MSH2 deficiency contributes to accelerated APC-mediated intestinal tumorigenesis.

Abstract

Accelerated intestinal tumorigenesis is probable in hereditary nonpolyposis colorectal cancer, a condition associated with germ line DNA mismatch repair (MMR) gene defects, and is believed to be caused by rapid accumulation of replication errors in critical genes, such as the APC (adenomatous polyposis coli) tumor suppressor gene. To study the potential contribution of MMR genes to accelerated intestinal tumorigenesis, we crossed the Min mouse, heterozygous for a germ line mutation of Apc, with an MMR gene (Msh2)-deficient mouse. MSH2 deficiency resulted in the development of many colonic aberrant crypt foci, as well as reduced survival of the mice, secondary to both a greater number and more rapidly developing adenomas. The mechanism of inactivation of the wild-type Apc allele depended on MSH2 status. In the presence of functional MSH2, all tumors demonstrated loss of heterozygosity. In contrast, whereas all adenomas were APC negative by immunostaining, only 5 of 34 adenomas from Apc+/-/Msh2-/- mice demonstrated loss of heterozygosity of the wild-type Apc allele, suggesting that somatic Apc mutations are responsible for the additional tumors. These findings provide evidence for the important role of MMR genes in accelerated intestinal tumorigenesis, thus supporting more aggressive surveillance strategies to prevent colorectal cancer in hereditary nonpolyposis colorectal cancer.

Full text

Page 1

1996;56:2922-2926. Published online July 1, 1996.Cancer Res


Armin H. Reitmair, Jian-Chun Cai, Matthew Bjerknes, et al.
Intestinal Tumorigenesis
MSH2 Deficiency Contributes to Accelerated APC-mediated



Updated Version
http://cancerres.aacrjournals.org/content/56/13/2922
Access the most recent version of this article at:


Citing Articles
http://cancerres.aacrjournals.org/content/56/13/2922#related-urls
This article has been cited by 33 HighWire-hosted articles. Access the articles at:


E-mail alerts
related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
.pubs@aacr.orgDepartment at
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
. permissions@aacr.orgDepartment at
To request permission to re-use all or part of this article, contact the AACR Publications
American Association for Cancer Research Copyright © 1996
on July 15, 2011 cancerres.aacrjournals.orgDownloaded from

