Bcl-2 expression delays hepatocyte cell cycle progression ... - Nature

ã

Oncogene (2002) 21, 1548 ± 1555 2002 Nature Publishing Group All rights reserved 0950 ± 9232/02 $25.00 www.nature.com/onc

Bcl-2 expression delays hepatocyte cell cycle progression during liver regeneration Mary E Vail1,2, Michelle L Chaisson1, James Thompson1 and Nelson Fausto*,1 1

Department of Pathology, University of Washington, Seattle, Washington, WA, 98195, USA; 2Molecular and Cellular Biology Program, University of Washington, Seattle, Washington, WA 98195, USA

Bcl-2 is the prototype of a family of genes that prevent apoptosis. However, several reports indicate that Bcl-2 may also act as a cell cycle modulator. In several human tumors, Bcl-2 expression correlates with a more favorable prognosis and lower tumor proliferative activity. We have shown that Bcl-2 expression delays liver tumor development in transgenic mice even when the gene is turned on shortly before the time of tumor development. We hypothesized that Bcl-2 may delay liver tumorigenesis by interfering with hepatocyte proliferation. To test whether Bcl-2 expression may act on hepatocyte replication we studied liver regeneration in Bcl-2 transgenic mice and wild-type littermates. DNA replication was delayed by approximately 8 h in Bcl-2 transgenic mice compared to the timing of the response in wild-type littermates. Cyclin D expression showed no alterations in the regenerating liver of Bcl-2 transgenic mice. In contrast, there was a delay in the expression of p107, cyclin E and in the activity of cyclin E/cdk 2 activity. These results show that Bcl-2 expression delays cell cycle progression in hepatocytes and suggests that it acts at a step involving cyclin E and p107. Oncogene (2002) 21, 1548 ± 1555. DOI: 10.1038/sj/ onc/1205212 Keywords: liver regeneration; cell cycle; Bcl-2; transgenic mice; hepatocyte Introduction Bcl-2 was ®rst identi®ed as an oncogene by its ability to inhibit lymphocyte cell death leading to the development of B-cell lymphoma (Korsmeyer, 1992). Based on this ®nding, it would be anticipated that overexpression of Bcl-2 contributes to tumor development. Recent studies have shown however that cancers of the breast and colon with high levels of Bcl-2 expression have a better prognosis than cancers with low levels of Bcl-2 expression (Silvestrini et al., 1996;

*Correspondence: N Fausto, Department of Pathology, University of Washington, K-078 Health Sciences Building, Box 357705, Seattle, WA 98195-7705, USA; E-mail: [email protected] Received 2 October 2001; revised 28 November 2001; accepted 28 November 2001

Sinicrope et al., 1995). Furthermore, Bcl-2 expression correlates with lower proliferative activity in intermediate and high grade non-Hodgkins lymphomas (Winter et al., 1998). Thus, at least for some tissues, the relationships between Bcl-2 expression, cell proliferation and cancer development are complex and may not conform to the expectation that Bcl-2 expression favors tumor development. We have shown that excess Bcl-2 expression in TGFa/Bcl-2 double transgenic mice delays the development of liver tumors induced by the growth factor (Vail et al., 2001). This observation is in agreement with data demonstrating that Bcl-2 inhibits c-mycinduced liver carcinogenesis (de La Coste et al., 1999). In our studies, Bcl-2 had an inhibitory eect in liver carcinogenesis even when its expression was turned on late during the process. This led us to suggest that Bcl2 may slow down hepatocyte proliferation in liver carcinogenesis. When stimulated to proliferate hepatocytes must emerge from quiescence (G0) and enter into the ®rst gap phase (G1) of the cell cycle prior to DNA synthesis (Fausto, 2000) Cell cycle progression is regulated by the activity of complexes consisting of cyclins and cyclin dependent kinases (cdks). Early G1 is associated with cyclin D/cdk 4, which in turn drives the activation of the cyclin E/cdk 2 complex (Sherr and Roberts, 1999). Together, cyclin/cdk complexes function in a coordinated manner to phosphorylate the retinoblastoma protein (Rb) and the related p107 and p130 proteins, thereby releasing the bound E2F transcription factors (Classon and Dyson, 2001; Dyson, 1998; Nevins, 1998). It has been reported that overexpression of Bcl-2 can delay cell cycle progression and DNA replication (Borner, 1996; Huang et al., 1997; Linette et al., 1996; Mazel et al., 1996; O'Reilly et al., 1996, 1997). Although the mechanisms by which Bcl-2 delays cell cycle progression are not exactly known, Vairo et al. (2000) showed that Bcl-2 overexpression led to sustained levels of the cdk inhibitor p27, and increased levels of the pocket protein p130, resulting in a prolonged G1 phase. Other studies have found eects of Bcl-2 on the Rb/E2F complex as well as on p21 (Truchet et al., 2000). Overexpression of Bcl-2 can also increase the rate at which cells withdraw from the cell cycle to become quiescent (Vairo et al., 1996). Based on these observations we analysed the eect of Bcl-2

