, . 184: 348–350 (1998)

EDITORIAL

WHY IS p53 PROTEIN STABILIZED IN NEOPLASIA?SOME ANSWERS BUT MANY MORE QUESTIONS!

1, 2 . 1*1Department of Molecular and Cellular Pathology, University of Dundee, Dundee, U.K.

2Department of Pathology, University of Umeå, Umeå, Sweden

SUMMARY

The p53 pathway provides a physiological system for integrating signals from diverse insults and eliciting adaptive cellular responsesthat include (but importantly are not restricted to) growth arrest and apoptosis. Defects in the pathway are prevalent in cancer, mostnotably being associated with mis-sense mutations in p53 itself. This leads to the inability of p53 to act as a transcription factor and thusto the non-occurrence of downstream events. Recent data indicate that the stability (and hence level) of p53 protein in cells is regulatedby its interaction with mdm2: this results in enhanced p53 degradation by ubiquitin-mediated events. Since mdm2 is itself regulated byp53, loss of function of p53 leads to lack of mdm2 and thus to p53 protein accumulation. This provides a mechanistic explanation forthe observation that p53 accumulation is associated with neoplasia. It may be that accumulation of p53 in the absence of p53 mutationcan occur as a consequence of mdm2 defects, as well as being a physiological response in many situations. Another recent developmentis the recognition of p53 homologues (p73 alpha, p73 beta, and KET) which have many sequence and probable structural features incommon with p53. It seems likely that this will reveal another layer of complexity in the control and regulation of p53 and its role inphysiology and pathology. ? 1998 John Wiley & Sons, Ltd.

J. Pathol. 184: 348–350, 1998.

KEY WORDS—p53; mdm2; protein–protein interaction; protein stability; p73; KET

The popularity of studies of p53 has been driven asmuch by the availability of reagents that allow thedetection of p53 protein accumulation in pathologicalstates as it has by the frequency of abnormalities of p53in cancer and its role in other physiological and patho-logical states. A key point of considerable utility andinterest is that p53 protein levels accumulate in cells aftera number of different external stresses and also as aconsequence of p53 gene mutation. Indeed, in many (butimportantly not all) circumstances, the immunohisto-chemical detection of p53 in tumours equates withmis-sense point mutation.1,2 A prospective study byDowell et al.3 as well as many retrospective analyses4have shown a clear association between p53 overexpres-sion and cancer, but the problems inherent in theimmunohistochemical detection of p53 are well docu-mented.1,2 Given that there is a very low but biochemi-cally detectable level of p53 in cells in vivo5 and thatinduction of p53 protein expression is seen in physiologi-cal and pathological states (the subject of many studiesby pathologists), the mechanistic basis of this proteinaccumulation warrants consideration. The paper byCruz et al. in the current issue of the journal6 highlights

the need to consider this issue, since the authors reportthat p53 accumulation may be seen in the suprabasalcells of oropharyngeal epithelium in pre-neoplasia andsuggest clinical utility of this phenotype when seen inconjunction with morphological abnormalities. The trueclinical value of this interesting observation awaits fur-ther study but recent reports are beginning to provide anintellectual framework for understanding how this andrelated phenomena occur. Concomitantly, yet otherreports illustrate just how complex the p53 pathwayreally is!Considerable data now place p53 at a nodal point in

the regulation of diverse stress responses.7–9 A range ofinsults, acting presumably via a series of different signaltransduction pathways, converge on p53 protein and, byits post-translational modification, alter its ability to actas a transcriptional regulator of downstream targetgenes. Some of these target genes are known but manyremain to be identified. However, the two best known(but certainly not the only) outcomes of p53 activationare cell-cycle arrest and apoptosis. Which of these eventsoccurs appears to be critically dependent on the actuallevel of p53 protein10,11 as well as cell type. Indeed, p53is a singularly dangerous protein because of its ability toinduce cell death and must be very, very tightly regu-lated. Multiple levels of control are known to influencethe activity of p53 as a transcriptional activator. Forexample, evidence points to the existence of p53 in bothactive and inactive forms.12 Furthermore, the ability of

*Correspondence to: Peter Hall, Department of Molecular andCellular Pathology, University of Dundee, Dundee, U.K. E-mail:p.a.hall@dundee.ac.uk

Contract grant sponsors: CRC; AICR; EU; Pathological Society ofGreat Britain and Ireland.

