Micropatterned matrix directs differentiation of human mesenchymal stem cells towards myocardial lineage

E XP E RI ME N T AL C E L L R E S EA RC H 31 6 ( 20 1 0) 1 1 5 9– 1 16 8

available at www.sciencedirect.com

www.elsevier.com/locate/yexcr

Research Article

Micropatterned matrix directs differentiation of human mesenchymal stem cells towards myocardial lineage Chor Yong Tay a , Haiyang Yu a , Mintu Pal a , Wen Shing Leong a , Nguan Soon Tan b , Kee Woei Ng a , David Tai Leong c , Lay Poh Tan a,⁎ a

School of Materials Science and Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore, 639798 School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, Singapore, 637551 c Cancer Science Institute of Singapore, National University of Singapore, Centre for Life Sciences, 28 Medical Drive, Singapore, 117456 b

A R T I C L E I N F O R M A T I O N

AB S TR AC T

Article Chronology:

Stem cell response can be influenced by a multitude of chemical, topological and mechanical

Received 16 September 2009

physiochemical cues. While extensive studies have been focused on the use of soluble factors to

Revised version received

direct stem cell differentiation, there are growing evidences illustrating the potential to modulate

5 February 2010

stem cell differentiation via precise engineering of cell shape. Fibronectin were printed on poly

Accepted 8 February 2010

(lactic-co-glycolic acid) (PLGA) thin film forming spatially defined geometries as a means to control

Available online 13 February 2010

the morphology of bone marrow derived human mesenchymal stem cells (hMSCs). hMSCs that were cultured on unpatterned substrata adhered and flattened extensively (∼ 10,000 μm2) while

Keywords:

cells grown on 20 μm micropatterend wide adhesive strips were highly elongated with much

Micro patterning

smaller area coverage of ∼ 2000 μm2. Gene expression analysis revealed up-regulation of several

Stem cell differentiation

hallmark markers associated to neurogenesis and myogenesis for cells that were highly elongated

Gene expression

while osteogenic markers were specifically down-regulated or remained at its nominal level. Even

Cell shape

though there is clearly upregulated levels of both neuronal and myogenic lineages but at the

Cytoskeletal rearrangement

functionally relevant level of protein expression, the myogenic lineage is dominant within the time scale studied as determined by the exclusive expression of cardiac myosin heavy chain for the micropatterned cells. Enforced cell shape distortion resulting in large scale rearrangement of cytoskeletal network and altered nucleus shape has been proposed as a physical impetus by which mechanical deformation is translated into biochemical response. These results demonstrated for the first time that cellular shape modulation in the absence of any induction factors may be a viable strategy to coax lineage-specific differentiation of stem cells. © 2010 Elsevier Inc. All rights reserved.

Introduction Over the past years, both embryonic and adult stem cells have been identified as major contenders as ideal cell sources in the field of tissue engineering and regenerative medicine [1]. This is largely due to its ability to self-renew and to differentiate into cells belonging to a multitude of tissue lineages. The key to capitalize on ⁎ Corresponding author. Fax: +65 6790 9081. E-mail address: [email protected] (L.P. Tan). 0014-4827/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2010.02.010

stem cells is the ability to direct its differentiation to the targeted cells of interest. Stem cell fate is controlled by a complex set of signals found in the cellular microenvironment [2,3]. These signals can exist in the form of soluble factors, the extracellular matrix (ECM), the biophysical environment and inter cellular contacts instructing whether the stem cell should undergo apoptosis, proliferate, differentiate or remain quiescent [2].

1160

E XP E RI ME N T AL C E L L R E SE A RC H 31 6 ( 20 1 0) 1 1 59 – 1 16 8

Typically, the classical method to control stem cell differentiation is by using biochemical factors. The general dogma is that these biochemical molecules will bind to the receptors located at the cell membrane resulting in the activation of second messenger such as cyclic adenosine monophosphate, or production of kinases that will in turn initiate a cascade of signaling pathways affecting cellular responses [4–6]. However, in recent years there have been growing evidences that the physical microenvironment the cell resides (stem cell niche) is also capable in directing stem cell fate [7]. Investigating the mechanotransduced responses of stem cells subjected to cyclical stretching [8,9] interactions with nano scale features [10–12], ECM stiffness [13] and even cell shapes [14] have emerged, suggesting that mechanical forces play a major regulatory role as well. Interestingly, regardless of the kind of stimuli the cells receive, cellular response is usually accompanied by alterations in cellular morphology [11,13–16]. In fact, changes in cell shape are often used as an early indicator of modulated cell functions [13,14]. A number of independent studies have shown that cell fate can be influenced via modulation of cell shape. By culturing embryonic stem cells on multi-perforated polycarbonate with holes of either 10 μm or 20 μm, Yang et al. [17] were able to alter the cell shape and showed that only cells that were elongated underwent myogenesis while cells that were rounded remained undifferentiated. McBeath et al. [14] showed that differentiation of human mesenchymal stem cells (hMSCs) can be modulated in part by cell shape and cytoskeleton tension. Human tendon fibroblast cultured on adhesive islands of varying aspect ratios showed increased expression of collagen I in those cells adopting a more elongated morphology, suggesting an existence of an optimized shape for specific cellular functions [18]. A recent study also linked the proliferative state of human aortic smooth muscle cells to the degree of elongation [19]. Taken together, these studies suggest that cell shape distortion is a potent regulator of cellular responses. In this study, we report a facile strategy to direct early stage stem cell lineage commitment by micro patterning hMSCs on 20 μm wide fibronectin strips printed on a poly (Lactic-co-glycolic acid) (PLGA) thin film. It was shown that modifications in cell morphology and cytoskeletal arrangement exerted significant effects on the early stages of stem cell lineage commitment. Elongated hMSCs expressed higher level of mRNAs that are usually associated with neurogenesis and myogenesis but specifically down regulated several osteogenic genes. Actin filament arrangement, vinculin distribution and cell nuclei of the elongated cells contrasted starkly from the unpatterned cells. Results presented in this study showed that modulation of hMSCs morphology alone may drive cell lineage commitment towards myocardial lineage. This method can potentially be used to improve graft take of many engineered tissues.

