Bioengineered Corneas: The Promise And The Challenge

The promise of bioengineered replacements for diseased or damaged tissues has become a reality, notably for skin and cartilage. The article by Tsai et al. in this issue of the Journal demonstrates the promise of a nascent form of technology that may provide a new tool for reconstructing damaged ocular surfaces that previously would have been unrepairable.

Conditions such as the Stevens-Johnson syndrome, cicatricial pemphigoid, and chemical burns, among others, can severely compromise ocular surfaces and cause catastrophic visual loss in otherwise healthy eyes; such problems afflict thousands of patients in North America every year. The global burden of blindness from these disorders is probably much greater because of the many cases of trachoma and the higher incidence of trauma outside the United States and Canada. A common pathogenic feature of this seemingly diverse group of disorders is the depletion of the stem-cell population responsible for repairing the damaged corneal epithelium.

Conventional corneal transplantation is simply not successful in patients with this type of chronic surface problem. The donor corneal epithelium is gradually replaced, and the remaining transplanted corneal stroma, which is immunologically nonreactive, must ultimately be resurfaced with epithelial cells derived from the recipient's corneal stem cells. An absence or deficiency of stem cells allows, or may even stimulate, conjunctival-cell ingrowth and its accompanying neovascularization and inflammation, resulting in failure of the corneal graft. Therapies aimed at replenishing the stem-cell population have evolved in tandem with increased knowledge of the biology of corneal stem cells.

As with other self-renewing epithelial tissues, the corneal epithelium is maintained by an ordered, hierarchical replication of stem cells, which reside in the basal layer of the limbus. It is clear from both experimental studies and studies of human ocular disorders that depletion of the limbal stem-cell pool results in an abnormal corneal surface, which cannot be normalized without the introduction of a new source of stem cells. The most direct approach to correcting corneal-surface disease due to stem-cell deficiency has been transplantation of a segment of the stem-cell-laden limbus, either as a contralateral autograft or, more recently, as an allograft from either a cadaveric donor or a living relative. However, this procedure requires the harvest of approximately half (or more) of the limbus, and the procedure therefore jeopardizes the donor eye, whether the transplant is autologous or allogeneic. An ideal approach would use a bioengineered replacement tissue that replenished the pool of stem cells without endangering the corneal stem cells of the donor eye.

To engineer a successful replacement tissue, it is necessary to provide an ex vivo environment for stem cells that maintains both the replicative function of the tissue and its differentiated phenotype. However, it is also necessary to create a biologically compatible scaffold or stroma for the tissue. Tsai and colleagues chose human amniotic membrane as the scaffold on which to grow their replacement corneal surface. That was a logical choice, given the already widespread use of human amniotic membrane for ocular-surface reconstruction. Numerous studies (reviewed in the article by Tsai et al.) have demonstrated that amniotic membrane facilitates epithelialization without allowing fibrovascular growth, supports epithelial-cell differentiation, is gradually resorbed in vivo, is nonantigenic, and contains extracellular-matrix components resembling those of conjunctival basement membrane.

Indeed, the successful outcomes in the patients into whose eyes Tsai et al. transplanted limbal epithelial cells cultured on amniotic membrane, as well as the patients who underwent a similar procedure in our own study, suggest the promise of the technique for patients with limbal stem-cell deficiency or disease. An advantage of the procedure is the reduction in the amount of donor limbal tissue required (2 mm² in both studies), which minimizes the possibility of damage to the donor eye. Combining the expansion of cells ex vivo with cultivation on a modified amniotic membrane has the advantage of ensuring a compatible extracellular matrix for the graft, thus increasing its durability and manipulability. This method is an advance over earlier attempts at the use of engineered corneal surfaces, in which fragile sheets of epithelial cells, with no substantial underlying stromal support, were transplanted.

Despite the promise of the procedure described by Tsai et al., challenges lie ahead. The difficulty remains of identifying stem cells and ensuring that enough of them are present in the bioengineered tissue to repopulate the ocular surface. Current techniques for identifying corneal stem cells rely on their ability to take up various substances or to serve as progenitor cells for large colonies in culture -- characteristics difficult to identify once the cells have been incorporated into a tissue prepared for human transplantation. The identification of cell-surface markers, such as those identified for keratinocyte stem cells, will help address this issue. A second challenge is to optimize the scaffold on which the replacement corneal surface is grown. Although amniotic membrane is currently the most widely used material for this purpose, a bioengineered laboratory equivalent for use as corneal stroma would minimize the possible risk, albeit remote, of the transmission of an as-yet-unidentified infectious disease through use of an allograft.

In addition, the persistence of transplanted donor stem cells, and their reintegration into the limbus of the recipient eye, have yet to be demonstrated in patients who have received bioengineered grafts. This issue is particularly important if the future of bioengineered corneal-surface tissue lies with allogeneic donor cells. Although the persistence of donor-specific DNA sequences has been detected as long as 30 months after conventional transplantation of allogeneic limbal tissue, no such analyses have been performed in recipients of bioengineered grafts. Evaluation of bioengineered composite skin grafts has revealed that, despite the long-term persistence of the allogeneic epithelial cells, the donor cells are slowly replaced by host cells. This finding suggests that, in some cases, the cultured allogeneic graft acts as a biologic dressing, providing the cytokines and growth factors required for tissue repair in the appropriate temporal sequence. Finally, answering the question of what will happen to the tissue or organ decades after it is implanted will require years of careful observation and follow-up.

Bioengineered or cultured tissue products are currently being produced to replace other tissues, and the progress with corneal-surface replacement indicates that such products are likely to revolutionize the treatment of many epithelial and even visceral diseases. Heartened as we should be by the rapid progress in this area, we must recognize that there is still much to be learned before bioengineered organs will be routinely available.

Ivan R. Schwab, M.D.

R. Rivkah Isseroff, M.D.

University of California, Davis, Medical School

Sacramento, CA 95817

Copyright © 2000 by the Massachusetts Medical Society.