Three-dimensional cell cultivation systems

© 2020 O. M. Sukach et al.; Published by the Institute of Molecular Biology and Genetics, NAS of Ukraine on behalf of Biopolymers and Cell. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited UDC 601.2:576.31:57.085.23

This review discusses the characteristics of three-dimensional cell culture systems on and without carriers (scaffolds). Scaffolds are used to simulate the extracellular matrix, as well as to reproduce the natural physical and structural microenvironment of cells, similar to living tissue. The review examines the types of scaffolds (hard and gel-like, natural and artificial, degradable and non-degradable), their characteristics, advantages and disadvantages, features of cell distribution in them. The use of decellularized and devitalized organs and tissues as scaffolds is discussed. The review also considers matrix-free cultivation of cells in the composition of three-dimensional multicellular structures -spheroids. The structure and biology of spheroids is discussed. The features of spheroid formation under static (self-assembly) and dynamic (under the influence of external forces) cultivation conditions are considered. The role of spheroid size for cell survival is discussed. For many decades, mammalian cell cultures have been an invaluable tool in biology. The prevalent type of cell cultivation systems is a monolayer two-dimensional (2D) cell culture grown on rigid substrates (glass, polystyrene). However, they do not fully reproduce in vivo conditions. Cells in two-dimensional cultures are generally flatter and more elongated compared to cells in vivo. Cell cultivation under 2D conditions is also accompanied by selection of a specific cell phenotype adapted to growth on culture plastics or glass [1,2], cell polarity displacement [3], lack of nutrients and oxygen metabolic gradients, decrease in number of intercellular and cell-matrix interactions.
Unlike 2D monolayer cultures, most of the cells in tissues and organs are in the threedimensional (3D) environment, which consists of cellular and non-cellular components that provide various biophysical, biochemical and mechanical signals and are involved in the regulation of somatic and stem cell functions in vivo [4,5]. The non-cellular component, called the extracellular matrix (ECM), is a three-dimensional gel-like structure formed mainly by structural proteins ( nectin and glycosaminoglycans). These proteins perform both structural and communicative functions, which ensure cellular interaction and maintenance of tissue specificity and homeostasis [6]. ECM also regulates the synthesis and release of soluble biomolecules and growth factors [6], provides the mechanical characteristics inherent in the cellular microenvironment [7,8], and participates in the spatiotemporal control of cell migration [7,8].
Cell interaction , as well as their interaction with ECM, are carried out through cell adhesion molecules located on the surface of the cell membrane and involved in numerous cellular processes, including recognition, adhesion, migration and cell differentiation. They are also responsible for transmitting information from ECM to the cell [9,10]. Depending on the structure and functions, cell adhesion molecules are divided into the immunoglobulin, integrin, cadherin and selectin families [11]. Intercellular interactions are mainly regulated by the cadherin protein family [12]. The proteins of integrin family are responsible for the binding of cells to ECM.
Noteworthy, the cellular microenvironment is a dynamic system that changes during the cells growth and development under the influence of spatio-temporal factors.
Thus the absence of three-dimensional cellular microenvironment results in the changes in morphology of cells [1,2], their polarity [3], division mode, signaling pathways [13,14], gene expression, biochemical processes [15,16] and loss of their phenotype [17,18]. Thereby at present the development of 3D culture methods that are maximally similar to in vivo conditions is an urgent task for researchers.
To fabricate the 3D microenvironment with physical, biological and mechanical characteristics of different types of tissues and cells, researchers use a wide range of materials and technological approach, which have both advantages and disadvantages. However, today there is no general approach for creating threedimensional cell models.
All existing 3D cultivation systems can be divided into 2 large groups: cell cultures on carriers or matrices and cell cultures without carriers.

