- Email: [email protected]
Investigation of cell–substrate interactions by focused ion beam preparation and scanning electron microscopy
Acta Biomaterialia 7 (2011) 2499–2507
Contents lists available at ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Investigation of cell–substrate interactions by focused ion beam preparation and scanning electron microscopy Andrea Friedmann ⇑, Andreas Hoess, Andreas Cismak, Andreas Heilmann Department of Biological and Macromolecular Materials, Fraunhofer Institute for Mechanics of Materials IWM, Walter-Hülse-Strasse 1, Halle 06120, Germany
a r t i c l e
i n f o
Article history: Received 3 November 2010 Received in revised form 21 January 2011 Accepted 15 February 2011 Available online 21 February 2011 Keywords: Focused ion beam Electron microscopy Cell–substrate interactions Nanoporous materials Biosensor
a b s t r a c t Cell–substrate interactions, which are an important issue in tissue engineering, have been studied using focused ion beam (FIB) milling and scanning electron microscopy (SEM). Sample cross-sections were generated at predefined positions (target preparation) to investigate the interdependency of growing cells and the substrate material. The experiments focus on two cell culturing systems, hepatocytes (HepG2) on nanoporous aluminum oxide (alumina) membranes and mouse fibroblasts (L929) and primary nerve cells on silicon chips comprised of microneedles. Cross-sections of these soft/hard hybrid systems cannot be prepared by conventional techniques like microtomy. Morphological investigations of hepatocytes growing on nanoporous alumina membranes demonstrate that there is in-growth of microvilli from the cell surface into porous membranes having pore diameters larger than 200 nm. Furthermore, for various cell cultures on microneedle arrays contact between the cells and the microneedles can be observed at high resolution. Based on FIB milled cross-sections and SEM micrographs cells which are only in contact with microneedles and cells which are penetrated by microneedles can be clearly distinguished. Target preparation of biological samples by the FIB technique especially offers the possibility of preparing not only soft materials but also hybrid samples (soft/hard materials). Followed by high resolution imaging by SEM, new insights into cell surface interactions can be obtained. Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction In biomedical research the development of new substrates for cell cultivation and tissue engineering is of great importance. Therefore, the interaction of cells or living tissues with the surfaces of artificial materials is relevant, both in research and practice [1]. These surface interactions can be used to guide the organization, growth and differentiation of cells [2]. Interactions between the material and the biosystem can be influenced by modification of the surface properties, in order to improve the biocompatibility of the substrate. Cell adherence and the formation of focal adhesion points especially have been the focus of such studies [3,4]. Usually these morphological details are observed by light microscopy or laser scanning microscopy. However, these techniques do not allow the direct visualization and observation of the interface between cells and the substrate. Other conventional high resolution microscopical methods with are scanning electron microscopy (SEM) and transmission electron microscopy (TEM), both of which require complex sample preparation and, especially for TEM, effective preparation of cross-sections . In the majority of cases cross-sections are generated with a microtome. Microtomy ⇑ Corresponding author. Tel.: +49 3455589258; fax: +49 3455589101. E-mail address: [email protected] (A. Friedmann).
or ultramicrotomy are widely used, but are unfavorable if the sample consists of a combination of soft biological and hard brittle materials, like cell cultures on ceramics or silicon. Usually the samples will be damaged during sectioning with a microtome and investigations of the interface between cells and the substrate material are no longer possible. Therefore, focused ion beam (FIB) technology, often referred to as FIB milling, offers a new means of cross-section preparation at preselected sample positions [5–7]. In a FIB workstation a focused gallium ion beam can be used for ablation and deposition of material on the surface of a specimen [8,9]. The resolution of the FIB is <10 nm. Cross-sections of a sample can be prepared with such beams. The secondary ions and electrons generated by the ion beam are also used for surface imaging, but nowadays the FIB source is incorporated into a high resolution scanning electron microscope (dual beam technology). The potential of this technique ranges from top to down structuring (etching or deposition of nanostructures) to three-dimensional (3D) tomographic characterization of complex microstructures and composite materials [10–12]. Further, FIB technology allows effective preparation of TEM lamellae on predefined areas (target preparation), a process that can be simultaneously monitored by SEM. Although FIB technology is widely used in the semiconductor industry, its usability for the preparation of solid polymer
1742-7061/$ - see front matter Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2011.02.024
2500
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
interfaces or cross-sectioning of adherent cells on scaffolds is still underestimated [12,13]. As a result of the limited availability of FIB technology so far, only a few studies of FIB processing of such materials have been carried out. In polymer material science FIB preparations are used for surface nanostructuring [12,14–16], as well as to investigate aging effects on polymer components [17]. The use of FIB for the preparation of polymeric samples is delicate, because it can induce amorphization of the sample surface, scission and/or cross-linking of polymer chains, shrinkage of the chains, and modification of the surface chemistry [12,13,18]. To suppress such damaging effects of the ion beam on polymers and biological specimens a well-defined set of processing parameters has to be considered [19,20]. The number of publications in the field of FIB milling of biological samples like cell cultures and tissue material has steadily increased over recent years. Based on the milling strategy, papers can be classified into two categories. On the one hand, FIB milling was used to prepare thin sections (lamella) of biological materials for high resolution TEM imaging [21–23]. On the other, crosssections of biological samples were prepared by FIB milling and directly investigated by SEM, using, for example, dual beam FIB/SEM instruments [5,7,21,25–29]. The latter technique was used in this study. Before ion beam milling prior sample preparation by plastic embedding, drying or freezing is required in both cases. In this study we present investigations using different FIB slicing techniques followed by high resolution imaging by SEM to observe cell–substrate interfaces. Two types of cell–material systems were selected. First, the morphology and adhesion of cells on nanoporous alumina membranes made by anodic oxidation will be described. These porous membranes can be applied for indirect coculture of different cell types on both membrane sides to investigate and stimulate cell–cell interactions. Secondly, cell cultures on microneedle chips were investigated. These microneedles are used as miniaturized patch-clamp electrodes [30,31]. Thus a detailed investigation of the coupling between or penetration by sensor needles and cultivated cells is required. Cells cultured in monolayers on nanoporous alumina or silicon microneedle arrays were prepared by conventional sample preparation for electron microscopy. The structure of the substrate surface as well as the cell morphology were imaged by SEM. Cell adhesion and cell–substrate interactions were characterized by FIB sectioning the samples at predefined positions and subsequent SEM imaging. FIB preparation was carried out using two different techniques, slice by slice preparation and pie slice preparation, to cross-section the target areas, as shown in Fig. 1.
The slice by slice preparation technique is carried out by stepwise material ablation in only one direction to create parallel cross-sections of the sample. During this stepwise process cell– substrate interactions as well as cellular focal adhesion sites can be continuously studied (Fig. 1a). With the resulting stack of images it is possible to generate a 3D reconstruction of the observed object at high resolution. In order to locate hidden objects or subsurface structures, like a microneedle below a cell, pie slice preparation can be applied (Fig. 1b). During this process the cutting direction of the ion beam is changed. With a freely selectable cutting direction specific structures inside the sample or hidden objects can be found and effectively prepared, which is obviously a significant advantage compared with conventional microtomy. In this way we introduce a technique able to obtain new insights into the growth behavior of adherent cells on different substrates. 2. Materials and methods 2.1. Preparation of and cell cultivation on nanoporous alumina membranes Self-supporting nanoporous alumina membranes were prepared using the well-known anodic oxidation of aluminum [32– 36]. The membranes exhibited parallel, open pores which were aligned perpendicular to the membrane surface and showed a narrow pore size distribution. Different electrolytes and anodization voltages were used to obtain self-supporting membranes with various pore diameters and membrane thicknesses [37,38]. In brief, electropolished aluminum plates (150 100 mm, 99.99% purity, Hydro Aluminum, Germany) were placed into electrolyte baths and connected as the anode. Anodization was carried out under constant potential conditions at 40 V (4 vol.% oxalic acid, membrane M1) and 150 V (1 vol.% phosphoric acid, membrane M2) for 24 h. Self-supporting nanoporous membranes were obtained by a stepwise voltage reduction, followed by mechanical detachment from the underlying aluminum substrate. For the cell cultivation experiments smaller membrane pieces (diameter 12.5 mm) were produced by laser cutting with an Nd-YAG laser at a wavelength of 1064 nm. The membranes were subsequently immersed in 5 vol.% H3PO4 solution for a given time, dependent on the electrolyte used during the anodization process. After cleaning in distilled water the membranes were sterilized in 70 vol.% ethanol for at least 24 h and dried under UV in a laminar flow hood. The cell culture experiments on nanoporous alumina membranes were performed with the human hepatoma cell line HepG2 (DSMZ GmbH, Germany). The HepG2 cells were cultured in RPMI
Fig. 1. Schematic diagram of the (a) slice by slice and (b) pie slice FIB preparation techniques.
