Development and Characterization of a Novel in vitro Progression Model for UVB-Induced Skin Carcinogenesis

Epidemiological studies suggest ultraviolet B (UVB) component (290–320 nm) of sun light is the most prevalent etiologic factor for skin carcinogenesis- a disease accounting for more than two million new cases each year in the USA alone. Development of UVB-induced skin carcinoma is a multistep and complex process. The molecular events that occur during UVB-induced skin carcinogenesis are poorly understood largely due to the lack of an appropriate cellular model system. Therefore, to make a progress in this area, we have developed an in vitro model for UVB-induced skin cancer using immortalized human epidermal keratinocyte (HaCaT) cells through repetitive exposure to UVB radiation. We demonstrate that UVB-transformed HaCaT cells gain enhanced proliferation rate, apoptosis-resistance, and colony- and sphere-forming abilities in a progressive manner. Moreover, these cells exhibit increased aggressiveness with enhanced migration and invasive potential and mesenchymal phenotypes. Furthermore, these derived cells are able to form aggressive squamous cell carcinoma upon inoculation into the nude mice, while parental HaCaT cells remain non-tumorigenic. Together, these novel, UVB-transformed progression model cell lines can be very helpful in gaining valuable mechanistic insight into UVB-induced skin carcinogenesis, identification of novel molecular targets of diagnostic and therapeutic significance, and in vitro screening for novel preventive and therapeutic agents.

to parental cells, which was more uniform, smaller in size and rounder in shape (Fig. 1B). Further, monitoring of these sublines in subsequent passages demonstrated no reversal of these morphological characteristics. Next, we examined the growth, clonogenicity and sphere formation ability of these sublines. We observed that the growth rate of UVB-irradiated HaCaT (3wk, 8wk, 12wk and 16wk) sublines was significantly higher as compared to the parental HaCaT cells ( Fig. 2A). Total number of cells on the 8 th day of culture indicated 10.6% (3wk), 28.1% (8wk), 38.8% (12wk) and 50.0% (16wk) increase in the growth of UVB-irradiated sublines as compared to the parental cells ( Fig. 2A,B). Furthermore, cell population doubling time (dt) estimated during the logarithmic growth phase was also progressively decreased in derived cell lines subjected to greater repetitive exposure to UVB radiation (Fig. 2B). Further, in our plating efficiency assay, we observed that UVB-irradiated HaCaT sublines (3wk, 8wk, 12wk and 16wk) exhibited significantly higher (1.6, 2.5, 4.0, and 5.5 folds, respectively) plating efficiency as compared to the parental HaCaT cells (Fig. 2C). More importantly, while parental HaCaT cells did not form any colony in anchorage-independent (soft agar) clonogenicity assay, all UVB-irradiated HaCaT sublines exhibited clonogenic ability (Fig. 2D). The number of colonies formed by HaCaT sublines was directly correlated with the UVB-irradiation treatment length. Moreover, we also examined the sphere formation ability of these sublines by growing in ultra-low attachment plates containing sphere formation medium. Parental HaCaT cells did not form any sphere but in contrast UVB irradiated HaCaT cells were seeded in 6-well plate and growth was monitored for 8 days by counting viable cells. Growth curve represents data from triplicate experiments (mean ± SEM; n = 3, *p < 0.05). (B) The population doubling time and percent growth increase on 8 th day was calculated as described in materials and methods section. (C) Cells (1000 cells/well) from each subline were plated in 6-well plate and allowed to form colonies. After 2 weeks, colonies were stained with crystal violet, visualized, photographed and counted using imaging system. (D) Control and UVB-transformed HaCaT sublines were seeded at a density of 2.5 × 10 4 cells/ ml in a 0.4% soft agar over a 0.8% agar bottom layer. After 3 weeks, colonies were visualized and counted using Nikon eclipse microscope. Bars represent mean ± SEM; n = 3, *p < 0.05. (E) Single-cell suspensions of HaCaT sublines (1 × 10 3 cells/well) were seeded in ultra-low attachment 6-well plate containing sphere formation medium (1:1, DMEM/F12) supplemented with B27, bFGF (10 ng/mL) and EGF (10 ng/mL). After 2 weeks, spheres were photographed and counted using Nikon Eclipse microscope. formed spheres. Importantly, the sphere formation ability of HaCaT sublines increased with time of UVB irradiation (Fig. 2E). Taken together, these findings clearly suggest that UVB-transformed cells have gained increased growth characteristics, clonogenic potential and sphere-formation ability as compared to parental HaCaT cells.
