Embed Size (px)
Citation preview
Nanostructured Selenium for Biomedical Applications:
from Theory to Practice
By Phong A. Tran
Brown University, 2010
A Dissertation Submitted in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
in the Department of Physics at Brown University
Providence, Rhode Island
May 2010
© Copyright 2010 by Phong A. Tran
iii
This dissertation by Phong A. Tran is accepted in its present form
by the Department of Physics as satisfying the
dissertation requirement for the degree of Doctor of Philosophy.
Date __________________ ___________________________
Thomas J. Webster, Advisor
Recommended to the Graduate Council
Date __________________ ___________________________
Thomas J. Webster, Reader
Date __________________ ___________________________
James M. Valles Jr., Reader
Date __________________ ___________________________
Derek Stein, Reader
Approved by the Graduate Council
Date _____________ __________________________________
Sheila Bonde, Dean of the Graduate School
iv
Curriculum Vitae
Phong A. Tran
Address: 38 East Street, Floor 3, Providence, RI 02906 Phone: (401) 855-4925 Email: [email protected] Date of birth: 01/04/1981 Place of birth: Bacgiang, Vietnam
EDUCATION Brown University, Providence, RI, USA 2005 - 2010 Ph.D. in Physics Area of focus: Nanostructured selenium for anti-cancer, anti-bacterial applications Advisor: Professor Thomas J. Webster, Division of Engineering and Department of Orthopedics, Brown University Co-Advisor: Professor Derek Stein, Department of Physics, Brown University Brown University M.Sc. in Electrical Sciences and Computer Engineering 2005 - 2009 Hanoi University of Technology, Hanoi, Vietnam 2000 - 2004 B.Sc. in Engineering Physics Honor Training Program Area of focus: Magnetic materials
HONORS AND AWARDS STAR (Student Travel Achievement Recognition) Honorable Mention, Society for Biomaterials (2010) Dissertation Fellowship, Brown University (2009-2010) Research image highlighted on front cover of Journal of Materials Chemistry, 19 (18) (2009) Invited instant insight: Bone repair breakthrough. Chemical Technology (http://www.rsc.org/Publishing/ChemTech/Volume/2009/04/bone_repair_breakthrough.asp) (2009) Paper (“Enhanced osteoblast adhesion on nanostructured selenium compacts for anti-cancer orthopedic applications,” International Journal of Nanomedicince, 3, 391-396) in the top 5% most downloaded for all articles published from that time in the journal (2008) Vietnam Education Foundation Fellowship. 1 of 51 selected from approximately 2000 applicants (2005 - 2010) Conference Travel Awards, awarded by the Vietnam Education Foundation (2007, 2008) Brown University Graduate School Fellowships (2005, 2006) Award for best senior research project, Hanoi University of Technology (2004) Awarded Highest Honors for Honor’s Thesis, Hanoi University of Technology (2004)
v
‘Rencontres du Vietnam’ scholarships (awarded to less than 1% of physics student body) (2001, 2002) Bronze Medal in National Mechanical Olympiad, Vietnam (2001) Hanoi University of Technology undergraduate scholarships (2000 - 2004)
RESEARCH EXPERIENCE Lab of Nanomedicine, Brown University Research Assistant 2006 - 2010 Conceptualize idea of combining chemopreventive properties of selenium and
mechanical properties of conventional orthopedic materials to create composites for anti-cancer orthopedic applications
Fabricate anti-cancer, anti-bacterial biomaterials by three methods: (i) pressing selenium particles into compacts, (ii) coating selenium nanoclusters on materials and (iii) synthesizing selenium nanoparticles
Characterize material structure and surface chemistry by scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM), energy dispersive X-ray spectroscopy (EDS), contact angle measurement, dynamic light scattering (DLS), Zeta potential, and X-ray diffraction (XRD)
Investigate functions of primary osteoblast cells on materials in vitro using cell assays (ELISA, CyQUANT, cell viability (MTT), Live/Dead, ALP activities, calcium deposition and protein synthesis)
Investigate functions of osteosarcoma cells on the materials using assays (Annexin V cell apoptosis, cell cycle arrest), fluorescence microscopy and flow cytometry
Investigate bacterial functions and formation of bacterial biofilms on the materials using Live/Dead assays, confocal microscopes, fluorescence microscopes, microplate readers and SEM
Develop fluorescence microscopy method and image processing programs to track the proliferation of primary osteoblast cells and osteosarcoma cells in co-culture
Create programs in ImageJ and Matlab to eliminate manual cell counting, reduce image processing, data processing time and increase the accuracy in determining cell count and cell spreading. The programs are being used across several labs
Design a multi-pressing mold that can be used to fabricate simultaneously multiple samples from powder or particles. The mold helped reduce sample preparation time approximately 12 times
Train 7 lab members on the mammalian co-culturing system and image/data processing programs
Modify the Cassie-Baxter model of contact angle on surface to quantify contribution of nanometer roughness to air-pocket formation during contact angle measurement
Develop program in Matlab to simulate adsorption of fibronectin on nanometer rough surfaces. The simulation results fit well with experimental results
Lab of Amorphous and Nanocrystalline Materials, Hanoi University of Technology Senior Research Project 2003, 2004 Designed and setup a system to measure giant magneto resistance (GMR) effects Measured GMR effects and hysteresis loops in rapidly quenched Cu-Co alloys
vi
Used Langevin theory of paramagnetism to estimate particle sizes from the measured hysteresis loops
Trained 3 lab members on using the GMR measuring system Supervised a group of two junior students to design a GMR-based magnetic sensor
PAPERS/BOOK CHAPTERS 1. P. A. Tran, L. Sarin, R. H. Hurt and T. J. Webster. Differential effects of nanoselenium doping on healthy and cancerous osteoblasts in co-culture on titanium. International Journal of Nanomedicine, (in press). 2. P. A. Tran, L. Sarin, R. H. Hurt and T. J. Webster. Titanium surfaces with adherent selenium nanoclusters as a novel anti-cancer orthopedic material. Journal of Biomedical Materials Research part A, online ahead of print, DOI: 10.1002/jbm.a.32631 (2009). 3. P. A. Tran, L. Zhang and T. J. Webster. Carbon nanofibers and carbon nanotubes in regenerative medicine. Advanced Drug Delivery Reviews, 61, 1097-111 (2009). 4. P. A. Tran, L. Sarin, R. H. Hurt and T. J. Webster. Opportunities for nanotechnology-enabled bioactive bone implants. Journal of Materials Chemistry, 19, 2653–2659 (2009). 5. P. A. Tran and T. J. Webster. Nanotechnologies for cancer diagnostics and treatment. In Charles J. Dixon and Ollin W. Curtines, editors, Nanotechnology: nanofabrication, patterning, and self assembly. Nova Science Publisher (2009). 6. P. A. Tran and T. J. Webster. Enhanced osteoblast adhesion on nanostructured selenium compacts for anti-cancer orthopedic applications. International Journal of Nanomedicince, 3, 391-396 (2008). 7. N. N. Hoang, P. A. Tran, et al. The influence of heat treatment on magnetoresistance effect in granular Cu-Co alloys prepared by rapid quenching. Adv. In Tech. of Mat. and. Mat. Proc.J (ATM).Vol.6 (1), 83-86 (2005).
CONFERENCE PROCEEDINGS 1. P. A. Tran, L. Sarin, R. H. Hurt and T. J. Webster. Selenium nanocluster coatings: transforming current orthopedic materials into inhibiting bone cancer. Materials Science Forum, 638-642: 718-723, Trans Tech Publications, Switzerland (2010). 2. P. A. Tran, E. Taylor, L. Sarin, R. H. Hurt and T. J. Webster. Selenium nanocluster coatings for anti-cancer, anti-bacterial orthopedic applications. Proceedings of 2009 AIChE Annual Meeting, Nashville, TN (in press). 3. P. A. Tran, L. Sarin, R. H. Hurt and T. J. Webster. Transforming orthopedic biomaterials into bone cancer inhibiting implants: The role of selenium nanoclusters. in Materials in Tissue Engineering, edited by Webster, T.J. (Mater. Res. Soc. Symp. Proc. Volume 1136E, Warrendale, PA (2009).
CONFERENCE PRESENTATIONS 1. P. A. Tran, E. Taylor, L. Sarin, R. H. Hurt and T. J. Webster. Novel anti-cancer, anti-bacterial coatings for biomaterial applications: selenium nanoclusters. 2009 Materials Research Society Meeting, Boston, MA (2009) 2. P. A. Tran, L. Sarin, R. H. Hurt and T. J. Webster. Imparting anti-cancer properties to orthopedic materials: The role of selenium nanoclusters. Society for Biomaterials 2009 Annual Meeting, San Antonio, TX (2009).
vii
3. P. A. Tran, L. Sarin, R. H. Hurt and T. J. Webster. Selenium nanocluster coatings for anti-cancer orthopedic applications. 35th Northeast Bioengineering Conference, Boston, MA (2009). 4. P. A. Tran, L. Sarin, R. H. Hurt and T. J. Webster. Promising orthopedic materials for bone cancer patients: titanium coated with selenium nanoclusters. 34th Northeast Bioengineering Conference, Providence, RI (2008). 5. P. A. Tran, L. Sarin, R. H. Hurt and T. J. Webster. Nanoselenium cluster coatings for anti-cancer orthopedic applications. Society for Biological Engineering’s 4th International Conference on Bioengineering and Nanotechnology Dublin, Ireland (2008). 6. P. A. Tran and T. J. Webster. Promising anti-cancer orthopedic material: Nanostructured selenium. Biomedical Engineering Society Fall Meeting, Los Angeles, CA (2007). 7. P. A. Tran and T. J. Webster. Novel anti-cancer orthopedic materials: Nanostructured selenium. 33rd Northeast Bioengineering Conference, Stony Brook, NY (2007).
COMPUTER SKILLS ImageJ, Image Pro Plus, Matlab, Assembly, Pascal, Microsoft Excel, Microsoft Word, Microsoft PowerPoint
viii
ACKNOWLEDGEMENTS
I am greatly grateful to my advisor, Professor Thomas J. Webster who has guided
me and provided financial support to me through my academic career at Brown
University. I am also grateful to my co-advisor, Professor Derek Stein who has given me
advice on my research and also on the career track in general. I deeply appreciate
Professor James M. Valles, Jr.’s interest in my research and serving in my preliminary
exam and thesis defense committees, helping me to finish my graduate career. I would
like to thank all members in the Laboratory of Nanomedicince Research for their support
and help during my time in the laboratory.
Most of all, I thank my parents who gave birth to me, raise me, love me and guide
me to the career path today. I also thank other members of my family, my sister, my
brother, who are always a constant source of inspiration for me to overcome challenges in
my studies.
I thank the Vietnam Education Foundation for their partial financial support for
my studies at Brown University.
ix
TABLE OF CONTENTS PAGE
LIST OF TABLES ............................................................................................................ xi LIST OF ILLUSTRATIONS............................................................................................ xii CHAPTER 1. INTRODUCTION ....................................................................................... 1
1.1. Bone cancers ............................................................................................................ 1 1.2. Implant bacterial infection ....................................................................................... 1 1.3. Selenium as an anti-cancer, anti-bacterial material ................................................. 4 1.4. Focus of thesis.......................................................................................................... 6
CHAPTER 2. SELENIUM NANOCLUSTER COATINGS ON METALS FOR ANTI-CANCER ORTHOPEDIC MATERIALS .......................................................................... 7
2.1. Introduction.............................................................................................................. 7 2.1.1. Problems with currently used orthopedic implants........................................... 7 2.1.2. The role of nano-structured surfaces on orthopedic implants........................... 8 2.1.3. Nano-structured selenium coatings: a novel approach of using selenium to create anti-cancer, anti-bacterial biomaterials ............................................................ 8
2.2. Materials and methods ............................................................................................. 9 2.2.1. Materials ........................................................................................................... 9 2.2.2. Material characterization ................................................................................ 11 2.2.3. In vitro biological assays for uncoated and selenium coated metallic substrates................................................................................................................................... 14 2.2.4. Fibronectin adsorption on uncoated and selenium coated titanium substrates19 2.2.5. Statistical analysis........................................................................................... 20
2.3. Results.................................................................................................................... 20 2.3.1. Surface characterization.................................................................................. 21 2.3.2. Cell adhesion and proliferation....................................................................... 28 2.3.3. Healthy osteoblast differentiation ................................................................... 35 2.3.4. Selenium release ............................................................................................. 37 2.3.5. Effects of selenium released into culture media on healthy and cancerous cells................................................................................................................................... 38 2.3.6. Intracellular thiol assays ................................................................................. 40 2.3.7. Fibronectin adsorption on uncoated and selenium-coated titanium substrates41
2.4. Results on stainless steel samples .......................................................................... 45 2.4.1. Material characterization ................................................................................ 45 2.4.2. Cell experiments on stainless steel substrates................................................. 48
2.5. Conclusions and summary ..................................................................................... 52 CHAPTER 3. COARSE - GRAINED MONTE CARLO COMPUTER SIMULATION OF FIBRONECTIN ADSORPTION ON NANOMETER ROUGH SURFACES .......... 54
3.1. Introduction............................................................................................................ 54 3.2. Protein adsorption onto biomaterials ..................................................................... 55 3.3. Roles of rough surfaces.......................................................................................... 58 3.4. Simple method to generate rough surfaces with pre-defined RMS roughness...... 59 3.5. RSA model of fibronectin adsorption on uncoated and selenium coated titanium surfaces ......................................................................................................................... 64 3.6. Summary and conclusions ..................................................................................... 76
x
CHAPTER 4. A MODIFIED CONTACT ANGLE MODEL TO FIT EXPERIMENTAL RESULTS ......................................................................................................................... 78
4.1. Introduction: role of contact angles in biomaterials .............................................. 78 4.2. Contact angle and Young’s equation ..................................................................... 81 4.3. Wenzel model versus Cassie-Baxter model on contact angles.............................. 82 4.4. The needs to modify the Wenzel model and Cassie-Baxter model to fit experimental results ...................................................................................................... 88 4.5. Modification of the Cassie-Baxter model .............................................................. 92 4.6. Summary and conclusions ..................................................................................... 96
CHAPTER 5. SELENIUM COATINGS ON POLYMERS FOR ANTI-INFECTION BIOMATERIAL APPLICATIONS.................................................................................. 97
5.1. Introduction............................................................................................................ 97 5.2. Materials and methods ........................................................................................... 99
5.2.1. Materials ......................................................................................................... 99 5.2.2. Material characterization .............................................................................. 100 5.2.3. In vitro assays for the polymeric substrate-bacterial experiments................ 101 5.2.4. Statistical analysis......................................................................................... 103
5.3. Results and discussion ......................................................................................... 103 5.3.1. SEM and EDS............................................................................................... 103 5.3.2. S. aureus response on uncoated and selenium-coated polymeric substrates 107 5.3.3. Thiol assays................................................................................................... 110
5.4. Discussion ............................................................................................................ 111 5.5. Conclusions.......................................................................................................... 112
CHAPTER 6. SELENIUM NANOPARTICLES FOR ANTI-BACTERIAL APPLICATIONS ............................................................................................................ 114
6.1. Introduction.......................................................................................................... 114 6.2. Materials and methods ......................................................................................... 115
6.2.1. Material synthesis ......................................................................................... 115 6.2.2. Material characterization .............................................................................. 115 6.2.3. Bacteria assays .............................................................................................. 116 6.2.4. Superoxide anion assays ............................................................................... 118 6.2.5. Statistical analysis......................................................................................... 119
6.3. Results.................................................................................................................. 119 6.3.1. TEM, DLS and XPS results on size distribution and oxidation state of selenium nanoparticles............................................................................................ 119 6.3.2. Effects of selenium nanoparticles on S. aureus in solution .......................... 122 6.3.3. Intracellular thiol assays ............................................................................... 122 6.3.4. Superoxide assays ......................................................................................... 124
6.4. Summary and conclusions ................................................................................... 125 CHAPTER 7. CONCLUSIONS ..................................................................................... 127
xi
LIST OF TABLES PAGE Table 2.1. Reagent volumes used in the colloidal synthesis of Se nanoclusters in the
presence of various base substrates............................................................................. 10 Table 2.2. Fitting parameter and the goodness of fitting (R2) for growth curve and
calculated doubling time of healthy osteoblasts cultured on uncoated and selenium-coated titanium substrates ........................................................................................... 33
Table 2.3.Fitting parameter and goodness of the fitting (R2) for growth curve and calculated doubling time of cancerous osteoblasts cultured on uncoated and selenium-coated titanium substrates. .......................................................................................... 35
Table 3.1. Convergence property of the proposed method to generate surface High-nSe-Ti with a defined RMS roughness. Dimensionless simulated roughness, Ro, is the square root of variance of the heights of all points on the simulated surface. Simulated roughness is then determined as R=RoDo (in this case, Do=9nm)............................... 63
Table 3.2. Parameters for simulation (a) calculated from experimental RMS roughness as described in section 3.5 and the RMS roughness of the simulated surfaces. The simulated surfaces resemble the experimental surfaces quite well in terms of RMS roughness. ................................................................................................................... 69
Table 4.1. The selenium surface fraction (ρ, determined from SEM images), relative surface roughness ratio (r, normalized by that of the uncoated substrates), the relative root mean square roughness (R, normalized to that of uncoated substrates), and cos(θ), of the substrates of interest to the studies. ...................................................... 90
Table 6.1. Superoxide anion assay reagent mixtures...................................................... 118
xii
LIST OF ILLUSTRATIONS PAGE Figure 1.1. A scanning electron microscopy image of a Staphylococcus biofilm on the
inner surface of a catheter showing bacteria and polymeric matrix (reprinted with permission from [11]). .................................................................................................. 3
Figure 2.1. Image of DSA100 (Updated Version of DSA-10, Kruss, Germany: Image Adopted from www.kruss.com). Capability of equipment: The range of contact angle measurement is between 0 and 180° with a resolution of ±0.1 and a range of surface tension between 10-2 and 100mJ/m2, with a resolution of 0.01mJ/m2. ....................... 13
Figure 2.2. Representative SEM images of: (A) uTi; (B) Low-nSe-Ti; (C) Medium-nSe-Ti and (D) High-nSe-Ti. The surface area coverage of selenium on the selenium coated substrates were determined (using an image processing program, ImageJ) to be 2.7%, 5.1% and 7.5% for Low-nSe-Ti, Medium-nSe-Ti and High-nSe-Ti, respectively. ................................................................................................................ 21
Figure 2.3. SEM image, taken at a 45° tilt, of High-nSe-Ti showing the hemispherical shape of the selenium nanoclusters on the titanium surface. ...................................... 22
Figure 2.4. Morphology of a selenium-coated titanium surface appeared the same before (A) and after sonication (B) and UV radiation (C). Bars=500 nm. ............................ 23
Figure 2.5. All coated substrates showed peaks of selenium that were not detected in uncoated substrate....................................................................................................... 24
Figure 2.6. XPS profiles of High-nSe-Ti showing the characteristic binding energy peak for elemental selenium at 55.2 eV. ............................................................................. 24
Figure 2.7. Representative AFM images and line analysis of uTi (A), Low-nSe-Ti (B), Medium-nSe-Ti (C) and High-nSe-Ti (D). Nano-scale roughness was created by the selenium coatings and increased with increasing selenium coating density............... 26
Figure 2.8. RMS of the substrates increased with increasing selenium coating density. Data = mean ± standard error of the mean; N=3; * p<0.05 compared to all other substrates; ** p<0.01 compared to Medium-nSe-Ti and High-nSe-Ti....................... 27
Figure 2.9. Water contact angles on the uncoated and coated Ti substrates. Contact angles increased on the substrates coated with selenium nanoclusters. Data = mean ± standard error of the mean; N=3; * p<0.05 compared to all the coated substrates. There was no significant difference among the contact angles on the coated substrates. ........................................................................................................ 28
Figure 2.10. Increased healthy osteoblast densities after 4 hrs, 1 day and 3 days. Data = mean ± standard error of the mean; N=3, * p<0.05 compared to Medium-nSe-Ti (compared at same time period), ** p<0.05 compared to uTi (compared at same time period), *** p< 0.05 compared to Low-nSe-Ti (compared at same time period), # p<0.05 compared to uTi (compared at same time period), & p =0.06 compared to Low-nSe-Ti. (compared at same time period). ........................................................... 29
Figure 2.11. Cancerous osteoblast densities after 4 hrs, 1 day, and 3 days. Decreased cancerous osteoblast densities on selenium coated titanium substrates after 4 hrs, 1 day and 3 days. Data = mean ± standard error of the mean; N=3, * p<0.05, ** p< 0.01 compared to High-nSe-Ti. (Compared at same time period); # p<0.1 compared to Medium-nSe-Ti. (Compared at same time period). There are no significant differences in cell densities among substrates after 1 day. ......................................... 30
xiii
Figure 2.12. Representative fluorescence microscopy images of healthy osteoblasts on: (A) uTi; (B) Low-nSe-Ti; (C) Medium-nSe-Ti and (D) High-nSe-Ti after 1 day (Magnification = 10X). Scale bars = 200 μm. ............................................................ 31
Figure 2.13. Representative fluorescence microscopy images of cancerous osteoblasts on: (A) uTi; (B) Low-nSe-Ti; (C) Medium-nSe-Ti and (D) High-nSe-Ti after 3 days (Magnification = 20X). Scale bars = 100 μm. ............................................................ 32
Figure 2.14. Fitting exponential growth curves to experiment data for densities of healthy osteoblasts cultured on uncoated titanium and on selenium-coated titanium substrates. Growth curves of healthy osteoblasts cultured on higher selenium coating densities had large slopes........................................................................................................... 33
Figure 2.15. Fitting exponential growth curves to experiment data for densities of cancerous osteoblasts cultured on uncoated titanium and selenium-coated titanium substrates. Curves of cancerous osteoblasts cultured on lower selenium coating densities had large slopes with the uncoated titanium substrates having the largest slope. ........................................................................................................................... 34
Figure 2.16. Increased alkaline phosphatase (ALP) activity of osteoblasts on High-nSe-Ti compared to all other substrates. Data = mean ± standard error of the mean; N=3, * p<0.01 compared to uTi, Low-nSe-Ti and Medium-nSe-Ti.................................... 36
Figure 2.17. Extracellular calcium deposition by osteoblasts on uTi, Low-nSe-Ti, Medium-nSe-Ti and High-nSe-Ti after 14 days. Data = mean ± standard error of the mean; N=3, * p<0.05 compared to uTi. ...................................................................... 37
Figure 2.18. Decreased thiol content of healthy osteoblasts cultured in supernatant from High-nSe-Ti compared to uTi. Data = mean ± standard error of the mean; N=3; * p<0.1 ........................................................................................................................... 40
Figure 2.19. Decreased thiol content of cancerous osteoblasts cultured in supernatant from High-nSe-Ti compared to uTi. Data = mean ± standard error of the mean; N=3; * p<0.05. ..................................................................................................................... 41
Figure 2.20. Increased fibronectin adsorption from DMEM onto Medium-nSe-Ti and High-nSe-Ti substrates. DMEM (supplemented with 10% FBS and 10% P/S) was estimated to have fibronectin concentration of 3µg/mL. Each sample was immersed in 0.5 mL of DMEM solution for 24 hrs under standard condition (37oC, 5% CO2, 95% humidified air) and amount of absorbed fibronectin was determined via optical density of ABTS solution using ELISA method. A standard curve relating optical density to amount of fibronectin was used to determine fibronectin concentration from optical densities of ABTS solution. Finally, amount of fibronectin was normalized to geometrical surface areas of the substrates. Data = mean ± standard error of the mean; N=3, * p<0.05 compared to uTi, ** p<0.05 compared to uTi and Low-nSe-Ti. ................................................................................................................ 43
Figure 2.21. Increased fibronectin adsorption from fibronectin solution (in PBS) on Medium-nSe-Ti and High-nSe-Ti substrates. Each sample was immersed in 0.5 mL of 5 µg/mL fibronectin solution for 24 hrs under standard condition (37oC, 5% CO2, 95% humidified air) and amount of absorbed fibronectin was determined via optical density of ABTS solution using ELISA method. A standard curve relating optical density to amount of fibronectin was used to determine fibronectin concentration from optical densities of ABTS solution. Finally, amount of fibronectin was normalized to geometrical surface areas of the substrates. Data = mean ± standard
xiv
error of the mean; N=3, * p<0.05 compared to uTi and Low-nSe-Ti. ** p=0.055 compared to Medium-nSe-Ti...................................................................................... 44
