1 Introduction

Low back pain (LBP) constitutes a major cause of disability to work with a global lifetime prevalence of up to 80% of the population [1–4]. It is a high socio-economic burden in western industrialized countries, implying direct costs for medical treatments and insurances [5–7]. In most cases, LBP is associated with intervertebral disc (IVD) degeneration [1, 2, 8], and may be related to genetic disorders, aging and declining cell nutrition [2, 8–10], or caused by physical trauma, abnormal weight-bearing forces and fatigue failure of the matrix [2, 9, 10].

The IVD is composed of: (i) a central part, the nucleus pulposus (NP); (ii) a peripheral region, the annulus fibrosus (AF); and (iii) cartilaginous endplates (CEPs), that superiorly and inferiorly confine the AF and NP and constitute an interface with the vertebral bodies (VBs) [3, 4, 6]. IVD degeneration is a process characterised by loss of proteoglycans and increased amount and crosslinking of type I collagen accompanied by a decrease of type II collagen (Col II) in the NP. This disturbed matrix turnover is followed by a reduction of tissue hydration, which hampers its mechanical properties. As the NP becomes unable to evenly re-distribute compressive forces between the VBs, those are transferred non-uniformly to the surrounding AF, culminating in a progressive structural IVD deterioration, accompanied by vascular and neural ingrowths [2, 3, 7, 11].

Nowadays, available treatments for IVD degeneration range from conservative to surgical treatments, depending on the degeneration stage [3, 12]. However, these options fail to restore IVD structure and mechanical function [3, 7], because they do not target the underlying disease mechanisms, neither do they cease the progression of IVD degeneration.

Novel strategies able to achieve a homeostatic adjustment and tissue regeneration are being actively pursued [3, 4, 13]. Strategies based on the injection of mitogens, morphogens, anti-catabolic and/or intercellular regulators to modulate resident cell proliferation/differentiation, matrix synthesis and/or down-regulate catabolic and inflammatory processes are being investigated [14, 15]. Examples of such active biomolecules include transforming growth factor beta-1 (TGF-β1) [16, 17] and platelet-rich plasma (PRP - a cocktail of growth factors) [18, 19]. On the other hand, small-scaled particles are starting to emerge as alternative approaches for IVD regeneration. Their size enables them to influence cellular processes (particularly at the nano-scale [20, 21]). Since they correlate with structural sizes of biological components [22] and are able to pass through biological barriers [23], being internalised into target cells [24]. Fullerol nanoparticles, in particular, were able to decrease IVD degenerative features in human cells and rabbit tissue [25]. In turn, poly(L-glycolic acid) microspheres coated with heparinized nanoparticles that release TGF-β3 and dexamethasone were shown to enhance PG accumulation in rat IVDs [26].

Moreover, the discovery of new biomaterials mimicking the properties of healthy IVD has also been encouraged [27]. Poly(γ-glutamic acid) (γ-PGA) is a poly(amino acid) that recently called attention for cartilage regeneration: γ-PGA-containing scaffolds were shown to support rat chondrocyte culture [28] and γ-PGA injections were patented for the treatment of degenerative joint diseases [29]. In addition, our group recently showed that γ-PGA-supplemented medium enhanced the production of cartilaginous matrix by human mesenchymal stem/stromal cells (MSCs) [30]. Since the IVD comprises chondrocyte-like cells and skeletal progenitors [31, 32] with close resemblance to bone marrow-derived MSCs [32–36], we hypothesise that γ-PGA can recover the healthy IVD ECM in a nucleotomised ex vivo IVD model. Other attractive features of γ-PGA (such as non-immunogenicity and biodegradability) have also been previously described [30]. On the other hand, γ-PGA/chitosan (Ch) polyelectrolyte complexes (PECs) have the ability to improve intrinsic properties shown by each counterpart on its own (e.g., cytocompatibility and hydrophilicity [37–41]), enhance bulk properties (e.g., mechanical properties [42, 43]) and also act as delivery systems with improved characteristics (e.g., pH-responsiveness [44–48], higher loading efficiency [44, 49] and effectiveness [40, 49]). As nanocomplexes (NCs), different γ-PGA/Ch PECs have shown an added value in different applications [50–53]. In particular, γ-PGA/Ch NCs previously developed by our group [54, 55] were recently proposed as an effective anti-inflammatory drug delivery system that successfully reduced macrophage activation [54].

In this regard, the present study explored the potential of γ-PGA (in its soluble form) and γ-PGA/Ch NCs self-assembled through electrostatic interactions, to regenerate a nucleotomised bovine IVD using an ex vivo organ culture system. The ex vivo model used was previously described [56, 57], and has the advantages of being highly reproducible and mimicking early IVD degeneration through loss of NP, while maintaining AF integrity [57].

2 Experimental section

2.1 Materials

Ch (France-Chitine) was purified as described in Antunes et al. [58]. Ch with a degree of acetylation (DA) of 12.8 ± 0.4% (determined by Fourier transform infrared spectroscopy with KBr pellets (FTIR-KBr)), and a molecular weight (Mw) of 538 ± 74 KDa, (determined by size-exclusion chromatography), was selected. γ-PGA has been produced by Bacillus subtilis and purified as described in Pereira et al. [55]. γ-PGA with low molecular weight (10–50 kDa), high purity grade (>99%), and a D-/L-glutamate ratio of 50–60/50–40% was used [58]. The endotoxin levels of Ch and γ-PGA, below the limit 0.5 EU/mL [59], were controlled by the Limulus Amebocyte Lysate (LAL) QCL-1000 kit (Cambrex/Lonza).

