Can robots, computers, and chip-making techniques save tissue engineering and bring internal organs to market?
Photo: Roger Wright/Getty Images; Illustration: Rob Magiera
Scott Burdette, all of nine years old, wanted to see what fire would do to a can of spray paint. When the can exploded, the left side of his body was covered with flaming paint.
He was flown by helicopter to the Children’s National Medical Center (Washington, D.C.), where his burned skin was replaced with a new covering of skin that quickly relieved his pain. He left the hospital eight days later and was back in school three weeks after the accident. Today, two years later, a casual observer would never know Scott had been burned.
Doctors often treat severe burns with grafts of a patient’s own skin. But the skin that helped Scott didn’t come from his or anyone else’s body. It came from the factory of Advanced Tissue Sciences Inc. (ATS, La Jolla, Calif.), a corporate pioneer in the fledgling field of tissue engineering, which seeks to provide off- the-shelf replacement body parts.
State of the industry
Biomedical science has made a lot of progress in understanding how cells grow into functioning tissue and what chemical and other cues they need to do it right. Tissue engineering is the application of that knowledge to the building or repairing of organs, including skin, the largest organ in the body. Generally, engineered tissue is a combination of living cells and a support structure called a scaffold. The scaffold, depending on the organ in production, can be anything from a matrix of collagen, a structural protein, to synthetic biodegradable plastic laced with chemicals that stimulate cell growth and multiplication. The “seed” cells that initiate this propagation come from laboratory cultures or from the patient’s own body.
But while tissue engineering has made great gains as a science, it has been much less successful as a business. Two leaders in the industry fell into financial trouble this fall. On 10 October, Scott Burdette’s supplier, ATS, filed for Chapter 11 bankruptcy, which allows the company to continue operations while figuring out how to restructure its finances. To keep the supply of skin coming, the company sold its stake in its skin-making operation to its more financially stable joint venture partner, the medical device maker Smith & Nephew PLC (London).
Weeks earlier, the other skin manufacturer, Organogenesis Inc. (Canton, Mass.), was forced into Chapter 11 bankruptcy, when its marketing and distribution partner, the drug company, Novartis International AG (Basel, Switzerland), refused to renegotiate how much it was paying Organogenesis for skin. It turned out that the contracted price was too low for the manufacturer to sustain itself. Industry insiders blame the bankruptcies on a combination of lackluster sales, the high cost of winning regulatory approval, and poor profit margins.
The bankruptcies are certainly a loss for the industry, but not a fatal one, says Michael J. Lysaght, biomedical engineering center director, School of Medicine, Brown University (Providence, R.I.) and keeper of tissue engineering industry statistics. Sales of engineered skin, about half of the industry’s US $50 million output for the year ending in June 2002, were on the rise; the rest of the total comes from engineered cartilage for joint replacement. Overall, companies involved in tissue engineering grew from 66 in 2000 to 99 in 2002, and investment in tissue engineering grew 14 percent to $675 million, Lysaght revealed at a joint meeting of the IEEE Engineering in Medicine and Biology Society and the Biomedical Engineering Society in October.
According to Gail K. Naughton, cofounder of ATS, the companies’ poor margins were largely due to a lack of automation in their manufacturing processes. For future ventures to succeed, she says, they will have not only to tackle the myriad scientific problems still left, but also to replace today’s largely manual processes with automated electronics-rich operations.
For one thing, precise sensors and control systems will be needed to create and maintain the biochemical and mechanical environments that nurture tissues like skin. Also, robotics and other automation will be needed to remove people from the tissue growth process. Already, fledgling firms and tissue engineering labs are borrowing some advanced engineering practices, like high-precision rapid prototyping and photolithography, as they strive to create engineered bone, cartilage, blood vessels, and internal organs. These technologies may be the forerunners of automated factories capable of mass-producing a head-to-toe variety of life-improving and life-saving body parts, including factory-built hearts and livers. This development could put an end to transplant waiting lists—and the suffering of those on them.
Only skin deep
Due in part to their simple structure, skin substitutes and cartilage replacements were the first engineered tissues to reach the market. One of these early products is TransCyte, the burn covering from ATS that came to the rescue of Scott Burdette. In cases like Scott’s, cadaver skin is sometimes used as a temporary covering to protect the patient’s recovering wounds. But cadaver skin is in limited supply, costly, and variable in quality, notes Michael Sabolinski, executive vice president of medical and regulatory affairs for Organogenesis, which makes a skin substitute called Apligraf. What’s more, Sabolinski adds, cadaver skin is usually rejected by the patient’s immune system within 20 days. But engineered skin isn’t rejected, say the manufacturers.
