Introducing dip pen nanolithography as a tool for controlling stem cell behaviour: unlocking the potential of the next generation of smart materials in regenerative medicine

Abstract

Reproducible control of stem cell populations, regardless of their original source, is required for the true potential of these cells to be realised as medical therapies, cell biology research tools and in vitro assays. To date there is a lack of consistency in successful output when these cells are used in clinical trials and even simple in vitro experiments, due to cell and material variability. The successful combination of single chemistries in nanoarray format to control stem cell, or any cellular behaviour has not been previously reported. Here we report how homogenously nanopatterned chemically modified surfaces can be used to initiate a directed cellular response, particularly mesenchymal stem cell (MSC) differentiation, in a highly reproducible manner without the need for exogenous biological factors and heavily supplemented cell media. Successful acquisition of these data should lead to the optimisation of cell selective properties of materials, further enhancing the role of nanopatterned substrates in cell biology and regenerative medicine. The successful design and comparison of homogenously molecularly nanopatterned surfaces and their direct effect on human MSC adhesion and differentiation are reported in this paper. Planar gold surfaces were patterned by dip pen nanolithography (DPN®) to produce arrays of nanodots with optimised fixed diameter of 70 nanometres separated by defined spacings, ranging from 140 to 1000 nm with terminal functionalities of simple chemistries including carboxyl, amino, methyl and hydroxyl. These nanopatterned surfaces exhibited unprecedented control of initial cell interactions and subsequent control of cell phenotype and offer significant potential for the future.

---

Introduction

Within the fields of regenerative medicine and cell biology significant research endeavour is focused on using minimally invasive methods to induce a specific cellular response. One way in which this could be achieved is by controlling the direct contact and interaction between a given cell type and a well defined material; for this strategy to prove successful the material must be homogenous. Also in addition to playing a role in controlling initial cell adhesion, the stimulus presented to the cells must be sufficient to control the cellular response over a prolonged period. This can either be by controlling what factors cells release, or by ensuring the long term stability and continuation of the spatial stimulus.

As research in this field develops it is clear that an array of material and in vitro factors can be utilised to induce a cellular response i.e. roughness, topography (micro and nano), surface energy, interactions, flow and bulk chemistry. To date research into the effect of isolated single variables has been limited as material production and modification methods have lacked the control required to produce surfaces that will allow investigations into specific interactions between cells and isolated variables i.e. a precisely defined chemical modification in a defined space with control over the induced change in topography and associated changes to surface energy. Therefore due to the phenomenological nature of current studies, at the single cell scale, the responses achieved have been heterogenous at both single cell and cell population levels;^1–8 therefore tissue resulting from these studies may have inherent weakness as it is not structurally correct. The nature of cellular interaction also inhibits advances in the research area as the heterogenous response of cells is amplified between research groups due to experimental variables introduced by local protocols and practice and the cells themselves. For certain applications i.e. bone contacting applications, this phenomenological response does provide a significant enhancement of the status quo, and so simple roughening of a surface (alumina blasted titanium) produces an implantable device that provides a significant improvement to the patient i.e. a stable hip replacement. In contrast for other tissue applications i.e. neurogenic, chondrogenic and myogenic, if the definition of a complete cell population cannot be provided then the potential of stem cells in regenerative medicine may not be realised.

MSCs are by definition heterogenous and it is now widely accepted that stem cells are often made up of a number of discrete sub-populations of committed progenitor cells.^9,10 Therefore to combat this inherent level of heterogeneity and utilise materials that will act efficiently and drive stem cells in one uniform behaviour, the stimulus from the material is required to be well defined and homogenous. Ultimately the material should also display cell selective properties that can further refine the heterogenous nature of the generic MSC population. This is particularly important for stem cells, as control over the initial attachment could dictate the ultimate function of the stem cell and result in the formation of de novo tissue that is far superior to any results that are currently achievable.

One way of controlling cell adhesion and subsequent morphology is by nanotopography, and research has proven that cells can detect and respond to an array of topographies and can be affected by the level of order of an induced topography, with clear effects on cell functionality.^11–15 Cell adhesion can also be controlled via manipulation of integrin/ligand binding.^16,17 These studies proved that control over integrin spacing and clustering dictated the formation and stability of focal adhesions and dictated the levels of cellular adhesion. RGD dots separated by a distance of 58 nm induced cell spreading, whilst RGD dots separated by a distance of 110 nm resulted in weakly attached cells due to the induced stability of focal contacts caused by the initial integrin/ligand interactions. The hypothesis tested in this research was that surface modification using chemical nanoarrays can efficiently control stem cell behaviour via control of focal contact formation and distribution; furthermore that there are optimum patterns of chemical arrays that will control cell function.

