Acta Biomaterialia 34 (2016) 53–59

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier.com/locate/actabiomat

Full length article Peptide patterning for controlling and monitoring cell growth q ⇑ Edith Lin, Adhirath Sikand, Jessica Wickware, Yubin Hao, Ratmir Derda

Department of Chemistry, University of Alberta, Edmonton, Alberta T6G 2G2, Canada article info abstract

Article history: The fate of cells is influenced by their microenvironment and many cell types undergo differentiation Received 22 July 2015 when stimulated by extracellular cues, such as soluble growth factors and the insoluble extracellular Received in revised form 9 January 2016 matrix (ECM). Stimulating differentiation by insoluble or ‘‘immobilized” cues is a particularly attractive Accepted 20 January 2016 method because it allows for the induction of differentiation in a spatially-defined cohort of cells within a Available online 21 January 2016 larger subpopulation. To improve the design of de novo screening of such insoluble factors, we describe a methodology for producing high-density peptide microarrays suitable for extended and flu- Keywords: orescence microscopy. As a model, we used a murine mammary gland cell line (NMuMG) that undergoes Peptide array epithelial to mesenchymal transition (EMT) in response to soluble transforming growth factor beta (TGF- DNA microarray printer b b b Peptides ) and surface-immobilized peptides that target TGF- receptors (TGF RI/II). We repurposed a well- Self-assembled monolayers established DNA microarray printing technique to produce arrays of micropatterned surfaces that dis- Epithelial to mesenchymal transition played TGFbRI/II-binding peptides and integrin binding peptides. Upon long-term culture on these arrays, Phenotypic assays only NMuMG cells residing on EMT-stimulating areas exhibited growth arrest and decreased E-cadherin expression. We believe that the methodology created in this report will aid the development of peptide- decorated surfaces that can locally stimulate defined cell surface receptors and control EMT and other well-characterized differentiation events.

Statement of Significance

Scope of work: This manuscript aims to accelerate the development of instructive biomaterials decorated with specific ligands that target cell-surface receptors and induce specific differentiation of cells upon con- tact. These materials can be used for practical applications, such as fabricating synthetic materials for large scale, stem cell culture, or investigating differentiation and asymmetric division in stem cells. Specifically, in this manuscript, we repurposed a DNA microarray printer to produce microarrays of peptide-terminated self-assembled monolayers (SAMs). To demonstrate the utility of these arrays in phenotypic assays with mammalian cells, we monitored the induction of epithelial to mesenchymal transition (EMT) in murine mammary epithelial cells using specific peptide ligands printed on these arrays. Novelty: We, and others, have published several strategies for producing peptide-based arrays suitable for long-term phenotypic assays. Many reports relied on patterning steps that made adaptation difficult. The use of a DNA microarray printer as the sole production tool simplified the production of peptide microarrays and increased the throughput of this technology. We confirmed that simplification in production did not compromise the performance of the array; it is still possible to study short-term adhesion, long-term growth, and complex phenotypic responses, such as EMT, in the cells. EMT was studied using immunofluorescent staining after four days of culture. Impact: This methodology will serve as a foundation for future screening of instructive biomaterials in our research group. As DNA printers are broadly available in academic institutions, we fore- see rapid adaptation of this approach by academic researchers. Ó 2016 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

q Part of the High Throughput Approaches to Screening Biomaterials Special Issue, edited by Kristopher Kilian and Prabhas Moghe. ⇑ Corresponding author. E-mail address: [email protected] (R. Derda). http://dx.doi.org/10.1016/j.actbio.2016.01.028 1742-7061/Ó 2016 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. 54 E. Lin et al. / Acta Biomaterialia 34 (2016) 53–59

