subclonal genomic architectures of primary and metastatic ... · biology of human tumors subclonal...

Biology of Human Tumors

Subclonal Genomic Architectures of Primary andMetastatic Colorectal Cancer Based onIntratumoral Genetic HeterogeneityTae-Min Kim1,2, Seung-Hyun Jung3,4, Chang Hyeok An5, Sung Hak Lee6, In-Pyo Baek3,4,Min Sung Kim7, Sung-Won Park3,4, Je-Keun Rhee1, Sug-Hyung Lee2,3,7, andYeun-Jun Chung2,3,4

Abstract

Purpose: The intratumoral heterogeneity (ITH) and the evo-lution of genomic architectures associated with the developmentof distant metastases are not well understood in colorectalcancers.

Experimental Design: We performed multiregion biopsies ofprimary and liver metastatic regions from five colorectal cancerswith whole-exome sequencing and copy number profiling.

Results: In addition to a substantial level of genetic ITH,multiregion genetic profiling identifies the subclonal mutationalarchitecture, leading to the region-based or spatial categorizationof somatic mutations and the inference of intratumoral evolu-tionary history of cancers. The universal mutations (thoseobserved in all the regional biopsies) are enriched in knowncancer genes such as APC and TP53 with distinct mutationalspectra compared with biopsy- or region-specific mutations,suggesting thatmajor operativemutationalmechanisms and their

selective pressures are not constant across the metastatic progres-sion. The phylogenies inferred from genomic data show branch-ing evolutionary patterns where some primary biopsies are oftensegregated with metastastic lesions. Our analyses also revealedthat copy number changes such as the chromosomal gains of c-MYC and chromothripsis can be region specific and the potentialsource of genetic ITH.

Conclusions:Ourdata show that the genetic ITH is prevalent incolorectal cancer serving as a potential driving force to generatemetastasis-initiating clones and also as a means to infer theintratumoral evolutionary history of cancers. The paucity ofrecurrent metastasis-clonal events suggests that colorectal cancerdistant metastases may not follow a uniform course of genomicevolution, which should be considered in the genetic diagnosisand the selection of therapeutic targets for the advanced colorectalcancer. Clin Cancer Res; 21(19); 4461–72. �2015 AACR.

IntroductionColorectal cancers are the third most common human malig-

nancy worldwide with a�5% lifetime risk (1). Distant metastasisis a major cause of cancer-related morbidity and mortality ofcolorectal cancer,mainly due to livermetastasis that is observed inapproximately 25% of primary diagnoses of colorectal cancer. Inspite of the clinical successes in early detection and prevention of

colorectal cancer that resulted in an overall reduction of colorectalcancer risks (2), metastatic colorectal cancer still remains one ofthe leading causes of cancer-related death, with few therapeuticoptions available. Therefore, a better understanding of themolec-ular and genetic mechanisms underlying the biological andphenotypic evolution of colorectal cancer during metastasis iscrucial to reduce morbidity and mortality of this disease.

As previously proposed (3), cancer evolution is described assuccessive cycles of somatic mutation acquisition and the naturalselection of the fittest subclones, eventually giving rise to meta-static subclones (4, 5). Assuming that the development andprogression of distant metastasis are genetic events, there havebeen efforts to identify causal genetic determinants responsiblefor distant metastasis. Recently, massively paralleled sequencingof breast cancer revealed that the majority of mutations observedin the metastatic lesions were also present, but often in low allelefrequencies in the primary tumors (6). This finding was alsosupported by a direct comparison of mutation profiles obtainedfrom primary and metastatic breast cancers (7). These resultssupport the presence of genetic evolution associated with thedevelopment of distant metastasis, also raising a possibility thatthe clonal ancestry of distant metastasis can be traced to primarytumors.

It is nowwell recognized that human cancers have a substantiallevel of intratumoral heterogeneity (ITH) andmutational ITH hasbeen demonstrated across multiple human cancer types (8–15).Genetic ITH is the manifestation of a driving force in cancer

1Department of Medical Informatics, College ofMedicine,TheCatholicUniversity of Korea, Seoul, South Korea. 2Cancer Evolution ResearchCenter, College of Medicine, The Catholic University of Korea, Seoul,SouthKorea. 3IntegratedResearchCenter forGenomePolymorphism,College of Medicine, The Catholic University of Korea, Seoul, SouthKorea. 4Department of Microbiology, College of Medicine, The Cath-olic University of Korea, Seoul, South Korea. 5Department of Surgery,College of Medicine, The Catholic University of Korea, Seoul, SouthKorea. 6Department of Hospital Pathology, College of Medicine, TheCatholic University of Korea, Seoul, South Korea. 7Department ofPathology, College of Medicine, The Catholic University of Korea,Seoul, South Korea.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

CorrespondingAuthors:Yeun-Jun Chung, The Catholic University of Korea, 222Banpo-daero Seocho-gu, Seoul 137-701, Korea. Phone: 82-2-2258-7343; Fax: 82-2- 537-0572; E-mail: [email protected]; or Sug-Hyung Lee,[email protected]

doi: 10.1158/1078-0432.CCR-14-2413

�2015 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 4461

on November 13, 2020. © 2015 American Association for Cancer Research.clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 15, 2015; DOI: 10.1158/1078-0432.CCR-14-2413

http://clincancerres.aacrjournals.org/

genomes for the emergence of new subclones that havemetastaticpotential. Thus, ITH should be taken into account when identi-fying a segregation pattern of metastasis-related mutations inprimary tumors (16). To account for the regional ITH of muta-tions, Yachida and colleagues (17) performed multiregion biop-sies in both primary and metastatic pancreatic tumors and iden-tified that a primary tumor mass is a mixture of geographicallydistinct subclones, some of which might specifically give rise todistant metastases. However, the extent of genetic ITH in primaryand metastatic colorectal cancer lesions and the subclonal archi-tectures specifically associated with the development of distantmetastasis are largely unknown.

To identify genetic alterations associated with the metastaticprogression of colorectal cancer, we performedmultiregion biop-sies in both primary andmetastatic colorectal cancer lesions fromfive colorectal cancer cases, and analyzed somatic mutations andcopy number alterations (CNA). In addition to the apparentgenetic ITH in both primary and metastatic lesions, we foundbranched evolutionary patterns in the colorectal cancer genomes,in which preexisting subclones in primary lesions often appearedresponsible for the development of distant metastases. Themuta-tions showed distinct mutational spectra, enriched molecularfunctions and aberration types (e.g., chromothripsis) accordingto their regional distribution and intratumoral recurrences (i.e.,spatial mutation categories) giving clues to the subclonal archi-tectures and potential metastasis-associated genomic changes ofthe corresponding tumors.

