REVIEWS

Clonal evolution of colorectal cancer

in IBD
Chang‑Ho R. Choi1,2*, Ibrahim Al Bakir1,2*, Ailsa L. Hart2 and Trevor A. Graham1
Abstract | Optimizing the management of colorectal cancer (CRC) risk in IBD requires a
fundamental understanding of the evolutionary process underpinning tumorigenesis. In IBD,
clonal evolution begins long before the development of overt neoplasia, and is probably
accelerated by the repeated cycles of epithelial wounding and repair that are characteristic
of the condition. Here, we review the biological drivers of mutant clone selection in IBD with
particular reference to the unique histological architecture of the intestinal epithelium coupled
with the inflammatory microenvironment in IBD, and the unique mutation patterns seen in
IBD-driven neoplasia when compared with sporadic adenomas and CRC. How these data can
be leveraged as evolutionary-based biomarkers to predict cancer risk is discussed, as well as how
the efficacy of CRC surveillance programmes and the management of dysplasia can be improved.
From a research perspective, the longitudinal surveillance of patients with IBD provides
an under-exploited opportunity to investigate the biology of the human gastrointestinal tract
over space and time.

1
Evolution and Cancer
Laboratory, Barts Cancer
Institute, Queen Mary
University of London,
Charterhouse Square,
London EC1M 6BQ, UK.
2
Inflammatory Bowel Disease
Unit, Level 4 St Mark’s
Hospital, Watford Road,
London HA1 3UJ, UK.
pacoblue@gmail.com;
i.albakir@qmul.ac.uk;
ailsa.hart@nhs.net;
t.graham@qmul.ac.uk
*These authors contributed
equally to this work.
doi:10.1038/nrgastro.2017.1
Published online 8 Feb 2017
IBD is a chronic relapsing–remitting disorder of the
gastrointestinal tract, consisting of two main subtypes:
Crohn’s disease and ulcerative colitis. The aetiology of
IBD is multifactorial, with the disease arising follow­
ing incompletely defined environmental triggers
in genet­ically predisposed individuals and resulting in
an a­ berrant immune-driven inflammatory response
towards an altered gut microbiota1.
The incidence of IBD is increasing worldwide,
with a prevalence now approaching 0.5% in the West2.
For patients with long-standing colonic inflammation,
colorectal cancer (CRC) is a recognized and feared
complication. Studies quantifying the increased risk
of CRC in patients with IBD have generated vari­able
results reflecting differences in country of origin,
study population and disease duration3. The first large
meta-­analysis assessing CRC risk in patients with
IBD showed a risk of 2% at 10 years after ulcerative
colitis diagnosis, 8% at 20 years and 18% at 30 years
after colitis onset 4. However, other studies suggest
that CRC rates in patients with ulcerative colitis might
be ­declining5.
Important clinical differences exist between colitis-­
associated CRC (CA‑CRC) arising in patients with
IBD, compared with sporadic CRC seen in the general
population. CA‑CRCs tend to affect younger patients6
(average age of 50–60 years in IBD7,8 compared with
65–75 years for sporadic CRCs in the general popula­
tion), they are more likely to be found in the proximal
colon in the presence of Crohn’s disease9 or primary
sclerosing cholangitis10, they are more commonly syn­
chronous (15–20% of CA-CRC11,12 compared with 3–5%
of sporadic CRC) and have an increased frequency of
mucinous or signet ring cell histology13.
Patients with long-standing extensive colonic IBD
are enrolled into surveillance programmes that aim
to detect precursor colitis-associated dysplasia, a pre-­
malignant equivalent of sporadic adenomas. In practice,
detecting dysplasia at endoscopy can be challenging.
These lesions are often flat with subtle borders that
are difficult to appreciate in the presence of concomi­
tant inflammation; they might be readily missed in a
background of extensive mucosal scarring and pseudo­
polyposis. Missed dysplasia might be a contributing
­factor in the stubbornly high rates of interval CA‑CRCs
that develop between scheduled endoscopies14.
When dysplasia is detected, patient management
remains fraught with challenges. The diagnosis of IBD-
associated dysplasia is confirmed through histo­logical
analysis with defining features including nuclear atypia,
mucin depletion and an irregular crypt architecture
with loss of basal‑to‑luminal epithelial cell matur­
ation15. Immunohistochemical staining for markers
including p53 and β‑catenin are sometimes used to
complement histological classification16. In practice,
dysplasia ­grading suffers from substantial inter­observer
variability and can be challenging to differentiate from
a regenerating epithelium17. With improved endoscopic

218 | APRIL 2017 | VOLUME 14 www.nature.com/nrgastro

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

Key points In IBD, the clonal evolution of cancer begins long

• Colorectal cancer development in IBD begins many years before the development
of neoplasia because of occult evolution within the inflamed bowel
• The cycles of wounding and repair characteristic of IBD provide a selective pressure
for mutant cells that are able to rapidly heal the mucosa and withstand the
inflammatory insult
• Measuring and modulating the occult evolutionary process offers new avenues for
effectively predicting and preventing colorectal cancer in IBD
• Repositories of IBD surveillance materials offer a surreptitious opportunity to study
in vivo clonal evolution in time and space in humans
imaging, areas of low-grade and indefinite dysplasia
are now detected in up to 10% of surveillance colono­
scopies 18. Some patients with IBD-associated low-
grade dysplasia are offered more intensive surveillance,
whereas others are offered the option of colectomy,
with all the associated risks of major surgery and con­
sequences of life with a stoma. Data from the largest
UK surveillance registry data has demonstrated that
the 10‑year CRC risk from low-grade dysplasia is only
30%11; others have shown that low-grade dysplasia
regresses in at least half of patients19 although the mech­
anism is unclear. Together, these findings indicate that
the current ­primary aim of dysplasia identification in
CRC surveillance is inadequate as a cancer-risk stratifi­
cation tool. A clear clinical need to better understand
the molecular aetiology of CA‑CRC remains, which
would enable the identifi­cation of more efficacious
biomarkers of c­ ancer risk. Recognizing the develop­
ment of CA‑CRC as an evolution­ary process prov­
ides a novel and powerful perspective to understand
this disease, in a manner that can be directly lever­
aged to improve the clinical care of this ­challenging
patient group.
This Review will summarize the evidence support­
ing the presence of extensive inflammation-driven clonal
evolution in IBD, long before the development of any
clinically apparent neoplasia, and contrast this under­
standing with our current knowledge of sporadic CRC
carcinogenesis. Finally, we discuss how knowledge of
this occult evolutionary process can aid prognostication
and ultimately improve the efficacy of IBD-associated
cancer surveillance.
Evidence of clonal evolution in IBD
Occult evolution in colitis
Carcinogenesis is a process of clonal evolution20–22.
Somatic cells acquire mutations that alter their pheno­
type, which might confer them and their progeny a
growth and/or survival advantage within their cur­
rent microenvironment, enabling the cells to persist
and clonally expand. A classic example is mutations
of APC in the context of colorectal adenoma forma­
tion that result in persistent β‑catenin signalling and
increased cellular proliferation23. With the occurrence
of further mutations, additional selection pressures
and clonal expansion could occur. In some cases, this
process eventually results in the development of a
malignant phenotype.
before the development of a true malignancy. Numerous
studies, using a variety of different molecular techniques,
have reported extensive genomic and epi­genomic alter­
ations in morphologically normal intestinal mucosa
(TABLE 1). The surprisingly high mutation burden of the
ostensibly normal IBD epithelium reveals the ‘occult
evolution’ that is occurring in the inflamed bowel.
These mutant clones clearly experience a pos­itive selec­
tive pressure (mutants cells are fitter than non­mutant
cells) as they can expand to fill large areas of the intes­
tinal mucosa24 and, in one extreme case, the entire
colon length25.
Mutant clones frequently bear mutations in key
tumour-suppressor genes including TP53 (encod­
ing p53) and CDKN2A (encoding p16), and in
the proto-­oncogene KRAS 24–27. Remarkably, these
cancer-­associated mutations do not alone cause neo­
plastic growth as the mutations are detected in non-­
neoplastic tissue 24–27. These ‘key’ mutations might
be necessary but insufficient for tumour growth, or
the phenotypic effects of these mutations are crit­
ically modu­lated by epigenetic and/or microenviron­
mental constraints present in non-neoplastic mucosa.
Nevertheless, the clonal expansion of these cancer-­
associated mutants provides a genetic foundation for the
phenomenon of ‘field cancer­ization’ (REF. 28), which is
the preconditioning of a large area of histologically nor­
mal epithelium to the future development of neoplastic
lesions29 (FIG. 1). Field cancerization probably explains
the high frequency of synchronous and metachronous
neoplastic lesions in patients with IBD11,12.
Inflammation accelerates evolution
The mutation burden in the morphologically normal
non-IBD colon is not well characterised; however, the
limited evidence available to date suggests that the nor­
mal mutation rate is less than in patients with IBD30,31
and that widespread clonal expansion of mutant cells
is rare32. Consequently, the inflamed bowel appears
to be a ‘hot bed’ of somatic clonal evolution, but the
­reasons why are currently unknown. Our hypothesis is
that the evolutionary process in IBD is accelerated by
long-standing repeated cycles of epithelial wounding
and repair that are characteristic of IBD. Mechanistically,
IBD-associated inflammation has the potential to
mediate clonal evolution by any, or all, of the following
three mechanisms: generating a mutagenic pressure;
providing a selective advantage to those clones able to
survive a (cytotoxic) inflammatory insult; providing a
selective advantage to those clones able to more rapidly
­repopulate the healing mucosa.
DSS (dextran sodium sulfate)-treated mice offer a
unique opportunity to assess flat and polypoid dys­plasia
in a single animal model; both these lesions develop as
a consequence of the induced chemical colitis. Here, the
severity of inflammation correlates with the propen­
sity to generate IBD-like flat dysplastic lesions; inflam­
mation scores were an order of magnitude higher in
colons containing flat dysplasia compared with those
with polypoid dysplasia33. Moreover, administration of

NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY VOLUME 14 | APRIL 2017 | 219

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Table 1 | Genetic and epigenetic changes detected in noncancerous IBD mucosa

Abnormality found Methods Spatial extent of change Tissue collection time Ref
DNA aneuploidy in progressors only Cytometry Multiple aneuploid clones co-existing At dysplasia or CRC 123
in the same colon; the largest clones
involve at least half the colon length
DNA aneuploidy Cytometry DNA aneuploidy is persistent and Before CRC, longest 110
spreads over time follow-up period was
8 years
TP53 LOH, DNA aneuploidy Cytometry, SSCP In one case almost the entire colonic At dysplasia or CRC 111
mucosa (95 of 100 biopsies) had
aneuploid DNA
DNA aneuploidy in most progressors Cytometry N/A Up to 2.5 years prior 96
prior to developing overt neoplasia to CRC
TP53 mutation, TP53 LOH Cytometry, SSCP TP53 mutations but not LOH found At dysplasia or CRC 64
in nondysplastic mucosa adjacent to
neoplastic areas
KRAS mutation in non-neoplastic tissue IHC, SSCP N/A At CRC 27
Microsatellite instability seen in up to Microsatellite genotyping Randomly selected non-neoplastic At CRC 124
50% of patients with ulcerative colitis mucosa
but not in patients with acute colitis
(ischaemic or infectious)
CNAs detected in all non-neoplastic FISH and CGH Largest clone extended from the At CRC 125
biopsies of progressors transverse colon to the caecum
KRAS and TP53 mutations, Cytometry, SSCP Up to one-third of the colon length At dysplasia or CRC 26
DNA aneuploidy
CNA at target loci on chromosomes 8, FISH Changes seen at both near At high-grade dysplasia 114
11, 17 and 18 (average 10 cm away) and distal or CRC
(average 50 cm away) tissue
CNAs detected much more commonly CGH and microsatellite The largest clone extended at least At CRC 126
than microsatellite instability instability analysis (10 loci) 36 cm from CRC
Increased ESR1, MYOD1, VCAN Methylation Changes seen in non-adjacent tissue At CRC 63
and CDKN2A methylation
Chromosome arm loss in non-neoplastic FISH N/A At CRC 127
tissue distinguishes progressors from
nonprogressors
In progressors, the degree of genomic PCR-based DNA Weak but significant correlation At dysplasia or CRC 128
instability in non-neoplastic tissue is fingerprinting between the degree of genomic
similar to that of neoplastic tissue in instability and distance from the
the same patient, and is greater than in neoplastic lesion
nonprogressors
Polyclonal TP53 mutations are present Genotyping of individual N/A At dysplasia or CRC 129
in regenerative mucosa and low-grade microdissected crypts
dysplasia; become monoclonal in
high-grade dysplasia and cancer
Increased DNA aneuploidy Cytometry Can involve most of the colon and At CRC; some patients 130
rectum had aneuploidy detected
prior to CRC
Increased mutation burden and clonal Hypermutable microsatellite Clonal patches up to 50 cm long At dysplasia or CRC 107
expansion in progressors locus genotyping (28 loci)
Increased RUNX3, APBA1, and COX2 Methylation Changes seen in non-adjacent tissue At CRC 131
methylation
Mutations in APC, TP53, KRAS, Genotyping of individial The largest clone was at least 14 cm, At CRC 24
chromosome 17p LOH microdissected crypts from which 3 spatially distinct
CRCs arose
Increased aneuploidy and increased Cytometry All patients with extensive aneuploidy Mean 9.2‑year patient 132
S‑phase fraction subsequently developed neoplasia follow‑up
The absolute number of regions with CGH, BAC array Clonal field sizes in progressors are At CRC 92
copy gains distinguishes non-neoplastic large, ranging from 49–161 cm
tissue of progessors from nonprogressors
Mutations in CDKN2A, TP53 and KRAS Genotyping of individual Pancolonic (clonal) 4 years prior to CRC 25
microdissected crypts

220 | APRIL 2017 | VOLUME 14 www.nature.com/nrgastro

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

Table 1 (cont.) | Genetic and epigenetic changes detected in noncancerous IBD mucosa
Abnormality found Methods Spatial extent of change Tissue collection time Ref
Clonal expansions are more frequent Microsatellite genotyping The largest clone involved most of the At CRC 133
in progressors than nonprogressors. In (17 loci) bowel length
nonprogressors, clonal expansions are
associated with older age
Aneuploidy of indefinite dysplastic Cytometry N/A 28‑month mean follow‑up 134
mucosa predictive of progression of indefinite dysplasia
The table shows a summary of genomic and epigenomic alterations found in nondysplastic colonic mucosa. These changes are evidence of clonal evolution in IBD
prior to cancer development. BAC, bacterial artificial chromosome; CGH, comparative genomic hybridisation; CNA, copy number alteration; CRC, colorectal
cancer; FISH, fluorescence in situ hybridisation; IHC, immunohistochemistry; LOH, loss of heterozygosity; SSCP, single-strand conformation polymorphism.

5‑aminosalicylates, the mainstay of treatment for ulcer­
ative colits, had a varying effect on the incidence and
progression of dysplastic lesions on the basis of mor­
phology, with fewer flat lesions that were unchanged in
size, but the same number of polypoid lesions that were
reduced in size34.
Patient-based studies confirm that inflammation has
a substantial cumulative role in increasing CRC risk in
IBD, with observational studies consistently demonstrat­
ing that CA‑CRC risk development is closely linked to
the extent35, duration4 and severity of inflammation7.
Histological studies suggest that neoplastic lesions arise
from a field of marked chronic inflammation, telomere
shortening, DNA damage and senescence, with ‘escape
from senescence’ a key step in the transition from low-
grade to high-grade dysplasia36. Thus, a comprehensive
understanding of carcinogenesis in IBD should not be
limited to the study of mutation generation and spread
via the epithelial crypt stem cell niche, but should also
include a concomitant analysis of the inflammatory stro­
mal microenvironment. Therapeutic interventions that
modulate cancer risk in IBD must target these ­aberrant
microenvironmental changes.
In addition to native cells of the epithelial layer,
immune cells and bacteria migrate up to the crypt as part
of IBD pathogenesis, inducing a myriad of inflammatory
cytokines and intracellular signalling pathways1 that are
beyond the scope of this Review37. These pathways act
directly on non-neoplastic and dysplastic epithelial cells,
which can potentially modulate cancer risk. Animal
model studies demonstrate how transcription signal­
ling by NF‑κB — a master regulator of inflammation38
that has a key role in IBD patho­genesis39 — ­promotes
the survival of pre-malignant epithelial clones40. IBD-
mediated inflammation also promotes β‑catenin
stabil­ity through aberrant PI3K–AKT41,42 and NF-κB43
signalling to further enhance canonical Wnt activity.
Epithelial STAT3 signalling is also upregulated in active
IBD44. Mouse models show that, although STAT3 path­
way activation helps amelior­ate the effects of chemical
colitis by reducing epithelial damage and inflam­mation,
it also promotes the survival and progression of pre-­
malignant epithelial clones45. Conceivably, non-cell-­
autonomous effects similar to those identified in other
systems (for example, in a breast cancer model46) might
also be involved. In this same context, the effect of an
altered immune system in IBD on immunosurveil­
lance necessary to limit CRC initi­ation and progression
­ arrants further investigation. For example, CA‑CRCs
w
and dysplastic lesions show a greater infiltration of CD3+
and CD8+ lymphocytes, but substantially less granzyme
B expression than ­sporadic CRCs, suggesting impaired
cytotoxic function47.
Emerging studies in the field of microbiome analy­sis
are revealing the role of the gut microbiota and intestinal
barrier function in tumorigenesis, and animal studies
are beginning to shed some light onto the complex and
dynamic interplay between the altered immune system,
the aberrant gut microbiome and cancer development
in IBD37. In practice, it remains difficult to determine
whether intestinal dysbiosis and mucosal barrier changes
drive mucosal pathology, or represent a secondary con­
sequence of disease. However, these options are not
mutually exclusive. A prominent example comes from
Arthur et al.48, who used Il10‑knockout mice to demon­
strate how chronic inflammation prevents the homeo­
static elimination of cancer-­associated Escherichia coli,
inducing the upregulation of the polyketide synthase
gene responsible for the genotoxic compound coli­
bactin. In turn, E. coli with genotoxic capabilities have
been shown to drive tumorigenesis in this mouse model,
in a manner independent of inflammation severity49.
This effect is probably dependent on a breakdown
in the mucosal barrier50, enabling the direct exposure of
the epithelium to colibactin. Subsequent human stud­
ies confirm that colibactin-equipped E. coli are mark­
edly more prevalent in colonic mucosa from patients
with sporadic CRC and patients with ulcerative colitis
compared with controls (patients with sporadic polyps
or IBS)49,51. Similarly, Fusobacterium have been found
to promote tumorigenesis in animal models52 and are
over-­represented in both sporadic adenomas and CRC53,
as well as in non-neoplastic IBD mucosa54. Finally, simi­
lar changes in colonic mucus affecting IBD, colon polyps
and colon cancer have been described55. A prominent
example is the re‑­expression of the mucin-associated
oncofoetal carbohydrate, sialyl-Tn‑antigen, at higher
rates in patients with ulcerative colitis who subsequently
progress to dysplasia or cancer than those that remain
cancer-free56. This increased expression seems to be
independent of inflammation severity.
Somatic mutations in CA‑CRC
The different evolutionary pressures offer an explan­
ation for the altered pattern of somatic mutations
seen in CA‑CRC compared with sporadic cancers57,58.

NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY VOLUME 14 | APRIL 2017 | 221

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Available data show that CA‑CRCs have increased muta­
tion frequencies of various intracellular and intercellu­
lar signalling molecules57,58. The most n­ otable mutations
include IL16, a gene encoding a chemo­attractant
cytokine that is overexpressed in IBD in an inflamma­
tion-dependent manner59. IL‑16 has a potential role in
directly mediating inflammation60 and so we specu­
late that a gene mutation encoding this protein could
provide a survival advantage in the inflamed bowel.
Another noteworthy mutation is in RADIL, a gene
encoding a modulator of Rho GTPase signalling in cell
migration, which might speculatively provide a selective
advantage in mucosal healing57.
Mutation signature analysis61 of CA-CRCs57 demon­
strate novel trinucleotide context changes not seen in
sporadic CRCs (an over-representation of A>C trans­
versions at AA dinucleotides). The predominance of
C>T transitions at CG dinucleotides in CA‑CRCs is in
keeping with ‘accelerated ageing’ through rapid cell turn­
over62 rather than the accumulation of direct genotoxic
hits caused by inflammation-associated carcinogens.
Indeed, colitis that progresses to high-grade dysplasia
or CA‑CRC demonstrates field changes in ostensibly
normal epithelium in keeping with accelerated age­
ing, including telomere shortening36 and age-related
CpG methylation63.
Although TP53 mutations are found at similar fre­
quencies in both sporadic CRC and CA‑CRC, TP53
alteration is an early event in CA‑CRC that is found
in non-tumour, and even nondysplastic, mucosa64.
Indeed, accumulating evidence suggests a role for TP53
mutations in promoting cell survival and growth65.
Interestingly, the increased frequency and early onset
of TP53 mutations might provide an additional explan­
ation for the fairly high proportion of flat dysplastic and
cancer­ous lesions seen in IBD. One study66 demonstrated
a significant variability in neoplasm morphology that is
dependent on Tp53 mutation status, with flat lesions
associated with Tp53−/− mice treated with DSS (~85%
flat lesions), and polypoid lesions significantly associated
with Tp53+/− and Tp53+/+ DSS-treated mice (~17% flat
lesions; P <0.0001).
Perhaps most notably, CA‑CRCs in humans show a
paucity of APC mutations57,58 compared with sporadic
adenomas and CRC, in which APC mutations are a
common, important and early ‘gatekeeper’ event67. APC
encodes a key regulator of the canonical Wnt pathway
driving β‑catenin signalling. Disruption of normal APC
function is sufficient to constitutively activate this path­
way and initiate sporadic adenomas in the noninflamed
bowel67,68. The lack of APC mutations in colitis raises the
possibility of IBD-generated inflammation-mediated
alterations of underlying Wnt signalling cascades that
render APC mutations dispensable for tumorigenesis.
Indeed, immunostaining studies of colonic epithelium
confirm high levels of Wnt signalling, as demonstrated
by increased nuclear to cytoplasmic β‑catenin levels in
human ulcerative colitis tissue69. The β‑catenin levels are
midway between normal noninflamed epithelium and
established CRC, thereby partially replicating the con­
stitutive canonical Wnt pathway activation of an APC
mutation. Furthermore, wounding of the intestinal epi­
thelium induces a host of canonical and noncanonical
Wnt signalling pathways by mesenchymal cells70.
Aside from these pertinent genomic differences, the
genomes of CA‑CRCs and sporadic CRCs in humans
interestingly show broad similarities. Both bear a high
frequency of TP53 alteration and mutations in the onco­
genes KRAS, BRAF, PIK3CA, CTNNB1 and the tumour
suppressors SMAD4 and FBXW7 are at relatively high,
albeit slightly different, frequencies57,58,71. Microsatellite
instability is also common to a subset of both cancer
types72,73. Together, these data imply some convergence
in the outcome of evolutionary processes of sporadic
CRC and CA‑CRC development. Dysregulation of key
pathways such as Wnt signalling and p53 function seem
to be necessary for tumour formation in the colon,
irrespec­tive of the presence or absence of a (prior)
inflammatory stimulus. Studies of clonal evolution at
the level of the crypt are discussed in the next section,
advancing our understanding of the survival advan­
tage these genomic changes provide, specifically in the
­context of IBD-mediated inflammation.
Mechanisms of clonal expansion
Colonic crypt architecture
The basic functional unit of the colon is the crypt,
a finger-­like invagination into the underlying lamina
propria lined by a single layer of columnar epithelial
cells. Under homeostatic conditions, the majority of
crypt cells migrate towards the top of the luminal sur­
face over the course of approximately a week, after which
they are shed in the lumen74. Migration out of the crypt
base is associated with differentiation into specialized
cell types, including absorptive enterocytes and sup­
porting secretory cell lineages75. These differentiated
cells are continuously replenished by a small number of
multi­potent stem cells located in the base of each crypt,
known as the stem cell niche76. In the presence of injury
to this actively cycling stem cell pool, cells from further
up the crypt axis fall into the crypt base and regain active
stem cell properties77. Carcinogenic mutations probably
first start accumulating in the long-lived stem cell line­
ages of the crypt, as the lifespan of non-stem cells is
too short for them to acquire the necessary mutations
before being shed. Although LGR5+ crypt base stem
cells are clearly implicated in mutant-APC-driven spor­
adic CRC and familial adenomatous polyposis78, the
precise stem cell(s) of origin in IBD-driven neoplasia
remains undefined.
Stem-cell-level evolution
Clonal expansion in the colon begins with the progeny
of a mutant cell repopulating an entire crypt in a process
referred to as monoclonal conversion of the crypt79. Even
in the absence of any mutant cells, the stem cells in a
crypt are randomly lost and replaced by a neighbour­ing
stem cell lineage in a process called drift. This process
eventually results in extinction of all stem cell popula­
tions except for those originating from the one ‘lucky’
stem cell clone that repopulates the whole crypt, in a
process referred to as niche succession79.

222 | APRIL 2017 | VOLUME 14 www.nature.com/nrgastro

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

Some mutations generated in IBD will bias the com­
petition between stem cells to make it more likely that a
particular mutant stem cell will go on to repopulate the
entire crypt. In vivo lineage tracing studies of intestinal
crypt cells in recombinant mouse models have been used
to measure the advantage of tumorigenic mutations.
Although a wild-type stem cell will replace another wild-
type stem cell 50% of the time, in keeping with neutral
competition80,81, a stem cell with a Kras mutation will
replace a wild-type neighbour stem cell in ~75% of com­
petitive divisions81. Interestingly, Tp53 mutations confer
no substantial competitive advantage in a healthy crypt:
these mutant cells will replace wild-type cells only 50% of
the time81. However, in the context of chemically induced
colitis (via DSS treatment), Tp53 mutation increases the
odds of lineage replacement to 58%81. The differing fate
of Tp53 mutations in an inflamed crypt compared with a
noninflamed crypt reflects the unique selective pressures
that exist in the inflamed bowel.
In the healthy bowel, tight morphogen gradients gen­
erated by cells in and adjacent to the crypt82 restrict the
size of the stem cell niche, keeping stem cell numbers
constant80. The epithelial wounding that is a hallmark
of colitis clearly disrupts these signalling gradients70,82.
Additionally, a study in transgenic mice shows how the
inflammation-mediated NF‑κB pathway can induce
dedifferentiation of non-stem cells, thereby increasing
the size of the pool of long-lived cell lineages that can
contribute to carcinogenesis43. Although the precise
mechanisms of morphogen gradient disruption are yet
to be characterized in human tissue, the altered cellular
composition of the supportive pericryptal fibroblasts in

a Normal colonic mucosa

b Inflamed colonic mucosa

Epithelial cell Stem cell Paneth cell Macrophage Fibroblast Altered microbiota
Mutant epithelial cell Mutant stem cell Mutant Paneth cell Neutrophil Mononuclear cell Ulceration

Figure 1 | Differences in clonal expansion between normal and inflamed Nature

mucosa. The accelerated rate of clonal evolution seen in IBD might be
explained by differences between normal and inflamed mucosa. a | In the
normal colon, a random mutation generated within a crypt stem cell
(first image, in red) results in an advantageous phenotype that eventually
results in the extinction of wild-type stem cells (second image). Eventually,
all cells of that crypt are formed by progeny of this mutant cell (monoclonal
conversion of a crypt). These same mutations often accelerate crypt
fission (third image), which Reviews
is the | Gastroenterology
main mechanism for clonal&expansion
Hepatology in
the intestine (fourth image). b | In IBD, the altered microenvironment
generates epithelial cells with different mutant signatures (red, first image)
compared with the normal colon. Active disease provides a selective
advantage for those mutant cells that can survive an inflammatory insult
(second image). The subsequent healing process selects for those clones
that can more rapidly repopulate the mucosa (third image). In this manner,
IBD accelerates clonal evolution and expansion (fourth image).

NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY VOLUME 14 | APRIL 2017 | 223

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

colitis is a notable example of a niche structure alter­
ation83 that has been associated with neoplastic progres­
sion in IBD84. Understanding the altered morphogen
gradients in IBD is key to understanding how and why
particular mutant cells have an evolutionary advantage
in IBD.
Crypt-level evolution
Once a mutant clone has repopulated an entire crypt,
further clone expansion is possible by crypt fission, the
mechanism by which human colonic crypts divide79.
Thus, the initial cell-level evolution within the crypt
switches to crypt-level evolution and, therefore, natural
selection of the crypt population drives field canceriza­
tion. Crypt fission is the main driver of clonal expan­
sion in the human colon85; indeed, the fission of crypts
surrounding a wound is the primary mechanism of epi­
thelial restitution in the intestine70. Our understanding
of the mechanistic drivers of fission is surprisingly lack­
ing74. 3D imaging studies suggest a critical role for the
relative arrangement of stem cells and their supportive
niche-providing Paneth cells86, and mathematical model­
ling indicates that a plausible mechanism for triggering
fission is exceeding a threshold number of stem cells87.
Crypt fission is a mechanism for mucosal healing70,
and the increase in frequency of crypt-fission events in
the IBD bowel is a reflection of the wounding or repair
that is characteristic of the disease. Cheng and col­
leagues88 found that, although <1 in 200 crypts in nor­
mal human colon are undergoing fission, this rate was
increased at least 60‑fold in IBD. Mutations that increase
the rate of crypt fission are probably positively selected
for in the IBD bowel, as these mutants are better able to
heal the damaged mucosa. In line with this hypothesis,
mutant Kras crypts divide 30‑fold faster than their wild-
type counterparts in the mouse intestine89, and Tp53
mutation is also reported to increase the crypt ­fission
rate in mice90. Thus, it is reasonable to expect that some
mutants which undergo substantial clonal expansion and
populate multiple crypts in the non-neoplastic epithe­
lium achieve this expansion by the acquisition of muta­
tion(s) that positively regulate crypt fission (FIG. 1). This
selection for mucosal healing rather than neoplasia per se
might be the underlying reason for the common occur­
rence of flat dysplastic lesions in IBD. Consequently, in
our opinion, we believe that assessing the inflamed epi­
thelium for clonally expanded mutations offers a ‘natural
experiment’ that can be exploited to identify new genes
that modulate crypt fission. This novel approach recog­
nizes one of the old adages of cancer biology: the overlap
between wound healing and carcinogenesis91.
Other potential mechanisms
One intriguing finding from IBD mapping studies is
the detection of mutant populations containing the
exact same point mutation (single nucleotide variant) or
copy number alteration (CNA), extending across most
of the colon length25,92. Although pancolonic selection
pressures could plausibly select for the same altered cel­
lular or crypt functions repeatedly through convergent
evolution, inducing the exact same point mutation is
highly unlikely and so the very large mutant populations
observed are likely to be clonal in origin. Additionally,
how mucosal repopulation by fission of adjacent crypts
alone could explain the same mutation spread across
the entire bowel in a matter of a few years is difficult
to comprehend. These findings raise the possibility
of unidenti­fied healing mechanisms in colitis. Mouse
model studies show that free-floating organoids intro­
duced intra­luminally can successfully engraft on to an
inflamed bowel to generate long-lasting functional crypt
populations, but not in the context of a non­inflamed
colon93. In light of the phenotypic elasticity of multiple
crypt cell populations and the rapid rate of epithelial
shedding in active IBD, this finding raises the intriguing
possibility of long-range stem cell migration following
shedding and distant stem cell re‑engraftment. However,
to date, no evidence exists for this speculative healing
mechanism in human studies or in IBD a­ nimal models.
Conversely, precedence for this process in other model
organisms does exist; long-distance stem cell migra­
tion has been noted to occur between widely separated
­ovarian ­follicles in Drosophila94.
Evolutionary management of cancer risk
Predicting cancer risk in IBD
Great interest exists for the development of cancer
biomarkers in IBD95, with DNA aneuploidy96 and/or
TP53 mutations64,97,98 as the most prominent examples
of  genetic biomarkers. However, extensive within-­
lesion genetic and phenotypic heterogeneity are now
considered hallmarks of both carcinogenesis99 and pre-­
malignant pathology22. Indeed, in the pre-­malignant
disease, Barrett oesophagus, which confers a ~30‑fold
increased risk of developing oesophageal adeno­
carcinoma100, sequencing studies of biopsies from non-­
neoplastic Barrett segments demonstrate cancer-related
mutations in a plethora of different genes. With the
exception of TP53, these mutations occur at similar
frequencies across nondysplastic through to high-grade
dysplasia and even adeno­carcinoma tissues101. In spor­
adic CRC, large-scale sequencing projects reveal only
seven genes that are recurrently mutated in >10%
of patients, meaning that the overwhelming majority of
mutations were unique to small subsets of patients71.
Collectively, these data indicate the improbability that a
single gene or small set of genetic changes can effectively
assay the vast array of different molecular pathways
that lead to cancer102, so it is unlikely that traditional
candidate-­gene biomarker approaches will yield clin­
ically applicable prognostic biomarkers. Nevertheless, as
the cost of sequencing continues to decrease, the option
to cost-effectively sequence a relatively large panel of
genes that encompass all the key pathways involved in
CA‑CRC development (once these have all been identi­
fied) could well be a p ­ ossibility — such an approach
already shows promise in acute myeloid leukaemia103.
A different approach involves looking for robust prog­
nostic biomarkers from measurements of the under­lying
evolutionary process itself. The rationale here is that,
although the evolutionary paths might differ between
patients, limiting the efficacy of candidate-gene-based

224 | APRIL 2017 | VOLUME 14 www.nature.com/nrgastro

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

studies, somatic evolution is always subject to the same
rules and the selective pressures of an inflamed bowel
are similar across patients. Thus, biomarkers based upon
measurements of these ­evolutionary rules could have
universal applicability.
One example of a candidate ‘evolutionary biomarker’
is clonal diversity. In ecological studies, the level of
diversity in a population is a major determinant of its
evolvability104. If there is no diversity, natural selection
cannot operate, whereas diverse populations are likely to
adapt and evolve quickly. Translated to IBD, the clonal
diversity within the bowel is a potential evolutionary
biomarker of cancer risk and, notably, measuring clonal
diversity requires no substantial prior knowledge of the
precise genetic and epigenetic pathways driving tumori­
genesis. To date, no studies have used genetic diversity
measures for CA‑CRC risk, but the principle has pre­
viously been applied to Barrett oesophagus. Genetic
diversity measures were predictors of cancer risk even
after controlling for age, Barrett segment length, TP53
mutation and aneuploidy status105, and were robust to
the molecular assay used to detect diversity106. A second
approach that also reflects evolvability would be detec­
tion of large clonal expansions as a measure of ongoing
evolution. Preliminary studies in patients with ulcer­
ative colitis found progressors (high-grade dysplasia or
cancer) have substantially larger clonal expansions than
nonprogressors, by using 4–5 biopsies (spaced ~20 cm
apart) for the detection of mutant clones with shared
point mutation(s)107. This work highlights a promising
avenue to explore. Additionally, changes over time in
the clonal composition of the epithelium might also be
prognostic, irrespective of what those changes might
be, because they are also an indicator of ongoing evolu­
tion. In Barrett oesophagus, neoplastic initiation seems
to be abrupt, but single nucleotide polymorphism array
studies suggest that considerable discriminating changes
in clonal composition occur 24 months before oesoph­
ageal adenocarcinoma detection108. Finally, analys­ing
changes in immune, mesenchyme and/or microbiota
microenvironmental composition (which form the
basis for future evolution), and/or their modu­lation
by the epithelial cells themselves109, offers yet another
approach. Concomitant genomic–­microenvironmental
evolutionary studies could also help answer a ‘chicken-
and-egg’ question in CA‑CRC: which came first, the
altered microenvironment that predisposes to cancer
risk, or the predisposed clones themselves?
Use of surveillance biopsies
In our opinion, IBD-associated cancer surveillance
is an ideal ‘model system’ in which to study in vivo
inflammation-associated carcinogenesis in humans.
Multiple biopsies are routinely taken at each surveil­
lance colonoscopy and subjected to histological assess­
ment for inflammation severity and endo­s copically
invisible dysplasia. This process creates an extensive
tissue library of archival specimens reflecting the
spatio­temporal clonal evolutionary changes that occur
during carcinogenesis. Thus, this rich resource can be
utilized to better understand the biology of the intestine
and how it is subverted during cancer development.
The spatial distribution of genetic alterations across
surveillance biopsies indicates the order of mutations
that led to the current clone composition: a ‘trunk’ or
‘founder’ mutation present in all biopsies arose before a
‘branch’ mutation that is limited to a subset of biopsies.
A phylogenetic analysis of multiple biopsies at a single
time-point (from an individual endoscopy) provides a
mechanism for inferring the p ­ attern of genetic changes
over space and time.
Several studies have used mapping biopsies to
­analyse the spatiotemporal changes in clonal popula­
tions, including the relationship between CNAs and key
gene mutations. In early studies, analysis of ulcerative
colitis colonoscopic biopsy specimens taken over an
8‑year period using flow cytometry showed that aneu­
ploidy can be detected before the onset of dysplasia in
at least 10% of patients, with CNAs becoming more
widespread over time in a subset of patients110. TP53
gene mutations can also be found in morphologically
normal mucosa and correlate strongly with the presence
of aneuploidy64. TP53 loss of heterozygosity was found
more commonly in dysplasia, suggesting that it might
be a later event111. Longitudinal analysis of three patients
with Crohn’s disease using targeted sequencing of TP53,
KRAS and CDKN2A revealed that some cancer cases
had arisen from a field of mutant clones that expanded
well beyond the cancer resection margin and formed
many years before cancer growth25. Similarly, large-scale
regional clonal expansions were associated with cancer
occurrence in a study that used hypermutable non­
coding polyguanine tracts as neutral lineage markers to
trace clonal expansions across colectomy specimens107.
A genome-wide comparative genomic hybridisation
array study found that CNAs of chromosomal segments
are present in most colons with pancolitis, and are at
increased frequency in those with neoplasia, without left
or right colon bias92. Interestingly, this study reported that
most CNAs (~85%) were unique to a single biopsy (with
substantial differences detected in CNA profiles from
biopsies as little as 2 cm apart) with no shared CNAs
between high-grade dysplasia and cancer, and the non­
dysplastic surrounding mucosa. This finding suggests
that genetic field cancerization does not universally
precede cancer development, and instead some patients
show focal genetic instability. Consequently, the detection
of these different occult evolutionary processes, includ­
ing selective mutant sweeps (in which a clone grows to
cover very large segments of the colon) and pancolonic
genomic instability (in which multiple different mutant
clones arise independently at different location), might
provide an indicator for CA‑CRC development risk.
Optimizing screening intervals
Endoscopic screening intervals (the time between suc­
cessive endoscopies) could potentially be optimized by
an underlying knowledge of the spatiotemporal dynam­
ics of clonal evolution occurring in IBD. The reason is
straightforward: if the minimum time it will take for a
cancer to evolve from the current composition of the
bowel is known, surveillance can be planned accordingly.

NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY VOLUME 14 | APRIL 2017 | 225

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Our current knowledge of the rates of clone growth
in the inflamed bowel is minimal. Basic questions, such
as the duration of time required for a clone to undergo
expansion substantial enough to be detectable with
biopsy, and the time required to establish a mutant
cancerization field, remain unanswered. Moreover,
the rate and consistency of mutation accumulation
is unknown. In our opinion, measuring these funda­
mental ­e volutionary parameters could have direct
clinical relevance.
A practical factor to consider when using IBD sur­
veillance is the fidelity of standard mapping biopsies at
reflecting true clonal composition of the entire colon.
Routine colonoscopic biopsies contain ~200 crypts112.
Owing to the sheer length of the whole colon, even the
most intensive IBD mapping biopsy protocols sample
<1% of the colonic mucosal surface113. Thus, endoscopy-­
based studies might result in considerable clonal popu­
lations being missed, unless clonal patch sizes truly
are often as large as those detected in some patients
with IBD25,92. Understanding the clonal mosaicism of
the bowel could help guide optimal biopsy protocols.
Moreover, as ulcerative colitis by definition starts in
the rectum and extends proximally, we speculate that the
rectum is the site where the most evolution has occurred,
suggesting that the evolutionary dynamics in the distal

a IBD nonprogression

* * * * * *
* No active
* *
inflammation
* * *
* * *
* * *
* * This single mutant clonal
*
population has not changed
* * significantly in size with time *
b IBD progression through clonal sweeps

* * * * * *
* Substantial * *
inflammation
* * *
* * *
* * *
* Dysplastic lesion is resected
* Dysplasia recurs, it shares
* Cancer arises
in a field of
from a field of successive rapidly the same truncal mutations rapid clonal
* expanding mutant clones * as the newly formed cancer * expansions

c IBD progression through clonal mosaicism

* * * * * *
* Substantial * *
inflammation
* * * Highest
Mutation burden
* * *
* * *
* * Two cancers and a dysplastic *
Multiple unrelated clonal lesion arise simultaneously through Lowest
* populations arise and
evolve independently * clonal mosaicism; they do not share
the same truncal mutations * Wild-type

Figure 2 | Different potential manifestations of mutant clonal evolution Nature

in IBD. Colours represent distinct clones and asterisks indicate biopsy sites.
a | In IBD nonprogression, although an occasional localized mutant clone
population might be detected, the overall clonal composition does not
change with time. b | In IBD progression with clonal expansions, a large
clonal population is already present, with a newly formed clone arising from
this field (left panel). Ongoing inflammation generates further successive
clonal expansions (centre panel). A dysplastic lesion in the rectum (white
star) is resected endoscopically. Biopsies of the surrounding mucosa and the
left colon confirm that the mutant clone is extensive. This finding suggests
a substantial risk of developing recurrent neoplasia (a consequence of field
cancerization). With ongoing Reviews | Gastroenterology
inflammation, & Hepatology
a cancer forms more proximally
(right panel) and a proctocolectomy might be considered. Here, large-scale
and/or rapid clonal expansion of cancer-associated mutations could be a
biomarker of high cancer risk. c | In IBD progression through clonal
mosaicism, multiple distinct clonal populations develop and progress
throughout the colon. The end result is two cancers and a dysplastic region,
all arising in the left colon, without sharing the same ‘truncal’ genomic
changes. In the absence of large clonal fields, a diagnosis of a ‘dangerous’
clone is challenging. However, high levels of clonal mosaicism and
substantial changes over time indicate ‘active evolution’ that could be
utilized as a biomarker of cancer risk.

226 | APRIL 2017 | VOLUME 14 www.nature.com/nrgastro

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

colon might have the most relevance for disease aetio­
logy. Indeed, preliminary evidence raises the possibility
that chromosomal instability of the rectum is a proxy
marker of CRC risk for the whole large intestine114.
Nevertheless, this approach might not apply for patients
with Crohn’s disease and so‑called skip-­lesions, or for
patients with primary sclerosing cholan­gitis and with a
propensity to develop right-sided colon cancer owing
to possible aberrant enterohepatic biliary circulation115.
Assessing the evolutionary dynamics of the colon
would enable us to define the window of opportunity
for cancer risk prognostication, rationalize colonoscopic
biopsy protocols (random versus segmental versus endo­
scopic areas of concern) and determine the appropriate
time intervals between endoscopies (FIG. 2). To ensure
clinical utility, such a biomarker should aim for maximal
sensitivity to avoid missing cancer cases.
Managing neoplastic lesions
Managing dysplasia remains the most challenging clin­
ical dilemma in IBD surveillance. Differentiating those
patients requiring immediate colectomy from those who
can be safely managed conservatively remains difficult.
Consensus guidelines based on dysplasia morphology
and endoscopic resectability aim to standardize clinical
approaches116. Although the majority of patients with
low-grade dysplasia will never develop cancer, 25% of
patients with low-grade dysplasia who are advised to
undergo colectomy will already have an established
­cancer in their surgical specimen11. Even when a ­cancer
is detected at colonoscopy, it is not uncommon for
patients to request a segmental resection rather than
proctocolectomy, to avoid living with a stoma and/or
ileo­anal pouch. Such requests go against standard clin­
ical advice; at least one-third of patients undergoing
colectomy for CA‑CRC detected at surveillance will have
a second s­ ynchronous neoplastic region11.
Dysplastic lesions offer an ideal opportunity to assess
the utility of evolutionary biomarkers by localizing the
site of a mutant cancerization field (FIG. 2). These lesions
and the surrounding mucosa can then be assessed for
markers of somatic evolution, such as the local burden of
genomic alteration (or mutational burden). The sizes
of clonally expanded patches and the degree of clonal
diversity might enable identification of high-risk lesions
that are likely to progress to CRC. This application is
particularly useful for patients with recurrent, invisible
or nonresectable low-grade dysplasia, whereas evolu­
tionary markers associated with a high risk of cancer
development could be used as an indication for colec­
tomy. Furthermore, analysing the remaining mucosa to
determine the spatial extent of these high-risk changes
might enable identification of a subset of patients with
limited cancerization field changes, who could be offered
a segmental colectomy that preserves anorectal function.
Optimizing treatment to prevent cancer
With inflammation a key driver of CA‑CRC risk in clin­
ical and laboratory studies, assessing how therapy modu­
lates occult evolution provides a means to understanding
how and why drug treatments affect cancer develop­
ment. Aspirin reduces both sporadic CRC incidence117
and precursor adenoma recurrence118, and evidence from
Barrett oesophagus shows that NSAID use slows the
rate of somatic mutation accumulation119, suggesting an
evolutionary mechanism of cancer prevention. Together,
these data imply that attenuation of inflammation can
have profound influence on the carcinogenic process.
Nevertheless, although drug treatment in patients with
IBD might modulate clonal evolution of the inflamed
bowel120, similar studies assessing evolution patterns
before and after therapy are needed to define and
rational­ize the most optimal medical intervention strat­
egy. As a clear majority of patients undergoing c­ ancer
surveillance show some evidence of active inflammation
on biopsy121, these studies might generate greater clinical
impetus towards achieving deep remission122 in high-risk
patients, rather than symptomatic relief alone.
Conclusion
Occult evolution is a feature of IBD-associated carcino­
genesis and begins long before the development
of clinically detectable neoplasia. Repeated episodes of
inflammation generate mutant clones and select for
those cells best adapted to this microenvironment, such
as clones resistant to apoptosis and those with acceler­ated
growth. These selective pressures, unique to IBD, might
explain the unique characteristics of IBD-associated
neoplasia in comparison to its sporadic counter­part.
Assaying this evolutionary process is an exciting new
opportunity for developing a new class of biomarkers
that are robust to the stochastic nature of tumorigenic
mutations, which has hampered the effective­ness of
traditional candidate-gene biomarkers. For clinicians,
such a marker would not only aid in cancer risk stratifi­
cation, but might even identify those patients requir­
ing surgery for dysplasia and/or cancer who would be
candidates for segmental resection rather than procto­
colectomy, thereby preserving some colo­rectal function.
For cancer biologists, the IBD surveillance protocol is
an ­u nderutilised model for ­u nderstanding human
intestinal biology.

