B Effector Cells Activated by a Chimeric Protein Consisting of IL-2 and the Ectodomain of Tgfb Receptor II Induce Potent Antitumor Immunity

Published OnlineFirst January 12, 2012; DOI: 10.1158/0008-5472.CAN-11-1659

Cancer Therapeutics, Targets, and Chemical Biology Research

B Effector Cells Activated by a Chimeric Protein Consisting of IL-2 and the Ectodomain of TGFb Receptor II Induce Potent Antitumor Immunity

Claudia Penafuerte1,2, Spencer Ng3,4, Norma Bautista-Lopez2, Elena Birman2, Kathy Forner2, and Jacques Galipeau2,4,5

Abstract We have previously shown that interleukin (IL)-2 receptor–expressing lymphoid cells stimulated with a chimeric protein linking IL-2 to the ectodomain of TGFb receptor II (also known as FIST) become resistant to TGFb-mediated suppression and produce significant amounts of proinflammatory cytokines. In this study, we have characterized the antigen presentation properties of FIST-stimulated B cells (hereafter inducible B effector cells, iBEC). FIST converts na€ve splenic B cells to B effector cells characterized by potent antigen presentation properties and production of TNFa and IFNg. iBECs display hyperphosphorylation of STAT3 and STAT5 downstream of the IL-2 receptor and upregulation of T-bet expression. iBECs maintain B-cell identity based on the expression of PAX5 and CD19 and overexpress Smad7, which confers resistance to TGFb-mediated suppression of B-cell activation. iBEC antitumor immunity was determined by a mouse model of lymphoma- expressing ovalbumin (E.G7-OVA) as a specific tumor antigen. OVA-pulsed iBECs function as antigen-presenting þ þ cells (APC) in vitro by inducing the activation of OVA-specific CD4 and CD8 T cells, respectively, and in vivo by conferring complete protective immunity against E.G7-OVA tumor challenge. In addition, OVA-pulsed iBECs promote tumor regression in immunocompetent C57Bl/6 mice bearing E.G7-OVA tumors. In conclusion, iBECs represent an entirely novel B cell–derived APC for immune therapy of cancer. Cancer Res; 72(5); 1–11. 2012 AACR.

Introduction able from nonstem cell source with small amounts of periph- Antigen presentation is required for the development of eral blood. fi effective cell-mediated immunity (1). Dendritic cells (DC) have A new B-cell classi cation takes into account: (i) the pattern been extensively used as cellular adjuvants to present antigen of polarized cytokines produced by B cells, (ii) the cytokine in vivo (2, 3). Although highly effective in their ability to induce environment in which they were stimulated during their T cell–mediated immunity, the clinical applicability of DCs has primary encounter with antigens, and (iii) the T-cell subsets encountered several disadvantages. First, they are relatively they interact with. Two effector B-cell subsets have been fi fi rare in peripheral blood (<1% of leukocytes) and are therefore identi ed based on this classi cation system. The B cell usually isolated from leukapheresis or marrow sources (4). stimulated by TH1 effectors were coined B effector 1 (Be1) Second, they are comprised by a heterogeneous population cells and B cells stimulated by TH2 effectors are known as B with distinctive functions and third, they are difficult to expand effector 2 (Be2) cells (6). Be1 cells mainly produce TH1 cyto- in vitro from a nonstem cell source (5). As an alternative source kines [IFNg and interleukin (IL)-12], whereas Be2 produce TH2 of antigen-presenting cells (APC), B cells are reliably expand- cytokines (IL-4 and IL-5). The combination of antigen stimulation and interaction with polarized effector TH1 and TH2 cells dictate the differentiation of na€ve B cells toward IFNg or IL-4– producing cells capable of triggering a T 1orT 2 differenti- Authors' Affiliations: 1Department of Experimental Medicine, McGill Uni- H H versity; 2Lady Davis Institute for Medical Research, Montreal, Canada; ation program in na€ve T cells, respectively (6, 7). 3Emory University Graduate Program in Immunology and Molecular Path- Like DCs, B cells can process and present antigens in the ogenesis of the Graduate Division of Biological and Biomedical Sciences; fi þ Departments of 4Hematology and Medical Oncology and 5Pediatrics, context of MHC class II molecules and recruit speci c CD4 Emory University Winship Cancer Institute, Atlanta, Georgia T cell help, which in turns stimulates B-cell proliferation and differentiation (8, 9). In addition, B cells express high levels of Note: Supplementary data for this article are available at Cancer Research þ Online (http://cancerres.aacrjournals.org/). MHC class I and therefore act as APC for CD8 T cells, which Corresponding Author: Jacques Galipeau, Winship Cancer Institute, 1365 stimulates IL-2 production and cytotoxic T cells (CTL) activity Clifton Road, Atlanta, GA 30322. Phone: 404-778-1779; Fax: 404-778- (10, 11). It has also been shown that B cell–derived APCs are 1267; E-mail: [email protected] comparable with mature DCs in regards to presentation doi: 10.1158/0008-5472.CAN-11-1659 function in vivo (12). However, the method required to convert 2012 American Association for Cancer Research. and expand na€ve B cells to APCs requires coculture with

www.aacrjournals.org OF1

Penafuerte et al.

