PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Cancer Cell. Author manuscript; available in PMC 2013 September 20.
Published in final edited form as:
PMCID: PMC3778921
NIHMSID: NIHMS398769

Integrative Analysis Reveals an Outcome-associated and Targetable Pattern of p53 and Cell Cycle Deregulation in Diffuse Large B-cell Lymphoma

Summary

Diffuse large B-cell lymphoma (DLBCL) is a clinically and biologically heterogeneous disease with a high proliferation rate. By integrating copy number data with transcriptional profiles and performing pathway analysis in primary DLBCLs, we identified a comprehensive set of copy number alterations (CNAs) that decreased p53 activity and perturbed cell cycle regulation. Primary tumors either had multiple complementary alterations of p53 and cell cycle components or largely lacked these lesions. DLBCLs with p53 and cell cycle pathway CNAs had decreased abundance of p53 target transcripts and increased expression of E2F target genes and the Ki67 proliferation marker. CNAs of the CDKN2A-TP53-RB-E2F axis provide a structural basis for increased proliferation in DLBCL, predict outcome with current therapy and suggest targeted treatment approaches.

Introduction

Diffuse large B-cell lymphoma (DLBCL) is the most common non-Hodgkin lymphoma in adults and a clinically and genetically heterogeneous disorder. With current immuno-chemotherapy, over 60% of patients with DLBCL can be cured; however, the remaining patients succumb to their disease (Friedberg, 2008). Despite recent advances in the molecular understanding of DLBCL pathogenesis, clinical risk factor models are still used to identify patients who are unlikely to be cured with current therapy. The most widely used model is the International Prognostic Index (IPI), an outcome predictor based on easily measurable clinical parameters including age, performance status, serum LDH, Ann Arbor stage and numbers of extranodal disease sites (Shipp et al., 1993). Although the IPI is robust and reproducible, the link between the included clinical parameters and underlying biology or targeted treatment remains to be defined.

In previous studies, increased cellular proliferation has also been associated with unfavorable outcome in DLBCL. Indirect indices of cellular proliferation included elevated serum LDH as a component of the IPI and increased expression of the Ki67 nuclear antigen (Grogan et al., 1988; Salles et al., 2011).

DLBCLs largely originate from germinal center (GC) B cells which have high growth rates and increased genomic instability (Klein and Dalla-Favera, 2008). GC B cells undergo somatic hypermutation (SHM) of their immunoglobulin variable region genes and class-switch recombination (CSR) to alter their immunoglobulin subtype. The rapid proliferation rate and errors in CSR and SHM predispose normal GC B cells to malignant transformation. As a consequence, DLBCLs exhibit multiple low frequency genetic alterations including chromosomal translocations, somatic mutations and copy number alterations (CNAs).

Given the numbers and types of genetic alterations in DLBCL, investigators have sought additional comprehensive classification systems to identify groups of tumors with similar molecular traits. Transcriptional profiling has been used to define DLBCL subsets that share certain features with normal B-cell subtypes (“cell-of-origin” classification, COO) (Lenz and Staudt, 2010). COO-defined DLBCLs include “germinal center B-cell” (GCB) and “activated B-cell” (ABC) types and an additional group of unclassified tumors. The COO-defined tumor groups are characterized by certain biological features, most notably increased NFκB activity and less favorable outcome in ABC-type DLBCLs (Lenz and Staudt, 2010). However, the outcome differences in GCB and ABC-type DLBCLs may be less striking in patients treated with current rituxan-containing combination chemotherapy regimens (Fu et al., 2008; Lenz et al., 2008a). An alternative transcriptional profiling classification, termed comprehensive consensus clustering (CCC), identifies DLBCL subtypes solely on the basis of distinctions within primary tumors and includes the 3 groups: “B-cell receptor” (BCR); “Oxidative Phosphorylation” (OxP); and “Host-response” (HR) (Chen et al., 2008; Monti et al., 2005).

To date, genetic alterations in DLBCL have largely been analyzed as single features or in association with the defined transcriptional subtypes (Bea et al., 2005; Lenz et al., 2008b). The platforms that were previously used to define CNAs in DLBCL had lower resolution and concordant assessments of transcript abundance and copy number were more limited. For these reasons, the precise boundaries of CNAs, the associated candidate “driver genes” and implicated pathways require further definition.

The earlier observations regarding cellular proliferation in DLBCL prompted additional analyses of certain individual cell cycle components and regulators. Cell cycle progression is controlled by series of cyclin-dependent kinases (CDKs) which are complexed with specific cyclins (Malumbres and Barbacid, 2009). The cyclin D-dependent kinases 4 and 6 (CDK4/CDK6) and the cyclin E-associated kinase 2 (CDK2) sequentially phosphorylate the retinoblastoma (RB) proteins, releasing the E2F transcription factors and promoting cell cycle progression. The A-type cyclins also activate CDK2 and CDK1 promoting S phase transition and the onset of mitosis. CDK activity is regulated by inhibitors such as the INK4 family member, p16INK4A (at the CDKN2A locus), and certain p53 targets such as p21, among others (Malumbres and Barbacid, 2009). Individual cell cycle components and regulators reported to be perturbed in small numbers of DLBCLs include CDKN2A (ARF and p16INK4A), p53 and its target, p21, cyclin D3 and RB1 (Jardin et al., 2010; Pasqualucci et al., 2011; Sanchez-Beato et al., 2003; Winter et al., 2010; Young et al., 2008). Recent deep sequencing analyses confirm earlier reports of TP53 somatic mutations in approximately 20% of DLBCLs (Morin et al., 2011; Pasqualucci et al., 2011), a much lower percentage than in certain non-hematologic malignancies (TCGA, 2008; TCGA, 2011). The relatively low frequency of TP53 somatic alterations in primary human DLBCLs suggests that additional bases of p53 deficiency remain to be defined.

Herein, we integrate copy number data with transcriptional profiles and perform pathway analyses to identify core deregulated and targetable pathways in primary DLBCLs.

RESULTS

Mapping recurrent copy number alterations in primary DLBCL

Recurrent CNAs in the 180 primary DLBCLs were detected using the GISTIC (Genomic Identification of Significant Targets in Cancer) algorithm. Within the identified regions of significant copy number (CN) gain or loss, narrower peaks of maximally significant CN change were identified (Supplemental Methods). We found 47 recurrent CNAs, including 21 copy gains and 26 copy losses, with frequencies of 4% to 27% (Figure 1 and Table S1). The GISTIC-defined CNAs range from narrow focal alterations such as amplification peak 2p16.1 to chromosome arm and whole-chromosome alterations, including gain of 1q, loss of 6q and gain of chromosome 7 (Figure 1).

