|Home | About | Journals | Submit | Contact Us | Français|
Hypertrophic scarring after burns is an unsolved problem and remains as devastating today as it was in the 40s and it may be that the main reason for this is the lack of an accepted, useful animal model. The female, red Duroc pig was described as a model of hypertrophic scarring nearly 30 years ago but then vanished from the literature. This seemed strange since the authors reported that 12 of 12 pigs developed thick scar. In the mid 90s we explored the model and found that, indeed, the red Duroc pig does make thick scar. Other authors have established that the Yorkshire pig does not heal in this fashion so there is the possibility of a same species control. We have continued to explore the Duroc/Yorkshire model and herein describe our experiences. Is it a perfect model of hypertrophic scarring? No. Is it a useful model of hypertrophic scarring? Time will tell. We have now obtained gene expression data from the Duroc/Yorkshire model and analysis is underway.
Hypertrophic scarring is a significant negative outcome of a burn injury. Hypertrophic scars (HSs) are hard, raised, red, itchy, tender, and contracted.1,2 These scars are ugly and uncomfortable and may diminish, but never completely go away. The resulting disfigurement and scarring affects quality of life, which, in turn, can lead to lowered self-esteem, social isolation, prejudicial societal reactions, and job discrimination.3–7 Scarring may also have profound rehabilitation consequences including loss of physical function, impairment, disability, and difficulties pursuing recreational and vocational pursuits.6,8,9
Literally hundreds of studies of collagen, fibroblasts, and growth factors in HSs have been performed over the past decades and yet the pathophysiology and treatment of this process are still essentially unknown.10–18 It can be argued that the reason that the etiology of human HS is unknown is the absence of a useful animal model.18–22 Mice, rats, rabbits, dogs, and cats have all failed to produce scar analogous to human HS. Repetitive literature searches have yielded few references to validated animal models of HS. Morris et al.21 reported a scar model in the rabbit ear following a small full-thickness wound, but this is quite different from the human situation that involves large, partial-thickness wounds so the relationship is somewhat questionable. Since originally reported, this model has been adopted by one other group.23 Xiang et al.24 created 2×5 cm full-thickness wounds on rabbit ears but again the full-thickness model seems to differ from the human condition. Models that include human HS tissue implanted into athymic rats and mice have also been described.22,25–31 Since originally reported, these athymic models have been adopted by two other groups in two studies,32,33 but this model seems very dissimilar to the clinical situations leading to HS formation in humans. Furthermore, the human transplanted tissue is established scar so any early morphogenic changes are missed. Yang et al.34 implanted normal human skin into nude mice and created wounds in the transplanted skin. But again the model seems quite dissimilar to the human situation. The authors recently reported that wounding the transplanted skin is not necessary for the development of hypertrophic scar.35 Aksoy et al.18 described a hypertrophic scar model in the albino, male guinea pig after excision of the panniculus carnosus and application of thermal injury or coal tar. Again, this seems quite dissimilar to the human condition and we could find no further use of this model. It would appear that no animal model of hypertrophic scarring has acquired the status of “gold standard” and become widely adopted.
Nearly 30 years ago, Silverstein and colleagues36,37 reported that deep partial thickness wounds in 12 of 12 female red Duroc pigs healed with “hypertrophic” scarring but was never published in a peer-reviewed journal. We could find no further studies of hypertrophic scarring utilizing this model. A manuscript by Aksoy et al.18 referenced an abstract presented at the 1981 meeting of the Plastic Surgery Research Council,38 which stated, “We initially performed several investigative studies in red Duroc pigs, including … various types of dermatome excisions … all without success”; however, no description of the methods nor actual results were provided. Furthermore, the abstract referenced a personal communication from B. Pruitt that the model was not reproduced in later studies. A review article by Kischer et al.26 also reported that the model had never been reproduced but gave no references.
We contacted Dr. Erk but it would appear that the details of the abstract are no longer available. We also discussed the abstract with Dr. Pruitt39,40 who could not recall this personal communication but did say “the scars were initially impressive, with time, spontaneous diminution occurred. Dr. Silverstein became discouraged when others criticized the model on the basis of that diminution ….” We also discussed the model with Dr. Silverstein who stated “While my scars did diminish somewhat over time, they never totally flattened out” and second “when Dr. Salisbury later tried the model on minipigs the wounds did not heal with thick scar.”41,42 Dr. Salisbury commented “We couldn’t duplicate Paul’s findings in the Duroc pig.”43 Dr. Robson also recalled that “over time the scars melted away.”44 Clearly there was some problem with this model but it seemed strange that the first 12 of 12 Durocs demonstrated thick scar at 5 months postwounding but then subsequent Durocs did not.
