A License to Kill: Understanding Natural Killer Cell Licensing to Fight Cancer

The Texas Medical Center Library DigitalCommons@TMC

The University of Texas MD Anderson Cancer Center UTHealth Graduate School of The University of Texas MD Anderson Cancer Biomedical Sciences Dissertations and Theses Center UTHealth Graduate School of (Open Access) Biomedical Sciences

12-2017

A LICENSE TO KILL: UNDERSTANDING NATURAL KILLER CELL LICENSING TO FIGHT CANCER

Jolie Schafer

Follow this and additional works at: https://digitalcommons.library.tmc.edu/utgsbs_dissertations

Part of the Medicine and Health Sciences Commons, and the Other Immunology and Infectious Disease Commons

Recommended Citation Schafer, Jolie, "A LICENSE TO KILL: UNDERSTANDING NATURAL KILLER CELL LICENSING TO FIGHT CANCER" (2017). The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences Dissertations and Theses (Open Access). 812. https://digitalcommons.library.tmc.edu/utgsbs_dissertations/812

This Dissertation (PhD) is brought to you for free and open access by the The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences at DigitalCommons@TMC. It has been accepted for inclusion in The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences Dissertations and Theses (Open Access) by an authorized administrator of DigitalCommons@TMC. For more information, please contact [email protected]. A LICENSE TO KILL: UNDERSTANDING NATURAL KILLER CELL

LICENSING TO FIGHT CANCER

Jolie Rae Schafer, B.S.

APPROVED:

______Shulin Li, Ph.D. Advisory Professor

______Dean Lee, M.D., Ph.D. Co-Mentor

______Michael Curran, Ph.D.

______Gregory Lizee, Ph.D.

______Annemieke Kavelaars, Ph.D.

______Silke Paust, Ph.D.

APPROVED:

______Dean, The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences

I A LICENSE TO KILL: UNDERSTANDING NATURAL KILLER CELL

LICENSING TO FIGHT CANCER

DISSERTATION

Presented to the Faculty of

The University of Texas

MD Anderson Cancer Center UTHealth

Graduate School of Biomedical Sciences

in Partial Fulfillment

of the Requirements

for the Degree of

DOCTOR OF PHILOSOPHY

Jolie Schafer, B.S.

Houston, Texas

December 2017

II Acknowledgements

I would like to acknowledge all of the many people who helped me and encouraged me along the way. My love for science began in 7th grade, when I learned about genetics from Mr. Flori.

Thank you Mr. Flori for instilling in me a facisination for science. My motivation for pursuing scientific research stemmed from my childhood friend, Chase McGowen, who has Cystic

Fibrosis. Chase, you are a miracle, your fighting spirit keeps me fighting to learn more. Thank you to all of my undergraduate professors, from Houston Baptist University, who encouraged me to apply to graduate school, Drs. Hannah Wingate, Susan Cook, Jackie Horn, Brenda

Whaley, Curtis Henderson, Rachel Hopp, and Saul Trevino. Thank you all for believing in me and seeing my potential. Thank you, Dr. Khandan Keyomarsi for my first laboratory position as a Cancer Prevention Research Institute of Texas undergraduate student, which turned into a yearlong research assistant position preceding my graduate work. Thank you to The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences for accepting me into this program. To Deans Barton and Blackburn, thank you for allowing me to have this opportunity.

To my amazing mentor, Dr. Dean Lee for mentoring me, teaching me, growing me, generously allowing me to keep my project upon your departure. Thank you, Dr. Lee for your support over the last four years. You are an incredible mentor and friend and I am grateful to have trained under you. Thank you, Dr. Shulin Li for graciously adopting me into your lab and letting me continue my dissertation project. Thank you to the amazing members of both Dr. Lee and Dr.

Shulin’s laboratories, your help and collaboration has been incredible. Thank you to Dr. Stefan

Ciurea for allowing me to collaborate on the NK cell clinical trial, I am excited to see whats in store for the Phase II study. Thank you to my great friend Dr. Ariany Aquino-Lopez for being III an amazing collegue and friend. Our journey through graduate school together is one that I will always cherish. Thank you to the Pediatics Research department for a great training environment throughout graduate school.

Thank you to my examining committee Drs. Steve Ullrich, Michael Curran, Neal Waxham,

Cao, Lizee, for challenging me and preparing me for finishing graduate school. Thank you to my wonderful advisory committee, Drs. Dean Lee, Shulin Li, Silke Paust, Michael Curran,

Gregory Lizee, and Annemieke Kavelaars. All of you have been a tremendous help and resource throughout graduate school. Thank you to the former Immunology Program Director,

Dr. Ben Zhu for your service to the Immunology Program and for giving me a great family of researchers to be a part of. Thank you to our new Immunology Director, Dr. Jagan Sastry, and co-director, Dr. Kimberly Schluns, for your support and dedication to the Immunology students.

Thank you, Dr. Melinda Yates, for your support and encouragement in the First-Generation

Student Association. The First-Gen group provided me with a great support system throughout graduate school. Thank you to Dr. Andrew Bean, Dr. Marenda Wilson-Pham, my mentors, deans, program, and graduate school for giving me the opportunity to participate in the summer graduate program in public policy with the Archer Center and to intern with the US Department of Health and Human Services in the Office of the Assistant Secretary for Preparedness and

Response. This was a once in a lifetime opportunity. Thank you to the Texas National Security

Network for providing me with a scholarship. Thank you to Cancer Answers for awarding me a scholarship in cancer research.

Thank you to my best friend, my husband, Dalton Schafer. Your support, encouragement, prayers, and undying love for me is what kept me from never giving up. You believed in me

IV even when I didn’t. Thank you for putting up with crazy experiments, with all of my faults and failures, and for allowing me to chase my dreams. Thank you to my incredible family, my mom, dad, step-mom, brother, sister in-laws, grandparents, mother and father in-law. Thank you for being my biggest cheerleaders and my encouragement, and for always believing in me. Thank you to my incredible friends both in graduate school and outside of graduate school. I have the best support system and I would not be where I am without all of you.

Thank you to God, for allowing me to be a part of discovering your Creation. My studies in science have only affirmed my faith.

V Abstract

A License to Kill: Understanding Natural Killer Cell Licensing to Fight Cancer

Jolie Rae Schafer, B.S.