Materials and MethodsTumor specimen

Colectomy tissues from five colorectal cancer patients used forthis study came from a university-affiliated hospital (EujeongbuSt. Mary Hospital, Korea). All of the patients were Koreans andweonly collected sporadic cases without any positive family historyof colorectal cancer. The hospital pathology department con-firmed pathologic features of the colorectal cancer. After thesurgery, two to six different tumor areas and one normal mucosalarea were picked from each fresh colectomy specimen (Supple-

mentary Fig. S1). Also, matched metastatic nodules in liver werepicked from the same patients along with primary colorectalcancer. Three patients (CRC1, CRC3, CRC4) had liver metastasesat initial surgeries. Two patients (CRC2, CRC5) had no livermetastasis at initial surgery, but later they were found to haveliver metastasis. All of the picked tissues from tumor and normalareas were frozen and stained with hematoxylin and eosin. Twopathologists selected cases with rich tumor cell populations (atleast 70%), which were subsequently used in the study. Forgenomic DNA extraction, we used the DNeasy Blood & TissueKit (Qiagen) according to the manufacturer's recommendation.To evaluate the status of microsatellite instability, we used fivemononucleotide repeats (BAT25, BAT26, NR-21, NR-24, andMONO-27; ref. 18). For all of the five colorectal cancer casesexamined, no markers showed instability; thus, all cases wereclassified as microsatellite-stable (MSS). The other clinical infor-mation of five colorectal cancer cases is available in Supplemen-tary Table S1.

Whole-exome sequencing and mutation analysesUsing genomic DNA from tumor and matched normal

samples, we performed exome-capture sequencing. Using Agi-lent SureSelect Human All Exome 50 Mb Kit (Agilent Technol-ogies), whole exome-sequencing was performed with IlluminaHiSeq2000 platform to generate 101 bp paired-end sequencingreads according to the manufacturer's instructions. Burrows-Wheeler aligner (19) was used to align the sequencing readsonto the human reference genome (hg19). We used GenomeAnalysis ToolKit for the local realignment and score recalibra-tion of the sequencing reads (20). We employed Picard (http://picard.sourceforge.net) and Samtools (21) for the basic proces-sing and the management of the sequencing data. MuTect (22)and SomaticIndelDetector (20) were used to call somatic pointmutations and indels by comparing the sequencing reads of thetumor and matched normal genomes. ANNOVAR package wasused to predict their functional consequences such as silent ornonsilent variants for somatic variants located in codingsequences (23). The evolutionary analyses based on mutationcalls can be substantially flawed when the mutations areindependently called from each of the intratumoral biopsiesand simply merged. This is mainly due to false negative calls oflow-frequent mutations or those in sparsely covered regionsleading to the overestimation of the ITH. To minimize suchbiases, we first collected index mutations with mutant allelefrequencies >5% and those called in the regions with sequenc-ing coverage (>20�). For each of the index mutations, weexamined the rest of the intratumoral biopsies for the presenceof any sequencing reads that support the corresponding muta-tion in other regions of the given case. The sequencing infor-mation, including the depth of sequencing for 35 tumor and 5normal genomes, is available in Supplementary Table S1.Sequences have been deposited to the SRA database and canbe accessed on Project ID of PRJNA271316.

Spatial mutation categories with respect to the regionaldistribution

We established schematics of mutation classification system asshown in Fig. 1A. The given example represents six regionalbiopsies from primary and metastatic lesions (P1–P3 and M1–M3, respectively) and the regional distribution of mutations thatcan be used to infer the phylogenetic relationship of regional

Translational Relevance

Metastasis, especially to liver, is a major cause of cancer-related death of colorectal cancer. Knowledge on genomicevolution from primary to metastatic colorectal cancer wouldbe crucial not only to understanding colorectal cancer path-ogenesis but also to biomarker development for diagnosis andprognosis. In this study, we identified that colorectal cancershowed branching evolutionary patterns rather than linearpatterns during the progression. Also the genetic ITH wasprevalent in colorectal cancer and might serve as a potentialdriving force to generate metastasis-initiating clones. The datasuggest that a single regional biopsy may underestimate colo-rectal cancermutations in clinic.Weobserved a relative paucityof recurrent metastasis-clonal event, suggesting that colorectalcancer distant metastases may not follow a uniform course ofgenomic evolution, which should be considered in the geneticdiagnosis and the selection of therapeutic targets for advancedcolorectal cancer.

Kim et al.

Clin Cancer Res; 21(19) October 1, 2015 Clinical Cancer Research4462




biopsies.We classified themutations into five spatial categories of"universal," "metastasis-clonal," "primary-private," "metastasis-private" mutations along with "unclassified." Universal muta-tions are those observed in all regional biopsies of a given tumor(red in Fig. 1A). The regionally clonal presentation (i.e., theregional commonality) of universal mutations suggests that theseevents have arisen earlier during tumor evolution and may

contain potential cancer initiators. Metastasis-clonal mutationsrepresent those that are regionally clonal only in the metastaticregions thereby excluding universal mutations. We used thismutation category in a broad sense to include the events witha regional clonality in metastatic lesions but partially present orabsent in primary lesions (orange and yellow in Fig. 1A, respec-tively). The latter might have occurred after the physical

Figure 1.Schematics of resolving somatic variants with respect to regional distribution and intratumoral recurrences. A, six regional biopsies from primary (P1–P3) andmetastatic tumors (M1–M3) and their inferred evolutionary relationships are shown in this exemplary case. The asterisk (P2�) indicates the regional biopsy of theprimary tumor as the potential ancestor of themetastatic clones. Mutations observed in all of the regional ormetastatic biopsies are classified into "universal" (red) or"metastasis-clonal" (orange and yellow), respectively. Metastasis-clonal mutations include those with partial presence (orange) or absence in the primary lesions(yellow). Mutations observed in specific primary or metastatic biopsies are primary-private or metastasis-private (blue and green, respectively). Mutationsthat are not classified according to this scheme are unclassified (gray). B, the regional distribution of somatic variants for the five colorectal cancer genomes is shown.For each of the five colorectal cancer genomes (CRC1–CRC5), the somatic variants for at least one regional biopsy in a given case are shown as rows on theheatmap with the number of somatic variants shown at the top. The mutations of primary and metastatic lesions are shown as black and purple on the heatmapand barplots above. At the right of each heatmapl, the classification of the somatic variants according to five spatial categories is shown with a color schemebelow. Nonsilent mutations in cancer census genes are shown for missense (blue) and nonsense point mutations (red) as well as frameshifting (orange) andnonframeshifting indels (yellow).