Materials and methods Preparation and characterization of poly (lactic-co-glycolic acid) thin film Poly (lactic-co-glycolic acid) (PLGA) ultrathin film (16.3 ± 1.5 μm) was prepared by the solvent cast method. PLGA (Boehringer

Ingelheim Pharma Chemicals, Resomer® LG824 S) was dissolved in dichloromethane (DCM) at a pre fixed ratio of 1 g PLGA: 15 ml of DCM. The polymer films were cast using a film applicator (Paul N. Gardner Company, Inc.). The cast film was then transferred to a vacuum oven to be dried for 5 days to ensure complete removal of residual solvent prior to use. The PLGA scaffolds were characterized with a thermogravimetric analyzer (TA instruments) to validate the complete removal of DCM (results not shown) and thickness gauge (Elcometer) to measure the thickness of the fabricated film.

Fabrication of patterned elastomeric stamp Elastomeric stamps made of polydimethylsiloxane (PDMS) were used to transfer print the ECM protein, human plasma derived fibronectin (BD, Biosciences) onto the PLGA scaffolds. The silicon master templates bearing the desired topographic features were fabricated via standard photolithography method. Briefly, a layer of photo resist was spin coated onto a 4′ silicon wafer and subsequently exposed to UV light through a patterned chrome mask. The non-cross linked polymer is then removed leaving behind the micro scale features. Liquid PDMS (1: 10 elastomeric base: curing agent, Sylgard 184, Dow Corning) was poured over the master, cured at 100 °C for 2 h and peeled away by hand. The PDMS stamp is then inked with fibronectin dissolved in phosphate buffer saline (PBS) solution (50 μg/ml) for 1 h, blown dry with pressurized purified nitrogen gas and placed in conformal contact with the PLGA substrate for ∼ 10 s. The non-printed region was then passivated with the tri block copolymer comprising of polypropylene (PPO)-poly ethylene oxide (PEO)-poly propylene (PPO) (Pluronic F 127, BASF) for 1 h at 37 °C to yield a non-biofouling surface. The substrate was then washed several times with PBS to remove the excess F 127 prior to cell seeding. Printed fibronectin lanes were immuno-labeled with rabbit polyclonal anti fibronectin antibody (Sigma) and counter stained with Cy3 (Molecular probes).

Cell culture hMSCs from an 18-year-old male donor and cell culture medium were both obtained from Lonza (Cambrex). The hMSCs expressed CD 105/+, CD166/+, CD 29/+, CD 44/+, CD 14/−, CD34/− and CD45/−. Information of the surface antigens were obtained via flow cytometry analysis provided in the company's data sheet. hMSCs were expanded and cultured in mesenchymal stem cell basal medium according to the vendor's instruction. For the micropatterning experiments, hMSCs were cultured in lowglucose Dulbecco's Modified Eagle's Medium (DMEM) containing L-glutamine (Sigma Aldrich) supplemented with 10% FBS (PAA) and 1% antibiotic/antimycotic solution (PAA). The cells were maintained at 37 °C in a humidified atmosphere of 5% CO2. Culture medium was changed every 2–3 days. PBS and 0.05% trypsin-EDTA obtained from Invitrogen were used for washing and cell detachment purposes respectively.

Micropatterning of hMSCs The printed and unprinted PLGA scaffolds (1 × 1 cm) were seated at the bottom of 24 well plates. hMSCs were seeded at a density of 2–3 × 103 cells per well. At the end of 1 1/2 h, unattached cells

E XP E RI ME N T AL C E L L R E S EA RC H 31 6 ( 20 1 0) 1 1 5 9– 1 16 8

were washed away with PBS followed by introduction of fresh cell culture medium. Culture medium was changed every 2–3 days.

Immunocytochemistry, microscopy and morphological analysis The actin cytoskeleton, focal adhesion protein, vinculin and cell nucleus was immuno-stained with actin cytoskeleton/ focal adhesion staining kit (Chemicon) according to the manufacturer's instructions. At various time points, cells were fixed with 4% paraformaldehyde for 15–20 min at 4 °C. These cells were then permeabilized with 0.1% Triton X-100 in PBS for 1–5 min at room temperature and blocked with 1% bovine serum albumin (BSA) in PBS (PAA) before the subcellular components were immunolabeled. Filamentous actin (F-actin), vinculin and cell nucleus were counter stained with Tetramethyl Rhodamine Iso-Thiocyanate (TRITC) conjugated-phalloidin (1:400, Chemicon), mouse monoclonal anti vinculin (1:400, Chemicon) counterstained with AlexaFluor 488 goat anti mouse IgG (Invitrogen) and 4′-6Diamidino-2-phenylindole (DAPI) respectively (Chemicon). Fluorescence images were visualized with a Nikon 80i eclipse (Nikon) upright microscope and captured with the Nikon DS-Fi1 (Nikon) using a 20X objective lens. Several parameters were introduced to quantify the morphological differences observed between the two experimental groups. Useful indicators such as the cell spreading area, bipolarity index (BI), 2D projection of cell nucleus area and nucleus shape index (NSI) were processed and analyzed with imageJ (rsbweb.nih.gov/ij/). The BI and NSI can be determined using the following expressions. BI ¼

