research article ERP

8/18/2019 research article ERP

1/12

Review article

Differentiation of mesenchymal stem cells for cartilage tissue

engineering: Individual and synergetic effects of three-dimensional

environment and mechanical loading

J.A. Panadero a,b,⇑, S. Lanceros-Mendez a,c, J.L. Gomez Ribelles b,d

a Center/Departament of Physics, Universidade do Minho, Campus de Gualtar, 4710-057 Braga, Portugalb Center for Biomaterials and Tissue Engineering, Universitat Politècnica de València, Camino de Vera s/n, 46022 Valencia, Spainc BCMaterials, Parque Científico y Tecnológico de Bizkaia, 48160-Derio, Spaind Networking Research Center on Bioengineering, Biomaterials and Nanomedicine (CIBER-BBN), Valencia, Spain

a r t i c l e i n f o

Article history:

Received 15 July 2015Received in revised form 17 December 2015Accepted 25 January 2016Available online 27 January 2016

Keywords:

Mesenchymal stem cellsChondrogenic differentiationCell adhesionMechanotransductionMechanical loading

a b s t r a c t

Chondrogenesis of dedifferentiated chondrocytes and mesenchymal stem cells is influenced not only bysoluble molecules like growth factors, but also by the cell environment itself. The latter is achievedthrough both mechanical cues – which act as stimulation factor and influences nutrient transport –and adhesion to extracellular matrix cues – which determine cell shape. Although the effects of solublemolecules and cell environment have been intensively addressed, few observations and conclusionsabout the interaction between the two have been achieved. In this work, we review the state of theart on the single effects between mechanical and biochemical cues, as well as on the combination of the two. Furthermore, we provide a discussion on the techniques currently used to determine themechanical properties of materials and tissues generated in vitro, their limitations and the futureresearch needs to properly address the identified problems.

Statement of Significance

The importance of biomechanical cues in chondrogenesis is well known. This paper reviews the existingliterature on the effect of mechanical stimulation on chondrogenic differentiation of mesenchymal stemcells in order to regenerate hyaline cartilage. Contradictory results found with respect to the effect of dif-ferent modes of external loading can be explained by the different properties of the scaffolding systemthat holds the cells, which determine cell adhesion and morphology and spatial distribution of cells, aswell as the stress transmission to the cells. Thus, this review seeks to provide an insight into the interplaybetween external loading program and scaffold properties during chondrogenic differentiation. Thereview of the literature reveals an important gap in the knowledge in this field and encourages newexperimental studies. The main issue is that in each of the few cases in which the interplay is investi-gated, just two groups of scaffolds are compared, leaving intermediate adhesion conditions out of study.The authors propose broader studies implementing new high-throughput techniques for mechanicalcharacterization of tissue engineering constructs and the inclusion of fatigue analysis as supportmethodology to more exhaustive mechanical characterization.

2016 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22. Differentiation in 3D culture without mechanical stimulus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2.1. 3D structure of cartilage in vivo, formation of cartilage during development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.2. Differentiation effects of 3D structures in vitro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

http://dx.doi.org/10.1016/j.actbio.2016.01.037

1742-7061/ 2016 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Center/Departament of Physics, Universidade do Minho, Campus de Gualtar, 4710-057 Braga, Portugal.

E-mail address: [email protected] (J.A. Panadero).

Acta Biomaterialia 33 (2016) 1–12

Contents lists available at ScienceDirect

Acta Biomaterialia

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / a c t a b i o m a t

http://dx.doi.org/10.1016/j.actbio.2016.01.037mailto:[email protected]://dx.doi.org/10.1016/j.actbio.2016.01.037http://www.sciencedirect.com/science/journal/17427061http://www.elsevier.com/locate/actabiomathttp://www.elsevier.com/locate/actabiomathttp://www.sciencedirect.com/science/journal/17427061http://dx.doi.org/10.1016/j.actbio.2016.01.037mailto:[email protected]://dx.doi.org/10.1016/j.actbio.2016.01.037http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://crossmark.crossref.org/dialog/?doi=10.1016/j.actbio.2016.01.037&domain=pdfhttp://-/?-http://-/?-


2/12

2.2.1. Pellets and micromass culture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.2.2. Encapsulation and scaffolding materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

3. Differentiation in 3D: mechanical loading effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.1. Mechanical loading in vivo. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.2. Mechanotransduction signaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.3. Mechanical loading in vitro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.4. Interaction of mechanical loading with 3D structure and cell adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

4. Mechanical measurements in cartilage tissue engineering: a necessity to complement the biochemical data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

5. Conclusions and open questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1. Introduction

The literature about cartilage engineering is large and somereviews can be found in references [1–9]. The chondrogenic capac-ity of mature chondrocytes dedifferentiating in monolayer expan-sion and mesenchymal stem cells has been demonstrated,chondrogenesis in vitro being ruled by the characteristics of theenvironment in which pluripotential cells are cultured. Three-dimensional scaffold materials, growth factor supply, cell–materialinteraction, hypoxia, mechanical stimulus and co-culture withmature cells have been emphasized as important factors drivingdifferentiation to the hyaline cartilage phenotype.

There are nevertheless important drawbacks in the literaturethat hinder reaching conclusions with respect to the influence of specific culture parameters on cell differentiation or cell biology.The simultaneous influence of many external factors on cell fatesometimes prevents reaching such conclusions. For instance, theeffect of mechanical stimulation is dependent on the scaffoldingmaterial in which cells are immersed, thus, simultaneous analysisof different studies carried out with particular supporting materi-als seems to be necessary to get general conclusions about thisfeature. The same can be stated about other culture parameters.

This reviewfocusses on the interplay betweenmechanical stim-uli and the 3D environment for chondrogenesis from mesenchymalstem cells. The main features of cell response when cultured in achondrogenic medium in a 3D environment in the absence of mechanical stimulus (individual effect of 3D structure) as well asthe general effects of mechanical loading have been described.Finally, the interplay between mechanical loading and biochemicalcues of the substrate is explored, as it demands more exhaustivestudies comparing materials with different cell adhesion patterns.This issue also demands new methodological approaches to iden-tify andunderstand that interplayand to manage the large quantityof generated information. Although high-throughput techniqueshave been implemented for the identification of extracellularmatrix components produced in chondrogenic differentiation

assays, which is useful to identify their role in mechanotransduc-tion processes, there is still lack of high-throughput techniquesfor mechanical characterization of tissue engineering constructs,as well as implementation of other relevant assays for the evalua-tion of mechanical properties, such as fatigue analysis. A discussionon these methodological needs and possible future solutions is alsoprovided in this work.

2. Differentiation in 3D culture without mechanical stimulus

2.1. 3D structure of cartilage in vivo, formation of cartilage during

development

In vivo, 3D structure is a regulator of cell shape and chondro-

genic differentiation during development in limb primordia. The

undifferentiated MSCs migrate to the sites of the developing carti-lage and condense through a specific combination of precartilagematrix and cell adhesion molecules, mainly N-cadherine and N-CAM [10]. This condensation allows essential cell–cell surfaceinteractions and signaling events that conclude in differentiationto hyaline chondrocytes. As differentiation progresses, the expres-sion of many focal adhesion receptors to early pericellular matrixmolecules [11,12], like fibronectin, is reduced. Consequently, mor-phological changes take place in the chondrogenic progenitors,from their fibroblastic-like shape to the more spherical morphol-ogy of hyaline chondrocytes. These morphological changes initiatea re-organization of cytoskeleton [13] organization, upregulatingthe synthesis of transcription factors (such as Sox5, Sox6 and spe-cially Sox9 [14]), which regulate the transcription and transductionof cartilage-specific extra-cellular matrix (ECM) molecules such ascollagen types II, IV, IX, and XI [15] and the highly-sulfated proteo-glycan aggrecan [16]. Then, the differentiated chondrocytes start tobe isolated by the ECM, whose molecules, like the proteoglycantenascin, favor the rounded shape of chondrocytes by steric hin-drance [17,18] of membrane receptors to molecules of the matrixlike fibronectin. At this stage of development, there are two possi-ble fates for hyaline chondrocytes: (a) in endochondral bone devel-

opment of long bones until adolescence, chondrocytes at the site of the growth plate become hypertrophic, produce alkaline phos-phatase and collagen type X and are eventually reabsorbed whilenew bone is formed, thus also being called transient chondrocytes[19] and (b) in the articular hyaline tissues, the chondrocytesremain with the mature phenotype for the rest of the lifespan of the organism, are separated in lacunae – formed by a highlyhydrated extracellular matrix (ECM) – and maintain the ECM of persistent hyaline cartilage [20]. Although the mechanisms deter-mining the two different fates remain unclear, it is known thatthe inhibition of N-cadherin and the route Wnt is a necessary stepto differentiate into cartilage after condensation [21]. In maturehyaline cartilage, each lacuna acts as an individual functional unitresponsible for maintaining the ECM metabolically. The ECM pro-

duced by the chondrocytes consists of 45–50% collagens (90% of which is collagen type II) and 20–25% proteoglycans (predomi-nantly aggrecan, decorin, and fibromodulin), whose negativelycharged glycosaminoglycans are responsible for swelling[20,22,23]. The relative values of these components, and even thespatial organization of them vary with the zone of cartilage [23].Moreover, the immediately surrounding matrix of the chondro-cytes, or pericellular matrix (PCM), is also enriched in some com-ponents that in the bulk ECM are less predominant, such ascollagen type VI [24], hyaluronan or biglycan, link protein,fibronectin and laminin [25–27].

In vivo, the ECM and the PCM drive the three-dimensional con-formation to their components by integrin-based cell adhesion.The importance of 3D environment configuration for chondrogen-

esis is such that even MSCs cultured in decellularized cartilage

2 J.A. Panadero et al. / Acta Biomaterialia 33 (2016) 1–12


3/12

in vitro are capable of chondrogenic differentiation without addi-tion of other exogenous factors [28–30]. As indicated above, themorphological change to the rounded shape and the subsequentreorganization of the cytoskeleton is a key component in the initi-ation of chondrogenic differentiation. The biophysical interactionbetween integrins with the collagens and proteoglycans in the car-tilage matrix [31,32], and the decrease in the interaction bindingswith other matrix components like fibronectin, non-specific of car-tilage, are responsible for the round shape. These membrane recep-tors to the matrix are linked to the three cytoskeletal networks:largely to actin microfilaments, and in a lesser extent, to interme-diate filaments and microtubules [16,33–35]. The interactionsdetermine the polymerization and organization of the cytoskeletalfilaments, and the dissociation of many receptor–matrix com-plexes in cartilage dissipates partially the isotonic tension betweenthe contractile forces of the cytoskeleton and the resistive force of the matrix [36]. This tension produces a signal propagation fromadhesion sites and transmit signals by linking to the nucleus[37], resulting in the control of protein expression and post-translational modifications. Thus, the interactions of the cells withtheir matrix involve intracellular signaling beyond just the mor-phology of the cells. Disruption of tubulin (microtubules) withchemical agents like colcemid and nocodazole does not lead tochondrogenic differentiation of stem cells [38]. Chondrogenic dif-ferentiation is only promoted by agents like cytochalasin D [33],which disrupt actin cytoskeleton through direct effects in severalparts of the network (capping of the endings of actin microfila-ments [39], hinder linking of integrin-binding with fibronectinand regulation of mRNA transcription in the nucleus [40]). Inhibi-tion of regulatory metabolites in some pathways like RhoA, whichis a negative actin cytoskeleton-regulating protein, enhances chon-drogenesis. However, the inhibition of the other downstreammetabolite in this pathway, Rho-associated protein kinase (ROCK),does not affect chondrogenesis, suggesting that there are unknownalternative pathways [41].

