Specific Recruitment of CC Chemokine Receptor 4–Positive Regulatory T Cells in Hodgkin Lymphoma Fosters Immune Privilege

Research Article

Takashi Ishida,1 Toshihiko Ishii,1 Atsushi Inagaki,1 Hiroki Yano,1 Hirokazu Komatsu,1 Shinsuke Iida,1 Hiroshi Inagaki,2 and Ryuzo Ueda1

Departments of 1Internal Medicine and Molecular Science and 2Clinical Pathology, Nagoya City University Graduate School of Medical Sciences, Kawasumi, Mizuho-chou, Mizuho-ku, Nagoya-shi, Aichi, Japan

Abstract CCL17 and MDC/CCL22, is expressed on regulatory T (Treg) cells Hodgkin lymphoma (HL) is characterized by the presence of (5–7). These observations prompted us to hypothesize that HL a small number of tumor cells in a rich background of tumor cells produce TARC/CCL17 and/or MDC/CCL22 that inflammatory cells, but the contribution of the abundant mediate trafficking of CCR4-positive Treg cells to the tumor, and nontumor cells to HL pathogenesis is poorly understood. We that this specific recruitment of Treg cells represents a mechanism showed that migratory CD4+ cells induced by HL cells were by which the tumor might be capable of suppressing the host hyporesponsive to T-cell receptor stimulation and suppressed antitumor response. In the present study, we examine the above the activation/proliferation of the effector CD4+ T cells in an hypothesis by using human peripheral blood mononuclear cells autologous setting. We further showed that HL cells in the (PBMC) and HL cell lines in vitro, and we also carry out double affected lymph nodes were surrounded by a large number of immunostaining with CCR4 and FOXP3, a hallmark of naturally occurring Treg cells (8–11), in affected lymph node tissues obtained lymphocytes expressing both CC chemokine receptor 4(CCR4) and FOXP3. These findings indicate that the migratory cells from patients with HL. induced by HL cells function as regulatory T (Treg) cells so We have recently developed a novel chimeric monoclonal anti- that these cells create a favorable environment for the tumor body (mAb), KM2760, which binds specifically to CCR4 (12, 13), cells to escape from host immune system. In addition, we and we are currently preparing for phase I and II clinical trials showed that a chimeric anti-CCR4monoclonal antibody of administration of this antibody in patients with CCR4-positive + T-cell lymphomas (14, 15). We have also reported that FOXP3 (mAb) could deplete CCR4 T cells and inhibit the migration + + + + in vitro mRNA was expressed in CD4 CCR4 T cells at a level nearly of CD4 CD25 T cells . Recognition of the importance of + + CCR4+ Treg cells in the pathogenesis of HL will allow rational comparable to that in CD4 CD25 T cells, and that KM2760 treat- design of more effective treatments, such as use of an anti- ment reduced the FOXP3 mRNA expression level in parallel with CCR4mAb, to overcome the suppressive effect of CCR4 + Treg the CCR4 mRNA level in PBMC obtained from healthy volunteers cells on the host immune response to tumor cells. (Cancer Res (12). These observations prompted us to hypothesize that KM2760 2006; 66(11): 5716-22) could be used as a novel strategy for treatment of patients with HL. The antibody could induce an effective antitumor immunity by reducing the concentration of CCR4-positive Treg cells in the HL Introduction tumor cell environment. In the present study, we also test in vitro Hodgkin lymphoma (HL) is characterized by the presence of a the ability of KM2760 to overcome the immunosuppressive small number of tumor cells in a rich background of T and B cells, environment around HL cells by measuring the antibody effect macrophages, and other inflammatory cells (1). The contribution of on migratory cells induced by the supernatants of the HL cell lines. these abundant nontumor cells to the pathogenesis of HL is still poorly understood. Because a small number of HL tumor cells can Materials and Methods survive the host anticancer immune response, it seems likely that HL cells may have developed mechanisms that inhibit development Cell lines. The HL cell lines, HDLM-2 and L-428, were kindly provided of a satisfactory immune response by the host against the tumor by Dr. Kunzo Orita (Director, Fujisaki Cell Center, Hayashibara Biochem- cells. Two significant findings may provide information about the ical Laboratories, Inc., Okayama, Japan) and KM-H2, L-540, and L-1236 possible underlying immunopathogenesis of HL: first, HL cells were purchased from Deutsche Sammlung von Mikroorganismen und Zellkulturen (Braunschweig, Germany). The anaplastic large cell lymphoma express high levels of thymus and activation-regulated chemokine (ALCL) cell line carrying the nucleophosmin-anaplastic lymphoma kinase (TARC/CCL17) and macrophage-derived chemokine (MDC/CCL22; fusion gene, SU-DHL-1, was kindly provided by Dr. Kunzo Orita and refs. 1–4), and an elevated serum level of TARC/CCL17 is an KARPAS-299 was purchased from Deutsche Sammlung von Mikroorganis- unfavorable prognostic factor in patients with HL (3); second, the men und Zellkulturen. CC chemokine receptor 4 (CCR4), the specific receptor for TARC/ Antibodies. Mouse anti-human TARC/CCL17 mAb (clone 54026.11) and mouse anti-human MDC/CCL22 mAb (clone 57226.11) were purchased from R&D Systems, Inc. (Minneapolis, MN). FITC-conjugated mouse anti-human Note: T. Ishida and T. Ishii contributed equally to this work. CCR4 mAb (FITC-KM2160) was previously described (14). Allophycocyanin- Requests for reprints: Takashi Ishida, Department of Internal Medicine and conjugated antihuman CD8 mAb (clone RPA-T8) and its allophycocyanin- Molecular Science, Nagoya City University Graduate School of Medical Sciences, 1 conjugated isotype control mAb were purchased from eBioscience (San Kawasumi, Mizuho-chou, Mizuho-ku, Nagoya, Aichi 467-8601, Japan. Phone: 81-52-853- Diego, CA). Peridinin chlorophyll protein (PerCP)-conjugated antihuman 8216; Fax: 81-52-852-0849; E-mail: [email protected]. I2006 American Association for Cancer Research. CD4 mAb (clone SK3), phycoerythrin (PE)-conjugated antihuman CD25 doi:10.1158/0008-5472.CAN-06-0261 mAb (clone M-A251), PE-conjugated anti-CCR4 mAb (clone 1G1), and their

