Adoptive cancer immunotherapy with human Vγ2vδ2 T cells

Transcription

1 University of Iowa Iowa Research Online Theses and Dissertations Fall 2016 Adoptive cancer immunotherapy with human Vγ2vδ2 T cells Mohanad Hameed Nada University of Iowa Copyright 2016 Mohanad Hameed Nada This dissertation is available at Iowa Research Online: Recommended Citation Nada, Mohanad Hameed. "Adoptive cancer immunotherapy with human Vγ2vδ2 T cells." PhD (Doctor of Philosophy) thesis, University of Iowa, Follow this and additional works at: Part of the Immunology of Infectious Disease Commons

2 ADOPTIVE CANCER IMMUNOTHERAPY WITH HUMAN Vγ2Vδ2 T CELLS by Mohanad Hameed Nada A thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree in Immunology in the Graduate College of The University of Iowa December 2016 Thesis Supervisor: Professor Craig T. Morita

3 Copyright by Mohanad Hameed Nada 2016 All Rights Reserved

4 Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL PH.D. THESIS This is to certify that the Ph.D. thesis of Mohanad Hameed Nada has been approved by the Examining Committee for the thesis requirement for the Doctor of Philosophy degree in Immunology at the December 2016 graduation. Thesis Committee: Craig T. Morita, Thesis Supervisor Vladimir P. Badovinac George J. Weiner Zuhair K. Ballas David M. Lubaroff

5 Dedicated to my first teachers, my parents, my beloved wife, Auter, my three boys Ayham, Anas, and Kinan, who are the joy of my life, my brothers and sisters, for their continued and unconditional love and support. ii

6 In the name of Allah, Most Gracious, Most Merciful iii

7 ACKNOWLEDGEMENTS I would like to express my gratitude to the people who contributed to this accomplishment. I want to thank first my mentor Dr. Craig Morita for his support and help throughout the past five years. Dr. Morita has been more than a mentor. He invested his time and effort in rising my scientific knowledge and research skills to the best possible level. It has been a truly rewarding experience working with Dr. Morita. I also want to thank my thesis committee members, Drs. Vladimir P. Badovinac, George J. Weiner, Zuhair K. Ballas, and David M. Lubaroff, for their valuable suggestions and feedbacks during our regular meetings. My thanks go to current and former members of Morita s lab, especially Dr. Hong Wang for technical advice and assistance. Also, I would like to thank the administration of the Immunology Graduate Program for keeping me on track with my requirements and meetings. For that I thank Dr. Steven Varga, Dr. Gail Bishop, Paulette Villhauer, and Joshua Lobb. I have received helpful assistance from the Flow Cytometry Facility, and I wish to thank Justin and Tom for their help in optimizing my flow cytometry skills. I would like to express my special thank to the Higher Committee of Education Development in Iraq for giving me the opportunity and financially supporting my doctoral study at the University of Iowa. Finally, I would like to sincerely thank my dear parents, Lamyaa and Hameed, for their continued support and unconditional love. Without you I would not have been where I am today. I ask Allah to grant you forgiveness and the highest level of paradise. I extend my thank to my brothers and sisters, Muthana, Muayad, Narjis, Sarab, Majid, Maeseer, Manhal, Reem and Maheoub, for their help and encouragements. My deepest thank to my beloved wife, Auter, for being on my side wherever life takes me. Thanks for your patience and support. Without you, none of this would have been possible. Thanks to my sons Ayham, Anas, and Kinan for giving me the motivation to work hard for a better future. iv

8 ABSTRACT Human γδ T cells expressing Vγ2Vδ2 TCRs monitor foreign- and self-prenyl pyrophosphate metabolites in isoprenoid biosynthesis to mediate immunity to microbes and tumors. Vγ2Vδ2 cells have been used for adoptive cancer immunotherapy with some partial and complete remissions. Most trials have used continuous zoledronate exposure to expand Vγ2Vδ2 cells. Zoledronate inhibits farnesyl pyrophosphate synthase causing isopentenyl pyrophosphate to accumulate that then stimulates Vγ2Vδ2 cells. Because zoledronate exposure is toxic, we hypothesized that a short period of exposure would reduce T cell toxicity but still be sufficient for monocytes uptake. Supporting this hypothesis, pulse zoledronate exposure with IL-2 resulted in more uniform expansion of Vγ2Vδ2 cells with higher purity and cell numbers as compared with continuous exposure. These Vγ2Vδ2 cells also had higher levels of CD107a and perforin and slightly increased tumor cytotoxicity. Importantly, adoptive immunotherapy with Vγ2Vδ2 cells derived by pulse stimulation controlled human PC-3 prostate cancer cells in immunodeficient NSG mice significantly better than those derived by continuous stimulation. Pulse zoledronate stimulation of Vγ2Vδ2 cells with IL-15 also resulted in higher purity and cell numbers. Like with CD8 αβ T cells, IL-15 preserved early memory Vγ2Vδ2 T cell subsets better than IL-2. However, despite this fact, adoptive immunotherapy with Vγ2Vδ2 cells derived with IL-15 showed similar inhibition of PC-3 tumor growth as those derived with IL-2. Thus, pulse zoledronate stimulation maximizes the purity, quantity, and quality of expanded Vγ2Vδ2 cells. This simple modification to existing protocols would likely enhance the effectiveness of adoptively transferred Vγ2Vδ2 T cells. v

9 PUBLIC ABSTRACT Cancer is the second leading cause of death worldwide. In 2015, cancer killed more than 600,000 people in US alone. Besides its impact on health, cancer has a huge economical impact with billions of dollars spent every year to provide health care for cancer patients. Today, there are 14 million cancer patients who live with the disease and hope that one day a cure for such a fatal disease can be a possibility. One of the most recent and promising advancement in the field of cancer treatment is using immune system to fight and kill cancer cells. Immune system consists of immune cells that are responsible of killing cancer cells once get activated. Among those cells, a group of cells called gamma delta (γδ) T cells that show a strong killing activity against wide range of cancer. Based on these results, γδ T cells have been used clinically to treat patients with various types of cancer. γδ T cell treatment has been shown to be safe with few major toxicities and has resulted in rare complete remissions, some partial remissions, and stabilization of disease. The clinical trials have established the safety of the treatment but also showed that the response rate has to be increased before its widespread use. In our studies, we have focused on enhancing the activity of γδ T cell cancer treatment by modifying current treatment protocols. By making a simple modification to existing protocols used to grow γδ T cells, we were able to significantly increase the number of γδ T cells generated and their killing of human cancer cells in laboratory studies. Furthermore, the γδ T cells showed improved ability to control human prostate cancer tumors in a pre-clinical mouse model as compared with γδ T cells grown conventionally. We expect that a similar improvement in efficacy would be observed in the treatment of cancer patients. We also discuss other potential ways to increase the effectiveness of γδ T cell therapy. Improving the effectiveness of γδ T cells could a major advance in the treatment of a variety of cancer types. vi

10 TABLE OF CONTENTS LIST OF TABLES... x LIST OF FIGURES... xi LIST OF ABBREVIATIONS... xiii CHAPTER I. Introduction... 1 Biology of human γδ T cells expressing Vγ2Vδ2 TCRs... 1 Cytokines and Vγ2Vδ2 T cells expansion... 2 Memory of Vγ2Vδ2 T cells... 3 BTN3A1 sensor function and stimulation of Vγ2Vδ2 T cells by prenyl pyrophosphates... 5 Functions of Vγ2Vδ2 T cells... 6 Tumor immunity of human Vγ2Vδ2 T cells demonstrated in preclinical mouse models of human cancer... 8 Anti-tumor response of Vγ2Vδ2 T cells in clinical trials In vivo immunization Adoptive cancer immunotherapy with Vγ2Vδ2 T cells generated ex vivo Rational and objectives CHAPTER II. Materials and Methods Reagents Ex vivo expansion of Vγ2Vδ2 T cells Purification of Vγ2Vδ2 T cells CD107a and intracellular cytotoxic protein and cytokine expression after stimulation of expanded Vγ2Vδ2 T cells Fluorometric assessment of T lymphocyte antigen specific lysis Human tumor xenograft and adoptive transfer of purified Vγ2Vδ2 T cells into immunodeficient mice Statistical analyses CHAPTER III A- Enhancing Adoptive Cancer Immunotherapy with Vγ2Vδ2 T Cells through pulse Zoledronate Stimulation and IL Rational vii

11 Results Pulse zoledronate stimulation improves the purity and number of expanded Vγ2Vδ2 T cells Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation exhibit increased degranulation and perforin expression compared with those expanded by continuous stimulation when stimulated by pamidronate-treated PC-3 cancer cells or Daudi Burkitt's lymphoma cells Cytotoxicity of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation against pamidronate-treated PC-3 prostate cancer cells Adoptive transfer of Vγ2Vδ2 T cells expanded using pulse zoledronate stimulation controlled PC-3 tumor growth in NSG mice Pulse zoledronate stimulation improves the purity and yield of Vγ2Vδ2 T cells cultured with IL IL-15 preserves higher proportions of early/central memory subsets of Vγ2Vδ2 T cells Vγ2Vδ2 T cells expanded using pulse zoledronate stimulation with IL-15 express CD107a, cytotoxic proteins, and cytokines in similar proportions to those expanded with IL Adoptive transfer of Vγ2Vδ2 T cells expanded using pulse stimulation with IL-15 controlled PC-3 tumor growth in NSG mice similarly to Vγ2Vδ2 T cells expanded with IL Vγ2Vδ2 T cells stimulated by pulse zoledronate expand to similar levels using OpTmizer TM media manufactured under cgmp and with large scale cultures B- The Role of IL-18 on the Expansion, Memory Subset Distribution, and Function of Vγ2Vδ2 T Cells Rational Results IL-18 with IL-2 or IL-15 supports the expansion of Vγ2Vδ2 T cells with zoledronate pulse stimulation IL-18 helps IL-2 to preserve more central/early memory subset of Vγ2Vδ2 T cells and helps IL-15 to maintains more late memory subset IL-18, along with IL-2 or 15, maintains high levels of cytokines and cytotoxic molecules under zoledronate pulse stimulation Discussion viii

12 CHAPTER IV. Future directions and conclusion Future directions Conclusions REFERENCES ix

13 LIST OF TABLES Table 1. Pre-clinical studies using Vγ2Vδ2 T cells to treat cancer in humanized mouse models Table 2. Clinical studies where Vγ2Vδ2 T cells have been activated in vivo to generate anti-tumor activity Table 3. Clinical studies using adoptively transferred Vγ2Vδ2 T cells generated ex vivo for cancer treatment x

14 LIST OF FIGURES Figure 1. Mevalonate pathway and key downstream branches in isoprenoid biosynthesis in humans Figure 2. Model of anti-tumor activity of Vγ2Vδ2 T cells Figure 3. Therapeutic strategies for Vγ2Vδ2 T cell immunotherapy Figure 4. Purity of Vγ2Vδ2 T cells after positive magnetic bead separation Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Expanding Vγ2Vδ2 T cells by pulse zoledronate stimulation with IL-2 increased their purity and number compared with continuous stimulation Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-2 exhibit slightly increased degranulation and expression of granzyme B and perforin compared with those expanded by continuous stimulation when exposed to pamidronate-treated PC-3 cancer cells Expansion of Vγ2Vδ2 T cells by pulse zoledronate stimulation increased degranulation as measured by the expression of CD107a in response to the stimulatory Burkitt's lymphoma cell line, Daudi Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation exhibit similar cytotoxicity against pamidronate-treated PC-3 cancer cells compared with Vγ2Vδ2 T cells expanded by continuous stimulation Adoptive transfer of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation in combination with pamidronate controlled PC-3 prostate tumor growth in NSG mice Figure 10. Adoptive transfer of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation in combination with pamidronate controlled PC-3 prostate tumor growth in NSG mice; tumor diameter data for Fig Figure 11. Expanding Vγ2Vδ2 T cells by pulse zoledronate stimulation with IL-15 increased their purity and numbers compared with continuous stimulation. 66 xi

15 Figure 12. IL-15 sustained higher proportions of early/central memory subsets among expanded Vγ2Vδ2 T cells compared to IL Figure 13. Functional capabilities of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-15 are similar to those expanded with IL Figure 14. Adoptive transfer of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-15 in combination with pamidronate controlled PC-3 prostate tumor growth in NSG mice similarly to Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL Figure 15. Adoptive transfer of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-15 in combination with pamidronate controlled PC-3 prostate tumor growth in NSG mice similarly to Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-2 (tumor diameter) Figure 16. Vγ2Vδ2 T cells stimulated by pulse zoledronate expand to similar levels using commercial cgmp grade OpTmizer TM media and with larger-scale cultures Figure 17. IL-18, in combination of IL-2 or IL-15, maintains high purity and increases the yield of Vγ2Vδ2 T cells stimulated by pulse zoledronate stimulation Figure 18. IL-18 alters the proportions of the memory subsets in expanded Vγ2Vδ2 T cells when added to IL-2 or IL-15 compared to IL-2 or IL-15 alone Figure 19. IL-18 maintains similar functional capabilities of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation in combination with IL-2 or IL-15, to those expanded with IL-2 or IL-15 alone Figure 20. Model for optimizing anti-tumor activity of Vγ2Vδ2 T cells through zoledronate pulse stimulation xii

16 LIST OF ABBREVIATIONS γδ T cells gamma delta T lymphocyte Vγ2Vδ2 Vgamma 2V delta 2 T cells Ab antibody ADCC antibody dependent cell-mediated cytotoxicity Akt protein kinase B AML acute myeloid leukemia APC antigen Presenting Cells APPPI triphosphoric acid 1-adenosin-5 -yl ester 3-(3-methylbut-3-enyl) ester ATCC american tissue culture association ATP adenosine triphosphate BCL-2 B cell lymphoma 2 gene BTLA B and T lymphocyte attenutter BTN3A butyrophilin 3A C-Media complete media CAR chimeric antigen receptors CCR c-c chemokine receptor CD cluster differentiation CFSE carboxyflurescein diacetate succinimidyl ester COX cyclooxgyenase CTLA-4 cytotoxic T lymphocyte associated protien-4 CXCR3 c-x-c chemokine receptor DC dendritic cells DNAM-1 DNAX accessory molecule -1 FasL fas ligand FATAL fluromatric assessment of T lymphocyte antigen specific lysis FCS fetal calf serum xiii

17 FDPS farnesyl Pyrophosphate Synthase FITC fluorescein isothiocyanate GD2 ganglioside GGPP geranylgeranyl pyrophosphate GM-CSF granulocyte-macrophage colony-stimulating factor GMP good manufacture practice HER-2 human epidermal growth factor receptor 2 HLA human leukocyte antigen HMBPP (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate HMG-CoA 3-hydroxy-3-methyl-glutaryl-coenzyme A IBL4 B cell lymphoma cell line ICAM-1 intracellular adhesion molecule-1 IDO indoleamine 2,3-dioxygenase IFN-α Interferon alpha IFN-γ Interferon gamma IL Interleukin IPP Isopentyel Pyrpphosphate LAG-3 lymphocyte Activation Gene 3 LAMP-1 lysosomal associated memberane protien -1 LCMV lymphocytic choriomeningitis virus LU lytic unit MEP methylerythritol 4-phosphate MHC major histocompetality complex molecule class MM multiple myeloma mtor mammalian target of rapamycin NB-1691 neuroblastoma cell line NHL non-hodgkin lymphoma NK natural killer cells NKG2D natural killer group 2D NSCLC non-small cell lung carcinoma xiv

18 NSG NOD.Cg-Prkdc scid Il2rg tm1wjl /SzJ PBMC peripheral blood mononuclear cells PC-3 human prostate cancer cell line PD-1 programed cell death-1 PEG2 progstoglandin PMA phorbol 12-myristate 13- acetate PSA prostate specific antigen SCID severe combined immunodeficiency, SH-SY-5Y human neuroblastoma cell line SLAM signaling lymphocyte activation molecule SW1353 human chondrosarcoma cell line T cm TCR T emra T fh TGF-β T central memory T cell receptor T effector memory T follicular helper cells transforming growth factor beta Th1 T helper 1 Th17 T helper 17 Th2 T helper 2 T cells TIGIT T-cell immunoreceptor with Ig and ITIM domains TIL Tumor infiltrated lymphocytes TIM-3 T cell immunoglubulin and mucine domain-3 TLR toll like receptor TNF-α TRAIL TRAILR T reg VEGF VISTA Tumor Necrosis Factor alpha TNF-related apoptosis-inducing ligand TRAIL receptor T regulatory cells vascular epithelial growth factor v-domain immunoglobulin (Ig)-containing suppressor of T-cell activation xv

