Functional roles of Akt signaling in mouse skin tumorigenesis

Article metrics

Abstract

The mouse skin carcinogenesis protocol is a unique model for understanding the molecular events leading to oncogenic transformation. Mutations in the Ha-ras gene, and the presence of functional cyclin D1 and the EGF receptor, have proven to be important in this system. However, the signal transduction pathways connecting these elements during mouse skin carcinogenesis are poorly understood. This paper studies the relevance of the Akt and ERK pathways in the different stages of chemically induced mouse skin tumors. Akt activity increases throughout the entire process, and its early activation is detected prior to increased cyclin D1 expression. ERK activity rises only during the later stages of malignant conversion. The observed early increase in Akt activity appears to be due to raised PI-3K activity. Other factors acting on Akt such as ILK activation and decreased PTEN phosphatase activity appear to be involved at the conversion stage. To further confirm the involvement of Akt in this process, PB keratinocytes were transfected with Akt and subsequently injected into nude mice. The expression of Akt accelerates tumorigenesis and contributes to increased malignancy of these keratinocytes as demonstrated by the rate of appearance, the growth and the histological characteristics of the tumors. Collectively, these data provide evidence that Akt activation is one of the key elements during the different steps of mouse skin tumorigenesis.

Introduction

The two-stage model of mouse skin carcinogenesis has provided an important instrumental framework for the understanding of many current concepts regarding human neoplasia, including the multistage nature of tumor development (reviewed in Conti, 1994; Yuspa, 1994). The development of squamous malignancy has been operationally divided into stages in this model: (i) Initiation, an irreversible and inheritable change with no phenotypic alteration. (ii) Promotion, a hyperproliferative stimulus that leads to the selection and expansion of the initiated cell population, leading to skin papillomas. Depending on their time of appearance and their histological characteristics, these papillomas can be subdivided into early (well differentiated with no signs of malignancy, appearing at weeks 10–15 of treatment), mid (appearing after 15–25 weeks) and late (appearing after 25–35 weeks, frequently with signs of malignant progression). (iii) Conversion, the transition from a premalignant to malignant phenotype characterized by the formation of invasive squamouse cell carcinomas (SCCs). Early studies of chemically-induced mouse skin carcinogenesis showed a high frequency of activation of the Harvey ras oncogene in premalignant papillomas initiated with dimethylbenzantracene (DMBA) and promoted with the phorbol ester 12-O-tetradecanoyl-porbol-13-acetate (TPA). Subsequently a codon 61 mutation of the Ha-ras gene has been established as the molecular signature of the initiation event (Yuspa, 1994).

The signal transduction events that take place during the process of mouse skin tumorigenesis are still poorly understood. Recently, it has been demonstrated that a functional cyclin D1 gene is essential in a pathway connected to the paracrine activation of the epidermal growth factor receptor (EGFR) (Robles et al., 1998). The importance of EGFR is also shown by the decreased size of papillomas after v-rasHa infection of EGFR-deficient primary keratinocytes (Dlugosz et al., 1997). This has been attributed to the capacity of the EGFR pathway to maintain a proliferative pool of basal cells and the prevention of premature terminal differentiation (Hansen et al., 2000). In addition, the development of spontaneous tumors in transgenic mice expressing a dominant form of SOS is inhibited in a null or mutated EGFR background (Sibilia et al., 2000). However, multiple signal transduction pathways are activated as a consequence of Ha ras mutations and/or EGFR activation (reviewed in Campbell et al., 1998; Vojtek and Der, 1998). Their connections to cyclin D1 expression are still a matter of controversy. In this regard, data obtained with cells in culture indicate a major involvement of the raf/MAPK/ERK kinase cascade in the expression of cyclin D1 (Albanese et al., 1995; Liu et al., 1995; Weber et al., 1997; Cheng et al., 1998; Lee et al., 2000). New data emerging from in vitro and transgenic mouse studies also suggest an important role for PI-3K and Akt signaling in this process and in mouse skin transformation (Gille and Downward, 1999; DiGiovanni et al., 2000; Sibilia et al., 2000).

The proto-oncogene serine/threonine kinase Akt, also known as PKB, is a well-known effector of PI-3K, although it can also be activated by other mechanisms (reviewed in Datta et al., 1999). The Akt kinase activity exerts anti-apoptotic and proliferative functions, and consequently its importance in tumor development is becoming increasingly recognized (see Datta et al., 1999, and references therein). PTEN tumor-suppressor gene lipid phosphatase activity acts in opposition to PI-3K function and downregulates Akt kinase activity (reviewed in Cantley and Neel, 1999; Di Cristofano and Pandolfi, 2000; Simpson and Parsons, 2001). In agreement, loss of PTEN function results in Akt activation, leading to protection from several apoptotic stimuli and increased proliferation (Stambolic et al., 1998; Lu et al., 1999; Paramio et al., 1999a; Weng et al., 2001).

This work seeks to determine the relevance of Akt and ERK activities during the different stages of mouse skin carcinogenesis. Biochemical data revealed a primary and major contribution of Akt in this system. In addition, the activation of Akt was seen to proceed through PI-3K activity during the early stages. Finally, the overexpression of Akt increased the tumorigenic properties of PB mouse keratinocytes and conferred a more aggressive malignant phenotype.

Results

Increased Akt activity during mouse skin carcinogenesis

To study the relative contribution of Akt and ERK in mouse skin carcinogenesis, their activities were biochemically determined in control and TPA-treated skin, papillomas arising at different times during the promotion stage (12, 20 and 30 weeks), and SCCs. A remarkable increase was found in the activity of Akt in cytoplasmic extracts (Figure 1A,B). Interestingly, such increase paralleled the process of tumor progression, and in early papillomas reached 10-fold that of normal or TPA-treated skin, and 40–50-fold in late papillomas and SCCs. The activation of Akt is associated with its translocation to the cell nucleus (Andjelkovic et al., 1997; Meier et al., 1997). The Akt activity present in nuclear extracts was therefore investigated, and compared to that observed in the cytoplasm. In all cases, Akt kinase activity was greater in the cytoplasm than in the nucleus (not shown), although the ratio between nuclear and cytoplasmic activities also increases during the process of tumor progression (Figure 1B′). Together, these results demonstrate that Akt activity increases in parallel with the promotion stages of mouse skin carcinogenesis, and remains high during the malignant conversion of papillomas to SCCs. ERK activity was found to significantly increase only in late (30 weeks) papillomas and in SCCs (Figure 1C,D) when compared to control and TPA-treated skin. This may indicate that ERK activity is mainly involved in the late stages of the process, coincident with the malignant conversion from mid to late papillomas or to SCCs. The activity patterns observed for ERK and Akt indicate that ERK activity might be a secondary event in the carcinogenesis process, whereas Akt activity might be a primary event.

