Skip Navigation LinksHome > May 1, 2010 - Volume 54 - Issue 1 > Recurrent Chromosomal Alterations in Molecularly Classified...
JAIDS Journal of Acquired Immune Deficiency Syndromes:
doi: 10.1097/QAI.0b013e3181d3d9eb
Basic and Translational Science

Recurrent Chromosomal Alterations in Molecularly Classified AIDS-Related Lymphomas: An Integrated Analysis of DNA Copy Number and Gene Expression

Deffenbacher, Karen E PhD; Iqbal, Javeed PhD; Liu, Zhongfeng MD; Fu, Kai MD, PhD; Chan, Wing C MD

Free Access
Article Outline
Collapse Box

Author Information

From the Departments of Pathology and Microbiology, University of Nebraska Medical Center, Omaha, NE.

Received for publication October 6, 2009; accepted January 5, 2010.

This work was supported in part by National Cancer Institute grant (5U01/CA114778), a supplement from the AIDS Malignancy Branch (5U01/CA114778-02S1), NIH grant (U01/CA84967), and Eppley Core Grant. The University of Nebraska Medical Center Microarray Core Facility is supported partially by NIH grant P20 RR016469 from the INBRE Program of the National Center for Research Resources.

Correspondence to: Wing C. Chan, MD, Codirector, Center for Research in Lymphoma and Leukemia, Department of Pathology and Microbiology, 983135 Nebraska Medical Center, Omaha, NE 68198-3135 (e-mail: jchan@unmc.edu).

Collapse Box

Abstract

HIV-infected individuals have a significantly increased risk of developing an aggressive B-cell Non-Hodgkin Lymphoma relative to HIV(−) persons. Due to their aggressive nature, AIDS-related lymphomas (ARL) can also be more difficult to classify. Genetic abnormalities are known to play a significant role in HIV(−) lymphomagenesis. To aid in case classification and identify key pathogenetic events in ARL, we analyzed gene expression data and somatic DNA copy number changes by high-resolution array comparative genomic hybridization in tumors from 20 B-cell derived ARL (B-ARL) patients. Gene expression-based predictors robustly classified the B-ARL cases, distinguishing Burkitt lymphoma and diffuse large B-cell lymphoma, and identifying activated B-cell like and germinal center B-cell like molecular subtypes of diffuse large B-cell lymphoma. Array comparative genomic hybridization analysis revealed 13 recurrent losses and 16 recurrent gains in the B-ARL cases, including gain of 19p13.2 and loss of 16q23, not previously reported in B-ARL. The WWOX tumor suppressor gene was characterized as a candidate gene for the 16q23.1 locus and showed gene silencing or truncated transcript in 9 of 16 cases. These data demonstrate the ability to molecularly classify B-ARL lymphomas by gene expression and identified DNA copy number alterations targeted in B-ARL.

Back to Top | Article Outline

INTRODUCTION

Non-Hodgkin lymphoma (NHL) is the second most common AIDS-defining malignancy with an overall incidence of 4%-10% in AIDS patients and accounting for about 10% of all NHL in the United States.1 Before the introduction of highly active antiretroviral therapy (HAART), the risk of developing an NHL was reported to be between 60-fold and 200-fold higher in AIDS patients relative to the risk among non-HIV-infected individuals.2 Despite an overall reduction in the incidence of AIDS-related NHL (ARL) in response to HAART, the incidence of systemic NHL remains higher in the AIDS population.3-6 The impact of HAART on lymphoma incidence varies widely by subtype, with the greatest response in immunoblastic lymphoma (IBL) and primary central nervous system lymphomas (PCNSL) and only minimal response in Burkitt lymphoma (BL).3,4 The vast majority of ARL are clinically aggressive, monoclonal B-cell neoplasms with about 80% of cases presenting systemically and ∼20% as PCNSL.1,2,7

The histopathologic spectrum of aggressive B-cell ARL includes: BL, diffuse large B-cell lymphoma (DLBCL) with centroblastic (CB) features, and DLBCL with IBL features.2,8 BL and DLBCL share overlapping morphologic and immunophenotypic features, which can make the distinction between these 2 lymphomas problematic. Even the t(8;14) translocation involving MYC, a hallmark of BL,9 is found in a subset of DLBCL cases.10 Classification based on morphology may be further complicated in the ARL setting due to an increase in overlapping features and often extensive necrosis due to an aggressive growth rate.1,2,7,11 Gene expression profiling (GEP) has been used in HIV(−) lymphomas to derive molecular predictors that reliably distinguish between BL and DLBCL.12,13 BL and DLBCL differ significantly in clinical course and disease management, underscoring the importance of a robust classifier. GEP studies have also defined 3 molecular subtypes of DLBCL: germinal center B-cell like (GCB), activated B-cell like (ABC), and primary mediastinal B-cell lymphoma.14 Comparative genomic hybridization (CGH) data further support that these subtypes reflect biologically distinct disease entities.14,15 Given their significantly different survival rates after multiagent chemotherapy,14,16-18 molecular classification of DLBCL subtypes also has important clinical implications. In this study, we determine if previously derived GEP-based predictors are able to molecularly classify B-ARL cases, to provide a robust means of distinguishing AIDS-lymphoma subtypes. GEP data were also used to define gene signatures and pathways that differ significantly between AIDS-NHL and HIV-negative NHL.

