Somatic TP53 Mutations in the Era of Genome Sequencingperspectivesinmedicine.cshlp.org/content/early/2016/08/...2016/08/06 · p53g) (Courtois et al. 2004; Khouryand Bour-don2010;Marceletal.2011).Althoughthepre-cise

Somatic TP53 Mutations in the Era of GenomeSequencing

Pierre Hainaut1 and Gerd P. Pfeifer2

1University Grenoble Alpes, Institut Albert Bonniot, Institut National de la Santé et de la RechercheMédicale (INSERM), 823 Grenoble, France

2Center for Epigenetics, Van Andel Research Institute, Grand Rapids, Michigan 49503

Correspondence: [email protected]

Amid the complexity of genetic alterations in human cancer, TP53 mutation appears as analmost invariant component, representing by far the most frequent genetic alteration overall.Compared with previous targeted sequencing studies, recent integrated genomics studiesoffer a less biased view of TP53 mutation patterns, revealing that .20% of mutations occuroutside the DNA-binding domain. Among the 12 mutations representing each at least 1% ofall mutations, five occur at residues directly involved in specific DNA binding, four affect thetertiary fold of the DNA-binding domain, and three are nonsense mutations, two of them inthe carboxyl terminus. Significant mutations also occur in introns, affecting alternative splic-ing events or generating rearrangements (e.g., in intron 1 in sporadic osteosarcoma). Inaggressive cancers, mutation is so common that it may not have prognostic value (all thesecancers have impaired p53 function caused by mutation or by other mechanisms). In severalother cancers, however, mutation makes a clear difference for prognostication, as, forexample, in HER2-enriched breast cancers and in lung adenocarcinoma with EGFR muta-tions. Thus, the clinical significance of TP53 mutation is dependent on tumor subtype andcontext. Understanding the clinical impact of mutation will require integrating mutation-specific information (type, frequency, and predicted impact) with data on haplotypes and onloss of heterozygosity.

In 1989, 10 years after the discovery of the p53protein, studies by Baker et al. (1989), Nigroet al. (1989), and Takahashi et al. (1989) re-vealed that the TP53 gene was frequently mu-tated in many forms of human cancer. Morethan 25 years later, in the era of genome-widesequencing of cancer DNA, TP53 is confirmedas the most frequently mutated gene associatedwith cancer in general. This finding is one ofthe most surprising lessons from cancer ge-

nome sequencing. There is no a priori reasonwhy candidate gene-centered sequencing ap-proaches would have correctly identified oneunique gene among 20,000 genes as the mostfrequently mutated one. Indeed, a comparisonof significantly mutated genes (SMGs) detectedby whole-exome analysis in 15 types of solidtumors shows that TP53 is the most fre-quently mutated gene in 11 of them (gliobla-stoma, clear cell renal cell carcinoma, head and

Editors: Guillermina Lozano and Arnold J. Levine

Additional Perspectives on The p53 Protein available at www.perspectivesinmedicine.org

Copyright # 2016 Cold Spring Harbor Laboratory Press; all rights reservedAdvanced Online Article. Cite this article as Cold Spring Harb Perspect Med doi: 10.1101/cshperspect.a026179

1

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

on June 5, 2021 - Published by Cold Spring Harbor Laboratory Press http://perspectivesinmedicine.cshlp.org/Downloaded from

mailto:[email protected]:[email protected]:[email protected]://www.perspectivesinmedicine.orghttp://www.perspectivesinmedicine.orghttp://www.perspectivesinmedicine.orghttp://perspectivesinmedicine.cshlp.org/

neck squamous cell cancer, serous ovarian car-cinoma, non-small-cell lung cancer, small-celllung cancer, gastric cancer, hepatocellular carci-noma [HCC], cholangiocarcinoma, breast can-cer [BC], prostate cancer) (average: 25%–60%),whereas it ranks second to KRAS in pancreaticand colorectal cancer, and third to BRAF andNRAF in melanoma (Watson et al. 2013). Inhematopoietic malignancies, TP53 mutationsare less frequent (10%–15%) but are amongthe 10 most frequent SMGs in five of six docu-mented malignancies (the exception is pediatricacute lymphoblastic leukemia [Zhang et al.2012]).

The first studies on TP53 mutations focusedon tumors with allelic deletions at the TP53locus (17p13.1) and showed that mutations re-sulting in single amino acid substitutions werefrequent on the allele that was retained. Thisobservation was consistent with the theoreticalhallmarkof a tumor-suppressor gene and Knud-sen’s two-hit hypothesis. Soon, however, it wasnoted that tumors that retained both parentalalleles also frequently carried point mutationson one allele, suggesting a form of dominanteffect of the mutation. Indeed, independentof the allelic context, nuclear accumulation ofmutant p53 protein was detected in a widespectrum of primary and metastatic lesions,suggesting a selective retention of mutant p53during tumor development and dissemination(Bartek et al. 1990). In 1991, distinct patternsof somatic TP53 mutations were revealed, re-spectively, in UV-induced nonmelanoma skincancers (Brash et al. 1991) and in aflatoxin-re-lated HCC (Bressac et al. 1991). These mutationpatterns were consistent with the ones inducedby these mutagens in experimental assays, con-firming the idea that “carcinogens [can] leavefingerprints” in TP53 sequence as evidence oftheir etiological role in carcinogenesis (Vogel-stein and Kinzler 1992).

In subsequent years, the results literallysparked an industry of sequencing TP53 exonsusing the Sanger technique of chain-terminat-ing inhibitors in cancer tissues. From a few hun-dred at the turn of the 1990s, the number ofsomatic TP53 mutations identified in cancergrew steadily to more than 10,000 by the end

of the millennium (Hainaut and Hollstein2000). Soon, computational resources were de-veloped to compile, retrieve, and compare thealready identified somatic mutations. In 1994,two databases collecting and annotating muta-tions detected by sequencing and published inthe peer-reviewed literature were established,and they have been since maintained continu-ously: the TP53 database at the InternationalAgency for Research on Cancer (IARC), initiat-ed by Monica Hollstein and Curtis Harris (p53.iarc.fr) (Olivier et al. 2002), and the UMD-p53database, initiated by Thierry Soussi (p53.fr/index.html) (Beroud and Soussi 1998). Thesedatabases currently contain more than 30,000annotated mutations and provide a numberof flexible tools to analyze the distribution ofmutations across cancers. They have been thebasis of several comprehensive reviews discuss-ing the variability and heterogeneity of TP53mutation patterns with respect to mutagenicprocesses, cancer etiology, or potential impactson cancer biology (Greenblatt et al. 1994; Hai-naut and Hollstein 2000; Petitjean et al. 2007;Soussi 2014).

In recent years, the development of second-generation sequencing technologies (also callednext-generation sequencing [NGS]) has en-abled the systematic analysis of cancer exomesand genomes, expanding our understanding ofmutation patterns far beyond the knowledgederived from single genes such as TP53. How-ever, despite numerous studies reporting anassociation between somatic TP53 mutationsand poor prognosis and unfavorable treatmentoutcomes (Olivier et al. 2010; Robles and Har-ris 2010), the only accepted recommendationfor clinical testing to date is for chronic lympho-cytic leukemia (CLL), in which somatic TP53mutation is predictive of poor response to con-ventional therapies (Stilgenbauer et al. 2014).Thus, the clinical significance of somatic TP53mutations remains elusive in most forms ofcancer. In this review, we revisit our under-standing of TP53 mutation patterns in the lightof recent data generated by next-generation se-quencing, and we discuss how somatic TP53mutation testing may contribute to inform clin-ical practice.

P. Hainaut and G.P. Pfeifer

2 Advanced Online Article. Cite this article as Cold Spring Harb Perspect Med doi: 10.1101/cshperspect.a026179

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org


http://p53.iarc.frhttp://p53.iarc.frhttp://p53.iarc.frhttp://p53.fr/index.htmlhttp://p53.fr/index.htmlhttp://p53.fr/index.htmlhttp://p53.fr/index.htmlhttp://perspectivesinmedicine.cshlp.org/

TOWARD AN UNBIASED SOMATIC TP53MUTATION SPECTRUM

TP53 occupies �19.14 kb of genomic DNA onchromosome 17p13.1 and is oriented on theminus strand, antisense to other genes in theneighborhood. It comprises 14 exons, including10 exons constituting the coding sequence of thecanonical, full-length p53 protein of 393 aminoacids, one noncoding exon 1, and three alterna-tive exons, exon 2/3, exon 9b, and exon 9g (exon9a being the regularly spliced form of exon 9)involved in the synthesis of alternatively splicedtranscripts encoding p53 protein isoforms (Fig.1A). The noncoding exon 1 is separated from therest of the coding sequence by a highly conservedintron that covers more than half of the gene(10,738 bp). Another gene, WRAP53, is locatedupstream of the TP53 coding sequence on theopposite (plus) strand. It overlaps with exon 1and the proximal part of intron 1 and encodesantisense transcripts that are complementary toexon 1 transcripts (Mahmoudi et al. 2009). Twomain promoter domains, P1 and P2, have beenidentified. P1 is the main promoter and is locat-ed upstream of exon 1. P2 is located within in-tron 1 at about 1000 bp of the 30 end of exon 1and regulates the expression of an intronic poly-adenylated transcript of 1125 b, Hp53int1(D17S2179E, GenBank: U58658.1) (Reismanet al. 1996). The p53 gene has a complex expres-sion pattern with a single major transcript (en-coding the full-length p53 protein) and morethan 12 documented transcripts of variableabundance generated by alternative splicingand/or internal promoter usage (a weak pro-moter has been identified in a region encom-passing exons 2–4 [Marcel et al. 2010]). Thesetranscripts encode protein isoforms that differfrom canonical p53 by the lack of amino-termi-nal domains (D40p53, lacking residues 1–39;D133p53, lacking residues 1–132) or by alter-native carboxyl-terminal domains (p53b andp53g) (Courtois et al. 2004; Khoury and Bour-don 2010; Marcel et al. 2011). Although the pre-cise functions of these isoforms is still poorlyunderstood, there is strong experimental evi-dence that amino-terminally truncated p53 var-iants may counteract the suppressor functions

of full-length p53 protein (reviewed in Marcelet al. 2011; see also Joruiz and Bourdon 2016).

