lable at ScienceDirect

Biomaterials xxx (2013) 1e9

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

Multifunctional polymer scaffolds with adjustable pore sizeand chemoattractant gradients for studying cell matrix invasion

Alexandra M. Greiner a,1, Maria Jäckel a,b,1, Andrea C. Scheiwe c,1, Dimitar R. Stamowd,Tatjana J. Autenrieth b, Joerg Lahann b, Clemens M. Franz d,**, Martin Bastmeyer a,b,*aDepartment of Cell and Neurobiology, Karlsruhe Institute of Technology (KIT), Haid-und-Neu-Straße 9, 76131 Karlsruhe, Germanyb Institute of Functional Interfaces (IFG), Karlsruhe Institute of Technology (KIT), 76344 Eggenstein-Leopoldshafen, Germanyc Institute of Applied Physics (APH), Karlsruhe Institute of Technology (KIT), Wolfgang-Gaede-Straße 1, 76131 Karlsruhe, GermanydDFG-Center for Functional Nanostructures (CFN), Karlsruhe Institute of Technology (KIT), Wolfgang-Gaede-Str. 1a, 76131 Karlsruhe, Germany

a r t i c l e i n f o

Article history:Received 21 August 2013Accepted 24 September 2013Available online xxx

Keywords:BiocompatibilityCell adhesionChemotaxisLaser manufacturingPhotopolymerizationLamin A/C

* Corresponding author. Department of Cell and Ntute of Technology (KIT), Haid-und-Neu-Straße 9,Tel.: þ49 (0) 608 43085.** Corresponding author. DFG-Center for FunctiKarlsruhe Institute of Technology (KIT), Wolfgang-GaGermany. Tel.: þ49 (0) 608 45804.

E-mail addresses: clemens.franz@kit.edu (C.M.kit.edu (M. Bastmeyer).

1 These authors contributed equally.

0142-9612/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.biomaterials.2013.09.095

Please cite this article in press as: Greiner AMfor studying cell matrix invasion, Biomateri

a b s t r a c t

Transmigrating cells often need to deform cell body and nucleus to pass through micrometer-sized poresin extracellular matrix scaffolds. Furthermore, chemoattractive signals typically guide transmigration,but the precise interplay between mechanical constraints and signaling mechanisms during 3D matrixinvasion is incompletely understood and may differ between cell types. Here, we used Direct LaserWriting to fabricate 3D cell culture scaffolds with adjustable pore sizes (2e10 mm) on a microporouscarrier membrane for applying diffusible chemical gradients. Mouse embryonic fibroblasts invade 10 mmpore scaffolds even in absence of chemoattractant, but invasion is significantly enhanced by knockout oflamin A/C, a known regulator of cell nucleus stiffness. Nuclear stiffness thus constitutes a major obstacleto matrix invasion for fibroblasts, but chemotaxis signals are not essential. In contrast, epithelial A549cells do not enter 10 mm pores even when lamin A/C levels are reduced, but readily enter scaffolds withpores down to 7 mm in presence of chemoattractant (serum). Nuclear stiffness is therefore not a primeregulator of matrix invasion in epithelial cells, which instead require chemoattractive signals. Micro-structured scaffolds with adjustable pore size and diffusible chemical gradients are thus a valuable tool todissect cell-type specific mechanical and signaling aspects during matrix invasion.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Cells embedded in tissues are exposed to a complex three-dimensional (3D) and multifunctional microenvironment contain-ing multiple extracellular matrix (ECM) components, other cellpopulations and soluble or ECM-bound cellular signaling factors.Cells constantly modify and restructure this environment and oftendisplay directional migration along a concentration gradient ofbioactive signaling molecules [1,2]. For instance, directed migrationis crucial for wound healing and the immune response as it enables

eurobiology, Karlsruhe Insti-76131 Karlsruhe, Germany.

onal Nanostructures (CFN),ede-Str. 1a, 76131 Karlsruhe,

Franz), martin.bastmeyer@

All rights reserved.

, et al., Multifunctional polymals (2013), http://dx.doi.org/1

immune cells to reach their targets [3e5]. Furthermore, cell migra-tion through tissues is involved inmany pathological processes, suchas chronic inflammation and cancer cell metastasis [6,7]. In thesecases, different aspects of cell behavior are also strongly influencedby the ease with which cells are able to migrate through the tissuemicroenvironment. For instance, the malignancy of cancer is, inparts, determined by the degree to which tumor cells can invadeneighboring tissues or distant organs [8e10]. In many cell types,interphase nuclei are about two to ten times stiffer than the sur-rounding cytoplasm [7,11e17], and the nucleus has therefore beenconsidered the rate-limiting organelle for cell migration throughmatrix pores [15,18,19]. In agreement, highly invasive cells, such ascancer cellsebut also cells of the immune systemeoften feature softand flexible nuclei [20e22]. Understanding the molecular mecha-nisms regulating nuclearmechanics is therefore important for betterunderstanding tissue invasion and tumor metastasis [7,15,18].

The nuclear lamina underlying the inner nuclear membraneprovides a dominant contribution to the mechanics of the inter-phase nucleus. Lamins, a family of intermediate filament proteins,are an integral component of the nuclear lamina and important

er scaffolds with adjustable pore size and chemoattractant gradients0.1016/j.biomaterials.2013.09.095

A.M. Greiner et al. / Biomaterials xxx (2013) 1e92

regulators of nuclear stiffness [14,15,23e25]. They include thelamin A/C subtypes and connect the nucleus to the cytoskeleton viathe LINC (Linker of Nucleus and Cytoskeleton) complex which inturn can interact with actin-, vimentin-, and tubulin-based fila-ments [7,13e15,26]. If this linkage is disturbed in cells lacking laminA/C, cells often have softer and misshaped nuclei and increasednuclear fragility leading to decreased cell viability [7,13e15,17,24,27,28]. Diverse mutations in the lmna gene or in lamin-binding proteins cause a group of human diseases jointly referredto as laminopathies, including Emery-Dreifuss muscular dystrophyand premature aging syndromes [11,12,14,17,24,29,30], underliningthe importance of lamin A/C for proper cell mechanics andbehavior.

