Leukemia (2011) 25, 1007–1016 & 2011 Macmillan Publishers Limited All rights reserved 0887-6924/11 www.nature.com/leu ORIGINAL ARTICLE

Proapoptotic signaling activity of the anti-CD40 dacetuzumab circumvents multiple oncogenic transformation events and chemosensitizes NHL cells

TS Lewis1, RS McCormick1, IJ Stone1, K Emmerton1, B Mbow2, J Miyamoto1, JG Drachman3, IS Grewal1 and C-L Law1

1Department of Pre-Clinical Research, Seattle Genetics, Inc., Bothell, WA, USA; 2Department of Experimental Therapeutics, Seattle Genetics, Inc., Bothell, WA, USA and 3Department of Research & Translational Medicine, Seattle Genetics, Inc., Bothell, WA, USA

Non-Hodgkin lymphoma (NHL) is a genetically heterogeneous oncogenic events, thereby reducing proliferation and resensitiz- disease with several oncogenic events implicated in the ing cells to apoptotic stimuli, would represent a significant transformation of normal developing B lymphocytes. The advance in the treatment of NHL. objective of this study was to elucidate the signal transduc- tion-based antitumor mechanism(s) of action for the anti-CD40 Lymphomas maintain many characteristics of the B-cell monoclonal antibody dacetuzumab (SGN-40) in NHL. We report differentiation state at which they originated, including surface that dacetuzumab activates two distinct proapoptotic signaling antigen expression. The receptor (TNFR) pathways, overcoming transformation events key to the superfamily member CD40 is expressed on B cells from the early pathogenesis of NHL. Dacetuzumab-mediated CD40 signaling pro-B phase through the fully differentiated memory B or constitutively activated the nuclear factor-jB and mitogen- secreting plasma cell phases.7,8 CD40 is a transmembrane activated protein kinase signaling pathways producing the sustained downregulation of B-cell lymphoma 6 (BCL-6), an protein with a large extracellular domain and a 42 amino acid oncoprotein implicated in lymphomagenesis. Loss of BCL-6 cytoplasmic TNFR-associated factor interacting domain, which resulted in c-Myc downregulation and activation of a transcrip- mediates downstream signaling through the nuclear factor kB tional program characteristic of early B-cell maturation, con- (NF-kB), mitogen-activated protein kinase (MAPK), phosphoi- comitant with reduced proliferation and cell death. In a second nositide 3-kinase and phospholipase Cg pathways.9 CD40 mechanism, dacetuzumab signaling induced the expression of signaling stimulated by binding to its trimeric ligand (CD40L the proapoptotic p53 family member TAp63a and downstream proteins associated with the intrinsic and extrinsic apoptotic or CD154) has an important role in normal B-cell differentiation, machinery. Dacetuzumab was synergistic in combination with and mutation at the CD40L locus is responsible for X-linked DNA-damaging chemotherapeutic drugs, correlating with hyper-IgM immunodeficiency syndrome.10 Interaction of a naı¨ve TAp63a upregulation. Furthermore, dacetuzumab augmented B cell with T-cell-expressed CD40L initiates proliferation and the activity of in combination with multiple chemo- formation of germinal centers (GCs), whereas CD40–CD40L therapies in the xenograft models of NHL. The ability of signaling in the GC light zone initiates a program to complete dacetuzumab signaling to circumvent oncogenic events and antibody affinity maturation and begin differentiation into mature potentiate the activity of regimens provides a 11 unique therapeutic approach to NHL. plasma cells or memory B cells. Leukemia (2011) 25, 1007–1016; doi:10.1038/leu.2011.21; CD40 is an attractive oncology target for therapeutic published online 11 March 2011 intervention commonly expressed on lymphoid malignancies Keywords: dacetuzumab; SGN-40; lymphoma; BCL-6; TAp63; (NHL, multiple myeloma, Hodgkin lymphoma and acute MAPK lymphoblastic leukemia) as well as on a variety of carcinomas.12 As with other TNFR superfamily members, signaling through CD40 activates pathways associated with both survival and apoptosis, which in NHL may produce different outcomes Introduction depending upon the lymphoma subtype.13 A prolonged cell survival response is observed in chronic lymphocytic leukemia, Non-Hodgkin lymphoma (NHL) is a genetically diverse subset of whereas an anti-proliferative apoptotic response is observed in lymphomas originating from developing B cells transformed at high-grade NHL such as diffuse large B-cell lymphoma various points in the normal B-cell maturation process. (DLBCL).14,15 This differing outcome to CD40 signaling in Oncogenic events occurring during VDJ recombination, somatic lymphoid malignancies likely correlates with the stage in B-cell hypermutation or immunoglobulin (Ig) class switching prevent maturation at which transformation occurred. apoptosis of B cells not activated by antigen binding to B-cell 1 At present, standard lymphoma therapy combines the anti- receptor, maintaining cells in an artificial proliferative state. CD20 monoclonal antibody rituximab with chemotherapy Genes deregulated or inactivated in NHL include overexpres- regimens such as CHOP (cyclophosphamide, doxorubicin, sion of B-cell lymphoma 6 (BCL-6), MYC or B-cell leukemia/ vincristine and prednisone) and ICE (ifosfamide, carboplatin lymphoma 2 (BCL-2) by translocation within Ig loci, FAS 16,17 2–6 and etoposide). A complementary approach to the treat- mutation and loss or mutation of TP53. A therapeutic ment of NHL is the targeting of CD40 with monoclonal approach capable of overcoming one or more of these antibodies or with recombinant human CD40L, both of which have been shown to induce apoptosis in aggressive lympho- Correspondence: Dr TS Lewis, Department of Pre-Clinical Research, mas.12,18 The humanized anti-CD40 monoclonal antibody Seattle Genetics, Inc., 21823–30th Drive Southeast, Bothell, WA dacetuzumab (SGN-40) mediates effector function activity, 98021, USA. E-mail: tlewis@.com antibody-dependent cellular cytotoxicity and antibody-depen- Received 19 April 2010; revised 23 November 2010; accepted 23 dent cellular phagocytosis, and is also able to directly induce 19,20 December 2010; published online 11 March 2011 apoptosis in cultured NHL cells. A comparison of Dacetuzumab proapoptotic signaling mechanisms in NHL TS Lewis et al 1008 dacetuzumab antitumor activity in severe combined immuno- were detected by western blotting with antibodies specific to deficient mice versus natural killer cell-deficient severe com- BCL-6 (Abcam, Cambridge, MA, USA), p53, BID, b-actin (Cell bined immunodeficient-beige mice demonstrated that efficacy Signaling Technology), BIM, c-Myc and p63 (BD Pharmingen, in certain lymphoma xenograft models is largely independent of San Jose, CA, USA). Fas upregulation was measured by flow effector function activity.19 This finding suggested that dacetu- cytometry with anti-Fas-FITC-conjugated antibody and FACScan zumab apoptotic signaling is an important component of its flow cytometer (BD Biosciences, San Jose, CA). in vivo antitumor activity in NHL. The goal of this study was to define non-effector function, RT-PCR analysis signal transduction-based mechanisms of action by which RNA was extracted with either RNA-Bee (Tel-Test, Friends- dacetuzumab induces apoptosis in NHL cell lines. We have wood, TX, USA) or RNeasy spin columns (Qiagen, Valencia, found that dacetuzumab stimulates signal transduction through CA, USA). Reverse transcription of 5.0 mg RNA per reaction was CD40 on lymphoma cells, activating the NF-kB, ERK1/2 MAPK, conducted using the SuperScript III First-Strand Synthesis System JNK1/2 MAPK and p38 MAPK pathways in a sustained manner. (Invitrogen). The resulting cDNA was PCR amplified with Dacetuzumab signaling resulted in downregulation of BCL-6 RedTaq (Sigma, St Louis, MO, USA) and gene-specific primers: oncoprotein and loss of c-Myc expression in GC-derived Blimp-1 (forward) 50-TGAGTAAAGAATACATACCAAAGGG-30 lymphoma cells, driving them from a transformed condition (reverse) 50-ATTCTTTGGGCAGAGTTCATTT-30, XBP-1 (forward) into a non-proliferative state resensitized to apoptotic stimuli. 50-CGCTGAGGAGGAAACTGAA-30,(reverse)50-GGGAGGCTG Furthermore, dacetuzumab upregulated the proapoptotic p53 GTAAGGAACT-30 and glyceraldehyde 3-phosphate dehydro- family member, TAp63a and downstream proteins involved in genase (forward) 50-CCACCCATGGCAAATTCCATGGCA-30 both the extrinsic and intrinsic apoptotic pathways. Dacetuzu- (reverse) 50-TCTAGACGGCAGGTCAGGTCCACC-30. Blimp-1, mab chemosensitized lymphoma cells in vitro and in mouse XBP-1 and glyceraldehyde 3-phosphate dehydrogenase PCR xenograft models, potentiating the activity of standard NHL products were separated on agarose gels, stained with ethidium chemotherapeutic regimens. These data suggest dacetuzumab- bromide and imaged and quantitated with a FluorChemQ imaging mediated signaling through CD40 can overcome key transfor- system (Alpha Innotech, San Leandro, CA, USA). mation events characteristic of NHL and improve the activity of existing rituximab-based combination NHL therapies. Combination index analysis Dacetuzumab and chemotherapeutic agents Bleomycin (Fauld- Materials and methods ing Labs, Paramus, NJ, USA), chlorambucil (LKT Labs, St Paul, MN, USA), carboplatin, doxorubicin and etoposide (Sigma) Cell lines, proliferation and apoptosis assays, and were serially diluted by twofold and added to SU-DHL-8 cells inhibitors in quadruplicate, either alone or in combination. Cells NHL cell lines Ramos, Toledo (ATCC, Manassas, VA, USA), RL, were treated for 72 h, and proliferation measured with the SU-DHL-4 and SU-DHL-8 (DSMZ, Braunschweig, Germany) 3H-thymidine incorporation assay. Dose–effect curves were were cultured in RPMI media þ 10% fetal bovine serum at generated and combination indices (CI) were calculated with 37 1C, 5% CO2. For cell proliferation assays, dacetuzumab was 0 CalcuSyn software (BioSoft, Ferguson, MO, USA). Combination crosslinked with a Fcg-specific goat anti-human IgG F(ab )2 indice values were expressed as the CI values at ED90,ED75 and fragment (Jackson ImmunoReasearch, Westgrove, PA, USA) (1:4 ED50±the s.e.m. from n ¼ 3 independent experiments. ratio), and plated onto cells in quadruplicate. Cells were treated for 72 h and then incubated for 8–24 h with 3H-thymidine and DNA-incorporated radioactivity was measured with a TopCount Immunohistochemistry NXT (PerkinElmer, Waltham, MA, USA). Cell death was Double-labeling immunohistochemistry was performed on measured using the annexin V-FITC apoptosis detection kit formalin-fixed, paraffin-embedded Ramos xenograft tumor (EMD Biosciences, San Diego, CA, USA). Small-molecule sections using the BondMax Autostainer (Leica Microsystems, signaling inhibitors U0126, JNK inhibitor VIII, p38 MAPK Bannockburn, IL, USA). Following heat-induced antigen retrie- inhibitor III and NF-kB activation inhibitor (6-amino-4-(4- val in EDTA, the Bond Polymer AP Red and Bond Intense R (for phenoxyphenethylamino)quinazoline) were obtained from biotinylated antibodies) kits were used for antibody detection EMD Biosciences, La Jolla, CA, USA. (Leica Microsystems). Tumor sections were blocked with Peroxidase Block (Dako, Carpinteria, CA, USA) and Avidin- Biotin Block (Vector Laboratories, Burlingame, CA, USA). Signaling studies and flow cytometry 5 Sections were first labeled with mouse anti-human p63 at NHL cells (3 Â 10 cells/ml in RPMI þ 2% fetal bovine serum), 0.07 mg/ml (Abcam), and detected using alkaline phosphatase- were treated with crosslinked hIgG, dacetuzumab or rhCD40L Fast Red. Next, sections were labeled with biotinylated goat (R&D Systems, Minneapolis, MN, USA), and collected by anti-human Fc at 5 mg/ml (Jackson ImmunoResearch Labora- centrifugation. Cells were lysed in protein extraction buffer tories), and detected with streptavidin-HRP/DAB. Tumor sec- (50 mM Tris (pH ¼ 7.5), 1% NP-40, 150 mM NaCl, 1 mM EDTA, tions were imaged using an Axiovert 200M fluorescence 1mM EGTA (pH ¼ 7.6), phosphatase inhibitors (EMD Bios- microscope (Zeiss, Oberkochen, Germany), then TAp63a- ciences San Diego, CA, USA) and protease inhibitor tablets specific staining was spectrally unmixed and quantitated using (Roche, Basel, Switzerland)), then the cell lysates were the Nuance multispectral imaging system (Cambridge Research centrifuged and the protein concentrations were measured. & Instrumentation, Woburn, MA, USA). Activation of MAP kinase, AKT and NF-kB pathways was detected by western blotting with antibodies specific to phospho-p38 MAP kinase (Thr180/Tyr182), phospho-p44/42 NHL xenograft studies MAP kinase (Thr202/Tyr204), phospho-JNK (Thr183/Tyr185), Severe combined immunodeficient mice were implanted with phospho-Akt (Ser473), IkB-a and IRF4 (Cell Signaling Technol- Ramos or SU-DHL-8 cells (5 Â 106 per mouse) then dosed ogy, Beverly, MA, USA). Survival and apoptosis-related proteins when average tumor volume reached 300 mm3 or 70 mm3,