19 Vδ1 XG-7 Zol V delta 1 T cells multiple myeloma cell line zoledronate xvi

20 CHAPTER I. Introduction Biology of human γδ T cells expressing Vγ2Vδ2 TCRs Human γδ T cells are small population of CD3 + T lymphocytes and consist 1-5% of T lymphocyte in healthy people. Due to the limited number of Vδ and Vγ genes, γδ T cells have a simpler repertoire than αβ T cells (1-3). In many species, γδ T cells are enriched in specific tissues such as the skin, intestine, liver, lung and reproductive organs. In humans, γδ can be subsetted based on their expression of specific Vγ and Vδ chain pairs. Vγ2Vδ2 T cells are the dominant γδ T cell subset in human blood; making up ~75% of human γδ T cells in blood and central lymphoid tissues (4, 5) whereas Vδ1 T cells are primarily tissue infiltrating cells (6, 7). Unlike αβ T cells, γδ T cells do not recognize peptide antigen presented by MHC class I or class II molecules. Rather, nonpeptide molecules such as IPP and HMBPP can stimulate Vγ2Vδ2 T cells in a TCRdependent manner (8-12). Vγ2Vδ2 T cells can also be activated in a TCR-independent manner through their NK receptors. The majority of Vγ2Vδ2 T cells express the NK activating receptor, NKG2D and can be partially stimulated upon receiving activating signals from target cells (13, 14). Vγ2Vδ2 T cells monitor self and foreign isoprenoid metabolism through their TCRs. There are two well-characterized isoprenoid metabolites that can trigger Vγ2Vδ2 T cell activation; isopentenyl pyrophosphate (IPP), a metabolite of the mevalonate pathway used by most eukaryotic cells (Fig. 1), and HMBPP, a metabolite of MEP pathway used by prokaryotic organism such as bacteria, the chloroplasts in plants, and Apicomplexan protozoan parasites. In normal human cells, the levels of IPP are usually not high enough to trigger Vγ2Vδ2 TCRs and induce expansion. However, aminobisphosphonate drugs (zoledronate, pamidronate, alendronate, etc.) can inhibit farnesyl pyrophosphate synthase (FDPS) causing the endogenous accumulation of IPP thereby activating Vγ2Vδ2 T cells (15-19) (Fig. 1). After activation, Vγ2Vδ2 T cells can 1

21 kill stressed cells or tumor cells and secrete a variety of proinflammatory cytokines (Fig. 2). Cytokines and Vγ2Vδ2 T cells expansion Like αβ T cells, cytokines play a fundamental role in the stimulation, differentiation, and function of Vγ2Vδ2 T cells. IL-2 has been shown to be critical for Vγ2Vδ2 T cells expansion. In fact, IL-2 alone, without antigen, can stimulate Vγ2Vδ2 T cells expansion up 15% due to their high expression of the IL-2β receptor (CD122). Besides IL-2, other common γ C chain family cytokines, such as IL-4, IL-7, and IL-15, have been reported to support Vγ2Vδ2 T cell expansion and development. Data presented in this thesis confirm that IL-15 supports the expansion of Vγ2Vδ2 T cells and has major effects on their memory development. The cytokines milieu and microenvironment play an important role in determining γδ T cells function (20, 21). While IL-2 or IL-15 is required for expansion, the presence of other cytokines can guide the development of Vγ2Vδ2 T cells into different functional lineages. There are two main subsets of Vγ2Vδ2 T cells: IFN-γ producing Vγ2Vδ2 T cells, which is the dominant subsets (>90%) and IL- 17 secreting Vγ2Vδ2 T cells (less than 2%) (21). While IL-12 along with IL-2 polarizes Vγ2Vδ2 T cells towards IFN-γ producing Vγ2Vδ2 T cells, the presence of IL-1β, IL-6 and TGF-β favors the development of IL-17 secreting Vγ2Vδ2 T cells (21). Furthermore, we and others have found that the addition of IL-18 to IL-2 enhances the expansion of Vγ2Vδ2 T cells and increases their IFN-γ production (22, 23). Due to their constitutive expression of the IL-21 receptor, Vγ2Vδ2 T cells also can develop into T FH cells and provide help to B cells in the presence of IL-21 cytokine (24, 25). Vγ2Vδ2 T cell proliferation can be inhibited by anti-inflammatory cytokines. through the secretion of IL-10, CD4 Treg cells suppress the expansion but not the cytokine production of Vγ2Vδ2 T cells (26, 27). Also, CD8 αβ T cells have been reported 2

22 to indirectly suppress Vγ2Vδ2 T cells by competing for IL-7 and IL-15 (28, 29). Given the great impact that cytokines have on expansion, development, and function of Vγ2Vδ2 T cells, finding the right cytokine combination could enhance the efficacy of Vγ2Vδ2 T cells in cancer immunotherapy. Memory of Vγ2Vδ2 T cells The development of immunological memory of Vγ2Vδ2 T cells has unique characteristics as compared to CD8 and CD4 αβ T cells (30). Whereas a large proportion of adult αβ T cells are naïve, the vast majority of adult Vγ2Vδ2 T cells are memory cells. Less than 2% of Vγ2Vδ2 T cells are naïve in healthy adults (31). The memory phenotype of Vγ2Vδ2 T cells are very similar to their counterparts in CD8 αβ T cells (31-33). One distinctive characteristic is that even at birth, many Vγ2Vδ2 T cells exhibit memory phenotypes with the expression of inflammatory chemokine receptors along with the production of inflammatory cytokines. The development of Vγ2Vδ2 T cell memory is at least partly due to environmental factors as evidenced by the large differences observed in the γδ TCR repertoire in some identical twin pairs (4). Moreover, the development of immunological memory has been demonstrated for Vγ2Vδ2 T cells in both human and non-human primates (34, 35). Previous studies have shown the importance of Vγ2Vδ2 memory cells in providing protection against infections (34-36). The phenotypic markers that best delineate the memory subsets of Vγ2Vδ2 T cells are controversial. Many studies have used CD45RA and CD27 to define naive cells (37). However, using CD27 and CD45R with quadrant gating overestimates the number of naive Vγ2Vδ2 T cells. In our laboratory, we have found that expression of CD27, CD28, and the absence of CD45RO allows the delineation of the major memory subsets. Early memory Vγ2Vδ2 T cells (central memory) express CD27 and CD28 with 3

23 CD45RO. Intermediate memory cells have lost expression of CD28 but retain expression of CD27 while expressing both CD45RA and CD45RO. Late memory cells or effector memory cells (also termed T EMRA cells) have lost expression of both CD27 and CD28 and have upregulated CD45RA expression while downregulating CD45RO. Based on these phenotypic markers, Vγ2Vδ2 T cells have been extensively studied in our laboratory. There are ~1200 genes whose expression significantly differs between early, intermediate, and late memory subsets. There are marked differences between the memory subsets in their proliferation capacity, cytotoxicity, and chemokine receptor expression (unpublished data). Early (central) memory Vγ2Vδ2 T cells have high proliferative capability but produce less cytotoxic mediators compared to terminally differentiated cells late memory T cells. Late memory cells tend to secret more IFN-γ and granulolysin compared to early memory cells(36). Although normal effector memory T cells can transiently express inhibitory receptors, exhausted or anergic T cells exhibit prolonged expression of inhibitory receptors such as PD-1, CTLA4, TIM3, and LAG3. Additionally, late memory cells tend to exhibit stronger effector functions such as cytotoxicity. For mediating cancer immunity, however, early/central memory cells are thought to be favored, presumably because of their ability to proliferate (38). Therefore, studies have focused on increasing early/central memory Vγ2Vδ2 T cells to optimize their anti-tumor response. The expression of chemokine receptors also differs between different memory Vγ2Vδ2 T cell subsets. Most Vγ2Vδ2 T cells express CXCR3 and CCR5 chemokine receptors and home to sites of inflammation rather than to lymphoid tissue due to their lack of CCR7 expression. A small percentage of Vγ2Vδ2 early memory cells express CCR7 and CCR6. Intermediate or late memory cells express CXCR1, CXCR2, and CX3CR1 (37, 39). Differences in the migratory capabilities of Vγ2Vδ2 memory subsets may determine the effectiveness of Vγ2Vδ2 T cells in cancer immunotherapy. 4

24 BTN3A1 sensor function and stimulation of Vγ2Vδ2 T cells by prenyl pyrophosphates One of the latest discoveries in the field of Vγ2Vδ2 T cells has been the identification of the critical role of butyrophilin 3A1 (BTN3A1 or CD277) in the stimulation of Vγ2Vδ2 T cells by isoprenoid metabolites. For more than two decades, the mechanism by which Vγ2Vδ2 T cells are activated by isoprenoid metabolites was unknown. Recent reports have shown that BTN3A1 is essential for this process (40-44). BTN3A1 is an immunoglobulin superfamily protein that is expressed on all lineages of human cells albeit at different levels. BTN3A1 is a member of the BTN3 family that is comprised of three isoforms, BTN3A1, BTN3A2, and BTN3A3. While the three isoforms share similar extracellular domains, only BTN3A1 and BTN3A3 have intracellular B30.2 domains. It was initially unclear where the isoprenoid metabolites were binding. One study identified a region in the IgV extracellular domain as the binding site (43) whereas our lab and Sandstrom et al. presented evidence that the intracellular B30.2 domain was the binding site (42, 44). Sandstrom et al. showed direct binding to the B30.2 domain but not to the extracellular domains and were able to crystallize the complex of the B30.2 domain of BTN3A1 and the HMBPP analog, HMBcPP (44). Moreover, mutations in the extracellular binding site had no effect on Vγ2Vδ2 T cells stimulation whereas mutation of residues in the B30.2 binding site totally abolished stimulation (45). Based on these studies, the "phosphoantigens" were not actually antigens but sensed molecules. Therefore, Vγ2Vδ2 T cells introduce a new paradigm of T cell activation in which a T cell can be activated through the action of a transmembrane sensor protein that binds stimulators intracellularly. BTN genes are located next to MHC genes on chromosome 6 suggesting that some may play a similar role to MHC molecules. Although essential for Vγ2Vδ2 T cells stimulation by isoprenoid metabolites, inducing the expression of BTN3A1 in murine and hamster APC failed to stimulate Vγ2Vδ2 T cells. Therefore, there is at least one other 5

25 critical molecule required for Vγ2Vδ2 T cells stimulation. Studies have shown that the other required molecule(s) are located on chromosome 6 (46). The presence of human chromosome 6 in a human:cho cell hybrid line was sufficient to support stimulation of Vγ2Vδ2 T cells by isoprenoid metabolites whereas expression of BTN3A1 alone was not. BTN3A1 is expressed on both normal and cancer cells. Our laboratory has not found a cancer cell line that does not express BTN3A1. However, the expression of BTN3A1 varies and low BTN3A1 levels can limit the ability of cancer cells to stimulate Vγ2Vδ2 T cells in response to IPP or HMBPP. Given that BTN3 genes play similar roles to MHC class I molecules, BTN3 expression might be enhanced by same cytokines that enhance MHC class I. In fact, our preliminary data showed that IFN-γ but not IFN-α or IL-12, increases the expression of BTN3 on tumor cells (unpublished data). Therefore, understanding the mechanism by which BTN3 expression is regulated will further benefit cancer immunotherapy with Vγ2Vδ2 T cells. Functions of Vγ2Vδ2 T cells Vγ2Vδ2 T cells likely play important roles in human immunity to microbes and tumors given that they can expand to up to 50% of more of circulating T cells. They have been difficult to study because none of the animal model species (e.g. mice, rats, guinea pigs, rabbits, fish, chickens, dogs, cats, and most ruminants) have a homologous subset of γδ T cells able to detect isoprenoid metabolites. However, knowledge about this subset of T lymphocytes and their roles in health and disease is slowly coming into focus. There are two main arms of immune response: innate immunity, where innate immune cells such as neutrophils and macrophages and other cells respond to pathogens and cell stress signals through the action of germline encoded receptors, and adaptive immunity, where lymphocytes respond specifically to certain antigens through the action of rearranging receptors. Vγ2Vδ2 T cells have been shown to play critical roles in both 6

26 responses (3, 47-50). One of the properties that makes Vγ2Vδ2 T cells unique is that they are multifunctional cells. Vγ2Vδ2 T cells have been shown to be inflammatory and/or inhibitory cells (39, 51-55). Vγ2Vδ2 T cells have the ability to function as antigen presenting cells (APCs). as they are phenotypically and functionally similar to dendritic cells (DC). Moreover, Vγ2Vδ2 T cells are able to cross-present antigens to CD8 αβ T cells inducing CD8 αβ T cell activation and expansion (56-60). Vγ2Vδ2 T cells can also contribute to DC maturation through the secretion of cytokines, primarily TNF-α and IFN-γ. Additionally, Vγ2Vδ2 T cells can recruit inflammatory cells to the site of infection (61). IL-17 producing Vγ2Vδ2 T cells have been shown to increase recruiting neutrophils and macrophages to site of bacterial infection (62-64). The role of IL-17 secreting Vγ2Vδ2 T cells in tumor immunity is controversial. While some studies have shown that IL-17 secretion by Vγ2Vδ2 T cells can initiate anti-tumor response through Th17 cells (65, 66), other studies have indicated that IL-17 producing γδ T cells could help tumor progression by polarizing neutrophils and recruiting immunosuppressive MDSCs, mainly via the secretion of granulocyte-macrophage colony-stimulating factor (GM-CSF) and TNF-α (67, 68). Importantly, Vγ2Vδ2 T cells have demonstrated a potent cytotoxicity against wide range of hematopoietic and non-hematopoietic tumors. While αβ T cells limit their anti-tumor response to immunogenic tumors, Vγ2Vδ2 T cells display a strong anti-tumor activity against tumor with low immunogenicity including tumors with low numbers of coding mutations (68-73). Tumor cells, in general, have to meet two requirements to be susceptible to Vγ2Vδ2 T cell killing activity; mevalonate pathway, which exists in every cancer cell, and expression of BTN3, which is expressed on all tumor cells. A variety of tumor cells have been tested and all tumors express BTN3 (40, 43). Tumor cells in general have an elevated level of IPP due to high metabolic rate. Also, treating tumor cells with aminobisphosphante compounds will sensitize tumor to Vγ2Vδ2 killing 7

27 activity by increasing the endogenous levels of IPP. Then, IPP will bind to the intracellular domain of BTN3A1 (B30.2) causing conformational change that provide required signal to trigger Vγ2Vδ2 TCR. Upon stimulation by isoprenoid metabolites, Vγ2Vδ2 T cells kill tumor cells directly through the secretion of perforin and granzymes or indirectly by inducing apoptosis through Fas-FasL or TRAIL-TRAILR interactions. Also, activated Vγ2Vδ2 T cells secrete large quantities of IFN-γ, a central cytokines for tumor immunity (68, 74-76) (Fig. 2). Furthermore, in vitro and in vivo studies have shown that tumor-killing activity of Vγ2Vδ2 T cells is strongly correlated with the level of mevelonate metabolites (IPP/APPPI) in the tumor cell targets (77). In addition, Vγ2Vδ2 T cells, through their expression of NKG2D, can sense and target some types of tumors the express low level of MHC class I. Some Vγ2Vδ2 T cells also express CD16 allowing them to target tumor cells in the presence of tumor-specific antibodies through ADCC. Therefore, monoclonal antibody such as rituximab (anti-cd20) and tratuzumab (anti-herneu) can increase the killing by Vγ2Vδ2 T cells in vitro and in vivo (78, 79). Thus, Vγ2Vδ2 T cells have unique functional properties that make them very interesting candidate for cancer immunotherapy. Vγ2Vδ2 T cell can function as effector and APC at the same time. They do not require MHC for activation, Vγ2Vδ2 T cells have no alloreactivity and can be adoptively transferred between unrelated donors. However, despite recent discoveries on the mechanism of activation of Vγ2Vδ2 T cells by prenyl pyrophosphates, a complete understanding of this process is still lacking. Tumor immunity of human Vγ2Vδ2 T cells demonstrated in preclinical mouse models of human cancer One of the largest challenges in studying the function of Vγ2Vδ2 T cells is the lack of a homologous γδ T cell subset in mice. However, immunodeficient mice can be used because they will accept xenographs from human Vγ2Vδ2 T cells and human 8

28 tumors allowing investigators to study anti-tumor immunity (80). In these preclinical mouse models, Vγ2Vδ2 T cells slow the growth of a wide range of hematopoietic and non-hematopoietic tumors (summarized in Table 1). The ability of Vγ2Vδ2 T cells to control the growth of hematopoietic tumor cells has been demonstrated in several studies. Malkovska et al. conducted one of the earliest studies on the response of Vγ2Vδ2 T cells against human Burkitt lymphoma cells. In their studies, Malkovska and colleagues showed that human Daudi Burkitt lymphoma cells can directly stimulate the proliferation of Vγ2Vδ2 T cells in vitro and in vivo. Human PBMC increase the survival of SCID mice innoculated with a lethal dose of Daudi cells. This anti-tumor response was due to Vγ2Vδ2 T cells and not NK cells. However, the anti-tumor response was restricted to Daudi lymphoma cells given that Vγ2Vδ2 T cells show no anti-tumor response against the non-stimulatory Raji lymphoma cells (81, 82). Besides exposure to Daudi tumor cells, there are two other methods to expand Vγ2Vδ2 T cells; exposure to aminobisphosphonate or prenyl pyrophosphates or to anti-γδ TCR antibodies. Both protocols can induce similar levels of Vγ2Vδ2 T cells proliferation and function. Zhou et al. showed that antibody-expanded Vγ2Vδ2 T cells also increase the survival of Daudi-bearing nude mice. Additionally, expanded Vγ2Vδ2 T cells have showed a promising response against different types of leukemia in preclinical models. Siegers et al. have shown that the adoptive transfer of antibody-expanded Vγ2Vδ2 T cells into leukemia-bearing NSG mice is well tolerated and reduces the leukemia burden although not prolonging survival. The route of Vγ2Vδ2 T cell administration seems to make a difference in the number of γδ T cells. Administering γδ T cells intraperitoneally in mice results in higher number of Vγ2Vδ2 T cells in the blood and other organs (83). Another study by Gertner-Dardenne et al. showed that Vγ2Vδ2 T cells expanded by bromohydrin pyrophosphate kill the acute myeloid leukemia (AML) cell line, U937, via the perforin/granzyme pathways (84). Expanded Vγ2Vδ2 T cells were able to home to bone marrow, reduce U937 tumor 9