Bcl-2 and liver regeneration ME Vail et al

1549

on the hepatocyte cell cycle in vivo during liver regeneration after partial hepatectomy (PH) in Bcl-2 transgenic mice. Surgical resection of two thirds of the liver triggers a synchronized proliferative response that involves more than 95% of hepatoctyes. Bcl-2 expression in the transgenic mice delayed DNA replication after PH. The delay was associated with eects on the expression of cyclins E and A, the pocket protein p107 and cyclin E/cdk2 activity.

Results Bcl-2 delays hepatocyte DNA synthesis both in vivo and in vitro We tested the hypothesis that Bcl-2 inhibits hepatocyte replication by analysing the timing of DNA replication after PH on Bcl-2 transgenic mice and wild-type (WT) littermates. We measured DNA replication at 12 h after PH and at 4-h intervals between 28 and 48 h after surgery in mice injected with BrdU 2 h prior to killing. DNA replication, as measured by BrdU incorporation, was ®rst detectable in hepatocytes of WT mice at 28 h after PH and reached a maximum at 36 h. Between 36 and 44 h, the percentage of BrdU labeled cells remained between 15 and 25% and decreased at 48 h (Figure 1a). In contrast only 5% of hepatocytes of Bcl-2 transgenic mice were labeled at 32 h. In these animals, DNA replication did not reach a peak until 44 h after PH, 8 h after the WT mice (Figure 1a). Despite the delay in DNA replication in Bcl-2 transgenics, the overall amount of BrdU positive hepatocytes (as determined by calculating the area under each curve) during the regenerative process was similar in the two groups of mice. These results demonstrate that Bcl-2 delays hepatocyte DNA replication in vivo in quiescent hepatocytes that receive a proliferative stimulus. Because the Bcl-2 transgene is under the control of the MT1 promoter and its expression can be regulated by the addition of zinc to the drinking water (Vail et al., 2001), WT and Bcl-2 transgenic mice received 25 mM zinc sulfate drinking water at least 1 week prior to surgeries. Expression of the Bcl-2 transgene was determined by Western blot analysis of liver lysates of non-PH mice and at 28 and 44 h after surgery (Figure 1b). As expected, Bcl-2 was not detected in the WT mice, whereas all Bcl-2 transgenic mice had high levels of Bcl-2 expression throughout the duration of the study. Quiescent hepatocytes isolated from the liver by collagenase perfusion can be stimulated to proliferate in culture. We wished to determine if Bcl-2 overexpression would also aect the proliferative response of hepatocytes maintained in primary culture. We measured DNA replication (as determined by [3H]thymidine incorporation) in cultured hepatocytes isolated from Bcl-2 transgenic and WT mice. Hepatocytes were plated in 24 well plates and allowed to attach for 4 h. Following attachment cells were maintained in culture media containing either 5% (Figure 2a) or 10% (Figure 2b) fetal bovine serum (FBS). After 27 h