Cancer Res 2006; 66: (11). June 1, 2006 5716 www.aacrjournals.org

PerCP- and PE-conjugated isotype control mAbs, and FITC-conjugated CD4-negative cells obtained from PBMC were irradiated (25 Gy) and used as isotype control mAb for KM2160 were purchased from BD Biosciences APC. The CD4+ cells (0.75 Â 104) placed in the upper wells or migratory (San Jose, CA). The chimeric anti-CCR4 mAb (KM2760) was previously cells (0.75 Â 104) in the lower wells were cultured in the presence of described (12, 13). Rituximab (16), which is an isotype-matched mAb 3.75 Â 104 APCs and soluble anti-CD3 mAb (clone Hit3a, BD Biosciences) with KM2760, was kindly provided by Chugai Pharmaceutical Co., Ltd. at a final concentration of 20 ng/mL in RPMI 1640 supplemented with (Tokyo, Japan). heat-inactivated 10% FBS for 96 hours in a total volume of 150 AL/well Human PBMC and flow cytometry analysis. Human PBMC were (U-bottomed 96-well plate). [3H]Thymidine diluted in RPMI 1640 supple- isolated from healthy volunteers using Ficoll-Paque (Pharmacia, Uppsala, mented with heat-inactivated 10% FBS (0.5 ACi/50 AL/well) was added Sweden). The volunteers gave informed written consent before the to each well for the last 20 hours of culture and incorporation of the sampling procedure according to the Declaration of Helsinki. The PBMC [3H]thymidine (cpm) was measured as an indicator of cell proliferation were cultured in RPMI 1640 containing 0.5% bovine serum albumin (BSA) using a scintillation counter (Perkin-Elmer, Wellesley, MA). All expressed in a 10-cm plastic culture dish for 1 hour to allow monocytes to adhere values were averages of triplicate experiments F SD. For measurement of and the nonadherent cells were recovered and used as human peripheral IFN-g production, 50 AL of RPMI 1640 supplemented with heat-inactivated blood lymphocytes (PBL). CD4-positive or -negative cells were isolated 10% FBS, instead of [3H]thymidine, were added to each well for the last from human PBMC using the CD4+ T Cell Biotin-Antibody Cocktail 20 hours and the supernatants were collected. IFN-g in the supernatant was Human (Miltenyi Biotec, Bergisch Gladbach, Germany) and Anti-Biotin measured by a Human IFN-g ELISA Kit (BioSource International, Camarillo, MicroBeads (Miltenyi Biotec) according to the instructions of the CA) according to the instructions of the manufacturer. All expressed values manufacturer. were averages of triplicate experiments F SD. In another experiment, PBMC obtained from 15 healthy adult volunteers In addition, the CD4+ cells placed in the upper wells and each of were incubated in RPMI 1640 supplemented with heat-inactivated 10% fetal the migratory cell populations in the lower wells were assessed for their j bovine serum (FBS) at 37 C, 5% CO2 for 12 hours with KM2760 at a ability to suppress the proliferation and IFN-g production activity of the 7 concentration of 10 Ag/mL per well in a 12-well culture plate (1.5 Â 10 CD4+ cells placed in the upper wells as an effector cell population. The CD4+ cells/3 mL). A control culture was incubated without KM2760 in the same cells (0.75 Â 104) placed in the upper wells or each migratory cell manner. At the end of the culture period, the cells were washed twice in PBS population (0.75 Â 104) was cocultured with the CD4+ cells placed in the and assessed by two-color flow cytometry analysis using PerCP-conjugated upper wells at a ratio of 1:1 in the presence of 3.75 Â 104 APCs and soluble anti-CD4 mAb and PE-conjugated anti-CCR4 mAb (1G1) or PerCP- and PE- anti-CD3 mAb, and each of the cocultured cell populations was assessed conjugated isotype control mAbs. Rituximab, which is an isotype-matched for proliferation and IFN-g production activity in the same manner. mAb control for KM2760, was incubated with the PBMC in the same Double immunostaining. The double immunostaining analysis was manner as KM2760. done as previously described (17). Briefly, the formalin-fixed, paraffin- Chemoattractant. HDLM-2, KM-H2, L-1236, L-428, L-540, KARPAS-299, embedded sections of lymph node tissues affected by HL were immunos- and SU-DHL-1 were suspended in RPMI 1640 containing 0.5% BSA at tained using mAbs against human CCR4 (KM2160) and human FOXP3 Â 5 j 3 10 cells/mL. The cells were cultured at 37 C, 5% CO2 for 96 hours, then (236A/E7; Abcam plc, Cambridge, United Kingdom). CCR4 protein at the each supernatant was collected and filtered through 0.22-Am pores to membrane was visualized in brown with 3,3¶ diaminobenzidine tetrahydro- remove the debris and then used as chemoattractant. The RPMI 1640 chloride and FOXP3 protein at the nucleus was visualized in blue with containing 0.5% BSA alone was treated in the same manner and the tetramethylbenzidine (True Blue, KPL, Gaithersburg, MD). The slides were supernatant was used as control chemoattractant. The TARC/CCL17 and counterstained with nuclear fast red and mounted. MDC/CCL22 concentrations in each supernatant were determined by an Statistical analysis. The significance of changes in the proportion of ELISA using the Human TARC/CCL17 or MDC/CCL22 Immunoassay Kit CD4+ T cells in the absence or presence of KM2760 or Rituximab was (R&D Systems), respectively, according to the instructions of the examined using the Wilcoxon signed-rank test. Data were analyzed with the manufacturer. All expressed values were averages of duplicate experiments. aid of StatView software (version 5.0, SAS Institute, Cary, NC). In this study, Recombinant human (rh)-TARC/CCL17 (R&D Systems) and/or rh-MDC/ P < 0.05 was considered significant. CCL22 (R&D Systems) were used as control chemoattractants. Chemotaxis assay. Human PBL (1 Â 106) suspended in 200 ALof RPMI 1640 containing 0.5% BSA were placed in the upper wells of a Chemotaxicell chemotaxis chamber with a 3-Am pore membrane (Kurabo, Results Osaka, Japan). The lower wells contained the chemoattractants (500 AL/ Production of TARC/CCL17 and MDC/CCL22 in HL and well). In some experiments, anti-TARC/CCL17 mAbs and/or anti-MDC/ ALCL cell lines. TARC/CCL17 and MDC/CCL22 concentrations in CCL22 mAbs were added to the lower wells at final concentrations of the supernatants of the HL and anaplastic large-cell lymphoma cell A 50 g/mL each or KM2760 was added to the upper wells at a final line cultures are shown in Table 1. All HL cell lines produced TARC/ A j concentration of 10 g/mL. After 4 hours of incubation at 37 C, 5% CO2, CCL17 and two of five cell lines produced MDC/CCL22. Neither of cells that migrated into the lower wells were recovered and stained with the ALCL cell lines produced detectable levels of TARC/CCL17 or PerCP-anti-CD4, allophycocyanin-anti-CD8, PE-anti-CD25, and FITC-anti- CCR4 (KM2160) and their isotype-matched mAbs. After staining, a MDC/CCL22. predetermined number of standard beads (Flow-Count, Beckman coulter, Chemotactic activity of the supernatants of HL cell lines. Fullerton, CA) were added and the number of cells in each stained cell Chemotactic responses of human PBL induced by the supernatants group was counted using FACSCalibur flow cytometer with the aid of of the HL and ALCL cell lines, rh-TARC/CCL17 and rh-MDC/ CellQuest software (BD Biosciences) according to the instructions of CCL22, were examined. Migratory cells (count/AL) collected in the the manufacturer. All expressed values were averages of triplicate lower well containing RPMI 1640 with 0.5% BSA were used as a experiments F SD. control; their migration was set as ‘‘one.’’ The numbers of migratory ; Cell proliferation and IFN- production assay. Human CD4-positive cells of each subset, collected in the lower wells containing the Â 6 A cells (1 10 ) suspended in 200 L of RPMI 1640 containing 0.5% BSA were chemoattractants, were determined relative to that of the control placed in the upper wells of a Chemotaxicell chemotaxis chamber with a cells. None of the supernatants of the HL cell lines showed any 3-Am pore membrane; the lower wells contained the chemoattractants (500 AL/well). The CD4+ cells placed in the upper wells and each migratory significant chemotactic activity for whole lymphocytes (Fig. 1A) + cell population in the lower wells were assessed for their proliferation and and CD8 cells (Fig. 1B) compared with control medium. In IFN-g production activity in response to T-cell receptor (TCR) stimulation contrast, the supernatants of the HL cell lines, HDLM-2, L-1236, in the presence of autologous antigen-presenting cells (APC). Autologous and L428, showed robust chemotactic activity for CD4+ cells. The www.aacrjournals.org 5717 Cancer Res 2006; 66: (11). June 1, 2006