Figure 2.22. SEM images of stainless steel substrates: (A) uncoated; (B), (C), and (D) coated with Se concentration in the final solution which equaled 0.42 mM, 0.83 mM ,and 1.25 mM, respectively. Some Se nanoclusters are indicated by arrows. These clusters are approximately 80nm in diameters. Bars = 1µm....................................... 46
Figure 2.23. Water contact angles on stainless steel substrates. Selenium nanocluster coating makes the stainless steel substrates become more hydrophobic. ................... 47
Figure 2.24. Maintained healthy osteoblast densities on selenium-coated stainless steel substrates compared to the uncoated substrate after 3 days of culturing. Data = mean ± standard error of the mean; N=3. ....................................................... 48
Figure 2.25. Representative fluorescence images of healthy osteoblast densities on uSS (A), Low-nSe-SS (B), Medium-nSe-SS (C) and High-nSe-SS (D). Magnification = 10X.............................................................................................................................. 49
Figure 2.26. Decreased cancerous osteoblast densities after 3 days of culturing on Se coated stainless steel. Cell densities on Medium-nSe-SS and High-nSe-SS were significantly less than those on uSS and Low-nSe-SS. * p<0.05 compared to uSS and Low-nSe-SS; ** p<0.01 compared to Medium-nSe-SS. Data = mean ± standard error of the mean; N = 3. ..................................................................................................... 50
Figure 2.27. Representative fluorescence images of cancerous osteoblast densities on uSS (A), Low-nSe-SS (B), Medium-nSe-SS (C) and High-nSe-SS (D). Magnification = 20X.............................................................................................................................. 51
Figure 3.1. Schematic structure of fibronectin showing two monomers linked by disulfide bonds (Reprinted with permission from [72]). ........................................................... 57
Figure 3.2. Representative regions on the four rough surfaces generated in this study with the RMS roughness values equal to experimental values (unit lengh Do=9nm, N=500). Colors are to aid in visualization only.......................................................... 64
Figure 3.3. Gallery of transmission electron microscopy images of individual fibronectin molecules freeze-dried in vacuum and shadowed by tungsten-tantalum. (Freeze-drying technique was used to preserve the native structure of fibronectin and tungsten-tantalum shadowing was used to give contrast for imaging). Most fibronectin molecules had elongated shapes with the length of ~ 15 nm and the width of ~9 nm. Magnification is 300,000. (Adapted with permission from [98]) . ........... 66
Figure 3.4. A 2D picture of fibronectin adsorbing onto a rough surface. Fibronectin was modeled as a dimmer consisting of 2 monomers and the surface was modeled as consisting of columns. Each point on the surface was defied by two coordinates (i,j). Fibronectin can adsorb on all the "unoccupied" sites (in this picture, the right and the left sides of point (i,j) are unoccupied, the top and bottom sides of point (i,j) are occupied)..................................................................................................................... 67
Figure 3.5. Examples of the simulation of protein adsorption onto a rough surface. For the case of the right side of the point specified by (i,j) (the bold side), there are two possible directions the protein can adsorb: horizontal direction (a) or vertical direction (b) (Note, horizontal adsorption to the left of the bold side is prohibited due to steric constraints). In this case, configuration (a) will have 4 contacts while (b) will have only 2 contacts upon adsorption, therefore, configuration (a) is more favorable for protein adsorption. For the case of the point on the far right (bold side), both the (c)
xv
and the (d) configuration will have the same number of contacts (3 contacts) upon adsorption, therefore, the protein will adsorb following either (c) or (d) with equal probability. .................................................................................................................. 68
Figure 3.6. Overhang is not allowed. Overhang would be immediately shifted downward as indicated by the dashed arrows............................................................................... 69
Figure 3.7. A snapshot of simulation process in 2D with N=500 and M=25000. Blue columns represent proteins. Grey columns represent rough surfaces. The adsorption did not reach saturation yet as some areas on the surface are still unoccupied. ......... 71
Figure 3.8. Simulation results of fibronectin saturated adsorption on all four substrates of interest. The protein amount was normalized to a monolayer of protein on a perfectly flat surface with the same geometrical surface area N x N. Substrates were generated to have experimental RMS roughness values. Protein adsorption simulation was implemented with the size of the substrates N x N=500x500 and the number of simulation M=6,250,000. Data = mean ± standard error of the mean (n=3 runs). ..... 72
Figure 3.9. Comparison of the increase in the amount of fibronectin adsorption (expressed as times increase compared to uTi) between simulation results and experimental results on uncoated and selenium coated titanium substrates. .............. 73
Figure 3.10. Simulation and experiment results on the amount of adsorbed fibronectin normalized by geometrical surface area. The simulation values of normalized fibronectin adsorption are greater than experimental values. ..................................... 74
Figure 3.11. Protein adsorption on a flat surface and a rough surface. Assuming the same "binding site" densities, the protein density (per unit surface area) d1 and d2 should be the same. ..................................................................................................................... 75
Figure 3.12. Computer simulation results on fibronectin density expressed as microgram of fibronectin per cm2 of the surface area. Surfaces with higher RMS roughness had lower fibronectin density. ........................................................................................... 75
Figure 3.13. An example illustrating that increased surface area does not necessarily lead to equal increase in protein adsorption. In this example, the rough surface area increases ~67% but the amount of fibronectin increases only 25%. .......................... 76
Figure 4.1. Atomic and ribbon chains of an extended human fibronectin molecule. Inset shows that hydrophilic parts are blue colored regions while hydrophobic parts are red colored in modules (Reprinted with permission from [100]). .................................... 80
Figure 4.2. Contact angle, θ, of a liquid on a solid surface depends on the interaction of
the liquid with the solid, liquid with vapor and solid with vapor. SL , SV , and LV are interfacial energies at the solid-liquid interface, solid-vapor interface and liquid-vapor interface, respectively. ................................................................................................ 81
Figure 4.3.Schematic demonstration of the derivation of Young's equation. When the solid-liquid interface is displaced by dx, the increase in liquid-vapor interface is dx.cos(θ)...................................................................................................................... 82
Figure 4.4. Schematic demonstration of the derivation of the Wenzel equation. When the solid-liquid interface is displaced by dx, the increase in liquid-vapor interface is dx.cos(θ), while the actual liquid-solid interface increase associated with this displacement is r.dx. ................................................................................................... 84
Figure 4.5. Homogenous wetting assumption in the Wenzel model. The liquid-solid interface is assumed to follow a solid surface. ........................................................... 85
xvi
Figure 4.6. A schematic demonstration of the derivation of the Cassie-Baxter model. When the solid-liquid interface increases by dx, the liquid-vapor interface increases by dx.cos(θ)................................................................................................................. 86
Figure 4.7. The Cassie-Baxter model of heterogeneous wetting. In this case, the liquid is considered to contact a surface consisting of the solid material and air. The width of spikes as well as the spacing between them is very small compared to the size of the liquid drop. The scale used in this picture is to aid in visualization. .......................... 87
Figure 4.8.Water contact angles on uncoated titanium and titanium coated with increasing selenium nanocluster densities. Data = mean ± standard error of the mean, n=5. ..... 89
Figure 4.9.The Cassie-Baxter model curve (dashed line) and the experimental curve (solid line) of the cosine of water contact angles on uncoated titanium substrates and titanium substrates coated with various amount of selenium nanoclusters. ............... 91
Figure 4.10. Possible mechanisms for the plateau in contact angles on the substrates with different selenium coating densities. Increasing coating density of selenium nanoclusters increased roughness that resulted in both increases in the roughness factor (which is the actual surface area to geometric area ratio) and the increase in the formation of air-pockets. The former made contact angles decrease while the latter made contact angles increase. The opposite contribution of the roughness factor and air-pockets is likely the mechanism for similar contact angles for the three selenium coating densities on titanium in here. ......................................................................... 92
Figure 4.11. Schematic representation of the modified Cassie-Baxter model on the wetting process. A drop of liquid makes contact to all three materials: titanium, selenium nanoclusters, and air when placed on a selenium-titanium composite surface. ........................................................................................................................ 93
Figure 5.1. SEM images of uncoated (left panel) and selenium (Se)-coated substrates (right panel). Arrows indicate selenium nanoclusters. The selenium nanoclusters had sizes ranging from approximately 80nm to 200nm and were uniformly coated on the substrates................................................................................................................... 104
Figure 5.2. EDS spectra (right panel) of an uncoated region (top) and a coated region (bottom) on a selenium-coated silicone substrate. Selenium peaks were detected in the coated region demonstrating that the clusters were selenium. Peaks of gold were from the sputter-coated layer. ............................................................................................ 105
Figure 5.3. EDS spectra (right panel) of an uncoated region (top) and a coated region (bottom) on a selenium-coated PU substrate. Selenium peaks were detected in the coated region demonstrating that the clusters were selenium. Peaks of gold were from the sputter-coated layer. ............................................................................................ 106
Figure 5.4. EDS spectra (right panel) of an uncoated region on selenium-coated PVC (top) and a coated region on selenium-coated PVC (bottom). Selenium peaks were detected in the selenium-coated PVC demonstrating that the clusters were selenium. Peaks of gold were from the sputter-coated layer..................................................... 107
Figure 5.5. Representative SEM images showing reduced S. aureus colonization on selenium-coated PU compared to uncoated PU after 8 hrs of inoculation. Arrows indicate cells.............................................................................................................. 108
Figure 5.6. Representative SEM images showing reduced S. aureus colonization on selenium-coated PVC compared to uncoated PVC. Arrows show bacteria. ............ 108
xvii
Figure 5.7. Representative SEM images showing reduced S. aureus colonization on selenium-coated compared to uncoated silicone. Arrows indicate cells................... 108
Figure 5.8. Decreased S. aureus densities on polymeric substrates coated with selenium nanoclusters. Data = mean ± standard error of the mean, N=3; *p<0.05 compared to the uncoated substrate (compared in the same material group), ** p<0.05 compared to silver-coated PVC. .................................................................................................... 109
Figure 5.9. Representative SEM images of S. aureus colonization on silver-coated and Se-coated PVC. Bacterial colonization on Se-coated PVC was reduced compared to silver-coated PVC after 8 hrs of incubation.............................................................. 110
Figure 5.10. Decreased intracellular thiol level in S. aureus cultured in the supernatant from selenium-coated PVC compared to control. Data = mean ± standard error of the mean; N=3; * p<0.05 compared to the control. ........................................................ 111
Figure 6.1. TEM image of selenium nanoparticles stabilized in BSA and dispersed in water. Selenium nanoparticles had average sizes of approximately 100 nm............ 120
Figure 6.2. Size-distribution profile of selenium nanoparticles in solution as measured by the dynamic light scattering technique. Particle sizes centered around 100 nm....... 121
Figure 6.3. XPS profile of selenium nanoparticles. Peak at 55.2 eV confirmed a zero oxidation state of selenium. ...................................................................................... 121
Figure 6.4. Inhibited growth of S. aureus in the presence of selenium nanoparticles at all three selenium nanoparticles concentrations: 12.5 µg/mL, 6.2 µg/mL and 3.1 µg/mL at 4 hrs, 12 hrs and 24 hrs. Data = mean ± standard error of the mean, N=3. *p<0.05 compared to bacteria treated with 6.2 µg Se/mL, 3.1 µg Se/mL and control (0 µg Se/mL) (compared at same time period); ** p<0.05 compared to bacteria treated with 3.1 µg Se/mL and control (0 µg Se/mL) (compared at same time period); ***p<0.05 compared to control (compared at same time period); #p<0.05 compared to control (compared at same time period). Dose-dependent inhibition was observed at 4hrs and 12hrs.......................................................................................................................... 122
Figure 6.5. Decreased intracellular thiol levels in S. aureus cultured with selenium nanoparticles compared to the control (i.e., S. aureus cultured in TSB without selenium nanoparticles). Bacteria were cultured at a density of 50,000 cells/mL for 4 hrs in either TSB or TSB added with selenium nanoparticle (at a concentration of 3.1µg/mL). Data = mean ± standard error of the mean; N=3; * p< 0.05........... 123
Figure 6.6. Superoxide anion assays for selenium particles (nSe). No significant difference was found among the test tubes. Selenium nanoparticles were mixed with GSH and luminol was used as the probes for superoxide anions. Superoxide anions oxidize luminol in a reaction that produces photons. An enhancer was used to amplify the chemiluminescence. Data = mean ± standard error of the mean, N=3. .............. 125
1
CHAPTER 1. INTRODUCTION
1.1. Bone cancers
It is estimated that 2,380 individuals (1,270 men and 1,110 women) will be
diagnosed with bone and joint cancer and 1,470 individuals will die from primary bone
and joint cancer in 2008 in the U.S. [1]. Primary bone cancer is rare as usually bone
cancer is a result of the spread of cancer from other organs (such as the lungs, breasts and
the prostate [2]). Because many deaths are officially attributed to the original cancer
source, the true numbers of bone-cancer deaths are underreported. A common technique
to treat bone cancer is the surgical removal of the cancerous tissue followed by insertion
of an orthopedic implant to restore patient function. Sometimes cancerous cells are not
completely removed, therefore, the remaining cancerous cells will proliferate and cancer
can reoccur at the implant site. In these cases, it would clearly be beneficial to have
implants specifically designed to not only promote healthy bone tissue growth but also to
prevent the reoccurrence of bone cancer.
1.2. Implant bacterial infection
Bacterial infection is a problem for all implants. For example, infection is one of
the most common causes for failure of a hip implant, responsible for 14% of the total
number of revision surgeries [3]. Another example of implant infection relates to
catheters which is the most common and complicated problem associated with catheter
2
usage. Infections are the most serious complication of tunneled dialysis catheters,
resulting in serious systemic infections, including endocarditis, osteomyelitis, epidural
abscess, septic arthritis, and even death [4]. Infections lead to implant failure, extended
hospital stay, and additional treatment/surgeries. Bacteria infect up to 54% of all catheters
[5] and cause many serious complications including patient death. For example, catheter
infection is associated with a mortality rate of 12% to 25% among critically ill patients
[6]. Catheter-associated urinary tract infection (CAUTI) is the most common type
(accounting for 40%) of hospital-acquired infection (“nosocomial infection”) resulting in
serious complications such as bloodstream infection, and even death [7]. Each year, in
U.S. acute-care hospitals and extended-care facilities, CAUTI affects approximately 1
million patients who then will have increased institutional death rates [8]. Chronic
indwelling urinary catheters also increase the risk of infection, accounting for 80% of all
nosocomial urinary tract infections [9]. Significantly, 14 % of deaths in people
undergoing dialysis in 1996 were due to infection [5]. The cost incurred by infections in
the U.S. is nearly $11 billion annually [10].
Implant associated infections are difficult to treat because of biofilm formation.
After implantation, bacteria (from the patient’s own skin, hospital personnel or
equipment) quickly attach and adhere irreversibly to the implant surface, secreting a
polymeric-like substance (composed of mostly polysaccharides) and form a biofilm
which consists of bacteria and a polymeric-like matrix (Figure 1.1).
3
Figure 1.1. A scanning electron microscopy image of a Staphylococcus biofilm on the inner surface of a catheter showing bacteria and polymeric matrix (reprinted with permission from [11]).
Bacteria in a biofilm can escape from the film and enter the blood, lungs, etc.,
causing serious problems. Biofilms tenaciously bind to surfaces. More importantly,
bacteria in biofilms are extremely resistant to antibiotic treatment due to the slow
transport of antibiotic molecules through the polymeric-like biofilm substance, altered
micro-environment within the biofilm and higher number of “persister” cells (cells that
are resistant to many types of stress) within the biofilm [compared to planktonic (free-
floating) cells] [12].
Among the most common pathogens found on infected implants are S.
epidermidis (comprising 24% of all bacteria for catheter infections and 15% for
orthopedic implant infections), S. aureus (20% for catheter infections, 35% for
orthopedic implant infections), and P. aeruginosa (25% for catheter infection) [11, 13,
14].
4
1.3. Selenium as an anti-cancer, anti-bacterial material
Selenium belongs to the group of metalloids, which are neither fully metals nor
non-metals but share characteristics of both [15]. In the periodic table of elements,
selenium is in the same column with oxygen and sulfur, and in the same row with
germanium, arsenic and bromine. The outermost electron configuration of selenium is
4s23d104p4. Selenium is naturally found in humans and animals and selenoproteins are
important proteins in antioxidant defense systems, thyroid hormone metabolism and
redox control of cell reactions [16]. Animal studies have shown that selenium intake in
excess of the nutritional requirement can inhibit and/or retard carcinogenesis [17].
Moreover, studies have shown that people in areas of low soil selenium (lower than 0.05
ppm) and people with decreased plasma selenium levels (below 128 ng/ml) have higher
cancer incidence and/or cancer mortality [18, 19]. High levels of selenium in the blood
(~154 µg/ml) have been correlated with reduced numbers of cancers including
pancreatic, gastric, lung, nasopharyngeal, breast, uterine, respiratory, digestive,
hematological and gynecological [20]. The strongest evidence for the effect of selenium
in reducing cancers was shown for lung cancer (46% lower incidence) [21], esophageal
and gastric-cardiac cancers [22] and especially prostate cancer (63% lower incidence)
[23, 24].
In vitro research has also shown the inhibitory effects of selenium on the growth
of many cancerous cell lines [25-29]. Apostolou and co-workers [30] added selenium (in
the form of either sodium selenite (Na2SeO3), selenomethionine (SeMet), or
selenocysteine (Sec)) to culture media and found a concentration-dependent inhibition of
malignant mesothelioma (MM) cells but normal mesothelial cells were not affected. They
5
also showed that selenium induced apoptosis in MM cells by a mechanism related to
SEP15 which is a gene encodinga 15-kDa selenium containing protein. Menter and
colleagues studied the effect of sodium selenite or SeMet on the growth of normal
primary prostate cells and prostate cancer cells [31] by adding either sodium selenite or
SeMet to culture media. They observed a dose-dependent growth inhibition and apoptosis
in the cancer cells treated with either sodium selenite or SeMet. Meanwhile, growth of
the normal cells was not suppressed by the selenium treatment. Narayan and co-workers
demonstrated that selenium (in the form of sodium selenite) caused growth inhibition and
apoptosis in human brain tumor cells. Interestingly, in spite of a great number of studies
on the effect of selenium on cancer, there are very few experiments focusing on the effect
of elemental selenium on cancer growth, especially the use of selenium in anti-cancer
applications. The mechanisms of selenium-based chemoprevention are also complex and
incompletely understood [32]. Selenium was shown to inhibit angiogenesis which is the
growth of blood vessels around the tumor without which the tumor will not be able to
grow [33]. Selenium was also demonstrated to induce apoptosis, which is a natural-cell-
death process [31, 34]. Evidence also exists to show that selenium produces superoxide
anions that kill cancer cells [35].
In the meantime, studies have also provided evidence of the anti-bacterial
properties of many selenium compounds. For example, selenium-enriched probiotics
have been shown to strongly inhibit the growth of pathogenic E. coli in vivo and in vitro
[36]. Specially, selenium-enriched probiotics (the selenium concentration was
approximately 0.5 µg/mg) were grown and co-cultured with E. coli solutions and were
shown to inhibit the growth of E. coli approximately 10 times compared to the control
6
(no probiotic treatment). In addition, mice fed with selenium-enriched probiotics had the
lowest mortality rate upon exposure to pathogenic E. coli [36]. A series of
organoselenium compounds (such as 2,4,6-tri-para-methoxyphenylselenopyrylium
chloride, 9-para-chlorophenyloctahydroselenoxanthene, perhydroselenoxanthene , and 9-
para-fluorophenyloctahydroselenoxanthene) have been synthesized and have shown anti-
bacterial activities in vitro, especially against S. aureus [37-39]. For example, all
organoselenium compounds synthesized and tested in [38] were shown to be as effective
as penicillin, a common antibiotic, in inhibiting S. aureus growth in solution in vitro. The
anti-bacterial properties of selenium are believed to be attributed to the ability of
selenium to generate superoxide radicals [35] and/or to catalyze oxidation of intracellular
thiol causing thiol depletion that leads to cell death [40].
1.4. Focus of thesis
This thesis is focused on exploring the potential of using nanometer scale
structured selenium for anti-cancer and anti-bacterial applications. The thesis is organized
as follows. Chapter 2 will describe selenium coatings on metallic orthopedic materials
(titanium and stainless steel) with a focus especially on titanium which is the most
common orthopedic material. Chapter 3 will propose a computer simulation method to
reproduce the experimental results on protein adsorption presented in Chapter 2. Chapter
4 will propose a modified contact angle model to fit the experimental results presented in
Chapter 2. Chapter 5 will describe the potential of using selenium coatings for anti-
bacterial applications. Chapter 6 will describe the use of selenium nanoparticles for anti-
bacterial applications. Finally, a conclusion chapter will summarize the content of the
thesis in the biomaterials field at large.
7
CHAPTER 2. SELENIUM NANOCLUSTER COATINGS ON METALS FOR ANTI-CANCER ORTHOPEDIC MATERIALS
2.1. Introduction
2.1.1. Problems with currently used orthopedic implants
The most commonly used materials for bone implants are metals (such as
titanium, stainless steel) and polymers (such as ultra high molecular weight polyethylene)
[41]. Several problems exist with current generation implants including: (i) insufficient
bonding between the implanted material and juxtaposed bone [42-45], (ii) different
mechanical properties between bone and the implant leading to stress shielding [43-45],
and (iii) wear debris generated at articulating surfaces of orthopedic implants that may
lead to cell and bone death [44, 45]. Insufficient new bone growth on implants is the main
cause of insufficient bonding between the implant and surrounding bone. Therefore, one
of the most important characteristics that the new generation of orthopedic implant
materials should have is better biocompatibility to support faster new bone growth. The
above-mentioned problems with implants usually lead to implant failure which results in
surgery to remove the failed implant and insertion of a new one (i.e., revision surgery).
Revision surgery is not only painful for patients but also costly. The number of revision
surgeries due to implant failure has increased steadily over the last decade and is
expected to continue increasing into the next decade.
8
2.1.2. The role of nano-structured surfaces on orthopedic implants
It is well known that tissues in the body possess nano-scale features. For example,
hydroxyapatite crystals, the major inorganic component of bone, are about 2 -5 nm in
width and 50 nm in length. Type I collagen, the major organic component of bone, has
fibrils 300 nm in length, 0.5 nm in width and has a periodicity of 67 nm [46]. Numerous
reports in the literature indicate that osteoblast function is greater on nano-structured
compared to current implant surfaces (which are micron-scale rough and nano-scale
smooth). Studies have shown optimal initial protein adsorption from serum onto nano-
structured ceramics [47, 48] and nano-phase metals [49] leading to greater osteoblast
functions. It has also been demonstrated that increased osteoblasts (i.e., bone-forming
cells) functions on nano-phase compared to micron, conventional ceramics is
independent of surface chemistry and material crystalline phase [50].
2.1.3. Nano-structured selenium coatings: a novel approach of using selenium to create
anti-cancer, anti-bacterial biomaterials
As discussed in Chapter 1, selenium has been shown to possess anti-cancer
properties. It has also been discussed in the previous section that nano-structured
materials can promote bone cell functions. Therefore, there has been efforts to create
nano-rough selenium by chemical etching from selenium compacts [51]. This mode of
selenium addition, however, can limit the mechanical properties of the implant as
selenium, being a metalloid, does not have sufficient mechanical strength needed for
orthopedic applications (such as hip implants). Moreover, this methodology does not
provide control over the release or the dose of selenium. Considering the toxicity of
9
selenium at high doses [52], stability and control over its release would be a very
desirable attribute. The colloidal synthesis of selenium nano-particles is known [53-56]
and yields stable dispersions of free nano-particles in the presence of surface stabilizing
agents [53-56]. It was hypothesized that the addition of nano-particles to implant surfaces
could provide a desirable nano-rough morphology, but would require strong adhesion to
fabricate a surface stable to vibration, sonication, or other common processing steps. A
second target would be to control the time release and the surface density of the selenium,
which determines the selenium dose to the adjacent tissue. To address these needs, a
novel fabrication method was developed here in which colloidal selenium synthesis was
carried out in the presence of substrates to create a coating of uniformly adherent
selenium nanoclusters instead of (or in addition to) free selenium nano-particles in
solution.
2.2. Materials and methods
A detailed fabrication and characterization method for coating titanium and
stainless steel with selenium nanoclusters is introduced here. Cell and bacterial assays to
investigate the functions of healthy osteoblasts and cancerous osteoblasts on the uncoated
and selenium coated materials is also described.
2.2.1. Materials
Metallic substrates (titanium, stainless steel substrates [Alfa Aesar]) were
individually degreased and sonicated in acetone and ethanol for 10 min according to
established procedures [57] with a slight modification comprising of an additional step in
which the substrates were hand-wiped with acetone soaked tissue before degreasing.