The initiating event in the pathogenesis of BL involves translocation of MYC, whereas alteration of BCL6 through translocation and/or mutation is believed to be a common primary event in DLBCL. These initial alterations precondition the B-cell to acquire additional genetic abnormalities that promote cellular transformation. Characterization of secondary alterations that give rise to distinct lymphoma subtypes have been undertaken through cytogenetic and CGH studies of DLBCL15and BL19 in immune competent patients, however, there are relatively few studies examining genetic abnormalities in ARL. All prior studies of DNA copy number changes (CNC) in ARL utilized CGH, which provides low resolution on the order of ∼10 megabase.20

Array-based CGH (aCGH) platforms provide resolution on the order of tens of kilobases, significantly refining CNC detected by cytogenetics. In the present study, we investigated 20 B-ARL cases by high resolution aCGH to identify recurrent genomic abnormalities targeted in AIDS lymphoma. We also sought to identify potential candidate oncogenes and tumor suppressor genes in regions of gain or loss, respectively, by correlating gene expression data with gene copy number (CN).

Back to Top | Article Outline

METHODS

Patients and Samples

Frozen tumor biopsies were obtained from 24 HIV-positive patients diagnosed with NHL between 1995 and 2005. Four cases were ascertained at the University of Nebraska Medical Center (UNMC), and 20 specimens were obtained from the National Cancer Institute AIDS and Cancer Specimen Resource tumor bank.21 All lymphoma specimens were reviewed and classified by hematopathologists both at UNMC or centrally at the AIDS and Cancer Specimen Resource. Available hematoxylin-eosin (H&E) stained sections were further reviewed by Wing C. Chan, and frozen sections were examined to determine that representative tumor tissues were included in the experiemental studies. Cases were classified according to the World Health Organization classification scheme as follows: 16 DLBCL, 2 Hodgkin lymphoma, 1 follicular lymphoma (FL), 1 true histiocytic, and 4 B-cell high-grade lymphoma, not otherwise specified (NOS). Only cases with an AIDS-defining B-cell NHL were included for further study, excluding the histiocytic, FL, and 2 Hodgkin lymphoma cases. Patient data are presented in Table 1. This study was reviewed and approved by the institutional review board at UNMC.

Table 1
Table 1
Image Tools
Back to Top | Article Outline
DNA and RNA Extraction

The Allprep DNA/RNA Mini Kit (Qiagen; Valencia, CA) was used to isolate genomic DNA and total RNA simultaneously from the same frozen tissue sample for subsequent aCGH/polymerase chain reaction (PCR) and GEP analysis, respectively.

Back to Top | Article Outline
EBV Detection by Quantitative PCR

EBV status of the cases was determined by Q-PCR amplification of the EBV nuclear antigen 1 (EBNA1) gene from genomic DNA. Amplification was done using the DyNAmo SYBR Green Q-PCR Kit (Finnzymes; Woburn, MA) according to manufacturer protocol and products were detected on an Opticon 2 instrument (MJ Research; Waltham, MA). 100 ng of genomic DNA was PCR amplified using previously published primers and conditions specific for a portion of the EBV EBNA1 gene.22 By aCGH analysis, the RPL13 locus was determined to have a normal diploid CN in all cases. A Q-PCR assay targeting the RPL13 gene was used to normalize for the number of cells amplified per reaction. The relative amount of EBNA1 was determined by the ΔΔCT method using RPL13 as a reference, and Namalwa genomic DNA as a standard and positive control. Namalwa is a human BL cell line known to harbor 2 integrated copies of the EBV genome/cell. Cases with relative expression values above the Namalwa reference were considered EBV positive.

Back to Top | Article Outline
WWOX Reverse Transcriptase-Polymerase Chain Reaction

cDNA was generated from total RNA isolated from B-ARL tumor specimens using the Superscript III First-Strand Synthesis System (Invitrogen, Carlsbad, CA) according to manufacturer's protocol. PCR amplification of the WWOX transcript from cDNA was done using previously published primers and conditions.23 Amplified products were examined by 2% agarose gel electrophoresis.

Back to Top | Article Outline
Gene Expression Profiling

Gene expression profiles were generated using the Human Genome U133 Plus 2.0 Array (Affymetrix, Santa Clara, CA). Sufficient amounts of RNA, at least 2 μg, were obtained from 17 of the 20 B-ARL specimens. RNA samples were prepared and hybridized to the arrays according to the manufacturer's protocol. The raw gene expression data have been deposited in the Gene Expression Omnibus (GEO) database24 and are available under accession number (GSE17189). Expression data were analyzed using BRB-ArrayTools version 3.7 software (http://linus.nci.nih.gov/BRB-ArrayTools.html).25 Array data were normalized by median centering of the log2 signal intensities.

Molecular classification of the cases was done using the class prediction tool in BRB Array Tools by the Bayesian compound covariate predictor method.18 A classifier distinguishing BL from DLBCL was previously generated using GEO dataset GSE4732.12 The subset of cases profiled on the U133 Plus 2.0 array were used as a training dataset to predict B-ARL class. We chose a criterion of ≥90% probability as the cutoff to classify cases. Cases classified as DLBCL were then subclassified using a 185 probe set multivariate predictor that distinguishes ABC and GCB subtypes of DLBCL.16 Gene set enrichment analysis (GSEA) was used to identify pathways and gene signatures that differed significantly between HIV-positive and HIV-negative lymphoma subgroups. The Curated gene sets collection in the Broad Institute's Molecular Signature Database (MsigDB) were used in GSEA analyses.