Figure 1 Delayed DNA synthesis following partial hepatectomy in Bcl-2 transgenic mice. (a) DNA synthesis in WT and Bcl-2 transgenic mice after PH as measured by BrdU incorporation. Surgeries were performed between 9:00 and 11:00 a.m. and animals were killed between 0 and 48 h. Two hours prior to killing, mice were injected with BrdU (50 mg/gr body weight). Tissue was ®xed in Methacarne, paran embedded and sections cut for IHC. BrdU IHC was performed and DNA replication was determined by counting BrdU positive hepatocytes. Four to Seven animals per time point were analysed and a minimum of 3000 cells per animal was counted. (b) Bcl-2 Western blot. Two WT and 2 Bcl-2 transgenic mice were examined at each time point at 0, 28 and 44 h after PH to verify Bcl-2 transgene expression. Error bars represent the s.e.m

[3H]thymidine was added to the culture media and cells were harvested thereafter at 30, 38, 46, 51, 66 and 90 h and the amount of [3H]thymidine incorporated into the DNA measured. When maintained in either 5% (Figure 2a) or 10% (Figure 2b) FBS-containing media, the hepatocytes from WT mice began to proliferate approximately 1 day before than those from Bcl-2 transgenic mice and the overall amount of [3H]thymidine incorporation was approximately 50% greater in WT hepatocytes at all times up to 99 h. Cyclin E and cyclin D expression and associated kinase activities following PH Because DNA replication was delayed in the regenerating livers of Bcl-2 transgenic mice, we investigated whether this delay was associated with a slower increase in the expression of cyclins D and E after Oncogene


1550

Figure 2 Delayed DNA replication in primary hepatocytes isolated from Bcl-2 transgenic and WT mice. Hepatocytes from Bcl-2 transgenic and WT mice were isolated following collagenase perfusion of livers. (a) After plating, hepatocytes were maintained in either 5% (a) or 10% (b) FBS. DNA replication was measured by [3H]thymidine incorporation. Error bars represent the s.e.m

PH. The liver samples selected for analysis were from WT and Bcl-2 transgenic mice in which the BrdU labeling indices were closest to the average labeling index for that sampling time. As determined by Western blot analysis, no dierences were detected between WT and Bcl-2 transgenic mice in the timing of cyclin D expression (Figure 3a,b). Cyclin D was increased in both WT and Bcl-2 transgenic mice 12 h after PH and the expression was sustained throughout the time course of the study (up to 48 h). The levels of cyclin D were actually higher in the Bcl-2 transgenic mice than in the WT mice. The same membranes used to detect cyclin D were then examined for cyclin E expression. Cyclin E was barely detectable in normal livers of WT and Bcl-2 transgenic mice (Figure 4a,b). In WT mice strong expression of the protein occurred at 28 h diminishing at 36 and 44 h. In contrast, cyclin E was only slightly elevated at 28 and 32 h in Bcl-2 transgenic mice and Oncogene

Figure 3 Eect of Bcl-2 overexpression on cyclin D expression following PH. Livers were harvested at 0 ± 44 h after PH. Two animals per time point were examined. (a) Expression of cyclin D as determined by Western blot analysis in WT and Bcl-2 transgenic mice following PH. (b) Quanti®cation of Western blots by densitometry. ADU, arbitrary densitometric units

strong induction did not occur until 36 h. In both the WT and the Bcl-2 transgenic mice, cyclin E induction correlated with entry into S phase. In summary, these results show that overexpression of Bcl-2 in the livers of transgenic mice delayed cyclin E but not cyclin D protein induction following PH. Although levels of cyclin E are low at 28 ± 32 h after PH in Bcl-2 transgenic mice there may be sucient protein to form active cyclinE/cdk2 complexes. We examined cyclin E/cdk2 kinase activity in Bcl-2 transgenic and WT mice 12, 28, 32 and 36 h after PH. For the kinase assay cyclin E was immunoprecipitated from whole cell protein lysates, incubated with 32 p-gATP and Histone H1 for the substrate. Cyclin E/ cdk 2 kinase activity after PH was delayed in Bcl-2 transgenic mice (Figure 5a,b). The histone H1 substrate was only minimally phosphorylated 12 h after PH in both WT and Bcl-2 transgenic mice. At 28 h histone H1 was strongly phosphorylated in the immunoprecipitates from WT but not Bcl-2 transgenic mice. Not until 36 h after PH was cyclin E-associated kinase activity elevated in the Bcl-2 transgenic mice.


