POTENTIAL OF NOTOCHORDAL CELLS WITHIN INJECTABLE BIOMATERIALS TO PROMOTE INTERVERTEBRAL DISC REGENERATION

.


Lower Back Pain and Intervertebral Disc Degeneration
Low back pain (LBP) is the biggest cause of disability worldwide; around 80 % of adults will suffer from LBP in their lifetime (Hartvigsen et al., 2018; Traeger et al., 2019).Most people experience mild pain and recover quickly, however, in some cases LBP lasting longer than 6 weeks progresses to chronic LBP (CLBP) which can contribute to lifelong disability and societal burden (Hartvigsen et al., 2018; Maetzel & Li, 2002).Current treatments to combat LBP may be pharmacological or nonpharmacological, such as: non-steroidal anti-inflammatory drugs, opioids, anti-depressants, exercise, massage, and manipulation (NICE, 2020; Qaseem et al., 2017).Whilst CLBP can be tackled with surgery in some situations, but this is invasive, expensive, and only targets the end-stage of disease in the spine having limited effectivity to manage CLBP in many patients (Bogduk, 2004; Foster et al., 2018; Phillips et al., 2003; Steffens et al., 2016).In around 40 % of cases, the underlying cause of CLBP is associated with degeneration of the intervertebral disc (IVD) (Luoma et al., 2000; Sakai & Andersson, 2015).Importantly, none of the current treatments target the regeneration of the IVD.

Intervertebral Disc and Notochordal Cells
The IVD permits range of motion and supports biomechanical forces applied to the spine (Risbud & Shapiro, European Cells and Materials Vol.47 2024 (pages 30-50) DOI: 10.22203/eCM.v047a03 2011).The core of the disc contains an aggrecan-rich gellike tissue called the nucleus pulposus (NP), which is enclosed circumferentially by the ligamentous annulus fibrosus (AF).The composition of the NP changes as a human IVD matures.The NP of a neonate is populated by notochordal (NC) cells, which are large, vacuolated, and morphologically distinct cells.In humans, some canine breeds, and other species, such as ovine and caprine, large vacuolated NCs (morphotypic NCs) gradually disappear during disc maturation, substituted by more smaller NP cells as the primary cell type (Bach et al., 2022; Sheyn et al., 2019).However, certain species retain morphotypic NCs within the NP region throughout most of their lifespan, including porcine, leporine, and murine species (Alini et al., 2008; Sheyn et al., 2019).Interestingly, an association between morphotypic NC loss and the onset of disc degeneration has been highlighted, where natural degeneration is only seen in species which lose their morphotypic NCs prior to adulthood (Bergknut et al., 2012; McCann & Séguin, 2016; Jill P.G.Urban & Roberts, 2003).As a result, NCs have become a trending research topic regarding their promise for therapeutic application to mediate disc regeneration (Bach et al., 2022; Humphreys et al., 2018).

Regenerating the Intervertebral Disc with Cell-biomaterial Treatments
Of the cell types which have been considered for the purposes of regenerating a degenerative IVD, less than 3 % of the papers utilised NCs as the choice of regenerative cell source in a recent review (Williams et al., 2021).The review highlights that NCs have been successfully extracted from several species, however, these studies have mainly focused on 3D in vitro culture, with a distinct lack of progression into in vivo studies (Williams et al., 2021).The limited use in studies so far may have been due to the difficulties in NC handling, such as: the inability to maintain their phenotype in monolayer culture (Gantenbein et al., 2014; Potier & Ito, 2014; Spillekom et al., 2014), harvesting insufficient numbers due to limited amplification and inability to passage whilst maintaining phenotype (Kim et al., 2009; Potier et al., 2014; Spillekom et al., 2014; Williams et al., 2023).Nonetheless, NCs could provide an excellent cell source for regenerating the degenerate IVD, due to their capabilities of synthesising extracellular matrix, being highly viable in the conditions within the IVD and producing anti-angiogenic, and anti-catabolic effects (Gantenbein et al., 2014; Potier et al., 2014; Purmessur et al., 2013; Spillekom et al., 2014), for a recent review on the potential of NCs please see Bach et al. (2022).
The limitation of NCs being maintained only in 3D culture models can be addressed by the seeding of cells directly into a biomaterial scaffold, facilitating cell growth and/or favourable differentiation, to provide mechanical support and aid the delivery of the cells into the disc (Peroglio et al., 2012; Thorpe et al., 2016).A proposed bio-material that enables liquid injection followed by in situ gelation, with additional attributes of biocompatibility, mechanical stability, compliance with regulatory pathway approvals and facile translation from research to future medical application, would be an ideal candidate for a therapeutic treatment.
Hydrogel-based biomaterial carriers that have previously been developed for NP regeneration, were investigated in this study to support NCs.The first biomaterial selected was a polyethylene glycol-crosslinked serum albumin/hyaluronan hydrogel, referred to as Albugel, which has been previously shown to support the survival of disc cells and disc healing and has been shown safe to use in human clinical trials regarding cartilage therapy in knee joints (Benz et al., 2010, 2012; Niemeyer et al., 2022).The second selected biomaterial was a synthetic Laponite® crosslinked poly N-isopropylacrylamide-co-N, N-dimethylacrylamide (NPgel) biomaterial (Boyes et al., 2021) which has been previously reported to induce human bone marrow progenitor cell differentiation into an NP-like phenotype with matrix deposition that mimics the native NP tissue (Thorpe et al., 2016; Thorpe et al., 2017; Vickers et al., 2019), and can restore degenerate discs in a goat organ culture model (Snuggs et al., 2023).The importance of cells to retain their NP phenotype will facilitate the regenerative potential of the cells, such as NP cell production of extracellular matrix.This hydrogel system has also undergone in vivo safety studies in rats, where it was demonstrated to not show any adverse events within a subcutaneous implantation model (Thorpe et al., 2018).The third and final biomaterial was selected to investigate whether the inclusion of the native extracellular matrix from the NCrich NP tissue improved phenotypic maintenance.Notochordal cell matrix (NCM)-based materials have been previously shown to promote anabolism within NP cells and degenerated IVD tissue (De Vries et al., 2019; Schmitz et al., 2022).Thus, here the hypothesis that the supplementation of the synthetic NPgel with the NCM would provide additional cues to the NCs supporting their phenotype (Bach et al., 2018; Bai et al., 2017; Cornejo et al., 2015; Vickers et al., 2019), specifically a decellularised NCM (dNCM) was utilised creating a bioactive functional biomaterial (NPgel/dNCM).
Altogether, this study investigated the use of porcine NCs (pNCs) in combination with three potential injectable biomaterials: Albugel, NPgel and NPgel/dNCM that were intentionally representative of either synthetic (NPgel), semi-synthetic (NPgel/dNCM) or primarily based on natural components (Albugel), as a first translation step of NCbased therapeutic strategies.

