Lineage tracing reveals multipotent stem cells maintainhuman adenomas and the pattern of clonal expansionin tumor evolutionAdam Humphriesa,b, Biancastella Cereserc, Laura J. Gayc, Daniel S. J. Millera, Bibek Dasa, Alice Gutteridgea,d,George Eliac, Emma Nyee, Rosemary Jefferya,f, Richard Poulsoma,f, Marco R. Novellig, Manuel Rodriguez-Justog,Stuart A. C. McDonaldc, Nicholas A. Wrighta,c,1, and Trevor A. Grahama,d,h,1

aHistopathology Laboratory and eExperimental Histopathology Laboratory, Cancer Research UK London Research Institute, London, WC2A 3LY, UnitedKingdom; bSt Mary’s Hospital, Imperial College Healthcare National Health Service Trust, London, W2 1NY, United Kingdom; cBarts Cancer Institute, Barts andthe London School of Medicine and Dentistry, Queen Mary University of London, London, ECM1 6BQ, United Kingdom; dCentre of Mathematics and Physics inthe Life Sciences and Experimental Biology, and gDepartment of Histopathology, University College London, London, WC1E 6BT, United Kingdom; fTheNational Centre for Bowel Research and Surgical Innovation, Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary Universityof London, London, E1 2AT, United Kingdom; and hCenter for Evolution and Cancer, University of California, San Francisco, CA 94143

Edited by Michael Wigler, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, and approved May 3, 2013 (received for review November 27, 2012)

The genetic and morphological development of colorectal cancer isa paradigm for tumorigenesis. However, the dynamics of clonalevolution underpinning carcinogenesis remain poorly understood.Here we identify multipotential stem cells within human colorectaladenomas and use methylation patterns of nonexpressed genes tocharacterize clonal evolution. Numerous individual crypts from sixcolonic adenomas and a hyperplastic polyp were microdissectedand characterized for genetic lesions. Clones deficient in cyto-chrome c oxidase (CCO−) were identified by histochemical stainingfollowed by mtDNA sequencing. Topographical maps of clonelocations were constructed using a combination of these data.Multilineage differentiation within clones was demonstrated byimmunofluorescence. Methylation patterns of adenomatous cryptswere determined by clonal bisulphite sequencing; methylation pat-tern diversity was compared with a mathematical model to infer toclonal dynamics. Individual adenomatous crypts were clonal formtDNA mutations and contained both mucin-secreting and neuro-endocrine cells, demonstrating that the crypt contained amultipotentstem cell. The intracrypt methylation pattern was consistent with thecrypts containing multiple competing stem cells. Adenomas wereepigenetically diverse populations, suggesting that they were rela-tively mitotically old populations. Intratumor clones typically showedless diversity in methylation pattern than the tumor as a whole.Mathematical modeling suggested that recent clonal sweeps encom-passing the whole adenoma had not occurred. Adenomatous cryptswithin human tumors contain actively dividing stem cells. Adenomasappeared to be relatively mitotically old populations, pocketed withoccasional newly generated subclones that were the result of recentrapid clonal expansion. Relative stasis and occasional rapid subclonegrowth may characterize colorectal tumorigenesis.

intratumor heterogeneity | tumor growth | tumor life-history |intestinal adenomas | cancer stem cells

The development of colorectal cancer along the adenoma–carcinoma pathway has become the archetypal model of solid

tumor evolution (1). Both the genetic lesions and morphologicalfeatures that evolve during colorectal carcinogenesis are wellcataloged (2, 3), but remarkably little is known about the dy-namics of the intratumor clones that bear these lesions. Fur-thermore, although mouse models point to the presence of stemcell compartments within adenomas (4), the cellular hierarchy ofhuman colorectal adenomas is undetermined. The dynamics ofthese intratumor stem cell clones critically inform the search foreffective biomarkers, provide a means to rationalize surveillancestrategies, and potentially guide the choice of therapeuticinterventions (5).Human adenomas have relatively low malignant potential:

Longitudinal studies have found that fewer than 1 in 10 adenomas

become malignant within 10 y of first detection (6). Estimates ofadenoma growth rates based on longitudinal endoscopic andbarium observational studies suggest that adenomas remain rela-tively static in size for many years, with a large proportion ofsmaller lesions even regressing over time (7–9). Modeling of therelative mutation burden of colorectal cancers versus adenomassuggested that it takes 17 y for a large adenoma to become ma-lignant (10). However, the clonal dynamics during this period ofcarcinogenesis are unclear. Intratumor clonal evolution maybe characterized by the independent evolution of many dif-ferent persistent subclones; alternatively, there may be ex-tensive clonal replacement by newly generated mutant clones(selective sweeps).Colorectal adenomas typically are composed of crypts (Fig. 1),

self-contained structures that are morphologically similar to theirnondysplastic counterparts in the normal colon. Mouse modelssuggest that the hierarchies of cell organization within adeno-matous crypts are caricatures of the normal intestine in whichrapidly cycling stem cells [expressing the leucine-rich repeat-containing G protein-coupled receptor 5 (Lgr5) gene] are locatedat the crypt base (4). However, the cellular hierarchy and lo-cation of stem cells within human adenomas is undetermined.Crypt fission—the bifurcation of a crypt into two daughtercrypts—is the predominant mechanism by which a mutated cryptcan expand clonally, both within the normal epithelium to form anadenoma and within early (noncancerous) adenomas themselves

Significance

The organization of cells within human colorectal adenomas,and specifically whether the tumors are maintained by stemcells, is unclear. Furthermore, the patterns of clonal evolutionleading to the development of a malignant tumor have notbeen determined. We performed lineage tracing in humanadenomas using a combination of nuclear and mitochondrialDNA lesions and epigenetic markers. Our data identify a stemcell population within adenomas and suggest that new growthof intratumor clones occurs infrequently, not as a steady con-tinual process as often is assumed. Our work offers a uniqueinsight into human cancer development.

Author contributions: A.H., S.A.C.M., N.A.W., and T.A.G. designed research; A.H., B.C., L.J.G.,D.S.J.M., B.D., A.G., G.E., E.N., R.J., R.P., and T.A.G. performed research; A.H., M.R.N., M.R.-J.,S.A.C.M., N.A.W., and T.A.G. analyzed data; and A.H., N.A.W., and T.A.G. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence may be addressed. E-mail: n.a.wright@qmul.ac.uk or trevor.graham@ucsfmedctr.org.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1220353110/-/DCSupplemental.

E2490–E2499 | PNAS | Published online June 13, 2013 www.pnas.org/cgi/doi/10.1073/pnas.1220353110

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 8,

202

1

(11, 12). Concurrently, the proportion of branching crypts in anadenoma is increased ∼80-fold compared with the normal co-lonic epithelium (11).Clonal expansion by fission produces contiguous patches of

clonal crypts (13). Intratumor clones thus can be demarcated byidentifying patches of crypts carrying a clone-defining geneticlesion (14). Additionally, deficiency of cytochrome c oxidase(CCO), readily detectable by histochemical staining, is a meansof visualizing intratumor clones (15). CCO deficiency usually isattributable to a mutation of the mtDNA where the gene isencoded; thus the shared ancestry of a patch of CCO− crypts canbe demonstrated by their having a clonal mutation. Furthermore,the presence of multiple cell lineages within a CCO− clone

demonstrates that the clone contains a multipotential stemcell (16).A serendipitous means to study dynamics and infer rates of

clonal expansion in human tissues is via analysis of methylationpatterns of CpG islands associated with nonexpressed genes (17).Methylation and demethylation at (some) nonfunctional locioccurs stochastically during DNA replication and is somaticallyinherited. Therefore, comparison of the methylation patternsbetween two somatic cells reveals their clonal relationship: Cellswith a recent common ancestor will tend to have similar meth-ylation patterns, whereas distantly related cells are unlikely toshare similar methylation patterns. In the normal human colon,small clonal patches of crypts tend to have dissimilar methylation

