Advanced Drug Delivery Reviews xxx (2014) xxx–xxx

ADR-12622; No of Pages 10

Contents lists available at ScienceDirect

Advanced Drug Delivery Reviews

j ourna l homepage: www.e lsev ie r .com/ locate /addr

Bioengineering 3D environments for cancer models☆

Mireia Alemany-Ribes, Carlos E. Semino ⁎Department of Bioengineering, IQS-School of Engineering, Ramon Llull University, Via Augusta 390, Barcelona 08017, Spain

Abbreviations: EMT, epithelial–mesenchymal transit2D, two-dimensional; 3D, three-dimensional; PEGpoly(lactide-co-glycolide); PLA, polylactic acid; RGD, argisequence;MMPs,matrixmetalloproteinases; KLD12, KLDLtide; RAD16-I, RADARADARADARADA self-assembling pnase; AV, arterio-venous; PDMS, polydimethylsiloxAdministration; NCI, National Cancer Institute; HTS, high☆ This review is part of the Advanced Drug Delivery Revieof tumor microenvironments”.⁎ Corresponding author at: Department of Bioeng

Laboratory, Institut Químic de Sarrià, Ramon Llull UniveBarcelona, Spain. Tel.: +34 93 267 2000; fax: +34 93 20

E-mail address: carlos.semino@iqs.url.edu (C.E. Semin

http://dx.doi.org/10.1016/j.addr.2014.06.0040169-409X/© 2014 Elsevier B.V. All rights reserved.

Please cite this article as:M. Alemany-Ribes, Cdx.doi.org/10.1016/j.addr.2014.06.004

a b s t r a c t

a r t i c l e i n f o

Available online xxxx

Keywords:TumorigenesisTumor tissue engineeringThree-dimensional cultureNanotechnologyBiomaterialsCancer modelsDrug screeningDrug resistance

Tumor development is a dynamic process where cancer cells differentiate, proliferate and migrate interactingamong each other and with the surrounding matrix in a three-dimensional (3D) context. Interestingly, theprocess follows patterns similar to those involved in early tissue formation by accessing specific genetic programsto grow and disseminate. Thus, the complex biological mechanisms driving tumor progression cannot easily berecreated in the laboratory. Yet, essential tumor stages, including epithelial–mesenchymal transition (EMT),tumor-induced angiogenesis and metastasis, urgently need more realistic models in order to unravel the under-lyingmolecular and cellularmechanisms that govern them. The latest implementation of successful 3Dmodels ishaving a positive impact on the fight against cancer by obtaining more predictive systems for pre-clinical re-search, therapeutic drug screening, and early cancer diagnosis. In this review we explore the latest advancesand challenges in tumor tissue engineering, by accessing knowledge and tools from cancer biology, materialscience and bioengineering.

© 2014 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 01.1. Paradigm shift: mimicking tumor progression through the third dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0

2. Biomaterials: applying biomimetic principles to study cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 03. Cancer models: recreating tumor progression step by step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0

3.1. Limitless cellular proliferation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 03.2. Sustained angiogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 03.3. Tissue invasion and metastasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0

4. Cells: maintaining tumor identity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 05. Unresolved issue: high throughput screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 06. Future directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0

ion; ECM, extracellular matrix;, polyethylene glycol; PLG,nine–glycine–aspartate peptideKLDLKLDL self-assembling pep-eptide; FAK, focal adhesion ki-ane; FDA, Food and Drugthroughput screening.ws theme issue on “Engineering

ineering, Tissue Engineeringrsity, Via Augusta 390, 08017,56266.o).

.E. Semino, Bioengineering 3D

1. Introduction

1.1. Paradigm shift: mimicking tumor progression through the thirddimension

Understanding the underlying biology in tumor initiation and pro-gression is the first step to a successful breakthrough in the develop-ment of new and efficient cancer therapies. To achieve this goal, thecomplex cellular microenvironment needs to be deconstructed intosimpler and more predictable systems. This approach helps researchersto identify and analyze the role of key chemical, mechanical and/orphysical factors that might drive human pathophysiology.

environments for cancermodels, Adv. DrugDeliv. Rev. (2014), http://

2 M. Alemany-Ribes, C.E. Semino / Advanced Drug Delivery Reviews xxx (2014) xxx–xxx

Following this framework, cancer research has traditionally relied ontwo-dimensional (2D) cultures [1,2]. However, it is commonly acceptedthat cells grow in non-physiologically constrained conditions on thesesurfaces. In particular, cells are attached to rigid and flat substrates,which force them to polarize and increase their exchange area to culturemedia. As a consequence, they are subjected to excessive nutrition andoxygenation and molecular gradients cannot be reproduced. Further-more, production of extracellular matrix (ECM) proteins is stronglymodified – in composition, configuration and amount – due to differ-ences in the surface receptors' orientation and clustering and, conse-quently, cells do not receive the proper signals that arise from naturalECM configuration [3–5]. Specifically in the field of cancer, poorly ad-herent cells – metastatic cells – cannot form tight focal adhesions and,as a consequence, are not easily cultured in classical cell culture dishes.The outcome obtainedwithmetastatic cells in drug screening processesunder 2D culture conditions is limited [6]. 2D cultures also activate animmortalization process through multiple passages, which result inthe selection of cancer cells that rapidly proliferate. These cells misrep-resent the whole tumor, since they are specifically susceptible to thera-pies that target rapidly dividing cells [7].

To avoid these experimental inconsistencies, it is essential todevelop models with a higher degree of complexity while retainingthe reproducibility and the capacity of cellular level imaging. Firststeps have been focused on the generation of multicellular spheroids,whichhave taken cancer biology to the third dimension (3D) and exem-plified the application of biomimetic principles in research. Spheroidcultures partially reestablish 3D tumor architecture patterns sincethey create hollow cores that contain quiescent and hypoxic cells. Inter-estingly, spheroids exhibit greater anticancer drug resistance as com-pared to conventional monolayer cultures [8,9]. However, they haveimportant limitations since they grow as independent cellular aggre-gates and show reduced interactions with the extracellular milieu [3,10]. Considering that microenvironment controls tumorigenesis, ECManalogs have been introduced as cell culture systems in order toembed cells in a 3D context and display the appropriate physical,

Fig. 1. Bioengineers have developed ECM analogs as 3D culture systems, applying biomimetismbetter comprehend disease pathogenesis of tumors. Nature is used as a source of ideas to obtaisignals that arise from the ECM (biomimetism). Due to the complexity and interaction amongvironment into simpler and predictable models that enable the analysis and identification of thImage of a disassembled telephone courtesy of J. P. Wiegmann and My Modern Metropolis.

Please cite this article as:M. Alemany-Ribes, C.E. Semino, Bioengineering 3Ddx.doi.org/10.1016/j.addr.2014.06.004

chemical and mechanical cues for cell fates (Fig. 1). Pioneering workhas been based on the use of biomaterials from natural origins,principally Matrigel and collagen. Experiments have revealed that phe-notypical differences betweenmalignant and normal epithelial cells canbe exclusively observed in 3D cultures, in whichmalignant cells lose tis-sue polarity and organization, phenomena not commonly detected in2D. Therefore, the remarkable plasticity of cancer cells under differentexperimental conditions can be easily reproduced by using 3D cultures,which enable reestablishment in vitro of crosstalk among neighboringcells and their surrounding stroma [11–13].

Cancer research has experienced a paradigm shift during the pasttwo decades. However, many groups from academia and the biomedicalindustry still routinely use 2D cultures, which provide unreliable dataand, thus, hamper the discovery and therapeutic assessment of cancerdrugs [14,15]. Multiple data illustrate the slow progress in cancer drugresearch and development. It is estimated that a 10 to 12 year cycle isneeded to develop a new cancer drug and candidates that enter clinicaltrials have only a 5% probability of receiving approval from theU.S. Foodand Drug Administration (FDA) [1,7,16]. 3D cultures may be a viablealternative to expedite theprocess frombench to bedside. The appropri-ate model design should help to identify key factors regulating tumordevelopment such as cell–matrix interaction receptors (i.e. integrins),cell–cell interaction receptors (i.e. cadherins) and cell growth factor re-ceptors as well as other modulators. As a result, the arsenal of cancertherapeuticswould strongly increase based on better-characterized sig-naling pathways related to the surrounding tumor microenvironmentthat may be used as new therapeutic targets. Furthermore, 3D modelscan also help in gaining a deeper understanding of the mechanismsthat confer multidrug resistance (expression of efflux transporters, de-regulation of cellular metabolism involving DNA repair, apoptosis orcell cycle signaling and decrease drug uptake) and, as a consequence,develop drugs capable of engaging, evading or exploiting them [17,18]. In this review, we describe the most representative 3Dbioengineered models for cancer, applying state of the art bioengineer-ing and biomaterial tools.

and deconstructivism as design principles, in order to produce valuablemodels that help ton biomaterials that can mimic as closely as possible the physical, chemical andmechanicalthese variables, tissue engineering focuses on deconstructing the in vivo cellular microen-e environmental signals that rule tumor initiation and progression (deconstructivism).

environments for cancermodels, Adv. DrugDeliv. Rev. (2014), http://

3M. Alemany-Ribes, C.E. Semino / Advanced Drug Delivery Reviews xxx (2014) xxx–xxx

2. Biomaterials: applying biomimetic principles to study cancer

Tumor progression is governed by the biochemical and mechanicalcues that arise from the ECM [19,20]. Specifically, ECM imparts biochem-ical signaling through twomechanisms: (i) the binding of a wide varietyof soluble effector molecules — controlling their diffusion and local con-centration and (ii) the exposure of specific motifs that are recognized bycellular adhesion receptors. As a result, ECM is dynamically integratedwith the intracellular signaling pathways that regulate gene expressionand participate in cell phenotype determination [21–25]. Additionally,cells are able to sense the matrix stiffness through their tension machin-ery (Rho/ROCK signaling pathway), which results in mechanical signal-ing. Cells routinely contract their actomyosin cytoskeleton in order topull on the milieu to which they are attached, generating internal stress.This mechanical stimulus is converted into a chemical response througha process known as mechano-transduction, which is reported toinfluence cell proliferation and differentiation [26–28]. It has beendemonstrated that some tumors are characterized as becoming a pro-gressively stiffer tissue, which is the case for breast cancer tissue thatcan be 10 times stiffer than healthy ones. This phenomenon is producedby an elevated deposition and remodeling of ECM components, mainlyfibrillar collagen and hyaluronic acid, secreted by cancer cells and resi-dent fibroblasts or connective tissue cells of the stroma [22,29].