2.2 Preparation of γ-PGA/Ch NCs at pH 5.0

γ-PGA/Ch NCs were prepared as previously described (as illustrated in Fig. 1) [54, 55]. Briefly, γ-PGA/Ch NCs were obtained through complex coacervation, by dropping γ-PGA solution (0.2 mg/mL) into a stirring Ch solution (0.2 mg/mL), both prepared in 0.05 M Tris-HCl buffer containing 0.15 M NaCl, using a 1 mL syringe in a syringe pump (KD Scientific Inc., Holliston, U.S.A.) at constant speed (3.6 μL/s) and room temperature (RT). Ch solution was obtained by diluting a Ch stock solution (5 mg/mL Ch in 0.2 M CH3COOH). The solutions pH was adjusted to 5.0 using HCl or NaOH before the assembly.

Fig.1
figure 1

a Molecular units of the polymers used to form the nanocomplexes (NCs): Chitosan (Ch, left), as a polycation, and poly(γ-glutamic acid) (γ-PGA, right), as a polyanion. b Schematics of γ-PGA/Ch NC preparation. NCs were obtained by complex coacervation driven by electrostatic interaction at pH 5.0, upon addition of a γ-PGA solution to a Ch solution under high stirring at RT. Counterions and polyions conformation were omitted for clarity (adapted from (55, 56) [59, 60]) (color figure online)

Full size image

2.3 Preparation of γ-PGA/Ch NCs at pH 7.4

γ-PGA/Ch NCs prepared at pH 5.0 were centrifuged (15000 rpm, RT) for 30 min, 2 h after initial preparation. γ-PGA/Ch NCs from the pellet were concentrated 10 times in the buffer solution at pH 5.0 (Supplementary Fig. S1). Concentrated γ-PGA/Ch NCs were then diluted 1 : 10 in cell culture medium (high-glucose Dulbecco’s Modified Eagle Medium (DMEM; Gibco)) supplemented with 2% Fetal Bovine Serum (FBS; Gibco), 1% penicillin/streptomycin (P/S; Gibco), 1 % Insulin-Transferrin-Selenium supplement (ITS; BD Becton Dickinson) and 0.1% Fungizone (Amphotericine B; PAA), a composition previously described for IVD organotypic cultures [55, 57]) at pH 7.4.

2.4 Characterisation of γ-PGA/Ch NCs by dynamic and electrophoretic light scattering (DLS and ELS) and transmission electron microscopy (TEM)

The size (d), polydispersity index (PdI), mean count rate (mcr) and surface electrical charge (ζ potential) were determined using a Zetasizer Nano ZS (Malvern Instruments, U.K.) equipped with a 4 mW HeNe laser beam with a wavelength of 633 nm and a scattering angle of 173° (for size measurements) and 13° (for ζ potential measurements). Size and ζ potential values were automatically calculated through the DTS Nano v.6.30 software, using the Stokes-Einstein equation or the Henry equation with the Smoluchowski approximation. The dispersion solution used was the solvent in which the NCs were formed, i.e., 0.05 M Tris-HCl buffer with 0.15 M NaCl. For the mean count rate, the attenuator and measurement position were fixed at the 8 and 3 mm, respectively, which were the automatically obtained values for the γ-PGA/Ch NCs prepared as described in Fig. 1. Disposable solvent resistant micro cuvettes or folded capillary cells with gold-plated electrodes (Malvern) were used. Data were acquired at 20 °C, with three repetitions for each sample. The morphology of the NCs was observed using a transmission electron microscope (TEM) JEM 1400 (80 kV; Jeol, Japan) with an OriusTM SC1000 Digital Camera (Gatan, U.S.A.). Samples were prepared by placing 10 μL of the NCs suspension onto a 400 μL mesh copper grid coated with carbon (EMS). Two minute after deposition, the grid was tapped with a filter paper and negatively stained with uranyl acetate (EMS) for 10 s. TEM images were obtained with different magnifications.

2.5 Preparation of fluorescent Ch (ftCh) and fluorescent γ-PGA/Ch NCs (γ-PGA/ftCh NCs)

Fluorescent nanoparticles were prepared according to Gonçalves et al. [54]. Briefly, Ch was labelled with fluorescein isothiocyanate (FITC) with 5% modification (5% of amine groups with FITC), whereby 100 mg of dried Ch were dissolved in 100 mL of 1% v/v CH3COOH at 4 °C until complete dissolution. FITC (11 mg to achieve 5% modification) was dissolved in 100 mL of methanol. Both solutions (Ch and FITC) were mixed at constant stirring, protected from light, for 3 h. The FITC-labelled Ch (ftCh) was then precipitated with 0.5 M NaOH and washed with ultrapure water until no fluorescence was detected in the supernatant. ftCh was then lyophilized, dried and weighted. γ-PGA/ftCh NCs (ftNCs) were prepared as previously described.