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
medium containing 10 vol.% fetal calf serum (FCS), 1 vol.% penicillin/streptomycin and 2 mM L-glutamine at 37 °C in an incubator with 5% CO2. The cells were grown until confluence in 25 cm2 tissue flasks, subsequently harvested with trypsin/EDTA solution and collected by centrifugation. After resuspension in fresh medium the cell concentration was determined by cell counting using a Neubauer chamber. The cells were seeded at a density of 5 104 cells cm 2 on the different substrates placed in 24-well tissue culture plates. For growth experiments the cell number was determined after specific culture durations using the CellTiterBlueÒ cell viability assay (Promega GmbH, Germany). 2.2. Silicon microneedle arrays Silicon-based microneedle arrays have been applied as cell–sensor hybrid structures to measure the intracellular potentials of cells [29,39]. The fabrication of electrode arrays with microneedles having diameters of around 1 lm was realized by a combination of isotropic and anisotropic plasma etching steps [30,31]. The chips contained an array of 64 microneedles occupying a total area of approximately 1 mm2. After fabrication the microstructured chips were seeded with cells. Cell cultivation on silicon microneedle arrays was carried out with L929 mouse fibroblasts and primary mouse cells (glia and neurons) [40]. Penetration of the needles into cell membranes was achieved by an electroporation technique (LOMINE) [30]. 2.3. Sample preparation for electron microscopy Preparation was performed on complete alumina membranes and on microneedle chips with open housing after cell cultivation. First, the cells were fixed with 2 vol.% glutardialdehyde in phosphate-buffered saline (PBS) for 2 h at room temperature. Afterwards the samples were post-fixed and stained in PBS containing 1 vol.% osmium tetroxide for 45 min. During the preparation of HepG2 cells grown on nanoporous alumina membranes the postfixation step was omitted in order to make the samples amenable to further investigation. Then the cells were dehydrated in a series of acetone/water mixtures (10, 30, 50, 70, 90, and 100 vol.%) for 10 min each and finally dried using a critical point dryer (CPD 030, BAL-TEC, Liechtenstein). When drying the samples with the CPD prior dehydration with acetone/water mixtures is advantageous because, compared with ethanol, dilution of acetone with liquid CO2 during critical point drying is better. Finally, the alumina membranes and microneedle chips were placed on SEM specimen holders and coated with a thin conducting platinum layer by magnetron sputtering or with a thin carbon layer by vacuum evaporation. 2.4. FIB preparation and SEM investigations All FIB preparations and SEM investigations were performed with a Quanta 3D FEG dual-beam apparatus from FEI (USA). In this device the gallium ion beam and the electron beam operate independently of each other. The point of coincidence of the two beams is located at a working distance of 10 mm. The angle between both beams is 52°. To allow vertical cutting with the FIB the sample was tilted by this angle. Thus observation of the cross-sections with the electron beam was also done at an angle of 52°. After screening the samples by conventional SEM at various magnifications the cells of interest were selected for cross-sectional preparation (target preparation). Before this target preparation an additional platinum layer was deposited on selected samples on top of the platinum or carbon layer (see above) using metal–organic chemical vapor deposition (MO-CVD) with a methylcyclopentadienyltrimethyl platinum source in the FIB device
2501
(30 kV for Ga+ ion acceleration, beam current 10 pA lm 2). This protective layer was deposited over only a small area at the point of interest and was done to protect the sample surface against redeposition of ablated atoms. Further, this platinum layer mechanically stabilizes sensitive samples. A smooth and uniform surface created by the platinum layer also reduces the curtain effect. As the first step in cross-section preparation at predefined areas of interest coarse material was ablated with gallium ions accelerated at 30 kV with ion currents between 5 and 7 nA. After this coarse milling step lower beam currents of 50–300 pA at lower acceleration voltages down to 5 kV were used to polish the crosssections. The samples were not usually cooled during FIB milling. The patterning conditions used conform to the standard patterning conditions for silicon materials. The SEM observation can be done during ion milling or after subsequent milling steps. The SEM images were generated in high vacuum mode using an acceleration voltage of 5 keV and an electron beam currents of 53 pA determined with an Everhart Thornley detector.
3. Results and discussion 3.1. Cell–substrate interactions of cell cultures on nanoporous alumina membranes It has already been demonstrated that the unique structure of nanoporous alumina can be beneficial for different cell culture applications [37,41]. Indirect co-cultivation of different cell types on the nanoporous membranes especially seems to be useful to stimulate and investigate cell–cell interactions and to influence cellular functions under in vitro culture conditions. In such a setup the nanoporous membrane acts as a physical barrier between the cells. Cell–cell communication is ensured by the diffusion of molecules through the parallel, open pores and can be controlled by the porosity, pore diameter or membrane thickness. For the cell growth experiments described below two different types of nanoporous alumina membranes were used. The main difference between these substrates was the pore diameter. The pore diameters (dP), determined by image analysis of SEM micrographs, were 40 ± 11 nm for membrane M1 and 270 ± 49 nm for membrane M2. Fig. 2 shows the initial adhesion (3 h after cell seeding) and proliferation over 1 week of HepG2 cells on the two types of nanoporous alumina membranes and conventional tissue culture polystyrene as a control substrate. The cell concentration was normalized to the growth areas provided by the porous substrates and the control surface, respectively. The results show that cell adhesion was highest on the control surface, with approximately 72% of the initially seeded cell concentration (5 104 cells cm 2). In comparison, 58% of the cells adhered to membrane M2 with pore diameters of 270 nm. The lowest cell adhesion of about 48% was detected on membrane M1 with a pore size of 40 nm. Within the following days of culture the cells proliferated on all three substrates, however, between culture days 3 and 5 cell growth was somewhat delayed. This was most noticeable on the membranes with a pore diameter of 40 nm. Additionally, the lowest cell concentrations were found on this substrate at the end of the culture period. In comparison, cell proliferation on membranes with s pore diameter of 270 nm was even better than on the control surface. After 7 days cell cultivation the cell–substrate interactions were studied by FIB preparation and SEM visualization. The membranes with HepG2 cells were prepared as described above. Fig. 3 shows SEM micrographs of HepG2 cells on the nanoporous alumina membranes M1 (Fig. 3a and c) and M2 (Fig. 3b and d).
2502
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
Fig. 2. Initial cell adhesion (3 h after cell seeding) and cell proliferation of HepG2 cells grown on nanoporous alumina membranes M1 (dP = 40 ± 11 nm) and M2 (dP = 270 ± 49 nm) as well as polystyrene (PS) as a control.
Fig. 3. SEM micrographs of HepG2 cells on nanoporous alumina membranes (a, c) M1 (dP 40 nm) and (b, d) M2 (dP 270 nm). (Upper) Overviews of single HepG2 cells on the different nanoporous alumina membranes; (lower) magnification of cell borders.
The overall cell morphology on both membrane types was comparable. The cells were in close contact with the porous substrates and showed a spreading morphology. The development of numerous microvilli was visible on the cell surfaces (Fig. 3a and b). Additionally, small cell extensions (filopodia) could be observed at the cell borders. However, higher magnifications showed that the cellular interactions with the underlying membranes were different and depended mainly on the pore diameter. The small filopodia of cells on membranes with a pore diameter of 40 nm simply lay
on the surface (Fig. 3c). In contrast, filopodia and also some bent microvilli from the cell surface penetrated into the pores of larger diameter (Fig. 3d). Recent studies have revealed that this phenomenon only occurs when the pore size is 200 nm or greater [37]. This can be related to the dimensions of the cellular features, with diameters between 100 and 150 nm. Cross-sectioning of the cells by FIB and subsequent SEM investigations confirmed good cell adhesion on the porous substrates (Fig. 4).
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
2503
Fig. 4. SEM micrographs of HepG2 cells on nanoporous alumina membranes (a, c) M1 (dP 40 nm) and (b, d) M2 (dP 270 nm) after FIB preparation. (Upper) Overview of cross-sections of HepG2 cells on the nanoporous alumina membranes; (lower) section enlargements.