10
Degreased titanium substrates were then sterilized by autoclaving at 121oC for 30 min.
Cleaned and sterilized metallic substrates (titanium or Ti, and stainless steel or SS) were
used as a base substrate for colloidal decoration with selenium nanoclusters. The
substrates were exposed to 4:1 molar mixtures of glutathione (GSH, reduced form, TCI
America) and sodium selenite (Na2SeO3, Alfa Aesar) in the concentration ranges shown
in Table 2.1.
Table 2.1. Reagent volumes used in the colloidal synthesis of Se nanoclusters in the presence of various base substrates.
Preparation Method
Reagent *[Se]=0.42mM “Low Dose”
[Se]=5mM “Medium Dose”
[Se]=11.7mM “High Dose”
Deionized water 14.5 ml 9 ml 1 ml 100 mM GSH 0.25 ml 3 ml 7 ml
25 mM Na2SeO3 0.25 ml 3 ml 7 ml Final volume 15 ml 15 ml 15 ml
*[Se] = Final concentration of selenium in the colloidal synthesis solution.
Specifically, the cleaned and sterilized substrates were first placed in a 50 ml
beaker with the side to be decorated facing upward. The reduced glutathione solution was
added to the beaker followed by the sodium selenite solution. Three different solution
concentrations (as shown in Table 2.1) were used to achieve different doses denominated
as low dose (LD), medium dose (MD) and high dose (HD). After a gentle mixing of the
solutions in the reaction beaker, 1mL of 1M NaOH was introduced to bring the pH into
the alkaline regime. The reaction mixture was once again gently mixed and left
undisturbed for 10 min. The substrates were removed from the beaker and rinsed in
deionized water. The uncoated and coated metallic substrates (Ti and SS) were exposed
11
to ultra-violet light for 24 hrs on each side to sterilize them before use in cell
experiments.
2.2.2. Material characterization
2.2.2.1. Topography and chemical characterization
Surfaces of uncoated and selenium coated metallic substrates (Ti and SS) were
visualized (without a conductive coating) using a scanning electron microscope (SEM,
LEO 1530VP FE-4800) with an accelerating voltage from 3 kV to 10 kV. SEM produces
images by scanning an electron beam across the surface of interest in a rectangular
pattern recording the reflected electrons. SEM can achieve much higher magnification
than conventional light microscopy because the wavelengths of electrons used in SEM
are much shorter than that of photons. For chemical characterization, X-ray photoelectron
spectroscopy (XPS, Perkin-Elmer PHI. 5500 Multi-Technique System) was used to
confirm the presence of selenium and relatively compare selenium concentrations on the
different substrates. In XPS, a sample surface is hit by a beam of X-rays that provide
energy to electrons in a thin top layer (about 1 to 10 nm) on the surface. Some electrons
receive enough energy to escape from the sample. XPS measures the kinetic energy and
the number of the escaped electrons that contain information about the elemental
composition, chemical state and quantity of the elements that exist in the material.
2.2.2.2. Strength of attachment of selenium on substrates
To test the strength of attachment, the selenium-coated substrates were placed in a
beaker filled with water. The beaker was subjected to sonication for 10 min at 90 W
12
(Ultrasonic Cleaner 75D, VWR). SEM was used to visualize the surfaces of the
substrates before and after sonication.
2.2.2.3. Surface roughness characterization
Surface roughness can affect protein adsorption and cellular adhesion. To
determine the surface roughness of the substrates (uncoated and selenium coated
substrates), atomic force microscopy (AFM, Autoprobe CP, Park Scientific Instrument)
was used with commercially available AFM tips (radius of tip curvature was less than 10
nm, NSC15/ALBS, Micro-Masch, OR) in tapping mode. The relatively high resolution of
~10nm in the lateral scale and 1 Ao in the vertical scale allows for accurate measurement
of nanometer dimensions (~40-90 nm) on the samples. AFM works by using a tip
attached to a cantilever to scan surface of sample and measure the deflection of the
cantilever caused by forces between atoms at the tip and those on the scanned area. The
scanning area was 5 µm x 5 µm and each substrate was scanned at 5 random regions. The
conventional roughness is defined as the deviation from median height (root mean
square):
2
1
( )N
i avei
Z ZRMS
N
(Equation 2.1)
Here, aveZ is the average Z value within a given area and N is the number of data points
within a given area.
13
2.2.2.4. Contact angle measurements
Contact angles contain information about hydrophilicity or hydrophobicity of a
surface which plays an important role in the interaction of the surface with biological
objects (such as proteins and cells). A drop shape analysis system (DSA-10, Kruss,
Germany) (Figure 2.1) with analysis software (DSA1 v 1.80) was used to determine the
surface contact angles for the samples.
Figure 2.1. Image of DSA100 (Updated Version of DSA-10, Kruss, Germany: Image Adopted from www.kruss.com). Capability of equipment: The range of contact angle measurement is between 0 and 180° with a resolution of ±0.1 and a range of surface tension between 10-2 and 100mJ/m2, with a resolution of 0.01mJ/m2.
To measure water contact angles, a 2µL drop of water was placed onto the
substrates and contact angles were recorded within 10 seconds at room temperature.
Pipettes with controlled suction speed (R-10, Rainin Instrument, MA) were used with
14
appropriately sized pipette tips (Optimum #7560-100) to ensure the accurate liquid
volumes. Measurements were repeated 9 times for each sample.
2.2.3. In vitro biological assays for uncoated and selenium coated metallic substrates
2.2.3.1. Bone cell adhesion and proliferation assays on uncoated and selenium coated Ti
substrates
To investigate osteoblast (bone-forming cell) functions on titanium substrates
coated with selenium, human osteoblasts (will be referred to as healthy osteoblasts, CRL-
11372; American Type of Culture Collection-ATCC, population numbers 6-10) in
Dulbecco's Modified Eagle Media (DMEM) supplemented with 10% fetal bovine serum
(FBS, Hyclone) and 1% Penicillin / Streptomycin (P/S, Hyclone) were seeded at a
density of 3500 cells/cm2 and placed in an incubator under standard cell culture
conditions (37oC, 5% CO2, 95% humidified air) for 4 hrs and 1 day. This human
osteoblast cell line was chosen because they have been shown to be able to differentiate
into mature osteoblasts expressing a normal osteoblast phenotype [58].
To study cancerous bone cell functions on the samples, mouse osteosarcoma
osteoblasts (will be referred to as cancerous osteoblasts; CRL-2837; ATCC, population
numbers 14-17) in DMEM supplemented with 10% FBS (Hyclone) and 1% P/S
(Hyclone) were seeded at a density of 3500 cells/cm2 and placed in an incubator under
standard cell culture conditions (37oC, 5% CO2, 95% humidified air) for 4 hrs as well as
1 and 3 days. Media was exchanged every other day. This mouse osteosarcoma cell line
was chosen because this cell line has been shown to be able to develop primary tumors
and pulmonary metastases with histology consistent with osteosarcoma in human patients
15
[59]. In addition, for this cell line, “expression of bone sialoprotein, biglyan, decorrin,
and osteopontin was suggestive of the bone lineage cells” [59].
After the desired time periods, cells were fixed using formaldehyde 4% (Sigma),
stained with 4',6-diamidino-2-phenylindole (DAPI) (ATCC) and counted under
fluorescence microscopy (Zeiss Axiovert 200M Light Microscope) in five random fields
and averaged for each substrate.
2.2.3.2. Healthy osteoblast differentiation assays
Healthy osteoblasts (100,000 cells/cm2) were seeded onto the substrates of
interest to the present study and cultured in DMEM (supplemented with 10% FBS, 1%
P/S, 50 μg/ml L-ascorbate (Sigma) and 10 mM β-glycerophosphate (Sigma)) under
standard cell culture conditions for 14 days. Medium was replaced every other day. At
the end of the prescribed time period, the substrates were rinsed with tris buffered saline
(pH 7.2, Sigma Aldrich) for three times. The remaining osteoblasts on the substrates were
lysed using 1 mL of deionized water and three freeze-thaw cycles. The supernatant
lysates were transferred into microtubes for determining intracellular protein synthesis
and alkaline phosphatase (ALP) activity, while the remaining substrates were used for
determining extracellular calcium deposition.
2.2.3.2.1. Total intracellular protein content
Total protein is proportional to total number of cells. Total protein content in the
cell lysates was determined spectrophotometrically using a commercially available kit
(Pierce Chemical Co.) and following manufacturer's instructions. For this purpose,
aliquots of each protein-containing, distilled-water supernatant were incubated with a
16
solution of copper sulfate and bicinchoninic acid at 37°C for 2 hrs. Light absorbance of
these samples was measured at 562 nm on a MR600 Spectrophotomectric Microplate
Reader (Dynatech). Total intracellular protein (expressed as mg) synthesized by
osteoblasts cultured on the substrates of interest to the present study was determined from
a standard curve of absorbance versus known concentrations of albumin run in parallel
with experimental samples. Total intracellular protein synthesized by osteoblasts cultured
on conventional, uncoated titanium served as controls.
2.2.3.2.2. Intracellular alkaline phosphatase (ALP) activity
Intracellular ALP activity is an indicator of how active the osteoblast cells are in
producing new bone. Intracellular human osteoblast ALP activity was analyzed by a
commercial ALP activity detection kit (Upstate Cell Signal Solutions) following
manufacturer's instructions. The ALP activity in cell lysates was spectrophotometrically
detected by adding a Malachite Green solution compared to a standard curve of
absorbance versus known concentrations of converted phosphate, which reported ALP
activity as the amount of converted phosphate. ALP activity was further normalized by
total intracellular protein content and substrate surface area.
2.2.3.2.3. Calcium deposition in the extracellular matrix
Calcium deposition is an important indicator of the bone-forming process by
osteoblasts. The remaining extracellular matrix on the substrates after 14 days of
culturing healthy osteoblasts was treated with 0.6 M HCL in an incubator (37oC) for 24 h.
After the prescribed time period, calcium present in the acidic supernatant was quantified
using a commercially available kit (Calcium Reagent Set, Pointe Scientific) following
17
manufacturer's instructions. Light absorbance of the acidic supernatant solution with the
addition of a calcium reagent was read at 570 nm. Calcium content was determined from
a standard curve of absorbance versus known concentrations of calcium run in parallel
with the acidic supernatant solution, and the results were normalized over the substrate
surface area.
2.2.3.3. Release of selenium in cell culture media
To study the release of selenium from the substrate surface into the cell culture
media, substrates (uncoated and low, medium and high selenium dose coated titanium)
were cultured without cells in DMEM supplemented with 1% P/S (Hyclone) for 3 days.
To quantify the release of selenium to the fluid phase, the media was exchanged after 2
days (to be consistent with the proliferation experiments where media was exchanged
every other day) and the spent media collected for graphite furnace atomic absorption
spectroscopy (AAS) using Zeeman background correction (41002L Perkin-Elmer GF-
AAS). The instrument was calibrated and tested using the samples with known selenium
concentrations. All the samples were acidified to 2% nitric acid prior to the GF-AAS run.
The wavelength was 196 nm with a bandwidth of 2 nm. A five-step graphite furnace
program was used: (1) 110oC, ramp for 1 s, hold for 20 s; (2) 130oC, ramp for 5 s, hold
for 30 s; (3) 650oC, ramp for 10 s, hold for 20 s; (4) 1750oC, hold for 6 s; and (5) 2400oC,
ramp for 1 s, hold for 2 s. To reduce any matrix effects, 4 µl of a matrix modifier ([1 µg
Pd + 0.6 µg Mg(NO3)2]/µl in 0.5 mol/l HNO3) was used per 20 µl of sample size.
18
2.2.3.4. Effects of selenium released into culture media on healthy and cancerous cells
Either uncoated titanium or titanium coated with the highest selenium density
were cultured (without cells) in DMEM supplemented with 10%FBS and 1% P/S for 2
days. After 2 days, the media (supernatant) was collected and used to culture either
osteoblast or osteosarcoma cells (described in 2.2.3.1) in the wells of 6 well-tissue culture
plates. Fresh DMEM supplemented with 10%FBS and 1% P/S was used as the control.
After 3 days of culture, cell densities were determined by the fluorescence microscopy
techniques as just described in 2.2.3.1.
2.2.3.5. Intracellular thiol assays
Intracellular thiol groups are crucial for the overall health of cells. Intracellular
thiols protect the cells from damages caused by free radicals, toxins, carcinogens, etc. To
investigate the influence of selenium released into the culture media on the level of
intracellular thiols, uTi and High-nSe-Ti were immersed in DMEM (supplemented with
10% FBS and 1% P/S) without cells for 2 days. After the indicated time, supernatants
collected and used to culture either osteoblast or osteosarcoma cells (described in 2.2.3.1)
for 3 days. After the indicated time, cells were collected and analyzed for thiol content
using a glutathione assay kit (CS1020, Sigma Aldrich) following manufacturer’s
instruction. The assay uses a thiol probe (monochlorobimane) which passes through cell
membranes to detect the level of reduced glutathione, the major free thiol in most living
cells. Glutathione is involved in many biological processes such as removal of
hydroperoxides and detoxification of xenobiotics. Intracellular reduced glutathione level
is a sensitive indicator of the overall health of a cell. If the probe binds to reduced
19
glutathione in cells, it forms a fluorescent product whose fluorescence intensity can be
measured. The unbound probe does not show fluorescence.
2.2.4. Fibronectin adsorption on uncoated and selenium coated titanium substrates
Fibronectin is one of important cell-adhesive proteins that regulate interaction
between osteoblasts and substrates. Adsorption of fibronectin on implant is, therefore,
important for cell adhesion, migration and spreading. Several methods exist to measure
protein adsorption on implant surfaces (such as XPS, Raman spectroscopy, fluorescence
spectroscopy, infrared spectroscopy, enzyme-linked immunosorbent assay (ELISA) or
surface plasmon resonance). Among all the methods, ELISA is simple, requiring less
equipment but still provide accurate results. In ELISA, fibronectin is directly linked to
antibodies which are then conjugated with fluorescent dyes and detected by a fluorescent
signal detector. Experiments were conducted to investigate adsorption of fibronectin from
either DMEM (supplemented with 10%FBS and 1% P/S) or 5µg/mL fibronectin in
phosphate-buffered solution (PBS, Sigma) onto the substrates of interest. Specifically,
ELISA experiments were conducted following these steps below:
1. Uncoated and selenium-coated titanium samples were immersed in 0.5 mL of
either DMEM (supplemented with 10% FBS and 1% P/S) or 5µg/mL fibronectin solution
(in PBS) in a 24 well culture dish (Corning) in an incubator (37oC, 5%CO2) for 24 hrs.
Samples were immersed in either DMEM (with no FBS or P/S supplement) or PBS (with
no fibronectin) were used as controls to subtract a nominal fluorescence signal later.
2. Samples were rinsed in a PBS three times. Bovine serum albumin (BSA,
Sigma) (2% by wt % in PBS) was used to block all substrate areas that did not react with
proteins.
20
3. A rabbit anti-bovine fibronectin antibody (AB2047, Chemicon) was added to
each well at concentration of 6 µg/mL (in 1% BSA) and was incubated for 1 hr to link to
fibronectin.
4. Samples were rinsed in 0.05% Tween 20 (Sigma) three times. A second
antibody, goat anti-rabbit conjugated with horseradish peroxidase (HRP) was added to
each well at a concentration of 10 µg/mL (in 1% BSA) and incubated for 1 hr.
5. Samples were transferred to new 24 well plates and rinsed twice in 0.05%
Tween 20. An 2,2'-azino-bis (3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) substrate kit
(SK-4500, Vector Laboratories) was used. ABTS reacts only with HRP to develop a
water soluble, green colored product. Different intensities of fluorescence (for each
sample) were detected by a spectrophotometer (SpectraMax 190, Molecular Devices) at
405nm. Fluorescence intensities were subtracted from those of the corresponding control.
2.2.5. Statistical analysis
Experiments were conducted in triplicate and repeated three times. Data were
collected and the significant differences were assessed with the probability associated
with a one- tailed Student’s t test.
2.3. Results
In this section, results from material characterization studies and cell tests on
titanium substrates will be presented first. Results on stainless steel substrates will follow
to demonstrate that the same synthesis method can be used on other orthopedic metals to
achieve similar results.
21
2.3.1. Surface characterization
Selenium nanoclusters were clearly present on titanium substrates following the in
situ reaction under the low, medium and high dose conditions (which corresponded to
increasing the overall reagent concentrations as detailed in Table 2.1) (Figure 2.2).
Figure 2.2. Representative SEM images of: (A) uTi; (B) Low-nSe-Ti; (C) Medium-nSe-Ti and (D) High-nSe-Ti. The surface area coverage of selenium on the selenium coated substrates were determined (using an image processing program, ImageJ) to be 2.7%, 5.1% and 7.5% for Low-nSe-Ti, Medium-nSe-Ti and High-nSe-Ti, respectively.
The clusters were approximately 80nm in diameter and had hemispherical shapes
(Figure 2.3). Higher surface densities of selenium nanoclusters were observed on the
titanium substrates for the higher dose cases (prepared at higher selenite concentrations in
the final synthesis solution). The substrates will be referred to as uTi, Low-nSe-Ti,
Medium-nSe-Ti and High-nSe-Ti for uncoated titanium, and low, medium and high
22
selenium nanocluster density coated titanium substrates, respectively. The surface
coverage of selenium determined from SEM images using an image processing program,
ImageJ (NIH), was approximately 2.7% for Low-nSe-Ti, 5.1% for Medium-nSe-Ti and
7.5% for High-nSe-Ti. The surface coverage of selenium is not linearly dependent on the
concentration of selenium in the synthesis solution (see Table 2.1)
Figure 2.3. SEM image, taken at a 45° tilt, of High-nSe-Ti showing the hemispherical shape of the selenium nanoclusters on the titanium surface.
The morphology in the high-magnification tilted SEM image (Figure 2.3) strongly
suggested the selenium was bonded to the substrate. Together with the fact that cleaning
of the titanium surface before coating was found to be crucial for forming of selenium
nanoclusters on the surfaces, it is speculated that selenium nanoclusters were formed
23
through nucleation processes. The nucleation sites on the titanium surface (e.g. defects,
impurities) served as high energy regions for the nucleation process to happen.
To test the strength of attachment, the composite substrate was subjected to
sonication for 10 min at 90 W (Ultrasonic cleaner 75D, VWR). The morphology of the
substrates and the densities of coatings appeared similar by SEM before and after
sonication. The morphology was also found to be unchanged after UV treatment for 24
hrs (Figure 2.4).
Figure 2.4. Morphology of a selenium-coated titanium surface appeared the same before (A) and after sonication (B) and UV radiation (C). Bars=500 nm.
The presence of elemental selenium on the titanium surfaces was further
confirmed by XPS analysis. All XPS profiles of the selenium coated substrates showed
peaks of selenium not observed for the conventional uncoated titanium substrates (Figure
2.5).
24
Figure 2.5. All coated substrates showed peaks of selenium that were not detected in uncoated substrate.
The position of the peak at a binding energy of 55.2 eV (Figure 2.6) confirmed
the elemental state of selenium [60].
Figure 2.6. XPS profiles of High-nSe-Ti showing the characteristic binding energy peak for elemental selenium at 55.2 eV.
The elemental form is also produced by the colloidal synthesis of free particles
[61], which indicated that the substrate did not alter the basic reaction chemistry except to
25
introduce nucleation sites. Selenium peaks were more pronounced both in height and in
number on the substrates coated with higher doses of selenium.
The AFM images and analysis (Figure 2.7) confirmed again the present and
hemispherical morphology of selenium nanoclusters on the coated Ti substrates (Figure
2.7).
26
Figure 2.7. Representative AFM images and line analysis of uTi (A), Low-nSe-Ti
(B), Medium-nSe-Ti (C) and High-nSe-Ti (D). Nano-scale roughness was created by the
selenium coatings and increased with increasing selenium coating density.
More importantly, all coated substrates possess nano-scale roughness (Figure 2.8)
contributed by the nanostructure of selenium coatings. The normalized RMS roughness
27
values were 8.07±0.71 nm for uTi, 9.72±0.41 nm for Low-nSe-Ti, 12.56±0.79 nm for
Medium-nSe-Ti and 14.11±1.08 nm for High-nSe-Ti.
Figure 2.8. RMS of the substrates increased with increasing selenium coating density. Data = mean ± standard error of the mean; N=3; * p<0.05 compared to all other substrates; ** p<0.01 compared to Medium-nSe-Ti and High-nSe-Ti.
Increased coating levels (provided by increasing selenium concentration in the
synthesis solution and as indicated by the densities of selenium clusters in SEM images
and in AFM images) created surfaces with increased nanometer scale roughness (Figures
2.7 and 2.8).
Uncoated Ti substrates were hydrophilic with an average water contact angle of
approximately 64 degrees. Importantly, coating Ti with selenium nanoclusters made the
Ti surfaces more hydrophobic with an average water contact angles of approximately 74
degrees (Figure 2.9). No significant difference was found among contact angles on all
selenium coated substrates.
28
Figure 2.9. Water contact angles on the uncoated and coated Ti substrates. Contact angles increased on the substrates coated with selenium nanoclusters. Data = mean ± standard error of the mean; N=3; * p<0.05 compared to all the coated substrates. There was no significant difference among the contact angles on the coated substrates.
The plateau of water contact angle values for different selenium coating densities
was hypothesized to be attributed to the formation of air-pockets during the contact angle
measurement. More detailed discussion of this will be presented in Chapter 4.
2.3.2. Cell adhesion and proliferation
Most importantly, after 4 hrs of culture, healthy osteoblast densities significantly
increased on the High-nSe-Ti compared to uTi and Low-nSe-Ti (Figure 2.10). After 4
hrs, healthy osteoblast density increased 12% on High-nSe-Ti compared to uTi. After 1
day, healthy osteoblast density increased 68% on High-nSe-Ti, 37% on Medium-nSe-Ti
29
and 13% on Low-nSe-Ti compared to uTi, respectively. After 3 days, healthy osteoblast
density increased 60% on High-nSe-Ti, 37% on Medium-nSe-Ti and 22% on Low-nSe-Ti
compared to uTi, respectively.
Figure 2.10. Increased healthy osteoblast densities after 4 hrs, 1 day and 3 days. Data = mean ± standard error of the mean; N=3, * p<0.05 compared to Medium-nSe-Ti (compared at same time period), ** p<0.05 compared to uTi (compared at same time period), *** p< 0.05 compared to Low-nSe-Ti (compared at same time period), # p<0.05 compared to uTi (compared at same time period), & p =0.06 compared to Low-nSe-Ti. (compared at same time period).
In contrast, for cancerous cells, after 4 hrs, cell densities were significantly higher
on the uTi and Low-nSe-Ti than on the High-nSe-Ti (Figure 2.11).
30
Figure 2.11. Cancerous osteoblast densities after 4 hrs, 1 day, and 3 days. Decreased cancerous osteoblast densities on selenium coated titanium substrates after 4 hrs, 1 day and 3 days. Data = mean ± standard error of the mean; N=3, * p<0.05, ** p< 0.01 compared to High-nSe-Ti. (Compared at same time period); # p<0.1 compared to Medium-nSe-Ti. (Compared at same time period). There are no significant differences in cell densities among substrates after 1 day.
Cancerous osteoblast densities after 4 hrs were 5365 ±145 cells/cm2 on uTi, 4779
± 443 cells/cm2 on Low-nSe-Ti, 4526 ± 308 cells/cm2 on Medium-nSe-Ti and 3547 ± 47
cells/cm2 on High-nSe-Ti. Cancerous osteoblast density decreased approximately 1.5
times on High-nSe-Ti compared to uTi.
After 1 day of cell culture, healthy osteoblast densities increased on Medium-nSe-
Ti (8165± 1509 cells/cm2) and High-nSe-Ti (9983± 1075 cells/cm2) compared to that on
uTi (5955± 561 cells/cm2) (Figure 2.12). Healthy cell density increased approximately
1.7 times on High-nSe-Ti and 1.4 times on Medium-nSe-Ti compared to that on uTi. No
significant difference was found for the cancerous osteoblast densities on the four types
of titanium substrates (i.e., uTi, Low-nSe-Ti, Medium-nSe-Ti, High-nSe-Ti) after one day
of culture.
31
Figure 2.12. Representative fluorescence microscopy images of healthy osteoblasts on: (A) uTi; (B) Low-nSe-Ti; (C) Medium-nSe-Ti and (D) High-nSe-Ti after 1 day (Magnification = 10X). Scale bars = 200 μm.
However, after 3 days of culture, cancerous osteoblast density on the High-nSe-Ti
was significantly reduced in comparison to all other substrates (Figure 2.13). Cancerous
osteoblast densities on Medium-nSe-Ti were also significantly lower than that on uTi.