After initial studies reported that somaticmutations were clustered over �1200 bases inconserved regions of exons 5–8, most studieshave sequenced only these regions. Convention-al sequencing methods are optimized for thedetection of small sequence alterations (sin-gle-base mutations, short indels). Larger indelsor rearrangements are not readily scored bythese techniques. These biases have caused anunderrepresentation of somatic mutations oc-curring outside exons 5–8, many of which areindels. Whole-genome sequencing (WGS) andwhole-exome sequencing (WES) mutation datagenerated by NGS may be, in principle, less bi-ased toward an a priori selection of “hotspot”mutation areas. The COSMIC database main-tained at the Sanger Institute (Hinxton, UnitedKingdom) is the largest public database de-signed to store and display somatic mutationinformation relating to cancer. The Whole Ge-nomes Resource (v73) of this database (cancer.sanger.ac.uk/wgs) compiles a total of �5000TP53 mutations detected in WGS/WES proj-ects, corresponding to 1206 different mutationevents (as of May 1, 2015). Of those, only 12mutations are highly recurrent, representingeach at least 1% of the entire mutation dataset (Table 1). These mutations represent muta-tional DNA “hotspots.” Of note, 11 of thesehotspots occur at CpG sites and constitute amolecular signature of the random deamina-tion of unstable 5-methylcytosine. Altogether,the 12 DNA hotspots represent �25%–30% ofthe total number of mutations. In contrast, 773mutations are rare events, each occurring onlyonce or twice in the data set, representing alto-gether �20% of the total number of mutations.Because several DNA hotspot mutations fallwithin the same codons (e.g., 245, 248, or273), only six “major hotspot” codons compriseeach at least 2% of all mutations (175, 213, 245,248, 273, and 282). Another 13 codons repre-sent “mini hotspots,” comprising each between1% and 2% of all mutations (158, 176, 179, 193,195, 196, 220, 249, 266, 278, 306, 337, and 342).Figure 1B shows the distribution of WGS/WESmutations along the p53 coding sequence and

Somatic TP53 Mutations in the Era of Genome Sequencing

Advanced Online Article. Cite this article as Cold Spring Harb Perspect Med doi: 10.1101/cshperspect.a026179 3

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org


http://cancer.sanger.ac.uk/wgshttp://cancer.sanger.ac.uk/wgshttp://cancer.sanger.ac.uk/wgshttp://cancer.sanger.ac.uk/wgshttp://perspectivesinmedicine.cshlp.org/

5

4

3

2

1

0

1

2

3

4

5

6

COSMIC WholeGenomes data set

B

WRAP53

Hp53int1

TP53

a2/3 a9βa9γ

(10 kb)1 2 3 4 5 6 7

19.136 kb

133

ATG

1 40BR

A

8 9 10 11

175

248

273

MissenseNonsenseIndels

Missense andnonsensemutations

176

158

193195196

220

249266 278

306 342

337

CTDODJDDBDPRRTAD2TAD1

213

245 282

179

7

6

7

IARC TP53 database

% o

f all

mut

atio

ns

Figure 1. TP53 mutation spectrum in human cancer. (A) TP53 locus on chromosome 17p13.1, showing mainintron/exon structure (coding exons, red; noncoding exon 1 and 30 UTR in exon 11, gray) and alternative exons2/3, 9a, and 9b (orange). The position of Hp53int1 in intron 1 is shown, as well as the region of exon1/intron 1overlapping with alternative exons 1 of WRAP53. All introns and exons are represented to scale, except intron1. Blue bar, location of rearrangement breakpoints in osteosarcoma. Red arrows, position of the main (þ1) andalternative (þ40, þ133) protein initiation sites. Green arrowheads, orientation of the coding sequences (TP53is located on the minus strand of DNA, WRAP53 on the plus strand). (Legend continues on following page.)



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org


http://perspectivesinmedicine.cshlp.org/

compares it with the distribution compiledfrom the IARC mutation data set in 2010, beforethe systematic usage of NGS (of note, the latestrelease of the IARC database [p53.iarc.fr] hasincluded 866 mutations identified by WGS/WES [3% of the data set]). Although these dis-tributions look very similar, the WGS/WESpattern shows that mutations outside exons5–8 represent 22.5% of the total, almost doublethat estimated by Sanger sequencing (10%–15%). The higher number of silent mutationsdetected by NGS explains only 1.5% of this dif-ference. Within the DNA-binding domain ofthe p53 protein, NGS data show that mutationsat codon 213 (80% nonsense, p.R213�) and co-don 220 (90% missense, p.Y220C) are hotspotsalmost as high as the well-defined hotspot co-dons 245 and 282. NGS data also show a fre-quently mutated area in the oligomerizationdomain (residues 326–355, 5% of all muta-

tions). Within this domain, codon 342 (91%nonsense, p.R342�) constitutes a minor hotspotcodon. Thus, the Sanger data set appears to bebiased toward underrepresentation of muta-tions outside exons 5–8 and overrepresentationof the mutations known as major hotspots(codons 175, 245, 248, 273, and 282), perhapsbecause of a tendency of investigators to notfully analyze exon 6 (which is GC-rich andhard to sequence) and to preferentially reportmutations detected at codons in exons 5, 7, and8. Interestingly, the newly scored hotspot co-dons 213 and 342 contain CpG dinucleotidesequences, as do the standard hotspot codons175, 245, 248, 273, and 282.

About 2% of all mutations occur at intron/exon boundaries and affect splicing donor oracceptor sites. Genome-wide splicing mutationanalysis has identified TP53 as the gene mostcommonly affected by splicing mutation in BC

Table 1. TP53 mutation “hotspots”: 12 mutations representing each of at least 1% of all mutations in the COSMICWhole Genome Dataset (cancer.sanger.ac.uk/wgs)

Position

(codon)

cDNA/base

change

Protein/amino

acid change Type of mutation %

CpG

site Mutation effect

175 c.524G.A p.R175H Substitution/missense 4.0 Yes Structural196 c.586C.T p.R196� Substitution/nonsense 1.4 Yes Structural213 c.637C.T p.R213� Substitution/nonsense 2.1 Yes Null220 c.659A.G p.Y220C Substitution/missense 1.5 No Structural245 c.733G.A p.G245S Substitution/missense 1.3 Yes Structural248 c.742C.T p.R248W Substitution/missense 2.1 Yes DNA-binding248 c.743G.A p.R248Q Substitution/missense 2.6 Yes DNA-binding273 c.818G.A p.R273H Substitution/missense 2.6 Yes DNA-binding273 c.817C.T p.R273C Substitution/missense 2.9 Yes DNA-binding282 c.844C.T p.R282W Substitution/missense 2.2 Yes DNA-binding306 c.916C.T p.R306� Substitution/nonsense 1.1 Yes Carboxy-terminal truncation342 c.1024C.T p.R342� Substitution/nonsense 1.3 Yes Carboxy-terminal truncation

Structural, mutation affecting the folding of the DNA-binding domain; null, mutation predicted to impair protein

synthesis; DNA-binding, mutation replacing an amino acid making direct and specific contact with DNA; carboxy-

terminal truncation, mutation predicted as generating a protein lacking the entire (p.R306�) or part of (p.R342�) carboxy-terminal oligomerization domain.

Figure 1. (continued) (B, top) Codon distribution of mutations (missense, nonsense, and indels) in the codingsequence of TP53, based on mutation data derived from integrated genomic studies compiled in the WholeGenomes Resource (v73) of the COSMIC mutation database (cancer.sanger.ac.uk/cosmic/signatures), showingthe position of major hotspots (.2% of all mutations) and mini hotspots (1%–2% of all mutations). (Bottom)Codon distribution of missense and nonsense mutationsbased on studies using conventional targeted sequencingapproaches and compiled from the International Agency for Research on Cancer (IARC) TP53 mutation database(version R.13, 2008 [Olivier et al. 2010]). The IARC distribution is displayed as a mirror image of the COSMICdistribution. TAD1, TAD2, Transcriptional activation domains 1 and 2; PRR, proline-rich region; DBD, DNA-binding domain; JD, junctional domain; OD, oligomerization domain; CTD, carboxy-terminal domain.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org


http://cancer.sanger.ac.uk/wgshttp://cancer.sanger.ac.uk/wgshttp://cancer.sanger.ac.uk/wgshttp://cancer.sanger.ac.uk/wgshttp://cancer.sanger.ac.uk/cosmic/signatureshttp://cancer.sanger.ac.uk/cosmic/signatureshttp://cancer.sanger.ac.uk/cosmic/signatureshttp://cancer.sanger.ac.uk/cosmic/signatureshttp://perspectivesinmedicine.cshlp.org/

(Dorman et al. 2014). However, significant so-matic variations occur in TP53 introns at sitesother than those implicated in splice junctions,which are not covered in many conventionalor exome-sequencing programs and thus arenot reported in mutation databases. In BC, mu-tations in intron 9 colocalizing with alternativeexons 9a and 9b have been detected using yeastfunctional assays (Iggo et al. 2013). These mu-tations alter the balance between fully splicedand alternatively spliced p53 transcripts, leadingto the preferential synthesis of a protein thatlacks part of the oligomerization domain. Inosteosarcoma, somatic rearrangements in in-tron 1 have been detected by WGS in �20%of the cases (Ribi et al. 2015). The rearrange-ment breakpoints are located across the entiresequence of intron 1, but most of them cluster ina region of 1.7 kb located immediately upstreamof the locus encoding the intronic Hp53int1transcript, suggesting that breakpoint may befacilitated by open chromatin conformation atthis locus (Fig. 1A) (Chen et al. 2014; Ribi et al.2015). So far, intron 1 rearrangements have notbeen observed in other sporadic cancers and thereason for the narrow tissue specificity is un-known. The impact of these rearrangementson p53 function remains to be analyzed, but itis noteworthy that an intron 1 rearrangement(455-kb inversion) cosegregates with inheritedrisk of several cancers in a Li–Fraumeni family.The tumors of these patients show impairedTP53 transcription and loss of heterozygosity(LOH) affecting the wild-type allele (Ribi et al.2015). Osteosarcoma has been considered as atype of cancer with low rate of TP53 mutationsin exons 2–11 (20%), partly balanced by fre-quent functional inactivation of p53 proteinthrough amplification of MDM2 (Ladanyiet al. 1993). The finding of intron 1 rearrange-ments in about half of sporadic osteosarcomasleads to a call for reconsidering them as a cancerwith a high rate of somatic TP53 mutations.