Back to Top | Article Outline
Array CGH

Genomic DNA isolated from 20 B-ARL tumor specimens were prepared according to the GeneChip Mapping 500K Assay protocol (Affymetrix Inc., Santa Clara, CA). Samples were hybridized to the 250K Nsp I array. Single nucleotide polymorphism (SNP) genotypes and CN data were generated using Genotyping Console 2.1 software from Affymetrix. CN status was determined by the log2 ratio of sample:global reference signal. The global reference was generated from 48 HapMap samples which were previously genotyped on the Nsp I array. The segment reporting tool was used to summarize CN data, in which a segment was defined as containing a minimum of 5 contiguous SNPs and spanning at least 100 kilobase. Regions of copy neutral loss of heterozygosity (LOH) [acquired uniparental disomy (aUPD] were also identified.

CN alterations from all cases were aligned to the genome to identify the minimal common region (MCR) for each recurrent abnormality. Abnormalities occurring in 3 or more cases (≥15%) were presumed to be nonrandom changes and were selected for further analysis. MCR physical coordinates were checked against the physical positions of copy number variants and segmental duplications deposited in the University of California Santa Cruz genome browser and the Database of Genomic Variants (http://projects.tcag.ca/variation/). MCR that either did not encompass any genes, or that showed 100% overlap with an annotated copy number variants of similar CN state were excluded from further analyses. All RefSeq genes within the physical boundaries of an MCR were compiled. Using the class comparison tool in BRB-ArrayTools, gene expression data for these genes were compared between cases exhibiting the copy change aberration and those with normal diploid CN. Criteria used to define differential gene expression between MCR+ and MCR− groups included: a P < 0.005 by the KS/LS random variance permutation test, which tests for a significant association of an MCR gene list with an abnormality; and a P < 0.05 under the random variance model univariate test, which determines significant group differences in the expression of individual genes.

Back to Top | Article Outline

RESULTS

Patient Characteristics

Table 1 summarizes case statistics. Included in the study were 16 pathologically defined cases of DLBCL, 6 of which showed prominent plasmacytic (PC) features by H&E stain. Four cases that could not be definitively diagnosed as DLBCL morphologically were also included as high-grade lymphoma NOS. All but 1 case were male, and the median age at diagnosis was 35.5 years with a range of 24-57 years. Typical DLBCL has a median age of onset of 60 years and a gender ratio of 1:1. As indicated in Table 1 and Figure 1, 11 of the 20 cases (55%) were EBV positive by our criteria. This included 1 of 4 molecular Burkitt lymphoma (mBL) cases and 10 of 16 DLBCL. All 6 DLBCL cases with PC features were EBV positive.

Figure 1
Figure 1
Image Tools
Back to Top | Article Outline
Molecular Classification of Cases by GEP

GEP was performed on 17 of the 20 cases which had sufficient RNA. We previously derived a GEP-based molecular classifier, which robustly distinguishes BL from DLBCL.12 Using the Bayesian prediction method in BRB Array Tools, the classifier identified 4 of the 17 B-ARL cases as mBL (P = 1), with the remaining 13 B-ARL cases classified as molecular DLBCL (mDLBCL) (Fig. 2A). A GEP-based molecular classifier, which distinguishes ABC and GCB DLBCL subtypes, was applied to the 13 B-ARL cases molecularly classified as DLBCL.18 The Bayesian compound covariate predictor identified 6 cases as ABC subtype and 3 GCB cases with probabilities greater than 90% (Fig. 2B). Four of the 13 mDLBCL cases were not classifiable (NC) by the subtype predictor. Five of 6 cases classified as ABC molecularly demonstrated significant PC features morphologically.

Figure 2
Figure 2
Image Tools

GSEA analysis was performed, comparing the 13 HIV+ mDLBCL cases with 60 HIV(−) DLBCL cases used in the ABC/GCB classification. Pathways that differed in B-ARL relative to the immune competent counterpart included enrichment for MYC targets, FAS pathway, mTOR downregulated genes, cell cycle genes, and PC and primary effusion lymphoma gene signatures. In the HIV(−) DLBCL, there was enrichment mostly for B-cell receptor (BCR) and T-cell receptor signaling pathways. Comparing HIV+ and HIV(−) ABC, GSEA analysis indicated enrichment for MYC targets, ARF pathway, and cell cycle components in HIV+ relative to HIV- ABC. Stratification of cases by EBV status indicated enrichment of IL12 and IL2RB pathways, Ras and p53 pathways, IL4R signaling, IL5 downregulated genes, and mTOR pathway in EBV+ cases.

Back to Top | Article Outline
aCGH Detection of CNCs and Association With GEP

aCGH analysis identified CNC in all 20 cases examined (Fig. 3). A number of MCR lacked genes within the interval and/or completely overlapped known CN polymorphisms and were excluded from further study. There were 13 recurrent losses (Table 2) and 16 recurrent gains (Table 3) with an average of 7.6 CNC per case and a range of 2-17 changes. In agreement with previous reports, gains were much more frequent than losses with an average of 4.8 gains and 2.8 losses per case. The total number of CNC was significantly elevated in the ABC subtype of DLBCL (mean = 12.3) relative to GCB DLBCL (mean = 2.3; P = 0.0016) and BL (mean = 7; P = 0.036) as assessed by 1-sided t test. There was no significant difference in number of CNC relative to EBV status, with EBV-positive cases averaging 8.8 CNC and an average of 6 CNC for EBV-negative cases (P = 0.096).

