Critical role of cyclin D3 in TSH-dependent growth of thyrocytes and in hyperproliferative diseases of the thyroid gland

Transcription

1 (2003) 22, & 2003 Nature Publishing Group All rights reserved /03 $ Critical role of cyclin D3 in TSH-dependent growth of thyrocytes and in hyperproliferative diseases of the thyroid gland Maria Letizia Motti 1, Angelo Boccia 1, Barbara Belletti 2, Paola Bruni 3, Giancarlo Troncone 4, Letizia Cito 1, Mario Monaco 2, Gennaro Chiappetta 2, Gustavo Baldassarre 1, Lucio Palombini 4, Alfredo Fusco 1 and Giuseppe Viglietto*,5 1 Dipartimento di Biologia e Patologia Cellulare e Molecolare L.Califano, Università Federico II, via S Pansini 5, Naples, Italy; 2 Servizio Oncologia Sperimentale E, Istituto Nazionale Tumori, via M Semmola, Naples, Italy; 3 Dipartimento di Biochimica e Biotecnologie Mediche, Università Federico II, via S Pansini 5, Naples, Italy; 4 Dipartimento di Scienze Morfologiche e Funzionali, Facoltà di Medicina e Chirurgia, Università Federico II, via S Pansini 5, Naples, Italy; 5 Istituto di Endocrinologia ed Oncologia Sperimentale del CNR c/o Dipartimento di Biologia e Patologia Cellulare e Molecolare L.Califano, Università Federico II, via S Pansini 5, Naples, Italy We report that cyclin D3 is rate limiting for G1 progression in thyroid follicular cells and that its constitutive upregulation by chronic stimulation of the TSH/cAMP pathway plays a role in human and experimental hyperproliferative diseases of the thyroid gland. These conclusions are supported by in vitro and in vivo studies. In rat thyrocytes (PC Cl 3 cells), cyclin D3 expression is enhanced in response to activation of the TSH/cAMP pathway. Interference with the expression of G1 cyclins (in particular cyclin D3) by the antisense methodology strongly reduced TSH-dependent proliferation of PC Cl 3 cells, indicating that proper progression through G1 requires cyclin D3. Accordingly, PC Cl 3 cells engineered to overexpress cyclin D3 (PC-D3 cells) show enhanced growth rate and elude hormone-dependence and contact inhibition. Using an animal experimental model of thyroid stimulation, we demonstrate that cyclin D3 is a key mediator of TSH-dependent proliferation of thyroid follicular cells also in vivo. Cyclin D3 protein levels were higher in the thyrocytes from glands of propylthiouraciltreated rats compared with control animals. The increase in cyclin D3 expression occurred after the propylthiouracil-induced increase in TSH levels and preceded the burst of cell proliferation. Finally, we found that cyclin D3 protein is expressed in a fraction of human goiters but it is strongly overexpressed in most follicular adenomas. (2003) 22, doi: /sj.onc Keywords: cyclin D3; thyroid; TSH; p27 kip1 Introduction Thyrocyte proliferation in vivo is essentially regulated by a single growth factor, that is thyroid-stimulating hormone (TSH) (Wilson and Foster, 1989). This *Correspondence: G Viglietto; viglietto@sun.ceos.na.cnr.it Received 24 January 2003; revised 22 July 2003; accepted 31 July 2003 glycoprotein hormone, which is secreted by the pituitary as a part of a negative feedback loop, contributes to the manteinance of normal systemic levels of thyroid hormones. The TSH receptor (TSHr) is a seventransmembrane, G protein-coupled receptor expressed mainly in thyroid cells (Davies et al., 2002). Activation of TSHr leads to stimulation of the cyclic AMP (camp) cascade that promotes activation of the camp-dependent protein kinase A (PKA) (Dremier et al., 2002). The key events in the complex signaling cascade(s) of TSH are poorly understood. However, these cascades are thought to modulate the level and activity of proteins that regulate the cell cycle machinery (Morgan, 1995; Ekholmand Reed, 2000). According to the current model, progression through G1 is determined by the sequential phosphorylation of the retinoblastoma susceptibility gene product, p105 Rb, whose phosphorylation by the cyclin D-Cdk4/6 and E-Cdk2 complexes, activates the E2F family of transcription factors (Weinberg, 1995). Studies of the TSH-dependent regulation of cell cycle progression suggested that cyclin D3 is a critical regulator of the thyroid follicular cell cycle (Kimura et al., 2001; Medina and Santisteban, 2000). Cyclin D3 is the predominant D-type cyclin in cultured thyrocytes of dogs and humans, and in the mouse thyroid in vivo (Coppee et al., 1998; Baldassarre et al., 1999; Kimura et al., 2001). Furthermore, inhibition of cyclin D3 function induced by neutralizing antibodies suppressed TSH-dependent proliferation of dog thyrocytes (Depoortere et al., 1998; Wollman et al., 1978). Conversely, in another rat cell line (the FRTL-5 cells), TSH-dependent proliferation occurs both through upregulation of D-type cyclins and downregulation of the cyclin-dependent kinase inhibitor p27 kip1 (Yamamoto et al., 1996; Carneiro et al., 1998; Medina et al., 1999). Because of its central role in thyroid cell function and proliferation, it is feasible that inappropriate stimulation of the camp cascade could lead to various thyroid diseases (Corvilain et al., 2001). The proliferation of thyroid cells is high during embryonal development but becomes negligible in the normal adult thyroid gland

2 (Dumont et al., 1992). However, thyroid follicular cells retain the capacity for rapid, self-limiting growth (Dumont et al., 1992). Most abnormalities in the control of thyroid growth result in thyroid enlargement (goiter) (Dumont et al., 1992; Hedinger et al., 1993). Goiter is invariably observed in patients with Graves disease, hyperfunctioning adenomas and adenomas caused by congenital defects. Nontoxic goiters are essentially diffuse in the early phase. However, diffuse goiters continue to grow and may become nodular, and finally autonomous. In the case of simple goiter, hyperplasia of follicular cells is explained by the classic concepts of thyroid regulation; that is, goiter is due to increased TSH secretion, which occurs as a response to congenital defects in iodine metabolism or dietary iodine insufficiency (Dumont et al., 1992). In patients with Graves disease, the development of stimulating autoantibodies against the TSHr (TSAb), results in activation of the receptor, thus stimulating the camp/pka pathway and the proliferation of thyroid cells (Dumont et al., 1992). Conversely, thyroid adenomas are well defined, encapsulated benign tumors characterized by TSHindependent function and growth (Dumont et al., 1992; Hedinger et al., 1993; Corvilain et al., 2001). Adenomas are characterized by activating mutations in the genes encoding the TSH receptor or the G protein a subunit (G as ), thereby resulting in the