1. Introduction

In this report, we developed a platform based on an array of self-assembled monolayers (SAMs) of peptide alkanethiols (ATs) on gold, that permits the investigation of physiological responses in cells cultured on diverse ligands that bind cell surfaces recep- tors. As a model system, we used a well-established cell line that is known to undergo epithelial to mesenchymal transition (EMT) in response to extracellular ligand stimulation. EMT is an essential process that was first recognized in organ formation during embry- onic development. This process is characterized by a shift in cell morphology from squamous, epithelial-like cells, to motile, spindle-like mesenchymal cells [1,2]. It promotes migration and increased invasiveness of cancer cells; carcinoma cells that undergo EMT become more motile and give rise to metastases [1–3]. The signaling pathway that drives EMT also controls the sig- naling that causes the formation and maintenance of highly Fig. 1. (A) Schematic of E-cadherin down regulation and epithelial to mesenchymal tumorigenic cells, known as cancer stem cells (CSCs) [4–6]. CSCs, transition (EMT) in NMuMG cells when cultured on a microarray of self-assembled by definition, are a subpopulation of cells within a tumor that have monolayers (SAMs) of peptides. EMT is only induced in cells grown on areas of a higher tumor forming capacity in xenograft models; CSCs were LTGKNFPMFHRN and not MHRMPSFLPTTL. (B) Design of microarray pattern: 9 Â 9 l also proposed to be the cells that form metastases after EMT has spots in 4 quadrants. The spots printed for all cell based experiments are 150 min diameter, printed with Telechem SMP3 Stealth Micro Spotting Pins at 0 ms contact occurred in vivo [3]. The non-stem and cancer stem cells intercon- time. (C) Customized stainless steel holder for gold coated slides. The holder is the vert reversibly via a process closely allied to EMT; this process is size of a microscope slide. (D) Spots with diameters of 150 lm and 400 lm were stochastic and only a small percentage of cancer cells undergo printed with Telechem SMP3 Stealth Micro Spotting Pins and SNS15 Stealth Solid EMT-like events that lead to de-differentiation to putative CSCs Pins, respectively. [7,8]. To aid the investigation of EMT and its connection to differen tiation/dedifferentiation events in a tumor, we developed a patterned areas to accommodate single or multiple cells. Although method that would permit the stimulation of EMT in cells in a spa- the technology for surface patterning has developed significantly tially localized manner. in the past 20 years, many techniques have been perfected to only The most characterized extracellular stimulus that initiates EMT yield high-fidelity binary patterns; modern technologies for pat- and enrichment of CSC-like cells, in multiple cancer cell lines, is terning multi-component surfaces relies on cumbersome serial transforming growth factor beta (TGF-b) [9]. The TGF-b signaling processing, accurate alignment, and access to instrumentation pathway is initiated by the dimerization of TGF-b serine/threonine not accessible outside of specialized laboratories [21]. The most kinase receptors (TGFbRI/II); this complex then phosphorylates scalable techniques for multi-component patterning of surfaces intracellular Smad, leading to the translocation of phospho-Smad are those that repurpose well-developed, high-throughput screen- into the nucleus [1,10]. In vivo, TGF-b acts as both a soluble ligand ing instruments, such as robotic spotters or microarray printers. and an extracellular matrix (ECM)-immobilized ligand [11]; immo- While these technologies have been applied to the fabrication of bilization of TGF-b on polymeric scaffolds in vitro is known to pre- arrays [22–24], their application to the fabrication of pep- serve its bioactivity and ability to elicit signaling in cells. Localized tide monolayers is limited [25,26]. immobilization of is a powerful tool in cell [12], In this report, we describe the use of a DNA microarray printer but further minimization of proteins to functional short peptides to pattern gold coated surfaces with micron-scaled islands of pep- b can help generalize this technology to bypass issues such as tides that bind to TFG RI/II and integrin receptors. The peptides are residue-specific immobilization and inactivation of proteins. conjugated to an alkanethiol linker for the creation of SAMs on the Specifically, Kiessling et al. previously demonstrated that EMT can gold-coated surfaces. In the future, this method can be used be induced in large scale populations of mammary epithelial cells to investigate the heterogeneity of EMT-responses and CSC- cultured on top of surface immobilized peptide analogs of TGF-b. dedifferentiation events in carcinoma cell lines in vitro (Fig. 1). The cells displayed signatures of EMT, such as down-regulation Contrary to the popular method of using soft lithography for print- and internalization of E-cadherin, increase in smooth muscle ing spatially defined SAMs, a DNA microarray printer provides an expression, changes in cell morphology, and growth arrest [2]. alternative, simple, robust, and reproducible method of generating arrays containing diverse surfaces. The size of the spots can be The use of chemically-modified surfaces for investigating fun- l damental questions dates back to 1978, when Folkman tuned within a range of 150–400 m in diameter to create arrays for monitoring the growth of single or multiple cells (Fig 1D). In et al. used surfaces of different hydrophobicity