Deregulated expression of cell-cycle proteins during ... - Nature

56 downloads 0 Views 443KB Size Report
7,12-dimethylbenz[a]anthracene (DMBA) induces a specific point mutation at codon 61 of the Ha-ras-1 gene (Balmain and Pragnell, 1983; Bizub et al., 1986;.
Oncogene (1998) 17, 2251 ± 2258 ã 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc

Deregulated expression of cell-cycle proteins during premalignant progression in SENCAR mouse skin Marcelo L Rodriguez-Puebla, Margaret LaCava, Irma B Gimenez-Conti, David G Johnson and Claudio J Conti The University of Texas M.D. Anderson Cancer Center, Science Park-Research Division, Smithville, Texas 78957, USA

It is now evident that several genes encoding regulatory activities that control the mammalian cell cycle, particularly some that control the progression of quiescent cells through G1 and into S phase, are targets for alterations that underlie the development of neoplasms. Here, we made a sequential study of alterations in cell cycle protein expression and complex formation among cyclin, cyclin dependent kinases (CDKs) and CDK inhibitors (CKIs) during premalignant progression in SENCAR mouse skin tumors. Changes in the level of expression were observed in positive (cyclin D1, D2, and E2F family members) and negative regulators (p16Ink4a, p57Kip2) of the cell cycle. Also, we observed the formation of cyclin/CDK/CKI complexes. The amounts of these proteins and complexes increased substantially at speci®c times during promotion but not during malignant conversion to carcinomas. These data show that deregulation of growth control occurs in benign tumors and that subsequent mutations not involved cell-cycle regulation are probably necessary to induce invasive behavior. Keywords: cyclins; cell-cycle; tumor papilloma; squamous cell carcinoma

progression;

Introduction Recently, how the cell cycle is regulated has begun to become clearer. Cyclins bind and activate cyclindependent kinases (CDKs)3, which then phosphorylate various proteins that are important for cell-cycle progression. In fact, cell-cycle progression is regulated by the combinations of cyclins and CDKs, that form complexes at speci®c phases of the cell cycle (Sherr, 1993). A key substrate for G1 cyclin/CDK complexes is the product of the retinoblastoma gene, pRb3. The phosphorylation of pRb, a tumor suppressor gene product, has been attributed to cyclin/CDK complexes and implicated in a proliferation-regulation mechanism in keratinocytes as well as other cell types (Demers et al., 1994; Koike et al., 1994). Loss of Rb function contributes to the loss of cell growth control in various tumors (Weinberg, 1991). There is considerable evidence that the E2F transcription factors are a critical target for pRb and pRb-related proteins (Bagchi et al., 1991; Bandara et al., 1991; Chellappan et al., 1991). The interaction of pRb with E2F

Correspondence: CJ Conti Received 1 April 1998; revised 18 May 1998; accepted 18 May 1998.

correlates with the capacity of pRb to arrest cell growth in G1 phase (Hiebert, 1993; Hiebert et al., 1992; Qian et al., 1992; Qin et al., 1992). Moreover, the oncogenic capacity of the E2F-1 gene was observed in primary rat embryo ®broblast cultures (Johnson et al., 1994). Ampli®cation and overexpression of cyclin D1 has been reported in the pathogenesis of several types of cancer, including carcinomas of human breast, and upper aerodigestive tract and head and neck squamous cell carcinomas (Bartkova et al., 1995b; Jiang et al., 1993; Lammie et al., 1991; Weinstat-Saslow et al., 1995). Early overexpression of cyclin D1 was also observed in the SENCAR mouse skin carcinogenesis model (Bianchi et al., 1993; Robles and Conti, 1995). Cyclin-dependent kinase inhibitors (CKIs) have been described as negative regulators of the cell cycle. These proteins can interact with and inhibit a wide variety of cyclin/CDK holoenzymes, and their overexpression blocks cells in G1 (Guan et al., 1994; Hannon and Beanch, 1994; Hirai et al., 1995). Several abnormalities of CKI genes have been reported. p57Kip2 gene was localized in a cluster of imprinted genes on mouse chromosome 7 (Hatada and Mukai, 1995), and loss of heterozygocity (LOH) in that region has been observed in chemically induced skin carcinogenesis in SENCAR mice (Bianchi et al., 1991). p16Ink4a gene has been found to be deleted and mutated in several tumor-derived cell lines, mouse skin squamous cell carcinomas (SCC)3, and some types of primary tumors (Linardopoulos et al., 1995; Mori et al., 1994; Zhang et al., 1994). The two-stage murine skin carcinogenesis model is a valuable system for studying precancerous changes and tumor progression. In this model, skin tumors are induced by topical treatment of a single dose of a genotoxic carcinogen (initiator) followed by multiple applications of a non-genotoxic tumor promoter (Slaga, 1989). The initiation-promotion protocol induces benign exophitic papillomas, some of which convert to SCC. The papillomas are believed to be clonal expansions of the initiated cells (Burns et al., 1978). It has been postulated that the malignant conversion of the benign papillomas occurs as a result of somatic mutations in the expanded population of initiated cells (Moolgavkar and Knudson, 1981; Potter, 1981). Initiation of tumorigenesis by the carcinogen 7,12-dimethylbenz[a]anthracene (DMBA) induces a speci®c point mutation at codon 61 of the Ha-ras-1 gene (Balmain and Pragnell, 1983; Bizub et al., 1986; Roop et al., 1986). Histopathological and cytogenetic studies demonstrated that relevant changes occur during papilloma progression. At 10 weeks of promotion, most papillomas are well di€erentiated and have no atypical cells. Important changes occur at

