fabrication and characterization of 3d printed poly ... phuc_dang_thesis.pdf · fabrication and...
TRANSCRIPT
FABRICATION AND CHARACTERIZATION OF 3D PRINTED
POLY(CAPROLACTONE) SCAFFOLDS WITH BIMODAL
POROSITY FOR LOCAL DRUG DELIVERY AND TISSUE
RECONSTRUCTION APPLICATIONS
Hoang Phuc Dang
B.Eng. (Biomedical Engineering)
Submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
School of Chemistry, Physics and Mechanical Engineering
Science and Engineering Faculty
Queensland University of Technology
2020
Keywords
Screw melt extrusion, porogen leaching, bimodal porosity, tissue reconstruction, local
drug delivery, chemotherapy
Abstract
Tissue reconstruction of large/critical‐size defects requires scaffolds that are able to
support tissue regeneration. In addition, tissue inflammation and disease recurrence are major
challenges after implantation. Therefore, there is an urgent need to develop new types of
scaffolds able to deliver drugs locally to prevent inflammation and disease recurrence as well as
support tissue reconstruction.
This PhD project focused on the fabrication and characterisation of a new type of three‐
dimensional (3D) scaffold having macroscale porosity generated from the printing pattern and
intra‐strut microscale porosity to be used as a dual‐function scaffold for local drug delivery and
tissue reconstruction. The applications of the scaffold for local drug delivery as well as tissue
regeneration were demonstrated in vitro and in vivo.
The fabrication of bimodal porous scaffolds utilising screw melt extrusion and porogen
leaching techniques was described. Physicochemical and in vitro characterisation demonstrated
that the bimodal porous scaffold had greater surface area and faster degradation compared with
conventional 3D‐printed scaffolds without microscale porosity.
The application of the bimodal porous scaffold in drug delivery was characterized. The
scaffold demonstrated different release profiles in vitro when loaded with drugs having different
charge and hydrophilic/hydrophobic properties. Drugs with more charge had more interaction
with the scaffold; hence, burst release was reduced and release was sustained. Conversely,
uncharged hydrophilic drugs had less interaction with the scaffold, and hence released faster and
might benefit from hydrogel coating. The antimicrobial effects and chemotherapeutic effects of
drug–loaded scaffolds were demonstrated, and showed that antibiotic–loaded scaffolds and
chemotherapeutic agent–loaded scaffolds maintained their bioactivity after fabrication and
could be used for infection prevention and anti‐cancer applications.
The bimodal porous scaffold was further investigated as a local drug implant in a mouse
orthotopic breast cancer model. Results demonstrated that the local treatment slowed down
tumour recurrence progression and reduced systemic cytotoxicity compared with a conventional
systemic administration method. In vivo and ex vivo bioluminescence imaging demonstrated that
doxorubicin (DOX)–loaded scaffolds reduced tumour growth compared with the control and
intravenous injection (I.V.) treatment. Weight monitoring confirmed that DOX–loaded scaffolds
had less cytotoxicity compared with I.V. treatment. Immunohistological staining was performed
to validate the chemotherapeutic effect and evaluate the cytotoxicity of the treatment. Overall,
DOX–loaded scaffolds were shown to be a potential substitute for systemic administration of
chemotherapeutic agents to improve treatment efficiency and reduce systemic toxicity.
Last, the application of the bimodal porous scaffold in tissue regeneration demonstrated
its ability to support new bone formation at the defect area in a rat critical‐size calvarial defect
model. The scaffold demonstrated a similar level of new bone formation and inflammatory
response to biomimetic calcium phosphate (CaP)‐coated scaffolds, which was confirmed by
micro‐computed tomography scanning and immunohistochemical staining. Further, the porous
scaffold was shown to have additional space inside struts for extracellular substance infiltration,
which could support and provide additional space for tissue formation.
In summary, in my PhD studies, I investigated the fabrication and validation of a new type
of 3D‐printed scaffold with macro/microscale porous structure for local drug delivery and tissue
reconstruction application. The scaffold was proved to provide superior treatment for breast
cancer in a mouse model compared with a systemic administration method as it used lower
quantities of drugs, reduced tumour recurrence and reduced toxicity. The scaffold was also able
to enhance tissue reconstruction in vivo in a rat model.
List of Publications
Peer‐reviewed journal articles related to this thesis
1. Luke E. Visscher*, Hoang Phuc Dang*, Mark A. Knackstedt, Dietmar W. Hutmacher,
and Phong A. Tran. "3D printed Polycaprolactone scaffolds with dual macro‐
microporosity for applications in local delivery of antibiotics." Materials Science and
Engineering: C 87 (2018): 78‐89. * Contributed equally.
2. Hoang Phuc Dang, Tara Shabab, Abbas Shafiee, Quentin C. Peiffer, Kate Fox, Nhiem
Tran, Tim R Dargaville, Dietmar Werner Hutmacher and Phong A Tran. “3D printed
dual macro‐, microscale porous network as a tissue engineering scaffold with drug
delivering function” Biofabrication 11 (2019): number 3.
3. Hoang Phuc Dang, Cedryck Vaquette, Tara Shabab, Román A. Pérez, Ying Yang, Tim
R. Dargaville, Dietmar W. Hutmacher, Abbas Shafiee, Phong A. Tran. “3D printed
scaffolds with dual macro‐, micro‐scale porous networks for bone tissue
regeneration” (submitted to Biomaterials Journal)
4. Hoang Phuc Dang, Abbas Shafiee, Tim R. Dargaville, Christoph A. Lahr and Phong A.
Tran. “Delivery of doxorubicin from the implanted bimodal porous
polycaprolactone scaffolds decreased systemic cytotoxicity and breast tumour
recurrence progression in mice” (under supervisor revision and targeted to submit
to Advanced Functional Materials Journal)
Peer‐reviewed journal articles unrelated to this thesis
5. Phong A. Tran, Hiep T. Nguyen, Philip J. Hubbard, Hoang Phuc Dang, and Dietmar W.
Hutmacher. "Mineralization of plasma treated polymer surfaces from super‐saturated
simulated body fluids." Materials Letters 230 (2018): 12‐15.
Peer‐reviewed book chapter unrelated to this thesis
6. Mina Mohseni, Nathan J. Castro, Hoang Phuc Dang, Tan Dat Nguyen, Hieu Minh Ho, Minh
Phuong Nam Tran, Thi Hiep Nguyen, and Phong A. Tran. "Adipose tissue regeneration:
Scaffold—Biomaterial strategies and translational perspectives." In Biomaterials in
Translational Medicine, pp. 291‐330. Academic Press, 2019.
Conference presentations (Oral and Poster)
7. Hoang Phuc Dang, Phong A Tran and Dietmar W Hutmacher. “The use of 3D printed
microporous‐strut polycaprolactone scaffolds for targeted local delivery of
chemotherapeutic agent for breast cancer application” The 7th International
Conference in Vietnam on the Development of Biomedical Engineering 2018. (oral
presentation)
8. Hoang Phuc Dang, Tara Shabab, Abbas Shafiee, Cedryck Vaquette, Ying Yang, Tim R.
Dargaville, Dietmar W. Hutmacher and Phong A. Tran. “Biomimetics multiscale
porous scaffolds for bone tissue engineering” the Biomimetics in Bioengineering
Conference 2019. (interactive poster presentation)
9. Hoang Phuc Dang, Abbas Shafiee, Dietmar W. Hutmacher and Phong A. Tran. “Local
delivery of doxorubicin loaded scaffolds decreasing breast tumor progression and
systemic cytotoxicity” the TERMIS‐AP + ABMC7 2019 Congress. (oral presentation)
10. Hoang Phuc Dang, Cedryck Vaquette, Abbas Shafiee, Dietmar W. Hutmacher and
Phong A. Tran. “Dual Scale Macro/micro Porous Scaffolds for Bone Tissue
Regeneration” the TERMIS‐AP + ABMC7 2019 Congress. (poster presentation)
Table of Contents
Keywords ................................................................................................................................................. ii
Abstract .................................................................................................................................................. iii
List of Publications .................................................................................................................................. v
Table of Contents .................................................................................................................................. vii
List of Figures .......................................................................................................................................... x
List of Tables ......................................................................................................................................... xvi
List of Supplementary Figures ............................................................................................................. xvii
List of Supplementary Tables ............................................................................................................... xix
List of Abbreviations .............................................................................................................................. xx
Statement of Original Authorship ...................................................................................................... xxiii
Acknowledgments .............................................................................................................................. xxiv
Note to the reader ................................................................................................................................... i
1 CHAPTER 1: INTRODUCTION ....................................................................................................... 2
1.1 Background .................................................................................................................................. 2
1.2 Literature Review ......................................................................................................................... 3 1.2.1 3D printing technology in biomedical application ............................................................ 3 1.2.2 Porosity in 3D printed scaffolds ........................................................................................ 4 1.2.3 Bimodal porosity ............................................................................................................... 5 1.2.4 3D printed polycaprolactone in tissue reconstruction application and local drug
delivery application .......................................................................................................... 7 1.2.4.1 Additive manufactured 3D polycaprolactone ................................................................... 7 1.2.4.2 3D printed polycaprolactone in tissue reconstruction application .................................. 7 1.2.4.3 3D printed polycaprolactone in local drug delivery application ....................................... 8
1.3 Research Gap ............................................................................................................................... 8
1.4 HypothesIs and objectives ......................................................................................................... 10
1.5 Scope of study and thesis Outline .............................................................................................. 10
2 CHAPTER 2: 3D PRINTED POLYCAPROLACTONE SCAFFOLDS WITH DUAL MACRO‐MICROPOROSITY FOR APPLICATIONS IN LOCAL DELIVERY OF ANTIBIOTICS .............................................................. 13
2.1 Abstract ...................................................................................................................................... 16
2.2 Introduction ............................................................................................................................... 16
2.3 Materials and Methods .............................................................................................................. 19
2.4 Results ........................................................................................................................................ 26 2.4.1 Scaffold manufacturing and characterization................................................................. 26 2.4.2 Degradation test. ............................................................................................................ 30 2.4.3 Loading and release of antibiotics and its bioactivity. .................................................... 32 2.4.4 Biocompatibility testing .................................................................................................. 37
2.5 Discussion .................................................................................................................................. 38
2.6 Conclusion .................................................................................................................................. 42
3 CHAPTER 3: 3D PRINTED DUAL MACRO‐, MICROSCALE POROUS NETWORK AS A TISSUE ENGINEERING SCAFFOLD WITH DRUG DELIVERING FUNCTION...................................................... 44
3.1 Abstract ...................................................................................................................................... 47
3.2 Introduction ............................................................................................................................... 48
3.3 Materials and Methods .............................................................................................................. 51 3.3.1 Scaffold fabrication ......................................................................................................... 51 3.3.2 Scaffold characterization ................................................................................................ 52 3.3.3 Drug loading, characterization and in vitro release ........................................................ 53 3.3.4 Antimicrobial assay ......................................................................................................... 55 3.3.5 In vitro chemotherapeutic effect .................................................................................... 55 3.3.6 Statistical analysis ........................................................................................................... 56
3.4 Results ........................................................................................................................................ 56 3.4.1 Material preparation, extrusion and scaffold characterization ...................................... 56 3.4.2 Micropores partially reduced compression Young’s modulus of scaffolds .................... 58 3.4.3 In vitro drug release ........................................................................................................ 59 3.4.4 Loading and release of doxorubicin as a water‐soluble and positively charged drug .... 59 3.4.5 Loading and release of paclitaxel as model insoluble non‐charged drug ....................... 63 3.4.6 Loading and release of Cefazolin as model water‐soluble, negatively charged drug –
PLGA overcoat can lead to sustained release ................................................................. 66 3.4.7 In vitro bioactivity of drug–loaded scaffolds .................................................................. 69 3.4.7.1 Antimicrobial effect of Cefazolin–loaded scaffolds ........................................................ 69 3.4.7.2 Effect of pMPCL scaffolds loaded with chemotherapeutic drugs against cancer cells ... 70
3.5 Discussion .................................................................................................................................. 72
3.6 Conclusion .................................................................................................................................. 76
4 CHAPTER 4: DELIVERY OF DOXORUBICIN FROM THE IMPLANTED BIMODAL POROUS POLYCAPROLACTONE SCAFFOLDS DECREASED SYSTEMIC CYTOTOXICITY AND BREAST TUMOUR RECURRENCE PROGRESSION IN MICE ........................................................................................... 77
4.1 Abstract ...................................................................................................................................... 80
4.2 Introduction ............................................................................................................................... 81
4.3 Materials and Methods .............................................................................................................. 84 4.3.1 Scaffold fabrication ......................................................................................................... 84 4.3.2 Drug loading ................................................................................................................... 84 4.3.3 In vitro chemotherapeutic effect of doxorubicin–loaded scaffolds against breast cancer
85 4.3.4 In vitro drug release ........................................................................................................ 86 4.3.5 Application of doxorubicin–loaded scaffolds in preventing breast cancer recurrence
and metastasis ................................................................................................................ 86 4.3.5.1 Cell culture ...................................................................................................................... 86 4.3.5.2 Tumour establishment.................................................................................................... 86 4.3.5.3 Treatment ....................................................................................................................... 87 4.3.5.4 In vivo drug release in blood ........................................................................................... 88 4.3.5.5 Histology and immunohistochemistry staining .............................................................. 88 4.3.5.6 Statistical analysis ........................................................................................................... 89
4.4 Results ........................................................................................................................................ 89 4.4.1 Scaffold characterization ................................................................................................ 89 4.4.2 Drug loading ................................................................................................................... 91 4.4.3 In vitro chemotherapeutic effect of doxorubicin–loaded scaffolds against breast cancer
92 4.4.4 Drug release behaviour of doxorubicin–loaded bimodal scaffolds ................................ 92 4.4.5 Application of doxorubicin–loaded scaffolds in preventing breast cancer recurrence
and metastasis ................................................................................................................ 93
4.4.5.1 Doxorubicin concentration in peripheral blood ............................................................. 93 4.4.5.2 Cytotoxicity of doxorubicin–loaded scaffolds. ................................................................ 93 4.4.5.3 Chemotherapeutic effect of doxorubicin–loaded scaffolds ........................................... 97
4.5 Discussion ................................................................................................................................ 102
4.6 Conclusion ................................................................................................................................ 106
5 CHAPTER 5: 3D PRINTED SCAFFOLDS WITH DUAL MACRO‐, MICRO‐SCALE POROUS NETWORKS FOR BONE TISSUE REGENERATION ............................................................................................. 108
5.1 Abstract .................................................................................................................................... 111
5.2 Introduction ............................................................................................................................. 111
5.3 Materials and Methods ............................................................................................................ 113 5.3.1 Preparation and characterization of the porogen, mPCL films and mPCL‐porogen films
113 5.3.2 Fabrication and characterization of mPCL scaffolds ..................................................... 114 5.3.3 Physicochemical characterization................................................................................. 116 5.3.4 Tensile strength characterization ................................................................................. 117 5.3.5 Accelerated degradation .............................................................................................. 118 5.3.6 In vitro hematoma model ............................................................................................. 118 5.3.7 Cell culture study .......................................................................................................... 119 5.3.8 Protein adsorption ........................................................................................................ 120 5.3.9 Cell proliferation ........................................................................................................... 120 5.3.10 In vivo study .................................................................................................................. 121 5.3.11 Statistical analysis ......................................................................................................... 122
5.4 Results ...................................................................................................................................... 123 5.4.1 Scaffold manufacturing ................................................................................................ 123 5.4.2 Porogen leaching .......................................................................................................... 124 5.4.3 Physicochemical properties .......................................................................................... 124 5.4.4 Tensile strength ............................................................................................................ 127 5.4.5 Microscale porosity accelerated degradation of mPCL scaffolds ................................. 127 5.4.6 In vitro hematoma model ............................................................................................. 130 5.4.7 Multiphasic scaffolds enhanced protein adsorption in vitro ......................................... 132 5.4.8 Multiphasic scaffolds had better support for cell proliferation in vitro ........................ 133 5.4.9 In vivo study .................................................................................................................. 133
5.5 Discussion ................................................................................................................................ 139
5.6 Conclusion ................................................................................................................................ 143
6 CHAPTER 6: CONCLUSION AND FUTURE DIRECTIONS ............................................................... 144
6.1 Conclusion ................................................................................................................................ 144
6.2 Limitations and Future Directions ............................................................................................ 145 6.2.1 Chapter 2 ...................................................................................................................... 146 6.2.2 Chapter 3 ...................................................................................................................... 147 6.2.3 Chapter 4 ...................................................................................................................... 147 6.2.4 Chapter 5 ...................................................................................................................... 147
APPENDICES .............................................................................................................................. 149
BIBLIOGRAPHY .......................................................................................................................... 160
List of Figures
Figure 1. (a) home‐built screw melt exutrsion device. (b) Schematic of fabrication process of PCL scaffolds using SME and salt leaching techniques. ......................................................................................... 20
Figure 2. Schematic of accelerated degradation experiment .................................................................... 23
Figure 3. Scaffold characterization. Overview images (insets) and SEM images of scaffolds at increasing magnifications for microporous scaffolds ((a1)‐(a3)), normal scaffolds (solid strut, (b1)‐(b3)) and GelMA coated microporous scaffolds ((c1)‐(c3)) and GelMA coated solid scaffolds ((d1)‐(d3)). Scale bars = 500 μm for (a1)‐(d1); 200 μm for (a2)‐(d2); 50 μm for (a3)‐(d3). Micro CT images of e) microporous scaffold and f) solid scaffold at two representative cross sectional planes. Microporosity analyzed by measuring in 2D slice images from micro CT is c.a. 41%. Circular non‐interconnected pore present in the solid scaffolds (f) are believed to result from air entrapment during the extrusion process. g) The leaching process of mPCL over 14 days indicating complete leaching of porogen was achieved. Scaffolds were dissolved in chloroform to make solutions of 10 %( w/v). Any porogen remained in the scaffold would give rise to the solution absorbance which was measured with a spectrophotometer. ............................................................................................ 27
Figure 4. Physicochemical characterization of microporous mPCL scaffolds compared with mPCL scaffolds showing mean + SEM for: a)& b) AFM imaging and surface roughness analysis showing non‐significant difference in topography; c) contact angles demonstrating similar hydrophilicity as expected from surface hydrolysis of mPCL during leaching d) porosity determined using scaffold dimensions, weight and mPCL density (n=4). Total porosity was significantly higher for microporous scaffolds at 80±1% compared with 62±0.3% for the solid scaffolds. e) surface areas measured with gas adsorption and analyzed with B.E.T method was also significantly higher for the microporous group (data is average of 2 measurements) and f) compressive modulus (MPa) over three compressing ‐ relax cycles (n=3) showing non‐significantly lower values for microporous scaffolds. g) Mean Zeta potential for solid and microporous scaffolds showing both having negative surface charges with no significant difference as expected from the same post‐printing treatment procedure to both scaffolds. (* p < 0.05). ........................................................................................................ 29
Figure 5. Accelerated degradation test. mPCL and microporous mPCL scaffolds were immersed in NaOH 2M in a shaker incubator (at 37oC, 121 rpm) for 48 hours. SEM imaging revealed the substantial increase in porosity both on the surface and cross section of the microporous scaffold (scale bars = 100 μm). (e): GPC analysis showed results of molecular weight changes after the test; mPCL scaffolds exhibited slight increase in Mw likely due to the removal of amorphous/less crystalline region during the degradation from the scaffolds. The reduction in Mw of microporous scaffolds is attributed to the bulk erosion as seen in SEM images on the surface and cross section of the scaffolds. ....................................................................................................................................... 31
Figure 6. Characterization of Cefazolin release kinetics from scaffolds. [A]: comparison of Cefazolin non‐cumulative release profiles for GelMA coated microporous mPCL and solid mPCL over the first 6 hours and next 7 days reported as mean±SEM (n=9) a) The released Cefazolin concentrations showed dose dependence in both normal and microporous scaffolds. The microporous scaffolds also showed consistently higher release up to day 6 (comparing within the same dose group). b) The total Cefazolin amount released at each time point (in μg) was consistently higher for microporous compared with normal scaffolds for 3 consecutive days. [B]. Comparison of Cefazolin release profiles in PBS for microporous mPCL scaffolds with and without GelMA coating (n=9) at high loading dose a) total release from scaffolds with a reduced burst release for GelMA coated scaffolds which released significantly less in the first 6 hours and b) release at each time point presented as percentage of total released, with the GelMA coated scaffolds showing a more gradual release indicating a reduced burst release. ................................................................................................................... 33
Figure 7. Antimicrobial activity of Cefazolin on S. aureus. [A] Effects of Cefazolin on planktonic S. aureus growth. a) Characterization of S. aureus growth demonstrating a typical growth curve of the bacteria including an initial lag phase (3 hours) followed by log growth phase (8 hours) (n=3). The stationary phase was not tested in this experiment. Bacteria in the log growth phase were then used for subsequent antibiotic testing experiments. b) Growth curves of S. aureus treated with increasing concentrations of Cefazolin, demonstrating the MIC50 = 0.1 ppm and MIC90=0.8 ppm where 50% and 80% of bacterial growth was inhibited respectively (n=3) [15]. [B] Bioactivity of Cefazolin eluates collected at different time points from microporous mPCL scaffolds loaded with Low, Medium and High Dose of Cefazolin determined as CFU normalized by that of untreated control. Data = mean ± SEM (N=3 repeats with 3 replicates in each repeat). The eluates demonstrated dose‐dependent antimicrobial activity from which MIC90 of eluted Cefazolin was calculated to be c.a 0.7 to 1.3 ppm, and an MIC50 of c.a. 0.3ppm which are similar to those of fresh Cefazolin in [A]b). ......................... 36
Figure 8. In vitro testing of biocompatibility and activity. Alamarblue assay on 3T3 fibroblast cells. a) Standard curve correlating fluorescence intensity (excitation: 544 nm, emission: 590 nm) and fibroblast cell number (data = mean± SEM, n=3) showing strong correlation (R2=0.999). b) Fibroblast cell number (data = mean± SEM, N=3) after 24 hour treatment with different concentrations of Cefazolin indicating no cytotoxicity of Cefazolin concentration of up to 100ppm. c) Agar diffusion test of Cefazolin–loaded microporous scaffolds (without GelMA coating) showing dose‐dependent zone of inhibition on S. aureus. * p<0.05, data = mean ± S.E.M (n= 8). d) In vitro blood clot formation on scaffolds. No significant difference in blood clot weight formed in presence of microporous PCL scaffold without (PCL scaffold) and with high dose of Cefazolin (PCL Scaffold with Cefazolin) compared with control group without scaffold (blood only) (n=5). .................................................. 37
Figure 9. Scanning electron microscopy (SEM) images showing surface pores on both non‐porous (nMPCL) and microporous (pMPCL) scaffolds (A, B, E, F) yet only pMPCL scaffold showed porous struts (G, H) compared with solid struts in nMPCL (C, D). The majority of intra‐strut micropores in pMPCL show interconnection with surrounding pores (arrows in G, H) and the surface pores on nMPCL are only artifacts of the extrusion (arrow in D). .......................................................... 56
Figure 10. Intra‐strut micropores reduced compression Young’s modulus of polycaprolactone (MPCL) scaffolds. (A): Effective compression Young’s modulus (data are expressed as mean ± s.e.m, n=9) showed no significant difference in the 0‐10% compression range. At 10‐15% and 15‐20% range, the micropores significantly reduced Young’s modulus of pMPCL. (B): SEM images of non‐microporous (nMPCL) and microporous (pMPCL) after the compression test showing the collapse of some micropores during the compression. ** is p<0.01 and from one‐way ANOVA test and Tukey post hoc test. ........................................................................................................ 59
Figure 11. DOX immobilized only on surface of nMPCL but both on surface and in intra‐strut pores of pMPCL. (A): bright field images of nMPCL and pMPCL before and after DOX–loading at 10 µg/mg. (B): Cross‐sections of nMPCL and pMPCL showing auto‐fluorescence of DOX on the surface of nMPCL and inside the struts of pMPCL. (C): FT‐IR spectra of DOX–loaded onto MPCL at increasing 3 loading doses (Low, Medium and High Doses). ................................................... 61
Figure 12. Intra‐strut micropores reduced burst release and prolonged release profile of drug–loaded microporous polycaprolactone (pMPCL) scaffolds. Drug release from scaffolds–loaded with increasing doses of DOX (Low, Medium and High doses) in PBS at pH 7.4 and pH 5.5 (data are expressed as mean ± s.e.m, n=6). pMPCL had less burst release and more prolonged elution compared with nMPCL. The release from DOX–loaded scaffolds increased under acidic conditions (pH 5.5). ...................................................................................................................... 62
Figure 13. Characterization of PTX–loaded pMPCL scaffolds. (A): Representative SEM micrographs of PTX–loaded pMPCL showing thin fiber ‐like morphology of immobilized PTX on the surface and inside micropores (white arrows) and their EDX analysis showing atomic C% (white numbers) higher than those of MPCL (85±1% for PTX and 75±1% for MPCL, data are expressed as mean ± s.e.m, n=6). (B): XPS analysis of PTX–loaded scaffolds showing N1s peaks from PTX even after 20 seconds of in situ ion beam etching for partial removal. (C): FT‐IR spectra of MPCL, PTX–loaded pMPCL scaffold (MPCL‐PTX) and PTX confirming the immobilization of PTX on MPCL scaffolds. ....................................................................................................................................... 64
Figure 14. Microporous structure prolonged and increased release of immobilized PTX on pMPCL. Drug was immobilized from solutions at 3 increasing doses (low, medium and high doses) by soak loading. Release at the first time point (24 hours) from pMPCL scaffolds was similar with nMPCL of the same dose but became significantly higher at later time points (data are expressed as mean ± s.e.m, n=6). ................................................................................................ 65
Figure 15. Immobilization of CEF on nMPCL and pMPCL scaffolds and the application of a PLGA overcoat layer. (A) and (B): SEM and EDX mapping of CEF–loaded scaffolds showing drug immobilized only on surface of nMPCL and also within pores inside the struts in pMPCL scaffolds. (C): SEM micrographs of drug–loaded pMPCL scaffolds with PLGA overcoat layers and EDX analysis showing differentiation of PLGA from the MPCL underneath which has less O content. ........................................................................................................................................ 67
Figure 16. Release of CEF from nMPCL and pMPCL scaffolds and the barrier effects of PLGA overcoat layers. (A): release profiles of scaffolds loaded with the drug at increasing doses (Low, Medium and High doses). (B): Release profiles of pMPCL scaffolds loaded with the drug at high dose with and without PLGA coating and model fitting of the release profiles. PLGA coating significantly slowed down the release of the CEF previously immobilized on the scaffolds. The pMPCL scaffolds showed first‐order like release and the application of the PLGA overcoat sustained the release further and achieved near zero‐order kinetic. ................................................................. 68
Figure 17. Cefazolin immobilized on microporous scaffolds retained its bioactivity. Zone of inhibition of CEF–loaded pMPCL scaffolds against S. aureus showing increasing zone diameter with increasing loading dose. pMPCL scaffolds without drug were used as control. Data are expressed as mean ± s.e.m (n=6). * is p<0.05 and ** is p<0.01 from one‐way ANOVA and Tukey post hoc test. ................................................................................................................................ 70
Figure 18. DOX and PTX immobilized on microporous scaffolds retained their bioactivity. (A): Metabolic activity of breast cancer (MDA‐MB‐231) cells in 3D culture treated with PTX–loaded scaffolds. Cells were encapsulated in methacrylated gelatin (GelMA) hydrogels and cultured in 24 well plate and PTX–loaded pMPCL scaffolds were then placed in the wells as treatment groups and Alamar Blue assay was performed on day 2, 5 and 8 of the treatment (data are expressed as mean ± s.e.m, n=8). (B): Bioactivity of DOX–loaded scaffolds. Cells were seeded in 24‐well plate, scaffolds were placed in the wells as treatment groups and intracellular DNA amounts were quantified after 2 days of treatment using PicoGreen assay (data are expressed as mean ± s.e.m, n=8). The results showed the effect of chemotherapeutic agent–loaded pMPCL scaffolds. * is p<0.05 and ** is p<0.01 from one‐way ANOVA and Tukey post hoc test. 71
Figure 19. Bimodal porous medical grade polycaprolactone (pmPCL) scaffolds fabricated using the screw melt extrusion method with interconnected pores. A: Representative bright field microscopic image of pmPCL scaffold. B and C: SEM micrographs of cross section of pmPCL scaffold strut before (B) and after (C) the porogen was leached out. D: µ‐CT micrograph of pmPCL scaffold after leaching. Black and white arrows: microscale pores on the surface of the strut. ............................................................................................................................................. 90
Figure 20. Loading of doxorubicin (DOX) on bimodal porous medical grade polycaprolactone scaffolds (pmPCL), dose‐dependent chemotherapeutic effect on breast cancer cells and in vitro linear sustained release of the loaded DOX. A: stereomicroscopic representative image of DOX–loaded pmPCL scaffold. B and C: representative images of surface (B) and cross section (C) of DOX–loaded pmPCL scaffold strut from spinning disc confocal microscope. Arrow: scaffold surface. D: Fourier‐transform infrared spectra of pmPCL scaffold, DOX and DOX–loaded pmPCL scaffold (DOX/pmPCL). E: in vitro cell viability of MDA‐MB‐231 breast cancer cell line at 2, 5 and 7 days of treatment with DOX–loaded pmPCL scaffolds at Low Dose (0.4 µg DOX/mg scaffold), High Dose (10 µg DOX/mg scaffold) and DOX in solution (Free DOX). Data are expressed as mean ± SEM (n=6). F and G: cumulative release of DOX–loaded from pmPCL at Low Dose (2 µg DOX/scaffold) and High Dose (8 µg DOX/scaffold) groups in PBS in the first 30 hours (F) and in 28 days (G). Data are expressed as mean ± SEM (n=3 scaffolds from 3 independent experiments). ** is p<0.01; One‐way ANOVA and Tukey post hoc test. ...................................... 91
Figure 21. The doxorubicin (DOX) was only detected in the peripheral blood of animals received the I.V. injection and reduced the body weight at week 1 post‐treatment. A: The concentration of DOX in the peripheral blood at day 1, 7 and 28 post‐treatment. B: Animal weight fold increment compared with the weight of animal at the time the treatment started. Data are expressed as mean ± SEM (n=6). ns: data is not statistically significant; @: Control group was significantly different from Low Dose and High Dose (p<0.01); #: Control group was significantly higher than Low Dose (p,0.01); * is p<0.05; ** is p<0.01 from one‐way ANOVA and Tukey post hoc test. .... 94
Figure 22. Cytotoxic effects in heart and liver of animals receiving intravenous (I.V.) doxorubicin (DOX) and medical‐grade polycaprolactone (pmPCL) scaffolds loaded with high dose, low dose and no DOX. Implantation of DOX‐loaded pmPCL scaffolds had similar mild cytotoxic effects on the liver and less cardio‐cytotoxicity compared with I.V. injection of DOX. A: H&E staining (top row) and Masson’s trichrome staining (middle row) of liver; congested central vein percentages measured from H&E staining and fibrotic tissue percentages measured from Masson’s trichrome staining of the Control (bimodal porous medical grade polycaprolactone (pmPCL) scaffolds without drug), I.V. (I.V. injection of doxorubicin (DOX) at 2 mg/kg), Low Dose and High Dose (DOX–loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups. B: Masson’s trichrome staining of heart (top and middle rows), and fibrotic tissue percentages measured from Masson’s Trichrome staining of the Control, I.V, Low Dose and High Dose groups. Data are expressed as mean ± SEM (n=6 mice for each group). ns: data is not statistically significant; * is p<0.05 from one‐way ANOVA and Tukey post hoc test. .............................................................. 96
Figure 23. Tumour recurrence in animals treated with intravenous (I.V.) doxorubicin (DOX) and medical‐grade polycaprolactone (pmPCL) scaffolds loaded with high dose, low dose and no DOX. Local drug delivery reduced the local tumour recurrence. A: Representative in vivo bioluminescence images of mice at days 0, 7, 14, 21 and 28 post treatment in Control (bimodal porous medical grade polycaprolactone (pmPCL) scaffolds without drug), I.V. (I.V. injection of doxorubicin (DOX) at 2 mg/kg dose), Low Dose and High Dose (DOX–loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups. B: quantitative in vivo IVIS signal over 21 days of treatment. C: ex vivo IVIS signal of local cancer recurrence at day 28 post treatment. Data are expressed as mean ± SEM (n=6 mice in each group). # is statistically significant compared with Control group, p<0.05; ns: data is not statistically significant; * is p<0.05 and ** is p<0.01 from one‐way ANOVA and Tukey post hoc test. ................................................................................... 98
Figure 24. Local drug delivery reduced the number and volume of the highly proliferative and vascularized local recurrent tumours. The representative bright field (A), H&E staining (B) and anti‐NuMA (human specific marker) staining images (C) of the recurrent tumour at the primary tumour site of the Control (bimodal porous medical grade polycaprolactone (pmPCL) scaffolds without drug), I.V. (I.V. injection of doxorubicin (DOX) at 2 mg/kg), Low Dose and High Dose (DOX–loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups. ................................. 99
Figure 25. Tumour metastasis in animals receiving intravenous (I.V.) doxorubicin (DOX) and medical‐grade polycaprolactone (pmPCL) scaffolds loaded with high dose, low dose and no DOX. Doxorubicin (DOX)–loaded bimodal porous medical grade polycaprolactone (pmPCL) scaffolds effectively reduced the breast tumor metastasis in lungs. Morphology (top row), anti‐NuMA staining (middle row), organ weight normalized to body weight, total flux from ex vivo BLI imaging, tumour cell count and micrometastasis count from NuMA staining images using ImageJ software of lung of the Control (bimodal porous medical grade polycaprolactone (pmPCL) scaffolds without drug), I.V. (I.V. injection of doxorubicin (DOX) at 2 mg/kg), Low Dose and High Dose (DOX–loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups. Arrows: tumour metastasis foci. The organs were weighted after fixation and normalized to the body weight of the animal at day 28. Data are expressed as mean ± SEM (n=6 mice in each experimental group). ns: data is not statistically significant; * is p<0.05; ** is p<0.01 from one‐way ANOVA and Tukey post hoc test. ........................................................................................ 101
Figure 26. The increased porogen concentration resulted in the higher intra‐strut microscale porosity, surface area and negatively charged group density. A and B: Representative SEM micrographs of strut’s surface and cross‐section, respectively. C: 3D reconstructed AFM micrographs of strut’s surface with their root‐mean‐square surface roughness (RMS). D: Representative µ‐CT micrographs
of scaffolds with their pore area. E: Intra‐strut porosity (determined from µ‐CT scans), surface area (measured by gas adsorption), water contact angles and surface charged group density (measured through TBO assay). RMS was measured from AFM scans of 20x20‐µm2 areas on struts’ surface. Data are expressed as mean ± s.e.m; n=6. ** is p<0.01 from one‐way ANOVA test and Tukey post hoc test. nmPCL, pmPCL17 and pmPCL44 are nonporous mPCL scaffolds, bimodal porous mPCL scaffolds prepared from mixture having 17% and 44% porogen, respectively. The scaffolds were extruded with the optimum flow rate and temperature to achieve the “superflow” extrusion state, create smooth and defect‐free struts [225]. As a result of the “superflow” state, only a small number of porogen was located at the surface of the struts; therefore, the surface porosity was significantly lower than intra‐strut porosity. .................................................................................................... 125
Figure 27. Microscale porosity accelerated degradation of mPCL scaffolds. A and B: Representative SEM micrographs from cross‐section (A) and surface (B) of nmPCL, pmPCL17 and pmPCL44 at 0, 24 and 48 hours in the accelerated conditions. C, D and E: Mass loss percentage (C), crystallinity measured from DSC (D) and molecular weight measured from GPC (E) of scaffolds after 48 hours of degradation. nmPCL, pmPCL17 and pmPCL44 are nonporous mPCL scaffolds, bimodal porous mPCL scaffolds prepared from mixture having 17% and 44% porogen, respectively. All microscale porous scaffolds showed evidence of degradation after 24 hours both on strut’s surface (rougher and enlarged pores) and throughout cross section (more small interconnected pores). After 48 hours, nmPCL showed only surface degradation; whereas, pmPCL44 degraded throughout the strut’s cross section. Data are expressed as mean ± s.e.m; n=6. ** is p<0.01 from one‐way ANOVA test and Tukey post hoc test. ............................................................................................................................... 128
Figure 28. In vitro blood clotting experiments. A and B: SEM micrographs of cross‐section (A) and surface (B) of hematoma encapsulated scaffolds. C: Wet weight of hematoma. D and E: TGF‐ß1 (D) and PDGF AB (E) release from hematoma after 24 hours incubation in the culture medium. F: Cell proliferation results (MTT assay) of MC3T3‐E1 cultured with hematoma supernatant after 1 and 3 days. Data are expressed as mean ± s.e.m; n=6. ns is not significant, * is p<0.05 and ** is p<0.01, from one‐way ANOVA test and Tukey post hoc test. nmPCL, nmPCL/CaP and pmPCL44 are nonporous mPCL scaffolds, calcium phosphate coated nmPCL scaffolds and bimodal porous mPCL scaffolds prepared from mixture having 44% porogen, respectively. ............................................. 131
Figure 29. Increased protein adsorption on pmPCL44 was observed compared with nmPCL and nmPCL/CaP. Metabolic activities of MC3T3‐E1 cells on nmPCL, nmPCL/CaP and pmPCL44 scaffolds after 1, 3 and 5 days. Results were normalized to control scaffold (nmPCL) at day 1. Data are expressed as mean ± s.e.m; n=6. ns is not significant, * is p<0.05 and ** is p<0.01 from one‐way ANOVA test and Tukey post hoc test. nmPCL, nmPCL/CaP and pmPCL44 are nonporous mPCL scaffolds, calcium phosphate coated nmPCL scaffolds and bimodal porous mPCL scaffolds prepared from mixture having 44% porogen, respectively. .......................................................................... 132
Figure 30. Bimodal porous scaffolds had a similar osteoconductivity compared with calcium phosphate coated scaffolds. A: µ‐CT‐based 3D reconstructed images of new bone formation at calvarial defects and their bone volume. nmPCL/CaP and pmPCL44 are calcium phosphate coated nonporous mPCL scaffolds and bimodal porous mPCL scaffolds prepared from mixture having 44% porogen, respectively. B: H&E staining at calvarial defects (boxed) of nmPCL/CaP and pmPCL44 with arrows showed microtissue inside the pmPCL44 struts. ob: old bone; nb: new bone; s: scaffold; sm: skin‐side membrane; and dm: dura‐side membrane; black arrow: infiltrated extracellular materials. C: Histomorphometrical analysis of bone area percentage and stroma tissue area percentage measured from scanned H&E stained slides using the Osteomeasure Analysis System. Data are expressed as mean ± s.e.m; n=6. ns is not significant from one‐way ANOVA test and Tukey post hoc test. ............................................................................................................................................. 134
Figure 31. Validation of new bone formation via immunohistological staining with osteoblast key markers, ALP (A) and Col I (B) at week 8 after scaffold implantation. Nonporous medical grade polycaprolactone ( nmPCL/CaP) and pmPCL44 are calcium phosphate coated nonporous mPCL scaffolds and bimodal porous mPCL scaffolds prepared from mixture having 44% porogen, respectively. Both nmPCL/CaP and pmPCL44 groups demonstrated similar levels of ALP and Col I expression. ob: old bone; nb: new bone; s: scaffold; sm: skin‐side membrane; and dm: dura‐side membrane. .................................................................................................................................. 136
Figure 32. Confirmation of blood vessel formation via immunohistological staining for CD31 (A) and vWF (B) at 8 weeks after scaffold implantation in rat critical size calvarial defect model. nmPCL/CaP and pmPCL44 are calcium phosphate coated nonporous mPCL scaffolds and bimodal porous mPCL scaffolds prepared from mixture having 44% porogen, respectively. The blood vessel formation was confirmed by positive staining for CD31 and vWF as indicated with brown color. ob: old bone; nb: new bone; s: scaffold; sm: skin‐side membrane; dm: dura‐side membrane; and black arrows: blood vessel. .......................................................................................................................................... 137
Figure 33. Immunohistological staining showed the presence of M1 macrophages (CD68+ and iNOS+) much less than M2 macrophages (CD68+ and CD163+) within the defect areas in both nmPCL/CaP (A) and pmPCL44 (B) groups. nmPCL/CaP and pmPCL44 are calcium phosphate coated nonporous mPCL scaffolds and bimodal porous mPCL scaffolds prepared from mixture having 44% porogen, respectively. ob: old bone; nb: new bone; s: scaffold; sm: skin‐side membrane; and dm: dura‐side membrane. Black arrows demonstrate positive cells (brown color). .............................................. 139
List of Tables
Table 1. Loading doses of doxorubicin (DOX) on bimodal porous medical grade polycaprolactone (pmPCL) scaffolds. ....................................................................................................................................... 85
Table 2. Treatment groups for in vivo implantation. ................................................................................ 87
Table 3. Porogen concentration in mixture with mPCL to make scaffolds. .............................................. 113
Table 4. Ultimate tensile strength, Young's modulus, and elongation at break of single struts with different porosity. Data are expressed as mean ± s.e.m; n=6. ** is p<0.01 from one‐way ANOVA test and Tukey post hoc test. nmPCL, pmPCL17 and pmPCL44 are nonporous mPCL scaffolds, bimodal porous mPCL scaffolds prepared from mixture having 17% and 44% porogen, respectively. ..................... 127
Table 5. Surface area of scaffolds measured by gas adsorption analysis. (data = mean ± s.e.m; n=3). *** is p < 0.001. nmPCL, nmPCL/CaP and pmPCL44 are nonporous mPCL scaffolds, calcium phosphate coated nmPCL scaffolds and bimodal porous mPCL scaffolds prepared from mixture having 44% porogen, respectively. .................................................................................................................. 130
List of Supplementary Figures
Figure S1. Rheology testing of mPCL and mPCL‐porogen for printing process showing amplitude sweep tests (a1‐a3) and temperature sweep (b1‐b3). For amplitude tests, temperature was kept constant at 110°C and angular frequency at 10 rad/sec. Each data point is the average of three experimental repeat (N=3); the error bars correspond to SD. For temperature sweep the temperature was varied from 125°C to 35°C. Strain was kept constant at 1% and angular frequency at 10 rad/sec. Each point is the average of three experimental repeat (N=3); the error bars correspond to SD. .................... 149
Figure S2. Microporous mPCL scaffolds – after compression – relaxation cycles in mechanical testing ... 150
Figure S3. mPCL scaffolds – after compression – relaxation cycles in mechanical testing ........................ 150
Figure S4. Morphology of nMPCL and pMPCL. (A): Representative bright field microscopic images of nMPCL and pMPCL. (B): AFM micrographs of surface of struts on nMPCL and pMPCL scaffolds and their root‐mean‐square surface roughness (RMS, data are expressed as mean ± s.e.m, n=6). (C) and (D): SEM and microCT images showing surface pores (short arrows) and connection to the inside pores through interconnectivity (long arrows). (E): Intra‐strut micropore size class distribution on surface and cross‐section of nMPCL and pMPCL. Data are expressed as mean ± s.e.m, n=6. .......... 151
Figure S5. Schematic of interaction between DOX and MPCL molecules. ............................................... 152
Figure S6. Drug loading efficiency of DOX, PTX and CEF on nPCL and pPCL. Data are expressed as mean ± s.e.m, n=6. ** is p<0.01 from one‐way ANOVA test and Tukey post hoc test. ................................ 153
Figure S7. Pore size distribution of bimodal porous medical grade polycaprolactone scaffolds. data were expressed as mean ± SEM, n = 4. .................................................................................................. 153
Figure S8. In vivo IVIS images of the control (A, bimodal porous medical grade polycaprolactone (pmPCL) scaffolds without drug), I.V. (B, I.V. injection of doxorubicin (DOX) at 2 mg/kg), low dose and high dose (C and D, DOX–loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups. ............... 154
Figure S9. Doxorubicin (DOX) –loaded bimodal porous medical grade polycaprolactone (pmPCL) scaffolds effectively reduced the breast tumor metastasis in liver and spleen. A and B: Morphology, organ weight normalized to body weight and total flux from ex vivo BLI imaging of liver (A) and spleen (B) of the Control (bimodal porous medical grade polycaprolactone (pmPCL) scaffolds without drug), I.V. (I.V. injection of doxorubicin (DOX) at 2 mg/kg), Low Dose and High Dose (DOX–loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups. The organs were weighed after fixation and normalized to the body weight of the animal at day 28. Data are expressed as mean ± SEM (n=6 mice in each experimental group). ns: data is not statistically significant; * is p<0.05; ** is p<0.01 from one‐way ANOVA and Tukey post hoc test. ............................................................................ 155
Figure S10. Porogen size class and viscoelasticity of MPCL and MPCL‐porogen films. A and B: Representative size class distribution by volume acquired by light scattering. C: SEM micrograph showing the morphology of PBS porogen. D, F and H: Storage modulus (G'), loss modulus (G'') and loss tangent result from amplitude sweep rheological measurement of PCL and porogen‐PCL composites at a temperature of 110°C and an angular frequency of 10 rad/sec. E, G and I: Storage modulus (G'), loss modulus (G'') and loss tangent (tan δ) results from temperature sweep rheological measurements at a strain of 1% and an angular frequency of 10 rad/sec. ..................................... 157
Figure S11. Validation of complete leaching of porogen. A: SEM micrographs of pMPCL44 before (left) and after (right) leaching. B: EDX analysis results of pPCL44 scaffolds after 0, 7 and 14 days leached in NaOH 0.01M in shaking incubator. C: EDX analysis results of all PCL scaffolds after leaching. D: UV‐Vis absorbance results of remained porogen percentage after leaching. Data are expressed as mean ± s.e.m; n=6. *: p<0.05 and **: p<0.01 from One‐way ANOVA test and Tukey post‐hoc test. ......... 158
Figure S12. Degradation of pMPCL compared with nMPCL. The crystalline/amorphous ratios of nPCL and pPCL44 scaffolds after 12, 24, 36 and 48 hours of degradation (A). GPC determination of molecular
weight and polydispersity of nPCL and pPCL44 scaffolds after 48 hours of accelerated degradation (B and C). Data are expressed as mean ± s.e.m; n=6. ** is p<0.01 from one‐way ANOVA test and Tukey post hoc test. ............................................................................................................................... 158
Figure S13. Representative SEM micrographs at low (A) and high (B) magnification showing surface morphology of nPCL/CaP scaffold. ................................................................................................ 159
List of Supplementary Tables
Table S1. List of primary antibodies used for immunohistochemically staining and their antigen retrieval conditions. ................................................................................................................................... 156
List of Abbreviations
ABS Acrylonitrile butadiene styrene
AFM Atomic force microscopy
ALP Alkaline phosphatase
ANOVA Analysis of variance
APS Ammonium persulfate
BET Brunanuer, Emmett and Teller
BLI Bioluminescent imaging
BSA Bovine serum albumin
CaP Calcium‐phosphate
CEF Cefazolin
CFU Colony forming units
Col I Collagen type I
DL Desolvation line
DMEM Dulbecco’s Modified Eagle’s Medium
DMSO Dimethyl sulfoxide
DNA Deoxyribonucleic acid
DOD Drop on demand
DOX Doxorubicin
DSC Differential scanning calorimetry
ECM Extracellular matrix
EDTA Ethylenediaminetetraacetic acid
EDX Energy‐dispersive X‐ray spectroscopy
ELISA Enzyme‐linked immunosorbent assay
EPR Enhanced permeability and retention
FBS Fetal bovine serum
FDM Fused deposition modelling
FT‐IR Fourier transform infrared spectroscopy
GelMA Gelatin methacryloyl
GPC Gel permeation chromatography
H&E Hematoxylin‐eosin
HPLC High performance liquid chromatography
IHC Immunohistochemical
I.V. Intravenous
IVIS In vivo imaging system
LCMS Liquid chromatography–mass spectrometry
MBG Mesoporous bioactive glass
MIC Minimum inhibitory concentration
Mn Number‐average molecular weight
mPCL Medical grade poly(ε‐caprolactone)
MTT 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyl tetrazolium bromide
Mw Weight‐average molecular weight
NIPS non‐solvent induced phase separation
PBE Phosphate buffered EDTA
PBS Phosphate buffered saline
PCL Poly(ε‐caprolactone)
PD Polydispersity
PDGF‐AB Platelet‐derived growth factor AB
PEO poly(ethylene glycol)
PLGA Poly(lactic‐co‐glycolic acid)
PLLA Poly(L‐lactic acid)
PPAA Polyacrylic acid
PTX Paclitaxel
PU Polyurethane
RMS Root‐mean‐square
RT Room temperature
SBF Simulated body fluid
SEM Scanning electron microscopy
SME Screw melt extrusion
SLS Selective laser sintering
SRM Selected reaction monitoring
STL Stereolithography
TBO Toluidine Blue O
TCP Tricalcium phosphate
TE Tissue engineering
TEMED Tetramethylethylenediamine
TGF‐β Transforming growth factor‐β
‐CT Micro‐computed tomography
UHPLC Ultra‐high performance liquid chromatography
UTS Ultimate tensile strength
UV Ultraviolet
vWF von Willebrand Factor
XPS X‐ray photoelectron spectroscopy
Statement of Original Authorship
The work contained in this thesis has not been previously submitted to meet
requirements for an award at this or any other higher education institution. To the best of my
knowledge and belief, the thesis contains no material previously published or written by another
person except where due reference is made.
