Download Repeated Sequences in CASPASE-5 and FANCD2 but not NF1 Are

Repeated Sequences in CASPASE-5 and FANCD2
but not NF1 Are Targets for Mutation in
Microsatellite-Unstable Acute Leukemia/
Myelodysplastic Syndrome
Judith Offman,1 Karen Gascoigne,1 Fiona Bristow,1 Peter Macpherson,1 Margherita Bignami,2
Ida Casorelli,2 Giuseppe Leone,3 Livio Pagano,3 Simona Sica,3 Ozay Halil,4
David Cummins,4 Nicholas R. Banner,4 and Peter Karran1
1
Cancer Research UK London Research Institute, Mammalian DNA Repair Laboratory, Clare Hall Laboratories,
South Mimms, Herts, United Kingdom; 2Istituto Superiore di Sanitá; 3Department of Hematology, Catholic
University, Rome, Italy; and 4Harefield Hospital, Harefield, Middlesex, United Kingdom
Abstract
Microsatellite instability (MSI) in tumors is diagnostic for
inactive DNA mismatch repair. It is widespread among
some tumor types, such as colorectal or endometrial
carcinoma, but is rarely found in leukemia. Therapyrelated acute myeloid leukemia/myelodysplastic
syndrome (tAML/MDS) is an exception, and MSI is
frequent in tAML/MDS following cancer chemotherapy or
organ transplantation. The development of MSI+ tumors
is associated with an accumulation of insertion/deletion
mutations in repetitive sequences. These events can
cause inactivating frameshifts or loss of expression of
key growth control proteins. We examined established
MSI+ cell lines and tAML/MDS cases for frameshift-like
mutations of repetitive sequences in several genes that
have known, or suspected, relevance to leukemia.
CASPASE-5, an acknowledged frameshift target in MSI+
gastrointestinal tract tumors, was frequently mutated in
MSI+ cell lines (67%) and in tAML/MDS (29%). Frameshiftlike mutations were also observed in the NF1 and
FANCD2 genes that are associated with genetic
conditions conferring a predisposition to leukemia. Both
genes were frequent targets for mutation in MSI+ cell
lines and colorectal carcinomas. FANCD2 mutations
were also common in MSI+ tAML/MDS, although NF1
mutations were not observed. A novel FANCD2
polymorphism was also identified.
(Mol Cancer Res 2005;3(5):251 – 60)
Introduction
Cancer arises by a process of somatic mutation and
selection, and its development is influenced by the rate of
spontaneous somatic mutation. DNA mismatch repair (MMR)
Received 10/27/04; revised 3/11/05; accepted 3/18/05.
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 in
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Peter Karran, Cancer Research UK London Research
Institute, Mammalian DNA Repair Laboratory, Clare Hall Laboratories, South
Mimms, Herts, United Kingdom EN6 3LD. Phone: 44-207-269-3870; Fax: 44207-269-3812. E-mail: [email protected]
Copyright D 2005 American Association for Cancer Research.
Mol Cancer Res 2005;3(5). May 2005
normally provides protection against mutation by correcting
DNA replication errors. As expected, MMR deficiency is
associated with an increased mutation rate and with cancer (for
review, see ref. 1). A significant fraction of certain tumors—
notably, but not exclusively, those of endometrial, colorectal,
and other gastrointestinal sites—are MMR defective. In the
MMR pathway, proteins encoded by the MSH2, MLH1, PMS2,
MSH6, and MSH3 genes form a series of heterodimeric
complexes that recognize promutagenic replication errors and
initiate their correction. A defect in any of these proteins
compromises repair efficiency (reviewed in ref. 2).
MMR is particularly efficient at reversing unpaired
nucleotides that arise during replication of DNA regions
comprising extensive mononucleotide or dinucleotide repeats
of the type found in microsatellites. If they remain
uncorrected, these structural aberrations cause insertion/
deletion mutations, which alter the length of the repeat.
MMR-deficient tumors contain many thousands of altered
microsatellite sequences (3, 4) and have been described as
having a microsatellite mutator phenotype or microsatellite
instability (MSI). In practice, MSI is usually defined as
alterations in a significant fraction of a standard panel of five
microsatellites in tumor DNA. This is considered diagnostic
for defective MMR.
MSI+ tumors do not, in general, exhibit widespread
chromosomal rearrangements or aneuploidy (5). This most
likely reflects at least partially distinct pathways of tumor
development that accompany MSI and chromosomal instability.
MSI is particularly associated with the accumulation of
inactivating frameshifts in repetitive elements, resembling
microsatellites but located within coding sequences of critical
tumor suppressor genes. Transforming growth factor-h (TGF-h)
provides a good example (6). Escape from the growth restraints
imposed by TGF-h is common in the development of colorectal
cancer. The TGFb-RII gene, which encodes a component of the
TGF-h receptor contains an A10 sequence. In MSI+ colorectal
tumors, TGFb-RII is almost invariably inactivated by a
frameshift in this repeat. These mutations are rare in MMRproficient, MSI colorectal tumors. The key gene products that
impose normal growth control are tissue dependent, and TGFbRII mutations are also infrequent in MSI+ tumors from tissues
that are not normally TGF-h responsive. Other examples of
251
252 Offman et al.
genes that are susceptible to inactivating frameshifts in MSI+
tumors include BAX (7), RAD50, and BLM (8). Each of these
contains a repeat of eight or more mononucleotides in coding
DNA sequences.
Extensive mononucleotide or dinucleotide repeats are rare in
exonic DNA sequences. Uninterrupted monotonic repeats occur
more frequently in the pyrimidine-rich tract that precedes
mRNA splice acceptor sites. Small insertions or deletions in
these intronic sequences can cause exon skipping and mRNA
instability. In one example, the single-base alterations in an
intronic T11 tract of the MRE11 cell cycle checkpoint/DNA
repair gene that are common in established MSI+ tumor cell
lines and, importantly, in MSI+ colorectal tumors (9), cause
defects in the DNA damage response. Mutated cells fail to
arrest DNA replication after exposure to ionizing radiation and
are sensitive to radiomimetic agents.