Cell-cycle regulation in mouse skin tumors ML Rodriguez-Puebla et al

2252

20 weeks, when most of the papillomas became moderately dysplastic and atypical cells appear in the basal and suprabasal layers, and at 30 ± 40 weeks, the papillomas are severely dysplastic (Aldaz et al., 1987). Concomitant with the histologic changes, papilloma progression results in chromosomal numerical alterations with non-random trisomization in chromosomes 6 and 7 (Aldaz et al., 1989; Bianchi et al., 1990). Although alteration in cell-cycle regulatory machinery in this model was previously reported (Zhang et al., 1997), this is the ®rst sequential study of the alterations in cell-cycle protein expression during the premalignant progression of epidermal lesions. In this study, we used the two-stage carcinogenesis model to determine the expression of cell-cycle proteins during promotion and papilloma progression in SENCAR mice. The levels of G1 cyclins, CDKs, CKIs, E2F transcription factors, pRb, and pRb-related proteins (p107 and p130) were analysed in papillomas during di€erent stages of tumor development, including early and late papillomas, SCC, normal skin (NS), and hyperplastic skin (HS). Increases in the levels of negative and positive regulators of the cell cycle were observed. Moreover, complex formation between D-type cyclins, CDKs, and CKIs was also induced, indicating that CDKs were activated at this time.

1). In fact, the levels of cyclins increase at 12 ± 15 h after TPA treatment in dorsal SENCAR mouse epidermis (Rodriguez-Puebla et al., 1998). Hence, in this particular model, we can consider the cyclin E and A levels to re¯ect cell proliferation but not changes associated with tumor development. These results showed that two D-type cyclins were overexpressed during papilloma progression and that cyclin D2 levels decreased signi®cantly during malignant conversion to SCC. Early studies demonstrated that cyclin D1 mRNA and protein are overexpressed in the proliferative compartment of papillomas, including small incipient papillomas. Immunohistochemical analysis showed that NS and HS are negative for cyclin D1 (Bianchi et al., 1993; Robles and Conti, 1995). Our studies also indicated that cyclin D2 might have an important role during mouse skin promotion. Expression of CKIs CKIs are proteins that bind and inhibit cyclin/CDK complexes. p16Ink4a belongs to a family of CKIs that speci®cally associate with CDK4 and CDK6 and inhibit their activity (Serrano et al., 1993; Xiong, 1996). p21Cip1, p27Kip1, and p57Kip2 are members of

Results Expression of cyclin and CDK proteins To study the expression of cell cycle proteins in the epidermal lesions of SENCAR mice, proteins were extracted from papillomas after 10, 20, 30 and 40 weeks of promotion as well as from NS (non-treated normal skin), HS (TPA treated skin), and SCC and used for Western blot analysis. For each time point, four animals were analysed and three papillomas per mouse were combined for protein extraction. SCC, NS, and HS were analysed individually. To determine the size of the proliferative compartment in non-neoplastic skin and tumors, PCNA protein expression was also analysed by the Western blot technique (van et al., 1991). PCNA has been shown to be expressed in all proliferative cells regardless of the stage of the cell cycle. The PCNA protein levels were similar in papillomas at di€erent stages of promotion and in SCC, NS, and HS (Figure 1). Our results support those of Fukuda et al. (1978), who found little di€erence in the proliferative compartments of papillomas and SCCs. Thus, the level of expression of PCNA in each pool was used to normalize the measurement of all cell cycle proteins. Relevant changes in the level of protein expression were detected at 20 weeks of promotion in cyclin D1 and cyclin D2, whereas the cyclin D3 levels remained the same. The cyclin D1 and cyclin D2 levels increased nine- and 2.5-fold, respectively, at 20 weeks, and the cyclin D2 levels decreased fourfold during malignant conversion to SCC. On the other hand, no signi®cant changes were observed between papillomas and carcinomas in levels of the three CDKs (CDK2, CDK4, and CDK6), cyclin E, and cyclin A. However, four- and 70-fold increases over the NS levels were observed in cyclins E and A in HS (Figure

Figure 1 Western blot analysis of cyclin and CDK expression during SENCAR mouse skin promotion. Protein lysates of NS and HS, papillomas at 10, 20, 30, and 40 weeks of promotion, and SCC were separated by SDS ± PAGE and blotted to a nitrocellulose membrane. Primary antibodies against cyclin D1, cyclin D2, cyclin D3, cyclin A, cyclin E, cdk2, cdk4, cdk6, and PCNA were used in the immunoblotting analysis (a). The levels of PCNA and cyclin D1, D2, and D3 were quanti®ed with a densitometer. The relative abundance of each cyclin at each stage was determined from the absorbance of each band and normalized against the PCNA expression in the same sample. Statistical analysis showed signi®cant di€erences in cyclin D1 (P50.0001) and cyclin D2 expression (P50.0001) during promotion and between papillomas at 10 and 20 weeks (cyclin D1 P40.0003; cyclin D2 P40.002) (b)

