Abstract

Background and aims . Tumor suppressor genes are often located in frequently deleted chromosomal regions of colorectal cancers (CRCs). In contrast to microsatellite stable (MSS) tumors, only few loss of heterozygosity (LOH) studies were performed in microsatellite instable (MSI) tumors, because MSI carcinomas are generally considered to be chromosomally stable and classical LOH studies are not feasible due to MSI. The single nucleotide polymorphism (SNP) array technique enables LOH studies also in MSI CRC. The aim of our study was to analyse tissue from MSI and MSS CRC for the existence of (frequently) deleted chromosomal regions and tumor suppressor genes located therein. Methods and results . We analyzed tissues from 32 sporadic CRCs and their corresponding normal mucosa (16 MSS and 16 MSI tumors) by means of 50K SNP array analysis. MSS tumors displayed chromosomal instability that resulted in multiple deleted (LOH) and amplified regions and led to the identification of MTUS1 (8p22) as a candidate tumor suppressor gene in this region. Although the MSI tumors were chromosomally stable, we found several copy neutral LOHs (cnLOH) in the MSI tumors; these appear to be instrumental in the inactivation of the tumor suppressor gene hMLH1 and a gene located in chromosomal region 6pter–p22. Discussion . Our results suggest that in addition to classical LOH, cnLOH is an important mutational event in relation to the carcinogenesis of MSS and MSI tumors, causing the inactivation of a tumor suppressor gene without copy number alteration of the respective region; this is crucial for the development of MSI tumors and for some chromosomal regions in MSS tumors.

Introduction

Colorectal cancer (CRC) is still one of the major causes of death from malignant diseases; >945  000, new cases are diagnosed annually worldwide ( 1 ). Tumors with similar clinical and pathological features show significant prognostic differences. These differences may be, in part, due to underlying molecular genetic changes, which include chromosomal instability (CIN) and microsatellite instability (MSI). MSI is due to a defect in DNA mismatch repair, resulting in an accumulation of somatic alterations in nucleotide repeat sequences called microsatellites ( 2 ). The majority of MSI tumors are chromosomally stable, and the majority of microsatellite stable (MSS) tumors are chromosomally instable ( 3–5 ).

Tumors with CIN frequently exhibit aneuploidy ( 6 ) as well as allelic imbalance resulting in the inactivation of tumor suppressor genes ( 7 , 8 ). Tumor suppressor genes can be inactivated by promoter silencing as a result of methylation or dual inactivation by either mutation or a combination of mutation of one allele and loss of the retained allele [loss of heterozygosity (LOH)]. The latter is the classical definition of a CIN tumor suppressor as proposed by Knudson in 1971 (the two-hit-hypothesis). According to this hypothesis, tumor suppressor genes are inactivated by an initial mutation of the first allele followed by deletion of the second allele ( 9 , 10 ). Copy neutral LOH (cnLOH) or alternatively uniparental disomy (UPD) is a recently described mechanism resulting in the inactivation of tumor suppressor genes as well as the activation of oncogenes in MSS and MSI tumors ( 11 , 12 ). Following initial mutational inactivation or activation of one allele, the remaining wild-type allele is deleted at the same time as the mutated allele is duplicated.

We have shown that UPD seems to be involved in the activation of early-acting tumor suppressor genes in MSS ( APC , CDKN 2a) and MSI ( hMLH 1, hMSH 2, APC , CDKN 2a) CRC cell lines ( 12 ). UPD/cnLOH has also been found in polycythemia vera, breast and lung cancer, leukemia and many other tumors [( 13–15 ); reviewed in ref. 16 ].

In contrast to MSS CRCs, only few studies of chrosomomal allelic losses (LOH/cnLOH) were performed in MSI carcinomas because they are infrequent (only 15% of sporadic CRCs) and generally considered to be chromosomally stable. Furthermore, classical LOH studies were not feasible due to the MSI and by comparative genomic hybridization and chromosomal-banding analysis cnLOH is not detected. Calhoun et al. ( 17 ) demonstrated that methods depending solely upon copy number analysis would theoretically identify a maximum of only 47.2% of all forms of LOH, suggesting the levels of LOH to be underestimated markedly in studies using comparative genomic hybridization. The recently developed single nucleotide polymorphism (SNP)-array technique allows the simultaneous analysis of copy number changes and genotypes in MSI and MSS tumors, thus allowing the detection of regions with copy neutral allelic losses (cnLOH) and classical LOH. We performed 50 K SNP arrays and mutational analysis of primary tumors (16 MSS tumors and 16 MSI tumors) to test the hypothesis that tumor suppressor genes/oncogenes are frequently inactivated/activated by cnLOH in MSS and MSI CRC.

Material and methods

Patient characteristics and DNA extraction

Between July 1999 and March 2004, a total of 165 CRC tissues and corresponding normal mucosa were collected during operations performed in the Department of Surgery, University of Wuerzburg. The collection was approved by the local Ethics Committee. Specimens containing at least 50% neoplastic cells (assessed by histological examination) were snap frozen in liquid nitrogen and subsequently stored at −80°C. In all cases, fresh specimens of appending normal colon mucosa were also collected and used as matching controls. DNA was extracted from tumors and matching normal sample tissues using a QIAGEN DNeasy tissue kit (QIAGEN, Hilden, Germany). The extracted DNA was diluted to a concentration of 100 ng/ml.