constitutive activation of the camp/pka pathway and TSHindependent clonal proliferation of mutated thyrocytes (Dumont et al., 1992; Corvilain et al., 2001). Mutations in the TSHr are also found in hyperfunctioning nodules frompatients with a multinodular goiter (Corvilain et al., 2001). Within this framework, we investigated the role of cyclin D3 in the regulation of thyroid proliferation and asked whether cyclin D3 is implicated in the molecular pathogenesis of thyroid hyperplastic diseases. Our results demonstrate that dysregulation of cyclin D3 expression plays a key role in the development of various human thyroid diseases. Results Effects of TSH on cell cycle regulators in rat thyrocytes in vitro To investigate the molecular mechanisms whereby TSH supports thyroid cell proliferation, we used the PC Cl 3 cultured rat thyroid cell line (Fusco et al., 1987; Berlingieri et al., 1990), which is ideal for studies of the mitogenic effect of TSH and camp (Kimura et al., 1999). In fact, rat thyroid PC Cl 3 cells are routinely maintained in 5% calf serum with a mixture of six hormones, of which only TSH and insulin are required for proliferation (see Materials and methods) (Fusco et al., 1987; Medina and Santisteban, 2000; Kimura et al., 2001). Although TSH alone does not induce DNA synthesis in this cell line (Florio et al., 2001), TSH associated with insulin or a low serumconcentration is mitogenic for PC Cl 3 cells (Kimura et al., 1999). Thus, we investigated the effects exerted by TSH on the expression and activity of cell cycle regulators. Cell extracts were prepared frompc Cl 3 cells starved in F12 medium supplemented with 5% calf serum and subsequently exposed to 10 mu/ml of TSH for 6, 12, 24 and 48 h in the presence of serum. Cell proliferation was measured with flow cytometry, and the expression of cyclins, Cdks and Cdk inhibitors with the Western blot technique. The results are shown in Figure 1. Thyroidstimulating hormone induced the expression of cyclin D1, cyclin D2 and especially of cyclin D3 after 6 h of treatment. The expression of cyclin E and A was slightly increased at h (Figure 1a). The expression of the Cdk inhibitors p21 cip1 and p27 kip1 was essentially unchanged. Similarly, p57 kip2 was expressed at very low levels after TSH stimulation (Figure 1b). Finally, TSH slightly increased Cdk2, Cdk4 and Cdk6 expression (Figure 1c). In parallel, cells were stained with propidiumiodide and analysed by flow cytometry with a FACScan flow cytometer (Becton Dickinson, San Jose, CA, USA) interfaced with Helwett-Packard computer (Palo Alto, CA, USA). Cell cycle data analysis was performed by CELL-FIT program(becton Dickinson). A representative experiment of two is reported in Figure 1d. Typically, PC Cl 3 cells remained in G0 when starved in 5% calf serumwithout the six hormones for 3 days (G1, 89%; S, 7% and G2/M, 4%) (Fusco et al, 1987; Florio et al., 2001). When stimulated to re-enter the cell cycle by the addition of TSH in the presence of serum, the fraction of cells in the G1 compartment decreased a TSH hours : cyc D1 cyc D2 cyc D3 cyc E cyc A b c p57 p21 p27 CDK2 CDK4 CDK6 d Events Events TSH : 0 TSH : 12 Events DNA Content DNA Content TSH : 48 TSH : S 19 S 16 G2/M 14 G2/M 9 Events DNA Content 89 G2/M S 7 4 DNA Content 85 G2/M S 8 7 Figure 1 Regulation of cyclin D3 by TSH in cultured thyroid cells. Western blot analysis of the effects of TSH on the expression of cyclins (a), and Cdk inhibitors (b) and Cdks (c) in PC Cl 3 cells. Lane 1:PC Cl 3 cells starved in F12 medium supplemented with 5% calf serum(0 h); lanes 2 5:PC Cl 3 cells treated with TSH for the indicated times in 5% calf serum. In each experiments the filters were stained with Red Ponceau to ensure uniformprotein loading and integrity. Cyclin D3 protein levels were quantified by densitometric analysis of the films and expressed as arbitrary units. The ratios of cyclin D3 protein in TSH-treated PC Cl 3 cells respect to starved PC Cl 3 cells (set 1 at 0 h) was 2-, and 4.4-fold after 6, 12, 24 and 48 h of TSH treatment, respectively. (d) Flow cytometry analysis of the effects exerted by TSH in PC Cl 3 cells. Cells were starved in F12 medium supplemented with 5% calf serumand treated with TSH for the indicated time 7577

3 7578 (67 and 75% at 24 and 48 h, respectively) and the fraction of cells in the S compartment increased (19 and 16% at 24 and 48 h, respectively). In conclusion, it appears that the pronounced TSH-dependent increase in cyclin D3 expression in PC Cl 3 cells is the earliest and most relevant effect exerted by TSH on cell cycle regulators, and that the increase in cyclin D3 expression precedes the entry of cells into S phase. Regulation of cyclin D3 by TSH in PC CL 3 cells TSH modulates thyroid functions by binding a highaffinity receptor of the seven-spanned transmembrane domains (TSHR) on thyrocyte cell surface, whose activation increases intracellular camp levels, which in turn activates the protein kinase A (Dumont et al., 1992). To investigate whether TSH-regulated induction of cyclin D3 expression could be mediated by the camp/pka, we evaluated whether cyclin D3 expression was induced by treatment with forskolin, a compound that increases the intracellular level of camp. As shown in Figure 2, treatment of PC Cl3 cells with forskolin, stimulated cyclin D3 protein expression in a manner similar to TSH, suggesting camp mediation in the induction of cyclin D3 protein expression exerted by TSH. Similar results were obtained with the membranepermeable camp agonist, adenosine-3 0,5 0 -cyclic monophosphorothioate Sp-isomer (Sp-cAMP) (data not shown). Subsequently, we investigated the mechanism whereby TSH upregulated cyclin D3 protein expression in PC Cl 3 cells. RNA prepared fromuntreated PC Cl 3 cell cultures contained moderate levels of cyclin D3 transcript. Treatment of cells with 10 mu/ml of TSH did not result in a significant increase in the level of cyclin D3 RNA (not shown), suggesting that in PC Cl 3 cells, TSH may regulate cyclin D3 expression at the protein level. Accordingly, treatment of TSH-starved PC Cl 3 cells with two highly specific proteasome inhibitors (the peptide aldheyde N-acetyl-leucyl-leucine norleucinal or LLnL and the inhibitor MG132) increased the level of cyclin D3 protein. Treatment of PC Cl 3 cells for 16 h with 50 mm of LLnL or 20 mm of MG132 resulted in 2.5- fold increase in the cyclin D3 level (Figure 2b, lane 2) compared to DMSO-treated cells (Figure 2b, lane 1), suggesting that in growth-arrested PC Cl 3 cells the proteasome-dependent pathway takes part in cyclin D3 turnover. Similar results were obtained with MG132 (not shown). As a complementary approach, we