Page 2

ICANCKR RESEARCH 56. 2922-2926.July 1. 1996]
Advances in Brief
MSH2 Deficiency
Tumorigenesis1
Contributesto AcceleratedAPC-mediated Intestinal
Armin H. Reitmair, Jian-Chun
Angela Mitri, Bharati V. Bapat, Tak W. Mak, and Steven Gallinger2
Cai, Matthew Bjerknes, Mark Redston, Hazel Cheng, Molly T. L. Find, Kazy Hay,
Amgen Insilimi-, Ontario Cancer Institute, Departments
Cell Biology.Universityof Toronto. Ontario A/55 IAS ¡M.B.. H. C.I: and Departments
Research Institute ¡M.R.. M. T. L P.. K. H., B. V. B., S. C.]. Mount Sinai Hospital. Toronto. Ontario MSG 1X5. Canada
of Medical Biophysicsunti Immunology.Toronto, Ontario M4X 1K9 [A. H. R., T. W.M.J:
of Pathology[J-C. C.. M. R.. A. M.. B. V. B.] and Surgery ¡S.G.I. Samuel Lunenfeld
Departmentof Anatomyanil
Abstract
Accelerated
polyposis
mismatch
rapid accumulation
(adenomatous
tial contribution
crossedthe Min mouse, heterozygous
with an \l.\IH gene i \/\/i2 i-di'lidrnl
the developmentof many colonie aberrant crypt foci, as well as reduced
survival of the mice, secondaryto both a greater
rapidly developing adenomas. The mechanism
type Ape alÃ-eledepended on MSI 12 status. In the presence of functional
\ISH2,all tumorsdemonstrated
whereas all adenomas were APC negative by immunostaining,
adenomas from Ape+'~/Msh2~'~ mice demonstrated
intestinaltumorigenesis
cancer, a condition
gene defects, and is believed
of replication errors in critical genes, such as the APC
polyposis coli) tumor suppressor
of MMR genes to accelerated
for a germ line mutation
mouse. MSII2 deficiency
is probable
associated
in hereditary
with germ line DNÂ
to be caused by
non-
colorectal
repair i\IMKi
gene. To study the poten
intestinal tumorigenesis, we
of Ape,
resulted in
numberand more
of the wild- of inactivation
loss of heterozygosity.In contrast,
only 5 of 34
loss of heterozygosity
of the wild-type
responsible for the additional tumors. These findings provide evidence for
the importantrole of MMR genes in accelerated
thus supporting more aggressivesurveillance
rectal cancer in hereditary nonpolyposis
Ape alÃ-ele,suggesting that somatic Ape mutations are
intestinal tumorigenesis,
strategies to prevent colo
colorectalcancer.
Introduction
Approximately
two autosomal
5% of cases of colorectal cancer can be attributed to
dominantinheritedsyndromes, FAP3 and HNPCC
(reviewed
APC tumor suppressor gene (2), and most patients with HNPCC have
germ line mutations of one of the MMR genes (3). Both FAP and
HNPCC are characterizedby the development of colorectal adenomas
and carcinoma at an early age. In FAP, and in most sporadic colorectal
cancers, inactivation of the APC gene appears to be critical. APC
mutations are among the earliest
premalignant microscopic ACF (4), whereas it is likely that mutations
of MMR genes, such as MSH2, are important during the apparent rapid
progression of the adenoma-to-carcinoma
Yet, the combinedmolecularcontribution
proteinsto colorectalpolyp development
known. In the present study, we have taken advantage
in Ref. 1). FAP is caused by germ line mutations of the
detectable lesions in potentially
sequence in HNPCC (5, 6).
of the APC and MMR
and progression is not
of mouse
Received 4/5/96; accepted 5/15/96.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement
18 U.S.C. Section 1734 solely to indicate this fact.
1Financial support for this work was provided
Canada (to M. R.. B. B.. T. W. M.. and S.G.)