(Advisor: Dean A. Lee, MD, PhD; Shulin Li, On-Site Advisor)

Natural killer (NK) cell education is an essential developmental process for NK cell effector function, that renders some NK cells “licensed” and others “unlicensed” (with heightened or lowered effector function potential, respectively) against tumor and targets lacking self- molecules. However, the underlying mechanisms responsible for the heightened effector responses of licensed cells remain unknown. Using NK cells derived from humans and expanded ex vivo we performed high-throughput protein expression analysis, and identified multiple proteins that are differentially regulated in licensed and unlicensed human NK cells before and after inhibition by killer-cell immunoglobulin-like receptors (KIRs) and activation by the NKp46 natural cytotoxicity receptor, including several related to cellular metabolic pathways. We explored cellular metabolism in the two subsets and found that licensed NK cells are highly glycolytic, and use glycolysis and mitochondrial respiration for cytolysis of leukemia targets, whereas unlicensed NK cells are dependent on mitochondrial respiration. We determined the metabolic pathways that are necessary for licensed and unlicensed NK cells to elicit a cytolytic response using metabolic inhibitors to inhibit glycolysis or mitochondrial respiration metabolic pathways in the NK cells during a cytotoxicity assay. We observed that licensed NK cells utilize both glycolysis and mitochondrial respiration to perform cytolysis whereas unlicensed NK cells only use mitochondrial respiration for their cytolytic response against leukemia targets. To our knowledge, this is the first description of the underlying mechanisms that explain the cytolytic differences between licensed and unlicensed NK cells. Our findings provide a groundbreaking platform to further explore and manipulate metabolism in licensed and unlicensed NK cells to improve NK cell immunotherapy.

VI Table of Contents

Acknowledgements ...... III

Abstract ...... VI

List of Illustrations ...... X

Figure Twenty-Four: Comparison of intracellular metabolites found in expanded licensed and

unlicensed single KIR positive NK cells...... XI

List of Tables ...... XII

Abbreviations ...... XIII

Chapter 1: Introduction ...... 1

1.1 Natural Killer Cell Biology ...... 1

1.2 NK cell Education ...... 4

1.3 NK cell Expansion ...... 9

1.4 NK cell Metabolism ...... 11

1.5 Regulation of Glycolysis ...... 13

Chapter 2: Materials and Methods ...... 15

Human Subjects ...... 15

NK Cell Isolation ...... 15

KIR and HLA Typing ...... 15

Licensed and Unlicensed NK cell Discrimination ...... 16

Flow Cytometry ...... 16

Cell Sorting ...... 17

Cell Expansion ...... 17

Cell Culture ...... 17

Cytotoxicity Assays ...... 17 VII Metabolic Inhibitors ...... 18

Mass Cytometry ...... 18

Telofish with CyTOF ...... 18

Extracellular Flux Assays ...... 19

siRNA Knockdown ...... 20

Cross-linking NK Cells for Protein Expression Analysis ...... 20

Reverse Phase Protein Array ...... 22

Western Blot ...... 34

NMR ...... 34

Statistics ...... 35

Chapter 3: Specific Aims ...... 37

Chapter 4: Understanding intrinsic mechanistic differences between licensed and unlicensed NK cells ...... 38

4.1 Licensed and unlicensed NK cell subset distribution before and after expansion ...... 39

4.2 Determining the proliferative capacity between licensed and unlicensed NK cells ...... 42

4.3 Determining the telomere length between licensed and unlicensed NK cells ...... 48

4.4 Determining differences in signaling pathways between licensed and unlicensed NK cells ... 53

Chapter 4 Discussion ...... 70

Chapter 5: Understanding the role of metabolism in licensed and unlicensed NK cell cytotoxicity ...... 75

5.1 Determining the effect AMPK signaling has on NK cell cytotoxicity ...... 76

5.2 Determining the effect AMPK signaling has on NK cell expansion ...... 78

5.3 Evaluation of licensed and unlicensed NK cell subsets utilization of glycolysis and

mitochondrial respiration ...... 80

VIII 5.4 Determining the role of glycolysis and OXPHOS in licensed and unlicensed NK cell cytolytic

effector function ...... 86

5.5 Determining the role of AMPKα and p38 signaling in NK cell glucose metabolism ...... 90

5.6 Comparison of intracellular metabolites in expanded licensed and unlicensed NK cells ...... 93

Figure Twenty-Four: Comparison of intracellular metabolites found in expanded licensed and

unlicensed single-KIR-positive NK cells...... 94

Chapter 5 Discussion ...... 96

Summary and Future Directions ...... 102

Bibliography ...... 110

Vita ...... 122

IX List of Illustrations

Figure One: Distribution of licensed and unlicensed NK cell subsets before and after expansion

Figure Two: Gating strategy for FACS of single KIR+ and KIR- NK cells

Figure Three: Licensed NK cells expand to greater numbers than unlicensed NK cells ex vivo

Figure Four: Licensed NK cells remain better killers than unlicensed NK cells after expansion for 3 weeks.

Figure Five: Spade analysis of Telomere Length Comparison between PBMC subsets.

Figure Six: Quantitative Analysis of Telomere Length Comparison between PBMC subsets.

Figure Seven: Licensed NK cells have shorter telomere lengths than unlicensed NK cells.

Figure Eight: NK cell single-KIR-positive populations after sorting on day 0 and expanding for 21 days.

Figure Ten: Baseline (isotype) signaling in expanded licensed and unlicensed NK cells.

Figure Eleven: Protein expression in expanded licensed and unlicensed NK cells after

NKp46 cross-linking.

Figure Twelve: Upregulated protein expression in licensed NK cells upon KIR cross- linking.

Figure Thirteen: Upregulated (A) or down regulated (B) protein expression in NK cells upon KIR cross-linking regardless of licensing. X Figure Fourteen: Validation of RRPA by flow cytometry and CyTOF.

Figure Fifteen: The Effect of AMPK Activation or Inhibition on licensed and unlicensed

NK cell cytotoxicity.

Figure Sixteen: The Effect of AMPK Activation or Inhibition on licensed and unlicensed

NK cell expansion.

Figure Seventeen: Expanded licensed versus unlicensed NK cell assessment of mitochondrial respiration.

Figure Eighteen: Expanded licensed versus unlicensed NK cell assessment of glycolysis.

