Biophysical microenvironment and 3D culture physiological relevance R e v ie w s � G E N E T O S C R E E N Drug Discovery Today �Volume 18, Numbers 11/12 �June 2013 REVIEWS Biophysical microenvironment and 3D culture physiological relevance Amish Asthana and William S. Kisaalita Cellular Bioengineering Laboratory, College of Engineering, Driftmier Engineering Center, University of Georgia, Athens, GA 30602, USA Force and substrate physical property (pliability) is one of three well established microenvironmental factors (MEFs) that may contribute to the formation of physiologically more relevant constructs (or not) for cell-based high-throughput screening (HTS) in preclinical drug discovery. In 3D cultures, studies of the physiological relevance dependence on material pliability are inconclusive, raising questions regarding the need to design platforms with materials whose pliability lies within the physiological range. To provide more insight into this question, we examine the factors that may underlie the studies inconclusiveness and suggest the elimination of redundant physical cues, where applicable, to better control other MEFs, make it easier to incorporate 3D cultures into state of the art HTS instrumentation, and reduce screening costs per compound. Conventionally, 3D cell culture simply refers to providing a 3D spatial microenvironment for the cells to grow in. However, in our recent work, the meaning of three-dimensionality has been extended to providing the total microenvironment that supports the formation of microtissue that exhibit ‘complex’ physiological relevance (CPR) or better emulation of the in vivo microtissue functionality in a manner not possible in 2D cultures [1]. A good example of CPR outcome is the formation of bile canaliculi-like structures by HepG2 hepatocyte cells (Fig. 1) is 3D but not in 2D culture formats. The literature has provided guidance that leads to three main categories or microenvironment factors (MEFs) or ‘three-dimensions’ of: (i) chemical or biochemical composition, (ii) spatial (geometric 3D) and temporal dimensions, and (iii) force and substrate physical properties [1–3]. However, as pointed out by Lai et al. [4], because of the lack of a quantifiable entity or biomarkers of three dimensionality, the optimum composition of the microenvironment that is required for the cells to provide a physiologically relevant response has remained elusive. It might be that one microenvironmental factor is more important than the other to emulate in vivo-like functionality or if the cells are provided with some initial cues they might be able to create their own endogenous microenvironment rendering the other exogen- ous factors less important [4]. Corresponding author: Kisaalita, W.S. (williamk@engr.uga.edu) 1359-6446/06/$ - see front matter � 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.drudis. Evidence in support of varying degrees of importance of MEFs comes from the success of the various commercially available 3D cell culture platforms that provide different MEFs that elicit similar functional or structural CPR from the cells. For instance, mam- mary epithelial cells (MCF-10A) grown in Matrigel (BD) have formed acini-like or hollow lumen, a structural element associated with glandular cells in vivo [5]. Similarly, a spheroid culture of MCF-7 cell line has also shown the presence differentiation fea- tures like lumen and budding formation [6]. Interestingly, both the growth platforms, although considered 3D, are very different in nature, with Matrigel providing all the three aforementioned microenvironmental cues to the cells while the spheroid culture just provides a 3D space for them to grow in. This begs the question as to the minimum level of exogenous MEFs that give rise to the in vivo conditions that emulate or produce a structurally and func- tionally analogous in vitro tissue model. It is reasonable to argue that the composition of the microenvironment is not standard, rather the optimum MEFs’ combination that depends on the application (e.g. cell type, relevant CPR, among others). While in the field of regenerative medicine, a precise emulation might be necessary as the construct is meant for implantation in vivo, on the contrary, in the field of drug discovery, exogenous emulation of the microenvironment to the level found in vivo (e.g. for CPR expression) may not be necessary. If this is true, the redundancy in the MEFs of interest may be eliminated which may result in 2012.12.005 www.drugdiscoverytoday.com 533 mailto:williamk@engr.uga.edu http://dx.doi.org/10.1016/j.drudis.2012.12.005 REVIEWS Drug Discovery Today �Volume 18, Numbers 11/12 �June 2013 Blood inflow Basement membrane Tight junction Bile canaliculus M mvG (b) (a) (c) BC bc SC tj tj Hepatocyte Sinusoidal endothelial cell Kupffer cell Stellate cell Drug Discovery Today FIGURE 1 Formation of in vivo-like bile canaliculi structures by HepG2 hepatocytes that is only observed in 3D but not 2D culture formats. (a) Illustrative schematic of liver tissue in vivo. (b) Transmission electron micrographs showing pericanalicular region of two adjoining periportal hepatocytes of adult rat liver tissue – labeled structures are the bile canaliculus (BC), mitochondria (M), Golgi apparatus (G) and tight junctional regions (arrowheads). (c) Almost identical transmission electron micrograph of HepG2 cells cultured on 3D porous polystyrene scaffolds (sc) for 21 days [61], exhibiting tight junction (tj) complexes between adjacent cell; The void formed in-between cells in (c) closely resembles the in vivo bile canaliculus (bc) in (b) and is similarly lined with microvilli (mv). Bar = 500 nm. R e v ie w s � G E N E T O S C R E E N reduction in cost and making it easier to configure for state of the art high-throughput screening (HTS) instrumentation. Herein we illustrate this ‘conjecture’ with a focus on substrate/scaffold plia- bility MEF, one of the three ‘dimensions.’ Cell–substrate interaction – biophysical constraints in 3D platforms Cells exert stress on their matrix during morphogenesis, tissue remodeling, differentiation, and normal physiological activities. The rigidity of the matrix along with the number of receptor- mediated adhesions formed by the cells with the microenviron- ment influence the extent to which the matrix can be contracted by them [7]. This in turn generates intracellular tension which leads to the formation of stress fibers in the cells. If the matrix is rigid, it is more difficult for the cells to contract it, resulting in differential cell functions [2], wherein lies the importance of providing an optimal biophysical microenvironment to the cells in vitro. Most of the 3D platforms that are commercially available (Table 1) can be broadly classified into three categories based on their rigidity: (i) Hydrogel-forming (alginate, agarose, chitosan, fibrin, hyaluronan and collagen to name a few) that are pliable and provide a ‘soft’ environment for the cells to grow in (Type I); (ii) Synthetic microporous (‘spongy’) scaffolds or constructs fabri- cated by freeform technology are generally made of stiffer materi- als having high modulus of elasticity such as polystyrene, PLLA 534 www.drugdiscoverytoday.com (Poly-L-Lactide Acid), PLGA (Poly(lactic-co-glycolic acid)) and PCL (Polycaprolactone) (Type II); (iii) Scaffold-free 3D formats such as spheroids or cellular aggregates, produced with hanging drop or the Rotating Wall Vessel (RWV) configurations that lack cell– exogenous material interaction (Type III). Advantages and disad- vantages of the three platform types are discussed below. Type I The pliability of hydrogels can be altered within the physiological range (100–10,000 Pa) by either changing the concentration of the polymer ((Poly(ethylene glycol))PEG, agarose), the extent of cross- linking or the proportion of the (Extracellular matrix) ECM pro- teins (hyaluronan, collagen, fibronectin, laminin) incorporated in the substrate. Modifying the concentration of the ECM proteins (also biochemical cues) might also lead to a change in the number of adhesion ligand sites present for the cells to bind to. However, this can be managed by mixing a more compliant ECM compo- nent like Matrigel with a stiffer constituent (collagen I) in defined proportions that allow a relatively constant chemical–ligand con- centration within a variable stiffness range. Furthermore, the rigidity of the hydrogels can also be increased by using the hydrogel in an ‘attached’ configuration, where the hydrogels are bound to the bottom of the culture dish and resist the forces that are exerted by the cells, rather than a floating raft or sus- pended mode. An increase in gel rigidity generally results in the Drug Discovery Today �Volume 18, Numbers 11/12 �June 2013 REVIEWS TABLE 1 Elastic moduli for commercially available 3D platforms Company Trade name Type and material Elastic moduli (kPa) Ref. 3DBiomatrix Perfecta3D plates Hanging drops NA Perfecta3D scaffolds Hydrogel NR InSphero GravityPlus plates Hanging drops NA BD Matrigel Laminin, collagen 0.45 [62] Glycosan Biosystems Extracel Hyaluronic acid and collagen 0.011–3.5 [63] GlobalCellSollutions/Hamilton GEM Magnetic alginate microcarrier 0.7% – 0.203 � 0.013 1.5% – 1.3 � 0.129 3.0% – 3.01 � 0.084 [64] Trevigen Cultrex 3D Matrix BME, laminin, collagen 0.45 [62] Sigma HydroMatrix Synthetic peptide hydrogel 1.59–14.7 [65] MaxGel Human ECM 120–380 [66] QGel MT 3D Matrix PEG hydrogel 0.448–5.408 (0–2% PEGDA) [67] Kollodis BioSciences MAPTrix HyGel Chemically defined hydrogel NR Synthecon Inc. BIOFELT PGA, PLLA, PLGA, custom NR Biomerix 3D Scaffold Polycarbonate polyurethane-urea NR Invitrogen Geltrex Laminin, collagen 0.45 [62] AlgiMatrix Alginate 0.7% – 0.203 � 0.013 1.5% – 1.3 � 0.129 3.0% – 3.01 � 0.084 [64] ZellWerk Sponceram Ceramic NR amsbio alvetex Polystyrene 77 [11] 3DM Inc. PuraMatrix Peptide 1.59–14.7 [65] Corning UltraWeb Polyamide 0.725 [68] 3DBiotek 3D Insert PCL Polycaprolactone 500 [69] 3D Insert PS Polystyrene 3,680 (fibrous) [70] 3D Insert PLGA Poly(DL-lactide-co-glycolide) 3,000 (porous) [71] b-TCP Disc b-Tricalcium phosphate 24.6 � 0.95 to 78.6 � 2.36 (�106) [72] MicroTissues Inc. 3D Petri Dish Agarose 3.2% – 294 3.9% – 496 6.7% – 626 [73] Abbreviations: BME: basement membrane extract; NA: not applicable; NR: not reported. R e v ie w s � G E N E T O S C R E E N enhancement of proliferation and inhibition of differentiation because of elevation in the phosphorylation of focal adhesion kinase and the formation of focal adhesions, as shown by Paszek et al. [8]. Even though such platforms provide an in vivo-like pliable environment for the cells to grow in, they lack a defined geometry and fail to impose any physical constraints on the size of the aggregates. As such, the microtissues formed range from being just a cluster of few cells to larger tissues that are above the crucial size for oxygen diffusion and this might generate an adulterated out- come in response to drug exposure [9]. Type II Because the pliability of these scaffolds is above the physiological range, it is usually assumed that they fail to provide the optimum biophysical cues for the cells, yet they have been successful as shown by their commercial adoption (Table 1). However, it should be noted that it is