ITH of Colorectal Cancers with Metastases

www.aacrjournals.org Clin Cancer Res; 21(19) October 1, 2015 4463




separation ofmetastasis-originating clones; however, it is possiblethat not all mutations in primary masses can be identified in ourexperimental settings due to regional biases or low allele frequen-cies. Thus, we collectively classified these events as metastasis-clonal mutations. These events correspond to "metastasis pro-gressors" in previous reports (17) and may include the geneticevents responsible for the initiation and development of distantmetastases.Hypothetically, the primary subclones fromwhich themetastasis-originating clones arise can be identified (annotated asP2� with asterisk in Fig. 1A). The other mutational categoriesinclude "metastasis-private" events that are present but not in allof themetastatic lesions (green in Fig. 1A). These eventsmay haveoccurred after the distant metastasis is established and acquiredduring the expansion of metastatic clones. In addition, "prima-ry—private" events that are present in primary but absent inmetastatic lesions (blue in Fig. 1A), representing the changesindependently acquired during the expansion of the primarytumor mass. Along with the four spatial mutational categories,we classified the mutations that cannot be classified into themutational categories discussed above as an additional categoryof "unclassified" (gray in Fig. 1A). We performed DAVID analyses(http://david.abcc.ncifcrf.gov; ref. 24) to identify the molecularfunctions enriched in the spatial category-specific mutations (i.e.,genes harboring nonsilent mutations only in one spatial categorybut not in other categories). We selected the top three significantfunctional clusters and their functional annotations from theDAVID output for each spatial category.

Copy number profilingWe used Agilent SurePrint G3 Human CGH Microarray 180 K

for copy number profiling as previously described (12). The log2ratio (tumor biopsy/matched normal) was segmented usingGLAD algorithm (25). We further employed ABSOLUTE algo-rithm to estimate the purity and ploidy for each copy numberprofile (26). The segmented log2 ratios and probe-level intensitieswere adjusted using the estimated purity and ploidy as describedpreviously (27). The microarray data have been submitted to theNCBI Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE58512. To compare the array-based copy number calls with those from exome sequencing, weused VarScan2 (28) to obtain the read depth differences betweenthe tumor andnormal sequencingdata. The concordance betweentwo platforms was estimated by calculating the Pearson correla-tion coefficient (r) between the GC-corrected, log2-transformedreaddepthdifferences and log2 ratio frommicroarray data. A goodconcordance level was observed (r¼ 0.75–0.98; median of 0.90),suggesting that DNA copy numbers estimated from two indepen-dent platforms were largely concordant.

ChromothripsisThe inference of chromothripsis was done as previously

described (29). For each of the chromosomes with at least 10breakpoints, we first evaluated whether the sizes of individualsegments are not substantially different from their neighboringsegments given that the random occurrence of breakpoints inchromothripsis would generatemultiple segmentswith their sizesroughly the same or at least at the same order of magnitude withtheir neighbors. If we suppose that a chromosome has n segments(sizes of s1, s2, . . ., sn), we can use Ri ¼ |log2(si/si�1)| as a measurefor the size difference between the neighboring segments. Then,weused the score S¼R/Re (R as themedianof theRi's and theRe as

the expected value of R under the hypothesis that breakpoints arerandomly distributed in the chromosome) to filter out the chro-mosomes that are unlikely to be involved with chromothripsis(S < 2.0). To further evaluate the oscillating pattern of copynumber states, we counted the number of peak-and-valley (i.e.,the segments showing copy number difference of log2 > 0.5 withtheir neighbors). The permutation was performed 10,000 timesand P value was assigned as the proportion of the permutationswhose peak and valley counts are greater than the observed countsof peak-and-valley. P < 0.05 was used to select the chromosomeswith the oscillating or alternating copy number changes.

Phylogeny estimation from mutations and copy numbersThe phylogenetic relationship between the regional biopsies of

a given sample was inferred by mutations and copy numberprofiles. For each colorectal cancer, maximum parsinomy treeswere inferred as previously described (11, 12) using branch-and-bound algorithm implemented in PHYLIP software package. Thephylogenetic relationship of intratumoral biopsies was alsoinferred from the copy number changes using the TuMultpackages (30). Among the GLAD-based segments, we selectedthose with log2 ratios of >0.2 or <�0.2 and used their boundariesas chromosomal breakpoints for the input of TuMult algorithm.As reference of Tumult, we downloaded the segmentation data of470 colorectal cancer genomes as available in TheCancerGenomeAtlas consortium (31). The copy number calls were also used totest for a star topology (i.e., whether a phylogeny agrees with alinear or a branched evolution) using MEDICCquant R package(32). Thenull hypothesis assuming a linear evolutionwas rejected(P < 0.05), so that the alternative hypothesis of a branchedevolution was selected for four of the five cases examined(CRC1–CRC4, except for CRC5).

ResultsMultiregion sequencing and mutation analyses of colorectalcancer genomes

In this study, we adopted a scheme of multiregion exomesequencing combined with spatial mutational profiling (12, 17)to reveal the mutational ITH of primary and metastatic colorectalcancers. Somatic mutational events were classified into five spatialmutation categories of "universal," "metastasis-clonal," "metasta-sis-private," "primary-private," and "unclassified" (Fig. 1A) withrespect to the regional distribution (i.e., the regional biases and thelevel of recurrences for given mutations). The details on the cate-gorization system and the interpretation of different spatial muta-tion categories are available in Materials and Methods.

We analyzed 35 multiregion biopsies of the primary andmetastatic lesions in five colorectal cancers (n ¼ 2–5 and n ¼2–6 for primary and metastatic biopsies per case, respectively;Supplementary Table S1). The geographical mapping of thebiopsy spots in primary colorectal cancers and liver metastaticlesions is shown in Supplementary Fig. S1. Overall, whole-exomesequencing of the 35 primary and metastatic colorectal cancergenomes identified 76 to 244 coding mutations and indels perbiopsy (median of 104 somatic variants). Mutation numbersbetween the primary andmetastatic biopsies of each patient werenot significantly different except one case with the largest numberof biopsies (CRC3;P¼4.8�10�5; t test). Themutations observedfrom the five colorectal cancer cases are listed with the detailedinformation of spatial categories and functional consequences in

Kim et al.





Supplementary Tables S2. The regional mutation profiles for thefive colorectal cancer cases with their spatial categories showed asubstantial level of ITH (Fig. 1B). We observed that 19.8% to53.9% of mutations in a given sample were universal, whereas46.1% to 80.2% of mutations were subclonal. Among the sub-clonal mutations, metastasis-clonal, metastasis-private and pri-mary-private categories comprised 0.7% to 15.6%, 2.4% to 40.7%and 13.8% to 56.0%ofmutations, respectively. In addition, 1.4%to 37.2% of mutations per biopsy were not observed in any otherregions of the given sample (biopsy-specific mutations, see Sup-plementary Fig. S2).