To attain an accurate recreation of the tumor in vivo microenviron-ment, the scaffoldmaterial of choice to replicate ECMwould need to sat-isfy two key requirements: molecular composition and stiffness. Thefirst successful approaches consisted in using scaffolds of natural origin,with collagen [30–32], Matrigel [33] and hyaluronic acid [34] being thegold standard in cancer research. They provide a broad range of chemi-cal cues, principally ECM binding motifs, which actively participate intumorigenesis. Thesematrices have produced important conceptual ad-vances, since they have achieved the overexpression of tumor genes(epithelial–mesenchymal transition [EMT]markers, matrixmetallopro-teinases [MMPs], pro-angiogenic factors, etc.) and the acquisition ofdrug resistance compared to 2D cultures, mimicking the in vivo cellularresponse [35–39]. On the other hand, the in vivo stiffness values can beeasily recreated by increasing concentration or cross-linking density ofthese biomaterials. Nevertheless, in parallel, they suffer modificationsof fiber architecture, adhesiveness and pore size, altering cell behaviorindependently of differences inmechanical properties [40–42]. Further-more, the composition and stiffness of the tissue can be a critical vari-able depending on the animal origin and the isolation and purificationprocedures, compromising assay reproducibility [3,4,10].

To overcome all these drawbacks, a further step in cancer biology in-volved the development of synthetic biomaterials. They provide areproducible cellular microenvironment and the flexibility to individu-ally tune a mechanical or chemical feature (with the aim of analyzingits specific role on the disease). Paradoxically, these advantages makethis class of biomaterials far more challenging because they do not con-tain signalingmotifs and, therefore, factors that drive tumor progressionhave to be identified and precisely incorporated. Indeed, unless surface-modifications are applied (adhesion of peptides or biological mole-cules), scaffolds only serve to hold and guide cells in 3D spaces untilthey produce their own physiological matrix environment [4,43](Fig. 2).

From the biochemical perspective, synthetic scaffolds can bedesigned to incorporate functional domains of ECM proteins and/orsoluble signaling molecules in a controlled spatiotemporal manner.Thus far, polymeric scaffolds have been extensively used as flexibleplatforms to examinemechanical and structural cues,mainly polyethyl-ene glycol (PEG) [10,44–46], poly(lactide-co-glycolide) (PLG) [47],poly(lactic-co-glycolic acid) (PLGA) [47], and polylactic acid (PLA) [48,49]. In addition, such polymeric scaffolds can be covalently functional-ized with integrin binding sites (arginine–glycine–aspartate, RGD se-quence) or proteolytic degrading sites (metalloproteinase targetsequences). The latter can be conjugated with specific soluble

Please cite this article as:M. Alemany-Ribes, C.E. Semino, Bioengineering 3Ddx.doi.org/10.1016/j.addr.2014.06.004

biomolecules (growth factors, angiogenic factors, cytokines, etc.),which can have subsequently controllable release based on proteinaseactivity [50,51].

To continue advancing in tumor tissue engineering, current effortsare being focused in twomain directions: (1) the systematic explorationof more ECM binding motifs and soluble signaling molecules related totumor progression and (2) the optimization of a new class of biomate-rials – synthetic peptide nanofibers – that serve as scaffolds to enablecancer cell growth in vitro. Synthetic scaffolds can be functionalizedwith the major recognition sequences for cellular adhesion, spreadingand migration. However, the role of many sequences, for example,laminin-binding sequences (YIGSR, IKVAV, PDSGR) and collagenbinding-sequences (PRGDSGYRGDS) [52,53] in tumor progression stillremains unclear. On the other hand, synthetic peptide nanofibersmimic the ECM physical properties more precisely than polymers doand, therefore, emerge as promising platforms to continue studyingthe influence of biochemical and mechanical signals. Indeed, the struc-ture of polymeric scaffolds such as PLA, PLG and PLGA consists of 50–500 μm fibers and 50–100 μm pores in diameter, which is in the sameorder of magnitude as mammalian cell dimensions. As a consequence,cells embedded in such polymers experience a truly 2Dmilieu environ-ment. In addition, biomolecules (growth factors and cytokines) are1000–10,000 times smaller and can diffuse away quickly. Instead,peptide scaffolds are characterized by −10 nm fibers and 5–200 nmpores in diameter (1000 times smaller than cells). They contain biolog-ically inspired sequences, with the most relevant examples being pep-tide amphiphiles [54], β-hairpin peptides [55] and self-assemblingpeptides, such as KLD-12 and RAD16-I (the second one commerciallyavailable under BD™ PuraMatrix™) [52,56,57].

Numerous publications report functionalization of synthetic peptidenanofibers through solid-phase synthesis extension at the C-terminiwith short peptide sequences, aiming to trigger different cell responses[43,58,59]. Therefore, all above-mentioned scaffolds are valuable tocancer research.

From themechanical perspective, the stiffness of synthetic biomate-rials can be precisely tuned by changing concentration or cross-linkingdensity, without introducing an array of confusing cues (i.e. changesin proteolytic sensitivity and cell ligand density). Until now, only afew studies based on varying stiffness of synthetic scaffolds have beenpublished therefore, more research is needed in this field. The reportedcases have used the polymer PEG [45,60] and the self-assembling pep-tide RAD16-I [61]. Interestingly, results have shown that ECM stiffnessper se could initiate tumor progression through modulation of integrindynamics.

Each scaffold thus provides specific biochemical and mechanicalcues, inducing a different cellular response. Hallmark experimentshave largely demonstrated that culturing the same cells in different bio-material results in changes in their morphology, proliferation rate, mi-gratory potential and EMT gene expression [62–64]. For this reason,the selection of the scaffold is a key point when planning the experi-ments and is dependent on multiple factors: (i) the application of the3Dmodel (the comprehension of disease mechanism, the developmentof drug screening processes, the establishment of gene signatures topredict the prognosis of personalized cancer cases, etc.), (ii) the tissueof origin and tumor etiology and (iii) the concrete step of tumor pro-gression to be recreated. The data gathered from 3D culture modelsshould be interpreted in the context of each experimental design.

3. Cancer models: recreating tumor progression step by step

Tumorigenesis is a multistage process characterized by: (i) limitlesscellular proliferation, (ii) sustained angiogenesis and (iii) tissue inva-sion and metastasis. Cells acquire these hallmark capabilities after theaccumulation of consecutive genetic alterations together with theevolving crosstalk with themicroenvironment [65]. Considering its def-inition, tumorigenesis cannot be entirely recreated in vitro as it engages

environments for cancermodels, Adv. DrugDeliv. Rev. (2014), http://

Fig. 2.Deconstructing the in vivo cellularmicroenvironment to comprehend the physical,mechanical and chemical signals that govern tumor progression at themolecular level. Syntheticbiomaterials consist of non-instructive building blocks,which can be tailoredwith bioactive epitopes (i.e. ECM functional peptides or soluble signalingmolecules). Cells aremixedwith theliquid solution of a synthetic scaffold that, subsequently, undergo an assembly process in response to controlled physical or chemical agents (pH, temperature, catalysts, etc.), rendering anordered structure at the micro- or nanomolecular level (hydrogels). As a consequence, cells are embedded within the synthetic biomaterial in a 3D arrangement and initiate self-organization programs. This is promoted by the environmental cues that arise from the extracellular milieu; for instance, they synthesize and deposit their own ECM, together withthe recruitment of morphogens. As a consequence, synthetic scaffolds can mimic specific cellular niches and activate programs, such as EMT in cancer pathophysiology.

4 M. Alemany-Ribes, C.E. Semino / Advanced Drug Delivery Reviews xxx (2014) xxx–xxx

multiple cellular programs. As a consequence, scaffolds are carefully de-signed depending on which step is modeled (Table 1).

3.1. Limitless cellular proliferation

Significantly, 90% of cancers have an epithelial origin. Epithelialtissues show some distinguishing microscopic features, such as cellpolarity, specialized intercellular interactions and attachment to an un-derlying basement membrane [66]. This ordered architecture is neces-sary for the proper control of cellular proliferation and differentiationand is disrupted during the pathogenesis of epithelial tumors [4,19,67]. When cancer cells grow in 2D cultures, they acquire upper (dorsal)and lower (ventral) surfaces and, thus, experience an artificial epithelialpolarity. Three-dimensionality can restore tumormorphology indepen-dently of matrix composition and stiffness [67].

Applying bioengineering tools enable the examination of the influ-ence of microenvironment on polarity disruption and abnormal cellulargrowth. Seminal experiments reveal that integrin blocking antibodiescan revert the malignant phenotype of cancer cells in 3D cultures [13,68]. Furthermore, it is demonstrated that cells undergo minimal or noproliferation when cultured in non-instructive PEG hydrogel comparedto PEG functionalized with integrin binding sites (RGD sequences) orMatrigel [44,45,53]. Therefore, tumor architecture and functionalityare orchestrated and maintained through ECM adhesion receptors. Inparticular, proliferation depends on the activation of the integrin β1family, which in turn phosphorylates the focal adhesion to the kinasefamily (FAK), signaling a pathway [13,67,69–72]. These findings areconsistent with previous work involving animals. For instance, in trans-genicmousemodels formammary or pancreatic beta cell cancer, knock-down of beta1 integrin results in the inhibition of the proliferation ofmammary cancer cells and senescence of pancreatic beta cancer cells[73,74]. Apart from matrix composition, matrix mechanical properties

Please cite this article as:M. Alemany-Ribes, C.E. Semino, Bioengineering 3Ddx.doi.org/10.1016/j.addr.2014.06.004

have a critical role on cellular proliferation, since they can per se initiatetumor progression through themodulation of integrin dynamics. Differ-ent studies show that high stiffness values result in an increase of colla-gen density, which promotes integrin clustering to enhance focaladhesion activation. As a result, ECM is remodeled to break down thehighly dense network that acts as a physical barrier to cellular growth[41,60,61,75,76].

