IL10 Receptor Is a Novel Therapeutic Target in Dlbcls

ORIGINAL ARTICLE IL10 receptor is a novel therapeutic target in DLBCLs

W Béguelin1, S Sawh1, N Chambwe2,3,4, FC Chan5, Y Jiang1, J-W Choo1, DW Scott5, A Chalmers1, H Geng6, L Tsikitas1,WTam7, G Bhagat8, RD Gascoyne5,9 and R Shaknovich1,8

Diffuse large B-cell lymphoma (DLBCL) is a biologically and clinically heterogeneous disease with marked genomic instability and variable response to conventional R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone) chemotherapy. More clinically aggressive cases of DLBCLs have high level of circulating interleukin 10 (IL10) cytokine and evidence of activated intracellular STAT3 (signal transducer and activator of transcription 3) signaling. We investigated the role of IL10 and its surface receptor in supporting the neoplastic phenotype of DLBCLs. We determined that IL10RA gene is ampliﬁed in 21% and IL10RB gene in 10% of primary DLBCLs. Gene expression of IL10, IL10RA and IL10RB was markedly elevated in DLBCLs. We hypothesized that DLBCLs depend for their proliferation and survival on IL10-STAT3 signaling and that blocking the IL10 receptor (IL10R) would induce cell death. We used anti-IL10R blocking antibody, which resulted in a dose-dependent cell death in all tested activated B-cell-like subtype of DLBCL cell lines and primary DLBCLs. Response of germinal center B-cell-like subtype of DLBCL cell lines to anti-IL10R antibody varied from sensitive to resistant. Cells underwent cell cycle arrest, followed by induction of apoptosis. Cell death depended on inhibition of STAT3 and, to a lesser extent, STAT1 signaling. Anti-IL10R treatment resulted in interruption of IL10-IL10R autostimulatory loop. We thus propose that IL10R is a novel therapeutic target in DLBCLs.

Leukemia (2015) 29, 1684–1694; doi:10.1038/leu.2015.57

INTRODUCTION autostimulatory loop via STAT3 signaling.5,13 IL10 can promote the Diffuse large B-cell lymphomas (DLBCLs) are a group of aggressive proliferation of normal B cells that are activated through BCR or 13,14 lymphomas and the most common form of non-Hodgkin CD40. IL10 is also necessary for the development and lymphoma.1 DLBCLs arise from germinal center (GC) and post- maintenance of B-cell lymphomas arising spontaneously in the 14 GC B cells and show marked genetic and epigenetic heterogeneity New Zealand black strain of mice. More recently, it has been and differences in clinical outcomes.2,3 Gene expression profiling shown that serum IL10 levels correlate with IL10 secreted by the 15 identified two main molecular subgroups of DLBCLs: activated primary DLBCL in patients. Most importantly, IL10 serum level is B-cell-like (ABC) and GC B-cell-like (GCB). The ABC subtype of a biomarker of clinical outcome in patients with primary DLBCLs: DLBCLs are more aggressive and are thought to arise from late GC higher IL10 in the serum predicts worse outcomes as shown by or preplasma cells and are characterized by low expression of Gupta et al.16 IL10 is a class 2 cytokine, which is a homodimer, and BCL6 and higher expression of plasma cell-associated transcription is produced by monocytes, type 2 T helper cells (TH2) CD4+CD25+ factors such as XBP1 and PRDM1, marked activation of nuclear Foxp3+ regulatory T cells and subsets of activated T and factor-κB (NF-κB) signaling and STAT3 (signal transducer and B cells.17,18 The main function of IL10 in normal immunity is activator of transcription 3) signaling, as compared with GCB immunosuppressive. This is mediated by the IL10 receptor (IL10R), subtype.2,4,5 Normal lymphocytes and lymphoid malignancies which consists of two subunits: IL10Rα and IL10Rβ.19,20 IL10 depend on cytokine stimulation for growth and survival via both suppresses inflammation by downregulating TH1 cytokines and autocrine and paracrine mechanisms. Cytokines such as major histocompatibility complex class II antigens, and by interleukin 4 (IL4), IL6, IL10 and IL21 have an important role in suppressing synthesis of proinflammatory cytokines such as the differentiation of B cells and in B-cell microenvironment IFN-γ, IL2, IL3 and tumor necrosis factor-α by macrophages and interaction.6–9 Hematologic tumors are known to 'hijack' normal T cells; it also regulates proliferation and survival of B cells. Binding biological processes, which confer a growth advantage. The best to cytokine-specific receptors results in autophosphorylation of studied example is activation of IL6 signaling in multiple the IL10Rα subunit, which in turn leads to the activation of either myeloma, a malignancy arising from a terminal stage of B-cell IL10Rα-associated Janus kinase 1 (JAK1) or IL10Rα-associated TYK2 differentiation.10–12 kinase.18 Both JAK1 and TYK2 activation in turn leads to IL10 signaling is known to have a crucial role in all DLBCLs, but downstream activation of a family of STAT (signal transducer particularly ABC DLBCLs, where IL10 has been shown to be and activator of transcription) proteins. IL10 has been shown to produced at higher levels.5 IL10 is a direct target of NF-κB preferentially signal through STAT3 and, to a much lesser extent, signaling, which is activated in ABC DLBCLs, and contributes to the through STAT1, STAT5 and STAT6.4 After phosphorylation on

