Collagen nanofibril self-assembly on a natural polymeric material for the osteoinduction of stem cells in vitro and biocompatibility in vivo

Abstract

This manuscript reports the characterization of molecularly self-assembled collagen nanofibers on a natural polymeric microporous structure and their ability to support stem cell differentiation in vitro and host tissue response in vivo. Specifically, cellulose acetate (CA) and poly(lactic acid-co-glycolic acid) (PLGA) produced two different microporous structures that were coated with self-assembled type I collagen (PLGAc and CAc). Though the total content of collagen was similar between the two materials after coating, the material chemistries significantly affected the molecular collagen self-assembly resulting in a more biomimetic nanofibrillar D-banding pattern on CA (mean fiber diameter of 80 nm) and a sheet like coating on PLGA (mean fiber diameter of 150 nm). Human mesenchymal stem cells (hMSCs) cultured on CA and CAc showed a significantly higher degree of in vitro osteoblastic progression, in contrast to PLGA and PLGAc. Furthermore, at 2 weeks post subcutaneous implantation, collagen coated CA materials showed increased matrix cellularization and enhanced biocompatibility. At 12 weeks both CA and CAc showed significantly greater matrix cellularity and immune acceptance compared to PLGA. This work illustrates the role of materials chemistry to dictate nanoscale protein assembly and its effect in terms of in vitro stem cell differentiation and in vivo host immune response. We also have proved that inclusion of nanoscale self-assembled ECM components such as collagen can enhance the stem cell inductive capabilities and biocompatibility of hydrophilic natural polymeric materials, making them viable alternatives to widely used synthetic polymers.

---

Introduction

Approximately half a million individuals suffer from fractures every year and the demand for bone graft procedures is continually increasing. Scaffold based bone tissue engineering (BTE) has made great progress in regenerating lost bone tissue. In BTE the scaffold plays an important role, but in order to achieve complete bone healing, the delivery of cells and/or growth factors¹ and/or small molecules is necessary.² While growth factors like BMP-2 (ref. 3), VEGF,⁴ FGF-2 (ref. 5) have been very effective in inducing tissue responses like bone formation, angiogenesis, and homing and proliferation of stem/progenitor cells, a number of constraints still exist in the delivery of these factors.² For instance, although BMP-2 is very effective in inducing bone formation, ectopic bone formation can also occur at non-target sites that can lead to pathological outcomes.⁶ Another limitation with delivering proteins such as growth factors is the challenge of retaining their biological activity during the delivery process. On the other hand, small molecules are easier to handle and pose less threat of denaturation or loss of function during processing, but may induce many different cellular signaling pathways and targeting a specific therapeutic outcome without side effects is a challenge.⁷ However, the delivery of autologous bone marrow derived stem cells is a simple and clinically relevant strategy to improve the outcome of bone healing.^8,9 It has been demonstrated that bone marrow derived mesenchymal stem cells (MSCs) are capable of differentiating in the osteochondral lineage.¹⁰

The physio-chemical nature of biomaterials can dictate the differentiation of stem cells into a specific lineage. Therefore, apart from serving as a delivery vehicle for cells, biomaterials also act as instructive guiding systems to direct stem cell differentiation for tissue regeneration. The various properties of the biomaterials that influence the differentiation of MSCs¹¹ into osteoblastic lineage include the strong mechanical properties of the matrix,¹² the presence of chemical functionalities,¹³ like –NH₂, –OH, –COOH and –CH₃ groups¹⁴ on the surface of the scaffold, optimal hydration properties,¹⁵ surface topography with micro and nanoscale roughness^16,17 and the presence of nanofibers.¹⁸ Many studies have used these mechanical and physio-chemical properties alone to differentiate stem cells in the osteochondral lineage. Though these studies have systematically examined each of the factors individually, it is well known that natural polymers like polysaccharides produced in biological systems inherently possess several of these features, such as favorable hydrophilicity, chemical functionalities and mechanical strength.^19–21 Hence these polysaccharides could serve as ideal biomaterials for bone.

In engineering a scaffold for BTE, it is crucial to achieve a mechanically strong, porous, three dimensional structure that allows the native cells of the surrounding tissues to infiltrate into the materials, as well as inducing stem cells to differentiate into a specific lineage. A sintered microsphere scaffold serves the needs of mechanical strength and material transfer 33 to 40% porosity is usually observed, in a 3D environment.^22,23 However, most bodily extracellular matrix (ECM) components are presented as nanofibers.¹⁸ A number of studies have shown the biological efficacy of nanofiber matrices in inducing a favorable tissue healing response.^24,25

Cells adhere to biomaterials through their integrin receptors. The types of integrin ligands (proteins) present on a biomaterial surface and the conformation of these ligands in turn dictate the cellular responses to the biomaterial.²⁶ The various responses of cell survival, adhesion, and differentiation on biomaterials are in turn driven by the integrin mediated signaling in the cells.²⁷ It has been shown in the literature that protein assembly on hydrophilic and polar surfaces may be more biomimetic than on hydrophobic, nonpolar surfaces.²⁸ Given the favorable characteristics of polysaccharides, as discussed above, we hypothesized that polysaccharide material chemistry would be conducive for the favorable presentation of integrin ligands, such as exogenously infused collagen. We also hypothesized that this biomimetic polymer system containing polysaccharide and protein nanofibers presented in a native conformation would induce the osteogenic differentiation of seeded stem cells and would be more biocompatible in the body in comparison to a model synthetic polymer (PLGA).

Taking the discussed factors into consideration, we formulated a two-component system made of a base of 3D porous sintered microsphere CA scaffold, further infused with collagen nanofibers.²² The principle of protein self-assembly was employed to achieve the formation of biomimetic collagen nanofibers on these 3D porous structures of polysaccharide, CA.²⁹ The synthetic polymer PLGA was used as the control material due to its widespread application in BTE.^30–32 PLGA has also been FDA approved for several applications such as drug delivery and suture formulations.¹ The PLGA materials were formulated into 3D porous sintered microspheres with similar dimensions as CA and likewise functionalized with collagen nanofibers. We have previously shown the feasibility of the formulation of cellulose acetate (CA) sintered 3D porous microstructures and shown their mechanical strength to be in the mid-range of human trabecular bone.³³ We also demonstrated the feasibility of coating these scaffolds with nanofibrillar collagen (CAc) and examined the mechanical properties, pore properties and the retainment of collagen in vitro on the materials over time. We have also previously shown the ability of CA and CAc to maintain the phenotype of cultured human osteoblasts in vitro.²²

In this work we examined the biological performance of cellulose acetate (CA) microstructure and cellulose acetate–collagen (CAc) micro-nano structures, in terms of their ability to dictate stem cell differentiation in vitro and their biocompatibility upon implantation. A well-established synthetic polymer poly(lactide-co-glycolide) (PLGA) microstructure and its collagen infused PLGAc micro-nano structures with similar pore properties served as controls. The effect of collagen coating on both CA and PLGA in terms of collagen content, morphology and material properties were also evaluated. Microspheres of CA and PLGA were formed using a standard oil-in-water emulsion system per our previous publications.^22,33 Microspheres of the selected size range for both polymers were sintered separately into porous microstructures with identical pore properties according to our published protocols.^22,33 These microstructures were incubated in a collagen solution to functionalize with collagen nanofibers.²² Scheme 1a illustrates the process of sintered microstructure fabrication of CA/PLGA, as well as collagen nanofiber functionalization. The red circle depicts an area of interest for the magnified view of the nanofibers. We further seeded these materials with human bone marrow derived mesenchymal stem cells (hMSCs) to observe the effect of the matrices on osteoblastic differentiation of the progenitor cells (Scheme 1b). To further examine the biocompatibility of the materials, we implanted them under the skin of rats and measured the immune outcomes over time (Scheme 1c). Our results indicate the greater potential of cellulose–collagen micro-nano structures to induce osteoblastic differentiation of progenitor cells and show greater biological compatibility of these structures over time.

Scheme 1 A schematic illustration of the preparation of micro-nano structured scaffolds and the examination of their osteoinductive properties in vitro and in vivo biocompatibility. (a) Cellulose acetate (CA) and poly(lactic-co-glycolic) acid (PLGA) microspheres were prepared by solvent evaporation from an oil in water emulsion of the polymer in a 1% PVA solution. The microspheres were sintered into three-dimensional (3D) scaffolds using a solvent/non-solvent combination in the case of CA and pressure and heat in the case of PLGA. These scaffolds were then incubated with collagen solution. Evaporation of the solution led to the self-assembly of collagen nanofibers. Scaffolds infused with collagen were exposed to a short wavelength UV light for twenty minutes on each side. This exposure may have caused partial cross-linking of the collagen nanofibers at the surface level. The red circle on the scaffolds shows a region of interest of the scaffold imaged to characterize the morphology of the self-assembled collagen nanofibers. (b) The ability of micro and micro-nano structured scaffolds to aid in differentiation of human mesenchymal stem cells (hMSC) into osteoblasts was examined under inductive conditions. (c) Subcutaneously implanted scaffolds in a rat were examined over time for tissue responses to determine the biocompatibility of the materials. Cellulose acetate (CA), poly(lactic-co-glycolic) acid (PLGA), cellulose acetate–collagen (CAc), poly(lactic-co-glycolic) acid-collagen (PLGAc), polyvinyl alcohol (PVA).

