《Nature》：确定出新的癌症基因

癌基因是一些在突变或功能障碍时会导致正常细胞发生癌变的基因。Memorial Sloan-Kettering癌症中心（MSKCC）的研究人员最近确定出了一种新的对癌症形成至关重要的癌基因。研究人员将这种新基因命名为POKEMON（POK Erythroid Myeloid Ontogenic factor）。这项研究刊登在2005年1月20日的《Nature》杂志上。

已经知道有一些基因能够导致癌症，但是Pokemon基因的独特之处为它是其他癌基因引发癌症所必须的。更重要的是，由于Pokemon蛋白在癌症的形成过程中扮演关键的角色，因此它将可能成为癌症新药物治疗的一个高效的靶标。

Pokemon能够控制将正常细胞变成癌细胞所需的途径。研究人员发现当将小鼠的Pokemon基因敲除时，这种转变过程就被抑制，因此细胞也不会发生癌变。能够抑制这种蛋白功能的药物将可能成为一种强有力的抗癌剂。

研究人员通过将这种癌基因插入小鼠的试验确证了Pokemon的致癌作用。Pokemon能够通过抑制其他蛋白（包括一种肿瘤抑制蛋白ARF）的功能来搞破坏。这些小鼠患上了侵略性的、致命类型的淋巴瘤。在进一步的研究中，研究人员利用组织芯片技术（tissue micro arrays）分析了来自不同类型癌症患者的样品。他们发现Pokemon在一定类型的B细胞和T细胞淋巴瘤中水平非常高。他们还发现具有高水平Pokemon蛋白的肿瘤恶化的可能性更大。

Pokemon蛋白是转录因子蛋白家族的一员，并且在人类癌症中发生了突变。这种蛋白很可能在固体肿瘤中处于重要地位，而且研究人员现在已经能够专一地干预这些转录因子的活性。这种新基因的确定将为癌症治疗提供一个新的靶标。

Nature 433, 278 - 285 (20 January 2005); doi:10.1038/nature03203

Role of the proto-oncogene Pokemon in cellular transformation and ARF repression

TAKAHIRO MAEDA1,2, ROBIN M. HOBBS1,2, TAHA MERGHOUB1,2, ILHEM GUERNAH1,2, ARTHUR ZELENT3, CARLOS CORDON-CARDO2, JULIE TERUYA-FELDSTEIN2 & PIER PAOLO PANDOLFI1,2

1 Cancer Biology and Genetics Program,
2 Department of Pathology, Memorial Sloan-Kettering Cancer Center, Sloan-Kettering Institute, 1275 York Avenue, New York, New York 10021, USA
3 Leukemia Research Fund Center at the Institute of Cancer Research, Chester Beatty Laboratories, Fulham Road, London SW3 6JB, UK

Correspondence and requests for materials should be addressed to P.P.P. (p-pandolfi@ski.mskcc.org).

Aberrant transcriptional repression through chromatin remodelling and histone deacetylation has been postulated to represent a driving force underlying tumorigenesis because histone deacetylase inhibitors have been found to be effective in cancer treatment. However, the molecular mechanisms by which transcriptional derepression would be linked to tumour suppression are poorly understood. Here we identify the transcriptional repressor Pokemon (encoded by the Zbtb7 gene) as a critical factor in oncogenesis. Mouse embryonic fibroblasts lacking Zbtb7 are completely refractory to oncogene-mediated cellular transformation. Conversely, Pokemon overexpression leads to overt oncogenic transformation both in vitro and in vivo in transgenic mice. Pokemon can specifically repress the transcription of the tumour suppressor gene ARF through direct binding. We find that Pokemon is aberrantly overexpressed in human cancers and that its expression levels predict biological behaviour and clinical outcome. Pokemon's critical role in cellular transformation makes it an attractive target for therapeutic intervention.

The POZ domain (for poxvirus and zinc finger; also known as the BTB domain) is a highly conserved protein–protein interaction domain found in about 250 known human proteins so far ( http://www.ebi.ac.uk/interpro/IEntry?ac=IPR000210). Members of the POK (POZ and Krüppel) family of proteins contain an amino-terminal POZ domain and a carboxy-terminal DNA-binding domain made of Krüppel-type zinc fingers and can act as potent transcriptional repressors through the recruitment of histone deacetylases (HDACs) and subsequent chromatin remodelling1, 2. POK proteins are important in embryonic development, differentiation and oncogenesis3-7. In particular, PLZF (for promyelocytic leukaemia zinc finger)8 and BCL6 (for B-cell lymphoma 6)9, two members of this POK family, are involved in chromosomal translocations associated with acute promyelocytic leukaemia and non-Hodgkin's lymphoma, respectively.

Pokemon (for POK erythroid myeloid ontogenic factor; also known as LRF10, OCZF11 and FBI-1 (ref. 12)) is a POK protein family member that has a critical and pleiotropic function in cellular differentiation and can physically interact with other POK family members such as BCL6 (ref. 10). Zbtb7 inactivation in mouse results in embryonic lethality and impaired cellular differentiation in multiple tissues including the B-cell compartment (T.M. and P.P.P., unpublished data). It could therefore be proposed that perturbation of Pokemon's transcriptional function impairs cellular differentiation, as is often observed in human cancer. In a non-mutually exclusive hypothesis, Pokemon could modulate cellular functions/pathways important for oncogenic transformation, thus taking a more direct role in tumorigenesis. To test this hypothesis we therefore assessed the role of Pokemon in cancer pathogenesis through a direct genetic approach.

