Wound healing of human skin grafted onto the nude mouse

To study the wound healing of full thickness wound, at two months after transplantation of human skin transplanted onto nude mice, animals were anesthetized and excisional wounds were made through the entire thickness of the graft, at the center of the graft, using a 2-mm punch. At various time intervals thereafter, ranging from 4 days to 3 months, the mice were sacrificed and healing grafts with surrounding mouse skin were harvested and processed for an immunohistological study with a large range of species and non-species specific antibodies.

The different technical steps to study the wound healing of human skin transplanted onto the nude mouse.

The reconstruction of transplanted human skin is made through a highly coordinated and complex series of processes that we will separately described but that are, in fact, closely interdependent. The main steps of the wound healing of full thickness wound, two months after transplantation of human skin transplanted onto nude mice are schematised in the following figures:

The different steps of the wound healing of human skin grafted onto the nude mouse

 1. The formation of the granulation tissue  

Immediately after the creation of the wound, a blood clot is made to re-establish a transient barrier with the external environment.

During the first week after wounding, mouse cells and blood capillaries invade the space under the blood clot and form the granulation tissue. This denomination originates from the observation, in an histological section, of numerous granules that correspond to cross-cut newly formed blood capillaries. At 7 days after wounding, the granulation tissue was negative with the human specific antibodies directed against elastic fibers, type I collagen or vimentin. however, it was positive with the mouse antibodies directed against type I collagen or vimentin. It was also rich in fibronectin and strongly positive for actin.

An electron microscopy   study shows the presence of “myofibroblasts” which are cells characteristic of the granulation tissue and that have been firstly described by Gabbiani in 1971. By using phalloidin, a toxin that binds actin and that has been coupled to rhodamin, we can confirm that the myofibroblasts contain a rich fibrillar network of actin. In electron microscopy, we also observe that these cells seem to establish gap junction-like linkages with each other and that they are individually surrounded by a basal membrane.This contractile cells have a major role in the wound healing process by giving the property to the granulation tissue to contract and concomitantly producing collagen , therefore reducing the size of the wound and stabilising the wound closure. Due to the retraction process, the collagen fibers are characteristically aligned transversally to the wound.

Indirect immunofluorescence labeling for actin and fibronectin of the granulation tissue, at 7 days and 3 weeks after injury at the center of human skin grafted onto the nude mouse.

At 7 days, the granulation tissue is rich in fibronectin and actin-positive cells while at 3 weeks, the fibronectin staining is more diffuse and the actin is only detected in the blood vessels and the epidermis.

By immunolabeling, we have observed that the entire tissue of granulation was stained with non species specific antibodies directed against type IV collagen, laminin, and heparan sulfate proteoglycan (HSPG), with a more intense labeling of the blood vessels. These three glycoproteins are probably constituents of the basement membrane surrounding the myofibroblasts-Human type IV collagen is only observed at the junction between the epidermis   and the granulation tissue. It is produced by the human keratinocytes. This result also confirms the mouse origin of the myofibroblasts.

 2. Migration of the epidermis

While the granulation tissue is being formed, major changes occur in the non-injured human epidermis. On either side of the skin wound, at 48h after injury, the epidermis is necrotic over a short distance (100-200um). Adjacent to the edge of the necrotic tissue, the human epidermis is thickened and forms an epidermal spur which has invaded the underlying connective tissue and demarcates necrotic tissue from human living connective tissue.

At 4 days after injury, human epidermal cells positive with an anti HLA-ABC antibody migrate into the wound. At 7 days, an epidermal tongue has penetrated into the wound, separating the living granulation tissue from the scab composed of the blood crust, the necrotic human tissue and mouse inflammatory cells.

Observations with transmission electron microscopy reveal that the migrating cells at the tip of the epidermal tongue exhibits extensive pseudopodial projections which vary from ordinary microvilli to large, blunt, pale projections of the cytoplasm containing no organelles except for ribosomes and a few tonofilaments and which are in contact with cells and fibrin of the granulation tissue. In this region, there is no lamina densa and no hemidesmosomes and the migrating keratinocytes are producing proteolytic enzymes and phagocyte part of the digested material in order to be able to progress through the inflammatory tissue.

We can observe that there is no lamina densa and that keratinocyte pseudopodial projections come in direct contact with fibrin and inflammatory cells of the granulation tissue.

By immunolabeling, we could observe that there were deposits of fibronectin and fibrin at the tip of the epidermal tongue while type IV collagen and laminin were absent.The bullous pemphigoid antigen (BP) is present till the tip of the epidermal tongue but remains intracellular. In zones more distant from the tip of the tongue, where the epidermis appears stabilised, type IV collagen and laminin are deposited, but in a fragmented fashion. By transmission electron microscopy, a fragmented lamina densa in front of hemidesmosomes is observed in this region.

Figures showing the extremity of the epidermal tongue and a non injuried zone of human skin grafted onto the nude mouse, at 7 days after injury. Indirect immunoflurescence labeling for fibronectin and fibrinogen that are revealed in white.