1551

Figure 4 Eect of Bcl-2 overexpression on cyclin E expression following PH. Livers were harvested at 0 ± 44 h after PH. Two animals per time point were examined. (a) Expression of cyclin E as determined by Western blot analysis in WT mice and Bcl-2 transgenic mice after PH. (b) Quanti®cation of Western blots by densitometry. ADU, arbitrary densitometric units

Overexpression of Bcl-2 can maintain or increase p27 levels, thereby inhibiting cyclin E/Cdk2 kinase activity (Vairo et al., 2000). However, p27 levels estimated by Western blot analysis in total tissue lysates were high and showed no variation after PH in both WT and Bcl-2 transgenic mice (data not shown). These data indicate that the alterations of cyclin E-associated kinase activity in Bcl-2 transgenic mice is most likely due to the low levels of cyclin E protein, rather than p27 inhibition. Expression of other cell cycle genes in Bcl-2 transgenic mice after PH We next examined the expression of other cell cycle proteins including p130, p107, PCNA and cyclin A. Normal (non-hepatectomized) livers of both WT and Bcl-2 transgenic mice expressed high levels of p130. Expression remained high up to 36 h after PH and diminished slightly there after regardless of genotype (Figure 6). Unlike p130, p107 is absent in quiescent cells and expressed as cells are preparing to enter S phase

Figure 5 Cyclin E/cdk 2 kinase activity. Livers were harvested 12 ± 36 h after PH. Total liver protein lysates (500 mg) were immunopreciptitated for cyclin E and assayed for kinase activity. The percentage of BrdU positive cells in each animal is noted. (a) Cyclin E associated kinase activity in WT (Bcl-2 7) and Bcl-2 (Bcl-2+) transgenic mice. (b) Quanti®cation of kinase activity by densitometry. ADU, arbitrary densitometric units

(Garriga et al., 1998; Nevins, 1998) Consistent with these observations, p107 was not detected until 28 h after PH in WT mice and remained high thereafter (Figure 6). In contrast, the increase in p107 expression after PH was delayed in Bcl-2 transgenic mice and expression remained low until 36 h after PH. We also analysed p107 expression by immunohistochemisitry (Figure 7) to evaluate the timing of nuclear expression in hepatocytes after PH. In both WT and Bcl-2 livers p107 was detected in hepatocyte nuclei. However, there was a major dierence in the timing of p107 expression after PH between WT and Bcl-2 transgenic mice. Consistent with the Western blot data, p107 expression Oncogene


1552

Figure 6 Western blot analysis of cell cycle proteins following PH. p107, cyclin A and PCNA expression is delayed in the Bcl-2 transgenics when compared to wild-type control littermates. Total liver protein (25 mg) was loaded per lane on a 4 ± 15% gradient gel

progressively increased from 28 to 36 h after PH and was predominantly localized to nuclei of periportal hepatocytes of WT mice (Figure 7a,c,e). In contrast p107 was barely detectable in hepatocytes of Bcl-2 transgenics at 28 and 32 h (Figure 7b,d) and strong nuclear staining was not detected until 36 h after PH (Figure 7f). In addition to p107, hepatic expression of PCNA and cyclin A were delayed in Bcl-2 transgenic mice following PH (Figure 6). PCNA was present at low levels in liver extracts from WT mice until 28 h after PH, and progressively increased thereafter. In contrast, induction of PCNA did not occur until 36 h after PH in liver extracts from Bcl-2 transgenic mice. Cyclin A was ®rst detected at 32 h after PH in liver extracts from WT mice but not until 36 h after PH in Bcl-2 transgenic mice (Figure 6). These data indicate that in addition to cyclin E, overexpression of Bcl-2 results in delayed expression of p107, PCNA and cyclin A in the liver after PH. Discussion Hepatocytes are quiescent cells that are capable of proliferation in response to loss of liver mass. Liver regeneration after PH is an excellent model to study hepatocyte cell cycle progression in vivo. We now show that Bcl-2 interferes with DNA replication after PH. Bcl-2 transgenic mice showed a delay in DNA replication after PH, a defect associated with delays Oncogene