Role of Pokemon in oncogenesis
We initially tested whether loss of Pokemon function would affect the transforming potential of primary cells. To this end, we performed classic cell growth and transformation assays by taking advantage of early-passage mouse embryonic fibroblast cells (MEFs). In this experimental setting, combinations of potent oncogenes such as E1A + H-rasV12, Myc + H-rasV12, T-antigen (T-Ag) + H-rasV12 normally elicit a marked proliferative response in wild-type (WT) MEFs13 (Fig. 1a). In addition, although overexpression of a single oncogene does not transform WT MEFs, combinations of oncogenes can induce cellular transformation as revealed by a colony formation assay in soft agar (Fig. 1b). Strikingly, Zbtb7-/- MEFs were completely unresponsive to both the proliferative stimulus triggered by these oncogene combinations and their transforming ability, whereas the expression levels of introduced oncogenes were comparable in WT and Zbtb7-/- MEFs (Supplementary Fig. 1a). Pokemon is therefore required for the growth-promoting and transforming activity of these combinations of oncogenes.

Figure 1 Pokemon is indispensable for cellular transformation and acts as a proto-oncogene. Full legend

High resolution image and legend (78k)

We next determined whether Pokemon would display proto-oncogenic activity when coexpressed in combination with other oncogenes. To this end, we infected WT MEFs with retroviruses containing Myc, H-rasV12, T-Ag and/or Pokemon genes and generated a growth curve. In this experimental condition, mock-infected control cells failed to proliferate, because cells were plated sparsely and lost cell–cell contact (not shown). As described previously, MEFs infected with Myc alone underwent apoptosis14, whereas expression of oncogenic ras resulted in premature senescence15 (not shown). As a consequence, MEFs infected singly with H-rasV12 or Myc failed to proliferate (Fig. 1c). In contrast, Pokemon coexpression blocked oncogene-induced apoptosis and senescence and, as a result, MEFs co-infected with Pokemon and Myc, H-rasV12 or T-Ag displayed a marked proliferative advantage (Fig. 1c). Furthermore, Pokemon-infected Rat1 MycErTM cells16 were more resistant to Myc-induced apoptosis than mock-infected cells (Supplementary Fig. 1c). Strikingly, MEFs co-infected with T-Ag, Myc, H-rasV12 and Pokemon readily formed colonies when plated in soft agar (Fig. 1d and Supplementary Fig. 1b). Pokemon also opposed E1A-induced apoptosis17, thus stimulating the growth of co-infected MEFs. However, E1A and Pokemon did not transform MEFs when co-infected (not shown).

Pokemon is a specific repressor of ARF
We set out to investigate the mechanisms by which Pokemon would exert its oncogenic activity. It has been reported that the N-terminal POZ domain of POK family members can mediate their physical interaction with a large co-repressor complex1, 2, and Pokemon is no exception to this rule (not shown). We therefore proposed that Pokemon might directly repress the expression of a pivotal tumour suppressor whose silencing would result in cellular transformation in combination with the aforementioned oncogenes. We therefore first characterized the putative Pokemon-binding sequence by CAST (for cyclic amplification and selection of targets; see Supplementary Fig. 2a, b) analysis18 followed by electrophoretic mobility-shift assay (EMSA). Pokemon directly and specifically bound oligonucleotides containing an identified Pokemon-binding sequence in vitro (Supplementary Fig. 2c). Furthermore, Pokemon could repress the activity of an artificial promoter containing Pokemon-binding sites (Supplementary Fig. 2d). To identify the core binding sequence in a Pokemon-binding site, we generated a series of mutated oligonucleotides and performed EMSA assays. Two cytosine residues in the centre of consensus sequence (underlined in Fig. 2a) were revealed to be essential for Pokemon binding (Supplementary Fig. 2e–g).

Figure 2 Pokemon is a key ARF transcriptional repressor. Full legend

High resolution image and legend (77k)

Subsequently, we searched the promoters of key tumour suppressor genes for putative Pokemon-binding sites. In this respect, p19Arf and p53 were legitimate candidates in view of the fact that their inactivation is known to cooperate with both Myc and oncogenic ras in neoplastic transformation14. We found several putative Pokemon-binding sites in the ARF promoter, and several transcriptional modulators, such as E2F1 (refs 19, 20), Bmi-1 (refs 21–26), TBX2, TBX3 (refs 27–29) and Dmp1 (ref. 30), have been reported to regulate the ARF promoter (Fig. 2b and Supplementary Fig. 3a).

