Sci. Aging Knowl. Environ., 20 August 2003
Vol. 2003, Issue 33, p. pe24
[DOI: 10.1126/sageke.2003.33.pe24]


The Persistence of Senescence

Norman E. Sharpless

The author is in the Departments of Medicine and Genetics, Lineberger Cancer Center, University of North Carolina School of Medicine, Chapel Hill, NC 27599, USA. E-mail: norman_sharpless{at};2003/33/pe24

Key Words: p16INK4a • p53 • Rb • senescence • p21 • Ink4a/Arf

When rogue elements are brought into check, the persistence of this control matters. Whether they be mad dogs, plutonium-producing nuclear reactors, or malignant cells, one sleeps more comfortably with the knowledge that these dangers are permanently, rather than temporarily, addressed. Senescence is a potent anticancer mechanism, representing a barrier that most, if not all, would-be tumor cells must traverse on their path to malignant transformation. Therefore, two recent publications (1, 2) regarding the durability of senescence are of interest not only to those who study aging, but to those who study cancer as well (see "Dangerous Liaisons").


The generally agreed-on model is that senescence is induced by a variety of stimuli, including telomeric shortening, some forms of DNA damage (for example, oxidative stress), and the activation of certain oncogenes [reviewed in (3) and see "More Than a Sum of Our Cells"]. Senescence requires the activation of the retinoblastoma (Rb) and/or p53 pathways, named for the tumor suppressor/transcriptional regulatory proteins Rb and p53, both of which inhibit cell proliferation. Regulators of these pathways, such as p16INK4a, p21, and ARF, are crucial in determining the onset of senescence (4-6) (Fig. 1) (see "Random Acts"). p53 induces senescence in part by stimulating the expression of p21, a cyclin-dependent kinase inhibitor (CDKI). CDKIs inhibit progression through the cell cycle by inhibiting kinases that phosphorylate and thereby inactivate Rb. The activity of p53 is predominantly mediated by inhibiting its MDM2-mediated degradation, and p53 is stabilized by diverse stimuli, including DNA damage signals (such as those resulting from oxidative stress or telomeric shortening) or oncogene activation. The stabilization of p53 by oncogenes is in part mediated by ARF (also designated p14ARF in the human or p19ARF in the mouse), which binds to MDM2, thereby inhibiting the destruction of p53 (7, 8). Another CDKI, p16INK4a, increases markedly in senescent cells and correlates with increasing Rb hypophosphorylation during this process (5, 6). The regulation of p16INK4a in senescence is not well understood, although it appears to be induced by several stimuli, including oncogene activation and cell culture. Therefore, p53-p21 and p16INK4a regulate Rb and the process of senescence in concert but likely in response to different stimuli.

View larger version (6K):
[in this window]
[in a new window]
Fig. 1. The p53 and Rb pathways mediate senescence. p53 is activated by various stimuli, including telomere shortening, certain forms of DNA damage, and p14ARF expression (which in turn results from oncogene activation). Increased p53 causes p21-dependent and -independent forms of growth arrest; for simplicity, only G1 arrest mediated by p21 expression is shown. Expression of p21 inhibits phosphorylation of Rb family members, with repression of E2F activity and senescence. Expression of p16INK4a is increased by telomere-independent signals such as MAPK activation. Other stimuli that induce (or inhibit repression of) p16INK4a are not fully characterized. Expression of p16INK4a likewise inhibits Rb phosphorylation with attendant induction of senescence. Various components of these pathways were targeted (shown in red) by Sage et al. (1) and Beausejour et al. (2). GSE, genetic suppressor element of p53; siRNA, small interfering RNA directed against p16INK4a.