3.2. Sustained angiogenesis

Angiogenesis is the formation of new capillaries from pre-existingones [77]. The process is activated when the tumor demand for oxygenand nutrients surpasses the local supply, typically at a diameter be-tween 1 and 2 mm. Tumor vasculature is required for cell growth anddissemination. Its architecture is characterized by a poor organizationdue to the imbalance between pro- and anti-angiogenic factors andthe stress generated by the uncontrolled proliferation rate that forcesvessels to move apart [78]. As a result, the tumor turns hypoxic with alow pH and a high interstitial fluid pressure, creating a hostile andresistant microenvironment. Intensive research is directed to the devel-opment of anti-angiogenic strategies in order to prevent cancerprogression (i.e., thalidomide, Herceptin, AZD2171, etc.) [79].

Multiple studies demonstrate that culture of cancer cells withintissue engineered scaffolds can neither promote nor support angiogen-esis per se, strongly suggesting that stromal cells are an essential re-quirement in the formation of a capillary-like microvasculature [80].3D cultures constitute an advantageous framework to co-culture cancercells with different cell types. In fact, culturing cancer cells, endothelialcells and fibroblasts within collagen result in the recruitment and differ-entiation of endothelial cells to develop blood vessel-resembling struc-tures. Cell–matrix interactions between endothelial cells and collageninduce the migration of endothelial cells towards the collagen-

environments for cancermodels, Adv. DrugDeliv. Rev. (2014), http://

Table 1The most representative 3D scaffolds used to model the different stages of tumor progression.

Tumor stage 3D cultures

Natural scaffolds Synthetic scaffolds

Limitlessproliferation

They provide a wide range of ECM receptors, basically adhesion and proteolytical-based re-modeling complexes:Matrigel [13,33,58,68], collagen [70,123] and hyaluronic acid [34,64,91].They enable the ability to modulate stiffness by increasing concentration of cross-linkingdensity. In parallel, they suffer changes in pore size, fiber architecture and adhesion sites:collagen [41,76] and hyaluronic acid [124,125].

They are blank environments that systematically explore the role ofECM receptors: PEG versus PEG–RGD sites and PEG–MMP sensitivesites [44,45,53].They can independently tune stiffness, without introducing anarray of confusing ECM signals: PEG [44,45] and RAD16-I [61].

Sustainedangiogenesis

Culturing cancer cellswithin 3Dmatrices is representative of pre-vascularized stages of tumorprogression.The interaction among cancer cells, endothelial cells and fibroblasts within a 3D collagen en-vironment is required for micro-vasculature formation [80,81].Alternatively, arterial explants can be implanted into 3D matrices of different origins(collagen, Matrigel) containing tumor cells [3,85].

They enable the immobilization and subsequent release ofangiogenic factors, in a controlled spatial and temporal manner:PLG [47,83,84].

ECMdegradationandmigration

They provide suitable platforms to perform invasion assays and study the expression of ECMdegrading enzymes: Matrigel [90], collagen [88–90] and hyaluronic acid [34,91].They can recreate micro-tunnels for proteolytically inactive cells and study alternativemechanisms to ECM remodeling: collagen [93].They can determine the stiffness contribution to tumor invasive phenotype, in parallel tophysical modifications of the matrix: collagen [41,76] and hyaluronic acid [124,125].

They are blank environments to systematically explore the role ofECM degrading enzymes: PEG versus PEG–MMP sensitive sites[44,45,53].They enable the investigation of the existence of alternative mi-gration mechanisms to ECM remodeling: PEG [60] and RAD16-I[62].The isolated impact of stiffness on migration can be studied: PEG[60] and RAD16-I [61].

Bloodcirculation

They are used in microfluidic platforms.Microchannels coated with stromal beds of Matrigel [94,95] and collagen [95–97] to evaluatethe impact of hydrodynamic forces on EMT.Endothelialized networks [100–103] within fluidic circuits to characterize adhesion betweencancer and endothelial cells and degradation of basement membrane.

They can enable analyzing what chemical cues participate in thecontrol of cancer cell circulation through the bloodstream.They can determine what adhesion complexes are needed toactivate extravasation.

Metastaticgrowth

The biomaterial used depends on the metastatic growth site.In general, the 3D matrix should contain bioactive motifs through which cells are able toadhere to or enter into a dormancy state [104].

They are blank environments that systematically explore the role ofadhesion sites: PEG versus PEG–RGD sites [44,45,53].

5M. Alemany-Ribes, C.E. Semino / Advanced Drug Delivery Reviews xxx (2014) xxx–xxx

embedded layer of fibroblasts and cancer cells. On the other hand, cell–cell interactions cause the differentiation of endothelial cells into capil-lary-like structures, through the delivery of pro-angiogenic factors in aspatiotemporal controlled manner [80–82]. Synthetic scaffolds likePEG and PLG are used for immobilization and subsequent release ofchemical cues that may be involved in angiogenesis, making such sys-tems powerful platforms to study tumor-dependent changes in angio-genic sprouting [47,83,84]. Finally, smart platforms located on thefrontier between in vitro and in vivo conditions are introduced. An im-portant case is the arterio-venous (AV) loop based on the microsurgicalimplantation of small caliber vessels in matrices of different composi-tions. For instance, arterial explants from umbilical cords are embeddedin Matrigel to study their interaction with cancer cells. These explantsled to capillary-like structures autonomously without stimulation withexogenous growth factors [3,85]. The dynamic observation of cancercells that recruit, interact and stimulate the growth of new vessels canpromote the understanding of tumor-driven angiogenesis.

3.3. Tissue invasion and metastasis

Metastasis is a poorly understood mechanism of tumor spreading,despite being responsible for 90% of all cancer deaths [66]. During met-astatic process, cells gain the capacity to degrade their basement mem-brane and invade surrounding tissue. Subsequently, metastatic cellsenter the lymphatic and circulatory systems in order to disseminateand undergo growth in distant parts of the body. This cascade of eventsis only possible if cells lose their epithelial phenotype, which results indisruption of basal–apical polarity, down-regulation of intercellular ad-hesions and a dramatic remodeling of the action cytoskeleton. As a con-sequence, cells acquire a mesenchymal phenotype that switches onproteolysis and motility programs. This conversion is known as EMTand involves changes in cellular architecture and function [65,86,87].Due to its clinical relevance, a major effort is directed to develop newmodels to capture specific steps of tumormetastasis and, therefore, dis-cover new insights of the molecular mechanism that drive metastasis.

One of the first questions to be addressed is how cells can migratethroughout the ECM. Experiments with natural scaffolds evidence that

Please cite this article as:M. Alemany-Ribes, C.E. Semino, Bioengineering 3Ddx.doi.org/10.1016/j.addr.2014.06.004

cancer cells overexpress ECMdegrading enzymes (MMPs, hyaluronases,etc.), compared to normal cells [34,88–91]. For this reason, syntheticbiomaterials like PEG are used as “neutral” state microenvironment toprecisely tune structural cues, in this case MMP-sensitive motifs [10,44,92]. Results show that cells are able to tunnel through the matrixvia proteolytic degradation executed by MMPs, resulting in its continu-ous remodeling. However, clinical studies demonstrate the incompletetherapeutic window covered byMMP's inhibitors in cancer progression.3D cultures are used to identify other mechanisms responsible for themigratory capacity of cancer cells. Notably, synthetic scaffolds (PEGand RAD16-I) [60,62] and natural scaffolds (collagen) [93] can recreatemicro-tunnels of the same size and topography as those produced bymetastatic cells. These opened paths are used byMMP deficient activitycells to migrate using an amoeboid phenotype, displacing matrix fibersby actomyosin-based mechanical forces. Therefore, bioengineeringtumor microenvironments can lead to the discovery of the synergismbetween proteolytic degradation and amoeboid movement formigration.

In a second stage, the molecular mechanisms that direct the estab-lishment of an endothelial network and the circulation of cancer cellsthrough the bloodstream are investigated using microfluidic platforms.Specifically, in these microfluidic devices, 3D cultures prepared withMatrigel [94,95] or collagen [95–99] scaffolds are subjected to a contin-uousflow (shear stress, interstitialfluid flow). These systems enable theanalysis of fluidic forces as modulators of EMT processes during tumor-igenesis, in a process called intravasation. Furthermore, methodologiesare developed to produce endothelialized networks within 3D scaffoldsin these microfluidic circuits [100–103]. Their main objective is tocharacterize the processes activated by cancer cells under shear stressconditions: adhesionwith endothelial cells and degradation of the base-ment membrane to undergo metastatic growth (extravasation).

Finally, colonization of cancer cells to a secondary metastatic site isevaluated bymimicking the host cellular niche. The location ofmetasta-sis is not random; each type of cancer tends to spread to a particulartissue or organ at a higher rate than expected by statistical chance. Itis postulated that a non-permissivemicroenvironment, in which cancercells are unable to properly adhere, triggers their dormancy [104]. For

environments for cancermodels, Adv. DrugDeliv. Rev. (2014), http://

6 M. Alemany-Ribes, C.E. Semino / Advanced Drug Delivery Reviews xxx (2014) xxx–xxx

this reason, the modeling of the newmicroenvironment is a useful toolfor understanding themechanismsmediated by the ECMand neighbor-ing cells that explain this specificity in metastasis location. For instance,the three most commonly diagnosed cancer types (prostate, lung andbreast) tend to metastasize to the bone. Consequently, research isdirected to the creation of biomimetic organic collagen [105] andinorganic hydroxyapatite (HA) [105,106] that form the bone. Resultsshow that cancer cells mineralize in an active and regulated processsimilar to osteoblasts. Therefore, cells possess osteomimetic capabilitiesthrough the expression of bone marker proteins that allow them toadapt and flourish within the bone microenvironment (Fig. 3).

