Fibroblasts maintained in 3 dimensions show a better differentiation state and higher sensitivity to estrogens

Toxicology and Applied Pharmacology 280 (2014) 421–433

Contents lists available at ScienceDirect

Toxicology and Applied Pharmacology journal homepage: www.elsevier.com/locate/ytaap

Fibroblasts maintained in 3 dimensions show a better differentiation state and higher sensitivity to estrogens Claudia Montani a, Nathalie Steimberg c, Jennifer Boniotti c, Giorgio Biasiotto a,b, Isabella Zanella a,b, Giuseppe Diafera d, Ida Biunno e,f, Luigi Caimi a,b, Giovanna Mazzoleni c, Diego Di Lorenzo a,⁎ a

Laboratory of Biotechnology, Department of Laboratory Medicine, Civic Hospital of Brescia, Italy Department of Molecular and Translational Medicine, University of Brescia, Brescia, Italy Laboratory of Tissue Engineering, Anatomy and Physiopathology Unit, Department of Clinical and Experimental Sciences, School of Medicine, University of Brescia, Italy d Integrated Systems Engineering (ISE), Milan, Italy e IRGB-CNR, Milan, Italy f IRCCS-Multimedica, Milan, Italy b c

a r t i c l e

i n f o

Article history: Received 17 January 2014 Revised 24 July 2014 Accepted 12 August 2014 Available online 28 August 2014 Keywords: Bioreactor Estrogen receptor 3 dimension Fibroblasts ECM Microgravity

a b s t r a c t Cell differentiation and response to hormonal signals were studied in a 3D environment on an in-house generated mouse fibroblast cell line expressing a reporter gene under the control of estrogen responsive sequences (EREs). 3D cell culture conditions were obtained in a Rotary Cell Culture System; (RCCS™), a microgravity based bioreactor that promotes the aggregation of cells into multicellular spheroids (MCS). In this bioreactor the cells maintained a better differentiated phenotype and more closely resembled in vivo tissue. The RCCS™ cultured fibroblasts showed higher expression of genes regulating cell assembly, differentiation and hormonal functions. Microarray analysis showed that genes related to cell cycle, proliferation, cytoskeleton, migration, adhesion and motility were all down-regulated in 3D as compared to 2D conditions, as well as oncogene expression and inflammatory cytokines. Controlled remodeling of ECM, which is an essential aspect of cell organization, homeostasis and tissue was affected by the culture method as assessed by immunolocalization of β-tubulin. Markers of cell organization, homeostasis and tissue repair, metalloproteinase 2 (MMP2) and its physiological inhibitor (TIMP4) changed expression in association with the relative formation of cell aggregates. The fibroblasts cultured in the RCCS™ maintain a better responsiveness to estrogens, measured as expression of ERα and regulation of an ERE-dependent reporter and of the endogenous target genes CBP, Rarb, MMP1 and Dbp. Our data highlight the interest of this 3D culture model for its potential application in the field of cell response to hormonal signals and the pharmaco-toxicological analyses of chemicals and natural molecules endowed of estrogenic potential. © 2014 Elsevier Inc. All rights reserved.

Introduction Whereas tissues and organs are three dimensionally organized (3D), the in vitro assessment of cell physiology, behavior and homeostasis and our ability to understand tissue formation, function and pathology has mostly depended on two dimensional (2D) cell culture systems (Yamada and Cukierman, 2007). Unfortunately, cell bi-dimensionality and microenvironment of conventional cultures often failed to mimic the dynamic of in vivo contexts, where cell behavior and homeostasis are regulated by systemic cues (Mazzoleni et al., 2009; Baker and Chen, 2012). As an intermediate model between cell cultures and

⁎ Corresponding author at: Laboratory of Biotechnology, Civic Hospital of Brescia, P.le Spedali Civili, 1, I-25123 Brescia, Italy. Fax: +39 030 307251. E-mail address: [email protected] (D. Di Lorenzo). URL: http://www.spedalicivili.brescia.it (D. Di Lorenzo).

http://dx.doi.org/10.1016/j.taap.2014.08.021 0041-008X/© 2014 Elsevier Inc. All rights reserved.

human tissues animals have been so far used to allow the definition and understanding of specific and complex bioprocesses (Yamada and Cukierman, 2007; Sonneveld et al., 2006; Cook et al., 2012). However, the ethical limit of performing animal based-experimentations and the always more stressing regulatory rules have rendered the search of alternative methods a high priority issue. The extrapolation of results to humans is often difficult due to the fact that animal models may not accurately reproduce some human features of physiological and pathological conditions (Ferrarini et al., 2013a; Bracken, 2009; Hartung, 2008; Mazzoleni and Steimberg, 2012). In light of these considerations, in the least decades, there was an increasing need to create new model systems able to mimic in vitro the features and functionality of living organs (Elsdale and Bard, 1972); for a number of tissue/cells, in vitro 3D tissue models provided a fruitful strategic approach that bridged the gap between traditional cell cultures and animal models (Ferrarini et al., 2013b; Griffith and Swartz, 2006; Rangarajan et al., 2004; Pampaloni et al., 2007; Hirschhaeuser et al., 2010).

422

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

When cultured on plastic dishes, cells face an artificial microenvironment where they are forced to grow in an abnormal bipolar organization. In such conditions, they dramatically differ in their morphology, cell–cell and cell–matrix interactions, behavior and differentiation, from those growing in more physiological 3D microenvironments as well as in tissue-specific context (Cukierman et al., 2002; Griffith and Swartz, 2006; Nelson and Bissell, 2006; Xu et al., 2009). The difference is not only in terms of cell morphology, adhesion and modality by which cells exert forces on their own surrounding extracellular matrix (ECM) and reciprocally the way by which ECM senses extrinsic forces and transmits them to the ECM-embedded cells (Pedersen and Swartz, 2005), but also in biological responses to microenvironmental stimuli such as biochemical, mechanical and biophysical factors (Silver et al., 2003; Huang et al., 2012). An appreciation is emerging for the extent to which the 3D environment governs the cell responses, particularly for cells naturally existing in 3D interstitial spaces like fibroblasts (Pedersen and Swartz, 2005). In fibroblasts, for example, whereas the 2D culture induces cell spreading with prominent cellular extensions, a 3D configuration favors a globular or bipolar dorsal–ventral or stellate shape according to the ECM characteristics, that governs the cell features (Beningo et al., 2004; Larsen et al., 2006) by affecting mechanotransduction of signals and specific intracellular signaling networks (Cukierman et al., 2002; Torres and Rosen, 2006). In vitro, the microenvironment properties (composition of ECM, dimensionality, architecture, topography, compliance of the ECM) may directly modulate the overall cell morphology, cell–matrix adhesion, and cytoskeletal organization (Giretti and Simoncini, 2008; Ananthakrishnan and Ehrlicher, 2007) and as a consequence they regulate cell phenotype survival, proliferation rate (Boateng et al., 2003), inhibition of proliferation (Sarasa-Renedo and Chiquet, 2005), cellular migration, cellular homeostasis and gene expression (Birgersdotter et al., 2005; Cukierman et al., 2002; Pedersen and Swartz, 2005). Thus, the maintenance of such 3D dynamic conditions is of importance when fundamental cell processes such as steroid hormone regulation are studied. Steroid hormones act as extranuclear signaling factors which regulate cell morphology (Naftolin and Malaspina, 2007) via rapid signaling to the actin cytoskeleton (Simoncini et al., 2006; Giretti et al., 2008; Sanchez et al., 2009; Louvet-Vallée, 2000; Tsukita and Yonemura, 1999). For instance, estrogens activate moesin in breast cancer, endothelial and neuronal cells through a rapid, extranuclear signaling cascade originated by the interaction of estrogen receptor alpha (ERα) with the G protein Ga13. This process leads to the recruitment of RhoA and of Rho-associated kinase, ROCK-2 and the activation of moesin which leads to modification of the interaction with the extracellular matrix and nearby cells (Sanchez and Simoncini, 2010; Fu and Simoncini, 2008). In this work we studied the role of the 3D microenvironment on fibroblast phenotype and more especially on a fundamental signaling process such as the response to estrogens. We maintained an in-house generated mouse fibroblast cell line expressing a reporter gene under the control of estrogen responsive sequences (EREs) in a dynamic 3D microenvironment where high mass transfer was allowed in the absence of significant shear stress. Such conditions were obtained in a Rotary Cell Culture System; (RCCS™), a microgravity based bioreactor that promotes the aggregation of cells into multicellular spheroids (MCS) by preventing cell adhesion to an artificial surface (Grun et al., 2009). This bioreactor has shown advantages over static and other dynamic tissue culture systems in that cells maintained in a laminar fluid state better express their differentiated phenotype and more closely resemble in vivo tissue equivalents (Ferrarini et al., 2013b; Hammond and Hammond, 2001; Unsworth and Lelkes, 1998; Mazzoleni and Steimberg, 2010; Mazzoleni et al., 2011; Steimberg et al., 2010). In parallel, the biological response of these cells to estrogen receptor activation was evaluated in the classical static 2D conditions. We here show that: a) the RCCS™ is suitable to maintain fibroblasts in a more differentiated state, demonstrated shown by the higher expression of genes