Figure Nineteen: Assessment of glycolysis in fresh NK cells.

Figure Twenty: Assessment of glycolysis in fresh NK cells after stimulation with PM21.

Figure Twenty-One: Inhibition of metabolic pathways in NK cells to determine metabolic pathways necessary for NK cell mediated lysis.

Figure Twenty-Two: Knockdown of AMPKα1/2 in licensed NK cells to determine the role

Figure Twenty-Three: Knockdown of p38 in expanded licensed NK cells to determine the role p38 plays in NK cell glycolysis.

Figure Twenty-Four: Comparison of intracellular metabolites found in expanded licensed and unlicensed single KIR positive NK cells.

Figure Twenty-Five: Summary of licensed and unlicensed NK cell characteristics.

XI List of Tables

Table 1: KIR and HLA donor typing

Table 2: Telofish and phospho-CyTOF antibody panel

Table 3: Reverse phase protein array antibody panel

XII Abbreviations

AKTS1 RAC-alpha serine/threonine-protein kinase substrate 1 AMP 5' adenosine monophosphate AMPK 5' adenosine monophosphate-activated protein kinase ATM Ataxia-telangiectasia mutated ATP Adenosine triphosphate BAD Bcl-2-associated death promoter BAP1 BRCA1 associated protein-1 BAX Bcl-2-associated X CD Cluster of differentiation CDK Cyclin dependent kinase Chk Checkpoint kinase CM Central Memory CyTOF Cytometry by time of flight E2F1 Target of retinoblastoma protein EM Effector memory EMA Mucin 1 FACS Fluorescence-activated cell sorting FasL Fas ligand FoxM1 Forkhead box protein M1 HER2 Human epidermal growth factor receptor 2 HLA Human leukocyte antigen HSCT Hematopoietic stem cell transplantation HSP Heat shock protein IFN Interferon ITIM Immunoreceptor tyrosine-based inhibition motif JAB1 B-7 KIR Killer cell immunoglobulin-like receptor MET Hepatocyte growth factor receptor MFI Mean fluorescence intensity MHC Major histocompatibility complex MIC MHC class I polypeptide-related sequence MIF Macrophage migration inhibitory factor MMI Mean metal intensity MSH2 MutS protein homolog MTCO2 Human Cytochrome C oxidase subunit II mTOR Mechanistic Target of Rapamycin OXPHOS Oxidative phosphorylation NDRG1 N-Myc downstream regulated 1 NK Natural killer XIII NFkB Nuclear factor kappa-light-chain-enhancer of activated B cells NMR Nuclear magnetic resonance spectroscopy PBMC Peripheral blood mononuclear cell Pdcd4 Programmed cell death protein 4 PHA Phytohaemagglutinin PI3K Phosphatidylinositol-4,5-bisphosphate 3-kinase PKM2 Pyruvate kinase isozymes M1/M2 PRAS Proline-rich Akt substrate PUMA p53 upregulated modulator of apoptosis RPPA Reverse phase protein array SCD Stearoyl-CoA desaturase SD Standard deviation SDHB Succinate dehydrogenase [ubiquinone] iron-sulfur subunit, mitochondria SH2 Src Homology 2 SHP Src homology region 2 domain-containing phosphatase Smad4 Mothers against decapentaplegic homolog SPADE Spanning-tree progression analysis of density-normalized events Src Proto-oncogene tyrosine-protein kinase Src TCA Tricarboxylic acid TFAM Mitochondrial transcription factor A TNF Tumor necrosis factor TRAIL TNF-related apoptosis-inducing ligand UGT1A UDP Glucuronosyltransferase Family 1 Member A Complex Locus ULBP UL16 binding protein 1 VTCN1 V-Set Domain Containing T-Cell Activation Inhibitor 1 YWHAZ 14-3-3 protein zeta/delta

XIV Chapter 1: Introduction

1.1 Natural Killer Cell Biology

Natural killer (NK) cells represent one component of the innate arm of the immune system.

NK cells are leukocytes that differentiate from common lymphoid progenitor cells. NK cells constitute between 10-15% of peripheral blood lymphocytes. NK cells are identified by the expression of CD56 and CD16, and lacking expression of CD3 and CD19, as well as other lineage markers. Peripheral blood NK cells are predominantly composed of the CD56dim

CD16bright mature highly cytolytic NK cells, along with a smaller percentage of CD56bright

CD16dim immature NK cells that are great cytokine producers. NK cells kill their target cells by secreting cytotoxic granules, perforin and granzyme B, or through engagement of TRAIL or

FASL death receptor pathways. NK cells also secrete pro-inflammatory cytokines, IFN-g and

TNF-a, and chemokines that aid in the recruitment of the adaptive immune system. NK cells are the first responders to viral infections and cancer or transformed cells.

NK cells express germline encoded cell surface receptors and do not express rearranged antigen specific receptors like other lymphocytes such as B cells and T cells. NK cells act to alert and recruit the adaptive immune system when danger is present. NK cells engage targets using an array of receptors on the cell surface that include both activating and inhibitory receptors. The engagement and strength of the signal from the various activating and inhibitory receptors determines whether a NK cell will become activated and elicit a response. Upon activation, a NK cell will release cytotoxic granules and or cytokines by degranulation in the direction of the target cell. The killer-cell immunoglobulin-like receptor (KIR) family is the

1 primary focus of my project. KIRs can be activating or inhibitory, however my project has focused on inhibitory KIR. Inhibitory KIRs provide two major functions for NK cells. First, they aid in discriminating self from non-self and induce tolerance to prevent NK cells from attacking self, thereby providing protection from autoimmunity. Second, inhibitory receptors provide an important education signal that “license” and educate NK cells to kill cells that are missing human leukocyte antigen (HLA) such as in the case of leukemia, other cancers and viral infections. NK cell education will be explained in further detail below.