Given the structural and chemical complexity of cellematrixinteractions, it is often difficult to obtain experimental access andstudy these processes in vivo. Therefore, artificial 3D systemsmimicking the native cell surrounding are highly desirable tools forstudying and analyzing the behavior of single cells in defined mi-croenvironments. A number of approaches for designing andfabricating such artificial non-cytotoxic 3D scaffolds have beendeveloped [31]. Direct Laser Writing (DLW) is a particularly ver-satile approach for building highly-defined and freely-scalable 3Dstructures for cell culture purposes [31e33]. DLW employs two-photon polymerization (2PP) to construct freeform 3D polymerstructures attached to a carrier substrate [31e35]. Briefly, in thissingle-step technique a femto-second laser beam is focused into aphotosensitivematerial using a high numerical aperture lens. In thefocal volume of the laser beam simultaneous absorption of twophotons leads to excitation of photosensitive molecules. This leadsto a localized chemical polymerization event that is confined to thefocal volume of the laser due to the non-linearity of the 2PP process[31e35].

In this study, we have used DLW to fabricate 3D cell culturescaffolds with different pore sizes on glass substrates or on a

Fig. 1. Microporous 3D cell culture scaffolds produced by direct laser writing (DLW). Sascaffolds with mesh sizes of 2 mm, 5 mm, and 10 mm fabricated by DLW.

Please cite this article in press as: Greiner AM, et al., Multifunctional polymfor studying cell matrix invasion, Biomaterials (2013), http://dx.doi.org/1

microporous membrane. 3D cell culture scaffolds on microporousmembranes can be combined with an external chemoattractantgradient (fetal calf serum) which is applied across the microporousmembrane. We hypothesize that 3D scaffolds on microporousmembranes can be used to study both biochemical aspects(chemotaxis) and biophysical parameters (nuclear mechanics)regulating scaffold invasion of specific cell types in a single exper-imental setup.

2. Material and methods

2.1. Cell culture

A549 cells (human lung carcinoma cells) were purchased from the AmericanTissue Culture Collection (ATCC, Manassas, VA, USA) and cultured in F-12K medium(Gibco). Wildtype mouse embryonic fibroblasts (MEF WT) and lmna-/- MEFs (MEFlmna KO) were a generous gift from Colin Stewart (Astar, Singapore) and cultured inDulbecco’s Modified Eagle’s Medium (DMEM, Gibco). Cell culture media were sup-plemented with 10% fetal calf serum (FCS, HyClone). Cells were cultivated in anincubator at 37 �C with 5% CO2 in a humidified environment and passaged threetimes a week.

2.2. Transfection

A549 cells (3 � 105) were seeded in a culture flask 24 h prior of transfection.Lipofectamine 2000 (25 ml, Invitrogen) and 6.25 ml of a 10 mM lamin A/C siRNA(Ambion) or GFP-labeled control siRNA (Qiagen) solution were added separately tovials containing 600 ml F-12K medium and then incubated for 5 min. Both solutionswere mixed and incubated for further 20 min. After the cells were rinsed withphosphate buffered saline (PBS, Gibco), siRNA solution and F-12K medium wereadded, resulting in a total volume of 5 ml. After 24 h, cells were rinsed first with PBSand then with F-12K medium containing 10% FCS. Cells were used for experimentsafter another 2 days of incubation.

2.3. Nuclear stiffness measurements

Atomic force microscopy (AFM) cell elasticity measurements were performedusing a NanoWizard II AFM (JPK Instruments, Germany) mounted on top of aninverted optical microscope (Carl Zeiss Axio Observer A1) and a designated PetriDish holder. Cells were cultured for 24 h in glass bottom dishes, rinsed with PBS andadapted to CO2-independent medium (Invitrogen, Germany) for 1 h at 37 �C and

cnning electron microscopy (SEM) images of 3D pentaerythritol tetraacrylate (PETTA)

er scaffolds with adjustable pore size and chemoattractant gradients0.1016/j.biomaterials.2013.09.095

A.M. Greiner et al. / Biomaterials xxx (2013) 1e9 3

finally subjected to colloidal force spectroscopymeasurements within 1 h of transferto the AFM. Colloidal probes were prepared by gluing single silica beads (KiskerBiotech, Germany) with a diameter of 10 mm to tipless V-shaped silicon nitridecantilevers (NPeO, Veeco, Germany) having a nominal force constant of 0.12 N/m.Individual cells were measured by positioning the bead over the middle of the cellnucleus as monitored using the optical microscope and extending the closed-loop z-scanner with a constant piezo extension rate of 1 mm/s until reaching a pre-set forceof 2 nN. This typically corresponded to isotropic cell indentation depths of 0.5e1.0 mm. Colloidal probes were calibrated before and after each measurement usingthe thermal noise method. Cell elasticity values were calculated from the obtainedforceedistance curves by applying a Hertzian mechanics model [6], approximatingthat the sample is isotropic and linearly elastic. To ensure the applicability of theHertz model, only the initial 500 nm of indentation of each force curve were fitted.

2.4. Direct Laser Writing

3D scaffolds with different mesh sizes were fabricated on glass cover slips (Roth)using Direct Laser Writing (DLW), a two photon polymerization (2PP) technique, asdescribed elsewhere [36,37]. As photoresist pentaerythritol tetraacrylate (PETTA,SigmaeAldrich) containing 3% Irgacure 379 photoinitiator (Ciba) was used. To create3D scaffolds on porous membranes, cut-outs of a filter plate from a disposableBoyden chamber system (Neuroprobe) with 3 mm pore diameter were used. First adrop of photoresist was placed on a glass slide. The porous film was then placed ontop of the photoresist and covered by another drop of photoresist and a second glassslide, yielding a sandwich system and a flattened porous film. Scaffolds were thenwritten onto the porousmembrane in an upsideedownposition into the photoresistlayer adjacent to the porous membrane.