Cell-cycle regulation in mouse skin tumors ML Rodriguez-Puebla et al

another family of CKIs that interact with and inhibit a wide variety of cyclin/CDK holoenzymes. Overexpression of these CKIs blocks cells in G1. We studied the pattern of expression of these proteins during mouse skin promotion by Western blot analysis. p21Cip1 and p27Kip1 protein levels did not change neither between NS and HS nor during tumor development (Figure 2). In contrast, p16Ink4a and p57Kip2 protein levels increased four- and threefold, respectively, at 20 weeks of promotion (Figure 2). These results showed that the levels of p16Ink4a and p57Kip2, two inhibitors that belong to di€erent families of CKIs, were elevated at the same stage during mouse skin papilloma development. It is noteworthy to mention that p16Ink4a is the product of a tumor suppressor gene that is mutated in several varieties of cancers (Hirama and Koe‚er, 1995), and the p57Kip2 gene is in a region of human chromosome 11 implicated in Beckwith-Wiedermann syndrome (Taniguchi et al., 1997). On the other hand, p21Cip1 and p27Kip1 have not yet been implicated in tumorigenesis. Formation of cyclin/CDK/CKI complexes We also examined the interaction of D-type cyclins with CDKs and CKIs during papilloma promotion. The pools from each time point were combined so that each sample represents papillomas of four mice at each stage of tumor development, NS, HS, and SCC. These samples were immunoprecipitated with antibodies against CDK4 and CDK6 and then analysed by Western blotting with cyclin D1, cyclin D2, cyclin

Figure 2 Western blot analysis of CKI expression during SENCAR mouse skin promotion. Protein lysates of NS and HS, papillomas at 10, 20, 30, and 40 weeks of promotion; and SCC were separated by SDS ± PAGE and blotted to a nitrocellulose membrane. Polyclonal anti-p21Cip1, p27Kip1, p57Kip2 and p16Ink4a antibodies were used for immunoblotting analysis (a). The relative abundances of the proteins were determined by densitometric analysis of the blots and normalized against the PCNA expression of the same sample. Statistical analysis showed signi®cant di€erences in p16Ink4a (P50.0001) and p57Kip2 (P50.003) protein levels during promotion and between papillomas of 10 and 20 weeks (p16Ink4a P40.002; p57Kip2 P40.02) (b)

D3, p21Cip1, p27Kip1 and p16Ink4a antibodies (Figure 3). Cyclin D1/CDK4 complexes were mostly induced after 20 weeks of promotion, although we also observed complexes in HS and early papillomas at 10 weeks of promotion. Complexes between CDK6 and cyclin D1 were not detected (Figure 3). Cyclin D3/CDK4 and cyclin D3/CDK6 complexes were mostly observed in HS and decreased during tumor development. We did not detect complexes with cyclin D2. Next we analysed the interaction of CKI proteins with CDK4 and CDK6. p16Ink4a formed complexes with both CDKs, whereas both p21Cip1 and p27Kip1 formed complexes only with CDK4. The levels of CDK4/ p27Kip1 complexes were similar in NS, HS, and papillomas at 10 weeks of promotion and then notably increased after 20 weeks of papilloma promotion (Figure 3). The CDK4/p21Cip1, CDK4/ p16Ink4a, and CDK6/p16Ink4a levels were very low in NS, HS, and early papillomas and then suddenly increased after 20 weeks of promotion (Figure 3). CDK4/cyclin D1 appeared to be the most abundant complex induced during promotion. To analyse in more detail the CDK4/cyclin D1 complex, we immunoprecipitated cyclin D1 and analysed the various complexes in which it was present by Western blotting (Figure 4). Again, the amounts of CDK4/ cyclin D1 complexes increased after 20 weeks and CDK6/cyclin D1 complexes were almost undetectable. The binding patterns of p21Cip1 and p27Kip1 were the same as previously observed (Figures 3 and 4) whereas

Figure 3 Coimmunoprecipitation and immunoblot analysis of cyclin D/CDK/CKI complexes formed during SENCAR mouse skin promotion. Fresh protein lysates of NS and HS, papillomas at 10, 20, 30, and 40 weeks of promotion; and SCC were immunoprecipitated with polyclonal anti-cdk4 (a) and anti-cdk6 (b) antibodies. The proteins were separated by SDS ± PAGE and blotted to a nitrocellulose membrane. Primaries antibodies against cyclin D1, cyclin D3, p21Cip1, p27Kip1, p16Ink4a, CDK4 and CDK6 were used for immunoblotting analysis. The negative control was immunoprecipitated with normal rabbit serum and lysates of papillomas at 40 weeks of promotion (NR). The protein lysate of a papilloma of 40 weeks of promotion was used as positive control (PL)

2253

Cell-cycle regulation in mouse skin tumors ML Rodriguez-Puebla et al

2254

p16Ink4a protein was not detected in these complexes. CDK4/cyclin D1 complex formation was induced after cyclin D1 protein expression increased at 20 weeks of promotion (Figures 1 and 4). These results showed that in later papillomas (after 20 weeks of promotion) and SCCs, there were two types of complexes with CDK4, cyclin D1/CDK4 and p16Ink4a/CDK4. Also, p16Ink4a protein expression was induced in papillomas at 20 weeks of promotion (Figure 2), so p16Ink4a/CDK4 complex formation may have been induced at this time. To study this, we immunoprecipitated the same samples with p16Ink4a antibody and analysed the precipitates by Western blotting for CDK4 and cyclin D1. In support of our hypothesis, p16Ink4a was complexed with CDK4 but not with cyclin D1 after 20 weeks of promotion (Figure 4). Expression of Rb family proteins and E2F transcription factors CDK/cyclin D complexes are known to phosphorylate pRb in vitro, and two known pRb-related proteins, p107 and p130, are also putative substrates of these CDK/cyclin complexes. After phosphorylation of pRb and pRb-related proteins, E2F proteins are released and activate the transcription of other genes involved in S phase progression. The interaction of pRb with E2F correlates with the capacity of pRb to arrest cell growth in G1 phase (Hiebert, 1993; Hiebert et al., 1992; Qian et al., 1992; Qin et al., 1992). Thus, we decided to study the expression of pRb, pRb-related proteins, (p107 and p130), and the E2F family in this