Cancerous osteoblast densities on the substrates after 3 days were 141882 ±36572
cells/cm2 for uTi, 111581±17101 cells/cm2 for Low-nSe-Ti, 72187±13702 cells/cm2 for
Medium-nSe-Ti and 31885±6933 cells/cm2 for High-nSe-Ti. Cancerous cell density
decreased approximately 4.5 times on High-nSe-Ti, 1.9 times on Medium-nSe-Ti
compared to that on uTi after 3 days.
32
Figure 2.13. Representative fluorescence microscopy images of cancerous osteoblasts on: (A) uTi; (B) Low-nSe-Ti; (C) Medium-nSe-Ti and (D) High-nSe-Ti after 3 days (Magnification = 20X). Scale bars = 100 μm.
To quantitatively assess the effects of selenium nanocluster coatings on growth of
healthy osteoblasts, exponential growth curves were fit to the experiment data (Figure
2.14)
33
Figure 2.14. Fitting exponential growth curves to experiment data for densities of healthy osteoblasts cultured on uncoated titanium and on selenium-coated titanium substrates. Growth curves of healthy osteoblasts cultured on higher selenium coating densities had large slopes.
The fitting parameter and doubling time of healthy osteoblast determined from
the exponential fitting curves were summarized in Table 2.2.
Table 2.2. Fitting parameter and the goodness of fitting (R2) for growth curve and calculated doubling time of healthy osteoblasts cultured on uncoated and selenium-coated titanium substrates
Fitting parameters for y~ebx b Doubling time (hours)
τ = ln(2)/b R2
uTi 0.0167 41.505 0.9125 Low-nSe-Ti 0.0193 35.914 0.9067
Medium-nSe-Ti 0.0198 35.007 0.8559 High-nSe-Ti 0.0208 33.324 0.8131
Doubling time of healthy osteoblasts decreased from ~ 41 hrs when cultured on
uncoated titanium to ~ 33 hrs when cultured on High-nSe-Ti, i.e., the healthy osteoblasts
divided ~ 1.2 times faster on High-nSe-Ti compared to uTi.
34
Exponential growth curves fitted from experiment data for densities of cancerous
osteoblast cultured on the uncoated and selenium coated substrates showed a reverse
trend (Figure 2.15); growth curves of cells cultured on selenium-coated surfaces had
smaller slopes.
Figure 2.15. Fitting exponential growth curves to experiment data for densities of cancerous osteoblasts cultured on uncoated titanium and selenium-coated titanium substrates. Curves of cancerous osteoblasts cultured on lower selenium coating densities had large slopes with the uncoated titanium substrates having the largest slope.
The fitting parameter and doubling time of cancerous osteoblast determined from
the exponential fitting curves were summarized in Table 2.3.
35
Table 2.3.Fitting parameter and goodness of the fitting (R2) for growth curve and calculated doubling time of cancerous osteoblasts cultured on uncoated and selenium-coated titanium substrates.
Fitting parameters for y~ebx b Doubling time (hours)
τ = ln(2)/b R2
uTi 0.0499 13.89 0.9794 Low-nSe-Ti 0.0478 14.5 0.9844
Medium-nSe-Ti 0.0471 14.716 0.99 High-nSe-Ti 0.0322 21.526 0.999
Doubling time of cancerous osteoblasts increased from ~ 14 hrs when cultured on
uncoated titanium to ~ 14.5 hrs on Low-nSe-Ti, 14.7 hrs on Medium-nSe-Ti and 21.5 hrs
when cultured on High-nSe-Ti.
2.3.3. Healthy osteoblast differentiation
To investigate the long-term functions of healthy osteoblasts on substrates coated
with Se, a variety of osteoblast differentiation makers (including intracellular protein
synthesis, ALP activity and extracellular calcium deposition) were examined in the
present study. Uncoated Ti substrates were used as the control. The value of ALP activity
(normalized to total protein content) was 136.78 ±3.7 pmol/µg.cm2 for uTi, 120.16 ±5.5
pmol/µg.cm2 for Low-nSe-Ti, 151.93 ±5.0 pmol/µg.cm2 for Medium-nSe-Ti and 226.75
±28.4 pmol/µg.cm2 for High-nSe-Ti. The results showed that after 14 days of culturing,
ALP activity of healthy osteoblasts on High-nSe-Ti was higher than on uTi, Low-nSe-Ti
and Medium-nSe-Ti (Figure 2.16). Compared to the uncoated substrate (uTi), the
substrate coated with highest amount of selenium (High-nSe-Ti) has ALP activity
increased approximately 1.7 times.
36
Figure 2.16. Increased alkaline phosphatase (ALP) activity of osteoblasts on High-nSe-Ti compared to all other substrates. Data = mean ± standard error of the mean; N=3, * p<0.01 compared to uTi, Low-nSe-Ti and Medium-nSe-Ti.
There was no significant difference between ALP activity on uTi, Low-nSe-Ti
and Medium-nSe-Ti. Higher ALP activity which was normalized by intracellular protein
content per unit area on High-nSe-Ti indicated that bone formation on High-nSe-Ti was
more active than on uTi, Low-nSe-Ti and Medium-nSe-Ti.
Calcium deposition in the extracellular matrix is one of the important markers of
bone formation and osteoblast differentiation. Calcium deposition (normalized to surface
area) was 0.97 ±0.03 (µg/cm2) for uTi, 1.003 ±0.029 (µg/cm2) for Low-nSe-Ti, 1.004
±0.038 (µg/cm2) for Medium-nSe-Ti and 1.093 ±0.04 (µg/cm2) for High-nSe-Ti.
Calcium deposition by osteoblasts on High-nSe-Ti was significantly greater than on uTi
(Figure 2.17). There were no significant difference between calcium deposition by
osteoblasts on uTi, Low-nSe-Ti and Medium-nSe-Ti.
37
Figure 2.17. Extracellular calcium deposition by osteoblasts on uTi, Low-nSe-Ti, Medium-nSe-Ti and High-nSe-Ti after 14 days. Data = mean ± standard error of the mean; N=3, * p<0.05 compared to uTi.
These results are consistent with the results of proliferation of healthy osteoblasts
after 1 day and 3 days (Figures 2.10 and 2.12) and suggest promoted active osteoblast
functions on High-nSe-Ti compared to uTi.
2.3.4. Selenium release
To test the release of selenium coatings into culture media, titanium substrates
coated with selenium were immersed in culture media (DMEM supplemented with 10%
FBS and 1% P/S) without cells for 3 days. Culture media were then used to analyze
selenium content. A dose dependent release of selenium from the substrates to cell
culture medium was observed by GF-AAS analysis (Figure 2.18).
38
Figure 2.18. Total selenium released into cell culture media after 2 and 3 days. Data = mean ± standard error of the mean; N=2, * p<0.0001 compared to High-nSe-Ti. (Compared at same time period); † p<0.0001 compared to Low-nSe-Ti. (Compared at same time period).
After 2 days, approximately 3.6% of the total selenium (estimated using SEM
images) was released for the High-nSe-Ti. Release then slowed greatly, with a negligible
selenium release occurring on the third day. The release from High-nSe-Ti was
significantly greater than that from Medium-nSe-Ti and that from Low-nSe-Ti (compared
at the same time period). The release from Medium-nSe-Ti was also significantly greater
than from Low-nSe-Ti (compare at the same time periods). The amount of selenium
released into culture media was also not linearly dependent on the coating density.
2.3.5. Effects of selenium released into culture media on healthy and cancerous cells
To further explore why proliferation of cancerous osteoblasts was inhibited on
selenium coated Ti, the supernatant experiments were conducted which consisted of
39
immersing the substrates in culture media with no cells, collecting the media (i.e.,
supernatant) after 2 days and culturing cells (either healthy osteoblasts or cancerous
osteoblasts) in the supernatant. The results (Figure 2.19) showed significantly lower
(approximately 1.5 times smaller) cancerous cell densities cultured with the supernatant
from High-nSe-Ti than cancerous cell cultured with fresh media (p<0.01) after 3 days of
culture. The density of cancerous cells cultured with the supernatant from High-nSe-Ti
was also smaller than that of cells cultured with the supernatant from uTi (p=0.12). On
the other hand, after 3 days, no significant difference was found among healthy osteoblast
densities cultured in fresh media, the supernatant from uTi or supernatant from High-nSe-
Ti (Figure 2.19). The release of selenium into culture media showed no toxic effects on
healthy osteoblasts in this study.
Figure 2.19. Decreased cancerous osteoblast and maintained healthy osteoblast densities in media collected from supernatant experiment with High-nSe-Ti. Data = mean ± standard error of the mean; N=3, ** p<0.01 compared to High-nSe-Ti. Density of cancerous cells cultured in media collected from the supernatant experiment with uncoated Ti was greater (but not significant, * p=0.12) than that of cells cultured in media collected from the supernatant experiment with High-nSe-Ti. There was no significant difference among the densities of healthy osteoblasts.
40
2.3.6. Intracellular thiol assays
To have a mechanistic understanding of the results presented in Figure 2.19,
intracellular thiol assays experiments were conducted for two types of substrates, uTi
(control) and High-nSe-Ti (which is the Se coated substrate that had the highest normal
healthy osteoblast density and the lowest cancerous osteoblast density from individual
cell experiments). Depletion of intracellular thiol content has been shown to cause cell
death (either apoptosis or necrosis or both) [62-65]. It has been shown that elemental
selenium in the form of nanoparticles caused thiol depletion in leukemia cells leading to
cell death [40]. Therefore, this part of the thesis tested the hypothesis that selenium
released into culture media caused depletion of intracellular thiols. For this, cells (either
healthy osteoblasts or cancerous osteoblasts) were cultured in supernatants (i.e., culture
media collected after 3 days of immersing in DMEM with no cells) from uTi or High-
nSe-Ti. After 1 day, cells were analyzed for content of intracellular thiols.
Thiol content of healthy osteoblasts cultured in supernatant from High-nSe-Ti was
lower than that of healthy osteoblasts cultured in supernatant from uTi (Figure 2.20).
Figure 2.18. Decreased thiol content of healthy osteoblasts cultured in supernatant from High-nSe-Ti compared to uTi. Data = mean ± standard error of the mean; N=3; * p<0.1
41
Similarly, content of intracellular thiols in cancerous osteoblasts cultured in
supernatant from High-nSe-Ti was lower than uTi (Figure 2.21).
Figure 2.19. Decreased thiol content of cancerous osteoblasts cultured in supernatant from High-nSe-Ti compared to uTi. Data = mean ± standard error of the mean; N=3; * p<0.05.
As can be seen from the results (Figures 2.20 and 2.21), thiol content was
depleted in both healthy osteoblasts and cancerous osteblasts when cultured in
supernatant from High-nSe-Ti. Since cancerous cells are more susceptible to oxidative
stress than their normal counterparts [66-68], it is likely that, in this study, the cancerous
cells were more sensitive to oxidative stress caused by thiol depletion compared to the
healthy cells. It is also possible that many cancer cells are more sensitive to thiol
depletion compared to normal counterparts because the level of thiols in those cancer
cells are close to the level required for cell survivals [69].
2.3.7. Fibronectin adsorption on uncoated and selenium-coated titanium substrates
Fibronectin is one of the most important cell adhesive proteins that regulate the
interactions between osteoblasts and surface of bone implants. To understand the
42
mechanism of improved healthy osteoblast functions (adhesion, proliferation and
differentiation) on the selenium-coated titanium substrates, fibronectin adsorption
experiments were conducted. Amount of fibronectin was determined using ELISA
method. In this method, fibronectin was attached to antibodies which were attached to a
“color” molecule. The “color” molecules were then solubilized by ABTS solution to give
a soluble green color product whose optical density can be measured. A standard curve
was constructed from a series of known fibronectin concentrations to give relationship
between optical density and amount of fibronectin. Results from experiments of
fibronectin adsorption from DMEM (supplemeted with 10% FBS and 10% P/S) onto
substrates of interests showed that fibronection adsorption was greater on Medium-nSe-
Ti and High-nSe-Ti compared to uTi while fibronectin adsorption on Low-nSe-Ti was
not significantly different from that on uTi (Figure 2.22).
43
Figure 2.20. Increased fibronectin adsorption from DMEM onto Medium-nSe-Ti and High-nSe-Ti substrates. DMEM (supplemented with 10% FBS and 10% P/S) was estimated to have fibronectin concentration of 3µg/mL. Each sample was immersed in 0.5 mL of DMEM solution for 24 hrs under standard condition (37oC, 5% CO2, 95% humidified air) and amount of absorbed fibronectin was determined via optical density of ABTS solution using ELISA method. A standard curve relating optical density to amount of fibronectin was used to determine fibronectin concentration from optical densities of ABTS solution. Finally, amount of fibronectin was normalized to geometrical surface areas of the substrates. Data = mean ± standard error of the mean; N=3, * p<0.05 compared to uTi, ** p<0.05 compared to uTi and Low-nSe-Ti.
Results on fibronectin adsorption from fibronectin solution (in PBS) onto the
uncoated and coated substrates also showed similar trend (Figure 2.21): Samples with
higher coating densities had higher amount of protein adsorbed.
44
Figure 2.21. Increased fibronectin adsorption from fibronectin solution (in PBS) on Medium-nSe-Ti and High-nSe-Ti substrates. Each sample was immersed in 0.5 mL of 5 µg/mL fibronectin solution for 24 hrs under standard condition (37oC, 5% CO2, 95% humidified air) and amount of absorbed fibronectin was determined via optical density of ABTS solution using ELISA method. A standard curve relating optical density to amount of fibronectin was used to determine fibronectin concentration from optical densities of ABTS solution. Finally, amount of fibronectin was normalized to geometrical surface areas of the substrates. Data = mean ± standard error of the mean; N=3, * p<0.05 compared to uTi and Low-nSe-Ti. ** p=0.055 compared to Medium-nSe-Ti.
It is noted that the concentration of fibronectin in DMEM is approximately the
same as the concentration of fibronectin in PBS used in adsorption experiments.
However, the amount of fibronectin adsorbed from fibronectin solution in PBS (no other
protein present) was approximately 10 times higher than that from DMEM. This is likely
due to the competitive adsorption of other proteins (such as albumin, laminin, vitronectin,
etc.) in the DMEM onto the substrate surface.
The increase in fibronectin adsorption onto the selenium nanocluster coated
surfaces can be attributed to factors such as surface area increase, surface roughness/
morphology, surface charges or surface hydrophilicity/hydrophobicity properties.
However, in this study, the surface area increase was not significant among the uncoated
45
and coated surfaces. In addition, hydrophobicity/hydrophilicity properties were not
significantly different among the coated surfaces. Therefore, it is hypothesized that the
increase in fibronectin adsorption on the higher density selenium-coated substrates was
due to the surface roughness/morphology. Detail discussion of this subject is in Chapter
3.
2.4. Results on stainless steel samples
Stainless steel is another common material used in orthopedics. As a
demonstration of the applicability of the fabrication method to create nanocluster
selenium coatings for anti-cancer orthopedic application, results on scanning electron
microscopy investigation of the stainless steel surfaces and results on culturing healthy
osteoblasts, cancerous osteoblasts on the substrates for 3 days were presented here.
2.4.1. Material characterization
Selenium nanoclusters were clearly present on stainless steel following the in situ
reaction under low, medium and high dose conditions, (which corresponded to increasing
overall reagent concentrations as detailed in Table 2.1) (Figure 2.22)
46
Figure 2.22. SEM images of stainless steel substrates: (A) uncoated; (B), (C), and (D) coated with Se concentration in the final solution which equaled 0.42 mM, 0.83 mM ,and 1.25 mM, respectively. Some Se nanoclusters are indicated by arrows. These clusters are approximately 80nm in diameters. Bars = 1µm.
Higher surface densities of Se nanoclusters were observed on the stainless steel
substrates with higher doses (prepared at higher selenite concentrations). The surface area
coverage of selenium nanoclusters (determined from SEM images using imageJ, an
image processing program) was approximately 0.6% in (B), 1.5% in (C) and 4.5% in (D).
The substrates will be referred to as uSS, Low-nSe-SS, Medium-nSe-SS and High-nSe-
SS for uncoated stainless steel, and low, medium and high Se nanocluster density coated
stainless steel substrates, respectively.
One can notice that, compared to the selenium coated titanium substrates, the
selenium coated stainless steel substrates prepared using the same concentration of
selenium in the synthesis solutions had lower coating densities of selenium nanoclusters.
This might be due to the difference in densities of nucleation sites present on the surfaces
47
of titanium and stainless steel substrates. The sizes of the selenium nanoclusters on
stainless steel substrates were similar to those on titanium substrates.
Water contact angles were measured on the stainless steel substrates. Uncoated
stainless steel substrate was hydrophilic and had a contact angle of 70.2±2 degrees.
Coating stainless steel substrates with selenium nanoclusters makes the substrates
become more hydrophobic (Figure 2.23).
Figure 2.23. Water contact angles on stainless steel substrates. Selenium nanocluster coating makes the stainless steel substrates become more hydrophobic.
Contact angles on the stainless steel substrates coated with three different
selenium nanocluster densities did not show significant difference. This trend was also
observed in titanium substrates coated with selenium nanoclusters. Again, this trend is
hypothesized to be attributed to the formation of air-pockets when a drop of liquid was
placed on the surface to measure the contact angle. Detailed discussion of the contact
angle results is presented in Chapter 4.
48
2.4.2. Cell experiments on stainless steel substrates
After 3 days of culture, there was no significant difference in non-cancerous
osteoblast densities among the substrates (i.e., selenium-coated stainless steel and
uncoated stainless steel) (Figures 2.26 and 2.27)
Figure 2.24. Maintained healthy osteoblast densities on selenium-coated stainless steel substrates compared to the uncoated substrate after 3 days of culturing. Data = mean ± standard error of the mean; N=3.
49
Figure 2.25. Representative fluorescence images of healthy osteoblast densities on uSS (A), Low-nSe-SS (B), Medium-nSe-SS (C) and High-nSe-SS (D). Magnification = 10X.
Compared to results on titanium substrates, this result showed that the stainless
steel substrates coated with selenium nanoclusters maintained densities of healthy
osteoblasts. This means that the coated stainless steel substrates had similar
biocompatibilities toward healthy bone cells compared to the uncoated substrate. It was
50
hypothesized that the selenium nanocluster coating densities were not high enough to
produce significantly promoted healthy osteoblast functions on stainless steel substrates.
In contrast, for cancerous osteoblasts, after 3 days of culture, cell densities were
significantly lower on High-nSe-SS and Medium-nSe-SS than on Low-nSe-SS, uSS
(Figures 2.28 and 2.29). Cell density was decreased approximately 1.5 times on High-
nSe-SS and 1.2 times on Medium-nSe-SS compared to that on uSS. In addition,
comparing between High-nSe-SS and Medium-nSe-SS substrates, High-nSe-SS had a
significantly lower number of cancerous osteoblasts than Medium-nSe-SS substrates.
Figure 2.26. Decreased cancerous osteoblast densities after 3 days of culturing on Se coated stainless steel. Cell densities on Medium-nSe-SS and High-nSe-SS were significantly less than those on uSS and Low-nSe-SS. * p<0.05 compared to uSS and Low-nSe-SS; ** p<0.01 compared to Medium-nSe-SS. Data = mean ± standard error of the mean; N = 3.
51
Figure 2.27. Representative fluorescence images of cancerous osteoblast densities on uSS (A), Low-nSe-SS (B), Medium-nSe-SS (C) and High-nSe-SS (D). Magnification = 20X.
This trend (i.e., decreased cancerous cell densities on substrates with increased
selenium coating densities) was similar to what observed in titanium samples. However,
the level of decrease in cancerous cell densities on selenium-coated stainless steel
substrates (compared to uncoated stainless steel substrate) was much less than the level of
decrease in cancerous cell densities on selenium-coated titanium substrates (compared to
52
uncoated titanium substrate); High-nSe-SS had cell density decreased 1.5 times compared
to uSS while High-nSe-Ti had cell density decreased 4.5 times compared to uTi. Again,
this difference in inhibition of cancerous cell growth between selenium-coated stainless
steel substrates and titanium substrate is likely due to the smaller coating densities on
stainless steel substrates.
2.5. Conclusions and summary
To investigate the potential of using selenium nanoclusters as a coating material
for orthopedic applications, selenium nanoclusters were coated on conventional
orthopedic materials including titanium and stainless steel. The coatings on titanium
samples were investigated in more details because titanium was chosen as the model base
material due to its popularity in orthopedics. Titanium samples with higher coating
densities of selenium nanoclusters showed greater osteoblast adhesion, proliferation and
differentiation compared to lower coating densities of selenium nanoclusters as well as
uncoated titanium. Proliferation of cancerous osteoblasts was reduced on the samples
coated with selenium nanoclusters. The inhibition of cancerous osteoblasts is likely due
to the release of selenium into the culture media from the coatings. The promoted
functions of healthy osteoblasts on titanium surfaces coated with selenium nanoclusters
can be correlated to the increased adsorption of fibronectin which is an important protein
mediating osteoblast adhesion on implant surface.
Stainless steel samples were also investigated to demonstrate that the coating
process can be applied to other orthopedic metals. Stainless steel samples were coated
with selenium nanoclusters with various coating densities. The selenium coated samples
showed maintained proliferation of healthy osteoblasts compared to uncoated samples.
53
Importantly, similar to the results obtained in selenium coated titanium samples,
proliferation of cancerous osteoblasts was inhibited on selenium coated stainless steel
samples.
Together, the results of this chapter suggest that selenium nanoclusters is a
promising coating material for promoting healthy bone growth while inhibiting cancer
growth for orthopedic applications. To fully understand the observed healthy osteoblast
functions, more intensive studies of protein adsorption on selenium coated surfaces from
other adhesion proteins such as vitronectin and laminin may be required because they are
also important adhesion proteins for osteoblast functions. To fully understand the
observed inhibition of cancerous osteoblasts functions on selenium-coated substrates,
more in-depth analysis of effects of selenium released into culture media on cancerous
osteoblasts should be conducted.
54
CHAPTER 3. COARSE - GRAINED MONTE CARLO COMPUTER SIMULATION OF FIBRONECTIN ADSORPTION ON NANOMETER
ROUGH SURFACES
3.1. Introduction
It was shown in Chapter 2 that increasing selenium nanocluster coating density
led to the increase in fibronectin adsorption and that the increased fibronectin adsorption
promoted healthy osteoblast functions. The objective of this part of the thesis is to
understand the factor that gave rise to increased fibronectin adsorption. The increase in
fibronectin adsorption might be attributed to factors such as substrate surface
hydrophobicity/hydrophilicity, morphology, roughness or a combination of these factors.
For example, elongated proteins (such as fibronectin) can adsorb in different directions.
They may adsorb onto a surface with the main axis oriented perpendicularly to or along
the surface. A rough surface with the rough features in the same length scale with the
proteins may have more proteins adsorbed perpendicularly to the surface because of the
increase in steric hindrance caused by the increased surface roughness. A protein
adsorbed perpendicularly to the surface will occupy less area on the surface. Therefore,
the surface with increased roughness may have increased protein adsorption compared to
a flat surface.
To further understand the influence of surface roughness on fibronectin
adsorption, in this chapter, a coarse-grained Monte Carlo computer simulation was used
to reproduce experimental results. First, a general
55
introduction of protein adsorption onto a biomaterial surface is presented, followed by a
quick summary of some of the established physical models of protein adsorption. Then a
simple computer method was used to simulate rough surfaces. Finally, surfaces with pre-
determined root mean square (RMS) roughness were generated and used for simulating
fibronectin adsorption. Simulation results and experimental results were then compared.
3.2. Protein adsorption onto biomaterials
Protein adsorption onto solid surfaces is a complex yet interesting problem for
many fields of studies (such as biotechnology, biomedicine, biology, biochemical
engineering etc.). Especially, in biomaterial applications, where a material is designed to
interact with biological objects, protein adsorption on a surface plays an important role as
it is the first event that occurs when a material is placed in contact with biological fluids.
This event will affect subsequent reactions of biological entities (such as cells and
tissues) with a material. For example, when a material is inserted into humans or animals,
proteins adsorb onto the surface of the material almost instantaneously (within seconds)
after implantation. Cells arrive to the surface after much longer times (on the order of
minutes). Therefore, cells will interact directly with the adsorbed protein layer instead of
the material. Thus, early cellular events (such as cellular adhesion and morphology)
depend largely on the adsorbed protein layers (specifically, what proteins are present,
their amount and bioactivity).