Fig. 1. Multipotent stem cells reside within clonal CCO− adenomatous crypts. (A) (i) H&E staining showing tubular adenoma C157 with low-grade dysplasia.(ii) CCO enzyme histochemistry identifying two patches of multiple, blue, CCO− crypts. (iii and iv) Laser-capture microdissection of areas in ii outlined in redand (v) mtDNA sequencing of single cells from multiple blue crypts and within the same blue crypts versus adjacent brown, wild-type crypts demonstratedthat all blue crypts shared a common, clonal point mutation in their mtDNA that was not present in adjacent brown crypts. (B) Immunofluorescence stainingof serial sections from adenoma 160. Clonal CCO− crypts contained cells positive for markers of neuroendocrine cells (chromogranin A) and secretory cells[mucin2 (MUC2) and mucin5AC (MUC5AC)], indicating these crypts contained a multipotential stem cell that had produced these distinct cell types. Detectionof CCO expression was conducted on the same section as chromogranin A after visualization for chromogranin A expression. Detection of MUC2 and MUC5ACexpression was conducted on the same section simultaneously. Negative controls were isotype-matched at the same concentration as the correspondingprimary antibody. Asterisks indicate the crypt enlarged in high-power images. (Scale bars: ∼50 μm in low-power images and 25 μm in high-power images. (C)LGR5 mRNA detection in FFPE tissue from patients with familial adenomatous polyposis (left pair) and sporadic adenocarcinoma (right pair). Expression isdetectable in bases of unaffected crypts, in patches of adenomatous epithelium, and, in this example, extensively through invasive adenocarcinoma. Bright-field and dark-field reflected light image pairs with Giemsa counterstain.

Humphries et al. PNAS | Published online June 13, 2013 | E2491

MED

ICALSC

IENCE

SPN

ASPL

US

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 8,

202

1

patterns, suggesting that clonal expansion rates are very slow in thenormal gut (18). Accordingly, methylation patterns offer a meansof infering the relative life history of tumor clones; epigeneticallydiverse tumors are likely to be relatively old clonal expansions,whereas relative homogeneity of the methylation pattern within aclone indicates de novo clonal expansion.Here, we have shown that human adenomatous crypts are

clonal populations maintained by multipotential stem cells. Wehave identified intratumor clones and characterized the evolu-tion of these clones and of the tumor itself by analyzing meth-ylation pattern diversity within these clones.

ResultsClonal and methylation analyses were performed on six snap-fro-zen sporadic human adenomas and one hyperplastic polyp. Ag-gregate DNA extracts from each tumor were screened foralterations in the genes frequently mutated in colorectal cancers(Table 1). Truncating mutations in adenomatous polyposis coli(APC) were detected in every adenoma but not in the hyperplasticpolyp, and v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog(KRAS) mutations were detected in three adenomas. No adeno-mas had detected mutations in exons 5–8 of tumour protein 53(TP53). No loss of heterozygosity (LOH) was detected on chro-mosome 5q close to APC or on 17p close to TP53 by LOH analysisof up to three microsatellite markers.Microsatellite instability (MSI), as identified by BAT25 and

BAT26 instability (not the full Bethesda panel), was not detectedin any of the adenomas, and all had wild-type v-raf murine sar-coma viral oncogene homolog B1 (BRAF), suggesting that theCpG island methylator phenotype (CIMP) was unlikely in theseadenomas (19–21). No significant difference in global genomicDNA methylation level, as measured by pyrosequencing of CpGswithin the long, interspersed nuclear element 1 (LINE-1) ele-ments, was seen between the normal and adenomatous epithe-lium of each patient (Fig. S1; paired Wilcox test, P = 0.5)suggesting that the rate of change in basal methylation withinadenomas was comparable to that of the surrounding morpho-logically normal tissue.Absence of CCO activity was used as a proxy to detect mtDNA

mutations (15). The mtDNA from cells from contiguous patchesof CCO-deficient (CCO−) adenomatous crypts then were PCR-sequenced to identify clonal mtDNA mutations. A clonal pointmutation was sought and found in the CCO− patch in adenoma157 and in four patches in the hyperplastic polyp; mtDNA se-quencing was not performed on other CCO− patches.

Adenomatous Crypts Are Clonal Populations Maintained by MultipotentStem Cells. CCO− crypts were identified by enzyme histochemistry(Fig. 1A). Multiple small areas, each approximating the area of asingle cell, were microdissected from CCO− crypts in the hyper-plastic polyp (sample 154) and in adenoma 157 and were sub-jected to whole-mtDNA sequencing. The same mtDNA pointmutation was present in all blue CCO− areas of each crypt butwas not present in the surrounding brown CCO-proficient(CCO+) crypts, confirming that the CCO− crypts were a clo-nal population. Adenoma 157 is shown in Fig. 1A.To demonstrate that the clonal CCO− crypts contained mul-

tipotential stem cell(s), immunofluorescence for markers of dif-ferent cell types (mucin 2 and mucin 5AC for secretory cells andchromogranin A for neuroendocrine cells) was performed oncrypts from formalin-fixed, paraffin-embedded (FFPE) blocks ofadenoma 160 and the hyperplastic polyp. All these cell typescould be detected within single CCO− crypts of adenoma 160,confirming that they were derived from the same multipotentialancestor stem cell (Fig. 1B). Goblet cells and neuroendocrinecells also were detected within single CCO− crypts of the hy-perplastic polyp (Fig. S2).In addition, isotopic in situ hybridization (ISH) was used to

detect LGR5 transcripts in an additional adenoma and alsoin a colorectal carcinoma. Adenomatous crypts containedLGR5-expressing cells and showed an expanded population ofTa

ble

1.Bulk-gen

otypes,ep

igen

etic

descriptionofea

chtumor,an

dpatientdetails

Aden

oma

Age,

ySe

xHistopathology

Size

,cm

LINE-1%

methylation

APC

KRAS

BRAF

MSI

TP53

%methylation

CSX

ACD

ICD

Norm

altissue

Aden

oma

1(148

)73

MTV

A,LG

D2.2

——

c.43

48C>

Tc.35

G>

AW

TMSS

WT

761.09

1.42

Ascen

dingco

lon

p.R14

50*

p.G12

D2(157

)66

MTA

,LG

D2.2

7672

c.39

64G

>T

WT

WT

MSS

WT

571.69

3.29

Sigmoid

p.E13

22*

3(158

)64

MTV

A,LG

D1.3

7072

c.39

56delC

c.38

G>

AW

TMSS

WT

870.96

1.53

Rectal

p.A12

99fs*1

p.G13

D4(160

)69

FTV

A,LG

D1.2

7263

c.41

18_4

126d

elCTG

AACAC

WT

WT

MSS

WT

940.68

0.87

Sigmoid

p.P13

71fs*1

05(162

)62

FTV

A,LG

D2

7370

c.44

74_4

475d

elGCinsT

WT

WT

MSS

WT

771.19

2.75

Splenic

flex

ure

p.A14

92fs*1

46(174

)60

MTA

,LG

D—

——

c.39

25delGAAAA

c.35

G>

AW

TMSS

WT

341.50

3.38

p.E13

09fs*4

p.G12

D7(154

)61

MHyp

erplastic

polyp

Sigmoid

1.1

7680

WT

WT

WT

MSS

WT

362.19

3.54

ACD,intracryptdistance

(ave

rageforallcrypts

inad

enoma);ICD,intercryptdistance

(ave

rageforallp

airs

ofcryp

tsin

thead

enoma);L

GD,low-gradedysplasia;

MSS

,microsatellite

stab

le;T

A,tubularad

enoma;

TVA,tubulovillu

sad

enoma;

–,nodataav

ailable.

E2492 | www.pnas.org/cgi/doi/10.1073/pnas.1220353110 Humphries et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 8,

202

1

LGR5-expressing cells when compared with nonadenomatouscrypts (Fig. 1C). In the carcinoma, LGR5 expression wassignificantly up-regulated: Nearly all malignant cells appearedto contain transcripts.