Experimental Design
To ensure biomaterials to be tested further with cell studies could withstand the shear stresses within the native www.ecmjournal.orgEuropean Cells and Materials Vol.47 2024 (pages 30-50) DOI: 10.22203/eCM.v047a03disc, initial preliminary rheological characterisation was performed, biomaterials which failed to withstand these forces would be excluded from further study.Optimisation experiments were then performed by seeding pNCs into a readily available biomaterial, which has been extensively characterised (NPgel) (Boyes et al., 2021; Thorpe et al., 2016; Thorpe et al., 2017; Vickers et al., 2019).These initial optimisation studies were used to determine optimal seeding density, and viability.Thereafter, pNCs were harvested from three independent biological porcine donors, the discs from each donor were pooled together prior to seeding into three different injectable biomaterials in at least duplicate per donor, total replicates: Albugel (n = 8), NPgel (n = 8) and NPgel/dNCM (n = 6), with all output measures performed on every sample.Cellbiomaterial constructs were cultured for 4 weeks in IVD conditions in relation to physioxia and low glucose conditions.Following culture viability was assessed using Calcein/Hoechst staining.Morphology and porous structure of the gel networks were investigated using scanning electron microscopy (SEM) and histology.Finally, phenotype and matrix production was assessed using immunohistochemistry and histological staining.

NPgel with dNCM Preparation
Decellularized NCM powder (dNCM) was prepared as previously described (Schmitz et al., 2022).Briefly, NC-rich NP tissue was harvested from 12-week-old porcine donors and processed by firstly freeze drying in a lyophilizer (Labconco, Kansas City, MO, USA) overnight at -80 °C, then for > 7 hours at ≤ -50 °C.The sample was then decellularized using a 200 U/mL benzonase (E1014; Merck KGaA, Darmstadt, Germany) in 50 mM Tris-HCl buffer (93313; Sigma, Zwijndrecht, The Netherlands), pH 7.5, 2.5 mM MgCl 2 (M8266; Sigma, Zwijndrecht, The Netherlands) at 0.01 mL buffer/mg dry weight tissue for 48 hours at 37 °C on a roller at 2 rpm; followed by pulverization to obtain a powder that can be kept a -80 °C.The dNCM was then sterilized using UV light and 0.5 % w/v was stirred to be dispersed within NPgel just prior to cell seeding, a single batch of dNCM was utilised for these studies.

Rheological Characterisation
The rheological behaviour of the biomaterials at different degrees of deformation was characterised using an Anton Paar 301 rheometer (Anton Paar, Graz, Austria) in parallel plates configuration.Circular samples were prepared using bespoke Polytetrafluoroethylene (PTFE) moulds (20 mm in diameter, 2 mm in height) and immersed in low glucose Dulbecco's modified eagles media (DMEM, 10567014; Gibco™ Life Technologies, Paisley, UK), with pyruvate pH adjusted to 6.8 to mimic the degenerate disc environment for up to 24 hours at 37 °C.To prevent sample slippage during measurement, sandpaper was attached to the measuring plates.A constant load of 1 N was applied on the sample throughout the measurement to maintain good contact between the sample and the top plate.Amplitude sweep measurements were conducted on the hydrogel samples between 0.1-50 % strain at 1 Hz frequency after 2 and 24 hours of immersion in media.A minimum of three replicates were conducted for each sample and the average moduli values are reported here.For samples that suffered from shrinkage, the resultant moduli values were corrected by the factor ( d d sample ) 4 where d is the diameter of the upper plate and d sample is the diameter of the measuring sample (Bron et al., 2009).