As the acquisition of tissue from human partial-thickness wounds in a systematic and controlled fashion is difficult, and our understanding of hypertrophic scarring is so minimal, we decided to attempt to validate an animal model. This Duroc model had appeal owing to the similarity of porcine skin to human skin, the similarity of the wounds to human wounds that result in HS, and the initial reported occurrence of thick scar in 12 of 12 animals. We have attempted to summarize our efforts in this review. We have also addressed the similarities and differences of pig and human skin and possible reasons that the Duroc model was not further utilized after the original description.
To validate an animal model, biologic findings in that model must be compared with the findings in the human condition. We reviewed the literature for human HS and enumerated the characteristic features. We felt that each criterion must be confirmed in the Duroc tissue to validate the model including (1) clinical appearance, (2) histologic appearance, (3) proteins and other biomolecules, (4) nerve fiber counts, (5) mast cell counts, (6) presence of collagen nodules, and (7) presence of myofibroblasts (Table 1).45–51
Our hypotheses regarding hypertrophic scarring include that both the deep dermis and the cones of the dermis47 play a role in hypertrophic scarring. Therefore, our wounds are deep partial thickness leaving the deepest portion of the dermis and the deep aspect of the cones. When the study requires the thickest scars, we leave very little dermis. When the study involves the deep aspects of the cones, we leave more of the deep dermis. At present, we are not studying full-thickness wounds where no dermis remains.
As previously reported,50 tangential wounds are created with a Padgett dermatome (Integra LifeSciences Corporation, Plainsboro, NJ) on the backs of 7-week-old female, red Duroc pigs. The dermatome is set to 0.015 in., 0.020 in., or 0.030 in. It is known that actual wound depth obtained with a dermatome is quite variable.52 Furthermore, for the deeper wounds two or more passes of the dermatome are necessary that introduces further error in wound depth. Therefore, we refer to total dermatome setting rather than wound depth. Eight 7×7 cm wounds are created on the back of each Duroc and the total dermatome settings are divided into six groups: 0.015 in., 0.030 in., 0.045 in., 0.060 in., 0.075 in., and 0.090 in. The wounds are allowed to heal without application of topical agents or dressings. Tissue samples are routinely harvested at 1, 2, and 3 weeks and 3 and 5 months after wounding. At 20 weeks, the pigs are returned to the producer.
At 3 weeks, shallow partial thickness wounds are usually healed, but deep partial thickness wounds are not. The gross clinical appearance of cutaneous scar on the female red Duroc after shallow partial thickness wounds is essentially the same as uninjured skin. The gross appearance of deep dermal wounds is raised, hard, hyperpigmented, and contracted, similar to human HS.47 It is not, however, as raised as in man. Whereas human HSs may reach 1–2 cm elevation above the surrounding skin, the scar elevation in the Duroc is at most 1–2 mm. Furthermore, the Duroc scars are not as red as in man.
After shallow injury, the histology of the healed wound at 3 weeks is essentially the same as uninjured skin. After deep partial thickness wounds, scars on the Duroc pig may reach 11 mm in thickness from hypodermis to stratum corneum compared with the normal skin thickness of ~3 mm (Figure 1).47,51 However, the shape of the scars is quite different from human HS and, as indicated above, only 1–2 mm of this scar extends above the surrounding uninjured skin, the remainder projects inward (Figure 2). There are large numbers of disorganized collagen fibers in the scar, in some places formed into whorls and nodules.47,51
TGFβ1 is a suspected mediator of fibrosis in a number of organs, including skin, lung, liver, and kidney.53–56 TGFβ1 protein is found in human HS54 and TGFβ1 mRNA is elevated in HS compared with uninjured tissue.57
After shallow wounds, TGFβ1 localization in the healing wounds was not immunohistochemically evident at any time. In contrast, in deep partial thickness wounds TGFβ1 was detectable at early time points in cells of the central scar mass and then declined to weakly detectable levels in the extracellular matrix in the depths of the wound at 5 months, probably in the residual dermis,47,49 which correlates with the findings reported for human HS.54 In both shallow and deep wounds, TGFβ1 mRNA was elevated at early time points and then declined to the levels of uninjured skin at 5 months,49 which also corresponds to the reports for human HS.57
Ghahary and others have suggested that IGF-1 may be involved in wound repair and associated with fibrosis.58–60 The authors reported that (1) IGF-1 protein was increased 78% in HS over uninjured skin61 and (2) IGF-1 mRNA is elevated early after wounding.