Adenomas Have Diverse Methylation Patterns. Individual crypts weremicrodissected from each adenoma, and clonal bisulphite se-quencing was used to determine the methylation patterns at CpGislands associated with the gene cardiac-specific homeobox [CSX(NKX2-5)] that is not expressed in the colon. Within each ade-noma, the variation in methylation patterns between the crypts wasmarked (Fig. 2 and Figs. S3 and S4), suggesting that each adenomawas of sufficient somatic age to have evolved distinct intratumoralpatterns of methylation. After correction for the degree of meth-ylation (SI Materials and Methods and Fig. S5), the diversity of theintercrypt methylation pattern was found to differ significantlyamong adenomas (F-test, P < 0.001; Fig. 2 B and C), suggestingdifferences in the evolutionary history of each adenoma.

Intratumor Clones Are Epigenetically Homogeneous. Intratumorclones were identified by their sharing a somatic mtDNA mu-

tation (adenoma 157 and the hyperplastic polyp; Fig. 1) not borneby the rest of the crypts in the adenoma, or, in the case of adenoma174, a KRAS point mutation present only in focal regions of thetumor (Fig. 2A). For four of five mtDNA-marked clones, the di-versity in the intercrypt methylation pattern in the crypts com-prising the intratumor clone was significantly less than the diversitybetween the clone and the rest of the tumor (Wilcox test, P < 0.05in four of five cases; Fig. 3), indicating that these clones repre-sented recent clonal expansions. Interestingly, the pattern ofmethylation diversity within the KRAS clone within adenoma 174was similar to the pattern in the bulk of the adenoma (Wilcox test,P = 0.8; Fig. 3), suggesting this KRAS-mutant clone was not arecent clonal expansion.To look for further evidence of recent subclone growth in the

adenomas, we considered the distribution of epigenetic distancesbetween crypts in the adenoma. The distribution of intercryptdistances (ICDs) within each adenoma was skewed with either anextended right or left tail (Fig. S6). Therefore, most crypts withineach adenoma were equally unrelated to one another. However,there were a few exceptional crypt-pairs in each adenoma thateither were particularly epigenetically dissimilar to the rest of the

APC/KRASAPC

67

2 3 4 5

12

17

1617

1 2 3 45

6

7

10

12

13

A

21 22

23

24

25262729

30

1

21 23

25

26

27

24

29

10

3016

13 22

B C

Fig. 2. Epigenetic diversity in adenomatous crypts.(A) Topographical maps showing crypts sampledfrom adenoma 174 (wholly APC-mutant, KRAS-mutant subclone). A pairwise comparison of adeno-matous crypts showed crypts had markedly differentmethylation patterns. Individual adenomatous cryptsshowed intracrypt epigenetic diversity. Boxes of dotsrepresent methylation patterns; each row is a tag(molecule), each column a CpG site. An open circledenotes an unmethylated site; a filled circle denotesa methylated site. (B and C) Pairwise differences be-tween each pair of crypts, sorted by adenoma. Insome adenomas (e.g., 160 and 158), all crypts hadsimilar methylation patterns, whereas in other ade-nomas (e.g., 162) some crypts had methylation pat-terns that were very different from those in themajority of crypts, illustrating the differences in thelife-history of each adenoma. (B) ICD and (C) mini-mum distance between crypts. Colors indicate thedominant genotype within the adenoma; blue: APC–mutant KRAS–wild-type, red: APC/KRAS-mutant,green: APC/KRAS–wild-type).

Humphries et al. PNAS | Published online June 13, 2013 | E2493

MED

ICALSC

IENCE

SPN

ASPL

US

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 8,

202

1

adenomatous crypt-pairs, or were particularly epigeneticallysimilar to one another. Finally, distance-based phylogenetic treeswere constructed (SI Materials and Methods) using the ICD ofmethylation patterns to define the distance between each of thecrypts. The trees had a nearly star-shaped structure, and theintratumor clones identified by genetic mutation tended to beclosely related (Fig. S7).

Relationship Between Spatial and Epigenetic Distance. To probefurther the evolutionary history of each tumor, the relationshipbetween the epigenetic distance and the spatial distance betweencrypts was examined. Straight-line distances between crypts weremeasured on a single slide for each adenoma. To facilitate com-parisons among adenomas, the distances were expressed in unitscorresponding to the average diameter of the crypts within theadenoma (averaged over at least 75 crypts when the adenoma wasof sufficient size; if the adenoma contained fewer than 75 crypts,all crypts in the section were measured). To compensate for therelatively sparse spatial sampling of the tumor, the average epi-genetic distance, g(r), of crypts that were within a radial distance of[r,Δr] of one another was considered (as per ref (22) where theinterval Δr was set at the average adenomatous crypt diameter,and the epigenetic distance between pairs of crypts was computedusing the metrics described above. There was no correlation be-tween the epigenetic and spatial distance of crypts in any of theadenomas (Fig. 4).

Tumor Growth Rate. The intercrypt diversity in the methylationpattern of adenomatous crypts within each adenoma was com-pared with the diversity of the same number of crypts in localizedpatches of normal colon (data reproduced from ref. 18), usinga case-resampling bootstrap method to compare equal numbers ofcrypts in each adenoma and the normal colon (SI Materials andMethods). After correction for the percentage methylation (SIMaterials and Methods), pairs of crypts from adenomas appeared,on average, only slightly more similar than pairs of normal crypts[linear regression, P < 0.01; ratio percent-methylation correctedmean ICD in normal tissue vs. mean ICD in adenoma, 1:1.01),suggesting that average crypt growth rates were not markedlygreater in the adenoma than in the normal colon.To assess further the relationship between intercrypt epige-

netic diversity and the ancestral relationship between crypts, theepigenetic diversity of a set of clonally derived crypts (e.g., asubset of crypts composing an adenoma) was simulated usinga mathematical model (Materials and Methods). Two models ofadenoma growth were considered: (i) the ”burst” model positingthat all adenomatous crypts formed at the same time froma single ancestral crypt (T years before sampling), and there wasno further crypt growth; and (ii) the “continual” model that

posited steady growth of crypts (each crypt divides every f years,with the first division occurring T years prior sampling). In bothmodels, the simulated adenomatous crypts were derived froma single ancestral (transformed) crypt. Epigenetic diversity wassuppressed in the continual growth model as compared with theburst model (Fig. 5) because continual growth led to crypts havinga more recent common ancestor than if the crypts were formed ina burst at the initiation of tumor growth. In the burst model, theepigenetic diversity of a population of crypts was strongly relatedto the time since the burst, or, equivalently to the time to the mostrecent common ancestor.The epigenetic diversity measured in adenoma 157, and the

CCO− subclone detected in this adenoma was compared with themodel’s predictions. The comparison showed that a clone thathad grown as a burst at least a decade previously could explainthe diversity observed in random samples of five crypts from thesubclone. The adenoma bulk was predicted to be significantlyolder. The same pattern was observed for the diversity of theintratumor clones in the other adenomas.

Stem Cell Dynamics Within Adenomatous Crypts. The clonal archi-tecture of individual adenomatous crypts was investigated byconsidering the diversity of the methylation pattern within in-dividual crypts. Individual crypts typically contained multiple dis-tinct methylation tags (Fig. 2). The diversity of the methylationpattern within individual adenomatous crypts was comparable tothe diversity of the methylation pattern observed in morphologi-cally normal crypts from nonpaired patients (Fig. 6A; normal cryptdata reproduced from ref. 18). After correcting for the percentageof methylation, diversity in adenomatous crypts was slightly but notstatistically significantly less than in normal crypts (linear model,P = 0.2), suggesting that adenomatous crypts have stem cell dy-namics similar to those of their normal counterparts. Comparisonof a stochastic model of stem cell turnover within crypts [previouslyproposed for normal crypts (17)] suggested that this level ofmethylation diversity is consistent with each crypt containinga number of stem cells that compete with one another for theirplace in the crypt (Fig. 6A). Further evidence for the existence ofmultiple long-lived stem cells in each crypt was derived from theobservation of adenomatous crypts composed of a mixture ofCCO− and CCO+ cells (Fig. S8); the CCO+ and CCO− cells de-rived from different stem cells.The diversity of methylation patterns observed within crypts

tended to be significantly lower than the diversity among all pairsof crypts within the adenoma (Fig. 6B), indicating that cryptsthemselves comprised a clone that had a more recent commonancestor than the most recent common ancestor of the tumor asa whole. In other words, the tumor population appeared mitot-ically older than any individual crypt.

