Peptide microarray patterning for controlling and monitoring cell growth

Accepted Manuscript Peptide microarray patterning for controlling and monitoring cell growth Edith Lin, Adhirath Sikand, Jessica Wickware, Yubin Hao, Ratmir Derda PII: DOI: Reference:

S1742-7061(16)30028-9 http://dx.doi.org/10.1016/j.actbio.2016.01.028 ACTBIO 4081

To appear in:

Acta Biomaterialia

Received Date: Revised Date: Accepted Date:

22 July 2015 9 January 2016 20 January 2016

Please cite this article as: Lin, E., Sikand, A., Wickware, J., Hao, Y., Derda, R., Peptide microarray patterning for controlling and monitoring cell growth, Acta Biomaterialia (2016), doi: http://dx.doi.org/10.1016/j.actbio. 2016.01.028

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Peptide microarray patterning for controlling and monitoring cell growth. Edith Lin, Adhirath Sikand, Jessica Wickware, Yubin Hao, Ratmir Derda* Department of Chemistry, University of Alberta, Edmonton, Alberta T6G 2G2, Canada * [email protected] Abstract The fate of cells is influenced by their microenvironment and many cell types undergo differentiation when stimulated by extracellular cues, such as soluble growth factors and the insoluble extracellular matrix (ECM). Stimulating differentiation by insoluble or “immobilized” cues is particularly attractive because it allows for the induction of differentiation in a spatiallydefined cohort of cells within a 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 cell culture and fluorescence microscopy. As a model, we used normal murine mammary gland (NMuMG) cells, which undergo epithelial-tomesenchymal transition (EMT) in response to soluble transforming factor beta (TGF-β) and surface-immobilized peptides that target TGF-β receptors (TGFβRI/II). We repurposed a wellestablished DNA microarray printing technique to produce arrays of micro patterned surfaces that displayed TGFβRI/II-binding peptides and integrin binding peptides. Upon long-term culture on these arrays, only NMuMG cells residing on EMT-stimulating areas exhibit growth arrest and a decrease in E-cadherin expression. We believe that the methodology developed in this report will aid the development of peptide-decorated surfaces that can locally stimulate defined cells surface receptors in cells and control EMT and other well-characterized differentiation events.

1

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 investigation of physiological responses of cells to diverse ligands for cell surfaces receptors. As a model response, we used a well-established cell line that is known to undergo epithelial to mesenchymal transition (EMT) in response to extracellular ligands stimulation. EMT is an essential process that was first recognized for organ formation in embryonic 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 signaling that causes the formation and maintenance of highly tumorigenic cells, known as cancer stem cells (CSCs) [4-6]. CSCs, by definition, are a subpopulation of cells within a tumor that have a higher tumor forming capacity in xenograft models; CSCs were also proposed to be the cells that form metastases after EMT has occurred in vivo [3]. The non-stem and cancer stem cells interconvert reversibly via a process closely allied to EMT; this process is stochastic and only a few percentage of the cancer cells undergo EMT-like events that lead to de-differentiation to putative CSCs [7, 8]. To aid the investigation of EMT and its connection to differentiation/dedifferentiation events in a tumor, we developed a method that would permit the stimulation of EMT in cells in a spatially localized manner. The most characterized extracellular stimulus that initiates EMT and enrichment of CSClike cells in multiple cancer cell lines is transforming growth factor beta (TGF-β) [9]. The TGF-β signaling pathway is initiated by dimerization of serine/threonine kinase receptors TGFβRI/II;

