Early Epidermal Destruction with Subsequent Epidermal Hyperplasia Is a Unique Feature of the Papilloma-Independent Squamous Cell Carcinoma Phenotype in PKCε Overexpressing Transgenic Mice

Abstract

Protein kinase C epsilon (PKCε) overexpressing transgenic (PKCεTg) mice develop papilloma-independent squamous cell carcinomas (SCC) elicited by 7,12-dimethylbenz[a]anthracene (DMBA) tumor initiation and 12-O-tetradecanoylphorbol-13-acetate (TPA) tumor promotion. We examined whether epidermal cell turnover kinetics was altered during the development of SCC in PKCε Tg mice. Dorsal skin samples were fixed for histological examination. A single application of TPA resulted in extensive infiltration of polymorphonuclear neutrophils (PMNs) into the epidermis at 24 h after TPA treatment in PKCε Tg mice while wild-type (WT) mouse skin showed focal infiltration by PMNs. Complete epidermal necrosis was observed at 48 h in PKCε Tg mice only; at 72 h, epidermal cell regeneration beginning from hair follicles was observed in PKCε Tg mice. Since the first TPA treatment to DMBA-initiated PKCε Tg mouse skin led to epidermal destruction analogous to skin abrasion, we propose the papilloma-independent phenotype may be explained by death of initiated interfollicular cells originally destined to become papillomas. Epidermal destruction did not occur after multiple doses of TPA, presumably reflecting adaptation of epidermis to chronic TPA treatment. Prolonged hyperplasia in the hair follicle may result in the early neoplastic lesions originally described by Jansen et al. (2001) by expanding initiated cells in the hair follicles resulting in the subsequent development of SCC.

Introduction

PKC is a family of phospholipid-dependent serine/threonine kinases with wide tissue distribution that mediates intracellular responses to a variety of stimuli, including chemical tumor promotion agents. PKC isoforms are classified into 3 subfamilies: classical PKCs (α, βI, βII, and λ) are activated by phosphatidylserine (PS), diacylglycerol (DAG), and calcium; novel PKCs (δ, ε, η,θ and μ) require PS and DAG, but are independent of calcium for activation; atypical PKCs (ξ, and ι/γ) need only PS (Newton, 1997; Mellor et al., 1998).

The mouse skin model of multistage carcinogenesis continues to serve as a major in vivo model for analyzing the sequential and stepwise evolution of the cancer process by chemical carcinogens. PKC is a major intracellular receptor for TPA. The phorbol ester TPA is a highly potent tumor promotor that activates some isoforms of the signaling enzyme PKC (Driedger and Blumberg, 1980; Ashendel et al., 1983; Kikkawa et al., 1983; Niedel and Vandenbark, 1983), which in turn may regulate both cell proliferation and cell differentiation (Jaken, 1996). Six PKC isoforms (α, δ, ε, η, ξ, and μ) have been identified in mouse skin epidermis (Leibersperger et al., 1991; Mills et al., 1992; Gschwendt et al., 1992; Osada et al., 1993; Wang et al., 1993; Rennecke et al., 1999). To determine the role of PKCε in skin tumor promotion by TPA, we generated transgenic mice overexpressing an epitope-tagged PKCε (T7-PKCε) under the control of the human keratin 14 promotor (Reddig et al., 2000; Jansen et al., 2001). Compared to WT mice, overexpression of T7-PKCε in the epidermis completely suppressed papilloma formation but resulted in the development of papilloma-independent metastatic squamous cell carcinomas (mSCC) elicited by the DMBA-TPA protocol (Reddig et al., 2000; Jansen et al., 2001).

In the present study, we examined early biological events after a single or multiple topical applications of TPA to DMBA-initiated skin of PKCε Tg mice and their WT littermates. We observed infiltration of polymorphonuclear neutrophils (PMNs) into the epidermis with subsequent extensive damage to the epithelial cells in the epidermis and in some hair follicles of PKCε Tg mice. Following extensive destruction of the interfollicular epidermis, cell proliferation within the hair follicles (Cotsarelis et al., 1990; Taylor et al., 2000) was demonstrated to result in reconstruction of the injured epidermis. During chronic tumor promotion, PKCε overexpression modulated keratinocyte proliferation with resultant epidermal hyperplasia in TPA-treated mice. Our results are discussed in terms of the possibility that epidermal destruction by PMNs and/or biochemical effects of PKCε overexpression in PKCε Tg mice after the first TPA treatment may explain the absence of papillomas in mice chronically treated with TPA. The role of chronic hyperplasia in the hair follicle of the skin in early cancer formation in PKCε transgenic mice is also discussed.