CD40L-expressing transfected cell lines or coculture with cytometers (BD) and analyzed by Cellquest software (BD) or T-helper cells (13), which may represent a substantial logistical Flow Jo. and regulatory obstacle to clinical translation of B cell–derived APCs in first-in-human use for immunotherapy purposes. Intracellular signaling, cytokine profile, and cell Here, we describe an entirely novel remedy to this conundrum proliferation which will allow further development of B cell–derived APCs as For signaling analysis, iBEC lysates were probed by Western a distinct APC for immunotherapy. blot with antiphosphorylated or antitotal STAT3, STAT1, We have previously shown that molecular genetic engineer- STAT5, BCL6, BLIMP-1, SMAD7, and PAX5 antibodies. Anti– ing of a fusion protein linking IL-2 to a TGFb decoy receptor b-tubulin and total protein antibodies were used as loading (14) adopts a hybrid IL-2-like cytokine function characterized controls. The concentration of cytokines produced by iBEC or by inhibition of TGF-b signaling and hyperactivation of IL-2 control-activated B cells (IL-5, IL-12, IL-17, IL-4, GM-CSF, signaling via STAT1. Here, we show that na€ve spleen-derived B TNFa, IFNg, and IL-6) was quantified by ELISA. The expression cells stimulated in vitro with FIST undergo a massive mitogenic of IFNg, IL-4, and IL-17 by CD19-positive cells was verified by response and adopt an entirely novel APC phenotype and flow cytometry with conjugated antibodies specific for each functional profile distinct from that previously described for cytokine. For T-bet, IFNg, IL-4, or IL-17 expression, iBEC or OT2 þ Be1-, Be2-, and CD40L-activated B cells. We further show that CD4 T cells were treated with Brefaldin A for 4 hours, these FIST-inducible B effector cells (iBEC) can lead to antigen- permeabilized, fixed, and stained with T-bet, IFNg, IL-4, or specific antitumor activity and immune protection against IL-17 conjugated antibodies or isotype controls for 1 hour. The tumor challenge in vivo. cell proliferation rate of iBECs by prestaining them with carboxyfluorescein succinimidyl ester (CFSE) and the percent- Materials and Methods age of proliferating cells were quantified by flow cytometry. Mice, reagents, and antibodies The data were analyzed by Cellquest software (Becton All experimental C57BL/6, transgenic mice (OT1 and OT2) Dickinson). þ þ and knockout mice (CD4 and CD8 T cell) were females of 6 In vitro to 8 weeks old (Jackson Laboratory). Antiphosphorylated and APC assay in vitro total length STAT3, STAT5, STAT1, BCL6, and BLIMP-1 anti- To assess the antigen presentation ability of iBECs , 4 bodies were obtained from Cell Signalling Technology; b-tubu- 1 10 B cells previously stimulated with FIST or control and lin and SMAD7 antibodies were obtained from Santa Cruz full-length ovalbumin (OVA)-pulsed B cells were washed 3 5 þ þ Biotechnology. Anti-mouse FcR III/II, CD79b (BCR subunit), times and cocultured with 1 10 CD4 or CD8 T cells T-bet, IFNg, IL-4, IL-17, MHC class I and II, CD19, CD4, CD45R isolated from OT2 and OT1 mice, respectively. After 48 hours, (B220), CD24, CD22.2, CD43, CD279, CD138, CD27, CD80, CD86, the supernatants were collected to quantify the concentration CD25, CD69, CD23, CCR7, FasL, TRAIL, CD95, and CD40 and of IL-2 and IFNg by mouse enzyme-linked immunosorbent fl assay (ELISA) kits (R&D). The proliferation rates of OVA- the isotype control antibodies for ow cytometry were þ þ specific CD4 or CD8 T cells were determined by culturing obtained from BD Biosciences. þ þ CFSE prestained CD4 or CD8 T cells with OVA-pulsed iBEC B-cell isolation, generation, and phenotypic analysis of or control-stimulated B cells for 24 hours, and the percentage FIST-stimulated B cells of proliferating cells was quantified by flow cytometry. The B cells were isolated from splenocytes of immunocompetent data were analyzed by Cellquest software (Becton Dickinson). C57BL/6 by magnetic separation with the EasySep Mouse B Cell Isolation Kit (Stemcell Technologies) according to the iBEC cytotoxicity and iBEC-dependent induction of CTL manufacturer's recommendations. B-cell population purity activity and natural killer cell activation assessed by flow cytometry was 96%. FIST or control-stimu- A total of 5 104 iBEC or control-activated B cells were lated B cells were generated by culturing B cells with 2 pmols cocultured with 5 104 E.G7 or TUBO cells for 5 hours to carry per milliliter of FIST or controls (IL-2, sTbRII, and IL-2 plus out an in vitro nonradioactive cytotoxicity assay according to sTbRII) for 3 to 4 days at 37C. FIST and control protein were the manufacturer's recommendations (Promega). To deter- obtained as previously described (14). For cell surface marker mine the induction of CTL activity by iBECs, 5 104 OVA- staining, iBECs were resuspended in PBS with 2% FBS, incu- pulsed iBEC or control-activated and OVA-pulsed B cells were bated with anti-human FcR III/II for 30 minutes and labeled cocultured with 1 105 OT1 CD8 T cells. After 24 hours of with conjugated antibodies specific for CD79b (BCR subunit), incubation, 5 104 E.G7 cells were added to the mixture of cells MHC class I and II, CD19, CD45R (B220), CD24, CD22.2, CD43, for another 5 hours to carry out an in vitro nonradioactive CD279, CD138, CD27, CD80, CD86, CD25, CD69, CD23, CD95, cytotoxicity assay according to the manufacturer's recommen- CCR7, FasL, TRAIL, and CD40. The expression of these cell dations (Promega). The percentage of cytotoxicity was calcu- surface markers was determined by FACS Calibur and Canto lated according with the following equation:

Experimental Effector spontaneous Target spontaneous % Cytotoxicity ¼ 100 Target maximum Target spontaneous

