Differential Tumor Biology Effects of Double-Initiation in a Mouse Skin Chemical Carcinogenesis Model Comparing Wild Type versus Protein Kinase Cepsilon Overexpression Mice

Abstract

Our previous studies showed that protein kinase Cepsilon (PKCɛ) verexpression in mouse skin resulted in metastatic squamous cell carcinoma (SCC) elicited by single 7,12-dimethylbenz(a)anthracene (DMBA)-initiation and 12-O-tetradecanoylphorbol-13-acetate (TPA)-promotion in the absence of preceding papilloma formation as is typically observed in wild type mice. The present study demonstrates that double-DMBA initiation modulates tumor incidence, multiplicity, and latency period in both wild type and PKCɛoverexpression transgenic (PKCɛ-Tg) mice. After 17 weeks (wks) of tumor promotion, a reduction in papilloma multiplicity was observed in double- versus single-DMBA initiated wild type mice. Papilloma multiplicity was inversely correlated with cell death indices of interfollicular keratinocytes, indicating decreased papilloma formation was caused by increased cell death and suggesting the origin of papillomas is in interfollicular epidermis. Double-initiated PKCɛ-Tg mice had accelerated carcinoma formation and cancer incidence in comparison to single-initiated PKCɛ-Tg mice. Morphologic analysis of mouse skin following double initiation and tumor promotion showed a similar if not identical series of events to those previously observed following single initiation and tumor promotion: putative preneoplastic cells were observed arising from hyperplastic hair follicles (HFs) with subsequent cancer cell infiltration into the dermis. Single-initiated PKCɛ-Tg mice exhibited increased mitosis in epidermal cells of HFs during tumor promotion.

Introduction

PKC is a family of phospholipid-dependent serine/threonine kinases. There are ten mammalian PKC isoforms that are divided into three groups (Newton, 2003). The conventional isoforms include PKC-α, -βI, -βII, and -γ. They are activated by diacylglycerol (DAG) or phorbol esters (e.g., TPA) in the presence of Ca²⁺ and anionic phospholipids. PKC-δ, -ɛ, -η, and -θ are novel PKCs. They can be activated by DAG or phorbol esters in the presence of anionic phospholipids but Ca²⁺ is not required. The third class of PKC isoforms is composed of the atypical isoforms PKC-ζ and -λ/τ Their activation does not require DAG, phorbol esters, or Ca²⁺ (Ohno and Nishizuka, 2002). PKC isoforms have been shown to have numerous biologic effects on cells, including modulation of the carcinogenesis process.

Our laboratory has previously demonstrated that DMBA/TPA treatment of transgenic mouse skin overexpressing PKCɛresulted in papilloma-independent SCC (Jansen et al., 2001). Further, our morphologic analysis of preneoplastic lesions suggested that the cell of origin of SCC was derived from the HF (Jansen et al., 2001). The mechanism(s) by which papillomas are prevented and carcinomas are induced are currently unknown. The present study using morphologic analysis of in vivo skin tissues was initiated to develop mechanistic insights into the PKCɛoverexpression phenotype. In the present study mice treated with either one or two doses of DMBA (single- versus double-initiation, respectively) were studied to determine if double-initiation would restore papilloma formation in PKCɛ-Tg mice. Contrary to our expectations, double-initiation had no effect on papilloma formation in PKCɛ-Tg mice (papillomas were not present in either single-or double-initiated mice), whereas double-initiated wild type mice had reduced papilloma formation. The present study analyzed epidermal cell kinetics and tissue morphology to try to explain these findings. Our results provide important new insights into the papilloma-independent SCC phenotype observed in PKCɛ-Tg mice, and we further obtained evidence suggesting papillomas derive primarily from the interfollicular epidermis in wild-type mice.

Materials and Methods

Chemicals

DMBA was purchased from Aldrich (Milwaukee, WI). TPA was purchased from Alexis (San Diego, CA). Karnovsky’s fixative and EMbed 812 were purchased from Electron Microscopy Sciences (Ft Washington, PA). TUNEL assay kit was purchased from Promega Corp. (Madison, WI). Prolong reagent was purchased from Molecular Probes (Eugene, OR).

