Oncogene (1997) 15, 2743 ± 2747  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

SHORT REPORT Di€erences in the mechanisms of growth control in contact-inhibited and serum-deprived human ®broblasts

Cornelia Dietrich, Katja Wallenfang, Franz Oesch and Raimund Wieser

Institute of Toxicology, Johannes-Gutenberg University, 55131 Mainz, Germany

In the present work we studied mechanisms of growth small inhibitory , known as p15, , p18, p19, control in contact-inhibited and serum-deprived human , p27 and p57 (for review see Sherr, 1994, 1996; diploid ®broblasts. The observation that the e€ects on Sherr and Roberts, 1995). [3H]thymidine incorporation and reduction of retinoblas- The activating pathway of mitogenic signals such as toma gene product-phosphorylation were additive when serum, growth factors or mitogenic hormones has contact-inhibition and serum-deprivation were combined been well established in the last few years. As a led us to the conclusion that the underlying mechanisms delayed early response, growth factors induce the might be di€erent. Both contact-inhibition and serum- expression and synthesis of D-type (D1, D2, deprivation led to a strong decrease of cdk4-kinase- D3) which are di€erentially and combinatorially activity and cdk2-phosphorylation at Thr 160, while the expressed in mammalian cells (Won et al., 1992; total amounts of cdk4 and cdk2 remained constant. In Matsuchime et al., 1991; Xiong et al., 1993). In contact-inhibited cells, we revealed a strong complex with their major catalytic partner cdk4, they accumulation of the cdk2-inhibitor p27 and a slight, but phosphorylate pRB in mid to late G1 (Matsushime et signi®cant increase of the cdk4-inhibitor p16. In serum- al., 1994; Xiong et al., 1993) which induces E deprived cells, the protein levels in p27 and p16 remained expression. At G1/S-transition, accumulates low. In contrast, we detected a rapid decrease of cyclin which in association with cdk2 also phosphorylates D1 and which did not occur in contact- pRB and, in addition, still unknown substrates inhibited cells. These results indicate that serum- (Hatakeyama et al., 1994). deprivation and contact-inhibition have di€erent mechan- In contrast, less is known about isms although they a€ect the same pathway ± pathways due to inhibiting signals such as contact- cdk4, pRB, cyclin E ± cdk2. inhibition and serum-deprivation. Polyak and co- workers (1994a) as well as Hengst and coworkers Keywords: Cyclin-dependent kinase regulation; contact- (Hengst et al., 1994; Hengst and Reed, 1996) could inhibition; serum-deprivation show an accumulation of the cdk-inhibitor p27. In contrast, ®broblasts from p27-knock out mice still show full contact-inhibition (Nakayama et al., 1996) which suggest the involvement of additional, still The decision of mammalian cells wether to replicate unknown mechanism(s). DNA and proliferate or to withdraw from In general, no mechanistic discrimination is made and go into quiescence is taken in late . This between quiescence due to high density or G1 checkpoint referred to as restriction (R) point has withdrawal. Since in vivo contact-inhibition, and not been ®rst described by Pardee (1989). It is now serum-deprivation, plays the fundamental role in generally believed that the retinoblastoma gene proliferation control (i.e. wound-healing), we postu- product (pRB) plays a crucial role in regulating the lated di€erences in the mechanisms of contact- restriction point. In its hypophosphorylated form pRB inhibition and serum-deprivation. In addition, in binds to the transcription factor hence inhibiting vitro, di€erences between contact-inhibition and ser- transcription. (Hyper)phosphorylation of pRB leads to um-deprivation have been described (Del Sal et al., its functional inactivation resulting in the loss of 1992; Gustincich and Schneider, 1993; Moreton et al, binding to E2F thus permitting entry into S-phase 1995a). In