evaluated the role of protein synthesis and stability in the determination of cyclin D3 abundance in PC Cl 3 cells. To inhibit protein synthesis, cells were incubated with 10 mm of the ribosomal complex inhibitor cycloheximide, a dose sufficient to block protein synthesis by >98%. PC Cl 3 cells were treated with TSH for 24 h either in the presence or in the absence of cycloheximide to block protein synthesis, and subsequently cyclin D3 expression Figure 2 In vitro regulation of cyclin D3 by TSH. Analysis of TSH- and forskolin-dependent regulation of cyclin D3 protein expression in PC Cl 3 cells. (a) Treatment of PC Cl 3 cells with 10 mu/ml of TSH or 10 mm forskolin (lane 1:untreated PC Cl 3 cells; lanes 2 and 4, TSH treatment for the indicated time; lanes 3 and 5, forskolin treatment for the indicated time). (b) Treatment of PC Cl 3 cells with 50 mm DMSO or LLnL. (c) PC Cl 3 cells were treated with TSH for 24 h either in the presence or in the absence of cycloheximide (10 mm) to block protein synthesis, and subsequently cyclin D3 expression levels were analysed at different times (16 and 24 h). (d) PC Cl 3 cells were starved for 3 days in 5% serumand treated with serumalone or TSH for 24 h (t ¼ 0). At this point, to determine the half-life of cyclin D3 protein, the translation of cyclin D3 was blocked with 10 mm cycloheximide and cell lysates were prepared at different times (1, 2, 4 and 8 h), loaded onto SDS PAGE and probed with an anticyclin D3 antibody

4 levels were analysed at different times (1, 6 and 24 h). As shown in Figure 2c, cycloheximide pretreatment of PC Cl 3 cells almost completely abolished the synthesis of cyclin D3 after 6 h of treatment, suggesting that protein synthesis was necessary for TSH-mediated induction of cyclin D3 in PC Cl 3 cells. Furthermore, to investigate whether TSH regulated the stability of cyclin D3 protein, PC Cl 3 cells were starved for 3 days in 5% serumand treated with serumalone or TSH for 24 h (t ¼ 0). At this point, translation of cyclin D3 was blocked with 10 mm cycloheximide at different times (1, 2, 4 and 8 h). Cell lysates were prepared and equivalent amounts of proteins were loaded on SDS PAGE and probed with an anticyclin D3 antibody. As shown in Figure 2d, we clearly show that the disappearance of cyclin D3 protein is faster in 3-day starved PC Cl 3 cells (Figure 2d, upper panels) than in TSH-induced cells (Figure 2d lower panels), suggesting a faster degradation of cyclin D3 in noninduced cells. After densitometry analysis of the bands, the half-life of cyclin D3 was estimated to be about 4 h in TSH-induced cells and 2 h in 3-day starved cells. In conclusion, it appears that the regulation of cyclin D3 protein levels by TSH is dependent on increased protein synthesis and decreased protein turnover. Interference with cyclin D3 increases growth rate in PC CL 3 cells We sought to obtain more direct evidence that cyclin D3 could be a downstreameffector of the TSH/cAMP signaling system. We investigated the effects of TSH on 3-day starved quiescent PC Cl 3 cells in the presence of antisense oligodeoxynucleotides which blocked the synthesis of cyclins D1 and D3. Using two different antisense oligonucleotides against the 5 0 region of cyclin D1 (D1-AS) and cyclin D3 (D3-AS) (see Materials and methods), we were able to drastrically reduce TSHinduced cyclin D1 and cyclin D3 expression (Figure 3b, c), whereas the same dose of a control oligodeoxynucleotides, with similar base composition but random sequence, had no effect on the cellular levels of all cyclins. To determine the effect exerted by blocking the expression of G1 cyclins on PC Cl 3 cells proliferation, PC Cl 3 cells were plated onto glass coverslips, starved in medium containing 5% serum for 36 h, and then transfected with cyclin D3 or cyclin D1 antisense or control oligonucleotides (1 mm). After 36 h, cells were stimulated with TSH for 24 h, incubated with 10 mm BrdU and processed for indirect immunofluorescence. In a typical experiment (Figure 3a), only 0.8% of 3-day starved PC Cl 3 cells incorporated BrdU. Transfection of control oligonucleotides with scrambled sequence into PC Cl3 cells had no effect on the fraction of cells that incorporated BrdU. In the presence of 5% serum, TSH induced 25.9% of cells to enter S phase. However, if TSH was administered in the presence of antisense oligonucleotides to cyclin D3, only 11.7% of PC Cl 3 cells entered S phase. Antisense oligonucleotides to cyclin D1 were less efficient in inhibiting TSH-induced Figure 3 Cyclin D3 is required for TSH-dependent growth stimulation of PC Cl 3 cells. To suppress the expression of G1 cyclins, PC Cl 3 cells were plated onto glass coverslips, starved in medium containing 5% serum for 36 h, and then transfected with cyclin D3 or cyclin D1 antisense or control oligonucleotides (1 mm) with oligofectamine. After 36 h, cells were stimulated with TSH for 24 h, incubated with 10 mm BrdU and processed for indirect immunofluorescence or Western blot. (a) Graph indicates the percentage of BrdU positive cells. (b) Western blot analysis of cyclin D3 expression induced by TSH in the presence of control or anticyclin D3 antisense oligonucleotides. (c) Western blot analysis of cyclin D1 expression induced by TSH in the presence of control or anticyclin D1 antisense oligonucleotides S phase entry of PC Cl 3 cells (17% of cells were induced to entry S phase). The results deriving fromthese antisense experiments point to a preminent role of cyclin D3 among D-type cyclins, which is in agreement with the relative levels of induction of these G1 cyclins by TSH observed in Figure 1a. As expected interference with both cyclin E and A resulted in the suppression of TSH stimulation of PC Cl 3 cell growth (not shown). Cyclin D3 overexpression increases growth rate and induces hormone-independent growth in PC CL 3 cells To probe further the role played by cyclin D3 in the control of rat thyroid cell cycle progression, we transfected PC Cl 3 cells with a CMV-based plasmid that carried the cdna encoding human cyclin D3 or with the backbone vector. Transfected cells were selected in G418-containing medium and several PC- CMV-neo and PC-D3 clones were isolated and expanded for further studies. The presence of the exogenous cyclin D3 protein in transfected cells was detected by Western blot analysis (not shown). Three PC-D3 clones were selected for biological studies. Since the same data were obtained with the three clones in all our experiments, we will refer them all as PC-D3 cells. We investigated whether the constitutive expression of cyclin D3 modified specific proliferative parameters of thyroid cells including proliferation rate, hormone requirement and contact inhibition of PC Cl 3 cells. To this aim, PC Cl 3 and PC-D3 cells were seeded at low density and grown for 7 days in the presence of 7579

