Functional Heterogeneity of PAX5 Chimeras Reveals Insight for Leukemia Development

Published OnlineFirst January 16, 2014; DOI: 10.1158/1541-7786.MCR-13-0337

Molecular Cancer Oncogenes and Tumor Suppressors Research

Klaus Fortschegger, Stefanie Anderl, Dagmar Denk, and Sabine Strehl

Abstract PAX5, a transcription factor pivotal for B-cell commitment and maintenance, is one of the most frequent targets of somatic mutations in B-cell precursor acute lymphoblastic leukemia. A number of PAX5 rearrangements result in the expression of in-frame fusion genes encoding chimeric proteins, which at the N-terminus consistently retain the PAX5 DNA-binding paired domain fused to the C-terminal domains of a markedly heterogeneous group of fusion partners. PAX5 fusion proteins are thought to function as aberrant transcription factors, which antagonize wild-type PAX5 activity. To gain mechanistic insight into the role of PAX5 fusion proteins in leukemogenesis, the biochemical and functional properties of uncharacterized fusions: PAX5–DACH1, PAX5–DACH2, PAX5– ETV6, PAX5–HIPK1, and PAX5–POM121 were ascertained. Independent of the subcellular distribution of the wild-type partner proteins, ectopic expression of all PAX5 fusion proteins showed a predominant nuclear localization, and by chromatin immunoprecipitation all of the chimeric proteins exhibited binding to endogenous PAX5 target sequences. Furthermore, consistent with the presence of potential oligomerization motifs provided by the partner proteins, the self-interaction capability of several fusion proteins was conﬁrmed. Remarkably, a subset of the PAX5 fusion proteins conferred CD79A promoter activity; however, in contrast with wild-type PAX5, the fusion proteins were unable to induce Cd79a transcription in a murine plasmacytoma cell line. These data show that leukemia-associated PAX5 fusion proteins share some dominating characteristics such as nuclear localization and DNA binding but also show distinctive features. Implications: This comparative study of multiple PAX5 fusion proteins demonstrates both common and unique properties, which likely dictate their function and impact on leukemia development. Mol Cancer Res; 12(4); 1– 12. 2014 AACR.

Introduction PAX5 alterations, including deletions, point mutations, fi PAX5 is a paired box family transcription factor, which is and ampli cations, occur in about 30% of B-cell precursor acute lymphoblastic leukemia cases, and chromosomal pivotal for B-cell development (1). Within the hematopoi- – – etic system PAX5 is exclusively expressed in the B-lymphoid rearrangements account for 2 3% (7 9). Most rearrange- lineage, where it is not only essential for pro-B-cell com- ments lead to the expression of in-frame fusion transcripts, mitment but also for the maintenance of B-cell identity until which encode chimeric proteins consisting of the N- the onset of plasma cell differentiation (1–3). PAX5 fulfills a terminal PAX5 portion and the C-terminal regions of a dual role by activating B-cell specific and concomitantly variety of markedly heterogeneous fusion partners ranging – from transcription factors over kinases to structural pro- repressing lineage-inappropriate genes (3 5), which is – achieved by the recruitment of distinct chromatin remodel- teins (8 16). ing, histone modifying, and basal transcription factor com- As a unifying feature, the DNA-binding paired domain is plexes to PAX5 target loci (6). always retained in the PAX5 fusion proteins, and for some of them DNA interaction has been demonstrated, suggesting Authors' Affiliation: Children's Cancer Research Institute (CCRI), St. Anna that they may act as trans–dominant-negative transcription Kinderkrebsforschung e.V., Vienna, Austria factors antagonizing wild-type PAX5 activity (8, 10, 17, 18). Note: Supplementary data for this article are available at Molecular Cancer This notion has been corroborated by reporter gene assays, in Research Online (http://mcr.aacrjournals.org/). which several PAX5 fusions competitively inhibited PAX5- – Klaus Fortschegger and Stefanie Anderl contributed equally to this work. mediated luciferase induction (8, 10, 12, 13, 17 19). Accordingly, ectopic expression of PAX5 fusion proteins Corresponding Author: Sabine Strehl, Children's Cancer Research Insti- tute (CCRI), St. Anna Kinderkrebsforschung e.V. Zimmermannplatz 10, led to altered levels of certain PAX5 target genes (8, 18, 20), 1090 Vienna, Austria. Phone: 43-1-40470-4020; Fax: 43-1-40470-7150; however, their majority remained unaffected (10, 13). Some E-mail: [email protected] PAX5 fusions have also been shown to self-interact, which doi: 10.1158/1541-7786.MCR-13-0337 has been suggested to influence their binding affinities and 2014 American Association for Cancer Research. repression capabilities (17, 18).

www.aacrjournals.org OF1

Fortschegger et al.