Figure 1
figure1

Increased Akt activity during promotion and progression stages of mouse skin tumorigenesis. The in vitro kinase activity of Akt (A) and ERK2 (C) was determined upon immunoprecipitation as described in Materials and methods. The mean value with respect to untreated skin (±s.d.) of the different experiments performed is summarized in (B) and (D). (B′) The in vitro kinase activity of Akt immunoprecipitated from nuclear fractions was also determined and compared to that obtained in cytoplasmic extracts. H2B denotes histone 2B and MBP denotes myelin basic protein, used as substrates in the kinase assays as stated

To determine if the observed differences in the kinetics of Akt and ERK activities were associated with increased expression, Western blot analyses were performed. In addition, cytoplasmic and nuclear extracts were analysed separately to monitor possible changes in their localization. Akt was mostly found in the cytoplasm of normal skin cells (Figure 2). TPA treatment led to an increase in both the cytoplasm and nucleus. Akt expression levels at early, mid or late stages of promotion, or in SCCs, showed no significant changes compared to TPA-treated skin. When the same blots were also used to analyse ERK expression, it was found that this kinase was almost exclusively nuclear. Its levels were very low in control and TPA-treated skin, and in early and mid papillomas, increasing only in late papillomas and SCCs. Finally, the blots were subsequently probed for cyclin D1 and Ha ras expression. In agreement with other authors (Rodriguez-Puebla et al., 1998), Cyclin D1 was exclusively detected in the nuclei of papillomas and SCCs, its levels becoming clearly greater in papillomas as promotion proceeded (Figure 2). An increase of Ha-ras was also observed in the cytoplasm (Figure 2). Similar increases in Ha ras expression during mouse skin tumor progression have previously been reported (Rodriguez-Puebla et al., 1999a) and may represent the increased amplification of mutant Ha ras alleles in SCCs (Bianchi et al., 1990).

Figure 2
figure2

Expression and cellular distribution of Akt, CycD1, Ha ras and ERK2 during mouse skin carcinogenesis. Nuclear and cytoplasmic extracts from the indicated samples including control skin, TPA-treated skin, papillomas obtained at different times after initiation, and SCCs, were analysed by Western blot for the expression of the indicated proteins using specific antibodies

Together, these results indicate that during the process of chemically-induced skin tumorigenesis, the increase in Akt activity is not due mainly to overexpression of the kinase, as its levels are similar in papillomas, SCCs and hyperplastic TPA-treated skin. It also precedes cyclin D1 up-regulation. This is followed by an increase in ERK activity, which could be due to increased Ha ras expression.

Increased Akt activity is due to raised PI-3K activity

The mechanism responsible for Akt activation during the promotion stages of mouse skin chemical carcinogenesis was also analysed. PI-3K activity has been reported the main activator of Akt (Burgering and Coffer, 1995, for a review see Datta et al., 1999). In addition, integrin-linked kinase (ILK) also phosphorylates and activates Akt (Delcommenne et al., 1998; Persad et al., 2000). Interestingly, ILK expression is up-regulated by erbB-2 (Xie et al., 1998) and appears to modulate the levels of cyclin D1 (D'Amico et al., 2000), suggesting a possible role for this enzyme in the process of mouse epidermal tumorigenesis. Finally, PTEN opposes PI-3K, decreasing Akt activity and inhibiting the Ha-ras-induced transformation of NIH3T3 cells (Tolkacheva and Chan, 2000). This may imply the presence of inactivating mutations in the PTEN tumor suppressor gene which affect mouse skin carcinogenesis.

The activity of PI-3K was studied in extracts representing the different promotion and progression stages of mouse skin tumorigenesis, and compared to that of normal or TPA-treated skin. PI-3K kinase activity was found to increase during the promotion stages (Figure 3A,B). Immunoprecipitation experiments demonstrated increased expression of the p85α regulatory subunit of PI-3K (Figure 3C). In addition, the increase in the amount of p85 translated into a greater number of functional p85/p110 dimers, as determined by corresponding Western analysis of p85 immunoprecipitates with p110α/β antibodies (Figure 3D). The kinase activity of ILK was significantly increased only in SCC samples in close parallelism with increased expression of this kinase (Figure 3E,F). Finally, to determine lipid phosphatase activity, PTEN was immunoprecipitated from mouse skin and mouse skin tumor samples, and assayed with radioactive PIP3 generated by transfecting 293T cells with p110α. Similar lipid phosphatase activity of PTEN was observed in control and TPA-treated skin and in papillomas, whereas a mild decrease was observed in some SCCs (Figure 3G,H). Together, these results indicate that increased PI-3K activity might be responsible for the observed activity of Akt in early and mid papillomas, whereas sustained activity in late papillomas and SCCs could result from multiple signals including ILK activation and decreased PTEN activity as well as PI-3K activity.