Results and discussion

Synthesis of the materials and experimental approach

Microspheres were fabricated by the standard oil in water emulsion/solvent evaporation method reported previously.^22,33 Microspheres of CA in the particle size range of 600–800 μm were filled in a Teflon mould and sintered together with the application of solvent/non-solvent combination as reported previously.^22,33 Molds filled with PLGA microspheres were sintered together at 90 °C for 2 h in an oven to obtain micro-porous scaffolds.^22,33 These micro-porous CA and PLGA scaffolds were incubated in collagen solution (0.1% weight/volume, pH 4.2) at 37 °C to allow collagen self-assembly. The process of microstructure fabrication and the infusion of collagen nanofibers is illustrated in Scheme 1a. The effect of collagen treatment (pH 4.2) on the CA/PLGA microsphere morphology during the coating process was studied using scanning electron microscopy (SEM). The red circles on the image present an area of interest view of the scaffold surface.

The ability of these micro-nano structures to induce the osteogenic differentiation of stem cells was examined by culturing human bone marrow derived mesenchymal stem cells (hMSCs) in osteogenic induction media on the materials for three weeks. The cell-seeded structures were evaluated for the degree of osteoblastic phenotype progression (Scheme 1b). Finally, the biocompatibility of the material was assessed using a subcutaneous rat model (Scheme 1c).

Characterization of materials properties

To investigate the potential changes to the material chemistry following acidic collagen solution treatment, scaffolds were examined with attenuated total reflection-Fourier transform infrared spectroscopy (ATR-FTIR). The control samples were CA/PLGA incubated in water (CA, PLGA), buffer without collagen (CA-buffer, PLGA-buffer), and collagen solution (CAc, PLGAc) which served as the test groups. The characteristic PLGA specific group frequencies, namely the alkyl (2800–2900 cm⁻¹), C–O ester (1050–1450 cm⁻¹) and C=O stretch (1750 cm⁻¹) remained unaltered. Noticeable changes were seen around 1600 cm⁻¹ and 3300 cm⁻¹ following treatment with the acidic collagen solution due to the cleavage of the ester bonds. The increase in band intensity around 1600 cm⁻¹ in PLGA is due to the cleavage of the ester bond resulting in carboxylic acid salts (O=C–ONa). Likewise, the increase in band intensity around 3300 cm⁻¹ is attributed to an increase in hydroxyl content as a result of ester bond cleavage.^34,35 The XRD spectra showed a steeper peak of the 2θ angle at 20° due to the increased PLGA crystallinity following acidic buffer treatment (Fig. S1, ESI†). Increased matrix crystallinity may be due to the leaching of amorphous oligomers, as well as PGA degradation, which leads to a proportional increase in the PLA content in the copolymer.³⁶ These observations are well supported by the literature, where PLGA degradation is influenced by factors such as temperature³⁷ and pH.^38,39 For instance, exposure of PLGA to both alkaline⁴⁰ and acidic⁴¹ environments, cleaves the ester bonds and results in carboxyl and hydroxyl functionalities. The presence of an acidic environment has been shown to catalyze the cleavage of the ester bond and degradation of PLGA,^42,43 resulting in matrix erosion.³⁸ It is also important to note that the crystalline oligomers present in the matrix fail to diffuse out due to their decreased solubility in acidic pH. Zolnik et al.⁴¹ reported a slower rate of PLGA microsphere degradation at a pH of 7.4 in contrast to the degree of degradation at pH 2.4, which is in line with our findings. It was seen that the mechanism of degradation varied with the change in pH. While the degradation of PLGA microspheres proceeded with pitting on the surface of the microsphere at neutral or alkaline pH, the surface of microspheres showed a smooth morphology during degradation at an acidic pH. However, the microspheres fractured much more readily under an acidic environment. It was seen that the oligomeric degradation products remained inside the microspheres at low pH conditions, but diffuse out in both alkaline and neutral pH due to their increased solubility.⁴⁴ In the current study, acidic pH exposure of PLGA in the coating process resulted in changes to the materials chemistry and moderately increased the crystallinity, but the microsphere morphology remained unaltered (Fig. 1a); a phenomenon consistent with the results of Zolnik et al.,⁴¹ who observed the smooth surface morphology of PLGA microspheres degraded in an acidic environment.

Fig. 1 Collagen nanofiber characterization. SEM images of (a) PLGA, PLGA-buffer treated and PLGAc, CA, CA-buffer treated and CAc scaffold interiors. Scale bar on top = 1 mm, middle = 50 μm, and bottom = 5 μm. (b) Percentage of total collagen on scaffolds after coating (coating efficiency), n = 4, an unpaired t-test was performed and no statistical significance was seen. (c) Diameter of collagen fibers on PLGAc and CAc at magnifications of 50 000×, n = 10, unpaired t-test performed and a *P < 0.001 was obtained. (d) SEM images of collagen nanofibers on PLGAc and CAc at 50 000× (D-band-like patterns indicated by arrows on CAc), scalebar = 2 μm and (e) second harmonic signals from collagen fibers on PLGAc and CAc at 10× (blue color indicated by arrows), scalebar = 100 μm.

Cellulose acetate also deteriorates upon exposure to both alkaline and acidic pH, through deacetylation of the acetate group. Alkaline pH as low as 0.01 N of NaOH is effective in altering the CA matrix strength, while a much higher HCl concentration and prolonged exposure is needed to observe similar degree of deterioration under acid conditions.⁴⁵ Enhanced deterioration at low pH is due to the deacetylation of the CA matrix, which remains stable at low pH. In the present study, CA specific group frequencies, namely ether (1040 cm⁻¹), acetyl ester (1220 cm⁻¹) and carbonyl (1740 cm⁻¹) groups remained unaffected following acidic collagen solution treatment,⁴⁵ indicating no considerable deacetylation. The presence of an additional band around 1600 cm⁻¹ on the CA surface following treatment was assigned to carboxylic acid salts from the buffer used for functionalization. Following treatment, CA became relatively less crystalline, as evidenced through the decrease in the steep peak at a 2θ angle of 20° in the XRD pattern (Fig. S2, ESI†). The highly amorphous nature of the polymer and the absence of crystalline peaks make it difficult to compute the accompanying changes in material crystallinity. Treatment of microspheres with the acidic collagen solution resulted in significant changes to the PLGA surface chemistry due to hydrolytically labile ester linkages in acidic environments as opposed to more stable acetyl groups in CA.

Acidic collagen treatment led to a change in material chemistry, but this change was not enough to account for a decline in the molecular weights for both the polymers. Though PLGA showed a significant decline in its weight average molecular weight (Mw) and Z-average molecular weight (Mz), it maintained its original polydispersity index (PDI). These results are consistent with the observations of Zolnik et al.⁴¹, who found similar molecular weight changes in PLGA incubated in acidic environment. This indicates the occurrence of PLGA microsphere surface erosion without a considerable effect on its bulk properties. Conversely, CA microspheres resulted in a decrease in the number average molecular weight (Mn) that affected the PDI. Changes in molecular weight for both the polymers following treatment were insignificant (Fig. S3, ESI†) and thus it is reasonable to adopt this methodology for the infusion of collagen nanofibers.

Characterization of collagen nanofibers

Inclusion of ECM components as part of the scaffold system has been widely adopted to improve the scaffold’s bioactivity and biological performance. For instance, the inclusion of non-specific tissue components like small intestinal submucosa (SIS) into poly(lactic acid) (PLA) scaffold resulted in improved biocompatibility.⁴⁶ However, this approach benefits from the addition of non-specific ECM components from a tissue. Using more well defined ECM components like collagen, fibronectin, and RGD peptides to control cellular events and to enhance bioactivity is advantageous over the use of non-specific components, as non-specific components could have more batch-to-batch variance and bring less consistency in the engineered scaffold system. Traditionally, collagen is a popular ECM component incorporated into polymeric^46,47 and ceramic⁴⁸ scaffolds to promote cell response and biocompatibility. Often, collagen and collagen containing scaffolds are stabilized by cross-linking and involve agents such as NHS/EDC coupling, glutaraldehyde⁴⁶ and chemical modifications.⁴⁹ Though cross-linking stabilizes the collagen structure, the process often compromises its degradation and causes inflammatory responses due to modifications of the native structure (for example – Healos®, INFUSE®, and Mastergraft Matrix®).⁵⁰ Self-assembled proteins^51,52 and peptides⁵³ offer the same benefits of specific ECM components without the need for crosslinking agents. The molecular self-assembly of proteins and peptides greatly depends on the interactions of the material with the protein or peptide. For instance, Razafiarison et al.⁵⁴ studies clearly show differences in collagen assembly on a hydrophobic surface versus a hydrophilic surface. The self-assembled collagen on a modified hydrophilic polymer surface was found to be osteoinductive and resulted in improved cell performance.⁵⁴ Self-assembled ECM protein nanofibers may present a more native structure and present a viable platform to coat scaffolds and implant materials. This study characterizes the self-assembly of collagen on the 3D-micro-structures of CA and PLGA in terms of its morphology and the possible accompanying changes to CA and PLGA base structures. The hydrophilic surface offered by the polysaccharide cellulose promoted collagen self-assembly that was more biomimetic. Furthermore, we have previously shown the stability of self-assembled nanofibers without any drastic cross-linking in an in vitro culture for up to 28 days.²²