Increased growth of UVB-transformed HaCaT sublines is associated with enhanced proliferation and apoptosis-resistance. Having observed higher growth and clonogenic potential of UVB-transformed HaCaT sublines, we next examined their cell cycle distribution and apoptotic-index. We observe that all UVB-transformed HaCaT sublines have significantly higher cell cycle progression as compared to the parent cell line as is evident from greater distribution of cells in the S phase of cell cycle (Fig. 3A). Conversely, data from apoptosis assay demonstrate that UVB-transformed HaCaT cells are relatively more resistant to apoptosis as compared to the parent HaCaT cells (Fig. 3B). Notably, increase in S-phase distribution of the derived sublines positively correlated with their UVB exposure frequency (2.15 fold-16wk; 1.93 fold-12wk; 1.74 fold-8wk and 1.45 fold-3wk) (Fig. 3A), while an inverse association was observed for the apoptotic index (2.23 fold-16wk; 1.83 folds-12wk; 1.44 fold-8wk and 1.2 fold-3wk) (Fig. 3B).
A number of proteins are involved in the regulation of cell cycle and apoptosis at the molecular level 15,16 . Therefore, we analyzed the expression profile of these proteins in HaCaT and its derived sublines by immunoblot assay. The data show a progressive decrease in the expression of cyclin dependent kinase inhibitor proteins (p21 and p27) and pro-apoptotic protein i.e. Bax in UVB-transformed HaCaT sublines, whereas the expression of cyclins (Cyclin D1 and Cyclin E) and anti-apoptotic proteins (Bcl2 and Bcl-XL) was increased as compared to the parental HaCaT cells (Fig. 3C). Together, these findings suggest that UVB-transformed HaCaT cell lines exhibit enhanced proliferation rate and more resistance to apoptosis as compared to the parental cells.
HaCaT sublines derived upon UVB-irradiation exhibit greater motility, invasiveness and epithelial-to-mesenchymal transition (EMT). In next set of experiments, we determined the motility and invasive potential of parental and UVB-transformed HaCaT sublines. Our data demonstrate that HaCaT sublines (3wk, 8wk, 12wk and 16wk) are more motile (2.6, 4.2, 6.8 and 8.7 folds, respectively) and invasive (3.5, 6.0, 10.9, and 12.9 folds, respectively) as compared to the parental HaCaT cells (Fig. 4A,B). Since acquisition of motile behavior and invasiveness is associated with transition to mesenchymal phenotype 17,18 , we next analyzed the cytoarchitecture of HaCaT and derived sublines by staining them with FITC-conjugated phalloidin. Microscopic examination demonstrated actin reorganization in UVB-transformed HaCaT sublines characterized by the presence of distinctive filopodia like structures (Fig. 5A). To further confirm the EMT of HaCaT sublines, we examined the expression of EMT-associated marker proteins. We observed a greater expression of mesenchymal markers (N-cadherin, slug and snail) and reduced expression of epithelial marker (E-cadherin) in HaCaT sublines as compared to parental cells (Fig. 5B). Together, these findings suggest that UVB-transformed sublines undergo EMT, and attain a more motile and invasive phenotype.

UVB-irradiated HaCaT cells exhibit tumorigenic potential in mice. Next, we conducted in vivo
study to examine if chronic UVB irradiation of HaCaT cells led to their oncogenic transformation. For this, we injected 16wk-HaCaT subline and parental HaCaT cell line subcutaneously into the immunocompromised nude mice and monitored the growth for 21 weeks. In accordance with the previously published studies 19-21 , we did not observe any tumor formation in mice injected with parental HaCaT cells even after 21 weeks, while there was 100% incidence of tumor formation in mice injected with 16wk-HaCaT subline (Fig. 6A,B). Tumor mass was visually evident at 10 week post-injection, which continue to increase until the experiment was ended (Fig. 6C). Average volume and weight of the developed tumors were 468.5 mm 3 (range from 180 to 1152 mm 3 ) and 0.346 g (range from 0.2 to 0.7 g), respectively (Fig. 6C,D). Histological examination of tumor by H&E staining revealed the feature of well differentiated squamous cell carcinoma with multilayered, hyper proliferative, stratified epithelium exhibiting prominent parakeratosis (Fig. 6E). Interestingly, in some sections, we also observed the presence of tumor cell nest into the basal connective tissue (Fig. 6E) suggesting the invasive nature of developed tumors. Together, these findings demonstrate that frequent exposure of UVB radiation causes oncogenic transformation of HaCaT cells, and derived subline forms well differentiated invasive squamous cell carcinoma.