29 burden, and increase survival of NOG mice. Furthermore, the response against AML cells by Vγ2Vδ2 T cells was γδ TCR-dependent and require DNAM-1 (84, 85). The anti-tumor response of Vγ2Vδ2 T cells against various types of solid tumors has also been studied in preclinical mouse models. Vγ2Vδ2 T cells exhibit anti-tumor activity against human breast cancer in mouse models (77, 86-88). The susceptibility of breast cancer cells to in vitro killing by Vγ2Vδ2 T cells strongly correlated with endogenous levels of IPP/APPPI in response to treatment with risesdronate (77). Similar results were noted in xenograft mouse models (86). Moreover, breast cancer cell lines with high level of IPP/APPPI attract activated Vγ2Vδ2 T cells (77). Besides the level of IPP/APPPI, high surface levels of ICAM-1 were associated with improved killing activity of Vγ2Vδ2 T cells against breast cancer (86). The traffic of bisphosphonate-expanded Vγ2Vδ2 T cells in healthy mice differs from tumor-bearing mice. Localization of activated Vγ2Vδ2 T cells to breast tissue has been shown to be TCR- dependent (88). In addition to direct anti-tumor responses, Vγ2Vδ2 T cells also enhance the efficacy of antibody treatment. Adoptive transfer of activated Vγ2Vδ2 T cells along with the anti- Her2/neu monoclonal antibody, trastuzumab, resulted in better control of human HER-2 + breast cancer in immunodeficient mice (87). The response of Vγ2Vδ2 T cells against lung cancer cells has been explored in several preclinical studies. Adoptive transfer of Zol-expanded Vγ2Vδ2 T cells 12 h after zoledronate treatment reduces the growth of small cell-lung carcinomas and increases survival (89). Adoptive transfer of γδ T cells expanded by either anti-cd3 or anti-γδ TCR antibodies also prolongs the survival of NSCLC-bearing mice (90, 91). In addition to lung cancer, activated Vγ2Vδ2 T cells also inhibit bladder cancer growth. Intravesical administration of Zol-expanded Vγ2Vδ2 T cells along with low doses of Zol also allows survival of mice with UM-UC-3Luc bladder cancer (92). Other studies also found that prolonged anti-tumor activity during treatment with Vγ2Vδ2 T cells requires repeated administration of expanded Vγ2Vδ2 T cells along with 10

30 a bisphosphonate. For instance, Kabelitz and colleagues demonstrated that Vγ2Vδ2 T cells given ip. along with alendronate prolong the survival of SCID mice bearing either human melanoma or pancreatic adenocarcinoma tumors (93). The reactivity of Vγ2Vδ2 T cells against melanoma and pancreatic adenocarcinoma was shown to be γδ TCRdependent. Santolaria et al. also demonstrated that repeat weekly treatments with Zolexpanded Vγ2Vδ2 T cells and pamidronate every week is required to control the growth of PC-3 prostate cancer tumors in NSG mice (94). Thus, optimal treatment of solid tumors is effected by repeat treatments with expanded Vγ2Vδ2 T cells and bisphosphonates. Vγ2Vδ2 T cells can also treat pediatric solid tumors in preclinical mouse models. These tumors are difficult to treat with monoclonal antibodies to checkpoint receptors or their ligands because they have few somatic mutations and therefore few neoantigens for endogenous αβ T cells to react with. Moreover, due to low MHC class 1 levels on neuroblastomas, cellular immunotherapy with CD8 αβ T cells has not been effective. Several groups have studied treatment of neuroblastoma with Vγ2Vδ2 T cells. Weekly treatments with Vγ2Vδ2 T cells and zoledronate only minimally extended survival of mice bearing the human neuroblastoma cell line, SH-SY-5Y (76). However, combination treatment with the h14.18 monoclonal antibody to the ganglioside, GD2, Vγ2Vδ2 T cells, and zoledronate (given 1 day earlier) greatly slowed the growth of the neuroblastoma cell line, Kelly, and the Ewing's sarcoma cell line, TC71 (95). Similarly, treatment with the h14.18 monoclonal antibody, Vγ2Vδ2 T cells, and the cytokine, Fc-IL-7 prolonged survival of mice that have been injected intravenously with the neuroblastoma cell line, NB-1691, even in the absence of a bisphosphonate (96). Vγ2Vδ2 T cells also slightly slowed the growth of the human chondrosarcoma cell line, SW1353, in mice repeatedly treated with Vγ2Vδ2 T cells and zoldronate (given 1 day earlier) (97). Therefore, Vγ2Vδ2 T cells could provide a route to immunotherapy of pediatric solid tumors 11

31 In addition to inducing Vγ2Vδ2 T cells proliferation, anti-γδ TCR antibodies can induce Vδ1 T cells proliferation. Vδ1 T cells appear to be more capable of infiltrating tissue in response to chemokines than Vδ2 T cells (98). Vδ1 T cells have been reported to control the growth in humanized mouse models of human colon cancer (99), ovarian cancer (100), and chronic lymphcytic leukemia (101). In contrast, other studies have shown that Vδ1 T cells contribute negatively to anti-tumor response and the number of Vδ1 T cells predicts poor prognosis. Inhibiting tissue-infiltration of Vδ1 T cells has led to enhanced tumor immunity by tumor specific T cells (102). Thus, clinical trials will be necessary to assess the role of Vδ1 T cells in treating malignancy. In summary, human Vγ2Vδ2 T cells represent a potential advance in cancer immunotherapy that can be used to treat a variety of tumors. Almost all kinds of tumor cells can trigger the activation of Vγ2Vδ2 T cells as long as they have elevated levels of IPP from treatment with an aminobisphosphonate. Pretreating tumor-bearing mice with an aminobisphosphonate along with the systemic administration of Vγ2Vδ2 T cells will ensure better killing of tumor cells if given repeatedly. Anti-tumor response of Vγ2Vδ2 T cells in clinical trials In vivo immunization Two main approaches to use Vγ2Vδ2 T cells to treat cancer have been studied in clinical trials. The first approach is to activate Vγ2Vδ2 T cells in vivo (Fig.3, left side). In this approach, patients with different types of cancer were treated with an aminobisphosphonate (mostly zoledronic acid) and IL-2 aiming for in vivo activation of Vγ2Vδ2 T cells. The advantage to using this approach is that it exploits other properties of zoledronic acid besides its stimulation of Vγ2Vδ2 T cells. Also, the cost of this approach is relatively low compared to adoptive immunotherapy. However, a major 12

32 drawback of the in vivo approach is that it can induce anergy of Vγ2Vδ2 T cells as well as their deletion (23, ). Wilhelm et al. was the first study to report the effects of in vivo activation of Vγ2Vδ2 T cells in patients with Non-Hodgkin lymphoma (NHL) and multiple myeloma (MM) (106). The study showed that in vivo administration of pamidronate along with IL- 2 significantly increases the number of Vγ2Vδ2 T cells and enhances their secretion of IFN-γ. As a result of the in vivo activation of Vγ2Vδ2 T cells, 33% of treated patients achieved a partial regression with an additional two patients obtained stable disease. Patients were selected based on their response to in vitro activation of Vγ2Vδ2 T cells by pamidronate and IL-2 (106). In a pilot study conducted by Dieli et al., nine patients with breast and prostate cancer metastatic to bone were studied (107). Treatment with zoledronate altered the memory subset profile of Vγ2Vδ2 T cells with the loss of naive and early memory cell. After 3 months of zoledronate treatment, 80% of Vγ2Vδ2 T cells population was effector cells versus less than 10% memory cells (107). In a phase I clinical trail, Dieli et al. treated nine patients with metastatic prostate cancer with zoledronate with or without IL-2. In vivo treatment with zoledronate and IL-2 resulted in better clinical outcomes compared to zoledronate alone. Treatment stabilized disease in 45% of patients with 22% achieving partial remissions at 12 months of the treatment. In addition, the number of in vivo activated Vγ2Vδ2 T cells, especially effector subsets, demonstrated strong correlation with serum level of prostate specific antigen (108). The feasibility and safety of in vivo treatment with zoledronate and low-level subcutaneous IL-2 was confirmed with patients with other solid tumors. In a phase I clinical trial, Meralviglia et al. treated patients with advanced breast cancer (109). In vivo activation of Vγ2Vδ2 T cells by zoledronate and IL-2 treatment was achieved in all patients. However, only 30% (3 patients) of the patients sustained normal Vγ2Vδ2 T cells whereas in the others, Vγ2Vδ2 T cells levels decreased to near zero after 12 months of 13

33 treatment. Of the three patients who maintained a stable level of blood Vγ2Vδ2 T cells, one achieved a partial remission whereas the other two had stable disease (109). To further assess the efficacy of in vivo activation approach with solid and hematopoietic tumors, Kuzmann et al. conducted a phase I/II trail with patients with metastatic renal cell carcinoma, melanoma, and AML (110). Although the treatment was well tolerated and induced in vivo Vγ2Vδ2 T cell activation, no clinical responses were observed in patients with solid tumors beyond stable disease in 6 patients. 25% of AML patients (2/8) achieved a partial remission. Surprisingly, patients treated with zoledronate and IL-2 showed elevated levels of vascular epithelial growth factor (VEGF), which correlates with a poor prognosis. This might explain why hematopoietic tumors appear to be more responsive to in vivo activation (110) compared with solid tumors. Finally, a pilot study by Lang et al with patients with renal cell carcinoma demonstrated that repeated treatment with zoledronate and IL-2 impaired the function and proliferation of Vγ2Vδ2 T cells on subsequent immunizations (111). No objective clinical responses were noted although two patients experience prolonged stable disease. Therefore, few clinical responses were noted except for the initial study with the weak aminobisphosphonate, pamidronate, where partial remissions were noted in patients with NHL. A major problem with this approach is the rapid induction of anergy and deletion of Vγ2Vδ2 T cells after the initial 1-3 cycles. Adoptive cancer immunotherapy with Vγ2Vδ2 T cells generated ex vivo The second approach being tested is adoptive immunotherapy where a patient's Vγ2Vδ2 T cells are expanded ex vivo and then reinfused. This avoids the anergy and deletion of Vγ2Vδ2 T cells that results from the in vivo immunization approach. For adoptive therapy, PBMCs are harvested from a patient by leukophoresis and their Vγ2Vδ2 T cells expanded in vitro with an aminobisphosphonate and IL-2 or with another 14

34 stimulator. The expanded Vγ2Vδ2 T cells are then reinfused into the same patient. This approach preserves the responsiveness of Vγ2Vδ2 T cells by banking the PBMCs by freezing in liquid nitrogen. Aliquots of the PBMC can then be thawed for use as needed. Drawbacks of this approach are the requirement for the use of Good Manufacturing Practices (GMP), the high technical expertise and special facilities needed for the production of clinical-grade cell products, and the high expense. Despite these obstacles, a number of clinical trails with Vγ2Vδ2 T cells have been conducted treating several types of cancer to assess the safety, feasibility, and efficacy of the adoptive immunotherapy approach. Bennouna et al. was the first group to perform a clinical trial with ex vivo activated Vγ2Vδ2 T cells (112). Patients with advanced renal cell carcinoma received multiple infusions of Vγ2Vδ2 T cells expanded ex vivo with or without low-dose IL-2. The treatment was well tolerated with few grade 3 or 4 toxicities mainly related to the use of IL-2. Treatment resulted in stable disease in 60% and progressive disease in 40% of the enrolled patients (112). Adoptive treatment with Vγ2Vδ2 T cells expanded ex vivo was also safe and well tolerated in six patients with multiple myeloma (113). For this study, Abe et al. used ex vivo stimulation with zoledronate and IL-2 to expand Vγ2Vδ2 T cells. although the expansions were modest so somewhat less numbers of Vγ2Vδ2 T cells were infused, the M proteins (a marker for the number of myeloma cells) stayed at baseline in 4/6 patients and increased in the other two patients (113). Adoptive immunotherapy with Vγ2Vδ2 T cells for advanced non-small cell lung cancer (NSCLC) has also been studied. In a phase I clinical trial, Sakamato et al. treated fifteen patients with NSCLC. Patients received between 3-12 cycles of treatment with most receiving six cycles of zoledronate-expanded Vγ2Vδ2 T cells alone every other week. Half of treated patients showed stabilized disease for four weeks after the last adoptive transfer where another six patients showed progressive disease. The number of blood Vγ2Vδ2 T cells in treated patients increased gradually with each cycle but there 15

35 was no correlation with the clinical response. The study concluded that treatment of NSCLC with ex vivo expanded Vγ2Vδ2 T cells was safe and may serve to help stabilize the disease (114). In a similar study, Okawaki et al. treated 228 patients with Vγ2Vδ2 T cells expanded ex vivo with zoledronate (115). Patients averaged 6.8 treatment cycles ranging from 1->30 with few severe toxicities. Three patients in whom oral fluorouracil was concomitantly administered achieved a complete response with partial responses in 6 other patients for a response rate of 13.2% (9/68 patients evaluated). In addition, 58.8% (40/68) had stable disease. Noguchi et al. treated an additional 25 patients with various solid tumors with Vγ2Vδ2 T cells expanded by zoledronate on 2 week cycles (116). Again, little toxicity was noted. Responses were modest with stable disease in 2 patients and 3/10 patients treated with a combination of chemotherapy and Vγ2Vδ2 T cells with partial responses. Thus, although adoptive therapy with ex vivo expanded Vγ2Vδ2 T cells alone is safe, feasible, and well tolerated, the clinical responses are modest and limited to stabilizing disease in a fraction of the patients ( ). To further increase the effectiveness of treatment with Vγ2Vδ2 T cells, in vivo treatment with zoledronate and IL-2 was added to the regimen. In phase I/II clinical trail, Kobyashi et al. reported on patients with metastatic renal carcinoma treated with ex vivo expanded Vγ2Vδ2 T cells and zoledronate and IL-2. Eleven patients were treated with intravenous zoledronate followed by adoptive transfer of Vγ2Vδ2 T cells and subcutaneous IL-2 daily for five days. 1-6 treatment cycles were completed. There was one durable complete remission (lasting at least three years) (121), 5 patients with stable disease, and 5 who progressed. For this study, Kobyahi et al used a synthetic pyrophosphomomester, 2-methyl-3-butenyl-1-pyrophosphate, to expand Vγ2Vδ2 T cells rather than zoledronate (117). Nicol et al. treated 18 patients with a variety of solid tumors with ex vivo expanded Vγ2Vδ2 T cells (up to cells) along with 1 mg zoledronate given 1 day 16

36 prior and 1 mg immediately before the cell infusion (118). Prior treatments were allowed to continue and the results evaluated after 6-8 treatment cycles. Again, no dose limiting toxicity was noted and the infused Vγ2Vδ2 T cells preferentially localized to peripheral tissue and tumor sites as assessed using indium III-oxine labeled T cells. Most patients progressed during treatment with 5 patients with stable disease. Notably, there was one complete remission and two partial remissions in patients who had continued previously ineffective therapies suggesting an additive effect. In conclusion, adoptive immunotherapy with Vγ2Vδ2 T cells expanded ex vivo is safe, feasible, and well tolerated but the clinical responses achieved thus far are modest. The best results were achieved when cell therapy was combined with zoledronate and IL- 2 or with concurrent chemotherapy or hormonal therapy. As of yet, clinical trials combining Vγ2Vδ2 T cells with checkpoint blockade, idoleamine 2,3 dioxygenases inhibitors, or immunostimulants has not been tried. These combination approaches could lead to breakthroughs in increasing the effectiveness of Vγ2Vδ2 T cell adoptive therapy for a variety of hematopoietic and non-hematopoietic tumors. Rational and objectives Vγ2Vδ2 T cells demonstrate a strong toxicity against wide range of tumor including tumors with low immunogenicity. Based on preclinical studies and clinical trails, using Vγ2Vδ2 T cells for cancer immunotherapy is promising and hold a great potential. the main approach to expand Vγ2Vδ2 T cells for clinical use is the ex vivo activation approach in which Vγ2Vδ2 T cells extracted from patients and expanded in vitro. However, using current protocol has been shown to impacts the expansion and function of Vγ2Vδ2 T cell resulting in suboptimal quantity and quality of Vγ2Vδ2 T cells. Optimizing the ex vivo approach of expansion Vγ2Vδ2 T cells is needed in order to improve the clinical response of cancer adoptive immunotherapy with Vγ2Vδ2 T cells. 17

