Introduction

Cancer of the pancreas is one of the most common cancer entities in human. Owing to the silent course of cancer development in the peritoneal cavity, this tumor gets diagnosed only at far advanced stages. This has hampered description of the cascade of genetic lesions that lead to final tumor development (Hruban et al., 2000; Bardeesy et al., 2001). To be able to investigate the different steps of genetic alterations occuring during cancer progression, a transgenic mouse tumor model was established, in which TGF is expressed under the regulation of the elastase promoter. This ensures TGF overexpression restricted to pancreatic acinar cells (Sandgren et al., 1990, 1991). It has already been shown that in this tumor model, the Ras/ERK pathway is constitutively active (Wagner et al., 2001). After a latency period for more than 1 year, this leads to pancreatic tumor development. When crossed into a background of p53^+/- heterozygous mice, the latency for tumor induction is reduced in a manner which turns this tumor model feasible for laboratory investigation. The tumors bear similar genetic changes, as were observed in human pancreatic cancer, which recommends this system as an in vivo model for pancreatic cancer (Wagner et al., 2001).

In this study, we describe the genomic changes in the cell line TD2 originating from this genetically engineered tumor model. A pattern of specific chromosomal aberrations that can be found in the original tumor tissue (Schreiner et al., submitted) was detected. These changes include gain of proximal chromosome 11, encompassing epidermal growth factor receptor (Egfr), combined with deletion of the distal part of the same chromosome, encompassing the wild-type p53 locus and hypermethylation of the p16^Ink4a locus. Further recurrent genetic changes were observed as gain of part of chromosome 15, where c-Myc has been mapped, or – alternatively – loss of distal chromosome 14 including the Rb1 locus. Thus, the cell line TD2 and the mouse tumor progression model represent powerful tools for future studies of pancreatic tumor growth in vitro and in vivo.

Results

Cytogenetic studies

Chromosome analysis of the TD2 cell line was performed at passage 23 and 103. No striking differences were noticed. The chromosome number ranged between 42 and 80. Of all the analysed metaphases, 87% were hypertriploid to hypotetraploid, containing 65–77 chromosomes. In all metaphases, one large marker chromosome with distinct bands in the proximal half and a varying number of several small marker chromosomes of uncertain origin were observed (Figure 1).

Figure 1.

Karyogram of a representative metaphase spread of cell line TD2 (passage 20) hybridized for SKY. Inverted DAPI stained and classified chromosomes in pseudocoloration are shown. Marker chromosomes are indicated by M

Full figure and legend (265K)

Analysis by spectral karyotypinq (SKY)

Spectral karyotyping (SKY) was performed at passage 20. One representative karyotype is shown in Figure 1, consisting of an inverted DAPI-banded karyotype together with the classified images in pseudocoloration. The chromosomal aberration pattern consists mostly of numerical aberrations. All chromosomes are represented three to four times except X and Y chromosomes, and chromosome 17, which is present in five copies. Besides several small marker chromosomes of varying number and undefined origin, one additional large marker chromosome is found in all metaphases. The proximal part of this marker chromosome is of chromosome 11 origin and consists of three conspicuous dark-staining bands as already noticed in G-banded metaphases. The distal part of the marker derives from chromosome 5. Conventional banding and SKY analysis revealed this marker chromosome together with the small unidentifiable marker chromosomes as the only structurally rearranged chromosomes.

CGH analysis

To allow comparison of the cell line with the original tumor sample, CGH analysis was performed with DNA extracted from the cell line at passage 25. Though the mean chromosome number indicates a highly aberrant karyotype, only two gains which surpassed the cutoff level of 1.25 are detected (Figure 2a). Almost the whole chromosome 17 and the proximal part of chromosome 11 are amplified. For MMU 15, a very specific CGH profile is noticed with gain of chromosomal material at the beginning of the second half of the chromosome.

Figure 2.