to show that the l area to which a cell spreads determines its survival [13]. In the this report, we primarily use areas of 150 m in diameter to 1990’s and early 2000’s, ‘‘binary” micropatterning of SAMs into accommodate multiple cells for the study of cell adhesion, spread- areas that promoted and suppressed attachment of cells made it ing, proliferation, and EMT in spatially defined areas in one com- possible to probe aspects of cellular adhesion [14], migration mon media. [15], and division [16]. In the 2000’s, several groups reported tech- nologies for co-patterning multiple chemically-defined SAMs for 2. Results and discussion co-culturing different cell lines [17]; these arrays of SAMs empow- ered the discovery of peptide-modified substrates that control dif- 2.1. Validating E-cadherin expression in NMuMG cells grown on large ferentiation of human embryonic stem cells [18]. Patterning surfaces of peptide SAMs technologies can dissect spatially-defined phenotypic transforma- tion in individual cells, such as branching morphogenesis [19],or We selected three peptides for the production of SAMs and asymmetric division [20]; technologies for this application should microarrays of SAMs with NMuMG cells. Peptides LTGKNFPMFHRN allow for the control of chemical composition, size, and shape of and MHRMPSFLPTTL were previously reported to cluster TFG-bRI/II E. Lin et al. / Acta Biomaterialia 34 (2016) 53–59 55 and potentiate the effects of soluble TFG-b at low concentrations the range of spot sizes that can be generated on the arrays, we [18]. The peptide sequence GRGDS, which binds to cell integrin printed spots with a diameter of 400 lm using SNS15 Stealth Solid receptors, was included as a negative control. As a substrate for Pins at 0 ms contact time (Fig. 1D). Shortly after printing, we SAMs, we used coverslips sputtered with a layer of translucent immersed the arrays into the solution of glucamine-alkanethiol gold to facilitate the analysis of these arrays via phase-contrast to form a protein- and cell-repelling background between the spots and fluorescent/confocal microscopy. NMuMG cells were moni- [27]. The best printing results were achieved on freshly-deposited tored for the following hallmarks of EMT: decreases in E- gold-coated surfaces (Supplementary Fig. S3). We attempted to cadherin levels, increases in smooth muscle actin levels, growth reuse gold-coated surfaces by stripping off the SAMs with a deter- arrest, and morphological changes. Gold exhibited no deleterious gent and plasma cleaning; upon reprinting the arrays, we noticed effect on NMuMG cells and cells cultured on surfaces displaying that the alignment of the printed spots highly varied on reused sur- GRGDS SAMs exhibited no changes in the basal level of E- faces. In addition to printing peptide-AT, we also printed rows of cadherin expression. KCN solution to produce a convenient visual reference of etched We validated that the level of E-cadherin expression in NMuMG gold (Fig. 2A) [28,29]. cells cultured on large areas SAMs displaying LTGKNFPMFHRN and We spiked the solutions of peptide-alkanethiols used for print- MHRMPSFLPTTL changed significantly when compared to cells cul- ing with fluorescein to visualize the locations of the printed areas tured on GRGDS. After four days of culture, E-cadherin levels, visu- before seeding cells (e.g. Fig. 2B). We observed minor variations in alized via immunofluorescent staining, were observed to be the the intensity of fluorescence, but these variations should not be highest in NMuMG cells grown on GRGDS and the lowest in interpreted as non-uniformity in SAMs, as printing deposits an NMuMG cells grown on LTGKNFPMFHRN (Supplementary excess amount of alkanethiols for the formation of SAMs. To con- Fig. S1). We also monitored these surfaces daily using phase- firm the functional uniformity of the monolayers, we produced contrast microscopy to check for morphological changes (Supple- three independent arrays and allowed for short term adhesion of mentary Fig. S2A). After four days, the cells cultured on SAMs of NMuMG cells (24 h) to minimize any variability from long-term LTGKNFPMFHRN and MHRMPSFLPTTL, and in media treated with growth (Supplementary Fig. S4). Counting the number of cells soluble TFG-b lost epithelial morphology and gained a spindle- per spot yielded only minor variations in the number of cells like appearance (Supplementary Fig. S2B). adhered to each peptide spot in three independent arrays of 36 replicates on the same array. 2.2. Optimization of printing SAMs 2.3. Optimization of cell concentrations for short-term growth For all microarray experiments, we used a pattern consisting of 4 identical quadrants of 9 Â 9 pattern (Fig. 1B). To integrate com- The optimal cell concentration for short-term growth experi- mercially available, small footprint microscope cover slides ments (0–5 days) should initially yield 1–5 cells adhering per area (18 Â 18 mm) with the DNA microarray spotter, we used custom- to allow cells to divide in the areas during the subsequent growth made stainless steel slide holders for printing (Fig. 1C). In all cell- period. We optimized conditions for NMuMG cells by seeding a based experiments in this manuscript, we used Telechem SMP3 suspension of 10,000, 25,000, or 50,000 cells in 3 mL to an array Stealth Micro Spotting Pins at 0 ms contact time to produce in a 6- well plate. After one hour of incubation, we rinsed off the peptide-modified spots of 150 lm in diameter. To demonstrate non-bound cells and supplemented the arrays with fresh NMuMG