2.2. Differentiation effects of 3D structures in vitro

2.2.1. Pellets and micromass culture

Cell culture in 2D does not resemble the in vivo situation, and itis known that mature chondrocytes de-differentiate [42,43].Further, MSCs take a spread, flattened morphology and develop afibrous phenotype, resembling, at the most, only the chondrocytesof the articular surface. 2D configuration enhances the natural ten-dency of MSCs to express relatively more collagen type I [44], thanGAG and collagen II. The first strategies for MSC differentiationtoward chondrogenic in 3D were designed to mimic the conditionsduring stem cell condensation: pellet and micromass cultures. Pel-let cultures consist of centrifuging the MSC suspension in a conicaltube and then incubation. After 24 h in culture, the cells aggregate

and form a round cell pellet. MSCs are capable of chondrogenic dif-ferentiation in pellet culture using serum-free medium containingglucocorticoids and TGF-b family [45]. However, it is hard to obtaina satisfactory cartilage in pellet culture. Cells are often found undif-ferentiated or necrotized in the central regionof the pellet andonlythe outside layer cells undergo chondrogenic differentiation [46–48]. Additionally, MSCs in the pellet culture show inducedfibrocartilage-like features such as expression of collagen I andhypertrophy, as shown by upregulation of collagen X [47].

In micromass cultures, a droplet of cell suspension is carefullyplaced in the center of each well of a multiwell plate. Cells areallowed to adhere at 37 C for some hours, followed by the additionof chondrogenic medium. After 24 h, the cells in every dropletmerge and form a spherical mass. All chondrogenic cultures are

performed for more than 21 days. It has been reported that micro-mass culture systems enhance chondrogenesis more than standard

pellet systems [49]. When compared with pellet cultures, in themicromasses the cartilage-like tissue is more homogenous andenriched in collagen II, decreasing the expression of fibrocartilagecollagen and collagen X. One reason for the existence of collagentype I and type X can be the lack of time regulation of the adhesionmolecules typical from condensation, which can result in cellsremaining in pre-cartilage stage. The nutrients and the oxygenreach the central zone of the micromasses easier than the pellets,which usually lead to necrotized tissue.

Both micromass and pellet cultures have been useful to showthat cell and nuclear shape are strong regulators of cell growthand physiology also in vitro, in particular, differentiation of adultor embryonic stem cells into a chondrocytic phenotype requiresa rounded cell shape [45,50,51]. In bone-marrow-derived MSCs itis shown that a more rounded nuclear shape was associated tothe larger expression of molecular chondrogenic markers [52].

2.2.2. Encapsulation and scaffolding materials

Although micromass culture being a suitable as model is able tomimic some of the conditions for chondrogenesis during develop-ment, it has two main problems as a therapy: it is not suitable forimplantation and produce necrotic problems, the latter attributedto too close contact, that, as it should be reminded, in adultsdisappears when the lacunae are formed.

Thus, the most common solution is to embed the cell suspen-sion in a surrounding environment capable to retain the cells,which can simulate the biomechanical environment existingin vivo. Two kinds of biomaterial supports are employed: hydro-gels, formed by suspension of cells within a solution that encapsu-lates them, and macroporous scaffolds. Both can surpass somelimitations of micromass and pellet systems. These scaffolds mustbe evaluated first in vitro to comprehend their effects in differen-tiation. 3D systems for in vitro culture provide more surface for celladhesion and proliferation. In vivo, scaffolds prevent the diffusionof transplanted cells and can improve integration.

A broad set of different materials have been used to produce

hydrogels for cartilage tissue engineering: proteins as collagentype I or collagen type II [53–55], fibrin [56–58], elastin-likepolypeptides [59], polysaccharides as hyaluronic acid [60–62], chi-tosan [63–65], chondroitin sulfate [61]; agar gel, gellan gum [66],synthetic hydrogels as crosslinked poly(ethylene glycol) [67],poly(vinyl alcohol) [68] and others. Hydrogels are formed bycross-linking of chain molecules, normally encapsulating the cellsat the beginning of the culture, which is an advantage for surgerysolutions. Hydrogels show interesting characteristics such as thepossibility of obtaining highly swollen structures approaching sim-ilar conditions than cartilage natural ECM and also the possibilityof obtaining cell homogeneous distributions.

Other non–hydrogel scaffolding systems are produced for carti-lage tissue engineering, mainly semicrystalline polymers, because

their viscoelastic behavior can be suitable for the mechanical envi-ronment of the cartilage. These polymers [69] include hydropho-bous biodegradable polyesters, as well as their copolymers suchas polylactide [70], polyglicolide [71], polycaprolactone [72–74],biodegradable poly(ether ester) multiblock copolymers [75], poly(3-hydroxybutyrateco-3-hydroxyhexanoate) [76], and also bio-stable acrylic polymers [77,78] among others [79]. Many combina-tions of these hydrogels with polymer scaffolds can exist, forexample fibrin with PCL [80], semicrystalline polymers providinga broader set of structures due to the extended plethora of processing methods [81].

2.2.2.1. Hydrogels. As cell shape is a factor of differentiation, theeffect of scaffolds on chondrogenesis is mainly regulated by cell

interactions with the matrix and the morphologic configurationacquired as a consequence [82,83]. In adherent 2D surfaces that

J.A. Panadero et al. / Acta Biomaterialia 33 (2016) 1–12 3

http://-/?-http://-/?-


4/12

favor integrin binding, such as fibrin or disperse patterned PDMS,MSCs adopt spread morphology and differentiate spontaneouslytoward myocites [84] in chondrogenic medium. Nevertheless, inless adherent 2D surfaces that promote more spherical morpholo-gies, MSCs differentiate toward chondrocytes in the same medium.In 3D, encapsulation of MSCs in non-adhesive hydrogels, such asPEG, alginate or agarose, is permissive to chondrogenesis becausethe cells adopt a rounded shape, which favors expression of chon-drogenic markers by direct cell–cell contact, similarly to pellet andmicromass systems and cell condensation. 3D adherent gels can bepotentially inductive of chondrogenesis by mimicking the interac-tion of the isolated chondrocytes with the components of the peri-cellular matrix of adult cartilage. Fibrin, a hydrogel with a highadhesive internal surface, is largely employed, and can be suitablefor chondrogenic differentiation in specific medium [85,86]. Itcould be expected that the differences in adhesion could lead todifferences in the commitment of MSCs toward chondrogenic lin-eage. Few experiments compare directly an adherent gel vs. non-adherent. Some comparisons between fibrin and agarose havebeen carried out for chondrogenic differentiation of MSCs, butthe different works are not conclusive. While one of them [87]shows fewer accumulation of GAG and collagen type II for adher-ent gels (fibrin) with respect to non-adherent gels, a recent study[88] shows no differences between the two types (alginate vs. fib-rin) in terms of chondrogenic marker expression. Only in a simu-lated osteochondral environment the agarose showed a superiorresponse. This finding can indicate that other environmental condi-tions in vitro, like soluble growth factors, may mask the effects of the 3D conformation, unlike in in vivo. Fibrin properties are highlydependent of the used concentrations of fibrinogen and thrombin[89–91]. The discrepancies among experiments could be causedby the difference in fibrin composition. Signaling complexity in3D also increases with respect to 2D. There are several metabolicpathways involved in the differentiation in both 2D and 3D sys-tems. However, the differences in these pathways between thetwo systems have not been identified yet [92]. Hyaluronic acid is

the natural backbone of the proteoglycans of hyaline cartilage.Hyaluronic acid gels are common gels used for chondrogenic dif-ferentiation, which present a lower cell adhesion degree than fibrin[93], but with interactions in some points through CD44 cell recep-tors [94], being more adhesive than agarose or PEG. This environ-ment improves chondrogenesis of MSCs compared to PEG gels inthe same conditions [95], allowing both interactions with thematrix and cell–cell contact [96]. The expression and synthesis of chondrogenic markers in hyaluronan gels is also dependent of the macromere density and cross-linking of the gel [97].

2.2.2.2. Macroporous scaffolds. Macroporous sponges fabricatedfrom molecules found in ECMs (e.g. collagens, GAG) or analogmolecules provide direct adhesion properties to cell ligands,

mainly integrins and cadherins. However, processing methods forthese materials are more limited and few structures can bedesigned. Thus, other non-bioactive materials are also used andthe cell response to their surface is not mediated by a direct con-tact, but rather through an interfacial layer formed on material sur-face once it is in contact with a physiological environment. In nodefined culture media, such a layer is created as result of non-specific adsorption of pericellular matrix proteins, like fibronectin,laminin and vitronectin, which interact with surface integrins [98].This protein deposition not only is fundamental for cell adhesion[99], but also influences posterior cell events like proliferation,migration and differentiation [100]. For chondrogenesis, usuallythis layer formation is controlled in some manner through func-tionalization with specific proteins and allowing non-specific

attaching before cell culture [101]. Scaffold chemistry influencessurface properties such as morphology, hydrophilicity, surface

energy and charge, which control this protein adsorption [69].Thus, by tuning these parameters through suitable physico-chemical modifications, the creation and characteristics of thislayer can be guided. Moreover, surfaces can be modified to attachspecific proteins in desired conformations, through grafting speci-fic peptide ligands, physically or covalently, such as RGD sequences[82]. In these sponges, when pore diameters are significantly largerthan the cell diameter (more than 100

lm), the surface presented

to cells ranges in the micrometric scale and the cells adhere in amanner resembling cell adhesion on 2D substrates [82,102]. Forexample, in polyurethane scaffolds, encapsulating the MSCs in fib-rin inside the scaffolds results in higher GAG and collagen type IIthan seeding the cells directly to the pore walls [86]. Reducing fiberdiameter to nanoscale can potentially enhance the chondrogenicdifferentiation of mesenchymal stem when compared with thesame material as a sponge of thicker fibers [103,104], althoughmore data are needed to reach consistent conclusions. The expla-nation can be that more binding points can be established byincreasing porosity, interconnectivity and permeability, parame-ters that influence differentiation through fluid flow, cell migra-tion, cell–cell contact and nutrient and soluble factors localconcentration [97,105–108].