1Department of Medicine, Division of Hematology and Oncology, Weill Cornell Medical College, New York, NY, USA; 2The HRH Prince Alwaleed Bin Talal Bin Abdulaziz Alsaud Institute for Computational Biomedicine, Weill Cornell Medical College, New York, NY, USA; 3Department of Physiology and Biophysics, Weill Cornell Medical College, New York, NY, USA; 4Tri-Instituitional Training Program in Computational Biology and Medicine, Weill Cornell Medical College, New York, NY, USA; 5Centre for Lymphoid Cancer, British Columbia Cancer Agency, Vancouver, BC, Canada; 6Department of Laboratory Medicine, University of California San Francisco, San Francisco, CA, USA; 7Department of Pathology and Laboratory Medicine, Weill Cornell Medical College, New York, NY, USA; 8Department of Pathology and Cell Biology, Columbia University Medical Center, New York, NY, USA and 9Department of Pathology, University of British Columbia, Vancouver, BC, Canada. Correspondence: Dr R Shaknovich, Department of Pathology and Department of Medicine, Division of Hematology and Oncology, Weill Cornell Medical College, 1300 York Avenue, Building C, Room 620C, New York, NY 10065, USA. E-mail: [email protected] Received 8 July 2014; revised 16 February 2015; accepted 19 February 2015; accepted article preview online 3 March 2015; advance online publication, 7 April 2015 Therapeutic targeting of IL10 receptor in DLBCLs W Béguelin et al 1685 Tyr705 and, to a lesser extent, on Ser737, STAT3 is translocated DLBCL patient samples to the nucleus, where it affects the expression of downstream Leftover, patient-deidentified tissue samples removed for diagnosis from genes containing consensus 'GAS' (gamma-activated site) DNA five patients with de novo DLBCL were obtained from New York motifs.18,21 The STAT3 gene promoter itself contains a GAS motif, Presbyterian Hospital. The use of human tissue was approved by the and thus its activation initiates an autostimulatory loop resulting research ethics board of New York Presbyterian Hospital and the Weill in increased amount of unphosphorylated-STAT3. Recently, it Cornell Medical Center. has been shown that not only phosphorylated-STAT3, but also unphosphorylated-STAT3 can translocate to the nucleus and Primary DLBCL cell culture 22,23 activate downstream genes. Irradiated HK cells were plated in Dulbecco's modification of Eagle's The importance of IL10-JAK1/2-STAT3 axis in DLBCL survival, medium with 10% fetal bovine serum and incubated for 24 h at 37 °C. The proliferation and treatment has been previously appreciated and media were then aspirated and patient DLBCL cells were resuspended in targeted therapeutically. Gupta et al.16 documented higher levels advanced RPMI with 10% human serum at 3 million cells per ml and plated of serum IL10 in patients with more aggressive disease, showing directly on attached HK cells. that IL10 alone, among cytokines, activates JAK2 and STAT3 in patient samples ex vivo, and showed that a JAK2-specific inhibitor Survival analysis can interrupt autocrine IL10 secretion and JAK2/STAT3 activation. Kaplan–Meier analysis was performed using the R Survival Package 4 5 dichotomizing the patients into high and low expression groups based Ding et al. and Lam et al. also demonstrated that STAT3 24 activation preferentially promotes the survival of ABC DLBCLs, and on a median threshold. We divided patients (n = 49; Shaknovich et al. ) into high and low expressers, based on whether they had IL10 and that the use of JAK inhibitors leads to apoptosis. Although prior IL10R levels above or below the median expression, respectively. A log- studies focused on IL10 secretion and signaling as a result of rank test was performed to assess survival differences between the two STAT3 activation in DLBCLs, attempting to target the signaling groups. pathways downstream of JAKs, we discovered that an IL10 autoregulatory loop is also activated upstream in DLBCLs with marked upregulation of IL10R expression, and consequent RESULTS activation of STAT3 signaling. We identified frequent upregulation IL10RA and IL10RB are markedly overexpressed and genomically of the IL10R via genomic amplification of IL10RA and IL10RB amplified in primary DLBCLs subunits in DLBCLs. We used this observation for therapeutic To understand the role of IL10 signaling in DLBCLs, we measured targeting and demonstrated that blocking IL10R results in the gene expression levels of IL10, IL10RA and IL10RB interruption of IL10 signaling. This leads to apoptosis in DLBCL using Affymetrix HG U133 plus 2.0 array in 52 primary DLBCLs.24 cell lines and in primary DLBCL cases treated ex vivo. We determined that all three genes were significantly overexpressed as compared with normal germinal center B cells (NGCBs) (all t-test, Po0.01) (Figure 1a). ABC DLBCLs have higher MATERIALS AND METHODS level of expression for all three genes when compared with GCB A more detailed description of the Materials and methods can be found in DLBCLs. RNA-Seq data from an independent cohort of 22 cases the Supplementary data. (dbGap (NCI) (CGCI) Study Accession: phs000532.v3.p1) confirmed this finding (Supplementary Figure 1A). We further confirmed that Cytokine and blocking antibodies IL10Rα protein expression is elevated using immunohistochem- Human IL10 (Antigenix America Inc., Huntington Station, NY, USA) was used at istry on tissue microarrays with primary DLBCLs (independent a concentration of 10 ng/ml. The following blocking anti-IL10Rα antibodies cohort, n = 38) (Supplementary Figure 1B). Normal lymphoid tissue were used at concentrations ranging from 0.1 to 10 μg/ml: clone 3F9-2 displayed the highest expression of IL10R in macrophages and low from Novus Biologicals (Littleton, CO, USA) (used in all neutralizing expression in NGCBs. Immunohistochemistry on tissue microarrays experiments, unless otherwise specified); C-20 and clone A-3, both from revealed that 45% of DLBCLs overexpressed IL10Rα several fold Santa Cruz Biologicals (Santa Cruz, CA, USA). As a control antibody, above the expression level in normal B cells (intensity 2+ and 3+ is anti-actin H-300 from Santa Cruz Biologicals was used. considered overexpression, as compared with intensity 1+ in NGCBs) (Table 1). Gene expression array and RNA-Seq To examine if DLBCL cell lines recapitulate the biology of IL10 Diagnostic DLBCL samples were collected from individuals who presented signaling, we used a panel of GCB (n = 8) and ABC (n = 4) DLBCL with de novo DLBCL at the British Columbia Cancer Agency (BCCA), cell lines and showed robust expression of all three genes of Canada. Detailed clinical information was published before and can be interest in all cell lines using quantitative real-time PCR (qRT-PCR) found in Shaknovich et al.24 RNA was hybridized to the Affymetrix chip (HG (Figure 1b). We confirmed the expression of IL10Rα and U133 plus 2.0). The raw data (.cel files) are available from gene expression IL10Rβ subunits on the cell membrane using flow cytometry omnibus database (GEO accession: GSE23501 for DLBCL samples and (Figure 1c). GSE15271 for normal GC B cells (NGCBs) from tonsils).24,25 Data were preprocessed using the Affymetrix pipeline and normalized using the robust multichip average approach.26 Data were summarized to the RefSeq custom CDF file to average expression intensities to RefSeq annotated Table 1. IL10Rα protein is overexpressed in 45% of DLBCLs mRNAs.27 We carried out differential expression analysis using a moderated t-test (limma package in R).28 We considered a gene significant Total Intensity 0 Intensity 1+ Intensity 2+ Intensity 3+ if the P-value was o0.05 and the magnitude of the log fold change |log FC| ⩾ 1.0, a twofold difference. The values were plotted using log 2 ABC DLBCL 14 4 (29%) 4 (29%) 2 (14%) 4 (28%) (normalized intensity). GCB DLBCL 24 4 (17%) 9 (37%) 4 (17%) 7 (29%) RNA-Seq libraries were prepared using TruSeq RNA Sample Kits All DLBCL 38 8 (21%) 13 (34%) 6 (16%) 11 (29%) (Illumina, San Diego, CA, USA) according to the manufacturer. Libraries Quantification of IL10Rα protein expression level by immunohistochem- were validated using the Agilent Technologies 2100 Bioanalyzer (Santa istry on a tissue microarray with primary DLBCLs. Intensity 1+ is Clara, CA, USA) and 8–10 pM sequenced on HiSeq2000 sequencer (Illumina) comparable to normal GC B cells. Abbreviations: ABC, activated B-cell- as 2 × 50. RNA-Seq data was aligned to transcripts using TopHat, gene like; DLBCL, diffuse large B-cell lymphoma; GCB, GC B-cell-like; IL, expression quantified in RPKM and differentially expressed genes interleukin. identified.

