Loss of Apc and the entire chromosome 18 but absence of mutations at the Ras and Tp53 genes in intestinal tumors from Apc1638N, a mouse model for Apc-driven carcinogenesis.

Abstract

The Apc1638N mouse carries a targeted mutant allele at the endogenous adenomatous polyposis coli (Apc) gene and represents a unique in vivo model to study intestinal tumor formation and progression. Heterozygous Apc+/Apc1638N mice progressively develop 5-6 adenomas and adenocarcinomas of the small intestine within the first 6 months of life following a histologic sequence similar to that observed in human intestinal tumors. Here, we present the somatic mutation analysis of a total of 57 tumors. The results indicate that in > or = 75% of the lesions tested the wild type copy of the Apc gene is lost and that this LOH event extends to the entire mouse chromosome 18. Unexpectedly, mutations at the K-, N- and H-ras genes have not been found in these tumors. Immunohistochemical analysis of the Apc1638N tumors failed to detect accumulation of the Tp53 protein. Also, no mutations have been found in exons 7 and 8 of the Tp53 gene. These results indicate that, although the genetic inactivation of Apc is involved in the initiating event of the human as well as murine intestinal tumorigenesis, tumor growth and progression follow different mutational pathways in these two species.

Full text

Page 1

carc$$0219
Carcinogenesis vol.18 no.2 pp.321–327, 1997
Loss of Apc and the entire chromosome 18 but absence of
mutations at the Ras and Tp53 genes in intestinal tumors from
Apc1638N, a mouse model for Apc-driven carcinogenesis
Ron Smits1, Alex Kartheuser2,
Shantie Jagmohan-Changur1, Veronique Leblanc1,
Cor Breukel1, Anja de Vries3, Henk van Kranen3,
J.Han van Krieken4, Sophia Williamson5,
Winfried Edelmann6, Raju Kucherlapati6, P.Meera Khan1
and Riccardo Fodde1,7
ized by the development of hundreds to thousands of colorectal
adenomatous polyps in the second to third decade of life (2,3).
The majority of these mutations lead to a premature stop
codon thought to result in stable truncated polypeptides. In
the majority of adenomatous polyps from FAP patients, either
an additional truncating mutation or, less frequently, an allelic
loss can be detected at the wild type APC allele (4). Mutations
of APC have also been found in most sporadic colon cancers
and in intestinal tumors from kindreds affected by hereditary
non-polyposis colorectal cancer (HNPCC) (5,6). In general,
the frequency of APC mutations is similar in early and
more advanced stages of tumor development indicating that
inactivation of both APC copies is an initiating and probably
obligatory step for tumor formation in the intestine (4,6).
Other somatic events, both genetic and epigenetic, are
necessary for the progression of colorectal tumors. A global
DNA hypomethylation is found in early tumor stages, whereas
mutations of the K-ras oncogene occur in ~50% of the early
to intermediate types of adenomas. More advanced stages of
intestinal tumors are characterized by mutations in the DCC
(deleted in colorectal cancer) and TP53 tumor suppressor genes
in respectively 70% and 75% of the cases (7).
Several mouse models for intestinal tumorigenesis have
been generated by introducing specific mutations into the
murine Apc gene. A severe phenotype is observed in mice
heterozygous for the Min (multiple intestinal neoplasia) and
Apc∆716mutation, characterized by truncated Apc polypeptides
of 95 and 80 kDa respectively (8–10). These mutations, when
present on an inbred C57BL/6J background, result in ~100
(Min) and 200–500 (Apc∆716) adenomas throughout the entire
intestinal tract, the great majority of which have lost the
remaining wild type copy of Apc (10–13). In the Min mouse,
this Apc allelic loss extends to the entire chromosome 18 (11)
where other putative tumor suppressor genes, i.e. Dcc, Mcc
and α-catenin, are also localized (14–16). These results show
that, as for most human adenomas, mutation of both Apc
copies is required for intestinal tumor formation in mouse.
