Introduction

Mantle cell lymphoma (MCL) is characterized by t(11;14)(q13;q32), which results in overexpression of CCND1, and is presumed to derive from naive pre-germinal center CD5⁺ B cells (Seto et al., 1992; Jaffe et al., 2001). The identification of this translocation in virtually all cases of MCL with CCND1 overexpression indicates that this genomic alteration is an important mechanism for its pathogenesis (Jaffe et al., 2001). Despite the presence of this common molecular marker, experiments with transgenic mice overexpressing CCND1 proved that this protein alone cannot induce lymphomas (Hinds et al., 1994; Lovec et al., 1994), so that genomic aberrations other than the 11q13 translocation must be involved in the development and progression of MCL. To identify such additional aberrations, several studies using comparative genomic hybridization (CGH) and chromosome banding analyses have been conducted (Monni et al., 1998; Beà et al., 1999; Cuneo et al., 1999; Bentz et al., 2000; Bigoni et al., 2001; Martinez-Climent et al., 2001; Allen et al., 2002). These studies showed that genomic imbalances, such as gain/amplification of 3q, 6p, 7p, 8q, 10p, 12q and 18q, and loss/deletion of 1p, 6q, 8p, 9p, 11q and 13q, frequently occur in MCL. Some genetic deregulations accompanying these genomic imbalances, such as BMI-1 from amplification of 10p12.2, p16^INK4a from deletion of 9p21.3 and ATM from deletion of 11q22.3 (Dreyling et al., 1997; Pinyol et al., 1997; Stilgenbauer et al., 1999; Schaffner et al., 2000; Beà et al., 2001; Rosenwald et al., 2003), were also detected. However, the target genes of these other amplification and deletion sites remain unknown, one of the reasons being the limited resolution of chromosomal banding analysis or conventional CGH, which can detect only DNA copy number aberrations greater than 10–20 mega bases (Mb).

Recently, a chip-based CGH approach with high resolution and accuracy, known as array-based CGH (array CGH), was developed (Pinkel et al., 1998). We established our own array CGH using a glass slide on which 2348 bacterial artificial chromosomes (BACs) and P-1-derived artificial chromosomes (PACs) were spotted in duplicate with an average resolution of 1.3 Mb. In addition, our array CGH could identify a novel tumor-related gene, C13orf25, at 13q31.3 in B-cell lymphomas (Ota et al., 2004). These results indicate that quantitative measurements of DNA copy number changes made with the array CGH can identify more accurately regions of genomic imbalance and that this procedure could thus be a useful tool for identification of tumor-related gene(s).

To gain a more accurate understanding of complex gene copy number changes and to identify key gain/loss regions in greater detail, we applied genome-wide array-based CGH to a panel of 29 patient samples and seven cell lines that derived from MCL.

Results

Genomic profiles of MCL patient samples and cell lines

Representative examples of the high-resolution analysis of a patient sample (G468) and SP-53 cell line are shown in Figure 1a and b, respectively. Array CGH could detect both small and whole-chromosome areas of gains and deletions as well as delineate amplification and deletion borders. A small amplicon involving clones containing known oncogenes was easily detected, as were small homozygous deletions.

Figure 1.

Representative genomic profiles for individual tumors. Whole genomic profiles are shown for a representative patient sample (a) and the SP-53 cell line (b). Log₂ ratios were plotted for all clones on the basis of chromosome position, with the vertical lines showing separation of chromosomes. The BACs and PACs are ordered according to their position in the genome from the 1p telomere on the left to the Xq telomere on the right. (a) Regions of copy number gain: 5p, 7p21.3, 17q21.31–q24.3 and X. Regions of copy number loss: 1p32, 2p11.2, 2q13, 8p12–p23.3, 8q12.3–q13.1, 9p24.3–q31.2, 11q22.3–q23.2, 13q14.3–q21.1 and 15. (b) Regions of copy number gain: 7p11.2–p22.3. Regions of copy number loss: 1p36.23–p36.32, 1p13.3–p31.2, 1q13.2–q44, 2p11.2–q14.3, 4p15.1–p16.1, 6q14.1–q21, 6q23.2–q26, 7q22.1–q32.3, 9p21.3–p22.1, 11q22.3–q23.2, 18q22.1 and 20q13.13–q13.2

Full figure and legend (62K)

Genomic imbalances of MCL patient samples

Gains or losses of genetic material shown by all 29 patient samples were subjected to data analysis. The entire tumor set involved copy number gains on an average of 130.6 Mb or 4.6%, and copy number loss on an average of 250.6 Mb or 8.8% of the genome. Total alterations averaged 3.8 regions of gain and 7.3 regions of loss. The genome-wide frequency of copy number alterations, both gains and losses, are shown in Figure 2. Regions of recurrent gain (five cases) involved chromosomes 3q13.11–q29, 6p21.32–p25.3, 7p14.3–p22.3, 8q13.2–q24.22, 10p12.1–p12.2 and 17q23.2–q24.1, and recurrent losses (five cases) localized at 1p36.23–p36.32, 1p11.2–p31.3, 1q42.2–q43, 2p11.2, 2q13, 5q21.1–q23.1, 6q16.2–q27, 8p12–p23, 9p24.3–q31.1, 10p12.31–p15.3, 11q14.3–q23.2, 13q13.2–q34, 15q13–q21.1, 17p11.2–p13.3, 19p13.2–p13.3 and 22q12.1–q13.1. The most frequent imbalances were gains of chromosomes 3q26.1 (48%), 7p21.1–p21.2 (34%), 6p25.3 (24%), 8q21.3–q24.21 (24%), 10p12.1–p12.2 (21%) and 17q23.2–q24.1 (17%), and losses of chromosome 2p11.2 (83%), 11q22.3–q23.1 (59%), 13q14.3–q21.1 (55%), 1p21.3–p22.1 (52%), 13q34 (52%), 9q22.33–q31.1 (45%), 17p13.3 (45%), 9p21.3–p22.1 (41%), 9p24.2–p24.3 (41%), 6q23.2–q24.1 (38%), 1p36.23–p36.32 (31%), 8p23.1–p23.3 (34%), 10p14–p15.3 (31%), 19p13.2 (28%), 5q22.1–q22.3 (21%), 22q12.2 (21%), 1q42.2–q43 (17%) and 2q13 (17%) (Table 1). Recurrent losses of 1p36.23–p36.32, 1q42.2–q43, 2p11.2, 2q13, 17p13.3 and 19p13.2–p13.3 were identified for the first time in this study, but no regions of gains were found other than those already listed in previous reports of studies using CGH for MCL.

