There is currently an explosion of interest in tissue engineering solutions for orthopaedic problems that have no reliable treatments. The excitement is predicated on work that has shown that the potential for the regeneration of many musculoskeletal tissues may be greater than we had ever expected and that cells from many of the musculoskeletal tissues can be isolated and grown in vitro. (Figure 1)

           Most non-osseous tissues are incapable of spontaneous regeneration and heal by forming dysfunctional fibrous tissue, or scar. In some nonvascular musculoskeletal tissues - such as articular cartilage - there may be an absence of healing altogether. The field of tissue engineering is based upon recent work that has shown that certain cells can proliferate and maintain their phenotype when cultured on specific 2-dimensional substrates or in specific 3-dimensional porous matrices and gels in vitro. Using these techniques it may be possible to form tissues in vitro and then implant them into a defect. Initially, the term "tissue engineering" was used principally to describe tissue produced in culture by cells seeded in porous absorbable matrices.1 More recently, however, the scope of tissue engineering has been widened to include the implementation of porous matrices - alone or seeded with cells - as implants to facilitate tissue regeneration in vivo.2

In Vivo Tissue Engineering

          Tissue engineered for clinical use can be produced in vitro for subsequent implantation or grown in vivo. In vitro tissue synthesis allows for analysis of structure and function prior to implantation, but there are several disadvantages to the in vitro approach. Growth in vitro deprives the tissue of the physiological mechanical environment that can profoundly influence the architecture of the tissue as it is forming. Because the mechanical environment that exists during the formation of musculoskeletal tissue in vivo is not well understood, it is not yet possible to recreate such an environment in vitro during the engineering of most tissues. Moreover, implanted-engineered tissue must become mechanically coupled to the surrounding host tissue. Union of the implanted tissue with the host organ requires degradation and new tissue formation at the interfaces of the implant with the host tissues.

The Development of Matrices (or Scaffolds) for In Vivo Tissue Regeneration

          Matrices for engineering the soft musculoskeletal tissues have included myriad synthetic (e.g., polylactic acid and polyglycolic acid) and natural polymers (e.g., collagen and fibrin ).3 The microstructure of the matrices needs to be able to accommodate the infiltration by parenchymal cells and support their functions. In this regard a porous structure is generally necessary. The required percent porosity and pore diameter, distribution, and orientation, might be expected to vary with tissue type. The chemical composition of the matrix is important with respect to its influence on cell adhesion and the phenotypic expression of the infiltrating cells. Moreover, because the objective is the regeneration of the original tissue, the scaffold needs to be absorbable. The degradation rate of the material generally may be determined based on the rate of new tissue formation and the normal period for remodeling of the tissue at the site of implantation. Of course, it is important to consider the effects of moieties released during degradation of the matrix on the host and regenerating tissue.

           A matrix can play several roles during the process of regeneration in vivo:

1)	It can serve to structurally reinforce the defect site so as to maintain the shape of the defect and prevent distortion of surrounding tissue.
2)	The matrix can serve as a barrier to the ingress of surrounding tissue that may impede the process of regeneration.
3)	The matrix can serve as a scaffold for migration and proliferation of cells in vivo or for cells seeded in vitro.
4)	The matrix can serve as an insoluble regulator of cell function through its interaction with certain integrins and other cell receptors.

          One design approach - the one being employed in the Orthopaedic Research Laboratory of the Brigham and Women's Hospital - has been to use matrices that serve as analogs of the extracellular matrix (ECM) of the tissue to be engineered. This concept recognizes that the molecular composition and architecture of the ECM displays chemical and mechanical properties required by the parenchymal cells and the physiological demands of the tissue. The ORL approach of using analogs of ECM as implants for the regeneration of the soft musculoskeletal tissues and nerve has employed porous collagen-glycosaminoglycan (GAG) copolymers. (Figure 2) Previous studies in the laboratory of our collaborator at MIT, Professor I.V. Yannas, found that regeneration of dermis in animals and human subjects and the reconnection of axons of cells in ruptured peripheral nerves in rats required certain tissue-specific pore characteristics and matrix degradation rates.4, 5 The MIT collagen-GAG material for dermal regeneration has been approved by the US Food and Drug Administration under the trade name Integra for therapeutic use in humans.

          Many methods have been employed to produce matrices for tissue engineering. In our own work, type I or type II collagen is precipitated from acid dispersion in the presence of chondroitin 6-sulfate or other glycosaminoglycans. The coprecipitate suspension can be injected into a tube or spread on a pan for immersion into a coolant bath and freeze-dried, to produce a porous architecture. The geometry of the mold for the collagen slurry, the temperature of the coolant bath, and the rate of immersion of the collagen suspension into the bath are determinants of the ice crystal formation and therefore the pore characteristics of the resultant sponge-like material. The matrices are subsequently cross-linked to control degradation rate and mechanical properties. These collagen-GAG sponge-like materials are nominally 95% porous and can be produced with an average pore diameter in the range of 30-120 µm.