BEYOND MUTATIONS: IMPACT OF TP53HAPLOTYPES

TP53 contains more than 100 validated natural-ly occurring single-nucleotide polymorphisms

(SNPs; see p53.iarc.fr/TP53GeneVariations.aspx) but only a few of them have been stud-ied for their effect on p53 functions. Thebest characterized exonic SNP is a nonsilentsubstitution at codon 72 in exon 4 (rs1042522,g.7520197G.C, p.R72P), which is present atminor allele (P) frequency between 10% (Cau-casians, Northern Europe) and 50% (Africans,Yoruba), depending on population (Whibleyet al. 2009). Several studies have shown thatthis SNP specifies protein variants with differ-ent functional properties in vitro and in ex-perimental in vivo models. The form of p53with R (arginine) at codon 72 (p53R72) moreeffectively induces p53-mediated apoptosisthan p53P72 (proline), partially through tar-geting p53 to the mitochondria (Dumont et al.2003). Meanwhile, p53P72, when comparedwith p53R72, more efficiently induces cell-cyclearrest and DNA repair (Pim and Banks 2004;Siddique and Sabapathy 2006). The ability ofmutant p53 to bind p73, neutralize p73-in-duced apoptosis, and experimentally transformcells in cooperation with EJ-Ras is enhancedwhen the mutant protein carries an arginine atposition 72 (Marin et al. 2000). In addition, R72alleles appear to be preferentially mutated andretained in squamous cell carcinomas arising inpatients who are R72P germline heterozygotes.Thus, this intragenic polymorphism may actas a modifier of mutant TP53 effect, suggest-ing that mutations and /or LOH may occurat different rates on different TP53 haplotypes.This hypothesis is supported by results fromMechanic et al. (2007) who analyzed 14 intra-genic TP53 SNPs in a case-control study oflung cancer and uncovered an association be-tween several combinations of SNPs and therisk of somatic mutation. The SNPs associatedwith mutation were located between intron2 and intron 7, including R72P/rs1042522(exon 4), rs9895829 (intron 4, c.376-125T.C),rs1625895 (intron 6, c.672þ62A.G), andrs12951053 (intron 7, c.782þ92T.G), defin-ing the haplotype containing R-T-A-G as morelikely to carry a mutation. The reason why thishaplotype is preferentially mutated is unknown.In HCC carrying the aflatoxin-induced p.R249Smutation, analysis of 19 SNPs spanning the



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org


http://p53.iarc.fr/TP53GeneVariations.aspxhttp://p53.iarc.fr/TP53GeneVariations.aspxhttp://p53.iarc.fr/TP53GeneVariations.aspxhttp://p53.iarc.fr/TP53GeneVariations.aspxhttp://perspectivesinmedicine.cshlp.org/

entire TP53 locus has shown a strong associa-tion between mutation and haplotypes thatcarry a combination of two SNPs (rs17882227and rs8064946) defining a linkage disequili-brium block extending from upstream of non-coding exon 1 to the first half of intron 1. Thisdomain contains two coding sequences over-lapping with TP53 (WRAP53 and Hp53int1)(Ortiz-Cuaran et al. 2013). It is possible thatthese SNPs may modulate the p53 antisenseactivity of these transcripts, thus having an im-pact on p53 expression levels and therefore onthe rate at which specific mutant alleles occurand are retained during tumor development. Asimple way to rationalize this notion is to con-sider that different haplotypes express p53 atdifferent levels, defining a continuum of func-tional p53 responses from “weak” to “strong”TP53 alleles, which may differ for different bi-ological p53 effects (e.g., cell-cycle arrest, apo-ptosis, control of metabolism, or DNA repair)(Fig. 2). Mutations on a “weak” haplotype maybe more likely to require the loss of the “strong”wild-type allele to exert a significant effect. Incontrast, mutation on a “strong” haplotype mayhave disrupting, dominant effects even if aweaker wild-type allele is present.

MUTATION PATTERNS AS SIGNATURESOF MUTAGENIC PROCESSES

TP53 mutations have been extensively used asclues for the etiology of endogenous and exog-enous mutagenic processes that operate duringcarcinogenesis in humans. Multiple indepen-dent studies have been conducted by Sangersequencing of TP53 in tumors selected fortheir suspected association with a mutagenor mutagenic processes. The spectra producedby the aggregation of these somatic mutationswere compared with the mutations experimen-tally generated by the suspected mutagensin vitro or in vivo systems (Hollstein et al.1999; Pfeifer 2015). Table 2 summarizes the“molecular signatures” identified in TP53,caused by specific agents or mutagenic process-es identified in TP53. This information has beenextensively discussed in previous reviews(Greenblatt et al. 1994; Hainaut and Hollstein

2000; Olivier et al. 2004, 2010; Robles and Har-ris 2010; Pfeifer 2015). Nonmelanoma skin can-cer (basal and squamous cell carcinoma) andmelanoma show a predominance of G:C.A:Ttransitions at dipyrimidine sites (mutated baseunderlined), including about 10% of CC.TTtandem mutations, a type of mutation resultingfrom UV-light-induced covalent coupling ofC¼C double bonds at adjacent pyrimidines.These mutations are not caused by other knowncarcinogens and are almost never observed intumors of internal organs (Luo et al. 2001). InHCC, a unique transversion at codon 249(p.R249S; G:C.T:A) is highly prevalent in geo-graphic areas in which the mycotoxin aflatoxinis a widespread contaminant of the food (partsof Africa, Eastern Asia, South America) (re-viewed in Gouas et al. 2009). In lung cancer,�30% of TP53 mutations are G:C to T:A trans-versions in smokers, but not in never-smokers.This class of mutation is the main mutageniceffect of polycyclic aromatic hydrocarbons(PAHs) from the tar fraction of cigarette smoke.Exposure of human bronchial cells to the diolepoxide of the PAH benzo[a]pyrene causesDNA damage at the same G:C pairs that arefrequently mutated into T:A in lung cancersfrom smokers (Denissenko et al. 1996; Tretya-kova et al. 2002). Recently, a mutational finger-print of aristolochic acid has been reportedin urothelial carcinomas of subjects with a his-tory of exposure to herbal remedies and foodproducts containing seeds of Aristolochia clema-titis (Hollstein et al. 2013; Poon et al. 2013).Last, all cancer types harbor at least 20% ofG:C.A:T mutations at hypermutable CpG di-nucleotides, attributed to the normal cellularevent of deamination of 5-methylcytosine tothymine. This process is enhanced by inflam-mation, resulting in a high prevalence of thesemutations in cancers associated with chronicinflammation (Ambs et al. 1999; Cooks et al.2014). CpG mutations are exceptionally com-mon in a few specific types of cancer includingcolorectal, gastric, and esophageal cancers (fre-quently occurring from precursor inflammato-ry lesions) as well as in brain cancers.

The introduction of NGS has provided astepping-stone for a giant leap in our under-



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



standing of the signatures of mutational pro-cesses in human cancer. In addition to accessto large data sets (between a few hundred and.10,000 mutations per tumor genome), NGSdata have supported the development of acomputational framework that allows decon-structing distinct but overlapping patternsfrom a set of cancer samples (Alexandrov et al.2013a,b). Somatic mutations in the genome ofany cancer are the result of cumulative muta-

genic processes encompassing endogenous andexogenous mechanisms, each operating withdifferent temporal patterns and with differentstrengths. Often, the cumulative pattern incancer does not match any of the known sin-gle-operative mutational processes. By usingalgorithms that decipher the minimal set ofmutagenic signatures that optimally explainthe proportion of each mutation type in eachcancer, it is now possible to estimate the contri-

MT WT

Mutation

MT WT

LOH

A

B

Figure 2. Variability and potential impact of TP53 haplotypes. (A) Linkage disequilibrium (LD) plot in theHapMap Panel for Caucasians, using 29 tag SNPs (tSNPs) for typing TP53 haplotypes (Garritano et al. 2010). Thedegree of LD is indicated by shades from black (strong LD) to white (no LD). This figure shows that most tSNPsin intron 1 and the TP53 promoter are in strong LD, forming a conserved haplotype block. (B) Model for loss ofheterozygosity (LOH) at the wild-type (WT) allele depending on the interplay between “strong” (red) and “weak”(green) mutant haplotypes. (Top row) If the mutation (MT) occurs on a “strong” haplotype, its effects maydominate the residual WT “weak” haplotype, making LOH not compulsory. In all other haplotype combinations,LOH is required to eliminate a WT haplotype stronger than or equally strong as the one carrying the mutation.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



Tabl

e2.

TP53

muta

tion

pat

tern

sas

asi

gnat

ure

ofm

uta

tional

pro

cess

es

Bas

ech

ange

(under

lined

)Ty

pe

Mec

han

ism

/

agen

t

Pro

muta

genic

lesi

on

Typic

alTP53

muta

tions

Can

cers

Ref

eren

ces

CC

.T

Tta

nd

emTr

ansi

tio

nSo

lar

UV

Dip

yrim

idin

ed

imer

sp

.S12

7F,

p.Q

136�

,p

.R15

8C,

p.H

179Y

,p

.R19

6�,

p.R

248W

,p

.R27

8F,

p.R

282W

No

nm

elan

om

ask

inB

rash

etal

.19

91

G:C

.T

:ATr

ansv

ersi

on

Afl

ato

xin

(AF

B1)

AF

B1-

N7-

guan

ine

add

uct

sp

.R24

9SL

iver

(hep

ato

cell

ula

rca

rcin

om

a)B

ress

acet

al.