37 Addressing the limitations of an exist protocol is the main object of this work. One of the main limitations of ex vivo approach is the long term exposure to zoledronate, which in turn inhibit Vγ2Vδ2 T cells by blocking their mevelonate pathway and impairing cell signaling. Therefore, the central hypothesis of this work is shortening zoledronate exposure will increase the quantity and the anti-tumor activity of ex vivo expanded Vγ2Vδ2 T cells. in addition, cytokines like IL-15 and IL-18 have shown to play a critical role in tumor immunity by influencing the proliferation, memory distribution, and function of T lymphocytes. Thus, finding a right cytokines combination could help further optimize the function of Vγ2Vδ2 T cells against tumor cells. 18

38 Table 1. Pre-clinical studies using Vγ2Vδ2 T cells to treat cancer in humanized mouse models Type of tumor being treated Mouse Strain Stimulator of γδ T cells Approach of using γδ T cells for therapy Number of injected γδ T cells Treatment outcome Reference First author Burkitt's lymphoma (Daudi) i.p. or i.v. (retroorbital sinus) SCID Daudi Enriched Vγ2Vδ2 T cells (52%) Vγ2Vδ2 T cells (45%) or Vγ2Vδ2 T cell clones i.v. Improved survival Malkovska (81) Burkitt's lymphoma (Daudi) SCID Daudi Human PBMC expanded in vivo i.p i.p. irradiated Daudi Improved survival Malkovska (82) Burkitt's lymphoma (Daudi) Ovarian Cancer (SKOV3) Nude mice IL-2 Purified γδ T cell TIL from Colorectal Cancer x 10 7 i.p. + ril-2 i.p. (x IU/mouse) Improved survival Chen (122) Burkitt's lymphoma (Daudi) i.v. Leukemia Ph + (EM-2eGFPluc) i.v. Acute myeloid leukemia (U937) i.v. Nude mice NSG mice (NOD.Cg- Prkdc scid Il2rg tm1wjl / SzJ, Jackson) cgy Irradiaiton NOG mice (NOD- SCID- γ c -/- Jackson) Purified Vγ2Vδ2 T cells coinjected with Daudi cells + ril-2 injected i.v. Purified γδ T cells cultured with irradiated PBMCs feeders + mitogen (concanavalin A 1 µg/ml) Ex vivo expansion with 3 µm BrHPP + IL-2 (100 IU/ml) + IL-15 (10 ng/ml) Immobilized anti-γδ TCR antibody + IL-2 (24 well plate coated with antiγδ TCR abs Vγ2Vδ2 T cells were purified (positive selection) and in vitro expanded for 8 d BrHPP-treated PBMC or purified Vγ2Vδ2 T cells Pronounce antilymphoma effect in vivo i.v. or i.p. + ril-2 i.p. (100 IU/mouse) (enriched (42%) or purified (98.8%) Vγ2Vδ2 T cells) No prolongation of survival but good tolerance to γδ T cell treatment Strong anti-tumor activity Zhou (98) Siegers (83) Gertner- Dardenn (84) 19

39 Table 1- continued Chronic myelogenous Leukemia (MM1Luc) NOD-SCID mice Ex vivo expansion with Zol (0.5 µm) and ril-2 (50 IU/ml) Purified Vγ2Vδ2 T cells (positive selection) i.v. + Zol (2 µg i.p.) + ril-2 i.p. (30,000 IU/mouse) Combined therapy showed strong anti-tumor activity D Asaro (75) Breast cancer (MDA-MB-231, 2Lmp/Luc subclone) s.c. NCR nu/nu Ex vivo expansion with anti-cd2 Purified γδ T cells were injected intravenously i.v. every three days Strong reduction of tumor growth Beck (88) Breast cancer HER2 + (SK-BR-3) s.c. in mammary area SCID/Beige (Female) Ex vivo expansion with BrHPP + IL-2 (300 IU/ml) Ex vivo stimulation with 3 µm of BrHPP for 14 d (not sorted but >95%) i.v. + Trastuzamab (10 mg/kg) i.p. Moderate reduction of tumor growth Capietto (87) Breast cancer (T47D) (BO2) NOD/SCID Ex vivo and in vivo expansion by Zol for 14 days Vγ2Vδ2 T cells were purified by positive selection or whole PBMC were injected in mice i.p. (PBMC) Vγ2Vδ2 T cells infiltrated and inhibited breast cancer cells that has high IPP/ApppL Benzaïd (77) Breast cancer (T47D) (BO2) NOD/SCID Ex vivo and in vivo expansion by Zol for 14 days Vγ2Vδ2 T cells were purified by positive selection or whole PBMC were injected in mice i.p. (PBMC) then Zol or Risedronate (100 µg/kg) + IL-2 (10,000 IU) + Doxorubicin (2 mg/kg) Vγ2Vδ2 T cells inhibited breast cancer tumor growth Benzaïd (86) NSCLC Lung cancer, squamous (NCI- H520) s.c. Nude mice Immobilized human anti-γδ TCR abs + IL-2 (PBMC from donors) Anti-γδ TCR expanded γδ T cell. Purified by positive selection i.p. 1 to 3 Moderate antitumor activity against lung cancer in vivo 3/8 survived Kang (91) NSCLC Lung cancer, large cell (H460) SCID mice Ex vivo expansion with anti-cd3 CD4, CD8 depleted PBMC were cultured with anti-cd (purified) Inhibit the growth of lung cancer in vivo Dokouhaki (90) Small cell lung cancer (SBC-5Luc) s.c. BALB/c nu/nu mice (male) Ex vivo expansion with Zol in the presence of IL-2 Enriched Vγ2Vδ2 T cells (positive selection, 99.8%) +/- Zol were injected i.v i.v. (once a week 3) 12 h after Zol (80 µg/kg i.v.) Strong cytotoxicity against prostate cancer Sato (89) 20

40 Table 1- continued Bladder cancer (UM-UC-3Luc and KU-7) intravesically SCID mice (male) Ex vivo expansion with 5 µm Zol for 14 days Vγ2Vδ2 T cells were purified by negative selection injected + Zol (5 µm 200 µl) intravesically In vivo inhibition of bladder cancer growth (100% remission with Vγ2Vδ2 T cells + Zol) Yuasa (92) Neuroblastoma (SH-Sy-5Y) into adrenal Balb/c nu/nu mice (Harlan) Ex vivo expansion with Zol (5 µm) and ril-2 (50 IU/ml) Vγ2Vδ2 T cells (>d14 and no purification but >90%) i.v. -/+ 16 h after Zol (150 µg/kg) Slight increase in survivial Di Carlo (76) Neuroblastoma (Kelly) Ewing's sarcoma (TC-71) NSG mice (NOD.Cg- Prkdc scid Il2rg tm1wjl / SzJ, Charles River) Ex vivo expansion with Zol (5 µm) and ril-2 (100 IU/ml) Purified Vγ2Vδ2 T cells (d14 positive selection) i.v. -/+ 16 h after Zol (120 µg/kg) + anti-gd2 (ch14.18/cho) Combination therapy limited tumor growth Fisher (95) Neuroblastoma (NM-1691) NOD-SCID mice Administer G-CSF to healthy donors Purified Vγ2Vδ2 T cells (positive selection) +/- hu14.18 antibody +/- Fc- IL Combination therapy enhanced survival and limited tumor growth Otto (96) Chondrosarcoma (SW1353-Luc) BALB/c nu/nu mice (male) Ex vivo expansion with Zol (1 µm) and ril-2 (100 IU/ml) Purified Vγ2Vδ2 T cells (d12-14 positive selection) i.v. -/+ 16 h after Zol (50 µg/kg) Combination therapy slowed tumor growth slightly Sun (97) Melanoma (MeWo) Pancreatic cancer (PancTu1) i.p. SCID (irradiated 300 rad + anti-asialo- GM1) BrHPP (200 nm) Alendronate (5 µm) Negative selection for expanded Vγ2Vδ2 T cell lines restimulated 2 with PHA after initial stimulation i.p. then varying amounts on d4, 10, 20, 30 + Alendronate 10 µg + ril-2 (300 ng) d0, 4, 11, 18, 25, 32, 39 Prolonged survival Kabelitz (93) Prostate cancer (PC-3) s.c. NSG mice (NOD.Cg- Prkdc scid Il2rg tm1wjl / SzJ, Charles River) Ex vivo expansion with pamidronate (200 µm) and ril-2 (60 ng/ml) out to d21 or more Resting Vγ2Vδ2 T cell lines expanded with pamidronate and IL-2 was used for adoptive transfer (70-95% Vγ2Vδ2 T cells) i.v. d15 then every week + pamidronate (50 µg/kg i.v.) Tumor growth is controlled by systemic administration with pamidronate and γδ T cells Santolaria (94) 21

41 Table 1- continued Cervical cancer SiHa s.c. SCID Purified γδ T cell TIL from colorectal cancer Immobilized anti-γδ TCR antibody + IL-2 (24 well plate coated with antiγδ TCR abs + IL-2 (400 IU/ml) s.c. + anti-galectin- 1 mab Improved survival moderately Li (123) Ovarian cancer (SKOV3Luc- D3) i.p. (IGROVLuc i.p. Ovarian cancer (CAROV3Luc) i.p. SCID/Beige NSG mice (NOD.Cg- Prkdc scid Il2rg tm1wjl / SzJ, Jackson) Ex vivo expansion with Zol (1 µg/ml) and ril-2 (100 IU/ml) + IL- 15 (10 ng/ml) Ex vivo expansion of purified γδ T cells followed by purification Purified Vγ2Vδ2 T cells (d15 positive selection) Ex vivo stimulation with artificial K562 APC + 50 IU/ml IL ng/ml IL i.v. + alendronate (100 µg.i.v. +/- liposomal) i.p. Prolonged survival, decreased tumor burden (no cures) Greatly prolonged survival but Vδ1 better than Vδ2 T cells Parente- Pereira (124) Deniger (100) Chronic lymphocytic leukemia (MEC1Luc) s.c. NSG mice (NOD.Cg- Prkdc scid Il2rg tm1wjl / SzJ, Charles River) Ex vivo expansion of purified γδ T cells followed by purifying Vδ1 T cells Purified γδ T cells stimulated by anti-cd3 (OKT3 70 ng/ml) + IL-4 + IFN-γ + IL-21 + IL-1β i.v. d i.v. d11 Prolonged survival by 15 days with 1 complete remission Almeida (101) Colon Carcinoma HT s.c. NOD/SCID Ex vivo expansion of Vδ1 T cells followed by purifying Vδ1 T cells Purified γδ T cells stimulated by PHA + IL i.v. Moderate antitumor activity Wu (99) 22

42 Table 2. Clinical studies where Vγ2Vδ2 T cells have been activated in vivo to generate anti-tumor activity Disease Vγ2Vδ2 activation Activating Agent(s) Patients (no) Progressive Disease (%) Stable Disease (%) Partial Remission (%) Reference NHL MM in vivo in vivo pamidronate + IL-2 pamidronate + IL Wilhelm (106) breast cancer prostate cancer in vivo in vivo ZA ZA 3 6 Dieli (107) prostate cancer prostate cancer in vivo in vivo ZA ZA + IL (at 12 mths) (at 12 mths) (at 12 mths) Dieli (108) prostate cancer in vivo ZA decreaesd PSA velocity Naoe (125) renal cell carcinoma in vivo BrHPP + IL-2 BrHPP + IL (7 wks) 50 (15 wks) Innate Pharma PR 2008 solid tumors in vivo BrHPP + IL-2 28 not assessed not assessed not assessed Bennouna (126) breast cancer in vivo ZA + IL Meraviglia (109) NHL in vivo BrHPP + IL-2 + Rituxan PR 26 CR Innate Pharma PR 2010 renal cell carcinoma in vivo ZA + IL Lang (111) renal cell carcinoma MM in vivo in vivo ZA + IL-2 ZA + IL Kunzmann (110) AML in vivo ZA + IL

43 Table 3. Clinical studies using adoptively transferred Vγ2Vδ2 T cells generated ex vivo for cancer treatment Disease Vγ2Vδ2 activation Activating Agent(s) Patients (no) Progressive Disease (%) Stable Disease (%) Partial Remission (%) Reference renal cell carcinoma ex vivo BrHPP + IL-2 IL-2 (in vivo) (100% progressed) Bennouna (112) multiple myeloma non-small cell lung cancer ex vivo ZA + IL ex vivo ZA + IL Abe (113) Sakamoto (114) solid tumors melanoma ovarian carcinona colon carcinoma breast carcinoma ex vivo ZA + IL-2 ZA (in vivo) ZA + IL-2 ZA + Add. Rx (in vivo) PR 33 CR Nicol (118) renal cell carcinoma ex vivo 2M3B1-PP + IL-2 ZA + IL-2 (in vivo) PR 9 CR Kobayashi (117) colorectal carcinoma ex vivo ZA + IL-2 6 6/6 maintained remission Izumi (119) gastric cancer malignant ascites ex vivo ZA + IL-2 ZA + IL-2 (in vivo) 1 d Wada (120) solid tumors ex vivo ZA + IL evaluable 28 (19 pts) 59 (40 pts) 8.8% PR (6 pts) 4.4% CR (3 pts) (with fluorouracil ) Okawaki (115) solid tumors ex vivo ZA + IL only with γδ T cells 10 combo PR Noguchi (116) 24

44 Figure 1. Mevalonate pathway and key downstream branches in isoprenoid biosynthesis in humans. Aminobisphosphonates block FPP synthase in the mevelonate pathway resulting in IPP accumulation which stimulates Vγ2Vδ2 T cells. HMG-CoA reductase is the rate-controlling enzyme in the mevalonate pathway and is subject to feedback regulation by downstream products. It is also inhibited by statins. FPP synthase converts IPP and DMAPP to GPP and FPP intermediates and is inhibited by aminobisphosphonates and alkylamines. Loss of FPP and GGPP leads to the loss of membrane anchoring of signaling proteins causing signaling defects and, in some cases, apoptosis. 25

45 26

46 Figure 2. Model of anti-tumor activity of Vγ2Vδ2 T cells. Tumor and stressed cells or cells treated with aminobisphosphonate or alkylamine compound have elevated levels of IPP which in turn is sensed by the B30.2 domains of BTN3A1 molecules leading to the activation of Vγ2Vδ2 TCR. Activated Vγ2Vδ2 T cells kill tumor cells either directly through the secretion of perforin and granzyme B, and Fas-FasL interaction or indirectly through cytokines (TNF-α and IFN-γ). Also, activated Vγ2Vδ2 T cells express CD16 that can facilitates the ADCC.. 27

47 28

48 Figure 3. Therapeutic strategies for Vγ2Vδ2 T cell immunotherapy. Strategies being pursued in clinical trials include in vivo immunization/activation (127) and adoptive immunotherapy (right). 29

49 30

50 CHAPTER II. Materials and Methods Reagents FITC-conjugated anti-human Vδ2 TCR (clone B6), allophycocyanin-cy7- conjugated anti-human CD3 (clone SK7), FITC-conjugated anti-human TCR γδ (clone B1), allophycocyanin-conjugated anti-human Vδ2 mab (clone B6), FITC- and PEconjugated anti-human CD3 (clone SP34), PerCP-Cy5.5 anti-human CD27 (clone MT271), PE-Cy7-conjugated anti-human CD28 (clone CD28.2), allophycocyaninconjugated anti-human CD45RO (clone UCHL1), PE-conjugated anti-human IL-21 (clone 3A3-N2.1), PE-Cy5- and PE-conjugated anti-human CD107a (clone H4A3), PEconjugated anti-human IFN-γ (clone 4S.B3), and allophycocyanin-conjugated anti-human TNF-α (clone MAB11) antibodies were purchased from BD Biosciences (San Jose, CA). PE-conjugated anti-human CD107a (clone H4A3), PerCP-Cy5.5-conjugated anti-human IL-17 (clone ebio64dec17), PE-conjugated anti-human granzyme B (clone GB11), PEconjugated anti-human IL-4 (clone 4D9-8), and PE-conjugated anti-human perforin (clone dg9) antibodies were purchased from ebioscience (San Diego, CA). The PEconjugated anti-human IL-22 (clone ) antibody was purchased from R&D Systems (Minneapolis, MN). Live cells were distinguished from dead cells by staining with Hoechst or Live/Dead Blue (ThermoFisher Scientific, Waltham, MA). Pamidronate (3 mg/ml) was from Hospira, Inc. (Lake Forest, IL). Zoledronate was provided by Dr. Eric Oldfield. Ex vivo expansion of Vγ2Vδ2 T cells PBMC from random healthy adult donors were obtained by density centrifugation over Ficoll-Hypaque (GE Healthcare Bio-Sciences Corp., Piscataway, NJ) of either random donor leukopaks (12 donors) (obtained from AllCells, LLC., Alameda, CA or the 31