Based on our previous work, in which we identified several the moloney murine leukemia virus ecotropic receptor (pro- novel PAX5 fusion partners (9, 15), the main aim of this vided by H. Strobl). First, Phoenix-GP cells were cotrans- study was to determine their functional features. We inves- fected with pBMN-eco receptor-IRES-mCD8a and gibbon tigated key properties of the so far uncharacterized fusion ape leukemia virus envelope constructs (26). NALM-6 cells proteins PAX5–DACH1, PAX5–DACH2, PAX5–HIPK1, were transduced thrice with the amphotropic virus using and PAX5–POM121 and compared them with PAX5– RetroNectin (Takara Bio) and centrifugation. After enrich- ETV6. Our data substantiate that PAX5 chimeras are ment by mCD8ahigh fluorescence-activated cell sorting, these nuclear proteins capable of binding to PAX5 target NALM-6-EcoR cells were transduced 3 times, and 558LmM sequences. Moreover, our findings confirm that self-inter- cells were transduced once with retroviruses from MSCV- action is a characteristic feature of a subset of PAX5 fusion MigR1 transfected Phoenix-E. 558LmM cells were analyzed proteins. However, in contrast to earlier studies, but in line on a LSR Fortessa flow cytometer (Becton Dickinson) using with recently described upregulation of certain PAX5 phycoerythrin-labeled anti-mouse IgM antibody (27). Before repressed targets by the PAX5–ETV6 fusion (21), we pro- further analyses, transduced NALM-6-EcoR and 558LmM vide evidence that some PAX5 fusion proteins may also cells were flow sorted for GFP positivity. activate target genes, arguing against a simplified trans– dominant-negative mode of action. Confocal microscopy Transduced NALM-6-EcoR cells were immobilized by Materials and Methods cytospin centrifugation (Thermo Scientific Shandon II) and Vectors and cloning fixed on Histobond slides (Marienfeld). Cells were permea- The coding sequences of PAX5, a DNA-binding–defi- bilized and incubated with HA primary (ab18181) and Alexa – Fluor 568 secondary antibodies. Nuclei were counterstained cient mutant (K67A/K87A/K89A), PAX5 DACH1, 0 PAX5–DACH2, coiled-coil (CC) region deleted mutants with 4 ,6-diamidino-2-phenylindole (DAPI) and mounted thereof, PAX5–ETV6, ETV6, PAX5–HIPK1, and PAX5– in Mowiol 4-88 (Sigma-Aldrich). For BiFC, HeLa cells were POM121 were cloned into pcDNA3 (Invitrogen), including transiently cotransfected with vectors containing the coiled- N-terminal myc- or tandem hemagglutinin (HA) tags. HA- coils of DACH1 or DACH2 fused to N- and C-terminal tagged coding sequences were further cloned into MSCV- yellow fluorescent protein (YFP) halves. The leucine zipper MigR1 (22), which contains an internal ribosome entry site of Jun (bJUN) along with wild-type or mutant Fos (bFOS or followed by green fluorescent protein (GFP) to allow for cell FOSDZIP) were used as positive and negative controls, sorting. For bimolecular fluorescent complementation respectively (23). Cells were fixed, DAPI stained and (BiFC), the coiled-coils of DACH1 and DACH2 were mounted in Mowiol. Confocal microscopy was performed cloned into pBiFC-YN155 and pBiFC-YC155 (23). on a TCS-SP5 (Leica Microsystems) equipped with an CD19-luc-1x containing one PAX5-binding site was derived HCX-PL-APO-CS-63.0 1.40-OIL objective. from luc-CD19 (24). The human CD79A promoter was cloned into pGL4.10 (Promega). As negative controls the Electrophoretic mobility shift assay PAX5-binding sites of the reporter vectors were mutated Nuclear extracts were prepared from transfected HEK293 employing QuikChange site-directed mutagenesis (Agilent). cells and tested for expression of the tagged proteins by Similarly, the coiled-coil domains of PAX5–DACH1 and immunoblotting. Assays with fluorescently labeled probes PAX5–DACH2 were deleted from the original fusion gene were performed using the Odyssey Infrared EMSA Kit (LI- constructs. The Renilla pRL-null vector (Promega) was used COR). In brief, nuclear extracts were incubated with 20 ng for normalization. PAX5–DACH2, DACH1 (25), MSCV- CD19-DY782 (28) or CD79A-DY682 (29) probe in bind- MigR1 and luc-CD19, pBiFC-YN155 and pBiFC-YC155 ing buffer. For probe competition experiments, 200 ng or vectors were kindly provided by C. Broccardo, R. Pestell, M. 1 mg of unlabeled wild-type or mutant oligonucleotides, and Busslinger, and T. Kerppola, respectively. Oligonucleotide for supershift assays 1 mg of mouse HA or myc antibody were and protein sequences are provided in the Supplementary added. Samples were separated on native 5% polyacrylamide Material. Tris/glycine gels, which were scanned on an Odyssey Infra- red scanner (LI-COR) at 800 or 700 nm. Cell lines and culture, transient transfection, and Competition assays were performed with nuclear extracts retroviral transduction and 20 fmol biotinylated CD19 probe in binding buffer. NALM-6 cells, purchased from DMSZ, and 558LmM, Samples were run on native 5% polyacrylamide gels in 0.5 provided by M. Reth, were cultured in RPMI 1640 with Tris-borate EDTA buffer and blotted to nylon membranes GlutaMAX (Invitrogen), 10% FBS and 1 penicillin/strep- (Amersham Hybond Nþ). Signals were detected using tomycin (PAA), and HEK293, HeLa, Phoenix-GP, and -E the Chemiluminescent Nucleic Acid Detection Module cells (both provided by H. Strobl) in Dulbecco's Modified (Pierce). Eagle Medium supplemented with 10% FBS and antibiotics. HeLa cells were transfected with Metafectene Pro (Bion- Co-immunoprecipitation and Western blotting tex), HEK293, and Phoenix cells with Lipofectamine 2000 Cotransfected cells were lysed in nondenaturing whole cell (Invitrogen) according to the manufacturers' protocols. For lysis buffer containing 100 U/mL benzonase (Novagen). As retroviral transduction, we generated NALM-6 cells carrying positive controls 5% of the cleared lysate inputs were