QUT Verified Signature
Acknowledgments
First, I would like to thank my supervisory team, Distinguished Professor Dietmar W.
Hutmacher, Associate Professor Tim Dargaville, Dr Phong A. Tran and Dr Abbas Shafiee. Thanks
for being patient with my inexperience, and the guidance and knowledge in design and
performing the experiment. I would like to thank Distinguished Professor Dietmar W. Hutmacher
for his motivation. I always felt inspired after discussions with him. I would like to thank Associate
Professor Tim Dargaville for emotional support after I finished my second year.
Next, I would like to thank QUT for giving me the QUTPRA scholarship and the
opportunity to pursue my research interest in Australia. I would like to thank the staff from the
Central Analytical Research Facility (CARF), Institute for Future Environment (IFE) and Medical
Engineering Research Facility (MERF) at QUT for their technical support, training and guidance.
I would like to thank Dr Mohit Chhaya, the first person I met and who trained me to
fabricate scaffolds. I would like to thank my friends Johann, Triet, Tara, Margaux and Jane for
sharing my feelings and lots of coffee and drinks. I would like to acknowledge the use of
professional editing service from Elite Editing TM (South Australia, Australia) to edit my thesis.
Last but not least, I would like to thank my wife Van for supporting me during my stressful
and difficult times and my daughter Thanh for being my motivation for what I am doing. Without
any of them, I believe I could not have made it this far.
Note to the reader
Chapters 2 and 3 have minor changes and from the published papers to reduce
confusion, fit into the thesis and follow the feedbacks of examiners.
Page 2
Page 2
1Chapter 1: Introduction
1.1 BACKGROUND
Tissue engineering using scaffold implants is a promising approach to treat large‐volume
tissue defects, such as those resulting from removal of bone or breast tissue with neoplasm
removal [1, 2]. The scaffolds give initial support for cell recruitment, attachment and proliferation
as well as blood vessel penetration and neovascularisation [3, 4]. However, scaffold implantation
has the risk of post‐surgery complications such as biofilm formation and infection [5].
Post‐implantation infection occurs in 1.2% of hip/knee replacement patients and 15% of
breast implant patients [6‐8]. Systemic administration of antibiotics before and after implantation
is the current approach to prevent post‐implantation infection. However, delivery efficiency is
exceedingly low as only a small amount of antibiotic reaches the target site, which may lead to
devastating side effects as the drug may accumulate in other undesired tissues [9]. Moreover,
frequent exposure of bacteria to antibiotics at less than minimum inhibitory concentration may
lead to antibiotic resistance [10].
In the case of patients undergoing tumour removal, local cancer recurrence occurs in 11–
12% of primary bone cancer patients and 10.2% of breast cancer patients [11, 12]. Similar to
antibiotics in infection prevention, chemotherapeutic agents are commonly used systemically to
prevent cancer recurrence. However, Harasym et al. used a mouse model to demonstrate that
only 10% of the drug accumulates at the target site [13]. Therefore, 85–90% of the administered
drug circulates in the blood system, accumulates in other organs and may cause systemic toxicity
[12‐14]. In addition, there is a limited amount of a drug that a patient can tolerate during his/her
lifetime.
Page 3
Page 3
Local drug delivery methods have been introduced as a promising approach to improve
the drug delivery efficacy and reduce the risk of drug resistance and systemic cytotoxicity. Shim
et al. mixed tobramycin with poly‐e‐caprolactone/poly(lactic‐co‐glycolic acid) to fabricate
antibiotic‐loaded scaffolds for local delivery purposes. However, heat exposure during the
fabrication process may affect the activity of the antibiotic [14‐16]. Conversely, many applications
of local chemotherapeutic delivery have been successfully commercialized, such as Viadur® to
prevent prostate cancer and Zoladex® to prevent endometriosis and breast cancer in women [17‐
20]. However, these implants are limited to a small volume and cannot be used as a scaffold for
tissue regeneration.
Given the current challenges and the high potential of local drug delivery, modification
of three‐dimensional (3D)‐printed scaffolds to enable them to serve as both a framework for
tissue reconstruction and a drug carrier for local drug delivery may be a promising solution. We
proposed a new approach to incorporate microscale porosity into solid struts of conventional
macroscale porous scaffolds. The intra‐strut microscale porosity can increase scaffold–protein
adsorption, stimulate cell attachment, and support cell penetration and vascularisation [21, 22].
Moreover, the microscale porosity can increase the interconnectivity of scaffolds, generating a
more complex structure, which can alter drug loading efficiency and affect the drug release
profile.
1.2 LITERATURE REVIEW
1.2.1 3D printing technology in biomedical application
3D printing technology is a very useful technique to fabricate 3D scaffolds with complex
architecture. Recently, 3D‐printed scaffolds have been manufactured and used as a carrier for
local drug delivery application [23, 24]. 3D‐printed scaffolds can be classified into different groups
according to their fabrication mechanism. The first group is photo‐crosslink polymers, in which
the polymers crosslink when exposed to a specific light source. The polymers can be in powder
Page 4
Page 4
form and spread on a collecting plate (stereolithography [STL]) or dissolved in solvent, displaced
on a collecting board and then crosslinked (material jetting or drop‐on‐demand method [DOD]).
The second group is thermal crosslink polymers, in which the materials melt and bind to each
other when exposed to a thermal source such as laser (selective laser sintering [SLS]). The last
group is molten polymers. The polymer filaments are fed into the printer, melted and discharged
onto a collecting plate (fused deposition modelling [FDM]); the polymers rapidly cool down and
solidify, forming the 3D structure. Another method using melting of polymers is screw melt
extrusion (SME), in which the polymers in any shape or form are melted in the heating chamber,
which connects directly to the extruder. The extruder extrudes materials with the screwing core
as well as compressed air and moves along X, Y and Z axes to form the 3D constructs.
Among these methods, SME is versatile and can be easily modified to fabricate and
characterize new types of material/scaffold. This technique can be applied to any type of material
that can be melted into liquid form and has suitable viscosity for printing.
1.2.2 Porosity in 3D printed scaffolds
One of the key parameters when designing 3D scaffolds is porosity. The porosity of
scaffolds can affect the interactions between the scaffold and the drugs, proteins and cells.
Macroscale pores (pores with the longest dimension of 100–1000 µm) are usually generated by
the printing pattern and play important roles in cell and tissue infiltration as well as blood vessel
ingrowth and neovascularisation [3, 22]. Conversely, microscale pores (pores with the longest
dimension of 1–100 µm) play essential roles in protein adsorption and cell attachment, migration
and proliferation. They also interact with extracellular substances and enhance
neovascularisation. Microscale pores usually cannot be designed by the printing pattern because
the precision of the technique is limited to 0.1 mm. They are usually generated by other
techniques, such as porogen leaching and foam casting. Samavedi et al. showed that microscale
pores can provide more surface area for protein adsorption, cell attachment and osteogenic
Page 5
Page 5
differentiation [21]. Moreover, microscale porosity plays an important role in haematoma
formation, which is the first critical step following tissue damage and an important process in
wound healing and tissue regeneration [25‐27].
1.2.3 Bimodal porosity
Macroscale porosity and microscale porosity can be combined to take advantage of both
structures. Wei et al. showed that successful bone tissue repair requires a combination of
macroscale porosity, microscale porosity and interconnectivity to support cell penetration and
vascularisation [22]. Other studies showed that macroscale porosity can support cell penetration,
infiltration and in‐growth as well as blood vessel ingrowth; whereas, microscale porosity
enhanced cell attachment and proliferation, blood vessel infiltration, protein penetration and
mineralization [28‐30]. The addition of microscale porosity into macroscale porous scaffolds can
increase bone density and uniformity.
The method of fabricating bimodal porous scaffolds can be classified into two groups:
moulding‐based and extrusion‐based methods. In the first group, the materials, which consist of
a combination of polymers and pore‐forming polymers/particles, are moulded. Then, the pore‐
forming materials are removed via a leaching process. Last, the samples are dried out or
lyophilized to obtain the final structure. Zhang et al. fabricated calcium silicate/calcium
phosphate scaffolds by applying this method and using 500 μm NaCl particles as the porogen.
The scaffolds were loaded with recombinant human bone morphogenetic protein‐2 (RhBMP‐2)
and showed support for myoblastic precursor cell attachment and proliferation [31]. Wei et al.
used a process of moulding, leaching and drying with a 400–500 μm NaCl porogen to fabricate
micro/macro magnesium calcium phosphate scaffolds. The scaffolds had increased cell
attachment, proliferation and ALP (alkaline phosphatase) activity compared with scaffolds with
macropores only [22]. Besides calcium phosphate (CaP), other materials can be used with this
method. Salerno et al. melt‐mixed polycaprolactone (PCL) with thermoplastic gelatin (TG), air
Page 6
Page 6
blowed the material to create micropores and finally leached TG to create interconnected
micropore channels [32]. The construct demonstrated support for human mesenchymal stem
cells (hMSCs) attachment. Akkouch et al. utilized a collagen/hydroxyapatite/poly(lactide‐co‐
caprolactone) composite, in which collagen served as the pore‐forming material and was
extracted by dissolving in water, to fabricate bimodal porous scaffolds that showed increased
osteoblast cell growth [33]. The benefit of bimodal porous scaffolds was also demonstrated in
vivo. Yan et al. fabricated silk fibroin/calcium phosphate scaffolds using 500–1000 μm NaCl
particles as the porogen [34]. The scaffolds improved the new bone formed in rat femur defects.
In the second group of fabrication method, which is extrusion based, the materials are
rearranged/mixed before extrusion. After extrusion, either into filaments or 3D scaffolds, the
pore‐forming materials are removed by leaching or solvent exchange to achieve the final bimodal
porous structure. Reignier et al. blended PCL with poly(ethylene oxide) (PEO) and NaCl crystals
before extrusion through a non‐restrictive extrusion die and leaching in water [35]. The scaffolds
showed complete leach‐out of NaCl, but there was some PEO entrapped. Further, the study
demonstrated that NaCl was not a good porogen as the 100–700 μm particles were broken into
two ranges of less than 10 μm and more than 100 μm, which makes it hard to control the pore
size as desired. Ahn et al. took a different approach by adding the materials layer by layer in the
feeding chamber before printing [36]. CaP/camphene was placed at the bottom and then
camphene was placed as a thick layer on top. When extruded into filaments, CaP/camphene
came out first before gradually mixing with camphene, creating a camphene gradient. The
camphene in the constructs was removed by sublimation, creating a bone‐like structure with
gradient porosity. Kim et al. took another approach to fabricate bimodal porous scaffolds by using
a non‐solvent‐induced phase separation (NIPS)‐based 3D plotting technique [37]. PCL was mixed
with hydroxyapatite (HA) particles in tetrahydrofuran (THF). The mixture was extruded by
pressurized air through a fine nozzle into an ethanol bath. The exchange between ethanol and
Page 7
Page 7
THF created a phase separation and resulted in microporous PCL/HA composite filaments. The
scaffolds demonstrated increased ALP activity of osteoblast cells.
1.2.4 3D printed polycaprolactone in tissue reconstruction application and local
drug delivery application
1.2.4.1 Additive manufactured 3D polycaprolactone
Among methods to fabricate 3D scaffold, Additive biomanufacturing (ABM) holds a
promising potential. Using this method, we can create scaffolds with desired architecture and
interconnected macroscale porous networks. ABM can also be used combined with other
techniques to fabricate multiscale porous construct for various applications. [38‐41]. ABM
combined with sintering method was used to fabricate titanium – polyvinyl alcohol (PVA)
scaffolds [38]. ABM combined with non‐solvent induced phase separation (NIPS)‐based plotting
was used to fabricate PCL/hydroxyapatite scaffolds [37].
Besides method, polymer is another important factor in scaffold fabrication. In this study,
medical grade polycaprolactone polymer was used because of its previous succeed in using for
FDA‐approved medical devices and versatile properties for printing and modification [58‐61]. The
material has low melting temperature (60oC), biocompatible and biodegradable.
1.2.4.2 3D printed polycaprolactone in tissue reconstruction application
ABM has been extensively studied for application in tissue engineering as implants to
support new tissue formation [23, 42‐44]. thanks to its ability to fabricate complex porous
structure with macropores for vascularization [45, 46]. Yet, the ideal scaffolds require both
macroscale pores as well as microscale pores to support tissue regeneration and remodelling [47‐
51]. The conventional method to address this issue include moulding combined with salt leaching,
use of sacrificed material or gas foaming to create dual scale porous structure [32, 52‐55]. Though
these methods can achieve having macroscale and microscale pores in single constructs, there
was limited controls over the shape, size and porosity. Therefore, ABM was employed. ABM
Page 8
Page 8
combined with sintering method was used to fabricate titanium – polyvinyl alcohol (PVA)
scaffolds [38]. ABM combined with non‐solvent induced phase separation (NIPS)‐based plotting
was used to fabricate PCL/hydroxyapatite scaffolds [37].
1.2.4.3 3D printed polycaprolactone in local drug delivery application
Biodegradable scaffolds, beside the primary role of supporting new tissue formation,
were proposed to be used as a local drug reservoir to deliver drug locally [23, 42‐44]. The use of
locally delivered antibiotics for surgical site prophylaxis or locally delivered chemotherapeutic
agents for post‐surgery recurrent cancer prevention is a promising approach for scaffold‐based
tissue engineering applications [23, 56]. The conventional method of preparing local drug delivery
device used mixing methods, in which, the bioactive agents were mixed prior to manufacturing.
Since the manufacturing involving heating, heat labile drugs can be affected and their bioactivities
may be reduced or lost. [14‐16]. Furthermore, the drug release in these cases depend on the
degradation of the device and may lead to treatment resistant due to frequent delivery of
underdosed drugs.
Despite the fact that PCL scaffolds have been extensively developed, they are mainly
design as a single‐function scaffolds to serve as either TE or DDS but not both. For diseases that
require a large volume of native tissue resection and post‐surgery treatment, the dual function
scaffolds are indeed add much more benefit for the treatment.
1.3 RESEARCH GAP
It is well known that the regeneration of large‐volume tissue defects requires
biocompatible implants with adjunct treatments to prevent complications or disease recurrence.
Regarding the biocompatible implants, bimodal porous scaffolds are promising as the scaffolds
have the advantages of both macroscale and microscale pores. However, recent methods to
fabricate the scaffolds still face challenges. The moulding‐based methods limit the construct to
the mould shape and size. In addition, NaCl is unsuitable as a porogen for formation of
Page 9
Page 9
macropores as it cannot maintain its size and shape after mixing, and micropores created by a
sintering or lyophilisation process are hardly controllable. In the case of the extrusion‐based
method, the macropores can be controlled by the printing pattern; however, the micropores
fabrication is not fully optimized. The use of PEO as a pore‐forming material leaves residual
material, and the use of the NIPS method leads to uncontrolled microscale porosity.
Regarding the adjunct treatments to be used with the implants, antibiotics and
chemotherapeutic agents have been commonly used to prevent infection and tumour
recurrence; however, they also have systemic side effects and toxicity. As mentioned in the
previous section, local drug delivery is a promising approach to prevent complications, reduce
systemic toxicity and enhance drug concentration at the site of implantation. 3D‐printed scaffolds
can serve as drug depots [23, 56]; however, existing methods in this area rely on mixing drugs
and biomaterials before manufacturing the mixture into scaffolds [23, 57, 58]. A major limitation
of direct mixing is that drug bioactivities can be altered by temperature and harsh solvents. For
example, biological activity loss of more than 10% for vancomycin and 20% for heparin was
reported when they were mixed with CaP for ink‐jet printing into 3D structures [57]. In addition,
being embedded inside the solid material, the drug is only released when the material degrades,
which might take a long time, necessitating much higher drug loading to have therapeutic effects
[59], which might create undesirable consequences such as persistent inflammation or treatment
resistance. Research groups have tried to address this issue by adding ingredients to the drug
biomaterial mixture, such as HA [59], cellulose [60] and chitosan [57], to facilitate fluid infiltration
and thus drug elution. This approach is cumbersome and requires further improvement.
Overall, there is a need for a new type of scaffold implant in which the microscale and
macroscale porosities can be controlled, and drugs can be loaded without altering drug
bioactivities.
Page 10
Page 10
1.4 HYPOTHESIS AND OBJECTIVES
Hypothesis: The introduction of microscale pores with controlled pore size into the solid
struts of 3D‐printed macroscale porous scaffolds will increase the scaffold surface area and
overall porosity to (1) support protein adsorption, cellular function and eventually tissue
formation and (2) efficiently adsorb drugs and control their release for anti‐infection and anti‐
cancer applications.
The hypothesis has been examined by achieving the following research objectives:
Objective 1: Fabricate and characterize bimodal porous PCL scaffolds.
Objective 2: Investigate the loading and release and bioactivities of antibiotics and
chemotherapeutic agents from the scaffolds.
Objective 3: Investigate the application of drug‐loaded bimodal porous scaffolds to treat
breast cancer in vivo.
Objective 4: Investigate the application of bimodal porous scaffolds for tissue
reconstruction in vivo.
1.5 SCOPE OF STUDY AND THESIS OUTLINE
The ultimate goal of this thesis is to demonstrate the versatility of the biomodal porous
scaffolds in various applications. As such this study was designed based on the viable product and
used it for a breath spectrum of application. Due to the limited timeframe of a PhD, this study
could not be able to go in dept characterization and optimization for each specific application.
Specifically, the feasibility of the new fabrication method using porogen‐polymer composites was
investigated; yet the optimized settings to fabricate scaffolds for specific needs, such as local
antibiotic delivery or breast tissue reconstruction, were not studied. Properties of scaffolds
including mechanical properties, physical properties and degradation properties were
characterized; yet the long‐term stability in term of structure and mechanical properties was not
characterized. For local drug delivery studies, different models drugs were used to characterize
Page 11
Page 11
the drug release profile, bioactivities of the loaded drugs; yet the studies only proposed the
interaction between drugs and polymers and the model fitting for each model drug to
characterize their release kinetic and diffusion was excluded from this studies. For local delivery
of chemotherapeutic drug to prevent breast cancer recurrence and metastasis, the drug loaded
scaffolds were used to compared with the conventional intravenous injection method; yet the
doses to achieve the most effective treatment were not optimized.
This report contains six chapters arranged as follows.
Chapter 1 introduces the background about large tissue/critical‐size tissue defects,
current challenges of methods used to prevent infection and cancer recurrence, a literature
review about 3D‐printed scaffolds in biomedical applications, requirements of porosity in tissue
reconstruction and drug delivery, and the challenges with current methods to fabricate bimodal
porous scaffolds, and proposes a novel approach, the hypothesis and the research objectives of
this thesis and the thesis outline.
Chapter 2 talks about preliminary investigations of scaffold fabrication and antibiotic
delivery using cefazolin as a drug model. Drug loading methods and drug delivery profiles of
antibiotic‐loaded PCL scaffolds were evaluated. The antimicrobial properties of drug‐loaded
scaffolds were demonstrated in vitro.
Chapter 3 further talks about characteristics of the bimodal porous scaffolds and release
behaviours of antibiotic and chemotherapeutic drugs with different characteristics and under
different conditions. The antimicrobial and chemotherapeutic properties of drug‐loaded
scaffolds were demonstrated in vitro.
Chapter 4 talks about the use of doxorubicin (DOX)‐loaded scaffolds as a local
chemotherapeutic implant in a mouse orthotopic breast cancer model. The chemotherapeutic
effects of the treatment against local recurrence and metastasis were evaluated. The cytotoxic
effects of the treatment were characterized and compared with a conventional intravenous (I.V.)
systemic administration method.
Page 12
Page 12
Chapter 5 talks about the use of bimodal porous scaffolds for tissue reconstruction. The
interactions between the scaffold and the proteins, haematoma and osteoblast cells were
characterized. Benefits derived from the scaffold tissue reconstruction were demonstrated in a
rat critical‐size defect model.
Chapter 6 summarizes the research findings and gives suggestions for future research.
Page 13
Page 13
2Chapter 2: 3D printed Polycaprolactone
scaffolds with dual macro‐microporosity for
applications in local delivery of antibiotics
Luke E. Visscher a, *, Hoang Phuc Dang b, c, *, Mark A. Knackstedt d, Dietmar W. Hutmacher b, c, Phong
A. Tran b, c
Published in Journal Materials Science and Engineering: C 87 (2018): 78‐89.
a School of Medicine, University of Queensland, QLD, Australia
b ARC Centre in Additive Biomanufacturing, Queensland University of Technology, Musk
Avenue, Kelvin Grove, Brisbane, Queensland 4059, Australia
c Queensland University of Technology (QUT), Brisbane, Queensland, Australia
d Australian National University, Canberra, Australia
* contributed equally
Page 14
Page 14
The authors listed below have certified that:
1. They meet the criteria for authorship in that they have participated in the conception,
execution, or interpretation, of at least that part of the publication in their field of expertise;
2. They take public responsibility for their part of the publication, except for the responsible
author who accepts overall responsibility for the publication;
3. There are no other authors of the publication according to these criteria;
4. Potential conflicts of interest have been disclosed to (a) granting bodies, (b) the editor or
publisher of journals or other publications, and (c) the head of the responsible academic unit,
and
5. They agree to the use of the publication in the student’s thesis and its publication on the QUT’s
ePrints site consistent with any limitations set by publisher requirements.
In the case of this chapter:
Published paper:
Luke E. Visscher, Hoang Phuc Dang, Mark A. Knackstedt, Dietmar W. Hutmacher and Phong A.
Tran. “3D printed Polycaprolactone scaffolds with dual macro‐microporosity for applications in
local delivery of antibiotics.” Materials Science and Engineering: C 87 (2018): 78‐89.
Contributor Statement of contribution
Hoang Phuc Dang
Conducted experiments and data analysis, wrote manuscript Signature
Date:
Luke E. Visscher Conducted experiments and data analysis, wrote manuscript
Mark A. Knackstedt Conducted experiments
Dietmar W. Hutmacher Experimental design, aided manuscript preparation
Page 15
Page 15
Phong A. Tran
Experimental design, aided data analysis and manuscript
preparation
Principal Supervisor Confirmation
I have sighted email or other correspondence from all Co‐authors confirming their
certifying authorship
Phong Tran 06/02/2020
Name Signature Date
QUT Verified Signature
Page 16
Page 16
2.1 ABSTRACT
Advanced scaffolds used in tissue regenerating applications should be designed to
address clinically relevant complications such as surgical site infection associated surgical
procedures. Recognizing that patient‐specific scaffolds with local drug delivery capabilities are a
promising approach, we combined 3D printing with traditional salt‐leaching techniques to
prepare a new type of scaffold with purposely designed macro‐ and micro‐porosity. The dual
macro/micro porous scaffolds of medical‐grade polycaprolactone (mPCL) were characterized for
their porosity, surface area, mechanical properties and degradation. The use of these scaffolds
for local prophylactic release of Cefazolin to inhibit S. aureus growth was investigated as an
example of drug delivery with this versatile platform. The introduction of microporosity and
increased surface area allowed for loading of the scaffold using a simple drop loading method of
this heat‐labile antibiotic and resulted in significant improvement in its release for up to 3 days.
The Cefazolin released from scaffolds retained its bioactivity similar to that of fresh Cefazolin.
There were no cytotoxic effects in vitro against 3T3 fibroblasts at Cefazolin concentration of up
to 100 µg/ml and no apparent effects on blood clot formation on the scaffolds in vitro. This study
therefore presents a novel type of scaffolds with dual macro‐ and micro‐porosity manufactured
by a versatile method of 3D printing combined with salt‐leaching. These scaffolds could be useful
in tissue regeneration applications where it is desirable to prevent complications using local
deliver of drugs.
2.2 INTRODUCTION
Scaffold‐based tissue engineering inherits a similar risk of surgical site infection as any
other implant based treatment. The use of implanted foreign body in surgery has been shown to
reduce the minimal infecting dose of S. aureus up to 1000 fold for formation of a permanent
abscess [61]. In fact, implants are recognized as a major contributing factor to these infections
because of the formation of bacterial biofilms, a self‐secreted polysaccharide matrix which
Page 17
Page 17
adheres to implant surfaces and protects the embedded bacteria from the immune system and
systemically administered antibiotics [5]. Surface modification and local delivery of antimicrobial
agents has been proposed as a promising approach in preventing biofilm formation [62‐66].
Infections associated with implants occur at significant rates (e.g. 0.8‐1.2% for total
hip/knee replacement [6], 5‐15% for several breast implants [7, 8]). They represent a challenging
problem and burden on healthcare systems in terms of morbidity, mortality and treatment costs
across all surgical disciplines. The two commensal bacteria normally found on patient’s skin
Staphylococcus aureus (S. aureus) and Staphylococcus epidermidis (S. epidermis) have been
identified as key pathogens in many of these infections [6]. Surgical site prophylaxis is currently
the standard approach for infection prevention with systemic antibiotic administration before
and after surgery. However, this mode of administration leads to low local antibiotic
concentrations in the desired tissue and side effects such as organ toxicity and altered host
microbiome[9]. It may also promote the development of antibiotic resistance in commensal
bacteria in the body, which can then themselves cause disease or spread their resistance via
plasmid transmission [10].
The use of locally delivered antibiotics for surgical site prophylaxis is a promising
approach for scaffold‐based tissue engineering applications. Traditional methods of mixing
antibiotic powder with polymers before manufacturing the mixture into cement for implantation
are simple but have major disadvantages when heat‐ or solvent labile antibiotics are used [14,
15]. A large number of commonly used antibiotics including all of the Cephalosporin family, which
are the mainstay of surgical site prophylaxis, have markedly reduced activity when exposed to
heat [16]. The same analysis also applied to local scaffold‐based delivery of other drugs and
molecules such as DNA, growth factors or chemotherapy drugs. Therefore, it is clear that
advanced scaffold technologies should have scaffold fabrication methods that are independent
from the drug loading process and that the latter should be designed to allow for in‐theatre
Page 18
Page 18
preparation. These criteria are important for the clinical translation of these scaffolds as they
permit treatment personalized to specific patients and surgical scenarios.
We hypothesized that microporous architecture in a scaffold would increase its surface
area and facilitate the incorporation and retention of drug from its solution through simple drop
or soak loading methods. 3D printing and screw melt extrusion (SME) techniques have emerged
as powerful tools to create patient‐specific scaffolds with macro‐porous architectures and
mechanical properties suitable for cellular infiltration, vascularization, tissue formation and
remodelling. SME is particularly versatile in terms of input materials and has been used to make
scaffolds with local drug delivery capability by modifying the input materials [59]. For example,
Hollander et al. made indomethacin‐PCL feeding filament by thermal extrusion before printed to
make T‐shaped scaffolds for drug delivery application in intrauterine system [67] . Yet the main
limitation of mixing the input materials with drugs is that the drugs have to go through the
material and scaffold fabrication processes, which often involve elevated temperature and/or
harsh chemicals, which can inactivate and/or degrade drugs/antibiotics. In this study, we thus
focus on employing SME technique to create scaffolds having intra‐bar/strut microporosity which
would allow for loading of heat/chemical sensitive drugs. However, introducing microporosity
into these scaffolds been reported only by a limited number of research groups and for only
certain specific materials [68, 69].
Here, we used a simple concept of combining SME with salt‐leaching and developed a
new type of 3D printed scaffold with dual micro‐and macro‐porosity that offers versatile drug
loading and controlled release capabilities. The objective of this study is to demonstrate the
feasibility of using PCL scaffolds with macroscale and microscale porosity for antibiotic delivery
to prevent infection and the potential of combining the new scaffolds with coating method to
control the loading and burst release of the chemotherapeutic and antibiotic agents. We
employed our home‐built screw melt extrusion device (Figure 1a) instead of coupling the
commercialized extruder [70]. Our device has a larger feeding chamber that feeds directly into
Page 19
Page 19
the extruder nozzle and mobilized nozzle and collecting board, it can be used for any materials
provided that they have low viscosity to print. The device changed from 2 step process, extruding
material filaments and then printing the scaffolds from the filaments, to 1 step process, printing
the scaffolds directly from the molten composite. We used polycaprolactone as a model polymer
because of its proven applications in many FDA‐approved medical devices, and its low melting
temperature (60oC) but the manufacturing method is designed as applicable to other
thermoplastic materials to fit various applications. Cefazolin was selected as a model antibiotic
because it is commonly used for prophylaxis against S. aureus infection and is particularly
sensitive to elevated heat and organic solvents making it unsuitable for commonly used
techniques [71].
2.3 MATERIALS AND METHODS
Scaffold Design & Manufacturing: Medical grade poly(ε‐caprolactone) (mPCL; Mn =70
kDa, Mw =120 kDa, Polydispersity = 1.69, in pellet form; from Purasorb PC12 Corbion, Purac) was
used to 3D print scaffolds. In order to achieve microporosity, mPCL was dissolved in chloroform
(Sigma Aldrich, St Louis) and mixed with phosphate buffered saline (PBS) particles that had been
ground from PBS tablets (Sigma Aldrich, St Louis), and sieved to size less than 75 μm at a weight
ratio of 5:4 (PCL: PBS). The porogen‐mPCL suspension in chloroform was then solvent‐casted to
create composite porogen‐mPCL films which were left in a fume hood to dry and stored in a
desiccator (Figure 1b). The porogen‐mPCL composite films were used to 3D print constructs
50mm x 50mm x 2mm in a laydown pattern of 0/60/120° as reported elsewhere[72]. Briefly, PCL‐
porogen films were heat up to 110oC for 30 mins and then screw‐extruded through a metal nozzle
of gauge 2 (inner diameter of 0.33 mm). The x‐y movement of the motorized stage and z‐
movement of the nozzle was computer‐controlled to produce 3 dimensional scaffolds with bar
diameter of 0.3 mm and bar distance of 1 mm. The scaffolds were then left in deionized water
for 4 weeks to allow leaching of the porogen to produce macroporous scaffolds with microporous
Page 20
Page 20
intrastrut/bar architecture (hereafter referred to as “microporous scaffold”). mPCL only was also
used to make scaffolds with solid strut/bar (hereafter referred to as “solid scaffold”) as reference
using the same manufacturing process. Both types of scaffolds were then treated with 1M NaOH
for 30 minutes on an orbital shaker to improve their hydrophilicity. Polycaprolactone is a
polyester which is hydrolyzed by strong base such as sodium hydroxide and this hydrolysis
introduces carboxylic and hydroxyl groups to the surface therefore increases the hydrophilicity.
A 5‐mm biopsy punch was used to produce scaffolds of 5 mm diameter x 1.5 ‐2 mm thickness as
samples for experiments. Samples were sterilized by ethanol followed by UV irradiation.
Figure 1. (a) home‐built screw melt exutrsion device. (b) Schematic of fabrication process of PCL
scaffolds using SME and salt leaching techniques.
After leaching of mPCL over 14 days, scaffolds were dissolved in chloroform to 10% (w/v) and
the solution’s cloudiness was measured at 600 nm using a spectrophotometer (Beckman
(b)
(a)
Temperature control unit
Voltage (speed) control unit
Collecting board
Nozzle
Heating coils
Feeding chamber
Pressurized airExtruding unit
Page 21
Page 21
Coulter Inc., California, US). The amounts of porogen residual were determined using the
standard curve made from porogen mixture in mPCL 1% solution in chloroform.
Cefazolin Loading: Scaffolds were loaded with Cefazolin (Sigma‐Aldrich, St Louis) via a “solution
drop loading” technique. Cefazolin was first dissolved in 67% ethanol to give solutions of 745
ppm, 1340 ppm and 3350 ppm. In a biosafety cabinet, scaffolds were placed in a 96‐well plate
16±0.3 mg scaffold each well and 5 μL droplet of solution was dropped onto each scaffold and
let dry. This loading process was repeated 3 times with at least 30 minutes in between to give
total loading doses of 11.17 μg (Low Dose), 20.1 μg (Med Dose) and 50.25 μg (High Dose) per
scaffold.
GelMA Coating: The scaffolds were dip‐coated using 2%(w/v) gelatin methacrylate (GelMA)
solution which was prepared as per a protocol published previously[73]. gelMA was prepared at
low concentration to achieve low viscosity that make it easily penetrate into the inner structure
of the scaffolds. The GelMA adsorbed on the scaffolds was then chemically crosslinked by
ammonium persulfate (APS) and Tetramethylethylenediamine (TEMED; Thermo Fisher
Scientific). An additional group (n=9) of each of the High Dose Solid and Microporous scaffolds
were left uncoated for comparison.
SEM Imaging: The scaffolds were gold‐sputter coated and examined using scanning
electron microscopy (Zeiss Sigma VP Field Emission SEM) operating at 5‐10 kV. X‐ray
microtomography (micro CT) imaging was conducted to characterize the scaffold
microarchitecture. X‐rays were generated by a GE‐Phoenix xs 180nf micro‐focus X‐ray source and
radiographs captured on a 3000x3000 silicon flat panel detector. X‐ray tube settings were 80 kV
and 110 microAmps and reconstruction undertaken using a helical cone‐bean filtered back
projection method [74]. The voxel size of the images was 2.2 m.
Porosity: The porosity was first calculated from the scaffold’s dimensions, weight and
PCL density. The total volume of a PCL cylindrical scaffold was calculated using the dimensions
Page 22
Page 22
measured with Vernier callipers, and the volume of PCL was calculated from scaffold’s weight
based on density of the mPCL (1.145 g/cm3 as per supplier). The porosity was calculated as below
and reported as mean ± s.e.m (n=4):
𝑝𝑜𝑟𝑜𝑠𝑖𝑡𝑦
, in which:
total scaffold volume =π x height x (diameter/2)2
mPCL volume = scaffold weight/mPCL density.
The porosity of the microporous scaffolds was also determined from micro CT data by
segmentation and image analysis of the 2D tomographic images. Each voxel of the sample was
assigned a phase based on the X‐ray attenuation corresponding to each voxel in the image. The
solid scaffold gives a strong bimodal distribution of attenuations associated with solid PCL and
open porosity between the struts. The microporous scaffold has additional porosity inside the
struts and bars.
Surface Area: The specific surface areas of the mPCL and microporous‐mPCL scaffolds
were measured by a surface area and porosity analyser (Tristar II 3020 Surface Area Analyser,
Micromeritics, Ottawa). Surface area was measured by performing N2‐adsorption–desorption
experiments at 77 K. The sampling tubs were immersed in liquid nitrogen and the samples were
degassed under N2 flow before analysis. Surface areas were calculated using Brunauer–Emmet–
Teller (B.E.T) model.
Mechanical Testing: Scaffold’s compressive modulus was determined using Instron 5567 (Instron,
Massachusetts) uniaxial testing system with a 5kN load cell and maximum displacement of 15%
(n=3).
Zeta Potential: A SurPASS electrokinetic analyzer (Anton Paar) was used for surface zeta
potential measurements of scaffolds. In each zeta potential recording, two identical samples
were attached to the adjustable gap cell to face each other with a 100 µm gap between them.
