In unilateral LSCD, or in bilateral LSCD, where a small portion of healthy limbus can be used as donor tissue for ex vivo expansion, autologous limbal grafts are advised [30, 32]. On the contrary, in total LSCD when the limbus is completely destroyed in both eyes, limbal tissue taken from a deceased donor or from a living relative can be used. In the literature, contrasting results have been reported on the use of allogeneic keratolimbal grafts, with an overall success rate of 73 % [1]. Both clinical successes and failures have been observed in the presence of systemic immunosuppressive therapy [6, 16, 23], while positive clinical results have been reported in the absence of immunosuppression [20, 33] and/or in the absence of allogeneic cell survival [2, 14]. In most cases, however, the interpretation of results has been hampered, either by the lack of a proper genetic evaluation of the presumptive long-term engraftment of allogeneic limbal grafts or by the inadequate length of follow-up. In the absence of demonstrated surviving donor cells, a possible explanation for clinical success is that patients with non-total limbal stem-cell deficiency have been included, and the grafted allogeneic limbal cells might have induced modification of the microenvironment and promoted proliferation of the patient’s own dormant stem cells, whose progeny gradually replaces donor cells. While remaining in situ in the injured eye, these limbal cells are evidently unable to generate corneal epithelium, either because of the lack of a suitable microenvironment for multiplication or because of fibrotic obstruction to their migration over the cornea. This would explain the mixed population of donor and recipient corneal cells observed at short-term follow-up. These findings are consistent with reports showing that clinical improvement observed following allogeneic keratolimbal grafts does not necessarily correlate with the long-term survival of donor cells [2, 14]. Similarly, cultured allogeneic epidermal keratinocytes do not engraft permanently, but provoke epidermal regeneration in partial-thickness skin burns, presumably by stimulating residual hair follicle stem cells [3].

Limbal stem-cell grafting is indicated to treat limbal stem-cell deficiency (LSCD) [8, 36]. As said above, LSCD includes heterogeneous diseases where the limbus has been damaged. The eyelids, conjunctiva, corneal stroma, nerves and endothelium, and immune and lacrimal systems can also be involved. Scrupulous step-by-step reconstruction should be planned, treating the structures involved separately, to prepare the best recipient bed for the cultivated cells. Eyelid malposition and malocclusion should first be treated. Conjunctival symblepharon should be then addressed using the appropriate procedures. Once the eyelids and conjunctiva have been treated, tear film and inflammation should be carefully evaluated. The minimum of tear film and the maximum inflammation allowing the successful long-term survival of the grafted stem cells are not clear. In our previous clinical trials [31, 32], we excluded patients with Schirmer’s test below 5 mm/5 min, but this was arbitrarily chosen, and one might suggest that the quality of tears might be even more important than the quantity. Unfortunately, at present there is still no valid method for its assessment. We do not include in our clinical protocol for limbal transplantation patients showing severe active inflammation. As for tear film, we are still far from having reproducible clinical assessment and inflammation grading, with the exception of redness scoring.

LSCD diagnosis is based on the evidence of a previous insult (cause) and peculiar clinical features (signs) and is eventually confirmed by instrumental tests [8, 36]. The causes of LSCD are shown in Table 16.1.

Various protocols for the cultivation of limbal stem cells for transplantation have been proposed and recently reviewed by Shortt et al. and Joe and Yeung, including methods to extract cells from the biopsy (mechanical disruption or enzymatic dissociation), substrates and carriers (fibrin sheet, amniotic membrane, polymers, contact lenses, collagen), or mediums with animal-derived components or xeno-free [18, 36]. Although good clinical outcomes have been reported with all of these different culture procedures, few studies have evaluated the clonal characteristics of the cultivated cells and their proliferative potential. When dealing with stem-cell-based therapies for diseases involving cell-renewing tissue, it should be mandatory to demonstrate the presence, survival, and concentration of stem cells in culture and in the graft and validate the procedure under GMP conditions [4, 28].

We previously showed, analyzing the proliferative potential and cloning characteristics, that corneal stem cells are segregated in the limbus, while conjunctival stem cells are uniformly distributed in the bulbar and forniceal conjunctiva. Moreover, conjunctival epithelial cells and goblet cells derive from a common bipotent progenitor [26]. We also showed that autologous limbal stem cells, cultivated on fibrin and 3 T3 feeder layer, maintain their properties and are able to restore corneal integrity in severe limbal stem-cell deficiency [31]. We later confirmed the long-term stability of the results, up to 10 years, and validated the procedure, comparing clinical results with the level of expression of ΔNp63α in culture [5, 29, 32]. Clinical success was statistically associated with the percentage of p63-positive cells in culture. Cultures in which p63-bright cells made up more than 3 % were associated with successful transplantation rate close to 80 %. In contrast, cultures with less than 3 % were associated with poor results, with successful transplantation in only 10 % of patients. On the basis of these data, only cultures that contain more than 3 % ΔNp63α cells are now grafted on patients.

We hereafter report our protocol: (i) biopsy, (ii) stem-cell expansion in culture, (iii) grafting, and (iv) postoperative management.