1991

G:C

.T

:ATr

ansv

ersi

on

Tob

acco

smo

kePA

H-N

2-gu

anin

ead

du

cts

p.V

157F

,p

.R15

8L,

p.G

245V

,p

.G24

5C,

p.R

248L

,p

.R24

9M,

p.R

273L

,p

.D28

1E,

p.E

298�

Bro

nch

us

and

lun

g,h

ead

and

nec

k,es

op

hag

us

Den

isse

nko

etal

.19

96;

Tret

yako

vaet

al.

2002

A:T

.T

:ATr

ansv

ersi

on

Ari

sto

loch

icac

idA

rist

ola

ctam

-DN

Aad

du

cts

p.Q

104L

,p

.N13

1Y,

p.K

139�

,p

.K16

4�,

p.R

209�

,p

.R28

0S,

p.K

291�

Bla

dd

er/u

roth

eliu

mH

oll

stei

net

al.

2013

G:C

.A

:Tat

Cp

GTr

ansi

tio

nR

and

om

dea

min

atio

n5-

met

hyl

cyto

sin

ep

.R17

5S,p

.R21

3�,p

.R24

8Q,p

.R24

8W,

p.R

282W

,p

.R30

6�,

p.R

342�

All

can

cers

,in

par

ticu

lar,

colo

n,

sto

mac

h,

bra

inA

mb

set

al.

1999



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



bution of different signatures to each cancermutation spectrum.

The first NGS studies focusing on muta-tion patterns were published in Nature in 2010and reported the patterns of somatic mutationsin malignant melanoma (Pleasance et al. 2010a)and small-cell lung carcinoma (Pleasance et al.2010b). These studies confirmed the wide-spread impact of UV light and tobacco carcin-ogens, respectively, in these cancers. Since then,many international consortia have undertakenthe sequencing of large numbers of cancer sam-ples, and mutational signature genomic pat-terns have been identified for many of them(Alexandrov et al. 2013a; Roberts and Gordenin2014). The COSMIC database website providesa resource documenting up to 30 mutationalsignatures based on pooled NGS data from10,952 exomes and 1048 whole genomes across40 different types of cancers (accessible atcancer.sanger.ac.uk/cosmic/signatures). Sever-al of these signatures confirm and refine mu-tational processes already described in TP53(COSMIC signature 1, spontaneous deamina-tion of 5meC; signature 4, tobacco smoke; sig-nature 7, UV light; signature 22, aristolochicacid; and signature 24, aflatoxin).

Surprisingly, there are only a few new mu-tational signatures informing on exogenous car-cinogen exposure in addition to those alreadydetected in TP53. One example is signature 11,which corresponds to a distinct C.T mutationpattern caused by the alkylating agent temozo-lomide in recurrent glioblastoma multiformeof patients that have undergone this type ofchemotherapy. On the other hand, NGS mu-tational signatures provide new insights intoendogenous mutation processes, including mu-tational signatures caused by the error-pronepolymerase h (signature 9, detected in chroniclymphocytic leukemia and malignant B-celllymphomas), by proofreading defects attribut-ed to polymerase 1 (signature 10, in subsets ofcolorectal and uterine cancers), by defectivemismatch repair (signatures 6, 15, 30, and 26,in gastric, colorectal, and endometrial cancer),and by defective repair of DNA strand breaksassociated with germline and somatic BRCA1and BRCA2 mutations (signature 3, breast, pan-

creatic, and ovarian cancers). Several new mu-tation signatures are still to be assigned to aspecific mechanism (signatures 5, 8, 12, 14,17–19, 21, 23, 25, 28, and 30). For example,esophageal adenocarcinomas carry a very un-usual type of mutation, A-to-C mutations at50 AA dinucleotides (Pfeifer 2015). There areno known mutational processes, either endog-enous or exogenous, that have been shown ex-perimentally to result in these types of muta-tions. Because these events are highly specificfor a tissue traversed by ingested materials, itis plausible that these mutations are caused byexposure to an exogenous mutagen. Further-more, in addition to the aforementioned expo-sures to aflatoxins, liver cancer genomes, unlikeother cancer genomes, have very diverse pat-terns of mutations targeting A/T base pairs.Because the liver is the primary organ for me-tabolism of xenobiotics, it would not be surpris-ing if these A/T-targeted mutations were causedby metabolites of environmental carcinogens.

Some of the most interesting new data pro-vided by NGS is the identification of an endog-enous pathway involving cytidine deaminasesof the AID/APOBEC family (signatures 2 and13). Abnormal targeting of cytidine deaminasesto random chromosomal locations appearsto cause frequent C.T mutations at the cyto-sine of 50TpC dinucleotides. The human ge-nome encodes eight APOBEC enzymes (sevenof them located at the APOBEC3 gene cluster)that normally serve to restrict viral infection andretrotransposon mobility by deaminating cyto-sines during the ssDNA stage of their replicationstages. These enzymes target TCA/TCT trinu-cleotides in ssDNA in which they deaminateC to U, which is either directly copied to Tor excised by uracil DNA glycosylase, followedby copying of the resulting abasic site into C.Tor C.G mutations. Because of the processivityof APOBEC enzymes, these mutations oftenoccur in clusters. AID/APOBEC signatures arepresent in many forms of cancer, and a germlinedeletion polymorphism affecting APOBEC3Aand APOBEC3B is associated with large num-bers of them (Nik-Zainal et al. 2014).

It is now possible to revisit TP53 mutationpatterns in the light of these extended and re-



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org


http://cancer.sanger.ac.uk/cosmic/signatureshttp://cancer.sanger.ac.uk/cosmic/signatureshttp://cancer.sanger.ac.uk/cosmic/signatureshttp://cancer.sanger.ac.uk/cosmic/signatureshttp://perspectivesinmedicine.cshlp.org/

fined mutational signatures. Recent studies haveidentified APOBEC mutation patterns in TP53in BC (Lindley 2013) and lung adenocarcinoma(Waters et al. 2015). Although these studiesare useful to better understand the sequenceand codon context of mutation formation inTP53, the amount of etiological informationthey reveal is limited compared with multigenestudies using NGS.

FUNCTIONAL DIVERSITY OF TP53MUTATIONS

The spectrum of TP53 mutation in cancer cellsand tissues is the result of combined processesacting as filters that select cancer mutationsamong a broad range of possible mutations.Sequence-context-dependent mutagenesis andDNA repair represent two of these filters. Thebiological selection of functionally “meaning-ful” mutations represents a third filter. Howev-er, what is a “meaningful” human cancer mu-tation is still largely unclear.

The lowest common denominator of cancermutations is loss of transcriptional function(LOF) altering the p53-dependent transcrip-tional repertoire. Not all mutants are equal inthis respect. The use of standardized yeast-basedassays to test for LOF at total of 2314 mutantsrepresenting all possible substitutions through-out the protein has shown that LOF is virtuallycomplete for frequently occurring mutants (in-cluding hotspots) but that some of the rarelyoccurring mutants retain quasi-wild-type func-tionality for transactivation (Kato et al. 2003).Most of these rare functional mutants and occuronly once or twice in the COSMIC WGS/WESdata set (representing ,3% of all mutations).They may simply represent passenger mutationswith no functional significance, detected onlybecause of their accidental presence in an ex-panded clone of transformed cells. An interest-ing parallel exists between these mutants andthose occasionally detected in noncancer tissuessuch as rheumatoid arthritis (RA), a noncancerprecursor chronic inflammatory disease charac-terized by oligoclonal proliferation of nontrans-formed synoviocytes. RA tissues frequently con-tain clusters of cell clones with atypical TP53

mutations that include a high proportion ofsilent mutations (14%), mutations retainingwild-type transactivation capacity (60%), andmutations rarely found in any cancer (70% ofthem are either absent or reported only once inthe WGS/WES COSMIC data set) (Firesteinet al. 1997; Yamanishi et al. 2002). Such mutantsare unlikely to be selected during carcinogenesisbut are carried over as passengers in expandingsubclones of RA synoviocytes.

There is compelling experimental evidencethat TP53 mutations induce functional changeswell beyond the scope of LOF. Many missensep53 mutants are expressed as stable proteinsthat exert dominant negative (DN) effects byinterfering with the product of the remainingwild-type allele. A “prion-like” effect of somemutant p53 over wild type has been shown invitro, in which the mutant enforces wild-typeprotein to adopt a denatured, mutant-like con-formation (Milner and Medcalf 1991). DNeffects may also result from a “sponge” effect ofstable mutant p53 protein, absorbing the low-abundance wild-type proteins into oligomersdominated by mutant forms. Aside from DNeffects, gain-of-function (GOF) effects havebeen demonstrated through which mutant p53exerts a number of prooncogenic biochemi-cal activities. GOF effects have been extensivelystudied in experimental cell systems and in ge-netically modified mouse models (Brosh andRotter 2009; Muller and Vousden 2014). A num-ber of mechanisms have been proposed for mu-tant p53 GOF, including direct binding to DNAchanging (rather than abolishing) gene expres-sion, binding to a large panel of transcriptionfactors enhancing or impairing their function(e.g., binding to the p53 family members p63and p73), binding to a number of proteins notdirectly involved in transcription (e.g., bindingto the MRE11-RAD50-NSB1 complex, disrupt-ing its function), and regulating microRNA net-works (e.g., miR130b, miR155, and miR205)(Brosh and Rotter 2009; Muller and Vousden2014). Recently, several p53 mutant proteinshave been shown to up-regulate a number ofchromatin regulatory genes, including themethyltransferases MLL1/KMTA2A and MLL2(KMT2D) and the acetyltransferase MOZ/KAT6A,



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



resulting in a global increase in histone methyl-ation and acetylation and suggesting that mutantp53 may induce GOF effects through sweepingchanges in the epigenetic control of chromatindynamics (Zhu et al. 2015). The functional con-sequences of these GOF effects encompass theentire spectrum of the “hallmarks of cancer”(Hanahan and Weinberg 2011; Solomon et al.2011). Thus, in contrast to LOF, there is no sim-ple standard assay for scoring GOF effects. Fur-thermore, there is no clear understanding of thestructural properties of mutant p53 that supportGOF effects. Given the multiplicity of GOFmechanisms, it is unlikely that all effects dependon a single structural feature. In addition, mu-tants with the most perturbed protein structureare not necessarily those with the strongest GOFeffects. In contrast, these effects are expected tobe subtle and context-dependent, because theyrequire interactions between mutant p53 andproteinsthatare themselves expressed in atissue-and context-dependent manner. The subtlety ofGOF effects is illustrated by p.R249S, the mutantcaused by aflatoxin-induced mutagenesis inHCC arising in a context of chronic hepatitis B(HB) infection. This mutant is a clear LOF mu-tant, but is not commonly considered as GOF inmost studies. However, its mechanism of actioninvolves the binding of the HB antigen HBx, ac-tivating it as a viral oncogene in liver cells (Jianget al. 2010; Gouas et al. 2012). This effect is atypical context-dependent GOF effect. Thebinding specificity of p.R249S to HBx may ex-plain the surprisingly narrow spectrum of afla-toxin-induced TP53 mutations in HCC, becauseother mutants potentially induced by aflatoxinmay not bind HBx with the same efficacy asp.R249S, preventing their selection during HB-dependent hepatocarcinogenesis (Denissenkoet al. 1998).