Materials and Methods

Chemicals and Antibodies

TPA was purchased from Alexis (San Diego, CA). DMBA was purchased from Aldrich (Milwaukee, WI). Karnovsky’s fixative was purchased from Electron Microscopy Sciences (Ft. Washington, PA). Monoclonal antibody directed against Keratin 10 (K10) was purchased from Capel (Aurora, Ohio). Monoclonal mouse anti-proliferating cell nuclear antigen (PCNA) was purchased from DAKO (Carpinteria, CA). Rabbit polyclonal antibody to human myeloperoxidase (MPO) was purchased from Biodesign (Saco, ME). Rabbit polyclonal antibody to PKCε was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Normal mouse serum, normal rabbit serum, and the immunoperoxidase LSAB+ kit were purchased from DAKO (Carpinteria, CA). 3,3′-diaminobenzidine tetra-hydrochloride (DAB/metal concentrate) was purchased from Pierce (Rockford, IL).

Mice and Treatment

PKCε Tg mice were generated as described (Reddig et al., 2000; Jansen et al., 2001). All experimental procedures conformed to institutional animal care guidelines. Transgenic mice were maintained by mating hemizygous transgenic mice with wild-type FVB/N mice. The mice were housed in groups of 2–3 in plastic bottom cages in light-, humidity-, and temperature-controlled rooms; food and water were available ad libitum. The animals were kept in a normal rhythm of 12 h light and 12 h dark periods. The transgene was detected by PCR analysis using genomic DNA isolated from 1 cm tail clips.

For mouse skin tumor initiation, a single dose of 100 nmol DMBA in 200 μl acetone was applied topically to the shaved backs of 10-week-old female mice. One week after initiation, a single or multiple doses of 5 nmol TPA in 200 μl acetone were applied twice weekly (72 h apart). Mice not treated with TPA served as control groups (0 h). A diagram depicting the treatment regime is shown in Figure 1. During tumor promotion, we observed that PKCε Tg mice lost hair on the dorsal skin where TPA was applied, whereas the WT mice regrew their hair (data not shown), so we could not accurately define the stage of the hair cycle when comparing these 2 groups. Mice were euthanatized at indicated time points (n = 3 per genotype per time point) after the last TPA treatment.

Immunohistochemistry (IHC)

The tissues to be examined were excised promptly after euthanasia and immediately placed in 10% neutral-buffered formalin (n = 3 per genotype per time point). Tissues were fixed for 2 h in formalin and processed for paraffin embedding following standard procedures. Tissues were sectioned (4 μm) for antibody staining. The primary antibodies used required antigen retrieval pretreatment by incubating samples in 90°C Tris (1 M)–urea (0.8 M) solution for 35 min. The slides were incubated overnight at 4°C with primary antibodies. K10 and PCNA antibodies were used at dilution of 1:10; PKCε antibody specificity was determined by examining immunoblots of epidermis (scraped off from dorsal skin of PKCε Tg and WT mice) extracts probed with anti-PKCε antibody. The results showed a single band of the appropriate molecular weight (90 KDa), and PKCε levels were significantly increased in PKCε Tg mice (Wheeler et al., 2003). PKCε antibody was used at dilution of 1:50 in IHC studies. Subsequent incubation steps were performed in a moist chamber at room temperature (RT). After intermediate washing steps in Tris buffered saline (TBS, 0.05 M, pH 7.4), the sections were incubated with biotin-labeled rabbit anti-mouse or goat anti-rabbit IgG (15 min, RT), and then with streptavidin-peroxidase complexes for 15 min at RT according to the manufacturer’s specifications (DAKO). Visualization was performed using DAB as a substrate for the peroxidase reactions, slides were transferred into tap water, and counter-stained with hematoxylin for 4 min. Negative controls were included for each study and used either mouse (monoclonal) or rabbit (polyclonal) normal serum. No immunoreactivity was observed in these control sections. Specimens were analyzed using an Olympus BX 51 microscope. Ten sites were randomly selected (2 sites from each strip of skin, 5–6 strips of skin for each mouse) for each mouse at each time point. Pictures were taken with a Nikon 35 mm camera.