Figure 3
figure3

Increased PI3-K activity in mouse skin tumors. The PI3-K activity (A,B) analysed from tumor extracts increased during the progression and conversion stages in a manner qualitatively similar to Akt activity (Figure 1A,B). The expression of p85α after immunoprecipitation (C), and the concomitant association between p110α/β and p85α (D) demonstrated by Western blot of the p85α immunoprecipitates, also demonstrated the increased PI3K activity. Increased ILK activity, which may be related to changes in its expression as demonstrated by Western blotting (E,F) and decreased PTEN lipid phosphatase (G,H) were only detected during the conversion stages in SCCs

The observed alterations in PTEN lipid phosphatase activity were then studied in more detail. Northern blot analyses demonstrated no major changes in PTEN gene expression (Figure 4A). In agreement, Western blots using an antibody against the amino terminus of PTEN detected no major changes in PTEN expression in papillomas (Figure 4B). However, in some SCCs (3/5), a fast migrating band was found (arrow in Figure 4B), together with the wt form. Given that the antibody used (N19) specifically recognizes the aminoterminus portion of PTEN, this might suggest the presence of mutations or deletions in the carboxyl terminus of the protein. Such mutations have been reported in some human tumors and appear to affect membrane ruffling but not Akt activity (Leslie et al., 2000). This aspect will be the subject of future investigations using genetic analysis and increased number of samples.

Figure 4
figure4

Expression of PTEN during mouse skin carcinogenesis. (A) The mRNA expression of PTEN and Akt were determined by Northern blotting as described (Paramio et al., 1999a). The multiple band corresponds to multiple polyadenylation signals. The arrow denotes the predominant transcript for PTEN (Furnari et al., 1997) (B). In addition, PTEN protein expression was determined from cytoplasmic extracts such as those shown in Figure 2. 7S rRNA (A) and keratin K14 expression (B) were used to normalize the loading. The arrow in B denotes the fast migrating PTEN suggestive of mutations

Immunohistochemical localization of p85α, Akt and PTEN in tumors

The expressions and localization of p85α, Akt and PTEN were studied in tumor sections representing the different stages of mouse skin carcinogenesis. In early and mid papillomas, p85α was found in the membrane of non-differentiating proliferative cells (Figure 5A). In agreement, Akt was localized in the nuclei of cells in non-differentiating proliferative areas of these tumors, and also in the cytoplasm and nuclei of the differentiating papilloma cells (Figure 5A′). Finally, PTEN immunoreactivity was clearly confined to differentiating areas of the papillomas (Figure 5A′′). In SCC samples a similar staining pattern was found for p85α, Akt and PTEN in their most differentiated areas (Figure 5B, B′ and B′′ respectively). However, in the infiltrative, non-differentiated areas, clear membrane staining for p85α was observed (Figure 5C), associated with strong cytoplasmic and nuclear expression of Akt (Figure 5C′). PTEN expression, however, was almost absent in these areas (Figure 5C′′).

Figure 5
figure5

Immunohistochemical detection of p85a, Akt and PTEN in mouse skin tumors. Formalin-fixed, paraffin-embedded sections from papillomas (A,A′,A′′) and SCCs (B–C′′) were stained for p85α (A,B,C), Akt (A′,B′,C′) and PTEN (A′′,B′′,C′′). Note that the PTEN staining is only lost in non-differentiating infiltrative areas of SCCs (C′′). p85α staining is detected in cell membranes in all tumors. Akt is detected in the cytoplasm and in the nucleus in all tumors. Brown color denotes positive staining in all the cases

Akt overexpression increases keratinocyte tumorigenesis

The above experiments clearly suggest that Akt-dependent signaling could be one of the major contributors to tumor progression and conversion in mouse skin carcinogenesis. To confirm this, in vivo tumorigenic assays were performed using mouse PB keratinocytes. These cells were selected on the basis that they are poorly tumorigenic and give rise to premalignant papillomas upon subcutaneous injection (Yuspa et al., 1986). In addition, in spite of their papilloma origin, they bear no mutations in the Ha ras gene (Harper et al., 1987), and therefore preclude the possible activation of endogenous Akt mediated by the activation of PI-3K by interaction with Ha ras (Rodriguez-Viciana et al., 1996; Rodriguez-Viciana and Downward, 2001).

PB keratinocytes were transfected with wild type Akt. It is worth to mention that wild type Akt was used instead of permanently active Akt (i.e.: myrAkt) to avoid the possible toxic effects observed by the expression of this active form in permanent transfections (C Murga and JS Gutkind, unpublished observations). After selection in G418, pooled clones (40–60) were injected subcutaneously into nude mice, and the rate of appearance of tumors, their growth and histopathological characteristics were studied. After 7 weeks (Figure 6A), vector-transfected control PB keratinocytes produced tumors in a minor percentage of the injected animals (28%, n=18), in agreement with the low reported tumorigenicity of these cells in vivo (Yuspa et al., 1986). However, all the mice injected with Akt-transfected PB keratinocytes developed tumors (n=8). In addition, Akt-transfected keratinocytes gave rise to detectable tumors after only 2–3 weeks (Figure 6A), whereas the control transfected cells required 4–5 weeks. To obtain a measure of the growth rate of the tumors, their volume was determined 26 and 28 days after injection, when control tumors are clearly detectable (Figure 6B). Tumors arising from control transfected cells nearly duplicated their volume over the 2 days (Figure 6B). In contrast, tumors derived from Akt-transfected cells quadruplicated their volume in the same period (Figure 6B), confirming their increased growth rate.

Figure 6
figure6

Increased tumorigenic potential of PB keratinocytes by Akt overexpression. PB keratinocytes were transfected with pcDNA3 or wt Akt. After selection, pooled clones (40–60) were amplified and subsequently injected subcutaneously into nude mice and the appearance of tumors monitored (A). The volume of tumors at day 26 (day 0) and 28 (day 2) was measured. (C–D) Histological appearance of tumors arising after injection of PB keratinocytes transfected with pcDNA3 (C) and Akt (D). Note the absence of differentiation areas, the aggressive morphology and the highly invasive behavior in (D)

Histological examination of the tumors obtained revealed that control-transfected PB keratinocytes gave rise to highly differentiated tumors (Figure 6C). However, those arising from the injection of Akt-transfected PB cells showed a highly malignant phenotype characterized by the absence of differentiation areas and increased invasive behavior. Biochemical analyses confirmed that tumors arising from the injection of Akt-transfected PB cells displayed increased expression of Akt together with increased amounts of active Akt, as demonstrated by Western blotting against Akt and Ser473-phosphorylated Akt (Figure 7A,B).

