Tissue Engineering
Polymer Powder Processing of Cryomilled Polycaprolactone for Solvent-free Generation of Homogeneous Bioactive Tissue Engineering Scaffolds Jing Lim, Mark Seow Khoon Chong, Jerry Kok Yen Chan, and Swee-Hin Teoh*
Synthetic
polymers used in tissue engineering require functionalization with bioactive molecules to elicit specific physiological reactions. These additives must be homogeneously dispersed in order to achieve enhanced composite mechanical performance and uniform cellular response. This work demonstrated the use of a solvent-free powder processing technique to form osteoinductive scaffolds from cryomilled polycaprolactone (PCL) and tricalcium phosphate (TCP). Cryomilling was performed to achieve micron-sized distribution of PCL and reduce melt viscosity, thus improving TCP distribution and improving structural integrity. A breakthrough was achieved in the successful fabrication of 70 weight percentage of TCP into a continuous film structure. Following compaction and melting, PCL/TCP composite scaffolds were found to display uniform distribution of TCP throughout the PCL matrix regardless of composition. Homogeneous spatial distribution was also achieved in fabricated 3D scaffolds. When seeded onto powder-processed PCL/ TCP films, mesenchymal stem cells were found to undergo robust and uniform osteogenic differentiation, indicating the potential application of this approach to biofunctionalize scaffolds for tissue engineering applications.
J. Lim, Dr. M. S. K. Chong, Prof. S.-H. Teoh Division of Bioengineering School of Chemical and Biomedical Engineering 70 Nanyang Drive, Singapore 637457, Singapore E-mail: [email protected] Dr. M. S. K. Chong, Prof. J. K. Y. Chan Experimental Fetal Medicine Group Department of Obstetrics and Gynaecology Yong Loo Lin School of Medicine National University of Singapore and National University Hospital System 5 Lower Kent Ridge Road, Singapore 119074, Singapore Prof. J. K. Y. Chan Department of Reproductive Medicine KK Women’s and Children’s Hospital Singapore, 100 Bukit Timah Road, Singapore 229899, Singapore Prof. J. K. Y. Chan Cancer and Stem Cell Biology Duke-NUS Graduate Medical School Singapore, 8 College Road, Singapore 169857, Singapore DOI: 10.1002/smll.201302389 small 2014, DOI: 10.1002/smll.201302389
1. Introduction Synthetic polymers are commonly used in regenerative medicine, particularly as scaffold material in tissue engineering. A major limitation lies in the passive nature of these materials, necessitating functionalization through incorporation of bioactive molecules, in order to elicit desirable biological responses. This may be achieved by dispersing bioactive ingredients as a second phase in polymer solutions[1–3] or polymer melts.[4] The use of high-heat and solvents, however, results in denaturation of the biomolecules, as well as aggregation and phase separation of additives during blending, compromising both efficacy and consistency of results. Current research focuses on combining the use of reduced solvents with low-temperature mixing,[5–7] but typically involves a compromise between maximum loading ability and homogeneity of dispersion. Apart from the concerns over the use of potentially toxic solvents and its residuals, homogeneity of dispersion is essential to attain high mechanical performances from composite materials, as highlighted by Coroller et al. recently.[8]
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1
full papers
J. Lim et al.
Figure 1. SEM images of (A) PCL (B) cPCL (C) Thickness of cPCL (D) Particle size and SAV ratio of PCL and cPCL. Significant particle size reduction was recorded (as recorded by laser diffraction scattering), with a corresponding increase in SAV ratio. SEM images were taken at 25× (inset), 50×, and 200× magnifications.
To circumvent these issues, polymer and bioactive ingredients may be homogeneously mixed in powdered phase and subsequently fused under high pressure to desired shape and form via a solvent-free approach. Polymer powder processing has been previously described for low temperature processing of pharmaceutical coatings,[9] as well as tissue engineering scaffolds.[10] Here, finely-powdered polycaprolactone (PCL) matrices carrying well-dispersed biofunctional molecules were generated using cryomilling, a polymer powder processing technique. PCL is a commonly used polymer for tissue engineering[11–13] and drug delivery applications.[14] To apply this in an orthopedic setting, tricalcium phosphate was incorporated into the polymer matrix due to its relevance to physiological bone tissue.[15] Tricalcium phosphate (TCP) is a bioactive inorganic salt known to be extremely potent for stimulating bone formation,[16–18] and has been commonly incorporated in polymer-based scaffolds[19,20] for bone tissue engineering (BTE). Using conventional techniques however, poor rheological properties limit loading of TCP, with aggregation and delamination commonly occurring at higher ratios.[21] Here, we employed a low-temperature, solvent-free fabrication technique for the incorporation of higher percentages of TCP. This method produces a consistently homogeneous distribution of TCP within PCL over a wide range of compositions (up to 70 wt% TCP). In addition, due to the homogeneous distribution of TCP, we were able to induce a robust cellular response from highly osteogenic mesenchymal stem cells (MSCs).
2 www.small-journal.com
2. Results Our approach starts with the generation of polymeric powders using cryogenic grinding, or cryomilling. Cryomilling is the mechanical attrition of materials at cryogenic temperatures, which occurs due to repeated fracturing and welding by collision with high energy milling balls in a vial. Traditionally, this method has been employed on metallic systems for fabricating nanocrystalline metals and alloys.[22,23] Here, we show that PCL can be similarly treated (Figure 1) to generate powders with a median micron-particulate size of 254 µm from their original sizes (Figure 1B-D), with a 60-fold increase in surface area-to-volume ratio (cPCL: 2218; PCL: 36.9) (Figure 1D). The ability to achieve particle size reduction has been well-studied in literature.[24] The underlying mechanism lies in processing polymers in their glassy state. PCL has a reported Tg of −60 °C,[14] therefore by processing at cryogenic temperature, brittle fracture of PCL occurs. As a result, a significant change in particle size could be achieved in a very short time (33 times reduction in 20 min of continuous milling), in line with similar studies describing an exponential decrease in grain sizes of cryomilled poly(ethylene terephthalate) (PET).[25] The influences of cryomilling on polymer properties were then studied. Percentage crystallinity (Xc), taken as a ratio of the area under the thermogram against the theoretical 100% crystalline value of PCL (136 J/g[26] indicate crystallinity to be unaffected by cryomilling (Table 1, p > 0.05) and further implies that molecular chain packing and folding were not
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201302389
Polymer Powder Processing of Cryomilled Polycaprolactone for Solvent-free Generation of Homogeneous Bioactive
Table 1. Thermal properties of pristine PCL and cPCL suggested minimal changes in crystallinity and peak melting temperature (p > 0.05). Molecular weight and number, and polydispersity index changes were also monitored. Significant reduction (20%, *p < 0.05) in molecular weight was recorded. Xc (%)
Tm (°C)
PCL
62.3 ± 0.4
81.1 ± 0.7
118 000 ± 1030
66 900 ± 3570
1.72 ± 0.09
cPCL
60.9 ± 1.4
78.6 ± 0.9
92 300 ± 680,*
65 700 ± 690
1.40 ± 0.01
affected by cryomilling. Importantly, retaining the crystalline nature of PCL suggests that in general, strength of PCL is retained. With reference to our earlier report, the tensile strength and elastic moduli of PCL films lie within the range of cancellous bone,[27–29] which is desirable when considering stiffness matching of biomaterials. To address the question on effects of cryomilling on the molecular weight of PCL, it was found that the gel permeation chromatography profiles of both PCL and cPCL were similarly unimodal, but a 21.7% decrease in the peak Mw was observed in cPCL (Table 1, *: p < 0.05). Molecular number (Mn) recorded minimal changes (Table 1). Additionally, the polydispersity index (n) for both PCL and cPCL were similar, suggesting that the distributions of Mw/ Mn were not affected significantly. In their study, Smith et al. investigated the effects of cryomilling on the Mw of several polymers, including poly(methyl methacrylate) (PMMA), polyisoprene (PI) and poly(ethylene-alt-propylene) (PEP). Their results indicated exponential Mw decrease profiles for PMMA and PEP after cryomilling.[30] Our results are in good agreement, with a 21.7% decrease in Mw after cryomilling. From the representative thermographs, it can be observed that the peak melting temperatures, as observed by the vertical dotted lines, are similar for PCL and cPCL (81.1 °C vs 78.6 °C). It was also observed that, apart from the physical reduction in size, chain scission and/or chemical crosslinking might occur.[31] Our results here suggest that PCL underwent primarily scission without cross-linking, implying improved chain mobility and consequently, reduced viscosity. Rheological examination of the cPCL indeed showed significantly lowered inherent viscosity. Across the wide range of oscillation frequencies tested, the viscosity of cPCL was consistently 18% – 22% lower than that of non-cryomilled PCL (Table 2). Additionally, loss tangent data suggested better elastic property of cPCL, and supported the claim that the inherent viscosity of cPCL was improved (Table 2).