Mice and Treatment

Transgenic mice on the FVB/N background overexpressing an epitope-tagged protein kinase Cɛ(T7-PKCɛ) in their epidermis using the human keratin 14 promoter were obtained from our own colony: Tg(KRT14-Prkce)215Akv (henceforth called PKCɛ-Tg). The generation of these animals was previously described (Reddig et al., 2000). Unpublished studies have documented that dorsal skin from these mice has normal morphology. The transgenic mice were maintained by mating hemizygous transgenic mice with wild-type FVB/N mice. The transgene was detected by polymerase chain reaction (PCR) analysis using genomic DNA isolated from 1-cm tail clips. Female mice at age of approximately 10 wks were shaved and used for the tumor promotion experiments in the present studies. DMBA (dose of 100 nmol in 200 μl acetone/mouse) and TPA (dose of 5 nmol in 200 μl acetone/mouse) were used as tumor initiator and promoter, respectively. In this communication we performed traditional and modified two-stage tumor promotion protocols on PKCɛ-Tg mice and their wild type littermates as illustrated in Figure 1, and the details of the experimental procedures are discussed in the following three sections. In each experiment, the mice were euthanized with halothane after the last treatment, and the dorsal skin was excised promptly. Animal care protocols and experimental procedures were approved by the University of Wisconsin Institutional Animal Care Committee.

Study Design

Short Term Single- and Double-Initiation Protocols

The single-initiation protocol shown in Figure 1A adopted the traditional two-stage tumor promotion regime: the mice were treated topically with a single dose of DMBA followed (1 wk later) by 1 × TPA treatment (designated as 1D-1T (D, DMBA; T, TPA), n = 3 per genotype per time point (24, 48, and 72 h)). The mice in control groups did not receive TPA treatment (designated as 1D, n = 3 per genotype). The double-initiation protocol in Figure 1A was a modified two-stage tumor promotion regime: the mice were initiated with one dose of DMBA, followed by one dose of TPA 1 wk later. These latter mice were rested for 1 wk and initiated with a second dose of DMBA, then treated with 1x TPA one wk later (designated as 1D-1T-1D-1T, n = 3 per genotype per time point (24, 48, and 72 h)). The mice in control groups did not receive further TPA treatment (designated as 1D-1T-1D, n = 3 per genotype).

Long Term (17 wks) Single- and Double-Initiation Protocols

In the single-initiation protocol shown in Figure 1B, the mice were treated topically with a single dose of DMBA followed (1 wk later) by TPA treatment twice weekly. This group was designated as 1D-nT (n, # of TPA treatments); 15 wild type and 17 PKCɛ-Tg mice were analyzed. Control groups for the single-initiation study were composed of mice without TPA treatment (designated as 1D, n = 3 per genotype). In the double-initiation protocol in Figure 1B, the mice were initiated with one dose of DMBA, followed by one dose of TPA 1 wk later. These latter mice were rested for 1 wk and initiated with a second dose of DMBA, followed (1 wk later) by TPA treatment twice weekly (designated as 1D-1T-1D-nT); 17 wild type and 21 PKCɛ-Tg mice were analyzed.

The mice in control groups did not receive further TPA treatment (designated as 1D-1T-1D, n = 3 per genotype). The experiments were terminated at 17 wks prior to the plateau phase of papilloma formation in the single DMBA-initiated wild type mice (the average number was approximately 20 papillomas/mouse after 22 wks of tumor promotion) (Reddig et al., 2000) because the PKCɛ-Tg mice suffered from severe dermatitis in the neck region, at least partially resulting from scratching. Since papillomas or carcinomas appeared only in the dorsal posterior region where DMBA/TPA had been applied, the dermatitis did not interfere with the analysis of results from the present study. Mice from single- or double-initiated groups started and terminated treatments concurrently; therefore, double-initiated mice received 4 doses of TPA less than the single-initiated mice since the second initiation event took 2 wks.

Long Term (9 or 20 wks) Single-Initiation Protocols

In these experiments, mice were treated topically with a single dose of DMBA followed (1 wk later) by TPA treatment twice weekly for 9 wks or 20 wks (Figure 1C). Mice were sacrificed at 45 min, 2, 6, 24, 48 and 72 h after the last TPA treatment (n = 3 per time point per genotype). Control groups were composed of mice without TPA treatment (n = 3 per genotype).