CGH analysis and real-time PCR of cell line TD2. (a) CGH: only chromosomes with conspicuous profiles are shown. Green lines on the left of the chromosome profile represent losses, red lines on the right indicate gains of chromosomal material. For each chromosomal profile, at least 20 chromosomes were calculated. (b) Quantitation by real-time PCR. The relative amplification rates of the tested genes of proximal chromosome 11 (dark gray) and chromosome 15 (light gray) are shown. The Gcnt2 gene was taken as internal reference

Full figure and legend (116K)

Compared to the original tumor sample of that cell line (not shown) the aberration pattern is widely confirmed. In the original tumor sample, an additional loss of distal chromosome 11 including the p53 locus is observed. In contrast to these findings, the CGH profile of distal MMU 11 of TD2 cells does not reach the cutoff level (Figure 2a). However, the loss of the wild-type p53 allele on MMU 11 in TD2 cells was proven by independent PCR analysis (not shown).

Calibration of DNA copy changes by real-time PCR

To define more exactly the length of the amplification unit and the degree of amplification, calibration was performed by real-time PCR (Figure 2b). For chromosome 11, the following genes mapping to the proximal 50 Mb of MMU 11 were tested: Egfr (9 cM; 16.8 Mb); reticuloendotheliosis oncogene, c-Rel (13 cM; 23.8 Mb); serine/threonine kinase 10, Stk10 (16 cM; 35.6 Mb); and FMS-like tyrosine kinase 4, Flt4 (25 cM; 50.0 Mb). Except Flt4, all analysed genes showed a relative amplification of 2.2–3.0 times. Flt4, which maps at 50 Mb, is not included in the amplification unit. Therefore, the amplification unit extends at least 35.6 Mb but at most 49 Mb.

From MMU 15, the genes myelocytomatosis oncogene, c-Myc (32.0 cM; 62.4 Mb), adenylate cyclase 8, Adcy8 (37.5 cM; 65.1 Mb) and protein tyrosine kinase 2, Ptk2 (42.0 cM; 73.7 Mb) were chosen. The copy number of the genes Adcy8 and Ptk2 is not affected in this cell line, but for c-Myc 15–16 times amplification was detected (Figure 2b).

Direct fluorescence in situ hybridization (FISH) analysis with gene-specific probes

Real-time PCR and CGH verified the amplification of the c-Myc locus, although this has not been observed by conventional cytogenetic and SKY analysis (Figure 1). To detect the localization of the amplified c-Myc gene, fluorescence in situ hybridization (FISH) with c-Myc-specific BAC probes (RPCI23-235H07, RPCI23-98D8, RPCI23-55F11) was performed. Many strong signals outside the chromosomes were observed (Figure 3a). These signals indicate the presence of double minutes (DMs) which are too small to get detected by conventional DAPI banding. Another type of c-Myc amplification is observed by hybridization of the same BAC probes to metaphases prepared from higher passage cultures. Metaphases with a homogeneous staining region (HSR) of c-Myc DNA existed in parallel with metaphases, in which c-Myc was present as DMs (Figure 3b). DM appeared in about 46% of all metaphases and a HSR is noticed in 54%.

Figure 3.

Hybridization of c-Myc-specific BACs (RPCI23-98D8; RPCI23-55F11; RPCI23-235H7) on metaphase spreads from cell line TD2. The c-Myc locus is amplified in the form of DMs (a), or an HSR chromosome (b). In (c) the endogenous c-Myc locus is shown in a partial metaphase

Full figure and legend (267K)

The original FISH signal from the endogenous c-Myc locus on MMU 15 is also detectable, but in metaphases with HSRs it is too weak to get detected simultaneously with the brilliant HSR signal. Therefore, a partial metaphase of TD2 is shown in Figure 3c.

The integration of the HSR does not occur on MMU 15 which harbors the endogenous c-Myc locus, but on another chromosome. By use of whole-chromosome painting probes, WCPs, the HSR-bearing chromosome was revealed as chromosome 6 in most cases. As is shown in Figure 4a, b, d and f, the integration of the HSR on chromosome 6 occurred at variable positions. In a minority of metaphases, the HSR is even integrated in further, hitherto unknown chromosomes (Figure 4c, e). The use of the WCP 6 probe further made clear that some chromosomal material of chromosome 6 is included in the DMs and is found interspersed in some HSRs (Figure 4b, c, g).