To determine whether Pokemon would regulate ARF gene expression, we first performed transactivation assays with ARF promoter constructs fused to luciferase reporters. Pokemon efficiently repressed both p19Arf and p14ARF basal reporter activity in a dose-dependent manner (Fig. 2c). We also determined that both the POZ and zinc finger domains are required for this repressive activity (Supplementary Fig. 3b). To identify the Pokemon-binding site essential for Pokemon's repressive activity upon the p19Arf promoter, we mutagenized the three putative binding sites in reporter constructs and performed luciferase assays (see Supplementary Fig. 3c). As shown in Fig. 2d, Pokemon was unable to repress mut2 reporter activity, which strongly indicates that the Pokemon-binding site located 50 base pairs upstream from the transcription start site is indispensable for p19Arf repression. Furthermore, we found that Pokemon efficiently abrogated E2F1-dependent ARF transactivation/de-repression in a dose-dependent manner (Supplementary Fig. 3d, and not shown). We next tested whether Pokemon binds the p19Arf promoter in vivo by chromatin immunoprecipitation (ChIP) assay and designed two different primer sets to amplify the p19Arf promoter region (Supplementary Fig. 3a). In these assays, the anti-Pokemon antibody specifically precipitated p19Arf promoter sequences from the extract prepared from WT MEFs but not from Zbtb7-/- MEFs (Fig. 2e). Furthermore, anti-Pokemon antibody was unable to precipitate -actin genomic sequences, which indicates that Pokemon binding might be specific for the p19Arf promoter. ChIP assays were also performed in transformed MEFs, confirming that Pokemon also binds to the p19Arf promoter under transformed conditions (not shown). Thus, Pokemon can bind to the p19Arf promoter in vivo and repress its activity.

The effect of Zbtb7 inactivation on p19Arf expression was then evaluated. p19Arf protein levels were remarkably higher in Zbtb7-/- MEFs starting from passage four, when p19Arf is normally induced due to culture shock14 (Fig. 2f). p19Arf upregulation also occurred at the mRNA level as indicated by real-time polymerase chain reaction (PCR) analysis (Supplementary Fig. 3e). The Ink4a-Arf locus encodes two tumour suppressors, p16Ink4a and p19Arf, which are homologous to human p16INK4A and p14ARF, respectively31. Both genes are able to induce cellular senescence in response to oncogenic stimuli, and their functional loss is frequently observed in human cancers32, 33. Interestingly, in Zbtb7-/- MEFs, only p19Arf but not p16Ink4a was constitutively upregulated, whereas gradual increases in p16Ink4a levels with passaging occurred comparably in Zbtb7-/- and WT MEFs (Fig. 2g, and not shown). Furthermore, Pokemon was unable to repress the mouse p16Ink4a promoter activity as revealed by luciferase assay (Supplementary Fig. 3f), which also supports the notion that Pokemon is a specific repressor for p19Arf but not for p16Ink4a. In complete agreement with these findings, p53 (a key downstream target gene of p19Arf) was also found overexpressed in Zbtb7-/- MEFs (Supplementary Fig. 3g).

We then examined whether Pokemon overexpression would repress p19Arf expression levels and whether this would also occur in transforming conditions when p19Arf is induced in primary cells. Indeed, Pokemon overexpression invariably resulted in decreased p19Arf levels under these conditions (Fig. 2h). p19Arf levels were also examined in protein extracts obtained from transformed colonies in these oncogene combinations. p19Arf and Pokemon protein expression levels were inversely correlated, whereas p16Ink4a expression was not affected (not shown). These data therefore demonstrate that Pokemon is an important and selective repressor of p19Arf expression.

Loss of Arf rescues Zbtb7-/- MEF transformability
Accumulation of p19Arf correlates with the onset of cellular senescence in MEFs after culture shock14. In full agreement with the finding that Pokemon represses p19Arf expression and that p19Arf upregulation is observed in Zbtb7-/- MEFs from about passage four, Zbtb7-/- MEFs proliferated normally until this passage number and their proliferative capacity rapidly declined at later passages (Fig. 3a). This was accompanied by a marked increase in cellular senescence as demonstrated by staining for the senescence marker -galactosidase34 (Fig. 3b). As expected, these cell growth defects were fully rescued by Pokemon add-back. In addition, Pokemon overexpression further enhanced the proliferative potential of WT MEFs (Fig. 3c).

Figure 3 p19Arf loss reverts premature senescence and refractoriness to oncogenic transformation in Zbtb7-/- MEFs. Full legend

High resolution image and legend (51k)

We next tested whether p19Arf loss could rescue the proliferative defect and premature senescence phenotype observed in Zbtb7-/- MEFs. To this end, Zbtb7 and p19Arf double-knockout MEFs (Zbtb7-/-p19Arf-/- MEFs) were established from double-mutant embryos. Indeed, the growth defects in Zbtb7-/- MEFs were fully reverted by p19Arf loss, and Zbtb7-/-p19Arf-/- MEFs proliferated indistinguishably from p19Arf-/- MEFs (Fig. 3d). Furthermore, p19Arf inactivation also reverted the refractoriness to oncogenic transformability observed in Zbtb7-/- MEFs. Zbtb7-/-p19Arf-/- MEFs were susceptible to transformation by oncogenic ras alone (Fig. 3e).