Scanning electron microscopy images of the PLGA groups showed a smoother microsphere surface while CA scaffolds presented a rougher surface with many features (Fig. 1a). Rougher surfaces have been shown to enhance osteoblastic phenotype development and maintenance when compared to smooth surfaces of the same material.^55,56 For example, titanium surfaces with greater micro⁵⁷ and nanoscale roughness⁵⁸ were seen to integrate better with the host bone by facilitating osteogenic differentiation of progenitors in contrast to smooth surfaces of the same material. Similar amounts of collagen were observed on both the PLGA and CA scaffolds, as determined by a protein assay (Fig. 1b). However, major differences were observed in terms of collagen assembly, fibril diameter and collagen distribution on PLGA and CA micro-nano structures. Collagen fibers on PLGA showed a wider range of fiber diameter distribution with a mean fiber diameter of 150 nm, while the collagen fibers on CA showed a narrow range of distribution with a mean fiber diameter of 80 nm (Fig. 1c). At 50 000× magnification, the collagen fibers of CAc presented a D-band like pattern of triple helical collagen. Such banding patterns were very sparse for the collagen fibers on PLGA (Fig. 1d). On subjecting the PLGAc and the CAc scaffolds to two-photon excitation, the second harmonic signal derived from the collagen fibers on CAc was stronger than that of PLGAc (Fig. 1e). In addition, the CAc group typically had collagen fibers arranged in co-operative sheets while the collagen fibers in the PLGAc group were more dispersed. Such observed differences in the collagen absorbance may be due to the material’s ability to interact with water and protein as they are inversely related.⁵⁹ The surface energy of the material may be a factor in determining protein assembly in a more biomimetic manner on the polysaccharide that presents a rougher surface than the polyester that presented a smooth surface.²⁸

Analysis of hydration property

The wettability of surfaces plays an important role in determining cellular responses like cell adhesion⁶⁰ and other cellular events. Cells adhere to materials through adhesion mediating proteins that adsorb onto biomaterial surfaces. While highly hydrophobic surfaces favor high levels of protein adsorption, they also mediate more material–protein and protein–protein interactions, but minimal protein–cell interaction due to the masking of active protein domains responsible for cell adhesion.¹⁵ This is explained by the rigid conformation of proteins achieved on hydrophobic non-polar surfaces against a more flexible protein conformation achieved on hydrophilic non-polar surfaces,²⁸ which may quantitatively adsorb less protein.¹⁵ On the other hand, highly hydrophilic surfaces may not allow for protein adhesion on their surfaces at all and likewise such surfaces do not favor cell adhesion on the surface.^61,62 Hence, the hydration behavior of a material would dictate protein adsorption and self-assembly on the material’s surface by influencing the surface energy of the material. Therefore the interaction of PLGA and CA scaffolds with water was examined to determine its influence during collagen treatment.

Scaffolds of PLGA, PLGAc, CA and CAc were freeze dried and subjected to water uptake studies over a period of 72 hours. Scaffolds treated with buffer at an acidic pH without collagen, namely PLGA-buffer and CA-buffer, served as controls. Fig. 2 shows the water uptake by various groups in phosphate buffer saline (PBS pH 7.4) at 37 °C. The water uptake by CA based structures was twice that of the PLGA based structures at all times (Fig. 2a). An increase in weight was seen with all PLGA groups until 24 hours, but this decreased at 72 hours when structures dropped to half of their gained weight at their equilibrium swelling point. All PLGA groups maintained similar weight gain profiles and acidic collagen treatment did not alter their water uptake (Fig. S4a, ESI†). All CA structures gained significant weight in the first 24 hours and achieved equilibrium weight by 72 hours. The CA-buffer and CAc groups showed progressive weight gain at all time points between 24–72 h. The addition of collagen in CAc (Fig. S4b, ESI†) and the presence of additional carboxylic acid groups in CA-buffer samples (Fig. S2, ESI†) increased the exposure to acidic pH, resulting in greater hydrophilicity and weight gain than CA structures. However, PLGA-buffer groups appear to be more hydrophobic as the acidic pH treatment may lead to degradation on the surface (Fig. S1, ESI†).

Fig. 2 Effect of hydration on the scaffold weight and volume changes over 72 hours of incubation in PBS at 37 °C. Comparison of PLGA and CA groups for (a) weight changes and (b) volume changes. Two-way ANOVA with Bonferroni post-test, with 95% confidence intervals, *P < 0.001, ^#P < 0.01, ^@P < 0.05.

The volumetric changes on hydration between PLGA and CA groups also showed a significant volume gain by all the CA groups, in contrast to PLGA (Fig. 2b). The volume changes on the PLGA groups that occurred in the first 24 h of hydration (Fig. S4c, ESI†), are also in agreement with the weight gain profiles (Fig. S4a, ESI†). Volumetric changes for both PLGA and CA based scaffolds were stabilized by 24 hours. However at 72 hours, only PLGAc and PLGA-buffer groups showed volumetric changes (Fig. S4c, ESI†), while the CA groups remained unchanged (Fig. S4d, ESI†). The hydrophilic characteristics of CA scaffolds presumably allowed the adsorbed collagen to be presented in a more native conformation than on PLGA.

Characterization of hMSC response

Cells adhere to materials through adsorbed ligands by binding to these ligands via their integrin receptors.⁶³ The classical view in material–protein interaction argues that a hydrophobic matrix, due to its favourable surface energy, adsorbs more proteins from the surrounding environment. These absorbed proteins play an important function in regulating cellular events.^59,64,65 However, the absolute quantity of adsorbed proteins onto a material could not be correlated to cell adhesion.⁶⁶ Many reports identify the importance of matrix hydrophilicity as an essential feature for achieving cell adhesion and long term biocompatibility.⁶⁷ Studies focused on improving the matrix hydrophilicity without altering its chemistry or surface roughness led to greater cellular adhesion, demonstrating that cellular responses were optimal on slightly hydrophilic materials.⁶⁸ Though experimental evidence is minimal, it is believed that hydrophilic matrices may lead to a more native-like conformation of the adsorbed proteins which in turn produce better cellular responses.⁶⁹ The differential response of collagen self-assembly on CA and PLGA polymers could be an attributing factor for the observed D-banding patterns on the CA matrix.

The chemical functional groups on a material can influence cellular responses immensely. For instance Curran et al. cultured bone marrow derived mesenchymal stem cells (MSCs) on glass substrates modified to contain the functional groups: –CH₃, –SH, –COOH, –OH, and –NH₂, under basal, osteogenic and chondrogenic conditions. They found that the –CH₃ functional groups helped maintain the multipotency of MSCs, while the –OH and –NH₂ groups were conducive for osteogenic differentiation, and the –COOH and –SH groups on the substrate supported chondrogenic differentiation. These differences were observed both under basal and stimulated conditions.¹⁴ Many other studies have used materials that varied only by the chemical functional groups on the biomaterial’s surface and all of them have shown that the chemical functionalities on the material surface can have a profound influence on stem cell behavior.^13,70 These changes can be attributed to the resulting material surface energy and hydration characteristics that determine the conformation of adsorbed proteins and, in turn, the cellular adhesion.²⁷ In the present study, polysaccharide CA provides the aforementioned functionalities and the control PLGA polyester lacks these groups to study the osteochondral phenotype progression of seeded stem cells. Apart from this, the observed micro scale roughness of CA groups over the smooth morphology of PLGA could also act as an inductive factor for osteogenic differentiation of stem cells.⁵⁷

In order to examine the biological implications of the micro nanostructures, we investigated their osteoinductive ability using human bone marrow derived mesenchymal stem cells (hMSCs) under osteoinductive media. Good cell viability on PLGA, PLGAc, CA and CAc from day 3 to day 21 was evident (Fig. S5, ESI†). The distribution of cells on CA was more uniform than on PLGA, where the cells were mostly confined to the microspheres. PLGA has a contact angle above 90°, but CA has a much lower contact angle of around 55°. It has been reported in several studies that a contact angle between 35–65° may be more conducive for cell adhesion than a contact angle above 90°.^71,72 Hence the CA materials showed greater cell adhesion and spread than PLGA, owing to their hydrophilic nature. In the current study all structures were seeded at a high cell density to study the osteogenic phenotype development and hence the cell proliferation rates were constant due to confluency (Fig. 3a–c and S6 ESI†).

Characterization of osteoblastic differentiation of hMSCs

The alkaline phosphatase activity (ALP), as a precursor of osteoblast maturation, and mineralization determined by alizarin red staining (ALZ), as a marker of late stages of osteoblastic maturation, were evaluated at 21 days. The ALP activity was highest on CAc and it was significantly greater than CA. However, this was not the case with PLGAc and PLGA as there was no difference between the two groups in ALP activity (Fig. 3d). Mineralization was higher on collagen-coated groups than on their uncoated counterparts. However mineralization on CAc was significantly greater than mineralization on PLGAc (Fig. 3e and f). These findings indicate that the osteoblastic phenotype progression may be greater on CAc than the other groups including PLGAc.

Fig. 3 In vitro phenotype development by hMSCs seeded on to scaffolds under osteoinduction. Viability of seeded cells over time on PLGA (top left), PLGAc (top right), CA (bottom left) and CAc (bottom right) with live (green)/dead (red) staining at (a) day 3 and (b) day 21. Scalebar = 200 μm. (c) DNA content by pico-green assay over 21 days in culture and (d) alkaline phosphatase activity at 21 days of osteoinduction on PLGA, PLGAc, CA and CAc. Mineralization at 21 days of osteoinduction on scaffolds measured by (e) visual micrographs of alizarin red stained scaffolds (PLGA, PLGAc, CA and CAc – top to bottom) and (f) colorimetric quantification of calcium deposition on PLGA, PLGAc, CA and CAc. One-way ANOVA with Tukey post-test, with 95% confidence intervals, *P < 0.001, ^#P < 0.01, ^@P < 0.05.