Table 2
Table 2
Image Tools
Table 3
Table 3
Image Tools
Figure 3
Figure 3
Image Tools

Most of the CNCs were previously reported in HIV-negative lymphomas, except gain of 19p13.2 and loss of 16q23.1. Although case numbers were too few to correlate specific abnormalities with lymphoma subtype, gain of 3q21.3 was solely found in BL cases, and +17q21.33, −3p14.2, and −16q23.1 were only found in ABC cases. We previously reported associations between DLBCL molecular subtypes and specific CN alterations.15 Corroborating these earlier findings, we detected 3 CNC enriched in the GCB subtype, including: losses of 1p36 and 13q34 and gain of 13q31.3. Loss of 9p21.3 is associated with the ABC subtype and was also found in the B-ARL cases. High-level amplification was found in 1 case each within the gain MCR for 4q, 12q, and 13q. Conversely, copy neutral LOH (aUPD) was found in one case each overlapping the 13q and 17p loss MCRs. Recurrent aUPD was found in 3 cases on chromosome 6p with an MCR spanning from the telomere to 35 megabase.

For each MCR, DNA CN status was correlated with gene expression for genes residing within the MCR. Candidate genes whose expression significantly correlated with DNA CN are indicated in Tables 2 and 3. These included previously reported candidate genes, such as: MIRH1, TP73, and MTAP. Other candidates demonstrating correlation between gene dosage and gene expression included: RASAL2, RCE1, SELPLG, HTATSF1, and NUP88. A number of genes within these MCR that did not show a clear correlation between CN and gene expression, however are strong candidates based on putative protein function, including: MAD1L1, CARD11, HGF, EIF3G, DNMT1, EDG5, FHIT, WWOX, and CDKN2A. The WWOX transcript was examined by reverse transcriptase-polymerase chain reaction and 1 of the 4 cases with 16q deletion expressed an aberrant, truncated transcript (Fig. 4). All 4 cases with 16q deletion were of the ABC subtype.

Figure 4
Figure 4
Image Tools
Back to Top | Article Outline

DISCUSSION

In this study, we have assembled 20 B-ARL cases which meet criteria for an AIDS-defining malignancy. All cases included in this study were systemic lymphomas with only about one-third involving the lymph node. This is in accordance with the known predilection of ARL for extranodal sites.7,26 The vast majority of systemic ARL fall into 3 histologic types of aggressive B-cell lymphomas, including BL and DLBCL with either CB or IBL features.1,7,26 DLBCL comprises about 70% and BL about 30% of systemic B-ARL cases,7 although these numbers are changing in the post-HAART era. Only H&E slides were available for pathology review, which identified 16 DLBCL specimens, 6 of which showed significant PC features, and 4 high-grade cases where a definitive morphologic diagnosis could not be made.

Diagnostic distinction of BL and DLBCL by histopathology could be difficult due to overlapping morphologic and immunophenotypic features. Recognizing this continuum of features, the World Health Organization 2008 Classification established a new diagnostic category of B-cell lymphoma unclassifiable, with features intermediate between diffuse large B-cell lymphoma and Burkitt lymphoma (DLBCL/BL). These cases harbor morphologic immunophenotypic and genetic features from both BL and DLBCL, thus excluding them from either category. Because optimal management of these 2 entities differs significantly, a GEP-based classifier which robustly distinguishes between BL and DLBCL molecularly is highly desirable.12 AIDS-related NHL frequently show increased mitotic figures, increased cellular debris, and extensive necrosis, indicating a high proliferative rate relative to corresponding HIV(−) NHLs.1,11 They also frequently demonstrate intermediate features, further compounding the difficulty in distinguishing AIDS-related BL and DLBCL by histopathology. In the absence of cytogenetic and immunophenotypic data, we cannot determine whether the 4 unclassifiable cases in this series belong to the new DLBCL/BL diagnostic category or whether the limited sample availability precluded diagnosis based on morphology alone. Given the increased difficulty in discerning these 2 entities in the AIDS setting, it would be interesting to examine whether there is an increased proportion of DLBCL/BL cases in ARL relative to HIV(−) lymphoma in a series of cases with available clinical data. GEP data could be useful in determining whether DLBCL/BL represents a true continuum with gene expression features of both entities or whether a unique gene signature can be derived from such cases.

We applied the Bayesian compound covariate predictor derived from HIV(−) BL and DLBCL cases12 to the B-ARL cases to determine whether they could be similarly classified molecularly. Of the 17 cases with GEP data, all were robustly assigned to either mBL or mDLBCL subtypes with a probability greater than 99% (Fig. 2). Four cases were defined as mBL. These 4 cases showed significantly increased expression of MYC target genes and decreased expression of MHC Class I molecules and NF-κB targets relative to the mDLBCL cases. Of these 4 cases, 1 was high-grade NOS by morphology; however, 3 were classified as DLBCL. Review was only by H&E, often with limited tissue or highly necrotic sections, which could possibly account for the discrepancy between the morphologic and molecular diagnoses. Yet, none of the cases have typical BL by histopathology underscoring the value of molecular diagnosis in identifying BL in B-ARL. The mBL cases also had an average of 7 CNC per case, consistent with the average number of 6.9 CNC found previously in mBL cases with discrepant morphology.19 In the post-HAART era, the incidence of AIDS-related DLBCL is declining, with IBL cases demonstrating the greatest response and BL the least.5,6 Definitive classification of B-ARL will allow further study of a larger homogeneous cohort to ascertain the mechanisms by which HIV augments BL incidence that paradoxically continues to be elevated under relatively mild immune suppression.