The cross-sections showed that the cells were connected tightly to the underlying membranes (Fig. 4a and b). No gaps between the cells and the porous membranes were found. Therefore, we suppose that the cells strongly adhere to the porous substrates. Also, there seems to be no influence of pore diameter or surface roughness on cell adhesion. However, higher magnifications of the cellular cross-sections showed that there was in-growth of cell protrusions from the cell bottom into the parallel aligned pores (Fig. 4d and insets). Because of their length, 1.5–2 lm, these structures are most likely microvilli on the cell membrane, comparable with those on the cell surface (Fig. 3). Again, this was observed only on membranes with the larger pore diameter (dP > 200 nm) and was not visible for membrane M1 with a pore diameter of approximately 40 nm (Fig. 4c). The porous, sponge-like appearance of the cellular cross-sections was probably caused by the preparation process, i.e. leaching of cytoplasmic components. Extension of the preparation procedure with further steps, as described in Leser et al. [5], may contribute to the reduction of such artifacts. In addition, a further effect of the FIB preparation, the so-called curtain effect, became visible. Changes in the sputtering yield as the ion beam passes over the interface of regions of different composition [6,42] influences the vertical ablation and sometimes a parallel pattern to the cellular cross-sections was found (see, for example, Fig. 4c and d). During the processing of polymers or biological materials especially this curtaining cannot totally be avoided. An additional platinum layer can be effective in reducing the curtain effect and protecting the sample surface from scattered ions. Fig. 4 demonstrates different protection strategies for the investigation of cell–surface interactions between HepG2 cells and nanoporous alumina. In Fig. 4a there is an additional rectangular protection layer (thickness 100 nm) all over the depicted sample area. In contrast, in Fig. 4b there is a protection layer (1 lm thick) formed as a bar across the cell. The main advantage of the rectan-
gular layer is the protection of a larger area, which is necessary when slicing whole cells with a FIB. Protection of the sample with a (thick) platinum bar alone has the advantage that the bar thickness reduces artifacts like curtaining. If necessary, the curtain can be reduced by digital image enhancement after taking the micrographs. In summary, neither type of nanoporous alumina membranes had a detrimental effect on cell viability. However, cell morphology, adhesion and proliferation seemed to be dependent on the membrane pore size. FIB preparation and SEM visualization demonstrated that cells on membrane M2 (dP 270 nm) showed ingrowth of filopodia and microvilli into the nanopores, which could be the reason for the better cell adhesion and proliferation observed. At the end of the culture period the cell concentration on these substrates was even higher than that on the control surface (see Fig. 2). Cells on membrane M1 with a pore diameter of 40 nm exhibited lower adhesion and slowed proliferation compared with the other substrates. The results support the usability of nanoporous alumina as a potential cell culture substrate. The open porosity provides nutrient supply to both cell sides, which can be favorable in culturing polar cell types, e.g. hepatocytes. The development of microvilli on the bottom of the cell on membranes with pores >200 nm is especially encouraging with respect to their use in co-culture applications. In this context SEM and FIB proved to be beneficial tools for the interpretation of cell–substrate interactions and their connection to the cellular behavior. 3.2. FIB investigations of cell cultures on silicon microneedle arrays Patch-clamp electrodes permit the detection of intracellular membrane potentials by insertion of microcapillary electrodes into living cells [43]. Patch-on-chip systems allow the parallel
2504
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
examination of a larger number of cells. Therefore, a novel cell chip design using microneedle-based electrodes has been developed [44]. The chips contain an array of 64 microneedle-based electrodes with diameters as small as 1 lm and with heights of up to 10 lm. The electrodes can be introduced into the cytoplasm of cells using local micro-invasive needle electroporation (LOMINE) [24]. LOMINE is a hybrid of the conventional electroporation method [45] and a new approach to patch-on-chip systems [46]. To test the chip devices for diverse applications different cell types, e.g. L929 fibroblasts and mouse primary cells (glia, neurons), were cultured on these electrodes. It was first necessary to prove the success of electroporation for the different cell–silicon hybrid structures. Up to now the only way of observing this has been the FIB preparation technique. In particular the slice by slice and pie slice preparation methods offer considerable advantages in localizing these electroporation events. As an example, Fig. 5a shows an SEM micrograph of a chip device with a fibroblast monolayer cell culture (L929) on a microneedle array produced by a special fabrication process [31]. As can be seen, the fibroblasts covered most of the microneedle electrodes (Fig. 5b), with no influence of the chip material on cell morphology and cell growth being observed. Due to the given layout, single microneedles can be addressed and the morphological investigations can be compared with the electroporation experiments. Based on SEM micrographs comparable with Fig. 5b, single cells were selected for cross-sectional preparation in order to investigate the microneedle–cell interactions. Figs. 6 and 7 depict SEM micrographs of two chips of different microneedle design seeded with L929 fibroblasts as well as FIB
cross-sectional preparations (slice by slice) to characterize the cell–microtip interface. Fig. 6a gives an overview of a region of interest with settled cells. Based on the layout of the microneedle chip it was supposed that the cell marked by an arrow was located very close to a microneedle tip. Fig. 6b shows this cell after FIB preparation. It is obvious that the tip was in direct contact with the cell membrane but did not perforate it. This was related to the dimensions and the blunt end of the microneedle. Additionally, different layers at the tip could be identified. The white layer indicates metallization of the needle surface by titanium tungsten (TiW) and platinum (Pt) in order to establish an electrical connection. In addition, an insulating layer consisting of silicon nitride (SiN) at the edges of the needle is visible. On top of the tip there is no insulating layer, but material which can be related to the cell membrane. In contrast, cells on chips with smaller microneedles show a completely different behavior (Fig. 7). In addition to a change in cell morphology, e.g. a lower number of microvilli (Fig. 7a), the cross-section of the investigated cell revealed penetration of the tip through the cell membrane (Fig. 7b). The cell cytoplasm was in direct contact with the metallic layer on the needle tip, and a good electrical contact could be expected. This shows that tapered microneedles can lead to easier electroporation of the cell membranes. The different cell morphologies between the cells shown in Figs. 6 and 7 could be due to various reasons. For example, an effect of tip penetration on cell morphology is possible. A more detailed study with a statistical evaluation is underway to confirm the effect of penetration on cell morphology and cell metabolism. Application of an additional protective platinum layer on top of
Fig. 5. SEM micrographs of L929 mouse fibroblasts growing on a silicon chip with a typical microneedle array. (a) Overview of the chip surface. (b) Magnified image of cells growing directly on the microneedle.
Fig. 6. L929 mouse fibroblasts on a silicon microneedle array (a) before and (b) after FIB preparation (slice by slice). Penetration of the cell (arrow) by the microneedle was not detectable however, higher magnifications reveal close contact of the microneedle with the cell as well as the functional layers of the tip (inset in b).
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
2505
Fig. 7. L929 mouse fibroblasts on a silicon microneedle array (a) before and (b) after FIB preparation (slice by slice). The microneedle perforates the cell (arrow) and is in close contact with the cytoplasm.
the cells was often omitted for efficient investigation of cell penetration on microneedle arrays. The preparation and observation focused on the microneedle–cell interface. As a further example the pie slice technique was applied to find microneedles, comparable with those shown in Fig. 7, below a confluent co-culture of glial cells and neurons from mice. In Fig. 8 different preparation steps and the final result of a FIB cross-sectional preparation can be seen. Fig. 8a shows a cross-section at the middle of a neuron and a glial cell. The thick layer on top of the cells is a platinum layer deposited by MO-CVD to protect the cellular structures during FIB processing. Up to this preparation step the microneedle was not visible. The cross-section indicates gaps at the interface between the insulation layer and the cells as well as between the glial cell and the neuron. Cell–cell contacts are visible,
especially between the different cell types. The cross-sectional investigations suggest that the neurons were not directly connected to the substrate. Since the microneedle was not exposed after the first processing step the preparation direction was changed by rotating the sample in order to prepare a pie slice cross-section (Fig. 8b and c). Fig. 8d shows a colored higher magnification of Fig. 8c. On this pie slice the total cross-section and all interfaces of the cell–silicon hybrid material become visible. Thus it was again possible to clearly identify the thin conduction layer (Pt/TiW) on the silicon tip and the insulation layer (SiN), which did not cover the tip but insulates one tip from another. Additionally, a thin part of the glial cell was in good contact with the insulation layer. The neuron was connected to the glial cell by focal adhesion points, but it did not
Fig. 8. FIB preparation of co-cultivated primary murine glia cells and neurons on a microneedle array (pie slice preparation).
