Machining the Body

TECHNOLOGY OVERVIEW

Though decidedly high-tech, tissue printers are based on long-proven technology. The solid freeform fabrication (SFF) and rapid prototyping (RP) methods have been used for more than a decade in engineering circles to build inexpensive three-dimensional models of parts, and in some cases, even to manufacture the parts themselves.

Unlike most 3-D manufacturing technologies, which remove material from a solid block to form the desired shape, rapid prototyping is an additive process. "The underlying concept in rapid prototyping is to go directly from a three-dimensional digital representation of an object that you want to build ... to a machine that can automatically build up that object in a layer-by-layer manner," explains Suman Das of the University of Michigan in Ann Arbor.

In tissue-engineering applications the 3-D data come from computed tomography or magnetic resonance imaging (MRI) scans, which are translated into micrometer-scale sections. The RP machine prints the sections sequentially, using a bioresorbable "ink"; newer technologies incorporate growth factors and even living cells. The end result: a three-dimensional, porous scaffold that degrades slowly in the body to allow healthy cells to migrate into a defect and repair it.

Three-dimensional printing (3DP), one of the first RP techniques to be used in tissue engineering, was originally developed in the 1990s by Emanuel Sachs of the Massachusetts Institute of Technology (MIT). The 3DP technique starts with a 3-D computer model sectioned into planes 100–200 μm thick. These layers are translated into a series of raster scans that are sent to the 3DP machine, which spreads a layer of organic polymers and water-soluble components onto a piston in dry powder form, followed by a binder which is selectively printed to hold the particles together in regions that represent the cross-sectional geometry. The piston is dropped, another layer is spread and printed, and the process repeats. "Each layer is supported by a previous layer that's solid, [which] means you can make overhangs and channels within a device with a lot of flexibility," says MIT's Linda Griffith, who has adapted 3DP for tissue-engineering applications.

Courtesy of Therics Inc.

Though organ printing is decades away from reality, the three-dimensional printing technology underlying it is in use in the clinic. Therics uses 3DP to produce synthetic bone graft substitutes that have been FDA-approved for the repair of small bone defects. Top Left: Internal architecture of tissue engineered bone scaffold. Top right: Tread features to anchor tissue. Center left, right: Graft resorption in canine defect. Bottom left, right: Bone ingrowth in graft.

The scaffold's feature size and porosity can be adjusted by varying the composition of the powder, explains MIT research scientist Jim Serdy. Bone-tissue engineering applications, which generally use materials such as polycaprolactone, tricalcium phosphate, or hydroxyapatite, require pore sizes of around 200 micrometers to allow bone cells to migrate into the scaffold. The transplant shape and gross features, such as vascular pathways, are generated by 3-D printing from an MRI image. Smaller features and surfaces for cell attachment and in-growth are generated through postprocessing steps.

MIT has licensed companies to use 3DP technology in a variety of fields. In May 2003 the US Food and Drug Administration granted one such licensee, Therics of Princeton, NJ, clearance to market its first 3DP-fabricated bone-graft substitute for use in filling relatively small (less than 5 cm) bone voids and defects.

Therics' 3DP product is composed of standard materials such as beta tricalcium phosphate, a bioresorbable ceramic. Lead scientist Sunil Saini says 3DP's ability to control both shape and porosity allows the designer to make a part's internal architecture resemble what native bone cells actually encounter inside the body. "The range of pore sizes ... within the device has been optimized to allow for very aggressive tissue in-growth once implanted into the defects," he says. Moreover, Saini says, a 3DP-manufactured synthetic bone graft substitute is faster to produce and does not rely on human donor tissue (which can be a problem due to limited supply).

One limitation of 3DP is that it requires some form of postprocessing to strengthen the part after fabrication. University of Michigan's Das uses a different RP technique, selective laser sintering (SLS), to develop polycaprolactone scaffolds in a single processing step.² In SLS, a focused laser beam scans a thin film of powder, fusing regions of the powder in a pattern that corresponds to the structure of the layer being built; the layer is lowered 100 microns, and the process is repeated for the next powder layer."

The technique allows the user to control scaffold porosity on two levels, Das says. "In addition to building three-dimensional, designed porous architectures into the scaffold, one can adjust or fine-tune the laser sintering process in a manner such that designed solid portions of the scaffold can be either porous or fully dense."

Other researchers, like Sun, are developing extrusion-based approaches. Jennifer Lewis of the University of Illinois, Urbana-Champaign, and colleagues, for instance, have developed a direct-write procedure in which polyelectrolyte filaments are printed by depositing a concentrated aqueous solution through a fine nozzle into a reservoir containing a solvent that induces coagulation or solidification.³ Her approach can be used to fabricate 3-D structures with submicron features. Similarly, Dietmar Hutmacher and colleagues at the National University of Singapore developed a fused deposition-modeling (FDM) instrument that melts and pumps a polymeric filamentous scaffold material through a nozzle and deposits it layer-by-layer onto a platform. Their technology is currently marketed by a spin-off company, Singapore-based Osteopore International, which manufactures scaffolds for bone repair.

DESIGNING FOR FUNCTION

Optimal tissue-engineering scaffolds not only mimic the shape of the part they are intended to replace, but also restore mechanical function. FDA-approved synthetic bone-graft substitutes cannot be used to repair defects in weight-bearing limbs, because the scaffold dissolves too quickly to be replaced by sufficiently dense bone. Load-bearing bone replacement currently involves the use of metal parts or orthologous bone grafts in which healthy bone from another part of the patient's body is used to fill the defect. Neither approach is ideal: Metal parts are harder than bone and can damage healthy bones by rubbing against them, while orthologous bone replacement requires healing of not only the initial defect but also the secondary site.