In AIDS-BL, compared with HIV(−) BL, GSEA indicated a significant decrease in p53 and its targets, MAPK and BCR signaling, numerous DNA repair genes, and altered expression of mTOR targets. Loss of p53 function is typical of BL27 but was more profound in the HIV+ cases. Differential expression of mTOR targets may reflect differences in the AKT pathway. It was recently shown that inhibition of BCR signaling is not detrimental to cell survival or proliferation in B-ARL, whereas HIV(−) DLBCL cells are highly sensitive to such inhibition by dasatinib.28 Lower BCR signaling in AIDS-BL also suggests that it is less dependant on BCR signaling. Loss of a significant number of DNA repair genes could lead to greater genomic instability, especially when SHM is active, and may explain the higher number of genomic abnormalities in AIDS-BL compared with BL in the non-AIDS setting. mBL with discrepant morphology has also been shown to harbor a significantly higher number of genomic abnormalities on average relative to classic BL,19 and it is possible that DNA repair pathways are also targeted in these intermediate cases.

Cases classified as mDLBCL were further classified into ABC and GCB molecular subtypes. Of these 13 cases, 3 were GCB, 6 were ABC, and 4 were not classifiable. All but 1 of the ABC-subtype cases demonstrated significant PC features by histology, but the reverse was not true, indicating the importance of molecular studies in classification. GSEA analysis was performed, comparing (1) HIV+ GCB with HIV(−) GCB cases and (2) HIV+ ABC with HIV(−) ABC cases. Both comparisons indicated diminished BCR signaling in HIV+ cases relative to HIV(−) cases, again suggesting that BCR signaling may be dispensable in B-ARL. Similar to the AIDS-BL comparison, the HIV+ ABC and GCB cases showed low MAPK signaling. The AIDS-ABC cases showed enrichment in MYC targets, ARF pathway, and cell cycle components relative to HIV(−) ABC cases, which may be molecular correlates of the higher proliferative rate in B-ARL. A subset of DLBCL cases harbor the t(8;14) translocation of MYC,10 and MYC deregulation has been postulated as an event leading to the transformation of a FL to a DLBCL.29 Enrichment for MYC targets in HIV+ ABC cases suggests that MYC deregulation may play a role in B-ARL as well.

The role of EBV in B-cell lymphomagenesis is well characterized, particularly within the context of immunocompromised patients.30 Both BL and GCB DLBCL can develop in the setting of mild immune deficiency with relatively high CD4+ T-cell counts. Only 30% of AIDS-related BL and 40% of AIDS-related GCB DLBCL are EBV+ with EBV+ cases failing to express viral antigens,31 which may be important in the presence of a relatively intact immune system. The pathogenesis of EBV(−) cases may be related to massive GCB cell proliferation upon chronic antigenic stimulation leading to the stochastic accumulation of genetic lesions. In contrast, IBL DLBCL develops under marked immune suppression with low CD4+ T-cell counts. About 90% of IBL cases are EBV+, and viral antigens are expressed,31 suggesting that EBV plays a crucial role in its pathogenesis. Ten of our 16 DLBCL cases were EBV+, and notably, all 6 cases demonstrating significant PC features were EBV+. Given the high rate of EBV positivity and the post-germinal center ABC-subtype derivation in all but 1 case with PC features, it is likely that these cases comprise IBL however, we cannot conclude this based on our morphologic data alone. Only 1 of 4 of the molecular BL cases was EBV+, consistent with the 30% incidence previously reported.

Using high-density aCGH, we also sought to determine the common genetic abnormalities involved in the development of a B-ARL. Prior CGH studies of ARL are relatively few and utilized low-resolution techniques. Genomic studies of DNA CNC in ARL include 1 case of BL,32 14 cases of DLBCL,33 8 cases of primary effusion lymphoma,34 and 1 study including 12 BL, 10 DLBCL, and 4 PCNSL.35 Previous studies identified gains on 2p, 4q, 7, 8p, 12q24, 16p, 17q, and X that were also found in our B-ARL cases and were refined in this study to ∼200 kilobase for most MCR. Similarly, losses of 1p, 9p, 17p, and 19q were previously reported for ARL and were significantly refined by our cases, providing potential target genes for these regions. Overall, aCGH analysis identified 29 recurrent CNC, including 13 losses and 16 gains. Interestingly, the gain of 3q21.3 was only found in AIDS-BL cases. This region has been implicated in other lymphoma types but not for BL. The MCR on 3q21.3 includes the RAB43, ISY1, and CNBP genes. CNBP is a STAT6 target gene that enhances lymphoid cell proliferation and survival and increases MYC activity through transcriptional regulation,36,37 making this gene a strong functional target for this locus and AIDS-BL.

We also compared recurrent abnormalities in B-ARL to those found in the corresponding HIV(−) lymphoma subtype.15,19 Gain of 13q, which is enriched in the GCB-subtype of DLBCL in HIV(−) patients, was also found in 2 DLBCL cases and amplified in 1 BL case from our B-ARL dataset. The putative target for this locus is the MIRH1 gene, harboring the miRNA 17-92. Losses of 1p36 and 13q34, which are enriched in the GCB-subtype, and loss of 9p21.3, which is enriched in the ABC-subtype, were also found in our B-ARL cases.