2.6 Quantification of γ-PGA/ftCh NCs internalisation by NP cells

To evaluate the capacity of ftNCs internalisation by IVD cells, NCs were incubated with IVD cells previously isolated from bovine IVD tissue. Six bovine caudal IVDs of a young adult animal (age < 2 years old) were collected within 3 h after slaughter from a local abattoir (Carnes Landeiro S.A., Nine, Portugal), with the ethical approval of the Portuguese National Authority for Animal Health. Each NP tissue sample was dissected into 2-3 mm3 fragments and enzymatically digested for 2 h in 1.0 mg/mL collagenase type I (Sigma-Aldrich) in DMEM, at 37°C, under agitation (50 rpm) and humidified atmosphere of 5% v/v CO2 in air. Supernatant was passed through a 40 μm filter (BD Falcon) to remove tissue debris. Cells were collected by centrifugation at 400 g for 7 min. ftNCs (0.7 mg/mL) were added to an IVD cell suspension at a concentration of 10% v/v in cell culture medium. Three hours later cells were fixed in 1% PFA and quantified by standard flow cytometry (FACSCalibur™ system, Becton Dickinson, NJ, U.S.A.) and imaging flow cytometry (ImageStream®, AmnisMillipore, MA, U.S.A.). Cell concentration and time point for analysis were selected based on our previous work with in vitro studies using an anti-inflammatory treatment on activated macrophages (55) [54]. Cells incubated without nanoparticles were used as control. ftCh fluorescence was assessed in Channel 2 (505–560 nm) and at least 5000 events were collected in a FACSCalibur. FACS data was analysed using FlowJo (version 8.7). For imaging flow cytometry, samples were collected in Image Stream (Amnis) and analysis was performed in IDEAS software (Amnis). An internalisation mask was applied to determine the percentage of cells with higher fluorescence on the inside rather than on the cell membrane.

2.7 Harvest of bovine IVD tissue

IVDs with CEPs were isolated from young (<1 year old) bovine tails (n = 6) obtained from the same slaughterhouse. Seven caudal IVDs from each bovine tail were harvested based on a protocol previously described [56, 60]. Briefly, the caudal IVDs with CEP were removed using a circular saw (Mini grinder DM-130A-1; Dremel, U.S.A.) with parallel cuts. The CEPs were washed afterwards with phosphate-buffered saline (PBS) solution (pH = 7.4) using a 50 mL syringe connected to a 21 G needle. Under sterile conditions, IVDs were incubated sequentially in 1/10/1% of penicillin/streptomycin (P/S; Gibco) in a PBS solution (pH = 7.4) for 1/10/1 min, respectively. Then, IVDs were placed in 6-well plates, incubated with the aforementioned cell culture medium and placed overnight in a humidified atmosphere with 5% CO2 at 37°C.

2.8 Ex vivo model of IVD organotypic culture

An ex vivo organotypic culture of IVD with CEP (as described in Pereira et al. [56]) was used to inject different solutions (Fig. 2). Briefly, a circular cavity of 4-8 mm3 was made in the CEP using a blade. The portion of each CEP was removed and kept in PBS with 1% P/S, while part of the NP (50-100 mm3 of tissue) was removed using another blade. The cavity created was filled (50 μL) with the following solutions:

  1. 1)

    γ-PGA/Ch NCs (0.4 mg/ml) in 0.05 M Tris-HCl with 0.15 M NaCl buffer (solvent pH 5.0);

  2. 2)

    γ-PGA/Ch NCs (0.4 mg/ml) in cell culture medium (solvent pH 7.4);

  3. 3)

    γ-PGA (0.2 mg/mL) in cell culture medium;

Fig. 2
figure 2

Intervertebral discs (IVDs) isolated from bovine tails. After removal of a portion of cartilaginous endplate (CEP) (~4-8 mm3) and nucleus pulposus (NP) (50-100 cm3), a cavity was created and filled with different solutions. Each removed CEP was then repositioned and sealed with cement. Scale bars represent 1 cm (adapted from (57, 58) [61, 62]) (color figure online)

Full size image

Additional experimental conditions included the respective controls with solvent at both pHs. The parameters evaluated were compared with those from intact IVDs obtained immediately after isolation (fresh).

Subsequently, each removed CEP was repositioned, sealed with cement (poly(methyl methacrylate) (PMMA), VertecemTM V Cement Kit, Synthes) and incubated for 30 min at 37°C, allowing the cement to dry. Next, IVDs were placed in 6-well plates with the previously mentioned cell culture medium for 2 or 7 days in a humidified atmosphere with 5% CO2 at 37°C. Media were changed every 2 days.

2.9 NP Haematoxylin-Eosin (HE) staining

IVD was sliced longitudinally and half of the disc was fixed in 3.7% PFA, washed in PBS and stored at 4°C. Subsequently, the fragments were placed into histology cassettes, embedded in paraffin wax, cut into 2-4 µm sections and mounted on glass slides (microscopic slides; MGF-slides), previously silanized with a 3-aminopropyltriethoxysilane treatment. Then, paraffin slices were immersed three times in xylene to remove paraffin, and then re-hydrated using decreasing concentrations of alcohol (twice in 100, 96, 70 and 50%) followed by distilled water. Next, IVD sections were incubated in Gill’s haematoxylin (Sigma-Aldrich) for 3 min to stain the cell nuclei. After washing in tap water and dehydrating through graded alcohol solutions (50, 70 and 96%, 3 min each), a counterstaining was performed in alcoholic eosin (Surgipath) for 1 min, to detect cell cytoplasm and most connective tissue fibres. After three quick passages in absolute alcohol and clearing in xylene, the slides were mounted with resinous mounting medium (Quick-D; Klinipath).

2.10 NP Safranin O-Light green staining

sGAG deposition was analysed by Safranin O-Light green staining of NP tissue. Paraffin-embedded IVD sections were incubated in Gill’s haematoxylin (Sigma-Aldrich) for 5 min to stain the cell nuclei. After washing in distilled water for 15 s, the samples were immersed in 0.4% Light green (Sigma) solution during 5 min to stain the cytoplasm. Following two washes (of 3 min each) in 1% acetic acid solution, slices were immersed for 4 min in 0.1% Safranin O (Sigma-Aldrich) solution to detect sGAG deposition. Then, samples were dehydrated through graded alcohol solutions (three quick dips in 96% and three times in 100% for 3 s each) and cleared in xylene. Finally, slides were mounted with resinous mounting medium (Quick-D, Klinipath).

