Induction of zonal-specific cellular morphology and matrix synthesis for biomimetic cartilage regeneration using hybrid scaffolds

Abstract

Cartilage is anisotropic in nature and organized into distinct zones. Our goal was to develop zonal-specific three-dimensional hybrid scaffolds which could induce the generation of zonal-specific cellular morphology and extracellular matrix (ECM) composition. The superficial and middle zones comprised two layers of hyaluronic acid (HA) hydrogel which enveloped specifically orientated or randomly arranged polylactic acid nanofibre meshes. The deep zone comprised a HA hydrogel with multiple vertical channels. Primary bovine chondrocytes were seeded into the individual zonal scaffolds, cultured for 14 days and then the ECM was analysed. The aligned nanofibre mesh used in the superficial zone induced an elongated cell morphology, lower glycosaminoglycan (GAG) and collagen II production, and higher cell proliferation and collagen I production than the cells in the middle zone scaffold. Within the middle zone scaffold, which comprised a randomly orientated nanofibre mesh, the cells were clustered and expressed more collagen II. The deep zone scaffold induced the highest GAG production, the lowest cell proliferation and the lowest collagen I expression of the three zones. Assembling the three zones and stabilizing the arrangement with a HA hydrogel generated aligned, randomly aggregated and columnar cells in the superficial, middle and deep zones. This study presents a method to induce zonal-specific chondrocyte morphology and ECM production.

1. Introduction

Articular cartilage (AC) is a thin hydrated tissue which covers articulating surfaces. Damaged AC has a limited potential for self-healing, and untreated cartilage defects often progress to osteoarthritis (OA) [1]. High hopes have been pinned on regenerative medicine strategies to meet the challenge of preventing progress towards OA. Clinically, a range of approaches are available to treat patients with small lesions, including microfracture, autologous chondrocyte implantation (ACI) and osteochondral auto- and allo-grafts [2]. The main limitations of these surgical approaches are that they are less effective for defects with a large surface area. Cartilage is difficult to engineer; one reason is that it is anisotropic in nature. With increasing depth from the articular surface, there are at least three distinct architectural zones (superficial, middle and deep) with striking variations in structure, chondrocyte phenotype and extracellular matrix (ECM) composition [3–5]. Collagen fibres in the superficial zone are densely packed and orientated parallel to the articular surface, which enables the dissipation of high tensile stress; in the middle zone, they have a random orientation to provide resistance to the multi-directional shear and compressive force; and in the deep zone they are orientated perpendicularly to the surface, which helps to facilitate high torsion and compressive stress resistance. The differences in the above zonal matrices induce chondrocytes to become elongated and to have a rounded-clustered and short columnar morphology in the three zones. This contributes significantly to the differences in the respective compressive properties along the depth and location of the AC [6–8].

If we are to repair a joint for patients with large cartilage defects, it is important to generate AC with full functional capacity. Replication of the zonal organization of tissue- engineered cartilage is one of the multiple strategies to generate functional cartilage. To date, there have been a limited number of studies describing zonal cartilage tissue engineering [9–15]. In cell-based methods, zonal chondrocytes have been isolated [9] and employed in specific regions of a construct to retain their zone-specific phenotype and secrete specific zonal ECM components, such as aggrecan, collagen II and collagen X [10–12]. Other studies have used bi- or multi-layered hydrogels to support cartilage production within different zonal sub-populations. For example, Ng et al. [13] built zonal populations of chondrocytes by altering the concentration of hydrogel scaffolds, e.g. agarose. Another strategy is the construction of hydrogel scaffolds with different chemical and mechanical properties for different zones. Nguyen et al. [14] constructed zones using a combination of chondroitin sulfate and poly- (ethylene glycol) containing matrix metalloproteinase-sensitive peptides. Steele et al. [15] created a multi-zone cartilage template by using electrospun polycaprolactone (PCL) nanofibres. Analysis of the multi-zone scaffolds demonstrated region-specific variations in chondrocyte number, ECM composition and chondrogenic gene expression [15]. However, the bilayered scaffolds lacked a suitable hydrated environment.