in increases of PCNA, cyclin E, cyclin E-associated kinase activity, cyclin A and p107. The G1 cyclins, cyclin D and cyclin E are the rate limiting cyclins necessary for the transition from G0/ G1 into S phase. Cyclin E expression and, as a result, the cyclin-associated kinase activity, were delayed in the Bcl-2 transgenic mice following PH. Despite the delay in DNA synthesis in Bcl-2 transgenic mice, the timing of cyclin D expression was unaected by Bcl-2 overexpression. Furthermore, the levels of cyclin D were comparable, if not higher in Bcl-2 transgenic mice than those found in the WT mice. This is in agreement with a study by Lin et al. (2001) which showed that overexpression of Bcl-2 correlated with increased levels of cyclin D. As the cyclin D antibody used in our experiments recognizes all D-type cyclins we cannot say for certain if just cyclin D1 is elevated or if cyclins D2 and D3 are also increased in the Bcl-2 transgenic mice. However, based upon the expression patterns of both cyclin D and cyclin E, our results suggest that Bcl-2 lengthens the G1 phase of the cell cycle in hepatocytes by delaying the expression of cyclin E. It should be noted that the target for Bcl-2 eects on the hepatocyte cell cycle may be upstream from cyclins. In this case, the delay in cyclin expression would be an indirect consequence of cell cycle inhibition. Overexpression of cyclin E and skp2, a ubiquitin ligase required for p27 degradation, can induce hepatocyte proliferation in the absence of cyclin D/ cdk4 kinase activity (Nelsen et al., 2001) both in vivo and in vitro. This suggests that the cyclin E/cdk2 complex is a primary driver of hepatocyte proliferation, an observation supported by our results. Thus, even though cyclin D is expressed at comparable levels in both WT and Bcl-2 transgenic mice, inhibition of cyclin E expression is sucient to delay entry into DNA synthesis in hepatocytes. The p130/E2F complex is a negative regulator of p107 gene expression with phosphorylation and subsequent degradation of p130 leading to E2Fmediated transcription of p107 (Classon and Dyson, 2001). Vairo et al. (2000) demonstrated that overexpression of Bcl-2 delayed phosphorylation and led to increased levels of p130 in ®broblasts stimulated to proliferate. Their results also showed that Bcl-2 increased p130 and reduced p107 expression in late G1. Inhibition of cell cycle progression by Bcl-2 required the presence of p27 and p130 but not of pRb. Our studies in the regenerating liver in vivo showed that in WT mice levels of both p27 and p130 remained unchanged, but p107 expression increased starting at approximately 28 h after PH, a time at which p130 hyperphosphorylation is detected (data not shown). Analysis of the timing of p107 and cyclin E activation in Bcl-2 transgenic mice suggests that Bcl-2 may slow down cell cycle progression of hepatocytes during liver regeneration by inhibiting the formation of a cyclin E-p107-E2F complex which accumulates in late G1 (Nevins, 1998). Huang et al. (1997) have suggested that the N-terminal region of Bcl-2 adjacent to tyrosine 28 is an important binding site for Bcl-2 cell


1553

Figure 7 Expression of p107 in WT (a,c,e) and Bcl-2 (b,d,f) transgenic mice as analysed by immunohistochemistry. Liver sections were harvested at 28 (a,b), 32 (c,d) and 36 (e,f) h following PH. Note number of p107 positive nuclei in WT but not Bcl-2 transgenic mice at 28 and 32 h

cycle inhibitory eects. Vairo et al. (2000) further suggested that the N-terminal region binds to a still unidenti®ed protein and regulates p130 and p27 levels by post-transcriptional mechanisms. We do not know at this time whether Bcl-2 N-terminal region binding to cell components also occurs in regenerating livers. The total number of proliferating hepatocytes in both WT and Bcl-2 transgenic mice were comparable. However, the Bcl-2 transgenic mice had a higher `peak' of DNA synthesis than the WT mice. The most likely explanation is that the block imposed by Bcl-2 provided a greater level of hepatocyte synchronization through the cell cycle. It is of interest to note that BclXL, an anti-apoptotic gene of the Bcl-2 family, is expressed as a delayed-early response gene between 4 ± 12 h after PH (Tzung et al., 1997). Although we have assumed that early Bcl-XL expression has an antiapoptotic function in the regenerating liver, it is possible that it may act to delay and synchronize hepatocyte cell cycle progression through the cell cycle. Our previous work demonstrated that Bcl-2 overexpression delays the development and reduces the frequency of TGFa - induced liver tumors in transgenic