5 7580 hormones. Fresh medium was provided every 48 h. Cells were harvested every 2 days, and the cell number was counted with a hemacytometer. As shown in Figure 4a, the proliferation of PC-D3 cells was higher than that of parental and PC-CMV-neo cells. The average population doubling time calculated for PC-D3 cells from growth curve experiments was less than 24 h, which is consistently shorter than the values calculated for PC Cl 3 cells, that is, between 24 and 36 h. Without TSH, PC Cl 3 cells cease proliferation and become quiescent (G0 phase). Upon stimulation with TSH in the presence of 5% serum, quiescent cells reenter the cell cycle and progress through G1 into S phase. D-type cyclins, and in particular cyclin D3, are induced during the delayed-early response to mitogenic stimulation (see Figure 1). To determine whether cyclin D3 is rate limiting for G1 progression in quiescent thyrocytes, we seeded PC Cl 3 and PC-D3 cells in medium supplemented with 5% serum and grew them for 7 days in the absence of hormones. Fresh medium was provided every 48 h. Cells were harvested every 2 days, and the cells were counted with a hemacytometer. As shown in Figure 4b, the constitutive expression of cyclin D3 reduced the requirement for TSH of rat thyrocytes. In the absence of hormones, the average population doubling time of PC-D3 cells was approximately 60 h versus more than 7 days for PC Cl 3 cells. Similar results were obtained when the growth properties of PC Cl 3, PC-CMV-neo and PC-D3 cells were a Cell Number 2500 G2/M S Cell Number S G2/M PC Cl PC D3 1 PC D H Days c +6H -6H S S G2/M 11.6 G2/M 2.0 PC Cl Days S G2/M b DNA Content PC D3 1 d PC Cl3 PC D3 1 PC D3 2 S G2/M PC Cl3 PC D3 1 Figure 4 Cyclin D3 overexpression increases growth rate and induces hormone-independent growth in PC Cl 3 cells. Effects of adaptive cyclin D3 expression on the growth rate and on the hormone dependence of rat thyroid PC Cl 3 cells. Growth rates of normal and transfected PC Cl 3 cells with (a) or without the six hormones (6 H) (b). PC Cl 3, PC-D3 clone 1 and PC-D3 clone 2 were grown in F12 medium supplemented with 5% calf serum and counted every 2 days for 7 days. (c) Flow cytometry analysis of proliferating PC Cl 3 and PC-D3 cells grown with (left panels, þ 6 H) or without the six hormones (right panels, 6 H). (d) PCCl 3 and PC-D3 cells were grown to confluence in medium supplemented with 5% calf serum and six hormones, and analysed by flow cytometry. Propidium iodide-labeled cells were analysed with FACScan using the CELL-FITT program. A representative experiment of two is reported analysed by flow-cytometry (Figure 3c). The fraction of cells in S phase was higher in PC-D3 cells than in vectortransfected PC Cl 3 cells both with and without hormones (Figure 4c). Finally, we observed that whereas PC Cl 3 and PC-CMV-neo cells accumulated in G1 at confluence (o5% of S phase cells at confluence), a significant percent (16%) of PC-D3 cells were in S phase (Figure 4d). Thus, in rat thyrocytes, cyclin D3 can act as a rate-limiting factor for TSH-dependent G1/ S transition and its unrestrained expression removes the hormone-dependency constraint and impairs contactinhibition-induced G1 arrest. We also determined whether unrestrained expression of cyclin D3 induced the acquisition of such other characteristics as proliferation in semisolid media and in vivo tumorigenicity. Soft-agar assays demonstrated that, like parental PC Cl 3 cells, PC-D3 cells did not proliferate in semisolid media (Table 1). Furthermore, inoculation of PC Cl 3 cells in athymic mice did not induce tumor growth (Table 1). Constitutive expression of cyclin D3 in cultured thyrocytes confers growth advantage and hormone independence but fails to confer a fully transformed phenotype. Cyclin D3 overexpression induces redistribution of p27 kip1 from cyclin E-Cdk and cyclin A-Cdk complexes to cyclin D3 Cdk complexes We investigated the molecular mechanism whereby cyclin D3 promotes cell growth in rat thyrocytes. D- type cyclins play a dual role in the regulation of cell cycle progression: they first increase the amount of cyclin D/Cdk4-6 complexes above a certain threshold to start p105 Rb phosphorylation, and then promote activation of cyclin E- and cyclin A-containing complexes by sequestering the cyclin-dependent kinase inhibitor p27 kipl (Bouchard et al., 1999; Sherr and Roberts, 1999; Perez- Roger et al., 1999). To determine the mechanism whereby cyclin D3 controls thyroid cell cycle progression, we compared the composition and activity of cyclin/cdk complexes in PC Cl 3 and PC-D3 cells. We immunoprecipitated proteic extracts frompc Cl 3 and PC-D3 cells with antibodies against cyclins D3, E and A, and evaluated the level of associated proteins in the immunoprecipitates by Western blot (Figure 5). We normalized the Table 1 Transformation markers of PC-D3 cells Cell type Soft-agar colony-forming efficiency a Tumor incidence b PC Cl 3 0 0/6 PC-D3 Cl 1 0 0/10 PC-D3 Cl 4 0 0/8 PC-D3 Cl 5 0 0/7 PC MPSV c 45 6/6 a Colonies larger than 64 cells scored after 3 weeks. Colony-forming efficiency was calculated with the formula: (number of colonies formed/number of plated cells) 100. b Tumorigenicity was assayed by injecting cells into nude mice that were subsequently palpated at the inoculation site each week. c Myeloproliferative sarcoma virus-transformed PC Cl 3 cells

6 7581 Figure 5 Cyclin D3 overexpression sequestrates p27 kip1 from cyclin E Cdk and cyclin A Cdk complexes. Analysis of the cyclin/cdk complexes in cells overexpressing cyclin D3. Total protein extracts (500 mg) were immunoprecipitated with antibodies to cyclin D3 (a), cyclin A (b) or cyclin E (c), as indicated, and analysed by immunoblot to determine the amount of associated cyclin, Cdk or p27 kip1. All immunoprecipitates were normalized against the levels of the immunoprecipitated proteins. Kinase activity was determined by using histone H1 (for cyclin E- and cyclin A-associated kinase activity) or prb peptide (cyclin D3-associated kinase activity) as substrates immunoprecipitates against the level of immunoprecipitated proteins. Analysis of cyclin immunoprecipitates demonstrated that the levels of cyclin D-associated p27 kip1 were higher, and the levels of cyclin E-and A- associated p27 kip1 were lower in PC-D3 cells than in PC Cl 3 cells (Figure 5a, b and c, respectively). This finding suggested that adoptive expression of cyclin D3 in PC Cl 3 cells promoted the formation of a