To further understand if there is a difference in the osteochondral progression of seeded hMSCs, the expression of osteochondral genes and proteins were analyzed (Fig. S7, ESI† and Fig. 4). Collage 1 (Coll1), an important ECM component of bone was significantly upregulated in all groups in contrast to PLGA. The CAc group had the highest collagen expression and it was significantly greater than PLGAc and CA (Fig. 4a). Osteonectin (ON), a protein associated with the collagen ECM, also showed a similar trend with cells on CAc showing the highest levels of expression (Fig. 4b). Bone sialoprotein (BSP), a mature marker of osteoblasts, was upregulated considerably on CA and CAc in contrast to PLGA and PLGAc. Interestingly CAc performed better than CA, but cells on PLGAc and PLGA showed no significant difference in BSP expression (Fig. 4c). Additionally, an important osteoblastic transcription factor Runx2 was seen to be significantly upregulated on CAc in contrast to all other groups (Fig. 4d).

Fig. 4 Osteoblast markers expressed by osteoinduced hMSCs on PLGA, PLGAc, CA and CAc at 21 days. Osteogenic gene expression of (a) collagen1A1 (Coll1A1), (b) osteonectin (ON), (c) bone sialoprotein (BSP) and (d) Runx2. Immunostaining for osteoblastic protein markers (marker – red/F-actin (phalloidin) – green/nucleus (Nucblue DAPI) – blue) staining of PLGA (top left), PLGAc (top right), CA (bottom left) and CAc (bottom right) (e) collagen1/F-actin/nuclei staining (f) bone sialoprotein/F-actin/nuclei and (g) no primary control (background – red/F-actin (phalloidin) – green/nuclei (DAPI) – blue) scalebar = 200 μm. One-way ANOVA with Tukey post-test, with 95% confidence intervals, *P < 0.001, ^#P < 0.01, ^@P < 0.05.

Furthermore we examined genes for osteochondral progression (Fig. S7, ESI†), such as alkaline phosphatase (ALPL), collagen III (Coll3A1), matrix metalloproteinase 13 (MMP13), Sox9 and collagen10A1 (Coll10A1). We saw an upregulation of these genes on the natural polymers (CA and CAc) in contrast to the synthetic polymers (PLGA and PLGAc). CAc showed the highest expression of all the genes indicating greatest osteoinduction of seeded hMSCs. The long bones of our body are formed by endochondral ossification,⁵⁸ in this process the MSCs undergo a transition through the formation of a condensed chondrogenic template to facilitate the formation of bone. Both CA and CAc promote better cell adhesion by integrin signaling due to their material properties, resulting in better cell–cell contact as well. This is likely to be more conducive for cellular condensation and cause the greater osteochondral progression seen on CA and CAc.

To visualize phenotype expression and cellularity within scaffolds, immunostains were performed for osteoblastic markers such as Coll1 and BSP along with phalloidin for cytoskeletal F-actin and Nucblue for cell nuclei. While the cells and markers were distributed evenly on the CA and CAc groups, PLGAc showed better cellular distribution than PLGA. The cells and markers were confined to the microspheres, with very little cellularity in the pores (Fig. 4e and f). Thus osteoblastic maturation was greatest on CAc followed by CA, indicating the effectiveness of natural polymeric micro-nano structured scaffolds towards the osteoinduction of stem cells. Though both CAc and PLGAc possessed collagen nanofibers, the collagen on CAc was seen to be more biomimetic than the collagen on PLGAc. It has been shown in the literature that denaturation of collagen leads to a loss of the mechanical stimuli needed for MSC’s osteoblastic differentiation.^73,74 It is also known that the Hedgehog signaling pathway, seen to play an important role in osteochondral progression^75,76 in the body, is upregulated by type one collagen with a native structure.⁷⁷ Hence we observed that the CAc micro-nanostructures were osteoinductive, to an extent greater than its PLGAc counterpart.

Biocompatibility analysis

Biocompatibility is a complex factor that involves the immune response by the body to the implanted foreign material. Though every material tends to elicit an immune response, it is important to make sure that the foreign body response (FBR) to a material is acceptable as per the norms laid out by regulatory agencies before implantation into the body. Though in vitro cell culture testing gives some indications for the material’s acceptability, implantation into the body is necessary to establish biocompatibility, understand the tissue response to the material and its safety as an implant material. The (IOS) 10993 criteria for biomaterial biocompatibility mandates implantation into animals and require that the implants be non-toxic on implantation.⁷⁸ Studies outlined here are designed to establish biocompatibility and tissue infiltration within the matrix over a period of time.

Subcutaneously implanted PLGA, CA and CAc scaffolds in Sprague Dawley® rats retrieved at week 2, 4, 8 and 12 were processed for histology. At the earliest time point, CA scaffolds had very little cellularity compared to the PLGA controls. However, the CAc scaffolds presented good cellularization at the early time point (2 weeks). The multinucleated giant cells (MNCs) representing a foreign body response, were high on CA, but considerably lower on CAc at 2 weeks (Fig. 5a). The collagen content of the tissue was also higher in CAc than CA (Fig. 5b). At later time points the response of all of the scaffolds seemed similar, with good cellularity in all cases (Fig. 5c). The collagen content in all groups had also increased (Fig. 5d). The progressive cellularization of CA groups and the increasing FBR on PLGA can be seen with the hematoxylin and eosin (H&E) stain (Fig. S8 and S9, ESI†). The histomorphometric analysis revealed a high FBR on PLGA at week 12, but a considerably lower immune response on CA and CAc. The tissue area on PLGA was reduced at week 12 in contrast to CA, which showed progressively greater cellularization with time and performed significantly better than PLGA (Fig. 5e, f and S10, ESI†).

Fig. 5 Hematoxylin and eosine (H&E) staining of subcutaneously implanted scaffolds at (a), (b) 2 weeks, and (c), (d) 8 weeks. Sections stained with H&E at (a) week 2, (c) week 8. Gomori trichrome staining of subcutaneously implanted scaffolds at (b) week 2, (d) week 8. Histomorphometric scores of (e) % tissue area in pores of the scaffold implant over time, (f) multinucleated giant cells (MNCs) in the implanted scaffolds over time. Scalebars at top panel (1×) = 3 mm, scalebar on CAc (middle panel) (20×) = 100 μm, all other 20× panels (middle panels) scalebar = 300 μm. Scalebar on CAc (lowest panel) (40×) = 60 μm, all other 40× (lowest panels) scalebar = 300 μm. Two-way ANOVA with Bonferroni post-test, with 95% confidence intervals, *P < 0.001, ^#P < 0.01, ^@P < 0.05.

Materials are exposed to body fluids following implantation and this leads to the adsorption of several proteins and opsonins which are also dependent on material hydrophilicity and hydrophobicity. A study by Elwin et al. showed the importance of protein absorption and stability to account for FBR. For instance, FBR response factor C3 compliment was adsorbed onto a highly hydrophobic material and resulted in denaturation of the adsorbed protein.⁷⁹ This led to exposure of the epitopes recognized by the antigens in the body and the ensuing FBR was greater. For example, hydrophilic contact lenses show lesser protein buildup and offer greater biocompatibility.⁸⁰ Hydrogels on the other hand absorb very little proteins and have shown greater biocompatibility and tissue integration responses.⁸¹ The present study showed an immune response to the CA scaffold at the early time points of 2 and 4 weeks in comparison to PLGA. However, the collagen coated CA and PLGA showed lower FBR response and resulted in matrix cellularization. The hydrophilic nature of CA resisted the initial adsorption of bodily proteins that leads to a lag in cell adhesion/infiltration. On the other hand the CAc scaffolds overcame the limitation associated with CA at early time points due to the presence of exogenous collagen. Both CA and CAc showed greater biocompatibility and tissue infiltration at 12 weeks, which differed from the PLGA matrices. For instance, the PLGA matrices resulted in an increase in MNCs, representative of an increasing FBR with time and reduced tissue infiltration. PLGA and synthetic polymers like PLLA may elicit an adverse immune response with time due to the change in the environmental pH associated with the degradation of the polymer matrix.⁸²

Hence, we have developed a polysaccharide-protein based natural polymeric, micro-nano structured porous scaffold system consisting of micron sized sintered CA microspheres decorated with collagen nanofibers. We compared it to the identical structure created by PLGA microsphere scaffolds, to determine the effect of the polymer on collagen coating, cellular response and biocompatibility. Though similar collagen loadings were seen on both PLGA and CA sintered microsphere scaffolds, nanofibers on CA had a smaller diameter and generated more SH signal along with greater D-banding patterns, similar to the native 3D structure of collagen found in tissues. The effect of substrate chemistry on protein adsorption and confirmation revealed greater protein adsorption on the hydrophobic materials.⁸³ However, many reports also show no difference in the conformation of adsorbed proteins on hydrophobic and hydrophilic matrices. Most of these studies report protein adsorption and confirmation in the solution phase within a short period of time,^84,85 which is different from what is presented in this study. The long term protein adsorption onto a matrix and following the conformational changes provides a better understanding of the biomaterial implant’s in vivo performance.⁸⁶ Though the absolute quantity of protein adsorbed onto a hydrophobic material may be higher, the conformation is less likely to be maintained close to the native protein structure,^83,87,88 while hydrophilic surfaces tend to allow the protein to maintain a structure similar to their native state once adsorbed. Also, it has been consistently seen in the literature that treatments such as plasma coating that increase materials hydrophilicity have led to long-term cell adhesion and biocompatibility.^89,90 This study showed that the polysaccharide chemistry of CA along with the rougher microsphere surface allowed retention of the native structure of self-assembled collagen in contrast to PLGA, a more hydrophobic, smooth surfaced matrix. Several other factors such as surface charges and functional groups on CA might have also contributed to the observed collagen conformation.