However, whether progression of these adenomas towards
malignancy is associated with mutation of the K-ras and Tp53
genes, as observed in human colorectal tumors, has not been
reported yet.
Here, we report the analysis of 42 adenomas and 15
adenocarcinomas, all resected from the intestine of Apc1638N
mice, for the presence of mutations in the Apc, H-ras, K-ras,
N-ras and Tp53 genes. The Apc1638N mouse model was
generated by gene targeting via embryonic stem cells by
introducing a neomycin expression cassette at codon 1638 of
the endogenous Apc gene. This results in a frameshift predicted
to lead to a truncated Apc polypeptide of ~185 kDa (17).
Surprisingly, no truncated protein was detectable by Western
analysis. However, immunoprecipitation studies have revealed
the presence of small amounts of the predicted mutant polypep-
tide (R.Smits, unpublished observations). Apc?/Apc1638N
mice show a significantly milder phenotype (5–6 small intest-
1MGC Department of Human Genetics, Leiden University, The Netherlands,
2Department of Surgery, University of Louvain, Bruxelles, Belgium,
3Laboratory of Carcinogenesis and Mutagenesis, RIVM, Bilthoven,
4Department of Pathology, Leiden University, The Netherlands,
5Department of Pathology, Royal Victoria Infirmary, University of
Newcastle upon Tyne, UK and6Department of Molecular Genetics,
Albert Einstein College of Medicine of Yeshiva University, New York, USA
7To whom correspondence should be addressed at: Department of Human
Genetics, Sylvius Laboratorium, Wassenaarseweg 72, 2333 AL Leiden,
The Netherlands
The Apc1638N mouse carries a targeted mutant allele at
the endogenous adenomatous polyposis coli (Apc) gene and
represents a unique in vivo model to study intestinal tumor
formation and progression. Heterozygous Apc?/Apc1638N
mice progressively develop 5–6 adenomas and adeno-
carcinomas of the small intestine within the first 6 months
of life following a histologic sequence similar to that
observed in human intestinal tumors. Here, we present the
somatic mutation analysis of a total of 57 tumors. The
results indicate that in ?75% of the lesions tested the wild
type copy of the Apc gene is lost and that this LOH event
extends to the entire mouse chromosome 18. Unexpectedly,
mutations at the K-, N- and H-ras genes have not been
found in these tumors. Immunohistochemical analysis of
the Apc1638N tumors failed to detect accumulation of the
Tp53 protein. Also, no mutations have been found in exons
7 and 8 of the Tp53 gene. These results indicate that,
although the genetic inactivation of Apc is involved in the
initiating event of the human as well as murine intestinal
tumorigenesis, tumor growth and progression follow differ-
ent mutational pathways in these two species.
Introduction
Several genes have been found to be implicated in the
progression from a normal human intestinal cell to a malignant
adenocarcinoma. One of the earliest events in this pathway
appears to be the mutation of the adenomatous polyposis coli
(APC*) gene. This gene encodes for a 312 kDa (2843 aa)
protein with unknown function, although its association with
catenins suggests a role in the regulation of cell adhesion
and in the intracellular transmission of adhesion signals (1).
Germline mutations at this gene lead to familial adenomatous
polyposis (FAP), an autosomal dominant condition character-
*Abbreviations:
matous polyposis; HNPCC, hereditary non-polyposis colorectal cancer; HE,
hematoxylin and eosin; CR, comparative ratio; ASO, allele-specific oligo-
nucleotides.
APC, adenomatous polyposis coli; FAP, familial adeno-
© Oxford University Press
321

Page 2

R.Smits et al.
LOH analysis of dinucleotide repeat markers
AmplificationandLOHanalysisofdinucleotiderepeatmarkerswereperformed
as described for the LOH analysis of the Apc locus, except for 1 pmol of
each primer. Data relative to the localization and primer sequence of the
employed markers were obtained from the Mouse Genome Database at the
Jackson Laboratory. Allele sizes were obtained from the MIT Database of
SSLP markers, except for the length of several 129/Ola alleles, kindly provided
by Dr A.R.Moser.