OF2 Mol Cancer Res; 12(4) April 2014 Molecular Cancer Research

PAX5 Fusion Proteins

retained. Lysates were incubated first with 1 mg of mouse binding–deficient (K67A/K87A/K89A) mutant (33), normal immunoglobulin G (IgG; sc-2025) or monoclonal PAX5–DACH1, PAX5–DACH2, PAX5–HIPK1, PAX5– HA (sc-7392X) or myc antibody (sc-40X) followed by anti- POM121, or PAX5–ETV6. The proteins were detected by mouse IgG M-280 Dynabeads (Invitrogen). The beads were indirect immunofluorescence using HA antibody followed washed 3 times with lysis buffer and boiled in SDS loading by confocal microscopy. Both wild-type and mutant PAX5 buffer to elute the precipitated proteins. localized exclusively to the nuclei of NALM-6-EcoR cells Samples were separated in 8% SDS-PAGE gels. After (Fig. 1B). Similarly, all fusion proteins were mainly detect- transfer to nitrocellulose membranes (GE Healthcare), blots able in the nuclei (Fig. 1B). Of note, for PAX5–ETV6 we were incubated with blocking reagent (Roche), and primary also observed a punctate staining pattern in the cytoplasm, at and secondary antibodies diluted in TBS-Tween 20, and least in a fraction of the cells (Fig. 1B). In HeLa cells, PAX5– then scanned using a LI-COR Odyssey imaging system. ETV6 and PAX5–HIPK1 partially localized to the cyto- Antibodies are listed in the Supplementary Material. plasm, whereas PAX5, PAX5–DACH1, PAX5–DACH2, and PAX5–POM121 were strictly nuclear (Supplementary Chromatin immunoprecipitation and quantitative PCR Fig. S1). Together, these findings support the notion that Chromatin immunoprecipitation (ChIP) was performed PAX5 fusions are nuclear proteins, which may consequently as described elsewhere (30) using 2 107 cells and 2 mg act as aberrant transcription factors. mouse HA, PAX5, or normal IgG antibody. Immunopre- cipitated DNA was purified using the QIAquick Kit (Qia- Some PAX5 fusion proteins self-interact whereas others gen). Selection of PAX5 target loci was based on published do not data (4–6, 31) and ENCODE ChIP-sequencing data for Several PAX5 fusion partner proteins contain potential GM12878 cells (32). Quantitative PCR (qPCR) was per- oligomerization domains and PAX5–C20orf112 and formedusingIQSYBRGreenmix(Bio-Rad)and200nmol/L PAX5–PML have already been demonstrated to form oli- primers on an iCycler (Bio-Rad). For ChIP, signals were gomers (17, 18). Therefore, we investigated whether the D individually normalized to input DNA using the 2 Ct PAX5 fusion proteins analyzed in this study also self-interact method. For mRNA quantification, RNA was isolated using (Fig. 1A). For this purpose, we simultaneously expressed TRizol (Invitrogen), reverse transcribed using MMLV RT HA- and myc-tagged protein versions in HEK293 cells and (Promega), oligo(dT)18, and random hexamers, and then performed co-immunoprecipitations (CoIP). As expected, qPCR was performed. Cd79a was normalized to Abl1 and PAX5, which is known to bind DNA as a monomer (24), did Gusb reference gene expression. Primers are listed in the not self-associate (Fig. 2A). Furthermore, neither for PAX5– Supplementary Material. HIPK1 nor for PAX5–POM121, which do not harbor any known oligomerization motifs, self-interaction was Luciferase reporter assay observed. In contrast, our data clearly demonstrated self- HEK293 cells were cotransfected in 24-well plates with association of PAX5– DACH1, PAX5–DACH2, and 200 ng CD19-luc-1x or CD79A-pGL4.10 reporter plasmid PAX5–ETV6. PAX5–ETV6 also interacted with wild-type containing either wild-type or mutated PAX5-binding sites, ETV6 (Fig. 2A), supporting the notion that the ETV6 sterile 600 ng pcDNA3 expression and 20 ng pRL-null vector. alpha domain mediates oligomerization of ETV6 and ETV6 Cells were lysed in passive lysis buffer and bioluminescence fusion proteins (34, 35). Likewise, bioinformatics analysis was measured using the Dual-Luciferase reporter assay predicted DACH1 and DACH2 homodimer formation via (Promega) on an Enspire plate reader (Perkin Elmer). parallel coiled-coils of about 100 amino acids (Fig. 1A). In Average fireflytoRenilla luciferase activity ratios were line with this prediction, PAX5–DACH1 and PAX5– normalized to the empty-vector control (fold activation). DACH2 did not interact with the respective coiled-coil– Significance levels were determined using two-way ANOVA deleted mutants (Fig. 2A), confirming that their oligomer- and Bonferroni test comparing wild-type reporters to the ization is mediated by the coiled-coil domains, which has mutated controls. Competition experiments with CD19- also been suggested for Drosophila Dac and murine DACH luc-1x or CD79A-pGL4.10 were performed in HEK293 proteins (36). cells by transfection of 150 or 200 ng of PAX5 vector, To further substantiate the self-interaction of DACH1 respectively, alone or with indicated fold amounts of PAX5 and DACH2 via the coiled-coil motifs, we conducted BiFC fusion vector. Empty pcDNA3 was added to maintain experiments (23) by co-expressing the coiled-coils fused to constant DNA amounts. Significance levels were calculated N- and C-terminal halves of YFP in HeLa cells. Expression of by applying repeated measures one-way ANOVA and Dun- the N- and C-terminal YFP fusions was monitored by nett tests. All assays were performed thrice independently immunofluorescence using antibodies against the Flag- and and each in triplicates. HA-tag, respectively (data not shown). BiFC by reconstitution of YFP was detected for the DACH1 and DACH2 Results coiled-coils and for the positive control bFOS/bJUN, but PAX5 fusion proteins localize mainly to the nucleus not for the negative control FOSDZIP/bJUN (Fig. 2B). To determine the subcellular localization of the investi- However, although the bFOS and bJUN YFP fusions gated PAX5 fusion proteins (Fig. 1A), NALM-6-EcoR cells localized to the nucleus, we noticed that the DACH CC- were retrovirally transduced with HA-tagged PAX5, a DNA- YFP fusions localized mainly to the cytoplasm despite both

www.aacrjournals.org Mol Cancer Res; 12(4) April 2014 OF3

Fortschegger et al.

Figure 1. Nuclear localization of PAX5 fusion proteins. A, schematic representation of the investigated PAX5 fusion proteins. PD, paired domain; 8, octapeptide motif; H, partial homeodomain; TA, transactivation domain; ID, inhibitory domain; CC, coiled-coil; SAM, sterile alpha motif; ETS, E26 transformation-speciﬁc domain; S, sumo interaction motif; aa, amino acids; arrowheads and crosses indicate nuclear localization signals and breakpoints, respectively. B, subcellular localization of the indicated exogenous HA-tagged proteins (red signals) in NALM-6-EcoR cells determined by indirect immunoﬂuorescence and confocal microscopy. White arrow points out cytoplasmic staining and white bars indicate 20 mm. Nuclei are counterstained with DAPI (blue). Pictures were acquired using identical settings.

containing a nuclear localization signal (NLS) (Fig. 1A). In competition experiments, an excess of unlabeled probe Nonetheless, the nuclear localization of PAX5–DACH1 and but not a mutated oligonucleotide inhibited the formation of PAX5–DACH2 (Fig. 1B) in conjunction with the positive labeled DNA/protein complexes of PAX5, PAX5–HIPK1, CoIP results (Fig. 2A) strongly suggests that both fusion PAX5–DACH1, and PAX5–DACH2, and less pronounced proteins self-interact in the nucleus. also of PAX5–POM121 (Fig. 3B). Furthermore, supershift experiments using HA antibody confirmed that the exoge- PAX5 fusion proteins bind to specific DNA sequences nous proteins were responsible for the observed mobility in vitro shifts (Fig. 3B). The highly reduced mobility of the CD19 The nuclear localization and the presence of the PAX5 probe in PAX5–DACH1 and PAX5–DACH2 complexes paired domain indicate DNA-binding properties of the most likely reflects the self-interaction of these proteins PAX5 fusion proteins. To address this issue, we conducted leading to an increased size and thus a stronger retardation electrophoretic mobility shift assays (EMSA) using an oli- (Fig. 3B). gonucleotide constituting a PAX5-binding site from human When EMSAs were conducted using a labeled probe CD19 (28) and HEK293 nuclear extracts containing HA- containing a CD79A promoter-derived PAX5-binding tagged PAX5 or the fusion proteins (Fig. 3A). Shifted probe site (29), shifted bands were of rather weak intensity (Sup- bands were detected with PAX5, PAX5–DACH1, PAX5– plementary Fig. S2), reflecting a significantly lower binding DACH2, PAX5–HIPK1, and PAX5–POM121 (Fig. 3B). affinity to the probe (37). Of note, we did not observe a