Molecular Weight (g/mol) Molecular Number (g/mol)
n
Based on the changes in Mw, which was approximately 21.7% lower after cryomilling (p < 0.05), the improvement in cPCL viscosity is in good agreement with the decrease in Mw, suggesting that viscosity is directly correlated to Mw. Taken together, these data suggest cPCL to be highly mobile in the melt solution, and consequently, contribute towards reduction of fusion defects during powder processing. By reducing the size disparity between the PCL and TCP particles, granular convective effects may be avoided and TCP is better dispersed.[32] Additionally, the improved rheological characteristics, as well as increased surface area to volume ratio[25,26] resulted in high processability with reduced fusion defects. This was demonstrated when PCL/ TCP powders were mixed using the following conditions and compared: cryomilling, and physical blending without cryomilling (Figure 2A, annotated as “control”). Under SEM, it could be observed that the TCP (distinguished as bright and round particles) agglomerated into isolated islands (indicated by arrows) in the control group. In contrast, the cryomilled PCL/TCP (65/35) (Figure 2B) presented a homogeneous distribution into the matrix of PCL (black arrows). Consequently, condensed film structures could be produced in a highly reproducible fashion, suggesting the lack of gross defects (Figure 2C, 2D, 2F). Salt leached scaffolds, suitable for bone tissue engineering, were also fabricated, and exhibit similarly uniform TCP dispersion through the polymer matrix. A comparison between films generated from native PCL and cryomilled PCL revealed the effect of moulding artifacts (Figure 2C, black arrow), suggesting impaired polymer flow, resulting in impaired structural integrity (Figure 2E). Conversely, continuous film structures could be obtained in the cryomilled PCL/TCP mix, with homogeneous distribution of TCP, whereas TCP particles were found to aggregate and separate in the control group, leading to delamination and spalling (Figure 2E, circled). Cross-sectional images of the films further illustrated this point, where agglomeration of
Table 2. Rheological properties of PCL and cPCL suggested improvement in cPCL viscosity, which was consistently 20% lower than PCL over the entire range of frequencies tested. Additionally, loss tangent data suggested the improved elastic response of cPCL, as evidenced by 15% – 20% decrease across most frequencies tested. Frequency (Hz)
0.20
0.40
0.80
1.59
3.18
6.34
η (PCL)
1456
1461
1441
1392
1302
1158
12.64 972
η (cPCL)
1143
1136
1123
1089
1033
938
800
η (cPCL)/η (PCL)
0.79
0.78
0.78
0.78
0.79
0.81
0.82
tanδ (PCL)
0.022
0.041
0.069
0.120
0.195
0.295
0.422
tanδ (cPCL)
0.021
0.032
0.058
0.099
0.164
0.253
0.375
tanδ (cPCL)/tanδ (PCL)
0.94
0.78
0.84
0.83
0.84
0.86
0.89
small 2014, DOI: 10.1002/smll.201302389
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
3
full papers
J. Lim et al.
Figure 2. SEM images representing (A) 65/35 control particles and (B) 65/35 particles are as shown. Clear agglomeration of TCP can be observed (black arrows) on (A), with a portion staying on PCL (circled region). 65/35 resulted in a good distribution of TCP in (B) (arrows). Gross morphology of (C) PCL (D) cPCL (E) 65/35 control (F) 65/35. Flow lines were visible in PCL (arrow), but were not as obvious in cPCL (arrow). SEM of cross-sections of the various films were conducted and represented in (G) – (I). Agglomeration of TCP could be observed in (H) 65/35 control, while (I) indicated a good distribution of TCP within PCL. PCL was represented by (G).
TCP particles could be seen in the control group (Figure 2H), while TCP particles were well-distributed in the cryomilled 65/35 film (Figure 2I). Thermogravimetric analysis (TGA) and micro-CT imaging (Figure 3) were then used to verify the distribution of TCP in the polymer matrix. In all cryomilled proportions of 75/25, 65/35, and 55/45 (Figure 3A), TGA analyses revealed measured PCL/TCP ratios of randomly-sized samples to be identical to predicted values, suggesting homogeneous distribution of the TCP through the polymeric phase. To demonstrate processability of the mixed powder and to be representative of scaffolds in BTE, porous blocks were also generated from cryomilled 75/25 (75/25 3D). When the films and scaffold were put through a micro-CT (Figure 3B), significant improvements in TCP distribution may be observed when compared to the control group. Notably, incorporation of physiological amounts of TCP (70 wt%) was successfully attained, and which also suggested homogeneous distribution of TCP. In the control group (65/35 control), less TCP was observed mainly due to the effect of spalling as shown earlier (Figure 2E, circled). Among the other test groups (75/25, 65/35, 55/45, 30/70), a homogeneous distribution of TCP could be observed. Here, the authors also fabricated a 3D scaffold (75/25 3D) using the same method, with the inclusion of salt particles that comprised 50 vol%. Following a salt leaching process in which all salt particles were removed, micro-CT results (Figure 3B, 75/25 3D) indicated the homogeneous distribution of TCP throughout the entire geometry, supporting
4 www.small-journal.com
our claim that cryomilling is able to achieve good particulate distribution in composite fabrication. In Demjén et al., interfacial interactions were found to play a significant role in modulating the mechanical properties of particulate-filled polymers.[33] The mechanical properties of PCL and PCL/TCP films are presented in Table 3 with representative stress-strain curves and a table summarizing yield stress, Young’s moduli, and yield strain. PCL and cPCL films indicated similar stress-strain curve profiles with similar yield stresses (PCL: 16.0 ± 1.4 vs. cPCL: 15.3 ± 0.7 MPa, p > 0.05). A lower Young’s modulus in cPCL (305 ± 24 MPa) resulted in an improvement in yield strain (PCL: 0.09 ± 0.02 vs. cPCL: 0.11 ± 0.01, *: p < 0.05). 65/35 films recorded lower yield stress as compared to PCL films, while Young’s modulus was significantly higher (492 ± 57 MPa, ^: p < 0.05). A corresponding decrease in yield strain accompanied the increase in Young’s modulus. No data was reported for the control group because the films were rendered unusable for meaningful tensile testing (Figure 2E). From our results, it is obvious that the reduced PCL particle size allowed better incorporation of TCP, verifying the work previously mentioned.[33] An interesting observation was the improved elastic strain by cPCL. From the rheological results obtained earlier (Table 2), the lowered loss tangent of cPCL could have played a possible role in the manifestation of improved elasticity (Table 3, *: p < 0.05) and possibly tensile strength, although the reduced tensile
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201302389
Polymer Powder Processing of Cryomilled Polycaprolactone for Solvent-free Generation of Homogeneous Bioactive
Figure 3. (A) TGA analyses of various PCL/TCP films. (B) Micro-CT evaluation of various PCL and PCL/TCP (up to 70 wt%). Better TCP distribution was observed in all cryomilled groups (75/25, 65/35, 55/45, 30/70) as compared to 65/35 control. Spatial distribution was suggested in 75/25 3D. Table 3. Summary of the mechanical characteristics of the various PCL films. Tensile strength of PCL and cPCL remained comparable while improvement in yield strain was achieved in cPCL (*: p < 0.05). 65/35 showed increased stiffness (^: p < 0.05), lower tensile strength and yield strain than both PCL and cPCL (#,*: p < 0.05). No meaningful data (ND) was obtained for 65/35 control. Composition
Yield Stress (MPa)
Young’s Modulus (MPa)
Yield Strain
PCL
16.0 ± 1.4
356 ± 66
0.09 ± 0.02
cPCL 65/35 65/35 control
15.3 ± 0.7
305 ± 24
0.11 ± 0.01,*
9.32 ± 0.9, #
492 ± 57, ^
0.032 ± 0.002,*
ND
ND
ND
strength in cPCL did not prove to be statistically meaningful. Importantly, we have demonstrated the possibility of tailoring the modulus of PCL/TCP by varying the composition of TCP. In order to substantiate our point that homogeneous distribution of bioactive TCP may lead to robust cellular responses, human fetal bone marrow-derived stem cells (hfMSCs), which are being studied for bone tissue engineering[34] were seeded on PCL films, 65/35 films or 65/35 control films and subsequently chemically-induced to form bone (Figure 4). Due to the extremely high osteogenic potential of these cells,[35] significant differentiation could be induced regardless of the substrate used. However, hfMSC mineralization was most pronounced and homogeneous in 65/35 (Figure 4A). 65/35 control films had uneven mineralization, which could be due to the uneven distribution of TCP. Additionally, the expression of late-stage osteogenic marker osteocalcin (OCN) was found to be small 2014, DOI: 10.1002/smll.201302389
expressed homogeneously on 65/35 films, while that of 65/35 control films resulted in hfMSC clustering and preferential differentiation around the TCP regions (Figure 4B, darker regions on 65/35 control). Taken together, we have shown that hfMSCs were able to differentiate homogeneously on 65/35 films, while non-cryomilled 65/35 control films resulted in preferential differentiation centered on the TCP regions (Figure 4).