3D Scaffold Cultivation Systems
Three-dimensional matrices (scaffolds) are used to mimic ECM, as well as to reproduce the natural physical and structural microenvironment of cells, similar to living tissue [19]. Wherein, ideal scaffolds should provide attachment and migration of the cells, retain the biochemical factors, provide the oxygen and carbon dioxide diffusion as well as diffusion of nutrients and products expressed in cells, and exert specific mechanical and biological effects on cells. Scaffolds must also be characterized by high porosity and have pore sizes, ensuring efficient colonization and migration of the cells throughout their structure. A pattern to strive for is a scaffold porosity of 90 % [20]. However, too high porosity can significantly impair the mechanical properties of the scaffold. The optimal pore size for different cells varies significantly [21]. For example, pores of 200-400 μm are effective for bone tissue formation [22], and pores of 50-200 μm are suitable for the growth of smooth muscle cells [23]. As a rule, the pores larger than 100 microns provide cell growth and scaffold vascularization. Too large pores (> 400 microns) reduce the number of intercellular contacts, since at this location the cells are closer to the conditions of 2D culture [24].
Depending on the physical properties and origin, scaffolds are divided into solid and gellike, as well as natural and artificial. Depending on the application [, the] scaffolds could be divided into degradable or non-degradable. The materials used for the manufacture of scaffolds must be non-toxic, biocompatible, non-immunogenic and thromboresistant.
Solid scaffolds are used in reconstructive medicine and preclinical in vitro testing of pharmaceutical preparations. In the first case, the cells are grown on carriers for possible in vivo transplantation to replace degenerative or altered tissue (bone, cartilage, ligaments, skin, blood vessels, and muscles) [25][26][27]. In the second case the scaffolds are used for modeling tumors or tissues in laboratory conditions [28].
Solid scaffolds are fabricated from a wide range of materials, including metals, ceramics, glass and polymers [29][30][31][32]. Depending on the structure, the solid scaffolds can be classified as networks, fibers, sponges, foams, etc. Such structures support a uniform spatial distribution of cells, their growth, diffusion of nutrients and metabolic products. For the production of solid scaffolds of various sizes, structures, stiffness, porosity and permeability, the polymers (Polyglycolide, Poly (ε-caprolactone), Poly (ethylene oxide), Polybutylene terephthalate, Poly (L-lactic acid)) and their derivatives are most often used [32]. The manufacturing process for solid polymer scaffolds depends on the bulk and surface properties of the material and their intended use. Most production methods include applying heat and/or pressure to the polymer or its dissolving in an organic solvent to impart the material of a required shape. For the manufacture of scaffolds, the solution casting [33], leaching [34], electromolding [35] and 3D printing [36] technologies are used.
When fabricating scaffolds, it is considered that the scale and topography of their internal structures are an important factor for threedimensional cell culture. In the organism, the ECM is a complex nanoscale structure involved in the control of cell behavior [37,38]. When bound to a scaffold, the cells usually spread and expand as if they were cultivated on a flat surface [39]. Therefore, even small changes in the nanoscale scaffold topography can have a significant effect on the cell behavior [40]. Cell attachment, growth and behavior are influenced not only by the scaffold microscale architecture and structure. The stiffness, permeability and mechanical properties of the scaffold, as well as the chemical properties of its surface, have a significant effect on these cell parameters [41]. Therefore, in order to improve cell adhesion, the surface of solid scaffolds is often modified by covering it with peptides (arginine-glycine-aspartic acid) [42], gelatin [43], and plasma proteins [44].
To colonize solid scaffolds with cells, two approaches are most often used: static and dynamic [45,46]. The most common methods of static colonization are surface seeding [47,48] and direct insertion of cells into the scaffold [49,50]. Dynamic methods of seeding include forced filling of the scaffold with cells (by passing a solution with cells under pressure through the scaffold), which provides better cell penetration into the scaffold as well as their subsequent better growth compared to static colonization [45,46].
To assess the effectiveness of scaffold colonization, the attached cells are trypsinized, collected and analyzed using the MTT test [51], or cell lysate DNA analysis [52]. Unfortunately, none of these methods can provide complete information about the extent of cell expansion and the efficiency of their penetration into scaffolds.
The main disadvantages of using solid scaffolds in cell biology are the limited visualization of cells in matrices and the difficulties in extracting cells from a solid matrix.
Gel-like scaffolds or hydrogels are hydrophilic polymer networks capable of swelling but not dissolving in water, which makes them similar to the soft tissues of the body and a perfect kind of materials for tissue engineering [53][54][55]. Hydrogels are also highly permeable to oxygen, nutrients and metabolites [54][55][56].