2.11 Metabolic activity of IVD organotypic cultures

Metabolic activity was evaluated through the resazurin assay that measures the bioreduction of resazurin to resorufin, which can be detected using fluorescence (530/590 nm). To analyse metabolic activity of IVD organ cultures treated with the different solutions, the CEPs of each IVD in culture were removed using a scalpel. Each IVD was cut sagittally in half (comprising AF and NP) and incubated in 8 mL of cell culture medium containing 10% resazurin solution (0.1 mg/mL resazurin sodium salt (C12H6NNaO4, Sigma) in PBS) for 3 h at 37°C, to guarantee linearity between IVD weight and rezasurin conversion (Supplementary Fig. S2 compared results of 1 IVD comprising AF, NP and CEPs, 1 IVD without CEPs, plus ¾ IVD, ½ IVD and ¼ IVD also with AF and NP tissue. Results showed that the IVD metabolic activity correlates with tissue wet weight from 0 to 2.6 g, or tissue dry weight from 0 to 0.3 g, corresponding to ¾ IVD tissue). A blank control comprising only medium was also included. For each experimental condition, three independent donors were selected. Metabolic cell activity was expressed as relative fluorescent units and normalised to the tissue wet weight.

2.12 DNA quantification of IVD organotypic cultures

To quantify DNA content, frozen NP was minced into very small pieces with the help of tweezers and a scalpel. After freeze-drying, the samples were digested overnight at 56°C with proteinase K (0.5 mg/mL in phosphate buffer containing 10.68 g/L NaH2PO4.2H2O, 8.45 g/L Na2HPO4.7H2O and 3.36 g/L disodium-EDTA in ultrapure water) [57]. pH was previously adjusted to 6.5. After vortexing and centrifugation, the cellularity of each sample was measured based on the DNA content using a CyQuant® kit (Invitrogen) with lambda DNA as a standard. DNA content was expressed as relative fluorescent units and normalised to the tissue dry weight (Supplementary Fig. S3). For each experimental condition, three independent donors were selected.

2.13 sGAG quantification in IVD organotypic cultures

sGAG content of IVD organotypic cultures was analysed on NP tissue by reaction with 1,9-Dimethyl-Methylene Blue zinc chloride double salt (DMMB; Sigma-Aldrich) dye reagent solution, containing 40 mM sodium chloride (NaCl; Roth), 40 mM Glycine (Roth) and 46 μM DMMB, previously adjusted to pH 3.0. Chondroitin sulphate A sodium salt from bovine trachea (Sigma) was used as standard. Results were normalised to the corresponding DNA content, quantified as previously described. For each experimental condition, three independent donors were selected.

2.14 Free active TGF-β1 quantification

TGF-β1 concentration in IVD organotypic culture supernatants was analysed by ELISA (LEGEND MAXTM Free Active TGF-β1 ELISA Kit, BioLegend®), according to the manufacturer’s instructions. Results were normalised to the DNA content of NP tissue. Three independent donors were selected.

2.15 Detection of Col II in the NP

Col II deposition in NP tissue was analysed by immunofluorescence (IF) staining and western blot analysis. For IF, antigen retrieval was performed in paraffin-embedded IVD slices through incubation with a 20 μg/mL proteinase K (Sigma-Aldrich) solution for 15 min at 37°C. A blocking step was introduced to reduce non-specific background staining with 1 h incubation in PBS containing 1% bovine serum albumin (BSA; Sigma-Aldrich) and 0.1% Triton X-100 (Sigma). Sections were then incubated for 2 h at 37°C with mouse anti-Col II (1 : 50) in the dark (monoclonal antibody (II-II6B3) developed by Dr. Thomas F. Linsenmayer, from the Developmental Studies Hybridoma Bank under the auspices of the NICHD and maintained at the University of Iowa, Department of Biology, Iowa City, IA, U.S.A.). A fluorescent secondary antibody (Invitrogen) was diluted 1 : 1000 and incubated during 1 h. After washing in PBS, the samples were mounted in Fluoroshield mounting medium with DAPI (Sigma). As positive controls, human nasal cartilage sections were used, while samples without primary antibody were chosen as negative controls.

Col II detection was also performed by western blot analysis using the same antibody. Frozen NP was minced into very small pieces with the help of tweezers and a scalpel. Proteins were extracted in a buffer containing 4 M guanidine hydrochloride (Sigma), 3 M sodium acetate (Merck) and 10 mM ethylenediaminetetraacetic acid (EDTA; Sigma-Aldrich), and enriched with 1 mM of phenylmethylsulfonyl fluoride (PMSF; Sigma), a protease and phosphatase inhibitor cocktail (Roche Diagnostics GmbH and Sigma, respectively) and 65 mM of freshly added dithiothreitol (DTT; Sigma) for 24 h at 4°C. Following centrifugation, ethanol precipitation and ressolubilization, proteins were quantified using 2-D Quant Kit (GE Healthcare) according to the manufacturer’s instructions. Protein samples (15 μg) were separated by sodium dodecyl sulphate (SDS)/9% polyacrylamide gel electrophoresis, and electroblotted onto a Hybond enhanced chemiluminescence (ECL) membrane (Amersham BioscienceseGE Healthcare). After incubation with the monoclonal antibody against Col II (1 : 2000), an anti-mouse (1 : 3000; Amersham Biosciences) horseradish peroxidase-conjugated secondary antibody was used. Then, ECL detection (Amersham Biosciences) was performed. Bands were quantified using Quantity One® 4.6.6 Software (Bio-Rad, Amadora, Portugal). Values were normalised to the density of each corresponding lane, following brief membrane staining with PageBlue-Coomassie Brilliant Blue G-250, Thermo Scientific) and by Col II expression levels of the control (fresh IVDs). Three independent donors were selected.