3. Differentiation in 3D: mechanical loading effects

3.1. Mechanical loading in vivo

Mechanical loading is also an important factor regulating andmaintaining the chondrocyte phenotype. In native hyaline carti-lage, organization of collagen, water uptake by GAG and othermatrix components provides mechanical stiffness able to resistmechanical loading caused by joint movement and weight bearing,uniform compressive normal stresses ranging from 5 to 10 MPa[109] and frequencies between 0.1 and 10 Hz [110]. From a mate-

rials science point of view, the cartilage behaves as a viscoelasticmaterial that deforms easily at small strains but stiffens whilestrain is increased [111,112]. Not only the mechanical propertiesof ECM allow the cartilage to accomplish its function, but they alsoregulate the stress transmission to the cells. There is a feedback sit-uation between loading and matrix synthesis, because the loadsthat must sustain the matrix act as signaling factor to the expres-sion and production of its components, which provide at the sametime the mechanical properties of cartilage. Mechanical loading isthus essential for proper musculoskeletal development [113].Therefore, the knowledge on mechanical properties of tissue envi-ronment and the ways in which loadings are transmitted to cells isfundamental.

The mechanical forces acting in the knee are varying within the

articular zones, due to differences in composition and structure. Inthe superficial zone, the highest strains can be found (up to 50%)as well as the highest fluid flow, being also the most resistant zoneto shear stress. This is the only zone where the interstitial fluid canflow out of the cartilage when it squeezes, through the surface,andthe hydrostatic pressure is the lowestof all cartilages. Also, thiszone sustains shear stress resulting from angular displacement of the two sides of the joint. In the middle zone the strains rangebetween 10%and 20%and there is less fluid flow and, from this zoneto the lower parts, it is limited to inside the cartilagematrix, as sub-chondralboneandadjacenttissueconfinetothesezones.Inthedeepzone, the strains are 0–5% and practically fluid movement does notexist. In this zone, the higher hydrostatic pressures are found[114]. This difference in loads results in anisotropic mechanical

properties throughthe cartilage,withthe elastic modulusof hyalinecartilage increasing with depth. For example in bovine, values of



5/12

modulus are 0.08 MPa at the surface and 2.1 MPa in the deep zone.The apparent modulus of the whole cartilage is 0.38 MPa [115].

3.2. Mechanotransduction signaling

Although the macroscopic effects of mechanical loading in vivoare well established [113], the molecular mechanisms still remainunclear. Recent research is unraveling that the mechanotransduc-tion in chondrocytes in vivo is mediated through mechanorecep-tors in the plasma membrane, mainly integrins, associated withstretch-activated ion channels and voltage-gated calcium channel,like TRPV4 [116]. As indicated before, integrins are binding pro-teins to the pericellular matrix, which play an important role inthe transmission of loading [24]. The biochemical transduction ismediated through the cytoskeleton also, as integrins interact withfocal adhesion kinase and the cytoskeleton. For example, vimentinintermediate filaments are thought to play a role in mechanosens-ing, specifically to strain deformations [117]. An immotile primarycilium on human mesenchymal stem cells [118], arthritic chon-droprogenitor cells (CPCs) and chondrocytes [119] has been alsoidentified, which is directly associated with many of themechanoreceptors mentioned previously[120]. Furthermore, theprimary cilium in chondrocytes also contains integrins [121],which bind with the pericellular matrix, and therefore also havea role in mechanotransduction [122,123]. Primary cilium trans-duces mechanical loading and regulates metabolic routes such asHedgehog [124], Wnt [125], and TGF/BMP [126]. In mature chon-drocytes, the cilium is disgregated with mechanical loading. Themean length of the cilium is different depending on the zone of the cartilage [127], because of the different mechanical environ-ment in each zone, which has also been verified in vitro. In refer-ence [128], it has been shown that the primary cilium lengthdepends on the strain applied over the cells. The length seems tobe a determinant factor of activation of the Hedgehog route, whichis an antagonist route of chondrogenesis that if expressed in chon-drogenic precursors leads them to a phenotype similar to the

hypertrophic phenotype of the end-stage differentiation in thegrowth plate, and finally, results in osteoarthritic cartilage [129].If the cilium is not disrupted, the route is activated. On the otherhand, the absence of cilium also activates this route [124,130]. Itsuggests the hypothesis that only a range of length is optimal fornon-hypertrophic chondrogenesis. As a possible mechanism, afterthe initial stages of cartilage development, the cilium length isreduced by loading, leading mature hyaline cartilage phenotype,until it finally shortens too much and becomes osteoarthritic. If this proposed mechanism could be corroborated, it would explainthe role of the cilium in mechanotransduction during differentia-tion. In mesenchymal stem cells, the primary cilium is related toits differentiation potential and its inhibition reduces the expres-sion of differentiation markers [118]. However, from all the lineage

commitments, the chondrogenic potential is the least affected bythis inhibition, because the levels of sox9 are the least reduced,suggesting additional mechanisms of regulation.

Although the influence of surface topography in primary ciliumlength has been studied in MSCs [131], it remains unclear howmechanical loading produces the changes in the length of primarycilium in MSCs, and howit regulatestheir chondrogenicdifferentia-tion. The primary cilium has also been proposed to transduce themechanical loading through the PKA route, which is a recent routeof mechanical transduction, through a cascade resulting in the acti-vation of sox9 transcription factor [132]. However, there is still noevidence.

The purinergic calcium signaling is a broad regulation mecha-nism of many biological processes and has been recently identified

also as a mechanotransduction pathway taking part not only in thephenotype regulation of mature chondrocytes, but also in the

chondrogenic differentiation of MSCs [133]. The calcium releaseis activated by purinergic receptors, composed of two families:P1 receptors, activated preferentially by adenosine, and P2 recep-tor, activated by a variety of nucleosides, which are subdivided inligand-gated ion channels (P2X)-activated by adenosine triphos-phate (ATP)- and G-protein coupled receptors (P2Y)-activated bynucleotides, di- or triphosphates, purines or pyrimidines. Recep-tors from both families have been found to participate in chondro-genic mechanotransduction. It seems that dynamic loadingdownregulates CD73 (50-nucleotidase), which dephosphorylatescyclic adenosine monophosphate (cAMP) to generate extracellularphosphate and adenosine, as well as the adenosine A2a receptorfrom the P1 family, leading to chondrogenic differentiation[134,135]. Dynamic loading can also trigger ATP release to theextracellular environment, activating calcium signaling via ATPbinding to purine P2 receptors [136] and chondrogenic differenti-ation of MSCs. However, the exact mechanisms are not elucidatedyet, because ATP release can occur by different routes. When MSCsare subjected to oscillatory fluid flow, vesicular release mecha-nisms [137] have been identified, but with dynamic compression,the purinergic response involves activation of connexin hemichan-nels and unidentified mechanisms. Purinergic receptors (ATP) canbe found also in the primary cilium of chondrocytes [138].

3.3. Mechanical loading in vitro

As previously stated, mechanical loading is an important signal-ing factor for the correct regulation of ECM in vivo. To investigateits effects in vitro, external mechanical stresses have been appliedthrough bioreactor devices, first in micromass cultures [139] andhydrogels with mature chondrocytes and more recently with mes-enchymal stem cells. The loadings in the hyaline cartilage are acomplex combination of tensile, shear, and compressive stressesand strains. In vitro, the ways of providing mechanical stressesare based in the reduction to single mechanical components –thoroughly reviewed by O’Conor et al. [140]: unconfined uniaxial

compression [141–143], direct shear stress [144–146] and perfu-sion (shear stress) [147–149] and hydrostatic pressure [150–153]. Each one of these mechanical solicitations simulate one orsome of the different components of mechanical stress in vivoand each one matches more closely to the conditions given inthe different zones of cartilage. Unconfined compression resemblesthe conditions found in the upper zones of the cartilage withhigher medium flow, low hydrostatic pressure, high stresses (indisplacement control) or low strains (in force control). Direct shearstress can be combined with unconfined compression. Hydrostaticpressure resembles better the condition in the deeper zones. Ide-ally, semi-confined compression would resemble a gradient of allthe zones, but has been used only recently [154] asit is hardto per-form for cell culture, requiringa permeable load plate over the load

zone and impermeable boundaries around the rest of the scaffold[114]. Some systems have combined compression and shear stress[155], and this combination enhances chondrogenic matrix pro-duction in MSCs with respect to each single loading mode [156].Moreover, the hypertrophic markers (col X) did not increase andthe GAG liberation is constant, altogether indicating that the com-bination of loading modes prevents hypertrophy.

Effects of compression and hydrostatic pressure have beenextensively studied for chondrogenic differentiation. It is knownthat for both compression and hydrostatic pressure in hydrogels,intermittent or dynamic application of mechanical stresses[141,150,151,157–159] produces higher expressions of collagen IIand proteoglycan expression than static loading [160,161] and bet-ter mechanical properties in mature chondrocytes in explants and

hydrogels. Success of application of dynamic loads depends on thefrequency of application [144], typically 1 Hz and the addition of


http://-/?-


6/12

repose periods is necessary. The period of continuous cycling load-ing in knee rarely exceeds 1 h and is dispersed through all day withresting periods, which reach at least 16 h [114].

3.4. Interaction of mechanical loading with 3D structure and cell

adhesion

In non-adherent hydrogels under cyclic hydrostatic pressure[162], the presence of a pericellular matrix, formed by the cellsor artificially recreated with coating, is mandatory for the differen-tiation of MSCs. Despite the lack of studies about the necessity of this pericellular matrix under cyclic compression, the findingsabout the resistance to non-cyclic compression of this matrix[163] suggest that it could show the same effect under cyclic com-pression. Although MSC cyclic loadings are capable to expresschondrogenic markers and to produce ECM in non-adherent hydro-gels, they need a pre-differentiation step and some initial pericel-lular matrix in order to sustain the same dynamic loads thandifferentiated chondrocytes for cultures until 21–28 days[141,164]. This effect is more noticeable in cultures until 42 dayswith daily loadings, in which levels of markers can drop below

free-swelling conditions [165]. If a pericellular matrix is allowedto be generated for at least 2 weeks, higher expression and matrixsynthesis is obtained. With respect to adherent hydrogels undercyclic hydrostatic pressure, it seems that the MSC responds posi-tively to this loading even if it is applied from the first days, theopposite to non-adherent gels. MSCs in fibrin accumulate moreGAG and reduce collagen type I under the effects of dynamichydrostatic pressure than non-adherent hydrogels, in whichhydrostatic pressure has little effect [162].