1. Kaser, A., Zeissig, S. & Blumberg, R. S. Inflammatory
bowel disease. Annu. Rev. Immunol. 28, 573–621
(2010).
2. Kaplan, G. G. The global burden of IBD: from 2015
to 2025. Nat. Rev. Gastroenterol. Hepatol. 12,
720–727 (2015).
3. Adami, H. et al. The continuing uncertainty about
cancer risk in inflammatory bowel disease. Gut 65,
889–893 (2016).
4. Eaden, J. A., Abrams, K. R. & Mayberry, J. F.
The risk of colorectal cancer in ulcerative colitis:
a meta-analysis. Gut 48, 526–535 (2001).
5. Castaño-Milla, C., Chaparro, M. & Gisbert, J. P.
Systematic review with meta-analysis: the declining
risk of colorectal cancer in ulcerative colitis.
Aliment. Pharmacol. Ther. 39, 645–659 (2014).
6. Jess, T., Rungoe, C. & Peyrin-Biroulet, L. Risk of
colorectal cancer in patients with ulcerative colitis:
a meta-analysis of population-based cohort studies.
Clin. Gastroenterol. Hepatol. 10, 639–645 (2012).
7. Rutter, M. D. et al. Thirty-year analysis of a
colonoscopic surveillance program for neoplasia in
ulcerative colitis. Gastroenterology 130, 1030–1038
(2006).
8. Baars, J. E. et al. Age at diagnosis of inflammatory
bowel disease influences early development of
colorectal cancer in inflammatory bowel disease
patients: a nationwide, long-term survey.
J. Gastroenterol. 47, 1308–1322 (2012).
9. Bansal, P. & Sonnenberg, A. Risk factors of
colorectal cancer in inflammatory bowel disease.
Am. J. Gastroenterol. 91, 44–48 (1996).
10. Mattar, M. C., Lough, D., Pishvaian, M. J.
& Charabaty, A. Current management of
inflammatory bowel disease and colorectal cancer.
Gastrointest. Cancer Res. 4, 53–61 (2011).

NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY VOLUME 14 | APRIL 2017 | 227