© 2015 Macmillan Publishers Limited Leukemia (2015) 1684 – 1694 1686 siao.Sria ifrne ewe h w ruswr acltduiglgrn test. log-rank using calculated were FI, groups three. low two of and the total between high a differences of into Survival representative estimator. divided is were shown experiment R-CHOP The lines. cell xrsinwsmaue sn fyerxH 13pu . ra aao 2piayDBL.ACDBL( DLBCL ABC DLBCLs. primary 52 on data array 2.0 plus to U133 HG classi Affymetrix non using measured was expression iue1. Figure ekma(05 1684 (2015) Leukemia GAPDH L0 shgl xrse nDBL n rdcswresria.( survival. worse predicts and DLBCLs in expressed highly is IL10R ﬁ nDBLcl ie.Vle naesona mean as shown are in Values lines. cell DLBCL in d( ed GAPDH n Cell number mRNA levels (relative to ) Gene expression levels =

aps42 oeo UH6 y0 TMD8 Ly10 SUDHL6 Toledo Karpas 422 (normalized intensity) ) GB omlG el ( cells B GC normal NGCB, 5); 1000 1500 2000 500 IL10R (FI) – 1694 hrpui agtn fI1 eetri DLBCLs in receptor IL10 of targeting Therapeutic GCB-DLBCL NGCB GCB IL10RA ** IL10RA Survival probability *** 0.0 0.4 0.8 1.0 0.2 0.6 DLBCL ABC NC 0 ABC Progression FreeSurvival p=0.02 y y Ly3 Ly1 Ly7 IL10R 2 4 n = Time (years) xrsinlvl,adporsinfe uvvl(F)wsaaye sn Kaplan using analyzed was (PFS) survival progression-free and levels, expression 100 200 300 400 )( 4) Béguelin W t ts,* -test, 6 8 10 NGCB GCB GCB-DLBCL ± *** * P ..( s.d. o IL10RB IL10 tal et *** IL10RB .5 ** 0.05, c IL10R ) DLBCL * ABC NC P o α a fl HBL1 opo fgn xrsinlvlof level expression gene of Boxplot ) n IL10R and oecneitniy ( intensity. uorescence ABC .1 *** 0.01, High Low 100 200 300 P IL10 IL10 β o NGCB GCB rti xrsinlvl by levels expression protein .0) ( 0.005). expression(n=25) expression(n=24) GCB-DLBCL ** *** *** b R-C of qRT-PCR ) d DLBCL IL10 ot-ieDBLptet rae with treated patients DLBCL Forty-nine ) ABC NC 05McilnPbihr Limited Publishers Macmillan 2015 © ABC n IL10RA = IL10RA 4;GBDBL( DLBCL GCB 14); fl , wctmtyi DLBCL in cytometry ow IL10RB , IL10RB control Isotype IL10R IL10R and and IL10 n IL10 = 3;NC, 33); relative – Gene . Meier Therapeutic targeting of IL10 receptor in DLBCLs W Béguelin et al 1687 To understand if genomic events may underlie the IL10 and commercially available anti-IL10Rα antibodies that we tested had a IL10R overexpression, we analyzed copy number changes using similar effect on cell viability (Supplementary Figure 2B). To confirm SNP array on a subset of 91 primary DLBCLs.29 We identified broad that cell death was due to blocking the receptor and not a non- regions of genomic amplifications in this cohort using the GISTIC specific effect, we used small hairpin RNAs (shRNAs) to knock down algorithm30 and determined that IL10RA was amplified in 21% and IL10R expression. We used three shRNAs against the IL10RA IL10RB in 10% of primary DLBCLs (Table 2). transcript and achieved marked downregulation, as determined by quantitative PCR and western blot analysis (Supplementary Higher expression of IL10 predicts worse progression-free survival Figures 2C and 2D; hairpin nos 1, 3 and 5). We tested the effect in in DLBCL patients two GCB (Ly1 and Ly7) and two ABC (TMD8 and HBL1) cell lines. To evaluate IL10 and IL10R expression as potential biomarkers of After infection with either control shRNA or shRNA 1, 3 or 5, cells were selected using puromycin for 8 days and cell surface clinical aggressiveness in DLBCLs, we performed survival analysis fl based on the gene expression levels of IL10 and IL10R. We divided expression of IL10R was assessed using ow cytometry, which patients (n = 49, Shaknovich et al.24) into high and low expressers, demonstrated marked downregulation of the receptor in all infected cell lines (Figure 2d). After puromycin selection cells were based on whether they had IL10 and IL10R levels above or 5 below the median expression, respectively. Cases with high plated at a density of 5 × 10 cells per ml and viability of infected expression of IL10 showed worse progression-free survival cell lines was assessed 3 days after plating. shRNA-dependent (log-rank test, P = 0.02, hazard ratio = 0.288, 95% confidence depletion of IL10R led to over 50% suppression of cell growth interval: 0.094–0.884) based on Kaplan–Meier survival analysis relative to the control shRNA in Ly1, Ly7, TMD8 and HBL1, but not (Figure 1d). Expression of IL10RA and IL10RB displayed a trend in in Ly8, Ly18 and Toledo (Figure 2e and Supplementary Figure 2E), α predicting the outcome of the disease, but did not reach statistical also shown to be resistant to anti-IL10R antibody treatment, significance. Larger sample size is necessary to achieve statistically indicating that the cell death observed because of blocking the significant differences in survival based on the expression of receptor by antibody treatment or silencing its expression using these genes. This finding confirms that IL10 expression is a shRNAs is a specific and selective effect. biomarker of disease aggressiveness, as demonstrated earlier by Gupta et al.16 based on the measurements in patient’s serum. Our Blocking IL10R leads to cell death through cell cycle arrest and data reveal that it is likely that a tumor-derived IL10 contributes induction of apoptosis to high serum levels. To determine the mechanism of cell death after IL10R blockade, we assessed cell cycle progression by flow cytometry analysis using DLBCLs are dependent on IL10 signaling and blocking the BrdU/7AAD (bromodeoxyuridine/7-amino-actinomycin D) staining, receptor kills DLBCL cell lines and we evaluated apoptosis using AnnexinV/DAPI (4',6-diamidino- To understand if all DLBCLs depend on IL10 signaling and to 2-phenylindole) staining and western blotting for poly (ADP-ribose) address the differences between GCB and ABC DLBCLs, we started polymerase cleavage. We found a significant increase in the G0–G1 by measuring IL10 secretion in the panel of cell lines using a phase and a large reduction in the S phase of the cell cycle 24 h commercial ELISA Kit. ELISA assays confirmed that all DLBCL cell after anti-IL10R antibody treatment as compared with the control lines tested secreted IL10, with ABC DLBCLs secreting more IL10 antibody treatment in all cell lines tested (Figures 3a and b). These than GCB DLBCLs. Amounts vary broadly from 60 pg/ml per million results indicate that anti-IL10R antibody induces cell cycle arrest in cells per 3 h on average by GCB DLBCLs to nearly 2000 pg/ml per the G0–G1 phase. We also observed significant cell death through million cells per 3 h on average by ABC DLBCLs (Figure 2a). To induction of apoptosis, as assessed by AnnexinV/DAPI staining understand how IL10 signaling, whether autocrine or exocrine, (Figures 3c and d) and poly (ADP-ribose) polymerase-1 cleavage affects DLBCL survival and proliferation, we decided to block IL10R (Figure 3e) by day 3 in Ly1, Ly7, TMD8 and HBL1 cells treated with on the cell surface. the anti-IL10R antibody, but not the control antibody. We took advantage of a published and commercially available Based on these data, we next examined whether anti-IL10R antibody against IL10Rα, which had been reported to have a treatment affects the cell viability of primary DLBCL specimens. blocking effect,31 and tested its effect on DLBCL viability using a Cells were isolated from lymphoma samples and plated onto a range of concentrations from 0.1 to 10 μg/ml (Figure 2b). We chose monolayer of irradiated HK cells (supportive follicular dendritic to block IL10Rα subunit also because the ligand binds to IL10Rα cells) without selection of B cells to optimize cell viability. After and not IL10Rβ subunit, and may be able to signal without 1 h of incubation in the growth medium, cells were treated with heterodimerization. Within 3 days of treatment with 5 μg/ml 5 μg/ml of anti-IL10R or control antibody. Primary DLBCL samples anti-IL10Rα antibody, most cell lines had reduced viability by are difficult to culture ex vivo, but we were able to maintain viable around 50%, and after 5 days, by more than 75% (Figure 2c). B cells in untreated and control antibody treated samples for more Notably, there were three GCB DLBCL cell lines, Ly8, Ly18 and than 3 days (Supplementary Figure 3A), as determined using Toledo, that were not affected by IL10Rα antibody treatment, even gating on B cells (CD20+ CD3-) and assessing AnnexinV/DAPI though they express IL10Rα (Supplementary Figure 2A). All three staining in that population of cells (Figure 3f). We found significant