the present work we studied the e€ects of (Mittnach and Weinberg, 1991; DeCaprio et al., 1989). contact-inhibition and serum-deprivation on closely pRB is phosphorylated by a speci®c family of serine/ connected upstream regulators of pRB in human threonine kinases, the cyclin-dependent kinases (cdks). diploid ®broblasts. Several mechanisms are employed to regulate cdk- Exponentially growing FH109 human embryonal activity: activation of a cdk, the catalytic subunit, is lung ®broblasts (Wieser et al., 1985) were seeded either dependent on the association with a cyclin, the sparsely or to high density in CG-medium (Vitromex) regulatory subunit. In addition, activation of the supplemented with 0.5% fetal calf serum (Gibco). catalytic subunit requires phosphorylation at a FH109 cells have been shown to highly sensitive to conserved Thr residue by cdk-activating kinase and cell-cell contacts with respect to growth-inhibition dephosphorylation at Thr14 and Tyr15. The activity of (Wieser et al., 1985, 1990; Wieser and Oesch, 1986). the complex is further modulated by the association of After 6 h the cells adhered to the plates (t=0), and the medium of sparsley seeded cells was changed to CG- medium without fetal calf serum for serum-deprivation Correspondence: R Wieser experiments. The cells were harvested after 24, 48 and Received 9 April 1997; revised 18 July 1997; accepted 21 July 1997 72 h. Growth control by contact-inhibition and serum-deprivation C Dietrich et al 2744 We ®rst investigated the e€ect of contact-inhibition clear shift to faster electrophoretic mobility. Phosphor- and serum-deprivation on proliferation of FH109 cells. ylation of pRB was even less when contact-inhibition and Proliferation was measured by [3H]thymidine incorpora- serum-deprivation were combined. The fact that the tion and pRB-phosphorylation. Phosphorylation of pRB antiproliferative e€ects of contact-inhibition and serum- from mid to late G1-phase is believed to be the critical deprivation were additive led to the assumption that step for G1/S-transition and hence re¯ects cell-cycle di€erent inhibitory pathways might be involved. progression in G1. After 24 h, both contact-inhibition Since phosphorylation of pRB is belived to be the and serum-deprivation reduced [3H]thymidine incor- ®nal control point for G1/S-transition we next poration to 20% compared to exponentially growing examined the upstream regulators of pRB-phosphor- cells whereas the combination resulted in a decrease to ylation, namely regulation of cdk4 and cdk2. The 5% (Figure 1a). [3H]Thymidine incorporation did not kinase cdk4 is activated by association of D-cyclins, decrease further after 48 h or 72 h. The inhibition of cdk2 is activated by cyclin E. While p16 speci®cally proliferation was also re¯ected by a reduction of pRB- blocks cdk4 (Serrano et al., 1993), p27 inhibits phosphorylation (Figure 1b). Using Western blot although not selectively but preferentially cdk2 analysis, we predominantly detected the slower migrat- (Harper et al., 1995; Soos et al., 1996). We ®rst ing species of pRB in proliferating cells which determined protein levels of cdk4 and cdk2, corresponds to the hyperphosphorylated, inactive pRB and D3 as well as p16 and p27. Protein levels of cdk4 (DeCaprio et al., 1989). In contact-inhibited or serum- were not a€ected, neither in response to serum- deprived cells, phosphorylation of pRB was strongly deprivation nor to contact-inhibition, and protein reduced and, concomitantly, we revealed an increase of levels of cdk2 also remained constant after contact- the hypophosphorylated, active species of pRB with a inhibition and for 48 h after serum-deprivation (Figure 2a). However, serum-deprivation resulted in a marked decrease of the cyclins D1 and D3 whereas they remained una€ected in contact-inhibited ®broblasts a (Figure 2b). This result is in accordance with the work of Won et al. (1992) who detected a 70%-