OF2 Cancer Res; 72(5) March 1, 2012 Cancer Research

IL-2 and sTbRII Fusion Protein and Antigen-Presenting B Cell

A B Media IL-2 IL-2 FIST

sTβRII IL-2 + sTβRII FIST

C Media IL-2 FIST 250K 250K 250K 200K 200K 200K 150K 150K 150K 100K 100K 100K SSC-A SSC-A SSC-A 50K 50K 50K 0 0 0 0102 103 104 105 0102 103 104 105 0102 103 104 105 Figure 1. FIST induces B-cell 800 800 800 activation and proliferation. A, B cells 600 9 600 11 600 72 were cultured with 2 pmol of FIST or 400 400 400 # Cells # Cells controls for 4 days and labeled with # Cells 200 200 200 0 0 0 CD19 antibody. B-cell size and 0102 103 104 105 0102 103 104 105 0102 103 104 105 granularity was assessed based on CFSE CFSE CFSE the FSC and SSC pattern determined D by ﬂow cytometry. B, hematoxylin Media IL-2 sTβRII IL-2 + sTβRII FIST and eosin staining of FIST- or control-activated B cells. C, B cells 529 536 510 504 647 prestaining with CFSE were

incubated with FIST or controls for 4 MHC II days. The decrease of CFSE intensity as indicative of cell proliferation was 623 613 624 610 determined by ﬂow cytometry. D, 642 iBEC or control-stimulated B cells analyzed for the expression of B-cell CD45R activation markers, costimulatory and MHC molecules. See also 282 290 292 303 418 Supplementary Fig. S1. CD86

153 141 152 159 174 CD27

179 168 175 160 251 CD25

155 141 132 134 182 CD69

172 204 192 228 393 MHC I

To determine whether iBECs induce natural killer (NK) cell magnetic cell separation with the EasySep Mouse NK Cell activation, 5 104 iBEC or control-stimulated B cells were Isolation Kit (Stemcell Technologies). After 24 hours of incu- cocultured with 1 105 NK cells. NK cells were isolated (92% of bation, NK cells were collected to quantify the expression of purity) from spleen of wild-type C57B/6 mice by immune- CD69 on the surface of NK1.1-positive cells by ﬂow cytometry.

www.aacrjournals.org Cancer Res; 72(5) March 1, 2012 OF3

Penafuerte et al.

The data were analyzed by Cellquest software (Becton Dickinson). A In vivo experiments IL-2 –+–+ – 5 In a prophylactic setting, 1 10 OVA-pulsed iBEC or sTβRII – – + –+ control-stimulated B cells were subcutaneously injected into FIST – – – +– C57BL/6 immunocompetent mice, and 3 weeks later mice were 5 boosted with 4 10 OVA-pulsed iBEC or control-stimulated B P-STAT 3 cells. One week after the second boost, mice were challenged with 5 105 E.G7 tumor cells. Similar experiments were þ þ repeated with CD4 and CD8 T cells knockout mice. Tumor T-STAT 3 volume and percentage of survival were determined over time. In a therapeutic setting, 5 105 E.G7 tumor cells were P-STAT 5 subcutaneously implanted into C57BL/6 immunocompetent mice and once mice developed palpable tumors, they were T-STAT 5 treated with 2 doses of 3 106 OVA-pulsed iBEC or control- stimulated B cells. Tumor volume and percentage of survival were determined over time. B

Statistical analysis PAX5 We used the 2-tailed unpaired Student t test to compare 2 experimental groups, the Dunnett multiple comparisons to SMAD7 compare 3 or more test groups.

Results BCL-6 FIST induces B-cell activation and differentiation to an APC effector phenotype B cells cultured with FIST for 3 to 4 days undergo a BLIMP1 differentiation process characterized by phenotypic changes fi including an increase in size and granularity de ned by side β-Tubulin scattergram profile (SSC) and forward scattergram profile (FSC; Fig. 1A) and hematoxylin and eosin staining (Fig. 1B). In addition, iBECs display a significant mitogenic response Figure 2. iBECs display a potent activation of JAK/STAT pathway and its compared with IL2-stimulated B cells in vitro (Fig. 1C). target gene expression. A total of 5 106 of iBEC or control-stimulated B sTbRII by itself or in combination with IL-2 does not induce cells were lysated and probed for (A) phosphorylated STAT3 and STAT5 and their respective anti–full-length protein antibodies and (B) PAX5, B-cell proliferation (Supplementary Fig. S1A). To character- SMAD7, BCL-6, BLIMP1, and b-tubulin as loading control. ize the phenotype of FIST-expanded cells (iBEC), we measured the expression levels of several B-cell receptors and costimulatory molecules including CD79b (BCR subunit), contrast, iBECs in culture for 3 days display a distinctive MHC class I and II, CD19, CD45R (B220), CD24, CD22.2, hyperactivation of both STAT3 and STAT5 (Fig. 2A). Pax5 is CD43, CD138, CD27, CD80, CD86, CD25, CD69, CD23, CD95, highly expressed in all the conditions tested whether with CCR7, FasL, and CD40. In comparison with control-stimu- iBEC or control-stimulated B cells. As previously observed for lated B cells (2 pmols/mL of IL-2, sTbRII or IL2 plus sTbRII), CTLL-2 and splenocytes (14), FIST also upregulates SMAD7 iBEC upregulate costimulatory molecules (CD86), markers expression in B cells. Paradoxically, iBECs display a robust of activation (CD69), B-cell memory marker (CD27), B-cell upregulation of BLIMP1 expression combined with BCL6 identity markers CD45R (B220) and CD19, IL-2 receptor downregulation (Fig. 2B) in the absence of the plasma cell alphasubunit(CD25),aswellasMHCclassIandIIexpres- marker CD138 (data not shown). sion (Fig. 1D). In addition, iBECs display significantly higher migration to the lymph nodes than resting B cells but similar FIST imprints a Be1-like phenotype migration as control-activated B cells (IL-2, sTbRII, and IL-2 To interrogate whether FIST promotes a Be1-like phenotype, combined with sTbRII; Supplementary Fig. S1B). T-bet expression and IFNg secretion were measured over time in B cells (from immunocompetent mice and STAT1 knockout FIST induces B-cell activation via Janus-activated mice) cultured with FIST. FIST induces a significant increase in kinase/STAT signaling pathway and Smad7 expression T-bet expression and IFNg production by both B cells from iBECs and B cells stimulated by equimolar concentrations of immunocompetent mice (Fig. 3A and B) and STAT1 knockout IL-2 combined with sTbRII for 30 minutes show similar mice (Supplementary Fig. S2A). IFNg expression by CD19- activation levels of STAT3 and STAT5 (data not shown), and positive cells was verified by intracellular staining with IFNg- no STAT1 phosphorylation was detected (data not shown). In specific conjugated antibodies (Supplementary Fig. S2B).