CLASSIFYING TP53 MUTATIONS FORCLINICAL USE

Although a multitude of experimental studieshave shown different effects of various mutantson proliferation, apoptosis, or responses to cy-totoxic drugs, there is surprisingly little clinicalevidence for association between different types

of mutations and clinicopathological variables,prognostics, or prediction of therapeutic re-sponses. This does not imply that such associ-ations do not exist. Perhaps the best evidence forclinically relevant differences between differentmutations is found in subjects who carry germ-line TP53 mutations. In a recent report on 322TP53 mutation carriers from French Li–Frau-meni families, the mean age of tumor onset wasstatistically different between carriers of mis-sense mutations (23.8 years) and those carryingnonsense mutations or genomic rearrange-ments (28.5 years). The difference was evenlarger when comparing hotspot mutations,such as p.R175H or p.R248W (mean age of tu-mor onset, 15–20 yr) with complete deletion ofthe TP53 gene (38.5 yr) (Bougeard et al. 2015).Assuming that complete deletion of TP53 rep-resents a “straight” LOF effect, these observa-tions suggest that the more severe clinical phe-notype of hotspot mutants is caused by theircapacity to impair the wild-type allele (DN ef-fects) and/or by specific GOF effects. It is likelythat similar genotype–phenotype correlationsoccur for somatic mutations.

Many clinical studies have tried to subdividesomatic mutations into severity classes basedon localization and predicted structural effects.Small clinical cohorts, suboptimal methodsused to assess TP53 mutations status, and lackof consistency in how the mutations are classi-fied have hindered a robust correlation withclinical features. Computational approaches us-ing sequence similarities such as SIFT (sortingintolerant from tolerant) generate relativelyaccurate prediction for loss of transactivationfunction in vitro (specificity, 90%; sensitivity,70% [Mathe et al. 2006]) but their clinical use-fulness is limited to the identification of rarepassenger mutations. Two simple classificationalgorithms have been used with some success(Fig. 3). The Poeta algorithm classifies muta-tions into two a priori classes, disruptive andnondisruptive mutations. Disruptive muta-tions combine predicted-null mutations (non-sense, frameshift, and splice site) and chemical-ly deleterious substitutions in loops formingthe DNA-binding surface of the protein (Poetaet al. 2007). Nondisruptive mutations include



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



Disruptive

No stableprotein

Stable mutant protein

0

0.0

0.2

0.4

0.6

0.8

Disruptive

0.0364NRNR27 41

76

Median timeto LRR (mo)

5-year freedomfrom LRR (%) P

Nondisruptive

Wild type

1.0

20 40

Follow-up (mo)

Pro

port

ion

of p

atie

nts

with

out L

RR

60 80 100

Non-missenseMissense DBM

Missense non-DBM

Mutations, codons 96–296

Olivier/hainaut algorithm (Olivier et al. 2006)

Poeta algorithm (Poeta et al. 2007)Locoregional recurrence, head and neck

squamous cell carcinoma

Nondisruptive

Changingcharge/polarity

Othermotifs

L2/L3loops

Introducing aSTOP

Nonsense Missense

Mutations, codons 96–296

Wild typeNondisruptiveDisruptive

A

B1.0

0.9

0.8

0.7

0.6

0.5 p < 0.0001

Pro

port

ion

surv

ivin

g

Non-missense

DBM

Non-DBM

Wild type

Overall survival, breast cancer

0

No. at risk:1492

57178

65

135749

145

47

120840

117

38

103030

101

27

7692771

20

5692349

16

4101539

12

20 40 60

Time in months

80 100 120

Nonsense,frameshiftsplicing

Figure 3. TP53 mutation classification algorithms. (A) The Poeta algorithm (left) (Poeta et al. 2007) classifiesmutations occurring in the DNA-binding domain in two groups, disruptive (D) and nondisruptive (ND)mutations. D mutations either preclude the synthesis of p53 (nonsense mutations) or cause a structurallyimportant amino acid change in the L2/L3 loops of the DNA-binding domain. Other missense mutationsare classified as ND. (Right) D and ND mutations have different prognostic value for locoregional recurrence(LRR) in a cohort of 74 patients with squamous head and neck cancer (Skinner et al. 2012). (B) The Olivier/Hainaut algorithm (left) classifies mutations in three groups: mutations predicting a null p53 protein (non-missense), missense mutations in DNA-binding motifs (missense DBM), and missense mutations outside theDNA-binding motif (missense non-DBM). (Right) These three categories show different prognostic values foroverall survival in a cohort of 1794 patients with breast cancer (Olivier et al. 2006).



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



any other missense mutation. In head and necksquamous cell cancer (HNSqCC), disruptivemutations are associated with decreased overallsurvival, whereas there is no significant associ-ation with nondisruptive mutations. Moreover,disruptive mutations predict locoregional re-currence and HNSqCC cell lines with disruptivemutations are significantly more radioresistantthan those expressing nondisruptive mutations(Skinner et al. 2012). The Olivier/Hainautalgorithm specifies three classes of mutations,predicted-null, missense mutation in the DNA-binding motif (DBM), and missense mutationin non-DNA-binding motif (non-DBM) (Oliv-ier et al. 2006). The distinction between DBMand non-DBM mutations is purely topologi-cal, based on the observation that interspeciesconservation of amino acids in the loops ofthe DBM is higher than in b-strands and shortintervening loops of the non-DBM. The non-DBM is thus predicted to be more tolerant tomutations than the DBM (Mathe et al. 2006). Ina large series of 1791 BCs from European pa-tients, the algorithm identified predicted-nullmutations as having the strongest associationwith poor survival, followed by DBM, whereasnon-DBM were just marginally worse thanwild-type TP53 (Olivier et al. 2006). In theMETABRIC BC cohort (1420 patients), no sta-tistical difference was found between the threemutation classes, although DBM mutations ap-pear to have a marginally more severe effect(Silwal-Pandit et al. 2014). Incidentally, thefact that the two different classes of missensemutations appear to have the same (or less)effects as predicted-null mutations arguesagainst DN or GOF effects of missense muta-tions in BC. When applied to non-small-celllung cancer (NSCLC), the algorithm does notidentify any prognostic value associated witheither mutation class. However, in a random-ized trial of platinium-based therapy after sur-gery (525 patients, 221 with mutation; 42%),non-DBM mutations were associated withpoor overall survival in the group that receivedchemotherapy but not in the group treated bysurgery only (Ma et al. 2014). This effect was notseen with predicted null or DBM mutations.This observation suggests that non-DBM may

exert DN and/or GOF effects that confer resis-tance to adjuvant therapy. It is interesting tonote that non-DBM mutations are frequent inNSCLC of smokers, because of the frequent tar-geting by tobacco carcinogens of codons encod-ing residues located in the hydrophobic core ofthe protein (Le Calvez et al. 2005). The selectionof these mutations may contribute to a form ofdrug resistance in tumor cells.

Overall, these results indicate that the clinic-al severity of mutations is tumor- and context-dependent. Thus, the same mutations do notcarry the same clinical consequences in differentcancers. Although it can be assumed thatthe LOF effects of mutations might be roughlyequivalent in different tissue contexts, DN ef-fects are dependent on the retention, expression,and activity of the wild-type allele, and GOF ef-fects may depend on factors that are expressed ina tissue- and context-dependent manner.

TP53 IN THE GENOMIC LANDSCAPEOF CANCER: SEEING THE NEEDLE WITHINTHE HAYSTACK

Until recently, TP53 mutations have been most-ly approached as stand-alone events in studiesof limited size and statistical power. The de-velopment of integrative genomics combiningWGS/WES with DNA copy number alterationsand studies on mRNA and protein expressionprovides an unprecedented view of the generalgenomic contexts in which TP53 mutations oc-cur (Kristensen et al. 2014). In recent months,large consortium studies have combined multi-dimensional molecular approaches to charac-terize the landscape of genomic alterations inhuman cancers. These studies are generating awealth of information that remains to be fullyexplored using TP53 mutation as a magnifyinglens. Here, we briefly highlight some of the newknowledge that integrative genomics is bring-ing to our understanding of TP53 mutationsin breast and non-small-cell lung cancer.

Breast Cancer (BC)

With mutation in 20%–25% of all cases, TP53ranks first in a series of seven genes that are



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



mutated in .10% of BCs (including PIK3CA,ERBB2, MYC, FGFR1/ZNF703, GATA3, andCCND1) (Stephens et al. 2012). A global studyof the genomic and transcriptomic architectureof breast cancers has identified up to 10 molec-ular subtypes (integrative clusters) in whichTP53 mutations occur at significantly differentfrequencies (Curtis et al. 2012). A detailed anal-ysis of the TP53 mutation spectrum in the 1420BCs of the METABRIC cohort has identifiedmutations in 28.3% of the cases (402 patients),with significant variations across BC subtypes(Silwal-Pandit et al. 2014). Mutations are fre-quent in basal-like and HER2-enriched tumors(65% and 53.4%, respectively), rare in luminalA and normal-like tumors (9.3% and 11%, re-spectively), and occur at an intermediate rate inluminal B tumors (24.8%). LOH at the TP53locus is observed in 80.9% of the mutated tu-mors, independently of molecular subtype.However, in TP53 wild-type cases, the frequen-cy of LOH varies across subtypes (from 24%in basal-like to 52% in luminal B). The distri-bution of mutations show striking differencesbetween subtypes, with hotspot mutationsdominating the spectrum of basal-like tumors,whereas a more uniform mutation distributionis seen in other subtypes.