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

11. Choi, C.‑H. R. et al. Forty-year analysis of colonoscopic
surveillance program for neoplasia in ulcerative colitis:
an updated overview. Am. J. Gastroenterol. 110,
1022–1034 (2015).
12. Lam, A. K.‑Y., Chan, S. S.‑Y. & Leung, M. Synchronous
colorectal cancer: clinical, pathological and molecular
implications. World J. Gastroenterol. 20, 6815–6820
(2014).
13. Itzkowitz, S. H. & Yio, X. Inflammation and cancer IV.
Colorectal cancer in inflammatory bowel disease:
the role of inflammation. Am. J. Physiol. Gastrointest.
Liver Physiol. 287, G7–G17 (2004).
14. Mooiweer, E. et al. Incidence of interval colorectal
cancer among inflammatory bowel disease patients
undergoing regular colonoscopic surveillance.
Clin. Gastroenterol. Hepatol. 13, 1656–1661 (2015).
15. Harpaz, N. & Polydorides, A. Colorectal dysplasia
in chronic inflammatory bowel disease: pathology,
clinical implications, and pathogenesis. Arch. Pathol.
Lab. Med. 134, 876–895 (2010).
16. van Schaik, F. D. M. et al. Endoscopic and pathological
aspects of colitis-associated dysplasia. Nat. Rev.
Gastroenterol. Hepatol. 6, 671–678 (2009).
17. DeRoche, T. C., Xiao, S.‑Y. & Liu, X. Histological
evaluation in ulcerative colitis. Gastroenterol. Rep. 2,
178–192 (2014).
18. Mooiweer, E. et al. Chromoendoscopy for surveillance
in inflammatory bowel disease does not increase
neoplasia detection compared with conventional
colonoscopy with random biopsies: results from
a large retrospective study. Am. J. Gastroenterol.
110, 1014–1021 (2015).
19. Zisman, T. L. et al. Prospective study of the
progression of low-grade dysplasia in ulcerative colitis
using current cancer surveillance guidelines.
Inflamm. Bowel Dis. 18, 2240–2246 (2012).
20. Nowell, P. C. The clonal evolution of tumor cell
populations. Science 194, 23–28 (1976).
21. Merlo, L. M. F., Pepper, J. W., Reid, B. J. & Maley, C. C.
Cancer as an evolutionary and ecological process.
Nat. Rev. Cancer 6, 924–935 (2006).
22. Greaves, M. & Maley, C. C. Clonal evolution in cancer.
Nature 481, 306–313 (2012).
23. Sansom, O. J. et al. Loss of Apc in vivo immediately
perturbs Wnt signaling, differentiation, and migration.
Genes Dev. 18, 1385–1390 (2004).
24. Leedham, S. J. et al. Clonality, founder mutations,
and field cancerization in human ulcerative
colitis–associated neoplasia. Gastroenterology
136, 542–550.e6 (2009).
25. Galandiuk, S. et al. Field cancerization in the
intestinal epithelium of patients with Crohn’s
ileocolitis. Gastroenterology 142, 855–864.e8
(2012).
26. Holzmann, K. et al. Comparative analysis of histology,
DNA content, 53 and Ki‑ras mutations in colectomy
specimens with long-standing ulcerative colitis.
Int. J. Cancer 76, 1–6 (1998).
27. Chaubert, P., Benhattar, J., Saraga, E. & Costa, J.
K‑Ras mutations and p53 alterations in neoplastic
and nonneoplastic lesions associated with longstanding
ulcerative colitis. Am. J. Pathol. 144, 767–775 (1994).
28. Slaughter, D. P., Southwick, H. W. & Smejkal, W.
‘Field cancerization’ in oral stratified squamous
epithelium. Clinical implications of multicentric origin.
Cancer 6, 963–968 (1953).
29. Braakhuis, B. J. M., Tabor, M. P., Kummer, J. A.,
Leemans, C. R. & Brakenhoff, R. H. A genetic
explanation of Slaughter’s concept of field
cancerization: evidence and clinical implications.
Cancer Res. 63, 1727–1730 (2003).
30. Zhu, D. et al. K‑Ras gene mutations in normal
colorectal tissues from K‑ras mutation-positive
colorectal cancer patients. Cancer Res. 57,
2485–2492 (1997).
31. Hsieh, J. C. F., Van Den Berg, D., Kang, H.,
Hsieh, C.‑L. L. & Lieber, M. R. Large chromosome
deletions, duplications, and gene conversion events
accumulate with age in normal human colon crypts.
Aging Cell 12, 269–279 (2013).
32. Nosho, K. et al. A prospective cohort study shows
unique epigenetic, genetic, and prognostic features
of synchronous colorectal cancers. Gastroenterology
137, 1609–1620.e3 (2009).
33. Cooper, H. S., Murthy, S., Kido, K., Yoshitake, H.
& Flanigan, A. Dysplasia and cancer in the dextran
sulfate sodium mouse colitis model. Relevance to
colitis-associated neoplasia in the human: a study
of histopathology, B‑catenin and p53 expression
and the role of inflammation. Carcinogenesis 21,
757–768 (2000).
34. Clapper, M. L. et al. 5‑aminosalicylic acid inhibits
colitis-associated colorectal dysplasias in the mouse
model of azoxymethane/dextran sulfate sodium-
induced colitis. Inflamm. Bowel Dis. 14, 1341–1347
(2008).
35. Ekbom, A., Helmick, C., Zack, M. & Adami, H. O.
Ulcerative colitis and colorectal cancer. A population-
based study. N. Engl. J. Med. 323, 1228–1233
(1990).
36. Risques, R. A. et al. Ulcerative colitis-associated
colorectal cancer arises in a field of short telomeres,
senescence, and inflammation. Cancer Res. 71,
1669–1679 (2011).
37. Grivennikov, S. I. Inflammation and colorectal cancer:
colitis-associated neoplasia. Semin. Immunopathol.
35, 229–244 (2013).
38. Ghosh, S. & Hayden, M. S. New regulators of
NF‑kappaB in inflammation. Nat. Rev. Immunol. 8,
837–848 (2008).
39. Atreya, I., Atreya, R. & Neurath, M. F. NF‑κB in
inflammatory bowel disease. J. Intern. Med. 263,
591–596 (2008).
40. Greten, F. R. et al. IKKβ links inflammation and
tumorigenesis in a mouse model of colitis-associated
cancer. Cell 118, 285–296 (2004).
41. Brown, J. B. et al. Mesalamine inhibits epithelial
beta‑catenin activation in chronic ulcerative colitis.
Gastroenterology 138, 595–605.e3 (2010).
42. Lee, G. et al. Phosphoinositide 3‑kinase signaling
mediates B‑catenin activation in intestinal epithelial
stem and progenitor cells in colitis. 139, 869–881
(2010).
43. Schwitalla, S. et al. Intestinal tumorigenesis initiated
by dedifferentiation and acquisition of stem-cell-like
properties. Cell 152, 25–38 (2013).
44. Mudter, J. et al. Activation pattern of signal
transducers and activators of transcription (STAT)
factors in inflammatory bowel diseases.
Am. J. Gastroenterol. 100, 64–72 (2005).
45. Bollrath, J. et al. gp130‑mediated Stat3 activation
in enterocytes regulates cell survival and cell-cycle
progression during colitis-associated tumorigenesis.
Cancer Cell 15, 91–102 (2009).
46. Marusyk, A., Tabassum, D. P., Altrock, P. M.,
Almendro, V. & Michor, F. Non-cell-autonomous
driving of tumour growth supports sub-clonal
heterogeneity. Nature 514, 54–58 (2014).
47. Michael-Robinson, J. M. et al. Characterization
of tumour-infiltrating lymphocytes and apoptosis
in colitis-associated neoplasia: comparison with
sporadic colorectal cancer. J. Pathol. 208, 381–387
(2006).
48. Arthur, J. C. et al. Microbial genomic analysis reveals
the essential role of inflammation in bacteria-induced
colorectal cancer. Nat. Commun. 5, 4724 (2014).
49. Arthur, J. C. et al. Intestinal inflammation targets
cancer-inducing activity of the microbiota. Science
338, 120–123 (2012).
50. Schwerbrock, N. M. J. et al. Interleukin 10‑deficient
mice exhibit defective colonic Muc2 synthesis before
and after induction of colitis by commensal bacteria.
Inflamm. Bowel Dis. 10, 811–823 (2004).
51. Prorok-Hamon, M. et al. Colonic mucosa-associated
diffusely adherent afaC+ Escherichia coli expressing
lpfA and pks are increased in inflammatory bowel
disease and colon cancer. Gut 63, 761–770 (2014).
52. Kostic, A. D., Xavier, R. J. & Gevers, D. The
microbiome in inflammatory bowel disease: current
status and the future ahead. Gastroenterology 146,
1489–1499 (2014).
53. McCoy, A. N. et al. Fusobacterium is associated with
colorectal adenomas. PLoS ONE 8, e53653 (2013).
54. Strauss, J. et al. Invasive potential of gut mucosa-
derived Fusobacterium nucleatum positively
correlates with IBD status of the host. Inflamm. Bowel
Dis. 17, 1971–1978 (2011).
55. Rhodes, J. M., Campbell, B. J. & Yu, L.‑G.
Lectin‑epithelial interactions in the human colon.
Biochem. Soc. Trans. 36, 1482–1486 (2008).
56. Itzkowitz, S. H. et al. Sialosyl‑Tn antigen is prevalent
and precedes dysplasia in ulcerative colitis:
a retrospective case-control study. Gastroenterology
110, 694–704 (1996).
57. Robles, A. I. et al. Whole-exome sequencing analyses
of inflammatory bowel disease-associated colorectal
cancers. Gastroenterology 150, 931–943 (2016).
58. Yaeger, R. et al. Genomic alterations observed in
colitis-associated cancers are distinct from those found
in sporadic colorectal cancers and vary by type of
inflammatory bowel disease. Gastroenterology 151,
278–287.e6 (2016).
59. Seegert, D. et al. Increased expression of IL‑16
in inflammatory bowel disease. Gut 48, 326–332
(2001).
60. Wang, P. et al. IL‑16 induces intestinal inflammation
via PepT1 upregulation in a pufferfish model:
new insights into the molecular mechanism of
inflammatory bowel disease. J. Immunol. 191,
1413–1427 (2013).
61. Alexandrov, L. B. et al. Signatures of mutational
processes in human cancer. Nature 500, 415–421
(2013).
62. Alexandrov, L. B. et al. Clock-like mutational processes
in human somatic cells. Nat. Genet. 47, 1402–1407
(2015).
63. Issa, J. P., Ahuja, N., Toyota, M., Bronner, M. P.
& Brentnall, T. A. Accelerated age-related CpG
island methylation in ulcerative colitis. Cancer Res. 61,
3573–3577 (2001).
64. Brentnall, T. et al. Mutations in the p53 gene: an early
marker of neoplastic progression in ulcerative colitis.
Gastroenterology 107, 369–378 (1994).
65. Kato, S. et al. Understanding the function-structure
and function-mutation relationships of p53 tumor
suppressor protein by high-resolution missense
mutation analysis. Proc. Natl Acad. Sci. USA 100,
8424–8429 (2003).
66. Chang, W. C. I. et al. Loss of p53 enhances the
induction of colitis-associated neoplasia by dextran
sulfate sodium. Carcinogenesis 28, 2375–2381
(2007).
67. Lamlum, H. et al. APC mutations are sufficient for the
growth of early colorectal adenomas. Proc. Natl Acad.
Sci. USA 97, 2225–2228 (2000).
68. Dow, L. E. et al. Apc restoration promotes
cellular differentiation and reestablishes crypt
homeostasis in colorectal cancer. Cell 161,
1539–1552 (2015).
69. Shenoy, A. K. et al. Transition from colitis to cancer:
high Wnt activity sustains the tumor-initiating
potential of colon cancer stem cell precursors.
Cancer Res. 72, 5091–5100 (2012).
70. Miyoshi, H. et al. Wnt5a potentiates TGF‑β signaling
to promote colonic crypt regeneration after tissue
injury. Science 338, 108–113 (2012).
71. The Cancer Genome Atlas Network. Comprehensive
molecular characterization of human colon and rectal
cancer. Nature 487, 330–337 (2012).
72. Schulmann, K. et al. Molecular phenotype of
inflammatory bowel disease-associated neoplasms
with microsatellite instability. Gastroenterology 129,
74–85 (2005).
73. van Dieren, J. M. et al. Chromosomal and
microsatellite instability of adenocarcinomas
and dysplastic lesions (DALM) in ulcerative colitis.
Diagn. Mol. Pathol. 15, 216–222 (2006).
74. Vermeulen, L. & Snippert, H. J. Stem cell dynamics
in homeostasis and cancer of the intestine. Nat. Rev.
Cancer 14, 468–480 (2014).
75. Barker, N. et al. Identification of stem cells in small
intestine and colon by marker gene Lgr5. Nature 449,
1003–1007 (2007).
76. Ritsma, L. et al. Intestinal crypt homeostasis revealed
at single-stem-cell level by in vivo live imaging. Nature
507, 362–365 (2014).
77. Tian, H. et al. A reserve stem cell population in
small intestine renders Lgr5‑positive cells dispensable.
Nature 478, 255–259 (2011).
78. Barker, N. et al. Crypt stem cells as the
cells‑of‑origin of intestinal cancer. Nature 457,
608–611 (2009).
79. Humphries, A. & Wright, N. A. Colonic crypt
organization and tumorigenesis. Nat. Rev. Cancer
8, 415–424 (2008).
80. Snippert, H. J. et al. Intestinal crypt homeostasis
results from neutral competition between
symmetrically dividing Lgr5 stem cells. Cell 143,
134–144 (2010).
81. Vermeulen, L. et al. Defining stem cell dynamics in
models of intestinal tumor initiation. Science 342,
995–998 (2013).
82. Biswas, S. et al. Microenvironmental control of stem
cell fate in intestinal homeostasis and disease.
J. Pathol. 237, 135–145 (2015).
83. Francoeur, C. et al. Degeneration of the pericryptal
myofibroblast sheath by proinflammatory cytokines
in inflammatory bowel diseases. Gastroenterology
136, 268–277.e3 (2009).
84. Yao, T. & Talbot, I. C. The demonstration of pericryptal
fibroblasts in background mucosa and dysplasia
complicating ulcerative colitis. Histopathology 28,
325–331 (1996).