Figure 4 Coimmunoprecipitation and immunoblot analysis of cyclin D1/CDKs and p16Ink4a/CDKs complexes formed during SENCAR mouse skin promotion. Fresh protein lysates of NS and HS; papillomas at 10, 20, 30, and 40 weeks of promotion; and SCC were immunoprecipitated with polyclonal anti-cyclin D1 (a) and anti-p16Ink4a (b) antibodies. The proteins were separated by SDS ± PAGE and blotted to a nitrocellulose membrane. Primaries antibodies against p21Cip1, p27Kip1, p16Ink4a, cyclin D1, CDK4, and CDK6 were used for immunoblotting analysis. The negative immunoprecipitation control was normal rabbit serum and lysates of papillomas at 40 weeks of promotion (NR). Protein lysate of a papilloma at 40 weeks of promotion was used as a positive control (PL)

model. p107 and p130 protein levels were analysed by Western blotting. In the case of pRb, samples were immunoprecipitated with an anti-Rb polyclonal antibody, and the immunoprecipitates were analysed by Western blotting with the same antibody. Figure 5 shows that pRb expression varied considerably in samples at the same stages of promotion, all samples had two bands with di€erent mobilities consistent with the hypophosphorylated and hyperphosphorylated forms of pRb. Unlike the other cell-cycle proteins, pRb shows a great variability in the levels of expression between samples of di€erent mice. Moreover, we did not observe pRb induction during promotion, as we did for cyclin D1, cyclin D2, p16Ink4a, and p57Kip2 (Figures 1 and 2). The p107 and p130 protein levels decreased somewhat at 20 and 30 weeks of promotion, the decrease in p130 being more notable (Figure 5). E2F transcription factors have been implicated as critical targets for the tumor suppressor pRb and pRbrelated proteins (Bagchi et al., 1991; Bandara et al., 1991; Chellappan et al., 1991). Five members of the E2F family have been described (E2F-1 to E2F-5) but only E2F-1 has been associated with tumorigenesis (Johnson et al., 1994; Singh et al., 1994; Yamasaki et al., 1996). Di€erent members of the E2F family appear

Figure 5 Western blot analysis of pRb and E2F family expression during SENCAR mouse skin promotion. Protein lysates of NS and HS; papillomas at 10, 20, 30, and 40 weeks of promotion; and SCC were separated by SDS ± PAGE and blotted to a nitrocellulose membrane. Primary antibodies against pRb, p107, p130, E2F-1, E2F-2, E2F-4, and E2F-5 were used for immunoblotting analysis (a). The levels of E2F-1, E2F-2, E2F-4, and E2F-5 were quanti®ed with a densitometer. The relative abundance of each protein at each stage was determined from the absorbance of each band and normalized against PCNA expression in the same sample. Statistical analysis showed signi®cant di€erences in the four members of the E2F family (E2F-1 P50.0001; E2F-2 P50.03; E2F-4 P50.0001, and E2F-5 P50.0001) during promotion and between papillomas of 10 and 20 weeks of promotion (E2F-1 P50.0001; E2F-2 P40.03, E2F-4 P40.001, and E2F-5 P40.001) (b)

Cell-cycle regulation in mouse skin tumors ML Rodriguez-Puebla et al

to play distinct roles by inducing the expression of di€erent genes that participate in G1/S phase transition (De Gregori et al., 1997; Zhang et al., 1997). We analysed the protein levels of E2F-1, E2F-2, E2F-4 and E2F-5 during mouse skin promotion by Western blotting. Expression was strongly induced at 20 weeks of promotion, increasing four-, three-, six- and 12-fold, respectively (Figure 5). No other signi®cant changes were observed during promotion, including the step of acquisition of the malignant phenotype to SCC. The role of E2F transcription factors as positive regulators of the cell cycle suggests that deregulation in their expression may have a role in the deregulated cell proliferation that occurs during tumorigenesis. Discussion The results presented here advance the study of cellcycle proteins that may participate in the expansion of initiated cell populations during premalignant progression. This study also supports previous reports of deregulated expression and ampli®cation of cyclin D1 in chemically induced skin tumors (Bianchi et al., 1993; Mitsunaga et al., 1995; Robles and Conti, 1995), human tumors (Bartkova et al., 1995a,b; Wang et al., 1996; Weinstat-Saslow et al., 1995) and in rat mammary tumors (Sgambato et al., 1995). Overexpression of cyclin D1 in the mammary glands of transgenic mice leads to hyperplasia and tumor formation (Wang et al., 1994). Also, our previous results have shown that overexpression of cyclin D1 under the control of the keratin 5 promotor leads to epidermal hyperproliferation and thymic hyperplasia (Robles et al., 1996). We determined that other positive and negative regulators of cell cycle are induced at speci®c times during tumor progression, and these alterations may be true changes associated with premalignant progression. Overexpression of cyclin D1, cyclin D2, and E2F could explain why cell proliferation becomes independent of external factors in late papillomas. In fact, cyclin D1, CDK4, and pRb have a functional interaction in which phosphorylation of pRb by the cyclin D/CDK complex releases E2F, which activates others genes involved in the G1/S phase transition (Chellappan et al., 1991; De Gregori et al., 1997; Dyson et al., 1993; Sherr, 1995). Overexpression of cyclin D1 and probably cyclin D2 may participate in this interaction, although we found no relevant di€erences in the pattern of pRb phosphorylation. However, the induction of E2F protein expression may overcome the necessity of pRb phosphorylation for cell-cycle progression. Thus, the increase in E2F protein may be the principal event responsible for cell-cycle progression and the promoterindependent stage. This hypothesis is supported by the results of Lukas et al. (1996), who showed that deregulated expression of E2F family members induces S-phase entry and overcomes p16Ink4a-mediated growth suppression. The expression levels of two pRb-related proteins, p130 and p107, did not change signi®cantly during papilloma promotion, although a slight decrease was observed at 20 weeks of promotion. Thus, the decreased levels of p107 and p130 may be involved in the overexpression of E2F-1 protein. This result is consistent with the role of these proteins in quiescent