Besides serving as burn coverings, engineered skin substitutes can help patients with diabetic foot ulcers. Today, most of these ulcers are treated with an approach that includes antibiotics, glucose control, special shoes, and frequent cleaning and bandaging. Despite this treatment, the ulcers are often slow to heal, and some don’t heal at all. If an ulcer fails to heal, the patient’s foot may have to be amputated. Of the more than 86 000 lower-extremity amputations in the United States each year, 85 percent are preceded by a diabetic foot ulcer, according to ATS. In a clinical trial, an ATS skin product designed for wound healing, Dermagraft, healed more chronic ulcers, and healed them faster, than conventional therapy alone, the firm reports.
The skin factory
With so much demand in view, firms developed the infrastructure to produce thousands of skin grafts every month. At Organogenesis’ factory, though, it was a costly, by-hand process.
The company made Apligraf in petri dishes filled with a liquid nutrient that stimulates the skin’s growth by mimicking its natural environment. The grafts start out as donated infant foreskin tissue taken from circumcisions. Such a source of cells is not just plentiful, it’s incredibly potent. A sample the size of a postage stamp will multiply to such a degree that several football fields of skin graft can be crafted from it.
To start the process, the foreskin cells are separated into two types: dermal fibroblasts and epidermal cells. Tissue growth occurs over 20 days in several stages. First, the lower, or dermal, layer of skin grows on a scaffold made of collagen (from cows), which is set in a shallow, round dish well-bathed in nutrient medium. Then a technician adds epidermal cells, which spread over the dermis to form the upper skin layer. Finally, the skin is lifted out of the culture bath so the top is exposed to air, which triggers the formation of a tough outer layer of dead cells. The process produces round skin patches measuring about 8 cm in diameter.
Using manual labor alone, Organogenesis could already produce about 50 000 skin grafts a year. But the firm planned to introduce industrial-scale automation. The first step would be to automate the addition and removal of nutrients, which is the most time-consuming operation. Without a periodic refresh, the growing skin will starve and poison itself on waste products. An electronically controlled system would make the process more precise and reproducible, while reducing costs and the likelihood of contamination, according to Leon Wilkins, vice president of technology development at Organogenesis. Those plans, like the company itself, are now on hold.
Organogenesis’ rival, ATS, had already incorporated some automated features into its tissue manufacturing. The automation starts with transforming the cells obtained from foreskin samples into the vast amounts needed to make living skin in bulk. Cells multiply inside nutrient-filled bottles, whose interiors are corrugated to increase the surface area. Every few days, a robotic system developed by ATS refreshes the nutrient medium. A robot arm picks up one bottle at a time, pours out the old liquid, and fills it with new medium.
While Organogenesis skin grows in petri dishes, ATS’s skin is grown in enclosed chambers called bioreactors. These are designed so that nutrients flow through them in a manner that enhances growth. Like the petri dishes, the bioreactors are kept in an incubator, where sensors monitor temperature, humidity, and gas concentrations. This data is fed into a controller that ensures that environmental variables remain within the narrow range demanded by growing skin. For example, temperature is maintained at 37 C ±0.1, while carbon dioxide concentration is kept at 5 ±0.1 percent. If a variable moves out of its specified range, an alarm alerts plant personnel.
Inside the bioreactor, the cells multiply and reorganize, eventually growing into functional tissue. The bioreactors double as shipment containers, so when the skin is ready, reactors are closed off and stored at low temperatures for transport. ATS normally keeps hundreds of skin units in storage and was geared up to make 200 000 per year. “Selling off-the-shelf products is our business model,” says Anthony Ratcliffe, the company’s vice president of research.
Besides skin, tissue-engineering companies have already commercialized bioreactor-built cartilage for use in joint replacement and are hoping to commercialize bone built in bioreactors as well. Another skin factory, IsoTis NV (Bilthoven, the Netherlands), expects the three million surgeries per year worldwide that could require bone-replacement tissue, such as plastic and reconstructive surgeries, will become a billion-dollar market. So the company plans to put 500 bioreactors in a new plant designed to produce up to 5000 engineered implants a year.
But unlike skin, three-dimensional organs and tissues like bone require complex scaffolds that give shape and support to growing tissue. To provide the best geometry for cell attachment, the support structures are normally fabricated with a high ratio of surface area to volume. Most scaffolds are also biodegradable, dissolving slowly inside the bioreactor as new tissue grows and becomes strong enough to support itself.
From a biochemical standpoint, collagen and biodegradable polyesters used as sutures in surgery were natural first choices for scaffolds. But to fabricate complicated tissues made up of many types of cells, more chemically complex scaffolds are needed. To that end, scientists are busy working out precisely what chemical cues a growing cell needs to receive and at what point it needs to receive them in order to correctly form tissue. The hope is that some of those chemical cues can be embedded into the scaffold in a way that releases them at just the right concentration and in just the right order. But it’s an uphill battle, because at the moment scientists believe they have identified only a small fraction of the hormones and other factors involved.
