Oncogene (2000) 19, 4746 ± 4753 ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc Reversal of an antiestrogen-mediated cell cycle arrest of MCF-7 cells by viral tumor antigens requires the retinoblastoma protein-binding domain

Hemant Varma1,3 and Susan E Conrad*,2,3

1Department of Biochemistry, Michigan State University, East Lansing, Michigan, MI 48824-1101, USA; 2Department of Microbiology, Michigan State University, East Lansing, Michigan, MI 48824-1101, USA; 3Cell and Molecular Biology Graduate Program, Michigan State University, East Lansing, Michigan, MI 48824-1101, USA

Proliferation of MCF-7 cells is estrogen dependent and that may lead to improvements in breast cancer antiestrogen sensitive. In the absence of estrogens or diagnosis and treatment. presence of antiestrogens MCF-7 cells arrest in the G1 Transit through G1 into S phase is regulated by phase of the cell cycle, and this arrest is associated with cyclin dependent kinases (CDKs), a family of protein an accumulation of the active, hypophosphorylated form kinases that are activated by association with cyclins. of the retinoblastoma protein (pRb). Because active pRb The CDKs that regulate passage through G1 include negatively regulates passage from G1 to S phase, this CDK4 and CDK6, which are activated by the D-type suggests that pRb is a crucial target of estrogen action, cyclins, and CDK2, which is activated by both cyclin E and that its inactivation might lead to antiestrogen and cyclin A (Sherr, 1993, 1994). One important target resistance. We tested this hypothesis by expressing viral of G1 cyclin/CDK complexes is the tumor suppressor tumor antigens (T antigens), which bind and inactivate protein pRb (Kato et al., 1993). In its hypopho- pRb, in MCF-7 cells, and determining the e€ects on cell sphorylated form, pRb associates with members of the proliferation in the presence of antiestrogens. The results E2F family of transcription factors, and prevents of these experiments demonstrate that T antigen transcription of genes required for entry into S phase. expression confers antiestrogen resistance to MCF-7 Upon phosphorylation of pRb, E2F transcription cells. Using a panel of mutant T antigens, we further factors are released and activate target genes that demonstrate that the pRb-binding, but not the p53 promote S phase entry (Figure 1a) (Hatakeyama and binding domain is required to confer antiestrogen Weinberg, 1995; Nevins et al., 1997). resistance. Thus, pRb is an important target of estrogen The e€ects of estrogen and antiestrogen treatments action, and its inactivation can contribute to the on cell cycle progression have primarily been investi- development of antiestrogen resistance. Oncogene gated in the MCF-7 cell line, which is an estrogen (2000) 19, 4746 ± 4753. receptor positive human breast cancer cell line. Proliferation of MCF-7 cells is estrogen-dependent Keywords: antiestrogen; breast cancer; cell cycle; pRb; and antiestrogen sensitive both in vitro and in vivo T antigen (Osborne et al., 1985; Soule and McGrath, 1980; Sutherland et al., 1983a), and in the presence of antiestrogens these cells accumulate in G1. Estrogen Introduction treatment of G1-arrested MCF-7 cells induces S phase entry, and this induction is associated with increases in Approximately 40% of human breast cancers require cyclin D1 protein levels, cyclin D1/CDK4 kinase estrogens such as 17b-estradiol for growth (Melora et activity, CDK2 kinase activity, and pRb phosphoryla- al., 1998). In the absence of estrogens or the presence tion (Foster and Wimalasena, 1996; Planas-Silva and of antiestrogens, proliferation of these tumor cells is Weinberg, 1997; Prall et al., 1997; Watts et al., 1995). inhibited and they accumulate in the G1 phase of the In addition, overexpression of cyclin D1 in the cell cycle (Sutherland et al., 1983a,b). Because of their presence of antiestrogens leads to pRb phosphoryla- ability to inhibit proliferation, antiestrogens are often tion, and is sucient to promote proliferation for at used to treat breast cancer patients. However, a least one cell cycle (Pacilio et al., 1998; Prall et al., signi®cant proportion of tumors that initially respond 1998). to antiestrogen therapy eventually progress to a One explanation for the ®ndings outlined above is resistant state in which cell proliferation is no longer that pRb is the ultimate target of the proliferative dependent on estrogens or inhibited by antiestrogens. e€ects of estrogen, and that overexpression of cyclin Understanding the molecular mechanisms by which D1 promotes proliferation in the presence of antiestro- estrogens and antiestrogens regulate passage from G1 gens by increasing pRb phosphorylation via the to S phase, and those by which cells become activation of cyclinD1/CDK4 complexes. However, it antiestrogen resistant, are therefore important goals is also possible that functions of cyclin D1 other than activating CDK4 contribute to or are responsible for antiestrogen resistance. For example, regulation of cyclin E/CKD2 kinase activity by estrogen and *Correspondence: SE Conrad, Department of Microbiology, Giltner antiestrogens occurs in the absence of changes in cyclin Hall, Michigan State University, East Lansing, Michigan, MI 48824- E or CDK2 protein levels. To account for this 1101, USA Received 17 April 2000; revised 31 July 2000; accepted 1 August observation, it has been proposed that induced or 2000 overexpressed cyclin D1 protein binds and sequesters Reversal of antiestrogen sensitivity by T antigens H Varma and SE Conrad 4747 protein is an important functional target of estrogen, and suggest that inactivation of normal pRb function could play a role in the development of antiestrogen resistance during tumor progression in vivo.