The streaming current was measured between two Ag/AgCl electrodes placed at both sides of
the samples. The measurements were performed using 1 × 10−3 m KCL solution as the electrolyte.
Page 23
Page 23
The VisioLab interface calculated the zeta potential from the streaming current measurements
according to the Fairbrother–Mastin approach (1)
(1)
Where dU/dp is the slope of the streaming current versus pressure, η is the viscosity of the
electrolyte, ε is the dielectric constant of the electrolyte, ε0 is the vacuum permittivity and KB is
the specific electrical conductivity of the electrolyte solution outside the capillary system.
Accelerated degradation: PCL is a polyester of slow hydrolysis‐based degradation rate;
therefore, we compared degradation properties of the PCL scaffolds using accelerated
degradation in NaOH (Figure 2). Samples were placed in NaOH 2M solution and incubated at 37oC
in shaking incubator at 200 rpm. The samples were recovered, washed with miliQ water and dried
in vacuum until stable weights. The molecular weight of samples was determined by gel
permeation chromatography (GPC).
Gel permeation chromatography: GPC was used to study the change in molecular weight
of PCL scaffolds after degradation in NaOH solution. Samples were dissolved in Chloroform with
a concentration of 0.5%w/v and the solutions were filtered with a 0.22 μm filter. The molecular
weights of samples were measured by Waters 1515 Isocratic HPLC Pump (Waters, Milford, MA)
equipped with Waters 2487 Dual Absorbance detector. Chloroform was used as a mobile phase
at a flow rate of 1 mL/min. 10 μL of sample was injected for each measurement and the
measurement was run for 40 minutes.
Figure 2. Schematic of accelerated degradation experiment
Page 24
Page 24
Cefazolin release from scaffold: Release of Cefazolin from scaffolds was tested in sterile
PBS with the PBS volume/scaffold weight ratio of 10 μL/mg in a shaker incubator set at 37°C and
225 rpm. The eluate was collected and replaced with fresh sterile PBS every hour for the first 6
hours and then every day for 7 days. The concentration of the Cefazolin in eluate was determined
by measuring absorbance at 272nm using a spectrophotometer (x‐Mark, BioRad, California) and
a standard curve of known Cefazolin concentrations.
Biocompatibility Testing: Cytotoxicity was tested using murine embryonic 3T3 fibroblasts
(CCL‐92, ATCC, VA, USA) and AlamarBlue assay. A standard curve was first created by culturing
cells to 70‐80% confluence in complete Modified Eagle Medium (αMEM, Thermo Fischer
Scientific) which was supplemented with 10%v/v fetal bovine serum (FBS, Sigma‐Aldrich) and 1%
penicillin–streptomycin (PS, Sigma‐Aldrich), cells were seeded in wells of a 96‐well plate with
decreasing density in 100 μL complete αMEM. The plates were incubated at 37°C for 12 hours,
then the culture medium removed and 100 μL of 10% Alamarblue in complete αMEM was added
for a further 4 hour incubation after which 50 μL from each well was transferred to a black 96‐
well plate and fluorescence intensity was measured using a plate reader (Polarstar Optima, BMG
Labtech, Ortenberg) with excitation of 544 nm and emission of 590 nm. For Cefazolin cytotoxicity
testing, the cells were seeded in a 96‐well plate at a density of 104 cells in 100 μL complete αMEM
and cultured for 24 hours at 37°C, culture media was removed and replaced with serially diluted
solutions of Cefazolin in complete αMEM . The plate was incubated for 24 hours at 37°C, then
medium removed and 100uL 10% Alamarblue in αMEM added, after a further 4 hours incubation
the fluorescence intensity of the solution was measured as above. Cells treated with 100%
Phenol and culture media only serve as positive and negative controls. Experiments were
repeated 3 times with n=3 replicates in each repeat (n=3).
Antibiotic Bioactivity Testing: S. aureus (ATCC29213) was used for the experiments.
Growth of the bacteria was first characterized. Bacteria colonies were selected from agar plates
and inoculated into 10 ml Lysogeny Broth (LB) and incubated for 12‐16 hours on a shaker
Page 25
Page 25
incubator set at 37°C and 225 rpm. The absorbance at 600 nm (OD600) of 100 μL bacterial
suspension was measured and correlated to the number of colony forming units (CFU)
determined by the agar plate spreading and counting method. From the overnight culture,
bacteria were diluted to 5x108 CFU/mL in LB and incubated in a shaker incubator at 37°C 225 rpm
and the OD600 and CFU counts were determined every hour for 6 hours to characterize the early
log phase bacterial growth. The late log phase growth of bacteria was characterized by diluting in
1:1 double strength cation‐adjusted Muller‐Hinton Broth (2xCAMHB, with 40 mg/L of CaCl2 and
20 mg/L of MgCl2) and PBS, and plating in a 96‐well plate to give 100 μL of 5x106 CFU/ml per well,
the plate was then incubated at 35°C statically and the OD600 measured and CFU counted by
plating method every hour for 8 hours.
The minimum inhibitory concentration (MIC) of fresh Cefazolin to kill 50% and 90%
(MIC50 and MIC90) of late log phase S. aureus was determined by adding 5x105 CFU bacteria into
each well of a 96‐well plate followed by adding serially diluted Cefazolin solutions and cultured
for 12‐16 hours 35°C statically. PBS solution was used as a negative control. Testing the bioactivity
of the Cefazolin in the eluates released from scaffolds was carried out using this protocol by
diluting eluates in double strength CAMHB containing 5x105 CFU bacteria and incubating for 8
hours in a shaker incubator at 35°C, 225 rpm. The absorbance at 600 nm (OD 600) was then
measured and the CFU/ml was calculated from the pre‐established standard curve and
normalized by that of PBS only negative control.
Agar diffusion testing: S. aureus was plated on Mueller‐Hinton agar, and scaffolds
(square, 2 mm x 2 mm x 1.5 mm) loaded with Low Dose, Medium and High doses of Cefazolin
were placed and gently pressed to have stable contact with agar to allow for diffusion of drug
into the agar. The plates were incubated at 37oC for 12 hours and zones of inhibition were
measured with a ruler.
Blood Clot Formation: Sheep whole blood was freshly collected into Vacuette® tube
(Greiner Bio‐one, Kremsmünster) 3.5mL 9NC coagulation sodium citrate 3.2% and stored at 4oC.
Page 26
Page 26
Blood and CaCl2 0.2M were warmed to 37oC for 1 hour prior to experiments. 190 µL blood was
added into each 2‐mL Eppendorf tube and a scaffold of known weight was added to the tube.
Then 10 µL of CaCl2 0.2M was added to each tube and gently mixed. The tubes were incubated
for 2 hours. Finally the clots were gently collected from the tubes and weighed. Solid scaffolds
and microporous scaffolds loaded with High Dose Cefazolin was compared with a control group
of no scaffolds (n=5).
Statistical Analysis: Data were reported as mean ± SEM unless otherwise indicated, and
statistical significance between groups was determined by two‐tailed student’s t test.
2.4 RESULTS
2.4.1 Scaffold manufacturing and characterization
The SEM images were taken at 3 magnifications to show the morphology of PCL
scaffolds, the cross sections and surfaces of the struts and bars. SEM imaging showed clear
presence of pores in micrometre scale within the microporous scaffold struts (Figure 3a and c) as
the result of the salt‐leaching process compared with solid struts of the normal mPCL scaffolds
(Figure 3b and d).
Page 27
Page 27
Figure 3. Scaffold characterization. Overview images (insets) and SEM images of scaffolds at
increasing magnifications for microporous scaffolds ((a1)‐(a3)), normal scaffolds (solid strut, (b1)‐
(b3)) and GelMA coated microporous scaffolds ((c1)‐(c3)) and GelMA coated solid scaffolds ((d1)‐
(d3)). Scale bars = 500 μm for (a1)‐(d1); 200 μm for (a2)‐(d2); 50 μm for (a3)‐(d3). Micro CT images
of e) microporous scaffold and f) solid scaffold at two representative cross sectional planes.
Microporosity analyzed by measuring in 2D slice images from micro CT is c.a. 41%. Circular non‐
interconnected pore present in the solid scaffolds (f) are believed to result from air entrapment
during the extrusion process. g) The leaching process of mPCL over 14 days indicating complete
leaching of porogen was achieved. Scaffolds were dissolved in chloroform to make solutions of
10 %( w/v). Any porogen remained in the scaffold would give rise to the solution absorbance
which was measured with a spectrophotometer.
The majority of micropores have relatively similar sizes ranging from c.a. 20 μm to c.a. 70
μm with few particularly larger pores which were likely the result of porogen aggregation during
Page 28
Page 28
the sample making processes. Microporosity was further characterized using micro‐CT which
interestingly showed occasional spherical pores in the solid scaffolds (Figure 3e and f). The
presence of these spherical pores in the struts and bars of the solid 3D scaffolds was previously
documented by our group and is likely the result of air bubbles entrapped during the melt‐
extrusion process, they are sparse and not interconnected. GelMA coating is visible as only a thin
material at the edge of the struts, most noticeable where the struts fuse (Figure 3c and d). The
leaching process demonstrated the progression of the removal of porogen during immersion and
was complete after 14 days (Figure 3g). This conclusion is also supported by examination of the
cross sections of 14 days leached scaffolds with SEM and micro CT showing the absence of
porogen particles (Figures 3 a1 and e).
The surfaces of microporous mPCL and mPCL scaffolds were found similar in terms of
microscopic morphology, roughness (Figure 4a and b). The surface charges of these scaffolds
were also not different (Figure 4g). The measured similar contact angles (Figure 4c) were
therefore expected.
The overall porosity (i.e., the total of microporosity and macroporosity) of the
microporous scaffolds calculated using mPCL density and scaffold dimensions was 80±1%
compared with 62±0.3% for the solid scaffolds (Figure 4d). Since the two types of scaffolds were
made using the same parameters of the 3D printing process, they were assumed to have the
same macroporosity; the microporosity within the scaffold struts in the microporous scaffold was
then calculated to be approximately 47%. This is in good agreement with the total microporosity
of 41±11% obtained from analysis of microCT images.
Page 29
Page 29
Figure 4. Physicochemical characterization of microporous mPCL scaffolds compared with mPCL
scaffolds showing mean + SEM for: a)& b) AFM imaging and surface roughness analysis showing
non‐significant difference in topography; c) contact angles demonstrating similar hydrophilicity
as expected from surface hydrolysis of mPCL during leaching d) porosity determined using
scaffold dimensions, weight and mPCL density (n=4). Total porosity was significantly higher for
microporous scaffolds at 80±1% compared with 62±0.3% for the solid scaffolds. e) surface areas
Page 30
Page 30
measured with gas adsorption and analyzed with B.E.T method was also significantly higher for
the microporous group (data is average of 2 measurements) and f) compressive modulus (MPa)
over three compressing ‐ relax cycles (n=3) showing non‐significantly lower values for
microporous scaffolds. g) Mean Zeta potential for solid and microporous scaffolds showing both
having negative surface charges with no significant difference as expected from the same post‐
printing treatment procedure to both scaffolds. (* p < 0.05).
Surface Area: The specific surface area of the microporous scaffolds was also
significantly higher for the microporous scaffolds at 4887±88 cm2/g compared with 1260±76
cm2/g of the solid scaffolds (Figure 4e).
Mechanical Testing: The microporous scaffolds showed lower (non‐significant)
compressive modulus compared with the solid mPCL scaffolds with the largest difference for the
first cycle of compression‐relaxing experiment (Figure 4f). This is expected as the intra‐strut
micropores would make the microporous scaffolds softer which could be beneficial for
applications such as in soft tissue regeneration where it is desirable to have scaffolds
mechanically similar to the native soft tissue.
2.4.2 Degradation test.
The accelerated degradation test demonstrated increased porosity in microporous
scaffolds evident on scanning electron microscope imaging around the perimeter and centrally
within the struts (Figures 5b1, b2, d1 and d2). The molecular weight increased for the solid mPCL
is attributed to the removal of amorphous/less crystalline regions of the polymer in the first steps
during degradation. In contrast, a decrease in Mw was seen for the microporous mPCL and this
is attributed to the fact that the increased surface areas and porosity accelerated the degradation
and has led to bulk erosion in the case of these scaffolds (Figure 5e). This conclusion is supported
Page 31
Page 31
by the stark difference in morphologies of mPCL and microporous mPCL scaffolds both on the
surface and in the cross sections in the SEM images (Figure 5 a, b, c and d)
Figure 5. Accelerated degradation test. mPCL and microporous mPCL scaffolds were immersed in
NaOH 2M in a shaker incubator (at 37oC, 121 rpm) for 48 hours. SEM imaging revealed the
substantial increase in porosity both on the surface and cross section of the microporous scaffold
(scale bars = 100 μm). (e): GPC analysis showed results of molecular weight changes after the
test; mPCL scaffolds exhibited slight increase in Mw likely due to the removal of amorphous/less
Page 32
Page 32
crystalline region during the degradation from the scaffolds. The reduction in Mw of microporous
scaffolds is attributed to the bulk erosion as seen in SEM images on the surface and cross section
of the scaffolds.
Weight loss measurement showed negligible reduction for mPCL scaffolds and almost
linear reduction of c.a. 65% in 48 hours in NaOH 2M.
This degradation data indicated that the microporous scaffolds facilated the degradation
of PCL which is preferable in many regenerative applications such as for soft tissue where the
current slow rate (~ up to 2 years for complete degradation) is a limitation for PCL.
2.4.3 Loading and release of antibiotics and its bioactivity.
Cefazolin release: Both types of scaffolds showed capability to load and subsequently
release Cefazolin. Microporous scaffolds demonstrated a clear dose‐dependent release
characteristic while solid scaffolds showed no significant difference between the Medium Dose
and High Dose in most of the time points tested which is likely due to a smaller surface area
compared with microporous mPCL scaffolds (Figure 6[A]).
Page 33
Page 33
Figure 6. Characterization of Cefazolin release kinetics from scaffolds. [A]: comparison of
Cefazolin non‐cumulative release profiles for GelMA coated microporous mPCL and solid mPCL
Page 34
Page 34
over the first 6 hours and next 7 days reported as mean±SEM (n=9) a) The released Cefazolin
concentrations showed dose dependence in both normal and microporous scaffolds. The
microporous scaffolds also showed consistently higher release up to day 6 (comparing within the
same dose group). b) The total Cefazolin amount released at each time point (in μg) was
consistently higher for microporous compared with normal scaffolds for 3 consecutive days. [B].
Comparison of Cefazolin release profiles in PBS for microporous mPCL scaffolds with and without
GelMA coating (n=9) at high loading dose a) total release from scaffolds with a reduced burst
release for GelMA coated scaffolds which released significantly less in the first 6 hours and b)
release at each time point presented as percentage of total released, with the GelMA coated
scaffolds showing a more gradual release indicating a reduced burst release.
The Cefazolin concentration the eluates from Cefazolin–loaded microporous scaffolds
ranged from a mean of c.a. 25 ppm in the first hour for the High Dose group to 0.25 ppm at day
7 for the Low Dose group (Figure 6[A]‐a). The High Dose microporous scaffolds produced a
concentration above the MIC90 for 4 consecutive days. The microporous scaffolds also released
significantly more drug than the solid scaffolds for 3 days (Figure 6[A]‐b). The mean release per
scaffold increased with loading concentration at the time points tested. The porous scaffolds
demonstrated an initial ‘burst release’ in the first few hours followed by a sustained release over
7 days. Importantly, the microporous scaffolds with GelMA‐coating exhibited a lower burst
release (Figure 6[B]‐a) (i.e., release in the first 6 hours) compared with scaffolds without GelMA
coating.
Loading efficacy: Loading efficacy was calculated as the ratio of total amount released to
the original drop–loaded Cefazolin amount. The loading efficacy was significantly higher for the
GelMA coated microporous scaffolds at 4.10.2, 5.60.5, 7.00.4% for High Dose, Med Dose and
Low Dose respectively compared with 1.50.1, 4.20.3 and 5.50.5% for the solid scaffolds. The
High Dose uncoated scaffolds showed higher loading efficacy of 7.41.7% and 17.61% for solid
Page 35
Page 35
and microporous respectively but this was mostly seen as a burst release in the first hours with
reduced subsequent release.
Bioactivity Testing: The S. aureus strain showed typical initial lag phase followed by logarithmic
early and late growth phase (Figure 7[A]‐a) which would typically be followed by a stationary
phase however this was not tested in our study. Testing using fresh Cefazolin solution showed
an MIC90 of c.a. 0.8 ppm and an MIC50 of c.a. 0.1 ppm (Figure 7[A]‐b). Bioactivity of Cefazolin
eluates collected at different time points from microporous mPCL scaffolds loaded with Low
Dose, Medium Dose and High Dose demonstrated dose‐dependent antimicrobial activity
(Figure 7[B]) from which MIC90 of eluted Cefazolin was calculated to be approximately from 0.7
to 1.3 ppm, and an MIC50 of c.a. 0.3 ppm which is similar to that of fresh Cefazolin (Figure 7[A]‐
b). The antimicrobial activity of the microporous scaffold loaded with Cefazolin was also
demonstrated through the zone of inhibition experiment (Figure 8c) which showed a dose‐
dependent zone of inhibition on the loading doses.
Page 36
Page 36
Figure 7. Antimicrobial activity of Cefazolin on S. aureus. [A] Effects of Cefazolin on planktonic S.
aureus growth. a) Characterization of S. aureus growth demonstrating a typical growth curve of
the bacteria including an initial lag phase (3 hours) followed by log growth phase (8 hours)
(n=3). The stationary phase was not tested in this experiment. Bacteria in the log growth phase
were then used for subsequent antibiotic testing experiments. b) Growth curves of S. aureus
treated with increasing concentrations of Cefazolin, demonstrating the MIC50 = 0.1 ppm and
MIC90=0.8 ppm where 50% and 80% of bacterial growth was inhibited respectively (n=3) [15].
[B] Bioactivity of Cefazolin eluates collected at different time points from microporous mPCL
scaffolds loaded with Low, Medium and High Dose of Cefazolin determined as CFU normalized
by that of untreated control. Data = mean ± SEM (N=3 repeats with 3 replicates in each repeat).
The eluates demonstrated dose‐dependent antimicrobial activity from which MIC90 of eluted
Cefazolin was calculated to be c.a 0.7 to 1.3 ppm, and an MIC50 of c.a. 0.3ppm which are similar
to those of fresh Cefazolin in [A]b).
Page 37
Page 37
2.4.4 Biocompatibility testing
Metabolic activity (shown to correlate well with cell number Figure 8a) assays showed
slight increase in cells treated with low Cefazolin concentration (< 1 ppm) and remained high for
concentrations up to 100 ppm (Figure 8b). This is in agreement with other studies which show
limited toxicity of Cefazolin below 100 ppm[75].
Formation of a stable blood clot in macroporous scaffolds after implantation has been
recognized recently by several groups as important for subsequent cellular infiltration and
vascularization [76, 77]. Here we observed no detrimental effects of microporosity and Cefazolin
loading on this process as indicated by similar clot weights (Figure 8d).
Figure 8. In vitro testing of biocompatibility and activity. Alamarblue assay on 3T3 fibroblast cells.
a) Standard curve correlating fluorescence intensity (excitation: 544 nm, emission: 590 nm) and
fibroblast cell number (data = mean± SEM, n=3) showing strong correlation (R2=0.999). b)
Page 38
Page 38
Fibroblast cell number (data = mean± SEM, N=3) after 24 hour treatment with different
concentrations of Cefazolin indicating no cytotoxicity of Cefazolin concentration of up to 100ppm.
c) Agar diffusion test of Cefazolin–loaded microporous scaffolds (without GelMA coating)
showing dose‐dependent zone of inhibition on S. aureus. * p<0.05, data = mean ± S.E.M (n= 8).
d) In vitro blood clot formation on scaffolds. No significant difference in blood clot weight formed
in presence of microporous PCL scaffold without (PCL scaffold) and with high dose of Cefazolin
(PCL Scaffold with Cefazolin) compared with control group without scaffold (blood only) (n=5).
2.5 DISCUSSION
It is being recognized that tissue engineering scaffolds should be designed to have
multiple functions to (i) facilitate cellular attachment, infiltration and vascularization and also (ii)
prevent complications such as infection to ensure successful treatment outcomes [78, 79].
Criterion (i) has been realized by using scaffold surface treatment and scaffold’s macro‐porosity
with interconnected pores of around 300 µm or above which is known to be crucial in
vascularization of implanted scaffolds in regeneration of clinically relevant volume of tissue [80].
Solid‐free form fabrication techniques such as SME are emerging techniques that allow for
generation of patient‐ specifically shaped constructs with such properties [81]. Criterion (ii) has
been often realized using drug‐delivery approach where drugs (such as antibiotics,
chemotherapeutics) are incorporated into the raw materials during the scaffold preparation [82,
83]. Yet, major limitations of this method include instability issue of the drugs and lack of control
on specificity in dosing.
This study aimed to develop a versatile scaffold platform to enable loading of drugs
independently from scaffold fabrication process. A traditional porogen leaching technique was
combined with SME to produce polymeric scaffolds with a macroporous struts/bars structure
that have intra‐strut/bar microporosity. The method of porogen leaching to create porogen was
Page 39
Page 39
chosen over other pore making methods thanks to its versatile properties. The main goal of this
study is to modify the PCL scaffolds, which have been previously proven to have good promise
for tissue reconstruction, to have the second function – drug delivery. The use of salt leaching can
utilize the intra‐strut volume to the PCL struts. Compared with other method, the microscale
pores inside the strut can be easily generated by a simple leaching process and pore size and
porosity of the intra‐strut structure can be easily altered by changing size, shape and
concentration of porogen.
The introduction of microporosity in macroporous polycaprolactone scaffolds allowed
for rapid and specific loading of model drug Cefazolin and its efficient release in a dose dependent
manner. This study therefore demonstrated the proof‐of‐principle that drugs such as antibiotics
could be readily loaded to scaffolds using simple solution‐based, in‐operation‐theather‐ready
loading method, and its release could be controlled and tailored for different applications in
surgical site infection prophylaxis.
A number of factors have been recognized to control the loading and release kinetics of
water soluble drugs such as Cefazolin from carrier devices [84‐86]. In this study, the loading of
microporous scaffolds was correlated with increased surface area, hydrostatic forces and surface
etching with NaOH. The microporous intrastrut/bar architecture was shown to increase the
release efficacy with significantly more Cefazolin released for 3 days, and a greater overall total
release. This is clinically relevant in the context of inhibiting the growth of bacteria introduced at
surgery for tissue engineering applications where there are often constraints such as on scaffold
weight and volume. These results also suggest that the drug release profile could be further
controlled by varying the microporosity and/or pore size through porogen concentration and size
during scaffold manufacturing.
The use of a GelMA barrier coating was shown to dampen the initial burst release, which
is recognized as undesirable in many cases because of the potential local toxicity and/or
subsequent release of sub‐therapeutic antibiotic levels. The combination of microporosity and
Page 40
Page 40
GelMA coating showed a smaller initial burst release over a few hours followed by an extended
release over the next few days. This release profile is highly desirable for an antibiotic eluting
device so as to target the risk of infection during initial inoculation and continue to prevent latent
infection [85]. The use of the GelMA is analogous to the barrier coating on a reservoir type drug
eluting device. These systems often include a drug‐containing polymer core and a drug free top
layer that acts as a rate‐limiting barrier leading to sustained drug release [87]. A small initial burst
release can also be observed in these systems due to the migration of the drug into the top layer
after preparation [84]. In this study, GelMA was chosen because of its versatility in production
and biological properties; it has been shown to encourage cellular adhesion and proliferation and
it can be readily photon – or chemically cross‐linked [88].
The released Cefazolin showed high retention of bioactivity which highlights an
important advantage over traditional systemic administration of sustained local delivery systems
where drug degradation in aqueous solution is greatly reduced[89]. In our experiments, there
was only slight decrease in MIC50 and MIC90 of Cefazolin released from scaffolds indicating that
the scaffold‐bound Cefazolin retained its bioactivity even when the scaffolds were immersed in
PBS. Another important advantage of scaffold‐based local delivery is that peak concentration of
Cefazolin around the scaffold is obtained immediately after implantation, which would be
effective in targeting bacteria inoculated during surgery. Reports in the literature show that after
systemic administration it takes up to 2 hours for tissue concentration of Cefazolin to reach peak
and the pharmacokinetics of single dose resulted in peak tissue concentrations of up to 24 ppm
with a half‐life of only 3 hours resulting in concentrations dropping below the MIC for S. aureus
in under 24 hours [90].
There are few reports in the literature focusing on local delivery of water soluble drugs
such as Cefazolin from polycaprolactone vehicles. One such study used PCL incorporating foam
pads made by freeze drying a solution of PCL and Cefazolin applied at the interface of a metal pin
to prevent infection in a rabbit tibia model [91]. This report showed strong in vitro release of
Page 41
Page 41
Cefazolin in the first day followed by more gradual release in days 2 and 3 and negligible release
afterward. The application of a Cefazolin–loaded pad in the rabbit reduced the rate of S. aureus
detection in the surgical wound. This Cefazolin‐PCL foam is potentially useful but this method of
incorporating Cefazolin has several limitations including limited solubility, potential degradation
of Cefazolin in the solvent and no system for achieving exact doses for specific scaffolds. The
method of blending/mixing PCL with a drug is perhaps the most common approach reported in
the literature to make drug delivery systems (scaffolds, meshes, membranes etc.). Teo et al.
blended Gentamicin sulphate with PCL and tricalcium phosphate (TCP) and extruded the mixture
to create mesh capable of local delivery of gentamicin to prevent wound infection [92]. However,
the elevated temperature used in such thermal blending process has the potential to cause
significant drug degradation. Another group 3D printed PCL/PLGA scaffolds loaded with
tobramycin and demonstrated bactericidal activity against S. aureus in vitro and successful
treatment of established bone infection in a rat model[14]. This experiment was aimed at treating
infection and used Tobramycin a known heat stable antibiotic, however this antibiotic is not
appropriate for prophylactic use and there are no antibiotics commonly used to prevent infection
which are heat stable [16]. Cryomilling could also be used for blending but it is an energy‐intensive
and lengthy process and obtaining the exact drug dosing for specific scaffolds is still challenging.
Here we developed a new strategy to independently control the scaffold manufacturing and drug
loading processes. Our novel scaffolds with micropores and increased surface area offer a
versatile platform for drug incorporation using a simple solution soak/drop loading method and
allow for specific dosing to suit different needs.
Both the microporous scaffolds with and without Cefazolin showed no apparent adverse
effects on blood clot formation. The formation of a haematoma after the surgical implantation of
a tissue engineering scaffold and establishment of a blood clot are recognized as essential for
cellular recruitment and the regeneration process in tissue engineering [93]. The stabilized clot
acts as a supporting matrix for cell migration, provides growth factors and pre‐cursor cells and
Page 42
Page 42
facilitates the formation of a vascular network [93]. It is therefore imperative that a tissue
engineering construct does not impede blood clot formation as was demonstrated with these
scaffolds.
A limitation in our experimental procedure was the fact that our methods of drop‐
loading of antibiotic onto scaffolds and dip coating with GelMA has not been optimized, which is
believed to have resulted in low overall loading efficacy (less than 10%). This drop loading method
was used so as to devise an in‐theatre‐preparation method which would allow Cefazolin to be
loaded onto previously prepared microporous scaffold with specific dosing based on patient
characteristics and surgical scenario. This would also be more versatile than pre‐produced
antibiotic–loaded scaffolds and negates issues with antibiotic stability in storage. Soak loading
microporous scaffolds in antibiotic solution may be a more preferable method to achieve more
consistent and uniform loading. Further, the dip coating process of GelMA has not been
optimized as significant drug loss occurred during this rather inefficient process. Hence, future
experiments will focus on optimising these loading and coating methods. Beside loading method,
the release profile should also be modelized to deeply understand the release kinetic and the
applied drugs.
2.6 CONCLUSION
3D printed scaffolds with intra‐strut microporosity were created by combining screw
melt extrusion and salt‐leaching methods for applications in tissue engineering. The
microporosity allowed for loading and subsequent sustained release of Cefazolin using simple
solution drop loading method. The increased surface area and the micropores facilitated
Cefazolin loading with specific doses and the application of GelMA coating reduced the burst
release allowing for more sustained release compared with scaffolds of solid struts. This type of
scaffold design with micro‐ and macro‐porosity is promising for several tissue engineering
Page 43
Page 43
applications where loading and local delivery of drugs, nanoparticles or small molecules is desired
to achieve improved therapy outcome.
Page 44
Page 44
3Chapter 3: 3D printed dual macro‐,
microscale porous network as a tissue
engineering scaffold with drug delivering
function
Hoang Phuc Dang a, b, Tara Shabab a, b, Abbas Shafiee b, c, Quentin C. Peiffer a, b, Kate Fox d, Nhiem
Tran e, Tim R. Dargaville f, Dietmar Werner Hutmacher a, b and Phong A. Tran a, b, g
Published in Biofabrication 11 (2019): number 3
a ARC Centre in Additive Biomanufacturing, Queensland University of Technology, Musk
Avenue, Kelvin Grove, Brisbane, Queensland 4059, Australia
b Centre in Regenerative Medicine, Institute of Health and Biomedical Innovation,
Queensland University of Technology, Brisbane, Australia
c The University of Queensland, UQ Diamantina Institute, Translational Research Institute,
Brisbane, Queensland, Australia
d Center for Additive Manufacturing, RMIT University, Melbourne, Victoria, Australia
e School of Science, RMIT University, Melbourne, Victoria, Australia
f Institute of Health and Biomedical Innovation, QUT, Brisbane, Queensland, Australia
g Interface Science and Materials Engineering group, School of Chemistry, Physics and
Mechanical Engineering, QUT, Brisbane, Queensland, Australia
Page 45
Page 45
The authors listed below have certified that:
6. They meet the criteria for authorship in that they have participated in the conception,
execution, or interpretation, of at least that part of the publication in their field of expertise;
7. They take public responsibility for their part of the publication, except for the responsible
author who accepts overall responsibility for the publication;
8. There are no other authors of the publication according to these criteria;
9. Potential conflicts of interest have been disclosed to (a) granting bodies, (b) the editor or
publisher of journals or other publications, and (c) the head of the responsible academic unit,
and
10. They agree to the use of the publication in the student’s thesis and its publication on the QUT’s
ePrints site consistent with any limitations set by publisher requirements.
In the case of this chapter:
Published paper:
Hoang Phuc Dang, Tara Shabab, Abbas Shafiee, Quentin C. Peiffer, Kate Fox, Nhiem Tran, Tim R.
Dargaville, Dietmar Werner Hutmacher and Phong A. Tran. “3D printed dual macro‐, microscale
porous network as a tissue engineering scaffold with drug delivering function.” Biofabrication 11
(2019): number 3.
Contributor Statement of contribution
Hoang Phuc Dang
Conducted experiments, data analysis, prepared the first draft of
manuscript Signature
Date:
Tara Shabab
Conducted experiments and data analysis, manuscript
preparation
Page 46
Page 46
Abbas Shafiee Aided data analysis and manuscript preparation
Quentin C Peiffer Conducted experiments
Kate Fox Conducted experiments
Nhiem Tran Aided manuscript preparation
Tim R Dargaville Aided manuscript preparation
Dietmar W. Hutmacher Experimental design, aided manuscript preparation
Phong A. Tran
Experimental design, aided data analysis and manuscript
preparation
Principal Supervisor Confirmation
I have sighted email or other correspondence from all Co‐authors confirming their
certifying authorship
Phong Tran 06/02/2020
Name Signature Date
QUT Verified Signature
Page 47
Page 47
3.1 ABSTRACT
Macroporous scaffolds are important in tissue engineering therapy for regeneration of large
volume defects resulting from diseases such as breast or bone cancers. Another important part
of the treatment of these conditions is adjuvant drug therapy where drugs are administered to
prevent disease recurrence or surgical site infection. In this study, we developed a new type of
macroporous scaffolds that have drug loading and release functionality to use in these scenarios.
3D printing allows for building macroporous scaffolds with deterministically designed complex
architectures for tissue engineering, yet they often have low surface areas thus limiting their drug
loading capability. In this proof‐of‐concept study, we demonstrated that microporosity
introduced into macroporous scaffolds allowed for efficient loading of various clinical drugs using
simple soak loading methods and sustained their release. Manufacturing of scaffolds having both
macroporosity and microporosity remains a difficult task. We combined porogen leaching and 3D
printing to achieve this goal. Porogen microparticles were mixed with medical grade
polycaprolactone (MPCL) and extruded into scaffolds having macropores of 0.7 mm in size. After
leaching, intra‐strut micropores were realized with pore size of 20 – 70 μm and a total
microporosity of nearly 40%. Doxorubicin (DOX), paclitaxel (PTX) and cefazolin (CEF) were chosen
as model drugs of different charges and solubilities to soak–load the scaffolds and achieved
loading efficiency of over 80%. The microporosity was found to significantly reduce the burst
release allowing the microporous scaffolds to release drugs up to 200, 500 and 150 hours for
DOX, PTX and CEF, respectively. Finally, cell assays confirmed the bioactivities and dose response
of the drug–loaded scaffolds. Together, the findings from this proof‐of‐concept study
demonstrate a new type of scaffolds with dual micro‐, macro‐porosity for tissue engineering
applications with intrinsic capability for efficient loading and sustained release of drugs to prevent
post‐surgery complications.
Page 48
Page 48
3.2 INTRODUCTION
Scaffold‐based tissue engineering is promising in regenerating large volume tissue
defects, such as those resulting from surgical removal of primary bone or breast tumors [1, 2, 42,
94‐96]. As with other implanted devices, these scaffolds are prone to post‐surgery complications
such as surgical site bacterial infections or bacterial biofilm formation on the scaffolds. In the case
of patients receiving these scaffolds following tumor removal, cancer can recur because some
cancer cells were incompletely removed [97‐101]. Therefore, current standard treatments still
involve systemic administration of antibiotics or chemotherapeutic agents post‐surgery.
However, systemic toxicity and low accumulation of the drug at the tumor site remain major
limitations [102‐104]. In addition, side effects caused by the solvents that are required to dissolve
poorly soluble drugs prior to treatment cause further burden to the patients. For example,
paclitaxel (PTX), a chemotherapeutic agent commonly used to treat breast, lung, and ovarian
cancer, is solubilized in solvents such as Cremophor EL and dehydrated ethanol in the commercial
product Taxol [105]. Cremophor EL has been clinically associated with serious side effects, such
as hypersensitivity, nephrotoxicity, and neurotoxicity [106, 107]. Furthermore, the low efficiency
of systemic treatments may result in multidrug resistant cancer and bacteria, leading to a
significantly higher mortality rate.
Many delivery systems have been developed to improve the localization and reduce
systemic toxicity of drugs. The existing systems can be divided into two groups based on their
working mechanisms. The first one relies on systemic delivery of drug–loaded nanomaterials
including polymers, lipids, dendrimers, and metals [108‐113]. These nanoparticles can reach and
be retained in tumors by both passively diffusing through the leaky vasculature system of the
tumors (enhanced permeability and retention [EPR] effect) and actively targeting by conjugated
with cancer cell‐specific molecules such as antibodies, folic acid, transferrin, and lactobionic acid
[114‐118]. Additionally, the nanoparticles can be engineered to respond to environmental stimuli
such as pH, temperature, light, and sound waves to provide controlled release of the drug [119,
Page 49
Page 49
120]. The first FDA approved nanomedicine Doxil® is doxorubicin (DOX) encapsulated in
PEGylated liposome [121] that was shown to significantly reduce cardio cytotoxicity while
maintaining the chemotherapeutic effect compared with conventional DOX [122]. Similarly,
nanoparticle‐based PTX treatments have demonstrated lower systemic toxicity and better
efficacy in combating breast, lung, and ovarian cancer compared with conventional formulation
[106, 123‐126]. Despite the promising treatment efficacy with lower systemic toxicity, localization
of drug is still challenging due to the removal and sequestration of nanoparticles by the body’s
reticuloendothelial system. Recent studies have suggested that the EPR effect works well in
rodents but not in human [127]. A retrospective literature analysis by Wilhelm et al.
demonstrated that only 0.7% (median) of the administered nanoparticle dose is found to be
delivered to a solid tumor [128].
The second group consists of drug depots that can be implanted intratumorally or in the
areas adjacent to cancerous tissues. This strategy may provide better drug localization compared
with nanoparticle‐based delivery systems. So far, a wide variety of drug–loaded structures
including gels, wafers, thin films, particles, and rods have been developed [129‐133]. These drug
depots, which are mostly made from biodegradable natural or synthetic polymers, can provide
sustained release of the drug into the tumor locally. Although natural polymers such as alginate,
dextran, and chitosan are abundant in nature, they are usually poorly soluble in organic solvents,
making it more difficult to incorporate water‐insoluble drugs such as PTX [129, 134, 135].
Furthermore, they suffer from batch to batch variation and their natural structures limit the
ability to customize their compositions to influence the degradation rate and drug release kinetics
[135, 136]. These weaknesses can be resolved by using synthetic polymers such as polyesters
based on lactide, glycolide, and caprolactone. Polycaprolactone, an FDA approved synthetic
polymer, is a promising candidate for the fabrication of drug depot.
In the context of tissue engineering (TE) where biodegradable scaffolds are implanted to
support formation of new tissue, it has been envisioned that these scaffolds could also act as a
Page 50
Page 50
local drug reservoir [23, 42‐44]. Yet a critical requirement for TE scaffolds, particularly for
regeneration of large volume tissue, is an interconnected network of macropores to allow
vascularization [45, 46]. 3D printing has emerged as a powerful technique to achieve this
requirement and produce scaffolds of deterministically complex architectures. It has then been
recently proposed that 3D printed scaffolds could also serve as drug depots [23, 56]. Yet the
macroporosity requirement, hence scaffold’s low surface areas, presents a challenge for
developing drug‐delivering TE scaffolds. As a result, most studies in this area demonstrated only
simple scaffold geometries [44, 57] or lacking macroporosity [44, 137]. Importantly, most of the
existing methods in this area relied on mixing of drugs and biomaterials before manufacturing
the mixture into scaffolds [23, 57, 58]. A major limitation of the direct mixing of biomaterials and
drugs is that it could affect drug’s bioactivities due to elevated temperature or harsh solvents. For
example, biological activity loss of more than 10% for vancomycin and 20% for heparin was
reported when they were mixed with calcium phosphate for ink‐jet printing into 3D structures
[57]. In addition, being embedded inside the solid material, the drug will also only release when
the material degrades which might take a long time, therefore will need much higher drug loading
to have therapeutic effects [59] and might create undesirable consequences such as persistent
inflammation or treatment resistance. Research groups have also tried to address this issue by
adding to the drug‐biomaterial mixture some ingredients (such as hydroxyapatite [59], cellulose
[60] or chitosan [57]) to facilitate fluid infiltration and thus drug elution. This approach is
cumbersome and requires further improvement.
Our work was based on 3D printed scaffolds with large interconnected pores
(approximately 0.7 mm) intended for regeneration of large volume tissue. We hypothesized that
we could employ these scaffolds themselves as a drug delivery vehicle to have additional benefit
of local delivery to prevent disease recurrence without systemic toxicity. We also hypothesized
that a drug immobilization process such as soak loading which is independent of scaffold
preparation would allow for versatile drug immobilization and retain drug full bioactivity strength.
Page 51
Page 51
Based on these hypotheses and inspired by the “sponge effect” where the small micrometer scale
pores facilitate capillary force to effectively absorb and retain liquid, we devised a new type of 3D
printed scaffolds having micropores that were loaded with drug through soak loading. We
envisioned that our approach would eventually allow for intraoperative loading with patient‐
specific doses by surgeons.
In this study, we demonstrated the proof of this concept using polymeric scaffolds
manufactured by screw melt extrusion (SME) technique. Medical grade PCL was used because of
its biocompatibility and degradation, from which many FDA‐approved devices have been
prepared and our extensive experience with PCL and manufacturing of PCL scaffolds [58‐61]. We
first mixed PCL with micro‐sized sodium phosphate‐based porogen particles, then manufactured
3D scaffolds by extruding the molten mixture through a fine nozzle to obtain a scaffold with
designed macro‐porosity. The scaffolds were then immersed in water to dissolute and leach out
porogen particles to realize the micro‐porosity inside the struts and bars of the scaffolds. The
physicochemical properties of the scaffolds were then extensively characterized and their
capability to load and release of various clinical drugs used for cancer treatment and infection
control was demonstrated. We used PCL and SME in this study yet this dual micro‐macro‐porous
scaffold concept can be extended to other 3D printing materials and techniques.
3.3 MATERIALS AND METHODS
3.3.1 Scaffold fabrication
The scaffolds were fabricated by combining SME and porogen leaching methods. Medical
grade PCL (MPCL; PURASORB® PC12) was dissolved in chloroform and mixed with porogen
(grounded disodium hydrogen phosphate and sodium chloride salts, sieved to less than 38 μm).
The mixture was cast on glass surfaces to form thin films (~ 150 µm thick) and dried in a fume
hood for solvent evaporation. The dried films were used as input material to an SME machine
Dual Bioextruder where the material was heated up to 100 ‐ 110oC and extruded through a nozzle
Page 52
Page 52
to build 3D scaffolds [62] using a design that has 0 – 90o lay‐down pattern, strut diameter of 0.3
mm and strut‐to‐strut spacing of was 0.6 mm. Scaffolds without microporosity (nMPCL) were
prepared from MPCL (without porogen) and used as a control. The scaffolds were then immersed
in NaOH 0.01 M solution in a shaker at 37oC for 14 days to leach out the porogen. Control scaffolds
were immersed in the same way to keep the treatment consistent. We found that the amount of
porogen needs to be high enough for its complete removal through leaching yet a too high
percentage of porogen would make extrusion impossible. The weight ratio of 5:4 (MPCL:porogen)
was found to be optimal through our pilot testing.