Assessment of Apoptotic and Mitotic Cells

Mouse skin samples were prepared for morphologic studies, and cell mitosis, apoptosis, and necrosis were counted as described previously (Li et al., 2005). Briefly, tissues fixed in Karnovsky’s fixative and embedded in EMbed 812 resin were sectioned (1 μm) and stained with toluidine blue. These semi-thin sections allowed more accurate assessment of mitosis or apoptosis than conventional hematoxylin and eosin (H&E) staining of 5 μm paraffin sections. A total of 500 cells (~50 cells/microscopic field at 400 × magnification) from interfollicular epidermis and 500 cells from follicular epidermis were counted for each mouse. The results are presented as percentages of total cells for mitosis, apoptosis, or cell death (apoptosis plus necrosis).

TUNEL Assay

To confirm the morphology studies, an alternative approach was used. Treated mouse skin was analyzed for the presence of apoptotic cells by the terminal deoxynucleotidyl transferase mediated dUTP nick end labeling (TUNEL) assay (Gavrieli et al., 1992). This assay provides a measure of apoptosis and readily identifies fragmented DNA. The TUNEL assay was performed using a commercial kit according to the manufacturer’s protocol, as described previously (Nelson et al., 1992). Briefly, 5 μm sections were deparaffinized and fixed in 4% paraformaldehyde at room temperature (RT) for 5 min. The slides were then treated with 20 μg/ml proteinase K for 10 min and permeabilized by incubation with 0.5% Triton X-100 in phosphate-buffered saline (PBS) for 5 min at RT. After being rinsed twice with PBS for 5 min, the slides were incubated with reaction buffer containing terminal deoxynucleotidyl transferase and fluorescein-12-dUTP in a humid atmosphere at 37°C for 1 h.

EDTA was added to the slide for 5 min to stop the reaction, and the slides were washed three times with PBS for 5 min and stained with 10 μg/ml propidium iodide for 10 min. Finally, after 3 washes with PBS for 5 min, coverslips were mounted with Prolong reagent to prevent fluorescence bleaching during analysis, and fragmented DNA was identified by measurement of incorporated fluorescein-12-dUTP. In negative controls, the TdT enzyme was omitted. The slides were examined with an Olympus Inverted System Microscope IX70 (Melville, NY).

Correlation Studies Between Cell Death and Papilloma Yields in Mouse Skin

Percent cell death in wild type and PKCɛ-Tg mice at 48 h after single- (1D-1T) or double (1D-1T-1D-1T) initiation were normalized to % cell death in wild type mice at 48 h after 1D-1T treatment and the ratios were calculated. The base-10 logarithm of the ratios was plotted against the yields of papillomas developed in single- or double-initiated wild-type or double-initiated PKCɛ-Tg mice. Statistical significance was determined by linear regression analysis (SPSS 13.0 software). Significance was set at p ≤ 0.05.

Statistical Analysis

Percent apoptosis, % mitosis, and % cell death were analyzed at different time points after various tumor promotion protocols, and expressed as mean ± SEM. Statistical significance was determined with analysis of variance ANOVA)/LSD tests (SAS system). Significance was set at p ≤ 0.05.

Results

Papilloma Formation was Suppressed in Double-Initiated Wild Type Mice and Absent in PKCɛ-Tg Mice

Our previous studies showed that treatment of PKCɛ-Tg mice with 1× DMBA-1x TPA (1D-1T) with one week between initiation and promotion resulted in denudation of the surface epidermis, an event that presumably eliminated DMBA-initiated cells that we and others have hypothesized to be the precursor cells for papillomas. Initial loss of the surface epidermis was followed by regeneration of a hyperplastic and poorly differentiated interfollicular epidermis within 72 h after TPA treatment (Li et al., 2005). In order to investigate whether challenging regenerated PKCɛ-Tg mouse skin with an additional DMBA dose (in an attempt to generate new initiated cells) would restore susceptibility to papilloma development, a double-initiation protocol was performed as described in Materials and Methods and illustrated in our study design (Figure 1). Tumor incidences/multiplicities were recorded every week. Once tumors had formed, the respective gross appearances of papillomas in wild type mice or carcinomas in PKCɛ-Tg mice were indistinguishable in single- vs. double-initiated mouse skin. The data for single-initiated wild type or PKCɛ-Tg mice were similar to data already published from our laboratory (Jansen et al., 2001; Reddig et al., 2000), demonstrating the reproducibility of our system.