Results

Inhibition of DNA synthesis by antiestrogens in MCF-7 cells Antiestrogens inhibit proliferation of MCF-7 cells. This is demonstrated in Figure 2a, in which the percentages of cells incorporating BrdUrd into DNA are compared in cultures incubated in basal medium containing 5% charcoal stripped serum (CSS), and in this medium supplemented with 17b-estradiol (E), 4-hydroxytamox- ifen (TAM) or ICI 182 780 (ICI). As previously reported by others, the dependence of MCF-7 cell proliferation on exogenous E in cell culture is not absolute (Lippman et al., 1976; Sutherland et al., 1983a; Wakeling et al., 1991). E stimulated prolifera- tion, but there was considerable BrdUrd incorporation Figure 1 Model for pRb inactivation by (a) estrogen treatment and (b) LT expression in MCF-7 cells in the basal medium in these assays. TAM is a non- steroidal antiestrogen with both agonist and antagonist activities. It acted as an antagonist in these assays, and decreased proliferation below the level observed in CSS the CDK inhibitor p21cip1, leading to activation of alone. ICI is a steroidal antiestrogen with no agonist cyclin E/CDK2 complexes (Planas-Silva and Weinberg, activity. As previously reported, it was a more potent 1997; Prall et al., 1997). Another potential mechanism inhibitor of MCF-7 cell proliferation than TAM of cyclin D1 action is via its interactions with cellular (Wakeling et al., 1991), and decreased the percentage transcription factors. Cyclin D1 binds to and mod- of cells incorporating BrdUrd to approximately 3 ± 5%. ulates the activity of several transcription factors, In addition, MCF-7 cell derivatives that are resistant to including the estrogen receptor (Neuman et al., 1997; ICI are cross-resistant to TAM, while TAMr cells are Zwijsen et al., 1997) and a myb-like transcription often sensitive to ICI (Brunner et al., 1993, 1997). The factor DMP1 (Hirai and Sherr, 1996; Inoue and Sherr, ability of a treatment to confer ICI resistance is 1998), and either of these activities could have therefore likely to also confer TAM resistance, while profound e€ects on cell proliferation. the converse is not true. For these reasons, ICI was the If pRb is the ultimate target of the proliferative antiestrogen used in the remainder of our experiments. e€ects of estrogen, one would predict that inactivation of this tumor suppressor protein would be sucient to Expression of SV40 large T antigen confers antiestrogen confer antiestrogen resistance to MCF-7 cells. Alter- resistance to MCF-7 cells, and the pRb binding function is natively, if functions of cyclin D1 other than required for this effect promoting pRb phosphorylation are critical for its e€ects on estrogen dependence, then inactivation of To determine the e€ect of SV-LT expression on pRb would not be sucient to promote antiestrogen antiestrogen sensitivity, plasmids encoding either wild resistance. To directly test the e€ect of pRb inactiva- type or mutant SV-LT proteins were introduced into tion on antiestrogen sensitivity, we have expressed the MCF-7 cells, and the ability of the transfected cells to simian virus 40 (SV) or polyoma virus (Py) large T incorporate BrdUrd into DNA in the presence of ICI antigens (LTs) in MCF-7 cells. Both SV-LT and Py-LT was determined as described in Materials and methods. bind to and inactivate pRb (Brodsky and Pipas, 1998; SV-LT is a multifunctional protein that interacts with a DeCaprio et al., 1988; Larose et al., 1991; Ludlow et number of cellular proteins, including the pRb family al., 1990), thereby allowing E2F to activate transcrip- (pRb, p107, p130), p53, p300 and TBP (Brodsky and tion and promote S phase entry (Figure 1b). SV-LT Pipas, 1998; Cole, 1996). The mutants used in these has several other activities in addition to pRb-binding, experiments (Table 1) were defective in one or more of including the ability to inactivate the tumor suppressor these activities that seemed likely to play a role in protein p53 (Lane and Crawford, 1979; Linzer and promoting antiestrogen resistance. As described in Levine, 1979). Py-LT shares most of these properties, Materials and methods, none of the constructs but di€ers from SV-LT in that it does not bind to p53 encoding SV-LT were capable of replicating extra- (Brodsky and Pipas, 1998; Cole, 1996). Using wild type chromosomally, and any BrdUrd incorporation de- and mutant forms of the viral proteins, we demonstrate tected therefore represented cellular DNA replication. that expression of either SV40 or polyoma virus LT is Representative micrographs of cells expressing SV- sucient to confer antiestrogen resistance to MCF-7 LT and/or incorporating BrUrd are shown in Figure 3. cells. This e€ect is dependent upon an intact pRb The percentages of SV-LT positive and SV-LT negative binding domain, but does not require a p53 binding cells in each transfected culture that were also BrdUrd domain. These data establish that pRb or a related positive was determined, and the results of this analysis