MSI analysis

The National Cancer Institute microsatellite panel (BAT25, BAT26, D2S123, D5S346 and D17S250) was used for MSI analysis ( 18 ). For polymerase chain reaction (PCR), one of the primers from each primer pair was labeled with a fluorescent dye (cy-5, cy5.5; TIB MoLBioL, Berlin, Germany) that enables detection of the resulting PCR product. PCR was performed in a 25 μl reaction mixture, containing 50 nM primers, 200 μM deoxynucleoside triphosphates, 1.5 mM MgCl 2 in ×1 PCR buffer, 50 ng template DNA and 0.4 U Taq polymerase. An initial 3 min denaturation at 96°C was followed by 33 cycles of 96°C for 20 s, 58°C for 20 s and 72°C for 30 s. Electrophoretic separation of the amplified PCR products was performed using a CEQ8000 (Beckman Coulter, Krefeld, Germany). An external size-standard facilitated size determination of the PCR products. MSI-L (low MSI) was defined as occurring when one marker out of five (BAT25, BAT26, D2S123, D5S346 and D17S250) was unstable. MSI-H (high MSI) was defined as occurring when more than two markers exhibited instability.

Immunohistochemistry

To assess the expression status of the mismatch repair proteins MLH1 (MutL protein homolog 1), MSH2 (mutS homolog 2), MTUS1 (mitochondrial tumor suppressor 1), NOL7 (nucleolar protein 7), ID4 (inhibitor of DNA binding 4), SOX4 (sex determining region Y) and E2F3 (E2F transcription factor 3) in tumor and normal tissue, immunophenotyping was performed on sections of formalin-fixed and paraffin-embedded tissue. The antibodies MLH1 (dilution 1:200; BD Bioscience, Heidelberg, Germany), MSH2 (dilution 1:40; Calbiochem, Schwalbach, Germany), MTUS1 (dilution 1:500; Abnova, Heidelberg, Germany), NOL7 (dilution 1:500; Abnova), ID4 (dilution 1:100; Abcam, Heidelberg, Germany), SOX4 (dilution 1:100; Abcam) and ESF3 (dilution 1:100; Abcam) were applied using a modified streptavidin–peroxidase antiperoxidase technique ( 19 ). In the case of a mismatch repair protein loss, no nuclear staining would be found in the tumor tissue, unlike the normal tissue.

Mutational analysis

Control fragments for use in PCR–single-strand conformation polymorphism analysis were cloned from human carcinoma cell lines with known mutations or were constructed by site-directed mutagenesis. The following mutations were used as positive controls: codon 600 (GTG→GAG) in the B-RAF gene; codon 33 (TCT→TAT), codon 41 (ACC→GCC and ACC→ATC), codon 45 (TCT→CCT and TCT→TTT) and del codon 45 (deletion of the complete codon) in CTNNB1 . Mutational analyses of B-RAF , CTNNB1 and APC genes were performed as described earlier ( 20 ). Briefly, for PCR–single-strand conformation polymorphism mutational analysis of B-RAF and CTNNB1 , a sample of 20 ng genomic DNA was amplified by conventional PCR with sequence-specific primers flanking codons 582–620 and 6–76, respectively. The amplicons were purified from a 2% agarose gel. Electrophoresis of 4 μl of the denatured product was performed using a 16 × 20 cm non-denaturing polyacrylamide gel in ×1 TBE (90 mM Tris base, 90 mM boric acid, 2 mM ethylenediaminetetraacetic acid; pH 8,0) with 20 mmol/l N -(2-hydroxyethyl)piperazine-1- N ′-(2-ethanesulfonic acid). In order to detect the B-RAF mutation, a 10% polyacrylamide gel was run at 18°C with 300 vol for 4.5 h. In the case of the CTNNB1 mutation analysis, a 13% gel was used at 26°C with 250 vol for 5 h and an 8% gel at 21°C with 200 vol for 5 h. Gels were silver stained and any shifted bands were excised, reamplified, purified and sequenced. For the mutational analysis of the APC gene all exons and exon–intron boundaries were amplified ( supplementary Table 1 is available at Carcinogenesis Online). PCR was performed in 25 μl reactions containing 2.5 μl of ×10 Taq polymerase buffer, a 0.2 mM concentration of each deoxynucleoside triphosphate, a 0.2 μM concentration of each primer, 1 U of Taq DNA polymerase and 20 ng of template DNA. DNA sequencing was performed on PCR products purified with ExoSAP-IT (USB Europe, Staufen, Germany). Sequence reactions were carried out using Quick-Start Dye Terminator Cycle Sequencing kit (Beckman Coulter) and analyzed on a CEQ8000 (Beckman Coulter) according to the manufacturer’s protocol. Identification of somatic mutations was followed by amplification and sequencing of corresponding parts of the APC gene using DNA extracted from normal tissue to exclude germ line mutations.

Table I.