3. Conclusion In this work, we have demonstrated the use of solvent-free, low-temperature polymer powder processing techniques to overcome limitations on distribution and loading of biologically active molecules. Further, a method was developed to reproducibly generate scaffolds containing biologically active TCP at physiologically relevant compositions (up to 70 wt%), with superior homogeneity over existing methods. Our findings will facilitate the development of tissue scaffolds with increased bioactive molecules leading to more robust or greater range of biological functions.
4. Experimental Section PCL (Osteopore International, Singapore) was used as received. All other materials were purchased from Sigma-Aldrich unless otherwise stated. Cryomilling: PCL/TCP composites were fabricated using cryomilling (Retsch®, Germany). PCL/TCP in three proportions (75/25, 65/35, 55/45) were weighed using a microbalance (Mettler Toledo,
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
5
full papers
J. Lim et al.
Figure 4. (A) von Kossa assessment of hfMSCs on PCL and PCL/TCP films in bone media (BM) and growth media (D10). Mineralization was observed in all groups in BM, with only 65/35 showing homogeneous mineralization over the entire film surface. Scale bar represents 500 µm. (B) Immunohistochemical (IHC) staining of osteocalcin (OCN) was conducted after 14 days of osteogenic incubation. hfMSC clustering and OCN expression was highly localized on the TCP regions (darkest regions) in 65/35 control, while hfMSC distribution and OCN expression was homogeneous on the 65/35 films. Images were taken at 4× magnification. Scale bar represents 250 µm.
USA), and loaded into the cryogenic vial with a ball-to-mass ratio of 30:1. The cryomilling protocol was: 6 – 8 min of pre-cooling in liquid nitrogen and 20 min of continuous milling for one cycle. Fabrication of PCL/TCP films and 3D scaffold: PCL/TCP films were thermally pressed into films of thickness approximately 30 – 60 µm. Briefly, a known mass of PCL/TCP was placed between two stainless steel sheets on the Carver system (Carver Inc, USA). Temperature was elevated to 100 °C and pressure added for 30 min. The pressed film was then allowed to cool to room temperature via normal convection cooling. 75/25 3D scaffold was fabricated using the same method, with the incorporation of 50 vol% of sodium chloride, which was subsequently salt-leached in water. Particle and Film Morphological analyses – Scanning Electron Microscopy: The morphologies of PCL, cryomilled-PCL (cPCL), and PCL/TCP composites were observed under the scanning electron microscope (SEM). The samples were platinum coated for 60 sec at 20 mA using the JFC-1600 Autofine Coater (JEOL, Japan) before
6 www.small-journal.com
loading into the chamber of the SEM 6390LA (JEOL, Japan). An acceleration voltage of 5 kV was applied with a working distance of 14 mm. Particle analysis was quantified using a laser diffraction system that measures the intensity of light scattered as the laser passes through a particulate sample (Malvern Mastersizer 2000 system, Malvern Instruments, USA). Molecular Weight – Gel Permeation Chromatography: Gel permeation chromatography (GPC) was performed by Supramolecular Biomaterials Lab (National University of Singapore, Singapore) using a Shimadzu SIL-10A and LC-20AD system equipped with two Phenogel 5µ 100 and 10000 Å columns (size: 300 × 4.6 mm) connected in series and a Shimadzu RID-10A refractive index detector. A solution of each sample was prepared using tetrahydrofuran (THF) as a solvent. 1.5 ml THF was added to approximately 5 mg of the sample and at least 30 mins was allowed to dissolve the sample completely. The solutions were then thoroughly mixed and filtered through a 0.45 µm phobic polytetrafluoroethylene filter. The mobile phase flow-rate used was 0.3 ml/min at a temperature
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201302389
Polymer Powder Processing of Cryomilled Polycaprolactone for Solvent-free Generation of Homogeneous Bioactive
of 40 °C. The data was collected and analyzed using LCsolution 1.22 software (Shimadzu, Japan). Mw and Mn of the PCL were calibrated using mono-dispersed polystyrene standard samples (n = 3). X-Ray Micro Computed Tomography (Micro-CT): High-resolution, qualitative visualization of TCP distribution in PCL matrix was conducted by micro-CT imaging (Skyscan in vivo microtomograph, 1076, Belgium). The films were placed on a sample holder and scanned through 180° at a spatial resolution of 9 µm. The image data from the scanned planes are subsequently thresholded and reconstructed to create 3-D images to visualize TCP distribution. Rheology: The rheological properties of PCL and PCL/TCP were investigated using the AR 2000ex rheometer (TA instruments, USA). Calibration of the equipment was conducted according to manufacturer’s instruction. A parallel plate configuration (25 mm steel plate, 0o) was chosen, and all tests were carried out at 100 °C. A frequency of 10 Hz was applied after a pre-test frequency sweep (1 – 500 Hz). The gap was maintained between 700 – 900 µm, and a normal contact force of 1.0 ± 0.5 N was maintained throughout. 30 sec was given to allow equilibration of the system, and 1 min was given to allow stabilization of the temperature. Mechanical Testing: The mechanical properties of the fabricated PCL and PCL/TCP films were tested on an Instron tester (Instron 5582, Norwood, MA) at a cross-head moving speed of 2 mm/min. The results presented were average values from six independent measurements (n = 6). Thermal Properties – Differential Scanning Calorimetry: The thermal properties (n = 3) were investigated using the Differential Scanning Calorimetry (DSC) (Diamond DSC, PerkinElmer, USA). The samples were heated from room temperature to 150 °C at a rate of 20 K/min. Crystallinity (Xc) was calculated by taking a ratio of the experimentally measured enthalpy heat of melting to the theoretical value of 100% crystalline PCL (136 J/g.[36] Results are presented as an average of three readings and their respective standard deviations. Thermogravimetric Analysis: Thermogravimetric analysis (TGA) (n = 3) was conducted using the TGA/DSC Mettler Toledo (Mettler Toledo, USA). PCL/TCP samples were heated at a rate of 20 K/min from room temperature until 450 °C, and held isotherm for 5 min. Samples and Ethics: Fetal tissue collection was conducted according to Zhang et al.,[37] which was approved by the Domain Specific Review Board of National University Hospital (D06/54), Singapore, in compliance with international guidelines regarding the use of fetal tissue for research.[38] Pregnant women gave separate written consent for the clinical procedure and for the use of fetal tissue for research purposes. Fetal tissues were collected after clinically-indicated termination of pregnancy. Fetal gestational age was determined by crown-rump length measurement. In this study, a fetal sample at 14+2 (weeks + days) gestations was utilized. Cell Source: hfMSCs (Passage 7) were isolated as previously described.[39] Cells were seeded in a flask (75 ml, Nunc, Rochester) with a cell density of 106/ml in DMEM (10% Fetal Bovine Serum/1% Penicillin-Streptomycin). Non-adherent cells were removed with the first medium change on day 3. Adherent hfMSCs were recovered from primary culture after one week. hfMSC Seeding on Films: PCL and PCL/TCP films were attached to 12 mm diameter glass coverslips soaked in 70% ethanol for 24 hours. Subsequently, the films were washed twice with phosphate buffer saline (PBS) and adherent hfMSCs (Passage 7) from small 2014, DOI: 10.1002/smll.201302389
primary culture were seeded with a density of 15,000 cells/cm2. Samples were kept in culture for 14 days. Immunohistochemical Staining of Osteocalcin and von Kossa: Immunohistochemical (IHC) assay was performed to investigate the expression level of osteocalcin (OCN), a late-stage differentiation marker. Cells were fixed, penetrated and blocked as reported in Wang et al. Briefly, hfMSC films were first fixed in 4% paraformaldehyde and treated with 0.1% Triton X-100 each for 15 mins, before treatment in 1% bovine serum albumin (BSA) for 24 hours. For antibody labeling, samples were first incubated with primary rabbit antibody (1:100, Abcam) followed by secondary goat anti-rabbit IgG conjugation to OCN (1:100, Molecular probes). 4′-6-diamidino-2-phenylindole (DAPI) was added last before mounting onto glass coverslips for detection of fluorescence emission at 488 nm using fluorescence light microscopy (CKX-41, Olympus). To ensure consistency, images were taken using the same parameters across all groups, at 4× magnification. For von Kossa, hfMSC films were gently rinsed twice with PBS, before being fixed in 10% formalin for 5 min, and subsequently washed twice with water. Finally, they were stained with freshly made 2% silver nitrate in water (w/v) for 10 min in the dark and exposed to strong light for 15 min. Statistical Analysis: The data obtained were statistically analyzed using the unpaired two-tailed student’s t-test to determine statistical significance with a confidence interval of 95%. Results with p < 0.05 were considered significant.