Leukemia Dacetuzumab proapoptotic signaling mechanisms in NHL TS Lewis et al 1009 respectively. Dacetuzumab and rituximab were dosed at 4.0 mg/kg the kinetics of BCL-6 downregulation by dacetuzumab in the in the Ramos model and at 10.0 mg/kg in the SU-DHL-8 model. Ramos cell line. Dacetuzumab treatment suppressed BCL-6 The CHOP chemotherapy regimen (cyclophosphamide, 30 mg/ protein levels in a sustained manner, with loss of expression first kg (q1d  1), i.p.; doxorubicin, 2.475 mg/kg (q1d  1), i.p.; detected at 8 h after treatment and maintained beyond 48 h vincristine, 0.375 mg/kg (q1d  1), i.p.; prednisone, 0.15 mg/kg (Figure 1b). Significantly, the onset of dacetuzumab-mediated (q1d  5), p.o.) was modified from Mohammad et al.21 ICE apoptosis corresponded with the timing of BCL-6 downregula- chemotherapy consisted of ifosfamide, 10 mg/kg (q1d  5), i.p.; tion (Supplementary Figure 1a). Loss of BCL-6 protein expres- mesna, 10 mg/kg (q1d  5), i.p.; carboplatin, 15 mg/kg (q1d  5), sion by ubiquitin-mediated proteolysis has been reported i.p.; etoposide, 10 mg/kg (q4d  3), i.p. Gemcitabine was dosed following activation of the B-cell receptor and phosphorylation at 100 mg/kg (q4d  3), i.p. Tumor progression rates were reported of sites within the PEST domains by ERK1/2 MAP kinases.25 as percent of mice having a tumor volume less than a defined end Examination of dacetuzumab-stimulated MAP kinase signaling point value over time, using Kaplan–Meier plots. P-values were kinetics revealed that the ERK1/2 pathway is constitutively generated using the log-rank (Mantel–Cox) test. All animal studies activated, and that the level of signaling significantly increases are representative of two independent experiments. over the course of several days, peaking at 43 h (Figure 1b). Dacetuzumab also activated the p38 MAPK pathway in a sustained manner, whereas activation of JNK signaling was Results biphasic, reaching peaks of activity at 2 h followed by reactivation at 18–43 h. BCL-6 downregulation in normal B Dacetuzumab is a signaling antibody in NHL cells and some lymphomas has been reported following NF-kB- Dacetuzumab produces an anti-proliferative and apoptotic mediated induction of IRF4, which transcriptionally represses response in cultured CD40 þ NHL cells, with increasing cell BCL-6 expression.23 Dacetuzumab activation of NF-kB signal- death occurring at 24 and 48 h after treatment in the Ramos cell ing was also sustained over several days as detected by line (Supplementary Figure 1a). To elucidate the mechanisms degradation of IkBa (Figure 1b, lower panels). These data behind dacetuzumab-mediated anti-proliferative and apoptotic indicated that the downregulation of BCL-6 by dacetuzumab in activity, we tested the ability of dacetuzumab to stimulate cell NHL cell lines correlates with its ability to constitutively activate signaling pathways following CD40 binding. Dacetuzumab, signaling through MAPK and NF-kB pathways. crosslinked with anti-human Fc antibody, activated the p42/44 To further link dacetuzumab signaling with BCL-6 down- (ERK1/2) and p38 MAP kinase pathways within 15 min in the RL regulation, pharmacological inhibitors of MAPK pathway lymphoma cell line relative to soluble dacetuzumab (Supple- activation were tested for their ability to prevent dacetuzu- mentary Figure 1b, lanes 5–7 versus lanes 2–4). NF-kB signaling mab-mediated suppression of BCL-6 expression. Pretreatment was strongly activated by dacetuzumab as detected by loss of its with the MKK1/2 inhibitor (U0126) or JNK inhibitor VIII both negative regulatory protein IkB-a to proteasomal degradation. partially blocked dacetuzumab-mediated BCL-6 degradation in Dacetuzumab produced only a minor increase in signaling Ramos and SU-DHL-4 cell lines (Figure 1c). Combining these through the AKT pathway. The enhanced signal transduction two inhibitors produced a greater protection of BCL-6 protein in observed with crosslinked dacetuzumab likely reflects the Ramos cells, suggesting the ERK1/2 and JNK MAPK pathways requirement for CD40 to multimerize on the cell surface and are important to the downregulation of BCL-6 by dacetuzumab. form active signaling complexes at the cytoplasmic domain In contrast, inhibition of the p38 MAPK pathway failed to through TNFR-associated factor proteins, a characteristic of prevent BCL-6 downregulation. Next, we examined the TNFR superfamily members.9 The signaling profile of dacetu- significance of ERK1/2 MAPK signaling and BCL-6 down- zumab was very similar to trimeric human CD40 ligand in regulation for dacetuzumab-mediated apoptosis. The percent NHL cells (Supplementary Figure 1b, lane 11). Dacetuzumab cell death induced by dacetuzumab was measured±U0126 in produced a dose-dependent anti-proliferative response in the BCL-6 þ lymphoma cell line RL (Figure 1d). Dacetuzumab Ramos cells, which correlated with its activation of ERK1/2 alone or with dimethyl sulfoxide carrier produced B50% total MAPK pathway signaling within a similar low ng/ml concentra- cell death, whereas RL cells pretreated with U0126 before tion range (Supplementary Figures 1a,c). addition of dacetuzumab had significantly reduced cell death (P ¼ 0.015; unpaired t-test). Increased RL cell viability correlated Dacetuzumab signaling downregulates BCL-6 with block of ERK1/2 activation and stabilization of BCL-6 oncoprotein protein from degradation (Figure 1d, lower panels). Similarly, in In normal B cells, BCL-6 proto-oncogene mRNA is down- the SU-DHL-4 BCL-6 þ cell line U0126 produced a partial but regulated from peak levels in the hyperproliferative GC significant (P ¼ 0.033) reduction of dacetuzumab-induced following CD40–CD40L signaling.22,23 Therefore, we tested apoptosis (Supplementary Figure 2a). IRF4 expression was not the hypothesis that dacetuzumab signaling would similarly affected by dacetuzumab in the RL, HT and Ramos cell lines, downregulate BCL-6 in GC-derived lymphomas in which the whereas dacetuzumab weakly induced IRF-4 expression in SU- BCL-6 gene is deregulated because of chromosomal transloca- DHL-4 cells (Figure 1d and Supplementary Figure 2b). tion or mutation in the 50-untranslated region. Three BCL-6- Collectively these data show that constitutive activation of expressing NHL cell lines Ramos (Burkitt’s lymphoma), MAP kinase signaling and the subsequent downregulation of SU-DHL-4 and RL (DLBCLs) and one BCL-6À DLBCL cell line BCL-6 protein is an important mechanism of dacetuzumab- (Toledo) were treated with dacetuzumab or hIgG control, and induced apoptosis in GC-derived lymphoma cell lines. the effect on BCL-6 protein level was examined by western blotting. Dacetuzumab treatment strongly downregulated BCL-6 protein in the three BCL-6 þ cell lines, almost completely Dacetuzumab induces plasma cell differentiation genes reducing detectable protein by 48 h (Figure 1a). and downregulates c-Myc BCL-6 is implicated in lymphomagenesis and its sustained BCL-6 transcriptional repressor activity prevents the differentia- inhibition or downregulation may produce an anti-proliferative tion of normal GC B cells to plasma cells until somatic or apoptotic effect in GC lymphomas.24 Therefore, we examined hypermutation is complete.26,27 Similarly in germinal center

