Structuring of Bioactive Glass Surfaces at the Micrometer Scale by Direct Casting Intended to Influence Cell Response**

Biomed. Glasses 2016; 2:63–71

Research Article Open Access

Barbara Pföss*, Miriam Höner, Monika Wirth, Andreas Bührig-Polaczek, Horst Fischer, and Reinhard Conradt Structuring of bioactive glass surfaces at the micrometer scale by direct casting intended to influence cell response**

DOI 10.1515/bglass-2016-0008 Received Aug 12, 2016; revised Oct 29, 2016; accepted Oct 31, 2016 1 Introduction

Abstract: Defect-free bioactive glass surfaces with a Large bone defects cannot be healed by the body itself. grooved microstructure at the low micrometer scale were Since the availability of autologous material is limited, achieved by a mold casting process. The process was synthetic materials offer a promising solution for bone applied to the well-known glass compositions 45S5 and replacement applications [1]. In this respect, bioactive 13–93. Such microstructured surfaces may exhibit espe- glasses exhibit excellent properties: They are cytocompat- cially favorable conditions for bone cell orientation and ible, bind quickly to bone and can stimulate bone regener- growth. The aim of the study was to assess the parameter ation [2]. range for a successful casting process and thus to produce Cellular behavior is crucially influenced not only by samples suitable to investigate the interaction between the chemistry of the underlying substrate but by the sur- structured surfaces and relevant cells. Viscous flow in its face topography and mechanical properties as well. Sur- temperature dependence and thermal analysis were ana- face structures can increase the bone–to–implant con- lyzed to identify a suitable process window and to design tact and thus the fixation in the host bone [3]. It has a manageable time-temperature process scheme. Coun- been shown that cells react to surface structures from the teracting effects such as formation of chill ripples, mold nanometer to the micrometer scale [4, 5]. Furthermore it sticking and build-up of permanent thermal stress in the is well known that cell adhesion and even differentiation glass had to be overcome. A platinum gold alloy was cho- of cells can be influenced by surface structures in the low sen as mold material with the mold surface bearing the micrometer range [6–8]. Parallel groove–ridge structures mother shape of the microstructure to be imprinted on promote contact guided alignment of different cell types the glass surface. First experiments studying the behavior with the groove depth having more impact as the grooves of osteoblast-like cells, seeded on these microstructured become deeper [7, 9–11]. More than 5 µm in depth are glass surfaces revealed excellent viability and an orienta- needed to sufficiently guide the cells [7, 10, 12]. In addition of the cells along the microgrooves. The presented re- tion, McBeath et al. and Kumar et al. have shown that sults show that direct casting is a suitable process to pro- elongated stem cells rather differentiate into osteoblasts duce defined microstructures on bioactive glass surfaces. whereas rounded cells become adipogenic cells [13, 14]. Photolithography and etching methods enabled the Keywords: casting; surface topography; groove structure; possibility to manufacture surface topographies at the mi- cell guidance crometer scale [15–18]. A wide variety of materials have thus become available to study the phenomenon of contact guidance [19, 20]. Still, due to their strong crystalliza- *Corresponding Author: Barbara Pföss: Institute of Mineral Engi- tion tendency, bioactive glasses are mostly ineligible for a neering, RWTH Aachen University, 52064 Aachen, Germany; Email: [email protected]; contributed equally to this work Miriam Höner: Dental Materials and Biomaterials Research, RWTH Aachen University Hospital, 52074 Aachen, Germany; contributed Reinhard Conradt: Institute of Mineral Engineering, RWTH Aachen equally to this work University, 52064 Aachen, Germany Monika Wirth, Andreas Bührig-Polaczek: Foundry Institute, ** Special Section: Larry L. Hench Memorial Symposium on Bioac- RWTH Aachen University, 52072 Aachen, Germany tive Glasses at the Annual Meeting of the Glass & Optical Materials Di- Horst Fischer: Dental Materials and Biomaterials Research, RWTH vision (GOMD) of the American Ceramic Society, 22nd–26th May 2016, Aachen University Hospital, 52074 Aachen, Germany Madison, Wisconsin, USA