a C SD CI kD 34 — cdk4

35 — cdk2

1 2 3 4 5 6 7

b C SD CI b kD 34 — D3 KD 117 — pRB*

105 — pRB 34 — D1 C SD CI CI + SD 1 2 3 4 5 6 7 Figure 1 Proliferation of exponentially growing (C), serum- deprived (SD), contact-inhibited (CI), or serum-deprived and contact-inhibited (CI+SD) FH109 cells. Proliferation was determined by [3H]thymidine incorporation (a) and phosphoryla- c tion of pRB by Western blot analysis (b). (a)56103 (sparsley) or C SD CI 56104 (high density) cells were seeded into microtiter plates, kD treated and cultured as mentioned. 24 h, 48 h and 72 h after cell 16 — p16 adhesion, the cells were labeled with 0.25 mCi/well of [3H]thymidine for 4 h. Incorporated radioactivity was determined by liquid scintillation spectrometry. (b)56105 (sparsley) or 3.56106 (high density) cells (60 mm-plates) were cultured as 27 — p27 described and solubilized 24 h after cell adhesion in 500 mlof 1 2 3 4 5 6 7 boiling SDS-sample bu€er (Laemmli, 1970). Protein determina- tion was performed according to (Smith et al., 1985), proteins Figure 2 Protein levels of cdk4 and cdk2 (a), cyclin D1 and D3 (20 mg) precipitated as described by (Wessel and FluÈ gge, 1984) (b), and p16 and p27 (c) in exponentially growing (c), serum- and solubilized in 20 ml of SDS-sample bu€er. After SDS- deprived (SD) or contact-inhibited (CI) FH109 cells. Total cell polyacrylamide gel electrophoresis (12.5%), proteins were extracts were harvested 24 h (lanes 1, 2, 5), 48 h (lanes 3 and 6) transferred onto Immobilon membrane (Millipore). Immuno- and 72 h (lanes 4 and 7) after cell adhesion followed by Western detection with anti-pRB-antibodies (0.1 mg/ml, Santa Cruz) was blot analysis as described in Figure 1 except that 50 mg of protein performed as previously described (Dietrich et al., 1996). was precipitated for p16-detection. The position of the molecular pRB*=(hyperphosphorylated) pRB weight marker proteins is shown on the left Growth control by contact-inhibition and serum-deprivation C Dietrich et al 2745 decrease of cyclin D1 and D3 mRNA in human diploid after 24 h which was more pronounced after 48 h and ®broblasts within 7 h after serum-depletion. 72 h. This result might point out that regulation of p16 In contrast, serum-deprived cells had constant levels expression is de®ned to a short period in G1-phase of the inhibitors p16 and p27 never exceeding control since the cells were asynchronously growing. We levels, but in contact-inhibited cells we revealed a con®rmed the hypothesis of p16-elevation in response strong and rapid accumulation of the inhibitor p27 to contact-inhibition in synchronized FH109 cells and (Figure 2c). The increase of p27 protein levels revealed a twofold elevation of p16 protein levels in con®rmed studies by Hengst et al. (1994) and Polyak mid G1-phase which resulted in a strong increase of et al. (1994a) who have recently described an p16 bound to cdk4 (manuscript in preparation). accumulation of the inhibitor p27 in contact-inhibited Conclusively, serum-deprivation led to a decrease of ®broblasts and epithelial cells, respectively. The the positive regulators cyclin D1 and D3 whereas accumulation is partly due to a translational upregula- contact-inhibition resulted in an increase of the tion, partly due to increased half-life of p27 in response inhibitory proteins p16 and p27. to decreased degradation by the - Reduction of cyclin D1 and D3 or accumulation of pathway (Hengst and Reed, 1996). The fact that p27 p16 and p27 should result in decreased kinase activities protein levels remained low in serum-deprived FH109 of cdk4 and cdk2. Using 10 ± 20% gradient polyacryl- cells is in contrast to the work of Kato et al. (1994), amide gels, it is possible to discriminate between the Nourse et al. (1994) and Coats et al. (1996). They activated form of cdk2 and the inactive species. described upregulation of p27 in CSF (colony Phosphorylation at Thr 160 and dephosphorylation at stimulating factor)-starved macrophages, in interleu- Thr14 and Tyr15, i.e. activation, results in a shift to kin2-starved T-lymphocytes and murine Balb/c-3T3 faster electrophoretic mobility (Dulic et al., 1992; Gu et ®broblasts. We conclude that the discrepancies occur al., 1992). Figure 3b demonstrates that the faster due to the di€erent cell systems and species, as, to our migrating species of cdk2 (33 kD) disappeared rapidly knowledge, upregulation of p27 has never been after serum-deprivation (by 50%) or contact-inhibition described in serum-deprived diploid ®broblasts. (by 60%). Since we previously have con®rmed that the In addition to p27-accumulation we detected a slight 33 kD form exerts kinase activity towards histone H1 increase of the inhibitor p16 in contact-inhibited cells in vitro (Dietrich et al., 1996), we concluded that the