not just the material that affects the pliability but also the form in which it is presented. For example, polystyrene in its bulk state as used in tissue culture plates has a very high elastic modulus (2–4 GPa; [10]) but when used to fabricate salt leached microporous scaffolds, it exhibits a considerably lower modulus (77 kPa; [11]). Also, ECM proteins can be coated on such scaffolds to provide adhesion sites for the cells to attach. The protein coating might provide a more compliant surrounding for the cells, however it has been argued that the coating is very thin and it is a known fact that cells can sense and adapt in response to the topography independent of the adsorbed proteins [12]. The major advantages of type II scaffolds are that they have a defined geo- metry and controllable pore sizes that provide a strict spatial control on the dimension of the microtissues [9] and are better suited for incorporation in HTS state of the art instrumentation. Type III Scaffold-free 3D culture production is either achieved in static conditions such as gravity-enforced hanging drops or dynamic conditions, such as RWV (Rotating Wall Vessel), Stirred Tank Reac- tors (CSTR), spinner vessels, and microfluidic chambers. In the dynamic configuration, shear stress generated because of the fluid flow constitutes the primary driving biophysical factor in absence of stress created because of the cells pulling onto to the material as in the previous cases. Microtissues generated by the above mentioned techniques have been used as in vivo surrogates far and wide [13,14], however, these systems are generally labor intensive and are difficult to adapt in state-of-the-art HTS instrumentation. www.drugdiscoverytoday.com 535 REVIEWS Drug Discovery Today �Volume 18, Numbers 11/12 �June 2013 R e v ie w s � G E N E T O S C R E E N Cellular complex physiological relevance and drug discovery outcomes Moving to second generation 3D cell culture platforms is being driven by the brief that the responses generated by the cells growing in a 3D format are not just ‘different,’ but are physiolo- gically more relevant, when compared to cells cultured on tradi- tional 2D surfaces. It is necessary to conclusively show that these responses produced in 3D formats are emulations of those that are seen in vivo. To be meaningful, these physiologically relevant outcomes (structural or functional) that are also known in vivo should be absent in 2D formats and such outcomes (CPRs) should be established for cells derived from the four major tissue types (epithelial, muscle, connective, and nerve). These outcomes can serve as a standard for determining how close a 3D culture is to its native tissue or which out of a given number of 3D platforms is better suited for a given application. Below, we explore some of the well established or provisional CPRs for cells derived from the three tissue types of most interest in preclinical drug discovery (epithelial, neuronal and cardiac). CPR in liver tissue-derived cells Because liver is the primary organ involved in the metabolism of xenobiotics, 3D constructs of hepatocytes are often used as a tool to screen for drug toxicity and test compounds that may affect liver cell function. As such, because of the vast amount of literature available, both structural and functional CPR responses for liver cells can be established with ease. One of the basic structural phenomena that distinguish liver cells in the native tissue from those cultured in 2D formats is polarity. While those in their natural environment possess structural and functional polarity [15,16], the ones that are isolated and cultured on most flat non- porous surfaces do not [17,18]. In their native conditions, hepa- tocytes maintain a cuboidal shape, with two to three basal surfaces facing the sinusoid. The lateral domain between adjacent cells is divided by a polygonal network of microvilli-lined bile canaliculi (Fig. 1) which is formed by membranes contributed from contig- uous cells and comprises the apical domain of cells. Furthermore, position specific processes are carried out in each domain. For instance, proteins involved in the shuttling metabolites from the blood capillaries are centralized to the basal surface, while those involved in bile acid transport are confined to the apical domain. In monolayer, the development of such bile canalicular networks is sparse, heterogeneous and transient [16,19] further differentiat- ing it from native tissue and substantiating this structural element as a valid CPR outcome. In terms of function, hepatocytes in their native environment display high levels of liver specific activities like high cytochrome P450 activity [20], transferrin secretion [21], albumin production [21,22], tyrosine aminotransferase induction [23] and ureagenesis [24] than their monolayer counterparts. Cells growing in a 2D format entirely lack or have very low levels of many Cytochrome P450 enzymes (CYPs) and transporters found in hepatocytes in vivo [25,26]. The CYPs are a family of phase 1 metabolizing enzymes that consist of 50 isoforms, six of which metabolize 90% of drugs [27]. The primary isoforms in human liver include CYP1A2, CYP2A6, CYP2C9, CYP2C19, CYP2D6, CYP2E1 and CYP3A4, making them a very important family of enzymes not only for screening purposes but also in validating the CPR of 3D liver cell cultures. 