2.16 Statistical analysis

Statistical analysis was performed using GraphPad Prism (version: 5.0a). The parametric distribution of the data was first evaluated by D’Agostino & Pearson omnibus normality test. Since the data followed a non-parametric distribution, statistical analysis was performed by the Friedman test followed by the Dunn’s multiple comparison post-test to compare more than 2 groups of paired samples. The Kruskal-Wallis test was used to compare more than two groups of unpaired samples. A confidence interval of at least 95% was chosen to define statistical significance (*p < 0.05, **p < 0.025, ***p < 0.001).

3 Results

3.1 Characterisation of γ-PGA/Ch NCs at pH 5.0 and pH 7.4

γ-PGA/Ch NCs were first prepared at pH 5.0, as previously reported [55]. To obtain γ-PGA/Ch NCs at physiological pH (with similar polymers concentration), NCs at pH 5.0 were concentrated in an acidic-buffered solution, and diluted in cell culture medium at pH 7.4. The comparison of NC features at both pHs is summarised in Fig. 3.

Fig. 3
figure 3

Characterisation of γ-PGA/Ch NCs. NCs were characterised immediately after preparation in 0.05 M Tris-HCl with 0.15 M NaCl buffer at pH 5.0 a and after incubation in cell culture medium (high-glucose Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 2% Fetal Bovine Serum (FBS), 1% penicillin/streptomycin (P/S), 1% Insulin-Transferrin-Selenium (ITS) supplement and 0,1% Fungizone (Amphotericine B)) at pH 7.4 b (n = 3). Size (d, nm), polydispersity index (PdI), ζ potential (ζ pot., mV) and mean count rate (mcr, Kcps) measurements, determined by dynamic and electrophoretic light scattering (DLS and ELS), were displayed for each step, together with particle size distribution plots obtained by intensity of the scattered light (using the same equipment) and representative transmission electron microscope (TEM) images (bars represent 500 and 100 nm, respectively) (color figure online)

Full size image

NCs obtained by complex coacervation at pH 5.0 resulted in spherical particles with occasional clusters of 2–3 individual NCs (d = 238.2 ± 39.1 nm). They presented a relatively monodisperse size distribution (PdI = 0.3 ± 0.0) with a ζ potential of 20.3 ± 0.5 mV (Fig. 3a). γ-PGA/Ch NCs were centrifuged and concentrated. Analysis of the resulting NC supernatant confirmed a low frequency of events with a size smaller than that of NCs, but also in a monomodal distribution. NCs were identified in the pellet resulting from centrifugation, with increased concentration, and in a homogeneous population, with identical ζ potential, confirming the success of NCs centrifugation step (Supplementary Fig. S1a and S1b). In turn, concentrated NCs diluted in cell culture medium (pH 7.4) resulted in a highly disperse size distribution (PdI = 0.9 ± 0.1) (Fig. 3b). And so, particles with three different sizes could be distinguished: the majority (61.1 ± 1.6%) with bigger size (767.2 ± 249.6 nm), and two other groups in lower percentages (27.1 ± 0.9 and 11.1 ± 1.2%) and lower sizes (11.6 ± 1.7 and 30.7 ± 1.3 nm, respectively). The number of total events detected was also smaller than those of Fig. 3a (NCs directly obtained from the complex coacervation method), 119.9 ± 2.6 kcps, and with a slightly negative average ζ potential (−9.4 ± 0.5 mV) (Fig. 3b). Cell culture medium alone was also analysed and small particles (13.3 ± 0.5 nm) were also identified, similarly to what was observed with NCs at pH 7.4 (Supplementary Fig. S1c).

Concerning NCs morphology, NCs presented a spherical shape in acidic pH, with the denser material (black) appearing to constitute the majority of the NCs, particularly their core. At pH 7.4, large aggregates of dense material were easily distinguished, with a less dense material surrounding them in an irregular and disperse manner. Still, the morphology of NCs at pH 5.0 (Fig. 3a) was not clearly detected at pH 7.4 (Fig. 3b).

3.2 Quantification of γ-PGA/ftCh NC internalisation by NP cells

γ-PGA/Ch NCs’ internalisation within NP cells was assessed through Ch labelling (ftCh) and further incubation of the γ-PGA/ftCh NCs with NP cells. By standard flow cytometry, 91.1% of the NP cells presented ftNCs (Fig. 4a). However, with this technique NC internalisation is not distinguishable from NC presence in the cell membrane. Therefore, NC internalisation was further assessed using image flow cytometry. With this technique it was possible to observe that from the 63.0% of NP cells with ftNCs (Fig. 4b), 34.9% (approximately 22% of total NP cells) presented higher fluorescence intensity inside the cell (R3 group), being the NCs effectively internalised, and approximately 62.6% of these cells presented NCs only associated with the cell membrane (R5 group) (Fig. 4c).