Physically, scaffolds range from simple sheets to multi-material structures riddled with pores and channels through which developing tissues grow. For quick manufacture of even the most complex scaffolds, which are sized and shaped for individual patients, Therics Inc. (Princeton, N.J.) has developed a super-precise fabrication machine based on technology licensed from the Massachusetts Institute of Technology (MIT, Cambridge). The company’s TheriForm machine operates, in principle, as other rapid prototyping machines do. Like an ink jet printer, it deposits two-dimensional patterns onto a stack of “pages.” But a TheriForm page is actually a layer of powder, and the ink acts as a binder. The pages are built atop one another to construct 3-D scaffolds sized and shaped for individual patients—an ear built of cartilage, for example, or a missing chunk of bone.
Therics bases each scaffold design on anatomical data taken from the patient’s X-ray, magnetic resonance imaging (MRI), or computed tomography (CT) scans. This data is digitized and fed into a computer-aided design (CAD) program. Therics uses the CAD software to create a 3-D computer model of the scaffold, which is sliced into sections corresponding to the layers that will be manufactured atop each other to yield the final product. Another program translates the final design into instructions for the TheriForm printer.
To start the manufacturing process, a roller pushes a layer of powder (usually a ceramic or polymer) only a few tens of micrometers thick across a flat base plate that sits on top of a piston. Then a servo-controlled print head deposits drops of binder onto the powder in a pattern dictated by the blueprint for that layer. The drops fuse the powder particles together to form a thin section of the scaffold. Unlike many rapid prototyping systems, which use heat to fuse particles, the TheriForm process relies on a chemical reaction to bind the powder. This so-to-speak cold process will not harm heat-sensitive materials that may be needed in scaffold manufacturing, notes Christopher Gaylo, vice president of engineering at Therics.
When the drop pattern has been printed, the piston lowers the base plate between 10 and 250 um (the thickness of an individual layer), and the machine makes another layer on top of the first. The sequence repeats over and over, with layers stacking up and binding to each other as the process goes on. At the end of the process, the excess powder around the solidified layers falls away, leaving the completed scaffold.
TheriForm print heads are fitted with an array of nozzles that can print a variety of materials onto each layer. Besides binders, the nozzles can deposit living cells and so-called growth factors, meaning biochemicals that start or assist cellular growth through the scaffold. The machine can operate up to 32 nozzles simultaneously, each of which deposits 800 to 1200 drops per second. To ensure that correct amounts of material are dispensed, a sensor counts the number of drops deposited in each location.
The machine places each drop within 10 um of its intended position, according to Gaylo. With its pinpoint accuracy, the TheriForm machine can turn out anatomically accurate products having precise internal micro-architectures, such as spaces for blood vessels or other patterns known to foster tissue growth.
Therics plans to introduce its first products early in 2003. The company’s initial offerings will be simple bone-replacement scaffolds for the millions of plastic and reconstructive surgeries performed each year.
A chip shot for capillaries
Part of the reason engineered skin has gotten as far as it has is that it doesn’t need its own built-in network of blood vessels to get nutrients and oxygen. Engineered skin, once grafted to the body, appears to exude hormones that recruit blood vessels to grow into it. But large, 3-D organs such as the liver need their own network designed into them. So some tissue engineers have turned their attention to getting blood to flow where it needs to go.
Blood vessels occupy a good bit of Joseph Vacanti’s time these days. One of tissue engineering’s most experienced practitioners, Vacanti is a transplant surgeon who knows firsthand about the lack of organs available for patients who need them.
“The organ shortage is the major problem in transplantation today,” says Vacanti, director of the pediatric transplantation program at Massachusetts General Hospital (Boston). “When I began tissue-engineering work [in the mid-1980s], there were probably less than 10 000 people on organ waiting lists in the United States alone. Now there are more than 80 000. So if tissue engineering works for a vital organ, it will instantly change how we do transplantation.”
Engineered organs will require complex blood vessel systems made up mostly of tiny capillaries. Without capillaries, oxygen diffuses only a few hundred micrometers into tissue. So Vacanti and his colleagues are trying to develop a scaffold that would support the growth of an organ along with its oxygen-supplying blood vessels. His thoughts turned to nearby Draper Laboratory Inc. (Cambridge, Mass.), which, among other things, uses microfabrication techniques to build missile guidance systems. Could Draper’s expertise in semiconductors and microelectromechanical systems (MEMS) be applied to the production of tiny blood vessels?
Vacanti discovered that Draper’s MEMS technology group routinely produces 1-um structural features. So the group was undaunted by the challenge of making capillaries with 10-um diameters. “That’s a chip shot for us,” says Jeffrey Borenstein, the Draper group’s leader and the biomedical engineering center’s director.