© 2016 B. Pföss et al., published by De Gruyter Open. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License. Bereitgestellt von | Universitätsbibliothek der RWTH Aachen Angemeldet Heruntergeladen am | 11.12.17 11:09 64 Ë B. Pföss et al. surface finish after manufacturing and therefore they are Combining the outstanding properties of bioactive rarely available for cell culture experiments regarding cell glass with those of a triggered surface topography is very behavior e.g. guidance phenomena on structured surfaces. promising and holds potential to further improve bone re- Itälä et al. manufactured micro-rough bioglass sur- placement applications. The presented results show that faces by chemical etching and found an improved os- a tailored casting process can be used to realize the de- teoblast attachment compared to polished control speci- manded glass surfaces. mens [21]. Following in vivo experiments even showed a higher amount of incorporated new bone in comparison to smooth implants [22]. Porous structures were realized 2 Methods by lithography-based additive manufacturing as well, cre- ating crystallized bioglass samples [23]. Efforts have also PtAu5 sheet metal with 8 mm thickness was used as sub- been made to structure bioglass, showing alignment of strate material, micrometer sized grooves were applied osteoblast-like MG-63 cells. But as the samples were sin- by a solid state ultrashort pulsed laser (Timebandwidth tered, the tested material became a glass-ceramic [24]. Sur- Duetto, Zurich, Switzerland). The applied groove depth to face structures on amorphous bioactive glasses have not be imprinted on the glass was 15 µm, the groove width been realized before. 30 µm and the ridge width 10 µm. To choose an appropriate testing pattern we took into Two bioactive glasses namely 45S5 [27] and 13–93 [28] consideration the results of above mentioned former stud- were synthesized via melt–quench technique. Batches of ies, where many cell types react to grooved substrates in the pure chemical components: Fused silica (Aachener the low micrometer range, mesenchymal stem cell size, as Quarz–Glas Technologie Heinrich, Aachen, Germany), well as the requirement that a surface structure on bioac- CaHPO4 (Merck, Darmstadt, Germany), CaCO3, Na2CO3 tive glass has to endure the different states of corrosion to (both AppliChem, Darmstadt, Germany) for 45S5, and ad- provide the intended biophysical cue throughout the for- ditionally MgO (VWR Chemicals, Darmstadt, Germany) mation of different layers after all. Moreover, when consid- and K2CO3 (Merck, Darmstadt, Germany) for 13–93, were ering the average viscosity of a glass melt, our calculations ∘ homogenized and then melted at 1400 C for 2 hours in a and experiments indicated that surface topographies at platinum crucible. Batch sizes yielding 300 g of glass were the scale of a few 10 µm suitable for casting. Finally a par- prepared and the melts were fritted into water. XRF and allel groove pattern of 30 µm groove, 10 µm ridge width, XRD analysis were carried out to ensure the correct com- and 15 µm height was applied to bioactive glasses in this position and amorphous nature of the material implying study. the described melting conditions. The glass frit was sieved A few mold materials were preliminary discussed and to different designated particle size fractions from 63 µm tested with minor success, as we decided to stay with plat- to 1 mm and provided for the following casting procedure inum gold. Platinum alloys are commonly used in combi- respectively. nation with glass melts and batches, as chemical reactions 15 g of glass granules were remelted in a small plat- are minimal up to high temperatures. Platinum gold is fur- inum crucible for 40 minutes at a temperature level Tcast ther known to be only poorly wetted by oxide glass melts, and then cast. The PtAu5 substrate bearing the microstruc- hence a favorable demolding behavior is expected. ture was preheated at a temperature T . To ensure the Discontinuous casting does not belong to the major mold correct mold temperature, the mold was stored in a sepa- forming methods for mass glass production, but it is typi- rate furnace wherein the casting was performed. The tem- cally used to manufacture bulk glass samples for a variety perature of the melt as well as the mold were varied in a se- of standard testing methods in glass science, or in the field ries of experiments. Demolding of the solidified glass was of glass art. accomplished by turning the substrate upside down and Microstructured components are for example fabri- carefully pounding from the back. The samples were then cated in the field of Micro-optics by using hot embossing annealed for 12 hours at a temperature level 20 K above the of borosilicate glass [25, 26]. Surface features of some mi- glass transition temperature Tg, and cooled down to room crometers are standard requirements for this application. temperature within 10 hours. Hot embossing requires heating the glass samples up to Polished bioactive glass 45S5 and 13–93 samples to be a temperature level sufficiently high enough for forming, used as control for cell culture experiments were cast from but due to the high crystallization tendency of bioactive ∘ 1400 C melt into graphite molds which were preheated at glasses, this process is hardly applicable to these glasses. ∘ 350 C as described elsewhere [29]. The cylinders were then