Figure 4.

Most frequent chromosomal markers in the TD2 cell line containing c-Myc amplification. Top column: schematic illustration of the different markers. The six most frequent marker chromosomes (a–f) and DM chromosomes (g) are shown. Green: chromosomal material from chromosome 6; Red: c-Myc-HSR. Middle column: FISH with WCP 6 (green signal) and c-Myc-specific BAC-clones (red signal). Of all metaphases, 66% with HSR marker chromosomes contain marker A, 13% marker B or C and 21% contain marker D, E, F or some more seldom variants. Also, in DMs, some residual DNA from chromosome 6 is detectable (g)

Full figure and legend (214K)

The analysis of the amplification unit on MMU 15 by real-time PCR revealed that genes mapping next to c-Myc are not amplified. To determine the length of the amplification unit, overlapping BACs from a contig of MMU 15D2-3 of the Ensembl Mouse Genome Server, which includes c-Myc-specific BACs, were hybridized to TD2 metaphases (Figure 5). According to these FISH analyses, the amplification unit spans about 1.65 Mb from 61.65 to 63.3 Mb. All BACs mapping between RPCI23-386K4 and RPCI23-23L17 were amplified together with the c-Myc locus. No difference with respect to the length of the amplification unit on DMs or HSR was observed.

Figure 5.

BACs from library RPCI23 or RPCI24 were chosen from contig 15006 (Ensembl mouse genome Server V. 7.3a.1) and aligned from centromere to telomere. BAC clones positive for c-Myc are indicated in red. BAC clones shown in green and the c-Myc-specific BAC clones (red) are included in the amplification unit on DMs and the HSR. BACs shown in blue are not included in the amplification unit

Full figure and legend (65K)

Also, the length of the amplification unit from proximal chromosome 11 was defined by analysis with chromosome-specific BAC probes. Probes mapping from the centromere down to 50 Mb were hybridized on metaphase spreads. Signals on the marker chromosome were detected from BACs RPCI23-282P12 (4 Mb) to RPCI23-273P11 (36.1 Mb) including Egfr, c-Rel and Stk10. All of these BAC probes revealed the same number of repetitions on the same marker chromosome (Figure 6). On elongated chromosomes, five signals were observed. Flt4-specific probes revealed signals only on the normal chromosome 11, but not on the marker chromosome. Therefore, Flt4 at 50.0 Mb is not part of the amplification unit. As shown in Figure 6 and illustrated in the accompanying scheme, five large amplification units in inverted order are integrated in tandem in the marker chromosome.

Figure 6.

SKY analysis and FISH of the typical marker chromosome of cell line TD2. (a) SKY: with pseudo-coloration, the different origin of chromosomal material of this marker could be classified. (b) Inverted DAPI staining image. (c) FISH with specific BAC probes for c-Rel (RPCI23-207N19; green) and Stk10 (RPCI23-47J10; red) together with a schematic drawing. Arrows on the right side of the scheme illustrate the head-to-head and tail-to-tail orientation of the amplification unit

Full figure and legend (139K)

Expression studies

The genes found amplified, Egfr and c-Rel from MMU 11 and c-Myc from MMU 15, were analysed for their state of gene expression by Northern blot analysis. A respective Northern blot is shown in Figure 7a–e. Only a basic expression level is detected for c-Rel (Figure 7d). Egfr is expressed at a low level, 1.8 times more than in NIH3T3 cells (Figure 7c), and 3.5 times when compared with wild-type pancreas (not shown). c-Myc is overexpressed three times compared to NIH3T3 cells (Figure 7a).

Figure 7.