Previous research has demonstrated a strong correlation between protein adsorption onto a surface, focal contact formation and subsequent MSC differentiation that is directly related to surface chemistry.^18,19 Modifying surfaces using bulk surface techniques with –CH₃, –NH₂, –OH and –COOH groups can induce stem cell differentiation,^18–20 but these results were heterogenous for the total area in terms of cell differentiation, largely due to the method of introducing the groups to the surface and the lack of control regarding spacing between groups and their distribution pattern. However, using both glass and polymer surfaces as the material substrate –CH₃ groups have been shown to enhance the MSC phenotype, whilst –NH₂ groups directed stem cells to become osteogenic i.e. bone forming. –OH and –COOH groups directed stem cells towards a cartilaginous tissue phenotype. These specific cellular effects were attributed to the chemical groups on the surfaces. The interaction between the surfaces and the cells drives the cellular response, eliminating the need for the use of additional biological factors, i.e. growth factors and cytokines or other cell culture supplements. Optimisation of these interactions could result in the development of highly defined, reproducible therapies in regenerative medicine and cell biology as a whole.

In parallel with advances in chemically defined material driven cell responses there is a wide body of emerging research evidence which has demonstrated that cells, more specifically MSC, can detect and respond to stimuli at the nanoscale i.e. nanotopography,^4,5,12 therefore the potential to combine a wide array of chemistries and defined nanoarrays is an area of research that is novel and previously unreported. Additionally the ability to control and enhance initial cell interactions to this degree using simple chemistries could result in the formation of new cost effective, stable surfaces that could offer an alternative technique for control of cell functionality. Dip pen nanolithography (DPN®) is an emerging technology that has an increasing use in the development of surfaces that can be used to investigate biological responses from the sub-micron level and upwards.^21,22 Within this study DPN® was used to produce a range of different and precisely defined large areas of chemically homogenous surface using four different chemical functionalities deposited at the nanometre scale using thiolated molecules on gold (which is an ideal flat, chemically appropriate surface to form nano-structured self-assembled monolayers (SAMs)). DPN® provides a suitable technique for the controlled deposition of a wide range of materials ranging from small organic molecules such as alkyl thiols, silazanes, Au(III) complexes, to large systems such as bio-molecules and nanoparticles by using a scanning probe microscope tip to deliver defined molecules in the form of an “ink” onto a surface via a water meniscus. The size of the feature being written is related to the interaction between the coated tip, surface, and DPN® conditions such as dwell time, temperature and humidity and these were optimised in this study. Areas of introduced modification were optimised to allow the formation of 1 focal contact on one area of modification i.e. dot size, dβ, was limited to 70–80 nm and focal contacts have been reported to be 60 nm.²³ DPN® allowed a top-down nanoscale deposition of a number of molecules that presented distinctly different chemical functionalities (–CO₂H, –NH₂, –CH₃, –OH) and as previously stated have a role in triggering MSC differentiation. Careful selection of the materials allowed deposition in the form of isolated nano-areas of SAMs onto gold surfaces via thiol linkages under low temperature and humidity conditions.

Preliminary screening experiments highlighted combinations of chemical nanoarrays that had a role in controlling cell adhesion, focal contact formation and cytoskeletal organization. In addition to highlighting the necessity to control material modifications at the nanometre scale for enhanced cell adhesive properties, this study also demonstrated the ability of specifically selected and homogenously modified surfaces to control MSC differentiation and phenotype and demonstrated a clear role for nanoarrays in the development of more efficient cell culture technologies. The study also proved that optimal spacing of chemical groups was dependent on the exact interaction induced by the chemical group, therefore enhancing the current perceptions of material mediated integrin responses.

Results and discussion

In order to investigate the effect of both the nanopattern and the terminating chemical functionality, a series of nanopatterned gold surfaces were prepared. Each nanopattern comprised a series of parallel dots spaced by a fixed distance (pitch, dα) and fixed diameter (dβ ≈ 65–70 nm) (Fig. 1). Preliminary studies tested an array of dβ values, ranging from 40–80 nm, these parameters were chosen based upon system limitations in terms of large scale production and accurate deposition of the pattern for all inks. These results showed no difference in initial focal contact formation and cellular response (data not shown) between 40 and 80 nm dot sizes, therefore to enable successful scale up and correlation between all inks a dot size of 70 nm was chosen for subsequent studies. Focal adhesions have previously been reported to be 60–67 nm in size, therefore this dot size provided for the formation of one focal contact per dot.²⁴ Spatial distribution of inks and the resultant effect on MSC adhesion were also compared using both square and hexagonal arrays. Nanopatterns using each ink were prepared using differing values for dα (140, 280, 1000 nm) with dot diameter and spacing confirmed by lateral force microscopy (LFM). These initial pitches were chosen for preliminary proof of concept studies based on published data regarding cellular response to various nanofeatures.^{11–13,16,24–29} Stability of the nanostructures under standard tissue culture conditions was established prior to cell seeding, therefore ensuring that material induced cell responses were not inhibited by degradation or desorption of the materials/molecules from the surface (Fig. 1). For this study into the use of DPN technologies in cell culture applications, the range of pitches were chosen based upon dimensions used in studies that focused on nanotopography and its role in controlling cell adhesion and focal adhesions,^15,23,25 and the workable dimensions of the DPN® systems ensuring the accurate deposition of each ink with the defined parameters. DPN® conditions were optimised to produce close to identical dot diameters for each ink.