2506
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
completely cover the microneedle. Hence, with the aid of the FIB pie slice preparation it becomes obvious that penetration of the microneedle through the cell membranes of this co-culture was not achieved. Based on the morphological investigations and on the prepared cross-sections the contact area between cells and microneedles on the silicon chips could be observed with high efficiency and precision. Penetrated cells and cells which were only in contact with a microneedle were clearly distinguishable. The prepared cross-sections can also give information about the inner structure of the cells, e.g. the location of organelles. Staining technologies need to be improved to produce better contrast between single cell constituents and to validate microneedle or drug modified cell function. The application of FIB to investigate biological samples has a number of advantages. These are, in brief: 1. The possibility to slice samples without prior embedding procedures. 2. The possibility to slice hybrid samples containing soft and hard materials. 3. Target preparation with precision in the micrometer range. 4. The possible to search of subsurface objects using different slicing directions. These preparation opportunities and the high resolution of the subsequent SEM investigations allow effective and comprehensive investigations, especially of soft/hard hybrid materials. To note a disadvantage, it should be mentioned that the sample or parts of the sample are destroyed after FIB/SEM investigation. However, using the slice and view feature the sample can be virtually reconstructed [22,27]. Using adequate software systems, 3D reconstruction is feasible, and will the topic of a forthcoming paper. 4. Conclusions With the help of different FIB preparation techniques combined with SEM visualization the cell–substrate interactions between cell cultures on ceramic membranes or semiconductor chips can be investigated. Compared with conventional preparation methods like microtomy, the FIB technology has advantages with respect to processing such hybrid materials. In the present study a number of new insights into cell growth and morphology could be gained. The in-growth of cellular structures like filopodia and microvilli into the pores (dP > 200 nm) of nanoporous alumina membranes was demonstrated. Additionally, this effect was associated with increased cell adhesion and, therefore, an increased cell growth rate. On the other hand, contact between microneedles and cells, as well as penetration of the tips through the cell membrane, can be demonstrated by FIB preparation. This is of great interest in describing the function of silicon microneedles as patch-on-chip electrodes. Based on these initial studies using a combination of FIB preparation and SEM further experiments are underway to obtain additional morphological information about microneedles and electroporated cells with respect to functional data on biosensors. This incorporates a 3D reconstruction of a cell placed on a microneedle tip. Acknowledgements The authors gratefully acknowledge Michel Simon and Frank Altmann for helpful discussions regarding the FIB technology. We thank Annika Thormann for producing the nanoporous alumina membranes. Gratitude is also expressed to Oliver Paul’s group at the Department of Microsystems Engineering (IMTEK) at the University of Freiburg for producing the microneedle chips and Wer-
ner Baumann’s group at the Department of Biophysics at the University of Rostock for cell cultivation on the microneedle chips. Parts of this work were supported by the Bundesministerium für Bildung und Forschung (BMBF) within the program bioMST (FKZ W3BIO033). References [1] Griffith LG, Naughton G. Tissue engineering–current challenges and expanding opportunities. Science 2002;295:1009–14. [2] Roach P, Eglin D, Rohde K, Perry CC. Modern biomaterials: a review–bulk properties and implications of surface modifications. J Mater Sci Mater Med 2007;18:1263–77. [3] Diener A, Nebe B, Lüthen F, Becker P, Beck U, Neumann HG, et al. Control of focal adhesion dynamics by material surface characteristics. Biomaterials 2005;26:383–92. [4] Chai F, Mathis N, Blanchemain N, Meunier C, Hildebrand HF. Osteoblast interaction with DLC-coated Si substrates. Acta Biomater 2008;4:1369–81. [5] Leser V, Drobne D, Pipan Z, Milani M, Tatti T. Comparison of different preparation methods of biological samples for FIB milling and SEM investigation. J Microsc 2009;233:309–19. [6] Drobne D, Milani M, Leser V, Tatti F. Surface damage induced by FIB milling and imaging of biological samples is controllable. Microsc Res Tech 2007;70:895–903. [7] Drobne D, Milani M, Zrimec A, Leser V, Zrimec MB. Electron and ion imaging of gland cells using the FIB/SEM system. J Microsc 2005;219:29–35. [8] Giannuzzi LA, Stevie FA. Introduction to focused ion beams: instrumentation, theory, techniques and practice 1st ed.. Berlin: Springer; 2005. [9] Volkert CA, Minor AM. Focused ion beam microscopy and micromaching. MRS Bull 2007;32:389–99. [10] Mayer J, Giannuzzi LA, Kamino T, Michael J. TEM sample preparation and FIBinduced damage. MRS Bull 2007;32:400–7. [11] Uchic MD, Holzer L, Inkson BJ, Principe EL, Munroe P. Three-dimensional microstructural characterization using focused ion beam tomography. MRS Bull 2007;32:408–16. [12] Moon MW, Lee DH, Sun JY, Oh KH, Vaziri A, Hutchinson JW. Wrinkled hard skins on polymers created by focused ion beam. Proc Natl Acad Sci USA 2007;104:1130–3. [13] Ionescu-Vasii LL, Wood-Adams PM, Duchesne E, l’Esperance G, Karjala T, Ansems P. Morphological analysis of highly filled propylene/ethylene copolymers. J Appl Polym Sci 2007;105:3758–72. [14] Aubry C, Trigaud T, Moliton JP, Chiron D. Polymer gratings achieved by focused ion beam. Synth Met 2002;127:307–11. [15] Pialat E, Trigaud T, Bernical V, Moliton JP. Milling of polymeric photonic crystals by focused ion beam. Mater Sci Eng C 2005;25:618–24. [16] Luchnikov V, Stamm M, Akhmadaliev C, Bischoff L, Schmidt B. Focused-ionbeam-assisted fabrication of polymer rolled-up microtubes. J Micromech Microeng 2006;16:1602–5. [17] Brunner S, Gasser P, Simmler H, Ghazi Wakili K. Investigation of multilayered aluminum-coated polymer laminates by focused ion beam (FIB) etching. Surf Coat Technol 2006;200:5908–14. [18] Beach E, Keefe M, Heeschen W, Rothe D. Cross-sectional analysis of hollow latex particles by focused ion beam-scanning electron microscopy. Polymer 2005;46:11195–7. [19] Stokes DJ, Morrissey F, Lich BH. A new approach to studying biological and soft materials using focused ion beam scanning electron microscopy (FIB SEM). J Phys Conf Ser 2006;26:50–3. [20] Stokes DJ, Vystavel T, Morrissey F. Focused ion beam (FIB) milling of electrically insulating specimens using simultaneous primary electron and ion beam irradiation. J Phys D Appl Phys 2007;40:874–7. [21] Martínez E, Engel E, López-Iglesias C, Mills CA, Planell JA, Samitier J. Focused ion beam/scanning electron microscopy characterization of cell behaviour on polymer micro/nanopatterned substrates: a study of cell–substrate interactions. Micron 2008;39:111–6. [22] Lich B. Site specific three-dimensional structural analysis in tissues and cells using automated dual beam slice and view. Microscopy Today 2007;2:26–30. [23] Obst M, Gasser P, Mavrocordatos D, Dittrich M. TEM-specimen preparation of cell/mineral interfaces by focused ion beam milling. Am Mineral 2005;90:1270–7. [24] Hayles MF, de Winter DAM, Schneijdenberg CTWM, Meeldijk JD, Luecken U, Persoon H, et al. The making of frozen-hydrated, vitreous lamellas from cells for cryo-electron microscopy. J Struct Biol 2010;172:180–90. [25] Knott G, Marchman H, Wall D, Lich B. Serial section scanning electron microscopy of adult brain tissue using focused ion beam milling. J Neurosci 2008;28:2959–64. [26] Drobne D, Milani M, Leser V, Tatti F, Zrimec A, Znidarsic N, et al. Imaging of intracellular spherical lamellar structures and tissue gross morphology by a focused ion beam/scanning electron microscope (FIB/SEM). Ultramicroscopy 2008;108:663–70. [27] Heymann JA, Hayles M, Gestmann I, Giannuzzi LA, Lich B, Subramaniam S. Sitespecific 3D imaging of cells and tissues with a dual beam microscope. J Struct Biol 2006;155:63–73. [28] Schroeder-Reiter E, Perez-Willard F, Zeile U, Wanner G. Focused ion beam (FIB) combined with high resolution scanning electron microscopy: a promising
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
[29]
[30]
[31]
[32]
[33] [34]
[35]
[36]
tool for 3D analysis of chromosome architecture. J Struct Biol 2009;165:97–106. Heilmann A, Altmann F, Cismak A, Baumann W, Lehmann M. Investigation of cell–sensor hybrid structures by focused ion beam (FIB) technology. In: MRS spring meeting San Francisco, materials research society symposia proceedings, 2006; 0983-LL03, pp. 1–6. Tautorat C, Koester PJ, Held J, Gaspar J, Ruther P, Paul O, et al. Intracellular potential measurement of adherently growing cells using microneedle arrays. San Diego: MicroTAS Conference; 2008. pp. 1777–80. Held J, Gaspar J, Koester PJ, Tautorat C, Cismak A, Heilmann A, et al. Microneedle arrays for intracellular recording applications. In: IEEE conference on micro electro mechanical systems (MEMS) Tucson, AZ; 2008. pp. 268–71. O’Sullivan JP, Wood GC. The morphology and mechanism of formation of porous anodic films on aluminium. Proc R Soc London Ser A 1970;317: 511–43. Martin CR. Nanomaterials: a membrane-based synthetic approach. Science 1994;266:1961–6. Thormann A, Teuscher N, Pfannmöller M, Rothe U, Heilmann A. Nanoporous aluminum oxide membranes for filtration and biofunctionalization. Small 2007;3:1032–40. Hanaoka TA, Heilmann A, Kroell M, Kormann HP, Sawitowski T, Schmid G, et al. Alumina membranes–templates for novel nanocomposites. Appl Organomet Chem 1998;12:367–73. Lee W, Scholz R, Nielsch K, Gösele U. A template-based electrochemical method for the synthesis of multisegmented metallic nanotubes. Angew Chem Int Ed 2005;44:6050–3.
2507
[37] Hoess A, Teuscher N, Thormann A, Aurich H, Heilmann A. Cultivation of hepatoma cell line HepG2 on nanaoporous aluminium oxide membranes. Acta Biomater 2007;3:43–50. [38] Hoess A, Staeudte A, Thormann A, Steinhart M, Heilmann A. Production of highly ordered nanoporous alumina and its application in cell cultivation. Mater Res Soc Symp Proc 2008;1093:CC04–16. [39] Hanein Y, Schabmueller CGJ, Holman G, Lücke P, Denton DD, Böhringer KF. High-aspect ratio submicrometer needles for intracellular applications. J Micromech Microeng 2003;13:91–5. [40] Koester P, Sakowski J, Baumann W, Glock HW, Gimsa J. A new exposure system for the in vitro detection of GHz field effects on neuronal networks. Bioelectrochemistry 2007;70:104–14. [41] Wolfrum B, Mourzina Y, Sommerhage F, Offenhäusser A. Suspended nanoporous membranes as interfaces for neuronal biohybrid systems. Nano Lett 2006;6:453–7. [42] Langford RM, Petford-Long AK. Preparation of transmission electron microscopy cross-section specimens using focused ion beam milling. J Vac Sci Technol 2001;19:2186–93. [43] Hamil OP, Marty A, Neher E, Sakmann B, Sigworth FJ. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflügers Arch 1981;391:85–100. [44] Held J, Gaspar J, Ruther P, Paul O. Systematic characterization of DRIE-based fabrication process of silicon microneedles. In: MRS fall meeting Boston, materials research society symposia proceedings, 2007; 1057-DD07, pp. 1–6. [45] Tsong TY. Electroporation of cell membranes. Biophys J 1991;60:297–306. [46] van Stiphout P, Knott T, Danker T, Stett A. 3D microfluidic chip for automated patch-clamping. Mikrosystemtechnik-Kongress Freiburg, 2005. pp. 435–8.