In wild type mice, papillomas were observed in single-and double-initiated groups, but malignant conversion of papillomas to SCC had not occurred by wk 17 in either group. A remarkable reduction of papilloma multiplicity (6 to 11-fold) and incidence was noted in double-initiated compared to single-initiated wild type mice (Figure 2A). At the time tumor promotion was terminated, single-initiated wild type mice had an average of 14.9 papillomas per mouse, whereas double-initiated wild type mice had an average of 3.6 papillomas per mouse. 100% of single-initiated wild type mice had papillomas by week 9 of tumor promotion, whereas only 86% of double-initiated wild type mice developed papillomas during the entire time course of the experiment.

After 17 wks of mouse skin tumor promotion, papillomas were not observed after either single- or double-initiation in PKCɛ-Tg. In contrast, extremely rapid development of SCC (as early as 4 doses of TPA following double-initiation) was observed in PKCɛ-Tg mice (Figure 2B), whereas the first SCC did not develop until 22 doses of TPA following single-initiation. At the time tumor promotion was terminated, 50% of double-initiated PKCɛ-Tg mice had SCC compared to 31% SCC incidence in the single-initiated PKCɛ-Tg mice, an overall increase of approximately 63% in SCC incidence (Figure 2B).

Increased Cell Death after the First TPA Treatment May Relate to Reduced Papilloma Formation

To address the question of why double-initiation of wild-type mice decreased the incidence and multiplicity of papillomas, we studied the morphology (Figure 3) and cell turnover kinetics (Figure 4) of skin from mice that received one dose of TPA after the last DMBA treatment. Four time points were examined: control (no TPA treatment), and 24, 48, and 72 h post-administration of the last dose of TPA. Images from the 48 h time point are shown in Figure 3.

In single-initiated wild type mice, control (1D) and single-TPA treated (1D–1T) groups showed low incidence of apoptosis (Figures 3A and 3B, respectively). However, in the double-initiated wild type mice apoptosis was significantly increased in both interfollicular and follicular epidermis of the control group (1D-1T-1D), as well as mice treated with an additional dose of TPA (1D-1T-1D-1T) (Figure 3C). Features of apoptosis, including cell shrinkage and chromatin condensation, were confirmed by electron microscopy (Figure 3C inset), and the terminal deoxynucleotidyl transferase mediated dUTP nick end labeling (TUNEL) assay previously described (Gavrieli et al. 1992) confirmed the presence of DNA fragmentation typical of apoptosis (data not shown). In single-initiated PKCɛ-Tg mice, skin from control mice (1D) (Figure 3D) did not show morphologic differences compared to their wild type control WT littermates (1D) shown in Figure 3A.

However, the skin from both wild type and PKCɛ-Tg mice following TPA treatment showed varying degrees of morphologic change. In single-initiated and single TPA treated PKCɛ-Tg mice (1D–1T), there was a tremendous amount of cell loss, predominantly via cell necrosis in both interfollicular and follicular epidermis at the 48 h time point (Figure 3E). However, the damaged epidermis was rapidly replaced by a hyperplastic and poorly differentiated epidermis at 72 h following the first TPA treatment (Li et al., 2005). In contrast to skin before DMBA/TPA treatment, regenerated hyperplastic skin was less sensitive to the damaging effects of further DMBA/TPA treatment since no significant epidermal destruction was observed following additional treatment with DMBA/TPA (1D-1T-1D-1T) at 48 h (Figure 3F).

Quantitation of mitosis and apoptosis in the interfollicular (Figure 4A) and follicular (Figure 4B) epidermis of wild type mice was performed manually with a light microscope. The percentage of mitotic cells did not change significantly in either the follicular or interfollicular epidermis of single- or double-initiated mice; however, the percentage of apoptosis in both the follicular and interfollicular epidermis increased significantly in double- vs. single-initiated mice at all time points. When comparing PKCɛ-Tg mice, the rate of cell mitosis in the interfollicular epidermis was not significantly different in single- vs. double-initiated mice (Figure 4C); however, mitotic rate in the follicular epidermis of double-initiated mice was significantly increased at most time points when compared to single-initiated mice (Figure 4D). Interestingly, there was a dramatic increase in mitotic rate at 72 h in both the interfollicular and follicular epidermis of single-initiated mice. The single-initiated PKCɛ-Tg mice at 48 h all showed a significant amount of cell damage and necrotic cell death.