           Current laboratory studies in the ORL are investigating these collagen-GAG matrices for tissue engineering of articular cartilage6, 7, meniscus8, 9, ligament, tendon, intervertebral disc10, and gingiva. This work is in collaboration with S.D. Martin, M.D. and R.M. Ozuna, M.D, of the orthopaedic surgery faculty at the Brigham and Women's Hospital. Orthopaedic residents (including Martha M. Murray and Sonu Ahluwahlia) and MIT graduate students (including doctoral students, Cynthia R. Lee and Dawn Hastreiter) are pursung this investigative work in their theses. ORL research faculty including Hu-Ping Hsu, MD and Sonya Shortkroff, MS and the technician staff and research fellows are also engaged in this research. These studies are employing cell and tissue culture and animal models to investigate the influence of the chemical composition, pore characteristics, mechanical properties, and degradation rate of the matrix on specific biological processes underlying tissue regeneration. The effects of selected growth factors on cell-matrix interactions are also being studied. Methodology includes histology, immunohistochemistry, electron microscopy, biochemistry and molecular biology techniques, and mechanical testing.

          The multidisciplinary nature of tissue engineering requires a team approach, engaging orthopaedics surgeons, engineers, and biomedical scientists. Moreover, the resources for such studies cannot all be found at the Brigham and Women's Hospital, thus requiring a multi-institutional approach in partnership with Harvard Medical School and MIT. Such a conjoint endeavor benefits from involvement of medical and engineering students along with orthopaedic residents and fellows. This is critical for advancing the education of the next generation of clinicians and laboratory investigators in tissue engineering approaches to orthopaedic problems.

Recent Work

The three basic components of tissue - matrix, cells, and soluble regulators--are the elements that can be manipulated in strategies to engineer tissue in vivo, or in vitro for subsequent implantation. Decisions as to which elements might be required for regeneration of tissue in vivo can be guided by an understanding of the deficits of the natural healing processes that prevent regeneration. Specific laboratory experiments can help characterize the factors that determine the quality of the tissue engineered.

Matrix
          The extracellular matrix can influence the behavior of cells in engineered tissues. Specific matrix characteristics favor certain cell behaviors. We investigated the response of chondrocytes to variations in collagen type and pore size in the collagen-GAG matrices on which they were cultured.6 There was a dramatic difference in the morphology of the cells that grew in the type I and type II collagen matrices. The cells in the type II collagen matrix retained their chondrocytic morphology and synthesized glycosaminoglycans, while in the type I matrix the chondrocytes displayed a fibroblastic morphology and limited synthesis of glycosaminoglycans. (Figure 3) The morphology of the cells grown in type I collagen matrices with a small pore diameter more closely resembled chondrocytes initially but became more fibroblastic with time.

          The marked influence of collagen type and pore characteristics on the phenotypic expression of seeded chondrocytes demonstrates that specific matrix characteristics favor certain cell behavior. This finding supports the concept that engineered tissues may more closely resemble the tissue they are mimicking if they are produced on matrices composed of analogs of the extracellular matrix of that tissue. For example, engineered articular cartilage may be best produced on a matrix made of type II collagen.

Cells
          Investigations using explants of human and animal articular cartilage, meniscus, and ligaments have demonstrated the capability of the parenchymal cells to migrate into the collagen matrices in vitro. This work is laying the foundation for the development of implants to use in future animal investigations.

           In another recent study of the injury and healing responses of several musculoskeletal tissues we have found that a percentage of nonvascular cells in ligament11, meniscus8, and intervertebral disk10 contain the contractile actin isoform normally found in smooth muscle cells, alpha-smooth muscle actin (SMA), and the percentage of such cells may increase after injury. SMA-positive cells facilitate wound healing in skin by contracting and narrowing the wound; however, this mechanism may be detrimental in the healing of musculoskeletal tissues where contraction of injured tissue can lead to retraction of the ruptured ends of the tissue, thereby creating a larger gap.

          The positive role of SMA-positive cells in normal musculoskeletal tissues may be to help impart and maintain the specific architecture of the tissue. We have found that cells isolated from these tissues are capable of contracting a collagen analog of extracellular matrix in vitro. The ability to inhibit this cell-mediated contraction in injured tissues might improve apposition of the wound surfaces and improve healing.

          These concepts impact tissue engineering techniques because the contractile phenotype can cause contracture of a cell-seeded matrix, distorting the pore openings and thereby impeding the cell proliferation and matrix synthesis required for regeneration.

Soluble Regulators
          We have found growth factors that modulate the expression of the SMA isoform. These growth factors may need to be incorporated in the tissue engineering strategy in order to limit unfavorable contraction of engineered tissues.

Conclusions

With the advent of cell-based therapies, there has come a broadening of the scope of tissue engineering to include the implantation of exogenous cells alone. Moreover, in many respects the implantation of autogenous and allogeneic tissue grafts also falls under the rubric of tissue engineering. In this light, tissue engineering is not so much of a new discipline but a new twist on a decades-old practice, and the impending "revolution" more of an "evolution" of orthopaedic surgery. Regardless of how the present work is viewed, the outlook for the future is encouraging in that it is likely that new treatment modalities will soon become available for the management of a wide variety of orthopaedic problems.

Myron Spector, PhD is Director of Orthopaedic Research at Brigham and Womens's Hospital, and Professor of Orthopaedic Surgery (Biomaterials) at Harvard Medical School

Address correspondence to:
Dr. Myron Spector; Department of Orthopaedic Surgery; Brigham and Women's Hospital; 75 Francis St.; Boston, MA 02115
e-mail: mspector@rics.bwh.harvard.edu