Assessment of allelic imbalance
Amplified fragments were resolved in a denaturing 6% polyacrylamide gel,
dried on paper and the number of counts per allele were determined on a
phosphor imager. Subsequently, the number of counts of the larger allele was
divided by the counts of the smaller allele to obtain an allelic ratio. A mean
allelic ratio was calculated for at least eight normal controls. This value was
used to generate a comparative ratio (CR) ?1.0 by dividing the tumor allelic
ratio by the mean normal allelic ratio (21). In this way, normalization of
allelic imbalances already observed in wild type controls due to preferential
amplification of a specific allele, is not necessary. A CR ?1.5 was interpreted
as significant, i.e. indicative of loss of the wild type allele.
H-, K- and N-ras mutation analysis
To detect mutations in the H-, K- and N-ras genes, exons 1 and 2 of each
gene were amplified, dot-blotted on nylon membranes and hybridized with
allele-specific oligonucleotides (ASO), recognizing all possible activating
substitutions at codons 12, 13 and 61. All exons were amplified in two PCR
rounds with primers described elsewhere (22). PCR conditions were essentially
as described for the Apc LOH analysis except for 0.2 mM of each dNTP,
3 mM MgCl2, 1 U Taq polymerase and 10 pmol of each primer in a total
volume of 50 µl. After an initial denaturation of 5 min at 94°C, each round
of amplification consisted of 30 cycles at 94°C for 60 s, 55°C for 90 s and
72°C for 120 s. Two microliters of the initial PCR were re-amplified with the
same primer pair. Subsequently, ~100 ng of PCR-product was denatured in
0.2 N NaOH and dot-blotted on nylon membrane. Each replicate filter
contained a positive control for the ASO, obtained by a second PCR round
on the wild-type amplification products with the respective ASO and an
opposing exon primer. ASO’s used for hybridization, were either 5?-end
labeled with [γ-32P]ATP or 3?-end labeled with digoxigenin-ddUTP according
to the manufacturer’s instructions (Boehringer Mannheim, Mannheim, Ger-
many). Hybridization and selective washing steps were performed as described
(23). Detection of the digoxigenin labeled oligos was accomplished using
anti-digoxigenin-alkaline phosphatase as conjugate and CSPD (Boehringer
Mannheim, Mannheim, Germany) as substrate, according to the manufacturer’s
instructions.
GC-clamped denaturing gradient gel electrophoresis (DGGE) of exons 7 and
8 of Tp53
A 242 bp and a 282 bp GC-clamped fragment encompassing Tp53 exons 7
and 8 respectively were obtained in two rounds of amplification. For the
second round of amplification nested primers were used, one of which
contained a 40-bp 5?-GC-clamp (5?-CGCCCGCCGCGCCCCGCGCCCG-
TCCCGCCGCCCCCGCCCC-3?). The position of the GC-clamped primer
and of the opposing primer were developed based on DNA melting behavior
simulations performed with the MELT87 program (24). PCR conditions were
as described for the ras mutation analysis, except for 2.5 mM MgCl2, an
annealing temperature of 60°C, and 2.5% formamide substituting for 10%
glycerol in the second round of amplification of exon 8. For each exon a
positive control with a known mutation was amplified to validate the DGGE
conditions used. The optimal ranges of denaturant used for DGGE analysis
and the primers used for amplification are listed in Table I. DGGE analysis
was performed as described elsewhere (25).
Immunohistochemistry for Tp53
Immunohistochemical staining with the polyclonal CM5 antibody (kindly
provided by Dr David Lane) was performed on 5 µm sections according to
the protocol described by Midgley et al. (26). As positive controls, sections
of UV-B induced skin tumors of hairless mice were used, in which previously
positive staining and a Tp53 mutation were detected.
inal tumors per mouse on the inbred C57BL/6J background)
when compared to Min and Apc∆716. As a consequence of the
milder phenotype and the longer lifespan of 8–16 months,
adenomas of Apc1638N are allowed more time to progress to
carcinomas and are therefore better suited to study the genetic
alterations associated with the adenoma-carcinoma sequence
in mouse.