Bereitgestellt von | Universitätsbibliothek der RWTH Aachen Angemeldet Heruntergeladen am | 11.12.17 11:09 Structuring of bioactive glass surfaces by direct casting Ë 65

Table 1: Temperatures (T in ∘C) crucial for the casting process; subscripts g = glass transition, x = main bulk crystallization peak; m = bulk melting peak, liq = calculated liquidus temperature; stick = sticking temperature, cast = optimal casting temperature, strain = strain point; mold = optimal mold temperature.

Glass Tg Tx Tm Tliq Tstick Tcast Tstrain Tmold Tcontact log η 13 9.8 1.3 14.5 45S5 536 691 1217 1198 592 1190 517 500 653 13–93 589 972 1141 1134 663 1417 580 500 704

annealed and cooled in the same way as the structured ters A, B, T0 were obtained as follows: A is the intercept of samples. Afterwards, they were then cut into disks using the fit, T0 = aVFT · Tg, B = (13 − A) · (Tg − T0). a diamond charged saw (Buehler, Düsseldorf, Germany) Live/dead stainings were conducted to estimate the and ground using SiC grinding paper up to grit 4000 with initial biological behavior of the microstructured samples a grinding machine (Saphir 520, Ruubin 500, both ATM, in contact with osteoblast-like MG-63 cells. Three struc- Mammelzen, Germany). tured and polished bioglass samples of each composi- The obtained surface structures of the samples were tion as well as borosilicate glass cover discs (Carl Roth characterized using laser scanning microscopy. Measure- GmbH+Co, Karlsruhe, Germany) according to ISO 8255-1 as ∘ ments were performed at 50x magnification using a control were sterilized at 200 C for 2 hours. After the steril- Keyence VK X100 series microscope (Neu–Isenburg, Ger- ization process, the cells were seeded on the glass samples many). A surface area of 200 × 275 µm was analyzed to at a concentration of 20,000 cells/cm2 and incubated at ∘ evaluate the surface quality in three dimensions. In ad- 37 C and 5% CO2 for 24 hours. A mixture of 600 µl Ringer dition, two separate measurements, one of the glass and solution (Delta-Pharma, Pfullingen, Germany), 10 µl of one of the mold surface, were compared in order to evalu- 5 mg/ml propidium iodide (Sigma-Aldrich, Steinheim, Ger- ate the shaping accuracy of the casting process: Two in- many) in Ringer solution and 10 µl of 5 mg/ml fluorescein dividual two–dimensional linescans of 250 µm were su- diacetate (Sigma-Aldrich, Steinheim, Germany) in acetone perimposed. Laser scanning microscopy was used as a was prepared. The cells were then stained by giving 20 µl non-destructive analyzing method so that analyzed sam- of the mixture onto the samples. They were then analyzed ples could be used for further testing. Cross sections were by fluorescence microscopy (AXIO Imager M2m, Zeiss, additionally studied with a Zeiss Leo 440i scanning elec- Wetzlar, Germany) to visualize the number of living as well tron microscope (Oberkochen, Germany). Microstructured as dead cells and their orientation. bioactive glass samples were cut and embedded orthogo- nally to the microgrooves in order to inspect the shape and size of the surface topography. 3 Results Thermal analysis of the two bioactive glasses with heating and cooling rates of 10 K/min were performed Characteristic temperatures of the two bioactive glasses with a Setaram Instrumentation Setsys DTA 16 apparatus were obtained from thermal analysis. These are the glass (Caluire, France) using an empty platinum crucible as ref- transition temperature, Tg (as established from the point erence. Data was adapted with a calibration measurement of inflection of the DTA curves), the bulk crystallization using empty crucibles. maximum temperatures Tx during up-scan, and the bulk Experimental viscosity data from Vedel et al. [30] were melting temperature, Tm. The range displaying endother- reevaluated on the basis of a VFT interpolation using the mal melting effects after crystallization comprises three numerical method of linearized coordinates. For this pur- and two consecutive individual peaks for the bioactive pose, a linear regression analysis of the experimental log η glasses 45S5 and 13–93 respectively as indicated in Fig- data vs. the auxiliary function ure 1. Individual peak minima were averaged to a single (︂ )︂ Tg Tg (1 − aVFT ) · value Tm, see Table 1. Note that Tm must be carefully dis- f = T (1) T Tg tinguished from the liquidus temperature, Tliq. Calculated 1 − aVFT · T values for Tliq using thermochemical modelling are also was performed under a numerical variation of the constant listed in Table 1. aVFT for minimum square deviation between the experi- The casting temperature was set to a level Tcast corre- mental data and the regression fit line. The VFT parame- sponding to log η = 1.3. This viscosity level was found

