Drug/Bioactive Eluting Chitosan Composite Foams for Osteochondral Tissue Engineering

Joint defects associated with a variety of etiologies often extend deep into the subchondral bone leading to functional impairment and joint immobility, and it is a very challenging task to regenerate the bone-cartilage interface offering significant opportunities for biomaterial-based interventions to improve the quality of life of patients. Herein drug-/bioactive-loaded porous tissue scaffolds incorporating nano-hydroxyapatite (nHAp), chitosan (CS) and either hydroxypropyl methylcellulose (HPMC) or Bombyx mori silk fibroin (SF) are fabricated through freeze drying method as subchondral bone substitute. A combination of spectroscopy and microscopy (Fourier transform infrared (FTIR) spectroscopy, scanning electron microscopy (SEM), X-ray diffraction (XRD), energy dispersive X-ray (EDX), and X-ray fluorescence (XRF) were used to analyze the structure of the porous biomaterials. The compressive mechanical properties of these scaffolds are biomimetic of cancellous bone tissues and capable of releasing drugs/bioactives (exemplified with triamcinolone acetonide, TA, or transforming growth factor-β1, TGF-β1, respectively) over a period of days. Mouse preosteoblast MC3T3-E1 cells were observed to adhere and proliferate on the tissue scaffolds as confirmed by the cell attachment, live-dead assay and alamarBlue TM assay. Interestingly, RT-qPCR analysis showed that the TA downregulated inflammatory biomarkers and upregulated the bone-specific biomarkers, suggesting such tissue scaffolds have long-term potential for clinical application.

(osteonecrosis, osteochondritis dissecans etc.) and often associated with swelling, pain and stiffness of the whole joint [1].The progressive deterioration of the defect leads to the involvement of subchondral bone where the existence of an interface (osteochondral interface) between the two dissimilar tissues becomes the main hurdle in its healing/treatment.A variety of conservative treatments (commonly pain management medication) and surgical treatments exist (ranging from the orthobiologics to advanced arthroscopic techniques) to treat joint defects [2,3].Several studies have been conducted to compare various treatments and understand which is the best treatment option, however, failure in treatment was observed in many patients at later stages of life [4,5], and there is therefore significant opportunity to improve long-term clinical outcomes [6][7][8][9].
The development of composite biomaterials for bone tissue engineering is an active area of biomaterials science and engineering [34,35,[40][41][42][43]. Polymers are widely employed in biomaterials due to their useful properties, for example, CS has antibacterial properties [44], and B. mori SF has pro-regenerative properties [45,46], moreover both polymers can be processed into various morphologies like films, fibers, hydrogels, and foams/sponges alone and as composite materials with mechanics mimetic of the biological niche in which they may be used [47][48][49].Such biomaterials may offer the biological/environmental factors that are instrumental to successful bone tissue engineering, with long term prospects for clinical application [50][51][52][53].It is now well established that regeneration of the bottom layer is more important than the cartilage layer alone hence layered scaffolds have also gained attention to facilitate the regeneration of interfaces where different tissue types having different physiology are connected.
Inflammation at defect sites may facilitate healing by generating a pro-regenerative response, however, severe inflammatory responses arising after surgery may hinder/halt the regeneration, consequently, it is important to effectively manage inflammatory responses in tissue engineering [54][55][56].Joint defects typically present elevated levels of tumor necrosis factor-α (TNF-α) and interleukins with inflamed synovium that downregulates extracellular matrix (ECM) production and the anabolic activity of the cells [57][58][59].To reduce inflammation in joints, triamcinolone acetonide (TA) can be administered [60][61][62][63][64]; and site-specific delivery of drugs from biomaterials circumvents systemic pharmacokinetic challenges and off-target effects [65].Scaffolds loaded with drugs/bioactives provide dual functionality in terms of local delivery of the payload in the extracellular environment, as well as serving as a support for the cells to organize in a 3D fashion [66][67][68].Here we report the development of novel composite biomaterials loaded with either TA or TGF-β1, where the elution of the drug/bioactive payload from the scaffolds is intended to control the inflammatory response and enhance the osteogenic response, using the mouse MC3T3 preosteoblast cell line for in vitro validation of the approach.Trypan blue was purchased from ACI Chemicals Ltd. (Dhaka, Bangladesh).All chemicals were either used as received and stored in line with the supplier's recommendations.