Notochordal Cell Extraction
pNC were obtained from lumbar IVDs harvested from young pigs (< 15 weeks old) (Medical Meat supplier, Rochdale, UK and Marr Grange Butchers, Doncaster, UK).Following the protocol from Williams et al. (2023), the spines were dissected, and the NP harvested under sterile conditions, the NP tissue from the same porcine donor were pooled and digested following a sequential treatment in 7 U/mL pronase (PRON-RO; Sigma, Dorset, UK) in α-MEM modified with ribonucleosides, deoxyribonucleosides, phe-European Cells and Materials Vol.47 2024 (pages 30-50) DOI: 10.22203/eCM.v047a03nol red, L-glutamine (12571063; Gibco™ Life Technologies, Paisley, UK) with 1 % v/v Penicillin-Streptomycin (P/S, 15140122; Gibco™ Life Technologies, Paisley, UK) for 30 minutes at 37 °C.Followed by a 16 hour digestion in 125 U/mL collagenase type II (17101015; Gibco™ Life Technologies, Paisley, UK) in α-MEM at 37 °C.The clusters of pNC were captured in a 40 µm cell strainer and further cultured as described below.
To assess initial cell viability, constructs were stained with 10 µM Calcein-AM (C1430; Invitrogen™, Carlsbad, CA, USA) and Hoechst 33342 (H3570; Invitrogen™, Carlsbad, CA, USA).Constructs following culture were firstly washed in PBS, next, cells were stained with 10 µM of Calcein-AM for 30 minutes at 37 °C, washed three times with PBS, followed by 5 pg/mL of Hoechst for 15 minutes at 37 °C.After staining culture media was re-added, and a small piece of the construct was taken and constrained between a microscope slide and coverslip and visualised on an Olympus BX60 microscope (Olympus Life Science, Southend-on-Sea, UK) using 494 nm filter (fluorescein isothiocyanate (FITC)) and 361 nm filter (4,6diamidino-2-phenylindole (DAPI)) for visualising Calcein-AM and Hoechst stains, respectively.After the culturing for up to 3 weeks constructs were harvested for assessment of cell morphology, matrix deposition and analysis of cellular phenotypic markers, along with acellular controls.

Scanning Electron Microscopy
Triplicate samples were removed from culture and flash frozen with liquid N 2 and stored at -80 °C.The samples were freeze dried overnight using a freeze dryer (FD-1A-50; Boyikang, Beijing, China) at -50 °C under vacuum.The samples were then fractured to expose the interior surface morphology, attached onto an aluminium stub, and then using a Quorum Technology 150 Q TES system (Quorum Technology, East Sussex, UK) (FEI NOVA, Oregon, USA).Secondary electron images were obtained using accelerating voltage 5 kV at various magnifications ranging from 600× to 6,000×.

Histological Analysis
NC phenotype and matrix deposition was investigated in acellular controls and cellular constructs following 2 and 4 weeks in culture.Triplicate samples per biomaterial were removed from culture and fixed in 10 % w/v neutral buffered formalin for 20 minutes and processed to wax using the TP1020 tissue processor (Leica Microsystems, Newcastle, UK).Following fixing and embedding, all samples were sectioned at 6 µm and mounted to X-tra® adhesive positively charged slides (Leica Microsystems, Newcastle, UK).Sections were dewaxed in Sub X (Leica Microsystems, Newcastle, UK) three times for 7 minutes and rehydrated in industrial methylated spirit (IMS; Fisher, Loughborough, UK; 10552904) for three times for 7 minutes and hydrated in running tap water for 5 minutes prior to histological staining.Sections were assessed using histological stains.Haematoxylin & Eosin (H&E): Mayer's Haematoxylin (MHS32; Sigma, Dorset, UK) for 1 minute, before being 'blued' in running water for 5 minutes and immersed in Eosin (3801601; Leica Microsystems, Newcastle, UK) for a further 1 minute; Alcian Blue (pH 2.5, 38016SS3; (Leica Microsystems, Newcastle, UK) for Weigert's Haematoxylin for 5 minutes, followed by 0.4 % v/v aqueous Fast green (F7252; Sigma, Dorset, UK) for 4 minutes, rinsed with 1 % v/v acetic acid (695092; Sigma, Dorset, UK) and counter stained with 0.125 % w/v Safranin O (S2255; Sigma, Dorset, UK).After specific histological staining, sections were dehydrated three times for 10minute washes in IMS, cleared in Sub-X (3803670E; Leica Microsystemcs, Newcastle, UK) three times for 10-minute and mounted in Pertex (3808707E; Leica Microsystems, Newcastle, UK).All slides were examined with an Olympus BX51 microscope and images captured by digital camera and Capture Pro OEM v8.0 software (Media Cybernetics, UK).Histological sections were analysed, and representative images captured to document their histological appearance and cellular staining patterns.