Figure 2.

Genome-wide frequency of copy number alterations for 29 patients. (a) Frequency of copy number gains. (b) Frequency of copy number losses. Clones are ordered from chromosome 1 to 22 and within each chromosome according to their Sanger Center mapping position, May 2004 version

Full figure and legend (143K)

Table 1 - Recurrent and most frequent regions of genomic gain and loss.

Full table

Recurrent regions of high-level gains (log₂ ratio >+1.0) were found at 10p12.2 (two cases, BMI-1 gene locus), and recurrent regions of homozygous loss (log₂ ratio <-1.0) at 2p11.2 (three cases, Ig gene locus) and 9p21.1–p24.1. As shown in Table 2, the most frequently homozygously lost clone at 9p21–p24 was RP11-149I2 (five cases), which contains the p16^INK4a tumor suppressor gene.

Table 2 - List of BAC clones showing homozygous loss (log₂ ratio <-1.0).

Full table

All the 59 clones on chromosome X were analysed separately because of sex mismatching, but the genomic alterations of X chromosomes were analysed only for the 17 male patients. Two cases showed low-grade copy number gains of Xq28, while three cases showed heterozygous (two cases) or homozygous (one case) loss of Xp21.3–p22.3 with the most frequent region at Xp22.31–p22.32.

Genomic imbalances of MCL cell lines

Genomic imbalances generally occurred more frequently in MCL cell lines than in patients. For example, gains of 7p21 and 8q24 (both n=3, 43%), and losses of 9p21 and 11q22 (both n=6, 86%), 1p22 (n=5, 71%) and of 6q, 8p23 and 13q21 (all n=3, 43%) were more frequently detected in MCL cell lines than in patient samples.

Recurrent regions of high-level gain were detected at 13q31.3 (n=2, C13orf25 gene locus) and at 18q21 (n=2, BCL2 gene locus).

Recurrent regions of homozygous loss were detected at 9p21.1–p24.1, 2p11.1 and 2q13 as seen in Table 2, which lists the homozygously lost clones of either patient samples or cell lines. Three cell lines (SP-53, Z-138 and Jeko-1) showed homozygous loss of 2q13 (log₂ ratio <-1.0), while two (REC-1 and NCEB-1) showed heterozygous loss at 2q13 (-1log₂ ratio-0.2). Five patient samples with 2q13 deletion also displayed a heterozygous loss pattern. Individual partial genomic profiles of a patient sample (G468), and of cell lines SP-53, Z-138 and Jeko-1 of chromosome 2, are shown in Figure 3, which clearly indicates that the lowest locus of loss of 2q was at BAC, RP11-438K19 (BAC438K19), which contains two genes, BIM and ACOXL (Acyl-CoA dehydrogenase gene). The former is a BH3-only Bcl-2 family member protein that promotes apoptosis (O'Conner et al., 1998), while the function of the latter remains unknown. As no information has been published regarding target gene(s) of homozygous loss at 2q13, we next searched for the minimum common region of loss of 2q13 to help us detect candidate target gene(s).

Figure 3.

Genomic profiles of chromosome 2q from a patient sample (G468) and from three MCL cell lines (SP-53, Z-138 and Jeko-1). Log₂ ratio over +0.2 represents genomic copy number gain, and a log₂ ratio below -0.2 genomic copy number loss. Physical distances (Mb) from the 2q centromere are indicated. The vertical lines indicate the lowest locus of chromosome 2 at BAC438K19 containing the BIM gene. Log₂ ratios were -0.59 (G468), -2.75 (SP-53), -1.71 (Z-138) and -1.76 (Jeko-1) at BAC438K19, suggesting that homozygous loss occurs at the BIM gene locus

Full figure and legend (123K)

Southern blot analyses

To detect the target gene of 2q13 loss, we performed Southern blot analyses of seven MCL cell lines using six genomic probes, which were designed from the genomic DNA of BAC438K19 (Figure 4a). The analyses using probes 1–6 demonstrated that partial exons of BIM were commonly deleted in the three cell lines SP-53, Z-138 and Jeko-1, and that the 'minimum common region' of homozygous loss is the BIM but not the ACOXL gene locus (Figure 4b). Here, 'minimum common region' represents the portion of the region that is aberrant in MCL cell lines with aberrations in a region. The minimum common region of homozygous loss of 2q13 of these three cell lines ranges at least from probes 2 to 3 (15 kb) and at most from probes 1 to 4 (45 kb). This region includes the open reading frame of BIM but no other gene according to the NIBC, Ensembl Genome Data Resources and UCSC Genome Bioinfomatics. Southern blot analyses were also performed with probes from BAC, RP11-368A17 (probe 7) and BAC, RP11-537E18 (probe 8) for seven MCL cell lines. BAC, RP11-368A17 (BAC368A17) is a clone with a 1.55 Mb telomeric to BAC438K19, and BAC, RP11-537E18 (BAC537E18) a clone with a 1.85 Mb centromeric to BAC438K10. Bands of probes 7 and 8 were positive in all seven MCL cell lines (data not shown), indicating that the region of homozygous deletion of each cell line (SP-53, Z-138 and Jeko-1) is at a maximum of 3.4 Mb.

Figure 4.

