- Email: [email protected]
Influence of nanoscale surface roughness on neural cell attachment on silicon
Nanomedicine: Nanotechnology, Biology, and Medicine 1 (2005) 125 – 129 www.nanomedjournal.com
Basic Research
Influence of nanoscale surface roughness on neural cell attachment on silicon Saida P. Khan, PhD(c), Gregory G. Auner, PhD, Golam M. Newaz, PhDT Wayne State University, Detroit, Michigan Received 2 December 2004; accepted 31 March 2005
Abstract
The adherence and viability of neural cells (primary cortical cells) from rat embryo on silicon wafers with varying surface roughness (10 to 250 nm) at the nano scale were investigated. The roughnesses were achieved by using chemical etching. Atomic force microscopy was utilized to determine surface roughness. We examined the adherence and viability of neural cells by using scanning electron microscopy and fluorescence immunoassaying. Antineuron-specific enolase antibody was used for immunostaining. Results from this investigation show that for these specific neural cells, there is an optimum surface roughness range, Ra = 20 to 100 nm that promotes cell adhesion and longevity. For silicon-based devices, this optimum surface roughness will be desirable as a suitable material/neuron interface. D 2005 Elsevier Inc. All rights reserved.
Key words:
Neuron; Nanoscale; Cell attachment; Roughness; Silicon
Design and development of neural implant device require the integration of electronics and biological systems. An electronic device can be used as a neural prosthesis in drug delivery, nerve regeneration, functional electrode stimulation, pain relief, and so on. But good incorporation of nerve cells and their viability with the device is a prerequisite for its proper function. A particular area of interest in this field of research is the study of the interface between a material surface and nerve cells in vitro. A neural implant device may have a complex design with various types of materials with different shapes and surfaces joined together. Cell adhesion and healthy growth on specific surfaces are very important for the integration and function of the device. A survey of the current literature shows that studies of cell adhesion and growth have been focused on two principal areas. The first is the application of various coatings or chemical modification of the surface, such as using polylysine, and the second is topological modification of the surface. Although chemical modification has No financial conflict of interest was reported by the authors of this paper. T Corresponding author. Mechanical Engineering Department, Wayne State University, 5050 Anthony Wayne Drive, Detroit, MI 48202. E-mail address: [email protected] (G. Newaz). 1549-9634/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.nano.2005.03.007
attained considerable success in good adhesion and directed growth of nerve cells for in vitro cases, further investigation is necessary to determine its long-term in vivo applications. On the other hand, topological surface modification facilitates studying the direct interface between the cell and the surface. Scientists have realized that topographical cues may play a crucial role in mediating cell orientation and biocompatibility [1,2]. It is well known that the surface microstructure of artificial biomedical materials has a role in cell attachment [3,4]. The effect of topological modification on various surfaces such as glass, quartz [5], metal oxides [6], and silicon (Si) [7-12] has been studied by several investigators. Silicon surfaces have been studied most extensively because their semiconductor properties make Si the most attractive material of choice for neural implants. It has been reported that porous Si enhances growth of neurons [7], and nanometer-scale pillars and wells produced by reactive iron etching can improve astrocyte adhesion on Si [9]. Fan et al [12] studied the quantitative relationship between the roughness of the Si surface and neural cell (substantia niagra) adherence. Responses of other types of neural cells and entities should also be studied. Fan et al [12] considered a specific cell type, while our investigation focused on mixed neuronal culture. In this investigation we
126
S. Khan et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 1 (2005) 125–129
Fig 1. Three-dimensional AFM topography of the surface of the Si wafer and the profile of the section. A, Si wafer with Ra = 64 nm. B, Si wafer with Ra = 18 nm.