Hydrogels are classified according to their ionic charge (neutral, cationic, anionic and ampholytic), structure (amorphous, semi-crystalline), and fabrication methods (homopolymer, copolymer, multipolymer) [54,57]. By the origin of the polymer, hydrogels are divided into natural, synthetic, and hybrid (synthetic/natural) [53]. Depending on the mechanism of bond formation between the polymer chains, hydrogels could be divided into physical (hydrogen or hydrophobic bonds) and chemical (covalent bonds) [54,58,59]. The type and degree of chains crosslinking affect the hydrogels swelling, elastic modulus, permeability and stiffness [57].
Natural hydrogels are biologically active, compatible and biodegradable [60]. Due to the presence of various endogenous factors, contributing to the maintenance of viability, proliferation and differentiation of many types of cells, natural hydrogels are capable of stimulating many cellular functions [61].
Protein-based hydrogels can be obtained by thermal exposure to proteins, as well as by the use of chemical crosslinking agents. Thus, Matrigel, being liquid at 4 °C, turns into a gel at 37 °C.
Polysaccharide hydrogels can be produced by covalent crosslinking, esterification and polymerization. The polysaccharides can also be combined with proteins [77,78].
DNA is able to form hydrogel networks through self-assembly, electrostatic interaction, chemical crosslinking, or enzymatic ligation [74,75]. The properties of these hydrogels depend on the initial concentrations and types of DNA monomers [79].
The disadvantages of natural hydrogels are their potential immunogenicity, probability of disease transmission, inconstancy of composition and relatively poor mechanical properties [1,2,18].
Synthetic hydrogels are fabricated completely from synthetic molecules such as polyethylene glycol (Poly (ethylene glycol), poly-vinyl alcohol (Polyvinyl alcohol), and poly-2-hydroxyethyl methacrylate, polylactic acid (Polylactide) [81][82][83][84]. They are biologically inert, but provide structural support for various types of cells. Compared to natural ones, synthetic hydrogels have more reproducible physical and chemical properties, which is crucial for tissue engineering. The synthetic polymers used in the production of hydrogels can be divided into non-biodegradable [85] and biodegradable [86].
Biodegradable synthetic hydrogels are used to make vascular structures or soft tissues, whereas non-biodegradable ones are used to construct bones and cartilages.
Although synthetic hydrogels can maintain the viability of encapsulated cells based on feasibility of ECM formation [87], most of them usually function only as passive scaffolds that do not promote active cellular interactions [88]. To eliminate this drawback, biologically active molecules and proteins (promoting cell adhesion, migration, proliferation, and differentiation) are included in the composition of synthetic hydrogel networks [58,89].
In hybrid hydrogels obtained by combining synthetic hydrogels with natural polymers, the synthetic unit provides customizable physical properties, and the natural one provides specific biological functions.
To colonize hydrogels, the cells are either mixed with the initial components of a liquid scaffold before its formation [3,90], or added to previously formed scaffolds.
After colonization, the cells can reconstruct their microenvironment by producing signaling molecules and ECM molecules. The cells are also capable of migration, proliferation, and differentiation. The ideal final result is the formation of stable homeostatic state of the cells, similar to intact tissue.
In contrast to solid scaffolds, the use of hydrogels allows the formation of multilayer tissue-like structures. Thus, for example, certain types of cells are embedded in separate hydrogel constructions, which are then superimposed on each other, thus forming layers similar to tissues in vivo [91].
In addition to artificially manufactured (scaffolds from natural or synthetic components), researchers use decellularized and devitalized organs and tissues as scaffolds [92]. In this type of natural tissue scaffolds, after removal of all cellular components that can cause an inflammatory reaction, the extant ECM retains its composition, architecture, integrity, biomechanical properties, biological activity, hemocompatibility as well as the ability to control cell migration, tissue-specific gene expression and cell fate. The decellularized material can maintain the integrity of the entire organ or its part, or can be subjected to further enzymatic treatment in order to transfer it to a liquid with subsequent formation of ECM-containing hydrogel. After decellularization and certain processing, these natural 3D scaffolds can be functionalized by repeated recellularization with the specific stem or somatic cells together with the necessary growth factors [92]. The sources for creating such natural scaffolds can be the organs and tissues of both humans and animals.

Matrix-Free 3D Cultivation Methods
In addition to the 3D cultivation systems using matrices (scaffolds), the matrix-free cell cultivation as a part of three-dimensional multi-cellular structures (aggregates and spheroids) has become widespread.
