Expression and enzyme activity of cytochrome P450 enzymes CYP3A4 and CYP3A5 in human skin and tissue‐engineered skin equivalents

CYP3A4 and CYP4A5 share specificity for a wide range of xenobiotics with the CYP3 subfamily collectively involved in the biotransformation of approximately 30% of all drugs. CYP3A4/5 mRNA transcripts have been reported in the skin, yet knowledge of their protein expression and function is lacking. In this study, we observed gene and protein expression of CYP3A4/5 in both human skin and tissue‐engineered skin equivalents (TESEs), and enzyme activity was detected using the model substrate benzyl‐O‐methyl‐cyanocoumarin. Mass spectrometric analysis of TESE lysates following testosterone application revealed a time‐dependent increase in metabolite production, confirming the functional expression of these enzymes in skin.


| BACKGROUND
Cytochrome P450 (CYP450) enzymes are a superfamily of haemoproteins that metabolize a multitude of endogenous and xenobiotic molecules and are essential to maintain homoeostasis. [1,2] As the most highly expressed xenobiotic enzymes in the liver, they are responsible for the majority of phase I reactions, metabolizing between 70% and 80% of all drugs. [3,4] The CYP3 subfamily of the CYP450 enzymes, comprising of CYP3A4, CYP3A5, CYP3A7 and CYP3A43, are collectively involved in the metabolism of over a third of these drugs. [4] CYP3A4, the most abundantly expressed isoform in the liver, [5] has a large active site that enables its interaction with a wide range of structurally diverse compounds and permits metabolism of such molecules and drugs as hormones (testosterone, oestrogen), anticancer drugs (paclitaxel, tamoxifen) and antifungal agents (ketoconazole) amongst others. [4,6] CYP3A5 shares 85% amino acid sequence homology with CYP3A4 and displays similar substrate specificity. [7] Gene expression of CYP3A4 and CYP3A5 has been detected in skin and tissue-engineered skin equivalents (TESEs) by genomic analysis in some studies [8,9,S1,S2] but not others, [S3] and to date, protein expression has not been detected. [S4,S5] Given the use of skin as a drug delivery route, it is extremely important to determine functional expression of CYP3A4/5. Moreover, the increasing use of TESE as an alternative to animal testing for drug irritation and sensitivity assays has led to their proposed use in drug-induced toxicity assays. However, data on functional xenobiotic enzyme activity in these models are lacking and urgently required.

| QUESTIONS ADDRESSED
The relative gene and protein expression of CYP3A4 and CYP3A5 in human skin in comparison with liver remains unclear. It is also unknown whether the expression is localized to the dermis or epidermis and whether the enzymes are functionally active and are able to metabolize clinically relevant molecules.

