Combinatorial conditioning of adipose derived‐mesenchymal stem cells enhances their neurovascular potential: Implications for intervertebral disc degeneration

ABSTRACT Background Mesenchymal stem cells (MSCs) are becoming an increasingly attractive option for regenerative therapies due to their availability, self‐renewal capacity, multilineage potential, and anti‐inflammatory properties. Clinical trials are underway to test the efficacy of stem cell‐based therapies for the repair and regeneration of the degenerate intervertebral disc (IVD), a major cause of back pain. Recently, both bone marrow‐derived MSCs and adipose‐derived stem cells (ASCs) have been assessed for IVD therapy but there is a lack of knowledge surrounding the optimal cell source and the response of transplanted cells to the low oxygen, pro‐inflammatory niche of the degenerate disc. Here, we investigated several neurovascular factors from donor‐matched MSCs and ASCs that may potentiate the survival and persistence of sensory nerve fibers and blood vessels present within painful degenerate discs and their regulation by oxygen tensions and inflammatory cytokines. Methods Donor‐matched ASCs and MSCs were conditioned with either IL‐1β or TNFα under normoxic (21% O2) or hypoxic (5% O2) conditions. Expression and secretion of several potent neurovascular factors were assessed using qRT‐PCR and human magnetic Luminex assay. Results ASCs and MSCs expressed constitutive levels of key neurotrophic factors; and stimulation of ASCs with hypoxia triggered increased secretion of both angiogenic factors (Ang‐2 and VEGF‐A) and neurotrophic (NGF and NT‐3) compared to MSCs. We also report increased transcriptional regulation of pain‐associated neuropeptides in hypoxia stimulated ASCs compared to those in normoxic conditions. We demonstrate transcriptional and translational upregulation of NGF, NT‐3, Ang‐1, and FGF‐2 in response to cytokines in ASCs in 21% and 5% O2. Conclusions This work highlights fundamental differences between the neurovascular secretome of donor‐matched ASCs and MSCs, demonstrating the importance of cell‐selection for tissue specific regeneration to reduce ectopic sensory nerve and blood vessel survival and improve patient outcomes.