Our studies on the osteoinductive ability of the scaffolds using hMSCs revealed the superior ability of CAc to favour osteogenic progression of progenitor cells, in contrast to the other scaffolds, both in terms of phenotype development and osteochondral gene and protein expression. These results emphasize the functional relevance of achieving a more biomimetic matrix to induce beneficial cellular responses as an inductive substitute for native ECM. Furthermore the implants showed great biocompatibility when placed in subcutaneous pouches, on par with the PLGA scaffolds. At twelve weeks the CA scaffolds were immunologically well tolerated in contrast to PLGA, which started showing FBR, probably due to degradation of the matrix. Collagen functionalized CAc scaffolds avoided the initial lag at two weeks that CA scaffolds exhibited in cellularization. These observations again highlight that a hydrophilic matrix like CA might lead to lower protein adsorption in the short term, but probably allow for greater conformational stability of adsorbed proteins from the body to be presented in a native state to avoid FBR on the long run. Using a native protein such as collagen to coat the scaffolds before implantation cued the body to accept the matrix more readily along with maintenance of biocompatibility over time. Thus the present study illustrates that combining natural polymers with micro and nanoscale features could be advantageous to achieve bone regeneration and tissue healing.

Experimental section

Materials

Cellulose acetate (Mw: 30k) (CA), and polyvinyl alcohol (30 000–70 000) (PVA) were procured from Sigma-Aldrich (St. Louis, MO, USA). Poly(lactic-co-glycolic acid) 85 : 15 (PLGA) was purchased from Lakeshore Biomaterials (Birmingham, USA). Acetone, dichloromethane, cyclohexane, paraformaldehyde and glutaraldehyde were purchased from Fisher Scientific (Fair Lawn, NJ, USA). BioRad alkaline phosphatase substrate kit (Hercules, CA, USA) and Invitrogen Quant-iT PicoGreen dsDNA assay kit (Eugene, Oregon, USA) were used in this study. Alizarin red and cetylpyridinium chloride was procured from Acros Organics (New Jersey, USA). Cell culture media was DMEM (HG) purchased from Lonza (Walkersville, USA).

Preparation of microspheres

Oil in water solvent-evaporation, followed by sintering of formed microspheres using a solvent/non-solvent mixture was used for producing the microspheres of (i) cellulose acetate (CA), (ii) poly(lactic-co-glycolic acid) 85 : 15 (PLGA). In brief, 13% (w/v) of CA polymer was dissolved in a solvent mixture containing methylene chloride and acetone at a ratio of 9 : 1 to produce microspheres. A 20% (w/v) solution in methylene chloride was used in the case of PLGA polymer to produce the microspheres. The polymer solutions were then poured in a thin stream into an aqueous media containing 1.25% (w/v) PVA as a surfactant, with constant stirring at 250 rpm to form an oil-in-water emulsion. These suspensions were stirred overnight to evaporate the solvent and obtain hardened microspheres. Isolated microspheres were washed repeatedly with deionized (DI) water, dried, and sieved into different microsphere sizes. Microspheres in the size range of 600–710 and 710–800 μm were mixed at a weight ratio of 1 : 1 and sintered into microporous scaffolds using either solvent-non-solvent or heat sintering based on the polymer.^22,33

Preparation of micro-porous sintered microsphere scaffolds

Teflon molds of different dimensions were filled with CA microspheres to which 200 μL of an optimized solvent/non-solvent composition, acetone : cyclohexane in the ratio of 3 : 1 (v/v), was added to cover the microspheres. The solvent was allowed to evaporate at room temperature to obtain sintered microsphere scaffolds. Cylindrical scaffolds measuring 5 mm diameter × 10 mm height were used for the characterization of hydration properties while 8 mm diameter × 2 mm thick tablets were used for cell studies and animal experiments. The control PLGA microporous scaffolds with identical micro-particle sizes were produced by heat sintering at 95 °C for 45 minutes using Teflon molds.³³

Preparation of micro-nano structured scaffolds

A modified biomimetic approach was used to functionalize microporous scaffolds with type I collagen. In brief, both control and test scaffolds were incubated in a 0.1% (w/v) collagen type I solution with a pH adjusted to 4.2 at 37 °C for 7 days to promote molecular collagen self-assembly. The dried scaffolds were treated with UV light for 30 min each side to achieve collagen nanofiber stability and washed repeatedly with DI water to remove buffer salts. Microspheres in the size range of 300–425 μm diameters were functionalized with collagen and used for material characterizations presented in this study. The effect of coating buffer pH was studied by incubating scaffolds in the buffer alone without collagen.²²

Characterization internal structure of scaffolds by SEM

Scanning electron microscopy (SEM) was used to characterize scaffold morphology and evaluate collagen fiber diameter and distribution. Scaffolds were sputter coated with Au/Pd using a Polaron E5100 sputtering system (Quorum Technologies, East Sussex, UK) to achieve an eighteen nanometer thick coating before viewing under SEM. The samples were viewed using FEI Nova NanoSEM 450 scanning electron microscope (FEI, Hillsboro, OR, USA) operated at an accelerating voltage of 2 kV at various magnifications. Collagen fibers at different locations were selected randomly on the functionalized samples and these images were used for measuring the fiber diameter using FIJI, NIH software and averaged (an average of 100 fibers).

Quantification of scaffold collagen content

Amount of collagen present on each scaffold was estimated with a calorimetry using a bicinchoninic acid (BCA) protein assay reagent kit (Pierce). Proteins form a purple colored chelation complex with BCA via reducing the cupric ions to cuprous ions. The purple color of this complex is directly proportional to the protein concentration and the absorbance was read at 562 nm using a BioTek plate reader. In brief, collagen coated composite scaffolds were transferred to a new 48 well plate and incubated with 500 μL of aqueous 1% acetic acid overnight with agitation followed by mixing with the aid of a pipette to extract all collagen from the scaffold. Similarly collagen remaining on the TCPS surface during the coating experiment was also extracted and quantified. A volume of 25 μL of the collagen extract was mixed with 200 μL of BCA reagent followed by 30 min incubation at 37 °C and analyzed at 562 nm using a plate reader. A sample size of n = 4 was used for all estimations. The absorbance of six known collagen concentrations 0, 62.5, 125, 250, 500, and 1000 μg mL⁻¹ were used to construct a standard curve to convert absorbance readings to collagen concentrations. The content of collagen on the scaffolds and the plates used for incubation were measured and the percentage of collagen on the scaffold against that on the plate was calculated to obtain the collagen coating efficiency on each scaffold.

Morphology of collagen by two-photon microscopy

Micro-nano structured scaffolds, namely PLGAc and CAc, were imaged at an excitation wavelength of 890 nm and the collagen second harmonic generation (SHG) signal along with the autofluorescent signal from the scaffold material were acquired with bandpass filters of 435–485 nm (SHG), 500–550 nm (scaffold autofluorescence), and 570–620 nm (scaffold autofluorescence). The SHG signal is generated from the triple helical structure of collagen.

Microsphere surface chemistry by ATR-FTIR analysis

A Fourier transform infrared spectroscope (FTIR), Nicolet Magna 560, using a ZnSe and Ge crystal (Nicolet Instrument Corporation, Madison, WI, USA) was used to measure the Attenuated Total Reflection (ATR)-infrared red (IR) spectra of the CA, CA-buffer, CAc and PLGA, PLGA-buffer, PLGAc treated microsphere surfaces. The Nicolet OMNIC® software was used to blank the runs every time before samples were placed on the pedestal for measurement of their spectra. Sample scans were run at 320 scans per specimen to extract the spectra in the range 400–4000 cm⁻¹.

Microsphere surface crystallinity by XRD analysis

A Bruker D2 Phaser powder diffractometer (Bruker axs, Inc., Madison, WI, USA), was used to measure the wide angle X-ray diffraction of CA, CA-buffer, CAc and PLGA, PLGA-buffer, PLGAc microspheres. A 2θ between 10–60° with an increment of 0.02° was captured, for all the polymers.

Polymer molecular weight by GPC analysis

A Waters GPC system with Jordi Gel fluorinated DVB columns (1–100k, 2–10k & 1–500 Å) was used for running the PLGA, PLGA-buffer and PLGAc microparticles. Tetrahydrofuran (THF) was used as a mobile phase along with Varian 380-LC Evaporative Light Scattering Detector (ELSD) detector. In brief, 0.1% (w/w) polymer solutions in THF were filtered through 0.2 μm nylon filter prior to testing. A flow rate of 1.25 mL min⁻¹ with an injection volume of 100 μL was used for the analysis. The molecular weight distributions were compared with polystyrene standards with molecular weight ranges 450–200 000 Da. For CA based samples, a Waters GPC system with two mixed bed Jordi Gel DVB columns were used with dimethylacetamide (DMAC) as the mobile phase. For this analysis 0.2% (w/w) CA in DMAC was used with at an injection volume 200 μL and a flow rate of 1.25 mL min⁻¹. Samples were analyzed using a Waters 2414 refractive index detector (RI) detector. The molecular weight distributions were compared with poly(methyl methacrylate) (PMMA) standards with molecular weight ranges 2000–1 100 000 Da.