For the cases in which direct cyclic compression is applied toMSCs, the difference between adherent and non-adherent gels isnot as clear. In a direct comparison between agarose and fibrin gels

[87], cyclic compression from day 0 inhibits GAG and total collagenaccumulation in both gels in short-term, but the accumulation isenhanced after 42 days. This indicates that fibrin may not be actingwith the same function as a pericellular matrix generated in a pre-differentiation step, and it does not improve early chondrogenicdifferentiation over agarose. Nevertheless, the delay of the com-pression regime in fibrin only enhances accumulation of GAGand not of collagen type II, which remains higher if the compres-sion is performed from the beginning. In any case, fibrin gels sup-port chondrogenic differentiation under cyclic compression [166],although it is not clear if the effects over chondrogenesis arecaused by the cell morphology conformation, or because of hin-drance of hydrodynamic and transport of nutrients [86]. In fibringels, when cyclic loading is applied, MSCs first keepun-differentiated traits, but in long-term culture, they increaseexpression of chondrogenic markers compared to unloaded gelsand inhibit myogenic markers even when loads are applied fromthe first day. In less adhesive gels, like hyaluronic acid gels,dynamic compression improves GAG and collagen deposition after70 days despite the expression levels remaining statistically simi-lar to free-swelling conditions. Unlike agarose hydrogels [167],the loading from early times (3 days) did not hinder chondrogene-sis [168], suggesting that the adhesion may provide a better envi-ronment for mechanotransduction. However, the differencesbetween hyaluronan and fibrin could suggest that the degree of this adhesion also influences the response to cyclic compression.

Recapitulating, under cyclic hydrostatic pressure, the disparityof the adhesion degree (non-adherent vs. highly-adherent gels)produces different responses, while under direct cyclic compres-sion, the differences are found in mid-term adhesive gels, leavingnon-adherent and highly-adherent gels with similar responses(Fig. 1). If the differences in chondrogenic differentiation are notcaused by other indirect factors like nutrient diffusion, this could

Fig. 1. General description of the effects of the different types of cyclic loadings on MSC chondrogenic differentiation, depending on the grade of adhesion provided by theencapsulating gel and according to the data of comparative studies between gels. Only the cases in which the cyclic loading has been applied from the beginning of cell

culturein differentiation mediumhave been taken intoconsideration. The graph shows the resultsin base of theincreasing trend of the gels to form cell–matrix bindings, dueto the absence of quantitative determination of the level of adhesion in the reviewed cases.



7/12

suggest that the mechanosensing of MSCs behaves differentlywhen the loading is homogeneous (hydrostatic pressure) thanwhen it is anisotropic (uniaxial compression). It would be interest-ing to study the effect of the adhesion in other types of cyclic load-ings like the combination of uniaxial compression and shear stress.

It should be noted that in native cartilage, chondrocytes in thedifferent zones show different morphology [169]: The typicalround shape found in the literature corresponds to the chondro-cytes from the middle and deep radial zone, but in the superficialzone, the chondrocytes are more spread. These spreading chondro-cytes can express collagen type I and lower amounts of collagentype II relative to proteoglycans than the round chondrocytes.Although there is no evidence that dynamic loading transforms cellshape in vitro, the different loading profiles in the different zonesand the varying matrix composition and structure may be relatedto the different morphology through different pathways of mechanotransduction. However, it is difficult to identify all therelations among the three parameters – cell shape, matrix andmechanical loading – and to categorize each one as cause or conse-quence of the other, due to the feed-back nature of the interactionsin cartilage. The results reviewed can be the first steps into unrav-eling these interactions.

Generally, it has been concluded that mesenchymal stem cellsneed a pre-differentiation step before applying the dynamic load-ing, in order to produce a pericellular matrix [170]. However, allthe cases that support this finding have been performed withnon-adherent hydrogels. The findings described in our review, forcyclic hydrostatic pressure, could suggest that the response of mes-enchymal stem cells with a pre-differentiation step is due to thegeneration of some pericellular matrix that could be acting as anadherent hydrogel, and not directly by the state of the cells. Thus,the presence of an adherent hydrogel from the beginning couldmake unnecessary a pre-differentiation step, although more evi-dence needs to be found to determine the optimal initial condi-tions in chondrogenic differentiation of MSCs under mechanicalstimuli. However, this hypothesis failed to be valid for cyclic com-

pression [87]. It is necessary to indicate that, in any case, the peri-cellular matrix plays an important role in retention of thesynthesized GAG, and fibrin for instance only retains 30–50% of GAGs [144]. Thus, the pericellular matrix could still be necessaryin chondrogenic differentiation for other reasons beyond a loadingshelter effect.

The cytoskeleton, which plays an important role in the adhe-sion, the morphology and the chondrogenic differentiation in theabsence of mechanical cues, as seen in previous sections, can alsobe involved in the response of chondrocytes to mechanical loading,suggest the fact that actin cytoskeleton of mature chondrocytesundergoes remodeling under mechanical loading [171]. However,no one has found a direct relation between the cytoskeleton state(which determines cell shape) and the response to mechanical

loading. Moreover, the role of cytoskeleton in mechanotransduc-tion of mature chondrocytes may be different to the cytoskeletonof undifferentiated MSCs. Actually, measurements of mechanicalproperties in individual cells find that undifferentiated MSCs havea higher elastic modulus than mature chondrocytes, as a conse-quence of the stronger cytoskeleton [172]. Thus, it is reasonableto think that MSCs should sustain mechanical loadings in differentways. Without mechanical loading, the effect of a rounded cellshape can resemble the situation of MSCs in thepre-condensation step, a stage of the development when themechanical cues are not present. It has been suggested thatdynamic compression can inhibit N-cadherin activity, the mainbinding protein during condensation [87]. Without loading,spreading increases N-cadherin expression, leading to the

up-regulation of myogenic genes. Like in vivo, the expression of N-cadherin is necessary at the beginning but it is reduced as

chondrogenesis progresses, thus in an in vitro culture the expres-sion should be reduced at the end. If this is correct, it could explainthe lack of differentiation response of MSCs to myogenic and chon-drogenic lineage at the beginning of the culture, but it would meanthat chondrogenesis would be induced by skipping the condensa-tion step, or simulating the final steps of condensation to maturecartilage. If cadherin is added to HA hydrogels, it promotes chon-drogenesis of MSCs but only for the early days [96].

The response of MSCs to mechanical loading seems also condi-tioned by other factors. For instance, under the same mechanicaland surrounding structure conditions, MSCs respond better withlower concentrations of TGF-b than in non-loaded cultures [173],what suggests that mechanical loading acts via similar pathways,and because of it, higher concentration of TGF-b masks the effectsof mechanical loading.

4. Mechanical measurements in cartilage tissue engineering: a

necessity to complement the biochemical data

Traditionally, in the experiments for chondrogenic differentia-tion, only the expression and translation predominant ECM mark-ers (collagen type II, aggrecan, sox9,

. . .) were analyzed, and other

components as collagens IX, XI were neglected, whose contributionto cartilage development is completely necessary, as could be seenin knock-out mice [174,175], in which the depletion of the expres-sion of these genes leads to loss of function. In the last years, theresources to describe chondrogenic differentiation have beenexpanded through the use of the ‘‘omics” high-throughput screen-ing techniques: transcriptomics (with microarray technology)[176–178] and proteomics (with mass spectrometry) [179]. Theseapproaches serve to obtain more accurate data of the MSC signal-ing and more detailed information about the ECM compositionafter differentiation. However, these techniques generate hugeamounts of data, which make it difficult to interpret the specificrelations among data. Although there have been efforts to separate

the effects of matrix components in cartilage by bottom-upapproaches of measuring single ECM components [180], the inte-gration of these data in a coherent model is defying. Nevertheless,the –omics have a great potential to describe complete profiles of the ECM and soluble components produced and their role in chon-drogenic differentiation.

The analysis of the cell-ECM-scaffold constructs is mandatory,because from a structural point of view, the mechanical propertiesof the constructs should be defined by the cell surrounding envi-ronment, which determine its functionality as medical substitutesonly if they are capable to resist the mechanical loading and trans-mit its signals properly. The analysis of mechanical propertiescomplements the data obtained from biochemical assays of matrixexpression and deposition, because a total correlation among

mechanical and biochemical data has not been found yet[142,179], and one cannot be inferred from the other. Despite theimportance of mechanical properties, we consider that, comparingwith the advance in the tools that have been developed for ECMscreening, there are still gaps in the implementation of mechanicalmeasurements of constructs after culture in vitro or after trans-plant in vivo.

The first gap is that more efforts are needed to develop non-destructive and high-throughput screening techniques formechanical properties. The techniques for –omics analysis requiredestruction of the samples. The usual techniques for mechanicalanalysis – summarized in Table 1 – are also destructive, whichadds the difficulty for working with a high number of samples, incell cultures that minimally require times of 28 days. However, if

the samples are employed for mechanical measurements with anon-destructive method, like indentation [181], they can be used



8/12

for other purposes, granted that sterile and clean conditions arekept. Compression tests could be non-destructive with the possi-bility of performing them at the same time of mechanical stimula-tion, via adding force sensors to the actuators of the bioreactors.But keeping the sensors in good working conditions in the harshoxidating conditions of cell culture media is challenging. Withrespect to high throughput screening mechanical analysis, onlyrecently a device has been developed for this purpose [182], cap-

able to measure the compressive equilibrium modulus of up to48 samples simultaneously. It still has a limitation of disparity onprecision (larger standard deviations compared to sample-by-sample testing), albeit, it has a strong potential if combinedwith –omics approaches. Furthermore, if an independent displace-ment or force control could be added, this device could serve asloading bioreactor and data collector at the same time.

The second gap is that the traditionally measured propertiesmay not be sufficient to fully characterize mechanically materialsthat are submitted to cyclic loading like in the knee. Hitherto, thestudy of mechanical properties uses parameters common in mate-rials science (Table 2). However fatigue, a phenomenon appearingin all materials when they are submitted to cyclic loading [183],has not been discussed properly for biomaterials and ECMyet. Fati-

gue in material sciences is understood as the irreversible structuraldamage produced in a material that is subjected to cyclic loading.