When examining the associations betweenmutation and prognosis, TP53 mutation ap-pears to be strongly correlated with poor overallsurvival in ER-positive patients but not in ER-negative patients. In fact, the presence of a mu-tation seems to wipe out the prognostic benefi-cial effect of ER positivity (Olivier et al. 2006;Silwal-Pandit et al. 2014). This observation sug-gests that p53 may function as a rate-limitingfactor for ER signaling in BC, a conjecture sup-ported by experimental data implicating p53in estrogen responses in breast cancer cells (Fer-nandez-Cuesta et al. 2011a,b). Patients withmutant TP53 may thus not respond as well toendocrine breast cancer therapy as patients withwild-type TP53. In a recent study, p53 proteinaccumulation has been identified as a strongpredictor of recurrence in ER-positive BC pa-tients treated by aromatase inhibitor. It remainsto be shown whether this accumulation is asso-ciated with mutant p53 (Yamamoto et al. 2014).

The association between TP53 mutation andBC prognosis is not uniform across subtypes.Mutations are significantly associated with poorsurvival in HER2-enriched, luminal B, and nor-mal-like but not in basal-like or luminal A, sug-gesting a diverse role and impact for TP53 muta-tions (Fig. 4) (Silwal-Pandit et al. 2014). In basal-like cancers, disruption of the p53 pathway maybe an obligate mechanism in virtually all tumors,either by mutation or byother mechanisms. Pres-ence of wild-type TP53 may therefore not implyan intact p53 pathway, explaining the lack ofprognostic value of TP53 mutations. In HER2-enriched cancers, mutation may disable the ca-pacity of p53 to operate a safeguard mechanismagainst oncogene-driven cell proliferation in-duced by activated HER2. In luminal B and nor-mal-like cancers, disruption of TP53 function isan optional mechanism and the occurrence of amutation and/or LOH makes a significant differ-ence for tumor growth and progression, thus en-tailing a clear negative prognostic value. In con-trast, in luminal A cancers, mutation is a rareevent that may be relatively unimportant. Thefact that luminal A mutations rarely occur at hot-spots suggest that manyof them might be passen-ger mutations. There is no significant differencein prognostic value between different classes ofmutations in any of the subtypes.

With respect to predictive value of muta-tions, it has long been suggested that docetaxelmay confer a greater therapeutic advantage overanthracyclines in BC with mutated comparedwith wild-type TP53. This hypothesis has notbeen substantiated in two large clinical trials,which have confirmed the prognostic value ofmutant TP53 but have not provided evidencefor a predictive effect of mutation on responseto therapy (Bonnefoi et al. 2011; Fernandez-Cuesta et al. 2012).

Non-Small-Cell Lung Cancer (NSCLC)

NSCLC comprises two major histological enti-ties, squamous cell carcinoma (SqCC) andadenocarcinoma (AdC), and several minor his-tological subtypes including large-cell carcino-mas (LCCs). WES reveals TP53 mutations in81% of SqCC, a frequency at the higher end of



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



those reported in studies using Sanger sequenc-ing (range, 50%–80%), in which TP53 muta-tion load in SqCC appeared to correlate withsmoking history (Le Calvez et al. 2005). TP53is by far the most frequently mutated gene inSqCC, ahead of MLL2 (20%), PI3KCA (16%),CDKN2A (15%), NFE2L2 (15%), and KEAP1(12%) (Clinical Lung Cancer Genome Project2013). There is no obvious difference in the pat-terns of mutated genes between TP53-mutatedand wild-type cases, suggesting that disruptionof the p53 pathway is an obligate mechanism inthe pathogenesis of SqCC. However, the highfrequency precludes using TP53 mutations as aclassifier for identifying molecular subtypes orfor investigating prognostic and predictive val-ue. In AdC, mutations occur in 46% of the cases,ahead of KRAS (33%), KEAP1 (17%), STK11(17%), EGFR (14%), NF1 (11%), and BRAF(10%). It is striking to note that TP53 is theonly gene that is commonly mutated in bothmajor histological types of NSCLC. In bothAdC and SqCC, the pattern of TP53 mutationsis characterized by a high rate of transversionsoccurring at G bases on the nontranscribedstrand, a typical signature of mutations inducedby tobacco smoke. This pattern is seen in all

histological subtypes but less marked in AdCthan SqCC. In AdC, this pattern is not detectedin EGFR-mutated cases, which mostly developin never smokers, and in which TP53 mutationsoccur at a lower frequency than in smoking-re-lated KRAS-mutated cases (30% vs. 60%, re-spectively). Independent of this associationwith smoke-related mutation signatures, TP53mutation does not appear to be associated withany specific molecular feature of AdC.

A multitude of cohort studies has failed toidentify a prognostic effect for TP53 mutationson NSCLC survival, even in AdC and after sep-arating TP53 into different classes (see, e.g.,Scoccianti et al. 2012; Ma et al. 2014). However,in a comprehensive analysis of 1255 clinicallyannotated lung cancers by the Clinical LungCancer Genome Project (CLCGP), TP53 muta-tion was found to be associated with a signif-icant negative prognostic value in AdC withmutated EGFR (Fig. 4) (Clinical Lung CancerGenome Project 2013). In fact, AdC with mu-tated EGFR seems to have a more favorableprognosis than AdC with other driver oncogenemutations, including KRAS. Although TP53mutation does not appear to have an effect onthe prognosis of AdC with mutated KRAS, it

0

0.0

0.2

0.4

BC

SS

pro

babi

lity

0.6

0.8

1.0

7385

6671

5248

3937

3430

2615

1614

96

wtmut

24 48

P = 0.030

TP53 mutant (e = 44)

TP53 wild type (e = 27)

72 96

Survival time (mo)

120 144 168 192

Breast cancerHER-2-enriched (n = 158)

1.0

0.8

0.6

0.4

0.2

0.0

0 10 20

12 15

16 TP53wt

TP53wt

Non-small-cell lung cancerEGFR mutant (n = 80)

TP53mut

TP53mut

StageI/II III/IV

37

P = 0.028 P = 0.01

30 40 50 60

Figure 4. Prognostic value of TP53 mutation in breast and lung cancer driven by oncogenes of the EGF receptorfamily. (Left) Prognostic value of TP53 mutations in patients with HER2-enriched breast cancer subtype (Silwal-Pandit et al. 2014). (Right) Prognostic value of TP53 mutations in patients with adenocarcinoma of the lung withmutant EGFR (Clinical Lung Cancer Genome Project 2013).



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



cancels the apparent beneficial effect of EGFRmutation. It should be noted that this beneficialeffect is not only attributable to the responseof these patients to targeted EGFR therapies. Itis also observed in a retrospective series of AdCpatients who have been treated with conven-tional therapies. These observations suggestthat wild-type p53 functions as a limiting factorin AdC driven by oncogenic activation of EGFR.There is a striking parallelism between the effectof TP53 mutation in EGFR-mutated lung AdCand in HER2-enriched breast cancer (Fig. 4).

TP53 mutations have been reported as hav-ing a predictive effect on the survival of patientstreated with adjuvant platinum-based therapy,a standard treatment regimen for stage II–IIIcompletely resected NSCLC (Ma et al. 2014).This effect predicts a marginal benefit of adju-vant therapy in patients with wild-type TP53and marginal detrimental effect in patientswith mutated TP53 mutation. There is no effectof TP53 mutation in patients who are treatedby surgery alone. As discussed above, this effectis significant only for non-DBM mutations, aclass of mutations occurring in regions encodingthe hydrophobic core structure of the p53 DNA-binding domain. Mutations in the hydrophobiccore domain are frequent in lung cancers ofsmokers, because of the targeting of codons inthis region by carcinogens from tobacco smoke.These mutations may carry GOF effect that, al-though having no impact on natural tumor de-velopment (no prognostic value), accelerates tu-mor progression or relapse under adjuvanttreatment (negative predictive value). This par-ticular situation is a rare illustration of a possibleGOF effect of mutant TP53 in a large random-ized clinical trial. Presence of a non-DBM mu-tation appears to be associated with an adverseeffect of adjuvant therapy, suggesting that thesetreatments should be restricted to wild-typeTP53 cases. The molecular basis of this effectremains to be identified.

CONCLUDING REMARKS: SOLVINGTHE TP53 MUTATION EQUATION

The biological and clinical significance of TP53mutation is the result of a complex equation that

combines structural and functional terms. Themain structural terms are the mutation itself(position, type, and effect on p53 protein struc-ture), the retention (or not) of a functionalwild-type allele (LOH at TP53 locus), and thehaplotype structure of both wild-type and mu-tant alleles, which may be critical in the fine bal-ancing act by which one allele may dominateover the other. Current sequencing strategiesare tailored to assess only the first of these terms,and only in an incomplete manner because moststudies remain focused on mutations occurringin exons. A recent initiative of researchers in theTP53 mutation field, led by Thierry Soussi, isrecommending that the entire sequence of theTP53 gene should be taken into considerationwhen developing mutational analysis strategiesto assess TP53 status. As a next step, it will beimportant to design algorithms that enable theassignment of haplotypes and LOH from se-quencing data, thus providing the technical basisfor an assay capable of scoring all the structuralterms of the mutant TP53 equation.