We observe a fibrin deposit under the epidermal tongue extremity while it is only present in the blood vessels in non injuried skin. Fibronectin is present through the whole thickness of the granulation tissue and of the non injuried dermis.

Figures showing the extremity of the epidermal tongue and a non injuried zone of human skin grafted onto the nude mouse, at 7 days after injury. Indirect immunoflurescence labeling for human type IV collagen and laminin that are revealed in white. The yellow arrow indicates the tip of the epidermal tongue.

We can observe that there is human type IV collagen and laminin all along the epidermal junction with the exception of the epidermal tongue tip. The blood vessels of the granulation tissue are labeled with the non-species specific anti-laminin antibody but are negative with the anti-human type IV collagen antibody, indicating their mouse origin.

We can observe hemides mosomes in front of fragments of lamina densa in construction and the presence of myofibroblasts with a well developed actin network.

Eventually, two weeks after injury, a new epidermis made of human cells has completely recovered the wound. It is at this stage that the crust will be discarded.

Therefore, epidermal cells of human skin transplanted onto nude mice are able to reconstruct an human epidermis above a mouse granulation tissue. For this purpose, they migrate on a matrix made of fibrin and fibronectin and devoid of type IV collagen and laminin. Similar observations have been made in studies of the wound healing of guinea pig or porcine skin. The ultrastructural changes observed during this process were very similar to the ones revealed during the wound healing of human skin “in situ”.

 3. Reconstruction of the dermal-epidermal junction

When the wound is completely re-epithelialised, the granulation tissue is going to be progressively remodeled. At three weeks after injury, a cross-section of the injuried zone shows an hyperplastic epidermis recovering a connective tissue constituted of a small number of cells dispersed in a well-developed extracellular matrix rich in type I collagen and relatively poor in fibronectin. Immunolabeling with species specific antibodies at this stage demonstrates that the epidermis is human while the connective tissue is of mouse origin.

At the junction between the two tissues, linear deposits of human type IV collagen and laminin are oserved while fibronectin is discretely deposited and fibrin is absent. Mouse type IV collagen was only detected in regions close to blood vessels.

Three weeks after injury, we observe continuous deposits of human type IV collagen, laminin, and BP antigen at the junction between the reconstructed human epidermis and the remodeled granulation tissue. Laminin is also expressed in the blood vessels while the human type IV collagen is absent.

In transmission electron microscopy, a basement membrane identical to the one observed in normal human skin is reconstructed. It is composed of hemidesmosomes, a lamina lucida, a continuous lamina densa and a sub-lamina densa with anchoring fibrils, microfibrillar bundles and collagen fibers.

These data therefore indicate that tissues originating from two differents species, mouse and human, are able to cooperate to reconstruct at their interface, a normal dermal-epidermal junction.

 4. Maturation of the epidermis

We have already reported that, during the phase of migration, the keratinocytes demonstrate characteristics of migrating cells. Afterwards, when the wound is completely re-epidermised, the healed epidermis must recover its normal architecture and the keratinocytes must follow a normal differentiation program that will end in the formation of corneocytes and of a mature stratum corneum  . In another article dedicated to the model of transplantaion of human skin transplanted on the nude mouse  , we have already shown that there are antibodies, such as those named AE1 and KL1 that allow to define that an epidermis is normally differentiating.

The expression of keratins   revealed with AE1 and KL1 antibodies and the distribution of involucrin  , a soluble precursor of the cross-linked envelope, strongly vary during the different phases of wound healing. In zones of grafted human skin distant from the wound, the normal staining with a strict labeling of the basal epidermal layer with AE1, of all the suprabasal layers with KL1 antibody, and of the upper part of the epidermis with the anti-involucrin antibody is observed.

At 48h after injury, in the thickened living epidermis adjacent to the wound, at 7 days in the migrating epidermal tongue and at 3 weeks after injury in the wounded epidermis, AE1 and anti-involucrin antibodies stain the suprabasal layers of the human epidermis while KL1 antibody labeled the entire epidermis.

KL1 labeling is suprasal in the epidermis of non injuried human skin and through the entire epidermis in the healing epidermis. Involucrin is detected in the upper stratum spinosum of the non injuried epidermis and in the suprabasal layers of the healing epidermis. Vimentin is present in the basal keratinocytes of the epidermal tongue and is detected in the mesenchymal dermal cells, and in dentritic epidermal cells (probably, human Langerhans cells and melanocytes) in the noninjuried epidermis.

At 6 weeks after injury, basal keratinocytes but also some suprabasal keratinocytes are labeled with AE1 antibody while KL1 stains strongly the suprabasal keratinocytes and less intensively the basal cells. At this stage, involucrin is detected in upper half of the epidermis.