Fracture-related failure has been assessed in cartilage and in some

hydrogels with techniques that evaluate resistance to a crack gen-erated in purpose, that grows depending on the test, analogous to aseries of standard methodology in material science [184]. How-ever, the material fabrication processes, and the macro- andmicro-pores in scaffolds themselves, introduce inherent imperfec-tions that can act as stress concentration points [185,186], whichcan initiate fatigue effects, where cyclic loadings and permanentdeformation can occur. Compression to failure in hydrogels orother biomaterials has been barely studied [187] and never undercyclic loading. Before a correlation between all relevant compo-nents of ECM under mechanical loading is found, it is importantto ensure the proper knowledge on the mechanical properties of the biomaterials and their relationship to overall cell response.Some preliminary works show that fatigue in scaffolds with

in vitro generated ECM provides information that other commonmeasurements miss. In a pioneering work [188], fatigue behavioris different for Poly-e-caprolactone (PCL) scaffolds with chondro-genic precursors seeded inside their pores – which produce ECM– that for the same scaffolds with other filler, despite that the ini-tial elastic modulus is the same. It is not clear why this phe-nomenon happens, but it could be related with the boundariesbetween the filler, that can act also as different stress concentra-tion points than other discontinuities. Fatigue analysis has thepotential to model predictions of the mechanical resistance of aconstruct that would be implanted in the knee. However, it is stillan opening discipline that needs to be improved, widening the per-formance parameters (frequency, strain, number of cycles, . . .),increasing the scaffolding materials tested and including other

extracellular matrices and native cartilage explants. Furthermore,because the model used in fatigue analysis in that work is a modeldeveloped for metals, it is necessary to create specific models morestatistically suitable for viscoelastic materials.

5. Conclusions and open questions

The differentiation of mesenchymal stemcells to obtain hyaline-like structures in vitro, requires seeding and culturing in 3D config-urations, which can adopt many forms depending on the geometryand the physico-chemical characteristics of the surrounding cells.The 3D configuration determines the cell commitment towardchondrogenic differentiation. However, the application of dynamic

loading, an epigenetical signaling also necessary as a stimulus, canalter the way the 3D configuration determines differentiation,

Table 1

Common tests to measure mechanical properties in hydrogels and macroporous scaffolds before and after cell culture in vitro for cartilage tissue engineering.

Technique Description Strengths Limitations

Extensiometry Two grips hold the material and tensile force isapplied by its movement

– Immediate obtaining of Young’s modulus, yieldstrength and ultimate tensile strength

– Only materials with shape of stripsor rings.

– Only uniaxial strains– Destructive

Compression test

(unconfined)

Two plates compress the material in between – Not limited by geometric shape – The material requires a flat surface

– DestructiveBulge test Air pressure inflates the material through a

windowin the substrate. The displacement canbemeasured using a CCD camera or a laser

– Able to characterize the residual stress, elasticmodulus, and other important parameters suchas yield strength and fracture toughness

– Best suitable for thin films

– Potential leakage of the sample– Difficulty to control applied

pressure– Dissolved air becoming trapped in

the solution– Requires finite

Microindentation(and nano)

A tip indents the material at a single point to apredetermined displacement depth andmeasuring the reaction force required

– Non-destructive– Allows quick– and real time measurements of materials– It can be used to measure localized mechanical– strength at different points on a material

surface

– The election of tip geometry playsan important role in determiningthe materials mechanical strength

Table 2

Common mechanical parameters measured in hydrogels and macroporous scaffolds

before and after cell culture in vitro for cartilage tissue engineering.

Parameter Physical meaning Formula

Young modulusor elasticmodulus (E )

Stiffness, as ratio of the stress along anaxis over the strain along that axis inthe range of stress in which Hooke’slaw is valid

E ¼ re

Poisson ratio (m) Negative ratio of transverse to axialstrain

m ¼ etranselong

Aggregatemodulus (H a)

Stiffness when fluid has stoppedflowing

H a ¼ E ð1mÞ

½ð1þmÞð12mÞ

Shear modulus(G)

Ratio of the shear stress to shear strain G ¼ sc

Storage and lossmodulus (E 0

and E 00 )

The storage modulus is a measure of the energy stored elastically duringdeformation, and the loss modulus is ameasure of the energy converted to

heat; in cyclical motions of strain andstress

Dynamiccompressivemodulus (E ⁄)

Ratio of stress to strain undervibratory conditions

E ¼ E 0 þ iE 00



9/12

reversing the effect of adherent and non-adherent surroundings,probably because of loading transmission effects. More evidenceabout the interaction between mechanical loading and 3D configu-ration is needed, likewise works comparing the response of undif-ferentiated mesenchymal stem cells to different cell adhesionpatterns in hydrogels and scaffolds. Besides, more complete profilesof the mechanical properties of the constructs formed by scaffolds,cells and cell produced ECM should be obtained, by measuring alsorelevant parameters, typically not considered, such as mechanicalfatigue. The additionally obtained relevant information can helpto understand the mechanical processes of the scaffolds in the kneecartilage loaded environment. A correlation between as muchparameters of mechanical behavior as possible with cell phenotypeand the histological characteristics of new-formed tissue would bea crucial step beyond the state of the art.

Acknowledgments

This work is funded by European Regional Development Fundthrough the ‘‘Programa Operacional Fatores de Competitividade –COMPETE” and by portuguese national funds arranged by FCT

(Foundation for the Science and Technology, Portugal), project ref-erence PEST-C/FIS/UI607/2014. This work was also funded by theSpanish Ministry of Economy and Competitiveness (MINECO)through the project MAT2013-46467-C4-1-R (including the FEDER financial support). CIBER-BBN is an initiative funded by the VINational R&D&i Plan 2008–2011, Iniciativa Ingenio 2010, Con-solider Program, CIBER Actions and financed by the Instituto deSalud Carlos III with assistance from the European Regional Devel-opment Fund.

The authors also thank support from the COST Action MP1206‘‘Electrospun Nano-fibres for bio inspired composite materialsand innovative industrial applications” and MP1301 ‘‘New Genera-tion Biomimetic and Customized Implants for Bone Engineering”.The authors also thank the financial support from the BasqueGovernment Industry Department under the ELKARTEK Program.

Juan Alberto Panadero thanks the FCT for the SFRH/BD/64586/2009 fellowship grant. Senentxu Lanceros-Méndez alsothanks the Diputación de Bizkaia (Spain) for financial supportunder the Bizkaia Talent program.

References

[1] W. Richter, Mesenchymal stem cells and cartilage in situ regeneration, J.Intern. Med. 266 (2009) 390–405.

[2] Z.Y. García-Carvajal, D. Garciadiego-Cázares, C. Parra-Cid, R. Aguilar-Gaytán,C. Velasquillo, C. Ibarra, J.S. Castro Carmona, Cartilage tissue engineering: therole of extracellular matrix (ECM) and novel strategies, in: Jose A. Andrades(Ed.), Regenerative Medicine and Tissue Engineering, 2013.

[3] J.S. Temenoff, A.G. Mikos, Review: tissue engineering for regeneration of articular cartilage, Biomaterials 21 (2000) 431–440.

[4] A.E. Beris, M.G. Lykissas, C.D. Papageorgiou, A.D. Georgoulis, Advances inarticular cartilage repair, Injury 36 (Suppl. 4) (2005) S14–S23 .

[5] C. Chung, J.A. Burdick, Engineering cartilage tissue, Adv. Drug Deliv. Rev. 60(2008) 243–262.

[6] E.M. Darling, K.A. Athanasiou, Articular cartilage bioreactors andbioprocesses, Tissue Eng. 9 (2003) 9–26.

[7] I. Martin, S. Miot, A. Barbero, M. Jakob, D. Wendt, Osteochondral tissueengineering, J. Biomech. 40 (2007) 750–765.

[8] J.F. Mano, R.L. Reis, Osteochondral defects: present situation and tissueengineering approaches, J. Tissue Eng. Regener. Med. 1 (2007) 261–273 .

[9] E.B. Hunziker, Articular cartilage repair: basic science and clinical progress. Areview of the current status and prospects, Osteoarthritis Cartilage/OARS,Osteoarthritis Res. Soc. 10 (2002) 432–463.

[10] A.M. DeLise, E. Stringa, W.A. Woodward, M.A. Mello, R.S. Tuan, Embryoniclimb mesenchyme micromass cultureas an in vitro model for chondrogenesisand cartilage maturation, Methods Mol. Biol. (Clifton, NJ) 137 (2000) 359–375.

[11] Y.S. Lee, N.S. Stott, T.X. Jiang, R.B. Widelitz, C. Chuon, Early events duringprecartilage condensation in limb bud micromass cultures, Cells Mater. 8

(1998) 19–32.

[12] J.L. Muschler, A.F. Horwitz, Down-regulation of the chicken alpha 5 beta 1integrin fibronectin receptor during development, Development (Cambridge,England) 113 (1991) 327–337.

[13] K. Daniels, M. Solursh, Modulation of chondrogenesis by the cytoskeleton andextracellular matrix, J. Cell Sci. 100 (Pt 2) (1991) 249–254.

[14] W. Bi, J.M. Deng, Z. Zhang, R.R. Behringer, B. de Crombrugghe, Sox9 is requiredfor cartilage formation, Nat. Genet. 22 (1999) 85–89.

[15] L.M. Hoffman, A.D. Weston, T.M. Underhill, Molecular mechanisms regulatingchondroblast differentiation, J. Bone Joint Surg. Am. 85-A (Suppl 2) (2003)124–132.

[16] A. Woods, G. Wang, F. Beier, Regulation of chondrocyte differentiation by theactin cytoskeleton and adhesive interactions, J. Cell. Physiol. 213 (2007) 1–8 .

[17] E. Ruoslahti, Proteoglycans in cell regulation, J. Biol. Chem. 264 (1989)13369–13372.

[18] R. Chiquet-Ehrismann, P. Kalla, C.A. Pearson, K. Beck, M. Chiquet, Tenascininterferes with fibronectin action, Cell 53 (1988) 383–390.

[19] F. Beier, Cell-cycle control and the cartilage growth plate, J. Cell. Physiol. 202(2005) 1–8.

[20] B.E. Bobick, F.H. Chen, A.M. Le, R.S. Tuan, Regulation of the chondrogenicphenotypein culture, Birth Defects Res. C: EmbryoToday 87 (2009)351–371.

[21] S.A. Oberlender, R.S. Tuan, Spatiotemporal profile of N-cadherin expression inthe developing limb mesenchyme, Cell Adhes. Commun. 2 (1994) 521–537.

[22] C. Csaki, P.R. Schneider, M. Shakibaei, Mesenchymal stem cells as a potentialpool for cartilage tissue engineering, Ann. Anat. = Anatomischer Anzeiger 190(2008) 395–412.