Figure 1
Recurrent CNAs in newly diagnosed DLBCLs

Comparison of CNAs in DLBCLs and non-hematological cancers

To distinguish between CNAs that are unique to DLBCL and those that are found in other tumors, we compared the DLBCL GISTIC analysis to that of 2433 non-hematological cancers (Beroukhim et al., 2010). The CNAs in DLBCLs and the non-hematologic cancers were visualized with a mirror plot (Figure 2) and the CNA overlap in the two series was computed (Figure S1A and Supplemental Experimental Procedures). Seven of 21 (33%) regions of copy gain and 16/26 (62%) regions of copy loss were common to both series; additional regions of copy gain exhibited partial overlap (Figure 2 and Figure S1A). Examples of shared alterations include gains of chromosome 7 and chromosome 1q and loss of chromosome 6q, suggesting a broader role for these alterations in multiple tumor types. In contrast, 9/21 (43%) regions of copy gain and 10/26 (38%) regions of copy loss were only identified in DLBCL, including gains of 2p16.1 and 19q13.42 (Figure 2 and Figure S1A). These DLBCL-selective CNAs were largely absent in a lymphoid malignancy of non-GC origin (Figure S1B).

Figure 2
Comparison of CNAs in primary DLBCLs and non-hematologic cancers

Integrative analysis of CNAs and transcript abundance

We anticipated that DLBCL CNAs would alter the corresponding gene transcript levels and prioritized genes with the most significant association between transcript abundance and CNA. All genes within the 47 defined CNA peaks and regions (Table S2) were analyzed for the association between transcript abundance and the presence/absence of the gene alteration (peak or region) across the DLBCL series. The “cis-signature” of a given CNA was defined as the set of within-peak (or within-region) genes with the most significant association between CN and transcript abundance (FDR q-values ≤ .25; top 5 peak transcripts, Figure 1 and Table S1; complete list, Table S3).

CNAs of genes with known roles in lymphomagenesis

The two genes most closely associated with the 6q21 and 6q23.3 copy loss were PRDM1 (BLIMP1) and TNFAIP3 (A20), respectively (Figure 1 and Table S1). Both genes are confirmed tumor suppressors that can be inactivated by several mechanisms, including copy loss (Calado et al., 2010; Kato et al., 2009; Pasqualucci et al., 2006). Deletion of the ubiquitin-editing enzyme, TNFAIP3, contributes to lymphoid transformation, in part, by deregulating NFκB signaling (Shembade et al., 2010). Inactivation of the PRDM1 transcriptional repressor promotes lymphomagenesis by blocking normal plasma cell differentiation (Mandelbaum et al., 2010).

The additional tumor suppressor genes, CDKN2A, RB1, FAS and TP53 were closely associated with 9p21.3, 13q14.2, 10q23.31 and 17p13.1 copy loss, respectively (Figure 1), consistent with earlier analyses (Jardin et al., 2010; Sanchez-Beato et al., 2003). Furthermore, two well-known oncogenes were tightly linked with amplification peaks, REL at 2p16.1 and BCL2 at 18q21.33 (Figure 1). Copy gains of 2p16.1/REL and 12q15 were more frequent in GCB DLBCLs whereas gains of 18q21.32/BCL2 and 19q13.42 were more common in ABC tumors, as described (Table S1) (Bea et al., 2005; Lenz et al., 2008b). Given the identification of known CNAs in DLBCL, the integrative analysis will likely define additional CNAs and genes with previously unappreciated roles in the disease.

CNAs of newly identified genes in DLBCL

The genes most closely associated with amplification of 1q23.3 (seen in 15% of DLBCLs) encode the low-affinity receptors for the IgG Fc receptors, FCGR2B (CD32B) and FCGR2C, and the related protein, FCRLA (FCRL1) (Figure 1). Increased FCGR2B expression was previously associated with adverse outcome in DLBCL (Camilleri-Broet et al., 2004) and FCGR2C copy number variation and overexpression was linked with certain autoimmune diseases (Breunis et al., 2008). In addition, FCRLA was preferentially expressed in B cells and postulated to be an activating co-receptor (Leu et al., 2005).

Genes associated with amplification of the 19q13.42 region include PRMT1 (protein arginine methyl tranferase 1) and BCL2L12 (Table S1). PRMT1 specifically dimethylates histone H4 at arginine 3 which generally serves as an activation signal (Nicholson et al., 2009). In addition, PRMT1 modifies transcription factors including FOXO1 (Yamagata et al., 2008) and signaling intermediaries such as the Igα subunit of the B-cell receptor (Infantino et al., 2010). BCL2L12 is an atypical BCL2 family member with cytoplasmic and nuclear roles. Cytoplasmic BCL2L12 inhibits caspases 3 and 7 whereas nuclear BCL2L12 interacts with p53 and inhibits its binding to target gene promoters (Stegh and DePinho, 2011).

CNAs of genes required for tumor immune recognition

In addition to identifying individual genes targeted by specific CNAs, we noted several alterations that perturbed genes required for tumor immune recognition. Copy loss of 6q21.33 decreased the abundance of the MHC class I molecules, HLA-B and HLA-C, at the peak and the MHC class I polypeptide-related sequences A and B, MICA and MICB, in the region (Figure 1 and Table S1). In addition, copy loss of 15q21.1 and 1p13.1 reduced the abundance of the peak β2 microglobulin (β2M) and CD58 transcripts, respectively (Figure 1) and 19p13.3 copy loss decreased the levels of the region TNFSF9 (CD137L) transcripts (Table S1).

The β2M polypeptide associates with HLA class I heavy chains on the cell surface to present antigen. In the absence of β2M, stable antigen-HLA class I complexes cannot be formed. Both HLA class I and B2M copy loss were previously described in large B-cell lymphomas of immunoprivileged sites (Booman et al., 2008; Jordanov et al., 2003) and inactivating mutations and deletions of B2M were recently reported in DLBCLs (Challa-Malladi et al., 2011; Pasqualucci et al., 2011).

The 6q21.33 region genes, MICA and MICB (Table S1), encode ligands of the activating NKG2D receptor which is expressed by NK cells and a subset of T cells (Raulet, 2003). Decreased expression of these NKG2D ligands likely limits an innate NK-cell mediated anti-tumor immune response.

The 1p13.1 peak gene, CD58 (LFA3) (Figure 1), encodes a member of the immunoglobulin superfamily that is a ligand for the co-stimulatory CD2 receptor on T and NK cells. CD58 was recently reported to be the target of inactivating somatic mutations in a small subset of DLBCLs (Challa-Malladi et al., 2011; Pasqualucci et al., 2011), providing additional evidence that CD58 loss promotes tumor immune escape.

The 19p13.3 region gene, TNFSF9 (Table S1), encodes the ligand for the CD137 costimulatory receptor which is expressed by follicular dendritic cells (FDC) and primed CD8+ memory T cells (Middendrop et al., 2009). Interactions between TNFSF9 on GC B cells and CD137 on FDC and T cells regulate the GC B-cell response and TNFSF9 loss promotes the development of GCB lymphomas (Middendrop et al., 2009).