Receptors

KIRs are found on human, primate and cattle NK cells. The murine counterpart of KIRs are Ly49 receptors. KIRs and Ly49 receptors recognize major histocompatibility complex class

I ligands (MHC I), also referred to as human leukocyte antigen (HLA) in humans. KIR expression is acquired later in NK cell development and is found primarily on mature NK cells

(Beziat et al., 2010). There is discrepancy in the field as whether KIR acquisition is stochastic or whether KIR acquisition is dependent upon MHC-I expression, as there is evidence for both of these hypotheses (Beziat et al., 2010; Sleiman et al., 2014). KIRs have heterogeneity at many levels; functional (inhibitory or activating), haplotype, expression, ligand, and allelic diversity, and all contribute to the complexity of KIRs. The KIR haplotypes are A and B (Uhrberg et al.,

1997). The A haplotype is characterized by the presence of all of the inhibitory KIR genes and no activating KIRs (Uhrberg et al., 1997). The B haplotype is defined by the presence of any combination of activating KIR genes (Uhrberg et al., 1997). For my project, I only used donors with KIR A haplotype to eliminate any influence activating KIRs may have on NK cell function. KIRs have expression diversity, NK cells express multiple combinations of different

KIRs and varying levels of KIR expression (Cheent and Khakoo, 2009). Each KIR has a

2 different HLA binding group ligand; KIR ligands will be discussed below. Lastly, there is allelic diversity; different KIR alleles have varying protection against diseases, variable expression and binding affinities to their ligands (Martin et al., 2007; Middleton et al., 2008; Moesta et al.,

2008).

Human Leukocyte Antigens

Human leukocyte antigens (HLA) are highly polymorphic glycoproteins and are normally constitutively expressed on all nucleated cells in the body. HLA molecules present self and non-self-peptides to T cells. HLA-class I molecules are further divided up into classical

(HLA-A, -B, -C) and non-classical molecules (HLA-E, -F, -G, UL16-binding proteins (ULBP) molecules, and major histocompatibility complex class I-related chain (MIC) A, MICB). HLA- class I classical molecules are the ligands for KIRs expressed on both NK cells and T cells

(Leone 2017 KIR-HLA). NKG2A (inhibitory) and NKG2C (activating) NK cell receptors both interact with HLA-E. KIRs recognize the HLA-I binding groups HLA Bw4, C1 and C2. Like

KIR, HLA also has allelic diversity. Various HLA alleles have varying binding affinity for KIR and have varying expression levels (Marra et al., 2015).

Tumor response

NK cells mount a cytolytic response against numerous tumor types (Aquino-Lopez et al., 2017). Tumor cells are most sensitive to NK cell lysis when the tumor cells have downregulated HLA-I molecules and have upregulated stress-ligands. Tumor cells downregulate

HLA-I as a way to escape T cell killing, however once HLA is downregulated tumor cells become susceptible to NK cell killing (Ljunggren and Karre, 1985; Marin et al., 2003; Demanet et al., 2004; Lanier, 2008; Aquino-Lopez et al., 2017). The importance of NK cells for tumor

3 surveillance has been demonstrated in patients who are deficient in NK cells by having an increased incidence of various cancers (Orange, 2013). Murine tumor studies have shown that the absence or depletion of NK cells leads to increased tumor growth and metastasis (Kim et al.,

2000; Hayakawa et al., 2002; Wu and Lanier, 2003; Hayakawa and Smyth, 2006). Autologous and allogeneic NK cells are currently used in the clinical setting for the treatment of leukemia and other cancers (Burns et al., 2003; Miller et al., 2005; Romee et al., 2016).

Viral response

NK cells elicit a potent effector response against virally infected cells and play an important role in viral clearance. Viral defense by NK cells is particularly important during hematopoietic stem cell transplant (HSCT). NK cells are the first lymphocytes to appear after HSCT and are critical for protecting patients against viral infections and tumor reoccurrence post-transplant.

Post-transplant, patients are highly susceptible to viral infections, the role of NK cells is extremely important to protect patients from viruses that can be lethal with a compromised immune system. Post-transplant NK cells help mediate graft versus leukemia effect as well as aiding in protection against graft versus host disease, which can be life threatening.

1.2 NK cell Education

NK cell education is an important developmental process that adjusts the alertness of NK cells to become functionally competent. NK cell functional competency is defined here by the ability to elicit a potent and quick cytokine or cytolytic response. This educational process also called licensing is dependent upon the host environment (Kim et al., 2005; Anfossi et al., 2006;

4 Raulet, 2006). NK cell education or licensing is important for NK cell function. In mice, unlicensed NK cells are insufficient at viral sensing and viral clearance (Wei et al., 2014).

NK cells become licensed when the inhibitory KIR on an NK cell interacts with the corresponding self-HLA ligand either on the NK cell itself (cis) or a bone marrow derived cell

(trans) during development (Bessoles et al., 2013; Ebihara et al., 2013). NK cell education is critical for effector NK cells to have heightened sensitivity against targets that have downregulated HLA. The well-characterized inhibitory KIR that contribute to NK cell education are

KIR2DL1, KIR2DL2/3 and KIR3DL1 and they interact specifically with HLA C2, HLA C1 and

HLA Bw4 ligands, respectively. Diversity also exists on the ligand level for NK cells because every individual can inherit HLA C1, HLA C2, HLA Bw4 and HLA Bw6 (for which there is no known inhibitory KIR) and various combinations of the four. This means that in some individuals, NK cells expressing KIR may be unlicensed if their corresponding self-ligand is not present, leaving those unlicensed NK cells hypo responsive to targets missing HLA. In the cancer setting if an NK cell is unlicensed, even when the tumor cell is missing the self-HLA ligand, unlicensed NK cells will be hypo responsive (Fernandez et al., 2005).

There are multiple proposed mechanisms that might explain NK cell education, the ones focused on here are the arming model, disarming model and the tuning model.

Arming model

The ‘Arming Model’ proposes that NK cells are initially hypo-responsive and require the interaction between KIR and HLA to acquire functional competency (Kim et al., 2005). In this model, KIR ligation with HLA induces inhibitory signaling and causes the phosphorylation of immunoreceptor tyrosine-based inhibitory motifs (ITIMs) leading to recruitment and

5 activation of the SH2-domain-containing protein tyrosine phosphatase 1 (SHP1). Through an unknown mechanism this inhibitory pathway leads to a heightened cytolytic response in the licensed NK cells against targets that lack HLA. Evidence for this model is supported by murine studies that have observed that NK specific deletion of ITIMs or SHP1 have hypo-responsive and uneducated NK cells (Kim et al., 2005; Wahle et al., 2007; Viant et al., 2014). Adoptive transfer of murine mature unlicensed NK cells into a licensing NK cell host showed gain of function in the previously unlicensed NK cells, after as few as 7 days post adoptive transfer

(Elliott et al., 2010). This finding suggests that licensing is plastic and that KIR-HLA interaction

“arms” the NK cell cytolytic response.

