Published OnlineFirst July 5, 2013; DOI: 10.1158/2326-6066.CIR-13-0049 Cancer Priority Brief Immunology Research Human Regulatory T Cells Kill Tumor Cells through Granzyme-Dependent Cytotoxicity upon Retargeting with a Bispecific Antibody Bryan D. Choi1,2, Patrick C. Gedeon1,3, James E. Herndon II4, Gary E. Archer1, Elizabeth A. Reap1, Luis Sanchez-Perez1, Duane A. Mitchell1,2,5, Darell D. Bigner2,5, and John H. Sampson1,2,5 Abstract A major mechanism by which human regulatory T cells (Treg) have been shown to suppress and kill autologous immune cells is through the granzyme-perforin pathway. However, it is unknown whether Tregs also possess the capacity to kill tumor cells using similar mechanisms. Bispecific antibodies (bscAb) have emerged as a promising class of therapeutics that activate T cells against tumor antigens without the need for classical MHC-restricted T- cell receptor (TCR) recognition. Here, we show that a bscAb targeting the tumor-specific mutation of the EGF þ þ þ receptor, EGFRvIII, redirects human CD4 CD25 FoxP3 Tregs to kill glioblastoma cells. This activity was significantly abrogated by inhibitors of the granzyme-perforin pathway. Notably, analyses of human primary þ glioblastoma also displayed diffused infiltration of granzyme-expressing FoxP3 T cells. Together, these data suggest that despite their known suppressive functions, tumor-infiltrating Tregs possess potent cytotoxic mechanisms that can be co-opted for efficient tumor cell lysis. Cancer Immunol Res; 1(3); 163–7. Ó2013 AACR. Introduction molecules that possess dual specificity for tumor-associated Despite maximal surgical resection, dose-intensive radio- surface antigens and the CD3 complex on T cells, which allows fi therapy and multimodal chemotherapy, glioblastoma remains them to divalently bind and afford potent, speci c target cell uniformly lethal with a median survival of less than 15 months lysis (8). Because CD3 is universally expressed among T cells, (1). Novel therapies are desperately needed to improve the BiTEs have the theoretical capacity to redirect and activate prognosis of this disease and immunologic targeting of tumor- even Tregs that are elevated and present in tumors of patients specific mutations has emerged as a promising approach (2). with cancer. fi A major barrier that has impeded translation of efficacious Among the few known tumor-speci c antigens, perhaps the immunotherapy is the inability to overcome profound immu- most widely characterized is the truncated mutant EGF recep- nosuppression associated with malignant disease (3). Regula- tor variant type III (EGFRvIII). EGFRvIII is a constitutively activated tyrosine kinase that is frequently expressed on the tory T cells (Tregs) in particular are thought to play a central role in tumor escape from immune-mediated rejection. One mech- surface of glioblastoma and other common neoplasms but is anism by which T are known to suppress and even kill completely absent from healthy tissues (9). regs fi autologous immune cells is through the granzyme-perforin In this study, we show that an EGFRvIII-speci c BiTE, fi pathway (4, 5). Despite this well-characterized cytotoxic capac- bscEGFRvIIIxCD3 (7), successfully redirects highly puri ed ity, whether T can co-opt cytotoxic mechanisms to kill Tregs and activates them in the presence of tumors expressing regs fi tumor cells has yet to be evaluated. EGFRvIII. Despite their known suppressive properties, puri ed þ þ dim/À fi T-cell–activating bispecific antibodies (bscAb), particularly CD4 CD25 CD127 T cells ef ciently lysed EGFRvIII- in vitro those of the bispecific T-cell engager (BiTE) subclass, represent expressing glioblastoma upon redirection and activa- a new therapeutic strategy that has the potential to treat even tion with bscEGFRvIIIxCD3. This activity was found to be bulky, invasive disease (6, 7). BiTEs are tandem single-chain dependent on the granzyme-perforin pathway. Immunohisto- chemical (IHC) analysis from human primary GBMs also displayed diffuse infiltration of activated, granzyme-producing þ FoxP3 cells, showing that T with potent effector functions Authors' Affiliations: 1Duke Brain Tumor Immunotherapy Program, Divi- regs sion of Neurosurgery, Department of Surgery, Departments of 2Pathology, may already be present in tumors even under natural 3Biomedical Engineering, and 4Biostatistics and Bioinformatics, and 5The conditions. Preston Robert Tisch Brain Tumor Center at Duke, Duke University Medical Center, Durham, North Carolina Previous efforts to enhance antitumor immunity via Treg depletion have been limited, in part due to an inability to Corresponding Author: John H. Sampson; Duke Brain Tumor Immuno- fi fi therapy Program, Division of Neurosurgery, Department of Surgery; Duke ef ciently eliminate suppressive