Oncogene Reversal of antiestrogen sensitivity by T antigens H Varma and SE Conrad 4748 are shown in Figure 2b. The percentage of SV-LT negative cells that incorporated BrdUrd in the transfected cultures (3 ± 5%) was similar to untrans- fected, ICI treated controls (Figure 2a,b), indicating that the transfection protocol did not cause prolifera- tion. In contrast, cells expressing wild type SV-LT incorporated BrdUrd at a level comparable to the cycling (E) population. These results indicate that SV- LT is able to completely overcome an ICI-mediated cell cycle arrest in this assay. The PVU1 mutant is defective in binding to pRb family members. In the presence of ICI, cells expressing PVU1 protein incorporated BrdUrd to the same low extent as untransfected cells (Figures 2b and 3). Both the transfection eciency of the PVU1 construct, and the level of expression of PVU1 protein in transfected cells were equivalent to the wild type (Figure 2c), indicating that the defect in this mutant was not due to a failure of the protein to be expressed. We therefore conclude that the pRb-binding domain of SV-LT is absolutely required to overcome an ICI induced cell cycle arrest. Two SV-LT mutants defective in p53 binding were also examined. One (2809) contains an insertion of four amino acids at position aa409, and the second (T1 ± 259) is a truncation mutant encoding the ®rst 259 amino acids of the protein. Both the 2809 and T1 ± 259 mutants were 60 ± 70% as e€ective as wild type SV-LT at inducing proliferation in the presence of ICI (Figure 2b). To investigate potential reasons for this small, but reproducible decrease in the ability of these mutants to overcome an ICI-induced arrest, both the transfection eciency of the DNA constructs and the level of expression of the encoded proteins in transfected cell populations were examined. As shown in Figure 2c, the percentage of cells expressing detectable SV-LT in 2809 and T1 ± 259 transfected cultures was approximately twofold lower than with the wild type, although the intensity of staining in LT positive cells was not detectably di€erent between the three constructs. In agreement with these observations, overall T1 ± 259 protein levels in transfected cultures were approxi- mately twofold lower than wild type as determined by Western blots. Interestingly, cultures containing the 2809 mutant expressed very low levels of full length LT protein, although they incorporated BrdUrd with approximately the same eciency as T1 ± 259. How- ever, it was noted that 2809-transfected cells expressed relatively high levels (equal to that of the T1 ± 259 Figure 2 Ability of wild type and mutant LT proteins to mutant) of a small protein of approximately 20 ± promote antiestrogen resistance. MCF7 cells were transfected with LT-encoding plasmids, labeled with BrdUrd, and stained for 25 kDa. Since the antibody used in these Western blots LT expression and BrdUrd incorporation as described in recognizes an N-terminal epitope of SV-LT (Harlow et Materials and methods. (a) The percentage of cells in al., 1981), the 20 ± 25 kDa peptide may be an N- untransfected cultures that incorporated BrdUrd during a 5 h terminal fragment that retains the pRb binding domain labeling in the presence of CSS, CSS+E, CSS+TAM or and has activity in proliferation assays. Alternatively, CSS+ICI. The results represent the mean+s.e. of an experiment done in triplicate. (b) The percentages of LT positive (®lled bars) all of the proliferative activity of the 2809 mutant may and LT negative (open bars) cells that incorporated BrdUrd in be due to the small amount of full-length protein the presence of ICI in transfected cultures. The percentage of cells present in the transfected cells. In summary, two that incorporated BrdUrd during a 5 h labeling in the presence of di€erent SV-LT mutants that are defective for p53 E or ICI in untransfected cultures are shown on the left side of the ®gure. The results represent the mean+s.e of at least three binding were able to induce proliferation in ICI treated independent experiments. (c) A representative Western blot MCF-7 cells. However, the magnitude of the e€ect was showing expression levels of wild type and mutant SV-LTs. signi®cantly less than that seen with the wild type Arrows point to the T1-259 truncation mutant, and to a smaller protein. Due to di€erences in the levels of expression of (20 ± 25 kDa) form of LT detected in cells transfected with the 2809 mutant. Values for the transfection eciency (percentage) of the encoded proteins, the data obtained did not allow each construct are represented as the mean+s.e. for at least three us to distinguish whether the decreased ability of the independent experiments 2809 and T1 ± 259 mutants to overcome an antiestro-

Oncogene Reversal of antiestrogen sensitivity by T antigens H Varma and SE Conrad 4749 Table 1 Characteristics of wild type and mutant viral T-antigens used in study LT protein pRb binding p53 binding Mutation Reference

SV-LT + + none (Zhu and Cole, 1989) PVU1 (SV) 7 + del aa 107/112 (Kalderon and Smith, 1984) 2809 (SV) + 7 ins 409 (4aa) (Zhu and Cole, 1989) T1-259 (SV) + 7 truncation (Dobbelstein et al., 1992) Py-LT + 7 none (Howes et al., 1996; Larose et al., 1990) dl-141 (Py) 77del aa 141/146 (Larose et al., 1991)

the di€erence was not statistically signi®cant. These results therefore indicate that Py-LT is as ecient as SV-LT in overcoming an antiestrogen-mediated arrest of MCF-7 cells, and establish that p53 inactivation is not required for this e€ect. The results presented above indicated that p53 inactivation was not required to confer antiestrogen resistance to cells under conditions where pRb family members were inactivated. To examine the e€ects of p53 inactivation alone on antiestrogen sensitivity, a plasmid encoding del239, a dominant negative (dn) p53 mutant (Gao et al., 1996), was transfected into MCF-7 cells. Expression of this protein did not increase BrdUrd incorporation above that seen in untransfected cells (Figure 4). In addition, del239 expression did not enhance the level of proliferation induced by Py-LT alone in co-transfection experiments. We therefore conclude that p53 inactivation alone is not sucient to confer antiestrogen resistance, and that it does not enhance the resistance conferred by pRb inactivation, in MCF-7 cells.