Disarming model

The “Disarming Model” proposes that NK cells have heightened effector potential, without needing an arming signal, and through overstimulation by activating receptors without inhibitory signals to counteract and balance the NK cell response, NK cells become “disarmed” and hypo-responsive (Raulet, 2004; Fernandez et al., 2005; Raulet, 2006). Support from this model comes from studies observing that continued engagement of activating receptors with their ligands in vivo and in vitro leads to NK cell hypo-responsiveness (Oppenheim et al., 2005;

Coudert et al., 2008; Tripathy et al., 2008; He and Tian, 2017). Additionally, the sustained engagement of the activating receptor KIR2DS1 leads to reduced NK cell responsiveness

(Fauriat et al., 2010; He and Tian, 2017).

Rheostat/Tuning model

The “Rheostat Model” proposes that NK cell responsiveness is dependent on the strength of inhibitory signals received by MHC-I (Brodin et al., 2009; Brodin et al., 2009;

6 Joncker et al., 2009; Kadri et al., 2016). This model also proposes that NK cell responses are constantly being tuned in mature NK cells by adjusting to their environment (Brodin et al.,

2009; Brodin et al., 2010; Hoglund and Brodin, 2010; Kadri et al., 2016). Evidence for the

“Rheostat Model” is supported by mouse and human studies showing that the number of self-

MHC-I receptors, Ly-49 and KIR in mice and humans respectively, is positively correlated with increased NK cell responsiveness (Brodin et al., 2009; Joncker et al., 2009; Sleiman et al.,

2014). The more educated inhibitory receptors present on the NK cell, the greater the NK cell effector response was observed. However, other studies in humans have also shown that there is no correlation between the quantity of self-KIR expression on the NK cell and NK cell responsiveness (Charoudeh et al., 2012; Beziat et al., 2013). Further studies have investigated

MHC-I expression levels in vitro and in vivo and whether or not MHC-I expression levels can tune the NK cell response depending on availability and quantity of MHC-I levels. However, several groups have obtained opposing findings, one group found that either low or normal expression of MHC-I leads to equal NK cell education, however another group reported that expression levels of MHC-I does matter and that lower levels of MHC-I leads to NK cells exhibiting lower responsiveness (Jonsson et al., 2010; Brodin et al., 2012). Due to opposing findings in support of or against the “Rheostat Model,” it is difficult to draw conclusions on whether or not this model holds up as NK cells undergo education in vivo. Further studies are needed to determine the validity of this model.

Non-classical MHC and MHC independent education.

In addition to classical MHC education from HLA and KIR interaction, there are observations of NK cell education and tuning of responses by inhibitory receptors that interact with non-classical MHC and MHC independent ligands for NK cell education. Human NK cell

7 inhibitory receptors that are reported to contribute to NK cell responsiveness include NKG2A,

TIGIT, 2B4 and interact with the ligands HLA-E, polio virus receptor CD155, and signaling lymphocyte activation molecule (SLAM)-related ligand CD48, respectively. For my dissertation, I am focusing on the contribution of human KIR receptors and their contribution to human NK cell education.

The standard approach for studying human NK cell education is using freshly isolated NK cells from the peripheral blood of donors for which HLA and KIR haplotypes are known.

However, the median NK cell content in peripheral blood is only 1.26 x 10^5 NK cells/mL

(Pittari et al., 2013), so that sorting pure populations of licensed and unlicensed NK cells from an entire buffy coat produces only 1,000,000, and 300,000 licensed and unlicensed NK cells respectively. Such low numbers make some mechanistic studies on these populations infeasible on freshly isolated cells unless apheresis products are used. To overcome this challenge, I expanded human NK cells ex vivo using an expansion platform developed in our lab as a way to generate large amounts of licensed and unlicensed pure NK cell subsets to perform analysis with.

It is important to distinguish which NK cell licensing model holds biological significance since the hypo responsiveness of the NK cells could potentially be altered in the case of the

NK cell “arming” or “rheostat” model.

1.3 NK cell Expansion

Studying human NK cell education and licensed versus unlicensed NK cell populations is difficult due to feasibility of studying these populations freshly isolated from the peripheral blood. Apheresis is likely the only way to obtain enough cells of each population to study.

However, the willingness and availability of apheresis donors is quite limited, and further limited by having to perform costly HLA and KIR genotyping on many potential apheresis donors in order to identify a few suitable donors. NK cell expansion from a small amount of peripheral blood is currently used as a way to generate high quantities of NK cells to be used for

NK cell adoptive therapy (Somanchi et al., 2011; Denman et al., 2012). We had previously performed KIR and HLA typing on 25 colleagues who were willing to donate peripheral blood for research studies under an IRB approved protocol.

Table 1: KIR and HLA donor typing

I utilized an NK cell expansion platform developed in our lab as a way to generate sufficient numbers of licensed and unlicensed NK cell populations to study mechanistic differences between the two subsets (Somanchi et al., 2011; Denman et al., 2012). Other methods of NK cell expansion have also been used, however the method developed by our laboratory achieved expansion of NK cells unmatched by all other NK cell expansion techniques. The expansion platform I used involves expanding licensed and unlicensed sorted NK cell populations with the

Clone9.mbIL21 K562 feeder cell line and low dose IL-2. The feeder cells express 4-1BBL and membrane bound IL-21 that aid in activating and propagating the cells ex vivo. This expansion 10 method generates a mean fold expansion of 48,000 with bulk NK cells isolated from the peripheral blood (Denman et al., 2012). While this is not the ideal model system to study NK cell education due to artificially manipulating the NK cells in vitro, many differences were observed in the licensed and unlicensed NK cell signaling pathways between the two subsets post ex vivo expansion.

