When trying to identify the beginnings of “tissue engineering” as a defined scientific field, the first major scientific meeting devoted to “tissue engineering” comes to mind. This meeting took place in 1988 at Granlibakken, Lake Tahoe, California. When looking through the proceedings of this meeting, one can see that many of the fundamental challenges of tissue engineering, still unresolved today, were articulated back at that time. As practiced today, the field of tissue engineering is based on a blending of engineering principles in the materials sciences with the principles of cell and tissue biology to create functional replacements for aged, traumadamaged, or diseased human tissues and organs.

The driving force behind the field of “tissue engineering” is the desire to replace defective human tissue with artificial substitutes to improve the quality of human life. This desire is not unique to our times but can be traced back several thousand years. The earliest reports of the use of artificial materials in wound healing are contained in the Edgar Smith papyrus. This historic document mentions the use of sutures and woundclosure devices around 4000 B.C. and the use of metals in bone repair around 2000 B.C. In ancient times, Native Americans apparently used silk and hair as suture materials.

In 1900 natural tissues like amnion and placenta were suggested as skin replacements, an approach that was obviously doomed to fail due to immunological reactions. The general lack of understanding of the immune response at the beginning of this century forced researchers and clinicians to abandon attempts to use natural tissues as medical implants and contributed to a focus on synthetic materials. Advances in metallurgy facilitated the introduction of metals for bone reconstruction and dental prosthesis, and the large number of casualties during World War I resulted in the establishment of stainless steel and other metals as the preferred orthopedic implant materials.

The success of tissue engineering rests on the ability to direct specific cell types to multiply, migrate, and express normal physiologic behaviors in order to yield a cellular organization that performs the functions of the desired tissue. For example the engineering of living bone to repair skeletal defects has focused on growing osteoblasts—the cells responsible for bone formation—on degradable polymer matrices in vitro. The polymer matrix initially serves as the scaffold for bone-cell proliferation and maturation. Ideally the cells form a bonelike tissue that after implantation is fully integrated into the patient's own bone, thus repairing the bone injury or defect. Soon thereafter, its function complete, the polymer scaffold resorbs away. Readily apparent is the crucial role the scaffold material occupies in tissue engineering since it serves as the template for cell growth and tissue formation. It is the interaction between the cell and the material that dictates whether the cells will proliferate, mature, and express the desired tissue characteristics.

A critical issue facing the biomedical industry today is the availability of raw materials for medical-device manufacture. Furthermore it is now recognized that the materials base of the medical-device industry is outdated. Metals and various industrial plastics (e.g., polysiloxanes, polyurethanes, Dacron®, Teflon®, polyethylene) are the most commonly used biomaterials. These biostable, synthetic implant materials lack the biological sequences and patterns crucial to normal cell function and can trigger aberrant cell responses. Likewise few degradable polymers are available to the medical-device designer and tissue engineer, representing another limitation of the materials base of the medical-device industry (Table I).

Tissue engineering offers the possibility to create completely natural tissue and replace failing or malfunctioning organs. In many cases, biocompatible, biodegradable polymers are utilized to either induce surrounding tissue and cell ingrowth or to serve as a temporary scaffold for transplanted cells to attach, grow, and maintain differentiated functions. Various processing techniques have therefore been developed to fabricate polymers with specific properties to meet the needs of a particular organ.

Polymer scaffolds must possess unique physical and chemical properties for specific applications and must satisfy some basic requirements for tissue engineering. These scaffolds may be implanted without cells, and the regeneration depends on ingrowth of surrounding tissue to such materials—a process known as tissue induction. Alternatively cells may be seeded into a porous polymer. The cell-polymer construct is then transplanted. In either case, one essential criterion for the scaffold is biocompatibility—that is, the polymer scaffolds and their degradation products should not invoke an adverse immune response or toxicity.

Because of the problems associated with long-term implants, such as infection, fibrous tissue formation, and the possible need for retrieval, the role of polymer scaffolds should only be a temporary one. The degradation rate is optimized to allow transplanted cells to proliferate and secrete their own extracellular matrix (ECM) while polymer scaffolds can vanish when necessary to leave enough space for new tissue growth. Biodegradability of a polymer is determined by its composition, molecular weight (MW), MW distribution, degree of crystallinity, and environmental conditions such as temperature and pH. Mechanical loading of the scaffold may affect its degradation.

Novel applications exist for biomedical materials that can undergo transitions in material properties in situ—that is, at the site of implantation in the body. Such transitions in polymeric biomaterials can be accomplished by crosslinking a material in situ, by heating or cooling to induce thermal transitions, and by precipitating polymer from solution in situ. This article will point out the need for materials that can be induced to undergo such transitions in situ and will describe selected tissue-engineering approaches that have been employed for this purpose.