Characterization of 32 tumors by MSI analysis, immunohistochemistry (IHC) of MLH1 and MSH2 and mutational analysis (Mut) of the B-RAF , CTNNB1 (β- Catenin ) and APC genes

ID Sex Age Stage (UICC) Grading MSI MLH1 (IHC) MSH2 (IHC) BRAF (Mut) CTNNB1 (Mut) APC (Mut) 
    MSS tumors 
21 79 G2 — pos pos WT WT WT 
29 56 G2 — pos pos WT WT WT 
33 44 G2 — pos pos WT WT c.3927 del AAGA 
51 77 G2 — pos pos WT WT c.3908 del AA 
123 67  — pos pos WT WT WT 
127 72 G2 — pos pos WT WT c.847 C>T 
132 71 G2 — pos pos WT WT c.1690 C>T 
143 73 G3 — pos pos WT WT c.4025 insT 
167 66 G3 — pos pos c.1799 T>A WT WT 
168 72 G3 — pos pos WT WT IVS10-2 A>G 
169 55 G3 — pos pos WT WT WT 
170 64 G2 — pos pos WT WT c.637C>T 
172 63 G2 — pos pos WT WT c.4057 G>T 
179 63 G2 — pos pos WT WT WT 
182 64 G2 — pos pos WT c.133 T>C c.4284 delA 
189 74 G2 — pos pos WT WT c.4416 ins 98 bp 
    MSI tumors 
15 66 G2 high neg pos WT c.133 T>G WT 
25 85 G3 high neg pos c.1799 T>A WT c.3779 A>G 
55 38 G2 high pos neg WT WT WT 
59 60 G2 high neg pos WT WT WT 
64 86 G3 high neg pos c.1799 T>A WT WT 
75 76 G2 high neg pos c.1799 T>A WT WT 
99 30 G2 high pos neg WT WT c.4666 ins A 
104 72 G3 high neg pos WT WT WT 
109 83 G3 high neg pos c.1799 T>A WT WT 
122 74 G2 high neg pos c.1799 T>A WT WT 
126 62 G2 high neg pos WT WT c.4660 G>A 
131 64 G2 high pos pos c.1799 T>A WT WT 
181 67 G3 high neg pos WT c.133 T>C WT 
186 73 G2 high neg pos c.1799 T>A WT WT 
32 76 G3 high neg pos c.1799 T>A WT WT 
98 80 G2 high neg pos c.1799 T>A WT WT 
ID Sex Age Stage (UICC) Grading MSI MLH1 (IHC) MSH2 (IHC) BRAF (Mut) CTNNB1 (Mut) APC (Mut) 
    MSS tumors 
21 79 G2 — pos pos WT WT WT 
29 56 G2 — pos pos WT WT WT 
33 44 G2 — pos pos WT WT c.3927 del AAGA 
51 77 G2 — pos pos WT WT c.3908 del AA 
123 67  — pos pos WT WT WT 
127 72 G2 — pos pos WT WT c.847 C>T 
132 71 G2 — pos pos WT WT c.1690 C>T 
143 73 G3 — pos pos WT WT c.4025 insT 
167 66 G3 — pos pos c.1799 T>A WT WT 
168 72 G3 — pos pos WT WT IVS10-2 A>G 
169 55 G3 — pos pos WT WT WT 
170 64 G2 — pos pos WT WT c.637C>T 
172 63 G2 — pos pos WT WT c.4057 G>T 
179 63 G2 — pos pos WT WT WT 
182 64 G2 — pos pos WT c.133 T>C c.4284 delA 
189 74 G2 — pos pos WT WT c.4416 ins 98 bp 
    MSI tumors 
15 66 G2 high neg pos WT c.133 T>G WT 
25 85 G3 high neg pos c.1799 T>A WT c.3779 A>G 
55 38 G2 high pos neg WT WT WT 
59 60 G2 high neg pos WT WT WT 
64 86 G3 high neg pos c.1799 T>A WT WT 
75 76 G2 high neg pos c.1799 T>A WT WT 
99 30 G2 high pos neg WT WT c.4666 ins A 
104 72 G3 high neg pos WT WT WT 
109 83 G3 high neg pos c.1799 T>A WT WT 
122 74 G2 high neg pos c.1799 T>A WT WT 
126 62 G2 high neg pos WT WT c.4660 G>A 
131 64 G2 high pos pos c.1799 T>A WT WT 
181 67 G3 high neg pos WT c.133 T>C WT 
186 73 G2 high neg pos c.1799 T>A WT WT 
32 76 G3 high neg pos c.1799 T>A WT WT 
98 80 G2 high neg pos c.1799 T>A WT WT 

UICC—Union for International Cancer Control (tumor staging), (‘—’, no instability; ‘high’, MSI-high). Expression analysis of MLH1 and MSH2 was performed by IHC (‘pos’, retained staining in tumor tissue; ‘neg’, absent staining in tumor tissue). BRAF, CTNNB1 and APC were analyzed by mutational analysis (Mut) (‘WT’, wild-type/No mutation found, ‘CdXXX Y->Z’—mutation in Codon XXX base X to Z, ‘Ins Y CdXXX’—insertion of base Y in Codon XXX).

SNP array analysis

SNP array analysis was performed according to standard protocols (Affymetrix, München, Germany; www.Affymetrix.com ). Briefly, a 250 ng sample of either amplified DNA or unamplified DNA was digested with XbaI and ligated to an XbaI adapter before subsequent PCR amplification using AmpliTaq Gold (Applied Biosystems, Foster City, CA). To obtain sufficient PCR products, four 100 μl PCRs were set up for each XbaI adapter-ligated DNA sample. The PCR products were pooled, purified, fragmented with DNase I and visualized on a 4% TBE agarose gel to confirm that the sizes ranged from 50 to 100 bp. Subsequently, the fragmented PCR products were end labeled with biotin and hybridized to the array. Detection was performed using the Affymetrix Fluidics station 450 and the GeneChip® scanner 3000. Scanned data files were generated using Affymetrix GeneChip Operating Software, version 1.2, and genotype calls were determined automatically using GeneChip DNA Analysis Software, version 3.0.