Figure 7
figure7

Increased Akt expression and activity in xenograft tumors. (A) Protein extracts from one tumor generated after injection of PB keratinocytes transfected with pcDNA3 (Control) and three obtained after injection of Akt transfected PB keratinocytes (Akt-1, Akt-2 and Akt-3) were analysed for the expression of Akt and Ser 473 phosphorylated (active) Akt by Western blotting. (B) Quantitative analysis of the data showed in (A). The phosphorimager values were normalized to the values corresponding to control tumor. Note the increased expression and increased active Akt in tumors from Akt transfected PB keratinocytes

The expression of several differentiation markers was studied in these tumors. It was observed that control cells gave rise to tumors expressing keratins K5 and K6 (Figure 8A,B). The high level of expression of keratin K10 and loricrin (Figure 8C,D) further confirmed the highly differentiated appearance of these tumors. In addition, keratin K13, an early marker of papilloma progression (Gimenez-Conti et al., 1990), was also expressed at a high level in the differentiating areas of these tumors (Figure 8F). Finally, keratin K8, a marker indicative of malignant conversion (Díaz-Guerra et al., 1992; Larcher et al., 1992) was absent in these tumors (Figure 8E). With respect to the tumors generated by injection of Akt transfected keratinocytes, clear expression of keratins K5 and K6 was observed (Figure 8A′,B′). However, in these tumors, the expression of K10 and loricrin was nearly absent (Figure 8C′,D′). Similarly, the expression of keratin K13 was severely reduced when compared to tumors from control-transfected PB keratinocytes (Figure 8F′). However, the vast majority of the cells were keratin K8-positive in the tumors generated by injection of Akt transfected keratinocytes (Figure 8E′). These data demonstrate that Akt overexpression confers not only increased tumorigenic potential to mouse PB keratinocytes, but also a more aggressive malignant phenotype with reduced epidermal differentiation and increased dysplastic and anaplastic areas.

Figure 8
figure8

Altered expression of differentiation markers in xenograft tumors. Tumors generated by subcutaneous injection of pcDNA3 (A–F)- or Akt (A′–F′)-transfected PB keratinocytes were fixed in 70% ethanol and paraffin-embedded. The expression of K5 (A,A′), K6 (B,B′), K10 (C,C′), loricrin (D,D′), K8 (E,E′) and K13 (F,F′) was subsequently monitored using specific antibodies. Note the absence of epidermal differentiation markers (K10 and loricrin) as well as the increased expression of K8, a marker of malignant conversion, in the tumors obtained with Akt-transfected keratinocytes. Brown color denotes positive staining in all the cases

The functions of Akt have been related to the negative control of apoptosis as well as positive modulation of proliferation. The apoptosis and proliferation rate of tumors obtained after injection of control or Akt transfected keratinocytes were therefore studied. Bromodeoxyuridine (BrdU) labeling experiments demonstrated increased proliferation in tumors from Akt-transfected keratinocytes (Figure 9B,E) compared to those from control cells (Figure 9A,E). In addition, the rate of apoptosis was significantly reduced in tumors from Akt-transfected keratinocytes (Figure 9C,D,E).

Figure 9
figure9

Increased proliferation and decreased apoptosis in tumors generated by Akt-transfected PB keratinocytes. The proliferation of tumors (A,B) was analysed by the ability of the cells to incorporate BrdU, whereas apoptosis induction was measured by TUNEL labeling (C,D). The quantitative analysis of both experiments, as well as the statistical significance (Student t-test) is shown in (E). At least six fields (as shown) per tumor of each type were scored for each analysis. Data in (E) are shown as mean±s.d.

Discussion

The use of molecular biology techniques in the study of mouse skin carcinogenesis has provided a more precise localization of the biochemical pathways that regulate tumor phenotype. Skin tumor promotion and progression stages are characterized by selective and sustained cell proliferation, alterations in differentiation and genetic instability, leading to the specific expansion of the initiated, putative stem cells into papillomas that may develop into carcinomas at the conversion stage. The Harvey allele of the ras gene family has been identified as a site frequently involved in initiating mutations (Quintanilla et al., 1986). Among the changes observed during the sequential stages, cycD1 overexpression has been determined as an early marker (Rodriguez-Puebla et al., 1998). This up-regulation has essential roles in mouse skin carcinogenesis since CycD1-deficient mice are highly resistant to carcinogenesis protocols (Robles et al., 1998). On the other hand, CycD1 overexpression in transgenic mice does not act as an oncogene independent of ras activity, and is not sufficient to effect mouse skin tumor development by itself (Rodriguez-Puebla et al., 1999b). This suggests the requirement of complementary factors. Knowing what they are, as well as an understanding of the biochemical pathways that may connect Ha ras and cycD1, are fundamental in comprehending skin carcinogenesis and vital to the development of therapeutic approaches to prevent or eradicate many types of epithelial cancer.

This paper investigates two possible biochemical pathways that might connect the cyclin D1 and Ha-ras mutations in mouse skin tumors: ERK and Akt. The data support the idea that Akt-dependent signaling is the most important, at least in the early stages. In fact, the increased activity of Akt compared to that of ERK (Figure 1), and the fact that PB keratinocytes transfected with Akt behaved in a more aggressive and tumorigenic fashion (Figures 6, 8 and 9), clearly support this hypothesis. In addition the up-regulation of cycD1 expression (Figure 2) is preceded by increased Akt activity (Figure 1), and is followed by increased ERK expression and activity (Figures 1 and 2), suggesting that cycD1 up-regulation is not due to ERK activity in the early stages. The relevance of ERK in this process has been deduced from observations in other systems in which the Ras–Raf–ERK pathway is the key for cycD1 expression (Albanese et al., 1995; Liu et al., 1995; Weber et al., 1997; Cheng et al., 1998; Lee et al., 2000). In some cultured cells, such expression requires PI-3K and Akt signaling (Gille and Downward, 1999), and transgenic mice models have also suggested an important role for this pathway in mouse skin transformation (DiGiovanni et al., 2000; Sibilia et al., 2000). However, the up-regulation of ERK expression and activity also suggests an important role for this cascade in the late stages of tumor development. In this respect, it has also been demonstrated that activated Ral-GDS and raf/MAPK/ERK may act in synergy with PI-3K to stimulate CycD1 transcription (Gille and Downward, 1999).