cells, in which they form inhibitory complexes that downregulate E2F-1 gene transcription (Cobrinik et al., 1993; Johnson, 1995). Also, a recent report showed that p107 and p130 double negative cells strongly derepressed E2F-1 gene expression in mouse embryonic ®broblasts (Hurford et al., 1997). Another unexpected result was the overexpression of p16Ink4a and p57Kip2, two negative regulators of the cell cycle. Several CKls have been implicated in tumorigenesis in which p16Ink4a deletion was detected in mouse skin tumors (Linardopoulos et al., 1995) and human tumors (Jadayel et al., 1997; Lilischkis et al., 1996; Musgrove et al., 1995; Schmit et al., 1994). However, the overexpression of p16Ink4a is consistent with a model in which the loss of pRb function results in the loss of a negative feedback loop. pRb has been shown to repress p16Ink4a expression, and loss of functional pRb results in p16Ink4a overexpression (Parry et al., 1995). Also, in primary lung cancers and cell lines, cyclin D1 overexpression and pRb inactivation result in p16Ink4a expression (Shapiro et al., 1995). Therefore, the overexpression of cyclin D1 and cyclin D2 can be responsible for pRb inactivation and overexpression of p16Ink4a. We observed that p16Ink4a displaced cyclin D1 from CDK4 and possibly from CDK6. No cyclin D/CDK complexes have been observed in cells lacking functional pRb and overexpressing p16Ink4a (Bates et al., 1994; Parry et al., 1995). Therefore, p16Ink4a protein expression and p16Ink4a/CDK complex formation are probably the consequence of cyclin D overexpression and pRb inactivation. It is noteworthy that only four of 20 studied tumors at 20 or more weeks of promotion and SCC had a pRb protein level comparable to that of HS, which increase the protein level and phosphorylation 12 ± 15 h after TPA treatment (Rodriguez-Puebla et al., 1998). While p21Cip1 and p27Kip1 protein levels remain constant throughout premalignant progression, both of these CKIs appear to bind to cyclin D1/CDK4 beginning at 20 weeks of promotion when cyclin D1 protein level increases. Thus, p21Cip1 and p27Kip1 protein levels may a€ect the kinase activity of this complex. Thus, our results could explain the reason that keratinocytes derived from p21Cip1 knockout mice, transformed with ras oncogene, exhibit a very aggressive behavior in nude mice (Missero et al., 1996), although opposed results were recently reported in a similar model (Weinberg et al., 1997). Our ®nding of CDK/cyclin complex formation during premalignant progression showed that cyclin D1 and cyclin D3 are not only expressed di€erently but also are probably activated at di€erent times. The relative abundance of cyclin D1 protein appears to regulate the formation of cyclin D1/CDK4 complexes at 20 weeks of promotion. Cyclin D3/CDK4 and cyclin D3/CDK6 complexes were strongly induced in HS, and low levels were detected in tumors. It is noteworthy that the levels of cyclin D3, CDK4, and CDK6 remained constant in NS, HS, and papillomas but that cyclin D3/CDK complex formation was mostly detected in HS. This suggests that events other than the simple abundance of these proteins regulate the formation of these complexes. Similarly, there is no correlation between protein steady-state level and cyclin D3/CDK complex formation after TPA treat-