involved in several pathways regulating cell differentiation and functions; b) the fibroblasts cultured in the RCCS™ maintain a better responsiveness to estrogens, thus suggesting that estrogen receptors and the estrogen signaling pathways are in a more functional situation than in 2D Petri dishes. Our data highlight the opportunity to further investigate this 3D culture model for its potential application in the field of research on nuclear receptor functions and for the pharmaco-toxicological analyses of chemicals and natural molecules endowed of hormonal potential. Material and methods Fibroblast immortalization Generation of immortalized embryonic mouse fibroblasts (MEFs) was performed on fibroblasts isolated from 13.5-day-old embryos from ERE-tK-Luciferase mice. Dermal fibroblasts were obtained from ventral embryo skin. The skin was cut and placed into Petri dishes maintaining the dorsal–ventral polarization. The tissue was maintained in DMEM (Lonza, Milan, Italy), which contained 20% fetal bovine serum (Invitrogen, Milan, Italy), 2 mM L-glutamine and antibiotic mix 1× (Sigma, Pomezia, Italy) (complete culture medium). After 10 days, confluent fibroblasts were transfected with the plasmid pSV40-Neo encoding for the SV40-large T antigen by mean of Lipofectamine 2000 (Invitrogen, Milan, Italy), as described by the manufacturer. Fortyeight hours after transfection, cells were selected for G418 resistance (0 .4 mg/ml), this selection was performed over three weeks. Foci of resistant cells were isolated using cloning rings and further expanded. The characterization of clones was based on their capacity to respond to estrogens, in DMEM without phenol red (Invitrogen, Milan, Italy) supplemented with 2% of fetal bovine serum (Invitrogen, Milan, Italy), 2 mM L-glutamine, 1× antibiotic mix (Sigma, Pomezia, Italy; 100 IU/ml penicillin, 100 μg/ml streptomycin and 0.25 μg/ml fungizone) and 1 mM Na-pyruvate (Sigma, Pomezia, Italy) (treatment medium). 104 dermal fibroblasts were seeded onto 35 mm-diameter Petri dishes in complete culture medium. Cells were daily harvested and their number determined using a hemocytometer. 3 dimensional cell culture in the RCCS™ The selected clone was expanded, harvested, resuspended in 10 ml culture medium and introduced into the 10 ml-HARV culture vessel (RCCS™ bioreactor) at the final concentration of 1.5 million of cells/ml (http://science.nasa.gov/NEWHOME/br/bioreactor.htm) (Houston, USA). The culture medium was the same as the 2D culture. Cells were grown for 2 or 7 days in 3D and then harvested. Immortalized fibroblasts were amplified in monolayer (2D) and subsequently cultured in the RCCS™ device to characterize the influence of the 3D culture conditions on the specific cell viability, phenotype and functions. Cells were harvested at different times of culture (from T0 up to 192 h) and immunolocalization of specific proteins was performed. Cell treatments Immortalized fibroblasts maintained either in 2D or in 3D conditions were treated with increasing doses of 17β-estradiol (Sigma-Aldrich), genistein (Sigma-Aldrich) or Bisphenol A (Sigma-Aldrich) for 24 h, in phenol free culture medium. After 24 h cells were harvested for further studies. Luciferase assay Luciferase assay was performed using a Luciferase Assay kit (Promega, Milan, Italy). Cells were lysed using the reporter lysis buffer 1×, previously added to the cells. Cells were collected, maintained on ice for 30 min. Then, luciferase activity was estimated and protein

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

concentration was determined by Bradford's assay. Luciferase enzymatic activity was measured mixing 20 μl of cell lysate and 100 μl of luciferase assay reagent. The intensity of the light was measured with a luminometer (Centro 960, Berthold, Germany) over 10 s and expressed as fold change (FC) of relative light units (RLU)/mg proteins. mRNA quantification Total RNAs from fibroblasts were extracted from 1 million of cells using the TRIzol reagent (Invitrogen, Milan, Italy). For each sample, RNAs were reversed transcribed using high capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA). Quantitative PCR was performed using Assay on Demand kits based on TaqMan chemistry. RT-PCR reactions were performed on an ABI PRISM 7000 Sequence Detection System instrument and data analysis was done with the ABI PRISM 7000 SDS software (Applied Biosystems, Foster City, CA). 18S RNA was used as the reference housekeeping gene. Calculations were done as described for the Comparative Method in the User Bulletin 2 of ABI PRISM sequence detection system. RNA expression analysis with TaqMan microfluidic cards: RNA for each sample was reversed transcribed as above, with a Master Mix containing 2.5 U/μl of MultiScribe Reverse Transcriptase and 1 μg of total RNA. Two microliters of single-stranded cDNA was mixed with 48 μl of nuclease-free water and 50 μl of TaqMan Universal PCR Master Mix. After loading 100 μl of the sample-specific PCR mixture into one sample port of the microfluidic cards, the cards were centrifuged twice for 1 min at 280 g and sealed to prevent well-to-well contamination. The cards were placed in the microfluidic card Sample Block of an ABI Prism 7900 HT Sequence Detection System (Applied Biosystems). The thermal cycling conditions were 2 min at 50 °C and 10 min at 95 °C, followed by 40 cycles of 30 s at 97 °C and 1 min at 59.7 °C. The assay for each gene was carried out in triplicate. The calculation of the threshold cycle (Ct) values was performed using the SDS 2.2 software (Applied Biosystems), after automatically setting the baseline and the threshold. The 96 genes Low Density Array (LDA) cards were designed containing key genes of the reported pathways. 18S RNA was used as the reference housekeeping gene. Specific oligonucleotide pairs were designed by the Applied Biosystems service. Gene expression profiles Cells were harvested in TRIzol reagent (Invitrogen, Milan, Italy) and stored at − 80 °C. RNAs were extracted following classical TRIzol– chloroform procedure and quality controlled on an Bioanalyzer 2100 (Agilent Technologies). All RNAs used in our experiments presented good quality and were processed according to Affymetrix (Santa Clara, CA) instructions for expression profiling. One microgram of total RNA was processed using the Affymetrix GeneChip one-cycle target labeling kit. The resulting biotinylated cRNA was fragmented and hybridized to the GeneChip Mouse Genome 430 2.0 Array. The arrays were washed, stained, and scanned using the Affymetrix Model 450 Fluidics Station and Affymetrix Model 3000 scanner. Array images were visually inspected to discount for signal artifacts, scratches or debris and Affymetrix quality metrics were analyzed using the Bioconductor (Gentleman et al., 2004) (yaqcaffy library). The raw data were normalized using RMA (Irizarry et al., 2003) and analyzed using linear models and empirical Bayesian methods (Smyth, 2004) as implemented in the limma package. FDR-adjusted p-values (Benjamini and Hochberg, 1995) had to be b0.05 and absolute values of log2 fold-changes N 1 for genes to be considered differentially expressed between conditions. In addition, KEGG (Kanehisa et al., 2010) pathways of interest were individually screened to assess the effect of cell culture in 3 dimensions.

423

(1× pH 7.4). The cell suspension was quickly poured in doughnutshaped agarose molds prepared as previously described (Moskaluk and Stoler, 2002) with few modifications. Briefly, 3% of standard agarose was poured into a 12 well dish and allowed to set. A cylindrical core of 6 mm-diameter was removed from the center to accommodate the low melting agarose cell suspension. The resulting molds were dislodged from the wells, placed in a tube with 1% PFA overnight at 4 °C to harden the agarose and embedded in paraffin using an automated tissue processor (Leica ASP300S). CMAs were made using the semiautomatic arrayer (Galileo CK3500, www.isenet.it) and 1.0-mm needle. Four sample cores were removed from each donor block and placed in a recipient block in a predesigned array coordinates. The resulting array was heated on a hot plate at 42 °C for 10 min and the surface of a clean glass slide was pressed on the block to homogenously integrate the sample cores into the paraffin. Four (4) μm-thick sections were sliced using an ordinary microtome and processed for immunofluorescence analysis. Immunofluorescence analysis Cells grown onto coverslips were fixed in cold 4% PFA for 20 min at 4 °C, washed in PBS and permeabilized with 0.1% Triton X-100 for 10 min at 4 °C. Deparaffinized sections were instead hydrated in graded alcohols and incubated in citrate buffer for standard heat-induced antigen retrieval. Slides were treated with blocking solution (2% donkey serum, 1.5% BSA, 0.5% fish gelatin) for 45 min and subsequently incubated with primary antibodies over night at 4 °C in blocking solution. Appropriate secondary antibodies conjugated with Rhodamine-Red (Jackson Immuno Research, West Grove, PA, USA) or Alexa Fluor 488 (Molecular Probes, Invitrogen, Carlsbad, California, USA) were used for the detection. Nuclei were counterstained with Hoechst 33258 and samples were mounted with GelMount aqueous mounting medium (SIGMA). Images were acquired using a LEICA DMI4000B inverted fluorescence microscope linked to a DFC360FX camera (LEICA Microsystems, Vienna, Austria). Protein isolation and western blot analysis. Cells were homogenized in ice-cold lysis buffer (50 mM tris pH 8, 150 mM NaCl, 1 mM EGTA, 100 mM NaF, 10% glycerol, 1 mM MgCl2 and 1% Triton X-100) containing protease inhibitors cocktail and incubated on ice for 30 min. Samples were centrifuged twice at 13000 ×g for 30 min at 4 °C and the protein content of the supernatant was determined using a Bio-Rad Protein assay (Bio-Rad). Forty milligrams of total cellular protein was resolved by SDS-PAGE on an 8% polyacrylamide gel and electrotransferred to PVDF membrane. Immunoblot analysis was done using the following antibodies: mouse anti-ERα (1:500, Santa Cruz Biotechnology); mouse anti-ERβ (1:1000, Santa Cruz Biotechnology) and goat antiactin (1:1000; Santa Cruz Biotechnology). Membranes were then reacted with secondary antibodies (1:3000, Santa Cruz Biotechnology) and developed using the ECL kit (Pierce). Densitometric quantitation relative to actin levels was performed using the ImageJ software (National Institutes of Health; http://rsbweb.nih.gov/ij/). Statistical analysis. Data are expressed as mean ± SE. Pairwise comparisons were made using Student's t-test. A probability level p b 0.05 was considered statistically significant. Results