Genotypes and probe intensities derived from normal tissue and cancer DNA were loaded into the dChip software package ( http://www.dchip.org/ ) for LOH and copy number analysis, as described previously ( 21 ). Briefly, the LOH call was assigned, for a pair of tumor and matched normal tissue samples, as LOH or allelic imbalance (heterozygosity in the normal tissue, homozygosity in the tumor), retention of heterozygosity (heterozygosity in both the normal tissue and the tumor) or non-informative (homozygosity in the normal tissue). The hidden Markov model from dChip was used to infer regional LOH with the LOH call threshold was set at 0.5. For calculating the copy number, data were normalized to a baseline array with median signal intensity, using the invariant set normalization method. The model-based (perfect match/mismatch) method was used to obtain the signal values after normalization. Normalized intensities in matched normal tissue DNA samples were used as the reference set to calculate the copy number of each marker in the tumor samples. The hidden Markov model was used to determine copy number change along each chromosome.

Results

Mutational and SNP array analysis of 16 MSS and 16 MSI CRC tissues and corresponding normal mucosa

MSI analysis and immunohistochemistry of MLH1 and MSH2 enabled identification of 16 patients with an MSI-H tumor. MLH1 was inactivated in 13 of the 16 tumors, MSH2 in 2 of the 16 and in 1 tumor neither MLH1 nor MSH2 was inactivated. The latter may be a case of dysfunctional protein that has retained antigenicity. Of the remaining 149 tumors without MSI-H, we randomly chose 16 tumors for further characterization. The resulting 32 tumors were further characterized by mutational analysis of the APC , B-RAF and CTNNB1 genes. Data regarding the patients, the tumors and the mutational profile are summarized in Tables I and II . Although B-RAF -mutations were almost exclusively found in the MSI-H tumors, APC -mutations were predominantly found in the MSS tumors; CTNNB1 mutations were rarely found in both subgroups. B-RAF -mutant tumors also generally display inactivation of hMLH1, and due to its strong association with silencing of hMLH1, B-RAF mutation can be considered a surrogate marker for promoter methylation of hMLH1 in sporadic MSI cancers.

Table II.

Analysis of allelic loss and minimal loss regions in 32 tumors by 50 K SNP array analysis

ID chromosome band Start End Size LOH cnLOH Allelic loss Gene 
    MSS CRC tissues ( n = 16)  
5q22.1–q22.3 110.827.000 113.651.000 2.824.000 3/16 5/16 8/16 APC 
8pter–p12 32.218.000 32.218.000 10/16 none 10/16 ??? 
17p13.2–p13.1 5.762.000 10.559.000 4.797.000 8/16 4/16 12/16 p53 
18q11.2–qter 22.361.000 76.900.000 54.539.000 13/16 none 13/16 SMAD4 
    MSI CRC tissues ( n = 16)  
6pter–p22.2 25.446.000 25.446.000 none 4/16 4/16 ??? 
3pter–p21.31 46.513.000 46.513.000 none 3/16 3/16 MLH1 
ID chromosome band Start End Size LOH cnLOH Allelic loss Gene 
    MSS CRC tissues ( n = 16)  
5q22.1–q22.3 110.827.000 113.651.000 2.824.000 3/16 5/16 8/16 APC 
8pter–p12 32.218.000 32.218.000 10/16 none 10/16 ??? 
17p13.2–p13.1 5.762.000 10.559.000 4.797.000 8/16 4/16 12/16 p53 
18q11.2–qter 22.361.000 76.900.000 54.539.000 13/16 none 13/16 SMAD4 
    MSI CRC tissues ( n = 16)  
6pter–p22.2 25.446.000 25.446.000 none 4/16 4/16 ??? 
3pter–p21.31 46.513.000 46.513.000 none 3/16 3/16 MLH1 

Monoallelic regions with reduced copy numbers represent regions with classical deletions (LOH). Monoallelic regions without changes in the copy number represent regions with copy neutral LOH (cnLOH). Allelic loss is defined as LOH or cnLOH. In the table, frequently affected regions are shown (>50% for MSS tumors, >1 for MSI tumors). Start, end and size defines the minimal lost chromosomal region found in all affected tumors. Gene–gene that is localized in the minimal loss region.

After mutational analysis, a 50 K SNP array analysis of the 32 tumors was performed. SNP array analysis using the ‘dchip’-Software allowed the identification of monoallelic chromosomal regions in MSS ( Figure 1A ) and MSI cell lines ( Figure 2A ) and facilitated the determination of their appropriate copy numbers ( Figures 1B and 2B ). Monoallelic regions with reduced copy numbers represent regions with classical deletions (LOH). Monoallelic regions without changes in the copy number represent regions with copy neutral LOH (cnLOH) ( Figures 1 and 2 ; Table II ).

Fig. 1.

SNP array ( A ) and copy number ( B ) analysis of MSS CRCs. Dark blue bands (A) represent allelic losses (LOH and cnLOH). In some chromosomal regions, allelic losses are incomplete (interrupted by non-allelic loss regions) and in some other regions, there are only scattered bands. This can be explained by a contamination of normal mucosa and/or different clones. In (B) a reduction of copy number is shown in blue color, a gain in red. Regions displaying cnLOH are represented by black bands and should match with the allelic loss regions in (A). In some chromosomal regions (e.g. 8p-tumor 170 and 18q-tumor 179), there are only scattered bands that do not match allelic losses in (A), so that these bands were not counted as cnLOH.

Fig. 1.