Several interconnected pathways may activate Akt (Datta et al., 1999), the activity of some of which are investigated in the present work. A clear induction of PI-3K activity was observed (Figure 3). This might be due to the increase of p85α expression, detected from the beginning of the process, together with a greater number of functional p85/p110 dimers (Figure 3). These data indicate that PI-3K activity is involved in the early activation of Akt, although other mechanisms may also be implicated. We have previously demonstrated that the expression of keratin K10 inhibits keratinocyte proliferation (Paramio et al., 1999b), and more recently have associated such inhibition to the repression of Akt activities (Paramio et al., 2001). Therefore, the increased activity of Akt could also be due to the loss of K10 expression in tumors.

It was also observed that ILK expression and activity increase in late papillomas and SCCs (Figure 3), and the lipid phosphatase of PTEN decreases in SCCs (Figure 3), indicating that the sustained activation of Akt in these late stages is also achieved by these activities. These data, in particular the reduced expression of ILK in early and mid papillomas as compared to skin, TPA-treated skin and SCCs (Figure 3E), are in disagreement with the reported increased expression of ILK in pretumoral lesions in transgenic mice expressing erbB2 under the K14 promoter (Xie et al., 1998). This suggests that different carcinogenic events take place in the conventional two-stage chemical carcinogenesis protocols and in these transgenic animals. On the other hand, the data are in agreement with the results of DiGiovanni et al. (2000) who reported that the expression of IGF in the epidermis of transgenic mice, which also confers increased tumor susceptibility, is associated with increased PI-3K and concomitantly Akt activities. Finally, the observed decrease in activity of PTEN lipid phosphatase at the conversion stage only suggests that PTEN is not important during the early stages of keratinocyte transformation. This is in agreement with findings in other types of tumors (see Cantley and Neel, 1999; Di Cristofano and Pandolfi, 2000; Simpson and Parsons, 2001 and references therein). However, the decreased lipid phosphatase activity observed in SCCs indicates that PTEN tumor suppressor functions are compromised at the conversion stage, in agreement with the finding that PTEN may inhibit the Ha-ras-induced transformation of NIH3T3 cells (Tolkacheva and Chan, 2000). These functions might be associated with increased proliferation or decreased apoptosis of tumor cells (Furnari et al., 1998, 1997; Paramio et al., 1999a). More recently, it has been shown that PTEN may also control the angiogenic switch (Wen et al., 2001). The elucidation of such mechanisms in mouse skin carcinogenesis requires further investigation.

Signaling through ILK has also been associated with different tumor types. Here we show increased expression and activity of ILK at the malignant conversion stage. This may mediate increased Akt activity, as ILK phosphorylates Akt and contributes to its activation (Persad et al., 2001). In addition, ILK may also increase beta-catenin-Lef/Tcf-dependent transcription, leading to decreased E-cadherin expression through the repression of Snail (Tan et al., 2001). In the context of mouse skin tumorigenesis, this observation is highly relevant, since the expression of E-cadherin leads to a partial suppression of tumorigenicity in transformed mouse keratinocytes (Navarro et al., 1991), and Snail is associated with the tumorigenic and invasive properties of mouse and human carcinoma cell lines and tumors (Cano et al., 2000).

The present biochemical data and in vivo tumorigenic experiments show the importance of Akt in the process of mouse skin carcinogenesis. The in vivo tumorigenic studies were performed using PB keratinocytes. These parental cells give rise to premalignant papillomas upon subcutaneous injection (Yuspa et al., 1986). However, despite their papilloma origin, they bear no mutations in the Ha ras gene (Harper et al., 1987), indicating that another, yet unknown mechanism(s) may contribute to the premalignant transformation of these cells. In this respect, the increased tumorigenic behavior upon transfection of the wt Akt suggests the existence of internal activities responsible for the activation of the transfected kinase (Figure 7A,B). Moreover, Akt overexpression leads to a more aggressive phenotype, with clear indications of its invasiveness and altered differentiation (Figures 6 and 8), increased proliferation and decreased apoptosis (Figure 9). The well-known functions of Akt in modulating cyclin D1, p21 or p27 levels or their cellular distribution, inducing angiogenesis or inhibits tumor cell apoptosis (Diehl et al., 1998; Lu et al., 1999; Paramio et al., 1999a, 2001; D'Amico et al., 2000; Graff et al., 2000; Collado et al., 2000; Zhou et al., 2001; Weng et al., 2001) may contribute to the observed increase in tumorigenicity. The careful analysis of these aspects may help to determine the molecular mechanisms responsible for the observed tumorigenic capacity of Akt.

Materials and methods

SENCAR mouse skin two-stage carcinogenesis

Dorsal skin of female SENCAR mice (NCI, Frederick Cancer Research Facility, Frederick, MD, USA) in the resting stage of the hair cycle (8–10 weeks of age) was shaved and initiated 3 days later with 7,12-dimethyl-benz[a]anthracene (DMBA) (2,5 μg/mouse in 200 μl of acetone). Promotion with 12-O-tetradecanoylphorbol-13-acetate (TPA) (2 μg/mouse, twice a week) was begun 2 weeks after initiation and continued for 50 weeks. Tumors were harvested 10–50 weeks after initiation and frozen in liquid nitrogen.