Fig. 2. (A) Modified design of microarray pattern for spotting LTGKNFPMFHRN, MHRMPSFLPTTL, and KCN. (B) Representative image of microarray immediately after printing. Printing solutions were supplemented with fluorescein to allow for visualization of the printed areas. (C) Notched boxplot generated by OriginLab software of the number of cells observed per spot after 24 hours when seeded with 10,000, 25,000, and 50,000 cells/well. 56 E. Lin et al. / Acta Biomaterialia 34 (2016) 53–59

Fig. 3. Overview of cells growing on individual peptide islands. (A) Tile scan of entire array and (B) Magnified view of the cells on individual peptide areas (C) By day 4, NMuMG cells have down-regulated the expression of E-cadherin on areas modified with peptides LTGKNFPMFHRN and MHRMPSFLPTTL, but not on areas modified with GRGDS.

media. Fig. 2C describes the number of cells on the arrays after 24 h and MHRMPSFLPTTL, whereas cells growing on GRGDS did not of culture. We found the optimal density of cells for short-term show any decrease in E-cadherin expression (Fig. 3). growth to be 3,300 cells/mL, or 10,000 cells/well (Fig. 2C).

2.4. E-cadherin expression of cells on printed SAMs 2.5. Monitoring growth rate and cell morphology

We cultured NMuMG cells on microarrays that displayed To detect growth arrest, we counted the number of cells grow- LTGKNFPMFHRN, MHRMPSFLPTTL, and GRGDS to identify condi- ing on individual peptide-modified areas after two, three, and four tions that induce EMT in cells in spatially defined areas. We antic- days in 35–40 replicates. We modified the original 9 Â 9 pattern to ipated EMT to be induced in cells grown on islands of SAMs with an asymmetrical pattern with distinct spaces for patterning LTGKNFPMFHRN and MHRMPSFLPTTL, but not on parallel islands LTGKNFPMFHRN, MHRMPSFLPTTL, GRGDS, and KCN (Fig. 2A). This of GRGDS. For one to three days, cells exhibited high levels of E- allowed for easier visual identification of peptide islands (Fig 2B). cadherin on all three peptides. On the fourth day, we detected On the second day, NMuMG cells adhered to only 50% of decreased levels of E-cadherin in cells growing on LTGKNFPMFHRN LTGKNFPMFHRN or MHRMPSFLPTTL modified areas, whereas cells E. Lin et al. / Acta Biomaterialia 34 (2016) 53–59 57