51 Blood Donor Center at the Dana-Farber Cancer Institute) or fresh blood (donors 2, 3, 4, and 6). Leukopak PBMCs were frozen prior to use whereas PBMC isolated from blood were used directly. Note, we primarily used frozen leukopak PBMC because this is the source of cells for most of the clinical trials assessing the adoptive transfer of Vγ2Vδ2 T cells. For most experiments, complete media (C-Media) was used and was prepared with 500 ml RPMI 1640 supplemented with 10 ml of 1 M HEPES, 5 ml of 200 mm L- glutamine, 5 ml of 100 mm sodium pyruvate, 5 ml of 50 MEM essential amino acids, 5 ml of 100 MEM non-essential amino acids, 0.5 ml of 55 mm 2-mercaptoethanol, 5 ml of 1000 penicillin/streptomycin (optional), 0.5 ml of 10 mg/ml gentamicin (optional) (all from Thermo Fisher Scientific, Waltham, MA), and 50 ml of fetal calf serum (FCS) and 12 ml of human AB + serum (Gemini Bio-Products, West Sacramento, CA). The media was adjusted to ph ~7.2 with ml of 2 M NaOH prepared with cell culture grade H 2 O and filtered through a 0.2 µm nylon filtration unit. The FCS was prescreened for support of T cell growth by assessing cloning efficiency of human T cell lines or clones. Human serum was prescreened by assessing support of Vγ2Vδ2 T cell expansion in response to HMBPP. A detailed protocol for this media is available on request. OpTmizer TM media was purchased from Thermo Fisher Scientific and used as directed. IL-2 was used at 1000 IU (60 ng/ml or 4 nm) (either as aldesleukin (Chiron, Emeryville, CA) or teceleukin (Roche Holding AG, Basel, Switzerland)). IL-15 was used at ng/ml and IL-18 was used at 20 ng/ml (both from PeproTech, Inc., Rocky Hill, NJ). For T cell functional assays and for culturing tumor cells, human serum was omitted and 12 ml of FCS was substituted. For expansions performed in 96-well plates, PBMCs in 100 µl media was added to equal volumes of media containing varying concentrations of 2 HMBPP or zoledronate per round bottom well. For pulsing, after 4 h the plates were centrifuged and the media with zoledronate removed by flicking and blotting any remaining media on a 32

52 sterile pad. The cells were washed three times and then 200 µl media was added. On day 3, 100 µl media was removed and replaced with media containing 2 IL-2, IL-15, IL-18, IL-2 + IL-18, or IL-15 + IL-18. For expansions performed in 24-well plates or 75 cm 2 flasks, PMBC were suspended at cells/ml in C-media lacking IL-2 with the indicated amount of HMBPP or zoledronate. For pulsing, the cells were incubated for 4 h at 37 C and 5% CO 2, washed three times with PBS, and resuspended in C-Media without zoledronate for use. On day 3, 50% of the media was removed and replaced with media containing 2 IL-2 or IL-15. PBMC were incubated for 14 d at 37 C and 5% CO 2. Cell growth was monitored by microscope examination and media color. Every 2-4 d, 50% of the media was changed for fresh C-Media (containing IL-2 or IL-15 but without zoledronate or HMBPP) and the cells were split 1:2 depending on cell density. On day 14, the cells were harvested, washed twice with PBS, and counted. Levels of γδ and Vγ2Vδ2 T cells and their differentiation state were assessed by flow cytometric analysis. Purification of Vγ2Vδ2 T cells After ex vivo expansion for 14 d, Vγ2Vδ2 T cells were positively purified using antibody-coated magnetic beads (MACS, Miltenyi Biotec, San Diego, CA). Expanded cells were washed once with PBS, counted, and then cells were resuspended in 0.1 ml buffer, and reacted with 10 µl of allophycocyanin-conjugated anti-human Vδ2 mab (clone B6) for 10 min on ice in the dark in ten 2 ml microcentrofuge tubes. Cells were then washed twice with purification buffer, resuspended in 80 µl of buffer, and 20 µl of anti-allophycocyanin magnetic beads added. The cells and beads were incubated on ice for 15 min and then washed twice cells were resuspended in 500 µl of purification buffer, and then loaded onto an LS column for positive selection. Retained Vδ2 T cells were washed on the column three times with 3 ml of buffer, removed from the magnetic field, and then eluted, washed, and tested either in vitro for their functional 33

53 activity or in vivo for their anti-tumor immunity. Cell purify was evaluated by flow cytometry and the Vγ2Vδ2 T cells used were at least 95% Vδ2 cells (most preparations were > 98% and examples of the purity of the Vγ2Vδ2 T cells are shown in Fig. 4). Note that an anti-vδ2 mab was used to determine Vγ2Vδ2 T cells after expansion by HMBPP or zoledronate because only γδ T cells expressing Vγ2Vδ2 TCRs respond and expand to prenyl pyrophosphates and aminobisphosphonates (15, 128, 129). After expansion, all Vδ2 chains are paired with Vγ2 chains because in adult PBMC, a Vδ2 chain is almost always paired with a Vγ2 chain. The converse is not always the case because Vγ2 chains sometimes pair with Vδ1 chains. CD107a and intracellular cytotoxic protein and cytokine expression after stimulation of expanded Vγ2Vδ2 T cells To assess the functional activity of Vγ2Vδ2 T cells expanded under different conditions, the surface mobilization of CD107a, cytotoxic protein levels, and cytokine levels were measured. CD107a surface mobilization was performed as previously described (44, 130). After 14 d of expansion, Vγ2Vδ2 T cells were either washed twice for use (unpurified cells) or purified using antibody-coated magnetic beads as described above. Human PC-3 prostate cancer cells (ATCC, Manassas, VA) were thawed and cultured in F12 media for 2 d. The PC-3 cells were treated overnight by culture with 200 µm pamidronate. The pamidronate-treated PC-3 cancer cells were washed and used to stimulate unpurified and purified Vγ2Vδ2 T cells. Daudi (ATCC, Manassas, VA) and Raji Burkitt's lymphoma cell lines were cultured in RPMI 1640, washed, and directly used to stimulate Vγ2Vδ2 T cells. For unpurified Vγ2Vδ2 T cells, E:T ratios were based on the number of Vγ2Vδ2 T cells. Cells were mixed at different E:T ratios (1:1 to 100:1) and cultured in C-media at 37 C with PE-Cy5- or PE-conjugated anti-cd107a mabs in the presence of monensin at 4 µl of GolgiStop (BD Biosciences) per 6 ml media. After 4 34

54 h, the cells were harvested, washed, stained with allophycocyanin-cy7- or PE-anti-CD3 and FITC-anti-Vδ2 mabs, and then staining was assessed by flow cytometry. To measure cytokine production and cytotoxic protein capabilities, unpurified or purified Vγ2Vδ2 T cells were stimulated with pamidronate-treated PC-3 cancer cells for 4 h in the presence of monensin at 4 µl of GolgiStop (BD Biosciences) per 6 ml media. Alternatively, purified Vγ2Vδ2 T cells were stimulated with ionomycin (2 µg/ml) and PMA (50 ng/ml) (both from Sigma-Aldrich, St. Louis, MO) for 4-6 h in the presence of monensin at 4 µl of GolgiStop (BD Biosciences) per 6 ml media. For flow cytometric analysis, PBMC were first stained with Live/Dead Blue (Invitrogen), to exclude dead cells followed by staining with allophycocyanin-cy7-conjugated anti-cd3 and FITCconjugated anti-vδ2 mabs. The cells were then washed, fixed, and permeabilized using the Cytofix/Cytoperm Kit (BD Biosciences) and then intracellularly stained with either PE-conjugated anti-ifn-γ, anti-granzyme B, anti-perforin, anti-il-22, anti-il-4, allophycocyanin-conjugated anti-tnf-α, or PerCP-Cy5.5-conjugated anti-il-17a mabs. Cytokine and cytotoxic protein levels of Vδ2 T cells were assessed by flow cytometry. Fluorometric assessment of T lymphocyte antigen specific lysis Cytotoxicity of expanded Vγ2Vδ2 T cells against tumor cells was determined using fluorometric assessment of T lymphocyte antigen specific lysis (FATAL) (131). Briefly, PC-3 prostate cancer cells were treated with 200 µm of pamidronate overnight. Pamidronate-treated PC-3 cells were then stained with PKH-26 using a kit (Sigma- Aldrich, St. Louis, MO). PC-3 cells were washed, resuspended in 75 µl of diluent C, mixed with 75 µl of PKH-26 dye ( M), and incubated for 4 min at room temperature. 150 µl of FCS was then added to stop the staining reaction, and the cells washed once in 10 ml of PBS. The cells were then resuspended in 75 µl PBS and mixed with 75 µl of 5 µm carboxyfluorescein diacetate succinimidyl ester (CFSE) dye (Vybrnat 35

55 CFDA SE Cell Racer Kit, ThermoFisher, Molecular Probes, Eugene, OR). 150 µl of FCS was immediately added to stop the staining reaction. The cells were washed twice with PBS and resuspended for use PKH-26/CFSE stained PC-3 cells were then incubated for 5 h with varying numbers of purified Vγ2Vδ2 T cells that had been expanded either by continuous or pulse stimulation. Tumor specific killing by Vγ2Vδ2 T cells was then assessed by flow cytometry. Target % survival was calculated as (mean CFSE hi percent of test well/ mean CFSE hi percent of spontaneous release) 100. Specific lysis was calculated as 100 % survival. Lytic units (LU) per T cells were calculated for 20% lysis as 10 7 / (( ) x) where x = E:T ratio corresponding to 20% specific target cell lysis (132). Human tumor xenograft and adoptive transfer of purified Vγ2Vδ2 T cells into immunodeficient mice To assess the anti-tumor activity of adoptively transferred Vγ2Vδ2 T cells, human PC-3 cancer cells were xenotransplanted into immunodeficient mice using a model developed by the Scotet laboratory (94) except that treatment was started one day earlier (day 13 rather than day 14) and purified Vγ2Vδ2 T cells were used. Note that for these experiments, human Vγ2Vδ2 T cells allogeneic to the tumor were used. Unlike αβ T cells, Vγ2Vδ2 T cells do not exhibit alloreactivity (133). Therefore, by using purified Vγ2Vδ2 T cells all potential alloreactive responses are avoided. Female, five-week old NOD.Cg-Prkdc scid Il2rg tm1wjl /SzJ (NSG) mice were purchased from the Jackson Laboratory (Bar Harbor, ME) and used at 6 weeks of age. For tumor xenotransplantation, human PC-3 cancer cells were suspended in 200 µl of sterile PBS and s.c. inoculated into the right flank of NSG mice on day 0. Each treatment group consisted of 8 mice. On day 13 (when tumor diameter had reached > 5 mm), mice were injected i.v. with pamidronate (50 µg/kg). The mice weighed 17 to 19 grams and therefore received 0.85 to 0.95 µg of pamidronate. This was followed on day 36

56 14 with the i.v. injection of purified Vγ2Vδ2 T cells that had been expanded either by continuous or pulse zoledronate stimulation with IL-2 or IL-15. For each treatment, Vγ2Vδ2 T cells were purified on day 14 from freshly expanded Vγ2Vδ2 T cells derived from frozen leukopak PBMC from the same donor. These treatments were repeated weekly until week 6. Control mice received only pamidronate treatments. Note that in this model, pamidronate by itself only minimally inhibits the growth of PC-3 tumors (94). Tumor size was assessed once weekly by external measurement of the longitudinal and transverse tumor diameter using a digital vernier caliper. Tumor volume was calculated using the modified ellipsoidal formula where tumor volume (mm 3 ) = (y x 2 ) / 2, where "y" is the longitudinal length and "x" is the transverse width. All experiments involving animals, including their housing and care in pathogen free conditions. Statistical analyses For statistical analyses, either the unpaired t test or the nonparametric Mann-Whitney U test was used as indicated with p < 0.05 considered statistically significant. Statistical analyses were done in Prism version 6.0g (GraphPad Software, La Jolla, CA). 37

57 CHAPTER III A- Enhancing Adoptive Cancer Immunotherapy with Vγ2Vδ2 T Cells through pulse Zoledronate Stimulation and IL-15 Rational Cancer is the second leading cause of deaths in the United States and is responsible for 25% of all deaths. Despite advances in our understanding of its causes, treatment had been limited for many tumor types. However, recent successes in cancer immunotherapy are revolutionizing treatment. Intrinsic T cell immunity against tumors can be released by using mabs to remove inhibition by checkpoint CTLA-4 and PD-1 receptors resulting in responses to many types of tumors. Adoptive immunotherapy with T cells expressing chimeric antigen receptors (CAR) or tumor-reactive αβ TCRs have resulted in cures. Yet, significant limitations exist for these therapies. CAR-T therapy is limited to tumors expressing proteins that allow their specific targeting. This has limited their use for solid, non-hematopoietic tumors (134). Tumor-specific αβ TCRs are difficult to identify and therapy must be individualized for each patient s MHC. Checkpoint blockade with anti- PD-1 does not work in >75% of lung cancer patients and is even less effective against other tumors, such as colorectal carcinomas, that have few neo-antigens due to coding mutations. Although cancer immunotherapy is a breakthrough therapy, additional approaches are needed to realize its full potential. Treatment with γδ T cells expressing Vγ2Vδ2 TCRs is one such therapy that shows promise. In contrast to αβ T cells, the response of human γδ T cells expressing Vγ2Vδ2 TCRs is not MHC restricted (135) but instead require the immunoglobulin superfamily protein, butyrophillin 3A1, that is expressed by all human cells tested (40-44). Thus, tumor cells and normal cells from all tissues can stimulate Vγ2Vδ2 cells. γδ T cells bridge innate and adaptive immunity by using their γδ TCRs in an innate fashion to recognize unconventional ligands associated with cell transformation, infections, and 38

58 inflammation. γδ T cells expressing Vγ2Vδ2 TCRs are found in primates but not rodents and play important roles in human immunity to microbes and tumors. Vγ2Vδ2 T cells expand to very high numbers during many infections (up to 1 in 2 circulating T cells) and these cells can kill infected cells and tumor cells as well as secrete inflammatory Th1 cytokines, chemokines, and growth factors. Vγ2Vδ2 T cells perform these functions by using their TCRs as pattern recognition receptors for self and foreign prenyl pyrophosphates. Prenyl pyrophosphates are essential intermediates in isoprenoid biosynthesis required by both microbes and humans. The major self-stimulator is isopentenyl pyrophosphate (IPP), an intermediate in the mevalonate pathway whereas the major microbial stimulator is (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate (HMBPP), an intermediate in the 2-C-methyl-D-erythritol-4-phosphate pathway (reviewed in Ref. 36). Vγ2Vδ2 T cells can also use their Vγ2Vδ2 TCRs to directly recognize certain malignant B cells such as the Burkitt's lymphoma line, Daudi (136), the plasmacytoma, RPMI 8226 (137), the myeloma cell line, XG-7 (138, 139), and the B cell lymphoma, IBL4 (140). Vγ2Vδ2 T cells also express a variety of NK and SLAM receptors that allow them to recognize and kill a number of different tumor cell lines (141). Treatment of tumor cells with aminobisphosphonates strongly increases this killing by blocking farnesyl pyrophosphate synthase (FPPS), which leads to the accumulation of its upstream metabolite, IPP. High levels of IPP stimulate Vγ2Vδ2 T cells through their TCRs (15, 19) in a process requiring BTN3A1. Several clinical trials have targeted Vγ2Vδ2 T cells either by stimulating Vγ2Vδ2 T cells in vivo (106, ) or by expanding them ex vivo for adoptive transfer ( ). Both approaches have had success in pilot studies treating patients with lymphoma, prostate cancer, renal cell cancer, multiple myeloma, and solid tumors. However, stimulation of Vγ2Vδ2 T cells in patients or primates with intravenous aminobisphosphonates or prenyl pyrophosphates and IL-2 results in large expansions of Vγ2Vδ2 T cells that wanes rapidly 39