Minimum common region of homozygous loss at 2q13 and expression of BIM. (a) Schematic illustration of BAC438K19, the BIM gene exons, and loss patterns of three cell lines (SP-53, Z-138 and Jeko-1). Gray boxes: exons (open reading frames) of BIM EL and BIM L. The open reading frame of BIM EL (597 bp) consists of three exons: exon 1 from 75 082 to 75 475 bp (394 bp) including the initiating codon (ATG), exon 2 from 49 074 to 49 177 bp (104 bp), and exon 3 from 34 990 to 35 088 bp (99 bp) including the termination codon (TGA), all on BAC438K19. Black and white circles: probes used for Southern blot analyses. Broken horizontal lines with white circles: homozygous loss (bands negative). Thick horizontal lines with black circles: no homozygous loss (bands positive). Thin horizontal lines: not confirmed whether homozygous loss or not. Bold broken horizontal arrows between probes 2 and 3 indicate the minimum common region of homozygous loss of 2q13. (b) Southern blot analyses using probes 1–6 for genomic DNAs of MCL cell lines. Lane 1, human placenta; lane 2, SP-53; lane 3, Granta 519 (G519); lane 4, Z-138; lane 5, REC-1; lane 6, NCEB-1; lane 7, Jeko-1; lane 8, JVM2. Bands of probe 1: human placenta (+), SP-53 (-), Granta 519 (+), Z-138 (+), REC-1 (+), NCEB-1 (+), Jeko-1 (-), and JVM2 (+). Bands of probes 2 and 3: human placenta (+), SP-53 (-), Granta 519 (+), Z-138 (-), REC-1 (+), NCEB-1 (+), Jeko-1 (-), and JVM2 (+). Bands of probe 4: human placenta (+), SP-53 (-), Granta 519 (+), Z-138 (-), REC-1 (+), NCEB-1 (+), Jeko-1 (+/-), and JVM2 (+). Bands of probes 5 and 6: human placenta (+), SP-53 (-), Granta 519 (+), Z-138 (-), REC-1 (+), NCEB-1 (+), Jeko-1 (+), and JVM2 (+). 'Control' indicates the representative control band of probe 3 (TCR probe) located under the bands of probe 6. (c) Northern blot analysis of BIM with seven MCL cell lines, B-cell lymphoma (Karpas 231) and Burkitt's lymphoma (Raji) cell lines. Control is -actin

Full figure and legend (161K)

Furthermore, we performed Southern blot analysis of patient samples for which materials were available (Figure 5a), and found a heterozygous deletion pattern in a patient sample (G468) that showed heterozygous deletion at 2q13.

Figure 5.

Southern blot and FISH analysis of a patient sample (G468). (a) Southern blot analysis. Lane 1: human placenta. Lanes 2, 3, 5: Patient samples without 2q13 deletion. Lane 4: G468 showing 2q13 loss by array CGH (see Figure 3). Lane 6: Jeko-1 cell line showing homozygous deletion at BIM locus. Probe 2 that contain BIM exon was used in this experiment. (b) Dual-color FISH analysis with probes A and B of G468. Probe A: BAC438K19; probe B: BAC368A17. Probe B is 1.55 Mb telomeric to probe A, and BAC438K19 contains the BIM gene. Interphase chromosomes have two pairs of red signals (probe B, red), and one pair of green signals (probe A, green), indicating heterozygous loss of probe A

Full figure and legend (119K)

Northern blot analysis

To examine the expression of BIM in MCL cell lines, Northern blot analysis was performed of seven MCL cell lines, one FCL cell line (Karpas 231) and one Burkitt's cell line (Raji). As shown in Figure 4c, three transcripts of BIM, one major (5.7 kb) and two minor (3.8 kb and 1.35 kb) bands, were observed in Granta 519, JVM2, Karpas 231 and Raji, whereas no or very weak expression was detected in SP-53, Z-138, Jeko-1, REC-1 and NCEB-1. Although array CGH data showed a heterozygous pattern at 2q13 in REC1 and NCEB1 cell lines, Northern blot analysis indicated that BIM mRNA in these two cell lines was clearly downregulated, which will be the result of a gene dosage effect.

Fluorescence in situ hybridization (FISH)

Dual-color FISH using a combination of BAC438K19 and BAC368A17 (1.55 Mb telomeric to BAC438K19) and one of BAC438K19 and BAC537E18 (1.85 Mb centromeric to BAC 438K10) was performed on three MCL cell lines (SP-53, Z-138 and Jeko-1) and a patient sample (G468). These three clones were placed contiguously on our array CGH glass slide. Results of dual-color FISH analysis using BAC438K19 and BAC368A17 for the patient sample (G468) are shown in Figure 5b. FISH results for these cell lines (data not shown) correlated well with the array CGH data. (i) In the SP-53 cell line, no signal of BAC438K19 was found, whereas two pairs of BAC368A17 signals, or one pair of BAC537E18 signals was observed, indicating homozygous deletion of the BAC438K19 clone (log₂ ratio=-2.74). (ii) In the Z-138 cell line, one pair of weak BAC438K19 signals was detected but two pairs of normal BAC368A17 signals or one pair of normal BAC537E18 signals was observed, suggesting intra-BAC438K19 deletion in this cell line (log₂ ratio=-1.71). (iii) In the Jeko-1 cell line, one pair of weak BAC438K19 signals but two pairs of normal BAC368A17 signals or one pair of weak BAC537E18 signals was observed, suggesting the deletion of intra-BAC438K19 in this cell lines (log₂ ratio=-1.76). These observations are concordant with the finding of total BAC438K19 deletion in SP-53 and partial BAC438K19 homozygous deletion in the Z-138 and Jeko-1 cell lines.