Embedding of cells suitable for cell microarray (CMA) construction

Growth state evaluation in the classical static 2D Petri dish and in 3 dimensional cell culture system RCCS™

Cultured cells were fixed in 4% PFA at 4 °C for 20 min and resuspended in 100 μl of warm 2% low melting agarose (Sigma) prepared in PBS

After seeding in 2D with a very short-time of latency (less than 24 h), cells grew exponentially and quickly from 48 h to 168 h

424

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

(doubling time about 18 h) (Fig. 1Aa). In order to evaluate the immortalized phenotype of SV40-transfected cells, cells were cultured with decreasing concentrations of serum (Fig. 1Ab). Results indicated that cell growth remained highly dependent on the serum presence (cells grow slower and entered a cell cycle arrest at 144 h for serum concentrations inferior to 20%). Moreover, low density seeding usually

used to evaluate the degree of transformation in vitro showed that cells cannot proliferate when seeded at a density lower than 100 cells (Fig. 1Ac). Under 3D conditions, cells spontaneously formed aggregates that grew in size (Fig. 1Ba) and size of aggregates progressively increased from about 60 cells/4 μm-section of aggregates at 24 h, to about 255 cells/4 μm-section of aggregates at 192 h (Fig. 1Bb).

Fig. 1. Dermal fibroblasts isolated from ERE-LUC fetal mice were transfected with the pSV-Neo plasmid harboring the SV40 Large T antigen. Growth properties of cells were assessed in different conditions of 2D culture (as a function of serum concentrations (a) and (b), or depending on seeding density, (c)). A. Cells were grown in 3D where they spontaneously form multicellular aggregates (a). Time-dependent cell proliferation (b). Immunofluorescence was performed on spheroid tissue arrays to evaluate apoptosis and necrosis processes. Cells were stained with Histone-3 (pH-3 green) antibodies for proliferation assessment and with active caspase-3 (Casp-3 red) antibodies to assess apoptotic states. Hoechst (Ho-blue) was used for the labeling of the nuclei. A time-course analysis was performed at 0–24–48–96 and 192 h.

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

Immunostaining in situ was performed on specific markers of cell proliferation (pH 3) and apoptosis (caspase3, Casp3). Cells were amplified in monolayer (2D), subsequently transferred to the RCCS™ device, and were harvested at different time points after starting the RCCS™ culture (up to 192 h) in order to characterize the influence of the culture conditions (3D and dynamic states) on cell growth, proliferation and apoptosis. Casp3 expression started to increase at 48 h and kept increasing up to 192 h, when cells were organized into a spheroid conformation (Fig. 1Bc). pH 3 was expressed at time 0 in 5% of the cells and started to decrease at 24 h (4%) falling down to 0.5% at 192 h, indicating a cell arrest in the G0 phase (Fig. 1Bc). Cell viability seems to progressively decrease with time in culture. Apoptotic processes were randomly localized inside the aggregate.

425

Since controlled remodeling of ECM is an essential aspect of cell organization, homeostasis and tissue repair in vivo and in vitro, the expression of metalloproteinase 2 (MMP2) and its physiological inhibitor (TIMP4), which are directly involved in these processes were further studied in both dimensionalities. Whereas in fibroblasts cultured in 2D, MMP2 seemed to be expressed in a random way, in 3D it follows a more time/density dependent pattern of expression. At the beginning of the 3D culture, cells expressed a very low level of MMP2, after 1 week of culture in the RCCS™, MMP2 level strongly increased, showing a clear distribution of the protein within the multicellular aggregates. On the contrary, TIMP-4 was detected almost only at the outer surface of the cell aggregates during the first 4 days, while it was also observed in the inner parts at day 8. In 2D, TIMP-4 expression increased after 4 days in culture and decreased after one week when only a few cells were still expressing it.

Cytoskeleton arrangement in 3D cultures by tissue arrays Analysis of differentiation pathways by microarrays The cytoskeleton arrangement was evaluated in the 2D and 3D microenvironments by immunolocalization of β-tubulin. As shown in Fig. 2A, β-tubulin expression increases in 2D, reaching maximal expression at 120 h when cells tend to reach confluence. In the RCCS™ (Fig. 2B), where fibroblasts aggregated gradually to form multicellular spheroids, which grew in size as a function of the time in culture, β-tubulin showed a peculiar pattern of expression within the whole aggregate. Whereas cells located in the inner parts of the aggregate slightly expressed β-tubulin, the outer cells showed higher expression. It seems that cells already stabilized in multicellular aggregates, loose their capacity to express β-tubulin.

Subsequently, we studied the effect of the 3D culture on gene expression by microarrays, to identify genes that are differentially regulated in each culture condition. We also studied changes in identified pathways before and after SV40-large T antigen immortalization of primary cells (Fig. 3). Four pathways of basal cell functions that show significant changes are represented as Volcano Plots (focal adhesion, actin cytoskeleton, cell cycle and TGF-beta signaling). The comparisons were made as follows: synchronized primary fibroblasts vs SV40-large T antigen immortalized fibroblasts in 2D (black circles), and, to evidence phenotype reversal toward more differentiated states, fibroblasts

Fig. 2. A. Immunolocalization through tissue arrays was performed onto cells maintained in 2D (A) and in 3D culture (B) for different periods of time. Different markers of cell morphology/ ECM were studied: β-Tubulin (β-Tub), MMP2 and TIMP4 (with red fluorophore: Rhodamine-Red) and histone (pH 3, green fluorophore: Alexa Fluor 488).

426

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

Fig. 3. Whole genome analyses were performed on immortalized fibroblasts vs primary fibroblasts, and in 2D vs 3D conditions. Patterns of gene expression were represented by Volcano Plots. This representation arranges genes along dimensions of biological and statistical significance. The x axis represents fold changes between the two groups (in log2). Positive values: upregulation. Negative values: downregulation. The y axis represents the p-value (t-test) of differences between samples on a negative log scale, to highlight small p-values. The lowest pvalues (higher significance) are toward the top of the figures. The experiments were performed in triplicate. Full circles (black) represent values from 3D cell cultures versus 2D. The different cells (immortalized vs primary cells) and culture conditions 3D vs 2D were compared as follows, white circles: 2D cultures vs 3D cultures; black circles: immortalized vs primary fibroblasts (*p b 0.05; **p b 0.01; ***p b 0.001).

cultured in 2D vs 3D cultures (white circles). Immortalization of fetal fibroblasts with the SV40-large T antigen induces an up-regulation of cell cycle related genes (Fig. 3D) (negative values are in the left side of the panel) and a down-regulation of focal adhesion, actin cytoskeleton and TGF-beta signaling pathway related genes (Figs. 3A, B, C) (positive values are in the left side of the panel). Immortalized fibroblasts transferred in the RCCS™ show an up-regulation of genes involved in focal adhesion, the actin cytoskeleton and TGF-β signaling compared to 2D (white circles) (Figs. 3A, B, C) (the right side of the panel indicates 2D vs 3D analysis), which probably reflect morphological remodeling and structural reorganization of the cells into spheroids, which appear to be composed of cells in a higher differentiation state. Quantification of differentially expressed genes The homeostatic state of cells in 3D culture versus cells in 2D was further assessed by quantifying the mRNA expression of different

genes involved in fibroblasts differentiation and function that showed modulation by the microarray analysis. After one week of culture, several genes of the pathways indicated in Fig. 4 are modulated, suggesting an effect of the 3D culture method on basic and highly specialized cell functions. With regard to the differentiation genes, we focus our attention on genes that regulate the more important function of fibroblasts, whatever their tissue origin, that is represented by synthesis of the extracellular matrix of connective tissues. Thus, metalloproteases (MMP) and type I collagen expression was examined and considered as differentiation markers. The transcription of MMP2 and COL1A1 mRNAs is highly up-regulated in 3D with respect to cells maintained in 2D culture (52- and 4-fold respectively) suggesting that the 3D microenvironment provided by the RCCS™ allows maintaining cells in a more differentiated state. The active regulation of fibroblast fates was also evidenced by the induction of genes such as FGF, MMP and PPARg. In 3D culture, the mRNA expression of FGF9 and FGF6, related to the paracrine FGF subfamily was highly induced (from 20 to 60 fold). The expression of

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

427

Fig. 4. mRNA quantification of genes involved in fibroblast differentiation and function that showed modulation by the microarray analysis. Expression of mRNAs was studied by qRT-PCR in 2D and 3D cultured cells. Gray bars show the expression of the same genes after seven days in the 3D culture (RCCS). 2D is set at 1 for comparison with the 3D cultures. Changes in gene expression are relative to the ribosomal housekeeping gene 18S RNA. The experiments were performed in triplicate (*p b 0.05; **p b 0.01; ***p b 0.001).

the MMP10 enzyme was about 60 fold increased, confirming the activation of ECM remodeling processes. PPARg, a modulator of fibrotic tissue processes and connective tissue turnover, was increased by 50 fold in 3D culture. Genes related to cell cycle, proliferation, cytoskeleton, migration, adhesion and motility were all down-regulated in 3D, as well as oncogene expression and inflammatory cytokines.