3.3.2 Scaffold characterization
The surface and strut microporosity of scaffolds were imaged using scanning electron
microscopy (SEM) and atomic force microscopy (AFM). Briefly, the leached scaffolds were
fractured in liquid nitrogen, and then the surface and cross‐section of struts were gold coated at
30 mA for 60‐75 seconds using a Leica EM SCD005 gold coater (Leica Microsystems GmbH,
Wetzlar, Germany) and imaged using a Zeiss FESEM (Carl Zeiss AG, Germany) at accelerating
voltage of 2 kV. The surface roughness of samples was quantified by scanning at different areas
of samples’ struts (20x20 μm2) using an NT‐MDT Solver SPM apparatus (NT‐MDT Spectrum
Instruments, Russia) equipped with uncoated Ted Pella Tap300‐G cantilevers (Ted Pella Inc., US)
[138]. To characterize microporous structure and micropore distribution, scaffolds were also
scanned using a μCT50 micro‐CT scanner (SCANCO Medical AG, Brüttisellen, Switzerland) at a
voltage of 45 kVp, a current of 133 μA and a voxel size of 5 μm. By measuring overall dimension
(V) and weight (W) of scaffolds, intra‐strut porosity was calculated using the following equation
and MPCL density (d) of 1.146 g/cm3 [139]):
𝑃𝑜𝑟𝑜𝑠𝑖𝑡𝑦 % 𝑉 𝑊/𝑑
𝑉∗ 100
Mechanical properties of scaffolds were evaluated using compression test. Samples (6
mm diameter and 1.6 mm height) were prepared and compressed at a rate of 0.01 mm/s until
Page 53
Page 53
reaching 20% compression using an Instron MicroTester 5848 instrument (Instron Ltd., UK). The
Young’s modulus was calculated from the slope of the stress‐strain curve.
To characterize the surface chemical properties of MPCL scaffolds, the density of
negatively charged groups on MPCL scaffold surfaces was quantified by Toluidine Blue O (TBO)
assay as described elsewhere [65‐68]. MPCL scaffolds were cut into pieces of 5x5x1.5 mm3 and
added into separate Eppendorf tubes. A volume of 2 mL TBO, 2 mM solution in NaCl 0.015 M at
pH 11, was added to each tube. The tubes were degassed for 5 minutes and shacked at RT for 2
hours. Scaffolds were washed twice with NaCl 0.015 M pH 11 and immersed in NaCl 0.015 M pH
11 for 4 hours under shaking conditions at RT. The scaffolds were then washed with NaCl 0.015
M pH 11 before placed in new Eppendorf tubes. A volume of 2mL of 70% acetic acid (v/v) solution
was added to each tube. The tubes were degassed for 5 minutes and left at RT for 1 hour under
shaking condition. The solutions were collected and read at 556 nm using a Bio‐Rad xMark™
microplate absorbance spectrophotometer (Bio‐Rad, California, US). The amounts of TBO on
scaffolds were quantified using the standard curve of TBO in 70% acetic acid solution.
3.3.3 Drug loading, characterization and in vitro release
Scaffolds were sterilized by immersing in 70% ethanol (v/v) with subsequent evaporation
in a laminar flow hood followed by ultraviolet (UV) light exposure. Drug solutions of DOX, PTX and
CEF were prepared in 100% methanol, 100% ethanol and 90% ethanol (v/v), respectively. All
drugs were loaded on scaffolds in 3 different doses: low dose (LD) = 0.4 µg/mg (drug/scaffold),
medium dose, MD = 2 µg/mg and high dose, HD = 10 µg/mg. Scaffolds were immersed in drug
solutions in 2‐mL microcentrifuge tubes in a sterile cabinet until solutions were completely
evaporated. Some CEF–loaded scaffolds were further coated with poly(lactic‐co‐glycolic acid)
(PLGA) for sustained release of CEF, due to the high water solubility of CEF. PLGA (85:15, Mw of
c.a. 66 kDa from Lakeshore Biomaterials, USA) was dissolved in acetonitrile to 5% w/v
Page 54
Page 54
concentration. Selected scaffolds were dip coated in the PLGA solution and dried in a sterile
cabinet.
Scaffolds were cryo‐fractured and analyzed using confocal fluorescence microscopy
(Leica, Japan), SEM combined with EDX, Fourier transform infrared spectroscopy (FT‐IR) and X‐
ray photoelectron spectroscopy (XPS) to confirm the presence of drugs on scaffolds. FT‐IR spectra
of MPCL scaffolds were collected by measuring at the setting of scan number of 64 and step of
1cm‐1 over a range from 400 to 4000 cm‐1 wave number by Nicolet FTIR spectrophotometer
(Nicolet Analytical Instruments, US). XPS data were collected in a Thermo‐Fisher K‐Alpha
spectrometer (10‐9 mbar) featuring a monochromated Al Kα X‐ray radiation source with a photon
energy of 1486.7 eV at a power of 300 W using a 400 μm spot size and survey spectra measured
at a 200 eV pass energy. High resolution spectra were measured at 50 eV pass energy. To restrict
sample charging, the low energy electron flood gun was used. Survey spectra were attained at a
pass energy of 200 eV, 10 ms dwell time and 0.1 eV step size. High resolution scans were attained
at a pass energy of 50 eV, 50 ms dwell time and 0.1 eV step size. Three samples were used for
both surveys and high resolution spectra with three spots average on each to determine the
atomic concentration of each element.
To characterize the effect of pH on the release profile of DOX–loaded scaffolds, the
release kinetics were evaluated in PBS at pH 7.4 and pH 5.5 in the shaking incubator at 37oC and
200 rpm. For PTX–loaded scaffolds, samples were placed in PBS (pH7.4) 0.1% w/v Tween 20 for
the release because of insolubility of PTX in water. The release profile of CEF–loaded scaffolds
was determined in PBS (pH 7.4).
At each time point, the supernatant solutions were collected and replaced with fresh
solutions. The absorbance of the collected solutions were respectively measured at 480 nm, 230
nm and 272 nm for DOX, PTX and CEF using an xMark microplate absorbance spectrophotometer
(Bio‐Rad, US) and the amounts of drug released in the solution were calculated using standard
curves.
Page 55
Page 55
3.3.4 Antimicrobial assay
The antimicrobial activity of CEF–loaded scaffolds was evaluated by agar diffusion assay
against S. aureus (ATCC 29213). Log phase S. aureus was plated on Mueller‐Hinton agar plates.
Scaffolds were placed and gently pressed to have stable contact and allow drug diffusing into the
agar. The plates were incubated at 37oC and inhibition zones were measured after 16 hours.
3.3.5 In vitro chemotherapeutic effect
MDA‐MB‐231 cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco,
US) supplemented with 10% fetal bovine serum (FBS, Gibco, US), 100 U/ml penicillin and 100
µg/mL streptomycin (Gibco, US). When the cells were approximately 80% confluent, cells were
collected and seeded in 24‐well plates with 1.5x104 cells/well cell density. After overnight
incubation, DOX–loaded scaffolds were placed individually into each well and incubated for 2
days. Then, picogreen assay was performed. In brief, scaffolds were collected, culture media were
discarded and 300 μL proteinase K (Invitrogen, US) was added into each well and the plates were
incubated at 37oC overnight. After that, the cell‐proteinase K suspensions were collected and
incubated at 56oC for 8 hours. The suspensions were diluted for 50 and 100 times in phosphate
buffered EDTA (PBE) buffer (2.84 g Na2HPO4, 4.14 g NaH2PO4, and 1.86 g Na2EDTA dissolved in
ultrapure water and adjusted to pH 7.1 using 6N NaOH) and transferred into black 96‐well plates.
Picogreen 1X (Invitrogen, US) was added into each well at the volume ratio of 1:1 and incubated
at RT under dark condition for 10 minutes. The plates were read at the excitation wavelength of
480 nm and the emission wavelength of 520 nm using a POLARstar Optima plate‐reader (BMG
Labtech, Germany).
For PTX–loaded scaffolds, MDA‐MB‐231 cells were encapsulated in gelatin methacryloyl
(GelMA) hydrogels as previously described [69] and cultured for 2 weeks. Then, the PTX–loaded
scaffolds with LD, MD and HD were placed into each well and cell viability was assessed after 2, 5
and 8 days using Alamar Blue assay (Invitrogen, US). Scaffolds without drug were used as control.
Page 56
Page 56
3.3.6 Statistical analysis
Data are expressed as mean ± standard error of means (s.e.m) and one‐way ANOVA test
and Tukey post hoc test were used to assess differences.
3.4 RESULTS
3.4.1 Material preparation, extrusion and scaffold characterization
In this study, porogen leaching method was used to create the micropores inside the
struts of 3D printed scaffolds to obtain drug delivery functionality. SEM analysis confirmed the
presence of micropores network within the struts, which were prepared with porogen mixed into
the MPCL and subsequently leached out (Figure 9E‐H).
Figure 9. Scanning electron microscopy (SEM) images showing surface pores on both
non‐porous (nMPCL) and microporous (pMPCL) scaffolds (A, B, E, F) yet only pMPCL scaffold
showed porous struts (G, H) compared with solid struts in nMPCL (C, D). The majority of intra‐
strut micropores in pMPCL show interconnection with surrounding pores (arrows in G, H) and the
surface pores on nMPCL are only artifacts of the extrusion (arrow in D).
Page 57
Page 57
Atomic force microscopy analysis revealed similarity of surface roughness of nMPCL and
pMPCL (Appendices, Figure S4B). The difference in surface morphology of nMPCL and pMPCL was
in macroscopic scale, hence did not strongly affect the surface roughness. The root‐mean‐square
(RMS) surface roughness of pMPCL (155±9 nm) was slightly decreased but not significantly
different from that of nMPCL (163±10 nm) (Appendices, Figure S4B). The µCT micrographs
displayed the uniform distribution of micropores throughout the pMPCL struts (Figure 9H). There
were also some pores on the surface of the struts in nMPCL scaffolds. These surface pores in non‐
porous scaffolds are identified as gaps between the surface PCL spherulites and thus are shallow
pores (Figure 9D). In contrast, majority of surface pores on the porous scaffolds are formed from
leaching of surface porogen and these surface pores created pathways for porogen embedded
deeper to leach out (Appendices, Figure S4‐C, D). Surface pores in pMPCL scaffolds were
significantly smaller (mostly ≤ 5 µm2 in cross‐section area) compared with intra‐strut pores
(mostly ≥ 100 µm2 in cross‐section area) (Appendices, Figure S4‐E).
Intra‐strut volume porosity of pMPCL scaffolds was determined from measuring strut’s
mass, dimension and calculated using the polymer’s density (1.146 g/cm3) [64] to be 31.7±1.5%
(mean ± standard error of means (s.e.m), n=4). Struts from non‐microporous MPCL scaffolds have
a measured density of 1.21± 0.02 g/cm3 which is close to the density of 1.146 g/cm3 of PCL
material [58] therefore have zero intra‐strut porosity as expected and evident through SEM and
micro‐CT images (Figure 9A‐D).
Interestingly, the surface of pMPCL struts represented lower porosity than in the cross‐
section. This difference can be explained by the extrusion of the molten salt porogen‐polymer
mixture through a nozzle in which the solid porogen particles would be pushed away from the
nozzle wall to reduce friction. As a result, fewer pores were present on the strut surface. SEM and
µCT analysis showed local interconnection of the micropores (Figure 9 G, H). Because these pores
were formed through the leaching of porogen, they must be connected to some surface pores.
This connection would allow drug loading by simply soaking in a drug solution.
Page 58
Page 58
3.4.2 Micropores partially reduced compression Young’s modulus of scaffolds
To characterize the effects of micropores on mechanical property of MPCL scaffolds,
compression test was performed. The compression Young’s moduli at strains of 0‐5%, 5‐10%, 10‐
15% and 15‐20% were 3.8±0.8 MPa, 10.2±0.9 MPa, 14.3±0.8 MPa and 13.0±0.9 MPa, respectively
for nMPCL and 3.5±1.5 MPa, 12.4±0.2 MPa, 10.4±0.7 MPa and 7.2±0.4 MPa, respectively for
pMPCL (Figure 10). There was a significant difference between nMPCL and pMPCL from 10% to
20% compression (p<0.01) which can be explained by the microporous strut pore network
contributing to a reduction in the material density and hence reduce the compressive modulus.
Page 59
Page 59
Figure 10. Intra‐strut micropores reduced compression Young’s modulus of
polycaprolactone (MPCL) scaffolds. (A): Effective compression Young’s modulus (data are
expressed as mean ± s.e.m, n=9) showed no significant difference in the 0‐10% compression
range. At 10‐15% and 15‐20% range, the micropores significantly reduced Young’s modulus of
pMPCL. (B): SEM images of non‐microporous (nMPCL) and microporous (pMPCL) after the
compression test showing the collapse of some micropores during the compression. ** is p<0.01
and from one‐way ANOVA test and Tukey post hoc test.
3.4.3 In vitro drug release
The micropores also increased the negatively charged group density in pMPCL (7.8±0.2
nmol/mg compared with 3.6±0.2 nmol/mg, n=6), which was measured by TBO assay. We
hypothesized that the release kinetics would then depend on the charge properties of the loaded
drugs. Therefore, we investigated the loading efficiency and release kinetics of three different
drugs with different solubility and charge characteristics at physiological pH: DOX was chosen as
a water‐soluble, positively charged drug; PTX as a nonsoluble and uncharged drug; and CEF as a
soluble and negatively charged drug.
3.4.4 Loading and release of doxorubicin as a water‐soluble and positively charged
drug
Doxorubicin (DOX) is a chemotherapeutic agent that is relatively water‐soluble (10
mg/mL at 25oC) and is positively charged at physiological pH (pH 7.4), which is attributed to its
aliphatic amine group with pKa around 9.93 [140]. DOX has been widely used to treat a number
of cancers and is known to have severe cytotoxicity [141, 142]. Therefore, application of lower
amount of drug using local delivery of DOX would be very beneficial for reducing the side effects.
Here, three concentrations of DOX were loaded into the scaffolds by soaking – 0.4 µg
drug/mg scaffold (low dose), 2 µg/mg (medium dose) and 10 µg/mg (high dose). Bright‐field
images showed relatively even distribution of DOX on the surface of scaffolds (Figure 11A).
Page 60
Page 60
pMPCL had higher drug loading efficiency (approximately 95%) compared with nMPCL
(approximately 80%) determined by measuring the amount of free drug remained after soak
loading (Appendices, Figure S6). For nMPCL, DOX only distributed on struts’ surfaces (Figure 11B).
In contrast, for pMPCL, DOX distributed not only on strut surface but also on intra‐strut micropore
surface which was confirmed by the fluorescence of DOX in the confocal images (Figure 11B).
Furthermore, the presence of DOX on MPCL was confirmed by FT‐IR (Figure 11C). The spectra of
DOX–loaded MPCL scaffolds showed the signature peaks of DOX (1612 cm‐1 and 1577 – 1579 cm‐
1 for the aromatic ring structure) in addition to peaks from MPCL (1163 cm‐1 for symmetric COC
stretching, 1239 cm‐1 for asymmetric COC stretching and 1721 cm‐1 for carbonyl stretching).
Page 61
Page 61
Figure 11. DOX immobilized only on surface of nMPCL but both on surface and in intra‐
strut pores of pMPCL. (A): bright field images of nMPCL and pMPCL before and after DOX–loading
at 10 µg/mg. (B): Cross‐sections of nMPCL and pMPCL showing auto‐fluorescence of DOX on the
surface of nMPCL and inside the struts of pMPCL. (C): FT‐IR spectra of DOX–loaded onto MPCL at
increasing 3 loading doses (Low, Medium and High Doses).
DOX was found to immobilize even on nMPCL surface and this immobilization is
attributed to hydrogen bonding or static attraction between the positively charged amine groups
Page 62
Page 62
on DOX and the –COOH groups on PCL (Appendices, Figure S5). The carboxyl groups were likely
created from ester cleavage of PCL in the scaffold surface by NaOH during the fabrication process.
To characterize the release of these systems where strong electrostatic interaction of
DOX and MPCL is expected, the experiments were carried out in PBS at pH 7.4 and pH 5.5 where
–COO‐ groups are expected to become protonated [143]. The release of DOX in PBS pH 7.4 at
different doses showed similar trends (Figure 12). pMPCL significantly reduced the burst release
of DOX from 0.09±0.02 µg/mg in LD nMPCL to 0.03±0.01 µg/mg in LD pMPCL and from 0.31±0.05
µg/mg in MD nMPCL to 0.14±0.02 µg/mg at 1 hour time point. The pMPCL had a more sustained
release profile for up to 120 hours compared with 48 hours for nMPCL. LD and MD of pMPCL
were effectively reduced the burst release from 22.5% (nMPCL) of total release amount to 5%
(pMPCL) and from 15% (nMPCL) to 7.5% (pMPCL). For HD, the burst releases of nMPCL and
pMPCL were similar (20%) (Figure 12).
Figure 12. Intra‐strut micropores reduced burst release and prolonged release profile of
drug–loaded microporous polycaprolactone (pMPCL) scaffolds. Drug release from scaffolds–
loaded with increasing doses of DOX (Low, Medium and High doses) in PBS at pH 7.4 and pH 5.5
(data are expressed as mean ± s.e.m, n=6). pMPCL had less burst release and more prolonged
Page 63
Page 63
elution compared with nMPCL. The release from DOX–loaded scaffolds increased under acidic
conditions (pH 5.5).
Comparing the release of DOX–loaded scaffolds under different pHs, the total amount of
DOX released in pH 5.5 was higher than in pH 7.4. Additionally, pMPCL groups with LD and MD of
DOX still maintained the low level of DOX release; whereas the remaining groups had their burst
release amounts increased. Furthermore, the release of DOX from pMPCL under acidic condition
was sustained for 240 hours. The increment of the total release amount of DOX under acidic
condition confirms the expected electrostatic interaction between amine group on DOX and
carboxyl groups on PCL resulted from ester cleavage during the scaffold preparation (i.e., leaching
in NaOH 0.01M) (Appendices, Figure S5).
3.4.5 Loading and release of paclitaxel as model insoluble non‐charged drug
PTX is a chemotherapeutic agent frequently used to treat ovarian, lung, and breast
cancers [144, 145]. PTX has zero net charge and poor water solubility at physiological pHs [146].
Therefore, Cremophor EL – a castor oil based formulation vehicle for poorly water‐soluble drug,
has been used to dissolve PTX for systemic injection [147]. However, application of Cremophor
EL results in further cytotoxicity to patient [106]. By delivering PTX locally using MPCL scaffolds as
a drug reservoir, we aim to minimize/avoid its systemic cytotoxicity.
The loading of PTX on MPCL scaffolds was confirmed by SEM, EDX, FT‐IR and XPS (Figure
13). The immobilization of PTX on MPCL was indicated by changes in atomic ratio of oxygen and
carbon in EDX analysis (Figure 13A). MPCL had atomic% Carbon/Oxygen ratio of 3.1±0.2;
whereas, that of PTX was higher (6.1±0.7). XPS also confirmed the immobilization of PTX in the
cross‐section of pMPCL by the presence of Nitrogen peak in XPS spectrum (Figure 13B). FT‐IR
spectra of PTX–loaded pMPCL scaffolds showed both characteristic peaks of MPCL (1721 cm‐1,
Page 64
Page 64
which corresponding to carbonyl groups of MPCL) and PTX (1635 cm‐1, which corresponding to
C=O‐N groups of PTX) (Figure 13C).
Figure 13. Characterization of PTX–loaded pMPCL scaffolds. (A): Representative SEM
micrographs of PTX–loaded pMPCL showing thin fiber ‐like morphology of immobilized PTX on
the surface and inside micropores (white arrows) and their EDX analysis showing atomic C%
(white numbers) higher than those of MPCL (85±1% for PTX and 75±1% for MPCL, data are
expressed as mean ± s.e.m, n=6). (B): XPS analysis of PTX–loaded scaffolds showing N1s peaks
from PTX even after 20 seconds of in situ ion beam etching for partial removal. (C): FT‐IR spectra
of MPCL, PTX–loaded pMPCL scaffold (MPCL‐PTX) and PTX confirming the immobilization of PTX
on MPCL scaffolds.
The elution of PTX from loaded scaffolds was then investigated. Overall, the release of
PTX was monotonic and changed from near zero‐order kinetic in low dose to first‐order kinetic in
Page 65
Page 65
high dose (Figure 14). Comparing release profiles of nMPCL and pMPCL, there was no difference
in term of burst release in all doses, yet the PTX release was significantly higher in pMPCL after
72 hours at medium and high doses and 400 hours for low dose. For nMPCL, the majority of the
drug was released in the first 165 hours as demonstrated through the initial big slope of the
release profile. After this period, the release was significantly reduced to nearly negligible as
evident from the diminishing slopes. On the other hand, pMPCL showed the trend of increasing
release up to 500 hours.
Figure 14. Microporous structure prolonged and increased release of immobilized PTX
on pMPCL. Drug was immobilized from solutions at 3 increasing doses (low, medium and high
doses) by soak loading. Release at the first time point (24 hours) from pMPCL scaffolds was similar
with nMPCL of the same dose but became significantly higher at later time points (data are
expressed as mean ± s.e.m, n=6).
As PTX is uncharged, it was expected that there are no strong interactions between PTX
and MPCL and the release of PTX was purely dissolution in the case of nMPCL. In the case of
pMPCL, the drug had to diffuse out of the microporous network; therefore, its release to the
outside was prolonged. The burst release of PTX from MPCL scaffolds into PBS solution was low
Page 66
Page 66
and similar for both nMPCL and pMPCL for the same reasons. Nonetheless, since pMPCL had
more surface area, the total final release was confirmed to be higher.
3.4.6 Loading and release of Cefazolin as model water‐soluble, negatively charged
drug –PLGA overcoat can lead to sustained release
Cefazolin (CEF) is a common antibiotic often used to prevent post‐surgery infection [148‐
151]. CEF is highly water‐soluble and negatively charged at physiological pH [152]. CEF is
traditionally administrated systematically yet it is well known that systemic administration of CEF
is not effective against bacteria biofilm formed locally on the surface of implant. There is also a
concern about increased risk of antibiotic resistance development after systemic administration.
Hence, we investigated the immobilization of CEF directly onto scaffolds to eventually prevent
bacterial infection at the surgical site.
Successful immobilization with c.a. 80% loading efficacy (Appendices, Figure S6) was
confirmed by EDX. In nMPCL, the peaks of Na and S corresponding to Na and S in the CEF were
only found on the surface (Figures 15A, B). In pMPCL, the peaks of Na and S were found in both
the surface and cross‐section indicating that CEF was loaded not only on surface but also into the
micropores of pMPCL scaffolds.
Page 67
Page 67
Figure 15. Immobilization of CEF on nMPCL and pMPCL scaffolds and the application of
a PLGA overcoat layer. (A) and (B): SEM and EDX mapping of CEF–loaded scaffolds showing drug
immobilized only on surface of nMPCL and also within pores inside the struts in pMPCL scaffolds.
(C): SEM micrographs of drug–loaded pMPCL scaffolds with PLGA overcoat layers and EDX
analysis showing differentiation of PLGA from the MPCL underneath which has less O content.
The elution of CEF from non‐porous nMPCL was purely burst release as expected from
the high water solubility and similar charge characteristics of CEF [83] (Figure 16A). Release of CEF
from microporous pMPCL was found significantly prolonged to c.a. 170 hours. In the first 24
hours, the burst release of both nMPCL and pMPCL were at high level (70% and almost 100% for
pMPCL and nMPCL, respectively). After the initial burst release, CEF was released gradually from
pMPCL until reaching 100% after 170 hours. In MD groups, similar trends were observed for both
nMPCL and pMPCL (75% and almost 100% of loaded drug amount released in the first 24 hours
for pMPCL and nMPCL, respectively).
Page 68
Page 68
Figure 16. Release of CEF from nMPCL and pMPCL scaffolds and the barrier effects of
PLGA overcoat layers. (A): release profiles of scaffolds loaded with the drug at increasing doses
(Low, Medium and High doses). (B): Release profiles of pMPCL scaffolds loaded with the drug at
high dose with and without PLGA coating and model fitting of the release profiles. PLGA coating
significantly slowed down the release of the CEF previously immobilized on the scaffolds. The
pMPCL scaffolds showed first‐order like release and the application of the PLGA overcoat
sustained the release further and achieved near zero‐order kinetic.
To further regulate CEF release from pMPCL scaffolds, a biodegradable over‐coat layer
of PLGA was applied to provide barrier effects to the CEF immobilized inside the micropores. The
presence of PLGA coating on pMPCL was confirmed by its distinctive morphology and its higher
oxygen content compared with PCL (Figure 16B). By applying the PLGA coating, the burst release
of pMPCL was significantly reduced from 70% to 30% (Figure 16A – High Dose). After burst
release, PLGA‐coated scaffolds were showed to have zero‐order like release kinetic (Figure 16B).
CEF has high water solubility and has a net negative charge [83]; hence its adsorption on
nMPCL, which also have surface negatively charged groups, was likely very weak and hence it
Page 69
Page 69
rapidly released into PBS by dissolution. In pMPCL, the unique architecture with low surface
porosity and high intra‐strut porosity entrapped CEF inside the struts and slowed down its release
to the outside of the scaffolds. In the time frame of the release studies, the drug release from
pMPCL can be considered as from non‐degradable matrices and was found to begin with some
burst release followed by first‐order like release kinetics. The release of CEF form pMPCL without
coating was fitted in first‐order release kinetic with R2 of 0.98 (Figure 16B). When an overcoat
layer of PLGA is applied to the microporous scaffolds, the system can be considered as a reservoir
with an inert coating functioning as a rate‐controlling barrier. The diffusions in these cases are
therefore more likely by Fickian driven mechanism, which meant the releasing rate remained
constant for a sustained period. The release of CEF from PLGA coated pMPCL scaffolds was fitted
in zero‐order release kinetic with R2 of 0.94.
3.4.7 In vitro bioactivity of drug–loaded scaffolds
The bioactivities of drug–loaded pMPCL scaffolds were performed to confirm that the
drugs retained their activities.
3.4.7.1 Antimicrobial effect of Cefazolin–loaded scaffolds
Antimicrobial activities of CEF–loaded pMPCL scaffolds with LD, MD and HD were
characterized by agar diffusion assay. pMPCL without drug was used as control. Samples were
placed on agar plates containing S. Aureus and incubated for 16 hours and the zones of inhibition
were measured. Overall, when the doses increased from 0.4 µg/mg (drug/MPCL, LD) to 2 µg/mg
(MD) and then 10 µg/mg (HD), the zone of inhibition significantly increased accordingly (Figure
17).
Page 70
Page 70
Figure 17. Cefazolin immobilized on microporous scaffolds retained its bioactivity. Zone
of inhibition of CEF–loaded pMPCL scaffolds against S. aureus showing increasing zone diameter
with increasing loading dose. pMPCL scaffolds without drug were used as control. Data are
expressed as mean ± s.e.m (n=6). * is p<0.05 and ** is p<0.01 from one‐way ANOVA and Tukey
post hoc test.
The dose‐dependent properties of CEF–loaded pMPCL over a broad working range from
0.4 µg/mg to 10 µg/mg could be beneficial. Since the volume of the scaffolds used for each
patient is different and the amount of antibiotics varies from patient to patient, therefore the
large working range of concentrations give doctors the option to adjust the doses.
3.4.7.2 Effect of pMPCL scaffolds loaded with chemotherapeutic drugs against
cancer cells
Cell culture on bottom of well plates are still standard condition for testing drug
bioactivity and 3D cell culture in hydrogels such as GELMA has recently been recognized as a more
sophisticated tool to better mimic the cell‐ECM interactions in vivo. We have developed extensive
expertise in using GelMA for 3D culture in our laboratories [73] and thus used it for testing of PTX.
MDA‐MB‐231 were encapsulated inside GelMA hydrogels [153] and treated with drug–loaded
pMPCL scaffolds. All PTX–loaded MPCL scaffolds reduced cell viability after 2 days of treatment
Page 71
Page 71
yet no significant difference was found among the loading doses except at day 5 (Figure 18A).
Medium and High dose groups showed a significantly stronger chemotherapeutic effect than Low
dose group. After 8 days, all drug–loaded scaffolds completely killed all the cancer cells (Figure
18A).
Figure 18. DOX and PTX immobilized on microporous scaffolds retained their bioactivity.
(A): Metabolic activity of breast cancer (MDA‐MB‐231) cells in 3D culture treated with PTX–
loaded scaffolds. Cells were encapsulated in methacrylated gelatin (GelMA) hydrogels and
cultured in 24 well plate and PTX–loaded pMPCL scaffolds were then placed in the wells as
treatment groups and Alamar Blue assay was performed on day 2, 5 and 8 of the treatment (data
are expressed as mean ± s.e.m, n=8). (B): Bioactivity of DOX–loaded scaffolds. Cells were seeded
in 24‐well plate, scaffolds were placed in the wells as treatment groups and intracellular DNA
amounts were quantified after 2 days of treatment using PicoGreen assay (data are expressed as
mean ± s.e.m, n=8). The results showed the effect of chemotherapeutic agent–loaded pMPCL
scaffolds. * is p<0.05 and ** is p<0.01 from one‐way ANOVA and Tukey post hoc test.
We also tested GelMA 3D culture for in vitro validation of DOX–loaded scaffolds, yet the
specific binding of DOX (positively charged) to GelMA (negatively charged) [140, 154] was found
interpreting the DOX’s concentrations on cell death. Therefore, we used the traditional cell
culture (on bottom of well plate) for testing of DOX–loaded scaffolds. DOX–loaded pMPCL
Page 72
Page 72
showed dose‐dependent chemotherapeutic effect against metastatic breast cancer cells, MDA‐
MB‐231 (Figure 18B). The DNA amounts of MDA‐MB‐231 cells treated with DOX–loaded scaffolds
showed significant decrease in DNA levels from 269.5±26.2 ng/well for untreated control group
to 116.5±14.5 ng/well for Low dose group and 44.5±10.1 ng/well for medium dose group and
57.0±9.2 ng/well for high dose group. There was no significant difference between medium and
high dose groups.
The results from all drug–loaded pMPCL scaffolds confirmed the bioactivity of the drugs
was maintained after being immobilized on the microporous scaffolds.
3.5 DISCUSSION
Our work was built upon the current scaffold‐based tissue engineering approach where
a scaffold with an interconnected network of macropores is implanted to allow tissue formation,
vascularization and tissue remodeling [45, 46]. We derived that the scaffolds themselves can be
developed to have local drug‐delivery capability to prevent recurrence of diseases or other
complications that affect tissue regeneration. Local drug delivery using implantable materials has
been recognized as a promising strategy to reduce systemic toxicity and enhance drug
concentration at the site of implantation. Various materials and material forms have been
investigated for development of local drug delivery vehicles such as metals, polymers, ceramics,
and hydrogels [129]. There are existing reports from other groups demonstrating the efficacy of
local drug‐delivery approach, yet they mostly were based on monolithic scaffolds [44, 57] which
lack interconnected macroporosity or designed complex internal architectures [44, 137].
Our group has extensive experience in polymer scaffolds prepared via additive
manufacturing (i.e., 3D printing) which allows for well defined, interconnected macro‐pores and
customized architectures. We aimed to build upon these scaffolds to achieve drug delivery
functionality. Current techniques to achieve drug delivery from 3D printed scaffolds rely on
mixing of drug with raw materials before printing [23, 57, 137] or absorbing drug to the scaffold’s
Page 73
Page 73
surface that has been treated to have high surface areas (such as etching or coatings) [42, 155,
156]. Inspired by the “sponge effects”, we hypothesized that introducing microporosity into these
macroporous scaffolds would allow the scaffolds to soak up and retain drug from a drug solution.
The novelty of our work is in (i) the generation of microporosity inside an interconnected
macroporous scaffold and (ii) the decoupling of scaffold manufacturing and drug loading.
Interconnected macroporosity in scaffolds is important for tissue regeneration and is currently
best achieved through 3D printing because of the ability in controlling scaffold’s architecture [45,
46]. Therefore, we built our work upon the advantages of 3D printed scaffolds. To incorporate
drugs to these scaffolds, we took a different approach than conventional drug‐mixing with raw
materials or surface‐etching/modification of scaffolds followed by drug absorption. We reasoned
that microporosity would play important roles in entrapping drugs from a solution through
increased surface area and capillary actions. This immobilization is also expected to be non‐
specific and therefore highly versatile and would allow for loading of patient‐specific dose
intraoperatively in the future.
In this proof‐of‐concept study, we created microporosity in our normal 3D printed
polycaprolactone (PCL) scaffolds and investigated their capability in drug absorption and
sustained release. Micro‐sized porogen particles were mixed with PCL and extruded via a
thermal‐extrusion process to build 3D scaffolds which were then immersed in water to leach out
the porogen. We then obtained scaffolds that have macroporosity (from the lay‐down pattern of
the struts and bars) and microporosity (from the voids created by porogen leaching).
The melt‐extrusion and porogen leaching resulted in a very interesting feature in these
scaffolds: the inside of the struts and bars have much larger and more pores compared with their
surface. The surface pores in non‐porous scaffolds are identified as gaps between the surface PCL
grains and thus are shallow pores (Figure 9D). On the contrary, the majority of surface pores on
the porous scaffolds are formed from leaching of surface porogen and these surface pores
created pathways for porogen embedded deeper to leach out. This connection allowed for
Page 74
Page 74
efficient drug entrapment from the solution (high loading efficacy of approximately 90%). The
small surface pores are also believed to be important in the reduced burst release of the
entrapped drugs from the inner pores in the release experiments (Figures 12, 14 and 16). Solid
PCL struts have a measured density of 1.21± 0.02 g/cm3 which is close to the density of 1.146
g/cm3 of PCL material [139] indicating that no diffusion of the entrapped drugs through PCL is
expected because of the high material density and this is in agreement with reports in literature
concerning impermeable PCL barrier [157, 158]. Control scaffolds without microporosity
absorbed drug from solutions onto the outer surface of their struts and into the shallow surface
pores and therefore released them much more quickly.
We demonstrated that this novel microporous scaffold design allowed immobilization of
three clinical drugs of different physical and chemical properties including cefazolin (water‐
soluble, negatively charged antibiotic), doxorubicin (water‐soluble, positively charged
chemodrug) and paclitaxel (insoluble, uncharged chemodrug) with high loading efficacies using
simple soak loading techniques. This result demonstrated the roles of the micropores in loading
and release of the drugs. Finally, we demonstrated the bioactivities and dose response of the
drug–loaded scaffolds in vitro.
This study presented a concept that can be expanded to other 3D printed materials and
techniques. Other polymers, ceramics, polymer‐ceramic composites, metals could be mixed with
appropriate micro‐sized porogen and manufactured into 3D scaffolds using techniques such as
extrusion, laser sintering or photopolymerization. The concept of microporosity‐enabled drug
loading and sustained release is envisioned to be extendable even to the emerging field of 4D
printing where the printed scaffolds also change their structures over time under external stimuli
[159, 160]. The stimuli‐triggered mechanical deformation in 4D printing has been exploited to
encapsulate and release drugs. Current designs focus strongly on using pH, temperature or
osmotic pressure changes to cause swelling/de‐swelling leading to folding and unfolding of
layered structures, therefore, capturing and releasing drug content [160‐162].
Page 75
Page 75
We believe the dynamic nature of the 4D printed scaffolds could present a challenge to
the introduction or maintaining of the microporosity because of the particular choice of scaffold
design or materials or manufacturing techniques. However, the dynamic nature could also
present a great opportunity for another level of control over drug entrapment into the
micropores and its release. That the loaded drug could be released on‐demand is a feature of
great importance in many applications. These smart materials and designs present another
advanced approach in local drug delivery and highlighted its importance in tissue engineering
[159]. Our work aimed to address this important issue using an approach that is envisioned to
complement the above mentioned 4D printing technologies.
In the current study, the porogen’s size and percentage were chosen based on our pilot
testing to achieve printability and complete porogen leaching. It is envisioned that these
parameters would significantly influence scaffold’s characteristics including drug loading and
release. This is a topic of our on‐going investigation. It is also believed that the surface pores play
an important role in drug loading and release from the fluid dynamic point of views. We purposely
prepared drug loading solutions to have high concentration of alcohols (such as ethanol and
methanol) to allow better infiltration into the micropores. Our loading of drugs from solutions
through soak loading is versatile and has potential to allow for intraoperative loading of patient‐
specific doses, yet further optimization is still needed. We believe that better drug solution
penetration into the micropores would be achieved with applications of negative (i.e. vacuum
assisted loading) or positive pressure (pressurized loading). Drug loading and delivery would also
depend on the size and shape of the scaffolds. For example, the drug loading per scaffold and
hence delivery dose would depend on the total surface areas and microporosity and thus is
strongly dictated by scaffold design. This is an on‐going investigation and optimization in our
group. The efficacy of the drug–loaded scaffolds in vivo is also a focus of our current work and
the results will be communicated in our next report.
Page 76
Page 76
3.6 CONCLUSION
This study demonstrated our new concept in creating microporosity inside macroporous
tissue engineering scaffolds to have the additional functionality of local drug delivery aiming at
preventing post‐surgery complications. We used SME scaffolds as a proof of concept. Through 3D
printing, we obtained scaffolds with highly interconnected macropores between the struts. By
introducing porogen microparticles into the polymer before printing and leach them out after
printing we created macroporous scaffolds with intra‐strut microporosity. The printing and
leaching processes resulted in a very particular intra‐strut micropore network where the surface
pores were much smaller in sizes and number than the inside. The micropore network formed
through complete porogen leaching allowed for efficient loading of three clinically relevant drugs
including chemotherapeutics (doxorubicin and paclitaxel) and antibiotic (cefazolin) via soaking in
the drug solutions. The drugs were also found to penetrate deep into the micropore network and
thus their release was significantly reduced. In vitro cancer cell culture and bacteria assays
confirmed the bioactivity of the drug–loaded scaffolds. This new type of macro‐microporous
scaffolds is envisioned to allow for simple intraoperative drug loading of specific doses for
applications in regenerating diseased tissue where the risks of complications such as surgical site
infection or cancer recurrence exist.
Page 77
Page 77
4Chapter 4: Delivery of doxorubicin from the
implanted bimodal porous polycaprolactone
scaffolds decreased systemic cytotoxicity and
breast tumour recurrence progression in mice
Hoang Phuc Dang a, b, Abbas Shafiee a, c, Christoph A. Lahr a, b, Tim R. Dargaville a, b, Phong A. Tran
a, b, d
Manuscript under supervisor revision; Target Journal: Advanced Functional Materials
a Institute of Health and Biomedical Innovation (IHBI), Queensland University of Technology,
Brisbane, Queensland, Australia
b ARC Centre in Additive Biomanufacturing, Queensland University of Technology, Brisbane,
Queensland, Australia
c UQ Diamantina Institute, Translational Research Institute, The University of Queensland,
Brisbane, Queensland, Australia
d Interface Science and Materials Engineering Group, School of Chemistry, Physics and
Mechanical Engineering, Queensland University of Technology, Brisbane, Australia
Page 78
Page 78
The authors listed below have certified that:
11. They meet the criteria for authorship in that they have participated in the conception,
execution, or interpretation, of at least that part of the publication in their field of expertise;
12. They take public responsibility for their part of the publication, except for the responsible
author who accepts overall responsibility for the publication;
13. There are no other authors of the publication according to these criteria;
14. Potential conflicts of interest have been disclosed to (a) granting bodies, (b) the editor or
publisher of journals or other publications, and (c) the head of the responsible academic unit,
and
15. They agree to the use of the publication in the student’s thesis and its publication on the QUT’s
ePrints site consistent with any limitations set by publisher requirements.
In the case of this chapter:
Manuscript under supervisor revision:
Hoang Phuc Dang, Abbas Shafiee, Christoph A. Lahr, Tim R. Dargaville and Phong A. Tran.
“Delivery of doxorubicin from the implanted bimodal porous polycaprolactone scaffolds
decreased systemic cytotoxicity and breast tumour recurrence progression in mice”.
Contributor Statement of contribution
Hoang Phuc Dang
Experimental design, conducted experiments and data analysis,
wrote the first draft of manuscript Signature
Date:
Abbas Shafiee
Performed experimental design, conducted experiment,
supervised the project, performed data analysis and contributed
in manuscript preparation
Page 79
Page 79
Christoph A. Lahr Conducted experiment
Tim R Dargaville Aided manuscript preparation
Phong A. Tran
Performed experimental design, supervised the project,
performed data analysis and contributed in manuscript
preparation
Principal Supervisor Confirmation
I have sighted email or other correspondence from all Co‐authors confirming their
certifying authorship
Phong Tran 06/02/2020
Name Signature Date
QUT Verified Signature
Page 80
Page 80
4.1 ABSTRACT
Breast augmentation combined with systemic chemotherapy is commonly used as a
conventional method to regain body image and prevent cancer recurrence after neoplasm
removal. However, this approach is limited due to the systemic cytotoxicity of chemotherapeutics
and the risk of implant removal due to the cancer recurrence. The use of drug–loaded resorbable
3D printed scaffolds is one strategy to locally deliver chemotherapeutic drugs and support breast
tissue regeneration; however, matching the drug release profile to the degradation of the
scaffolds is challenging. Here, we demonstrate novel 3D extruded scaffolds with bimodal porous
structure—macroscale pores from the printing pattern and microscale pores inside the struts—
that can be used to modulate drug release and degradation profiles. The scaffolds were
manufactured from a composite of a medical grade polycaprolactone (pmPCL) and salt
microparticles using a screw melt extrusion (SME) method with a designed macroscale
architecture. The porogen was subsequently leached to generate microscale pores with an
average size of 5—35 µm. In vitro tests demonstrated that DOX–loaded on the bimodal porous
scaffolds had a linear sustained release profile of up to 28 days and that the DOX–loaded scaffolds
possessed a dose‐dependent chemotherapeutic effect against the breast cancer cell line MDA‐
MB‐231. Using a mouse orthotopic breast cancer model, we demonstrated that the DOX–loaded
scaffolds, despite using only 5 and 20% of the amount of the intravenous injection (i.v.) (or 2 and
8 µg/scaffold/animal), significantly improved the treatment efficacy against breast cancer local
recurrence and metastasis progression in lung, liver and spleen. Moreover, the use of drug–
loaded scaffolds reduced the cardio—cytotoxicity of DOX compared with i.v. administration.