mice (Vail et al., 2001). Inhibition of liver tumorigenesis occurred even when Bcl-2 expression was turned on at 8 months of age. Other studies showed that Bcl-2 expression inhibits liver tumor development in c-myc transgenic mice (de La Coste et al., 1999) as well as in mice injected with the carcinogen diethylnitrosamine (M Vail and N Fausto, in preparation). In many human cancers Bcl-2 expression correlates with a better prognosis or low tumor proliferative activity (Kuwashima et al., 1997; Silvestrini et al., 1996; Sinicrope et al., 1995; Winter et al., 1998). In a mouse model of mammary carcinogenesis, Bcl-2 inhibited both cell proliferation and apoptosis at the early stages of tumorigenesis but at the later stages, only the antiapoptotic activity persisted (Furth et al., 1999). Mammary gland speci®c Bcl-2 expression in transgenic mice treated with dimethylbenz(a)anthracene also delayed mammary tumor development (Murphy et al., 1999) The data presented here demonstrates that Bcl-2 overexpression extends the length of time that hepatocytes stay in the G0/G1 phase of the cell cycle when stimulated to proliferate. These ®ndings support Oncogene


1554

the hypothesis that the delay in hepatocarcinogenesis in Bcl-2 transgenic mice is predominantly due to a direct eect of Bcl-2 on hepatocyte cell cycle progression. We suggest that Bcl-2 may have a similar eect in tumor development in other tissues.

Materials and methods Transgenic mice Male 8 to 10 week-old-heterozygous Bcl-2 transgenic mice and wild-type littermates were utilized for all experiments in this study. The genetic background of the mice (CD16C57B6C3H) is the same as that previously described (Vail et al., 2001). The Bcl-2 transgene is under the control of the mouse metallothionein promoter, which can be induced by the addition of 25 mM ZnSO4 to the drinking water (zinc water). Mice were maintained on zinc water for a minimum of 1 week prior to surgery. Genotypes were determined by PCR ampli®cation of genomic DNA obtained from mouse tails. Mice were maintained in speci®c-pathogen-free housing and cared for in accordance with NIH guidelines for animal care.

Immunohistochemistry

Partial hepatectomies were performed between 9 a.m. and noon to minimize diurnal eects. Surgical procedures as described previously (Webber et al., 1994) were followed resulting in removal of approximately 2/3 of the liver mass. Animals were killed 0, 12, 28, 32, 36, 40, 44 and 48 h postoperatively. Between 4 and 7 Bcl-2 transgenic and wildtype mice each were used per time point and were injected with 50 mg/g BrdU 2 h prior to death. At the time of death, the remaining liver was removed and portions ®xed in 10% buered formalin and methacarne for immunohistochemistry or snap frozen in liquid nitrogen for later protein isolation.

Portions of livers were ®xed in either neutral buered formalin for 24 h for routine histology and immunohistochemistry (IHC) or in methacarne (60% methanol, 30% chloroform and 10% acetic acid) for 1 h and transferred to 100% methanol for 24 h for bromo-deoxyuridine (BrdU) IHC. After ®xation, tissue was embedded in paran. For routine histological analysis, 5-mm sections were cut from paran-embedded blocks and stained with hematoxylin and eosin. For p107 IHC, sections were deparanized, rehydrated, and treated for 30 min with 0.03% H2O2 to inactivate endogenous peroxidase activity. Sections were microwaved for 10 min in 10 mM NaCitrate buer, pH 6.0 prior to incubation in primary antibody (rabbit anti-p107, Santa Cruz Biotechnology) diluted 1 : 500 in phosphate buered saline (PBS) containing 5% horse serum. Staining was detected using the Vectastain elite ABC kit (Vector Laboratories, Burlingame, CA, USA). BrdU IHC was performed as described above with the following modi®cations: Trypsin digestion (slides were incubated in 1 mg/ml trypsin for 10 min at room temperature) was used in place of microwave antigen retrieval, followed by incubation in 2.5 M HCl for 10 min at 378C. Mouse anti-BrdU diluted 1 : 40 was used as the primary antibody (Dako, Carpinteria, CA, USA).