Protein extracts

For whole-cell extracts, mouse skin and mouse skin tumors were ground in a mortar in liquid nitrogen and homogenized in buffer P (Tris pH 7.5, NaCl 150 mM, EDTA 1 mM, EGTA 1 mM, β-glycerophosphate 40 mM, sodium orthovanadate 1 mM, PMSF 0.1 mM, aprotinin 2 μg/ml, leupeptin 2 μg/ml, NP-40 1%). After centrifugation at 10 000 r.p.m. for 20 min at 4°C, the supernatants were stored at −70°C or used directly. For nuclear and cytoplasmic fractions, mouse skin and mouse skin tumors were ground in buffer A (10 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM EDTA and 0.1 mM EGTA) containing DTT and protease inhibitors. Cytosolic proteins were obtained by shaking samples on ice for 15 min and separating by brief centrifugation at 2000 r.p.m. To isolate nuclear proteins, the above-obtained pellet was resuspended in hypertonic buffer C (20 mM HEPES pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 25% glycerol) with DTT and protease inhibitors. Nuclear proteins were extracted by shaking samples vigorously on ice for 15 min and separated by centrifugation at 4000 r.p.m. for 5 min. The supernatant was used as the nuclear protein extract. The purity of the cytoplasmic and nuclear fractions was confirmed by Western blotting against β-tubulin (not shown). Protein concentration in each sample was determined using the Bio-Rad protein assay system (Bio-Rad, Richmond, CA, USA).

Northern blotting

Total RNA from freshly harvested mouse epidermis and frozen tumors was isolated by guanidine isothyocianate-phenol-chloroform extraction. Northern blots containing total RNA (15 μg/lane) were probed for expression of Akt and PTEN. DNA probes were prepared by random primed reactions using the complete sequences as previously described (Paramio et al., 1999a). The membranes were also hybridized with a 7S RNA probe to verify that equal amounts of mRNA were loaded and transferred.

Western blotting

Immunoprecipitates, whole-cell proteins, cytosolic and nuclear extracts from mouse skin and mouse skin tumors were resolved in SDS–PAGE gels and transferred to nitrocellulose (Amersham, IL, USA). Ponceau and Laemmli staining were used to normalize the protein amounts loaded per lane. Membranes were blocked with 5% non-fat milk in TBS, and incubated with appropriate antibodies diluted in TBSTB (0.5% BSA, 0.1% Tween 20 in TBS). The following antibodies were used: anti Akt (1/1000 dilution of c-20 antibody, Sta. Cruz Biotechnology, Sta. Cruz, CA, USA), anti Akt phosphorylated in Ser 473 (1/1000 dilution; Cell Signaling Technology Inc., Beverley MA, USA), anti p85α (1/200; Upstate Biotechnology, Lake Placid, NY, USA), anti ERK2 (1/1000 dilution of c-14 antibody; Sta. Cruz Biotechnology), anti Cyclin D1 (1/1000 dilution of H295 antibody, Sta. Cruz Biotechnology), anti ILK (1/1000 dilution, Upstate Biotechnology, Lake Placid, NY), anti PTEN (1/2000 dilution of N19 antibody, Sta. Cruz Biotechnology), anti ras (1/500 dilution of pan ras F132 antibody, Sta. Cruz Biotechnology) and anti p110α/β (a mixture of S-19 and N-16 antibodies diluted 1/1000, Sta. Cruz Biotechnology). Secondary antibodies were anti-rabbit, anti-mouse or anti-goat IgG, purchased from Jackson Immunoresearch (PA) and used 1/2000 in TBSTB. WestPicoSignal (Pierce, Rockford, IL, USA) was used to detect the bands according to the manufacturer's recommendations.

ILK, Akt and MAPK assays

ILK, MAPK and Akt activity in mouse skin or mouse skin tumors was determined after immunoprecipitation with anti ILK (Upstate Biotechnology, Lake Placid, NY, USA; 1 μl/25 μg protein) anti ERK (Sta. Cruz C-14 antibody 1 μl/25 μg protein), and anti Akt (Sta. Cruz, C-20 antibody 1 μl/25 μg protein) essentially as previously described (Coso et al., 1995; Murga et al., 1998), using myelin basic protein (MBP; Sigma, St. Louis, MO, USA) or histone 2B (H2B; Roche Molecular Biochemicals) as substrates for ILK and ERK, and Akt respectively. Autoradiograms were scanned and subsequently quantified using NIH Image version 1.61 software or a Phosphorimager (Bio-Rad).

PI-3K assays

Mouse skin and mouse skin tumors samples were immunoprecipitated with agarose-conjugated 4G10 antiphosphotyrosine antibody (Upstate Biotechnology, Lake Placid, NY, USA). The PI-3K reaction was subsequently performed as previously described (Gutkind et al., 1990). After thin layer chromatography on LK6D plates (Whatman, Ltd., Maidstone, UK), 32P-labeled phospholipids were visualized by autoradiography and quantified using a Phosphorimager (Bio-Rad).

PTEN lipid phosphatase assay

The lipid phosphatase activity of PTEN was assayed upon immunoprecipitation of PTEN, using N-19 antibody (Sta. Cruz Biotechnology, Sta. Cruz, CA, USA; 1 μl/25 μg protein) from mouse skin and mouse skin tumor samples, employing 32P-labeled phosphatidylinositol-3-phosphate (PIP3) as a substrate. Radiolabeled PIP3 was generated upon transfection of 293T cells with cDNA from the catalytic subunit of human PI3Kβ as previously described (Murga et al., 2000).

Cell culture and transfection

PB keratinocytes (Yuspa et al., 1986) were cultured in Dulbecco's modified Eagle's medium (Life Technologies) supplemented with 10% fetal calf serum (BioWhitakker, Walkersville, MD, USA). Transfection was performed by the calcium phosphate method essentially as described by Paramio et al. (1999). The plasmid expressing the wt form of Akt has been previously described (Murga et al., 1998). Upon transfection, clones were selected by growing in the presence of G418 (0.5 mg/ml) for 3–4 weeks.

In vivo tumorigenic assays

Pooled PB keratinocyte clones (40–80 different clones) were grown. Upon trypsinization, cells were washed twice, resuspended in PBS (1×106 cells per 0.1 ml) and subcutaneously injected into nu/nu mice (5–6 weeks females, c57Bl/6 strain background). For each mouse, the right flank was injected with PB keratinocytes transfected with Akt, while the left flank received PB keratinocytes transfected with control vector (pcDNA3). Thus, each mouse received two injections of 1×106 cells. Tumor growth was monitored for up to 7 weeks, at which time the mice were sacrificed. Tumors that formed were measured, excised and processed for histopathological and biochemical analysis.