SNP array ( A ) and copy number ( B ) analysis of MSS CRCs. Dark blue bands (A) represent allelic losses (LOH and cnLOH). In some chromosomal regions, allelic losses are incomplete (interrupted by non-allelic loss regions) and in some other regions, there are only scattered bands. This can be explained by a contamination of normal mucosa and/or different clones. In (B) a reduction of copy number is shown in blue color, a gain in red. Regions displaying cnLOH are represented by black bands and should match with the allelic loss regions in (A). In some chromosomal regions (e.g. 8p-tumor 170 and 18q-tumor 179), there are only scattered bands that do not match allelic losses in (A), so that these bands were not counted as cnLOH.

Fig. 2.

SNP array ( A ) and copy number ( B ) analysis of MSI CRCs. Dark blue bands (A) represent allelic losses (LOH and cnLOH). In some chromosomal regions, allelic losses are incomplete (interrupted by non-allelic loss regions) and in some other regions, there are only scattered bands. This can be explained by a contamination of normal mucosa and/or different clones. In (B) a reduction of copy number is shown in blue color, a gain in red. In tumor 99, there seem to be a gain of almost all chromosomal regions that can be interpreted as variation of signal intensity in this specific array.

Fig. 2.

SNP array ( A ) and copy number ( B ) analysis of MSI CRCs. Dark blue bands (A) represent allelic losses (LOH and cnLOH). In some chromosomal regions, allelic losses are incomplete (interrupted by non-allelic loss regions) and in some other regions, there are only scattered bands. This can be explained by a contamination of normal mucosa and/or different clones. In (B) a reduction of copy number is shown in blue color, a gain in red. In tumor 99, there seem to be a gain of almost all chromosomal regions that can be interpreted as variation of signal intensity in this specific array.

Monoallelic chromosomal fragments in MSS colorectal carcinomas

Monoallelic regions were frequently (>50%) found in the 5q22.1–q22.3, 8pter–8p12, 17p13.2–13.1 and 18q11.2–qter regions ( Figure 1 , Table II ). In these regions, genes associated with colon carcinogenesis could be assigned to APC (5q21–22), TP53 (17p13.1), and SMAD4 (18q21.1). More than half of the tumor regions (five of eight), displaying loss of heterozgosity in 5q22.1–q22.3, exhibited no concomitant copy number alterations (cnLOH). Mutations in the APC gene were found in all tumors displaying either LOH with concomitant deletion or cnLOH in 5q22.1–q22.3. Twelve of 16 tumors exhibited allelic loss in 17p13.2–13.1 and 30% of these allelic losses were copy neutral events. In contrast to 5q22.1–q22.3 and 17p13.2–13.1, 8pter–p12 and 18q11.2–qter were frequently lost, resulting in a monoallelic region with a reduced copy number. SMAD4 is located in 18q11.2–qter, but so far no tumor suppressor gene associated with colon cancer carcinogenesis has been identified in the 8p region. MTUS1 (‘mitochondrial tumor suppressor 1’) is a recently described candidate tumor suppressor gene, located at 8p22. In order to determine the relevance of MTUS1 as a potential target of this deletion, we performed an immunohistochemical analysis of the MTUS1 expression on 10 of the 16 tumors with a deletion in 8pter–p12 and corresponding normal mucosa as reference tissue ( supplementary Figure 1 is available at Carcinogenesis Online). A complete absence of staining was found in 4 of the 10 tumors, two others exhibited significant reduction in the staining intensity, two exhibited the same level of staining of MTUS1 as in the reference tissue and two could not be evaluated because there was insufficient material. A sequencing analysis of the three tumors with a complete absence of staining was performed, but no mutation in the MTUS1 gene could be identified.

Monoallelic chromosomal fragments in MSI colorectal carcinomas

In comparison with MSS tumors, fewer chromosomal alterations were found in MSI tumors ( Figure 1 , Table II ). Monoallelic regions in MSI tumors were always reduplicated after an initial deletion, resulting in copy neutral LOH. Most frequently altered were the 6pter–p22.2 (4/16) and 3pter–p21.31 (3/16) regions ( Table III ). Unique allelic losses were found in chromosomal regions 4q24–qter, Chromosome 9, 13q12.11–qter and 15q12–qter, 19p ( Figure 2 ). In 3pter–p21.31, the hMLH1 gene is localized (3p21.3). The absence of staining of hMLH1 occurred in all tumors with allelic loss in 3pter–p21.31. A tumor suppressor gene in 6pter–p22.2 has not yet been described for colorectal carcinomas. We chose ID4 (6p22.3), SOX4 (6p22.2), NOL7 (6p23) und E2F3 (6p22.3) as candidate genes in this region (because of potential relevance for CRC and available antibodies). A cnLOH was found in 4 of the 16 MSI tumors. NOL7, SOX4 and ID4 exhibited no reduction in the staining intensity in tumors and we found only one tumor in which staining was absent for E2F3 ( supplementary Figure 2 is available at Carcinogenesis Online).

Table III.