The interaction between proteins and a solid surface depends on both properties of
the protein (such as hydrophobicity/hydrophilicity, charge, conformation, etc.) and the
properties of the surface (such as the hydrophobicity/hydrophilicity, elemental
composition, charge, roughness, morphology, etc.). Protein adsorption is a very complex
56
process driven by many surface-protein interactions (such as hydrophobic/hydrophilic
interactions, electrostatic interactions, van der Waals interactions and formation of ionic,
hydrogen or covalent bonds [70]). For example, hydrophobic regions on a protein adsorb
to hydrophobic surface regions and similarly, hydrophilic regions of a protein adsorb to
hydrophilic surface regions. Since proteins usually have regions with different charges
(even if the whole protein is neutral), electrostatic interactions play an important role if
the solid surface acquires some surface charge in solution. The electrostatic interaction
usually becomes complex if there are ions (such as calcium ions) in solution which will
also interact with the protein and the solid surface.
Fibronectin is one of the most important adhesive proteins that mediate
anchorage-dependent cell (such as osteoblasts, fibroblasts, endothelial cells, etc.)
adhesion. Fibronectin is a high molecular weight extra-cellular matrix protein consisting
of two nearly identical ~ 250kDa monomers connected by disulfide bonds [71] (Figure
3.1).
57
Figure 3.1. Schematic structure of fibronectin showing two monomers linked by disulfide bonds (Reprinted with permission from [72]).
Within seconds after implantation, proteins (including fibronectin) arrive and
adsorb onto the surface of the implant. Anchorage-dependent cells arrive at the implant
surface later and, therefore, will interact directly with the adsorbed protein layer. Among
many proteins that can adsorb on the surface of an orthopedic implant material (such as
albumin, laminin, vitronectin, etc.), adhesive proteins (such as fibronectin and
vitronectin) play an important role for cell adhesion which is a crucial prerequisite for
subsequent cell functions (such as proliferation, and for osteoblasts, synthesis of
extracellular matrix proteins and formation of calcium-containing mineral deposits).
These adhesive proteins have certain peptide sequences that specific cells can recognize
and adhere to. For example, osteoblasts recognize and adhere to the arginine-glycine-
aspartic acid (RGD) peptide sequence in fibronectin. Studies have shown that interrupting
58
the interactions between osteoblasts and fibronectin in cell culture media results in a
sharp decrease in osteoblast functions. For example, osteblasts were grown in culture
media added with anti-fibronectin (an antibody to fibronectin at a concentration of
100µg/mL) and showed a 10% reduction of mineralization by osteoblasts compared to
cells grown in control conditions (i.e., culture media without anti-fibronectin) [73].
Interfering with fibronectin-osteoblast interactions was also reported to result in reduced
osteoblast attachment [74] and even complete nonadherence of osteoblasts to implant
materials [75]. Coating a material surface with fibronectin (at a concentration of
~10µg/mL) has been a method to increase the attachment of osteoblasts to an implant
surface [74, 76-79]. The concentration of fibronectin in plasma is from 300-400 µg/mL
[80]. Osteoblasts have also been shown to adhere more strongly to higher density
fibronectin-coated surfaces [81, 82]. Specifically, using the micropipette aspiration
technique, osteoblasts were determined to adhere 80% more strongly on a titanium
surface coated with fibronectin compared to an uncoated substrate [82].
3.3. Roles of rough surfaces
Rough surfaces are encountered more often than flat surfaces. More strictly
speaking, there is hardly any surface which is perfectly flat. For many problems where
interaction of molecules (such as proteins) with a surface is investigated using computer
simulation, there is a great interest in generating rough surfaces with pre-determined
properties such as the dimension of surface structures (for example, pores, groves, etc.),
hydrophobicity/hydrophilicity properties, the roughness of the surfaces, etc. Among
many surface properties that require simulating, surface roughness is a very important
factor. Surface roughness can affect electrical properties in semiconductor materials
59
(such as electrical capacity, electronic conductivity, peak electric field and sheet
resistance [91]). In biomaterials, surface roughness can influence the adsorption,
conformation, and denaturation of proteins when the material is placed in contact with
biological fluids (as this thesis has demonstrated). In addition, there is a growing interest
in computer simulation of the interactions of biological entities (such as proteins and
cells) with biomaterial surfaces because it is important to have a mechanistic
understanding of these interactions in order to engineer desired materials. However, most
of the methods employed in the literature (for example, the Fast Fourier Transform (FFT)
method and artificial neural networks method) are complex and involve complicated
mathematic manipulations. In the next section, a simple method to generate rough
surfaces with pre-determined RMS values is proposed.
3.4. Simple method to generate rough surfaces with pre-defined RMS roughness
To generate a rough surface having a random morphology, but defined RMS
roughness, the idea of variance of uniformly-distributed random numbers was employed
here. Consider a set of random variables, X, which have a uniform continuous distribution
in the interval [a,b]: {X}=U[a,b]:
bxaforab
bxoraxforxf1
0)(
(Equation 3.1)
where f(x) is the probability function, a and b are the two values defining the interval
(a<b).
60
The variance of X, defined as Var(X)=E[X-E(X) 2], where E(X) is the mean of X,
is:
12
)(2
)(
abXVar uniform
(Equation 3.2)
If instead of a continuous uniform distribution, X has a discrete uniform
distribution in the interval [a,b]; i.e., {X}=Udiscrete[a,b]:
bkaforn
bkorakforkp1
0)((Equation 3.3)
where n = b-a+1 and p(k) is the probability function, the variance of X is given as:
12
1)(
2
)(
nXVar discrete
(Equation 3.4)
Now, the explicit expression of the variance of discrete, uniformly distributed
variables X is:
N
xxXVar
N
imeani
1
2)()(
(Equation 3.5)
Now, consider the definition of RMS:
61
N
xxRMS
N
imeani
1
2)((Equation 3.6)
It can be seen that:
)(XVarRMS (Equation 3.7)
Therefore, a random, rough surface with a pre-defined RMS roughness can be
generated by using the concept of uniformly distributed discrete variables. Specifically,
the problem of generating a 3D rough surface with RMS roughness values of R using
Matlab consists of the following steps:
Step 1: Determine the interval [0, a] of random numbers to be generated. The
value of a is determined from the requirement that the RMS is equal to R:
12)(
aXVarRMSR
(Equation 3.8 )
Step 2: Define a dimensionless roughness parameter, r, which is the normalization
(i.e., dividing) of the actual RMS value R by the dimension Do of an appropriately chosen
parameter. For example, if adsorption of a specific protein onto the surface is the problem
of interest, Do can be the smallest dimension of the protein. Do is the unit length in the
simulation (i.e., the smallest length scale in the problem).
0D
ar
(Equation 3.9)
62
Step 3: Generate a matrix NxN of random numbers, X, that have a uniform,
continuous distribution in [0, r].
Step 4: Convert X into X* whose elements are the nearest integers greater than or
equal to elements in X.
If N is large enough, the X* “surface” will have a random morphology with the
RMS value r.
The method described above was used here to create the High-nSe-Ti surface in
this study with an RMS value, R, equal to 14.113 nm. From the value of R, the value of a
was calculated to be equal to 48.878nm from the equation R2=a2/12. If the problem of
interest in this study is the adsorption of a globular protein whose diameter Do is 9nm, the
dimensionless parameter, r, is calculated by dividing a by Do, or r = 5.431. Now, a N x N
matrix of random variables X whose distribution is uniform in the interval [0, 5.431] can
be generated by using the rand() function in Matlab: X=5.431 rand(N,N). However, in
order to use the surface in computer simulation, the height of the points on the surface
needs to be discretized, this can be done by using the ceil() function in Matlab which
rounds up the value of the height to the nearest integer greater than or equal to the height:
X*=ceil(X). As can be from Table 3.1, the method converges as N increases.
63
Table 3.1. Convergence property of the proposed method to generate surface High-nSe-Ti with a defined RMS roughness. Dimensionless simulated roughness, Ro, is the square root of variance of the heights of all points on the simulated surface. Simulated roughness is then determined as R=RoDo (in this case, Do=9nm).
N Dimensionless simulated
roughness, Ro Simulated
roughness, R(nm) Actual
roughness (nm)50 1.594 14.346
100 1.576 14.184 250 1.575 14.175 500 1.570 14.130
14.113
This method was used to generate surfaces having RMS roughness values the
same as the experimental substrates (i.e., uTi, Low-nSe-Ti, Medium-nSe-Ti and High-
nSe-Ti) with unit length Do=9nm and N=500. The generated surfaces are referred to as
S-uTi, S-Low-nSe-Ti, S-Medium-nSe-Ti and S-High-nSe-Ti. Figure 3.2 shows
representative regions on the generated surfaces.
64
Figure 3.2. Representative regions on the four rough surfaces generated in this study with the RMS roughness values equal to experimental values (unit lengh Do=9nm, N=500). Colors are to aid in visualization only.
3.5. RSA model of fibronectin adsorption on uncoated and selenium coated titanium
surfaces
The ELISA method used in this study to detect and quantify the amount of the
adsorbed fibronectin on substrate surfaces relies on the recognition and attachment of
antibodies to proteins. As the result, the ELISA method detects only the top layer of
adsorbed proteins (if there is multi-layer adsorption). Therefore, the simulation in this
study will assume a monolayer adsorption of fibronectin on the substrates.
65
The concentration of fibronectin used in this study (i.e., adsorption onto surfaces
of fibronectin from the solution of fibronectin in PBS) was 5µg/mL. This value of
fibronectin concentration is smaller than physiological concentrations of fibronectin
(which is ~ 300 µg/mL in blood plasma) but is close to the concentration of fibronectin in
the cell culture media (i.e., DMEM supplemented with 10% FBS, 1% P/S, which is the
optimal medium for culturing osteoblasts in vitro has approximately 2.5 µg/mL of
fibronectin). This concentration of fibronectin (i.e., 5µg/mL) is in the range of
concentrations that the RSA model fits best [89, 92]. In addition, many reports in the
literature use the RSA model to describe the adsorption of proteins onto solid surfaces
with good agreement between experimental results and predictions of the RSA model
[93-95]. The size of fibronectin depends on factors such as its folding and the
surrounding environment ionic strength, therefore, dimensions of fibronectin reported in
the literature vary significantly. Unfolded fibronectin is 140nm in length and 2nm in
diameter [96]. Fibronectin is in more compact form in near - physiological conditions
[97]. Fibronectin were reported to be in a compact, elongated structure with the length of
approximately 15nm and the width of approximately 9nm (Figure 3.3)[98].
66
Figure 3.3. Gallery of transmission electron microscopy images of individual fibronectin molecules freeze-dried in vacuum and shadowed by tungsten-tantalum. (Freeze-drying technique was used to preserve the native structure of fibronectin and tungsten-tantalum shadowing was used to give contrast for imaging). Most fibronectin molecules had elongated shapes with the length of ~ 15 nm and the width of ~9 nm. Magnification is 300,000. (Adapted with permission from [98]) .
An elongated shape of fibronectin molecule in physiological pH (7.4) with the
length of ~ 24 nm and the width of ~ 16 nm was also observed using transmission
electron microscopy [99]. Therefore, in this thesis, fibronectin molecule was modeled as
a hard rod with dimensions of 18nm x 9nm x 9nm. The smallest dimension value, 9nm,
was chosen to be the width of fibronectin reported in [98] and the length of the rod,
18 nm, was chosen to be near the average of the two values of the length of fibronectin
molecule reported in [98] (15 nm) and [99] (24 nm). Desorption and diffusion of
67
adsorbed protein on the surface was assumed to be negligible. Fibronectin is assumed to
adsorb randomly on the surface with equal probability, i.e., any point on the surface can
be the point for protein adsorption. For simulation, a fibronectin molecule was modeled
as a dimmer consisting of 2 monomers (see Figure 3.4).
Figure 3.4. A 2D picture of fibronectin adsorbing onto a rough surface. Fibronectin was modeled as a dimmer consisting of 2 monomers and the surface was modeled as consisting of columns. Each point on the surface was defied by two coordinates (i,j). Fibronectin can adsorb on all the "unoccupied" sites (in this picture, the right and the left sides of point (i,j) are unoccupied, the top and bottom sides of point (i,j) are occupied).
In a 3D simulation, each point on the surface was specified by 3 coordinates
(x,y,z) where z is the height. Fibronectin was assumed to adsorb sequentially onto the
surface. Each adsorption trial started by randomly choosing a point (specified by x, y and
z) on the rough surface for one end of the protein to “attach” to (this end will then have a
“contact”). The direction of the protein was then varied. If the adsorption of one of the
monomers was impossible due to steric constraints, the trial was rejected. If steric
constraints were satisfied, proteins would adsorb in the direction that had more contacts.
Figure 3.5 demonstrates examples to clarify this description.
68
Figure 3.5. Examples of the simulation of protein adsorption onto a rough surface. For the case of the right side of the point specified by (i,j) (the bold side), there are two possible directions the protein can adsorb: horizontal direction (a) or vertical direction (b) (Note, horizontal adsorption to the left of the bold side is prohibited due to steric constraints). In this case, configuration (a) will have 4 contacts while (b) will have only 2 contacts upon adsorption, therefore, configuration (a) is more favorable for protein adsorption. For the case of the point on the far right (bold side), both the (c) and the (d) configuration will have the same number of contacts (3 contacts) upon adsorption, therefore, the protein will adsorb following either (c) or (d) with equal probability.
“Overhang” is not allowed, i.e., it is prohibited to have protein adsorption at a site
in which the site immediately underneath is not occupied. Overhang will be immediately
shifted downward to reach the nearest occupied site. Figure 3.6 demonstrates examples of
overhangs which would be immediately shifted downward (dashed arrows).
69
Figure 3.6. Overhang is not allowed. Overhang would be immediately shifted downward as indicated by the dashed arrows.
To generate surfaces of uncoated and selenium-coated titanium substrates, the
method described in section 3.5 was employed here. Specifically, with the RMS
roughness determined experimentally, the parameters for the surface were calculated and
summarized in (Table 3.2).
Table 3.2. Parameters for simulation (a) calculated from experimental RMS roughness as described in section 3.5 and the RMS roughness of the simulated surfaces. The simulated surfaces resemble the experimental surfaces quite well in terms of RMS roughness.
Sample Experimental RMS roughness (nm)
a Simulated RMS roughness (dimensionless) L=500x500
Simulated RMS roughness (nm) L=500x500
uTi 8.07 27.97 0.88 7.93 Low-nSe-Ti 9.72 33.67 1.08 9.73 Medium-nSe-Ti 12.56 43.53 1.34 12.51 High-nSe-Ti 14.11 48.88 1.58 14.22
70
As mentioned, the generated surfaces will be referred to as S-uTi (simulated-uTi),
S-Low-nSe-Ti, S-Medium-nSe-Ti and S-High-nSe-Ti. Simulation was implemented with
size of the substrate NxN=500x500 and the total number of adsorption trials,
M=6,250,000. With this choice of N and M, each point on the surface of all the substrates
had at least 5 protein adsorption trials. This choice of M was shown (by simulation) to be
efficient enough to achieve saturation adsorption on all the surfaces (as indicated by less
than 2.6% of the surface area unoccupied after M adsorption trials).
The amount of adsorbed protein was normalized by a monolayer of proteins
adsorbed on a perfectly flat surface which had the same geometric surface area.
Simulation was implemented first in 2D to verify the validity of the algorithm. A
snapshot of the surfaces during the simulation is shown in Figure 3.7.
71
Figure 3.7. A snapshot of simulation process in 2D with N=500 and M=25000. Blue columns represent proteins. Grey columns represent rough surfaces. The adsorption did not reach saturation yet as some areas on the surface are still unoccupied.
72
With the choice of N x N and M indicated, simulations in 3D were implemented
with 3 runs for each substrate; results reported as the average of 3 runs are shown in
Figure 3.8.
Figure 3.8. Simulation results of fibronectin saturated adsorption on all four substrates of interest. The protein amount was normalized to a monolayer of protein on a perfectly flat surface with the same geometrical surface area N x N. Substrates were generated to have experimental RMS roughness values. Protein adsorption simulation was implemented with the size of the substrates N x N=500x500 and the number of simulation M=6,250,000. Data = mean ± standard error of the mean (n=3 runs).
Simulation results showed a small increase in fibronectin adsorption on the Low-
nSe-Ti substrate compared to uTi. However, fibronectin adsorption on Medium-nSe-Ti
and High-nSe-Ti significantly increased compared to both uTi and Low-nSe-Ti. This
trend is similar to the trend in the experimental results (Figure 2.23 in Chapter 2).
To better compare the simulation results with experimental results, the results
were expressed as the increase compared to uTi (the number of times the amount of
fibronectin increased compared to the uncoated substrate). As shown in Figure 3.9,
computer simulation results fit quite well to the experimental results.
73
Figure 3.9. Comparison of the increase in the amount of fibronectin adsorption (expressed as times increase compared to uTi) between simulation results and experimental results on uncoated and selenium coated titanium substrates.
The good agreement between computer simulation and experimental results
indicated that the simple RSA model employed in this study is appropriate to describe the
adsorption process of fibronectin on nanometer rough surfaces. It is appropriate here to
evaluate some other physical quantities from the simulation (such as the amount of
adsorbed fibronectin normalized to geometrical surface area of the substrate). Recall that
the unit length in the simulation corresponded to 9nm in the experiment. The molecular
weight of fibronectin is approximately 440kDa. From the total protein adsorbed on the
substrates, the normalized protein amount (expressed as µg/cm2) was calculated and
presented together with experimental results for comparison in Figure 3.10.
74
Figure 3.10. Simulation and experiment results on the amount of adsorbed fibronectin normalized by geometrical surface area. The simulation values of normalized fibronectin adsorption are greater than experimental values.
As can be seen from Figure 3.10, simulation results are approximately 80% higher
than corresponding experimental results. The difference in the values for the simulation
and experiment results is an indication of the simplicity of the RSA model employed in
this study. However, the trend is similar for both simulation and experiment results (as
can be seen from Figure 3.8 and previously analyzed and presented in Figure 3.9). It is
likely that the higher values of the simulation results on fibronectin adsorption compared
to experimental results are due to the fact that desorption of adsorbed fibronectin was
ignored in the simulation model.
Another interesting physical quantity that can be extracted from the computer
simulation is the amount of adsorbed protein per unit area of the rough surface. As the
roughness of a surface is increased, the surface area of the surface also increases. Assume
that the “binding site” densities on a rough surface and a perfectly flat surface (of the
same material and the same geometrical area) are the same. Intuitively, it is expected that
75
the densities of a protein adsorbed on the flat surface and the rough surface (per unit
surface area) are the same (Figure 3.11).
Figure 3.11. Protein adsorption on a flat surface and a rough surface. Assuming the same "binding site" densities, the protein density (per unit surface area) d1 and d2 should be the same.
However, as seen in Figure 3.12, computer simulation predicts that surfaces with
increased roughness will have decreased adsorbed protein density per unit surface area.
Figure 3.12. Computer simulation results on fibronectin density expressed as microgram of fibronectin per cm2 of the surface area. Surfaces with higher RMS roughness had lower fibronectin density.
This prediction by computer simulation can be understood as follows. The model
employed in this study assumes that proteins will try to maximize contact with the
76
substrate during adsorption. In addition, in this model, proteins can only adsorb in one of
two directions: horizontal direction and vertical direction. Therefore, an increase in
surface area will not necessarily lead to an equal increase in protein adsorption. An
example showing that increases in surface area do not necessarily lead to equal increases
in protein adsorption is shown in Figure 3.13.
Figure 3.13. An example illustrating that increased surface area does not necessarily lead to equal increase in protein adsorption. In this example, the rough surface area increases ~67% but the amount of fibronectin increases only 25%.
Future experiments that can accurately determine the density of adsorbed
fibronectin per unit surface area (rather than per unit geometrical surface area) need to be
implemented to verify this prediction.
3.6. Summary and conclusions
In this chapter, a simple method for generating rough surfaces with pre-
determined RMS roughness was proposed using the idea of uniformly-distributed random
variables. The method was used to generate four substrates with RMS roughness values
equal to the experimental values of the uncoated and selenium coated titanium substrates.
A simple random sequential adsorption model was used to simulate the adsorption of
77
fibronectin on the generated surfaces. Simulation results of the increase in fibronectin
adsorption on coated surfaces compared to uncoated surfaces were found to be in a good
agreement with experimental results. Further development is needed to generate rough
surfaces, for example, to simulate not only random but also ordered-rough surfaces. The
RSA model also needs to be improved to include more complex process in the adsorption
(such as desorption, diffusion, and denaturation of the adsorbed proteins). For example,
to account for denaturation process, the adsorbed proteins can be modeled to have a size
that is larger than that of the non-adsorbed ones.
78
CHAPTER 4. A MODIFIED CONTACT ANGLE MODEL TO FIT EXPERIMENTAL RESULTS
In this chapter, a brief introduction to the roles of contact angles in biomaterials
will be presented and followed by an introduction to two theoretical models of contact
angles (i.e., the Wenzel model and the Cassie-Baxter model). After that, the needs to
modify the Wenzel model and Cassie-Baxter model to fit the experimental results on
contact angles presented in Chapter 2 will be discussed. Finally, a modified contact angle
model will be proposed to fit the experimental results.
4.1. Introduction: role of contact angles in biomaterials
Contact angles easily and quickly identify the hydrophobicity/hydrophilicity of a
proposed material for medical applications; this is of paramount importance for the
interaction of a material with biological entities in the body. When an implant is inserted
into humans or animals, proteins absorb onto the surface of the implant within seconds
after implantation. Cells arrive at the implant surface much later (within minutes),
therefore, directly interacting with such absorbed proteins. As the result, early responses
of cells to a biomaterial (such as adhesion, migration, spreading, etc.) depend largely on
the adsorption of proteins (i.e., which type, amount and consequent bioactivity) onto the
material surface which, in turn, depends on the specific properties of the material surface
(such as hydrophilicity/hydrophobicity, surface charge, surface morphology, surface
roughness, etc.). Among the protein-surface interactions, hydrophobic/hydrophilic
79
interactions play an important role. Hydrophilicity and hydrophobicity are clearly the
water-attracting and water-repelling properties of a material. If the material is water-
repelling, it is hydrophobic. If it is water-attracting, it is hydrophilic. A more rigorous
definition of hydrophilic and hydrophobic materials will be given in the next section.
Nanostructured materials, especially materials having surface with nanofeatures have
increased surface area, different surface morphology, surface roughness, etc., and,
importantly, will have different hydrophilicity properties compared to flat surfaces (of the
same chemistry).
Proteins usually have both hydrophobic regions and hydrophilic regions (for
example, see Figure 4.1 for the hydrophobic and hydrophilic regions in a fibronectin
molecule). Hydrophobic regions on a protein attach to hydrophobic regions on a material
surface and, vice versa, hydrophilic regions attach to hydrophilic regions on material
surface.
80
Figure 4.1. Atomic and ribbon chains of an extended human fibronectin molecule. Inset shows that hydrophilic parts are blue colored regions while hydrophobic parts are red colored in modules (Reprinted with permission from [100]).
In this part of the study, an introduction to contact angle fundamentals and the
Young’s equation will be discussed first followed by a brief introduction into two
physical models of contact angles on rough and composite surfaces. Finally, a modified
contact angle model will be proposed and discussed to fit the experimental results from
this study.
81
4.2. Contact angle and Young’s equation
The hydrophilicity/ hydrophobicity of a material is usually determined by
measuring contact angles of drops of liquid on the material surface (as illustrated in
Figure 4.2).
Figure 4.2. Contact angle, θ, of a liquid on a solid surface depends on the interaction of
the liquid with the solid, liquid with vapor and solid with vapor. SL , SV , and LV are interfacial energies at the solid-liquid interface, solid-vapor interface and liquid-vapor interface, respectively.
If the material is water-repelling (hydrophobic), a water drop on the surface of a
material will have a contact angle, θ, greater than 90 degrees. Otherwise, if the material is
water-attracting (hydrophilic), the contact angle will be less than 90 degrees. Consider a
drop of liquid on a solid substrate. The contact angle at equilibrium is θ. Imagine if the
liquid-solid interface is displaced by dx, the increase in the liquid-vapor interface is
dx.cos(θ) (Figure 4.3).
82
Figure 4.3.Schematic demonstration of the derivation of Young's equation. When the solid-liquid interface is displaced by dx, the increase in liquid-vapor interface is dx.cos(θ).
The change in total energy associated with this displacement is
)cos()( dxdxdE LVSVSL (Equation 4.1)
A stable contact angle corresponds to dE/dx=0:
0)cos()( LVSVSL(Equation 4.2)
or
LV
SLSV
)cos((Equation 4.3)
again where SL , SV , and LV are interfacial energies at the solid-liquid interface, solid-
vapor interface and liquid-vapor interface, respectively. This is Young’s equation of
contact angle.