Fig. 1 (A) Nanopatterned surfaces used for MSC control and differentiation showing dot to dot pitch (dα) and dot diameter (dβ), (B) LFM image of small area 280 nm pitch array, (C) LFM image of 140 nm pitch array. (D–F) AFM topographical image of an alkanethiol resist array fabricated on gold surface after chemical etching. The cursor profile shows an average diameter feature (dβ) of 70 nm.

Well characterised human MSCs were cultured in contact with the model surfaces and the effects of the chemical functionality and nanometre pitch and distribution on focal contact formation and cytoskeletal organisation were determined after 24 hours. Three nanopatterns were prepared for each of the four inks and the direct contact stem cell experiments repeated 4 times. Based on data received from these preliminary experiments combinations of chemistry and pitch that demonstrated a significant effect on initial cell adhesion were investigated further in terms of cellular responses over a prolonged period of time in basal conditions using a combination of qualitative (Flow Cytometry) and quantitative methods (immunohistochemistry, analysis of an array of markers associated with MSC differentiation).

Previously we have reported on the use of bulk coated –CH₃ surfaces as promoters of the MSC phenotype, a hypothesis that was supported by the results obtained in this study. Additionally results obtained also proved that combinations of chemistry and pitch can control initial cell attachment by dictating the distribution of initial focal contact formation. When cultured in contact with –CH₃ presenting surfaces hexadecane thiol (HDT) (Fig. 2) only supported viable cell attachment at a 280 nm pitch. Within the area of modification, cells attached and formed dense cell clusters with focal contacts evident throughout the bulk of the cell cluster indicative that the cell cluster was well attached to the surface. The distribution of cell contacts and differences in cell contact shape and density were dictated by the underlying nanopattern and closer observation proved that dense mature focal contacts²³ were directly associated with the nanopatterned area. Focal contacts associated with the edge of the cell cluster and the outer extremities of the area of nanopatterning were more immature, less dense and elongated in shape, and associated with individual stress fibres, associated with motile and less well attached areas of the cells (Fig. 2G). Evaluation of the distribution of F-actin throughout the cells was indicative of highly motile cells, subsequent cell migration assays proved that cells preferentially attached and moved into the –CH₃ 280 nm pitch patterned area from adjacent unmodified areas and cells that were initially attached in the modified area were highly active in terms of lamellipodia formation and contraction, well attached to the surface and did not move out of the area of modification, confirming the chemotactic properties of the surface (Fig. 2H and ESI Video S1†).

Fig. 2 (A–F) MSC cultured in contact with HDT and MHA modified surfaces for 24 hours and stained with Oregon Green Phalloidin (green-stress fibres), vinculin (red-focal contacts) and DAPI (blue-nuclei). Only 280 nm pitch modified surfaces supported viable cell adhesion. (G) High magnification image of cells clustered on ODT 280 nm pitch, arrow heads show motile focal adhesions associated with the periphery of the patterned area, whilst focal adhesions inside the cell cluster are dense and well established, depicting the strong binding of the cell via focal adhesions to the underlying nanoarray. (H) Single frame from time lapse microscopy experiment demonstrating the chemotactic nature of the modified surface. Over a 12 hour time period highlighted cells moved from the periphery of the modified area (where cell bodies were in contact with control and modified areas) into the centre of the modified area (CH₃ 280 nm pitch, inside the smaller white square) from the adjacent unmodified areas, full time lapse analysis is available in ESI Video S1†. The images collected were from low cell density experiments to allow investigation into single cell interactions with the surface, the same chemotactic properties were observed when the cell density was increased.

Material characterisation data, LFM images in Fig. 1B and C, demonstrated the change in surface features between a pitch of 280 nm and 140 nm. At 280 nm the individual features will be easily recognisable to the cells, as the outer extremities of each field of modification will be separated by at least 210 nm, whilst at a pitch of 140 nm, where pitch refers to a centre-to-centre spacing, the dots will have a separation value of just 70 nm. Stable focal contact formation relies on integrin formation and clustering, a lack of integrin clustering results in weakly attached cells, which in the case of MSC would lead to non-viable cells. Previous studies based on integrin spacing on RGD patterned surfaces have demonstrated that on ordered surfaces a spacing of approximately 70 nm is essential for determining the ability of a material to support cell adhesion.^27–29 At distances greater than and equal to 67 nm, focal contact formation is inhibited, a phenomena observed on the 140 nm pitch HDT substrates, which in turn have an end-to-end spacing of approximately 70 nm. Fitting well with previous data the surface control and definition used in this research could be used to inhibit focal contact formation via the controlled spacing of integrin clustering at a critical level. Furthermore these data were indicative that control of cell adhesion by –CH₃ presenting surfaces was similar to RGD induced cell adhesion i.e. the cells preferentially bound to the –CH₃ presenting areas, a response that was also replicated by –COOH presenting surfaces. At a pitch of 1 micron spacing between individual focal contacts would have been too great to support cell adhesion, once again indicating that when cultured in contact with –CH₃ presenting surfaces, focal contacts preferentially form on the –CH₃ areas of modification.