Materials and methods
Animals, tumor samples and histopathological classification
All the C57BL/6JIco-Apc1638N (here referred to as B6-Apc1638N) animals
employed in this study were derived by backcrossing the original B6 ? 129/
Ola-Apc1638N to C57BL/6J for 7–8 generations.
To obtain intestinal tumor material, heterozygous Apc1638N mice were
killed at 8–16 months by euthanization with halothane, after which the entire
intestine was removed and opened longitudinally. Sections of ~10 cm were
spread out flat on filter paper and fixed overnight at 4°C in Notox®(Earth
Safe Industries, Inc., Bellemead, NJ). Tumors were resected, transferred to
70% ethanol and embedded in paraffin.
In total, 57 tumors were dissected from heterozygous Apc?/Apc1638N mice
on two different genetic backgrounds: 35 directly from B6-Apc1638N and 22
from F1[C3H/HeOuJIco ? C57BL/6JIco-Apc1638N] (here referred to as
C3HxB6-Apc1638N).LocalizationoftheB6-Apc1638Ntumorswasasfollows:
22 in the periampullary region, eight elsewhere in the duodenum, two in the
jejunum, and three in the ileum. Of the 22 C3HxB6-Apc1638N tumors, six
were localized in the periampullary region, another six elsewhere in the
duodenum and 10 in the jejunum. Histopathological analysis of the tumors
was performed using standard criteria for the classification of adenomas and
the assessment of the degree of dysplasia (18). The diagnosis of adeno-
carcinoma was made when glandular structures with severely abnormal shapes
and atypical epithelial cells, or small groups of epithelial cells, were obviously
surrounded by stroma showing a desmoplastic response. In many cases
invasion of muscle was seen. In some cases these abnormalities were seen
within polyps and invasion of muscle was not seen. These cases were classified
as adenocarcinomas; it is possible that some of these cases represented
intramucosal carcinomas (19). Of the 57 lesions, 15 were adenocarcinomas
and 40 tubular adenomas mainly characterized by moderate or severe dysplasia
or both, although a few mildly dysplastic adenomas were also observed. Two
out of the 57 tumors were tubulovillous adenomas with moderate and
severe dysplasia.
Tumor DNA isolation
Five to ten 25 µm thick paraffin sections were placed on glass slides,
deparaffinized, rehydrated and briefly stained with hematoxylin. The area
containing the tumor cells was subsequently scraped under a dissection
microscope using as reference a hemotoxylin and eosin (HE) stained 5 µm
section. DNA was isolated from the scraped material essentially as described
elsewhere (20). Briefly, the isolated neoplastic tissue was transferred in 250 µl
extraction buffer (10 mM Tris?HCl pH 8.0, 100 mM NaCl, 25 mM EDTA,
0.5% SDS, 300 µg/ml Proteinase K) and incubated at 55°C for 40 h. A second
aliquot of 60 µg of Proteinase K was added after 20 h of incubation. Cellular
proteins were removed by a phenol/chloroform and chloroform extraction
followed by a 30 min precipitation at ?20°C with 125 µl 7.5 M NH4Ac,
20 µg glycogen and 625 µl ethanol. The precipitate was suspended in 150 µl
TE-4(10 mM Tris?HCl pH 8.0, 0.1 mM EDTA).
Loss of heterozygosity (LOH) analysis of the Apc locus
Loss of the wild type Apc allele was analysed by amplifying in a single
reaction both the wild type and targeted Apc alleles with three primers: a
common forward primer (Apc-A2, 5?-TCAGCCATGCCAACAAAGTCA-3?)
and two allele-specific reverse primers (Apc-C2, 5?-GGAAAAGTTTAT-
AGGTGTCCCTTCT-3?, and Neo3, 5?-CACTTCATTCTCAGTATTGTTTTG-
3?). Amplification from primers Apc-A2 and Apc-C2 results in a 216 bp
product diagnostic of the wild type allele, while the combination Apc-A2 and
Neo3 yields a product of 227 bp diagnostic of the targeted Apc1638N allele.
