Printing and Prototyping of Tissues and Scaffolds

See allHide authors and affiliations

Science  16 Nov 2012:
Vol. 338, Issue 6109, pp. 921-926
DOI: 10.1126/science.1226340


New manufacturing technologies under the banner of rapid prototyping enable the fabrication of structures close in architecture to biological tissue. In their simplest form, these technologies allow the manufacture of scaffolds upon which cells can grow for later implantation into the body. A more exciting prospect is the printing and patterning in three dimensions of all the components that make up a tissue (cells and matrix materials) to generate structures analogous to tissues; this has been termed bioprinting. Such techniques have opened new areas of research in tissue engineering and regenerative medicine.

The development of implantable medical devices and organ transplantation has radically changed the scope of medical intervention to deal with chronic medical problems and potential end-of-life situations. However, issues of device failure, the limited supply of donor organs, and the need for immunosuppressants are ongoing problems. The broad areas of tissue engineering and regenerative medicine seek to avoid some of these problems though the augmentation of the healing process or the fabrication of autologous tissue grafts that will not cause an immune response in the host. It has been suggested that we will be able to build complete organs from scratch, and the term “organ printing” has been coined and used by the popular press. Although such a goal is unlikely to be attainable in the near future, printing and related technologies are of current use in the areas of tissue engineering and regenerative medicine.

An important concept in tissue engineering is the scaffold, a three-dimensional (3D), highly porous substrate. Cells donated by the patient are expanded in culture and are then transferred to the scaffold. The scaffold provides a surface on which cells adhere, thrive, multiply, and generate the extracellular matrix (ECM) of structural and functional proteins and saccharides that make up living tissue. Both the scaffold material composition and its internal architecture (dimensions of the struts, walls, pores, or channels) control the behavior and well-being of the cells seeded inside.

Within the body, cells must be very close to the capillary network that supplies oxygen and nutrients. Similarly, within a tissue scaffold, all cells must be supplied with the means to maintain life, and this is achieved initially by providing a highly porous open structure to allow the uninterrupted flow and access of culture media in a bioreactor. In many ways, the engineering component of tissue engineering lies in the design and manufacture of the scaffold. Hutmacher (1) gave an early analysis of the topological requirements of a suitable scaffold material and reviewed the manufacturing routes that could be used to achieve the desired structure. Traditionally, porous scaffolds have been made by a number of routes that result in a foam-like internal structure with a random architecture and a limited control of scale. However, the development of rapid prototyping techniques since the 1980s has enabled fabrication of fine-scale internal porous structures with the desired complexity, allowing a true engineering of the scaffold (1, 2).

Rapid prototyping methods, sometimes referred to as 3D manufacturing or solid freeform manufacture, produce complex objects from a 3D design file by decomposing the object’s shape into a series of parallel slices. The shape is then fabricated by reproducing these slices a layer at a time, building up the structure. The philosophy of the method is to create objects by adding material layer-by-layer; hence, it is now generally referred to as additive manufacture (AM) to distinguish it from conventional machining, which removes material in a subtractive manner. With design files now held in digital format as standard, it is relatively easy to convert an arbitrary object into the slices required by AM fabrication tools. These slice design files are used to generate solid layers using a range of manufacturing techniques, including selective polymerization, selective sintering/melting, building solids through the laying down of viscous threads, and 3D inkjet printing.

These techniques have been used to manufacture laboratory components (3) or cell microwells (4) to order. The hardware cost for these fabrication tools has also decreased, and open-source equipment is now available from a number of groups (5, 6). Other workers have “hacked” commercial desktop printers to fabricate 2D and 3D objects from engineering and biological materials (710). These techniques and their applications in tissue engineering have been reviewed extensively (1, 11). Early scaffolds were fabricated by AM from a single biocompatible material. It is now possible to engineer materials that contain biomimetic components to control the cellular environment (12, 13). The scaffold should be a temporary feature and disappear, through dissolution or degradation, as the cells produce the ECM that defines the tissue. The materials that make up the scaffold must be designed to fulfill this function and be compatible with the fabrication route.

Another approach to tissue engineering has been proposed that does not require a solid scaffold structure. Some workers have championed what has been identified as “bottom-up” or “scaffold-free” tissue engineering (14), which starts from the belief that tissue is a cell-generated material, and thus the direct manipulation and control of cells is of greater importance than the provision of structural, mechanical, and chemical cues via the intermediary of a scaffold. In such an approach, cells are formed into clusters (15), aggregates (16, 17), or 2D sheets (18). These are then manipulated or positioned into 3D cellular constructs by using clusters as “bricks” or sheets as laminates. This methodology has a number of potential advantages over scaffolds:

• Because no scaffold is present, there are no problems of material or degradation product compatibility.