We demonstrate transcriptional and translational upregulation of NGF, NT-3, Ang-1, and FGF-2 in response to cytokines in ASCs in 21% and 5% O 2 .
Conclusions: This work highlights fundamental differences between the neurovascular secretome of donor-matched ASCs and MSCs, demonstrating the importance of cell-selection for tissue specific regeneration to reduce ectopic sensory nerve and blood vessel survival and improve patient outcomes. others. There is a growing clinical database and available results support the preliminary conclusion that these strategies can provide an effective regenerative or disease-modifying outcome. A definitive conclusion will become available when late stage clinical trials are completed. Pain is a key feature of many musculoskeletal conditions, including, for example, chronic severe lower back pain associated with intervertebral disc (IVD) disease (IDD) or chronic joint pain in osteoarthritis. The impact on patients is immense and leads to reduced quality of life and sometimes serious addictions to opioid medications.
Low back pain (LBP) is an increasingly common health problem and a major cause of disability worldwide, resulting in its inclusion in the World Health Organization's list of 12 priority diseases. Total annual costs associated with back pain are estimated in excess of €12 billion across Europe. Both the UK and the US report LBP as being the most common cause of disability amongst young adults, leading to a loss of more than 100 and 149 million workdays per year, respectively. 1 It is starkly clear that as the world population ages, the disease burden of LBP will increase substantially. The causes of LBP remain multifactorial, although approximately 40% of all cases are attributed to the degeneration of one or more IVDs. 2 Current treatments include conservative symptomatic pain relief or end-stage surgical treatments that are often ineffective as they are symptom-, rather than disease-modifying. [3][4][5] Thus, regenerative therapies targeting the degenerate IVD have emerged at the forefront of research within the spinal field; with considerable attention focusing on the development of cell-based therapeutics that address the underlying pathogenesis of degenerative disc disease. [6][7][8] Many groups have studied the potential use of MSCs from bone-marrow or adipose (ASCs) to repair and regenerate the damaged IVD, either alone or in combination with native NP cells, sometimes delivered using biomaterial scaffolds and carriers. 9,10 Stem cells are an attractive option for regenerative therapies due to their availability, self-renewal capacity, multilineage potential, and immunosuppressive properties. One particular question is the source of stem cells, and various studies have investigated the differences between MSCs and ASCs for their regenerative potential.
Clinical trials are underway to elucidate a successful stem cellbased therapy for the repair and regeneration of IVD tissue, with preclinical and phase I data showing promising results for the safety and efficacy of allogeneic and autologous ASCs and MSCs. [11][12][13][14][15] Although cell-therapies have been reported to increase hydration and flexibility of the affected disc, reduce pain, and in some instances, increase disc height, there is a subset of patients who do not show improvement in pain score, thus failing to meet primary endpoints. The causes of patient nonresponse are inconclusive, yet it could be suggested that the delivered MSCs respond negatively to the degenerate microenvironment resulting in persistent pain following delivery. 11,13 While several in vivo studies have demonstrated MSC engraftment and long-term survival in the harsh environment of normal disc tissue, [16][17][18][19][20] the capacity of MSCs and ASCs to stimulate the survival and persistence of nerve and blood vessels within painful discs has not been addressed.
This study aimed to test the hypothesis that hypoxic and proinflammatory conditioning mimicking that of the degenerate IVD, would modify the expression and secretory profile of donor-matched ASCs and MSCs by enhancing the expression of potent angiogenic and neurotrophic factors, thus exacerbating pain in patients with innervated discs.
Our objectives were to establish whether there is a difference in expression of angiogenic and neurotrophic factors between undifferentiated ASCs and MSCs and whether the expression of these factors is affected by culture in a hypoxic or inflammatory environment alone, or in combination.
To the authors' knowledge, this is the first study to investigate and report fundamental differences in the neurovascular expression and secretory profile of ASCs and MSCs, identifying an optimal cell source for regenerative therapies of the IVD to reduce the occurrence of neural and vascular growth and improve clinical outcomes. Cells were isolated from bone marrow and adipose tissue as previously described. 21 Briefly, bone marrow aspirates were washed with PBS and then centrifuged to obtain a cell pellet, and the pellet was resuspended in Minimum Essential Medium α-modification (αMEM; Sigma-Aldrich) which included nonessential amino acids, 110 mg/L sodium pyruvate, 1000 mg/L glucose supplemented with final concentrations of 100 U/mL penicillin, 100 μg/mL streptomycin and 0.25 μg/mL amphotericin, 2 mM GlutaMAX (Life Technologies), and 20% (vol/vol) fetal bovine serum (FBS). MSCs were isolated using density gradient centrifugation as previously described. 21,22 Adipose tissue was finely minced with a scalpel and then ASCs were isolated enzymatically as previously described. 23 Both cell types were maintained at 37 C in a humidified atmosphere containing 5% (vol/vol) CO 2 ; the culture medium was changed after 5 days and cells cultured to 80% confluence in complete αMEM as above supplemented with 10% FBS. MSC and ASC phenotype was confirmed as previously described (data not shown). 21,23,24 2.2 | Cytokine regulation of angiogenic factors, neurotrophic factors, and pain peptides in donormatched adipose and bone-marrow-derived MSCs Following expansion to passage three, ASCs and MSCs were plated into wells of a 12-well culture plate at a density of 50 000 cells per well. The following day cells were washed and exposed to low serum medium (complete medium as above supplemented with 2% FCS) either alone or with the addition of IL-1β (10 ng/mL) or TNFα (10 ng/mL) for 48 hours in either 21% or 5% O 2 . All conditions were performed in technical triplicate. Following 48 hours conditioned medium was collected and stored at −80 C until required for further analysis.

| Metabolic activity of MSCs under different oxygen conditions and cytokine stimulations
The metabolic cellular activity of ASCs and MSCs cultured at either 21% or 5% O 2 was assessed using Alamar Blue assay (Thermo) in low serum media (2% FBS) at day 0 and day 2 as per the manufacturers protocol. The absorbance was recorded using a fluorescence microplate reader (TECAN) fluorescence excitation and emission wavelengths of 590/540 nm. Fluorescence was recorded for cellular constructs and normalized to the fluorescence of day 0 samples as an indication of cytotoxicity/proliferation.