The ability to form three-dimensional multicellular structures is based on the tendency of cells with three equivalent degrees of freedom towards aggregation in the absence of substrates, and on the subsequent active formation of intercellular contacts and (micro)structures with a minimum surface/volume ratio, i.e. spheroids. Spheroids are characterized by diffusion of oxygen, nutrients and metabolic products inherent to living tissues [93]. In spheroids, the microenvironment signals are restored simultaneously with a decrease in metabolic rates and reduction in the consumption of ATP and oxygen [94]. This is associated with an increase in cell survival in the composition of spheroids. In addition to an increase in the survival time of functionally active cells, the configuration of spheroids (due to the minimum surface to volume ratio) allows one to achieve cell density per unit volume comparable with cell density in organs [95]. Moreover, many cells within the spheroid take the most energy-efficient spherical shape [96], which, according to some researchers, can induce the reexpression of earlier genes, as well as the activation of cell regeneration genes [97].
The advantage of spheroids over more complex 3D scaffold-based matrix systems is the simpleness of their analysis by visualization using light, fluorescence and confocal microscopy.
Spheroid Formation Methods. Almost all methods for generating spheroids include 3 main stages: formation of loose cell aggregates; induction of cadherin expression, during which cell aggregates condense; compaction of aggregates resulting in formation of spheroids. The formation of loose aggregates, as a rule, occurs during cultivation at high cell density (0.5-4×10 6 cells/ml) under conditions that prevent or impede the attachment of cells to the substrate. In this case, the formation of loose aggregates occurs with the participation of both cadherins and integrins. The intercellular contacts formed during the aggregation process lead to an increase of cadherin expression, which accumulates on the membrane surface. Further, due to the intensification of cadherin mediated intercellular interactions, a morphological transition occurs from loose cell aggregates to compact spheroids [98].
The methods for the formation of spheroids can be divided into static -cell self-assembly, and dynamic -assembly of cells based on forced collisions (Table). Self-assembly is a process that takes place in a static environment in which cells cannot attach to the surface and thus come into contact with each other, forming aggregates. The formation of spheroids during self-assembly can occur during cell cultivation on adhesive [99] and non-adhesive [100][101][102][103] surfaces, as well as in the absence of an attachment surface (hanging drop culture, [104,105] emulsion technologies [106]).
In dynamic conditions, the formation of spheroids occurs under the influence of external forces [107][108][109][110][111], when, preventing the attachment of cells to the substrate, their collision and subsequent adhesion are initiated. Gravitational and centrifugal forces [107,108,110,112], magnetic [109] and electric [113] fields, as well as acoustic waves [111] (Table) can be used as external forces. At the same time, as previously noted, a prerequisite for spheroid formation, regardless of the methods used, is cultivation at high cell density (0.5-4×10 6 cells/ml).
All methods for the formation of spheroids have both advantages and disadvantages (Table). Among the disadvantages, it should be noted the difficulty of obtaining a sufficiently large number of spheroids (standard in size and shape) during short time, at low cost of labor and means.
The methods of self-assembly on adhesive and non-adhesive surfaces are simple, inexpensive, well reproducible, allowing the production of a large number of spheroids. However, they are characterized by high variability of spheroids' shape and size and are applicable only to certain types of cells.
Using the "hanging drop" method and emulsion technologies allows obtaining of standard size spheroids in a controlled microenvironment, but at the same time these methods are highly labor intensive, require special equipment, and the forming spheroids are characterized by small sizes. The "hanging drop" method also does not allow obtaining a sufficiently large number of spheroids [104,105].
The use of various types of rotation and rocking are simple and high productivity methods that provide good conditions for delivery of nutrients and oxygen, removal of waste products, possibility of long-term cultivation in a controlled microenvironment [107,108]. However, these methods do not allow obtaining spheroids of standard size and shape. A significant part of the spheroids is characterized by large sizes, which is accompanied by the death of a part of the cells.
The use of magnetic and electric fields, as well as acoustic waves makes it possible to obtain spheroids of a controlled size and shape from various types of cells in a short period of time [109,111,113]. However, the cell aggregates may lose their integrity after the removal of "force". Additionally, the influence of "external force" can lead to cell damage.