OF4 Mol Cancer Res; 12(4) April 2014 Molecular Cancer Research

PAX5 Fusion Proteins

Figure 2. Self-association of PAX5 fusion proteins. A, analysis of PAX5 and PAX5 fusion protein self- association capability, and of the heterologous interactions between PAX5–ETV6 and ETV6, and between PAX5–DACHs and the respective CC deletion (DCC) mutants by CoIP and Western blotting (WB) using the indicated antibodies. Normal IgG immunoprecipitates served as negative, 5% input as positive controls. Asterisks indicate unspeciﬁc bands. One representative of at least 3 replicates is shown. B, BiFC of DACH1 and DACH2 CCs. Reconstitution of YFP ﬂuorescence (yellow signals) indicates self-interaction. The bJUN/bFOS leucine zipper heterodimer served as positive, bJUN and bFOS lacking parts of the zipper domain (FOSDZIP) as negative control, respectively. Nuclei are counterstained with DAPI (blue). White bars indicate 20 mm.

specific band shift for PAX5–ETV6–binding reactions, FLT3, CD79A, and PARP1, and were validated by ChIP which may be caused by aggregation of the fusion protein using NALM-6 cells and PAX5 antibody (Fig. 4B). under native, nondenaturing conditions. Higher amounts of target DNA were detected in anti-HA Furthermore, we asked whether the fusion proteins com- immunoprecipitates of cells expressing PAX5 fusion proteins pete with wild-type PAX5 for DNA binding. Therefore, we as compared with empty-vector control cells or control IgG performed EMSAs with a biotinylated CD19 probe, con- immunoprecipitates (Fig. 4C) clearly showing that PAX5 stant PAX5 and increasing amounts of fusion proteins. To fusion proteins bind to several endogenous PAX5 target sites. properly separate the complexes, we used differently tagged Of note, also PAX5–ETV6, which failed to show band shifts proteins and supershifted the PAX5 fusion protein/DNA in the EMSAs, exhibited binding and displayed remarkably complexes with myc antibody. Increasing amounts of high signals for the CD79A and PARP1 TSSs (Fig. 4C). In PAX5–DACH1, PAX5–DACH2, or PAX5–HIPK1 indeed addition, despite the rather low PAX5–POM121 protein resulted in a decreased intensity of the PAX5/DNA complex levels, its ChIP signals were at least 3 times higher than bands confirming competitive binding of the proteins (Fig. empty vector or IgG control immunoprecipitate levels, 3C). Taken together, the EMSA data verify that PAX5 verifying its DNA-binding capacity. However, because of fusion proteins bind DNA in vitro and compete with PAX5. the variable protein levels of the introduced PAX5 fusions (Fig. 4A), which are probably due to differences in the PAX5 fusion proteins occupy endogenous genomic transduction efficiencies, protein sizes, or stabilities, no final PAX5 target sites conclusions regarding their binding affinities may be drawn. Next, we investigated whether the DNA-binding capability of PAX5 fusion proteins also applies to endogenous Certain PAX5 fusion proteins activate a CD79A PAX5 target sites. For this purpose, NALM-6-EcoR cells promoter reporter expressing the tagged proteins of interest (Fig. 4A) were Based on the observation that PAX5 fusion proteins subjected to ChIP using HA antibody followed by qPCR specifically bind to genomic PAX5 sites, we wanted to know with locus-specific primers. The PAX5 target sequences whether they also influence target gene expression. There- included the 30-end of VPREB3, the first intron of fore, we conducted luciferase reporter assays using HEK293 MIR17HG, and the transcription start sites (TSS) of CD19, cells and a CD19-luc-1x firefly vector containing either one

www.aacrjournals.org Mol Cancer Res; 12(4) April 2014 OF5

Fortschegger et al.

Figure 3. In vitro DNA-binding capability of PAX5 fusion proteins. A, Western blot for HA-tagged PAX5 and PAX5 fusion protein expression was performed with HEK293 nuclear extracts used in B using an HA antibody. B, ﬂuorescently labeled DY782-CD19 probe and nuclear extracts containing indicated HA-tagged proteins were incubated for EMSA without (), with unlabeled wild- type (com) or mutated (mut) CD19 competitor in 50-fold excess. Supershifts were conducted by adding HA antibody sc-7392X (AB). Shifted, supershifted, and unspeciﬁc bands are marked by regular, broad arrows, and asterisks, respectively. One representative experiment of 3 is depicted. C, competition EMSA with PAX5 and PAX5–DACH1, PAX5–DACH2, or PAX5–HIPK1. Indicated amounts of nuclear extract containing HA-PAX5 and myc-tagged fusions were incubated with biotinylated CD19 probe. Myc antibody was used to supershift the PAX5 fusion protein/ DNA complex. Unbound probe, shifted, and supershifted protein/ DNA-complexes are indicated. One representative of 3 experiments is shown.

PAX5-binding site or a mutated site as a negative control. (Supplementary Fig. S3). This might be because of the rather Western blot analysis of the lysates using HA antibody artificial nature of these constructs, which, next to a b-globin confirmed the expression of the respective exogenous pro- minimal promoter, contain 1 or 3 modified PAX5-binding teins. Although PAX5 clearly activated CD19-luc-1x com- sites. Therefore, we cloned the PAX5-activated human pared with the mutated reporter, the PAX5 DNA-binding– CD79A promoter into the firefly vector pGL4.10 and, as deficient mutant did not (Fig. 5A). Furthermore, none of the a negative control, we introduced mutations, which signif- PAX5 fusion proteins was capable of activating CD19-luc-1x icantly decrease the PAX5-binding affinity (38). (Fig. 5A). However, this reporter and its parent luc-CD19 Remarkably, like wild-type PAX5, some fusions, in par- are not activated by endogenous PAX5 in NALM-6 cells ticular PAX5–DACH1 and PAX5–POM121, induced