Leukemia Dacetuzumab proapoptotic signaling mechanisms in NHL TS Lewis et al 1010

+++++ ++++ Ramos SU-DHL-4 RL Toledo Dacetuzumab 1 2 3 1 2 3 1 2 3 1 2 3 BCL-6 BCL-6 -Actin -Actin

Ramos SU-DHL-4

Dacetuzumab Signaling Kinetics (Ramos) RL cells 2 4 8 18 24 32 43 48 56 Hours 1 Control hIgG CDCDCDCDCD CDCDCDC D 50 2 Dacetuzumab BCL-6 3 Dacetuzumab + DMSO 40 4 U0126 5 Dacetuzumab + U0126  -Actin 30 6 Untreated

20 p-ERK1/2 Cell Death (%) 10

p-JNK 0 123456 p-p38 123456 IB- p-ERK1/2 -Actin BCL-6

IRF-4

-Actin

Figure 1 Dacetuzumab downregulates BCL-6 through sustained MAPK signaling and dacetuzumab-mediated apoptosis is ERK1/2 MAP kinase dependent. (a) Ramos, SU-DHL-4, RL and Toledo cells were either untreated (lane 1) or treated with crosslinked control hIgG (lane 2) or crosslinked dacetuzumab (lane 3) for 48 h at 1.0 mg/ml. BCL-6 protein expression was detected by western blotting, and b-actin was analyzed as a protein load control. (b) Ramos cells were treated with crosslinked hIgG control ‘C’ or dacetuzumab ‘D’ at 1.0 mg/ml over a 56 h time course, and BCL-6 protein levels were observed by western blotting (top panels). Under identical conditions, the kinetics of ERK1/2, JNK and p38 MAP kinase pathway activation by dacetuzumab was examined using phospho-p44/42 MAPK (Thr202/Tyr204), phospho-JNK MAPK (Thr183/Tyr185) and phospho-p38 MAPK (Thr180/Tyr182)-specific antibodies, respectively, (lower panels). Constitutive activation of the NF-kB signaling pathway was detected by western blotting for IkB-a.(c) Ramos and SU-DHL-4 cell lines were pretreated for 1 h with U0126 MKK1/2 inhibitor, JNK inhibitor VIII, or p38 inhibitor III (10 mM), then treated with dacetuzumab (1.0 mg/ml) for 36 h. BCL-6 protein expression was detected by western blotting. (d)RL cells were pretreated for 1 h with U0126 (10 mM) and then treated±dacetuzumab (1.0 mg/ml) for 69 h. Total cell death for each treatment condition was determined by flow cytometry using the Annexin V-FITC apoptosis detection assay and standard error (s.e.m.) reported. (Lower panel) Activation of the ERK1/2 MAPK pathway by dacetuzumab±U0126 and the effect on BCL-6 and IRF4 protein expression were examined by western blotting. RL cells were collected at 40 h post dacetuzumab treatment in parallel with the apoptosis studies.