536 www.drugdiscoverytoday.com CPR in nerve tissue-derived cells The literature on cells of neuronal origin has not provided a consensus toward establishing a neuronal microtissue CPR, however, several phenomena with high CPR outcome potential are more worthy of further exploration. For example, intracel- lular calcium oscillations are an innate characteristic of neural cells in vivo and have a pivotal role in synaptic signal transmis- sion. Although calcium oscillations have been observed in both 2D and 3D nerve cell cultures, we submit that there should be differences in the nature of oscillations between the two cul- tures. The frequencies of these oscillations found in 3D cultures are considerably lower (e.g. 3.42/600 s, brain slices [28]; 8/600 s, NP cells derived from neurospheres [29]) than those found in 2D cultures (e.g. 60/600 s [30], 60/600 s [31]) and is closer to many in vivo experiences (e.g. 3.6/600 s [32]). Interestingly, the differ- ences in voltage-gated calcium channel (VGCC) function in 2D and 3D cultures might partly be the underlying cause, because these channels are central to the movement of calcium into and out of cells [33]. It has been shown that intracellular calcium transients are significantly lower in 3D as compared with 2D cultures and this is more representative of the in vivo situation [11,13]. Also, L-type VGCC agonists and antagonists have pre- viously been shown to enhance and abolish calcium oscilla- tions, respectively, in nerve cells [34–37], substantiating involvement of VGCC in the differential frequency of calcium oscillations found in different culture formats. These findings suggest potential for calcium transients and/or oscillations to serve as nerve cell CPR outcome. CPR in muscle tissue-derived cells Myocardial microtissues are of particular importance as they find application in HTS Q-T elongation assays, required for screening every drug in development [38]. Some recent exam- ples of drugs that have been withdrawn from various markets because of Q-T interval prolongation effects include Droperidol (Inapsine; Akorn; 2001), Dofetilide (Tikosyn; Pfizer; 2004) and Thiordazine (Mellari; Novartis; 2005). As such, it is essential to establish a consensus for the CPR of cardiac microtissue to provide an in vivo-like tissue model for drug development and screening. A couple of phenomena that are characteristic of the native cardiac tissue in vivo, found in 3D cultures but are lacking in traditional monolayers include beat frequency and contraction force. In a study by Kelm et al. [39], neonatal rat cardiomyocytes (NRC) showed rhythmic contractions at a beat frequency of 60 beats per minute (bpm) when they were grown as spheroids in a hanging drop culture. Similar beat frequencies (43 � 21 bpm) have been seen when NRC were transplanted in vivo in adult rat and formed a microtissue [40] which is considerably lower as compared with 2D culture (83.4 � 4.5 bpm [41], 85.6 � 9.3 [42]). This is consistent with the fact that cells in monolayers exhibit exaggerated responses, specifically VGCC function which is responsible for cardiac cell contraction and has been discussed earlier. However, there are a few studies that have suggested otherwise, such as Zimmer- man et al. [43] who reported the beat frequency of NRC growing as Engineered Heart Tissue (EHT) of 180 bpm. It should be noted that hearts from 2-day-old rats have a beat rate of 135–155 bpm [44]. Drug Discovery Today �Volume 18, Numbers 11/12 �June 2013 REVIEWS R e v ie w s � G E N E T O S C R E E N CPR in HTS 3D platform validation The two examples of liver and muscle cell culture used above, to illustrate the CPR concept, have traditionally been associated with low throughput later phases of discovery applications like toxicol- ogy. A relevant question is how relevant the examples are in early HTS phases of discovery. In response to this question, we submit that as long as the cells can express a target of interest, their 3D assay can be used for both HTS and low throughput discovery applications. For example, there is no reason why HepG2 (human hepatocellular carcinoma) cannot be used in both toxicology (low throughput) and chemotherapeutics HTS. Therefore, CPR should not be thought about only in the context of low throughput applications. As a matter of fact, 3D culture platforms are slowly finding their way into HTS laboratories. For example, studies of HTS assay robustness (in terms of Z0 factor) with 3D cultures in high well-density plates (e.g. 96–384 wells) are beginning to appear [45,46]. A drug exposure might generate a different, and in some cases a ‘better,’ response, such as differentiation toward a particular line- age, from cells growing on a 3D format as compared with those in a monolayer. However, in the absence of a validated CPR outcome, there would be no way of telling if the outcome is predictive of the in vivo response. For example, Tung et al. [47] showed that two cytotoxic drugs had different effects on cells growing as spheroids in hanging drops and those in 2D. Had it been shown that the 3D spheroids produced in this particular platform emulated the native tissue and were validated using a particular CPR outcome, then the effect of the drugs on the spheroids would have been considered more predictive of its in vivo effect. Therefore, it is more desirable TABLE 2 Liver cell studies with different platforms, but similar CPR outcome Cell line and type Scaffold type and material Structural CPR Rat small hepatocytes (SHs) Stacked layers on microporous membranes Bile canaliculi w tight junctions, HHY41 Alginate beads canaliculi with desmosomes a HepG2 and HHY41 Alginate beads Desmosomes, j and canaliculi l Lig-8 cell line (adult rat liver progenitor) Peptide hydrogel (Puramatrix) Primary rat hepatocytes, HepG2 Chitosan-collagen coated PET mesh scaffold microvilli Primary rat hepatocytes Nanofibrous/porous PLLA scaffolds Tight