**4. Discussion**

High radiation doses such as those delivered during radiation therapy (RT) produce pathological changes in mesenchymal tissues with long-term alterations of fibroblast phenotypic and functional characteristics that may impair the quality of life of treated cancer patients [16,26,27], whereas the pathophysiology of epithelial complications of RT has been investigated less [17]. Furthermore, the effects of very low doses of genotoxic stress are poorly documented [19,28], although dysplasia has been reported in the skin of radiotherapy patients [18]. We have here addressed the question of possible adverse reactions subsequent to the exposition of keratinocyte stem and progenitor cells to low radiation doses. The dose analyzed here (50 mGy) is in the range of those delivered to normal tissues adjacent to the target tumor volume during radiotherapy, and is relevant for biomedical diagnostic procedures, notably scanner imaging. It is much lower than the dose limit accepted for the induction of carcinoma, which has been proposed to be around 500 mGy, notably based on the Japanese atomic bomb survivor study [29]. The key contribution of the present study was the demonstration that dysplasia and epithelial-to-mesenchymal transition (EMT) develop in epidermises generated by keratinocyte stem and precursor cells exposed to a single low dose of γ irradiation, thus documenting a micro-environment favoring the development of skin cancer.

Epidermal holoclone keratinocytes provided a model representative of an immature population of cultured precursors containing functional stem and progenitor cells. These cells correspond to the clonal progeny of single keratinocyte stem cells that were functionally characterized according to an extensive growth potential exceeding 100 population doublings through successive subcultures, and the capacity for epidermis reconstruction in vitro and regeneration in vivo [11,12]. Importantly, the stem cell status attributed to holoclones has been demonstrated in vivo by cellular tracing in the entirely regenerated epidermis of an epidermolysis bullosa patient engrafted with an autologous, genetically corrected skin substitute [30]. Moreover, the fibrin-based epidermis organoids that were used in this study corresponded to an adaption from a clinically relevant model of bioengineered skin substitute [20,31]. The present epidermis regeneration approach permits modeling of skin stem and progenitor cell properties and potentialities in conditions characterized by a higher level of cell proliferation and metabolic activities than in the context of healthy skin homeostasis. These conditions probably exacerbate radiosensitivity and thus allow observation of cell and tissue responses to low stress levels.

The semi-quantitative imaging approaches that were set-up to characterize the epidermal distribution patterns of molecular markers provided clues for the understanding of genotoxic stress-induced epidermal dysplasia and EMT development. Firstly, the marked alteration of VANGL2 patterns that was detected in association with the impaired orientation of basal keratinocyte nuclei constituted a relevant parameter, due to the involvement of this membrane protein in the regulation of cell polarity and migration [23]. VANGL2 is a central component of the planar cell polarity signaling pathway (PCP), which is essential for correct epidermal development and morphogenesis [32,33], as well as epidermal wound repair [34]. The loss of epithelial polarity is a key process in the early steps of EMT. Weakening and disruption of cell–cell contacts, which were documented here by the presence of non-cohesive spaces and a local decrease in E-cadherin level, are typical characteristics of the pathophysiological process of EMT [25]. Finally, detection of an ectopic expression of the mesenchymal markerα-SMA in keratinocytes consolidated the EMT-like phenotypic switch occurring in radiation-induced dysplastic areas (DAs) [35]. Among the various transcription factors involved in EMT, ZEB1 has been described as a major early player in its development, later favoring epithelial tumor progression in association with E-cadherin suppression [36]. We show here the appearance of ectopic ZEB1 protein in keratinocytes of dysplastic epidermis, thus providing a link between low-dose irradiation of holoclone keratinocyte precursor cells and potential initiation sites of the carcinoma development cascade. Notably, perturbated functions of the 'wingless' (WNT) signaling pathways, as suggested here by the loss of the β-catenin protein, associated with focal dysplasia, might constitute a promoter event of the pathophysiological processes described here.

In conclusion, we have developed an approach based on in vitro bioengineering of human skin organoids, coupled with in vivo xenografting in immune-deficient mice, to explore the pathophysiological consequences of low-dose γ-irradiation exposure of epidermal stem and progenitor cells on their subsequent regenerative capacity. We have observed that a single 50 mGy radiation dose was sufficient to promote local perturbations in regenerated epidermises with cellular and molecular characteristics of dysplasia and EMT, which may constitute an initial risk for the future development of carcinomas. Interestingly, this approach is directly applicable to other biomedical research domains, for example characterization of the skin's defenses and responses to pollution [37].

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4409/9/8/1912/s1, Figure S1: Absence of epidermis mixing between recipient mice skin and regenerated human epidermis grafts. Figure S2 and Table S1: Absence of p16INK4a in human epidermises regenerated by irradiated and non-irradiated keratinocyte precursors. Supplementary Methods.

**Author Contributions:** S.C.: investigation, methodology, resources, formal analysis, validation and original draft preparation; R.N.G.: investigation, methodology, resources, formal analysis and validation; D.S.: investigation, methodology, resources and formal analysis; P.A.: investigation, methodology and resources; M.T.M.: conceptualization, methodology, formal analysis, visualization, validation, writing—original draft preparation and writing—review and editing, supervision, funding acquisition; N.O.F.: conceptualization, methodology, formal analysis, visualization, validation, writing—original draft preparation and writing, review and editing, supervision, funding acquisition. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by grants from: CEA and INSERM (UMR967); "Délégation Générale de l'Armement" (DGA) grants; FUI-AAP13 and the "Conseil Général de l'Essonne" within the STEMSAFE grant; EURATOM (RISK-IR, FP7, grant 323267); and "Electricité de France" (EDF). Genopole® (Evry, France) also provided support for equipment and infrastructures.

**Acknowledgments:** We wish to thank S. Bouet and A. Boukadiri (histology platform, UMR 1313 GABI, INRA/CEA, Jouy en Josas) for technical assistance. Our thanks also go to H. Serhal and Y. Diaw of the Clinique de l'Essonne (Evry), who kindly provided human skin samples from healthy donors. We thank J.-J. Lataillade and M. Trouillas (IRBA, INSERM U1197, Clamart) for helpful discussions on the skin regeneration model.

**Conflicts of Interest:** The authors declare no conflict of interest.