| RESULTS
Gene expression levels of CYP3A4 and CYP3A5 were in much greater abundance in the liver than skin with expression being several hundred-fold greater for CYP3A4 and ninefold for CYP3A5, respectively ( Figure 1A,B). In human liver, gene expression of CYP3A4 was significantly greater (P < .001) than for CYP3A5, whereas, in skin, gene expression of CYP3A5 was significantly greater (P < .001) than that of CYP3A4 ( Figure 1A,B). Immunoblot analysis revealed that both enzymes were also expressed at markedly lower levels in skin compared to liver ( Figure 1C,D), confirming the qPCR data. An immunopositive band corresponding to CYP3A5 was not detected in skin at exposure times that were saturating for liver samples; however, when probed in isolation, CYP3A5 was detected when the exposure time was extended to increase assay sensitivity ( Figure 1E). mRNA transcripts for CYP3A4 and CYP3A5 were detected in TESE.
Gene expression levels for CYP3A4 were very low and several hundred-fold lower than for CYP3A5, and similar to native skin, TESE expressed significantly more mRNA transcripts for CYP3A5 than CYP3A4 (P < .001) ( Figure 1F), and significantly more CYP3A4 (P < .05) and CYP3A5 (P < .01) was expressed in the epidermis compared to the dermis ( Figure 1G,H). Immunohistochemical analysis showed that CYP3A4 was equally distributed throughout the epidermis although absent in the cornified layers of both skin and TESE ( Figure S1A,B). CYP3A4 staining was also observed in the majority of stromal fibroblasts and was prominent in the endothelium lining the blood vessels within the connective tissue of native skin ( Figure   S1A). In contrast, staining was not observed in the fibroblasts populating the collagen hydrogel in the TESE ( Figure S1B), supporting the qPCR data that showed very low mRNA expression in the TESE dermis. In native skin, expression of CYP3A5 was largely restricted to F I G U R E 1 Comparison of gene and protein expression levels between human liver, skin and TESE for CYP3A4 and CYP3A5. cDNA from human liver and skin was subjected to qPCR analysis for CYP3A4 (blue) and CYP3A5 (red) and gene expression relative to GAPDH determined for (A) liver and (B) skin. Protein expression in liver and skin was resolved by immunoblotting for (C) CYP3A4 and (D) CYP3A5. (E) CYP3A5 expression in skin following prolonged exposure during infrared detection of immunopositive bands; β-actin was used as a loading control, and immunoblots are representative of 3 independent experiments. F, Relative gene expression of CYP3A4 (blue) and CYP3A5 (red) in whole EpiDerm-FT400™ TESE by qPCR analysis. RNA from the epidermis and dermis of EpiDerm-FT400™ TESE was isolated, reverse transcribed and subjected to qPCR analysis for (G) CYP3A4 and (H) CYP3A5. Data are relative expression to GAPDH. Data are mean ± SD for n = 3 independent experiments and 3 technical repeats per experiment. *P < .05, **P < .01, ***P < .001 by ANOVA F I G U R E 2 Functional activity of CYP3A in (A) liver, (B) skin and (C) TESE and testosterone metabolism by TESE. Protein extracts were generated from human liver, skin and EpiDerm-FT400™ TESE and subjected to kinetic enzyme analysis using increasing concentrations of the model substrate benzyl-O-methyl-cyanocoumarin (BOMCC). A Nelder-Mead algorithm was used to fit the enzyme activity data, and all data sets were fit with the assumption that K m (the binding affinity for the substrate at half the maximum velocity) was the same in each assay (K m = 10 μmol/L). Maximum velocity (V max ) was calculated as specific metabolic activity of CYP3A in μmol/min/mg. Data are from three independent experiments performed in triplicate. Goodnessof-fit R 2 values are as follows: (A) .8855, (B) .9779 and (C) .9962. Testosterone was added either to (D) the surface of the epidermis to mimic topical drug deliver or (E) added to the medium that bathes the dermis to mimic systemic drug delivery. The production of the specific testosterone metabolite 6β-OH-testosterone was measured by mass spectrometry in epidermal and dermal extracts as well as the tissue culture medium after 8 and 24 hours. Data are from 2 independent experiments and 3 technical repeats per experiment the stratum basale ( Figure S1C) whereas expression in TESE models was extremely weak throughout the entire epithelium ( Figure S1D).
The discordant in immunostaining and qPCR data is likely due to low antibody affinity for CYP3A5. Both native skin and TESE revealed no staining in the dermis, suggesting that dermal fibroblasts do not express CYP3A5.
Liver extracts exhibited a maximal reaction velocity (V max ) of 6694 μmol/min/mg whilst native human skin extracts exhibited a lower V max of 3215 μmol/min/mg. The V max of TESE was approximately half that of native skin (1788 μmol/min/mg). Assuming that the enzyme catalytic rate (K cat ) for CYP3A for this substrate is the same in liver, skin and TESE, we can use these V max estimates to calculate the relative tissue/extract enzyme expressions, namely skin:liver = 0.48:1; TESE:liver = 0.267:1; TESE:skin = 0.556:1 (Figure 2A-C).
Testosterone was applied to the stratum corneum or to the culture medium of TESE for 8 and 24 hours to mimic topical or systemic delivery. TESE models were separated into epidermal and dermal components, and extracts of these along with the conditioned medium were analysed by mass spectrometry for presence of the CYP3Agenerated metabolite, 6β-OH-testosterone. When testosterone was applied topically, 6β-OH-testosterone was detected in the epidermis, dermis and culture medium at both 8 and 24 hours. At 8 hours, 6β-OH-testosterone levels were similar in the epidermis, dermis and medium, whereas after 24 hours, the levels of 6β-OH-testosterone found in the epidermis were markedly lower, in the dermis were similar and in the medium had increased fivefold compared to levels at 8 hours (P < .001) ( Figure 2D). A markedly different pattern of metabolite production was observed when testosterone was added to mimic systemic delivery. In these models, the presence of 6β-OHtestosterone was not detected in the epidermal or dermal compartments at 8 or 24 hours, whereas after 8 hours, 6β-OH-testosterone was detected in the culture medium and these levels were significantly increased after 24 hours (P < .001) ( Figure 2E).

| CONCLUSIONS
In this study, we show both gene and protein expression of predominantly CYP3A5 but also CYP3A4 in native skin and TESE. These data have important clinical implications for dermal drug delivery of existing or new compounds, in particular those with structural components that are likely to be metabolized by CYP3A4/5. They also show that, although expressed at lower levels than skin, TESEs are able to replicate the kinetics of skin metabolism of topically and systemically delivered drugs.

CONFLICT OF INTERESTS
The authors have declared no conflicting interests.

AUTHOR CONTRIBUTIONS
SAS, HEC, KMS, RS-Y and PS performed the experiments and analysed the data. SDW and AS analysed the enzyme kinetic data. CM, SAS and HEC designed the research study and wrote the manuscript. CM and SDW supervised the research.