2

this complex then phosphorylates intracellular Smad, leading to translocation of phospho-Smad to the nucleus [1, 10]. In vivo, TGF-β acts both as soluble ligand and ECM-immobilized ligand [11]; immobilization of TGF-β protein on polymeric scaffolds in vitro is known to preserve its bioactivity and ability to elicit TGF-β signaling in cells. Localized immobilization of proteins is a powerful tool in cell biology [12], but further minimization of proteins to functional short peptides can help generalize this technology to bypass issues such as residue-specific immobilization and inactivation of proteins. Specifically, Kiessling et al. previously demonstrated that EMT can be induced in large scale populations of mammary epithelial cells when cultured on top of surface immobilized peptide analogs of TGF-β. The cells display signatures of EMT, such as down-regulation and internalization of E-cadherin, increase in smooth muscle actin expression, change in cell morphology, and growth arrest [2]. The use of chemically-modified surfaces for investigating fundamental cell biology questions dates back to 1978, when Folkman et al. used surfaces of different hydrophobicity to show that the area to which the cell spreads determines its survival [13]. In 1990 and the early 2000’s, “binary” micropatterning of self-assembled monolayers (SAMs) into areas that promoted and suppressed attachment of cells made it possible to probe aspects of cellular adhesion [14], migration [15], and division [16]. In the 2000’s, several groups reported technologies for copatterning multiple chemically-defined SAMs for co-culturing different cells [17]; these arrays of SAMs empowered the discovery of peptide-modified substrates that control differentiation of human embryonic stem cells [18]. Patterning technologies can dissect spatially-defined phenotypic transformation in individual cells, such as branching morphogenesis [19], or asymmetric division [20]; technologies for this application should allow for the control of chemical composition, size, and shape of patterned areas to accommodate single or multiple cells.

3

Although technology for surface patterning developed significantly in the past 20 years, many techniques have been perfected to yield high-fidelity binary patterns only; modern technology for multi-component patterning surfaces relies on cumbersome serial processing, accurate alignment, and access to instrumentation not accessible outside of specialized laboratories [21]. The most scalable techniques for multi-component patterning of surfaces are those that repurpose welldeveloped, high-throughput screening instruments, such as robotic spotters or microarray printers. While these technologies have been applied to fabrication of protein arrays [22-24], their application to fabrication of peptide monolayers is limited [25, 26]. In this report, we describe the use of a DNA microarray printer to pattern gold coated surfaces with micron-scaled islands of peptides that bind to TFGβRI/II and integrin receptors. The peptides are conjugated to an alkanethiol linker for the creation of SAMs on gold-coated glass surfaces. In the future, this method can be used to investigate the heterogeneity of EMTresponses and CSC-dedifferentiation events in carcinoma lines in vitro (Fig. 1). Contrary to the popular method of using soft lithography for printing spatially defined SAMs, a DNA microarray printer provides an alternative simple, robust, and reproducible method of generating arrays containing diverse surfaces. The size of the spots can be tuned within a range of 150-400 µm in diameter to create arrays for monitoring the growth of single or multiple cells (Fig 1D). In this report, we primarily use areas of 150 µm in diameter to accommodate multiple cells for the study of cell adhesion, spreading, proliferation, and EMT in spatially defined subpopulations of cells in one common media. 2. Results and Discussion 2.1 Validating E-cadherin expression in NMuMG cells grown on large surfaces of peptide SAMs

4

We selected three peptides for production of SAMs and microarrays of SAMs with NMuMG cells. Peptides LTGKNFPMFHRN and MHRMPSFLPTTL were previously reported to cluster TFG-β receptors I and II and potentiate the effects of soluble TFG-β at low concentrations [18]. The peptide sequence GRGDS, which binds to cell integrin receptors, was included as a negative control. As a substrate for SAMs, we used coverslips sputtered with a layer of translucent gold to facilitate analysis of these arrays via phase-contrast and fluorescent/confocal microscopy. NMuMG cells were monitored for the following hallmarks of EMT: decreases in E-cadherin levels, increases in smooth muscle actin levels, growth arrest, and morphological changes. Gold exhibited no deleterious effect on NMuMG cells and cells cultured on gold coated coverslips displaying GRGDS SAMs exhibited no changes in the basal level of expression of E-cadherin. We validated that the basal level of E-cadherin in NMuMG cells cultured on large areas SAMs displaying peptides 1 or 2 changed significantly when compared to cells growing on GRGDS SAMs. After four days of culture, E-cadherin levels, visualized via immunofluorescent staining, were observed to be the highest in NMuMG cells grown on GRGDS and the lowest in NMuMG cells grown on LTGKNFPMFHRN (Supplementary Fig. S1). We also monitored on these surfaces daily with a phase-contrast microscopy to check for morphological changes (Supplementary Fig. S2A). After four days, the cells cultured on SAMs of peptides 1 and 2, and in media treated with soluble TFG-β, lost epithelial morphology and gained spindle like appearance (Supplementary Fig. S2B). 2.2 Optimization of printing SAMs For all microarray experiments, we used a pattern consisting of 4 identical quadrants of 9x9 pattern (Fig. 1B). To integrate commercially available, small footprint microscope cover