Advanced Ceramics Research of Tucson, Ariz., is currently developing an RP-fabricated bone graft substitute called Plasti-Bone, in which a substrate material called PBT (a plastic similar to that used in milk containers and which dissolves over several years) is covered with hydroxyapatite or tricalcium phosphate, which dissolves within six to eight weeks and acts to guide bone cells into the implant. "After two or three years, this whole piece that you are placing into the missing area will get completely replaced by the patient's own natural bone," says Ranji Vaidyanathan, the company's director of polymer products. Advanced Ceramics Research has performed successful tests of Plasti-Bone in rats and dogs and is currently seeking FDA approval to test the material in humans.

While Plasti-Bone tackles the problem of function from the materials side, other researchers approach it at the level of design. Scott Hollister of the University of Michigan develops computer-aided design algorithms to optimize a scaffold's internal architecture so that it can handle loads while at the same time remaining permeable enough for cell infiltration. Balancing these two goals can be tricky, Hollister explains: "The more porous you can make the scaffold, the better you can deliver the biologic and the better tissue in-growth you get, but at the same time you make the scaffold much weaker in performing function."

Hollister also acknowledges that even if a computer model finds an optimal design, its implementation is limited by the RP technologies themselves, most of which currently cannot print features at the nanoscale or even microscale level. "When you do these optimization schemes, they may end up creating structural features in the design that are smaller than 200 microns. So if you cannot replicate that on the fabrication system, then you won't be able to realize that structure and you won't be able to get the performance that the model has predicted," Hollister says.

Scaffolds with smaller feature sizes are in the making. Hutmacher is developing a robot to stack "microparts" measuring 100 to 200 micrometers into 3-D scaffolds, each of which can be seeded with cells prior to stacking. "The advantage of using the robot is that you have now four degrees of freedom in your design parameters. Normally with rapid prototyping, we build layer by layer, so we have two degrees of freedom only," Hutmacher says.

PRINTING CELLS

In most RP-fabricated tissue-engineering scaffolds, cells are either seeded into the finished device or migrate naturally into them when implanted in the body. Some tissue engineers argue, however, that such constructs are not ideal. Brian Derby of the University of Manchester, UK, explains that the problem with seeding a prebuilt scaffold is that cells will penetrate only a few millimeters beyond the scaffold's surface. "So you end up with a sort of a crust around the scaffold where the cells adhere themselves, and then they take a very long time to fill the interior," says Derby.

© 2004 Nature Publishing Group

(A) The ink-deposition process. A concentrated polyelectrolyte ink is housed in a syringe immersed in a coagulation reservoir and deposited onto a glass substrate. (B) Optical image acquired in situ during deposition reveals the features drawn in (A), including the deposition nozzle that is patterning a three-dimensional lattice, as well as a completed radial array. (scale bar: 100 μm). (C) Three-dimensional periodic structure with a face-centred tetragonal geometry (filament diameter: 1 μm; 10 layers; scale bar: 10 μm). (D) Three-dimensional radial array (filament diameter: 1 μm; 5 layers; scale bar: 10 μm). (Reprinted with permission from G. Gratson, et al., Nature, 428:386, 2004.)

To circumvent this problem, several groups, including Derby's, have developed methods to print cells directly into the scaffold during fabrication. Thomas Boland and colleagues at Clemson University in South Carolina used a modified Hewlett Packard thermal inkjet printer to print mammalian cell-based "bioinks" – Chinese Hamster Ovary (CHO) cells and rat embryonic motor neurons – onto gel-based "biopapers." They demonstrated that only 10 percent of the printed cells were damaged by the procedure.⁴

Boland and colleagues had previously shown that 3-D tubular structures could be fabricated in a similar manner using cell aggregates as a bioink,⁵ and that these structures remained functional (i.e., heart cells "beat" like normal heart tissue). In unpublished research, he has constructed 3-D parts from single-cell suspensions by alternately printing cell and scaffold layers. But these constructs have limited mechanical stability, he says. "These are tubular structures or branched structures that have very thin walls ... and are hollow on the inside. And so these work for soft tissues including the heart," says Boland, but not for more structurally rigid tissues.

Derby and colleagues use a piezoelectric-based inkjet printer to print human skin cells into a biodegradable plastic matrix several millimeters thick; they are testing their device with other cell types, including bone and cartilage. Though every cell type they have tested appears to survive printing, a critical question is whether the cells retain their phenotype after the procedure. "So far, a large number of cells appear to retain their character, but it's possible that some specialist cells required for tissues may not retain their character after printing," Derby says.

Cell printing has some inherent difficulties, notes Griffith. The feature-size resolution of cell printing is one challenge, as build rates slow down considerably as resolution becomes finer (Boland's inkjet printers print droplets between 10 and 50 microns in diameter, roughly the size of a cell), and most cells are too fragile to be processed by this technique. "Anything that involves processing and handling of cells inherently adds risk to a therapeutic application, and cells are very fussy ... all the applications we work with use primary cells, [and] they are not super-happy to be in some kind of processing machine for [the time] it would take to make a device," says Griffith. Ultimately Griffith found that it was more practical to build a scaffold first and let the cells organize themselves within it.

Derby agrees that cell handling is a bottleneck for 3-D cell printing. In his method, cells are suspended in a culture medium that keeps them alive for the several minutes it takes to print out a sample construct. But he cautions that for the method to go from the laboratory to the clinic, "We'll have to devise methods that will keep cells alive for hours, because it will take a significant amount of time to build structures large enough to be of clinical interest. And that's one of the bottlenecks at the moment, developing this way of keeping the cells alive for longer periods of time."