Fig. 3. Design parameters for 3D scaffolds depend on the tumor stage and the corresponding cecharacterized by different stages with particular genetic and epigenetic alteration hallmarks: (1characterized by ECM degradation and migration and epithelial–mesenchymal transition (EMmesenchymal–epithelial transition (MET) and (6) sustained angiogenesis. Tissue invasion andvation of EMT, the intravasation into the lymph and blood vessels to allow their passive translymph and blood vessels to colonize a secondary site. Importantly, the angiogenesis process ismost relevant biomaterials (from both natural and synthetic origins) used to date to model ea

Please cite this article as:M. Alemany-Ribes, C.E. Semino, Bioengineering 3Ddx.doi.org/10.1016/j.addr.2014.06.004

4. Cells: maintaining tumor identity

Apart from the continuous research in the field of 3D scaffolds, amajor challenge in cancer biology is the optimization of the cell sourcein order to provide an accurate tissue engineering toolbox for identifica-tion and characterization of new therapeutic approaches. Commonly,in vitro models have been based on immortal tumor-derived cell lines.They constitute the most common means of studying cancer patho-physiology due to their accessibility, ease of culture and homogeneity.However, genomic studies demonstrate that cell lines only reflect alimited part of the gene expression profile that characterize the tumor

llular programs that want to be precisely recreated. Tumorigenesis is a multistage process) normal tissue; (2) primary tumorwith limitless cellular proliferation; (3) invasive tumorT); (4) blood circulation; (5) intravasation, extravasation and metastasis process throughmetastasis comprise the ECM degradation, the migration of tumor cells through the acti-port to distant organs, and the extravasation to degrade the basement membrane of theperformed within primary, invasive and metastatic tumor sites. The figure illustrates thech tumor stage.

environments for cancermodels, Adv. DrugDeliv. Rev. (2014), http://

7M. Alemany-Ribes, C.E. Semino / Advanced Drug Delivery Reviews xxx (2014) xxx–xxx

in vivo, emerging as unreliable predictive preclinical models [14,107].There are multiple reasons that explain this misrepresentation. Celllines have been in culture for several years or decades in 2D conditions,which have imposed a strong selective pressure on them, makingadaptation only possible by changes at the DNA level. Moreover, asthey are distributed in many laboratories, the same cell line mighthave undergone various selection steps due to differences in feedingand passage techniques [107]. Secondly, the stroma is an active partici-pant in tumor progression, a framework that includes the ECM as wellas cellular components such as fibroblasts, blood-vessel cells (endothe-lial, pericytes and smooth muscle cells) and immune cells (lympho-cytes, macrophages and mast cells) [108,109]. For this reason,enrichment ofmalignant cells and lack of stroma in cell lines significant-ly change gene expression profiles compared to in vivo situations.Finally, analyses reveal a high degree of genomic heterogeneity acrossthe human cancer patient population and the crucial role that it has inthe variable clinical-response to treatment. A unique cell line does notcapture this complexity [14].

The scientific community currently prefers to work with morereliable cellular approaches, including mainly clinical biopsies or theparallel analysis of large panels of cell lines. Biopsies are obtained atthe time of surgical resection, when they are freshly harvested andcultured in 2D or 3D conditions. The complex mixture of cells that arepart of the tumor milieu is preserved and, therefore, recreate thecrosstalk between stromal and cancer cells more precisely. This newmethodology is the first step to personalized medicine, enabling thescreening on a patient-by-patient basis for drug sensitivity and resis-tance and, therefore, the selection of the best suitable treatment. How-ever, this platform is not oriented to be representative of a class oftumor, since they only incorporate the genetic information of an indi-vidual patient [110,111]. The second strategy is focused on the estab-lishment of cellular platforms in which each type of tumor isrepresented by a large number of cell lines. The most popular are theNational Cancer Institute 60 (NCI60) and the Center for Molecular Ther-apeutics 1000 (CMT1000) platforms. These centers have 6 and 51 celllines, respectively, for studying breast cancer, for example. This cell col-lection captures the genomic diversity of human cancer with the mainpurpose of correlating underlying genotypes with drug response. Theobtained data permits to create molecular signatures that are clinicallyuseful for both predicting drug sensitivity and elucidating mechanismsof drug action. Themajor limitation is the logistical challenge associatedwith the culture of large panels of cancer cell lines, considerably limitingthe throughput of the platform with respect to the number of com-pounds that can be realistically tested in a given period of time [14].

Fig. 4. Rethinking drug screening processes. Tissue engineering provides 3D cultures that recreato the incorporation ofmultiple physical,mechanical and chemical cues that arise from ECM–cenot capture important facets of human behavior and they are not feasible for HTS applications

Please cite this article as:M. Alemany-Ribes, C.E. Semino, Bioengineering 3Ddx.doi.org/10.1016/j.addr.2014.06.004

5. Unresolved issue: high throughput screening

3D cultures are an indispensable tool in the advancement of basiccancer research. The next step consists of incorporating them as predic-tive models in massive drug screening processes, shifting fromacademia to the pharmaceutical industry. 3D cultures can both contrib-ute critically to the reduction in animal testing and allow for economicalsavings as well as become a powerful model to optimize drug candi-dates for enhanced tissue distribution and efficacy (Fig. 4). To achievethis goal, 3D cultures should fulfill a key requirement: high throughputapplicability. Major efforts are being directed to (i) the development ofreproducible biomaterials capable of accurately recreating the in vivotumor microenvironment, (ii) the establishment of simple and stan-dardized culture protocols and (iii) the attachment of automated cellu-lar level imaging and analysis techniques.

According to literature data, it is expected that the vast majority ofdrug candidates lose efficacy in a 3D environment, compared to 2D.These results are explained due to the existence of molecular gradientsthat challenge drug delivery and the acquisition of a major cellularmalignancy as confirmed by the overexpression of tumor genes[112–114]. There are, however, some cases in which the drug is moreeffective in 3D because the molecular target is expressed only or inparticular in these conditions [36].

Nowadays, the most popular current methodology to obtain high-throughput in 3D cultures is cell patterning. It is based on producingminiature culture arrays in the tissue-engineered scaffold with a 96-or 384-well plate format using techniques such as soft lithography. Inthis way, a single cell or suspension is microinjected in each pattern ofthe scaffold, being localized in spatially controlled coordinates. Suspen-sion formation can occur for a broad range of cells, independent of theirability to form spontaneous cell–cell contacts due to the presence of anextracellular milieu. Moreover, overlapping phenomena between sus-pensions is avoided and, thereby, it is possible to perform individualanalysis. Importantly, to avoid batch-to-batch variability in scaffoldproperties, natural biomaterials are submitted to chemical cross-linking stabilization or synthetic biomaterials are used [15,115]. Untilnow, relevant examples comprise micro-patterned collagen [111] andMatrigel [116,117]. These platforms provide a rapid readout for theaverage response of all cells within the 3D scaffold. However, they donot generate information about cells in distinct regions of the construct,which are physiologically and metabolically different, because they areexposed to gradients of oxygen, nutrients and signaling molecules. Anew technique is being developed in which a continuous 3D tissue isproduced by stacking different layers of 200 μm-thick paper, each one

te the complex cellularmicroenvironmentmore precisely than traditional 2D cultures, duell and cell–cell interactions. At the opposite end of the experimental continuum, animals do. Therefore, 3D cultures can bridge the gap between 2D cultures and animal models.

environments for cancermodels, Adv. DrugDeliv. Rev. (2014), http://

8 M. Alemany-Ribes, C.E. Semino / Advanced Drug Delivery Reviews xxx (2014) xxx–xxx

supporting 96 individual hydrophilic zones, impregnated with cell-containing Matrigel. These layers can easily be destacked to performanalysis of cells in specific regions of the 3D construct [118].

As biomaterials and culture procedures are being optimized forhigh-throughput screening applications (HTS), automated and high-resolution analysis techniques have to be adapted from 2D to 3D cul-tures. These tests are based on monitoring suspension cellular growthkinetics and migration. Particularly, they are driven by automaticallytracking labeling cells through high-resolution of phase-contrast orfluorescence microscopy. The processing steps involve performingcorrelation of the cell template with each observed volume (templatematching step) and quantitatively reconstructing cell trajectories orvolumes from recorded image sequences under computer-assistedmicroscopy (mean-shift process step) [119]. These techniques canonly be applied in the case of transparent matrix gels. More advancedanalysis approaches are being developed. For instance, simple tests aredesigned to monitor EMT phenomena and thus identify antimetastaticagents for the treatment of cancer. In particular, they incorporate a lucif-erase reporter in the genetic sequence of prognostic EMT biomarkers, asvimetin. Therefore, its expression is a direct readout of the mesenchy-mal phenotype of cancer cells [120].

6. Future directions

It is our hope that this review serves as inspiration for researchersworking in cancer biology field in order to re-define or improve theircell culture systems. It is also our intention to encourage them to usemore predictable human models that could reduce animal experimen-tation, which has been shown to be informative but not necessarilyrepresentative of the current human need in therapeutic practice. In ad-dition, the bottleneck in cancer drug screening is based on 1) the use ofnon-well representative cancermodels that provide unreliable data and2) the use of predictable ones but with the additional limitation in theircapacity to be scaled up for HTS. This is the case for instance when wecompared traditional 2D platforms with high-tech microfluidic systemdevices. The first, clearly are simple and feasible for HTS but with lowrepresentation of the real situation and, as a consequence, low efficacy.The reason is based on the fact that 2D cultures fail to integrate themul-tiple signals that arise from the extracellular milieu. The second onecould represent the tumor scenario but would be difficult to adapt to a96–384-multiwell plate due to its highly sophisticated platform.Furthermore, an additional challenge is that animals, in general, donot capture important human facets for drug cancer screening. Thepossible solution we emphasize in this review is to develop 3D culturesystems that can be precisely tailored depending on the origin of thetumor tissue and the concrete stage to model (limitless proliferation,sustained angiogenesis or metastasis). Moreover, the 3D platformshould accommodate automation capacity with robotics (drug screen-ing in dose–effect prediction, easy culture maintenance, good signal-to-background readout and efficacy).