A radioactive PCR was performed on 1 µl of tumor DNA or 1 µl normal
DNA samples, the latter isolated from normal intestinal paraffin sections and
tail samples from Apc?/Apc1638N mice. Duplicate reactions were performed
for each DNA sample. Amplifications were performed in a 10 µl volume,
containing 10 mM Tris?HCl, pH 8.9, 50 mM KCl, 2.5 mM MgCl2, 10%
glycerol, 200 µg/ml BSA, 0.01% gelatin, 0.2 mM of each dATP, dTTP and
dGTP, 0.05 mM dCTP, 0.25 µCi [32P]-dCTP (3000 Ci/mmol), 0.2 U Taq
polymerase and 10 pmol of each primer. The reactions were heated for 5 min
at 94°C followed by 29 PCR cycles at 94°C for 30 s, 55°C for 60 s and 72°C
for 90 s.
Results
General strategy
Forty-two adenomas mainly characterized by moderate or
severedysplasiaorboth,and15adenocarcinomaswereresected
from the small intestine of Apc1638N mice. Tumor DNA was
isolated and screened for the presence of mutations and/or
LOH at the Apc, H-ras, K-ras, N-ras and Tp53 genes. The
322

Page 3

A mouse model for APC-driven carcinogenesis
Table I. PCR primer and DGGE parameters for the mutation analysis of the Tp53 gene
Tp53 exon PCR product Denaturing gradientPrimer sequences
7–1st 257 bpforward: 5?-TCCCCGGCTGCTGCAGGTCA-3?
reverse: 5?-GGCGGGACTCGTGGAACAGA-3?
forward: 5?-TGTAGTGAGGTAGGGAGCGACTT-3?
reverse: 5?-GCclamp-TGGGGAAGAAACAGGCTAA-3?
forward: 5?-ATGACAAGAGGGGTTGGGAACA
reverse: 5?-AGGAGAGAGCAAGAGGTGACT
forward: 5?-GCclamp-TAGTTTACACACAGTCAGGATG-3?
reverse: 5?-GGCTCCTCCGCCTCCTTGGT-3?
7–2nd242 bp 45–65%
8–1st313 bp
8–2nd282 bp 50–70%
29 out of 35 (83%) B6-Apc1638N tumors and 14 out of 22
(64%) C3HxB6-Apc1638N tumors (Table II). No significant
loss was seen in the remainder of the tumors, i.e. 17% (n ?
6) and 36% (n ? 8) of the B6-Apc1638N and C3HxB6-
Apc1638N tumors respectively. In total, the CRs were
indicative of a loss at the wild type Apc allele in 75% (43 out
of 57) of the tumors analysed.
Analysis of chromosome 18 loss
To assess whether the observed loss at the Apc gene extends
to the entire chromosome 18 as it has been previously observed
in the Min model (11), we analysed the 22 C3HxB6-Apc1638N
tumors for the presence of LOH at four chromosome 18
dinucleotide repeat markers, known to be polymorphic between
the C3H, C57BL/6J and/or 129/Ola inbred strains. As shown
in Table IIA, of the 14 tumors where the Apc gene was found
to be lost, 13 showed LOH at all chromosome 18 markers. In
the remaining tumor (no. C14 in Table IIA) no loss at any of
the employed markers could be detected. Three tumors for
which the comparative ratios at the Apc gene were between
1.3 and 1.5 (tumors no. C15, C16 and C17 in Table IIA),
showed evidence of LOH at some of the chromosome 18
markers tested, while other markers were again close to the
significant value of 1.5. In fact, two of these three lesions
were found to be characterized by infiltration of significant
amounts (?75%) of normal cells (lymphocytes and stromal
cells) which might have interfered with the LOH analysis
(Figure 2). In the remaining five tumors where no Apc LOH
could be detected, both alleles at each chromosome 18 marker
were retained. Again, the latter lesions were characterized by
a significant infiltration of normal cells.