major trimeric complex (i.e., the cyclin D3-Cdk-p27 kip1 complex). Kinase assays with GST-RB or histone H1 as substrates demonstrated much lower cyclin D3-associated kinase activity in PC-D3 cells, in agreement with the finding that approximately threefold more p27 kip1 was coimmunoprecipitated with cyclin D3 (Figure 5a). Conversely, with respect to PC Cl 3 cells, in PC-D3 cells cyclin E- and A-associated p27 kip1 levels decreased, parallel to the increased activity of cyclin A- and (to a lesser extent) cyclin E-associated kinases in PC-D3 cells (Figure 5b, c, respectively). These results suggest that cyclin D3 expression in thyroid cells engineered to constitutively express high levels of cyclin D3, promotes cell cycle progression by sequestering p27 kip1 and thus activating cyclin E- and cyclin A-dependent kinases. Subsequently, we determined whether the transient increase in the expression of endogenous cyclin D3 induced by TSH treatment of PC Cl 3 cells, also promoted proliferation of PC Cl 3 cells by activation of cyclin E- and cyclin A-dependent kinases through p27 kip1 sequestration (Figure 6). To this end, we compared the composition and the activity of cyclin/cdk complexes in PC Cl 3 cells starved for 3 days in 5% serumand in the same cells treated for 24 and 48 h with TSH. Proteic extracts fromstarved and TSH-treated PC Cl 3 cells were immunoprecipitated with antibodies against cyclins D3, E and A, and the level of associated proteins in the immunoprecipitates were evaluated by Western blot. Immunoprecipitates were normalized against the level of immunoprecipitated proteins. Analysis of cyclin immunoprecipitates demonstrated that in 3-day starved PC Cl 3 cells, p27 kip1 was associated preferentially with cyclin E and cyclin A whereas its association with cyclin D3 was Figure 6 Cyclin D3 expression induced by TSH sequestrates p27 kip1 fromcyclin E Cdk and cyclin A Cdk complexes. Analysis of the cyclin Cdk complexes in PC Cl 3 cells induced by TSH. Total protein extracts (800 mg) were immunoprecipitated with antibodies to cyclin D3 (a), cyclin A (b) or cyclin E (c), as indicated, and analysed by immunoblot to determine the amount of associated cyclin, Cdk or p27 kip1. All immunoprecipitates were normalized against the levels of the immunoprecipitated proteins. Kinase activity was determined by using prb peptide (cyclin D3- associated kinase activity) or histone H1 (for cyclin E- and cyclin A-associated kinase activity) as substrates relatively scarce (Figure 6a, b and c, respectively). Upon treatment with TSH, cyclin D3 expression increases and in the same manner increased the amount of p27 kip1 associated to cyclin D3 (Figure 6a). In parallel, we observed a reduction in the level of p27 kip1 associated with cyclin E and cyclin A (Figure 6b, c, respectively). This finding suggested that the transient expression of cyclin D3 induced by TSH in PC Cl 3 cells promoted the formation of a major trimeric complex (the cyclin D3- Cdk-p27 kip1 complex) after 24 h of TSH stimulation, reducing the amount of p27 kip1 associated with cyclin E- and cyclin A-containing complexes. Accordingly, kinase assays with GST-RB or histone H1 as substrates demonstrated lower cyclin D3-associated activity in PC Cl 3 cells stimulated with TSH for h compared with starved PC Cl 3 cells, in agreement with the finding that approximately twofold more p27 kip1 was coimmunoprecipitated with cyclin D3 (Figure 6a). Conversely, TSH treatment induced a decrease in cyclin E- and A- associated p27 kip1 levels, and an increase in the activity of cyclin A- and cyclin E-associated kinases (Figure 6b, c, respectively). These results indicated that in PC Cl 3 cells, cyclin D3 induced by TSH treatment, promoted proliferation by activating cyclin E- and cyclin A-dependent kinases through sequestration of p27 kip1. Regulation of cyclin D3 expression by TSH in vivo To verify further whether cyclin D3 is involved in the TSH-dependent growth of thyroid follicular cells, we used an experimental in vivo model of TSH-dependent goitrogenesis that has been successfully used in a number of thyroid studies (Wollman et al., 1978; Wynford-Thomas et al., 1982; Smeds and Wollman, 1983; Rognoni et al., 1984; Sato et al., 1995; Viglietto et al., 1997; Viglietto et al., 1999). We administered the goitrogen drug propylthiouracyl (PTU) to rats to increase the serumlevels of TSH. The significantly increased serumlevel of TSH in PTU-fed rats was accompanied by marked enlargement of the thyroid gland, and by a decrease in the levels of thyroid hormones T 3 and T 4 (Table 2). Proliferation of thyroid

7 7582 Values Table 2 Propylthiouracil treatment of rats Days T3 (ng/ml) T4 (mg/dl) TSH (ng/ml) PCNA (%) o Thyroid weight (mg) Cyclin D3 a +/ a Measured with the Western blot technique (see text) and expressed as arbitrary units follicular cells in PTU-fed rats was evident fromday 5 and peaked between days 8 and 16 (Table 2). The thyroid glands of control rats expressed low levels of cyclin D3 protein (Figure 7a, lane 1). The expression of cyclin D3 protein started to increase on day 3 of PTU-treatment and increased steadily (Figure 7a, lanes 2 and 3). Cyclin D3 upregulation induced by increased TSH levels in PTU-fed rats preceded PCNA-labeling of thyroid follicular cell nuclei, which suggests that cyclin D3 is in the proliferation of TSH-stimulated thyroid cells. Finally, immunohistochemical staining of thyroid glands of untreated rats with anticyclin D3 antibodies demonstrated low-to-moderate cyclin D3 expression in scattered cells (Figure 7b). In the hyperplastic thyroids of PTU-treated rats, not only was staining for cyclin D3 much more intense at a single-cell level, but many more cells stained positive for the antibodies after 5 days of treatment (Figure 7c). Control reactions, performed for each experiment by omitting the anticyclin D3-specific antibodies, showed no staining (data not shown). Figure 7 In vivo regulation of cyclin D3 by TSH. (a) Analysis of protein expression of cyclin D3 in the thyroid gland of PTU-treated rats, by Western blot. Lane 1: thyroid glands fromcontrol rats; lanes 2 and 3: thyroid glands fromrats treated with PTU for 5 and 10 days, respectively. Proteins (40 mg) were loaded in each lane. The bottompart of the figure shows hybridization with g-tubulin to ensure protein uniformloading and integrity. The ratios of cyclin D3 protein in the thyroid gland of PTU-treated rats cells respect to control rats (set 1 at day 0) were 1.8- and 4.1-fold after 5 and 10 days of PTU treatment, respectively. (b) Immunostaining analysis of cyclin D3 expression in the thyroid glands of untreated rats. (c) Immunostaining