5

slides (18x18mm) with the DNA microarray spotter, we used custom-made stainless steel slide holders for printing (Fig. 1C). In all cell-based experiments in this manuscript, we used Telechem SMP3 Stealth Micro Spotting Pins at 0 ms contact time to produce peptide-modified spots of 150 micron in diameter. To demonstrate the range of spot sizes that can be generated on the arrays, we printed spots of 400 µm with SNS15 Stealth Solid Pins at 0ms contact time (Fig. 1D). Shortly after printing, we immersed the arrays into the solution of glucamine-alkanethiol to form a protein- and cell-repelling background between the spots [27]. The best printing results were achieved on freshly-deposited gold-coated surfaces (Supplementary Fig. S3). We attempted to re-use gold-coated surfaces by stripping the SAMs with detergent and plasma cleaning; upon reprinting the arrays, we noticed that the alignment of the printed spots highly varied specifically on reused surfaces. In addition to printing peptide-AT, we also printed rows with KCN solution to produce a convenient visual reference of etched gold (Fig. 2A) [28, 29]. We spiked the solutions of peptide-alkanethiols used for printing with fluorescein to visualize the locations of the printed areas before seeding cells (e.g. Figure 2B). We observed minor variations in the intensity of fluorescence, but these variations should not be interpreted as non-uniformity in SAMs, as printing deposits an excess amount of alkanethiols for the formation of SAMs. To confirm the functional uniformity of the formed monolayers, we produced three independent arrays and allowed for short term adhesion of NMuMG cells (24 hours) to the array to minimize any variability from long-term growth (Supplementary Fig. S4). Counting the number of cells per spot yielded only minor variations in the number of cells adhered to each peptide spot in three independent arrays or 36 replicates on the same array. 2.3 Optimization of cell concentrations for short-term growth

6

The optimal cell concentration for short-term growth experiments (0-5 days) should initially yield 1-5 cells adhering per area to allow cells to divide in the areas during the subsequent growth period. We optimized conditions for NMuMG cells by seeding a suspension of 10,000, 25,000, or 50,000 cells in 3 mL to an array in a 6 well plate. After one hour of incubation, we rinsed the non-bound cells and supplemented the arrays with fresh NMuMG media. Fig. 5 describes the number of cells on arrays after 24 hours of culture. We found the optimal density of cells for short-term growth to be 3,300 cells/mL, or 10,000 cells/well (Fig. 2C).

2.4 E-cadherin expression of cells on printed SAMs We cultured NMuMG cells on microarrays that displayed peptides 1, 2, and GRGDS to identify conditions that induce EMT in cells in spatially defined areas. We anticipated EMT to be induced in cells grown on islands of SAMs with peptides 1 and 2, but not on parallel islands of GRGDS. For one to three days, cells exhibited high levels of E-cadherin on all three peptides. On the fourth day, we detected decreased levels of E-cadherin in cells growing on LTGKNFPMFHRN and MHRMPSFLPTTL, whereas cells growing on GRGDS did not show any decrease in E-cadherin expression (Fig. 3). 2.5 Monitoring growth rate and cell morphology To detect growth arrest, we counted the number of cells growing on individual peptidemodified areas after 2, 3, and 4 days in 35-40 replicates. We modified the original 9X9 pattern to an asymmetrical pattern with distinct spaces for patterning peptides 1, 2, GRGDS, and KCN (Fig. 2A). This allowed for easier visual identification of peptide islands (Fig 2B). On the second day, NMuMG cells only adhered to 50% of LTGKNFPMFHRN or MHRMPSFLPTTL modified