As mentioned before, during the analysis of the C3HxB6
tumors, we observed that a high number of tumors had retained
the 129/Ola alleles at several chromosome 18 markers. This
enabled us to perform the chromosome 18 LOH analysis on
all B6-Apc1638N tumors with dinucleotide repeat markers
known to be polymorphic between the 129/Ola and C57BL/
6J inbred strains (Table IIB). In the 29 tumors where the Apc
gene was found to be lost, marker D18Mit64, localized 13 cM
proximal to Apc, was informative in 26 out of 29 tumors.
Marker D18Mit58, localized 10 cM distal to Apc, was inform-
ative in 28 out of 29 tumors. Each of the 25 tumors informative
for both markers showed loss of the C57BL/6J alleles
diagnostic of the chromosome containing the wild type Apc
allele. No LOH at these markers was observed in the tumors
without a detectable loss of the wild type Apc allele. An
example of a LOH analysis with marker D18Mit64 is shown
in Figure 3. Although the number of chromosome 18 markers
employed in the latter study is limited, the results clearly
indicate a similar chromosomal loss as observed in the
C3HxB6 tumors.
Fig. 1. LOH analysis of nine B6-Apc1638N tumors at the Apc locus. In the
normal controls (C1 to C3) the 216 bp wild type allele (Apc?) amplifies
more efficiently than the 227 bp targeted Apc1638N allele (mean allelic
ratio obtained from eight control samples was 0.69). This normal
amplification bias is inverted in the tumors resulting in comparative ratios
indicative of a loss of the wild type allele (CR ?1.5, see Materials and
methods). Tumor numbers are depicted on top of each lane. Comparative
ratios are reported under each lane.
extent of the Apc LOH events on mouse chromosome 18 was
also examined.
Tumors were isolated from heterozygous Apc?/Apc1638N
mice on two different genetic backgrounds: C57BL/6J (B6-
Apc1638N) and an F1 resulting from a cross between the
B6-Apc1638N mice and the C3H inbred strain (C3HxB6-
Apc1638N); the latter were generated to facilitate the loss of
chromosome 18 analysis. However, during the analysis of the
C3HxB6-Apc1638N and B6-Apc1638N tumors, it became
clear that a large number of the markers from the targeted
chromosome 18 had retained the original 129/Ola alleles from
the embryonic stem cell line E14, employed for the gene
targeting experiments. Nevertheless, we were able to take
advantage of the allelic differences among C3H, B6 and 129/
Ola at several chromosome 18 markers to evaluate the extent
of the LOH events.
The Apc and Tp53 genes were analysed for loss of hetero-
zygosity, while the ras and Tp53 genes were examined for
point mutations by ASO hybridization and DGGE respectively.
Since, due to technical difficulties, only Tp53 exons 7 and 8
could be analysed by DGGE, and to exclude the presence of
mutations elsewhere in the gene, sections of the tumors were
also analysed by immunohistochemistry.
Loss of the wild type Apc allele in Apc1638N tumors
Somatic loss of the wild type Apc allele appears to be an
obligatory step for adenoma formation in the Min and Apc∆716
models (10–13). To determine whether a loss of the remaining
wild type Apc allele also characterizes Apc1638N tumori-
genesis, we selectively amplified in a single PCR reaction both
the wild type and targeted Apc allele from tumor DNA samples.
An example of the Apc LOH analysis of nine tumor samples
and three normal controls is shown in Figure 1. A significant
loss of the wild type allele (CR ?1.5) could be detected in
323

Page 4

R.Smits et al.
Table IIA. Histopathological analysis and chromosome 18 loss of the C3HxB6-Apc1638N tumors
Tumor no. C1C2 C3C4 C5C6 C7C8 C9C10 C11 C12 C13 C14 C15 C16 C17C18 C19 C20 C21 C22
Ad/Car
Dysplasia
Contamination
D19Mit19 (2 cM)
Apc (15 cM)
D18Mit35 (24 cM)
D18Mit81 (42 cM)
D18Mit80 (50 cM)
Ad
1
?