Scaffold water uptake by gravimetric analysis

Scaffolds of 10 mm × 5 mm size were used for analysis of the weight and volume changes through hydration over 72 hours. The dry weight and dimensions of the scaffolds were noted prior to hydration testing (0 h). All groups, namely CA and PLGA (treated with DI water), CA-buffer/PLGA-buffer (treated with buffer without collagen) and CAc/PLGAc (treated with collagen solution) were individually placed in microcentrifuge tubes with 1 mL of PBS (pH 7.4) (n = 6 per group) and incubated at 37 °C. At the end of 24 h, 48 h and 72 h, the scaffolds were isolated and the changes in weight were noted with a Mettler Toledo® (XP56C-Mass Comparator) balance with an accuracy of 0.01 mg and dimensional changes were measured with a digital calipers (at six different points on the scaffold). The volume changes in the individual cylindrical scaffolds were calculated using the relation = π × (diameter/2)² × height.

In vitro human mesenchymal stem cells (hMSCs) culture on scaffolds

Human bone marrow derived mesenchymal stem cells were purchased from Lonza (Lonza, Walkersville, USA) and expanded as per the protocol provided by the supplier. Cells used for all experiments were at passage 5. The basal media consisted of DMEM (HG) (Lonza, Walkersville, USA) supplemented with 10% FBS and 1% antibiotics (penicillin–streptomycin). The osteogenic media was composed of basal media with 0.2 mM L-ascorbic acid and 7.0 mM glycerol 2-phosphate disodium salt and 0.1 μM dexamethasone.

Scaffolds were soaked in 70% ethanol for 20 min and allowed to dry in the cell culture hood. Each side of the scaffold was exposed to UV light for 20 minutes in the tissue culture hood. Scaffolds were soaked in basal media overnight prior to seeding with the hMSCs. Each scaffold was seeded with 250 000 cells in a 50 μL cell suspension and incubated for 4 h before additional media was added. A total of 500 μL basal media was added to the samples in a 48 well plate and then switched to osteogenic media following 24 h. The media was changed every other day and cultures were maintained for 21 days. One set of scaffolds was also cultured in basal media up to 21 days to evaluate hMSC proliferation using pico-green assay. All studies were done in triplicate for each time point and each group of scaffolds.

Cell viability

Viability of hMSCs on the scaffolds was analyzed with a live/dead cell viability kit. In brief, calcein AM enters live cells and reacts with intracellular esterase to produce a bright green fluorescence, while ethidium homodimer-1 enters only dead cells with damaged membranes and produces a bright red fluorescence upon binding to nucleic acids. Scaffolds were imaged on 3, 7, 14 and 21 days of osteoinduction using a Zeiss 780/Laser Scanning Confocal Microscope (LSCM) at magnifications of 10× to view cells independently and along with the scaffolds. 3D reconstruction of the confocal stacks was done using Imaris software (Bitplane).

Cell proliferation

The rate of hMSCs proliferation after transferring to osteogenic media was quantified by measuring the amount of cellular DNA content at various culture points using a Picogreen dsDNA assay. In brief, at different culture time of 3, 7, 14 and 21 days, the cellular constructs were washed twice with PBS, transferred to new well plates and 1 mL of 1% Triton X-100 solution was added to lyse the cells. The well plates underwent three freeze–thaw cycles, between −70 °C and room temperature, and mixed with the aid of a pipette to extract cell lysate from the 3D scaffolds prior to analysis. 125 μL of sample DNA was transferred into a new well plate to which 375 μL (component B) and 500 μL (component A) kit reagents were added. Well plates were covered with aluminum foil to prevent light exposure and incubated for 5 min. A BioTek plate reader was used to read fluorescence (485 nm/535 nm). Optical readings were converted in DNA concentration using a standard curve.

Alkaline phosphatase activity

Alkaline phosphatase activity of hMSCs on the scaffolds was evaluated as a marker of osteoblast phenotype progression using an ALP substrate kit. 100 μL of cell lysate was transferred into a well plate to which 400 μL of p-NPP (para-nitro phenol phosphate) substrate and buffer solution were added and incubated at 37 °C for 30 min. After 30 minutes, 500 μL of 0.4 N of sodium hydroxide was added to stop the reaction. The intensity of the color produced though the reaction is proportional to the ALP activity. The optical density of the solution was measured at 405 nm using a BioTek plate reader. The results for ALP activity were optical density and these were normalized to scaffold volume.

Mineralized matrix deposition by cells

Mineralized matrix deposition by osteoinduced hMSCs on the scaffolds were evaluated as marker of mature osteoblast phenotype using an alizarin red staining method for calcium deposition at 21 days of osteoinduction. This colorimetric analysis is based on solubilizing the red matrix precipitate with cetylpyridinium chloride (CPC) to yield a purple solution. In brief, at 21 days of osteoinductive cell culture, cellular constructs were fixed in 70% ethanol for 1 h, at room temperature and then stained with 40 mM alizarin red (Sigma) solution for 10 min at room temperature. After washing 5–10 times with distilled water to remove the adsorbed/absorbed dye, chemically bound red matrix precipitate was solubilized in 1 mL of 10% CPC until color was stable. The optical density of the solution was read at 562 nm using BioTek plate reader. The results for calcium deposition were also normalized to scaffold volume.

Gene expression

The RNA from the scaffolds was extracted using RNeasy Plus Universal Mini Kit (Quiagen, Valencia, CA, USA) as per the manufacturer’s protocol. Briefly, each scaffold was washed with sterile PBS (pH 7.4) and then RNA extraction was carried out using RNeasy Plus Universal® kit (Quiagen, Valencia, CA, USA). Briefly, Quiazol®/chloroform extraction yielded an aqueous layer which was loaded onto columns and washed successively using buffers to collect RNA which was suspended in 30 μL of RNase-free water in a microcentrifuge tube. The RNA was measured using a nanodrop spectrophotometer and 1.5 μg RNA was used for DNAse treatment to eliminate any remnant genomic DNA contamination. A DNA-free™ Kit (Ambion, Life Technologies, Foster City, CA, USA) was used and the manufacturer’s protocol was used to digest the cDNA.

This DNAse treated RNA was used to reverse transcribe into cDNA using iSript advanced reverse transcription kit (Bio-rad, Hercules, California, U.S.A.), using the manufacturer’s protocol. To each sample, 4 μL of reaction buffer and 1 μL of advanced reverse transcriptase enzyme were added. Following incubation at 42 °C for 30 minutes and enzyme inactivation at 85 °C for 5 minutes, the cDNA was ready for gene expression analysis.

Prime-PCR (Bio-Rad, Hercules, California, U.S.A.) gene arrays were performed where primers of osteochondral genes were used for the gene expression analysis. Validated SYBR green primers of the following genes were analyzed: GAPDH, TBP, HPRT1 (internal control genes expressions were normalized to), Coll1A1, Coll1A2, Coll3A1, Coll2A1, BGLAP, MMP-13, Coll10A1, Sox9, RUNX-2, ALPL, IBSP, BMP-2, DMP-1, SPARC (ON), MMP-9, MMP-13, IHH. To each of the test wells, 10 μL of 2× Sos advanced SYBR green mix, 10 μL of cDNA with ultrapure water (equivalent to 12.5 ng) was added. All controls to check for purity of starting material including gDNA contamination, and RNA integrity were included and passed the quality control criteria. Samples used were cells seeded and osteoinduced on PLGA, PLGAc, CA and CAc. PLGA was kept at 1.0 and the gene expression fold changes were with reference to PLGA.

Osteogenic marker immunostaining

The scaffolds seeded with hMSCs were harvested at 21 days of osteoinduction and washed with PBS (pH 7.4) prior to fixation with 4% paraformaldehyde (PFA) in PBS (pH 7.4) at room temperature for 10 minutes. The samples were then washed with PBS (pH 7.4) and permeabilized using 0.1% Triton X-100 in 1× PBS (pH 7.4) for 10 minutes. Samples were washed again with PBS (pH 7.4) and subsequently incubated in 3% bovine serum albumin (BSA) in 1× PBS (pH 7.4) solution for 30 min to block non-specific antibody binding. Primary antibodies at specific concentrations (Colleen Rabbit anti human (Abcam), Cambridge, MA, U.S.A.) at 1 : 200 dilution; and bone sialoprotein rabbit anti human (Millipore, Billerica, MA, U.S.A. at a dilution of 1 : 100) were dissolved in 1% BSA/PBS (pH 7.4) buffer. Scaffolds were incubated in primary antibody solution for one hour at room temperature. This was followed by rinsing of samples with PBS (pH 7.4). Dylight 594-goat-anti-rabbit secondary antibody (Jackson immune, West Grove, PA, U.S.A.) at a dilution of 1 : 400 was dissolved in 1% BSA/PBS (pH 7.4) buffer along with Alexafluor 488-Phalloidin (Invitrogen, Eugene, Oregon, USA) at a 1 : 40 dilution were added to scaffold construct for one hour. This was followed by rinsing of samples 3 times with PBS (pH 7.4). NucBlue (DAPI) at a 2 drops per mL of PBS (pH 7.4) was used for staining the cellular nuclei. This was followed by confocal imaging with z-stacking for each of the scaffolds for visualization of the marker (red)/cytoskeleton (green) and nuclei (blue).