228 | APRIL 2017 | VOLUME 14 www.nature.com/nrgastro

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

85. Greaves, L. C. et al. Mitochondrial DNA mutations are
established in human colonic stem cells, and mutated
clones expand by crypt fission. Proc. Natl Acad.
Sci. USA 103, 714–719 (2006).
86. Langlands, A. J. et al. Paneth cell-rich regions
separated by a cluster of Lgr5+ cells initiate crypt
fission in the intestinal stem cell niche. PLoS Biol. 14,
e1002491 (2016).
87. Loeffler, M., Bratke, T., Paulus, U., Li, Y. Q.
& Potten, C. S. Clonality and life cycles of intestinal
crypts explained by a state dependent stochastic
model of epithelial stem cell organization. J. Theor.
Biol. 186, 41–54 (1997).
88. Cheng, H., Bjerknes, M., Amar, J. & Gardiner, G.
Crypt production in normal and diseased human
colonic epithelium. Anat. Rec. 216, 44–48 (1986).
89. Snippert, H. J., Schepers, A. G., Van Es, J. H.,
Simons, B. D. & Clevers, H. Biased competition
between Lgr5 intestinal stem cells driven by oncogenic
mutation induces clonal expansion. EMBO Rep. 15,
62–69 (2014).
90. Pruitt, S. C., Freeland, A. & Kudla, A. Cell cycle
heterogeneity in the small intestinal crypt and
maintenance of genome integrity. Stem Cells 28,
1250–1259 (2010).
91. Schäfer, M. & Werner, S. Cancer as an overhealing
wound: an old hypothesis revisited. Nat. Rev. Mol. Cell
Biol. 9, 628–638 (2008).
92. Lai, L. A. et al. Pan-colonic field defects are detected
by CGH in the colons of UC patients with dysplasia/
cancer. Cancer Lett. 320, 180–188 (2012).
93. Yui, S. et al. Functional engraftment of colon
epithelium expanded in vitro from a single adult Lgr5+
stem cell. Nat. Med. 18, 618–623 (2012).
94. Nystul, T. & Spradling, A. An epithelial niche in the
Drosophila ovary undergoes long-range stem cell
replacement. Cell Stem Cell 1, 277–285 (2007).
95. Thorsteinsdottir, S., Gudjonsson, T., Nielsen, O. H.,
Vainer, B. & Seidelin, J. B. Pathogenesis and
biomarkers of carcinogenesis in ulcerative colitis.
Nat. Rev. Gastroenterol. Hepatol. 8, 395–404
(2011).
96. Rubin, C. E. et al. DNA aneuploidy in colonic
biopsies predicts future development of dysplasia
in ulcerative colitis. Gastroenterology 103,
1611–1620 (1992).
97. Lashner, B. A., Shapiro, B. D., Husain, A.
& Goldblum, J. R. Evaluation of the usefulness of
testing for p53 mutations in colorectal cancer
surveillance for ulcerative colitis. Am. J. Gastroenterol.
94, 456–462 (1999).
98. Gerrits, M. M. et al. Biomarker-based prediction
of inflammatory bowel disease-related colorectal
cancer: a case-control study. Cell. Oncol. (Dordr.) 34,
107–117 (2011).
99. Marusyk, A., Almendro, V. & Polyak, K. Intra-tumour
heterogeneity: a looking glass for cancer? Nat. Rev.
Cancer 12, 323–334 (2012).
100. Solaymani-Dodaran, M., Logan, R. F. A., West, J.,
Card, T. & Coupland, C. Risk of oesophageal cancer
in Barrett’s oesophagus and gastro-oesophageal
reflux. Gut 53, 1070–1074 (2004).
101. Weaver, J. M. J. et al. Ordering of mutations
in preinvasive disease stages of esophageal
carcinogenesis. Nat. Genet. 46, 837–843 (2014).
102. Ein-Dor, L., Zuk, O. & Domany, E. Thousands of
samples are needed to generate a robust gene list
for predicting outcome in cancer. Proc. Natl Acad.
Sci. USA 103, 5923–5928 (2006).
103. Papaemmanuil, E. et al. Genomic classification and
prognosis in acute myeloid leukemia. N. Engl. J. Med.
374, 2209–2221 (2016).
104. Masel, J. & Trotter, M. V. Robustness and evolvability.
Trends Genet. 26, 406–414 (2010).
105. Maley, C. C. et al. Genetic clonal diversity predicts
progression to esophageal adenocarcinoma.
Nat. Genet. 38, 468–473 (2006).
106. Martinez, P. et al. A dynamic clonal equilibrium
predicts cancer risk in Barrett’s oesophagus.
Nat. Commun. 7, 12158 (2016).
107. Salk, J. J. et al. Clonal expansions in ulcerative colitis
identify patients with neoplasia. Proc. Natl Acad.
Sci. USA 106, 20871–20876 (2009).
108. Li, X. et al. Temporal and spatial evolution of somatic
chromosomal alterations: a case-cohort study of
Barrett’s esophagus. Cancer Prev. Res. (Phila.) 7,
114–127 (2014).
109. Itzkowitz, S. H. et al. Sialosyl‑Tn antigen: initial report
of a new marker of malignant progression in long-
standing ulcerative colitis. Gastroenterology 109,
490–497 (1995).
110. Lofberg, R., Brostrom, O., Karlen, P., Ost, A.
& Tribukait, B. DNA aneuploidy in ulcerative-colitis —
reproducibility, topographic distribution, and relation to
dysplasia. Gastroenterology 102, 1149–1154 (1992).
111. Burmer, G. C. et al. Neoplastic progression in
ulcerative colitis: histology, DNA content, and loss
of a p53 allele. Gastroenterology 103, 1602–1610
(1992).
112. Greaves, L. C. et al. Clonal expansion of early to
mid‑life mitochondrial DNA point mutations drives
mitochondrial dysfunction during human ageing.
PLoS Genet. 10, e1004620 (2014).
113. Shergill, A. K. et al. The role of endoscopy in
inflammatory bowel disease. Gastrointest. Endosc. 81,
1101–1121.e13 (2015).
114. Rabinovitch, P. S. et al. Pancolonic chromosomal
instability precedes dysplasia and cancer in ulcerative
colitis. Cancer Res. 59, 5148–5153 (1999).
115. Torres, J., de Chambrun, G. P., Itzkowitz, S.,
Sachar, D. B. & Colombel, J.‑F. Review article:
colorectal neoplasia in patients with primary
sclerosing cholangitis and inflammatory bowel disease.
Aliment. Pharmacol. Ther. 34, 497–508 (2011).
116. Laine, L. et al. SCENIC international consensus
statement on surveillance and management of
dysplasia in inflammatory bowel disease.
Gastroenterology 148, 639–651.e28 (2015).
117. Algra, A. M. et al. Effects of regular aspirin on long-
term cancer incidence and metastasis: a systematic
comparison of evidence from observational studies
versus randomised trials. Lancet Oncol. 13, 518–527
(2012).
118. Asano, T. K. & McLeod, R. S. Non steroidal
anti‑inflammatory drugs (NSAID) and aspirin for
preventing colorectal adenomas and carcinomas.
Cochrane Database Syst. Rev. 2, CD004079 (2004).
119. Kostadinov, R. L. et al. NSAIDs modulate clonal
evolution in Barrett’s esophagus. PLoS Genet. 9,
e1003553 (2013).
120. Zhao, L.‑N. et al. 5‑aminosalicylates reduce the risk
of colorectal neoplasia in patients with ulcerative
colitis: an updated meta-analysis. PLoS ONE 9,
e94208 (2014).
121. Gupta, R. B. et al. Histologic inflammation is a risk
factor for progression to colorectal neoplasia in
ulcerative colitis: a cohort study. Gastroenterology
133, 1099–1105 (2007).
122. Zallot, C. & Peyrin-Biroulet, L. Deep remission
in inflammatory bowel disease: looking beyond
symptoms. Curr. Gastroenterol. Rep. 15, 315
(2013).
123. Levine, D. S. et al. Distribution of aneuploid cell
populations in ulcerative colitis with dysplasia or
cancer. Gastroenterology 101, 1198–1210 (1991).
124. Brentnall, T. A. et al. Microsatellite instability in
nonneoplastic mucosa from patients with chronic
ulcerative colitis. Cancer Res. 56, 1237–1240
(1996).
125. Willenbucher, R., Zelman, S., Ferrell, L., Moore, D.
& Waldman, F. Chromosomal alterations in ulcerative
colitis-related neoplastic progression.
Gastroenterology 113, 791–801 (1997).
126. Willenbucher, R. F. et al. Genomic instability is
an early event during the progression pathway
of ulcerative-colitis-related neoplasia. Am. J. Pathol.
154, 1825–1830 (1999).
127. O’Sullivan, J. N. et al. Chromosomal instability in
ulcerative colitis is related to telomere shortening.
Nat. Genet. 32, 280–284 (2002).
128. Chen, R. et al. DNA fingerprinting abnormalities can
distinguish ulcerative colitis patients with dysplasia
and cancer from those who are dysplasia/cancer-free.
Am. J. Pathol. 162, 665–672 (2003).
129. Yoshida, T., Mikami, T., Mitomi, H. & Okayasu, I.
Diverse p53 alterations in ulcerative colitis-associated
low-grade dysplasia: full-length gene sequencing
in microdissected single crypts. J. Pathol. 199,
166–175 (2003).
130. Sjöqvist, U. et al. Colorectal cancer in colonic
Crohn’s disease — high frequency of DNA-aneuploidy.
Anticancer Res. 25, 4393–4397 (2005).
131. Garrity-Park, M. M., Loftus, E. V. J., Sandborn, W. J.,
Bryant, S. C. & Smyrk, T. C. Methylation status
of genes in non-neoplastic mucosa from patients
with ulcerative colitis-associated colorectal cancer.
Am. J. Gastroenterol. 105, 1610–1619 (2010).
132. Söderlund, S. et al. Colitis-associated DNA
aneuploidy and dysplasia in Crohn’s disease and
risk of colorectal cancer. Inflamm. Bowel Dis. 17,
1101–1107 (2011).
133. Salk, J. J. et al. Clonal expansions and short telomeres
are associated with neoplasia in early-onset, but not
late-onset, ulcerative colitis. Inflamm. Bowel Dis. 19,
2593–2602 (2013).
134. Choi, W.‑T. et al. Outcome of ‘indefinite for dysplasia’
in inflammatory bowel disease: correlation with
DNA flow cytometry and other risk factors of
colorectal cancer. Hum. Pathol. 46, 939–947 (2015).
Acknowledgements
The authors are grateful for funding from Cancer Research
UK, the Medical Research Council, the Derek Willoughby
Fund for Inflammatory Research, the St Mark’s Hospital
Foundation and Barts Charity.
Author contributions
The authors contributed equally to the review.
Competing interests statement
The authors declare no competing interests.

NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY VOLUME 14 | APRIL 2017 | 229

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.