CCC 0022–3417/98/040348–03 $17.50 Received 7 October 1997? 1998 John Wiley & Sons, Ltd. Accepted 20 October 1997

p53 to transactivate downstream genes depends not juston p53, but on the sequences on which it acts, since notall p53-responsive elements are functionally equivalent.The level of p53 within a cell is determined by very

many factors, only some of which are well documented.The responsiveness of different cell types in vivo toinduction of p53 appears to be very tightly regulated,5 apoint not apparent from previous cell culture studies.Whether this profile is the same for other inducingagents or is determined by both cell type and insultremains uncertain. There is general agreement that p53level is not regulated by transcription of the p53gene—although in fact few data exist that pertain to thein vivo situation in the context of stress responses.Alternative splicing of p53 mRNA has been reported inmouse p53, but its functional significance and relevanceto man are not clear. Regulation of mRNA stability andcontrol of p53 translation is considered to be importantby some authors and is a burgeoning area.13,14 Post-translational modification of p53 by covalent linkage ofRNA and O-glycosylation is described but is of uncer-tain significance, although the latter is typical of a rangeof transcription factors. The modification of a numberof serine and threonine residues by phosphate is welldocumented and certainly of considerable importance inregulating the function of p53.7,15 The localization ofp53 protein within a cell is of potential significance.16,17Observations in pathological material showed that p53staining could be cytoplasmic rather than nuclear.Rather than being an artifact of fixation or processing,this now appears to be a highly significant observation.Moreover, it may be that the sub-cellular localization ofp53 protein is regulated, perhaps by interaction withother proteins or with RNA, and contributes to theglobal regulation of the activity of p53: if the protein isin the cytoplasm it cannot act as a transcription factor.18In many respects, there are similarities with the NF-êBtranscription factor that is clearly regulated by sub-cellular localization, with cytoplasmic retention beingmediated by interaction with the cytoplasmic proteinIêB. However, the physiological role of sub-cellularlocalization regulation of p53 is not yet fully elucidated.Interaction of p53 with other proteins is importantand very many potential interactions have beendocumented.7–9 Again the physiological relevance ofmany of these interactions remains uncertain, as are thepossible interactions and competitions between factors.One protein–protein interaction that is clearly critical

in the regulation of p53 function is with the oncogenemdm2.19 mdm2 protein interacts with a region at theN-terminus of p5320 and can block the ability of p53 totranscriptionally activate target genes.21,22 In addition,recent data indicate that this interaction targets p53protein to ubiquitin-mediated degradation.23,24 Sincemdm2 is itself a target for downstream activation byp53, there exists an exquisitely sensitive feedback loop(reviewed in Lane and Hall25). It would seem that basallevels of p53 activate mdm2 transcription and the result-ant mdm2 protein binds to p53, inactivating its tran-scriptional activating properties and targeting p53 fordegradation. This tonic loop can be perturbed by eventsthat inhibit the binding of mdm2 to p53, perhaps by

phosphorylation of either moiety by signal transductionpathways that derive from stress signals. This wouldlead to activation of further downstream genes (as theinhibition of mdm2 is lost) and increased levels of p53protein. Presumably at some later time the activation islost and the system returns to baseline, with mdm2inhibiting p53 function and targeting the degradation ofp53. This model then explains why p53 protein oftenaccumulates in tumours. Where there are mutations ofp53 and it cannot act as a transcription factor, mdm2 isnot up-regulated. In the absence of the mdm2-dependentstimulus to ubiquitin-mediated degradation, p53 proteinlevels accumulate. Direct evidence for this model comesfrom the elegant experiments of Midgley and Lane26 incell culture systems. Of course, in other situations thepathway may be compromised by other events such asabnormalities of mdm2, which are increasingly beingrecognized,27 or by involvement of other interactionsand events.The central role of mdm2 in p53 homeostasis is an

important observation and may indeed have therapeuticconsequences.28 The functions of mdm2 are complex.The N-terminus is clearly the p53 interaction domain.mdm2 also has the ability to bind RNA via the ringfinger domain at the C-terminus and has previouslyunrecognized nuclear export signals as well as the well-characterized nuclear import signal.29 The consequenceof this is that there is clear potential for mdm2 to act asa shuttle protein, being cycled in and out of the nucleus.The significance of this remains unclear as does therecent identification of an mdm2 homologue termedmdmX.30 What is more certain is that mdm2 (andpresumably mdmX) has roles in differentiation,development, and oncogenesis.31,32 Indeed, it is increas-ingly hard to think about the p53 pathway withoutconsidering mdm2/mdmX.If all the above was not sufficiently complex, the past