junctions junctions HepG2 Porous PS scaffold Tight junctions Primary rat hepatocytes PMMA or PC polymer scaffold Junctional com Fetal porcine hepatocytes PLLA scaffolds Primary rat astrocytes Spheroids Bile canaliculi, m Abbreviations: PET: Poly(ethylene terephthalate); PC: Propylene carbonate; PMMA: Poly(methy that the existing and the upcoming novel 3D culture platforms are validated with respect to a CPR outcome known in the native tissue and then the results of a drug exposure would be a more physiologically relevant outcome that should lead to a better success rate in the drug development process. Relationship between CPR and biophysical factors – are they independent? Cases where the biophysical MEFs do not seem to influence the CPR outcome have raised the question as to how much importance they command or to what extent they need to be exogenously included in construct design for HTS. Summarized in Table 2 are studies where liver cells were grown in type I, II and III scaffolds that provided vastly different biophysical cues with no difference in cellular function outcomes. Encapsulation of primary human hepatocytes (HHY41) in alginate beads (pliable; type I) has been shown to promote the growth of these cells leading to the forma- tion of 3D aggregates and upregulation of liver specific functions [48]. Cells grown in this particular 3D format have prominent structural CPR features, such as junctional complexes and micro- villi-lined network of canaliculi along with improved secretion of liver specific proteins, cytochrome P450 function, and urea synth- esis. As the pliability of alginate beads is comparable to that of native hepatic tissue (bovine liver: 0.62 � 0.24 kPa and 0.94 � 0.65 kPa by ultrasound and Instron Young’s modulus, respectively [49]) and their resultant biophysical MEF can be considered optimal. Considering this as a standard, a porous polystyrene surface (rigid; 77 kPa) that has an elastic modulus considerably above the physiological range can be considered s Functional CPR Ref. ith luminal microvilli, desmosomes Albumin secretion, tyrosine aminotransferase expression [74] network of microvilli; nd junctional complexes Albumin secretion, fibrinogen, a-1-antitrypsin production, cytochrome P450 1A1 activity, urea synthesis [48] unctional complexes ined with microvilli Albumin secretion, CYP1A1, CYP1A2 cytochrome p450 activity [75] Albumin secretion, CYP1A1, CYP1A2, and CYP2E1 cytochrome p450 activity [76] Albumin secretion [77] , bile canaliculi, gap Glycogen storage, HNF-4 positive, albumin secretion [54] , channels with microvilli Albumin secretion [50,51] plexes, luminal microvilli Albumin secretion, cytochrome P450 activity, tyrosine aminotransferase induction [78] Albumin secretion, cytochrome P450 1A1/2 capacity, ammonia removal, urea synthesis [79] icrovilli [57] l methacrylate); PS: Polystyrene. www.drugdiscoverytoday.com 537 REVIEWS Drug Discovery Today �Volume 18, Numbers 11/12 �June 2013 R e v ie w s � G E N E T O S C R E E N sub-optimal and the liver cells grown on it are expected to produce an aberrant or at least a significantly different outcome. However, studies by Bokhari et al. [50,51] have shown that HepG2 cells grown on porous polystyrene scaffolds (type II) exhibit CPRs similar to those shown by microtissues grown on softer materials like higher viability, structural integrity and formation of bile canaliculi, enhanced liver function and drug response comparable to in vivo activity. One might argue that the type of cells (e.g. primary hepatocytes versus hepatocellular carcinoma) and substrate rigidity might be a factor in screening results as cancerous cells are known to have a higher elastic modulus than the non-malignant phenotype and can easily adapt to a more rigid environment or substrate (e.g. standard polystyrene plates). However, the modulus of HepG2 cells (1.1, 1.6 and 1.4 kPa cultured on Collagen I-, Laminin-, and Matrigel-coated substrates, respectively [52], 2 kPa by micropipette aspiration [53]) is not considerably higher than that of native tissue (bovine liver – 0.62 � 0.24 kPa and 0.94 � 0.65 kPa by ultra- sound and Instron YM, respectively [48]), while the modulus of polystyrene is well beyond the adaptability of the cell, thus ruling out the difference in cell phenotype and respective optimal growth environment elasticity requirement as major factors. This is further substantiated by the fact that primary hepatocytes grown in nanofibrous PLLA scaffolds have also exhibited comparable morphological and functional CPR outcomes [54]. Interestingly, the level of albumin produced by cells cultured on this scaffold for a day was found to be 70 mg/106 cells (10 mg/day/106 cells in 2D) which is considerably higher than that found in alignate beads (55 mg/day/106 cells in 3D, 10 mg/day/106 cells in 2D [48]) and closer to that in vivo (140 mg/day/106 cells in adult normal rat liver [55]). Furthermore, to totally eliminate the effects of substrate pliability and cell–material interactions, spheroid systems that lack any physical scaffolding for cells (type III) can be considered. Such systems, whether employed in static (hanging drops) or dynamic state (RWV [56] or spinner flasks [57]) have yielded CPR outcomes similar to the previous cases. Hence, it can be inferred from these studies that neither the pliability (soft, rigid or scaffold-free) nor the structure (microporous or nanofibrous) of the substrate has a major effect on the CRP phenomena exhibited by the liver cells grown therein. A similar analysis is needed if this is also the case for cells derived from other tissue types. A possible explanation for the above