Pathway enrichment analyses reveal coordinate deregulation of p53 signaling and cell cycle

After identifying CNAs of several genes required for tumor immune recognition, we sought a more comprehensive method to characterize additional pathways perturbed by CNAs in DLBCL. We first defined global cis-acting peak or region signatures as the union of all individual cis-acting peak or region signatures (Figure 3Aa). Thereafter, we performed pathway enrichment of the global signatures using a curated series of gene sets and ranked the results by FDR (Figures 3Aa and B, top pathways; Table S4, full analysis). In the global peak signature, 13 of 15 of the most significantly enriched gene sets reflect related aspects of p53 signaling, apoptosis and cell cycle regulation (Figure 3B, top panel, FDR < .10). Although the gene sets have different names, they include common genes that are targeted by CNAs – TP53, CDKN2A, RB1 and RBL2 (all copy loss) and BCL2 (copy gain) (Figure 3B, top panel).

Figure 3
Pathway and transcription factor (TF) binding site enrichment

In the global region signature, the most significantly enriched gene set is the “p53 signaling pathway” (Figure 3B, bottom panel, FDR .0003). Additional p53 pathway components altered by CNAs include the p53 modifiers, MDM2, MDM4, RFWD2 (COP1) (all copy gain); p53 targets, PERP, SCOTIN, TNFRSF10 (DR5/TRAIL receptor) and FAS (all copy loss); and critical cell cycle regulators, CCND3 (cyclin D3), CDK4, CDK6 and CDK2 (all copy gain) (Figure 3B, bottom panel).

Components of the p53, apoptotic and cell cycle pathways perturbed by CNAs

CNAs of p53, apoptotic and cell cycle pathway members are illustrated in Figure 4.

Figure 4
Components of the p53, apoptotic and cell cycle pathways perturbed by CNAs

p53 pathway

CNAs of p53 pathway components all had the same predicted downstream effect – decreased abundance of functional p53 and reduced levels of associated p53 targets. Copy loss of CDKN2A, at 9p21.3, occurs in 24% of DLBCLs (Figure 4). The two alternative transcripts derived from the CDKN2A locus, p16INK4A and ARF, have complementary roles in p53 signaling and cell cycle regulation. ARF interferes with binding of the MDM2 E3 ligase to p53, decreasing its ubiquitylation and proteasomal degradation (Brooks and Gu, 2006). As a consequence, CDKNA2 deletion (ARF loss) and MDM2 (12q15) amplification both increase the ubiquitylation and subsequent degradation of p53 (Figure 4). Two additional E3 ligases with complementary but largely non-overlapping functions in destabilizing cellular p53 levels, MDM4 and RFWD2 (COP1), are increased by 1q23.3 copy gain (Figure 4) (Dornan et al., 2004).

Moreover, TP53 itself and two positive p53 modifiers, RPL26 and KDM6B (JMJD3), are targeted by 17p13.1 copy loss (Figure 4). The H3K27 demethylase, KDM6B, participates in the active removal of the repressive methyl mark from p16INK4A-ARF, contributing to its transcriptional activation (Agger et al., 2009). Therefore, KDM6B copy loss represents an additional mechanism of indirectly reducing functional p53 activity (Figure 4). KDM6B also directly modulates p53 methylation, cellular distribution and function (Sola et al., 2011). The other positive modifier of p53 activity, RPL26, binds to the 5’ UTR of TP53, promotes its translation and significantly increases stress-induced p53 levels (Chen and Kastan, 2010; Takagi et al., 2005) (Figure 4). RPL26 is also a target of MDM2 which polyubiquitylates the ribosomal protein and enhances its proteasomal degradation (Ofir-Rosenfeld et al., 2008) (Figure 4). In addition, the recently identified negative modulator of p53 transcriptional activity, BCL2L12 (at 19q13.42), is amplified in a subset of DLBCLs (Figure 4).

Apoptotic pathways

Independent of its role in regulating p53, BCL2L12 amplification limits apoptosis by blocking the effector caspases 3 and 7 (Figure 4). An additional means of perturbing the intrinsic apoptotic pathway is BCL2 copy gain (18q21.33) (Figure 4). Copy loss also decreases the abundance of several p53 targets that promote apoptosis, including the extrinsic apoptotic pathway components, FAS, TNFRF10B, SCOTIN and PERP (Figure 4) (Beaudry et al., 2010; Bourdon et al., 2002; Wilson et al., 2009).

Cell cycle degregulation

The loss of p16INK4A and decreased abundance of p53 targets such as p21 and GADD45 relieve repression of the cell cycle components, CCND3 (cyclin D3), CDK2 and CDK1, respectively (Figure 4). In addition, CDK2, CCND3 and the cyclin D-associated CDKs, CDK4 and CDK6 are increased by copy gain (Figure 4). In addition, RB1 and the related RB locus, RBL2 (p130), are targeted by copy loss in a subset of DLBCLs (Figure 4). RB1 is also a recognized target of the MDM2 E3 ligase (Polager and Ginsberg, 2009).

Signature of E2F activation

We next sought an unbiased approach to assess the relationship between CNA-dependent changes and the abundance of E2F target genes. Because transcription factors (TF) such as E2F will target genes outside the identified CNAs, we first defined the “trans-acting signature” of each CNA (those genes outside the CNA with the most significant association between transcript abundance and the CNA, Figure 3Ab). The union of the cis- and trans-acting signatures, termed the “global cis/trans-acting transcriptional signature”, was then tested for enrichment of genes with common TF binding sites (Figure 3Ab). The “global cis/trans-acting transcriptional signature” was significantly enriched for genes containing E2F binding sites; specifically, 7/7 of top-ranked binding sites were either E2F, E2F/DP1 or E2F/DP2 (Figure 3C, full list in Table S4). Therefore, DLBCL CNAs are tightly associated with cell cycle deregulation and increased abundance of E2F target genes.

Patterns of CNAs of pathway components

The analysis of CNAs that perturb p53 signaling, apoptosis and cell cycle regulation also illustrates four important principles. First, a single CNA may alter several genes which synergistically modulate the same pathway, as in 17p13.1 copy loss decreasing expression of p53 itself and the p53 modifiers, RPL26 and KDM6B (JMJD3) (Figure 4). Second, several CNAs may modify the same pathway. For example, 1q23.3 copy gain (MDM4 and RFWD2), 9p21.3 copy loss (CDKN2A), 12q15 copy gain (MDM2), 17p13.1 copy loss (TP53, RPL26 and KDM6B) and 19q13.42 copy gain (BCL2L12) all function to decrease p53 activity (Figure 4). Third, certain single CNAs may alter complementary pathways such as 12q15 amplification (CDK2, CDK4 and MDM2) enhancing cell cycle progression and reducing p53 activity (Figure 4). Fourth, multiple CNAs may modify complementary pathways such as p53 signaling, apoptosis and cell cycle regulation (Figure 4).