Estrogen receptor expression We next evaluated if the 3D induced-phenotype modifies the responsiveness of the cells to estrogens. Thus we estimated the expression level of the two estrogen receptor subtypes (ERα and ERβ) in 2D and 3D at two different time points in culture (0, 3 and 7 days) (Fig. 5). ERα is significantly more expressed in cells in 3D culture at day 7. ERβ appeared generally to be more expressed in 3D, without however reaching a significant difference compared to the cells in 2D. ERα protein levels were also increased at 7 days in 3D cultures, while ERβ showed no significant changes (Fig. 5B).

Comparison of estrogen responsiveness in 3D versus 2D cell cultures After having assessed the higher expression of ERα in the cells maintained in 3D culture, we next evaluated its ability to respond to estradiol (E2). Cell were placed in 2D and 3D conditions and treated with 1 nM and 10 nM E2 for 24 h. Estrogen receptor activation was evaluated through the response of an ERE-tK-Luciferase reporter stably integrated. E2 was able to modulate the expression of the reporter in both culture conditions. At 1 nM, in 3D cultures E2 induced the reporter up to 4.9folds, whereas in 2D, such concentration did not induce any significant response. At the 10 nM dose a 3-fold induction was achieved in 2D and a 2-fold induction in 3D (Fig. 6A). These results suggested a better response to E2 when cells are in 3D. With the aim of studying the application of the RCCS™ cell culture device to the assessment of the activity of estrogenic chemicals of different nature, we next evaluated the estrogenic activity of the food chemical genistein and the industrial chemical Bisphenol A. As shown in Fig. 6B, genistein induced the expression of the reporter in a dose-dependent manner in monolayer (p b 0.05 at 10 μM) and at

428

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

A

2D 0 days

A

* 2

2D 3D

*

4

**

3 2 1 0

+ICI

Contr.

1

B

+ICI

+ICI

E2 1nM

+ICI

C 1 nM

**

0 0

3

7

0

3

7

10 nM 1 µM 10 µM

days 12

**

ERß

ERα α

2D

3D

E2

100-fold ex.

2D 3D BPA

2 2D D 3D 3 D 7 days

2D 3D 7 days

be regulated by estrogens from in-house studies (Aldh1a7, RARβ) or from the literature (FGF2, MMP1, Dbp and CBP) (Tozlu et al., 2006; Fogarty et al., 2012; Pirani et al., 2013; Skrzypczak et al., 2013). These genes were used to further assess the effect of the culture conditions on the extent of responsiveness to estrogens (E2 and genistein). Gene expression was quantified by low density microfluidic LDA cards and each gene showed regulation at different doses of hormones. Among these genes, FGF2 and Aldh1a7 showed much higher levels of basal expression in 3D, whereas other genes (CBP, Rarb, MMP1 and Dbp) were all repressed in 3D culture (Fig. 7A). When fibroblasts were stimulated with estrogens, all these genes showed a good responsiveness. However, CBP, Rarb, MMP1 and Dbp were more sensitive to estrogens in 3D, while Aldh1a7 and FGF2 were better induced in 2D. Genistein was effective in both culture models, although, as for E2, its effect was more pronounced in 2D on Aldh1a7 and FGF2 while, for the other studied genes, the 3D conditions conferred more sensitivity to these estrogens.

6 5 4 3 2 1 0

ERß actin Time 0

2D 3D Genistein

Fig. 6. A. Activity of 17β-estradiol (E2) in ERE-tK-LUC fibroblasts in 2D and 3D cell cultures. Cells were treated for 24 h with 1 nM or 10 nM of E2. Controls were incubated with vehicle (ethanol). Luciferase activity is expressed as relative light units normalized to protein concentration. Values are expressed as fold-induction. The experiments were repeated three times. Values with asterisk are significantly different from control (*p b 0.05; **p b 0.001). B. Dose–response analysis of estradiol (E2), genistein and Bisphenol A on ERE-tK-LUC fibroblasts cultured in 2D or 3D conditions. Compounds were mixed in a red phenolfree media and 5% charcoal stripped fetal bovine serum. Compound concentrations were 0, 1 nM, 10 nM, 1 μM and 10 μM. Data are expressed as fold change. Values with asterisk are significantly different from control (*p b 0.05; **p b 0.01).

actin

Arbitrary Units

1

0

Time 0 Time

*

+ICI

*

* *

+ICI

*

*

+ICI +ICI

6 5 4 3 2 1 0

*

* 2

+ICI

Arbitrary Units

B

Fold change

3

+ICI

ER

100-fold ex.

E2 10 nM

+ICI

Fold change

3D 7 days

Fold change

3D 3 days 3

** 5

Fig. 5. Effect of 3D culture conditions on the expression of ERα and ERβ genes at 3 days and 7 days. White bars represent the expression of each gene in 2D at time 0. The value is set at 1 for comparison with the expression at days 3 and 7. Dark gray bars show gene expression after 3 days in 3D culture. Black bars represent gene expression at 7 days in 3D culture. ERα and ERβ expression is shown relative to the ribosomal housekeeping gene 18S RNA. B) Effect of the RCCS™ culture system of the protein level of ERα and ERβ at time 0 (starting 2D culture) and 7 days. Densitometric quantitation relative to actin levels was performed using the ImageJ software (National Institutes of Health; http://rsbweb. nih.gov/ij/) and is representative of experiments performed in triplicate. Data are shown as means ± SEM for each group. Values with asterisk are significantly different from T0 (*p b 0.05).

10 nM in 3D (p b 0.05), Bisphenol A induced an upregulation of the reporter at 10 μM in each system. Regulation of metabolic genes From the microarrays data we identified and selected a few genes that showed differential expression in 3D vs 2D and that are known to

Discussion Skin represents a target organ for cosmetics, xenobiotics, drugs, pollutants, chemicals as well as radiations. Among all the cell types and structures that form the skin, and more distinctly dermal connective tissue, fibroblasts represent an attractive object of study since, together with other cell types, it contributes to important biological processes, such as wound healing and inflammation, (Shaw and Martin, 2009). Fibroblasts are effectively responsible for collagen, elastin and other matrix protein syntheses, ECM turn-over and the buildup of the connective

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

A

429

*** 300

***

Fold change

3

2

1

Fgf2

Aldh1a7

0

*

* * * CBP MMP1 Dbp

Rarβ 2D

B CBP ***

30

15 1 0

*

0

0,1

1

*

10 0,1

1

50

* *

10

0

1

10 1

*

* 1 0

10

0

0,1

E2 (nM)

Aldh 1A7

10

Rarβ

100

Gen (µM)

Fold change

3D

***

1

10 0,1

1

*

10

0

1

Gen (µM)

***

5

50

*

* 1 0

10

***

FGF2

100

10 1

E2 (nM)

0

0,1

1

10 0,1

1

10

Gen (µM)

0

1

10 1

1 0

10

0,1

1

10 0,1

1

10

0

Gen (µM)

E2 (nM)

MMP1 ***

20

0

1

10 1

Dbp ***

50

10

E2 (nM)

***

**

** **

10

1 0

25

*

0

0,1

1

0,001 10 0,01 0,1

Gen (µM)

1

0

1

10 1

E2 (nM)

10

1 0

*

*

0

0,1

1

*

*

10 0,001 0,01 0,1

Gen (µM)

1 0 10

1

10 1

10

E2 (nM)

Fig. 7. ERE-tK-LUC fibroblasts were cultured under 2D and 3D conditions and gene expression was analyzed. Relative mRNA quantification for specific genes was based on TaqMan, LDA microfluidic cards. A few genes which expression is known to be depend on estrogens were chosen either to evaluate the influence of the 3D microenvironment on their regulation. The concentration of chemicals used was in the following ranges: 1 and 10 nM for estradiol and 1, 10 and 100 nM and 1 and 10 μM for genistein. The expression of target genes was normalized to the expression of 18S reference gene (relative quantification). Values represent the fold induction/inhibition over the control. The experiments were performed in triplicate. Bars represent the average ± SEM (*p b 0.05; **p b 0.01; ***p b 0.001).

tissue. Furthermore, they are fundamental for normal tissue homeostatic processes such as tissue repair in response to injury, development and regulation of cell growth (Hinz and Gabbiani, 2003; Desmoulière et al., 2005; Broughton et al., 2006), and they are associated to the pathogenesis of fibrosis related diseases (Kisseleva and Brenner, 2008a, 2008b) and cancer (Beacham and Cukierman, 2005; Kalluri and Zeisberg, 2006; Pavlakis et al., 2008; Castelló-Cros and Cukierman, 2009; Rhee, 2009; Ostman and Augsten, 2009). It might be difficult to define fibroblast differentiation state also because of the incomplete knowledge about this cell type, either in in vivo or in vitro conditions. This cell type is in realty

composed of a very heterogeneous population of cells, which phenotype varies according to their organ/tissue origin, their localisation within a tissue, their physiological or pathological state, their stimulation by microenvironment and their mesenchymal plasticity (Chang et al., 2002; Sorrell and Caplan, 2009). However, because the main and common function of fibroblasts is to maintain connective tissue homeostasis by regulating synthesis and breakdown of the ECM components, we decided to focus our attention on type I collagen and metalloproteases (as well as their tissue inhibitors, TIMP). Thus, in our manuscript the notion of “more differentiated state” is referred to the capacity of the 3D