Canada (to M. B.. H. C.. and T. W. M.). A. H. R. is supported by the German Academic
Exchange Service.
2To whom requests for reprints should be addressed, at Department of Surgery. Mt.
Sinai Hospital. Suite 1225. 600 University Avenue. Toronto. Ontario M5G 1X5. Canada.
1The abbreviations used are: FAP. familial adenomatous
in accordance with
by the National Cancer Institute of
and the MedicalResearchCouncil of
polyposis; HNPCC. hered
itary nonpolyposis
RER. replication error; LOH. loss of heterozygosity.
colorectal cancer; MMR. mismatch repair; ACF, aberrant crypt focus;
models harboring germ line mutations of Ape and Msh2 to investigate
the relationship of APC and MMR genes to intestinal tumorigenesis.
Materials and Methods
Mice. Min mice (Apc*'~; Ref. 7) have been bred on the C57BL/6J strain in
our facility (original breeding pair from The Jackson Laboratory. Bar Harbor, ME)
tor the past 18 months. The generation of our Msh2 knockout mouse (Msh2~'~)
has been described recently (8). Male Min mice were crossed
Ms/i2+'~ mice to generate Apc+'~IMsh2*'~
were then bred with female Msh2*'~
Ape*'' IMshl*'-',andApc+'-lMsh2-'-
with female
mice. Male Ape+'~IMsh2*'~
mice to generate
offspring.
mice
Apc+'~/MsH2+>+,
Enumeration
sacrificed
and flushed
of IntestinalAdenomas
and intestines
solution to remove
and Colonie
were immediately
food and fecal debris.
ACF. Mice were
by cervical dislocation,
with Ringer's
removed
The
intestines were opened longitudinally,
for at least 3 h in 10% neutral buffered formalin. Fixed intestines were stained
with méthylène blue and examinedfor tumors and ACF by gross inspection
and light microscopy, as described (4).
Mathematical Models of Adenoma and ACF Accumulation.
were assumed to result from a Poisson-distributed
mutant clones present at birth and from postnatal somatic mutations that occur
randomly throughout life at rate A. The mutant clones are assumed to behave
independentlyand to require a random period of time to develop into adeno
mas (with cumulative density function F(i) assumed to be a y distribution).
With these assumptions,the number of visible adenomas at time t is Poisson-
distributed with the mean4
laid flat on Whatman paper, and fixed
Adenomas
(with mean i) number of
Note that for simplicity
in this analysis
tumors arise early in life. ACF accumulation
except that it was assumed
adenomas.Then the number of ACF at time t is Poisson-distributed
the mean
i-
. l r(x) d<c where r(t) =
Ja
is the probabilitydensity function for the time spent before a mutant clone
converts into an ACF and F2(t) is the cumulative density function for the time
required for an ACF to develop into a visible adenoma. Similarly, the number
of adenomas expected is Poisson-distributed
we have ignored potential
(9). It turns out to be a good approximation
effects of the crypt cycle
since most
was modeled in the same way,
that ACF. after a random period of time, form
r
f,(x)(\ -
with
J t
with the meanj
LS(/) + A l s(x)d\
where s(t) =lJ
J nJ9
,(r -x)dx
The times required for ACF formation and adenoma conversion
to be exponentiallydistributed.
MicrosatelliteInstability and Ape LOH Assay.
paraffin blocks of normalliver and intestinal
were assumed
DNA was extracted from
adenomasby proteinaseK
4 M. Bjerknes. unpublishedobservation.
2922
American Association for Cancer Research Copyright © 1996
on July 15, 2011 cancerres.aacrjournals.orgDownloaded from