C CN W C CN W C1

CN

1

C2

CN

2

C4aC

N4a

C4b

CN

4b

W

02

46

8IC

D

174 (KRAS)

157 (mtDNA) 154 (mtDNA)

*** ** *** *

Fig. 3. Intercrypt diversity of the methylation patterns from intratumor clones compared with the diversity of the bulk of the adenoma. Methylationpatterns of subclones tended to be significantly less diverse than the bulk of the tumor cell population. C, intraclone diversity (red); CN, diversity betweenclone and nonclone regions of the tumor (blue); W, diversity of whole tumor (green); ***P <.001, **P < 0.01, *P < 0.05 for Wilcoxon test between clone andnonclone; numbers at the top of the plot indicate the adenoma along with the mutation denoting the subclone.

E2494 | www.pnas.org/cgi/doi/10.1073/pnas.1220353110 Humphries et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 8,

202

1

DiscussionThe clonal evolution of cancer remains poorly understood, de-spite extensive knowledge about the genetic and morphologicalchanges that occur during the progression to malignancy (2).Here we investigated the clonal dynamics of adenomatous cryptpopulations within human colorectal adenomas. We found lim-ited examples of clonal expansions within adenomas; instead,relative stasis of adenomatous crypt clones appeared to be thenorm. These data are consistent with a model for tumor growthin which the time to the most recent common ancestor of eachcrypt-pair is greater than the time needed for the methylationpatterns of two clonally derived daughter crypts to diverge. As-suming that the time needed for methylation patterns of daughtercrypts to diverge is on the order of a decade [as appears reason-able in the normal colon (18)] and that the adenomas studied arearound a decade old (10), our results suggest that the majority of

the adenomatous crypt population is produced at around the sametime, near the onset of tumor formation, and then this initial periodof growth is followed by relative stasis (Fig. 7). Further supportingthis notion is our observation that, on average, the methylationpatterns of crypts that were located close together in the adenomawere no more similar than those of crypts that were disparatelyspaced (Fig. 4). For comparison, in a tumor composed of a mosaicof rapidly growing clones, individual clones would likely be identi-fiable as patches of crypts with relatively similar patterns of meth-ylation (23). Interestingly, similar patterns of epigenetic diversityhave been reported in advanced colorectal cancers: Typically thelate-stage lesions showed little evidence of recent clonal expan-sions, and all pairs of cancer crypts appeared equally unrelated bytheir methylation patterns (24). Taken together, these datasuggest that relative stasis, in which clonal outgrowth occurs ata slower rate than the rate of methylation pattern divergencemay be characteristic of all stages of colorectal carcinogenesis.It is instructive to consider the evolutionary selective forces

that are acting on the tumor. Supposing that strong selection fornew mutants would lead to rapid clonal expansion, growingintratumor clones would be expected to be relatively epigeneti-cally homogeneous. The high diversity observed in the adenomassuggests only weak selection for new mutants (presuming thatthese new mutants are present) in early adenomas. Conse-quently, it is reasonable to suppose that the founder clone of theadenoma is sufficiently well adapted so that additional mutantsconfer relatively little fitness benefit (25, 26). Clonal evolution,analogous to tumor progression, therefore may occur in fits andstarts (27), perhaps driven by the infrequent emergence ofa particularly well-adapted clone that has undergone significantgenetic evolution. Clearly, genome-wide approaches are re-quired to test this hypothesis. The textbook model of carcino-genesis suggests that the progression to malignancy proceeds ina regular stepwise fashion, with repeated rounds of mutation andclonal expansion of the newly mutated clone (28, 29). Insteada punctuated model in which clonal expansion is rare (and per-haps slow), and clonal stasis is the norm, may explain our data. Ifthe initial growth of an adenoma is rapid, perhaps because theresource that restricts growth in later adenomas has not yet be-come limiting, then the initial growth phase may representa period when new mutations are likely to become established inthe adenoma population. A mutant crypt produced in this initialgrowth phase will likely expand beyond the point where it islikely to go extinct by occasional (random) crypt extinctions.When the period of relative stasis is reached (and the phase ofnet adenoma growth is over), clonal expansion may be relativelymore difficult, restricting the ability of new mutants to geta foothold within the tumor. This model explains the observationof the KRAS-mutant subclone in adenoma 174 being as epige-netically diverse as the rest of the adenoma bulk. Such a mode ofclonal evolution may go some way to explaining the long interval[which may be as long as 17 y (10)] between the establishment ofthe adenoma and invasion. Moreover, if similar considerations

02

46

8

spatial distance

ICD

A

0 10 20 30 40 50 60

158157148174160154162

05

1020

30

spatial distance

cum

ulat

ive

ICD

B

0 10 20 30 40 50 60

158157148174160154162

Fig. 4. Correlation between spatial and epigeneticdistance in adenomas. (A) Spatial distance versusepigenetic distance for whole adenomas. No cor-relation between spatial distance and epigeneticdistance was observed. Spatial distances betweencrypts were binned into intervals of five units, witha unit corresponding to the average crypt diameter.The average ICD of all pairs of crypts within theadenoma that fell within the spatial interval wascomputed. Each colored line represents a differentadenoma. (B) Spatial distance versus cumulativeepigenetic distance. The straight lines suggest thatphysically close crypt-pairs are no more similar intheir methylation patterns than physically distantpairs of crypts. If physically close crypts were moreepigenetically similar on average than distant pairs of crypts, the cumulative epigenetic distance would increase exponentially. The cumulative epigeneticdistance was computed from the data in A to compensate for the sparse spatial sampling of adenomas.

age of clone (years)

mea

n in

ter−

cryp

t dis

tanc

e

0 5 10 15 20 25 30

Model of crypt fission every 2 yearsModel of initial burst growth

01

23

4

subclonediversity

bulkdiversity

Fig. 5. Modeling epigenetic diversity during clonal expansions. The di-versity of methylation patterns between a patch of five clonally derivedcrypts was simulated with a mathematical model. Two modeling assump-tions were tested: (i) that the patch formed as burst (with very rapid divisionof the ancestral crypt) at the onset of tumor growth, and (ii) that the cryptswithin a patch continued to divide every f years. The epigenetic diversity ofthe crypts was greater in the initial burst model (cyan line) than in thecontinual growth model (orange line, shown for crypt divisions every f = 2 y),and this diversity increased as the clone aged. Dashed lines represent 95%quantiles of the simulated values. Boxplots represent the mean ICD ofsamples of repeated five crypts taken from the CCO− clone in adenoma 157(red box) or from the nonclonal area (blue box). The diversity of the subclonewas significantly less than that of the nonclone, suggesting that the sub-clone shared a much more recent common ancestor than the bulk of theadenoma.