An important parameter of the formed spheroids is their size, which, depending on the production methods, can vary over a wide range -from 50 to 1000 μm. It was found that cell viability decreases with an increase of the spheroid size [114]. This is explained by the fact that an increase of spheroid size is accompanied by formation of gas and substance distribution gradients [115,116]. This leads to both the impaired diffusion of oxygen and nutrients to the central cells, and the accumulation of carbon dioxide and waste products in them. Such processes occurring in large spheroids lead to the formation of several cell layers: a necrosis zone, located in the center and consisting of dead cells; a zone of living cells in a resting state; and an outer zone consisting of viable, metabolically active cells [117]. Thus, there is a necessity to determine the sizes of spheroids that ensure maximum cell viability and productivity. Numerous studies have shown that the spheroids with a diameter up to 100 μm are characterized by a low risk of developing hypoxic conditions [115,116,118]. In contrast, the spheroids with a diameter exceeding 200 μm are at an increased risk of oxygen deficiency in the central zone, leading to cell death [117,118].
The dynamic cultivation conditions contribute to an increase of the diffusion rates of oxygen, nutrients and metabolites, which allows enlarging the size of spheroids consisting entirely of viable cells. In order to ensure continuous supply of oxygen to the spheroids during cultivation, special gas-permeable chips have also been developed [119]. They provide an opportunity to increase the size of spheroids consisting of viable cells up to 600 μm.

Conclusion
Thus, the development of three-dimensional cultivation systems has become an important step towards the creation of cell models that, by their characteristics, are closer to intact tissues in vivo. However, unlike 2D cultivation systems, there is still no standard approach for 3D cultivation systems. Noteworthy, the de-velopment of such a universal approach is challenging because of the complexity of ECM composition and structure, the physiological characteristics of different types of cells, the specificity of mechanical properties, biochemical signals, intercellular and cell-matrix connections in different tissues.
The matrix 3D cultivation systems allow the development of structures that, due to their mechanical properties, porosity and biological activity, mimic living tissues in vivo. Nevertheless, biochemical signals and, mainly intercellular and cell-matrix bonds in these structures can differ significantly from natural ones. Moreover, the cell cultivation in scaf- Emulsion technology [106] Dynamic Action of external forces Rotating cell culture system [107,108] -simplicity of the method -high productivity -long-term cultivation -control of microenvironment -MCS shape and size variability -MCS damage -lack of cell-matrix interaction Rocked cell culture system [110] Magnetic field [109] -MCS size and shape control -applicability to various types of cells for MCS formation -co-cultivation of various types of cells for MCS formation -fast MCS formation -loss of MCS integrity after the removal of "power" -cell damage due to external force -lack of cell-matrix interaction Electric field [113] Acoustic waves [111] folds can often be characterized as 2D in 3D, since the cells in three-dimensional surface of the scaffolds attach, spread and migrate in the same way as the cells cultured under adherent two-dimensional conditions. In multicellular spheroids, the intercellular contacts similar to tissues are reproduced as well as the gradients of oxygen, carbon dioxide, nutrients and metabolic products. However, in spheroids, as a rule, there is a lack of cellmatrix interaction. It is not completely clear which cells assemble into spheroids and what is the degree of conformity between the structure and composition of ECM formed in spheroids and the natural one. In spheroids, as well as in matrix cultures, the cellular signaling pathways differ from signaling pathways in tissues in vivo, which also causes differences in the behavior of cells in spheroids compared with the behavior of living tissues cells. The optimal size of the spheroids, on which the metabolism, gene expression, and stem characteristics of cells depend, is also indeterminate. The influence of spheroid formation methods on the state and properties of cells is inexplicit. Thus, for example, the spontaneously formed spheroids probably consist of comparably homogeneous population of cells capable of aggregation, whereas the forcedly generated spheroids apparently contain different types of cells.
Currently, the efforts of numerous researchers are aimed at resolving these complex issues. New 3D cultivation systems and new approaches are being developed, including the combining of matrix cultivation systems with matrixless ones, as well as the bioprinting using spheroids, natural and artificial matrices. Organic solvent free preparation of porous scaffolds вості утворення сфероїдів в статичних (самозбірка) і динамічних (під впливом зовнішніх сил) умовах культивування. Обговорюється роль розміру сфероїдів для виживання клітин.