1.4 NK cell Metabolism

Generating energy is important for cell functions such as proliferation and effector functions. There are two major energy-producing pathways glycolysis and mitochondrial respiration also known as oxidative phosphorylation (OXPHOS) , that I focused on for my dissertation. Glycolysis involves the conversion of intracellular glucose into pyruvate. Pyruvate has two major fates: the first is that pyruvate can be further converted into lactate and shuttled out of the cell as lactic acid, this is a short and quick way of producing energy and 2 molecules of ATP are generated per one glucose molecule. This short and quick way of metabolizing glucose into lactic acid is conventionally performed in the absence of oxygen (anaerobic conditions), whereas OXPHOS requires oxygen. However, highly proliferative cells such as tumor cells and immune cells also undergo glycolysis to produce lactic acid even in the presence of oxygen (aerobic), this is called the Warburg effect and was first described by Otto Warburg

(Warburg et al., 1927; Warburg et al., 1958). In addition to energy aerobic glycolysis can generate biosynthetic precursors that facilitate the survival of proliferating cells (Romero-Garcia et al., 2016). The second fate of pyruvate is to be converted into acetyl-coA and shuttled into the tricarboxylic acid cycle (TCA). The TCA cycle generates NADH, which is then used in

OXPHOS to generate 34 ATP molecules in addition to 2 ATP molecules produced from glycolysis and 2 ATP molecules produced in the TCA cycle. Traditionally, OXPHOS is able to 11 produce 70-95% of the cells energy (Vander Heiden et al., 2009; Nsiah-Sefaa and McKenzie,

2016). While OXPHOS is a very efficient metabolic pathway to produce energy, it is also very slow. Aerobic glycolysis can be quicker at generating a large amount of ATP than mitochondrial respiration.

Several studies in T cells have shown that proliferation is dependent upon glucose metabolism (Newsholme et al., 1985; Newsholme et al., 1985; Jacobs et al., 2008; Wang et al.,

2011; van der Windt and Pearce, 2012). Upon T cell activation, GLUT1glucose transporters are up regulated along with an increase in glucose uptake and glycolysis (Frauwirth et al., 2002;

Jacobs et al., 2008). Few metabolic studies have been performed in NK cells, and only bulk NK cell analysis has been possible because of the infrequent numbers of NK cells per subset that exist within an individual (Donnelly et al., 2014; Marcais et al., 2014; Marcais and Walzer,

2014; Keppel et al., 2015). Glycolysis and OXPHOS are critical for murine and bulk human NK cell cytokine production (Donnelly et al., 2014; Keppel et al., 2015; Keating et al., 2016).

Resting NK cells exhibit low levels of OXPHOS and glycolysis (Keppel et al., 2015). One study compared human NK cell subsets CD56dim mature NK cells to CD56bright immature NK cells and found that the CD56bright NK cells were more metabolically active and have increased glucose metabolism compared to the CD56dim NK cells (Keating et al., 2016). In the same study, OXPHOS metabolism was found to be important for NK cell cytotoxicity and IFN-γ production. Lastly in the CD56bright NK cells glycolysis was a key regulator of IFN- γ production as these cells are the main cytokine producers (Keating et al., 2016).

12 1.5 Regulation of Glycolysis

Glycolysis is a tightly regulated pathway and can be regulated by different proteins in various cell types. T cell activation through CD28 signaling leads to PI3K/Akt/mTOR signaling cascade, leading to an increase in aerobic glycolysis (Frauwirth et al., 2002). Akt and mTOR signaling act to induce glycolysis post-translationally by increasing glucose transporters on the cell surface, activating p70S6 kinase leading to the activation of glycolysis and activating glycolysis enzymes (Miyamoto et al., 2008; Tandon et al., 2011; Gerriets and Rathmell, 2012).

AMP-activated protein kinase (AMPK) can act as both an inhibitor and an activator of glycolysis. AMPK becomes activated when the ratio of AMP to ATP is high, thus sensing low cellular energy levels. In a study of T cell lymphoblastic leukemia, AMPK signaling inhibited glycolysis through a mTORC1 dependent manner (Kishton et al., 2016). In a separate study,

AMPK signaling through activation of PFKFB3, a glycolysis enzyme, acted to increase glycolysis leading to survival of breast cancer cell lines (Domenech et al., 2015). Another study observed that AMPK signaling leads to activation of glycolysis during mitochondrial stress (Wu and Wei, 2012). In one study, AMPK activation occurs after T cell activation via TCR stimulation by a Ca2+ -calmodulin-dependent kinase kinase kinase (CAMKK) dependent mechanism (Tamas et al., 2006). AMPK signaling can also lead to activation of the glucose transporter GLUT1 in a rat liver epithelial cell line (Barnes et al., 2002). The importance of

AMPK signaling in metabolism lead us to investigate the role of AMPK signaling in NK cell glycolysis.

P38, a member of the mitogen activated protein kinase (MAPK) protein family was shown to play a role in glycolysis activation and inhibition. P38 in response to stress controls transcription of immediate-early genes. In one study, p38 signaling led to a signaling cascade

13 that resulted in PFKFB3 activation leading to an increase in glycolysis (Novellasdemunt et al.,

2013).

14 Chapter 2: Materials and Methods

Human Subjects

NK cells were derived from peripheral blood or buffy coats of normal healthy volunteer donors under the MD Anderson Cancer Center IRB-approved protocol LAB07-0296 or IRB exemption

PA13-0978. Donors were selected by KIR and HLA genotyping. Only donors with KIR A haplotype were used, to eliminate all activating KIRs from influencing any data. 5 donors were used total, and replicates for each donor are described in the figure legends.

NK Cell Isolation

NK cells were isolated from both buffy coats and whole blood using the previously described method (Somanchi et al., 2011). Briefly, PBMCs from buffy coat or whole blood diluted with

PBS at a 2:1 PBS to blood ratio was layered over ficoll hypaque in 50mL conical tubes and centrifuged for 20 min at 1200 RPM with no break. PBMCs were collected and washed 2 times with PBS. PMBCs were counted and RBCs were added at ratio of 100:1 RBC: WBC. RBCs and

WBCs were then incubated with Rosette Sep NK cell cocktail for 20 min. Cells were then layered over ficoll hypaque in 50mL conical tubes and centrifuged for 20 min at 1200 RPM no break. Purified NK cells were then washed 2 times with PBS and RBCs were lysed when necessary. Purity of NK cells was assessed by flow cytometry.

KIR and HLA Typing

DNA was isolated from NK cells and submitted to the HLA laboratory at the University of

Texas MD Anderson Cancer Center for HLA and KIR typing. HLA typing was performed at the intermediate-resolution level at HLA-A, HLA-B, and HLA-C loci by PCR amplification and oligonucleotide hybridization (One Lambda, Canoga Park, CA). KIR genotyping was performed 15 with reverse sequence-specific oligonucleotide methodology using fluorescently labeled beads conjugated to oligonucleotide probes (One Lambda). The HLA group for KIR binding was determined using the KIR Ligand Calculator maintained by the European Bioinformatics

Institute of the European Molecular Biology Labs (EMBL-EBI)

(http://www.ebi.ac.uk/ipd/kir/ligand.html).