2.5. 3D cell culture experiments

PETTA scaffolds were sterilized with 70% ethanol (Roth), washed with PBS,coated with 10 mg/ml fibronectin (SigmaeAldrich) for 30 min, and rinsed againwithPBS. When using 3D scaffolds on glass slides, cells were seeded in medium con-taining 10% FCS. Incubation times were 16 h for A549 cells or 24 h for MEFs. Ex-periments with PETTA scaffolds on porous films, were performed using the wells ofthe disposable chemotaxis system. The lower well was filled with 30 ml F-12K me-dium containing 10% FCS. The porous film was then placed on top of the well con-taining medium supplemented with FCS. To avoid physical displacement of theporous film relative to the well, it was fixed with a polydimethylsiloxane (PDMS)ring and a small amount of vaseline (Roth) to the wall of the well. Inside the PDMSring 50 ml of cell suspension in F-12K medium without FCS was added and the cellswere incubated for 16 h under standard cell culture conditions.

2.6. Immunofluorescence

Cells were fixed for 10 min with 4% paraformaldehyde (SigmaeAldrich) in PBS.After permeabilization with PBS containing 0.1% Triton (SigmaeAldrich), cells wereincubated with phalloidin Alexa Fluor488 (Molecular Probes) at a dilution of 1:60,DAPI (Roth) at a dilution of 1:1000, or DRAQ5 (1:800, BioStatus Limited) in 1% BSA inPBS at room temperature for 1 h. Finally, samples were embedded in Mowiol(Hoechst) containing 1% N-propylgallate (SigmaeAldrich).

2.7. Microscopy

Samples were analyzed using a laser scanning microscope (LSM 510Meta, Zeiss)equipped with a 63 � /1.4 Oil DIC objective (Zeiss). 3D reconstructions of confocalimage stacks were prepared with the Volocity Software Version 4.3.2 (Perkin Elmer).Scanning electron microscopy (SEM) images were obtained with a Supra 55 scan-ning electron microscope (Zeiss).

Fig. 2. Atomic force microscopy (AFM) approach to measure the interphase nuclear stiffncell (I) and the cell nucleus is subsequently indented (II). (B) Measuring the Young’s modullmna KO) revealed significantly softer nuclei of MEF lmna KO cells (567 Pa) compared to nindependent experiments).

Please cite this article in press as: Greiner AM, et al., Multifunctional polymfor studying cell matrix invasion, Biomaterials (2013), http://dx.doi.org/1

2.8. Western blot

Western blotting was performed to identify lamin A/C knockdown levels inA549 cells. Blotting for glutaraldehyde 3-phosphate dehydrogenase (GAPDH) servedas a protein loading control. Briefly, total protein was isolated from cells in trizolreagent and the protein concentration was determined using the Bradford assay(Bio-Rad). Equal amounts of protein were homogenized and separated by electro-phoresis on 12% sodium dodecyl sulfate (SDS)-polyacrylamide gels. Proteins weretransferred to a polyvinylidene difluoride (PVDF) membrane, and the blot mem-branes were incubated in 5% non-fat milk in TBS buffer as blocking solution. Primaryantibodies were polyclonal rabbit lamin A/C (Santa Cruz Biotechnologies) and rabbitpolyclonal GAPDH (SigmaeAldrich). Blots were developed using a protein A/alkalinephosphatase-conjugated secondary antibody (SigmaeAldrich) strategy and devel-oped on the blotting membrane with 3% 4-Nitro blue tetrazolium chloride (NBT)/2%5-Bromo-4-chloro-3-indolyl-phosphate (BCIP) solution. Band densities weremeasured using the Image J software (http://rsb.info.nih.gov/ij/) and normalized toGAPDH levels.

2.9. Statistics

Nucleus stiffness (AFMmeasurements) and cell invasion data were expressed asmean� s.e.m. OriginPro 8G software (OriginLab Cooperation, Northampton, USA) orExcel (Microsoft) was used for statistical analysis (Chi square test after Brandt andSnedecor or t test). The type of test applied in each experiment is indicated in thefigure legend together with the information about the number of experiments ornumber of cells tested. Differences were considered statistically significant (*) whenthe calculated p value was <0.05.

3. Results

3.1. 3D cell culture scaffolds with adjustable mesh size

Cell matrix invasion is an essential part of many normal andpathological processes, including leukocyte extravasation andcancer cell metastasis [10,18,20,38]. A prerequisite for cells to enterand transmigrate through dense ECM scaffolds is their ability totranslocate through micrometer-sized pores formed by themacromolecular matrix network. Artificial cell culture scaffoldswith adjustable pore sizes are useful tools to elucidate mechanicalmechanisms underlying cell matrix invasion. To emulate a matrixwith different pore sizes, we used DLW to generate well-defined 3Dscaffolds with different mesh sizes on a glass carrier substrate(Fig. 1). DLW-fabricated microstructures are freely-scalable and weused this advantage to vary the scaffold grid dimension to yieldmesh sizes of 2 mm, 5 mm, and 10 mm. The resulting pore dimensionsare below typical diameters of mammalian cells and nuclei, therebycreating a physical obstacle to scaffold invasion similar to size-limited micropores in dense ECM networks [19,39]. The scaffoldswere produced from pentaerythritol tetraacrylate (PETTA), a non-toxic polymer previously used as a material for cell culture sub-strates [37]. For subsequent cell experiments, the microporousscaffolds were coated with fibronectin, an abundant ECMmolecule,to enhance cell attachment and scaffold invasion.

ess. (A) A bead glued to an AFM cantilever is positioned over an interphase nucleus of aus of wildtype mouse embryonic fibroblasts (MEF WT) and lmna knockout MEFs (MEFuclei of WT cells (777 Pa) (t test; *, p, <0.05; n >215 cells tested per cell type from 3

er scaffolds with adjustable pore size and chemoattractant gradients0.1016/j.biomaterials.2013.09.095

A.M. Greiner et al. / Biomaterials xxx (2013) 1e94

3.2. Lamin A/C knockout and nuclear stiffness of MEF cells

Cells migrating through the extracellular matrix duringembryogenesis, wound healing, or metastasis usually need todeform their cell body and nucleus to pass through micrometer-sized pores within the matrix scaffold. In fact, nuclear stiffnessmay constitute a major barrier to matrix invasion [7,8,14,19].Here, we intended to investigate specifically the role of nuclearstiffness on cell matrix invasion through size-limited pores.Several factors are known to affect the stiffness of nuclei,including chromatin content and the expression levels of the

Fig. 3. Fibroblast invasion into 3D scaffolds on glass. (A, B) Top and side view 3D reconst(MEF lmna KO) on 3D scaffolds with a mesh size of 10 mm. The cell nuclei of MEF WT andembedded within the 3D structure (nuclei ¼ red, F-actin ¼ green, 3D scaffold ¼white). (C) SiWT cells (chi square test; *, p <0.05; n >72 cells from 5 independent experiments) (For interpversion of this article.).