CNAs of p53 pathway and cell cycle components in individual primary DLBCLs

After comprehensively defining CNAs that perturb p53 signaling and cell cycle pathways in DLBCLs, we assessed the patterns and combinations of alterations that occur in individual tumors. When the primary DLBCLs were clustered in the space of the CNAs that alter p53 pathway and cell cycle components, 66%(118/180) of tumors had multiple alterations (termed “complex”) whereas the remaining 34% of tumors lacked these lesions (designated “clean”, Figure 5A). Primary DLBCLs with single copy loss of 17p13.1 (TP53/RPL26/KDM6B) often had CNAs perturbing an additional p53 modifier – 9p21.3 (CDKN2A/ARF), 19q13.42 (BCL2L12), 12q15 (MDM2) or 1q23.3 (MDM4/RFWD2) (Figure 5A). Of interest, CNAs of the respective p53 modifiers, CDKN2A (ARF, 9p21.3), MDM2 (12q15) and MDM4/RFWD2 (1q23.3) occurred in largely separate groups of tumors (Figure 5A). DLBCLs with CNAs of p53 pathway members frequently exhibited concurrent alterations of additional cell cycle components such as CCND3 (6p21.32), CDK6 (7q22.1), CDK2/CDK4 (12q15) and/or RB1 (13q14.2) or RBL2 (16q12.2) (Figure 5A). Tumors with “complex” patterns of p53 pathway and cell cycle components also had more total CNAs than DLBCLs with “clean” p53/cell cycle signatures (Figure 5A, bottom panel, Σ all CNAs, “complex” vs. “clean” p < .0001 and Figure S2A) and more frequent TP53 mutations (Figure 5A top panel, “complex” 22% vs. “clean” 7%, p < 0.005, Figure S2 and Table S5). The patterns of “complex” vs. “clean” CNAs of p53 pathway and cell cycle components and the association between “complex” signature and total CNAs were confirmed in an independent series of 79 primary DLBCLs (Figure S2B).

Figure 5
CNAs of p53 pathway and cell cycle components in individual primary DLBCLs and association with outcome

To further characterize “complex” vs. “clean” tumors, we performed gene set enrichment analysis (GSEA) with publicly available series of p53 target genes and a RB-deficiency gene set which included multiple E2F targets (Knudsen and Knudsen, 2008). The GSEA computational method identifies statistically significant, concordant differences in the transcript abundance of a previously defined set of genes (such as p53 targets) in two biological states (ie, “clean” versus “complex” primary DLBCLs) (Subramanian et al., 2005). The p53 target transcripts were significantly less abundant in “complex” DLBCLs, directly linking their genetic signature of p53 deficiency with decreased p53 activity (Figure 5B and Figure S2C). Furthermore, the RB-deficiency gene set was significantly enriched in “complex” DLBCLs suggesting that these tumors had increased E2F-mediated cell cycle progression (Figure 5C). Consistent with these observations, DLBCLs with “complex” CNA patterns also had significantly higher proliferation indices as determined by Ki67 immunostaining (Figure 5D).

Structural complexity as a significant predictor of outcome

We next assessed the prognostic significance of the “complex” CNA pattern in the subset of patients who were treated with R-CHOP (rituxan, cyclophosphamide, adriamycin, oncovin, prednisone) and had long-term follow up (Tables S6 and S7). R-CHOP treated patients with “complex” CNA patterns had a 5-year overall survival of only 62% whereas those with “clean” CNA signatures were all cured (Figure 6A, p = .001). The association between CN complexity and outcome was independent of transcriptional COO categories (Figure S3).

Figure 6
Prognostic significance of “complex” vs. “clean” CNA pattern in DLBCLs

We next assessed the relationship of CN complexity and the clinical IPI risk model. Although the IPI was highly predictive of outcome (Low/Low-Intermediate vs. High-Int/High, Figure 6B, left panel), the CNA pattern significantly increased prognostic accuracy (Figure 6B, middle and right panel). In both the Low/Low-Intermediate and High-Intermediate/High-risk groups, patients whose tumors had “complex” CNAs had significantly shorter overall survivals whereas all patients with “clean” CNA patterns were cured (Figure 6B, middle and right panel). The contribution of the CNA pattern to IPI outcome stratification was also confirmed by a Cox-proportional hazard model (p<.001, Supplemental Methods). Taken together, these data provide a structural basis for deregulated cell cycle, increased cellular proliferation and unfavorable outcome in DLBCL.

Targeting deregulated cell cycle with broad-acting CDK inhibitors

The predictive value of the “complex” CNA pattern and its association with deregulated cell cycle and increased activation of CDK4/6, CDK2 and, likely CDK1 (Figure 4) prompted us to assess the activity of a broad-acting CDK inhibitor such as flavopiridol (Lapenna and Giordano, 2009) in DLBCL. We used a panel of DLBCL cell lines derived from patients with relapsed/refractory disease; all lines have decreased or absent p53 activity and CNAs of cell cycle components including CDKN2A, CCND3, CDK4, CDK6, CDK2 and/or copy loss of RB1 (Figure S4A). Flavopiridol, which inhibits CDKs 4/6, 2 and 1 (and CDK9), decreased the cellular proliferation of the DLBCL cell lines at nanomolar doses (Figure 7A). Similar results were obtained with a second pan-CDK inhibitor, AT-7519 (Figure S4B). Of interest, a DLBCL cell line with single copy RB1 loss (DHL7), was less sensitive to lower doses of flavopiridol (Figure 7A) consistent with RB1 being downstream of the targeted CDKs.

Figure 7
Targeting deregulated cell cycle with a pan-CDK inhibitor

In these DLBCL cell lines, treatment with the pan-CDK inhibitor decreased S phase and induced cell cycle arrest (Figure 7B). In addition, the broad-acting CDK inhibitor increased apoptosis, as assessed by subG1 peaks and Annexin V/7-AAD staining (Figure 7B and C), and decreased the phosphorylation of RB1 at CDK4/6 and CDK2-specific sites (pS780 and pT821, respectively) (Figure 7D and Figure S4B). In multiple DLBCL xenograft models, flavopiridol treatment significantly reduced tumor growth and lymphoma infiltration of bone marrow and spleen (Figures 8A–C). Taken together, these data suggest that genetically driven cell cycle deregulation in DLBCL may be amenable to targeted therapy.

Figure 8
In vivo efficacy of a pan-CDK inhibitor in DLBCL xenografts

Discussion

Using a combination of HD-SNP arrays, gene expression profiling and pathway analyses, we have comprehensively defined CNAs, associated candidate driver genes and perturbed signaling pathways in a large series of newly diagnosed DLBCLs. The precision of the HD-SNP platform allowed us to precisely determine the boundaries of recurrent CNAs and distinguish alterations that were unique to DLBCL from ones that were shared with non-hematologic malignancies. The multiple low-frequency CNAs prompted us to systematically evaluate the alterations and associated genes with pathway analyses. The approach revealed a large complementary set of CNAs that decreased p53 activity and perturbed cell cycle regulation. The CNA-associated signature of p53 deficiency and cell cycle deregulation was highly predictive for outcome and potentially amenable to targeted therapy.