a

pRB

C SD CI

b

kD 35 — cdk2 33 — cdk2*

C SD CI Figure 3 Kinase activity of cdk4 (a) and activation of cdk2 (b) of exponentially growing (c), serum-deprived (SD), or contact- inhibited (CI) FH109 cells. Cells were harvested 24 h after cell adhesion. (a) Kinase activity of cdk4-immunoprecipitates was measured as previously described (Dietrich et al., 1996) using a pRB-fusion protein as substrate. The ®gure on the right shows cdk4- activity in % of control values (c=100%). (b) Total cell extracts (20 mg of protein) were separated on 10 ± 20% gradient SDS- polyacrylamide gels. Subsequent Western blot analysis was performed with anti-cdk2-antibodies as mentioned in Figure 1. The position of the molecular size marker proteins is shown on the left. The ®gure on the right shows cdk2-phosphorylation (Thr160) in % of control values (c=100%). cdk2*=phosphorylated (Thr160) cdk2 Growth control by contact-inhibition and serum-deprivation C Dietrich et al 2746 complexes. Since a reduction of cyclin D after serum- deprivation increases the amount of free cdk4 and since p27 has a higher anity to cdk-cyclin complexes than to either subunit alone (Harper et al., 1995), we assume according to the work of Polyak et al. (1994a) that p27 associates with and inhibits cdk2-cyclin E. In conclu- sion, contact-inhibition and serum-deprivation inhibit cdk4- and cdk2-kinase activities by di€erent mechan- isms, but converge at the restriction point, i.e. they ®nally result in decreased pRB-phosphorylation and hence cell cycle arrest in G1-phase. This conclusion is in agreement with earlier published results which also describe di€erences in the growth control of contact-inhibited and serum- deprived cells. For example, Gustincich and Schneider (1993) have shown that the sdr (serum deprivation response) gene is activated in NIH3T3 cells only in Figure 4 Proposed model for the mechanisms of growth control response to serum-deprivation and not to contact- in contact-inhibited and serum-deprived human ®broblasts. inhibition. Protein expression and kinase activity of Serum-deprivation leads to reduction of the positive regulator protein kinase C (PKC) subspecies are also di€erently cyclin D which in turn decreases cdk4-activity. As a result (not shown) p27 binds to and inhibits the cdk2-cyclin E complex. In regulated. In contact-inhibited cells, PKC a and d contrast, contact-inhibition acts via inhibitory proteins: p16 increase whereas PKC g and e decrease. In contrast, no blocks activity of cdk4, p27 inhibits cdk2-activity change can be detected in serum-deprived cells (Moreton et al., 1995a). Expression of MARCKS protein (myristoylated alanine rich C kinase sub- decrease of the phosphorylated form concomitantly strate), a substrate of PKC, accumulates when cells re¯ected a decrease of cdk2 activity. Although cdk4 is reach con¯uence but is reduced as cells become activated similarly by phosphorylation at Thr172 and quiescent by serum-deprivation (Moreton et al., dephosphorylation at Thr15 and Tyr14, it is ± to our 1995b). In addition, elevation of membrane tyrosine knowledge ± not possible to separate the activated phosphatase activity in murine ®broblasts is only form from inactive cdk4 by gel electrophoresis. We coupled to contact-inhibition but not to serum- therefore studied kinase activity in vitro by phosphor- deprivation (Pallen and Tong, 1991). In our study, ylation of a pRB-fusion protein. cdk4-activity was we have shown for the ®rst time, that contact- inhibited by 91% after serum-deprivation and by 73% inhibition and serum-deprivation inhibit pRB-phos- after contact-inhibition (Figure 3a). phorylation by distinct mechanisms although a€ecting We conclude that, in FH109 cells, contact-inhibition the same pathway. Serum-deprivation results in a leads to an increase of the inhibitors p16 and p27 decrease of the positive regulators cyclin D1 and which block cdk4 and cdk2, respectively. p16 acts by cyclin D3 whereas contact-inhibition acts via regula- dissociating cyclin D from cdk4, p27 inhibits cdk2- tion of the inhibitory proteins p16 and p27. cyclin E directly by binding to the complex (Polyak et al. 1994b) and decreased phosphorylation at Thr160 by blocking cdk-activating kinase (Aprelikova et al., Acknowledgements 1995). In contrast, serum-deprivation causes a This work was supported by a grant of the Deutsche decrease of the cdk4-activators cyclin D1 and cyclin Forschungsgemeinschaft WI 727/3-1 and is part of the MD D3 hence reducing the amount of active cdk4-cyclin D thesis of KW