Discussion
Epidermal keratinocytes are the most predominant (∼ 95%) cell type present in the outermost layer of skin and therefore are the prime target of UVB radiations 22,23 . The present study developed and characterized a unique and stable in vitro progression model of UVB-induced skin carcinogenesis by using HaCaT cells. The most common scenario of UV exposure that could be associated with skin carcinogenesis in human is the chronic/repetitive exposure of keratinocytes to UV from sunlight during recreational sun bathing activities or from tanning beds 24,25 . According to published studies, around 40 mJ/cm 2 of UVB is a minimal erythemal dose 26,27 , which causes DNA damage and significant apoptosis in exposed keratinoytes 28 . Therefore, we chose a sub-erythemal dose (30 mJ/cm 2 ) of UVB and allowed repetitive exposure of keratinocytes to reflect a practical scenario of repetitive DNA damage and repair. Consequently, our treatment led progressive visible changes in cellular morphology as would be expected during oncogenic transformation 29 . Besides alteration in morphology, UVB-exposed sublines also exhibited enhanced growth, clonogenic potential and sphere-forming capacity as compared to the control HaCaT cells.
Published study suggest that UVB at low doses (2.5 to 10 mJ/cm 2 ) induces HaCaT cell proliferation without noticeable cell death 30 , whereas a single exposure at higher does (> 20 mJ/cm 2 ) inhibits proliferation and survival 23,31 . However, we did not see any growth stimulation in our study (data not shown) even at low UVB doses, but observed noticeable death of HaCaT cells at higher doses (20-40 mJ/cm 2 ). Therefore, enhanced growth of remaining HaCaT cells upon repetitive UVB-exposure in our study is likely due to the accumulation of deleterious mutations and/or activation of tumor-promoting signaling pathways in these cells.
We also observed increased cell-cycle progression and apoptosis resistance, the key characteristics of a tumor cell, in UVB-transformed HaCaT sublines. Our results are consistent with several previously published reports that also highlight the association of increased cell-cycle progression and apoptotic resistance with the enhanced growth potential of a cell 30,32,33 . At molecular level cell-cycle process is tightly regulated by specific proteins cyclins, cyclins-dependent kinases (CDKs) and inhibitors of CDKs 15,16 . Similarly, a fine balance of pro-and anti-apoptotic proteins regulates the cellular apoptosis 34 . Higher expression of cyclins (Cyclin D1 and E) and anti-apoptotic proteins (Bcl2 and Bcl-XL) and reduced expression of CDK inhibitors (p21 and p27) and pro-apoptotic proteins (Bax) in UVB-transformed cells may thus, provide the molecular basis for the enhanced proliferation and survival in these cells. In corroboration to these observations, Han and He (2009) have also demonstrated the increased G1-S phase cell-cycle progression in human keratinocytes upon UVB exposure that correlated with the increased expression of cyclin D1 30 . A study performed by Liang et al. (2000) provides direct support to our findings 35 , suggesting that increased expression of cyclin D1 is an early event in skin cancer, and its overexpression was further suggested to be associated with sun exposure. Moreover, altered expression of Bcl2 family member proteins including Bcl2 and Bax in skin tumors as compared to case-matched nonneoplastic skin samples has been observed, clearly suggesting their critical roles in skin carcinogenesis 36 . Furthermore, the mesenchymal phenotype is directly associated with the aggressiveness of tumor cells and transition of a cell from epithelial to mesenchymal stage is considered as an important phenomena of tumor progression 17,37 . In agreement to this, we also observed that UVB-transformed HaCaT cells exhibiting aggressive phenotypes also had mesenchymal characteristics including high expression of N-cadherin, Slug and Snail as compared to their normal counterparts. It has been reported that UVB exposure results in loss of E-cadherin and compromised E-cadherin-beta-catenin signaling in HaCaT cells 38 . Further, role of Snail and Slug, the transcriptional repressors of E-cadherin, in promoting EMT has been well documented in a number of tumors including skin cancer 39,40 . Choudhary et al. (2013) investigated the effect of UVB-irradiation on EMT in UVB-induced tumors in SKH-1 hairless mice and found the elevated level of mesenchymal markers like N-cadherin, snail, slug and twist while reduced expression of epithelial marker E-cadherin in UVB-induced tumors 41 . Together, these studies clearly support our findings that UVB causes EMT and thus enhances aggressive phenotypes in HaCaT cells.