Table 2. IL10R gene is ampliﬁed in primary DLBCL samples

RefSeq Chromosomal location Gain minimal common regions Amplified in Amplified in Amplified in Amplified in all DLBCL ABC DLBCL GCB DLBCL unclassified (n = 91) (n = 12) (n = 32) DLBCL (n = 47)

IL10RA NM_001558 Chr11: 117 362 316–117 377 408 Chr11: 117 062 750–117 400 137 19 (21%) 2 (17%) 8 (25%) 9 (19%) IL10RB NM_000628 Chr21: 33 560 535–33 591 409 Chr21: 32 341 542–34 104 468 9 (10%) 2 (17%) 2 (6%) 5 (11%) Analysis of copy number changes using SNP array on a subset of 91 primary DLBCLs. Genomic ampliﬁcations were identiﬁed using the GISTIC algorithm. Chromosomal locations are based on NCBI35/HG17 Human Genome Assembly. Abbreviations: ABC, activated B-cell-like; Chr, chromosome; DLBCL, diffuse large B-cell lymphoma; GCB, GC B-cell-like; IL, interleukin; SNP, single-nucleotide polymorphism.

© 2015 Macmillan Publishers Limited Leukemia (2015) 1684 – 1694 Therapeutic targeting of IL10 receptor in DLBCLs W Béguelin et al 1688 cell death through induction of apoptosis at day 3 in all patient Cell death after anti-IL10R treatment is due to the interruption of samples treated with the anti-IL10R antibody, but not the control IL10-JAK-STAT signaling antibody (Figures 3f, g and h and Supplementary Figure 3B). The We next investigated whether anti-IL10R treatment has a speciﬁc IL10R antibody treatment of primary DLBCL B cells led to an effect on downstream signaling cascades. We conﬁrmed that ABC average of 30% reduction of cell viability relative to control DLBCLs have higher baseline expression levels of STAT1 and antibody (Figure 3h; t-test, P = 0.015). STAT3 proteins as compared with GCB DLBCLs, and that IL10

0.1 g/ml Ly1 1 g/ml Ly7 10 g/ml HBL1 TMD8 Ly3 IL10 secretion (pg/ml)

0 3 6 24 Time (h)

GCB ABC

2 days 3 days 5 days

GCB-DLBCL ABC-DLBCL

shControl shIL10RA #3 shIL10RA #1 shIL10RA #5

Ly1 Ly7 Ly1 Ly7 TMD8 HBL1 120 120 120 120 * 100 * 100 ** 100 100 ** 80 80 80 80 60 60 60 60 40 40 40 40

TMD8 HBL1 cells % Viable 20 20 20 20 0 0 0 0 Max

% #1 #3 #5 #1 #3 #5 #1 #3 #5 #1 #3 #5 shIL10RA shIL10RA shIL10RA shIL10RA shControl shControl shControl shControl Cell number IL10R (FI) Figure 2. Blocking IL10Rα inhibits DLBCL cell line proliferation. (a) IL10 secretion was measured by ELISA 3, 6 and 24 h after plating the cells. (b) DLBCLs were exposed to anti-IL10Rα antibody (ab) or control ab (anti-actin antibody) at the indicated concentrations for 5 days, and viability was measured using cell titer blue. (c) DLBCLs were exposed to 5 μg/ml anti-IL10Rα ab or control ab for 2, 3 and 5 days, and viability was measured as in (b). (d) IL10Rα protein expression levels by flow cytometry in DLBCL cells expressing three independent IL10RA shRNAs or control. (e) Viability of DLBCLs from (d) was evaluated 11 days after infection, using cell titer blue. The experiment shown is representative of a total of three (t-test, *Po0.05, **Po0.01). Values in (a)–(c) and (e) are shown as mean of three to five replicate experiments ± s.d. FI, fluorescence intensity.

Leukemia (2015) 1684 – 1694 © 2015 Macmillan Publishers Limited Therapeutic targeting of IL10 receptor in DLBCLs W Béguelin et al 1689 treatment induces higher levels of phosphorylated species of Y705 residue in both ABC and GCB DLBCL cell lines within 1 h of these two signaling molecules4,5 (Figure 4a). We found that treatment (Figure 4a and Supplementary Figure 4A). We also blocking IL10R results in a loss of phosphorylation of STAT3 at observed that phosphorylation of STAT1Y701 was selectively

Ly1 Ly7 Ly3 TMD8 HBL1 Ly1 Ly7 Ly3 100% S 21% S 45% S 35% S 40% S 27% ** * * ** Control * ** 50% *** ab G2-M G2-M G2-M G2-M G2-M G0-G1 19% 15% G0-G1 11% 10% G0-G1 16% 58% G0-G1 39% 52% G0-G1 48% 53% 0% * ** C I C I C I S 16% S 19% S 1.4% S 3% S 11% TMD8 HBL1 Sub-G0 IL10R 100% ab *** ** G0-G1 G2-M G2-M G2-M G2-M G2-M *** *** G0-G1 10% G0-G1 12% G0-G1 7% G0-G1 14% G0-G1 9% 50% S 65% 67% 79% 74% 72% G2-M BrdU 0% 7AAD C I C I

Ly1 Ly7 TMD8 HBL1 Ly1 Ly7 TMD8 HBL1 100% Live 2 5 8 12 *** ** ** *** *** *** Early Control * apoptotic 89 50% ab 79 78 60 *** Late *** *** *** apoptotic 8 27 15 13 0% C I C I C I C I 11 19 31 14

IL10R PARP-1 ab 48 29 36 29 40 51 32 56

DAPI Actin AnnexinV

CD20+ NT C I Average patients Non treated Control ab IL10R ab 100% * 30 40 61 100% Patient #1 50% 20.9 75% 35 28 21 35 14 0.7 0% 50% 35 38 68 100% 25% * Patient #2 50% 33.9 0% 23 21 20 NT C I 36 32 7 0% 17 20 69 100% Live

Patient #3 75 Early 50% apoptotic 24 16 11 57 59 13 Late 0% apoptotic 23 22 38 100% Patient #4 21.6 50% 25 16 30

CD3 44 47 24 DAPI 0% CD20 AnnexinV Figure 3. Inhibition of IL10Rα kills primary DLBCLs and cell lines through induction of apoptosis. (a and b) Cells were exposed to 5 μg/ml IL10Rα antibody (ab) or control ab for 24 h. G0–G1, G2–M and S cell cycle phases were measured by ﬂow cytometry using BrdU (bromodeoxyuridine) and 7AAD (7-amino-actinomycin D) staining. (c–e) Cells were exposed to 5 μg/ml IL10Rα ab or control ab for 3 days. (c and d) Viable, early and late apoptotic cells were measured by ﬂow cytometry using annexinV and DAPI staining. (e) Immunoblotting from whole-cell lysates. (f–h) Primary human DLBCL cells were cocultured with HK follicular dendritic cells and treated with 5 μg/ml IL10Rα ab or control ab for 3 days. Viable, early and late apoptotic B cells (CD20+, CD3 − ) were measured as in (c). Numbers are shown for individual patients (g) and average of all patient samples (h)(t-test, *P = 0.015 relative to control ab). Values in (b), (d) and (h) are shown as mean of triplicate experiments ± s.d. (t-test, *Po0.05, **Po0.01, ***Po0.005 relative to control ab). C, control ab; I, IL10Rα ab; NT, non-treated.