Efforts should continue in the optimization of material and cellsource. The field is advancing exponentially in terms of technologicalplatforms, which indicates that soon most of the desired 3D cancermodels will be up and running in academic and pharmaceutical indus-trial settings. The goal for the future will consist of obtaining highly ef-fective models to identify and predict pre-cancer cells (early stages oftumorigenesis) as a way to have early diagnosis, since disease preven-tion is infinitivelymore effective than its treatment. Secondly, therapeu-tic strategies can be re-thought. Nanoparticles are gaining importanceas efficient delivery systems for anticancer drugs and their structurecan be changed by using a stimulus property of tumor microenviron-ment, such as lowpH, lowpartial oxygenpressure or high concentrationof proteases. Designing efforts are also focused in nanoparticles thatmodulate the mechanical or chemical properties of tumor microenvi-ronment in order to normalize it. For instance, the ever-changing

Please cite this article as:M. Alemany-Ribes, C.E. Semino, Bioengineering 3Ddx.doi.org/10.1016/j.addr.2014.06.004

vascular system, immune response and ECM structure can be therapeu-tic targets [121,122].

Acknowledgments

The authors want to thank Jessica Reik for her help in writing thismanuscript and Mercedes Balcells-Camps for her critical revision andhelpful comments. This research was supported by the IQS-School ofEngineering, Bioengineering Department Budget to C.E.S (Grant num-ber 982). M.A.-R. Thanks to the Comissionat per a Universitats i Recercadel Departament d'Innovació, Universitats i Empresa de la Generalitatde Catalunya i del Fons Social Europeu for a predoctoral fellowship.

References

[1] F. Pampaloni, E.G. Reynaud, E.H.K. Stelzer, The third dimension bridges the gap be-tween cell culture and live tissue, Nat. Rev.Mol. Cell Biol. 8 (2007) 839–845, http://dx.doi.org/10.1038/nrm2236.

[2] K.S.M. Smalley, M. Lioni, M. Herlyn, Life isn't flat: taking cancer biology to the nextdimension, In Vitro Cell. Dev. Biol. Anim. 42 (2006) 242–247.

[3] D.W. Hutmacher, R.E. Horch, D. Loessner, S. Rizzi, S. Sieh, J.C. Reichert, et al., Trans-lating tissue engineering technology platforms into cancer research, J. Cell. Mol.Med. 13 (2009) 1417–1427, http://dx.doi.org/10.1111/j.1582-4934.2009.00853.x.

[4] L.G. Griffith, M.A. Swartz, Capturing complex 3D tissue physiology in vitro, Nat.Rev. Mol. Cell Biol. 7 (2006) 211–224, http://dx.doi.org/10.1038/nrm1858.

[5] C. Castells-Sala, M. Alemany-Ribes, T. Fernández-Muiños, L. Recha-Sancho, P.López-Chicón, C. Aloy-Reverté, et al., Current applications of tissue engineeringin biomedicine, J. Biochip Tissue Chip. S2 (2013) 1–14, http://dx.doi.org/10.4172/2153-0777.S2-004.

[6] L.A. Gurski, N.J. Petrelli, X. Jia, M.C. Farach-Carson, 3Dmatrices for anti-cancer drugtesting and development, Oncol. Issues (2010) 20–25.

[7] E. Burdett, F.K. Kasper, A.G. Mikos, J.A. Ludwig, Engineering tumors: a tissue engi-neering perspective in cancer biology, Tissue Eng. B Rev. 16 (2010) 351–359.

[8] R.M. Sutherland, J.A. McCredie, W.R. Inch, Growth of multicell spheroids in tissueculture as a model of nodular carcinomas, J. Natl. Cancer Inst. 46 (1971) 113–120.

[9] R.M. Sutherland, B. Sordat, J. Bamat, H. Gabbert, B. Bourrat, Oxygenation and differ-entiation in multicellular spheroids of human colon carcinoma, Cancer Res. 46(1986) 5320–5329.

[10] D. Loessner, K.S. Stok, M.P. Lutolf, D.W. Hutmacher, J.A. Clements, S.C. Rizzi,Bioengineered 3D platform to explore cell–ECM interactions and drug resistanceof epithelial ovarian cancer cells, Biomaterials 31 (2010) 8494–8506, http://dx.doi.org/10.1016/j.biomaterials.2010.07.064.

[11] M.J. Bissell, D. Radisky, Putting tumours in context, Nat. Rev. Cancer 1 (2001)46–54, http://dx.doi.org/10.1038/35094059.

[12] O.W. Petersen, L. Rønnov-Jessen, A.R. Howlett, M.J. Bissell, Interaction with base-ment membrane serves to rapidly distinguish growth and differentiation patternof normal and malignant human breast epithelial cells, Proc. Natl. Acad. Sci. U. S. A.89 (1992) 9064–9068.

[13] V.M. Weaver, O.W. Petersen, F. Wang, C.A. Larabell, P. Briand, C. Damsky, et al., Re-version of the malignant phenotype of human breast cells in three-dimensionalculture and in vivo by integrin blocking antibodies, J. Cell Biol. 137 (1997)231–245.

[14] S.V. Sharma, D.A. Haber, J. Settleman, Cell line-based platforms to evaluate thetherapeutic efficacy of candidate anticancer agents, Nat. Rev. Cancer 10 (2010)241–253, http://dx.doi.org/10.1038/nrc2820.

[15] S. Breslin, L. O'Driscoll, Three-dimensional cell culture: themissing link in drug dis-covery, Drug Discov. Today (2012), http://dx.doi.org/10.1016/j.drudis.2012.10.003.

[16] www.cancer.gov, (n.d.).[17] G. Szakács, J.K. Paterson, J.A. Ludwig, C. Booth-Genthe, M.M. Gottesman, Targeting

multidrug resistance in cancer, Nat. Rev. Drug Discov. 5 (2006) 219–234, http://dx.doi.org/10.1038/nrd1984.

[18] K. Moitra, H. Lou, M. Dean, Multidrug efflux pumps and cancer stem cells: insightsinto multidrug resistance and therapeutic development, Clin. Pharmacol. Ther. 89(2011) 491–502, http://dx.doi.org/10.1038/clpt.2011.14.

[19] K.M. Yamada, E. Cukierman, Modeling tissue morphogenesis and cancer in 3D, Cell130 (2007) 601–610, http://dx.doi.org/10.1016/j.cell.2007.08.006.

[20] D.W. Hutmacher, Biomaterials offer cancer research the third dimension, Nat.Mater. 9 (2010) 90–93, http://dx.doi.org/10.1038/nmat2619.

[21] C. Frantz, K.M. Stewart, V.M.Weaver, The extracellularmatrix at a glance, J. Cell Sci.123 (2010) 4195–4200, http://dx.doi.org/10.1242/jcs.023820.

[22] P. Lu, V.M. Weaver, Z. Werb, The extracellular matrix: a dynamic niche in cancerprogression, J. Cell Biol. 196 (2012) 395–406, http://dx.doi.org/10.1083/jcb.201102147.

[23] R.L. Juliano, S. Haskill, Signal transduction from the extracellular matrix, J. Cell Biol.120 (1993) 577–585.

[24] S.H. Jang, M.G. Wientjes, D. Lu, J.L.S. Au, Drug delivery and transport to solid tu-mors, Pharm. Res. 20 (2003) 1337–1350.

[25] S.-H. Kim, J. Turnbull, S. Guimond, Extracellular matrix and cell signalling: the dy-namic cooperation of integrin, proteoglycan and growth factor receptor, J.Endocrinol. 209 (2011) 139–151, http://dx.doi.org/10.1530/JOE-10-0377.

environments for cancermodels, Adv. DrugDeliv. Rev. (2014), http://

9M. Alemany-Ribes, C.E. Semino / Advanced Drug Delivery Reviews xxx (2014) xxx–xxx

[26] M.A. Wozniak, C.S. Chen, Mechanotransduction in development: a growing role forcontractility, Nat. Rev. Mol. Cell Biol. 10 (2009) 34–43, http://dx.doi.org/10.1038/nrm2592.

[27] M.S. Samuel, J.I. Lopez, E.J. McGhee, D.R. Croft, D. Strachan, P. Timpson, et al.,Actomyosin-mediated cellular tension drives increased tissue stiffness and β-catenin activation to induce epidermal hyperplasia and tumor growth, CancerCell 19 (2011) 776–791, http://dx.doi.org/10.1016/j.ccr.2011.05.008.

[28] P.A. Janmey, R.T. Miller, Mechanisms of mechanical signaling in development anddisease, J. Cell Sci. 124 (2011) 9–18, http://dx.doi.org/10.1242/jcs.071001.

[29] M.R. Ng, J.S. Brugge, A stiff blow from the stroma: collagen crosslinking drivestumor progression, Cancer Cell 16 (2009) 455–457, http://dx.doi.org/10.1016/j.ccr.2009.11.013.

[30] L. Chen, Z. Xiao, Y. Meng, Y. Zhao, J. Han, G. Su, et al., The enhancement of cancerstem cell properties of MCF-7 cells in 3D collagen scaffolds for modeling of cancerand anti-cancer drugs, Biomaterials 33 (2012) 1437–1444, http://dx.doi.org/10.1016/j.biomaterials.2011.10.056.

[31] C.S. Szot, C.F. Buchanan, P. Gatenholm, M.N. Rylander, J.W. Freeman, Investigationof cancer cell behavior on nanofibrous scaffolds, Mater. Sci. Eng. C 31 (2011)37–42, http://dx.doi.org/10.1016/j.msec.2009.12.005.