Pokemon is oncogenic in vivo
To determine whether Pokemon would act as a genuine proto-oncogene in vivo, we generated a transgenic mouse model in which Pokemon is overexpressed in immature T and B lymphoid lineage cells taking advantage of an lckEµ enhancer/promoter transgenic construct35 (Fig. 4a). Positive transgenic founders were identified and the expression levels of the transgene in F1 mice from the various founders were assessed by semiquantitative polymerase chain reaction with reverse transcription (RT–PCR) from splenocyte RNA (Supplementary Fig. 4a). We selected two of the highest-expressing lines (numbers 16 and 26) and the F1 and F2 generations of each line (litters obtained by intercross with negative mice) were used for further study.
Figure 4 Pokemon transgenic mice develop pre-T LBL. Full legend

High resolution image and legend (43k)

Strikingly, both transgenic lines developed aggressive tumours. The hallmark of the disease in these lckEµ-Pokemon mice was massive thymic enlargement, accompanied by lymphadenopathy, splenomegaly, hepatomegaly and tumour infiltration into bone marrow (Fig. 4b). Histopathological examinations of the tumours revealed that the thymus and lymph node architecture was completely effaced and showed a starry-sky pattern with tingible-body macrophages. The lymphomas contained a monotonous population of immature lymphoid cells. The spleen showed expansion of the white pulp and effacement and infiltration of the red pulp by tumour cells. The liver showed prominent periportal, lobular and sinusoidal infiltration by tumour cells (Fig. 4c).

To further characterize the phenotype of these lymphomas, we performed flow cytometry analysis. The lymphomas typically showed an immature T-cell immunophenotype (CD3+, CD4++, CD8++ CD44++, CD25-; Fig. 4d). Furthermore, Wright–Giemsa staining of the bone marrow tumour cells showed large lymphoblasts with round nuclei, fine chromatin and scanty cytoplasm, which is fully consistent with a mouse precursor T-cell lymphoblastic lymphoma/leukaemia (pre-T LBL)36 (Fig. 4d, bottom right). Taking advantage of a specific anti-Pokemon monoclonal antibody 13E9 (Supplementary Fig. 4b), immunohistochemistry of lymphoma sections showed intense nuclear positivity for Pokemon (Supplementary Fig. 4c). However, in normal thymus, Pokemon is expressed mainly in medullary epithelial cells and Hassle's corpuscles, which are negative for the T-cell linage marker CD3 and positive for the epithelial cell marker cytokeratin AE1:AE3 (Supplementary Fig. 5a, d). In addition, lymphomas were highly positive for terminal deoxynucleotidyl transferase, a diagnostic marker for lymphoblastic lymphoma/leukaemia (Supplementary Fig. 4c)37, 38. The increased Pokemon protein expression in thymic lymphomas was also confirmed by western blot analysis (Supplementary Fig. 4d). To examine whether the lymphomas were monoclonal in origin, we extracted DNA from six different thymic lymphomas and analysed D1-to-J1 rearrangement of the T-cell antigen receptor- locus. Four thymic DNA samples from WT mice were used as controls. As shown in Fig. 4e, either a single non-rearranged (or rearranged in other D and J elements) band or a predominant single rearranged band was amplified in lymphoma DNA, whereas PCR products from the four WT thymic DNA demonstrated five possible rearranged and one germline non-rearranged band, as described previously39. These data indicate that the lckEµ-Pokemon lymphomas were monoclonal. These lymphomas were fatal. Specifically, 16 of 27 transgene-positive mice from the number 26 line (the line expressing Pokemon most) developed tumours, and these mice either died or became terminally ill (and hence were killed) between 9 and 40 weeks of age (median 14 weeks) (Fig. 4f). Transgenic mice from the number 16 line also gave rise to pre-T LBL, but at a lower penetrance (4 of 18 mice).

Aberrant Pokemon expression in human cancer
As Pokemon expression proved essential for cellular transformation and this transcription factor is oncogenic when overexpressed both in vitro and in vivo, we analysed the levels of Pokemon expression in human cancers performing tissue microarray analysis (TMA)40 with an anti-Pokemon monoclonal antibody 13E9. We first analysed human T-cell and B-cell lymphomas because lckEµ-Pokemon mice developed T-cell lymphomas and Pokemon is normally coexpressed with BCL6 in the B-cell within the germinal centre (Supplementary Fig. 5b, c, and not shown), and might therefore cooperate with BCL6 in lymphomagenesis. We initially performed TMA on small cohorts of T-cell and B-cell lymphoma cases (Supplementary Fig. 6a, b). Interestingly, Pokemon is highly expressed in a subset of T-cell lymphomas, although Pokemon was barely detectable in normal human thymic CD3-positive cells. Furthermore, Pokemon is strongly expressed in diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma (FL), in whose pathogenesis BCL6 is known to be involved.

We subsequently analysed and scored 130 cases of DLBCL and 290 cases of FL for Pokemon protein expression levels (for patient's clinical and immunohistochemical features, see Supplementary Tables 1 and 2 and Supplementary Fig. 6c). Pokemon was highly expressed in 25–35% and moderately expressed in 40–50% of both DLBCL and FL clinical samples (Fig. 5a).