Fig. 4
figure 4

Flow cytometry analysis of fluorescent γ-PGA/Ch NCs (ftNCs) internalization by isolated bovine IVD cells from a pool of 6 discs. a Dot plot and histogram profiles for the ftNCs internalisation analysis by standard flow cytometry. b Dot plot profiles for the ftNCs internalisation analysis by imaging flow cytometry. c Histogram profile, after application of an internalisation mask, showing cells negative (R5) and positive (R3) for ftNCs internalisation. Each cell is represented by a row of three images acquired simultaneously in flow - from left to right: brightfield (grey), FITC fluorescence (green) from the NCs, merged image (color figure online)

Full size image

3.3 Impact of γ-PGA and γ-PGA/Ch NCs at different pHs (pH 5.0 and pH 7.4) on tissue architecture, metabolic activity and DNA content of ex vivo IVD organotypic cultures

The first impact of γ-PGA and γ-PGA/Ch NC treatments was evaluated in ex vivo IVD organotypic cultures after 2 days in culture. The general architecture of the NP central region, next to the cavity created, was visualised using HE histological staining (Fig. 5). An extensive presence of extracellular matrix (ECM) was clearly distinguished within fresh IVDs. The matrix contained some diffuse flashes of lighter eosinophilia, along with a low cell density. All cells presented a round-shaped morphology with lacunae formation (chondrocytic cells). The majority of the cells were found in 2-cell clusters, but larger groups of cells in close proximity were also frequently observed. In IVDs with cavity, other differences were detected: thin basic fibres could be perceived in some areas of the NP core through the eosin staining. The majority of the cells presented a round-shaped morphology, but some cells had a more elongated shape. Some 2-cell clusters were found within the matrix, along with single cells.

Fig. 5
figure 5

NP architecture of NP from IVDs after 2 days in culture, obtained by haematoxylin/eosin histological staining. A fresh IVD was compared with IVDs with cavities filled with: γ-PGA/Ch NCs at pH 5.0 and pH 7.4 (NCs), the respective solvents, and γ-PGA. One representative donor is presented. Columns give detailed images of distinct regions of each NP centre (bars represent 25 μm). Inserts represent general overviews over the latter regions (bars represent 500 μm). Cells are highlighted with black circles (occasional elongated cells; singular round-shaped cells with lacunae formation, double-eyed or 2-cell clusters) and black large arrowheads (groups of 3-8 round-shaped cells). The black thinner arrows highlight intense basic extracellular fibres (color figure online)

Full size image

With acidic solutions’ (solvent or γ-PGA/Ch NCs, at pH 5.0) injection, the NPs’ ECM appeared to be sparse: diffusely distributed shades of pale pink could be seen and fewer cells were present. Some cells appeared more elongated, while others seemed smaller and with an irregularly shaped cytoplasm, contrarily those seen with the other conditions. Most cells were found isolated and 2-cell clusters were rarely observed. When NCs were included, intensely stained bundles of basic fibrillar structures appeared arranged as a network. Cells, which displayed chondrocyte-like features with prominent lacunae, were mostly placed between those fibres. On the other hand, at pH 7.4 (solvent, γ-PGA/Ch NCs or γ-PGA, at pH 7.4), NP ECM seemed to be more organised. More intense bundles of eosinophilic ECM appeared to be observed in IVDs treated with NCs and γ-PGA. With NCs, the ECM looked more heterogeneous, with a few large acidophilic areas next to densely packed fibre bundles oriented as networks. In turn, with γ-PGA, the oriented structures were the dominant feature. At the cellular level, the main differences at this pH were occasional double-eyed chondrocytes sharing the same territorial matrix and several groups with more than 3 cells in close proximity.

Metabolic activity and DNA content of organotypic cultures, when in contact with solvent, γ-PGA or γ-PGA/Ch NCs were also assessed after 2 days in culture (Fig. 6). Both acidic conditions—solvent (*p < 0.05) or γ-PGA/Ch NCs (**p < 0.025), at pH 5.0 - led to a reduced metabolic activity of IVD cells when compared with fresh IVDs. When γ-PGA/Ch NCs were injected at pH 7.4, in cell culture medium, this effect was not observed, although γ-PGA injection at physiological pH also resulted significant metabolic activity decrease (*p < 0.05), as observed with acidic conditions.

Fig. 6
figure 6

a Metabolic activity of IVD organ cultures in the presence of γ-PGA and γ-PGA/Ch NCs after 2 days b NP DNA content of IVD organ cultures in the distinct settings after 2 days. The following groups were compared: fresh, solvent at pH 5.0 (0.05 M Tris-HCl with 0.15 M NaCl buffer) and γ-PGA/Ch NCs prepared in the buffered solution (NCs), and at pH 7.4: solvent (cell culture medium), solvent supplemented with γ-PGA/Ch NCs (NCs), or with γ-PGA. Results are shown as mean ± SD from three independent donors. Statistical significance was considered for at least *p < 0.05, (**p < 0.025, ***p < 0.001), using the Friedman paired test followed by the Dunn’s multiple comparison post-test

Full size image

NP DNA content was also analysed and its reduction was more evident when acidic solutions were injected into the IVD (**p < 0.025), in comparison with solutions at physiological pH (Fig. 6b). DNA content normalisation by dry weight was selected since the water content is the higher contributor to NP’s weight. So, small variations in NP tissue sampling can result in pieces of tissue with different water retention capacity, and a higher error associated with this normalisation (Supplementary Fig. S3). Thus, although IVD wet weight is widely used in the literature as a measure of IVD size or cell number [60–64], it is not the most accurate method for normalisation, since small variations in the water content of this tissue (that can be related to different IVD sizes, different region within the tissue(s), etc.) can have a pronounced effect on IVD wet weight.

In this study, the quantification of IVD cell metabolic activity could not be directly related with DNA content, since the former was analysed in an IVD fragment comprising both AF and NP, while DNA was quantified in the NP tissue near the cavity region of each IVD.

3.4 Impact of γ-PGA and γ-PGA/Ch NCs, at pH 7.4, on the ECM of ex vivo IVD organotypic cultures

To evaluate if NCs could have an impact on the production of cartilaginous matrix deposition in the IVDs with time, organ cultures were extended to 7 days. Since γ-PGA/Ch NCs at pH 7.4 had a significantly lower impact on tissue viability, only studies conducted at this pH were carried out and compared to the solvent at pH 7.4.

