BackgroundDescemet's membrane endothelial keratoplasty (DMEK) might be a promising technique for future xeno-corneal transplantation due to its ultrathin graft, extremely low rejection occurrence, suture-free graft fixation, and minimal immunosuppressive regime usage. The aim of this study is to explore the feasibility and efficacy of preparing porcine DMEK grafts by 2 techniques and investigate the graft ultrastructure. MethodsTwo mainstream techniques, mechanical stripping technique and liquid bubble technique, were modified to prepare the porcine DMEK grafts. In all, 40 corneas harvested from WZS-pigs (aged 10-12months) were subjected to the techniques (20 corneas for each technique). The success rate, time consumption, and endothelial cell density (ECD) before and after preparation were recorded and compared between the 2 techniques. And the ultrastructure of the porcine DEMK graft was investigated by transmission electron microscope. In addition, 9 WZS-pigs with different ages were sacrificed to explore the correlation between the thickness of Descemet's membrane and porcine age. ResultsAfter modifying several technical details, the porcine DMEK grafts were successfully prepared by either mechanical stripping technique or liquid bubble technique, and the mark technique to distinguish the 2 sides of the graft was also explored. In all, 13 DMEK grafts (65%) were prepared successfully by the mechanical stripping technique, whereas 14 successful cases (70%) were prepared by the liquid bubble technique. The success rates between the 2 techniques showed no significant difference (P=.847). However, the mechanical stripping technique was significantly time-consuming when compared with the liquid bubble technique (P<.0001). The ECDs reduced significantly after preparation no matter what techniques were used (P<.0001), but the ECD after the liquid bubble preparation was significantly higher than the ECD after mechanical stripping (P=.032). The ECD reduction positively correlated to the time consumption for both mechanical stripping technique (P=.0014, R-2=0.621) and liquid bubble technique (P=.013, R-2=0.412). The ultrastructure showed the graft was comprised of stromal residuals, non-banded layer, and endothelial layer. Unlike human Descemet's membrane (DM), anterior banded layer was not observed. The thickness of porcine DM increased with the age, and a significant positive correlation between them was found (P<.0001, R-2=0.949), and the predict equation was Y=0.3764*X+7.378 (Y indicates the thickness, whereas X indicates the age). ConclusionsPorcine DMEK grafts could be prepared either by mechanical stripping technique or liquid bubble technique, and the liquid bubble technique seems superior over the mechanical stripping technique regarding time consumption and ECD preservation. Although there are several technical barriers to overcome, xeno-DMEK might be a promising direction for future xeno-corneal transplantation.