Immunohistochemistry and Immunofluorescence Analysis
Immunohistochemistry (IHC) and immunofluorescence were also performed on porcine disc samples and pNCs cultured in biomaterials.Specific antibodies were used to target antigens in relation to NP matrix, NC and NP phenotype (Table 1).Samples embedded in paraffin wax were dewaxed in Sub X for three times for 7 minutes and rehydrated in IMS for three times for 7 minutes.For IHC, as previously reported (Binch et al., 2020), endogenous peroxidase activity was blocked for 30 minutes using IMS containing 3 % w/w H 2 O 2 (8.22287;Sigma, Dorset, UK) and 5 drops of 37 % hydrochloric acid.Followed by three times for 5 minutes washes in tris-buffered saline (TBS; 20 mM tris (10528830; Fisher, Loughborough, UK), 150 mM sodium chloride (10092740; Fisher, Loughborough, UK), pH 7.5).Sections were subjected to antigen retrieval methods as detailed in Table 1.For heat antigen retrieval, sections were immersed in pre-heated (60 °C) antigen retrieval buffer (0.05 M tris, pH 9.5) and irradiated for 5 minutes at 40 % power in a microwave oven (Sanyo 800 Watt) and then again for 5 minutes at 20 % power.Sections were left to stand at room temperature for 15 minutes.For enzyme antigen retrieval, sections were placed in a preheated (37 °C) TBS buffer with 0.01 % w/v α-chymotrypsin (C4129; Sigma, Dorset, UK) in 0.1 % w/v CaCl 2 for 30 minutes at 37 °C; or no retrieval was required.Sections were washed in TBS and blocked for 2 hours with normal serum from which the animal in which the secondary antibody was raised in; 5 % v/v bovine serum albumin (BSA, A7030; Sigma, Dorset, UK) w/v in 75 % TBS and 25 % normal serum v/v (Table 1).Primary antibody was applied overnight at 4 °C according to Table 1 alongside equivalent concentration IgG controls.On the second day, the sections were washed in TBS followed by appropriate biotinylated secondary antibody for 30 minutes.Sections were subject to avidin-biotin-complex (ABC) elite reagent for 30 minutes (PK-7100; Vector Laboratories, Newark, CA, USA), 3,3-diaminobenzidine tetrahydrochloride (DAB) solution (D5637; Sigma, Dorset, UK) for 20 minutes and washed in running tap water for 5 minutes prior to counterstaining with Mayer's Haematoxylin for 1 minute and blued in running tap water for 5 minutes.Sections were dehydrated in IMS three times for 10 minutes, cleared in Sub-X three times for 10 minutes, and mounted in Pertex.

Statistical Analysis
Statistics were performed in GraphPad prism v8.4.1 (GraphPad Software, Inc., San Diego, CA, USA).The normality of the data was assessed using the Shapiro-Wilk test.The cell distribution when seeded within the biomaterials was not normally distributed, therefore analysed with a Kruskal  Primary antibody target, clonality, dilutions optimised and antigen retrieval.Along with the Secondary antibodies and dilutions, that was used alongside for immunohistochemical staining the staining methods.

Rheological Properties
Storage moduli of the three biomaterials (postgelation) were monitored over a range of strain values (Fig. 1).During the strain sweep (or amplitude sweep) experiment, all three hydrogels behaved like a viscoelastic solid characterised by a higher storage modulus (G ′ ) than loss modulus (G ′′ ).A gradual downturn of G ′ can be seen in every sample as the strain value was increased.However, the elastic portion of these materials prevailed the viscous portion throughout the strain range tested here (no G ′ and G ′′ crossover).In both acellular NPgel and acellular NPgel/dNCM systems, the storage moduli increased after 24 hours, reaching an average G ′ of ~14 kPa and ~3 kPa (at 1 Hz), respectively.These two hydrogel systems exhibited modest shrinkage (up to 20 % by volume) after 24 hours of media immersion.The slight change in volume was associated with the release of water as a result of interaction with the salts present in the media, creating a denser scaffold with higher G ′ .Albugel did not display shrinkage during the 24-hour period and maintained a G ′ value of ~0.8 kPa at both time points.The strain-independent region of an amplitude sweep, also known as the linear viscoelastic (LVE) region, describes the limits within which a viscoelastic material can be deformed non-destructively.Whilst all three biomaterials displayed linear viscoelastic properties within physiological strain ranges and thus were progressed for further cell testing, the LVE limit of Albugel (up to 40 % strain) exceeded that of NPgel (7 % strain) and NPgel/dNCM (10 % strain) (Fig. 1).

Optimising Cell Seeding Density within NPgel as an Example Biomaterial
An initial optimisation experiment was conducted to determine seeding density and to observe if different cell densities influenced the morphology and survival of pNC in vitro following 2 weeks culture within NPgel as the example biomaterial.Cells were visible in the 1 × 10 6 cells/mL constructs, but they were sparse, and presented as single cells (Fig. 2).Cell clustering was observed in biomaterial constructs seeded with 4 × 10 6 cells/mL, 1 × 10 7 cells/mL and 2 × 10 7 cells/mL.There was no observable difference between the 4 × 10 6 cells/mL and 1 × 10 7 cell/mL, and in the constructs seeded at 2 × 10 7 cells/mL displayed smaller clusters with dispersed single cells (Fig. 2).As clusters were first observed in 4 × 10 6 cells/mL, further experimental set ups were performed with 4 × 10 6 cells/mL seeding density in biomaterials.