[23] R.M. Schulz, A. Bader, Cartilage tissue engineering and bioreactor systems forthe cultivation and stimulation of chondrocytes, Eur. Biophys. J. 36 (2007)539–568.

[24] F. Guilak, L.G. Alexopoulos, M.L. Upton, I. Youn, J.B. Choi, L. Cao, L.A. Setton,M.A. Haider, The pericellular matrix as a transducer of biomechanical andbiochemical signals in articular cartilage, Ann. N.Y. Acad. Sci. 1068 (2006)498–512.

[25] C.A. Poole, Review. Articular cartilagechondrons: form, function andfailure, J.Anat. 191 (1997) 1–13.

[26] C.A. Poole, S. Ayad, J.R. Schofield, Chondrons from articular cartilage: I.Immunolocalization of type VI collagen in the pericellular capsule of isolatedcanine tibial chondrons, J. Cell Sci. 90 (Pt 4) (1988) 635–643.

[27] N. Miosge, K. Flachsbart, W. Goetz, W. Schultz, H. Kresse, R. Herken, Light andelectron microscopical immunohistochemical localization of the smallproteoglycan core proteins decorin and biglycan in human knee jointcartilage, Histochem. J. 26 (1994) 939–945.

[28] N.-C. Cheng, B.T. Estes, H.A. Awad, F. Guilak, Chondrogenic differentiation of adipose-derived adult stem cells by a porous scaffold derived from nativearticular cartilage extracellular matrix, Tissue Eng. Part A 15 (2009) 231–241.

[29] C.-C. Chen, C.-H. Liao, Y.-H. Wang, Y.-M. Hsu, S.-H. Huang, C.-H. Chang, H.-W.Fang, Cartilage fragments from osteoarthritic knee promote chondrogenesisof mesenchymal stem cells without exogenous growth factor induction, J.

Orthop. Res. 30 (2012) 393–400.[30] N.C. Cheng, B.T. Estes, T.H. Young, F. Guilak, Genipin-crosslinked cartilage-derived matrix as a scaffold for human adipose-derived stem cellchondrogenesis, Tissue Eng. Part A 19 (2013) 484–496.

[31] S.H. Kim, J. Turnbull, S. Guimond, Extracellular matrix and cell signalling: thedynamic cooperation of integrin, proteoglycan and growth factor receptor, J.Endocrinol. 209 (2011) 139–151.

[32] M. Shakibaei, C. Csaki, A. Mobasheri, Diverse roles of integrin receptors inarticular cartilage, Adv. Anat. Embryol. Cell Biol. 197 (2008) 1–60.

[33] N.C. Zanetti, M. Solursh, Induction of chondrogenesis in limb mesenchymalculturesby disruption ofthe actincytoskeleton, J. Cell Biol. 99(1984)115–123.

[34] C. Farquharson, D. Lester, E. Seawright, D. Jefferies, B. Houston, Microtubulesare potential regulators of growth-plate chondrocyte differentiation andhypertrophy, Bone 25 (1999) 405–412.

[35] E.J. Blain, S.J. Gilbert, A.J. Hayes, V.C. Duance, Disassembly of the vimentincytoskeleton disrupts articular cartilage chondrocyte homeostasis, MatrixBiol. 25 (2006) 398–408.

[36] K. Burridge, K. Fath, T. Kelly, G. Nuckolls, C. Turner, Focal adhesions:transmembrane junctions between the extracellular matrix and the

cytoskeleton, Annu. Rev. Cell Biol. 4 (1988) 487–525.[37] D.E. Ingber, D. Prusty, Z. Sun, H. Betensky, N. Wang, Cell shape, cytoskeletal

mechanics, and cell cycle control in angiogenesis, J. Biomech. 28 (1995)1471–1484.

[38] G. Yourek, M.A. Hussain, J.J. Mao, Cytoskeletal changes of mesenchymal stemcells during differentiation, ASAIO J. (American Society for Artificial InternalOrgans: 1992) 53 (2007) 219–228.

[39] M. Schliwa, Action of cytochalasin D on cytoskeletal networks, J. Cell Biol. 92(1982) 79–91.

[40] T. Takasuka, S. Ishibashi, T. Ide, Expression of cell-cycle-dependent genes inserum stimulated cells whose entry into S phase is blocked by cytochalasin D,Biochim. Biophys. Acta 909 (1987) 161–164.

[41] M.J. Kim, S. Kim, Y. Kim, E.J. Jin, J.K. Sonn, Inhibition of RhoA but not ROCKinduces chondrogenesis of chick limb mesenchymal cells, Biochem. Biophys.Res. Commun. 418 (2012) 500–505.

[42] B. Ma, J.C. Leijten, L. Wu, M. Kip, C.A. van Blitterswijk, J.N. Post, M. Karperien,Gene expression profiling of dedifferentiated human articular chondrocytesin monolayer culture, Osteoarthritis Cartilage/OARS, Osteoarthritis Res. Soc.

21 (2013) 599–603.