Routine Electron Microscopy

Procedures for routine electron microscopy (EM) have been previously described in detail (Gonzalez et al., 1989). Briefly, tissues were fixed in Karnovsky’s fixative and post-fixed in osmium tetroxide. Tissues were then dehydrated in an ethanol series and embedded in Epon 812. Thin sections for electron microscopy were cut with a LKB (Bromma, Sweden) ultramicrotome. Copper grids were stained with lead citrate and uranyl acetate and photographed with a Hitachi (Tokyo, Japan) electron microscope.

Assessment of Apoptotic and Mitotic Cells Using Epon-Embedded Semithin Sections

Mouse skin was excised immediately after euthanasia, and ~1 mm width strips were cut with razor blades. Approximately 30 strips were randomly selected and fixed in Karnovsky’s fixative, and post-fixed in osmium tetroxide. Tissues were then dehydrated in an ethanol series and 7 strips were randomly selected for embedding in Epon 812. Three Epon embedded skin samples were randomly selected for semithin sectioning (~1 μm) using a LKB (Bromma, Sweden) ultramicrotome. The sections were stained with toluidine blue and mitosis or cell death counted in a blinded fashion using a light microscope. Tissues were scanned at ×100 magnification, and an area randomly selected for cell counting at ×400 magnification. In the interfollicular epidermis, the numbers of apoptosis/necrosis/mitosis were analyzed randomly from stratum basale, stratum spinosum, and stratum granulosum. A total of 500 cells in the interfollicular epidermis were counted for each mouse (n = 3 per genotype per time point). In the follicular epidermis, cells from both cross sections and longitudinal sections were analyzed; a total of 500 cells for each mouse was obtained from at least 10 hair follicles (n = 3 per genotype per time point). Characteristics defining apoptosis were typical morphological changes including shrinkage of the cell and hypercondensation and fragmentation of nuclear chromatin. Cells in all phases of mitosis (prophase, metaphase, anaphase, telophase, and cytokinesis) were included in mitotic counts. Necrosis was classified as cells with cytoplasmic swelling and nuclear karyolysis.

Assessment of Skin Thickness

For the epidermal hyperplasia study, Epon-embedded skin samples from different time points were examined. Vertical sections were cut and stained with toluidine blue; digital pictures of 10 arbitrarily selected fields were taken at a magnification of ×400. For each site of skin, the thickness of the epidermis from the basal layer up to but excluding the stratum corneum was measured using “Microsoft Photo Editor” software; if skin was embedded in an imperfect orthogonal plane position, we used right triangle trigonometry (a²+ b² = c², angle C is a right angle) to determine the correct numbers. If the mice developed skin papillomas or carcinomas, only uninvolved skin thickness was measured. Ten measurements (unit: pixel) per mouse were obtained; n = 3 per genotype per time points were analyzed. To convert pixels to μm, 10 pictures of a transparent ruler were taken at ×20, ×40, ×100, respectively, and the factor between true length and digital file size in pixel numbers was determined; at ×400 magnification, 1 mm in length was found to equal 3810 pixels.

Statistical Analysis

Values obtained at different time points (mean ± SEM) were independent of each other and so standard analyses could be performed. Comparison of cell death or mitosis was made between the following pairs: results from single-TPA treated WT mice at various time points (12, 24, 48, 72, and 96 h) versus results from WT control group (0 h); results from single-TPA treated PKCε Tg mice at various time points (12, 24, 48, 72, and 96 h) versus results from PKCε Tg control group (0 h); results from WT versus PKCε Tg mice at matched time points. Cell kinetics were analyzed by ANOVA/LSD tests (SAS system), while the ratio between cell mitosis and cell death was analyzed by R software. p ≤ 0.05 were considered significant.