Primary hepatocyte cultures

Kinase assays

Hepatocytes were isolated from 8-10-week-old WT and Bcl-2 transgenic mice by collagenase perfusion of the liver, as described in (Wu et al., 1994). Hepatoctyes were plated (2.56104 cells per well) in 24 well plates and allowed to attach for 4 h in attachment media (Dulbeco's modi®ed Eagle's medium/Ham's F-12) supplemented with a mixture of insulin, transferrin, and selenium (ITS), 0.1 mM dexamethasone and gentamicin at 50 mg/ml. After attachment, media was replaced with attachment media containing either 5 or 10% fetal bovine serum (FBS) and 27 h later [3H]thymidine (5 m Ci-NEN) was added to each well. Incorporation of [3H]thymidine was assayed by precipitation of nucleic acids with TCA.

Portions of snap-frozen liver were homogenized in IP lysis buer (50 mM HEPES, pH 7.5, 150 mM NaCl, 1 mM EDTA, 2.5 mM EGTA, 1 mM DTT, 0.1% Tween-20) containing 10% glycerol and protease inhibitors (0.1 mM PMFS, 10 mg/ml leupeptin, 20 U/ml aprotinin, 10 mM b-glycerophosphate, 1 mM NaF, 0.1 mM sodium orthovanadate) and brie¯y sonicated. Lysates (500 mg) were immunoprecipitated with cyclin E (2 mg/ml; same as listed above for the Western blots) and protein A/G plus agarose beads (Santa Cruz). Following immunoprecipitation (IP), beads were washed four times with IP lysis buer and twice with histone wash buer (25 mM Tris HCL pH 7.5, 70 mM NaCl, 10 mM MgCl2, 1 mM DTT). The cyclin E IPs were incubated at 308 in a 30-ml reaction mix containing histone wash buer and 0.1 mg/ml histone H1, 10 mM ATP and 0.2 mCi/ml 32P-gATP (6000 Ci/mmol. NEN, DuPont, Wilmington, DE, USA). The entire reaction mix was separated on 10% SDS ± PAGE gels and analysed by autoradiography of the dried gel. Samples of lysates for the assays were normalized based on total cellular protein content.

Partial hepatectomies

Western blot analysis Total liver protein was isolated from snap-frozen tissue by homogenization in RIPA buer (1.0% NP40, 0.5% sodium deoxycholate and 0.1% SDS containing the following protease inhibitors: 1 mM dithiothreitol (DTT), 0.5 mM paminoethylbenzenesulfonyl ¯uoride. Two mg/ml aprotinin, 2 mg/ml pepstatin, 2 mg/ml leupeptin and 10 mg/ml soybean trypsin inhibitor. Protein concentrations were determined using the Bradford protein assay (Bio-Rad, Hercules, CA, USA). Total protein (25 mg) was separated on 12% SDS ± PAGE gels for the cyclin D, cyclin E and Bcl-2 Western Oncogene

blots, and on 4 ± 15% tris HCL gradient gels (Bio-Rad) for the p130, p107, PCNA, cyclin A and a,-tubulin blots. Proteins were transferred to nylon membranes, and blocked in 5% milk containing 0.1% Tween-20 prior to incubation with antibody. Antibodies against Bcl-2 (Santa Cruz #sc492, 1 : 2000 dilution), p107 (Santa Cruz #sc-318, 1 : 1000 dilution), p130 (Santa Cruz #sc-317, 1 : 1000 dilution), cyclin E (Santa Cruz #sc-198, 1 : 500 dilution), Cyclin A (Santa Cruz #sc-596, 1 : 500 dilution), PCNA (Santa Cruz #¯-26,1 1 : 1000 dilution), and cyclin D (Pharmigen, 1 : 1000 dilution) were used. Enhanced chemiluminescence (ECL; Santa Cruz Biotechnology) was used for detection.

