Characterization of target genes at the 2p15–16 amplicon in diffuse large B‐cell lymphoma

Abstract

Amplification of 2p has been observed as a recurrent alteration in diffuse large B‐cell lymphoma (DLBCL). Whereas two candidate oncogenes, REL and BCL11A, have been investigated as targets for 2p amplification, the question remains as to whether the true target gene in the amplicon is REL, BCL11A or both. We previously identified frequent genomic gains of chromosomal 2p in 25 out of 99 DLBCL cases by means of genome‐wide array comparative genomic hybridization (CGH). All of these 25 cases included recurrent copy number gain at 2p15–16. In the study presented here, cases were analyzed in greater detail by means of contig bacterial artificial chromosome (BAC) array CGH for the 4.5‐Mb region at 2p15–16, which contained 33 BAC clones. We confined the minimal common region to 500‐kb in length, where only the candidate oncogene REL, and not BCL11A, is located. Real‐time quantitative PCR was carried out to investigate the correlation between genomic gain and expression. It showed a significant correlation for both genes, indicating that these two genes are common targets for the 2p15–16 amplicon. However, given the fact that REL is more frequently amplified than BCL11A, the REL gene may play a more important role than BCL11A in the pathogenesis of DLBCL. (Cancer Sci 2006; 97: 499 – 504)

Diffuse large B cell lymphoma (DLBCL) is the most common type of malignant lymphoma, accounting for 30% of adult non‐Hodgkin's lymphoma.^{( 1 )} However, clinicopathological and genetic heterogeneities in this entity have suggested that further refinement of its subgroups is required.^{( 2} , ³ , ⁴ , ^{5 )} Array‐based comparative genomic hybridization (array‐CGH) analysis is a powerful tool to identify genomic imbalances characteristic of distinct subgroups in DLBCL.^{( 6} , ^{7 )} It has been useful not only for genome scanning of tumor cells but also for identification of novel oncogenes and suppressor genes.^{( 8 )}

We previously used a genome‐wide array‐CGH to identify a gain of 2p15–16 in 25 out of 99 DLBCL cases.^{( 7 )} Amplification at the 2p arm has been reported in B‐cell lymphomas, such as DLBCL,^{( 9} , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ^{15 )} classical Hodgkin's lymphoma (cHL),^{( 16} , ^{17 )} follicular lymphoma (FL)^{( 18 )} and primary mediastinal B‐cell lymphoma (PMBCL).^{( 11} , ¹⁴ , ^{15 )} The two candidate genes, REL and BCL11A, have been mapped within this 2p15–16 amplicon. The REL proto‐oncogene, which encodes a member of the NF‐κB transcription factor family, has frequently been found amplified in B‐cell lymphomas. BCL11A, which is located quite near REL on chromosome 2p15–16, is coamplified with REL in DLBCL and cHL.^{( 12} , ¹⁵ , ^{16 )} In spite of numerous studies of this region, the question remains whether the target gene in the 2p15–16 amplification is REL, BCL11A, or both.

For the study presented here, we carried out a contig array‐CGH using glass slides on which contiguously ordered bacterial artificial chromosome (BAC) clones were spotted throughout 4.5 Mb of the 2p15–16 genome to confine the minimal common region of amplification at 2p15–16 in DLBCL cases. Real‐time quantitative–polymerase chain reaction (RQ‐PCR) analysis was then used to further investigate the relationship between genomic amplification and expression.