Results

Greater Basal and TPA-Induced Levels of PKCε in PKCε Tg than in WT Mouse Epidermis

We reported previously that FVB/N transgenic mice over-expressed (~18-fold) epitope-tagged PKCε (T7-PKCε) immunoreactive protein in scraped epidermis as analyzed by western blot analysis (Reddig et al., 2000). TPA, a stable diacylglycerol analog, is a potent PKCε activator. We examined whether a single TPA treatment modulated the expression and/or subcellular distribution of PKCε. An immunohistochemistry study using PKCε antibody was performed on formalin-fixed, paraffin-embedded mouse skin samples. As shown in Figure 2 (A and B), PKCε immunoreactivity was observed in both epidermal cell layer and hair follicle keratinocytes, the latter predominantly in keratinocytes of the inner root sheath. Connective tissue and dermis were negative. Interfollicular and follicular epidermis appeared to stain equally in each strain of mice, but the staining intensity was greater in PKCε Tg than WT mouse epidermis at all time points (interfollicular and follicular epidermis).

WT mouse skin had only trace labeling at 0 h. After the first TPA treatment, PKCε expression level was increased in both genotypes, with a peak time point between 48 and 72 h, and the immunoreactive protein staining decreased at 96 h after TPA application. At 96 h in PKCε Tg mouse skin, follicular epidermis showed significantly more PKCε immunostaining than interfollicular epidermis. Preliminary quantitative RT-PCR studies performed on liquid nitrogen frozen-mouse skin samples (n = 2 per genotype per time point) also confirmed that a single TPA application up-regulated PKCε mRNA expression in PKCε Tg mouse skin at 24 and 48 h, but the expression level declined at 72 h; in WT mice, PKCε mRNA was increased at 24, 48, and 72 h, but was at lower levels compared to the levels observed in PKCε Tg mouse skin samples (data not shown).

To further clarify the relationships between the first TPA treatment, skin hyperplasia, and PKCε transcription/translation, epidermal thickness was measured (Figure 2C). In WT mice, epidermal thickness increased as a function of time after 1× TPA treatment (24, 48, and 72 h), then decreased at 96 h to a level similar to that of control groups. In PKCε Tg mice, epidermis thickness was comparable to WT epidermis at the beginning time points (0 and 24 h). However, at 48 h, the skin thickness of PKCε Tg mice was dramatically decreased. Skin thickness of PKCε Tg mice increased to statistically significant levels at 72 and 96 h following TPA stimulation in comparison to skin thickness of WT mice. Preliminary data showed that skin thickness returned to the levels of their corresponding control groups within 2 weeks in both genotypes (data not shown). These results suggested that both hyperplasia and TPA stimulation contributed to the increased levels of PKCε protein observed in mouse skin of both genotypes.

A Single Topical Application of DMBA/TPA Resulted in PMN Infiltration of the Epidermis in PKCε Tg Mice with Subsequent Epidermal Destruction

Topical application of TPA to the dorsal epidermis of mice during the process of tumor promotion has been characterized by several different types of responses, with some studies describing infiltration of leukocytes into the epidermis (Reynolds et al., 1998). In our study, we observed varying levels of inflammation in the epidermis after TPA applications in WT and PKCε Tg mouse skin. Morphologic (electron microscopy studies, data not shown) and MPO immunohistochemistry (data not shown) studies demonstrated that the inflammatory cells were PMNs. MPO is an abundant protein in neutrophils, monocytes, and subpopulations of tissue macrophages (Klebanoff et al., 1970), and is believed to play a critical role in host defenses and inflammatory tissue injury. MPO is stored in PMN granules and secreted into the extracellular space and the phagolysosomal compartment following activation. WT mouse epidermis treated with a single application of TPA showed occasional patches of MPO-positive regions primarily on the epidermal surface at 12 and 24 h, but no staining was observed after 48 h. In PKCε Tg mice, the staining intensity within the epidermis was greatly enhanced at 12 and 24 h, and at 48 h the entire interfollicular epidermis was MPO-positive; after 72 h, no staining was detected within the epidermis as the PMNs themselves underwent apoptosis and sloughed off along with the dead epidermis. By gross examination, the whole epidermis was destroyed at 48 h, resulting in a denuded dermis on the backs of the treated animals. The destruction of epidermis was confirmed by electron microscopy as shown in Figure 3; there was evidence of extensive epidermal cell death. After 72 h epidermal regeneration was observed (see later).