Humphries et al. PNAS | Published online June 13, 2013 | E2495

MED

ICALSC

IENCE

SPN

ASPL

US

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 8,

202

1

apply to other preinvasive lesions, such as ductal carcinoma insitu, then the interval before invasion could be explained simi-larly (30). We note that it is challenging to prove this hypothesisconclusively with our data, because the expansion of cloneswithin the tumor may be so slow as to be undetectable by ourmethylation-pattern molecular clock assay.It is interesting that we have detected the presence not only of

multipotential stem cells but also of clonal expansion withina hyperplastic polyp: Such solitary lesions have not been repor-ted to show evidence of progression potential (31), and althoughclonal genetic changes are observed in some hyperplastic polypswithin the context of hyperplastic polyposis (32) associated withBRAF mutations (33), this lesion was microsatellite stable andBRAF wild type. Nevertheless, it is clear that mutations drivinga clonal expansion can be selected within such solitary lesions,possibly supporting the view that hyperplastic polyps may befundamentally neoplastic rather than hyperplastic, even in theabsence of microsatellite instability and/or BRAF mutation (34).Mathematical modeling suggested that the significantly lower

diversity typically observed in demarcated subclones was causedby the subclone’s having a common ancestor that was signifi-

cantly more recent than the common ancestor of the bulk of theadenoma. The model was predicated on the assumption ofsteady rates of cell division (one division per day) and a constantrate of methylation change (2 × 10−5·CpG site−1·division−1).Because of the uncertainty surrounding these rate estimates, andindeed because of the constancy of the rates, coupled withuncertainties about the number and patterns of division of ad-enomatous crypt stem cells, the estimates of time derived fromthe model can be considered only illustrative. Indeed, methyla-tion at nonexpressed loci tends to be higher in cancers than inage-matched normal mucosa, suggesting carcinogenesis-specificdifferences in the evolution of the methylation pattern (35).What is clear from these data, however, is that adenomas aremitotically old populations, and the observed clonal expansionsare rare and relatively slow events. Only two scenarios of ade-noma clone growth were considered in our model: continualsteady growth or an initial burst followed by relative quiescence.Biological reality may a complex hybrid of these two growthpatterns, but we believe that our data are insufficient to distin-guish the precise clonal expansion rates.

��

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

���

�

�

�

�

�

�

�

�

�

�

���

�

�

�

�

�

�

�

�

�

�

�

�

�

60 65 70 75 80 85

01

23

4

age

ACD

A

A A A A A A A

02

46

8

dive

rsity

I I I I I I I

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

���

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

� �

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

��

�

�

�

�

�

�

�

��

�

�

�

�

��

�

�

�

�

�

�

�

�

���

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

� �

�

�

��

�

��

��

�

��

���

�

�

�

��

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

��

�

�

�

��

�

�

�

�

�

��

�

�

�

�

�

� �

��

�

��

�

�

�

�

�

�

�

��

�

�

��

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

���

��

��

�

�

�

�

�

�

�

�

��

�

�

�

� ��

� �

�

�

�

�

�

�

�

�

�

�

�

�

�

��

��

�

�

�

�

�

�

�

�

�

�

� �

�

�

�

�

��

�

� �

�

��

��

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

�

�

�

��

�

�

�

�

��

�

�

�

�

�

�

��

�

�

�

���

�

�

�

�

�

��

�

�

�

�

� ��

�

�

�

��

�

�

�

���

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

�

� �

�

�

�

�

��

��

�

�

�

�

�

��

�

�

��

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

� �

�

�

�

�

���

�

�

�

�

�

� �

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

� �

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

� ���

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

��

�

�

�

�

�

�

�

�

�� ��

�

�

�

�

�

��

��

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

� �

�

�

�

�

�

� �

�

�

� �

���

�

�

��

�

�

�

��

��

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

� �

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

� �

��

�

�

��

�

��

�

�

�

� �

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

��

��

� ��

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

��

�

�

�

�

�

��

��

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

��

� �

�

�

�

�

�

�

�

�

��

�

�

�

� �

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

��

�

�

�

�

� �

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�� �

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

��

�

�

�

�

�

�

�

�

�

�

�

�

� �

�

�

�

�

�

�

��

�

�

�

� ��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

� �

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

��

�

�

�

�

� �

�

��

�

��

�

148 157 158 160 162 174 154** ** ** * **

B

Fig. 6. Diversity of the methylation pattern of individual adenomatous crypts compared with the adenoma bulk. (A) Intracrypt diversity of normal and adeno-matous crypts as a function of age is consistent with amodel of neutral stem cell competitionwithin crypts. Points are intracrypt diversity measured in normal crypts(blue circles) or adenomatous crypts (red triangles); the x-axis (age) jitter is artificial to allowall datapoints tobe seen. Solid black line showsmean intracryptdiversityas predicted by the mathematical model with parameters as described in Materials and Methods; dashed lines are 95% quantiles for the simulation results. (B)Intracrypt distances (A) versus ICDs (I) for each adenoma. The mean ICD was greater than the intracrypt distance for all adenomas. Statistical comparison wasperformed using a Wilcox test: **P < 0.001, *P < 0.05. Unmarked pairs did not show a statistically significant difference.

MRCA

MRCA

sampled genotypes

sampled genotypes

time/divisions/methylation pattern divergence

time/divisions/methylation pattern divergence

A B

time/divisions/methylation pattern divergence

exte

nt o

f clo

nein

ade

nom

a successive wavesof clonal expansion

tumourinitiatingclonal expansion

time/divisions/methylation pattern divergence

exte

nt o

f clo

nein

ade

nom

a

clone generated early in tumorigenesisundergoes limited clonal expansion

infrequent clonegeneration

Fig. 7. Cartoon of tumor evolution. (A) A stepwise model oftumor evolution in which sequential mutations trigger exten-sive clonal expansions of the new mutant clone within thetumor mass. (B) A model for the relative stasis scenario of co-lorectal adenoma evolution. Here, intratumor clones can formnear the onset of neoplasia but do not sweep through thetumor and so appear as spatially localized clones with di-vergent intraclone methylation patterns. Rare subclones formlater in tumor development, have relatively homogeneousmethylation patterns, and also occupy only focal regions of thetumor. Colors denote distinct clones. Within a clone, methyl-ation patterns diverge as the clone ages. MRCA, most recentcommon ancestor.

E2496 | www.pnas.org/cgi/doi/10.1073/pnas.1220353110 Humphries et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 8,

202

1

Theoretical modeling of the relationship between spatial andgenetic distance suggests that a tumor composed of a patchworkof competing clones would show an asymptotic relationship be-tween the average genetic distance and spatial distance betweencrypts (22). This relationship would be the result of clonalcompetition whereby the growth of a clone is slowed by its havingto compete with its comparably fit neighboring clones. Geneticdistance between two clones was defined as the number of lociwhere the genotype of the two clones differed (the Hammingdistance). In our study, we saw no relationship between thespatial distance and the average epigenetic distance of non-selected loci of individual crypts, and consequently it is temptingto conclude that the adenoma is not composed of a patchwork ofcompeting clones. However, there are caveats to this conclusion.First, the relationship between the spatial distance and geneticdistance of crypts is dependent on the mutation rate. When themutation rate is high, the dependence occurs over only shorttime scales, because for high mutation rates, which better ap-proximate the methylation error rate, records of ancestry are lostrapidly. Second, it is conceivable that the early adenomas may benearly monoclonal populations composed a single clone drivenonly by an APC mutation (36). Clonal competition indeed mayoccur when a clonal outgrowth happens, but our data may recordfew occurrences of clonal outgrowth.Both the percentage of methylation and the diversity of the

methylation pattern varied significantly among tumors. Thesemarked differences suggest that the pattern of growth differedamong tumors: Diverse tumors likely were older neoplasms thantheir more genetically homogeneous counterparts. However, wecannot discount the hypothesis that the variation in the evolu-tionary history of each adenoma is explained by inherent dif-ferences between the clonal dynamics of the normal mucosa ofeach patient that are inherited by the adenoma. The apparentdifferences in the evolution of each tumor were notably notreflected by differences in stage and grade: All adenomas showedlow-grade dysplasia. Thus, histopathological classifications maynot represent precise measures of the state of tumor evolution.Interestingly, KRAS-mutant adenomas tended to be morehypermethylated than their KRAS wild-type counterparts (84 vs.65%, respectively; Wilcox test, P < 0.01) although the robustnessof this association is unclear because of the small size of ourtumor cohort. The percentage of methylation presumably reflec-ted the methylation status of the initial clone that formed thetumor. Methylation accrues with mitotic aging (17), and so it istempting to suggest that KRAS-mutant tumors may have formedfrom a clone that was mitotically older than the initiating clone ofKRAS wild-type tumors. Activation of oncogenic KRAS in theotherwise normal intestinal epithelium causes hyperplasia andincreases proliferation with crypts (37) but is insufficient to initiatetumor growth (38). Thus, hypermethylated KRAS-mutant adeno-mas may have grown from a crypt that acquired an activatingKRAS mutation before the gatekeeper APC mutation.Next-generation sequencing provides a powerful means for