P53 deficiency

The CNA-associated pattern of deregulated p53 signaling was detected in 66% of newly diagnosed DLBCLs, of note because somatic inactivating mutations of p53 are much less common in DLBCLs than in multiple epithelial malignancies. For example, only 16% of tumors in the current series of primary DLBCLs exhibited hemizygous TP53 mutations and the majority of these were in “complex” tumors with additional CNAs of p53 pathway members. All of the CNAs of p53 modulators and signaling pathway components had the same functional effect – decreased abundance of functional p53 and reduced levels of p53 targets. In addition to identifying previously described CNAs of p53 modifiers such as CDKN2A (ARF) and MDM2 and TP53 itself, we found CNAs of the p53 regulators, MDM4, RFWD2 and BCL2L12 in DLBCL. We also defined a “deletion block” on chromosome 17p13 that includes two additional p53 modifiers, KDM6B and RPL26, as well as TP53. The concurrent loss of TP53, RPL26 and KDM6B may perturb p53 signaling to a greater degree than anticipated in tumors with hemizygous 17p13 deletions. The gain of both MDM4 and RFWD2 at 1q23.3 delineates an additional “amplicon block” that serves to decrease p53 activity. These insights regarding genetic mechanisms that reduce normal p53 activity in DLBCL may inform targeted treatment strategies. For example, two recently developed p53 inhibitors are predicated on disrupting the interaction between functional p53 and the p53 modifiers, MDM2 and MDM4 (Bernal et al., 2010; Shangary and Wang, 2008).

Perturbed cell cycle regulation

Besides copy loss of the cyclin D-dependent kinase inhibitor, p16INK4A, we identified copy gain of CDK4, CDK6 and CCND3, the most abundant and essential D-type cyclin in germinal center B cells (Cato et al., 2011). In addition to the likely relief of p53/p21-dependent CDK2 inhibition, we also found copy gain of CDK2 in association with CDK4 (and MDM2) in a chromosome 12q15 “amplicon block” and copy loss of both RB1 and RBL2. There was a highly significant CNA-associated signature of increased E2F transcriptional activity underscoring the functional consequences of these genetic alterations.

The p53 and cell cycle component CNAs occur together in a comprehensive “complex” pattern in 66% of the primary DLBCLs; the remaining tumors have only rare CNAs. GSEA revealed that DLBCLs with “complex” CNAs had significantly less abundant expression of p53 target genes, directly linking their genetic signature of p53 deficiency with decreased p53 activity. In addition, these “complex” tumors exhibited enrichment of E2F targets by GSEA and increased cellular proliferation by Ki67 immunostaining. Most importantly, the “complex” CNA pattern is highly predictive for outcome in R-CHOP treated DLBCL patients. These findings, which provide a mechanistic basis for previous observations regarding the prognostic significance of cellular proliferation in DLBCL (Broyde et al., 2009; Grogan et al., 1988; Salles et al., 2011), should be further validated in future DLBCL series.

The current study highlights the value of a comprehensive approach to identify CNA-defined alterations of p53 and cell cycle regulatory pathways, some of which have been characterized on an individual or selective basis and associated with outcome in earlier studies (Faber and Chiles, 2007; Jardin et al., 2010; Sanchez-Beato et al., 2003; Winter et al., 2010; Young et al., 2008). We find that a single CNA (17p13.1) targets several p53 modulators, multiple CNAs perturb p53 activity (1q23.3, 9p21.3, 12q15, 17p13.1 and 19q13.42) and a single CNA (12q15/MDM2, CDK2 and CDK4) modulates both p53 signaling and cell cycle progression. Because many of these CNAs are shared with additional non-hematologic malignancies (Figure 2), these findings may also be applicable to other tumor types. In fact, an array CGH-defined “complex” pattern of copy gains and losses was recently associated with high mitotic counts and TP53 alterations in breast cancer (reviewed in (Kwei et al., 2010)).

Genomic instability in the subset of DLBCLs with perturbed p53 signaling and cell cycle deregulation

In our DLBCL series, tumors with “complex” CNAs of p53 and cell cycle components also had significantly more of the additional recurrent CNAs including focal and regional alterations and gains or losses of half or whole chromosomes (Figure S2A). The basis for the increased genomic instability in these “complex” DLBCLs remains to be defined but may be linked to the deficiencies in p53 signaling and perturbed cell cycle regulation. Numerical and structural chromosome instability (CIN) is better tolerated in a p53-deficient background and alterations of TP53, MDM2, MDM4 (MDMX), the CDK2 partner, CCNE1 (cyclin E1), and RB1 all foster CIN (Hernando et al., 2004; Matijasevic et al., 2008; Shlien et al., 2008; Thompson et al., 2010; Wang et al., 2008). In the setting of hyperactive CDKs and DNA damage, cell cycle progression further increases genomic instability (Malumbres and Barbacid, 2009).

In addition to CNAs of the p53 apoptotic pathway, DLBCLs with the “complex” pattern exhibit alterations of other apoptotic members including BCL2/18q21.33, FAS/10q23.32 and TNFRF10B/8p21.3 (Figure S2A). CNAs of immune recognition molecules, including HLA-B, HLA-C, MICA and MICB (6q21.33), B2M (15q21.1), CD58/1p13.1 and TNFSF9 (19p13.3), also largely occur in DLBCLs with “complex” patterns (Figure S2A). These data highlight the importance of evaluating specific genetic alterations in the context of a more comprehensive assessment of CNAs and associated genomic instability.

Clinical significance

The prognostic value of the perturbed p53 signaling/cell cycle deregulation signature prompted us to evaluate the activity of pan-CDK inhibitors in DLBCL. Following treatment, DLBCL cell lines with CNAs of p53 signaling and cell cycle components (with or without additional p53 mutations) exhibited decreased proliferation and RB1 phosphorylation and increased apoptosis in vitro and significantly reduced tumor growth in vivo. Therefore, prognostically significant, genetically driven cell cycle deregulation in DLBCL may be amenable to targeted treatment.

Experimental Procedures

Patients and primary tumor samples

High molecular weight DNA and total RNA were extracted from frozen biopsy specimens of newly diagnosed, previously untreated primary DLBCLs with ≥ 80% tumor involvement according to Institutional Review Board-approved protocols from three institutions (Mayo Clinic, Brigham & Women Hospital, and Dana-Farber Cancer Institute). For one subset of patients, informed consent was obtained (Mayo). For other patients, a waiver to obtain informed consent was granted by the local IRBs because otherwise discarded tissue was used. The series included 72 DLBCLs from patients who were treated with a rituxan-containing, anthracycline-based combination chemotherapy regimen (R-CHOP-like) and had long-term follow up; 68 of these patients had available information on all clinical parameters in the IPI (Table S6).