(a–e) Gene expression tested by Northern blot analysis of c-Myc (a), Egfr (c), c-Rel (d) and -actin (b, e) as control and NIH3T3 cells as reference. (f) Analysis of the methylation status of the p16^Ink4a-promoter region. Primers for the methylated and unmethylated sequences were used for cell line TD2, normal pancreatic cells (NP) and normal DNA which was not treated with bisulfite (NPnt); (N) negative control. (g) EMSA of TD2 cells showing NF B/Rel activity (lane 1). By supershift assay using antibodies against RelA and NFB1 binding specificity was confirmed

Full figure and legend (74K)

The p16^INK4a locus has been suggested to play a specific role during progression of pancreatic adenocarcinomas (Caldas et al., 1994). However, CGH and LOH analyses in TD2 cells were unconspicuous for this locus on chromosome 4. Analysis of the methylation status of the p16^Ink4a promoter region using a methylation-specific PCR assay showed hypermethylation of this sequence (Figure 7f), as was already observed in the original tumor sample. Also, some unmethylated p16^Ink4a promoter sequences were detectable by a nested PCR procedure specific for the unmethylated promoter sequence (Figure 7f). To prove the successful conversion of cytosin to uracil residues after bisulfite treatment, a primer pair for the unmodified sequence was used and no amplification product obtained (data not shown).

Electromobility shift assay (EMSA)

In the large amplification unit of MMU 11, Egfr and also c-Rel is included. Therefore, the possible effects of c-Rel amplification on NF-B-binding activity were studied by performing EMSA. Nuclear extracts of TD2 cells (passage 105) were incubated with a ³²P-labeled oligonucleotide containing the sequence of the NF-B recognition site. A strong binding activity was revealed (Figure 7g). To determine to which NF-B proteins the binding activity is due, supershift assays were performed using antibodies against all members of the NF-B transcription factor family: RelA, NF-B1, NF-B2, c-Rel and RelB. A supershift band was observed for RelA and NF-B1, but not for c-Rel (Figure 7g).

Discussion

The TGF/p53^+/- transgenic mouse has been shown to be an informative tumor model for pancreatic adenocarcinoma (Wagner et al., 2001). During analysis of the secondary genetic changes, which are induced during tumor progression, it was surprising to observe that this genetically modified system canalyses a relative strict cascade of consecutive genetic changes (Schreiner et al., submitted). A common finding is amplification of proximal chromosome 11 including the TGF receptor Egfr, combined with deletion of distal chromosome 11 including the p53 locus and hypermethylation of the p16^Ink4a locus on chromosome 4. This specific chromosomal aberration pattern is found complemented with amplification of the c-Myc locus on chromosome 15. From one of the original tumors, the cell line TD2 was established (Greten et al., 2002).

As proven by the combined molecular–cytogenetic analysis, the original pattern of chromosomal changes has been preserved (Figure 2). According to PCR analysis also the wild-type p53 locus is deleted, though the CGH profile did not reach the cutoff level for the deletion of distal chromosome 11. In a minority of the original tumor samples, LOH for the p16^Ink4a locus was present; in other tumors, p16^Ink4a was inactivated by promoter methylation (Schreiner et al., submitted). In the TD2 cell line, the regulatory 5' region of the p16^Ink4a locus is hypermethylated, though upon thorough analysis some unmethylated p16^Ink4a alleles were also found (Figure 7f). p53 and p16^Ink4a deletion/inactivation are combined in the same tumor cell line. This cell line reflects all genetic alterations observed in the primary tumor.

It was surprising to observe mostly numerical chromosomal changes. Without regard to the unidentifiable small marker chromosomes (Figure 1), only one aberrant chromosome was present. This unusual pattern of changes with a low number of structural chromosomal aberrations has been observed in further murine tumor cell lines (Weaver et al., 1999; Wu et al., 2002). Whereas in human pancreatic adenocarcinoma cell lines various structural chromosomal aberrations are present (Curtis et al., 1998), in murine tumors nondisjunction seems to be the preferred type of chromosomal aberration. In spite of this, the pattern of genetic changes, as determined by CGH, is very similar when human and mouse cell lines are compared.