When cultured in contact with mercaptohexadecanoic acid (MHA), –COOH presenting nanopatterns viable cell adhesion was only observed at the 280 nm pitch. Levels of cell adhesion associated with MHA 280 nm pitch were far less than previously observed with the HDT surfaces, but the distribution of focal contacts was similar within the areas of viable cell adhesion i.e. at the centre of the area of interest there was evidence of dense mature focal contacts, whilst evidence of more motile immature focal contacts was found at the periphery of the cells, once again these focal contacts were associated directly with individual stress fibres, Fig. 2E. Closer examination of the MHA 1 µm patterned area showed evidence of very weakly attached cells, with no evidence of focal contacts formation or subsequent organised actin networks, Fig. 2D. The inability to form stable focal contacts on these surfaces ultimately leads to the lack of viable cell adhesion.

Previously we have reported on the ability of bulk coated –OH groups to trigger chondrogenic differentiation via their ability to induce minimal focal contact formation on a surface whilst maintaining viable cell adhesion. It is widely accepted that optimal chondrogenic culture conditions minimise cell contact formation to a surface, ultimately chondrocyte attachment to a substrate can lead to dedifferentiation, characterised by the loss of collagen II expression and an increase in collagen I expression. More recently it has been proven that changes in stem cell shape can optimise the potential of chondrogenic differentiation in combination with exogenous biological factors.³⁰ Results obtained in this study regarding control of focal adhesion formation were amplified when combining the introduction of –OH groups at specific pitches and arrays, Fig. 3A–C. When cultured in contact with the mercaptoundecanol (MUO), –OH presenting surfaces, viable cell adhesion was observed at all pitches, but patterns of focal contact formation and actin distribution were affected by changes in pitch, Fig. 3. At a phenomenological level these data were indicative that –OH group induced cell binding was controlled by a different mechanism than the previously discussed “RGD” hypothesis and associated integrin clustering, as cells could adhere and form actin networks at all pitches. When cultured in contact with 280 nm and 140 nm pitches actin fibre formation was evident throughout the entirety of the cell body, and the cells appeared more well spread on the 280 nm pitch, Fig. 3B. When cultured in contact with the 140 nm pitch there was evidence of concentrated areas of actin associated with the periphery of cell bodies, and areas of concentrated dense focal contact formation at distinct areas under the cell body. Overall the number of individual focal contact formations was less than previously observed on octadecanethiol (ODT) and mercaptoundecylamine (AUT) surfaces, but there was evidence of focal contact clustering, resulting in viable cell adhesion. When cultured in contact with 1 µm pitch MUO surfaces, in addition to the previously described actin budding, there was also evidence of actin ring formation at the periphery of individual cells (Fig. 3A), a response that has been previously observed when MSCs are actively undergoing chondrogenic differentiation.¹⁸ Focal contacts were reduced on the 1 µm surface compared to the 280 nm and 140 nm pitches, but there was evidence of focal contact clustering which anchored the cells to the surface, minimising parallel stress fibre formation throughout the body of the cell and supporting viable cell adhesion with a morphology representative of chondrogenic differentiation at a very early time point.

Fig. 3 (A–C) MSC cultured in contact with MUO (–OH) and with AUT (–NH₂) modified surfaces for 24 hours and stained for F-actin (green), vinculin (red) and cell nuclei (blue). When cultured in contact with 1 µm spaced MUO groups, there was evidence of rounded cell morphology and concentrated clusters of focal adhesions. Cells cultured in contact with 280 nm and 140 nm pitched surfaces showed parallel fibres of F-actin and an elongated cell morphology with minimal focal contact formation. There was also evidence of concentrated buds of actin on these surfaces. When MSCs were cultured in contact with –NH₂ modified (D–F) surfaces for 24 hours cells were elongated and aligned when cultured in contact with 280 nm pitch surfaces (E), cells in the upper right and left corners are at the outer extremities of the patterned area, therefore show signs of polarisation but are not as strong as cells in the centre of the pattern where the entirety of the cell body receives the same stimulus. Time lapse microscopy (G–H) confirmed the –NH₂ 280 nm pitch surfaces ability to induce elongation. (G) Single frame from 280 nm pitch –NH₂, all of the field of view is patterned with –NH₂, there is clear evidence of cell clustering and elongation of the well attached cell clusters, single frame taken from ESI Video S2†. When the pitch was increased by 100 nm (H) the effect on cell adhesion and morphology was significant, at a pitch of 395 nm the cells could not attach properly or elongate and resulted in rounded poorly attached cells, the area of modification is located within the four marker points, double letters contained within a square bracket, located at the corners of the field of view, single frame taken from ESI Video S3†.