| Quantitative real-time PCR
After 48 hours in normoxic or hypoxic culture with or without cytokine stimulation RNA was extracted using PureLink RNeasy Spin Columns (Invitrogen) as per manufacturer's instructions and quantification performed using the NanoDrop (Thermo). cDNA was synthesized using the High-Capacity cDNA Reverse Transcription Kit Reactions were prepared in duplicate using Fast SYBR Green Master Mix (Applied Biosystems, UK) and predesigned KiCqStart Primers (Sigma-Aldrich; Table 1) to a total volume of 10 μL, containing 10 ng cDNA and 600 nM each primer. Data were analyzed according to the 2 −ΔCt method, with expression normalized to 18S reference gene.

| Magnetic Luminex Assay analysis of conditioned medium of cytokine stimulated donormatched MSCs under normoxic and hypoxic conditions
Levels of Ang-1, Ang-2, FGF-2, and VEGF-A, NGF, BDNF, and NT-3, were measured in cell culture supernatant from three patients in technical duplicates using quantitative Luminex assays (R&D Systems) according the manufacturer's instructions, and analyzed using the Biorad Bioplex 200 multiplex ELISA system.

| Enzyme-linked immunosorbent assay
Substance P levels were measured in cell culture supernatant from three patients in technical duplicate using an ELISA for detection of human substance P (Abcam) as per manufacturer's instructions and the absorbance read at 450 nm and 570 nm using a Victor X3 plate reader (PerkinElmer).

| Statistical analysis
Data were assessed for normality using the Shapiro-Wilks test and found to be nonparametric. Statistical significance determined using two-way ANOVA with Tukey's multiple comparisons test, * = P ≤ .05.  Forward primer (5 0 à3 0 ) Reverse Primer (5 0 à3 0 ) Accession number

| ETHICAL APPROVAL AND CONSENT TO PARTICIPATE
F I G U R E 1 Metabolic activity of ASCs and MSCs exposed to cytokines IL-1β or TNFα under 21% or 5% oxygen.  TNFα stimulation with an increase in NGF, but a smaller increase was observed following IL-1β stimulation, yet this change failed to reach significance ( Figure 3A). At protein level, NGF secretion into the culture media was significantly increased following TNFα stimulation in both 21% and 5% O 2 from ASCs; however, levels were significantly higher from ASCs under 5% O 2 compared to 21% O 2 ( Figure 3B).
BDNF gene expression levels were significantly higher in ASCs under 5% O 2 compared to ASCs in 21% O 2 ( Figure 3C). Gene expression was significantly downregulated by IL-1β and TNFα stimulation in 5% O 2 compared to the corresponding unstimulated control ( Figure 3C). BDNF expression from MSCs was not altered by cytokine stimulation or O 2 levels at gene level ( Figure 3C), yet secreted levels of BDNF were significantly upregulated by IL-1β and TNFα in MSCs at both 21% and 5% O 2 compared to unstimulated controls ( Figure 3D).
BDNF secretion from ASCs was not altered by cytokine stimulation, yet unstimulated ASCs under 5% O 2 secreted significantly higher levels of BDNF than unstimulated ASCs under 21% O 2 ( Figure 3D).
Gene expression of NT-3 was not altered by cytokine stimulation in either cell type ( Figure 3E), yet levels of secreted NT-3 were significantly increased upon IL-1β stimulation in ASCs at 21% and 5% O 2 ( Figure 3F).