• Cells are cultured in conditions more similar to the 3D environment of the body, allowing better intercellular communication.

• Clusters made of cells that serve complex functions in organs and are very sensitive to the local environment (e.g., hepatocytes) are less likely to redifferentiate and lose function (15).

However, there are other aspects of tissue engineering where the presence of a scaffold may be beneficial:

• Scaffolds provide mechanical strength while ECM is being generated.

• Assembly of large cellular constructs from sheets or clusters is limited by transport of oxygen, nutrients, and waste products until a vascular system develops.

Thus, a hybrid approach intermediate to the scaffold-directed approach (seeded cells) and scaffold-free approach (organized cellular subcomponents) is likely to be developed. This is certainly the case where complex organ-like structures are required; in these instances, the strategy of “let the cells decide” used in the scaffold-free approach will probably need to be assisted by the use of biochemical cues to direct the desired outcome.

The concept of bioprinting, which is essentially an extension of the philosophy that uses AM methods to build complex scaffold structures, can be thought of as a combination of (i) different types of cells in defined locations, (ii) supporting matrix or scaffolds (if required), and (iii) biochemical cues to control behavior. A co-deposition strategy involving chosen combinations of these three elements can be achieved with the use of printing or AM tools to define their 3D arrangement—either topologically or by providing channels for nutrient access (19). Although bioprinting has its origins in the area of tissue engineering and is sometimes described as organ printing (10), there are other application areas where a printed or artificially fabricated tissue analog structure is useful, including cell-based sensors (20), drug/toxicity screening (21), and tissue or tumor models (22).

A key feature of bioprinting is that the deposition process must be cytocompatible, as it requires the dispensing of cell-containing media. This reduces the range of AM techniques that are feasible because of the need to work in an aqueous or aqueous-gel environment at temperatures from room temperature to 38°C. Hence, most bioprinting work reported in the literature has used filament extrusion, inkjet printing, or laser forward transfer (Fig. 1).

Fig. 1

In additive manufacture (AM), a rapid prototyping technique used in bioprinting applications, solids are produced through the sequential deposition of solid layers or slices. (A and B) Inkjet printing accurately positions drops, which may contain cells in suspension, in the volume range 1 to 100 pl. Drops are generated either by heating and vaporizing the ink so that a bubble forces out a liquid drop (A) or by mechanical actuation (B). (C) Microextrusion or filament plotting deposits a continuous thread of material to build up a layer. (D) With laser forward transfer, material is initially deposited on a transparent ribbon in gel form. A pulsed laser beam vaporizes a small quantity of material; this ejects the remainder of the material illuminated, which is transferred to the substrate across a small air gap. [Panels A and B reproduced with permission from (24)]

Controlling 3D Architectures

In manufacturing engineering, AM frees the designer from the constraints of molds and dies, thereby drastically reducing the cost of producing a single, one-off design. It also allows the exploration of design variables. In tissue engineering, this enables the easy variation of the internal architecture of a scaffold, allowing the manufacture of a range of structures to allow optimization, or libraries of structures that can be used in rapid-throughput screening for a given application. Melchels et al. (23) used stereolithography (a well-established AM method) to produce a range of scaffold designs with different internal architectures and porosity (Fig. 2) from a biocompatible and degradable polymer. By altering the internal architecture, it is possible to vary the specific surface area, scaffold mechanical response to external stress, and flow resistance in a bioreactor. Thus, variations in complex internal scaffold architecture can be easily explored as part of an experimental study. Another advantage of architectural control of a porous structure is the ability to better select and control mechanical properties such as stiffness and strength. Conventional tissue scaffolds are fabricated through the foaming of fluid precursors or leaching of pore-generating particles around which a scaffold has been formed (1); this produces a fairly random structure when compared with a scaffold fabricated through AM methods (Fig. 2).