It is shown here that only the combined analysis by cytogenetic and molecular techniques is sufficient to detect all relevant tumor-specific chromosomal changes. The classical cytogenetic banding analysis and also the elegant SKY analysis did not reveal c-Myc amplification. This became evident only by CGH analysis and here again, only due to the fact that amplification was really high. Bei real-time PCR c-Myc amplification up to 15–16 times was observed. On a nearly triploid chromosomal background – as is given for cell line TD2 – about 45–48 extra copies of c-Myc are present. By FISH analysis with c-Myc-specific genomic probes, the c-Myc amplification became visible in the form of DMs, or as an HSR (Figure 3). The DMs are far too small to get detected by conventional cytogenetic analysis and also, a HSR chromosome may be difficult to recognize as such in an aberrant mouse karyotype. Upon continuous growth, we observed that the presence of c-Myc amplification in the form of DMs or HSR is variable. At low passage, only DMs were observed. At higher passage, metaphases with an HSR chromosome dominated, but no combination of DMs and HSR chromosome in one and the same metaphase was detected (Figure 3).

The HSR integration did not take place on the original chromosome 15, but on another chromosome. This is the usual finding for this type of gene amplification (Wahl, 1989; Schwab, 1999). In most metaphases, the HSR is integrated on chromosome 6. But as is evident from Figure 4, the integration on chromosome 6 is variable. Further evidence for variation between the presence of c-Myc amplification as DMs or HSR provides hybridization of respective metaphases with the WCP 6 probe. Also on the extrachromosomal DMs the probe detects material from chromosome 6. The most simple interpretation of this unusual observation is that these DMs in their former history did integrate on MMU 6, but later switched over to the extrachromosomal DM state again. This interpretation is in contrast to the commonly held view of the emergence of DMs and HSRs (Wahl, 1989). This is seen as a one-way process from extrachromosomal episome formation, via DM development to stable integration as HSR into another chromosome. But, obviously, the c-Myc locus is subject to enhanced mobility in the cell line described here.

The length of the c-Myc amplification unit has been determined up to now only in human tumor cell lines (Maurer et al., 1987). FISH with BAC clones mapped to chromosome 15 indicates a length of the c-Myc amplification unit of 1.65 Mb. This amplification unit includes the Pvt1 locus (Figure 5). Therefore, the unit is considerably larger than the human c-Myc amplification units in the cell lines HL-60 and COLO320, which are 120–250 kb long (Maurer et al., 1987).

The organization of the amplification unit on proxi-mal chromosome 11 is quite different. Here, a large unit extending at least from Egfr at 9 cM (16.8 Mb) to Stk10 at 16.0 cM (35.6 Mb) is amplified. The amplified unit is integrated in tandem and in inverted order on a marker chromosome. This marker chromosome is made up of chromosome 11 in its proximal part and of chromosome 5 in its distal part. Therefore, Egfr amplification takes place on the endogenous chromosome, on which Egfr maps. The large amplification unit is visible even by conventional cytogenetic techniques by a periodic ladder-like banding pattern on an easily identifiable large marker chromosome (Figure 6). This marker chromosome is present in one copy and was observed in each metaphase every time during passaging the cell line.