[32] C.S. Szot, C.F. Buchanan, J.W. Freeman,M.N. Rylander, 3D in vitro bioengineered tu-mors based on collagen I hydrogels, Biomaterials 32 (2011) 7905–7912, http://dx.doi.org/10.1016/j.biomaterials.2011.07.001.

[33] O. Zschenker, T. Streichert, S. Hehlgans, N. Cordes, Genome-wide gene expressionanalysis in cancer cells reveals 3D growth to affect ECM and processes associatedwith cell adhesion but not DNA repair, PLoS ONE 7 (2012) e34279, http://dx.doi.org/10.1371/journal.pone.0034279.

[34] L. David, V. Dulong, D. Le Cerf, C. Chauzy, V. Norris, B. Delpech, et al., Reticulatedhyaluronan hydrogels: a model for examining cancer cell invasion in 3D, MatrixBiol. 23 (2004) 183–193, http://dx.doi.org/10.1016/j.matbio.2004.05.005.

[35] K. Chitcholtan, E. Asselin, S. Parent, P.H. Sykes, J.J. Evans, Differences in growthproperties of endometrial cancer in three dimensional (3D) culture and 2D cellmonolayer, Exp. Cell Res. 319 (2013) 75–87, http://dx.doi.org/10.1016/j.yexcr.2012.09.012.

[36] L.A. Gurski, A.K. Jha, C. Zhang, X. Jia, M.C. Farach-Carson, Hyaluronic acid-basedhydrogels as 3D matrices for in vitro evaluation of chemotherapeutic drugs usingpoorly adherent prostate cancer cells, Biomaterials 30 (2009) 6076–6085, http://dx.doi.org/10.1016/j.biomaterials.2009.07.054.

[37] S. Dangi-Garimella, M.J. Strouch, P.J. Grippo, D.J. Bentrem, H.G. Munshi, Collagenregulation of let-7 in pancreatic cancer involves TGF-β1-mediated membranetype 1-matrix metalloproteinase expression, Oncogene 30 (2011) 1002–1008,http://dx.doi.org/10.1038/onc.2010.485.

[38] S. Dangi-Garimella, S.B. Krantz, M.R. Barron, M.A. Shields, M.J. Heiferman, P.J.Grippo, et al., Three-dimensional collagen I promotes gemcitabine resistance inpancreatic cancer through MT1-MMP-mediated expression of HMGA2, CancerRes. 71 (2011) 1019–1028, http://dx.doi.org/10.1158/0008-5472.CAN-10-1855.

[39] B. Fallica, J.S. Maffei, S. Villa, G. Makin, M. Zaman, Alteration of cellular behavior andresponse to PI3K pathway inhibition by culture in 3D collagen gels, PLoS ONE 7(2012) e48024, http://dx.doi.org/10.1371/journal.pone.0048024.

[40] P.P. Provenzano, D.R. Inman, K.W. Eliceiri, J.G. Knittel, L. Yan, C.T. Rueden, et al., Col-lagen density promotes mammary tumor initiation and progression, BMC Med. 6(2008) 11, http://dx.doi.org/10.1186/1741-7015-6-11.

[41] P.P. Provenzano, D.R. Inman, K.W. Eliceiri, P.J. Keely, Matrix density-inducedmechanoregulation of breast cell phenotype, signaling and gene expressionthrough a FAK–ERK linkage, Oncogene 28 (2009) 4326–4343, http://dx.doi.org/10.1038/onc.2009.299.

[42] K.R. Levental, H. Yu, L. Kass, J.N. Lakins, M. Egeblad, J.T. Erler, et al., Matrixcrosslinking forces tumor progression by enhancing integrin signaling, Cell 139(2009) 891–906, http://dx.doi.org/10.1016/j.cell.2009.10.027.

[43] E. Genové, C. Shen, S. Zhang, C.E. Semino, The effect of functionalized self-assembling peptide scaffolds on human aortic endothelial cell function, Biomate-rials 26 (2005) 3341–3351, http://dx.doi.org/10.1016/j.biomaterials.2004.08.012.

[44] B.J. Gill, D.L. Gibbons, L.C. Roudsari, J.E. Saik, Z.H. Rizvi, J.D. Roybal, et al., A syntheticmatrix with independently tunable biochemistry and mechanical properties tostudy epithelial morphogenesis and EMT in a lung adenocarcinoma model, CancerRes. 72 (2012) 6013–6023, http://dx.doi.org/10.1158/0008-5472.CAN-12-0895.

[45] S. Sieh, A.V. Taubenberger, S.C. Rizzi, M. Sadowski, M.L. Lehman, A. Rockstroh, et al.,Phenotypic characterization of prostate cancer LNCaP cells cultured within abioengineered microenvironment, PLoS ONE 7 (2012) e40217, http://dx.doi.org/10.1371/journal.pone.0040217.

[46] D. Loessner, S.C. Rizzi, K.S. Stok, T. Fuehrmann, B. Hollier, V. Magdolen, et al., Abioengineered 3D ovarian cancer model for the assessment of peptidase-mediated enhancement of spheroid growth and intraperitoneal spread, Biomate-rials 34 (2013) 7389–7400, http://dx.doi.org/10.1016/j.biomaterials.2013.06.009.

[47] C. Fischbach, R. Chen, T. Matsumoto, T. Schmelzle, J.S. Brugge, P.J. Polverini, et al.,Engineering tumors with 3D scaffolds, Nat. Methods 4 (2007) 6–11, http://dx.doi.org/10.1038/NMETH1085.

[48] J.L. Horning, S.K. Sahoo, S. Vijayaraghavalu, S. Dimitrijevic, J.K. Vasir, T.K. Jain, et al.,3-D tumor model for in vitro evaluation of anticancer drugs, Mol. Pharm. 5 (2008)849–862.

[49] S.K. Sahoo, A.K. Panda, V. Labhasetwar, Characterization of porous PLGA/PLA mi-croparticles as a scaffold for three dimensional growth of breast cancer cells,Biomacromolecules 6 (2005) 1132–1139.

[50] S.C. Rizzi, M. Ehrbar, S. Halstenberg, G.P. Raeber, H.G. Schmoekel, H. Hagenmüller,et al., Recombinant protein-co-PEGnetworks as cell-adhesive and proteolytically de-gradable hydrogelmatrixes. Part II: biofunctional characteristics, Biomacromolecules7 (2006) 3019–3029, http://dx.doi.org/10.1021/bm060504a.

Please cite this article as:M. Alemany-Ribes, C.E. Semino, Bioengineering 3Ddx.doi.org/10.1016/j.addr.2014.06.004

[51] M.P. Lutolf, J.L. Lauer-Fields, H.G. Schmoekel, A.T. Metters, F.E. Weber, G.B. Fields,et al., Synthetic matrix metalloproteinase-sensitive hydrogels for the conductionof tissue regeneration: engineering cell-invasion characteristics, Proc. Natl. Acad.Sci. U. S. A. 100 (2003) 5413–5418, http://dx.doi.org/10.1073/pnas.0737381100.

[52] F. Gelain, D. Bottai, A. Vescovi, S. Zhang, Designer self-assembling peptide nanofi-ber scaffolds for adult mouse neural stem cell 3-dimensional cultures, PLoS ONE1 (2006) e119, http://dx.doi.org/10.1371/journal.pone.0000119.

[53] M.S. Weiss, B.P. Bernabé, A. Shikanov, D.A. Bluver, M.D. Mui, S. Shin, et al., The im-pact of adhesion peptides within hydrogels on the phenotype and signaling of nor-mal and cancerous mammary epithelial cells, Biomaterials 33 (2012) 3548–3559,http://dx.doi.org/10.1016/j.biomaterials.2012.01.055.

[54] J.D. Hartgerink, E. Beniash, S.I. Stupp, Self-assembly and mineralization of peptide–amphiphile nanofibers, Science 294 (2001) 1684–1688, http://dx.doi.org/10.1126/science.1063187.

[55] J.P. Schneider, D.J. Pochan, B. Ozbas, K. Rajagopal, L. Pakstis, J. Kretsinger, Respon-sive hydrogels from the intramolecular folding and self-assembly of a designedpeptide, J. Am. Chem. Soc. 124 (2002) 15030–15037.

[56] C.E. Semino, Can we build artificial stem cell compartments? J. Biomed. Biotechnol.2003 (2003) 164–169, http://dx.doi.org/10.1155/S1110724303208019.

[57] S. Zhang, M. Altman, Peptide self-assembly in functional polymer science and en-gineering, React. Funct. Polym. 41 (1999) 91–102, http://dx.doi.org/10.1016/S1381-5148(99)00031-0.

[58] S. Zhang, T.C. Holmes, C.M. DiPersio, R.O. Hynes, X. Su, A rich, self-complementaryoligopeptide matrices support mammalian cell attachment, Biomaterials 16(1995) 1385–1393.

[59] E. Genové, S. Schmitmeier, A. Sala, S. Borrós, A. Bader, L.G. Griffith, et al., Function-alized self-assembling peptide hydrogel enhance maintenance of hepatocyte activ-ity in vitro, J. Cell. Mol. Med. 13 (2009) 3387–3397, http://dx.doi.org/10.1111/j.1582-4934.2009.00970.x.

[60] P. Soman, J.A. Kelber, J.W. Lee, T.N. Wright, K.S. Vecchio, R.L. Klemke, et al., Cancercell migration within 3D layer-by-layer microfabricated photocrosslinked PEGscaffolds with tunable stiffness, Biomaterials 33 (2012) 7064–7070, http://dx.doi.org/10.1016/j.biomaterials.2012.06.012.

[61] Y. Miroshnikova, D. Jorgens, L. Spirio, M. Auer, A. Sarang-Sieminski, V.M. Weaver,Engineering strategies to recapitulate epithelial morphogenesis within syntheticthree-dimensional extracellular matrix with tunable mechanical properties, Phys.Biol. 8 (2011) 026013.