studying tumor evolution and permits assaying many more loci inthe genome, and at greater depth, than is possible with the clonalbisulphite sequencing technique that we applied. Indeed, suchanalysis has suggested punctuated evolution of breast cancers(39, 40) and found extensive genetic heterogeneity in renal car-cinomas (41, 42) and B-cell chronic lymphocytic leukemia (43).Although our study lacks the extensive genetic characterizationof those approaches, our methods have two benefits. First, bysequencing individual crypts, we retain spatial information andrestrict ourselves to the study of clonal units, readily allowing thespatial extent of clonal proliferations to be determined. Second,the methylation patterns of nonexpressed genes are selectivelyneutral, removing the potential confounding imposed by selectionin the interpretation of lesion accumulation at nonneutral loci.The morphological basis of clonal evolution is not discussed

often. If there is relative stasis within an adenoma, how is itmanifested? Intestinal tumors are hyperproliferative at the level ofindividual cells (11), although this cellular proliferation is unlikely

to relate to the growth of an adenoma as a whole, because it maysignal only a rapid turnover of cells within individual crypts, notnet growth. Correspondingly, apoptosis at the crypt base, the likelylocation of the stem cell compartment (4), is a prominent featureof small adenomas [those less than 5 mm in diameter (44)], sup-porting our conclusion of turnover within adenomatous crypt stemcells and perhaps within crypts also. Furthermore, because there islittle change in the adenoma size over many years (8), any clonalexpansion (via crypt fission) within a tumor must be approximatelybalanced by an equal amount of crypt extinction. Adenomas havea dramatically increased proportion of branching crypts comparedwith the normal colonic epithelium (11); we note that this increasedoes not necessarily imply an increased crypt growth rate, becausethe branched crypts could represent “stalled” fission events orthere could be extensive crypt death leading to a high rate of cryptturnover and no net growth. Although directly observing cryptdeath events presents experimental challenges, we note that lu-minal apoptosis often is observed in both small and larger ade-nomas (44) and may be a hallmark of crypt death, and, in turn,clone extinction.Adenomatous crypts were shown to be clonal populations

containing multipotent stem cells. Thus, the clonal structure ofhuman adenomatous crypts matches that of the adenomatouscrypts in the mouse: Using their Lgr5/Apc/R26R-Confetti mice,Schepers et al. (4) were able, through a second tamoxifen in-jection, to perform lineage retracing of Lgr5+ stem cells in theadenoma, which were shown to be multipotential. Similarly, Fig.S8 shows adenomatous crypts partially occupied by CCO− mu-tant cells (which later convert wholly to CCO− cells, in which wehave shown multilineage differentiation). Such partially mutantcrypts also indicate the presence of multiple stem cells in eachadenomatous crypt.It has been proposed that LGR5, an accepted marker of in-

testinal stem cells in the mouse, also identifies human intestinalstem cells in both normal and neoplastic tissues (45–47). Here wehave shown an expanded population of cells expressing LGR5mRNA in an adenoma as compared with nonadenomatous crypts(Fig. 1C), similar to that reported in Apcmin adenomas (4). In thecarcinoma, LGR5 expression was detected in nearly all cells; ifindeed LGR5 marks human cancer stem cells, this observationwould suggest that most, if not all, colon carcinoma cells are stemcells, with few in the non-stem cell compartment. The non–self-renewing non-stem cell population represents an evolutionarydead-end, and so the apparent up-regulation of stemness observedin the cancer may be a consequence of selection for cells capableof self-renewing during cancer evolution.Stem cell dynamics were inferred by considering the diversity

in intracrypt methylation and were found to be consistent witha model of neutral stem cell competition (drift) within eachcrypt. Neutral competition between stem cells appears to bea hallmark of normal intestinal crypts (17, 48), and our datasuggest that the same dynamics persist in human adenomatouscrypts. Indeed, the diversity of adenomatous crypts was compa-rable to that observed in normal crypts, implying that the dy-namics of stem cell turnover in the normal and adenomatouscrypt are comparable. Furthermore, analysis of the variance inthe microsatellite pattern in mismatch repair-deficient adenomas(49) and in the methylation patterns in colorectal cancers (24)suggests that cell hierarchies may persist throughout carcino-genesis, albeit with altered numbers of stem cells.In summary, we have shown that human colorectal adenomas

are epigenetically diverse populations that show occasional sig-natures of recent clonal expansions. Colorectal adenomatouscrypts contain actively dividing multipotent stem cells.

Materials and MethodsTumor Selection, Histology, and Bulk DNA Extraction. Seven neoplasms wereobtained from seven patients according to institutional guidelines, and allpatients gave informed consent (Research Ethics Committee). Half of eachneoplasm was snap-frozen, and the other half was processed for diagnosticpurposes as FFPE blocks. Paired histologically normal margin tissue for each

Humphries et al. PNAS | Published online June 13, 2013 | E2497

MED

ICALSC

IENCE

SPN

ASPL

US

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 8,

202

1

tumor also was available in FFPE blocks. The frozen tumor tissue was seriallysectioned onto PALM Membrane Slides (Zeiss) to facilitate laser-capturemicrodissection at a thickness of 10 μm; typically ∼100 sections were takenfrom each tumor. FFPE tissue was serially sectioned (6 × 5μm sections) ontoplain glass slides for histochemical analysis. Histopathological classificationof the tumor and margins was assessed using a serial H&E slide by patholo-gists (N.A.W., M.R.-J., and M.R.N.).

Two additional neoplasms (one tubular adenoma, one moderately well-differentiated adenocarcinoma) processed as FFPE blockswere used for in situRNA detection (see below). Two further tubular adenomas were subjectedsolely to CCO staining (see below).

Bulk DNA extraction was performed by needle dissection. In the case ofFFPE material, all six serial sections were used for DNA extraction, and theserial H&E staining was used as a guide to enrich epithelial cell extraction.For adenomas, three entire serial sections were used for DNA extraction.DNA extraction was performed by overnight incubation in Picopure Pro-teinase K DNA extraction buffer (Arcturus Bioscience).

CCO Histochemistry. Frozen tumor sections were assayed for CCO activityusing a two-color enzyme histochemical staining as described by Taylor et al.(15). CCO− cells stained blue, and CCO-normal cells stained brown.

Immunofluorescence for Cell-Type–Specific Markers. Adenoma FFPE sectionswere probed with mouse anti-human CCO subunit IV (2 μg/mL) (Invitrogen),mouse anti-human Muc5AC (82 μg/mL) (Binding Site), rabbit anti-humanchromogranin A (46 μg/mL) (DAKO), and Muc2 (40 μg/mL) (AbCam)primary antibodies.

Antigen retrieval was performed in the microwave on rehydrated FFPEsections in boiling sodium citrate buffer, pH 6.0, for 10min (CCO) or Tris-EDTAbuffer, pH 9.0, for 15 min (Muc2 and Muc5AC). All sections were blockedusing a serum-free protein block (DAKO) for 10 min followed by incubationwith the primary antibody or isotype-matched negative control (DAKO) for35 min at room temperature. Sections were washed three times for 5 mineach washing in PBS followed by incubation with Alexa Fluor-conjugatedsecondary antibody for 35min (goat anti-mouse 488 for CCO andMuc5AC; goatanti-rabbit 555 for chromogranin A and donkey anti-rabbit 594 for Muc2)(Invitrogen). Sections were mounted with VECTASHIELD HardSet mountingmedium with DAPI (Vector Labs).

ISH for LGR5 RNA. ISH for LGR5 expression was performed using a riboprobecorresponding to LGR5 that was located 566 bp from the 5′ UTR to exon5(University of California, Santa Cruz chr12:70,120,102–70,233,231, intronsexcluded, cloned into pGEM3Z) (kind gift of Stefania Segditsas, WellcomeTrust Centre for Human Genetics, Oxford, United Kingdom). Hybridizationand detection was as previously described (50).