We next investigated recurrent nonsilent mutations (i.e., thoseobserved in �3 of the five colorectal cancer cases). Figure 2Ashows the nine genes with suchmutations, including well-knownthe KRAS mutations. All of the APC, KRAS, and TP53 mutationswere identified as universal events except for a metastasis-clonalAPC mutation in one case. Eight APC mutations were all trun-cating (either nonsense substitutions or frameshift indels) andwere often biallelic (e.g., CRC2 harbored one nonsense mutationand another frameshift mutations). Two KRAS missense muta-tions occurred at amino acid residue of position 12 (p.G12S andp.G12D). Two of five TP53 mutations were truncating and theother three TP53missensemutations occurred at (p.R213X and p.R273C) or near the knownmutation hotspots (p.R267P; ref. 33).See also Supplementary Tables S2, for the details on the functionalconsequences and the amino acid residues affected for othermutations. The other recurrent mutations besides APC, TP53,and KRAS were dispersed across different spatial categories. Wealso performed gene set enrichment analysis to infer the unifyingmolecular terms of the nonsilent mutations that exclusively

belong to each of the spatial categories (i.e., those observed inonly one of five spatial categories) or the category-specific muta-tions. The functional annotations of "cell adhesion" are com-monly enriched in universal, metastasis-clonal and primary-pri-vate categories (Fig. 2B).

Distinct mutational spectrum of spatial mutation categoriesWe evaluated the properties of mutations according to the

spatial categories (Fig. 3). Figure 3A shows the number of muta-tions across five spatial categories for five colorectal cancer cases.We observed that frameshift and inframe indels were enriched tothemetastasis-private category in two cases (CRC2 andCRC4; Fig.3B). Although C:G>T:A transitions at CpG dinucleotides weresignificantly enriched to universal category in three colorectalcancers (Fig. 3C), the enrichment of C:G>A:T and T:A>A:T trans-versions were only observed in lesion-specific (metastasis-privateand primary-private) categories in more than two cases. Analysesof mutant allele frequencies revealed that mutations frommetas-tasis biopsies had significantly higher allele frequencies comparedwith those from primary biopsies (P ¼ 2.2 � 10�16; t test;Supplementary Fig. S3). This finding suggests that metastaticlesions were relatively clonal due to the evolutionary bottleneckand restriction of diversity. Regarding spatialmutation categories,universal mutations had the highest allele frequencies, whereasprimary- and metastasis-private mutations had the lowest (Sup-plementary Fig. S3).

Phylogeny and copy number profiles of regional biopsiesPhylogenetic trees inferred from genomic data may represent

the evolutionary life history of individual tumors as discussed in

Figure 2.Recurrent mutations and pathway analyses. A,mutations with no less than three occurrencesacross the cases are shown with four cancercensus genes (APC, TP53, KRAS, and MAML2) atthe top (gray shaded). Subcolumns representingCRC1 to CRC5 are in each of the five columnscorresponding to the spatial mutationcategories. Nonsilent mutations are marked by acolor scheme (right). Redundant mutations arealso shown. For example, CRC3 harbored twouniversal mutations on APC, one nonsensemutation (red) and one frameshift indel(orange). B, the results of the DAVID pathwayanalyses are shown for the five spatial mutationcategories. The top three functional clusters areshown for each spatial category with the x-axisrepresenting the log-transformed P values.






previous literatures (11, 12). In this study, the phylogeneticrelationship between the regional biopsies of a given samplewas inferred from mutation data using maximum parsimonyand chromosomal breakpoints in DNA copy number profilesas described previously (11, 12). Figure 4 shows phylogenicprofiles (mutation- and copy number-based) of the five -s alongwith the genome-wide heatmap of CNAs. The inferred evolu-tionary patterns appeared as a branched evolution rather than aconstant or linear evolution. For example, statistical analysis ofthe phylogeny patterns (32) indicated that the phylogeny offour (CRC1–CRC4) of the five -s was branched; that is, the nullhypothesis of the test (a linear or constant evolution) wasrejected for these four colorectal cancers except for CRC5.

In either mutation- or copy number–based phylogenies,most regional biopsies frommetastases were cosegregated apartfrom those of primary lesions indicating that they were moregenetically similar to each other than the primary lesions(Fig. 4). Most CNAs were shared by both primary and meta-static lesions, representing early or universal genomic events.Interestingly, in two cases (CRC2 and CRC3), mutation-basedphylogenies segregated some primary subclones to those ofmetastatic tumors rather than those of other primary tumors (e.g., CRC2-P1 and CRC3-P3/P4; those with asterisks in Fig. 4).This phenomenon was consistently observed in copy number–based phylogenies (Fig. 4). These observations suggest that theorigins of distant metastases could also be traceable in thesubclonal architecture of primary colorectal cancers as previ-ously reported in pancreatic cancers (17).

Copy number alterations containing well-known cancer genesWell-known cancer genes that were reported to be frequently

altered in colorectal cancer genomes such as APC, PTEN, IGF2,and SMAD4 (31) were found to be universally altered across thewhole biopsy sites in this study (Fig. 5A). We also observed the

universal focal deletion in between VTI1A and TCF7L2 inCRC2, which was reported to be responsible for the recurrentfusion between these genes in colorectal cancers (34). Most of theuniversal focal changes involving potential driver genes showedexactly coinciding boundaries across the regional biopsies indic-ative of their clonal origins (Fig. 5A). By contrast, genes located infragile regions such as MACROD2, FHIT, WWOX (all on CRC2),and PARK2 (on CRC3) showed deletions whose boundarieswere irregular across the regional biopsies of given tumors,although they were universal (Fig. 5B) indicative of ongoinggenomic instability in these fragile loci. It is worth noting thatthe copy number gains involving c-MYC oncogene were observedin all five colorectal cancers analyzed (one focal and four 8q arm-level change cases; Fig. 5C). Universal or metastasis-clonal copynumber gains of 8q encompassing c-MYCwere observed inCRC1,CRC2, CRC4, and CRC5, whereas a focal gain of c-MYC as ametastasis-clonal event was observed in CRC3. In CRC2, inaddition to arm-level metastasis-clonal 8q gain, a focal gainencompassing c-MYC was also observed in a primary lesion(CRC2-P4), suggesting that the evolutionary path to c-MYC copynumber gains may be diverse even within a single cancer mass.