When cultured in contact with AUT (amino groups) viable cell adhesion was recorded with well orientated cytoskeletons and focal contact formation at all pitches. Changing the pitch between the dots of the amino functionality controlled cell spreading and orientation (Fig. 3D–F). The formation and distribution of focal contacts on the AUT at 280 nm and 140 nm pitches were reflective of highly motile cells on the surface, for the 280 nm pitch this was accompanied by elongated cells that showed evidence of polarisation and alignment, Fig. 3E. When cultured in contact with the 1 µm AUT pitch there was a mixture of stable mature focal contacts, associated with cell nuclei and immature focal contacts associated with individual stress fibres, Fig. 3D. No evidence of cell alignment was observed on the AUT 1 µm or 140 nm pitches. Additional cell migration assays (time lapse microscopy, ESI Video S2†) confirmed the cell elongation properties of the –NH₂ 280 nm pitch surfaces (Fig. 3G) and the importance of pitch. Cell migration assays proved that cells attached well to the 280 nm AUT surface, initially cell clusters were formed but cells were highly motile on the surface and resulted in polarisation and elongation of the cell clusters (ESI Video S2†). When the pitch was increased to 395 nm (Fig. 3H and ESI Video S3†) there was a profound effect on cell adhesion and behaviour. When cultured in contact with the 395 nm array initial cell attachment in the modified area was much reduced compared to the 280 nm pitch. Cells remained rounded and were not able to spread out, there was no evidence of lamellipodia formation or contraction on the surface resulting in weakly attached cells, after a prolonged time in culture there was a minimal level of attachment that was not sufficient to support viable cell adhesion after 24 hours in vitro. Overall the phenomena observed on the 395 nm pitch surfaces followed the previously discussed “RGD” binding phenomena, where critical distances between sites of integrin clustering and subsequent focal contact formation determined the stability of the cell on the surface. The data collected within this research have proven that both chemistry and pitch play a role in determining the initial integrin binding and subsequent focal adhesion formation.

Results from the AUT modified surfaces proved that both chemistry and pitch play a role in determining the focal contact formation and distribution throughout a cell body, subsequent adhesion and spreading. The combination of control of focal contact formation with material induced cell signalling, attributed to a specific surface chemistry, provides a unique way to amplify previous material induced differentiation, via control of focal contact formation.

As stated previously the role of specific nanoarrays and their effect on MSC phenotype were evaluated further over a prolonged culture period in vitro. Novel DPN® systems were incorporated to produce homogenous nanopatterned surfaces, with no edge effects, therefore ensuring that every cell in a given test received the same spatial stimulus, minimizing uncontrolled cell surface interaction and enhancing the effect of the chemical nanoarrays on a cell response. More specifically optimum –CH₃ presenting nanoarrays were evaluated in terms of promoting/maintaining the MSC phenotype for prolonged periods of time in vitro. MSCs were cultured in contact with –CH₃ 280 nm square array surfaces for up to 28 days on modified and control surfaces. The expression of a panel of MSC markers was quantified by fluorescent activated cell sorting (FACS) analysis. To maximise the relevance of the results MSCs isolated from 4 separate donors (data provided by Lonza, UK) were used in the experiments. Results obtained after 28 days proved that –CH₃ modified surfaces maintained and enhanced the expression of CD29, CD73, CD90, CD105 and CD166 compared to unmodified substrates. Of significance was the fact that the –CH₃ modified surfaces maintained an enhanced expression of the phenotypic markers compared to traditional TCPS (tissue culture polystyrene), this was statistically significant for CD29, 73, 90 and 166. The expression of CD34 (hematopoietic stem cell marker and negative for MSC) was significantly lower on –CH₃ surfaces compared with TCPS controls, proving that chemical nanoarrays can be used to produce surfaces that can enhance current protocols used for the maintenance and purification of MSC phenotypes in vitro (Fig. 4). The maintenance of the panel of markers associated with the MSC phenotype was solely attributable to the combination of –CH₃ at a 280 nm square array, as when cultured in contact with –NH₂ 280 nm square array patterns the expression of CD29, CD73, CD90, CD105 dropped to below 70% by day 14 (data not shown), also there was a loss of STRO-1 expression from as early as day 7. Therefore whilst no distinct pattern of differentiation was observed on the –NH₂ patterned surfaces the surfaces did not maintain the MSC phenotype. Correlation of the data supported the original hypothesis that –CH₃ chemical modifications could be used to promote and maintain the MSC phenotype in vitro, combining the introduction of this group with chemical patterning resulted in surfaces that performed more efficiently than TCPS in MSC culture, in addition to being chemotactic to MSC.

