Unravelling the role of Mesenchymal Stem/Stromal Cells Secretome in intervertebral disc degeneration in a pro-inflammatory/degenerative ex vivo model CURRENT STATUS: POSTED

Background: Mesenchymal stem/stromal cells (MSCs) have been increasingly used in clinical trials for intervertebral disc (IVD) degeneration. Here, we aimed to evaluate the potential of a cell-free approach to degenerated IVD, testing if MSCs secretome can stimulate a regenerative response by modulating the IVD inflammatory cascade. Methods: Human bone marrow-derived MSCs were preconditioned with IL-1β (10 ng/mL) and low oxygen (6% O2). The secretome of MSCs (MSCsec) was collected after 48h. Bovine IVD tissue explants cultured in pro-inflammatory/degenerative conditions (needle puncture + IL-1β) were treated with MSCsec or co-cultured with MSCs. Results: MSCsec obtained upon IL-1β-pre-conditioning, as well as MSCs co-culture, down-regulated gene expression of pro-inflammatory cytokines, bIL-6 and bIL-8 after 48h, in IVD. IVD matrix degrading enzymes, bMMP1 and bMMP3, were downregulated and upregulated, respectively, in the presence of MSCsec, but not MSCs. After 14 days, MSCsec-treated IVDs revealed increased aggrecan content at the protein level, contrarily to MSCs/IVD co-cultures. Interestingly, IL-1β-preconditioning only, but not IL-1β-IVD, increased gene expression of hADAMTS5 and hTIMP-1in MSCs. Additionally, conditioned medium from MSCsec-treated IVDs did not promote angiogenesis or neurogenesis. In MSCsec-treated IVD, an increase in MCP-3 and GCP-2 was observed, while SDF-1α, TNF-α, IGF-1, Eotaxin 3, FGF-9, MIP-1δ, IFN-γ, IL-5, TNF-β, IL-4, TGF-β1, IL-16, IGFBP-3 and IGFBP-4 were decreased, compared with MSCs/IVD co-cultures. Conclusions: MSCsec obtained upon IL1β-preconditioning, present an immunomodulatory role in degenerated IVD, as well as MSCs. Nevertheless, MSCsec but not MSCs, potentiate aggrecan deposition in IVD in pro-inflammatory/degenerative conditions. This finding can open new perspective on the use of MSCsec as a cell-based/cell-free approach to LBP.


Pro-inflammatory/degenerative Bovine IVD Organ Culture with MSCsec
Bovine IVDs were isolated from tails of young animals (<1 year old) up to 3 h post-slaughter and cultured, as previously described [17]. At the fifth day of culture, IVDs were needle-punctured with a 21G needle and the culture medium was supplemented with 10 ng/mL of recombinant human IL-1β (PeproTech). Three hours later, the culture medium was exchanged for MSCsec further supplemented with 10 ng/mL of IL-1β. For short-term IVDs organotypic culture (48h), the supernatants of MSCs, IVDs and co-culture were collected. For long-term IVDs organotypic culture (14 days), IVDs were exposed to MSCsec following the same aforementioned protocol with the exception that the MSCsec used was obtained after the preconditioning of 2x10 5 MSCs and culture medium was exchanged (25%) every 4-5 days. At day 14, IVDs were collected for histological analysis. Non-stimulated IVDs, IVDs punctured and stimulated with 10 ng/mL of IL-1β and IVDs with non-preconditioned MSCs seeded on the membrane filter inserts were used as controls in all experiments. All the IVDs cultures were maintained in culture at 37 o C, 6% O 2 and 8,5% CO 2 . The complete experimental setup is summarized in Figure 1. The protocols used for each assay are extensively described in the Supplementary Materials and Methods section.

IVD and MSCs analysis
After 48h, IVD and MSCs total RNA was extracted and reverse transcribed into cDNA for gene expression analysis of bovine IL-6, IL-8, MMP1, MMP3, collagen type II (COL2), aggrecan (ACAN), ADAMTS5 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH), and human MMP2, MMP3, MMP9, ADAMTS4, ADAMTS5, TIMP1 and GAPDH (Table S.1). After 14 days of culture, IVDs were fixed and embedded in paraffin for ECM analysis, as previously described [16]. Collagen type II distribution was analysed by immunofluorescence (IF). Aggrecan and Collagen type I distributions were analysed by immunohistochemistry (IHC), using the Novolink TM Polymer Detection Kit (Leica Biosystems) according to the manufacturer's instructions. Sulfated glycosaminoglycan (sGAG) content of the IVDs was also assessed, by the 1,9-dimethyl-methylene blue zinc chloride double salt (DMMB) assay and normalized by DNA content.