The proportion of sporadic tumors that are MSI+ varies with
the tumor type and location. For example, 15% to 20% of
sporadic colorectal or endometrial carcinomas are MSI+,
whereas the phenotype is notably rare in leukemia and affects
<5% of cases. Therapy-related acute leukemia is an important
exception (reviewed in ref. 10). Probably more than half of
acute myeloid leukemia/myelodysplastic syndrome (AML/
MDS) cases that occur secondary to successful chemotherapy
are MSI+ (11-14). We recently reported a high incidence of MSI
among AML/MDS cases from solid organ transplant patients
who received prolonged treatment with immunosuppressive
drugs (15). We have suggested that MMR defects in
chemotherapy or immunosuppression-related AML/MDS
(tAML/MDS) reflect the intrinsic drug resistance of MSI+
myeloid cells (16).
If loss of MMR is an early step in tAML/MDS, frameshift
mutations in key genes regulating normal hematopoiesis may
be important in the development of these secondary
malignancies. To understand how MSI+ tAML/MDS might
develop, it is necessary to define the relevant frameshift
targets. Here, we show that CASPASE-5, a gene acknowledged
to be a frameshift target in other tumor types, is also mutated
in MSI+ tAML/MDS. In addition, we identify an intronic T26
sequence in the NF1 gene as a target for insertion/deletion
mutations in MSI+ tumors. Mutations in this sequence were
frequent in MSI+ colorectal carcinoma but not in tAML/MDS.
A mononucleotide intronic repeat in the FANCD2 gene, an
important determinant of genetic stability, is identified as
target for insertion/deletions in colorectal carcinoma and
tAML/MDS. A novel polymorphism in this sequence is also
described.
Results
Putative Frameshift Targets
Table 1 lists the genes that were analyzed and Fig. 1
indicates the location and nature of each repeat sequence. Most
of the repetitive elements comprise z7 mononucleotides. We
also examined a (TG)5 repeat in FANCG and a (T)6 sequence in
NF1. The putative targets fall into three broad classes: genes
known to be mutated by insertion or deletion in MSI+ tumors
(e.g., CASPASE-5), genes implicated in the development of
AML or associated with particular susceptibility to AML/MDS
(e.g., FANCD2 and NF1), and tumor suppressor genes that have
not been identified previously as frameshift targets (e.g., RB).
Sequences were identified from the ENSEMBL database
(http://www.ensembl.org/).
CASPASE-5 and NF1 Mutations in MSI+ Tumor Cell Lines
Twelve established MSI+ tumor cell lines and 2 MMRdeficient cell lines selected by in vitro treatment (RajiF12 and
A2780-MNUcl1) were compared with 6 MMR-proficient lines
(Table 2). The regions of interest were amplified with
fluorescent primers and the length of the fluorescent products
was used to define a genotype. The approach was validated
using the A10 repeat of CASPASE-5, which is a known target
for mutation in MMR-deficient cells. Examples of genotyping
and direct sequencing are shown in Fig. 2. The CASPASE-5 A10
sequence was intact in all five MMR-proficient cell lines we
examined. Eight of 12 (67%) MMR-defective cell lines
contained a one-base deletion in at least one CASPASE-5 T10
allele (Table 2). In two cases, only the T9 allele was observed,
consistent with biallelic mutations or loss of heterozygosity.
The high frequency of CASPASE-5 alterations in these cell lines
is consistent with published reports of frequent CASPASE-5
mutations in MSI+ gastrointestinal, endometrial, and other
tumors. The perfect correspondence between the sequencing
data and genotyping indicates that the fluorescent PCR-based
method provides a sensitive and accurate analysis of frameshift
target sequences.
An uninterrupted T26 repeat precedes the splice acceptor site
in intron 24/25 of the NF1 gene on chromosome 17 (Fig. 1).
We amplified a 232-nucleotide region that included the T26
repeat. The PCR products from the four MMR-proficient cell
lines we examined were all closely similar with two
predominant peaks of similar intensity that corresponded to
232 and 231 nucleotides (Fig. 3A). The shorter product is
consistent with errors during PCR of this long repeat. This
region was altered in five of five MMR-defective cell lines
Table 1. Target Genes, Location, and PCR Primers
Target
Gene
Type of Location
Repeat
CASPASE-5 (A)10
CHK1
(A)9
FAS
(T)9
RB
(T)9
FANCD2
(T)10
FANCE
(C)7
FANCG
(TG)5
NF1
(T)26
(C)7
(T)6
PCR
Primers
Exon 2
CAGAGTTATGTCTTAGGTGAAGG
AACATGAAGAACATCTTTGCCCAG
Exon 7
CTGCCATGCCTATCCTGATT
TCACACAATGATGAAACCACCT
Intron 4/5
ATCACCTGGCCATTTTCTTG
GGTTGGGGGAAAGGAGAATA
Intron 6/7
AAGAAAGAAAATCTTTACCATGCTG
CAGCCTTAGAACCATGTTTGG
Intron 5/6
CTTGCAAAGAGCCATCTGCT
ATCCTGTGTTCCCGCTATTT
Exon 4
GGGAGCTTCTCCACTGTCTG
TGAGTCCTTTCTGCGTTTCC
Intron 3/4
AGGAAAGCCAGAGTGTGTGG
ACTGCAGCTGGAGAGAAAGG
Intron 24/25 TCCCCATTTGAGATGATTTTG
AACTACTGCTTCCCATGCTTGC
Exon 18
CCCAAGTTGCAAATATATGTC
TACCTGTTGCAAATATATTGC
Exon 36
CAGTAGACAACATAAAGCCTC
ATTCCTGTTAAGTCAACTGGG
Mol Cancer Res 2005;3(5). May 2005
Frameshift Targets in Therapy-Related AML
FIGURE 1. Location of target
sequences and lengths of PCR
products. Exonic sequences
are boxed. Repeat sequences
and splice site AG sites are
shown in bold. Arrows, positions
of PCR primers. The length of
the expected PCR product is
indicated (bp).
consistent with considerable instability of the T26 repeat
(Table 2). In each case, the mutated alleles were shorter and
the extent of the change ranged from f3 to 9 nucleotides
(Table 2; examples in Fig. 3A).