Acknowledgments We thank Gretchen Argast for additional help with partial hepatectomies; John Brooling for technical assistance with


the animals and members of the Fausto lab for helpful discussions. This work was supported by grant CA74131 from the National Cancer Institute (NCI). Further support

was provided to M Vail and M Chaisson by PHS National Research Grant T32 GMO7270 and to M Chaisson by National Research Service Award CA 09437.

1555

References Borner C. (1996). J. Biol. Chem., 271, 12695 ± 12698. Classon M and Dyson N. (2001). Exp. Cell Res., 264, 135 ± 147. de La Coste A, Mignon A, Fabre M, Gilbert E, Porteu A, Van Dyke T, Kahn A and Perret C. (1999). Cancer Res., 59, 5017 ± 5022. Dyson N. (1998). Genes Dev., 12, 2245 ± 2262. Fausto N. (2000). J. Hepatol., 32, 19 ± 31. Furth PA, Bar-Peled U, Li M, Lewis A, Laucirica R, Jager R, Weiher H and Russell RG. (1999). Oncogene, 18, 6589 ± 6596. Garriga J, Limon A, Mayol X, Rane SG, Albrecht JH, Reddy EP, Andres V and Grana X. (1998). Biochem. J., 333, 645 ± 654. Huang DC, O'Reilly LA, Strasser A and Cory S. (1997). EMBO J., 16, 4628 ± 4638. Korsmeyer SJ. (1992). Cancer Surv., 15, 105 ± 118. Kuwashima Y, Kobayashi Y, Kurosumi M, Tanuma J, Shiromizu K and Kishi K. (1997). Anticancer Res., 17, 3773 ± 3776. Lin HM, Lee YJ, Li G, Pestell RG and Kim HR. (2001). Cell Death Dier, 8, 44 ± 50. Linette GP, Li Y, Roth K and Korsmeyer SJ. (1996). Proc. Natl. Acad. Sci. USA, 93, 9545 ± 9552. Mazel S, Burtrum D and Petrie HT. (1996). J. Exp. Med., 183, 2219 ± 2226. Murphy KL, Kittrell FS, Gay JP, Jager R, Medina D and Rosen JM. (1999). Oncogene, 18, 6597 ± 6604. Nelsen CJ, Hansen LK, Rickheim DG, Chen C, Stanley MW, Krek W and Albrecht JH. (2001). Oncogene, 20, 1825 ± 1831.

Nevins JR. (1998). Cell Growth Dier, 9, 585 ± 593. O'Reilly LA, Harris AW and Strasser A. (1997). Int. Immunol., 9, 1291 ± 1301. O'Reilly LA, Huang DC and Strasser A. (1996). EMBO J., 15, 6979 ± 6990. Sherr CJ and Roberts JM. (1999). Genes Dev., 13, 1501 ± 1512. Silvestrini R, Benini E, Veneroni S, Daidone MG, Tomasic G, Squicciarini P and Salvadori B. (1996). J. Clin. Oncol., 14, 1604 ± 1610. Sinicrope FA, Hart J, Michelassi F and Lee JJ. (1995). Clin. Cancer Res., 1, 1103 ± 1110. Truchet I, Jozan S, Guerrin M, Mazzolini L, Vidal S and Valette A. (2000). Exp. Cell Res., 254, 241 ± 248. Tzung SP, Fausto N and Hockenbery DM. (1997). Am. J. Pathol., 150, 1985 ± 1995. Vail ME, Pierce RH and Fausto N. (2001). Cancer Res., 61, 594 ± 601. Vairo G, Innes KM and Adams JM. (1996). Oncogene, 13, 1511 ± 1519. Vairo G, Soos TJ, Upton TM, Zalvide J, DeCaprio JA, Ewen ME, Ko A and Adams JM. (2000). Mol. Cell Biol., 20, 4745 ± 4753. Webber EM, Wu JC, Wang L, Merlino G and Fausto N. (1994). Am. J. Pathol., 145, 398 ± 408. Winter JN, Andersen J, Reed JC, Krajewski S, Variakojis D, Bauer KD, Fisher RI, Gordon LI, Oken MM, Jiang S, Jeries D and Domer P. (1998). Blood, 91, 1391 ± 1398. Wu JC, Merlino G and Fausto N. (1994). Proc. Natl. Acad. Sci. USA, 91, 674 ± 678.

Oncogene