Young’s equation provides a quantitative relationship between the contact angles
and physical properties of a solid, liquid and vapor. In Young’s equation, the solid
surface is assumed to be perfectly flat.
4.3. Wenzel model versus Cassie-Baxter model on contact angles
To describe the effects of different surface roughness on the wetting behavior of a
solid surface, Wenzel considered a drop of water on a solid surface. The surface energy
83
of the wetted area (i.e., under the drop), σwetted, will be, in general, different from that of
the non-wetted area (i.e., not covered by the drop), σnon-wetted. Let’s assume for a moment
σwetted < σnon-wetted . In this case, the wetting process is involved in the decrease of the
surface energy or wetting is energetically favorable, in another words, the surface is
hydrophilic. Now, if the surface under the drop is rough instead of flat, the decrease in
surface energy of the wetting process will be greater. Therefore, rough surfaces will be
more easily wetted, or more hydrophilic.
The same reasoning applies to the case of a surface that has σwetted > σnon-wetted , or
a water-repelling (hydrophobic) surface. There will be a net increase in surface energy
during the wetting progresses. If the surface under the drop is rough instead of flat, the
actual wetted surface area will be greater than the geometric area. As the result, a net
increase in surface energy during wetting will be greater. In this case, roughness will
make the surface more hydrophobic.
Wenzel introduced the concept of a “roughness factor”, r, that describes the
increase in surface area of a rough surface compared to the geometric area:
areasurfacegeometric
areasurfaceactualr
(Equation 4.4)
where r is always greater than 1 except in the case of a perfectly flat surface where the
actual surface and geometric surface are identical.
Now, consider a drop of a liquid on a rough surface which has roughness factor of
r. At equilibrium, the contact angle is θ. Imagine if the liquid-solid interface is displaced
84
by dx, the increase in the liquid-vapor interface is dx.cos(θ), while the actual liquid-solid
interface increase associated with this displacement is r.dx (Figure 4.4).
Figure 4.4. Schematic demonstration of the derivation of the Wenzel equation. When the solid-liquid interface is displaced by dx, the increase in liquid-vapor interface is dx.cos(θ), while the actual liquid-solid interface increase associated with this displacement is r.dx.
The total energy change is then:
)cos()( dxrdxdE LVSVSL (Equation 4.5)
where E is the total energy of the system, SL , SV , and LV are the surface tensions at
the solid-liquid interface, solid-vapor interface and liquid-vapor interface, respectively.
A stable contact angle θ* corresponds to dE/dx=0; or
0)cos()( * LVSVSL r (Equation 4.6)
or:
85
)cos()(
)cos( *
rr
LV
SLSV
(Equation 4.7)
where θ is the contact angle of the same solid, but with a smooth surface.
The above equation indicates that if the surface which is hydrophilic (contact
angle θ < 90 degrees, cos(θ) >0) is roughened, cos(θ*) will be greater than cos(θ), or θ*
will be smaller than θ or the surface becomes more hydrophilic after being roughened. In
the other words, hydrophilic surfaces become more hydrophilic after being roughened.
Similarly, if θ is greater than 90 degrees, or cos(θ) < 0, cos(θ*) will be more negative
(compared to cos(θ)), or θ* will become greater than θ (and greater than 90 degrees). In
the other words, hydrophobic surfaces become more hydrophobic after being roughened.
The underlying assumption in the Wenzel model is that the liquid-solid interface
follows the solid surface, i.e., wetting is homogeneous (Figure 4.5).
Figure 4.5. Homogenous wetting assumption in the Wenzel model. The liquid-solid interface is assumed to follow a solid surface.
86
Cassie and Baxter proposed a model for the wetting on a heterogeneous surface
consisting of two different materials. The derivation of Cassie-Baxter equation is
summarized here. Consider a composite surface of two species with a surface fraction of
f1 and f2 (f1+f2=1). Suppose that the liquid drop is at equilibrium with a contact angle θ.
Now imagine the interfacial boundary is displaced by a small amount, dx (Figure 4.6).
Figure 4.6. A schematic demonstration of the derivation of the Cassie-Baxter model. When the solid-liquid interface increases by dx, the liquid-vapor interface increases by dx.cos(θ).
The energy change associated with this increase will be:
)cos()()( ,2,22,1,11 dxdxfdxfdE LVSVSLSVSL (Equation 4.8)
where SL , SV , LV are surface tensions at the solid-liquid interface, solid-vapor
interface and liquid-vapor interface, respectively. Indices 1 and 2 denote species 1 and 2,
respectively.
A stable contact angle θ corresponds to dE/dx=0, or
0)cos()()( ,2,22,1,11 LVSVSLSVSL ff (Equation 4.9)
87
The above equation can be re-written to obtain the Cassie-Baxter equation as
follows:
0)cos()()( ,2,2
2,1,1
1
LV
SVSL
LV
SVSL ff(Equation 4.10)
or
)cos()cos()cos( 2211 ff (Equation 4.11)
The Cassie-Baxter model of wetting can be applied to the case of a rough surface
where the roughness can create air-pockets when wetting happens (Figure 4.7).
Figure 4.7. The Cassie-Baxter model of heterogeneous wetting. In this case, the liquid is considered to contact a surface consisting of the solid material and air. The width of spikes as well as the spacing between them is very small compared to the size of the liquid drop. The scale used in this picture is to aid in visualization.
88
In this case, because the contact angle of a liquid drop on air is considered be 180
degrees, the above equation can be re-written as:
)180cos().1()cos(.)cos( 0111 ff (Equation 4.12)
or
1))cos(1()1()cos(.)cos( 11111 fff (Equation 4.13)
where f1 is the area fraction of the solid-liquid interface.
It was shown that the Wenzel model has been typically applied to surfaces with
low or moderate roughness (r less than 1.2) [101, 102]. However, the range of roughness
ratio, r, for the Cassie-Baxter “regime” is less consistent among different reports. Some
have reported the Cassie-Baxter regime at r values greater than 1.2 [101], others have
reported the regime at r values greater than 2.3 [101, 102]. In practical term, the latter is
usually used for surfaces that have spike-structures where air can be trapped between the
spikes.
4.4. The needs to modify the Wenzel model and Cassie-Baxter model to fit experimental
results
As presented in chapter 2, contact angle measurements on uncoated titanium and
selenium-coated titanium substrates possessed some interesting findings (Figure 4.8).
89
Figure 4.8.Water contact angles on uncoated titanium and titanium coated with increasing selenium nanocluster densities. Data = mean ± standard error of the mean, n=5.
Specifically, uncoated titanium substrates were hydrophilic (i.e., had contact
angle less than 90 degrees). Coating titanium substrates with selenium increased the
substrate roughness (r greater than 1). Therefore, if differences in chemistry between
selenium and titanium are ignored, the coated substrates should become more hydrophilic
(i.e., the contact angle should decrease) according to the Wenzel theory. However, the
experimental data showed an opposite trend. Therefore, the differences in chemistry of
selenium and titanium should be taken into account. Thus, the Cassie-Baxter model,
which describes the contact angle of a composite surface as a function of contact angles
of the components, was used here to explain the wetting behavior of the selenium coated
surfaces:
)cos()cos()cos( 2211 ff (Equation 4.14)
where is the water contact angle on the selenium-coated titanium substrates, 1 is the
water contact angle on the uncoated titanium substrate, 2 is the water contact angle on
the selenium substrate, f1 is the surface fraction of titanium, and f2 is the surface area
90
coverage fraction of selenium, ρ, (f1 +ρ =1); ρ was determined from SEM images and is
given in Table 4.1.
Table 4.1. The selenium surface fraction (ρ, determined from SEM images), relative surface roughness ratio (r, normalized by that of the uncoated substrates), the relative root mean square roughness (R, normalized to that of uncoated substrates), and cos(θ), of the substrates of interest to the studies.
Substrate Selenium surface fraction, ρ, (%)
Relative surface roughness ratio, r
Relative RMS, R
cos(θ)
uTi 0 1 1 0.433 Low-nSe-Ti 2.4 1.0373 1.204 0.274
Medium-nSe-Ti 5.0 1.050 1.556 0.262 High-nSe-Ti 6.2 1.054 1.748 0.220
The relative surface roughness ratio was calculated by dividing the surface
roughness ratio of the sample by the surface roughness ratio of the uncoated substrates.
Similarly, the relative RMS was calculated by dividing the RMS of the sample by the
RMS of the uncoated substrate. 2 was measured on a flat elemental selenium surface
using the same method as described in section 2.2.2.4. 2 was determined to be 83.5±1.4
degrees.
Using Equation 4.12, a cos( )-versus-f2 curve was constructed based on the
Cassie-Baxter model on the selenium-titanium composite surfaces. However, the
experiment data did not fit well to this curve (Figure 4.9). The experimental values of
cos( ) were lower than predicted by the Cassie-Baxter model.
91
Figure 4.9.The Cassie-Baxter model curve (dashed line) and the experimental curve (solid line) of the cosine of water contact angles on uncoated titanium substrates and titanium substrates coated with various amount of selenium nanoclusters.
From the comparison of the cos(θ)-versus-selenium surface percentage curves
between the Cassie-Baxter model and the experiment data, it can be speculated that
during contact angle measurements on the selenium coated surfaces, there was formation
of air-pockets on the selenium coated samples that made the contact angles larger
(consequentially, smaller values of cos(θ)). It can be imagined that the higher selenium
coating density had higher RMS roughness values that, according to the Wenzel theory,
made the contact angles decrease. However, the surfaces with higher roughness could
also have more air-pockets that made the contact angles increase. Therefore, with the
contribution from both roughness and air-pockets, the contact angles of the three different
coating densities were not significantly different (Figure 4.10).
92
Figure 4.10. Possible mechanisms for the plateau in contact angles on the substrates with different selenium coating densities. Increasing coating density of selenium nanoclusters increased roughness that resulted in both increases in the roughness factor (which is the actual surface area to geometric area ratio) and the increase in the formation of air-pockets. The former made contact angles decrease while the latter made contact angles increase. The opposite contribution of the roughness factor and air-pockets is likely the mechanism for similar contact angles for the three selenium coating densities on titanium in here.
To quantify the contribution of each factor to contact angles, in this part of the
chapter, a modified Cassie-Baxter model will be introduced that will take into account the
contribution of surface roughness, selenium chemistry and the contribution of air-
pockets.
4.5. Modification of the Cassie-Baxter model
It can be seen from the discussion in the Cassie-Baxter model that air-pockets
formed during contact angle measurements will increase the contact angle compared to
the case of no air-pockets formed. Therefore, from the comparison of the cos(θ)-versus-
selenium surface percentage curves between the Cassie-Baxter model and experimental
data, it can be speculated that there were air-pockets which increased contact angles
(consequently, decreased values of cos(θ)). Also, because the roughness ratio, r, of the
selenium coated substrates were smaller than 1.1, it is reasonable to adapt the Wenzel
model to the Cassie-Baxter model as follows:
93
)cos()]cos()cos([)cos( 332211 fffr (Equation 4.15)
where r is the surface roughness ratio as defined by Equation 4.2; f1, f2 and f3 are the
surface area fractions of titanium, selenium and air, respectively; and θ1 ,θ2 , and θ3 are
contact angles on titanium, selenium and air, respectively (f1 + f2 + f3 =1 and θ3 =180o).
In this modified model the liquid drop contacts with all three materials: titanium,
selenium and air (as seen in Figure 4.10).
Figure 4.11. Schematic representation of the modified Cassie-Baxter model on the wetting process. A drop of liquid makes contact to all three materials: titanium, selenium nanoclusters, and air when placed on a selenium-titanium composite surface.
Clearly, f2 now is not necessarily the same as the surface fraction, ρ, of the
selenium nanoclusters as determined from SEM pictures. In the most general sense, all
three parameter (f1, f2 and f3) are dependent on ρ, r and R. To simplify this problem, it is
assumed that f2 and ρ are equal and f3 is a function of ρ, R and r. To further simplify the
94
model, since the values of r do not vary significantly among the substrates (see Table
4.1), it was assumed that f2 had no dependence on r and had the following form:
RCRBARff )1(),(3(Equation 4.16)
where A, B and C are fitting parameters. Because ρ, r and R are dimensionless by their
definitions (see Table 4.1), A, B and C are also dimensionless. Physically, A and B
represents the contribution to the air-pocket formation from (i) the surface coverage of
selenium on titanium and (ii) the increased RMS roughness of the composite surface,
respectively; C represents the inter-play of both (i) and (ii) to air-pocket formation
because they are not independent (specifically, increasing ρ increases R). The factor (R-1)
comes from the requirement that f3 is 0 when ρ=0 and R=1.
Now f1 can be calculated by:
))1((11 321 CRRBAfff (Equation 4.17)
Equation 4.13 can now be re-written as:
)cos(1
)cos())]cos()(cos()[cos()1(
1
121
r
rCRRBA
(Equation 4.18)
Solving a set of 3 linear equations with 3 unknowns one can determine that
A = 0.089, B = -0.184 and C = -0.030.
Recall that:
])cos()cos([)cos( 32211 fffr (Equation 4.19)
95
and
RCRBARff )1(),(3(Equation 4.20)
The positive value of A and negative value of B indicate the different contribution
to the air-pocket formation and, eventually, to the value of cos(θ). This different (in fact,
counteractive) contribution explains the plateau of the values of cos(θ) for the three
coating densities in this study. However, the negative value of B implies that increasing
surface roughness led to a decrease in air-pocket formation which is non-intuitive
because one expects there would be more air-pockets trapped between the gaps created
by the rough features on the selenium coated surface. This analysis can be understood as
follows. The underlying assumption of the Wenzel model is that the liquid-solid interface
follows the solid surface. In the meantime, the underlying assumption in the Cassie-
Baxter model for the case of air-pocket formation is that the liquid-solid interface does
not follow the solid surface. Therefore, the formation of air-pockets can be considered as
the indication of the degree of “non-Wenzel properties” in the contact angle
measurement. In other words, f3 also contains information about the “Wenzel properties”.
The fact that B is negative is in an agreement with this statement; increasing surface
roughness (i.e., increasing R) led to decreased f3, therefore, increased value of cos(θ) or
decreased contact angle.
Similarly, negative value of C indicates another contribution of Wenzel model in
the modified model proposed in this thesis.
96
4.6. Summary and conclusions
In this chapter, a modified Cassie-Baxter model for contact angles was proposed
to fit the experimental data. Air-pockets were hypothesized to form when measuring
contact angles of water on the selenium-coated substrates. The modified model was used
to quantify the contribution of surface chemistry, surface area and surface roughness to
the formation of air-pockets. It was shown that the formation of air-pockets was
attributed mostly to the surface roughness brought about by the selenium coatings.
However, it should be noted that air-pockets can also occur on the uncoated titanium
substrate when measuring water contact angles. Experimental methods should be
developed to visualize and quantify air-pockets formed during the contact angle
measurements.
97
CHAPTER 5. SELENIUM COATINGS ON POLYMERS FOR ANTI-INFECTION BIOMATERIAL APPLICATIONS
In previous chapters, it has already been shown that selenium can be readily
coated on a substrate surface to create various densities of nanoclusters. In addition to the
anti-cancer properties of selenium reported in Chapter 2 and in the literature, selenium
has also been shown to have anti-bacterial properties. Therefore, in this chapter, the use
of selenium nanocluster coatings for anti-bacterial applications was investigated. This
chapter is focused on bacterial infection on polymeric catheters as catheter-related
infection is the leading cause for extended hospital stays and increased cost of treatment.
5.1. Introduction
As discussed in Chapter 1, implant infection is a serious and complicated
problem. A common method to treat implant infection is though systemic antibiotic
therapy. However, this method has not shown satisfactory results to date. For example,
systemic antibiotic (vancomycin and gentamicin) administration alone without catheter
removal is only 22 to 37% effective in treating blood infection associated with using
long-term (more than 2 weeks) central venous catheters [4]. In addition, this mode of
antibiotic administration has the disadvantage of causing side effects and creating a pool
of antibiotic-resistant bacteria [10]. Another treatment method is the local delivery of
antibiotics. In this method, antibiotic molecules are incorporated (for example, by
adsorption or impregnation) onto the surface or into coating of the implants so that they
98
will slowly release locally over an extended period of time after implantation. Because of
the use of antibiotics, this method also has significant draw backs including
ineffectiveness against antibiotic-resistant bacteria (such as methicillin-resistant
staphylococcus aureus or MRSA, important bacterium that infects all implants). An
emerging approach to prevent implant infection is engineering raw implant surfaces to
resist bacterial attachment and colonization.
Since the implant surface itself is an important source of infection and bacterial
adhesion to the implant surface, is important in the pathogenesis of infection, the most
promising and straight forward strategy towards decreasing infection is to fabricate
implant materials that resist bacteria attachment [103]. However, currently used materials
used as implant have been shown to be ineffective to resist bacterial infection [104-107].
For example, silver and silver compounds (such as chlorohexidine-silver sulfadiazine),
one of the most widely studied materials for anti-bacterial coatings on catheters or for
impregnating catheters, have yielded mixed results towards preventing infection. Some
studies as early as the 1950s and later in the 1970s and 1990s, showed that silver coatings
could reduce the incidence of catheter-associated infection [108-111]. In contrast, a large
number of recent studies have demonstrated that catheters coated or impregnated with
silver or silver-compounds either did not reduce the incidence of bacterial colonization
and catheter-associated infection [104-107] or lacked efficacy [7, 112]. Lastly, coating
catheters with silver is expensive; endotracheal tubes coated with silver can cost as much
as $100 per tube.
As discussed in Chapter 1, selenium has been shown to have anti-bacterial
properties. Therefore, the specific objective of this part of the study was to test the anti-
99
bacterial properties of selenium nanocluster coatings on polymers that were created using
the novel method described in Chapter 2. Due to the popularity of catheter usage and
common catheter-related infections, this part of study focused on using selenium for anti-
bacterial catheter applications. Selenium nanoclusters were coated on three materials
commonly used for catheters: polyvinyl chloride (PVC), polyurethane (PU) and silicone,
using the same fabrication technique described in Chapter 2 and bacteria functions on the
selenium coated materials were determined.
The rest of this chapter is organized into sections on materials and methods,
results and conclusions. The materials and methods section describes the methods used
for creating, characterizing and testing the polymers coated with selenium. In the results
section, the highly anti-bacterial property of selenium nanocluster coatings was
demonstrated as direct evidence of the potential use of the coatings for anti-infection
catheters.
5.2. Materials and methods
A detailed fabrication and characterization method for coating PVC, PU and
silicone with selenium nanoclusters is introduced here. Bacterial culture assays to
investigate functions of bacteria on the polymer selenium coated materials is also
described.
5.2.1. Materials
Polymeric substrates (PVC, PU and silicone, in the form of discs 6mm in
diameter and approximately 2mm in height) were cut from PU (U.S. Plastic, catalog
number 48222), commercial PVC endotracheal tubes (Hudson RCI, catalog number 5-
100
10412), and commercial silicone tracheostomy tubes (Bivona, catalog number 60PFS30).
The polymeric substrates were then immersed in ethanol for 10 min, sonicated in ethanol
for 10 min for sterilization and air-dried in a sterile hood overnight. Cleaned and
sterilized polymeric substrates (PVC, PU, and silicone) were used as base substrates for
the colloidal decoration with selenium nanoclusters. For this, 10ml of deionized water
and 10ml of 100mM glutathione (GSH, reduced form, TCI America) were added to a
beaker followed by 10ml of 25mM sodium selenite (Na2SeO3, Alfa Aesar). After gentle
mixing, 2ml of 1N NaOH was added to the beaker to bring the pH of the mixture into the
alkaline regime. The beaker was left untouched for 10 minutes. The substrates were
removed and rinsed for 24 hrs in deionized water. The uncoated and coated polymeric
substrates (PVC, PU, and silicone) were immersed in ethanol for 30 min for sterilization
and air-dried in a sterile hood overnight. Bard’s (Covington, GA) commercially available
Agento I.C. Silver-Coated (Bard Medical, catalog number 500375) endotracheal tubes
were also used as a comparison.
5.2.2. Material characterization
Polymeric substrates (PVC, PU, and silicone) were sputter-coated with a thin
layer (approximately 90 Ao) of palladium and observed under SEM at 5 kV. For chemical
characterization, energy dispersed X-ray spectroscopy (EDS) with an acceleration voltage
of 20kV was employed. EDS was used instead of XPS because it is more convenient and
can be done at the same time as SEM surface characterization. For EDS, a beam of
electrons is focused onto the sample’s surface. The energy of the electron beam excites
electrons in the inner shells of atoms and creates holes in the inner shells. Electrons from
the outer shells (higher energy levels) follow a transition to fill the holes and the energy is
101
released in the form of X-rays. The energy of the X-rays is characteristic of the elemental
composition being analyzed.
5.2.3. In vitro assays for the polymeric substrate-bacterial experiments
5.2.3.1. SEM visualization of bacteria on the uncoated and coated substrates
S. aureus (ATCC, 25923), an important bacteria infecting all catheters was
prepared from a frozen vial received from the vendor. A sterile loop was used to
withdraw some of the bacteria from the frozen vial, streaked onto tryptic soy agar plates
(30g Tryptone [MP Biomedicals] and 15g agar [Sigma] per litter of distilled water) and
incubated for 12 hrs. Another sterile loop was used to select one colony from the agar
plates to place into 3ml trypic soy broth (TSB, 30g/L) and was incubated for 16 hrs.
Uncoated and selenium-coated substrates were placed into the wells of 24 well plates.
The bacteria were then seeded in TSB at approximately 20,000,000 bacteria per well and
inoculated for 8 hrs in a standard bacteria culture incubator (37ºC, 95% humidified air,
5% CO2 environment, non-shaking). At the end of the indicated time, the substrates were
removed from culture and treated according to standard protocols for SEM visualization.
Briefly, substrates were rinsed in 0.1M sodium cacodylate buffer (SCB) solution, fixed in
2.5% glutaraldehyde in 0.1M SCB for 30 min at room temperature (RT), rinsed in SCB,
dehydrated in a series of ethanol solutions (30%, 50%, 70%, 90%, 95%, 100% ethanol,
10 min each), left in 100% ethanol for 15 min, critically-point dried, sputter-coated with
gold and finally visualized under SEM (LEO 1530VP) at a voltage from 3 to 10 kV.
102
5.2.3.2. Crystal violet assays for bacteria quantification
To quantitatively compare the adherent bacteria on the different substrates, crystal
violet assays were used. After incubation of bacteria for 8 hrs, substrates were rinsed in
1% NaCl solution to remove all unattached bacteria. The adherent bacteria on the
surfaces were then stained with crystal violet (CV) which penetrated cell walls and cell
membranes. For this, substrates were immersed in 300µL of 0.01% CV (in water) for 15
minutes. After the indicated time, unbound CV was washed using 1% NaCl. The
substrates were then left air-dried overnight to make sure all the liquid was evaporated.
The bound CV (to cells) was solubilized in 100% ethanol. For this, 300µL of 100%
ethanol was added to each substrate and incubated at room temperature for 30 minutes.
Optical densities (ODs) of the soluble CV solutions (which correlate to total cell number)
were measured using a microplate reader (SpectraMax, Molecular Devices) and
compared among the substrates.
5.2.3.3. Intracellular thiol assays
Uncoated and selenium-coated PVC substrates were immersed in 1ml TSB for 8
hrs in an incubator (37o, 5% CO2). After the indicated time, the supernatant was collected
and used for culturing S. aureus in 15ml centrifuge tubes. For this, S. aureus (prepared as
described in 5.2.3.1) were cultured in the supernatant at the concentration of 10,000,000
bateria/mL for 8 hrs. After the indicated time, bacteria were collected and analyzed for
intracellular thiol content using a glutathione assay kit (CS 1020, Sigma) following
manufacturer’s instructions. The assay uses a thiol probe (monochlorobimane) which
passes through cell membranes to detect the level of glutathione, the major free thiol in
103
most living cells. If the probe binds to reduced glutathione in cells, it forms a fluorescent
product. The unbound probe does not show fluorescence.
5.2.4. Statistical analysis
Experiments were conducted in triplicate and repeated three times. Data were
collected and the significant differences were assessed with the probability associated
with one-tailed Student’s t tests.
5.3. Results and discussion
In this section, the results from SEM visualization of the selenium coated
polymers as well as EDS analysis are presented. Bacteria (S. aureus) response on the
substrates is also discussed.