59 on subsequent immunizations (23, 103). In contrast, adoptive immunotherapy with Vγ2Vδ2 T cells avoids this loss of responsiveness by using frozen lymphocytes obtained by leukapheresis prior to intravenous zoledronate therapy. This approach also allows patients to be easily screened to select those that exhibit responsiveness and the majority of patients respond. Very large numbers of Vγ2Vδ2 T cells can be obtained with increases in Vγ2Vδ2 T cell numbers of 400- to 10,000-fold. To date, adoptive γδ immunotherapy has resulted in a durable remission in a patient with metastatic renal cancer (121), a complete remission in a patient with metastatic breast cancer (118), and stable disease in 50% of advanced lung cancer patients (142). For most clinical trials and preclinical studies, continuous exposure of PBMC to zoledronate with IL-2 was used to expand Vγ2Vδ2 T cells where zoledronate is slowly diluted over the course of the culture (113, 114, 118, 120, 143, 144). Previously, we found that continuous exposure to zoledronate is toxic to Vγ2Vδ2 T cells and other T cells (19). This toxicity results in the suboptimal expansion of Vγ2Vδ2 T cells with relatively narrow dose ranges for all aminobisphosphonates studied. In contrast, a short period of exposure (pulse) of PBMC to aminobisphosphonates results in uniform expansions of Vγ2Vδ2 T cells over a 100-fold concentration range (19). The cellular toxicity of aminobisphosphonates has been extensively studied and is due to the loss of the FPPS downstream metabolites, farnesyl pyrophosphate (FPP) and geranylgeranyl pyrophosphate (GGPP) and the production of a toxic ATP analog, ApppI. The loss of FPPS metabolites impairs the transfer of farnesyl or geranylgeranyl chains to the C-termini of small GTPases, such as RAS, RAP, RAB, and RHO, and the γ subunit of G protein-coupled receptors that allows them to anchor to the inner leaflet of membranes and traffic to their proper subcellular locations to function in signal transduction. This loss of normal signaling leads to impaired function and apoptotic cell death and can be partially reversed by the addition of farnesol and/or geranylgeraniol ( ). Similarly, the addition of GGPP and IL-18 to purified Vγ2Vδ2 T cells prevents zoledronate toxicity 40

60 and allows their stimulation (148). Unlike prenyl pyrophosphates, zoledronate is not catabolized so its toxicity persists for the culture period as it is slowly diluted. Additionally, the accumulation of IPP leads to the production of APPPI that can directly induce apoptosis through inhibition of the mitochondrial adenine nucleotide translocase (149, 150). Besides bisphosphonate toxicity, an additional factor that could influence the results of adoptive therapy with Vγ2Vδ2 T cells is the γ C growth cytokine used during culture. Present adoptive immunotherapy trials have used IL-2 with zoledronate for Vγ2Vδ2 T cell expansion. However, in mice, CD8 αβ T cells expanded ex vivo using IL- 15 (152, 153) or IL-7/IL-15 (154) rather than IL-2, mediate increased tumor immunity. Increases in tumor immunity have been postulated to be due to increases in early/central memory CD8 αβ T cells given that these cells provide better anti-tumor immunity compared to late memory cells (155). IL-15 also supports Vγ2Vδ2 T cell proliferation in response to IPP and, in combination with IL-12, increases IFN-γ production (156). Thus, expanding Vγ2Vδ2 T cells with IL-15 rather than IL-2 could have benefits but this possibility has not been tested in vivo in humanized mouse models. In this study, we have compared expanding Vγ2Vδ2 T cells by pulse zoledronate exposure to expanding Vγ2Vδ2 cells by continuous exposure with IL-2 or IL-15. We find that expanding Vγ2Vδ2 T cells by pulse zoledronate exposure results in higher purity, numbers, and quality of Vγ2Vδ2 T cells. The Vγ2Vδ2 T cells were significantly more effective at mediating tumor immunity in adoptive immunotherapy halting tumor growth and decreasing tumor volume by 50% compared with Vγ2Vδ2 T cells expanded by continuous zoledronate exposure. pulse zoledronate stimulation similarly improved expansion of Vγ2Vδ2 T cells with IL-15. Interestingly, IL-15 maintains high proportion of central/early memory Vγ2Vδ2 T cells compared to IL-2. However, adoptive transfer of Vγ2Vδ2 T cells expanded with IL-15 did not result in improved tumor immunity compared to those expanded with IL-2 41

61 Results Pulse zoledronate stimulation improves the purity and number of expanded Vγ2Vδ2 T cells In our previous study, we found that continuous aminobisphosphonate exposure during culture inhibits the proliferation of Vγ2Vδ2 T cells as well as other γδ and αβ T cells (19). This toxicity results in the suboptimal expansion of Vγ2Vδ2 T cells over a narrow aminobisphosphonate concentration range whereas pulse exposure results in larger expansions over a wide concentration range (19). To assess the effect on PBMC of pulse versus continuous exposure to zoledronate, PBMC were cultured with varying concentrations of zoledronate either continuously or for a 4 h pulse period. Note that during continuous exposure, zoledronate was slowly diluted over the course of culture with half of the media changed on day 3 and every 2-4 d thereafter with media without zoledronate. Vγ2Vδ2 T cells were then measured on day 14. Most donors responded to continuous zoledronate exposure with the highest levels of Vγ2Vδ2 T cells between 3-10 µm (Fig. 5A, right panels, open circles). For six of the eight donors, higher zoledronate concentrations greatly decreased responsiveness although 2 donors (Donors 4 and 7) responded even at higher concentrations. In contrast, all eight donors responded to pulse zoledronate exposure similarly over a fold range from ~ µm (Fig. 5A, right panels, closed circles). HMBPP shows no toxic effect on Vγ2Vδ2 T cells even at high concentrations (Fig. 5A, left panels). To assess Vγ2Vδ2 T cell numbers after expansion, zoledronate at 1 µm and 10 µm was chosen for continuous stimulation of PBMC to be consistent with clinical trials. Pulse stimulation with zoledronate was done at 100 µm. Despite achieving similar levels of purity in many cases (Fig. 5A, right panels), the number of Vγ2Vδ2 T cells at day 14 expanded by pulse stimulation with zoledronate (100 µm) in 96-well plates averaged 8.6- fold greater than the number of cells derived by continuous stimulation with zoledronate (10 µm) (Fig. 5B, p < 0.01). Increases in the number of Vγ2Vδ2 T cells were also noted 42

62 when PBMC were cultured in 24-well plates with an average of 1.9-fold greater number of cells (Fig. 5C, p < 0.01). All donors exhibited increases in purity (mean Vγ2Vδ2 T cells were 58.8% for continuous stimulation versus 80.4% for pulse stimulation, Fig. 5D, left panel) and in Vγ2Vδ2 T cell numbers (mean cell numbers were cells for continuous stimulation versus cells for pulse stimulation, Fig. 5D, right panel). Increases in cell numbers between donors ranged from 1.4- to 5.4-fold. Thus, exposure of PBMC to zoledronate for a short 4 h period resulted in increased purity and yield of Vγ2Vδ2 T cells compared with continuous zoledronate exposure. Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation exhibit increased degranulation and perforin expression compared with those expanded by continuous stimulation when stimulated by pamidronate-treated PC-3 cancer cells or Daudi Burkitt's lymphoma cells To assess the function of expanded Vγ2Vδ2 T cells derived using pulse stimulation, their surface expression of CD107a and intracellular expression of cytotoxic proteins and cytokines were measured in response to TCR stimuli. CD107a (LAMP-1) is a lysosomal membrane protein that is mobilized to the surface when cytotoxic T and NK cells degranulate for killing (53, 130, 151). When cultured with pamidronate-treated PC-3 prostate cancer cells that have high intracellular levels of IPP, a higher proportion of Vγ2Vδ2 T cells expanded by pulse stimulation expressed CD107a compared with Vγ2Vδ2 cells expanded by continuous stimulation (50.9% versus 38.2% for purified Vγ2Vδ2 T cells, Fig. 6A, p < 0.05). Similar results were noted with stimulatory Daudi lymphoma cells with a higher proportion of Vγ2Vδ2 T cells expanded by pulse stimulation expressing CD107a than Vγ2Vδ2 cells expanded by continuous stimulation (31.9% versus 19.5%, Fig. 7A, B, p = ) Consistent with their high level of CD107a, Vγ2Vδ2 T cells expanded by pulse stimulation had significantly higher proportions expressing perforin compared with 43

63 Vγ2Vδ2 T cells derived with continuous stimulation. Other cytotoxic proteins and cytokines were expressed at similar levels with Vγ2Vδ2 T cells from both conditions producing high levels of IFN-γ and granzyme B. There were low proportions of Vγ2Vδ2 T cells producing IL-17A (as observed previously in Ref. 21) and IL-21 (Fig. 6B-F). IL- 10 was not detected in expanded Vγ2Vδ2 T cells (data not shown). Activation by the mitogen, ionomycin, with PMA, stimulated IFN-γ and granzyme B to similar very high proportions of Vγ2Vδ2 T cells expanded using either condition (Fig. 7C). Thus, a significantly higher proportion of Vγ2Vδ2 T cells degranulated in response to pamidronate-treated PC-3 prostate cancer cells and Daudi lymphoma cells when derived using pulse stimulation compared with those derived by continuous stimulation. Cytotoxicity of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation against pamidronate-treated PC-3 prostate cancer cells Cytotoxicity of Vγ2Vδ2 T cells for tumor cells was assessed by fluorometric measurement of tumor cell lysis. In this assay, tumor cells are stained with fluorescent dyes to allow their identification and to evaluate their viability after incubation with cytotoxic T or NK cells (131). For our experiments, purified Vγ2Vδ2 T cells expanded using either pulse or continuous zoledronate stimulation were mixed at different E:T ratios (1:1 to 100:1) with pamidronate-treated PC-3 tumor cells. Specific killing of PC-3 cells was then assessed by flow cytometric analysis. Vγ2Vδ2 T cells expanded by pulse stimulation killed PC-3 prostate cancer cells more efficiently than those expanded by continuous zoledronate stimulation. Vγ2Vδ2 T cells expanded by pulse stimulation had LU compared with 61.3 LU (2.5-fold lower) for Vγ2Vδ2 T cells expanded by continuous stimulation (Fig. 8A, B). Thus, consistent with the CD107a expression results, Vγ2Vδ2 T cells expanded by pulse stimulation exhibited higher levels of cytotoxicity for tumor cells than those expanded using continuous zoledronate stimulation. 44

64 Adoptive transfer of Vγ2Vδ2 T cells expanded using pulse zoledronate stimulation controlled PC-3 tumor growth in NSG mice To assess the in vivo anti-tumor activity of Vγ2Vδ2 T cells expanded under the two conditions, a humanized immunodeficient mouse model developed by the Scotet laboratory was used (94). In this model, human PC-3 prostate cancer cells are xenotransplanted into the flanks of highly immunodeficient NSG mice followed 2 weeks later by 4 weekly treatment with the aminobisphosphonate, pamidronate, and adoptively transferred Vγ2Vδ2 T cells (schema shown in Fig. 9A). The only modifications to their protocol that were made were that mice were treated with pamidronate on day 13 instead of day 14 and that freshly purified Vγ2Vδ2 T cells (>98% purity, Fig. 4) from 14 day cultures were used for each adoptive transfer rather than unpurified Vγ2Vδ2 T cells from cultures of varying ages. The use of purified Vγ2Vδ2 T cells allowed us to directly compare the anti-tumor activity of Vγ2Vδ2 T cells expanded under the two conditions using identical numbers of transferred T cells. Vγ2Vδ2 T cells were expanded from frozen leukopak PBMC derived from a single donor to more closely approximate present clinical treatment protocols. Adoptive transfer of Vγ2Vδ2 T cells expanded under either condition greatly slowed tumor growth with large decreases in tumor volume and tumor diameter compared with pamidronate treatment alone (Fig. 9B and Fig. 10). These results are identical to those reported by Santolaria et al. (94). Pamidronate treatment alone had no effect on tumor growth whereas mice treated with pamidronate followed by Vγ2Vδ2 T cells derived using continuous zoledronate stimulation slowed tumor growth with only slight increases in tumor diameter (Fig. 10). In contrast, tumor growth stopped in mice treated with Vγ2Vδ2 T cells expanded using pulse zoledronate stimulation, resulting in significantly lower tumor volumes (Fig. 9B, p < 0.01) and tumor diameters (Fig. 10B, p < 0.01) compared with treatment with Vγ2Vδ2 T cells expanded using continuous zoledronate stimulation. Tumor volume was reduced by 42% at week 7 (Fig. 9B, right 45

65 panel) and averaged 46% reductions for weeks 5-7. Thus, while adoptive transfer of Vγ2Vδ2 T cells expanded under either condition greatly slowed tumor growth, tumors in mice treated with Vγ2Vδ2 T cells expanded using pulse stimulation were about half the size of tumors in mice treated with Vγ2Vδ2 T cells expanded using continuous stimulation. This better tumor control is consistent with the increased cytotoxicity of these cells in vitro. Pulse zoledronate stimulation improves the purity and yield of Vγ2Vδ2 T cells cultured with IL-15 IL-15 is a cytokine supporting T cell growth that is a member of the common γ C cytokine family that also includes IL-2, IL-7, and IL-21. Tumor-reactive murine CD8 αβ T cells expanded ex vivo using IL-15 (152, 153) or IL-7/IL-15 (154) mediate increased tumor immunity in vivo as compared to cells grown in IL-2. However, comparable studies have not been reported for Vγ2Vδ2 T cells. Therefore, we assessed whether pulse zoledronate stimulation would also improve expansion of Vγ2Vδ2 T cells when IL-15 was used. Similar to stimulation with IL-2, Vγ2Vδ2 T cells expanded using continuous zoledronate stimulation in 96-well plates showed maximal expansion between 1-10 µm with higher concentrations greatly decreasing expansion (Fig. 11A, right panels). In contrast, expansion of Vγ2Vδ2 T cells using pulse zoledronate stimulation showed expansion over a broader range of up to 100-fold (Donors 4 and 8). Maximal purity levels were similar between the various conditions and to levels observed with HMBPP stimulation (except for Donor 9) (Fig. 11A, left panels). Both the purity and number of Vγ2Vδ2 T cells was higher with pulse zoledronate stimulation compared with continuous stimulation with levels similar to those observed with HMBPP (Fig. 11B). All donors except one, exhibited increases in purity (Fig. 11C, left panel) and in Vγ2Vδ2 T cell numbers (Fig. 11C, right panel) with pulse zoledronate 46

66 stimulation. Increases in cell numbers ranged from 1.1- to 4.8-fold, averaging 2.4-fold for the 8 donors. These findings were consistent with those observed with Vγ2Vδ2 T cell expansion with IL-2. IL-15 preserves higher proportions of early/central memory subsets of Vγ2Vδ2 T cells Previous studies with αβ T cells have shown that IL-15 specifically sustains the early/central memory subsets of CD8 αβ T cells (157, 158) and in combination with IL-7 maintain stem αβ T cells (159). Early/central memory CD8 αβ T cells confer stronger anti-tumor immunity upon adoptive immunotherapy compared with late/effector T cells in mouse models (155, 160). To determine the effect of IL-15 on Vγ2Vδ2 T cells, we compared the proportion of memory subsets in the presence of either IL-2 or IL-15 following zoledronate or HMBPP stimulation. CD28 and CD27 were used as phenotypic markers for different Vγ2Vδ2 T cell memory subsets because these receptors distinguish early/central memory (CD28 + CD27 +/ ) from intermediate (CD28 CD27 + ) or late/effector memory (CD28 CD27 ) cells (data not shown and 161). The gating strategy is shown for Vγ2Vδ2 T cells from PBMC (Fig. 12A). Note that in adults, almost all Vγ2Vδ2 T cells are memory cells with the naive population being very small (< 2%). When Vγ2Vδ2 T cells were expanded by zoledronate or by HMBPP stimulation in the presence of IL-15, the majority of Vγ2Vδ2 T cells were early/central memory cells (colored green (CD28 + CD27 + ) or yellow (CD28 + CD27 ) in Fig. 12B, right panels) whereas the majority of Vγ2Vδ2 T cells expanded in the presence of IL-2 were late/effector memory cells (colored red (CD28 CD27 ) or blue (CD28 CD27 + ) in Fig. 12B, left panels). There were significantly more early/central memory T cells when Vγ2Vδ2 T cells were expanded by pulse zoledronate stimulation in IL-15 and more late/effector memory T cells with pulse zoledronate stimulation in IL-2 (Fig. 12C, right 47

67 panel). Similar results were noted with continuous stimulation (Fig. 12C, left panel). Therefore, IL-15 maintained a higher proportion of early/central memory subsets of Vγ2Vδ2 T cells than IL-2. Vγ2Vδ2 T cells expanded using pulse zoledronate stimulation with IL-15 express CD107a, cytotoxic proteins, and cytokines in similar proportions to those expanded with IL-2 To assess the effect of IL-15 on Vγ2Vδ2 T cell function, the expression of CD107a, cytotoxic proteins, and cytokines were measured in Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-15 in comparison to Vγ2Vδ2 T cells expanded with IL-2. PC-3 cancer cells were treated with pamidronate to increase intracellular IPP levels. The treated PC-3 cells were then used to stimulate purified Vγ2Vδ2 T cells that were derived using pulse zoledronate stimulation with either IL-2 or IL-15. CD107a, granzyme B, perforin, IFN-γ, TNF-α, and IL-4 expression was then assessed by flow cytometry. No significant difference were noted in the proportion of Vγ2Vδ2 T cells expressing CD107a, granzyme B, perforin, IFN-γ, and TNF-α between Vγ2Vδ2 T cells expanded with IL-15 to those expanded with IL-2 (Fig. 13). There were somewhat higher proportions of Vγ2Vδ2 T cells producing IL-4 when expanded with IL-15 but these cells constituted a small population (<6%) and are of uncertain significance in vivo. Thus, Vγ2Vδ2 T cells expanded with IL-15 have similar functional capabilities to those expanded with IL-2. Adoptive transfer of Vγ2Vδ2 T cells expanded using pulse stimulation with IL-15 controlled PC-3 tumor growth in NSG mice similarly to Vγ2Vδ2 T cells expanded with IL-2 To assess the effect of IL-15 on anti-tumor activity, Vγ2Vδ2 T cells expanded 48