Bereitgestellt von | Universitätsbibliothek der RWTH Aachen Angemeldet Heruntergeladen am | 11.12.17 11:09 66 Ë B. Pföss et al.

Figure 1: DTA up-scan analysis of bioactive glasses 45S5, sample mass 39.109 mg (left) and 13–93, sample mass 32.490 mg (right), performed with Setsys TG-DTA 16 apparatus, heating rate 10 K/min. Tg = glass transition temperature from point of inflection, Tx = maximum crystallization temperature, Tm = average melting temperature resulting from individual peak minima.

Figure 2: 2D-linescans of glass surfaces from samples cast from varying viscosity levels. 1420 ∘C: log(η) = 1.29; 1400∘C: log(η) = 1.35; 1380∘C: log(η) = 1.42; η in dPas. to be sufficiently low to allow a rapid spread of themelt and a reliable filling of the microgrooves. Figure 2 compiles Figure 3: Viscosity-temperature relation from VFT interpolation for the results of a series of different casting temperatures and bioactive glasses 45S5 and 13–93; data taken from Vedel et al. [30]; therefore the impact of the viscosity level of the melt on temperatures Tstrain,Tstick, and Tcast correspond to viscosity levels form filling. It is clearly visible that a temperature differ- of log η = 14.5, 9.8, 1.3, respectively; η in dPas. ence of only 40 K has a significant impact on the achieved ridge height of the microstructure. With a casting viscosity was avoided by taking into account the results by Rieser of log(η) = 1.42 (η in dPas), only very weak impressions et al. [32]. They identified so-called contact sticking condi- of the microstructure could be found on the glass surface, tions, suggesting that mold sticking during demolding is whereas with log(η) = 1.29 (η in dPas) the grooves in the safely avoided at temperatures below a viscosity level Tstick mold were sufficiently filled and a ridge structure onthe corresponding to log η = 9.8 (η in dPas). The temperature glass with more than 10 µm in height was generated. level Tstick addresses the contact temperature Tcontact at the If heat dissipation is too fast, concentric waves can ap- interface between glass melt and mold material. Tcontact pear at the glass surface. The effect of the so called chill was calculated according to Rieser [33]. Thus, sticking was ripples [31] constrains the filling of the microgrooves and indeed safely avoided by setting the mold temperature therefore has to be avoided. This could be achieved by ∘ well below Tstick at 500 C for both glasses. The significant choosing a mold pre-heating temperature T at an el- ∘ mold effect of a variation in the mold temperature from 500 C evated temperature. Yet further constraints for the pro- ∘ to 580 C can be seen in Figure 3. The higher mold temper- cess temperatures resulted from requirements of demold- ature leads to sticking of the glass samples and complica- ing and thermal stress release. Sticking during demolding

