Temporal Dissection of Tumorigenesis in Primary Cancers Research Brief RESEARCH BRIEF

Published OnlineFirst June 29, 2011; DOI: 10.1158/2159-8290.CD-11-0028

Temporal Dissection of Tumorigenesis in Primary Cancers research brief RESEARCH BRIEF

Temporal Dissection of Tumorigenesis in Primary cancers

Steffen Durinck 1,*, Christine Ho2, *, Nicholas J. Wang1, *, Wilson Liao3 , Lakshmi R. Jakkula1 , Eric A. Collisson1 , Jennifer Pons3 , Sai-Wing Chan3 , Ernest T. Lam3 , Catherine Chu3 , Kyunghee Park5 , Sung-woo Hong5 , Joe S. Hur6 , Nam Huh5 , Isaac M. Neuhaus3 , Siegrid S. Yu3 , Roy C. Grekin3 , Theodora M. Mauro3 , James E. Cleaver3 , Pui-Yan Kwok3 , Philip E. LeBoit4 , Gad Getz7 , Kristian Cibulskis7 , Jon C. Aster8 , Haiyan Huang2 , Elizabeth Purdom2 , Jian Li9, 10, Lars Bolund9, 10, Sarah T. Arron3 , Joe W. Gray 1, 11, Paul T. Spellman1, †, and Raymond J. Cho3, †

ABSTRACT Timely intervention for cancer requires knowledge of its earliest genetic aberrations. Sequencing of tumors and their metastases reveals numerous abnormalities occurring late in progression. A means to temporally order aberrations in a single cancer, rather than inferring them from serially acquired samples, would define changes preceding even clinically evident disease. We integrate DNA sequence and copy number information to reconstruct the order of abnormalities as individual tumors evolve for 2 separate cancer types. We detect vast, unreported expansion of simple mutations sharply demarcated by recombinative loss of the second copy of TP53 in cutaneous squamous cell carcinomas (cSCC) and serous ovarian adenocarcinomas, in the former surpassing 50 mutations per megabase. In cSCCs, we also report diverse secondary mutations in known and novel oncogenic pathways, illustrating how such expanded mutagenesis directly promotes malignant progression. These results reframe paradigms in which TP53 mutation is required later, to bypass senescence induced by driver oncogenes.

siGnificance: Our approach reveals sequential ordering of oncogenic events in individual cancers, based on chromosomal rearrangements. Identifying the earliest abnormalities in cancer represents a critical step in timely diagnosis and deployment of targeted therapeutics. Cancer Discovery; 1(2); 137–43. © 2011 AACR.

Authors’ Affiliations: 1 Life Sciences Division, Lawrence Berkeley National INTRODUCTION Laboratories; 2 Department of Statistics, University of California, Berkeley; 3 Molecular characterization of human cancers usually Department of Dermatology, University of California, San Francisco, and 4 San Francisco Dermatopathology Service, San Francisco, California; profiles a single point in time, yielding a catalog of genomic 5 Emerging Technology Research Center, Samsung Advanced Institute of and epigenetic abnormalities reflecting years of somatic Technology, and 6 Samsung Electronics Headquarters, Seoul, Korea; 7 The change. Although recent efforts reveal some mutations as- Eli and Edythe L. Broad Institute of Harvard and MIT, Cambridge; sociated with metastasis and recurrence ( 1 ), timing of events 8 Department of Pathology, Brigham and Women’s Hospital, Boston, Massachusetts; 9 Beijing Genomics Institute–Shenzhen, Shenzhen, China; early in tumorigenesis remains difficult. Precursor dysplas- 10 Institute of Human Genetics, Aarhus University, Aarhus, Denmark; tic lesions may be sampled and compared against invasive 11 Biomedical Engineering Department, Oregon Health and Science malignancies, but in many cancer types, early lesions are University, Portland, Oregon not clinically identifiable, nor is it obvious which lesions *These primary authors contributed equally to this article. will actually progress. In addition, the increasingly apparent † These senior authors contributed equally to this article. heterogeneity of human cancers suggests such comparisons Image rendered by Grace Yi-Wen Chang; “Leap II,” 2009, Monotype, will require very large sample sizes to reconstruct progres- 22 x 30 inches (http://www.gracechangstudio.com). sion ( 2 ). Note: Supplementary data for this article are available at Cancer Primary cutaneous squamous cell carcinomas (cSCC) Discovery Online (http://www.cancerdiscovery.aacrjournals.org). rank among the most common human malignancies, with Corresponding Author: Raymond J. Cho, Department of Dermatology, University of California, San Francisco, 1701 Divisadero Street, an annual incidence in Caucasians of >150 in 100,000 in- 4th Floor, San Francisco, CA 94115. Phone: 415-353-7880; Fax: 415- dividuals ( 3 ). These tumors arise in anatomic sites and with 353-7870; E-mail: [email protected] a demographic proportionate to sunlight exposure and doi: 10.1158/2159-8290.CD-11-0028 acquire a mutational spectrum reflecting significant UV ©2011 American Association for Cancer Research. radiation damage ( 4 ). Although many are excised without