Cell length Cell width

ð1Þ

[A higher BI value indicate a more elongated cell] sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  2 AB ð0VNSI < 1Þ NSI = 1 − CD

ð2Þ

where AB and CD refer to the lengths of the minor and major axes of the cell nucleus, respectively. [NSI closer to 0 indicates a rounder nucleus; NSI closer to 1 indicates a more elliptical nucleus]. 40 cells per experimental group from 3 separate images were analyzed to compute the mean and standard deviation of the respective morphological indices.

1161

amplified product. Primers specific to the targeted genes (Table 1) were obtained from primerbank (http://pga.mgh. harvard.edu/primerbank/) an online public resource for PCR primers. Specificity of the chosen primers to the gene of interest was examined by performing a BLAST (Basic Local Alignment Search Tool) search. Melt curve analysis was included to assure that only one PCR product was formed. Relative quantification of gene expression was analyzed with the Relative Expression Software Tool (REST) software [20]. Results presented are fold change expression normalized against the calibrators: GAPDH (glyceraldehyde-3-phosphate dehydrogenase) and β-actin as the endogenous housekeeping genes.

Dual immuno-fluorescent labeling of myogenic and neurogenic differentiation markers Cells at the respective time points were fixed with 4% paraformaldehyde in PBS for 5 min at room temperature, washed 3 times with PBS–0.02% tween and permeabilized with PBS–0.1% triton-X 100 at room temperature for 10 min. After three washes with PBS– 0.02% tween, the cells were blocked with 5% normal goat serum (Vector Labs) for 1 h at room temperature followed by addition of the primary antibodies. The primary antibodies used are as follows: mouse monoclonal anti cardiac myosin heavy chain (ab15) (1:100, abcam) and chicken polyclonal anti MAP2 (ab5392) (1:10000, abcam). Alexa Fluor secondary antibodies 488 goat anti mouse IgG, 568 goat anti chicken IgG, (1:400, Molecular probes, Invitrogen), were used. Samples were incubated with the respective secondary antibodies for 1 h at room temperature. Neural stem cell line (Millipore) and C2C12 murine myoblast cell line (ATCC) was used as positive controls. Negative controls (in the absence of primary antibodies) were performed in all cases to validate specific bindings of the secondary antibodies. Nuclei were counter stained with DAPI (Chemicon). Images of the immune-labeled cells were captured with either the Nikon eclipse 80i upright microscope.

Statistical analysis All data are presented as mean ± standard deviation (SD). Significant differences (p < 0.01) between experimental groups were determined with Student t-test.

Quantitative real time polymerase chain reaction (qRT-PCR)

Results

Total RNA (tRNA) was extracted with RNeasy mini kit (Qiagen). The concentration and quality of the extracted tRNA was determined spectrophotometrically (Nano drop-N100, Thermo Scientific). First strand complementary DNA (cDNA) was synthesized with oligo dT and ImProm-II™ reverse transcription system (Promega) according to the manufacturer's protocol. qRT-PCR was performed on a CFX96 real time PCR detection system (BioRad Laboratories, Inc, USA). The reaction mix comprises of 10 μl of KAPA SYBR FAST master mix (2×) universal, 0.4 μl of forward and reverse primers (10 μM) respectively, 1 μl of diluted cDNA and PCR grade water to make up a final volume of 20 μl. The following cycling thermal profile was employed: enzyme activation at 95 °C for 3 min, 45 cycles of 95 °C for 3 s and 60 °C for 20 s followed by a dissociation step to analysis the melt curve of the

Bone marrow MSCs adopted drastically different morphology and cytoskeletal configuration when plated on cell-shape defining substrates Micro features on the elastomeric pDMS stamps were validated and characterized using conventional phase contrast microscopy. As shown in Figs. 1A and B, the fabricated micro features on the PDMS surface comprises of 40 μm wide and 1 μm deep grooves separated by elevated lanes of 20 μm in width. Highly regular and spatially defined immuno-labeled fibronectin adhesive strips on PLGA substrate are shown in Fig. 1C. Typically, the printed lanes are of 20 μm in width separated by a gap width of 40 μm nonadhesive region. Attachment of hMSCs onto the printed proteins can be observed as early as 5 min after cell seeding using time

1162

E XP E RI ME N T AL C E L L R E SE A RC H 31 6 ( 20 1 0) 1 1 59 – 1 16 8

Table 1 – Compiled list of gene targets probed in this study. Gene bank accession number NM_001101 NM_002046 NM_004348 BC021289 NM_003118 NM_000711 NM_002052 NM_002478 NM_000257 NM_000364 X65964 NM_002500 NM_002374 NM_002055

Gene target

Sequence (5′-3′)

Amplicon length (bp)

β-actin GAPDH Runt related transcription factor 2 (RUNX2) Akaline phosphotase (ALPL) Secreted protein, acidic, cysteine-rich (SPARC) (osteonectin) Osteocalcin (OCal) GATA-4 Myoblast differentiation protein 1 (MyoD1) β-Myosin heavy chain (β-MHC) Cardiac troponin T (cTnT) Nestin (NES) Neurogenic differentiation 1 (NeuroD1) Microtubule Associated protein (MAP2) Glial fibrillary acidic protein (GFAP)