months have seen frenetic activity in p53 laboratories.Two groups, neither in the p53 field, have identified p53homologues in man, one termed p7333 and the otherKET.34 Both groups, using degenerate PCR approachesand library screening methods in a search for signaltransduction components, serendipitously identifiedhomologues of p53, something that many had suspectedand sought but none had previously found. Intriguingly,the homologues closely resemble p53 in having (1) apredicted N-terminal activation domain with conserva-tion of critical mdm2 contact residues, (2) a centralDNA binding domain (again with conservation ofknown critical DNA binding contact residues which arehot spots for mutation in p53), and (3) a C-terminalregion homologous to the oligomerization domain inp53. Both p73 and KET have additional C-terminalextensions of unknown function and both resemble amollusc sequence which may be ancestral to p53 as wellas p73 and KET. At this point, life becomes verycomplex as it seems that the homologues (and there maywell be others) have some restriction in tissue expression,KET, for example, being restricted to squamous epi-thelium by RT-PCR studies. It may be that hetero-oligomers may exist and that the homologues canpotentially interact with mdm2 (and maybe with

349EDITORIAL

? 1998 John Wiley & Sons, Ltd. , . 184: 348–350 (1998)

mdmX). Few functional data are yet available, but p73can bind DNA and thus transactivate the p53-responsive element in the p21waf1 gene, and has beenshown to induce apoptosis.35 Whether the profile ofDNA binding is the same as, or different from p53 willbe interesting to determine: do these homologues acti-vate the same, an overlapping, or a different set ofdownstream genes, and what is the transactivation spec-trum of potential hetero-oligomers? What the physi-ological activators of these homologues will be remainsuncertain, although preliminary data suggest that atleast p73 is not DNA damage-inducible. New reagentsthat allow the homologues to be assayed in a wide rangeof systems, including histology, are urgently needed toallow us to address these and the many related issues.Given these dramatic new observations, our picture of

p53, which was seemingly beginning to become clearer(!), is again thrown into great confusion. How theobservations of Cruz et al.6 should be mechanisticallyinterpreted is now perhaps very much less clear,although the operational clinical perspective of thisreport is certainly worthy of further study. The majorityof the extant data on the role of p53 and mdm2 in theprocesses of physiological adaptation, development, dif-ferentiation, pre-neoplasia, and neoplasia must now becritically re-examined and re-interpreted. Who couldhave guessed this twist in the p53 tail!

ACKNOWLEDGEMENTS

V. Save is supported by a grant to P. A. Hall from theDepartment of Health, and K. Nylander by the Lion’sCancer Research Foundation, Umeå University. Workin P. A. Hall’s laboratory is also supported by the CRC,the AICR, the EU, and the Pathological Society ofGreat Britain and Ireland.

REFERENCES1. Wynford-Thomas D. p53 in tumour pathology: can we trust immunohisto-

chemistry? J Pathol 1992; 166: 329–330.2. Hall PA, Lane DP. p53 in tumour pathology: can we trust

immunohistochemistry?—revisited! J Pathol 1994; 172: 1–4.3. Dowell SP, Derias N, Wilson POG, Lane DP, Hall PA. The clinical utility

of the immunocytochemical detection of p53 protein in cytological speci-mens. Cancer Res 1994; 54: 2914–2918.

4. Harris CC, Hollstein M. Clinical implications of the p53 tumor-suppressorgene. N Engl J Med 1993; 329: 1318–1327.

5. MacCallum DE, Hupp TR, Midgley CA, et al. The p53 response to ionisingradiation in adult and developing murine tissues. Oncogene 1996; 13:2575–2587.

6. Cruz IB, Snijders PJ, Meijer CJ, et al. p53 expression above the basal celllayer in oral mucosa is an early event of malignant transformation and haspredictive value for developing oral squamous cell carcinoma. J Pathol1998; 184..

7. Hall PA, Meek D, Lane DP. p53: integrating the complexity. J Pathol 1996;180: 1–5.

8. Ko LJ, Prives C. p53: puzzle and paradigm. Genes Dev 1996; 10: 1054–1072.9. Levine AJ. p53, the cellular gatekeeper for growth and division. Cell 1997;

88: 323–331.10. Chen X, Ko LJ, Jayararman L, Prives C. p53 levels, functional domains and

DNA damage determine the extent of the apoptotic response of tumourcells. Genes Dev 1996; 10: 2438–2451.

11. Lassus P, Ferlin M, Piette J, Hibner U. Anti-apoptotic activity of low levelsof wild-type p53. EMBO J 1996; 15: 4566–4573.