Of the 29 CNC, 2 regions, gain of 19p13.2 and loss of 16q23.1, have not been previously reported for ARL. Gain of 19p13 occurred in 25% of cases, including 1 BL and 4 DLBCL. Candidate genes in this interval include: ANGPTL6, EIF3G, DNMT1, and EDG5. Loss of 16q23.1 occurred in four DLBCL cases, all of ABC-subtype, with the MCR harboring only the single WWOX gene. Three of the 4 cases with WWOX loss demonstrated concurrent loss of FHIT on 3p14.2, with FHIT also the sole gene in this MCR. Both FHIT and WWOX span fragile sites and encode proteins with tumor suppressor activity. FHIT and WWOX expression is positively associated and coordinate loss of these genes are common in several types of cancers.38 FHIT has been shown to be deleted in 30% of DLBCL39 and silenced either through deletion or hypermethylation in 30% of BL.40 Decreased FHIT expression associates with worse prognosis and overall survival in DLBCL.41WWOX is proapoptotic, involved in the DNA damage response checkpoint,42 and regulates TP73 activity,43 a gene that is frequently lost on 1p36 in the B-ARL cases and in numerous other lymphoma subtypes. Gene expression of TP73 correlated with deletion of 1p36 in the B-ARL cases. Similar in function to WWOX, TP73 is also proapoptotic in response to DNA damage. WWOX has not been previously characterized as a target in lymphoma, although recently, mice hypomorphic for this gene showed an increased propensity to develop spontaneous B-cell lymphomas.44 By reverse transcriptase-polymerase chain reaction, we detected expression of an aberrant truncated transcript in 1 case harboring a 16q deletion. Expression of aberrant WWOX transcripts has previously been reported in a number of solid tumors.45

Most MCRs harbored few genes in the interval. DNA CN status was compared with gene expression for genes residing in the MCR intervals to determine functional status of these genes. Potential candidate genes showing a correlated change in gene expression with DNA CN are shown in bold in Tables 2 and 3. Both RASAL2 and RCE1 are involved in the regulation and activation of Ras. Gains of 1q have been reported in both BL and AIDS-BL,32,46 an entity in which alterations in RAS activity and RAS gene mutations have been reported. As RASAL2 is the only gene within the 1q25.2 MCR and gene expression correlates with CN, it is a strong candidate gene for this region. EBV LMP2A can promote B-cell survival through constitutive activation of the RAS/PI3K/AKT pathway,30 suggesting a more prominent role for RAS pathway genes in the AIDS setting where EBV likely contributes to lymphomagenesis. Deletion and methylation of 9p21 occur in a variety of tumors. This region harbors the CDKN2A gene, encoding p16INKA and p14ARF, and MTAP. CDKN2A is thought to be the target of this locus, as p16 and p14 both have well-characterized tumor suppressor activities. However, codeletion of CDKN2A and MTAP frequently occurs and has been associated with histologic transformation of NHLs.47MTAP gene deletion and lack of protein expression was also associated with poor prognosis in Mantle Cell Lymphoma.48 In the B-ARL cases, MTAP gene expression correlated with DNA CN, whereas CDKN2A expression did not. It is possible that gene expression measurements may not correlate well with gene dosage if the gene is also expressed by nontumor tissues. This may be particularly true for deletions.

For this reason, other potential candidate genes based on putative function may not demonstrate a correlation of CN with expression. These include gains of HGF on 7q21 and TAC4 on 17q. HGF/Met signaling controls B-cell survival and proliferation and is thought to contribute to lymphomagenesis in DLBCL. HGF positively correlates with tumor load and aggressiveness and with worse prognosis in DLBCL.49TAC4 or hemokinin-1 is a B-cell survival factor and is transcriptionally regulated by NFκB.50

In conclusion, we have shown that GEP data can be used to robustly classify B-ARL cases, where morphologic classification can be difficult. BL and DLBCL have different prognoses and treatment strategies, underscoring the importance of molecular classification. Similarly, ABC and GCB subtypes of DLBCL have distinct pathogenetic mechanisms and disease course. A number of unique differences have been identified in B-ARL relative to tumors from nonimmunocompromised patients that could be validated with a larger sample size. Although most CNA in this study were previously reported and not specific to ARL, we identified 2 novel loci, a gain of 19p13.2 and loss of 16q23.1, not previously described in ARL. The high-resolution platform used in this study resolved most of the gains and losses to a few hundred kilobase with limited number of genes residing in the intervals. Identification of the underlying candidate genes in these intervals is further facilitated by the simultaneous availability of GEP data. Further studies on a larger number of cases will lead to a more comprehensive characterization of the pathways and mechanisms of lymphomagenesis in the different subtypes of B-ARL.

Back to Top | Article Outline

ACKNOWLEDGMENTS

The authors thank Martin Bast for the clinical data coordination, Lisa Bough for technical assistance, and the AIDS and Cancer Specimen resource for providing samples and clinical data.

Back to Top | Article Outline

REFERENCES

1. Krause J. AIDS-related non-Hodgkin's lymphomas. Microsc Res Tech. 2005;68:168-175.

2. Carbone A, Gloghini A. AIDS-related lymphomas: from pathogenesis to pathology. Br J Haematol. 2005;130:662-670.

3. Bonnet F, Chene G. Evolving epidemiology of malignancies in HIV. Curr Opin Oncol. 2008;20:534-540.

4. Silverberg MJ, Abrams DI. AIDS-defining and non-AIDS-defining malignancies: cancer occurrence in the antiretroviral therapy era. Curr Opin Oncol. 2007;19:446-451.