Immunohistochemistry

Formalin fixed, paraffin-embedded tumor samples were sectioned (6 μm thick). Sections were de-waxed and stained with anti-p85α (1/200), Akt (1/200) and anti PTEN (1/600) overnight at 4°C. After exhaustive washing in PBST, sections were incubated with appropriate HRP-labeled secondary antibodies (all 1/1000 in PBST). Expression of epidermal differentiation markers was performed in ethanol-fixed samples using K8.60 (anti K10 mAb, 1/40; Sigma) anti K5, K6 or loricrin (all 1/500; Babco) Troma1 (anti K8 rat mAb, neat supernatant) or rabbit anti K13 (1/500, a generous gift of Dr DR Roop). For in vivo BrdU labeling, mice received intraperitoneal injections of BrdU (60 mg/g body weight) 1 h before tumor harvesting. Detection was performed by immunohistochemical procedures as previously described (Paramio et al., 1999a,b). Apoptosis was measured by TUNEL analysis using Apoptag (Oncor) following the manufacturer's recommendations. Positive staining was determined using diaminobenzidine (DAB) as a substrate (DAB kit vector, Burlingame, CA, USA) following the manufacturer's recommendations. Sections were then counterstained with hematoxylin and mounted. Observations were performed in a Zeiss Axiophot photomicroscope. At least six different mouse skin or mouse skin tumors of each type were analysed. Controls omitting primary antibodies, or after the preincubation of the antibodies with the immunizing peptide, were routinely performed.

References

  1. Albanese C, Johnson J, Watanabe G, Eklund N, Vu D, Arnold A, Pestell RG . 1995 J. Biol. Chem. 270: 23589–23597

  2. Andjelkovic M, Alessi DR, Meier R, Fernandez A, Lamb NJC, Frech M, Cron P, Cohen P, Lucocq JM, Hemmings BA . 1997 J. Biol. Chem. 272: 31515–31524

  3. Bianchi AB, Aldaz CM, Conti CJ . 1990 Proc. Natl. Acad. Sci. USA 87: 6902–6906

  4. Burgering BM, Coffer PJ . 1995 Nature 376: 599–602

  5. Campbell SL, Khosravi-Far R, Rossman KL, Clark GJ, Der CJ . 1998 Oncogene 17: 1395–1413

  6. Cano A, Perez-Moreno MA, Rodrigo I, Locascio A, Blanco MJ, del Barrio MG, Portillo F, Nieto MA . 2000 Nat. Cell. Biol. 2: 76–83

  7. Cantley LC, Neel BG . 1999 Proc. Natl. Acad. Sci. USA 96: 4240–4245

  8. Cheng M, Sexl V, Sherr CJ, Roussel MF . 1998 Proc. Natl. Acad. Sci. USA 95: 1091–1096

  9. Collado M, Medema RH, Garcia-Cao I, Dubuisson ML, Barradas M, Glassford J, Rivas C, Burgering BM, Serrano M, Lam EW . 2000 J. Biol. Chem. 275: 21960–21968

  10. Conti CJ . 1994 Handbook of mouse mutations with skin and hair abnormalities Sundberg JP (ed) CRC Press, Inc., Boca Raton, Fla pp. 39–46

  11. Coso OA, Chiariello M, Yu JC, Teramoto H, Crespo P, Xu N, Miki T, Gutkind JS . 1995 Cell 81: 1137–1146

  12. D'Amico M, Hulit J, Amanatullah DF, Zafonte BT, Albanese C, Bouzahzah B, Fu M, Augenlicht LH, Donehower LA, Takemaru KI, Moon RT, Davis R, Lisanti MP, Shtutman M, Zhurinsky J, Ben-Ze'evm A, Troussard AA, Dedhar S, Pestell RG . 2000 J. Biol. Chem. 275: 32649–32657

  13. Datta SR, Brunet A, Greenberg ME . 1999 Genes Dev. 13: 2905–2927

  14. Delcommenne M, Tan C, Gray V, Rue L, Woodgett J, Dedhar S . 1998 Proc. Natl. Acad. Sci. USA 95: 11211–11216

  15. Díaz-Guerra M, Haddow S, Bauluz C, Jorcano JL, Cano A, Balmain A, Quintanilla M . 1992 Cancer Res. 52: 680–687

  16. Di Cristofano A, Pandolfi PP . 2000 Cell 100: 387–390

  17. Diehl JA, Cheng M, Roussel MF, Sherr CJ . 1998 Genes Dev. 12: 3499–3511

  18. DiGiovanni J, Bol DK, Wilker E, Beltran L, Carbajal S, Moats S, Ramirez A, Jorcano J, Kiguchi K . 2000 Cancer Res. 60: 1561–1570

  19. Dlugosz AA, Hansen L, Cheng C, Alexander N, Denning MF, Threadgill DW, Magnuson T, Coffey Jr RJ, Yuspa SH . 1997 Cancer Res. 57: 3180–3188

  20. Furnari FB, Lin H, Huang HS, Cavenee WK . 1997 Proc. Natl. Acad. Sci. USA 94: 12479–12484

  21. Furnari FB, Huang HJ, Cavenee WK . 1998 Cancer Res. 58: 5002–5008

  22. Gille H, Downward J . 1999 J. Biol. Chem. 274: 22033–22040

  23. Gimenez-Conti I, Aldaz CM, Bianchi AB, Roop DR, Slaga TJ, Conti CJ . 1990 Carcinogenesis 11: 1995–1999

  24. Graff JR, Konicek BW, McNulty AM, Wang Z, Houck K, Allen S, Paul JD, Hbaiu A, Goode RG, Sandusky GE, Vessella RL, Neubauer BL . 2000 J. Biol. Chem. 275: 24500–24505