12. Hupp TR, Sparks A, Lane DP. Small peptides activate the latent sequence-specific DNA binding function of p53. Cell 1995; 83: 237–245.

13. Mosner J, Mummenbrauer T, Bauer C, Sczakiel G, Grosse F, Deppert W.Negative feedback regulation of wild-type p53 biosynthesis. EMBO J 1995;14: 4442–4449.

14. Fu L, Benchimol S. Participation of the human p53 3*UTR in translationalrepression and activation following gamma-irradiation. EMBO J 1997; 16:4117–4125.

15. Meek DW. Post-translational modification of p53. Semin Cancer Biol 1994;5: 203–210.

16. Ostermeyer AG, Runko E, Winkfield B, Ahn B, Moll UM. Cytoplasmicallysequestered wild-type p53 protein in neuroblastoma is relocated to thenucleus by a C-terminal peptide. Proc Natl Acad Sci USA 1996; 93:15 190–15 194.

17. Moll UM, Ostermeyer AG, Haladay R, Winkfield B, Frazier M, ZambettiG. Cytoplasmic sequestration of wild-type p53 protein impairs the G1checkpoint after DNA damage. Mol Cell Biol 1996; 16: 1126–1137.

18. Martinez JD, Craven MT, Joseloff E, Bowden GT. Regulation of DNAbinding and transactivation of p53 in nuclear localisation and phosphoryla-tion. Oncogene 1997; 14: 2511–2520.

19. Piette J, Neel H, Marechal V. mdm2: keeping p53 under control. Oncogene1997; 15: 1001–1010.

20. Chen J, Marechal V, Levine AJ. Mapping of the p53 and mdm2 interactiondomains. Mol Cell Biol 1993; 13: 4107–4114.

21. Momand J, Zambetti GP, Olson DC, George D, Levine AJ. The mdm-2oncogene product forms a complex with the p53 protein and inhibitsp53-mediated transactivation. Cell 1992; 69: 1237–1245.

22. Wu X, Bayle JH, Olson D, Levine AJ. The p53–mdm-2 autoregulatoryfeedback loop. Genes Dev 1993; 7: 1126–1132.

23. Kubbutat MH, Jones SN, Vousden KH. Regulation of p53 stability bymdm2. Nature 1997; 387: 299–303.

24. Haupt Y, Maya R, Kazaz A, Oren M. mdm2 promotes the rapid degrada-tion of p53. Nature 1997; 387: 296–299.

25. Lane DP, Hall PA. MDM2: arbiter of p53’s destruction. Trends BiochemSci. 1997; 22: 372–374.

26. Midgley CA, Lane DP. p53 dependent stability in tumour cells is notdetermined by mutation but is dependent on mdm2 binding. Oncogene. (inpress).

27. Schlott T, Reimer S, Jahns A, et al. Point mutations and nucleotideinsertions in the mdm2 zinc finger structure of human tumours. J Pathol1997; 182: 54–61.

28. Bottger A, Bottger V, Garcia-Echeverria C, et al. Molecular characteriza-tion of the mdm2–p53 interaction. J Mol Biol 1997; 269: 744–756.

29. Elenbaas B, Dobbelstein M, Roth J, Shenk T, Levine AJ. The MDM2oncoprotein binds specifically to RNA through its ring finger domain. MolMed 1996; 2: 439–451.

30. Shvarts A, Steegenga WT, Riteco N, et al. MDMX: as novel p53-bindingprotein with some functional properties of MDM2. EMBO J 1996; 15:5349–5357.

31. Lundgren K, Montes de Oca Luna R, McNeill YB, et al. Targetedexpression of MDM2 uncouples S phase from mitosis and inhibits mam-mary gland development independent of p53. Genes Dev 1997; 11: 714–725.

32. Dazard J-E, Augias D, Neel H, et al. MDM-2 protein is expressed indifferent layers of normal human skin. Oncogene 1997; 14: 1123–1128.

33. Kaghad M, Bonnet H, Yang A, et al. Monoallelically expressed gene relatedto p53 at 1p36, a region frequently deleted in neuroblastoma and otherhuman cancers. Cell 1997; 90: 809–819.

34. Schmale H, Bamburger C. A novel protein with strong homology to thetumour suppressor p53. Oncogene 1997; 15: 1363–1367.

35. Jost CA, Marin MC, Kaelin WG Jr. p73 is a human p53 related protein thatcan induce apoptosis. Nature 1997; 389: 191–194.

350 EDITORIAL

? 1998 John Wiley & Sons, Ltd. , . 184: 348–350 (1998)