Acknowledgements The authors of this work would like to express their gratitude to Dr. Zhang Qinyuan (NUS) for her help in conducting the GPC tests, and Dr. Chang Lei (NUS) for her help in running the micro-CT. Special thanks to Mr. Jason Feng (NUS) for his assistance in providing IHC stains. Financial assistance was provided by the NTU internal funding.
[1] H. Cao, N. Kuboyama, Bone 2010, 46, 386–395. [2] S. K. Misra, T. I. Ansari, S. P. Valappil, D. Mohn, S. E. Philip, W. J. Stark, I. Roy, J. C. Knowles, V. Salih, A. R. Boccaccini, Biomaterials 2010, 31, 2806–2815. [3] Y. Liu, J. Lim, S.-H. Teoh, Biotechnol. Adv. 2013, 31, 688– 705. [4] I. Zein, D. W. Hutmacher, K. C. Tan, S. H. Teoh, Biomaterials 2002, 23, 1169–1185. [5] A. Rakovsky, I. Gotman, E. Rabkin, E. Y. Gutmanas, J. Mech. Behav. Biomed. Mater. 2013, 18, 37–46. [6] C. Makarov, I. Gotman, X. Jiang, S. Fuchs, C. J. Kirkpatrick, E. Y. Gutmanas, J. Mater. Sci.: Mater. Med. 2010, 21, 1771– 1779. [7] M. Bernstein, I. Gotman, C. Makarov, A. Phadke, S. Radin, P. Ducheyne, E. Y. Gutmanas, Adv. Eng. Mater. 2010, 12, B341–B347. [8] G. Coroller, A. Lefeuvre, A. Le Duigou, A. Bourmaud, G. Ausias, T. Gaudry, C. Baley, Composites, Part A 2013, 51, 62–70. [9] S. Bose, R. H. Bogner, Pharm. Dev. Technol. 2007, 12, 115–131.
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
7
full papers
J. Lim et al.
[10] F. E. Wiria, K. F. Leong, C. K. Chua, Rapid Prototyping J. 2010, 16, 400–410. [11] J. H. Ashton, J. A. M. Mertz, J. L. Harper, M. J. Slepian, J. L. Mills, D. V. McGrath, J. P. Vande Geest, Acta Biomater. 2011, 7, 287–294. [12] S. C. Neves, L. S. Moreira Teixeira, L. Moroni, R. L. Reis, C. A. Van Blitterswijk, N. M. Alves, M. Karperien, J. F. Mano, Biomaterials 2011, 32, 1068–1079. [13] K. W. Ng, H. N. Achuth, S. Moochhala, T. C. Lim, D. W. Hutmacher, J. Biomater. Sci. Polym. Ed. 2007, 18, 925–938. [14] M. A. Woodruff, D. W. Hutmacher, Prog. Polym. Sci. 2010, 35, 1217–1256. [15] T. Dvir, B. P. Timko, D. S. Kohane, R. Langer, Nat. Nanotechnol. 2010, 6, 13–22. [16] N. Fujita, T. Matsushita, K. Ishida, K. Sasaki, S. Kubo, T. Matsumoto, M. Kurosaka, Y. Tabata, R. Kuroda, J. Tissue Eng. Regener. Med. 2012, 6, 291–298. [17] K. Thakare, M. Bhongade, P. Charde, P. Jaiswal, N. Shah, A. Deshpande, Case Reports in Dentistry 2013, 4. [18] K. Busuttil Naudi, D. Lappin, J. McMahon, A. Ayoub, L. Di Silvio, J. Appl. Biomater. Biomech. 2012, 40, e461. [19] S. Ozkan, D. M. Kalyon, X. Yu, J. Biomed. Mater. Res., Part A 2010, 92, 1007–1018. [20] J.-P. Chen, M.-J. Tsai, H.-T. Liao, Colloids Surf., B 2013. [21] M. G. Yeo, G. H. Kim, Chem. Mater. 2011, 24, 903–913. [22] K. S. Kumar, H. Van Swygenhoven, S. Suresh, Acta Mater. 2003, 51, 5743–5774. [23] E. J. Lavernia, B. Q. Han, J. M. Schoenung, Mater. Sci. Eng., A 2008, 493, 207–214. [24] A. P. Smith, H. Ade, C. C. Koch, S. D. Smith, R. J. Spontak, Macromolecules 2000, 33, 1163–1172. [25] Y.-g. Zhu, Z.-Q. Li, D. Zhang, T. Tanimoto, J. Polym. Sci., Part B: Polym. Phys. 2006, 44, 986–993. [26] Y. He, Y. Inoue, Polym. Int. 2000, 49, 623–626.
8 www.small-journal.com
[27] J. Lim, M. S. Chong, E. Y. Teo, G. Q. Chen, J. K. Chan, S. H. Teoh, J. Biomed. Mater. Res. B Appl. Biomater. 2013, 101, 752–761. [28] G. Perego, G. D. Cella, C. Bastioli, J. Appl. Polym. Sci. 1996, 59, 37–43. [29] Z. Y. Wang, E. Y. Teo, M. S. Chong, Q. Y. Zhang, J. Lim, Z. Y. Zhang, M. H. Hong, E. S. Thian, J. K. Chan, S. H. Teoh, Tissue Eng., Part C 2013, 4, 4. [30] A. P. Smith, J. S. Shayb, R. J. Spontaka, C. M. Balika, H. Adec, S. D. Smith, C. C. Kocha, Polymer 2000, 41, 6271–6283. [31] A. P. Smith, H. Ade, C. M. Balik, C. C. Koch, S. D. Smith, R. J. Spontak, Macromolecules 2000, 33, 2595–2604. [32] G. Marosi, G. Bertalan, P. Anna, I. Rusznak, J. Polym. Eng. 1993, 12, 33–62. [33] Z. Demjén, B. Pukánszky, J. Nagy, Composites, Part A 1998, 29, 323–329. [34] Z. Y. Zhang, S. H. Teoh, M. S. K. Chong, E. S. M. Lee, L. G. Tan, C. N. Mattar, N. M. Fisk, M. Choolani, J. Chan, Biomaterials 2010, 31, 608–620. [35] Z.-Y. Zhang, S.-H. Teoh, M. S. K. Chong, J. T. Schantz, N. M. Fisk, M. A. Choolani, J. Chan, Stem Cells 2009, 27, 126–137. [36] Q. Guo, C. Harrats, G. Groeninckx, H. Reynaers, M. H. J. Koch, Polymer. 2001, 42, 6031–6041. [37] Z.-Y. Zhang, S. H. Teoh, E. Y. Teo, M. S. Khoon Chong, C. W. Shin, F. T. Tien, M. A. Choolani, J. K. Y. Chan, Biomaterials 2010, 31, 8684–8695. [38] G. Britainand, J. Polkinghorne, National government publication, CM series, London: HM Stationary Office 1989. [39] M. S. Chong, J. Chan, Methods Mol Biol. 2010, 614, 135–147.