Bereitgestellt von | Universitätsbibliothek der RWTH Aachen Angemeldet Heruntergeladen am | 11.12.17 11:09 Structuring of bioactive glass surfaces by direct casting Ë 67

itself. The periodical and defect-free micro topography was also confirmed by cross sectional SEM analysis (Figure 7). The comparison of the laser scanning results from a series of cast glass samples verifies the stability of the process. Figure 8 shows how the microstructure could be re- produced on the surface of ten different 13–93 glass samples manufactured under identical conditions. Figure 4: Surfaces of two cast 13–93 glass samples with varying ∘ Live/dead stainings (Figure 9) show predominately mold temperature. Tmold = 580 C (left) led to sticking whereas Tmold = 500∘C (right) ensured proper demolding and therefore a high green fluorescent cells, indicating a high cell viability on quality glass surface. the structured as well as on the polished bioactive glasses (here: 45S5). In addition, when compared to cells on polished bioactive glass samples, the cells spread and formed long directional extensions. Notably, the cellular morphology appeared polarized and oriented in the direction of the grooves, revealing that microstructured bioactive glass surfaces influence cellular guidance. The cell culture experiments using 13–93 showed similar results.

Figure 5: Surfaces of glass 13–93 after remelting 125–180 µm particle size precursor material (left) yielding defects from remaining 4 Discussion bubbles and 355–1000 µm particle size precursor material (right) achieving a high quality surface. A successful casting process depends on the choice of crucial temperature levels in a most sensitive way. As typi- tions in demolding which results in a poor surface qual- cal in glass technology, these temperature levels are com- ∘ ity. Setting Tmold at 500 C allowed to demold quickly and municated in terms of temperatures at which specific vis- therefore prevent destruction due to thermal stress. Tem- cosity values are reached, making these settings useful perature levels below the strain point at log η = 14.5 (η in for glasses of different chemical compositions and viscos- dPas) are critical and swiftly transferring the cast sample ity temperature curves. A casting temperature Tcast corre- to an annealing furnace kept at Tg + 20 K is essential. sponding to a viscosity level of log η = 1.3 (η in dPas) Further process parameters, such as Tcast,Tstick and was found to be sufficiently high to ensure mold filling. Tstrain related to viscosity levels are marked in Figure 4 and The mold sticking level of log η = 9.8, originally obtained a complete set of relevant temperatures yielding from ther- for smooth mold surfaces [31], was also applicable to mold mal analysis and the viscosity temperature curve is com- structures at the 10 µm scale. The remaining temperatures piled in Table 1. related to viscosity, i.e.,Tg and Tstrain are commonplace lev- Further attention was given to the effect of the quality els of glass technology and processing in general [34, 35]. ∘ of the glass to be remelted from powders with constant par- Tmold at 500 C was a compromise found for both glass ticle size fractions. In accordance with the intended rheo- compositions regarding chill ripples, sticking and crack- logical parameter and melting time, glass granules from ing due to thermal stress. 355–1000 µm were found to yield a bubble free melt. This Taking into consideration that the platinum gold alloy is essential for the quality of the surface obtained. Figure 5 is known to be only poorly wetted by glass melts and the compares the quality of the glass surface when using 125– temperature dependence of the surface tension of the melt 180 µm and 355–1000 µm granules of the 13–93 glass com- is not significant, we conclude that viscosity is the decisive position as precursor material. Holes on the surface result- parameter for form filling on the low micrometer scale. Fig- ing from bubbles in the glass melt appeared for the smaller ure 2 confirms the high sensitivity of the process ofform particle size fraction. filling on viscosity. 2D laser scanning microscopy linescans of glass sur- It comes to notice, that the temperature corresponding faces and mold and 3D images confirm the accuracy of the to the casting viscosity Tcast for bioactive glass 45S5 is very developed casting process, see Figure 6. The acute peaks in close to the liquidus temperature of the crystallized mate- the 2D linescans, especially at the edges of the grooves, are rial and below the experimentally derived parameter Tm. no real surface features but artifacts of the laser technique Comparable rheological conditions for both glasses were

Bereitgestellt von | Universitätsbibliothek der RWTH Aachen Angemeldet Heruntergeladen am | 11.12.17 11:09 68 Ë B. Pföss et al.

Figure 6: 2D laser scanning profile of the surface microstructure on bioactive glass 45S5 (top) and 13–93 (bottom) superimposed with2D scan of the mold surface; as well as 3D laserscanning profile of the glass surfaces (right).