Evidences of region-specific chromothripsisOf the five colorectal cancers examined, two candidate

chromothripsis events were observed as one universal (onchromosome 2 of CRC3) and one metastasis-private event (onchromosome 16 of CRC5). In the chromothripsis of CRC3,breakpoints in deleted segments exactly overlapped across allthe regional biopsies, suggesting that this event occurred earlyin tumor development (Fig. 5D). However, the interveningsegments in between the copy number losses often showedneutral states or copy number gains in the primary tumorswhile they were consistently copy number gained in the met-astatic lesions. This result suggests that additional copy number

Figure 3.The mutational features of five spatial mutationcategories. A, the number of mutations in fivespatial mutation categories are shown for thefive colorectal cancer genomes according to thecolor scheme shown in the inset. B, for each ofthe spatial mutation categories, the relativefractionof six functional categories (SC) is shownin percentage (%; y-axis). Asterisks indicatesignificant enrichment asmeasured by the Fisherexact test (P < 0.05). For example, nonsensemutations were relatively enriched in themetastasis-clonal mutations of CRC2 given thetotal number of mutations and nonsensemutations for the other spatial categories in thesample. C, similarly, significant enrichment is alsoindicated by asterisks for the sequencecomposition of mutations. C:G>T:A changes(C>T) are further distinguished into thosespecific to CpG dinucleotides or not.

Kim et al.





changes were acquired in the rearranged chromosomal seg-ments during metastatic progression after the initial universalchromothripsis event. In CRC5, we observed a metastasis-

private chromothripsis event that must have occurred after thedivergence of metastatic clones from the primary cancer mass(Fig. 5E).

Figure 4.Phylogenetic trees of the regional biopsies. For each case (CRC1 to CRC5 from top), two phylogenetic trees are shown (one on the left and the other in the middle)along with genome-wide plots of DNA copy number changes (right). Phylogenetic trees constructed using maximum parsimony from somatic mutations are shownon the left with the length of branches in the tree proportional to the number of mutations in the corresponding branches. The phylogenetic trees inferredfrom chromosomal breakpoints in DNA copy number profiles using TuMult are shown in themiddle. Asterisks indicate the primary biopsies that aremore geneticallysimilar to metastatic biopsies than those of remaining primary biopsies. The nodes representing normal ("Norm") and common precursors ("Pre") are shownwith the biopsies from primary (P; gray) andmetastatic lesions (M; red). The order of the regional biopsies in the heatmap (right) is according to the tree constructedfrom chromosomal breakpoints as shown with dotted lines. In the heatmap, chromosomal gains and losses are shown as red and blue, respectively.






Potential genetic determinants of colorectal cancer metastasesTheoretically, metastasis-clonal events are potential candidates

for metastasis determinants. We first investigated apparent loss-of-function events among the metastasis-clonal mutations,including frameshift indels (GDPD1, NPS, and SLK) and non-sense mutations (APC, ASCC3, CDH12, ENTPD4, GYLTL1B,

NEDD4L, PCDHB5, and TK2). Through capillary sequencing ofseven lesions (four primary and three metastatic) of the corre-sponding case (CRC1), the ITH of the NEDD4L nonsensemutation was confirmed (data not shown). In the followingimmunohistochemical analysis for the CRC1, primary lesionswithout the NEDD4L mutation showed positive NEDD4L

Figure 5.Copy number alterations in cancer genes and chromothripsis. A, focal and universal copy number alterations are shown for APC, PTEN, VTI1A-TCF7L2, IGF2, andSMAD4. Heatmaps of the selected loci are shown for probe-level log2 intensities. The case number andgenomic coordinates are shown for each heatmapwith arrowsindicating the locus of genes. B, the probe-level intensities are shown for FHIT, WWOX, MACROD2, and PARK2. C, the copy number changes are shown forchromosome 8 for all the regional biopsies (left). Focal chromosomal gains involving MYC are marked and separately shown on the right. CRC2 and CRC3 casesharbored primary-private and metastasis-clonal focal chromosomal gains involving the MYC locus (arrow). D, whole chromosome-scale chromothripsis onchromosome 2 is shown (CRC3, left). Focal scale chromothripsis on chromosome 17 (CRC1; E) and biopsy-specific chromothripsis on chromosome 16 (CRC5) areshown (F).

Kim et al.





immunostaining, whereas metastatic liver lesions with theNEDD4L mutation showed decreased NEDD4L immunostain-ing (Fig. 6A and B). Furthermore, in an independent set of 17colorectal cancers with liver metastasis, seven liver metastases(41.2%) even without NEDD4L mutation showed decreasedNEDD4L expression compared with the corresponding primarycolorectal cancers (Fig. 6C and D), suggesting that liver metas-tasis-specific NEDD4L downregulation might be caused bynonmutational as well as mutational events. Among the focalCNAs, we selected metastasis-clonal deletions of PCDH9 as acandidate genetic event for metastasis. The ITH of PCDH9 wasconfirmed by real-time qPCR in the corresponding case(CRC2). In this case, primary colorectal cancer sites withoutthe PCDH9 deletion showed positive immunostaining, where-as two liver metastasis lesions showed much weaker immu-nostaining in the cancer cells, indicating that the metastasis-clonal pattern of the PCDH9 deletion was recapitulated at theprotein level (Supplementary Fig. S4).

DiscussionIn this study, multiregion biopsies from primary colorectal

cancer and liver metastases revealed a substantial level of ITH inboth primary and matched liver metastases. Our results, 46% to80% subclonal mutation fractions (i.e., those not detectableacross all the regional biopsies), are similar to the previousestimates of other cancer types (10–12), further supporting thenotion that ITH is a general phenomenon in cancers. AlthoughITH at the genomic level has been observed across many cancertypes, including a report of colon adenomas (35), to our knowl-edge, our study is the first report on the ITH of colorectal cancergenomes with a detailed view on spatial mutation categories andphylogenetic structures of primary and metastatic lesionsobtained from the same individuals.

There havebeen concerns onpreviousmetastasis-specificmuta-tion identification that might generate bias. For example, studiesbased on "index" mutations solely identified from metastatictumors (17, 36) might not only ignore a substantial level of ITH,but also can be biased by the selection ofmetastatic biopsies used

to generate the indexmutations. In present study, we demonstratethat multiregion sequencing across multiple primary and meta-static lesions provides more information (e.g., the extent of ITH,the categorization of the mutations according to the differentevolutionary stages and the evolutionary relationship between theprimary and metastatic lesions) compared with a simple primaryversus metastatic lesions (37, 38).