7

areas, whereas cells adhered to 75% GRGDS modified areas. On the fourth day, dense populations of cells were found on over 60% of GRGDS modified areas. In contrast, both LTGKNFPMFHRN and 2 areas contained biphasic populations; dense populations were observed in approximately two-thirds of the areas, and sparse populations were observed on one third of the areas (Supplementary Fig. S5). Changes in cell morphology were difficult to deconvolute due to the combination of both growth and packing of cells in an area of limited size. NMuMG cells growing on GRGDS had significantly smaller spreading area by day 4 than cells growing on peptides 1 and 2 (Supplementary Fig. S6). For all three peptides, the size of individual cells doubled over the course of 4 days; cells spread from an average area of 1988 µm2 on day 2 to 3745 µm2 on day 4. Visual inspection showed that NMuMG cells grown on GRGDS islands yielded dense colonies resulting from rapid cell growth. The average area of a cell on a constrained growth island, thus, was small. Cell colonies were less dense on peptide1 and 2 islands due to reduced growth and division yielding, on average, larger cells that spread through the entire area. The observed reciprocal relation between cell number and cell area arises because cells grow and spread on the constant area S. Individual cells are capable of spread over the entire area exhibiting cells spreading area as large as S, but as cells divide to produce N cells, the areas of individual cells is reduced to ~S/N. 2.6 Culturing MDA-MB-231 and NMuMG cells on peptides found through phage display In vitro display technologies, such as phage-[30], bacteria-[31] and cell-surface displays, are powerful technologies for de novo discovery of ligands that target cell surface receptors and control differentiation of cells [32]. We previously developed a series of peptides that bind to MDA-MB-231 cells [33] and further demonstrated that many of these peptides can support

8

adhesion of closely and distally related epithelial tumour cell lines (manuscript in review). We sought to test whether the identified peptides support adhesion of normal epithelial cells, and whether adhesion of cells to these peptides can induce EMT. Here, we show that peptide arrays can be used as a medium-throughput phenotypic screening method for validating of peptides discovered by phage display. We previously identified peptides STASYTR, GKPMPPM, AMSSRSL, IPAPLRS, and HAIYPRH by phage-display library selection against intact MDAMB-231 basal breast cancer cells. Cellulose peptide arrays, created using the SPOT peptide synthesis method, validated the ability of these peptides to support short-term adhesion of these MDA- MB-231 GFP cells [33]. When printed on the SAMs-array, all peptides supported shortterm adhesion of MDA-MB-231 cells as well as their growth over the course of three days (Fig. 4A). As these peptides bind to one of the receptors on mesenchymal-like cancer cells, we investigated whether any of these peptides can stimulate EMT in epithelial cells. NMuMG cells adhered to four out of five peptides. Over the course of 4 days, we observed decrease in E-cadherin expression in NMuMG cells on surfaces that displayed IPAPLRS- and GKPMPPM-terminated SAMs to the levels compatible to those observed on control TGFβR-binding peptides (LTGKNFPMFHRN and MHRMPSFLPTTL). In contrast, level of immunostaining with E-cadherin antibody in cells growing on HAIYPRH and AMSSRSL SAMs remained similar to that observed on cells cultured atop GRGDS SAMs (Fig. 4B). We also inspected the increase of smooth muscle actin (SMA) expression in NMuMG cells grown on peptide arrays, displaying TGFβRI/II binding, phage derived, and integrin binding peptides, for 4 days. We observed the highest SMA expression in cells cultured on LTGKNFPMFHRN, TGFβRI/II binding peptide, and no increased expression of SMA on