Discussion

In the study reported here, high-resolution mapping of copy number changes was achieved for the entire MCL genome. Frequent gains and losses could be identified with high resolution by means of array analysis, which allows for precise mapping of genomic aberrations. Although numerous genomic changes of MCL were identified by array CGH, many of them were the same as those previously listed in reports of studies using chromosomal CGH (also known as conventional CGH). Several authors (Monni et al., 1998; Beà et al., 1999; Bentz et al., 2000; Martinez-Climent et al., 2001; Allen et al., 2002) reported recurrent regions of gain as 3q (40–70%), 6p (20%), 7p (27%), 8q (20–30%), 10p (20%), 12q (20–30%), 18q21 (20%) and recurrent regions of loss as 1p (24–33%), 6q (27–37%), 8p (20–30%), 9p (16–30%), 11q (22–30%) and 13q (40–60%). However, the incidence of genomic aberrations identified by array CGH was generally higher than that reported in chromosomal CGH studies. For example, our data show more frequent losses of 1p21–p22 (52%) and 9p21 (41%, INK4/ARF locus), as well as of 11q22 (55%, ATM locus), than the corresponding losses previously detected by chromosomal CGH. Among these frequent losses, although the candidate target gene of 1p22 loss could not be identified, our array CGH analysis showed a most frequent region of loss within the 5.8 Mb region of 1p21.3–p22.1.

Recently, Kohlhammer et al., 2004 reported the results of their study of 49 patients for which they used array-based (matrix) CGH with glass slides on which 812 artificial chromosomes were spotted, and found higher frequencies of genomic alterations of MCL than those seen in chromosomal CGH data. Their patient characteristics (e.g. percentage of stage III/IV, poor performance status, high LDH level, leukemic MCL and extra-nodal involvement) were almost the same as those of our series, as were the frequencies of genomic alterations. However, they could not identify several regions of loss detected by us, such as 1p36, 1q42.2–q43, 2p11.2, 2q13, 17p13.3 and 19p13.2–p13.3, because their clones were not selected from throughout the genome. The superior resolution of our study can thus be attributed to the unbiased selection of artificial chromosome clones from throughout the genome.

Of the six novel genomic regions of loss detected in our study, loss of 17p13.3 deserves special comment because it is highly interesting in that our array CGH analysis of 17p showed the most frequently deleted region(s) at 17p13.3, which suggests the existence of an additional tumor suppressor gene(s) distal to the SP53 gene. Frequent allelic loss at 17p13.3 independent of the SP53 locus has also been found in a variety of other human malignancies including lung, breast, ovarian and hepatocellular carcinomas as well as neural tumors (Fujimori et al., 1991; Cogen et al., 1992; Saxena et al., 1992; Phillips et al., 1996; Schultz et al., 1996; Konishi et al., 2002). Although SP53 mutation in MCL is a well-known genomic alteration and is associated with variant cytology and poor prognosis (Greiner et al., 1996; Hernandez et al., 1996), our finding indicated that other candidate tumor suppressor gene(s) at 17p13.3 may also be involved in the lymphomagenesis of MCL.

The key biological value of high-resolution array CGH lies in its ability to detect small, high-level gains in copy numbers and homozygous deletions that are capable of harboring specific oncogenes and tumor suppressor genes. Recurrent regions of high-level copy number gains have been identified as 10p12.2 (BMI-1), 13q31.3 (C13orf25) and 18q21 (BCL2) (Beà et al., 2001; Hofmann et al., 2001; Martínez et al., 2003; Ota et al., 2004). Since biallelic (homozygous) loss is considered to be a hallmark of chromosomal regions harboring tumor suppressor genes (Knudson, 1971), the detection of recurrent regions of homozygous loss at 2p11, 2q13 and 9p21–p24 is significant. Loss of 2p11 may be due to immunoglobulin gene rearrangement, but the loss region of 9p21–p24 covers nearly 15 Mb, making it difficult to identify the responsible gene(s) even though this region features the most frequently and homozygously deleted clone, RP11-149I2, which contains the p16^INK4a gene, which may well be the candidate gene for this region of homozygous loss of MCL (Dreyling et al., 1997; Pinyol et al., 1997).

While no previous studies have reported any candidate target gene of 2q13 among the three homozygous loss regions, our study showed that the minimum common region of 2q13 loss contains partial exons of BIM but no other genes or ESTs. This suggests that BIM appears to be the most likely target of this region of loss. It has recently become known that disturbances of pathways associated with apoptosis also contribute to the development of MCL (Hofmann et al., 2001; Martínez et al., 2003). Another study found that BIM is a proapoptotic BCL2 family member and a major physiological antagonist of BCL2, particularly in hematopoietic systems (Bouillet et al., 2002), and Enders et al. (2003) recently reported that B lymphocytes lacking Bim are refractory to apoptosis induced by B-cell receptor ligation in vitro. Finally, Egle et al. (2004) using Bim^-/- and Bim^+/- E-Myc mice, demonstrated that the loss of Bim was related to the onset of oncogenesis. These findings strongly suggest that BIM could be a tumor suppressor gene.

We demonstrated that the minimum common region of loss of 2q13 in MCL cell lines occurred at the BIM locus. Furthermore, we confirmed that the BIM expression of five out of seven MCL cell lines was downregulated, while normal expression was found in two MCL cell lines without deletion of 2q13. These results constitute a powerful indication that BIM is the most likely candidate target gene of 2q13 loss/deletion and that its downregulation may contribute to tumorigenesis of MCL.

In summary, the use of high-resolution array CGH technology for a detailed study of MCL allowed for an accurate identification of genomic aberrations and identification of BIM as a possible novel candidate tumor suppressor in MCL.

Materials and methods

MCL patients and samples

Tumor specimens obtained from 29 MCL cases, comprising 16 males and 13 females all from the Aichi Cancer Center, were included in the study. In all, 24 cases were classified as typical and five as blastoid variants. The median age of the patient was 67 years (49–92 years old). Out of 27 cases, 18 (67%) were leukemic (data of 27 cases were available), 27 out of 29 cases (93%) were in an advanced stage III–IV, eight out of 27 cases (30%) had elevated LDH (data of 27 cases were available), five out of 28 cases (18%) had a poor performance status (data of 28 cases were available) and 21 out of 27 cases (78%) had more than one extranodal site of involvement (data of 27 cases were available). The immunophenotype of the tumors was determined by immunohistochemistry for tissue sections and/or flow cytometry for cell suspensions. These studies used Ig light and heavy chains, several B-cell (CD19, CD20, CD22, CD45RA and CD79a) and T-cell (CD2, CD3, CD5, CD7, CD4, CD8, CD45RO and CD43) markers, CD10 and CD23. CCND1 expression was examined in all cases by Northern blot analysis and/or immunohistochemistry (Suzuki et al., 1999). All tumors included in this study had a B-cell phenotype, co-expressed CD5 and showed CCND1 overexpression.