B-cell-like (GCB) subtype lymphomas, BCL-6 deregulation of the NHL lines in which BCL-6 expression was lost and Blimp- maintains cells in a hyperproliferative non-differentiated state, 1 expression was activated, whereas only a minor decrease was unresponsive to apoptotic stimuli such as DNA damage.28,29 observed in the Toledo cell line (Figure 2b, lane 3). Down- Therefore, we tested whether dacetuzumab-mediated down- regulation of c-Myc was observed in one cell line with a regulation of BCL-6 affected downstream differentiation and MYC-IG translocation (Ramos) and two cell lines lacking proliferation pathways in NHL cells. Reverse transcriptase-PCR MYC-IG translocation (RL and SU-DHL-4) (Supplementary was used to detect transcript levels of Blimp-1 and XBP-1, key Table 1). These data suggest that BCL-6-expressing lymphoma transcription factors that control plasma cell differentiation, cells can be driven by dacetuzumab-activated signaling into a which are negatively regulated by BCL-6. In the three BCL-6 þ hypoproliferative state with early characteristics of plasma cell lymphoma lines examined (Ramos, SU-DHL-4 and RL), differentiation via sustained downregulation of BCL-6, as dacetuzumab treatment upregulated Blimp-1 mRNA levels by hypothesized in Figure 2c. 1.7-, 3.8- and 5.4-fold, respectively, relative to control hIgG (Figure 2a, lane 3). However, in the BCL-6À cell line (Toledo) Blimp-1 transcript was constitutively expressed and unaffected Dacetuzumab upregulates the proapoptotic p53 family by dacetuzumab treatment. Similarly, unspliced XBP-1 mRNA member, TAp63a was moderately upregulated by dacetuzumab in SU-DHL-4 We have previously shown that the cytotoxic, anti-proliferative (2.5-fold) and RL (1.5-fold) cell lines, but not in Toledo. Total effect of dacetuzumab is not restricted to BCL-6 þ lymphoma levels of XBP-1 message were not altered in Ramos cells, rather cell lines, suggesting that dacetuzumab signaling can mediate dacetuzumab signaling produced a twofold accumulation of the antitumor response through mechanisms independent of BCL-6. fully spliced active form of XBP-1 (lower band). Another feature CD40L signaling has been shown to regulate the expression of of normal B-cell differentiation is loss of c-Myc expression, a the p53-related transcription factor p73,30 therefore, we target of Blimp-1 transcriptional repression activity. Dacetuzu- examined the effect of dacetuzumab signaling on p53 family mab treatment downregulated c-Myc protein expression in each members. In Ramos cells, dacetuzumab produced a minor

Leukemia Dacetuzumab proapoptotic signaling mechanisms in NHL TS Lewis et al 1011 Dacetuzumab Ramos SU-DHL-4 RL Toledo CD40 132 13 2 13 2 13 2 ERK1/2 - Blimp-1 MAPKs unspliced BCL-6 XBP-1 spliced -- GAPDH Blimp-1 Ramos SU-DHL-4 RL Toledo -

132 13 2 132 13 2 XBP-1 c-Myc c-Myc + + -Actin Hypoproliferation, Hyperproliferation, Early Differentiation Survival

Figure 2 Dacetuzumab upregulates Blimp-1 and XBP-1 plasma cell differentiation regulatory genes, but downregulates c-Myc. (a) Ramos, SUDHL-4, RL and Toledo NHL cells (3 Â 106) cultured in RPMI þ 2% fetal bovine serum were either untreated (lane 1), treated with crosslinked control hIgG (lane 2) or treated with crosslinked dacetuzumab (lane 3) at 2.0 mg/ml for 48 h. Blimp-1 and XBP-1 mRNA expression was analyzed by RT-PCR with glyceraldehyde 3-phosphate dehydrogenase included as a normalization control. Unspliced XBP-1 mRNA (upper band) and its active splice variant (lower band) encoding the transactivating domain were detected in Ramos and Toledo cell lines. (b) Downregulation of c-Myc protein by dacetuzumab was detected by western blotting. Treatments conditions and corresponding lane numbers are the same as described above. (c) A model of dacetuzumab’s signaling mechanism of action in BCL-6 þ NHL cell lines. Dacetuzumab binding to CD40 activates ERK1/2 signaling and suppresses BCL-6 expression. Loss of BCL-6 reverses the transcriptional repression of plasma cell differentiation genes Blimp-1 and XBP-1 and inhibits cell proliferation by downregulating c-Myc.

Kinetics of TAp63 Induction (Ramos) tially lower TAp63a levels in response to dacetuzumab 2481824 32 43 48 56 72 Hours treatment. Endogenous TAp63a levels were below the lower CDCDCDCDCD C DCDCDCDCD limit of detection by western blotting. Transcriptional regulation p53 of p63 isoforms reportedly involves the NF-kB pathway, and a TAp63 single NF-kB-binding site has been identified within the p63 promoter.31 To determine whether dacetuzumab activation of -Actin the NF-kB pathway is important for upregulation of TAp63a,we used a potent small-molecule NF-kB transcriptional activation Ramos RLSU-DHL-4 SU-DHL-8 inhibitor. Pretreatment of Ramos cells with NF-kB transcrip- 123123 123 123 tional activation inhibitor produced a dose-dependent decrease  TAp63 in TAp63a protein levels induced by dacetuzumab (Figure 3c). -Actin Dacetuzumab stimulation of the NF-kB pathway is therefore required for complete upregulation of TAp63a.

- - NF-κB Activation Inhibitor (nM) C ++++++Dacetuzumab TAp63 Dacetuzumab has chemosensitization activity in NHL p53 is frequently mutated in NHL cell lines (Supplementary -Actin Table 1), and dacetuzumab activation of TAp63a expression may partially restore p53-regulated proliferation and apoptosis Figure 3 Dacetuzumab upregulates the proapoptotic p53 family checkpoints, such as response to cytotoxic drugs. To explore this member TAp63a in a NF-kB-dependent manner. (a) Expression of p53 possibility, dacetuzumab was evaluated in combination with and TAp63a proteins was monitored by western blotting following treatment with dacetuzumab over a 72 h time course. Ramos cells DNA-damaging chemotherapeutic drugs in a NHL cell line with were treated with 1.0 mg/ml crosslinked hIgG control ‘C’ or dacetu- high levels of inducible TAp63a. Combination index analysis 32 zumab ‘D’. (b) Ramos, RL, SU-DHL-4 and SU-DHL-8 NHL cell lines performed using the Chou–Talalay method in the SU-DHL-8 were untreated (lane 1), treated with hIgG control (lane 2), or treated cell line showed that dacetuzumab is moderately synergistic with dacetuzumab (lane 3) for 24 h (1.0 mg/ml), and TAp63a (CI 0.7–0.85) with chlorambucil, doxorubicin and etoposide expression was detected by western blotting. (c) Ramos cells were and synergistic (CI 0.3–0.7) with bleomycin and carboplatin pretreated (1 h) with a 10-fold dilution series of NF-kB activation inhibitor and then treated with 1.0 mg/ml dacetuzumab or control hIgG (Figure 4a and Supplementary Figure 3a). These data indicate ‘C’ for 24 h. TAp63a protein expression was detected by immunoblot- that dacetuzumab has chemosensitization activity. TAp63a 33 ting. expression is induced by a variety of chemotherapeutic drugs, therefore we compared the ability of dacetuzumab to upregulate TAp63a with that of cytotoxic drugs. Dacetuzumab produced a upregulation of p53 protein; however, a transcriptionally active far stronger induction of TAp63a in Ramos cells than either form of p63 (TAp63a) was strongly induced by dacetuzumab at chlorambucil or gemcitabine at 20 mM or 20 nM, respectively 18–72 h after treatment (Figure 3a). Other NHL cell lines treated (Figure 4b, lanes 2–4 versus lanes 7 and 10). The ability of with dacetuzumab also showed inducible expression of dacetuzumab to induce TAp63a to higher levels than observed TAp63a, with the amount of upregulation varying between cell with chemotherapeutic drugs alone is one potential mechanism lines (Figure 3b). The highest induction was observed in Ramos through which NHL cells are sensitized by dacetuzumab to and SU-DHL-8 cells, whereas SU-DHL-4 and RL had substan- cytotoxic damage.