To investigate whether the NF1 microsatellite mutations
caused exon skipping, we examined mRNA from LoVo or
SW48 cells, both of which contained mutated NF1 alleles,
and SW480 and HeLa cells, in which the NF1 genotype was
normal. A skipped NF1 exon 25 would result in a deletion of
117 nucleotides. Reverse transcription-PCR (RT-PCR) amplification of the region between mid-exons 23 and 26 produced
the expected 383-nucleotide product from MMR-proficient
SW480 and HeLa cells. A product of apparently identical size
was amplified from MMR-defective LoVo and SW48 and
there was no evidence of a truncated 266-nucleotide species
(Fig. 3B). The presence of an intact exon 25 in LoVo and
SW48 NF1 mRNA was confirmed by direct sequencing (data
not shown). We conclude that the intronic T26 microsatellite in
intron 24/25 of NF1 is a frequent target of mutation in MMRdefective cells but that these mutations do not cause deletion
of exon 25.
Mol Cancer Res 2005;3(5). May 2005
Table 2. Mutations in MMR-Proficient and MMR-Deficient
Cell Lines
Cell Line
MMR
CASPASE-5 (A)10
NF1 (T)26
FANCD2 (T)10
HeLa
Raji
SW480
Ramos
BL2
A2780-SCA
Jurkat
SW48
HCT116
LoVo
DLD1
Hec1A
REH
Molt4
CCRF-CEM
AN3CA
DU145
SKOV3
A2780-MNUcl1
RajiF12
+
+
+
+
+
+
(A)10
(A)10
(A)10
(A)10
(A)10
ND
(A)9
(A)9/(A)10
(A)9/(A)10
(A)9/(A)10
(A)9/(A)10
(A)10
(A)10
(A)9/(A)10
(A)9
(A)10
(A)9/(A)10
(A)10
ND
ND
(T)26
ND
(T)26
(T)26
(T)26
ND
ND
(T)20
ND
(T)17
(T)23
(T)22
ND
ND
(T)23
ND
ND
ND
ND
ND
(T)8/(T)10
(T)8/(T)10
(T)10
(T)10
(T)10
(T)8/(T)10
(T)9/(T)10
(T)10/(T)11
(T)10/(T)11
(T)10
(T)8/(T)10
(T)10
(T)10/(T)11
(T)8
(T)10
(T)8
(T)10
(T)8/(T)10
(T)8/(T)10
(T)8/(T)10
NOTE: ND, not determined.
253
254 Offman et al.
FIGURE 2. CASPASE-5
repeats in tumor cell lines. A.
Genotyping. The CASPASE-5
repeat was amplified from Raji,
SW48, or Jurkat cell DNA using
fluorescent PCR primers. The
length of the PCR product was
determined. The sequences
have been aligned and the vertical line indicates the position of
the wild-type A10 sequence. B.
Sequence analysis. The same
region was amplified by PCR
and sequenced.
The two short NF1 exonic pyrimidine repeats, C7 in exon 18
and T6 in exon 36 (Fig. 1), were not mutated in our panel of
MMR-defective cell lines (data not shown).
No mutations in the A9 or A7 repeats of the CHK1 gene, the
T9 repeat of FAS, and the T9 repeat of RB were detected in any
of the MSI+ cell lines (data not shown). For this reason, these
sequences were not considered further.
CASPASE-5 and NF1 Mutations in MSI+ Malignancies
In a previous study, 16 MSI+ chemotherapy-related
leukemias were identified among a total of 25 cases (14).
We examined the A10 repeat of CASPASE-5 in 22 of the
cases for which sufficient DNA remained. Three MSI+ cases
were heterozygous for a single-base deletion (A9/A10).
The remaining 11 and 8 MSI cases all had the wild-type
A10 genotype (Table 3). Interestingly, one of the CASPASE-5
mutations was in case 8 that we previously designated MSI
stable (14) based on analysis of the BAT26 microsatellite.
Recently, doubts have been raised as to the informativeness
of BAT26 as a unique diagnostic marker for MSI in
tAML (17). In view of this, and the finding that both the
CASPASE-5 A10 tract and the FANCD2 T10 tract (see below)
are altered in case 8, it seems appropriate to reclassify it
as MSI+.
We also examined CASPASE-5 in six cases of MSI+ AML/
MDS from organ transplant patients (15). A sample of the
explanted heart or lung provided a source of nontumor DNA.
All nontumor tissues had the expected A10 genotype (Table 4).
Two AML/MDS cases were heterozygous for A10/A9. For
case 2, the T10/T9 genotype was confirmed in a second,
independent biopsy. Because CASPASE-5 mutations were
found in 5 of 17 (29%) of the cancer therapy-related and
immunosuppression-related tAML/MDS cases, we conclude
that frameshifts in the CASPASE-5 A10 repeat are common in
this group of diseases.
No mutations in the NF1 T26 sequence were observed
among nine chemotherapy-related AMLs examined. Four
MSI+ cases and 5 microsatellite-stable cases were indistinguishable from 10 normal blood donors. All 19 samples
produced a similar 231/232-nucleotide pattern characteristic
of the T26 sequence in the microsatellite-stable cell lines
(examples in Fig. 3A). This consistent pattern indicates that
the NF1 T26 microsatellite is likely to be monodisperse in
the normal population. There was no indication of instability
at this locus in five MSI+ transplant-related AMLs (Table 4).
Because the apparent stability of the NF1 T26 microsatellite
in the tAML/MDS cases was surprising in view of its
frequent mutation in the MSI+ cell lines, we also analyzed a
series of colorectal carcinoma DNAs. The normal 231/232nucleotide pattern was observed in 17 of 17 MSI samples.
In four of five MSI+ colorectal carcinoma, there was
evidence of instability and we observed deletions ranging
Mol Cancer Res 2005;3(5). May 2005
Frameshift Targets in Therapy-Related AML
from 2 to 10 nucleotides (examples of a stable case and a
MSI+ case are shown in Fig. 3A). These data confirm that
the NF1 T26 sequence is a frequent mutational target in
MSI+ tumors as well as established cell lines. Surprisingly,
these NF1 mutations seem uncommon in MSI+ tAML/MDS
cases.