Patients, Materials and Methods

Tumor samples

Tumor samples were obtained from 99 patients under a protocol approved by the International Review Board of the Aichi Cancer Center. Informed consent was obtained in accordance with the Declaration of Helsinki. All of the DNA and RNA samples were obtained from tumors at the time of diagnosis before the administration of any treatment. DNA was extracted with a standard phenol chloroform method from lymphoma specimens of the tumors. Normal DNA was prepared from peripheral‐blood lymphocytes of healthy male donors. Total RNA was extracted using the standardized guanidium isothiocyanate and cecium chloride method from lymphoma specimens taken from the tumors. Data for genomic gains and losses at the chromosome 2 region of the 99 DLBCL have been reported previously.^{( 7 )}

Genome‐wide array‐based CGH

DNA preparation, labeling, array fabrication and hybridization were carried out as described elsewhere.^{( 6} , ⁸ , ^{19 )} Briefly, the array consisted of 2213 BAC and P‐1‐derived artificial chromosome (PAC) clones, covering the whole human genome with a resolution of 1.3 Mb, from library RP11, 13 for BAC clones and RP1, and 3, 4, 5 for PAC clones. Of the 2213 clones spotted on the glass slides, 188 were of chromosome 2, and of these 188 BAC/PAC clones, 75 were of the short arm of chromosome 2 (2p). These clones were obtained from the BAC/PAC Resource Center at the Children's Hospital Oakland Research Institute, Oakland, CA, USA (http://bacpac.chori.org/). The thresholds for the log₂ ratio of gains and losses were set at the log₂ ratios of +0.2 and −0.2, respectively. High‐level copy number gain (amplification) was defined as log₂ ratio ≥+1 and low‐level copy number gain as +0.2 ≤ log₂ ratio < +1.0.^{( 8 )}

Contig array‐based CGH

Detailed analysis with contig array‐CGH was carried out for cases with amplification (n = 3) and gain (n = 4, cases with available RNA and restricted region of gain) at 2p15–16. Selection of 33 BAC clones of 2p15–16 was based on information from National Center of Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/). Each clone was placed contiguously between BAC clones RP11–518G12 and RP11–511I11 according to the mapping position obtained from the NCBI. These clones were obtained from the BAC/PAC Resource Center at the Children's Hospital Oakland Research Institute. BAC were then isolated from their bacterial cultures with the relevant antibiotics and extracted with a plasmid Mini‐Kit (Qiagen, Valencia, CA, USA). The exact location of each clone was determined by means of standard fluorescence in site hybridization (FISH) analysis. Degenerate oligonucleotide‐primed polymerase chain reaction (DOP‐PCR)^{( 20 )} was carried out on the DNA of BAC clones, as described elsewhere.^{( 8 )} DOP‐PCR products were dissolved in 30 µL of TE (100 mM Tris‐HCl and 1 mM ethylenediaminetetracetic acid, pH 7.5) buffer, and 10 µL of Solution I (Takara Bio, Tokyo, Japan) was added to each of the products, which were then spotted in triplicate onto Hubble‐activated slides (Takara Bio) using the Stampman Arrayer (Nippon Laser and Electronics Laboratory, Nagoya, Japan) with a split pin. Slides were fixed in 0.2% sodium dodecylsulfate for 2 min and in 0.3 M NaOH for 5 min, then dehydrated with 100% cold ethanol for 3 min, and finally air‐dried. DNA preparation, labeling, array fabrication and hybridization were carried out according to methods described previously.^{( 6} , ⁸ , ²¹ , ^{22 )} The Agilent Micro Array Scanner (Agilent Technologies, Palo Alto, CA, USA) was used for scanning analysis. The array images thus acquired were further analyzed with Genepix Pro 4.1 (Axon Instruments, Foster City, CA, USA).

Reverse transcription–polymerase chain reaction analysis for screening of candidate genes

