Method Special: Bioprinting

Bioprinting in the Lab and the Doctor’s Surgery
by Steven Buckingham, Labtimes 06/2017

3D printing is the Next Big Thing since the printing press. When applied to living tissues, it offers the exciting promise of printed tissues and organs to order. But is it versatile enough to deliver a consistent and flexible platform for making organs?

Photo: Wake Forest Institute for Regenerative Medicine

It’s December, 2067, your grand-daughter, Sophie (now 65 years old), is attending her appointment for a liver replacement. Now a routine doctor’s job, she watches a 3D bioprinter through the window that opens into the lab next door. A high resolution scan has modelled her own liver, down to the very details of the vasculature and the network of ducts. An algorithm designed to recreate what the liver would look like, were it only 20 years old, turns the clock back half a century. Pluripotent stem cells derived from her own body have been reprogrammed and printed, layer by layer, as her brand new, custom-built liver builds up. In a couple of days, once the artificial organ has matured, she will have the liver implanted.

Back to 2017, where this is still just a dream. But one, if the enthusiasm of many is to be believed, that is sure to become reality.

3D bioprinting is the application of 3D printing techniques to biological materials. The basic idea behind 3D printing is really quite simple. All you do is print a series of thin sheets, each representing a 2D slice through the 3D object you are trying to make, and then simply glue them all together. For example, to print a 3D cylinder you would glue a series of discs into a stack. Make the discs progressively smaller and you get a cone. You get the idea.

What about overhangs?

3D printing yields ultra-fine polymer slices and then automatically fuses them together. Ah, you say, what about overhangs? Yes, that indeed has been a serious challenge. But there is a trick (and it will be useful to remember this when we talk about bioprinting): you print your section where there is an overhang (so this will, in fact, be two bits with a space between them), then fill in the gaps with some sacrificial material. Glue it all together and when it is set, dissolve away the sacrificial material. Or you can print using an ink that can be polymerised, using a laser or other external input.

Applying 3D printing to bioprinting is conceptually very simple. Here, the “ink” is some sort of cell suspension and, usually, another ink is used to provide support. But hold on before you rush out and try it with your HP inkjet. Your desktop printer can be very rough with its ink and will not respect the narrow tolerance of living cells. High temperatures and shear forces alone would be fatal. Bioprinters have to solve the challenge of reconciling the engineering needs of printing with the biological needs of cells. 3D printers work with hard, durable plastic – biologists work with soft, living cells. And once you have solved that problem, you are left with the task of developing inks and delivery strategies that will construct useful 3-dimensional structures.

Smart hydrogels

How do you make a bioink? In essence, you need cells, of course, plus some sort of supporting material. This material has to be both biologically and mechanically favourable. Nature has once again been kind to us, providing us not only with an example (the natural extracellular cell matrix) but also the starting materials. People have made bioinks from naturally-derived colloids, such as Matrigel (R), an extracellular matrix material derived from mice tumours or more simple, extracellular matrix proteins like collagen or hyaluronic acid. Finding the right one for any specific application is a matter of educated trial and error.

Several labs have developed bespoke “smart” hydrogels, such as using a hydrogel backbone that includes a degradable peptide motif. Many factors can determine the success or otherwise of the hydrogel used to embed the cells, including the mechanical properties, such as the stiffness. Having so many parameters for the choice of hydrogel has proved a blessing and a curse – a blessing in providing a wide space of engineering possibilities, a curse when it comes to optimising.

While the range of bioink designs has been expanding, so have the strategies for using the inks to build up a 3D structure. Two broad approaches have emerged – extrusion and lithography. In extrusion methods, bioink is forced out through a nozzle in a continuous ribbon. Either the nozzle or the product is moved in three dimensions to build up the shape. Alternatively, the material can be ejected as droplets, just like a desktop printer.

New bioinks and printing techniques

All extrusion approaches have to somehow solidify the ink after it has been deposited. Usually, some sort of chemical cross-linking is done. This can be simply by printing at a temperature high enough to preserve the material in a liquid phase but low enough not to damage the cells, followed by cooling to produce the gel state. In practice, however, a more robust cross-linking is done chemically (forming covalent bonds) or physically (using hydrophobic or ionic interactions). It all boils down to being able to apply the ink as a liquid and turning it as quickly as possible into a solid – all in a cell-friendly environment.

The extrusion approach has the same constraints regarding overhangs as with all 3D printing and the solutions to the problem are, likewise, similar. 3D bioprinting of unconstrained shapes is often done by replacing the nozzle with a needle and injecting the ink into a supporting gel. Once the shape has been created, it is solidified by polymerisation and the supporting gel released by, for example, an up-shift in temperature.

Printing cells layer by layer

Some labs have used lithography to build up a 3D shape. Layers of bioink are deposited and retain their liquid phase until illuminated by a laser to photopolymerise the hydrogel. Alternatively, the photopolymerisation can be controlled by applying a precisely-patterned illumination, a bit like projecting a picture.