Fig. 2

Recent advances in AM technology, tomographic reconstruction, and numerical modeling methods have allowed workers to design a range of complex internal scaffold architectures with a range of length scales. These can be fabricated with a minimum feature length of <100 μm and the resulting structures inspected for perfection. This figure shows the design file, image of fabricated scaffold, and 3D tomographic reconstruction for three different designed scaffold architectures and a salt-leached structure formed by a random foaming process. [Reproduced with permission from (24)]

With 3D manufacture, rapid-throughput screening of structural variation can be difficult because each structure requires a finite volume of study. However, considerable progress can be made through the use of highly structured surfaces (so-called “2½D” objects), where micromanufacturing methods can be used to provide a large range of structures with relatively high aspect ratios by varying a small number of structural units or building blocks. Unadkat et al. have demonstrated a high-throughput method of studying the influence of surface topology on the fate of human mesenchymal stem cells (24). They used conventional silicon lithographic methods to form a master stamp to pattern a thermoplastic biomaterial, poly(lactic acid). Because the variation in surface topology was spatially correlated, they were able to use optical probes and image analysis techniques to correlate cell response with local topology. An example of their “topochip” is illustrated in Fig. 3.

Fig. 3

Example of a “topochip”—a method of producing a variation in surface structure that can be identified by spatial correlation. (A and B) Examples of different surface topologies at different locations on the surface (scale bar, 50 μm). (C) TopoChip (with dimensions 2 cm × 2 cm) and its carrier system. (D and E) Local cell behavior can be identified and correlated to the known position of each surface topology unit (280 × 280 μm). [Reproduced with permission from (25)]

As well as allowing the exploration of experimental variables in scaffold design and providing controlled features to study cell interactions, AM-produced tissue scaffolds have recently progressed from the laboratory toward clinical application. Two recent studies have demonstrated substantial progress. Reichert et al. carried out a large-scale trial using a sheep model for hard-tissue regeneration in long bones (25). This study compared the performance of a tissue engineering solution, using polycaprolactone-hydroxyapatite composite scaffolds fabricated by a filament deposition method (fused deposition modeling), to that of an autologous graft. The implant-tissue integration was assessed biomechanically and by x-ray computed tomography; the behavior of the AM-produced implant was equivalent to that of the autologous graft. Probst et al. have reported the successful outcome of a trial clinical animal model translation study that used a designed porous scaffold fabricated by AM (26). It is highly likely that AM scaffolds will be used clinically in hard-tissue reconstructive surgery in the near future, notwithstanding the considerable regulatory hurdles that still need to be overcome.

A key problem in the manufacture of scaffold structures by AM methods is the need to fabricate internal cavities via a layer-by-layer additive process. Such cavities or pore networks are required to assist nutrient transport in a bioreactor and to provide a framework to encourage neovascularization as ECM is deposited. Self-evidently it is difficult to deposit a material above an empty cavity, and this problem is overcome through the use of either temporary support structures or sacrificial materials that are leached away after manufacture to form cavities (2, 27). The use of leaching to form cavities is limited by the need for cytocompatible solvents and leaching products; recent work by Miller et al. has proposed an elegant solution to this problem by developing a soluble carbohydrate glass formulation (28) that can be readily dispensed by filament microextrusion to form a network that can be leached with water to form a series of interconnected fine cylindrical channels (Fig. 4).

Fig. 4

Three-dimensional lattice fabricated from a carbohydrate glass structure using a filament extrusion AM technique. A photopolymerizable cell suspension is cast around this structure, which is subsequently removed by leaching to provide a vascular network. Scale bar, 1 mm. [Reproduced with permission from (29)]


Bioprinting has been defined as the positioning of biochemicals, biological materials, and living cells (19). It is distinct from AM technologies used to fabricate scaffolds because it allows the deposition of different materials via the same equipment (e.g., the fabrication of structural materials and the simultaneous or sequential deposition of cells) (4). By analogy with color printing, we can use the technology to print discrete patches of different materials (or gradients and transitions between materials) to explore their influence in more detail. Printing can be conveniently divided into analog and digital printing. Analog printing is normally the direct transfer of an image by some contact process onto a substrate. Common high-resolution contact printing technologies used for bioprinting applications include microcontact printing and dip-pen lithography. These have both been used extensively to pattern surfaces with biochemicals and adhesion modifiers (29, 30), chiefly to study how surface features affect the behavior of cells. Both technologies sacrifice speed of application to achieve very high spatial resolution, typically <<1 μm, and neither can be used to deposit cells. This review is limited to computer-controlled and digital printing processes. In digital printing, the image is converted into pixels and the print consists of a digital pixel on/off application of material. Many forms of digital printing are contactless; the most common is inkjet printing. Another digital printing technology, laser forward transfer, has been used extensively for bioprinting applications. The most important computer-controlled deposition process is plotting or microextrusion.