Py-LT confers antiestrogen resistance in long term assays It was previously reported that over-expression of Figure 3 Double indirect immuno¯uorescence of transfected cyclin D1 conferred antiestrogen resistance to MCF-7 cells. MCF7 cells were transfected with wild type or mutant cells in short term assays, but was not sucient to LT-encoding plasmids, incubated in medium containing ICI, and promote long term growth in the presence of labeled with BrdUrd as described in Materials and methods. SV- antiestrogen (Pacilio et al., 1998). We therefore wanted LT was detected using a rhodamine-conjugated secondary anti- body (red ¯uorescence), and BrdUrd incorporation with an to determine if LT expression was sucient to confer FITC-conjugated secondary antibody (green ¯uorescence). Py- antiestrogen resistance to MCF-7 cells in long term LT was detected using an FITC-conjugated secondary antibody, proliferation assays. Constructs encoding either a wild and in this case BrdUrd incorporation was detected with a type or pRb binding mutant (dl141) of Py-LT were rhodamine-conjugated secondary antibody. Overlay images were generated to show nuclei that were positive for both LT and introduced into MCF-7 cells using the pneo-LT1 BrdUrd, which appear yellow in the photograph vector, which contains a G418r gene for selection of stably transfected cells. Individual G418r colonies were selected, expanded into cell lines, and tested for gen-induced arrest was due to decreased expression of expression of Py-LT protein by Western blotting. these proteins, or if they were truly less active than Two cell lines expressing the wild type protein (WT1 wild type SV-LT. and WT5) and one expressing the pRb binding mutant (dl 141) were chosen for further analysis (Figure 5a). One G418r cell line that did not express detectable Inactivation of p53 does not confer antiestrogen resistance levels of Py-LT protein (WT4), was used as a negative To further investigate the possibility that p53 inactiva- control. All of these cell lines grew well in the presence tion played a role in SV-LT's ability to confer of 5% FBS, and contained similar amounts and ratios antiestrogen resistance, we examined the ability of of hyper- and hypo-phosphorylated pRb in this Py-LT to promote proliferation of ICI-arrested MCF-7 medium (Figure 5a). cells. Py-LT binds to pRb, but does not bind or The e€ects of ICI treatment on proliferation of the inactivate p53 (Brodsky and Pipas, 1998; Cole, 1996). Py-LT expressing cells were evaluated in several ways. A plasmid encoding wild type Py-LT (pCMVLT) was In Figure 5b, pRb phosphorylation and cyclin A transfected into MCF-7 cells, and proliferation was expression were examined by Western blotting after 2 assayed in the presence of ICI as described above. days of ICI treatment. Previous studies have shown Representative ¯uorescence micrographs are shown in that both pRb phosphorylation and cyclin A expres- Figure 3, and summaries of the experiments are shown sion are inhibited by ICI treatment of the parental in Figures 2b and 4. Although the average induction of MCF-7 cell line (Watts et al., 1995). In agreement with proliferation was slightly lower than seen with SV-LT, this, hyperphosphorylated pRb and cyclin A protein

Oncogene Reversal of antiestrogen sensitivity by T antigens H Varma and SE Conrad 4750 binding domain is absolutely required for viral tumor antigens to promote proliferation in the presence of ICI. The e€ects of ICI treatment on pRb phosphoryla- tion and cyclin A expression were paralleled by changes in cell proliferation, as measured by either [3H]thymidine or BrdUrd incorporation (data not shown). Taken together, the results in stable transfec- tants con®rm those seen in transient assays, and indicate that expression of Py-LT confers antiestrogen resistance to MCF-7 cells.