The RC‐K8 cell line was established from histiocytic lymphoma cells (kindly provided by I. Kubonishi, Kochi, Japan).^{( 23 )} The L428 cell line was established from Hodgkin's lymphoma (German Collection of Microorganisms and Cell Cultures, Braunschweig, Germany).^{( 24 )} The Karpas‐1106p cell line was established from mediastinal lymphoblastic B‐cell lymphoma (kindly provided by A. Karpas, Cambridge, UK).^{( 25 )} These cell lines and human placenta^{( 26 )} were subjected to reverse transcription–polymerase chain reaction (RT‐PCR) analysis. SuperScript II (Gibco‐BRL, Gaithersburg, MD, USA) was used for cDNA. Each 5 µg of total RNA was reverse‐transcribed into cDNA in a volume of 40 µL distilled water. RT‐PCR was carried out for nine genes by using the specific corresponding primers. The names and accession numbers of the genes were: LOC442017 (XM_497839), LOC130865 (XM_497840), ATP1B3P1 (NG_000849), PAPOLG (NM_022894), LOC400957 (XM_379097), REL (NM_002908), LOC344423 (AK124741), FLJ32312 (NM_144709) and PEX13 (NM_002618). The primers used for RT‐PCR are shown in Table 1. Each primer was designed so that the melting temperature (Tm) value would be between 58°C and 63°C. Amplifications were carried out using a Thermal Cycler (Perkin‐Elmer, Norwalk, CT, USA), and RT‐PCR was carried out using the touchdown PCR method. The reactions comprised 10 cycles of denaturation (94°C, 0.5 min), annealing (65°C, 0.5 min, 1°C decrease per 2 cycles) and extension (72°C, 2.5 min), followed by 35 cycles of denaturation (94°C, 0.5 min), annealing (60°C, 0.5 min) and extension (72°C, 2.5 min), and a final extension of 5 min at 72°C. In the cases of LOC130865 and LOC442017, annealing temperature of the reaction ranged from 63 to 58°C.

Table 1.

Reverse transcription–polymerase chain reaction analysis of genes within the 2p15–16 amplification region of RC‐K8, L428 and Karpas‐1106p

Gene	Forward primer	Reverse primer	Human placenta	RC‐K8	L428	Karpas 1106p
LOC442017	5′‐ctgtccaaaccgtcttcactc	5′‐acttggcagtggaggcgtag	+/–	–	–	+/–
LOC130865	5′‐cccacactcgcagaaagatt	5′‐acagggacagctatgctgttag	+/–	+/–	+/–	+
ATP1B3P1	5′‐gcacacgatgaagaaggagtc	5′‐gcttgaagtaacgaaatgggagat	+/–	+/–	+/–	–
PAPOLG	5′‐attgacgccatgaaaccatt	5′‐gcttctcaggcgagagtcgt	+	+	+	–
LOC400957	5′‐actatagccggatacagggaga	5′‐cgtagcacccgttacacaga	+/–	+	+/–	–
REL	5′‐gaaactgtgccaggatcacg	5′‐ccaacaggtattctcaggaatgg	+	++	++	++
LOC344423	5′‐catgatggcctagcatatgaa	5′‐gctcttcttgacacctcatcaa	+	++	++	++
FLJ32312	5′‐tctgttagtcatgttggaagcact	5′‐ggctttcataactgccattcta	+/–	–	+	+
PEX13	5′‐acaaccgcctccgtgtaga	5′‐ctctggcaactacatggtcatc	+	++	++	++

Detected by electrohphoresis: ++, positive thick band; +, positive thin band; +/–, very weakly positive band; –, no band.

Real‐time quantitative–polymerase chain reaction

Expression levels of REL, BCL11A, LOC344423 and PEX13 mRNA were measured by means of real‐time fluorescence detection using a previously described method.^{( 22} , ^{27 )} Briefly, the primers of REL were sense: 5′‐cccacgctcaggcaataca‐3′ and antisense: 5′‐tggtgggataccttgcgaat‐3′, those of BCL11A were sense: 5′‐aaaaagagagaaacaaaaaagtgtgaca‐3′ and antisense: 5′‐catccatgtgacattctagcagg‐3′, those of LOC344423 were sense: 5′‐ gccatgcactagagggtactca‐3′ and antisense: 5′‐gctcttcttgacacctcatcaa‐3′, and those of PEX13 were sense: 5′‐aggaccgagcagctacctca‐3′ and antisense: 5′‐tggcaactacatggtcatcctc‐3′. Real‐time PCR using SYBR® Green1 and primers was carried out with a Smart Cycler System (Takara Bio) according to the manufacturer's protocol. G6PDH served as an endogenous control, whereas the expression levels of REL, BCL11A, LOC344423 and PEX13 mRNA in each sample were normalized on the basis of the corresponding G6PDH content and recorded as relative expression levels. Overexpression was defined as the mean of the relative expression plus two or more standard deviation units.