Monoallelic chromosomal fragments in MSI and MSS colorectal carcinomas and corresponding genes inactivated by LOH/cnLOH

MSS tumors 
ID 8pter–p12 MTUS1 (IHC) 5q22.1–q22.3 APC (Mut) 
21 LOH na      
29 LOH neg      
33   LOH c.3927 del AAAGA 
51 LOH ↓↓      
123        
127 LOH na LOH c.847C>T 
132   cnLOH c.690C>T 
143 LOH neg cnLOH c.4025insT 
167        
168 LOH pos cnLOH IVS10-2 A>G 
169        
170 LOH neg cnLOH c.637C>T 
172 LOH ↓↓ LOH c.4057G>T 
179 LOH neg      
182   cnLOH c.4284delA 
189 LOH pos      
MSI tumors 
ID 3pter–p21.31 MLH1 (IHC) 6pter–p22.2 NOL7 (IHC) ID4 (IHC) E2F3 (IHC) SOX4 (IHC) 
15 cnLOH neg      
25  neg      
55  pos      
59 cnLOH neg cnLOH pos pos pos pos 
64  neg      
75  neg      
99  pos      
104  neg      
109  neg cnLOH pos pos pos pos 
122  neg      
126  neg      
131  pos cnLOH pos pos pos pos 
181 cnLOH neg      
186  neg cnLOH pos pos pos pos 
32  neg      
98  neg      
MSS tumors 
ID 8pter–p12 MTUS1 (IHC) 5q22.1–q22.3 APC (Mut) 
21 LOH na      
29 LOH neg      
33   LOH c.3927 del AAAGA 
51 LOH ↓↓      
123        
127 LOH na LOH c.847C>T 
132   cnLOH c.690C>T 
143 LOH neg cnLOH c.4025insT 
167        
168 LOH pos cnLOH IVS10-2 A>G 
169        
170 LOH neg cnLOH c.637C>T 
172 LOH ↓↓ LOH c.4057G>T 
179 LOH neg      
182   cnLOH c.4284delA 
189 LOH pos      
MSI tumors 
ID 3pter–p21.31 MLH1 (IHC) 6pter–p22.2 NOL7 (IHC) ID4 (IHC) E2F3 (IHC) SOX4 (IHC) 
15 cnLOH neg      
25  neg      
55  pos      
59 cnLOH neg cnLOH pos pos pos pos 
64  neg      
75  neg      
99  pos      
104  neg      
109  neg cnLOH pos pos pos pos 
122  neg      
126  neg      
131  pos cnLOH pos pos pos pos 
181 cnLOH neg      
186  neg cnLOH pos pos pos pos 
32  neg      
98  neg      

In MSS tumors, a chromosomal loss of the region 5q22.1–22.3 is associated with a mutation of the APC gene. Expression of MTUS1 (8p22) is downregulated or absent in 6/10 tumors with a loss of 8pter–8p12 (IHC analysis: ‘pos’, retained staining in tumor tissue; ‘neg’, absent staining in tumor tissue; ‘na’ - not analysed because of insufficient material). In MSI tumors, a loss of the region 3pter–p21.31 was associated with an absent expression of MLH1 in 3/3 affected tumors. IHC analysis of NOL7, SOX4 and ID4 exhibited no reduction in the staining intensity in tumors and we found only one tumor in which staining was absent for E2F3.

Discussion

We found recurrent cnLOH and LOH regions in 5q22.1–q22.3, 17p13.2–13.1, 18q11.2–qter and 8pter–p12 in MSS carcinomas. In 5q22.1–q22.3 and 17p13.2–13.1, either LOH or cnLOH is frequently found (8 and 12 of 15 samples, respectively). There is a trend for reduplication of the 5q region resulting in cnLOH in five of the eight cases examined, so that the cancer cells may have a growth advantage if the copy number in 5q22.1–q22.3 is preserved. Where-as tumor suppressor genes located in 5q22.1–q22.3, 17p13.2–13.1 and 18q11.2–qter are already known ( APC , TP53 , SMAD4 ), no tumor suppressor gene that is associated with colon cancer carcinogenesis has yet been identified in the 8pter–p12 region. MTUS1 (mitochondrial tumor suppressor 1) is a recently described candidate tumor suppressor gene, located at 8p22 ( 22 ). In order to test the relevance of MTUS1 as a potential target gene of this alteration, we performed immunophenotyping ( supplementary Figure 1 is available at Carcinogenesis Online) of the region in 10 of 15 tumors with a deletion in 8pter–p12 and recorded the absence or marked reduction of staining of MTUS1 in six of these, associated with an allelic loss of 8pter–p12 in these tumors. In a recent article Zuern et al. ( 23 ) also demonstrated a downregulation of MTUS1 expression in colon cancer tissues and detected no mutations of MTUS1 in the coding sequence or the predicted promoter region in cancer tissues. Furthermore, they found no difference of CpG methylation in the predicted promotor region, but an altered CpNpG methylation. Thus, while further analysis of the mechanisms of downregulation is required, MTUS1 seems to be an interesting candidate tumor suppressor gene in the 8p22 region.

In MSI carcinomas, frequent allelic losses were found in 3pter–p21.31 (3/16) and, even more frequently, in 6pter–p22.2 (4/16). In 3pter–p21.31, hMLH1 is localized (3p21.3). In all tumors with cnLOH of 3pter–p21.31, we recorded the absence of staining of hMLH1 during immunophenotyping, suggesting that cnLOH plays a role in the inactivation of hMLH1 . Nevertheless, no cnLOH was found in 10 of the 13 tumors without hMLH1 staining, so that mechanisms other than cnLOH must be involved in the inactivation of hMLH1 . Frequent allelic losses in 6pter–p22.2 in MSI carcinomas have also been described by van Puijenbroek et al. ( 24 ). To date, no tumor suppressor gene has been identified in 6pter–p22.2. The affected region spans 25 Mb and contains 106 genes. In this study, we analyzed four genes that were interesting candidate tumor suppressors [ NOL7 (13723538-13729106; 6p23), ID4 (19846839-20047640; 6p22.2), SOX4 (21701951-21706828; 6p22.2) and E2F3 (20510116-20601924; 6p22.2)]. NOL7 , SOX4 , ID4 and E2F3 showed no alteration in tumors. Consequently, these genes seem to be of no or little relevance for tumorigenesis, even though they could be truncated or altered in another way that could not be identified by immunohistochemistry. There are few reports of an involvement of the 6p chromosomal region in colorectal carcinogenesis. Honchel et al. ( 25 ) performed microsatellite analysis of 64 CRC tissue samples and found frequent allelic losses distal of the marker D6S271 (6p21.1). Studies with a larger series of MSI carcinomas have to be performed in order to define a more precise minimal deleted/altered region in 6pter–p22.2.