Therefore, we have marked the mitosis data at this time point as “N.D.” to represent “Not Detected” for these mice. Because the cell death in both the interfollicular and follicular epidermis of single-initiated PKCɛ-Tg mice was primarily necrotic in nature, we have labeled this group of data as “Cell Death” in order to encompass both necrotic and apoptotic events of cell death. For the most part in both the interfollicular and follicular epidermis of PKCɛ-Tg mice, the amount of cell death/apoptosis was significantly greater in the double-initiated compared to single-initiated mice at matched time points (Figure 4C, 4D).

In comparisons between wild type and PKCɛ-Tg mice, the mitotic rate of single-initiated wild type mice were significantly higher in both the follicular and interfollicular epidermis at 24 and 48 h compared to PKCɛ-Tg mice (Figure 4). Interestingly, the mitotic rate in both the follicular and interfollicular epidermis of single-initiated PKCɛ-Tg mice increased significantly at 72 h when compared to wild type mice. There were few significant differences in mitotic rate in either the follicular or interfollicular epidermis of double-initiated PKCɛ-Tg mice compared to wild type; mitotic rate in the follicular epidermis of PKCɛ-Tg mice showed a small significant increase compared to wild type in control (1D) and 72 h (1D-1T-1D-1T) mice. When comparing rates of apoptosis/cell death, both single- and double-initiated PKCɛ-Tg mice had significantly higher rates of apoptosis/cell death in the interfollicular epidermis compared to time- and treatment-matched wild type mice. The follicular epidermis of PKCɛ-Tg mice showed increased incidence of apoptosis/cell death at 24 and 48 h in the single-initiated and at 24 h in the double-initiated mice compared to wild type (Figure 4).

We hypothesized that increased cell death (apoptosis in wild type mice; necrosis and/or apoptosis in PKCɛ-Tg mice) observed in PKCɛ-Tg mice and double-initiated wild type mice at early stages of tumor promotion may result in increased loss of initiated cells, which in turn would lead to the absence (in PKCɛ-Tg mice) or reduction (in double-initiated wild type mice) of papilloma formation. Based on quantitative data presented in Figure 2A and Figure 4, we analyzed the correlation between % cell death and number of papillomas in mouse skin (Figure 5). This figure demonstrates an inverse relationship between induction of cell death in the interfollicular compartment and the average number of papillomas formed, with a correlation coefficient r ² equal to 0.99 (p ≤ 0.05). These results suggested that reduced papilloma formation was due to increased keratinocyte cell death in both wild type and PKCɛ-Tg mice during tumor promotion, and is also consistent with the hypothesis that the majority of papillomas in wild type mice are derived from the interfollicular epidermis (Morris et al., 2000).

Neoplasia Originated From HF Regions in PKCɛ-Tg mice, and Double-Initiation Treatment Caused Follicular Hyperplasia and Accelerated Carcinoma Formation

To identify the cell of origin of carcinoma and better understand the rapid formation of SCC in double-initiated PKCɛ-Tg mouse skin, histologic examination of paraffin-embedded, H&E stained skin samples from three double-initiated PKCɛ-Tg mice that developed grossly visible cancer at early times after TPA treatments was performed. A representative image of adjacent skin uninvolved by cancer is shown in Figure 6A. Small cancers were focally observed directly arising from hyperplastic HFs via light microscopy, and these squamous cell carcinomas invaded deep into the dermis (Figure 6B, C). Neoplastic lesions were observed immediately adjacent to the HF regions as indicated by arrowheads in Figure 6B, 6C. Cancers often penetrated into muscle (Figure 6C), metastasized to distant tissues such as lymph nodes (Jansen et al., 2001), or grew outward to become visually identifiable gross lesions on the surface of dorsal skin (Figure 6D). Except for the more rapid time course of appearance of carcinoma in double-initiated PKCɛ-Tg mouse skin, the series of pathologic events observed during carcinoma formation was similar if not identical in single- (Jansen et al., 2001) versus double-initiated (this study) mouse skin.