9

surfaces modified with IPAPLRS when compared to cells grown on areas modified with RGD peptide. 3. Conclusion In phenotypic cell-based assays, the performance of microarray-printed peptide arrays is similar to the performance of previously reported peptide arrays. [15, 31]. These arrays integrate seamlessly with standard cell culture, imaging, and immunofluorescence technologies. Regular arrangement of peptide areas and visual references, such as a grid of etched gold-free zones, facilitate both automatic analysis and daily manual monitoring of the assays. We believe that the major benefit of the microarray peptide SAMs described in this report, when compared to the previous reports, is simplicity in fabrication and scalability in production. Specifically, it is possible to scale up both the number of different peptides printed on the surface and the number of arrays printed in a single session. Most DNA microarray printers are designed for high-speed production of up to a hundred arrays at once. One of the remaining limitations is the synthesis of peptide-alkanethiols (AT); every peptide-AT must be purified and characterized independently. Although the direct synthesis of peptides on the surface could potentially yield larger number of peptides and bypass the need for purification, there is a clear compromise between the number of peptides and quality. In our experience, even optimized synthesis of short peptides on surface can yield products with less than 50% purity [34]. Another limitation is the analysis of the arrays; this problem can be avoided by using specialized, high-content analysis hardware and advanced image recognition software. An alternative solution is imaging of the arrays on a fluorescent gels scanner, or DNA array scanner. In conclusion, we believe that this printed SAMarray will be an important tool in validation of small-molecule and peptide ligands that induce differentiations and phenotypic transformations in cells.

10

4. Materials and Methods 4.1 Synthesis of peptides on an alkanethiol linker The alkanethiol linker was synthesized and conjugated to polystyrene monomethoxytrityl chloride (MMT-Cl) resin in preparation for Fmoc solid phase peptide synthesis. The peptides LTGKNFPMFHRN and MHRMPSFLPTTL, STASYTR, GKPMPPM, AMSSRSL, IPAPLRS, and HAIYPRH were synthesized on AT-modified resin to make peptide-conjugated alkanethiols as described in our previous publication [19, 29]. Amino acids were purchased from Chempep. The peptides were purified using high performance liquid chromatography (HPLC) and verified through matrix-assisted laser desorption/ionization (MALDI) and liquid chromatography mass spectrometry (LCMS). Peptides were stored at -20 °C as lyopholized powder and as a 1 mM solution in milli-Q water.

4.2 Making and storing gold chips 99.99% pure gold (Gold Maple Leaf pure bullion coin) was evaporated onto 18X18 mm glass cover slides using a thermal evaporation system from Torr International Inc., model THEUPG, to create the platform for self assembled monolayers. Immediately before deposition, the glass cover slides were cleaned by soaking in piranha solution (1 H2O2: 3 H2SO4) for 20 minutes, extensively rinsed with milli-Q H2O, and dried with argon gas. Warning: Piranha solution is extremely corrosive and potentially explosive if in contact with organic matter. 5 nm of Chromium was first evaporated onto the glass surfaces, followed by 20 nm of gold. The gold chips were stored in a clean polystyrene petri dish for up to 3 weeks. 4.3 Microarray printing of peptides and KCN onto gold surfaces