Figure 5 Cooperative roles of Pokemon and BCL6 in lymphomagenesis. Full legend

High resolution image and legend (74k)

With the use of DNA microarray and TMA analysis, it has recently been proposed that DLBCL can be divided into prognostically significant subgroups with a germinal-centre B-cell-like (GCB) and an activated B-cell-like phenotype41-43. We therefore scored and analysed the expression of GCB markers in our DLBCL cohort and correlated these with Pokemon expression levels. Pokemon expression was not significantly correlated with the GCB phenotype, although Pokemon is expressed in the germinal centre in normal reactive tonsil (Supplementary Table1).

Interestingly, the vast majority of BCL6-positive DLBCL and FL were found to be Pokemon positive (Supplementary Fig. 6d). In addition, lymphomas positive for both proteins displayed a higher proliferative index than single-negative or double-negative tumours, as revealed by Ki-67 staining, indicating a possible functional cooperation between Pokemon and BCL6 in tumorigenesis (Fig. 5b). BCL6 can immortalize primary cells and fully transform them in cooperation with oncogenic ras44. Pokemon proved essential for BCL6 immortalizing and transforming activities in MEFs (Fig. 5c, d). Pokemon overexpression not only rescued these defects in Zbtb7-/- cells, but also markedly potentiated the growth-promoting activity of BCL6 (Fig. 5c). This cooperative crosstalk between Pokemon and BCL6 in tumorigenesis indicated, in turn, that Pokemon expression and/or Pokemon/BCL6 coexpression might identify a specific lymphoma subtype and possibly predict its clinical outcome. Strikingly, Pokemon positivity was predictive of better overall survival in DLBCL (Fig. 5e, left). Furthermore, Pokemon and BCL6 double positivity predicted an even better clinical outcome for overall survival (Fig. 5e, right).

Discussion
We have demonstrated that Pokemon expression levels are critical determinants of the cellular response to oncogenic transformation. We show that the function of this novel member of the POK family of transcriptional repressors is necessary for oncogenic transformation whereas, conversely, Pokemon can act as a genuine proto-oncogene when overexpressed (Fig. 5f). Importantly, taking advantage of a transgenic mouse model, we demonstrate that Pokemon can indeed act as a proto-oncogene in vivo.

We identify Pokemon as a potent transcriptional repressor of the p19Arf tumour suppressor gene. Pokemon is, to our knowledge, the first ARF-specific transcriptional repressor to be identified. Furthermore, p19Arf inactivation rescued the refractoriness to oncogenic transformation of Zbtb7-/- cells. Taken together, these data strongly indicate that, at least in certain cell types, one of the main oncogenic functions of Pokemon consists of its ability to repress p19Arf expression. Interestingly, real-time RT–PCR analysis of a cohort of DLBCL cases (n = 37) revealed that high Pokemon gene expression is generally correlated with low expression of the p14ARF gene (Supplementary Fig. 6f). Our findings underscore the potential importance of ARF suppression in human tumorigenesis.

We show that Pokemon is expressed at very high levels in a subset of human lymphomas. TMA analyses in breast, lung, colon, prostate and bladder carcinomas also reveal high levels of Pokemon expression in a subset of these tumours (T.M. and P.P.P., unpublished data). In this respect, it is worth noting that the genomic region where the ZBTB7 gene resides (chromosome 19p13.3) is a hotspot for chromosomal translocations (CGAP; http://cgap.nci.nih.gov/Chromosomes). We show that Pokemon expression or its coexpression with BCL6 in lymphoma predicts clinical outcome. This is in keeping with the fact that Pokemon is required for BCL6 oncogenic activity, and that high expression of potent proto-oncogenes such as BCL6 and LMO2 also identifies cohorts of patients who respond better to current treatment modalities41, 43, 45. Cooperative tumorigenesis between these two proto-oncogenes could be exerted through distinct transcriptional programmes or pathways46, but also through co-regulation of common target genes in view of their ability to interact physically. In this respect, it is of interest that BCL6 has recently been reported to repress p53 expression directly47.

Our findings provide a further rational for transcription-based therapeutic modalities and identify Pokemon as an important target for therapy on the basis of its key role in oncogenesis.

Methods
Growth curves, transformation and senescence assays MEFs were prepared from embryos at 13.5 days post coitus. MEFs at passage one or two were used for all experiments. Growth curves after oncogene infections were generated by seeding 3,000 cells (per well) in 12-well plates in triplicate48. For transformation assays, 5 104 cells were resuspended in 1.5 ml complete Dulbecco's modified Eagle's medium containing 0.3% agarose and seeded into 35-mm plates containing a 3-ml layer of solidified 0.6% agar in complete medium. Foci were scored two or three weeks later. For growth curves of unmanipulated MEFs, 9 105 cells were plated in 10-cm dishes and passaged every fourth day. For senescence assays, -galactosidase-positive cells were scored at passage six in accordance with the manufacturer's specification (Senescence detection kit; Oncogene). The Rat1 MycErTM cells were kindly provided by G. I. Evan. The mock-infected (pWZL-Hygro) cells and the Pokemon-infected (pWZL-Flag-Pokemon) cells were generated by us. After 24 h of serum deprivation, Myc-induced apoptosis was subsequently triggered by the addition of 100 nM 4-hydroxytamoxifen (Sigma) in medium containing 0.1% serum16, and apoptotic cells were scored with the trypan blue exclusion method on the indicated days.