Given the pitfalls described above, our aim of this study is to develop three novel hybrid three-dimensional (3D) zonally distinctive scaffolds through spatially organized nanofibre meshes in combination with a hyaluronic acid (HA) hydrogel to induce zonally distinct chondrocyte differentiation and orientation. Previously, Wise et al. [16] used natural and synthetic electrospun polymers to mimic the topography of the existing ECM. Polylactic acid (PLA) is a US Food and Drug Administration-approved polymer widely used for medical devices and in the field of musculoskeletal tissue engineering [17]. Herein, we applied our expertise in producing electrospun PLA nanofibres to induce the expected orientation of chondrocyte alignment and corresponding ECM secretion. We combined our PLA with HA because HA is a major component of synovial fluid and cartilage, and is of great importance to the maintenance of chondrocyte phenotype and functions [18]. Initially, we studied the influence of our individual scaffolds on both chondrocyte morphology and ECM production. Aligned and randomly arranged nanofibres in combination with HA hydrogel were used to create the superficial and middle zones, respectively. The HA hydrogel included defined micro-channels to mimic the characteristics of the native deep zone. Following careful testing, all three zones were combined in a single construct to create a full 3D zonal construct with biomimetic zonal organization in order to attempt to recreate native AC ECM.

2. Material and methods

2.1. Isolation of chondrocytes

Full-depth AC was dissected from the articulating surface of the humeral trochlea of 18-month-old cows. Four separate isolations were performed, each using one humerus. Chondrocyte isolations were performed with a minor modification of our previously described protocol [19]. Briefly, diced cartilage was sequentially digested with 0.1% (w/v) proteinase K for 1 h and then 0.3% (w/v) collagenase type IA for 3 h. Chondrocytes from the digestion supernatant were strained through a 70 µm cell sieve and washed in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) fetal calf serum (FCS). The filtrate was centrifuged at 750g. The cell pellets were washed three times with DMEM.

2.2. Fabrication of polylactic acid nanofibres

PLA nanofibres were generated by electrospinning as previously described [20]. Briefly, 0.2 ml PLA solution (2% (w/v), poly-l,d-lactic acid (96% l/4% d; Purac BV, The Netherlands)) was delivered at a rate of 0.025 ml min⁻¹ by a syringe pump through an 18G needle. The needle and the collector were connected to a power supply charged at ±6 kV (Spellman HV, Pulborough, UK).

In order to produce highly aligned nanofibre meshes with low line density to replicate the collagen fibre distribution within the superficial zone of the AC, a rectangular portable collector was used [21]. A metal ring collector was used to collect random nanofibres to replicate the ECM within the middle zone of the AC. After electrospinning, the produced aligned nanofibre mesh was transferred to a cellulose acetate frame to make it portable. The mesh was sterilized under ultraviolet light for 90 s three times.

2.3. Preparation of hyaluronic acid gel

HA sodium salt powder with an average molecular weight of 1.5 × 10⁶ Da was supplied by Shangdong Freda Biopharm Co., Ltd. China. HA gel was prepared by crosslinking HA molecules with the crosslinker 1,4-butanediol diglycidyl ether (BDDE; Sigma), according to our established protocol [22]. Briefly, HA powder was dissolved in 1% NaOH at a concentration of 10%, after which BDDE was added to the HA solution with stirring to a final concentration of 0.4%. The solution was then allowed to crosslink at 40°C for 5 h followed by drying at room temperature for 3 days. For rehydration, phosphate-buffered saline (PBS) was then added to the crosslinked HA. The swollen HA was dialysed against deionized water and then against PBS to remove any residual BDDE.

The crosslinked HA gel was used to generate the deep zone or the base for individual zones. To enable the crosslinked HA gel to be formed into the other zones, it was pulverized with a homogenizer to obtain mini-gel particles of 0–400 µm. The mixture of gel particles was mixed with 2% (w/v) agarose hydrogel in a ratio of 9 : 1.