11

Peptides 1 and 2 were dissolved in milli-Q water at 1mM concentration. We supplemented each solution with 0.2 mM fluorescein; this concentration was necessary to visualize the location of the spot. The KCN solution for etching gold was made by dissolving NaOH and KCN in milli-Q water to yield a 1 mM of KCN in 1M NaOH solution. Warning: this solution is extremely toxic and corrosive. The three solutions were printed onto the gold surfaces by using the DNA microarrayer in the Microarray and Proteomics Facility (MAF) at the University of Alberta. Telechem SMP3 pins from Arrayit and 0 ms contact time were used to deposit the solutions onto the surface. After printing, the surface of the gold slide was placed on top of a solution of glucamineAT to form a cell-repellent layer in between printed areas. Specifically, 30 uL of the solution was deposited in a sterile petri dish and the array was inverted onto the glucamine solution, forming a uniform layer of liquid sandwiched in between the array and the surface of the petri dish. A wet napkin was placed on the lid of the petri dish and the lid was sealed with parafilm to prevent evaporation of the glucamine-AT solution. The blocking process was allowed to proceed for 1224 hours at room temperature. Prior to cell-adhesion studies, the arrays were sterilized with 95% ethanol three times for 20 minutes. The arrays were transferred to a sterile 6 well plate in a laminar flow hood and rinsed 3 times with 1X PBS and 3 times with MEM. 4.4 Cell Culture Conditions of NMuMG cells NMuMG cells were maintained in DMEM supplemented with 10% FBS, non-essential amino acids, GlutaMax™, and 10 ug/mL of human insulin. For cell adhesion and growth experiments on peptide arrays, the cells were detached from the culture surface with trypsin (5 min), rinsed and resuspended with growth medium, and seeded onto the arrays in six well plates.

12

The arrays were incubated at 37 °C with 5% CO2 for 2 hours before the media was replaced with growth media supplemented with 1ng/mL of soluble TGF-β. The cells were cultured for 4 days at 37 °C with 5% CO2. 4.5 Immunofluorescent Staining Cells on the gold surfaces were fixed with 2% paraformaldehyde (PFA) in PBS and stored at 4 °C for a minimum of 30 minutes. PFA was aspirated from the wells and the cell membrane was permeabilized by 0.1% TritonX-100 in PBS for 15 minutes. The arrays were incubated in a blocking buffer (10% calf bovine serum [CBS] and 10% TritonX-100 in PBS) for 1 hour. For E-cadherin staining, the arrays were incubated with monoclonal rat antiuvomorulin/E-cadherin (Sigma), diluted 1:200 in PBS, for 24 hours at 4 °C. The samples were rinsed 3 x 5 min with PBS and incubated with Alexa Fluor 488 conjugated rabbit anti-rat (Invitrogen) at 1:500 dilution for 1 hour at room temperature. For smooth muscle actin (SMA) staining, the arrays were incubated with rabbit anti-alpha smooth muscle actin (Abcam), diluted to 5 mg/mL, for 24 hours at 4°C. The samples were then rinsed 3 x 5 min with PBS and incubated with Alexa Fluor 488 Donkey Anti Rabbit IgG (Invitrogen) at 1:500 dilution for 1 hour at room temperature. All arrays were then rinsed 3 x 5 min with 1X PBS and incubated with DAPI (1 mg/mL in 1X PBS) for 30 minutes at room temperature to stain the cells’ nuclei. The arrays were imaged using a confocal fluorescence microscope (Zeiss LSM 700) with 20X objective or 64X oil immersion objective. 4.6 Area analysis and cell count Zen software was used for image processing. Cell peripheries were traced to define the area each cell spread to. The cell count was performed by manually counting the number of

13

nuclei, stained with DAPI dye, per peptide area. Multi-layered, over populated peptide spots were excluded from data analysis for efficient analysis. Boxplot graphs were created using Origin 8.6 software (OriginLab, Northampton, MA, USA), which show relative increase of cell number or area over time. Acknowledgements This work was supported by the Canadian Institute of Health Research (CIHR) Bridge Funding. E.L. acknowledges Alberta Innovates Health Solutions (AIHS) for the Summer Studentship; A.S. acknowledges Mitacs Globalink Research Internship. J.W. acknowledges the Undergraduate Research Initiative (URI) for the summer studentship. Infrastructure support was provided by Canadian Foundation for Innovation (CFI). Microarray printing was provided by Troy Locke in the Microarray and Proteomics Facility (MAF) at the University of Alberta. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at