2255

Cell-cycle regulation in mouse skin tumors ML Rodriguez-Puebla et al

2256

ment of mouse dorsal skin (Rodriguez-Puebla et al., 1998). Our results revealed two stages of papilloma development at which the levels of some cell-cycle proteins increased. In HS, cyclin E, and cyclin A protein expression and cyclin D3/CDK complex formation were induced, but in tumors there was no further increases. This result is consistent with the roles of those proteins in cell proliferation but not in the neoplastic process. The second time of change was during premalignant progression; the levels of positive and negative regulators of the cell cycle increased from 10 to 20 weeks of promotion. This is the time when these tumors become dysplastic and develop chromosomal numerical abnormalities (Aldaz et al., 1987). We speculate that the changes observed in the expression of cell-cycle proteins are connected with the histopathologic changes and chromosomal alterations observed in this model (Aldaz et al., 1987, 1989; Bianchi et al., 1990). In this report, we showed that deregulated growth control occurred during premalignant progression, but no other changes were observed in malignant SCCs in comparison with late papillomas. Our observations suggest that early events in carcinogenesis activate speci®c signaling pathways (e.g., the ras pathway after the carcinogen DMBA induces a speci®c point mutation at codon 61) and deregulation of cell-cycle machinery promotes cell growth that becomes independent of growth factors. Subsequent mutations could induce invasive behavior. It is therefore of considerable interest that adhesion receptors have been shown to function as tumor suppressors (Frixen et al., 1991; Gianconi & Ruoslani., 1990; Tennenbaum et al., 1992). Also, in the mouse carcinogenesis skin model, proteases such as transin and urokinases that induce malignant conversion are secreted in fos transformed cells (Yuspa et al., 1990). This model therefore proposes that deregulation of the cell cycle provides growth advantage as well as genomic instability, and then another mutation(s) leads to invasive behavior. In the absence of the cell-cycle deregulation, the later mutation(s) would be disadvantageous or perhaps even lethal to the a‚icted cell. This notion is consistent with a model in which mutations do not occur in a strict order but changes in certain genes generally occur early in tumorigenesis. Materials and methods Experimental animals Young adult female SENCAR mice were obtained from the National Cancer Institute (Frederick, MD, USA). The mice's dorsal sides were shaved, and the mice were allowed to rest for 2 days. Four animals were topically treated with TPA (Sigma Chemical Co., St. Louis, MO, USA) (1 mg in 0.2 ml of acetone) twice a week for 10 weeks, and then dorsal skins were used as HS, whereas the dorsal skins of untreated mice were used as NS. Initiation was performed with DMBA (Sigma Chemical Co.) (10 nmol in 0.2 ml of acetone) which was topically applied to each mouse's shaved dorsal skin. Two weeks after DMBA initiation, TPA (1 mg in 0.2 ml of acetone) was applied twice a week on each mouse's dorsal skin. Papillomas appeared after 8 weeks of continuous TPA treatment. The mice were killed after 10, 20, 30, and 40 weeks of TPA treatment, and papillomas and SCCs were immediately frozen

at 7708C. The SCCs appeared at approximately 35 ± 40 weeks of promotion. Three papillomas of each stage from each animal were combined for protein extraction, so each papilloma pool represents one stage of one animal. SCC, NS, and HS samples were also obtained from each animal. Mouse skin and tumors were ground with a mortar in liquid nitrogen and homogenized in homogenization bu€er (60 mM Tris-HCl, pH 8.6; 5 mM EDTA, 5 mM EGTA, 300 mM sucrose; 5 mM dithiothreitol; 2 mM phenylmethylsulfonyl ¯uoride; 10 mM sodium molibdate; 20 mg/ml aprotinin; 20 mM NaF; and 100 mM sodium orthovanadate) with a Polytron PT10 homogenizer (three 15-s bursts at setting 6 on ice). The homogenates were sonicated with a Branson soni®er 450 (three 8-s bursts at speed 2, duty cycle 30% on ice) and centrifuged at 2800 g for 10 min. The supernatants were collected and used directly for western blot analysis or stored at 7708C. Western blot analysis and immunoprecipitation assay The protein concentration in each lysate was measured with the Bio-Rad protein assay system (Bio-Rad Laboratories, Richmond, CA, USA). Protein lysates (25 mg from each sample) were electrophoresed through 7, 8, 10, or 12% acrylamide gels and electrophoretically transferred onto nitrocellulose membranes. After being blocked with 5% nonfat powdered milk in Dulbecco's phosphatebu€ered saline (Sigma Chemical Co.), the membranes were incubated with 1 mg/ml speci®c antibodies. The following antibodies were used: polyclonal antibodies against PCNA (PC10), cyclin D1 (C-20), cyclin D2 (M20), cyclin D3 (C-16), cyclin E (M-20), cyclin A (C-19), CDK4 (C-22), CDK6 (C-21), CDK2 (M-2), p21 (C-19), p27 (C-19), p57 (M-20), p16 (M156), E2F-1 (C-20), E2F-2 (L20), pRb (C-15), p107 (C-18), and p130 (C-20) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) and monoclonal antibodies against E2F-4 (RK13), E2F-5 (MH5) (Santa Cruz Biotechnology, Inc.). Horseradish peroxidase-conjugated secondary antibody (Amersham Corp., Arlington Heights, IL, USA), and enhanced chemiluminescense (ECL Detection Kit) (Amersham Corp.) were used for immunoblotting detection. To study cyclin D/CDK/CKI complex formation, four pools of samples of each stage of promotion and SCC, NS, and HS were combined and used for immunoprecipitation. We used polyclonal anti-p16, anti-cyclin D1, anti-CDK4, and anti-CDK6 antibodies conjugated with protein G-Sepharose beads (Life Technologies Inc., Grand Island, NY, USA) to immunoprecipitate protein lysates overnight at 48C with constant rotation. After washing, Western blot analysis was performed as described above with monoclonal antibodies against cyclin D1 (C-20), cyclin D3 (C-16), p16 (M156), p21 (C-19), p27 (C-19), CDK4 (C-22), and CDK6 (C-21) (Santa Cruz Biotechnology, Inc.). Bio-image analysis was used to quantitate the expression levels of the proteins. Abbreviations CDK, cyclin-dependent kinase; CKI, cdk inhibitor; DMBA, 7,12-dimethylbenz[a]anthracene; pRb, retinoblastoma protein; NS, normal skin; HS, hyperplastic skin; SCC, squamous cell carcinoma; TPA, 12-o-tetradecanoylphorbol-13-acetate; PCNA, proliferating cell nuclear antigen. Acknowledgements We thank the Science Park histology service for assistance with tumor preparations, the Science Park animal facility personnel; Melissa Bracher for secretarial assistance, Sharon L Stockman for preparation of the manuscript, and Maureen Goode for editing the paper.