CATGTACGTTGCTATCCAGGC CTCCTTAATGTCACGCACGAT CATGAGAAGTATGACAACAGCCT AGTCCTTCCACGATACCAAAGT TCCTATGACCAGTCTTACCCCT GGCTCTTCTTACTGAGAGTGGAA

250 113 190

CTCTCCAAGACGTACAACACC AATGCCCACAGATTTCCCAGC AGCACCCCATTGACGGGTA GGTCACAGGTCTCGAAAAAGC

201 105

CACTCCTCGCCCTATTGGC GCCTGGGTCTCTTCACTACCT CCCAGACGTTCTCAGTCAGTG GCTGTTCCAAGAGTCCTGCT CGGCGGAACTGCTACGAAG GCGACTCAGAAGGCACGTC

138 146 172

CACTGATAACGCTTTTGATGTGC TAGGCAGACTTGTCAGCCTCT TCTCCGAAACAGGATCAACGA GCCCGGTGACTTTAGCCTT CAACAGCGACGGAGGTCTC CCTCTACGCTCTCTTCTTTGAGT GCCTTGCTATTCTAAGACGCA GTGGGTTGGGATAAGCCCTT CAGGAATTGACTCCCTCTACAGC TCTTCACCAGGCTTACTTTGC ATCGAGAAGGTTCGCTTCCTG TGTTGGCGGTGAGTTGATCG

165 68 163 156 80 151

lapse video microscope (not shown) and the cells adopted an elongated shape exclusively within the printed regions (Fig. 1D). Apart from the visible differences in cell shape, the cytoskeletal arrangement, focal adhesions distribution (as represented by localization of the integrin associated adapter protein, vinculin) and nuclei were immuno-stained in order to investigate how the aforementioned subcellular components were influenced. Generally, the unpatterned cells were well spread, with thick bundles of well defined F-actin or stress fibers observed. The stress fibers were either aligned in a single direction or arranged circumferentially around the cell nucleus (Fig. 2A). Distinct vinculin staining showed that focal adhesion complexes were localized mostly at the cell periphery, at the ends of stress fibers or ventral to the cell

nucleus. DAPI staining showed that cell nuclei remained round (Fig. 2A). Patterned cells exhibited an elongated morphology. Stress fibers aligned parallel along the longitudinal axis of the cells and were less developed compared to the cells cultured on the unpatterned substrate. Vinculin appeared to be less defined compared to the unpatterned samples. Notably, cell nuclei were highly elliptical (Figs. 2B and C). Taken together, hMSCs were coerced to adopt a highly stretched and thin morphology that conformed to the printed fibronectin strips (Figs. 2B and C) By day 14, the cells had become confluent as shown in Fig. 3. hMSCs cultured on the unpatterned substrate assumed a spread morphology with distinct well developed stress fibers (Figs. 3A

Fig. 1 – Top (A) and cross sectional view (scale bar = 100 mm) (B) of the fabricated PDMS master stamp consisting of micro sized lanes of 20 μm width separated by 40 μm wide grooves (scale bar = 25 μm). Height of the elevated lanes (black arrow) is 1–2 μm as measured with imageJ. Transferred printed lanes of fibronectin on PLGA scaffold were labeled with anti-fibronectin antibody and fluorescently stained with cy3 (C) (scale bar = 100 μm). Functionality of the printed proteins was confirmed by culturing hMSCs on the scaffold (D). Image of patterned cells was acquired 4 days after cell seeding. (Scale bar = 100 μm).

E XP E RI ME N T AL C E L L R E S EA RC H 31 6 ( 20 1 0) 1 1 5 9– 1 16 8

1163

Fig. 2 – TRITC-Phalloidin labeled F-actin (red), AlexaFluor 488 labeled vinculin (green), DAPI nuclear staining (blue) and overlaid fluorescent image of immuno-stained cellular components (merged) for the unpatterned (A) and patterned hMSCs (B and C). Samples were cultured in DMEM supplemented with 10% FBS and 1% antibiotic/antimycotic solution for 4 days before they were fixed and stained. All images were taken with a 20 × objective lens. (Scale bar = 100 μm).

and B). The extent of cell spreading was greater for the samples when compared with cells on day 4. Regions of localized alignment of hMSCs can be observed on the unpatterned substrate as depicted in Fig. 3A. In contrast, the patterned cells exhibited a higher degree of alignment and were more compacted in the transverse direction, giving rise to higher cell numbers per unit area (Fig. 3C). For both experimental groups, majority of the vinculin staining was localized to apical end of the cells (Fig. 3).

Cellular morphological indices analysis Statistical analysis was conducted to quantify the morphological differences for the cells belonging to the two experimental groups. The spreading areas of the cell, BI, 2D projection of the cell nucleus area as well as the NSI were assessed. By day 4, cells that were cultured on the unpatterned substrate adopted a larger spreading area of 11212.5 ± 5494.7 μm2, a BI value of ∼ 2.1 indicative of a weak bipolarity correlation (Figs. 4A and B), a large nucleus area of 415.0 ± 150.5 μm2 and an NSI value of ∼ 0.65 (Figs. 4C and D). On the other hand, the morphological indices for the patterned cells contrasted starkly to that of the unpatterned cells. The patterned cells adopted a much smaller spreading area of 2167.1 ± 945.1 μm2 (∼80% reduction in cell area), a BI value of ∼ 13.6 suggesting that the cells adopted a more elongated morphology relative to the unpatterned cells, a projected cell nucleus area of 236.3 ± 81.8 μm2 and NSI ∼0.80, indicating that cell nuclei were highly elliptical (Fig. 4). Morphological indices obtained on day 14 revealed a similar trend. Altogether, hMSCs cultured on micropatterned

surface exhibited smaller spreading area, lower CSI value, smaller and elliptical nuclei when compared with hMSCs cultured on unpatterned surface.

