All authors have declared that no competing interests exist. The Mayo Clinic is a nonprofit medical center and therefore should not be considered a commercial company. All the authors from Mayo Clinic have declared that no competing interests exist for this manuscript.
Conceived and designed the experiments: AJW MQD. Performed the experiments: AJW RAH NZ QY YH EB RF EDM AD. Analyzed the data: AJW RAH EB RF EDM AD MQD. Contributed reagents/materials/analysis tools: HL JZ LdL PGI AW EB RF EDM AD. Wrote the paper: AJW MQD.
¶ These authors also contributed equally to this work.
Splenic marginal zone lymphoma (SMZL) is an indolent B-cell lymphoproliferative disorder characterised by 7q32 deletion, but the target genes of this deletion remain unknown. In order to elucidate the genetic target of this deletion, we performed an integrative analysis of the genetic, epigenetic, transcriptomic and miRNomic data. High resolution array comparative genomic hybridization of 56 cases of SMZL delineated a minimally deleted region (2.8 Mb) at 7q32, but showed no evidence of any cryptic homozygous deletion or recurrent breakpoint in this region. Integrated transcriptomic analysis confirmed significant under-expression of a number of genes in this region in cases of SMZL with deletion, several of which showed hypermethylation. In addition, a cluster of 8 miRNA in this region showed under-expression in cases with the deletion, and three (miR-182/96/183) were also significantly under-expressed (
Splenic marginal zone lymphoma (SMZL) is a rare indolent B-cell non-Hodgkin’s lymphoma, being less than 2% of lymphoid malignancies
The 7q deletion may play an important role in the pathogenesis of SMZL. To investigate this, we searched for evidence of any potential tumour suppressor genes in the MDR. The classic tumour suppressor genes are often inactivated on both alleles by multiple mechanisms including homozygous deletion, heterozygous deletion and mutation, and transcriptional repression by promoter methylation. To ascertain the gene or genes targeted by the 7q deletion in SMZL, we sought evidence of homozygous deletion by gene resolution array CGH of chromosome 7. We also assessed the expression of miRNA and coding genes in the MDR and correlated their expression levels with the 7q deletion status. For the coding genes, we further correlated their expression with the methylation status. Finally, genomic sequencing was performed on the miRNA and key candidate genes in the MDR to search for inactivating mutation.
A total of 95 cases of SMZL were analysed in this study: 35 cases were the subject of a previous oligonucleotide aCGH study by the Mayo Clinic group
A total of 56 case of SMZL were investigated by aCGH using oligonucleotide array platform in this study. A series of 35 SMZL (including 10 with 7q deletion) were examined using the Agilent aCGH 244A platform (6.4 Kb resolution) (Agilent Technologies, Palo Alto, California, USA) by the Mayo Clinic group in a previous study (
The oligonucleotide array CGH was carried out essentially according to the manufacturer’s instructions. Briefly, genomic DNA was extracted from frozen tissues containing >60% tumour cells based on histological estimation with the assistance of CD20 and CD3 staining, together with retrospective review of the CGH profile
Copy-number abnormalities (CNA) were calculated using the aberration detection module (ADM)-2 algorithm. Copy number variations (CNV) were identified and excluded from the analysis by reference to the database of genetic variation (Build GRCh37, Feb 2009:
A total of 48 cases of SMZL (including 15 with 7q deletion) were analysed using the Affymetrix HG-U133 Plus 2.0 platform (Affymetrix, Santa Clara, California, USA). Arrays were performed according to the manufacturer’s instructions. Briefly, RNA was extracted from snap frozen tissues with >60% tumour cells using the RNeasy extraction kit (Qiagen) and subjected to DNAse treatment (Turbo DNAse kit, Ambion). RNA integrity was assessed using an Agilent 2100 Bioanalyzer. cDNA synthesis was carried out with 2 µg RNA using the GeneChip® One-Cycle cDNA Synthesis Kit (Affymetrix), followed by
Raw gene expression data from Affymetrix.CEL files were uploaded to Bioconductor where a combined MAS5 and gcRMA normalization procedure was performed and used to filter out non-variant probes across all samples as described previously
Epigenetic methylation analysis was performed using the Infinium® Human Methylation 27 array (Illumina, San Diego, California, USA). The array contained 27,568 CpG islands within the proximal promoter regions of transcription start sites of 14,475 RefSeq genes, including 12,883 well annotated genes (NCBI CCDS database: Build 36). The methylation array was carried out in 12 SMZL (6 cases with 7q deletion), 6 follicular lymphomas (FL) and 6 mantle cell lymphomas (MCL) as per the manufacturer’s instructions. Briefly, 2 µg genomic DNA extracted from frozen tissues with >60% tumour cells was bisulphite modified using the EZ DNA Methylation™ Kit (Zymo Research Corporation). Bisulphite modified DNA was then amplified using the MSM master mix (Illumina) and incubated at 37°C for 22 hours. Amplified DNA was then fragmented and hybridised to BeadChips in an Illumina Hybridisation Oven at 48°C for 18 hours. Following hybridisation, single base extension of hybridised DNA using hapten labelled bases was performed. Staining was then developed using immunochemical stains catalysed by the haptens, and the arrays washed. The chips were scanned using the BeadArray™ Reader (Illumina) and the BeadScan™ software (Illumina) using the Infinium Methylation Scan setting. The scanned data was then analysed in GenomeStudio™ (Illumina) using the Methylation analysis module. The data are available from GEO (GSE21554).
The scanned data was spatially normalised to correct for background noise using the Smethillium algorithm (
Total RNA was extracted from frozen tissues of 18 SMZL including 8 cases with 7q deletion, and 15 other low grade B-cell lymphomas including 7 FL, 4 MCL and 4 mucosa associated lymphoid tissue (MALT) lymphomas using the miRNEasy® kit (Qiagen). In each case, the frozen tissue used for RNA extraction contained at least 60% of tumour cells. The expression level of mature miR-593, miR-129, miR-182, miR-96, miR-183, miR-335, miR-29a, and miR-29b1 and their minor variants was investigated by qRT-PCR using the Qiagen miScript® system (Supplementary
Mutation analysis of miRNA was performed in 60 cases of SMZL. Primers flanking the predicted primary miRNA (pri-miRNA) (approximately 100 bp 5¢ and 3¢ of the predicted loop) were designed for miR-593, miR-129, miR-182, miR-96, miR-183, miR-335, miR-29a, and miR-29b1 (
Similarly,
CGH using the chromosome 7 (Ch7) BAC tile-path array, Agilent 244A oligonucleotide array platform and chromosome 7 gene resolution oligonucleotide array maps the MDR to a 2.8 Mb region at 7q32 (Panel A). As shown by Affymetrix gene expression analyses, the genes on 7q, particularly those within the MDR are significantly under-expressed in SMZL with 7q deletion than those without the deletion. Panel B shows the genes significantly differentially expressed between SMZL with and without 7q deletion according to