focused on primary cortical cells from the rat embryo, which were directly inoculated onto the surface of Si wafers with 3 different average roughnesses ranging from 18 to 201 nm. No coating or chemical modification was used before cell plating. Examination of cell attachment and viability were studied using scanning electron microscopy (SEM) and neuron-specific enolase (NSE) immunoassaying. Silicon wafers The double-sided polished surface of Si (111) wafers (Institute of Electronic Materials Technology, Poland) was used for this surface roughness investigation. The wafers were cleaned using the standard method of washing with acetone and distilled water, and then blowing dry with nitrogen gas. Wafers used as controls (untreated) were rinsed with deionized water. Surface roughness of the etched wafers was controlled by variable exposure time to HF/HNO3/H2O
(1:1:10, v/v) solutions. Thus, Si wafers with 3 different surface roughnesses (with mean Ra) were prepared. In this experiment all prepared Si wafers were exposed to the air to form a stable oxide (SiO2) layer before further use. Both sides of Si wafers were sterilized via autoclaving at 1218C for 40 minutes before being inoculated with neural cells. The topography of Si wafers was measured by atomic force microscopy (AFM) (Autoprobe, Park Scientific Instruments) using the contact mode in air. By using AFM, surface topography was resolved on a nanometer scale with a scan range of 1600 1600 nm2. The Ra was evaluated using the surface profile data extracted from AFM images. Raman spectra were used to check the chemical groups on the wafer surface before and after etching. The laser raman spectrum was obtained using a Renishaw Invia Raman Microscope, at a wavelength range of 400 to 2000/cm. Cell culture Neural cells were cultured using a protocol modified from Brewer et al [13]. Rat cortical neurons were collected
Fig 2. Raman spectra of the surface of Si wafers with different roughnesses are showing no chemical change due to etching.
Fig 3. Dependence of cell viability (6 days after inoculation) on Ra of Si wafers. The cell coverages from 3 separate samples are shown for each average roughness.
S. Khan et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 1 (2005) 125–129
127
Fig 4. Scanning electron microscopy images of the primary cortical neuron culture at 6 days. A, Surface Ra = 18 nm (scale bar = 30 Am). B, Surface Ra = 64 nm (scale bar = 30 Am). C, Magnified image of a cluster of neurons on surface Ra = 64 nm (scale bar = 25 Am). D, Magnified image of a growth cone extending from a neuron (scale bar = 10 Am). E, Surface Ra = 204 nm (scale bar = 60 Am).
from the cerebral cortex of prenatal E19 Sprague Dawley rats under sterile conditions. The 1-cm2 Si samples were placed in 12 well culture dishes. Cells were inoculated on the surface of the prepared wafers. Some empty dishes were used as controls; neuron cells were directly inoculated on the culture plastic. Cells were cultured in neurobasal (GIBCO) medium supplemented with glutamine, B27, and 4% fetal bovine serum (FBS), and then incubated in a humid 5% CO2 atmosphere at 378C. The cells were counted using trypan blue in a hemocytometer on the day of inoculation. Cells were inoculated at a density of 2.28 105 cells per square centimeter on the surface of the prepared Si wafers. Wells without Si wafers and where the neuron cells were inoculated directly on the culture plastic were regarded as controls. Cell adherence and viability After 6 days, cell cultures used for SEM observation were fixed in 4% paraformaldehyde using 0.1 mol phosphate buffered saline (PBS) at 378C for 30 minutes. These samples were processed according to a standard procedure: After gradual dehydration in ethanol, they were critical-point dried and spattered with gold coating for conduction. The cultures were then observed (Hitachi S 2400, Hitachi Instruments) with an acceleration voltage of 25 keV. For immunostaining, cell cultures were fixed with 4% paraformaldehyde in 0.1 mol PBS at 378C for 30 minutes, and then rinsed 3 times in 0.1 mol PBS. Cultures were permeabilized with 0.1% Triton X-100 in
0.1 mol PBS (PBS-t) for 5 minutes at room temperature. To avoid nonspecific binding, the cells were first washed twice with 5% FBS in HankTs balanced salt solution (HBSS) for 10 minutes. Anti-rat NSE in HBSS was used on fixed cells as the primary antibody, and the cells were kept at 48C overnight. Before the secondary antibody was used, the cells were washed again in 5% FBS in HBSS for 10 minutes 3 times. A 0.02% Na azide solution was used as an antibacterial agent. As a secondary antibody, antimouse IgG fluorescein isothiocyanate (Sigma) was used, at a dilution of 1:100 for 1 hour. After immunostaining, cultures were observed with a Diaphot 300 inverted microscope (Nikon). For statistical analysis and reproducibility verification, each experiment was performed with 3 samples. Results and discussion Surface topography The surface roughness of all Si wafers was measured quantitatively by AFM. The 3 different Ra of the Si wafersT surfaces, produced by various etching conditions, were 18, 64, and 204 nm. The AFM surface topography and profiles of 2 examples of Si wafers with differing surface roughness are shown in Figure 1. The first surface of the Si wafers was produced by etching in 2:3:10 of HF/HNO3/H2O solution, with an Ra = 18 nm (Figure 1, A). When wafers were etched in 2:3:10 of HF/HNO3/H2O solution for 10 minutes, the Ra increased to 204 nm.