Isolated Notochordal Cell Morphology in NPgel as a Model Biomaterial
Optimisation experiments were conducted to determine whether pNCs could be maintained within a biomaterial system.Phase contrast images demonstrated the presence of large cell clusters within NPgel, with predominant vacuoles seen up to 2 weeks in culture, which are still visible at 3 weeks (Fig. 3; vacuoles indicated by black arrowheads).

Porcine Notochordal Cells in Albugel, NPgel and NPgel/dNCM
Live-cell imaging with Calcein-AM and Hoechst staining, demonstrated that pNCs remained present and viable up to 2 weeks in all biomaterials.The maintenance of NC morphology and the retention of vacuole-like structures (indicated with white arrows) were observed in the 1-week NPgel and up to 2-week in the Albugel samples seeded with pNCs, although quantification was not possible given the clustered appearance of cells (Fig. 4).NPgel/dNCM with pNC showed less notochordal-like morphological phenotype with no vacuoles present at any timepoints (Fig. 4).

Scanning Electron Microscopy Morphology
SEM analysis of acellular constructs demonstrated a porous structure within all biomaterials.Albugel demonstrated a dense structure with fibrous pores (Fig. 5A).Whilst NPgel displayed a more honeycomb porous structure (Fig. 5C), which following inclusion of dNCM showed the appearance of thinner strands (Fig. 5E).Whilst constructs containing pNCs demonstrated apparent cellular structures within all gel systems (Fig. 5).Although constructs were heterogeneous, with some areas filled with cellular structures, other areas were devoid of cells (Fig. 5B,D,E), suggesting an uneven distribution of pNCs within the constructs.The majority of cells within all three biomaterials were seen in small clusters (Fig. 5B,D,E), although some single/duplicate cells were observed in Albugel constructs (Fig. 5Biii).Within NPgel constructs some clusters displayed a visible 'membrane' which appeared to wrap the entire pNC cluster (Fig. 5D; white arrowheads).

H+E staining of pNC & NPgel and pNC &
NPgel/dNCM demonstrated the seeded cells remained in cluster formation, whereas the pNCs in the Albugel construct appeared more dispersed (Fig. 6).Alcian Blue demonstrated GAG staining around the cell clusters (stained blue) in pNC & NPgel constructs (Fig. 6).In the NPgel and NPgel/dNCM constructs (Fig. 6) Masson's Trichrome blue staining surrounding the cells indicates collagen expression in the surrounding material (Fig. 6).Collagen expression is also indicated in Safranin Orange stain by pericellular green/turquoise staining (Fig. 6).Albugel constructs failed to demonstrate clear matrix indication when stained with Alcian Blue or Masson's Trichrome, however Safranin Orange stain indicated slight green/turquoise staining surrounding the pNC (Fig. 6).
The number of cells per cluster was determined fol-European Cells and Materials Vol.47 2024 (pages 30-50) DOI: 10.22203/eCM.v047a03lowing 4 weeks in culture.The number of cells in clusters were significantly different between each biomaterial, with the number of pNCs in clusters when cultured in NPgel, were significant greater when compared to NPgel/dNCM (p ≤ 0.01) and significantly greater when compared to Albugel (p ≤ 0.0001); NPgel/dNCM were significantly greater when compared to Albugel (p ≤ 0.001; Fig. 7).With the highest number of cells observed per cluster of 40 cells in NPgel, 26 in NPgel/dNCM and 14 cells in Albugel, with the median number of cells observed per cluster was 5 in NPgel, 2 in NPgel/dNCM and 1 in Albugel (Fig. 7).Altogether, this demonstrated that pNCs present themselves in larger clusters in NPgel and NPgel/dNCM biomaterials, whereas in Albugel they were mostly single cells.

Notochordal Cell Characterisation and Phenotypic Analysis within Novel Biomaterials
Immunohistochemical staining was further used to analyse and characterise the pNC that were cultured in the biomaterials for 4 weeks (Fig. 8A).The pNCs seeded into all three biomaterials showed nuclear and cytoplasmic positive staining for the NC/NP cell markers brachyury and cytokeratin 8/18/19, which when quantified showed no significant difference in expression between the biomaterials (p > 0.05) (Fig. 8B).Extracellular immunohistochemical staining for aggrecan was shown in the pNCs seeded into Albugel, NPgel and NPgel/dNCM, with extracellular immunohistochemical staining for collagen type II only distinguishable in the pNC & NPgel constructs (Fig. 9).How- ever, staining was also observed in the acellular NPgel and NPgel/dNCM construct, due to the nature of biomaterials retaining DAB (Fig. 9).
GAGs released into the culture media from pNC and biomaterial constructs was assessed throughout the 4week culture period and analysed using DMMB assay (Fig. 10).The acellular NPgel/dNCM constructs had a significantly higher GAG release into the culture media compared to acellular NPgel and Albugel biomaterials at 1 week (NPgel/dNCM vs Albugel, p ≤ 0.001; NPgel/dNCM vs NPgel, p ≤ 0.001), 2 week (NPgel/dNCM vs Albugel, p ≤ 0.0001; NPgel/dNCM vs NPgel, p ≤ 0.01) and 3week (NPgel/dNCM vs Albugel, p ≤ 0.01; NPgel/dNCM vs NPgel, p ≤ 0.01) culture time points but was reduced following 4 weeks in culture (p > 0.05).This may suggest release of the natural GAGs coming from the dNCM within the biomaterial, with no further increase of GAG release into media within NPgel/dNCM materials seeded with pNCs.No significant difference in GAG release was observed between acellular controls and pNC-containing constructs at any time-point (Fig. 10).