Overall, this study showed the proof‐of‐concept that tissue engineered scaffolds can be
manufactured to have drug delivery functionality using microscale porosity for potential
applications such as breast tissue reconstruction after neoplasm removal.
Page 81
Page 81
4.2 INTRODUCTION
Breast cancer is the most commonly diagnosed cancer in women worldwide [163‐165]
and is often treated by mastectomy or lumpectomy combined with adjuvant chemotherapy or
radiotherapy [166‐173]. However, attempts to completely remove the neoplasm surgically
usually result in residual malignant cells, which may induce cancer local recurrence (45.5%) and
metastasis (10.1–39.4%) [174]. Further, the use of adjuvant treatments including radiotherapy
and chemotherapy may lead to unavoidable side effects and systemic cytotoxicity.
Recently, local chemotherapy has been introduced as a promising treatment for cancer
patients. Local chemotherapy aims to effectively combat cancer by delivering the
chemotherapeutic agents directly into the malignant tissues and to reduce the side effects
associated with systematic delivery. To date, several chemotherapeutic implants have been
commercialized to treat prostate cancer (Viadur® and Vantas®), malignant glioma (Gliadel®
wafer), and endometriosis (Zoladex®) [17‐20, 175, 176].
For breast cancer, a few local drug delivery methods have been developed to deliver the
chemotherapy drug DOX utilising soft implants, such as hydrogel and electrospun scaffolds, as
drug carriers to match the mechanical properties of the mammary tissue [177, 178]. For example,
Seib et al. developed DOX–loaded silk hydrogel films and evaluated them in a mouse orthotopic
breast cancer model. After 6 weeks, they observed a significant reduction in the tumor
recurrence compared with i.v. injection using the same dose [179]. Wu et al. combined DOX with
polydopamine coated‐gold nanoparticles in thermoresponsive supramolecular poly(N‐acryloyl
glycinamide‐co‐acrylamide) hydrogels with the aim of having both thermal and
chemotherapeutic effects. Local injection of the hydrogel showed an increase in recurrence‐free
survival rate compared with I.V. injection [177]. Qi et al. employed DOX–loaded hexapeptide
hydrogel and showed reduced tumour progression and lung metastasis in animals subjected to
local delivery compared with I.V. injection of DOX [178]. Application of DOX–loaded
Page 82
Page 82
photothermal therapeutic hydrogels to reduce tumour recurrence in a mouse model mice by half
compared with local injection of DOX [180].
Thermoplastic scaffolds have also been developed to locally deliver chemotherapeutics
to reduce tumour recurrence. Yuan et al. encapsulated DOX–loaded mesoporous silica
nanoparticles into electrospun poly(lactic‐co‐glycolic acid) (PLGA) nanofibers and observed the
sustained release of DOX over 120 days. In vivo data showed a significant increase in cancer cell
apoptosis, as indicated by the increased level of caspase‐3 [181, 182]. Ding et al. showed poly‐
D,L‐lactide electrospun nanofibers–loaded with docetaxel, one of the commonly used
chemotherapeutic agents for breast cancer, increased recurrence‐free survival rate compared
with I.V. injection of docetaxel with the same doses [183].
Despite promising preliminary outcomes from the hydrogels and electrospun implants,
the scaffolds in the systems described above were mostly designed to serve solely as drug carriers
and do not consider the possibility of tissue regeneration. Therefore, in the context of tissue
reconstruction for breast cancer patients, new type of implant should be able to support breast
tissue regeneration and gain the body shape after surgery. Thus, it has been suggested that the
next generation of scaffolds used in such reconstructive surgeries should be designed to both
support tissue formation and deliver anti‐cancer drugs locally.
Additive manufacturing techniques have been widely employed to manufacture
deterministically complex scaffolds for large—volume tissue regeneration [3]. The application of
these scaffolds in drug delivery has also been demonstrated. For example, Min et al. loaded
isoniazid, an anti‐tuberculosis drug, on 3D‐printed mesoporous bioceramic scaffolds bound with
poly (3‐hydroxybutyrate‐co‐3‐hydroxyhexanate) and demonstrated sustained release above the
minimum inhibitory concentration for up to 12 weeks in rabbits [56]. Hollander et al. pre‐mixed
polycaprolactone (PCL) with indomethacin before extrusion into a T shape structure for
intrauterine system application [184]. Recently, Ahangar et al. locally applied a 3D printed DOX–
loaded PCL disk to inhibit prostate cancer bone metastases [185]. The system had a linear release
Page 83
Page 83
profile for 7 days in vitro. However, with the current approach of mixing bioactive agents with
polymers/copolymers before printing using 3D printing (3DP) for drug delivery, the bioactivities
of the loaded agents may be altered because of the elevated heat or chemicals used during the
printing process and the release of the agents depends on the degradation rate of the polymers,
which may be exceedingly low.
Recently, we designed a research programme focusing on a new type of bimodal porous
biodegradable polymeric scaffold that has dual microscale and macroscale porosity, prepared
using a melt extrusion additive manufacturing technique in combination with salt leaching. Using
PCL as a model polymer and simply mixing it with a phosphate salt microparticle porogen before
extrusion, we demonstrated that we can manufacture scaffolds of deterministically complex
macroscale architecture that contain microscale pores of controlled size after salt leaching. We
showed that the scaffold (1) supported bone tissue formation in a rat critical‐size calvarial defect
model [186] and (2) was readily loaded with antibiotics and chemotherapeutic drugs through
simple soak loading to have extended drug release [41].
The central hypothesis in our research programme is that these scaffolds can be readily
loaded with drugs that will release locally to prevent complications (such as cancer recurrence),
and that after the drugs have been exhausted, tissue regeneration will occur with the support
from the bimodal scaffold architecture. In the current study, we aimed to investigate the efficacy
of these bimodal porous biodegradable scaffolds for local delivery of chemotherapeutic agent,
DOX, to prevent/reduce cancer recurrence and the metastasis burden in vivo. We used pmPCL as
a model biodegradable polymer because this material has been commonly used for implant
products. The scaffolds were manufactured by extruding the pmPCL–porogen composite into 3D
constructs, and then the porogen was leached out and intra‐strut microscale porosity was
achieved. The bioactivities of DOX loading on pmPCL scaffolds were characterized in vitro and the
application of DOX‐loaded pmPCL scaffolds to prevent/reduce breast cancer recurrence and
metastasis was demonstrated in a mouse orthotopic breast cancer model.
Page 84
Page 84
4.3 MATERIALS AND METHODS
4.3.1 Scaffold fabrication
The bimodal porous pmPCL scaffolds were fabricated as described elsewhere [187].
Briefly, 38 μm porogen (microparticles grinded and sieved from phosphate buffered saline [PBS]
tablets [Dulbecco A, OXOID, UK]) was mixed with pmPCL (PURASORB® PC12, Corbion,
Netherlands) at a 5:4 weight ratio in chloroform, cast onto glass surfaces and left overnight for
solvent evaporation. The composite was loaded into an in‐house built 3D printer—a computer‐
controlled SME device—as described elsewhere [187] and printed at 110°C with 0.3 mm struts,
1 mm inter‐strut distance, 0.75 mm thickness and 0/90° layout pattern. The porogen in the
printed constructs was leached out in NaOH (0.1 M) in 14 days.
The morphology of the pmPCL scaffolds was characterized by stereomicroscopy,
scanning electron microscope (SEM) imaging and micro‐computed tomography (μ‐CT) scanning.
For SEM imaging, samples were fractured in liquid nitrogen, gold coated and imaged using an
SEM (Zeiss FESEM, Carl Zeiss AG, Oberkochen, Germany). For μ‐CT scanning, X‐rays were
generated by a GE‐Phoenix xs 180 nf micro‐focus X‐ray source (General Electric, Massachusetts,
US) and radiographs were captured on a 3000x3000 silicon flat panel detector. X‐ray tube settings
were 80 kV, 110 μA and 2.2 μm voxel size. The 3D reconstruction was undertaken using a helical
cone‐beam filtered back projection method.
4.3.2 Drug loading
To prepare the DOX‐loaded scaffolds, the pmPCL scaffolds were punched using a 6 mm
biopsy punch, sterilized in ethanol 70% and left drying out overnight in sterile conditions. DOX
(Sigma‐Aldrich, US) was dissolved in 100% methanol at different concentrations and added into
Eppendorf tubes containing the sterile scaffolds, as detailed in Table 1. To characterize the
chemotherapeutic effect in vitro, three doses of DOX were used for cell viability assays (Table 1).
For implantation study, DOX was loaded on scaffolds at two doses (Table 1), which were 5%
Page 85
Page 85
(2 ug/scaffold) and 20% (8 ug/scaffold) of the I.V. injection dose (2 mg/kg). DOX was loaded on
pmPCL scaffolds by immersing the scaffolds in DOX solutions and leaving under sterile conditions
overnight for drug adsorption and solvent evaporation. The drug‐loaded scaffolds were
transferred to fresh tubes for further studies.
Table 1. Loading doses of doxorubicin (DOX) on bimodal porous medical grade
polycaprolactone (pmPCL) scaffolds.
Experiment Group Drug loading concentration Doses
In vitro cell viability
assay
Low Dose 20 µg/mL 0.4 µg DOX/mg scaffold
High Dose 500 µg/mL 10 µg DOX/mg scaffold
In vitro drug release
study &
In vivo implantation
study
Low Dose 20 µg/mL 2 µg DOX/scaffold
High Dose 100 µg/mL 8 µg DOX/scaffold
The loading of DOX on scaffolds was confirmed by stereomicroscopy, fluorescence
imaging (Nikon spectral spinning disc confocal microscope) and Fourier‐transform infrared
spectroscopy (FT‐IR). FT‐IR analyses were carried out with a Nicolet FTIR spectrophotometer
(Nicolet Analytical Instruments, US) at the following settings: scan number, 64; resolution, 1 cm‐
1; and wavenumber range, 4000–400 cm‐1.
4.3.3 In vitro chemotherapeutic effect of doxorubicin–loaded scaffolds against
breast cancer
To investigate in vitro chemotherapeutic effects of DOX‐loaded scaffolds against breast
cancer, MDA‐MB‐231 cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco,
US) supplemented with 10% fetal bovine serum (FBS, Gibco), 100 U/mL penicillin and 100 µg/mL
streptomycin (Gibco). Cells were collected at 80% confluence and seeded in 24‐well plates with
Page 86
Page 86
a cell density of 1.5 x 104 cells/well. After overnight incubation, one DOX‐loaded scaffold was
added into each well. A group with DOX in solution (Free DOX), at an amount equal to that loaded
on the High Dose scaffolds, was used to compare different treatment methods and investigate
the bioactivity of the loaded drug compared with the free drug. Cell viability was assessed at Days
2, 5 and 7 post treatment using Alamar Blue assays (Invitrogen, US). Data were presented as the
percentage of viable cells compared with the Control group (cells with no treatment).
4.3.4 In vitro drug release
Samples were placed in 2 mL Eppendorf tubes containing 200 μL PBS. Then, the
samples were placed in a shaking incubator at 37°C at 200 rpm, and the release solutions
were completely collected and refreshed at each timepoint for measurement. The
absorbance of the drug releasing solutions was measured using an xMark microplate
absorbance spectrophotometer (Bio‐Rad, US) at 480 nm wavelength, and the released
amounts were calculated using the standard curve.
4.3.5 Application of doxorubicin–loaded scaffolds in preventing breast cancer
recurrence and metastasis
4.3.5.1 Cell culture
Luciferase‐expressing MDA‐MB‐231‐bo (TxSA) cells were cultured in DMEM,
supplemented with 10% FBS, 100U/ml penicillin (Invitrogen) and 100ug/mL streptomycin
(Invitrogen) until reaching 80% confluence.
4.3.5.2 Tumour establishment
The animal experiments were approved by the Animal Ethics Unit of the University of
Queensland’s Animal Ethics Committees (approval number: AE21569). In this study, 28 female
NOD‐SCID IL2Rγnull (NSG) mice (4‐week‐old) were purchased and housed in groups of five in
individual cages at the Translational Research Institute (TRI, Brisbane, Australia) for 2 weeks prior
to the experiment. Orthotopic breast tumour was established by injecting 5 × 105 TxSA cells
Page 87
Page 87
suspended in 50 μL PBS into the fourth and fifth right mammary fat pads of the animals. The
tumour growth was monitored by bioluminescent imaging (BLI) for 2 weeks. Images were
acquired 10 minutes after intraperitoneal injection of 200 μL XenoLight D‐Luciferin Potassium Salt
(PerkinElmer, US) at 1.5 mg/L using a Xenogen IVIS Spectrum (PerkinElmer, US). The total flux
signals from cancer cells were quantified using the Living Image v4.5.2 software (PerkinElmer,
US).
4.3.5.3 Treatment
Two weeks after inoculation, the tumours were resected. For the Control group (scaffold
without drug), Low Dose and High Dose groups, the scaffolds were implanted at the resected site.
For the I.V. group, DOX was administrated only one time by tail vein injection. The treatments
(Table 2) were continued for 4 weeks, during which time the animals were weighed weekly and
tumour recurrence was monitored weekly by BLI as previously mentioned. At 4 weeks post
treatment, the animals were euthanized; the implants and organs were retrieved. The tumour
recurrence and cancer metastasis were determined by ex vivo BLI as described before. The
samples were fixed in 4% paraformaldehyde for 24 hours at 4°C and placed in 80% ethanol for
further analysis.
Table 2. Treatment groups for in vivo implantation.
Group Treatment
Control Bimodal porous medical grade polycaprolactone (pmPCL) scaffold without
drug
I.V. Intravenous tail vein injection of doxorubicin (DOX) at 2 mg/kg
Low Dose DOX–loaded pMPCL scaffold at 2 μg DOX/scaffold (5% of I.V. dose)
High dose DOX–loaded pMPCL scaffold at 8 μg DOX/scaffold (20% of I.V. dose)
Page 88
Page 88
4.3.5.4 In vivo drug release in blood
To evaluate the release behaviour of DOX from DOX‐loaded scaffolds, DOX concentration
in the serum harvested from peripheral blood was quantified. At Days 1, 7 and 28 of the
treatment, a volume of 100 μL blood was collected from each mouse; then, samples were left at
room temperature (RT) for 20 minutes and centrifuged at 350 × g for 5 minutes. The blood serum
was collected and mixed with the internal control (vancomycin 0.002 μg/mL) at 1:1 volume ratio.
Then, 20 μL of the mixture was added into 80 μL extraction solution (5 mM ammonium acetate
pH 5 and acetonitrile at 2:3 volume ratio), vortexed, placed on ice for 10 minutes and centrifuged
at 15000 × g for 10 minutes. The supernatant was collected and analysed using a Nexera X2
UHPLC system coupled to a triple quadrupole mass spectrometer (Shimadzu LCMS‐8050,
Shimadzu, Japan). The separations were carried out at a flow rate of 0.3 mL/minute using a
Kinetex® 2.6 μm EVO C18 100 Å LC column 100 x 2.1 mm (Phenomenex, Japan), with mobile
phase A consisting of water/0.1% formic acid and mobile phase B consisting of acetone
nitrile/0.1% formic acid. The auto‐sampler was set at 5°C with an injection volume of 5 μL. DOX
was ionized using a positive‐ion electrospray with nebulising gas flow of 3 L/minute, heating gas
flow of 10 L/minute, drying gas flow of 10 L/minute, interface temperature of 300°C, desolvation
line (DL) temperature of 250°C and heat block temperature of 400°C. Quantitative analysis was
carried out using the selected reaction monitoring (SRM) mass transition from 544.20 m/z to
397.00 m/z with a dwell time of 22 ms per ion and 1 ms delay between each measurement.
4.3.5.5 Histology and immunohistochemistry staining
Samples were sectioned at 5 μm thickness, deparaffinated in xylene, rehydrated through
a series of ethanol (100%, 90% and 70%) and stained with haematoxylin–eosin (H&E, Sigma). For
immunohistological analyses to detect the human breast cancer cells after transplantation, anti‐
nuclear‐mitotic apparatus protein antibody (NuMA, ab97585, Abcam, UK) was used. After the
rehydration step, slides were placed in antigen retrieval buffer (Tris‐sodium citrate pH 6) and heat
treated in a decloaking chamber at 95°C for 5 minutes. Then, the slides were incubated with H2O2
Page 89
Page 89
for 5 minutes at RT, permeabilized with 0.1% Triton X‐100 for 5 minutes, blocked with 2% bovine
serum albumin (Sigma‐Aldrich, US) at RT for 30 minutes and incubated with primary antibodies
for 1 hour at RT. Then, slides were incubated with DAKO Envision+ Dual Link System HRP for 45
minutes. Last, the slides were incubated with DAKO Liquid DAB+ Substrate Chromogen for colour
development.
Masson’s trichrome staining was performed for heart and liver sections to evaluate the
level of fibrosis after treatment. Slides were deparaffinated and rehydrated as explained before.
Then, slides were placed in 1.3% aqueous picric acid (Sigma‐Aldrich, US) at 60°C for 1 hour, and
next placed in running tap water for 5 minutes. After that, the slides were stained with Weigert’s
haematoxylin (Acros Organics, US) for 10 minutes and Ponceau‐Fuchsin working solution
(Masson’s Red, Sigma‐Aldrich, US) for 5 minutes, and differentiated in phosphotungstic acid (PTA,
Sigma‐Aldrich, US) for 6–10 minutes, with 1‐minute rinsing in water between each staining step.
Last, the slides were transferred to methyl blue (Alfa Aesar, US) for 1–2 minutes to stain the
collagen fibres blue.
4.3.5.6 Statistical analysis
Data are expressed as mean ± SEM. The significant differences between groups were
analysed by one‐way ANOVA test and Tukey post hoc test using SigmaPlot 13.0 software.
Significant differences were recorded when p < 0.05. *, **, *** and ns in figures indicate
p < 0.05, p < 0.01, p < 0.001 and data not statistically significant, respectively.
4.4 RESULTS
4.4.1 Scaffold characterization
Bright field images showed a uniform macroporous structure for the printed scaffolds
(Figure 19A). The incorporated porogen microparticles (Figure 19B) were completely leached out,
thereby creating microscale pores with a size of 1–100 µm (Figure 19C). The µ‐CT micrographs
(Figure 19D) showed the uniform distribution of microscale pores throughout the scaffolds, with
Page 90
Page 90
pore area of 5–840 μm2 (Supplementary data, Figure S7). Moreover, the SEM and µ‐CT
micrographs showed that the large microscale pores were mainly located inside the struts (20–
70 µm), while smaller microscale pores were located close to the surface of the struts (2–10 µm)
(Figure 19D, arrows). The pores inside the struts were interconnected, with open pores on the
surface of struts indicating the interconnectivity of the microscale porous networks.
Figure 19. Bimodal porous medical grade polycaprolactone (pmPCL) scaffolds fabricated
using the screw melt extrusion method with interconnected pores. A: Representative bright field
microscopic image of pmPCL scaffold. B and C: SEM micrographs of cross section of pmPCL
scaffold strut before (B) and after (C) the porogen was leached out. D: µ‐CT micrograph of pmPCL
scaffold after leaching. Black and white arrows: microscale pores on the surface of the strut.
Page 91
Page 91
4.4.2 Drug loading
After loading of the scaffolds with DOX, they became red in colour (Figure 20A), indicating
homogeneous and uniform distribution of the DOX in the scaffolds. The distribution of DOX was
further measured by fluorescence imaging, which showed DOX both on the strut surface and
inside the intra‐strut microscale pores (Figure 20B and C). The presence of DOX on the pmPCL
scaffolds was also validated by its signature peaks (1612 cm‐1 and 1577–1579 cm‐1 for the
aromatic ring structure) in the FT‐IR spectrum (Figure 20D).
Figure 20. Loading of doxorubicin (DOX) on bimodal porous medical grade
polycaprolactone scaffolds (pmPCL), dose‐dependent chemotherapeutic effect on breast cancer
cells and in vitro linear sustained release of the loaded DOX. A: stereomicroscopic representative
image of DOX–loaded pmPCL scaffold. B and C: representative images of surface (B) and cross
section (C) of DOX–loaded pmPCL scaffold strut from spinning disc confocal microscope. Arrow:
scaffold surface. D: Fourier‐transform infrared spectra of pmPCL scaffold, DOX and DOX–loaded
pmPCL scaffold (DOX/pmPCL). E: in vitro cell viability of MDA‐MB‐231 breast cancer cell line at 2,
5 and 7 days of treatment with DOX–loaded pmPCL scaffolds at Low Dose (0.4 µg DOX/mg
Page 92
Page 92
scaffold), High Dose (10 µg DOX/mg scaffold) and DOX in solution (Free DOX). Data are expressed
as mean ± SEM (n=6). F and G: cumulative release of DOX–loaded from pmPCL at Low Dose (2 µg
DOX/scaffold) and High Dose (8 µg DOX/scaffold) groups in PBS in the first 30 hours (F) and in 28
days (G). Data are expressed as mean ± SEM (n=3 scaffolds from 3 independent experiments). **
is p<0.01; One‐way ANOVA and Tukey post hoc test.
4.4.3 In vitro chemotherapeutic effect of doxorubicin–loaded scaffolds against
breast cancer
An in vitro cell viability assay with the MDA‐MB‐231 breast cancer cell showed no
significant difference in cell survival rate among groups at day 2 (58.9±11.0%, 56.2±21.9% and
39.9±14.9% for Low Dose, High Dose and Free DOX groups, respectively, Figure 2E). At day 5, DOX
loaded on the scaffolds maintained its dose‐dependent chemotherapeutic effect against the
cancer cells, indicated by a significant decrease in cell viability from 22.7±5.1% in the Low Dose
(0.4 µg/mg) group to 0.4±0.03% in the High Dose (10 µg/mg) group at day 5 (Figure 20E). There
was also no statistically significant difference in terms of cell viability between the High Dose and
Free DOX groups (0.9±0.7%). The dose‐dependent trend was maintained up to day 7. At day 7,
High Dose DOX/pmPCL groups displayed reduced cell viability to less than 0.5%, which was
significantly less than Low Dose (8.1±1.1%) (Figure 20E; p<0.01) but similar to Free DOX.
4.4.4 Drug release behaviour of doxorubicin–loaded bimodal scaffolds
The drug release profile of DOX–loaded scaffolds was investigated prior to the
implantation experiments. The doses used for drug release study were the same as those used
for the in vivo implantation studies, which were 2 and 8 µg DOX/scaffold for Low Dose and High
Dose, respectively.
The drug release study showed that in both Low Dose and High Dose groups, DOX–
loaded on pmPCL scaffolds had a linear release profile for more than 28 days with constant rates
Page 93
Page 93
of 0.1 and 0.4 ng/scaffold/hour for Low Dose and High Dose, respectively. The increase in the
amount of the loaded drug did not prolong the release further but increased the amount of drug
released from 0.04 µm/scaffold to 0.2 µm/scaffold after 1 hour, from 0.06 µm/scaffold to 0.3
µm/scaffold after 24 hours (Figure 20F) and from 0.13 µm/scaffold to 0.53 µm/scaffold after 672
hours (Figure 20G). Overall, the pmPCL scaffolds were able to release the DOX at a constant rate.
4.4.5 Application of doxorubicin–loaded scaffolds in preventing breast cancer
recurrence and metastasis
4.4.5.1 Doxorubicin concentration in peripheral blood
The concentration of DOX from pmPCL scaffolds into the peripheral blood after
implantation was quantified using HPLC‐MS by quantifying the level of DOX in the serum and
compared with I.V. injection of DOX. HPLC‐MSC results showed the DOX level was only detectable
in the I.V. group at day 1 and 7 after injection (Figure 21A). For animals treated with DOX–loaded
scaffolds the concentration was below the limits of detection (was <1.5 ppb) (Figure 21A). Overall,
the pmPCL scaffolds were able to release the DOX at a constant rate in vitro but for in vivo
experiment, the concentrations of DOX in the blood serum was limited to be detected.
Conversely, in I.V. group, DOX can be detected up to 7 days after injection.
4.4.5.2 Cytotoxicity of doxorubicin–loaded scaffolds.
The cytotoxicity of DOX–loaded scaffolds was evaluated by monitoring the changes in
animal weight during the treatment as well as by histological and immunohistochemical staining.
Regarding the weight, as showm in Figure 21B, significant weight loss was recorded in the I.V.
group (animals that received one I.V. injection of DOX) compared with the other groups at week
1 after treatment. At the second week, animal weight was increased in the I.V., Low Dose and
High Dose groups, but less than that in the Control group.
Page 94
Page 94
Figure 21. The doxorubicin (DOX) was only detected in the peripheral blood of animals
received the I.V. injection and reduced the body weight at week 1 post‐treatment. A: The
concentration of DOX in the peripheral blood at day 1, 7 and 28 post‐treatment. B: Animal weight
fold increment compared with the weight of animal at the time the treatment started. Data are
expressed as mean ± SEM (n=6). ns: data is not statistically significant; @: Control group was
significantly different from Low Dose and High Dose (p<0.01); #: Control group was significantly
higher than Low Dose (p,0.01); * is p<0.05; ** is p<0.01 from one‐way ANOVA and Tukey post
hoc test.
With histological and immunohistochemical staining, hematoxylin‐eosin (H&E) and
Masson’s trichrome staining were performed to investigate the cytotoxic effects of DOX in liver
and heart. In liver, histological analysis revealed no difference in the percentage of congested
Page 95
Page 95
central veins among all the groups (Figure 22A). Further, Masson's trichrome staining results
showed no significant difference in the fibrotic tissue area percentage among all the groups
(Figure 22A). Taken together, no hepatotoxicity effects were detected in all of the experimental
groups at day 28 of the treatment.
In heart, Masson’s trichrome staining confirmed the presence of fibrotic tissue in all the
experimental groups (Figure 22B; low and high magnifications). Quantification of fibrotic tissue
area showed no statistically significant difference between the Control group and the treatment
groups. However, both Low Dose and High Dose groups showed significantly less fibrotic tissue
area than did the I.V. group (7.3±0.6% and 7.1±0.5% for Low Dose and High Dose, respectively,
compared with 9.2±0.3% for I.V., Figure 22B). Taken together, the local delivery of DOX using
pmPCL scaffolds reduced cardio cytotoxicity compared with the I.V. injection method. Overall,
the DOX–loaded pmPCL scaffolds, containing less amount of DOX, significantly reduced the
cytotoxic effects compared with I.V. group, which was indicated by significant weight loss in
animals at week 1 after I.V. injection of DOX as well as increased fibrotic tissue area percentage
in the heart.
Page 96
Page 96
Figure 22. Cytotoxic effects in heart and liver of animals receiving intravenous (I.V.)
doxorubicin (DOX) and medical‐grade polycaprolactone (pmPCL) scaffolds loaded with high dose,
low dose and no DOX. Implantation of DOX‐loaded pmPCL scaffolds had similar mild cytotoxic
effects on the liver and less cardio‐cytotoxicity compared with I.V. injection of DOX. A: H&E
staining (top row) and Masson’s trichrome staining (middle row) of liver; congested central vein
percentages measured from H&E staining and fibrotic tissue percentages measured from
Masson’s trichrome staining of the Control (bimodal porous medical grade polycaprolactone
Page 97
Page 97
(pmPCL) scaffolds without drug), I.V. (I.V. injection of doxorubicin (DOX) at 2 mg/kg), Low Dose
and High Dose (DOX–loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups. B:
Masson’s trichrome staining of heart (top and middle rows), and fibrotic tissue percentages
measured from Masson’s Trichrome staining of the Control, I.V, Low Dose and High Dose groups.
Data are expressed as mean ± SEM (n=6 mice for each group). ns: data is not statistically
significant; * is p<0.05 from one‐way ANOVA and Tukey post hoc test.
4.4.5.3 Chemotherapeutic effect of doxorubicin–loaded scaffolds
In vivo BLI showed tumour recurrence in the Control and I.V. groups at 1 week earlier
than in the Low Dose and High Dose groups (Figure 23A). At Day 14, recurrent tumour was
detectable in the Control and I.V. groups using IVIS. At Day 21, tumour recurrence was detected
in all groups, and recurrent tumours had enlarged at Day 28. Quantification of in vivo total flux
normalized to Day 0 of the treatment showed that at Day 7, only the Low Dose group had
significantly reduced tumour growth (2.1 ± 0.2 fold increase, n = 6) compared with the Control
group (6.0 ± 1.5 fold increase, n = 6) (Figure 23B). There was no statistically significant difference
between the I.V. (19.9 ± 6.3 fold increase, n = 10) or High Dose (3.0 ± 0.9 fold increase, n = 6)
group and the Control group. At Day 14, both Low Dose and High Dose groups (3.9 ± 1.4 and
81.9 ± 53.5 fold increase, respectively) had significantly less tumour growth than did the Control
group (377.9 ± 98.3 fold increase). At Day 21, the tumour growth in the Control group was
significantly greater than that in the treatment groups. Ex vivo total flux quantification of local
recurrent tumours at Day 28 of the treatment showed that the recurrent tumours in the Control
group (335.7 ± 125.3) were larger than in the I.V., Low Dose and High Dose groups (134.5 ± 103.5,
21.7 ± 12.1 and 17.6 ± 5.9, respectively) (Figure 23C). Both Low Dose and High Dose groups had
significantly less tumour growth than did the I.V. and Control groups.
Page 98
Page 98
Figure 23. Tumour recurrence in animals treated with intravenous (I.V.) doxorubicin
(DOX) and medical‐grade polycaprolactone (pmPCL) scaffolds loaded with high dose, low dose
and no DOX. Local drug delivery reduced the local tumour recurrence. A: Representative in vivo
bioluminescence images of mice at days 0, 7, 14, 21 and 28 post treatment in Control (bimodal
porous medical grade polycaprolactone (pmPCL) scaffolds without drug), I.V. (I.V. injection of
doxorubicin (DOX) at 2 mg/kg dose), Low Dose and High Dose (DOX–loaded pmPCL with 2 and 8
µg DOX/scaffold, respectively) groups. B: quantitative in vivo IVIS signal over 21 days of
treatment. C: ex vivo IVIS signal of local cancer recurrence at day 28 post treatment. Data are
expressed as mean ± SEM (n=6 mice in each group). # is statistically significant compared with
Control group, p<0.05; ns: data is not statistically significant; * is p<0.05 and ** is p<0.01 from
one‐way ANOVA and Tukey post hoc test.
Page 99
Page 99
Bright field images showed that the recurrent tumours in all groups were mature and
vascularized (Figure 24A). Images from H&E staining (Figure 24B) showed that there was multiple
tumour recurrence occurred locally in all animals in the Control group. The malignant tissues in
the Control group had different stages of tumour development, some of the tumour tissue was
dense, while some were loose at the centre. In the other groups, all animals exhibited only single
local recurrent tumours. In the Low Dose and High Dose groups, the tumours only grew next to
the scaffolds but did not encapsulate or infiltrate the inner pores of the scaffolds. NuMA staining
(Figure 24C) images showed that positive cells were observed in all groups, indicating that the
recurrent tumours were from the injected human breast cancer cells. Overall, the histological and
immunohistochemical staining showed that all recurrent tumours were highly proliferative and
vascularized with single recurrence occurred in all animals of the treatment groups and multiple
recurrence occurred in all animal of the Control group.
Figure 24. Local drug delivery reduced the number and volume of the highly proliferative
and vascularized local recurrent tumours. The representative bright field (A), H&E staining (B) and
anti‐NuMA (human specific marker) staining images (C) of the recurrent tumour at the primary
tumour site of the Control (bimodal porous medical grade polycaprolactone (pmPCL) scaffolds
Page 100
Page 100
without drug), I.V. (I.V. injection of doxorubicin (DOX) at 2 mg/kg), Low Dose and High Dose (DOX–
loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups.
The chemotherapeutic effects of the treatments on metastases were investigated by
weighing and ex vivo BLI analysis of different organs. In lung, the macroscopic metastatic foci
were only observed in the Control and I.V. groups (Figure 25, top row). The presence of metastatic
cells in lung was confirmed by immunostaining using human specific NuMA antibody (Figure 25).
Moreover, NuMA staining showed more tumour cells (stained brown) in the I.V. group than in
the other groups (Figure 25, second row). The metastasis might have increased the weight of the
lungs in the Control and I.V. groups compared with the Low Dose and High Dose groups (Figure
25). Quantification of total flux showed that the Low Dose and High Dose groups had less signals
than did the Control and I.V. groups, but the data were not statistically significant. NuMA staining
revealed that the tumour cells were distributed as clusters/micrometastases in the I.V. and
Control groups. In the other groups, there were more single tumour cells than micrometastases.
Quantification of number of tumour cells and micrometastases confirmed that both the Low
Dose group and the High Dose group had significantly less tumour cells and micrometastases
compared with the I.V. group (Figure 25). There was no significant difference between the I.V.
group and the Control group. Overall, it can be said that the local treatment using DOX‐loaded
scaffolds significantly reduced breast tumour metastasis in lung compared with the I.V. injection
method.
In liver, no noticeable difference in surface morphology was found among all groups
(Supplementary data, Figure S9A), which was confirmed by similarity in organ weight (Figure 25B,
bottom row). Ex vivo total flux quantification showed similar trends with lung signals. In the Low
Dose and High Dose groups, metastasis progression, which was indicated by total flux signal, was
less than in the Control and I.V. groups.
Last, analysis of spleen showed that animals in both the Low Dose group and the High
Dose group had smaller spleen size than did those in the I.V. and Control groups (Supplementary
Page 101
Page 101
data, Figure S9B). Regarding the ex vivo total flux quantification, the Low Dose and High Dose
groups had smaller signals than did the Control and I.V. groups (Supplementary data, Figure S9B).
Overall, the ex vivo BLI signal quantification showed that the Low Dose and High Dose
groups reduced tumour progression in lung, liver and spleen compared with both the Control and
the I.V. groups. Conversely, there was no statistically significant difference found between the
I.V. group and the Control group.
Figure 25. Tumour metastasis in animals receiving intravenous (I.V.) doxorubicin (DOX)
and medical‐grade polycaprolactone (pmPCL) scaffolds loaded with high dose, low dose and no
DOX. Doxorubicin (DOX)–loaded bimodal porous medical grade polycaprolactone (pmPCL)
scaffolds effectively reduced the breast tumor metastasis in lungs. Morphology (top row), anti‐
NuMA staining (middle row), organ weight normalized to body weight, total flux from ex vivo BLI
imaging, tumour cell count and micrometastasis count from NuMA staining images using ImageJ
software of lung of the Control (bimodal porous medical grade polycaprolactone (pmPCL)
scaffolds without drug), I.V. (I.V. injection of doxorubicin (DOX) at 2 mg/kg), Low Dose and High
Page 102
Page 102
Dose (DOX–loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups. Arrows: tumour
metastasis foci. The organs were weighted after fixation and normalized to the body weight of
the animal at day 28. Data are expressed as mean ± SEM (n=6 mice in each experimental group).
ns: data is not statistically significant; * is p<0.05; ** is p<0.01 from one‐way ANOVA and Tukey
post hoc test.
4.5 DISCUSSION
Adjuvant systemic chemotherapy is used to eliminate residual cancerous tissue after
surgery but is known to be associated with systemic cytotoxicity [188]. Localized chemotherapy
is a promising approach to enhance treatment effectiveness and reduce systemic toxicity. Several
products are available for local treatment of various types of cancer, such as subcutaneous
implantation of Viadur® (leuprolide acetate implant) to treat prostate cancer, local implantation
of Gliadel® wafer (carmustine implant) to treat malignant glioma and subcutaneous implantation
of Zoladex® drug reservoir to treat endometriosis [17‐20, 175, 176]. However, the products in
these methods are only implanted subcutaneously in the arm or abdomen area, and have a
simple structure and are monofunctional as a drug reservoir. Further, treatment for breast cancer
is still under development. The use of implants as drug carriers is promising as it can provide the
advantages of localized therapy as well as the tissue reconstruction function of scaffolds;
however, little is known about using scaffolds as dual‐function devices for localized therapy and
tissue regeneration, and existing studies mainly focus on bone cancer and bone regeneration.
Therefore, we aimed to develop a dual‐function 3D‐printed scaffold with bimodal porosity as a
local drug carrier and tissue‐supportive scaffold for breast cancer applications.
In this study, we developed a drug‐loaded scaffold and investigated its efficacy in
preventing cancer recurrence. The scaffolds were fabricated by the SME method combined with
porogen leaching to achieve bimodal porosity. The combination of SME and porogen leaching
gives more control over porosity. The macroscale pore shape and size are controlled by the
Page 103
Page 103
printing pattern, and the microscale pore shape and size are controlled by the input porogen.
Further, as this method is based on molten polymer, the use of solvent can be removed, giving it
a non‐toxic fabrication process.
After fabrication, the scaffolds were loaded with DOX by using a soaking method, taking
advantage of the capillary action of the microscale pores. Because the drug only adsorbs on the
surface instead of being trapped inside a solid block of polymer, its release mechanism is not
dependent on the degradation of the scaffold (PCL takes more than 2 years to degrade in vivo
[189]). This means that after the drug is completely released, the scaffold still maintains its shape,
and mechanical and material properties, to support tissue reconstruction. This feature is difficult
to achieve by conventional methods of mixing drug with biomaterials before making into
scaffolds reported in the literature [56, 184, 185].
In vitro cell viability assays showed that DOX‐loaded scaffolds showed a dose‐dependent
chemotherapeutic effect in vitro against the breast cancer cell line MDA‐MB‐231. Yuan et al. also
studied the chemotherapeutic effects of DOX‐loaded nanoparticles embedded in hydrogels in
vitro [181]. In their study, the drugs loaded on hydrogels were released for 72 hours and the
release solution was used as the treatment for the cell viability study. MDA‐MB‐231 cells were
seeded in 96‐well plates at 104 cells/well density and cultured with the drug release solution for
24, 48 and 72 hours. Compared with their results, in which the cell viability of MDA‐MB‐231 cells
reduced to only 50% after 72 hours of treatment, our results showed a better outcome. Another
experiment by Seib et al. studied the chemotherapeutic effects of DOX‐loaded silk film (40
µg/film) in vitro [179]. In their study, MDA‐MB‐231 cells were seeded in the transwell insert at
8000 cells/well and the film was placed at the bottom of each well. The treatment was carried
out for 72 hours before MTT assays were performed. Compared with their results, our scaffolds
showed less chemotherapeutic effect. However, it is worth noting that their dose was fivefold
that in our High Dose group. In addition, in both studies, the experiments were designed to
characterize the effects of the release solution on cells, and not the direct contact of cancer cells
Page 104
Page 104
with the implant. These designs were well suited for in vitro study. However, to correlate between
in vitro and in vivo results, the chemotherapeutic effects of drug‐loaded scaffolds on breast
cancer cells via direct contact has to be considered. Therefore, the method of direct culture of
DOX‐loaded scaffolds with breast cancer cells gives a better predictive result for in vivo study.
After demonstration of the chemotherapeutic effects of DOX loaded on bimodal porous
scaffolds against breast cancer in vitro, we demonstrated the application of DOX‐loaded scaffolds
to prevent cancer recurrence and metastasis in vivo.
Drug release characterisation showed that DOX‐loaded pmPCL scaffolds used for
implantation can release DOX at a constant rate for up to 28 days in vitro. This release behaviour
is due to the effect of the intra‐strut microporous structure, which has been documented in our
previous study [190]. This effect is beneficial as it can be used to maintain the drug release
amount at the effective concentration throughout the implantation experiment, as opposed to
the single treatment in the I.V. group. Compared with other methods of loading drug by mixing
directly in the polymer solution for electrospun scaffolds or making hydrogels [181, 191‐193], our
method, besides having homogeneous distribution of drug throughout the scaffold, has a release
profile that is independent of the scaffold degradation. Therefore, we can alter the release profile
while maintaining the designed structure and properties of the scaffold for tissue regeneration.
The DOX‐loaded scaffolds were then used in a mouse orthotopic breast cancer model. In
our study, human breast cancer cells were injected into the mammary fat pad and inoculated for
2 weeks. After that, the tumour was resected, and the scaffold was implanted at the resected
site. The implantation of DOX‐loaded scaffolds was compared with I.V. injection of DOX in terms
of chemotherapeutic effects against local recurrence and metastasis as well as systemic cytotoxic
effects on organs. It is important to mention that the scaffolds were only loaded with 5%
(2 μg/scaffold, Low Dose) and 20% (8 μg/scaffold, High Dose) of the amount of the I.V. injection
dose (2 mg/kg).
Page 105
Page 105
We achieved increased treatment efficiency with reduced cytotoxicity when using
scaffolds loaded with only 5% or 20% of the amount used in the I.V. injection group. Specifically,
local drug delivery from the scaffolds significantly decreased the level of DOX circulated in the
blood system, which reduced the systemic cytotoxicity. In the I.V. group, only a small percentage
of the DOX accumulated at the target site; meanwhile, in the DOX‐loaded scaffold groups, the
drug mainly released at the target site and only a fraction of the DOX was diffused into the blood
system and circulated in the body. By delivering DOX locally, its cardio‐cytotoxicity was
significantly decreased, which were indicated by the less development of fibrotic tissue area and
in agreement with literature [194, 195]. In liver, the similarity in fibrotic tissue percentage, as well
as congested central vein percentage, which were reported as side effects of systematic delivery
of DOX [196, 197], among all groups can be explained by two reasons: first, the effect of DOX on
liver is relatively mild, and second, the regeneration rate of liver tissue is higher than that in other
organs.
Regarding the chemotherapeutic effect, the results from BLI showed that the I.V.
group with single administration of DOX had less chemotherapeutic effect to prevent cancer
recurrence than did the DOX‐loaded scaffold groups, which had DOX release for more than 28
days. The better chemotherapeutic effect of the local treatment groups can be explained by the
maintenance of effective concentrations of DOX due to the sustained release of DOX from the
scaffolds. Conversely, the I.V. group only had a short period of treatment (up to 7 days based on
the amount detected in blood); hence, effective concentrations could not be maintained to
prevent tumour recurrence. Of note, in the BLI result, the insignificant differences among the
Control, I.V and DOX‐loaded scaffold groups could have been due to biological and surgical
variations, which means that tumour recurrence might have been exhibited at different times in
individual mice. Because of this limitation, we suggest using the starting point of local recurrence
as the time to deliver the treatment and only choose the animals that have the least variation.