Preparation of porous composite foam tissue scaffolds
Porous composite foam tissue scaffolds with variable amounts of the polymers (i.e., SF, CS, HPMC) and nHAp were prepared by freeze drying (an adaptation of our previously reported protocol [69].Briefly, a 2% acetic acid solution was used to dissolve CS (typically taking 4-6h) at room temp, yielding a 33 wt% aqueous CS solution; nHAp was added to this solution with the help of an ultrasonicator to form a uniformly dispersed slurry.A solution of SF denaturant was prepared by dissolution of anhydrous calcium chloride (111 g, 1 mole) in distilled water (144 g, 144 mL, 8 moles), and to this solution ethanol (92.1 g, 116.8 mL, 2 moles) was added; 50 mL of this solution was used to dissolve 15 g of SF by stirring at 80°C for 1 hr., followed by filtration over a glass frit and dialysis (MWCO 12kDa) against concentrated PEG solution for 48 hrs., in accordance with the literature [70].Optionally both the CS and SF solutions were mixed as required and crosslinking was induced as a result of hydrogen bonding by the addition of HPMC, along with: unloaded TA, TA loaded polycaprolactone microspheres (prepared according to a reported literature protocol for providing prolong release of the drug from scaffolds [71], drug loading in polycaprolactone microspheres was determined by UV-vis spectroscopy, and observed to be 28% by weight; on the basis of this data, the microspheres were added to the scaffolds (i.e., 5% drug and 18% PCL by weight; sample B3 in Table 1)) and/or TGF-β1; the slurries were transferred into molds and stored overnight at -40°C prior to lyophilization for 48h to obtain the porous composite tissue scaffolds.The precise compositions of the biomaterials in their dry state are shown in Table 1.The control composite scaffolds (B1) were fabricated using the same preparation method without the addition of drugs.The addition of microspheres into the scaffolds was also performed by a post seeding method rather than in situ loading, where the microspheres were dispersed in deionized water and added to the surface of tissue scaffolds where both gravitational force and flow of liquid across the pores led to the distribution of microspheres on the polymeric surfaces inside the tissue scaffolds.

Scanning electron microscopy (SEM) and energy dispersive X-ray (EDX)
spectroscopy SEM images and EDX spectra were recorded using a Tescan VEGA3 equipped with EDX (Brno -Kohoutovice, Czech Republic).Rectangular blocks of the scaffolds after gold sputtering were analyzed for microstructure and elemental distribution, and further processed using Fiji ImageJ.

Fourier transform infrared (FTIR) spectroscopy
FTIR spectra were recorded using ThermoFisher scientific FTIR (Nicolet-6700 TM , USA) coupled with an attenuated total reflectance (ATR) accessory in the spectral region of 650-4000 cm -1 at room temperature.The analysis was carried out at a resolution of 8 cm -1 accumulating a total of 128 number of scans per sample.

X-ray diffraction (XRD) analysis
X-ray diffractograms were collected using a Rigaku Smartlab powder diffractometer (Rigaku Ltd., Kent, UK) equipped with a DTex250 one-dimensional (1D) detector, irradiating the films at a wavelength of 0.15418 nm, from Cu K radiation.The Cu source was operated at 45 kV and 2.6.X-ray fluorescence (XRF) analysis X-ray fluorescence measurements were performed on a Shimadzu EDX-8000 instrument.The samples were gently crushed using a stainless-steel spatula and placed on a thin film of Mylar, and irradiated with a beam of 0.5mm radius.The counting time was 180 seconds and the elemental intensities were analyzed using the Fundamental Parameter (FP) method.No balance medium was used in the analysis, and so only elements above 0.5 weight % in the final output are considered.
Interpretation of these weight % is only qualitative but confirms the presence of Ca and P, with possible contaminants of Si and S.However, it is noted that Si and S are often contaminants of the Mylar film used to mount the samples.