The exonic C7 or T6 repeats of NF1 were not mutated in
any of the MSI+ or MSI tumors (data not shown).
FANCD2
A FANCD2 Polymorphism. Fanconi anemia is associated
with a high incidence of AML. We therefore sought potential
frameshift targets in the Fanconi anemia genes. The ENSEMBL
database revealed three possibilities: a FANCE C7 repeat, (TG)5
in FANCG, and T10 immediately adjacent to the splice acceptor
site for intron 5/6 of FANCD2 (Fig. 1).
No mutations were observed in the FANCE or FANCG
repeats in any of the cell lines irrespective of MMR status (data
not shown). Direct sequencing of the FANCD2 T repeat
confirmed the presence of the T10 tract in MMR-proficient cell
lines and normal individuals. In some cases, however,
sequencing data suggested the presence of a shorter allele.
Subsequent genotyping confirmed the presence of a T8 allele in
these cases. Figure 4A shows examples of the T10/T8 and T10
genotypes from microsatellite-stable cells together with the
corresponding sequencing data. When present, the T8 allele was
consistently observed in repeat determinations, indicating that it
is unlikely to be a PCR or sequencing artifact. Among 10
normal blood donors, 11 EBV-transformed cell lines established
from normal individuals, 6 MSI tumor cell lines, and 19 MSI
colorectal carcinomas, about half of the samples had the T10
genotype and the remainder had T10/T8. The single exception
was GM0621—an EBV-transformed cell line from a clinically
normal adult—in which only the T8 allele was present. These
data are summarized in Table 5. Thus, of 46 MMR-proficient
cells or tumors, 28 (61%) were T10, 17 (37%) were T10/T8, and
the single example of T8 represents 2% of the total. If each case
of T10 or T8 contains two copies, the allele frequencies are
f0.8 for T10 and 0.2 for T8. This suggests that the T8 allele is a
relatively infrequent polymorphism.
FANCD2 Mutations in MSI + Cell Lines. The FANCD2 T10
sequence was a mutational target in MSI+ cells (Table 2).
Among the 14 MMR-defective cell lines, novel alleles were
observed in Jurkat (T9), SW48 (T11), HCT116 (T11), and REH
(T11). Examples of altered alleles in MSI+ cells are shown in
Fig. 4B. There were also two examples of apparent homozygosity for T8 in this series. The remaining eight cell lines were
either T10 or T10/T8. It was noteworthy that all the lines retained
at least one T10 or T8 allele. The T11 and T9 alleles together
comprise f15% of the total among the MMR-deficient cells.
The T8 allele frequency was f0.3 in these cell lines, close to
the frequency in repair-proficient samples. In view of the
instability of this tract, it is possible that some of these T8
Table 3. Mutations in Chemotherapy-Related AML Samples
FIGURE 3. NF1 genotype and mRNA expression. A. NF1 genotype.
DNA from cells or tumors was genotyped for the T26 intronic repeat. Left,
established MMR-proficient (HeLa and SW480 ) or MMR-deficient (LoVo
and SW48 ) cells. Center, normal individuals (BD63 and BD829 ), 1344c
(MSI colorectal carcinoma), and 930c (MSI+ colorectal carcinoma). Right,
normal tissue (A3b) and MSI+ AML/MDS (A1a) from organ transplant
patients. MSI+ (T12 ) and MSI (T10 ) chemotherapy-related AML. The
sequences have been aligned and the vertical line indicates the position
of the 232-bp product. B. NF1 mRNA. A fragment of the NF1 transcript
spanning exons 5 to 9 was amplified by RT-PCR. The PCR products were
separated on a 2% agarose gel and visualized with ethidium bromide.
Arrow, expected position for a product with a deleted exon 25.
Mol Cancer Res 2005;3(5). May 2005
AML
Case
MSI
CASPASE-5
FANCD2
AML
Case
2
3
5
6
7
8
9
12
13
17
18
20
21
23
+
+
+
+
+
+
+
+
+
+
+
+
+
+
(A)9/(A)10
(A)10
(A)10
(A)10
(A)10
(A)9/(A)10
(A)10
(A)10
(A)10
(A)10
(A)10
(A)10
(A)10
(A)9/(A)10
(T)9
(T)10
(T)10
(T)10
(T)10
(T)8/9/10
(T)10
(T)10
(T)8
(T)8/10
(T)10
(T)10
(T)8/10
(T)8/10
1
10
11
14
19
22
24
25
NOTE: Case numbers are taken from ref. 14.
MSI
CASPASE-5
FANCD2
(A)10
(A)10
(A)10
(A)10
(A)10
(A)10
(A)10
(A)10
(T)10
(T)10
(T)8/10
(T)10
(T)10
(T)10
(T)10
(T)10
255
256 Offman et al.
Table 4. Mutations in Transplant-Related AML Samples
AML Case
Tissue
CASPASE-5
FANCD2
A1a
A1b
A2a
A2b
A3a
A3b
A4a
A4b
A5a
A5b
A7
BM
Heart
BM1
Lung
BM
Heart
BM
Heart
BM
Heart/lung
BM
(A)10
(A)10
(A)9/(A)10
(A)10
(A)9/(A)10
(A)10
(A)10
(A)10
(A)10
(A)10
(A)10
(T)8/(T)10
(T)8/(T)10
(T)8
(T)8/(T)10
(T)8/(T)10
(T)10
(T)8/(T)10
(T)8/(T)10
(T)8/(T)10
(T)8/(T)10
(T)8/(T)10
NOTE: BM, bone marrow. Case numbers taken from ref. 15.
alleles arose by mutation, however. The frequent occurrence of
novel T9 or T11 alleles indicates that this repeat is a significant
target for mutation in MSI+ cells.