2.4. Fabrication and assembly of individual zone constructs

Superficial zone construct: an aligned PLA nanofibre mesh was placed on a 1 cm² slice of HA gel to form the hybrid scaffold. Chondrocytes (1 × 10⁵ per sample) were seeded onto this hybrid scaffold. After 2 h, the reconstituted HA gel (100 µL) was applied to the scaffold surface to stabilize the cell and nanofibre positions, thus forming the superficial zone construct.

Middle zone construct: using a randomly aligned PLA nanofibre mesh, the middle zone was constructed in the same way as the superficial zone.

Deep zone construct: the crosslinked HA hydrogel (1 cm² × 0.5 mm) was first prepared and then 10 microscale channels per sample were formed vertically within the HA hydrogel using a micro-needle (400 µm). Chondrocytes (1 × 10⁴ cells per channel) were seeded in the scaffold to form the deep zone construct.

Control construct: chondrocytes (1 × 10⁵ per sample) were mixed with the reconstituted HA gel without nanofibre or channels to provide the negative control.

All constructs were cultured with DMEM supplemented with 10% (v/v) FCS, 1% (v/v) l-glutamine and 1% penicillin–streptomycin (v/v) for 14 days at 37°C in 5% CO₂. The medium was exchanged every three days. At the end of the culture period, zonal scaffolds were freeze-dried, and digested individually with 300 µl papain solution per sample overnight at 60°C. Papain solution was prepared by dissolving 125 µg papain in 0.1 M sodium phosphate, 5 mM EDTA and 5 mM cysteine–HCl at pH 6.5.

2.5. Fabrication and assembly of the multiple zone construct

Multiple zonal constructs were formed by the layer-by-layer assembly method [21]. Briefly, the hydrogel with micro-channels was placed on a polytetrafluoroethylene plate, which was seeded with cells at a cell density of 1 × 10⁴ cells per channel. The cells were allowed to attach at 37°C, 5% CO₂ for 2 h. PLA random nanofibre meshes were placed on top of the hydrogel and 1 × 10⁵ cm⁻² chondrocytes were seeded onto the nanofibre meshes. The cells were allowed to attach at 37°C for 2 h, and 100 µl of the reconstructed HA hydrogel was then loaded on top of the construct. The aligned nanofibre meshes were placed directly on top of the hydrogel and chondrocytes were seeded onto the nanofibre meshes at a cell density of 1 × 10⁵ cm⁻². The cells were allowed to attach in the incubator for 2 h. The 3D constructed samples generated were sealed with HA hydrogel and cultured as individual constructs in 24-well plates. To visualize the cell locations in the assembled constructs, Hoechst 4′,6-diamidino-2-phenylindole (DAPI) dye (1 µg ml⁻¹ PBS) was added to samples for 10 min at room temperature, and then washed twice with PBS. A confocal laser scanning microscope (Olympus Fluoview FV 1200 with Fluoview software (4.1 version)) was then used to observe the intact constructs.

2.6. Scaffold characterization and assessment of cell behaviour within the scaffolds

2.6.1. Morphology of nanofibres and micro-channels

Nanofibre scaffolds were imaged using a scanning electron microscope (SEM; Hitachi S4500). Specimens were coated with gold using an Emscope 200 (Emscope, UK) sputter coater for 2 min and imaged at an accelerating voltage of 15 kV. Fibre diameters were determined using ImageJ (1.51j 4) [20]. Briefly, a calibrated line drawing tool was used to measure cross-sections of individual fibres. Two separate nanofibre sub-samples were examined in a minimum of three different areas; approximately 50–200 fibres were measured for each sample.

Optical coherence tomography (OCT; Telesto II, Thorn lab, USA) was used to observe the micro-channel morphology and the dimensions of the assembled 3D hybrid zonal scaffold. OCT utilized the wavelength centred at 1300 nm, providing approximately 1 mm image penetration.