[62] K. Mi, G. Wang, Z. Liu, Z. Feng, B. Huang, X. Zhao, Influence of a self-assemblingpeptide, RADA16, compared with collagen I and Matrigel on the malignant pheno-type of human breast-cancer cells in 3D cultures and in vivo, Macromol. Biosci. 9(2009) 437–443, http://dx.doi.org/10.1002/mabi.200800262.

[63] S. Krause, M.V. Maffini, A.M. Soto, C. Sonnenschein, The microenvironment deter-mines the breast cancer cells' phenotype: organization ofMCF7 cells in 3D cultures,BMC Cancer 10 (2010) 263–276, http://dx.doi.org/10.1186/1471-2407-10-263.

[64] S. Pedron, E. Becka, B.A.C. Harley, Regulation of glioma cell phenotype in 3D matri-ces by hyaluronic acid, Biomaterials 34 (2013) 7408–7417, http://dx.doi.org/10.1016/j.biomaterials.2013.06.024.

[65] D. Hanahan, R.A. Weinberg, Hallmarks of cancer: the next generation, Cell 144(2011) 646–674, http://dx.doi.org/10.1016/j.cell.2011.02.013.

[66] G. Christofori, New signals from the invasive front, Nature 441 (2006) 444–450,http://dx.doi.org/10.1038/nature04872.

[67] J. Debnath, J.S. Brugge, Modelling glandular epithelial cancers in three-dimensionalcultures, Nat. Rev. Cancer 5 (2005) 675–688, http://dx.doi.org/10.1038/nrc1695.

[68] B. Weigelt, A.T. Lo, C.C. Park, J.W. Gray, M.J. Bissell, HER2 signaling pathway activa-tion and response of breast cancer cells to HER2-targeting agents is dependentstrongly on the 3D microenvironment, Breast Cancer Res. Treat. 122 (2010)35–43, http://dx.doi.org/10.1007/s10549-009-0502-2.

[69] T. Shibue, R.A. Weinberg, Integrin β1-focal adhesion kinase signaling directs theproliferation of metastatic cancer cells disseminated in the lungs, Proc. Natl.Acad. Sci. U. S. A. 106 (2009) 10290–10295, http://dx.doi.org/10.1073/pnas.0908506106.

[70] E. Dhimolea, M.V. Maffini, A.M. Soto, C. Sonnenschein, The role of collagen reorga-nization on mammary epithelial morphogenesis in a 3D culture model, Biomate-rials 31 (2010) 3622–3630, http://dx.doi.org/10.1016/j.biomaterials.2010.01.077.

[71] C.C. Park, H. Zhang, M. Pallavicini, J.W. Gray, F. Baehner, C.J. Park, et al., β1 integrininhibitory antibody induces apoptosis of breast cancer cells, inhibits growth, anddistinguishes malignant from normal phenotype in three dimensional culturesand in vivo, Cancer Res. 66 (2006) 1526–1535, http://dx.doi.org/10.1158/0008-5472.CAN-05-3071.

[72] C.C. Park, H.J. Zhang, E.S. Yao, C.J. Park, M.J. Bissell, β1 integrin inhibition dramati-cally enhances radiotherapy efficacy in human breast cancer xenografts, CancerRes. 68 (2008) 4398–4405, http://dx.doi.org/10.1158/0008-5472.CAN-07-6390.

[73] D.E. White, N.A. Kurpios, D. Zuo, J.A. Hassell, S. Blaess, U. Mueller, et al., Targeteddisruption of β1-integrin in a transgenic mouse model of human breast cancer re-veals an essential role in mammary tumor induction, Cancer Cell 6 (2004)159–170.

[74] A. Kren, V. Baeriswyl, F. Lehembre, C. Wunderlin, K. Strittmatter, H. Antoniadis,et al., Increased tumor cell dissemination and cellular senescence in the absenceof β1-integrin function, EMBO J. 26 (2007) 2832–2842, http://dx.doi.org/10.1038/sj.emboj.7601738.

[75] K. Bott, Z. Upton, K. Schrobback, M. Ehrbar, J.A. Hubbell, M.P. Lutolf, et al., The effectof matrix characteristics on fibroblast proliferation in 3D gels, Biomaterials 31(2010) 8454–8464, http://dx.doi.org/10.1016/j.biomaterials.2010.07.046.

[76] M.J. Paszek, N. Zahir, K.R. Johnson, J.N. Lakins, G.I. Rozenberg, A. Gefen, et al., Ten-sional homeostasis and the malignant phenotype, Cancer Cell 8 (2005) 241–254,http://dx.doi.org/10.1016/j.ccr.2005.08.010.

environments for cancermodels, Adv. DrugDeliv. Rev. (2014), http://

10 M. Alemany-Ribes, C.E. Semino / Advanced Drug Delivery Reviews xxx (2014) xxx–xxx

[77] J. Folkman, Tumor angiogenesis, Cancer Med. (2000) 132–152.[78] A.I. Minchinton, I.F. Tannock, Drug penetration in solid tumours, Nat. Rev. Cancer 6

(2006) 583–592, http://dx.doi.org/10.1038/nrc1893.[79] D. Fukumura, R.K. Jain, Tumormicrovasculature andmicroenvironment: targets for

anti-angiogenesis and normalization, Microvasc. Res. 74 (2007) 72–84, http://dx.doi.org/10.1016/j.mvr.2007.05.003.

[80] L.C. Kimlin, G. Casagrande, V.M. Virador, In vitro three-dimensional (3D) models incancer research: an update, Mol. Carcinog. 52 (2013) 167–182, http://dx.doi.org/10.1002/mc.21844.

[81] O. Velazquez, R. Snyder, Z. Liu, R. Fairman, M. Herlyn, Fibroblast-dependent differ-entiation of human microvascular endothelial cells into capillary-like 3-dimensional networks, FASEB J. 16 (2002) 1316–1318.

[82] J.M. Chan, I.K. Zervantonakis, T. Rimchala, W.J. Polacheck, J. Whisler, R.D. Kamm,Engineering of in vitro 3D capillary beds by self-directed angiogenic sprouting,PLoS ONE 7 (2012) e50582, http://dx.doi.org/10.1371/journal.pone.0050582.

[83] S.S. Verbridge, E.M. Chandler, C. Fischbach, Tissue-engineered three-dimensionaltumor models to study tumor angiogenesis, Tissue Eng. A 16 (2010) 2147–2152,http://dx.doi.org/10.1089/ten.TEA.2009.0668.

[84] R.R. Chen, E. a Silva, W.W. Yuen, A. a Brock, C. Fischbach, A.S. Lin, et al., Integratedapproach to designing growth factor delivery systems, FASEB J. 21 (2007)3896–3903, http://dx.doi.org/10.1096/fj.06-7873com.

[85] G. Seano, G. Chiaverina, P.A. Gagliardi, L. di Blasio, R. Sessa, F. Bussolino, et al.,Modeling human tumor angiogenesis in a three-dimensional culture system,Blood 121 (2013) e129–e137, http://dx.doi.org/10.1182/blood-2012-08-452292.

[86] J.P. Thiery, J.P. Sleeman, Complex networks orchestrate epithelial–mesenchymaltransitions, Nat. Rev. Mol. Cell Biol. 7 (2006) 131–142, http://dx.doi.org/10.1038/nrm1835.

[87] A. Wodarz, I. Näthke, Cell polarity in development and cancer, Nat. Cell Biol. 9(2007) 1016–1024.

[88] K. Wolf, Y.I. Wu, Y. Liu, J. Geiger, E. Tam, C. Overall, et al., Multi-step pericellularproteolysis controls the transition from individual to collective cancer cell invasion,Nat. Cell Biol. 9 (2007) 893–904, http://dx.doi.org/10.1038/ncb1616.

[89] J. Oyanagi, T. Ogawa, H. Sato, S. Higashi, K. Miyazaki, Epithelial–mesenchymal tran-sition stimulates human cancer cells to extend microtubule-based invasive protru-sions and suppresses cell growth in collagen gel, PLoS ONE 7 (2012) e53209,http://dx.doi.org/10.1371/journal.pone.0053209.

[90] K.-V. Nguyen-Ngoc, K.J. Cheung, A. Brenot, E.R. Shamir, R.S. Gray, W.C. Hines, et al.,ECM microenvironment regulates collective migration and local dissemination innormal and malignant mammary epithelium, Proc. Natl. Acad. Sci. U. S. A. 109(2012) E2595–E2604, http://dx.doi.org/10.1073/pnas.1212834109.

[91] L.A. Gurski, X. Xu, L.N. Labrada, N.T. Nguyen, L. Xiao, K.L. van Golen, et al.,Hyaluronan (HA) interacting proteins RHAMM and hyaluronidase impact prostatecancer cell behavior and invadopodia formation in 3D HA-based hydrogels, PLoSONE 7 (2012) e50075, http://dx.doi.org/10.1371/journal.pone.0050075.

[92] M.P. Schwartz, B.D. Fairbanks, R.E. Rogers, R. Rangarajan, H. Zaman, K.S. Anseth, Asynthetic strategy for mimicking the extracellular matrix provides new insightabout tumor cell migration, Integr. Biol. 2 (2010) 32–40, http://dx.doi.org/10.1039/b912438a.

[93] C. Kraning-Rush, S. Carey, M. Lampi, C. Reinhart-King, Microfabricated collagentracks facilitate single cell metastatic invasion in 3D, Integr. Biol. 5 (2013)606–616.

[94] I. Rizvi, U.A. Gurkan, S. Tasoglu, N. Alagic, J.P. Celli, L.B. Mensah, Flow induces epi-thelial–mesenchymal transition, cellular heterogeneity and biomarker modulationin 3D ovarian cancer nodules, Proc. Natl. Acad. Sci. U. S. A. 110 (2013)E1974–E1983, http://dx.doi.org/10.1073/pnas.1216989110/-/DCSupplemental.www.pnas.org/cgi/doi/10.1073/pnas.1216989110.