References

Aprelikova O, Xiong Y and Liu ET. (1995). J. Biol. Chem., Harper JW, Elledge SJ, Keyomarsi K, Dynlacht B, Tsai LH, 270, 18195 ± 18197. Zhang PM, Dobrowolski S, Bai C, Connellcrowley L, Coats S, Flanagan WM, Nourse J and Roberts J. (1996). Swindell E, Fox MP and Wei N. (1995). Mol. Biol. Cell., 6, Science, 272, 877 ± 880. 387 ± 400. DeCaprio JA, Ludlow JW, Lynch D, Furukawa Y, Grin J, Hatakeyama M, Brill JA, Fink GR and Weinberg R. (1994). Piwinica-Worms H, Huang C-M and Livingston DM. Genes & Dev., 8, 1759 ± 1771. (1989). Cell, 58, 1085 ± 1095. Hengst L, Dulic V, Slingerland JM, Lees E and Reed SI. Del Sal G, Ruaro ME, Philipson L and Schneider C. (1992). (1994). Proc. Natl. Acad. Sci. USA, 91, 5291 ± 5295. Cell, 70, 595 ± 607. Hengst L and Reed SI. (1996). Science, 271, 1861 ± 1864. Dietrich C, Bartsch T, Oesch F and Wieser RJ. (1996). Proc. Kato J, Matsuoka M, Polyak K, Massague J and Sherr C. Natl. Acad. Sci. USA, 93, 10815 ± 10819. (1994). Cell, 79, 487 ± 496. Dulic V, Lees E and Reed SI. (1992). Science, 257, 1958 ± Laemmli UK. (1970). Nature, 227, 680 ± 685. 1961. Matsushime H, Quelle D, Shurtle€ SA, Shibuya M, Sherr CJ Gu Y, Rosenblatt J and Morgan D. (1992). EMBO J., 11, and Kato J-Y. (1994). Mol. Cell. Biol., 14, 2066 ± 2076. 3995 ± 4005. Matsushime H, Roussel MF, Ashmun RA and Sherr CJ. Gustincich S and Schneider C. (1993). Cell Growth Di€., 4, (1991). Cell, 65, 701 ± 713. 753 ± 760. Growth control by contact-inhibition and serum-deprivation C Dietrich et al 2747 Mittnacht S and Weinberg RA. (1991). Cell, 65, 381 ± 393. Sherr CJ. (1994). Cell, 79, 551 ± 555. Moreton K, Turner R, Blake N, Paton A, Groome N and Sherr CJ. (1996). Science, 274, 1672 ± 1677. Rumsby M. (1995a). FEBS Lett., 372, 33 ± 38. Sherr CJ and Roberts JM. (1995). Genes & Dev., 9, 1149 ± Moreton K, Turner R, Paton A, Groome N and Rumsby M. 1163. (1995b). Biochem. Soc. Trans., 23, 446S. Smith PK, Krohn RJ, Hermanson GT, Mallia AK, Gartner Nakayama K, Ishida N, Shirane M, Inomata A, Inoue T, FH, Provenzano MD, Fujimoto EK, Goeke NM, Olson Shishido N, Horii I, Loh DY and Nakayama K. (1996). BJ and Klenk DC. (1985). Anal. Biochem., 150, 76 ± 85. Cell, 85, 707 ± 720. Soos TJ, Kiyokawa H, Yan JS, Rubin MS, Giordano A, Nourse J, Firpo E, Flanagan WM, Coats S, Polyak K, Lee Deblasio A, Bottega S, Wong BM, Mendelsohn J and Ko€ M-H, Massague J, Crabtree G and Roberts J. (1994). A. (1996). Cell Growth Di€., 7, 135 ± 146. Nature, 372, 570 ± 573. Wessel D and FluÈ gge UJ, (1984). Anal. Biochem., 138, 141 ± Pallen CJ and Tong PH (1991). Proc. Natl. Acad. Sci. USA, 143. 88, 6996 ± 7000. Wieser RJ, Heck R and Oesch F. (1985). Exp. Cell Res., 158, Pardee A. (1989). Science, 246, 603 ± 608. 493 ± 499. Polyak K, Kato JY, Solomon MJ, Sherr CJ, Massague J, Wieser RJ and Oesch F. (1986). J. Cell Biol., 103, 361 ± 367. Roberts JM and Ko€ A. (1994a). Genes & Dev., 8, 9 ± 22. Wieser RJ, SchuÈ tz S, Tschank G, Thomas H, Diener H-P and Polyak K, Lee M-H, Erdjument-Bromage H, Ko€ A, Oesch F. (1990). J. Cell. Biol., 111, 2681 ± 2692. Roberts J, Tempst P and Massague J. (1994b). Cell, 78, Won KA, Xiong Y, Beach D and Gilman MZ. (1992). Proc. 59 ± 66. Natl. Acad. Sci. USA, 89, 9910 ± 9914. Serrano M, Hannon GJ and Beach D. (1993). Nature, 366, Xiong Y, Zhang H and Beach D. (1993). Cell, 71, 505 ± 514. 704 ± 707.