FANCD2 Mutations in MSI + Malignancies. The FANCD2
repeat was also mutated in MSI+ tumors. There were two
examples of T9 among 5 MSI+ colorectal cancers (40%). In
one of these, T9 was the only allele observed. Among 22
chemotherapy-related AML samples, all MSI cases were
either T10 (7 cases) or T10/T8 (1 case; Table 3). Of the 14
MSI+ cases, 8 were T10 and 3 were T10/T8. There was a single
example each of apparent homozygosity for T9 and for T8. In
one sample (case 8; see above), all three alleles were
repeatedly observed. Patterns consistent with mutation were
observed in the transplant-related AML/MDS samples
(Table 4). In case A3, the explanted tissue had a T10
genotype. In contrast, the AML/MDS was T10/T8, suggesting
that the T8 allele arose as the result of a mutation. In case A2,
the constitutive genotype was T10/T8, whereas the AML/MDS
was T8. This change is also consistent with a two-base
deletion, although it could also have occurred by loss of the
T10 allele.
In summary, a (T8) variant of the T10 mononucleotide
repeat that precedes the intron 4/5 splice acceptor site of
FANCD2 seems to be a novel polymorphism that is relatively
rare in the normal population. In a MMR-defective background, the FANCD2 repeat is a target for insertion or
deletion mutation. The frequency of these mutations is high
and they were observed in 15% to 40% of MSI+ cell lines and
tumors.
Effect of Repeat Length on the Function of FANCD2.
Representatives of each of the FANCD2 genotypes were
analyzed by RT-PCR. In each case, the major product was of
the length expected for a full-length FANCD2 mRNA (Fig. 4C).
A minor, shorter product was also present (Fig. 4C, arrow).
This was most apparent in LoVo and AN3CA, although it was
detected in all of the samples, including MSI ones, by more
sensitive staining (data not shown). These truncated cDNA
products were sequenced and alignment by BLAST 2 search
indicated that they were identical to FANCD2 with a deleted
exon 8 (data not shown). Skipping exon 8 generates a
premature stop codon in exon 9. Because this aberrant splice
variant is present in all cell lines, irrespective of MMR status,
we conclude that it probably is not a consequence of mutation
and is unlikely to be physiologically relevant. There was no
evidence of a cDNA product corresponding to a skipped exon
6, the expected outcome of altered splicing due to alterations in
the intronic T10/T8 sequence.
Homozygous inactivation of FANCD2 causes a dramatic
increase in sensitivity to DNA cross-linking agents, such as
mitomycin C. To assess whether homozygosity for the less
frequent T8 allele, or the presence of a mutant allele affected
sensitivity to cross-linking agents, we examined the growth of
Raji (T10/T8), Molt4 (T8), and REH (T10/11) after treatment with
mitomycin C. Although there were slight differences, there was
no indication of dramatic mitomycin C sensitivity that would be
consistent with a severely compromised FANCD2 function
(Fig. 5A).
The FANCD2 protein exists in two isoforms (FANCD2-S
and FANCD2-L) that can be resolved by Western blotting.
Under normal conditions, FANCD2-S predominates. Following
DNA damage, FANCD-2 is monoubiquitinated and converted
to the more slowly migrating FANCD2-L form (18). We
investigated whether conversion to the FANCD2-L form was
impaired in cells of the T8 genotype. Figure 5B shows that
exposure to ionizing radiation induced conversion of FANCD2S to FANCD2-L in Raji (T10/T8), GM00621(T8), and Molt4
(T8) cells. Taken together with the absence of acute sensitivity
to mitomycin C, these findings suggest that the T8 isoform is
not associated with a severe impairment of FANCD2 protein
function.
Discussion
The identification of mutational targets for insertion/deletion
mutations in MSI+ tAML is a prerequisite for understanding
how MSI+ tAML develops. Our data indicate that an
acknowledged frameshift target in CASPASE-5 and a previously
unidentified target in FANCD2 are mutated in treatment-related
MSI+ AML cases. Similar CASPASE-5 and FANCD2 mutations
are also frequent in established MSI+ cell lines. The absence of
similar insertion/deletions in analogous repeats within the
CHK1, FAS, FANCE, FANCG, and RB genes in established cell
lines is in general agreement with other findings (ref. 19; but
see also ref. 20) and underlines the possible significance of the
CASPASE-5 and FANCD2 changes. A NF1 intronic mononucleotide microsatellite was also identified as a frequent target of
mutation in established MSI+ cell lines and colorectal tumors.
Surprisingly, similar mutations were not found in MSI+ tAML.
CASPASE-5 was first identified as a frameshift target in
MSI+ endometrial and gastrointestinal tumors (21) and singlebase deletions were present in 67% of our MSI+ tumor cell lines
Table 5. FANCD2 Allele Distribution in MicrosatelliteStable Samples
T10/T10
T10/T8
T8/T8
Total
Normal
Lymphoblastoid
Cell Lines
Blood
Donors
MMR+
Tumor Cell
Lines
MMR+
8
2
1
11
5
5
0
10
3
3
0
6
12
7
0
19
Total
Colon
Carcinomas
28
17
1
46
Mol Cancer Res 2005;3(5). May 2005
Frameshift Targets in Therapy-Related AML
FIGURE 4. FANCD2 genotype and mRNA expression. A. Genotype of
MMR-proficient cells and individuals. The T10 FANCD2 repeat was
analyzed by genotyping (left ) or sequencing (right ). Four examples of
DNA from normal individuals (blood donors, BD1-BD4 ) and the MSI cell
line (Raji ) are shown. Data have been aligned and the vertical lines
represent the lengths expected from 301 nucleotides (T10) or 299
nucleotides (T8). B. Genotype of MMR-defective cell lines. Four examples
of DNA from MSI+ cell lines are shown. C. FANCD2 mRNA. A fragment of
the FANCD2 transcript spanning exons 5 to 9 and a h-actin mRNA
fragment were coamplified by RT-PCR. The PCR products were
separated on a 2% agarose gel and visualized with ethidium bromide.
Arrow, position of a minor cDNA product that was obtained from all cell
lines assayed.
that included several of colorectal origin. About 20% of our
MSI+ tAML cases had similar CASPASE-5 mutations in broad
agreement with other reports (22, 23), although Olipitz et al.
(24) found no CASPASE-5 mutations in five therapy-related
Mol Cancer Res 2005;3(5). May 2005
leukemias. The precise function of the CASPASE-5 protein is
not known. Although they are involved in apoptosis, proteases
of the caspase group to which CASPASE-5 belongs seem to be
involved in regulating inflammatory responses (25). The
frequent occurrence of frameshifts in the A10 repeat of
CASPASE-5 in MSI+ tumors suggests that CASPASE-5 may
protect against tumorigenesis. Because 20% of MSI+ tAMLs
also contained CASPASE-5 mutations, this protection also
seems to be significant in hematologic malignancy.