of migratory cells collected in the lower wells is shown in Fig. 2. Table 1. Production of TARC/CCLI7 and MDC/CCL22 in + HL and ALCL cell lines The migratory CD4 cells induced by the HL cell lines, HDLM-2, L-1236, and L-428, consisted of a large majority of CD25+CCR4+ + TARC MDC cells. The proportions of migratory CD4 cells, as determined by (ng/mL) (ng/mL) CD25 and CCR4 expression, induced by rh-TARC/CCL17 and rh- MDC/CCL22 were almost the same as those of the three HL cell HL cell lines HDLM-2 124.304 2.453 lines, HDLM-2, L-1236, and L-428. The proportion of migratory KM-H2 0.942 <0.063 CD4+ cells induced by control medium was almost the same as that L-1236 6.399 1.116 of the PBL placed in the upper well. The proportions of migratory L-428 107.733 <0.063 CD4+ cells induced by the ALCL cell lines, KARPAS-299 and SU- L-540 0.166 <0.063 DHL-1, were also almost the same as that of PBL in the upper well ALCL cell lines KARPAS-299 <0.031 <0.063 (data not shown). SU-DHL-1 <0.031 <0.063 Treg cell function of migratory CD4+ cells induced by HL cell Control medium <0.031 <0.063 lines. The CD4+ cells placed in the upper wells and the migratory cells collected in the lower wells were assessed for their proliferation activity in response to TCR stimulation in the presence of autologous APC and for their ability to suppress the relative migratory cell counts of the HL cell lines, HDLM-2, KM-H2, proliferation activity of the CD4+ cells in the upper wells (Fig. 3A). L-1236, L-428, and L-540, compared with control medium (1.0 F When the CD4+ cells in the upper wells were stimulated with 0.4) were 9.9 F 0.9, 2.4 F 1.3, 5.8 F 2.3, 4.7 F 1.0, and 1.4 F 0.6, soluble anti-CD3 mAb in the presence of autologous APC, respectively (Fig. 1C). The supernatants of the ALCL cell lines, they responded to this TCR stimulation with a robust proliferation KARPAS-299 and SU-DHL-1, did not show any chemotactic activity (the [3H]thymidine incorporation value was 3,519 F 1,273 cpm). On for CD4+ cells (Fig. 1C). Furthermore, the supernatants of the HL the other hand, when the migratory cells in the lower well induced cell lines, HDLM-2, L-1236, and L428, showed robust chemotactic by the HL cell lines, HDLM-2, L-1236, and L-428, were stimulated in activity for CD4+CD25+CCR4+ cells. The relative migratory cell the same manner, they exhibited hypoproliferation responses to counts of HDLM-2, KM-H2, L-1236, L-428, and L-540 compared TCR stimulation (447 F 49, 875 F 63, and 771 F 277 cpm, with those in control medium (1.0 F 0.4) were 22.4 F 2.9, 3.3 F 1.4, respectively). The migratory cells in the lower well induced by 100 9.9 F 4.5, 10.0 F 2.4, and 2.0 F 0.7, respectively (Fig. 1D). The ng/mL rh-TARC/CCL17 or 100 ng/mL rh-MDC/CCL22 also supernatants of the ALCL cell lines, KARPAS-299 and SU-DHL-1, exhibited hypoproliferation responses to TCR stimulation (532 F did not show any chemotactic activity for CD4+CD25+CCR4+ cells 97 and 697 F 126 cpm, respectively; Fig. 3A, left). When the two cell (Fig. 1D). Rh-TARC/CCL17 and MDC/CCL22 showed robust populations were cocultured at a ratio of 1:1 with soluble anti-CD3 chemotactic activity for CD4+ and CD4+CD25+CCR4+ cells but mAb in the presence of autologous APC, the migratory cells not for whole lymphocytes and CD8+ cells (Fig. 1A-D). induced by the HL cell lines, HDLM-2, L-1236, and L-428, Flow cytometry analysis of PBL placed in the upper well dramatically suppressed the proliferation of the CD4+ cells. When and migratory cells in the lower well. A four-color flow the two populations were cocultured at a ratio of 1:1 in the same cytometry analysis of human PBL placed in the upper wells and manner, migratory cells induced by rh-TARC/CCL17 and rh-MDC/