After 7 days in culture, a small increase in the DNA content of IVDs cultured with γ-PGA/Ch NCs and γ-PGA could be observed, although not statistically significant (Fig. 7a). Furthermore, the sGAGs of IVDs cultured with γ-PGA seemed to be more uniformly distributed throughout the IVD sections and in higher amounts than at day 2 (Figs. 7b and 7c). This observation was made after comparison with fresh IVDs. The possible interference of γ-PGA in the process of sGAG staining was discarded, after analysing AF from the same tissue sections and confirming that no red/orange staining was observed in neither of the samples (Supplementary Fig. S4). Endogenous free active TGF-β1 was quantified in culture supernatants (Fig. 6d), since TGF-β1 is known to inhibit matrix degradation, being associated with higher NP cell proliferation [65, 66], PG synthesis [16, 67] and Col II production when supplemented to the culture medium [68]. However, although slightly higher values were detected in the presence of γ-PGA after 7 days in culture, no statistical differences were observed between the conditions under study.

Fig. 7
figure 7

Evaluation of ECM synthesis of IVD organ cultures in the presence of γ-PGA and γ-PGA/Ch NCs at pH 7.4 during 7 days of culture. a DNA content of NP tissue; sGAG deposition b and distribution c (Safranin O/Light green histological staining). Representative donors are presented here. Columns give general overviews over distinct regions of each NP centre (bars represent 500 μm); inserts represent detailed images of the latter regions (bars represent 25 μm). d Free active transforming growth factor beta 1 (TGF-β1) concentration in culture supernatants. a, c and d Results are represented as mean ± SD from three independent donors/time point. Statistical significance was considered for at least *p < 0.05, (**p < 0.025, ***p < 0.001), using the Kruskal-Wallis test followed by the Dunn’s multiple comparison post-test (color figure online)

Full size image

Col II distribution within IVD organ cultures was assessed by IF during 2 and 7 days and its production quantified by western blot analysis after day 7. To compare the different conditions, image acquisition settings were maintained among the different samples (Fig. 8a). Col II labelling (red) appeared to be enhanced with γ-PGA and NC treatments, in respect to the corresponding solvent controls. Nevertheless, some heterogeneity (particularly with NCs) was displayed in regard to Col II distribution and intensity. Col II was also quantified by western blot, after 7 days in culture by western blot (Fig. 8b). When compared with fresh IVDs, a significant decrease (*p < 0.05) in Col II content could be observed in IVDs treated with solvent alone. For IVDs cultured in the presence of γ-PGA/Ch NCs, a recovery of Col II to levels close to those of freshly isolated IVDs was observed. On the other hand, a clear significant augment of Col II content was reached with γ-PGA-treated IVDs (*p < 0.05) in detriment of the solvent treatment.

Fig. 8
figure 8

Distribution (red) a and quantification b of type II collagen (Col II) deposition (red) in NP tissue cultured for 7 days in culture with different treatments at pH 7.4: cell culture medium (solvent), solvent supplemented with γ-PGA/Ch NCs (NCs), or with γ-PGA. First column contains low magnification images (bars represent 250 μm). The 2 columns to the right are detailed images of distinct areas of the cavity region (bars represent 25 μm). White arrows point to more intense Col II fibre bundles. IVDs from representative donors are presented here. For Col II quantification the results are represented as mean ± SD from three independent donors. Statistical significance was considered for at least *p < 0.05, (**p < 0.025, ***p < 0.001), using the Friedman paired test followed by the Dunn’s multiple comparison post-test. A representative immunoblot is shown (color figure online)

Full size image

4 Discussion

This study addresses the application of γ-PGA or γ-PGA/Ch NCs to IVD regeneration. Recently, our group has shown that γ-PGA increases and accelerates cartilaginous ECM production in MSC pellet in vitro cultures when used as a supplement in cell culture medium [30]. Herein, γ-PGA/Ch NCs, previously developed by our group [55], were also tested as a supplement in cell culture medium while injected directly into the cavity, an approach that promotes the closest contact possible with the damaged tissue. Hence, NCs were tested in the cell culture medium at pH 7.4. Notwithstanding, NCs were additionally evaluated in the solvent in which they were built - the aforementioned buffered solution at pH 5.0 - to be able to infer on their impact on degenerated IVDs.

Stable polyelectrolyte NCs were produced through electrostatic interactions promoted by an acidic pH, as previously described [54, 55]. Positive charge and TEM analysis suggest that the protonated Ch, the larger polyion, has a major contribution for polymers’ assembly (as suggested by Lin and colleagues [69]). To inject NCs at physiological pH, these were concentrated by centrifugation and resuspended in cell culture medium, at pH 7.4. Nevertheless, NCs’ dilution in the cell culture medium resulted in a higher polydispersed size distribution, in which the major contribution to the detected events came from larger compositions that indicated swelling and/or agglomeration of NCs. Particle agglomeration and swelling have already been reported by Hadju et al. [70]. This was expected due to deprotonation of the amine groups of Ch amine groups and γ-PGA chain expansion at neutral pH [70].

Both γ-PGA and γ-PGA/Ch NCs at pH 5.0 and 7.4 were injected in a nucleotomised IVD ex vivo culture model [56, 57, 60]. Injection of Ch alone was not explored in this study, since this polymer is not soluble at physiological pH [71]. Moreover, being Ch the only naturally occurring cationic GAG analogue [72–75], it has already been widely described for cartilage [75–77] and IVD [78–82] regeneration.