http://refhub.elsevier.com/S1742-7061(16)30037-X/h0005http://refhub.elsevier.com/S1742-7061(16)30037-X/h0005http://refhub.elsevier.com/S1742-7061(16)30037-X/h0010http://refhub.elsevier.com/S1742-7061(16)30037-X/h0010http://refhub.elsevier.com/S1742-7061(16)30037-X/h0010http://refhub.elsevier.com/S1742-7061(16)30037-X/h0010http://refhub.elsevier.com/S1742-7061(16)30037-X/h0015http://refhub.elsevier.com/S1742-7061(16)30037-X/h0015http://refhub.elsevier.com/S1742-7061(16)30037-X/h0020http://refhub.elsevier.com/S1742-7061(16)30037-X/h0020http://refhub.elsevier.com/S1742-7061(16)30037-X/h0025http://refhub.elsevier.com/S1742-7061(16)30037-X/h0025http://refhub.elsevier.com/S1742-7061(16)30037-X/h0030http://refhub.elsevier.com/S1742-7061(16)30037-X/h0030http://refhub.elsevier.com/S1742-7061(16)30037-X/h0035http://refhub.elsevier.com/S1742-7061(16)30037-X/h0035http://refhub.elsevier.com/S1742-7061(16)30037-X/h0035http://refhub.elsevier.com/S1742-7061(16)30037-X/h0040http://refhub.elsevier.com/S1742-7061(16)30037-X/h0040http://refhub.elsevier.com/S1742-7061(16)30037-X/h0040http://refhub.elsevier.com/S1742-7061(16)30037-X/h0045http://refhub.elsevier.com/S1742-7061(16)30037-X/h0045http://refhub.elsevier.com/S1742-7061(16)30037-X/h0045http://refhub.elsevier.com/S1742-7061(16)30037-X/h0050http://refhub.elsevier.com/S1742-7061(16)30037-X/h0050http://refhub.elsevier.com/S1742-7061(16)30037-X/h0050http://refhub.elsevier.com/S1742-7061(16)30037-X/h0050http://refhub.elsevier.com/S1742-7061(16)30037-X/h0055http://refhub.elsevier.com/S1742-7061(16)30037-X/h0055http://refhub.elsevier.com/S1742-7061(16)30037-X/h0055http://refhub.elsevier.com/S1742-7061(16)30037-X/h0060http://refhub.elsevier.com/S1742-7061(16)30037-X/h0060http://refhub.elsevier.com/S1742-7061(16)30037-X/h0060http://refhub.elsevier.com/S1742-7061(16)30037-X/h0065http://refhub.elsevier.com/S1742-7061(16)30037-X/h0065http://refhub.elsevier.com/S1742-7061(16)30037-X/h0065http://refhub.elsevier.com/S1742-7061(16)30037-X/h0070http://refhub.elsevier.com/S1742-7061(16)30037-X/h0070http://refhub.elsevier.com/S1742-7061(16)30037-X/h0075http://refhub.elsevier.com/S1742-7061(16)30037-X/h0075http://refhub.elsevier.com/S1742-7061(16)30037-X/h0075http://refhub.elsevier.com/S1742-7061(16)30037-X/h0080http://refhub.elsevier.com/S1742-7061(16)30037-X/h0080http://refhub.elsevier.com/S1742-7061(16)30037-X/h0085http://refhub.elsevier.com/S1742-7061(16)30037-X/h0085http://refhub.elsevier.com/S1742-7061(16)30037-X/h0090http://refhub.elsevier.com/S1742-7061(16)30037-X/h0090http://refhub.elsevier.com/S1742-7061(16)30037-X/h0090http://refhub.elsevier.com/S1742-7061(16)30037-X/h0095http://refhub.elsevier.com/S1742-7061(16)30037-X/h0095http://refhub.elsevier.com/S1742-7061(16)30037-X/h0100http://refhub.elsevier.com/S1742-7061(16)30037-X/h0100http://refhub.elsevier.com/S1742-7061(16)30037-X/h0105http://refhub.elsevier.com/S1742-7061(16)30037-X/h0105http://refhub.elsevier.com/S1742-7061(16)30037-X/h0105http://refhub.elsevier.com/S1742-7061(16)30037-X/h0110http://refhub.elsevier.com/S1742-7061(16)30037-X/h0110http://refhub.elsevier.com/S1742-7061(16)30037-X/h0110http://refhub.elsevier.com/S1742-7061(16)30037-X/h0110http://refhub.elsevier.com/S1742-7061(16)30037-X/h0115http://refhub.elsevier.com/S1742-7061(16)30037-X/h0115http://refhub.elsevier.com/S1742-7061(16)30037-X/h0115http://refhub.elsevier.com/S1742-7061(16)30037-X/h0120http://refhub.elsevier.com/S1742-7061(16)30037-X/h0120http://refhub.elsevier.com/S1742-7061(16)30037-X/h0120http://refhub.elsevier.com/S1742-7061(16)30037-X/h0120http://refhub.elsevier.com/S1742-7061(16)30037-X/h0125http://refhub.elsevier.com/S1742-7061(16)30037-X/h0125http://refhub.elsevier.com/S1742-7061(16)30037-X/h0125http://refhub.elsevier.com/S1742-7061(16)30037-X/h0130http://refhub.elsevier.com/S1742-7061(16)30037-X/h0130http://refhub.elsevier.com/S1742-7061(16)30037-X/h0130http://refhub.elsevier.com/S1742-7061(16)30037-X/h0135http://refhub.elsevier.com/S1742-7061(16)30037-X/h0135http://refhub.elsevier.com/S1742-7061(16)30037-X/h0135http://refhub.elsevier.com/S1742-7061(16)30037-X/h0135http://refhub.elsevier.com/S1742-7061(16)30037-X/h0140http://refhub.elsevier.com/S1742-7061(16)30037-X/h0140http://refhub.elsevier.com/S1742-7061(16)30037-X/h0140http://refhub.elsevier.com/S1742-7061(16)30037-X/h0145http://refhub.elsevier.com/S1742-7061(16)30037-X/h0145http://refhub.elsevier.com/S1742-7061(16)30037-X/h0145http://refhub.elsevier.com/S1742-7061(16)30037-X/h0145http://refhub.elsevier.com/S1742-7061(16)30037-X/h0150http://refhub.elsevier.com/S1742-7061(16)30037-X/h0150http://refhub.elsevier.com/S1742-7061(16)30037-X/h0150http://refhub.elsevier.com/S1742-7061(16)30037-X/h0155http://refhub.elsevier.com/S1742-7061(16)30037-X/h0155http://refhub.elsevier.com/S1742-7061(16)30037-X/h0155http://refhub.elsevier.com/S1742-7061(16)30037-X/h0160http://refhub.elsevier.com/S1742-7061(16)30037-X/h0160http://refhub.elsevier.com/S1742-7061(16)30037-X/h0165http://refhub.elsevier.com/S1742-7061(16)30037-X/h0165http://refhub.elsevier.com/S1742-7061(16)30037-X/h0170http://refhub.elsevier.com/S1742-7061(16)30037-X/h0170http://refhub.elsevier.com/S1742-7061(16)30037-X/h0170http://refhub.elsevier.com/S1742-7061(16)30037-X/h0175http://refhub.elsevier.com/S1742-7061(16)30037-X/h0175http://refhub.elsevier.com/S1742-7061(16)30037-X/h0175http://refhub.elsevier.com/S1742-7061(16)30037-X/h0175http://refhub.elsevier.com/S1742-7061(16)30037-X/h0180http://refhub.elsevier.com/S1742-7061(16)30037-X/h0180http://refhub.elsevier.com/S1742-7061(16)30037-X/h0180http://refhub.elsevier.com/S1742-7061(16)30037-X/h0185http://refhub.elsevier.com/S1742-7061(16)30037-X/h0185http://refhub.elsevier.com/S1742-7061(16)30037-X/h0185http://refhub.elsevier.com/S1742-7061(16)30037-X/h0190http://refhub.elsevier.com/S1742-7061(16)30037-X/h0190http://refhub.elsevier.com/S1742-7061(16)30037-X/h0190http://refhub.elsevier.com/S1742-7061(16)30037-X/h0195http://refhub.elsevier.com/S1742-7061(16)30037-X/h0195http://refhub.elsevier.com/S1742-7061(16)30037-X/h0200http://refhub.elsevier.com/S1742-7061(16)30037-X/h0200http://refhub.elsevier.com/S1742-7061(16)30037-X/h0200http://refhub.elsevier.com/S1742-7061(16)30037-X/h0205http://refhub.elsevier.com/S1742-7061(16)30037-X/h0205http://refhub.elsevier.com/S1742-7061(16)30037-X/h0205http://refhub.elsevier.com/S1742-7061(16)30037-X/h0205http://refhub.elsevier.com/S1742-7061(16)30037-X/h0210http://refhub.elsevier.com/S1742-7061(16)30037-X/h0210http://refhub.elsevier.com/S1742-7061(16)30037-X/h0210http://refhub.elsevier.com/S1742-7061(16)30037-X/h0210http://refhub.elsevier.com/S1742-7061(16)30037-X/h0210http://refhub.elsevier.com/S1742-7061(16)30037-X/h0210http://refhub.elsevier.com/S1742-7061(16)30037-X/h0210http://refhub.elsevier.com/S1742-7061(16)30037-X/h0210http://refhub.elsevier.com/S1742-7061(16)30037-X/h0205http://refhub.elsevier.com/S1742-7061(16)30037-X/h0205http://refhub.elsevier.com/S1742-7061(16)30037-X/h0205http://refhub.elsevier.com/S1742-7061(16)30037-X/h0200http://refhub.elsevier.com/S1742-7061(16)30037-X/h0200http://refhub.elsevier.com/S1742-7061(16)30037-X/h0200http://refhub.elsevier.com/S1742-7061(16)30037-X/h0195http://refhub.elsevier.com/S1742-7061(16)30037-X/h0195http://refhub.elsevier.com/S1742-7061(16)30037-X/h0190http://refhub.elsevier.com/S1742-7061(16)30037-X/h0190http://refhub.elsevier.com/S1742-7061(16)30037-X/h0190http://refhub.elsevier.com/S1742-7061(16)30037-X/h0185http://refhub.elsevier.com/S1742-7061(16)30037-X/h0185http://refhub.elsevier.com/S1742-7061(16)30037-X/h0185http://refhub.elsevier.com/S1742-7061(16)30037-X/h0180http://refhub.elsevier.com/S1742-7061(16)30037-X/h0180http://refhub.elsevier.com/S1742-7061(16)30037-X/h0180http://refhub.elsevier.com/S1742-7061(16)30037-X/h0175http://refhub.elsevier.com/S1742-7061(16)30037-X/h0175http://refhub.elsevier.com/S1742-7061(16)30037-X/h0175http://refhub.elsevier.com/S1742-7061(16)30037-X/h0170http://refhub.elsevier.com/S1742-7061(16)30037-X/h0170http://refhub.elsevier.com/S1742-7061(16)30037-X/h0170http://refhub.elsevier.com/S1742-7061(16)30037-X/h0165http://refhub.elsevier.com/S1742-7061(16)30037-X/h0165http://refhub.elsevier.com/S1742-7061(16)30037-X/h0160http://refhub.elsevier.com/S1742-7061(16)30037-X/h0160http://refhub.elsevier.com/S1742-7061(16)30037-X/h0155http://refhub.elsevier.com/S1742-7061(16)30037-X/h0155http://refhub.elsevier.com/S1742-7061(16)30037-X/h0155http://refhub.elsevier.com/S1742-7061(16)30037-X/h0150http://refhub.elsevier.com/S1742-7061(16)30037-X/h0150http://refhub.elsevier.com/S1742-7061(16)30037-X/h0150http://refhub.elsevier.com/S1742-7061(16)30037-X/h0145http://refhub.elsevier.com/S1742-7061(16)30037-X/h0145http://refhub.elsevier.com/S1742-7061(16)30037-X/h0145http://refhub.elsevier.com/S1742-7061(16)30037-X/h0145http://refhub.elsevier.com/S1742-7061(16)30037-X/h0140http://refhub.elsevier.com/S1742-7061(16)30037-X/h0140http://refhub.elsevier.com/S1742-7061(16)30037-X/h0140http://refhub.elsevier.com/S1742-7061(16)30037-X/h0135http://refhub.elsevier.com/S1742-7061(16)30037-X/h0135http://refhub.elsevier.com/S1742-7061(16)30037-X/h0135http://refhub.elsevier.com/S1742-7061(16)30037-X/h0135http://refhub.elsevier.com/S1742-7061(16)30037-X/h0130http://refhub.elsevier.com/S1742-7061(16)30037-X/h0130http://refhub.elsevier.com/S1742-7061(16)30037-X/h0130http://refhub.elsevier.com/S1742-7061(16)30037-X/h0125http://refhub.elsevier.com/S1742-7061(16)30037-X/h0125http://refhub.elsevier.com/S1742-7061(16)30037-X/h0120http://refhub.elsevier.com/S1742-7061(16)30037-X/h0120http://refhub.elsevier.com/S1742-7061(16)30037-X/h0120http://refhub.elsevier.com/S1742-7061(16)30037-X/h0120http://refhub.elsevier.com/S1742-7061(16)30037-X/h0115http://refhub.elsevier.com/S1742-7061(16)30037-X/h0115http://refhub.elsevier.com/S1742-7061(16)30037-X/h0115http://refhub.elsevier.com/S1742-7061(16)30037-X/h0110http://refhub.elsevier.com/S1742-7061(16)30037-X/h0110http://refhub.elsevier.com/S1742-7061(16)30037-X/h0110http://refhub.elsevier.com/S1742-7061(16)30037-X/h0110http://refhub.elsevier.com/S1742-7061(16)30037-X/h0110http://refhub.elsevier.com/S1742-7061(16)30037-X/h0105http://refhub.elsevier.com/S1742-7061(16)30037-X/h0105http://refhub.elsevier.com/S1742-7061(16)30037-X/h0100http://refhub.elsevier.com/S1742-7061(16)30037-X/h0100http://refhub.elsevier.com/S1742-7061(16)30037-X/h0095http://refhub.elsevier.com/S1742-7061(16)30037-X/h0095http://refhub.elsevier.com/S1742-7061(16)30037-X/h0090http://refhub.elsevier.com/S1742-7061(16)30037-X/h0090http://refhub.elsevier.com/S1742-7061(16)30037-X/h0085http://refhub.elsevier.com/S1742-7061(16)30037-X/h0085http://refhub.elsevier.com/S1742-7061(16)30037-X/h0080http://refhub.elsevier.com/S1742-7061(16)30037-X/h0080http://refhub.elsevier.com/S1742-7061(16)30037-X/h0075http://refhub.elsevier.com/S1742-7061(16)30037-X/h0075http://refhub.elsevier.com/S1742-7061(16)30037-X/h0075http://refhub.elsevier.com/S1742-7061(16)30037-X/h0070http://refhub.elsevier.com/S1742-7061(16)30037-X/h0070http://refhub.elsevier.com/S1742-7061(16)30037-X/h0065http://refhub.elsevier.com/S1742-7061(16)30037-X/h0065http://refhub.elsevier.com/S1742-7061(16)30037-X/h0060http://refhub.elsevier.com/S1742-7061(16)30037-X/h0060http://refhub.elsevier.com/S1742-7061(16)30037-X/h0060http://refhub.elsevier.com/S1742-7061(16)30037-X/h0055http://refhub.elsevier.com/S1742-7061(16)30037-X/h0055http://refhub.elsevier.com/S1742-7061(16)30037-X/h0055http://refhub.elsevier.com/S1742-7061(16)30037-X/h0050http://refhub.elsevier.com/S1742-7061(16)30037-X/h0050http://refhub.elsevier.com/S1742-7061(16)30037-X/h0050http://refhub.elsevier.com/S1742-7061(16)30037-X/h0050http://refhub.elsevier.com/S1742-7061(16)30037-X/h0045http://refhub.elsevier.com/S1742-7061(16)30037-X/h0045http://refhub.elsevier.com/S1742-7061(16)30037-X/h0045http://refhub.elsevier.com/S1742-7061(16)30037-X/h0040http://refhub.elsevier.com/S1742-7061(16)30037-X/h0040http://refhub.elsevier.com/S1742-7061(16)30037-X/h0035http://refhub.elsevier.com/S1742-7061(16)30037-X/h0035http://refhub.elsevier.com/S1742-7061(16)30037-X/h0030http://refhub.elsevier.com/S1742-7061(16)30037-X/h0030http://refhub.elsevier.com/S1742-7061(16)30037-X/h0025http://refhub.elsevier.com/S1742-7061(16)30037-X/h0025http://refhub.elsevier.com/S1742-7061(16)30037-X/h0020http://refhub.elsevier.com/S1742-7061(16)30037-X/h0020http://refhub.elsevier.com/S1742-7061(16)30037-X/h0015http://refhub.elsevier.com/S1742-7061(16)30037-X/h0015http://refhub.elsevier.com/S1742-7061(16)30037-X/h0010http://refhub.elsevier.com/S1742-7061(16)30037-X/h0010http://refhub.elsevier.com/S1742-7061(16)30037-X/h0010http://refhub.elsevier.com/S1742-7061(16)30037-X/h0010http://refhub.elsevier.com/S1742-7061(16)30037-X/h0010http://refhub.elsevier.com/S1742-7061(16)30037-X/h0010http://refhub.elsevier.com/S1742-7061(16)30037-X/h0005http://refhub.elsevier.com/S1742-7061(16)30037-X/h0005http://-/?-