lack of pliability effect comes from the ‘cell-on-cell’ hypothesis [58], where cell-to-cell contacts appear to have a more pivotal role as neighboring cells provide a soft ‘stroma’ for surrounding cells and produce responses similar to those seen when cells are grown in softer gels. For example, prominent actomyosin striation can be seen in myotubes 538 www.drugdiscoverytoday.com when they are cultured on top of a layer of muscle cells. The lower layer of myotubes that adheres more strongly to the rigid substrate shows formation of ample stress fibers, however, myotubes in the upper layers differentiate to a more physiological, striated state with an elastic modulus in a range similar to that of gels suited for differentiation and also the native muscle tissue [58]. Similarly, a basal layer of astrocytes grown on glass provides a pliable envir- onment optimal for branching of neurons, which is similar to gels having brain-like pliability [59]. Also, when endothelial cells seeded at a high density become confluent, they have similar morphologies on both pliable and rigid substrates [60], which differ from cells in a monolayer that are attached only to an underlying rigid surface that cause actin cytoskeletal remodeling and spreading. A similar phenomenon might be responsible for the above discussed liver cell CPR outcomes from cells grown on rigid porous scaffolds (type II/high muduli) where only the out- ermost layer of the microtissue comes in contact and adapts to the pliability of the substrate while the cells in the core grow in the softer environment provided by peripheral cells. However, whether it is only the peripheral layer that is affected by substrate rigidity or a radial gradient in elastic modulus exists within the microtissue is not yet known and requires further investigation. Concluding remarks With the field of 3D cell culture moving into its second generation products, the significance of translational research is increasing. Up until now, much thought and effort has been put towards engineering an optimal 3D microenvironment that can emulate the characteristics of the native milieu. However, adapting state- of-the-art cell-based HTS platforms to accommodate 3D cultures requires a balance between simplistic architecture, control over microenvironmental parameters, structural and functional integ- rity of microtissues and cost effectiveness. Based on the current state of knowledge, it is fair to suggest that in some cases, biophy- sical factors might not be necessary for obtaining CPR outcomes in vitro, which are predictive of treatment/drug efficacy in vivo as the microtissues might be creating their own physical domain (endo- genous ECM) rendering the exogenously incorporated factors less important. Elimination of such redundant physical cues might lead to a better control over important parameters like aggregate size and hypoxia, easier adaptability to automated handling and a reduction in high costs that are currently associated with 3D cell culture platforms. However, more translational research spanning different platforms, constituting a vast array of biophysical factors such as pliability (soft, rigid or scaffold free) and structure (micro- porous or nanofibrous) is required to establish a design philosophy consensus. References 1 Kisaalita, W.S. (2010) 3D Cell-Based Biosensors in Drug Discovery Programs: Microtissue Engineering for High Throughput Screening. CRC Press, Taylor & Francis Group, Boca Raton, FL 2 Griffith, L.G. and Swartz, M.A. (2006) Capturing complex 3D tissue physiology in vitro. Nat. Rev. Mol. Cell Biol. 7, 211–224 3 Green, J.A. and Yamada, K.M. (2007) Three-dimensional microenvironments moderate fibroblast signaling responses. Adv. Drug Deliv. Rev. 59, 1289–1293 4 Lai, Y. et al. (2011) Biomarkers for simplifying HTS 3D cell culture platforms for drug discovery: the case for cytokines. Drug Discov. Today 16, 293–297 5 Debnath, J. and Brugge, J.S. (2005) Modelling glandular epithelial cancers in three- dimensional cultures. Nat. Rev. Cancer 5, 75–688 6 do Amaral, J.B. et al. (2010) Cell death and lumen formation in spheroids of MCF-7 cells. Cell Biol. Int. 34, 267–274 7 Grinnell, F. et al. (1999) Differences in the regulation of fibroblast contraction of floating versus stressed collagen matrices. J. Biol. Chem. 274, 918–923 8 Paszek, M.J. et al. (2005) Tensional homeostasis and the malignant phenotype. Cancer Cell 8, 241–254 Drug Discovery Today �Volume 18, Numbers 11/12 �June 2013 REVIEWS R e v ie w s � G E N E T O S C R E E N 9 Asthana, A. and Kisaalita, W.S. (2012) Microtissue size and hypoxia in HTS with 3D cultures. Drug Discov. Today http://dx.doi.org/10.1016/j.drudis.2012.03.004 10 Chen, A.A. et al. (2009) Modulation of hepatocyte phenotype in vitro via chemomechanical tuning of polyelectrolyte multilayers. Biomaterials 30, 1113–1120 11 Lai, Y. and Kisaalita, W.S. (2012) Performance evaluation of 3D polystyrene 96-well plates with human neural stem cells in a calcium assay. J. Lab. Autom. 17, 284–292 12 Lim, J.Y. and Donahue, H.J. (2007) Cell sensing and response to micro- and nanostructured surfaces produced by chemical and topographic patterning. Tissue Eng. 13, 1879–1891 13 Cheng, K. et al. (2008) Three-dimensional polymer scaffolds for high throughput cell-based assay systems. Biomaterials 29, 2802–2812 14 Lai, Y. et al. (2011) Neural cell 3D microtissue formation is marked by cytokines’ up- regulation. PLoS ONE 6, E26821 15 Hubbard, A.L. et al. (1994) Hepatocyte surface polarity. In The Liver: Biology and Pathobiology (Arias, I.M., Boyer, J.L., Fausto, N., Jakoby, W.B., Schachter, D., Shafritz, D., eds), Raven Press, New York 16 Maurice, M. et al. (1988) Formation of plasma membrane domains in rat hepatocytes and hepatoma cell lines in culture. J. Cell Sci. 90, 79–92 17 Moghe, P.V. et al. (1996) Culture matrix configuration and composition in the maintenance of hepatocytes polarity and function. Biomaterials 17, 373–385 18 LeCluyse, E.L. et al. (1994) Formation of extensive canalicular networks by rat hepatocytes cultured in collagen-sandwich configuration. Am. J. Physiol. 