5.3.1. SEM and EDS
Importantly, this study elucidated a process to coat all three polymeric substrates
with selenium nanoclusters. At the same concentration of selenium in the synthesis
solution, coating densities on 3 different types of polymeric substrates (silicone, PU and
PVC) were different as shown by SEM images (Figure 5.1). This difference is likely due
to the difference in nucleation sites (e.g. defects) present on the surfaces of the materials
to be coated. Compared to the selenium coatings on titanium and stainless steel in
Chapter 2, the morphology of selenium nanoclusters on the polymeric substrates
appeared similar. The size of the selenium nanoclusters on polymeric substrates was
approximately 150 nm in diameter which was larger than on titanium and stainless steel
substrates (which was ~80nm in diameter). The surface coverage of selenium
(determined from SEM images using an image processing program, imageJ (National
104
Institute of Health)) was approximately 10% on PVC substrates, 4% on PU substrates and
7% on silicone substrates.
Figure 5.1. SEM images of uncoated (left panel) and selenium (Se)-coated substrates (right panel). Arrows indicate selenium nanoclusters. The selenium nanoclusters had sizes ranging from approximately 80nm to 200nm and were uniformly coated on the substrates.
EDS spectra were used to confirm the chemistry of selenium nanoclusters.
Selenium peaks observed in the area coated with selenium nanoclusters confirmed that
the clusters were selenium (Figures 5.2, 5.3, and 5.4).
105
Figure 5.2. EDS spectra (right panel) of an uncoated region (top) and a coated region (bottom) on a selenium-coated silicone substrate. Selenium peaks were detected in the coated region demonstrating that the clusters were selenium. Peaks of gold were from the sputter-coated layer.
106
Figure 5.3. EDS spectra (right panel) of an uncoated region (top) and a coated region (bottom) on a selenium-coated PU substrate. Selenium peaks were detected in the coated region demonstrating that the clusters were selenium. Peaks of gold were from the sputter-coated layer.
107
Figure 5.4. EDS spectra (right panel) of an uncoated region on selenium-coated PVC (top) and a coated region on selenium-coated PVC (bottom). Selenium peaks were detected in the selenium-coated PVC demonstrating that the clusters were selenium. Peaks of gold were from the sputter-coated layer.
5.3.2. S. aureus response on uncoated and selenium-coated polymeric substrates
The uncoated substrates were colonized with S. aureus after 8 hrs of inoculation
while selenium-coated substrates had very little S. aureus colonization (Figures 5.5, 5.6
and 5.7).
108
Figure 5.5. Representative SEM images showing reduced S. aureus colonization on selenium-coated PU compared to uncoated PU after 8 hrs of inoculation. Arrows indicate cells.
Figure 5.6. Representative SEM images showing reduced S. aureus colonization on selenium-coated PVC compared to uncoated PVC. Arrows show bacteria.
Figure 5.7. Representative SEM images showing reduced S. aureus colonization
on selenium-coated compared to uncoated silicone. Arrows indicate cells.
109
Bacteria densities determined by crystal violet assays showed that the bacterial
colonization on the selenium coated substrates significantly decreased compared to
controls (to 80%, 20% and 40% of the control, i.e., uncoated substrates, for PU, PVC and
silicone, respectively) (Figure 5.8).
Figure 5.8. Decreased S. aureus densities on polymeric substrates coated with selenium nanoclusters. Data = mean ± standard error of the mean, N=3; *p<0.05 compared to the uncoated substrate (compared in the same material group), ** p<0.05 compared to silver-coated PVC.
Importantly, compared to the commercial Bard’s silver-coated PVC endotracheal
tube, selenium nanocluster-coated PVC tubes were shown to be approximately twice
more effective in preventing bacteria attachment after 8 hrs (Figures 5.8 and 5.9).
110
Figure 5.9. Representative SEM images of S. aureus colonization on silver-coated and Se-coated PVC. Bacterial colonization on Se-coated PVC was reduced compared to silver-coated PVC after 8 hrs of incubation.
For the first time, selenium nanocluster coatings were demonstrated to inhibit S.
aureus colonization. Especially, selenium nanocluster coatings were approximately twice
more effective than silver coatings in preventing S. aureus colonization after 8 hrs.
Research has shown that elemental selenium could catalyze the oxidation of intracellular
thiols resulting in cellular thiol depletion that causes cell death [40]. It has also been
demonstrated (in Chapter 2) that selenium was released from selenium coatings on
titanium and that the released selenium induced cellular thiol depletion in normal
osteoblasts and cancerous osteoblasts. Therefore, this part of the study also tested if the
selenium nanocluster coatings on polymers inhibited S. aureus growth via the thiol-
depletion mechanism. The results will be presented in the following section.
5.3.3. Thiol assays
To test the hypothesis that selenium nanoclusters caused intracellular thiol
depletion in S. aureus, the bacteria were cultured in the supernatant from either uncoated
PVC or selenium coated PVC substrates. After 8 hrs of incubation, intracellular thiol
levels in bacteria were determined using the Glutathione assay kit (Sigma). S. aureus
111
cultured with fresh bacterial culture media (i.e., tryptic soy broth) were used as controls.
As can be seen from Figure 5.10, the thiol level in the bacteria cultured in the supernatant
from the selenium-coated PVC was significantly lower than that of bacteria cultured in
the supernatant from uncoated PVC.
Figure 5.10. Decreased intracellular thiol level in S. aureus cultured in the supernatant from selenium-coated PVC compared to control. Data = mean ± standard error of the mean; N=3; * p<0.05 compared to the control.
Therefore, it is likely that selenium in the selenium nanocluster coatings released
into the bacteria culture media and this released selenium induced depletion of
intracellular thiol which caused bacteria death.
5.4. Discussion
Attaching selenium to a material to transform an otherwise bacterial colonizing
surface into a material that is anti-bacterial has also been explored by other research
groups. For example, a group at Texas Tech University (Lubbock, TX) developed a
technology that covalent bonds organo-selenium molecules (whose name was unrevealed
for proprietary reasons) to silicone hydrogel contact lenses to transform the lenses into
112
anti-bacterial materials [113]. Specifically, the contact lenses were coated with the
organo-selenium molecules and tested against colonization of Pseudomonas aeruginosa
and S. aureus. The results showed little or no bacterial biofilm formation on the coated
lenses after 4 days of incubating the lenses in bacteria solutions. In contrast, uncoated
lenses showed extensive bacterial biofilm formation. The safety of the organo-selenium
coated lenses was also tested by placing them into the eyes of New Zealand albino rabbits
for up to 2 months. There was no sign of toxicity for the eyes wearing the coated lenses
as well as the eyes wearing uncoated lenses. However, because the method used by this
research group relies on the covalent bonding between selenium compounds and the
material, their method can only be used to impart selenium onto organic materials (such
as polymers). The method used in this thesis, relying on the nucleation of selenium onto
surfaces, creates selenium coatings not only on polymers but also on metals (as shown in
Chapter 2). Therefore, the process employed in this thesis has the ability to create
selenium coatings on a broad spectrum of biomaterials.
5.5. Conclusions
Selenium nanoclusters were coated on polymeric substrates using the same
synthesis methods utilized to coat selenium nanoclusters on metallic substrates in Chapter
2. Similar morphology and chemistry of the coatings (compared to the coating on
titanium and stainless steel substrates) was observed. Compared to the uncoated
substrates, substrates coated with selenium nanoclusters had significantly lower
attachment of S. aureus after 8 hrs of incubation. Importantly, compared to commercially
available silver coated endotracheal tubes, the selenium coated tubes were approximately
twice as more effective at reducing S. aureus attachment. The decreased bacterial
113
attachment on selenium coated substrates was likely due to the released selenium that
depleted thiol levels in the cells causing cell death. This result, for the first time,
demonstrates the potential of using selenium nanoclusters as a coating material for
numerous anti-bacterial applications.
114
CHAPTER 6. SELENIUM NANOPARTICLES FOR ANTI-BACTERIAL APPLICATIONS
6.1. Introduction
In the previous chapters, it has been demonstrated that selenium nanocluster
coatings on traditional implant materials promoted healthy bone growth while inhibiting
bone cancer growth as well as inhibiting bacterial growth in vitro. However, how
selenium nanoparticles act as opposed to adherent nanoclusters on implants on bacteria
(such as S. aureus, S. epidermidis) remains largely unknown.
Nanotechnology has enable researchers to synthesize nano-size particles (that is,
particles that have sizes less than 100nm in at least 1 dimension) for a wide range of
applications. Nanoparticles have increased surface areas, therefore, have potentially
increased interactions with biological targets (such as bacteria) compared to
conventional, micro-size particles. As a result, nano-antibacterial particles will likely
exert a stronger effect on bacteria than their microparticle-counterparts.
Therefore, the objective of this part of the study was to examine the functions of
S. aureus in the presence of selenium nanoparticles. In doing so, this study hopes to
reveal a new type of anti-bacterial nanoparticles for decreasing S. aureus adhesion and
colonization.
This chapter is organized into a materials and methods section where the methods
to synthesize and characterize selenium nanoparticles are described, a results section and
a summary and conclusion section.
115
6.2. Materials and methods
6.2.1. Material synthesis
Selenium nanoparticles were synthesized by reduction of sodium selenite (Alfa
Aesar) by glutathione (reduced form, GSH, TCI America) stabilized by bovine serum
albumin (BSA, Sigma Aldrich). Specifically, 3ml of 25mM Na2SeO3, 3ml of 100mM
GSH and 0.15g BSA were added to 9ml double distilled water in a sterile cabinet. BSA
was used as a surfactant to keep the selenium nanoparticles well dispersed. All solutions
were made in a sterile environment by using a sterile cabinet and double distilled water.
After the reactant solution was mixed, 1M NaOH was added to bring the pH of the
solution to the alkaline regime. Selenium nanparticles were formed immediately
following the addition of NaOH as visualized by a color change of the reactant solution
from clear white to clear red. Selenium nanoparticles were then collected by centrifuging
at 13000 rounds per minute and were re-suspended in deionized water five times before
using in bacteria experiments.
6.2.2. Material characterization
The size and morphology of selenium nanoparticles were investigated using a
transmission electron microscope (TEM). For this, the dispersions of nanoparticles in
deionized water were allowed to slowly dry on formvar-coated copper grids. All imaging
was carried out using a Philips 410 TEM (New York, NY, USA) at 80kV.
The size distribution of selenium nanoparticles was investigated by dynamic light
scattering (DLS) technique using a Zetasizer-Nano-S90 (Malvern Instrument).
Chemistry of selenium nanoparticles was characterized using X-ray photoelectron
116
spectroscopy (XPS, Perkin-Elmer PHI. 5500 Multi-Technique System). For this, a drop
of selenium nanoparticle solution was placed on a glass coverslip, air-dried and analyzed
under XPS.
6.2.3. Bacteria assays
6.2.3.1. Bacteria preparation
Bacteria solutions were prepared as described in section 5.2.3.1 in Chapter 5.
Briefly, a bacterial cell line of biofilm-producing S. aureus was obtained in freeze-dried
form from the American Type Culture Collection (catalog number 25923). The cells were
propagated in 30mg/mL tryptic soy broth (TSB, MP Biomedicals). A sterile loop was
used to withdraw some of the bacteria from the frozen vial, streaked onto tryptic soy agar
plates (30g Tryptone [MP Biomedicals] and 15g agar [Sigma] per litter of distilled water)
and incubated for 12 hrs. Another sterile loop was used to select one colony from the agar
plates to place into 3ml trypic soy broth (TSB, 30g/L) and was incubated for 16 hrs.
Bacteria concentration was assessed by measuring optical density of bacterial solution at
562nm and using a standard curve correlating optical densities and bacterial
concentrations. The bacterial solution was prepared at a concentration of 50,000 cells/mL
for all experiments.
6.2.3.2. Effects of selenium nanoparticles on planktonic S. aureus
Three concentrations of selenium (Se) nanoparticles were tested against S. aureus
growth: 3.1µg Se/mL, 6.2 µg Se/mL and 12.5 µg Se/mL with TSB (0 µg Se/mL) as the
control. Selenium nanoparticles were mixed with bacterial solutions and cultured for 4,
12 and 24 hrs in an incubator (37°C, humidified, 5% CO2) shaking at 250 rounds per
117
minute. Blank solutions were prepared by adding selenium nanoparticles into TSB
without bacteria at the above concentrations of particles. Blank solutions of TSB without
bacteria and without selenium nanoparticles were used as the blank for the controls.
At the end of the prescribed time periods, optical densities (or the degree of
“cloudiness”) of the bacteria solutions (which is proportional to bacteria densities) were
measured. For this, 200µL of the bacteria solution, control, or blank were added to wells
of a 96 well plate and optical densities were measured at 562nm. Measured optical
densities of bacterial solutions were subtracted by that of the corresponding blanks to
remove the optical density resulting from the nanoparticles alone.
6.2.3.3. Intracellular thiol assays
S. aureus (at a density of ~50,000 cells/mL) were cultured with selenium
nanoparticles (at a concentration of 3.1 µg/mL in tryptic soy broth) for 4 hrs under
standard conditions (37oC, 5%CO2, 95% humidified air) shaking at 250 rounds per
minute. After the indicated time, cells were collected and analyzed for thiol content using
a glutathione assay kit (CS1020, Sigma Aldrich) following manufacturer’s instruction. S.
aureus cultured in normal cell culture media without selenium nanoparticles were used as
a control. The glutathione assay kit uses a thiol probe (monochlorobimane) which passes
through cell membranes to detect the level of glutathione, the major free thiol in most
living cells. Glutathione is involved in many biological processes such as removal of
hydroperoxides and detoxification of xenobiotics. Intracellular reduced glutathione level
is a sensitive indicator of the overall health of a cell. If the probe binds to reduced
glutathione in cells, it forms a fluorescent product whose fluorescence intensity can be
measured. The unbound probe does not show fluorescence.
118
6.2.4. Superoxide anion assays
Some reports in the literature show that selenium (in the presence of thiols) can
generate superoxide anion which is a short-lived (half-life time ~ 0.05 seconds) free
radical that kills bacteria [35, 114, 115]. Therefore, in this chapter, superoxide anion
assays were conducted to test if selenium nanoparticles in this thesis generated
superoxide anion. The assays were conducted using a superoxide anion kit (LumiMax
Superoxide Anion Detection kit, Strategene, catalog number 204525). Reagents for
superoxide anion assays were prepared following manufacturer’s instruction with some
modifications. Briefly, three types of reagents were prepared: test samples which had
luminol, enhancer, glutathione (GSH) and selenium nanoparticles (referred to as nSe),
control 1 which did not have nSe and control 2 which did not have the enhancer solution
(Table 6.1).
Table 6.1. Superoxide anion assay reagent mixtures
Assay
medium (µL)
Luminol solution
(µL)
Enhancer (µL)
12.5 µg/mL nSe solution
(µL)
1µg/mL GSH (µL)
Test sample 89 5 5 5 5 Control 1
(without selenium) 94 5 5 0 5
Control 2 (without enhancer)
94 5 0 5 5
The mixture was mixed and immediately measured for photon emission using an
illuminometer (Lumat LB9510, Berthold Technologies). Due to the short life-time of
superoxide anions (~0.05 seconds), photon emission was read every 10 seconds for 30
119
seconds immediately after mixing. Photon emission was then reported as the total relative
light unit emitted during these 30 seconds.
6.2.5. Statistical analysis
All experiments were conducted in triplicate and repeated three times. Data were
collected and the significant differences were assessed with the probability associated
with one- tailed Student’s t tests.
6.3. Results
6.3.1. TEM, DLS and XPS results on size distribution and oxidation state of selenium
nanoparticles
TEM images of selenium nanoparticles show that the particles were spherical and
about 40-100 nm in diameter (Figure 6.1). Further investigation of size distribution of the
selenium nanoparticles by DLS revealed that most of the particles had diameters of
100nm (Figure 6.2). The sizes observed by DLS were larger than those determined by
TEM images because water molecules bound to the surface of the nanoparticles creating
a “hairy layer” that made the particles appear larger [116, 117].
120
Figure 6.1. TEM image of selenium nanoparticles stabilized in BSA and dispersed in water. Selenium nanoparticles had average sizes of approximately 100 nm.
121
Figure 6.2. Size-distribution profile of selenium nanoparticles in solution as measured by the dynamic light scattering technique. Particle sizes centered around 100 nm.
The size-distribution profile demonstrated that the synthesis method yielded
selenium nanoparticles of a narrow size range and stable in water. Nanometer scale sizes
of the synthesized nanoparticles rendered the nanoparticles large surface areas that are
important towards increasing the particles’ interaction with bacteria.
The peak at 55.2 eV (Figure 6.3) demonstrated that the selenium nanoparticles
had a zero oxidation state (i.e., they were elemental selenium).
Figure 6.3. XPS profile of selenium nanoparticles. Peak at 55.2 eV confirmed a zero oxidation state of selenium.
122
6.3.2. Effects of selenium nanoparticles on S. aureus in solution
When the selenium nanoparticles were mixed with the bacterial solution, the
growth of bacteria was significantly inhibited after 4 hrs (compared to the controls -
0µg/mL). The inhibitory effects continued after 24 hrs (Figure 6.4). A dose-dependent
inhibition of bacterial growth was observed at 4 hrs and 12 hrs.
Figure 6.4. Inhibited growth of S. aureus in the presence of selenium nanoparticles at all three selenium nanoparticles concentrations: 12.5 µg/mL, 6.2 µg/mL and 3.1 µg/mL at 4 hrs, 12 hrs and 24 hrs. Data = mean ± standard error of the mean, N=3. *p<0.05 compared to bacteria treated with 6.2 µg Se/mL, 3.1 µg Se/mL and control (0 µg Se/mL) (compared at same time period); ** p<0.05 compared to bacteria treated with 3.1 µg Se/mL and control (0 µg Se/mL) (compared at same time period); ***p<0.05 compared to control (compared at same time period); #p<0.05 compared to control (compared at same time period). Dose-dependent inhibition was observed at 4hrs and 12hrs.
6.3.3. Intracellular thiol assays
As it was demonstrated in Chapter 5 that selenium can be released from selenium
coatings on polymers into the bacteria culture media causing depletion of intracellular
123
thiol in S. aureus. To test if selenium nanoparticles inhibited S. aureus in this study via a
similar mechanism, thiol assays were conducted to determine thiol content in S. aureus
cultured with selenium nanoparticles. S. aureus cultured in normal bacteria culture media
(tryptic soy broth (TBS)) were used as the control. The results (Figure 6.5) showed that
the intracellular thiol levels in S. aureus cultured with selenium nanoparticles was
significantly decreased (to approximately 90% of the control) compared to that of the
control. This means that S. aureus treated with selenium nanoparticles had depleted thiol
levels which is known to cause cell death.
Figure 6.5. Decreased intracellular thiol levels in S. aureus cultured with selenium nanoparticles compared to the control (i.e., S. aureus cultured in TSB without selenium nanoparticles). Bacteria were cultured at a density of 50,000 cells/mL for 4 hrs in either TSB or TSB added with selenium nanoparticle (at a concentration of 3.1µg/mL). Data = mean ± standard error of the mean; N=3; * p< 0.05.
124
Similar effects (i.e., selenium induced intracellular thiol depletion) were observed
in selenium nanocluster coatings on metals (Chapter 2) and polymers (Chapter 5). These
results indicated that intracellular thiol depletion is likely the mechanisms of anti-cancer
and anti-bacterial properties of nanometer structured selenium in this thesis.
6.3.4. Superoxide assays
Some reports in the literature demonstrate that selenium, in the presence of thiols
(such as glutathione), can generate superoxide anion ( 2O ) that kills bacteria [114, 115].
To test if the selenium nanoparticles in this study generated superoxide anions which
killed S. aureus, superoxide anion assays were conducted. The assay used luminol which
reacts with superoxide anions to generate photons of light which can be measured using
an illuminometer. The assays also used an enhancer solution which amplifies the
luminescence signal to increase the sensitivity of the measurement. For this, selenium
nanoparticles (referred to as nSe) were mixed with reduced glutathione (GSH, Sigma),
luminol and enhancer solutions and the photon emission from this mixture was measured.
Two controls (luminol + enhancer and luminol + GSH+ nSe) were used for comparisons.
Results showed that there was no significant difference among the tested sample (i.e.,
nSe) and controls (i.e., luminol + enhancer or luminol + GSH + nSe) (Figure 6.6).
Therefore, it was concluded that elemental selenium nanoparticles in this study did not
generate superoxide anions.
125
Figure 6.6. Superoxide anion assays for selenium particles (nSe). No significant difference was found among the test tubes. Selenium nanoparticles were mixed with GSH and luminol was used as the probes for superoxide anions. Superoxide anions oxidize luminol in a reaction that produces photons. An enhancer was used to amplify the chemiluminescence. Data = mean ± standard error of the mean, N=3.
6.4. Summary and conclusions
In this chapter, it has been demonstrated that selenium nanoparticles were
synthesized with a narrow size distribution centered around the value of 100nm. More
importantly, selenium nanoparticles were shown to significantly inhibit the growth of S.
aureus in solution. It was further shown that selenium nanoparticles induced thiol
depletion that caused death of S. aureus. Future studies should evaluate the anti-bacterial
properties of selenium nanoparticles against other bacteria (such as Pseudomonas
aeruginosa and S. epidermidis) to fully explore the potential of using selenium
nanoparticles for anti-infection applications. Toxicity studies should also be conducted to
determine safety concentrations of selenium nanoparticles to use in humans. Specifically,
in vitro studies should be implemented to evaluate response of mammalian cells (for
example fibroblasts) to selenium nanoparticles at different concentrations. Further studies
126
should then be completed to evaluate the anti-bacterial properties as well as safety
concentration of selenium nanoparticles in animals and humans.
127
CHAPTER 7. CONCLUSIONS
This thesis presented a comprehensive study of the potential of nanometer
structured selenium for various biomedical applications. Specifically, in this study,
selenium nanoclusters were coated on orthopedic implant metals (such as titanium and
stainless steel). The coatings were fully characterized using techniques such as electron
scanning microscopy, X-ray photoelectron spectroscopy, atomic force microscopy, and
contact angle measurements. Most importantly, compared to uncoated metals, the
coatings on orthopedic metals promoted healthy bone cell functions. This result by itself
suggests much promise for selenium coated metals for any orthopedic application.
Moreover, cancerous bone cell functions were inhibited on selenium nanocluster coatings
compared to uncoated metals. This result suggests that selenium coated implants should
be tested for numerous cancer applications, in particular, bone cancer applications. This
study also examined the possible mechanisms of the promoted healthy bone-cell
properties of selenium nanocluster coatings on metals. It was shown that the increased
nanometer roughness of the selenium coated surfaces promoted the adsorption of
fibronectin (which is one of the most important cell adhesive proteins that regulate
interactions between osteoblasts and an implant surface). The experimental results of
protein adsorption were complimented by results from computer simulation. The anti-
cancer properties of selenium nanocluster coatings were attributed in this thesis to
128
selenium released from the coatings into cell culture media to induce cellular thiol
depletion which caused cell death in cancerous osteosarcoma cells.
To further demonstrate the versatility of nanometer scale structured selenium in
biomedical applications, the selenium nanoclusters were also coated on common catheter
polymers (such as polyvinyl chloride, polyurethane and silicone). The selenium coatings
were shown to be highly effective against the colonization of S. aureus, an important
bacterium that infects all implants, especially catheters. This result opens the possibility
of using nanostructured selenium for any anti-bacterial application. This thesis also
demonstrated that the selenium released from the coatings induced cellular thiol depletion
which caused cell death in S. aureus.
The study also investigated the potential of using selenium nanoparticles in
various anti-bacterial medical applications. Selenium nanoparticles were synthesized and
characterized by techniques such as transmission electron microscopy, X-ray
photoelectron microscopy and dynamic light scattering. Selenium nanoparticles were
demonstrated to be highly effective at inhibiting the growth of S. aureus. Intracellular
thiol levels of the bacteria treated with selenium nanoparticles also decreased compared
to those of the bacteria treated with normal cell culture media. This result opens the
possibility of using nanoparticulate selenium in numerous infection applications (such as
implantation, catheterization, wound repair, etc.)
In conclusion, this thesis demonstrated a number of promising directions for using
nanometer scale structured selenium in diverse biomedical applications (including bone-
promoting, cancer-inhibiting and bacteria-inhibiting applications). The anti-cancer
properties of nanometer structured selenium should be further tested on other cancerous
129
cell types (such as breast cancer, lung cancer, etc.) to fully evaluate its applications in
cancer treatment. The anti-bacterial properties of nanometer structured selenium should
also be tested on other types of bacteria (such as P. aeruginosa, S. epidermidis, etc.) to
fully investigate its anti-infection properties and its anti-infection applications. In vivo
studies should be conducted to further investigate the potential of this biomaterial.