5. Barbaro G, Barbarini G. HIV infection and cancer in the era of highly active antiretroviral therapy (Review). Oncol Rep. 2007;17:1121-1126.

6. Biggar RJ, Chaturvedi AK, Goedert JJ, et al. AIDS-related cancer and severity of immunosuppression in persons with AIDS. J Natl Cancer Inst. 2007;99:962-972.

7. Gaidano G, Carbone A, Dalla-Favera R. Pathogenesis of AIDS-related lymphomas: molecular and histogenetic heterogeneity. Am J Pathol. 1998;152:623-630.

8. Carbone A, Gloghini A, Volpe R, et al. The pathology spectrum of AIDS-related non-Hodgkin's lymphoma [in Italian]. Pathologica. 1998;90:763-770.

9. Neri A, Barriga F, Knowles DM, et al. Different regions of the immunoglobulin heavy-chain locus are involved in chromosomal translocations in distinct pathogenetic forms of Burkitt lymphoma. Proc Natl Acad Sci U S A. 1988;85:2748-2752.

10. Kramer MH, Hermans J, Wijburg E, et al. Clinical relevance of BCL2, BCL6, and MYC rearrangements in diffuse large B-cell lymphoma. Blood. 1998;92:3152-3162.

11. Knowles DM. Immunodeficiency-associated lymphoproliferative disorders. Mod Pathol. 1999;12:200-217.

12. Dave SS, Fu K, Wright GW, et al. Molecular diagnosis of Burkitt's lymphoma. N Engl J Med. 2006;354:2431-2442.

13. Hummel M, Bentink S, Berger H, et al. A biologic definition of Burkitt's lymphoma from transcriptional and genomic profiling. N Engl J Med. 2006;354:2419-2430.

14. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403:503-511.

15. Lenz G, Wright GW, Emre NC, et al. Molecular subtypes of diffuse large B-cell lymphoma arise by distinct genetic pathways. Proc Natl Acad Sci U S A. 2008;105:13520-13525.

16. Lenz G, Wright G, Dave SS, et al. Stromal gene signatures in large-B-cell lymphomas. N Engl J Med. 2008;359:2313-2323.

17. Rosenwald A, Wright G, Chan WC, et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. N Engl J Med. 2002;346:1937-1947.

18. Wright G, Tan B, Rosenwald A, et al. A gene expression-based method to diagnose clinically distinct subgroups of diffuse large B cell lymphoma. Proc Natl Acad Sci U S A. 2003;100:9991-9996.

19. Salaverria I, Zettl A, Bea S, et al. Chromosomal alterations detected by comparative genomic hybridization in subgroups of gene expression-defined Burkitt's lymphoma. Haematologica. 2008;93:1327-1334.

20. Bentz M, Plesch A, Stilgenbauer S, et al. Minimal sizes of deletions detected by comparative genomic hybridization. Genes Chromosomes Cancer. 1998;21:172-175.

21. Ayers LW, Silver S, McGrath MS, et al. The AIDS and cancer specimen resource: role in HIV/AIDS scientific discovery. Infect Agent Cancer. 2007;2:7.

22. Ryan JL, Fan H, Glaser SL, et al. Epstein-Barr virus quantitation by real-time PCR targeting multiple gene segments: a novel approach to screen for the virus in paraffin-embedded tissue and plasma. J Mol Diagn. 2004;6:378-385.

23. Fabbri M, Iliopoulos D, Trapasso F, et al. WWOX gene restoration prevents lung cancer growth in vitro and in vivo. Proc Natl Acad Sci U S A. 2005;102:15611-15616.

24. Edgar R, Domrachev M, Lash AE. Gene expression omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 2002;30:207-210.

25. Simon R, Lam A, Li MC, et al. Analysis of gene expression data using BRB-array tools. Cancer Inform. 2007;3:11-17.

26. Carbone A. AIDS-related non-Hodgkin's lymphomas: from pathology and molecular pathogenesis to treatment. Hum Pathol. 2002;33:392-404.

27. Gaidano G, Ballerini P, Gong JZ, et al. p53 mutations in human lymphoid malignancies: association with Burkitt lymphoma and chronic lymphocytic leukemia. Proc Natl Acad Sci U S A. 1991;88:5413-5417.

28. Lu P, Yang C, Guasparri I, et al. Early events of B-cell receptor signaling are not essential for the proliferation and viability of AIDS-related lymphoma. Leukemia. 2009;23:807-810.

29. Davies AJ, Rosenwald A, Wright G, et al. Transformation of follicular lymphoma to diffuse large B-cell lymphoma proceeds by distinct oncogenic mechanisms. Br J Haematol. 2007;136:286-293.

30. Pietersma F, Piriou E, van Baarle D. Immune surveillance of EBV-infected B cells and the development of non-Hodgkin lymphomas in immunocompromised patients. Leuk Lymphoma. 2008;49:1028-1041.

31. Ballerini P, Gaidano G, Gong J, et al. Molecular pathogenesis of HIV-associated lymphomas. AIDS Res Hum Retroviruses. 1992;8:731-735.

32. Zunino A, Viaggi S, Ottaggio L, et al. Chromosomal aberrations evaluated by CGH, FISH and GTG-banding in a case of AIDS-related Burkitt's lymphoma. Haematologica. 2000;85:250-255.