Microdissection. Laser-capture microdissection was performed using eitherPALM (Zeiss) or Leicamicrodissection systems. For mtDNA analysis, small areasof crypts, approximating a single cell, were microdissected from a single slide.For genomic andmethylation analysis entire crypts were microdissected fromthree or more serial sections. DNA extraction was performed by incubationin Picopure Proteinase K DNA extraction buffer (Arcturus Bioscience) fora minimum of 3 h. Tubes with only digestion buffer served as controls fordetecting contamination.

Mutation Detection. Bulk DNA extracts were screened for somatic mutationsin genes frequently mutated in colorectal cancers: the APC mutation clusterregion (approximately codons 1,250–1,450), KRAS exon 1 (codons 12–13), BRAF(the region around codon 600), and TP53 exons 5–9. Primers and conditions forscreening the mutation cluster region of APC are listed in Table S1; otherwiseprimers and conditions were as previously described (51). Sequences wereobtained using BigDye 3.1 cycle sequencing and were run on an ABI 3100DNA analyzer (Applied Biosystems).

Individual crypt DNA lysates were sequenced for mutations detected in thebulk-DNA sample from that adenoma.

Microsatellite LOH andMSI Analysis. LOH analysis was performed on individualcrypt lysates using up to six informative microsatellite markers close to theAPC gene (D5S346, D5S2001, and D5S489) and markers on 17p (D17S1832)and 18q (D18S474 and D18S58). Amplification was performed using a mul-tiplexed microsatellite PCR kit (Qiagen) using the primers previously de-scribed (51). Fragment analysis was performed on an ABI3100 system(Applied Biosystems), followed by analysis with the Peak Scanner software(Applied Biosystems). LOH was considered present if the area under one

allelic peak was more than twice that of the other, after normalizing peakareas relative to constitutional DNA. MSI was assessed using multiplexedassay for the BAT-25 and BAT-26 mononucleotide repeats as previously de-scribed (51).

Bisulphite Sequencing. Methylation pattern analysis was performed as pre-viously described (18). Briefly, individual crypt DNA lysates were treated withbisulphite using an Epitect bisulphite conversion kit (Qiagen). After conver-sion, a nested PCR protocol was used to amplify the CpG island locus withinCSX. PCR products then were cloned using the pGEM-T system (Promega) andwere sequenced to obtain single-strand resolution.

Detecting mtDNA Mutations. The whole mitochondrial genome was amplifiedusing a nested two-stage PCR protocol as previously described by Taylor andcolleagues (15). Only samples with a blank negative control were takenforward for sequencing.

LINE-1 Pyrosequencing. Bulk DNA extracts of adenoma tissue and pairedadjacent normal tissue (a histopathologically normal resection margin) weretreated with bisulphite using an Epitect bisulphite conversion kit (Qiagen) asdescribed above. Treated DNA was amplified using the PyroMark CpG LINE-1kit (Qiagen) and pyrosequenced on a PyroMark Q96MD sequencer (Qiagen).The kit assays four CpG sites within the LINE-1 repeated element; the per-centage ofmethylation at each CpG site was recorded and used in subsequentstatistical analysis.

To check whether stromal tissue within the adenomamight be confoundingthe LINE-1 methylation assessment, for adenoma 157 both macrodissectedadenoma and laser-capture microdissected (LCMD) adenomatous crypts onlywere run in parallel; the mean methylation level across the four CpG sites was72.1% for the macrodissected tissue and 72.2% for the LCMD sample, sug-gesting that any stromal cells present within the macrodissected adenomasamples did not affect the pyrosequencing analysis significantly.

Statistical Description of Methylation Patterns. A tag denoted the methylationpattern of one allele (one sequenced clone). To summarize the tags present ina single crypt, two summary statistics were used. The percentage of meth-ylation was the percentage of methylated CpG sites averaged across all sitesand tags. The intracrypt distance was the average pairwise distance betweenall pairs of tags in the crypt. The distance between two tagswas defined as thenumber of CpG sites in which the methylation status of the two tags differed.

To characterize the degree of variability in methylation pattern betweenpairs of crypts within an adenoma, two summary statistics were used. The ICDis the average pairwise difference between all the tags of one crypt comparedwith the tags of the other crypt. The minimum distance is the average dif-ference between each tag in the first crypt compared with its most similar tagin the second crypt.

Mathematical definitions of these metrics are as previously stated (18).

Mathematical Modeling of Methylation Pattern Dynamics.Methylation patterndiversity within crypts and in patches of clonally derived crypts was simulatedusing a stochastic probabilistic model as previously described (17, 18). Briefly,each crypt was assumed to contain a fixed number of stem cells denoted byN. At each (synchronous) round of division, each stem cell could produce no,one, or two stem cell offspring, and a corresponding two, one, or no non-stem cell offspring with probability p0, p1, and p2, respectively, constrainedso that the total number N stem cells remained after division. Each cellcontained two CpG island loci, each containing M = 8 CpG sites (denotingthe two alleles of CSX). Changes in the methylation status (methylation ordemethylation) of each site occurred with probability μ = 2 × 10−5 per di-vision; otherwise daughter cells inherited the methylation patterns of theirparent cell. Crypt fission was simulated by dividing the N stem cells equallybetween the two daughter crypts. The full stem cell complement of the newcrypts was formed by instantaneous division of the N/2 cells. For simplicity,simulations began with loci being randomly 50% methylated to account foran accumulation of methylation during an occult initiation phase of tumorgrowth (assumed to occur because of the high percentage methylationobserved in the adenomas). For each parameter set, at least 10,000 simula-tion repeats were performed, and averages and quantiles of these simu-lations were reported. Simulations were performed by proprietary C codethat is available on request.

Statistical Modeling. Statistics were performed in R using parametric ornonparametric tests as appropriate. Individual tests used are stated in themain text. Measures of methylation pattern diversity correlated with the

E2498 | www.pnas.org/cgi/doi/10.1073/pnas.1220353110 Humphries et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 8,

202

1

average percentage of methylation of the tags being compared. To accountfor this confounding factor, diversity measures were corrected for the per-centage of methylation when appropriate (see SI Materials and Methodsfor details).

ACKNOWLEDGMENTS. We thank the members of the Equipment Parkand Experimental Histology Laboratory, London Research Institute, Can-cer Research UK, and the histopathology core service at the Blizard In-

stitute, Queen Mary University of London, for technical assistance. TheLGR5 ISH probe was cloned and given to us by Dr. Stefania Segditsas.T.A.G., A.H., L.J.G., and N.A.W. were funded by Cancer Research UK (quin-quennial grant). T.A.G. was additionally supported by NIH-R01 GrantCA140657. A.H. received additional funds from The Jean Shanks Foundationand The R L St J Harmsworth Memorial Research Fund. S.A.C.M. and B.C. werefunded by the Medical Research Council (Grant 90901178). M.R.-J. is sup-ported by University College London Hospitals/University College LondonComprehensive Biomedical Research Centre.

1. Kinzler KW, Vogelstein B (1996) Lessons from hereditary colorectal cancer. Cell 87(2):159–170.

2. Fearon ER (2011) Molecular genetics of colorectal cancer. Annu Rev Pathol 6:479–507.3. Anonymous; Cancer Genome Atlas Network (2012) Comprehensive molecular

characterization of human colon and rectal cancer. Nature 487(7407):330–337.4. Schepers AG, et al. (2012) Lineage tracing reveals Lgr5+ stem cell activity in mouse

intestinal adenomas. Science 337(6095):730–735.5. Greaves M, Maley CC (2012) Clonal evolution in cancer. Nature 481(7381):306–313.6. Stryker SJ, et al. (1987) Natural history of untreated colonic polyps. Gastroenterology

93(5):1009–1013.7. Welin S, Youker J, Spratt JS, Jr. (1963) The Rates and Patterns of Growth of 375

Tumors of the Large Intestine and Rectum Observed Serially by Double ContrastEnema Study (Malmoe Technique). Am J Roentgenol Radium Ther Nucl Med 90:673–687.

8. Hofstad B, et al. (1996) Growth of colorectal polyps: Redetection and evaluation ofunresected polyps for a period of three years. Gut 39(3):449–456.