Page 3

ACCELERATEDTUMORIGENESISIN MSH2-DEFICIENTMin MICE
digestion
Microsatellite
bility of amplified normal and tumor DNA using primers from loci on mouse
chromosomes1, 2, 5, 6, and 10, as described (8).
The PCR-based Ape locus quantitation
Luongo et al. (10). Briefly, following digestion with Hindltt, the 144-bp PCR
product from the ApcM<"alÃ-eleand 123-bp product from the Apc+ alÃ-elewere
(8). Care was taken to avoid nonneoplastic
instability was detected by comparison
areas in adenomas.
of electrophoreticmo
assay was performed as described by
resolved on 6% polyacrylamide
samples were amplified in duplicate,
10%. An Ape* a\ìeìeiApcM'"
gels, followed by densitometric
and ratios of pairs differed by less than
alÃ-eleratio of 0.85 is expected in normal tissue
analysis. All
(no LOH) based on the difference in the number of radioactive deoxycytosine
residues in the two alÃ-elesfollowing digestion with HindiU.
Immunostainingof Intestinal Adenomas for the APC COOH Terminus.
Sectionsfrom intestinaladenomaswere incubated
antibodiesspecific for the COOHterminus
Biotechnology,Inc., Santa Cruz, CA) and then stained with Cy3-conjugated
goat anti-rabbit IgG(JacksonImmunoResearch
Grove, PA). Controls, which were negative, included omission of the anti-APC
antibody and competition with the peptide against which the antibodies were
generated(human APC amino acids 2824-2843).
with rabbitpolyclonal
(Santa Cruzof APC protein
Laboratories,Inc., West
Results and Discussion
MMR Deficiency Results in Accelerated Intestinal Tumorigen-
esis in Apc+'~/Msh2~'~ Mice Compared toApc+'~/Msh2+/+ and
Apc+'~/Msh2+'~Mice. Min mice bred in our facility developed
approximately 100 small intestinal adenomas and 5 colorectal adeno
mas by 160-180 days, at which time they became moribund and died
from anemia and intestinal obstruction. Msh2~'~ mice are susceptible
to the development of lymphoma by 2 months of age (8, 11),whereas
Msh2+'~ mice arehealthy up to at least 1year of age. Min mice were
bred with MSH2-deficient mice to examine the effects of MMR on
intestinal adenoma formation. Animals were killed serially beginning
at 10days of age, and all adenomas in small and large intestines were
counted. The number of adenomas were compared in Apc+'~/
Msh2+/+, Apc+'~/Msh2+'~, and Apc+/~/Msh2~'~ mice. Dysplastic
adenomas were present from 10 days in all three genotypes, and the
tumors were histopathologically similar. However, striking numbers
of small and large bowel adenomas soon emerged in the Apc^'~l
Msh2~'~ mice in comparison with the other genotypes. Forexample,
two of three Apc+l~IMsh2~'~ mice had 8 colon adenomas at 27 days,
whereas Apc+'~/Msh2*'~ and Ape +'~IMsh2+/+ mice never devel
oped more than five colon adenomas until at least 70 days of age.
Even more remarkable was the very large number of small bowel
adenomas that arose inApc+/~/Msh2~'~ mice after 47 days. Between
47-78 days, an average of 333 adenomas were scored in 10Apc+'~/
Msh2~'~ mice compared to a mean of 48 and 42 small bowel
adenomas in similar age Apc+'~/Msh2+/+ (n = 9) and Apc+/~l
Msh2+'~ (n = 8) mice, respectively (Fig. \,A andB). No carcinomas
were detected in any of the animals. The increased number of colon
and small bowel adenomas in Apc+'~/Msh2~'~ mice contributed to
B
0 2040 6080100 120 140 160180
age(d)
age(d)
Fig. 1. A. counts of small intestinal adenomas in Apc*'~ IMsh2~'~
of adenoma accumulation, as described in "Materials and Methods." Note the plateau in adenoma accumulation in all genotypes, indicating that adenomas arise early, with little postnatal
contribution. B. colon adenomas in 27-day-old Apc*'~ IMsh2~'~mouse. Macroscopicphotograph of colon with four large and three small easily identifiable colonie tumors (bar. 5
(*). Apc^'~IMsh2*f+ ( 0 ). and Apc^~'~/Msh2+i~(O) mice. Lines represent maximum likelihood fits of a model
mm). Typical lobulated surface contour is visible in larger tumors. The segment of colon involved by tumors is dilated. Histológica! examination
adenomas. C counts of colon adenomas (O) and ACF (+) in Apc+l~IMsh2~'~
revealed that all seven tumors were
to the ACF data only. The mice. The increasing salid line is a fit of a model of ACF accumulation
dolled line represents the number of adenomas expected to derive from ACF. The third line (at the haitom) represents a fit of the adenoma accumulation
Colon adenomaaccumulationplateaus, while ACF show a steady increase, indicating
Apc*'~IMsh2~'~mice. Photomicrograph of méthylène blue-stained colon of 78-day-old mouse (bar. 250 jim). Mucosa is viewed en face, revealing cecal folds. Crypt lumina (arrows)
model to the adenoma data.
found in colons of that most ACF result from postnatal events. £>. ACF are frequently
are identified as small white dots where light is easily transmitted through the colonie wall. A three-crypt ACF is characterized
surface and have thicker and more darkly staining epithelial outlines. £. photomicrograph
epithelial cells of ACF have mucin depletion with stratified, enlarged, and hyperchromatic
by dilated crypts that project slightly from the mucosal
section of ACF (bar. 50 urn). Compared to surrounding nonneoplastic crypts,
nuclei typical of dysplasia/microadenoma.
of H&E-stained
2923
American Association for Cancer Research Copyright © 1996
on July 15, 2011cancerres.aacrjournals.orgDownloaded from