128
S. Khan et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 1 (2005) 125–129
Fig 5. Fluorescence images of the fluorescein isothiocyanate-labeled immunohistologically stained primary cortical neurons cultured on (A) Si wafer Ra = 64 nm, and (B) Si wafer Ra = 204 nm. The gray emission corresponds to anti-NSE immunostaining (scale bar = 15 Am).
To check for differences in the chemical status of the surface of Si wafers before any preparation and after etching in HF/HNO3/H2O solutions, raman spectroscopy was used. There were no differences among Raman spectra of Si wafers with differing surface roughness (Figure 2). Cell adherence and viability Scanning electron microscopy and immunohistochemical observations showed that the adhesion and viability of primary cortical neurons were significantly dependent on the surface roughness of the Si wafer. Counting cells was difficult because of cell clusters. Therefore, we used a different approach than Fan et al [12]. We calculated the percent coverage of the surface by the neurons at different surface roughnesses. Our observations were fairly consistent with Fan et al [12]. We found that at an average roughness of 64 nm, cells cover a greater surface area than a surface roughness of 18 nm or 204 nm. A summary of average cell density on Si wafers by surface Ra is shown in Figure 3. Cell coverages were counted in at least 3 random 0.5 0.5 mm2 fields in each of the 3 samples, and then an average was taken. It is apparent from this figure that optimum neural adhesion occurred on Si wafers with surface Ra in the neighborhood of 64 nm, in cases where surface coverage was measured 5 days after inoculation. On surfaces with Ra of less than about 60 nm, and on rough surfaces with Ra above 70 nm, cell adhesion was much lower. These results are similar to those of Fan et al [12]. Figure 4 shows SEM images of the cells at different surface roughnesses. Contact adhesion is a precondition for the neurite growth of neuronal cells. The outgrowth of axons on the modified silicon as shown in Figure 4, D indicates good adherence on the surface. Some matrix-like matter, which should be some protein excreted by the living cells adsorbed by the surface for attachment, can be seen around the cells in Figure 4, B and C. Figure 4, A and E show poor cell attachment for surfaces with roughness of 18 nm and 204 nm, respectively. Figure 5 shows the fluorescence images of stained neurons. To avoid excessive astrocyte growth, a low concentration of serum (4%) was used in the media. This
low concentration may have caused the abovementioned neuron clustering as a result of low adhesion. Serum contains various growth factors and proteins that promote cell adhesion. Therefore, common practice in neuronal culture methods is to apply 10% to 15% serum in growth media. With the plated cell density, an average surface coverage should be about 10%, while in our surfaces the range of surface coverage was 1.1% to 1.9%. It is apparent that without use of any adhesion layer (eg, polylysine), initial attachment is poor. Si and SiO2 surfaces do not have cytotoxicity to neural cells [7], but a low affinity for cell adherence was reported for B104 neuroblastoma cells on SiO2 substrate [14]. A study of vascular systems [15] showed that when an artificial system interacts with a biological system, proteins and other macromolecules adsorb to the surface. This protein layer may act as an adhesion layer between the cell and the surface. It has also been shown that surface composition and properties influence this adsorption [16]: Low-energy surfaces promote protein adsorption more than high-energy surfaces. Surface roughness is correlated to interfacial energy [17] and wettability of the surface. The surface roughness can affect the adhesion of the cell due to increased or decreased contact area, which is proportionate to the interfacial adhesive force. Therefore, surface roughness may influence the contact area of the cell membrane with the substrate. It seems that at the optimum surface roughness, the cell body contacts the surface in the most suitable manner required for proper attachment and subsequent axonal proliferation.
