- Email: [email protected]
Inhibitors of cyclin-dependent kinases induce features of replicative senescence in early passage human diploid fibroblasts
Brief Communication
351
Inhibitors of cyclin-dependent kinases induce features of replicative senescence in early passage human diploid fibroblasts Beth B. McConnell, Maria Starborg, Sharon Brookes and Gordon Peters
Address: Imperial Cancer Research Fund, P.O. Box 123, 44 Lincoln’s Inn Fields, London WC2A 3PX, UK. Correspondence: Gordon Peters E-mail: [email protected] Received: 3 December 1997 Revised: 16 January 1998 Accepted: 28 January 1998 Published: 2 March 1998 Current Biology 1998, 8:351–354 http://biomednet.com/elecref/0960982200800351 © Current Biology Ltd ISSN 0960-9822
Results and discussion TIG-3 strain human diploid fibroblasts, which typically undergo 60 to 70 PDs in tissue culture, were infected at PD 40–44 with recombinant retroviruses encoding different Cdk inhibitors (see Materials and methods). These included p21CIP1 and p27KIP1, which interact with multiple cyclin–Cdk complexes, plus p16INK4a and p15INK4b, two members of the INK4 family which specifically inhibit the
Figure 1 (b)
15 Relative number of cells
Ve D cto ay r D 6 ay Se 12 ne sc e
nt
(a)
p15INK4b p16INK4a p21CIP1
Vector p15INK4b p16INK4a p21CIP1 p27KIP1
10
5
0 6
p27KIP1
9
12
Days post-infection
IP 1
ppRb pRb
(d)
Proportion of cells incorporating BrdU (%)
(c)
Ve p1 cto 5 r p1 INK4 p26 INK b 1 4a p2 CIP1 7K
30
20
10
P1
KI
a
1C p2 IP1 7
p2
b
K4
K4
p1
p1
Current Biology
6 IN
ct o 5 IN r
0
Ve
After a limited number of population doublings (PDs), cultures of normal mammalian diploid cells undergo an irreversible growth arrest known as replicative senescence [1]. As well as contributing to cellular ageing, senescence is viewed as an important mechanism of tumour suppression by preventing the emergence of immortal cell clones [2–4]. Senescent cells have a number of characteristics that distinguish them from cycling or quiescent cells including elevated levels of two cyclin-dependent kinase (Cdk) inhibitors, p16INK4a and p21CIP1 [5–11]. Here, we demonstrate that both of these Cdk inhibitors, as well as other members of their protein families (the INK4 and CIP/KIP families, respectively [12]), induce several facets of the senescent phenotype when ectopically expressed in young human diploid fibroblasts. These include a reduced proliferative capacity, an altered size and shape, the presence of underphosphorylated retinoblastoma protein (pRb), increased expression of plasminogen activator inhibitor (PAI-1) and the appearance of senescence-associated b-galactosidase (SA-b-gal) activity [2,3,13–15]. A 20 amino acid peptide from p16INK4a that inhibits Cdks active in the G1 phase of the cell cycle [16] produces similar effects in a dosedependent manner suggesting that, in primary fibroblasts, inhibition of G1-specific Cdk activity is sufficient to induce phenotypic changes that normally occur at the end of their finite lifespan.
Growth arrest of human diploid fibroblasts by ectopic expression of Cdk inhibitors. (a) Lysates were prepared from TIG-3 cells (PD 40–44) 6 or 12 days after infection with recombinant retroviruses encoding p15INK4b, p16INK4a, p21CIP1 or p27KIP1 and the level of the expressed Cdk inhibitor in each lysate was compared to those present in cells infected with the empty vector control (Vector) and from TIG-3 cells that had reached the end of their finite lifespan (~65 PD; Senescent). Equivalent amounts of cellular proteins were fractionated by gel electrophoresis and analysed by immunoblotting [21] with primary antibodies against p15INK4b (Santa Cruz #sc-612), p16INK4a (DCS50), p21CIP1 (Santa Cruz #sc-397) and p27KIP1 (Santa Cruz #sc528). (b) Cells were plated at 1 × 104 cells per well, in a 24-well plate, and at 6, 9 and 12 days after infection with the same retroviruses as in (a), triplicate wells were fixed and stained with 10% formalin and 0.1% crystal violet. The intracellular stain was eluted with 10% acetic acid and relative cell numbers determined by measuring the optical density of the eluates at 590 nm [15]. The results shown were derived from two separate experiments. (c) Immunoblotting for pRb (using the Pharmingen G3-245 antibody) was performed on lysates made six days after infection with the indicated retroviruses. The hyperphosphorylated form of pRb is indicated as ppRb. (d) Cells were labelled with 10 µM bromodeoxyuridine (BrdU; Amersham) for 4 h 6 days after infection with the indicated retroviruses, stained with fluorescein isothiocyanate (FITC)-conjugated anti-BrdU antibody and analysed by two-dimensional flow cytometry.