The recurrent mutations on known cancer genes (e.g., APC,KRAS, and TP53; ref. 39) were almost exclusively observed in theuniversal category consistent with recent reports (37, 38). In caseof APC mutations, three of five cases showed "second-hit" muta-tions on APC, probably representing biallelic mutations or theoccasional third hit on APC locus (40). To get a clearer under-standing, biallelic APC mutations in one case (one universalframeshift indel and one metastasis-clonal nonsense mutationof CRC2) were validated using Sanger sequencing (Supplemen-tary Fig. S5). The nonsense APC mutation in this case wasclassified as metastasis-clonal because of the absence of themutations in a primary region (CRC2-P1) that also exclusivelyharbors an entire loss of chromosome 5. Thus, it is expected thattwo APCmutations in this case were biallelic, but the subsequentchromosomal loss led to the loss of only one of the biallelicmutations (nonsensemutation) while retaining the other (frame-shift indel). It is also notable that this lesion-specific chromo-somal loss altered the spatial category of the mutation (universalto metastasis-clonal), which may be a potential source of unclas-sified mutations.

The mutations that cannot be explained according to theevolutionary model (Fig. 1A) were left in "unclassified" category.We propose several assumptions for these evolutionarily unex-plainable mutations (e.g., those subclonal both in primary andmetastatic biopsies). First, the false-negative calls of mutationsdue to a low sequencing depth or other technical issues may leadto a misclassification of mutations and can be the source of"unclassified" mutations. Second, unclassified mutations mayhave arisen with subsequent chromosomal changes such asloss-of-heterozygosity as shown in the example of biallelic APCmutations in CRC2. Third, these mutations may have beenacquired independently in primary and metastatic lesions

Figure 6.Visualization of NEDD4L expression in coloncancer tissues and corresponding livermetastasis by immunohistochemistry. A coloncancer case with NEDD4L nonsense mutationshows positive NEDD4L immunostaining inprimary colon tissues (A), but very weakNEDD4L immunostaining inmetastasized cancercells from the liver of the same patient (B; frozentissues). Another colon cancer case in avalidation set shows positive NEDD4Limmunostaining in primary colon tissues (T in C),whereasmetastasized colon cancer in the liver ofthe same patient shows very weak NEDD4Limmunostaining (T in D). Adjacent normal colonmucosal (N in C) and liver (N in D) cells showpositive NEDD4L immunostaining.






representing a convergent evolution (homoplasy). Althoughhomoplasy events have been previously reported in terms ofgene-level convergent events (e.g., SETD2 frameshift indels andsplicingmutations independently acquired by distinct biopsies ina cancer mass; ref. 10), we believe that convergent mutations withidentical genomic coordinates would be rare. But the mutationswith knownhotspotswill beworthy of further investigation as thepotential source ofmisclassification. Finally, a recent report aboutthe colorectal cancer evolution proposed that the subclonalmutational architecture as configured early in the evolution ofcancers, may persist during the growth of cancer mass (41). Thepresence of "pervasive" mutations (i.e., subclonal mutationspresent across the cancer mass), especially those with very lowmutation frequencies, may be the source of error in the accurateinferring of the phylogeny.

The first issue may be overcome by using "high-confident"somatic mutations (i.e., those called in genomic regions withsufficient sequencing coverage across all the regional biopsies). Totest this, we selected somatic mutations called with sequencingdepth >50� in all the regional biopsies and inferred the phylog-eny (Supplementary Fig. S6). Although the phylogenies inferredfrom the mutations with or without this stringent filtering arelargely concordant with each other, it should be further investi-gated whether the use of high-confident mutations may performbetter in inferring the evolutionary relationship of intratumoralbiopsies. In addition, low-confidentmutations in particular biop-sies can be selectively ignored to preserve the number of muta-tions as evolutionary markers.

We observed the preponderance of indels in late (metastasis-private and primary-private) genomic events, suggesting thatcolorectal cancer genomes seem more vulnerable to differenttypes of somatic variants during their evolution. It is alsopossible that these unique genomic footprints might reflectdifferent fixation rates of the mutation types, for example,highly deleterious events such as indels may be negativelyselected in the early evolutionary phase. We also disclosed themutation spectra associated with colorectal cancer develop-ment and progression stages (e.g., C:G>T:A transitions enrichedin universal mutations), indicative of different mutationalprocesses dominant in the early and late evolutionary phasesduring colorectal cancer progression.

The overall higher allele frequencies of mutations in the met-astatic tumors compared with those in primary tumors wasindicative of thepresence of an evolutionary bottleneck.However,the allele frequencies of the five spatial mutation categoriesrevealed that primary- or metastasis-private mutations had thelowest allele frequencies, indicating that these lesion-specificmutations could be largely late events. Cares should be taken forthis interpretation since this comparison as well as other analysesbased on allele frequency can be biased by the purity and ploidylevel of the selected tumor lesions. For example, low tumor puritywith a high level of normal tissue contamination will lead to theunderestimation of tumor allele frequencies of mutations. More-over, we selected tumor-rich regions (>70%) for regional biopsyin this study, but additional source of low tumor purity (e.g., theselection of tumor invasive fronts) would still exist.

The phylogenetic trees inferred from somatic mutations andDNA copy number changeswere largely concordant and providedthe evolutionary history. First, the metastatic clones are cosegre-gated apart from those of primary lesions indicative of theirgenetic similarity. Second, mutation profiles in some primary

lesions were genetically more similar to those of metastasis thanthe other primary lesions suggesting that theymight be a potentialmetastatic source. In spite of the overall similarities, some dis-crepancies were observed between the trees from somatic muta-tions and DNA copy number changes. Potential causes for thediscrepancies may include those of "unclassified" mutations aspreviously addressed as well as technical issues in calling theCNAs. Among the unclassifiedmutations, those in CRC1 were allobserved on chromosome 2 (i.e., four unclassifiedmutations thatare present in all of the regional biopsies except for onemetastaticbiopsy of CRC1-M3). It is speculated that those mutations areuniversal events ab initio but subsequent chromosomal eventssuch as loss-of-heterozygosity in CRC1-M3 led to the biopsy-specific loss of mutational calls. Moreover, the chromosome 2 ofCRC1-M3 does not show apparent copy number changes suggest-ing that this chromosomal event is likely to be copy numberneutral loss-of-heterozygosity. Supporting our speculation, allthree universal mutation events in this chromosome showedsignificantly highermutational allele frequencies (P <0.05; pairedt test) for CRC1-M3 compared with those of other regionalbiopsies.

Although a prevailing view has supported early occurrences ofchromothripsis, especially those that are cancer drivers (42), wefound a colorectal cancer case that harbored a metastasis-privatechromothripsis event, supporting another report that chromo-thripsis does not necessarily represent an initiating event (43).Our results suggest that chromothripsis may occur at any devel-opmental stage in colorectal cancer genomes, and the incidence ofchromothripsis estimated from primary colorectal cancer casesmight be underestimated. Additional complexity of chromothrip-sis in terms of the CNAs of affected chromosomal segmentsfurther suggests that this chromosomal event may be accompa-nied by subsequent chromosomal changes.