© 2015 Macmillan Publishers Limited Leukemia (2015) 1684 – 1694 Therapeutic targeting of IL10 receptor in DLBCLs W Béguelin et al 1690 affected in ABC DLBCL cell lines. Considering the sensitivity of GCB in all cell lines between 3 and 24 h of treatment (Figure 5a). DLBCLs to IL10R inhibition despite the lower levels of STAT3 Furthermore, we found that anti-IL10R treatment led to the expression, we decided to additionally test their dependence on downregulation of IL10 and/or IL10R transcription (Figure 5b and JAK-STAT signaling by treating cell lines with JAK1/2 inhibitor. We Supplementary Figure 5). These results were confirmed by used a panel of six GCB DLBCLs and three ABC DLBCL cell lines knocking down IL10RA using shRNAs in HBL1 and Ly7 cell lines and treated them with two JAK inhibitors: the non-selective drug (Figure 5c). AG490 and the selective JAK1/2 inhibitor INC18424.32 All cell lines These findings suggest that the effect of IL10R inhibition is due responded to similar concentrations of INC18424 with marked not only to the inhibition of IL10R downstream signaling but also cell death and decrease in STAT3Y705 phosphorylation to the transcriptional silencing of IL10 autostimulatory loop. (Supplementary Figures 4B–D). Two GCB DLBCL cell lines, SUDHL4 and WSU-DLCL2, showed less sensitivity to AG490 as compared DISCUSSION with ABC DLBCLs. We further evaluated the dependence of DLBCLs on JAK-STAT signaling by treating three GCB and three Our investigation of IL10 signaling in B cells led to the discovery of ABC DLBCL cell lines with 5,15-DPP (5,15-diphenylporphyrin), a frequent genomic amplifications and marked overexpression selective STAT3 inhibitor and STAT3-SH2 antagonist. STAT3 of IL10Rα and IL10Rβ in DLBCLs. We used this novel observation inhibition led to cell death in both ABC and GCB DLBCLs at to target IL10R therapeutically using blocking antibody. Although variable concentrations (Supplementary Figures 4E and 4F), previous efforts in understanding the role of IL10 signaling in consistent with the effects triggered by the selective JAK1/2 DLBCLs have focused downstream of the receptor, we made an inhibitor INC18424. As AKT and ERK have been implicated in important novel observation that both IL10 and IL10R are signaling downstream of IL10R,33,34 we evaluated their phosphor- amplified and overexpressed in DLBCLs, and contribute to a ylation status. We determined that blocking IL10R does not affect constitutively activated IL10-dependent autostimulatory loop. The AKT and ERK activation in DLBCL cell lines (Supplementary Figure importance of IL10 signaling in DLBCL biology with resultant 4G). To evaluate if STAT proteins were transcriptionally inactive in activation of JAK-STAT intracellular signaling has been studied 4,5 cells treated with the IL10R antibody, we assessed their nuclear before and key downstream targets of STAT3 that contribute to translocation and binding to DNA consensus sites using a cell proliferation and survival have been identified. While IL10 commercial kit (TransAM; Active Motif, Carlsbad, CA, USA). For gene is also a transcriptional target of NF-κB, and may be this, nuclear extracts were prepared from cell lines treated for 1 h overexpressed, at least in part, as a consequence of the activation with 10 ng/ml IL10 to stimulate signaling, and followed with 2 h of of NF-κB signaling, we made a novel observation that the initiating treatment with 5 μg/ml of the anti-IL10R antibody or the control event in the upregulation of IL10 signaling may be genomic antibody. Nuclear extracts were then incubated with wild-type or amplification and transcriptional deregulation that is independent mutant oligonucleotides to assure specificity of binding to DNA of NF-κB and STAT3. A stable inhibitor of NF-κB suppressed consensus site. The oligonucleotides were immobilized on the IL10 secretion in only one of two cell lines studied, suggesting that other activators of IL10 expression have important roles in ABC plate, and after incubation with nuclear extracts, the amount of 5 bound STAT transcription factors was determined using specific DLBCLs. Here we show that treatment of DLBCL cell lines with antibodies and colorimetric assay. DNA binding by STAT3 was IL10 resulted in the upregulation of the downstream targets, increased in Ly3, TMD8, HBL1 and Ly7 after IL10 stimulation, and which included IL10 itself and STAT3, supporting the idea of an completely abolished after anti-IL10R treatment (Figure 4b). There autostimulatory loop that exists in DLBCLs. It has been postulated was also a small increase in selective STAT1 DNA binding in the before that ABC DLBCLs are more dependent on IL10 signaling, in same cell lines, which was completely inhibited by the anti-IL10R part, because these cell lines secrete higher levels of IL10, have antibody treatment (Figure 4c). IL10 and anti-IL10R treatment had higher levels of both unphosphorylated-STAT3 and STAT3Y705 4 fi no effect on STAT5A and STAT5B signaling (Supplementary Figure and are more sensitive to JAK1 inhibitor. Our study con rmed fi 4H). We assessed STAT3 phosphorylation in nuclear extracts used that ABC DLBCLs secrete signi cantly more IL10 and have higher in DNA binding assays, and we found high levels of nuclear STAT3 levels of phosphorylated and unphosphorylated-STAT3, and also phosphorylation at residue Y705 in IL10-treated Ly3, Ly7, HBL1 demonstrated that inhibiting IL10R results in cell cycle arrest, and TMD8 cells (Figure 4d). Anti-IL10R antibody treatment blocked induction of apoptosis and cell death in all 5 ABC DLBCLs and 9 out of 12 GCB DLBCL cell lines tested. The cellular killing effect was nuclear STAT3 phosphorylation (Figure 4d), consistent with fi abrogation of STAT3 binding to DNA consensus sites. not signi cantly different between ABC- and GCB-sensitive cell To understand how transcriptional targets of STAT3 are affected lines. In addition, a selective JAK1/2 inhibitor INC18424 and STAT3 inhibitor 5,15-DPP killed both ABC and GCB cell lines. These by treatment, we performed qRT-PCR of STAT3 targets that are fi crucial for maintaining cell cycle and proliferation, as well as ndings suggest that despite the fact that ABC DLBCLs have antiapoptotic genes.5 Anti-IL10R treatment resulted in inhibition higher basal level of STAT3 transcription factor and secrete much of transcription of CCND1, CCND2 and CCND3, which are required higher levels of IL10, survival of both subtypes depends on IL10 for cell cycle G1/S transition, consistent with the cell cycle arrest signaling. IL10R blockade resulted in nearly complete block of that we previously observed. Anti-IL10R treatment also inhibited STAT3 phosphorylation, nuclear translocation and DNA binding the transcription of antiapoptotic factors, such as BCLXL and MCL1, based on our results in both molecular subtypes, suggesting that decreasing signaling below the minimal required threshold drive proto-oncogenes, such as cMYC, JUNB, PIM1 and BCL3 and STAT3 cells into cell cycle arrest and apoptosis. Even though signaling itself (Figure 4e and Supplementary Figures 4I and 4J). downstream of IL10 involves primarily activation of JAK1/JAK2 and Altogether, these results indicate that blocking IL10 cell surface STAT3, it is possible that other signaling molecules engaged in receptor results in induction of cell cycle arrest and apoptosis GCB DLBCL subtype may also be important. The potential through inhibition of IL10-JAK-STAT signaling. candidates that may be affected by IL10R inhibition are STAT1, and members of the TYK2-MAPK axis. As significant loss of cellular Anti-IL10R inhibition results in the interruption of IL10 viability resulted from inhibiting the IL10R, our findings suggest autostimulatory loop that DLBCLs may be ‘addicted’ to chronic IL10 signaling. This is To gain insight into how anti-IL10R treatment might alter similar to the described addiction to the oncogenic transcription IL10 secretion, we performed ELISA assays in DLBCL cell lines factors BCL6 or NF-κB.35–39 treated with the control or anti-IL10R antibody. Anti-IL10R The proposed novel therapeutic approach of targeting IL10R treatment resulted in marked downregulation of IL10 secretion has advantages over other therapies that target the JAK-STAT axis