Figure 1. Chemotactic activity of the supernatants of HL cell lines. The chemotactic responses of human PBL induced by the supernatants of HL and ALCL cell lines, rh-TARC/CCL17 and rh-MDC/CCL22. The number of cells migrated (count/AL) into the lower well containing control medium is set as ‘‘one’’ and the relative numbers of cells of each subset migrated into the lower well containing the chemoattractants are presented. Horizontal dotted line, base unit. The supernatants of the HL cell lines did not show any chemotactic activity for whole lymphocytes (A) and CD8+ cells (B) compared with control medium. The supernatants of the HL cell lines, especially HDLM-2, L-1236, and L428, showed robust chemotactic activity for CD4+ cells (C) and CD4+CD25+CCR4+ cells (D) compared with control medium. Columns, mean of triplicate experiments; bars, SD.

Cancer Res 2006; 66: (11). June 1, 2006 5718 www.aacrjournals.org

Figure 2. Flow cytometry analysis of PBL placed in the upper well and migratory cells in the lower well. Four-color flow cytometry analysis of PBL placed in the upper well and migratory cells in the lower well using PerCP-anti-CD4, allophycocyanin-anti-CD8, PE-anti-CD25, and FITC-anti-CCR4 (KM2160) and their isotype matched mAbs. Top, dot plot analyses determined by CD4 and CD8 expressions. Bottom, dot plot analyses determined by CCR4 and CD25 expressions in the CD4+ cells; the percentages of CD4+ cells in each quadrant determined by CCR4 and CD25 expressions are presented below. The migratory CD4+ cells induced by the HL cell lines, HDLM-2, L-1236, and L-428, consisted of a large majority of CD25+CCR4+ cells.

CCL22 also dramatically suppressed the proliferation of the CD4+ well (1.0 F 0.0, and 1.1 F 0.2, respectively; Fig. 3B, left). When the cells (Fig. 3A, right). two cell populations were cocultured at a ratio of 1:1 with soluble The CD4+ cells in the upper wells and the migratory cells in the anti-CD3 mAb in the presence of autologous APC, the migratory lower wells were also assessed for their IFN-g production activity cells induced by the HL cell lines HDLM-2, L-1236, and L-428 in response to TCR stimulation in the presence of autologous dramatically suppressed the IFN-g production of the CD4+ cells. APC and for their ability to suppress IFN-g production activity of When the two populations were cocultured at a ratio of 1:1 in the the CD4+ cells (Fig. 3B). When the CD4+ cells in the upper wells same manner, the migratory cells induced by rh-TARC/CCL17 or were stimulated with soluble anti-CD3 mAb in the presence of rh-MDC/CCL22 also dramatically suppressed the IFN-g produc- autologous APC, they produced IFN-g at a concentration of 3.2 F tion of the CD4+ cells (Fig. 3B, right). 1.6 pg/mL. On the other hand, when the migratory cells in the Double immunostaining analysis for CCR4and FOXP3 lower well induced by the HL cell lines, HDLM-2, L-1236, and expression in affected lymph nodes obtained from patients L-428, were stimulated in the same manner, they only produced with HL. We did double-immunostaining analysis for CCR4 and small amounts of IFN-g in each well (0.9 F 0.8, 1.2 F 0.2, and FOXP3 using affected lymph node tissues obtained from the HL 0.8 F 0.7 pg/mL, respectively). The migratory cells in the lower patients. A representative picture from a patient with mixed wells induced by 100 ng/mL rh-TARC/CCL17 or 100 ng/mL cellularity HL is shown in Fig. 4. In this case, nearly 80% of the rh-MDC/CCL22 also produced small amounts of IFN-g in each reactive nontumor lymphocytes surrounding the HL tumor cells

Figure 3. Regulatory T-cell function of migratory CD4+ cells induced by HL cell lines. The CD4+ cells placed in the upper wells and the migratory cells in the lower wells were assessed for their proliferation and IFN-g production activity in response to TCR stimulation in the presence of autologous APC and for their ability to suppress the proliferation and IFN-g production activity of the CD4+ cells placed in the upper wells. Columns, mean of triplicate experiments; bars, SD. A, the migratory cells induced by the HL cell lines, HDLM-2, L-1236, and L-428, exhibited hypoproliferation responses to TCR stimulation (left). When the two populations were cocultured at a ratio of 1:1, migratory cells induced by HDLM-2, L-1236, and L-428 dramatically suppressed the proliferation of the CD4+ cells in the upper well (right). B, the migratory cells induced by HDLM-2, L-1236, and L-428 only produced small amounts of IFN-g in response to TCR stimulation (left). When the two cell populations were cocultured at a ratio of 1:1, migratory cells induced by HDLM-2, L-1236, and L-428 dramatically suppressed the IFN-g production of the CD4+ cells in the upper well (right).