The functional terms of the equation areeven more complex. Although the major gener-ic term is LOF, DN and GOF represent addi-tional variables that may be extremely context-dependent. There is currently no simple androbust assay to score DN and GOF functionsin a clinical setting. This complexity is com-pounded by the fact that, in many cancers, thep53 pathway may be disrupted by other mech-anisms than structural alteration of the TP53gene, so that tumors that apparently retain in-tact alleles actually do not have an intact p53pathway. Many studies have analyzed the tran-scriptomic patterns of TP53 dysfunction incancer cells and tissues but so far no robustconsensus signature has emerged. Although anumber of confirmed p53 target genes exist,their aberrant expression in cancer tissue maynot be detectable without an inducer or treat-ment, or the expression of these genes in can-cer tissue may be altered by p53-independentmechanisms. Efforts should be continued toidentify surrogate biomarkers of a disruptedp53 pathway. A promising approach is to inves-tigate p53-regulated long and short noncodingRNA networks (Donzelli et al. 2014).



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



Integrative genomics studies are confirmingthat somatic TP53 mutation lies at the very cen-ter of the patterns of alterations that character-ize the genomic landscape of cancer. They alsoconfirm that the biological and clinical signifi-cance of TP53 mutation is very context-depen-dent, with large differences among histologicalsubtypes of the same cancers. We propose todistinguish three patterns of association be-tween TP53 and genomic landscapes, highlight-ing their contrasting clinical messages.

High Mutation Frequency in Tumors withProfoundly Rearranged Genomes

In many subtypes of cancer (in particular insolid tumors), the mutation frequency is sohigh that it can be assumed that deregulationof the p53 pathway is an obligate mechanism forcarcinogenesis. These cancers often show verydisorganized genomes with high overall muta-tion load, often associated with intense expo-sure to endogenous or exogenous mutagens.They are poorly responsive to either convention-al or pathway-targeted therapies. This is the sit-uation observed in basal-like breast cancer or inSqCC of the lung. The overarching role of TP53mutation in these cancers might be to promoteand maintain metabolic and/or epigenetic pro-grams essential for the expansion of aggressivecancer cells with stem-like phenotypes. In turn,loss of p53 function may promote genomic in-stability by favoring the occurrence of massivecatastrophic genomic rearrangements such aschromothripsis (Rausch et al. 2012). In thesetumors, TP53 mutation in itself may not be ofmuch clinical significance for subclassifyingcases or stratifying patients. However, a deeperevaluation of all structural terms of the TP53mutation equation may uncover significant dif-ferences in the clinical behavior of cancers withdifferent mutants. Despite having an overallpoor prognosis, these tumors with frequentTP53 mutation show a range of clinical behav-iors that may depend on factors that modulatethe penetrance of TP53 mutation, such as LOH,haplotypes, and alteration of regulators of p53function. In breast cancer, combining TP53mutation, LOH, and MDM2 overexpression re-

veals an additive negative prognostic effect, sug-gesting that several mechanisms of p53 inacti-vation may cooperate within the same tumor(Silwal-Pandit et al. 2014).

Intermediate Mutation Frequency inOncogene-Driven Cancers

Knowledge of TP53 mutation status is provingof more direct relevance in tumors with inter-mediate TP53 mutation frequencies, such asEGFR-mutated lung AdC and HER2-enrichedBC. In these oncogene-driven cancers, TP53mutation occurs at frequencies between 30%and 50% and is a clear indicator of poor prog-nosis. In both cases, the driving event is theoncogenic activation of a cell-surface tyrosinekinase receptor of the EGF superfamily and itis plausible that TP53 mutation has similar con-sequences in both cancers. An interesting hy-pothesis is that this effect is caused by the in-volvement of p53 in regulating cell responsesto oncogenic signaling through the p14arf–Mdm2 pathway. Inactivation of p53 by muta-tion may switch off this pathway, thus eliminat-ing a natural safeguard against excessive cellproliferation (Lomazzi et al. 2002). In supportof this hypothesis, we have reported that p14arfexpression is often down-regulated in lungAdC with mutated EGFR and wild-type TP53,suggesting that this pathway is under strong neg-ative selection pressure in EGFR-mutated lungcancers (Mounawar et al. 2007; Cortot et al.2014). However, why this effect is seen withEGFR mutation or HER2 enrichment and notwith KRAS mutation is not clear. Manipulatingthe p53 pathway to increase and stabilize wild-type responses may prove a valuable companionapproach in the treatment of these cancers, inparticular with targeted therapies.

Low Mutation Frequencies in Tumorswith Actionable Wild-Type TP53

A broad range of cancers has rare TP53 muta-tions, in particular at early stages. These cancersinclude lesions of overall good prognosis suchas adenoma of the colon or luminal A breastcancers, in which rare TP53 mutations might



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



just be occasional passengers. Most of these le-sions are adequately curated by surgery. How-ever, TP53 mutations are also rare in severalcancer types requiring more aggressive treat-ment, such as melanoma (6%) or osteosarcoma(10%–12%). In these cases, a reassessment ofTP53 mutation status is warranted. It is possiblethat significant structural alterations in TP53have been missed in conventional studies be-cause they occur in regions of the gene thatare not routinely screened. A dramatic exampleof such a situation is observed in osteosarcoma,in which frameshift alterations/translocationsare detected in intron 1 in �50% of the caseswith no mutation in the coding sequence (Chenet al. 2014; Ribi et al. 2015). Another potentialintragenic mechanism of inactivation is over-expression of dominant-negative p53 isoforms,either caused by deregulation of transcriptionand/or splicing, or by mutations in intronsthat activate the expression of naturally occur-ring isoforms (Marcel et al. 2011). For melano-ma, the frequent deletion or mutation of theCDKN2A locus, which includes alterationsof p14/ARF, may be a contributing factor forthe lower mutation prevalence in the TP53gene itself (Hodis et al. 2012). The fact thatmany of these tumors have in common the re-tention of a potentially functional p53 pathwaysuggests that it may be possible to pharmaco-logically “resuscitate” a suppressor function. Aproof of principle has been given in cutaneousmelanoma, in which overexpression of MDM4promotes the survival of human metastatic mel-anoma by antagonizing p53 proapoptotic func-tion (Gembarska et al. 2012). Inhibition of theMDM4–p53 interaction restores p53 functionin melanoma cells, resulting in increased sensi-tivity to cytotoxic chemotherapy and to inhib-itors of the BRAF (V600E) oncogene, thus iden-tifying MDM4 as a promising target for therapyin melanoma.

In the coming years, TP53 mutation datawill continue to grow at an unprecedented, ex-ponential pace because at least the frequentlymutated exons of TP53 are included in mostcurrent next-generation sequencing panels.During the weeks of redaction of this article(May–July 2015), the COSMIC WGS/WES

TP53 mutation data set grew by more than1000 units. A large concerted effort is neededto mine and annotate this enormous mass ofinformation and to turn it into a source of de-tailed knowledge for the biologist, the clinicianand, ultimately, for the benefit of the patients.

REFERENCES�Reference is also in this collection.

Alexandrov LB, Nik-Zainal S, Wedge DC, Aparicio SA, Beh-jati S, Biankin AV, Bignell GR, Bolli N, Borg A, Borresen-Dale AL, et al. 2013a. Signatures of mutational processesin human cancer. Nature 500: 415–421.

Alexandrov LB, Nik-Zainal S, Wedge DC, Campbell PJ,Stratton MR. 2013b. Deciphering signatures of muta-tional processes operative in human cancer. Cell Rep 3:246–259.

Ambs S, Bennett WP, Merriam WG, Ogunfusika MO, OserSM, Harrington AM, Shields PG, Felley-Bosco E, Hus-sain SP, Harris CC. 1999. Relationship between p53 mu-tations and inducible nitric oxide synthase expression inhuman colorectal cancer. J Natl Cancer Inst 91: 86–88.

Baker SJ, Fearon ER, Nigro JM, Hamilton SR, Preisinger AC,Jessup JM, vanTuinen P, Ledbetter DH, Barker DF, Naka-mura Y, et al. 1989. Chromosome 17 deletions and p53gene mutations in colorectal carcinomas. Science 244:217–221.

Bartek J, Bartkova J, Vojtesek B, Staskova Z, Rejthar A, Ko-varik J, Lane DP. 1990. Patterns of expression of the p53tumour suppressor in human breast tissues and tumoursin situ and in vitro. Int J Cancer 46: 839–844.

Beroud C, Soussi T. 1998. p53 gene mutation: Software anddatabase. Nucleic Acids Res 26: 200–204.

Bonnefoi H, Piccart M, Bogaerts J, Mauriac L, Fumoleau P,Brain E, Petit T, Rouanet P, Jassem J, Blot E, et al. 2011.TP53 status for prediction of sensitivity to taxane versusnon-taxane neoadjuvant chemotherapy in breast cancer(EORTC 10994/BIG 1-00): A randomised phase 3 trial.Lancet Oncol 12: 527–539.

Bougeard G, Renaux-Petel M, Flaman JM, Charbonnier C,Fermey P, Belotti M, Gauthier-Villars M, Stoppa-LyonnetD, Consolino E, Brugieres L, et al. 2015. Revisiting Li–Fraumeni syndrome from TP53 mutation carriers. J ClinOncol 33: 2345–2352.

Brash DE, Rudolph JA, Simon JA, Lin A, McKenna GJ,Baden HP, Halperin AJ, Ponten J. 1991. A role for sunlightin skin cancer: UV-induced p53 mutations in squamouscell carcinoma. Proc Natl Acad Sci 88: 10124–10128.

Bressac B, Kew M, Wands J, Ozturk M. 1991. Selective G to Tmutations of p53 gene in hepatocellular carcinoma fromsouthern Africa. Nature 350: 429–431.

Brosh R, Rotter V. 2009. When mutants gain new powers:News from the mutant p53 field. Nat Rev Cancer 9: 701–713.

Chen X, Bahrami A, Pappo A, Easton J, Dalton J, Hedlund E,Ellison D, Shurtleff S, Wu G, Wei L, et al. 2014. Recurrentsomatic structural variations contribute to tumorigenesisin pediatric osteosarcoma. Cell Rep 7: 104–112.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



Clinical Lung Cancer Genome Project (CLCGP). 2013. Agenomics-based classification of human lung tumors. SciTransl Med 5: 209ra153.