OF6 Mol Cancer Res; 12(4) April 2014 Molecular Cancer Research

PAX5 Fusion Proteins

Figure 4. Binding of PAX5 fusion proteins to endogenous PAX5 target sites. A, expression levels of exogenous proteins and endogenous PAX5 in the NALM-6-EcoR cells used in C. Areas were sliced from one blot incubatedwithanantibodyrecognizingthePAX5 N-terminus, which allows for the concomitant detection of exogenous and endogenous proteins. B, validation of ChIP primers speciﬁc for the indicated gene loci. NALM-6 chromatin was subjected to immunoprecipitation using normal control sc-2025 (IgG, light gray bars) or PAX5 C-terminus–speciﬁc antibody sc-13146X D (PAX5, dark gray bars). Quantitative PCRwasperformedwithprecipitatedDNAandnormalizedtotheinputusingthe2 Ct method. Mean percentages of input in immunoprecipitates SD of technical triplicates from one representative experiment of 3 are shown. C,ChIPwasperformedwith NALM-6-EcoR cell lines, which stably express the indicated HA-tagged proteins. Normal control sc-2025 (IgG, light gray bars) or HA antibody sc-7392X (dark gray bars) and conditions as in B were used.

www.aacrjournals.org Mol Cancer Res; 12(4) April 2014 OF7

Fortschegger et al.

A B

Figure 5. Luc-CD19 and CD79A promoter reporter assays. Luciferase assays in HEK293 using CD19-luc-1x (A) or CD79A-pGL4.10 (B) containing a functional (dark gray bars) or mutated PAX5 binding site (light gray bars). Fireﬂy activity was measured and normalized to Renilla activity as well as empty pcDNA3 vector control. Dashed lines represent basal levels. Mean fold activation SD (n ¼ 3) is plotted. Two-way ANOVA and Bonferroni test were applied to compare wild-type to mutated reporter activation. Representative immunoblots using HA antibody sc-7392 or GAPDH antibody sc-32233 are shown below. C, competition experiments performed with CD19-luc-1x vector. Numbers indicate fold amounts of PAX5 fusion added to a constant amount of PAX5 vector. Mean fold activation levels SD (n ¼ 3) are depicted. Repeated measures one-way ANOVA and Dunnett tests were applied to compare to PAX5 alone. Signiﬁcance levels: , 0.05 > P > 0.01; , 0.01 > P > 0.001; , P < 0.001. Representative Western blots detecting HA-tagged PAX5 and fusion proteins or GAPDH are shown below. D, competition experiments performed as in C with CD79A-pGL4.10 vector.

OF8 Mol Cancer Res; 12(4) April 2014 Molecular Cancer Research

PAX5 Fusion Proteins

Figure 6. PAX5 fusion proteins fail to activate Cd79a and sIgM in 558LmM cells. A, ﬂow cytometry using PE-labeled antibody against mouse sIgM was performed 4 days after transduction of 558LmM cells with indicated MSCV-MigR1 vectors. Intact cells were gated according to forward and side scatter, and the sIgM-positive percentages of the GFP-positive cells were calculated. B, summary derived from 3 independent experiments performed as in A (mean sIgM-positive percentage SD are plotted; repeated measures one-way ANOVA and Dunnett tests were applied to compare the PAX5 proteins to the empty vector; , P < 0.001). C, quantitative RT-PCR was performed with cDNA of GFP-sorted 558LmM cells from the experiments in B. Cd79a was normalized to Gusb and Abl1 expression and to the empty- vector control. The same statistical tests as in B were applied.

reporter expression driven by the CD79A promoter (Fig. was transduced with PAX5, its DNA-binding–deficient 5B). In contrast, the PAX5 DNA-binding–deficient mutant mutant or the fusions. In about 20% of the GFP-positive and PAX5–ETV6 did not activate the CD79A promoter PAX5 expressing cells, via activation of Cd79a, surface IgM construct (Fig. 5B). Although the mutated reporters were (sIgM) was present (Fig. 6A–C). In contrast, neither the still slightly activated by PAX5, as well as by some of the DNA-binding–deficient PAX5 mutant nor any fusion pro- fusion proteins, we attribute the potent reporter induction to tein significantly induced sIgM (Fig. 6A and B). Although promoter recruitment of the introduced proteins and to the expression level of wild-type PAX5 and most of the association of the paired domain with the target sequence. fusion proteins was comparable (Supplementary Fig. S4), Competition experiments confirmed that PAX5 fusion only PAX5 activated Cd79a transcription (Fig. 6C). Con- proteins inhibit PAX5-mediated activation of CD19-luc-1x sequently, in contrast to the CD79A-pGL4.10 reporter in (Fig. 5C). In addition, the induction of CD79A-pGL4.10 HEK293 cells, the PAX5 fusions were unable to activate by PAX5 was also slightly repressed by PAX5–ETV6, but, in endogenous Cd79a and sIgM presentation in 558LmM cells. stark contrast, was markedly increased by PAX5–DACH1 and PAX5–DACH2 (Fig. 5D). Together, these data provide Discussion first evidence that certain PAX5 fusion proteins may activate PAX5-activated target genes such as CD79A. In this study, we investigated the biochemical and functional characteristics of a number of PAX5 fusion proteins. PAX5 fusion proteins do not induce surface IgM in As a unifying feature, PAX5 chimeras share the N-terminal 558LmM cells DNA-binding paired domain, indicating that they may act To further investigate the activation potential of PAX5 in a trans–dominant-negative mode by competing for bind- fusions on the endogenous Cd79a promoter, the murine ing to PAX5 target sites and thereby antagonizing wild-type 558LmM plasmacytoma cell line, which expresses 3 of the 4 PAX5 activity. However, the fusion partner proteins differ components of the B-cell receptor, except for CD79A (27), considerably in terms of their function, localization, and

www.aacrjournals.org Mol Cancer Res; 12(4) April 2014 OF9

Fortschegger et al.