In contrast to our previous study ( 12 ), we found cnLOH-regions not only limited to regions where early-acting tumor suppressor genes (MLH1, APC) are located, but also in the genomic region of TP53, which is a late acting gene in the adenoma-carcinoma sequence of MSS colorectal carcinomas. Consequently cnLOH seems not to be limited to early tumor development but is a mutational mechanism affecting multiple regions in all tumor stages and tumor subgroups (MSI) and conferring a growth advantage by creating a stable copy number in the affected region.

There are three other studies that deal with SNP array analysis and CRC. Andersen et al. ( 11 ) examined 15 laser-microdissected CRC tissues and the corresponding normal mucosa. All carcinomas were sporadic and MSS. They also found commonly affected regions (>25% of the tumors analyzed) in 8q, 5q, 17p and 18q. In addition, they identified 3p, 13q, 14q, 15q, 20p and 22q as recurrently altered regions. In a similar approach, Kurashina et al. ( 26 ) analyzed 94 colorectal carcinomas. They found recurrent copy number abnormalities in CRC associated with chromosomes 7, 8, 13, 18 and 20, and recurrent deletions in chromosomes 1p, 4q, 5q, 8p, 11q, 14q, 15q, 17p, 18 and 22. In contrast to our study, they found frequent UPD not only in 5q and 17p but also in chromosomes 8p and 18q. Van Puijenbroek et al. ( 24 ), using 6 K SNP arrays, analyzed 10 sporadic MSI-H carcinomas with hMLH1 promoter hypermethylation from 10 patients. In 6 of the 10 sporadic MSI-H carcinomas, few small copy number changes were detected. Three of these tumors exhibited one copy number abnormality and the other three displayed two changes. The affected segments ranged in size from 8 to 47 sub-bands of the genome of these carcinomas, affecting chromosomes 1q, 4, 6, 8p, 9, 10 and 12. Amplification of the entire chromosome 12 occurred in two cases; all additional copy number changes were unique. The genomic locus of MLH 1 showed neither physical loss nor cnLOH in eight tumors that could be tested. CnLOH was observed in 3 of the 10 carcinomas (30%). Two tumors exhibited one segment of cnLOH (12q, 6p25.3–p22.1) and the other tumor displayed two segments (6p25.3–q25.2, 19p13.3–p13.11). The region most frequently affected was 6pter–p22.2, which corresponds exactly to the minimal deleted region identified in our study.

Conclusion

In this study, we found allelic losses in MSS and MSI tumors. In MSS tumors, more losses than in MSI tumor were identified. Nethertheless, the results suggest that cnLOH is an important mechanism involved in the carcinogenesis of MSS and MSI tumors, enabling the inactivation of a tumor suppressor gene without copy number alteration of the respective region; this may be crucial for MSI tumors (3pter–p21.31— MLH1 ) and for some chromosomal regions in MSS tumors (5q22.1–q22.3— APC and 17p13.2–13.1— TP53 ). SNP array analysis resulted in the identification of MTUS1 (8p22) as a candidate tumor suppressor gene in the frequently lost 8pter–p12 region in MSS colorectal carcinomas. In addition, the chromosomal region 6pter–p22.2 was identified as a frequently altered region in MSI carcinomas. In further studies, a greater number of MSI tumors need to be analyzed in order to narrow the minimal deleted region in chromosome 6p to facilitate identification of the tumor suppressor gene located there. A mutational analysis of tumors exhibiting loss of MTUS1 expression is required and functional studies need to be performed to confirm that MTUS1 is a tumor suppressor gene with relevance for colon cancer carcinogenesis.

Supplementary material

Supplementary Table 1 and Figures 1 and 2 can be found at http://carcin.oxfordjournals.org/

Funding

Interdisciplinary Center for Clinical Science, Wuerzburg, Germany.

Abbreviations

    Abbreviations
  • CIN

    chromosomal instability

  • CRC

    colorectal cancer

  • MSS

    microsatellite stable

  • LOH

    loss of heterozygosity

  • MSI

    microsatellite instability

  • SNP

    single nucleotide polymorphism

  • PCR

    polymerase chain reaction

  • UPD

    uniparental disomy

Conflict of Interest Statement: None declared.