cyclin-D-dependent kinases Cdk4 and Cdk6 (reviewed in [12]). Retroviruses based on the pBABE-puro or pBABEhygro vectors [17] allowed selection of transduced cells within 4 to 6 days and infection efficiencies were between 50% and 75% as judged by the percentage of cells surviving
352
Current Biology, Vol 8 No 6
Figure 2
Vector
p2 7
C
KI P1
IP 1
(b)
p2 1
Ve
p1 6
IN K4 b IN K4 a
ct o Se r ne sc p1 en t 5
(a)
PAI-1
GAPDH
(c)
p16INK4a
100 Proportion of cells with SA-β-gal activity (%)
Expression of senescence markers in human diploid fibroblasts expressing Cdk inhibitors. (a) Total RNA was isolated from senescent cells or cells 6 days after infection either with a retrovirus encoding the indicated Cdk inhibitor or with the empty vector control (Vector) and 8 µg of each sample was separated by electrophoresis, blotted and hybridised with a PAI-1 cDNA probe. Transcript levels determined by phosphorimaging analysis were normalised to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) levels. (b) Cells were assayed for SA-β-gal activity in situ as previously described [14]. The examples shown are cells expressing vector alone, p16INK4a or p21CIP1 as indicated. Following staining for SA-β-gal activity, cell nuclei were visualised with Hoechst 33528 dye (5 µg/ml for 3 min) to determine the total number of cells in each field. (c) The proportion of cells expressing SA-β-gal activity was calculated by observing ~400 cells per sample in three separate experiments.
75 p21CIP1
50
25
selection. Expression of p15INK4b, p16INK4a, p21CIP1, and p27KIP1 was monitored by immunoblotting cell lysates taken at 6 and 12 days post-infection. The levels of ectopic expression were compared with the endogenous levels present in cells transduced with the empty vector and in TIG-3 cells that had reached the end of their finite lifespan (Figure 1a). In each case, ectopic expression of the Cdk inhibitor was at least 10-fold higher than in the vector-only controls and cells infected with the p16INK4a or p21CIP1 viruses were producing amounts of protein roughly equivalent to those found in senescent cells. As judged by cell proliferation assays, cells expressing the Cdk inhibitors had a significantly lower rate of division than cells transduced with the empty vector (Figure 1b). Fibroblasts expressing INK4 inhibitors were more effectively arrested than those expressing CIP/KIP proteins (Figure 1b) with p15INK4b causing complete growth arrest over the timescale of the experiment. Similarly, whereas pRb was predominantly in its hyperphosphorylated state in proliferating control cells, judged by its mobility in sodium dodecyl sulphate (SDS)-containing acrylamide gels, the Cdk inhibitors caused accumulation of the more rapidly migrating, underphosphorylated form (Figure 1c). Again, the INK4 proteins had a more pronounced effect than the CIP/KIP proteins. The same trend was noted in bromodeoxyuridine (BrdU) incorporation assays: only 3% of the cells expressing p15INK4b or p16INK4a were able to
IP
1
4a C
IP
p2 1 7K p + 16 IP1 p2 IN 1 K
a
1C
p2
K4
K4
5 IN
6 IN
p1
p1
Ve
ct
or
b
0
Current Biology
enter the DNA synthesis (S) phase of the cell cycle during the 4 hour labelling period compared with 15% and 8% of the cells expressing p21CIP1 and p27KIP1, respectively (Figure 1d). These differences in the extent to which cells were arrested by the Cdk inhibitors could be attributable to different levels of expression within the pools of infected cells or to differences in the molar concentration of each protein required to fully block G1 Cdk activity. As well as being growth arrested, the majority of the infected cells were enlarged and flattened in appearance (data not shown). Such morphological changes, which are often associated with senescence [2–4], were noticeable within 6 days of infection and persisted. The cells also had increased expression of the mRNA for PAI-1, which is involved in regulating the turnover of extracellular matrix proteins (Figure 2a). Levels of expression of the two PAI-1 mRNA transcripts, and in particular the smaller of the two, have been shown to be higher in senescent cells [13,15]. The amounts of the smaller PAI-1 transcript were 3–4-fold higher in cells expressing the Cdk inhibitors than in vectoronly controls and approached those found in newly senescent TIG-3 cells (Figure 2a). Infected cells were also tested for expression of SA-β-gal activity, a convenient marker that can distinguish between cells arrested in a senescent versus a quiescent state [14]. As shown in Figure 2b,c, ectopic expression of any of the Cdk inhibitors was sufficient to induce positive staining for SA-β-gal activity, although with
353
Brief Communication
The data we describe here reinforce the proposed connection between replicative senescence and the Cdk inhibitors p16INK4a and p21CIP1 by providing the first direct evidence for their ability to induce senescence-like features in early passage human diploid fibroblasts. Serrano et al. [15] recently described analogous phenotypic changes in human diploid fibroblasts transduced with retroviruses encoding
(a)
(c)
(d)
(b)
Untreated 2.5 µM
7.5 µM
5.0 µM
10.0 µM
(e) 80 Proportion of cells with SA-β-gal activity (%)
These results strongly suggest that the inhibition of G1 Cdk activity is a critical determinant of several facets of the senescent phenotype. To confirm this proposition, we used a 20-residue synthetic peptide corresponding to amino acids 84–103 of p16INK4a to directly inhibit Cdk4/Cdk6 activity in young TIG-3 cells. When coupled to a 16-amino-acid domain from the Drosophila Antennapedia protein, the p16INK4a peptide has been shown to penetrate biological membranes and to block the ability of Cdk4 and Cdk6 to phosphorylate pRb [16]. The Antennapedia–p16 peptide (Ant–p16 peptide), tagged with biotin at its amino terminus, was added to the medium of young TIG-3 cells at a range of concentrations (0–10 µM). Immunofluorescence detection of the biotin tag confirmed that the peptide was present in all cells (Figure 3a). As with introduction of the retrovirally expressed Cdk inhibitors, cells treated with the Ant–p16 peptide had a significantly slower rate of division whereas control cells treated with equimolar amounts of p16 and Antennapedia peptides, synthesised separately and mixed, grew at a normal rate (data not shown). The Ant–p16 peptide also inhibited BrdU incorporation in the treated cells in both a time-dependent and dose-dependent manner (Figure 3b). When cells were stained for SA-β-gal activity at various times following addition of the Ant–p16 peptide, the proportion of positive cells again depended on the concentration of added peptide, with approximately 80% of the cells staining blue after 6 days in 10 µM peptide (Figure 3c–e). At these high concentrations, however, there was a tendency for cells to aggregate and detach from the culture dish, an effect that was not apparent with the separate Antennapedia and p16 peptides. Nevertheless, these results imply that the total concentration of Cdk inhibitor has a direct impact on the completeness of the G1 arrest and the subsequent expression of senescence markers.