| Neuropeptides were transcriptionally regulated by low oxygen in ASCs and MSCs
Here, we aimed to investigate the potential of MSCs to produce pain related peptides in response to low oxygen and inflammatory ing heightened tube formation in vitro compared to MSCs. 31 Other studies largely conclude that the therapeutic potential of stem cells is attributed to their differentiation into vascular endothelial cells but also to their ability to produce various angiogenic factors, including VEGF and FGF-2. [31][32][33] In the current study, we report increased transcriptional and translational expression of pro-angiogenic VEGF-A in response to reduced oxygen in ASCs and MSCs with higher levels evident in ASCs. We also report increased secretion of VEGF-A from ASCs stimulated with IL-1β under 5% oxygen. This study demonstrated significantly higher levels of other pro-angiogenic factors including Ang-1, Ang-2, and FGF-2 at gene level from ASCs when exposed to 5% oxygen compared to 21% oxygen. The presence of receptors. 37 Together with the data presented in this study, it could be further hypothesized that ASCs may alter their receptor profiles under different oxygen tensions and therefore become more responsive to cytokines, eliciting increased neurotrophic capabilities compared to MSCs.
Neuropeptide substance P is recognized as a pleiotropic molecule able act on sensory nerve fibers as a classical nociceptor, but more recently reported for its ability to enhance the stem cell phenotype and homing of MSCs to the site of injury. 38 Neuropeptides are produced by degenerate NP cells [39][40][41][42][43] and their regulation by proinflammatory cytokines IL-1β and TNFα is well documented. 40,41 Here, we report significant transcriptional regulation of substance P by low oxygen in ASCs yet this was not evident in the conditioned medium. Induction of substance P gene expression and release has been documented previously in transdifferentiated neural MSCs stimulated with IL-1α, 44 suggesting the induction of a neural phenotype is required for the production and release of substance P into the environment.
The data presented here highlights increased responsiveness of ASCs to cytokine stimulation under low oxygen by enhancing secretion of factors involved in neurogenesis and neo-angiogenesis within the IVD; a process that could lead to persistent discogenic pain. Although we report that naïve ASCs have a greater potential to increase their expression and secretion of factors involved in nerve survival and vascularization under hypoxic and inflammatory conditions, the minimally invasive availability of these stem cells is still an attractive source for therapy. ASCs retrieved from adipose tissue following liposuction plastic surgeries have been shown to be up to 500 times more prevalent than MSCs when comparing an equivalent volume of tissue. 45 ASCs may benefit from being predifferentiated or preconditioned prior to delivery to reduce their responsiveness to low oxygen levels and the inflammatory milieu ( Figure 5). Clarke et al were the first to report the use of growth differentiation factor-6 (GDF-6) as a differentiation factor promoting NP-like differentiation of ASCs in replicating matrix synthesis and gene expression of classical NP markers. 23 It could therefore be hypothesized that GDF-6 pre-differentiated ASCs could enhance NP cell repair as well as attenuating the inflammatory and neurovascular environment as recently reported by Miyazacki et al. 46 While this study utilized monolayer culture of human MSCs, it forms a basis for future studies into the mechanistic actions of MSCs when exposed to inflammatory and oxygen depleted environments such as the IVD; and future studies should include 3D cultures of MSCs and NP cells to assess the interaction between different cell types following delivery and their effect on preexisting neural and vascular structures. The IVD is an extremely complex tissue undergoing progressive changes and so it is important to consider all aspects of the target tissue to ensure correct cell types and delivery systems are in place to ensure optimal success rates, including but not limited to a soft, hydrated 3D environment subjected to mechanical load.
Therefore, further research should be invested into elucidating the mechanisms by which MSC-based therapies aid regeneration of the IVD tissue, and avoid the promotion of ectopic structures that may hinder clinical outcomes.

| CONCLUSION
To conclude, this is the first report to highlight fundamental differences in neurovascular profiles of donor-matched ASCs and MSCs exposed to inflammatory conditioning under low oxygen tensions in order to provide insights into the importance of stem cell selection for regenerative therapies. This includes other tissues, which have a hypoxic and inflammatory environment in response to injury or disease and which conversely to IVD require angiogenesis and innervation for healing, for example, bone lesions. When considering therapies for regeneration of the IVD, selecting the optimal cell source to repair the tissue as well as preventing neo-angiogenesis and ectopic sensory nerve fiber ingrowth is crucial to reducing discogenic pain and improving patient outcomes. These observations have allowed us to propose a neurovascular response network associated with cell delivery to the degenerating disc. It appears that NGF, NT-3, VEGF-A, and Ang-1 are key players in this network, and future work should investigate the effect of the ASC/MSC secretome on native NP cells to ensure clinical efficacy of cell-based regenerative therapies.

ACKNOWLEDGMENTS
The authors would like to thank Pauline Baird for her technical assistance at The University of Manchester. We would like to thank Dr F I G U R E 5 The role of inflammatory cytokines on ASCs under low oxygen environment and their implications in regulating neurovascular persistence and survival within the degenerate IVD.
The highly inflammatory nature of the degenerate IVD will expose delivered stem cells to multiple pro-inflammatory cytokines. TNFα and IL-1β increased expression of neurotrophic and angiogenic factor expression and secretion from ASCs under low oxygen compared to MSCs. The production of these neurovascular factors may lead to the sensitization of ectopic sensory nerve fibers within the degenerate IVD resulting in sustained pain in these patients. Therefore, we hypothesize that pre-stimulation or pretreatment of ASCs with GDF-6 may induce an NP-like phenotype and reduce their responsiveness to the degenerate microenvironment upon delivery