Leukemia Dacetuzumab proapoptotic signaling mechanisms in NHL TS Lewis et al 1012 Synergistic Etoposide Dacetuzumab Chlorambucil Gemcitabine

Doxorubicin 12345678910

Chlorambucil TAp63

Carboplatin

Bleomycin -Actin

0.0 0.5 1.0 1.5 Combination Index (CI)

0 hrs 24 hrs 48 hrs 450 450 450 400 400 400 hIgG (isotype control) 350 350 350 300 300 300 300 Dacetuzumab (isotype control) 250 250 250 200 200 200 Fas hIgG (anti-Fas) Events 150 150 Fas 150 150 100 100 Dacetuzumab (anti-Fas) 100 50 50 50 0 0 0 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 Fluorescence Intensity

2 4 81824 32 42 48 56 72 Hours CDCCCC D D D D CDCCCC D D D D

BimEL

BimL

BimS

-Actin

Figure 4 Dacetuzumab is synergistic with DNA-damaging drugs and induces proteins in the extrinsic and intrinsic apoptotic pathways. (a) SU- DHL-8 cells were treated with crosslinked dacetuzumab±chemotherapy drug for 96 h and proliferation measured by 3H-thymidine incorporation. Dose effect curve data from three independent experiments was analyzed to derive combination index (CI) values using CalcuSyn software. CI o0.9 is synergistic, CI41.1 is antagonistic and CI ¼ 0.9–1.1 is additive. (b) Ramos cells were treated with control hIgG (lane 1) or with threefold dilutions of crosslinked dacetuzumab (0.11, 0.33 and 1.0 mg/ml; lanes 3–4), chlorambucil (2.2, 6.7, and 20 mM; lanes 5–7) and gemcitabine (2.2, 6.7 and 20 nM; lanes 8–10) for 24 h. Induction of TAp63a protein was detected by western blotting. (c) Ramos cells were treated with crosslinked dacetuzumab or control hIgG for 24 or 48 h and Fas expression was analyzed by flow cytometry. (d) Expression of the three major Bim isoforms (BimEL, BimL and BimS) was examined by western blotting following hIgG control ‘C’ or dacetuzumab ‘D’ treatment of Ramos cells (1.0 mg/ml) over a 72 h time course.

Analysis of TAp63a-regulated genes in hepatocellular carcino- and R-Gem in xenograft models of NHL. Kaplan–Meier tumor ma has identified apoptosis-associated genes positively regulated progression plots showing time to reach a tumor volume end by this transcription factor, including the death receptor Fas point were used to present in vivo efficacy data in order to more (CD95) and the proapoptotic BCL-2 family member Bim.33 Fas accurately present statistical differences between the groups was strongly upregulated on the surface of Ramos cells 24–48 h (Figures 5a–c); corresponding tumor volume plots for each study after treatment with dacetuzumab, consistent with the kinetics of are provided in Supplementary Figure 4. Dacetuzumab alone TAp63a induction (Figure 4c). In addition, dacetuzumab induced delayed the growth of subcutaneous Ramos tumors in the severe protein expression of BimEL,BimL and BimS isoforms in Ramos combined immunodeficient mice, and dacetuzumab had greater cells 18–72 h after treatment (Figure 4d). The BimL isoform was efficacy than rituximab, CHOP and ICE (Figures 5a and b). The upregulated by 2.8-fold above endogenous levels 72 h after antitumor activity of R-CHOP was significantly improved when treatment. A second proapoptotic BH3-only BCL-2 family combined with dacetuzumab (D þ R-CHOP) in the Ramos member, Bid, was similarly induced by dacetuzumab (Supple- model (P ¼ 0.0005), producing complete remissions in 8/10 mentary Figure 3b). Therefore, dacetuzumab positively regulates animals (Figure 5a). Direct comparison of D-CHOP antitumor proapoptotic components in both the mitochondrial and death activity with that of R-CHOP demonstrated that dacetuzumab receptor pathways in NHL cell lines. potentiates the activity of this chemotherapy regimen more strongly than rituximab in the Ramos model (P ¼ 0.0052). Furthermore, the efficacy observed in the D þ R-CHOP combi- Dacetuzumab augments NHL chemotherapeutic nation exceeded the dacetuzumab þ rituximab combination, regimens suggesting that dacetuzumab-mediated chemosensitization of Frontline, second-line and third-line DLBCL therapies consist of Ramos cells to CHOP, not augmented antibody-dependent rituximab combined with CHOP (R-CHOP), ICE (R-ICE) and cellular cytotoxicity effector activity, is the dominant mechan- gemcitabine (R-Gem) chemotherapy regimens, respectively. To ism. The non-steroidal cytotoxic components of CHOP were determine whether dacetuzumab potentiates standard lympho- individually combined with dacetuzumab in the Ramos model, ma therapies in vivo, we evaluated the antitumor activity of and the DNA-damaging drug cyclophosphamide was the most dacetuzumab alone and in combination with R-CHOP, R-ICE active agent (Supplementary Figure 5).

Leukemia Dacetuzumab proapoptotic signaling mechanisms in NHL TS Lewis et al 1013 Dacetuzumab + R-CHOP (Ramos) Untreated CHOP 100 Untreated

3 Dacetuzumab 80 R x Rituximab 60 Dacetuzumab + Rituximab CHOP 40 D + CHOP (2 cycles: Day-1, 17) R + CHOP (2 cycles: Day-1, 17) hIgG hIgG + CHOP 20

Tumor < 1,000 mm D + R-CHOP (2 cycles: Day-1, 17) Percent Animals with

0 R-CHOP vs. D + R-CHOP 0 10 20 30 40 50 60 70 80 90 100 110 P = 0.0005 Days Post Tumor Implant R-CHOP vs. D-CHOP P = 0.0052

Dacetuzumab + R-ICE (Ramos) Dacetuzumab Dacetuzumab + CHOP 100 Untreated

3 80 Dacetuzumab Rx Rituximab 60 ICE Rituximab + ICE (R-ICE) 40 Dacetuzumab + ICE Dacetuzumab + R-ICE 40000 20 Tumor < 1,000 mm Percent Animals with Dacetuzumab vs. D + ICE 30000 0 P = 0.0019 0 10 20 30 40 50 60 70 80 90 100110 120 R-ICE vs. D + R-ICE P < 0.0001 20000 Days Post Tumor Implant Expression

 Dacetuzumab + R-Gem (SU-DHL-8) 10000

100 (Optical Density) Untreated TAp63 Rituximab 0 80 R x Dacetuzumab hlgG 60 Gemcitabine CHOP Untreated D + CHOP Dacetuzumab + Rituximab 40 DacetuzumabhlgG + CHOP Rituximab + Gemcitabine (R-Gem) 1 234567891011 20 Dacetuzumab + Gemcitabine Dacetuzumab + R-Gem TAp63

Increase in Tumor Volume 0 Gem vs. D-Gem

Percent Animals with < 8-Fold 0 5 10 15 20 25 30 35 P= 0.0056 Days Post Tumor Implant R-Gem vs. D + R-Gem -Actin P < 0.0001

Figure 5 Dacetuzumab improves the activity of rituximab-based chemotherapeutic regimens and induces TAp63a in vivo.(a) SCID mice (n ¼ 10 per group) bearing subcutaneous Ramos tumors (300 mm3) were dosed (arrow) with dacetuzumab (4 mg/kg; q4d  8; i.p.), rituximab (4 mg/kg; q4d  8; i.p.), or CHOP chemotherapy alone or in combination. Two cycles of CHOP were administered in the D-CHOP, R-CHOP and D-R- CHOP groups as indicated. (b) SCID mice (n ¼ 10 per group) with subcutaneous Ramos tumors (300 mm3) were dosed with dacetuzumab (4 mg/kg; q4d  4; i.p.), rituximab (4 mg/kg; q1d  1; i.p.) or ICE chemotherapy alone or in combination. (c) SCID mice (n ¼ 10 per group) subcutaneously implanted with SU-DHL-8 tumor cells were dosed with dacetuzumab (10 mg/kg; q4d  4; i.p.), rituximab (10 mg/kg; q4d  4; i.p.) or gemcitabine (100 mg/kg; q4d  3; i.p.) alone or in combination when average tumor volume reached 70 mm3. Tumor progression was plotted as percent of mice with less than eightfold increase in tumor volume (o560 mm3) over time. (d) (Top panel) SCID mice with subcutaneous Ramos tumors (250 mm3) were dosed with dacetuzumab or hIgG (4 mg/kg; q1d  1; i.p.)±CHOP and collected at 96 h after treatment. Formalin-fixed, paraffin- embedded tumor sections underwent immunohistochemistry staining for nuclei (blue), tumor bound antibody (brown) and TAp63a (magenta).  40 bright field images were acquired with the Nuance multispectral imaging system. Arrows indicate representative TAp63a-positive tumor cell nuclei, with regions of interest enlarged (boxes). (Middle panel) TAp63a tumor expression (optical density) was quantitated from n ¼ 8 images from two tumors per treatment condition; s.e.m. reported. (Bottom panel) TAp63a expression in Ramos tumors was detected by western blotting protein extracts from homogenized tumors (n ¼ 2 per treatment group). Untreated (lane-1), CHOP (lanes 2–3), hIgG (lanes 4–5), dacetuzumab (lanes 6–7), hIgG þ CHOP (lanes 8–9) and dacetuzumab þ CHOP (lanes 10–11).