68 using pulse zoledronate stimulation with either IL-15 or IL-2 were adoptively transferred into NSG mice bearing PC-3 prostate tumors (Fig. 14A). Both Vγ2Vδ2 T cells expanded with IL-15 or with IL-2 efficiently controlled PC-3 tumors stopping their growth but not eliminating the tumors. There were no significant differences in tumor diameter (Fig. 15B) or tumor volume (Fig. 14B) when Vγ2Vδ2 T cells expanded with IL-15 (open circles) were compared to those expanded with IL-2 (closed circles). However, two weeks after adoptive transfer of Vγ2Vδ2 T cells had stopped (week 8), tumor growth resumed in both groups without significant differences albeit at a slower rate of growth than PC-3 tumors treated with pamidronate alone. Therefore, unlike murine studies, expansion of Vγ2Vδ2 T cells in IL-15 did not improve immunity against prostate tumors in our humanized mouse model. This result is consistent with the absence of significant differences in the in vitro function of Vγ2Vδ2 T cells expanded by IL-15 compared to those expanded by IL-2. Vγ2Vδ2 T cells stimulated by pulse zoledronate expand to similar levels using OpTmizer TM media manufactured under cgmp and with large scale cultures To determine if pulse zoledronate stimulation could improve current practices in clinical trials, we examined whether similar results could be obtained using a commercial media used for T cell expansions that is produced under current good manufacturing practices (cgmp) and with larger scale expansions. To meet regulatory requirements, culture media used for clinical trials must meet cgmp standards. Therefore, we compared the enriched RPMI 1640 media used in our experiments (termed C-media) with OpTmizer TM media, a media meeting cgmp standards. Expansion of Vγ2Vδ2 T cells in OpTmizer TM media (open bars) resulted in comparable levels of Vγ2Vδ2 T cell purity (Fig. 16A) and numbers (Fig. 16B) to those achieved using C-media (solid bars). Consistent with experiments using C-media, expansion of Vγ2Vδ2 T cells in 49

69 OpTmizer TM media using pulse zoledronate stimulation resulted in higher purity (71% vs. 42% Vγ2Vδ2 T cells) and cell numbers ( versus Vγ2Vδ2 T cells, 2.2- fold higher) compared with expansion using continuous stimulation (Fig. 16A, B). Similar increases in cell numbers were noted with larger scale expansions. Pulse zoledronate stimulation of thawed leukopak PBMC in flasks resulted in an 858-fold increase in Vγ2Vδ2 T cell numbers by day 14 ( cells increasing to cells, Fig. 16C). This expansion was 1.9-fold higher than that achieved by continuous zoledronate stimulation. Consistent with our experiments in 24-well plates, expansion of Vγ2Vδ2 T cells using pulse zoledronate stimulation with IL-2 in flasks gave similar cell numbers when IL-15 was used (774-fold with IL-2 versus 738-fold with IL-15, Fig. 16D). These results demonstrate that pulse zoledronate stimulation resulted in increased Vγ2Vδ2 T cell purity and numbers with cgmp media and larger scale expansions. Similar benefits are likely, not only for preclinical in vitro studies and humanized mouse models, but also for clinical trials employing adoptive transfer of Vγ2Vδ2 T cells. 50

70 Figure 4. Purity of Vγ2Vδ2 T cells after positive magnetic bead separation. Vγ2Vδ2 T cells expanded by either continuous or pulse zoledronate stimulation for 14 d and then purified by magnetic bead separation. Expanded Vγ2Vδ2 T cells (top panels) were washed, resuspended to cell/ml, and reacted with APCconjugated anti-human Vδ2 mab (clone B6) for 20 min. Cells were then washed twice with purification buffer and anti-apc magnetic beads were added. After incubation on ice for 15 min, the cells were washed twice, resuspended in 500 µl of purification buffer, and loaded onto LS columns for positive selection. Purified Vδ2 T cells were removed from the magnet, eluted, washed, and then counted for use for adoptive transfer or other functional assays. Cell purity was evaluated by flow cytometry (bottom panels) and was at least > 95% for use. 51

71 52

72 Figure 5. Expanding Vγ2Vδ2 T cells by pulse zoledronate stimulation with IL-2 increased their purity and number compared with continuous stimulation. a Expansion of Vγ2Vδ2 T cells in response to varying doses of HMBPP and zoledronate with IL-2. Human Vγ2Vδ2 T cells were expanded ex vivo from PBMC by exposure to HMBPP (left panels) or zoledronate (right panels). Human PBMC were cultured in 96-well plates either continuously with varying starting concentrations of HMBPP or zoledronate (open circles) or pulsed with zoledronate for 4 h (closed circles) followed by washing twice. IL-2 was added to 1000 IU on day 3. Thereafter, media was changed every 2-3 d depending on cell growth. On day 14, Vγ2Vδ2 T cell numbers were determined by flow cytometric analysis. b, c Numbers of Vδ2 T cells after 14 d expansion under the indicated conditions with IL-2 for triplicate cultures in either 96-well plates (b) (results for eight donors) or 24-well plates (c) (results for nine donors). PBMC were cultured for 14 d in the presence of IL-2 at cells/well for 96-well plates or cells/well for 24-well plates. **p < 0.01 compared to 10 µm zoledronate continuous stimulation using the Mann-Whitney U test. Mean and standard deviation is shown. d Comparison of the purity (left panel) and number (right panel) of Vδ2 T cells expanded by continuous (10 µm) or pulse (100 µm) stimulation with zoledronate with IL-2 in 24-well plates. Each line connects the results for an individual donor. Results for nine donors from (c). 53

73 54

74 Figure 6. Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-2 exhibit slightly increased degranulation and expression of granzyme B and perforin compared with those expanded by continuous stimulation when exposed to pamidronate-treated PC-3 cancer cells. Human PBMCs were cultured either continuously with zoledronate (5 µm) or pulsed with zoledronate (100 µm) for 4 h and then washed twice before culture with IL-2 (1000 IU/ml). After 14 d, expanded Vγ2Vδ2 T cells were purified by positive selection or left unpurified. PC-3 cancer cells were treated overnight with pamidronate (200 µm) and then washed. Pamidronate-treated PC-3 cells were incubated with unpurified or purified Vγ2Vδ2 T cells for 4 h in duplicate samples followed by surface and intracellular mab staining. CD107a, IFN-γ, granzyme B, perforin, IL-17A, and IL-21 levels on or in Vγ2Vδ2 T cells were assessed by flow cytometric analysis. Mean ± SD is shown. Representative of two experiments. *p < 0.05 compared to 5 µm zoledronate continuous stimulation using the unpaired t-test. 55

75 Vδ2 CD107a IFN-γ Granzyme B Perforin Vδ Pulse Continuous CD3 Vδ2 56

76 Figure 7. Expansion of Vγ2Vδ2 T cells by pulse zoledronate stimulation increased degranulation as measured by the expression of CD107a in response to the stimulatory Burkitt's lymphoma cell line, Daudi. a Increased expression of CD107a by Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation when exposed to the stimulatory Daudi cell line. Vγ2Vδ2 T cells were expanded from PBMCs by either continuous or pulse zoledronate stimulation for 14 d. Vγ2Vδ2 T cell numbers were assessed by flow cytometric analysis. Unpurified Vγ2Vδ2 T cells were then mixed with Daudi or Raji Burkitt's lymphoma cells at 10:1 (E:T) ratio for 4 h in the presence of monesin and PE-Cy7-anti-CD107a. The cultures were then washed and stained with PE-anti-CD3 and FITC-anti-Vδ2 mabs and analyzed by flow cytometry. b Mean % of Vδ2 T cells expressing CD107 after culturing with Daudi cells (n = 3 experiments). **p = using the unpaired t-test. c Similar levels of IFN-γ and granzyme B in Vγ2Vδ2 T cells expanded either by continuous or pulse zoledronate stimulation or by culturing in IL-2 after ionomicin/pma stimulation for 4 h. Levels were determined by intracellular staining and flow cytometric analysis. 57

77 58

78 Figure 8. Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation exhibit similar cytotoxicity against pamidronate-treated PC-3 cancer cells compared with Vγ2Vδ2 T cells expanded by continuous stimulation. Vγ2Vδ2 T cells were expanded as described in Fig. 6. Vγ2Vδ2 T cells expanded by continuous or by pulse zoledronate stimulation were then purified for use. PC-3 cancer cells were treated overnight with pamidronate (200 µm) and then washed twice. To assess cytotoxicity, the FATAL assay was used. Pamidronate-treated PC-3 cells were stained with PHK-26 and CFSE and incubated with purified Vγ2Vδ2 T cells for 5 h at various E:T ratios in duplicate. Cells were then fixed with 1% paraformaldehyde followed by flow cytometric analysis. a CFSE levels of PC-3 prostate cancer cells incubated with varying numbers of purified Vγ2Vδ2 T cells expanded by continuous or pulse zoledronate stimulation. b Cytolytic activity of Vγ2Vδ2 T cells expanded either by continuous or pulse zoledronate stimulation against pamidronate-treated PC-3 cancer cells as determined by the FATAL assay. Representative of two experiments. 59

79 60

80 Figure 9. Adoptive transfer of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation in combination with pamidronate controlled PC-3 prostate tumor growth in NSG mice. a Schema of treatment protocol used to evaluate the anti-tumor efficacy of Vγ2Vδ2 T cells. Immunodeficient NSG mice were s.c. inoculated with human PC-3 prostate cancer cells on day 0. On day 13, pamidronate (50 µg/kg) was given i.v. On day 14, purified Vγ2Vδ2 T cells expanded using either continuous or pulse zoledronate stimulation were inoculated i.v. Treatments were repeated weekly until week 6. Logitudinal and transverse diameters of the tumors were measured weekly. b, left panel Vγ2Vδ2 T cells stimulated by pulse zoledronate exposure exhibit significantly better anti-tumor immunity compared with those expanded by continuous zoledronate exposure. Mean PC-3 tumor volume ± SD is shown for 7-8 mice per group treated with either pamidronate alone (open triangles), pamidronate with purified Vγ2Vδ2 T cells derived by continuous zoledronate stimulation (open circles), or pamidronate with purified Vγ2Vδ2 T cells derived by pulse zoledronate stimulation (closed circles). **p < 0.01, ***p < compared with tumor volume of mice treated with Vγ2Vδ2 T cells derived by pulse zoledronate stimulation using the Mann-Whitney U test. Right panel, Tumor volume at week 7 of individual mice treated with pamidronate alone (open triangles), pamidronate with Vγ2Vδ2 T cells derived by continuous zoledronate stimulation (open circles), or pamidronate with Vγ2Vδ2 T cells derived by pulse zoledronate stimulation (closed circles). Bars represent mean values. **p < 0.01, ***p < using the Mann-Whitney U test. 61

81 62

82 Figure 10. Adoptive transfer of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation in combination with pamidronate controlled PC-3 prostate tumor growth in NSG mice; tumor diameter data for Fig. 9. a Schema of treatment protocol used to evaluate the anti-tumor efficacy of Vγ2Vδ2 T cells. Immunodeficient NSG mice were s.c. inoculated with human PC-3 prostate cancer cells on day 0. On day 13, pamidronate (50 µg/kg) was given i.v. On day 14, purified Vγ2Vδ2 T cells expanded either by continuous or by pulse zoledronate stimulation were inoculated i.v. Treatments were repeated weekly until week 6. Logitudinal and transverse diameters of the tumors were measured weekly. b, left panel Vγ2Vδ2 T cells stimulated by pulse zoledronate exposure exhibit significantly better anti-tumor immunity compared with those expanded by continuous zoledronate exposure. Mean PC-3 tumor diameter ± SD is shown for 7-8 mice per group treated with either pamidronate alone (open triangles), pamidronate with purified Vγ2Vδ2 T cells derived by continuous zoledronate stimulation (open circles), or pamidronate with purified Vγ2Vδ2 T cells derived by pulse zoledronate stimulation (closed circles). **p < 0.01, ***p < compared with tumor volume of mice treated with Vγ2Vδ2 T cells derived by pulse zoledronate stimulation using the Mann-Whitney U test. Right panel, Tumor volume at week 7 of individual mice treated with pamidronate alone (open triangles), pamidronate with Vγ2Vδ2 T cells derived by continuous zoledronate stimulation (open circles), or pamidronate with Vγ2Vδ2 T cells derived by pulse zoledronate stimulation (closed circles). Bars represent mean values. **p < 0.01, ***p < using the Mann-Whitney U test. 63

83 64

84 Figure 11. Expanding Vγ2Vδ2 T cells by pulse zoledronate stimulation with IL-15 increased their purity and numbers compared with continuous stimulation. a Expansion of Vγ2Vδ2 T cells in response to varying doses of HMBPP and zoledronate with IL-15. Human Vγ2Vδ2 T cells were expanded ex vivo from PBMC by exposure to HMBPP (left panels) or zoledronate (right panels). Human PBMC were cultured in 96-well plates either continuously with varying starting concentrations of HMBPP or zoledronate (open circles) or pulsed with varying concentrations of zoledronate for 4 h (closed circles) followed by washing twice. IL-15 was added to 100 ng/ml on day 3. Thereafter, media was changed every 2-3 d depending on cell growth. On day 14, Vγ2Vδ2 T cell numbers were determined by flow cytometric analysis. Results for 8 donors are shown. b, upper panel Purity of Vδ2 T cells under the indicated conditions with IL-15 (100 ng/ml) after 14-day culture in 24-well plates. Each point represents one donor (n = 8). ** p = 0.003, Mann-Whitney U test. Lower panel, Number of Vδ2 T cells under the indicated conditions with IL-15 (100 ng/ml) after 14- day culture in 24-well plates. Each point represents one donor (n = 8). ** p = 0.007, Mann-Whitney U test. c Comparison of the purity (left panel) and number (right panel) of Vδ2 T cells expanded by continuous (10 µm) or pulse (100 µm) stimulation with zoledronate with IL-15 in 24-well plates. Each line connects the results for an individual donor from (b). 65

85 66

86 Figure 12. IL-15 sustained higher proportions of early/central memory subsets among expanded Vγ2Vδ2 T cells compared to IL-2. a Gating strategy for delineating memory subsets of Vγ2Vδ2 T cells from blood. Memory status of Vγ2Vδ2 T cells was assessed by expression of CD27 and CD28. CD28 + /CD27 + and CD28 + /CD27 - cells are early/central memory cells that express CD45RO. CD28 - /CD27 + cells are intermediate memory cells. CD28 - /CD27 - cells are late memory cells that express CD45RA. b Proportions of human Vγ2Vδ2 T cell memory subsets after expansion with the indicated concentrations of HMBPP or zoledronate by continuous and pulse stimulation and in the presence of IL-2 (left panels) or IL-15 (right panels) for 14 days. Data for four donors is shown. c, left panel Comparison of the proportion of memory subsets of Vγ2Vδ2 T cells expanded by continuous zoledronate stimulation (10 µm) with IL-2 or IL-15. Mean ± SEM for 4 individuals. Right panel, Comparison of the proportion of memory subsets of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation (100 µm) with IL-2 or IL-15. Mean ± SEM for 4 individuals. ** p < 0.01 using the unpaired t-test. 67

87 68

88 Figure 13. Functional capabilities of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-15 are similar to those expanded with IL-2. Human PBMCs were pulsed with zoledronate (100 µm) for 4 h and then washed twice before re-culture either with IL-15 (50 ng/ml) or IL-2 (1000 IU/ml) for 14 d. Expanded Vγ2Vδ2 T cells were purified by positive selection or left unpurified. PC-3 cells were treated overnight with pamidronate (200 µm) and then washed. Pamidronate-treated PC-3 cells were incubated with unpurified or purified Vγ2Vδ2 T cells for 4 h in duplicate samples with monensin followed by surface and intracellular mab staining. Staining was assessed by flow cytometric analysis. Mean ± SD is shown. Representative of two experiments. *p < 0.05 using the unpaired t-test. 69

