In Denmark, at least 500 mice per year are used for skin infection studies, corresponding to around 25,000 mice throughout EU. In vitro skin models, such as artificial human skin grown in a Petri dish, are used to test cytotoxicity and allergenicity of topical drugs. However, the efficacy of antimicrobial drugs for skin infections are currently tested in vivo, as no in vitro model for this purpose exists. Thus, the aim of this project has been to develop, implement and evaluate an in vitro skin infection model for use as an alternative to animal models for test of the efficacy of topical antimicrobial compounds and investigation of bacterial load by different pathogens.
The in vitro wound infection model is based on commercially available in vitro skin (EpiDerm-FT, MatTek). The skin consists of epidermal keratinocytes and dermal fibroblasts, which have been cultured at the air/liquid interface to form a multilayered model of the human skin, including a fully developed basement membrane. A biopsy punch is used to expose the dermis while leaving the basement membrane intact, thereby resembling a superficial skin wound (3 mm in diameter). Then the wound is infected with bacteria, and the infection is allowed to develop (18 h), whereupon topical treatment with antimicrobial ointment is performed twice daily.
We have used the model for evaluation of treatment efficacy of two frequently used topical antimicrobials; fusidic acid and mupirocin for treatment of infection with Methicillin Resistant Staphylococcus aureus (MRSA). We have also measured the effect of infection and treatment of the infection upon the level of 29 cytokines and chemokines, and the skin pieces has been subjected to histology.
The level of bacteria in the wounds treated with fusidic acid and mupirocin were, respectively, 10.000 and 200.000 lower than in untreated infected wounds. This treatment effect is comparable to what is found using an in vivo mouse wound infection model investigating the same two topical antimicrobials. The cytokine and chemokine profile indicated a markedly lower inflammation level for the infected wounds treated with the two antimicrobials than for the untreated infected wounds. The histology showed that the bacteria remaining in the treated wounds were primarily located in the wound periphery, maybe because the bacteria here are protected against the treatment by small cracks in the skin.
We also used the wound infection model to investigate the infection level (without treatment) of three different Staphylococcus strains. We found that S. pseudintermedius (an opportunistic pathogen of domestic animals) infected the skin model almost to the same level as the two S. aureus strains (one methicillin sensitive and one methicillin resistant strain), however the S. pseudintermedius caused a lower inflammation level (based on the cytokine and chemokine profile) than the two S. aureus strains. Infection level and cytokine and chemokine profile were not significantly different for the two S. aureus strains.
Thus we show here the in vitro wound infection model developed can be used for examining the bacterial load with or without treatment with topical antimicrobials. Moreover, the model can be used to examine the bacterial load of different strains of bacteria. Investigation of infections based on histology and cytokine profile can also be performed using the in vitro wound infection model. Thus the model offers a strong alternative to animal models for research in wound infections, including but not limited to investigations of the efficacy of topical antimicrobial compounds.