structure (8–16), raising the question whether these differ- Collectively, these data support the current notion that ences have an impact on the actual function of the individual PAX5 fusion proteins bind to PAX5 target loci without chimera. providing normal regulatory functions, thereby antagoniz- Independent of the original subcellular localization of the ing PAX5 activity provided by the wild-type allele. However, partner proteins and their highly diverse C-terminal in contrast to PAX5–ETV6, which has been demonstrated domains, all PAX5 fusion proteins predominantly localize to primarily act as a transcriptional repressor (8, 20), PAX5– to the nucleus (Fig. 1B). Their nuclear import seems to be DACH1, PAX5–DACH2, and PAX5–POM121 showed mainly dictated by the PAX5 paired domain and at least one induction of the CD79A promoter reporter (Fig. 5B–D). NLS provided by PAX5 itself or the individual partner Notably, DACH1 has been suggested to play a role as a co- proteins (Fig. 1A). However, PAX5–C20orf112, PAX5– activator in myeloid cells (44, 45) and there is growing FOXP1, and PAX5–ETV6 have been detected in both the evidence that nuclear pore components, such as POM121, nuclear and cytoplasmic fractions (13), and, as shown herein, additionally localize to the nucleoplasm and activate tran- also PAX5–HIPK1 partially localizes to the cytoplasm scription (46, 47). However, none of the PAX5 fusion (Supplementary Fig. S1), suggesting a tendency to shuttle proteins exhibited Cd79a transactivation capability in between cellular compartments. In this regard, nucleo-cyto- murine 558LmM cells (Fig. 6C). plasmic shuttling has been demonstrated for both wild-type The apparent discrepancy between the luciferase assay and HIPK1 and ETV6 (39, 40). However, the significance of this the 558LmM cell system is supposedly because of different finding and whether PAX5 fusion proteins might also exert target readouts, one being an episomal reporter gene tran- specific functions in the cytoplasm remains to be determined. siently expressed in human embryonic kidney-derived cells, Nevertheless, all PAX5 fusion proteins are capable of the other constituting a chromatinized silent endogenous binding to genomic DNA, but supposedly with variable locus in murine plasmacytoma cells. On the one hand, the affinities. Remarkably, in ChIP experiments PAX5–ETV6 reporter construct might lack more distal regulatory regions, bound rather strongly to the CD79A and PARP1 promoters which in the plasmacytoma cells might be activated by wild- (Fig. 4C), which may be explained by additional protein/ type PAX5 but not the fusions. On the other hand, to exert DNA interactions with adjacent sites via the ETV6 ETS their activation potential, PAX5 fusion proteins may require domain. Although a link between PAX5 and ETV6 has not factors absent in 558LmM cells, and that either, like ETS1, yet been described, cooperative binding to the murine EBF1, TCF3, and chromatin remodeling/modifying com- Cd79a promoter of PAX5 and a similar transcription factor, plex components, cooperate with PAX5 (6, 33, 42, 48, 49) namely ETS1 (41), has already been demonstrated (42). or with the fusion partner, for example the DACH proteins PAX5–FOXP1 and PAX5–ZNF521 also contain additional (50). In this context, luciferase assays in HeLa cells and target DNA-binding domains provided by the partner proteins (8), gene regulation in DG75 lymphoma cells ectopically expres- and thus might bind to distinct spectrums of target sing PAX5–ELN as well as PAX5–ELN–positive primary sequences. Together, these data suggest that at least some leukemia cells also showed remarkable differences (10). PAX5 fusion proteins may not only interfere with the Together, these data suggest that the impact of PAX5 expression of PAX5 targets but also with that of their fusion fusion proteins is critically dependent on the cellular context partner proteins (19, 21). and differentiation state, which is supposedly due to differ- It has also been suggested that self-interaction of PAX5 ences in the protein repertoire as well as chromatin archi- fusion proteins alters their DNA affinities (17, 18). Although tecture (3, 6, 33). This notion is further supported by recent it has been proposed that PAX5–PML has a lower DNA- data showing that introduction of PAX5 into Hodgkin binding affinity than wild-type PAX5 (18), for PAX5– lymphoma cells is unable to reestablish a mature B-cell C20orf112 the opposite—that is, a highly stable chromatin expression signature (51). interaction—has been reported (17). In summary, our study confirms that PAX5 fusion pro- As shown herein, 3 additional fusion proteins, namely teins share common features, including nuclear localization PAX5–ETV6, PAX5–DACH1, and PAX5–DACH2 self- and DNA-binding capability. However, our data also pro- associate (Fig. 2A). Of note, oligomerization of ETV6 fusion vide evidence that the heterogeneous C-terminal domains proteins, in particular of tyrosine kinase chimeras, has been may modulate their DNA-binding affinities and their reg- shown to be crucial for oncogenic transformation (43). ulatory potentials to activate or repress target genes, which Intriguingly, PAX5 fusions, such as PAX5–ETV6, may also rather argues for a more complex mode of action of the fusion interact with, and influence the wild-type partner proteins proteins than for a simple PAX5 antagonizing function. (Fig. 2A), which may also play a role in the development of the respective leukemia. In line with this notion, a dual Disclosure of Potential Conflicts of Interest fl dominant-negative mode has been proposed for PAX5– No potential con icts of interest were disclosed. PML, which also forms homo- and heterodimers (18, – Authors' Contributions 19). However, this concept does not apply to PAX5 Conception and design: K. Fortschegger, S. Strehl DACH1 and PAX5–DACH2 because both DACH1 and Development of methodology: K. Fortschegger, S. Anderl, D. Denk DACH2 are not expressed in B-cells and DACH1 is also not Acquisition of data (provided animals, acquired and managed patients, provided – – facilities, etc.): K. Fortschegger, S. Anderl expressed in primary PAX5 DACH1 positive leukemia Analysis and interpretation of data (e.g., statistical analysis, biostatistics, compu- (data not shown). tational analysis): K. Fortschegger, S. Anderl, D. Denk, S. Strehl

OF10 Mol Cancer Res; 12(4) April 2014 Molecular Cancer Research

PAX5 Fusion Proteins

Writing, review, and/or revision of the manuscript: K. Fortschegger, S. Anderl, Grant Support D. Denk, S. Strehl This study was supported by the Austrian Science Fund (FWF Project No. Study supervision: S. Strehl P21554-B19) and the St. Anna Kinderkrebsforschung e.V. The costs of publication of this article were defrayed in part by the payment of page Acknowledgments charges. This article must therefore be hereby marked advertisement in accordance with The authors thank W. Mensing, K. Nebral, and D. Printz for their assistance, 18 U.S.C. Section 1734 solely to indicate this fact. M. Busslinger, M. Reth, R. Pestell, C. Broccardo, H. Strobl, and T. Kerppola for providing material, and A.M. Dohnal and D.N. Aryee for their critical review of Received June 26, 2013; revised December 2, 2013; accepted December 18, 2013; the article. published OnlineFirst January 16, 2014.