References [1] [2] [3] [4] [5]

R. Kalluri, R.A. Weinberg, The basics of epithelial-mesenchymal transition. J Clin Invest. 119 (2009) 1420-1428. J. Yang, R.A. Weinberg, Epithelial-Mesenchymal Transition: At the Crossroads of Development and Tumor Metastasis. Cell Press: Dev Cell. 14 (2008) 818-829. C. Scheel, R.A. Weinberg, Cancer stem cells and epithelial-mesenchymal transition: concepts and molecular links. Semin Cancer Biol. 22 (2012) 396-403. P.B. Gupta, C.L. Chaffer, R.A. Weinberg, Cancer stem cells: mirage or reality? Nat Med. 15 (2009) 1010-1012. M.J. Meyer, J.M. Fleming, M.A. Ali, M.W. Pesesky, E. Ginsburg, B.K. Vonderhaar, Dynamic regulation of CD24 and the invasive, CD44posCD24neg phenotype in breast cancer cell lines. Breast Cancer Res. 11 (2009) R82.

14

[6]

[7]

[8]

[9]

[10] [11] [12]

[13] [14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

A.P. Morel, M. Lièvre, C. Thomas, G. Hinkal, S. Ansieau, A. Puisieux, Generation of breast cancer stem cells through epithelial-mesenchymal transition. PLoS One. 3 (2008) e2888. C.L. Chaffer, N.D. Marjanovic, T. Lee, G. Bell, C.G. Kleer, F. Reinhardt, A.C. D'Alessio, R.A. Young, R.A. Weinberg, Poised chromatin at the ZEB1 promoter enables breast cancer cell plasticity and enhances tumorigenicity. Cell. 154 (2013) 61-74. P.B. Gupta, C.M. Fillmore, G. Jiang, S.D. Shapira, K. Tao, C. Kuperwasser, E.S. Lander, Stochastic state transitions give rise to phenotypic equilibrium in populations of cancer cells. Cell. 146 (2011) 633-644. M.M. Kohl, O.A. Shipton, R.M. Deacon, J.N.P. Rawlins, K. Deisseroth, O. Paulsen, Hemisphere-specific optogenetic stimulation reveals left-right asymmetry of hippocampal plasticity. Nat Neurosci. 14 (2011) 1413-1415. R.J. Akhurst, R. Derynck, TGF-beta signaling in cancer--a double-edged sword. Trends Cell Biol. 11 (2001) S44-S51. R.G. Wells, D.E. Discher, Matrix elasticity, cytoskeletal tension, and TGF-beta: the insoluble and soluble meet. Sci Signal. 1 (2008) pe13. 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. J. Folkman, A. Moscona, Role of Cell-Shape in Growth-Control, Nat. 273 (1978) 345349. M. Mrksich, L.E. Dike, J. Tien, D.E. Ingber, G.M. Whitesides, Using microcontact printing to pattern the attachment of mammalian cells to self-assembled monolayers of alkanethiolates on transparent films of gold and silver. Exp Cell Research. 235 (1997) 305-313. 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 (2005) 975978. 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 Biol. 7 (2005) 947-953. W.S. Dillmore, M.N. Yousaf, M. Mrksich, A photochemical method for patterning the immobilization of ligands and cells to self-assembled monolayers. Langmuir. 20 (2004) 7223-7231. L.Y. Li, J.R. Klim, R. Derda, A.H. Courtney, L.L. Kiessling, Spatial control of cell fate using synthetic surfaces to potentiate TGF-beta signaling. Proc. Natl. Acad. Sci. U.S.A. 108 (2011) 11745-11750. 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. 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. S. Ankam, B.K. Tea, M. Kukumberg, E.K. Yim, High throughput screening to investigate the interaction of stem cells with their extracellular microenvironment. Organogenesis. 9 (2013) 128-142.