2.6.2. Mechanical testing of scaffolds

The mechanical properties of the individual and assembled zonal scaffolds, the bovine cartilage tissue (positive control) and the HA gel alone construct (negative control) were measured by uniaxial compression testing using an Electro force Model 3200 testing machine (BOSE), equipped with a 22 N load cell operated at a crosshead speed of 0.05 mm s⁻¹. The specimen dimension was measured as 1 × 1×0.05 cm for individual zonal scaffolds and 1 × 1 × 0.06 cm for the 3D full zonal scaffold. The compression modulus was determined from the linear region of the stress–strain curve (between 0.1 and 0.5 strain), in which the applied forces had been converted into stress with the samples' area (stress = force/area). The height difference between the samples was taken into consideration when calculating the strain values obtained from the displacement (ΔL) divided by the initial height (L) (strain = ΔL/L). The ultimate compression strength was taken as the maximum stress.

2.6.3. Cell viability

The live chondrocyte morphology within the 3D constructs was observed using transmitted light microscopy (Olympus CKX41; Olympus, UK). The images were taken at different culture time points for cells in separated and assembled constructs. The viability of the cells was determined using a viability/cytotoxicity kit (Invitrogen, Paisley, UK). Briefly, cell culture medium was removed from the samples. The samples were washed with PBS and immersed in a solution containing 10 mM calcein-AM and 1 mM propidium iodide at 37°C for 30 min in the dark. Calcein-AM labels viable cells green and propidium iodide labels dead cells red. Cell staining was observed using a confocal microscope. The visualization of the 3D samples was achieved by scanning sections from the bottom to the top at 10 µm intervals. Individual images were then reconstructed into 3D models using Imaris 8.1 analysis software (Bitplane, UK).

2.6.4. Cell number

Cell numbers were determined using Picogreen^® fluorescent DNA quantification with lambda DNA as the standard (1 ng ml⁻¹–1 µg ml⁻¹). Duplicate samples and standards were transferred to a 96-well plate and fluorescence was read at an excitation wavelength of 480 nm and an emission wavelength of 520 nm on a BioTec Synergy 2 microplate reader. The cell number was calculated using the widely reported value of 7.7 pg of DNA per chondrocyte [23].

2.6.5. Total sulfated glycosaminoglycan

Total sulfated glycosaminoglycan (sGAG) was assessed using the 1,9-dimethylmethylene blue (DMMB) dye assay as previously described [24]. All reagents were obtained from Sigma. A 4× DMMB solution (32 mg DMMB, 1.52 g glycine, 1.19 g NaCl, 47.5 ml of 0.1 M HCl, in 200 ml H₂O, pH 3.0) was prepared. Bovine tracheal chondroitin sulfate standards (0–200 µg ml⁻¹) were prepared in distilled water. Duplicates (50 µl) of each papain-digested sample and standard were added to a 96-well plate. The above DMMB solution (200 µl well⁻¹) was added to all wells and the plate was immediately read at 530 nm on a BioTec Synergy 2 plate reader.

2.6.6. Western blotting

After 14 days of culture, the protein in each sample was extracted by digestion with radioimmunoprecipitation lysis buffer (Sigma, UK). The protein concentration was quantified by the bicinchoninic acid method [25]. Cell lysate containing 50 µg of protein was electrophoresed by 10% sodium dodecyl sulfate polyacrylamide gel and then transferred to a polyvinylidene fluoride membrane (Millipore Corp., MA, USA). The membrane was blocked with 5% (w/v) skimmed milk for 1 h at room temperature and incubated overnight at 4°C with primary antibodies against collagen type II (Santa Cruz; 1 : 400), collagen type I (Abcam; 1 : 400) and aggrecan (Santa Cruz; 1 : 400).

2.6.7. Immunostaining

Three samples for each culture group were fixed with 4% (w/v) paraformaldehyde at room temperature for 30 min. Immunostaining was performed using primary antibodies against the following proteins: collagen type II (Santa Cruz), collagen type I (Abcam) and aggrecan (Santa Cruz). All samples were subjected to an unmasking treatment prior to the staining in accordance with an established protocol [26]. Briefly, for detection of collagen type II and collagen type I, samples were initially treated with 2 mg ml⁻¹ testicular hyaluronidase (Sigma). For detection of aggrecan, samples were pre-treated with 25 mU ml⁻¹ chondroitinase ABC (Sigma). After pre-treatment, the samples were incubated with the primary antibodies and then labelled with tetramethylrhodamin-conjugated secondary antibody for collagen type II and collagen type I and fluorescein isothiocyanate-conjugated secondary antibody for aggrecan.