Please cite this article in press as: Greiner AM, et al., Multifunctional polymfor studying cell matrix invasion, Biomaterials (2013), http://dx.doi.org/1

class V intermediate filament lamin A/C [7,11,12,14,16,17]. Therole of lamin A/C in regulating cell matrix invasion is still underinvestigation [11,12,14,17]. To assess the effect of lamin A/C onnuclear stiffness and invasive behavior, we compared mouseembryonic fibroblast wildtype cells (MEF WT) and MEFknockout cells for the lamin A gene (MEF lmna KO). Deletion ofthe lamin A gene leads to loss of expression of both lamin A andC, since both proteins are encoded by a single gene and lamin Cis a splicing variant of the lamin A protein Refs. [28,40,41].Lamin A/C is a proposed regulator of nuclear stiffness [27] andwe therefore expected the nuclei of lamin A/C knockout cells to

ructions of wildtype mouse embryonic fibroblasts (MEF WT) and lmna knockout MEFsMEF lmna KO either localized on top of the 3D scaffold, or they were partially or fullygnificantly more cell nuclei were partially or fully embedded for the KO cell line than forretation of the references to color in this figure legend, the reader is referred to the web

er scaffolds with adjustable pore size and chemoattractant gradients0.1016/j.biomaterials.2013.09.095

A.M. Greiner et al. / Biomaterials xxx (2013) 1e9 5

be softer than those of MEF WT cells. We characterized thestiffness of interphase nuclei of MEF WT and MEF lmna KO cellsusing atomic force microscope (AFM) indentation experiments.In these experiments, a micron-sized bead glued to a V-shapedcantilever is positioned above the cell nucleus and then loweredto indent the nucleus (Fig. 2A) [27]. The elastic modulus wasderived from applying a Hertz model fit to the indentation forcecurve. The elasticity measurements confirmed that nuclei ofMEF lmna KO cells are significantly softer (567 Pa) than those ofMEF WT cells (777 Pa) (Fig. 2B). These results emphasized theimportance of lamin A/C for nuclear stiffness and demonstratedthat MEF lmna KO cells are a suitable model system for cellswith reduced nucleus stiffness.

Fig. 4. Invasion of epithelial cells into 3D scaffolds on glass. (A) Lamin A/C knockdown incontrol cells (Ctrl) or cells transfected with control siRNA (si-Ctrl) (t test; z, p >0.05; *, pmeasurements showed that the nucleus of A549 KD cells was significantly softer (280 Pa) thp <0.05; n >215 per cell type, cells from 3 independent experiments). (C) A549 WT and A549structure (nuclei ¼ red, F-actin ¼ green, 3D scaffold ¼ white) (For interpretation of the refarticle.).

Please cite this article in press as: Greiner AM, et al., Multifunctional polymfor studying cell matrix invasion, Biomaterials (2013), http://dx.doi.org/1

3.3. Invasion of lamin A/C knockout MEF cells into 3D scaffolds

To investigate whether reduced nucleus stiffness promotesscaffold invasion, MEF WT or MEF lmna KO cells were seeded ontop of the scaffolds and incubated for 24 h. Afterwards, cells werefixed and labeled with DRAQ5 to visualize nuclei and phalloidin tostain the actin cytoskeleton (Fig. 3A and B). The degree of scaffoldinvasion was assessed from confocal image stacks. A cell wasconsidered to have become “fully embedded” if its nucleus wascompletely inserted into the scaffold. Cells with partially insertednuclei into the scaffold were considered “partially embedded”,while cells with nuclei located completely outside the scaffoldwere termed “on top”. A fraction of MEF WT and MEF lmna KO

epithelial A549 cells (KD) halved the lamin A/C protein level compared to non-treated<0.05; n ¼ 4 independent blots per condition). (B) Atomic force microscopy (AFM)an the nucleus of Ctrl and si-Ctrl cells (490 Pa Ctrl, 450 Pa si-Ctrl) (t test; z, p >0.05; *,KD cells rarely invade 10 mm scaffolds, instead forming a dense cell layer on top of the

erences to color in this figure legend, the reader is referred to the web version of this

er scaffolds with adjustable pore size and chemoattractant gradients0.1016/j.biomaterials.2013.09.095

Fig. 5. 3D setup for chemically controlled cell invasion. (A) (i) Direct laser writing (DLW) was applied to fabricate non-cytotoxic 3D scaffolds on porous membranes. Briefly, theporous polymeric membrane was mounted in a photoresist and placed between two glass cover slips. A femto-second laser beam was focused onto the photosensitive liquidmaterial, excited photoinitiator molecules by two photon absorption and causing a highly localized chemical polymerization reaction strictly confined to the focal volume of thelaser. (ii) Side view of the experimental setup. The porous membrane equipped with the 3D scaffold is placed over a reservoir containing medium supplemented with fetal calfserum (FCS). The 3D scaffold is surrounded by a ring made of polydimethylsiloxane (PDMS) and covered with media containing no FCS. Cells are seeded onto the 3D scaffold and cellinvasion into the 3D scaffold is studied (iii) SEM overview, side view and top view images of a 3D scaffold with a mesh size of 10 mm resting on a membrane with a pore diameter of3 mm. (B) The majority of A549 wildtype (WT) and lamin A/C knockdown (KD) cells did not invade 3D woodpile scaffold written on glass (no chemoattractant). Instead, cell invasioninto the scaffold was achieved by using 3D structures fabricated on a porous membrane and exposure of this setup to a soluble chemical gradient (Boyden chamber approach). A549