Contents lists available at ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Investigation of cell–substrate interactions by focused ion beam preparation and scanning electron microscopy Andrea Friedmann ⇑, Andreas Hoess, Andreas Cismak, Andreas Heilmann Department of Biological and Macromolecular Materials, Fraunhofer Institute for Mechanics of Materials IWM, Walter-Hülse-Strasse 1, Halle 06120, Germany
a r t i c l e
i n f o
Article history: Received 3 November 2010 Received in revised form 21 January 2011 Accepted 15 February 2011 Available online 21 February 2011 Keywords: Focused ion beam Electron microscopy Cell–substrate interactions Nanoporous materials Biosensor
a b s t r a c t Cell–substrate interactions, which are an important issue in tissue engineering, have been studied using focused ion beam (FIB) milling and scanning electron microscopy (SEM). Sample cross-sections were generated at predefined positions (target preparation) to investigate the interdependency of growing cells and the substrate material. The experiments focus on two cell culturing systems, hepatocytes (HepG2) on nanoporous aluminum oxide (alumina) membranes and mouse fibroblasts (L929) and primary nerve cells on silicon chips comprised of microneedles. Cross-sections of these soft/hard hybrid systems cannot be prepared by conventional techniques like microtomy. Morphological investigations of hepatocytes growing on nanoporous alumina membranes demonstrate that there is in-growth of microvilli from the cell surface into porous membranes having pore diameters larger than 200 nm. Furthermore, for various cell cultures on microneedle arrays contact between the cells and the microneedles can be observed at high resolution. Based on FIB milled cross-sections and SEM micrographs cells which are only in contact with microneedles and cells which are penetrated by microneedles can be clearly distinguished. Target preparation of biological samples by the FIB technique especially offers the possibility of preparing not only soft materials but also hybrid samples (soft/hard materials). Followed by high resolution imaging by SEM, new insights into cell surface interactions can be obtained. Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction In biomedical research the development of new substrates for cell cultivation and tissue engineering is of great importance. Therefore, the interaction of cells or living tissues with the surfaces of artificial materials is relevant, both in research and practice [1]. These surface interactions can be used to guide the organization, growth and differentiation of cells [2]. Interactions between the material and the biosystem can be influenced by modification of the surface properties, in order to improve the biocompatibility of the substrate. Cell adherence and the formation of focal adhesion points especially have been the focus of such studies [3,4]. Usually these morphological details are observed by light microscopy or laser scanning microscopy. However, these techniques do not allow the direct visualization and observation of the interface between cells and the substrate. Other conventional high resolution microscopical methods with are scanning electron microscopy (SEM) and transmission electron microscopy (TEM), both of which require complex sample preparation and, especially for TEM, effective preparation of cross-sections . In the majority of cases cross-sections are generated with a microtome. Microtomy ⇑ Corresponding author. Tel.: +49 3455589258; fax: +49 3455589101. E-mail address: [email protected] (A. Friedmann).
or ultramicrotomy are widely used, but are unfavorable if the sample consists of a combination of soft biological and hard brittle materials, like cell cultures on ceramics or silicon. Usually the samples will be damaged during sectioning with a microtome and investigations of the interface between cells and the substrate material are no longer possible. Therefore, focused ion beam (FIB) technology, often referred to as FIB milling, offers a new means of cross-section preparation at preselected sample positions [5–7]. In a FIB workstation a focused gallium ion beam can be used for ablation and deposition of material on the surface of a specimen [8,9]. The resolution of the FIB is <10 nm. Cross-sections of a sample can be prepared with such beams. The secondary ions and electrons generated by the ion beam are also used for surface imaging, but nowadays the FIB source is incorporated into a high resolution scanning electron microscope (dual beam technology). The potential of this technique ranges from top to down structuring (etching or deposition of nanostructures) to three-dimensional (3D) tomographic characterization of complex microstructures and composite materials [10–12]. Further, FIB technology allows effective preparation of TEM lamellae on predefined areas (target preparation), a process that can be simultaneously monitored by SEM. Although FIB technology is widely used in the semiconductor industry, its usability for the preparation of solid polymer
1742-7061/$ - see front matter Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2011.02.024
2500
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
interfaces or cross-sectioning of adherent cells on scaffolds is still underestimated [12,13]. As a result of the limited availability of FIB technology so far, only a few studies of FIB processing of such materials have been carried out. In polymer material science FIB preparations are used for surface nanostructuring [12,14–16], as well as to investigate aging effects on polymer components [17]. The use of FIB for the preparation of polymeric samples is delicate, because it can induce amorphization of the sample surface, scission and/or cross-linking of polymer chains, shrinkage of the chains, and modification of the surface chemistry [12,13,18]. To suppress such damaging effects of the ion beam on polymers and biological specimens a well-defined set of processing parameters has to be considered [19,20]. The number of publications in the field of FIB milling of biological samples like cell cultures and tissue material has steadily increased over recent years. Based on the milling strategy, papers can be classified into two categories. On the one hand, FIB milling was used to prepare thin sections (lamella) of biological materials for high resolution TEM imaging [21–23]. On the other, crosssections of biological samples were prepared by FIB milling and directly investigated by SEM, using, for example, dual beam FIB/SEM instruments [5,7,21,25–29]. The latter technique was used in this study. Before ion beam milling prior sample preparation by plastic embedding, drying or freezing is required in both cases. In this study we present investigations using different FIB slicing techniques followed by high resolution imaging by SEM to observe cell–substrate interfaces. Two types of cell–material systems were selected. First, the morphology and adhesion of cells on nanoporous alumina membranes made by anodic oxidation will be described. These porous membranes can be applied for indirect coculture of different cell types on both membrane sides to investigate and stimulate cell–cell interactions. Secondly, cell cultures on microneedle chips were investigated. These microneedles are used as miniaturized patch-clamp electrodes [30,31]. Thus a detailed investigation of the coupling between or penetration by sensor needles and cultivated cells is required. Cells cultured in monolayers on nanoporous alumina or silicon microneedle arrays were prepared by conventional sample preparation for electron microscopy. The structure of the substrate surface as well as the cell morphology were imaged by SEM. Cell adhesion and cell–substrate interactions were characterized by FIB sectioning the samples at predefined positions and subsequent SEM imaging. FIB preparation was carried out using two different techniques, slice by slice preparation and pie slice preparation, to cross-section the target areas, as shown in Fig. 1.
The slice by slice preparation technique is carried out by stepwise material ablation in only one direction to create parallel cross-sections of the sample. During this stepwise process cell– substrate interactions as well as cellular focal adhesion sites can be continuously studied (Fig. 1a). With the resulting stack of images it is possible to generate a 3D reconstruction of the observed object at high resolution. In order to locate hidden objects or subsurface structures, like a microneedle below a cell, pie slice preparation can be applied (Fig. 1b). During this process the cutting direction of the ion beam is changed. With a freely selectable cutting direction specific structures inside the sample or hidden objects can be found and effectively prepared, which is obviously a significant advantage compared with conventional microtomy. In this way we introduce a technique able to obtain new insights into the growth behavior of adherent cells on different substrates. 2. Materials and methods 2.1. Preparation of and cell cultivation on nanoporous alumina membranes Self-supporting nanoporous alumina membranes were prepared using the well-known anodic oxidation of aluminum [32– 36]. The membranes exhibited parallel, open pores which were aligned perpendicular to the membrane surface and showed a narrow pore size distribution. Different electrolytes and anodization voltages were used to obtain self-supporting membranes with various pore diameters and membrane thicknesses [37,38]. In brief, electropolished aluminum plates (150 100 mm, 99.99% purity, Hydro Aluminum, Germany) were placed into electrolyte baths and connected as the anode. Anodization was carried out under constant potential conditions at 40 V (4 vol.% oxalic acid, membrane M1) and 150 V (1 vol.% phosphoric acid, membrane M2) for 24 h. Self-supporting nanoporous membranes were obtained by a stepwise voltage reduction, followed by mechanical detachment from the underlying aluminum substrate. For the cell cultivation experiments smaller membrane pieces (diameter 12.5 mm) were produced by laser cutting with an Nd-YAG laser at a wavelength of 1064 nm. The membranes were subsequently immersed in 5 vol.% H3PO4 solution for a given time, dependent on the electrolyte used during the anodization process. After cleaning in distilled water the membranes were sterilized in 70 vol.% ethanol for at least 24 h and dried under UV in a laminar flow hood. The cell culture experiments on nanoporous alumina membranes were performed with the human hepatoma cell line HepG2 (DSMZ GmbH, Germany). The HepG2 cells were cultured in RPMI
Fig. 1. Schematic diagram of the (a) slice by slice and (b) pie slice FIB preparation techniques.