Molecular analysis Western blotting was performed in accordance with the standard protocol with the following antibodies: E1A (M58; NeoMarkers), T-Ag (SV40 Ab-1; NeoMarkers), Myc (no. 06-340; Upstate), Ras (Ab-3; Oncogene), Xpress (Invitrogen), p19Arf (Ab-1 PC435 lot no. D18714-1; Oncogene), p16Ink4a (M-156; Santa Cruz), p53 (Ab-7; Oncogene) and BCL6 (N-3; Santa Cruz). Anti-Pokemon hamster monoclonal antibody (clone 13E9) was raised against a peptide derived from the Pokemon N-terminal region (amino acids 1–20, MAGGVDGPIGIPFPDHSSDI), fully conserved between human and mouse proteins. Loading was assessed by -actin (A-5316; Sigma), -tubulin (B-5-1-2; Sigma) or Hsp-90 (no. 610419; BD). CAST analysis and EMSA were performed as described18. For reporter assays, 105 NIH 3T3 cells per well were plated into 12-well plates in triplicate 16 h before transfection, and cells were transfected with Lipofectamine 2000. At 24 h after transfection, cells were assayed with a dual luciferase assay kit (Promega).

Plasmids Mouse Zbtb7 complementary DNA was amplified by PCR from C57BL/6 mouse bone marrow total cDNA and subcloned into the pcDNA3.1His vector (containing the Xpress tag; Invitrogen), pSG5 vector (with Flag sequence; Stratagene), pWZL-Hygro and MSCV-PIG vectors (Puro-IRES-GFP, provided by S. W. Lowe). Pokemon deletion mutant constructs used for reporter assays (pcDNA3.1His-POZ, POZ, Zn and Zn) were generated from the pcDNA3.1His-Pokemon vector. Retroviral vectors, expression vectors and luciferase constructs were gifts from A. Koff (pWZL-Hygro-human-c-Myc and pSG5-T-Ag), R. Bernards (pBabe-Hygro-BCL6), S. W. Lowe (pBabe-puro-H-rasV12 and pWZL-Hygro-E1A), W. Kaelin (pRc-CMV-HA-E2F1); S. W. Hiebert (human p14ARF-luciferase reporter plasmid), M. F. Roussel (mouse p19Arf-luciferase reporter plasmid), B. A. Mock (mouse p16Ink4a-luciferase reporter plasmid) and R. M. Perlmutter (p1026X vector). Mutated p19Arf-luciferase reporters were constructed by using a QuikChange Mutagenesis Kit (Stratagene). pWZL-Hygro-T-Ag was generated from the pSG5-T-Ag vector. pWZL-Hygro-H-rasV12 was generated from the pBabe-puro-H-rasV12 vector. 3POKBB-Luc was constructed by insertion of three copies of the Pokemon-binding sequence (5'-GGTTAAAAGACCCCTCCCCGAATTCGGATC-3') into a luciferase reporter vector TK-LUC from R. Dalla-Favera.

Generation of Zbtb7-/-p19Arf-/- mutants Zbtb7+/- mice and p19Arf+/- mice (provided by C. J. Sherr) were crossed to generate Zbtb7+/-p19Arf+/- double-heterozygous mice, which were subsequently intercrossed to generate Zbtb7-/-p19Arf-/- embryos and MEFs along with WT and single-mutant littermates.

Generation and characterization of lckEµ-Pokemon mice The Flag-Pokemon transgene was produced by ligating the 1.7-kilobase Flag-Pokemon cDNA sequence (from pSG5 Flag-Pokemon vector) into the BamH1 cloning site of the p1026X vector35. For generating transgenic mice, a purified Spe1 fragment containing lckEµ-Pokemon was injected into fertilized (C57BL/6J CBA/J) F2 eggs at the MSKCC Transgenic Mouse Core Facility. Transgenic founders were detected by Southern blotting of tail DNA with a Flag-Pokemon cDNA fragment (generated from pSG5 Flag-Pokemon) as a probe. Positive founders were intercrossed with negative founders to establish F1 generations, and the colony of two highest copy founder lines (nos 16 and 26, revealed by Southern blotting and RT–PCR) were expanded and used for further characterization. To examine the monoclonality of thymic lymphomas, DNA was extracted from WT thymus and thymic lymphomas and the D-to-J rearrangements of the T-cell antigen receptor- locus were analysed as described previously, using primers 1 and 4 (ref. 39).

TMA analysis The study cohort comprised 130 DLBCL and 290 FL biopsies consecutively ascertained at the Memorial Sloan-Kettering Cancer Center (MSKCC) between 1984 and 2000. Patient anonymity was ensured and the study received a waiver by the Institutional Review Board. All DLBCL biopsies were obtained at their first evaluation at MSKCC and all DLBCL patients received an anthracyclin-based chemotherapy regimen. The initial histological diagnosis was based on haematoxylin–eosin staining and immunophenotyping results. Biopsies were reviewed and reclassified histologically in accordance with the World Health Organization classification38. TMAs were constructed and stained in the Pathology and Molecular Cytology Core Facilities as described previously49. Further detailed information, including staining conditions and antibody dilutions, is described in Supplementary Methods.