AC was used as a gold standard positive control. The samples without primary antibodies were used as negative controls (electronic supplementary material, figure SI). All cells were evaluated using the same exposure time, gain and offset camera settings in the confocal microscope so that the immunofluorescence intensity was directly comparable for each given antibody across different groups. 3D visualization of samples was achieved by scanning sections from bottom to top. Individual images were then reconstructed to 3D models, using Imaris 8.1 analysis software.

2.7. Statistical analysis

For each analysis, data are expressed as the mean ± standard error of the mean (s.e.m.). Statistical analyses were performed with SPSS statistical software (15.0). All experimental groups had a sample size of a minimum of three replicates for biochemical analyses and five replicates for mechanical property analyses. Three independent experiments were undertaken. Statistical significance was determined by performing one-way ANOVA, followed by Tukey's post hoc test. Significance was accepted at a p-value < 0.05.

3. Result

3.1. Scaffold morphology and construct mechanical properties

SEM images of the highly aligned and randomly arranged nanofibre meshes are shown in figure 1a(i,ii), respectively. The fibre diameter between the two types of fibres did not differ significantly (518 ± 60 nm versus 486 ± 80 nm; table 1). It was demonstrated clearly by visible light and OCT images (figure 1a(iii),b(iii)) that the cylinder-like channels were formed into the HA gel with an average channel diameter of 489 µm ± 100 (table 1). The individual scaffolds fabricated through a combination of electrospun nanofibres and HA hydrogel are illustrated in figure 1b. The cross-section of the assembled 3D zonal scaffold with zonal organization (superficial, middle and deep zone) is illustrated by an OCT imaging technique (figure 2b).

Figure 1.

Images of individual 3D zonal scaffolds. (a) Top view; (b) side view. (i) Aligned nanofibre for superficial zone; (ii) random nanofibre for middle zone; (iii) channels in HA hydrogel for deep zone. (i,ii) SEM images; (iii) light microscope and OCT images. Stars denote the HA hydrogel layer. Scale bar is 100 µm.

Table 1.

Physical parameters of the hybrid zonal scaffolds, HA gel alone scaffold and bovine cartilage tissue.

scaffold cartilage mimicking	scaffold orientation	fibre and channel size (µm)	ultimate compression stress (kPa)	compression modulus (kPa)
superficial	aligned (horizontal)	0.518 ± 0.06	143 ± 32.35	15.47 ± 3.5
middle	random	0.486 ± 0.08	149 ± 28.11	16.16 ± 2.9
deep	vertical channels	489 ± 100	153 ± 12.41	15.85 ± 4.2
full scaffold construction	—	—	209 ± 40.35	25.12 ± 4.4
HA gel only scaffold	—	—	145 ± 31.83	17.67 ± 3.8
bovine cartilage tissue	—	—	3410 ± 983	570 ± 164

Figure 2.

(a) Schematic illustration of the assembled three separately cultured zonal scaffolds with corresponding thickness. (b) An OCT image of the assembled 3D zonal scaffold showing aligned nanofibres in the superficial zone, random nanofibres in the middle zone and vertical channels in the deep zone. Scale bar is 250 µm. (Online version in colour.)

It was demonstrated by compression modulus measurement of individual scaffolds that there was little variation between the superficial, middle and deep zones and the HA gel alone scaffold, but the full 3D zonal scaffold had a higher compression modulus than each of the individual scaffolds. However, all scaffolds had a lower compression modulus than native cartilage tissue (table 1). The representative load deformation curves are displayed in the electronic supplementary material, figure SII.