In addition to Borenstein’s group, Vacanti enlisted the help of researchers at MIT working on a computer model of a blood vessel network. The 3-D computer model is divided into many horizontal layers, each of which serves as a blueprint. Following the blueprints, Borenstein’s group uses photolithography to draw blood vessel patterns on silicon wafers. These wafers serve as molds for the scaffold layers.
To produce the layers, the team pours a biodegradable polymer onto the molds. When the material solidifies, the casts are peeled from the molds and stacked up to form a plastic version of the computer model. A scaffold for a human-size engineered organ might include thousands of stacked layers, each 50 to 100 um thick.
If the team were making a liver, the stacked scaffold would be seeded with cells by pumping them into the organ through the vascular structure. The result should be layers of liver cells alternated with layers of blood vessel cells, thus ensuring that no liver cells would be too far from the blood supply. In a bioreactor, liver and blood vessel cells would grow and multiply, lining the sides of the scaffold’s capillaries. Eventually, the biodegradable scaffold would dissolve, leaving behind a liver with its own blood vessel network.
In the laboratory, the layered approach to organ building has had some success. One blood vessel structure, implanted in a rat, circulated blood with no leakage or obstruction of the flow, according to Vacanti. But he’s not ready to say when the process will yield a functioning human organ, or even if it ever will. “It could be a dead end,” he admits. “So all our energies are going into building a prototypic system that demonstrates that our concept is sound.”
Blood vessels for bypass
Engineer organs, of course, are not the only ones in need of blood vessels. Too often a person’s own organs could do with an off-the-shelf replacement artery. In fact, diseases of the blood vessels—small- and medium-bore arteries, in particular—account for most deaths in the United States. Usually when these are replaced, as in coronary artery bypass surgery, the substitutes are taken from elsewhere in the patient’s body. For instance, in bypass surgery, a leg vein may be removed and installed as an artery in the heart, a procedure that’s painful for the patient and time-consuming for the surgeon.
At Duke University Medical Center (Durham, N.C.), Laura Niklason fabricates blood vessels from pig and cow cells. Similar to real arteries, these conduits comprise a layer of so-called endothelial cells, which make up the inner lining of veins and arteries, surrounded by a layer of smooth-muscle cells, the kind that are not under conscious control. Both layers are embedded in a biodegradable tube-shaped scaffold hooked up to a small pump. The pump beats like a heart, pulsing nutrient fluid through the blood vessels-to-be.
The pulsing seems to help strengthen and differentiate the growing tissue, in keeping with the widely held belief that exposing tissue to the kind of environment it will find in the body helps it to grow properly. So far, results have been encouraging, reports Niklason, an assistant professor of biomedical engineering at Duke. Vessels grown from cow cells “have excellent mechanical properties,” she says. “They’re fully as strong as native arteries.”
Niklason is also looking into the effects of electrical stimulation on the growth and development of blood vessel cells. In living creatures, cardiovascular tissues receive electrical stimulation from the heart. “We don’t know yet, but mimicking these electrical environments may improve the growth and differentiation of tissues,” she says.
Ultimately, Niklason would like to develop a system for mass-producing human blood vessels, and that is clearly on the agenda at tissue engineering companies. ATS was granted a patent on the formation of “tubular structures,” including blood vessels, in 1999. But Niklason thinks it will be a decade or two before she—or anyone else—makes human vessels that are ready for the market. “If tissue-engineered blood vessels rupture or tear, the patient could die,” she says. “So they have to clear a much higher bar than some other types of tissue.”
Tissues to come
What will future tissue engineering firms be capable of making? Don’t expect too much too soon, experts caution. Even before manufacturing and automation come into play, scientists need to work out a number of critical issues, including finding good sources of cells for the engineered organs, dealing with the immune system’s response to them, building better scaffolds, and learning how to better preserve the organs once they’re constructed.
“It’s very difficult to replace what biology has taken millions of years to develop. So we have to be realistic,” ATS’s Ratcliffe says. “We should be able to return people to normal lives. But can we make things as good as or better than biology has made them? That’s a very long-term goal.”
So, in all likelihood, is the goal of growing vital organs for implantation in humans. Niklason thinks it could take tissue engineers as long as 40 years to clear all the scientific and regulatory hurdles on the road to products like tissue-engineered hearts and livers.
All the same, it may be only a matter of time until tissue engineers can replicate virtually every part of the human body. “It’s not outlandish to think it could be done at some point,” says Robert Langer, a professor of chemical and biomedical engineering at MIT and a pioneer in the field. “The issue is how long it takes to get there. I don’t know if it’ll be 20 years or 500 years.”