Figure 8: Ten superimposed 2D-linescans of different 13–93 glass samples manufactured with identical conditions, showing the high reproducibility.

perature conduction and mold melt interaction direct cast- Figure 7: SEM micrograph of the cross section of cast bioactive ing can be used to produce bioactive glass samples with glass 45S5 showing the surface microstructure; BSE mode, 15 kV an oriented microtexture on the surface. Results from ther- accelerating voltage. mal analysis and the viscosity temperature curve are very convenient tools to set process parameters. Continuance requested, however the corresponding viscosity levels of in the provided glass material for remelting is also deci- glass 13–93 are at significantly higher temperatures than sive to prevent remaining bubbles. A particle size fraction those of glass 45S5. Though setting the temperature at Tcast of 355–1000 µm was found to be suitable. for glass 45S5 and preventing crystallization was possible Our cell culture results on 45S5, which is a well-known by first heating above the bulk melting temperature Tm and cytocompatible material and the most widely used bioac- cooling down to Tcast after that. tive glass composition [36], are in accordance with ear- The results show that with sufficient considerations lier observations in literature. All cells appeared green, regarding the material preparation, as well as time tem- indicating they were all alive, showing an excellent cy-

Bereitgestellt von | Universitätsbibliothek der RWTH Aachen Angemeldet Heruntergeladen am | 11.12.17 11:09 Structuring of bioactive glass surfaces by direct casting Ë 69

Figure 9: Live/dead staining of MG-63 cells on polished (left) and microstructured surface (right) of bioactive glass 45S5; the cells are oriented along the microstructures; active viable cells are stained with a green fluorescent dye, necrotic cells would have appeared in red (none are detected). tocompatibility of the material. It is known that different struction of the sample by thermal stresses could be con- cell types react to patterned surfaces: Roughened bioma- trolled and mostly eliminated. The initial response of the terial surfaces in the micrometer range can enhance cell structured bioactive glass surfaces on osteoblast-like cells adhesion in vitro [37, 38] and bone-to-implant contact in was also estimated. The cellular morphology appeared ori- vivo [3]. Itälä et al. have shown that microroughening of ented in the direction of the grooves known as contact porous bioactive glass surfaces enhance the biological ac- guidance. Though our investigations are still fundamental tivity of certain bioactive glass compositions [22]. In this and restricted to planar samples, the derived results indi- study first live/dead experiments on structured bioactive cate that the microstructuring of bioactive glass surfaces glasses have shown that almost all cells appear green, dis- might offer a promising method to enhance bone healing playing alive cells. Furthermore, most cells are well spread processes. and oriented along the structures, known as contact guidance [19]. As cell shape and orientation are associated with Acknowledgement: Thanks to Univ.-Prof. Dr. med. dent. changes in the differentiated phenotype of cells [13, 14, 39], Hendrik Meyer-Lückel and PD Dr. Marcella Esteves- we conclude that a tailored guidance of cells by melt- Oliveira, Department of Operative Dentistry, Periodon- derived microstructured bioactive glasses might be possi- tology and Preventive Dentistry, RWTH Aachen Univer- ble. sity Hospital for their help with the experiments on the laser scanning microscope. We also thank Andreas Dohrn, Fraunhofer ILT Aachen for his help concerning structuring 5 Conclusion of the molds with laser. We are grateful to the Deutsche Forschungsgemein- schaft (DFG) for financial support (grant # FI 975/21-1; CO We elaborated process parameters allowing to generate 249/14-1, BU 1072/36-1). periodically and defect-free structured bioactive glass surfaces at the 10 µm scale by direct casting. Casting temperature and mold temperature were defined under consideration of viscosity levels and crystallization temperatures. References As a result, the negative effects of chill ripple formation, crystallization, mold sticking, residual bubbles, and de- [1] Oryan A., Alidadi S., Moshiri A. and Maffulli N., Bone regener- ative medicine: classic options, novel strategies, and future di-

Bereitgestellt von | Universitätsbibliothek der RWTH Aachen Angemeldet Heruntergeladen am | 11.12.17 11:09 70 Ë B. Pföss et al.