Cell Turnover Kinetics after Chronic TPA Treatment

PKCɛhas been implicated in antiapoptotic function (Basu et al., 2002). To determine whether HF hyperplasia was caused by increased mitosis and/or decreased cell death, percentages of mitosis and apoptosis were calculated using a long-term (chronic) initiation-promotion model (shown in Figure 1C). The chronic model consisted of a 1 × DMBA treatment followed by treatment with TPA twice a week for 9 (18T) or 20 (40T) weeks, with mice being sacrificed at predetermined time points following the final TPA treatment. Apoptotic cells in the interfollicular epidermis of the 18T group increased dramatically in both wild type and PKCɛ-Tg mice, with the percent apoptosis being greatest at 6 h following the final TPA treatment (almost 5-fold higher in wild type (Figure S1A) and 3-fold higher in PKCɛ-Tg (Figure S1B) mice compared to Control); both wild type and PKCɛ-Tg showed a decrease in apoptosis in the remainder of the 18T time course, dropping to levels similar to those found in Control mice.

The % apoptotic cells in the interfollicular epidermis of the 40T treatment groups increased dramatically within 45 min of the final TPA treatment in both wild type (Figure S1A) and PKCɛ-Tg (Figure S1B) mice when compared to Controls; levels of apoptosis began to fall slightly in the PKCɛ-Tg mice, however, they remained significantly higher than the Controls throughout the remainder of the time course. Percentages of mitosis did not increase significantly in 18T or 40T treated wild type mice compared to untreated Controls until 24 or 48 hours (Figure S1A), while the amount of mitosis in the PKCɛ-Tg mice did not significantly change at any time point compared to Controls in either treatment group (Figure S1B). The percentages of apoptosis and mitosis in the follicular epidermis of wild type mice in both the 18T and 40T TPA treatment groups increased almost 6-fold (apoptosis) and 3-fold (mitosis) compared to untreated Control mice and remained elevated at most time points throughout the remainder of the time course (Figure S2A). The percentages of apoptosis in the follicular epidermis of PKCɛ-Tg mouse skin in both the 18T and 40T treatment groups increased significantly compared to Controls at early time points and then dropped back near Control levels at the later time points (Figure S2B). Mitosis was not present in significant percentages in the 18T treatment group until 24 h, while the percent mitosis was greatest at 45 min in the 40T treatment group and remained significantly higher than Controls throughout the 48 h time point.

Cell turnover kinetics rate expressed as the ratio of % mitosis to % apoptosis are graphed in Figure 7. In these analyses, faster cell turnover kinetics rate was indicated as a ratio >1, which correlated with increased cell growth with resultant skin hyperplasia. In contrast, slower cell turnover kinetics rate was indicated as a ratio <1, which correlated with increased cell loss and resultant skin atrophy or injury.

Wild type mice (Figure 7A) showed a faster cell turnover kinetics rate at 72 h after 18T treatment, and at 48 h after 40T treatment in the interfollicular epidermis compared to PKCɛ-Tg mice (Figure 7B) at matched time points. Also, the interfollicular epidermal cells showed a faster turnover rate compared to the follicular epidermal cells of wild type mice at 48 h after 40T treatment. Significantly faster cell turnover kinetics rate occurred in the follicular epidermis of the PKCɛ-Tg mice at 24 and 72 h after 18T treatment, and 45 min, 2 and 48 h after 40T treatment compared to both the PKCɛ-Tg interfollicular epidermis and time- and treatment-matched wild type mice (Figure 7A). The follicular epidermal cells showed faster cell turnover kinetics compared to the interfollicular epidermal cells in the same PKCɛ-Tg mice, with statistical significance achieved at 24 and 72 h after 18T, and 45 min, 2 h, and 48 h after 40T (Figure 7B). Our results are consistent with the hypothesis that the tumor promotion protocol increased cell growth of the interfollicular epidermal cells in wild type mice to a modest extent at selected time points, but increased cell growth of follicular epidermal cells in PKCɛ-Tg mice to a more significant degree over a more prolonged period of time.

Analysis of mutation of ras gene

Evaluation of possible genetic mechanisms for altered cell growth kinetics in transgenic versus wild-type mice was performed by examining the mutation of ras gene using the Xba I restriction fragment length polymorphism analysis PCR technique described previously (Nelson et al. 1992). Mutations of ras were not observed in DNA isolated from mouse skin after a single TPA treatment (Figure S3), and papillomas or carcinomas from chronic TPA treated mice also did not demonstrate ras mutations (data not shown).

Discussion

In this communication, we used a double-initiation approach with subsequent histopathologic and cell turnover kinetics analyses to determine the role of PKCɛin papilloma suppression and cancer formation, as well as to obtain further insights into the biology of SCC formed in PKCɛ-Tg mice after various treatments, and compare the results to those obtained in wild type mice.