ABC-DLBCL GCB-DLBCL Ly3 TMD8 HBL1 Ly7 Ly1 IL10 --+ + ++++- +++ - +++ - +++ Control ab --+ - - --+ --+ - ---+ - --+ IL10R ab ---+ ---+ ---+ - --+ ---+ pSTAT3 Y705

STAT3

pSTAT1 Y701

STAT1

11.3 1 0.3 1 1.5 1.4 0.6 1 1.1 1.2 0.4 1111 pSTAT1/STAT1

IL10+Control ab IL10+Control ab IL10+ IL10R ab STAT3 STAT1 IL10+ IL10R ab

** * *

STAT activation STAT ** * (relative to control)

WT oligo +++++- - - -- +++++- - - -- Mutant oligo --+++++- - - --+++++- - -

Ly3 TMD8 HBL1 Ly1 Ly7 Ly3 TMD8 HBL1 Ly1 Ly7

IL10+Control ab Nuclear extracts IL10+IL10R ab Ly1 Ly3 Ly7 HBL1 TMD8 IL10+Control ab + --+ + - + - + - IL10+IL10R ab - + - + - + - + - +

*** pSTAT3 Y705 *** mRNA levels mRNA * STAT3 ** (relative to non treated) * ** Histone H3 * ** ** ** PIM1 BCL3 JUNB MCL1 cMYC STAT3 BCLXL CCND1 CCND3 CCND2 Figure 4. Blocking IL10Rα inhibits STAT3 and STAT1 signaling. (a) Immunoblotting from whole-cell lysates from cells pretreated with 10 ng/ml IL10 for 1 h and treated with 5 μg/ml IL10Rα antibody (ab) or control ab for 1 h. The experiment shown is representative of a total of four. (b and c) Transcriptional activation of STAT3 (b) and STAT1 (c) was measured with TransAM Kit (Active Motif) using nuclear extracts from cells pretreated with 10 ng/ml IL10 for 1 h and treated with 5 μg/ml IL10Rα ab or control. (d) Immunoblotting from nuclear extracts used in (b). The experiment shown is representative of a total of three. (e) qRT-PCR of STAT3 target genes relative to GAPDH and relative to untreated cells in HBL1 cells pretreated with 10 ng/ml IL10 for 1 h and treated with 5 μg/ml IL10Rα ab or control ab for 6, 12 or 24 h (CCND1, CCND2, CCND3, BCLXL, cMYC and JUNB:6h;PIM1 and STAT3:12h;MCL1 and BCL3: 24 h). Values in (b), (c) and (e) are shown as mean of triplicate experiments ± s.d. (t-test, *Po0.05, **Po0.01, ***Po0.005 relative to control ab).

Ly1 Ly7 Ly3 TMD8 HBL1 Control ab IL10R ab

** * *** *** *

IL10 secretion (pg/ml) *

0 3 6 24 0 3 6 24 0 3 6 0 3 6 24 0 3 6 24 Time (h) Time (h) Time (h) Time (h) Time (h)

IL10+Control ab IL10+IL10R ab IL10 IL10RB Ly1 Ly7Ly3 TMD8 HBL1 HBL1 Ly7

** * * * * ** ** * * * mRNA levels mRNA * * levels mRNA ** ** ** ** ** ** ** (relative to non treated) (relative to shControl) ** **

IL10 IL10 IL10 IL10 IL10 #1 #3 #5 #1 #3 #5 IL10RA IL10RA IL10RA IL10RA IL10RA IL10RA IL10RB IL10RB IL10RB IL10RB IL10RB shIL10RA shIL10RA shControl shControl Figure 5. Blocking IL10Rα interrupts IL10 autostimulatory loop. (a) IL10 secretion was measured by ELISA after treating cells with 5 μg/ml IL10Rα antibody (ab) or control ab for 3, 6 and 24 h. (b) qRT-PCR of IL10RA, IL10RB and IL10 relative to GAPDH and relative to non-treated cells in DLBCL cells pretreated with 10 ng/ml IL10 for 1 h and treated with 5 μg/ml IL10Rα ab or control ab for 6, 12 or 24 h (IL10 in Ly7 and HBL1, IL10RA in Ly7, Ly3, TMD8 and HBL1, IL10RB in Ly1, Ly3 and TMD8: 6 h; IL10 in TMD8, IL10RA in Ly1, IL10RB in Ly7 and HBL1: 12 h; IL10 in Ly1 and Ly3: 24 h). (c) qRT-PCR of IL10RB and IL10 relative to GAPDH and relative to control in Ly7 cells expressing three independent IL10RA shRNAs or control. Values are shown as mean of triplicate experiments ± s.d. (t-test, *Po0.05, **Po0.01, ***Po0.005 relative to control ab or shControl).

in lymphomas: the cell surface localization of the receptor allows dependent immunosuppressive paracrine effect; and possibly easy targeting and assessment of IL10R on lymphoma cells using decreasing the number of Breg cells, creating a more permissive flow cytometry that can be incorporated into routine clinical environment for the development of Treg and TH1 cells. Matsuda diagnostic practice. Rituximab or anti-CD20 treatment became et al.41 demonstrated that pretreatment of melanoma cells with part of the accepted first-line therapy for patients with DLBCL, and IL10 results in the downregulation of major histocompatibility proved to be efficacious despite the CD20 epitope is not normally complex class I expression and complete inhibition of autologous overexpressed as compared with normal B cells. Overexpression of CTL-dependent tumor lysis.41 Thus, anti-IL10R therapy promises to IL10R in lymphoma cells promises to offer a more selective effect combat lymphoma directly by enabling the antitumor immune on lymphoma cells than on normal B cells. Regulatory B cells response. (Bregs) are the normal B-cell subset with the highest expression of It is likely that single-agent therapy, no matter how successful, IL10 and IL10R. Bregs or B10 cells are defined by IL10 production will require combination with other drugs to overcome or prevent and profound immunosuppressive effects on TH1, TH17 and Treg therapy resistance. Rational candidates for such combinations may cells, thus having an important role in the prevention of represent molecules targeting the same pathway to achieve a autoimmune disease and on the promotion of allograft faster response, or alternatively molecules that target different tolerance.6,40 Bregs resemble innate immune B-1a cells and are pathways, and allow for synergism of action. We demonstrated thought to arise from T2 marginal zone precursor B cells. that inhibition of IL10R results in effects downstream of JAK1 with Interaction of precursor B10pro cells with TLR4 and TLR9 results decreased phosphorylation and transcriptional activation of in the highest production of IL10 and maturation into B10 cells.40 STAT3. To completely inhibit JAK-STAT3 signaling, it is possible Signaling through MyD88, which is frequently mutated in DLBCLs, to use the drug INCB18424, which is a great candidate for enhances IL10 production, but is not required for cell maturation. combinatorial therapy, as it is an orally bioavailable small- The link between Breg and DLBCLs is not known, but the molecule inhibitor of JAK1 and JAK2, but not of JAK3, limiting its immunomodulatory function of these cells includes profound toxicity.42,43 This drug was approved by the FDA in 2011 and later suppressive effect on TH1, TH17 and Treg cells and raises the in Europe and Canada to treat Myelofibrosis. It has also been question: how will anti-IL10R treatment affect the host immune tested in the setting of Rheumatoid Arthritis.44 Other selective response to tumors? Anti-IL10R treatment may potentially boost JAK1/2 inhibitors include CYT387,45 CP-690550 (refs. 6,47 )or host antitumor immune response by eliminating the tumor cells INCB16562.48 An alternative approach for selecting an agent for that produce high levels of IL10, thus eliminating the IL10- combination therapy is targeting a different pathway that is