Printing Surfaces and Structures to Control the Distribution of Cells

Inkjet printing is ideally suited to the deposition of biological materials because it is designed for the generation and precise positioning of picoliter (pl) volumes of fluid (31); these liquid drops can be aqueous solutions and can thus be used for the deposition of most biological species (19). Inkjet printing was used as a tool for the deposition of cells and biological materials in 1988 by Klebe, who used a Hewlett-Packard desktop printer to deposit collagen and fibronectin suspensions as well as cells to form 3D simple tissue analog structures (9). This study used an unmodified printer and simply replaced the graphics ink in a cartridge, after appropriate cleaning, with a biological fluid. Desktop printers with unmodified ink delivery but with a new ink formulation have been used to deposit functional inorganic materials (7, 8) as well as cells and biological materials (32), illustrating the versatility of this fabrication tool.

The surface resolution attainable by inkjet printing is controlled by the size of a printed drop in flight and its contact angle on the surface after impact (33). The smallest drops available through inkjet printing are ~1 pl in volume (radius ≈ 6.2 μm) and thus, on mildly hydrophilic surfaces, they have a minimum feature dimension of >10 μm, similar to that of whole cells. Using electrohydrodynamic drop generation, it is possible to produce linear features ~2 μm in width (34). Early work featured the use of surface patterning to produce large areas on a surface to selectively control cell adhesion (35).

One advantage of inkjet printing over analog printing methods is that it is possible to engineer variation in surface concentration through overprinting at different drop densities (Fig. 5). This has been exploited by Campbell et al., who demonstrated that the response of cells could be directly correlated the surface concentration of printed hormones (36). They were also able to show that surface concentrations of growth factors are able to control stem cell fate and that different patterns could be used to differentiate cell fate in the same culture dish (37). These studies used flat polystyrene surfaces for printing and subsequent cell culture. In tissue, cells experience a 3D environment, so the bioprinting results may not be truly representative. Thus far, fabricating 3D distributions of growth factors in matrices to probe cell behavior has not been achieved. However, Campbell and colleagues have attempted to address some of these issues by using nanofibrous scaffolds onto which growth factors are printed (38).

Fig. 5

Illustration of how digital printing methods can be used to print spatially defined concentration gradients of biochemicals. (A) Schematic showing how gradients of fibroblast growth factor–2 (FGF-2) are printed with different surface concentrations formed by overprinting multiple drops in the same location on the surface. (B) Example of an FGF-2 gradient prepared using sequential overprinting of biotinylated FGF-2, which was labeled with streptavidin conjugate quantum dots. [Reproduced with permission from (36)]

Direct Cell Printing

If the goal of bioprinting is to reproduce tissue structures, it must be capable of fabricating complex heterogeneous architectures—either by positioning different cell types in desired locations, or by inducing progenitor cells to differentiate into the desired type in specific locations. Achieving these tissue analog or tissue progenitor structures requires a 3D fabrication facility with multiple material capability in order to fabricate a temporary structural support with an open structure to allow the flow of nutrients, along with the ability to position cells and biochemicals with controlled delivery (39). Considering that our ability to fabricate and characterize simple single-material scaffolds is relatively recent, the target of a printed tissue is highly ambitious. Three main technologies are used for deposition and patterning with cells: inkjet printing (9, 32), microextrusion or filament plotting (9, 40), and laser forward transfer (41). Similar technologies have also been used to transport and place clusters or aggregates of cells (10). These techniques have comparable levels of maximum attainable spatial resolution, ~100 μm. Much higher resolution can be attained using techniques such as electrohydrodynamic jetting (34), which has been demonstrated with cell suspensions (42). However, given that typical cell densities used during culture are ~107 cells/ml, an individual volume of 100 pl (drop radius ≈ 70 μm) will contain on average a single cell; this limits the effective accuracy of deposition for cell placement.

One concern is the level of stress that cells experience during deposition processes. Each method for fabricating cell-containing structures has different requirements for the material surrounding the cell. Inkjet printing requires a low-viscosity environment to allow efficient drop ejection (31), microextrusion methods have a very wide range of fluid properties that are compatible with the process but offer a lower spatial resolution, and laser forward transfer requires the cells to be immobilized in a gel (41). Early reports on inkjet printing of cells reported poor viability after deposition (32); however, subsequent work found cell survival rates consistent with those of unprinted controls (i.e., survival >95%) with appropriate choice of printing conditions (4345), which suggests that the precise formulation of the ink is important. It is noteworthy that Cui et al. (45) found that after printing, the cell membrane had a transient increase in density of nanopores; this allowed a greater transfection rate than with unprinted cells and may indicate the mechanism of cell damage that occurs during printing. Studies of damage to cells during microextrusion experiments have been less systematic. Chang and Sun have reported cell survival rates in the range of 40 to 80% after extrusion of HEPG2 cells, with the survival rate decreasing with increasing extrusion pressure (46). This trend of increasing mechanical stress reducing survival after deposition is consistent with Saunders’ data for cell survival after inkjet printing, where an increase in the actuating voltage used to generate the drops led to a decrease in cell survival (44). The same trend is seen with laser cell transfer, where a greater amount of optical energy is seen to decrease cell survival rates after deposition (47).