Fig. 4. Cells cultured on arrays of peptides identified by phage-display panning on MBA-MB-231 cells (A) An array displaying five phage-derived peptides and three control peptides was seeded with 10,000 MDA-MB-231-GFP cells. The cells were allowed to grow for 3 days. The central image represents the tile scan of the entire array; the intensity of greyscale is proportional to GFP fluorescence. Insets describe individual array elements at 20Â magnification. (B) An array identical to that in (A) was seeded with 10,000 NMuMG cells; cells were allowed to grow for 3 days, then fixed and stained with E-cadherin and DAPI. The intensity of greyscale is proportional to E-cadherin fluorescence. The images represent selected areas of 3 peptide zones. Insets describe individual array elements at 20Â magnification. adhered to 75% of the GRGDS modified areas. On the fourth day, test whether the identified peptides support adhesion of normal dense populations of cells were found on over 60% of GRGDS mod- epithelial cells, and whether adhesion of cells to these peptides ified areas. In contrast, both LTGKNFPMFHRN and MHRMPSFLPTTL can induce EMT. Here, we show that peptide arrays can be used areas contained biphasic populations; dense populations were as a medium-throughput phenotypic screening method for validat- observed in approximately two-thirds of the areas, and sparse pop- ing peptides discovered by phage display. We previously identified ulations were observed on one third of the areas (Supplementary the peptides STASYTR, GKPMPPM, AMSSRSL, IPAPLRS, and Fig. S5). HAIYPRH by phage-display library selection against intact MDA- Changes in cell morphology were difficult to deconvolute due to MB-231 basal cells. Cellulose peptide arrays, created the combination of both growth and packing of cells in an area of using the SPOT peptide synthesis method, validated the ability of limited size. NMuMG cells growing on GRGDS had a significantly these peptides to support short-term adhesion of MDA-MB-231 smaller spreading area by day 4 than cells growing on GFP cells [33]. When printed on the SAMs-array, all peptides sup- LTGKNFPMFHRN and MHRMPSFLPTTL (Supplementary Fig. S6). ported short-term adhesion of MDA-MB-231 cells as well as their For all three peptides, the size of individual cells doubled over growth over the course of three days (Fig. 4A). NMuMG cells the course of four days; cells spread from an average area of adhered to four of the five peptides. As these peptides bind to 1988 lm2 on day 2 to 3745 lm2 on day 4. Visual inspection one of the receptors on mesenchymal-like cancer cells, we investi- showed that NMuMG cells grown on GRGDS islands yielded dense gated whether any of these peptides can stimulate EMT in epithe- colonies resulting from rapid cell growth. Thus, the average area of lial cells. a cell on a constrained growth island was small. Cell colonies were Over the course of 4 days, we observed a decrease in E-cadherin less dense on LTGKNFPMFHRN and MHRMPSFLPTTL islands due to expression in NMuMG cells on surfaces that displayed IPAPLRS- reduced growth and division, yielding, on average, larger cells that and GKPMPPM-terminated SAMs to levels comparable to those spread through the entire area. The observed reciprocal relation- observed on control TGFbR-binding peptides (LTGKNFPMFHRN ship between cell number and cell area arises because cells grow and MHRMPSFLPTTL). In contrast, the level of immunostaining and spread on the constant area, S. Individual cells are capable of with E-cadherin antibody in cells growing on HAIYPRH and spreading over the entire area, exhibiting cells spreading an area AMSSRSL modified areas remained similar to that observed on cells as large as S. As cells divide to produce N daughter cells, the areas cultured atop areas of GRGDS (Fig. 4B). of each individual cell is reduced to S/N. We also inspected the increase of smooth muscle actin (SMA) expression in NMuMG cells grown on peptide arrays displaying b 2.6. Culturing MDA-MB-231 and NMuMG cells on peptides found TGF RI/II binding, phage derived, and integrin binding peptides through phage display for 4 days. We observed the highest SMA expression in cells cul- tured on LTGKNFPMFHRN, but no increased expression of SMA In vitro display technologies, such as phage-[30], bacteria-[31] on surfaces modified with IPAPLRS when compared to cells grown and cell-surface displays, are powerful for de novo discovery of on areas modified with GRGDS. ligands that target cell surface receptors and control differentiation of cells [32]. We previously developed a series of peptides that bind 3. Conclusion to MDA-MB-231 cells [33] and further demonstrated that many of these peptides can support adhesion of closely and distally related In phenotypic cell-based assays, the performance of microarray- epithelial tumor cell lines (manuscript in review). We sought to printed peptide arrays is similar to the performance of previously 58 E. Lin et al. / Acta Biomaterialia 34 (2016) 53–59 reported peptide arrays [15,31]. These arrays integrate seamlessly by dissolving NaOH and KCN in milli-Q water to yield a 1 mM of with standard cell culture, imaging, and immunofluorescence tech- KCN in 1 M NaOH solution. Warning: this solution is extremely nologies. Regular arrangement of peptide areas and visual refer- toxic and corrosive. The three solutions were printed onto the gold ences, such as a grid of etched, gold-free zones, facilitate both surfaces using the DNA microarray printer in the Microarray and automatic analysis and daily manual monitoring of the assays. Proteomics Facility (MAF) at the University of Alberta. Telechem We believe that the major benefit of the microarray peptide SAMs SMP3 pins from Arrayit and 0 ms contact time were used to deposit described in this report, when compared to previous reports, is the solutions onto the surface. simplicity in fabrication and scalability in production. Specifically, After printing, the surface of the gold slide was placed on top of it is possible to scale up both the number of different peptides