Within the MDR, there were 6 miRNA (miR-593, miR-129, miR-182, miR-96, miR-183, miR-335) with miR-29a/29b1 polycistron distal to the MDR, and 38 potential coding genes (
Gene | Full name | Mutation frequency | Nature of mutation | Cancer type | |
|
staphylococcal nuclease and tudor domain containing 1 | 4/93 | 4.3% | Missense,inframe insertion | Ovarian carcinoma’ primitive neuroestodermal tumour- medulloblastoma, laryngeal carcinoma |
|
leucine rich repeat containing 4 | 4/154 | 2.6% | Missense,inframe deletion | Glioma, breast carcinoma, colon carcinoma, pancreatic carcinoma |
|
leptin | 0/90 | 0% | N/A | N/A |
|
RNA binding motif protein 28 | 2/92 | 2.2% | Missense, | Ovarian carcinoma |
|
IMP (inosine 5'-monophosphate) dehydrogenase 1 | 5/231 | 2.2% | Missense, | Skin melanoma, glioma, lung carcinoma,Ovarian carcinoma |
|
methyltransferase like 2A | 2/250 | 0.8% | Missense, | Mouth squamous carcinoma, pharyngeal carcinoma |
|
family with sequence similarity 71, member F1 | 2/70 | 2.8% | Missense, nonsense | Skin melanoma, mouth squamous carcinoma |
|
2/25 | 8% | Missense, | Ovarian carcinoma | |
|
filamin C, gamma | 7/142 | 4.9% | Missense, nonsense | Skin melanoma, breast carcinoma, primitive neuroestodermal tumour-medulloblastoma, colon carcinoma, ovarian carcinoma |
|
kielin/chordin-like protein | 0/68 | 0% | N/A | N/A |
|
interferon regulatory factor 5 | 0/217 | 0% | N/A | N/A |
|
transportin 3 | 2/219 | 0.9% | Missense, nonsense | Skin melanoma, lung carcinoma |
|
ATPase, H+ transporting, lysosomal 14 kDa, V1 subunit F | 1/91 | 1% | Missense, | Ovarian carcinoma |
|
tetraspanin 33 | 1/91 | 1% | Missense, | Ovarian carcinoma |
|
smoothened, frizzled family receptor | 32/2436 | 1.3% | Missense,Inframe insertion | Skin basal cell carcinoma, primitive neuroestodermal tumour- medulloblastoma, lung carcinoma, colon carcinoma, glioma, bile duct carcinoma |
|
adenosylhomocysteinase-like 2 | 3/71 | 4.2% | Missense, nonsense | laryngeal carcinoma, ovarian carcinoma |
|
nuclear respiratory factor 1 | 1/91 | 1% | Missense | Ovarian carcinoma |
|
ubiquitin-conjugating enzyme E2H | 1/217 | 0.5% | N/A | N/A |
|
zinc finger, C3HC-type containing 1 | 0/90 | 0% | N/A | N/A |
|
kelch domain containing 10 | 0/90 | 0% | N/A | N/A |
|
transmembrane protein 209 | 2/92 | 2.2% | Missense, | Ovarian carcinoma, colon carcinoma |
|
carboxypeptidase A2 (pancreatic) | 1/92 | 1% | Missense, | Mouth squamous carcinoma |
|
carboxypeptidase A4 | 3/517 | 0.6% | Missense, nonsense | laryngeal carcinoma, ovarian carcinoma |
|
carboxypeptidase A5 | 2/93 | 2.1% | Missense, nonsense | Ovarian carcinoma |
|
mesoderm specific transcript homolog | 2/516 | O.6% | Missense, nonsense | Ovarian carcinoma |
|
testis specific, 13 | 3/93 | 3.2% | Missense, | laryngeal carcinoma, ovarian carcinoma |
|
Kruppel-like factor 14 | 1/91 | 1% | Missense, | Mouth squamous carcinoma |
From the COSMIC database (
We first investigated whether the expression of the miRNA within the 7q32 MDR was impaired by 7q deletion by comparing their expression between SMZL with (n = 8) and without 7q deletion (n = 10) using qRT-PCR. In general, 1) there was a significant correlation between the expression of major and minor species of the miRNA (
The results show that the expression of these mature miRNAs is in general lower in SMZL with 7q deletion than those without the deletion, although not statistically significant with the exception of miR-29a. *minor miRNA species, FL: follicular lymphoma; MCL: mantle cell lymphoma; MALT-L: MALT lymphoma.