www.aacrjournals.org 5719 Cancer Res 2006; 66: (11). June 1, 2006

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2006 American Association for Cancer Research. Cancer Research expressed CCR4 (visualized in brown) at the membrane and nearly 30% of the CCR4 positive reactive lymphocytes simultaneously expressed FOXP3 (visualized in blue) at the nucleus. Any lymphocyte expressing FOXP3 at the nucleus simultaneously expressed CCR4 at the membrane. KM2760-induced effect on the proportion of CD4+ T cells in PBMC. We next examined whether KM2760 treatment would influence the proportions of CD4+ cells and CD4+CCR4+ T cells in fresh PBMC and of CD4+CCR4+ T cells in fresh CD4+ T cells obtained from 15 healthy adult volunteers (Fig. 5A-C). A two-color flow cytometry analysis was done using PerCP-conjugated antihuman CD4 mAb and PE-conjugated antihuman CCR4 mAb (1G1) and their isotype control mAbs. Importantly, the binding of 1G1 to CCR4 was not affected by the presence of KM2760 (data not shown). A 12-hour treatment with KM2760 significantly decreased the proportion of CD4+ T cells in the PBMC (P = 0.0031; Fig. 5A). In contrast, Rituximab significantly increased the proportion of CD4+ T cells in the PBMC (P = 0.0045; Fig. 5A). A 12-hour treatment with KM2760 dramatically decreased the proportion of CD4+CCR4+ T cells in the PBMC in all 15 cases. This KM2760-induced effect was statistically significant (P = 0.0007; Fig. 5B). In contrast, Rituximab significantly increased the proportion of CD4+CCR4+ T cells in the PBMC (P = 0.0031; Fig. 5B). A 12-hour treatment with KM2760 also decreased the proportion of CD4+CCR4+ T cells in the CD4+ T cells in all 15 cases. This KM2760-induced effect was statistically significant (P = 0.0007; Fig. 5C). In contrast, Rituximab did not significantly influence the proportion of CD4+CCR4+ T cells in the CD4+ T cells (Fig. 5C). Figure 5. The attribution of the chemotactic activity of the supernatant of the HL cell line to TARC/CCL17 and MDC/CCL22 and KM2760-induced The attribution of the chemotactic activity of the super- inhibition of the migration of CD4+CD25+ T cells. A to C, percentages were natant of the HL cell line to TARC/CCL17 and MDC/CCL22. determined of CD4+ cells in PBMC, CD4+CCR4+ cells in PBMC, and + + + To determine the source of the chemotactic activity in the super- CD4 CCR4 cells in CD4 cells, obtained from 15 healthy adult volunteers, incubated with KM2760 or Rituximab or without any antibody. Each closed circle natants of the HL cell lines, a chemotaxis assay was done with hu- represents the percentage of CD4+ cells in each PBMC, CD4+CCR4+ cells in man PBL and the neutralizing antibodies antihuman TARC/CCL17 each PBMC, and CD4+CCR4+ cells in each CD4+ cell population. Points, mean; mAb and antihuman MDC/CCL22 mAb (Fig. 5D). The number of bars, SD. *, P < 0.05, significant differences between each group (n.s., not significant). A 12-hour treatment with KM2760 significantly decreased the migratory cells in the lower wells was measured in a two-color flow proportions of CD4+ cells in PBMC, CD4+CCR4+ cells in PBMC, and cytometry analysis using PerCP-conjugated antihuman CD4 mAb CD4+CCR4+ cells in CD4+ cells. D, the source of the chemotactic activity in the supernatants of the HL cell lines was examined by a chemotaxis assay of human and PE-conjugated antihuman CD25 mAb. The results are PBL using the neutralizing antibodies antihuman TARC/CCL17 and antihuman presented as percent migration in which the migratory cell number MDC/CCL22 mAbs. Antihuman TARC/CCL17 alone or antihuman MDC/CCL22 mAb alone did not completely inhibit the migration of the CD4+CD25+ T cells induced by the supernatant of one of the HL cell lines, HDLM-2, which produces both TARC/CCL17 and MDC/CCL22. In contrast, the presence of both antibodies together caused an almost complete inhibition of the CD4+CD25+ T-cell migration induced by the supernatant of HDLM-2. KM2760 partly inhibited the migration of the CD4+CD25+ T cells induced by rh-TARC/CCL17, rh-MDC/ CCL22, and the supernatant of HDLM-2. Columns, mean of triplicate experiments; bars, SD.

was set as 0% in the absence of chemoattractant and as 100% in the presence of each chemoattractant alone. The horizontal dotted line on each vertical bar graph represents the 100% level. Concen- trations of 50 Ag/mL of either antihuman TARC/CCL17 mAb or antihuman MDC/CCL22 mAb almost completely inhibited the migration of CD4+CD25+ T cells induced by 100 ng/mL of rh-TARC/ CCL17 (À5 F 4%) or rh-MDC/CCL22 (4 F 7%), respectively. Concentrations of 50 Ag/mL antihuman TARC/CCL17 or antihuman MDC/CCL22 mAb alone did not fully inhibit the migration of the CD4+CD25+ T cells induced by the supernatant of the HDLM-2 Figure 4. Double immunostaining analysis for CCR4 and FOXP3 expression cell line (66 F 2% and 63 F 18%, respectively). When 50 Ag/mL of in affected lymph node obtained from a patient with HL. The formalin-fixed, each of the two antibodies were added together to the lower well, paraffin-embedded sections of lymph node tissues affected by HL were double + + immunostained using mAbs against human CCR4 and human FOXP3. Brown, they almost completely inhibited the migration of the CD4 CD25 CCR4 protein at the membrane; blue, FOXP3 protein at the nucleus. The slides T cells induced by the supernatant of HDLM-2 (9 F 4%). were counterstained with nuclear fast red and mounted. A representative + + picture from a patient with mixed cellularity HL. Arrows, lymphocytes that KM2760-induced inhibition of the migration of CD4 CD25 expressed both CCR4 and FOXP3; asterisks, HL tumor cells. T cells. The KM2760-induced inhibition of the migration of