Fig. 4 Human MSCs were cultured in contact with ODT (CH₃ nanoarray), plain gold and standard TCPS substrates for 28 days in basal conditions, n = 4. FACS analysis showed a significant increase in expression of CD29, CD73, CD90 and CD166 on ODT compared to standard TCPS substrates. Expression of CD34 on ODT substrates was also significantly reduced on ODT substrates compared to TCPS.

Experimental section

Materials and methods

Creation of homogenous small area nanopatterned surfaces via DPN for initial cell culture studies. DPN patterning of the surface was achieved using an NScriptor (NanoInk, Skokie, IL) system in contact mode. The AFM cantilever probes were coated by dipping the chip into a saturated solution of thiol molecules (3–5 mM). For the 1 µm pitch arrays an A-type cantilever with spring constant of 0.1 N m⁻¹ was used (NanoInk) to fabricate an area equivalent to the full range of the scanner (∼90 × 90 µm²). Optical alignment with surface marks allowed patterning over larger areas. For the smaller pitch patterns (dα = 140 and 280 nm) multiple pen arrays consisting of 24 tips (D type with a 35 µm tip to tip pitch) were used to pattern areas of 700 × 700 µm². Calibration and determination of the diffusion coefficient for each ink were carried out prior to array fabrications. The diffusion coefficient was checked again at the end of each deposition cycle to ensure that homogeneous arrays were fabricated. Dwell times varied according to ink type. Typically tip dwell times of 0.1 s (HDT and MHA) or 0.2 s (MUO and AUT) were used to produce dots of dβ = 65–70 nm. Small quality control areas were deposited on the same substrate and examined using lateral force mode (LFM) scanning. Samples were rejected prior to cell deposition if array characteristics were varied significantly or if non-specific deposition of ink occurred via airborne diffusion from the non-tip portion of the cantilever to the substrate surface during the DPN experiment.

Creation of homogenous nanopatterened surfaces for long term cell culture studies. For higher throughput and large area fabrication of nanosized features a two-dimensional pen array (2D nano PrintArray™) was used. The 2D NPA consists of 55 000 tips in a 1 cm² chip. The chip was vapour-coated under vacuum with thiol molecules under 300 motors and 80 °C for 3 hours, followed by gradual cooling to 25 °C for 6 hours. This was done to lower the melting point of the thiol molecules to facilitate the evaporation of the thiol compounds into the gas phase. This process was repeated twice to ensure homogeneous coating of the tips. The coated chip is then loaded into an NScriptor and levelled with respect to the substrate surface, thereby providing uniform contacts between the cantilevers and the surface which lead to reproducible, accurate, and homogeneous edge-to-edge patterning of nanostructures on surfaces across large areas. In order to achieve uniform patterning across the surface several DPN parameters were controlled, such as homogeneous tip coating, temperature and humidity. In addition, the substrate was heated or cooled depending on the thiol ink.³¹ The substrate temperature was controlled to control the dot sizes, several substrates were prepared with different dot sizes ranging from 50 to 200 nm. These substrates were characterized and tested, the best results were obtained for the 70 nm dot diameters. For every three substrates one was used as a quality control to ensure homogeneous edge-to-edge nanostructures fabrication with 70 nm spot diameters (Fig. 1E) using chemical wet-etching.

Cell culture. Well characterised (CD29⁺, CD73⁺, CD90⁺ and CD105⁺) Human MSCs (Lonza, UK) were introduced to the surface at a cell concentration of 5 × 10⁴ cells ml⁻¹, 1 ml of cell suspension (basal medium Lonza, UK) was added to each material in a 24 well plate. As a screening process to determine optimum combinations of chemistry and pitch cells were cultured for 24 hours at 37 °C, 5% CO₂ at which time samples were washed and fixed and stained for focal adhesions and F-actin. Selected surfaces that supported viable cell adhesion were progressed through to long term culture and analysis experiments.

Focal adhesion and F-actin staining. Samples were dual labelled with vinculin (focal adhesions) and F-actin (stress fibres) using previously described methods.¹⁸ Briefly samples were removed from culture, washed with PBS and fixed with 4% formaldehyde and 2% sucrose. After washing samples were permeabilised, 0.05% Triton ×100 solution, washed and incubated with primary mouse anti-human vinculin, 0.22 mg ml⁻¹ (Sigma UK), followed by a rhodamine-conjugated IgG goat anti-mouse secondary antibody, 10 µg ml⁻¹ (ICN). Samples were washed with a 0.1% Tween 20 solution in PBS and incubated with 0.005 mg ml⁻¹ Oregon Green Phalloidin (Invitrogen), 30 minutes at 4 °C. Samples were mounted with Vectashield (Vector) and visualised using Confocal Laser Scanning Microscopy (CLSM 510, Zeiss).