A.M. Greiner et al. / Biomaterials xxx (2013) 1e96

Please cite this article in press as: Greiner AM, et al., Multifunctional polymer scaffolds with adjustable pore size and chemoattractant gradientsfor studying cell matrix invasion, Biomaterials (2013), http://dx.doi.org/10.1016/j.biomaterials.2013.09.095

A.M. Greiner et al. / Biomaterials xxx (2013) 1e9 7

cells invaded scaffolds with 10 mm pores (Fig. 3A and B). Impor-tantly, MEF WT and MEF lmna KO cells displayed significant dif-ferences in invasive behavior at this pore size. Whilew55% of MEFWT cells were able to enter the scaffold, w80% of MEF lmna KOcells were located partially or fully within the scaffold (Fig. 3C),indicating enhanced scaffold invasion of MEF lmna KO cells. Thedistribution of the remaining (partially or fully embedded) cellswas also significantly shifted between MEF WT and MEF lmna KOcells. A larger percentage of MEF lmna KO cells were fully orpartially embedded within the scaffold compared to WT cells(Fig. 3B and C). Thus, AFM and invasion experiments established alink between reduced nucleus stiffness and enhancedscaffold population. Enhanced scaffold invasion by MEF lmna KOcells also correlated with morphological changes of these cells.The nuclei of MEF lmna KO cells were smaller and deformed moreseverely upon invasion into the scaffold (Fig. 3B), indicating anadditional correlation between nuclear deformability and scaffoldinvasion.

3.4. Impact of lamin A/C levels and nuclear stiffness for epithelialcell invasion

The previous experiments established a link between lamin A/C-dependent nucleus stiffness and scaffold invasion in fibroblastcells. To complement these experiments with a different cell type,we investigated the effect of lamin A/C depletion on nuclearstiffness and scaffold invasion in an epithelial cell line. The A549cells are alveolar cancer cells of epithelial origin and arefrequently used as a model cell line with epithelial markers [42].In these cells we used a siRNA approach to reduce lamin A/C levelsand nuclear stiffness. Western blot analysis demonstrated aw50%reduction of lamin A/C protein levels in A549 knockdown (KD)cells compared to non-treated wildtype (WT) control (Ctrl) orcontrol siRNA-treated (si-Ctrl) cells (Fig. 4A). Analyzing nuclearstiffness by AFM revealed that interphase nuclei of A549 KD cellswere significantly softer (280 Pa) than nuclei of A549 WT cells(490 Pa Ctrl; 450 Pa si-Ctrl) (Fig. 4B). Thus, the partial lamin A/Cknockdown in epithelial A549 cells had a comparable relativeeffect on nuclear stiffness as the lamin A/C knockout in MEF cells.The clear correlation between protein levels and stiffness valuesconfirmed the role of lamin A/C for nuclear stiffness across celltypes. However, nuclei of A549 WT cells (490 Pa) were softer thannuclei of MEF WT cells (777 Pa). Likewise, the Young’s modulus ofA549 KD cell nuclei was about half the value (280 Pa) of MEF lmnaKO nuclei (567 Pa), demonstrating differences in absolute stiffnessbetween fibroblast and epithelial cell lines. Neither A549 WT norA549 KO cells invaded scaffolds with 2 mm and 5 mm pores(Supplemental Figure S1), indicating that these pore sizes arelimiting for cell invasion for both cell lines. After culturing A549WTor A549 KD cells on 10� 10 mm FN-modified scaffolds for 16 h,themajority of cells was located on top of the structure, while onlya low number of cells had migrated fully into the scaffold (Fig. 4C).Thus, in contrast to fibroblast cells, lamin A/C depletion did notenhance scaffold invasion of epithelial cells at 10 mm pore size.Likewise, A549 WT and A549 KD cells displayed no differences innuclear morphology residing on top of the woodpile scaffold(Fig. 4C). These results indicated that in epithelial cells nuclearstiffness plays a less important role in impeding pore invasioncompared to fibroblasts.

WT and A549 KD cells both invaded the composite scaffolds when FCS was present as a chesize (10 mm, 8 mm, 7 mm) but was significantly higher than into 3D scaffolds without chemoatglass; þ, p <0.05 compared to the KD on glass; n >83 cells from 4 independent experimentsmesh sizes (nuclei ¼ red, 3D scaffold ¼ white) (For interpretation of the references to colo

Please cite this article in press as: Greiner AM, et al., Multifunctional polymfor studying cell matrix invasion, Biomaterials (2013), http://dx.doi.org/1

3.5. A 3D setup for chemotaxis-driven cell invasion

The inability of A549 cells to invade scaffolds with large poresizes indicated that these cells require additional cues for matrixinvasion. It is commonly assumed that epithelial cancer cellmigration is directed by pre-existing spatial gradients of chemo-kines and growth factors in the target tissue [43]. To study suchchemotaxis effects in an in vitro model system, cell motility isoften stimulated by applying a chemoattractant gradient throughplanar 2D filters containing micrometer-sized pores (e.g., Boydenchamber assays). However, these planar systems do not reflect thethree-dimensionality of the in vivo cell tissue microenvironment.To incorporate a chemotaxis assays into our matrix invasionmodel, we combined our 3D scaffolds with a Boyden chambersystem. In this approach, the PETTA scaffolds were fabricated byDLW directly on commercially available microporous polymericfilms using a newly-developed sandwich approach (left scheme inFig. 5A). 3D structures with different mesh sizes (7 mm, 8 mm, and10 mm) were written onto a film with a pore size of 3 mm, yieldingmicroporous 2D/3D composite structures (SEM images in Fig. 5A)suitable for chemotaxis and invasion studies. After FN-coating, theporous film carrying the 3D scaffold was placed onto a reservoircontaining cell culture medium supplemented with fetal calfserum (FCS) as a chemoattractant (right scheme in Fig. 5A). Toachieve a chemoattractive gradient, the 2D/3D composite as-sembly was then overlaid with serum-free cell culture medium(right scheme in Fig. 5A) and scaffold invasion of A549 WT orA549 KD cells was studied with respect to the scaffold pore size asbefore.