Results

Recurrent amplification detected by genome‐wide array‐CGH at 2p15–16 in DLBCL

Genome‐wide array‐CGH analysis at a resolution of 1.3 Mb showed that 25 of the 99 DLBCL cases (25%) had copy number gains on chromosome 2p. All cases included recurrent copy number gain at 2p15–16, with three of the 25 cases showing genomic amplification (log₂ ratio ≥ 1). Individual genomic profiles of these three tumors (D768, D778 and D792) showed that every amplification was located at 2p15–16 (Fig. 1). The common recurrent region of the 25 cases was confined to the 4.5‐Mb region at 2p15–16 (Fig. 2).

Figure 1.

Representative individual genomic profiles of chromosome 2 in diffuse large B‐cell lymphoma. Genome‐wide array‐based comparative genomic hybridization profiles of three cases with 2p amplification: (a) D768, (b) D778 and (c) D792. Dots represent the log₂ ratio of bacterial artificial chromosome/P‐1‐derived artificial chromosome clones, which are shown in order from the p telomere to the q telomere. The vertical arrow above each profile indicates the region of amplification. The threshold for gain and loss was defined as the log₂ ratio of +0.2 and −0.2, respectively.

Figure 2.

Illustration of genomic amplification at chromosome 2p. Summary of genome‐wide array‐based comparative genomic hybridization profiles at 2p. Copy number gains were detected in 25 of 99 diffuse large B‐cell lymphoma cases. Thin lines: low copy number gain (+0.2 ≤ log₂ ratio < +1.0); thick lines: high copy number gain (amplification, log₂ ratio ≥ +1.0). Genomic amplifications were observed at 2p15–16 in three cases (D768, D778 and D792). The gray area represents the most common recurrent region in the 25 cases, which is 4.5 Mb in length.

Determination of minimal common region by contig array‐CGH

We speculated that the target genes of 2p15–16 amplification were located within the common recurrent region between BAC clones RP11–518G12 and RP11–511I11. To specify the alterations of 2p15–16 in more detail, we prepared high‐resolution contig array glass slides containing 33 BAC clones, which were placed contiguously throughout the 4.5‐Mb region. Contig array‐CGH was conducted for seven cases, for which the genome‐wide array‐CGH showed gain or amplification at 2p15–16 (Fig. 3). The profiles showed that the common region for three cases with amplification (log₂ ratio  1.0) was 1.6 Mb located between BAC clones RP11–440P5 and RP11–813L21. Further analysis for four cases with genomic gain confined the minimal common region to 500 kb between BAC clones RP11–416L21 and RP11–373L24. The region contained only REL, not BCL11A, as the candidate oncogene.

Figure 3.

Contig array‐based comparative genomic hybridization (array‐CGH) profiles at 2p15–16 in diffuse large B‐cell lymphoma. Contig array‐CGH containing 33 bacterial artificial chromosome (BAC) clones were constructed and placed contiguously at the most common recurrent region (4.5 Mb) identified with genome‐wide array‐CGH. (a) Genomic profiles of three cases with high copy number gain (log₂ ratio ≥ +1.0) at 2p15–16. (b) Genomic profiles for four cases with low copy number gain (+ 0.2 ≤ log₂ ratio < +1.0) at 2p15–16. Dots represent the log₂ ratio of BAC clones, which are shown in order from the p telomere to the centromere. Vertical lines: log₂ ratio. The thresholds for gain and loss were defined as the log₂ ratio of +0.2 and −0.2, respectively. The BAC clones are all shown contiguously. The underlined BAC clones were used for genome‐wide array‐CGH. The gray area represents the minimum common region in the 25 cases, which is 500‐kb in length between BAC clones RP11–416L21 and RP11–373L24.