Figure 3
Proportion of cells incorporating BrdU (%)
varying degrees of effectiveness. SA-β-gal activity was detectable at 9 days post-infection and maximal by day 12. A higher proportion of cells positive for SA-β-gal was obtained with the INK4 proteins than with p21CIP1 or p27KIP1, suggesting that the ability of the Cdk inhibitors to induce senescence-associated markers was directly related to the extent to which they reduced cell growth and blocked entry into S phase. Significantly, co-infecting cells using retroviruses encoding p16INK4a and p21CIP1 with different selectable markers had an additive effect, with a higher proportion of cells positive for SA-β-gal than with cells expressing either inhibitor on its own (Figure 2c).
30
20 10 0
60 40 20 0
1
3 6 Days of incubation
Untreated 2.5 µM 5.0 µM 7.5 µM 10.0 µM
0
1 2 3 4 5 Days of incubation
6
7
Current Biology
Induction of senescence markers using the Ant–p16 peptide. (a) TIG-3 cells were treated for 24 h with 10 µM biotinylated Ant–p16 peptide and fixed with methanol/acetone (50:50) for 5 min at –20°C. Intracellular peptide was visualised using fluorescein-conjugated streptavidin at a dilution of 1:40. (b) Cells were either left untreated or treated with 2.5, 5.0, 7.5 or 10 µM Ant–p16 peptide, with fresh medium and peptide added every second day. On days 1, 3 and 6 after peptide addition, the cells were labelled with 10 µM BrdU for 4 h and assayed for BrdU incorporation by immunofluorescence. (c) Untreated TIG-3 cells and (d) cells treated with 5 µM Ant–p16 peptide for 6 days were assayed in situ for SA-β-gal activity as described [14]. (e) Cells were either left untreated or incubated with 2.5, 5.0, 7.5 or 10 µM Ant–p16 peptide and the proportion of cells with SA-β-gal activity was estimated 1, 3 and 6 days after peptide addition. Each point was calculated from counting 400 cells in two independent experiments.
activated Ras and ascribed the effects to the upregulation of p53, p21CIP1 and p16INK4a. Similar effects have also been reported in some cancer cell lines ectopically expressing p16INK4a alone [18] but the majority of established cell lines do not respond in this way, raising interesting questions regarding the difference between the G1 arrest commonly associated with ectopic expression of Cdk inhibitors and the panoply of changes induced by Cdk inhibitors in human diploid fibroblasts. One possibility is that such effects are only observed in diploid or ‘near-diploid’ cells and that most tumour cell lines have sustained a sufficient number of additional mutations to obscure or obviate the effects. Another intriguing point is the timescale required to develop some of the phenotypic characteristics of senescence. For example, with retroviral delivery of the inhibitors, SA-β-gal staining was not apparent until 9–10
354
Current Biology, Vol 8 No 6
days post-infection as previously noted by Serrano et al. [15] whereas the changes in cell size and expression of PAI-1 occurred earlier. Such observations suggest a need for the accumulation or decay of some intracellular factor(s). The interpretation of these events, however, depends on whether the ectopic expression of Cdk inhibitors caused a generalised slowing of the cell cycle in all infected cells or whether the variable levels of expression within the pools of cells caused complete arrest in some cells but allowed others to cycle normally. The additive effect of retroviruses encoding p16INK4a and p21CIP1 could be due to an ability of these inhibitors to target different cyclin–Cdk complexes or to their different modes of action. Indeed, Alcorta et al. [7] reported that in senescent human diploid fibroblasts, most of the Cdk4/Cdk6 is associated with p16INK4a while the remainder is found in complexes with p21CIP1. The most logical explanation for our results, however, would be that it is the degree to which Cdk activity can be inhibited that is critical. Thus, there appears to be a good correlation between the degree of growth arrest and inhibition of DNA synthesis observed with different inhibitors and the extent to which the cells stain positively for SA-β-gal. Moreover, simply increasing the dose of the Ant–p16 peptide can achieve a more complete arrest and a higher proportion of SA-β-gal positivity. Nevertheless, as both p21CIP1 and p16INK4a accumulate in senescent cells, we favour a model in which both of these Cdk inhibitors are critical executors of the Hayflick limit [1] and presumably of a mechanism that acts in vivo to prevent immortalisation and neoplasia. The data also suggest that features previously regarded as peculiar to senescent cells can be reproduced by imposing a sustained G1 arrest in the continued presence of serum mitogens.