Applying a chemoattractant significantly enhanced A459 cellWT invasion into 10 mm pore scaffolds (Fig. 5B and C). While only34% of WT cells were partially or fully embedded without chemo-attractant, this value rose to 82% in presence of the chemo-attractant. Decreasing the pore size to 8 mm and 7 mm progressivelyreduced the number of invaded WT cells, consistent with anincreasing geometrical constraint to cell migration. While 18% of allcells were on top at a mesh size of 10 mm, more than 35% were ontop at amesh size of 8 mmand about half of all cells at a mesh size of7 mm (Fig. 5B and C). Adding a chemoattractant also stimulatedA549 KD cell invasion in a similar fashion as A549 WT cells. Inter-estingly, no significant differences in invasion behavior were foundbetween A549 WT and A549 KD cells in presence of chemo-attractant. At a respective pore size, both cell types invaded thescaffolds to a similar degree (Fig. 5B and C). Thus, nuclear stiffnessapparently plays a less important role in regulating epithelial cellinvasion. However, most A549 WT and A549 KD cells that hadinvaded the scaffold showed deformed cell nuclei (Fig. 5C), sug-gesting that nuclear deformation is a result from but not a pre-requisite for scaffold invasion.

4. Discussion

The cell nucleus is a major mechanosensory integrator andimposes a physical challenge for cells moving in 3D microenvi-ronments. In agreement, nuclei often deform severely when cellsmigrate through narrow capillaries in vivo [26]. Better under-standing the contribution of nuclear stiffness to matrix invasion indifferent cell types is an important goal of biomedical research andrequires well-defined cell culture scaffolds.

mical attractant. The rate of migration into the scaffold decreased with decreasing poretractant at corresponding pore sizes (chi square test; *, p <0.05 compared to the WT on). (C) Transverse section of A549 WT and A549 KD populating 3D scaffolds of differentr in this figure legend, the reader is referred to the web version of this article.).

er scaffolds with adjustable pore size and chemoattractant gradients0.1016/j.biomaterials.2013.09.095

A.M. Greiner et al. / Biomaterials xxx (2013) 1e98

To study the effect of nuclear stiffness on cell invasion with theadditional option to apply a chemoattractant gradient, we havefabricated multifunctional 3D scaffolds on microporous films,creating composite 2D/3D microporous cell culture scaffolds withdifferent mesh sizes. In this sandwich approach, the microporousfilm is embedded in a photoresist and the 3D scaffold structure isthen generated by DLW in an upside down orientation directly onthe microporous membrane. To the best of our knowledge, DLW orother 2PP techniques have so far mainly been employed using solidglass carriers [31,32,35,44]. The combination of a planar micropo-rous carrier membrane and a 3D scaffold with adjustable mesh sizeexpands the versatility of artificial cell culture substrates forstudying chemotaxis-driven cell matrix invasion.

Several previous studies report the combination of cell culturescaffolds with chemotaxis assays. For instance, Hebeiss et al.introduced channel-like deformations into a porous polymer foil[45]. In contrast, we fabricated true 3D scaffolds with tunable meshsizes on top of a porous membrane carrier. A complementaryapproach combining 3D microstructures with a chemoattractantgradient has also been introduced by Tayalia et al. [34]. The authorsof this study created 2PP-produced 3D woodpile scaffolds withpore sizes of 25, 50, and 75 mm, which are not size-limiting fortranslocating mammalian cells. Tayalia et al. placed the 3D struc-tures in an L-shaped PDMS chamber and then positioned collagengel discs equipped with a chemokine at one end of the L-shapedchamber as a chemoattractant source. Diffusion out of the collagengel establishes a chemical gradient towards the 3D scaffold, simi-larly as our approach. While we have studied the role of nuclearmechanics on cell invasion into 3D scaffolds in the presence orabsence of a chemical gradient, Tayalia et al. used cell activation andchemoattractant exposure to quantitatively control directed cellmigration in the scaffolds [34]. Thus, by adjusting the microstruc-ture of the scaffolds and the nature of chemoattractant application,specific aspects of cell behavior can be analyzed.

Different factors are known to affect nucleus stiffness, amongthem the expression of class V intermediate filaments of the laminfamily. Reducing the lamin A/C protein content in A549 cells had asimilar effect on nuclear stiffness as a lamin A/C knockdown in MEFcells, clearly establishing the importance of lamins for nuclearmechanics in different cell types. MEF lmna KO cells invaded porousscaffolds more efficiently, which we attributed to the reduced nu-clear stiffness of these cells. In addition to nuclear stiffness, bothelasticity and viscosity of the cytoplasm are reduced in MEF lmnaKO cells [46], whichmay further help these cells tomigrate throughnarrow confinements in 3D microenvironments. Together with thehigher deformability of the nucleus and reduced mechanical sup-port of the nuclear lamina, these factors seem to favor the migra-tion of mesenchymal-derived cells into the 3D woodpile scaffolds.In agreement, a study by Khatau et al. showed that MEF lmna KOand MEF WT cells migrate similarly on planar 2D surfaces, butdifferently when cultured in a 3D collagen I matrix. The authorsconcluded that nucleo-cytoskeletal connections play a central roleformigrationwithin physiologically relevant 3D environments [47].

Although lamin A/C knockdown decreased nuclear stiffness inA549 cells, this did not enhance scaffold invasion in presence orabsence of a chemoattractant. Overall, A459 WT cell nuclei weresofter than MEF WT and even MEF lmna KO cells, further sup-porting that low nuclear stiffness does not correlate with efficientinvasion in certain cell types, such as epithelial cells. A549 cellslocated primarily on top of the scaffolds as expected for epithelialcells normally forming cell sheets [48]. However, A549 cell invasionwas strongly enhanced by a chemical gradient independently ofnuclear stiffness. In contrast, MEF cell invasion was sensitive tonuclear stiffness, but occurred independently of a soluble chemicalattractant. Nuclear stiffness may be an important factor specifically

Please cite this article in press as: Greiner AM, et al., Multifunctional polymfor studying cell matrix invasion, Biomaterials (2013), http://dx.doi.org/1

affecting mesenchymal-derived cells when migrating into a 3Dscaffold independently of chemical attractants. Thus evencomparatively simple artificial 3D scaffolds which structurallymimick an ECM-inspired microenvironment can reveal differencesin cell migration and invasion between different cell types.