Biocompatibility by subcutaneous implantation

Sprague Dawley rats of 250–300 grams of body weight were purchased from Charles River (Wilmington, MA, USA). All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC), at the UConn Health in accordance with the Public Health Service Policy on Humane Care and Use of Laboratory Animals (PHS 1986), the USDA Animal Welfare Act/Regulations (CFR 1985) and related guides. Isoflurane (3%) in a gaseous mix with oxygen was used as the anesthetic. Each rat was implanted with six scaffolds, two of each kind (PLGA, CA and CAc). Briefly, the dorsal side of the rats were shaved and prepped with betadine and alcohol. A 2.5 cm long incision was created to make a subcutaneous pouch and individual scaffolds (8 × 2 mm) were implanted into pouches. The pouches were sutured and animals were monitored regularly. At each of the 2, 4, 8 and 12 weeks of implantation time points, two rats were sacrificed using carbon dioxide overdose and four scaffolds (n = 4) of each type (PLGA, CA and CAc) were removed and collected for evaluation.

Histological staining

The scaffold samples were harvested along with the surrounding subcutaneous tissue, washed in PBS (pH 7.4) and fixed in 4% PFA/PBS (pH 7.4) overnight at 4 °C. The scaffolds were rinsed in PBS (pH 7.4) after fixation and transferred to a 1 : 1 mix of OCT and 30% sucrose solution in PBS (pH 7.4). The samples were incubated for 24 h at 4 °C. The samples were then transferred to OCT alone at room temperature for an hour. The samples were flash frozen and sectioned into 10 μm thick sections and captured on cryo-films. The cryo films were fixed onto slides using UV-curing adhesive. The samples were hydrated in water and hematoxylin and eosin staining and Gomori trichrome staining was performed.

Histomorphometric analysis

The hematoxylin and eosin stained samples were used for characterization of cell types infiltrating the implants. Two independent evaluators scored the samples for multinucleated giant cells (MNC) characteristic of foreign body response, macrophages, fibroblasts and blood vessels based on cell morphology, in consultation with a clinical pathologist. The Gomori trichrome samples were used as guidelines for the confirmation of the cellular phenotype. FIJI (NIH) software was used to count the total number of cells of each type and to measure the tissue area in the pore of each group of scaffold. All the scores were normalized to the tissue area fraction. Tissue area fraction was equal to tissue area in section (i.e. tissue area in the pores of the scaffold)/total area of the section. An n = 3 was used for this analysis.

Statistical analysis

All data are reported as the mean ± standard deviation (SD) of results from at least three independent runs. In case of experiments with two groups, a Student t-test was performed. In the case of experiments with multiple groups and multiple time points a 2-way-ANOVA with Bonferroni post-test was performed. In the case of experiments that focused on one time point and several groups, a one-way-ANOVA with Tukey post-test was performed. 95% confidence interval was employed to arrive at the p values. The denotation of significances given by p values are – *p < 0.001 (extremely significant), ^#p < 0.01 (very significant), ^@p < 0.05 (significant). Analyses were performed using GraphPad Prism Software (GraphPad Software, San Diego, CA).

Conclusions

This manuscript described the role of material–protein and material–cell interactions to promote tissue regeneration. In summary, we conclude that natural polymeric micro-nano structured 3D porous structures are superior to synthetic polymers of similar design in inducing osteoblastic differentiation of stem cells and eliciting greater biocompatibility in the body. We were able to obtain a more biomimetic presentation of protein (collagen) on the more hydrophilic natural polymers than on the synthetic polymers, that in turn translated into a superior biological performance, as shown by the ability to induce osteoblastic differentiation of stem cells and induce a favourable immune response in the host tissue on implantation of these matrices. Moreover, we have developed a completely natural polymeric scaffold platform that functions better than synthetic polymers in inducing favourable cellular differentiation and tissue healing response. Finally, we propose that natural polymeric micro-nano structured scaffolds, which are cost effective, osteoinductive and biocompatible can be viable substitutes to synthetic polymers such as PLGA for bone tissue engineering.

Acknowledgements

These studies were supported by a grant from the National Science Foundation (IIP-1355327, IIP-1311907 and EFRI-1332329), and Connecticut Regenerative Medicine Research Fund-15-RMB-UCHC-08. Principal Investigator, Prof. Kumbar has an equity interest in the Natural Polymer Devices, Inc. The terms of this arrangement have been reviewed and approved by the University of Connecticut, Storrs in accordance with its conflict of interest policies. The authors thank Dr Gloria Gronowicz (professor emerita) UConn Health, for their inputs in the study design and outcome evaluations.