In shallow wounds, IGF-1 localization at early times was found in the epidermis, neo-vessels, and inflammatory cells. After 30 days, the staining was restricted to the epidermis and endothelial cells as in uninjured skin. In deep wounds at 10 days, a very strong expression was found throughout the wound that declined to 150 days when the expression was found only in a few fibroblasts47,49, which correlates with the published reports for human HS.61 IGF-1 mRNA as detected by in situ hybridization in shallow wounds at 10 days was found in inflammatory cells and by 30 days was essentially gone. In deep wounds, the expression at 10 days was overwhelming throughout the wound bed. This was greatly diminished at 5 months but still greater than in uninjured skin.47,49 With qRTPCR, in shallow wounds IGF-1 mRNA was elevated at early times and declined to the levels of uninjured skin. In deep wounds, the levels were significantly elevated at day 10 and declined thereafter. There was a second lesser rise at 3 months. Levels were still elevated, however, at 5 months compared with shallow wounds.49 These findings also correlate with reported findings for human HS.62 The second rise at 3 months also correlates with the findings of Gallant et al.63 who has also extensively evaluated scarring in red Duroc pigs.63–66
Decorin is the most abundant proteoglycan in normal dermis67 and has been reported to promote formation of correct collagen fibrils.68 It is possible, then, that in the absence of decorin, the structure and organization of collagen fibrils become distorted into whorls and nodules. It has been reported that decorin expression is reduced in HSs.69,70
In shallow Duroc wounds, decorin localization was reduced at 10 days but returned to normal at 30 days. In deep wounds, decorin staining was consistently reduced even at 5 months.47,50 This pattern is similar to that reported in human HS.69,70
Versican is highly expressed in fast growing tissue and cells, occupies the space between the collagen fibrils, and interferes with the assembly of fibrils into bundles and fibers.70–72 This large proteoglycan has 12–15 chondroitin sulfate chains, which are largely responsible for the water-holding capacity of connective tissue.73 Versican has been found to be virtually undetectable in normal dermis70 and increased in HS.70,74
In shallow wounds at all time points, we found no versican immunostaining. In deep wounds, versican appeared at 30 days in the scar tissue layer and the staining increased thereafter to 150 days.47,50 These findings are similar to those reported for human HS.70,72
NO production both by inducible NO synthase and constitutive endothelial NO synthase plays many important roles in wound healing from the early inflammatory phase through the process of scar remodeling.75 In addition, it has been suggested that NO may be involved in hypertrophic scarring.76
NO levels in deep Duroc wounds were higher than in shallow wounds at 10 days and then declined to levels significantly lower than in shallow wounds at 5 months.50 These low levels are similar to those reported for human HS.76
Recently, the nerve system of skin has been implicated in wound healing, perhaps via the biological effects of neuropeptides including the proliferation of epithelial, vascular, and connective tissue.77–82 The debilitating itching and pain associated with HS implicates sensory nerves within this aberrant healing process. Several investigators have studied nerves in HS. Crowe et al.83 demonstrated an increase in the number of neuropeptide-containing nerves in human HS compared with uninjured skin. Zhang and Laato84 reported that HSs are characterized by extensive nerve fasciculi identified by immunofluorescence with antineurofilament antibodies.
There are some differences in our “Duroc model” and that used by these authors. Our wounds are 6×6 up to 8×8 cm compared with the 2×2 cm wounds the authors reported. In addition, we follow the wounds for 20 weeks compared with the author’s 10 weeks. It is possible that wound geometry and time after wounding may alter the outcomes. Furthermore, the authors use a dermatome setting of 1.8 mm or 0.072 in., whereas we set the dermatome at 0.5 mm or 0.020 in. This difference could introduce differences in actual wound depth. It may be that there are Duroc models of scarring rather than one Duroc model.
Nevertheless, the authors reported gross and histologic findings similar to ours with the exception of the following. The authors reported that at 10 weeks the wounds were no longer raised or overtly fibrotic. As mentioned above, the thick scars in the Durocs are far less raised than HS in man. Usually the deep partial thickness wounds in the Durocs are raised ~1 mm compared with 1–2 cm for HS in man.
In summary, the authors have conducted extensive studies of healing in the Duroc pig and confirmed many findings. In addition, they added an extensive array of gene expression data. The comparable findings between the two studies are similar and seem to demonstrate that healing in the Duroc differs from that of the Yorkshire and is more similar to fibroproliferative scarring.