Conclusions Surface roughness has considerable influence on neural cell attachment on bare Si. It appears that at the nano-scale level, neuronal cells recognize a surface roughness range for optimum attachment. Neurons do not readily attach on very smooth or rough surfaces. From an average roughness of 0 to 64 nm, cell adherence increases with roughness; but at roughness of around 204 nm or more, roughness negatively affects attachment. This investigation was an initial study to determine the relative influence of roughness on cell attachment. From this study, it can be concluded that without use of proper adhesive coating, cell attachment is poor on bare Si surfaces, but this can be considerably moderated by changing surface roughness.
References [1] Clark P, Britland S, Connolly P. Growth cone guidance and neuron morphology on micropatterned laminin surface. J Cell Sci 1993;105: 203 - 12. [2] von Recum AF, den Braber ET, de Ruijter JE, Smits HTJ, Ginsel LA, Jansen JA. Quantitative analysis of cell proliferation orientation on substrata with uniform parallel surface micro-grooves. Biomaterials 1996;17:1093 - 9.
S. Khan et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 1 (2005) 125–129 [3] Curtis AG, Wilkinson CDW. Reaction of cell to topography. J Biomater Sci Poly Edu 1998;9:1313 - 29. [4] Russell TP, Kim HC. Tack—a sticky subject. Science 1999;285: 1219 - 21. [5] Rajnicek A, Britland S, McCaig C. Contact guidance of CNS neuritis on grooved quartz, neuronal age and cell type. J Cell Sci 1997;110: 2905 - 13. [6] Kawana A, Torimitsu K. Selective growth of sensory nerve fibers on mental oxide pattern in culture. Dev Brain Res 1990;51:128. [7] Bayless SC, Bucberry LD, Fletcher I, Tobin MJ. The culture of neurons on silicon. Sensor Actuators A 1999;74:139 - 42. [8] Turner JN, Shain W, Szarowski H, Anderson M, Martins M, Isaacson M, et al. Cerebral astrocyte response to micromachined silicon implants. Exp Neurol 1999;156:33 - 49. [9] Turner AMP, et al. Attachment of astroglial cells to microfabricated pillar arrays to different geometries. J Biomed Res 2000;51:4305. [10] Craighead HG, et al. Chemical and topographical patterning for directed cell attachment. Curr Opin Sol State Mat Sci 2001;5:177. [11] Fan YW, et al. Improvement of neural cell adherence to silicon surface by hydroxyl ion implantation. Surface Coating Technol 2000;131:355.
129
[12] Fan YW, Cui FZ, et al. Culture of neural cells on silicon wafers with nanoscale surface topograph. J Neurosci Methods 2002;120: 17 - 23. [13] Brewer GJ, Torricelli JR, Evege EK, Price PJ. Optimized survival of hippocampal neurons in B27-supplemented Neurobasalk, a new serum-free medium combination. J Neurosci Res 1993;35:567 - 76. [14] Corey JM, Brunette AL, Chen MS, Weyhenmeyer JA, Brewer GJ, Wheeler BC. Differentiated B104 neuroblastoma cells are a highresolution assay for micropatterned substrates. J Neurosci Methods 1997;75:91 - 7. [15] Vormman L, Adams AL, Klings M, Fischer GC, Munoz PC, Solensky R. Reactions of formed elements of blood with plasma proteins at interfaces. Ann N Y Acad Sci 1977;283:65 - 76. [16] Hickman JJ, Stenger DA. Interaction of cultured neurons with defined surfaces. In: Stenger A, Mckenna TM, editors. Enabling technologies for cultured neural networks. New York7 Academic Press; 1994. p. 51 - 76. [17] Schakenraad JM, Ranter BD, Hoffman AS, Schoen FJ, Lemons JE, editors. Biomaterials science. New York7 Academic Press; 1977. p. 141 - 5.