rections, J Orthop Surg Res, 2014, 9, 1–27. [20] Walboomers X. F., Croes H. J., La Ginsel and Jansen J. A., Contact [2] Hench L. L., Splinter R. J., Allen W. C. and Greenlee T. K., Bond- guidance of rat fibroblasts on various implant materials, Journal ing mechanisms at the interface of ceramic prosthetic materi- of Biomedical Materials Research, 1999, 47, 204–212. als, Journal of Biomedical Materials Research, 1971, 5, 117–141. [21] Itälä A., et al., Characterization of microrough bioactive glass [3] Hallgren C., Reimers H., Chakarov D., Gold J. and Wennerberg surface: Surface reactions and osteoblast responses in vitro, J. A., An in vivo study of bone response to implants topographi- Biomed. Mater. Res., 2002, 62, 404–411. cally modified by laser micromachining, Biomaterials, 2003, 24, [22] Itälä A., Koort J., Ylänen H. O., Hupa M. and Aro H. T., Biologic 701–710. significance of surface microroughing in bone incorporation of [4] Wittenbrink I., et al., Low-aspect ratio nanopatterns on bioinert porous bioactive glass implants, J. Biomed. Mater. Res., 2003, alumina influence the response and morphology of osteoblast- 67, 496–503. like cells, Biomaterials, 2015, 62, 58–65. [23] Tesavibul P., et al., Processing of 45S5 Bioglass® by [5] Neuss S., et al., Adhesion of human mesenchymal stem cells lithography-based additive manufacturing, Materials Let- can be controlled by electron beam-microstructured titanium al- ters, 2012, 74, 81–84. loy surfaces during osteogenic differentiation, Biomedical Engi- [24] Detsch R., Guillon O., Wondraczek L. and Boccaccini A. R., Initial neering / Biomedizinische Technik, 2015, 60. Attatchment of rMSC and MG-63 Cells on Patterned Bioglass® [6] Matsuzaka K., Walboomers X. F., Ruijter J. E. de and Jansen J. A., Substrates, Adv. Eng. Mater., 2012, 14, B38–B44. The effect of poly-L-lactic acid with parallel surface micro groove [25] Schubert A., Edelmann J. and Burkhardt T., Micro structuring on osteoblast-like cells in vitro, Biomaterials, 1999, 20, 1293– of borosilicate glass by high-temperature micro-forming, Mi- 1301. crosyst Technol, 2006, 12, 790–795. [7] Clark P., Connolly P., Curtis A. S., Dow J. A. and Wilkinson C. D., [26] Edelmann J., Worsch C., Schubert A. and Rüssel C., Micro struc- Topographical control of cell behaviour. I. Simple step cues, De- turing of inorganic glass by hot embossing of coated glass velopment, 1987, 99, 439–448. wafers, Microsyst Technol, 2010, 16, 553–560. [8] Fadeeva E., Schlie S., Koch J. and Chichkov B. N., Selective cell [27] Hench L. L., Chronology of Bioactive Glass Development and control by surface structuring for orthopedic applications, Jour- Clinical Applications, NJGC, 2013, 03, 67–73. nal of Adhesion Science and Technology, 2010, 24, 2257–2270. [28] Brink M., Turunen T., Happonen R. and Yli-Urpo A., Composi-