Materials and methods Cell culture and retroviral infection TIG-3 cells were cultured in DMEM containing 10% foetal calf serum and passaged at a 1:4 split ratio. To permit uptake of ecotropic retroviruses, cells were infected with an amphotropic retrovirus (pWXLNeo-Eco) encoding the mouse basic amino acid transporter, which serves as the cell surface receptor for mouse ecotropic retroviruses [15,19], and selected in 200 µg/ml G418. Pools of G418-resistant cells were infected with ecotropic retroviruses prepared by transient transfection of BOSC 23 cells with pBABE-puro or pBABE-hygro vectors [17,20] and selected 48 h post-infection in medium containing 2.5 µg/ml puromycin or 75 µg/ml hygromycin.
Acknowledgements We are very grateful to Manuel Serrano and Ignacio Palmero for openly discussing their data and providing the PAI-1 probe and retroviral vector encoding the ecotropic receptor. We also thank Bruno Amati and Hartmut Land for recombinant viruses and Nic Jones and Harmut Land for helpful comments on the manuscript. M.S. is supported by a grant from the Swedish Medical Research Council.
References 1. Hayflick L, Moorhead PS: The serial cultivation of human diploid cell strains. Exp Cell Res 1961, 25:585-621.
2. Goldstein S: Replicative senescence: the human fibroblast comes of age. Science 1990, 249:1129-1133. 3. Stein GH, Dulic V: Origins of G1 arrest in senescent human fibroblasts. BioEssays 1995, 17:537-543. 4. Smith JR, Pereira-Smith OM: Replicative senescence: implications for in vivo aging and tumor suppression. Science 1996, 273:63-67. 5. Noda A, Ning Y, Venable SF, Pereira-Smith OM, Smith JR: Cloning of senescent cell-derived inhibitors of DNA synthesis using an expression screen. Exp Cell Res 1994, 211:90-98. 6. Tahara H, Sato E, Noda A, Ide T: Increase in expression level of p21sdi1/cip1/waf1 with increasing division age in both normal and SV40-transformed human fibroblasts. Oncogene 1995, 10:835-840. 7. Alcorta DA, Xiong Y, Phelps D, Hannon G, Beach D, Barrett JC: Involvement of the cyclin-dependent kinase inhibitor p16 (INK4a) in replicative senescence. Proc Natl Acad Sci USA 1996, 93:13742-13747. 8. Hara E, Smith R, Parry D, Tahara H, Stone S, Peters G: Regulation of p16CDKN2 expression and its implications for cell immortalisation and senescence. Mol Cell Biol 1996, 16:859-867. 9. Loughran O, Malliri A, Owens D, Gallimore PH, Stanley MA, Ozanne B, et al.: Association of CDKN2A/p16INK4A with human head and neck keratinocyte replicative senescence: relationship of dysfunction to immortality and neoplasia. Oncogene 1996, 13:561-568. 10. Reznikoff CA, Yeager TR, Belair CA, Savelieva E, Puthenveettil JA, Stadler WM: Elevated p16 at senescence and loss of p16 at immortalisation in human papillomavirus 16 E6, but not E7, transformed human uroepithelial cells. Cancer Res 1996, 56:2886-2890. 11. Robetorye RS, Nakanishi M, Venable SF, Pereira-Smith OM, Smith JR: Regulation of p21Sdi1/Cip1/Waf1/mda-6 and expression of other cyclin-dependent kinase inhibitors in senescent human cells. Mol Cell Diff 1996, 4:113-126. 12. Sherr CJ, Roberts JM: Inhibitors of mammalian G1 cyclindependent kinases. Genes Dev 1995, 9:1149-1163. 13. Mu X-C, Higgins PJ: Differential growth state-dependent regulation of plasminogen activator inhibitor type-1 expression in senescent IMR-90 human diploid fibroblasts. J Cell Physiol 1995, 165:647-657. 14. Dimri GP, Lee X, Basile G, Acosta M, Scott G, Roskelley C, et al.: A novel biomarker identifies senescent human cells in culture and in aging skin in vivo. Proc Natl Acad Sci USA 1995, 92:9363-9367. 15. Serrano M, Lin AW, McCurrach ME, Beach D, Lowe SW: Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell 1997, 88:593-602. 16. Fåhraeus R, Paramio JM, Ball KL, Laín S, Lane DP: Inhibition of pRb phosphorylation and cell-cycle progression by a 20-residue peptide derived from p16CDKN2/INK4A. Curr Biol 1996, 6:84-91. 17. Morgenstern JP, Land H: Advanced mammalian gene transfer: high titre retroviral vectors with multiple drug selection markers and a complementary helper-free packaging cell line. Nucleic Acids Res 1990, 18:3587-3596. 18. Uhrbom L, Nistér M, Westermark B: Induction of senescence in human malignant glioma cells by p16INK4A. Oncogene 1997, 15:505-514. 19. Wang H, Kavanaugh MP, North RA, Kabat D: Cell-surface receptor for ecotropic murine retroviruses is a basic amino-acid transporter. Nature 1991, 352:729-732. 20. Pear, WS, Nolan, GP, Scott, ML, Baltimore, D: Production of hightiter helper-free retroviruses by transient transfection. Proc Natl Acad Sci USA 1993, 90:8392-8396. 21. Palmero I, McConnell B, Parry D, Brookes S, Hara E, Bates S, et al.: Accumulation of p16INK4a in mouse fibroblasts as a function of replicative senescence and not of retinoblastoma gene status. Oncogene 1997, 15:495-503.