Similarly, dacetuzumab potentiated the antitumor activity of model, whereas gemcitabine delayed tumor progression without ICE chemotherapy in the Ramos model, and D-ICE efficacy was producing complete remissions (Figure 5c). In combination with greater than R-ICE (Figure 5b). The antitumor activity of R-ICE gemcitabine, dacetuzumab significantly improved the activity of was significantly improved in combination with dacetuzumab this cytotoxic drug in vivo (P ¼ 0.0056). Furthermore, the (D þ R-ICE) (Po0.0001), producing complete remissions in 3/10 efficacy of R-Gem was significantly improved in combination animals. D þ R-ICE was administered in a single cycle, not with dacetuzumab (D þ R-Gem) (Po0.0001), producing non- allowing for a direct comparison with D þ R-CHOP efficacy, durable complete remissions in 7/9 animals. which was dosed in two cycles. The ability of dacetuzumab to augment the activity of Gemcitabine alone produced durable complete remissions in chemotherapeutic regimens above the level achieved with the Ramos model, therefore we used a more chemoresistant rituximab suggests that dacetuzumab has in vivo chemosensi- DLBCL xenograft model, SU-DHL-8, to examine the activity of tization activity. Furthermore, in vitro data indicated that in dacetuzumab in combination with R-Gem. SU-DHL-8 subcuta- certain NHL cell lines TAp63a upregulation might be important neous tumors are extremely aggressive with a tumor quadru- for increased response to cytotoxic drugs (Figures 4a,b). We pling time of B4 days. Dacetuzumab and rituximab dosed therefore tested whether TAp63a is upregulated in vivo alone or in combination had minimal antitumor activity in this following dosing with dacetuzumab in the Ramos xenograft

Leukemia Dacetuzumab proapoptotic signaling mechanisms in NHL TS Lewis et al 1014 model. Immunohistochemistry staining of TAp63a was per- explaining how dacetuzumab signaling through CD40 produces formed on Ramos tumor sections collected at 96 h after an anti-proliferative, proapoptotic response in NHL cells as a treatment with CHOP chemotherapy, control hIgG±CHOP or single agent and in combination with cytotoxic drugs. dacetuzumab±CHOP (Figure 5d). As expected, dacetuzumab The POZ/zinc-finger transcriptional repressor BCL-6 is a key localized to the cell membrane of Ramos tumor cells, and the regulator of GC formation, B-cell differentiation and survival.34 strongest anti-human Fc staining was detected at the tumor Activation of CD40-CD40L signaling in immature B cells in the periphery or near blood vessels. TAp63a is a nuclear protein, GC activates a differentiation program to halt somatic hyper- and anti-TAp63a staining was detected in a small population of mutation, class switching and proliferation and initiate B-cell lymphoma cell nuclei in control groups. However, in the outer maturation processes. By suppressing p53 checkpoint and the region of the Ramos tumors in which dacetuzumab staining was DNA-damage sensor ATR, BCL-6 protects developing B cells strongest, the number of lymphoma cells expressing nuclear during somatic hypermutation and Ig class switching from TAp63a was distinctly elevated. Quantitation of the TAp63a apoptosis induced by DNA damage.27,28 Similarly, in GC- chromaphore (fast red) with a multispectral imaging camera derived lymphomas, including Burkitt’s lymphoma, DLBCL and demonstrated that TAp63a levels were 43.5-fold higher in follicular lymphoma; BCL-6 maintains a non-differentiated dacetuzumab treated tumors±CHOP versus control treatment hyperproliferative state analogous to a proliferating B cell in groups. The in vivo upregulation of TAp63a was confirmed by the GC. Mouse models mimicking the t(3;14) (q27:q32) western blotting of Ramos tumor extracts (Figure 5d, lower translocation of BCL-6 found in certain DLBCL patients have panels). These data are direct evidence that dacetuzumab can been used to implicate BCL-6 deregulation in the formation of stimulate signal transduction in lymphoma cells in vivo, B-cell lymphomas.35,36 BCL-6 is an appealing new target in the inducing a protein with proapoptotic and chemosensitization treatment of NHL, which is currently being targeted by activities. inhibitory peptides, acetylation inhibitors, Hsp90 inhibitors and RNA interference.37,38 In a collaborative research effort, dacetuzumab has been shown to have greater cytotoxic activity Discussion against NHL cell lines of the BCL-6 þ GCB subtype than those of the BCL-6À activated B-cell (ABC) subtype.39 Our finding that We previously showed that the anti-CD40 monoclonal antibody dacetuzumab downregulates BCL-6 in a sustained manner dacetuzumab triggers apoptosis in NHL cell lines.19 In this suggests that anti-CD40 signaling antibodies may be a novel study, we identified two mechanisms that contribute to therapeutic approach to overcome this key oncogenic event in dacetuzumab-mediated apoptosis and arrested proliferation in GCB subtype lymphomas. NHL cells; summarized in (Figure 6). We demonstrate that The capability of CD40-mediated signaling to downregulate dacetuzumab is a signaling antibody, capable of sustained BCL-6 is likely due to both transcriptional and post-translational activation of the ERK1/2, JNK, p38 MAP kinase and NF-kB modification mechanisms. BCL-6 is reported to be transcription- signaling cascades. Furthermore, constitutive dacetuzumab ally downregulated following CD40L signaling in GC B cells signaling through MAPK pathways downregulated BCL-6 and some lymphoma cell lines through NF-kB upregulation of expression in GC-derived lymphoma cells, inducing B-cell the IRF4 transcription factor.23 However, our observations differentiation markers and ultimately triggering apoptosis. In indicate that BCL-6 downregulation by dacetuzumab can occur addition, dacetuzumab signaling chemosensitized p53-mutant in the absence of IRF4 upregulation. Furthermore, mutations or NHL cells concurrent with the upregulation of the p53 family chromosomal translocations blocking the IRF4 binding site member TAp63a. These observations provide a working model within the BCL-6 promoter prevents downregulation of BCL-6

Dacetuzumab FAS Bim CD40 Bax + Bid Chemotherapy

IB

NF-B

26S Proteasome PPP - + 1) Intrinsic BCL-6 TAp63 2) Extrinsic Apoptosis Differentiation, Prosurvival Genes Proapoptotic Genes Blimp-1, XBP-1, c-Myc Fas, Bim, Bid

Figure 6 A model of dacetuzumab-mediated proapoptotic signaling and chemosensitization in NHL cell lines. Dacetuzumab binding to CD40 triggers sustained intracellular signaling through MAP kinase and NF-kB signaling pathways mediated by cytoplasmic TNFR-associated factor (TRAF) proteins. Constitutive phosphorylation of BCL-6 by ERK1/2 MAPKs activates ubiquitin-mediated proteolysis. BCL-6 downregulation in GC-derived lymphomas relieves the transcriptional repression of Blimp-1 and XBP-1 plasma cell differentiation genes and reverses c-Myc-driven hyperproliferation. Upregulation of the p53 family member TAp63a by dacetuzumab is BCL-6 independent and is reliant upon NF-kB pathway signaling. The downstream proapoptotic TAp63a transcriptional targets Fas and Bim are upregulated. Loss of the BCL-6 survival program and its repression of DNA damage checkpoint resensitizes lymphoma cells to apoptotic stimuli such as cytotoxic drugs. Similarly, in p53 pathway- defective lymphoma cells, upregulation of TAp63a may restore sensitivity to DNA damage caused by chemotherapeutic drugs.