Relationship between genomic alteration and expression of REL and BCL11A

Four of the seven cases analyzed by contig array‐CGH showed gain or amplification for both REL and BCL11A genes, whereas the remaining three demonstrated gain or amplification of REL alone (Fig. 4a). To investigate the relationship between genomic amplification and gene expression, we conducted RQ‐PCR analysis for both candidate genes in six of the seven cases with 2p15–16 gain whose RNA was available, and in seven cases without any change at 2p15–16 (Fig. 4b). Three cases with copy number gains for both genes (D778, D856 and D1767) demonstrated overexpression for both, and two cases with a copy number gain of REL alone (D768 and D1664) showed significant overexpression of REL, but the remaining case (D754) showed an expression level beyond the average for normal controls, although the difference was not significant. These three cases showed neither genomic gain nor overexpression of BCL11A (D768, D1664 and D754). Correlation coefficients for the log₂ ratio of DNA and the relative amount of mRNA were 0.76 for REL and 0.89 for BCL11A. These data indicate that overexpression of REL and BCL11A correlates well with the level of genomic amplification of both genes. Thus, both of these two genes are targets at the 2p15–16 region, whose expression is altered by genomic amplification in DLBCL. Furthermore, we also conducted RQ‐PCR in an additional six DLBCL cases with 2p gain, for which contig array‐CGH could not be carried out. All of the additional cases except one showed REL overexpression in accordance with genomic gain, whereas only four cases showed BCL11A overexpression (data not shown).

Figure 4.

Relationship between genomic amplification and expression of REL, BCL11A, PEX13 and LOC344423. (a) Mean log₂ ratio of all bacterial artificial chromosome (BAC) clones containing the REL locus and the BCL11A locus obtained from contig array‐based comparative genomic hybridization (array‐CGH). The thresholds for gain and loss were defined as the log₂ ratio of +0.2 and −0.2, respectively. White bars: REL locus. Black bars: BCL11A locus. (b) Relative expression levels of REL and BCL11A determined by real‐time quantitative–polymerase chain reaction (RQ‐PCR) in cases with 2p15–16 gain and those without any change at 2p15–16 (normal). For relative expression levels, each expression was normalized on the basis of the corresponding G6PDH content. White bars: REL. Black bars: BCL11A. *ND: RQ‐PCR not done for D792 because RNA was not available. (c) Mean log₂ ratio of all BAC clones containing the PEX13 locus and the LOC344423 locus obtained from contig array CGH. The thresholds for gain and loss were defined as the log₂ ratio of +0.2 and −0.2, respectively. Dotted bars: PEX13 locus. Striped bars: LOC344423 locus. (d) Relative expression levels of PEX13 and LOC344423 detected by RQ‐PCR in cases with 2p15–16 gain and of cases without any change at 2p15–16 (normal). For relative expression levels, each expression was normalized on the basis of the corresponding G6PDH content. Dotted bars: PEX13. Striped bars: LOC344423. *ND: RQ‐PCR not done for D792 because RNA was not available.

RT‐PCR analysis of the genes within the 2p15–16 amplicon, and RQ‐PCR for LOC344423 and PEX13