Swelling ratio analysis
The swelling ratio of a tissue scaffold (i.e., its capacity to absorb water after immersion in PBS solution (pH 7.4) at 37 ± 0.5°C) for different time intervals was recorded (n=3).The dry weight (Wd) of the scaffolds was recorded before immersion of the scaffolds in PBS after which it was stored, then blotted on paper before recording the wet (Ws) weight of the blotted samples.The % water uptake was calculated using the following formula:

Degradation testing via mass loss
Degradation of the composite scaffolds was determined by immersion of a pre-weighed sample (n=3) (W0) in PBS (pH 7.4) maintained at 37°C for various time intervals.At the end of every time point the samples were removed from the medium, dried in an oven at 65°C overnight and their weight was again recorded as W1.The percent degradation was then calculated using the following formula:

Porosity testing
A liquid displacement method employing ethanol as the displacement solvent was used to find out the porosity of composite scaffolds.For this purpose, a pre-weighed sample was immersed in a known volume (V1) of ethanol in a graduated cylinder (n=3).The final volume with scaffold was noted as V2 and after removal of the sample the final volume was recorded as V3.The porosity was calculated using the following formula:

Mechanical testing
Mechanical testing of samples (n=3) was carried out at room temperature in accordance with the ASTM D5024-95a guidelines.An electrodynamic fatigue testing machine (LFV-E 1.5kN) with a load cell of 1.5kN was used to compress the rectangular scaffold blocks having dimensions of (5x3x3mm) at a strain rate of 1x10 -3 /s, and crosshead speed between 0.2-0.35mm/min up to 25% deformation.The obtained results were used to calculate mean elastic modulus and universal compressive strength.

In vitro drug release studies and release kinetics
To carry out the drug release studies, scaffolds measuring an equal amount of drug were added into 10 mL PBS solution maintained at 37 ± 0.5°C in a water bath (n=3).At predetermined time intervals an aliquot volume approximating 1 mL was withdrawn, filtered and analyzed to quantify the amount of released drug, i.e., TA at a wavelength of 236 nm using UV-vis spectrophotometer and TGF-β1 at 490 nm using ELIZA kit following the manufacturer's specification with the help of a microplate reader (Bio-Rad PR4100).After each withdrawal an equal volume of pre-warmed fresh PBS was added/replenished to keep the sink conditions constant.From the data obtained, the cumulative drug release was then calculated.Afterward the data was fitted into various kinetic models to study the release pattern from the matrices.

Cell culture
Mouse MC3T3-E1 pre-osteoblast cells were cultured using fresh medium supplemented with the requisite ingredients.Cells of passage (P=5) were trypsinized with trypsin/EDTA before seeding and incubated at 37°C in a 5% CO2 incubator.Cells from fresh culture were seeded on presterilized scaffolds (Gamma radiated at a dose of 2.5 K Grey) (n=3) and incubated in a 5% CO2 incubator.

Cell attachment assay
To study the behavior of cells towards the scaffolds an adhesion/attachment assay was performed.
Exponentially growing cells at a density of 2x10 4 cells/well were seeded on the scaffolds in a 24 well plate.The samples after incubation for 3-days in a 5% CO2 incubator at 37°C were washed thoroughly with cell culture grade PBS and fixed with 4% PFA solution for 30 min, followed by dehydration using a serial dilution (50%, 60%, 70%, 80%, 90% and 100%) of ethanol.The samples after air drying were analyzed via SEM.

alamarBlue TM assay
A nondestructive alamarBlue TM assay was carried out to quantitatively measure cell viability.
Exponentially growing cells at a density of 2x10 4 cells per well were aseptically seeded onto the scaffolds in a 24 well plate and incubated at 37°C for specified time points (i.e., 2 & 8 days) in a 5% CO2 incubator.200 µL of alamarBlue TM working standard was added to each well containing in the dark inside a 5% CO2 incubator.Bio-Rad PR4100 absorbance microplate reader was then used to measure absorbance at a wavelength of 550 and 620nm.Cells grown on tissue culture plates were used as a positive control.

Live/dead assay
Samples incubated with MC3T3 preosteoblast cells at 37°C in a 24 well plate for 9 days in a 5% CO2 incubator were estimated for number of live and dead cells.The assay was carried out by careful aspiration of the growth medium and incubation of the samples with 3µM Calcein AM and 2.5-5µM propidium iodide in DPBS for 30 min in dark.400 µL of this solution was added into each well of 24 well plate and were observed using fluorescence microscope supplied with green and red filters.The obtained images were further processed in Fiji ImageJ.

RT-qPCR
The gene expression analysis of the TA loaded scaffolds was carried out in accordance to a previously established protocol [72] using real-time quantitative PCR (Qiagen Rotor-Gene ® Q 5Plex HRM) provided with 2.3.5 software from the manufacturer.Briefly, the MC3T3 cells after interaction with the scaffolds for 12h were lysed using Trizol reagent to isolate RNA.The isolated RNA was quantified using nanodrop and used to synthesize cDNA library according to the manufacturer's protocol.The transcription level of four genes including TNF-α, IL-6, RUNX2 & ALP were quantified according to the reference gene GAPDH with the help of SYBR green dye.
The primer sequences of mouse genes used in the study are given in Table 2.