JULY 2011 CANCER DISCOVERY | 137

research brief Durinck et al. complication, cSCCs sometimes behave aggressively, with loss. CN-LOH events at 17p represent 2% of coding sequence recurrence and regional spread, especially in immunosup- and show normalized mutation frequencies reflective of the pressed and repair-deficient genetic backgrounds (3). We remainder of the exome (Supplementary Fig. S3). Although hypothesized that the long-term mutational stress on these studies establish some p53 mutations as gain-of-function tumors might offer unique insight into the progressive events with respect to cancer type (10), or biochemically dominant determining a cancer’s individuality. negative, ours is the first to report that the vast majority of simple mutations—tens of thousands genome-wide in the case of cSCCs—appear sharply gated by elimination of the second RESULTS copy of TP53. We further detect at least partial persistence of Exome-level sequencing of 8 primary cSCCs and matched active DNA repair, suggesting that a profound loss of damage normal tissue revealed a very large mutation burden of ap- surveillance contributes to the high number of observed mu- proximately 1,300 somatic single-nucleotide variants per cSCC tations (Supplementary Fig. S1; ref. 11). exome (1 per ∙30,000 bp of coding sequence; Supplementary Three samples without CN-LOH at 17p show at least 2 dis- Fig. S1), making cSCCs among the most highly mutated hu- tinct TP53 mutations, presumably causing biallelic mutation. man malignancies. Of mutations assessed by capillary sequenc- In the sample in which TP53 mutations were not detected, a ing across our series, 75 of 75 (100%) confirmed the originally regionally duplicated mutation in the ATM kinase domain detected mutations, including 4 instances of dinucleotide sub- was observed, suggesting an alternative means of escaping stitution. C > T transition base substitutions at dipyrimidine damage surveillance mechanisms during telomere crisis (12). sites were by far the most common change (>85%), consistent We sought to validate our observations in an additional with UV damage. Past analysis of selected TP53 exons sug- cancer type. Recently, full genomic sequence and copy num- gests that the gene is mutated in 50–90% of cSCCs (5). Our ber changes were determined for 10 ovarian serous adeno- study identified TP53 mutations in 7 of 8 skin cancers, all co- carcinomas by The Cancer Genome Atlas Project. Ovarian inciding with previously reported changes in the Catalogue of cancers generally show more complex karyotypic abnormali- Somatic Mutations in Cancer (COSMIC) database (6). Known ties than do cSCCs (13). In the 3 samples with a clear, infor- changes were also found in CDKN2A, encoding the p16/p14ARF mative CN-LOH event at 17p, we again found solid evidence bifunctional tumor suppressor, the HRAS small GTPase, and for complete loss of TP53 as the earliest event (Fig. 1D). These instances of COSMIC mutations not previously described initial events in ovarian tumorigenesis could not have been in cSCCs (Table 1; full list of base substitutions provided in determined through sequencing of precursor lesions and in- Supplementary Table S1). We also detected 42 discrete chro- vasive cancers (1, 14), as the asymptomatic nature of early mosomal abnormalities, about 5 per sample (Supplementary disease precludes tissue collection. Table S2). Integrative analyses of copy number and exome sequence The ability to temporally order successive molecular also reveal information about the temporal order of chro- changes within an individual tumor, beginning in the initial mosomal abnormalities within an individual cancer (7). As stages of tumorigenesis, would allow discrimination between described above, the ratio of heterozygous to homozygous mutations forming precancerous lesions from those produc- mutations ρ, in a given region of CN-LOH, provides a direct ing invasive carcinomas. The high prevalence of both simple measure of the relative age of the duplication (Fig. 2A). In mutations and copy number abnormalities in cSCCs and other words, duplications with higher ρ occur earlier than ovarian cancers enabled us to reconstruct the evolutionary those with lower ρ. We found that ρ varied widely among re- order of some somatic changes based on the following idea: If gions of CN-LOH (Figs. 2B–D and 3) and could statistically a mutation precedes a regional duplication, its copy number distinguish the temporal order of aberrations within a sam- is doubled, whereas mutations following a duplication event ple (Fig. 2). Overall, 7 informative duplications co-occurring appear in haploid copy number (7). Therefore, (1) simple mu- with 17p CN-LOH all showed a substantially lower relative ρ tations preceding a chromosomal duplication event show (Supplementary Table S2) and thus likely occurred after 17p discretely higher copy numbers compared with those occur- duplication. Therefore, loss of the second TP53 allele appears ring after duplication (Fig. 1); and (2) the ratio of heterozy- to precede not only a vast expansion of simple mutations but gous to homozygous mutations ρ, in a region of copy-neutral also the development of chromosomal aberrations. As a gen- LOH (CN-LOH), directly measures the age of the duplication eral principle, any regional copy gain acquires a heterozygote in evolutionary time (Fig. 2; Supplementary Fig. S2). mutation frequency uniquely reflective of the time of gain. For We first used this principle to investigate the specific tem- selected instances in our series, extension of this principle en- poral order of mutations in areas of CN-LOH, in which a abled temporal dissection of more complex copy gains (Fig. 3), regional chromosomal duplication replaces the matching revealing that these alterations also follow complete TP53 loss. portion of the paired chromosome (8, 9). Of the cSCCs in In cSCCs, we found 486 nonsynonymous mutations that our series, 4 of 8 showed CN-LOH at chromosome 17p, all were sequenced deeply enough to determine copy num- harboring TP53 mutations reported in COSMIC. Remarkably, ber (>50 independent reads) and that fell at least once in all 4 TP53 mutations were present at high allelic abundance in a region of CN-LOH. These included known mutations in the CN-LOH region, compared with other somatic mutations, CDKN2A, WT1, and HRAS (Table 1), each of which showed indicating that TP53 mutations occurred and were duplicated multiple instances of either wild-type allele loss or biallelic before other mutations arose (Fig. 1B and C). In aggregate, mutation, as seen for TP53. Of interest, this pattern of recur- 59 of 63 mutations in 17p appear after loss of the second rent biallelic inactivation was also detected at high prevalence TP53 wild-type allele, 15-fold greater than those preceding for the suspected epithelial tumor suppressors NOTCH1 and