Cancer Res 2006; 66: (11). June 1, 2006 5720 www.aacrjournals.org

CD4+CD25+ T cells was examined via the PBL chemotaxis assay consistent with the reports by other investigators that reactive (Fig. 5D). A concentration of 10 Ag/mL KM2760 in the upper well lymphocytes in affected lymph nodes from patients with HL mainly partly inhibited the migration of the CD4+CD25+ T cells induced by consisted of Th2 cells (23, 24). rh-TARC/CCL17 (41 F 11%), rh-MDC/CCL22 (55 F 13%), and the Clinical observations by van den Berg et al. (2) also support the supernatant of HDLM-2 (46 F 7%). Importantly, KM2760 did not present study: tumor cells from HL patients express TARC/CCL17 block the interaction between CCR4 and TARC/CCL17 or between and reactive lymphocytes surrounding the HL tumor cells express CCR4 and MDC/CCL22 (data not shown). CCR4. A situation similar to that of HL has been shown in another hematologic neoplasm, B-cell chronic lymphocytic leukemia (B-CLL). B-CLL cells produced MDC/CCL22 and attracted CCR4+ Discussion T cells (25) and patients with B-CLL showed significantly increased In the present study, ALCL cell lines were used as controls for HL frequencies of CD4+CD25high Treg cells in their peripheral blood cell lines because HL tumor cells and ALCL cells have a similar CD4+ T cells compared with the healthy individuals (26). Fur- morphology and surface immunophenotype (1, 18). With regard to thermore, Curiel et al. (6) have recently reported that ovarian chemokine expression, these two types of lymphoma cells are quite cancer cells and microenvironmental macrophages produce the different, as shown in the present study. The production of TARC/ MDC/CCL22, which mediates trafficking of CCR4+ Treg cells to the CCL17 and MDC/CCL22 in HL cells that we found in the present tumor, and that this specific recruitment of CCR4+ Treg cells study is consistent with the reports by other investigators (1–4). represents a mechanism by which tumors may gain immune Treg cells exist in a CD4+CD25+ cell subpopulation, and FOXP3 privilege. In addition, Nakayama et al. (27) have recently reported can be specifically expressed in CD4+CD25+ Treg cells and is that EBV-infected B cells acquire the ability to produce TARC/ associated with their development and function (8–11). Notably, we CCL17 and MDC/CCL22 and suggested that the production of showed here that HL cell lines attracted CD4+CD25+CCR4+ T TARC/CCL17 and MDC/CCL22, which attract CCR4+ Treg cells, cells, but CD8+ T cells and ALCL cell lines attracted neither may help EBV-infected B cells evade immune surveillance by the CD4+CD25+CCR4+ T cells nor CD8+ T cells. In addition, most host immune system. importantly, this is the first report showing that migratory cells, The recognition of the importance of Treg cells in HL induced by the supernatants of HL tumor cells, are hyporesponsive pathogenesis will allow rational design of more effective treat- to TCR stimulation and suppress the activation/proliferation of the ments. For instance, experimental depletion of Treg cells in mice effector T cells in an autologous setting. These findings show that with tumors improved immune-mediated tumor clearance (28) and the migratory cells induced by HL cells function as Treg cells so enhanced the response to immune-based therapy (29). In addition, that these cells create a favorable environment for the tumor cells a clinical trial of a fully human mAb that blocks CTLA-4 has to escape from host immune system. Our data thus provide a already established the possibility of effecting substantial tumor strong rationale for the ineffective immune clearance of HL cells by destruction, although at the expense of some autoimmunity (30). In the host immune system. We showed here that rh-TARC/CCL17 the present study, KM2760 treatment significantly reduced the and rh-MDC/CCL22 exhibited robust chemotactic activity for proportion of CD4+CCR4+ T cells in PBMC and CD4+CCR4+ T cells CD4+CD25+CCR4+ cells and that the migratory cells induced by in CD4+ T cells obtained from 15 healthy adult volunteers. We have these two specific ligands for CCR4 also were hyporesponsive to previously reported that KM2760 induces a robust antibody- TCR stimulation and suppressed the activation/proliferation of the dependent cellular cytotoxicity (ADCC) against CCR4-positive effector T cells in an autologous setting. Furthermore, antihuman lymphoma cells but does not induce any complement-dependent TARC/CCL17 or MDC/CCL22 mAb alone did not fully inhibit the cytotoxicity or direct antiproliferation activity (12). Based on these migration of the CD4+CD25+ T cells induced by the supernatant of observations, we can therefore conclude that the observed CCR4+ one of the HL cell lines, HDLM-2, which produced both TARC/ T-cell reduction in a PBMC culture incubated with KM2760 was CCL17 and MDC/CCL22, whereas both mAbs together almost induced by ADCC. In this situation, natural killer (NK) cells and completely inhibited the migration of the CD4+CD25+ T cells in monocytes in the PBMC play a role as effector cells in the ADCC. response to the HDML-2 supernatant. These findings indicate that We also previously reported that the majority of CD4+CCR4+ T cells the migration of CD4+CD25+ Treg cells induced by HDLM-2 was exist in a CD25+ subpopulation of PBMC from healthy volunteers attributable to the interaction between CCR4 expressed on Treg (12) and we showed here in the chemotaxis assay that the cells and TARC/CCL17 or MDC/CCL22. The double immunostain- migratory CD4+ cells induced by HL cells consisted of a large ing analysis for CCR4 and FOXP3 in affected lymph node tissues majority of CD25+CCR4+ cells. These observations indicate that the obtained from the HL patients confirmed the in vitro observations KM2760-induced inhibition of the migration of CD4+CD25+ T cells mentioned above. That is to say, HL tumor cells were actually in the present study was due to KM2760-induced lysis of surrounded by a large number of lymphocytes that expressed both CD4+CD25+CCR4+ T cells during the 4-hour incubation. Successful CCR4 and FOXP3. In contrast with the results obtained using treatment of HL patients with KM2760 would require depletion of affected lymph nodes from patients with HL, we have previously the CCR4+ Treg cells not only in the periphery but also in the reported that in non-tumor-affected reactive lymph nodes, only a affected lymph nodes. KM2760 should be able to accomplish this small number of interfollicular T cells were positive for CCR4 by in vivo because NKcells or macrophages, which can mediate immunostaining, and these comprised 3% to 7% of all cells in the ADCC, partly constitute the abundant infiltrating nontumor cells in paracortex area (14). In affected lymph nodes from patients with the affected lymph nodes with HL (1, 31). In addition, the Fc region HL, any lymphocyte expressing FOXP3 at the nucleus simulta- in KM2760 is artificially defucosylated to enhance ADCC by neously expressed CCR4 at the membrane whereas some increasing its binding affinity to the Fcg receptor on effector cells lymphocytes expressing CCR4 did not express FOXP3. Because (12, 13). Collectively, these findings suggest that KM2760 could be the CCR4 is also known to be expressed on Th2 cells (19–22), these used as a novel strategy for treatment of HL patients and function À reactive CCR4+FOXP3 lymphocytes seem to be Th2 cells, which is as a novel inducer of effective anticancer immunity by depleting www.aacrjournals.org 5721 Cancer Res 2006; 66: (11). June 1, 2006