Time lapse microscopy. Human MSCs were resuspended in basal medium (Lonza, UK) and labelled with DiI (Vybrant cell-labelling, Invitrogen) according to the manufacturers guidelines. Labelled cells were introduced to the substrate in suspension and monitored over a 12 hour period using confocal laser microscopy (Zeiss). Images were taken at 60 second intervals.

Quantitative FACS analysis. Human MSCs from 4 different donors were purchased from Lonza, UK and expanded to 4P.

Cells were stained with CD29 (Cy5-PE conjugated, BD), CD90 (FITC conjugated BD), CD73 (PE conjugated, BD), CD105 (FITC conjugated, BD), CD166 (PE conjugated, BD), CD34 (FITC conjugated, BD) and CD45 (FITC conjugated, BD) prior to introduction to the test substrates to confirm the MSC phenotype of the cells. MSCs were cultured in contact with the 5 mm² test substrates surfaces in 48 well plates at an optimised concentration of 2 × 10⁴ cells per well in basal medium (Lonza, UK) for 28 days. Cells were fed twice weekly and the expression of the panel of CD markers was quantified at the end of the test period. Statistical Analysis, ANOVA Tukey models, were incorporated to establish the statistical significance of the results, p = 0.05.

Materials and abbreviations.

Conclusion

The data presented in this study provided evidence that direct functional chemical modifications can themselves control the mechanism of focal contact formation; with one of the mechanisms being through controlling specific integrin clustering. This control of initial cell signalling events can then be used to direct the cellular response i.e. controlling phenotype and function. Chemistry can control specific integrin binding; this is dependent on the type of integrin interaction and the spacings between integrin clusters, as this is critical in determining the nature of the cell adhesion and subsequent cell signalling events.

It has been proven that the combination of specific chemistry with defined positioning and density had a direct effect on cell adhesion and function. For –CH₃ and –COOH groups it was established that the cell responses followed the previously published research data for RGD nanopatterns. –NH₂ chemistries initiated different cellular responses in terms of pitch induced cellular adhesion, this effect was most notable on the –NH₂ modified surfaces for which increasing the pitch by 100 nm essentially inhibited good cell attachment and spreading on the surface, such a pronounced effect will ultimately control cell function. The weak level of attachment associated with the 395 nm pitch would limit the use of these surfaces in a number of applications i.e. osteogenic and for any surfaces that require cells to withstand a shear force.

Control is provided by controlling initial adhesion, integrin formation and clustering, focal contact formation, distribution and spreading of cells on surfaces. There are many important subtle variations in adhesion between well adhered and fully spread to non-adherent cells within which cells can be directed to the desired phenotype and function. Even significant inhibition of cell adhesion using chemistry and spacing to inhibit focal contact formation can be utilised to control cell populations, for example –OH chemistry can provide the optimal chemistry for chondrogenic and possibly adipogenic applications, –NH₂ for bone formation and possibly neurogenesis due to cellular elongation. This range of functional chemistries provide the proof that cell growth can be controlled using substrate definition by defined nanopatterns through the use of specific chemical functionalities which can be chosen for different cell phenotypes and functions for both in vitro assays and in vivo cell based therapies.

It has been demonstrated that surfaces modified using DPN® provided the required resolution and definition for these chemical functionalities. This approach was highly reproducible and provided for the control of MSC adhesion. There are potentially an almost unlimited number of inks that can be delivered to a surface using this approach, unlocking a vast array of options for the production of materials that can be used both in vivo and in vitro in cell biology, clinical diagnostics and cell therapies, by applying the production of surfaces that not only control cell function but also have the ability to act as cell selection agents. In parallel the new generation of DPN systems and the increasing availability of compatible “inks” have provided the successful scale up of materials and prolonged cell culture, unlocking the potential of these materials in all areas of cell biology and regenerative medicine.