analysis of cyclin D3 expression in the thyroid glands of rats treated with PTU for 5 days Figure 8 Cyclin D3 expression in human non-neoplastic thyroid diseases. (a) Western blot analysis of cyclin D3 expression in human thyroid diseases. Total proteins (50 mg) fromgoiters (lanes 1 7) or adenomas (lanes 8 12) were loaded in each lane. The bottompanel shows hybridization with b-tubulin to ensure uniformprotein loading and integrity. (b d) Cyclin D3 expression in goiter. (b) or in follicular adenomas ((c, 40) and (d, 100)) Role of cyclin D3 in human thyroid diseases Our observation that cyclin D3 is a key regulator of thyroid cell proliferation in experimental in vitro and in vivo models raises the possibility that upregulation of cyclin D3 induced by chronic activation of the TSHr/ camp pathway may be involved in the development of human thyroid benign diseases such as goiters and adenomas. To address this hypothesis, we determined cyclin D3 expression by Western blot hybridization on bioptic specimens from 32 patients affected by goiters, and 10 by thyroid adenomas. Normal human thyroid gland expresses almost undetectable levels of cyclin D3 protein. Conversely, cyclin D3 protein was increased by about 1.5- to 2.2-fold on average in about half patients with goiters (15/32 cases) and by about 3.5- to sevenfold in all patients with adenomas (10/10), indicating that cyclin D3 overexpression is a frequent feature of human hyperproliferative diseases of the thyroid gland, and in particular of adenomas. A representative experiment is reported in Figure 8a. To confirmthe relevance of cyclin D3 in human thyroid pathology, we evaluated the expression of cyclin D3 protein by immunostaining in normal thyroid gland (n ¼ 2), in multinodular goiters (n ¼ 13) and in adenomas (n ¼ 10) and correlated it to the evaluation of Ki-67 staining as a proliferative index (results are reported in Table 3). The Labeling index (LI) for cyclin D3 was

8 Table 3 Summary of cyclin D3 expression in thyroid goiters and adenomas Histology Cyclin D3 b Ki-67 b Multinodular goiter, median (range) 0.8 ( ) 0.2 ( ) Follicular cell adenomas, median (range) 8.5 ( ) 1.9 ( ) a The expression of cyclin D3 and Ki-67 proteins was determined by immunostaining as described in Materials and methods. Data are expressed as median value and range. b Po0.001 obtained by assessing at least 1000 epithelial follicular cells, and was expressed as a percentage of the total cell population. In normal thyrocytes, cyclin D3 immunoreactivity was very low and was observed only in few scattered follicular cells (LIo0.1%). In agreement with the results obtained with the Western blots, goiters presented cyclin D3 staining in a greater fraction of cells (n ¼ 13) (see for an example Figure 8b) whereas adenomas (10/10) stained consistently for cyclin D3 expression (Figure 8c). LI for cyclin D3 in goiters ranged between 0.2 and 1.2% with a median value of 0.8%. Conversely, in adenomas the median value of LI for cyclin D3 was 8.5% (ranging from7.0 to 11.0%). The nonparametric Mann Whitney U-test was used to compare differences in LIs for cyclin D3 in multinodular goiter and in follicular cell adenomas: the differences observed were statistically significant (Po0.001). To determine the relevance of cyclin D3 overexpression in the proliferation of thyrocytes in vivo, we correlated cyclin D3 expression in goiters and adenomas with the proliferation rate of thyrocytes, as measured by Ki-67 staining. As shown in Table 3, in multinodular goiters, thyrocyte proliferation is low (the median value obtained was of 0.2, ranging between 0.1 and 0.3) whereas in adenomas thyrocyte proliferation was drastically enhanced (the median value obtained was of 1.9, with a range of ). By using the Spearman rank order correlation, we determined that the association between cyclin D3 and Ki-67 expression in thyroid hyperproliferative diseases was statistically significant. In fact, the value of the Spearman R was 0.87 (Po0.001). Taken together these results indicate that cyclin D3 overexpresion is correlated with the proliferative status of thyrocytes and that cyclin D3 may play a crucial role in driving the growth of thyroid adenomas. Discussion Most thyroid diseases characterized by thyroid enlargement (goiter) are consequent to abnormalities in thyroid growth control. In the initial stage of goiter development, the gland is hypertrophic and hyperplastic due to chronic TSHr stimulation (as, for example, in patients affected by nontoxic goiters or Graves disease). Conversely, nodular goiters and adenomas possibly implicate the development of functional autonomy, that is, independence from TSH stimulation in one or more areas of the gland consequent to activating mutations in the genes encoding TSHr or Gsa (Corvilain et al., 2001). Constitutive activation of the TSHr/cAMP pathway has repeatedly been implicated in the pathogenesis of most thyroid hyperproliferative diseases (Dumont et al., 1992). However, the precise mechanisms whereby chronic hyperactivation of the TSHr/cAMP pathway deregulate the thyrocyte cycle during the development of goiters and adenomas are unknown. Here, we provide compelling evidence that cyclin D3, a G1 cyclin, is an important target of the TSHr/cAMP pathway in thyrocytes, and that camp-dependent upregulation of cyclin D3 is involved in the pathogenesis of human thyroid goiter and adenoma. We also show that thyroid tissues derived frompatients affected by multinodular goiter (n ¼ 32) contained moderate levels of cyclin D3 protein whereas follicular adenoma (n ¼ 10) contained higher levels of cyclin D3 protein. On average, thyroids frompatients with adenoma or nodular hyperplasia expressed higher levels of cyclin D3 protein than did normal thyroids. Moreover, the results and the statistical analysis performed to validate such results, clearly demonstrated that cyclin D3 expression was adenomas>goiters>normal thyroid and that the level of cyclin D3 protein was correlated with the values of cell proliferation in the thyroid lesion, as measured by Ki-67 staining. Using an experimental animal model of TSH-dependent thyroid goitrogenesis that mimics some aspects of the human disease and by in vitro experiments with the rat cultured cell line PC Cl 3, we confirmthat cyclin D3 is crucial for the proliferation of thyroid follicular cells during goiter development in humans. Our in vivo results demonstrate that cyclin D3 expression is markedly increased by the treatment of animals with a goitrogenous compound (PTU). The increase occurs after serum TSH levels increase (at days 1 3), but precedes the onset of thyrocyte proliferation (at days 3 