The contig array‐CGH showed that nine genes were located within the 2p15–16 minimum common region (500 kb in length). RT‐PCR analysis was used to evaluate expression of these nine genes (Table 1). The cell lines used for screening were B‐cell lymphoma cell lines with 2p amplification (RC‐K8, L428^{( 28 )} and Karpas‐1106p^{( 29 )}) and human placenta. The sizes of all products obtained by RT‐PCR were confirmed by electrophoresis and were as expected (data not shown). Expression of three genes (REL, LOC344423 and PEX13) could be detected in all B‐cell lymphoma cell lines with 2p amplification. We conducted RQ‐PCR analysis for LOC344423 and PEX13 in the same six cases with 2p15–16 gain to investigate the relationship between genomic amplification and gene expression (Fig. 4c,d). All cases but one showed overexpression of LOC344423 and PEX13 in a similar fashion to that of REL. Correlation coefficients for the log₂ ratio of DNA and the relative amount of mRNA were 0.73 for LOC344423 and 0.97 for PEX13. Thus, both of these genes are also possible targets in the 2p15–16 region in DLBCL, although functional study is needed.

Discussion

Gain of chromosome 2p has been identified as a recurrent alteration in 20% of DLBCL,^{( 9} , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ^{15 )} 50% of cHL,^{( 16} , ^{17 )} and 3–47% of PMBCL cases.^{( 11} , ¹⁴ , ^{15 )} In previous studies, CGH and FISH analyses were used as tools to investigate genomic alterations. Array‐CGH is superior to these two methods in terms of higher resolution and the ability to directly map the copy number changes to the genome sequence. The genome‐wide array‐CGH with a resolution of 1.3 Mb used for the 99 DLBCL cases enrolled in our study identified a common recurrent 4.5‐Mb region at 2p15–16. By carrying out high‐resolution contig array‐CGH in the recurrent region, we were able to restrict the minimal common region to 500 kb in length.

The REL proto‐oncogene, which is located in the 2p15–16 region, has been investigated as a target of the region in many reported studies. As BCL11A was identified as another oncogene in this region,^{( 30 )} the two oncogenes have been examined together. In most cHL cases, coamplification of both gene loci was shown by fluorescence immunophenotyping and interphase cytogenetics (FICTION). Bea et al. detected simultaneous overexpression of both genes in all DLBCL cases with 2p amplification by means of RQ‐PCR.^{( 12 )}

In the present study, we were able to identify the minimally targeted genomic regions of 2p15–16 amplification where REL, not BCL11A, is located as the only candidate oncogene. Martin‐Subero et al. investigated amplification of both gene loci, REL and BCL11A, in cHL tumors by means of FICTION and reported that one case displayed selective amplification of the REL locus and two cases showed signal patterns suggesting breakpoints within the REL locus.^{( 16 )} They concluded that the target of 2p alterations might be closely associated with the REL locus, consistent with the region we identified as a minimum common region, although they did not investigate gene expression.

It has been reported that a high copy number of REL correlates with extranodal disease in DLBCL.^{( 9 )} A recent multicenter cooperative study demonstrated that expression of REL and BCL11A was frequently recognized in the germinal center like B cell (GCB)‐DLBCL subgroup that is known to have favorable prognosis. However, the number of cases with 2p gain or amplification in our cohort was too small to draw any definitive conclusions in this respect, but it has been reported that the correlation of 2p gain and favorable outcome could not be demonstrated.^{( 15 )}

To investigate the correlation between genomic amplification and expression, we carried out RQ‐PCR analysis of the candidate genes REL and BCL11A and found a significant correlation between genomic alteration and expression in both genes. This indicates that both genes are common targets for genomic amplification in this region. However, given the fact that REL is more frequently amplified than BCL11A, it is speculated that REL may play a more important role than BCL11A in the development of lymphoma.

Our results identified two additional candidate genes as targets of the 2p15–16 amplicon and demonstrated for the first time the overexpression of LOC344423 and PEX13 in primary tumors. Because these two genes and REL are located within the same BAC (RP11–373L24), it is speculated that the LOC344423 and PEX13 loci are also amplified in cases with REL locus amplification. However, the coding frame of LOC344423 is missing nor regulatory RNA, such as micro RNA, is found, and PEX13 is a structural gene of the peroxisome. This suggests that neither LOC344423 nor PEX13 are likely candidate genes for oncogenesis.