Bioprinting, which began as a concept for cell manipulation (9) and is now a viable technique for patterning with cells (32, 40, 46), has been combined with a number of biological materials to directly produce hybrid cell-containing structures. D’Lima et al. used an aqueous solution of poly(ethylene glycol) dimethacrylate that contained chondrocytes in suspension and printed this into a model osteochondral defect before using a photolytic cross-linker to form a hybrid cell-containing hydrogel (48). After a number of days in culture, the printed structure appeared to have integrated into the surrounding tissue (Fig. 6), demonstrating the feasibility of this approach. Although this study demonstrates great promise, the general use of photoinitiated cross-linking of cell suspensions must address the potential cytotoxicity of the photoinitiator. Williams et al. carried out a systematic study of the influence of a number of different photoinitiators on cell populations and found that the relative toxicity of each chemical studied depended on cell type (49); hence, great care must be taken in selecting appropriate polymer-photoinitiator combinations in the presence of cells.

Fig. 6

(A) Schematic of bioprinting a cartilage analog structure, combining inkjet printing with a poly(ethylene glycol) dimethacrylate (PEGDMA) solution containing cells in suspension with a simultaneous photopolymerization process. (B) Light microscopy image of cell-containing polyethylene hydrogel printed into a defect formed in an osteochondral plug (scale bar, 2 mm). After culture, the cells within the printed material express ECM similar to those in the adjacent tissue. [Reproduced with permission from (48)]

Prospects for the Future

We are still a long way from organ printing. Although current deposition and fabrication technologies allow us to build structures that are analogous to tissue in their composition, the development of fully functioning tissue is a much greater step. The use of AM to produce scaffold structures, mainly for bone tissue engineering, is now almost routine but is still only possible with a limited set of biomaterials. The use of AM-fabricated scaffolds appears to be making important steps toward translation (25, 26), and for hard-tissue applications, the barriers are now regulatory rather than scientific or technical. The prospects are less good for soft connective-tissue applications where more compliant materials (e.g., hydrogels) are required. Here the main technical challenges are to find suitable biomaterials with appropriate mechanical properties that can be deposited using AM methods. The challenge here is in the field of developing new biomaterials. To date, fine-detailed hydrogel or biological material scaffolds have been fabricated using indirect methods, where a sacrificial mold is built using AM (28, 29).

Bioprinting offers promise for the fabrication of structures modeled on tissue architectures. The majority of published work to date is at a relatively low level of technological readiness and has used a very limited range of materials: sodium alginate, modified diblock copolymers, and photocured acrylates. The range of available materials is severely constrained by the need to develop cytocompatible gelation mechanisms that can be delivered by AM methods using fluid delivery (inkjet and filament extrusion) and can produce a cell-containing matrix with an appropriate range of mechanical properties. There is also still considerable uncertainty concerning the level of cell damage that occurs during cell deposition by all bioprinting methods. It is clear that much further work will be needed in this area before regulatory approval can be obtained for translational studies. What is more likely is that we will use these tissue analog structures for applications such as toxicity screening and drug testing (21, 39). One interesting development is the use of microfabrication technology to construct tumor models, allowing variation in physiological conditions in vitro (50).

For any new technology to have an impact in scaffold manufacture or tissue printing, further consideration of generic problems must also be considered. The concept of a scaffold requires an understanding of the behavior of an implanted structure within the body and its interactions at a cellular and tissue scale. The rate of scaffold degradation and ECM formation must be both understood and controllable. However, AM methods provide an important advance in scaffold design because many of the parameters of architecture that control the scaffold’s physical, mechanical, and degradation properties are adjustable within fairly broad limits. Thus, as discussed in detail by Melchels (24), it is now possible to model the influence of tissue architecture and validate the model by experiments on a range of structures. A challenge here is to fully ascertain whether designed scaffold structures are reproduced by AM techniques, and the recent increase in availability of computed tomography imaging will aid this effort. Tomography will again assist in monitoring degradation in vitro, and possibly functional nuclear magnetic resonance may be useful in imaging fluid flows and degradation of scaffolds in vivo. Predictive modeling or validation of tissue engineering solutions is likely to be the area of immediate impact for these new manufacturing technologies.

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article