Licensed and Unlicensed NK cell Discrimination

To avoid confounding from expression of multiple KIRs, here we identified licensed and unlicensed NK cells as those expressing a single inhibitory KIR in an individual who has or lacks, respectively, the relevant cognate We selected the following KIR AA haplotype available individuals from table 1; donors 13, 18, and 19. Donors 8 and 5 were also used for preliminary experiments. HLA ligand and lacks a complementary activating KIR (e.g.,

KIR2DL1+KIR2DL2/3−KIR3DL1− NK cells in an HLA-C*0201 individual who lacks the

KIR2DS1 gene). Licensed status was determined irrespective of NKG2A expression.

Flow Cytometry

NK cells were labeled with fluorescently conjugated human antibodies to determine licensed and unlicensed NK cell populations: CD56 PerCP CY5.5 (Biolegend), CD3 APC Cy7

(Biolegend), KIR2DL1 FITC (R&D), KIR2DL2/3 APC (Miltenyi), and KIR3DL1 PE

(Beckman Coulter). Cells were stained with antibodies at concentrations according to manufacturer guidelines and titrated when necessary. For intracellular staining, cells were washed with staining buffer containing phosphate buffered saline (PBS) and 5% fetal bovine serum (FBS). Cells were fixed in 2% formaldehyde for 10 minutes, then washed and permeablized with methanol on ice for 10 minutes. NK cells were then washed with staining buffer, followed by staining for p-38 MAPK (D13E1) antibody (Cell Signaling Technologies).

16 Flow cytometry was performed on an LSR Fortessa (BD Biosciences).

Cell Sorting

Cells were labeled with the following fluorescently conjugated human antibodies: CD56 PerCP

CY5.5 (Biolegend), CD3 APC Cy7 (Biolegend), KIR2DL1 FITC (R&D), KIR2DL2/3 APC

(Miltenyi), and KIR3DL1 PE (Beckman Coulter). Cells were stained with antibodies at concentrations according to manufacturer guidelines and titrated when necessary. Cells were sorted using a FACSAria IIu sorter (BD Biosciences).

Cell Expansion

Each NK cell population positive for a single KIR were expanded in the presence of K562

Clone9.mbIL21 artificial activating and propagating cells for 14-21 days as previously described (Denman et al., 2012). Cells were counted using a hemocytometer. Briefly, purified

NK cell populations were expanded for 7 days with a 2:1 K562 Clone9.mbIL21feededer cell:

NK cell ratio. Fresh IL-2 at 50 IU/mL was added every other day along with fresh media.

721.221 cell line was used in the expansion of bulk NK cells from typed donors to determine the percentages of licensed and unlicensed NK cells before and after expansion.

Cell Culture

Mononuclear cells from peripheral blood were cultured in RPMI 1640 Medium supplemented with 50 IU/mL recombinant human IL-2, 10% FBS, L-glutamine, and penicillin/streptomycin, as previously described (Denman et al., 2012). The medium was changed every other day.

Cytotoxicity Assays

Frozen PBMCs or expanded NK cells were thawed two days prior to effector assays.

CalceinAM was used to measure cytotoxicity as previously described(Somanchi et al., 2011).

17 Effector target ratios are described in figure legends. Briefly, 721.221 cells were loaded with calcein for 1hr. 721.221 calcein loaded target cells were incubated with NK cells at varying effector target ratios for 4-6 hours as indicated in figure legends.

Metabolic Inhibitors

The following metabolic inhibitors were used during cytotoxicity assays and expansion protocols as indicated: 500µM AICAR, 5µM shikonin, 2mM 2-DG, and 40nM oligomycin (all from Sigma). Glucose-free medium was made using glucose-free RPMI supplemented with

10% dialyzed FBS, 50 IU/mL recombinant human IL-2, L-glutamine, and penicillin/streptomycin.

Mass Cytometry

Frozen primary PBMCs (for Telo-FISH) and expanded NK cells (for phospho-CyTOF staining) were thawed 2 days prior to staining to allow for cell recovery. Antibodies were labeled with heavy metals using Maxpar-X8 labeling reagent kits (DVS Sciences) according to the manufacturer’s instructions and were titrated to determine the optimal concentration. Live cell discrimination was performed by washing cells twice in serum-free medium for 5 minutes at

500 × g. Cells were then labeled with 5 µM cisplatin for 1 minute followed by two washes in complete medium for 5 minutes at 500 × g. Cells were stained as previously described, and data were acquired on a CyTOF instrument (DVS Sciences; (Bendall et al., 2011)). The antibody clones and their respective heavy metal are in Table 1.

Telofish with CyTOF

PBMCs were first stained for live cell discrimination with 5uM cisplatin in serum free RMPI media for 1 minute. Cells were then washed and incubated with Fc blocker for ten minutes and then stained for surface receptors. Next cells were fixed with 2% formaldehyde for ten minutes 18 at room temperature. Cells were then permeabilized with methanol for ten minutes on ice.

Intracellular staining was then performed for 30 minutes. Cells were then crosslinked by BS3

(Thermo scientific) 200uL of 10mM BS3 solution added drop wise to cell pellet and incubated at

4oC for 30 minutes (Schmid and Jamieson, 2004). 20uL of 1M TrisCl, pH 8.0 was added to the cell solution, mixed well and incubated at room temperature for 30 minutes. Cells were washed with PBS and centrifuged 500 g for 5 minutes. Cells were resuspended in PBS and counted using a hemocytometer. Cells were centrifuged at 500 g for 5 minutes in 500uL of PBS containing 2% BSA in a 1.5mL eppendorf tube. 500uL of deionized formamide was added to the cell suspension and mixed gently by pipetting. Cells were incubated for ten minutes at room temperature, fifteen minutes at 87oC and then ten minutes at room temperature. Cells were centrifuged and washed two times with 2% BSA in PBS. After the final spin, supernatant was

removed and 2% BSA in PBS containing 2.5nM PNA probe Cy5-OO-(CCCTAA)3 in ratio of

100uL per one million cells was added. Cells were incubated overnight at room temperature in the dark with gentle rotation. Cells were washed twice with 2% BSA in PBS. Cells were stained with an anti-Cy5 metal conjugated antibody (same as intracellular staining above). Cells were then stained with Ir intercalator after intracellular staining as previously described (Bendall et al., 2011).