Statistical analysis
Each sample data represents triplicate run (n=3) as the mean ± standard deviation (SD).Analysis was performed using GraphPad Prism-8 software.One-way analysis of variance (ANOVA) statistics followed by Post hoc Dunnett test was used.p-value ≤ 0.05 was considered as statistically significant.

Scaffold preparation and microscopic and compositional analysis
The porous composite tissue scaffolds were prepared by a simple, scalable freeze-drying method (depicted in Scheme 1), with their compositions clarified in Table 1.Surface characteristics of a biomaterials play an important role in their interactions with the external environment.The SEM images of the porous composite scaffolds as displayed in Figure 1 show the pore walls to have both smooth and rough surfaces due to the presence of the various polymeric/particulate components in the composites in line with the literature [73].The mean pore diameter of the scaffold was found to be 64 ± 1.06 µm.An interconnected and stratified 3D porous structure with both micropores and macropores can be seen in the cross-sectional images which is supportive of cell migration/infiltration and nutrient/waste transport/diffusion.The overall porosity of the scaffolds remained ≥ 60% in all samples, in line with the porosity measurement data obtained from liquid infiltration.A uniform distribution of the nHAp crystals in the form of white beads in the pore walls (highlighted by yellow arrow heads) (Figure 1), as confirmed by EDX analysis.
Moreover, uniformly distributed microspheres can be seen embedded in the polymeric matrices (highlighted by blue arrow heads), which suggests that in situ loading of microspheres could be an optimum method as compared to the post seeding method (where the microspheres were not well integrated in the polymeric matrix of the porous scaffolds).Interestingly, the distribution of microspheres inside the scaffolds using the post seeding technique was correlated to the size of the microspheres, and if the size of microspheres was less than pore size they infiltrated the scaffolds otherwise they sat on top of the scaffolds [74].It was found that the type and ratio of polymers affects the surface characteristics, i.e., the SF based scaffolds (Figure 1: B4, B7 & B8) had a smooth surface, whereas the CS-based scaffolds (Figure 1: B1-B3, B5 & B6) had rough surfaces which is in line with the literature [75].
Elemental analysis of the composite scaffolds was assessed via EDX analysis (Figure 2).Peaks at values of 0.3, 0.38, 0.6, 2.1 and 3.7 KeV were observed for both CS-based and SF-based tissue scaffolds.The peaks both at 0.38 and 3.7 KeV are characteristic of Ca 2+ , and the peak at 2.1 KeV is characteristic of P in PO4 3-, thereby confirming the nHAp is present in the scaffolds in line with the literature [76].The peaks at 0.3 and 0.6 are the representative peaks of C and O which are present in the polymeric components (CS/SF/HPMC) in the samples.In case of SF-based tissue scaffolds another peak pointing at 1.15 KeV is characteristic of Na as a trace residue from PBS.
Complementary elemental analysis via XRF (Table 1) confirmed the presence of Ca and P in all samples.

FTIR analysis
The FTIR spectra of unloaded, TA-loaded and TGF-β1-loaded scaffolds (B1-B8) are shown in Figure 3 whereas the individual components are shown in Figure 4.The spectra of B1-B3 shows all the characteristic peaks for CS and/or HPMC; i.e., the peak at 1046 cm -1 is characteristic of the C-O-C stretching of glucosamine residues, whereas the peaks at 1420 cm -1 , 1550 cm -1 and 1680 cm -1 are characteristic of the C=O stretching of the amide bond and N-H of the amino group of CS [77].As both CS and HPMC are polysaccharides and have similar structures, the bands overlap each other.A slight change in peak intensity and position is characteristic of new non-covalent interactions, most likely hydrogen bonds [78], and peaks between 1420 cm -1 and 1680 cm -1 in case of B2 and B3 may arise from the drug (TA).
In the SF-based scaffolds, the absorptions at 1611 cm -1 , 1507 cm -1 and 1224 cm -1 arise from the amide I (C=O stretching), amide II (C≡N stretching) and amide III bands, respectively.A slight change in peak intensity and shift from 1611 cm -1 to 1620 cm -1 for the amide I (characteristic of the presence of β-sheets), and 1224 cm -1 to 1240 cm -1 for the amide III were seen in the scaffolds B4, B7 and B8.These slight changes may be a result of the intermolecular interaction of the polypeptide chain with the solvent altering the degree of crystallinity and/or intermolecular/intramolecular hydrogen bonding between polypeptide chains (e.g., β-sheet formation, however, complex composites containing more than one biopolymer with amides makes this difficult to quantify).The nHAp exhibit typical non-degenerating symmetric stretching of P=O bond at 950 cm -1 , and degenerated asymmetric stretching of P-O bond appears at 1060 cm - 1 and 1015 cm -1 , respectively [79].The appearance of peaks between 1338 cm -1 and 1473 cm -1 can also be seen in the scaffolds B4, B7 and B8.The broad peak in the range of 3000 cm -1 -3600 cm - 1 was common in almost all scaffolds which corresponds to the overlapping of free hydroxyl group and stretching vibrations of N-H of amino group from CS and HPMC.It can be seen that the addition of TGF-β1 into these scaffolds had no observable effect on the scaffolds.