3D-bioprinted coronary artery. Though researchers made considerable progress in recent years, they are still a long way off functional bioprinted tissues and organs. Photo: Carnegie Mellon University

You can also build a 3D shape by dissolving away sacrificial material. It is common practice to make a vascular network pervading a tissue, by printing the tissue layer by layer with two inks: one a hydrogel with the cells, another a sacrificial ink forming the vasculature. After the hydrogel is cured, the sacrificial material is dissolved away and can then be filled with a bioink, containing cells needed for laying down vasculature.

Choosing the right material

The material you print your cells onto, which provides the solid supporting structure for your cells, is critical. And if you are going to risk a patient's life with a 3D-printed implant, the interaction with the host has to be taken into account, too. Aside from the obvious danger of an immune response, the material must have the right mechanical properties. Too flexible and it might rupture. Too stiff and you could evoke fibrosis.

One way around this is to avoid using such materials altogether. Narotushi ­Hibino of Johns Hopkins University, Baltimore, found a way of printing human heart tissue without any artificial bio-materials, at all. Their approach was to generate cell spheroids from a mixture of cardiomyocytes, fibroblasts and endothelial cells, and then use a robotic manipulator to stick them onto an array of pins. Imagine something like a circus “bed of nails” with a sphere stuck onto the end of each nail. Over time, the spheroids fused into a homogeneous tissue, with many of the properties of cardiac tissue. Coordinated contractions were visible, cardiac-associated gap junctions were expressed at cell junctions and an action potential closely resembling a cardiac action potential was propagated around the tissue.

But 3D printing is limited in its resolution. It is suited to making tissues bigger than 1 cm³ and rarely achieves a resolution less than 250 µm, far from the resolution you would need to print realistic tissues. There is a limit to how fine a nozzle you can use in bioprinting. Make the nozzle too fine, and the shear stresses tear your cells apart. True, there are alternatives to nozzles but each of them comes with other costs – problems with clogging, limited cell density and droplet size, for instance. Happily, Hagan Bayley from the University of Oxford, who revolutionised DNA sequencing with his invention of the Nanopore system, has now come to the rescue with a novel form of bio-inkjet printing that has taken the resolution up a notch (Scientific Reports 7:7004).

Printing with droplets

Instead of printing a bioink onto a surface, Bayley’s approach is to make cell-containing, aqueous droplets injected into oil. A cell-favourable medium containing the cells is inkjet-sprayed into oil, where the physical properties of the oil/water interface help the droplets retain their integrity. The droplets settle onto a glass surface at the bottom of the oil and the lipid ­monolayer that surrounds them helps them to stick to one another. As the droplets pack themselves into a monolayer, they take on a hexagonal shape, which keeps them in position.

Hagan Bayley’s group invented a droplet-based bioprinter, capable of printing different types of mammalian cells with high spatial control. Photo: Alexander Graham

High resolution printing

Contrast this with the usual method of spraying droplets onto a surface, where the droplets spread out, undermining the printing resolution. In Bayley’s approach, layers of droplets are stacked on top of each other to form a 3D “tissue”. When the printing is complete, the tissue is cooled to form a gel and the lipid washed off. The whole structure is then embedded in a cell-free bioink and transferred to a culture medium. Bayley's method broke the 200 µm resolution barrier with a printed tissue that was able to produce something very much like cartilage.

So-called 4D bioprinting extends 3D ­bioprinting into the time dimension. This involves printing a 3D structure that somehow changes shape over time, either as a maturational process or, more excitingly, as a responsive machine. 4D printing has been used to overcome the limitations of 3D bioprinting. While there is an ever-growing inventory of tips and tricks to produce hollow channels and overhangs, there are limits to what can be achieved with these tricks.

These limits get pushed further back, by making use of materials that change shape in response to an environmental stimulus. Several labs have printed cells onto surfaces that roll up on exposure to water, in order to make hollow vessels too small to be made by printing alone. David Gracias’ group from the Department of Chemical and Biomolecular Engineering at the Johns Hopkins University, printed cells onto a double layer of polyethylene-glycols (PEGs) of different molecular weights (Adv. Healthcare Mater. 2, 1142-50). The two PEG layers absorbed water at different rates, causing the surface to roll up, forming a narrow tube. The structures made in this way can be much smaller than those made by extrusion printing, which is limited in resolution to some hundredths of micrometres by the constraints on the nature of the ink.

Maturational strategy

Another way in which bioprinting is taken into the fourth dimension, is in exploiting the maturational processes that take place after printing has been completed. Living things change with time and a successful bioprinting strategy aligns these changes to the purpose at hand, often exploiting the surprising willingness of cells to self-organise.

Jennifer Lewis’ and David Kolesky’s work at Harvard University used an ingenious maturational strategy to produce thick, vascularised tissue (PNAS, 113, 3179-84). They began the process by using two inks, one to make the blood vessels, the other to make the deep tissue. The blood vessel ink was a sacrificial ink, based on pluronic acid and contained thrombin. The deep tissue ink was based on gelatin and contained human mesenchymal stem cells, neonatal dermal fibroblasts and, critically, fibrinogen. The “blood vessels” were printed as a regular lattice surrounded by the deep tissue ink, with a gap in between them to keep them apart.

