2017).In vitro, only an increase in KRT19 protein expression was detected in both canine and human NPCs, whereas other phenotypic markers were downregulated in hNPCs.This discrepancy could be due to the used culture system, since it is known that NPCs don't maintain their specific morphology and phenotype in two-dimensional (monolayer) culture (Wang et al., 2001).Furthermore, the absence of phenotypic markers could also be due to passage number used in the current study (P2 for cNPCs and P3 for hNPCs).
Altogether, the results from our study show that the anabolic effects and inhibition of proinflammatory mediators, detected after intradiscal NCM injection in vivo (Bach et al., 2018), are accompanied by an improved NPC phenotype in addition to upregulation of DUSP protein expression.A limitation of the current study is that it only focused on the NP, whereas the AF and the EPs are also known to be involved in the inflammatory process during IVD degeneration (Lai et al., 2023;Yamagishi et al., 2022).

Clinical perspective and conclusions
Further research should focus on improving the clinical applicability of NCM.Extracellular DNA containing porcine endogenous retroviruses is problematic within the NCM (Wilson, 2008), since they have been shown to infect human cells in vitro and thus pose a risk for patients (Martin et al., 2000).Also, fragmented cell-free DNA may cause inflammation (Motwani et al., 2019;Poli et al., 2017).Prior to utilization in a clinical setting, processing of the NCM to remove DNA is thus required (Schmitz et al., 2022).
In conclusion, NCM initially stimulated proinflammatory mediators in vitro, but thereafter exerts its effects by influencing the MAPK pathway.The latter leads to the reduced expression of inflammatory cytokines and improved NPC phenotype after prolonged treatment (Fig. 11).The work presented enhance our understanding of how NCM instructs the degenerate NPCs and provide for new hypothesis to study in NC-based therapeutic strategies employing cell-free or cell-based approaches.
Editor's note: The Scientific Editor responsible for this paper was Sibylle Grad.
,l).Since NCM induced DUSP5 expression in both hNPCs and cNPCs, DUSP5 immunopositivity was determined in degenerated canine IVDs treated with NCM in vivo (Bach et al., 2018).Confirming the DigiWest results, canine IVDs treated with 2× NCM demonstrated increased DUSP5 immunopositivity compared with control and 1× NCM IVDs, in both mildly (no NX-IVDs) or moderately degenerated discs (NX-IVDs) (p < 0.05, Fig. 3m,n).Fig. 4 depicts tables and volcano plots for top proteins differentially expressed in 72 hours NCM-treated cNPCs (Fig. 4a) and hNPCs (Fig. 4b) versus controls, as determined by DigiWest technology.MAPK signaling proteins are evolutionarily conserved serine/threonine kinases, which regulate signal transduction and cell responses, i.e. cell proliferation, differentiation, survival, death, and inflammation.An overview of all significantly up/downregulated proteins in the MAPK pathway by NCM treatment is given in Fig.5

Fig. 5 .
Fig. 5. Mitogen-activated protein kinase (MAPK) signaling pathway as influenced by NCM treatment in canine and human NPCs from degenerate IVDs.Red and green indicates that these signaling proteins are down-and upregulated, respectively, in NCM-treated NPCs compared with controls.Yellow: upregulated in canine, but downregulated in human NPCs.Proteins illustrated in light blue were not significantly affected by NCM treatment.This schematic picture is based on previous publications MAPK: Mitogen-activated protein kinase, MAP4K1: Mitogen-activated protein kinase kinase kinase kinase 1 (HPK1), MAP3K: Mitogen-activated protein kinase kinase kinase, MAP2K: Mitogen-activated protein kinase kinase, ERK: extracellular signal-regulated kinase, JNK/SAPK: cJun N-terminal kinase/stress-activated protein kinase, PKC: protein kinase C, DUSP: Dual-Specificity Phosphatase, S6: S6 ribosomal protein.

Fig. 7 .
Fig. 7. IL-1β and IL-1R immunohistochemical results of in vivo canine IVDs treated with no (control), one (1× NCM) or two (2× NCM) intradiscal injections of 50 µL of 10 mg/mL NCM.Samples were collected 6 months after the first intradiscal injections.Representative samples of all 3 conditions (in this case for NX-IVDs) are depicted.NX: partial NP removal to induce moderate instead of mild IVD degeneration.n = 5. n.s.: not significantly affected.

Fig. 11 .
Fig. 11.Main conclusion of the current study.NCM exerts its long-term anti-inflammatory effects mainly by increasing cytosolic DUPS6 and nuclear DUSP1 and DUSP5 expression, which together dephosphorylate p-ERK1/2, p-JNK, and p-PKC (all involved in the MAPK signaling pathway) and by directly dephosphorylating p-PKC.Decreased p-ERK and p-JNK expression leads to inhibited inflammatory cytokine production.

Table 2 . Primers used for quantitative PCR of canine and human samples. All
primers were designed in-house using Perlprimer.Bp: base pairs.
NPCs.Relative gene expression (E ΔΔCT ) of canine NPCs for (a) IL-6, (b) IL8, (c) COX2 and human NPCs for (d) IL-6, (e) IL-8, and (f) COX2 after 6, 24, and 72 hours monolayer culture with basal culture medium with/without 10 mg/mL NCM.Log10 scale is used for the y-axis, n=4.(g) n-Fold change (x control) of induction of an NF-κB chondrosarcoma reporter cell line after 6 hours of culture with basal culture medium with/without 10 mg/mL NCM of 3 different porcine donors measured in quadruplicate.

Table 4 . Reactome pathway analysis of canine NPCs treated with NCM.
Head pathways (bold) and sub-pathways (non-bold) are listed together in one row.

Table 5 . Reactome pathway analysis of human NPCs treated with NCM.
Head pathways (bold) and sub-pathways (non-bold) are listed together in one row.The fold enrichment of the proteins in the pathways is represented compared with expected values from the reference list.The pathways in the table indicates over-representation of this category in the analyzed list: more genes are observed than expected on the basis of the reference list (for this category, the number of genes in the list is greater than the expected value)(Mi et al., 2019).Pos: positively influenced pathway, Neg: negatively influenced pathway, Pos/neg: both positively and negatively regulated proteins in this pathway.