Senescence differs from other forms of growth arrest, such as quiescence, in two important ways (Fig. 2). First, senescence in somatic cells is thought to be irreversible, and it therefore represents a specialized form of terminal differentiation. Second, it encompasses certain phenotypic alterations such as characteristic morphological changes and the expression of senescence-associated-{beta}-galactosidase (SA-{beta}-Gal) activity. Recently, senescence has been shown to correlate with the establishment of an unusual form of heterochromatin that is present in discrete nuclear foci [called SA-heterochromatic foci (SAHF)] (9). In aggregate, these data suggest that senescence results from the durable repression of promoters associated with growth. This repression is enforced by the construction of stable heterochromatin-like complexes, the formation of which is directed in part by hypophosphorylated Rb.

View larger version (8K):
[in this window]
[in a new window]
Fig. 2. Properties of senescence versus quiescence. Note that senescence is considered to be "permanent" in the setting of persistent and unperturbed Rb and p53 function, as discussed in the text.

The Role of Rb

To determine whether senescence requires persistent Rb function, members of the Jacks and Campisi labs created cultures of senescent cells by standard methods. In these cultures, the authors then inactivated the p16INK4a-Rb pathway using several different approaches (Fig. 1). Sage et al. (1) deleted Rb through the expression of CRE recombinase in murine embryo fibroblasts (MEFs) that harbor LoxP-flanked Rb alleles. When Rb is excised, senescent MEFs begin to proliferate again, escaping senescence. The authors used video microscopy to document the fact that senescent cells actually reentered the cell cycle upon loss of Rb, excluding the possibility that the process merely facilitated the escape of nonsenescent clones. It is important to note that in addition to Rb, there are two other Rb family members, p107 and p130, which also regulate E2F activity and can partially compensate for the loss of Rb function (10) (Fig. 1). E2F is a member of a family of transcription factors that function with Rb, p130, and p107 in the regulation of cell proliferation, apoptosis, and differentiation. Acute excision of the Rb gene in senescent MEFs caused p107 expression to increase rapidly, and in a minority of these MEF cultures, the cells eventually reentered a p107-mediated senescence. Perhaps most worrisome, though, is that the majority of formerly senescent cultures that underwent Rb excision did not reenter senescence, but, rather, continued to proliferate and were presumably immortal. In some of these senescence-resistant cultures, inactivation of the ARF-p53 pathway was demonstrated--a frequent event in cultured MEFs and a common mechanism of their immortalization (4). Therefore, even if p107 can compensate for Rb in some settings, nascent cancer cells that are forced into an Rb-dependent senescence might likewise acquire other immortalizing mutations during a proliferative burst that occurs after Rb inactivation.

In the work by Beausejour et al. (2), on the other hand, the organism appears a somewhat safer place. Here, the authors used high-efficiency lentiviral vectors to deliver various oncogenic molecules to senescent human cells. Lentiviral vectors were used because they are capable of infecting nondividing cells as opposed to other more traditionally used retroviral vectors. Using this approach, the authors discovered that senescence appears to come in at least two varieties, which can be distinguished by the presence or absence of significant p16INK4a expression. Some senescent lines, such as human BJ fibroblasts (low p16INK4a), could be driven back into the cell cycle by perturbation of p53 signaling alone. The authors believe that these cells are limited from further growth by the antiproliferative activities of p53 resulting from telomeric shortening. In agreement with these findings, p53 abrogation in either human (11) or murine (12) fibroblasts by other mechanisms has similarly been shown to reinduce proliferation. In contrast, other senescent lines, such as human WI-38 cells (high p16INK4a), could not be persuaded to proliferate, even when a highly oncogenic molecule such as SV40 large T antigen (T Ag) was expressed. T Ag inactivates p53 and Rb (as well as the other Rb family members, p107 and p130), yet senescent WI-38 cells expressing T Ag synthesized DNA (that is, entered S phase) but were unable to complete the cell cycle. The authors then demonstrated that the form of senescence that is resistant to Rb inactivation results from the expression of p16INK4a. As proof of this, the authors showed that WI-38 cells could be rendered resistant to permanent arrest by expressing in these cells a small interfering RNA that blocked expression of p16INK4a before the onset of senescence. In this case, WI-38 cells behaved like BJ fibroblasts; that is, they senesced with low p16INK4a and high p21, and this arrest could be reversed by p53 inactivation. From these results, the authors conclude that p16INK4a-mediated senescence differs qualitatively from that induced by p53 and its downstream effector p21.