It is thus likely that 7q deletion underlies the reduced expression of these miRNA. Interestingly, the expression of these miRNA was also different among SMZL, FL, MCL and MALT lymphoma entities (
To further search for evidence that miRNA may be the target of 7q deletion in SMZL, we performed PCR and sequenced the miR-593, miR-129, miR-182, miR-96, miR-183, miR-335, miR-29a and miR-29b1 genes in 23 cases of SMZL. This initial screening revealed a SNP, rs76481776 (G to A substitution at nucleotide position 106), in the primary miR-182 in 5 cases of SMZL but showed no nucleotide alterations in any of the other primary miRNA investigated. In addition, two novel SNPs were observed in the 3¢ region distal to the pri-miR-182 hairpin. Subsequent screening showed the rs76481776 (G106A) variant in 12 (20%) of the 60 SMZL investigated. There was no association between the rs76481776 substitution and 7q deletion, or immunoglobulin heavy chain somatic mutation status. The rs76481776 substitution was also seen frequently in FL (20%), HCL (13%), MALT lymphoma (13%), PTCL-NOS (9%), less frequently in CLL (5%), DLBCL (4%) and MCL (3%), but not in lymphoplasmacytic lymphoma and 3 lymphoma cell lines (
To investigate whether any of the coding genes at 7q32 may be the target of 7q deletion, we compared the expression of the genes on chromosome 7 between SMZL with (n = 15) and without 7q deletion (n = 33) by analysis of the Affymetrix gene expression profile. In general, the genes on 7q, particularly those within the MDR, were under-expressed in cases with 7q deletion, in keeping with the relatively large and heterogeneous regions affected by the deletion (
Transcriptional repression by promoter methylation is one of the common mechanisms underlying the inactivation of tumour suppressor genes. To search for potential evidence of gene inactivation by promoter methylation, we correlated the levels of promoter methylation and expression of the genes within the MDR in SMZL. Among the 37 genes which were represented in both the methylation and Affymetrix expression arrays,
Among the coding genes within the MDR,
By extensive high-resolution array CGH analyses of a large cohort of SMZL, the present study mapped the MDR of 7q deletion in SMZL to a region of 2.8 Mb (127,287795-130,078068 Mb, GRCh37). There are several features associated with the deletion, which are worth pointing out in considering the candidate genes targeted by the deletion. First, there was no evidence of recurrent breakpoints or cryptic deletion, which might indicate a particular single gene as the target of the deletion. This finding raises the probability that multiple genes are targeted by the deletion. Second, there was no evidence of homozygous deletion at 7q32 in SMZL. Thus, it is unknown whether the gene targeted by 7q deletion is a classic tumour suppressor gene and required to be inactivated on both alleles, and if so, an alternative mechanism of inactivation must be responsible. Finally, the SMZL with 7q deletion do not constitute a distinct subgroup as shown by integrated clinicopathological and gene expression microarray analyses
As shown by qRT-PCR, there was a clear trend of under-expression of all the miRNA within the MDR in SMZL with 7q deletion as compared with those without the deletion. In addition, the expression of miR-29a, which was located distal to the MDR, was significantly reduced in SMZL with 7q deletion than those without the deletion. Under-expression of miR-29a in SMZL has also been reported in previous studies
The normal cell counterpart of SMZL is unknown, so it is not possible to directly compare miRNA expression between SMZL and the cell of its origin. Several studies have investigated the miRNome of normal B-cell subsets by expression microarrays and shown variable expression of miR-182/96/183 among different B-cell subsets, being expressed highly in the germinal centre B-cells, but moderately in the naïve and memory B-cells
Alteration in miRNA sequences is another mechanism underlying miRNA mediated oncogenesis
Among of the 37 coding genes within the MDR, 18 were significantly under-expressed in SMZL with the 7q deletion in comparison with those without the deletion. In 10 of these genes, their reduced expression appeared to be associated with epigenetic methylation as shown by integrated analyses of the gene expression and methylation microarray data, suggesting that promoter methylation may underpin their reduced expression. However, similar correlations were also seen in FL and MCL. It remains to be investigated whether these correlations are associated with lymphoma development or a common feature of their normal cell counterparts. In view of lack of clear evidence that pinpoints the genes likely targeted by the deletion, we further searched for somatic mutations in a favourite candidate gene
To help the identification of the target genes of 7q deletion in SMZL, we searched the somatic mutation data in the COSMIC (Catalogue of Somatic Mutations in Cancer) database (
In summary, our integrated analyses of 7q deletion by genomic, expression and methylation profiling showed that the expression of several genes within the MDR was significantly affected by the deletion. Our sequencing analyses also excluded the possibility of several potential candidate genes as the target of 7q deletion. The data presented in this study provide valuable guidance for further characterisation of 7q deletion.
(TIF)
(TIF)
(TIF)
(TIF)
(TIF)
(XLS)
(DOC)
(DOC)
(DOC)
(DOC)
The authors would like to thank Dr. Ian McFarlane, the Microarray CoreLab, National Institute of Health Research, Cambridge Comprehensive Biomedical Research Centre for his help in DNA sequencing.