Discussion

Previous evidence linking estrogen and antiestrogen Figure 4 E€ects of Py-LT and dominant negative p53 expres- sion on ICI sensitivity in MCF-7 cells. MCF7 cells were treatments to pRb phosphorylation suggested that pRb transfected with plasmids encoding Py-LT, a dn p53 (del239), might be a crucial target of estrogen, and that or both. BrdUrd incorporation in the presence of ICI was inactivation of this tumor suppressor protein might determined in Py-LT or del239 positive cells (®lled bars) and Py- lead to antiestrogen resistance. We tested this possibi- LT or del239 negative cells (open bars), and is expressed as a lity by expressing viral T antigens in MCF-7 cells, and percentage of the total cells. BrdUrd incorporation was also measured in cycling (E) cells as a positive control. Results determining their e€ects on antiestrogen sensitivity. represent the mean+s.e. of three independent experiments Expression of either the SV40 or polyoma virus large T antigen promotes MCF-7 cell proliferation in the presence of the pure antiestrogen ICI 182,780. Furthermore, this e€ect is dependent upon the ability of the viral proteins to bind and inactivate pRb, but is independent of their ability to bind p53. Thus, pRb is an important target of estrogens in regulating MCF-7 cell proliferation, and its inactivation can contribute to the development of antiestrogen resistance. The fact that both pRb inactivation and over- expression of cyclin D1 (Pacilio et al., 1998; Prall et al., 1998) promote MCF-7 cell proliferation in the presence of antiestrogens con®rms an important role for the cyclin D/pRb pathway in estrogen dependence. It also suggests that the primary mechanism by which cyclin D1 overcomes an antiestrogen-mediated arrest is by activating CDK 4/6, which in turn phosphorylate pRb. As discussed previously, CDK2 activity is inhibited in antiestrogen treated MCF-7 cells, and it has been proposed that a second role for cyclin D1 in promoting antiestrogen resistance is to sequester p21cip1, thereby leading to CDK2 activation. The results Figure 5 E€ects of Py-LT expression on ICI sensitivity in stably reported here suggest that either antiestrogen treatment transfected cells. MCF-7 cells were transfected with vectors fails to inhibit CDK2 activity in MCF-7 cells encoding Py-LT, and G418r colonies were selected and expressing T antigens, or that the requirement for r propagated as described in Materials and methods. (a) G418 CDK2 activity is bypassed in these cells. Since p21cip1 cell lines were assayed for Py-LT expression and pRb phosphor- transcription is induced by p53 (el-Deiry et al., 1993), ylation by Western blotting. Cell lines WT1 and WT5 expressed wild type Py-LT. dl-141 cells expressed a pRb binding mutant of one mechanism by which SV-LT could lead to CDK2 Py-LT. WT4 was a G418r clone that did not express detectable activation would be by binding and inactivating p53, Py-LT, and served as a negative control. Hyperphosphorylated leading to decreases in p21cip levels. However, this does pRb (ppRb) migrates more slowly than the hypophosphorylated not appear to play a role in promoting antiestrogen form (pRb) on these gels. (b) Cell lines were plated in medium containing 5% FBS, treated with either ICI (+) or vehicle (7) resistance, since LTs that do not inactivate p53 induce for 48 h, and cell extracts were prepared and analysed for pRb proliferation in the presence of ICI, and since a dn p53 phosphorylation and cyclin A expression by Western blotting. mutant does not induce proliferation in the presence of The blots were also probed for actin as a control for protein ICI. The mechanisms by which LTs overcome the loading inhibition of CDK2 activity by antiestrogens therefore remain to be elucidated. Although our results clearly demonstrate that an were undetectable in ICI-treated WT4 control cells, intact pRb-binding domain is required, they do not which do not express Py-LT. In contrast, WT1 and establish if pRb inactivation alone is sucient to WT5 cells, which express wild type Py-LT, contained confer antiestrogen resistance to MCF-7 cells. The Rb both hyperphosphorylated pRb and cyclin A when family also contains the related proteins p107 and grown in the presence of ICI. Cells expressing dl141, p130. We cannot eliminate a role for p107 and/or p130 the pRb binding mutant of Py-LT, had a similar inactivation in conferring antiestrogen resistance, since phenotype to WT4 cells, con®rming that the pRb- both SV-LT and Py-LT bind these proteins, and since

Oncogene Reversal of antiestrogen sensitivity by T antigens H Varma and SE Conrad 4751 the PVU1 and dl-141 mutations eliminate binding to defective in promoting proliferation in the presence of all members of the pRb family (Brodsky and Pipas, ICI. Thus, ST expression is not sucient to confer 1998; Cole, 1996). Since pRb is known to be antiestrogen resistance. In addition, the CMVLT vector phosphorylated in response to estrogen treatment, used to express Py-LT in transient transfections and since its role in regulating cell proliferation is well contains a LT cDNA, and does not encode ST. This established, it is likely to be the major target of plasmid confers ICI resistance to MCF-7 cells with the estrogen action. However, it would be interesting to same eciency as SV-LT, indicating that ST expression determine if other pRb family members also contribute is not required. to estrogen/antiestrogen sensitivity in breast cancer We attempted to determine whether LT expression cells. confers antiestrogen resistance to MCF-7 cells in long SV-LT possesses additional functions that might term proliferation assays. Although overexpression of contribute to antiestrogen resistance, including the cyclin D1 conferred antiestrogen resistance to MCF-7 ability to bind p300 and components of the basal cells in short term assays, it was not sucient to transcription apparatus (Brodsky and Pipas, 1998; promote the long term growth of these cells in the Cole, 1996). It is possible that LT binding to presence of antiestrogen (Pacilio et al., 1998). Stably components of the basal transcription apparatus plays transfected cells expressing Py-LT were resistant to ICI a role in promoting antiestrogen resistance, since this treatment, as demonstrated by the facts that they function maps to N-terminal domains of SV-LT that contained cyclin A and hyperphosphorylated pRb, and are maintained in the T1-259 mutant (Damania and synthesized DNA in the presence of ICI. However, this Alwine, 1996; Johnston et al., 1996). We are not aware resistance was not complete. As shown in Figure 5b, of any reports that Py-LT also binds to basal both pRb phosphorylation and cyclin A protein levels transcription factors, but it may do so. In contrast, were reduced somewhat by ICI treatment of WT1 and the p300 binding activity of LT is unlikely to be WT5 cells, although to a lesser extent than in the involved in promoting antiestrogen resistance, since control WT4 or dl141 cells. Similar e€ects were seen on several reports indicate that C-terminal sequences that BrdUrd and [3H]thymidine incorporation (data not are missing in the T1-259 mutant are required for p300 shown), suggesting that proliferation of Py-LT expres- binding by both SV-LT and Py-LT (Lill et al., 1997; sing MCF-7 cells is not completely resistant to Nemethova and Wintersberger, 1999). As discussed antiestrogen treatment. However, when the stably previously, the p53 binding/inactivation function of transfected cell lines were examined for Py-LT SV-LT is not required to confer antiestrogen resistance, expression by immuno¯uorescence, only a subset of since it is missing in both the T1-259 and 2809 cells had detectable levels of LT, even during selective mutants, and in Py-LT. growth in G418 (data not shown). The apparent partial It is interesting to compare the LT domains required ICI-sensitivity seen in these experiments may therefore to overcome cell cycle arrests in di€erent systems. be due to variable expression of Py-LT in the Expression of SV-LT can overcome a TGF-b1 transfected cells, rather than to an inability of LT to mediated G1 arrest in keratinocytes (Pietenpol et al., confer long term antiestrogen resistance. 1990) and mink lung epithelial cells (Laiho et al., Finally, the fact that both pRb inactivation and 1990), and this e€ect is dependent upon the pRb- cyclin D1 overexpression promote antiestrogen resis- binding domain. It is also probably independent of p53 tance in MCF-7 cells suggests that deregulation of the binding, since the HPV E7 protein, which does not pRb pathway might occur during the development of bind p53, has a similar e€ect (Laiho et al., 1990). antiestrogen resistance in vivo. Several indirect pieces of However, the absolute requirement for a pRb-binding evidence support this suggestion. Although 20 ± 25% of domain, and the lack of a role for p53 binding, in breast cancer cell lines examined lack functional pRb, promoting proliferation is not universal. For example, all estrogen-responsive breast cancer cell lines identi®ed at least four independent functions of SV-LT, including to date contain wild type pRb (Lee et al., 1988; T'Ang pRb binding, p53 binding, and an unidenti®ed function et al., 1988). In addition, in one study of primary abolished by the S189N mutation, contribute to its breast cancers a signi®cant correlation was observed ability to promote proliferation in serum starved CV1P between the absence of pRb expression and a failure to cells (Dickmanns et al., 1994; Dobbelstein et al., 1992). respond to antiestrogen therapy (Anderson et al., Mutations in any one of these functions, including pRb 1996). Together, these data suggest that hormone binding, had little e€ect on LT's ability to promote dependence cannot be maintained in the absence of proliferation in this system. In contrast, double pRb function, and that changes in pRb status or mutants, including pRb binding/p53 binding double regulation could be involved in the progression of mutants, showed signi®cant defects. This variability in tumors to antiestrogen resistance. the requirements for pRb and/or p53 inactivation in overcoming cell cycle arrests may be due to di€erences in the mechanisms by which treatments such as ICI, TGF-b1 and serum starvation induce cell cycle arrests, Materials and methods or may represent di€erences in the requirements for pRb and/or p53 inactivation between cell lines. Cell culture We have considered the possibility that small T MCF-7 cells were obtained from Dr Michael Johnson antigen (ST) plays a role in conferring antiestrogen (Lombardi Cancer Center) and were maintained in IMEM resistance, since the SV-LT encoding vectors used in media (Bio¯uids Inc., Rockville, MD, USA) supplemented these experiments contain the entire early region of the with 5% fetal bovine serum (Hyclone), penicillin (100 units/ viral genome, and therefore also encode ST. The PVU1 ml) and streptomycin (100 mg/ml). To study the e€ects of mutant encodes a wild type ST, yet it is completely estrogen and antiestrogens, cells were cultured in phenol red