Enforced cell shape upregulates myogenic and neurogenic gene expression of hMSCs Myogenic and neurogenic genes To obtain clues on whether the altered cell morphology was accompanied by stem cell differentiation, we performed qRT-PCR of selected genes known to be involved in myogenesis, osteogenesis and neurogenesis. The basic helix-loop-helix (bHLH) transcription factors GATA4 and MyoD1 are key early regulators of cardiomyogenesis [21–23]. The expressions of these two genes were up-regulated in the patterned samples when compared with unpatterned samples. The product of β-MHC and cTnT genes is essential for the normal function of the contractile machinery of skeletal and cardiac muscle [24]. In addition, expression of the β-MHC gene is in part regulated by GATA 4 [25] during cardiomyogenesis. In concordance, the mRNA expression level of β-MHC and cTnT were increased after day 14 of subculturing on patterned surface when compared to unpatterned surface (Fig. 5C). Overall, mRNAs that are normally associated with myogenesis were all up regulated for the patterned cells. NeuroD1 and nestin are both useful early markers for pro neural cells preparing for neural development. NeuroD1 is the first of the NeuroD family (D1–D3) to be expressed during neural

1164

E XP E RI ME N T AL C E L L R E SE A RC H 31 6 ( 20 1 0) 1 1 59 – 1 16 8

Fig. 3 – Phalloidin labeled F-actin (red), AlexaFluor 488 labeled vinculin (green), DAPI nuclear staining (blue) and overlaid fluorescent image of immuno-stained cellular components (merged) for the unpatterned (A and B) and patterned hMSCs (C). Samples were cultured in DMEM supplemented with 10% FBS and 1% antibiotic/ antimycotic solution for 14 days before they were fixed and stained. All images were taken with a 20 × objective lens. (Scale bar = 100 μm).

development [26] while nestin is a transient neurogenic marker for multipotent neural stem cell [27]. Both the chosen neural early markers were up regulated on the 4th and 14th day of culture with the NeuroD1 gene showing the greatest increment at day 14 (∼ 3.2 fold) while nestin remained up regulated at a consistent level of ∼2 times on both days (Fig. 5B). GFAP is a type III intermediate filament found in mature astrocytes in the central nervous system (CNS) [27] and the MAP2 gene encodes the microtubule associated protein which is a marker for mature neurons believed to be essential for the formation and maintenance of neurites [28]. Gene expression study revealed that hMSCs patterned in an elongated/ spindle like shape consistently displayed a higher expression for neurogenic markers (Fig. 5B). Similarly the GFAP gene and MAP2 genes were both up regulated clearly demonstrating that genes associated with neurogenesis can be modulated by controlling the cell shape.

Reduced osteogenic-related genes in hMSCs cultured on patterned surface Alkaline phosphatase, a tissue non-specific isozyme encoded by the ALPL gene (tissue non-specific precursor) is believed to be critical for bone matrix mineralization [29]. By day 4, ALPL transcription level for the patterned hMSCs was reduced by 7.3 times and maintained at its diminished level till the 14th day of cell culture (Fig. 5A). RUNX2, a basic helix loop helix (bHLH) transcription factor is a key regulator of osteocalcin, is also indicative of osteogenic differentiation as osteocalcin is expressed at high level from the onset of bone mineralization [30–32]. At

day 4, RUNX2 transcript level was ∼4.7 times lower than the nominal level while differential expression of osteocalcin for both patterned and unpatterned samples was marginal. By the 14th day of cell culture, there was an increase of RUNX2. For other markers, there are either no or not statistically significant upregulation as compared to control. Taken together, expression of osteogenic genes were either down modulated or maintained near nominal level for cells that adopts an elongated or spindlelike morphology.

Micropattern directs hMSCs differentiation into myocyte like cells As in the case of the micropatterned cells, we have observed the upregulation of both neuronal and myocardial mRNA transcripts level relative to the control samples. Therefore it will be of interest to determine if the increase in mRNA transcripts level is significant enough to drive differentiation towards the particular tissue lineage at the protein level and whether differentiation markers belonging to different tissue lineages can co-exist within the same cell. Experiments were designed specifically to gain insights into the aforementioned questions. Mature markers of myogenesis, cardiac MHC and neurogenesis MAP2 were chosen to be examined via immunocytochemistry because expression of either marker is specific and distinct to the respective lineage (Fig. 6, panels A to D). Interestingly, significant expression of cardiac MHC was only detected for the micropatterned cells but not the unpatterned cells at day 4 of

E XP E RI ME N T AL C E L L R E S EA RC H 31 6 ( 20 1 0) 1 1 5 9– 1 16 8

1165

Fig. 4 – Morphological parameters such as the cell spreading area (A), BI (B), 2D projection of cell nucleus area (C) and NSI were examined from fluorescent images of hMSCs cultured under two different in vitro conditions (Unpatterned cells : white bars, patterned cells : Shaded bars) for 4 and 14 days. Bars represent mean ± standard deviation derived from 40 data points (n = 40) per experimental groups. Data were collected from at least two experiments. Single and double asterisk indicates statistical significance difference of p < 0.01 and p < 0.05, respectively, between the two experimental groups.

cell culture. By day 14, hMSCs cultured on the unpatterned substrata approaches confluence, weak cardiac MHC expression can be observed for a sub population of the cells adopting elongated morphology. MAP2 expression was not detected for both the patterned and the unpatterned hMSCs at all time points. Performance of the cardiac MHC and MAP2 antibodies was validated using C2C12 cells (panel E) and neural stem cells respectively (panel F). Collectively, upregulation of β-MHC gene level and protein expression of late molecular signature associated with myogenesis suggests that the micropatterned cells are directed towards the myogenic lineage at the time scale being examined.