Analysis of MSCsec and IVD conditioned media
To assess the angiogenic and neurogenic potential of the conditioned media from IVD cultures, a cell sprouting assay using HUVECs was conducted [22], as well as an axonal growth assay using murine dorsal root ganglion (DRG) culture, as previously described [23]. In addition, the cytokine content of MSCsec and IVD conditioned media was analysed using the Human Cytokine Array C5 (RayBiotech), following the manufacturer's protocol. For that, a pool of MSCsec from 4 different donors and a pool of conditioned media collected from 4 distinct IVDs cultures stimulated with: i) needle-puncture + IL-1β (IL-1β), ii) needle-puncture + IL-1β + MSCs (IL-1β + MSCs) or iii)needle-puncture + IL-1β + MSCsec (IL-1β + MSCsec) were mixed and analysed in the membrane array. Chemiluminescent signal was detected in a Chemidoc XRSþ (Bio-Rad) with an exposure time of 5.0 sec. Membrane array images were analysed using ImageLab 6.0 software (Bio-Rad). A heatmap representative of the relative expression of all the molecules in the array was produced using Morpheus software (Figure 7).

Statistical Analysis
Results are presented in box and whiskers plots, as median interquartile range (IQR). Since the data from this work did not pass the normality test (D'Agostine and Pearson normality test), nonparametric tests were applied. When more than two groups were compared either unpaired Kruskal-Wallis test or paired Wilcoxon test, with Dunn multiple comparison test as post hoc, were used. When only two groups were compared, statistical analysis was performed using nonparametric test Mann-Whitney. A confidence level of at least 95% (p < 0.05) was set. Graph Pad Prism v6.01 software was used for the analysis.

Results
MSCs secretome upon cells pre-conditioning in a hypoxic, pro-inflammatory microenvironment, has an immunomodulatory effect on the degenerating IVD MSCs were previously shown to have an immunomodulatory effect in the context of the degenerative IVD [16], by experiments performed in cells' migration-enabling conditions (8-μm pore size cell culture insert on top of the IVD) [16]. So, to further confirm the MSCs paracrine mode of action, we repeated the experiments with MSCs seeded on migration-disabling transwells (1-μm pore size), and a similar immunomodulatory effect was observed (supplementary data, Figure S1). Therefore, in the present study, we evaluated whether the secretome produced by MSCs, pre-conditioned with microenvironmental cues from the pro-inflammatory/degenerative stimuli of the IVD organ culture previously established by us [17], would recapitulate the immunomodulatory role observed with the MSCs viability, metabolic activity, phenotype and inflammatory cytokine content were evaluated in the presence of different oxygen levels (6% and 21%) and the presence/absence of the proinflammatory cytokine IL-1β, after 48 h of exposure ( Figure 2A). MSCs remained highly viable and metabolic active in all the conditions tested ( Figure 2A) and their phenotype (cell surface markers CD105, CD90 and CD73 defined by the International Society for Cellular Therapy) remained unaltered ( Figure 2B). Nevertheless, IL-1β but not hypoxia (6% O 2 ) significantly (*, p<0.05) promoted the secretion of pro-inflammatory cytokines (IL-6, IL-8, RANTES, PGE2, MCP-1) by MSCs ( Figure 2C). In an attempt to use the preconditioning protocol that more closely mimics the microenvironment found by the MSCs in the degenerative IVD, from this point onwards, only secretome produced under hypoxia and IL-1β (H + IL-1β) was used and will be, therefore, referred to as "MSCsec".
Using MSCsec to treat IVD organ cultures in pro-inflammatory/degenerative conditions, we observed that the gene expression levels of the pro-inflammatory markers, bIL-6 and bIL-8, upregulated in the degenerative/pro-inflammatory IVD model compared to non-stimulated discs, were significantly downregulated (*, p<0.05) in the presence of either MSCs or MSCsec, reaching levels similar to the control IVDs ( Figure 2D). This reveals that the immunomodulatory role of MSCs in the degenerated IVD is independent of their presence.

MSCs secretome impacts on IVD matrix remodelling in a distinct way from MSCs
Considering the modulation of the inflammatory cascade in the IVD by MSCsec, we further explored their role on IVD ECM remodelling. First, gene expression of most abundant ECM proteins and proteases expressed by IVD cells was analysed, 48 h post-treatment ( Figure 3A). Results showed a downregulation of gene expression levels of the IVD matrix components bCOL2 and bACAN in the degenerative IVD model, comparatively to the non-stimulated discs, and an increase in bMMP1 and bMMP3 as previously reported [17]. No significant differences were observed in bADMATS5 (bADAMTS4 was also analysed but was expressed at very low levels, data not shown). Once more, the presence of MSCs significantly decreased bCOL2 and bACAN (*, p<0.05), as previously described [16].
MSCsec had a similar effect, but decreased significantly only bACAN (*, p<0.05). Concerning bMMPs gene expression, MSCsec significantly down-regulated bMMP1 (-5,04 fold-change, *, p<0,05) and upregulated bMMP3 (1,78 fold-increase,*, p<0.05) relative to the pro-inflammatory/degenerative IVD model, which was not observed in the presence of MSCs. To confirm these observations at the protein level, the IVDs explants were analysed 14 days post-treatment ( Figure 4B, 4C). All the proinflammatory/degenerative IVD organ cultures presented loss of sGAG, when compared with the control, an effect that could not be reversed by the presence of MSC or secMSC ( Figure 3B). At tissue level, Collagen type I and II did not present significant changes in the different conditions analysed, but aggrecan levels were decreased on the pro-inflammatory/degenerative IVD compared to the control tissue, independently of the presence of MSCs ( Figure 3C). Surprisingly, in the IVDs treated with MSCsec, aggrecan deposition significantly increased (* p<0.05), 3.6-fold increase compared with the degenerated IVD ( Figure 3C).
To better understand these results, MSCs pre-conditioned with IL-1β+6%O 2 or in co-culture with the IVD were analysed regarding the gene expression levels of enzymes relevant for the matrix degradation: MMP2/3/9, ADAMTS4/5 and TIMP1 ( Figure 4). The results showed no significant differences between levels of expression of human MMPs and ADAMTS4 in the different MSCs cultures, even though preconditioned MSCs seemed to express slightly lower levels of ADAMTS4 (decrease of 0.3-fold, p=0.2063). However, MSCs preconditioned with IL-1β expressed higher levels of ADAMTS5 (2.5-fold increase; p=0.0556) and TIMP-1 (1.7-fold increase, *p=0.0159), compared with MSCs+IVD+IL-1β suggesting that these are key proteases on MSCsec potential to promote aggrecan deposition on IVD.