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
medium containing 10 vol.% fetal calf serum (FCS), 1 vol.% penicillin/streptomycin and 2 mM L-glutamine at 37 °C in an incubator with 5% CO2. The cells were grown until confluence in 25 cm2 tissue flasks, subsequently harvested with trypsin/EDTA solution and collected by centrifugation. After resuspension in fresh medium the cell concentration was determined by cell counting using a Neubauer chamber. The cells were seeded at a density of 5 104 cells cm 2 on the different substrates placed in 24-well tissue culture plates. For growth experiments the cell number was determined after specific culture durations using the CellTiterBlueÒ cell viability assay (Promega GmbH, Germany). 2.2. Silicon microneedle arrays Silicon-based microneedle arrays have been applied as cell–sensor hybrid structures to measure the intracellular potentials of cells [29,39]. The fabrication of electrode arrays with microneedles having diameters of around 1 lm was realized by a combination of isotropic and anisotropic plasma etching steps [30,31]. The chips contained an array of 64 microneedles occupying a total area of approximately 1 mm2. After fabrication the microstructured chips were seeded with cells. Cell cultivation on silicon microneedle arrays was carried out with L929 mouse fibroblasts and primary mouse cells (glia and neurons) [40]. Penetration of the needles into cell membranes was achieved by an electroporation technique (LOMINE) [30]. 2.3. Sample preparation for electron microscopy Preparation was performed on complete alumina membranes and on microneedle chips with open housing after cell cultivation. First, the cells were fixed with 2 vol.% glutardialdehyde in phosphate-buffered saline (PBS) for 2 h at room temperature. Afterwards the samples were post-fixed and stained in PBS containing 1 vol.% osmium tetroxide for 45 min. During the preparation of HepG2 cells grown on nanoporous alumina membranes the postfixation step was omitted in order to make the samples amenable to further investigation. Then the cells were dehydrated in a series of acetone/water mixtures (10, 30, 50, 70, 90, and 100 vol.%) for 10 min each and finally dried using a critical point dryer (CPD 030, BAL-TEC, Liechtenstein). When drying the samples with the CPD prior dehydration with acetone/water mixtures is advantageous because, compared with ethanol, dilution of acetone with liquid CO2 during critical point drying is better. Finally, the alumina membranes and microneedle chips were placed on SEM specimen holders and coated with a thin conducting platinum layer by magnetron sputtering or with a thin carbon layer by vacuum evaporation. 2.4. FIB preparation and SEM investigations All FIB preparations and SEM investigations were performed with a Quanta 3D FEG dual-beam apparatus from FEI (USA). In this device the gallium ion beam and the electron beam operate independently of each other. The point of coincidence of the two beams is located at a working distance of 10 mm. The angle between both beams is 52°. To allow vertical cutting with the FIB the sample was tilted by this angle. Thus observation of the cross-sections with the electron beam was also done at an angle of 52°. After screening the samples by conventional SEM at various magnifications the cells of interest were selected for cross-sectional preparation (target preparation). Before this target preparation an additional platinum layer was deposited on selected samples on top of the platinum or carbon layer (see above) using metal–organic chemical vapor deposition (MO-CVD) with a methylcyclopentadienyltrimethyl platinum source in the FIB device
2501
(30 kV for Ga+ ion acceleration, beam current 10 pA lm 2). This protective layer was deposited over only a small area at the point of interest and was done to protect the sample surface against redeposition of ablated atoms. Further, this platinum layer mechanically stabilizes sensitive samples. A smooth and uniform surface created by the platinum layer also reduces the curtain effect. As the first step in cross-section preparation at predefined areas of interest coarse material was ablated with gallium ions accelerated at 30 kV with ion currents between 5 and 7 nA. After this coarse milling step lower beam currents of 50–300 pA at lower acceleration voltages down to 5 kV were used to polish the crosssections. The samples were not usually cooled during FIB milling. The patterning conditions used conform to the standard patterning conditions for silicon materials. The SEM observation can be done during ion milling or after subsequent milling steps. The SEM images were generated in high vacuum mode using an acceleration voltage of 5 keV and an electron beam currents of 53 pA determined with an Everhart Thornley detector.
3. Results and discussion 3.1. Cell–substrate interactions of cell cultures on nanoporous alumina membranes It has already been demonstrated that the unique structure of nanoporous alumina can be beneficial for different cell culture applications [37,41]. Indirect co-cultivation of different cell types on the nanoporous membranes especially seems to be useful to stimulate and investigate cell–cell interactions and to influence cellular functions under in vitro culture conditions. In such a setup the nanoporous membrane acts as a physical barrier between the cells. Cell–cell communication is ensured by the diffusion of molecules through the parallel, open pores and can be controlled by the porosity, pore diameter or membrane thickness. For the cell growth experiments described below two different types of nanoporous alumina membranes were used. The main difference between these substrates was the pore diameter. The pore diameters (dP), determined by image analysis of SEM micrographs, were 40 ± 11 nm for membrane M1 and 270 ± 49 nm for membrane M2. Fig. 2 shows the initial adhesion (3 h after cell seeding) and proliferation over 1 week of HepG2 cells on the two types of nanoporous alumina membranes and conventional tissue culture polystyrene as a control substrate. The cell concentration was normalized to the growth areas provided by the porous substrates and the control surface, respectively. The results show that cell adhesion was highest on the control surface, with approximately 72% of the initially seeded cell concentration (5 104 cells cm 2). In comparison, 58% of the cells adhered to membrane M2 with pore diameters of 270 nm. The lowest cell adhesion of about 48% was detected on membrane M1 with a pore size of 40 nm. Within the following days of culture the cells proliferated on all three substrates, however, between culture days 3 and 5 cell growth was somewhat delayed. This was most noticeable on the membranes with a pore diameter of 40 nm. Additionally, the lowest cell concentrations were found on this substrate at the end of the culture period. In comparison, cell proliferation on membranes with s pore diameter of 270 nm was even better than on the control surface. After 7 days cell cultivation the cell–substrate interactions were studied by FIB preparation and SEM visualization. The membranes with HepG2 cells were prepared as described above. Fig. 3 shows SEM micrographs of HepG2 cells on the nanoporous alumina membranes M1 (Fig. 3a and c) and M2 (Fig. 3b and d).
2502
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
Fig. 2. Initial cell adhesion (3 h after cell seeding) and cell proliferation of HepG2 cells grown on nanoporous alumina membranes M1 (dP = 40 ± 11 nm) and M2 (dP = 270 ± 49 nm) as well as polystyrene (PS) as a control.
Fig. 3. SEM micrographs of HepG2 cells on nanoporous alumina membranes (a, c) M1 (dP 40 nm) and (b, d) M2 (dP 270 nm). (Upper) Overviews of single HepG2 cells on the different nanoporous alumina membranes; (lower) magnification of cell borders.
The overall cell morphology on both membrane types was comparable. The cells were in close contact with the porous substrates and showed a spreading morphology. The development of numerous microvilli was visible on the cell surfaces (Fig. 3a and b). Additionally, small cell extensions (filopodia) could be observed at the cell borders. However, higher magnifications showed that the cellular interactions with the underlying membranes were different and depended mainly on the pore diameter. The small filopodia of cells on membranes with a pore diameter of 40 nm simply lay
on the surface (Fig. 3c). In contrast, filopodia and also some bent microvilli from the cell surface penetrated into the pores of larger diameter (Fig. 3d). Recent studies have revealed that this phenomenon only occurs when the pore size is 200 nm or greater [37]. This can be related to the dimensions of the cellular features, with diameters between 100 and 150 nm. Cross-sectioning of the cells by FIB and subsequent SEM investigations confirmed good cell adhesion on the porous substrates (Fig. 4).
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
2503
Fig. 4. SEM micrographs of HepG2 cells on nanoporous alumina membranes (a, c) M1 (dP 40 nm) and (b, d) M2 (dP 270 nm) after FIB preparation. (Upper) Overview of cross-sections of HepG2 cells on the nanoporous alumina membranes; (lower) section enlargements.