Oncogene Reversal of antiestrogen sensitivity by T antigens H Varma and SE Conrad 4752 free IMEM supplemented with 5% charcoal stripped fetal idine (BrdUrd) for 5 h. For control proliferating cultures, bovine serum (Hyclone), with or without the addition of cells were grown in medium containing charcoal stripped 0.2 nM 17b-estradiol (Sigma), 1 mM 4-hydroxytamoxifen serum supplemented with 0.2 nM 17b-estradiol. For stable (Sigma), or 100 nM ICI 182 780 (Zeneca Pharmaceuticals). transfections, 1 mg of DNA (pneoLT1 or dl141) was introduced into MCF7 cells using Lipofectin reagent (Gibco BRL). Colonies were selected for G418 resistance in medium Antibodies and plasmids containing 400 mg/ml G418 and expanded into clonal cell The antibody to SV-LT was a supernatant of the L19 lines. Py-LT expression in these cell lines was assayed by hybridoma cell line, which was a gift of Dr E Harlow Western blotting. (Harlow et al., 1981). A rat monoclonal antibody, Ab-1 (Oncogene Science) was used to detect Py-LT by indirect Indirect immunofluorescence immuno¯uorescence. A polyclonal rat antiserum was used for Py-LT immunoblotting, and was a gift of Dr Michele Fluck. Forty-eight hours after transfection, cells on coverslips were Antibodies for pRb, cyclin A, and p53 were from ®xed with ice cold acetone/methanol (50 : 50 v/v), and washed Pharmingen. The mouse a-BrdUrd monoclonal antibody with phosphate bu€ered saline (PBS). The following was obtained from Boehringer Mannheim, and sheep a- procedures were used for antibody staining, with PBS washes BrdUrd was from Fitzgerald Industries, Concord, MA, USA. (three times) being conducted between each treatment. FITC conjugated goat a-rat, donkey a-sheep, and goat a- Coverslips containing ®xed cells were incubated with a mouse antibodies were obtained from Sigma Biochemicals. primary antibody against SV-LT or Py-LT for 45 min at Rhodamine Red-X conjugated donkey a-mouse antibody was 378C, then with the appropriate ¯uorochrome-conjugated from Jackson Immunoresearch. secondary antibody for 1 h at room temperature. Following The plasmids encoding wild type SV-LT, PVU1, and 2809 this procedure, cells were re-®xed in a 50 mM glycine, 50% were obtained from Dr Charles Cole (Zhu and Cole, 1989; ethanol solution (pH 2.1) for 45 min at room temperature. Zhu et al., 1991). The SV40 genome in all of these plasmids Samples were then treated with 4 N HCl for 15 min to contains a 6 bp deletion at the viral origin of replication that denature the DNA. The acid was aspirated, and coverslips renders them replication incompetent (Gluzman et al., 1980; were washed in PBS to neutral pH (7.0). Samples were then Keller and Alwine, 1984; Zhu et al., 1991). The truncation incubated with an appropriate anti-BrdUrd antibody (sheep mutant (T1 ± 259) was provided by Dr Ellen Fanning for SV-LT and mouse for Py-LT) for 45 min at 378C, (Dobbelstein et al., 1992). The T antigen encoded by this washed, and incubated with a second ¯uorescent antibody mutant lacks helicase and ATPase activity and will therefore (donkey anti-sheep FITC for SV-LT and donkey anti-mouse not promote replication from the viral origin (Clark et al., rhodamine Red X for Py-LT). Finally, the coverslips were 1983; Cole et al., 1986; Pipas et al., 1983). The plasmid extensively washed with PBS, mounted on glass slides, and pCMVLT, which contains a cDNA encoding Py-LT, was viewed under a ¯ourescence microscope (Olympus). The obtained from Dr Brian Scha€hausen (Howes et al., 1996). percentage of cells that were positive for LT (or p53) alone, The pneoLT1 and dl-141-encoding vectors used for stable BrdUrd alone, or both, were counted in each transfection. transfections were provided by Dr Michael Bastin (Larose et More than 100 LT-positive and 100 LT-negative cells were al., 1990, 1991). The p53 dominant negative mutant (del239) counted in each experiment. The images shown in Figure 2 was a gift of Dr John Minna (Takahashi et al., 1989). were taken on a Meridian Instruments confocal microscope, and processed with Insight IQ computer software to generate single color and overlay images. Transfections For transient transfections, 56105 MCF-7 cells were plated in 60 mm tissue culture dishes containing 2 ± 18 mm cover slips and incubated at 378C overnight. Plasmid DNA was Acknowledgments added to the cells using the Superfect reagent (Qiagen) The authors would like to thank Dr Michele Fluck, Dr following the protocol suggested by the manufacturer. SV40- Richard Miksicek, Andrew Skildum and Christopher LT transfections were done using 2.5 and 5 mg of DNA per Ontiveros for helpful comments on the manuscript. We dish, and pCMV-LT transfections were done using 0.3 mg also thank Drs Charles Cole, Ellen Fanning, Brian DNA per dish. These amounts were chosen because they gave Sha€hausen, Michael Bastin, and John Minna for provid- the highest transfection eciencies in pilot experiments. After ing plasmids used in this research. This work was 3 h, the transfection medium was removed and replaced with supported in part by grants from the National Institutes phenol red free IMEM supplemented with 5% charcoal of Health (CA76647, SE Conrad) and the Elsa U Pardee stripped serum and 100 nM ICI 182 780. Cells were incubated Foundation (SE Conrad). Additional funds were provided for 43 h and then labeled with 25 mM 5-bromo-2'-deoxyur- by the Cancer Center at Michigan State University.