?
?
?
?
?
Ad
1–2
?
?
?
?
?
?
Ad
1–2
?
?
?
?
NI
?
Ad
1–2
?
?
?
?
?
?
Ad
2
?
?
?
?
NI
?
Ad
2–3
?
?
?
?
NI
?
Ad
2–3
?
?
?
?
NI
?
Ad
2–3
?
?
?
?
NI
?
Ad
2–3
?
?
?
?
NI
?
Ad
2–3
?
?
?
?
NI
?
Car CarCarAd
2
?? ?
?
?
?
NI
?
Ad
1–2
Ad
2
?? ??? ?? ?? ?? ?? ???
????
????
????
?
NINI
?
????
CarAd
1–2
Ad
2
Ad
2–3
CarCar
?
?
?
?
?
?
?? ?
?
?
?
NI
?
?
?
?
NI
?
?
?
?
NI
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
Table IIB. Histopathological analysis and chromosome 18 loss of the B6-Apc1638N tumors
Tumor no123456789101112 131415 1617 18
Ad/Car
Dysplasia
Contamination
D18Mit64 (2 cM)
Apc (15 cM)
D18Mit58 (25 cM)
Ad
1
?
?
?
?
Ad
2
?
?
?
?
Ad
2
??
?
?
?
Ad
2
??
NI
?
?
Ad
2–3
?
?
?
?
Ad
2–3
?
?
?
?
Ad
2–3
?
?
?
?
Ad
2–3
?
?
?
?
Ad
2–3
??
?
?
?
Ad
2–3
??
?
?
?
Ad
2–3
??
?
?
?
Ad
2–3
??
?
?
?
Ad
2–3
??
NI
?
?
Ad
2–3
??
NI
?
?
Ad
2–3
??
?
?
NI
Ad
3
?
?
?
?
Ad
3
?
?
?
?
Ad
3
?
?
?
?
Tumor no. 1920212223 24252627 28 2930 313233 3435
Ad/Car
Dysplasia
Contamination
D18Mit64 (2 cM)
Apc (15 cM)
D18Mit58 (25 cM)
Ad
3
??
?
?
?
Ad
3
??
?
?
?
Ad
3
??
?
?
?
CarCarCar CarCarCarCarCarAd
2
?
?
?
?
Ad
2–3
?
?
?
?
Ad
2–3
?
?
?
?
Ad
2–3
??? ??? ???
??
??
??
Ad
2–3
Car
?
?
?
?
?
?
?
?
??
?
?
?
??? ?
?
?
?
?
?
?
?
??
?
?
?
??
?
?
?
?
?
?
?
?
?
Histological classification: adenoma (Ad), adenocarcinoma (Car).
Degrees of dysplasia among adenomas: mild (1), moderate (2) and severe (3).
Contamination of the tumor region with normal cells: ?40% (?), 40–75% (??), ?75% (???).
Comparative ratios indicative of a loss of an allele [CR ?1.5, see Materials and methods, are depicted as (?) while values ?1.5 are depicted as (?)].
Markers are ordered from centromere to telomere; the genetic distance from the centromere is indicated between parentheses. Non-informative markers for
each tumor are indicated as NI.
Ras mutation analysis
Mutationalactivationof theK-rasgeneis afrequentlyobserved
event in human colorectal malignancies and appears to con-
tribute to adenoma progression (7). We have screened all
Apc1638N tumors for the presence of activating mutations at
codons 12, 13 and 61 in the murine K-ras homologue by ASO
hybridization. Surprisingly, no mutation was detected in the
57 tumors examined. To ascertain whether another member of
the ras family fulfils a similar role in murine intestinal
tumorigenesis as the K-ras gene in man, a subset of 43 tumors
was analysed for the presence of mutations in H- and N-ras
at codons 12, 13 and 61 (Figure 4). As in the case of the
K-ras, no mutations were identified at these genes, clearly
indicating that ras mutational activation is very rarely associ-
ated with intestinal tumor progression in the Apc1638N model.