How to reconcile the two views? First, it is formally possible that T Ag is not functionally equivalent to Rb deletion. It may be that T Ag cannot completely inactivate Rb function, and, therefore, some residual Rb pathway function in WI-38 cells responds to the high level of p16INK4a expression, thereby blocking cell cycle reentry. More likely, however, is that the result reflects differences between the senescence of murine and human fibroblasts. In both species, senescence shares the common features of flattened morphology, expression of SA-{beta}-Gal activity, and SAHF formation. In MEFs, however, senescence is a more dynamic process characterized by the frequent emergence of spontaneously immortalized clones ("escape"), persistent DNA synthesis even in the setting of G1 arrest, and p53-mediated growth suppression that requires both DNA damage signals and ARF expression (4, 13, 14; see also Hornsby Perspective). Although an increase in p16INK4a expression does correlate with the onset of their senescence, MEFs lacking p16INK4a enter senescence with kinetics identical to those of wild-type cells (15, 16). Therefore, MEF senescence appears to be more similar to senescence of the p53-mediated type of BJ fibroblasts, rather than the p16INK4a-dependent senescence of WI-38 cells. This p16INK4a-mediated senescence in human cells appears to be a more static phenomenon, with a greatly decreased frequency of spontaneous escape as compared to MEFs and no detectable ongoing DNA synthesis. Consistent with this view, the appearance of SAHF in murine cells and BJ fibroblasts is decreased relative to that seen in p16INK4a-dependent senescence (9). This result also helps to explain the observation that although p16INK4a-deficient cells enter senescence in a way identical to wild-type cells, p16INK4a-deficient MEFs escape senescence with an increased frequency (16). Therefore, p53 and p21 may induce a potentially reversible senescence, the stability of which can be enhanced by p16INK4a expression.

The Sage/Jacks (1) data strongly suggest that the reversible, p53-mediated form of senescence requires ongoing Rb synthesis. The role of Rb in the irreversible, p16INK4a-mediated form is unclear and likely depends on the status of Rb within those cells. Perhaps Rb is incorporated into complex chromatin in permanently senesced cells and, in addition, perhaps such Rb is highly stable. In such a setting, therefore, little or no new production of Rb would be required to maintain growth arrest. Likewise, Rb in this setting might be unavailable for T Ag binding, and thus this type of senescence would be resistant to that highly oncogenic molecule. Another possibility, however, is that Rb is no longer required for growth arrest after the establishment of dense SAHF. It has been previously demonstrated that cells lacking all three Rb family members are highly resistant to the induction of nearly any type of growth arrest (17, 18), but this does not mean that the expression of these proteins is required for the maintenance of this p16INK4a-mediated senescence of human cells.

Different Kinds of Senescence In Vivo?