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Received: August 5, 2013 Revised: December 26, 2013 Published online:
small 2014, DOI: 10.1002/smll.201302389
Polymer Powder Processing of Cryomilled Polycaprolactone for Solvent-free Generation of Homogeneous Bioactive Tissue Engineering Scaffolds Jing Lim, Mark Seow Khoon Chong, Jerry Kok Yen Chan, and Swee-Hin Teoh*
Synthetic
polymers used in tissue engineering require functionalization with bioactive molecules to elicit specific physiological reactions. These additives must be homogeneously dispersed in order to achieve enhanced composite mechanical performance and uniform cellular response. This work demonstrated the use of a solvent-free powder processing technique to form osteoinductive scaffolds from cryomilled polycaprolactone (PCL) and tricalcium phosphate (TCP). Cryomilling was performed to achieve micron-sized distribution of PCL and reduce melt viscosity, thus improving TCP distribution and improving structural integrity. A breakthrough was achieved in the successful fabrication of 70 weight percentage of TCP into a continuous film structure. Following compaction and melting, PCL/TCP composite scaffolds were found to display uniform distribution of TCP throughout the PCL matrix regardless of composition. Homogeneous spatial distribution was also achieved in fabricated 3D scaffolds. When seeded onto powder-processed PCL/ TCP films, mesenchymal stem cells were found to undergo robust and uniform osteogenic differentiation, indicating the potential application of this approach to biofunctionalize scaffolds for tissue engineering applications.
J. Lim, Dr. M. S. K. Chong, Prof. S.-H. Teoh Division of Bioengineering School of Chemical and Biomedical Engineering 70 Nanyang Drive, Singapore 637457, Singapore E-mail: [email protected] Dr. M. S. K. Chong, Prof. J. K. Y. Chan Experimental Fetal Medicine Group Department of Obstetrics and Gynaecology Yong Loo Lin School of Medicine National University of Singapore and National University Hospital System 5 Lower Kent Ridge Road, Singapore 119074, Singapore Prof. J. K. Y. Chan Department of Reproductive Medicine KK Women’s and Children’s Hospital Singapore, 100 Bukit Timah Road, Singapore 229899, Singapore Prof. J. K. Y. Chan Cancer and Stem Cell Biology Duke-NUS Graduate Medical School Singapore, 8 College Road, Singapore 169857, Singapore DOI: 10.1002/smll.201302389 small 2014, DOI: 10.1002/smll.201302389
1. Introduction Synthetic polymers are commonly used in regenerative medicine, particularly as scaffold material in tissue engineering. A major limitation lies in the passive nature of these materials, necessitating functionalization through incorporation of bioactive molecules, in order to elicit desirable biological responses. This may be achieved by dispersing bioactive ingredients as a second phase in polymer solutions[1–3] or polymer melts.[4] The use of high-heat and solvents, however, results in denaturation of the biomolecules, as well as aggregation and phase separation of additives during blending, compromising both efficacy and consistency of results. Current research focuses on combining the use of reduced solvents with low-temperature mixing,[5–7] but typically involves a compromise between maximum loading ability and homogeneity of dispersion. Apart from the concerns over the use of potentially toxic solvents and its residuals, homogeneity of dispersion is essential to attain high mechanical performances from composite materials, as highlighted by Coroller et al. recently.[8]
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1
full papers
J. Lim et al.
Figure 1. SEM images of (A) PCL (B) cPCL (C) Thickness of cPCL (D) Particle size and SAV ratio of PCL and cPCL. Significant particle size reduction was recorded (as recorded by laser diffraction scattering), with a corresponding increase in SAV ratio. SEM images were taken at 25× (inset), 50×, and 200× magnifications.
To circumvent these issues, polymer and bioactive ingredients may be homogeneously mixed in powdered phase and subsequently fused under high pressure to desired shape and form via a solvent-free approach. Polymer powder processing has been previously described for low temperature processing of pharmaceutical coatings,[9] as well as tissue engineering scaffolds.[10] Here, finely-powdered polycaprolactone (PCL) matrices carrying well-dispersed biofunctional molecules were generated using cryomilling, a polymer powder processing technique. PCL is a commonly used polymer for tissue engineering[11–13] and drug delivery applications.[14] To apply this in an orthopedic setting, tricalcium phosphate was incorporated into the polymer matrix due to its relevance to physiological bone tissue.[15] Tricalcium phosphate (TCP) is a bioactive inorganic salt known to be extremely potent for stimulating bone formation,[16–18] and has been commonly incorporated in polymer-based scaffolds[19,20] for bone tissue engineering (BTE). Using conventional techniques however, poor rheological properties limit loading of TCP, with aggregation and delamination commonly occurring at higher ratios.[21] Here, we employed a low-temperature, solvent-free fabrication technique for the incorporation of higher percentages of TCP. This method produces a consistently homogeneous distribution of TCP within PCL over a wide range of compositions (up to 70 wt% TCP). In addition, due to the homogeneous distribution of TCP, we were able to induce a robust cellular response from highly osteogenic mesenchymal stem cells (MSCs).
2 www.small-journal.com
2. Results Our approach starts with the generation of polymeric powders using cryogenic grinding, or cryomilling. Cryomilling is the mechanical attrition of materials at cryogenic temperatures, which occurs due to repeated fracturing and welding by collision with high energy milling balls in a vial. Traditionally, this method has been employed on metallic systems for fabricating nanocrystalline metals and alloys.[22,23] Here, we show that PCL can be similarly treated (Figure 1) to generate powders with a median micron-particulate size of 254 µm from their original sizes (Figure 1B-D), with a 60-fold increase in surface area-to-volume ratio (cPCL: 2218; PCL: 36.9) (Figure 1D). The ability to achieve particle size reduction has been well-studied in literature.[24] The underlying mechanism lies in processing polymers in their glassy state. PCL has a reported Tg of −60 °C,[14] therefore by processing at cryogenic temperature, brittle fracture of PCL occurs. As a result, a significant change in particle size could be achieved in a very short time (33 times reduction in 20 min of continuous milling), in line with similar studies describing an exponential decrease in grain sizes of cryomilled poly(ethylene terephthalate) (PET).[25] The influences of cryomilling on polymer properties were then studied. Percentage crystallinity (Xc), taken as a ratio of the area under the thermogram against the theoretical 100% crystalline value of PCL (136 J/g[26] indicate crystallinity to be unaffected by cryomilling (Table 1, p > 0.05) and further implies that molecular chain packing and folding were not
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201302389
Polymer Powder Processing of Cryomilled Polycaprolactone for Solvent-free Generation of Homogeneous Bioactive
Table 1. Thermal properties of pristine PCL and cPCL suggested minimal changes in crystallinity and peak melting temperature (p > 0.05). Molecular weight and number, and polydispersity index changes were also monitored. Significant reduction (20%, *p < 0.05) in molecular weight was recorded. Xc (%)
Tm (°C)
PCL
62.3 ± 0.4
81.1 ± 0.7
118 000 ± 1030
66 900 ± 3570
1.72 ± 0.09
cPCL
60.9 ± 1.4
78.6 ± 0.9
92 300 ± 680,*
65 700 ± 690
1.40 ± 0.01
affected by cryomilling. Importantly, retaining the crystalline nature of PCL suggests that in general, strength of PCL is retained. With reference to our earlier report, the tensile strength and elastic moduli of PCL films lie within the range of cancellous bone,[27–29] which is desirable when considering stiffness matching of biomaterials. To address the question on effects of cryomilling on the molecular weight of PCL, it was found that the gel permeation chromatography profiles of both PCL and cPCL were similarly unimodal, but a 21.7% decrease in the peak Mw was observed in cPCL (Table 1, *: p < 0.05). Molecular number (Mn) recorded minimal changes (Table 1). Additionally, the polydispersity index (n) for both PCL and cPCL were similar, suggesting that the distributions of Mw/ Mn were not affected significantly. In their study, Smith et al. investigated the effects of cryomilling on the Mw of several polymers, including poly(methyl methacrylate) (PMMA), polyisoprene (PI) and poly(ethylene-alt-propylene) (PEP). Their results indicated exponential Mw decrease profiles for PMMA and PEP after cryomilling.[30] Our results are in good agreement, with a 21.7% decrease in Mw after cryomilling. From the representative thermographs, it can be observed that the peak melting temperatures, as observed by the vertical dotted lines, are similar for PCL and cPCL (81.1 °C vs 78.6 °C). It was also observed that, apart from the physical reduction in size, chain scission and/or chemical crosslinking might occur.[31] Our results here suggest that PCL underwent primarily scission without cross-linking, implying improved chain mobility and consequently, reduced viscosity. Rheological examination of the cPCL indeed showed significantly lowered inherent viscosity. Across the wide range of oscillation frequencies tested, the viscosity of cPCL was consistently 18% – 22% lower than that of non-cryomilled PCL (Table 2). Additionally, loss tangent data suggested better elastic property of cPCL, and supported the claim that the inherent viscosity of cPCL was improved (Table 2).