Nevertheless, the present result reflects the real‐life application in which the scaffolds would be
Page 106
Page 106
used as a preventive treatment and we cannot predict which patients will have tumour
recurrence.
Regarding metastasis, it occurred at the highest level in lung, indicated by the presence
of micrometastasis foci on the surface. Despite the fact that only a small amount of DOX
circulated in the blood system in the local treatment groups, the drug‐loaded scaffolds effectively
reduced metastasis progression. It is likely that because the drug‐loaded scaffolds reduced the
primary source of tumour cells at the primary site, their metastasis was reduced.
In both case of preventing cancer recurrence and metastasis, the LD group was appeared
to be more effective than HD group but not significant different. This can be explained as there
was no difference in term of chemotherapeutic effect between LD (2ug/mg) and HD (8ug/mg)
and the differences was due to variation. However, there was an alternative explanation that the
systemic cytotoxicity of DOX released from scaffolds in HD group was higher than LD group, hence
reduced treatment effectiveness. The latter hypothesis needs further investigation in future
study to validate it.
Finally, the bimodal porous scaffolds, after the completion of drug release (of which the
releasing time varies from drug to drug and was expected to be less than 1 year), can maintain
their structure, thanks to its low degradation rate (more than 2 years), to support tissue
reconstruction. When then scaffolds start to degrade, their by products were showed to be
hemocompatible and non‐toxic, hence may not interfere with the tissue regeneration [198].
In the current study, we only focused on the prevention of cancer recurrence. The tissue
reconstruction aspect of these scaffolds is the focus of our future study.
4.6 CONCLUSION
In this study, a new type of 3D‐printed scaffold with intra‐strut porosity was introduced
as a potential method to prevent breast cancer recurrence. DOX‐loaded bimodal porous scaffolds
demonstrated chemotherapeutic effects against breast cancer MDA‐MB‐231 cells in vitro. In vivo
Page 107
Page 107
implantation of DOX‐loaded scaffolds showed a significant decrease in cardio‐cytotoxicity
compared with the I.V. injection method. Further, the scaffolds reduced breast tumour
recurrence and metastasis progression compared with I.V. injection. Overall, this study
demonstrated that local delivery using drug‐loaded scaffolds with only 5–20% of the I.V. injection
dose can not only reduce cytotoxicity but also improve chemotherapeutic effects against tumour
recurrence and metastasis.
Page 108
Page 108
5Chapter 5: 3D printed scaffolds with dual
macro‐, micro‐scale porous networks for
bone tissue regeneration
Hoang Phuc Dang a, b, Cedryck Vaquette c, Tara Shabab a, b, Román A. Pérez d, Ying
Yang e, Tim R. Dargaville a, Dietmar W. Hutmacher a, b, Abbas Shafiee b, f, Phong A. Tran
a, b, g
Submitted to Biomaterials
a ARC Centre in Additive Biomanufacturing, Queensland University of Technology, Musk
Avenue, Kelvin Grove, Brisbane, Queensland 4059, Australia
b Centre in Regenerative Medicine, Institute of Health and Biomedical Innovation,
Queensland University of Technology, Brisbane, Australia
c School of Dentistry, The University of Queensland, Brisbane, Queensland, Australia
d Institute of Bioengineering, School of Dentistry, Universitat Internacional de Catalunya,
Barcelona, Spain
e Australia‐China Centre for Tissue Engineering and Regenerative Medicine (ACCTERM),
Queensland University of Technology, Brisbane, Australia
f UQ Diamantina Institute, Translational Research Institute, The University of Queensland,
Brisbane, Queensland, Australia
g Interface Science and Materials Engineering Group, School of Chemistry, Physics and
Mechanical Engineering, Queensland University of Technology, Brisbane, Australia
Page 109
Page 109
The authors listed below have certified that:
16. They meet the criteria for authorship in that they have participated in the conception,
execution, or interpretation, of at least that part of the publication in their field of expertise;
17. They take public responsibility for their part of the publication, except for the responsible
author who accepts overall responsibility for the publication;
18. There are no other authors of the publication according to these criteria;
19. Potential conflicts of interest have been disclosed to (a) granting bodies, (b) the editor or
publisher of journals or other publications, and (c) the head of the responsible academic unit,
and
20. They agree to the use of the publication in the student’s thesis and its publication on the QUT’s
ePrints site consistent with any limitations set by publisher requirements.
In the case of this chapter:
Submitted manuscript:
Hoang Phuc Dang, Cedryck Vaquette , Tara Shabab, Román A. Pérez, Ying Yang , Tim R. Dargaville,
Dietmar W. Hutmacher , Abbas Shafiee and Phong A. Tran. “3D printed scaffolds with dual macro‐
, micros‐cale porous networks for bone tissue regeneration.”
Contributor Statement of contribution
Hoang Phuc Dang
Performed experimental design, conducted experiments, data
analysis, and wrote the first draft of manuscript Signature
Date:
Cedryck Vaquette Conducted experiments
Tara Shabab Conducted experiments
Román A. Pérez
Page 110
Page 110
Ying Yang Conducted experiments
Tim R Dargaville Aided manuscript preparation
Dietmar W. Hutmacher
Performed experimental design, supervised the project,
performed data analysis and contributed in manuscript
preparation
Abbas Shafiee
Performed experimental design, conducted experiment,
supervised the project, performed data analysis and contributed
in manuscript preparation
Phong A. Tran
Performed experimental design, conducted experiment,
supervised the project, performed data analysis and contributed
in manuscript preparation
Principal Supervisor Confirmation
I have sighted email or other correspondence from all Co‐authors confirming their
certifying authorship
Phong Tran 06/02/2020
Name Signature Date
QUT Verified Signature
Page 111
Page 111
5.1 ABSTRACT
In this study, we combined 3D printing with porogen leaching to develop scaffolds with multiscale
porosity and investigated their regenerative capacity in a rat calvarial defect model. The scaffolds
were additively manufactured from medical grade polycaprolactone (mPCL) doped with porogen
microparticles having an average size of 22 μm, which were subsequently leached to create
microscale porosity. Morphological analysis revealed an interconnected macroscale porosity of
about 60% with an average pore size of 700 µm and intra‐strut microscale pores with a porosity
of nearly 40% and average pore size of 20–70 µm. The microscale porosity resulted in a 3‐fold
increase in the scaffolds’ surface area, a 2‐fold enrichment in negatively charged surface groups,
which did lead to significantly increased protein adsorption and faster hydrolysis‐driven
degradation in vitro. Higher protein adsorption led to increased cell proliferation. An in vitro blood
clot assay demonstrated an increased TGF‐β1 release. In a rat calvarial defect, bone formation
was found in both the macro‐ and microscale pores and was at a similar level when compared
with biomimetic calcium phosphate coated mPCL scaffolds.
5.2 INTRODUCTION
Scaffold‐guided bone regeneration (SGBR) [47, 48] requires a minimal set of design
requirements, of which the combination of a macro‐ and microscale porosity have been shown
to affect tissue regeneration and remodeling [49‐51]. However, only a few techniques are able to
achieve reproducible manufacturing of this type of scaffold porosity designs. Conventional
approach uses molding and salt leaching where pores are created both from the molding and
drying/lyophilization and from phase separation or solvent extraction or porogen leaching [32,
52]. Using this approach, Du et al. fabricated scaffolds with a macro/microscale porous
architecture by casting silk fibroin solution in Teflon molds together with paraffin spheres
followed by freeze‐drying which led to the formation of microscale pores from silk fibroin
network. The paraffin was then leached out to produce macroscale pores [53]. Duan et al. soaked
Page 112
Page 112
a solution of mesoporous bioactive glass (MBG) and polyacrylic acid (PAA) microscale spheres
into polyurethane (PU) sponges [54]. The specimens were dried and heated at 600oC for 6 hours
to burn out PU sponges, which created macroscale pores, and PAA microspheres, which formed
microscale pores. Reed et al. used acrylonitrile butadiene styrene (ABS) plastic to make negative
molds and Poly(L‐lactic acid) (PLLA) to make positive molds with pre‐defined macroscale porous
structures. Then, chitosan‐alginate solutions were either poured into the negative molds or
infused into the positive molds by centrifugation. The constructs were crosslinked with CaCl2 and
lyophilized to create microscale porous networks [55]. Overall, these techniques were limited in
achieving deterministically complex macroscale pores and controlled microscale pore sizes.
Additive biomanufacturing (ABM) has emerged as a unique technology to create
scaffolds with deterministically designed, interconnected macroscale porous networks. A small
number of studies have reported on ABM methods to produce dual macro‐/microscale porous
scaffolds for tissue engineering[38‐41]. Maleksaeedi et al. utilized inkjet 3D printing to fabricate
titanium – polyvinyl alcohol (PVA) scaffolds, in which the macroscale pores were generated by
design and in the intra‐strut pores were generated by sintering process [38]. Zhang et al. also
used inkjet 3D printing and sintering process to fabricate CaP bimodal porous scaffolds [40]. Kim
et al. utilized non‐solvent induced phase separation (NIPS)‐based plotting method to fabricate 3D
polycaprolactone (PCL)/hydroxyapatite scaffolds with macroscale pores generated from the
printing design and microscale pores generated from solvent exchange [37]. However, in these
methods, the pore size was not easily controlled and in the lastest method, it relies on the usage
of toxic solvent. One might conclude the fundamental effects of microscale porosity on tissue
formation have not been well characterized.
Our group has extensive experience in fabricating a wide range of scaffolds using ABM
[187, 199‐201]. By combining ABM with porogen leaching, we were able to produce macroscale
porous scaffolds having intra‐strut microscale pores with controllable pore sizes [41, 202]. In
current study, we hypothesized that the intra‐strut microscale porosity with increased surface
Page 113
Page 113
areas would enhance protein adsorption and stimulate cellular attachment, hematoma
formation and ultimately bone regeneration. Hematoma formation is the first step following
implantation and has been recognized as an important process which guides the subsequent
tissue regeneration [25‐27]. Hematomas contain a fibrin network, inflammatory cells and growth
factors (such as platelet‐derived growth factor AB (PDGF‐AB) and transforming growth factor‐β
(TGF‐β)) with angiogenic and osteogenic potentials that ultimately accelerate bone healing
process [25, 203, 204]. The importance of the of macro‐ and micro‐scale porosity in hematoma
formation and stabilization has been only recently discussed, but the feature of having the
combined interplay of macro and micro porosity has not been addressed [26, 205]. In this study,
we manufactured bimodal porous scaffolds from medical grade polycaprolactone (mPCL) – a
material that has been successfully used to develop a number of FDA‐approved medical devices
[206] and investigated their physicochemical properties, the effects on in vitro protein
adsorption, cellular functions, hematoma formation and eventually in vivo bone regeneration.
5.3 MATERIALS AND METHODS
5.3.1 Preparation and characterization of the porogen, mPCL films and mPCL‐porogen films
Microscale porogen was ground from Dulbecco A’s phosphate buffered saline tablets
(PBS, OXOID, UK) and sieved to obtain particles smaller than 38 µm. The porogen was mixed with
medical grade polycaprolactone pellets (mPCL, PURASORB® PC12, Purac Biomaterials,
Netherlands) in chloroform (mPCL 20%w/v) according to the weight ratios shown in table 3. The
mPCL concentration was chosen based on our pilot study to achieve the viscosity that can
maintain the porogen in the suspension. The mixtures were stirred for 1 hour on a hot plate with
the temperature set at 100oC, cast onto glass surfaces and left in a fume hood for solvent
evaporation overnight at room temperature (RT). The solvent‐cast films (approximately 50 ‐
150µm thick) were collected stored in a low humidity cabinet at RT for further use.
Table 3. Porogen concentration in mixture with mPCL to make scaffolds.
Page 114
Page 114
Group Description porogen concentration
(%w/w) in the mixture
nmPCL nonporous mPCL scaffold 0
pmPCL17 Bimodal porous mPCL scaffolds prepared from
mixture having 17% porogen
17
pmPCL44 Bimodal porous mPCL scaffolds prepared from
mixture having 44% porogen
44
Volume‐average size class of porogen was measured using a Malvern Mastersizer 3000
laser light scattering unit (Malvern Instruments Ltd, Malvern, UK) by dispersing the porogen in
100% ethanol and sonicating for 5 minutes prior to measurement at 25°C.
The influence of porogen on the viscoelasticity and printability of mPCL‐porogen film was
investigated by amplitude sweep and temperature sweep rheological measurements using an
Anton Parr M302 Rheometer (Anton Paar USA Inc., Virginia, US) equipped with a cylindrical
parallel plate geometry having 25‐mm diameter plate and 1 mm gap. The linear viscoelastic
regions in amplitude sweep tests were measured by melting samples at 110°C for 5 minutes and
increasing the shear strain from 0.01% to 150% with a constant angular speed of 10 rad/second.
The viscoelastic behaviors of materials in temperature sweep tests were determined by
maintaining an angular speed of 10 rad/second and a shear strain of 1%, melting samples at
125°C, then gradually decreasing at 5°C steps and equilibrating for 5 minutes prior to each
measurement until reaching the final temperature of 40°C. Graphs of storage (G') and loss (G'')
moduli and loss tangent (tan δ) were plotted for interpretation and analysis.
5.3.2 Fabrication and characterization of mPCL scaffolds
mPCL scaffolds were fabricated using an in‐house built 3D printer as described elsewhere
[187]. Briefly, mPCL pellets or mPCL‐porogen films were melted at 100‐110°C for 30 minutes and
then extruded through a metal nozzle (gauge 22) with 0/90° laydown pattern, size of 40x40x1.5
Page 115
Page 115
mm3 and strut‐to‐strut distance of 1 mm. The porogen (in pmPCL17 and pmPCL44) was leached
out over 14 days in NaOH 0.01M at 37°C under shaking condition. Nonporous scaffolds (nmPCL)
were also treated under the same condition. The theoretical macroscale porosity generated from
the printing setting was calculated by the following equation:
𝑃𝑜𝑟𝑜𝑠𝑖𝑡𝑦 % 𝑉 𝑉𝑠𝑡𝑟𝑢𝑡
𝑉∗ 100
Where V is the overall volume of scaffolds and Vstrut is the total volume of struts
measured from diameter, length and number of struts.
For protein adsorption, blood clot experiment, cell culture and in vivo study, a positive
control group (a well studied calcium‐phosphate coated scaffolds (nmPCL/CaP) [207‐210]), was
used for benchmarking against the experimental group. CaP coating was performed as described
by Vaquette et al. [156]. Briefly, the scaffolds were immersed in 70% ethanol under vacuum for
10 minutes and then treated with 2M NaOH at 37oC for 5 minutes under vacuum. After that, the
scaffolds were rinsed with ultrapure water and immersed in highly saturated simulated body fluid
(SBF 10x) pH 6, vacuum treated for 5 minutes and placed in water bath at 37oC for 30 minutes.
Finally, the scaffolds were rinsed with ultrapure water, treated with 0.5M NaOH at 37oC for 30
minutes and washed with ultrapure water.
Surface and cross‐section morphology of mPCL struts were analyzed by fracturing
scaffolds in liquid nitrogen, gold coating using a Leica EM SCD005 gold coater (Leica Microsystems
GmbH, Wetzlar, Germany) and imaging using a scanning electron microscope (SEM, Zeiss FESEM,
Carl Zeiss AG, Oberkochen, Germany) combined with energy‐dispersive X‐ray spectroscopy (EDX).
Porogen residues were detected by EDX analysis at an accelerating voltage of 10 kV. Additionally,
to quantify unleached porogen residues, the mPCL component of the leached scaffolds was
dissolved in chloroform (1%w/v) and the porogen/chloroform suspensions were quantified via
turbidity measurement at 600 nm using a Beckman Coulter DU 800 spectrophotometer
(Beckman Coulter Inc., California, US).
Page 116
Page 116
5.3.3 Physicochemical characterization
The porosity of scaffolds was characterized using SEM (see section 2.2) and micro‐
computed tomography (‐CT). X‐rays were generated by a GE‐Phoenix xs 180 nf micro‐focus X‐
ray source (General Electric, Massachusetts, US) and radiographs were captured on a 3000x3000
silicon flat panel detector. X‐ray tube settings were 80 kV, 110 μA and 2.2 μm in voxel size. The
reconstruction was undertaken using a helical cone‐bean filtered back projection method. The
intra‐strut microscale porosity of microscale porous scaffolds was measured based on their
masses and dimensions. mPCL struts were cut into small specimens in order to measure the
length, radius and mass. The porosity of samples was calculated using the following equation:
𝑝𝑜𝑟𝑜𝑠𝑖𝑡𝑦 % 𝜋𝑟 𝑙
𝑚𝑑
𝜋𝑟 𝑙 100
where r, l and m are the radius, length and mass of strut, respectively, and d is the density
of PCL (1.145g/cm3).
The surface areas of nonporous (nmPCL) and microscale porous scaffolds (pmPCL17 and
pmPCL44) were measured using a Micromeritics ASAP 2020 accelerated surface area and
porosimetry system (Micromeritics, Georgia, US) and the data was analyzed following Brunanuer,
Emmett and Teller (BET) adsorption isotherm equation as described elsewhere [211, 212].
To analyze the surface roughness, strut surfaces were scanned by an atomic force
microscope NT‐MDT Solver SPM apparatus (NT‐MDT Spectrum Instruments, Moscow, Russia)
equipped with uncoated Ted Pella Tap300‐G cantilevers (Ted Pella Inc., US) [138].
The crystallinity was measured using differential scanning calorimetry (DSC). The
temperature was changed from ‐60 to 100oC at a rate of 10oC/minute to determine the melting
enthalpy. The crystallinity was calculated by comparison with the heat of fusion of 100%
crystalline PCL of 139.5 J/g [213].
Wettability of both groups was evaluated using contact angle measurement. For this
experiment, samples were prepared as films to eliminate the effects of architecture. Briefly,
solvent cast nmPCL, pmPCL17 and pmPCL44 films were cut into specimens of 1x1 cm2, placed on
Page 117
Page 117
glass slides and melted at 110oC for 3‐5 minutes, immersed in NaOH 0.01 M for 14 days for
leaching, washed with ultrapure water and dried prior to measurement. Images of contact angle
measurement were recorded and analyzed using First Ten Angstroms FTA 200 analyzer (First Ten
Angstroms Inc., Virginia, US). A volume of 20 µL ultrapure water was dropped on the surface of
samples and the contact angle was measured after 5 seconds.
To determine the density of negatively charged groups, Toluidine Blue O (TBO) assay
(Sigma‐Aldrich, US) was performed [214‐217]. Specimens of 5x5 mm2 were prepared and placed
separately into Eppendorf tubes containing 2 mL of TBO solution (2 mM in NaCl 0.015 M, pH 11).
After a 5‐minute sonication treatment, samples were kept at RT for 2 hours under shaking
conditions to facilitate the solution penetration into the pores. TBO solutions were removed and
samples were washed by immersing in NaCl 0.015 M, pH 11 and agitating for 4 hours. Then, the
samples were transferred into Eppendorf tubes containing 70% acetic acid, sonicated for 5
minutes and agitated for 1 hour at RT. The solutions were then collected for measuring the
absorbance at wavelength 556 nm using a Bio‐Rad xMark™ microplate absorbance
spectrophotometer (Bio‐Rad, California, US). TBO standard curve was used to determine the
amount of TBO attached on scaffolds, which correspond to the amount of surface negatively
charged groups.
5.3.4 Tensile strength characterization
To determine the effect of the microscale porous structure on the mechanical behavior,
the tensile properties of single struts were studied[218]. Since the mechanical behavior of
materials is different under dry and wet conditions, PBS solution was used to create an
environment similar to body fluid. The specimens were mounted vertically onto MicroTester
5848 instrument (Instron Ltd., High Wycombe, UK) equipped with 500 N load cell to have a 12.5‐
mm gap between the top and bottom grips, saturated with 37oC PBS solution for 1 minute and
elongated at a rate of 1 mm/s for 30 mm for analysis.
Page 118
Page 118
5.3.5 Accelerated degradation
Accelerated degradation studies were performed as previously reported for macroscale
mPCL scaffolds [219‐221]. Samples were placed in NaOH 2M solution and incubated at 37oC in
shaking incubator at 200 rpm. At each time point, the samples were taken out, washed with
ultrapure water and dried in vacuum at 37oC overnight before weighing and imaging by SEM (as
described in section 2.2).
The crystallinity of samples after degradation was evaluated by DSC (as defined in section
2.3) and Fourier‐transform infrared spectroscopy (FT‐IR). FT‐IR spectra of mPCL scaffolds were
collected by measuring at the setting of scan number of 64 and step of 1cm‐1 over a wave number
range from 400 to 4000 cm‐1 by Nicolet FTIR spectrophotometer (Nicolet Analytical Instruments,
Madison, WI, US). Ratios of the area under the peaks at wavenumber 1720cm‐1/1734cm‐1 were
recorded as the crystalline/amorphous ratios.
Gel permeation chromatography (GPC) was used to measure the Number‐average
molecular weight (Mn), Weight‐average molecular weight (Mw) and Polydispersity of the
degraded samples [219]. The analyses were performed on Waters gel permeation
chromatography system equipped with Waters 1515 Isocratic HPLC Pump, Waters 2707 auto‐
sampler with a 100 µL injection loop, a column heater and a Waters 2487 Dual Absorbance
Detector (λ=254 nm) in series with a Waters 2414 refractive index detector (Waters,
Massachusetts, US). Three Phenomenex phenogel columns installed in series (in order 104 Å, 103
Å and 50 Å pore size) and preceded by a Waters guard column, operating at 30°C and using
tetrahydrofuran as eluent at a flow rate of 1 mL/min. Sample injections comprised of 10 µL
solution (2.5mg/mL in THF and filtered through 0.22 µm PTFE filters) and molecular weights were
calculated relative to polystyrene standards.
5.3.6 In vitro hematoma model
Heparinized human whole blood sample was kindly provided by Red Cross Blood Bank
(Brisbane, Australia). The tests were performed following the approval from the Human Ethics
Page 119
Page 119
Committee of the Queensland University of Technology (Human Ethics Approval number
1400000023). Scaffolds were cut in 6 mm in diameter, pre‐weighed and sterilized by immersing
in 70% ethanol for 1 hour and dried under sterile condition. The blood sample was pre‐mixed
with CaCl2 (1080 mg/L, Unilab, Ajax Finechem, Australia) and thrombin (200 g/mL, Banksia
Scientific Company, Australia). Then, a volume of 100 L of pre‐mixed blood was added to each
2 mL Eppendorf tube containing the sample and incubated at 37oC for 2 hours to achieve fully
clotted blood [222]. Finally, samples were fixed in 4% glutaraldehyde overnight. The fixed samples
were dehydrated using a series of ethanol (50%, 70%, 90%, 100% for 1 hour each), and
hexamethyldisilazane (HMDS) for 1 hour, weighed and gold coated for SEM imaging.
Growth factor release from hematoma lysis: The influences of different scaffolds on the
secretion of platelet‐derived growth factor‐AB (PDGF‐AB) and transforming growth factor‐β1
(TGF‐β1) from the hematoma were evaluated by ELISA. The samples containing hematoma were
immersed in ACK Lysis Buffer for 24 hours at 37oC. Then the solutions were collected and
centrifuged at 1000g for 15 minutes. Afterward, quantitative measurements of PDGF‐AB and
TGF‐β1 in supernatants were performed according to the manufacturer’s instructions using ELISA
kits for PDGF‐AB (ab100623, Abcam®, Australia) and TGF‐β1 (ab100647, Abcam®, Australia).
5.3.7 Cell culture study
To explore the influence of the released growth factors from the clots on osteoblast’s
growth, a metabolic activity assay was performed. Newborn mouse calvaria‐derived MC3T3‐E1
subclone 14 pre‐osteoblasts were sourced from the Shanghai Cell Bank of the Chinese Academy
of Sciences. Cells were expanded in Dulbecco's Modified Eagle's medium (DMEM; Gibco®, Life
Technologies Pty Ltd., Australia) supplemented with 10% fetal bovine serum (FBS; In
Vitro Technologies, Australia) and 1% penicillin/streptomycin (50 U/ml and 50 μg/ml; P/S;
Gibco®). The cells were seeded in 24–well plate with a density of 2 × 104 cells/cm2. After 24
hours, the medium was replaced with the medium containing the blot clot exudate at a 9:1 ratio
(900 μL fresh medium and 100 μL exudate). Afterward, the metabolic activity of cells was
Page 120
Page 120
evaluated at day 1 and 3 by MTT [3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyl tetrazolium bromide]
assay (R&D Systems). In brief, the medium was removed, the cells were washed with PBS, then
500 μL of 0.5 mg/mL of MTT/DMEM solution was added into each well. After 4 hours, MTT‐
DMEM solution was carefully removed and 300 μL of dimethyl sulfoxide (DMSO, Univar USA) was
added to dissolve formazan crystals. The absorbance was then read at 570 nm in 100 μL in
triplicate using a microplate spectrophotometer [223].
5.3.8 Protein adsorption
The protein adsorption onto of mPCL scaffolds was evaluated using bovine serum
albumin (BSA) as a model protein according to published protocols [224]. Samples were cut into
specimens of 6 mm in diameter weighted and pre‐wetted by immersion in 100% ethanol for 1
hour; subsequently, the samples were immersed twice in a 1x PBS solution for 30 minutes under
vacuum condition and finally incubated in PBS at 37oC overnight. Each scaffold was placed in a 2‐
mL Eppendorf tube containing 200 µL of BSA solution (100 µg/mL in PBS) and incubated for 4
hours under shaking conditions at 37oC for protein adsorption. Scaffolds were then washed with
PBS to remove any non‐adsorbed proteins and placed in the tubes containing 400 µL of sodium
dodecyl sulfate 5% w/v for 2 hours under shaking conditions at 37oC to recover the adsorbed
proteins. The amounts of proteins were measured using Pierce™ BCA Protein Assay Kit (Thermo
Fisher, US) according to the manufacturer's protocol.
5.3.9 Cell proliferation
Mouse pre‐osteoblast cells were expanded as mentioned in section 2.7. Their suspension
was prepared at a density of 1x106 cell/mL and seeded on scaffolds with a density of 50000
cell/scaffold. The cell seeded scaffolds were then incubated for 2 hours and then 1 mL of culture
medium was added into each well. The cell‐seeded scaffolds were cultured overnight and
transferred into new plates to be ready for cell proliferation characterization by Alamar Blue assay
at days 1, 3 and 5. At each time point, the culture medium was removed, cells were washed with
PBS and incubated with Alamar Blue 10% v/v in full culture medium (500 µL/well) for 4 hours.
Page 121
Page 121
Their fluorescence intensity was measured at an excitation wavelength of 544 nm and an
emission wavelength of 590 nm using a POLARstar Optima plate‐reader (BMG Labtech,
Germany). The cells were rinsed with PBS and further cultured until the subsequent time point.
The measurements were reported as cell percentage normalized to the control (nmPCL) on day
1.
5.3.10 In vivo study
In vivo experiments were approved by the University Animal Ethics Committee (UAEC)
of Queensland University of Technology (Animal Ethics Approval number: 1600000821).
nmPCL/CaP and pmPCL44 scaffolds were sterilized by immersing in 70% ethanol for 30 minutes,
dried out under sterile condition and UV‐irradiated for 20 minutes. Twelve Sprague Dawley male
rats were sourced from the Animal Resources Centre, Canning Vale, WA, Australia. The calvarial
defects were created as described elsewhere [156, 225]. The animals were anesthetized with
isoflurane. The dorsal part of the cranium was shaved and disinfected with 50 mg/mL povidone‐
iodine (Betadine, Mundipharma BV, The Netherlands). A sagittal incision was performed through
the skin to expose the periosteum of the calvaria and the cranial vertex. After raising the full
thickness periosteal flap, the circular osseous defects were created on the calvaria by means of a
trephine bur (internal diameter of 5.0 mm) with copious isotonic solution (0.9% saline) irrigation.
An 8 mm mPCL membrane was placed on top of the exposed dura mater in order to prevent soft
tissue infiltration in the defect. A 5‐mm scaffold (nmPCL/CaP or pmPCL44) was placed into each
defect. A thin electrospun mPCL membrane was used to cover the defect, stabilize the scaffolds
and provide soft tissue occlusion. The wounds were closed in layers using resorbable sutures
(Vicryl 5.0 and 4.0, Ethicon, Germany). After 8 weeks, the animals were euthanized and the
implants were retrieved, fixed in 4% paraformaldehyde for 24 hours at 4°C and placed in PBS at
pH 7.4.
‐CT scanning was used to evaluate new bone formation within the defect site. Samples
were scanned at 55 kWp, 145 A, 8W and a voxel size of 30 µm using an ‐CT40 (Scanco Medical
Page 122
Page 122
AG, Switzerland). The 3D reconstructed images of defects were generated by using ‐CT system
software with a grayscale thread hold of 220.
After ‐CT scanning, samples were decalcified in 5% formic acid at RT for 1 month prior
to paraffin embedding. Samples were sectioned at a 5 m thickness, deparaffinized, and stained
with hematoxylin‐eosin (H&E). To quantify the levels of new bone formation and stromal tissue
areas, histomorphometric analyses were performed according to standard protocols using the
Osteomeasure Analysis System (Osteometrics, Atlanta, GA, USA) on three representative tissue
sections per animal, each was at least 10 sections apart from each other.
For Immunohistological analysis, anti‐Collagen I antibody (Col I, ab34710, Abcam, UK),
anti‐CD31 antibody (CD31, ab182981, Abcam, UK), anti‐Alkaline Phosphatase antibody (ALP,
ab108337, Abcam, UK), anti‐Human von Willebrand Factor (vWF, IR52761‐2, Agilent, US) anti‐
CD163 antibody (CD163, ab182422, Abcam, UK), anti‐CD68 antibody (CD68, ab125212, Abcam,
UK) and anti‐iNOS antibody (iNOS, ab15323, Abcam, UK) were used. The sections were
deparaffinized in xylene, rehydrated through a series of ethanol (100%, 90% and 70%). The slides
were then placed in antigen retrieval buffers (Tris‐sodium citrate pH 6 for Col I, ALP, CD163, CD68
and iNOS; Tris‐EDTA pH 9 for CD31 and vWF) and heat treated in decloaking chamber at 95oC for
5 minutes. Then, the slides were incubated with H2O2 for 5 minutes at RT, blocked with 2%
bovine serum albumin (Sigma‐Aldrich, US) at RT for 30 minutes and incubated with primary
antibodies for 2 hours at RT (ALP – 1:500 dilution, Col I – 1:100 dilution, vWF and CD163 – 1:100
dilution) or overnight at 4oC (CD31 – 1:500 dilution, CD68 – 1:100 dilution and iNOS – 1:100
dilution). Then, the slides were incubated with DAKO Envision+ Dual Link System HRP for 45
minutes. Lastly, the slides were incubated with DAKO liquid DAB+ substrate Chromogen for color
development.
5.3.11 Statistical analysis
Data are expressed as mean ± standard error of mean (s.e.m). All in vitro experiments
were performed with at least three replicates and repeated at least three times. The significant
Page 123
Page 123
differences between groups were analyzed by one‐way ANOVA test and Tukey post hoc test using
SigmaPlot 13.0 software. The significant difference was recorded when p<0.05.
5.4 RESULTS
5.4.1 Scaffold manufacturing
The viscoelasticity properties of the mPCL‐porogen composites were evaluated to
determine the impact of porogen concentrations on the ability of polymer melts to be extruded.
Amplitude sweep results (Supplement data, Figure S10D, F and H) showed that both mPCL and
composite materials maintained the linear viscoelasticity up to 0.1% shear strain. mPCL and
pmPCL17 had higher viscosity (with tan (δ) values around 7.5 to 8.5) compared with pmPCL44
(with tan (δ) values around 3). From 0 to 50% shear strain, while nmPCL maintained the linear
viscoelasticity property, the viscoelasticity properties of composite materials gradually increased
and finally reached to the same viscoelasticity state observed for nmPCL. Overall, the results
indicated that under high strain (more than 20%), the viscous properties of the composite
materials can match those of mPCL.
The effects of temperature on the rheological properties of the composite melts were
also studied (Supplement data, Figure S10E, G and I). The increase in porogen contents, a low
specific heat capacity material[226, 227], may have protected the mPCL from temperature
changes, provided some levels of thermal insulation and decreased the thermal sensitivity of
mPCL. All samples maintained their linear viscoelasticity from 25°C to 45oC. From 45°C to 130oC,
the rise in temperature increased the viscous modulus, and the viscoelasticity of all samples
increased differently (Supplement data, Figure S10E and G). While in both nmPCL and pmPCL17,
the δ‐value reached c.a. 8, pmPCL44 had δ‐value of c.a. 4 (Supplement data, Figure S10I).
Composite materials with a δ ‐value above 1 indicating a fluid‐like behavior and were printable.
Page 124
Page 124
5.4.2 Porogen leaching
The completion of porogen leaching was confirmed by SEM imaging, EDX analysis and
spectrophotometry. There was no visible porogen on the surfaces and in the cross sections of the
scaffolds as confirmed by SEM (Supplement data, Figure S11A). Only pmPCL17 group had
porogen residuals (less than 2%w/w) (Supplement data, Figure S11 B, C and D). Collectively, the
results showed that the leaching process was sufficient to remove the porogen from the scaffolds
with 44% w/w porogen (pmPCL44). For scaffolds with lower porogen concentration (i.e., 17%),
the amount of porogen was not high enough to allow for complete leaching.
5.4.3 Physicochemical properties
The scaffolds were designed to have macroscale porosity of 50% ‐ 60%, which was
confirmed based on the dimensions of the scaffolds after manufacturing. Regarding the
microscale porosity, SEM micrographs showed remarkable differences in pore size and geometry
on strut’s surface. (Figure 26A). Pores on the surface of nmPCL were identified as gaps between
PCL spherulites; whereas, pores on the surface of pmPCL scaffolds also included the smaller pores
from leaching of porogen near the surface. There were no significant differences in surface pore
areas among all groups. In the bimodal porous groups, pmPCL44 scaffolds were showed to have
bigger intra‐strut microscale pores and higher porosity when compared with pmPCL17 (Figure
26B). There were no differences in surface roughness among all groups (Figure 26C).
Page 125
Page 125
Figure 26. The increased porogen concentration resulted in the higher intra‐strut microscale
porosity, surface area and negatively charged group density. A and B: Representative SEM
micrographs of strut’s surface and cross‐section, respectively. C: 3D reconstructed AFM
micrographs of strut’s surface with their root‐mean‐square surface roughness (RMS). D:
Representative µ‐CT micrographs of scaffolds with their pore area. E: Intra‐strut porosity
(determined from µ‐CT scans), surface area (measured by gas adsorption), water contact angles
Page 126
Page 126
and surface charged group density (measured through TBO assay). RMS was measured from AFM
scans of 20x20‐µm2 areas on struts’ surface. Data are expressed as mean ± s.e.m; n=6. ** is
p<0.01 from one‐way ANOVA test and Tukey post hoc test. nmPCL, pmPCL17 and pmPCL44 are
nonporous mPCL scaffolds, bimodal porous mPCL scaffolds prepared from mixture having 17%
and 44% porogen, respectively. The scaffolds were extruded with the optimum flow rate and
temperature to achieve the “superflow” extrusion state, create smooth and defect‐free struts
[228]. As a result of the “superflow” state, only a small number of porogen was located at the
surface of the struts; therefore, the surface porosity was significantly lower than intra‐strut
porosity.
The intra‐strut microscale porosity (measured by the pore area) increased from
21.2±0.5% for pmPCL17 to 40.6±0.5% for pmPCL44 (mean ± s.e.m; n=6). The microscale pores
had a relatively uniform distribution (Figure 26D). The presence of circular pores in nmPCL as well
as in pmPCL17 and pmPCL44 was attributed to air entrapment during the extrusion of molten
polymer. The amount of porogen in the pmPCL17 was not sufficient to allow the porogen
particles to be completely leached out (appearing as bright dots in the µ‐CT images; Figure 26D).
The porosity of the experimental group was also calculated based on their mass, volume
and density. The intra‐strut porosity increased from 10.7±0.7% for pmPCL17 to 35.2±1.4% for
pmPCL44 (Figure 26E). Microscale pores also increased surface area approximately 3 folds from
718±47 cm2/g (nmPCL) to 2457±36 cm2/g (pmPCL44) (mean ± s.e.m; n=6, Figure 26E).
DSC analysis showed no significant differences in the crystallinity among all the groups
(61.9±1.5%, 59.5±1.2% and 60.8±1.0% (mean ± s.e.m; n=6) for nmPCL, pmPCL17 and pmPCL44,
respectively). Static contact angle measurement showed that all groups had hydrophilic surface
with contact angle of about 50o (Figure 26E). The increase in microscale porosity also led to
significant increase in the density of negatively charged groups, 3.6±0.2, 5.0±0.3 and 7.9±0.2
nmol/mg (mean ± s.e.m; n=6) for nmPCL, pmPCL17 and pmPCL44, respectively (Figure 26E).
Page 127
Page 127
5.4.4 Tensile strength
Young’s modulus and ultimate tensile strength (UTS) were influenced by the intra‐strut
microscale pores. Young modulus decreased from 357.5±31.6 to 276.2±30.1 and 261.6±23.8 MPa
for nmPCL, pmPCL17 and pmPCL44, respectively. The UTS decreased from 19.4±0.8 MPa (nmPCL)
to 14.2±1.1 MPa (pmPCL17) and 7.4±0.4 MPa (pmPCL44) (mean ± s.e.m; n= 6, p<0.01, Table 2).
Introduction of microscale porosity resulted in more flexible scaffolds, yet toughness decreased
compared with the nonporous scaffolds (Table 4).
Table 4. Ultimate tensile strength, Young's modulus, and elongation at break of single
struts with different porosity. Data are expressed as mean ± s.e.m; n=6. ** is p<0.01 from one‐
way ANOVA test and Tukey post hoc test. nmPCL, pmPCL17 and pmPCL44 are nonporous mPCL
scaffolds, bimodal porous mPCL scaffolds prepared from mixture having 17% and 44% porogen,
respectively.
nmPCL pmPCL17 pmPCL44
Ultimate tensile strength (MPa) 19.4±0.8** 14.2±1.1** 7.4±0.4**
Young’s modulus (MPa) 357.5±31.6 276.2±30.1 261.6±23.8
Elongation at break (%) >240** 24.4±0.2 36.0±8.9
5.4.5 Microscale porosity accelerated degradation of mPCL scaffolds
PCL is an aliphatic polyester, which has a fairly slow hydrolysis rate [221]. Therefore, it
was expected that microscale porosity would enhance the scaffold’s degradation process of
microscale porous scaffolds thanks to the increased surface area.
SEM micrographs (Figure 27A‐B) showed stark differences in the degradation among all
the groups. After 24 hours of degradation in the accelerated condition, in pmPCL44 group, both
strut’s surface and cross sections showed evidence of degradation (the surface was rougher,
surface pores were widened and more interconnected intra‐strut pores formed); whereas, in
nmPCL and pmPCL17 groups, only the surface showed clear evidence of degradation (rougher
Page 128
Page 128
surface). After 48 hours, pmPCL44 struts were thoroughly degraded and deformed; whereas, in
nmPCL group, only the surface were significantly degradaded (Figure 27A‐B).
Figure 27. Microscale porosity accelerated degradation of mPCL scaffolds. A and B:
Representative SEM micrographs from cross‐section (A) and surface (B) of nmPCL, pmPCL17 and
pmPCL44 at 0, 24 and 48 hours in the accelerated conditions. C, D and E: Mass loss percentage
(C), crystallinity measured from DSC (D) and molecular weight measured from GPC (E) of scaffolds
after 48 hours of degradation. nmPCL, pmPCL17 and pmPCL44 are nonporous mPCL scaffolds,
bimodal porous mPCL scaffolds prepared from mixture having 17% and 44% porogen,
respectively. All microscale porous scaffolds showed evidence of degradation after 24 hours both
Page 129
Page 129
on strut’s surface (rougher and enlarged pores) and throughout cross section (more small
interconnected pores). After 48 hours, nmPCL showed only surface degradation; whereas,
pmPCL44 degraded throughout the strut’s cross section. Data are expressed as mean ± s.e.m;
n=6. ** is p<0.01 from one‐way ANOVA test and Tukey post hoc test.
The mass loss percentage showed that the microscale porous scaffolds degraded faster
than the nonporous scaffolds (Figure 27C). After 48 hours, the mass loss percentage of 1.8±0.1%,
19.8±1.1% and 66.8±3.7% (mean ± s.e.m; n=6) were found in nmPCL, pmPCL17 and pmPCL44,
respectively.
The mass loss was proportional with the increased crystallinity (Figure 27C and D). The
mass loss of pmPCL44 increased sharply during 48 hours of degradation, and its crystallinity also
significantly increased. The continuous bulk erosion of pmPCL44 led to the continuous increase
in the crystallinity of the remaining of pmPCL44 samples. In the nmPCL group, there was no
significant change in the scaffold mass after 48 hours of degradation. The initial increase in
crystallinity of nmPCL could be because of the degradation and removal of the less crystalline
fraction on the surface. A small decrease in the crystallinity from 36 hours to 48 hours from 65%
to 62% could be caused by bulk erosion on surface and at the deepened pores, which degraded
and removed crystalline regions. The crystalline/amorphous ratios were calculated from FT‐IR
data and confirmed the changes in crystallinity of pmPCL44 scaffolds during the degradation
process (Supplementary data, Figure S12A).
To better understand the degradation process of the scaffolds, the molecular weights of
nmPCL and pmPCL44 were determined by GPC. The number average molecular weight (Mn) of
nmPCL scaffolds remained almost the same after 24h degradation (Figure 27E), which can be
explained as the degradation of nmPCL scaffolds occurred on the surface only and the
degradation products were removed by rinsing of the scaffolds. However, there were significant
decreases in Mn of all microscale porous groups (Figure 27E). The microscale porous groups had
Page 130
Page 130
a higher surface area compared with nmPCL scaffolds and had both bulk and surface erosion; as
a result, regions of both low and high Mn could be removed and long polymer chains could be
readily hydrolyzed into shorter chains, leaving the remaining scaffolds with decreased Mn.