Basic Research
Influence of nanoscale surface roughness on neural cell attachment on silicon Saida P. Khan, PhD(c), Gregory G. Auner, PhD, Golam M. Newaz, PhDT Wayne State University, Detroit, Michigan Received 2 December 2004; accepted 31 March 2005
Abstract
The adherence and viability of neural cells (primary cortical cells) from rat embryo on silicon wafers with varying surface roughness (10 to 250 nm) at the nano scale were investigated. The roughnesses were achieved by using chemical etching. Atomic force microscopy was utilized to determine surface roughness. We examined the adherence and viability of neural cells by using scanning electron microscopy and fluorescence immunoassaying. Antineuron-specific enolase antibody was used for immunostaining. Results from this investigation show that for these specific neural cells, there is an optimum surface roughness range, Ra = 20 to 100 nm that promotes cell adhesion and longevity. For silicon-based devices, this optimum surface roughness will be desirable as a suitable material/neuron interface. D 2005 Elsevier Inc. All rights reserved.
Key words:
Neuron; Nanoscale; Cell attachment; Roughness; Silicon
Design and development of neural implant device require the integration of electronics and biological systems. An electronic device can be used as a neural prosthesis in drug delivery, nerve regeneration, functional electrode stimulation, pain relief, and so on. But good incorporation of nerve cells and their viability with the device is a prerequisite for its proper function. A particular area of interest in this field of research is the study of the interface between a material surface and nerve cells in vitro. A neural implant device may have a complex design with various types of materials with different shapes and surfaces joined together. Cell adhesion and healthy growth on specific surfaces are very important for the integration and function of the device. A survey of the current literature shows that studies of cell adhesion and growth have been focused on two principal areas. The first is the application of various coatings or chemical modification of the surface, such as using polylysine, and the second is topological modification of the surface. Although chemical modification has No financial conflict of interest was reported by the authors of this paper. T Corresponding author. Mechanical Engineering Department, Wayne State University, 5050 Anthony Wayne Drive, Detroit, MI 48202. E-mail address: [email protected] (G. Newaz). 1549-9634/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.nano.2005.03.007
attained considerable success in good adhesion and directed growth of nerve cells for in vitro cases, further investigation is necessary to determine its long-term in vivo applications. On the other hand, topological surface modification facilitates studying the direct interface between the cell and the surface. Scientists have realized that topographical cues may play a crucial role in mediating cell orientation and biocompatibility [1,2]. It is well known that the surface microstructure of artificial biomedical materials has a role in cell attachment [3,4]. The effect of topological modification on various surfaces such as glass, quartz [5], metal oxides [6], and silicon (Si) [7-12] has been studied by several investigators. Silicon surfaces have been studied most extensively because their semiconductor properties make Si the most attractive material of choice for neural implants. It has been reported that porous Si enhances growth of neurons [7], and nanometer-scale pillars and wells produced by reactive iron etching can improve astrocyte adhesion on Si [9]. Fan et al [12] studied the quantitative relationship between the roughness of the Si surface and neural cell (substantia niagra) adherence. Responses of other types of neural cells and entities should also be studied. Fan et al [12] considered a specific cell type, while our investigation focused on mixed neuronal culture. In this investigation we
126
S. Khan et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 1 (2005) 125–129
Fig 1. Three-dimensional AFM topography of the surface of the Si wafer and the profile of the section. A, Si wafer with Ra = 64 nm. B, Si wafer with Ra = 18 nm.