  25. Gutkind JS, Lacal PM, Robbins KC . 1990 Mol. Cell. Biol. 10: 3806–3809

  26. Hansen LA, Woodson 2nd RL, Holbus S, Strain K, Lo YC, Yuspa SH . 2000 Cancer Res. 60: 3328–3332

  27. Harper JR, Reynolds SH, Greenhalgh DA, Strickland JE, Lacal JC, Yuspa SH . 1987 Carcinogenesis 8: 1821–1825

  28. Larcher F, Bauluz C, Diaz-Guerra M, Quintanilla M, Conti CJ, Ballestin C, Jorcano JL . 1992 Mol. Carcinog. 6: 112–121

  29. Lee RJ, Albanese C, Fu M, D'Amico M, Lin B, Watanabe G, Haines GK, Siegel PM, Hung MC, Yarden Y, Horowitz JM, Muller WJ, Pestell RG . 2000 Mol. Cell Biol. 20: 672–683

  30. Leslie NR, Gray A, Pass I, Orchiston EA, Downes CP . 2000 Biochem. J. 346: 827–833

  31. Liu JJ, Chao JR, Jiang MC, Ng SY, Yen JJ, Yang-Yen HF . 1995 Mol. Cell Biol. 15: 3654–3663

  32. Lu Y, Lin YZ, LaPushin R, Cuevas B, Fang X, Yu SX, Davies MA, Khan H, Furui T, Mao M, Zinner R, Hung MC, Steck P, Siminovitch K, Mills GB . 1999 Oncogene 18: 7034–7045

  33. Meier R, Alessi DR, Cron P, Hemmings BA . 1997 J. Biol. Chem. 272: 30491–30497

  34. Murga C, Fukuhara S, Gutkind JS . 2000 J. Biol. Chem. 275: 12069–12073

  35. Murga C, Laguinge L, Wetzker R, Cuadrado A, Gutkind JS . 1998 J. Biol. Chem. 273: 19080–19085

  36. Navarro P, Gomez M, Pizarro A, Gamallo C, Quintanilla M, Cano A . 1991 J. Cell Biol. 115: 517–533

  37. Paramio JM, Casanova ML, Segrelles C, Mittnacht S, Lane EB, Jorcano JL . 1999b Mol. Cell. Biol. 19: 3086–3094

  38. Paramio JM, Navarro M, Segrelles C, Gomez-Casero E, Jorcano JL . 1999a Oncogene 18: 7462–7468

  39. Paramio JM, Segrelles C, Ruiz S, Jorcano JL . 2001 Mol. Cell. Biol. 21: 7449–7459

  40. Persad S, Attwell S, Gray V, Delcommenne M, Troussard A, Sanghera J, Dedhar S . 2000 Proc. Natl. Acad. Sci. USA 97: 3207–3212

  41. Persad S, Attwell S, Gray V, Mawji N, Deng JT, Leung D, Yan J, Sanghera J, Walsh MP, Dedhar S . 2001 J. Biol. Chem. 276: 27462–27469

  42. Quintanilla M, Brown K, Ramsden M, Balmain A . 1986 Nature 322: 78–80

  43. Robles AI, Rodriguez-Puebla ML, Glick AB, Trempus C, Hansen L, Sicinski P, Tennant RW, Weinberg RA, Yuspa SH, Conti CJ . 1998 Genes Dev. 12: 2469–2474

  44. Rodriguez-Puebla ML, LaCava M, Bolontrade MF, Russell J, Conti CJ . 1999a Mol. Carcinog. 26: 150–156

  45. Rodriguez-Puebla ML, LaCava M, Conti CJ . 1999b Cell Growth Differ. 10: 467–472

  46. Rodriguez-Puebla ML, LaCava M, Gimenez-Conti IB, Johnson DG, Conti CJ . 1998 Oncogene 17: 2251–2258

  47. Rodriguez-Viciana P, Downward J . 2001 Methods Enzymol. 333: 37–44

  48. Rodriguez-Viciana P, Warne PH, Vanhaesebroeck B, Waterfield MD, Downward J . 1996 EMBO J. 15: 2442–2451

  49. Stambolic V, Suzuki A, de la Pompa JL, Brothers GM, Mirtsos C, Sasaki T, Ruland J, Penninger JM, Siderovski DP, Mak TW . 1998 Cell 95: 29–39

  50. Sibilia M, Fleischmann A, Behrens A, Stingl L, Carroll J, Watt FM, Schlessinger J, Wagner EF . 2000 Cell 102: 211–220

  51. Simpson L, Parsons R . 2001 Exp. Cell. Res. 264: 29–41

  52. Tan C, Costello P, Sanghera J, Dominguez D, Baulida J, de Herreros AG, Dedhar S . 2001 Oncogene 20: 133–140

  53. Tolkacheva T, Chan AM . 2000 Oncogene 19: 680–689

  54. Vojtek AB, Der CJ . 1998 J. Biol. Chem. 273: 19925–19928

  55. Weber JD, Hu W, Jefcoat Jr SC, Raben DM, Baldassare JJ . 1997 J. Biol. Chem. 272: 32966–32971

  56. Wen S, Stolarov J, Myers P, Su DJ, Wigler MH, Tonks NK, Durden DL . 2001 Proc. Natl. Acad. Sci. USA 98: 4622–4627

  57. Weng LP, Brown JL, Eng C . 2001 Hum. Mol. Genet. 10: 599–604

  58. Xie W, Li F, Kudlow JE, Wu C . 1998 Am. J. Pathol. 153: 367–372

  59. Yuspa SH . 1994 Cancer Res. 54: 1178–1189

  60. Yuspa SH, Morgan D, Lichti U, Spangler EF, Michael D, Kilkenny A, Hennings H . 1986 Carcinogenesis 7: 949–958

  61. Zhou BP, Liao Y, Xia W, Spohn B, Lee MH, Hung MC . 2001 Nat. Cell Biol. 3: 245–252

Download references

Acknowledgements

The authors are grateful to M Isabel de los Santos and Pilar Hernández for their excellent work in the histology core facility. This work was partially supported by grants PM98-0039 and SAF 98-0047 from the Spanish Ministry of Science and Technology, 1RO1CA79065-01, and CA79065-02 from NCI/NIH.

Author information

Correspondence to Jesús M Paramio.

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • skin carcinogenesis
  • signal transduction
  • ras
  • Akt
  • PTEN

Further reading