138 | CANCER DISCOVERY JULY 2011 www.aacrjournals.org

Temporal Dissection of Tumorigenesis in Primary Cancers research brief

Table 1. Description of sample cases and selected mutations of interest

Immune Other known Sex Age Site status TP53 CDKN2A NOTCH1 NOTCH2 NOTCH3 NOTCH4 COSMIC mutations 1. M 76 Scalp + r248W P135L Q610* W330*, EP300 , WT1 R1838* 2. M 87 Scalp + e285K P1771S, P226S WT1 R1595Q 3. M 84 Left dorsal + e224 C478F R1333C hand (Splice site) 4. F 61 Left cheek + Y220n W309* PIK3CG 5. M 83 Left cheek + H179Y, W1769* Q1634*, S1602F P278S T2278I 6. M 85 Right temple + P142N, P133L (Splice site) S1836F, EZH2 H179Y E297K 7. M 58 Left aural helix - E286K, Q1924* Q1616*, HRAS T329I, E349* G488D 8. M 63 Lower lip + none a HSPB2

NOTE: Boldface denotes loss of wild-type allele; * denotes nonsense mutations; shaded mutations are found in the COSMIC database. TP53 mutations in samples 1–4 are duplicated through CN-LOH on chromosome 17p. aIn sample 8, no mutation in TP53 was seen, but early, duplicated mutation was detected in ATM at G2024R, in the PI kinase–related domain. Given their sheer number, some TP53 mutations may act in true dominant negative fashion but were not detected in our series. COSMIC Notch family mutations are activating in hematologic malignancies; the truncations and disabling mutations reported in this study have not been previously described.

NOTCH2 and the polycystic kidney disease gene PKHD1 (See commonly form in sun-exposed skin ( 19 ; 20 ). Subsequent Supplementary Methods). NOTCH1 shows 3 instances of elimination of the second TP53 allele, often through recom- early truncation, 2 of which show wild-type loss; 1 case of bination, sharply demarcates a vast expansion in simple multiple mutation; and 2 other mutations, 1 of which occurs mutations, in cSCCs reaching 50 per megabase (150,000 per in a splice site ( Table 1 ). NOTCH2 shows multiple mutations genome) and making them the most mutagenized human in 4 of 8 samples, and 3 of these contain at least 1 truncating cancers known. Because DNA repair remains at least partially mutation. active, this vast mutation burden might result from the col- laborative effects of ongoing DNA damage (from intrinsic and exogenous insults) coupled with disabled DNA damage– DISCUSSION induced apoptosis. The reproduction of this phenomenon We trace the mutational evolution of individual tumors, in ovarian adenocarcinoma suggests that the vast majority using a novel, sequence-based assessment strategy and, in of mutations follow second TP53 allele loss, irrespective of doing so, provide a patient-centric complement to more tra- mode of DNA damage or tissue of origin. ditional “mutation-by-stage” approaches ( 2 , 14 ). Our results Classic studies report that precursor lesions and invasive illuminate key aspects of timing in cancer evolution without cancers both carry mutated driver oncogenes, but find TP53 requiring large sample series, for which precursor lesions inactivation more frequent in invasive disease ( 15 , 21 ), sug- are often inaccessible. TP53 is often mutated in precursor le- gesting p53 inactivation to be a late event. Activation of a key sions, but paradigms of oncogenesis propose p53 loss as a oncogene prior to biallelic TP53 loss in our series is formally late requirement, overcoming senescence programs activated possible, but few coding mutations precede 17p duplication, by prior activation of driver oncogenes ( 15 ; 16 ) and enabling and none recur in established oncogenes. Although appar- survival through telomere crisis ( 17 ). Furthermore, biallelic ently contradictory, these findings could be reconciled by TP53 loss occurs frequently, despite evidence that p53 mu- a temporal requirement that TP53 mutation precede driver tants behave dominantly both structurally and functionally oncogene mutation in precursor lesions destined to prog- with respect to phenotypes such as tumor formation ( 10 , 18 ). ress to invasive cancer. In this model, precursor lesions that Our data reveal that decades of UV damage and in- activate oncogenes first (before TP53 inactivation) fail to activation of a single TP53 allele result in only about 100 progress, but would nonetheless be detected in “mutation mutations in the epithelial exome. This tenacious genetic frequency-by-stage” surveys. Alternatively, different cancer stability explains the benign behavior of clonal keratinocyte types might exhibit distinct temporal ordering of key muta- proliferations, harboring heterozygous TP53 mutation, that tions. Application of our approach to sequence data from