Discussion hMSCs have the capability to differentiate into numerous cells belonging to the three germ layers and this property makes them a valuable cell source in the field of regenerative medicine. In recent years we have seen a surge in the number of publications utilizing chemicals, biological factors and nanotechnologies to regulate stem cell differentiation [8–14]. Regardless of the methods employed, cell shape alteration is often observed leading up to the resultant cell fate, suggesting that cell shape may be a driving factor in stem cell differentiation. Indeed, our study herein showed

that hMSCs coerced to adopt and proliferate on a micropatterned fibronectin-PLGA surface favored the up-regulation of genes associated with neurogenesis and myocardiogenesis, even when cultured in non-differentiating condition. hMSCs cultured on the micropatterned surface are highly aligned and stretched, with much smaller cell spreading area compared to the cells cultured on surface with uniformly distributed fibronectin presentation. Within minutes of cell seeding, hMSCs form focal adhesions on the substrate which they are cultured on. Micro-printed lanes of fibronectin control the spatial distribution of these focal adhesions, confining the cells to attach and grow within these regions resulting in large scale deformation of the cells. The scale of distortion is then conveyed to the rest of the cell as a result of mechanical continuity between the focal adhesions, cytoskeleton and cell nucleus scaffold as proposed by the tensegrity model for cells [33–35]. It is known that the cytoskeleton plays an important role in signal transduction. Changes made to the spatial distribution or cytoskeletal re-organization can therefore provide a physical impetus by which mechanical deformation is translated into biochemical responses [35]. One proposed mechanism for this is via the deformation of the cell nucleus which is physically linked to the cytoskeleton by intermediate filaments. Such deformations has been shown to alter gene transcription profiles and thereby modify cellular functions [36].

1166

E XP E RI ME N T AL C E L L R E SE A RC H 31 6 ( 20 1 0) 1 1 59 – 1 16 8

Fig. 5 – Relative gene level expression analyses of osteogenic (A), neurogenic (B) and myogenic (C) markers. White bar and shaded bar represent samples retrieved on the 4th and 14th day of culture respectively. Bars represent mean ± standard deviation. Relative gene expression levels (I.e. ΔΔCT) were normalized to the control samples indicated by the dashed line in red. The asterisks indicates statistical significance (p < 0.05) difference between the two experimental groups (patterned and unpatterned samples).

Progress of stem cell differentiation can also be inferred from the mechanical properties of the nucleus suggesting that the nucleus can potentially function as an important mechanosensor or modulator of cell differentiation [37]. Several mechanisms of nuclear mechanotransduction have been proposed [36]. For instance, mechanical tugging of the nucleus membrane can directly affect gene activity via altering differential accessibility of the transcriptional factors to the genes and also modulating the nuclear pore size thus influencing regulatory molecules (e.g. mRNA, DNA) or ion flux. However, it seems that alterations to nuclei shape plays a lesser role in guiding hMSCs differentiation into the lineages examined here since similarities in the NSI did not correlate well with the up-regulation of gene expression at day 14. As the cells proliferated, they were able to secrete their own ECM proteins and desorption of F127 (anti fouling agent) may take place. On top of that, cell confluence may also limit the elongation of cells. As a result, the measured cell nucleus indices for the patterned cells after the 14th day of cell culture is significantly lesser (less elliptical) as compared to the cells on the 4th day of cell culture.

The ability of the micropatterned substrate to direct hMSCs lineage specification towards neurogenesis and myogenesis was further ascertained by immuno-detection of lineage-specific differentiation markers such as cardiac MHC (myogenic) as well as MAP2 (neurogenic). While negligible expression was observed for markers associated with neurogenesis (i.e. MAP2), significant expression of the myogenic mature marker cardiac MHC was perceived only for the micropatterned cells. Collectively, with results obtained for the gene expression study, we reasoned that the stretched and longitudinal cell shape appears to be conducive for both myogenic and neurogenic differentiation gene expression. However, at the functionally relevant level of protein expression, the myogenic lineage is dominant within the time scale studied. This area opens up a very interesting extension to our project that post transcriptional control drives hMSCs in our patterning system towards a predominant myogenic lineage at the time scale we looked at. Indeed, the unique mechanism by which shape can induce the differentiation of stem cells is an area that warrants future studies.

E XP E RI ME N T AL C E L L R E S EA RC H 31 6 ( 20 1 0) 1 1 5 9– 1 16 8

1167

Fig. 6 – Dual immuno-detection of myogenic and neurogenic differentiation markers. Cells were costained with distinctive and mature markers of myogenesis and neurogenesis using mouse monoclonal anti heavy chain cardiac myosin (green) and chicken polyclonal anti MAP2 (red) (panels A–D). Experiments were performed in triplicate for each experimental group. Cells were cultured in low glucose DMEM supplemented with 10% FBS and 1% antibiotic/antimycotic solution. Positive controls were used to verify the specificity of the primary antibodies. C2C12 were induced to differentiate in 2% horse serum and stained for cardiac MHC (panel E). Neural stem cells were induced to differentiate in DMEM F-12 and stained for MAP2 (panel F). (Scale bar = 100 μm).