Supplementary information accompanies this paper.

Received 17 September 2004;accepted 18 November 2004

References 1. Lin, R. J. et al. Role of the histone deacetylase complex in acute promyelocytic leukaemia. Nature 391, 811−814 (1998)
2. Melnick, A. et al. Critical residues within the BTB domain of PLZF and Bcl-6 modulate interaction with corepressors. Mol. Cell. Biol. 22, 1804−1818 (2002)
3. Barna, M., Hawe, N., Niswander, L. & Pandolfi, P. P. Plzf regulates limb and axial skeletal patterning. Nature Genet. 25, 166−172 (2000)
4. Ye, B. H. et al. The BCL-6 proto-oncogene controls germinal-centre formation and Th2-type inflammation. Nature Genet. 16, 161−170 (1997)
5. Adhikary, S. et al. Miz1 is required for early embryonic development during gastrulation. Mol. Cell. Biol. 23, 7648−7657 (2003)
6. Carter, M. G. et al. Mice deficient in the candidate tumor suppressor gene Hic1 exhibit developmental defects of structures affected in the Miller−Dieker syndrome. Hum. Mol. Genet. 9, 413−419 (2000)
7. Chen, W. Y. et al. Heterozygous disruption of Hic1 predisposes mice to a gender-dependent spectrum of malignant tumors. Nature Genet. 33, 197−202 (2003)
8. Chen, Z. et al. Fusion between a novel Kruppel-like zinc finger gene and the retinoic acid receptor-alpha locus due to a variant t(11;17) translocation associated with acute promyelocytic leukaemia. EMBO J. 12, 1161−1167 (1993)
9. Ye, B. H. et al. Alterations of a zinc finger-encoding gene, BCL-6, in diffuse large-cell lymphoma. Science 262, 747−750 (1993)
10. Davies, J. M. et al. Novel BTB/POZ domain zinc-finger protein, LRF, is a potential target of the LAZ-3/BCL-6 oncogene. Oncogene 18, 365−375 (1999)
11. Kukita, A. et al. Osteoclast-derived zinc finger (OCZF) protein with POZ domain, a possible transcriptional repressor, is involved in osteoclastogenesis. Blood 94, 1987−1997 (1999)
12. Pessler, F., Pendergrast, P. S. & Hernandez, N. Purification and characterization of FBI-1, a cellular factor that binds to the human immunodeficiency virus type 1 inducer of short transcripts. Mol. Cell. Biol. 17, 3786−3798 (1997)
13. Weinberg, R. A. The cat and mouse games that genes, viruses, and cells play. Cell 88, 573−575 (1997)
14. Zindy, F. et al. Myc signaling via the ARF tumor suppressor regulates p53-dependent apoptosis and immortalization. Genes Dev. 12, 2424−2433 (1998)
15. Serrano, M., Lin, A. W., McCurrach, M. E., Beach, D. & Lowe, S. W. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell 88, 593−602 (1997)
16. Evan, G. I. et al. Induction of apoptosis in fibroblasts by c-myc protein. Cell 69, 119−128 (1992)
17. de Stanchina, E. et al. E1A signaling to p53 involves the p19(ARF) tumor suppressor. Genes Dev. 12,2434−2442 (1998)
18. Wright, W. E., Binder, M. & Funk, W. Cyclic amplification and selection of targets (CASTing) for the myogenin consensus binding site. Mol. Cell. Biol. 11, 4104−4110 (1991)
19. Bates, S. et al. p14ARF links the tumour suppressors RB and p53. Nature 395, 124−125 (1998)
20. Rowland, B. D. et al. E2F transcriptional repressor complexes are critical downstream targets of p19(ARF)/p53-induced proliferative arrest. Cancer Cell 2, 55−65 (2002)
21. Jacobs, J. J. L., Kieboom, K., Marino, S., DePinho, R. A. & van Lohuizen, M. The oncogene and Polycomb-group gene bmi-1 regulates cell proliferation and senescence through the ink4a locus. Nature 397, 164−168 (1999)
22. Jacobs, J. J. et al. Bmi-1 collaborates with c-Myc in tumorigenesis by inhibiting c-Myc-induced apoptosis via INK4a/ARF. Genes Dev. 13, 2678−2690 (1999)
23. Smith, K. S. et al. Bmi-1 regulation of INK4A-ARF is a downstream requirement for transformation of hematopoietic progenitors by E2a-Pbx1. Mol. Cell 12, 393−400 (2003)
24. Kranc, K. R. et al. Transcriptional coactivator Cited2 induces Bmi1 and Mel18 and controls fibroblast proliferation via Ink4a/ARF. Mol. Cell. Biol. 23, 7658−7666 (2003)
25. Kim, J. H. et al. The Bmi-1 oncoprotein is overexpressed in human colorectal cancer and correlates with the reduced p16INK4a/p14ARF proteins. Cancer Lett. 203, 217−224 (2004)
26. Vonlanthen, S. et al. The bmi-1 oncoprotein is differentially expressed in non-small cell lung cancer and correlates with INK4A-ARF locus expression. Br. J. Cancer 84, 1372−1376 (2001)
