Author Response: Donor Cell Survival in Corneal Grafts

We welcome the comments regarding our two recent articles ^1,2 detailing our observations of donor and recipient endothelial cell populations in the human corneal graft and are pleased to have the opportunity to further discuss and clarify our reported findings. 

In both studies, we used fluorescent in situ hybridization (FISH) of the sex chromosomes in sex-mismatched corneal grafts, a mode of investigation pioneered by Wollensak et al. ^3,4 In the first such study published by that group in 1999, FISH analysis of paraffin-embedded sections from 14 corneal buttons revealed the persistence of a proportion of donor keratocytes (15%–26%) for at least 4.5 years, whereas no donor endothelial cells were detected in any sample. ³ The authors acknowledged an individual variability in replacement of donor keratocytes but observed early and complete replacement of donor endothelium (and epithelium). 

In our first study, using a broad sample of 52 buttons and the same technique, we demonstrated not only the presence of donor endothelial cells, but that they could be found in large numbers (up to 100% donor endothelium) and could remain for long periods in the graft (up to 32 years). ¹ Our results were similar for donor keratocytes. To overcome the sampling limitations inherent in the analysis of thin tissue sections (particularly regarding corneal endothelial cells), in our second study we devised a technique for two-dimensional analysis of the origin of endothelial cells. ² This second study confirmed the results of the first—namely, that donor endothelial cells could persist for long periods within the (clear) graft and that there is substantial variability in cell replacement. As suggested, our present findings require that the results of the 1999 study ³ be viewed in a broader context. 

Rather than perpetuating a prejudice regarding donor cell survival, our intent in these studies was to highlight that donor cells could persist in the graft without any obvious tendency toward eventual replacement by recipient cells and that there is a large variability in cell persistence and replacement. Although indeed some samples revealed that complete donor endothelial cell replacement was compatible with clear grafts, opposing evidence could also be cited, since other clear grafts had significant proportions of surviving donor endothelial cells (cases 26, 40, and 49 in the first study ¹ and cases 4, 8, and 19 in the second study ² ). Evidently, neither dogma—that of donor cell survival or that of eventual donor cell replacement—can be upheld. 

The answer to the question of the fate of cells in the corneal transplant is likely more nuanced than has been previously recognized. It is our hope that the long-standing debate can now be shifted from “do the transplanted cells survive?” to “what are the factors underlying the great variability in donor cell survival?” Once we concede that the question of donor cell survival is a complex one, we can begin to focus on elucidating the factors influencing cell motility after transplantation as a first step toward improving transplant survival. 

With regard to Professor Wollensak's specific comments:

1.  
   In both studies, surgeons were instructed to send corneal buttons in sex-mismatched cases, where donor sex was known from records. In several cases with long postoperative intervals, donor sex was unknown because of the unavailability of records. In these cases, donor sex was either confirmed by signals in stromal cells (first study), or the sample was excluded from the study population (second study; 13 of 49 samples were unsuitable for final analysis, one reason being our inability to confirm donor sex).
2.  
   In the second study, we found that the proportion of donor endothelial cells was reduced in grafts older than the median age of 5 years compared with younger grafts, with this result just reaching significance. As suggested, a very large number of samples would be necessary to detect significant relationships in which a wide variability exists. Also, we concur with the suggestion of a multiplicity of causes of donor cell attrition after transplantation. As mentioned in our first article, a limitation of our study spanning multiple centers and with long postoperative intervals was the scarcity of postoperative records. In addition, differences in terminology and descriptions of postoperative complications were used by different surgeons. For these reasons, we could not reliably determine which grafts had rejections, but it is reasonable to assume, on the basis of data from the Swedish Corneal Register, ⁵ that several of the decompensated grafts in the studies would have had an episode of rejection.
3.  
   In our second study, we indicated that, in the 1999 study by Wollensak and Green, ³ only failed grafts were examined (and as the authors indicated, three of them were compensated). We wished to highlight the cases in our studies that were clinically successful, totally transparent grafts, without recurrence of dystrophy. In these cases (five in the first study and four in the second), the grafts were removed for refractive reasons (astigmatism).
4.  
   One argument often cited in support of the donor cell replacement hypothesis is that of recurrence of dystrophies in the graft. Although dystrophies often recur in grafts during the early postoperative years, long periods until recurrence have been reported, ⁶ and notably, in our second study, three cases of lattice dystrophy were included, with recurrence after 13, 18, and 30 years. ² In addition, with the naive assumption of a constant rate of annual donor endothelial cell loss, we found that 37% of grafts in our first study and 19% of grafts in our second study had a donor endothelial cell loss rate of 5% or less per year. Based on our findings, we suspect that long-term survival of donor cells is not only a feature in a few isolated cases, but is a more prevalent phenomenon. We acknowledge that evidence can be found in favor of opposite viewpoints and reiterate our earlier statement that we must instead address the question of why cells persist for long periods in some cases and not in others.
5.  
   Our results appear to support the notion that endothelial cell density (or total cell number) may be more important than cell origin in influencing graft decompensation. The factors mentioned may play an important role in producing the patterns of cell replacement observed in our second study.
6.  
   Cases of complete or nearly complete replacement of donor endothelium by recipient cells can be compatible with a transparent and successful graft, as we have noted in both our studies. ^1,2 As indicated by Professor Wollensak, in such cases, endothelial cell density may remain well above the threshold for proper endothelial function. In addition, endothelial cell density may not be uniform across the entire endothelial surface, and the presence of endothelial stem cells at the limbus cannot be ruled out. Such factors may also contribute to continued function of the graft endothelium, despite total or substantial donor cell replacement.
7.  
   We acknowledge that arguments can also be made in favor of the alternative hypothesis of donor endothelial cell survival in the graft. The power of the technique developed by Wollensak et al. ^3,4 is that we no longer have to assume either survival or replacement of cells, as we can observe these cells directly in sex-mismatched cases. If one accepts our study findings and if the results are representative of transplants in general, then donor cell survival is variable, and postoperative time is not a good predictor of cell fate in the graft. We hope that future studies with larger populations, as well as work in the area of endothelial stem cells, will shed more light on this discussion.
8.  
   Our studies have shown that the corneal endothelium after transplantation is a dynamic environment. We suggest that the utmost care be taken to minimize the loss of endothelial cells (and their function) during every step of the corneal harvesting, banking, transplantation, and postoperative process. Special attention must be given to minimizing surgical trauma and the circular scar and to promoting rapid healing and engraftment, to give endothelial cells (on both sides of the scar) the best chance of survival. In some cases, however, a desire to minimize disturbance of the endothelium may be incompatible with the clinical goal—for example, the necessity of using relatively large-diameter transplants for flattening the keratoconic cornea. We are hopeful that in cases with good recipient endothelium, techniques such as deep anterior lamellar keratoplasty (DALK) ⁷ will be used more routinely to spare the endothelium entirely.