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2006 American Association for Cancer Research. Cancer Research the CCR4+ Treg cells. On the other hand, it was necessary to be used as a novel strategy for treatment of a variety of other diseases, alert to possible KM2760-induced autoimmunity, as reported in the such as HL, B-CLL, ovarian cancer, and EBV-associated disease, to anti-CTLA-4 mAb study (30); thus, KM2760 underwent extensive overcome the suppressive effect of CCR4+ Treg cells on the host evaluation in cynomolgus monkeys and did not cause any notable immune response to tumor or virus-infected cells. Especially in HL, acute or chronic clinical toxicity.3 therapy with an anti-CCR4 mAb combined with tumor-specific In conclusion, we have clarified here how the abundant non- CTLs (33, 34), which are often impaired by Treg cells (35), will offer tumor cells in the environment of a small number of HL tumor a novel treatment approach worthy of pursuit. Furthermore, com- cells contribute to the immunopathogenesis. The CD4+CD25+ bination treatment with this mAb and conventional chemotherapy, CCR4+ Treg cells migrate to the HL cells in a process mediated such as ABVD, seems to be very promising, as observed in the great by the chemokines TARC/CCL17 and/or MDC/CCL22, which are success of Rituximab treatment (16). produced by the HL cells and are capable of suppressing the host antitumor response. We have also shown here that KM2760 Acknowledgments can deplete the CCR4+ T cells and inhibit the migration of CD4+ + Received 1/25/2006; revised 3/24/2006; accepted 3/30/2006. CD25 T cells induced by the HL cells. In this connection, it is Grant support: Grant-in-Aids for General Scientific Research (S. Iida and R. Ueda), also relevant to note that Rituximab, a chimeric anti-CD20 mAb, Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Culture, Science, Sports, and Technology (R. Ueda), and Grant-in-Aids for Cancer combined with the chemotherapy regimen, cyclophosphamide- Research from the Ministry of Health, Labor, and Welfare, Japan (S. Iida and R. Ueda). doxorubicin-vincristine-prednisone, has changed the standard The costs of publication of this article were defrayed in part by the payment of page therapy in elderly patients with diffuse large B-cell lymphoma charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. (16). Now, an anti-CCR4 mAb could be an ideal treatment modality We thank Dr. Toshitada Takahashi, president of Aichi Cancer Center Hospital and for patients with CCR4+ neoplasms (14, 15, 32) and could also be Research Institute, for his critical review of the manuscript; Kyowa Hakko Kogyo Incorporation (Tokyo) for providing us with anti-CCR4 mAb (KM2160) and chimeric anti-CCR4 mAb (KM2760) and for letting us view the data from the KM2760 study in cynomolgus monkeys; Seizo Nagaya and Chiori Fukuyama for their skillful technical 3 Unpublished data from Kyowa Hakko Kogyo Incorporation. assistance; and Takashi Sato for helpful advice and discussions.