a solution of glucamine-AT to form a cell-repellent layer in printed on the surface and the number of arrays printed in a single between the printed areas. Specifically, 30 lL of the solution was session. Most DNA microarray printers are designed for high-speed deposited in a sterile Petri dish and the array was inverted onto production of up to a hundred arrays at once. One of the remaining the glucamine solution, forming a uniform layer of liquid sand- limitations is the synthesis of peptide-alkanethiols (AT); every wiched in between the array and the surface of the Petri dish. A peptide-AT must be purified and characterized independently. wet napkin was placed on the lid of the Petri dish and the lid Although the direct synthesis of peptides on the surface could was sealed with Parafilm to prevent evaporation of the potentially yield larger number of peptides and bypass the need glucamine-AT solution. This blocking process was allowed to pro- for purification, there is a clear compromise between the number ceed for 12–24 hours at room temperature. of peptides and quality of the array. In our experience, even an Prior to cell-adhesion studies, the arrays were sterilized with optimized method of synthesizing short peptides on a surface 95% ethanol three times for 20 min. The arrays were transferred can yield products with less than 50% purity [34]. Another limita- to a sterile 6-well plate in a laminar flow hood and rinsed 3 times tion is the analysis of the arrays; this problem can be avoided by with 1Â PBS and 3 times with MEM to eliminate traces of ethanol. using specialized, high-content analysis hardware and advanced image recognition software. An alternative solution is imaging of the arrays on a fluorescent gel or DNA array scanner. In conclusion, 4.4. Cell culture conditions of NMuMG cells we believe that this printed SAM-array will be an important tool in validation of small-molecule and peptide ligands that induce dif- NMuMG cells were maintained in DMEM supplemented with TM ferentiation and phenotypic transformations in cells. 10% FBS, non-essential amino acids, GlutaMax , and 10 ug/mL of human insulin. For cell adhesion and growth experiments on pep- tide arrays, the cells were detached from the culture surface with 4. Materials and methods trypsin (5 min), rinsed, and resuspended with growth medium, and seeded onto the arrays in six well plates. The arrays were incu- 4.1. Synthesis of peptides on an alkanethiol linker bated at 37 °C with 5% CO2 for 2 h before the media was replaced with growth media supplemented with 1 ng/mL of soluble TGF-b. The alkanethiol linker was synthesized and conjugated to poly- The cells were cultured for 4 days at 37 °C with 5% CO2. styrene monomethoxytrityl chloride (MMT-Cl) resin in preparation for Fmoc solid phase peptide synthesis. The peptides LTGKNFPMFHRN, MHRMPSFLPTTL, STASYTR, GKPMPPM, AMSSRSL, 4.5. Immunofluorescent staining IPAPLRS, and HAIYPRH were synthesized on AT-modified resin to make peptide-conjugated alkanethiols as described in our previous Cells on the gold surfaces were fixed with 2% paraformaldehyde publication [19,29]. Amino acids were purchased from Chempep. (PFA) in PBS and stored at 4 °C for a minimum of 30 min. PFA was The peptides were purified using high performance liquid chro- aspirated from the wells and the cell membrane was permeabilized matography (HPLC) and verified through matrix-assisted des- using 0.1% Triton X-100 in PBS for 15 min. The arrays were incu- orption/ionization (MALDI) and liquid chromatography mass bated in a blocking buffer (10% calf bovine serum [CBS] and 10% spectrometry (LCMS). Peptides were stored at À20 °C as lyopho- Triton X-100 in PBS) for 1 h. For E-cadherin staining, the arrays lized powder and as 1 mM solutions in Milli-Q water. were incubated with monoclonal rat anti-uvomorulin/E-cadherin (Sigma), diluted 1:200 in PBS, for 24 h at 4 °C. The samples were rinsed 3 Â 5 min with PBS and incubated with Alexa Fluor 488 con- 4.2. Making and storing gold chips jugated rabbit anti-rat (Invitrogen) at 1:500 dilution for 1 h at room temperature. For smooth muscle actin (SMA) staining, the 99.99% pure gold (Gold Maple Leaf pure bullion coin) was evap- arrays were incubated with rabbit anti-alpha smooth muscle actin orated onto 18 Â 18 mm glass cover slides using a thermal evapo- (Abcam), diluted to 5 mg/mL, for 24 h at 4 °C. The samples were ration system from Torr International Inc., model THEUPG, to then rinsed 3 Â 5 min with PBS and incubated with Alexa Fluor create the platform for self assembled monolayers. Immediately 488 Donkey Anti Rabbit IgG (Invitrogen) at 1:500 dilution for 1 h before deposition, the glass cover slides were soaked in piranha at room temperature. All arrays were then rinsed 3 Â 5 min with solution (1 H2O2:3 H2SO4) for 20 min, extensively rinsed with 1Â PBS and incubated with DAPI (1 mg/mL in 1Â PBS) for 30 min milli-Q H2O, and dried with argon gas. Warning: Piranha solution at room temperature to stain the cells’ nuclei. The arrays were is extremely corrosive and potentially explosive when in contact imaged using a confocal fluorescence microscope (Zeiss LSM 700) with organic matter. 5 nm of chromium was first evaporated onto at 20Â objective. the glass surfaces, followed by 20 nm of gold. The gold surfaces were stored in a clean polystyrene Petri dish for up to 3 weeks. 4.6. Area analysis and cell count 4.3. Microarray printing of peptides and KCN onto gold surfaces Zen software was used for image processing. Cell peripheries All peptides used for printing were dissolved in Milli-Q water at were traced to define the area to which each cell spread. The cell 1 mM concentration. We supplemented each solution with 0.2 mM count was performed by manually counting the number of nuclei fluorescein; this concentration was necessary to visualize the loca- stained with DAPI dye, per peptide area. Multi-layered, over popu- tion of each spot. The KCN solution used for etching gold was made lated peptide spots were excluded from . E. Lin et al. / Acta Biomaterialia 34 (2016) 53–59 59