In the remaining of the study, we used pmPCL44 scaffold and compared it with nmPCL
scaffold. Since we aimed to evaluate the scaffolds for bone regeneration, we also included a new
group which was nmPCL that had biomimetic calcium phosphate coating (CaP) as a reference
(nmPCL/CaP). The CaP coating on polycaprolactone scaffolds and on other materials have been
shown in the literature to significantly improve bone formation [210, 229, 230]. The surface areas
of scaffolds are shown in Table 5.
Table 5. Surface area of scaffolds measured by gas adsorption analysis. (data = mean ±
s.e.m; n=3). *** is p < 0.001. nmPCL, nmPCL/CaP and pmPCL44 are nonporous mPCL scaffolds,
calcium phosphate coated nmPCL scaffolds and bimodal porous mPCL scaffolds prepared from
mixture having 44% porogen, respectively.
Scaffold nmPCL nmPCL/CaP pmPCL44
Surface area (cm2/g) 718±47*** 3642±82*** 2457±36***
5.4.6 In vitro hematoma model
SEM analysis showed blood cells penetrated into the intra‐strut microscale pores in
pmPCL44 scaffolds; whereas, on nmPCL and nmPCL/CaP, the blood cells were remained only on
the surface (Figure 28A and B). Moreover, fibrin network were appeared to be denser on
pmPCL44 and nmPCL/CaP than on nmPCL. There was no significant difference in the weight of
hematomas which formed in all the groups (Figure 28C).
Page 131
Page 131
Figure 28. In vitro blood clotting experiments. A and B: SEM micrographs of cross‐section (A) and
surface (B) of hematoma encapsulated scaffolds. C: Wet weight of hematoma. D and E: TGF‐ß1
(D) and PDGF AB (E) release from hematoma after 24 hours incubation in the culture medium. F:
Cell proliferation results (MTT assay) of MC3T3‐E1 cultured with hematoma supernatant after 1
and 3 days. Data are expressed as mean ± s.e.m; n=6. ns is not significant, * is p<0.05 and ** is
p<0.01, from one‐way ANOVA test and Tukey post hoc test. nmPCL, nmPCL/CaP and pmPCL44
are nonporous mPCL scaffolds, calcium phosphate coated nmPCL scaffolds and bimodal porous
mPCL scaffolds prepared from mixture having 44% porogen, respectively.
The microscale porosity significantly influenced the growth factor releases in the
hematoma experiments. Both nmPCL/CaP and pmPCL44 increased the release of TGF–ß1 and
Page 132
Page 132
PDGF AB (Figure 28D and E). pmPCL44 significantly increased the release of TGF–ß1 compared
with nmPCL. There was also a statistically non‐significant increase in the release of PDGF AB in
pmPCL44 compared with nmPCL and nmPCL/CaP scaffolds (Figure 28E).
The conditioned media containing the released growth factors from hematoma were
used for cell culture experiments to investigate their bioactivity on osteoblast precursor cells,
MC3T3‐E1. Proliferation assay showed that supernatant from pmPCL44 significantly enhanced
osteoblast cell growth after 1 day (Figure 3F). Cell growth in pmPCL44 at day 3 was also higher
but not statistically significant than that in nmPCL/CaP (Figure 3F).
5.4.7 Multiphasic scaffolds enhanced protein adsorption in vitro
The interactions between scaffolds and proteins play important roles in cell behavior and
tissue regeneration [231]. In this study, albumin was used as a model protein to study protein
adsorption. pmPCL44 showed the highest amount of protein adsorption, followed by nmPCL/CaP
and then nmPCL (Figure 29A). The amount of protein adsorbed on pmPCL44 was 1.8±0.1
µg/scaffold; whereas, those of nmPCL/CaP and nmPCL were 1.3±0.1 µg/scaffold and 0.5±0.3
µg/scaffold, respectively.
Figure 29. Increased protein adsorption on pmPCL44 was observed compared with nmPCL and
nmPCL/CaP. Metabolic activities of MC3T3‐E1 cells on nmPCL, nmPCL/CaP and pmPCL44
scaffolds after 1, 3 and 5 days. Results were normalized to control scaffold (nmPCL) at day 1. Data
are expressed as mean ± s.e.m; n=6. ns is not significant, * is p<0.05 and ** is p<0.01 from one‐
Page 133
Page 133
way ANOVA test and Tukey post hoc test. nmPCL, nmPCL/CaP and pmPCL44 are nonporous mPCL
scaffolds, calcium phosphate coated nmPCL scaffolds and bimodal porous mPCL scaffolds
prepared from mixture having 44% porogen, respectively.
5.4.8 Multiphasic scaffolds had better support for cell proliferation in vitro
The ability of scaffolds to support cell attachment and proliferation was assessed by
seeding mouse pre‐osteoblast cells on scaffolds and performing Alamar Blue metabolic activity
assay at days 1, 3 and 5 to characterize the cell viability. pmPCL44 scaffolds significantly increased
the cell attachment compared with nmPCL and nmPCL/CaP, this was indicated by the increased
metabolic activity at day 1 (Figure 29B). After 3 and 5 days, the cells cultured on pmPCL44 still
had the highest metabolic activity compared with the cells on nmPCL and nmPCL/CaP.
5.4.9 In vivo study
From the results of the in vitro experiments, pmPCL44 and nmPCL/CaP were selected to
evaluate their regenerative potential in a rat critical size calvarial defect model. Both nmPCL/CaP
and pmPCL44 supported new bone formation into the scaffolds at similar levels (8.7±1.8 mm3
for nmPCL/CaP and 10.7±2.4 mm3 for pmPCL44, n=6; p=0.520). Figure 30B, left column showed
the position of scaffolds (labeled s) in the calvarial defects (black box sandwiched between the
skin‐side and dura‐side membranes (labeled sm and dm, respectively), and the formation of new
bone tissue (labeled nb) next to old bone (labeled ob) inside the nmPCL/CaP and pmPCL44
scaffolds. At high magnification (Figure 30B, middle column), significant new bone formation in
the periphery and center of the defect in both nmPCL/CaP and pmPCL44 was observed. Due to
the presence of intra‐strut microscale pores, pmPCL44 demonstrated additional support for new
bone formation, as infiltration of extracellular substances into the middle and at the boundary of
the struts were observed only in pmPCL44 group (Figure 30B, right column, black arrows).
Page 134
Page 134
Figure 30. Bimodal porous scaffolds had a similar osteoconductivity compared with calcium
phosphate coated scaffolds. A: µ‐CT‐based 3D reconstructed images of new bone formation at
calvarial defects and their bone volume. nmPCL/CaP and pmPCL44 are calcium phosphate coated
nonporous mPCL scaffolds and bimodal porous mPCL scaffolds prepared from mixture having
44% porogen, respectively. B: H&E staining at calvarial defects (boxed) of nmPCL/CaP and
pmPCL44 with arrows showed microtissue inside the pmPCL44 struts. ob: old bone; nb: new
bone; s: scaffold; sm: skin‐side membrane; and dm: dura‐side membrane; black arrow: infiltrated
extracellular materials. C: Histomorphometrical analysis of bone area percentage and stroma
Page 135
Page 135
tissue area percentage measured from scanned H&E stained slides using the Osteomeasure
Analysis System. Data are expressed as mean ± s.e.m; n=6. ns is not significant from one‐way
ANOVA test and Tukey post hoc test.
The histomorphometrical analysis was performed to quantify the levels of bone and
stroma tissue areas in the implantation site. The percentage of bone area in the defect showed a
slight increase in pmPCL44 group; whereas, the percentage of stroma tissue area in the
nmPCL/CaP was higher with no statistically significant difference (Figure 30C; n=6).
IHC staining was used to validate the presence of new bone formation in the defect area.
In both nmPCL/CaP and pmPCL44 groups, the IHC analyses confirmed bone‐like tissue formation
next to the struts (Figure 31). The new bone formation was confirmed by positive staining for ALP
and Col‐I markers (Figure 31). Both nmPCL/CaP and pmPCL44 groups demonstrated similar levels
of ALP and Col‐I expression. These data suggest that the stromal tissue was being remodelled in
the process of bone regeneration.
Page 136
Page 136
Figure 31. Validation of new bone formation via immunohistological staining with osteoblast key
markers, ALP (A) and Col I (B) at week 8 after scaffold implantation. Nonporous medical grade
polycaprolactone ( nmPCL/CaP) and pmPCL44 are calcium phosphate coated nonporous mPCL
scaffolds and bimodal porous mPCL scaffolds prepared from mixture having 44% porogen,
respectively. Both nmPCL/CaP and pmPCL44 groups demonstrated similar levels of ALP and Col I
expression. ob: old bone; nb: new bone; s: scaffold; sm: skin‐side membrane; and dm: dura‐side
membrane.
Page 137
Page 137
Furthermore, IHC was also used to confirm vascularization throughout the scaffolds at
the defects (Figure 32). The vascularization was observed at the interphase of scaffolds and the
membranes as well as in the stroma and new bone areas, which was confirmed by positive
staining for CD31 and vWF markers (key endothelial cell markers).
Figure 32. Confirmation of blood vessel formation via immunohistological staining for CD31 (A)
and vWF (B) at 8 weeks after scaffold implantation in rat critical size calvarial defect model.
Page 138
Page 138
nmPCL/CaP and pmPCL44 are calcium phosphate coated nonporous mPCL scaffolds and bimodal
porous mPCL scaffolds prepared from mixture having 44% porogen, respectively. The blood
vessel formation was confirmed by positive staining for CD31 and vWF as indicated with brown
color. ob: old bone; nb: new bone; s: scaffold; sm: skin‐side membrane; dm: dura‐side
membrane; and black arrows: blood vessel.
The effect of intra‐strut microscale porosity on inflammatory responses was evaluated
by immunostaining for macrophages (Figure 33). In both nmPCL/CaP and pmPCL44 groups, the
macrophages infiltrated into the macroscale pores of scaffolds, as indicated by the presence of
CD68+ cells. There was a similar level of macrophage infiltration into the defect areas in both
nmPCL/CaP and pmPCL44 scaffolds (Figure 33A and B). The macrophages mainly distributed in
the stroma and in the space between the membranes and the defects. The number of CD68+
cells (macrophages) in the bone area was limited. Further staining for M1 (iNOS+ cells; pro‐
inflammatory) and M2 (CD163+ cells; pro‐remodeling) macrophages showed there was
significantly higher M2 macrophages at the defect sites, indicating the bone regeneration process
were occurring [232, 233]. The presence of M1 macrophage in a small quantity compared with
M2 macrophage demonstrated low level of inflammation in the defect sites.
Page 139
Page 139
Figure 33. Immunohistological staining showed the presence of M1 macrophages (CD68+ and
iNOS+) much less than M2 macrophages (CD68+ and CD163+) within the defect areas in both
nmPCL/CaP (A) and pmPCL44 (B) groups. nmPCL/CaP and pmPCL44 are calcium phosphate
coated nonporous mPCL scaffolds and bimodal porous mPCL scaffolds prepared from mixture
having 44% porogen, respectively. ob: old bone; nb: new bone; s: scaffold; sm: skin‐side
membrane; and dm: dura‐side membrane. Black arrows demonstrate positive cells (brown
color).
5.5 DISCUSSION
Macroscale porosity is as a critical factor in the scaffold guided tissue formation to
achieve sufficient vascularization and tissue remodeling [24, 234, 235]. Moreover, the essential
role of microscale porosity in enhancing protein adsorption, hematoma formation, cellular
attachment and cell infiltration has been recently recognized [236‐238]. The microporous
environment supports the infusion of nutrients and oxygen and diffusion of waste products, cell
Page 140
Page 140
migration. It also increases surface area, which increases cell – material interaction [239]. The
increase in surface area from microporous structure also leads to more protein adsorption, which
increases bone‐related cell proliferation and calcium deposition [236, 240]; the capillary force
generated from that structure enhanced the migration of bone related cells [236]. However, the
contribution of macroscale and microscale porosity in a single scaffold on cellular behaviour and
tissue regeneration has not been well characterized. In this study, we used a combination of ABM
and porogen leaching techniques to obtain bimodal macro‐, micro‐ porous scaffolds.
Specifically, we extruded a mixture of mPCL and porogen (phosphate salt microparticles),
which were then leached out to create intra‐strut microscale pores. During the extrusion process,
the polymer melt flow instability occurred because of the difference in local stress concentration
between the material at the core and the material near the nozzle and the stick‐slip condition
when the extrudate existed from the nozzle, which resulted in surface distortions i.e. surface
roughness (non‐periodic distortion) and “sharkskin” effect – periodic small‐amplitude distortion
[228]. To eliminate the distortion, we increased the flow rate of the extrudate to achieve the
“superflow” state and matching the flow rate with the speed of the collecting board to achieve
defect‐free struts. When extruded, the scaffolds were immersed to leach out the porogen and
create microscale pores.
The increase in the surface area, contributed by the intra‐strut microscale porous
network, significantly accelerated the degradation of microscale porous scaffolds. While
conventional macro‐scale porous scaffolds (nmPCL) slowly degraded from the surface toward the
inner structure, microscale porous scaffolds were degraded throughout the thickness of the
struts. Because PCL has a relatively slow degradation (approximately 2‐4 years for complete
degradation in vivo depending on the molecular weight and implant volume [189]), the
introduction of intra‐strut microscale pores can be utilized to achieve a faster degradation [241‐
243].
Page 141
Page 141
It is well established that blood clotting (hematomas) is an important step triggering
tissue regeneration [203, 244, 245]. In the case of bone, the hematomas induce the infiltration of
inflammatory cells and release of key cytokines such as TGF–ß1 and PDGF–AB [244]. The release
of cytokines from the initial hematoma induces cell migration and regulates the tissue hemostasis
[246]. Subsequently, the released cytokines stimulate the proliferation and differentiation of
osteoblasts, induce matrix proteins synthesis and inhibit bone resorption by suppressing the
osteoclast precursor activities resorption [247]. In our study, scaffolds with dual porosity
(pmPCL44) showed a higher growth factor release from the hematoma, which was found to be
correlated with the increase in the proliferation of osteoblast progenitor cells. Our results are in
agreement with previous reports showing microscale porosity could enhance the entrapment
and release of growth factors, such as TGF–ß1 and rhBMP‐2 [248, 249]. The increase in growth
factor release in pmPCL44 despite the similarity in blood clot weights can be attributed to the
entrapment of blood cells within the microscale pores. In this study, only PDGF AB and TGF‐ ß1
were measured; however, the contribution of other growth factors on the results needs to be
investigated [250, 251].
The increase in surface roughness generated from different fabrication processes
increases surface area for protein – material and cell – material interactions. The difference in
surface roughness (pattern, level of roughness) can lead to the enhancement of different types
of cells [240, 252]. In this study, regarding protein – polymer interaction, albumin was used as a
protein model. The adsorption mechanism between protein and polymer was thanks the
wettability of the polymer and non‐specific to any particular protein; therefore, the level of
albumin adsorption could represent the level of proteins, which have similar hydrophilicity to
albumin, adsorb to the scaffolds. It was proven that level of albumin adsorbed on the scaffolds
was directly proportional to the level of mouse‐calvaria‐derived pre‐osteoblastic cells
proliferation and calcium deposition [240]. The fact that protein adsorbed more on pmPCL44
than calcium phosphate coated scaffolds (nmPCL/CaP) despite the latter had higher surface area
Page 142
Page 142
can be attributed to the intra‐strut microscale pore architecture in pmPCL44 group which could
entrap more protein solution through capillary actions. In fact, microscale pores and their
associated capillary forces have been found to result in higher cell adhesion, attachment strength,
proteins adsorption and cell penetration [231, 236, 253].
We further implanted pmPCL44 scaffolds in a rat critical size calvarial defect to
investigate the bone regeneration potential in vivo. As a reference group, we used biomimetic
CaP‐ coated mPCL scaffold. Previous work from our group and others demonstrated that the CaP
coatings significantly increased the tissue mineralization and new bone formation [210, 225].
The rat critical size calvarial defect model has been successfully used by our group and
others to specifically investigate the new bone formation from native bone toward the scaffolds
[254‐258]. In our study, we further utilized the occlusive electro‐spun membranes (minimum
pore size of less than 1 µm) [259] [260] at the top and bottom of the defects (containing the
scaffolds) to minimize penetration of cells from the dura and skin and maintain space for bone
formation.
After 8 weeks of implantation, the μ‐CT results showed ossification occurred from the
defect’s edge toward the center of the defect suggesting the new bone tissues were originated
from the host bone and not from mesenchymal precursor cells present in the dura. This result
illustrate the use of occlusive membranes to highly control and evaluate osteoconductivity of the
scaffolds in skull defect model. Both nmPCL/CaP and pmPCL44 had similar osteoconductivity,
which was demonstrated by a similar level of bone formation. IHC staining showed a low level of
macrophages in both groups. M1 macrophages, which are associated with inflammation, were
found to be in much lower numbers than M2 macrophages in the defect area. M2 macrophages
are an indicator that regeneration was occurring in the defect region [232, 233]. Additionally, H&E
staining revealed extracellular matrix (ECM) materials infiltrated into the microscale pores
supposedly through the capillary forces from the pores. We suggest that this ECM could
Page 143
Page 143
contribute to tissue formation by the entrapment and subsequent release of signaling molecules
[261].
The manufacturing parameters in this study, such as the porogen size, concentration and
strut diameter, were chosen based on our pilot trials to achieve maximum porosity while the
polymer‐porogen mixture can still be extruded. In this study, we used solvent to dissolve mPCL
and mixed in porogen particles. We are moving toward using melt‐mixing methods to completely
avoid toxic solvents in the preparation of the composite material for being used on the ABM
platform. It is envisioned that scaffold macroscale pore designs (e.g., pore size and pore
architecture) and microscale pore sizes could be further optimized to enhance in vivo hematoma
formation and stabilization, blood vessel penetration, and ultimately in vivo tissue formation and
remodelling.
5.6 CONCLUSION
This study aimed to investigate the application of 3D printed scaffolds with dual
macro/microscale porosity for bone tissue regeneration. The combination of screw‐driven melt
extrusion with porogen leaching technique was employed to manufacture mPCL scaffolds with
dual macro‐, microscale porosity. These new scaffolds showed higher surface area and surface
negatively charged groups compared with scaffolds having only macroscale porosity. The
scaffolds also enhanced blood clot formation and cell attachment in vitro. In vivo experiments
showed pmPCL44 scaffolds had a similar level of osteoconductivity and bone formation
compared with the scaffolds having biomimetic calcium phosphate coating.
Page 144
Page 144
6Chapter 6: Conclusion and future directions
6.1 CONCLUSION
This thesis presents the fabrication and characterisation of a new type of 3D‐printed
scaffold with bimodal porosity to be used as a dual‐function device for tissue reconstruction and
local drug delivery. The main conclusions are summarized below.
At the outset, the method of fabricating bimodal porous scaffolds was introduced in
Chapter 2. The scaffolds were prepared by using an SME system to build the scaffolds with a
designed lay‐down pattern. Porogen was leached from the extruded scaffolds to create
microscale pores. The microscale pores increased the drug loading capacity by providing more
surface area for drug attachment. The scaffolds demonstrated a reduced burst release and
prolonged release profile of cefazolin when combined with GelMA coating compared with
GelMA‐coated solid‐strut scaffolds. Last, the antimicrobial property of cefazolin‐loaded scaffolds
was also demonstrated in vitro.
In Chapter 3, the application of bimodal porous scaffolds for drug delivery was further
characterized. The ionic bonding interaction of DOX was previously shown by Jassal et al. [217]
and Niza et al. [262]. The slow diffusion rate of PTX due to its hydrophobic property was shown
by Innocente et al. [263] and Zhu et al. [264]. The fast diffusion rate of highly hydrophilic CEF
resulting fast release was shown by Mutsuzaki et al. [91] and Rath et al. [265]. Positively charged
hydrophilic drug (DOX) had more interaction with pmPCL scaffolds, resulting in a reduced burst
release and sustained release profile. Non‐charged hydrophobic drug (PTX) also showed a
sustained release profile because of the hydrophobic property of PTX. For non‐charged
hydrophilic drug (CEF), the burst release was reduced compared with the solid‐strut scaffolds;
however, it was still high because there was no electrostatic interaction between drug and
Page 145
Page 145
pmPCL, and the drug was easy to dissolve in water. In addition, drug‐loaded bimodal porous
scaffolds demonstrated in vitro that the drug retained its bioactivities (antimicrobial and anti‐
cancer properties).
In Chapter 4, the bimodal porous scaffolds were fabricated and loaded with DOX to
demonstrate, as a proof‐of‐concept, their potential application in breast cancer treatment. The
DOX‐loaded scaffolds demonstrated their dose‐dependent chemotherapeutic effects in vitro
against MDA‐MB‐231 cells with the dose range of 0.4–10 μg DOX/mg pmPCL scaffold. The
implanted DOX‐loaded scaffolds, which used 5% and 20% of the I.V. injection dose, demonstrated
reduced cytotoxicity and enhanced chemotherapeutic effects on local recurrent tumour
compared with I.V. injected DOX. However, further dose optimisation needs to be done to
balance the chemotherapeutic effect and cytotoxicity.
Last, in Chapter 5, the application of the bimodal porous scaffolds for tissue regeneration
was demonstrated. The scaffolds demonstrated their biocompatibility by enhanced protein
adsorption, cell attachment and proliferation compared with solid‐strut scaffolds and CaP‐coated
scaffolds. Finally, their osteoconductivity was demonstrated in vivo with a similar level of new
bone formation to CaP‐coated scaffolds.
All in all, this thesis introduced and proposed a new type of scaffold with distinct porous
architecture as a dual‐function scaffold for tissue reconstruction and local drug delivery to
enhance the treatment efficiency and reduce the complications and side effects of conventional
treatment. The findings in this thesis can serve as a proof‐of‐concept and a foundation for future
research regarding dual‐function devices for tissue engineering and drug delivery applications.
6.2 LIMITATIONS AND FUTURE DIRECTIONS
This thesis serves as a proof‐of‐concept for bimodal porous scaffolds to be used as dual‐function
scaffolds. The scaffolds were developed as a viable product and to validate its feasibility in a broad
application. As such, further optimization and deeply characterization need to be done for
Page 146
Page 146
speficic purpose. Therefore, further investigation needs to be done before achieving the final
goal.
This thesis demonstrated a possibility of improving the commonly used 3D PCL scaffolds
in tissue engineering field to become dual‐functional scaffolds that can be used for different
applications. Due to the limited timeframe of a PhD, certain in‐dept characterization and
optimization experiments have been excluded to prioritize for the demonstration of a broad
application of this scaffold. Herein, we would like to discuss about limitations and provide
recommendations for further study following this project.
6.2.1 Chapter 2
Firstly, in this study, the porogen used was sieved from PBS tablets to a specific size range.
This can be improved by using commercialized water‐soluble porogen with specific size classes
and shapes to evaluate how the pore size and shape affect the release profile. Furthermore, after
preparing the porogen‐polymer composite for printing, the rheological measurement was only
used to validate the printability of limited conditions. It is suggested that the temperature, air
pressure and other contribution factors should be considered and well characterized for different
porogen‐polymer ratio and printing condition (different nozzle size and printing pattern). The
porogen‐polymer ratio can affect the viscosity of the composite while the nozzle size and printing
pattern may determine the cooling time of the printed strut and determine if there was enough
cooling time between different layers. In the next step of drug loading method, the drop loading
combined with gelMA coating showed very low loading efficiently, which mainly due to the drug
lost during the coating process. This can be improved by using other method to create the coating
barrier for the scaffolds after drug loading, such as using electrospinning to fabricate thin layer of
polymers, covering the scaffolds and hence reduce the release of the drug. Lastly, for the drug
release study, the drug release should be modelized to understand the release kinetic of the
applied drugs
Page 147
Page 147
6.2.2 Chapter 3
Secondly, in this study, the physicochemical and mechanical properties characterization and the
degradation test were performed; however, the long‐term stability e.g. properties of the
scaffolds after degraded or after long period of time were not characterized. Beside long‐term
stability, the influence of porogen, fabrication process, leaching and drug loading on
crystallization, crystallinity and mechanical properties should be tested. In this experiment
design, the polymer used was pmPCL, which has a charged surface and interacts with charged
molecules, thereby altering the release profile. Other polymers can be used to characterize and
achieve a specific need of interacting with the used drug and achieve the desired release profile.
Furthermore, the strength of interaction between drug and polymers should be quantified for
comparison.
6.2.3 Chapter 4
Thirdly, in application of local delivery of chemotherapeutic drug to prevent breast
cancer recurrence and metastasis, the chemotherapeutic effect of HD group was more effective
than LD group in in vitro setting but slightly lower in in vivo setting. We suggested to repeat the
experiment with better selection of animal used right before stating the treatment as well as
optimize the doses to balance between the chemotherapeutic effect and the systemic
cytotoxicity. In this study, we used immunocompromised animals. It would be beneficial in our
future study to use immunocompetent animals. The characterization of tissue regeneration after
transplantation and the used of combined agents loaded scaffolds are of interest for our future
study.
6.2.4 Chapter 5
Lastly, in application of bone reconstruction, the study showed that the bimodal porous
scaffolds could regenerate bone in the calvaria defect model as good as CaP coating scaffolds.
Those the effect of osteoinduction was claimed for the intrastrut microscale porous structure,
Page 148
Page 148
the mechanism of why such structure can enhance new bone formation should be the focus of
the next study.
Page 149
Page 149
Appendices
Figure S1. Rheology testing of mPCL and mPCL‐porogen for printing process showing amplitude
sweep tests (a1‐a3) and temperature sweep (b1‐b3). For amplitude tests, temperature was kept
constant at 110°C and angular frequency at 10 rad/sec. Each data point is the average of three
experimental repeat (N=3); the error bars correspond to SD. For temperature sweep the
temperature was varied from 125°C to 35°C. Strain was kept constant at 1% and angular
frequency at 10 rad/sec. Each point is the average of three experimental repeat (N=3); the error
bars correspond to SD.
Page 150
Page 150
Figure S2. Microporous mPCL scaffolds – after compression – relaxation cycles in mechanical
testing
Figure S3. mPCL scaffolds – after compression – relaxation cycles in mechanical testing
Page 151
Page 151
Figure S4. Morphology of nMPCL and pMPCL. (A): Representative bright field microscopic images
of nMPCL and pMPCL. (B): AFM micrographs of surface of struts on nMPCL and pMPCL scaffolds
and their root‐mean‐square surface roughness (RMS, data are expressed as mean ± s.e.m, n=6).
(C) and (D): SEM and microCT images showing surface pores (short arrows) and connection to the
inside pores through interconnectivity (long arrows). (E): Intra‐strut micropore size class
distribution on surface and cross‐section of nMPCL and pMPCL. Data are expressed as mean ±
s.e.m, n=6.
Page 152
Page 152
Figure S5. Schematic of interaction between DOX and MPCL molecules.
Page 153
Page 153
Figure S6. Drug loading efficiency of DOX, PTX and CEF on nPCL and pPCL. Data are expressed as
mean ± s.e.m, n=6. ** is p<0.01 from one‐way ANOVA test and Tukey post hoc test.
Figure S7. Pore size distribution of bimodal porous medical grade polycaprolactone scaffolds.
data were expressed as mean ± SEM, n = 4.
Page 154
Page 154
Figure S8. In vivo IVIS images of the control (A, bimodal porous medical grade polycaprolactone
(pmPCL) scaffolds without drug), I.V. (B, I.V. injection of doxorubicin (DOX) at 2 mg/kg), low dose
and high dose (C and D, DOX–loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups.
Page 155
Page 155
Figure S9. Doxorubicin (DOX) –loaded bimodal porous medical grade polycaprolactone (pmPCL)
scaffolds effectively reduced the breast tumor metastasis in liver and spleen. A and B:
Morphology, organ weight normalized to body weight and total flux from ex vivo BLI imaging of
liver (A) and spleen (B) of the Control (bimodal porous medical grade polycaprolactone (pmPCL)
scaffolds without drug), I.V. (I.V. injection of doxorubicin (DOX) at 2 mg/kg), Low Dose and High
Dose (DOX–loaded pmPCL with 2 and 8 µg DOX/scaffold, respectively) groups. The organs were
weighed after fixation and normalized to the body weight of the animal at day 28. Data are
expressed as mean ± SEM (n=6 mice in each experimental group). ns: data is not statistically
significant; * is p<0.05; ** is p<0.01 from one‐way ANOVA and Tukey post hoc test.
Page 156
Page 156
Table S1. List of primary antibodies used for immunohistochemically staining and their antigen
retrieval conditions.
Catalogue
number
Company Antigen retrieval
solution
Antigen retrieval
condition
Dilution
Anti‐NuMA
antibody
ab97585 Abcam,
UK
Tris‐sodium citrate
pH 6
95oC for 5 minutes 1:200
Anti‐Ki67
antibody
ab15580 Abcam,
UK
Tris‐sodium citrate
pH 6
90oC for 20 minutes 1:200
Page 157
Page 157
Figure S10. Porogen size class and viscoelasticity of MPCL and MPCL‐porogen films. A and B:
Representative size class distribution by volume acquired by light scattering. C: SEM micrograph
showing the morphology of PBS porogen. D, F and H: Storage modulus (G'), loss modulus (G'')
and loss tangent result from amplitude sweep rheological measurement of PCL and porogen‐PCL
composites at a temperature of 110°C and an angular frequency of 10 rad/sec. E, G and I: Storage
modulus (G'), loss modulus (G'') and loss tangent (tan δ) results from temperature sweep
rheological measurements at a strain of 1% and an angular frequency of 10 rad/sec.
Page 158
Page 158
Figure S11. Validation of complete leaching of porogen. A: SEM micrographs of pMPCL44 before
(left) and after (right) leaching. B: EDX analysis results of pPCL44 scaffolds after 0, 7 and 14 days
leached in NaOH 0.01M in shaking incubator. C: EDX analysis results of all PCL scaffolds after
leaching. D: UV‐Vis absorbance results of remained porogen percentage after leaching. Data are
expressed as mean ± s.e.m; n=6. *: p<0.05 and **: p<0.01 from One‐way ANOVA test and Tukey
post‐hoc test.
Figure S12. Degradation of pMPCL compared with nMPCL. The crystalline/amorphous ratios of
nPCL and pPCL44 scaffolds after 12, 24, 36 and 48 hours of degradation (A). GPC determination
of molecular weight and polydispersity of nPCL and pPCL44 scaffolds after 48 hours of accelerated
degradation (B and C). Data are expressed as mean ± s.e.m; n=6. ** is p<0.01 from one‐way
ANOVA test and Tukey post hoc test.
Page 159
Page 159
Figure S13. Representative SEM micrographs at low (A) and high (B) magnification showing
surface morphology of nPCL/CaP scaffold.
Page 160
Page 160
Bibliography
1. Ratheesh, G., et al., 3D Fabrication of Polymeric Scaffolds for Regenerative Therapy. ACS Biomaterials Science & Engineering, 2017. 3(7): p. 1175‐1194.
2. Hutmacher, D.W., Scaffolds in tissue engineering bone and cartilage. Biomaterials, 2000. 21: p. 2529‐43.
3. Chhaya, M.P., et al., Sustained regeneration of high‐volume adipose tissue for breast reconstruction using computer aided design and biomanufacturing. Biomaterials, 2015. 52: p. 551‐60.
4. Chhaya, M.P., et al., Transformation of Breast Reconstruction via Additive Biomanufacturing. Sci Rep, 2016. 6: p. 28030.
5. Song, Z., et al., Prosthesis infections after orthopedic joint replacement: the possible role of bacterial biofilms. Orthopedic reviews, 2013. 5: p. 65‐71.
6. Lamagni, T., Epidemiology and burden of prosthetic joint infections. Journal of Antimicrobial Chemotherapy, 2014. 69: p. 5‐10.
7. Phillips, B.T., et al., A systematic review of antibiotic use and infection in breast reconstruction: what is the evidence? Plastic and reconstructive surgery, 2013. 131: p. 1‐13.
8. Feldman, E.M., et al., Breast implant infections: is cefazolin enough? Plastic and reconstructive surgery, 2010. 126: p. 779‐785.
9. Huiras, P., et al., Local antimicrobial administration for prophylaxis of surgical site infections. Pharmacotherapy:The Journal of Human Pharmacology & Drug Therapy, 2012. 32: p. 1006‐1019.
10. Marshall, B.M., D.J. Ochieng, and S.B. Levy, Commensals: Underappreciated Reservoir of Antibiotic Resistance. Microbe, 2009. 4: p. 231‐238.
11. Colleoni, M., et al., Annual Hazard Rates of Recurrence for Breast Cancer During 24 Years of Follow‐Up: Results From the International Breast Cancer Study Group Trials I to V. J Clin Oncol, 2016. 34(9): p. 927‐35.
12. Cipriano, C., et al., Developing an Evidence‐based Followup Schedule for Bone Sarcomas Based on Local Recurrence and Metastatic Progression. Clin Orthop Relat Res, 2017. 475(3): p. 830‐838.
13. Harasym, T.O., P.R. Cullis, and M.B. Bally, Intratumor distribution of doxorubicin following i.v. administration of drug encapsulated in egg phosphatidylcholine/cholesterol liposomes. Cancer Chemother Pharmacol, 1997. 40(4): p. 309‐17.
14. Shim, J.‐H., et al., Three‐dimensional printing of antibiotics‐loaded poly‐ε‐caprolactone/poly(lactic‐co‐glycolic acid) scaffolds for treatment of chronic osteomyelitis. Tissue Engineering and Regenerative Medicine, 2015. 12(5): p. 283‐293.
15. Repka, M.A., et al., Applications of hot‐melt extrusion for drug delivery. Expert Opin Drug Deliv, 2008. 5(12): p. 1357‐76.
16. Traub, W.H. and B. Leonhard, Heat stability of the antimicrobial activity of sixty‐two antibacterial agents. J Antimicrob Chemother, 1995. 35(1): p. 149‐54.
17. Anselmo, A.C. and S. Mitragotri, An overview of clinical and commercial impact of drug delivery systems. J Control Release, 2014. 190: p. 15‐28.
Page 161
Page 161
18. Moul, J.W. and K. Civitelli, Managing advanced prostate cancer with Viadur (leuprolide acetate implant). Urologic nursing, 2001. 21(6): p. 385.
19. Jonat, W., et al., Goserelin versus cyclophosphamide, methotrexate, and fluorouracil as adjuvant therapy in premenopausal patients with node‐positive breast cancer: The Zoladex Early Breast Cancer Research Association Study. J Clin Oncol, 2002. 20(24): p. 4628‐35.
20. Rock, J.A., et al., Zoladex (Goserelin Acetate Implant) in the Treatment of Endometriosis: A Randomized Comparison With Danazol. Obstetrics & Gynecology, 1993. 82(2): p. 198‐205.
21. Samavedi, S., A.R. Whittington, and A.S. Goldstein, Calcium phosphate ceramics in bone tissue engineering: a review of properties and their influence on cell behavior. Acta Biomater, 2013. 9(9): p. 8037‐45.
22. Wei, J., et al., Hierarchically microporous/macroporous scaffold of magnesium‐calcium phosphate for bone tissue regeneration. Biomaterials, 2010. 31(6): p. 1260‐9.
23. Zhang, J., et al., 3D‐printed magnetic Fe3O4/MBG/PCL composite scaffolds with multifunctionality of bone regeneration, local anticancer drug delivery and hyperthermia. J. Mater. Chem. B, 2014. 2(43): p. 7583‐7595.
24. Liu, X. and P.X. Ma, Polymeric scaffolds for bone tissue engineering. Annals of biomedical engineering, 2004. 32(3): p. 477‐486.
25. Schell, H., et al., The haematoma and its role in bone healing. J Exp Orthop, 2017. 4(1): p. 5.
26. Shiu, H.T., P.C. Leung, and C.H. Ko, The roles of cellular and molecular components of a hematoma at early stage of bone healing. J Tissue Eng Regen Med, 2018. 12(4): p. e1911‐e1925.
27. Pajarinen, J., et al., Mesenchymal stem cell‐macrophage crosstalk and bone healing. Biomaterials, 2019. 196: p. 80‐89.
28. Freeman, F.E., et al., Biofabrication of multiscale bone extracellular matrix scaffolds for bone tissue engineering. Eur Cell Mater, 2019. 38: p. 168‐187.
29. Gupta, D., et al., Multiscale Porosity in Compressible Cryogenically 3D Printed Gels for Bone Tissue Engineering. ACS Appl Mater Interfaces, 2019. 11(22): p. 20437‐20452.
30. Rustom, L.E., et al., Multiscale Porosity Directs Bone Regeneration in Biphasic Calcium Phosphate Scaffolds. ACS Biomaterials Science & Engineering, 2016. 3(11): p. 2768‐2778.
31. Zhang, J., et al., RhBMP‐2‐loaded calcium silicate/calcium phosphate cement scaffold with hierarchically porous structure for enhanced bone tissue regeneration. Biomaterials, 2013. 34(37): p. 9381‐9392.
32. Salerno, A., et al., Engineered mu‐bimodal poly(epsilon‐caprolactone) porous scaffold for enhanced hMSC colonization and proliferation. Acta Biomater, 2009. 5(4): p. 1082‐93.
33. Akkouch, A., Z. Zhang, and M. Rouabhia, A novel collagen/hydroxyapatite/poly(lactide‐co‐ε‐caprolactone) biodegradable and bioactive 3D porous scaffold for bone regeneration. Journal of Biomedical Materials Research Part A, 2011. 96A(4): p. 693‐704.
34. Yan, L.‐P., et al., De novo bone formation on macro/microporous silk and silk/nano‐sized calcium phosphate scaffolds. Journal of Bioactive and Compatible Polymers, 2013. 28(5): p. 439‐452.
35. Reignier, J. and M.A. Huneault, Preparation of interconnected poly(ε‐caprolactone) porous scaffolds by a combination of polymer and salt particulate leaching. Polymer, 2006. 47(13): p. 4703‐4717.
Page 162
Page 162
36. Ahn, M.‐K., et al., Design and Production of Continuously Gradient Macro/Microporous Calcium Phosphate (CaP) Scaffolds Using Ceramic/Camphene‐Based 3D Extrusion. Materials, 2017. 10(7): p. 719.
37. Kim, J.W., et al., Production of Poly(epsilon‐Caprolactone)/Hydroxyapatite Composite Scaffolds with a Tailored Macro/Micro‐Porous Structure, High Mechanical Properties, and Excellent Bioactivity. Materials (Basel), 2017. 10(10).
38. Maleksaeedi, S., et al., Toward 3D Printed Bioactive Titanium Scaffolds with Bimodal Pore Size Distribution for Bone Ingrowth. Procedia CIRP, 2013. 5: p. 158‐163.
39. Martins, A., R.L. Reis, and N.M. Neves, Micro/Nano Scaffolds for Osteochondral Tissue Engineering. Adv Exp Med Biol, 2018. 1058: p. 125‐139.
40. Zhang, B., et al., Porous bioceramics produced by inkjet 3D printing: Effect of printing ink formulation on the ceramic macro and micro porous architectures control. Composites Part B: Engineering, 2018. 155: p. 112‐121.
41. Dang, H.P., et al., 3D printed dual macro‐,micro‐porous network as a tissue engineering scaffold with drug delivery function. Biofabrication, 2019.
42. Liaw, C.Y. and M. Guvendiren, Current and emerging applications of 3D printing in medicine. Biofabrication, 2017. 9(2): p. 024102.
43. Habraken, W.J., J.G. Wolke, and J.A. Jansen, Ceramic composites as matrices and scaffolds for drug delivery in tissue engineering. Adv Drug Deliv Rev, 2007. 59(4‐5): p. 234‐48.
44. Palama, I.E., et al., Therapeutic PCL scaffold for reparation of resected osteosarcoma defect. Sci Rep, 2017. 7(1): p. 12672.
45. Feng, B., et al., The effect of pore size on tissue ingrowth and neovascularization in porous bioceramics of controlled architecture in vivo. Biomed Mater, 2011. 6(1): p. 015007.
46. Liu, X. and P.X. Ma, Polymeric scaffolds for bone tissue engineering. Ann Biomed Eng, 2004. 32(3): p. 477‐86.
47. Cipitria, A., et al., Porous scaffold architecture guides tissue formation. J Bone Miner Res, 2012. 27(6): p. 1275‐88.
48. Hutmacher, D.W., et al., State of the art and future directions of scaffold‐based bone engineering from a biomaterials perspective. J Tissue Eng Regen Med, 2007. 1(4): p. 245‐60.
49. Chan, B.P. and K.W. Leong, Scaffolding in tissue engineering: general approaches and tissue‐specific considerations. Eur Spine J, 2008. 17 Suppl 4: p. 467‐79.
50. Chiu, Y.C., et al., The role of pore size on vascularization and tissue remodeling in PEG hydrogels. Biomaterials, 2011. 32(26): p. 6045‐51.
51. Jones, J.R., Scaffolds for tissue engineering, in Biomaterials, artificial organs and tissue engineering. 2005, Woodhead Publishing. p. 201‐214.
52. Sosnowski, S., P. Wozniak, and M. Lewandowska‐Szumiel, Polyester scaffolds with bimodal pore size distribution for tissue engineering. Macromol Biosci, 2006. 6(6): p. 425‐34.
53. Du, L., et al., Hierarchical macro/micro‐porous silk fibroin scaffolds for tissue engineering. Materials Letters, 2019. 236: p. 1‐4.
54. Duan, B., et al., Microporous density‐mediated response of MSCs on 3D trimodal macro/micro/nano‐porous scaffolds via fibronectin/integrin and FAK/MAPK signaling pathways. Journal of Materials Chemistry B, 2017. 5(19): p. 3586‐3599.
55. Reed, S., et al., Macro‐ and micro‐designed chitosan‐alginate scaffold architecture by three‐dimensional printing and directional freezing. Biofabrication, 2016. 8(1): p. 015003.
56. Zhu, M., et al., 3D‐printed hierarchical scaffold for localized isoniazid/rifampin drug delivery and osteoarticular tuberculosis therapy. Acta Biomater, 2015. 16: p. 145‐55.
