The research of tissue-engineered skin has made a great development in plastic surgery field, but as above-mentioned, the tissue-engineered skin and skin substitutes reported in literature or used in a commercialized   market have a lot of disadvantages. So, in this experiment, we evaluated the effect of a new kind of tissue-engineered skin, ADM-KGF, and epidermal stem cells in repairing skin defects of nude mice. We also gave detailed discussions on the results, advantages and problems. KGF which belongs to fibroblast growth factor family (FGF-7) is a kind single polypeptide chain synthesized by fibroblasts and purified from fibroblast culture medium. Different from other members of FGF, KGF acts on epithelial cells exclusively as a paracrine regulation factors promoting the growth of epidermal cells. Local application of KGF can keep the proliferation and differentiation of epidermal cells and complete the covering and healing of wound surfaces through epithelization^{[6, 18-24]}. KGF is massively produced in the early stages of wound healing^[25]. Keratinocyte proliferation, migration, and cytoskeletal rearrangement are largely controlled by KGF that is released by the underlying fibroblasts^[26]. The growth curve in this study showed that the growth of epidermal cells added with KGF was better than the group without KGF, which is similar to the results of previous research and confirms the proliferation effect of KGF again^{[6, 18-19, 25-27]}. KGF was often added in culture medium or sprayed on defect surfaces directly in the previous experiments. Because of the short half-time, KGF needs multiple doses or continuous administration to keep its concentration, which increase the burden of laboratory technicians and the tediousness of treatment process, and meanwhile, some KGF cannot act directly on epidermal cells, which also induce waste to some extent. Therefore, KGF release system is needed and the nanocapsule is a nice choice to construct controlled release system of growth factors^[20-21]. In this study, we first encapsulated KGF in the PLGA nanocapsules, cross-linked the capsules with ADM, and then epidermal cells were seeded on it. All of these can make KGF release last for 4 weeks or more so that KGF can act on cells slowly and thoroughly. This method has not been reported yet and our study showed that tissue-engineered skin loading KGF nanocapsules was better than blank group and control group in the repair of skin defects, which indicates that KGF encapsulated in the nonocapsules can have a continuous and slow action.  Based on our previous research^[17], we constructed KGF nanocapsules successfully. The capsules have an integrated shape and a smooth surface, what is more, they also have an ability of connecting tightly with ADM without exfoliation. Most of disadvantages of polypeptide growth factors, such as short half-time, discharged and diluted by wound exudates, degraded by protease easily and decreased efficacy, can be solved through this method. Because the biocompatibility of macromolecular chemical PLGA which degrade into CO₂ and H₂O, this ideal release system can release growth factors continuously and without poison and harm to the human body. In addition, the use of ADM is based on the following experimental and clinic reason^{[3, 22-23]}: without the keratinocytes, langerhans cells and other cells that have immunogenicity, the ADM is suitable to be dermal scaffold of tissue-engineered skin which has no immunogenicity and cannot induce immunological rejection. This is why we choose ADM as the dermal scaffold to utilize its native structure and molecular information. As seed cells of tissue-engineered skin, the epidermal stem cells have two significant biological characteristics^[5]: slow cell cycle, shorter time for cell division; and powerful ability of proliferation, clonal growth in vitro culture. The DNA in cells has lest mutation and has enough time for self-repair and self-renew.  Some research has indicated that β1-integrin ^[24] and K19^[7] are two most reliable surface markers for cell selection. So in this study, we used β1-integrin to confirm the existence of epidermal stem cell and used K10 as a marker of mature keratinocyte. Testified by immunofluorescence assay and flow cytometry detection, there were a lot of epidermal stem cells^{[1, 10-11]} in epidermal cells obtained through collagenase type I combined with trypsin digestion method. The epidermal stem cells seeded on the KGF-ADM grew well and could be affected by KGF continuously. They were not only connected with ADM tightly and had an active proliferation, but also combined with PLGA tightly and grew well on it. This phenomenon is not seen in previous report and manifests that our method is a reliable way and the extracellular matrix constructed is suitable for cell adhesion or growth. The KGF in nanocapsules can act on epidermal cells and epidermal stem cells adhered on the cyst wall directly through the ways like leaching or infiltration of cyst wall, dissolution of PLGA and others. The experimental findings also prove that PLGA has a good biocompatibility and the method is completely feasible and effective by which we combined natural biomaterials ADM with chemosynthetic macromolecular degradable biomaterials PLGA to construct the dermal matrix loading nanocapsules.In the experiment, we used the tissue-engineered skin constructed according to the above-mentioned method to repair full-thickness skin defects of nude mice. We had a 6-8 weeks continuous observation on the restoration of skin transplantation. The experiment was carried out in three groups: one was experimental group which was applied with tissue-engineered skin lading KGF nanocapsules, one was blank group which was applied with tissue-engineered skin without KGF nanocapsules, and the last one was control group in which autologous full-thickness skin of nude mice was used. Two weeks after transplantation, the repairing areas of three groups were healed basically. At 4 weeks, the color of skin grafts in each group was as same as the surrounding tissues, but there were some contractions in the three groups, especially in the control group. Six weeks later, the color of skin grafts in each group had no difference from the normal skin, and the skin graft in the experimental group had no contraction anymore, but in the control group the contraction was still obvious. Under the microscope, the experimental group had the best healing: the layer number of epidermal cells was similar to the normal skin, the structure was similar to the stratified squamous epithelium, the connection between cells or epidermal cells and dermis was tight, the stratum corneum which could play a barrier function of skin was produced and the new collagen in the dermis was arranged orderly. These results indicate that epidermal cells could proliferate and differentiate continuously since they were affected by KGF. On the other hand, there was abscission in some epidermal cell layer areas of blank group which was probably due to the lack of KGF. Detected by anti-human K10 FITC-immunofluorescence, the transplanted skin in experimental group had good cell layers in the epidermis and expressed normal keratin with green fluorescence, confirming that the defect areas were repaired by the tissue-engineered skin. We detected skins graft at 8 and 10 weeks after transplantation using anti-human β1-integrin-immunofluorescence, and found β1-integrin positive cells in some sections. Based on this, we inferred that there were still epidermal stem cells or transient amplifying cells which had the ability of proliferation and differentiation. This result reveals again that the epidermal stem cells as seed cells can keep the self-renew and growth of tissue-engineered skin for a long time. In the traditional skin transplantation model, autologous full-thickness skin is always used as “gold standard”, but in this study, we found the repairing effect of experimental group was better than the skin autograft. The tissue-engineered skin combined with epidermal stem cells maybe hold a promise of being a new kind biomaterial to repair skin defects completely. In addition, we also observed that the contraction of auto-skin transplantation group was worse than the other two groups, which is probably because the skin autograft of nude mice was looser than the ADM used in the experimental and blank groups. As we known, ADM is much thicker than the full-thickness skin of nude mice. When taking bandaging fixation, the ADM is unease to contract because of some supporting effect. So, the long term effect of ADM is better than the skin autograft.In short, the findings of this experiment and our previous research^{[1, 10-11, 17]} all show that: we can construct a new kind tissue-engineered skin which is suitable for skin burn treatment and defects which can be repaired within 1-2 weeks by using KGF nanocapsules controlled release system and epidermal stem cells. The effect of this kind tissue-engineered skin in repairing full-thickness skin defect is better than common tissue-engineered skin and similar to auto full-thickness skin. In the meanwhile, there are still epidermal stem cells or transient amplifying cells existing in the tissue-engineered skin that have proliferative and differentiative potentiality to keep long-term autogenous renew and metabolism of the transplanted skin successively.