430

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

microenvironment to favor the expression of ECM related genes (type I collagen and MMP) as compared to the 2D flat environment. Other investigated genes that referred to proliferation, oncogenes, and cell cycle pathways were included because they reflect the capacity of the cell to divide/proliferate and the process that is often accompanied by a loss of differentiated functions. 2D culture was already shown to represent an optimal model to study cell proliferation as compared to 3D systems, whereas 3D culture is more adequate for differentiated phenotype studies (Weigelt et al., 2014). In our 3D culture conditions in the RCCS, results confirmed that the 3D conformation induced a down-regulation of all these proliferation pathways in fibroblasts whereas tissue specific proteins are expressed. When placed onto 2D culture substrates, fibroblasts loose the highly polar apical–basal organization of the in vivo situation and acquire an upper (dorsal) and lower (ventral) surface, from where, structures for cell adhesion to the substrate, emerge. The same cells lose this artificial dorsal–ventral polarity when placed back into a mesenchymal 3D matrix, where they regain their in vivo morphology (Amatangelo et al., 2005; Cukierman et al., 2001; Grinnell et al., 2003), thus threedimensionality can physiologically reprogram fibroblasts (Zaman et al., 2006). Moreover, fibroblasts by regulating ECM homeostasis, control elasticity and strength of the tissue mechanical properties, that in turn influence fibroblast behavior. In 3D, such biochemical, mechanical and physical cues were also shown to be fundamental for sustaining cell homeostasis (Chiquet, 1999; Chiquet et al., 2003, 2009; Farran et al., 2010; Hakkinen et al., 2011; Lu et al., 2011). The dynamic conditions as compared to the static ones creating a shear stress and/or fluid flow, can also regulate cell behavior either in 2D or in 3D (Kinney et al., 2011; Wang and Tarbell, 2000; Ng and Swartz, 2003; Shi and Tarbell, 2011). In this work, fetal mouse fibroblasts were immortalized with a SV40-large T antigen and a cell line was cloned and amplified. Its growth characteristics indicated that it expresses an immortalized behavior, in vitro, but not a transformed phenotype as demonstrated by their relative dependence to serum concentration and seeding density. These results were confirmed by the absence of significant tumorigenicity in nude mice (data not shown). These dermal fibroblasts were isolated from ERE-tK-Luciferase mice and were also characterized for their sensitivity to estrogens and endocrine interferents. Immortalized fibroblasts were cultured either in static 2D conditions (Petri dishes) or in the RCCS™ bioreactor that provides an optimum mass transfer (nutrient supply, waste removal) with very low shear stress, often delirious for cells/multicellular aggregates (Freed and Vunjak-Novakovic, 1995; Navran, 2008; Steimberg et al., 2010). Here we show that the organization of fibroblasts grown in the RCCS™, achieves threedimensional structures in which the cells are clearly in a higher steady and differentiation state as indicated by the increased expression in specific differentiation genes and the low level of cell death at one week of culture. Gene expression profiles showed typical expression trends evidenced by the upregulation of genes of focal adhesion, actin cytoskeleton and TGFβ indicating that immortalized fibroblasts grown in this 3D bioreactor revert to differentiated states more similar to those of primary fibroblasts. This is also demonstrated by the expression of typical markers of skin cell differentiation such as MMP2 and TIMP4. The data suggest that in the RCCS™ bioreactor, multicellular aggregates secrete ECM macromolecules and, as the period of culture progresses, they start to remodel the newly synthesized ECM. This is evidenced by the fact that at the spheroids' surface, where no ECM is accumulated, MMP2 is absent since it is not required for remodeling, where instead TIMPs are expressed. At 24 h, whereas cells began to aggregate, ECM probably starts to increase, paralleled by an increased expression of MMP2 and expression of TIMP4 in the outer layer of the spheroid. After one week of culture in the RCCS™, when aggregated cells formed about 1- to 2-mm (diameter) multicellular spheroids, TIMP-4 begin to be re-expressed in some inner areas of the aggregate, suggesting that here the cells and their neighboring ECM components reached a steady state (homeostasis) where the old matrix undergoes degradation to

leave space for the newly synthesized matrix. The synthesis of a skin specific ECM is conformed by the increase in the expression of collagen genes (COL1A1, COL1A2 and COL3A1) (not shown). Thus, the RCCS™ system appears to favor the natural ECM features found in vivo (as it has been observed for other more investigated systems) (Kim, 2005; Lee et al., 2008; Rhee and Grinnell, 2007; Yamada and Cukierman, 2007; Carlson et al., 2008). These results thus confirmed that the 3D organization of fibroblasts in aggregates and their intercellular and cell-ECM interactions play an important role in ECM homeostasis (Langholz et al., 1995; Zigrino et al., 2001). The effects of 3D environments on fibroblasts morphology, including cytoskeletal organization, cell adhesion, cell polarity and intracellular signaling characteristics, have also been shown in other systems such as on cells grown on 3D matrices (Yamada and Cukierman, 2007; Walpita and Hay, 2002). In our 3D RCCS™-based model, the rearrangement of fibroblasts cytoskeleton is also confirmed by the expression of microtubules thought to be essential for the functionality of the cytoskeleton, cell shape, division and growth, and that increase in a time-dependent manner, showing a rearrangement that concerns almost all cells until aggregates, was more stabilized. Subsequently, tubulin expression involves more specifically, even if not exclusively, cells located at the periphery of the spheroid. Several authors have reported that gene expression profiles of cells grown in 3D culture distinctly differs from those found after conventional 2D cultures and more closely resembles those seen in vivo (Kenny et al., 2007; Härmä et al., 2010). In particular, in addition to the spatial arrangement of the cells, it has become evident that the ECM plays a defining role in organizing communication between cells, controlling cell differentiation and modulating response to biochemical signals from the cellular microenvironment (reviewed in Huang and Li, 2011; Kim et al., 2011). In immortalized ERE-LUC-fibroblasts kept in our dynamic 3D culture conditions, genes related to cell cycle, proliferation, cytoskeleton, migration, adhesion and motility were all down-regulated as well as oncogene expression and inflammatory cytokines. Among the induced genes, the pro-apoptotic Bid gene, MMP10 (stromelysin2), an important metalloprotein usually synthesized by keratinocytes involved in skin wound healing and cellular and the PPARg nuclear receptor which plays a role in the regulation of the pro-fibrotic response and inhibits TGF-beta induced collagen gene expression (Ghosh et al., 2004; Sha et al., 2012). Interestingly the growth factors, FGF9 (associated with wound healing,) and FGF6 (mainly involved in myogenesis, osteogenesis, muscle regeneration, bone remodeling and muscle resistance to mechanical stress), were strongly induced, probably as a result of their influence on cell proliferation/differentiation homeostasis. In the actin cytoskeleton organization, MMPs and related enzymes can degrade components of the ECM. In addition, MMPs acts on non-matrix substrates by modifying and/or activating other MMPs, growth factors, cytokines, chemokines (Imai et al., 1995; Karsdal et al., 2002; Schönbeck et al., 1998) and other growth regulators that are matrix-bound or present on cell surface (Li et al., 2002). Moreover, the family of TIMP factors, including TIMP4, has been shown to contribute not only to MMP inhibition, but also to the regulation of apoptosis (Guo et al., 2004; Tummalapalli et al., 2001). The activity of these enzymes and relative inhibitors is thus critical to the maintenance and turnover of connective tissues (Howard et al., 1983). These were maintained in our 3D culture conditions as a function of ECM–cell interactions and homeostasis. The acquisition of a greater differentiation state also affects the ability of cells to better respond to external stimuli thus promoting advances also in the development of more reliable systems for studying cellular signaling and toxicology (Mazzoleni and Steimberg, 2010). This would also be applied to the generation of new screening methods for therapeutics and xenobiotics (Yamada and Cukierman, 2007). As a matter of fact, the development of 3D model systems in pharmacotoxicology has recently emerged as important tools to capture in vitro the complex responses to physiological signals and xenobiotics of the 3D tissue