It is known that organ cultures provide a more physiological environment than in vitro conditions, in this case maintaining IVD cells in a microenvironment closer to that of an in vivo scenario (hypoxia, limited nutrition, etc.) [83–85]. The bovine model was selected due to having anatomic proportions close to the human IVD, similar matrix composition, cell phenotype and nutrient transport [83, 86]. IVDs injected with solutions at pH 5.0, particularly the solvent alone, showed the presence of cells with an irregular cytoplasm within the NP centre (possibly necrotic cells). On the other hand, the IVDs with solutions at pH 7.4 presented more cells and a denser matrix. These cells presented typical chondrocyte morphology and banded basic fibrillar structures (stained in pink). Such staining is related with positively charged structures like cationic amine groups of Col, suggesting the existence of a Col network [3, 4, 10]. Ch itself could also take part in these networks, as defended by Roughley et al. [82].

Previous studies by our group have shown no cytoxicity of γ-PGA and γ-PGA/Ch PECs on human MSCs/chondrocytes [30] and fibroblasts/macrophages [54, 58] cultures, respectively. Herein, no major cytotoxicity was observed in IVD, although a decline in bovine NP cell metabolic activity has been reported by pH reduction to 6.2 within 24 h of in vitro culture, particularly under nutrient deficit conditions (low oxygen and low glucose concentrations) [87]. γ-PGA/Ch NCs were shown to be internalised by NP cells, although we cannot exclude that the biological effect of these NCs could be due to interaction with membrane receptors, as suggested by other works [40, 88–90]. In this work, imaging flow cytometry technique was used due to its potential as a high throughput method compared with standard flow cytometry (which fails to accurately quantify internalised vs. cell bound particles due to low resolution [54]), and regular fluorescence microscopy (as shown by Phanse et al. [91] for cellular internalisation of nanoparticles).

A significant reduction in DNA content of IVDs treated with NCs at pH 5.0 was also observed. However, at pH 7.4 this effect was not observed for all the solutions tested (solvent, γ-PGA/Ch NCs or γ-PGA).

In light of the results obtained, given that it is known that bovine NP cells do not survive in vitro at pH 6.0 for more than 12 days [92, 93], show detrimental features (water loss, elevated intracellular ionic strength, subsequent cell shrinkage with physical distortion of the nuclear matrix and cytoskeleton disorganisation [94], reduction of NP cell ability to synthesise novel DNA [94, 95], increase of cell senescence and decrease of cell viability [96–99]) after exposure to high salt concentrations (as with the Tris-HCl with NaCl buffer solution), only the effect of γ-PGA and γ-PGA/Ch NCs in the solvent at pH 7.4 was investigated in IVD ECM after 7 days in culture. Moreover, since increased cartilaginous matrix formation could be observed in MSC pellet in vitro cultures in which γ-PGA-supplemented cell culture medium was used [30], the chosen solvent for the present analyses was also cell culture medium, here adapted to IVD organotypic cultures. The latter solvent has been considered an ideal solution to maintain cell viability, ensure the feasibility of whole-organ ex vivo cultures and rely on the results obtained [56, 57, 100, 101]. An improvement to its effect would be an achievement worth exploring. Notwithstanding, the previous treatments were not all compared at this time point also due to the limited availability of IVDs from the same bovine tail.

Interestingly, at pH 7.4, an increase in PG staining was observed in the presence of γ-PGA and an evident augment in Col II synthesis was detected in the presence of both γ-PGA and γ-PGA/Ch NCs, but particularly with γ-PGA treatment. These results show that γ-PGA, besides promoting Col II production by human MSCs [30], can have a similar effect in the NP. Of notice, all the outputs obtained in this study were analysed at the protein-level; an added value compared with most IVD studies that only show outputs at gene expression level [57, 61, 102]. Although the specific mechanism of action of γ-PGA remains to be investigated, we hypothesise that the glutamate pool generated upon γ-PGA biodegradation may activate the glutamate signalling pathway [103]. Glutamate signalling is known to modulate chondrocyte proliferation [104] and their response to mechanical stimulation [105]. Despite the fact that no specific data has been reported for Ch’s mechanism of action in IVD cells, other polysaccharides (of herbal origin [106, 107]) have been shown to promote rat chondrocyte proliferation [106, 107] and increase Col II expression [107] through the Wnt/β-catenin signalling pathway. On the other hand, Peng and colleagues [88] have shown that γ-PGA/Ch NCs are internalised by human fibrosarcoma cells (HT1080) via non-specific charge-mediated interaction between the NCs (positively charged) and the cell membrane components (negatively charged proteoglycans via a specific protein-mediated endocytosis).

Moreover, both γ-PGA solution and γ-PGA/Ch NC dispersion are liquid solutions and easy to inject. This is an advantage compared with highly viscous hydrogel solutions. Although we cannot exclude the risk of leakage, a beneficial effect was observed with these solutions, using a single-dose injection. Therefore, this effect could be further explored by evaluating multiple injections, different polymer/NC concentrations and biomolecules’ delivery, among others.

5 Conclusions

The application of γ-PGA and γ-PGA/Ch NCs on nucleotomised IVD was screened for the first time. γ-PGA/Ch NC injection at acidic pH was detrimental to IVD organ cultures, decreasing the DNA content of the tissue(s). Nevertheless, these effects were overcome by NC injection at pH 7.4. In this condition, γ-PGA alone and γ-PGA/Ch NCs were able to promote Col II production after 7 days in culture. This study is the first proof of concept that γ-PGA-based injections can be used in a degenerated IVD context. Promising outcomes can be perceived for the future use of γ-PGA-based strategies in its soluble form or in the form of NCs that may additionally deliver biologically active molecules into the IVD.