MCL cell lines

The seven MCL-derived cell lines used were SP-53, Granta 519, Z-138, REC-1, NCEB-1, Jeko-1 and JVM2. All these MCL cell lines have been thoroughly characterized in terms of morphology, immunophenotype and/or interphase cytogenetics (detection of t(11;14)(q13;q32)) (Saltman et al., 1988; Jeon et al., 1998; Amin et al., 2003). The JVM2 cell line, derived from a prolymphocytic leukemia and carrying t(11;14)(q13;q32), was also included in the study. Z-138, NCEB-1, Granta 519, REC-1 and JVM2 were kindly provided by Dr Martin Dyer of Leicester University, UK. Karpas 231 derived from follicular lymphoma (FCL) and carrying t(14;18)(q32;q21) was kindly provided by Dr Abraham Karpas of the Medical Research Council Center, Hills Road, Cambridge, UK (Nacheva et al., 1993). The Raji cell line was derived from Burkitt's lymphoma.

DNA and RNA samples

High molecular weight DNA was extracted from 29 lymph nodes using standard Proteinase K/RNAse treatment and phenol–chloroform extraction. Normal DNA was obtained from male and healthy blood donors. RNAs were prepared from cell lines by homogenization in guanidinium thiocyanate and centrifugation through cesium chloride.

Array-based CGH

The array consisted of 2348 BAC and PAC clones, covering the human genome at a resolution of roughly 1.3 Mb, from libraries RP11 and RP13 for BAC clones, and RP1, RP3, RP4 and RP5 for PAC clones. BAC and PAC clones were selected from the information in NCBI (http://www.ncbi.nlm.nih.gov/), Ensembl Genome Data Resources (http://www.ensembl.org/) and UCSC genome Bioinfomatics (http://www.ncbi.nlm.nih.gov/), and obtained from the BACPAC Resource Center at the Children's Hospital (Oakland Research Institute, Oakland, CA, USA). Clones were ordered from chromosomes 1 to 22 and X within each chromosome on the basis of Ensembl Genome Data Resources from the Sanger Center Institute, February 2004 version. The locations of all the clones used for array CGH were confirmed by fluorescence in situ hybridization (FISH). Clone names and their chromosome locations are available on request. The template for degenerate oligonucleotide-primed PCR (DOP-PCR) consisted of 10 ng of BAC (or PAC) DNA. DOP-PCR products were ethanol precipitated and dissolved in DNA spotting solution DSP0050 (Matsunami, Osaka, Japan) and robotically spotted in duplicate onto CodeLink™ activated slides (Amersham Biosciences, Piscataway, NJ, USA) using the inkjet technique by a ceramic actuator (NGK, Nagoya, Japan). Fabrication and validation of the array, hybridization methods and analytical procedures have been described elsewhere in detail (Ota et al., 2004). Briefly, 1 g of tested (tumor or normal) and of referenced (normal) DNA was digested with DpnII and labeled with the BioPrime DNA labeling system (Invitrogen Life Technologies, Inc., Tokyo, Japan) using Cy3-dUTP and Cy5-dUTP (Amersham Pharmacia Biotech, Piscataway, NJ, USA) for the tested and referenced DNA, respectively. Test and reference DNAs were then mixed with 100 g of Cot-1 DNA (Life Technologies, Inc., Gaithersburg, MD, USA), precipitated and resuspended in 45 l of a hybridization solution (50% formamide, 10% dextran sulfate, 2 SCC, 4% SDS and 100 g tRNA) and hybridized onto a glass slide. After 48–66 h hybridization, the slide was washed and scanned with an Agilent Micro Array Scanner (Agilent Technologies, Palo Alto, CA, USA) and the acquired array images were analysed with Genepix Pro 4.1 (Axon Instruments, Inc., Foster City, CA, USA). After automatic segmentation of the DNA spots and subtraction of the local background, intensities of the signals were determined. Subsequently, ratios of the signal intensity of two dyes (Cy3 intensity/Cy5 intensity) were calculated for each spot and converted into log₂ ratios on an Excel sheet in the order of chromosomal position. For the array, six simultaneous hybridizations of normal male versus normal male were performed to define the normal variation for the log₂ ratio. A total of 113 clones with less than 10% of the mean fluorescence intensity of all the clones, with the most extreme average test over reference ratio deviations from 1.0 and with the largest s.d. in this set of normal controls were excluded from further analyses. Thus, we analysed a total of 2235 clones (covered 2988 Mb, 1.3 Mb of resolution) for further analysis. Out of 2235, 2176 clones (covered 2834 Mb) were from chromosome 1p telomere to 22q telomere. 59 out of 2235 clones were from chromosome X. Since more than 96% of the measured fluorescence log₂ ratio values of each spot (2 2235 clones) ranged from +0.2 to -0.2, the thresholds for the log₂ ratio of gains and losses were set at the log₂ ratios of +0.2 and -0.2, respectively. Regions of low-level gain/amplification were defined as log₂ ratio +0.2 to +1.0, those suggested of containing a heterozygous loss/deletion as log₂ ratio -1.0 to -0.2, those showing high-level gain/amplification as log₂ ratio >+1.0, and those suggested of containing a homozygous loss/deletion as log₂ ratio <-1.0.