Molecular Weight (g/mol) Molecular Number (g/mol)
n
Based on the changes in Mw, which was approximately 21.7% lower after cryomilling (p < 0.05), the improvement in cPCL viscosity is in good agreement with the decrease in Mw, suggesting that viscosity is directly correlated to Mw. Taken together, these data suggest cPCL to be highly mobile in the melt solution, and consequently, contribute towards reduction of fusion defects during powder processing. By reducing the size disparity between the PCL and TCP particles, granular convective effects may be avoided and TCP is better dispersed.[32] Additionally, the improved rheological characteristics, as well as increased surface area to volume ratio[25,26] resulted in high processability with reduced fusion defects. This was demonstrated when PCL/ TCP powders were mixed using the following conditions and compared: cryomilling, and physical blending without cryomilling (Figure 2A, annotated as “control”). Under SEM, it could be observed that the TCP (distinguished as bright and round particles) agglomerated into isolated islands (indicated by arrows) in the control group. In contrast, the cryomilled PCL/TCP (65/35) (Figure 2B) presented a homogeneous distribution into the matrix of PCL (black arrows). Consequently, condensed film structures could be produced in a highly reproducible fashion, suggesting the lack of gross defects (Figure 2C, 2D, 2F). Salt leached scaffolds, suitable for bone tissue engineering, were also fabricated, and exhibit similarly uniform TCP dispersion through the polymer matrix. A comparison between films generated from native PCL and cryomilled PCL revealed the effect of moulding artifacts (Figure 2C, black arrow), suggesting impaired polymer flow, resulting in impaired structural integrity (Figure 2E). Conversely, continuous film structures could be obtained in the cryomilled PCL/TCP mix, with homogeneous distribution of TCP, whereas TCP particles were found to aggregate and separate in the control group, leading to delamination and spalling (Figure 2E, circled). Cross-sectional images of the films further illustrated this point, where agglomeration of
Table 2. Rheological properties of PCL and cPCL suggested improvement in cPCL viscosity, which was consistently 20% lower than PCL over the entire range of frequencies tested. Additionally, loss tangent data suggested the improved elastic response of cPCL, as evidenced by 15% – 20% decrease across most frequencies tested. Frequency (Hz)
0.20
0.40
0.80
1.59
3.18
6.34
η (PCL)
1456
1461
1441
1392
1302
1158
12.64 972
η (cPCL)
1143
1136
1123
1089
1033
938
800
η (cPCL)/η (PCL)
0.79
0.78
0.78
0.78
0.79
0.81
0.82
tanδ (PCL)
0.022
0.041
0.069
0.120
0.195
0.295
0.422
tanδ (cPCL)
0.021
0.032
0.058
0.099
0.164
0.253
0.375
tanδ (cPCL)/tanδ (PCL)
0.94
0.78
0.84
0.83
0.84
0.86
0.89
small 2014, DOI: 10.1002/smll.201302389
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
3
full papers
J. Lim et al.
Figure 2. SEM images representing (A) 65/35 control particles and (B) 65/35 particles are as shown. Clear agglomeration of TCP can be observed (black arrows) on (A), with a portion staying on PCL (circled region). 65/35 resulted in a good distribution of TCP in (B) (arrows). Gross morphology of (C) PCL (D) cPCL (E) 65/35 control (F) 65/35. Flow lines were visible in PCL (arrow), but were not as obvious in cPCL (arrow). SEM of cross-sections of the various films were conducted and represented in (G) – (I). Agglomeration of TCP could be observed in (H) 65/35 control, while (I) indicated a good distribution of TCP within PCL. PCL was represented by (G).
TCP particles could be seen in the control group (Figure 2H), while TCP particles were well-distributed in the cryomilled 65/35 film (Figure 2I). Thermogravimetric analysis (TGA) and micro-CT imaging (Figure 3) were then used to verify the distribution of TCP in the polymer matrix. In all cryomilled proportions of 75/25, 65/35, and 55/45 (Figure 3A), TGA analyses revealed measured PCL/TCP ratios of randomly-sized samples to be identical to predicted values, suggesting homogeneous distribution of the TCP through the polymeric phase. To demonstrate processability of the mixed powder and to be representative of scaffolds in BTE, porous blocks were also generated from cryomilled 75/25 (75/25 3D). When the films and scaffold were put through a micro-CT (Figure 3B), significant improvements in TCP distribution may be observed when compared to the control group. Notably, incorporation of physiological amounts of TCP (70 wt%) was successfully attained, and which also suggested homogeneous distribution of TCP. In the control group (65/35 control), less TCP was observed mainly due to the effect of spalling as shown earlier (Figure 2E, circled). Among the other test groups (75/25, 65/35, 55/45, 30/70), a homogeneous distribution of TCP could be observed. Here, the authors also fabricated a 3D scaffold (75/25 3D) using the same method, with the inclusion of salt particles that comprised 50 vol%. Following a salt leaching process in which all salt particles were removed, micro-CT results (Figure 3B, 75/25 3D) indicated the homogeneous distribution of TCP throughout the entire geometry, supporting
4 www.small-journal.com
our claim that cryomilling is able to achieve good particulate distribution in composite fabrication. In Demjén et al., interfacial interactions were found to play a significant role in modulating the mechanical properties of particulate-filled polymers.[33] The mechanical properties of PCL and PCL/TCP films are presented in Table 3 with representative stress-strain curves and a table summarizing yield stress, Young’s moduli, and yield strain. PCL and cPCL films indicated similar stress-strain curve profiles with similar yield stresses (PCL: 16.0 ± 1.4 vs. cPCL: 15.3 ± 0.7 MPa, p > 0.05). A lower Young’s modulus in cPCL (305 ± 24 MPa) resulted in an improvement in yield strain (PCL: 0.09 ± 0.02 vs. cPCL: 0.11 ± 0.01, *: p < 0.05). 65/35 films recorded lower yield stress as compared to PCL films, while Young’s modulus was significantly higher (492 ± 57 MPa, ^: p < 0.05). A corresponding decrease in yield strain accompanied the increase in Young’s modulus. No data was reported for the control group because the films were rendered unusable for meaningful tensile testing (Figure 2E). From our results, it is obvious that the reduced PCL particle size allowed better incorporation of TCP, verifying the work previously mentioned.[33] An interesting observation was the improved elastic strain by cPCL. From the rheological results obtained earlier (Table 2), the lowered loss tangent of cPCL could have played a possible role in the manifestation of improved elasticity (Table 3, *: p < 0.05) and possibly tensile strength, although the reduced tensile
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201302389
Polymer Powder Processing of Cryomilled Polycaprolactone for Solvent-free Generation of Homogeneous Bioactive
Figure 3. (A) TGA analyses of various PCL/TCP films. (B) Micro-CT evaluation of various PCL and PCL/TCP (up to 70 wt%). Better TCP distribution was observed in all cryomilled groups (75/25, 65/35, 55/45, 30/70) as compared to 65/35 control. Spatial distribution was suggested in 75/25 3D. Table 3. Summary of the mechanical characteristics of the various PCL films. Tensile strength of PCL and cPCL remained comparable while improvement in yield strain was achieved in cPCL (*: p < 0.05). 65/35 showed increased stiffness (^: p < 0.05), lower tensile strength and yield strain than both PCL and cPCL (#,*: p < 0.05). No meaningful data (ND) was obtained for 65/35 control. Composition
Yield Stress (MPa)
Young’s Modulus (MPa)
Yield Strain
PCL
16.0 ± 1.4
356 ± 66
0.09 ± 0.02
cPCL 65/35 65/35 control
15.3 ± 0.7
305 ± 24
0.11 ± 0.01,*
9.32 ± 0.9, #
492 ± 57, ^
0.032 ± 0.002,*
ND
ND
ND
strength in cPCL did not prove to be statistically meaningful. Importantly, we have demonstrated the possibility of tailoring the modulus of PCL/TCP by varying the composition of TCP. In order to substantiate our point that homogeneous distribution of bioactive TCP may lead to robust cellular responses, human fetal bone marrow-derived stem cells (hfMSCs), which are being studied for bone tissue engineering[34] were seeded on PCL films, 65/35 films or 65/35 control films and subsequently chemically-induced to form bone (Figure 4). Due to the extremely high osteogenic potential of these cells,[35] significant differentiation could be induced regardless of the substrate used. However, hfMSC mineralization was most pronounced and homogeneous in 65/35 (Figure 4A). 65/35 control films had uneven mineralization, which could be due to the uneven distribution of TCP. Additionally, the expression of late-stage osteogenic marker osteocalcin (OCN) was found to be small 2014, DOI: 10.1002/smll.201302389
expressed homogeneously on 65/35 films, while that of 65/35 control films resulted in hfMSC clustering and preferential differentiation around the TCP regions (Figure 4B, darker regions on 65/35 control). Taken together, we have shown that hfMSCs were able to differentiate homogeneously on 65/35 films, while non-cryomilled 65/35 control films resulted in preferential differentiation centered on the TCP regions (Figure 4).