References 13. Niwa R, Shoji-Hosaka E, Sakurada M, et al. De- 25. Ghia P, Strola G, Granziero L, et al. Chronic fucosylated chimeric anti-CCR4 IgG1 with enhanced lymphocytic leukemia B cells are endowed with the 1. Stein H, Mann R, Delsol G, et al. Classical Hodgkin antibody-dependent cellular cytotoxicity shows potent capacity to attract CD4+, CD40L+ T cells by producing lymphoma. In: Jaffe ES, Harris NL, Stein H, Vardiman JW, therapeutic activity to T cell leukemia and lymphoma. CCL22. Eur J Immunol 2002;32:1403–13. editors. Pathology and genetics tumors of haemato- Cancer Res 2004;64:2127–33. 26. Beyer M, Kochanek M, Darabi K, et al. Reduced poietic and lymphoid tissues. Lyon (France): IARC Press; 14. Ishida T, Utsunomiya A, Iida S, et al. Clinical frequencies and suppressive function of CD4+CD25hi 2001. p. 244–53. significance of CCR4 expression in adult T-cell leuke- regulatory T cells in patients with chronic lymphocytic 2. van den Berg A, Visser L, Poppema S. High expression mia/lymphoma (ATLL): Its close association with skin leukemia after therapy with fludarabine. Blood 2005;106: of the CC chemokine in Reed-Sternberg cells. A possible involvement and unfavorable outcome. Clin Cancer Res 2018–24. explanation for the characteristic T-cell infiltrate in 2003;9:3625–34. 27. Nakayama T, Hieshima K, Nagakubo D, et al. Hodgkin’s lymphoma. Am J Pathol 1999;154:1685–91. 15. Ishida T, Inagaki H, Utsunomiya A, et al. CXCR3 and Selective induction of Th2-attracting chemokines CCL17 3. Weihrauch MR, Manzke O, Beyer M, et al. Elevated CCR4 expression in T-cell and NK-cell lymphomas with and CCL22 in human B cells by latent membrane protein serum levels of CC thymus and activation-related chemo- special reference to clinicopathologic significance for 1 of Epstein-Barr virus. J Virol 2004;78:1665–74. kine (TARC) in primary Hodgkin’s disease: potential for a peripheral T-cell lymphoma, unspecified. Clin Cancer 28. Shimizu J, Yamazaki S, Sakaguchi S. Induction of prognostic factor. Cancer Res 2005;65:5516–9. Res 2004;10:5494–500. tumor immunity by removing CD4+CD25+ T cells: a 4. Maggio EM, van den Berg A, Visser L, et al. Common 16. Coiffier B, Lepage E, Briere J, et al. CHOP chemo- common basis between tumor immunity and autoim- and differential chemokine expression patterns in re therapy plus rituximab compared with CHOP alone in munity. J Immunol 1999;163:5211–8. cells of NLP, EBV positive and negative classical Hodgkin elderly patients with diffuse large-B-cell lymphoma. 29. Steitz J, Bruck J, Lenz J, Knop J, Tuting T. Depletion of lymphomas. Int J Cancer 2002;99:665–72. N Engl J Med 2002;346:235–42. CD25+CD4+ T cells and treatment with tyrosinase- 5. Iellem A, Mariani M, Lang R, et al. Unique chemotactic 17. Tsuboi K, Iida S, Inagaki H, et al. MUM1/IRF4 related protein 2-transduced dendritic cells enhance response profile and specific expression of chemokine expression as a frequent event in mature lymphoid the interferon a-induced, CD8+ T-cell-dependent im- receptors CCR4 and CCR8 by CD4+CD25+ regulatory malignancies. Leukemia 2000;14:449–56. mune defense of B16 melanoma. Cancer Res 2001;61: T cells. J Exp Med 2001;194:847–53. 18. Delsol G, Ralfkiaer E, Stein H, Wright D, Jaffe ES. 8643–6. 6. Curiel TJ, Coukos G, Zou L, et al. Specific recruitment Anaplastic large cell lymphoma. In: Jaffe ES, Harris NL, 30. Sanderson K, Scotland R, Lee P, et al. Autoimmunity of reglatory T cells in overian carcinoma fosters immune Stein H, Vardiman JW, editors. Pathology and genetics in a phase I trial of a fully human anti-cytotoxic T- privilege and predicts reduced survival. Nat Med 2004; tumors of haematopoietic and lymphoid tissues. Lyon lymphocyte antigen-4 monoclonal antibody with multi- 10:942–9. (France): IARC Press; 2001. p. 230–5. ple melanoma peptides and Montanide ISA 51 for 7. Lee I, Wang L, Wells AD, Dorf ME, Ozkaynak E, 19. Yoshie O, Imai T, Nomiyama H. Chemokines in patients with resected stages III and IV melanoma. J Clin Hancock WW. Recruitment of Foxp3+ T regulatory cells immunity. Adv Immunol 2001;78:57–110. Oncol 2005;23:741–50. mediating allograft tolerance depends on the CCR4 20. Imai T, Nagira M, Takagi S, et al. Selective 31. Ortac R, Aktas S, Diniz G, Erbay A, Vergin C. Prog- chemokine receptor. J Exp Med 2005;201:1037–44. recruitment of CCR4-bearing Th2 cells toward nostic role of natural killer cells in pediatric mixed 8. Hori S, Nomura T, Sakaguchi S. Control of regulatory antigen-presenting cells by the CC chemokine cellularity and nodular sclerosing Hodgkin’s disease. T cell development by the transcription factor FoxP3. thymus and activation-regulated chemokine and Anal Quant Cytol Histol 2002;24:249–53. Science 2003;299:1057–61. machrophage-derived chemokine. Int Immunol 1999; 32. Ishida T, Inagaki H, Kusumoto S, et al. CC chemokine 9. Walker MR, Kasprowicz DJ, Gersuk VH, et al. 11:81–8. receptor 4-positive diffuse large B-cell lymphoma Induction of FoxP3 and acquisition of T regulatory 21. Bonecchi R, Bianchi G, Bordignon PP, et al. involving the skin: a case report. Int J Hematol 2005; À activity by stimulated human CD4+CD25 Tcells. J Clin Differential expression of chemokine receptors and 82:148–51. Invest 2003;112:1437–43. chemotactic responsiveness of type 1 helper cells 33. Gottschalk S, Edwards OL, Sili U, et al. Generating 10. Yagi H, Nomura T, Nakamura K, et al. Crucial role of (Th1s) and Th2s. J Exp Med 1998;187:129–34. CTLs against the subdominant Epstein-Barr virus LMP1 FOXP3 in the development and function of human 22. Sallusto F, Lenig D, Mackay CR, Lanzavecchia A. antigen for the adoptive immunotherapy of EBV- CD25+CD4+ regulatory T cells. Int Immunol 2004;11: Flexible programs of chemokine receptor expression on associated malignancies. Blood 2003;101:1905–12. 1643–56. human polarized T helper 1 and 2 lymphocytes. J Exp 34. Bollard CM, Aguilar L, Straathof KC, et al. Cytotoxic 11. O’Garra A, Vieira P. Regulatory T cells and mecha- Med 1998;187:875–83. T lymphocyte therapy for Epstein-Barr virus+ Hodgkin’s nism of immune system control. Nat Med 2004;10:801–5. 23. Skinnider BF, Mak TW. The role of cytokines in disease. J Exp Med 2004;200:1623–33. 12. Ishida T, Iida S, Akatsuka Y, et al. The CC chemokine classical Hodgkin lymphoma. Blood 2002;99:4283–97. 35. Dannull J, Su Z, Rizzieri D, et al. Enhancement of receptor 4 as a novel-specific molecular target for 24. Poppema S, van den Berg A. Interaction between vaccine-mediated antitumor immunity in cancer immunotherapy in adult T-cell leukemia/lymphoma. host T cells and Reed-Sternberg cells in Hodgkin patients after depletion of regulatory T cells. J Clin Clin Cancer Res 2004;10:7529–39. lymphomas. Semin Cancer Biol 2000;10:345–50. Invest 2005;115:3623–33.

Cancer Res 2006; 66: (11). June 1, 2006 5722 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2006 American Association for Cancer Research. Specific Recruitment of CC Chemokine Receptor 4−Positive Regulatory T Cells in Hodgkin Lymphoma Fosters Immune Privilege

Takashi Ishida, Toshihiko Ishii, Atsushi Inagaki, et al.

Cancer Res 2006;66:5716-5722.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/66/11/5716

Cited articles This article cites 32 articles, 18 of which you can access for free at: http://cancerres.aacrjournals.org/content/66/11/5716.full#ref-list-1

Citing articles This article has been cited by 48 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/66/11/5716.full#related-urls

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/66/11/5716. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.