JULY 2011 CANCER DISCOVERY | 139

research brief Durinck et al.

B 1

TP53 1 1

1 TP53

TP53TP53 0.5 Allele frequency 0.5 0.5 0.5 Allele frequency 0 1 Allele frequency Allele frequency 0

C 1 ALKBH5

0 TP53 1 0

1 ALKBH5 TP53 0.5 ALKBH5 TP53 Allele frequency 0.5 ALKBH5 TP53 0.5 Allele frequency 01 0.5 Allele frequency Non–coding

01 Non–coding

Allele frequency TP53 Non–coding

01 Non–coding D TP53 Non–coding

0.5 Non–coding 01 TP53 Allele frequency

Non–coding0.5 Non–coding

Allele frequency TP53 0.5 0 0 10 20 30 40 50 60 70 80 Allele frequency Position (MB) 0 0.5 0 10 20 30 40 50 60 70 80 Position (MB) CNLOH region Nonheterozgous mutation, non–CNLOH region 0 Allele frequency 0 Single10 nucleotide polymorphism20 30 40 Estimated50 homozygous 60allele frequency 70 80 Homozygous mutation, CNLOH region Estimated heterozygous allele frequency CNLOH region Nonheterozgous mutation, non–CNLOH region Heterozygous mutation, CNLOH region Position (MB)Centromere Single nucleotide polymorphism Estimated homozygous allele frequency Heterozygous mutation, non–CNLOH region Homozygous mutation, CNLOH region Estimated heterozygous allele frequency

0 HeterozygousCNLOH region mutation, CNLOH region CentromereNonheterozgous mutation, non–CNLOH region 0 Heterozygous10Single nucleotide mutation,20 polymorphism non–CNLOH30 region 40 Estimated50 homozygous allele60 frequency 70 80 Homozygous mutation, CNLOH region Position (MB)Estimated heterozygous allele frequency Heterozygous mutation, CNLOH region Centromere Heterozygous mutation, non–CNLOH region CNLOH region Nonheterozgous mutation, non–CNLOH region Single nucleotide polymorphism Estimated homozygous allele frequency Homozygous mutation, CNLOH region Estimated heterozygous allele frequency Figure 1. LossHeterozygous of second TP53 mutation, wild-type CNLOH allele region precedes simple mutationsCentromere in cSCCs and ovarian cancers. A, schematic for acquisition of discretely higherHeterozygous copy number mutation, for mutations non–CNLOH preceding region a regional duplication, enabling temporal ordering of mutations during tumorigenesis. For representative samples, mutant allele frequency (y-axis) is shown plotted against physical position on chromosome 17 (x-axis) for cSCCs (B and C) and ovarian cancers (D). More than 50 independent reads were required for inclusion of a mutation. Estimated allele frequencies are shown as lines for heterozygous (solid) and homozygous (dotted) mutations. Regions without CN-LOH in ovarian cancers often show complex ploidy, generating greater variance in mutation frequency; these simple frequency estimates were therefore not applied. Mutation frequencies are centered below 1.0 for homozygotes and 0.5 for heterozygotes because of the effects of partial dilution with nontumorous cells (see Supplementary Methods for more details).