Cell-cycle regulation in mouse skin tumors ML Rodriguez-Puebla et al

This work was supported by DHS grants CA 42157 and CA 57596, and American Cancer Society (ACS) CN152 to DGJ and by institutional support grants CA 16672 to M.D.

Anderson Cancer Center for the animal facility and Center grant P30-E507784-01 for the histology service.

References Aldaz CM, Conti CJ, Klein-Szanto AJP and Slaga T. (1987). Proc. Natl. Acad. Sci. USA, 84, 2029 ± 2032. Aldaz MC, Trono D, Larcher F, Slaga T and Conti CJ. (1989). Molecular Carcinogenesis, 2, 22 ± 26. Bagchi S, Weinmann R and Raychaudhuri P. (1991). Cell, 65, 1063 ± 1072. Balmain A and Pragnell I. (1983). Nature, 303, 72 ± 74. Bandara L, Adamczewski J, Hunt T and La TN. (1991). Nature, 352, 249 ± 251. Bartkova J, Lukas J, Muller H, Strauss M, Gusterson B and Bartek J. (1995a). Cancer Res, 55, 949 ± 956. Bartkova J, Lukas J, Strauss M and Bartek J. (1995b). Oncogene, 10, 775 ± 778. Bates S, Parry D, Bonetta L, Vousden K, Dickson C and Peters G. (1994). Oncogene, 9, 1633 ± 1640. Bianchi A, Aldaz MC and Conti CJ. (1990). Proc. Natl. Acad. Sci. USA, 87, 6902 ± 6906. Bianchi AB, Fischer SM, Robles AI, Rinchik EM and Conti CJ. (1993). Oncogene, 8, 1127 ± 1133. Bianchi AB, Navone NM, Aldaz CM and Conti CJ. (1991). Proc. Natl. Acad. Sci. USA, 88, 7590 ± 7594. Bizub D, Wood A and Skalka A. (1986). Proc. Natl. Acad. Sci. USA, 83, 6048 ± 6052. Burns FJ, Vanderlaan M, Snyder E and Albert RE. (1978). Carcinogenesis, Vol. 2. In: Slaga, TJ, Sivak A and Boutwell RR (eds). Raven: New York, pp. 91 ± 96. Chellappan, S, Hiebert S, Mudryj M, Horowitz J and Nevins J. (1991). Cell, 65, 1053 ± 1061. Cobrinik D, Whyte P, Peeper D, Jacks T and Weinberg R. (1993). Genes Dev, 7, 2392 ± 2404. De Gregori J, Leone G, Miron A, Jakoi L and Nevins J. (1997). Proc. Natl. Acad. Sci. USA, 94, 7245 ± 7250. Demers GW, Foster SA, Halbert CL and Calloway DA. (1994). Proc. Natl. Acad. Sci. USA, 91, 4382 ± 4386. Dyson N, Dembski M, Fattaey A, Ngwu C, Ewen M and Helin K. (1993). J. Virol., 67, 7641 ± 7647. Frixen UH, Beherens J, Schas M, Eberle G, Voss B, Wanda A, Lochmer D and Birchmeier W. (1991). J. Cell Biol., 112, 173 ± 185. Fukuda M, Okamura K, Rohrbrach R, Bohm, N and Fujita S. (1978). Cell Tiss. Kinet., 11, 611 ± 621. Gianconi FG and Ruoslani E. (1990). Cell, 60, 849 ± 859. Guan KL, Jenkins CW, Li, Y, Nichols MA, Wu X, O'Keefe C, Matera AG and Xiong Y. (1994). Genes Dev., 8, 2939 ± 2952. Hannon GJ and Beanch D. (1994). Nature, 371, 257 ± 261. Hatada I and Mukai T. (1995). Nat. Genet., 11, 204 ± 206. Hiebert S. (1993). Mol. Cell. Biol., 13, 3384 ± 3391. Hiebert S, Chellappan S, Horowitz J and Nevins J. (1992). Genes Dev., 6, 177 ± 185. Hirai H, Roussel MF, Kato JY, Ashmun RA and Sherr CJ. (1995). Mol. Cell. Biol., 15, 2672 ± 2681. Hirama T. and Koe‚er HP. (1995). Blood, 86, 841 ± 854. Hurford R, Cobrinik D, Lee M-H and Dysion N. (1997). Genes Dev., 11, 1447 ± 1463. Jadayel D, Lukas J, Nacheva E, Bartkova J, Stranks G, De SP, Lens D, Bartek J, Dyer M, Kruger A and Catovsky D. (1997). Leukemia, 11, 64 ± 72. Jiang W, Zhang Y, Kahn S, Hollstein M, Santella R, Lu S, Harris C, Montesano R and Weinstein I. (1993). Proc. Natl. Acad. Sci. USA, 90, 9026 ± 9030. Johnson D. (1995). Oncogene, 11, 1685 ± 1692. Johnson DG, Cress WD, Jakoi L and Nevins JR. (1994). Proc. Natl. Acad. Sci. USA, 91, 12823 ± 12827.