Leukemia Dacetuzumab proapoptotic signaling mechanisms in NHL TS Lewis et al 1015 by CD40 signaling in a subset of DLBCL. BCL-6 phosphorylation xenograft models, improving the activity of rituximab in within its PEST domains by ERK1/2 MAP kinases following combination with CHOP, ICE and gemcitabine. This chemo- activation of B-cell receptor signaling leads to its proteasomal sensitization activity can be accounted for by mechanisms degradation.25 Our data shows that dacetuzumab-mediated utilizing BCL-6 depletion and/or induced expression of p53 apoptosis and downregulation of BCL-6 can be attenuated by family members. Upregulation of TAp63a may substitute for p53 inhibiting the ERK1/2 and JNK MAPK pathways, suggesting that DNA damage response in lymphoma cells with mutant or dacetuzumab uses a post-translational mechanism to suppress deleted p53, improving the potency of DNA-damaging drugs. In BCL-6 expression. Therapeutically downregulating BCL-6 ex- addition, CD40-mediated downregulation of BCL-6 restores pression through sustained activation of MAPK signaling is sensitivity to DNA damage in lymphoma cells,27 and dacetu- advantageous in those GCB subtype DLBCLs in which the BCL-6 zumab may similarly sensitize lymphoma cells to cytotoxic promoter is mutated and cannot be downregulated through IRF4 drugs by constitutively downregulating BCL-6 expression. transcriptional repression alone. Interestingly, clinical data suggest that BCL-6 þ lymphoma Our data suggests that dacetuzumab downregulation of BCL-6 patients did not benefit from the addition of rituximab to CHOP drives lymphoma cells into a hypoproliferative state resensitized chemotherapy, whereas BCL-6À patients showed improved to apoptotic stimuli analogous to the differentiation signal response.50 This raises the possibility an anti-CD40 signaling CD40–CD40L signaling triggers on GC B cells. MAPK signaling antibody, such as dacetuzumab, could provide additional is often associated with mitogenic effects, nevertheless, there are benefit to BCL-6 þ NHL patients. documented instances in which sustained activation of MAPK In summary, we provide the first mechanistic data describing signaling induces transformed cells to halt proliferation and how dacetuzumab signaling through CD40 produces anti- undergo partial differentiation,40,41 and MEK1 has recently been proliferative, apoptotic and chemosensitization responses in identified as a tumor suppressor in a mouse lymphoma model.42 NHL cells. Dacetuzumab’s ability to counteract oncogenic The BCL-6-controlled transcription factors Blimp-1 and XBP-1 events implicated in lymphomagenesis distinguishes it from are important regulators of plasma cell differentiation,43–46 and other NHL therapeutic antibodies with effector function based both were found to be upregulated in BCL-6 þ cell lines or ligand blocking mechanisms of action. following dacetuzumab treatment. Furthermore, repression of c-Myc by Blimp-1 is another important step in normal B-lymphocyte differentiation.47 c-Myc is frequently deregulated Conflict of interest in lymphomas and dacetuzumab signaling repressed c-Myc protein expression in BCL-6 þ cell lines, correlating with All authors are employees and shareholders of Seattle Genetics, induced Blimp-1 expression. Interestingly, dacetuzumab signal- Inc. ing downregulated c-Myc in cell lines lacking a MYC-IG translocation as well as in a cell line having the t(8;14) Acknowledgements MYC–IGH translocation. Blockade of BCL-6 and c-Myc expres- sion by dacetuzumab provides a mechanism for how lymphoma We thank Dr Hans-Peter Gerber for optimization of in vivo cell lines of GC origin are driven into a non-proliferative state chemotherapy regimens. susceptible to apoptosis in the absence of prosurvival signaling. The apoptotic activity of dacetuzumab is not restricted to BCL-6-deregulated lymphoma cell lines, therefore, other apop- References totic signaling pathways are activated by dacetuzumab. We found that dacetuzumab upregulated the proapoptotic p53 1 Kuppers R. Mechanisms of B-cell lymphoma pathogenesis. Nat family member TAp63a, a transcriptionally active isoform of Rev 2005; 5: 251–262. p63, which has apoptotic activity when overexpressed and 2 Ye BH, Rao PH, Chaganti RS, Dalla-Favera R. Cloning of bcl-6, the chemosensitizes carcinoma cells in vitro.33,48 Dacetuzumab locus involved in chromosome translocations affecting band 3q27 induction of TAp63a is NF-kB pathway dependent; however, in B-cell lymphoma. Cancer Res 1993; 53: 2732–2735. 3 Weiss LM, Warnke RA, Sklar J, Cleary ML. Molecular analysis of TAp63a is a target of ubiquitin-mediated proteolysis and the t(14;18) chromosomal translocation in malignant lymphomas. dacetuzumab signaling may also upregulate this protein post N Engl J Med 1987; 317: 1185–1189. translationally.49 The abrupt upregulation of TAp63a from very 4 Ladanyi M, Offit K, Jhanwar SC, Filippa DA, Chaganti RS. MYC low basal levels following dacetuzumab treatment is analogous rearrangement and translocations involving band 8q24 in diffuse to the upregulation of wild-type p53 in normal tissues large cell lymphomas. Blood 1991; 77: 1057–1063. responding to cytotoxic stress. Significantly, upregulation of 5 Gronbaek K, Straten PT, Ralfkiaer E, Ahrenkiel V, Andersen MK, Hansen NE et al. Somatic Fas mutations in non-Hodgkin’s nuclear TAp63a was detected in subcutaneously implanted lymphoma: association with extranodal disease and autoimmunity. Ramos tumor cells following dosing of dacetuzumab in mice, Blood 1998; 92: 3018–3024. demonstrating that dacetuzumab signaling occurs in vivo at the 6 Koduru PR, Raju K, Vadmal V, Menezes G, Shah S, Susin M et al. tumor site. TAp63a positively regulates numerous genes Correlation between mutation in P53, p53 expression, cytoge- involved in mitochondrial and death receptor apoptotic path- netics, histologic type, and survival in patients with B-cell non- ways,33 including the dacetuzumab-induced proteins Fas and Hodgkin’s lymphoma. Blood 1997; 90: 4078–4091. 7 Biancone L, Cantaluppi V, Camussi G. CD40-CD154 interaction in Bim. The proapoptotic BH3-only family members Bim and Bid experimental and human disease (review). Int J Mol Med 1999; 3: were both upregulated by dacetuzumab, and this may mitigate 343–353. the prosurvival advantage of BCL-2 overexpression in certain 8 van Kooten C, Banchereau J. CD40-CD40 ligand. J Leukoc Biol lymphomas. 2000; 67: 2–17. The signaling characteristics of dacetuzumab may also 9 Harnett MM. CD40: a growing cytoplasmic tale. Sci STKE 2004; account for its ability to sensitize lymphoma cell lines to 2004: pe25. 10 Ramesh N, Morio T, Fuleihan R, Worm M, Horner A, Tsitsikov E chemotherapeutic drugs. We demonstrated that dacetuzumab is et al. CD40-CD40 ligand (CD40L) interactions and X-linked synergistic with a panel of cytotoxic DNA-damaging drugs hyperIgM syndrome (HIGMX-1). Clin Immunol Immunopathol in vitro and potentiated chemotherapeutic regimens in NHL 1995; 76(3 Part 2): S208–S213.