References 1. Medvedovic J, Ebert A, Tagoh H, Busslinger M. Pax5: a master 17. Kawamata N, Pennella MA, Woo JL, Berk AJ, Koeffler HP. Dominant- regulator of B cell development and leukemogenesis. Adv Immunol negative mechanism of leukemogenic PAX5 fusions. Oncogene 2011;111:179–206. 2012;31:966–77. 2. Cobaleda C, Schebesta A, Delogu A, Busslinger M. Pax5: the guardian 18. Qiu JJ, Chu H, Lu X, Jiang X, Dong S. The reduced and altered activities of B cell identity and function. Nat Immunol 2007;8:463–70. of PAX5 are linked to the protein-protein interaction motif (coiled-coil 3. Revilla-i-Domingo R, Bilic I, Vilagos B, Tagoh H, Ebert A, Tamir IM, et al. domain) of the PAX5-PML fusion protein in t(9;15)-associated acute The B-cell identity factor Pax5 regulates distinct transcriptional pro- lymphocytic leukemia. Oncogene 2011;30:967–77. grammes in early and late B lymphopoiesis. EMBO J 2012;31:3130– 19. Kurahashi S, Hayakawa F, Miyata Y, Yasuda T, Minami Y, Tsuzuki S, 46. et al. PAX5-PML acts as a dual dominant-negative form of both PAX5 4. Schebesta A, McManus S, Salvagiotto G, Delogu A, Busslinger GA, and PML. Oncogene 2011;30:1822–30. Busslinger M. Transcription factor Pax5 activates the chromatin of key 20. Fazio G, Palmi C, Rolink A, Biondi A, Cazzaniga G. PAX5/TEL acts as a genes involved in B cell signaling, adhesion, migration, and immune transcriptional repressor causing down-modulation of CD19, function. Immunity 2007;27:49–63. enhances migration to CXCL12, and confers survival advantage in 5. Delogu A, Schebesta A, Sun Q, Aschenbrenner K, Perlot T, Busslinger pre-BI cells. Cancer Res 2008;68:181–9. M. Gene repression by Pax5 in B cells is essential for blood cell 21. Fazio G, Cazzaniga V, Palmi C, Galbiati M, Giordan M, Te Kronnie G, homeostasis and is reversed in plasma cells. Immunity 2006;24: et al. PAX5/ETV6 alters the gene expression profile of precursor B cells 269–81. with opposite dominant effect on endogenous PAX5. Leukemia 6. McManus S, Ebert A, Salvagiotto G, Medvedovic J, Sun Q, Tamir I, 2012;27:992–95. et al. The transcription factor Pax5 regulates its target genes by 22. Pear WS, Miller JP, Xu L, Pui JC, Soffer B, Quackenbush RC, et al. recruiting chromatin-modifying proteins in committed B cells. EMBO Efficient and rapid induction of a chronic myelogenous leukemia-like J 2011;30:2388–404. myeloproliferative disease in mice receiving P210 bcr/abl-transduced 7. Familiades J, Bousquet M, Lafage-Pochitaloff M, Bene MC, Beldjord bone marrow. Blood 1998;92:3780–92. K, De Vos J, et al. PAX5 mutations occur frequently in adult B-cell 23. Hu CD, Chinenov Y, Kerppola TK. Visualization of interactions among progenitor acute lymphoblastic leukemia and PAX5 haploinsufficiency bZIP and Rel family proteins in living cells using bimolecular fluores- is associated with BCR-ABL1 and TCF3-PBX1 fusion genes: a cence complementation. Mol Cell 2002;9:789–98. GRAALL study. Leukemia 2009;23:1989–98. 24. Czerny T, Busslinger M. DNA-binding and transactivation properties of 8. Mullighan CG, Goorha S, Radtke I, Miller CB, Coustan-Smith E, Dalton Pax-6: three amino acids in the paired domain are responsible for the JD, et al. Genome-wide analysis of genetic alterations in acute lym- different sequence recognition of Pax-6 and BSAP (Pax-5). Mol Cell phoblastic leukaemia. Nature 2007;446:758–64. Biol 1995;15:2858–71. 9. Nebral K, Denk D, Attarbaschi A, Konig M, Mann G, Haas OA, et al. 25. Wu K, Yang Y, Wang C, Davoli MA, D'Amico M, Li A, et al. DACH1 Incidence and diversity of PAX5 fusion genes in childhood acute inhibits transforming growth factor-beta signaling through binding lymphoblastic leukemia. Leukemia 2009;23:134–43. Smad4. J Biol Chem 2003;278:51673–84. 10. Bousquet M, Broccardo C, Quelen C, Meggetto F, Kuhlein E, Delsol G, 26. Jorgl A, Platzer B, Taschner S, Heinz LX, Hocher B, Reisner PM, et al. et al. A novel PAX5-ELN fusion protein identified in B-cell acute Human Langerhans-cell activation triggered in vitro by conditionally lymphoblastic leukemia acts as a dominant negative on wild-type expressed MKK6 is counterregulated by the downstream effector PAX5. Blood 2007;109:3417–23. RelB. Blood 2007;109:185–93. 11. Cazzaniga G, Daniotti M, Tosi S, Giudici G, Aloisi A, Pogliani E, et al. 27. MaierH,ColbertJ,FitzsimmonsD,ClarkDR,HagmanJ.Activation The paired box domain gene PAX5 is fused to ETV6/TEL in an acute of the early B-cell-specific mb-1 (Ig-alpha) gene by Pax-5 is depen- lymphoblastic leukemia case. Cancer Res 2001;61:4666–70. dent on an unmethylated Ets binding site. Mol Cell Biol 2003; 12. Coyaud E, Struski S, Prade N, Familiades J, Eichner R, Quelen C, 23:1946–60. et al. Wide diversity of PAX5 alterations in B-ALL: a Groupe Fran- 28. Kozmik Z, Wang S, Dorfler P, Adams B, Busslinger M. The promoter of cophone de Cytogenetique Hematologique study. Blood 2010; the CD19 gene is a target for the B-cell-specific transcription factor 115:3089–97. BSAP. Mol Cell Biol 1992;12:2662–72. 13. Kawamata N, Ogawa S, Zimmermann M, Niebuhr B, Stocking C, 29. Bougel S, Renaud S, Braunschweig R, Loukinov D, Morse HC 3rd, Sanada M, et al. Cloning of genes involved in chromosomal transloca- Bosman FT, et al. PAX5 activates the transcription of the human tions by high-resolution single nucleotide polymorphism genomic telomerase reverse transcriptase gene in B cells. J Pathol 2010; microarray. Proc Natl Acad Sci U S A 2008;105:11921–6. 220:87–96. 14. Lee ST, Ji Y, Kim HJ, Ki CS, Jung CW, Kim JW, et al. Sequential array 30. Fortschegger K, de Graaf P, Outchkourov NS, van Schaik FM, Timmers comparative genomic hybridization analysis identifies copy number HT, Shiekhattar R. PHF8 targets histone methylation and RNA poly- changes during blastic transformation of chronic myeloid leukemia. merase II to activate transcription. Mol Cell Biol 2010;30:3286–98. Leuk Res 2012;36:418–21. 31. Pridans C, Holmes ML, Polli M, Wettenhall JM, Dakic A, Corcoran LM, 15. Nebral K, Konig M, Harder L, Siebert R, Haas OA, Strehl S. Identifi- et al. Identification of Pax5 target genes in early B cell differentiation. cation of PML as novel PAX5 fusion partner in childhood acute J Immunol 2008;180:1719–28. lymphoblastic leukaemia. Br J Haematol 2007;139:269–74. 32. Wang J, Zhuang J, Iyer S, Lin X, Whitfield TW, Greven MC, et al. 16. Denk D, Nebral K, Bradtke J, Pass G, Moricke A, Attarbaschi A, et al. Sequence features and chromatin structure around the genomic PAX5-AUTS2: a recurrent fusion gene in childhood B-cell precursor regions bound by 119 human transcription factors. Genome Res acute lymphoblastic leukemia. Leuk Res 2012;36:e178–81. 2012;22:1798–812.