These observations bring into sharp relief a nettlesome question: If all cyclin-dependent kinase inhibitors effect growth arrest by blocking the phosphorylation of Rb family proteins, how can the result of p16INK4a expression differ from that of p21 (Fig. 3)? It has been noted that p21 concentrations decline in the final stages of senescence, whereas those of p16INK4a markedly increase (5, 6). Therefore, one possible explanation for the differing effects of p21 and p16INK4a is that the expression of p21 is more transient than that of p16INK4a (Fig. 3A). That is, perhaps temporary Rb hypophosphorylation, which would occur while p21 concentrations are still high, represses E2F without wholesale chromatin alterations, causing a reversible growth arrest. More durable Rb activation, via p16INK4a according to this model, would lead to the establishment of repressive chromatin structures that "lock in" the repression of E2F and engender permanent senescence. Although a minimum period of p16INK4a expression in growth-arrested cells is probably required to establish permanent senescence (19), this temporal model cannot be wholly correct. For example, we have observed that density-arrested murine cells demonstrate Rb hypophosphorylation but will reenter the cell cycle when split to subconfluence, even after weeks of growth arrest (20). Likewise, in vivo, it is believed that many cells, such as memory T cells, can spend years in G1 arrest, only to reenter the cell cycle when called on by antigen stimulation. In these settings, a reversible but prolonged period of Rb hypophosphorylation does not appear sufficient to induce senescence. In fact, such types of long-lived but reversible growth arrest are likely crucial to normal stem cell function and organ homeostasis (21). Alternatively, it is possible that p16INK4a is a more effective inhibitor of E2F than p21 (Fig. 3B), or that either p16INK4a or p21 exerts Rb-independent effects on cell growth (Fig. 3C). At present, however, there are little data to support either of these models, and the mechanistic basis for the difference between the types of senescence described by Beausejour et al. (2) remains unexplained.

View larger version (6K):
[in this window]
[in a new window]
Fig. 3. Models to explain the difference between p53/p21-mediated and p16INK4a-mediated senescence. (A) Temporal model. The expression of p16INK4a is more durable than that of p21, allowing not only for E2F inhibition, but also for the formation of extensive alterations of chromatin architecture, resulting in permanent senescence. (B) Efficacy model. p16INK4a more effectively represses E2F than does p21, implying that permanent senescence requires the surpassing of some threshold of Rb hypophosphorylation. (C) Novel targets model. Either p21 or p16INK4a inhibits (activates) some pathway (designated "X" or "Y") that is independent of Rb. In this model, senescence would result from the combined effects on Rb and either X or Y.

The role of senescence in an intact organism remains controversial. Except for the predisposition to neoplasia, animals lacking p16INK4a are seemingly normal (15, 16), which suggests that senescence plays little or no developmental role, instead serving predominantly to prevent malignancy. This assertion has been considered problematic because it has been argued that evolution should not favor the selection of anticancer mechanisms, as cancer is predominantly a disease of older organisms that ensues well after their reproductive primes. Startling evidence born of attempts in clinical oncology to measure minimal residual disease in patients treated for leukemia/lymphoma, however, has called that notion into question. Using polymerase chain reaction (PCR)-based methods to detect oncogenic fusions such as BCR-ABL, IGH-BCL2, and AML-ETO, several groups have demonstrated that as many as one in three of the "negative controls" for those assays (that is, healthy cancer-free individuals) harbor oncogenic translocations, yet the vast majority of these individuals never go on to develop cancer (22-26). These observations imply that anticancer mechanisms, such as senescence, are at work even in neonates (26), in the absence of aging or carcinogen exposure. Furthermore, it has been suggested that the expression of these anticancer mechanisms, in their effort to stem malignancy, may compromise regenerative potential and organismal fitness, thereby contributing to aging (27, 28). The marked accumulation of p16INK4a in a variety of aged tissues suggests that senescence may contribute to aging, but to date this accumulation has only been shown to be correlative. The demonstration by Sage et al. (1) that some forms of senescence can be reversed by Rb loss may help explain how some premalignant cells escape senescence to produce cancer. The demonstration by Beausejour et al. (2), that p16INK4a appears to enforce a specialized, permanent form of senescence, makes one wonder further if this molecule does, in fact, contribute to human aging.