focused on primary cortical cells from the rat embryo, which were directly inoculated onto the surface of Si wafers with 3 different average roughnesses ranging from 18 to 201 nm. No coating or chemical modification was used before cell plating. Examination of cell attachment and viability were studied using scanning electron microscopy (SEM) and neuron-specific enolase (NSE) immunoassaying. Silicon wafers The double-sided polished surface of Si (111) wafers (Institute of Electronic Materials Technology, Poland) was used for this surface roughness investigation. The wafers were cleaned using the standard method of washing with acetone and distilled water, and then blowing dry with nitrogen gas. Wafers used as controls (untreated) were rinsed with deionized water. Surface roughness of the etched wafers was controlled by variable exposure time to HF/HNO3/H2O
(1:1:10, v/v) solutions. Thus, Si wafers with 3 different surface roughnesses (with mean Ra) were prepared. In this experiment all prepared Si wafers were exposed to the air to form a stable oxide (SiO2) layer before further use. Both sides of Si wafers were sterilized via autoclaving at 1218C for 40 minutes before being inoculated with neural cells. The topography of Si wafers was measured by atomic force microscopy (AFM) (Autoprobe, Park Scientific Instruments) using the contact mode in air. By using AFM, surface topography was resolved on a nanometer scale with a scan range of 1600 1600 nm2. The Ra was evaluated using the surface profile data extracted from AFM images. Raman spectra were used to check the chemical groups on the wafer surface before and after etching. The laser raman spectrum was obtained using a Renishaw Invia Raman Microscope, at a wavelength range of 400 to 2000/cm. Cell culture Neural cells were cultured using a protocol modified from Brewer et al [13]. Rat cortical neurons were collected
Fig 2. Raman spectra of the surface of Si wafers with different roughnesses are showing no chemical change due to etching.
Fig 3. Dependence of cell viability (6 days after inoculation) on Ra of Si wafers. The cell coverages from 3 separate samples are shown for each average roughness.
S. Khan et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 1 (2005) 125–129
127
Fig 4. Scanning electron microscopy images of the primary cortical neuron culture at 6 days. A, Surface Ra = 18 nm (scale bar = 30 Am). B, Surface Ra = 64 nm (scale bar = 30 Am). C, Magnified image of a cluster of neurons on surface Ra = 64 nm (scale bar = 25 Am). D, Magnified image of a growth cone extending from a neuron (scale bar = 10 Am). E, Surface Ra = 204 nm (scale bar = 60 Am).
from the cerebral cortex of prenatal E19 Sprague Dawley rats under sterile conditions. The 1-cm2 Si samples were placed in 12 well culture dishes. Cells were inoculated on the surface of the prepared wafers. Some empty dishes were used as controls; neuron cells were directly inoculated on the culture plastic. Cells were cultured in neurobasal (GIBCO) medium supplemented with glutamine, B27, and 4% fetal bovine serum (FBS), and then incubated in a humid 5% CO2 atmosphere at 378C. The cells were counted using trypan blue in a hemocytometer on the day of inoculation. Cells were inoculated at a density of 2.28 105 cells per square centimeter on the surface of the prepared Si wafers. Wells without Si wafers and where the neuron cells were inoculated directly on the culture plastic were regarded as controls. Cell adherence and viability After 6 days, cell cultures used for SEM observation were fixed in 4% paraformaldehyde using 0.1 mol phosphate buffered saline (PBS) at 378C for 30 minutes. These samples were processed according to a standard procedure: After gradual dehydration in ethanol, they were critical-point dried and spattered with gold coating for conduction. The cultures were then observed (Hitachi S 2400, Hitachi Instruments) with an acceleration voltage of 25 keV. For immunostaining, cell cultures were fixed with 4% paraformaldehyde in 0.1 mol PBS at 378C for 30 minutes, and then rinsed 3 times in 0.1 mol PBS. Cultures were permeabilized with 0.1% Triton X-100 in
0.1 mol PBS (PBS-t) for 5 minutes at room temperature. To avoid nonspecific binding, the cells were first washed twice with 5% FBS in HankTs balanced salt solution (HBSS) for 10 minutes. Anti-rat NSE in HBSS was used on fixed cells as the primary antibody, and the cells were kept at 48C overnight. Before the secondary antibody was used, the cells were washed again in 5% FBS in HBSS for 10 minutes 3 times. A 0.02% Na azide solution was used as an antibacterial agent. As a secondary antibody, antimouse IgG fluorescein isothiocyanate (Sigma) was used, at a dilution of 1:100 for 1 hour. After immunostaining, cultures were observed with a Diaphot 300 inverted microscope (Nikon). For statistical analysis and reproducibility verification, each experiment was performed with 3 samples. Results and discussion Surface topography The surface roughness of all Si wafers was measured quantitatively by AFM. The 3 different Ra of the Si wafersT surfaces, produced by various etching conditions, were 18, 64, and 204 nm. The AFM surface topography and profiles of 2 examples of Si wafers with differing surface roughness are shown in Figure 1. The first surface of the Si wafers was produced by etching in 2:3:10 of HF/HNO3/H2O solution, with an Ra = 18 nm (Figure 1, A). When wafers were etched in 2:3:10 of HF/HNO3/H2O solution for 10 minutes, the Ra increased to 204 nm.