XRD analysis
Figure 5 displays the XRD patterns of the various scaffolds.XRD data for samples with CS showed characteristic broad peaks at ca. 2θ ≈ 10° and 20°; XRD data for samples with HPMC showed the characteristic broad peak at ca. 2θ ≈ 20°; XRD data for samples with SF were largely amorphous (broad peak at 2θ ≈ 20°), with evidence of crystalline content (peaks at 2θ = 19.9°and 24.0°) suggesting β-sheet content in line with FTIR measurements; the characteristic peaks for nHAp (at 2θ = 31.8°,32.6°, 32.9° and 33.1°) are evident in most XRD patterns albeit somewhat obscured in the samples containing SF.

Swelling analysis
Ability of biomaterials to absorb water and retain its structure is important for its function as a tissue scaffold.Swelling depends on the internal crosslinking of the polymers and the greater the degree of crosslinking, the less swelling, and vice versa.Figure 6 shows the swelling behavior of all scaffolds after a 24 h period in PBS.It can be seen that the SF-based scaffolds i.e., B4 and B8 exhibited less swelling (i.e., 4.73 ± 0.03% and 4.54 ± 0.31%, respectively), than the CS/HPMCbased scaffolds.This difference in the swelling of SF-based scaffolds is in part due to the intermolecular interactions between the protein chains lacking between the polysaccharides CS and HPMC [80].Another factor affecting water uptake is the degree of porosity, as the porous structure can hold large volumes of liquid within the empty spaces.
In the case of B7, a slight increase (5.37 ± 0.24%) in swelling was seen which can be again attributed to the addition of CS to the SF solution.the β-sheet content of SF influences water solubility, thereby reducing swelling and slowing degradation, while concomitantly increasing mechanical properties [81].A very minor difference in the swelling of B5 and B6 was noted (i.e., 7.23 ± 0.04% and 7.83 ± 0.12%, respectively).The highest values for swelling ratio with an increasing order were noted for B3 < B1 < B2, with values of 8.44 ± 0.38%, 8.55 ± 0.16% and 8.62 ± 0.21%, respectively.Increased liquid uptake facilitates transport of nutrients and waste to/from resident cells, whereas, increased swelling ratio slows cell infiltration within the scaffold.the addition of the ceramic (nHAp) to the scaffolds was observed to reduce the swelling ratio which is in accordance with the literature [82].These results show that SF-based scaffolds exhibit the optimum swelling behavior which is useful for tissue engineering.