5) and the increase in thyroid weight (observed at day 5). Thus, we conclude that the constitutive upregulation of cyclin D3 induced by chronic TSHr stimulation drives thyrocytes proliferation and hence goiter development. This conclusion is in agreement with another animal model of camp-driven thyroid disease, that is, transgenic mice expressing the adenosine A 2a R serpentine receptor under the control of the thyroglobulin promoter (Ledent et al., 1992). In these transgenic mice, camp accumulation induces pronounced cyclin D3 expression and promotes the development of thyroid goiter and severe hyperthyroidismthat resembles the human disease (Ledent et al., 1992). Cyclin D3 is a member of the D-type family of G1 cyclins, a family of proteins that regulate G1 progression in mammalian cells (Morgan, 1995). Our in vivo and in vitro results suggest that in rat PC Cl 3 cells, cyclin D3 may function as the sensor of serum TSH levels by linking the TSHr/cAMP pathway to the cell cycle machinery. On this basis, cyclin D3 represents a rate-limiting factor for the progression of PC Cl 3 cells fromg1 to S. In fact, its constitutive expression abolishes the hormone-dependence constraint and 7583

9 7584 contact inhibition. These results are complementary to the mrna interference experiments, which showed that cyclin D3 expression was apparently more relevant than cyclin D1 for TSH-dependent proliferation of PC Cl 3 cells. Furthermore, these results are reminiscent of those obtained with single-cell microinjection, showing that cyclin D3 was necessary for G1 progression of dog thyrocytes (Depoortere et al., 1998). In the FRTL-5 rat thyroid cell line model, TSH and/or IGF-1, accelerate G1 phase by inducing the expression of cyclins D1, D3 and E and by decreasing p27 kip1 expression (Yamamoto et al., 1996; Carneiro et al., 1998; Medina et al., 1999). Within this framework, cyclin D3 is a likely candidate for targeting by deregulation of the TSHr/cAMP pathway during the development of human thyroid diseases. We propose that hyperactivation of the TSHr/ camp pathway, both in vitro and in vivo, induces cyclin D3 expression, which in turn, may cause unrestrained proliferation of thyrocytes with subsequent thyroid enlargement and autonomous growth. This conclusion, however, does not rule out the possibility that other pathways that are activated in thyroid cells during the development of hyperplasias, such as the EGF/EGF receptor and the IGF-1/IGF-1 receptor, may have a role in the observed upregulation of cyclin D3 (Pedrinola et al., 2001). Dysregulation of cyclin D3 can be a primary oncogenic event in human lymphoid and myeloid tumors (Shaughnessy et al., 2001; Sonoki et al., 2001). However, as expected froma genetic alteration that occurs in hyperplastic diseases or benign tumors, thyrocytes engineered to overexpress cyclin D3 show dysregulated proliferation, but neither growth in soft agar nor tumorigenicity in nude mice. This conclusion agrees with the concept that cyclin D3 functions to limit the rate of G1 phase progression in fibroblasts (Herzinger and Reed, 1998). Lineage-specific differences in the role of cyclin D3 or a different degree of susceptibility to the transforming potential of unrestrained cyclin D3 expression may account for the differences observed. Growth factors regulate mammalian cell cycle entry by stimulating accumulation of D-type cyclins (cyclins D1, D2 and D3), which then assemble into holoenzymes with Cdk4 or Cdk6. In turn, cyclin D-Cdk4/6 complexes promote progression through G1 by two distinct but complementary mechanisms. Complexes of cyclin D- Cdk4/6 phosphorylate the retinoblastoma susceptibility gene product p105 Rb, and activate the E2F family of transcription factors, which in turn activate the transcription of genes required for S-phase entry, including cyclin E and cyclin A. Alternatively, cyclin D-Cdk4/6 complexes sequester the cyclin-dependent kinase inhibitor p27 kip1 fromcyclin E-Cdk2, thus releasing cyclin E-Cdk2 fromits inhibition (Bouchard et al., 1999; Perez-Roger et al., 1999; Sherr and Roberts, 1999). In PC Cl 3 cells, the excess of cyclin D3 appears to titrate p27 kip1 fromcyclin E- or cyclin A-containing complexes thereby leading to Cdk2 activation. Sequestration of p27 kip1 by cyclin D3/Cdk complexes in PC Cl 3 cells could be necessary because, unlike FRTL-5 and WRT cells, p27 kip1 is not downregulated in PC Cl 3 cells in response to TSH or insulin (Carneiro et al., 1998; Medina et al., 1999). In conclusion, these data suggest that cyclin D3 plays a pathogenic role in the first steps of hyperproliferative diseases of the thyroid gland. However, it is necessary to point out that cyclin D3 may not be the only determinant of thyrocyte proliferation in human thyroid hyperproliferative diseases, since previous works have demonstrated that cyclin D1 is overexpressed in hyperplastic nodules and adenomas (Saiz et al., 2002). Materials and methods Tissue samples and immunostaining Samples of thyroid goiters were obtained after resection from patients who had undergone surgery at the National Cancer Institute Fondazione Pascale, Naples, Italy or at the Dipartimento di Scienze Morfologiche e Funzionali of University Federico II of Napoli. Samples of thyroid multinodular goiter (n ¼ 32), follicular adenoma (n ¼ 10) and normal tissues (n ¼ 5) were used for protein expression by Western blot and tissues fromgoiters (n ¼ 13), follicular adenoma (n ¼ 10) and normal thyroid (n ¼ 2) were evaluated for immunohistochemical study. For immunohistochemistry analysis, 5 6 mmparaffin sections were deparaffinized and incubated overnight with primary antibody (DCS22 monoclonal antibody from Neomarkers) diluted 1 : 1000 in PBS. Antigen retrieval was performed by heat. The Ki-67 antibody was the MIB-1 monoclonal antibody (Novocastra). It was used at a dilution of 1 : 50. Immunostaining was revealed by incubating the slides in diaminobenzidine (DAB-DAKO) solution containing 0.06 mm DAB and 2 mm hydrogen peroxide in 0.05% PBS ph 7.6 for 5 min. Micrographs were taken on Kodak Ektachrome film with a photo Zeiss system. LIs for Cyclin D3 and Ki-67 were obtained by assessing at least 1000 epithelial follicular cells, and expressed as a percentage of the total cell population. In the experiments with PTU-treated rats, cyclin D3 expression was determined in at least three animals for each experimental point. The statistical analysis was performed using SPSS s ver for Windows s. Data in the text and tables are expressed as median value and range. The nonparametric Mann Whitney U-test was used to compare differences in LIs for Cyclin D3 and Ki-67 in multinodular goiter and in follicular cell adenomas. The Spearman rank order correlation was used to verify the association between cyclin D3 and Ki-67. A P-value less that 0.05 was considered statistically significant. Cell lines and culture The rat thyroid cell line PC Cl 3 used in this study is a thyroid epithelial cell line derived from18-month-old Fisher rats that are described elsewhere (Fusco et al., 1987; Berlingieri et al., 1990). PC Cl 3 cells were grown in F12 medium (Sigma) supplemented with 5% calf serum (Life Technologies) additioned with a mix of six hormones TSH, insulin, somatostatin, growth hormone, glycil-hystidil-lysine and transferrin (Sigma). In vivo studies with propylthiouracil-fed rats A total of 30 Fisher rats (8-week-old) were fed low-iodide rat chow. The goitrogen agent propyl thiouracil (PTU) was added