15

[22] [23] [24] [25]

[26]

[27] [28] [29]

[30]

[31]

[32] [33]

C.J. Flaim, D. Teng, S. Chien, S.N. Bhatia, Combinatorial signaling microenvironments for studying stem cell fate. Stem Cells Dev. 17 (2008) 29-39. 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) 4324. G. MacBeath, S.L. Schreiber, Printing proteins as microarrays for high-throughput function determination. Science. 289 (2000) 1760-1763. J.R. Falsey, M. Renil, S. Park, S. Li, K.S. Lam, Peptide and small molecule microarray for high throughput cell adhesion and functional assays. Bioconjugate Chem. 12 (2001) 346-353. S. Li, N. Marthandan, D. Bowerman, H.R. Garner, T. Kodadek, Photolithographic synthesis of cyclic peptide arrays using a differential deprotection strategy. Chem Commun (Camb), 5 (2005) 581-583. B.P. Orner, R. Derda, R.L. Lewis, J.A. Thomson, L.L. Kiessling, Arrays for the combinatorial exploration of cell adhesion. J. Am. Chem. Soc. 126 (2004) 10808-10809. R. Derda, D.J. Wherritt, L.L. Kiessling, Solid-phase synthesis of alkanethiols for the preparation of self-assembled monolayers. Langmuir. 23 (2007) 11164-11167. Y.N. Xia, X.M. Zhao, E. Kim, G.M. Whitesides, A selective etching solution for use with patterned self-assembled monolayers of alkanethiolates on gold. Chem of Mater. 7 (1995) 2332-2337. R. Derda, S. Musah, B.P. Orner, J.R. Klim, L. Li, L.L. Kiessling, High-throughput discovery of synthetic surfaces that support proliferation of pluripotent cells. J. Am. Chem. Soc. 132 (2010) 1289-1295. L.E. Little, K.Y. Dane, P. Daugherty, K.E. Healy, D.V. Schaffer, Exploiting bacterial peptide display technology to engineer biomaterials for neural stem cell culture. Biomaterials. 32 (2011) 1484-1494. B.P. Gray, K.C. Brown, Combinatorial peptide libraries: mining for cell-binding peptides. Chem Rev. 114 (2014) 1020-1081. F. Deiss, W.L. Matochko, N. Govindasamy, E.Y. Lin, R. Derda, Flow-through synthesis on Teflon-patterned paper to produce peptide arrays for cell-based assays. Angew Chem Int Ed Engl. 53 (2014) 6374-6377.

16

Figures and Legends

Figure 1: (A) Schematic of E-Cadherin down regulation and epithelial to mesenchymal transition (EMT) in NMuMGcells when cultured on microarray of self-assembled monolayer (SAM) of peptides. EMT is only induced in cells grown on areas of LTGKNFPMFHRN and not MHRMPSFLPTTL. (B) Design of microarray pattern: 9x9 spots in 4 quadrants. The spots printed for all cell based experiments are 150 µm in diameter, printed with Telechem SMP3 Stealth Micro Spotting Pins at 0 ms contact time. (C) Stainless steel holder for gold coated slides. The holder is the size of a microscope slide. (D) Spots of 150 µm and 400 µm diameter were printed with Telechem SMP3 Stealth Micro Spotting Pins and SNS15 Stealth Solid Pins, respectively.

17

Figure 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 seeding with 10000, 25000, and 50000 cells/well.

18

Figure 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) NMuMG downregulate the expression of E-cadherin by day 4 on areas that contain LTGKNFPMFHRN and MHRMPSFLPTTL but not on areas that contain GRGDS peptide.

19

Figure 4: Cells cultured on arrays of peptides identified by phage-display panning on MBA-MB231 cells (A) 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 image represents the tile scan of entire array; the intensity of greyscale is proportional to GFP fluorescence. Insets describe individual array elements at 20x magnification. (B) 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 Ecadherin fluorescence. The images represent selected areas of 3 peptide zones. Insets describe individual array elements at 20x magnification.

20

21