Koike M, Ishino K, Huh N and Kuroki T. (1994). Biochem. Biophys. Res. Commun., 201, 673 ± 681. Lammie G, Fantl V, Smith R, Schuuring E, Brookes S, Michalides R, Dickson C, Arnold A and Peters G. (1991). Oncogene, 6, 439 ± 444. Lilischkis R, Sarcevic B, Kennedy C, Warlters A and Sutherland R. (1996). Int. J. Cancer, 66, 249 ± 254. Linardopoulos S, Street A, Quelle D, Parry D, Peters G, Sherr C and Balmain A. (1995). Cancer Res., 55, 5168 ± 5172. Lukas J, Petersen B, Holm K, Bartek J and Helin K. (1996). Mol. Cell. Biol., 16, 1047 ± 1057. Missero C, Di Cunto F, Kiyokawa H, Ko€ A and Dotto GP. (1996). Genes Dev., 10, 3065 ± 3075. Mitsunaga S, Zhang S and Rugger B. (1995). Carcinogenesis, 16, 1629 ± 1635. Moolgavkar S and Knudson AJ. (1981). J. Natl. Cancer Inst., 66, 1037 ± 1052. Mori T, Miura K, Aoki T, Nishihira T, Mori S and Nakamura Y. (1994). Cancer Res., 54, 3396 ± 3397. Musgrove E, Lilischkis R, Cornish A, Lee C, Setlur V, Seshadri R and Sutherland R. (1995). Int. J. Cancer, 63, 584 ± 591. Parry D, Bates S, Mann DJ and Peters G. (1995). EMBO. J., 14, 503 ± 511. Potter V. (1981). Carcinogenesis, 2, 1375 ± 1379. Qian Y, Luckey C, Horton L, Esser M and Templeton D. (1992). Mol. Cell. Biol., 12, 5363 ± 5372. Qin X, Chittenden T, Livingston D and Kaelin WJ. (1992). Genes Dev., 6, 953 ± 964. Robles AI and Conti CJ. (1995). Carcinogenesis, 16, 781 ± 786. Robles AI, Larcher F, Whalin RB, Murillas R, Richie E, Gimenez-Conti IB, Jorcano JL and Conti CJ. (1996). Proc. Natl. Acad. Sci. USA, 93, 7634 ± 7638. Rodriguez-Puebla ML, Robles AI, Johnson DG, LaCava M and Conti CJ. (1998). Cell Growth Di€er., 9, 31 ± 39. Roop D, Lowy D, Tambourin P, Strickland J, Harper J, Balaschak M, Spangler E and Yuspa S. (1986). Nature, 323, 822 ± 824. Schmit E, Ichimura K and Reifenberger G. (1994). Cancer Res., 54, 6321 ± 6324. Serrano M, Hannon GJ and Beach D. (1993). Nature, 366, 704 ± 707. Sgambato A, Han E, Zhang Y, Moon R, Santella R and Weinstein I. (1995). Carcinogenesis, 16, 2193 ± 2198. Shapiro GI, Edwards CD, Kobzik L, Godleski J, Richards W, Sugarbaker DJ and Rollins BJ. (1995). Cancer Res., 55, 505 ± 509. Sherr CJ. (1993). Cell, 73, 1059 ± 1065. Sherr CJ. (1995). Trends Biochem. Sci., 20, 187 ± 190. Singh P, Wong S and Hong W. (1994). EMBO J., 13, 3329 ± 3338. Slaga TJ. (1989). In: Carcinogenesis. A Comprehensive Survey, Vol. 11. Contin CJ, Slaga TJ and Klein-Szanto AJ (eds). Raven Press: New York, pp. 1 ± 18. Taniguchi T, Okamoto K and Reeve A. (1997). Oncogene, 14, 1201 ± 1206. Tennenbaum T, Yuspa S, Grover A, Castronovo V, Sobel M, Yamada Y and DeLuca L. (1992). Cancer Res., 52, 2966 ± 2976. van DJ, Wijsman J, Keijzer R, van dVC and Cornelisse C. (1991). Am. J. Pathol., 138, 1165 ± 1172.

2257

Cell-cycle regulation in mouse skin tumors ML Rodriguez-Puebla et al

2258

Wang A, Yoshimi N, Suzui M, Yamauchi A, Tarao M and Mori H. (1996). J. Cancer Res. Clin. Oncol., 122, 122 ± 126. Wang T, Cardi€ R, Zukerberg L, Lees E, Arnold A and Schmidt E.(1994). Nature, 369, 669 ± 671. Weinberg R. (1991). Science, 254, 1138 ± 1146. Weinberg WC, Montano NE and Deng C. (1997). Oncogene, 15, 685 ± 690. Weinstat-Saslow D, Merino M, Manrow R, Lawrence J, Bluth R, Wittenbel K, Simpson J, Page D and Steeg P. (1995). Nat. Med., 1, 1257 ± 1260. Xiong Y. (1996). Biochim. Biophys. Acta, 1288, 1 ± 5.

Yamasaki L, Jacks T, Bronson R, Goillot E, Harlow E and Dyson N. (1996). Cell, 85, 537 ± 548. Yuspa S, Hennings H, Roop D, Strickland J and Greenhalgh D. (1990). Environ. Health Perspect., 88, 193 ± 195. Zhang S, Klein-Szanto A, Sauter E, Shafarenko M, Mitsunaga S, Nobori T, Carson D, Ridge J and Goodrow T. (1994). Cancer Res., 54, 5050 ± 5053. Zhang S-Y, Liu S-C, Goodrow T, Morris R and KleinSzanto AJP. (1997). Molecular Carcinogenesis, 18, 142 ± 152.