The cross-sections showed that the cells were connected tightly to the underlying membranes (Fig. 4a and b). No gaps between the cells and the porous membranes were found. Therefore, we suppose that the cells strongly adhere to the porous substrates. Also, there seems to be no influence of pore diameter or surface roughness on cell adhesion. However, higher magnifications of the cellular cross-sections showed that there was in-growth of cell protrusions from the cell bottom into the parallel aligned pores (Fig. 4d and insets). Because of their length, 1.5–2 lm, these structures are most likely microvilli on the cell membrane, comparable with those on the cell surface (Fig. 3). Again, this was observed only on membranes with the larger pore diameter (dP > 200 nm) and was not visible for membrane M1 with a pore diameter of approximately 40 nm (Fig. 4c). The porous, sponge-like appearance of the cellular cross-sections was probably caused by the preparation process, i.e. leaching of cytoplasmic components. Extension of the preparation procedure with further steps, as described in Leser et al. [5], may contribute to the reduction of such artifacts. In addition, a further effect of the FIB preparation, the so-called curtain effect, became visible. Changes in the sputtering yield as the ion beam passes over the interface of regions of different composition [6,42] influences the vertical ablation and sometimes a parallel pattern to the cellular cross-sections was found (see, for example, Fig. 4c and d). During the processing of polymers or biological materials especially this curtaining cannot totally be avoided. An additional platinum layer can be effective in reducing the curtain effect and protecting the sample surface from scattered ions. Fig. 4 demonstrates different protection strategies for the investigation of cell–surface interactions between HepG2 cells and nanoporous alumina. In Fig. 4a there is an additional rectangular protection layer (thickness 100 nm) all over the depicted sample area. In contrast, in Fig. 4b there is a protection layer (1 lm thick) formed as a bar across the cell. The main advantage of the rectan-
gular layer is the protection of a larger area, which is necessary when slicing whole cells with a FIB. Protection of the sample with a (thick) platinum bar alone has the advantage that the bar thickness reduces artifacts like curtaining. If necessary, the curtain can be reduced by digital image enhancement after taking the micrographs. In summary, neither type of nanoporous alumina membranes had a detrimental effect on cell viability. However, cell morphology, adhesion and proliferation seemed to be dependent on the membrane pore size. FIB preparation and SEM visualization demonstrated that cells on membrane M2 (dP 270 nm) showed ingrowth of filopodia and microvilli into the nanopores, which could be the reason for the better cell adhesion and proliferation observed. At the end of the culture period the cell concentration on these substrates was even higher than that on the control surface (see Fig. 2). Cells on membrane M1 with a pore diameter of 40 nm exhibited lower adhesion and slowed proliferation compared with the other substrates. The results support the usability of nanoporous alumina as a potential cell culture substrate. The open porosity provides nutrient supply to both cell sides, which can be favorable in culturing polar cell types, e.g. hepatocytes. The development of microvilli on the bottom of the cell on membranes with pores >200 nm is especially encouraging with respect to their use in co-culture applications. In this context SEM and FIB proved to be beneficial tools for the interpretation of cell–substrate interactions and their connection to the cellular behavior. 3.2. FIB investigations of cell cultures on silicon microneedle arrays Patch-clamp electrodes permit the detection of intracellular membrane potentials by insertion of microcapillary electrodes into living cells [43]. Patch-on-chip systems allow the parallel
2504
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
examination of a larger number of cells. Therefore, a novel cell chip design using microneedle-based electrodes has been developed [44]. The chips contain an array of 64 microneedle-based electrodes with diameters as small as 1 lm and with heights of up to 10 lm. The electrodes can be introduced into the cytoplasm of cells using local micro-invasive needle electroporation (LOMINE) [24]. LOMINE is a hybrid of the conventional electroporation method [45] and a new approach to patch-on-chip systems [46]. To test the chip devices for diverse applications different cell types, e.g. L929 fibroblasts and mouse primary cells (glia, neurons), were cultured on these electrodes. It was first necessary to prove the success of electroporation for the different cell–silicon hybrid structures. Up to now the only way of observing this has been the FIB preparation technique. In particular the slice by slice and pie slice preparation methods offer considerable advantages in localizing these electroporation events. As an example, Fig. 5a shows an SEM micrograph of a chip device with a fibroblast monolayer cell culture (L929) on a microneedle array produced by a special fabrication process [31]. As can be seen, the fibroblasts covered most of the microneedle electrodes (Fig. 5b), with no influence of the chip material on cell morphology and cell growth being observed. Due to the given layout, single microneedles can be addressed and the morphological investigations can be compared with the electroporation experiments. Based on SEM micrographs comparable with Fig. 5b, single cells were selected for cross-sectional preparation in order to investigate the microneedle–cell interactions. Figs. 6 and 7 depict SEM micrographs of two chips of different microneedle design seeded with L929 fibroblasts as well as FIB
cross-sectional preparations (slice by slice) to characterize the cell–microtip interface. Fig. 6a gives an overview of a region of interest with settled cells. Based on the layout of the microneedle chip it was supposed that the cell marked by an arrow was located very close to a microneedle tip. Fig. 6b shows this cell after FIB preparation. It is obvious that the tip was in direct contact with the cell membrane but did not perforate it. This was related to the dimensions and the blunt end of the microneedle. Additionally, different layers at the tip could be identified. The white layer indicates metallization of the needle surface by titanium tungsten (TiW) and platinum (Pt) in order to establish an electrical connection. In addition, an insulating layer consisting of silicon nitride (SiN) at the edges of the needle is visible. On top of the tip there is no insulating layer, but material which can be related to the cell membrane. In contrast, cells on chips with smaller microneedles show a completely different behavior (Fig. 7). In addition to a change in cell morphology, e.g. a lower number of microvilli (Fig. 7a), the cross-section of the investigated cell revealed penetration of the tip through the cell membrane (Fig. 7b). The cell cytoplasm was in direct contact with the metallic layer on the needle tip, and a good electrical contact could be expected. This shows that tapered microneedles can lead to easier electroporation of the cell membranes. The different cell morphologies between the cells shown in Figs. 6 and 7 could be due to various reasons. For example, an effect of tip penetration on cell morphology is possible. A more detailed study with a statistical evaluation is underway to confirm the effect of penetration on cell morphology and cell metabolism. Application of an additional protective platinum layer on top of
Fig. 5. SEM micrographs of L929 mouse fibroblasts growing on a silicon chip with a typical microneedle array. (a) Overview of the chip surface. (b) Magnified image of cells growing directly on the microneedle.
Fig. 6. L929 mouse fibroblasts on a silicon microneedle array (a) before and (b) after FIB preparation (slice by slice). Penetration of the cell (arrow) by the microneedle was not detectable however, higher magnifications reveal close contact of the microneedle with the cell as well as the functional layers of the tip (inset in b).
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
2505
Fig. 7. L929 mouse fibroblasts on a silicon microneedle array (a) before and (b) after FIB preparation (slice by slice). The microneedle perforates the cell (arrow) and is in close contact with the cytoplasm.
the cells was often omitted for efficient investigation of cell penetration on microneedle arrays. The preparation and observation focused on the microneedle–cell interface. As a further example the pie slice technique was applied to find microneedles, comparable with those shown in Fig. 7, below a confluent co-culture of glial cells and neurons from mice. In Fig. 8 different preparation steps and the final result of a FIB cross-sectional preparation can be seen. Fig. 8a shows a cross-section at the middle of a neuron and a glial cell. The thick layer on top of the cells is a platinum layer deposited by MO-CVD to protect the cellular structures during FIB processing. Up to this preparation step the microneedle was not visible. The cross-section indicates gaps at the interface between the insulation layer and the cells as well as between the glial cell and the neuron. Cell–cell contacts are visible,
especially between the different cell types. The cross-sectional investigations suggest that the neurons were not directly connected to the substrate. Since the microneedle was not exposed after the first processing step the preparation direction was changed by rotating the sample in order to prepare a pie slice cross-section (Fig. 8b and c). Fig. 8d shows a colored higher magnification of Fig. 8c. On this pie slice the total cross-section and all interfaces of the cell–silicon hybrid material become visible. Thus it was again possible to clearly identify the thin conduction layer (Pt/TiW) on the silicon tip and the insulation layer (SiN), which did not cover the tip but insulates one tip from another. Additionally, a thin part of the glial cell was in good contact with the insulation layer. The neuron was connected to the glial cell by focal adhesion points, but it did not
Fig. 8. FIB preparation of co-cultivated primary murine glia cells and neurons on a microneedle array (pie slice preparation).