27. Jacobs, J. J. et al. Senescence bypass screen identifies TBX2, which represses Cdkn2a (p19(ARF)) and is amplified in a subset of human breast cancers. Nature Genet. 26, 291−299 (2000)
28. Lingbeek, M. E., Jacobs, J. J. & van Lohuizen, M. The T-box repressors TBX2 and TBX3 specifically regulate the tumor suppressor gene p14ARF via a variant T-site in the initiator. J. Biol. Chem. 277, 26120−26127 (2002)
29. Brummelkamp, T. R. et al. TBX-3, the gene mutated in Ulnar-Mammary Syndrome, is a negative regulator of p19ARF and inhibits senescence. J. Biol. Chem. 277, 6567−6572 (2002)
30. Inoue, K., Roussel, M. F. & Sherr, C. J. Induction of ARF tumor suppressor gene expression and cell cycle arrest by transcription factor DMP1. Proc. Natl Acad. Sci. USA 96, 3993−3998 (1999)
31. Quelle, D. E., Zindy, F., Ashmun, R. A. & Sherr, C. J. Alternative reading frames of the INK4a tumor suppressor gene encode two unrelated proteins capable of inducing cell cycle arrest. Cell 83, 993−1000 (1995)
32. Sherr, C. J. & McCormick, F. The RB and p53 pathways in cancer. Cancer Cell 2, 103−112 (2002)
33. Lowe, S. W. & Sherr, C. J. Tumor suppression by Ink4a-Arf: progress and puzzles. Curr. Opin. Genet. Dev. 13, 77−83 (2003)
34. Dimri, G. P. et al. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. Proc. Natl Acad. Sci. USA 92, 9363−9367 (1995)
35. Iritani, B. M., Forbush, K. A., Farrar, M. A. & Perlmutter, R. M. Control of B cell development by Ras-mediated activation of Raf. EMBO J. 16, 7019−7031 (1997)
36. Morse, H. C. III et al. Bethesda proposals for classification of lymphoid neoplasms in mice. Blood 100, 246−258 (2002)
37. Donlon, J. A., Jaffe, E. S. & Braylan, R. C. Terminal deoxynucleotidyl transferase activity in malignant lymphomas. N. Engl. J. Med. 297, 461−464 (1977)
38. Jaffe, E. S., Harris, N. L., Stein, H. & Vardiman, J. W. Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues (IARC Press, Lyon, France, 2001)
39. Whitehurst, C. E., Chattopadhyay, S. & Chen, J. Control of V(D)J recombinational accessibility of the D beta 1 gene segment at the TCR beta locus by a germline promoter. Immunity 10, 313−322 (1999)
40. Gurrieri, C. et al. Loss of the tumor suppressor PML in human cancers of multiple histologic origins. J. Natl Cancer Inst. 96, 269−279 (2004)
41. Alizadeh, A. A. et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403, 503−511 (2000)
42. Rosenwald, A. et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. N. Engl. J. Med. 346, 1937−1947 (2002)
43. Hans, C. P. et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood 103, 275−282 (2004)
44. Shvarts, A. et al. A senescence rescue screen identifies BCL6 as an inhibitor of anti-proliferative p19ARF-p53 signaling. Genes Dev. 16, 681−686 (2002)
45. Lossos, I. S. et al. Prediction of survival in diffuse large-B-cell lymphoma based on the expression of six genes. N. Engl. J. Med. 350, 1828−1837 (2004)
46. Sanchez-Beato, M., Sanchez-Aguilera, A. & Piris, M. A. Cell cycle deregulation in B-cell lymphomas. Blood 101, 1220−1235 (2003)
47. Phan, R. T. & Dalla-Favera, R. The BCL6 proto-oncogene suppresses p53 expression in germinal-center B cells. Nature 432, 635−639 (2004)
48. Carnero, A., Hudson, J. D., Price, C. M. & Beach, D. H. p16INK4A and p19ARF act in overlapping pathways in cellular immortalization. Nature Cell Biol. 2, 148−155 (2000)
49. Hedvat, C. V. et al. Application of tissue microarray technology to the study of non-Hodgkin's and Hodgkin's lymphoma. Hum. Pathol. 33, 968−974 (2002)
Acknowledgements. We thank D. Yao, C. Hedvat, M. Dudas, J. Qin and A. Wilton for assistance on TMA preparation, staining and statistical analyses; S. Hasan for assistance with data input and management; K. Manova, C. Farrell and other Molecular Cytology Core Facility members for advice and assistance with IHC; J. Overholser and other Monoclonal Antibody Core Facility staff for help with antibody generation; G. Cattoretti and R. Dalla-Favera for advice; and L. Khandker, L. Dong, M. Hu, L. DiSantis and other P.P.P. laboratory members for assistance and discussion. This work is supported in part by an NCI grant to P.P.P.