HD SNP array analysis and expression profiling

Primary DLBCL DNA samples and normal DNA specimens were profiled on Affymetrix HD-SNP arrays 6.0. (Supplemental Experimental Procedures). For the detection of CN alterations, the SNP array 6.0 data was processed through a previously described analytical pipeline (TCGA, 2008). Across-sample GISTIC analysis of the segmented data was carried out to identify statistically significant CNAs (Supplemental Experimental Procedures) (Beroukhim et al., 2007). Alteration regions with FDR q-values below .25 were considered significant. Within each region, a peak (or peaks) was identified as the contiguous set (or sets) of loci with highest q-values. To visualize the distribution of alterations across samples, we created a matrix with each entry indicating the presence/absence of an alteration (row) in a given sample (column).

RNA samples from 169 of the primary DLBCLs were transcriptionally profiled and the data were processed using Affymetrix’ MAS5 summarization method (Supplemental Experimental Procedures).

Integrative analysis

Cis-acting alteration signatures

The genes within the peak (region) of each GISTIC-identified alteration were tested for an association between their expression (transcript abundance) and the presence/absence of the harboring alteration by a two-group t-statistic with unequal variance. The cis-acting alteration signature for a given alteration was then defined as the set of within-peak (-region) transcripts with FDR q-values ≤ .25.

Trans-acting alteration signatures

The transcripts from genes which were outside an alteration peak were also evaluated for an association between their expression and the respective copy number alteration. The top 6000 transcripts ranked by across-sample median absolute deviation (MAD) were used as the candidate list. The trans-acting alteration signature for an alteration was defined as the set of outside-peak transcripts with FDR q-values ≤ .25 and fold change ≥ 1.3.

Pathway and Transcription Factor (TF) Binding Site Enrichment analysis

The global cis-acting signature, defined as the union of all of the individual cis-acting alteration signatures, was analyzed for pathway enrichment by testing the signature against curated gene sets from the MSigDB repository (C2 collection, version 2.5) (Supplemental Experimental Procedures). FDR-corrected q-values were computed based on the hypergeometric distribution.

The global trans-acting signature was defined as the union of all trans-acting alteration signatures. The union of the global cis-acting and trans-acting signatures was then analyzed for enrichment of targets of specific TFs using the curated sets of TF targets in the MSigDB C3 collection (Supplemental Experimental Procedures). FDR-corrected q-values were computed based on the hypergeometric distribution.

Highlights

  • A complementary set of CNAs decrease p53 activity and deregulate cell cycle.
  • DLBCLs either have multiple CNAs of p53/cell cycle components or lack these lesions.
  • The p53/cell cycle CNAs predict outcome and suggest targeted treatment approaches.
  • DLBCLs with p53/cell cycle CNAs have increased genomic instability.

Significance

In spite of advances in the molecular understanding of DLBCL pathogenesis, clinical models are still used to identify high-risk patients who then receive empiric therapy. In DLBCLs, which have infrequent inactivating somatic mutations of TP53 and RB1, the current studies define an alternative copy number-dependent mechanism of deregulating p53 and cell cycle. This genetic signature predicts outcome and suggests targeted approaches to treatment such as pan cyclin-dependent kinase (CDK) inhibition.

Supplementary Material

01

03

04

Acknowledgements

This work was supported by NIH PO1CA092625. BC was supported by a grant from the German Research Foundation (DFG Ch 735/1-1).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Accession Number

The HD-SNP 6.0 and gene expression data are available through the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under the accession number GSE34171.

Xenograft models

All animal studies were performed according to Dana-Farber Cancer Institute Institutional Animal Care and Use Committee (IACUC)-approved protocols.

Supplemental Experimental Procedures include bioinformatic analyses, GSEA, TP53 sequencing, immunohistochemical analyses, immunoblotting, cell lines and culture conditions, in vitro assays of proliferation, cell cycle and apoptosis following CDK inhibition and xenograft models.