References

Anderson JJ, Tiniakos DG, McIntosh GG, Autzen P, Henry Cole CN. (1996). Fundamental Virology, Bernard N, Fields JA, Thomas MD, Reed J, Horne GM, Lennard TW, Dmk, Peter M, Howley (ed.). Lippincott-Raven, pp 917 ± Angus B and Horne CH. (1996). J. Pathol., 180, 65 ± 70. 945. Brodsky JL and Pipas JM. (1998). J. Virol., 72, 5329 ± 5334. Cole CN, Tornow J, Clark R and Tjian R. (1986). J. Virol., Brunner N, Boysen B, Jirus S, Skaar TC, Holst-Hansen C, 57, 539 ± 546. Lippman J, Frandsen T, Spang-Thomsen M, Fuqua SA Damania B and Alwine JC. (1996). Genes Dev., 10, 1369 ± and Clarke R. (1997). Cancer Res., 57, 3486 ± 3493. 1381. Brunner N, Frandsen TL, Holst-Hansen C, Bei M, DeCaprio JA, Ludlow JW, Figge J, Shew JY, Huang CM, Thompson EW, Wakeling AE, Lippman ME and Clarke Lee WH, Marsilio E, Paucha E and Livingston DM. R. (1993). Cancer Res., 53, 3229 ± 3232. (1988). Cell, 54, 275 ± 283. Clark R, Peden K, Pipas JM, Nathans D and Tjian R. (1983). Mol. Cell Biol., 3, 220 ± 228.