Boxplot graphs were created using Origin 8.6 software (Origi- [13] J. Folkman, A. Moscona, Role of cell-shape in growth-control, Nature 273 nLab, Northampton, MA, USA), which show relative increase of cell (1978) 345–349. [14] M. Mrksich, L.E. Dike, J. Tien, D.E. Ingber, G.M. Whitesides, Using microcontact number or area over time. printing to pattern the attachment of mammalian cells to self-assembled monolayers of alkanethiolates on transparent films of gold and silver, Exp. Cell Acknowledgments Res. 235 (1997) 305–313. [15] X. Jiang, D.A. Bruzewicz, A.P. Wong, M. Piel, G.M. Whitesides, Directing cell migration with asymmetric micropatterns, Proc. Natl. Acad. Sci. U.S.A. 102 This work was supported by the Canadian Institute of Health (2005) 975–978. Research (CIHR) Bridge Funding. E.L. acknowledges Alberta Inno- [16] M. Théry, V. Racine, A. Pépin, M. Piel, Y. Chen, J.B. Sibarita, M. Bornens, The extracellular matrix guides the orientation of the cell division axis, Nat. Cell vates Health Solutions (AIHS) for the Summer Studentship; A.S. Biol. 7 (2005) 947–953. acknowledges Mitacs Globalink Research Internship. J.W. acknowl- [17] W.S. Dillmore, M.N. Yousaf, M. Mrksich, A photochemical method for edges the Undergraduate Research Initiative (URI) for the summer patterning the immobilization of ligands and cells to self-assembled monolayers, Langmuir 20 (2004) 7223–7231. studentship. Infrastructure support was provided by Canadian [18] L.Y. Li, J.R. Klim, R. Derda, A.H. Courtney, L.L. Kiessling, Spatial control of cell Foundation for Innovation (CFI). Microarray printing was provided fate using synthetic surfaces to potentiate TGF-beta signaling, Proc. Natl. Acad. by Troy Locke in the Microarray and Proteomics Facility (MAF) at Sci. U.S.A. 108 (2011) 11745–11750. the University of Alberta. [19] C.M. Nelson, M.M. VanDuijn, J.L. Inman, D.A. Fletcher, M.J. Bissell, Tissue geometry determines sites of mammary branching morphogenesis in organotypic cultures, Science 314 (2006) 298–300. Appendix A. Supplementary data [20] S.J. Habib, B.C. Chen, F.C. Tsai, K. Anastassiadis, T. Meyer, E. Betzig, R. Nusse, A localized Wnt signal orients asymmetric stem cell division in vitro, Science 339 (2013) 1445–1448. Supplementary data associated with this article can be found, in [21] S. Ankam, B.K. Tea, M. Kukumberg, E.K. Yim, High throughput screening to the online version, at http://dx.doi.org/10.1016/j.actbio.2016.01. investigate the interaction of stem cells with their extracellular microenvironment, Organogenesis 9 (2013) 128–142. 028. [22] C.J. Flaim, D. Teng, S. Chien, S.N. Bhatia, Combinatorial signaling microenvironments for studying stem cell fate, Stem Cells Dev. 17 (2008) References 29–39. [23] A. Ranga, S. Gobaa, Y. Okawa, K. Mosiewicz, A. Negro, M.P. Lutolf, 3D niche microarrays for systems-level analyses of cell fate, Nat. Commun. 5 (2014) [1] R. Kalluri, R.A. Weinberg, The basics of epithelial-mesenchymal transition, J. 4324. Clin. Invest. 119 (2009) 1420–1428. [24] G. MacBeath, S.L. Schreiber, Printing proteins as microarrays for high- [2] J. Yang, R.A. Weinberg, Epithelial-mesenchymal transition: at the crossroads of throughput function determination, Science 289 (2000) 1760–1763. development and tumor metastasis, Cell Press: Dev. Cell 14 (2008) 818–829. [25] J.R. Falsey, M. Renil, S. Park, S. Li, K.S. Lam, Peptide and small molecule [3] C. Scheel, R.A. Weinberg, Cancer stem cells and epithelial-mesenchymal microarray for high throughput cell adhesion and functional assays, transition: concepts and molecular links, Semin. Cancer Biol. 22 (2012) 396– Bioconjugate Chem. 12 (2001) 346–353. 