266, C1764–C1774 19 Knop, E. et al. (1995) Ultrastructural and functional differentiation of hepatocytes under long-term culture conditions. Anat. Rec. 242, 337–349 20 Wu, F.J. et al. (1999) Enhanced cytochrome P450 1A1 activity of self-assembled rat hepatocyte spheroids. Cell Transplant. 8, 233–246 21 Tong, et al. (1990) Long-term culture of rat liver cell spheroids in hormonally defined media. Exp. Cell Res. 189, 87–92 22 Koide, N. et al. (1990) Formation of multicellular spheroids composed of adult rat hepatocytes in dishes with positively charged surfaces and under other nonadherent environments. Exp. Cell Res. 186, 27–235 23 Landry, J. et al. (1985) Spheroidal aggregate culture of rat liver cells: histotypic reorganization, biomatrix deposition, and maintenance of functional activities. J. Cell Biol. 101, 914–923 24 Wu, F.J. et al. (1996) Efficient assembly of primary rat hepatocyte spheroids for tissue engineering applications. Biotechnol. Bioeng. 50, 404–415 25 Wilkening, S. et al. (2003) Comparison of primary human hepatocytes and hepatoma cell line HepG2 with regard to their biotransformation properties. Drug Metab. Dispos. 31, 1035–1042 26 Ek, M. et al. (2007) Expression of drug metabolizing enzymes in hepatocyte-like cells. Chem. Pharmacol. 74, 496–503 27 Lynch, T. and Price, A. (2007) The effect of cytochrome P450 metabolism on drug response, interactions and adverse effects. Am. Fam. Phys. 76, 391–396 28 Kawamura, M., Jr and Kawamura, M. (2011) Long-term facilitation of spontaneous calcium oscillations in astrocytes with endogenous adenosine in hippocampal slice cultures. Cell Calcium 49, 249–258 29 Striedinger, K. et al. (2007) Exocytosis of ATP from astrocyte progenitors modulates spontaneous Ca2+ oscillations and cell migration. Glia 55, 652–662 30 Numakawa, T. et al. (2002) Brain-derived neurotrophic factor-induced potentiation of Ca(2+) oscillations in developing cortical neurons. J. Biol. Chem. 277, 6520–6529 31 Hemstapat, K. et al. (2004) Measurement of intracellular Ca2+ in cultured rat embryonic hippocampal neurons using a fluorescence microplate reader: potential application to biomolecular screening. J. Pharmacol. Toxicol. Methods 49, 81–87 32 Kuribara, M. et al. (2010) BDNF stimulates Ca2+ oscillation frequency in melanotrope cells of Xenopus laevis: contribution of IP3-receptor-mediated release of intracellular Ca2+ to gene expression. Gen. Comp. Endocrinol. 169, 123–129 33 Dunn, T.A. and Feller, M.B. (2008) Imaging second messenger dynamics in developing neural circuits. Dev. Neurobiol. 68, 835–844 34 Charles, A.C. et al. (1996) Intercellular calcium waves in neurons. Mol. Cell. Neurosci. 7, 337–353 35 Costantin, J.L. and Charles, A.C. (1999) Spontaneous action potentials initiate rhythmic intercellular calcium waves in immortalized hypothalamic (GT1-1) neurons. J. Neurophysiol. 82, 429–435 36 Tomić, M. et al. (1999) Characterization of plasma membrane calcium oscillator in rat pituitary somatotrophs. J. Biol. Chem. 274, 5693–35702 37 Simeone, K.A. et al. (2011) L-Type calcium channel blockade reduces network activity in human epileptic hypothalamic hematoma tissue. Epilepsia 52, 531–540 38 Fermini, B. and Fossa, A.A. (2003) The impact of drug-induced QT interval prolongation on drug discovery and development. Nat. Rev. Drug Discov. 2, 39–447 39 Kelm, J.M. et al. (2004) Design of artificial myocardial microtissues. Tissue Eng. 10, 201–214 40 Li, R.K. et al. (1996) In vivo survival and function of transplanted rat cardiomyocytes. Circ. Res. 78, 283–288 41 Er, F. et al. (2003) Dominant-negative suppression of HCN channels markedly reduces the native pacemaker current If and undermines spontaneous beating of neonatal cardiomyocytes. Circulation 107, 485–489 42 Michels, G. et al. (2006) Long-term and immediate effect of testosterone on single T- type calcium channel in neonatal rat cardiomyocytes. Endocrinology 147, 5160–5169 43 Zimmermann, W.H. et al. (2000) Three-dimensional engineered heart tissue from neonatal rat cardiac myocytes. Biotechnol. Bioeng. 68, 106–114 44 Wiechert, S. et al. (2003) 24-h Langendorff-perfused neonatal rat heart used to study the impact of adenoviral gene transfer. Am. J. Physiol. Heart Circ. Physiol. 285, H907– H914 45 Hsiao, A.Y. et al. (2012) 384 hanging drop arrays give excellent Z-factors and allow versatile formation of co-culture spheroids. Biotechnol. Bioeng. 109, 1293–1304 46 Lai, Y. and Kisaalita, W. (2012) Performance evaluation of 3D polystyrene 96-well plates with human stem cells and the calcium assay. J. Lab. Autom. 17, 284–292 47 Tung, Y.C. et al. (2011) High-throughput 3D spheroid culture and drug testing using a 384 hanging drop array. Analyst 136, 473–478 48 Selden, C. et al. (1999) Three-dimensional in vitro cell culture leads to a marked upregulation of cell function in human hepatocyte cell lines – an important tool for the development of a bioartificial liver machine. Ann. NY Acad. Sci. 875, 353–363 49 Chen, E. et al. (1996) Young’s modulus measurements of soft tissues with application to elasticity imaging. IEEE Trans. Ultra Ferro. Freq. Control 43, 191–194 50 Bokhari, M. et al. (2007) Culture of HepG2 liver cells on three dimensional polystyrene scaffolds enhances cell structure and function during toxicological challenge. J. Anat. 211, 567–576 51 Bokhari, M. et al. (2007) Novel cell culture device enabling three-dimensional cell growth and improved cell function. Biochem. Biophys. Res. Commun. 