References

1.
Weitz
J
, et al.  . 
Colorectal cancer
Lancet
 , 
2005
, vol. 
365
 (pg. 
153
-
165
)
2.
Thibodeau
SN
, et al.  . 
Microsatellite instability in cancer of the proximal colon
Science
 , 
1993
, vol. 
260
 (pg. 
816
-
819
)
3.
Cahill
C
, et al.  . 
Mutations of mitotic checkpoint genes in human cancers
Nature
 , 
1998
, vol. 
392
 (pg. 
300
-
303
)
4.
Melcher
R
, et al.  . 
Spectral karyotype analysis of colon cancer cell lines of the tumour suppressor and mutator pathway
Cytogenet. Genome Res.
 , 
2002
, vol. 
98
 (pg. 
22
-
28
)
5.
Abdel-Rahman
WM
, et al.  . 
Spectral karyotyping suggests additional subsets of colorectal cancers characterized by pattern of chromosome rearrangement
Proc. Natl Acad. Sci. USA.
 , 
2001
, vol. 
98
 (pg. 
2538
-
2543
)
6.
Lengauer
C
, et al.  . 
Genetic instabilities in human cancers
Nature
 , 
1998
, vol. 
396
 (pg. 
643
-
649
)
7.
Kinzler
KW
, et al.  . 
Lessons from hereditary colorectal cancer
Cell
 , 
1996
, vol. 
87
 (pg. 
159
-
170
)
8.
Lengauer
C
, et al.  . 
Genetic instability in colorectal cancers
Nature
 , 
1997
, vol. 
386
 (pg. 
623
-
627
)
9.
Knudson
AG
Mutation and cancer: statistical study of retinoblastoma
Proc. Natl Acad. Sci. USA
 , 
1971
, vol. 
68
 (pg. 
820
-
823
)
10.
Moolgavkar
SH
, et al.  . 
Mutation and cancer: a model for human carcinogenesis
J. Natl Cancer Inst.
 , 
1981
, vol. 
66
 (pg. 
1037
-
1052
)
11.
Andersen
CL
, et al.  . 
Frequent occurrence of uniparental disomy in colorectal cancer
Carcinogenesis
 , 
2007
, vol. 
28
 (pg. 
38
-
48
)
12.
Melcher
R
, et al.  . 
SNP-Array genotyping and spectral karyotyping reveal uniparental disomy as early mutational event in MSS- and MSI-colorectal cancer cell lines
Cytogenet. Genome Res.
 , 
2007
, vol. 
118
 (pg. 
214
-
221
)
13.
Kralovics
R
, et al.  . 
Acquired uniparental disomy of chromosome 9p is a frequent stem cell defect in polycythemia vera
Exp. Hematol.
 , 
2002
, vol. 
30
 (pg. 
229
-
236
)
14.
Zhao
X
, et al.  . 
An integrated view of copy number and allelic alterations in the cancer genome using single nucleotide polymorphism arrays
Cancer Res.
 , 
2004
, vol. 
64
 (pg. 
3060
-
3071
)
15.
Raghavan
M
, et al.  . 
Genome-wide single nucleotide polymorphism analysis reveals frequent partial uniparental disomy due to somatic recombination in acute myeloid leukemias
Cancer Res.
 , 
2005
, vol. 
65
 (pg. 
375
-
378
)
16.
Tuna
M
, et al.  . 
Uniparental disomy in cancer
Trends Mol. Med.
 , 
2009
, vol. 
15
 (pg. 
120
-
128
)
17.
Calhoun
ES
, et al.  . 
Copy-number methods dramatically underestimate loss of heterozygosity in cancer
Genes Chromosomes Cancer
 , 
2006
, vol. 
45:
 (pg. 
1070
-
1071
)
18.
Boland
CR
, et al.  . 
A National Cancer Institute Workshop on Microsatellite Instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer
Cancer Res.
 , 
1998
, vol. 
58
 (pg. 
5248
-
5257
)
19.
Rudiger
T
, et al.  . 
Differential diagnosis between classic Hodgkin's lymphoma, T-cell-rich B-cell lymphoma, and paragranuloma by paraffin immunohistochemistry
Am. J. Surg. Pathol.
 , 
1998
, vol. 
22
 (pg. 
1184
-
1191
)
20.
Scholtka
B
, et al.  . 
A gene marker panel covering the Wnt and the Ras-Raf-MEK-MAPK signalling pathways allows to detect gene mutations in eighty percent of early (UICC I) colon cancer stages in humans
Cancer Epidemiol.
 , 
2009
, vol. 
33
 (pg. 
123
-
129
)
21.
Lin
M
, et al.  . 
dChipSNP: significance curve and clustering of SNP-array-based loss-of-heterozygosity data
Bioinformatics
 , 
2004
, vol. 
20
 (pg. 
1233
-
1240
)
22.
Seibold
S
, et al.  . 
Identification of a new tumor suppressor gene located at chromosome 8p21.3-22
FASEB J.
 , 
2003
, vol. 
17
 (pg. 
1180
-
1182
)
23.
Zuern
C
, et al.  . 
Down-regulation of MTUS1 in human colon tumors
Oncol. Rep.
 , 
2010
, vol. 
23
 (pg. 
183
-
189
)
24.
van Puijenbroek
M
, et al.  . 
Genome-wide copy neutral LOH is infrequent in familial and sporadic microsatellite unstable carcinomas
Fam. Cancer
 , 
2008
, vol. 
7
 (pg. 
319
-
330
)
25.
Honchel
R
, et al.  . 
Tumor necrosis factor-alpha allelic frequency and chromosome 6 allelic imbalance in patients with colorectal cancer
Cancer Res.
 , 
1996
, vol. 
56
 (pg. 
145
-
149
)
26.
Kurashina
K
, et al.  . 
Chromosome copy number analysis in screening for prognosis-related genomic regions in colorectal carcinoma
Cancer Sci.
 , 
2008
, vol. 
99
 (pg. 
1835
-
1840
)