www.aacrjournals.org Mol Cancer Res; 12(4) April 2014 OF11

Fortschegger et al.

33. He T, Hong SY, Huang L, Xue W, Yu Z, Kwon H, et al. Histone 42. Garvie CW, Hagman J, Wolberger C. Structural studies of Ets-1/Pax5 acetyltransferase p300 acetylates Pax5 and strongly enhances complex formation on DNA. Mol Cell 2001;8:1267–76. Pax5-mediated transcriptional activity. J Biol Chem 2011;286: 43. De Braekeleer E, Douet-Guilbert N, Morel F, Le Bris MJ, Basinko A, De 14137–45. Braekeleer M. ETV6 fusion genes in hematological malignancies: a 34. Golub TR, Goga A, Barker GF, Afar DE, McLaughlin J, Bohlander SK, review. Leuk Res 2012;36:945–61. et al. Oligomerization of the ABL tyrosine kinase by the Ets protein TEL 44. Lee JW, Kim HS, Hwang J, Kim YH, Lim GY, Sohn WJ, et al. Regulation in human leukemia. Mol Cell Biol 1996;16:4107–16. of HOXA9 activity by predominant expression of DACH1 against C/ 35. Jousset C, Carron C, Boureux A, Quang CT, Oury C, Dusanter-Fourt EBPalpha and GATA-1 in myeloid leukemia with MLL-AF9. Biochem I, et al. A domain of TEL conserved in a subset of ETS proteins Biophys Res Commun 2012;426:299–305. defines a specific oligomerization interface essential to the mito- 45. Lee JW, Kim HS, Kim S, Hwang J, Kim YH, Lim GY, et al. DACH1 genic properties of the TEL-PDGFR beta oncoprotein. EMBO J regulates cell cycle progression of myeloid cells through the control of 1997;16:69–82. cyclin D, Cdk 4/6 and p21Cip1. Biochem Biophys Res Commun 36. Okamoto N, Nishimori Y, Nishimura T. Conserved role for the Dachs- 2012;420:91–5. hund protein with Drosophila Pax6 homolog eyeless in insulin expres- 46. Kalverda B, Pickersgill H, Shloma VV, Fornerod M. Nucleoporins sion. Proc Natl Acad Sci U S A 2012;109:2406–11. directly stimulate expression of developmental and cell-cycle genes 37. Fitzsimmons D, Hodsdon W, Wheat W, Maira SM, Wasylyk B, Hagman inside the nucleoplasm. Cell 2010;140:360–71. J. Pax-5 (BSAP) recruits Ets proto-oncogene family proteins to form 47. Capelson M, Liang Y, Schulte R, Mair W, Wagner U, Hetzer MW. functional ternary complexes on a B-cell-specific promoter. Genes Chromatin-bound nuclear pore components regulate gene expression Dev 1996;10:2198–211. in higher eukaryotes. Cell 2010;140:372–83. 38. Wheat W, Fitzsimmons D, Lennox H, Krautkramer SR, Gentile LN, 48. Gao H, Lukin K, Ramirez J, Fields S, Lopez D, Hagman J. Opposing McIntosh LP, et al. The highly conserved b-hairpin of the paired DNA- effects of SWI/SNF and Mi-2/NuRD chromatin remodeling complexes binding domain is required for assembly of Pax-Ets ternary complexes. on epigenetic reprogramming by EBF and Pax5. Proc Natl Acad Sci Mol Cell Biol 1999;19:2231–41. U S A 2009;106:11258–63. 39. Hanson CA, Wood LD, Hiebert SW. Cellular stress triggers TEL nuclear 49. Sigvardsson M, Clark DR, Fitzsimmons D, Doyle M, Akerblad P, Breslin export via two genetically separable pathways. J Cell Biochem T, et al. Early B-cell factor, E2A, and Pax-5 cooperate to activate the 2008;104:488–98. early B cell-specific mb-1 promoter. Mol Cell Biol 2002;22:8539–51. 40. Li X, Zhang R, Luo D, Park SJ, Wang Q, Kim Y, et al. Tumor necrosis 50. Ikeda K, Watanabe Y, Ohto H, Kawakami K. Molecular interaction and factor alpha-induced desumoylation and cytoplasmic translocation of synergistic activation of a promoter by Six, Eya, and Dach proteins homeodomain-interacting protein kinase 1 are critical for apoptosis mediated through CREB binding protein. Mol Cell Biol 2002;22: signal-regulating kinase 1-JNK/p38 activation. J Biol Chem 2005; 6759–66. 280:15061–70. 51. Dimitrova L, Seitz V, Hecht J, Lenze D, Hansen P, Szczepanowski M, 41. Wei GH, Badis G, Berger MF, Kivioja T, Palin K, Enge M, et al. Genome- et al. PAX5 overexpression is not enough to reestablish the mature wide analysis of ETS-family DNA-binding in vitro and in vivo. EMBO J B-cell phenotype in classical Hodgkin lymphoma. Leukemia 2014; 2010;29:2147–60. 28:213–6.

OF12 Mol Cancer Res; 12(4) April 2014 Molecular Cancer Research

Functional Heterogeneity of PAX5 Chimeras Reveals Insight for Leukemia Development

Klaus Fortschegger, Stefanie Anderl, Dagmar Denk, et al.

Mol Cancer Res Published OnlineFirst January 16, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-13-0337

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2014/01/17/1541-7786.MCR-13-0337.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/early/2014/04/01/1541-7786.MCR-13-0337. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.