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

physiology (Grinnell, 2003; Griffith and Swartz, 2006) (Krause et al., 2008; Mazzoleni et al., 2009; Mazzoleni and Steimberg, 2012; Lan and Starly, 2011; Breslin and O'Driscoll, 2013). This has been investigated here for estrogens, which compose the largest family of exogenous hormonally active compounds of great interest for the therapy of several diseases, as food bioactives and as contaminating chemicals of industrial production. Our approach was based on the observation from other authors, who have shown that the treatment of cell lines with physiological doses of estradiol lead to a rapid remodeling of the actin cytoskeleton with a loss of stress fibers (Giretti et al., 2008). The higher expression of ERα in our 3D culture conditions compared to cells grown in 2D and the greater sensitivity to low physiological doses of estradiol (1 nM), as also assessed on a consensus estrogen-dependent reporter (ERE-Luciferase), support the fact that ER activity tends toward optimization when 3D cell structures are better expressed and functional. Infact, while E2 showed no effect on the several analyzed genes in 2D, it had the ability to upregulate several of them in the RCCS™, including CBP, Rarb, MMP1 and Dbp. The FGF2, MMP1 and Dbp genes are known to respond to estrogens through consensus sequences, while CBP is an nuclear receptor coactivator for which a molecular mechanism of regulation by estrogens in not yet illustrated (Tozlu et al., 2006; Fogarty et al., 2012; Pirani et al., 2013; Skrzypczak et al., 2013). The regulation of Aldh1a7 and RARβ was identified by us, but no molecular studies are yet available. The strong basal induction of the Aldh1a7 and FGF2 genes already observed as a result of the culture method (Fig. 7A) might have masked a further upregulation by E2 in 3D, while it was well detected in 2D. However, we cannot attribute the higher sensitivity to E2 solely to the slight higher expression and synthesis of ERα. Knowing the complexity of the ER-regulated transcription machinery the expression of several ER co-regulators not investigated in this work can play a significant role (Feng and O'Malley, 2014). Also genistein, a food estrogen, was efficient in inducing the same reporter through an inverted U-shape dose-dependent mechanism as was observed for the E2, (typical of ligands of nuclear receptors) (Taylor et al., 2012; Mlynarcikova et al., 2013; Penza et al., 2006; Montani et al., 2008) and the same genes Thus, while the mechanisms modulating the different regulations of estrogen regulated genes in the two culture systems remain to be finely elucidated, it becomes evident that when performing estrogenicity tests, cell sensitivity to estrogen is related to the state of cell differentiation and is thus of critical importance. This might prone us to reconsider the reliability, specificity and sensitivity of existing and validated in vitro methods for assessing hormone regulation. The results shown in this work confirms that 3D culture conditions, in addition to keep cells in a more differentiated state, may better regulate cell sensitivity to hormones and/or xenobiotics (Meng et al., 2006; Yin et al., 2009; Weigelt et al., 2010; Tung et al., 2011). In summary, we show here that differentiated functions of cells usually lost either in 2D cultures or as a result of cell immortalization, can be restored when the cells are placed in a 3D microenvironment that favors high mass transfer and low shear stress. Moreover, in such conditions, we demonstrated that estrogen receptor functions are more sensitively modulated by ligands, indicating that the cells are in a state of higher responsiveness to estrogens. Acknowledgments We are grateful to Alessandro Bulla and Francesco Sainato for their helpful English writing and secretarial assistance. This work was supported in part by the European Union Grants QLK4-CT-2002-02221 (EDERA) and LSHB-CT-2006-037168 (EXERA) from the Ministero dell'Università e della Ricerca Scientifica and FONDAZIONE CARIPLO (Grant 2006) (Grant n° 2006.0762 / 11.6457). Conflict of interest statement The authors declare no conflicts of interest related to this work.

431

References Amatangelo, M.D., Bassi, D.E., Klein-Szanto, A.J., Cukierman, E., 2005. Stroma-derived three-dimensional matrices are necessary and sufficient to promote desmoplastic differentiation of normal fibroblasts. Am. J. Pathol. 167, 475–488. Ananthakrishnan, R., Ehrlicher, A., 2007. The forces behind cell movement. Int. J. Biol. Sci. 3, 303–317. Baker, B.M., Chen, C.S., 2012. Deconstructing the third dimension: how 3D culture microenvironments alter cellular cues. J. Cell Sci. 125, 3015–3024. Beacham, D.A., Cukierman, E., 2005. Stromagenesis: the changing face of fibroblastic microenvironments during tumor progression. Semin. Cancer Biol. 15, 329–341. Beningo, K.A., Dembo, M., Wang, Y.L., 2004. Responses of fibroblasts to anchorage of dorsal extracellular matrix receptors. Proc. Natl. Acad. Sci. U. S. A. 101, 18024–18029. Benjamini, Y., Hochberg, Y., 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B 57. Birgersdotter, A., Sandberg, R., Ernberg, I., 2005. Gene expression perturbation in vitro—a growing case for three-dimensional (3D) culture systems. Semin. Cancer Biol. 15, 405–412. Boateng, S.Y., Hartman, T.J., Ahluwalia, N., Vidula, H., Desai, T.A., Russell, B., 2003. Inhibition of fibroblast proliferation in cardiac myocyte cultures by surface microtopography. Am. J. Physiol. Cell Physiol. 285, C171–C182. Bracken, M.B., 2009. Why animal studies are often poor predictors of human reactions to exposure. J. R. Soc. Med. 102, 120–122. Breslin, S., O'Driscoll, L., 2013. Three-dimensional cell culture: the missing link in drug discovery. Drug Discov. Today 18, 240–249. Broughton, G., Janis, J.E., Attinger, C.E., 2006. Wound healing: an overview. Plast. Reconstr. Surg. 117, 1e-S–32e-S. Carlson, M.W., Alt-Holland, A., Egles, C., Garlick, J.A., 2008. Three-dimensional tissue models of normal and diseased skin. Curr. Protoc. Cell Biol. suppl. 41, 1–17 (Chapter 19, Unit 19.19). Castelló-Cros, R., Cukierman, E., 2009. Stromagenesis during tumorigenesis: characterization of tumor-associated fibroblasts and stroma-derived 3D matrices. Methods Mol. Biol. 522, 275–305. Chang, H.Y., Chi, J.T., Dudoit, S., Bondre, C., van de Rijn, M., Botstein, D., Brown, P.O., 2002. Diversity, topographic differentiation, and positional memory in human fibroblasts. Proc. Natl. Acad. Sci. U. S. A. 99, 12877–12882. Chiquet, M., 1999. Regulation of extracellular matrix gene expression by mechanical stress. Matrix Biol. 18, 417–426. Chiquet, M., Renedo, A.S., Huber, F., Flück, M., 2003. How do fibroblasts translate mechanical signals into changes in extracellular matrix production? Matrix Biol. 22, 73–80. Chiquet, M., Gelman, L., Lutz, R., Maier, S., 2009. From mechanotransduction to extracellular matrix gene expression in fibroblasts. Biochim. Biophys. Acta 1793, 911–920. Cook, N., Jodrell, D.I., Tuveson, D.A., 2012. Predictive in vivo animal models and translation to clinical trials. Drug Discov. Today 17, 253–260. Cukierman, E., Pankov, R., Stevens, D.R., Yamada, K.M., 2001. Taking cell–matrix adhesions to the third dimension. Science 294, 1708–1712. Cukierman, E., Pankov, R., Yamada, K.M., 2002. Cell interactions with three-dimensional matrices. Curr. Opin. Cell Biol. 14, 633–639. Desmoulière, A., Chaponnier, C., Gabbiani, G., 2005. Tissue repair, contraction, and the myofibroblast. Wound Repair Regen. 13, 7–12. Elsdale, T., Bard, J., 1972. Collagen substrata for studies on cell behavior. J. Cell Biol. 54, 626–637. Farran, A.J., Teller, S.S., Jha, A.K., Jiao, T., Hule, R.A., Clifton, R.J., Pochan, D.P., Duncan, R.L., Jia, X., 2010. Effects of matrix composition, microstructure, and viscoelasticity on the behaviors of vocal fold fibroblasts cultured in three-dimensional hydrogel networks. Tissue Eng. Part A 16, 1247–1261. Feng, Q., O'Malley, B.W., 2014. Nuclear receptor modulation — role of coregulators in selective estrogen receptor modulator (SERM) actions. Steroids (Jun 16. pii: S0039128X(14)00147-0). http://dx.doi.org/10.1016/j.steroids.2014.06.008. Ferrarini, M., Mazzoleni, G., Steimberg, N., Belloni, D., Ferrero, E., 2013a. Innovative models to assess multiple myeloma biology and the impact of drugs. In: P.R.H. (Ed.), Multiple Myeloma — A Quick Reflection on the Fast Progress. InTech (pp.). Ferrarini, M., Steimberg, N., Ponzoni, M., Belloni, D., Berenzi, A., Girlanda, S., CaligarisCappio, F., Mazzoleni, G., Ferrero, E., 2013b. Ex-vivo dynamic 3-D culture of human tissues in the RCCS™ bioreactor allows the study of multiple myeloma biology and response to therapy. PLoS One 8, e71613. Fogarty, E.A., Matulis, C.K., Kraus, W.L., 2012. Activation of estrogen receptor α by raloxifene through an activating protein-1-dependent tethering mechanism in human cervical epithelial cancer cells: a role for c-Jun N-terminal kinase. Mol. Cell. Endocrinol. 1, 331–338. Freed, L.E., Vunjak-Novakovic, G., 1995. Cultivation of cell–polymer tissue constructs in simulated microgravity. Biotechnol. Bioeng. 46, 306–313. Fu, X.D., Simoncini, T., 2008. Extra-nuclear signaling of estrogen receptors. IUBMB Life 60, 502–510. Gentleman, R.C., Carey, V.J., Bates, D.M., Bolstad, B., Dettling, M., Dudoit, S., Ellis, B., Gautier, L., Ge, Y., Gentry, J., Hornik, K., Hothorn, T., Huber, W., Iacus, S., Irizarry, R., Leisch, F., Li, C., Maechler, M., Rossini, A.J., Sawitzki, G., Smith, C., Smyth, G., Tierney, L., Yang, J.Y., Zhang, J., 2004. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 5, R80. Ghosh, A.K., Bhattacharyya, S., Lakos, G., Chen, S.J., Mori, Y., Varga, J., 2004. Disruption of transforming growth factor beta signaling and profibrotic responses in normal skin fibroblasts by peroxisome proliferator-activated receptor gamma. Arthritis Rheum. 50, 1305–1318. Giretti, M.S., Simoncini, T., 2008. Rapid regulatory actions of sex steroids on cell movement through the actin cytoskeleton. Steroids 73, 895–900.