References

1. F. B. Basmanav, G. T. Kose and V. Hasirci, Biomaterials, 2008, 29, 4195–4204.
2. K. Lee, E. A. Silva and D. J. Mooney, J. R. Soc., Interface, 2011, 8, 153–170.
3. C. Li, C. Vepari, H.-J. Jin, H. J. Kim and D. L. Kaplan, Biomaterials, 2006, 27, 3115–3124.
4. W. L. Murphy, M. C. Peters, D. H. Kohn and D. J. Mooney, Biomaterials, 2000, 21, 2521–2527.
5. I. Martin, R. Suetterlin, W. Baschong, M. Heberer, G. Vunjak-Novakovic and L. Freed, J. Cell. Biochem., 2001, 83, 121–128.
6. P. C. Bessa, M. Casal and R. Reis, J. Tissue Eng. Regener. Med., 2008, 2, 81–96.
7. A. Aravamudhan, D. M. Ramos, J. Nip, A. Subramanian, R. James, M. D. Harmon, X. Yu and S. G. Kumbar, Curr. Pharm. Des., 2013, 19, 3420–3428.
8. F. Granero-Moltó, J. A. Weis, M. I. Miga, B. Landis, T. J. Myers, L. O’Rear, L. Longobardi, E. D. Jansen, D. P. Mortlock and A. Spagnoli, Stem Cells, 2009, 27, 1887–1898.
9. J. Shao, W. Zhang and T. Yang, Biol. Res., 2015, 48, 1.
10. S. Tholpady, A. Katz and R. Ogle, Anat. Rec., Part A, 2003, 272, 398–402.
11. F. Guilak, D. M. Cohen, B. T. Estes, J. M. Gimble, W. Liedtke and C. S. Chen, Cell Stem Cell, 2009, 5, 17–26.
12. A. J. Engler, S. Sen, H. L. Sweeney and D. E. Discher, Cell, 2006, 126, 677–689.
13. D. S. Benoit, M. P. Schwartz, A. R. Durney and K. S. Anseth, Nat. Mater., 2008, 7, 816–823.
14. J. M. Curran, R. Chen and J. A. Hunt, Biomaterials, 2006, 27, 4783–4793.
15. E. Filová, E. Brynda, T. Riedel, L. Bačáková, J. Chlupáč, V. Lisá, M. Houska and J. E. Dyr, J. Biomed. Mater. Res., Part A, 2009, 90, 55–69.
16. E. K. Yim, E. M. Darling, K. Kulangara, F. Guilak and K. W. Leong, Biomaterials, 2010, 31, 1299–1306.
17. R. L. Sammons, N. Lumbikanonda, M. Gross and P. Cantzler, Clin. Oral. Implants. Res., 2005, 16, 657–666.
18. P. X. Ma, Adv. Drug Delivery Rev., 2008, 60, 184–198.
19. A. Aravamudhan, D. Ramos, A. Nada and S. Kumbar, Nat. Synth. Biomed. Polym., 2014, 67–89.
20. G. G. d’Ayala, M. Malinconico and P. Laurienzo, Molecules, 2008, 13, 2069–2106.
21. O. Felt, P. Buri and R. Gurny, Drug Dev. Ind. Pharm., 1998, 24, 979–993.
22. A. Aravamudhan, D. M. Ramos, J. Nip, M. D. Harmon, R. James, M. Deng, C. T. Laurencin, X. Yu and S. G. Kumbar, J. Biomed. Nanotechnol., 2013, 9, 719–731.
23. M. D. Kofron, J. A. Cooper Jr, S. G. Kumbar and C. T. Laurencin, J. Biomed. Mater. Res., Part A, 2007, 82, 415–425.
24. J. M. Holzwarth and P. X. Ma, Biomaterials, 2011, 32, 9622–9629.
25. M. M. Stevens, Mater. Today, 2008, 11, 18–25.
26. L. Wang, G. Zhao, R. Olivares-Navarrete, B. F. Bell, M. Wieland, D. L. Cochran, Z. Schwartz and B. D. Boyan, Biomaterials, 2006, 27, 3716–3725.
27. B. G. Keselowsky, D. M. Collard and A. J. García, J. Biomed. Mater. Res., Part A, 2003, 66, 247–259.
28. G. Anand, S. Sharma, A. K. Dutta, S. K. Kumar and G. Belfort, Langmuir, 2010, 26, 10803–10811.
29. H. W. Kim, L. H. Li, E. J. Lee, S. H. Lee and H. E. Kim, J. Biomed. Mater. Res., Part A, 2005, 75, 629–638.
30. M. V. Jose, V. Thomas, K. T. Johnson, D. R. Dean and E. Nyairo, Acta Biomater., 2009, 5, 305–315.
31. P. Gentile, V. Chiono, I. Carmagnola and P. V. Hatton, Int. J. Mol. Sci., 2014, 15, 3640–3659.
32. J. M. Anderson and M. S. Shive, Adv. Drug Delivery Rev., 2012, 64, 72–82.
33. S. Kumbar, U. Toti, M. Deng, R. James, C. Laurencin, A. Aravamudhan, M. Harmon and D. Ramos, Biomed. Mater., 2011, 6, 065005.
34. J. B. Lee, Y.-G. Ko, D. Cho, W. H. Park, B. N. Kim, B. C. Lee, I.-K. Kang and O. H. Kwon, J. Nanomater., 2015, 2015, 136.
35. D. Naumann, D. Helm, H. Labischinski and P. Giesbrecht, Mod. Tech. Rapid Microbiol. Anal., 1991, 43–96.
36. C. Chu, Polymer, 1985, 26, 591–594.
37. B. S. Zolnik, P. E. Leary and D. J. Burgess, J. Controlled Release, 2006, 112, 293–300.
38. N. Faisant, J. Siepmann and J. Benoit, Eur. J. Pharm. Sci., 2002, 15, 355–366.
39. F. Alexis, Polym. Int., 2005, 54, 36–46.
40. L. Li and S. P. Schwendeman, J. Controlled Release, 2005, 101, 163–173.
41. B. S. Zolnik and D. J. Burgess, J. Controlled Release, 2007, 122, 338–344.
42. K. Fu, D. W. Pack, A. M. Klibanov and R. Langer, Pharm. Res., 2000, 17, 100–106.
43. A. G. Ding and S. P. Schwendeman, Pharm. Res., 2008, 25, 2041–2052.
44. K. Makino, H. Ohshima and T. Kondo, J. Microencapsulation, 2008, 3, 203–212.
45. Y. Yamashita and T. Endo, J. Appl. Polym. Sci., 2004, 91, 3354–3361.
46. G. Khang, J. M. Rhee, P. Shin, I. Y. Kim, B. Lee, S. J. Lee, Y. M. Lee, H. B. Lee and I. Lee, Macromol. Res., 2002, 10, 158–167.
47. P. Datta, P. Ghosh, K. Ghosh, P. Maity, S. K. Samanta, S. K. Ghosh, P. K. D. Mohapatra, J. Chatterjee and S. Dhara, J. Biomed. Nanotechnol., 2013, 9, 870–879.
48. S. Teixeira, M. Fernandes, M. Ferraz and F. Monteiro, J. Biomed. Mater. Res., Part A, 2010, 95, 1–8.
49. S. M. Oliveira, R. A. Ringshia, R. Z. Legeros, E. Clark, M. J. Yost, L. Terracio and C. C. Teixeira, J. Biomed. Mater. Res., Part A, 2010, 94, 371–379.
50. C. Kraiwattanapong, S. D. Boden, J. Louis-Ugbo, E. Attallah, B. Barnes and W. C. Hutton, Spine, 2005, 30, 1001–1007.
51. W. Zhang, S. Liao and F. Cui, Chem. Mater., 2003, 15, 3221–3226.
52. F. W. Kotch and R. T. Raines, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 3028–3033.
53. S. Zhang, Nat. Biotechnol., 2003, 21, 1171–1178.
54. T. Razafiarison, U. Silván, D. Meier and J. G. Snedeker, Adv. Healthcare Mater., 2016, 5, 1481–1492.
55. B. Boyan, S. Lossdorfer, L. Wang, G. Zhao, C. Lohmann, D. Cochran and Z. Schwartz, Eur. Cells Mater., 2003, 6, 22–27.
56. G. Schneider, H. Perinpanayagam, M. Clegg, R. Zaharias, D. Seabold, J. Keller and C. Stanford, J. Dent. Res., 2003, 82, 372–376.
57. M. O. Klein, A. Bijelic, T. Ziebart, F. Koch, P. W. Kämmerer, M. Wieland, M. A. Konerding and B. Al-Nawas, Clin. Implant. Dent. Relat. Res., 2013, 15, 166–175.
58. R. A. Gittens, T. McLachlan, R. Olivares-Navarrete, Y. Cai, S. Berner, R. Tannenbaum, Z. Schwartz, K. H. Sandhage and B. D. Boyan, Biomaterials, 2011, 32, 3395–3403.
59. E. A. Vogler, Biomaterials, 2012, 33, 1201–1237.
60. L. Bačáková, M. Lapčíková, D. Kubies and F. Rypáček, in Tissue Engineering, Stem Cells, and Gene Therapies, Springer, 2003, pp. 179–189.
61. L. Grausova, A. Kromka, L. Bacakova, S. Potocky, M. Vanecek and V. Lisa, Diamond Relat. Mater., 2008, 17, 1405–1409.
62. L. Grausova, L. Bacakova, A. Kromka, M. Vanecek, B. Rezek and V. Lisa, Diamond Relat. Mater., 2009, 18, 258–263.
63. A. J. García and D. Boettiger, Biomaterials, 1999, 20, 2427–2433.
64. D. L. Elbert and J. A. Hubbell, Annu. Rev. Mater. Sci., 1996, 26, 365–394.
65. A. Krishnan, P. Cha, Y.-H. Liu, D. Allara and E. A. Vogler, Biomaterials, 2006, 27, 3187–3194.
66. J. Comelles, M. Estévez, E. Martínez and J. Samitier, Nanomedicine, 2010, 6, 44–51.
67. J. D. Bryers, C. M. Giachelli and B. D. Ratner, Biotechnol. Bioeng., 2012, 109, 1898–1911.
68. A. Ranella, M. Barberoglou, S. Bakogianni, C. Fotakis and E. Stratakis, Acta Biomater., 2010, 6, 2711–2720.
69. Z. Ma, Z. Mao and C. Gao, Colloids Surf., B, 2007, 60, 137–157.
70. J. E. Phillips, T. A. Petrie, F. P. Creighton and A. J. García, Acta Biomater., 2010, 6, 12–20.
71. K. Webb, V. Hlady and P. A. Tresco, J. Biomed. Mater. Res., 1998, 41, 422.
72. J. Wei, M. Yoshinari, S. Takemoto, M. Hattori, E. Kawada, B. Liu and Y. Oda, J. Biomed. Mater. Res., Part B, 2007, 81, 66–75.
73. S.-W. Tsai, H.-M. Liou, C.-J. Lin, K.-L. Kuo, Y.-S. Hung, R.-C. Weng and F.-Y. Hsu, PLoS One, 2012, 7, e31200.
74. L. Xu, A. L. Anderson, Q. Lu and J. Wang, Biomaterials, 2007, 28, 750–761.
75. B. St-Jacques, M. Hammerschmidt and A. P. McMahon, Genes Dev., 1999, 13, 2072–2086.
76. C. Scotti, B. Tonnarelli, A. Papadimitropoulos, A. Scherberich, S. Schaeren, A. Schauerte, J. Lopez-Rios, R. Zeller, A. Barbero and I. Martin, Proc. Natl. Acad. Sci., 2010, 107, 7251–7256.
77. S. M. Zunich, M. Valdovinos, T. Douglas, D. Walterhouse, P. Iannaccone and M. L. Lamm, Mol. Cancer, 2012, 11, 1.
78. B. D. Ratner, Sci. Transl. Med., 2015, 7, 272–274.
79. H. Elwing, B. Nilsson, K.-E. Svensson, A. Askendahl, U. R. Nilsson and I. Lundström, J. Colloid Interface Sci., 1988, 125, 139–145.
80. K. L. Menzies and L. Jones, Optom. Vis. Sci., 2010, 87, 387–399.
81. L. Zhang, Z. Cao, T. Bai, L. Carr, J.-R. Ella-Menye, C. Irvin, B. D. Ratner and S. Jiang, Nat. Biotechnol., 2013, 31, 553–556.
82. O. Böstman, J. Bone. Joint. Surg., 1998, 80, 333–338.
83. V. Hlady and J. Buijs, Curr. Opin. Biotechnol., 1996, 7, 72–77.
84. E. Gurdak, P. G. Rouxhet and C. C. Dupont-Gillain, Colloids Surf., B, 2006, 52, 76–88.
85. E. Pamuła, V. De Cupere, Y. F. Dufrêne and P. G. Rouxhet, J. Colloid Interface Sci., 2004, 271, 80–91.
86. G. Gao, D. Lange, K. Hilpert, J. Kindrachuk, Y. Zou, J. T. Cheng, M. Kazemzadeh-Narbat, K. Yu, R. Wang and S. K. Straus, Biomaterials, 2011, 32, 3899–3909.
87. C. A. Haynes and W. Norde, J. Colloid Interface Sci., 1995, 169, 313–328.
88. D. Kowalczyk, S. Slomkowski and F. W. Wang, J. Bioact. Compat. Polym., 1994, 9, 282–309.
89. C. M. Alves, Y. Yang, D. Carnes, J. Ong, V. Sylvia, D. Dean, C. Agrawal and R. Reis, Biomaterials, 2007, 28, 307–315.
90. A. Hezi-Yamit, C. Sullivan, J. Wong, L. David, M. Chen, P. Cheng, D. Shumaker, J. N. Wilcox and K. Udipi, J. Biomed. Mater. Res., Part A, 2009, 90, 133–141.

---

Footnote

† Electronic supplementary information (ESI) available: Material characterization, stem cell viability, stem cell proliferation, histology and histomorphometric scores at all implantation time points. See DOI: 10.1039/c6ra15363a

---