Discussion
NCs have attracted considerable interest due to their potential application for regeneration of the intervertebral disc (Bach et al., 2022), but to date, to the authors knowledge only one study investigated the potential of one biomaterial to support their viability and matrix production with limited phenotypic investigations made, with studies mainly investigating simple 3D culture systems to maintain phenotype for in vitro applications.
This study investigated three potential biomaterials which are all injectable and could provide potential mechanical support to the IVD whilst delivering the regenerative cell source.The selected biomaterials represented a naturally inspired cross linked hydrogel (Albugel), a fully synthetic hydrogel which gels based on entanglement of polymer chains (NPgel) (Boyes et al., 2021), and the supplementation of the synthetic hydrogel with naturally derived dNCM (NPgel/dNCM) (Schmitz et al., 2022).All three acellular biomaterials demonstrated viscoelastic stability within the physiological strain range observed for IVDs.NPgel exhibited a higher gel strength (characterised here by G ′ ) compared to NPgel/dNCM and Albugel.Furthermore, NPgel and to a lesser extent NPgel/dNCM showed an increase in G ′ following 24 hrs in culture media which mimics the degenerate IVD.The G ′ of NPgel following 24 hrs in degenerate media reached ~14 KPa which is similar to that of human NP tissue reported previously of 7-21 kPa (Iatridis et al., 1997; Iatridis et al., 1996).We have previously reported other rheological properties of NPgel  including its gelation kinetics (Boyes et al., 2021).Initial studies with pNCs utilised NPgel to determine the seeding density for future studies.Porcine NCs were seeded at different cell densities ranging between 1 × 10 6 cells/mL to 2 × 10 7 cells/mL into NPgel, where pNCs presented mainly in clusters and easier to identify within sectioned materials at 4 × 10 6 cells/mL and 1 × 10 7 cells/mL.Whilst lower densities of 1 × 10 6 cells/mL displayed mainly single cells and the highest cell density of 2 × 10 7 cells/mL displayed both clustered and single cells but showed no advantage over lower seeding densities.The native cell density of the NP has been reported to be 1.3 × 10 5 cells/cm 3 in nonchondrodystrophic dogs (Hunter et al., 2003), and 4 × 10 6 cells/cm 3 in humans (Maroudas et al., 1975).The importance of preservation the pNC clusters phenotype was also hypothesised to support NC survival and function (Aguiar et al., 1999; Gantenbein et al., 2014; Humphreys et al., 2018; Hunter et al., 2003; Spillekom et al., 2014).Additional, NCs are also reported to be highly metabolically active and sensitive to nutrient deprivation (Guehring et al., 2009; Spillekom et al., 2014), therefore the optimum cell density was selected as 4 × 10 6 cells/mL, as cluster formation was retained without excessive cell density which could provide nutrient deprivation over long term culture and is also in agreement with the cell density reported in human NP (Maroudas et al., 1975; Roughley, 2004).Based on brightfield images pNCs showed clear cluster and vacuole structures up to at least 2 weeks in culture, thus following these successful initial results, the complete screening experiment was performed, seeding pNC into three biomaterials that have shown potential for disc regeneration.