cells that in ltrate tumor University Medical Center; Box 3050; Durham, NC 27710. Phone: 919-684- tissue (10). Alternatively, our data suggest that BiTEs can 9043; Fax: 919-419-1741; E-mail: [email protected] activate even suppressive Tregs to lyse tumor cells by redirect- doi: 10.1158/2326-6066.CIR-13-0049 ing their natural granzyme-mediated cytotoxic potential. Ó2013 American Association for Cancer Research. These findings not only highlight a new mechanism by which www.aacrjournals.org 163 Downloaded from cancerimmunolres.aacrjournals.org on September 26, 2021. © 2013 American Association for Cancer Research. Published OnlineFirst July 5, 2013; DOI: 10.1158/2326-6066.CIR-13-0049 Choi et al. BiTEs may circumvent certain aspects of Treg-mediated added to each well of a 96-well round-bottom plate for an suppression, but also have broader implications with regard additional 24 hours and cells were collected by a cell harvester. to the natural functional role of activated, tumor-infiltrating Counts were conducted using a Wallac 1450 Microbeta Trilux Tregs that express granzyme and perforin in the tumor Liquid Scintillation/Luminescence Counter (Perkin-Elmer). microenvironment. Similarly, cytotoxicity assays were conducted by a standard chromium release assay as previously described by us (7). In Materials and Methods brief, target cells were labeled with 51Cr and incubated with Tumor cell lines and reagents EGFRvIII BiTE and effector cells [E:T ratio, 20:1; incubation The human glioma cell line U87MG and its subline time, 18 hours at 37C; (bscEGFRvIIIxCD3), 10 mg/mL] in a U87MG.DEGFR, which expresses EGFRvIII, are described else- total volume of 200 mL. Following incubation, supernatants where and were kind gifts from Dr. Webster Cavenee at Uni- were removed and measured by gamma counter. Where noted, versity of California, San Diego (11). EGFRvIII expression was for inhibition of the granzyme-perforin axis, concanamycin A verified by flow cytometry using the L8A4 antibody (12). These (CMA, 100 nmol/L; Sigma), Granzyme B Inhibitor I (Z-AAD- cell lines were regularly confirmed to be Mycoplasma-free by CMK, 50 mmol/L; Calbiochem), or ethylene glycol tetracetic nucleic acid hybridization assay (GEN-PROBE MTC-NI). Char- acid (EGTA, 4 mmol/L; Calbiochem) was added to each well. acterization, production, and purification of bscEGFRvIIIxCD3 Inhibitor concentrations were chosen based on previously as well as control antibody constructs were conducted as published studies to establish dependence of cell-mediated previously described by us (7). Antibodies to CD4 (RPA-T4), cytotoxicity on perforin-granzyme activity (4, 13, 14). Before CD25 (M-A251), CD69 (L78), CD152 (BNI3), Granzyme A (CB9), use, each inhibitor was found to have insignificant effects on Granzyme B (GB11), and Perforin (dG9) were purchased from the viability of Tregs following 18 hours incubation at 37 Cas BD Biosciences. Anti-FoxP3 antibody (PCH101) and the FoxP3 assessed by LIVE/DEAD Fixable Violet Dead Cell Stain Kit Staining Buffer Set were purchased from eBioscience and (Invitrogen). Blockade of FasL- and TRAIL-mediated apoptosis intracellular staining was conducted according to manufac- was carried out with antibody clones NOK-1 (25 mg/mL; turer's instructions. Antibodies against human FoxP3 (259D, BD Biosciences) and RIK-2 (25 mg/mL; BD Biosciences), BioLegend) and granzyme B (Cat No. ab4059) were used for respectively. IHC staining. Immunohistochemistry Treg isolation and preparation Tissues for IHC analysis were derived from human brain All human samples were obtained at Duke University Med- tumor biopsy material. Paraffin-embedded tissues were fixed, ical Center (Durham, NC) from individuals who had given cut into 5 mm sections, mounted on glass slides and subjected written, informed consent. Human PBMCs were prepared by to primary and secondary staining using the MACH 2 Double density gradient centrifugation from buffy coats of healthy Stain 1 polymer Detection Kit (Biocare Medical) according to donor leukapheresis. Highly purified regulatory T cells were the manufacturer's instructions. Diaminobenzidine (DAB) and þ isolated from PBMCs by magnetic separation using the CD4 AminoEthyl Carbazole (AEC) were used to detect granzyme B þ À CD25 CD127dim/ Regulatory T Cell Isolation Kit II and and FoxP3, respectively. AutoMACS Separator (Miltenyi Biotec) according to the man- ufacturer's instructions. Following isolation, purity was con- Statistical analysis firmed by flow cytometric analysis to be more than 95% Groups were compared using a two-sample two-tailed t test þ þ (CD25 FoxP3 ). Where noted, purified Tregs were prepared and statistical significance was determined at a value of P < 0.05. with the T Expansion Kit (Miltenyi Biotec) at a MACSiBead reg Results