354, 1095–1100 52 Sharma, S. et al. (2009) Influence of substrates on hepatocytes: a nanomechanical study. J. Scann. Probe Microsc. 4, 7–16 53 Yin, C. et al. (2003) Adhesion contact dynamics of HepG2 cells on galactose- immobilized substrates. Biomaterials 24, 837–850 54 Bierwolf, J. et al. (2011) Primary rat hepatocyte culture on 3D nanofibrous polymer scaffolds for toxicology and pharmaceutical research. Biotechnol. Bioeng. 108, 141–150 55 Uygun, B.E. et al. (2010) Organ reengineering through development of a transplantable recellularized liver graft using decellularized liver matrix. Nat. Med. 16, 814–820 56 Chang, T.T. and Hughes-Fulford, M. (2008) Monolayer and spheroid culture of human liver hepatocellular carcinoma cell line cells demonstrate distinct global gene expression patterns and functional phenotypes. Tissue Eng. A 14, 1–9 57 Abu-Absi, S.F. et al. (2002) Structural polarity and functional bile canaliculi in rat hepatocyte spheroids. Exp. Cell Res. 274, 56–67 58 Discher, D. et al. (2005) Tissue cells feel and respond to the stiffness of their substrate. Science 310, 1139–1143 59 Georges, P.C. et al. (2006) Matrices with compliance comparable to that of brain tissue select neuronal over glial growth in mixed cortical cultures. Biophys. J. 90, 3012–3018 60 Yeung, T. et al. (2005) Effects of substrate stiffness on cell morphology, cytoskeletal structure, and adhesion. Cell Motil. Cytoskeleton 60, 24–34 61 Crawford, J.M. et al. (1995) Imaging biliary lipid secretion in the rat: ultrastructural evidence for vesiculation of the hepatocyte canalicular membrane. J. Lipid Res. 36, 2147–2163 62 Soofi, S.S. et al. (2009) The elastic modulus of Matrigel as determined by atomic force microscopy. J. Struct. Biol. 167, 216–219 63 Vanderhooft, J.L. et al. (2009) Rheological properties of cross-linked hyaluronan- gelatin hydrogels for tissue engineering. Macromol. Biosci. 9, 20–28 64 West, E.R. (2007) Physical properties of alginate hydrogels and their effects on in vitro follicle development. Biomaterials 28, 4439–4448 65 Leon, E.J. et al. (1998) Mechanical properties of a self-assembling oligopeptide matrix. J. Biomater. Sci. Polym. Ed. 9, 297–312 66 Choi, J.S. et al. (2010) Fabrication of porous extracellular matrix scaffolds from human adipose tissue. Tissue Eng. C 16, 387–396 67 Peyton, S.R. et al. (2006) The use of poly(ethylene glycol) hydrogels to investigate the impact of ECM chemistry and mechanics on smooth muscle cells. Biomaterials 27, 4881–4893 68 Ayres, V.M. et al. (2010) Scanning probe recognition microscopy investigation of neural cell prosthetic properties. Int. J. Nanomanuf. 6, 279–290 69 Olah, L. et al. (2006) Synthesis, structural and mechanical properties of porous polymeric scaffolds for bone tissue regeneration based on neat poly(epison- caprolactone) and its composites with calcium carbonate. Polym. Adv. Technol. 17, 889–897 www.drugdiscoverytoday.com 539 http://dx.doi.org/10.1016/j.drudis.2012.03.004 REVIEWS Drug Discovery Today �Volume 18, Numbers 11/12 �June 2013 R e v ie w s � G E N E T O S C R E E N 70 Baker, S.C. et al. (2006) Characterisation of electrospun polystyrene scaffolds for three-dimensional in vitro biological studies. Biomaterials 27, 3136–3146 71 Zhang, J.C. et al. (2006) Fabrication of three dimensional polymeric scaffolds with spherical pores. J. Mater. Sci. 41, 1725–1731 72 Wang, C.X. et al. (2004) Influence of sintering temperatures on hardness and Young’s modulus of tricalcium phosphate bioceramic by nanoindentation technique. Mater. Charact. 52, 301–307 73 Ripoche, A.C. et al. (2006) Evaluation of nisin diffusion in a polysaccharide gel: influence of agarose and fatty content. Innov. Food Sci. Emerg. Technol. 7, 107–111 74 Sudo, R. et al. (2005) Reconstruction of 3D stacked-up structures by rat small hepatocytes on microporous membranes. FASEB J. 19, 1695–1717 75 Khalil, M. et al. (2001) Human hepatocyte cell lines proliferating as cohesive spheroid colonies in alginate markedly upregulate both synthetic and detoxificatory liver function. J. Hepatol. 297, 68–77 540 www.drugdiscoverytoday.com 76 Semino, C.E. et al. (2003) Functional differentiation of hepatocyte-like spheroid structures from putative liver progenitor cells in three-dimensional peptide scaffolds. Differentiation 71, 262–270 77 Risbud, M.V. et al. (2003) Hydrogel-coated textile scaffolds as candidate in liver tissue engineering. II. Evaluation of spheroid formation and viability of hepatocytes. J. Biomater. Sci. Polym. Ed. 14, 719–731 78 Eschbach, E. et al. (2005) Microstructured scaffolds for liver tissue cultures of high cell density: morphological and biochemical characterization of tissue aggregates. J. Cell Biochem. 95, 243–255 79 Huang, H. et al. (2006) Enhanced functional maturation of fetal porcine hepatocytes in three-dimensional poly-L-lactic acid scaffolds: a culture condition suitable for engineered liver tissues in large-scale animal studies. Cell Transplant. 15, 799–809 Outline placeholder Cell-substrate interaction - biophysical constraints in 3D platforms Type I Type II Type III Cellular complex physiological relevance and drug discovery outcomes CPR in liver tissue-derived cells CPR in nerve tissue-derived cells CPR in muscle tissue-derived cells CPR in HTS 3D platform validation Relationship between CPR and biophysical factors - are they independent? Concluding remarks References