References

1. G. Balasundaram and T. J. Webster, J. Biomed. Mater. Res., Part A, 2007, 80a, 602–611.
2. R. Barbucci, D. Pasqui, A. Wirsen, S. Affrossman, A. Curtis and C. Tetta, J. Mater. Sci., 2003, 14, 721–725.
3. J. Blümmel, N. Perschmann, D. Aydin, J. Drinjakovic, T. Surrey, M. Lopez-Garcia, H. Kessler and J. P. Spatz, Biomaterials, 2007, 28, 4739–4747.
4. A. S. Curtis, M. J. Dalby and N. Gadegaard, J. R. Soc. Interface, 2006, 3, 393–398.
5. M. J. Dalby, N. Gadegaard, R. Tare, A. Andar, M. O. Riehle, P. Herzyk, C. D. Wilkinson and R. O. Oreffo, Nat. Mater., 2007, 6, 997–1003.
6. M. J. Ernsting, R. S. Labow and J. P. Santerre, J. Biomed. Mater. Res., Part A, 2007, 83a, 759–769.
7. K. Kimura, A. Hattori, Y. Usui, K. Kitazawa, M. Naganuma, K. Kawamoto, S. Teranishi, M. Nomizu and T. Nishida, Invest. Ophthalmol. Visual Sci., 2007, 48, 1110–1118.
8. J. Salber, S. Grater, M. Harwardt, M. Hofmann, D. Klee, J. Dujic, H. Jinghuan, J. Ding, S. Kippenberger, A. Bernd, J. Groll, J. P. Spatz and M. Moller, Small, 2007, 3, 1023–1031.
9. I. Majore, P. Moretti, R. Hass and C. Kasper, Cell Communication and Signaling, 2009, 7, 6.
10. M. Kucia, R. Reca, V. R. Jala, B. Dawn, J. Ratajczak and M. Z. Ratajczak, Leukemia, 2005, 19, 1118–1127.
11. M. J. Dalby, Nanomedicine, 2009, 4, 247–248.
12. M. J. Dalby, A. Andar, A. Nag, S. Affrossman, R. Tare, S. McFarlane and R. O. Oreffo, J. R. Soc. Interface, 2008, 5, 1055–1065.
13. M. J. Dalby, M. J. Biggs, N. Gadegaard, G. Kalna, C. D. Wilkinson and A. S. Curtis, J. Cell. Biochem., 2007, 100, 326–338.
14. M. J. Dalby, N. Gadegaard, A. S. Curtis and R. O. Oreffo, Curr. Stem Cell Res. Ther., 2007, 2, 129–138.
15. M. J. Dalby, N. Gadegaard, R. Tare, A. Andar, M. O. Riehle, P. Herzyk, C. D. Wilkinson and R. O. Oreffo, Nat. Mater., 2007, 6, 997–1003.
16. E. A. Cavalcanti-Adam, D. Aydin, V. C. Hirschfeld-Warneken and J. P. Spatz, HFSP J., 2008, 2, 276–285.
17. E. A. Cavalcanti-Adam, T. Volberg, A. Micoulet, H. Kessler, B. Geiger and J. P. Spatz, Biophys. J., 2007, 92, 2964–2974.
18. J. M. Curran, R. Chen and J. A. Hunt, Biomaterials, 2005, 26, 7057–7067.
19. J. M. Curran, R. Chen and J. A. Hunt, Biomaterials, 2006, 27, 4783–4793.
20. Y. J. Ren, H. Zhang, H. Huang, X. M. Wang, Z. Y. Zhou, F. Z. Cui and Y. H. An, Biomaterials, 2009, 30, 1036–1044.
21. D. K. Hoover, E. J. Lee, E. W. Chan and M. N. Yousaf, ChemBioChem, 2007, 8, 1920–1923.
22. K. Salaita, Y. Wang and C. A. Mirkin, Nat. Nanotechnol., 2007, 2, 145–155.
23. B. Wehrle-Haller and B. Imhof, Trends Cell Biol., 2002, 12, 382–389.
24. K. Poole, K. Khairy, J. Friedrichs, C. Franz, D. A. Cisneros, J. Howard and D. Mueller, J. Mol. Biol., 2005, 349, 380–386.
25. E. A. Cavalcanti-Adam, A. Micoulet, J. Blummel, J. Auernheimer, H. Kessler and J. P. Spatz, Eur. J. Cell Biol., 2006, 85, 219–224.
26. E. A. Cavalcanti-Adam, I. M. Shapiro, R. J. Composto, E. J. Macarak and C. S. Adams, J. Bone Miner. Res., 2002, 17, 2130–2140.
27. J. Huang, S. V. Grater, F. Corbellini, S. Rinck, E. Bock, R. Kemkemer, H. Kessler, J. Ding and J. P. Spatz, Nano Lett., 2009, 9, 1111–1116.
28. C. Selhuber-Unkel, T. Erdmann, M. Lopez-Garcia, H. Kessler, U. S. Schwarz and J. P. Spatz, Biophys. J., 2010, 98, 543–551.
29. C. Selhuber-Unkel, M. Lopez-Garcia, H. Kessler and J. P. Spatz, Biophys. J., 2008, 95, 5424–5431.
30. L. Gao, R. McBeath and C. S. Chen, Stem Cells, 2010, 28(3), 564–572.
31. R. G. Sanedrin, N. A. Amro, J. Rendlen and M. Nelson, Temperature controlled dip-pen nanolithography, Nanotechnology, 21(11) DOI:10.1088/0957–4484/21/11/115302.

---

Footnote

† Electronic supplementary information (ESI) available: Time lapse microscopy videos taken when cells were cultured in contact with; 1. –CH₃ patterned surfaces, 2. and 3. –NH₂ patterned surfaces. See DOI: 10.1039/c004149a

---