Southern blot analyses

To detect the target gene of 2q13 loss, probes 1–6 were designed from genomic DNA on BAC 438K19. The length of BAC 438K19 (Accession number: AC096670) is 179 497 bp. Probes 7 and 8 were designed from genomic DNA on BAC 368A17 (1.55 Mb telomeric to BAC438K19) and BAC537E18 (1.85 Mb centromeric to BAC 438K19), respectively. Probes 1–8 used by Southern blot analysis were amplified with the PCR method using eight primer pairs from human placenta DNA. The primer pairs used for PCR were Probe 1 (850 bp): sense (BAC438K19: 30 851–30 874 bp), 5'-ttgcacaagtaaagtggcaattac-3'; antisense (BAC438K19: 31 700–31 677 bp), 5'-atccctgacaactcagcgtttaga-3', Probe 2 (837 bp): sense (BAC438K19: 34 214–34 237 bp), 5'-acgaatggttatcttacgactgtt-3'; antisense (BAC438K19: 35 050–35 027 bp), 5'-atctatgcatctgagtccagactg-3', Probe 3 (850 bp): sense (BAC438K19: 49 071–49 094 bp), 5'-taccctccttgcatagtaa gcgtt-3'; antisense (BAC438K19: 49 920–49 897 bp), 5'-tagtgacag cttaatgaaagggca-3', Probe 4 (811 bp): sense (BAC438K19: 75 127–75 150 bp), 5'-gggtttgtgttgatttgtcacaac-3'; antisense (BAC438K19: 75 937–75 914 bp), 5'-tgctgccctcagcattttcggcaa-3', Probe 5 (1095 bp): sense (BAC438K19: 80 501–80 524 bp), 5'-ggg tttgtgtt gatttgtcacaac-3'; antisense (BAC438K19: 81 595–81 572 bp), 5'-cc gcgctggagttacaaactctat-3' and Probe 6 (890 bp): sense (BAC438K19: 177 361–177 384 bp), 5'-cattccccagaaacagatctcgtt-3'; antisense (BAC438K19: 178 250–178 227 bp), 5'-catagcattatcaatgccatcgat-3'. Probe 7 (820 bp): sense (BAC368A17: 34 301–34 320 bp), 5'-ccatagttaatgtacacagc-3'; antisense (BAC368A17: 35 101–35 120 bp), 5'-tcgcaaaccattagga actg-3'. Probe 8 (500 bp): sense (BAC537E18: 191 071–191 094 bp), 5'-ttggagccaaggtaggattaaaca-3'; antisense (BAC537E18: 191 487–191 570 bp), 5'-ctggaggaatagtgcttccagatg-3'. Probes 2–4 included an open reading frame of BIM. BIM (BIM EL) has several splice variants such as BIM L, BIM alpha, BIM beta and BIM gamma (O'Connor et al., 1998; U et al., 2001; Liu et al., 2002), and the open reading frame of BIM EL (597 bp) includes exons of these variants. Probe 4 includes the initiating codon (ATG) of BIM (BIM EL and BIM L), and Probe 2 the termination codon (TGA) of BIM. Amplifications were performed on a Thermal Cycler (Perkin-Elmer Corporation, Norwalk, CT, USA). PCR was conducted with the touchdown PCR method described elsewhere (Motegi et al., 2000). Briefly, the reactions consisted of 10 cycles of denaturation (94°C, 0.5 min), annealing (63°C, 0.5 min, 1°C decrease per two cycles), and extension (72°C, 2.5 min), followed by 25 cycles of denaturation (94°C, 0.5 min), annealing (58°C, 0.5 min), and extension (72°C, 2.5 min), and a final extension of 5 min at 72°C. The basic annealing temperature of the reaction ranged from 63 to 58°C. All PCR products were separated by electrophoresis and purified with the QIA Quick™ Gel Extraction Kit (Qiagen). TA cloning to purified PCR products was performed with the aid of pBluescriptII SK (-), and sequenced with the ABI PRISM™ 310 Genetic Analyzer (Applied Biosystems, Foster City, VA, USA). In all, 10 g of each genomic DNA sample was restriction digested for 16 h with BamH1 (for probes 1 and 6) or HindIII (for probes 2–4) and electrophoresed on a 0.8% agarose gel in 1 TBE. Gels were sequentially immersed in 0.25 M HCl for 30 min, 1.5 M NaCl/0.5 M NaOH for 30 min and 0.5 M Tris (pH 7.4)/1.5 M NaCl for 30 min. Electrophoresed DNA was then transferred onto Hybond N⁺ membranes (Amersham Pharmacia Biotech, Tokyo, Japan), washed and hybridized overnight at 65°C with [-³²P]-dCTP-labeled probes 1–8. It was then washed, first with 2 SSC and then with diminishing concentrations of SSC-0.1% N-lauryl sarcosine at 65°C, and finally exposed to BioMax™ MS films (EKC, Rochester, NY, USA).

Northern blot analysis

Northern blotting was performed with BIM EL cDNA against seven MCL cell lines, Karpas 231 (FCL) and Raji (Burkitt's lymphoma). Probes used for Northern blot analysis were amplified with the RT–PCR method using a primer pair: sense, 5'-atggcaaagcaaccttctgatgta-3'; antisense, 5'-tcaatgcattctccacaccaggcg-3'. cDNA (open reading frame, 597 bp) of BIM EL was generated from fetal brain cDNA. Total cellular RNA (10 g) was size-fractioned on a 1% agarose/0.66 M formaldehyde gel and transferred onto a Hybond-N⁺ nylon membrane (Amersham Pharmacia Biotech, Tokyo, Japan). The membranes were then hybridized overnight at 42°C with [-³²P]-dCTP-labeled probes, washed and finally exposed to BioMax™ MS films.

Fluorescence in situ hybridization

Interphase chromosomes were prepared from paraffin-embedded sample (G468) and cell lines. Dual-color FISH analysis was conducted as described previously (Nomura et al., 2003; Zhang et al., 2004). Probes used in this experiment were probe A: BAC438K19 (green) and probe B: BAC368A17 (red).