By real-time PCR, a threefold amplification of the Egfr locus was observed (Figure 2). On the triploid chromosomal background of cell line TD2, this indicates the presence of five to six extra copies of Egfr, c-Rel and Stk10, though on the single marker chromosome only three units get detectable at first sight (Figure 1). On enforced analysis of elongated metaphase chromosomes, the green signals of the c-Rel probe and the red signals of the Stk10 probe were sometimes observed as double signals. This would be in line with an inverted order of the amplification units on the marker chromosome, as is shown in the schematic drawing (Figure 6c). An inverted order is due to an amplification mechanism, which differs from that just described for the c-Myc amplification. These differences are: (1) the amplified sequences remain on the original Egfr-encoding chromosome 11; (2) the amplification unit is extremely large, encompassing more than a cytogenetic band with at least 35.6 Mb; and (3) the degree of amplification itself is limited; only five extra copies of the Egfr unit are present. These are characteristics of an amplification model according to the breakage/fusion/bridge model, as extensively outlined by Ma et al. (1993). According to this model, amplification occurs by an intrachromosomal mechanism and leads to an inverted repeat organization of the amplification unit. This inverted organization causes the described ladder-like appearance of the FISH signals on the marker chromosome (Toledo et al., 1992; Hellman et al., 2002). It is obvious that both of these different amplification mechanisms – intrachromosomal by a breakage/fusion/bridge cycle for the Egfr region and extrachromosomal via DMs for c-Myc – are realized in the same cell line. This is only possible in a cell line with enhanced genetic instability. In cell line TD2, this is thought to be due to the loss of p53 activity at an early stage during tumor induction (Paulson et al., 1998).

It was remarkable to observe such a large amplification unit around the Egfr locus, which raises the suspicion that besides Egfr there might be another gene on proximal chromosome 11, whose activity is beneficial for tumor growth. One attractive candidate gene would be c-Rel. In pancreatic tumor tissue, NF-B has been shown to be active (Wang et al., 1999). One of the several different dimerization partners of NF-B is c-Rel (Gilmore, 1999). To prove if c-Rel or other members of this transcription factor family bind to the NF-B-affinity sequence, EMSAs were performed on nuclear extracts of TD2 cells. These assays show a constitutive activity of NF-B; however, the complex seems to consist of RelA/NF-B1 dimers (Figure 7g).

The cell line established here from murine pancreatic adenocarcinomas bears all genetic changes of the original tumor. The TD2 cell line has been shown to be also a potent inducer of pancreatic adenocarcinomas in vivo (Greten et al., 2002). In future, this cell line will prove a valuable in vitro model for tumor therapeutic studies.

Materials and methods

Establishment of the cell line TD2

The tumor cell line TD2 is derived from a pancreatic tumor of a male TGF/p53^+/- transgenic mouse (Greten et al., 2002).

Cytogenetic analysis

Chromosomes were prepared at passage no. 23, 30, 103 and 119. GTG banding was done by standard techniques. WCPs were used according to the manufacturer's protocol (Cambio Ltd, Cambridge, UK). Bacterial artificial chromosomes (BACs) of MMU 11 were chosen from the Baylor College of Medicine database: 'Mouse Chromosome 11 Bac mapping' (www.mousegenome.bcm.tmc.edu) and further gene-specific BACs from the Ensembl Mouse Genome Server (www.ensembl.org/Mus_musculus/) version 3.1.1. BACs from MMU 15D2-3 were elected according to the map provided by the Ensembl Mouse Genome Server, contig 15006. All BACs are from library RPCI-23 or RPCI-24 and purchased at the BACPAC resource center (www.chori.org/bacpac). For FISH, standard techniques were applied.

Spectral karyotyping (SKY)

The chromosome preparations were hybridized with the mouse SKY probe mixture (SkyPaint^™, Applied Spectral Imaging Ltd, Israel), according to the manufacturer's instructions. The SKY probe mixture was applied to the preparations and hybridized for 48 h in a humidified chamber. Washing and detection was applied using standard techniques. For image acquisition, the SpectraCube system SD200 including a triple-band-pass filter was connected to a fluorescence microscope (Axiophot, Zeiss, Germany), assisted by Spectral Imaging Software, version 2.5. The final analysis of the metaphases was performed with the help of the Sky View program, version 1.6.1 (Applied Spectral Imaging Ltd, Israel).

Comparative genomic hybridization (CGH)

DNA from the cell line TD2 (passage 25) and normal female mouse tissue as reference DNA were used. CGH was performed as described (Kallioniemi et al., 1992), with slight modifications (Schreiner et al., in preparation). The DNA from cell line TD2 was labeled with biotin-16-dUTP, the normal reference DNA with digoxigenin-11-dUTP, using standard nick translation.