Page 4

ACCELERATEDTUMORIGHNESISIN MSH2-DEFICIENTMin MICE
Table I Maximum iikelihtxxi estimates of parametersfor the adenoma accumulation
model
model of macroscopic adenoma accumulation (see
yielded maximumlikelihood estimates of the average time
The application of the mathematical
"Materials and Methods")
required for a nascent lumor to develop into a macroscopically
model, some nascent tumors arise from mutant clones present at birth, whereas others
arise from postnatalsomatic mutationsthai occur randomly
mutant clones/day. The number of perinatal mutant clones was assumed Poisson-disirih-
uicd with mean i mutant clones. The analysis also provided maximum likelihood estimules
A and t for A and i. respectively.Note that for all genotypes,
tumor initiation was found (A = 0)
visible adenoma. In the
throughout life at rale À
no evidence of postnatal
Avg. time 10
GenotypeadenomaSmall
intestineApc+/-/Msh2-/-
daysApc+/-/Msh2+/-
daysColonApc+/-/Msh2-/-
42
66zndApf+/-/Msh2+/+
21
100daysi
daysAi>c+/-/M5h2+/-
andApc+/-/Msh2+/+
(clones) A(clones/day)350
095
012
05
0
their early death from anemia and bowel obstruction (mean age when
moribund: Apc+'~/Msh2+/+,163 days; Apc+'~/Msh2+/~,
Apc+'~IMsh2~'~,82 days).
159 days;
Having shown that MSH2 deficiency results in the development of
more adenomas in Ape ' 'mice, we sought to understand the dynam
ics of intestinal tumorigenesis
tions. We developed a mathematical
lihood estimates of the timing and frequency of initiating mutations,
as well as estimates of the rate of adenoma growth, in each of the
genotypes. The model assumes that some adenomas are present in
nascent form perinatally (e.g.. clones derived from prenatal somatic
mutations), whereas others arise at random times throughout life. The
analysis revealed three main conclusions that apply to both the large
and small intestine: (a) in all three genotypes, the bulk of adenomas
arise from mutant clones that are present perinatally (few arise post-
natally); (/?) there are about three times more nascent tumors
Apc+'~/Msh2~'~ mice than in the other genotypes;
progression is accelerated in the Ape ' ' IMsh2 '
and the genetic basis of our observa
model to obtain maximum like
in
and (c) tumor
mouse (Fig. \A\
Table I). Therefore,
more perinatal mutant clones, which expand into tumors more rapidly.
The observationthat tumors arise primarily
supports the findings of others (12) but is surprising
deficiencywould be expectedto contribute
replication errors in genes such as Ape, resulting in additional intes
tinal neoplasms during the life ofApc+'fMsh2'
MSH2 deficiencyleads to the development of
perinatallyin mice
since MSH2
postnatalto ongoing
mice. To address
this issue, we used the ACF as an intermediate biomarker to determine
the effects of MSH2 deficiency on postnatal genetic events. ACF have
been proposed as precursors to colonie adenomas in man and carcin
ogen-treatedrodents, based on their high frequency
FAP patients, and by the presence of somatic APC and K-ra.v muta
tions (13, 14). We counted dysplastic ACF in méthylène blue-stained
colons from each of the genotypes. ACF were quite rare in Ape*"1 I
Msh2 ' 'and Ape *~'~IMsh2*~/+mice but were abundant in Apc+/~/
Msh2~'~ mice (Fig. I, C-E). We then developed
in colons from
a more elaborate
mathematical
tween nascent colonie tumors and adenoma and fitted the model to the
ACF data alone (Fig. 1C). We found that for all three genotypes, in
contrast to adenomas, most ACF arise from postnatal events. Further
more,ACF ariseaboutI(X) times
Msh2"'- (2.9 clones/day)than in Apc+'^/Msh2^'