Type 1 neurofibromatosis (NF1) is a common autosomal
dominant disorder characterized by neurofibromas and multiple
café au lait spots. The NF1 gene is considered to be a tumor
suppressor in myeloid cells and NF1 individuals have an
estimated 200- to 500-fold elevated risk of myeloid malignancy
and myelodysplasia (reviewed in ref. 26). Concerning tAML/
MDS, it is particularly interesting that alkylating agent treatment
accelerates the development of myeloid malignancy in Nf1
heterozygous mice (27). Defective MMR is associated with
NF1. Homozygosity for the MSH2 or MLH1 mutations that
predispose heterozygous individuals to hereditary nonpolyposis
colorectal carcinoma in adulthood results in a childhood
condition with the clinical features of NF1 and hematologic
malignancies (28-30). This suggests that NF1 might be
particularly susceptible to mutation in MMR-deficient cells.
Consistent with this view, NF1 mutations were found to be more
frequent in MSI+ than in microsatellite-stable tumors (31). In
addition, a MMR defect significantly accelerates the development of myeloid leukemia in Nf1 heterozygous mice (32). We
examined three potential frameshift targets: the short exonic C7
and T6 repeats and the T26 intronic microsatellite. No mutations
were found in the short repeats, although a single-base deletions
have been reported in the MSI+ colorectal carcinoma HCT116
(-C) and LoVo (-T) cell lines (31). (We did not observe the T
deletion in our LoVo cells, suggesting that the reported mutation
had occurred during culture in vitro and was not present in the
original tumor.) The T26 repeat was a frequent mutational target
in MSI+ cell lines and colorectal carcinomas. RT-PCR provided
no evidence of mRNA splicing defects in cell lines with a
mutated T26, although this analysis does not preclude the
possibility of subtle effects. Surprisingly, in view of the
association between defective MMR and neurofibromatosis,
all MSI+ AML/MDS cases we examined retained wild-type NF-1.
Fanconi anemia is an autosomal recessive disorder comprising eight complementation groups. In addition to lifethreatening pancytopenia, Fanconi anemia individuals have a
very high risk of developing AML. Cells from Fanconi anemia
patients are extremely sensitive to DNA cross-linking agents,
such as mitomycin C or diepoxybutane (for recent reviews, see
refs. 33, 34). There are interesting parallels between the AML
in Fanconi anemia patients and tAML/MDS. Cytogenetic
abnormalities (5q-, monosomy 7, and 7q-) commonly found
in AML of Fanconi anemia patients (35) are strikingly similar
to those found in tAML/MDS. These changes are relatively
infrequent in de novo AML (33, 34). In addition, AML in
Fanconi anemia patients is generally preceded by a myelodysplastic phase (36), again more closely resembling tAML/MDS
rather than de novo disease. One possible implication of these
similarities is that acquired defects in the Fanconi anemia
pathway of DNA damage signaling and repair might contribute
257
258 Offman et al.
to the development of tAML/MDS. Because the majority of
tAML/MDS cases are MSI+, we examined putative frameshift
targets in Fanconi anemia genes. Although no mutations were
identified in a C7 repeat in FANCE or a (TG)5 repeat in FANCG
in any of the MSI+ cell lines, frequent addition/deletion
mutations were found in the intronic T repeat of FANCD2.
Although FANCD2 is clearly a target for mutation in MSI+
tumors, mRNA splicing, drug sensitivity, and DNA damage –
dependent FANCD2 protein ubiquitination were not detectably
impaired in mutated cells. The RT-PCR method was, however,
sufficiently sensitive to detect a rare FANCD2 splice variant
that results from skipping exon 8. We note that none of the cell
lines were homozygous for the mutant T9 allele, although this
genotype was observed in tumors. It would be of interest to
examine mRNA expression, drug sensitivity, and FANCD2
ubiquitination in cells with the T9 genotype.
The FANCD2 T8 variant seems to be a previously unidentified
polymorphism. Among normal individuals, MMR-proficient
tumor cell lines, and colon cancers (46 samples in total), the
frequencies of the T10 and T8 alleles were f0.8 and 0.2,
respectively. We found only a single example of apparent
homozygosity for T8 among non-MSI+ cells. Because this was in
an EBV-transformed lymphoblastoid cell line that has been
maintained in culture for many years, the true frequency of the T8/
T8 genotype in the normal population remains to be determined.
In summary, we have identified targets for insertion/deletion
mutations in MSI+ AML/MDS. One of these is an acknowledged target in the CASPASE-5 gene. The other is an intronic
repeat within FANCD2 in which we also identified a rare
polymorphism. Mutations in the FANCD2 repeat were
identified in MSI+ AML/MDS cases as well as in colorectal
tumors and cell lines. A more limited analysis indicated that an
intronic T26 microsatellite in NF1 is a common target in MSI+
colorectal carcinomas but not in tAML/MDS. Whether this
reflects selection against NF-1 mutations in developing
myeloid malignancy is an interesting possibility confirmation
of which awaits analysis of more MSI+ AML/MDS cases. None
of the intronic mutations caused severe impairment of mRNA
splicing and no gross alterations in sensitivity to a DNA crosslinking agent could be ascribed to heterozygosity for mutated
FANCD2 or homozygosity for the rare allele. The possibility of
subtle defects resulting from two mutated FANCD2 alleles or
singly in combination with hypomorphic mutations or polymorphisms in other DNA repair functions is not excluded,
however, and might repay further investigation.
Materials and Methods
Patient Material
Genomic DNA was obtained from chemotherapy-related
(14) and transplant-related AMLs (15) as described previously.
FIGURE 5. FANCD2 functionality. A. Mitomycin C sensitivity. Exponentially growing Raji (left ), Molt4 (center), or REH (right ) cells were treated with
mitomycin C and cell growth was monitored by cell counts at 24-hour intervals as shown. Cells were diluted as appropriate to maintain exponential growth. y,
untreated; n, 10 nmol/L; E, 30 nmol/L; , 100 nmol/L; *, 300 nmol/L; ., 1,000 nmol/L. B. FANCD2 ubiquitination. Extracts prepared from cells unirradiated or
following exposure to 15 Gy ionizing radiation (IR ) were analyzed by Western blotting.