Delivering materials to the body in one form and utilizing them in another form after a transition at the site of implantation has, generally speaking, two potential advantages: (1) the ability to match the morphology of a material implant to a complex tissue shape and (2) the ability to deliver a large device through a small hole in the body. With respect to the former motivation, tissue shapes in the body have an enormous range of complexity: a region of a blood vessel—for example with a curved central axis, a diameter that varies along the length, an eccentric diseased plaque, and numerous side branches. One could develop a variety of implant shapes—arterial stents in this example (for use in structurally supporting arteries after balloon angioplasty, thereby creating a larger cross section for blood flow above the diseased plaque)—and then select the most appropriate implant shape after detailed imaging of the tissue site. One can alternatively attempt to employ some material transformation to deliver a precursor to the final shape of the implant, utilizing the tissue shape to obtain the proper final implant morphology. With regard to the second motivation, it may be desirable to deliver a large object through a small hole, utilizing material transformations. Advances in surgery have focused on manipulating (cutting, coagulating, suturing, stapling) large tissue sites through small holes in the body via minimally invasive surgery. Using such approaches, it has become possible to perform many complex surgical procedures in the joints, abdominopelvic cavity, thoracic cavity, and nasal sinuses, for example, using surgical instruments that are manipulated through surgical access holes less than 1 cm in diameter. Even procedures as complex as coronary-artery bypass surgery have been performed in this way. It still remains generally impossible however to implant devices in the body through such holes unless these implants are very small. If such devices were for example able to be delivered as liquids and then shaped into devices at the implant site, such minimally invasive surgical-device placement could be envisioned.

Recent advances in materials science and biotechnology have given birth to the new and exciting field of tissue engineering, in which the two normally disparate fields are merging into a profitable matrimony. In particular the use of biomaterials capable of initiating new bone formation via a process called osteoinduction is leading to quantum leaps for the tissue engineering of bone.

The classic work of Marshall R. Urist and A. Hari Reddi opened the field of osteoinductive biomaterials. Urist discovered that, upon implantation of devitalized, demineralized bone matrix in the muscle of experimental animals, new bone formation occurs within two weeks, a phenomenon he described as bone formation by induction. The tissue response elicited by implantation of demineralized bone matrix in muscle or under the skin includes activation and migration of undifferentiated mesenchymal cells by chemotaxis, anchoragedependent cell attachment to the matrix, mitosis and proliferation of mesenchymal cells, differentiation of cartilage, mineralization of the cartilage, vascular invasion of the cartilage, differentiation of osteoblasts and deposition of bone matrix, and finally mineralization of bone and differentiation of marrow in the newly developed ossicle.

The osteoinductive ability of the extracellular matrix of bone is abolished by the dissociative extraction of the demineralized matrix, but is recovered when the extracted component, itself inactive, is reconstituted with the inactive residue—mainly insoluble collagenous bone matrix. This important experiment showed that the osteoinductive signal resides in the solubilized component but needs to be reconstituted with an appropriate carrier to restore the osteoinductive activity. In this case, the carrier is the insoluble collagenous bone matrix—mainly crosslinked type I collagen.

The majority of musculoskeletal injuries involve the soft connective tissues including cartilage, meniscus, tendon, and ligament. In contrast to bone tissue, which regenerates after injury, these soft tissues tend to heal poorly through formation of weak, disorganized scar tissue. In fact certain soft-tissue injuries cannot heal on their own so surgical repair or complete reconstruction is required to restore function. Biological grafts (autografts and allografts) and “permanent” synthetic prostheses are currently used to reconstruct large soft-tissue defects. Autografts (a patient's own tissue transferred to another site) can be successful, but as you “rob Peter to pay Paul,” complications at the donor site can be significant. Allografts (transplants from another human) are in short supply and carry the risk of rejection or disease transmission. In general, permanent synthetic prostheses fail to withstand long-term mechanical loading associated with the musculoskeletal system. New strategies are clearly needed to manage musculoskeletal soft-tissue reconstruction.

Several tissue-engineering strategies are potentially useful to enhance repair of these troublesome injuries or to induce regeneration. One interesting approach (developed for skin and cartilage regeneration) combines scaffolds, cells, and cell signals to grow living tissue analogues in the laboratory. Scaffolds are natural or synthetic biomaterials upon which the cells attach and grow. The cells may be fibroblasts, chondrocytes, etc., depending on the specific application. Influential cell signals can include attachment factors, growth factors, and mechanical loads.

Our laboratory is developing a similar approach for ligament reconstruction. Ligaments are the strong, flexible bands of collagenous tissue that connect bone-to-bone to provide a delicate balance of stability and flexibility to the joints of the body. The anterior cruciate ligament (ACL) is the primary stabilizer of the knee and is frequently injured in sporting activities and accidents. Serious ACL injuries can lead to disability and progressive degeneration of cartilage and meniscus that normally serve as shock absorbers in the joint. Anterior-cruciate-ligament ruptures do not heal without surgical intervention, due to a poor intrinsic healing response and poor vascularity. To restore normal joint function, ruptured ACLs can be surgically reconstructed using a patellar-tendon autograft or allograft. While significant progress has been made in understanding ACL anatomy, structure, biomechanics, and healing, there is still no biological graft or biomaterial ideally suited for ACL reconstruction.