Regenerated Epidermis in PKCε Tg Mice Was Less Differentiated and Hyperplastic

Histology examination of PKCε Tg mouse skin revealed that the first TPA treatment caused epidermal destruction-reconstruction events. We hypothesized that the regenerated epidermis might be different from the original epidermis in PKCε Tg mice. We performed immunohistochemistry analysis using the differentiation marker Keratin 10 antibody. Epidermal K10 immunoreactive protein levels were drastically reduced in PKCε Tg mouse skin at 24, 48, 72, and 96 h in interfollicular epidermis (Figure 4B), whereas in WT mice, interfollicular epidermis was stained equally and there were no visual staining differences throughout this time course (Figure 4A). The less differentiated state of skin of PKCε Tg mice was confirmed by EM techniques. This latter analysis showed that there were significantly more nucleated layers in the epidermis of PKCε Tg vs. WT mouse skin as shown in Figure 4C, E at ×2,600 magnification, and PKCε Tg keratinocytes expressed lesser amounts of tonofilaments in their cytoplasm than WT keratinocytes as shown in Figure 4D, F at ×25,900 magnification, indicating less differentiation in PKCε Tg epidermis. These micrographs (C vs. E or D vs. F) were at the same magnification, allowing direct comparisons between groups.

TPA Treatment Caused an Imbalance between Cell Death and Cell Proliferation

We previously reported that PKCε Tg mice developed papilloma-independent mSCC (Jansen et al., 2001). To better understand why PKCε Tg mice did not develop papillomas and yet were more susceptible to the development of mSCC, we analyzed both cell death (apoptosis and/or necrosis) and cell birth (mitosis) rates of WT and PKCε Tg mice after TPA treatments.

Cell turnover kinetics (mitosis or apoptosis/necrosis) were analyzed using routine light microscopy in Epon-embedded 1-μm sections. In WT mice, necrotic cells were not detected at any time points examined; the percentage of cell death was expressed as the number of dead cells (apoptosis without necrosis) per 500 total cells. In comparison to the basal level (WT-0 h), no significant changes in the number of dead cells were observed at 12, 24, 48, 72, and 96 h after the first TPA treatment. However, significant increases (p ≤ 0.05) in mitosis were observed at 24 h in the interfollicular compartment and at 48 h in the follicular compartment in WT mouse skin (Figure 5A). In analysis of ratio data from WT mice, except at 12 h after the first TPA treatment, the overall mitotic rates were greater than cell death rates (ratio >1) from time course studies, but did not reach statistical significance in comparison to WT-0 h (Figure 5B).

In PKCε Tg mice, both apoptosis and necrosis were observed at 12, 24, and 48 h after the first TPA treatment. The presence of necrotic keratinocytes was confirmed by electron microscopy. At 48 h, there was evidence of extensive epidermal cell death; though apoptosis was identified, extensive EM studies showed that cell death was predominantly due to necrosis as demonstrated in Figure 3. To simplify, we defined the total percentage of cell death at 48 h in the interfollicular compartment as 100%. After 48 h, necrotic cells were not observed and epidermal regeneration was present. Combining apoptosis and necrosis to define total cell death (apoptosis% + necrosis% = total cell death%), the overall percentages of dead cells from a total of 500 cells counted in PKCε Tg interfollicular epidermis at 0, 12, 24, 48, 72, and 96 h were calculated (n = 3 per genotype per time point).

As shown in Figure 5A, cell death rates in PKCε Tg interfollicular epidermis were significantly increased after the first TPA treatment at 12, 24, 48, and 72 h compared to PKCε Tg 0 h (p ≤ 0.05), with return to baseline level (0 h) at 96 h. In the follicular compartment, compared to PKCε Tg 0 h, the cell death data showed significant increases (p ≤ 0.05) at 12 and 24 h, with a peak at 48 h, and a return to near baseline levels at 72 h. The extensive cell loss in the epidermis of PKCε Tg mouse skin was compensated by cell proliferation at 72 and 96 h (p ≤ 0.05). In analysis of ratio data from time-course studies of PKCε Tg mouse skin, the increase in cell proliferation response (72 and 96 h) lagged behind the cell death response (12, 24, and 48 h) in epidermis of PKCε Tg mice, especially at 96 h in the follicular region, a time point at which the mitosis rate greatly exceeded the cell death rate (Figure 5B). The cell turnover kinetics data analyzed after the first TPA treatment demonstrated that overexpression of PKCε in the mouse epidermis could modify keratinocyte biology as demonstrated by the presence of cell destruction at early time points and stimulation of cell proliferation at later time points in comparison to cell turnover kinetics data from WT mice.