140 | CANCER DISCOVERY JULY 2011 www.aacrjournals.org

Temporal Dissection of Tumorigenesis in Primary Cancers research brief

A EarlyEarly copy-neutral copy-neutral LOH LOH EarlyEarly copy-neutral copy-neutral LOH

NoNo copy-neutral copy-neutral LOH LOH NoNo copy-neutral copy-neutral LOH

HomozygousHomozygous mutation mutation HomozygousHomozygous mutation mutation HeterozygousHeterozygous mutation mutation HeterozygousHeterozygous mutation mutation LateLate copy-neutral copy-neutral LOH LOH LateLate copy-neutral copy-neutral LOH

cSCCcSCC #1, #1, Chromosome Chromosome 17 17 B cSCC #1, Chromosome 17 1 cSCC1 #1, Chromosome 17 1 1 TP53TP53 TP53 TP53 0.5 0.5 0.5 Counts Counts Counts 0.5 Allele frequency Allele frequency Allele frequency Counts Allele frequency

TP53TP53 0 0 0 2 4 6 8 TP53 0 2 4 6 8 0 00 1010 20 20 30 30 40 40 5050 6060 70 70 80 80 00 0.50.5 TP5311 0 2 4 6 8 0 0 10 20 30 40 50 60 70 80 0 0.5 1 0 2 4 6 8 PositionPosition (MB) (MB) AlleleAllele frequency frequency 0 10 20 30Position 40 (MB) 50 60 70 80 0 Allele 0.5frequency 1 C cSCCcSCC #1, #1, Chromosome Chromosome 2 2 Position (MB) Allele frequency 1 cSCC1 #1, Chromosome 2 cSCC1 #1, Chromosome 2 COL4A4COL4A4 TANC1TANC1 COL4A4 1 GG 2 2 TANC1 PLCL1PLCL1 COL6A3COL6A3 G 2 PLCL1 COL4A4COL6A3 TANC1IFIH1IFIH1 C2orf57C2orf57 IFIH1 LRP2LRP2 C2orf57ITM2CITM2C IL1RL2IL1RL2G 2 LRP2PLCL1 COL6A3CAB39ITM2CCAB39 IL1RL2 IFIH1 C2orf57CAB39 0.5 0.5 LRP2 ITM2C 0.5 Counts IL1RL2 CAB39 Counts Counts Allele frequency Allele frequency Allele frequency 0.5 Counts Allele frequency 0 0 0 2 4 6 8 10 0 2 4 6 8 10 0 00 5050 100100 150150 200200 00 0.50.5 11 0 2 4 6 8 10 0 50 100 150 200 0 0.5 1 PositionPosition (MB) (MB) AlleleAllele frequency frequency 0 Position (MB) Allele frequency 0 2 4 6 8 10 0 50 100 150 200 0 0.5 1 D cSCCcSCC #1, #1, Chromosome Chromosome 14 14 Allele frequency 1 cSCC1 #1, Chromosome 14 Position (MB) CNLOHCNLOH region region 1 CNLOHSingleSingle nucleotideregionnucleotide cSCC #1, Chromosome 14 C14orf115C14orf115 SNAPC1SNAPC1SPTBSPTB C14orf115 Single nucleotide SLC7A7SLC7A7 AKAP6AKAP6 SOS2SOS2 polymorphismpolymorphism 1 SNAPC1 SPTBGALNTL1GALNTL1PCNXPCNX SLC7A7 ADCY4ADCY4 AKAP6 SOS2 polymorphismCNLOH region ADCY4 SOCS4SOCS4 GALNTL1PCNXHEATR4HEATR4 HomozygousHomozygous mutation, mutation, SALL2SALL2 CPNE6CPNE6 PYGLPYGL DAAM1DAAM1KCNH5KCNH5 BTBD7BTBD7 BAG5BAG5 Homozygous mutation, SOCS4 C14orf115HEATR4 Single nucleotide SALL2 CPNE6 PYGL DAAM1KCNH5 BTBD7 BAG5 CNLOHCNLOH region region SNAPC1 SPTB NEK9NEK9 BCL11BBCL11B CNLOH region SLC7A7OR4M1OR4M1 AKAP6 SOS2 SAMD4ASAMD4A polymorphismHeterozygousHeterozygous mutation, mutation, OR4M1ADCY4 GALNTL1PCNXNEK9 BCL11B 0.5 0.5 SAMD4A Heterozygous mutation, SOCS4 HEATR4 HomozygousCNLOHCNLOH region region mutation, 0.5 CPNE6 Counts SALL2 Counts PYGL DAAM1KCNH5 BTBD7 BAG5 CNLOHHeterozygousHeterozygous region mutation, mutation,

Counts CNLOH region Allele frequency Allele frequency NEK9 BCL11B Heterozygous mutation, OR4M1Allele frequency Heterozygousnon–CNLOHnon–CNLOH region region mutation, SAMD4A non–CNLOH region 0.5 CNLOHEstimatedEstimated region homozygous homozygous

Counts Estimated homozygous Heterozygousalleleallele frequency frequency mutation, Allele frequency alleleEstimatedEstimated frequency heterozygous heterozygous