[9] Boyan B. D., Hummert T. W., Dean D. D. and Schwartz Z., Role tional dependence of bioactivity of glasses in the system Na2O- of material surfaces in regulating bone and cartilage cell re- K2O-MgO-CaO-B2O3-P2O5-SiO2, Journal of Biomedical Materi- sponse, Biomaterials, 1996, 17, 137–146. als Research, 1997, 37, 114–121. [10] Martínez E., Engel E., Planell J. and Samitier J., Effects of artificial [29] Plewinski M., Schickle K., Lindner M., Kirsten A., Weber M. and micro- and nano-structured surfaces on cell behaviour, Annals Fischer H., The effect of crystallization of bioactive bioglass of Anatomy - Anatomischer Anzeiger, 2009, 191, 126–135. 45S5 on apatite formation and degradation, Dental materials : [11] Mwenifumbo S., Li M., Chen J., Beye A. and Soboyejo W., oflcial publication of the Academy of Dental Materials, 2013, 29, Cell/surface interactions on laser micro-textured titanium- 1256–1264. coated silicon surfaces, Journal of Materials science: Materials [30] Vedel E., Arstila H., Ylänen H., Hupa L. and Hupa M., Predict- in medicine, 2007, 18, 9–23. ing physical and chemical properties of bioactive glasses from [12] Matsuzaka K., Frank Walboomers X., Yoshinari M., Inoue T. and chemical composition. Part 1: Viscosity Jansen J. A., The attachment and growth behavior of osteoblast- characteristics, Glass Technology - European Journal of Glass like cells on microtextured surfaces, Biomaterials, 2003, 24, Science and Technology Part A, 2008, 49, 251–259. 2711–2719. [31] Brückner R., Studie über das Zusammenspiel von Gren- [13] McBeath R., Pirone D. M., Nelson C. M., Bhadriraju K. and Chen zflächenspannungen und Fließverhalten beim Gießen von C. S., Cell shape, cytoskeletal tension, and RhoA regulate stem Schmelzen und keramischen Suspensionen (Study of the inter- cell lineage commitment, Developmental cell, 2004, 6, 483– action between surface tension and viscosity during casting of 495. melts and ceramic suspensions), Berichte Deutsche Keramis- [14] Kumar G., et al., The determination of stem cell fate by 3D scaf- che Gesellschaft, 1964, 41, 534–541. fold structures through the control of cell shape, Biomaterials, [32] Rieser D., Spieß G. and Manns P., Investigations on glass- 2011, 32, 9188–9196. to-mold sticking in the hot forming process, Journal of Non- [15] Lenhert S., Meier M.-B., Meyer U., Chi L. and Wiesmann H. P., Crystalline Solids, 2008, 354, 1393–1397. Osteoblast alignment, elongation and migration on grooved [33] Rieser D., Manns P., Spieß G. and Kleer G., Investigations on polystyrene surfaces patterned by Langmuir– Blodgett lithog- sticking temperature and wear of mold materials and coatings, raphy, Biomaterials, 2005, 26, 563–570. Advances in Fusion and Processing of Glass III, Volume 141, [16] Clark P., Cell behaviour on micropatterned surfaces, Biosensors 2004, 281–289. and Bioelectronics, 1994, 9, 657–661. [34] Jebsen-Marwedel H. and Brückner R., Glastechnische Fab- [17] Curtis A. and Wilkinson C., Topographical control of cells, Bio- rikationsfehler, “Pathologische” Ausnahmezustände des Werk- materials, 1997, 18, 1573–1583. stoff es Glas und ihre Behebung; eine Brücke zwischen Wis- [18] Riedel N. A., Williams J. D. and Popat K. C., Ion beam etching senschaft, Technologie und Praxis, Springer, Berlin, New York, titanium for enhanced osteoblast response, Journal of materials 2011. science, 2011, 46, 6087–6095. [35] Scholze H., Glas, Natur, Struktur und Eigenschaften, Springer, [19] Weiss P., Cell contact, International review of cytology, 1958, 7, Berlin, Heidelberg, 1988. 391–423. [36] Jones J. R., Review of bioactive glass: From Hench to hybrids, Acta Biomaterialia, 2013, 9, 4457–4486.

Bereitgestellt von | Universitätsbibliothek der RWTH Aachen Angemeldet Heruntergeladen am | 11.12.17 11:09 Structuring of bioactive glass surfaces by direct casting Ë 71

[37] Recum A. F. V., Shannon C. E., Cannon C. E., Long K. J., van [39] Meyer U., Buchter A., Wiesmann H. P., Joos U. and Jones D. B., Kooten T. G. and Meyle J., Surface roughness, porosity, and Basic reactions of osteoblasts on structured material surfaces, texture as modifiers of cellular adhesion, Tissue engineering, Eur Cell Mater, 2005, 9, 39–49. 1996, 2, 241–253. [38] Lincks J., et al., Response of MG63 osteoblast-like cells to titanium and titanium alloy is dependent on surface roughness and composition, Biomaterials, 1998, 19, 2219–2232.

Bereitgestellt von | Universitätsbibliothek der RWTH Aachen Angemeldet Heruntergeladen am | 11.12.17 11:09