OF4 Cancer Res; 72(5) March 1, 2012 Cancer Research

IL-2 and sTbRII Fusion Protein and Antigen-Presenting B Cell

A B Media IL-2 sTβRII IL-2+sTβRII FIST 3,000 70 2,500 60 * 50 * 2,000 40 1,500 30 IFN γ pg/mL 1,000 T-bet MFI T-bet 20 10 500 0 0 024487296 Figure 3. FIST imprints a Be1-like 02448 72 96 Hours phenotype. A, T-bet expression over Hours time by 1 105 of iBEC or control- C D stimulated B cells. B, IFNg 5 expression over time by 1 10 of 500 * 300 iBEC or control-stimulated B cells. 400 250 *

T-bet and IFN expression over time cells/72 h g 5 200 cells/72 h

5 300 was determined by intracellular 150 staining and the mean fluorescence 200 100 intensity (MFI) values were 100 50 fl measured by ow cytometry. TNFa 0 0 (C), GM-CSF (D), and IL-6 (E) 5 IL-2 βRII IL-2 production by 1 10 FIST or TNF α pg/1 × 10 FIST FIST Media sTβRII Media sTβRII control-stimulated B cells for 72 GM-CSF pg/1 × 10 hours. F, IL-17 expression IL-2 + sT IL-2 + sTβRII measured by intracellular staining of E CD19-positive cells previously 300 * stimulated with FIST or controls for /72 h 250 5 72 hours. G, IL-4 expression 200 measured by intracellular staining of 150 B220-positive cells previously 100 stimulated with FIST or controls for 50 72 hours. Significant differences IL-6 pg/1 × 10 0 between FIST and controls in all RII fi IL-2 β FIST gures are denoted by asterisks: RPMI sTβRII , P < 0.05. These results are representative of 3 independent IL-2 + sT experiments carried out in F Isotype Media IL-2 sTβRII IL-2 + sTβRII FIST 4 4 4 4 triplicates. See also Supplementary 4 4 10 10 10 10 Fig. S2. 10 10 0% 0% 0% 0% 0% 0 CD19 0 0 0 0 0 CD19FITC CD19FITC CD19FITC CD19FITC CD19FITC rlgG2 α FITC 10 10 10 10 10 0 1 2 3 4 10 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 100 101102 103104 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 100 101 102 103 104 rlgG1 PE IL-17 PE IL-17 PE IL-17 PE IL-17 PE IL-17 PE IL-17 β β G 4 Isotype 4 Media 4 IL-2 4 sT RII 4 IL-2 + sT RII 4 FIST 10 10 10 10 10 10 0% 0% 0% 0% 0% 0 0 0 IL-4 0 0 0 IL-4 FITC IL-4 FITC IL-4 FITC IL-4 FITC IL-4 FITC 10 10 10 10 10 rat lgG1-FITC rat 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 100 101102 103104 rat lgG2a PE B220 PE B220 PE B220 PE B220 PE B220 PE B220

þ In addition, iBECs secrete high levels of proinﬂammatory context of I-Ab, respectively. iBECs only induce OT2 CD4 cytokines (GM-CSF, TNFa, and IL-6; Fig. 3C and D) and no T-cell activation when they were previously pulsed with OVA detectable levels of IL-5, IL-15, and IL-12 measured by ELISA antigens, whereas no IL-2 or IFNg was detected in the absence (data not shown). There was no detectable IL-4 and IL-17 by of OVA loading (Supplementary Fig. S3A and S3B). OVA-pulsed þ þ intracellular cytokine staining of CD19- or B220-positive cells iBECs promote signiﬁcantly greater CD4 and CD8 T-cell (Fig. 3E and F). activation and TH1 cytokine (IL-2 and IFNg) production than OVA-pulsed B cells previously stimulated with equimolar iBECs promote T-cell activation and polarization concentration of cytokine controls (Fig. 4A and B). To char- OT1 and OT2 T cells were designed to recognize OVA 257 to acterize the type of T helper–mediated immunity induced by þ 264 residues in the context of H2Kb or 323 to 339 residues in the iBECs, OT2 CD4 T cells were cocultured with OVA-pulsed

www.aacrjournals.org Cancer Res; 72(5) March 1, 2012 OF5

Penafuerte et al.

A B OT2 mFIST Figure 4. mFIST OT1 * iBECs induce T-cell * * activation, proliferation, and * IL-2 + sTβRII IL-2 + sTβRII polarization into TH1 effector cells. 4 β After 4 days in culture, 1 10 β sT RII sT RII iBECs were pulsed with chicken IL-2 IL-2 OVA for additional 24 hours, washed, and cocultured with Media Media 1 105 OT2 CD4þ T cells or 1 105 OT1 CD8þ T cells for 48 hours. IL-2 0 200 400 600 800 0 200 400 600 800 (A) and IFNg (B) concentrations IL-2 pg/mL IFNγ pg/mL produced by OT1 and OT2 T cells primed by OVA-pulsed iBECs. C, C IFNg, IL-4, and IL-17 expressed by 50 OT2 CD4þ T previously primed by IFNγ IL-4 IL-17 * 45 OVA-pulsed iBEC or control- 40 stimulated B cells were determined 35 by intracellular staining with 30 fluorescent-conjugated specific 25 antibodies and the mean MFI fluorescence intensity (MFI) values 20 were measured by flow cytometry. 15 þ þ D, OT1 CD8 T cells and OT2 CD4 10 T cells prestaining with CFSE were 5 cocultured with OVA-pulsed 0 iBECs. The decrease of CFSE Media IL-2 sTβRII IL-2 + sTβRII FIST intensity as indicative of cell proliferation was assessed by flow D cytometry. Significant differences Media IL-2 sTβRII IL-2 + sTβRII FIST between FIST and controls in all figures are denoted by asterisks: 2% 4% 4% 7% 15% , P < 0.05. The results are OT1 Counts Counts Counts Counts Counts representative of 3 independent 0 200 0 200 0 200 0 200 0 200 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 experiments carried out in CFSE CFSE CFSE CFSE CFSE triplicates (data are shown as the 4% 4% 6% 7% 15% mean SD). See also Supplementary Fig. S3. OT2 Counts Counts Counts Counts Counts 0 200 0 200 0 200 0 200 0 1 2 3 4 0 200 10 10 10 10 10 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 CFSE CFSE CFSE CFSE CFSE