Degradation analysis
Degradation assessed by the percentage weight loss over time was analyzed with a view to understanding prospects for successful integration of the scaffolds into native tissues.If a scaffold degrades at a very slow rate that may impede ingrowth of new tissue, whereas if it degrades too quickly it may not offer mechanical support for long enough.Ideally the degradation rate of the scaffold should coincide with the formation of native tissue which is dependent on the type and location of the defect [83].Consequently, time dependent degradation of the scaffolds was analyzed at five different time points (i.e., day 1, day 3, day 7, day 10 and day 14), with representative results for day 1 & 14 shown in Figure 7 (a complete set of data is presented in the supplementary information, Figure S1).Degradation of the scaffolds starts on day 1 and carries on over the course of the experiment.The rate of degradation of B5 was significant (i.e., 5.2 ± 0.3% (p<0.01)) after day 1 when compared to the control B1.Non-significant weight loss as compared to B1 was exhibited by composites B2, B3, B6 and B7 after day 1 (having values of 4.5 ± 0.3%, 3.5 ± 0.3%, 4.1 ± 0.2% and 4.4 ± 0.3%, respectively); significantly less degradation loss was noted for composites B4 and B8 (2.0 ± 0.2% and 2.1 ± 0.4% weight loss, respectively).
Weight loss continued with time for all composites as compared to B1 (control) with different rates.The non-significant difference in degradation of the scaffolds after the addition of TA or TA-loaded microspheres became significant after day 14, but for composite B2 it increased in rate, and for composite B3 it decreased in rate, relative to the control.For all other composites a similar trend was seen after each time point.The percent degradation of the composites after 14 days of incubation in the medium in increasing order was B4 < B8 < B7 < B3 < B1 < B6 <B2 < B5 (with experimental values 10.1 ± 0.3%, 10.4 ± 0.4%, 13.6 ± 0.2%, 16.9 ± 0.4%, 18.7 ± 0.4%, 19.8 ± 0.4%, 20.0 ± 0.2% and 20.7 ± 0.2%, respectively).It can be seen that the CS-based scaffolds exhibited a total weight loss of around 20%, whereas SF-based scaffolds showed ca.10% weight loss, suggesting the SF-based scaffolds would degrade more slowly than CS-based scaffolds.

Porosity analysis
Porosity is very important as the highly porous structure facilitates cells infiltration which in turn compromises the mechanical properties of the scaffold.It has been previously suggested that a scaffold for bone tissue regeneration should have a porosity of more than 60% [24,84,85].
Scaffold porosity determined by solvent displacement (Figure 8) shows the composites have similar porosities.The unloaded scaffold (B1) used as a control had a porosity of 65.7 ± 2.1%.
Addition of drug to the composite (B2) has resulted in increased porosity with value of 68.8 ± 1.3%, whereas addition of TA-loaded microspheres to composite (B3) resulted in non-significant decrease (p>0.05) in porosity as compared to the control (62.0 ± 1.0%).
The addition of TGF-β1 to the scaffolds has no effect on the porosity as can be seen from the porosity data of composite B1, B5 and B6 with values 65.8 ± 2.1%, 66.8 ± 3.6 and 65.6 ± 0.6% respectively.A significant increase in porosity was observed for composites B4 and B8 (80.3 ± 1.8% (p<0.0001) and 77.0 ± 1.4% (p<0.001),respectively).The highly porous structure is beneficial for tissue engineering in terms of cell transplantation and cultivation [86], and the presence of SF encourages highly porous structures through freeze drying (which plays a role in β-sheet formation) [87].Porosity is linked to the mechanical properties of materials, however, the irregular pore shapes/sizes in the composite scaffolds described herein make it illogical to attempt to draw conclusions about correlation between porosity and mechanics which are important for their success [88].

Mechanical properties
Mechanical properties play a pivotal role in the effective design of tissue scaffolds for tissue regeneration, wherein biomaterials with biomimetic properties are desirable [89], composite materials can be prepared where the ratio of components control the mechanical properties of the biomaterials [90,91].The compressive properties of the biomaterials studied herein were observed to be of a similar order of magnitude (low MPa regime) to natural human tissue (e.g., soft bone tissue, cartilage tissue, etc. [92,93] ) as shown in Figure 9.The elastic modulus of B1, B2 and B3 was found to be 3.17 ± 0.04 MPa, 3.11 ± 0.09 MPa and 3.70 ± 0.28 MPa, respectively.The compressive strength of B1 was 0.21 ± 0.02 MPa and B2 was 0.27 ± 0.02 MPa, showing a statistically non-significant difference (p≥0.05);however, a significant difference was seen when comparing to B3 and B6 which have values of 0.31 ± 0.03 MPa (p=0.0084) and 0.32 ± 0.03 MPa (p=0.0063),respectively.Both the elastic modulus and compressive strength was significantly higher (p<0.0001)for composites B4, B5, B7 and B8.The elastic modulus values were 4.67 ± 0.19 MPa, 4.84 ± 0.21 MPa, 3.98 ± 0.13 MPa and 4.98 ± 0.12 MPa, whereas the compressive strength values were 0.81 ± 0.06 MPa, 0.83 ± 0.01 MPa, 0.73 ± 0.03 MPa and 0.85 ± 0.05 MPa respectively.Composite B6 was observed to have the most significant increase in both elastic modulus and mechanical strength (3.84 ± 0.24 MPa and 0.32 ± 0.03 MPa (p=0.0063),respectively).
Interestingly, the SF-based tissue scaffolds showed greater increases in mechanical strength than CS-based tissue scaffolds, potentially because of its propensity for intermolecular and intramolecular interactions (e.g., β-sheet formation).