Tp53 mutation analysis
The Tp53 gene, as is the case of K-ras, is frequently involved
in the progression of human intestinal cancer (7). We employed
denaturing gradient gel electrophoresis to detect point
mutations in exons 7 and 8 of the murine Tp53 gene. In all
57 tumors analysed, no aberrant DGGE patterns indicative of
a mutation could be detected. LOH-analysis of the 22 C3HxB6-
Apc1638N tumors with markers D11Mit86 and D11Mit320,
respectively 11 cM proximal and 5.5 cM distal to Tp53,
showed no significant loss of the C3H or B6 allele (results
not shown).
Since LOH analysis and screening of exons 7 and 8 alone
might result in an underestimation of the involvement of Tp53
in Apc1638N intestinal tumorigenesis, and because point
mutations at this gene often result in the stabilization of the
protein making it detectable by immunohistochemistry, we
employed the polyclonal CM5 raised against mouse Tp53 for
the immunostaining of paraffin embedded Apc1638N tumor
sections (26). No specific staining of tumor cells was detectable
in any of the tumors. These results show that, similar to the
ras genes, mutation of Tp53 is a very infrequent event in
intestinal tumor progression in the Apc1638N mouse model.
Discussion
Somatic mutation analysis of the intestinal Apc1638N tumors
has revealed that, as in the case of Min and Apc∆716, the
majority (75%) of the tumors are characterized by a detectable
loss of the wild type copy of Apc. This loss appears to extend
to the entire chromosome 18, as observed in Min. The results
here reported also indicate that 25% (n ? 14) of the Apc1638N
tumors analysed did not undergo LOH at the Apc locus.
However, it should be said that several of the tumors analysed
in the present study (11 out of 14) were characterized by
infiltrationofthetumorwithsignificantamountsofcontaminat-
ing lymphocytes and stromal cells which, upon isolation, may
obscure the LOH event. A high number of normal cells in
neoplastic areas is most frequently associated with more
324

Page 5

A mouse model for APC-driven carcinogenesis
Fig. 2. Examples of H&E stained small intestinal Apc1638N tumors. A tubular adenoma with moderate dysplasia (A) and an adenocarcinoma (B) are shown
(magnification ?25). Note the presence of large numbers of lymphocytes around the glands of the carcinoma in (B).
advanced stages of tumor progression in Apc1638N (see also
Table II). An illustrative example of lesions with mild and
severe contamination of normal cells is shown in Figure 2.
This high number of normal cells observed in several of
the Apc1638N lesions makes it technically more difficult to
selectively isolate the tumor cells with the method employed
here. The detectable loss of Apc in 75% of the Apc1638N
tumors may therefore represent an underestimation of the
actual LOH frequency. Nevertheless, in three of the 14 tumors
showing no Apc loss, the HE sections used as reference showed
no prominent number of normal cells in the neoplastic region
(tumors no. 30, 31, 32 in Table IIB). Therefore, alternative
routes for adenoma formation and progression may exist
independently of a complete Apc inactivation due to large
deletionsorchromosomallosses,i.e.truncatingpointmutations
or small intragenic deletions and insertions. Interestingly,
absence of LOH in tumors of the Min mouse has also been
reported for four out of 32 tumors studied by Laird et al. (13).
No mutations of the K-ras, H-ras, N-ras and Tp53 genes
were detected in any of the Apc1638N tumors. This result is
in contrast with the situation observed in the majority of
human colorectal lesions characterized by high frequencies
of K-ras and TP53 mutations. Several explanations can be
envisaged for the virtual absence of ras and Tp53 gene
mutations as reported in this study. Although DGGE and
ASO-probe hybridization are generally considered sensitive
techniques (27,28), it is possible that a mutation present in a
minority of the cells is not detected. As far as Tp53 is
concerned, analysis of exons 7 and 8 by DGGE and immuno-
histochemistry will not detect mutations present at other exons
not resulting in accumulation of the protein. It appears very
unlikely, however, that a more detailed and extensive investi-
325