Leukemia (2015) 1684 – 1694 © 2015 Macmillan Publishers Limited Therapeutic targeting of IL10 receptor in DLBCLs W Béguelin et al 1693 implicated in IL10 signaling, for example, NF-κB signaling. NF-κBis 12 Sati HI, Apperley JF, Greaves M, Lawry J, Gooding R, Russell RG et al. Interleukin-6 is crucial for DLBCL survival and is activated in ABC DLBCLs, thus expressed by plasma cells from patients with multiple myeloma and monoclonal making it an attractive target in lymphoma. NF-κB inhibitor gammopathy of undetermined significance. Br J Haematol 1998; 101:287–295. DHMEQ (dehydroxymethylepoxyquinomicin) is designed to bind 13 Rousset F, Garcia E, Defrance T, Peronne C, Vezzio N, Hsu DH et al. Interleukin 10 is a potent growth and differentiation factor for activated human B lymphocytes. members of the Rel family and inhibit their DNA-binding Proc Natl Acad Sci USA 1992; 89: 1890–1893. 41,49–52 activity. 14 Czarneski J, Lin YC, Chong S, McCarthy B, Fernandes H, Parker G et al. Studies in In conclusion, we identified marked upregulation of IL10 and NZB IL-10 knockout mice of the requirement of IL-10 for progression of B-cell IL10R in both GCB and ABC subtypes of DLBCLs and determined lymphoma. Leukemia 2004; 18: 597–606. that blocking the cell surface receptor results in induction of cell 15 Voorzanger N, Touitou R, Garcia E, Delecluse HJ, Rousset F, Joab I et al. Interleukin death by cell cycle arrest and apoptosis. This is a novel approach (IL)-10 and IL-6 are produced in vivo by non-Hodgkin's lymphoma cells and act as cooperative growth factors. Cancer Res 1996; 56:5499–5505. that has not been used before. This strategy offers the advantage 16 Gupta M, Han JJ, Stenson M, Maurer M, Wellik L, Hu G et al. Elevated serum IL-10 of a targeted therapy directed at an easily accessible cell surface levels in diffuse large B-cell lymphoma: a mechanism of aberrant JAK2 activation. molecule and promise of having a beneficial impact on the tumor Blood 2012; 119: 2844–2853. immune microenvironment. We propose that IL10 and IL10R can 17 Zdanov A. Structural analysis of cytokines comprising the IL-10 family. Cytokine be used as biomarkers in DLBCLs: serum IL10 can be easily Growth Factor Rev 2010; 21:325–330. 18 Moore KW, de Waal Malefyt R, Coffman RL, O'Garra A. Interleukin-10 and the measured in the peripheral blood, whereas cell surface expression 19 – α β fl interleukin-10 receptor. Annu Rev Immunol 2001; : 683 765. of IL10R and IL10R can be evaluated by ow cytometry and 19 Matesanz-Isabel J, Sintes J, Llinas L, de Salort J, Lazaro A, Engel P. New B-cell CD incorporated into routine diagnostic practice. molecules. Immunol Lett 2011; 134: 104–112. 20 Yoon SI, Logsdon NJ, Sheikh F, Donnelly RP, Walter MR. Conformational changes mediate interleukin-10 receptor 2 (IL-10R2) binding to IL-10 and assembly of the CONFLICT OF INTEREST signaling complex. J Biol Chem 2006; 281: 35088–35096. The authors declare no conflict of interest. 21 Akira S, Nishio Y, Inoue M, Wang XJ, Wei S, Matsusaka T et al. Molecular cloning of APRF, a novel IFN-stimulated gene factor 3 p91-related transcription factor involved in the gp130-mediated signaling pathway. Cell 1994; 77:63–71. ACKNOWLEDGEMENTS 22 Yang J, Liao X, Agarwal MK, Barnes L, Auron PE, Stark GR. Unphosphorylated STAT3 accumulates in response to IL-6 and activates transcription by binding to This work was supported by the start up fund from the Department of Medicine NFkappaB. Genes Dev 2007; 21: 1396–1408. WCMC and a generous gift from Bob Horne and Laurie Lindenbaum. We thank Robert 23 Yoshida Y, Kumar A, Koyama Y, Peng H, Arman A, Boch JA et al. Interleukin 1 Bichard for editing the paper. activates STAT3/nuclear factor-kappaB cross-talk via a unique TRAF6- and p65- dependent mechanism. J Biol Chem 2004; 279:1768–1776. 24 Shaknovich R, Geng H, Johnson NA, Tsikitas L, Cerchietti L, Greally JM et al. DNA AUTHOR CONTRIBUTIONS methylation signatures define molecular subtypes of diffuse large B-cell lymphoma. Blood 2010; 116:e81–e89. WB designed and performed research and analyzed data; SS, JWC, AC and LT 25 Caron G, Le Gallou S, Lamy T, Tarte K, Fest T. CXCR4 expression functionally collected data and performed research; NC, FCC, YJ and HG analyzed data and discriminates centroblasts versus centrocytes within human germinal center performed statistical analysis; DS, WT, GB and RDG contributed with patient B cells. J Immunol 2009; 182: 7595–7602. samples and interpreted data; RS designed research, interpreted data and 26 Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U et al. wrote the manuscript. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 2003; 4:249–264. 27 Dai M, Wang P, Boyd AD, Kostov G, Athey B, Jones EG et al. Evolving gene/ REFERENCES transcript definitions significantly alter the interpretation of GeneChip data. Nucleic Acids Res 2005; 33: e175. fi 13 – 1 Coif er B. Diffuse large cell lymphoma. Curr Opin Oncol 2001; :325 334. 28 Smyth GK. Linear models and empirical bayes methods for assessing differential 2 Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A et al. Distinct types expression in microarray experiments. Statist Appl Genet Mol Biol 2004; 3: Article3. fi fi of diffuse large B-cell lymphoma identi ed by gene expression pro ling. Nature 29 Chambwe N, Kormaksson M, Geng H, De S, Michor F, Johnson NA et al. Variability 2000; 403: 503–511. in DNA methylation defines novel epigenetic subgroups of DLBCL associated with 3 Rosenwald A, Wright G, Chan WC, Connors JM, Campo E, Fisher RI et al. The use of different clinical outcomes. Blood 2014; 123: 1699–1708. molecular profiling to predict survival after chemotherapy for diffuse large-B-cell 30 Beroukhim R, Getz G, Nghiemphu L, Barretina J, Hsueh T, Linhart D et al. Assessing lymphoma. N Engl J Med 2002; 346: 1937–1947. the significance of chromosomal aberrations in cancer: methodology and appli- 4 Ding BB, Yu JJ, Yu RY, Mendez LM, Shaknovich R, Zhang Y et al. Constitutively cation to glioma. Proc Natl Acad Sci USA 2007; 104: 20007–20012. activated STAT3 promotes cell proliferation and survival in the activated B-cell 31 Brooks DG, Trifilo MJ, Edelmann KH, Teyton L, McGavern DB, Oldstone MB. subtype of diffuse large B-cell lymphomas. Blood 2008; 111: 1515–1523. Interleukin-10 determines viral clearance or persistence in vivo. Nat Med 2006; 12: 5LamLT,WrightG,DavisRE,LenzG,FarinhaP,DangLet al. Cooperative signaling 1301–1309. through the signal transducer and activator of transcription 3 and nuclear factor-{kappa} 32 Ganetsky A. Ruxolitinib: a new treatment option for myelofibrosis. Pharma- B pathways in subtypes of diffuse large B-cell lymphoma. Blood 2008; 111:3701–3713. cotherapy 2013; 33:84–92. 6 Goode I, Xu H, Ildstad ST. Regulatory B cells: the new 'it' cell. Transplant Proc 2013; 33 Bai W, Liu H, Ji Q, Zhou Y, Liang L, Zheng R et al. TLR3 regulates mycobacterial 46:3–8. RNA-induced IL-10 production through the PI3K/AKT signaling pathway. Cell 7 Lu X, Nechushtan H, Ding F, Rosado MF, Singal R, Alizadeh AA et al. Distinct IL-4- Signal 2014; 26:942–950. induced gene expression, proliferation, and intracellular signaling in germinal 34 Strle K, Zhou JH, Broussard SR, Venters HD, Johnson RW, Freund GG et al. IL-10 center B-cell-like and activated B-cell-like diffuse large-cell lymphomas. Blood promotes survival of microglia without activating Akt. J Neuroimmunol 2002; 122: 2005; 105: 2924–2932. 9–19. 8 Deenick EK, Avery DT, Chan A, Berglund LJ, Ives ML, Moens L et al. Naive and 35 Ci W, Polo JM, Cerchietti L, Shaknovich R, Wang L, Yang SN et al. The BCL6 memory human B cells have distinct requirements for STAT3 activation to dif- transcriptional program features repression of multiple oncogenes in primary B ferentiate into antibody-secreting plasma cells. J Exp Med 2013; 210: 2739–2753. cells and is deregulated in DLBCL. Blood 2009; 113:5536–5548. 9 Rutsch S, Neppalli VT, Shin DM, DuBois W, Morse HC III, Goldschmidt H et al. IL-6 and 36 Cerchietti LC, Yang SN, Shaknovich R, Hatzi K, Polo JM, Chadburn A et al. A MYC collaborate in plasma cell tumor formation in mice. Blood 2010; 115:1746–1754. peptomimetic inhibitor of BCL6 with potent antilymphoma effects in vitro and 10 Zheng MM, Zhang Z, Bemis K, Belch AR, Pilarski LM, Shively JE et al. The systemic in vivo. Blood 2009; 113: 3397–3405. cytokine environment is permanently altered in multiple myeloma. PLoS One 37 Lim KH, Yang Y, Staudt LM. Pathogenetic importance and therapeutic 2013; 8: e58504. implications of NF-kappaB in lymphoid malignancies. Immunol Rev 2012; 246: 11 Soutar RL, Dillon JM, Brown D, Ralston SH. Cytokine expression in multiple 359–378. myeloma and monoclonal gammopathy: analysis by reverse transcription/poly- 38 Staudt LM. Oncogenic activation of NF-kappaB. Cold Spring Harbor Perspect Biol merase chain reaction and quantitative PCR. Leuk Lymphoma 1996; 24:111–120. 2010; 2: a000109.