References

1. Allen JE, Hough RE, Goepel JR, Bottomley S, Wilson GA, Alcock HE, Baird M, Lorigan PC, Vandenberghe EA, Hancock BW & Hammond DW. (2002) Br. J. Haematol. 116: 291−298. | Article | PubMed | ISI | ChemPort |
2. Amin HM, McDonnell TJ, Medeiros LJ, Rassidakis GZ, Leventaki V, O'Connor SL, Keating MJ & Lai R. (2003) Arch. Pathol. Lab. Med. 127: 424−431. | PubMed |
3. Beà S, Ribas M, Hernández JM, Bosch F, Pinyol M, Hernández L, García JL, Flores T, González M, López-Guillermo A, Piris MA, Cardesa A, Montserrat E, Miró R & Campo E. (1999) Blood 93: 4365−4374. | PubMed | ISI | ChemPort |
4. Beà S, Tort F, Pinyol M, Puig X, Hernández L, Hernández S, Fernández PL, Lohuizen M, Colomer D & Campo E. (2001) Cancer Res. 61: 2409−2412. | PubMed | ISI | ChemPort |
5. Bentz M, Plesch A, Bullinger L, Stilgenbauer S, Ott G, Muller-Hermelink HK, Baudis M, Barth TF, Moller P, Lichter P & Dohner H. (2000) Genes Chromosomes Cancer 27: 285−294. | Article | PubMed | ISI | ChemPort |
6. Bigoni R, Cuneo A, Milani R, Roberti MG, Bardi A, Rigolin GM, Cavazzini F, Agostini P & Castoldi G. (2001) Leuk. Lymphoma. 40: 581−590. | PubMed | ISI | ChemPort |
7. Bouillet P, Purton JF, Godfrey DI, Zhang LC, Coultas L, Puthalakath H, Pellegrini M, Cory S, Adams JM & Strasser A. (2002) Nature 415: 922−926. | Article | PubMed | ISI | ChemPort |
8. Cogen PH, Daneshvar L, Metzger AK, Duyk G, Edwards MS & Sheffield VC. (1992) Am. J. Hum. Genet. 50: 584−589. | PubMed | ChemPort |
9. Cuneo A, Bigoni R, Rigolin GM, Roberti MG, Bardi A, Piva N, Milani R, Bullrich F, Veronese ML, Croce C, Birg F, Dohner H, Hagemeijer A & Castoldi G. (1999) Blood 93: 1372−1380. | PubMed | ISI | ChemPort |
10. Dreyling MH, Bullinger L, Ott G, Stilgenbauer S, Muller-Hermelink HK, Bentz M, Hiddemann W & Dohner H. (1997) Cancer Res. 57: 4608−4614. | PubMed | ISI | ChemPort |
11. Egle A, Harris AW, Bouillet P & Cory S. (2004) Proc. Natl. Acad. Sci. USA 101: 6164−6169. | Article | PubMed | ChemPort |
12. Enders A, Bouillet P, Puthalakath H, Xu Y, David M, Tarlinton DM & Strasser A. (2003) J. Exp. Med. 198: 1119−1126. | Article | PubMed | ISI | ChemPort |
13. Fujimori M, Tokino T, Hino O, Kitagawa T, Imamura T, Okamoto E, Mitsunobu M, Ishikawa T, Nakagama H, Harada H, Yagura M, Matsubara K & Nakamura Y. (1991) Cancer Res. 51: 89−93. | PubMed | ISI | ChemPort |
14. Greiner TC, Moynihan MJ, Chan WC, Lytle DM, Pedersen A, Anderson JR & Weisenburger DD. (1996) Blood 87: 4302−4310. | PubMed | ISI | ChemPort |
15. Hernandez L, Fest T, Cazorla M, Teruya-Feldstein J, Bosch F, Peinado MA, Piris MA, Montserrat E, Cardesa A, Jaffe ES, Campo E & Raffold M. (1996) Blood 87: 3351−3359. | PubMed | ISI | ChemPort |
16. Hinds PW, Dowdy SF, Eaton EN, Arnold A & Weinberg RA. (1994) Proc. Natl. Acad. Sci. USA 91: 709. | PubMed | ChemPort |
17. Hofmann W-K, de Vos S, Tsukasaki K, Wachsman W, Pinkus GS, Said JW & Koeffler P. (2001) Blood 98: 787−794. | Article | PubMed | ISI | ChemPort |
18. Jaffe ES. (ed). (2001) World Health Classification of Tumors: Pathology & Genetics of Tumours of Haematopoietic and Lymphoid Tissues IARC Press: Lyon and Washington.
19. Jeon HJ, Kim CW, Yoshino T & Akagi T. (1998) Br. J. Haematol. 102: 1323−1326. | Article | PubMed | ChemPort |
20. Knudson AG, Jr. (1971) Proc. Natl. Acad. Sci. USA 68: 820−823. | PubMed |
21. Kohlhammer H, Schwaenen C, Wessendorf S, Holzmann K, Kestler HA, Kienle D, Barth T, Moller P, Ott G, Kalla J, Radlwimmer B, Pscherer A, Stilgenbauer S, Dohner H, Lichter P & Bentz M. (2004) Blood 104: 795−801. | Article | PubMed | ChemPort |
22. Konishi H, Nakagawa T, Harano T, Mizuno K, Saito H, Masuda A, Matsuda H, Osada H & Takahashi T. (2002) Cancer Res. 62: 271−276. | PubMed | ChemPort |
23. Liu J-W, Chandra D, Tang S-H, Chopra D & Tang DG. (2002) Cancer Res. 62: 2976−2981. | PubMed | ISI | ChemPort |
24. Lovec H, Grzeschiczek A, Kowalski B & Moroy T. (1994) EMBO J. 13: 3487−3495. | PubMed | ISI | ChemPort |
25. Martinez-Climent JA, Vizcarra E, Sanchez D, Blesa D, Rubio-Moscardo F, Albertson DG, Garcia-Conde J, Dyer MJS, Levy R, Pinkel D & Lossos IS. (2001) Blood 98: 3479−3482. | Article | PubMed | ISI | ChemPort |