130
References:
1. ACS. Cancer Statistics. 2008 [cited 2009 2/12]; Available from: http://www.cancer.org/docroot/STT/STT_0.asp
2. Miller D WT. Anticancer orthopedic implants. In: Nalwa H WTJ, editor. Cancer nanotechnology. California: American Scientific, 2007. p. 307-316.
3. Orthopedics today; 29:6 2009 [cited 2009 2/7/2010]; Available from: http://www.orthosupersite.com/view.asp?rID=44153
4. Krishnasami Z, Carlton D, Bimbo L, Taylor ME, Balkovetz DF, Barker J, et al. Management of hemodialysis catheter-related bacteremia with an adjunctive antibiotic lock solution. Kidney Int 2002 Mar;61(3):1136-1142.
5. Trerotola SO. Hemodialysis catheter placement and management. Radiology 2000 Jun;215(3):651-658.
6. Sanders J, Pithie A, Ganly P, Surgenor L, Wilson R, Merriman E, et al. A prospective double-blind randomized trial comparing intraluminal ethanol with heparinized saline for the prevention of catheter-associated bloodstream infection in immunosuppressed haematology patients. J Antimicrob Chemother 2008 Oct;62(4):809-815.
7. Johnson JR, Kuskowski MA, Wilt TJ. Systematic review: antimicrobial urinary catheters to prevent catheter-associated urinary tract infection in hospitalized patients. Ann Intern Med 2006 Jan 17;144(2):116-126.
8. Maki DG, Tambyah PA. Engineering out the risk for infection with urinary catheters. Emerg Infect Dis 2001 Mar-Apr;7(2):342-347.
9. Roe D, Karandikar B, Bonn-Savage N, Gibbins B, Roullet JB. Antimicrobial surface functionalization of plastic catheters by silver nanoparticles. J Antimicrob Chemother 2008 Apr;61(4):869-876.
10. Schierholz JM, Beuth J. Implant infections: a haven for opportunistic bacteria. Journal of Hospital Infection 2001;49(2):87-93.
11. Donlan RM. Biofilms and device-associated infections. Emerg Infect Dis 2001 Mar-Apr;7(2):277-281.
12. Davies D. Understanding biofilm resistance to antibacterial agents. Nat Rev Drug Discov 2003;2(2):114-122.
13. Chastre J, Fagon J-Y. Ventilator-associated Pneumonia. Am J Respir Crit Care Med 2002 April 1, 2002;165(7):867-903.
14. Sanderson PJ. Infection in orthopaedic implants. Journal of Hospital Infection 1991;18(Supplement 1):367-375.
15. Styblo M, Thomas DJ. Selenium modifies the metabolism and toxicity of arsenic in primary rat hepatocytes. Toxicol Appl Pharmacol 2001 Apr 1;172(1):52-61.
16. McDowell L. Minerals in human nutrition. Amsterdam:: Elsevier, 2003. 17. Combs GFJ, Combs S. The role of selenium in nutrition. Orlando: Academic press, 1986. 18. Clark LC, Cantor KP, Allaway WH. Selenium in forage crops and cancer mortality in
U.S. counties. Arch Environ Health 1991 Jan-Feb;46(1):37-42. 19. Combs GF, Jr., Gray WP. Chemopreventive agents: selenium. Pharmacol Ther 1998
Sep;79(3):179-192. 20. Navarro-Alarcon M, Lopez-Martinez MC. Essentiality of selenium in the human body:
relationship with different diseases. Sci Total Environ 2000 Apr 17;249(1-3):347-371.
131
21. Clark LC, Combs GF, Jr., Turnbull BW, Slate EH, Chalker DK, Chow J, et al. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group. JAMA 1996 December 25, 1996;276(24):1957-1963.
22. Wei W-Q, Abnet CC, Qiao Y-L, Dawsey SM, Dong Z-W, Sun X-D, et al. Prospective study of serum selenium concentrations and esophageal and gastric cardia cancer, heart disease, stroke, and total death. Am J Clin Nutr 2004 January 1, 2004;79(1):80-85.
23. Clark LC, Dalkin B, Krongrad A, Combs GF, Jr., Turnbull BW, Slate EH, et al. Decreased incidence of prostate cancer with selenium supplementation: results of a double-blind cancer prevention trial. Br J Urol 1998 May;81(5):730-734.
24. Rayman MP. Selenium in cancer prevention: a review of the evidence and mechanism of action. Proceedings of the Nutrition Society 2005;64(04):527-542.
25. Lu J, Jiang C, Kaeck M, Ganther H, Ip C, Thompson H. Cellular and metabolic effects of triphenylselenonium chloride in a mammary cell culture model. Carcinogenesis 1995 Mar;16(3):513-517.
26. Ronai Z, Tillotson JK, Traganos F, Darzynkiewicz Z, Conaway CC, Upadhyaya P, et al. Effects of organic and inorganic selenium compounds on rat mammary tumor cells. Int J Cancer 1995 Nov 3;63(3):428-434.
27. Redman C, Scott JA, Baines AT, Basye JL, Clark LC, Calley C, et al. Inhibitory effect of selenomethionine on the growth of three selected human tumor cell lines. Cancer Lett 1998 Mar 13;125(1-2):103-110.
28. Kaeck M, Lu J, Strange R, Ip C, Ganther HE, Thompson HJ. Differential induction of growth arrest inducible genes by selenium compounds. Biochem Pharmacol 1997 Apr 4;53(7):921-926.
29. Lu J, Pei H, Ip C, Lisk DJ, Ganther H, Thompson HJ. Effect on an aqueous extract of selenium-enriched garlic on in vitro markers and in vivo efficacy in cancer prevention. Carcinogenesis 1996 Sep;17(9):1903-1907.
30. Apostolou S, Klein JO, Mitsuuchi Y, Shetler JN, Poulikakos PI, Jhanwar SC, et al. Growth inhibition and induction of apoptosis in mesothelioma cells by selenium and dependence on selenoprotein SEP15 genotype. Oncogene 2004 Jun 24;23(29):5032-5040.
31. Menter DG, Sabichi AL, Lippman SM. Selenium effects on prostate cell growth. Cancer Epidemiol Biomarkers Prev 2000 Nov;9(11):1171-1182.
32. Combs GF, Jr. Impact of selenium and cancer-prevention findings on the nutrition-health paradigm. Nutr Cancer 2001;40(1):6-11.
33. Cheng J, Weiqin J, Clement I, Howard G, Junxuan L. Selenium-induced inhibition of angiogenesis in mammary cancer at chemopreventive levels of intake. Molecular Carcinogenesis 1999;26(4):213-225.
34. Jiang C, Wang Z, Ganther H, Lu J. Caspases as Key Executors of Methyl Selenium-induced Apoptosis (Anoikis) of DU-145 Prostate Cancer Cells. Cancer Res 2001 April 1, 2001;61(7):3062-3070.
35. Spallholz JE. On the nature of selenium toxicity and carcinostatic activity. Free Radic Biol Med 1994 Jul;17(1):45-64.
36. Yang J, Huang K, Qin S, Wu X, Zhao Z, Chen F. Antibacterial action of selenium-enriched probiotics against pathogenic Escherichia coli. Dig Dis Sci 2009 Feb;54(2):246-254.
132
37. Ratushnaya EV, Kirova YI, Suchkov MA, Drevko BI, Borodulin VB. Synthesis and Antibacterial Activity of Organoselenium Compounds. Pharmaceutical Chemistry Journal 2002;36(12):652-653.
38. Pietka-Ottlik M, Wojtowicz-Mlochowska H, Kolodziejczyk K, Piasecki E, Mlochowski J. New organoselenium compounds active against pathogenic bacteria, fungi and viruses. Chem Pharm Bull (Tokyo) 2008 Oct;56(10):1423-1427.
39. Kucukbay H, Durmaz R, Orhan E, Gunal S. Synthesis, antibacterial and antifungal activities of electron-rich olefins derived benzimidazole compounds. Farmaco 2003 Jun;58(6):431-437.
40. Sieber F, Günther WHH, Daziano J-p, Krieg-kowald M, Bousbaa J, Bula RJ, inventors. Method of making, and the use of cytotoxic agents containing elemental selenium, 2007.
41. Crompton K. Joint replacements become more routine here. J Business-Spokane 2004;19:10.
42. Buser D, Nydegger T, Oxland T, Cochran DL, Schenk RK, Hirt HP, et al. Interface shear strength of titanium implants with a sandblasted and acid-etched surface: a biomechanical study in the maxilla of miniature pigs. J Biomed Mater Res 1999 May;45(2):75-83.
43. Webster T. Nanophase ceramics: the future of orthopedic and dental implant material. In: Ying J, editor. Nanostructured Materials. New York: Academy Press, 2001. p. 125-166.
44. Kaplan FS HW, Keaveny TM., Boskey A, Einhorn TA, Iannotti JP. Biomaterials. In: SP S, editor. Orthopedic Basic Science. Columbus: American Academy of Orthopedic Surgeons, 1994. p. 460-478.
45. Kaplan FS LW, Keaveny TM, Boskey A, Einhorn TA, Iannotti JP. . Forms and functions of bone. In: SP S, editor. Orthopedic basic science. Columbus: American Academy of Orthopedic Surgeons, 1994. p. 127-185.
46. Park JB, RS L. Biomaterials-An introduction. New York: Plenum Press, 1992. 47. Webster TJ, Ergun C, Doremus RH, Siegel RW, Bizios R. Enhanced functions of
osteoblasts on nanophase ceramics. Biomaterials 2000 Sep;21(17):1803-1810. 48. Webster TJ, Siegel RW, Bizios R. Osteoblast adhesion on nanophase ceramics.
Biomaterials 1999 Jul;20(13):1221-1227. 49. Webster TJ, Ejiofor JU. Increased osteoblast adhesion on nanophase metals: Ti,
Ti6Al4V, and CoCrMo. Biomaterials 2004 Aug;25(19):4731-4739. 50. Webster TJ, Ergun C, Doremus RH, Siegel RW, Bizios R. Specific proteins mediate
enhanced osteoblast adhesion on nanophase ceramics. J Biomed Mater Res 2000 Sep 5;51(3):475-483.
51. Perla V, Webster TJ. Better osteoblast adhesion on nanoparticulate selenium- A promising orthopedic implant material. J Biomed Mater Res A 2005 Nov 1;75(2):356-364.
52. Whanger P, Vendeland S, Park YC, Xia Y. Metabolism of subtoxic levels of selenium in animals and humans. Ann Clin Lab Sci 1996 Mar-Apr;26(2):99-113.
53. Zhang S-Y, Zhang J, Wang H-Y, Chen H-Y. Synthesis of selenium nanoparticles in the presence of polysaccharides. Materials Letters 2004;58(21):2590-2594.
54. Kopeikin V, Valueva S, Kipper A, Borovikova L, Filippov A. Synthesis of selenium nanoparticles in aqueous solutions of poly(vinylpyrrolidone) and morphological characteristics of the related nanocomposites. Polymer science Series A 2003(45(4)):374-379.
133
55. Lin Z-H, Chris Wang CR. Evidence on the size-dependent absorption spectral evolution of selenium nanoparticles. Materials Chemistry and Physics 2005;92(2-3):591-594.
56. Wang H, Wei W, Zhang SY, Shen YX, Yue L, Wang NP, et al. Melatonin-selenium nanoparticles inhibit oxidative stress and protect against hepatic injury induced by Bacillus Calmette-Guerin/lipopolysaccharide in mice. J Pineal Res 2005 Sep;39(2):156-163.
57. Thapa A, Miller DC, Webster TJ, Haberstroh KM. Nano-structured polymers enhance bladder smooth muscle cell function. Biomaterials 2003 Aug;24(17):2915-2926.
58. ATCC. 2010 [cited 2010 2/27]; Available from: http://www.atcc.org/ATCCAdvancedCatalogSearch/ProductDetails/tabid/452/Default.aspx?ATCCNum=CRL-11372&Template=cellBiology
59. Khanna C, Prehn J, Yeung C, Caylor J, Tsokos M, Helman L. An orthotopic model of murine osteosarcoma with clonally related variants differing in pulmonary metastatic potential. Clinical and Experimental Metastasis 2000;18(3):261-271.
60. Moulder JF, Stickle WF, Sobol PE, Bomben KD. Handbook of X-ray photoelectron spectroscopy: A reference book of standard spectra for identification and interpretation of xps data. Minnesota: Perkin-Elmer, 1992.
61. Zhang JS, Gao XY, Zhang LD, Bao YP. Biological effects of a nano red elemental selenium. Biofactors 2001;15(1):27-38.
62. Yang CF, Shen HM, Ong CN. Ebselen induces apoptosis in HepG2 cells through rapid depletion of intracellular thiols. Archives of Biochemistry and Biophysics 2000;374(2):142-152.
63. Iwata S, Hori T, Sato N, Hirota K, Sasada T, Mitsui A, et al. Adult T cell leukemia (ATL)-derived factor/human thioredoxin prevents apoptosis of lymphoid cells induced by L-cystine and glutathione depletion: possible involvement of thiol-mediated redox regulation in apoptosis caused by pro-oxidant state. The Journal of Immunology 1997;158(7):3108.
64. Borutaite V, Brown GC. Nitric oxide induces apoptosis via hydrogen peroxide, but necrosis via energy and thiol depletion. Free Radical Biology and Medicine 2003;35(11):1457-1468.
65. Aoshiba K, Yasui S, Nishimura K, Nagai A. Thiol depletion induces apoptosis in cultured lung fibroblasts. American journal of respiratory cell and molecular biology 1999;21(1):54.
66. Oberley TD, Oberley LW. Antioxidant enzyme levels in cancer. Histology and histopathology 1997;12(2):525.
67. Hileman EO, Liu J, Albitar M, Keating MJ, Huang P. Intrinsic oxidative stress in cancer cells: a biochemical basis for therapeutic selectivity. Cancer chemotherapy and pharmacology 2004;53(3):209-219.
68. Coursin DB, Cihla HP, Sempf J, Oberley TD, Oberley LW. An immunohistochemical analysis of antioxidant and glutathione S-transferase enzyme levels in normal and neoplastic human lung. Histology and histopathology 1996;11(4):851.
69. Anderson M, Meister A, inventors. Method for depletion of glutathione, 1995. 70. Zhdanov VP, Kasemo B. Van der Waals interaction during protein adsorption on a solid
covered by a thin film. Langmuir 2001;17(18):5407-5409.
134
71. Besenbacher F, Chevallier J, Hovgaard MB, Rechendorff K, Foss M. Fibronectin Adsorption on Tantalum: The Influence of Nanoroughness. Journal of Physical Chemistry B 2008;112(28):8241-8249.
72. Magnusson MK, Mosher DF. Fibronectin: structure, assembly, and cardiovascular implications. Arteriosclerosis, thrombosis, and vascular biology 1998;18(9):1363.
73. Moursi AM, Damsky CH, Lull J, Zimmerman D, Doty SB, Aota S, et al. Fibronectin regulates calvarial osteoblast differentiation. Journal of Cell Science 1996;109(6):1369.
74. Puleo DA, Bizios R. Mechanisms of fibronectin-mediated attachment of osteoblasts to substrates in vitro. Bone and mineral 1992;18(3):215-226.
75. Lamponi S, Forbicioni M, Barbucci R. The role of fibronectin in cell adhesion to spiral patterned TiO2 nanoparticles. J Appl Biomater Biomech 2009 May-Aug;7(2):104-110.
76. Verrier S, Pallu S, Bareille R, Jonczyk A, Meyer J, Dard M, et al. Function of linear and cyclic RGD-containing peptides in osteoprogenitor cells adhesion process. Biomaterials 2002;23(2):585-596.
77. Ponik SM, Pavalko FM. Formation of focal adhesions on fibronectin promotes fluid shear stress induction of COX-2 and PGE2 release in MC3T3-E1 osteoblasts. Journal of Applied Physiology 2004;97(1):135.
78. Schneider G, Burridge K. Formation of focal adhesions by osteoblasts adhering to different substrata. Experimental cell research 1994;214(1):264-269.
79. Klee D, Ademovic Z, Bosserhoff A, Hoecker H, Maziolis G, Erli HJ. Surface modification of poly (vinylidenefluoride) to improve the osteoblast adhesion. Biomaterials 2003;24(21):3663-3670.
80. Mosher DF. Plasma fibronectin concentration: a risk factor for arterial thrombosis? Arteriosclerosis, thrombosis, and vascular biology 2006;26(6):1193.
81. Altankov G, Groth T. Reorganization of substratum-bound fibronectin on hydrophilic and hydrophobic materials is related to biocompatibility. Journal of Materials Science: Materials in Medicine 1994;5(9):732-737.
82. Nugiel DJ, Wood DJ, Sung KLP. Quantification of adhesiveness of osteoblasts to titanium surfaces in vitro by the micropipette aspiration technique. Tissue Engineering 1996;2(2):127-140.
83. Feder J. Random sequential adsorption. Journal of Theoretical Biology 1980;87(2):237-254.
84. Pomeau Y. Some asymptotic estimates in the random parking problem. Journal of Physics A: Mathematical and General 1980;13:L193-L196.
85. Schaaf P, Voegel J-C, Senger B. From Random Sequential Adsorption to Ballistic Deposition: A General View of Irreversible Deposition Processes. The Journal of Physical Chemistry B 2000;104(10):2204-2214.
86. Talbot J, Tarjus G, Van Tassel PR, Viot P. From car parking to protein adsorption: an overview of sequential adsorption processes. Colloids and Surfaces A: Physicochemical and Engineering Aspects 2000;165(1-3):287-324.
87. Elaissari A, Haouam A, Huguenard C, Pefferkorn E. Kinetic factors in polymer adsorption at solid/liquid interfaces. Methods of study of the adsorption mechanism. Journal of Colloid and Interface Science 1992;149(1):68-83.
88. Ramsden JJ. Concentration scaling of protein deposition kinetics. Physical Review Letters 1993;71(2):295.
135
89. Van Tassel PR, Guemouri L, Ramsden JJ, Tarjus G, Viot P, Talbot J. A particle-level model of irreversible protein adsorption with a postadsorption transition. Journal of Colloid and Interface Science 1998;207(2):317-323.
90. Elaissari A, Pefferkorn E. Polyelectrolyte adsorption at solid/liquid interfaces: A simple model for the structural relaxation and excluded area effects. Journal of Colloid and Interface Science 1991;143(1):85-91.
91. Patrikar RM. Modeling and simulation of surface roughness. Applied Surface Science 2004;228(1-4):213-220.
92. Hedsten K, Fonollosa J, Enoksson P, Modh P, Bengtsson J, Sutherland DS, et al. Optical Label-Free Nanoplasmonic Biosensing Using a Vertical-Cavity Surface-Emitting Laser and Charge-Coupled Device. Analytical Chemistry:442-651.
93. Dimilla PA, Albelda SM, Quinn JA. Adsorption and elution of extracellular matrix proteins on non-tissue culture polystyrene petri dishes. Journal of Colloid and Interface Science 1992;153(1):212-225.
94. Guemouri L, Ogier J, Zekhnini Z, Ramsden JJ. The architecture of fibronectin at surfaces. The Journal of Chemical Physics 2000;113:8183.
95. Hemmersam AG, Rechendorff K, Foss M, Sutherland DS, Besenbacher F. Fibronectin adsorption on gold, Ti-, and Ta-oxide investigated by QCM-D and RSA modelling. Journal of Colloid and Interface Science 2008;320(1):110-116.
96. Erickson HP, Carrell N, McDonagh J. Fibronectin molecule visualized in electron microscopy: a long, thin, flexible strand. J Cell Biol 1981 December 1, 1981;91(3):673-678.
97. Rocco M, Infusini E, Daga MG, Gogioso L, Cuniberti C. Models of fibronectin. The EMBO journal 1987;6(8):2343.
98. Victor EK, Marina AG, Marina VB, Vladimir NS, Valdimir VF, Olga MZ, et al. A Study of the Structure of Fibronectin. European Journal of Biochemistry 1981;119(3):619-624.
99. Tooney NM, Mosesson MW, Amrani DL, Hainfeld JF, Wall JS. Solution and surface effects on plasma fibronectin structure. J Cell Biol 1983 December 1, 1983;97(6):1686-1692.
100. Khang D. adhesion protein adsorption and bone cell growth on carbon nanotube composite materials: Perdue University; 2003.
101. Yoshimitsu Z, Nakajima A, Watanabe T, Hashimoto K. Effects of Surface Structure on the Hydrophobicity and Sliding Behavior of Water Droplets. Langmuir 2002;18(15):5818-5822.
102. Ishino C, et al. Wetting transitions on rough surfaces. EPL (Europhysics Letters) 2004;68(3):419.
103. Maki DG, Stolz SM, Wheeler S, Mermel LA. Prevention of Central Venous Catheter-Related Bloodstream Infection by Use of an Antiseptic-Impregnated Catheter: A Randomized, Controlled Trial. Ann Intern Med 1997 August 15, 1997;127(4):257-266.
104. Yahav D, Rozen-Zvi B, Gafter-Gvili A, Leibovici L, Gafter U, Paul M. Antimicrobial lock solutions for the prevention of infections associated with intravascular catheters in patients undergoing hemodialysis: systematic review and meta-analysis of randomized, controlled trials. Clin Infect Dis 2008 Jul 1;47(1):83-93.
105. Trerotola SO, Johnson MS, Shah H, Kraus MA, McKusky MA, Ambrosius WT, et al. Tunneled hemodialysis catheters: use of a silver-coated catheter for prevention of infection--a randomized study. Radiology 1998 May 1998;207(2):491-496.
136
106. Logghe C, Van Ossel C, D'Hoore W, Ezzedine H, Wauters G, Haxhe JJ. Evaluation of chlorhexidine and silver-sulfadiazine impregnated central venous catheters for the prevention of bloodstream infection in leukaemic patients: a randomized controlled trial. J Hosp Infect 1997 Oct;37(2):145-156.
107. Dunser MW, Mayr AJ, Hinterberger G, Florl CL, Ulmer H, Schmid S, et al. Central venous catheter colonization in critically ill patients: a prospective, randomized, controlled study comparing standard with two antiseptic-impregnated catheters. Anesth Analg 2005 Dec;101(6):1778-1784.
108. Winson L. Catheterization: a need for improved patient management. Br J Nurs 1997 Nov 27-Dec 10;6(21):1229-1232, 1234, 1251-1222.
109. Akiyama H, Okamoto S. Prophylaxis of indwelling urethral catheter infection: clinical experience with a modified Foley catheter and drainage system. J Urol 1979 Jan;121(1):40-42.
110. Liedberg H, Lundeberg T. Silver alloy coated catheters reduce catheter-associated bacteriuria. Br J Urol 1990 Apr;65(4):379-381.
111. Kumon H, Hashimoto H, Nishimura M, Monden K, Ono N. Catheter-associated urinary tract infections: impact of catheter materials on their management. International journal of antimicrobial agents 2001;17(4):311-316.
112. Bach A, Eberhardt H, Frick A, Schmidt H, Bottiger BW, Martin E. Efficacy of silver-coating central venous catheters in reducing bacterial colonization. Crit Care Med 1999 Mar;27(3):515-521.
113. Mathews SM, Spalholz JE, Dubielzig RR, Grimson MJ, Reid TW. The Efficacy of an Organo-Selenium Coating on Silicone Hydrogel Contact Lenses. Investigative Ophtalmology and Visual Science 2002;43(12):3109.
114. Spallholz JE, Shriver BJ, Reid TW. Dimethyldiselenide and methylseleninic acid generate superoxide in an in vitro chemiluminescence assay in the presence of glutathione: implications for the anticarcinogenic activity of L-selenomethionine and L-Se-methylselenocysteine. Nutrition and cancer 2001;40(1):34-41.
115. Tran PL, Hammond AA, Mosley T, Cortez J, Gray T, Colmer-Hamood JA, et al. An Organo-Selenium Coating on Cellulose Inhibits Pseudomonas aeruginosa and Staphylococcus aureus Biofilm Formation. Appl Environ Microbiol 2009 April 3, 2009:AEM.02683-02608.
116. Brouwer WM, Zsom RLJ. Polystyrene latex particle surface characteristics. Colloids and Surfaces 1987;24(2-3):195-208.
117. Seebergh JE, Berg JC. Evidence of a hairy layer at the surface of polystyrene latex particles. Colloids and Surfaces A: Physicochemical and Engineering Aspects 1995;100:139-153.
![[320] Web 3: Selenium · for Selenium Java module for Selenium Ruby module for Selenium JavaScript mod for Selenium Chrome Driver Firefox Driver Edge Driver. Examples. Starter Code](https://img.pdfslide.us/doc/110x75/5eadce82cc4f0d7405687f01/320-web-3-selenium-for-selenium-java-module-for-selenium-ruby-module-for-selenium.jpg)