© 2015 Macmillan Publishers Limited Leukemia (2015) 1684 – 1694 Therapeutic targeting of IL10 receptor in DLBCLs W Béguelin et al 1694 39 Davis RE, Brown KD, Siebenlist U, Staudt LM. Constitutive nuclear factor kappaB 46 Williams NK, Bamert RS, Patel O, Wang C, Walden PM, Wilks AF et al. Dissecting activity is required for survival of activated B cell-like diffuse large B cell specificity in the Janus kinases: the structures of JAK-specific inhibitors complexed lymphoma cells. J Exp Med 2001; 194: 1861–1874. to the JAK1 and JAK2 protein tyrosine kinase domains. J Mol Biol 2009; 387: 40 Kalampokis I, Yoshizaki A, Tedder TF. IL-10-producing regulatory B cells (B10 cells) 219–232. in autoimmune disease. Arthritis Res Ther 2013; 15:S1. 47 Manshouri T, Quintas-Cardama A, Nussenzveig RH, Gaikwad A, Estrov Z, Prchal J 41 Matsuda M, Salazar F, Petersson M, Masucci G, Hansson J, Pisa P et al. Interleukin et al. The JAK kinase inhibitor CP-690,550 suppresses the growth of human 10 pretreatment protects target cells from tumor- and allo-specific cytotoxic polycythemia vera cells carrying the JAK2V617F mutation. Cancer Sci 2008; 99: T cells and downregulates HLA class I expression. J Exp Med 1994; 180: 1265–1273. 2371–2376. 48 Lichtenstein A, Tu Y, Fady C, Vescio R, Berenson J. Interleukin-6 inhibits apoptosis 42 Harrison C, Mesa R, Ross D, Mead A, Keohane C, Gotlib J et al. Practical man- of malignant plasma cells. Cell Immunol 1995; 162:248–255. agement of patients with myelofibrosis receiving ruxolitinib. Expert Rev Hematol 49 Umezawa K. Possible role of peritoneal NF-kappaB in peripheral inflammation and 2013; 6: 511–523. cancer: lessons from the inhibitor DHMEQ. Biomed Pharmacother Biomed Phar- 43 Verstovsek S, Mesa RA, Gotlib J, Levy RS, Gupta V, DiPersio JF et al. A double-blind, macother 2011; 65:252–259. placebo-controlled trial of ruxolitinib for myelofibrosis. N Engl J Med 2012; 366: 50 Horiguchi Y, Kuroda K, Nakashima J, Murai M, Umezawa K. Antitumor effect of a 799–807. novel nuclear factor-kappa B activation inhibitor in bladder cancer cells. Expert 44 Yamaoka K, Tanaka Y. Jak inhibitor; possibility and mechanism as review of anticancer therapy 2003; 3:793–798. a new disease modifying anti-rheumatic drug. Jpn J Clin Immunol 2009; 32: 51 Yasuda A, Kondo S, Nagumo T, Tsukamoto H, Mukudai Y, Umezawa K et al. Anti- 85–91. tumor activity of dehydroxymethylepoxyquinomicin against human oral squamous 45 Pardanani A, Lasho T, Smith G, Burns CJ, Fantino E, Tefferi A. CYT387, a selective cell carcinoma cell lines in vitro and in vivo. Oral oncology 2011; 47:334–339. JAK1/JAK2 inhibitor: in vitro assessment of kinase selectivity and preclinical stu- 52 Shimada C, Ninomiya Y, Suzuki E, Umezawa K. Efficient cellular uptake of the dies using cell lines and primary cells from polycythemia vera patients. Leukemia novel NF-kappaB inhibitor (-)-DHMEQ and irreversible inhibition of NF-kappaB in 2009; 23: 1441–1445. neoplastic cells. Oncology research 2010; 18:529–535.

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