August 20, 2003
  1. J. Sage, A. L. Miller, P. A. Perez-Mancera, J. M. Wysocki, T. Jacks, Acute mutation of retinoblastoma gene function is sufficient for cell cycle re-entry. Nature 424, 223-228 (2003). [CrossRef][Medline]
  2. C. M. Beausejour, A. Krtolica, F. Galimi, M. Narita, S. W. Lowe, P. Yaswen, J. Campisi, Reversal of human cellular senescence: Roles of the p53 and p16 pathways. EMBO J. 22, 4212-4222 (2003).[Abstract/Free Full Text]
  3. J. Campisi, Cellular senescence as a tumor-suppressor mechanism. Trends Cell Biol. 11, S27-S31 (2001).[CrossRef][Medline]
  4. T. Kamijo, F. Zindy , M. F. Roussel, D. E. Quelle, J. R. Downing, R. A. Ashmun, G. Grosveld, C. J. Sherr, Tumor suppression at the mouse INK4a locus mediated by the alternative reading frame product p19ARF. Cell 91, 649-659 (1997).[CrossRef][Medline]
  5. G. H. Stein, L. F. Drullinger, A. Soulard, V. Dulic, Differential roles for cyclin-dependent kinase inhibitors p21 and p16 in the mechanisms of senescence and differentiation in human fibroblasts. Mol. Cell. Biol. 19, 2109-2117 (1999).[Abstract/Free Full Text]
  6. D. A. Alcorta, Y. Xiong, D. Phelps, G. Hannon, D. Beach, J. C. Barrett, Involvement of the cyclin-dependent kinase inhibitor p16 (INK4a) in replicative senescence of normal human fibroblasts. Proc. Natl. Acad. Sci. U.S.A. 93, 13742-13747 (1996).[Abstract/Free Full Text]
  7. J. Pomerantz, N. Schreiber-Agus, N. J. Liegeois, A. Silverman, L. Alland, L. Chin, J. Potes, K. Chen, I. Orlow, H. W. Lee, C. Cordon-Cardo, R. A. DePinho, The Ink4a tumor suppressor gene product, p19Arf, interacts with MDM2 and neutralizes MDM2's inhibition of p53. Cell 92, 713-723 (1998).[CrossRef][Medline]
  8. Y. Zhang, Y. Xiong, W. G. Yarbrough, ARF promotes MDM2 degradation and stabilizes p53: ARF-INK4a locus deletion impairs both the Rb and p53 tumor suppression pathways. Cell 92, 725-734 (1998).[CrossRef][Medline]
  9. M. Narita, S. Nunez, E. Heard, M. Narita, A. W. Lin, S. A. Hearn, D. L. Spector, G. L. Hannon, S. W. Lowe, Rb-mediated heterochromatin formation and silencing of E2F target genes during cellular senescence. Cell 113, 703-716 (2003).[CrossRef][Medline]
  10. M. Classon, E. Harlow, The retinoblastoma tumour suppressor in development and cancer. Nat. Rev. Cancer 2, 910-917 (2002).[CrossRef][Medline]
  11. V. Gire, D. Wynford-Thomas, Reinitiation of DNA synthesis and cell division in senescent human fibroblasts by microinjection of anti-p53 antibodies. Mol. Cell. Biol. 18, 1611-1621 (1998).[Abstract/Free Full Text]
  12. A. M. Dirac, R. Bernards, Reversal of senescence in mouse fibroblasts through lentiviral suppression of p53. J. Biol. Chem. 278, 11731-11734 (2003).[Abstract/Free Full Text]
  13. K. M. Frank, N. E. Sharpless, Y. Gao, J. M. Sekiguchi, D. O. Ferguson, C. Zhu, J. P. Manis, J. Horner, R. A. DePinho, F. W. Alt, DNA ligase IV deficiency in mice leads to defective neurogenesis and embryonic lethality via the p53 pathway. Mol. Cell 5, 993-1002 (2000).[CrossRef][Medline]
  14. S. Parrinello, E. Samper, A. Krtolica, J. Goldstein, S. Melov, J. Campisi, Oxygen sensitivity severely limits the replicative lifespan of murine fibroblasts. Nat. Cell Biol. 5, 741-747 (2003).[CrossRef][Medline]