Page 163
Page 163
57. Vorndran, E., et al., Simultaneous Immobilization of Bioactives During 3D Powder Printing of Bioceramic Drug‐Release Matrices. Advanced Functional Materials, 2010. 20(10): p. 1585‐1591.
58. Martinez‐Vazquez, F.J., et al., Fabrication of novel Si‐doped hydroxyapatite/gelatine scaffolds by rapid prototyping for drug delivery and bone regeneration. Acta Biomater, 2015. 15: p. 200‐9.
59. Water, J.J., et al., Three‐Dimensional Printing of Drug‐Eluting Implants: Preparation of an Antimicrobial Polylactide Feedstock Material. Journal of pharmaceutical sciences, 2015. 104(3): p. 1099‐1107.
60. Boetker, J., et al., Modifying release characteristics from 3D printed drug‐eluting products. Eur J Pharm Sci, 2016. 90: p. 47‐52.
61. Zimmerli, W., Prosthetic joint associated infections. Best Practice and Research: Clinical Rheumatology, 2006. 20: p. 1045‐1063.
62. Tran, N., et al., Silver doped titanium oxide–PDMS hybrid coating inhibits Staphylococcus aureus and Staphylococcus epidermidis growth on PEEK. Materials Science and Engineering: C, 2015. 49: p. 201‐209.
63. Tran, P.A., D.M. Hocking, and A.J. O'Connor, In situ formation of antimicrobial silver nanoparticles and the impregnation of hydrophobic polycaprolactone matrix for antimicrobial medical device applications. Materials Science and Engineering: C, 2015. 47: p. 63‐69.
64. Tran, P.A. and T.J. Webster, Antimicrobial selenium nanoparticle coatings on polymeric medical devices. Nanotechnology, 2013. 24(15): p. 155101.
65. van Heerden, J., et al., Antimicrobial coating agents: can biofilm formation on a breast implant be prevented? Journal of Plastic, Reconstructive & Aesthetic Surgery, 2009. 62(5): p. 610‐617.
66. Tran, N. and P.A. Tran, Nanomaterial‐based treatments for medical device‐associated infections. ChemPhysChem, 2012. 13(10): p. 2481‐2494.
67. Holländer, J., et al., Three‐dimensional printed PCL‐based implantable prototypes of medical devices for controlled drug delivery. Journal of pharmaceutical sciences, 2016. 105(9): p. 2665‐2676.
68. Malinauskas, M., et al., 3D microporous scaffolds manufactured via combination of fused filament fabrication and direct laser writing ablation. Micromachines, 2014. 5(4): p. 839‐858.
69. Lam, C.X.F., et al., Scaffold development using 3D printing with a starch‐based polymer. Materials Science and Engineering: C, 2002. 20(1): p. 49‐56.
70. Zhang, J., et al., Coupling 3D printing with hot‐melt extrusion to produce controlled‐release tablets. Int J Pharm, 2017. 519(1‐2): p. 186‐197.
71. Traub, W.H. and B. Leonhard, Heat stability of the antimicrobial activity of sixty‐two antibacterial agents. Journal of Antimicrobial Chemotherapy, 1995. 35(1): p. 149‐154.
72. Hutmacher, D.W., et al., Mechanical properties and cell cultural response of polycaprolactone scaffolds designed and fabricated via fused deposition modeling. J Biomed Mater Res, 2001. 55(2): p. 203‐16.
73. Loessner, D., et al., Functionalization, preparation and use of cell‐laden gelatin methacryloyl‐based hydrogels as modular tissue culture platforms. Nature protocols, 2016. 11: p. 727‐46.
74. Varslot, T., et al., High‐resolution helical cone‐beam micro‐CT with theoretically‐exact reconstruction from experimental data. Med Phys, 2011. 38(10): p. 5459.
75. Yazdi, I.K., et al., Cefazolin‐loaded mesoporous silicon microparticles show sustained bactericidal effect against Staphylococcus aureus. J Tissue Eng, 2014. 5: p. 2041731414536573.
Page 164
Page 164
76. Meijer, G.J., et al., Cell‐based bone tissue engineering. PLoS medicine, 2007. 4(2): p. e9.
77. Fisher, J.P., et al., Soft and hard tissue response to photocrosslinked poly (propylene fumarate) scaffolds in a rabbit model. Journal of Biomedical Materials Research Part A, 2002. 59(3): p. 547‐556.
78. Mouriño, V. and A.R. Boccaccini, Bone tissue engineering therapeutics: controlled drug delivery in three‐dimensional scaffolds. Journal of the Royal Society Interface, 2009: p. rsif20090379.
79. Kim, K., et al., Incorporation and controlled release of a hydrophilic antibiotic using poly (lactide‐co‐glycolide)‐based electrospun nanofibrous scaffolds. Journal of Controlled Release, 2004. 98(1): p. 47‐56.
80. Loh, Q.L. and C. Choong, Three‐dimensional scaffolds for tissue engineering applications: role of porosity and pore size. Tissue Eng Part B Rev, 2013. 19(6): p. 485‐502.
81. Hutmacher, D.W. and S. Cool, Concepts of scaffold‐based tissue engineering ‐ The rationale to use solid free‐form fabrication techniques. Journal of Cellular and Molecular Medicine, 2007. 11: p. 654‐669.
82. Prabu, P., et al., Preparation and drug release activity of scaffolds containing collagen and poly (caprolactone). Journal of Biomedical Materials Research Part A, 2006. 79(1): p. 153‐158.
83. Miyai, T., et al., Antibiotic‐loaded poly‐ε‐caprolactone and porous β‐tricalcium phosphate composite for treating osteomyelitis. Biomaterials, 2008. 29(3): p. 350‐358.
84. Raval, A., J. Parikh, and C. Engineer, Mechanism of Controlled Release Kinetics From Medical Devices. Brazilian Journal of Chemical Engineering, 2010. 27: p. 211‐225.
85. Elsner, J.J., et al., Highly porous drug‐eluting structures: from wound dressings to stents and scaffolds for tissue regeneration. Biomatter, 2012. 2: p. 239‐270.
86. Mouriño, V. and A.R. Boccaccini, Bone tissue engineering therapeutics: controlled drug delivery in three‐dimensional scaffolds. Journal of the Royal Society, Interface / the Royal Society, 2010. 7: p. 209‐27.
87. Abizaid, A. and J.R. Costa, New drug‐eluting stents an overview on biodegradable and polymer‐free next‐generation stent systems. Circulation: Cardiovascular Interventions, 2010. 3: p. 384‐393.
88. Yue, K., et al., Synthesis, properties, and biomedical applications of gelatin methacryloyl (GelMA) hydrogels. Biomaterials, 2015. 73: p. 254‐271.
89. Zeller, V., et al., Ctontinuous cefazolin infusion to treat bone and joint infections: Clinical efficacy, feasibility, safety, and serum and bone concentrations. Antimicrobial Agents and Chemotherapy, 2009. 53: p. 883‐887.
90. Douglas, A., et al., Plasma and tissue pharmacokinetics of cefazolin in patients undergoing elective and semielective abdominal aortic aneurysm open repair surgery. Antimicrob Agents Chemother, 2011. 55(11): p. 5238‐42.
91. Mutsuzaki, H., et al., Cefazolin‐containing poly(ε‐caprolactone) sponge pad to reduce pin tract infection rate in rabbits. Asia‐Pacific Journal of Sports Medicine, Arthroscopy, Rehabilitation and Technology, 2014. 1(2): p. 54‐61.
92. Teo, E.Y., et al., Polycaprolactone‐based fused deposition modeled mesh for delivery of antibacterial agents to infected wounds. Biomaterials, 2011. 32(1): p. 279‐87.
93. Reichert, J.C., et al., A Tissue Engineering Solution for Segmental Defect Regeneration in Load‐Bearing Long Bones. Science Translational Medicine, 2012. 4: p. 1‐9.
94. Wang, M., et al., Cold atmospheric plasma (CAP) surface nanomodified 3D printed polylactic acid (PLA) scaffolds for bone regeneration. Acta Biomater, 2016. 46: p. 256‐265.
Page 165
Page 165
95. Bauermeister, A.J., A. Zuriarrain, and M.I. Newman, Three‐Dimensional Printing in Plastic and Reconstructive Surgery: A Systematic Review. Ann Plast Surg, 2015.
96. Melchels, F., et al., CAD/CAM‐assisted breast reconstruction. Biofabrication, 2011. 3(3): p. 034114.
97. Winter, J., S. Shiga, and A. Islur, The Complications of Polyacrylamide Hydrogel Augmentation Mammoplasty. Plastic Surgery Case Studies, 2017. 3: p. 2513826X1769382.
98. Nava, M.B., et al., How to prevent complications in breast augmentation. Gland Surg, 2017. 6(2): p. 210‐217.
99. Wei, W., Treatment of complications from polyacrylamide hydrogel breast augmentation. Exp Ther Med, 2016. 12(1): p. 173‐176.
100. Ogita, M., et al., Risk factors for complications among breast cancer patients treated with post‐mastectomy radiotherapy and immediate tissue‐expander/permanent implant reconstruction: a retrospective cohort study. Breast Cancer, 2017. 19: p. 1‐9.
101. Visscher, L.E., et al., Breast Augmentation and Reconstruction from a Regenerative Medicine Point of View: State of the Art and Future Perspectives. Tissue Eng Part B Rev, 2017. 23(3): p. 281‐293.
102. Gehl, J., et al., Combined doxorubicin and paclitaxel in advanced breast cancer: Effective and cardiotoxic. Annals of Oncology, 1996. 7: p. 687‐693.
103. Swain, S.M., F.S. Whaley, and M.S. Ewer, Congestive heart failure in patients treated with doxorubicin: a retrospective analysis of three trials. Cancer, 2003. 97(11): p. 2869‐79.
104. Wang, F., et al., Preclinical development of drug delivery systems for paclitaxel‐based cancer chemotherapy. J Control Release, 2017.
105. Feng, L. and R.J. Mumper, A critical review of lipid‐based nanoparticles for taxane delivery. Cancer Lett, 2013. 334(2): p. 157‐75.
106. Gelderblom, H., et al., Cremophor EL: the drawbacks and advantages of vehicle selection for drug formulation. Eur J Cancer, 2001. 37(13): p. 1590‐8.
107. Sparreboom, A., et al., Nonlinear pharmacokinetics of paclitaxel in mice results from the pharmaceutical vehicle Cremophor EL. Cancer Res, 1996. 56(9): p. 2112‐5.
108. Fonseca, C., S. Simoes, and R. Gaspar, Paclitaxel‐loaded PLGA nanoparticles: preparation, physicochemical chracterization and in vitro anti‐tumoral activity. Journal of Controlled Release, 2002. 83: p. 273‐286.
109. Crosasso, P., et al., Preparation, characterization and properties of sterically stabilized paclitaxel‐containing liposomes. J Control Release, 2000. 63(1‐2): p. 19‐30.
110. Khandare, J.J., et al., Dendrimer versus linear conjugate: Influence of polymeric architecture on the delivery and anticancer effect of paclitaxel. Bioconjug Chem, 2006. 17(6): p. 1464‐72.
111. Sun, T., et al., Engineered nanoparticles for drug delivery in cancer therapy. Angew Chem Int Ed Engl, 2014. 53(46): p. 12320‐64.
112. Gibson, J.D., B.P. Khanal, and E.R. Zubarev, Paclitaxel‐functionalized gold nanoparticles. J Am Chem Soc, 2007. 129(37): p. 11653‐61.
113. Tran, N., et al., Non‐lamellar lyotropic liquid crystalline nanoparticles enhance the antibacterial effects of rifampicin against Staphylococcus aureus. J Colloid Interface Sci, 2018. 519: p. 107‐118.
114. Zhai, J., et al., Epidermal growth factor receptor‐targeted lipid nanoparticles retain self‐assembled nanostructures and provide high specificity. Nanoscale, 2015. 7(7): p. 2905‐13.
115. Wu, J., Q. Liu, and R.J. Lee, A folate receptor‐targeted liposomal formulation for paclitaxel. Int J Pharm, 2006. 316(1‐2): p. 148‐53.
116. Zhong, Y., et al., Ligand‐directed active tumor‐targeting polymeric nanoparticles for cancer chemotherapy. Biomacromolecules, 2014. 15(6): p. 1955‐69.
Page 166
Page 166
117. Choi, C.H., et al., Mechanism of active targeting in solid tumors with transferrin‐containing gold nanoparticles. Proc Natl Acad Sci U S A, 2010. 107(3): p. 1235‐40.
118. Zhang, L., et al., Tailored Synthesis of Octopus‐type Janus Nanoparticles for Synergistic Actively‐Targeted and Chemo‐Photothermal Therapy. Angew Chem Int Ed Engl, 2016. 55(6): p. 2118‐21.
119. Gao, W., J.M. Chan, and O.C. Farokhzad, pH‐Responsive nanoparticles for drug delivery. Mol Pharm, 2010. 7(6): p. 1913‐20.
120. Mura, S., J. Nicolas, and P. Couvreur, Stimuli‐responsive nanocarriers for drug delivery. Nat Mater, 2013. 12(11): p. 991‐1003.
121. Park, J.W., Liposome‐based drug delivery in breast cancer treatment. Breast Cancer Res, 2002. 4(3): p. 95‐9.
122. O'Brien, M.E.R., Reduced cardiotoxicity and comparable efficacy in a phase III trial of pegylated liposomal doxorubicin HCl (CAELYXTM/Doxil") versus conventional doxorubicin for first‐line treatment of metastatic breast cancer. Annals of Oncology, 2004. 15(3): p. 440‐449.
123. Miele, E., et al., Albumin‐bound formulation of paclitaxel (Abraxane® ABI‐007) in the treatment of breast cancer. International Journal of Nanomedicine, 2009. 4: p. 99‐105.
124. Green, M.R., et al., Abraxane, a novel Cremophor‐free, albumin‐bound particle form of paclitaxel for the treatment of advanced non‐small‐cell lung cancer. Ann Oncol, 2006. 17(8): p. 1263‐8.
125. Micha, J.P., et al., Abraxane in the treatment of ovarian cancer: the absence of hypersensitivity reactions. Gynecol Oncol, 2006. 100(2): p. 437‐8.
126. Zhai, J., et al., Paclitaxel‐Loaded Self‐Assembled Lipid Nanoparticles as Targeted Drug Delivery Systems for the Treatment of Aggressive Ovarian Cancer. ACS Appl Mater Interfaces, 2018. 10(30): p. 25174‐25185.
127. Danhier, F., To exploit the tumor microenvironment: Since the EPR effect fails in the clinic, what is the future of nanomedicine? J Control Release, 2016. 244(Pt A): p. 108‐121.
128. Wilhelm, S., et al., Analysis of nanoparticle delivery to tumours. Nature Reviews Materials, 2016. 1(5).
129. Wolinsky, J.B., Y.L. Colson, and M.W. Grinstaff, Local drug delivery strategies for cancer treatment: gels, nanoparticles, polymeric films, rods, and wafers. J Control Release, 2012. 159(1): p. 14‐26.
130. De Souza, R., et al., Polymeric drug delivery systems for localized cancer chemotherapy. Drug Deliv, 2010. 17(6): p. 365‐75.
131. Bastiancich, C., et al., Anticancer drug‐loaded hydrogels as drug delivery systems for the local treatment of glioblastoma. J Control Release, 2016. 243: p. 29‐42.
132. Weinberg, B.D., et al., Antitumor efficacy and local distribution of doxorubicin via intratumoral delivery from polymer millirods. J Biomed Mater Res A, 2007. 81(1): p. 161‐70.
133. Boateng, J.S., et al., Characterisation of freeze‐dried wafers and solvent evaporated films as potential drug delivery systems to mucosal surfaces. Int J Pharm, 2010. 389(1‐2): p. 24‐31.
134. Sionkowska, A., Current research on the blends of natural and synthetic polymers as new biomaterials: Review. Progress in Polymer Science, 2011. 36(9): p. 1254‐1276.
135. Uhrich, K.E., et al., Polymeric Systems for Controlled Drug Release. Chemical Reviews, 1999. 99(11): p. 3181‐3198.
136. Pillai, O. and R. Panchagnula, Polymers in drug delivery. Curr Opin Chem Biol, 2001. 5(4): p. 447‐51.
137. Inzana, J.A., et al., 3D printed bioceramics for dual antibiotic delivery to treat implant‐associated bone infection. Eur Cell Mater, 2015. 30: p. 232‐47.
Page 167
Page 167
138. Jeon, H., H. Lee, and G. Kim, A surface‐modified poly(varepsilon‐caprolactone) scaffold comprising variable nanosized surface‐roughness using a plasma treatment. Tissue Eng Part C Methods, 2014. 20(12): p. 951‐63.
139. Sigma‐Aldrich. 440752 Aldrich Polycaprolactone. Available from: https://www.sigmaaldrich.com/catalog/product/aldrich/440752?lang=en®ion=AU.
140. Wunderlich, B., Heat capacity of polymers. Handbook of thermal Analysis and Calorimetry, 2002. 3(1): p. 47.
141. Thorn, C.F., et al., Doxorubicin pathways: pharmacodynamics and adverse effects. Pharmacogenet Genomics, 2011. 21(7): p. 440‐6.
142. Von Hoff, D.D., et al., Risk factors for doxorubicin‐induced congestive heart failure. Ann Intern Med, 1979. 91(5): p. 710‐7.
143. Leenheer, J.A., R.L. Wershaw, and M.M. Reddy, Strong‐Acid, Carboxyl‐Group Structures in Fulvic‐Acid from the Suwannee River, Georgia .2. Major Structures. Environmental Science & Technology, 1995. 29(2): p. 399‐405.
144. Rowinsky, E.K. and R.C. Donehower, Paclitaxel (taxol). N Engl J Med, 1995. 332(15): p. 1004‐14.
145. Perez, E.A., Paclitaxel in Breast Cancer. The Oncologist, 1998. 3(6): p. 373‐389. 146. Ma, P. and R.J. Mumper, Paclitaxel Nano‐Delivery Systems: A Comprehensive Review.
J Nanomed Nanotechnol, 2013. 4(2): p. 1000164. 147. Chao, T.C., et al., Paclitaxel in a novel formulation containing less Cremophor EL as
first‐line therapy for advanced breast cancer: A phase II trial. Investigational New Drugs, 2005. 23(2): p. 171‐177.
148. Periti, P., G. Stringa, and E. Mini, Comparative multicenter trial of teicoplanin versus cefazolin for antimicrobial prophylaxis in prosthetic joint implant surgery. Italian Study Group for Antimicrobial Prophylaxis in Orthopedic Surgery. Eur J Clin Microbiol Infect Dis, 1999. 18(2): p. 113‐9.
149. Allababidi, S. and J.C. Shah, Efficacy and pharmacokinetics of site‐specific cefazolin delivery using biodegradable implants in the prevention of post‐operative wound infections. Pharmaceutical Research, 1998. 15(2): p. 325‐333.
150. Adams, W.P., Jr., et al., Optimizing breast‐pocket irrigation: the post‐betadine era. Plast Reconstr Surg, 2001. 107(6): p. 1596‐1601.
151. Tacconelli, E., et al., Vancomycin versus cefazolin prophylaxis for cerebrospinal shunt placement in a hospital with a high prevalence of meticillin‐resistant Staphylococcus aureus. J Hosp Infect, 2008. 69(4): p. 337‐44.
152. Jiang, B. and B. Li, Polypeptide nanocoatings for preventing dental and orthopaedic device‐associated infection: pH‐induced antibiotic capture, release, and antibiotic efficacy. J Biomed Mater Res B Appl Biomater, 2009. 88(2): p. 332‐8.
153. Loessner, D., et al., Functionalization, preparation and use of cell‐laden gelatin methacryloyl‐based hydrogels as modular tissue culture platforms. Nat Protoc, 2016. 11(4): p. 727‐46.
154. Burgess, D.J. and J.E. Carless, Microelectrophoretic Studies of Gelatin and Acacia for the Prediction of Complex Coacervation. Journal of Colloid and Interface Science, 1984. 98(1): p. 1‐8.
155. Lee, J. and D.‐W. Cho, 3D Printing Technology Over a Drug Delivery for Tissue Engineering. Current Pharmaceutical Design, 2015. 21(12): p. 1606‐1617.
156. Mathew, A., et al., Antimicrobial and Immunomodulatory Surface‐Functionalized Electrospun Membranes for Bone Regeneration. Adv Healthc Mater, 2017. 6(10).
157. Beeley, N.R., et al., Fabrication, implantation, elution, and retrieval of a steroid‐loaded polycaprolactone subretinal implant. J Biomed Mater Res A, 2005. 73(4): p. 437‐44.
Page 168
Page 168
158. Gorrasi, G., et al., Vapor barrier properties of polycaprolactone montmorillonite nanocomposites: effect of clay dispersion. Polymer, 2003. 44(8): p. 2271‐2279.
159. Gao, B., et al., 4D Bioprinting for Biomedical Applications. Trends Biotechnol, 2016. 34(9): p. 746‐756.
160. Wu, J.‐J., et al., 4D Printing: History and Recent Progress. Chinese Journal of Polymer Science, 2017. 36(5): p. 563‐575.
161. Gladman, A.S., et al., Biomimetic 4D printing. Nat Mater, 2016. 15(4): p. 413‐8. 162. He, H., J. Guan, and J.L. Lee, An oral delivery device based on self‐folding hydrogels. J
Control Release, 2006. 110(2): p. 339‐346. 163. Siegel, R.L., K.D. Miller, and A. Jemal, Cancer statistics, 2015. CA Cancer J Clin, 2015.
65(1): p. 5‐29. 164. National Cancer Institute, SEER Cancer Statistics Review: 1975‐2012. 2015. 165. Australian Institute of Health and Welfare, Australian Cancer Incidence and Mortality
(ACIM) books: breast cancer. 2016. 166. Paredes, et al., The impact of breast reconstruction on the quality of life of patients
after mastectomy at the Plastic Surgery Service of Walter Cantídio University Hospital. Revista Brasileira de Cirurgia Plástica, 2013. 28(1): p. 100‐4.
167. Nissen, M.J., K.K. Swenson, and E.A. Kind, Quality of life after postmastectomy breast reconstruction. Oncol Nurs Forum, 2002. 29(3): p. 547‐53.
168. Stavrou, et al., Quality of life after breast cancer surgery with or without reconstruction. ePlasty: Open Access Journal of Plastic Surgery, 2009. 9.
169. Dean, C., U. Chetty, and A.P. Forrest, Effects of immediate breast reconstruction on psychosocial morbidity after mastectomy. Lancet, 1983. 1(8322): p. 459‐62.
170. Anampa, J., D. Makower, and J.A. Sparano, Progress in adjuvant chemotherapy for breast cancer: an overview. BMC Med, 2015. 13: p. 195.
171. Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15‐year survival: an overview of the randomised trials. The Lancet, 2005. 366(9503): p. 2087‐2106.
172. Polgar, C., et al., Sole brachytherapy of the tumor bed after conservative surgery for T1 breast cancer: five‐year results of a phase I‐II study and initial findings of a randomized phase III trial. J Surg Oncol, 2002. 80(3): p. 121‐8; discussion 129.
173. de la Puente, P. and A.K. Azab, Delivery systems for brachytherapy. J Control Release, 2014. 192: p. 19‐28.
174. Onitilo, A.A., et al., Breast cancer subtypes based on ER/PR and Her2 expression: comparison of clinicopathologic features and survival. Clin Med Res, 2009. 7(1‐2): p. 4‐13.
175. Schlegel, P.N., Efficacy and Safety of Histrelin Subdermal Implant in Patients With Advanced Prostate Cancer. The Journal of Urology, 2006. 175(4): p. 1353‐1358.
176. Westphal, M., et al., Gliadel wafer in initial surgery for malignant glioma: long‐term follow‐up of a multicenter controlled trial. Acta Neurochir (Wien), 2006. 148(3): p. 269‐75; discussion 275.
177. Wu, Y., et al., An Injectable Supramolecular Polymer Nanocomposite Hydrogel for Prevention of Breast Cancer Recurrence with Theranostic and Mammoplastic Functions. Advanced Functional Materials, 2018. 28(21): p. 1801000.
178. Qi, Y., et al., Injectable Hexapeptide Hydrogel for Localized Chemotherapy Prevents Breast Cancer Recurrence. ACS Appl Mater Interfaces, 2018. 10(8): p. 6972‐6981.
179. Seib, F.P. and D.L. Kaplan, Doxorubicin‐loaded silk films: drug‐silk interactions and in vivo performance in human orthotopic breast cancer. Biomaterials, 2012. 33(33): p. 8442‐50.
180. Li, Q., et al., Graphene‐Nanoparticle‐Based Self‐Healing Hydrogel in Preventing Postoperative Recurrence of Breast Cancer. ACS Biomaterials Science & Engineering, 2019. 5(2): p. 768‐779.
Page 169
Page 169
181. Yuan, Z., et al., Doxorubicin‐loaded mesoporous silica nanoparticle composite nanofibers for long‐term adjustments of tumor apoptosis. Nanotechnology, 2016. 27(24): p. 245101.
182. Yuan, Z., et al., In situ adjuvant therapy using a responsive doxorubicin‐loaded fibrous scaffold after tumor resection. Colloids Surf B Biointerfaces, 2017. 158: p. 363‐369.
183. Ding, Q., et al., Preparation and therapeutic application of docetaxel‐loaded poly(d,l‐lactide) nanofibers in preventing breast cancer recurrence. Drug Deliv, 2016. 23(8): p. 2677‐2685.
184. Hollander, J., et al., Three‐Dimensional Printed PCL‐Based Implantable Prototypes of Medical Devices for Controlled Drug Delivery. J Pharm Sci, 2016. 105(9): p. 2665‐2676.
185. Ahangar, P., et al., Nanoporous 3D‐Printed Scaffolds for Local Doxorubicin Delivery in Bone Metastases Secondary to Prostate Cancer. Materials (Basel), 2018. 11(9).
186. Dang, H.P., et al., 3D printed biodegradable scaffolds with dual macro‐,micro‐scale porous networks for bone tissue regeneration. Biomaterials, 2019. Submitted.
187. Dang, H.P., et al., 3D printed dual macro‐, microscale porous network as a tissue engineering scaffold with drug delivering function. Biofabrication, 2019.
188. Untch, M. and G. von Minckwitz, Recent advances in systemic therapy: advances in neoadjuvant (primary) systemic therapy with cytotoxic agents. Breast Cancer Res, 2009. 11(2): p. 203.
189. Sun, H., et al., The in vivo degradation, absorption and excretion of PCL‐based implant. Biomaterials, 2006. 27(9): p. 1735‐40.
190. Dang, H.P., et al., 3D printed dual macro‐, microscale porous network as a tissue engineering scaffold with drug delivering function. Biofabrication, 2019. 11(3): p. 035014.
191. Defail, A.J., et al., Controlled release of bioactive doxorubicin from microspheres embedded within gelatin scaffolds. J Biomed Mater Res A, 2006. 79(4): p. 954‐62.
192. Bala Balakrishnan, P., et al., Star poly(epsilon‐caprolactone)‐based electrospun fibers as biocompatible scaffold for doxorubicin with prolonged drug release activity. Colloids Surf B Biointerfaces, 2018. 161: p. 488‐496.
193. Sedlacek, O., et al., Poly(2‐ethyl‐2‐oxazoline) conjugates with doxorubicin for cancer therapy: In vitro and in vivo evaluation and direct comparison to poly[N‐(2‐hydroxypropyl)methacrylamide] analogues. Biomaterials, 2017. 146: p. 1‐12.
194. Henninger, C., et al., Chronic heart damage following doxorubicin treatment is alleviated by lovastatin. Pharmacol Res, 2015. 91: p. 47‐56.
195. Huang, W., et al., Surfactin‐based nanoparticles loaded with doxorubicin to overcome multidrug resistance in cancers. Int J Nanomedicine, 2018. 13: p. 1723‐1736.
196. Alkreathy, H.M., et al., Enhanced antitumour activity of doxorubicin and simvastatin combination loaded nanoemulsion treatment against a Swiss albino mouse model of Ehrlich ascites carcinoma. Clin Exp Pharmacol Physiol, 2019. 46(5): p. 496‐505.
197. Pawar, S., G. Shevalkar, and P. Vavia, Glucosamine‐anchored doxorubicin‐loaded targeted nano‐niosomes: pharmacokinetic, toxicity and pharmacodynamic evaluation. J Drug Target, 2016. 24(8): p. 730‐43.
198. V, S.S. and V.M. P, Degradation of Poly(epsilon‐caprolactone) and bio‐interactions with mouse bone marrow mesenchymal stem cells. Colloids Surf B Biointerfaces, 2018. 163: p. 107‐118.
199. Carter, S.D., et al., Additive Biomanufacturing: An Advanced Approach for Periodontal Tissue Regeneration. Ann Biomed Eng, 2017. 45(1): p. 12‐22.
200. Hutmacher, D.W., M. Sittinger, and M.V. Risbud, Scaffold‐based tissue engineering: rationale for computer‐aided design and solid free‐form fabrication systems. Trends Biotechnol, 2004. 22(7): p. 354‐62.
Page 170
Page 170
201. Cipitria, A., et al., BMP delivery complements the guiding effect of scaffold architecture without altering bone microstructure in critical‐sized long bone defects: A multiscale analysis. Acta Biomater, 2015. 23: p. 282‐294.
202. Visscher, L.E., et al., 3D printed Polycaprolactone scaffolds with dual macro‐microporosity for applications in local delivery of antibiotics. Materials Science and Engineering: C, 2018. 87: p. 78‐89.
203. Kolar, P., et al., Human early fracture hematoma is characterized by inflammation and hypoxia. Clin Orthop Relat Res, 2011. 469(11): p. 3118‐26.
204. Claes, L., S. Recknagel, and A. Ignatius, Fracture healing under healthy and inflammatory conditions. Nature Reviews Rheumatology, 2012. 8: p. 133.
205. Claes, L., S. Recknagel, and A. Ignatius, Fracture healing under healthy and inflammatory conditions. Nat Rev Rheumatol, 2012. 8(3): p. 133‐43.
206. Manavitehrani, I., et al., Biomedical Applications of Biodegradable Polyesters. Polymers, 2016. 8(1): p. 20.
207. A review paper on biomimetic calcium phosphate coating. The Open Biomedical Engineering Journal, 2015. 9: p. 56‐64.
208. Dadsetan, M., et al., Effect of calcium phosphate coating and rhBMP‐2 on bone regeneration in rabbit calvaria using poly(propylene fumarate) scaffolds. Acta Biomater, 2015. 18: p. 9‐20.
209. Surmenev, R.A., M.A. Surmeneva, and A.A. Ivanova, Significance of calcium phosphate coatings for the enhancement of new bone osteogenesis‐‐a review. Acta Biomater, 2014. 10(2): p. 557‐79.
210. Vaquette, C., et al., Effect of culture conditions and calcium phosphate coating on ectopic bone formation. Biomaterials, 2013. 34(22): p. 5538‐51.
211. Cacicedo, M.L., et al., Modified bacterial cellulose scaffolds for localized doxorubicin release in human colorectal HT‐29 cells. Colloids Surf B Biointerfaces, 2016. 140: p. 421‐9.
212. Kosik‐Kozio, A., et al., Surface modification of 3D printed polycaprolactone constructs via a solvent treatment: impact on physical and osteogenic properties. ACS Biomaterials Science & Engineering, 2018.
213. Crescenzi, V., et al., Thermodynamics of fusion of poly‐β‐propiolactone and poly‐ϵ‐caprolactone. Comparative analysis of the melting of aliphatic polylactone and polyester chains. European Polymer journal, 1972. 8: p. 15.
214. Tiraferri, A. and M. Elimelech, Direct quantification of negatively charged functional groups on membrane surfaces. Journal of Membrane Science, 2012. 389: p. 499‐508.
215. Ma, Z., Z. Mao, and C. Gao, Surface modification and property analysis of biomedical polymers used for tissue engineering. Colloids Surf B Biointerfaces, 2007. 60(2): p. 137‐57.
216. Foo, H.L., et al., Catalytic surface modification of roll‐milled poly(ε‐caprolactone) biaxially stretched to ultra‐thin dimension. Materials Science and Engineering: C, 2007. 27(2): p. 299‐303.
217. Jassal, M., S. Sengupta, and S. Bhowmick, Functionalization of electrospun poly(caprolactone) fibers for pH‐controlled delivery of doxorubicin hydrochloride. J Biomater Sci Polym Ed, 2015. 26(18): p. 1425‐38.
218. Standard Test Method for Tensile Properties of Plastics. 2014, ASTM International. 219. Lam, C.X., et al., Evaluation of polycaprolactone scaffold degradation for 6 months in
vitro and in vivo. J Biomed Mater Res A, 2009. 90(3): p. 906‐19. 220. Lam, C.X.F., S.H. Teoh, and D.W. Hutmacher, Comparison of the degradation of
polycaprolactone and polycaprolactone–(β‐tricalcium phosphate) scaffolds in alkaline medium. Polymer International, 2007. 56(6): p. 718‐728.
Page 171
Page 171
221. Lam, C.X., et al., Dynamics of in vitro polymer degradation of polycaprolactone‐based scaffolds: accelerated versus simulated physiological conditions. Biomed Mater, 2008. 3(3): p. 034108.
222. Shiu, H.T., et al., Controlling whole blood activation and resultant clot properties by carboxyl and alkyl functional groups on material surfaces: a possible therapeutic approach for enhancing bone healing. J. Mater. Chem. B, 2014. 2(20): p. 3009‐3021.
223. Sadowska, J.M., et al., Effect of nano‐structural properties of biomimetic hydroxyapatite on osteoimmunomodulation. Biomaterials, 2018. 181: p. 318‐332.
224. Woo, K.M., V.J. Chen, and P.X. Ma, Nano‐fibrous scaffolding architecture selectively enhances protein adsorption contributing to cell attachment. J Biomed Mater Res A, 2003. 67(2): p. 531‐7.
225. Dinarvand, P., et al., New approach to bone tissue engineering: simultaneous application of hydroxyapatite and bioactive glass coated on a poly(L‐lactic acid) scaffold. ACS Appl Mater Interfaces, 2011. 3(11): p. 4518‐24.
226. Wen, J., Heat capacities of polymers, in Physical Properties of Polymers Handbook. 2007. p. 145‐154.
227. Archer, D.G., Thermodynamic Properties of the NaCl + H2O System. I. Thermodynamic Properties of NaCl(cr). journal of physical and chemical reference data, 1991. 21(1): p. 1‐21.
228. Kishore, V., et al. Predicting sharkskin instability in extrusion additive manufacturing of reinforced thermoplastics. in The 28th Annual International Solid Freeform Fabrication Symposium ‐ An Additive Manufacturing Conference. 2017.
229. Choong, C., J.T. Triffitt, and Z.F. Cui, Polycaprolactone Scaffolds for Bone Tissue Engineering. Food and Bioproducts Processing, 2004. 82(2): p. 117‐125.
230. Poh, P.S.P., et al., In vitro and in vivo bone formation potential of surface calcium phosphate‐coated polycaprolactone and polycaprolactone/bioactive glass composite scaffolds. Acta Biomater, 2016. 30: p. 319‐333.
231. Amani, H., et al., Controlling Cell Behavior through the Design of Biomaterial Surfaces: A Focus on Surface Modification Techniques. Advanced Materials Interfaces, 2019: p. 1900572.
232. Palmer, J.A., et al., Macrophage phenotype in response to implanted synthetic scaffolds: an immunohistochemical study in the rat. Cells Tissues Organs, 2014. 199(2‐3): p. 169‐83.
233. Mantovani, A., et al., Macrophage polarization: tumor‐associated macrophages as a paradigm for polarized M 2 mononuclear phagocytes. TRENDS in Immunology, 2002. 23(11): p. 549‐55.
234. Feng, B., et al., The effect of pore size on tissue ingrowth and neovascularization in porous bioceramics of controlled architecture in vivo. Biomedical Materials, 2011. 6(1): p. 015007.
235. Shafiee, A., et al., Fetal Bone Marrow‐Derived Mesenchymal Stem/Stromal Cells Enhance Humanization and Bone Formation of BMP7 Loaded Scaffolds. Biotechnol J, 2017. 12(12).
236. Zhang, K., et al., Effect of microporosity on scaffolds for bone tissue engineering. Regen Biomater, 2018. 5(2): p. 115‐124.
237. Rustom, L.E., M.J. Poellmann, and A.J. Wagoner Johnson, Mineralization in micropores of calcium phosphate scaffolds. Acta Biomater, 2019. 83: p. 435‐455.
238. Perez, R.A. and G. Mestres, Role of pore size and morphology in musculo‐skeletal tissue regeneration. Mater Sci Eng C Mater Biol Appl, 2016. 61: p. 922‐39.
239. Turnbull, G., et al., 3D bioactive composite scaffolds for bone tissue engineering. Bioact Mater, 2018. 3(3): p. 278‐314.
Page 172
Page 172
240. Khampieng, T., et al., Protein adsorption and cell behaviors on polycaprolactone film: The effect of surface topography. Advances in Polymer Technology, 2018. 37(6): p. 2030‐2042.
241. De Witte, T.M., et al., Bone tissue engineering via growth factor delivery: from scaffolds to complex matrices. Regen Biomater, 2018. 5(4): p. 197‐211.
242. Stevens, M.M., Biomaterials for bone tissue engineering. Materials Today, 2008. 11(5): p. 18‐25.
243. Zhang, N., et al., Assessment of the degradation rates and effectiveness of different coated Mg‐Zn‐Ca alloy scaffolds for in vivo repair of critical‐size bone defects. J Mater Sci Mater Med, 2018. 29(9): p. 138.
244. Shiu, H.T., et al., Formation of blood clot on biomaterial implants influences bone healing. Tissue Eng Part B Rev, 2014. 20(6): p. 697‐712.
245. Hoff, P., et al., Human immune cells' behavior and survival under bioenergetically restricted conditions in an in vitro fracture hematoma model. Cell Mol Immunol, 2013. 10(2): p. 151‐8.
246. Shah, P., L. Keppler, and J. Rutkowski, A review of platelet derived growth factor playing pivotal role in bone regeneration. J Oral Implantol, 2014. 40(3): p. 330‐40.
247. Bonewald, L.F. and G.R. Mundy, Role of Transforming Growth Factor Beta in Bone Remodeling: A Review. Connective Tissue Research, 2009. 23(2‐3): p. 201‐208.
248. Tang, W., et al., Bioinspired trimodal macro/micro/nano‐porous scaffolds loading rhBMP‐2 for complete regeneration of critical size bone defect. Acta Biomater, 2016. 32: p. 309‐323.
249. Sohier, J., et al., Tailored release of TGF‐beta1 from porous scaffolds for cartilage tissue engineering. Int J Pharm, 2007. 332(1‐2): p. 80‐9.
250. Durante, C., et al., Growth factor release from platelet concentrates: analytic quantification and characterization for clinical applications. Vox Sang, 2013. 105(2): p. 129‐36.
251. Schar, M.O., et al., Platelet‐rich concentrates differentially release growth factors and induce cell migration in vitro. Clin Orthop Relat Res, 2015. 473(5): p. 1635‐43.
252. Petrochenko, P. and R.J. Narayan, Novel approaches to bone grafting: porosity, bone morphogenetic proteins, stem cells, and the periosteum. J Long Term Eff Med Implants, 2010. 20(4): p. 303‐15.
253. Microporosity enhances bioactivity of synthetic bone draft substitutes. Journal of Materials Science: Materials in Medicine, 2005. 16: p. 467‐475.
254. Woodruff, M.A., et al., Sustained release and osteogenic potential of heparan sulfate‐doped fibrin glue scaffolds within a rat cranial model. J Mol Histol, 2007. 38(5): p. 425‐33.
255. Schantz, J.T., et al., Repair of calvarial defects with customised tissue‐engineered bone grafts II. Evaluation of cellular efficiency and efficacy in vivo. Tissue Eng, 2003. 9 Suppl 1: p. S127‐39.
256. Schantz, J.T., et al., Repair of calvarial defects with customised tissue‐engineered bone grafts ‐ II. Evaluation of cellular efficiency and efficacy in vivo. Tissue Engineering, 2003. 9: p. S127‐S139.
257. Berner, A., et al., Effects of scaffold architecture on cranial bone healing. Int J Oral Maxillofac Surg, 2014. 43(4): p. 506‐13.
258. Sawyer, A.A., et al., The stimulation of healing within a rat calvarial defect by mPCL‐TCP/collagen scaffolds loaded with rhBMP‐2. Biomaterials, 2009. 30(13): p. 2479‐88.
259. Vaquette, C., et al., A biphasic scaffold design combined with cell sheet technology for simultaneous regeneration of alveolar bone/periodontal ligament complex. Biomaterials, 2012. 33(22): p. 5560‐5573.
Page 173
Page 173
260. Mathew, A., et al., Antimicrobial and Immunomodulatory Surface‐Functionalized Electrospun Membranes for Bone Regeneration. Advanced healthcare materials, 2017. 6(10): p. 1601345.
261. Somo, S.I., et al., Pore Interconnectivity Influences Growth Factor‐Mediated Vascularization in Sphere‐Templated Hydrogels. Tissue Eng Part C Methods, 2015. 21(8): p. 773‐85.
262. Niza, E., et al., Assessment of doxorubicin delivery devices based on tailored bare polycaprolactone against glioblastoma. Int J Pharm, 2019. 558: p. 110‐119.
263. Innocente, F., et al., Paclitaxel‐eluting biodegradable synthetic vascular prostheses: a step towards reduction of neointima formation? Circulation, 2009. 120(11 Suppl): p. S37‐45.
264. Zhu, Y., et al., A highly flexible paclitaxel‐loaded poly(epsilon‐caprolactone) electrospun fibrous‐membrane‐covered stent for benign cardia stricture. Acta Biomater, 2013. 9(9): p. 8328‐36.
265. Rath, G., et al., Fabrication and characterization of cefazolin‐loaded nanofibrous mats for the recovery of post‐surgical wound. Artif Cells Nanomed Biotechnol, 2016. 44(8): p. 1783‐1792.