Systematic sequencing analyses of primary and metastaticcancer genomes have shown that a small number of mutationsare required for an invasive colorectal cancer to metastasize (36).Moreover, even newly acquired mutations during metastaticprogression arenot overlapped among individual patient samples(17). Consistent with these findings, our study identified metas-tasis-clonal mutations only in 0.7% to 15.6% of total mutationsin the given cases. These data suggest that essential mutations forcancer progression may be obtained in primary cancer genomesbefore metastasis. In line with this hypothesis, colon adenomagenomeswere reported tobenearly as old as invasive cancers (44),suggesting the time required formalignant transformationmayberelatively short in the entire life history of colorectal cancer. Thedriving force for colorectal cancer metastases may be beyond thescope of our study, for example, mutations occurring in noncod-ing sequences (45) or epigenetic alterations (46).

Finally, we validated some of metastasis-clonal mutations,including a NEDD4 nonsense mutation and PCDH9 deletion(both as singletons) by resequencing or qPCR inour test sets alongwith immunostaining in an independent cohort. Metastasis-clon-al nonsense mutation of NEDD4L is interesting, because theencoded protein NEDD4L interacts with Smad proteins (47) andcan regulate TGFb signaling that has been implicated in cancerinvasion and metastasis (48). However, the potential role ofNEDD4 as metastatic determinants along with that of PCDH9requires further investigation. Although we did not find noveldriver genes of colorectal cancer metastasis in the metastasis-clonal category, integration of these intermittent findings will be

Kim et al.





needed to elicit biologically relevant insights to improve diagno-sis and treatment of colorectal cancer patients with metastasis.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: S.-H. Lee, Y.-J. ChungDevelopment of methodology: Y.-J. ChungAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): S.-H. Jung, S.H. Lee, M.S. Kim, S.-W. Park, Y.-J. ChungAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): T.-M. Kim, S.-H. Jung, S.H. Lee, I.-P. Baek, J.-K. Rhee,Y.-J. Chung

Writing, review, and/or revision of the manuscript: T.-M. Kim, S.-H. Jung,J.-K. Rhee, S.-H. Lee, Y.-J. ChungAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): S.-W. ParkStudy supervision: C.H. An, S.-H. Lee, Y.-J. Chung

Grant SupportThis study was supported by a grant from the National Research Foundation

of Korea (2012 R1A5A2047939).The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisementin accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received September 23, 2014; revised March 24, 2015; accepted April 22,2015; published OnlineFirst May 15, 2015.

References1. DeSantis CE, Lin CC, Mariotto AB, Siegel RL, Stein KD, Kramer JL, et al.

Cancer treatment and survivorship statistics. CA Cancer J Clin2014;64:252–71.

2. Winawer SJ, Zauber AG, Fletcher RH, Stillman JS, O'brien MJ, Levin B,et al. Guidelines for colonoscopy surveillance after polypectomy: aconsensus update by the US Multi-Society Task Force on ColorectalCancer and the American Cancer Society. CA Cancer J Clin 2006;56:143–59.

3. Nowell PC. The clonal evolution of tumor cell populations. Science1976;194:23–8.

4. Valastyan S, Weinberg RA. Tumor metastasis: molecular insights andevolving paradigms. Cell 2011;147:275–92.

5. Yates LR, Campbell PJ. Evolution of the cancer genome. Nat Rev Genet2012;13:795–806.

6. Shah SP, Morin RD, Khattra J, Prentice L, Pugh T, Burleigh A, et al.Mutational evolution in a lobular breast tumour profiled at single nucle-otide resolution. Nature 2009;461:809–13.

7. Ding L, Ellis MJ, Li S, Larson DE, Chen K, Wallis JW, et al. Genomeremodelling in a basal-like breast cancer metastasis and xenograft. Nature2010;464:999–1005.

8. Bashashati A, Ha G, Tone A, Ding J, Prentice LM, Roth A, et al.Distinct evolutionary trajectories of primary high-grade serous ovariancancers revealed through spatial mutational profiling. J Pathol 2013;231:21–34.

9. de Bruin EC, McGranahan N, Mitter R, Salm M, Wedge DC, Yates L, et al.Spatial and temporal diversity in genomic instability processes defines lungcancer evolution. Science 2014;346:251–6.

10. Gerlinger M, Rowan AJ, Horswell S, Larkin J, Endesfelder D, Gronroos E,et al. Intratumor heterogeneity and branched evolution revealed by multi-region sequencing. N Engl J Med 2012;366:883–92.

11. Gerlinger M, Horswell S, Larkin J, Rowan AJ, Salm MP, Varela I,et al. Genomic architecture and evolution of clear cell renal cellcarcinomas defined by multiregion sequencing. Nat Genet 2014;46:225–33.

12. Kim TM, Jung SH, Baek IP, Lee SH, Choi YJ, Lee JY, et al. Regional biases inmutation screening due to intratumoural heterogeneity of prostate cancer.J Pathol 2014;233:425–35.

13. Lindberg J, Klevebring D, Liu W, Neiman M, Xu J, Wiklund P, et al. Exomesequencing of prostate cancer supports the hypothesis of independenttumour origins. Eur Urol 2013;63:347–53.

14. Navin N, Kendall J, Troge J, Andrews P, Rodgers L, McIndoo J, et al.Tumour evolution inferred by single-cell sequencing. Nature2011;472:90–4.

15. Zhang J, Fujimoto J, Zhang J, Wedge DC, Song X, Zhang J, et al. Intratumorheterogeneity in localized lung adenocarcinomas delineated by multi-region sequencing. Science 2014;346:256–9.

16. Greaves M, Maley CC. Clonal evolution in cancer. Nature 2012;481:306–13.

17. Yachida S, Jones S, Bozic I, Antal T, Leary R, Fu B, et al. Distant metastasisoccurs late during the genetic evolution of pancreatic cancer. Nature2010;467:1114–7.

18. Patil DT, Bronner MP, Portier BP, Fraser CR, Plesec TP, Liu X. A five-markerpanel in a multiplex PCR accurately detects microsatellite instability-highcolorectal tumors without control DNA. Diagn Mol Pathol 2012;21:127–33.

19. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 2009;25:1754–60.

20. DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. Aframework for variation discovery and genotyping using next-generationDNA sequencing data. Nat Genet 2011;43:491–8.

21. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. TheSequence Alignment/Map format and SAMtools. Bioinformatics 2009;25:2078–9.