5. Conclusion

In this study we have used DLW to fabricate well-defined 3Dscaffolds with adjustable mesh sizes on glass substrates or onmicroporous polymer membranes. The resulting cell culture scaf-folds provide a versatile tool for studying cell migration and inva-sion in a defined 3-dimensional microenvironment modified withdifferent ECM molecules and within gradients of soluble chemicalattractants. Our approach establishes more physiological culturesystems for studying cellular processes and may help to elucidatewhich factors are operative in ECM-mimicking 3D microenviron-ments. This tunable system may also be useful for understandingdifferences in cellular behavior in 3D vs. 2D environments on acellular and molecular level in more detail. Thus, it has significantvalue for a range of biomedical applications including advancedresearch tools or medical diagnostics. For instance, multifunctional3D scaffolds may allow for cell screening and cancer grading basedon differences in nuclear mechanics and invasion potential.Furthermore, composite scaffolds may provide informationregarding the differentiation state of stem cells, as it is known thatthe nuclear stiffness of these cells depends on their level of dif-ferentiation [26,46,49]. Therefore, the composite scaffold systemdeveloped in this study may provide an experimental platformbasis for future studies focusing on laminopathies, cancer metas-tasis, inflammatory disorders, tissue engineering and post-implantation wound healing.

Acknowledgments

A.M.G. acknowledges financial support by the Baden-Württem-berg Stiftung. M.J., A.C.S., C.M.F., and D.R.S. acknowledge financialsupport provided by the Deutsche Forschungsgemeinschaft (DFG)and the State of Baden-Württemberg through the DFG-Center forFunctional Nanostructures (CFN). The PhD education of A.C.S. is inthe framework of the Karlsruhe School of Optics & Photonics (KSOP).C.M.F. and D.R.S. thank Bianca Schatz for her assistance with elas-ticity measurements. The support by the BioInterface platform atKIT is highly acknowledged.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.biomaterials.2013.09.095.

References

[1] Jin T, Xu X, Hereld D. Chemotaxis, chemokine receptors and human disease.Cytokines 2008;44:1e8.

[2] Weiner OD. Regulation of cell polarity during eukaryotic chemotaxis: thechemotactic compass. Curr Opin Cell Biol 2002;14:196e202.

[3] Levayer R, Lecuit T. Biomechanical regulation of contractility: spatial controland dynamics. Trends Cell Biol 2012;22:61e81.

[4] Gardel ML, Schneider IC, Aratyn-Schaus Y, Waterman CM. Mechanical inte-gration of actin and adhesion dynamics in cell migration. Annu Rev Cell DevBiol 2010;26:315e33.

[5] Petrie RJ, Doyle AD, Yamada KM. Random versus directionally persistent cellmigration. Nat Rev 2009;10:538e49.

[6] Guck J, Lautenschlager F, Paschke S, Beil M. Critical review: cellular mecha-nobiology and amoeboid migration. Integr Biol (Camb) 2010;2:575e83.

[7] Chow KH, Factor RE, Ullman KS. The nuclear envelope environment and itscancer connections. Nat Rev Cancer 2012;12:196e209.

[8] Friedl P, Alexander S. Cancer invasion and the microenvironment: plasticityand reciprocity. Cell. 2011;147:992e1009.

er scaffolds with adjustable pore size and chemoattractant gradients0.1016/j.biomaterials.2013.09.095

A.M. Greiner et al. / Biomaterials xxx (2013) 1e9 9

[9] Kedrin D, van Rheenen J, Hernandez L, Condeelis J, Segall JE. Cell motility andcytoskeletal regulation in invasion and metastasis. J Mammary Gland BiolNeoplasia 2007;12:143e52.

[10] Kumar S, Weaver VM. Mechanics, malignancy, and metastasis: the forcejourney of a tumor cell. Cancer Metastasis Rev 2009;28:113e27.

[11] Hampoelz B, Lecuit T. Nuclear mechanics in differentiation and development.Curr Opin Cell Biol 2011;23:668e75.

[12] Gruenbaum Y, Margalit A, Goldman RD, Shumaker DK, Wilson KL. The nuclearlamina comes of age. Nat Rev Mol Cell Biol 2005;6:21e31.

[13] Wang N, Tytell JD, Ingber DE. Mechanotransduction at a distance: mechani-cally coupling the extracellular matrix with the nucleus. Nat Rev Mol Cell Biol2009;10:75e82.

[14] Dahl KN, Ribeiro AJ, Lammerding J. Nuclear shape, mechanics, and mecha-notransduction. Circ Res 2008;102:1307e18.

[15] Friedl P, Wolf K, Lammerding J. Nuclear mechanics during cell migration. CurrOpin Cell Biol 2011;23:55e64.

[16] Versaevel M, Grevesse T, Gabriele S. Spatial coordination between cell and nu-clear shape within micropatterned endothelial cells. Nat Commun 2012;3:671.

[17] Shimi T, Butin-Israeli V, Goldman RD. The functions of the nuclear envelope inmediating the molecular crosstalk between the nucleus and the cytoplasm.Curr Opin Cell Biol 2012;24:71e8.

[18] Pathak A, Kumar S. Biophysical regulation of tumor cell invasion: movingbeyond matrix stiffness. Integr Biol (Camb) 2011;3:267e78.

[19] Wolf K, Te Lindert M, Krause M, Alexander S, Te Riet J, Willis AL, et al. Physicallimits of cell migration: control by ECM space and nuclear deformation andtuning by proteolysis and traction force. J Cell Biol 2013;201:1069e84.

[20] Suresh S. Biomechanics and biophysics of cancer cells. Acta Biomater 2007;3:413e38.

[21] Xu W, Mezencev R, Kim B, Wang L, McDonald J, Sulchek T. Cell stiffness is abiomarker of the metastatic potential of ovarian cancer cells. PloS One 2012;7:e46609.