3. Conclusion In this work, we have demonstrated the use of solvent-free, low-temperature polymer powder processing techniques to overcome limitations on distribution and loading of biologically active molecules. Further, a method was developed to reproducibly generate scaffolds containing biologically active TCP at physiologically relevant compositions (up to 70 wt%), with superior homogeneity over existing methods. Our findings will facilitate the development of tissue scaffolds with increased bioactive molecules leading to more robust or greater range of biological functions.
4. Experimental Section PCL (Osteopore International, Singapore) was used as received. All other materials were purchased from Sigma-Aldrich unless otherwise stated. Cryomilling: PCL/TCP composites were fabricated using cryomilling (Retsch®, Germany). PCL/TCP in three proportions (75/25, 65/35, 55/45) were weighed using a microbalance (Mettler Toledo,
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
5
full papers
J. Lim et al.
Figure 4. (A) von Kossa assessment of hfMSCs on PCL and PCL/TCP films in bone media (BM) and growth media (D10). Mineralization was observed in all groups in BM, with only 65/35 showing homogeneous mineralization over the entire film surface. Scale bar represents 500 µm. (B) Immunohistochemical (IHC) staining of osteocalcin (OCN) was conducted after 14 days of osteogenic incubation. hfMSC clustering and OCN expression was highly localized on the TCP regions (darkest regions) in 65/35 control, while hfMSC distribution and OCN expression was homogeneous on the 65/35 films. Images were taken at 4× magnification. Scale bar represents 250 µm.
USA), and loaded into the cryogenic vial with a ball-to-mass ratio of 30:1. The cryomilling protocol was: 6 – 8 min of pre-cooling in liquid nitrogen and 20 min of continuous milling for one cycle. Fabrication of PCL/TCP films and 3D scaffold: PCL/TCP films were thermally pressed into films of thickness approximately 30 – 60 µm. Briefly, a known mass of PCL/TCP was placed between two stainless steel sheets on the Carver system (Carver Inc, USA). Temperature was elevated to 100 °C and pressure added for 30 min. The pressed film was then allowed to cool to room temperature via normal convection cooling. 75/25 3D scaffold was fabricated using the same method, with the incorporation of 50 vol% of sodium chloride, which was subsequently salt-leached in water. Particle and Film Morphological analyses – Scanning Electron Microscopy: The morphologies of PCL, cryomilled-PCL (cPCL), and PCL/TCP composites were observed under the scanning electron microscope (SEM). The samples were platinum coated for 60 sec at 20 mA using the JFC-1600 Autofine Coater (JEOL, Japan) before
6 www.small-journal.com
loading into the chamber of the SEM 6390LA (JEOL, Japan). An acceleration voltage of 5 kV was applied with a working distance of 14 mm. Particle analysis was quantified using a laser diffraction system that measures the intensity of light scattered as the laser passes through a particulate sample (Malvern Mastersizer 2000 system, Malvern Instruments, USA). Molecular Weight – Gel Permeation Chromatography: Gel permeation chromatography (GPC) was performed by Supramolecular Biomaterials Lab (National University of Singapore, Singapore) using a Shimadzu SIL-10A and LC-20AD system equipped with two Phenogel 5µ 100 and 10000 Å columns (size: 300 × 4.6 mm) connected in series and a Shimadzu RID-10A refractive index detector. A solution of each sample was prepared using tetrahydrofuran (THF) as a solvent. 1.5 ml THF was added to approximately 5 mg of the sample and at least 30 mins was allowed to dissolve the sample completely. The solutions were then thoroughly mixed and filtered through a 0.45 µm phobic polytetrafluoroethylene filter. The mobile phase flow-rate used was 0.3 ml/min at a temperature
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201302389
Polymer Powder Processing of Cryomilled Polycaprolactone for Solvent-free Generation of Homogeneous Bioactive
of 40 °C. The data was collected and analyzed using LCsolution 1.22 software (Shimadzu, Japan). Mw and Mn of the PCL were calibrated using mono-dispersed polystyrene standard samples (n = 3). X-Ray Micro Computed Tomography (Micro-CT): High-resolution, qualitative visualization of TCP distribution in PCL matrix was conducted by micro-CT imaging (Skyscan in vivo microtomograph, 1076, Belgium). The films were placed on a sample holder and scanned through 180° at a spatial resolution of 9 µm. The image data from the scanned planes are subsequently thresholded and reconstructed to create 3-D images to visualize TCP distribution. Rheology: The rheological properties of PCL and PCL/TCP were investigated using the AR 2000ex rheometer (TA instruments, USA). Calibration of the equipment was conducted according to manufacturer’s instruction. A parallel plate configuration (25 mm steel plate, 0o) was chosen, and all tests were carried out at 100 °C. A frequency of 10 Hz was applied after a pre-test frequency sweep (1 – 500 Hz). The gap was maintained between 700 – 900 µm, and a normal contact force of 1.0 ± 0.5 N was maintained throughout. 30 sec was given to allow equilibration of the system, and 1 min was given to allow stabilization of the temperature. Mechanical Testing: The mechanical properties of the fabricated PCL and PCL/TCP films were tested on an Instron tester (Instron 5582, Norwood, MA) at a cross-head moving speed of 2 mm/min. The results presented were average values from six independent measurements (n = 6). Thermal Properties – Differential Scanning Calorimetry: The thermal properties (n = 3) were investigated using the Differential Scanning Calorimetry (DSC) (Diamond DSC, PerkinElmer, USA). The samples were heated from room temperature to 150 °C at a rate of 20 K/min. Crystallinity (Xc) was calculated by taking a ratio of the experimentally measured enthalpy heat of melting to the theoretical value of 100% crystalline PCL (136 J/g.[36] Results are presented as an average of three readings and their respective standard deviations. Thermogravimetric Analysis: Thermogravimetric analysis (TGA) (n = 3) was conducted using the TGA/DSC Mettler Toledo (Mettler Toledo, USA). PCL/TCP samples were heated at a rate of 20 K/min from room temperature until 450 °C, and held isotherm for 5 min. Samples and Ethics: Fetal tissue collection was conducted according to Zhang et al.,[37] which was approved by the Domain Specific Review Board of National University Hospital (D06/54), Singapore, in compliance with international guidelines regarding the use of fetal tissue for research.[38] Pregnant women gave separate written consent for the clinical procedure and for the use of fetal tissue for research purposes. Fetal tissues were collected after clinically-indicated termination of pregnancy. Fetal gestational age was determined by crown-rump length measurement. In this study, a fetal sample at 14+2 (weeks + days) gestations was utilized. Cell Source: hfMSCs (Passage 7) were isolated as previously described.[39] Cells were seeded in a flask (75 ml, Nunc, Rochester) with a cell density of 106/ml in DMEM (10% Fetal Bovine Serum/1% Penicillin-Streptomycin). Non-adherent cells were removed with the first medium change on day 3. Adherent hfMSCs were recovered from primary culture after one week. hfMSC Seeding on Films: PCL and PCL/TCP films were attached to 12 mm diameter glass coverslips soaked in 70% ethanol for 24 hours. Subsequently, the films were washed twice with phosphate buffer saline (PBS) and adherent hfMSCs (Passage 7) from small 2014, DOI: 10.1002/smll.201302389
primary culture were seeded with a density of 15,000 cells/cm2. Samples were kept in culture for 14 days. Immunohistochemical Staining of Osteocalcin and von Kossa: Immunohistochemical (IHC) assay was performed to investigate the expression level of osteocalcin (OCN), a late-stage differentiation marker. Cells were fixed, penetrated and blocked as reported in Wang et al. Briefly, hfMSC films were first fixed in 4% paraformaldehyde and treated with 0.1% Triton X-100 each for 15 mins, before treatment in 1% bovine serum albumin (BSA) for 24 hours. For antibody labeling, samples were first incubated with primary rabbit antibody (1:100, Abcam) followed by secondary goat anti-rabbit IgG conjugation to OCN (1:100, Molecular probes). 4′-6-diamidino-2-phenylindole (DAPI) was added last before mounting onto glass coverslips for detection of fluorescence emission at 488 nm using fluorescence light microscopy (CKX-41, Olympus). To ensure consistency, images were taken using the same parameters across all groups, at 4× magnification. For von Kossa, hfMSC films were gently rinsed twice with PBS, before being fixed in 10% formalin for 5 min, and subsequently washed twice with water. Finally, they were stained with freshly made 2% silver nitrate in water (w/v) for 10 min in the dark and exposed to strong light for 15 min. Statistical Analysis: The data obtained were statistically analyzed using the unpaired two-tailed student’s t-test to determine statistical significance with a confidence interval of 95%. Results with p < 0.05 were considered significant.