iBECs for 24 hours and the intracellular expression of IL-4, OVA-pulsed iBEC induce protective immunity against þ IL-17, or IFNg in OT2 CD4 T cells were measured by flow OVA-expressing tumor challenge – cytometry. OVA-pulsed iBECs induce a TH1 cell mediated OVA-pulsed iBEC conferred complete protection to mice immunity based on a significant upregulation of IFNg expres- from E.G7-OVA tumor challenge. In contrast, the majority sion as the main TH1 polarized cytokine (Fig. 4C). Both OVA- (60%–80%) of mice previously treated with nonstimulated or þ þ specific CD4 and CD8 T cells proliferate in vitro in response control-stimulated and OVA-pulsed B cells developed tumors to OVA presented by iBECs (Fig. 4D). and were sacrificed (Fig. 6A, B, and C). To characterize the antigen presentation properties of OVA-pulsed iBECs in vivo, þ þ similar experiments were carried out in CD4 and CD8 T cell iBECs as effector and inducer of CTL activity and NK cell knockout mice. Although both strains developed tumors, activation indicating that OVA-pulsed iBECs failed to induce protective þ iBECs upregulate TRAIL expression (Fig. 5A), but they do not antitumor immunity, tumor grew faster in CD4 T cell knock- þ display significant in vitro cytotoxic activity against either a out mice. Consequently, CD4 T cell knockout mice display a þ DR5- null (EG7) or a DR5 (TUBO) tumor cell line (data not percentage of survival similar to control groups (Fig. 6D and E). þ þ shown). In contrast, iBECs promote CTL activity in OT1 CD8 On the contrary, in CD8 T cell knockout mice, the tumors T cells in an antigen-dependent manner. CTL cells induced by grew with significant delay compared with the control group iBECs display significant cytotoxicity (55% versus no killing by and 20% of the mice display protective immunity against control-stimulated cells) against E.G7 tumor cells (Fig. 5B). In tumor challenge (Fig. 6F and G). Nine weeks after tumor addition, iBECs induce NK cell activation based on the expres- implantation, treated mice sera were collected to quantify sion of CD69 (Fig. 5C). Activated NK cells display an increase in OVA-specific antibody titer by ELISA. No significant differ- size and granularity defined by SSC and FSC(Fig. 5D). ences were observed in the level of OVA-specific antibodies

OF6 Cancer Res; 72(5) March 1, 2012 Cancer Research

IL-2 and sTbRII Fusion Protein and Antigen-Presenting B Cell

A Media IL-2 FIST

1,500 1,500 1,500 MFI = 715 MFI = 752 MFI = 806 1,000 1,000 1,000 # Cells # Cells # Cells 500 500 500

0 0 0 0102 103 104 105 0102 103 104 105 0102 103 104 105 TRAIL PE TRAIL PE TRAIL PE B Figure 5. iBECs as effector and 70 inducer of CTL activity and NK cell activation. A, TRAIL expression on 60 * CTL cells FIST or control-stimulated B cells. B, 5 50 cytotoxicity activity of OT1 CD8 T cells activated by iBEC or control- 40 stimulated B cells against E.G7 30 tumor cells. C, CD69 expression on NK cells previously activated by iBEC 20 or control-stimulated B cells. D, NK 10 cell size and granularity was assessed based on the FSC and SSC 0 pattern determined by flow of 1 × 10 % of cytotoxicity IL-2 FIST cytometry. Significant differences in Media sTβRII all figures are denoted by asterisks: , P < 0.05. The results are IL-2 + sTβRII representative of 2 independent CDNK cells + control B cells experiments carried out in triplicates Isotype Media IL-2

(data are shown as the mean SD). 1,000

0% 1% 2% SSC-H Counts Counts Counts 0200 0200 0200 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 0 Ham IgG-Cy7 CD69-Cy7 CD69-Cy7 0 200 400 600 800 1,000 FSC-H β β sT RII IL-2 + sT RII FIST NK cells + iBEC cells 1,000 1% 2% 6% Counts Counts Counts 0200 0200 0200

0 1 2 3 4 SSC-H 10 10 10 10 10 100 101 102 103 104 100 101 102 103 104 CD69-Cy7 CD69-Cy7 CD69-Cy7 0 0 200 400 600 800 1,000 FSC-H

obtained from mice treated with OVA-pulsed iBECs and IL-2 of 7 months (data not shown). Overall, mice treated with plus sTbRII-stimulated B cells (Supplementary Fig. S4). OVA-pulsed iBECs displayed significant lower rate of tumor growth over time (Fig. 7E). OVA-pulsed iBEC treatment significantly inhibits tumor growth C57BL/6 mice with preestablished tumors (previously Discussion injected with 5 105 E.G7-OVA tumor cells) were treated B cells, T cells, and NK cells do not respond to IL-2 in the with OVA-pulsed iBECs or control-stimulated B cells. Two same fashion. In T cells and NK cells, the stimulation of IL-2 intravenous injections of 3 106 B cells were given a week receptor induces cell activation and proliferation, which is apart and tumor volume was determined over time (Fig. 7A). driven by activation of STAT3, STAT5, and STAT1. In contrast, Mice bearing palpable tumors (previously implanted with na€ve B cells do not proliferate in the presence of IL-2. In T and 5 105 E.G7 cells) and treated with OVA-pulsed B cells NK cells, FIST induces distinctive STAT3 and STAT1 activation developed tumors that progressed in a similar manner to the compared with equimolar concentrations of IL-2 or IL-2 nontreated control group (Fig. 7B and C). In contrast, mice combined with sTbRII, which suggest that FIST may bind with treated with OVA-pulsed iBECs displayed a delay of tumor higher affinity or avidity to the IL-2 receptor (14). Similarly, in B growth and 20% of treated mice showed complete tumor cells, FIST induces aberrant hyperactivation of STAT3 and rejection (Fig. 7D), which remained tumor free for a period STAT5, as well as upregulation of their target gene expression.

www.aacrjournals.org Cancer Res; 72(5) March 1, 2012 OF7

Penafuerte et al.