9. Tada M, Misaki F, Kawai K (1984) Growth rates of colorectal carcinoma and adenomaby roentgenologic follow-up observations. Gastroenterol Jpn 19(6):550–555.

10. Jones S, et al. (2008) Comparative lesion sequencing provides insights into tumorevolution. Proc Natl Acad Sci USA 105(11):4283–4288.

11. Wong W-M, et al. (2002) Histogenesis of human colorectal adenomas andhyperplastic polyps: The role of cell proliferation and crypt fission. Gut 50(2):212–217.

12. Preston SL, et al. (2003) Bottom-up histogenesis of colorectal adenomas: Origin in themonocryptal adenoma and initial expansion by crypt fission. Cancer Res 63(13):3819–3825.

13. Greaves LC, et al. (2006) Mitochondrial DNA mutations are established in humancolonic stem cells, and mutated clones expand by crypt fission. Proc Natl Acad Sci USA103(3):714–719.

14. Thirlwell C, et al. (2010) Clonality assessment and clonal ordering of individualneoplastic crypts shows polyclonality of colorectal adenomas. Gastroenterology138(4):1441–1454.

15. Taylor RW, et al. (2003) Mitochondrial DNA mutations in human colonic crypt stemcells. J Clin Invest 112(9):1351–1360.

16. Fellous TG, et al. (2009) A methodological approach to tracing cell lineage in humanepithelial tissues. Stem Cells 27(6):1410–1420.

17. Yatabe Y, Tavaré S, Shibata D (2001) Investigating stem cells in human colon by usingmethylation patterns. Proc Natl Acad Sci USA 98(19):10839–10844.

18. Graham TA, et al. (2011) Use of methylation patterns to determine expansion of stemcell clones in human colon tissue. Gastroenterology 140(4):1241-1250.e1249.

19. Samowitz WS, et al. (2005) Evaluation of a large, population-based sample supportsa CpG island methylator phenotype in colon cancer. Gastroenterology 129(3):837–845.

20. Weisenberger DJ, et al. (2006) CpG island methylator phenotype underlies sporadicmicrosatellite instability and is tightly associated with BRAF mutation in colorectalcancer. Nat Genet 38(7):787–793.

21. Ang PW, et al. (2010) Comprehensive profiling of DNA methylation in colorectalcancer reveals subgroups with distinct clinicopathological and molecular features.BMC Cancer 10:227–234.

22. Martens EA, Kostadinov R, Maley CC, Hallatschek O (2011) Spatial structure increasesthe waiting time for cancer. New J Phys 13:13–37.

23. Siegmund KD, Marjoram P, Tavaré S, Shibata D (2009) Many colorectal cancers are“flat” clonal expansions. Cell Cycle 8(14):2187–2193.

24. Siegmund K, Marjoram P, Woo Y, Tavaré S, Shibata D (2009) Inferring clonalexpansion and cancer stem cell dynamics from DNAmethylation patterns in colorectalcancers. Proc Natl Acad Sci USA 106(12):4828–4833.

25. Siegmund KD, Marjoram P, Tavaré S, Shibata D (2011) High DNA methylation patternintratumoral diversity implies weak selection in many human colorectal cancers. PLoSONE 6(6):e21657.

26. Bozic I, et al. (2010) Accumulation of driver and passenger mutations during tumorprogression. Proc Natl Acad Sci USA 107:18545–18550.

27. Eldredge N, Gould SJ (1972) Punctuated equilibria: An alternative to phyleticgradualism. Models in Paleobiology, ed Thomas JM (Freeman, Cooper and Co, SanFrancisco), pp 82–115.

28. Maley CC, et al. (2004) Selectively advantageous mutations and hitchhikers inneoplasms: p16 lesions are selected in Barrett’s esophagus. Cancer Res 64(10):3414–3427.

29. Fearon ER, Vogelstein B (1990) A genetic model for colorectal tumorigenesis. Cell61(5):759–767.

30. Leonard GD, Swain SM (2004) Ductal carcinoma in situ, complexities and challenges.J Natl Cancer Inst 96(12):906–920.

31. Jin Y, Sun A, Noriki S, Imamura Y, Fukuda M (2007) Detection of cancer clones inhuman colorectal adenoma as revealed by increased DNA instability and other bio-markers. Eur J Histochem 51(1):1–10.

32. Leggett BA, et al. (2001) Hyperplastic polyposis: Association with colorectal cancer.Am J Surg Pathol 25(2):177–184.

33. Ahn HS, et al. (2012) Hyperplastic Polyposis Syndrome Identified with a BRAFMutation. Gut Liver 6(2):280–283.

34. Iino H, et al. (1999) DNA microsatellite instability in hyperplastic polyps, serratedadenomas, and mixed polyps: A mild mutator pathway for colorectal cancer? J ClinPathol 52(1):5–9.

35. Woo Y-J, Siegmund KD, Tavaré S, Shibata D (2009) Older individuals appear toacquire mitotically older colorectal cancers. J Pathol 217(4):483–488.

36. Lamlum H, et al. (2000) APC mutations are sufficient for the growth of earlycolorectal adenomas. Proc Natl Acad Sci USA 97(5):2225–2228.

37. Feng Y, et al. (2011) Mutant KRAS promotes hyperplasia and alters differentiation inthe colon epithelium but does not expand the presumptive stem cell pool.Gastroenterology 141(3):1003–1013.

38. Sansom OJ, et al. (2006) Loss of Apc allows phenotypic manifestation of thetransforming properties of an endogenous K-ras oncogene in vivo. Proc Natl Acad SciUSA 103(38):14122–14127.

39. Navin N, et al. (2011) Tumour evolution inferred by single-cell sequencing. Nature472(7341):90–94.

40. Nik-Zainal S, et al.; Breast Cancer Working Group of the International Cancer GenomeConsortium (2012) The life history of 21 breast cancers. Cell 149(5):994–1007.

41. Gerlinger M, et al. (2012) Intratumor heterogeneity and branched evolution revealedby multiregion sequencing. N Engl J Med 366(10):883–892.

42. Xu X, et al. (2012) Single-cell exome sequencing reveals single-nucleotide mutationcharacteristics of a kidney tumor. Cell 148(5):886–895.

43. Campbell PJ, et al. (2008) Subclonal phylogenetic structures in cancer revealed byultra-deep sequencing. Proc Natl Acad Sci USA 105(35):13081–13086.

44. Brodie CM, Crotty PL, Gaffney EF (2004) Morphologically distinct patterns ofapoptosis correlate with size and high-grade dysplasia in colonic adenomas.Histopathology 44(3):240–246.

45. Dalerba P, et al. (2011) Single-cell dissection of transcriptional heterogeneity inhuman colon tumors. Nat Biotechnol 29(12):1120–1127.

46. Merlos-Suárez A, et al. (2011) The intestinal stem cell signature identifies colorectalcancer stem cells and predicts disease relapse. Cell Stem Cell 8(5):511–524.

47. Kemper K, et al. (2012) Monoclonal antibodies against Lgr5 identify human colorectalcancer stem cells. Stem Cells 30(11):2378–2386.

48. Lopez-Garcia C, Klein AM, Simons BD, Winton DJ (2010) Intestinal stem cell replacementfollows a pattern of neutral drift. Science 330(6005):822–825.

49. Tsao JL, et al. (1998) Tracing cell fates in human colorectal tumors from somaticmicrosatellite mutations: Evidence of adenomas with stem cell architecture. Am JPathol 153(4):1189–1200.

50. Poulsom R, Longcroft JM, Jeffery RE, Rogers LA, Steel JH (1998) A robust method forisotopic riboprobe in situ hybridisation to localise mRNAs in routine pathologyspecimens. Eur J Histochem 42(2):121–132.

51. Galandiuk S, et al. (2012) Field cancerization in the intestinal epithelium of patientswith Crohn’s ileocolitis. Gastroenterology 142(4):855–864.

Humphries et al. PNAS | Published online June 13, 2013 | E2499

MED

ICALSC

IENCE

SPN

ASPL

US

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 8,

202

1