2506
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
completely cover the microneedle. Hence, with the aid of the FIB pie slice preparation it becomes obvious that penetration of the microneedle through the cell membranes of this co-culture was not achieved. Based on the morphological investigations and on the prepared cross-sections the contact area between cells and microneedles on the silicon chips could be observed with high efficiency and precision. Penetrated cells and cells which were only in contact with a microneedle were clearly distinguishable. The prepared cross-sections can also give information about the inner structure of the cells, e.g. the location of organelles. Staining technologies need to be improved to produce better contrast between single cell constituents and to validate microneedle or drug modified cell function. The application of FIB to investigate biological samples has a number of advantages. These are, in brief: 1. The possibility to slice samples without prior embedding procedures. 2. The possibility to slice hybrid samples containing soft and hard materials. 3. Target preparation with precision in the micrometer range. 4. The possible to search of subsurface objects using different slicing directions. These preparation opportunities and the high resolution of the subsequent SEM investigations allow effective and comprehensive investigations, especially of soft/hard hybrid materials. To note a disadvantage, it should be mentioned that the sample or parts of the sample are destroyed after FIB/SEM investigation. However, using the slice and view feature the sample can be virtually reconstructed [22,27]. Using adequate software systems, 3D reconstruction is feasible, and will the topic of a forthcoming paper. 4. Conclusions With the help of different FIB preparation techniques combined with SEM visualization the cell–substrate interactions between cell cultures on ceramic membranes or semiconductor chips can be investigated. Compared with conventional preparation methods like microtomy, the FIB technology has advantages with respect to processing such hybrid materials. In the present study a number of new insights into cell growth and morphology could be gained. The in-growth of cellular structures like filopodia and microvilli into the pores (dP > 200 nm) of nanoporous alumina membranes was demonstrated. Additionally, this effect was associated with increased cell adhesion and, therefore, an increased cell growth rate. On the other hand, contact between microneedles and cells, as well as penetration of the tips through the cell membrane, can be demonstrated by FIB preparation. This is of great interest in describing the function of silicon microneedles as patch-on-chip electrodes. Based on these initial studies using a combination of FIB preparation and SEM further experiments are underway to obtain additional morphological information about microneedles and electroporated cells with respect to functional data on biosensors. This incorporates a 3D reconstruction of a cell placed on a microneedle tip. Acknowledgements The authors gratefully acknowledge Michel Simon and Frank Altmann for helpful discussions regarding the FIB technology. We thank Annika Thormann for producing the nanoporous alumina membranes. Gratitude is also expressed to Oliver Paul’s group at the Department of Microsystems Engineering (IMTEK) at the University of Freiburg for producing the microneedle chips and Wer-
ner Baumann’s group at the Department of Biophysics at the University of Rostock for cell cultivation on the microneedle chips. Parts of this work were supported by the Bundesministerium für Bildung und Forschung (BMBF) within the program bioMST (FKZ W3BIO033). References [1] Griffith LG, Naughton G. Tissue engineering–current challenges and expanding opportunities. Science 2002;295:1009–14. [2] Roach P, Eglin D, Rohde K, Perry CC. Modern biomaterials: a review–bulk properties and implications of surface modifications. J Mater Sci Mater Med 2007;18:1263–77. [3] Diener A, Nebe B, Lüthen F, Becker P, Beck U, Neumann HG, et al. Control of focal adhesion dynamics by material surface characteristics. Biomaterials 2005;26:383–92. [4] Chai F, Mathis N, Blanchemain N, Meunier C, Hildebrand HF. Osteoblast interaction with DLC-coated Si substrates. Acta Biomater 2008;4:1369–81. [5] Leser V, Drobne D, Pipan Z, Milani M, Tatti T. Comparison of different preparation methods of biological samples for FIB milling and SEM investigation. J Microsc 2009;233:309–19. [6] Drobne D, Milani M, Leser V, Tatti F. Surface damage induced by FIB milling and imaging of biological samples is controllable. Microsc Res Tech 2007;70:895–903. [7] Drobne D, Milani M, Zrimec A, Leser V, Zrimec MB. Electron and ion imaging of gland cells using the FIB/SEM system. J Microsc 2005;219:29–35. [8] Giannuzzi LA, Stevie FA. Introduction to focused ion beams: instrumentation, theory, techniques and practice 1st ed.. Berlin: Springer; 2005. [9] Volkert CA, Minor AM. Focused ion beam microscopy and micromaching. MRS Bull 2007;32:389–99. [10] Mayer J, Giannuzzi LA, Kamino T, Michael J. TEM sample preparation and FIBinduced damage. MRS Bull 2007;32:400–7. [11] Uchic MD, Holzer L, Inkson BJ, Principe EL, Munroe P. Three-dimensional microstructural characterization using focused ion beam tomography. MRS Bull 2007;32:408–16. [12] Moon MW, Lee DH, Sun JY, Oh KH, Vaziri A, Hutchinson JW. Wrinkled hard skins on polymers created by focused ion beam. Proc Natl Acad Sci USA 2007;104:1130–3. [13] Ionescu-Vasii LL, Wood-Adams PM, Duchesne E, l’Esperance G, Karjala T, Ansems P. Morphological analysis of highly filled propylene/ethylene copolymers. J Appl Polym Sci 2007;105:3758–72. [14] Aubry C, Trigaud T, Moliton JP, Chiron D. Polymer gratings achieved by focused ion beam. Synth Met 2002;127:307–11. [15] Pialat E, Trigaud T, Bernical V, Moliton JP. Milling of polymeric photonic crystals by focused ion beam. Mater Sci Eng C 2005;25:618–24. [16] Luchnikov V, Stamm M, Akhmadaliev C, Bischoff L, Schmidt B. Focused-ionbeam-assisted fabrication of polymer rolled-up microtubes. J Micromech Microeng 2006;16:1602–5. [17] Brunner S, Gasser P, Simmler H, Ghazi Wakili K. Investigation of multilayered aluminum-coated polymer laminates by focused ion beam (FIB) etching. Surf Coat Technol 2006;200:5908–14. [18] Beach E, Keefe M, Heeschen W, Rothe D. Cross-sectional analysis of hollow latex particles by focused ion beam-scanning electron microscopy. Polymer 2005;46:11195–7. [19] Stokes DJ, Morrissey F, Lich BH. A new approach to studying biological and soft materials using focused ion beam scanning electron microscopy (FIB SEM). J Phys Conf Ser 2006;26:50–3. [20] Stokes DJ, Vystavel T, Morrissey F. Focused ion beam (FIB) milling of electrically insulating specimens using simultaneous primary electron and ion beam irradiation. J Phys D Appl Phys 2007;40:874–7. [21] Martínez E, Engel E, López-Iglesias C, Mills CA, Planell JA, Samitier J. Focused ion beam/scanning electron microscopy characterization of cell behaviour on polymer micro/nanopatterned substrates: a study of cell–substrate interactions. Micron 2008;39:111–6. [22] Lich B. Site specific three-dimensional structural analysis in tissues and cells using automated dual beam slice and view. Microscopy Today 2007;2:26–30. [23] Obst M, Gasser P, Mavrocordatos D, Dittrich M. TEM-specimen preparation of cell/mineral interfaces by focused ion beam milling. Am Mineral 2005;90:1270–7. [24] Hayles MF, de Winter DAM, Schneijdenberg CTWM, Meeldijk JD, Luecken U, Persoon H, et al. The making of frozen-hydrated, vitreous lamellas from cells for cryo-electron microscopy. J Struct Biol 2010;172:180–90. [25] Knott G, Marchman H, Wall D, Lich B. Serial section scanning electron microscopy of adult brain tissue using focused ion beam milling. J Neurosci 2008;28:2959–64. [26] Drobne D, Milani M, Leser V, Tatti F, Zrimec A, Znidarsic N, et al. Imaging of intracellular spherical lamellar structures and tissue gross morphology by a focused ion beam/scanning electron microscope (FIB/SEM). Ultramicroscopy 2008;108:663–70. [27] Heymann JA, Hayles M, Gestmann I, Giannuzzi LA, Lich B, Subramaniam S. Sitespecific 3D imaging of cells and tissues with a dual beam microscope. J Struct Biol 2006;155:63–73. [28] Schroeder-Reiter E, Perez-Willard F, Zeile U, Wanner G. Focused ion beam (FIB) combined with high resolution scanning electron microscopy: a promising
A. Friedmann et al. / Acta Biomaterialia 7 (2011) 2499–2507
[29]
[30]
[31]
[32]
[33] [34]
[35]
[36]
tool for 3D analysis of chromosome architecture. J Struct Biol 2009;165:97–106. Heilmann A, Altmann F, Cismak A, Baumann W, Lehmann M. Investigation of cell–sensor hybrid structures by focused ion beam (FIB) technology. In: MRS spring meeting San Francisco, materials research society symposia proceedings, 2006; 0983-LL03, pp. 1–6. Tautorat C, Koester PJ, Held J, Gaspar J, Ruther P, Paul O, et al. Intracellular potential measurement of adherently growing cells using microneedle arrays. San Diego: MicroTAS Conference; 2008. pp. 1777–80. Held J, Gaspar J, Koester PJ, Tautorat C, Cismak A, Heilmann A, et al. Microneedle arrays for intracellular recording applications. In: IEEE conference on micro electro mechanical systems (MEMS) Tucson, AZ; 2008. pp. 268–71. O’Sullivan JP, Wood GC. The morphology and mechanism of formation of porous anodic films on aluminium. Proc R Soc London Ser A 1970;317: 511–43. Martin CR. Nanomaterials: a membrane-based synthetic approach. Science 1994;266:1961–6. Thormann A, Teuscher N, Pfannmöller M, Rothe U, Heilmann A. Nanoporous aluminum oxide membranes for filtration and biofunctionalization. Small 2007;3:1032–40. Hanaoka TA, Heilmann A, Kroell M, Kormann HP, Sawitowski T, Schmid G, et al. Alumina membranes–templates for novel nanocomposites. Appl Organomet Chem 1998;12:367–73. Lee W, Scholz R, Nielsch K, Gösele U. A template-based electrochemical method for the synthesis of multisegmented metallic nanotubes. Angew Chem Int Ed 2005;44:6050–3.
2507
[37] Hoess A, Teuscher N, Thormann A, Aurich H, Heilmann A. Cultivation of hepatoma cell line HepG2 on nanaoporous aluminium oxide membranes. Acta Biomater 2007;3:43–50. [38] Hoess A, Staeudte A, Thormann A, Steinhart M, Heilmann A. Production of highly ordered nanoporous alumina and its application in cell cultivation. Mater Res Soc Symp Proc 2008;1093:CC04–16. [39] Hanein Y, Schabmueller CGJ, Holman G, Lücke P, Denton DD, Böhringer KF. High-aspect ratio submicrometer needles for intracellular applications. J Micromech Microeng 2003;13:91–5. [40] Koester P, Sakowski J, Baumann W, Glock HW, Gimsa J. A new exposure system for the in vitro detection of GHz field effects on neuronal networks. Bioelectrochemistry 2007;70:104–14. [41] Wolfrum B, Mourzina Y, Sommerhage F, Offenhäusser A. Suspended nanoporous membranes as interfaces for neuronal biohybrid systems. Nano Lett 2006;6:453–7. [42] Langford RM, Petford-Long AK. Preparation of transmission electron microscopy cross-section specimens using focused ion beam milling. J Vac Sci Technol 2001;19:2186–93. [43] Hamil OP, Marty A, Neher E, Sakmann B, Sigworth FJ. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflügers Arch 1981;391:85–100. [44] Held J, Gaspar J, Ruther P, Paul O. Systematic characterization of DRIE-based fabrication process of silicon microneedles. In: MRS fall meeting Boston, materials research society symposia proceedings, 2007; 1057-DD07, pp. 1–6. [45] Tsong TY. Electroporation of cell membranes. Biophys J 1991;60:297–306. [46] van Stiphout P, Knott T, Danker T, Stett A. 3D microfluidic chip for automated patch-clamping. Mikrosystemtechnik-Kongress Freiburg, 2005. pp. 435–8.