Oncogene Reversal of antiestrogen sensitivity by T antigens H Varma and SE Conrad 4753 Dickmanns A, Zeitvogel A, Simmersbach F, Weber R, Nemethova M and Wintersberger E. (1999). J. Virol., 73, Arthur AK, Dehde S, Wildeman AG and Fanning E. 1734 ± 1739. (1994). J. Virol., 68, 5496 ± 5508. Neuman E, Ladha MH, Lin N, Upton TM, Miller SJ, Dobbelstein M, Arthur AK, Dehde S, van Zee K, DiRenzoJ,PestellRG,HindsPW,DowdySF,BrownM Dickmanns A and Fanning E. (1992). Oncogene, 7, 837 ± and Ewen ME. (1997). Mol. Cell Biol., 17, 5338 ± 5347. 847. Nevins JR, Leone G, DeGregori J and Jakoi L. (1997). J. el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Cell Physiol., 173, 233 ± 236. Trent JM, Lin D, Mercer WE, Kinzler KW and Vogelstein Osborne CK, Hobbs K and Clark GM. (1985). Cancer Res., B. (1993). Cell, 75, 817 ± 825. 45, 584 ± 590. Foster JS and Wimalasena J. (1996). Mol. Endocrinol., 10, Pacilio C, Germano D, Addeo R, Altucci L, Petrizzi VB, 488 ± 498. Cancemi M, Cicatiello L, Salzano S, Lallemand F, Gao Q, Hauser SH, Liu XL, Wazer DE, Madoc-Jones H and Michalides RJ, Bresciani F and Weisz A. (1998). Cancer Band V. (1996). Cancer Res., 56, 3129 ± 3133. Res., 58, 871 ± 876. Gluzman Y, Frisque RJ and Sambrook J. (1980). Cold PietenpolJA,SteinRW,MoranE,YaciukP,SchlegelR, Spring Harb. Symp. Quant. Biol., 44, 293 ± 300. Lyons RM, Pittelkow MR, Munger K, Howley PM and Harlow E, Crawford LV, Pim DC and Williamson NM. Moses HL. (1990). Cell, 61, 777 ± 785. (1981). J. Virol., 39, 861 ± 869. Pipas JM, Peden KW and Nathans D. (1983). Mol. Cell Biol., Hatakeyama M and Weinberg RA. (1995). Prog. Cell Cycle 3, 203 ± 213. Res., 1, 9 ± 19. Planas-Silva MD and Weinberg RA. (1997). Mol. Cell Biol., Hirai H and Sherr CJ. (1996). Mol. Cell Biol., 16, 6457 ± 17, 4059 ± 4069. 6467. Prall OW, Rogan EM, Musgrove EA, Watts CK and Howes SH, Bockus BJ and Scha€hausen BS. (1996). J. Sutherland RL. (1998). Mol. Cell Biol., 18, 4499 ± 4508. Virol., 70, 3581 ± 3588. Prall OWJ, Sarcevic B, Musgrove EA, Watts CKW and Inoue K and Sherr CJ. (1998). Mol. Cell Biol., 18, 1590 ± Sutherland RL. (1997). J. Biol. Chem., 272, 10882 ± 10894. 1600. Sherr CJ. (1993). Cell, 73, 1059 ± 1065. Johnston SD, Yu XM and Mertz JE. (1996). J. Virol., 70, Sherr CJ. (1994). Cell, 79, 551 ± 555. 1191 ± 1202. Soule HD and McGrath CM. (1980). Cancer Lett., 10, 177 ± Kalderon D and Smith AE. (1984). Virology, 139, 109 ± 137. 189. Kato J, Matsushime H, Hiebert SW, Ewen ME and Sherr CJ. Sutherland RL, Green MD, Hall RE, Reddel RR and Taylor (1993). Genes Dev., 7, 331 ± 342. IW. (1983a). Eur. J. Cancer Clin. Oncol., 19, 615 ± 621. Keller JM and Alwine JC. (1984). Cell, 36, 381 ± 389. Sutherland RL, Reddel RR and Green MD. (1983b). Eur. J. Laiho M, DeCaprio JA, Ludlow JW, Livingston DM and Cancer Clin. Oncol., 19, 307 ± 318. Massague J. (1990). Cell, 62, 175 ± 185. TakahashiT,NauMM,ChibaI,BirrerMJ,RosenbergRK, Lane DP and Crawford LV. (1979). Nature, 278, 261 ± 263. Vinocour M, Levitt M, Pass H, Gazdar AF and Minna JD. Larose A, Dyson N, Sullivan M, Harlow E and Bastin M. (1989). Science, 246, 491 ± 494. (1991). J. Virol., 65, 2308 ± 2313. T'Ang A, Varley JM, Chakraborty S, Murphree AL and Larose A, St-Onge L and Bastin M. (1990). Virology, 176, Fung YK. (1988). Science, 242, 263 ± 266. 98 ± 105. Wakeling AE, Dukes M and Bowler J. (1991). Cancer Res., Lee EY, To H, Shew JY, Bookstein R, Scully P and Lee WH. 51, 3867 ± 3873. (1988). Science, 241, 218 ± 221. Watts CK, Brady A, Sarcevic B, deFazio A, Musgrove EA Lill NL, Tevethia MJ, Eckner R, Livingston DM and and Sutherland RL. (1995). Mol. Endocrinol., 9, 1804 ± Modjtahedi N. (1997). J. Virol., 71, 129 ± 137. 1813. Linzer DI and Levine AJ. (1979). Cell, 17, 43 ± 52. Zhu JY and Cole CN. (1989). J. Virol., 63, 4777 ± 4786. Lippman M, Bolan G and Hu€ K. (1976). Cancer Res., 36, Zhu JY, Rice PW, Chamberlain M and Cole CN. (1991). J. 4595 ± 4601. Virol., 65, 2778 ± 2790. Ludlow JW, Shon J, Pipas JM, Livingston DM and Zwijsen RM, Wientjens E, Klompmaker R, van der Sman J, DeCaprio JA. (1990). Cell, 60, 387 ± 396. Bernards R and Michalides RJ. (1997). Cell, 88, 405 ± 415. Melora DB, Allred CC and O'Conell P. (1998). Principles of Molecular Medicine, 1st edn. Humana Press.

Oncogene