0 2 4non–CNLOH 6 8 10 12 region 0 2 4 6 8 10 12 0 0 Estimated heterozygous 0 2 4 6 8 10 12 0 2020 30 30 40 40 50 50 6060 7070 80 80 90 90 100 100 00 0.50.5 11 Estimatedalleleallele frequency frequency homozygous 20 30 40 50 60 70 80 90 100 0 0.5 1 alleleCentromereCentromere frequency PositionPosition (MB) (MB) AlleleAllele frequency frequency Centromereallele frequency Position (MB) Allele frequency Estimated heterozygous 0 2 4 6 8 10 12 0 20 30 40 50 60 70 80 90 100 0 0.5 1 allele frequency Centromere Figure 2. Conceptual framework for Positiontemporal (MB) ordering of chromosomal aberrations. A, schematicAllele frequency for differential mutant allele frequencies in areas of CN-LOH, based on relative age of chromosomal event during tumorigenesis. Earlier events allow greater time for accumulation of heterozygous mutations, whereas later events are predominated by homozygous mutations. Three events are ordered in a single sample, with the earliest on chromosome 17 (B), later on chromosome 2 (C), and last on chromosome 14 (D). Relative mutant allele frequency (y-axis) is plotted against chromosomal physical position (x-axis) for cSCCs. To the right of each panel, a histogram of mutant allele frequencies for each CN-LOH event shows density (y-axis) plotted against allele frequency (x-axis). A modified binomial distribution function (see Supplementary Methods) distinguishes the proportion of heterozygote and homozygote mutant frequencies ρ, ordering duplications in chromosome 17 (0.00–0.12), chromosome 2 (0.17–0.45), and chromosome 14 (0.65–0.94).

JULY 2011 CANCER DISCOVERY | 141

research brief Durinck et al.

Figure 3. For 2 regional chromosomal aberrations, mutant allele frequencies reveal distinct evolutionary histories. On chromosome 11, 2 peaks in mutation frequency are seen, 1 for heterozygous and 1 for homozygous mutations; these reflect a simple regional gain, including a COSMIC mutation in WT1. In contrast, 3 separate frequencies are detected on chromosome 9, culminating in a triploid state for mutant CDKN2A and loss of all wild-type copies. Three discrete allele frequencies can be achieved only by 2 gains and 1 loss, with the highest frequency (triploid) peak representing the first gain and the diploid peak containing mutations from the interval preceding the second gain. The disproportionate size of the triploid peak (P < 0.03) indicates that the first gain occurred more than halfway through tumor evolution, as measured by the evolutionary clock, and the second gain even later (see Supplementary Methods for more detail).

other cancer types, such as colon adenocarcinoma, should genomic aberration history can be applied immediately to help distinguish these possibilities. any cancer for which sequence data and copy number are In selected instances, we are able to show mutant TP53 du- available. plication occurring before dosage changes in mutant alleles, such as for CDKN2A and WT1 (Fig. 3). The consequences METHODS of such expanded mutagenesis and chromosomal instabil- We obtained 8 matched cSCCs and normal tissue samples as part ity emerge dramatically in the Notch signaling pathway, in of a skin cancer study protocol, with all subjects providing informed which multiple family members develop mutations and wild- consent according to procedures approved by the University of type alleles are lost frequently. Constitutive Notch protein California, San Francisco (UCSF), Committee on Human Research, activation drives subsets of acute leukemias (22), but a clear San Francisco, California. Diagnosis of cSCCs was confirmed for all tumor suppressor phenotype has also been established in ke- tumors via histologic examination of a standard biopsy specimen ratinocytes (23), with attentuated expression producing pro- by a board-certified dermatopathologist. DNA was extracted from liferation and invasive morphologies. Further study should tumor and control samples, and allele-specific copy number analysis clarify the breadth of epithelial cancers harboring these so- was performed using Affymetrix Genome-Wide Human SNP Array matic changes, as well as their specific functional effects. Our 6.0 chips. Approximately 40 megabases of coding region were iso- data also confirm low-prevalence activation of known onco- lated from each sample, using oligonucleotide-based hybrid capture and sequenced with the Illumina sequencing-by-synthesis platform genes such as Ras in human cSCCs, raising the possibility (Supplementary Fig. S4). Mutation detection was performed as previ- that numerous mutations in other pathways may serve as the ously described (see Supplementary Methods for detail). Seventy-five functional oncogene in this setting (24). mutations were independently validated using Sanger sequencing; Taken together, these insights imply that targeting acti- 100% confirmed the originally identified somatic change. Sequences vated oncogenes (e.g., those with small molecule inhibitors) for all nonsynonymous mutations will be deposited in the database fails to address a fundamental, detectable abnormality in of Genotypes and Phenotypes. Patient information and genomic cancer genomes that accelerates evolution toward clinical profiling for ovarian serous adenocarcinomas analyzed in this study resistance. Temporal dissection of tumorigenesis provides have been described previously (25). early, assayable diagnostic markers and illuminates the spe- Chromosomal regions with aberrant copy number, including cific biological consequences of these aberrations. We show CN-LOH and simple copy gains and losses, were identified based on discrete shifts in single nucleotide polymorphism (SNP) copy the utility of this method for the CN-LOH and copy gains number from both SNP array and exome sequencing data. The that constitute most chromosomal aberrations. Because type of abnormality was further confirmed by assessing raw copy many cancer types carry rearrangement of substantial pro- number depth from SNP array data. Mutations were called after portions of the genome, especially those spanning key on- alignment of sequence reads to a reference genome (NCBI36). cogenes, extension of this method should rapidly reveal The fraction of chromosomal copies carrying a mutation was esti- additional ordered events. The described reconstruction of mated as the fraction of all independent sequence reads containing