Msh2 ' ' ' mice (0.028 clones/day).
Apc+'~ mice, MSH2 deficiency yields frequent mutant clones (ACF),
model incorporating an intermediateACF stage be
morefrequentlyin Apc+'~/
and Apc+'~l
Thus, it appears that in adult
but these do not contribute significantly
scopic colon adenomas. What limits the progression into adenomas is
unclear. It may simply be a lack of time (mosM/w+/~
to the population of macro
IMsh2~'mice
die at <IOO days), or conditions for rapid clonal expansion may be
most favorable in the perinatal period for hormonal or morphogenetic
reasons.
Infrequent Microsatellite Instability and LOH of Ape in Ade
nomas from Apc+'~/Msh2~'~ Mice. We addressed possible genetic
mechanismsof accelerated intestinal
Msli2~'~mice. Since cancers from HNPCC patients with germ line
tumorigenesis in Ape +l~/
MMR gene mutations are distinguished
microsatelliteinstability (also referred to as the RER+
Rets. 15 and 16), we searched for shifts in electrophoretic
microsatellite sequences in tumors from tas Ape
by their high frequency of
phenotype;
mobility in
/Mxh2 'mice.
We examined amplified sequences from a total of 24 small bowel and
colonie adenomas from 11 Apc+'~/Msh2~'~mice, using up to five
dinucleotide
were heterozygous and informative with a 2-8-bp separation between
the two alÃ-eles,only two RER+
adenomasfrom 2 mice of a total of 72 PCR reactions.
lymphomas were observed in two of seven Ape*' ~IMsh2~'~ mice,
repeat loci/sample. Although the majority of these loci
patterns were observed in 2 colon
Thymic
which were 70-102 days old. These lymphomas were both positive at
the DI locus (Fig. 2). Our observation
adenomas from Apc^'~/Msh2 ''mice were RER negative is not too
that the majority of intestinal
surprising since many human colorectul adenomas as well as normal
tissues from Mslû ' mice also do not display microsatellite
bility when analyzed with standard
niques (8, 17-19). Thus, widespread microsatellite
not a major factor leading to accelerated ACF and adenoma develop
ment in Ape*' IUsH2 ' mice. It seems probable that perhaps only
insta
tech(non-dilution) detection
instability is likely
one or a few critical genes manifest replication errors and are thus
responsible for the Apc+/ IMsk2 '
investigatethe contribution of APC to tumorigenesis
Mxh2 ~'~ mice.
phenotype. This prompted us to
in Apc+'~/
Loss of both copies of the APC gene, either by somatic mutations
or by inactivation of the second alÃ-ele,as demonstrated
appears to be a prerequisitefor intestinal
humans and Min mice (10, 14, 20). Although
mutations and LOH have been shown in human tumors, all adenomas
from Min mice demonstrate LOH (10, 14, 20). We examined
frequency of LOH of Ape in small bowel and colon adenomas from
mice of each genotype. Ratios of Apc+ allele:/4/?fMl" alÃ-ele for normal
liver in Ape*'' IMsh2*'^ mice were 0.75 ±0.08 (mean
by LOH.
tumor development
both somatic APC
in
the
±SD;
n = 5), and for small bowel (n = 7) and colon adenomas (n = 7)
together were 0.24 ±0.13. Mean ratios for normal liver in Apc+'~/
NHflll*M
10 11
colon
adenomas
small bowel
adenomas
mice are RER negative. Microsatellitc Fig. 2. Intestinal adenomas in Ape ' ' /Mxh2 '
analysis of normal tissues, lymphoma. and small and large bowel adenomas in Apc*'~I
Msh2
Lane 2, liver; Lane 3, thymic lymphoma; Lanes 4-fi. colon adenomas; Lanes 7-11. small
bowel adenomas. Note that only the lymphoma is RER positive (arrow).
mice. Shown are representativeresults for the DIMiiJlocus. Ltine I. spleen;
2924
American Association for Cancer Research Copyright © 1996
on July 15, 2011cancerres.aacrjournals.org Downloaded from