Our study also provided a convincing evidence of malignant transformation of HaCaT cells that were chronically exposed to UVB radiation. All the mice injected with UVB irradiated HaCaT subline developed tumors, whereas no tumor formation was observed in any of the mice implanted with control HaCaT cells. This is consistent with published data demonstrating non-tumorigenic nature of HaCaT cells 20,26 . On the other hand, studies have shown that long-term thermal stress or activation of stromal cells could potentially induce tumorigenic conversion of HaCaT cells 20,42 . The tumors developed by UVB-transformed HaCaT cells in our study were histologically well differentiated SCCs and aggressively invaded into the surrounding tissue. Similarly, earlier studies also showed tumorigenic conversion of HaCaT cells by chronic UVA exposure or deregulated Nf-κ B signaling form well differentiated SCC and invasive tumors in nude mice 29,43 . However, only benign tumor formation has been reported in elevated temperature transformed HaCaT cells without any signs of local invasion that suggests the more aggressive nature of UVB transformed HaCaT cells 20 .
In conclusion, we have developed a novel in vitro model system for the UVB-induced skin carcinogenesis. Our data provides evidence that repetitive exposure of UVB alone at sub-erythemal doses can cause malignant transformation of human epidermal keratinocytes. This novel in vitro cell line model mimicking in vivo tumor promotion can be useful in dissecting progressive changes in genes and molecular signaling pathways responsible for initiation; progression and development of UVB-induced skin carcinogenesis. The model can also facilitate identification of novel molecular diagnostic/therapeutic/ preventive targets, and be useful for the in vitro screening for the agents against UVB-induced skin malignancies.

Materials and Methods
Cell culture. Immortalized human epidermal keratinocyte (HaCaT) obtained from German Cancer Research Center (Heidelberg, Germany) were maintained as monolayer culture in a humidified atmosphere with 5% CO 2 at 37 °C in Dulbecco Modified Eagle Medium (DMEM; Invitrogen, Carlsbad, CA) supplemented with 10% (v/v) fetal bovine serum (FBS; Atlanta Biologicals, Lawrenceville, GA), penicillin (100 units/mL) and streptomycin (100 μ g/mL) (Invitrogen). For sphere formation assay, cells were grown in DMEM/F12 medium (1:1, Invitrogen) supplemented with B-27, bFGF and EGF (Life technology ™ , Carlsbad, CA). Cells were routinely tested and determined to be free from mycoplasma contamination using MycoSensorPCR assay kit (Stratagene, La Jolla, CA) as per manufacturer's protocol. Short tandem repeats (STR) genotyping was used as a way to authenticate the cell lines. The STR profiling was matched to the Cell Line Integrated Molecular Authentication database (CLIMA) version 0.1.201406 (http://bioinformatics.istge.it/clima/clima_search.php) 44 .

UVB-irradiation of HaCaT cells. For the development of UVB-transformed cell line model, HaCaT
cells were seeded (1 × 10 6 cells/plate) in 60-mm glass Petri-dishes and allowed to grow for 24 h. Thereafter, medium was replaced with phosphate buffer saline (PBS) and cells were exposed to UVB radiation (30 mJ/cm 2 ) using UVA/UVB Research Irradiation Unit (Daavlin, Bryan, OH) once a week. Treatment was continued for various time periods viz. 3, 8, 12 and 16 weeks and after each time points, a set of cells treated cells were separated and maintained in regular medium. Derived sublines after attaining stable features were designated as per treatment period viz. 3wk, 8wk, 12wk and 16wk.
Growth kinetics assay. Parental (NT) and UVB-transformed HaCaT sublines (3wk, 8wk, 12wk and 16wk) were seeded in six-well plates (1 × 10 4 cells/well). Number of cells was counted using Countess ® Automated Cell Counter (Life technology ™ ) on each day up to 8 th day to determine the growth rate.
Thereafter, growth curve was plotted and cell population doubling time (dt) during exponential growth phase (72-120 h) was calculated using the following formula: dt = 0.693 t/ln (Nt/N0), where t is time (in h), Nt is the cell number at time t, and N0 is the cell number at initial time.