[22] Zhang H, Labouesse M. Signalling through mechanical inputs: a coordinatedprocess. J Cell Sci 2012;125:3039e49.

[23] Hutchison CJ. Lamins: building blocks or regulators of gene expression? NatRev Mol Cell Biol 2002;3:848e58.

[24] Prokocimer M, Davidovich M, Nissim-Rafinia M, Wiesel-Motiuk N, Bar DZ,Barkan R, et al. Nuclear lamins: key regulators of nuclear structure and ac-tivities. J Cell Mol Med 2009;13:1059e85.

[25] Schape J, Prausse S, Radmacher M, Stick R. Influence of lamin A on the me-chanical properties of amphibian oocyte nuclei measured by atomic forcemicroscopy. Biophys J 2009;96:4319e25.

[26] Versaevel M, Riaz M, Grevesse T, Gabriele S. Cell confinement: putting thesqueeze on the nucleus. Soft Matter 2013. http://dx.doi.org/10.1039/c3sm00147d.

[27] Broers JL, Peeters EA, Kuijpers HJ, Endert J, Bouten CV, Oomens CW, et al.Decreased mechanical stiffness in LMNA-/- cells is caused by defective nucleo-cytoskeletal integrity: implications for the development of laminopathies.Hum Mol Genet 2004;13:2567e80.

[28] Lammerding J, Fong LG, Ji JY, Reue K, Stewart CL, Young SG, et al. Lamins A and Cbut not lamin B1 regulate nuclear mechanics. J Biol Chem 2006;281:25768e80.

[29] Zwerger M, Ho CY, Lammerding J. Nuclear mechanics in disease. Annu RevBiomed Eng 2011;13:397e428.

Please cite this article in press as: Greiner AM, et al., Multifunctional polymfor studying cell matrix invasion, Biomaterials (2013), http://dx.doi.org/1

[30] Kaufmann A, Heinemann F, Radmacher M, Stick R. Amphibian oocyte nucleiexpressing lamin A with the progeria mutation E145K exhibit an increasedelastic modulus. Nucleus 2011;2:310e9.

[31] Greiner AM, Richter B, Bastmeyer M. Micro-engineered 3D scaffolds for cellculture studies. Macromol Biosci 2012;12:1301e14.

[32] Fischer J, Wegener M. Three-dimensional optical laser lithography beyond thediffraction limit. Laser Photonics Rev 2012.

[33] Weiss T, Schade R, Laube T, Berg A, Hildebrand G, Wyrwa R, et al. Two-Photonpolymerization of biocompatible photopolymers for microstructured 3Dbiointerfaces. Adv Eng Mater 2011;13:B264e73.

[34] Tayalia P, Mazur E, Mooney DJ. Controlled architectural and chemotacticstudies of 3D cell migration. Biomaterials 2011;32:2634e41.

[35] Ovsianikov A, Deiwick A, Van Vlierberghe S, Dubruel P, Moller L, Drager G,et al. Laser fabrication of three-dimensional CAD scaffolds from photosensi-tive gelatin for applications in tissue engineering. Biomacromolecules2011;12:851e8.

[36] Klein F, Striebel T, Fischer J, Jiang Z, Franz CM, von Freymann G, et al. Elasticfully three-dimensional microstructure scaffolds for cell force measurements.Adv Mater 2010;22:868e71.

[37] Klein F, Richter B, Striebel T, Franz CM, von Freymann G, Wegener M, et al.Two-component polymer scaffolds for controlled three-dimensional cellculture. Adv Mater 2011;23:1341e5.

[38] Lammermann T, Sixt M. Mechanical modes of ’amoeboid’ cell migration. CurrOpin Cell Biol 2009;21:636e44.

[39] Wolf K, Friedl P. Extracellular matrix determinants of proteolytic and non-proteolytic cell migration. Trends Cell Biol 2011;21:736e44.

[40] Lammerding J, Schulze PC, Takahashi T, Kozlov S, Sullivan T, Kamm RD, et al.Lamin A/C deficiency causes defective nuclear mechanics and mechano-transduction. J Clin Invest 2004;113:370e8.

[41] Ho CY, Lammerding J. Lamins at a glance. J Cell Sci 2012;125:2087e93.[42] Forbes B, Ehrhardt C. Human respiratory epithelial cell culture for drug de-

livery applications. Eur J Pharm Biopharm 2005;60:193e205.[43] Scherber C, Aranyosi AJ, Kulemann B, Thayer SP, Toner M, Iliopoulos O, et al.

Epithelial cell guidance by self-generated EGF gradients. Integr Biol (Camb)2012;4:259e69.

[44] Lindenmann N, Balthasar G, Hillerkuss D, Schmogrow R, Jordan M, Leuthold J,et al. Photonic wire bonding: a novel concept for chip-scale interconnects. OptExpress 2012;20:17667e77.

[45] Hebeiss I, Truckenmuller R, Giselbrecht S, Schepers U. Novel three-dimensional Boyden chamber system for studying transendothelial trans-port. Lab Chip 2012;12:829e34.

[46] Lee JSH, Hale CM, Panorchan P, Khatau SB, George JP, Tseng Y, et al. Nuclearlamin A/C deficiency induces defects in cell mechanics, polarization, andmigration. Biophys J 2007;93:2542e52.

[47] Khatau SB, Bloom RJ, Bajpai S, Razafsky D, Zang S, Giri A, et al. The distinctroles of the nucleus and nucleus-cytoskeleton connections in three-dimensional cell migration. Sci Rep 2012;2:488.

[48] McCaffrey LM, Macara IG. Epithelial organization, cell polarity and tumori-genesis. Trends Cell Bio 2011;21:727e35.

[49] Solovei I, Wang AS, Thanisch K, Schmidt CS, Krebs S, Zwerger M, et al. LBR andlamin A/C sequentially tether peripheral heterochromatin and inverselyregulate differentiation. Cell. 2013;152:584e98.

er scaffolds with adjustable pore size and chemoattractant gradients0.1016/j.biomaterials.2013.09.095