403. [26] S. Li, N. Marthandan, D. Bowerman, H.R. Garner, T. Kodadek, Photolithographic [4] P.B. Gupta, C.L. Chaffer, R.A. Weinberg, Cancer stem cells: mirage or reality?, synthesis of cyclic peptide arrays using a differential deprotection strategy, Nat Med. 15 (2009) 1010–1012. Chem. Commun. (Camb.) 5 (2005) 581–583. [5] M.J. Meyer, J.M. Fleming, M.A. Ali, M.W. Pesesky, E. Ginsburg, B.K. Vonderhaar, [27] B.P. Orner, R. Derda, R.L. Lewis, J.A. Thomson, L.L. Kiessling, Arrays for the Dynamic regulation of CD24 and the invasive, CD44posCD24neg phenotype in combinatorial exploration of cell adhesion, J. Am. Chem. Soc. 126 (2004) breast cancer cell lines, Breast Cancer Res. 11 (2009) R82. 10808–10809. [6] A.P. Morel, M. Lièvre, C. Thomas, G. Hinkal, S. Ansieau, A. Puisieux, Generation [28] R. Derda, D.J. Wherritt, L.L. Kiessling, Solid-phase synthesis of alkanethiols for of breast cancer stem cells through epithelial-mesenchymal transition, PLoS the preparation of self-assembled monolayers, Langmuir 23 (2007) 11164– One 3 (2008) e2888. 11167. [7] C.L. Chaffer, N.D. Marjanovic, T. Lee, G. Bell, C.G. Kleer, F. Reinhardt, A.C. [29] Y.N. Xia, X.M. Zhao, E. Kim, G.M. Whitesides, A selective etching solution for D’Alessio, R.A. Young, R.A. Weinberg, Poised chromatin at the ZEB1 use with patterned self-assembled monolayers of alkanethiolates on gold, enables breast cancer cell plasticity and enhances tumorigenicity, Cell 154 Chem. Mater. 7 (1995) 2332–2337. (2013) 61–74. [30] R. Derda, S. Musah, B.P. Orner, J.R. Klim, L. Li, L.L. Kiessling, High-throughput [8] P.B. Gupta, C.M. Fillmore, G. Jiang, S.D. Shapira, K. Tao, C. Kuperwasser, E.S. discovery of synthetic surfaces that support proliferation of pluripotent cells, J. Lander, Stochastic state transitions give rise to phenotypic equilibrium in Am. Chem. Soc. 132 (2010) 1289–1295. populations of cancer cells, Cell 146 (2011) 633–644. [31] L.E. Little, K.Y. Dane, P. Daugherty, K.E. Healy, D.V. Schaffer, Exploiting bacterial [9] M.M. Kohl, O.A. Shipton, R.M. Deacon, J.N.P. Rawlins, K. Deisseroth, O. Paulsen, peptide display technology to engineer biomaterials for neural stem cell Hemisphere-specific optogenetic stimulation reveals left-right asymmetry of culture, Biomaterials 32 (2011) 1484–1494. hippocampal plasticity, Nat. Neurosci. 14 (2011) 1413–1415. [32] B.P. Gray, K.C. Brown, Combinatorial peptide libraries: mining for cell-binding [10] R.J. Akhurst, R. Derynck, TGF-beta signaling in cancer–a double-edged sword, peptides, Chem. Rev. 114 (2014) 1020–1081. Trends Cell Biol. 11 (2001) S44–S51. [33] F. Deiss, W.L. Matochko, N. Govindasamy, E.Y. Lin, R. Derda, Flow-through [11] R.G. Wells, D.E. Discher, Matrix elasticity, cytoskeletal tension, and TGF-beta: synthesis on teflon-patterned paper to produce peptide arrays for cell-based the insoluble and soluble meet, Sci. Signal. 1 (2008) pe13. assays, Angew. Chem. Int. Ed. Engl. 53 (2014) 6374–6377. [12] F. Deiss, A. Mazzeo, E. Hong, D.E. Ingber, R. Derda, G.M. Whitesides, Platform for high-throughput testing of the effect of soluble compounds on 3D cell cultures, Anal. Chem. 85 (2013) 8085–8094.