Taken together, we have demonstrated for the first time that cellular shape modulation in the absence of any induction media may be a viable strategy to coax lineage-specific differentiation of stem cells. Immunocytochemical analysis revealed that cells cultured on the micropatterned substrata favor the differentiation of hMSCs into myocyte like cells. This novel approach that induced expression of differentiation markers not only complements present stem cell differentiation procedure, the biocompatibility and biodegradable properties of PLGA also makes it suitable for tissue engineering applications [38–40]. The molecular events responsible for the observed phenomenon would be of interest for future studies and from an application standpoint, micro patterning hMSCs on PLGA scaffold could enable large scale alignment of cells required for certain applications like cardiac muscle tissue engineering, among others.

Acknowledgments The authors would like to thank A⁎Star (SERC Grant No: 072 101 0021), Singapore, NUS Academic Research Fund (R-364-000-089112) and the Lee Kuan Yew Fellowship (to D.L.) for their financial support.

REFERENCES

[1] C. Conrad, R. Huss, Adult stem cell lines in regenerative medicine and reconstructive surgery, J. Surg. Res. 124 (2005) 201–208. [2] C. Metallo, J. Mohr, C. Detzel, J.d. Pablo, B.V. Wie, S. Palecek, Engineering the stem cell microenvironment, Biotechnol. Prog. 23 (2007) 18–23.

[3] M. Lutolf, J. Hubbell, Synthetic biomaterials as instructive extracellular microenvironments for morphogenesis in tissue engineering, Nat. Biotechnol. 23 (2005) 47–55. [4] K. Claffey, W. Wilkison, B. Spiegelman, Vascular endothelial growth factor. Regulation by cell differentiation and activated second messenger pathways, J. Biol. Chem. 267 (1992) 16317–16322. [5] A. Polesskaya, P. Seale, M. Rudnicki, Wnt signaling induces the myogenic specification of resident CD45+ stem cells during muscle regeneration, Cell (2003) 113. [6] S. Neves, P. Tsokas, A. Sarkar, E. Grace, P. Rangamani, S. Taubenfeld, C. Alberini, J. Schaff, R. Blitzer, I. Moraru, R. Iyengar, Cell shape and negative links in regulatory motifs together control spatial information flow in signaling networks, Cell (2008) 133. [7] F. Guilak, D. Cohen, B. Estes, J. Gimble, W. Liedtke, C. Chen, Control of stem cell fate by physical interactions with the extracellular matrix, cell stem cell 5 (2009) 17–26. [8] J. Park, J. Chu, C. Cheng, F. Chen, D. Chen, S. Li, Differential effects of equiaxial and uniaxial strain on mesenchymal stem cells, Biotechnol. Bioeng. 88 (2004) 359–368. [9] K. Kurpinski, J. Chu, C. Hashi, S. Li, Anisotropic mechanosensing by mesenchymal stem cells, Proc. Nat. Acad. Sci. U.S.A. 103 (2006) 16095–16100. [10] S. Park, S. Park, S. Namgung, B. Kim, J. Im, J. Kim, K. Sun, K. Lee, J. Nam, Y. Park, S. Hong, Carbon nanotube monolayer patterns for directed growth of mesenchymal stem cells, Adv. Mater. 19 (2007) 2530–2534. [11] E. Lim, S. Pang, K. Leong, Synthetic nanostructures inducing differentiation of human mesenchymal stem cells into neuronal lineage, Exp. Cell Res. 313 (2007) 1820–1829. [12] C.Y. Tay, H. Gu, W.S. Leong, H. Yu, H.Q. Li, B.C. Heng, H. Tantang, S.C.J. Loo, L.J. Li, L.P. Tan, Cellular behavior of human mesenchymal stem cells cultured on single-walled carbon nanotube film, Carbon 48 (2010) 1095–1104. [13] A. Engler, S. Sen, H. Sweeney, D. Discher, Matrix elasticity directs stem cell lineage specification, Cell 126 (2006) 677–689.