To further confirm our morphologic cell turnover kinetics results, we performed immunoperoxidase analysis using formalin-fixed, paraffin-embedded skin samples from the same mice with specific antibody against PCNA, which identified cells in the cell cycle by the presence of nuclear staining (Figure 5C, D). Mitotic cells in semithin sections and PCNA-immunostained cells in paraffin-embedded sections showed similar but not identical anatomic distribution. Mitotic cells were present primarily in basal layer and hair follicle regions in both genotypes before and after TPA treatment. At 0 h in both transgenic and WT mouse skin, PCNA immunostaining was mainly detected in nuclei of cells in the hair follicles, and basal layer of interfollicular epidermis, and sporadically detected in nuclei of suprabasal cells of interfollicular epidermis. After a single TPA treatment, increased numbers of PCNA-positive keratinocytes were observed in the hair follicles, basal, and suprabasal layers of interfollicular epidermis. No PCNA immunostaining was observed in dermis and connective tissues before or after TPA treatment. In the PKCε Tg mice, we observed higher (Figure 5D) and prolonged proliferation in epidermis at 72 and 96 h compared to skin of WT littermates at the same time points (Figure 5C). Especially at 48 h, even though severe epidermal destruction was identified, epithelial cells in the hair follicles were morphologically normal and exhibited positive nuclear PCNA staining. We hypothesize that these surviving cells gave rise to a renewed epidermis after 48 h. Histopathological examination of the skin revealed that the regenerated epidermis in PKCε Tg mice was more hyperplastic than epidermis from WT mice.

Adaptive Response of Keratinocytes to Chronic TPA Treatments

To determine the temporal regulation of cell turnover kinetics, we performed a time-course experiment on mouse skin after multiple TPA treatments (4×, 18×, or 40× TPA treatments). We performed morphologic analyses and investigated cell turnover kinetics in the epidermis at the indicated time points following the last TPA treatments in PKCε Tg mice and their WT littermates: 0 h, 45 min, 2 h, 6 h, 24 h, 48 h, and 72 h. Compared to results following the first TPA treatment, less MPO positive staining was observed at 6 and 24 h in both WT and PKCε Tg mouse skin chronically treated with TPA, and PMNs were predominantly present only on the surface of the epidermis, not in an intra-epithelial location. No MPO staining was observed after 48 h in either WT or PKCε Tg mice (data not shown). Furthermore, an almost complete absence of epidermal destruction (apoptosis was the predominant form of cell death at a rate of less than 10%) was observed in PKCε Tg mice treated with 4×, 18×, and 40× TPA topical applications. The apoptosis rate was not significantly different between WT and PKCε Tg mouse skin (data not shown). These observations indicated that the regenerated skin in PKCε Tg mice was less susceptible to TPA-induced inflammatory damage.

PCNA staining (and mitotic cell counting) of chronically TPA treated mouse skin revealed that the keratinocyte proliferation rate was increased in both WT and PKCε Tg mice with increasing topical TPA applications; however, the percentages of PCNA positive cells in follicular or interfollicular epidermis (500 cells counted in each skin compartment) did not show dramatic differences between the two genotypes at various time points (data not shown).

To evaluate the biologic outcome of the dynamics of cell turnover kinetics in these mice, epidermis thickness was quantified (Figure 6). During chronic treatment, hyperplastic epidermis was observed in both WT and PKCε Tg mice. In WT mice, after 4×TPA treatment, epidermis thickness was significantly increased at 6, 24, and 48 h compared to WT-0 h, and decreased at 72 h. After 18× TPA treatments, epidermis thickness was significantly increased at 6 and 24 h compared to WT-0 h, and decreased at 48 h. After 40× TPA treatments, epidermis thickness was significantly increased at 6 h, and decreased at 24 h. These data indicated tight regulation of epidermal thickness in WT mouse skin. In PKCε Tg mouse, after 4× TPA treatments, epidermis thickness was significantly increased at 6, 24, and 48 h compared to PKCε Tg-0 h. The values were also significantly higher at 6 and 24 h when compared to the WT mice at matched time points.