Page 5

ACCELERATEDTUMOR1GENESISIN MSH2-DEFICIENTMin MICE
Ape
Genotype
Msh2
L AdL AdL Ad
144 bp«»«» «••«» «•»tf»
123 bp,», w» -~«»tf»
RatioApc+:ApcMìn 055 0.16 0.77 0.17 051056
LOH++
Fig. 3. A, LOH occurs in all intestinal adenomas from Apc+f~/Msh2+/+and Apc+/~/Msh2 +f~ mice hut infrequently in adenomasfrom Apc*'~¡Msh2~'~ mice. Shown are
representative
for Ape* allele:.A/«'
from an Apc+'~IMsh2~'~
results of radioactive PCR-amplified
alÃ-eleare indicated helttw each lane. B, absence of APC COOH terminus in adenomas. Confocal fluorescence
mouse, incubated with rabbit polyclonal antibodies specific for the COOH terminus of APC protein and then stained with Cy3-conjugated
Ape loci in DNA from normal liver (L) and colon adenomas (Ad) following digestion with //mdlll.Ratios of densitometric values
micrograph of a section of a colon adenoma,
goat anti-rabbit
IgG (bur, 50 /¿m).Note the wild-type staining of normal epithelium surrounding nonstaining dysplastic epithelium (top left). Similar sections from six small intestinal and eight colon
adenomas, from Apc*'~IMsh2~' mice, demonstratedlack of staining of dysplastic epithelium,as did two colon adenomasfrom Min mice. Wild-typestaining is abolished by
competition with the COOH-terminalpeptide, against which the antibody was generated (data not shown).
Msh2+l
and colon adenomas (n = 7) together were 0.19 ±0.13. Therefore,
LOH was observed in all (27 of 27) adenomas from Apc+'~/Msh2+l+
and Apc+/~/Msh2+'~mice. Mean
Apc+/~/Msh2~'~ mice were 0.77 ±0.07 (n = 7), and ratios for small
mice were 0.72 ± O.IO (n = 6), and for small bowel (n = 6)
ratiosfornormal liverin
bowel
0.77 ±0.14, except for 3 colon adenomas,
based on ratios of 0.04, 0.16, and 0.02 and two small bowel adenomas
with ratios of 0.33 and 0.34. These data indicate that only 5 of 34
intestinal adenomas from 7 Apc+'~/Msh2~'~
(/( = 20) and colonadenomas(n = 14) together
which displayed
were
LOH
mice had lost the wild-
type Ape alÃ-ele(Fig. 3A). Failure to detect loss of the wild-type APC
alÃ-elein adenomasfrom Apc+'~/Msh2~'~ mice was not due to a
difference
compared to those from the other two genotypes. Thus, the lack of
LOH in most adenomas from Apc+'~/Msh2 ~'~ mice prompted us to
in the fraction of nonneoplastic cells in these adenomas
confirm APC deficiency
bowel and colon adenomas from Apc+'~/Msh2~'~
in these tumors. Immunostainingof small
mice, using anti
bodies generated against the COOH terminus of APC protein, dem
onstrated absence of the APC COOH terminus in dysplastic compared
to normal tissue (Fig. 3ß). Therefore, adenoma formation in MSH2-
deficient Min mice still involves inactivation of both copies of the Ape
gene, probably resulting from somatic mutation of the wild-type Ape
alÃ-ele,an hypothesis we are currently testing.
Our results provide strong biological evidence that DNA MMR defi
ciency leads to accelerated intestinal tumorigenesis.
perspective, it is believed that more frequent surveillance colonoscopy is
necessary to detect rapidly growing adenomatous polyps in at-risk indi
viduals within HNPCC families (1,5, 6). Indeed, the finding of acceler
ated colonie adenoma growth in the Apc+'~/Msh2~'~
From a clinical
mouse supports
the hypothesis that adenoma progression is more rapid in HNPCC.