128
S. Khan et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 1 (2005) 125–129
Fig 5. Fluorescence images of the fluorescein isothiocyanate-labeled immunohistologically stained primary cortical neurons cultured on (A) Si wafer Ra = 64 nm, and (B) Si wafer Ra = 204 nm. The gray emission corresponds to anti-NSE immunostaining (scale bar = 15 Am).
To check for differences in the chemical status of the surface of Si wafers before any preparation and after etching in HF/HNO3/H2O solutions, raman spectroscopy was used. There were no differences among Raman spectra of Si wafers with differing surface roughness (Figure 2). Cell adherence and viability Scanning electron microscopy and immunohistochemical observations showed that the adhesion and viability of primary cortical neurons were significantly dependent on the surface roughness of the Si wafer. Counting cells was difficult because of cell clusters. Therefore, we used a different approach than Fan et al [12]. We calculated the percent coverage of the surface by the neurons at different surface roughnesses. Our observations were fairly consistent with Fan et al [12]. We found that at an average roughness of 64 nm, cells cover a greater surface area than a surface roughness of 18 nm or 204 nm. A summary of average cell density on Si wafers by surface Ra is shown in Figure 3. Cell coverages were counted in at least 3 random 0.5 0.5 mm2 fields in each of the 3 samples, and then an average was taken. It is apparent from this figure that optimum neural adhesion occurred on Si wafers with surface Ra in the neighborhood of 64 nm, in cases where surface coverage was measured 5 days after inoculation. On surfaces with Ra of less than about 60 nm, and on rough surfaces with Ra above 70 nm, cell adhesion was much lower. These results are similar to those of Fan et al [12]. Figure 4 shows SEM images of the cells at different surface roughnesses. Contact adhesion is a precondition for the neurite growth of neuronal cells. The outgrowth of axons on the modified silicon as shown in Figure 4, D indicates good adherence on the surface. Some matrix-like matter, which should be some protein excreted by the living cells adsorbed by the surface for attachment, can be seen around the cells in Figure 4, B and C. Figure 4, A and E show poor cell attachment for surfaces with roughness of 18 nm and 204 nm, respectively. Figure 5 shows the fluorescence images of stained neurons. To avoid excessive astrocyte growth, a low concentration of serum (4%) was used in the media. This
low concentration may have caused the abovementioned neuron clustering as a result of low adhesion. Serum contains various growth factors and proteins that promote cell adhesion. Therefore, common practice in neuronal culture methods is to apply 10% to 15% serum in growth media. With the plated cell density, an average surface coverage should be about 10%, while in our surfaces the range of surface coverage was 1.1% to 1.9%. It is apparent that without use of any adhesion layer (eg, polylysine), initial attachment is poor. Si and SiO2 surfaces do not have cytotoxicity to neural cells [7], but a low affinity for cell adherence was reported for B104 neuroblastoma cells on SiO2 substrate [14]. A study of vascular systems [15] showed that when an artificial system interacts with a biological system, proteins and other macromolecules adsorb to the surface. This protein layer may act as an adhesion layer between the cell and the surface. It has also been shown that surface composition and properties influence this adsorption [16]: Low-energy surfaces promote protein adsorption more than high-energy surfaces. Surface roughness is correlated to interfacial energy [17] and wettability of the surface. The surface roughness can affect the adhesion of the cell due to increased or decreased contact area, which is proportionate to the interfacial adhesive force. Therefore, surface roughness may influence the contact area of the cell membrane with the substrate. It seems that at the optimum surface roughness, the cell body contacts the surface in the most suitable manner required for proper attachment and subsequent axonal proliferation.