33. Tiirikainen MI, Mullaney BP, Holly EA, et al. DNA copy number alterations in HIV-positive and HIV-negative patients with diffuse large-cell lymphomas. J Acquir Immune Defic Syndr. 2001;27:272-276.

34. Mullaney BP, Ng VL, Herndier BG, et al. Comparative genomic analyses of primary effusion lymphoma. Arch Pathol Lab Med. 2000;124:824-826.

35. Vaghefi P, Martin A, Prevot S, et al. Genomic imbalances in AIDS-related lymphomas: relation with tumoral Epstein-Barr virus status. AIDS. 2006;20:2285-2291.

36. Michelotti EF, Tomonaga T, Krutzsch H, et al. Cellular nucleic acid binding protein regulates the CT element of the human c-myc protooncogene. J Biol Chem. 1995;270:9494-9499.

37. Tuomela S, Rautajoki KJ, Moulder R, et al. Identification of novel Stat6 regulated proteins in IL-4-treated mouse lymphocytes. Proteomics. 2009;9:1087-1098.

38. Sbrana I, Veroni F, Nieri M, et al. Chromosomal fragile sites FRA3B and FRA16D show correlated expression and association with failure of apoptosis in lymphocytes from patients with thyroid cancer. Genes Chromosomes Cancer. 2006;45:429-436.

39. Kameoka Y, Tagawa H, Tsuzuki S, et al. Contig array CGH at 3p14.2 points to the FRA3B/FHIT common fragile region as the target gene in diffuse large B-cell lymphoma. Oncogene. 2004;23:9148-9154.

40. Hussain A, Gutierrez MI, Timson G, et al. Frequent silencing of fragile histidine triad gene (FHIT) in Burkitt's lymphoma is associated with aberrant hypermethylation. Genes Chromosomes Cancer. 2004;41:321-329.

41. Chen PM, Yang MH, Hsiao LT, et al. Decreased FHIT protein expression correlates with a worse prognosis in patients with diffuse large B-cell lymphoma. Oncol Rep. 2004;11:349-356.

42. Yang J, Zhang W. WWOX tumor suppressor gene. Histol Histopathol. 2008;23:877-882.

43. Aqeilan RI, Pekarsky Y, Herrero JJ, et al. Functional association between Wwox tumor suppressor protein and p73, a p53 homolog. Proc Natl Acad Sci U S A. 2004;101:4401-4406.

44. Ludes-Meyers JH, Kil H, Nunez MI, et al. WWOX hypomorphic mice display a higher incidence of B-cell lymphomas and develop testicular atrophy. Genes Chromosomes Cancer. 2007;46:1129-1136.

45. Pimenta FJ, Cordeiro GT, Pimenta LG, et al. Molecular alterations in the tumor suppressor gene WWOX in oral leukoplakias. Oral Oncol. 2008;44:753-758.

46. Garcia JL, Hernandez JM, Gutierrez NC, et al. Abnormalities on 1q and 7q are associated with poor outcome in sporadic Burkitt's lymphoma. A cytogenetic and comparative genomic hybridization study. Leukemia. 2003;17:2016-2024.

47. Dreyling MH, Roulston D, Bohlander SK, et al. Codeletion of CDKN2 and MTAP genes in a subset of non-Hodgkin's lymphoma may be associated with histologic transformation from low-grade to diffuse large-cell lymphoma. Genes Chromosomes Cancer. 1998;22:72-78.

48. Marce S, Balague O, Colomo L, et al. Lack of methylthioadenosine phosphorylase expression in mantle cell lymphoma is associated with shorter survival: implications for a potential targeted therapy. Clin Cancer Res. 2006;12:3754-3761.

49. Kawano R, Ohshima K, Karube K, et al. Prognostic significance of hepatocyte growth factor and c-MET expression in patients with diffuse large B-cell lymphoma. Br J Haematol. 2004;127:305-307.

50. Tran AH, Berger A, Wu GE, et al. Regulatory mechanisms in the differential expression of Hemokinin-1. Neuropeptides. 2009;43:1-12.

Cited By:

This article has been cited 3 time(s).

Archives of Pathology & Laboratory Medicine
Lymphoid Proliferations Associated With Human Immunodeficiency Virus Infection
Chadburn, A; Abdul-Nabi, AM; Teruya, BS; Lo, AA
Archives of Pathology & Laboratory Medicine, 137(3): 360-370.
10.5858/arpa.2012-0095-RA
CrossRef
Retrovirology
B cell lymphoma in hiv transgenic mice
Curreli, S; Krishnan, S; Reitz, M; Lunardi-Iskandar, Y; Lafferty, MK; Garzino-Demo, A; Zella, D; Gallo, RC; Bryant, J
Retrovirology, 10(): -.
ARTN 92
CrossRef
Seminars in Diagnostic Pathology
Immunodeficiency-associated lymphoid proliferations (ALPS, HIV, and KSHV/HHV8)
Chadburn, A
Seminars in Diagnostic Pathology, 30(2): 113-129.
10.1053/j.semdp.2012.08.006
CrossRef
Back to Top | Article Outline
Keywords:

AIDS-related lymphoma; array-based comparative genomic hybridization; Burkitt lymphoma; Diffuse Large B-Cell lymphoma; gene expression profiling

© 2010 Lippincott Williams & Wilkins, Inc.

Login

Search for Similar Articles
You may search for similar articles that contain these same keywords or you may modify the keyword list to augment your search.