432

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433

Giretti, M.S., Fu, X.D., De Rosa, G., Sarotto, I., Baldacci, C., Garibaldi, S., Mannella, P., Biglia, N., Sismondi, P., Genazzani, A.R., Simoncini, T., 2008. Extra-nuclear signalling of estrogen receptor to breast cancer cytoskeletal remodelling, migration and invasion. PLoS One 3, e2238. Griffith, L.G., Swartz, M.A., 2006. Capturing complex 3D tissue physiology in vitro. Nat. Rev. Mol. Cell Biol. 7, 211–224. Grinnell, F., 2003. Fibroblast biology in three-dimensional collagen matrices. Trends Cell Biol. 13, 264–269. Grinnell, F., Ho, C.H., Tamariz, E., Lee, D.J., Skuta, G., 2003. Dendritic fibroblasts in threedimensional collagen matrices. Mol. Biol. Cell 14, 384–395. Grun, B., Benjamin, E., Sinclair, J., Timms, J.F., Jacobs, I.J., Gayther, S.A., Dafou, D., 2009. Three-dimensional in vitro cell biology models of ovarian and endometrial cancer. Cell Prolif. 42, 219–228. Guo, Y.H., Gao, W., Li, Q., Li, P.F., Yao, P.Y., Chen, K., 2004. Tissue inhibitor of metalloproteinases-4 suppresses vascular smooth muscle cell migration and induces cell apoptosis. Life Sci. 75, 2483–2493. Hakkinen, K.M., Harunaga, J.S., Doyle, A.D., Yamada, K.M., 2011. Direct comparisons of the morphology, migration, cell adhesions, and actin cytoskeleton of fibroblasts in four different three-dimensional extracellular matrices. Tissue Eng. Part A 17, 713–724. Hammond, T.G., Hammond, J.M., 2001. Optimized suspension culture: the rotating-wall vessel. Am. J. Physiol. Renal Physiol. 281, F12–F25. Härmä, V., Virtanen, J., Mäkelä, R., Happonen, A., Mpindi, J.P., Knuuttila, M., Kohonen, P., Lötjönen, J., Kallioniemi, O., Nees, M., 2010. A comprehensive panel of threedimensional models for studies of prostate cancer growth, invasion and drug responses. PLoS One 5, e10431. Hartung, T., 2008. Food for thought … on animal tests. ALTEX 25, 3–16. Hinz, B., Gabbiani, G., 2003. Mechanisms of force generation and transmission by myofibroblasts. Curr. Opin. Biotechnol. 14, 538–546. Hirschhaeuser, F., Menne, H., Dittfeld, C., West, J., Mueller-Klieser, W., Kunz-Schughart, L. A., 2010. Multicellular tumor spheroids: an underestimated tool is catching up again. J. Biotechnol. 148, 3–15. Howard, G.A., Turner, R.T., Puzas, J.E., Nichols, F., Baylink, D.J., 1983. Bone cells on microcarrier spheres. JAMA 249, 258–259. Huang, N.F., Li, S., 2011. Regulation of the matrix microenvironment for stem cell engineering and regenerative medicine. Ann. Biomed. Eng. 39, 1201–1214. Huang, G., Wang, L., Wang, S., Han, Y., Wu, J., Zhang, Q., Xu, F., Lu, T.J., 2012. Engineering three-dimensional cell mechanical microenvironment with hydrogels. Biofabrication 4, 042001. Imai, K., Yokohama, Y., Nakanishi, I., Ohuchi, E., Fujii, Y., Nakai, N., Okada, Y., 1995. Matrix metalloproteinase 7 (matrilysin) from human rectal carcinoma cells. Activation of the precursor, interaction with other matrix metalloproteinases and enzymic properties. J. Biol. Chem. 270, 6691–6697. Irizarry, R.A., Hobbs, B., Collin, F., Beazer-Barclay, Y.D., Antonellis, K.J., Scherf, U., Speed, T.P., 2003. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4, 249–264. Kalluri, R., Zeisberg, M., 2006. Fibroblasts in cancer. Nat. Rev. Cancer 6, 392–401. Kanehisa, M., Goto, S., Furumichi, M., Tanabe, M., Hirakawa, M., 2010. KEGG for representation and analysis of molecular networks involving diseases and drugs. Nucleic Acids Res. 38, D355–D360. Karsdal, M.A., Larsen, L., Engsig, M.T., Lou, H., Ferreras, M., Lochter, A., Delaissé, J.M., Foged, N.T., 2002. Matrix metalloproteinase-dependent activation of latent transforming growth factor-beta controls the conversion of osteoblasts into osteocytes by blocking osteoblast apoptosis. J. Biol. Chem. 277, 44061–44067. Kenny, P.A., Lee, G.Y., Myers, C.A., Neve, R.M., Semeiks, J.R., Spellman, P.T., Lorenz, K., Lee, E.H., Barcellos-Hoff, M.H., Petersen, O.W., Gray, J.W., Bissell, M.J., 2007. The morphologies of breast cancer cell lines in three-dimensional assays correlate with their profiles of gene expression. Mol. Oncol. 1, 84–96. Kim, J.B., 2005. Three-dimensional tissue culture models in cancer biology. Semin. Cancer Biol. 15, 365–377. Kim, S.H., Turnbull, J., Guimond, S., 2011. Extracellular matrix and cell signalling: the dynamic cooperation of integrin, proteoglycan and growth factor receptor. J. Endocrinol. 209, 139–151. Kinney, M.A., Sargent, C.Y., McDevitt, T.C., 2011. The multiparametric effects of hydrodynamic environments on stem cell culture. Tissue Eng. Part B Rev. 17, 249–262. Kisseleva, T., Brenner, D.A., 2008a. Fibrogenesis of parenchymal organs. Proc. Am. Thorac. Soc. 5, 338–342. Kisseleva, T., Brenner, D.A., 2008b. Mechanisms of fibrogenesis. Exp. Biol. Med. (Maywood) 233, 109–122. Krause, S., Maffini, M.V., Soto, A.M., Sonnenschein, C., 2008. A novel 3D in vitro culture model to study stromal–epithelial interactions in the mammary gland. Tissue Eng. Part C Methods 14, 261–271. Lan, S.F., Starly, B., 2011. Alginate based 3D hydrogels as an in vitro co-culture model platform for the toxicity screening of new chemical entities. Toxicol. Appl. Pharmacol. 256, 62–72. Langholz, O., Röckel, D., Mauch, C., Kozlowska, E., Bank, I., Krieg, T., Eckes, B., 1995. Collagen and collagenase gene expression in three-dimensional collagen lattices are differentially regulated by alpha 1 beta 1 and alpha 2 beta 1 integrins. J. Cell Biol. 131, 1903–1915. Larsen, M., Artym, V.V., Green, J.A., Yamada, K.M., 2006. The matrix reorganized: extracellular matrix remodeling and integrin signaling. Curr. Opin. Cell Biol. 18, 463–471. Lee, M.Y., Kumar, R.A., Sukumaran, S.M., Hogg, M.G., Clark, D.S., Dordick, J.S., 2008. Threedimensional cellular microarray for high-throughput toxicology assays. Proc. Natl. Acad. Sci. U. S. A. 105, 59–63. Li, Q.L., Wang, H.M., Lin, H.Y., Liu, D.L., Zhang, X., Liu, G.Y., Qian, D., Zhu, C., 2002. Expression of gelatinases and their tissue inhibitors in rat corpus luteum during pregnancy and postpartum. Mol. Reprod. Dev. 63, 273–281.