22. Cibulskis K, Lawrence MS, Carter SL, Sivachenko A, Jaffe D, Sougnez C,et al. Sensitive detection of somatic point mutations in impure andheterogeneous cancer samples. Nat Biotechnol 2013;31:213–9.

23. WangK, LiM,HakonarsonH.ANNOVAR: functional annotationof geneticvariants from high-throughput sequencing data. Nucleic Acids Res2010;38:e164.

24. Huang dW, Sherman BT, Lempicki RA. Systematic and integrative analysisof large gene lists using DAVID bioinformatics resources. Nat Protoc2009;4:44–57.

25. Hupe P, StranskyN, Thiery JP, Radvanyi F, Barillot E. Analysis of array CGHdata: from signal ratio to gain and loss of DNA regions. Bioinformatics2004;20:3413–22.

26. Carter SL, Cibulskis K, Helman E, McKenna A, Shen H, Zack T, et al.Absolute quantification of somatic DNA alterations in human cancer. NatBiotechnol 2012;30:413–21.

27. Landau DA, Carter SL, Stojanov P, McKenna A, Stevenson K, Lawrence MS,et al. Evolution and impact of subclonal mutations in chronic lymphocyticleukemia. Cell 2013;152:714–26.

28. Koboldt DC, Zhang Q, Larson DE, Shen D, McLellan MD, Lin L, et al.VarScan 2: somatic mutation and copy number alteration discovery incancer by exome sequencing. Genome Res 2012;22:568–76.

29. Kim TM, Xi R, Luquette LJ, Park RW, Johnson MD, Park PJ. Functionalgenomic analysis of chromosomal aberrations in a compendium of 8000cancer genomes. Genome Res 2013;23:217–27.

30. Letouze E, Allory Y, Bollet MA, Radvanyi F, Guyon F. Analysis of the copynumber profiles of several tumor samples from the same patient reveals thesuccessive steps in tumorigenesis. Genome Biol 2010;11:R76.

31. The Cancer Genome Atlas Consortium. Comprehensive molecularcharacterization of human colon and rectal cancer. Nature 2012;487:330–7.

32. Schwarz RF, Trinh A, Sipos B, Brenton JD, Goldman N, Markowetz F.Phylogenetic quantification of intra-tumour heterogeneity. PLoS ComputBiol 2014;10:e1003535.

33. Olivier M, Hollstein M, Hainaut P. TP53 mutations in human cancers:origins, consequences, and clinical use. Cold Spring Harb Perspect Biol2010;2:a001008.

34. Bass AJ, Lawrence MS, Brace LE, Ramos AH, Drier Y, Cibulskis K, et al.Genomic sequencing of colorectal adenocarcinomas identifies a recurrentVTI1A-TCF7L2 fusion. Nat Genet 2011;43:964–8.






35. Thirlwell C, Will OC, Domingo E, Graham TA, McDonald SA, Oukrif D,et al. Clonality assessment and clonal ordering of individual neoplasticcrypts shows polyclonality of colorectal adenomas. Gastroenterology2010;138:1441–54.

36. Jones S, Chen WD, Parmigiani G, Diehl F, Beerenwinkel N, Antal T, et al.Comparative lesion sequencing provides insights into tumor evolution.Proc Natl Acad Sci U S A 2008;105:4283–8.

37. Lee SY, Haq F, Kim D, Jun C, Jo HJ, Ahn SM, et al. Comparative genomicanalysis of primary and synchronous metastatic colorectal cancers. PLoSOne 2014;9:e90459.

38. Vakiani E, Janakiraman M, Shen R, Sinha R, Zeng Z, Shia J, et al. Com-parative genomic analysis of primary versus metastatic colorectal carcino-mas. J Clin Oncol 2012;30:2956–2962.

39. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell1990;61:759–67.

40. Segditsas S, Rowan AJ, Howarth K, Jones A, Leedham S, Wright NA, et al.APC and the three-hit hypothesis. Oncogene 2009;28:146–55.

41. Sottoriva A, Kang H, Ma Z, Graham TA, Salomon MP, Zhao J, et al. A BigBang model of human colorectal tumor growth. Nat Genet 2015;47:209–16.

42. Forment JV, Kaidi A, Jackson SP. Chromothripsis and cancer: causes andconsequences of chromosome shattering. Nat Rev Cancer 2012;12:663–70.

43. Kloosterman WP, Hoogstraat M, Paling O, Tavakoli-Yaraki M, Renkens I,Vermaat JS, et al. Chromothripsis is a common mechanism drivinggenomic rearrangements in primary and metastatic colorectal cancer.Genome Biol 2011;12:R103.

44. Tsao JL, Yatabe Y, Salovaara R, Jarvinen HJ, Mecklin JP, Aaltonen LA, et al.Genetic reconstruction of individual colorectal tumor histories. Proc NatlAcad Sci U S A 2000;97:1236–41.

45. Huang FW, Hodis E, Xu MJ, Kryukov GV, Chin L, Garraway LA. Highlyrecurrent TERT promoter mutations in human melanoma. Science2013;339:957–9.

46. Lao VV, Grady WM. Epigenetics and colorectal cancer. Nat Rev Gastro-enterol Hepatol 2011;8:686–700.

47. Gao S, Alarcon C, Sapkota G, Rahman S, Chen PY, Goerner N, et al.Ubiquitin ligase Nedd4L targets activated Smad2/3 to limit TGF-b signal-ing. Mol Cell 2009;36:457–68.

48. Oft M, Heider KH, Beug H. TGFb signaling is necessary for carcinoma cellinvasiveness and metastasis. Curr Biol 1998;8:1243–52.


Kim et al.




2015;21:4461-4472. Published OnlineFirst May 15, 2015.Clin Cancer Res Tae-Min Kim, Seung-Hyun Jung, Chang Hyeok An, et al. Colorectal Cancer Based on Intratumoral Genetic HeterogeneitySubclonal Genomic Architectures of Primary and Metastatic

Updated version

10.1158/1078-0432.CCR-14-2413doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2015/05/20/1078-0432.CCR-14-2413.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/21/19/4461.full#ref-list-1

This article cites 48 articles, 10 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/21/19/4461.full#related-urls

This article has been cited by 22 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/21/19/4461To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-14-2413

http://clincancerres.aacrjournals.org/content/suppl/2015/05/20/1078-0432.CCR-14-2413.DC1

http://clincancerres.aacrjournals.org/content/21/19/4461.full#ref-list-1

http://clincancerres.aacrjournals.org/content/21/19/4461.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/21/19/4461


subclonal genomic architectures of primary and metastatic ... · biology of human tumors subclonal...

Documents