Leukemia Dacetuzumab proapoptotic signaling mechanisms in NHL TS Lewis et al 1016 11 Guzman-Rojas L, Sims-Mourtada JC, Rangel R, Martinez-Valdez 32 Chou TC, Talalay P. Quantitative analysis of dose-effect relation- H. Life and death within germinal centres: a double-edged sword. ships: the combined effects of multiple drugs or enzyme inhibitors. Immunology 2002; 107: 167–175. Adv Enzyme Regul 1984; 22: 27–55. 12 Vonderheide RH. Prospect of targeting the CD40 pathway for 33 Gressner O, Schilling T, Lorenz K, Schulze Schleithoff E, Koch A, cancer therapy. Clin Cancer Res 2007; 13: 1083–1088. Schulze-Bergkamen H et al. TAp63alpha induces apoptosis by 13 Aggarwal BB. Signalling pathways of the TNF superfamily: a activating signaling via death receptors and mitochondria. EMBO J double-edged sword. Nat Rev Immunol 2003; 3: 745–756. 2005; 24: 2458–2471. 14 Furman RR, Asgary Z, Mascarenhas JO, Liou HC, Schattner EJ. 34 Kelly KF, Daniel JM. POZ for effectFPOZ-ZF transcription factors Modulation of NF-kappa B activity and apoptosis in chronic in cancer and development. Trends Cell Biol 2006; 16: 578–587. lymphocytic leukemia B cells. J Immunol 2000; 164: 2200–2206. 35 Baron BW, Anastasi J, Montag A, Huo D, Baron RM, Karrison T 15 Funakoshi S, Longo DL, Beckwith M, Conley DK, Tsarfaty G, et al. The human BCL6 transgene promotes the development of Tsarfaty I et al. Inhibition of human B-cell lymphoma growth by lymphomas in the mouse. Proc Natl Acad Sci USA 2004; 101: CD40 stimulation. Blood 1994; 83: 2787–2794. 14198–14203. 16 Coiffier B. Rituximab therapy in malignant lymphoma. Oncogene 36 Cattoretti G, Pasqualucci L, Ballon G, Tam W, Nandula SV, 2007; 26: 3603–3613. Shen Q et al. Deregulated BCL6 expression recapitulates the 17 Zelenetz AD, Hamlin P, Kewalramani T, Yahalom J, Nimer S, pathogenesis of human diffuse large B cell lymphomas in mice. Moskowitz CH. Ifosfamide, carboplatin, etoposide (ICE)-based Cancer Cell 2005; 7: 445–455. second-line chemotherapy for the management of relapsed and 37 Cerchietti LC, Yang SN, Shaknovich R, Hatzi K, Polo JM, Chadburn refractory aggressive non-Hodgkin’s lymphoma. Ann Oncol 2003; A et al. A peptomimetic inhibitor of BCL6 with potent antilym- 14(Suppl 1): i5–10. phoma effects in vitro and in vivo. Blood 2009; 113: 3397–3405. 18 Tong AW, Stone MJ. Prospects for CD40-directed experimental 38 Cerchietti LC, Lopes EC, Yang SN, Hatzi K, Bunting KL, Tsikitas LA therapy of human cancer. Cancer Gene Ther 2003; 10: 1–13. et al. A purine scaffold Hsp90 inhibitor destabilizes BCL-6 and has 19 Law CL, Gordon KA, Collier J, Klussman K, McEarchern JA, specific antitumor activity in BCL-6-dependent B cell lymphomas. Cerveny CG et al. Preclinical antilymphoma activity of a Nat Med 2009; 15: 1369–1376. humanized anti-CD40 monoclonal antibody, SGN-40. Cancer 39 Burington B, Yue P, Shi X, Advani R, Yu N, Lau J et al. CD40 Res 2005; 65: 8331–8338. Pathway activation status predicts response to CD40 targeted 20 Oflazoglu E, Stone IJ, Brown L, Gordon KA, van Rooijen N, Jonas therapy in diffuse large B cell lymphoma. Sci Transl Med 2011 M et al. Macrophages and Fc-receptor interactions contribute to (in press). the antitumour activities of the anti-CD40 antibody SGN-40. Br J 40 Marshall CJ. Specificity of receptor tyrosine kinase signaling: Cancer 2009; 100: 113–117. transient versus sustained extracellular signal-regulated kinase 21 Mohammad RM, Wall NR, Dutcher JA, Al-Katib AM. The addition activation. Cell 1995; 80: 179–185. of bryostatin 1 to cyclophosphamide, doxorubicin, vincristine, and 41 Whalen AM, Galasinski SC, Shapiro PS, Nahreini TS, Ahn NG. prednisone (CHOP) chemotherapy improves response in a CHOP- Megakaryocytic differentiation induced by constitutive activation resistant human diffuse large cell lymphoma xenograft model. Clin of mitogen-activated protein kinase kinase. Mol Cell Biol 1997; Cancer Res 2000; 6: 4950–4956. 17: 1947–1958. 22 Allman D, Jain A, Dent A, Maile RR, Selvaggi T, Kehry MR et al. 42 Bric A, Miething C, Bialucha CU, Scuoppo C, Zender L, Krasnitz A BCL-6 expression during B-cell activation. Blood 1996; 87: et al. Functional identification of tumor-suppressor genes through 5257–5268. an in vivo RNA interference screen in a mouse lymphoma model. 23 Saito M, Gao J, Basso K, Kitagawa Y, Smith PM, Bhagat G et al. A Cancer Cell 2009; 16: 324–335. signaling pathway mediating downregulation of BCL6 in germinal 43 Shapiro-Shelef M, Lin KI, Savitsky D, Liao J, Calame K. Blimp-1 is center B cells is blocked by BCL6 gene alterations in B cell required for maintenance of long-lived plasma cells in the bone lymphoma. Cancer Cell 2007; 12: 280–292. marrow. J Exp Med 2005; 202: 1471–1476. 24 Jardin F, Ruminy P, Bastard C, Tilly H. The BCL6 proto-oncogene: 44 Shaffer AL, Lin KI, Kuo TC, Yu X, Hurt EM, Rosenwald A et al. a leading role during germinal center development and lympho- Blimp-1 orchestrates plasma cell differentiation by extinguishing magenesis. Pathol Biol (Paris) 2007; 55: 73–83. the mature B cell gene expression program. Immunity 2002; 17: 25 Niu H, Ye BH, Dalla-Favera R. Antigen receptor signaling induces 51–62. MAP kinase-mediated phosphorylation and degradation of the 45 Iwakoshi NN, Lee AH, Vallabhajosyula P, Otipoby KL, Rajewsky BCL-6 transcription factor. Genes Dev 1998; 12: 1953–1961. K, Glimcher LH. Plasma cell differentiation and the unfolded 26 Tunyaplin C, Shaffer AL, Angelin-Duclos CD, Yu X, Staudt LM, protein response intersect at the transcription factor XBP-1. Calame KL. Direct repression of prdm1 by Bcl-6 inhibits Nat Immunol 2003; 4: 321–329. plasmacytic differentiation. J Immunol 2004; 173: 1158–1165. 46 Reimold AM, Iwakoshi NN, Manis J, Vallabhajosyula P, Szomola- 27 Ranuncolo SM, Polo JM, Dierov J, Singer M, Kuo T, Greally J et al. nyi-Tsuda E, Gravallese EM et al. Plasma cell differentiation requires Bcl-6 mediates the germinal center B cell phenotype and the transcription factor XBP-1. Nature 2001; 412:300–307. lymphomagenesis through transcriptional repression of the DNA- 47 Lin KI, Lin Y, Calame K. Repression of c-myc is necessary but not damage sensor ATR. Nat Immunol 2007; 8: 705–714. sufficient for terminal differentiation of B lymphocytes in vitro. 28 Phan RT, Dalla-Favera R. The BCL6 proto-oncogene suppresses p53 Mol Cell Biol 2000; 20: 8684–8695. expression in germinal-centre B cells. Nature 2004; 432: 635–639. 48 Yang A, Kaghad M, Wang Y, Gillett E, Fleming MD, Dotsch V et al. 29 Phan RT, Saito M, Basso K, Niu H, Dalla-Favera R. BCL6 interacts p63, a p53 homolog at 3q27–29, encodes multiple products with with the transcription factor Miz-1 to suppress the cyclin- transactivating, death-inducing, and dominant-negative activities. dependent kinase inhibitor p21 and cell cycle arrest in germinal Mol Cell 1998; 2: 305–316. center B cells. Nat Immunol 2005; 6: 1054–1060. 49 Li Y, Zhou Z, Chen C. WW domain-containing E3 ubiquitin 30 Dicker F, Kater AP, Prada CE, Fukuda T, Castro JE, Sun G et al. protein ligase 1 targets p63 transcription factor for ubiquitin- CD154 induces p73 to overcome the resistance to apoptosis of mediated proteasomal degradation and regulates apoptosis. Cell chronic lymphocytic leukemia cells lacking functional p53. Blood Death Differ 2008; 15: 1941–1951. 2006; 108: 3450–3457. 50 Winter JN, Weller EA, Horning SJ, Krajewska M, Variakojis D, 31 Lantner F, Starlets D, Gore Y, Flaishon L, Yamit-Hezi A, Dikstein R Habermann TM et al. Prognostic significance of Bcl-6 protein et al. CD74 induces TAp63 expression leading to B-cell survival. expression in DLBCL treated with CHOP or R-CHOP: a Blood 2007; 110: 4303–4311. prospective correlative study. Blood 2006; 107: 4207–4213.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

Leukemia