351
Inhibitors of cyclin-dependent kinases induce features of replicative senescence in early passage human diploid fibroblasts Beth B. McConnell, Maria Starborg, Sharon Brookes and Gordon Peters
Address: Imperial Cancer Research Fund, P.O. Box 123, 44 Lincoln’s Inn Fields, London WC2A 3PX, UK. Correspondence: Gordon Peters E-mail: [email protected] Received: 3 December 1997 Revised: 16 January 1998 Accepted: 28 January 1998 Published: 2 March 1998 Current Biology 1998, 8:351–354 http://biomednet.com/elecref/0960982200800351 © Current Biology Ltd ISSN 0960-9822
Results and discussion TIG-3 strain human diploid fibroblasts, which typically undergo 60 to 70 PDs in tissue culture, were infected at PD 40–44 with recombinant retroviruses encoding different Cdk inhibitors (see Materials and methods). These included p21CIP1 and p27KIP1, which interact with multiple cyclin–Cdk complexes, plus p16INK4a and p15INK4b, two members of the INK4 family which specifically inhibit the
Figure 1 (b)
15 Relative number of cells
Ve D cto ay r D 6 ay Se 12 ne sc e
nt
(a)
p15INK4b p16INK4a p21CIP1
Vector p15INK4b p16INK4a p21CIP1 p27KIP1
10
5
0 6
p27KIP1
9
12
Days post-infection
IP 1
ppRb pRb
(d)
Proportion of cells incorporating BrdU (%)
(c)
Ve p1 cto 5 r p1 INK4 p26 INK b 1 4a p2 CIP1 7K
30
20
10
P1
KI
a
1C p2 IP1 7
p2
b
K4
K4
p1
p1
Current Biology
6 IN
ct o 5 IN r
0
Ve
After a limited number of population doublings (PDs), cultures of normal mammalian diploid cells undergo an irreversible growth arrest known as replicative senescence [1]. As well as contributing to cellular ageing, senescence is viewed as an important mechanism of tumour suppression by preventing the emergence of immortal cell clones [2–4]. Senescent cells have a number of characteristics that distinguish them from cycling or quiescent cells including elevated levels of two cyclin-dependent kinase (Cdk) inhibitors, p16INK4a and p21CIP1 [5–11]. Here, we demonstrate that both of these Cdk inhibitors, as well as other members of their protein families (the INK4 and CIP/KIP families, respectively [12]), induce several facets of the senescent phenotype when ectopically expressed in young human diploid fibroblasts. These include a reduced proliferative capacity, an altered size and shape, the presence of underphosphorylated retinoblastoma protein (pRb), increased expression of plasminogen activator inhibitor (PAI-1) and the appearance of senescence-associated b-galactosidase (SA-b-gal) activity [2,3,13–15]. A 20 amino acid peptide from p16INK4a that inhibits Cdks active in the G1 phase of the cell cycle [16] produces similar effects in a dosedependent manner suggesting that, in primary fibroblasts, inhibition of G1-specific Cdk activity is sufficient to induce phenotypic changes that normally occur at the end of their finite lifespan.
Growth arrest of human diploid fibroblasts by ectopic expression of Cdk inhibitors. (a) Lysates were prepared from TIG-3 cells (PD 40–44) 6 or 12 days after infection with recombinant retroviruses encoding p15INK4b, p16INK4a, p21CIP1 or p27KIP1 and the level of the expressed Cdk inhibitor in each lysate was compared to those present in cells infected with the empty vector control (Vector) and from TIG-3 cells that had reached the end of their finite lifespan (~65 PD; Senescent). Equivalent amounts of cellular proteins were fractionated by gel electrophoresis and analysed by immunoblotting [21] with primary antibodies against p15INK4b (Santa Cruz #sc-612), p16INK4a (DCS50), p21CIP1 (Santa Cruz #sc-397) and p27KIP1 (Santa Cruz #sc528). (b) Cells were plated at 1 × 104 cells per well, in a 24-well plate, and at 6, 9 and 12 days after infection with the same retroviruses as in (a), triplicate wells were fixed and stained with 10% formalin and 0.1% crystal violet. The intracellular stain was eluted with 10% acetic acid and relative cell numbers determined by measuring the optical density of the eluates at 590 nm [15]. The results shown were derived from two separate experiments. (c) Immunoblotting for pRb (using the Pharmingen G3-245 antibody) was performed on lysates made six days after infection with the indicated retroviruses. The hyperphosphorylated form of pRb is indicated as ppRb. (d) Cells were labelled with 10 µM bromodeoxyuridine (BrdU; Amersham) for 4 h 6 days after infection with the indicated retroviruses, stained with fluorescein isothiocyanate (FITC)-conjugated anti-BrdU antibody and analysed by two-dimensional flow cytometry.