Angiogenic and neurogenic potential of the pro-inflammatory/degenerative IVDs treated with MSCsec
In this study we evaluated the angiogenic and neurogenic potential of degenerated IVDs treated with MSCsec using pre-established in vitro models. For angiogenesis, we used an in vitro culture of HUVECs in a 3D matrix environment (Matrigel), which enables the formation of tubular structures by these cells in response to soluble factors in the culture medium [24]. The vascular networks formed in the presence of the different stimuli are represented in Figure 5A. Both IVD basal medium with IL-1β and MSCsec seemed to inhibit angiogenesis, even though results were not statistically significant. Neither IVD conditioned medium, with or without the inflammatory stimuli, or treated with MSCs or MSCsec, promoted angiogenesis, suggesting that MSCsec could be a safe approach concerning angiogenesis.
Regarding neurogenesis, a DRG explant culture was conducted, as previously established [25]. The results showed that the IVD medium+IL-1β slightly promoted axonal growth, meaning it is neurogenic, when compared to the neurogenic NGF-supplemented medium (0,45 fold-increase) ( Figure 5B). Nevertheless, IVDs conditioned medium, either in basal or pro-inflammatory conditions, did not stimulate axonal outgrowth. Similarly, MSCsec did not demonstrate a higher neurogenic potential than the NGF-containing neurobasal control (0,13 fold change). Once more, the results obtained suggest that MSCsec could be a safe approach concerning neurogenesis.

Prediction of mechanisms of action of MSCsec in pro-inflammatory/degenerative IVD organ culture
To unravel the main mediators in the crosstalk between MSCs and the IVD cells, the conditioned medium from the pro-inflammatory/degenerative IVDs organ cultures after treatment with MSCsec was screened for its protein content using an antibody membrane array, and compared with those from IVDs+MSCs and from MSCsec by itself. The molecules most differently expressed (up-regulation >5 and down-regulation <-5) ( Figure 6B), were analysed with the Functional Annotation Clustering Tool from the DAVID Database (Table S2). By analysis of the molecules upregulated (>5 fold change) in the MSCs co-culture with the DAVID Functional Clustering Tool, a significant cluster was found that included terms implicated in angiogenesis, neutrophil chemotaxis and the NOD-like receptor signalling pathway. In contrast, regarding the proteins upregulated in the presence of the MSCsec, a cluster was found with all the terms identified with MSCs, except for angiogenesis. Instead, MSCsec exclusively affect proteins in the TNF signalling pathway. Relatively to the downregulated molecules, a significant cluster was only detected with MSCsec, and is associated with inflammation, in particular with the T cell receptor signalling pathway.
Overall, treatment with MSCsec led to a more widespread downregulation of cytokines compared with co-culture with the MSCs (Figure 6C). Only two molecules were up-regulated in result of the treatment with MSCsec, but not MSCs, the MCP-3 and GCP-2, while fourteen molecules were down-regulated MSCs preconditioning with IL-1β and 6% O2 did not seem to significantly alter cellular behaviour and secretory profile at this level. Considering that the basal medium used in every condition is baseline stimulatory but none of the other conditions mimics this tendency, it could also be hypothesized that this model is somewhat inhibitory to the DRGs axonal outgrowth. In this context, it is reasonable to assume that other factors inherent to the microenvironment can be associated with the increasing axonal outgrowth, such as the higher osmolarity of the medium or slight changes in pH due to incubation at low oxygen pressure.
In order to understand the mechanisms linking the various processes in IVD degeneration and its treatment outcome with MSCsec, we analysed the molecular content of the IVDs' culture conditioned medium. The membrane-based array analysis allowed us to evaluate the dynamics of the degenerative disc microenvironment in the presence of MSCs or their secretome, although we were also acknowledge FCT -Fundação para a Ciência e a Tecnologia, in the framework of the PhD grant of JRF (PD/BI/128357/2017).     Schematic representation of factors altered in response to degenerated IVD co-culture with

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.