3.2. Chondrocyte morphology in individual and multiple constructs

The morphology of the chondrocytes on each 3D zonal scaffold is illustrated in figure 3 by light microscopic observation. Within the superficial zone, it can be seen that the individual chondrocytes appear to stretch along the aligned nanofibres from the first day of culture and further developed into a highly aligned morphology at day 7 of culture. At 2 weeks, these cells proliferated into highly organized bundle structures along the nanofibres. By contrast, chondrocytes maintained their round shape and random distribution in the randomly oriented nanofibres within the middle zone during the 2 weeks of culture, while cell proliferation could still be ascertained as the density increased or as the cell aggregates noticeably appeared. Within the HA channels of the deep zone, the chondrocytes appeared to form vertical round-shaped cells which stacked up within the channels. The chondrocytes in the control constructs (HA hydrogel alone) exhibited a round shape with random distribution or clusters at late culture time points.

Figure 3.

Microscopic images of the live chondrocytes in the three separately cultured zonal scaffolds and HA gel alone constructs showing their morphology and orientation during culture at different time points (n = 3). Scale bar is 150 µm.

3.3. Cell viability

Cell viability in the different 3D zonal hybrid scaffolds was demonstrated using live/dead cell staining and observed by confocal microscopy, as shown in figure 4. All 3D zonal scaffolds had high amounts of viable cells (predominantly green-stained cells) with only a few dead cells (a few red-stained cells), but the superficial zone scaffold had the highest number of viable cells and the lowest quantity of dead cells in comparison with the middle zone and deep zone scaffolds. High cell viability was demonstrated in the control constructs.

Figure 4.

The live and dead staining images of chondrocytes in three separately cultured scaffolds and HA gel alone construct taken at day 14 of culture. The insert is the reconstructed z-stage image showing the cells within the vertical channel, with the channel direction indicated by an arrow. Green indicates live cells, and red dead cells. Scale bar is 100 µm (n = 3).

3.4. Cell number

A DNA quantitative assay was used to show the cell number change over time for each zonal scaffold, as shown in figure 5a. It was demonstrated that, at all culture time points, the scaffolds forming the HA gel alone, the deep zone and the middle zone had almost the same cell number, which was lower than those in the superficial zone scaffolds (p < 0.05).

Figure 5.

Biochemical analysis of the three separately cultured scaffolds and HA gel alone constructs. (a) Cell number; (b) total content of GAG; (c) normalization of GAG content, μg per cell, after 14 days of culture. Data are expressed as the mean ± s.d. (n = 3). * indicates statistical significance by one-way ANOVA analysis.

3.5. Total sulfated glycosaminoglycan content

The total amount of sulfated GAG accumulated in the different 3D zonal scaffolds was quantified as shown in figure 5b. The HA gel alone, deep and middle zone scaffolds had the highest GAG content when compared with the superficial zone scaffold (p < 0.05). The normalized GAG content (μg) per cell in each different 3D nanofabricated zonal construct is illustrated in figure 5c. The GAG content with respect to the DNA present was highest in the deep zone scaffold compared with the other zones, especially at later time points (p < 0.05).

3.6. Extracellular matrix components

AC-specific ECM markers, collagen II and aggrecan within each zone construct were characterized by immunostaining, as shown in figure 6. Immunostaining for collagen I was also carried out to determine cell phenotype. It can be seen that the cells in the superficial zone sample had the highest collagen I production but the lowest collagen II and aggrecan. The middle and deep zone samples demonstrated higher collagen II and aggrecan production than the superficial zone. With regards to collagen I, the staining intensity decreased with the depth-dependent constructs, which was in contrast to the aggrecan, which increased through the depth-dependent scaffold. For collagen II, the highest staining was seen in the middle zone samples and the lowest was observed in the superficial zone samples. It was noted that collagen II was present in all scaffold types, which was not the case for collagen I staining. The staining images of native bovine cartilage samples showed the depth variation of collagen I, II and aggrecan with collagen I in the superficial zone only; there was collagen II staining across all zones and more intense aggrecan staining in the middle and deep zones (electronic supplementary material, figure SIII).