cyclin-D-dependent kinases Cdk4 and Cdk6 (reviewed in [12]). Retroviruses based on the pBABE-puro or pBABEhygro vectors [17] allowed selection of transduced cells within 4 to 6 days and infection efficiencies were between 50% and 75% as judged by the percentage of cells surviving
352
Current Biology, Vol 8 No 6
Figure 2
Vector
p2 7
C
KI P1
IP 1
(b)
p2 1
Ve
p1 6
IN K4 b IN K4 a
ct o Se r ne sc p1 en t 5
(a)
PAI-1
GAPDH
(c)
p16INK4a
100 Proportion of cells with SA-β-gal activity (%)
Expression of senescence markers in human diploid fibroblasts expressing Cdk inhibitors. (a) Total RNA was isolated from senescent cells or cells 6 days after infection either with a retrovirus encoding the indicated Cdk inhibitor or with the empty vector control (Vector) and 8 µg of each sample was separated by electrophoresis, blotted and hybridised with a PAI-1 cDNA probe. Transcript levels determined by phosphorimaging analysis were normalised to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) levels. (b) Cells were assayed for SA-β-gal activity in situ as previously described [14]. The examples shown are cells expressing vector alone, p16INK4a or p21CIP1 as indicated. Following staining for SA-β-gal activity, cell nuclei were visualised with Hoechst 33528 dye (5 µg/ml for 3 min) to determine the total number of cells in each field. (c) The proportion of cells expressing SA-β-gal activity was calculated by observing ~400 cells per sample in three separate experiments.
75 p21CIP1
50
25
selection. Expression of p15INK4b, p16INK4a, p21CIP1, and p27KIP1 was monitored by immunoblotting cell lysates taken at 6 and 12 days post-infection. The levels of ectopic expression were compared with the endogenous levels present in cells transduced with the empty vector and in TIG-3 cells that had reached the end of their finite lifespan (Figure 1a). In each case, ectopic expression of the Cdk inhibitor was at least 10-fold higher than in the vector-only controls and cells infected with the p16INK4a or p21CIP1 viruses were producing amounts of protein roughly equivalent to those found in senescent cells. As judged by cell proliferation assays, cells expressing the Cdk inhibitors had a significantly lower rate of division than cells transduced with the empty vector (Figure 1b). Fibroblasts expressing INK4 inhibitors were more effectively arrested than those expressing CIP/KIP proteins (Figure 1b) with p15INK4b causing complete growth arrest over the timescale of the experiment. Similarly, whereas pRb was predominantly in its hyperphosphorylated state in proliferating control cells, judged by its mobility in sodium dodecyl sulphate (SDS)-containing acrylamide gels, the Cdk inhibitors caused accumulation of the more rapidly migrating, underphosphorylated form (Figure 1c). Again, the INK4 proteins had a more pronounced effect than the CIP/KIP proteins. The same trend was noted in bromodeoxyuridine (BrdU) incorporation assays: only 3% of the cells expressing p15INK4b or p16INK4a were able to
IP
1
4a C
IP
p2 1 7K p + 16 IP1 p2 IN 1 K
a
1C
p2
K4
K4
5 IN
6 IN
p1
p1
Ve
ct
or
b
0
Current Biology
enter the DNA synthesis (S) phase of the cell cycle during the 4 hour labelling period compared with 15% and 8% of the cells expressing p21CIP1 and p27KIP1, respectively (Figure 1d). These differences in the extent to which cells were arrested by the Cdk inhibitors could be attributable to different levels of expression within the pools of infected cells or to differences in the molar concentration of each protein required to fully block G1 Cdk activity. As well as being growth arrested, the majority of the infected cells were enlarged and flattened in appearance (data not shown). Such morphological changes, which are often associated with senescence [2–4], were noticeable within 6 days of infection and persisted. The cells also had increased expression of the mRNA for PAI-1, which is involved in regulating the turnover of extracellular matrix proteins (Figure 2a). Levels of expression of the two PAI-1 mRNA transcripts, and in particular the smaller of the two, have been shown to be higher in senescent cells [13,15]. The amounts of the smaller PAI-1 transcript were 3–4-fold higher in cells expressing the Cdk inhibitors than in vectoronly controls and approached those found in newly senescent TIG-3 cells (Figure 2a). Infected cells were also tested for expression of SA-β-gal activity, a convenient marker that can distinguish between cells arrested in a senescent versus a quiescent state [14]. As shown in Figure 2b,c, ectopic expression of any of the Cdk inhibitors was sufficient to induce positive staining for SA-β-gal activity, although with
353
Brief Communication
The data we describe here reinforce the proposed connection between replicative senescence and the Cdk inhibitors p16INK4a and p21CIP1 by providing the first direct evidence for their ability to induce senescence-like features in early passage human diploid fibroblasts. Serrano et al. [15] recently described analogous phenotypic changes in human diploid fibroblasts transduced with retroviruses encoding
(a)
(c)
(d)
(b)
Untreated 2.5 µM
7.5 µM
5.0 µM
10.0 µM
(e) 80 Proportion of cells with SA-β-gal activity (%)
These results strongly suggest that the inhibition of G1 Cdk activity is a critical determinant of several facets of the senescent phenotype. To confirm this proposition, we used a 20-residue synthetic peptide corresponding to amino acids 84–103 of p16INK4a to directly inhibit Cdk4/Cdk6 activity in young TIG-3 cells. When coupled to a 16-amino-acid domain from the Drosophila Antennapedia protein, the p16INK4a peptide has been shown to penetrate biological membranes and to block the ability of Cdk4 and Cdk6 to phosphorylate pRb [16]. The Antennapedia–p16 peptide (Ant–p16 peptide), tagged with biotin at its amino terminus, was added to the medium of young TIG-3 cells at a range of concentrations (0–10 µM). Immunofluorescence detection of the biotin tag confirmed that the peptide was present in all cells (Figure 3a). As with introduction of the retrovirally expressed Cdk inhibitors, cells treated with the Ant–p16 peptide had a significantly slower rate of division whereas control cells treated with equimolar amounts of p16 and Antennapedia peptides, synthesised separately and mixed, grew at a normal rate (data not shown). The Ant–p16 peptide also inhibited BrdU incorporation in the treated cells in both a time-dependent and dose-dependent manner (Figure 3b). When cells were stained for SA-β-gal activity at various times following addition of the Ant–p16 peptide, the proportion of positive cells again depended on the concentration of added peptide, with approximately 80% of the cells staining blue after 6 days in 10 µM peptide (Figure 3c–e). At these high concentrations, however, there was a tendency for cells to aggregate and detach from the culture dish, an effect that was not apparent with the separate Antennapedia and p16 peptides. Nevertheless, these results imply that the total concentration of Cdk inhibitor has a direct impact on the completeness of the G1 arrest and the subsequent expression of senescence markers.