1168

E XP E RI ME N T AL C E L L R E SE A RC H 31 6 ( 20 1 0) 1 1 59 – 1 16 8

[14] R. McBeath, D. Pirone, C. Nelson, K. Bhadriraju, C.S. Chen, Cell shape, cytoskeletal tension, and RhoA regulate stem cell lineage commitment, Dev. Cell (2004) 6. [15] J. Sanchez-Ramos, S. Song, F. Cardozo-Pelaezb, C. Hazzi, T. Stedeford, A. Willingd, T.B. Freemand, S. Saporta, W. Janssen, N. Patel, D.R. Cooper, P.R. Sanberg, Adult bone marrow stromal cells differentiate into neural cells in vitro exp, Neurol. 164 (2000) 247–256. [16] W.Y. Yeong, H. Yu, K.P. Lim, K.L.G. Ng, Y.C.F. Boey, S.S. Venkatraman, L.P. Tan, Multi-scale topological guidance for cell alignment via direct laser writing on biodegradable polymer, Tissue Eng. Part C (2010), http://www.liebertonline.com/doi/ abs/10.1089/ten.tec.2009.0604. [17] Y. Yang, N. Relan, D. Przywara, L. Schuger, Embryonic mesenchymal cells share the potential for smooth muscle differentiation: myogenesis is controlled by the cell's shape, Development 126 (1999) 3027–3033. [18] F. Li, B. Li, Q. Wang, J.H. Wang, Cell shape regulates collagen type I expression in human tendon fibroblasts, Cell Motil. Cytoskeleton 65 (2008) 332–341. [19] R.G. Thakar, Q. Cheng, S. Patel, J. Chu, M. Nasir, D. Liepmann, K. Komvopoulos, S. Li, Cell-shape regulation of smooth muscle cell proliferation, Biophys. J. 96 (2009) 3423–3432. [20] M. Pfaffl, G. Horgan, L. Dempfle, Relative Expression Software Tool (REST) for group-wise comparison and statistical analysis of relative expression results in Real-Time PCR, Nucleic Acids Res. 30 (2002) E36. [21] Y. Lee, T. Shioi, H. Kasahara, S. Jobe, R. Wiese, B. Markham, S. Izumo, The cardiac tissue-restricted homeobox protein Csx/Nkx2.5 physically associates with the zinc finger protein GATA4 and cooperatively activates atrial natriuretic factor gene expression, Mol. Cell. Biol. 18 (1998) 3120–3129. [22] R. Patient, J. McGhee, The GATA family (vertebrates and invertebrates), Curr. Opin. Genet. Dev. 12 (2002) 416–422. [23] R. Davis, H. Weintraub, A. Lassar, Expression of a single transfected cDNA converts fibroblast to myoblasts, Cell 51 (1987) 987–1000. [24] L. Thierfelder, H. Watkins, C. Macrae, R. Lamnas, W. Mckenna, H. Vosberg, J. Seidman, C. Seidman, Alpha-tropomyosin and cardiac troponin-T mutations cause familial hypertrophic cardiomyopathy—a disease of the sarcomere, Cell 77 (1994) 701–712. [25] E. Morkin, Control of cardiac myosin heavy chain gene expression, Microsc. Res. Tech. 50 (2000) 522–531. [26] M.B. McCormick, R.M. Tamimi, L. Snider, A. Asakura, D. Bergstrom, S.J. Tapscott, NeuroD2 and neuroD3: distinct expression patterns and transcriptional activation potentials within the neuroD gene family, Mol. Cell. Biol. 16 (1996) 5792–5800.

[27] J. Dahlstrand, L. Zimmerman, R. McKay, U. Lendahl, Characterization of the human nestin gene reveals a close evolutionary relationship to neurofilaments, J. Cell Sci. 103 (1992) 589–597. [28] L. Dehmelt, S. Halpain, Actin and microtubules in neurite initiation: are MAPs the missing link, J. Neurobiol. 58 (2004) 18–33. [29] Y.F. Zhou, V. Sae-Lim, A.M. Chou, D.W. Hutmacher, T.M. Lim, Does seeding density affect in vitro mineral nodules formation in novel composite scaffolds, J. Biomed. Mater. Res., Part A 78A (2006) 183–193. [30] C. Banerjee, S.W. Hiebert, J.L. Stein, J.B. Lian, G.S. Stein, An AML-1 consensus sequence binds an osteoblast-specific complex and transcriptionally activates the osteocalcin gene, Proc. Natl. Acad. Sci. U. S. A. 93 (1996) 4968–4973. [31] C. Banerjee, L. McCabe, J. Choi, S. Hiebert, J. Stein, G. Stein, J. Lian, Runt homology domain proteins in osteoblast differentiation: AML3/CBFA1 is a major component of a bone-specific complex, J. Cell. Biochem. 66 (1997) 1–8. [32] J. Lian, G. Stein, J. Stein, A. vanWijnen, Osteocalcin gene promoter: unlocking the secrets for regulation of osteoblast growth and differentiation, J. Cell. Biochem. Suppl. 30 (1998) 62–72. [33] D. Ingber, Tensegrity-based mechanosensing from macro to micro, Prog. Biophys. Mol. Biol. 97 (2008) 163–179. [34] A. Maniotis, C. Chen, D. Ingber, Demonstration of mechanical connections between integrins cytoskeletal filaments, and nucleoplasm that stabilize nuclear structure, Proc. Natl. Acad. Sci. U. S. A. 94 (1997) 849–854. [35] D. Ingber, Tensegrity: the architectural basis of cellular mechanotransduction, Annu. Rev. Physiol. 59 (1997) 575–599. [36] N. Wang, J.D. Tytell, D.E. Ingber, Mechanotransduction at a distance: mechanically coupling the extracellular matrix with the nucleus, Nat. Rev. Mol. Cell Biol. 10 (2009) 75–82. [37] J. Pajerowski, K. Dahl, F.Z. FL, P. Sammak, D. Discher, Physical plasticity of the nucleus in stem cell differentiation, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 15619–15624. [38] M. Mishali, J. Zoldan, S. Levenberg, Effect of scaffold stiffness on myoblast differentiation, Tissue. Eng. Part. A 15 (2009) 935–944. [39] T. McDevitt, J. Angello, M. Whitney, H. Reinecke, S. Hauschka, C. Murry, P. Stayton, In vitro generation of differentiated cardiac myofibers on micropatterned laminin surfaces, J. Biomed. Mater. Res. 60 (2002) 472–479. [40] H. Park, M. Radisic, J. Lim, B. Chang, G. Novakovic, A novel composite scaffold for cardiac tissue engineering, In Vitro Cell. Dev. Biol. Anim. 41 (2005) 188–196.