Topical application of DMBA to mouse skin results in transformation of keratinocytes. This initiation process typically involves mutation and activation of the H-ras gene (Balmain and Brown, 1988; Balmain et al., 1984). Mathematical modeling of DMBA-induced tumorigenesis predicts that approximately 1% of skin cells may become initiated (Kopp-Schneider and Portier, 1992). After applying the tumor promoter TPA, clones of carcinogen-altered cells are selected and expanded because of resistance to induction of terminal differentiation and stimulation of proliferation (Yuspa et al., 1981). These events have been shown to be involved in the production of papillomas. However, in PKCɛ-Tg mice (single- or double-initiated), or in their wild type littermates after double-initiation, increased cell loss after the first TPA treatment was observed, and may therefore eliminate initiated keratinocytes originally destined to become papillomas.

The difference between wild-type and PKCɛ-Tg mice in the induction of cell death at early stages of tumor promotion may determine differences in papilloma multiplicity. To test this hypothesis, we collected SCC samples and analyzed initiation events. Six frank carcinomas were excised after 14 wks of tumor promotion from single- or double-initiated PKCɛ-Tg mice, and DNA was extracted to study ras gene mutation status using the highly sensitive technique of allele-specific PCR. The PCR results were negative for both ras alleles (H-ras and K-ras; data not shown). One possible explanation for these observations may be epidermal destruction at an early stage of TPA tumor promotion resulting in the elimination of the majority of the ras-mutated cells, which would make the PCR technique unable to detect ras mutations.

Alternatively, ras-mutated cells may be truly absent in PKCɛ-Tg mice and some other mutation(s) may cause initiation. However, conventional PCR studies demonstrated that wild type ras mRNA was elevated in PKCɛ-Tg whole mouse skin (including both dermis and epidermis) in comparison to wild type mouse skin after 1x DMBA, or at 24, 48 and 96 h after 1 × DMBA-1 × TPA treatment (data not shown). Also more wild type ras protein was detected in the HFs as assessed by immunohistochemistry in PKCɛ-Tg as compared to wild type mouse skin after single-initiation followed by chronic TPA treatments (18 × or 40 × TPA, data not shown). These preliminary data may partially explain altered cell kinetics in PKCɛ-Tg versus wild type mouse HFs.

The present study provides evidence that cell kinetics (rate of cell birth in relation to rate of cell death) determines at least partially whether papillomas or carcinomas are formed. In wild type mice cell death in the interfollicular epidermis correlated strongly with the number of papillomas formed. In PKCɛ-Tg mice subjected to single initiation with chronic tumor promotion, more prolonged cell proliferation was observed in the follicular epidermis, contributing to accelerated carcinoma formation. Thus, while the biochemical mechanism(s) by which cell kinetics are affected in different skin compartments depending on PKCɛ levels are unknown, our results suggest that cell kinetics have an effect on tumor biology outcome (papillomas versus carcinomas, tumor incidence and latency).

The biochemical mechanisms(s) by which double DMBA initiation suppresses papilloma formation in wild type and/or PKCɛ-Tg mouse skin is unknown. Certainly DMBA is known to cause DNA damage. However, DMBA is also known to result in the production of oxidative damage products (Oberley et al., 2004), and may kill papilloma precursors via this mechanism(s). Future studies will be necessary to prove the biochemical mechanism(s) by which DMBA double-initiation prevents papilloma formation in wild type mice, yet stimulates carcinoma formation in PKCɛ-Tg mice.

We demonstrated that SCCs rapidly developed in PKCɛ-Tg mice using a double-initiation approach. All early lesions examined by light microscopy appeared to originate from in situ lesions in the HFs in regions of significant hyperplasia. A rapidly enlarging tumor mass often with evidence of extension beyond the anatomic confines of skin (i.e. involving lymph nodes) was observed in both single- and double-initiated PKCɛ-Tg mice. Our data suggest carcinomas from PKCɛ-Tg mice derive from keratinocyte stem cells in the HF bulge regions, whereas carcinomas from wild type mice most likely derive from preexisting papillomas, whose precursor cells probably originated in the interfollicular epidermis.

Our data demonstrates that PKCɛis an important modulator of both proliferation and cell death of keratinocytes. Multiple PKCɛdownstream effectors contributing to formation of SCC in the PKCɛ-Tg mouse skin will be investigated in the future.