Liang et al.85 repeated the histologic studies and found thick scar at 5 months and that the cones exist in Duroc skin and are severely amputated in deep wounds.
Given the difficulty working with a large animal model, it is worth evaluating the benefits of porcine models. The skin of the pig is known to be similar to human skin. We will summarize some of the similarities and differences.47,86–98
Pigs can sustain sunburn, as can man, and both species rely on fat, not fur, for insulation. Both have sparse hair over most of the body. The epidermis of both is thick with distinct rete pegs and dermal papillae and the dermal–epidermal ratio varies from 10:1 to 13:1. The total turnover time of the epidermis in both is ~30 days and the epidermis of both contains Langerhans cells. Both contain elastic fibers. The collagenous tissue framework of the dermis and the adipose chambers of the hypodermis are quite similar. Immunohistochemical staining of human and porcine skin shows similar patterns for keratins 10 and 16, collagen IV, fibronectin, and vimentin. Collagen structure in porcine skin is remarkably similar to human collagen and evokes a minimal immune response in man. Both skins have cones in the dermis.47
There are dissimilarities however. Porcine epidermis includes only three layers (germinativum, granulosum, and corneum) and stratum corneum is thicker and very compact. Over the body surface, the pig has only apocrine sweat glands that are not involved in thermal regulation. Eccrine sweat glands in the pig are found only in the snout, lips, and “carpal organ.” The epidermis of the pig is high in alkaline phosphatase, whereas the endothelium of the surface capillaries of the pig contains no alkaline phosphatase. The hairs of the pig have practically no medulla and, in active colored hair follicles, melanocytes are present not only in the bulb, above the critical level, but a special population of them is also found in the matrix. The sebaceous glands of the pig contain much alkaline phosphatase and no glycogen. The pig displays a rapid and marked mast cell sensitivity to stress.
As a result of these general similarities, as mentioned by Rothschild99,100 the pig has assumed an important place in wound healing research and has been used in many studies, with many from the laboratories of Eriksson, Hart, Nanney, and Davis. Most studies used the young female Yorkshire pig. We found only two groups using the Duroc101–103 and two using the Yucatan minipig104,105 before the studies referenced in this manuscript.
It is strange that the first authors reported thick scar in 12 of 12 pigs or 100% and then did not study the model further. It is also strange that the first 12 of 12 pigs produced thick scar and subsequent Durocs studied by other authors did not. It seems clear from the history we obtained by personal communication that the problem has to do with the thickness of the scar at 5 months postwounding. We suspect that this has to do with wounding and wound depth. Healing with thick scar only occurs with deep wounds. Full-thickness wounds are quite easy to make. But then all of the dermis is removed and our HS hypothesis is that the deep dermis is required. Leaving a small portion of the dermis is difficult. It may be that the model was discarded because too much dermis was left in the wounds resulting in normotrophic healing. In addition, at that time very few antibodies and reagents were available for pig tissues. And finally for other reasons, as described below, the model is not perfect.
No. In addition to the problem of wound creation described above, Duroc pigs are not easy to obtain, as they are not common in the food industry. In addition, they are large (average 200# at 5 months of age) and expensive. Furthermore, their life span is 6–8 years and they achieve sexual maturity in 5–6 months so it is difficult to compare the biologic clock of the pig with the human.
Porcine models may be our best hope for recreating human wounds using animal models. We and others have recently revived the female red Duroc pig model of hypertrophic scarring that Silverstein and colleagues36,37 originally described in 12 of 12 Duroc pigs in the 1970s. As the scars, in our experience with 12 animals, are not as thick as human HSs or as raised above the surrounding skin, we refer to the scar in Durocs as “thick” rather than as “hypertrophic.” Nevertheless, our data combined with the work of Gallant and colleagues suggest that many similarities exist between the thick scars of the female red Duroc pig and human HS. As porcine microarrays are now available, and given the devastating nature of hypertrophic scarring, our ignorance as to cause and treatment, and the paucity of accepted animal models, we have now used the Duroc/Yorkshire model to obtain gene expression data in shallow and deep wounds at 1, 2, and 3 weeks and 3 and 5 months. Analysis is underway.
This work was funded by (1) Washington State Council of Firefighters Burn Foundation; (2) National Institute on Disability and Rehabilitation Research/Office of Special Education and Rehabilitation Services/US Department of Education; (3) National Institutes of Health; and the Northwest Burn Foundation. In addition, the authors are indebted to Dr. Edward E. Tredget, MD, for invaluable advice on how to validate the model.