Clonogenicity and sphere-formation assays. Anchorage -dependent and -independent clonogenicity assays were performed as described previously 45,46 . For sphere-formation assay, single-cell suspensions of HaCaT sublines (1 × 10 3 cells/well) were seeded in ultra-low attachment 6-well plates (Corning, Inc., Corning, NY) containing sphere-formation medium [DMEM/F12 (1:1), supplemented with B27, bFGF (10 ng/mL) and EGF (10 ng/mL)] and allowed to form spheres for 2 weeks. Following Cell cycle analysis. Cells (5 × 10 5 cells/well) were synchronized by culturing them in serum-free media as described previously 47 . Subsequently, cells were grown in regular media for 24 h, washed, trypsinized and fixed with 70% ethanol overnight at 4 °C. Post fixation, cells were washed and stained with Propidium Iodide using PI/RNase kit (BD Bio Sciences, San Jose, CA) and analyzed by flow-cytometry on a BD-FACS Canto ™ II (BD Bio Sciences). The percentage of cell population in various phases of cell cycle was calculated using Mod Fit LT software (Verity Software House, Topsham, ME). Apoptosis assay. Apoptosis assay was performed using PE Annexin V apoptosis detection kit (BD Biosciences) as described earlier 45 . Briefly, control and UVB-transformed HaCaT cells (5 × 10 5 cells/ well) were seeded in 6-well plate. After 24 h, cells were replenished with fresh medium and further allowed to grow for 96 h. Thereafter, cells were harvested, stained with PE Annexin V and 7AAD (7-Amino-Actinomycin-d) solution as per manufacture's protocol and analyzed by flow cytometry.
Motility and invasion assays. Cells (2 × 10 5 and 5 × 10 4 for migration and invasion, respectively) were seeded in the top chamber of non-coated (for migration) and Matrigel-coated (for invasion) transwell chamber (BD Biosciences), respectively. Medium supplemented with 10% FBS was added to the lower chamber as a chemoattractant. After 16 h of incubation, non-migrated/invaded cells on the upper surface of the membrane were removed and the migrated/invaded cells on the bottom surface were fixed, stained with Diff-Quick cell staining kit (Dade Behring, Newark, DE), mounted on slide and counted under the microscope in 10 random fields at 200X. Phalloidin staining. Phalloidin staining for actin filaments was performed according to previously described procedure 49,50 . Briefly, Cells (1 × 10 4 ) grown on FluoroDish (World Precision Instruments., Sarasota, FL) for overnight were fixed in 4% formaldehyde in PBS for 10 min at room temperature. Next, cells were washed with PBS and permeabilized in 0.1% Triton X-100 prepared in PBS, for 5 min and blocked with antibody diluent for 45 min. F-actin was selectively labeled with Alexafluor 488 phalloidin (Molecular Probes, Invitrogen, Eugene, OR) for 20 min. After washing cells were mounted using Vectashield mounting medium with DAPI. Immunostaining was observed under Nikon Eclipse TE2000-U fluorescent microscope (Nikon Instruments Inc, Melville, NY).

Subcutaneous xenograft mouse tumor model and histological analysis.
Animal studies were carried out in accordance with the standard principles and procedures approved by the Institutional Animal Care and Use Committee (IACUC) of University of South Alabama. Immunodeficient nude female mice (4-to 6-week old) were purchased from Harlan Laboratories; Prattville, AL. Parental and UVB-transformed (16wk) HaCaT cells (1 × 10 6 suspended in 50 μ L of HBSS medium) were injected into the left flank region of mice (n = 5, each group) to test their tumorigenic potential. Tumor growth was measured using Vernier Caliper once a week up to 21 weeks. At the end point, mice were sacrificed by CO 2 asphyxiation and autopsied. Tumors recovered from sacrificed mice were weighed and tumor volume was calculated using the formula: π /6 × (smaller diameter) 2 × (larger diameter). Tissue sections of formalin fixed, paraffin embedded tumors were stained with Hematoxylin and Eosin (H&E) and observed under microscope for histological examination.
Statistical analysis. All the experiments were performed at least three times independently and data expressed as "mean ± SEM". Wherever appropriate, the data were also subjected to unpaired two tailed Student's t-test. p < 0.05 was considered statistically significant.