Further experimentation with the Apc+'~/Msh2~'~ mouse may prove
useful in understanding hereditary colorectal cancer syndromes, includ
ing the evaluation of environmental and dietary factors thought to play a
role in the disease. In addition, crossing MSH2-deficient mice with other
mutant strains should be useful in studying the effects of targeted inac
tivation of other tumor suppressor genes.
Acknowledgments
We thank L. Siminovitch
script.
and M. Meuth for critical reading of the manu
References
1. Rustgi, A. K. Hereditary
N. Engl. J. Med.. 331: 1694-1702,
2. Groden. J., Thliveris.
Joslyn. G., Stevens. J., Spirio. L.. Robertson. M.. Sargeant. L.. Krapcho. K.. Wolff. E..
Bun. R.. Hughes, J. P.. Warrington. J.. McPherson. J.. Wasmuth. J.. Le Paslier. D.,
Abderrahim. H.. Cohen. D.. Leppert. M.. and White. R. Identification of the familial
adenomatous polyposis coli gene. Cell. 66: 589-6(X).
3. Liu, B.. Parsons, R.. Papadopoulos. N.. Nicolaides. N. C. Lynch. H. T.. Watson, P.,
Jass, J. R.. Dunlop. M., Wyllie, A.. Peltomäki.P.. de la Chapelle. A., Hamilton. S. R.,
Vogclstein.B.. and Kinzler. K. W. Analysis of mismatch repair genes in hereditary
non-polyposis colorectal cancer patients. Nat. Med.. 2: 169-174.
4. Smith. A. J.. Stern. H. S., Penner. M.. Hay, K.. Mitri. A.. Bapat. B. V.. and Gallinger.
S. Somalie APC and K-ra.s codon 12 mutations in aberrant crypt foci from human
colons. Cancer Res.. 54: 5527-5530. 1994.
5. Järvinen.H. J.. Mecklin. J. P.. and Sistonen. P. Screening reduces colorectal cancer
rate in hereditary nonpolyposiscolorectal cancer (HNPCC) families. Gastroenterol-
ogy. 108: 1405-1411.1995.
6. Lynch. H. T. Is there a role for prophylactic
nonpolyposiscolorectal cancer gcrmline mutation carriers? Dis. Colon Rectum. 39:
109-110,1996.
7. Su. L-K.. Kinzler. K. W., Vogelstein. B.. Preisinger. A. C.. Moser, A. R.. Luongo. C..
Gould. K. A., and Dove, W. F. Multiple intestinal neoplasia caused by a mutation in the
murine homolog of the APC gene. Science (Washington DC). 256: 668-670.
8. Reitmair. A. H., Schmits, R., Ewel. A.. Bapat, B.. Redston. M., Mitri. A.. Waterhouse.
P.. Mittrücker.H-W.. Wakeham. A., Liu. B.. Thomason. A.. Griesser. H.. Gallinger.
S.. Ballhausen. W. G., Fishel. R.. and Mak. T. W. MSH2 deficient mice are viable and
susceptible to lymphoid tumours. Nat. Genet.. //: 64-70.
9. Bjerknes,M. The crypt cycle and the asymptotic
differently sized mutant crypt clones in the mouse intestine. Proc. R. Soc. Lond. Ser.
B Biol. Sci.. 260: 1-6, 1995.
10. Luongo, C., Moser. A. R.. Gledhill. S.. and Dove, W. F. Loss of Ape* in intestinal
adenomas from Min mice. Cancer Res., 54: 5947-5952,
11. de Wind. N., Dekker. M.. Berns. A.. Radman. M.. and te Riele. H. Inaclivalion of the
mouse Msh2 gene resultsin mismatch
hyperrecombination.and predisposition
gastrointestinalpolyposisand nonpolyposis syndromes.
1994.
W.. Carlson. A.. Samowitz.M.. Gelbert, L.. Albertsen,H.,
1991.
1996.
subtotal colectomy among hereditary
1992.
1995.
dynamics of the proportionof
1994.
repair deficiency,
to cancer. Cell. K2: 321-330.
methylationtolerance,
1995.
2925
American Association for Cancer Research Copyright © 1996
on July 15, 2011cancerres.aacrjournals.orgDownloaded from