After 40× TPA treatments, epidermis thickness was significantly increased at 24, 48, and 72 h compared to PKCε Tg-0 h, and the values were significantly higher than the values obtained from WT mouse epidermis at matched time points. However, epidermal thickness changes after 18× TPA treatment were quite different from patterns observed after 4× and 40× TPA treatments, and did not reach statistical significance until the 72-h time point compared to results from PKCε Tg-0 h as well as WT mice at the 72-h time point. One explanation for this phenomenon would be that after 18× TPA treatment, both cancer and papillomas started to develop. These pathologic changes at least partially affected skin biology in both genotypes. Our data suggested that inducible hyperplasia is a temporary phenomenon in WT mouse epidermis, whereas hyperplasia is prolonged in PKCε Tg mouse epidermis; although levels do fluctuate after each TPA treatment analyzed in PKCε Tg mice, epidermis from PKCε Tg mice was always (except at 24 h after 18× TPA) thicker than epidermis from WT mice after varying numbers of TPA treatments.

Discussion

It has been reported that overexpression of PKCε in NIH 3T3 cells increased growth rates, resulted in higher cell densities in monolayer culture, and contributed to neoplastic transformation (Mischak et al., 1993). To better understand the role of PKCε in mouse skin carcinogenesis, we have previously generated PKCε Tg mice and found a high incidence of squamous cell carcinomas but complete suppression of papillomas in littermates heterozygous for the PKCε overexpressing allele (Jansen et al., 2001), suggesting direct involvement of PKCε (combination of endogenous and T7-tagged PKCε) (Reddig et al., 2000) in carcinogenesis. In this communication we focused on the relative importance of PKCε in early events in mouse skin tumor promotion after a single DMBA/TPA treatment to try to understand why PKCε Tg mice develop papilloma-independent carcinomas.

We observed that a single application of DMBA followed 1 week later by a single TPA treatment resulted in marked infiltration of skin by PMNs with subsequent epidermal destruction at 48 h, followed by regeneration of the epidermis at 72 h. While it is tempting to speculate that PMN activation in PKCε Tg but not WT mice resulted in epidermal destruction, we cannot rule out the possibility that PKCε overexpression can also contribute directly to the massive epidermal destruction observed. The regenerated epidermis expressed high immunoreactive protein levels of both the Keratin 6 (data not shown) and PCNA proliferation markers, whereas the differentiation marker K10 was absent. The first TPA treatment resulted in more cell death relative to the chronic TPA treatments in both genotypes (data not shown). We hypothesized that the regenerated skin from hair follicle stem cells were the populations that survived following the first TPA treatment and replaced the denuded skin. Their hierarchical progeny may have had the ability to withstand further TPA treatment. However, we do not have biochemical data demonstrating differences between the original skin and regenerated skin. Cell turnover kinetics after chronic TPA treatments demonstrated that PKCε functions as a positive regulator of keratinocyte proliferation in vivo. Enhanced cancer formation may be closely associated with prolonged hyperplasia induced by topical TPA treatments.

In the 2-stage DMBA/TPA protocol, initiation, as achieved by DMBA application, induces mutations in the Ha-ras gene at codon 61, which results in constitutive RAS activity (Bizub et al., 1986; Quintanilla et al., 1986). Although we had reported previously that DMBA-acetone is sufficient to induce carcinoma development in PKCε Tg mice, papillomas or carcinomas were not observed in mice treated with a single dose of DMBA (n = 3) at 20 weeks (defined as control groups) in the current studies, possibly due to the small number of mice analyzed. In the experimental groups, promotion with repetitive TPA treatment facilitates expression of the initiated phenotype by inducing proliferation and ultimately results in the induction of benign papillomatous lesions (Balmain and Harris, 2000). In the later phases of promotion, TPA-induced proliferation results in the accumulation of additional mutations, and/or chromosomal deletions and amplifications that result in the progression of some benign papillomas into squamous cell carcinomas.