Mol Cancer Res 2005;3(5). May 2005
Frameshift Targets in Therapy-Related AML
Cell Lines
All cell lines used were obtained from Cancer Research UK
London Research Institute Central Cell Repository. The MMRdefective variants A2780-MNUcl1 (37) and RajiF12 (38) have
been described previously.
Sensitivity to Mitomycin C
Exponentially growing cells were seeded at a concentration of
5 105/mL in 24-well plates and grown in the continuous
presence of mitomycin C. Cell numbers were determined by
daily hemocytometer counts. Throughout the experiment,
exponential growth was maintained by appropriate dilution into
growth medium.
Preparation of Total RNA
Cultured cells were harvested during exponential growth.
RNA was prepared using a Qiagen (Crawley, United Kingdom)
RNeasy Mini kit according to the manufacturer’s instructions.
Samples were homogenized using Qiashredders (Qiagen). A
DNase digestion step was included using the Qiagen RNaseFree DNase Set according to the manufacturer’s instructions.
Target Gene Analysis
Genotyping. The CASPASE-5, CHK1, FAS, RB, NF1,
FANCD2 , FANCE, and FANCG repeat sequences were
amplified by PCR using fluorescent primers (Table 1). The
sense primer of each primer set was FAM labeled at its 5V
terminus. Fragments were amplified as multiplex PCRs:
CASPASE-5 and FANCD2; CHK1, FAS, and RB; FANCE and
FANCG; or individually (NF1). Genomic DNA (10 pg-20 ng)
was amplified with 10 pmol of each primer. After initial
denaturation at 95jC for 15 minutes, 30 to 40 amplification
cycles were followed by a final extension at 72jC for
10 minutes. Each amplification cycle was 30 seconds at 94jC,
90 seconds at 55jC to 63jC, and 60 seconds at 72jC. For DNA
from clinical samples, the annealing time was increased from
90 seconds to 3 minutes. Products were analyzed either on an
ABI Prism 377 sequencer or an ABI Prism 3100 capillary
system. Data were analyzed using the Genotyper 1.1 software.
DNA Sequencing. PCR reactions were done as described
above but using nonfluorescent primers. PCR products were
separated by electrophoresis through 2% agarose gels. The PCR
products were excised, purified with the Qiagen Gel Purification kit, and sequenced using the BigDye Terminator version
3.1 Cycle Sequencing Kit (Applied Biosystems, Warrington,
United Kingdom) according to the manufacturer’s instructions.
Data were analyzed using the Genotyper 1.1 software.
Reverse Transcription-PCR
Total cellular RNA (1 Ag) was reverse transcribed in
ImProm-II Reaction Buffer (Promega, Southampton, United
Kingdom) containing 20 units RNasin RNase inhibitor
(Ambion, Huntington, United Kingdom), 10 mmol/L deoxynucleotide triphosphates, 3 mmol/L MgCl2, 0.5 Ag pd(N)6
random hexamer primers (Amersham Pharmacia Biotech,
Amersham, United Kingdom), and 1 AL ImProm-II Reverse
Transcriptase according to the manufacturer’s instructions.
FANCD2 cDNA was coamplified with h-actin as a positive
Mol Cancer Res 2005;3(5). May 2005
control. The reverse-transcribed product was amplified in 1
Multiplex PCR Master Mix containing 10 pmol appropriate
primer pairs. After initial denaturation at 95jC for 15 minutes, 35
cycles of amplification were followed by a final extension at
72jC for 10 minutes. Each cycle was 30 seconds at 94jC, 90
seconds at 57jC, and 60 seconds at 72jC. The NF1 and FANCD2
products were purified and sequenced as described above.
Acknowledgments
We thank our colleagues in the Cancer Research UK London Research Institute
Oligonucleotide Synthesis, Cell Services, and Equipment Park for their
invaluable help.
References
1. Peltomäki P. Role of DNA mismatch repair defects in the pathogenesis of
human cancer. J Clin Oncol 2003;21:1174 – 9.
2. Jiricny J, Nyström-Lahti M. Mismatch repair defects in cancer. Curr Opin
Genet Dev 2000;10:157 – 61.
3. Ionov Y, Peinado MA, Malkhosyan S, Shibata D, Perucho M. Ubiquitous
somatic mutations in simple repeated sequences reveal a new mechanism for
colonic carcinogenesis. Nature 1993;363:558 – 61.
4. Aaltonen LA, Peltomäki P, Leach FS, et al. Clues to the pathogenesis of
familial colorectal cancer. Science 1993;260:812 – 6.
5. Lengauer C, Kinzler KW, Vogelstein B. Genetic instability in colorectal
cancers. Nature 1997;386:623 – 6.
6. Markowitz S, Wang J, Myerhoff L, et al. Inactivation of the type II TGF-h
receptor in colon cancer cells with microsatellite instability. Science
1995;268:1336 – 8.
7. Rampino N, Yamamoto H, Ionov Y, et al. Somatic frameshift mutations in the
BAX gene in colon cancers of the microsatellite mutator phenotype. Science
1997;275:967 – 9.
8. Kim N-G, Choi YR, Baek MJ, et al. Frameshift mutations at coding
mononucleotide repeats of the hRAD50 gene in gastrointestinal carcinomas with
microsatellite instability. Cancer Res 2000;61:36 – 8.
9. Giannini G, Ristori E, Cerignoli F, et al. Human MRE11 is inactivated in
mismatch repair-deficient cancers. EMBO Rep 2002;3:248 – 54.
10. Bignami M, Casorelli I, Karran P. Mismatch repair and response to DNAdamaging antitumor therapies. Eur J Cancer 2003;39:2142 – 9.
11. Ben-Yehuda D, Krichevsky S, Caspi O, et al. Microsatellite instability and
p53 mutations in therapy-related leukemia suggest mutator phenotype. Blood
1996;88:4296 – 303.