Louvet-Vallée, S., 2000. ERM proteins: from cellular architecture to cell signaling. Biol. Cell 92, 305–316. Lu, F., Ogawa, R., Nguyen, D.T., Chen, B., Guo, D., Helm, D.L., Zhan, Q., Murphy, G.F., Orgill, D.P., 2011. Microdeformation of three-dimensional cultured fibroblasts induces gene expression and morphological changes. Ann. Plast. Surg. 66, 296–300. Mazzoleni, G., Steimberg, N., 2010. 3D culture in microgravity: a realistic alternative to experimental animal use. ALTEX 57, 321–324. Mazzoleni, G., Steimberg, N., 2012. New models for the in vitro study of liver toxicity: 3D culture systems and the role of bioreactors. In: Tyshenko, D.M.G. (Ed.), The Continuum of Health Risk Assessments. InTech, pp. 162–194. Mazzoleni, G., Di Lorenzo, D., Steimberg, N., 2009. Modelling tissues in 3D: the next future of pharmaco-toxicology and food research? Genes Nutr. 4, 13–22. Mazzoleni, G., Boukhechba, F., Steimberg, N., Boniotti, J., Bouler, J., Rochet, N., 2011. Impact of dynamic culture in the RCCS™ bioreactor on a three-dimensional model of bone matrix formation. International Congress on the Mechanical Behavior of Materials (ICM11), Como Lake, Italy (pp.). Meng, Q., Zhang, G., Shen, C., Qiu, H., 2006. Sensitivities of gel entrapped hepatocytes in hollow fibers to hepatotoxic drug. Toxicol. Lett. 166, 19–26. Mlynarcikova, A., Macho, L., Fickova, M., 2013. Bisphenol A alone or in combination with estradiol modulates cell cycle- and apoptosis-related proteins and genes in MCF7 cells. Endocr. Regul. 4, 189–199. Montani, C., Penza, M., Jeremic, M., Biasiotto, G., La Sala, G., De Felici, M., Ciana, P., Maggi, A., Di Lorenzo, D., 2008. Genistein is an efficient estrogen in the whole-body throughout mouse development. Toxicol. Sci. 103, 57–67. Moskaluk, C.A., Stoler, M.H., 2002. Agarose mold embedding of cultured cells for tissue microarrays. Diagn. Mol. Pathol. 11, 234–238. Naftolin, F., Malaspina, D., 2007. Estrogen, estrogen treatment and the post-reproductive woman's brain. Maturitas 57, 23–26. Navran, S., 2008. The application of low shear modeled microgravity to 3-D cell biology and tissue engineering. Biotechnol. Annu. Rev. 14, 275–296. Nelson, C.M., Bissell, M.J., 2006. Of extracellular matrix, scaffolds, and signaling: tissue architecture regulates development, homeostasis, and cancer. Annu. Rev. Cell Dev. Biol. 22, 287–309. Ng, C.P., Swartz, M.A., 2003. Fibroblast alignment under interstitial fluid flow using a novel 3-D tissue culture model. Am. J. Physiol. Heart Circ. Physiol. 284, H1771–H1777. Ostman, A., Augsten, M., 2009. Cancer-associated fibroblasts and tumor growth— bystanders turning into key players. Curr. Opin. Genet. Dev. 19, 67–73. Pampaloni, F., Reynaud, E.G., Stelzer, E.H., 2007. The third dimension bridges the gap between cell culture and live tissue. Nat. Rev. Mol. Cell Biol. 8, 839–845. Pavlakis, K., Messini, I., Vrekoussis, T., Yiannou, P., Keramopoullos, D., Louvrou, N., Liakakos, T., Stathopoulos, E.N., 2008. The assessment of angiogenesis and fibroblastic stromagenesis in hyperplastic and pre-invasive breast lesions. BMC Cancer 8, 88. Pedersen, J.A., Swartz, M.A., 2005. Mechanobiology in the third dimension. Ann. Biomed. Eng. 33, 1469–1490. Penza, M., Montani, C., Romani, A., Vignolini, P., Pampaloni, B., Tanini, A., Brandi, M.L., Alonso-Magdalena, P., Nadal, A., Ottobrini, L., Parolini, O., Bignotti, E., Calza, S., Maggi, A., Grigolato, P.G., Di Lorenzo, D., 2006. Genistein affects adipose tissue deposition in a dose-dependent and gender-specific manner. Endocrinology 12, 5740–5751. Pirani, T., Chen, J., Vieira, A., 2013. Effects of estradiol on the endocytic transport of vitamin D carrier protein in hepatocytes. Biochim. Biophys. Acta 6, 3421–3426. Rangarajan, A., Hong, S.J., Gifford, A., Weinberg, R.A., 2004. Species- and cell type-specific requirements for cellular transformation. Cancer Cell 6, 171–183. Rhee, S., 2009. Fibroblasts in three dimensional matrices: cell migration and matrix remodeling. Exp. Mol. Med. 41, 858–865. Rhee, S., Grinnell, F., 2007. Fibroblast mechanics in 3D collagen matrices. Adv. Drug Deliv. Rev. 59, 1299–1305. Sanchez, A.M., Simoncini, T., 2010. Extra-nuclear signaling of ERalpha to the actin cytoskeleton in the central nervous system. Steroids 75, 528–532. Sanchez, A.M., Flamini, M.I., Fu, X.D., Mannella, P., Giretti, M.S., Goglia, L., Genazzani, A.R., Simoncini, T., 2009. Rapid signaling of estrogen to WAVE1 and moesin controls neuronal spine formation via the actin cytoskeleton. Mol. Endocrinol. 23, 1193–1202. Sarasa-Renedo, A., Chiquet, M., 2005. Mechanical signals regulating extracellular matrix gene expression in fibroblasts. Scand. J. Med. Sci. Sports 15, 223–230. Schönbeck, U., Mach, F., Libby, P., 1998. Generation of biologically active IL-1 beta by matrix metalloproteinases: a novel caspase-1-independent pathway of IL-1 beta processing. J. Immunol. 161, 3340–3346. Sha, W., Thompson, K., South, J., Baron, M., Leask, A., 2012. Loss of PPARγ expression by fibroblasts enhances dermal wound closure. Fibrogenesis Tissue Repair 5, 5. Shaw, T.J., Martin, P., 2009. Wound repair at a glance. J. Cell Sci. 122, 3209–3213. Shi, Z.D., Tarbell, J.M., 2011. Fluid flow mechanotransduction in vascular smooth muscle cells and fibroblasts. Ann. Biomed. Eng. 39, 1608–1619. Silver, F.H., Siperko, L.M., Seehra, G.P., 2003. Mechanobiology of force transduction in dermal tissue. Skin Res. Technol. 9, 3–23. Simoncini, T., Mannella, P., Genazzani, A.R., 2006. Rapid estrogen actions in the cardiovascular system. Ann. N. Y. Acad. Sci. 1089, 424–430. Skrzypczak, M., Kapka-Skrzypczak, L., Cyranka, M., Treeck, O., Wrobel, A., Matosiuk, D., 2013. Nuclear estrogen receptors co-activation mechanisms. Curr. Med. Chem. 27, 3317–3338. Smyth, G.K., 2004. Linear models and empirical Bayes methods for assessing differential expression in microarray experiments. Stat. Appl. Genet. Mol. Biol. 3 (Article3). Sonneveld, E., Riteco, J.A., Jansen, H.J., Pieterse, B., Brouwer, A., Schoonen, W.G., van der Burg, B., 2006. Comparison of in vitro and in vivo screening models for androgenic and estrogenic activities. Toxicol. Sci. 89, 173–187. Sorrell, J.M., Caplan, A.I., 2009. Fibroblasts—a diverse population at the center of it all. Int. Rev. Cell Mol. Biol. 276, 161–214.

C. Montani et al. / Toxicology and Applied Pharmacology 280 (2014) 421–433 Steimberg, N., Boniotti, J., Mazzoleni, G., 2010. 3D culture of primary chondrocytes, cartilage, and bone/cartilage explants in simulated microgravity. In: Tim Maguire, E.N. (Ed.), Methods in Bioengineering: Alternative Technologies to Animal Testing. Artech House, Boston, USA, pp. 205–212. Taylor, J.A., Richter, C.A., Suzuki, A., Watanabe, H., Iguchi, T., Coser, K.R., Shioda, T., vom Saal, F.S., 2012. Dose-related estrogen effects on gene expression in fetal mouse prostate mesenchymal cells. PLoS One 10, e48311. Torres, E., Rosen, M.K., 2006. Protein-tyrosine kinase and GTPase signals cooperate to phosphorylate and activate Wiskott–Aldrich syndrome protein (WASP)/neuronal WASP. J. Biol. Chem. 281, 3513–3520. Tozlu, S., Girault, I., Vacher, S., Vendrell, J., Andrieu, C., Spyratos, F., Cohen, P., Lidereau, R., Bieche, I., 2006. Identification of novel genes that co-cluster with estrogen receptor alpha in breast tumor biopsy specimens, using a large-scale real-time reverse transcription-PCR approach. Endocr. Relat. Cancer 4, 1109–1120. Tsukita, S., Yonemura, S., 1999. Cortical actin organization: lessons from ERM (ezrin/ radixin/moesin) proteins. J. Biol. Chem. 274, 34507–34510. Tummalapalli, C.M., Heath, B.J., Tyagi, S.C., 2001. Tissue inhibitor of metalloproteinase-4 instigates apoptosis in transformed cardiac fibroblasts. J. Cell. Biochem. 80, 512–521. Tung, Y.C., Hsiao, A.Y., Allen, S.G., Torisawa, Y.S., Ho, M., Takayama, S., 2011. Highthroughput 3D spheroid culture and drug testing using a 384 hanging drop array. Analyst 136, 473–478. Unsworth, B.R., Lelkes, P.I., 1998. Growing tissues in microgravity. Nat. Med. 4, 901–907. Walpita, D., Hay, E., 2002. Studying actin-dependent processes in tissue culture. Nat. Rev. Mol. Cell Biol. 3, 137–141.

433

Wang, S., Tarbell, J.M., 2000. Effect of fluid flow on smooth muscle cells in a 3-dimensional collagen gel model. Arterioscler. Thromb. Vasc. Biol. 20, 2220–2225. Weigelt, B., Lo, A.T., Park, C.C., Gray, J.W., Bissell, M.J., 2010. HER2 signaling pathway activation and response of breast cancer cells to HER2-targeting agents is dependent strongly on the 3D microenvironment. Breast Cancer Res. Treat. 122, 35–43. Weigelt, B., Ghajar, C.M., Bissell, M.J., 2014. The need for complex 3D culture models to unravel novel pathways and identify accurate biomarkers in breast cancer. Adv. Drug Deliv. Rev. 69, 42–51. Xu, R., Boudreau, A., Bissell, M.J., 2009. Tissue architecture and function: dynamic reciprocity via extra- and intra-cellular matrices. Cancer Metastasis Rev. 28, 167–176. Yamada, K.M., Cukierman, E., 2007. Modeling tissue morphogenesis and cancer in 3D. Cell 130, 601–610. Yin, J., Meng, Q., Zhang, G., Sun, Y., 2009. Differential methotrexate hepatotoxicity on rat hepatocytes in 2-D monolayer culture and 3-D gel entrapment culture. Chem. Biol. Interact. 180, 368–375. Zaman, M.H., Trapani, L.M., Sieminski, A.L., Siemeski, A., Mackellar, D., Gong, H., Kamm, R. D., Wells, A., Lauffenburger, D.A., Matsudaira, P., 2006. Migration of tumor cells in 3D matrices is governed by matrix stiffness along with cell–matrix adhesion and proteolysis. Proc. Natl. Acad. Sci. U. S. A. 103, 10889–10894. Zigrino, P., Drescher, C., Mauch, C., 2001. Collagen-induced proMMP-2 activation by MT1-MMP in human dermal fibroblasts and the possible role of alpha2beta1 integrins. Eur. J. Cell Biol. 80, 68–77.