Reference

  • Agger K, Cloos PAC, Rudkjaer L, Williams K, Andersen G, Christensen J, Helin K. The H3K27 me3 demethylase JMJD3 contributes to the activation of the INK4A-ARF locus in response to oncogene- and stress-induced senescence. Genes Dev. 2009;23:1171–1176. [PubMed]
  • Bea S, Zetti A, Wright G, Salaverria I, Jehn P, Moreno VCB, Ott G, Xavier P, Yang L, et al. Diffuse large B-cell lymphoma subgroups have distinct genetic profiles that influence tumor biology and improve gene-expression-based survival prediction. Blood. 2005;106:3183–3190. [PubMed]
  • Beaudry VG, Jiang D, Dusek RL, Park EJ, Knezevich S, Ridd K, Vogel H, Bastian BC, Attardi LD. Loss of the p53/p63 regulated desmosomal protein Perp promotes tumorigenesis. PloS Genetics. 2010;6:e1001168. [PMC free article] [PubMed]
  • Bernal F, Wade M, Godes M, Davis TN, Whitehead DG, Kung AL, Wahl GM, Walensky LD. A stapled p53 helix overcomes HDMX-mediated suppression of p53. Cancer Cell. 2010 [PMC free article] [PubMed]
  • Beroukhim R, Getz G, Nghiemphu L, Barretina J, Hsueh T, Linhart D, Vivanco I, Lee JC, Huang JH, Alexander S, et al. Assessing the significance of chromosomal aberrations in cancer: methodology and application in glioma. Proc Natl Acad Sci U S A. 2007;104:20007–20012. [PubMed]
  • Beroukhim R, Mermel CH, Porter D, Wei G, Raychaudhuri S, Donovan J, Barretina J, Boehm JS, Dobson J, Urashima M, et al. The landscape of somatic copy-number alteration across human cancers. Nature. 2010;463:899–905. [PMC free article] [PubMed]
  • Booman M, Szuhai K, Rosenwald A, Hartmann E, Kluin-Nelemans HC, de Jong D, Schuuring E, Kluin PM. Genomic alterations and gene expression in primary diffuse large B-cell lymphomas of immune-privileged sites: the importance of apoptosis and immunomodulatory pathways. J Pathol. 2008;216:209–217. [PubMed]
  • Bourdon J-C, Renzing J, Robertson PL, Fernandes KN, Lane DP. Scotin, a novel p53-inducible proapoptotic protein located in the ER and the nuclear membrane. J Cell Biol. 2002;158:235–246. [PMC free article] [PubMed]
  • Breunis WB, van Mirre E, Bruin M, Geissler J, de Boer M, Peters M, Roos D, de Haas M, Koene HR, Kuijpers TW. Copy number variation of the activating FCGR2C gene predisposes to idiopathic thrombocytopenic purpura. Blood. 2008;111:1029–1038. [PubMed]
  • Brooks CL, Gu W. p53 ubiquitination: Mdm2 and beyond. Mol Cell. 2006;21:307–315. [PMC free article] [PubMed]
  • Broyde A, Boycov O, Strenov Y, Okon E, Shpilberg O, Bairey O. Role and prognostic significance of the Ki-67 index in non-Hodgkin's lymphoma. Amer J Hematol. 2009:338–343. [PubMed]
  • Calado DP, Zhang B, Srinivasan L, Sasaki Y, Seagal J, Unitt C, Rodig SJK, Tarakhovsky A, Schmidt-Supprian M, Rajewsky K. Constitutive canonical NF-kappaB activation cooperates with disruption of BLIMP1 in the pathogenesis of activated B cell-like diffuse large cell lymphoma. Cancer Cell. 2010;18:580–589. [PMC free article] [PubMed]
  • Camilleri-Broet S, Cassard L, Broet P, Delmer A, Le Touneau A, Diebold J, Fridman WH, Molina TJ, Sautes-Fridman C. FcgammaRIIB is differentially expressed during B cell maturation and in B-cell lymphomas. Br J Haematol. 2004;124:55–62. [PubMed]
  • Cato MH, Chintalapati SK, Yau IW, Omori SA, Rickert RC. Cyclin D3 is selectively required for proliferative expansion of germinal center B cells. Mol Cell Biol. 2011;31:127–137. [PMC free article] [PubMed]
  • Challa-Malladi M, YK L, Califano O, Holmes AB, Bhagat G, Murty VV, Dominguez-Sola D, Pasqualucci L, Dalla-Favera R. Combined genetic inactivation of beta2-microglobulin and CD58 reveals frequent escape from immune recognition in diffuse large B cell lymphoma. Cancer Cell. 2011;20:728–740. [PMC free article] [PubMed]
  • Chen J, Kastan MB. 5'-3'-UTR interactions regulate p53 mRNA translation and provide a target for modulating p53 induction after DNA damage. Genes Dev. 2010;24:2146–2156. [PubMed]
  • Chen L, Monti S, Juszczynski P, Chen W, Kutok JL, Shipp MA. SYK-dependent tonic B-cell receptor signaling is a rational treatment target in diffuse large Bcell lymphoma. Blood. 2008;111:2230–2237. PMCID: PMC2234057. [PubMed]
  • Dornan D, Bheddah S, Newton K, Ince W, Frantz GD, Dowd P, Koeppen H, Dixit VM, French DM. COP1, the negative regulator of p53, is overexpressed in breast and ovarian adenocarcinomas. Cancer Res. 2004;64:7226–7230. [PubMed]
  • Faber AC, Chiles TC. Inhibition of cyclin-dependent kinase-2 induces apoptosis in human diffuse large B-cell lymphomas. Cell Cycle. 2007;6:2982–2989. [PubMed]
  • Friedberg JW. Diffuse large B-cell lymphoma. Hematol Oncol Clin North Am. 2008;22 941-ix. [PMC free article] [PubMed]
  • Fu K, Weisenburger DD, Choi WW, Perry KD, Smith LM, Shi X, Hans CP, Greiner TC, Bierman PJ, Bociek RG, et al. Addition of rituximab to standard chemotherapy improves the survival of both the germinal center B-cell-like and non-germinal center B-cell-like subtypes of diffuse large B-cell lymphoma. J Clin Oncol. 2008;26:4587–4594. [PubMed]
  • Grogan T, Kippman S, Spier C, Slymen D, Rybski J, Rangel C, Richter L, Miller T. Independent prognostic significance of a nuclear proliferation antigen in diffuse large cell lymphomas as determined by the monoclonal antibody Ki-67. Blood. 1988;71:1157. [PubMed]
  • Hernando E, Nahle Z, Juan G, Diaz-Rodriguez E, Alaminos M, Hermann M, Michel L, Mittal V, Gerald W, Benezra R, et al. Rb inactivation promotes genomic instability by uncoupling cell cycle progression from mitotic control. Nature. 2004;430:797–802. [PubMed]
  • Infantino S, Benz B, Waldmann T, Jung M, Schneider R, Reth M. Arginine methylation of the B cell antigen receptor promotes differentiation. J Exp Med. 2010;207:711–719. [PMC free article] [PubMed]
  • Jardin F, Jais J-P, Molina T-J, Parmentieer F, Picquenot J-M, Ruminy P, Tilly H, Bastard C, Salles G-A, Feugier P, et al. Diffuse large B-cell lymphomas with CDKN2A deletion have a distinct gene expression signature and a poor prognosis under R-CHOP treatment: a GELA study. Blood. 2010;116:1092–1104. [PubMed]
  • Jordanov ES, Riemersma SA, Philippo K, Schuuring E, Klun PM. Beta 2-microglobulin aberrations in diffuse large B-cell lymphoma of the testis and the central vervous system. Int J Cancer. 2003;103:393–398. [PubMed]
  • Kato M, Sanada M, Kato I, Sato Y, Takita J, Takeuchi K, Niwa A, Chen Y, Nakazaki K, Nomoto J, et al. Frequent inactivation of A20 in B-cell lymphomas. Nature. 2009;459:712–716. [PubMed]
  • Klein U, Dalla-Favera R. Germinal centres: role in B-cell physiology and malignancy. Nat Rev Immunol. 2008;8:22–33. [PubMed]
  • Knudsen ES, Knudsen KE. Tailoring to RB: tumour suppressor status and therapeutic response. Nat Rev Cancer. 2008;8:714–724. [PMC free article] [PubMed]
  • Kwei KA, Kung Y, Salari K, Holcomb IN, Pollack JR. Genomic instability in breast cancer: Pathogenesis and clinical implications. Molecular Oncology. 2010;4:255–266. [PMC free article] [PubMed]