An important problem in skin cancer research is the identification of the target cells for chemical and physical carcinogens. Most evidence has placed these target cells in the hair follicles (Binder et al., 1998). To determine the origin of skin tumors, Morris et al. (2000) completely removed the interfollicular epidermis of carcinogen-initiated mice by an abrasion technique known to leave the hair follicles undisturbed. Their results suggested that the target cells for carcinomas and for many papillomas reside in the hair follicles but that target cells for some papillomas are present in the interfollicular epidermis or in the upper infundulum of the hair follicles. In our PKCε Tg mouse model, the first TPA treatment during tumor promotion caused a heavy inflammatory response in the epidermis, and eventual complete destruction of the interfollicular epidermis and partial loss of the follicular epidermis. One attractive hypothesis is that the infiltration of PMNs after the first TPA treatment resulted in elimination of DMBA-initiated keratinocytes via cell death and thus prevented papilloma formation. A further extension of this hypothesis is that some of the precursor cells destined for carcinoma formation reside within the hair follicle bulge region and did not undergo cell death but rather accumulated further epigenetic and genetic damages, eventually giving rise to carcinomas. A previous study has demonstrated putative early neoplastic lesions arising from the hair follicle of PKCε Tg mouse skin (Jansen et al., 2001).

TPA is a pro-inflammatory agent. The activity of TPA appears to directly mimic the natural response of the skin to injury, including the induction of both cytokine release from keratinocyte stores and de novo gene expression, which can mediate or participate in dermatotoxic responses such as inflammation, hyperkeratosis, hypersensitivity, and skin cancer (Oberyszyn et al., 1993; Katayama et al., 1994; Fischer et al., 1995). The present studies showed focal accumulation of PMNs primarily on the epidermal surface in WT mice but within the epidermis in PKCε Tg mice at 12 and 24 hrs after a single TPA treatment. PMNs were biologically activated and targeted to epidermis only in PKCε Tg mice as demonstrated by extensive keratinocyte cell death occurring in PKCε Tg mice at 48 h after a single TPA treatment. The explanation of why TPA caused heavy infiltration of PMNs producing a cytotoxic effect in follicular and interfollicular epidermis in PKCε Tg mice only is not known and is an area for further research. Our results suggest that overexpression of PKCε not only influences epidermal differentiation and proliferation, but also regulates activation of PMNs, leading to the destruction of the epidermis.

PMNs may play a complex and significant role in the carcinogenesis process at least partially because PMNs are a potential source of genotoxic reactive oxygen and/or nitrogen species (Wiseman and Halliwell, 1996). Chemicals found in or released by PMNs may be mutagenic and contribute to the burden of genetic abnormalities associated with tumor progression. As shown in animal models, inflammation enhances the development of colon tumors (Chester et al., 1989). Human diseases involving inflammation of the gastrointestinal tract, such as chronic viral hepatitis, ulcerative colitis, and Barrett’s esophagus, lead to an increased risk of cancer in the corresponding organ sites (Coussens and Werb, 2002; Orlando, 2002; Wang et al., 2002). Our studies showed that PMN infiltration was present following chronic treatments (up to 40× TPA, data not shown) in both genotypes of mice, but did not produce epidermal destruction as observed after the first TPA treatment in PKCε Tg mice. Taken together, we suggest the possibility that inflammation caused by a single TPA application could deplete cellular anti-oxidants and enhance the oxidation of DNA and other cellular constituents in PKCε Tg mice epidermis, and predispose mice to carcinogenesis following chronic TPA treatment. We hypothesize a role for PMN inflammation as an endogenous tumor promoter during the early stages of tumor formation in PKCε Tg mice. Further studies are required to define the linkage between inflammation and genotoxicity and the mutation frequency of sensitive genes such as p53 and ras.

The present study indicates that TPA may cause both apoptosis and cell proliferation (Figure 3 and 5). This result may seem paradoxical but has been demonstrated by another laboratory using a different strain of mouse (Zhao et al., 2002). In this latter study, PMN infiltration was not observed, so the effect of TPA on cell fates can occur independent of PMNs. Further, papillomas, but not squamous cell carcinomas, were observed in this previous study. In the PKCε Tg mouse model, PMNs may further amplify effects on cell death and cell proliferation.

Our studies provide new insights into the possible role of PKCε in modulation of cutaneous inflammation as well as keratinocyte proliferation and differentiation. Evidence indicates that abnormalities of keratinocyte growth and differentiation, together with an influx of inflammatory cells, are found in a number of skin diseases, including psoriasis. Furthermore, alterations in the PKC signal transduction pathways have been identified in psoriasis. The PKCε Tg mice may provide a model to investigate the role of inflammation in skin diseases and tumorigenesis.