Real-time PCR

DNA amplification was quantified using real-time PCR analysis (TaqMan, ABI PRISM 7700 Sequence Detection System, PE Applied Biosystems, Norwalk, CT, USA). PCR reaction, denaturation at 95°C for 2 min followed by 40 cycles of 95°C for 15 s and 60°C for 1 min (Sybr Green PCR Core Reagents, PE Applied Biosystems) were performed using genomic DNA extracted from TD2 cells as well as normal DNA as reference. The relative quantitation of the target DNA level was calculated with the comparative C_T method according to the manufacturer's protocol.

Electromobility shift assay (EMSA)

Nuclear extracts were prepared from TD2 cells according to standard techniques. Protein concentration was determined by Bradford assay (Bio-Rad Laboratories, Munich). The oligonucleotide used for binding reaction contained the high-affinity B sequence of mouse as present in the -light-chain enhancer. The oligonucleotide sequence, EMSA and supershift assays are described elsewhere (Steinle et al., 1999).

Northern blot analysis

Total cellular RNA was isolated from TD2 cells at cell culture passage 86 and 136, as well as from NIH3T3 cells with the Qiagen RNeasy purification kit. About 25 g RNA was used for Northern blot analysis, which was conducted according to standard protocol. Probes for c-Myc, Egfr and c-Rel were generated by PCR. The -actin probe was purchased from Clontech, CA, USA. All probes were labeled with the rediprime II random prime labeling system (Amersham Biosciences, UK) in the presence of [-³²P]dCTP. Autoradiographs were obtained by exposing the membrane for 8–24 h at -70°C. The signals were detected by Fujifilm FLA-300 and BAS Reader (V 2.26, raytest, Straubenhardt, Germany) and quantified by Aida Image Analyzer (V 2.31).

R16^ink4a methylation assay

Total genomic DNA (700 ng) was restricted with EcoRI at 37°C for at least 6 h. The bisulfite treatment was performed as described by Hajkova et al. (2002). The methylated sequence was amplified with the primers p16-M1 (5'CGA TTG GGC GGG TAT TGA ATT TTC GC 3') and p16-M2 (5'CAC GTC ATA CAC ACG ACC CTA AAC CG 3') (Bardeesy et al., 2002). To amplify the unmethylated allele, a nested PCR reaction was necessary. First, the PCR reaction was performed with primers FM6 (5'TTT TTA GAG GAA GGA AGG AGG GAT TT 3') and p16-UB (5'ACC CAC ACA TCA ACA CAA CCC TAA 3') (Patel et al., 2000). In the second PCR assay, the nested primers Un1-p16 (5'GTA ATT GGG TGG GTA TTG AAT TTT TGT G 3') and Un2-p16 (5'CAC ACA TCA TAC ACA CAA CCC TCC ACC/T A3') were used.