Figure 6.

Immunostaining images of chondrocytes cultured in the three separately cultured zonal scaffolds and HA gel alone constructs. (a) Superficial zone; (b) middle zone; (c) deep zone; (d) HA hydrogel alone. Blue, nuclei; green, aggrecan; red, collagen type II or collagen type I. Scale bar is 100 µm (n = 3, three independent experiments).

Western blotting (figure 7; scrap-book images in electronic supplementary material, figure SIV) showed that the superficial zone sample produced a higher amount of collagen type I and a lower amount of collagen type II and aggrecan. The collagen type II exhibited a higher level in the middle zone scaffold with lower collagen type I production. Aggrecan, one of the main markers of AC, was increasingly exhibited with depth. The deep zone had the highest production of aggrecan and lowest amount of collagen type I. These results concurred with the immunostaining data.

Figure 7.

Western blotting of collagen type I, II and aggrecan production of chondrocytes in the three separately cultured zonal constructs after culturing for 14 days.

To demonstrate that the chondrocytes could be incorporated and maintained in the individual zones after the three zones were assembled, DAPI was used to label the cells. The three different zoned constructs were seeded with DAPI-labelled cells and then cultured for 14 days. The cross-sectional reconstructed 3D images of the assembled construct are depicted in figure 8. This result suggests that the zonal structure was well maintained and that the cells were well distributed across the zones. The live/dead staining assay showed a high cell viability within the assembled constructs (figure 8a,b).

Figure 8.

Reconstructed 3D images of DAPI-labelled chondrocytes in the assembled 3D zonal scaffolds alongside live/dead kit staining covering the superficial, middle and deep zones (presenting one vertical channel only). (a) Live cells (green); (b) dead cells (red); (c) DAPI (blue). Scale bar is 250 µm.

4. Discussion

The zonal organization of scaffolds that mimic the organization and structure of AC tissue and potentially induce the synthesis of appropriate ECM within the separate zones could lead to long-term functionality in cartilage regeneration. There is no doubt that this is a complex challenge to tackle. Our data and other studies confirm that native cartilage has a different ECM distribution alongside the depth (electronic supplementary material, figure SIII). In our study, we have fabricated hybrid cartilage scaffolds, comprising PLA nanofibres and HA hydrogel, which appear to provide appropriate zone-specific parameters mimicking the microstructural organization and inducing ECM production. The complex and novel combination of multiple imaging, biochemical and biomechanical assessments has revealed that seeded bovine chondrocytes can respond to our scaffolds' spatial orientation and arrangement. Our model with its hybrid zonal-specific arrangement has induced some sort of formation of biomimetic zonal organization and composition of ECM. However, HA gel alone scaffolds only achieved randomly distributed round chondrocytes with homogeneous ECM distribution. Our findings are very encouraging and serve to demonstrate that we might have a model which has the potential to in part recreate native AC.

It has been reported that cytoskeletal morphology and the orientation of the ECM is tightly interrelated [27]. The control of the orientation and the morphology of cells can be used to control the architecture of secreted ECM because the orientation of newly produced ECM follows the cytoskeletal shape. Inversely, artificial ECM with specific patterns on the cell culture substrates can induce the aligned cellular patterns through ‘contact guidance' [28]. Dalby et al. [27] have documented that the ECM architectures and cytoskeletal orientation can be controlled by nanometre-scale structures [28]. Work from our own group [20] has further demonstrated that electrospun nanofibres facilitate the adhesion of chondrocytes and the regulation of their morphology. Herein, we have provided further evidence to show that the aligned and randomly orientated fibre morphology can result in elongated and small clusters of round cell morphology, respectively, in both the superficial and the middle zones. Different from other reports [10–12], our work has generated hydrated and full-thickness constructs with distinct ECM architecture and composition with depth.