Tissue engineering involves the application of physical and life sciences to develop functional substitutes for dysfunctional organs or tissue structures. From an engineering standpoint, tissues contain two basic components—the cells that are organized into proper units, and the material surrounding the cells, called the extracellular matrix (ECM). A third, frequently overlooked feature essential to the maintenance of the activity of the engineered tissue is the three-dimensional architecture of the cell-matrix composite.

A comprehensive review of the scope and impact of tissue engineering has previously appeared. Tissue-engineered devices have the potential to reduce the annual health-care cost related to tissue loss and end-stage organ failure to the order of $400 billion, eight million invasive surgical procedures, and 65 million hospital days. A common approach to engineer a functional tissue is to place cells derived from a healthy organ or tissue type (identical or similar to the dysfunctional tissue/organ) on or within matrices analogous to host-tissue ECM. These systems can then be enclosed in appropriate membranes that isolate cells from immune rejection following implantation or can be transplanted directly with the administration of drugs that prevent the immune rejection. Another application of these systems is for extracorporeal (outside the patient's body) device support of a dysfunctional organ. In either instance, the success of the engineered tissue depends critically on the interactions of cells with the tissue analogues. The objective of this article is to outline the simplest matrix-design parameters to control these interactions. While organs are comprised of very different tissue types, for the sake of simplicity, this article is primarily pertinent to the tissue engineering of one major organ, the liver. The choice of this tissue type is intended to serve as a comprehensive generalization of many different cell types since in the diversity and complexity of its activities, the liver has few parallels. The development of an artificial liver is also critically awaited, as in the United States alone, 35,000 people, including the many wait listed for the exorbitant liver organ transplants ($300,000), die each year of chronic liver disorders. In many other countries, liver disease is the second leading cause of death.

The nervous system of the human body can be divided into two parts: the central and the peripheral nervous system. While the central nervous system (CNS) includes the brain and the spinal cord, the peripheral nervous system (PNS) consists of all the nerve branches exiting the spinal cord or brain stem that process information from our environment to the CNS and vice versa. The main extension that passes all incoming information to the next neuron or a peripheral target is called an axon. Axons can be as long as 1 meter, as in the case of axons sending electrical signals to the toes. All incoming information is received by smaller structures called dendrites. Axons are wrapped with glial cells, Schwann cells in the PNS, and oligodendrocytes in the CNS. These cells are responsible for the deposition of myelin, a powerful insulator. Each glial cell is covering an axonal segment of about 1-2 mm. An unmyelinated gap separates two glial cells. Action potentials are generated at those gaps. This electrical signal jumps from one unmyelinated gap to the next, giving rise to conduction velocities as fast as 1 m/s. In the PNS, axons are grouped together into fascicles, consisting of a layer of connective tissue surrounding bundles of axons. Several of these fascicles form a peripheral nerve surrounded by another connective tissue layer named epineurium. The epineurium is responsible for the integrity of the overall structure of a peripheral nerve. While the PNS shows regenerative capabilities, this is in general not the case for the CNS. The mechanisms explaining the poor regeneration capabilities of the CNS have started to be elucidated. The expression of inhibitory molecules in the mammalian-adult nervous system seems to be an essential component in the inherent lack of CNS regeneration. Examples for possible applications of materials systems addressing certain issues of PNS and CNS regeneration are presented in the following sections.

Soluble signaling proteins called growth factors execute critical functions during the formation of specialized tissues throughout the developing embryo. When growth factors are provided to adult animals, they often encourage regeneration or repair of organs damaged by disease or trauma: Basic fibroblast growth factor (bFGF) and transforming growth factor ß1 (TGF-ß1) encourage wound healing hematopoetic growth factors stimulate the production of blood cells, bone morphogenetic proteins (BMPs) induce bone formation, nerve growth factor (NGF) enhances the survival of degenerating cholinergic neurons, and angiogenic growth factors activate new blood-vessel growth. Our understanding of the role of growth factors in development and regeneration should continue to expand dramatically over the next decade, inasmuch as new molecules (and new activities for known molecules) are appearing at a rapid rate.

Protein growth factors may be useful in augmenting the new approaches for tissue engineering. Modern biotechnology permits the large-scale manufacture of highly purified proteins so that large quantities can be produced for use in humans. However proteins are often exceedingly difficult to administer, particularly if sustained levels are required. Most protein growth factors have short half-lives after intravenous injection, with their biological activity lasting only a few minutes in the circulation, so that injection must be repeated frequently to obtain sustained blood levels (Table I). Since these molecules are large, they penetrate tissue barriers, such as the capillary wall, very slowly. In addition, growth factors are extremely potent, often possessing biological activity at a number of tissue sites throughout the body. Therefore systemic administration can lead to toxicity. In view of these difficulties, new methods for growth-factor delivery are needed. The most promising new methods involve polymers, which can be engineered to provide precisely controlled, prolonged growth-factor delivery at a localized site.