Retention of Notochordal Cell Cytomophology in Biomaterials
From the histological images, it was promising to observe that pNCs survived for up to 4 weeks in all three biomaterials, however the morphotypic pNC was not retained in all biomaterials.In Albugel, Calcein-AM images, SEM and histology, showed a lack of large clusters, with most cells observed in single cells.However, some preservation of vacuole-like structures was observed after 2-weeks and key NC/NP markers were retained, including brachyury and keratin 8/18/19.In native NP tissue, single cells are usually smaller and resemble more the mature NP cells often seen in a more mature and degenerate disc (Chen et al., 2006; Hunter et al., 2003), suggesting that the pNC in Albugel were differentiating into these mature NP cells.Within the literature their remains some debate to whether the population of smaller mature NP cells are derived from the juvenile NCs or replaced from alternative progenitor cells (Kim et al., 2009), whereas these results support the perspective that NCs have the capability to differentiate into mature NP cells, which can retain their expression of brachyury and cytokeratin (Minogue et al., 2010; Risbud et al., 2010).In NPgel, the pNCs predominately presented in clusters and contained vacuolated structures evident in Calcein-AM stained live-cell images, within this biomaterial clusters containing up to 40 cells per cluster were maintained.However, following long term culture of 4-weeks the vacuole-like phenotype was lost despite the pNC remaining in clusters.Similar findings were also observed during a longer-term culture of NCs in alginate beads, despite the maintenance of phenotypic markers, the morphology of the cells changed from large NC clusters to NCs with smaller vacuoles and mature NP cells (Arkesteijn et al., 2015).The loss of vacuole morphology along with loss of phenotypic markers was also noted when NC were cultured in monolayer (Arkesteijn et al., 2015) and when injected directly into NP tissue (Arkesteijn et al., 2017).In NPgel/dNCM, histological images, SEM and cell cluster size analysis demonstrated the presence of clusters and vacuole like structures were observed, although this was not visualised within live-cell imaging.A study by Ma et al. (2013) utilised rabbit NCs seeded in a link-N selfassembling peptide scaffold, demonstrating that rabbit NCs viability was maintained and could deposit aggrecan and type II collagen after 14 days of 3D culture and reported the presence of phenotypic vacuoles at 7 days of 3D culture but did not study further their phenotype using other phenotypic markers.

Retention of Notochordal Cell Phenotype in Biomaterials
The responded to the turquoise GAG stain within Alcian Blue (Leung et al., 2009).Interestingly, collagen type II staining was only strongly observed in pNC cultured within NPgel.IHC demonstrated the retention of brachyury and cytokeratin NC/NP phenotypic markers.Studies using NCs in alginate beads have also shown cell survival during culture, which corresponded with brachyury and cytokeratin expression and an increase in GAG detection (Aguiar et al., 1999; Arkesteijn et al., 2017; Gantenbein-Ritter & Chan, 2012).Within the current study there was limited GAG release evidenced from pNCs cultured within the biomaterials, with only Albugel demonstrating a difference between acellular and cellular constructs.Potentially suggesting the hyaluronic acid within the Albugel increases GAG production.Alternatively, the lack of GAG release into the media in NPgel constructs could indicate greater GAG retention in the NPgel constructs containing pNCs, as histological staining did display cellular GAG staining using both Alcian Blue and Safranin O. Whilst NPgel supplemented with dNCM demonstrated high secretion of GAGs into the media even in acellular controls over the first 3 weeks, this decreased after 4 weeks in culture.Suggesting loss of the dNCM (which is GAG rich) into the media over the initial time course.The decreased GAG content in the media after 4 weeks may indicate exhaustion of the GAG content from the dNCM or potentially the formation of a more complex matrix network leading to the trapping of the GAGs at later time course.However, given the low collagen staining observed with Masson's Trichrome staining and immunohistochemical staining for collagen type II it is more likely the dNCM had leached out over the first 3 weeks.The incorporation of dNCM into NPgel is based on a physical entanglement rather than any physical cross linkages and thus it may be necessary to develop biomaterials which enable chemical cross linking of NCM into the materials to ensure longer term retention.Additionally, in Ma et al. (2013) a Link-N self-assembled scaffold biomaterial promoted the accumulation of aggrecan and collagen type II when seeded with rabbit NCs (Ma et al., 2013).Furthermore, when such a system is injected into the NP tissue space in a whole IVD the GAG release would likely be retained in the disc and thus this leaching out may not be a detriment follow-  et al. (2007) would enable a better understanding whether the observed matrix staining was produced from the pNCs themselves rather than a remnant from the biomaterial.

Vacuoles within Cultured Notochordal Cells
The seeding of pNC in the three biomaterials resulted in the presentation of morphologically different cells, which was characterised by the gradual loss of vacuoles, but a retention of NC cell marker expression.The purpose of NC vacuolation in embryonic development of the spine is focused on retaining hypertonic tension (Hunter et al., 2007).Taking NCs into an in vitro setting does have the limitation of altering the cells environment; in this study the constructs were maintained in physiologically relevant hypoxic conditions of 5 % oxygen (Chen et al., 2014; Soukane et al., 2005), and using media recommended for IVD 3D culture (Basatvat et al., 2023).Nevertheless, with the isolation of cells, there is a change of environment such as the loss of mechanical stress and extracellular matrix which otherwise would have promoted water retention, although by maintaining cell clusters the pericellular matrix should be retained (Oegema, 1993; Urban et al., 1978).All three biomaterial systems investigated here are hydrogelbased biomaterials, which have high water-soluble polymeric materials.In the biomaterial that contains NC matrix, (NPgel/dNCM), there was some evidence of the retention of vacuole-like structures.In previous studies where altered media osmolarity has been tested, from standard 300 to 500 mOsm/L this resulted in a sustained canine NPC population with improved phenotype, irrespective of oxygen conditions (Laagland et al., 2022; Spillekom et al., 2014), suggesting that there could be many factors that can affect the NC phenotype.It may be that the lack of external mechanical and hydrostatic tension caused the difference in NC morphology, as vacuoles function to withstand compression forces (Bagwell et al., 2020; Ellis et al., 2013; Hong et al., 2018; Wang et al., 2017).Although compression loading has also been shown to be detrimental to NC health (Guehring et al., 2010; Hong et al., 2018; Spillekom European Cells and Materials Vol.47 2024(pages 30-50) DOI: 10.22203/eCM.v047a03 et al., 2014; Yurube et al., 2014), whilst other studies have shown no effect on NC viability and phenotype (Saggese et al., 2020).Thus, it would be important to investigate if these pNC seeded biomaterial constructs subjected to physiological compression and loading, would affect morphological NC and their phenotype.