It is only at 9 weeks after injury that the wounded epidermis becomes normal with a strict labeling of the basal epidermal layer with AE1, of all the suprabasal layers with KL1 antibody, and of the upper stratum spinosum and stratum granulosum with the anti-involucrin antibody.

Three months after injury, the expression of involucrin and of keratins detected by KL1 antibody ( in yellow) are normal in the wounded epidermis. The whole tissue appears in orange/red due to the Blue Evans counterstaining.

Similar observations have been made by JP Ortonne and JP Lacour in studies of the wound healing of split-thickness wounds with human skin “in situ”.

In conclusion, it appears that the epidermis is reconstructed at 3 weeks after injury but its normal characteristics are only restored at 9 weeks after injury. As we are going to see now, the normalisation of the epidermal differentiation is correlated with the appearance of a human neodermis in place of the mouse granulation tissue.

 5. Formation of a human neodermis

We have seen that, at 7 days after injury, a mouse granulation tissue is present under the crust from which it is separated by a migrating human epidermal tongue. Between the first and the sixth week after injury, this granulation tissue is the site of major changes. The number of cells strongly decreases to be replaced by a well developed extracellular matrix rich in collagen.This remodeled granulation tissue is constantly negative with human specific antibodies directed against type I collagen or vimentin and is strongly positive with the antibodies directed against the same antigens but of mouse origin. On the other hand, at 9 weeks after injury, a band of tissue containing vimentin-positive human cells and human type I collagen is detected between the reconstructed human epidermis and the remodeled granulation tissue positive for mouse type I collagen.

It is only at 9 weeks after injury that human vimentin positive fibroblasts are migrating under the healed epidermis, in the mouse granulation tissue where they produce human type I collagen and form a human neodermis.

At 3 months after injury, the mouse granulation tissue has nearly completely disappeared to be replaced by a human neodermis labeled with antibodies directed against Type I, III, or VI human collagen, against Human elastic fibers, or against human vimentin.

In the neodermis, human type I collagen has progressively substituted the mouse type I collagen of the remodeled granulation tissue.

These data therefore indicate that, after remodelling, the granulation tissue produced by mouse cells is progressively replaced by a human neodermis made by cells originating from the non-injuried human dermis. A double staining with antibodies directed against human vimentin or mouse type I collagen shows that this phenomena results from the progressive invasion of the granulation tissue by human cells coming from the periphery of the wound and migrating just beneath the reconstructed epidermis.

This last observations are of major relevance since they demonstrate that the reconstruction of the dermis in a skin wound is made in two steps. Firstly, a transient tissue, the granulation tissue is rapidly produced by cells the origin of which are yet not clearly defined but which are not cells from the surrounding non-injuried dermis. Then, this tissue is completely replaced by a human neodermis made by human cells coming from this non-injuried dermis.

 6. Langerhans cells   and melanocytes

In another article, we have already reported that human Langerhans cells and human melanocytes ar present in human skin transplanted onto the nude mouse. It was therefore interesting to study their fate during the wound healing process.

By using monoclonal antibodies OKT6 and anti-HLA-DR or by transmission electron microscopy, we could observe that Langerhans cells colonize the wounded epidermis but with a slight delay in comparison with the migration of the keratinocytes. At 7 days after injury they were already observed in the epidermal tongue at a some distance from the tip.

A Langerhans cell observed with a transmission electron microscope in the healed epidermis at three weeks after injury at the center of human skin transplanted onto the nude mouse.

Observations with transmission electron microscopy also show the presence of melanocytes in the wounded epidermis at 3 weeks after injury. Melanocytes contain vimentin. At this stage, no epidermal cells are labeled with an antibody directed against mouse vimentin. On the other hand, cells, a part of them being in basal position, are positive for human vimentin. It is therefore likely that the melanocytes observed in transmission electron microscopy are human.

A human melanocyte observed with a transmission electron microscope in the healed epidermis at three weeks after injury at the center of human skin transplanted onto the nude mouse.

Since we never observe human vimentin-positive cells in the granulation tissue at the early stages of the process, the migration of human Langerhans cells and human melanocytes that repopulate the wounded epidermis likely occurs through the healing epidermis from the pre-existing pool of cells from the non-injuried human epidermis.

 References

Demarchez M, Sengel P, Prunieras M. Wound healing of human skin transplanted onto the nude mouse. I. An immunohistological study of the reepithelialization process. Dev Biol. 1986 Jan ;113(1):90-6. PubMed PMID : 2417903.

Démarchez M, Hartmann DJ, Herbage D, Ville G, Pruniéras M. Wound healing of human skin transplanted onto the nude mouse. II. An immunohistological and ultrastructural study of the epidermal basement membrane zone reconstruction and connective tissue reorganization. Dev Biol. 1987 May ;121(1):119-29. PubMed PMID : 3552786.

Demarchez M, Desbas C, Prunieras M. Wound healing of human skin transplanted on to the nude mouse. Br J Dermatol. 1985 Jul ;113 Suppl 28:177-82. PubMed PMID : 3893519.