Conclusions Surface roughness has considerable influence on neural cell attachment on bare Si. It appears that at the nano-scale level, neuronal cells recognize a surface roughness range for optimum attachment. Neurons do not readily attach on very smooth or rough surfaces. From an average roughness of 0 to 64 nm, cell adherence increases with roughness; but at roughness of around 204 nm or more, roughness negatively affects attachment. This investigation was an initial study to determine the relative influence of roughness on cell attachment. From this study, it can be concluded that without use of proper adhesive coating, cell attachment is poor on bare Si surfaces, but this can be considerably moderated by changing surface roughness.
References [1] Clark P, Britland S, Connolly P. Growth cone guidance and neuron morphology on micropatterned laminin surface. J Cell Sci 1993;105: 203 - 12. [2] von Recum AF, den Braber ET, de Ruijter JE, Smits HTJ, Ginsel LA, Jansen JA. Quantitative analysis of cell proliferation orientation on substrata with uniform parallel surface micro-grooves. Biomaterials 1996;17:1093 - 9.
S. Khan et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 1 (2005) 125–129 [3] Curtis AG, Wilkinson CDW. Reaction of cell to topography. J Biomater Sci Poly Edu 1998;9:1313 - 29. [4] Russell TP, Kim HC. Tack—a sticky subject. Science 1999;285: 1219 - 21. [5] Rajnicek A, Britland S, McCaig C. Contact guidance of CNS neuritis on grooved quartz, neuronal age and cell type. J Cell Sci 1997;110: 2905 - 13. [6] Kawana A, Torimitsu K. Selective growth of sensory nerve fibers on mental oxide pattern in culture. Dev Brain Res 1990;51:128. [7] Bayless SC, Bucberry LD, Fletcher I, Tobin MJ. The culture of neurons on silicon. Sensor Actuators A 1999;74:139 - 42. [8] Turner JN, Shain W, Szarowski H, Anderson M, Martins M, Isaacson M, et al. Cerebral astrocyte response to micromachined silicon implants. Exp Neurol 1999;156:33 - 49. [9] Turner AMP, et al. Attachment of astroglial cells to microfabricated pillar arrays to different geometries. J Biomed Res 2000;51:4305. [10] Craighead HG, et al. Chemical and topographical patterning for directed cell attachment. Curr Opin Sol State Mat Sci 2001;5:177. [11] Fan YW, et al. Improvement of neural cell adherence to silicon surface by hydroxyl ion implantation. Surface Coating Technol 2000;131:355.
129
[12] Fan YW, Cui FZ, et al. Culture of neural cells on silicon wafers with nanoscale surface topograph. J Neurosci Methods 2002;120: 17 - 23. [13] Brewer GJ, Torricelli JR, Evege EK, Price PJ. Optimized survival of hippocampal neurons in B27-supplemented Neurobasalk, a new serum-free medium combination. J Neurosci Res 1993;35:567 - 76. [14] Corey JM, Brunette AL, Chen MS, Weyhenmeyer JA, Brewer GJ, Wheeler BC. Differentiated B104 neuroblastoma cells are a highresolution assay for micropatterned substrates. J Neurosci Methods 1997;75:91 - 7. [15] Vormman L, Adams AL, Klings M, Fischer GC, Munoz PC, Solensky R. Reactions of formed elements of blood with plasma proteins at interfaces. Ann N Y Acad Sci 1977;283:65 - 76. [16] Hickman JJ, Stenger DA. Interaction of cultured neurons with defined surfaces. In: Stenger A, Mckenna TM, editors. Enabling technologies for cultured neural networks. New York7 Academic Press; 1994. p. 51 - 76. [17] Schakenraad JM, Ranter BD, Hoffman AS, Schoen FJ, Lemons JE, editors. Biomaterials science. New York7 Academic Press; 1977. p. 141 - 5.