A Tumor 3 weeks 1 week measurement

1 × 105 OVA FIST B cells 4 × 105 OVA FIST B cells E.G7 tumor implantation

BC β β Media IL-2 sT RII IL-2+sT RII FIST 100 * Figure 6. iBECs promote complete 400 protective immunity against tumor 80 challenge. A, B-cell therapy 300 * 60 scheme for the prophylactic setting. A total of 1 105 OVA- 200 40 pulsed iBECs were intravenously n ¼ 100 20 injected in C57Bl/6 mice ( 5); 3 weeks later the mice were boosted 0 of survival Percentage 0 with 4 105 OVA-pulsed iBEC, and 10 15 20 25 0 20406080100 1 week after the boost the mice Days posttumor implantation Days posttumor implantation were subcutaneously challenged with 5 105 E.G7 tumor cells DE expressing OVA. Tumor volume (B) ControlB cell FIST B cell and percentage of survival (C) in 4,500 100 CD4T cell KO CD4T cell KO immunocompetent C57Bl/6 mice 4,000 was monitored over time. Tumor 3,500 80 3,000 volume (D) and percentage of 60 2,500 survival (E) was monitored over þ 2,000 40 time in CD4 T cell knockout mice 1,500 (n ¼ 5). Tumor volume (F) and Tumor volume Tumor 1,000 20 percentage of survival (G) was 500 þ Percentage of survival Percentage monitored over time in CD8 T cell 0 0 knockout mice (n ¼ 5). Signiﬁcant 5678910111213141516 04010 20 30 differences in all ﬁgures are Days posttumor implantation Days posttumor implantation denoted by asterisks: , P < 0.05. The results are representative of 2 FG independent experiments (data are 4,500 CD8T cell KO100 CD8T cell KO shown as the mean SD). See also 4,000 Supplementary Fig. S4. 3,500 80 3,000 * 2,500 60 2,000 1,500 40 Tumor volume Tumor 1,000 500 20 Percentage of survival Percentage 0 0 5678910111213141516 0405010 20 30 Days posttumor implantation Days posttumor implantation

Consequently, FIST converts na€ve B cells to a new type of B- upregulates BLIMP1 gene expression promoting plasma cell effector characterized by the appearance of phenotypic cell differentiation, while downregulating BCL6 expression and functional features distinct from Be1 or Be2 cells as (17). In contrast, STAT5 activation leads to self-renewal and previously described. A dramatic upregulation of CD25, acti- inhibition of plasma B-cell differentiation due to upregulation vation marker (CD69), costimulatory molecule (CD86), of BCL6 expression (15, 18). Interestingly, iBECs display hyper- and MHC class (I and II) molecule expression are indicative activation of both STAT3 and STAT5, the sum of whose effects of B-cell responsiveness to FIST stimulatory effects. iBECs are downregulation of BCL6 expression coordinated with a increase in size and granularity, but they do not differentiate robust BLIMP1 upregulation that paradoxically coexists with into plasma cells because they keep their B-cell identity based high levels of Pax5 and MHC class II expression in proliferating on CD19 and B220 expression, proliferate, and express high cells. Therefore, despite STAT3 activation, iBECs do not levels of MHC class II molecules; features that typically acquire a plasma cell phenotype. disappear once B cells are fully differentiated into plasma cells Furthermore, FIST-mediated STAT5 activation may be par- (15, 16). tially antagonized by FIST-mediated STAT3 activation because STAT3 and STAT5 activation in B cells play opposite roles in STAT5 activation fails to inhibit STAT3-dependent BLIMP1 the regulation of B- cell fate. Speciﬁcally, STAT3 activation upregulation and BCL-6 downregulation. Notwithstanding

OF8 Cancer Res; 72(5) March 1, 2012 Cancer Research

IL-2 and sTbRII Fusion Protein and Antigen-Presenting B Cell

Tumor A measurement

5 days 1 week

Figure 7. iBEC treatment promotes 5 × 105 E.G7 cells 3 × 106 OVA FIST B cells 3 × 106 OVA FIST B cells tumor regression in a therapeutic setting. A, B-cell therapy scheme for BD the therapeutic setting. Immunocompetent C57Bl/6 mice 6,000 Control 6,000 OVA FIST B cells n ¼ ( 5) with preestablished E.G7 5,000 5,000 tumors (previously injected with 5 4,000 4,000 5 10 E.G7) were treated with 3 106 OVA-pulsed iBEC or control- 3,000 3,000 stimulated B cells and 1 week later, 2,000 2,000 Tumor volume Tumor the mice were boosted with volume Tumor 1,000 1,000 3 106 OVA-pulsed iBEC or control- 0 stimulated cells. Tumor volume over 0 10 20 30 40 10 20 30 40 time per mouse was measured in nontreated group (B), OVA-pulsed Days posttumor implantation Days posttumor implantation B cell–treated group (C), and OVA- pulsed iBEC-treated group (D). E, CE average of tumor volume over time 6,000 OVA B cells per group. Signiﬁcant differences in 5,000 Control 5,000 all ﬁgures are denoted by asterisks: 4,000 OVA B cells , P < 0.05. The results are 4,000 OVA FIST B cells representative of 3 independent 3,000 3,000 * experiments (data are shown as the 2,000 * mean SD). volume Tumor 2,000 1,000 0 1,000