  • Lapenna S, Giordano A. Cell cycle kinases as therapeutic targets for cancer. Nat Rev Drug Discov. 2009;8:547–566. [PubMed]
  • Lenz G, Staudt LM. Aggressive lymphomas. New Engl J Med. 2010;362:1417–1429. [PubMed]
  • Lenz G, Wright G, Dave SS, Xiao W, Powell J, Zhao H, Xu W, Tan B, Goldschmidt N, Iqbal J, et al. Stromal Gene Signatures in Large-B-Cell Lymphomas. New England Journal of Medicine. 2008a;359:2313–2323. [PubMed]
  • Lenz G, Wright GW, Emre NC, Kohlhammer H, Dave SS, Davis RE, Carty S, Lam LT, Shaffer AL, Xiao W, et al. Molecular subtypes of diffuse large B-cell lymphoma arise by distinct genetic pathways. Proc Natl Acad Sci USA. 2008b;105(36):13520–13525. [PubMed]
  • Leu CM, Davis RS, Gartland LA, Fine WD, Cooper MD. FcRH1: an activation coreceptor on human B cells. Blood. 2005;105:1121–1126. [PubMed]
  • Malumbres M, Barbacid M. Cell cycle, CDKs and cancer: a changing paradigm. Nature Rev Cancer. 2009;9:153–166. [PubMed]
  • Mandelbaum J, Bhagat G, Tang H, Mo T, Brahmachary M, Shen Q, Chadburn A, Rajewsky K, Tarakhovsky A, Pasqualucci L, Dalla-Favera R. BLIMP1 is a tumor suppressor gene frequently disrupted in activated B cell-like diffuse large B cell lymphoma. Cancer Cell. 2010;18:568–579. [PMC free article] [PubMed]
  • Matijasevic Z, Krzywicka-Racka A, Sluder G, Jones SN. MdmX regulates transformation and chromosomal stability in p53-deficient cells. Cell Cycle. 2008;7:2967–2973. [PMC free article] [PubMed]
  • Middendrop S, Xiao Y, Song JY, Peperzak V, Krijger PH, Jacobs H, Borst J. Mice deficient for CD137 logand are predisposed to develop germinal center-derived B-cell lymphoma. Blood. 2009;114:2280–2289. [PubMed]
  • Monti S, Savage KJ, Kutok JL, Feuerhake F, Kurtin P, Mihm M, Wu B, Pasqualucci L, Neuberg D, Aguiar RC, et al. Molecular profiling of diffuse large B-cell lymphoma identifies robust subtypes including one characterized by host inflammatory response. Blood. 2005;105:1851–1861. [PubMed]
  • Morin RD, Mendez-Lago M, Mungall AJ, Goya R, Mungall KL, Corbett RD, Johnson NA, Severson TM, Chiu R, Field M, et al. Frequent mutation of histone-modifying genes in non-Hodgkin lymphoma. Nature. 2011 advance online publication. [PMC free article] [PubMed]
  • Nicholson TB, Chen T, Richard S. The physiological and pathophysiological role of PRMT1-mediated protein arginine methylation. Pharmacol Res. 2009;60:466–474. [PubMed]
  • Ofir-Rosenfeld Y, Boggs K, Michael D, Kastan MB, Oren M. Mdm2 regulates p53 mRNA translation through inhibitory interactions with ribosomal protein L26. Mol Cell. 2008;32:180–189. [PMC free article] [PubMed]
  • Pasqualucci L, Compagno M, Houldsworth J, Monti S, Grunn A, Nandula SV, Aster JC, Murty VV, Shipp MA, Dalla-Favera R. Inactivation of the PRDM1/BLIMP1 gene in diffuse large B cell lymphoma. J Exp Med. 2006;203:311–317. [PMC free article] [PubMed]
  • Pasqualucci L, Trifonov V, Fabbri G, Ma J, Rossi D, Chiarenza A, Wells VA, Grunn A, Messina M, Elliot O, et al. Analysis of the coding genome of diffuse large B-cell lymphoma. Nature Genetics. 2011;43:830–837. [PMC free article] [PubMed]
  • Polager S, Ginsberg D. p53 and E2f: partners in life and death. Nature Rev Cancer. 2009;9:738–748. [PubMed]
  • Raulet DH. Roles of the NKG2D immunoreceptor and its ligands. Nat Rev Immunol. 2003;3:781–790. [PubMed]
  • Salles G, de Jong D, Xie W, Rosenwald A, Chhanabhai M, Gaulard P, Klapper W, Calaminici M, Sander B, Thorns C, et al. Prognostic significance of immunohistochemical biomarkers in diffuse large B-cell lymphoma: a study from the Lenenburg Lymphoma Biomarker Consortium. Blood. 2011;117:7070–7078. [PubMed]
  • Sanchez-Beato M, Sanchez-Aguilera A, Piris MA. Cell cycle deregulation in B-cell lymphomas. Blood. 2003;101:1220–1235. [PubMed]
  • Shangary S, Wang S. Targeting the MDM2-p53 interaction for cancer therapy. Clin Cancer Res. 2008;14:5318–5324. [PMC free article] [PubMed]
  • Shembade N, Ma A, Harhaj EW. Inhibition of NF-kappaB signaling by A20 through disruption of ubiquitin enzyme complexes. Science. 2010;327:1135–1139. [PMC free article] [PubMed]
  • Shipp M, Harrington D, Chairpersons, Anderson J, Armitage J, Bonadonna G, Brittinger G, Cabanillas F, Canellos G, Coiffier B, Connors J, et al. A predictive model for aggressive non-Hodgkin's lymphoma: The International NHL Prognostic Factors Project. N Engl J Med. 1993;329:987–994. [PubMed]
  • Shlien A, Tabori U, Marshall CR, Pienkowska M, Feuk L, Novokmet A, Nanda S, Druker H, Scherer SW, Malkin D. Excessive genomic DNA copy number variation in the Li-Fraumeni cancer predisposition syndrome. PNAS. 2008;105:11264–11269. [PubMed]
  • Sola S, Xavier JM, Santos DM, Aranha MM, Morgado AL, Jepsen K, Rodrigues CMP. p53 interaction with JMJD3 results in its nuclear distribution during mouse neural stem cell differentiation. PloS. 2011;6:1–10. [PMC free article] [PubMed]
  • Subramanian A, Tamayo P, Mootha VK, et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. PNAS. 2005;102:15545–15550. [PubMed]
  • Takagi M, Absalon MJ, McLure KG, Kastan MB. Regulation of p53 translation and induction after DNA damage of ribosomal protein L26 and nucleolin. Cell. 2005;123:49–63. [PubMed]
  • TCGA. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature. 2008;455:1061–1068. [PMC free article] [PubMed]
  • TCGA. Integrated genomic analyses of ovarian carcinoma. Nature. 2011;474:609–615. [PMC free article] [PubMed]
  • Thompson SL, Bakhoum SF, Compton DA. Mechanisms of chromosomal instability. Current Biol. 2010;20:R285–R295. [PMC free article] [PubMed]
  • Wang P, Lushnikova T, Odvody J, Greiner TC, Jones SN, Eischen CM. Elevated Mdm2 expression induces chromosomal instability and confers a survival and growth advantage in B cells. Oncogene. 2008;27:1590–1598. [PubMed]
  • Wilson NS, Dixit V, Ashkenazi A. Death receptor signal transducers: nodes of coordination in immune signaling networks. Nature Immunol. 2009;10:348–355. [PubMed]
  • Winter JN, Li S, Aurora V, et al. Expression of p21 protein predicts clinical outcome in DLBCL patients older than 60 years treated with R-CHOP but not CHOP: A prospective ECOG and Southwest Oncology Group Correlative Study on E4494. Clin Cancer Res. 2010;16:2435–2442. [PMC free article] [PubMed]
  • Yamagata K, Daltoku H, Takahashi Y, Namiki K, Hisatake K, Kako K, et al. Arginine methylation of FOXO transcription factors inhibits their phosphorylation by Akt. Mol Cell. 2008;32:221–231. [PubMed]
  • Young KH, Leroy K, Moller MB, Colleoni GWB, Sanchez-Beato M, Kerbauy FR, Haioun C, Eickhoff JC, Young AH, Gaulard P, et al. Structural profiles of TP53 gene mutations predict clinical outcome in diffuse large B-cell lymphoma: an international collaborative study. Blood. 2008;112:3088–3098. [PubMed]