89 70

90 Figure 14. Adoptive transfer of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-15 in combination with pamidronate controlled PC-3 prostate tumor growth in NSG mice similarly to Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-2. a Schema of treatment protocol used to evaluate the anti-tumor efficacy of Vγ2Vδ2 T cells. Immunodeficient NSG mice were s.c. inoculated with human PC-3 prostate cancer cells on day 0. On day 13, pamidronate (50 µg/kg) was given i.v. On day 14, purified Vγ2Vδ2 T cells expanded using pulse zoledronate stimulation with either IL-15 or IL-2 were inoculated i.v. Treatments were repeated weekly until week 6. Logitudinal and transverse diameters of the tumors were measured once weekly until week 9. b, left panel Vγ2Vδ2 T cells stimulated by pulse zoledronate exposure with IL-15 showed similar anti-tumor immunity compared with those expanded with IL-2. Mean PC-3 tumor volume ± SD is shown for 7-8 mice per group treated with either pamidronate alone (open triangles), pamidronate with purified Vγ2Vδ2 T cells derived by pulse zoledronate stimulation with IL-15 (open circles), or pamidronate with purified Vγ2Vδ2 T cells derived by pulse zoledronate stimulation with IL-2 (closed circles). ***p < compared with mean tumor volume of mice treated with Vγ2Vδ2 T cells derived by pulse zoledronate stimulation with IL-2 using the Mann-Whitney U test. Right panel, Tumor volume at week 7 of individual mice treated with pamidronate alone (open triangles), pamidronate with Vγ2Vδ2 T cells derived by pulse zoledronate stimulation with IL-15 (open circles), or pamidronate with Vγ2Vδ2 T cells derived by pulse zoledronate stimulation with IL-2 (closed circles). Bars represent mean values. ***p < using the Mann-Whitney U test. 71

91 72

92 Figure 15. Adoptive transfer of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-15 in combination with pamidronate controlled PC-3 prostate tumor growth in NSG mice similarly to Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with IL-2; tumor diameter data for Fig. 14. a Schema of treatment protocol used to evaluate the anti-tumor efficacy of Vγ2Vδ2 T cells. Immunodeficient NSG mice were s.c. inoculated with human PC-3 prostate cancer cells on day 0. On day 13, pamidronate (50 µg/kg) was given i.v. On day 14, purified Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation with either IL-15 or IL-2 were inoculated i.v. Treatments were repeated weekly until week 6. Logitudinal and transverse diameters of the tumors were measured twice weekly until week 9. b, left panel Vγ2Vδ2 T cells stimulated by pulse zoledronate exposure with IL-15 showed similar anti-tumor immunity compared with those expanded with IL-2. Mean PC-3 tumor diameter ± SD is shown for 7-8 mice per group treated with either pamidronate alone (open triangles), pamidronate with purified Vγ2Vδ2 T cells derived by pulse zoledronate stimulation with IL-15 (open circles), or pamidronate with purified Vγ2Vδ2 T cells derived by pulse zoledronate stimulation with IL-2 (closed circles). ***p < compared with mean tumor volume of mice treated with Vγ2Vδ2 T cells derived by pulse zoledronate stimulation with IL-2 using the Mann-Whitney U test. Right panel, Tumor volume at week 7 of individual mice treated with pamidronate alone (open triangles), pamidronate with Vγ2Vδ2 T cells derived by pulse zoledronate stimulation with IL-15 (open circles), or pamidronate with Vγ2Vδ2 T cells derived by pulse zoledronate stimulation with IL-2 (closed circles). Bars represent mean values. ***p < using the Mann-Whitney U test. 73

93 74

94 Figure 16 Vγ2Vδ2 T cells stimulated by pulse zoledronate expand to similar levels using commercial cgmp grade OpTmizer TM media and with larger-scale cultures. a, b Similar purity (a) and yield (b) of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation in cgmp grade OpTimizer TM culture media as compared with supplemented RPMI 1640 C-media. Vγ2Vδ2 T cells from frozen leukopak PBMC were expanded by pulse zoledronate stimulation with IL-2 for 14 d in either OpTimizer TM culture media or the supplemented RPMI 1640 C-media used in this study. Vγ2Vδ2 T cell numbers were determined by flow cytometric analysis. Representative of two experiments. c, d Expansion of Vγ2Vδ2 T cells by pulse zoledronate culture under larger-scale cultures yielded similar enhancements in yield as noted with small-scale cultures. For largerscale expansion of Vγ2Vδ2 T cells, leucopak PBMCs were stimulated by either continuous (5 µm) or pulse zoledronate (100 µm) exposure (c) as detailed in Fig. 1 with IL-2 in C-Media in 75 cm 2 cell culture flasks in triplicate cultures. For (d), Vγ2Vδ2 T cells were expanded by pulse zoledronate stimulation with either IL-2 or IL-15 in C-Media in 75 cm 2 cell culture flasks in triplicate cultures. After 14 days, Vγ2Vδ2 T cell numbers were determined by flow cytometric analysis. Starting number of Vγ2Vδ2 T cells was for each group. Mean ± SD is shown. **p < 0.01, ***p < 0.001, ****p < using the unpaired t-test. 75

95 76

96 B- The Role of IL-18 on the Expansion, Memory Subset Distribution, and Function of Vγ2Vδ2 T Cells Rational IL-18 is an proinflamtory cytokine that was first discovered in 1995 as an IFN-γ inducing factor (162). IL-18 belongs to IL-1 family cytokines and shares some structural and functional similarities. IL-18 encoding genes are located on chromosome 11 in human. The production of active IL-18 is caspase dependent. IL-18 is synthesized as a 23 kda inactive precursor that is activated by caspase-1 cleavage (163). Upon stimulation, IL-18 can be produced by several types of cells such as dendritic cells, macrophages, osteoblasts, Kupffer cells, synovial fibroblasts, keratinocytes, and chondrocytes (164, 165). IL-18 binds to IL-18R, which is a heterodimeric receptor that consists of IL-18Rα (IL-18 binding receptor) and IL-18Rβ (the signaling peptide). IL-18Rα is expressed on naïve and mature T cells and NK cells. In addition, IL-18Rα is expressed on macrophages, neutrophils, B cells, basophils, mast cells, and endothelial cells (164, 166). Upon binding to its receptor, IL-18 induces the secretion of IFN-γ and other inflammatory cytokines through MyD88-NF-kB pathway. Studies have shown that other cytokines, particularly IL-2, IL-15, and IL-12, influence the expression of IL-18R and modulate its function (164, 167). IL-18 plays an important role in both innate and adoptive immunity, mainly thorough its effect on IFN-γ production. Several TLRs agonists, especially agonists for TLR-2 or TLR-4, stimulate IL-18 production (164, 168). Also, in some cases, NLR activation induces IL-18 production. Besides inducing IFN-γ production, IL-18 also promotes the expression of GM-CSF, TNF-α, IL-1β, and IL-8 (169). IL-18, alone or in combination with other cytokines, influences the differentiation and the activity of different immune cells. Along with IL-12, IL-18 plays a critical role in polarizing CD4 T cells towards Th1 cells. IL-18 can also increase the secretion of Th2 cytokines (170). IL- 77

97 18 along with IL-15 can increase NK activity. More recent studies have shown that IL- 18, in combination with IL-2, promotes the expansion of Vγ2Vδ2 T cells (22, 23, 171). IL-18 also enhances the cytotoxicity of CD8 αβ T cells (164, 172). During cancer development, IL-18 plays both positive and negative roles. IL-18 can promote tumor progression or metastasis, mainly through its effect on VCAM-1 and VEGF production (173, 174). In contrast, IL-18 can also enhance immunity to several kinds of tumors including melanoma and liver cancer (164, 175, 176). Due to its impact on IFN-γ production, IL-18 enhances anti-tumor activity by NK, Th1, and Vγ2Vδ2 T cells (22, 164). We have shown in previous sections that pulse stimulation with IL-2 or IL-15 can greatly enhance the expansion, and anti-tumor activity of Vγ2Vδ2 T cells. Given the importance of IL-18 on the expansion, differentiation, and function of NK and Th1 cells, we hypothesized that expansion and the function of Vγ2Vδ2 T cells can be further optimized using IL-18 in combination with IL-2 or IL-15 cytokines. Results IL-18 with IL-2 or IL-15 supports the expansion of Vγ2Vδ2 T cells with zoledronate pulse stimulation Previous studies have shown that IL-18 increases the expansion of CD4 αβ, Th1, CD8 αβ, and NK cells. However, the role of IL-18 on expansion of human Vγ2Vδ2 T cells has not been fully explored. Therefore, we hypothesized that IL-18 could increase the expansion of Vγ2Vδ2 T cells. To assess the role of IL-18 on the expansion of Vγ2Vδ2 T cells, PBMCs were pulsed with zoledronate and then cultured with IL-18 alone or in combination with IL-2 or IL-15. After 14 days of culture, the purity and yield of Vγ2Vδ2 T cells was assessed by flow cytometric analysis. IL-18 alone does not support the expansion of Vγ2Vδ2 T cells (Fig. 17). The percentage of Vγ2Vδ2 T cells with IL-18 alone was similar to Vγ2Vδ2 T cells expanded with no cytokines (25% vs. 78

98 20%) indicating that IL-18 alone has no effect on Vγ2Vδ2 T cells expansion. When added to IL-2 or IL-15, IL-18 did not increase the purity of Vγ2Vδ2 T cells which is near maximum (> 90%) with either IL-2 or IL-15 alone. However, IL-18 does enhance the expansion of Vγ2Vδ2 T cells as determined by their absolute numbers. IL-18 significantly increased the number of Vγ2Vδ2 T cells when added to IL-2 or IL-15. As expected, IL-18 alone did not affect the yield of Vγ2Vδ2 T cells (Fig. 17). Taken together, in combination with IL-2 or IL-15, IL-18 can further enhance the expansion of Vγ2Vδ2 T cells through zoledronate pulse stimulation. IL-18 helps IL-2 to preserve more central/early memory subset of Vγ2Vδ2 T cells and helps IL-15 to maintains more late memory subset Studies with CD8 T cells have reported that central memory subsets provide better immunity against cancer (38). As our previous results showed, IL-15 significantly increases the central early memory subsets of Vγ2Vδ2 T cells whereas IL-2 maintains high proportion of late memory cells. In our effort to enhance the quality of memory subsets of Vγ2Vδ2 T cells, we sought to investigate the role of IL-18 on the distribution of memory subsets of Vγ2Vδ2 T cells, as the role of IL-18 on memory cells of Vγ2Vδ2 T cells is not clear yet. To assess the role of IL-18 on memory status of Vγ2Vδ2 T cells, PBMCs were cultured under zoledronate pulse stimulation in the presence of IL-18 alone or in combination with IL-2 or IL-15. Using CD27 and CD28 expression, the memory subsets of Vγ2Vδ2 T cells at day 14 of culture were determined by flow cytometry. IL-18 by itself has no effect on the proportions of the Vγ2Vδ2 T cell memory subsets (Fig. 18). However, when added to IL-2 or IL-15, IL-18 does alter the proportions of the memory subsets. IL-18 plays different roles in each cases. When added to IL-2, IL-18 maintains more early/central memory cells, while when IL-18 was added to IL-15, more late 79

99 memory cells were observed (Fig. 18). These results suggest that IL-18 can alter the proportions of the different memory Vγ2Vδ2 T subsets. IL-18, along with IL-2 or 15, maintains high levels of cytokines and cytotoxic molecules under zoledronate pulse stimulation IL-18 has been known to greatly enhance the function of NK, Th1, and CD8 T cells. Also, IL-18 plays a major role in IFN-γ production by NK cells. Other cytokines and cytotoxic molecules, such as perforin, also have been shown to be regulated by IL- 18. Few studies have explored the effect of IL-18 on the function of Vγ2Vδ2 T cells, and it is still not clear whether adding IL-18 on top of IL-2 or 15 would enhance the function of Vγ2Vδ2 T cells stimulated by zoledronate pulse stimulation. To assess the role of IL- 18 on the function of Vγ2Vδ2 T cells, PBMCs were stimulated by pulse zoledronate exposure in the presence of IL-18 alone or in combination with IL-2 or IL-15. At day 14, purified Vγ2Vδ2 T cells from each condition were co-cultured overnight with pamidronate-treated PC-3 human prostate cancer cells. Then, CD107a expression, cytokine levels (IFN-γ, perforin, granzyme B, and IL-4), and cytotoxic molecule (perforin and granzyme B) production was measured. Consistent with our previous findings, IL-18 alone has no effect on CD107, cytokine or cytotoxic molecule production (Fig. 20). Also, adding IL-18 to IL-2 or IL-15 does not increase the percentage of Vγ2Vδ2 T cells expressing CD107a, IFN-γ, and TNF-α. Compared to IL-15 alone, IL-18 increased granzyme B expression slightly. Interestingly, IL-18 moderately reduced perforin expression in all conditions (Fig. 19). Also, IL-18 alone significantly increased IL-4 expression in Vγ2Vδ2 T cells. Taken together, IL-18 addition to IL-2 or IL-15 maintains Vγ2Vδ2 T cell function after co-culture with cancer cells. The already high levels of cytokines and cytotoxic molecules expressed by Vγ2Vδ2 T cell may explain the minimal impact of IL-18 on Vγ2Vδ2 T cell function. 80

100 Figure 17 IL-18, in combination of IL-2 or IL-15, maintains high purity and increases the yield of Vγ2Vδ2 T cells stimulated by pulse zoledronate stimulation. (Top panel) shows the percentage of Vγ2Vδ2 T cells at indicated cytokines. (Bottom panel) represents the absolute cell number (yield) of Vγ2Vδ2 T cells cultured with the different cytokines. For this study, leucopak Vγ2Vδ2 T cells were expanded by pulse zoledronate stimulation with either IL-2, IL-15, IL-18, IL-2 + IL-18, and IL-15 + IL-18 in C-Media in 75 cm 2 cell culture flasks in triplicate cultures. After 14 days, Vγ2Vδ2 T cell numbers were determined by flow cytometric analysis. Starting number of Vγ2Vδ2 T cells was for each group. Representative of two experiments. Mean ± SD is shown. **p < 0.01, ***p < 0.001, ****p < using the unpaired t-test. 81

101 82

102 Figure 18. IL-18 alters the proportions of the memory subsets in expanded Vγ2Vδ2 T cells when added to IL-2 or IL-15 compared to IL-2 or IL-15 alone. The proportion of each memory Vγ2Vδ2 T cell subset was assessed by expression of CD27 and CD28. CD28 + /CD27 + and CD28 + /CD27 - cells are early/central memory cells that express CD45RO. CD28 - /CD27 + cells are intermediate memory cells. CD28 - /CD27 - cells are late memory cells that express CD45RA. (Top panel) Proportions of human Vγ2Vδ2 T cell memory subsets after expansion under zoledronate pulse stimulation in the presence of IL-2, IL-15, IL-18, IL-2 + IL-18, or Il-15 + IL-18 IL-15 for 14 days. Data for one donor in triplicates is shown. (Bottom panels) Comparison of the different cytokines on the proportion of memory subsets of Vγ2Vδ2 T cells expanded by zoledronate pulse stimulation. Mean ± SEM for 4 individuals. ** p < 0.01 using the unpaired t-test. 83

103 84

104 Figure 19. IL-18 maintains similar functional capabilities of Vγ2Vδ2 T cells expanded by pulse zoledronate stimulation in combination with IL-2 or IL-15, to those expanded with IL-2 or IL-15 alone. Human PBMCs were pulsed with zoledronate (100 µm) for 4 h and then washed twice before re-culture either with IL-15 (50 ng/ml), IL-2 (1000 IU/ml), or IL-18 (20 ng/ml) for 14 d. Expanded Vγ2Vδ2 T cells were purified by positive selection or left unpurified. PC-3 cells were treated overnight with pamidronate (200 µm) and then washed. Pamidronate-treated PC-3 cells were incubated with unpurified or purified Vγ2Vδ2 T cells for 4 h in duplicate samples with monensin followed by surface and intracellular mab staining. Staining was assessed by flow cytometric analysis. Mean ± SD is shown. Representative of two experiments. *p < 0.05 using the unpaired Student's t-test. 85

105 86

106 Figure 20. Model for optimizing anti-tumor activity of Vγ2Vδ2 T cells through zoledronate pulse stimulation. Top panel During continuous stimulation with zoledronate and IL-2, monocytes (the primary APC inpbmc) take up zoledronate via pinocytosis. Zoledronate then blocks FPPS resulting in the endogenous accumulation of IPP. IPP then stimulates Vγ2Vδ2 T cells through their TCRs in a process dependent on BTN3A1. Due to the effect of IL-2, most of the Vγ2Vδ2 T cells will downregulate CD28 and CD27. Vγ2Vδ2 T cells expanded under this condition showed the ability to secrete IFN-γ, granzyme B, and perforin and to kill tumor cells as shown by their expression of CD107a upon co-incubation.these cells showed good control of PC-3 tumor growth in preclinical models. However, activated Vγ2Vδ2 T cells will also take up zoledronate due to its continued presence thereby negatively affecting thier survival and function. Bottom panel Zoledronate pulse stimulation limits zoledronate exposure to 4 h. This enhances the yield and purity of Vγ2Vδ2 T cells. These Vγ2Vδ2 T cells still produce high level of IFN-γ and granzyme B but have significantly higher levels of perforin and CD107a expression. Vγ2Vδ2 T cells from pulse stimulation show higher killing activity in vitro. Moreover, in the PC-3 preclinical model of human prostate cancer, they display significantly better tumor control on a per cell basis. Although replacing IL-2 with IL-15 resulted in more early/central memory Vγ2Vδ2 T cells, this did not improve their anti-tumor activity in vitro or in vivo in the PC-3 model. 87

107 88