References

1. Bardeesy N, Morgan J, Sinha M, Signoretti S, Srivastava S, Loda M, Merlino G and DePinho RA. (2002). Mol.Cell. Biol., 22, 635–643.
2. Bardeesy N, Sharpless NE, DePinho RA and Merlino G. (2001). Semin. Cancer Biol., 11, 201–218. | Article | PubMed | ISI | ChemPort |
3. Caldas C, Hahn SA, da Costa LT, Redston MS, Schutte M, Seymour AB, Weinstein CL, Hruban RH, Yeo CJ and Kern SE. (1994). Nat. Genet., 8, 27–32. | Article | PubMed | ISI | ChemPort |
4. Curtis LJ, Yong L, Gerbault-Sereau M, Kuick R, Dutrillaux A-M, Goubin G, Fawcett J, Cram S, Dutrillaux B, Hanash S and Muleris M. (1998). Genomics, 53, 42–55. | Article | PubMed | ISI | ChemPort |
5. Gilmore TD. (1999). Oncogene, 18, 6842–6844. | Article | PubMed | ChemPort |
6. Greten FR, Weber CK, Greten TF, Schneider G, Wagner M, Adler G and Schmid RM. (2002). Gastroenterology, 123, 2052–2063. | Article | PubMed | ChemPort |
7. Hajkova P, EI Maarri O, Engemann S, Oswald J, Olek A and Walter J. (2002). Methods Mol. Biol., 200, 143–154. | PubMed | ChemPort |
8. Hellman A, Zlotorynski E, Scherer SW, Cheung J, Vincent JB, Smith DI, Trakhtenbrot L and Kerem B. (2002). Cancer Cell, 1, 89–97. | Article | PubMed | ISI | ChemPort |
9. Hruban RH, Goggins M, Parsons J and Kern SE. (2000). Clin. Cancer Res., 6, 2969–2972. | PubMed | ISI | ChemPort |
10. Kallioniemi A, Kallioniemi OP, Sudar D, Rutovitz D, Gray JW, Waldman F and Pinkel D. (1992). Science, 258, 818–821. | Article | PubMed | ISI | ChemPort |
11. Ma C, Martin S, Trask B and Hamlin JL. (1993). Genes Dev., 7, 605–620. | PubMed | ISI | ChemPort |
12. Maurer BJ, Lai E, Hamkalo BA, Hood L and Attardi G. (1987). Nature, 327, 434–437. | Article | PubMed | ISI | ChemPort |
13. Patel AC, Anna CH, Foley JF, Stockton PS, Tyson FL, Barrett JC and Devereux TR. (2000). Carcinogenesis, 21, 1691–1700. | Article | PubMed | ISI | ChemPort |
14. Paulson TG, Almasan A, Brody LL and Wahl GM. (1998). Mol. Cell. Biol., 18, 3089–3100. | PubMed | ISI | ChemPort |
15. Sandgren EP, Luetteke NC, Palmiter RD, Brinster RL and Lee DC. (1990). Cell, 61, 1121–1135. | Article | PubMed | ISI | ChemPort |
16. Sandgren EP, Quaife CJ, Paulovich AG, Palmiter RD and Brinster RL. (1991). Proc. Natl. Acad. Sci. USA, 88, 93–97. | PubMed | ChemPort |
17. Schwab M. (1999). Semin. Cancer Biol., 9, 319–325. | Article | PubMed | ISI | ChemPort |
18. Steinle AU, Weidenbach H, Wagner M, Adler G and Schmid RM. (1999). Gastroenterology, 116, 420–430. | Article | PubMed | ISI | ChemPort |
19. Toledo F, Le Roscouet D, Buttin G and Debatisse M. (1992). EMBO J., 11, 2665–2673. | PubMed | ISI | ChemPort |
20. Wagner M, Greten FR, Weber CK, Koschnick S, Mattfeldt T, Deppert W, Kern H, Adler G and Schmid RM. (2001). Genes Dev., 15, 286–293. | Article | PubMed | ISI | ChemPort |
21. Wahl GM. (1989). Cancer Res., 49, 1333–1340. | PubMed | ISI | ChemPort |
22. Wang W, Abbruzzese JL, Evans DB, Larry L, Cleary KR and Chiao PJ. (1999). Clin. Cancer Res., 5, 119–127. | PubMed | ISI | ChemPort |
23. Weaver ZA, McCormack SJ, Liyanage M, du MS, Coleman A, Schrock E, Dickson RB and Ried T. (1999). Genes Chromosom. Cancer, 25, 251–260. | PubMed |
24. Wu Y, Renard CA, Apiou F, Huerre M, Tiollais P, Dutrillaux B and Buendia MA. (2002). Oncogene, 21, 1518–1526. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We thank Antje Kollak, Alexandra Kilian and Beate Knobl for excellent technical assistance. This work was supported by grants from the Deutsche Forschungsgemeinschaft SFB 518 (Teilprojekt B6) and IZKF Ulm (Teilprojekt C4) to RMS.