Future Directions
This study was conducted within an in vitro setting, to test the potential of biomaterials to retain NC phenotype, as potential delivery systems for regenerative strategies.Such systems will need to be further tested within conditions which mimic more closely the degenerative environment of the disc, such as presence of catabolic cytokines especially IL-1β (Phillips et al., 2015).Furthermore, other features of the degenerate IVD include low glucose which was mimicked in these studies, however low glucose is also accompanied by low pH, the effects of which remain to be determined.If these cell-seeded biomaterials were to become clinically translated as an injectable treatment for disc regeneration, the next phase of testing would be to determine whether the cells could be maintained within a degenerative disc micro-environment.Such a study would need to investigate the complex mico-environment, beyond simply investigating physioxia and low glucose, as investigated in this study.Future testing should also involve analysing mechanical properties of the whole cell-biomaterial construct ex vivo, through applying a physiological loading regime (e.g., simulated daily loading).An example of this type of regime would be described as dynamic loading in a human physiological range of between 0.2-1 Mpa disc pressure at 1 Hz (Korecki et al., 2008; Korecki et al., 2010; Chan et al., 2011; Wilke et al., 1999) and may involve a complex loading regime similar to that reported in Le Maitre et al. (2009), such studies will provide a more through understanding of the mechanical stability of the materials in combination with cells within a physiologically relevant environment.

Translational Potential
We report here excellent cytocompatibility of the biomaterials screened and provide an insight on how the biomaterials will perform with NCs.If the cell-seeded biomaterial is to be used in a clinical setting as a novel injectable therapy for disc regeneration, it would be regulated as an advanced medicinal product (ATMP).The upscaling of a product to clinic involves attaining predictable robustness, batch-to-batch variability, safety profiling, reproducibility, and proficiency.Up-scaling the NC-based therapy co-delivered with the seeded biomaterial would involve generating a working cell bank and the involvement of allogeneic cells not only being a far better 'off-the-shelf' business model but a necessity since the human species lose their phenotypic NCs in childhood.As such the options for NC sources include fetal/baby human NCs which comes with significant ethical challenges and are limited in supply.Alternatively, the used of primary xenographic cells from pigs is a possibility, such an approach is growing in clinic applications, as safety is improving with the use of crossspecies cells (Sachs, 1994).Based on the successful use of porcine islets cells transplantation in a primate diabetes model in clinical trials, which is accomplished by knocking down the gene for alpha-1,3-galactosyltransferase lacking alpha 1,3-gal sugars on the porcine cell surfaces, that reacts to the innate immune system (Lai & Prather, 2002; Kuwaki et al., 2005; Thompson et al., 2011; Shin et al., 2015).

Conclusions
Three biomaterials, Albugel, NPgel and NPgel/dNCM representing differing classes of natural (Albugel), synthetic (NPgel), and semi-synthetic (NPgel/dNCM) biomaterials were investigated, to determine if they would maintain pNCs phenotype as a first stage to develop a biomaterial delivery system for the IVD.The results determined that pNC can retain viability and maintain NC cell markers in biomaterials, Albugel, NPgel and NPgel/dNCM for up to 4 weeks in physiological 5 % v/v oxygen, however loss of clustering was observed within Albugel and loss of the morphotypic vacuolated NC phenotype was lost or decreased in all three biomaterials.Therefore, the biomaterials selected have the potential to harness NC regenerative properties, however future work is needed to determine whether retention of the morphotypic phenotype is required to provide regenerative properties and mechanical testing of biomaterials following cell cultures is required to understand the true biomechanical stability following application.

Fig. 1 .
Fig. 1.Storage moduli of the three biomaterials (post-gelation).Amplitude sweeps of the NPgel, NPgel/dNCM and Albugel after 2and 24-hours immersion in culture media which mimics the degenerate conditions of the disc (OSM, pH, glucose).Experimental setup is shown on the right-hand side.

Fig. 3 .
Fig. 3. Brightfield live cell imaging of porcine notochordal cells cultured in NPgel, for up to 3 weeks in a hypoxia unit of 5 % v/v oxygen at 37 °C.Scale bar 100 µm.Black arrowheads indicate vacuoles.

Fig. 4 .
Fig. 4. Immunofluorescence staining of live cell imaging of porcine notochordal cells cultured in NPgel, NPgel/dNCM and Albugel, for up to 2 weeks in a hypoxia unit of 5 % v/v oxygen.Images were captured using Calcein-AM and Hoechst 33342.White arrowheads indicate vacuoles.