26. Martínez N, Camacho FI, Algara P, Rodríguez A, Dopazo A, Ruíz-Ballesteros E, Martín P, Martínez-Climent JA, García-Conde J, Menárguez J, Solano F, Mollejo M & Piris MA. (2003) Cancer Res. 63: 8226−8232. | PubMed |
27. Monni O, Oinonen R, Elonen E, Franssila K, Teerenhovi L, Joensuu H & Knuutila S. (1998) Genes Chromosomes Cancer 21: 298−307. | Article | PubMed | ISI | ChemPort |
28. Motegi M, Yonezumi M, Suzuki H, Suzuki R, Hosokawa Y, Hosaka S, Kodera Y, Morishima Y, Nakamura S & Seto M. (2000) Am. J. Pathol. 156: 807−812. | PubMed | ISI | ChemPort |
29. Nacheva E, Dyer MJ, Fischer P, Stranks G, Heward JM, Marcus RE, Grace C & Karpas A. (1993) Blood 82: 231−240. | PubMed | ChemPort |
30. Nomura K, Yoshino T, Nakamura S, Akano Y, Tagawa H, Nishida K, Seto M, Nakamura S, Ueda R, Yamagishi H & Taniwaki M. (2003) Cancer Genet. Cytogenet. 140: 49−54. | Article | PubMed | ChemPort |
31. Ota A, Tagawa H, Karnan S, Karpas A, Kira S, Yoshida Y & Seto M. (2004) Cancer Res. 64: 3087−3095. | PubMed | ChemPort |
32. O'Connor L, Strasser A, O'Reilly LA, Hausmann G, Adams JM, Cory S & Huang DC. (1998) EMBO J. 17: 384−395. | PubMed |
33. Phillips NJ, Ziegler MR, Radford DM, Fair KL, Steinbrueck T, Xynos FP & Donis-Keller H. (1996) Cancer Res. 56: 606−611. | PubMed | ISI | ChemPort |
34. Pinkel D, Segraves R, Sudar D, Clark S, Poole I, Kowbel D, Collins C, Kuo WL, Chen C, Zhai Y, Dairkee SH, Ljung BM, Gray JW & Albertson DG. (1998) Nat Genet. 23: 41−46.
35. Pinyol M, Hernandez L, Cazorla M, Balbín M, Jares P, Fernandez PL, Montserrat E, Cardesa A, Lopez-Otín C & Campo E. (1997) Blood 89: 272−280. | PubMed | ISI | ChemPort |
36. Rosenwald A, Wright G, Wiestner A, Chan WC, Connors JM, Campo E, Gascoyne RD, Grogan TM, Muller-Hermelink HK, Smeland EB, Chiorazzi M, Giltnane JM, Hurt EM, Zhao H, Averett L, Henrickson S, Yang L, Powell J, Wilson WH, Jaffe ES, Simon R, Klausner RD, Montserrat E, Bosch F, Greiner TC, Weisenburger DD, Sanger WG, Dave BJ, Lynch JC, Vose J, Armitage JO, Fisher RI, Miller TP, LeBlanc M, Ott G, Kvaloy S, Holte H, Delabie J & Staudt LM. (2003) Cancer Cell 3: 185−197. | Article | PubMed | ISI | ChemPort |
37. Saltman DL, Cachia PG, Dewar AE, Ross FM, Krajewski AS, Ludlam C & Steel CM. (1988) Blood 72: 2026−2030. | PubMed | ISI | ChemPort |
38. Saxena A, Clark WC, Robertson JT, Ikejiri B, Oldfield EH & Ali IU. (1992) Cancer Res. 52: 6716−6721. | PubMed | ISI | ChemPort |
39. Schaffner C, Idler I, Stilgenbauer S, Dohner H & Lichter P. (2000) Proc. Natl. Acad. Sci. USA 97: 2773−2778. | Article | PubMed | ChemPort |
40. Schultz DC, Vanderveer L, Berman DB, Hamilton TC, Wong AJ & Godwin AK. (1996) Cancer Res. 56: 1997−2002. | PubMed | ISI | ChemPort |
41. Seto M, Yamamoto K, Iida S, Akao Y, Utsumi KR, Kubonishi I, Miyoshi I, Ohtsuki T, Yawata Y, Namba M & Ueda R. (1992) Oncogene 7: 1401−1406. | PubMed | ChemPort |
42. Stilgenbauer S, Winkler D, Ott G, Schaffner C, Leupolt E, Bentz M, Moller P, Muller-Hermelink HK, James MR, Lichter P & Dohner H. (1999) Blood 94: 3262−3264. | PubMed | ISI | ChemPort |
43. Suzuki R, Kuroda H, Komatsu H, Hosokawa Y, Kagami Y, Ogura M, Nakamura S, Kodera Y, Morishima Y, Ueda R & Seto M. (1999) Leukemia 13: 1335−1342. | PubMed | ChemPort |
44. U M, Miyashita T, Shikama Y, Tadokoro K & Yamada M. (2001) FEBS Lett. 509: 135−141. | Article | PubMed | ChemPort |
45. Zhang X, Karnan S, Tagawa H, Suzuki R, Morishima Y, Nakamura S & Seto M. (2004) Cancer Science 95: 809−814. | PubMed | ChemPort |

Acknowledgements

The outstanding technical assistance of Ms H Suzuki and Y Kasugai is very much appreciated. We also wish to thank Dr Ryuzo Ohno, the chancellor of Aichi Cancer Center, for his general support. This work was supported in part by a Grant-in-Aid from the Ministry of Health, Labour and Welfare, a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology, a Grant-in-Aid (B2) from the Japan Society for the Promotion of Science, a Grant-in-Aid from the Foundation of Promotion of Cancer Research, a Grant-in-Aid for Cancer Research from the Princess Takamatsu Cancer Research Fund and a Grant-in-Aid from the China Japan Medical Association.