Cooks T, Harris CC, Oren M. 2014. Caught in the cross fire:p53 in inflammation. Carcinogenesis 35: 1680–1690.

Cortot AB, Younes M, Martel-Planche G, Guibert B, Isaac S,Souquet PJ, Commo F, Girard P, Fouret P, Brambilla E,et al. 2014. Mutation of TP53 and alteration of p14arf

expression in EGFR- and KRAS-mutated lung adenocar-cinomas. Clin Lung Cancer 15: 124–130.

Courtois S, Caron de Fromentel C, Hainaut P. 2004. p53protein variants: Structural and functional similaritieswith p63 and p73 isoforms. Oncogene 23: 631–638.

Curtis C, Shah SP, Chin SF, Turashvili G, Rueda OM, Dun-ning MJ, Speed D, Lynch AG, Samarajiwa S, Yuan Y, et al.2012. The genomic and transcriptomic architecture of2,000 breast tumours reveals novel subgroups. Nature486: 346–352.

Denissenko MF, Pao A, Tang M, Pfeifer GP. 1996. Preferen-tial formation of benzo[a]pyrene adducts at lung cancermutational hotspots in P53. Science 274: 430–432.

Denissenko MF, Koudriakova TB, Smith L, O’Connor TR,Riggs AD, Pfeifer GP. 1998. The p53 codon 249 muta-tional hotspot in hepatocellular carcinoma is not relatedto selective formation or persistence of aflatoxin B1 ad-ducts. Oncogene 17: 3007–3014.

Donzelli S, Strano S, Blandino G. 2014. microRNAs: Shortnon-coding bullets of gain of function mutant p53 pro-teins. Oncoscience 1: 427–433.

Dorman SN, Viner C, Rogan PK. 2014. Splicing mutationanalysis reveals previously unrecognized pathways inlymph node–invasive breast cancer. Sci Rep 4: 7063.

Dumont P, Leu JI, Della Pietra AC, George DL, Murphy M.2003. The codon 72 polymorphic variants of p53 havemarkedly different apoptotic potential. Nat Genet 33:357–365.

Fernandez-Cuesta L, Anaganti S, Hainaut P, Olivier M.2011a. p53 status influences response to tamoxifen butnot to fulvestrant in breast cancer cell lines. Int J Cancer128: 1813–1821.

Fernandez-Cuesta L, Anaganti S, Hainaut P, Olivier M.2011b. Estrogen levels act as a rheostat on p53 levelsand modulate p53-dependent responses in breast cancercell lines. Breast Cancer Res Treat 125: 35–42.

Fernandez-Cuesta L, Oakman C, Falagan-Lotsch P, SmothKS, Quinaux E, Buyse M, Dolci MS, Azambuja ED, Hai-naut P, Dell’orto P, et al. 2012. Prognostic and predictivevalue of TP53 mutations in node-positive breast cancerpatients treated with anthracycline- or anthracycline/taxane-based adjuvant therapy: Results from the BIG02-98 phase III trial. Breast Cancer Res 14: R70.

Firestein GS, Echeverri F, Yeo M, Zvaifler NJ, Green DR.1997. Somatic mutations in the p53 tumor suppressorgene in rheumatoid arthritis synovium. Proc Natl AcadSci 94: 10895–10900.

Garritano S, Gemignani F, Palmero EI, Olivier M, Martel-Planche G, Le Calvez-Kelm F, Brugieres L, Vargas FR,Brentani RR, Ashton-Prolla P, et al. 2010. Detailed hap-lotype analysis at the TP53 locus in p.R337H mutationcarriers in the population of Southern Brazil: Evidencefor a founder effect. Hum Mutat 31: 143–150.

Gembarska A, Luciani F, Fedele C, Russell EA, Dewaele M,Villar S, Zwolinska A, Haupt S, de Lange J, Yip D, et al.2012. MDM4 is a key therapeutic target in cutaneousmelanoma. Nat Med 18: 1239–1247.

Gouas D, Shi H, Hainaut P. 2009. The aflatoxin-inducedTP53 mutation at codon 249 (R249S): Biomarker ofexposure, early detection and target for therapy. CancerLett 286: 29–37.

Gouas DA, Villar S, Ortiz-Cuaran S, Legros P, Ferro G, KirkGD, Lesi OA, Mendy M, Bah E, Friesen MD, et al. 2012.TP53 R249S mutation, genetic variations in HBX and riskof hepatocellular carcinoma in The Gambia. Carcinogen-esis 33: 1219–1224.

Greenblatt MS, Bennett WP, Hollstein M, Harris CC. 1994.Mutations in the p53 tumor suppressor gene: Clues tocancer etiology and molecular pathogenesis. Cancer Res54: 4855–4878.

Hainaut P, Hollstein M. 2000. p53 and human cancer: Thefirst ten thousand mutations. Adv Cancer Res 77: 81–137.

Hanahan D, Weinberg RA. 2011. Hallmarks of cancer: Thenext generation. Cell 144: 646–674.

Hodis E, Watson IR, Kryukov GV, Arold ST, Imielinski M,Theurillat JP, Nickerson E, Auclair D, Li L, Place C, et al.2012. A landscape of driver mutations in melanoma. Cell150: 251–263.

Hollstein M, Hergenhahn M, Yang Q, Bartsch H, Wang ZQ,Hainaut P. 1999. New approaches to understanding p53gene tumor mutation spectra. Mutat Res 431: 199–209.

Hollstein M, Moriya M, Grollman AP, Olivier M. 2013.Analysis of TP53 mutation spectra reveals the fingerprintof the potent environmental carcinogen, aristolochicacid. Mutat Res 753: 41–49.

Iggo R, Rudewicz J, Monceau E, Sevenet N, Bergh J, SjoblomT, Bonnefoi H. 2013. Validation of a yeast functional assayfor p53 mutations using clonal sequencing. J Pathol 231:441–448.

Jiang W, Wang XW, Unger T, Forgues M, Kim JW, HussainSP, Bowman E, Spillare EA, Lipsky MM, Meck JM, et al.2010. Cooperation of tumor-derived HBx mutantsand p53–249ser mutant in regulating cell proliferation,anchorage-independent growth and aneuploidy in a tel-omerase-immortalized normal human hepatocyte-de-rived cell line. Int J Cancer 127: 1011–1020.

� Joruiz SM, Bourdon JC. 2016. p53 isoforms: Key regulatorsof the cell fate decision. Cold Spring Harb Perspect Meddoi: 10.1101/cshperspect.a026039.

Kato S, Han SY, Liu W, Otsuka K, Shibata H, Kanamaru R,Ishioka C. 2003. Understanding the function–structureand function–mutation relationships of p53 tumor sup-pressor protein by high-resolution missense mutationanalysis. Proc Natl Acad Sci 100: 8424–8429.

Khoury MP, Bourdon JC. 2010. The isoforms of the p53protein. Cold Spring Harb Perspect Biol 2: a000927.

Kristensen VN, Lingjaerde OC, Russnes HG, Vollan HK,Frigessi A, Borresen-Dale AL. 2014. Principles and meth-ods of integrative genomic analyses in cancer. Nat RevCancer 14: 299–313.

Ladanyi M, Cha C, Lewis R, Jhanwar SC, Huvos AG, HealeyJH. 1993. MDM2 gene amplification in metastatic oste-osarcoma. Cancer Res 53: 16–18.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



Le Calvez F, Mukeria A, Hunt JD, Kelm O, Hung RJ, TaniereP, Brennan P, Boffetta P, Zaridze DG, Hainaut P. 2005.TP53 and KRAS mutation load and types in lung cancersin relation to tobacco smoke: Distinct patterns in never,former, and current smokers. Cancer Res 65: 5076–5083.

Lindley RA. 2013. The importance of codon context forunderstanding the Ig-like somatic hypermutation strand-biased patterns in TP53 mutations in breast cancer. CancerGenet 206: 222–226.

Lomazzi M, Moroni MC, Jensen MR, Frittoli E, Helin K.2002. Suppression of the p53- or pRB-mediated G1checkpoint is required for E2F-induced S-phase entry.Nat Genet 31: 190–194.

Luo JL, Tong WM, Yoon JH, Hergenhahn M, Koomagi R,Yang Q, Galendo D, Pfeifer GP, Wang ZQ, Hollstein M.2001. UV-induced DNA damage and mutations in Hupki(human p53 knock-in) mice recapitulate p53 hotspotalterations in sun-exposed human skin. Cancer Res 61:8158–8163.

Ma X, Rousseau V, Sun H, Lantuejoul S, Filipits M, Pirker R,Popper H, Mendiboure J, Vataire AL, Le Chevalier T, etal. 2014. Significance of TP53 mutations as predictivemarkers of adjuvant cisplatin-based chemotherapy incompletely resected non-small-cell lung cancer. MolOncol 8: 555–564.

Mahmoudi S, Henriksson S, Corcoran M, Mendez-Vidal C,Wiman KG, Farnebo M. 2009. Wrap53, a natural p53antisense transcript required for p53 induction uponDNA damage. Mol Cell 33: 462–471.

Marcel V, Vijayakumar V, Fernandez-Cuesta L, Hafsi H,Sagne C, Hautefeuille A, Olivier M, Hainaut P. 2010.p53 regulates the transcription of its D133p53 isoformthrough specific response elements contained within theTP53 P2 internal promoter. Oncogene 29: 2691–2700.

Marcel V, Dichtel-Danjoy ML, Sagne C, Hafsi H, Ma D,Ortiz-Cuaran S, Olivier M, Hall J, Mollereau B, HainautP, et al. 2011. Biological functions of p53 isoformsthrough evolution: Lessons from animal and cellularmodels. Cell Death Differ 18: 1815–1824.

Marin MC, Jost CA, Brooks LA, Irwin MS, O’Nions J, TidyJA, James N, McGregor JM, Harwood CA, Yulug IG, et al.2000. A common polymorph

Documents

Somatic TP53 Mutations in the Era of Genome Sequencingperspectivesinmedicine.cshlp.org/content/early/2016/08/...2016/08/06 · p53g) (Courtois et al. 2004; Khouryand Bour-don2010;Marceletal.2011).Althoughthepre-cise