  15. P. Krimpenfort, K. C. Quon, W. J. Mooi, A. Loonstra, A. Berns, Loss of p16Ink4a confers susceptibility to metastatic melanoma in mice. Nature 413, 83-86 (2001).[CrossRef][Medline]
  16. N. E. Sharpless, N. Bardeesy, K. H. Lee, D. Carrasco, D. H. Castrillon, A. J. Aguirre, E. A. Wu, J. W. Horner, R. A. DePinho, Loss of p16Ink4a with retention of p19Arf predisposes mice to tumorigenesis. Nature 413, 86-91 (2001).[CrossRef][Medline]
  17. J. H. Dannenberg, A. van Rossum, L. Schuijff, H. te Riele, Ablation of the retinoblastoma gene family deregulates G(1) control causing immortalization and increased cell turnover under growth-restricting conditions. Genes Dev. 14, 3051-3064 (2000).[Abstract/Free Full Text]
  18. J. Sage, G. J. Mulligan, L. D. Attardi, A. Miller, S. Chen, B. Williams, E. Theodorou, T. Jacks, Targeted disruption of the three Rb-related genes leads to loss of G(1) control and immortalization. Genes Dev. 14, 3037-3050 (2000).[Abstract/Free Full Text]
  19. C. Y. Dai, G. H. Enders, p16 INK4a can initiate an autonomous senescence program. Oncogene 19, 1613-1622 (2000).[CrossRef][Medline]
  20. N. E. Sharpless, unpublished observation.
  21. T. Cheng, N. Rodrigues, H. Shen, Y. Yang, D. Dombkowski, M. Sykes, D. T. Scadden, Hematopoietic stem cell quiescence maintained by p21cip1/waf1. Science 287, 1804-1808 (2000).[Abstract/Free Full Text]
  22. S. Bose, M. Deininger, J. Gora-Tybor, J. M. Goldman, J. V. Melo, The presence of typical and atypical BCR-ABL fusion genes in leukocytes of normal individuals: biologic significance and implications for the assessment of minimal residual disease. Blood 92, 3362-3367 (1998).[Abstract/Free Full Text]
  23. Y. Liu, A. M. Hernandez, D. Shibata, G. A. Cortopassi, BCL2 translocation frequency rises with age in humans. Proc. Natl. Acad. Sci. U.S.A. 91, 8910-8914 (1994).[Abstract/Free Full Text]
  24. K. E. Summers, L. K. Goff, A. G. Wilson, R. K. Gupta, T. A. Lister, J. Fitzgibbon, Frequency of the Bcl-2/IgH rearrangement in normal individuals: implications for the monitoring of disease in patients with follicular lymphoma. J. Clin. Oncol. 19, 420-424 (2001).[Abstract/Free Full Text]
  25. C. Biernaux, M. Loos, A. Sels, G. Huez, P. Stryckmans, Detection of major bcr-abl gene expression at a very low level in blood cells of some healthy individuals. Blood 86, 3118-3122 (1995).[Abstract/Free Full Text]
  26. H. Mori, S. M. Colman, Z. Xiao, A. M. Ford, L. E. Healy, C. Donaldson, J. M. Hows, C. Navarrete, M. Greaves, Chromosome translocations and covert leukemic clones are generated during normal fetal development. Proc. Natl. Acad. Sci. U.S.A. 99, 8242-8247 (2002).[Abstract/Free Full Text]
  27. N. E. Sharpless, R. A. DePinho, p53: good cop/bad cop. Cell 110, 9-12 (2002).[CrossRef][Medline]
  28. J. Campisi, Cancer and ageing: rival demons? Nat. Rev. Cancer 3, 339-349 (2003).[CrossRef][Medline]
Citation: N. E. Sharpless, The Persistence of Senescence. Sci. SAGE KE 2003 (33), pe24 (2003).

The Search for Biomarkers of Aging: Next Stop INK4a/ARF Locus.
G. P. Dimri (2004)
Sci. Aging Knowl. Environ. 2004, pe40
   Abstract »    Full Text »

Science of Aging Knowledge Environment. ISSN 1539-6150