10/12

[43] G. Schulze-Tanzil, Activation and dedifferentiation of chondrocytes:implications in cartilage injury and repair, Ann. Anat. = AnatomischerAnzeiger 191 (2009) 325–338.

[44] N. Huebsch, P.R. Arany, A.S. Mao, D. Shvartsman, O.A. Ali, S.A. Bencherif, J.Rivera-Feliciano, D.J. Mooney, Harnessing traction-mediated manipulation of the cell/matrix interface to control stem-cell fate, Nat. Mater. 9 (2010) 518–526.

[45] B. Johnstone, T.M. Hering, A.I. Caplan, V.M. Goldberg, J.U. Yoo, In vitrochondrogenesis of bone marrow-derived mesenchymal progenitor cells, Exp.Cell Res. 238 (1998) 265–272.

[46] W. Kafienah, S. Mistry, S.C. Dickinson, T.J. Sims, I. Learmonth, A.P. Hollander,Three-dimensional cartilage tissue engineering using adult stem cells fromosteoarthritis patients, Arthritis Rheum. 56 (2007) 177–187.

[47] M.B. Mueller, R.S. Tuan, Functional characterization of hypertrophy inchondrogenesis of human mesenchymal stem cells, Arthritis Rheum. 58(2008) 1377–1388.

[48] A.D. Murdoch, L.M. Grady, M.P. Ablett, T. Katopodi, R.S. Meadows, T.E.Hardingham, Chondrogenic differentiation of human bone marrow stem cellsin transwell cultures: generation of scaffold-free cartilage, Stem Cells(Dayton, Ohio) 25 (2007) 2786–2796.

[49] L. Zhang, P. Su, C. Xu,J. Yang, W. Yu,D. Huang,Chondrogenic differentiation of human mesenchymal stem cells: a comparison between micromass andpellet culture systems, Biotechnol. Lett. 32 (2010) 1339–1346.

[50] G.R. Erickson, J.M. Gimble, D.M. Franklin, H.E. Rice, H. Awad, F. Guilak,Chondrogenic potential of adipose tissue-derived stromal cells in vitro andin vivo, Biochem. Biophys. Res. Commun. 290 (2002) 763–769.

[51] G.M. Hoben, E.J. Koay, K.A. Athanasiou, Fibrochondrogenesis in twoembryonic stem cell lines: effects of differentiation timelines, Stem Cells(Dayton, Ohio) 26 (2008) 422–430.

[52] S.H. McBride, T. Falls, M.L. Knothe Tate, Modulation of stem cell shape andfate B: mechanical modulation of cell shape and gene expression, Tissue Eng.Part A 14 (2008) 1573–1580.

[53] U. Noth, L. Rackwitz, A. Heymer, M. Weber, B. Baumann, A. Steinert, N.Schutze, F. Jakob, J. Eulert, Chondrogenic differentiation of humanmesenchymal stem cells in collagen type I hydrogels, J. Biomed. Mater.Res., Part A 83 (2007) 626–635.

[54] S.M. Vickers, L.S. Squitieri, M. Spector, Effects of cross-linking type II collagen-GAG scaffolds on chondrogenesis in vitro: dynamic pore reduction promotescartilage formation, Tissue Eng. 12 (2006) 1345–1355.

[55] M. Zscharnack, C. Poesel, J. Galle, A. Bader, Low oxygen expansion improvessubsequent chondrogenesis of ovine bone-marrow-derived mesenchymalstem cells in collagen type I hydrogel, Cells Tissues Organs 190 (2009) 81–93.

[56] T.A. Ahmed, E.V. Dare, M. Hincke, Fibrin: a versatile scaffold for tissueengineering applications, Tissue Eng. Part B: Rev. 14 (2008) 199–215 .

[57] D. Eyrich, F. Brandl, B. Appel, H. Wiese, G. Maier, M. Wenzel, R. Staudenmaier,A. Goepferich, T. Blunk, Long-term stable fibrin gels for cartilage engineering,Biomaterials 28 (2007) 55–65.

[58] G.M. Peretti, J.W. Xu, L.J. Bonassar, C.H. Kirchhoff, M.J. Yaremchuk, M.A.Randolph, Review of injectable cartilage engineering using fibrin gel in miceand swine models, Tissue Eng. 12 (2006) 1151–1168.

[59] C. Hrabchak, J. Rouleau, I. Moss, K. Woodhouse, M. Akens, C. Bellingham, F.Keeley, M. Dennis, A. Yee, Assessment of biocompatibility and initialevaluation of genipin cross-linked elastin-like polypeptides in thetreatment of an osteochondral knee defect in rabbits, Acta Biomater. 6(2010) 2108–2115.

[60] C. Chung, J. Mesa, M.A. Randolph, M. Yaremchuk, J.A. Burdick, Influence of gelproperties on neocartilage formation by auricular chondrocytesphotoencapsulated in hyaluronic acid networks, J. Biomed. Mater. Res., PartA 77 (2006) 518–525.

[61] L.H. Nguyen, A.K. Kudva, N.L. Guckert, K.D. Linse, K. Roy, Unique biomaterialcompositions direct bone marrow stem cells into specific chondrocyticphenotypes corresponding to the various zones of articular cartilage,Biomaterials 32 (2011) 1327–1338.

[62] H. Tan, C.R. Chu, K.A. Payne, K.G. Marra, Injectable in situ formingbiodegradable chitosan-hyaluronic acid based hydrogels for cartilage tissueengineering, Biomaterials 30 (2009) 2499–2506.

[63] J.H. Cho, S.H. Kim, K.D. Park, M.C. Jung, W.I. Yang, S.W. Han, J.Y. Noh, J.W. Lee,Chondrogenic differentiation of human mesenchymal stem cells using athermosensitive poly(N-isopropylacrylamide) and water-soluble chitosancopolymer, Biomaterials 25 (2004) 5743–5751.

[64] R. Jin, L.S. Moreira Teixeira, P.J. Dijkstra, M. Karperien, C.A. van Blitterswijk, Z.Y. Zhong, J. Feijen, Injectable chitosan-based hydrogels for cartilage tissueengineering, Biomaterials 30 (2009) 2544–2551.

[65] H.Sa-Lima, S.G.Caridade,J.F. Mano, R.L.Reis,Stimuli-responsive chitosan-starchinjectablehydrogels combinedwith encapsulatedadipose-derivedstromal cellsfor articular cartilage regeneration, Soft Matter 6 (2010) 5184–5195.

[66] J.T. Oliveira, L. Martins, R. Picciochi, P.B. Malafaya, R.A. Sousa, N.M. Neves, J.F.Mano, R.L. Reis, Gellan gum: a new biomaterial for cartilage tissueengineering applications, J. Biomed. Mater. Res., Part A 93 (2010) 852–863 .

[67] S.J. Bryant, K.S. Anseth, D.A. Lee, D.L. Bader, Crosslinking density influencesthe morphology of chondrocytes photoencapsulated in PEG hydrogels duringthe application of compressive strain, J. Orthopaedic Res. 22 (2004) 1143–1149.

[68] S.A. Maher, S.B. Doty, P.A. Torzilli, S. Thornton, A.M. Lowman, J.D. Thomas, R.

Warren, T.M. Wright, E. Myers, Nondegradable hydrogels for the treatment of focal cartilage defects, J. Biomed. Mater. Res., Part A 83 (2007) 145–155.

[69] D. Puppi, F. Chiellini, A.M. Piras, E. Chiellini, Polymeric materials for bone andcartilage repair, Prog. Polym. Sci. 35 (2010) 403–440.

[70] J.C. Antunes, J.M. Oliveira, R.L. Reis, J.M. Soria, J.L. Gomez-Ribelles, J.F. Mano,Novel poly(L -lactic acid)/hyaluronic acid macroporous hybrid scaffolds:characterization and assessment of cytotoxicity, J. Biomed. Mater. Res., PartA 94 (2010) 856–869.

[71] E. López-Ruiz, M. Perán, J. Cobo-Molinos, G. Jiménez, M. Picón, M.Bustamante, F. Arrebola, M.C. Hernández-Lamas, A.D. Delgado-Martínez, E.Montañez, J.A. Marchal, Chondrocytes extract from patients withosteoarthritis induces chondrogenesis in infrapatellar fat pad-derived stem

cells, Osteoarthritis Cartilage 21 (2013) 246–258.[72] N. Garcia-Giralt, R. Izquierdo, X. Nogues, M. Perez-Olmedilla, P. Benito, J.L.

Gomez-Ribelles, M.A. Checa, J. Suay, E. Caceres, J.C. Monllau, A porous PCL scaffold promotes the human chondrocytes redifferentiation and hyaline-specific extracellular matrix protein synthesis, J. Biomed. Mater. Res., Part A85 (2008) 1082–1089.

[73] J.L. Ivirico, M. Salmeron-Sanchez, J.L. Ribelles, M.M. Pradas, J.M. Soria, M.E.Gomes, R.L. Reis, J.F. Mano, Proliferation and differentiation of goat bonemarrow stromal cells in 3D scaffolds with tunable hydrophilicity, J. Biomed.Mater. Res. B Appl. Biomater. 91 (2009) 277–286.

[74] R. Izquierdo, N. Garcia-Giralt, M.T. R

Documents

research article ERP