The chondrocyte phenotype is defined by a change in morphology and the alteration of key ECM components, including collagen type II and aggrecan [29]. Our data showed that the organization of cell aggregates appeared to regulate the ECM. For example, within the superficial zone, we observed that aligned nanofibres resulted in elongated chondrocyte morphology. The resulting chondrocytes produced collagen I, which is associated with a fibroblast-like phenotype [29]. By contrast, our middle zone scaffold maintained the round shape of chondrocytes, which generally demonstrated higher collagen II production. These data suggest that the orientation of electrospun fibres could be used to regulate the chondrocytes' phenotype. A review of the current literature suggests that one reason for our findings could be that the randomly aligned nanofibre surfaces reduced the extent of cell spreading [30], allowing multiple focal adhesions to be established [20].

Hydrogels have high water content, tuneable physical properties, homogeneous cell distribution and a high permeability for nutrients and waste products of metabolism [31]. Thus hydrogels are ideal components to mimic the chemical and physical environments of the ECM and, therefore, are ideal cellular microenvironments for cell proliferation and differentiation [32]. One requisite of hydrogel scaffolds for cartilage regeneration is to maintain chondrogenesis (cartilage growth). HA, found in native cartilage, has been researched for decades [33]. In situ, AC chondrocytes exhibit a rounded morphology and cluster in small groups [34], where they interact with HA via various surface receptors. The interaction with HA triggers a sophisticated signalling pathway, allowing chondrocytes to retain their original morphology and phenotype, an area that is still not fully understood [33]. In the current study, we used crosslinked and reconstructed HA gel as the stable hydrate environment for chondrocytes. Our data illustrate that overall HA can be used to control both ECM production and chondrocyte morphology when combining local cellular morphological control but it is still a long way from being fine-tuned.

The mechanical properties of the constructs in this study were inferior to the native counterparts. Increase of culture duration, HA concentration in the constructs and application of mechanical conditions for the constructs which accelerate the ECM production are possible approaches to improve the mechanical properties of the constructs before they become clinically useful products.

This study adapted a smart hybrid and sandwich-style fabrication for zonally distinct constructs, which separately and synergistically has the potential to regulate the chondrocytes' phenotype. In this way, we have shown that our model can enable chondrocyte elongation (superficial zone) or chondrocyte aggregation (random clusters in the middle zone and columnar clusters in the deep zone). The underlying HA hydrogel restricted cells from attaching, which drove cells to adhere and orientate along the nanofibre meshes, while the overlying HA hydrogel (reconstructed HA gel) provided the necessary support to stabilize the orientation of the chondrocytes and nanofibre meshes. Altogether, combining a highly hydrated hydrogel, which has low protein affinity, with nanofibres, which have high cell attachment capacity, has the potential to lead to a clinically beneficial product in cell therapy for cartilage treatment. Hence, adaptation of the two-level control strategies in scaffold fabrication has the potential to create multiple zone cartilage regeneration.

5. Conclusion

This feasibility study confirmed that PLA nanofibre meshes with different alignments and micro-channels in an HA hydrogel can be used to create a hybrid scaffold model. Our model has been demonstrated to (1) induce distinguished chondrocyte alignment and (2) encourage ECM production within AC zones. The combination of the regulation of a chondrocyte's aggregation state and skeleton morphology by nanofibre and micro-channels and the maintenance of these morphologies in a highly hydrated gel can be used as a facile technique to replicate zonal-specific cartilage constructs. The three zonal constructs can be assembled and kept intact during culture. Thus, this study presents new hybrid scaffolds and a simple method for biomimetic cartilage regeneration.

Data accessibility

All experimental data have been included in the manuscript. Data for the control sample assessment are included in the electronic supplementary material.

Authors' contributions

H.A.O. carried out the laboratory work, participated in data analysis and drafted the manuscript; R.Y., L.C. and N.J.K. participated in data analysis and manuscript correction; Y.Y. conceived, designed and coordinated the study. All authors gave final approval for publication.

Competing interests

The authors declare no competing financial interest.

Funding

The authors are grateful for the British Council Newton Fund: PhD Placement grant for H.A.O. and the ‘Open Funding Project of the State Key Laboratory of Bioreactor Engineering’ funding from ECUST.