Then the assembly was in-filled with a gel containing transglutaminase. This brought the two inks into diffusional contact, with the result that thrombin, fibrinogen and glutaminase mix in together. The thrombin causes the fibrinogen to form fibrin, which the transglutaminase cross-links stably to the gelatin. The assembly was then cooled, which converts the sacrificial ink in the “blood vessels” to go from a gel state to a liquid, which is then flushed out of the channels. The now empty channels were perfused with a suspension of human umbilical vein endothelial cells (HUVECs). Shortly after, the HUVECs did their appointed job of lining the vessels, with the result that they could be perfused with a medium that not only supported the embedded cells but could also be supplied with growth factors to drive the formation of bone tissue.

Creating micromachines

Another reason for developing 4D bioprinting is to make responsive micro-machines. For example, Leonid Ionov’s group, then at the University of Dresden, deposited yeast cells onto a double layer of gels that expanded and shrank at different rates with changes in temperature (Soft Matter 7, 3277-79). The gels had a flower-like structure that opened up as the temperature increased, releasing the cells. This points to exciting possibilities for 4D bioprinting in drug delivery, where cells and materials are combined to produce programmable nano-machines. The technical requisites for such devices are already becoming available. Hagan Bayley showed how printing micro-droplets with tuned osmotic properties can fuse, to form complex sheets that can be manipulated by changing the osmotic pressure (Science 340:48-52). These networks can have biological proteins inserted into their membranes to make them do useful things.

3D bioprinting is already earning its keep in the lab. We all know just how much the development of 2D cell culture has contributed to bioscience. But 2D culture has almost passed its sell-by date. The limitations of growing cells as a flat sheet are becoming clearer. This is particularly felt in the area of cancer, where it is emerging that the micro-environment of cells within a tumour is one of the biggest factors that determine how each cell will behave, including whether it decides to metastasise. Cells in big lumps talk to each other and what they say, depends on where they are in spatial relation to each other. The cells in a tumour agree with John Donne: no cell is an island entire of itself (or something like that). They are more like citizens in a city, each dependent of the rest. Oncologists are sociologists.

Controlled pattern of cells

Cancer researchers have long wanted to be able to grow model tumours in 3D and using 3D bioprinting to make more realistic models is a hot area of research. Utkan Demirci’s lab at Harvard found a way to automate the bioprinting of human ovarian cancer cells mixed with normal fibroblasts in a way that makes it amenable to high-throughput methods (Biotechnol J, 6: 204-12).

3D-printed tumour cells open new possibilities to establish more realistic cancer models. Photo: Institute of physics

The key to their success was the ability afforded by 3D bioprinting to pattern the cells in a controlled manner at determined densities. Growing tumours encourages their own vascularisation to meet their metabolic demands, and Shaochen Chan and colleagues at the University of California, San Diego, used a device that employs a bank of mirrors to focus a polymerising light beam to produce, in seconds, an arbitrary user-designed pattern onto a bio-gel. This system produced tumour models with perfusion channels, with life-like branching patterns (Biomed Microdevices 16: 127-32).

The need for an alternative to donated organs is, of course, a pressing one. Not only could waiting for suitable donors be eliminated but organs printed, using the patient’s own cells, would have no risk of rejection. There are other benefits to 3D bioprinting. Tissues or organs created from a patient's own cells could be used to predict the individual effects of medical treatment specific to that patient.

Printed organs still out of sight

Is 3D bioprinting ready for making new organs, like Sophie's liver? No, not yet. Making homogeneous tissues is (relatively) easy, structured tissues are an occasional success, whole organs beyond our current horizon.

All the same, there is an awful lot of good you can do, even with a homogeneous printed tissue. When the cardiac tissue made by Hibino’s bed-of-nails trick was transplanted into rat hearts, it fused successfully with the host tissue and induced the host to provide new blood vessels. Meanwhile, Vladimir Mironov at the Laboratory for Biotechnological Research, Moscow, took advantage of the fact that the thyroid gland has no ductal system.

He printed 3D spheroids from thyrocytes and endothelial cells and implanted them into hypothyroid mice, and found that they brought blood thyroxine levels and body temperature back to normal (Biofabrication 9:034105). And while Lewis and Kolesky's printed bone may not yet be as strong as real bone, there is reasonable hope that a combination of optimisation and felicitous opportunity will overcome the barriers.

Big challenges ahead

The big challenge for 3D bioprinting is affordability and versatility. Impressive as the breakthroughs have been, they tend to depend on a set of circumstances being just right – having just the right combination of materials, cell types and structural features. Will the future of 3D bioprinting be limited by a combination of serendipity and laborious optimisation, or will we be surprised by the willingness of cells and materials to self-organise? We are far from finding a unifying set of design principles for arbitrary organ, or even tissue, construction. Will we ever find them? Looks like we will have to use our time machine again.

Last Changed: 28.11.2017