10 to the drinking water at a concentration of 2 mg/ml. A total of 30 Fisher rats (8-week-old) fed normal chow served as controls. After 1, 3, 5, 10 days, the rats were anesthetized with ether, blood samples were taken from the jugular vein and the rats were killed. TSH, T 3 and T 4 hormones were evaluated by radioimmunoassay with a rat-specific kit (Amersham Pharmacia Biotech). Thyroid glands were removed and divided in two equally representative parts: one was weighed and frozen in liquid nitrogen until required for RNA extraction, the other fixed in 4% paraformaldehyde for 15 h at 41C, embedded in paraffin and processed for immunoperoxidase staining. Protein extraction, Western blot and antibodies The antibodies used in this work were obtained fromsanta Cruz Inc. (anticyclin D3 C16, anticyclin D1 HD11 and anticyclin D2 C-17), Science (anticyclin A AB-2 and anti-p21 AB-1), Pharmingen (anti-p16, anticyclin B1 anticyclin D1, anticyclin E HE12, anti-cdk1 A17 and antiprb G3-245, anticyclin A BF638, anticyclin E HE67 and anti- Cdk4) and Transduction Laboratories (anti-p27, anti-cdk2). Cells were scraped in ice-cold phosphate-buffered saline (PBS) and lysed in cold Nonidet-P40 (NP-40) lysis buffer (0.5% NP- 40, 50 mm HEPES ph 7, 250 mm NaCl, 5 mm EDTA, 50 mm NaF, 0.5 mm Na 3 VO 4, 0.5 mm PMSF, 5 mg/ml of aprotinin and 5 mg/ml of leupeptin. Proteins (50 mg) were separated on polyacrylamide gel, transferred to nitrocellulose filters membranes (Hybond C, Amersham Pharmacia Biotech). Membranes were blocked in 5% nonfat dry milk for 2 h at room temperature and subsequently incubated with primary antibodies for 2 h at room temperature. Membranes were incubated with secondary antibodies for 1 h at roomtemperature and revealed by enhanced chemiluminescence (Amersham Pharmacia Biotech). References Baldassarre G, Belletti B, Bruni P, Boccia A, Trapasso F, Pentimalli F, Barone MV, Chiappetta G, Vento MT, Spiezia S, Fusco A and Viglietto G. (1999). J. Clin. Invest., 104, Berlingieri MT, Akamizu T, Fusco A, Grieco M, Colletta G, Cirafici AM, Ikuyama S, Kohn L and Vecchio G. (1990). Biochem. Biophys. Res. Comm., 173, Bouchard C, Thieke K, Maier A, Saffrich R, Hanley-Hyde J, Ansorge W, Reed S, Sicinski P, Bartek J and Eilers M. (1999). EMBO J., 18, Carneiro C, Alvarez CV, Zalvide J, Vidal A and Dominguez F. (1998)., 16, Coppee F, Depoortere F, Bartek J, Ledent C, Parmentier M and Dumont JE. (1998)., 17, Corvilain B, Van Sande J, Dumont JE and Vassart G. (2001). Clin. Endocrinol., 55, Davies T, Marians R and Latif R. (2002). J. Clin. Invest., 110, Depoortere F, Van Keymeulen A, Lukas J, Costagliola S, Bartkova J, Dumont JE, Bartek J, Roger PP and Dremier S. (1998). J. Cell Biol., 140, Immunoprecipitation and kinase assay Cells were lysed in (NP-40) lysis buffer, and the protein concentration of lysates was determined. Proteins (400 mg) were immunoprecipitated with 1 2 mg of the indicated antibodies for 60 min at 41C. Immunoprecipitated proteins were collected on protein A/G-Sepharose (Santa Cruz). Nine-tenths of the immunoprecipitated proteins were resolved on polyacrylamide denaturing gels, transferred onto nitrocellulose filters and incubated with primary antibodies as described above. One-tenth of immunoprecipitates were resuspended in 25 ml of kinase buffer containing 20 mm MOPS, ph 7.2, 25 mm b-glycerophosphate, 5 mm EGTA, 1 mm sodiumorthovanadate, 1 mm DTT, 7.5 mm MgCl 2,50mM ATP and 10 mci of [g- 32 P] datp and 5 mg of Histone H1 (Upstate Biotechnology) (for cyclin E- and cyclin A-associated kinase activity) or 1 mg of GST-pRB 769 (Santa Cruz Inc.) (for cyclin D3-associated activity) and incubated for 15 min at 301C. Reactions were stopped and incorporation of radioactive phosphate was determined by SDS PAGE. Dried gels were analysed with a Phosphorimager (GS-525 Biorad), interfaced with Hewlett- Packard computer. Reagents and antisense oligonucleotides Antisense phosphorothioate oligonucleotides used in this studies are: cyclin D CCAAGCGGAGCAAACTCT-3 0 ; cyclin D CAGTCTTAAGCATGGCTC-3 0. Oligonucleotides were used at a concentration of 1 mm and were delivered by the Oligofectamine reagent (Invitrogen). Indirect immunofluorescence Cells were grown to subconfluence on coverslips, fixed in 3% paraformaldehyde, and permeabilized with 0.2%. Triton X bromo-3 0 -deoxyuridine (BrdU) was added to the culture medium to a final concentration of 10 mm and allowed to react for 1 h. A 5 0 -bromo-3 0 -deoxyuridine Labeling and Detection Kit fromboehringer Mannheimwas used to identify S-phase cells. Cell nuclei were identified by HOECHST staining. Fluorescence was viewed with a Zeiss 140 epifluorescent microscope equipped with filters that allowed discrimination between Texas red and FITC. Acknowledgements This work was supported by the Progetto Finalizzato Biotecnologie del CNR to GV and fromassociazione Italiana Ricerca sul Cancro (AIRC) and Ministero dell Istruzione e Ricerca Scientifica (MIUR) to AF. We are indebted to Jean Ann Gilder for editing the text. Dremier S, Coulonval K, Perpete S, Vandeput F, Fortemaison N, Van Keymeulen A, Deleu S, Ledent C, Clement S, Schurmans S, Dumont JE, Lamy F, Roger PP and Maenhaut C. (2002). Ann. N. Y. Acad. Sci., 968, Dumont JE, Lamy F, Roger P and Maenhaut C. (1992). Physiol. Rev., 72, EkholmSV and Reed SI. (2000). Curr. Opin. Cell Biol., 12, Florio T, Arena S, Thellung S, Iuliano R, Corsaro A, Massa A, Pattarozzi A, Bajetto A, Trapasso F, Fusco A and Schettini G. (2001). Mol. Endocrinol., 15, Fusco A, Berlingieri MT, Di Fiore PP, Portella G, Grieco M and Vecchio G. (1987). Mol. Cell. Biol., 7, Hedinger C, Willians ED and Sobin LH. (1993). International Histological Classification of Tumours, World Health Organization. 2nd edn. Springer-Verlag: Berlin, Heidelberg, New York, London, Paris, Tokyo. Herzinger T and Reed SI. (1998). J. Biol. Chem., 273,