142 | CANCER DISCOVERY JULY 2011 www.aacrjournals.org

Temporal Dissection of Tumorigenesis in Primary Cancers research brief

that mutation. (A detailed description of patient consent, meth- 3. Madan V, Lear JT, Szeimies R. Non-melanoma skin cancer. Lancet ods, and reagents used for tissue acquisition, genomic profiling, 2010;375:673–85. and statistical analysis of mutational evolution is provided in the 4. Ziegler A, Jonason AS, Leffell DJ, Simon JA, Sharma HW, Supplementary Methods.) Kimmelman J, et al. Sunburn and p53 in the onset of skin cancer. Nature 1994;372:773–6. 5. Leffell DJ. The scientific basis of skin cancer. J Am Acad Dermatol Disclosure of Potential Conflicts of Interest 2000;42:18–22. No potential conflicts of interest were disclosed. 6. Forbes SA, Tang G, Bindal N, Bamford S, Dawson E, Cole C, et al. COSMIC (the Catalogue of Somatic Mutations in Cancer): a resource to investigate acquired mutations in human cancer. Nucleic Acknowledgments Acids Res 2010;38:D652–57. We thank Allan Balmain, Boris Bastian, Jeffrey Cheng, Douglas 7. Pleasance ED, Cheetham RK, Stephens PJ, McBride DJ, Humphray Brash, and Dennis Oh for early support and helpful discussions; and SJ, Greenman CD, et al. A comprehensive catalogue of somatic muta- Henrik Bengtsson, Pierre Neuvial, Hubert Stoppler, Matthew Akana, tions from a human cancer genome. Nature 2010;463:191–6. Connie Ha, Lauren Lee, Annie Poon, and Eric Dybbro for technical 8. O’Keefe C, McDevitt MA, Maciejewski JP. Copy neutral loss of het- assistance. erozygosity: a novel chromosomal lesion in myeloid malignancies. Blood 2010;115:2731–9. Grant Support 9. Tuna M, Knuutila S, Mills GB. Uniparental disomy in cancer. Trends Mol Med 2009;15:120–8. W. Liao is supported by a Dermatology Foundation Psoriasis 10. Olive KP, Tuveson DA, Ruhe ZC, Yin B, Willis NA, Bronson RT, et al. Career Development Award and the National Institute of Arthritis, Mutant p53 gain of function in two mouse models of Li-Fraumeni Musculoskeletal, and Skin Diseases (K08AR057763); E.A. Collisson syndrome. Cell 2004;119:847–60. by NIH/NCI K08 CA137153; J.E. Cleaver by the University of 11. Wu X, Levine AJ. p53 and E2F-1 cooperate to mediate apoptosis. California Cancer Coordinating Committee; S.T. Arron by NIH/ Proc Natl Acad Sci U S A 1994;91:3602–6. National Center for Research Resources/OD UCSF–Clinical & 12. Cimprich KA, Cortez D. ATR: an essential regulator of genome integ- Translational Science Institute Grant KL2 RR024130, a Canary rity. Nat Rev Mol Cell Biol 2008;9:616–27. Foundation/American Cancer Society Postdoctoral Fellowship 13. Cooke SL, Ng CKY, Melnyk N, Garcia MJ, Hardcastle T, Temple J, for the Early Detection of Cancer, and a Dermatology Foundation et al. Genomic analysis of genetic heterogeneity and evolution in Career Development Award in Dermatologic Surgery; and R.J. Cho high-grade serous ovarian carcinoma. Oncogene 2010;29:4905–13. by a Dermatology Foundation Career Development Award and as a 14. Ding L, Ellis MJ, Li S, Larson DE, Chen K, Wallis JW, et al. Genome Samsung Biotechnology Scholar-in-Residence. remodelling in a basal-like breast cancer metastasis and xenograft. This research was supported under J.W. Gray [by the Director, Nature 2010;464:999–1005. Office of Science, Office of Biological and Environmental Research, 15. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigen- U.S. Department of Energy, under Contract DE-AC02-05CH11231; esis. Cell 1990;61:759–67. by NIH, National Cancer Institute (NCI) Grants P50 CA 58207, 16. Hruban RH, Goggins M, Parsons J, Kern SE. Progression model for U54 CA 112970, and NHGRI U24 CA 126551; by the Department pancreatic cancer. Clin Cancer Res 2000;6:2969–72. 