10 20 30 40 of tumor volume Average 0 Days posttumor implantation 12 16 19 22 27 36 Days posttumor implantation

antagonism with STAT3 activation, iBECs display upregulation such as STAT3 may act as transducers of IFNgR signaling in B of STAT5 target genes such as B-cell identity marker (CD19), cells. CD25, and Pax5 expression as well as cell proliferative capacity As observed in Be1 cells (6), iBEC produce IFNg, TNFa, and (18). IL-6, but not IL-4, IL12, or IL-17. Similarly to Be1- and CD40- FIST antagonizes TGFb signaling by acting as decoy recep- activated B cells, iBECs pulsed with antigens are able to € tor trapping active TGFb and inducing Smad7 expression (14). polarize nave T cells into TH1 cells, as well as promote the TGFb increases the rate of apoptosis of normal resting B cells generation of antigen-specific CTL. via caspase 3 (19) and inhibits CIITA and MHC class II Although iBECs exhibit similar migration capacity to the expression (20). FIST-mediated STAT3 activation may also lymph nodes as control-activated B cells, they promote more þ þ induce Smad7 expression which directly inhibits the TGF-b potent activation and proliferation of CD4 and CD8 pathway in iBECs (21–23). Therefore, iBECs are also resistant T cells in an antigen-dependent manner, as observed in to TGFb-dependent suppression of activation. This functional vitro. In addition, iBECs promote polarization of T cells into TGFb signaling blockade in combination with potent Janus- TH1-like cells that produce high levels of TH1polarized activated kinase (JAK)/STAT activation induces iBECs to cytokines such as IFNg and IL-2 as well as CTL activity produce significant quantities of IFNg, which correlate with against tumor cells expressing T cell–specificantigens. progressive induction of the master regulator of IFNg gene Furthermore, iBECs induce the activation of NK cells, which transcription, T-bet. Previous reports indicate that T-bet and display features characteristic of differentiated lymphokine- IFNgR expression are required for IFNg production by Be1 activated killer cells such as an increase in size and gran- cells. In contrast to T cells, IFNg production in Be1 cells is not ularity. In addition, iBECs display upregulation of TRAIL critically dependent on STAT1 coupled to IFNgR (24). In expression, which suggest that they may directly act as accordance with previous reports, iBECs from STAT1 knock- effector cells by inducing apoptosis of susceptible cells out mice produce similar amounts of IFNg as iBEC from through TRAIL receptor engagement (25). However, we normal mice. This confirms that IFNg production in B cells found that iBECs do not display significant in vitro cytotoxic þ is regulated by a STAT1-independent mechanism and activity against either a DR5 null (EG7) or a DR5 (TUBO) suggests that other transcription factors or STAT proteins tumor cell lines. The sum of these data suggest that TRAIL

www.aacrjournals.org Cancer Res; 72(5) March 1, 2012 OF9

Penafuerte et al.

and IFNg production by iBECs may be insufficient on their kine and chemokine production circumvent the requirement own to mediate direct cytolysis of tumor cells expressing of TH1 polarized cells, transfected feeder cells, and (or not) DR5. cytokines needed to differentiate B cells into Be1 cells. In a cancer vaccine setting, OVA-pulsed iBECs confer Furthermore, this is the first in vivo study that shows the complete protective immunity against tumor cells expres- effectiveness of ex vivo activated B cells to induce antitumor sing specific OVA antigen. Because tumor growth rates from protective immunity. In conclusion, iBECs display a novel both OVA-pulsed iBEC and control groups are similar in B effector phenotype and act as potent APCs that protect þ CD4 T cell knockout mice, we conclude that the potent mice from tumor challenge in an antigen-dependent fashion. protective immunity observed in immunocompetent mice Therefore, FIST could serve as a tool for ex vivo B-cell þ requires help from antigen-specificCD4 T cells primed by expansion and activation for the cell-based therapy of OVA-pulsed iBECs. Although, the OVA-pulsed iBEC-treated cancer. þ group developed tumors in CD8 T cell knockout mice, their tumors grew with a significant delay compared with the Disclosure of Potential Conflicts of Interest nontreated group and 20% of treated mice displayed pro- No potential conflicts of interest were disclosed. þ tective immunity. These results indicate that, besides CD8 T cells, other immune effectors may also contribute to the Grant Support in vivo This work was supported by a Canadian Institute for Health Research antitumor response observed .Indeed,iBECsalso operating grant MOP-15017. induce NK cell activation, which may occur by a cell con- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked tact-dependent mechanism via costimulatory molecule advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this interactions or by the production of proinflammatory cyto- fact. kines by iBECs (26, 27). The use of FIST as a tool to generate and expand Be1- Received May 16, 2011; revised December 27, 2011; accepted December 27, like cells with potent antigen-presenting properties, cyto- 2011; published OnlineFirst January 12, 2012.