Figure 3
Proportion of cells incorporating BrdU (%)
varying degrees of effectiveness. SA-β-gal activity was detectable at 9 days post-infection and maximal by day 12. A higher proportion of cells positive for SA-β-gal was obtained with the INK4 proteins than with p21CIP1 or p27KIP1, suggesting that the ability of the Cdk inhibitors to induce senescence-associated markers was directly related to the extent to which they reduced cell growth and blocked entry into S phase. Significantly, co-infecting cells using retroviruses encoding p16INK4a and p21CIP1 with different selectable markers had an additive effect, with a higher proportion of cells positive for SA-β-gal than with cells expressing either inhibitor on its own (Figure 2c).
30
20 10 0
60 40 20 0
1
3 6 Days of incubation
Untreated 2.5 µM 5.0 µM 7.5 µM 10.0 µM
0
1 2 3 4 5 Days of incubation
6
7
Current Biology
Induction of senescence markers using the Ant–p16 peptide. (a) TIG-3 cells were treated for 24 h with 10 µM biotinylated Ant–p16 peptide and fixed with methanol/acetone (50:50) for 5 min at –20°C. Intracellular peptide was visualised using fluorescein-conjugated streptavidin at a dilution of 1:40. (b) Cells were either left untreated or treated with 2.5, 5.0, 7.5 or 10 µM Ant–p16 peptide, with fresh medium and peptide added every second day. On days 1, 3 and 6 after peptide addition, the cells were labelled with 10 µM BrdU for 4 h and assayed for BrdU incorporation by immunofluorescence. (c) Untreated TIG-3 cells and (d) cells treated with 5 µM Ant–p16 peptide for 6 days were assayed in situ for SA-β-gal activity as described [14]. (e) Cells were either left untreated or incubated with 2.5, 5.0, 7.5 or 10 µM Ant–p16 peptide and the proportion of cells with SA-β-gal activity was estimated 1, 3 and 6 days after peptide addition. Each point was calculated from counting 400 cells in two independent experiments.
activated Ras and ascribed the effects to the upregulation of p53, p21CIP1 and p16INK4a. Similar effects have also been reported in some cancer cell lines ectopically expressing p16INK4a alone [18] but the majority of established cell lines do not respond in this way, raising interesting questions regarding the difference between the G1 arrest commonly associated with ectopic expression of Cdk inhibitors and the panoply of changes induced by Cdk inhibitors in human diploid fibroblasts. One possibility is that such effects are only observed in diploid or ‘near-diploid’ cells and that most tumour cell lines have sustained a sufficient number of additional mutations to obscure or obviate the effects. Another intriguing point is the timescale required to develop some of the phenotypic characteristics of senescence. For example, with retroviral delivery of the inhibitors, SA-β-gal staining was not apparent until 9–10
354
Current Biology, Vol 8 No 6
days post-infection as previously noted by Serrano et al. [15] whereas the changes in cell size and expression of PAI-1 occurred earlier. Such observations suggest a need for the accumulation or decay of some intracellular factor(s). The interpretation of these events, however, depends on whether the ectopic expression of Cdk inhibitors caused a generalised slowing of the cell cycle in all infected cells or whether the variable levels of expression within the pools of cells caused complete arrest in some cells but allowed others to cycle normally. The additive effect of retroviruses encoding p16INK4a and p21CIP1 could be due to an ability of these inhibitors to target different cyclin–Cdk complexes or to their different modes of action. Indeed, Alcorta et al. [7] reported that in senescent human diploid fibroblasts, most of the Cdk4/Cdk6 is associated with p16INK4a while the remainder is found in complexes with p21CIP1. The most logical explanation for our results, however, would be that it is the degree to which Cdk activity can be inhibited that is critical. Thus, there appears to be a good correlation between the degree of growth arrest and inhibition of DNA synthesis observed with different inhibitors and the extent to which the cells stain positively for SA-β-gal. Moreover, simply increasing the dose of the Ant–p16 peptide can achieve a more complete arrest and a higher proportion of SA-β-gal positivity. Nevertheless, as both p21CIP1 and p16INK4a accumulate in senescent cells, we favour a model in which both of these Cdk inhibitors are critical executors of the Hayflick limit [1] and presumably of a mechanism that acts in vivo to prevent immortalisation and neoplasia. The data also suggest that features previously regarded as peculiar to senescent cells can be reproduced by imposing a sustained G1 arrest in the continued presence of serum mitogens.