Acknowledgements The authors of this work would like to express their gratitude to Dr. Zhang Qinyuan (NUS) for her help in conducting the GPC tests, and Dr. Chang Lei (NUS) for her help in running the micro-CT. Special thanks to Mr. Jason Feng (NUS) for his assistance in providing IHC stains. Financial assistance was provided by the NTU internal funding.
[1] H. Cao, N. Kuboyama, Bone 2010, 46, 386–395. [2] S. K. Misra, T. I. Ansari, S. P. Valappil, D. Mohn, S. E. Philip, W. J. Stark, I. Roy, J. C. Knowles, V. Salih, A. R. Boccaccini, Biomaterials 2010, 31, 2806–2815. [3] Y. Liu, J. Lim, S.-H. Teoh, Biotechnol. Adv. 2013, 31, 688– 705. [4] I. Zein, D. W. Hutmacher, K. C. Tan, S. H. Teoh, Biomaterials 2002, 23, 1169–1185. [5] A. Rakovsky, I. Gotman, E. Rabkin, E. Y. Gutmanas, J. Mech. Behav. Biomed. Mater. 2013, 18, 37–46. [6] C. Makarov, I. Gotman, X. Jiang, S. Fuchs, C. J. Kirkpatrick, E. Y. Gutmanas, J. Mater. Sci.: Mater. Med. 2010, 21, 1771– 1779. [7] M. Bernstein, I. Gotman, C. Makarov, A. Phadke, S. Radin, P. Ducheyne, E. Y. Gutmanas, Adv. Eng. Mater. 2010, 12, B341–B347. [8] G. Coroller, A. Lefeuvre, A. Le Duigou, A. Bourmaud, G. Ausias, T. Gaudry, C. Baley, Composites, Part A 2013, 51, 62–70. [9] S. Bose, R. H. Bogner, Pharm. Dev. Technol. 2007, 12, 115–131.
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
7
full papers
J. Lim et al.
[10] F. E. Wiria, K. F. Leong, C. K. Chua, Rapid Prototyping J. 2010, 16, 400–410. [11] J. H. Ashton, J. A. M. Mertz, J. L. Harper, M. J. Slepian, J. L. Mills, D. V. McGrath, J. P. Vande Geest, Acta Biomater. 2011, 7, 287–294. [12] S. C. Neves, L. S. Moreira Teixeira, L. Moroni, R. L. Reis, C. A. Van Blitterswijk, N. M. Alves, M. Karperien, J. F. Mano, Biomaterials 2011, 32, 1068–1079. [13] K. W. Ng, H. N. Achuth, S. Moochhala, T. C. Lim, D. W. Hutmacher, J. Biomater. Sci. Polym. Ed. 2007, 18, 925–938. [14] M. A. Woodruff, D. W. Hutmacher, Prog. Polym. Sci. 2010, 35, 1217–1256. [15] T. Dvir, B. P. Timko, D. S. Kohane, R. Langer, Nat. Nanotechnol. 2010, 6, 13–22. [16] N. Fujita, T. Matsushita, K. Ishida, K. Sasaki, S. Kubo, T. Matsumoto, M. Kurosaka, Y. Tabata, R. Kuroda, J. Tissue Eng. Regener. Med. 2012, 6, 291–298. [17] K. Thakare, M. Bhongade, P. Charde, P. Jaiswal, N. Shah, A. Deshpande, Case Reports in Dentistry 2013, 4. [18] K. Busuttil Naudi, D. Lappin, J. McMahon, A. Ayoub, L. Di Silvio, J. Appl. Biomater. Biomech. 2012, 40, e461. [19] S. Ozkan, D. M. Kalyon, X. Yu, J. Biomed. Mater. Res., Part A 2010, 92, 1007–1018. [20] J.-P. Chen, M.-J. Tsai, H.-T. Liao, Colloids Surf., B 2013. [21] M. G. Yeo, G. H. Kim, Chem. Mater. 2011, 24, 903–913. [22] K. S. Kumar, H. Van Swygenhoven, S. Suresh, Acta Mater. 2003, 51, 5743–5774. [23] E. J. Lavernia, B. Q. Han, J. M. Schoenung, Mater. Sci. Eng., A 2008, 493, 207–214. [24] A. P. Smith, H. Ade, C. C. Koch, S. D. Smith, R. J. Spontak, Macromolecules 2000, 33, 1163–1172. [25] Y.-g. Zhu, Z.-Q. Li, D. Zhang, T. Tanimoto, J. Polym. Sci., Part B: Polym. Phys. 2006, 44, 986–993. [26] Y. He, Y. Inoue, Polym. Int. 2000, 49, 623–626.
8 www.small-journal.com
[27] J. Lim, M. S. Chong, E. Y. Teo, G. Q. Chen, J. K. Chan, S. H. Teoh, J. Biomed. Mater. Res. B Appl. Biomater. 2013, 101, 752–761. [28] G. Perego, G. D. Cella, C. Bastioli, J. Appl. Polym. Sci. 1996, 59, 37–43. [29] Z. Y. Wang, E. Y. Teo, M. S. Chong, Q. Y. Zhang, J. Lim, Z. Y. Zhang, M. H. Hong, E. S. Thian, J. K. Chan, S. H. Teoh, Tissue Eng., Part C 2013, 4, 4. [30] A. P. Smith, J. S. Shayb, R. J. Spontaka, C. M. Balika, H. Adec, S. D. Smith, C. C. Kocha, Polymer 2000, 41, 6271–6283. [31] A. P. Smith, H. Ade, C. M. Balik, C. C. Koch, S. D. Smith, R. J. Spontak, Macromolecules 2000, 33, 2595–2604. [32] G. Marosi, G. Bertalan, P. Anna, I. Rusznak, J. Polym. Eng. 1993, 12, 33–62. [33] Z. Demjén, B. Pukánszky, J. Nagy, Composites, Part A 1998, 29, 323–329. [34] Z. Y. Zhang, S. H. Teoh, M. S. K. Chong, E. S. M. Lee, L. G. Tan, C. N. Mattar, N. M. Fisk, M. Choolani, J. Chan, Biomaterials 2010, 31, 608–620. [35] Z.-Y. Zhang, S.-H. Teoh, M. S. K. Chong, J. T. Schantz, N. M. Fisk, M. A. Choolani, J. Chan, Stem Cells 2009, 27, 126–137. [36] Q. Guo, C. Harrats, G. Groeninckx, H. Reynaers, M. H. J. Koch, Polymer. 2001, 42, 6031–6041. [37] Z.-Y. Zhang, S. H. Teoh, E. Y. Teo, M. S. Khoon Chong, C. W. Shin, F. T. Tien, M. A. Choolani, J. K. Y. Chan, Biomaterials 2010, 31, 8684–8695. [38] G. Britainand, J. Polkinghorne, National government publication, CM series, London: HM Stationary Office 1989. [39] M. S. Chong, J. Chan, Methods Mol Biol. 2010, 614, 135–147.
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Received: August 5, 2013 Revised: December 26, 2013 Published online:
small 2014, DOI: 10.1002/smll.201302389