References 1. Steinman RM. The dendritic cell system and its role in immunogenicity. class I- and class II-binding peptides. Immunology 2008;124: Ann Rev Immunol 1991;9:271–96. 129–40. 2. Hsu FJ, Benike C, Fagnoni F, Liles TM, Czerwinski D, Taidi B, et al. 13. Schultze JL, Michalak S, Seamon MJ, Dranoff G, Jung K, Daley J, Vaccination of patients with B-cell lymphoma using autologous anti- et al. CD40-activated human B cells: an alternative source of highly gen-pulsed dendritic cells. Nat Med 1996;2:52–8. efficient antigen presenting cells to generate autologous antigen- 3. Nestle FO, Alijagic S, Gilliet M, Sun Y, Grabbe S, Dummer R, et al. specific T cells for adoptive immunotherapy. J Clin Invest 1997;100: Vaccination of melanoma patients with peptide- or tumor lysate- 2757–65. pulsed dendritic cells. Nat Med 1998;4:328–32. 14. Penafuerte C, Bautista-Lopez N, Bouchentouf M, Birman E, Forner K, 4. Siena S, Di NM, Bregni M, Mortarini R, Anichini A, Lombardi L, et al. Galipeau J. Novel TGF-beta antagonist inhibits tumor growth and Massive ex vivo generation of functional dendritic cells from mobilized angiogenesis by inducing IL-2 receptor-driven STAT1 activation. CD34þ blood progenitors for anticancer therapy. Exp Hematol J Immunol 2011;186:6933–44. 1995;23:1463–71. 15. Piskurich JF, Lin KI, Lin Y, Wang Y, Ting JP, Calame K. BLIMP-I 5. Ardeshna KM, Pizzey AR, Thomas NS, Orr S, Linch DC, Devereux S. mediates extinction of major histocompatibility class II transactivator Monocyte-derived dendritic cells do not proliferate and are not expression in plasma cells. Nat Immunol 2000;1:526–32. susceptible to retroviral transduction. Br J Haemato 2000;108: 16. Shaffer AL, Lin KI, Kuo TC, Yu X, Hurt EM, Rosenwald A, et al. 817–24. Blimp-1 orchestrates plasma cell differentiation by extinguishing 6. Harris DP, Haynes L, Sayles PC, Duso DK, Eaton SM, Lepak NM, et al. the mature B cell gene expression program. Immunity 2002;17: Reciprocal regulation of polarized cytokine production by effector B 51–62. and T cells. Nat Immunol 2000;1:475–82. 17. Diehl SA, Schmidlin H, Nagasawa M, van Haren SD, Kwakkenbos MJ, 7. Harris DP, Goodrich S, Mohrs K, Mohrs M, Lund FE. Cutting edge: the Yasuda E, et al. STAT3-mediated up-regulation of BLIMP1 is coordi- development of IL-4-producing B cells (B effector 2 cells) is controlled nated with BCL6 down-regulation to control human plasma cell dif- by IL-4, IL-4 receptor alpha, and Th2 cells. J Immunol 2005;175: ferentiation. J Immunol 2008;180:4805–15. 7103–7. 18. Scheeren FA, Naspetti M, Diehl S, Schotte R, Nagasawa M, Wijnands 8. Lanzavecchia A. Antigen-specific interaction between T and B cells. E, et al. STAT5 regulates the self-renewal capacity and differentiation Nature 1985;314:537–9. of human memory B cells and controls Bcl-6 expression. Nat Immunol 9. Rock KL, Benacerraf B, Abbas AK. Antigen presentation by hapten- 2005;6:303–13. specific B lymphocytes. I. Role of surface immunoglobulin receptors. 19. Lomo J, Blomhoff HK, Beiske K, Stokke T, Smeland EB. TGF-beta 1 J Exp Med 1984;160:1102–13. and cyclic AMP promote apoptosis in resting human B lymphocytes. 10. Ke Y, Kapp JA. Exogenous antigens gain access to the major histo- J Immunol 1995;154:1634–43. compatibility complex class I processing pathway in B cells by recep- 20. Dong Y, Tang L, Letterio JJ, Benveniste EN. The Smad3 protein is tor-mediated uptake. J Exp Med 1996;184:1179–84. involved in TGF-beta inhibition of class II transactivator and class II 11. Barnaba V, Franco A, Alberti A, Benvenuto R, Balsano F. Selective MHC expression. J Immunol 2001;167:311–9. killing of hepatitis B envelope antigen-specific B cells by class 21. Jenkins BJ, Grail D, Nheu T, Najdovska M, Wang B, Waring P, et al. I-restricted, exogenous antigen-specific T lymphocytes. Nature 1990; Hyperactivation of Stat3 in gp130 mutant mice promotes gastric 345:258–60. hyperproliferation and desensitizes TGF-beta signaling. Nat Med 12. Ahmadi T, Flies A, Efebera Y, Sherr DH. CD40 Ligand-activated, 2005;11:845–52. antigen-specific B cells are comparable to mature dendritic cells in 22. Zhang S, Fei T, Zhang L, Zhang R, Chen F, Ning Y, et al. Smad7 presenting protein antigens and major histocompatibility complex antagonizes transforming growth factor beta signaling in the nucleus

OF10 Cancer Res; 72(5) March 1, 2012 Cancer Research

IL-2 and sTbRII Fusion Protein and Antigen-Presenting B Cell

by interfering with functional Smad-DNA complex formation. Mol Cell 25. Wang S. TRAIL: a sword for killing tumors. Curr Med Chem Biol 2007;27:4488–99. 2010;17:3309–17. 23. Itoh S, ten DP. Negative regulation of TGF-beta receptor/Smad signal 26. Blanca IR, Bere EW, Young HA, Ortaldo JR. Human B cell activation by transduction. Curr Opin Cell Biol 2007;19:176–84. autologous NK cells is regulated by CD40-CD40 ligand interaction: 24. Harris DP, Goodrich S, Gerth AJ, Peng SL, Lund FE. Regulation role of memory B cells and CD5þ B cells. J Immunol 2001;167:6132–9. of IFN-gamma production by B effector 1 cells: essential roles 27. Lenz P, Gessner JE, Sautes C, Schmidt RE. Fc gamma-receptor III (CD for T-bet and the IFN-gamma receptor. J Immunol 2005;174: 16) is involved in NK-B cell interaction. Immunobiology 1996;196: 6781–90. 387–98.

www.aacrjournals.org Cancer Res; 72(5) March 1, 2012 OF11

B Effector Cells Activated by a Chimeric Protein Consisting of IL-2 and the Ectodomain of TGF β Receptor II Induce Potent Antitumor Immunity

Claudia Penafuerte, Spencer Ng, Norma Bautista-Lopez, et al.

Cancer Res Published OnlineFirst January 12, 2012.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-11-1659

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2012/01/11/0008-5472.CAN-11-1659.DC1

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

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

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2012/02/06/0008-5472.CAN-11-1659. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.