Materials and methods Cell culture and retroviral infection TIG-3 cells were cultured in DMEM containing 10% foetal calf serum and passaged at a 1:4 split ratio. To permit uptake of ecotropic retroviruses, cells were infected with an amphotropic retrovirus (pWXLNeo-Eco) encoding the mouse basic amino acid transporter, which serves as the cell surface receptor for mouse ecotropic retroviruses [15,19], and selected in 200 µg/ml G418. Pools of G418-resistant cells were infected with ecotropic retroviruses prepared by transient transfection of BOSC 23 cells with pBABE-puro or pBABE-hygro vectors [17,20] and selected 48 h post-infection in medium containing 2.5 µg/ml puromycin or 75 µg/ml hygromycin.
Acknowledgements We are very grateful to Manuel Serrano and Ignacio Palmero for openly discussing their data and providing the PAI-1 probe and retroviral vector encoding the ecotropic receptor. We also thank Bruno Amati and Hartmut Land for recombinant viruses and Nic Jones and Harmut Land for helpful comments on the manuscript. M.S. is supported by a grant from the Swedish Medical Research Council.
References 1. Hayflick L, Moorhead PS: The serial cultivation of human diploid cell strains. Exp Cell Res 1961, 25:585-621.
2. Goldstein S: Replicative senescence: the human fibroblast comes of age. Science 1990, 249:1129-1133. 3. Stein GH, Dulic V: Origins of G1 arrest in senescent human fibroblasts. BioEssays 1995, 17:537-543. 4. Smith JR, Pereira-Smith OM: Replicative senescence: implications for in vivo aging and tumor suppression. Science 1996, 273:63-67. 5. Noda A, Ning Y, Venable SF, Pereira-Smith OM, Smith JR: Cloning of senescent cell-derived inhibitors of DNA synthesis using an expression screen. Exp Cell Res 1994, 211:90-98. 6. Tahara H, Sato E, Noda A, Ide T: Increase in expression level of p21sdi1/cip1/waf1 with increasing division age in both normal and SV40-transformed human fibroblasts. Oncogene 1995, 10:835-840. 7. Alcorta DA, Xiong Y, Phelps D, Hannon G, Beach D, Barrett JC: Involvement of the cyclin-dependent kinase inhibitor p16 (INK4a) in replicative senescence. Proc Natl Acad Sci USA 1996, 93:13742-13747. 8. Hara E, Smith R, Parry D, Tahara H, Stone S, Peters G: Regulation of p16CDKN2 expression and its implications for cell immortalisation and senescence. Mol Cell Biol 1996, 16:859-867. 9. Loughran O, Malliri A, Owens D, Gallimore PH, Stanley MA, Ozanne B, et al.: Association of CDKN2A/p16INK4A with human head and neck keratinocyte replicative senescence: relationship of dysfunction to immortality and neoplasia. Oncogene 1996, 13:561-568. 10. Reznikoff CA, Yeager TR, Belair CA, Savelieva E, Puthenveettil JA, Stadler WM: Elevated p16 at senescence and loss of p16 at immortalisation in human papillomavirus 16 E6, but not E7, transformed human uroepithelial cells. Cancer Res 1996, 56:2886-2890. 11. Robetorye RS, Nakanishi M, Venable SF, Pereira-Smith OM, Smith JR: Regulation of p21Sdi1/Cip1/Waf1/mda-6 and expression of other cyclin-dependent kinase inhibitors in senescent human cells. Mol Cell Diff 1996, 4:113-126. 12. Sherr CJ, Roberts JM: Inhibitors of mammalian G1 cyclindependent kinases. Genes Dev 1995, 9:1149-1163. 13. Mu X-C, Higgins PJ: Differential growth state-dependent regulation of plasminogen activator inhibitor type-1 expression in senescent IMR-90 human diploid fibroblasts. J Cell Physiol 1995, 165:647-657. 14. Dimri GP, Lee X, Basile G, Acosta M, Scott G, Roskelley C, et al.: A novel biomarker identifies senescent human cells in culture and in aging skin in vivo. Proc Natl Acad Sci USA 1995, 92:9363-9367. 15. Serrano M, Lin AW, McCurrach ME, Beach D, Lowe SW: Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell 1997, 88:593-602. 16. Fåhraeus R, Paramio JM, Ball KL, Laín S, Lane DP: Inhibition of pRb phosphorylation and cell-cycle progression by a 20-residue peptide derived from p16CDKN2/INK4A. Curr Biol 1996, 6:84-91. 17. Morgenstern JP, Land H: Advanced mammalian gene transfer: high titre retroviral vectors with multiple drug selection markers and a complementary helper-free packaging cell line. Nucleic Acids Res 1990, 18:3587-3596. 18. Uhrbom L, Nistér M, Westermark B: Induction of senescence in human malignant glioma cells by p16INK4A. Oncogene 1997, 15:505-514. 19. Wang H, Kavanaugh MP, North RA, Kabat D: Cell-surface receptor for ecotropic murine retroviruses is a basic amino-acid transporter. Nature 1991, 352:729-732. 20. Pear, WS, Nolan, GP, Scott, ML, Baltimore, D: Production of hightiter helper-free retroviruses by transient transfection. Proc Natl Acad Sci USA 1993, 90:8392-8396. 21. Palmero I, McConnell B, Parry D, Brookes S, Hara E, Bates S, et al.: Accumulation of p16INK4a in mouse fibroblasts as a function of replicative senescence and not of retinoblastoma gene status. Oncogene 1997, 15:495-503.