[95] C.P. Huang, J. Lu, H. Seon, A.P. Lee, L.A. Flanagan, H. Kim, et al., Engineering micro-scale cellular niches for three-dimensional multicellular co-cultures, Lab Chip 9(2009) 1740–1748, http://dx.doi.org/10.1039/b818401a.

[96] W.J. Polacheck, J.L. Charest, R.D. Kamm, Interstitial flow influences direction oftumor cell migration through competing mechanisms, Proc. Natl. Acad. Sci. U. S. A.108 (2011) 11115–11120, http://dx.doi.org/10.1073/pnas.1103581108.

[97] V. Vickerman, J. Blundo, S. Chung, R.D. Kamm, Design, fabrication and implementa-tion of a novel multi-parameter control microfluidic platform for three-dimensional cell culture and real-time imaging, Lab Chip 8 (2008) 1468–1477.

[98] R. Hernández Vera, E. Genové, L. Alvarez, S. Borrós, R.D. Kamm, D. Lauffenburger,et al., Interstitial fluid flow intensity modulates endothelial sprouting in restrictedSrc-activated cell clusters during capillary morphogenesis, Tissue Eng. A 15 (2009)175–185, http://dx.doi.org/10.1089/ten.tea.2007.0314.

[99] C.E. Semino, R.D. Kamm, D.A. Lauffenburger, Autocrine EGF receptor activationme-diates endothelial cell migration and vascular morphogenesis induced by VEGFunder interstitial flow, Exp. Cell Res. 312 (2006) 289–298, http://dx.doi.org/10.1016/j.yexcr.2005.10.029.

[100] M.K. Shin, S.K. Kim, H. Jung, Integration of intra- and extravasation in one cell-based microfluidic chip for the study of cancer metastasis, Lab Chip 11 (2011)3880–3887, http://dx.doi.org/10.1039/c1lc20671k.

[101] V.L. Cross, Y. Zheng, N. Won Choi, S.S. Verbridge, B.A. Sutermaster, L.J. Bonassar,et al., Dense type I collagen matrices that support cellular remodeling andmicrofabrication for studies of tumor angiogenesis and vasculogenesis in vitro,Biomaterials 31 (2010) 8596–8607, http://dx.doi.org/10.1016/j.biomaterials.2010.07.072.

Please cite this article as:M. Alemany-Ribes, C.E. Semino, Bioengineering 3Ddx.doi.org/10.1016/j.addr.2014.06.004

[102] Y. Zheng, J. Chen, M. Craven, N.W. Choi, S. Totorica, A. Diaz-Santana, et al., In vitromicrovessels for the study of angiogenesis and thrombosis, Proc. Natl. Acad. Sci.U. S. A. 109 (2012) 9342–9347, http://dx.doi.org/10.1073/pnas.1201240109.

[103] J.S. Jeon, I.K. Zervantonakis, S. Chung, R.D. Kamm, J.L. Charest, In vitro model oftumor cell extravasation, PLoS ONE 8 (2013) e56910, http://dx.doi.org/10.1371/journal.pone.0056910.

[104] D. Barkan, J.E. Green, A.F. Chambers, Extracellular matrix: a gatekeeper in the tran-sition from dormancy to metastatic growth, Eur. J. Cancer 46 (2010) 1181–1188,http://dx.doi.org/10.1016/j.ejca.2010.02.027.

[105] R.F. Cox, A. Jenkinson, K. Pohl, F.J. O'Brien,M.P. Morgan, Osteomimicry ofmammaryadenocarcinoma cells in vitro; increased expression of bone matrix proteins andproliferation within a 3D collagen environment, PLoS ONE 7 (2012) e41679,http://dx.doi.org/10.1371/journal.pone.0041679.

[106] S.P. Pathi, D.D.W. Lin, J.R. Dorvee, L.A. Estroff, C. Fischbach, Hydroxyapatitenanoparticle-containing scaffolds for the study of breast cancer bone metastasis,Biomaterials 32 (2011) 5112–5122, http://dx.doi.org/10.1016/j.biomaterials.2011.03.055.

[107] A. Birgersdotter, R. Sandberg, I. Ernberg, Gene expression perturbation in vitro — agrowing case for three-dimensional (3D) culture systems, Semin. Cancer Biol. 15(2005) 405–412, http://dx.doi.org/10.1016/j.semcancer.2005.06.009.

[108] N.A. Bhowmick, E.G. Neilson, H.L. Moses, Stromal fibroblasts in cancer initiationand progression, Nat. Rev. Cancer 432 (2004) 332–337.

[109] M.M. Mueller, N.E. Fusenig, Friends or foes — bipolar effects of the tumour stromain cancer, Nat. Rev. Cancer 4 (2004) 839–849, http://dx.doi.org/10.1038/nrc1477.

[110] A.D. Leeper, J. Farrell, L.J. Williams, J.S. Thomas, J.M. Dixon, S.E. Wedden, et al., De-termining tamoxifen sensitivity using primary breast cancer tissue in collagen-based three-dimensional culture, Biomaterials 33 (2012) 907–915, http://dx.doi.org/10.1016/j.biomaterials.2011.10.028.

[111] H.H. Truong, J. de Sonneville, V.P.S. Ghotra, J. Xiong, L. Price, P.C.W. Hogendoorn,et al., Automated microinjection of cell–polymer suspensions in 3D ECM scaffoldsfor high-throughput quantitative cancer invasion screens, Biomaterials 33 (2012)181–188, http://dx.doi.org/10.1016/j.biomaterials.2011.09.049.

[112] S.-F. Lan, B. Starly, Alginate based 3D hydrogels as an in vitro co-culture modelplatform for the toxicity screening of new chemical entities, Toxicol. Appl.Pharmacol. 256 (2011) 62–72, http://dx.doi.org/10.1016/j.taap.2011.07.013.

[113] Z. Yang, X. Zhao, A 3D model of ovarian cancer cell lines on peptide nanofiber scaf-fold to explore the cell–scaffold interaction and chemotherapeutic resistance of an-ticancer drugs, Int. J. Nanomedicine 6 (2011) 303–310, http://dx.doi.org/10.2147/IJN.S15279.

[114] H.K. Dhiman, A.R. Ray, A.K. Panda, Three-dimensional chitosan scaffold-basedMCF-7 cell culture for the determination of the cytotoxicity of tamoxifen, Biomate-rials 26 (2005) 979–986, http://dx.doi.org/10.1016/j.biomaterials.2004.04.012.

[115] Y. Markovitz-bishitz, Y. Tauber, E. Afrimzon, N. Zurgil, M. Sobolev, Y. Shafran, et al.,A polymer microstructure array for the formation, culturing, and high throughputdrug screening of breast cancer spheroids, Biomaterials 31 (2010) 8436–8444,http://dx.doi.org/10.1016/j.biomaterials.2010.07.050.

[116] F. Xu, J. Celli, I. Rizvi, S. Moon, T. Hasan, U. Demirci, A three-dimensional in vitroovarian cancer coculture model using a high-throughput cell patterning platform,Biotechnol. J. 6 (2011) 204–212, http://dx.doi.org/10.1002/biot.201000340.

[117] T.R. Sodunke, K.K. Turner, S.A. Caldwell, K.W. Mcbride, M.J. Reginato, H. Moses,Micropatterns of Matrigel for three-dimensional epithelial cultures, Biomaterials28 (2007) 4006–4016, http://dx.doi.org/10.1016/j.biomaterials.2007.05.021.

[118] F. Deiss, A. Mazzeo, E. Hong, D.E. Ingber, R. Derda, G.M. Whitesides, Platform forhigh-throughput testing of the effect of soluble compounds on 3D cell cultures,Anal. Chem. 85 (17) (2013) 8085–8094, http://dx.doi.org/10.1021/ac400161j.

[119] I. Adanja, O. Debeir, V. Mégalizzi, R. Kiss, N. Warzée, C. Decaestecker, Automatedtracking of unmarked cells migrating in three-dimensional matrices applied toanti-cancer drug screening, Exp. Cell Res. 316 (2010) 181–193, http://dx.doi.org/10.1016/j.yexcr.2009.10.004.

[120] Q. Li, C. Chen, A. Kapadia, Q. Zhou, M.K. Harper, J. Schaack, et al., 3D models ofepithelial–mesenchymal transition in breast cancer metastasis: high-throughputscreening assay development, validation, and pilot screen, J. Biomol. Screen. 16(2011) 141–154, http://dx.doi.org/10.1177/1087057110392995.

[121] M. Kanapathipillai, A. Brock, D.E. Ingber, Nanoparticle targeting of anti-cancerdrugs that alter intracellular signaling or influence the tumor microenvironment,Adv. Drug Deliv. Rev. (2014), http://dx.doi.org/10.1016/j.addr.2014.05.005.

[122] Y. Chen, D.-Y. Gao, L. Huang, In vivo delivery of miRNAs for cancer therapy:challenges and strategies, Adv. Drug Deliv. Rev. (2014), http://dx.doi.org/10.1016/j.addr.2014.05.009.

[123] M. Egeblad, M.G. Rasch, V.M. Weaver, Dynamic interplay between the collagenscaffold and tumor evolution, Curr. Opin. Cell Biol. 22 (2010) 697–706, http://dx.doi.org/10.1016/j.ceb.2010.08.015.

[124] B. Ananthanarayanana, Y. Kima, S. Kumara, Elucidating the mechanobiology of ma-lignant brain tumors using a brain matrix–mimetic hyaluronic acid hydrogel plat-form, Biomaterials 32 (2011) 7913–7923.

[125] Y. Lianga, J. Jeonga, R.J. DeVoldera, C. Chac, F. Wangd, Y.W. Tonge, et al., A cell-instructive hydrogel to regulate malignancy of 3D tumor spheroids with matrixrigidity, Biomaterials 32 (2011) 9308–9315.

environments for cancermodels, Adv. DrugDeliv. Rev. (2014), http://