In vitro drug release studies and release kinetics
The cumulative release data of both TA and TGF-β1 from the respectively loaded scaffolds (B2, B3 & B5-B8) shows release from the matrices over the period of 72 hours (Figure 10).The release of TA from B2 is greater than B3, and this is observed to increase over time because of the incorporation of the drug inside microspheres in the case of B3 (consequently the drug has to diffuse through the polymer matrix into the sink medium), the release of the drug followed a concentration dependent mechanism and was affected by the degradation of the polymer along with the diffusion mechanism; the composites released 5.2 ± 1.3% and 2.7 ± 1.8% of the drug, respectively, after 2h in the release medium.A clear decline in the release rate of both composites can be seen after 12 h and the bend in curve has continued towards x-axis along with time; the composites released 70.6 ± 4.9% and 52.4 ± 2.3% of drug after 72 h, respectively.
By comparison with the release of TA, the release of TGF-β1 from composites (B5-B7) was slower, with a gradual increase after 24 h incubation.The only composite that has shown a slow and prolonged release among the TGF-β1-loaded scaffolds was B8.The mathematical values of cumulative TGF-β1 released from composites B5, B6, B7 and B8 was 23.19 ± 0.99%, 20.38 ± 0.88%, 19.20 ± 0.67% and 15.13 ± 1.88%, respectively.The cumulative release after 72 h was in the order B8 (40.28 ± 0.60%) < B7 (67.48 ± 2.02%) < B5 (76.32 ± 1.70%) < B6 (82.42 ± 4.78%), which shows that the SF-based scaffolds have the ability to prolong the release of the drug more than CS/HPMC-based scaffolds (correlating to the degradation studies).A concentration independent release mechanism following both dissolution and diffusion process was exhibited by the composite scaffolds (a complete set of data for drug release kinetics is presented in the supplementary information, Table S1).

Cell attachment assay
Adhesion of the cells on any surface is important for tissue scaffolds, and the attachment behavior of mouse MC3T3 preosteoblast cells after 3 days of culture on the scaffolds was studied (Figure 11).The cytoplasmic extensions from the cells (highlighted by the yellow arrows) secured the cells on the surface of the biomaterials.The attachment of cells only on the surface of a scaffold and forming a scar tissue layer at its surface limits the infiltration of cells within the scaffolds, resulting in rejection of the graft as a mechanically non-functional necrotic tissue; from the SEM results it is evident that the cells have effectively infiltrated the pores in the biomaterials (evidenced by formation of active and randomly distributed cytoskeletal extensions).The blue arrow heads are indicative of areas of the scaffolds with onset/observable degradation over time which is beneficial for the success of an implanted scaffold.No effect of the addition of TA or TGF-β1 on cell adhesion.

alamarBlue TM (Resazurin) assay
The metabolic activity of a cell is the primary marker to estimate its viability.Figure 12 shows bar graphs of cells on scaffolds at two different time points (i.e., at day 2 and day 8) against the positive control (i.e., tissue culture plastic, TCP).On day 2, a significantly reduced viability (p<0.001) can be seen for samples B1, B2, B3 and B6 (83.5 ± 2.0%, 81.9 ± 2.2%, 91.5 ± 1.5% and 84.4 ± 2.6%, respectively); however, the other samples had comparable viability to the control (non-significant difference, p≥0.05, attributed to the presence of SF).By comparison, after 8 days of incubation all of the samples have shown either non-significant (p≥0.05)difference in viability or significantly (p<0.001)improved viability as compared to the control.This clearly shows that the number of viable cells has been increased over time and resulted in enhanced metabolic activity.The scaffolds with significantly improved cell viability were B4, B5 and B8, which had values of 109.9 ± 2.2% (p=0.0101),110.2 ± 2.3% (p=0.0079) and 112.2 ± 3.3% (p=0.0016),respectively.Reduced metabolic activity is an early indicator of poor cell health that may result in compromised cell membrane integrity causing cell death.