17. Chin L, Artandi SE, Shen Q, Tam A, Lee SL, Gottlieb GJ, et al. p53 of the Army, Award W81XWH-07-1-0663 (The U.S. Army Medical deficiency rescues the adverse effects of telomere loss and cooper- Research Acquisition Activity, Fort Detrick, Maryland, is the award- ates with telomere dysfunction to accelerate carcinogenesis. Cell ing and administering acquisition office); and by the Stand Up To 1999;97:527–38. Cancer–American Association for Cancer Research Dream Team 18. Milner J, Medcalf EA. Cotranslation of activated mutant p53 with Translational Cancer Research Grant SU2C-AACR-DT0409. The wild type drives the wild-type p53 protein into the mutant confor- content of this information does not necessarily reflect the position mation. Cell 1991;65:765–74. or the policy of the federal government, and no official endorsement 19. Jonason AS, Kunala S, Price GJ, Restifo RJ, Spinelli HM, Persing should be inferred]; P.T. Spellman (by NIH/NCI U24 CA1437991); JA, et al. Frequent clones of p53-mutated keratinocytes in normal R.J. Cho (by an unrestricted gift grant from the Samsung Advanced human skin. Proc Natl Acad Sci U S A 1996;93:14025–9. Institute of Technology); T.M. Mauro (by NIH Grants AR051930 20. Ren ZP, Ahmadian A, Pontén F, Nistér M, Berg C, Lundeberg J, et al. and R01AG028492, and the Medical Research Service, Department Benign clonal keratinocyte patches with p53 mutations show no of Veterans Affairs); and A. Balmain (by NIH/National Institute of genetic link to synchronous squamous cell precancer or cancer in Arthritis and Musculoskeletal and Skin Diseases Program Project human skin. Am. J. Pathol. 1997;150:1791–1803. Grant 5-P01-AR050440-05). 21. Baker SJ, Preisinger AC, Jessup JM, Paraskeva C, Markowitz S, Willson JK, et al. p53 gene mutations occur in combination with 17p Received March 7, 2011; revised May 12, 2011; accepted May 17, allelic deletions as late events in colorectal tumorigenesis. Cancer Res 2011; published OnlineFirst June 29, 2011. 1990;50:7717–22. 22. Weng AP, Ferrando AA, Lee W, Morris JP, Silverman LB, Sanchez- Irizarry C, et al. Activating mutations of NOTCH1 in human T cell acute lymphoblastic leukemia. Science 2004;306:269–71. 23. Lefort K, Mandinova A, Ostano P, Kolev V, Calpini V, Kolfschoten I, References et al. Notch1 is a p53 target gene involved in human keratinocyte . 1 Campbell PJ, Yachida S, Mudie LJ, Stephens PJ, Pleasance ED, tumor suppression through negative regulation of ROCK1/2 and Stebbings LA, et al. The patterns and dynamics of genomic instabil- MRCKα kinases. Genes Dev 2007;21:562–77. ity in metastatic pancreatic cancer. Nature 2010;467:1109–13. 24. Simon MM, Sliutz G, Luger TA. Proto-oncogenes and oncogenes in 2. Bozic I, Antal T, Ohtsuki H, Carter H, Kim D, Chen S, et al. epidermal neoplasia. Exp Dermatol 1995;4:65–73. Accumulation of driver and passenger mutations during tumor pro- 25. The Cancer Genome Atlas Group. Integrated genomic analyses of gression. Proc Natl Acad Sci U S A 2010;107:18545–50. ovarian carcinoma. Nature 2011;474:609–15.

JULY 2011 CANCER DISCOVERY | 143

Temporal Dissection of Tumorigenesis in Primary Cancers

Steffen Durinck, Christine Ho, Nicholas J. Wang, et al.

Cancer Discovery 2011;1:137-143. Published OnlineFirst June 29, 2011.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-11-0028

Supplementary Access the most recent supplemental material at: Material http://cancerdiscovery.aacrjournals.org/content/suppl/2011/06/29/2159-8290.CD-11-0028.DC1

Cited articles This article cites 25 articles, 8 of which you can access for free at: http://cancerdiscovery.aacrjournals.org/content/1/2/137.full#ref-list-1

Citing articles This article has been cited by 27 HighWire-hosted articles. Access the articles at: http://cancerdiscovery.aacrjournals.org/content/1/2/137.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department Subscriptions at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerdiscovery.aacrjournals.org/content/1/2/137. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.