12. Das-Gupta EP, Seedhouse CH, Russell NH. Microsatellite instability occurs
in defined subsets of patients with acute myeloblastic leukaemia. Br J Haematol
2001;114:307 – 12.
13. Sheikhha MH, Tobal K, Yin JAL. High level of microsatellite instability but
not hypermethylation of mismatch repair genes in therapy-related and secondary
acute myeloid leukaemia and myelodysplastic syndrome. Br J Haematol
2002;117:359 – 65.
14. Casorelli I, Offman J, Mele L, et al. Drug treatment in the development of
mismatch repair defective acute leukemia and myelodysplastic syndrome. DNA
Repair 2003;2:547 – 59.
15. Offman J, Opelz G, Doehler B, et al. Defective DNA mismatch repair in
acute myeloid leukemia/myelodysplastic syndrome after organ transplantation.
Blood 2004;104:822 – 8.
16. Karran P, Offman J, Bignami M. Human mismatch repair, drug-induced
DNA damage, and secondary cancer. Biochimie 2003;85:1149 – 60.
17. Faulkner RD, Seedhouse CH, Das-Gupta EP, Russell NH. BAT-25 and
BAT-26, two mononucleotide microsatellites, are not sensitive markers of
microsatellite instability in acute myeloid leukaemia. Br J Haematol 2004;124:
160 – 5.
18. Garcia-Higuera I, Taniguchi T, Ganesan S, et al. Interaction of the Fanconi
anemia proteins and BRCA1 in a common pathway. Mol Cell 2001;7:249 – 62.
19. Abdel-Rahman WM, Arends MJ, Wyllie AH, Morris RG, Ramadan ME.
Death pathway genes Fas (Apo-1/CD95) and Bik (Nbk) show no mutations in
colorectal carcinomas. Cell Death Differ 1999;6:387 – 8.
20. Yamamoto H, Gil J, Schwartz S, Perucho M. Frameshift mutations in Fas ,
Apaf-1 , and Bcl-10 in gastro-intestinal cancer of the microsatellite mutator
phenotype. Cell Death Differ 2000;7:238 – 9.
21. Schwartz S, Yamamoto H, Navarro M, Maestro M, Reventos J, Perucho M.
259
260 Offman et al.
Frameshift mutations at mononucleotide repeats in caspase-5 and other target
genes in endometrial and gastrointestinal cancer of the microsatellite mutator
phenotype. Cancer Res 1999;59:2995 – 3002.
30. Gallinger S, Aronson M, Shayan K, et al. Gastrointestinal cancers and
neurofibromatosis type 1 features in children with a germline homozygous MLH1
mutation. Gastroenterology 2004;126:576 – 85.
22. Takeuchi S, Takeuchi N, Fermin AC, Taguchi H, Koeffler HP. Frameshift
mutations in caspase-5 and other target genes in leukemia and lymphoma cell
lines having microsatellite instability. Leuk Res 2003;27:359 – 61.
31. Wang Q, Montmain G, Ruano E, et al. Neurofibromatosis type 1 gene as a
mutational target in a mismatch repair-deficient cell type. Hum Genet 2003;
112:117 – 23.
23. Scott S, Kimura T, Ichinohasama R, et al. Microsatellite mutations of
transforming growth factor-b receptor type II and caspase-5 occur in human
precursor T-cell lymphoblastic lymphomas/leukemias in vivo but are not
associated with hMSH2 or hMLH1 promoter methylation. Leuk Res 2003;27:
23 – 34.
32. Gutmann DH, Winkeler E, Kabbarath O, et al. Mlh1 deficiency accelerates
myeloid leukemogenesis in neurofibromatosis 1 (Nf1) heterozygous mice.
Oncogene 2003;22:4581 – 5.
24. Olipitz W, Hopfinger G, Aguiar RCT, et al. Defective DNA mismatch repair:
a potential mediator of leukemogenic susceptibility in therapy-related myelodysplasia and leukemia. Genes Chromosomes Cancer 2002;34:243 – 8.
34. Alter BP. Cancer in Fanconi anemia, 1927-2001. Cancer 2003;97:425 – 40.
25. Martinon M, Tschopp J. Inflammatory caspases: linking an intracellular
innate immune system to autoinflammatory diseases. Cell 2004;117:561 – 74.
26. Dasgupta B, Gutmann DH. Neurofibromatosis 1: closing the GAP between
mice and men. Curr Opin Genet Dev 2003;13:20 – 7.
27. Mahgoub N, Taylor BR, Beau MML, et al. Myeloid malignancies induced by
alkylating agents in Nf1 mice. Blood 1999;93:3617 – 23.
33. D’Andrea AD, Grompe M. The Fanconi anemia/BRCA pathway. Nat Rev
Cancer 2002;3:23 – 34.
35. Thurston VC, Ceperich TM, Vance GH, Heerema NA. Detection of
monosomy 7 in bone marrow by fluorescence in situ hybridization. A study of
Fanconi anemia patients and review of the literature. Cancer Genet Cytogenet
1999;109:154 – 60.
36. Alter BP, Caruso JP, Drachtman RA, Uchida T, Velagaleti GVN, Eleghetany
MT. Fanconi anemia: myelodysplasia as a predictor of outcome. Cancer Genet
Cytogenet 2000;117:125 – 31.
28. Ricciardone MD, Ozcelik T, Cevher B, et al. Human MLH1 deficiency
predisposes to hematological malignancy and neurofibromatosis type 1. Cancer
Res 1999;59:290 – 3.
37. Branch P, Masson M, Aquilina G, Bignami M, Karran P. Spontaneous development of drug resistance: mismatch repair and p53 defects in resistance to cisplatin in human tumor cells. Oncogene 2000;19:
3138 – 45.
29. Whiteside D, McLeod R, Graham G, et al. A homozygous germ-line
mutation in the human MSH2 gene predisposes to hematological malignancy and
multiple cafe-au-lait spots. Cancer Res 2002;62:359 – 62.
38. Branch P, Aquilina G, Bignami M, Karran P. Defective mismatch binding
and a mutator phenotype in cells tolerant to DNA damage. Nature 1993;362:
652 – 4.
Mol Cancer Res 2005;3(5). May 2005