Live/dead assay
Figure 13 shows the result of the live/dead assay of the mouse MC3T3 preosteoblast cells cultured on the scaffolds after 3 days of incubation.It is the cell membrane integrity that is used as the base for the Live/Dead assay, rather the metabolic activity of the cells that reduces the resazurin dye.It is evident form the images that viable cells (stained with green color) were present in each sample, whereas very few dead cells (stained red) were observed in the images which clearly shows the biocompatible nature of the developed scaffolds.

RT-qPCR
The severity of inflammation at the defect site strongly depends on the expression or suppression of the inflammatory markers.Herein the mRNA transcript expression level using real-time PCR after 24 h culturing of the samples with MC3T3 mouse preosteoblast cell line was studied, using four targets: TNF-α and IL-6 (inflammatory markers, Figure 14) and RUNX2 and ALP (bone markers, Figure 15).Composites B2 and B3 containing the anti-inflammatory drug (TA) have significantly (p<0.05) down regulated the target proteins.A 10-fold reduction in the expression of TNF-α can be seen in case of B2, compared to B3 where the fold decrease is 6; this difference in gene suppression is because of the incorporation of drug directly into scaffolds in case of B2 as compared to B3 (where the incorporated drug is included in the form of microspheres and the drug needs to elute from the microspheres).Similarly, a 10-fold decrease in the expression of IL-6 can be seen for both B2 and B3, which is significantly (p<0.05)lower than the control (B1).These results are showing a significant difference in the down regulation of target proteins by the composites hence suggesting an anti-inflammatory potential of the scaffolds if implanted in an animal model.
In addition to the anti-inflammatory potential, the osteogenic potential of the same scaffold composites i.e., B2 and B3 was evaluated by quantifying the mRNA transcript level of two bone markers.B2 was observed to have a 10-fold increase, whereas B3 had a 9-fold increase in the upregulation of RUNX2 as compared to the control (B1) which is statistically significant (p<0.05).
There is a small fold increase (i.e., 2.5 and 3-fold) for upregulation of ALP by sample B2 and B3 respectively.

Conclusion
Tissue scaffolds incorporating CS, nHAp and either HPMC or SF were prepared by a simple freeze-drying method, and microscopy and spectroscopy showed the components to be distributed evenly through the matrix of the porous tissue scaffolds.The tissue scaffolds had biomimetic mechanical properties suitable for trabecular/spongy bone tissue.The drug release data showed that loaded drugs and growth factor were released at a controlled rate instead of burst release which might be helpful in reducing the administration of postsurgical medication, and a reduction in release rate can be achieved through incorporation of the drug and/or growth factor inside microspheres.The cytocompatibility of the tissue scaffolds was shown by cell adhesion and viability tests, and PCR data demonstrated that the addition of anti-inflammatory drug can reduce the chances of inflammation at the defect site, and pro-osteogenic activity can be enhanced by the incorporation of both TGF-β1 and TA in the tissue scaffolds.On the basis of the results presented, we conclude such biomaterials have long-term potential for clinical applications in the field of subchondral tissue regeneration.

Future Prospects
The work presented in this study is part of an ongoing project which is going to be extended to the development of bilayered scaffolds for osteochondral regeneration and their in vivo evaluation using a small mammal (e.g., rabbit) model.

Figure 4 .
Figure 4. FTIR spectra of the individual components of the scaffolds.

Figure 6 .
Figure 6.Swelling ratio of the unloaded, TA loaded and TGF-β1 loaded composite scaffolds

Figure 7 .
Figure 7. Degradation behavior of the unloaded, TA loaded and TGF-β1 loaded bottom layer

Figure 8 .
Figure 8. Percent porosity of the unloaded, TA loaded and TGF-β1 loaded composite scaffold

Figure 9 .
Figure 9. Mechanical properties of the unloaded, TA loaded and TGF-β1 loaded bottom layer

Figure 10 .
Figure 10.Cumulative drug release behavior from composite tissue scaffolds.TA-loaded

Figure 11 .
Figure 11.SEM images MC3T3 pre-osteoblast cells on the composite tissue scaffolds.Yellow

Figure 15 .
Figure 15.Osteogenic effect of the TA loaded scaffolds against bone markers RUNX2 (A) and

Table 1 .
Composition of the composite tissue scaffolds.

Table 2 .
Primer sequence used in the RT-qPCR analysis