Extracellular Flux Assays

All reagents were from Agilent Technologies unless otherwise stated. Mito stress and glycolysis stress tests were performed to measure extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) using the Seahorse XFe96 analyzer. NK cells were seeded (300,000 cells per well) on 96-well flux plates coated with Cell-Tak (Corning) according to manufacturer guidelines. Glycolysis was determined in the presence of saturating amounts of glucose in the glycolysis stress test. Glycolytic Capacity is the maximum glycolytic rate after mitochondrial 19 respiration is inhibited in the mitochondrial stress test. Glycolytic reserve represents the maximum glycolytic function of a cell and equals Glycolytic Capacity - Glycolysis (Agilent

Technologies). NK cells were assayed in triplicate in XF Base Medium at pH of 7.35. For the glycolysis stress test, NK cells were supplemented with 2 mM glutamine, 10 mM glucose, 1 µM oligomycin, and 50 mM 2-DG. For the mitochondrial stress test, XF Base Medium was supplemented with 25 mM glucose, 1 mM pyruvate, and 2 mM glutamine, and the following drug concentrations were used: 1 µM oligomycin, 1.5 µM FCCP, and 0.5 µM rotenone and

Antimycin A. siRNA Knockdown

Expanded licensed NK cells were subjected to siRNA knockdown using ON-TARGETplus

Non-Targeting Pool, ON-TARGETplus human MAPK14 siRNA, ON-TARGETplus PRKAA1 siRNA and ON-TARGETplus PRKAA2 (all from Dharmacon). siRNA knockdown was performed using a Human NK Cell Nucleofector Kit (Lonza) by electroporation with an Amaxa

Nucleofector Device as previously described (Phatarpekar et al., 2016).

Cross-linking NK Cells for Protein Expression Analysis

1x107 cells were used for cross-linking experiments. Cells were washed in NK cell medium containing 10% FBS for blocking. NK cells were incubated for 30 minutes on ice with the primary antibody at a concentration of 10µg/mL. Cells were then washed twice. To cross-link the cells, the secondary antibody was added at a concentration of 1µg/mL for 10 min at 37°C.

Cells were then pelleted for protein lysate purification. Primary antibodies include: KIR3DL1

(BD Bioscience), KIR2DL1 (Biolegend), KIR2DL2/3 (Biolegend), human NKp46 (Biolegend).

Isotype control antibody was Mouse IgG1 (Tonbo). Secondary antibody was goat-anti-mouse

IgG (Biolegend).

20 Table 1: Telofish and phospho-CyTOF antibody panel.

Marker Metal or clone source Catalog # Fluorophore Idu I127Di Sigma I7125-5G PE* 141Pr PE001 BioLegend 408102 CD19 Nd142Di HIB19 BioLegend 302202 FITC* 144Nd FIT-22 DVS-Fluidigm 3144006B CD4 Nd145 RPA-T4 BioLegend 300502 CD8a Nd146Di RPA-T8 BioLegend 301002 CD20 Sm147Di 2H7 DVS-Fluidigm 3147001B CD45RA Eu153 HI100 DVS-Fluidigm 3153001B CD45 Sm154Di HI30 DVS-Fluidigm 3154001B CD33 Gd158Di WM53 DVS-Fluidigm 3158001B Cy5 Tb159Di CY5-15 Abcam ab52061 CD14 Gd160Di M5E2 DVS-Fluidigm 3160001B p-p38 (180/182) 160Dd D3F9 CST 4511BF NKp80 161Dy 5D12 BioLegend 346702 CD56 Dy162Di NCAM16.2 BD 559043 p-AMPK (T172) 164Dy 40H9 CST 2535BF CD16 Ho165Di 3G8 DVS-Fluidigm 3165001B NKG2A 166Er 131411 R&D MAB1059 NKG2C 169Tm 134522 R&D MAB1381 APC* 170Er A85-1 BD 560089 CD62L Tb171Di DREG-56 BioLegend 304802 CD57 Yb172Di HCD57 DVS-Fluidigm 3712009B CD3 Lu175Di UCHT1 BioLegend 300443 CyclinB1 Yb176Di GNS-1 BD 554177 CD45 QDOT655 HI30 Thermo Fisher Q22154 Scientific KIR3DL/DS1 PE Z27.3.7 Beckman 41116015 Coulter KIR2DL1 FITC 143211 R&D FAB1844F KIR2DL2/DL3 APC DX27 Miltenyi 130-092-617

21 Reverse Phase Protein Array

Cells were first expanded for 3 weeks, as described above. A reverse phase protein array was performed by the University of Texas MD Anderson Cancer Center RPPA core. In brief, NK cell lysates were serially diluted two-fold, five times, and arrayed on nitrocellulose-coated slides. Samples were probed with antibodies using a tyramide-based signal amplification approach and were visualized using the diaminobenzidine colorimetric reaction. Slides were scanned on a flatbed scanner to produce a 16-bit tiff image. Spots from the images were identified, and the optical density was quantified with an Array-Pro Analyzer

(MeidaCybernetics, Rockville, Maryland). Relative protein levels for each sample were determined by interpolation of each dilution curve from the “standard curve” (SuperCurve) of the slide (antibody). The SuperCurve was constructed by a script in R. All data points were normalized for protein loading and transformed to linear values and then to log2 values.

22 Table 2: Reverse phase protein array antibody Panel.

# Official Ab Name Ab Name Reported on Gene Name Company Catalog # Dataset

1 14-3-3 beta 14-3-3-beta YWHAB Santa Cruz sc-628 2 14-3-3 epsilon 14-3-3-epsilon YWHAE Santa Cruz sc-23957 3 14-3-3 zeta 14-3-3-zeta YWHAZ Santa Cruz sc-1019 4 4E-BP1 4E-BP1 EIF4EBP1 CST 9452 5 4E-BP1 (phospho S65) 4E-BP1_pS65 EIF4EBP1 CST 9456 6 4E-BP1 (phospho T37/T46) 4E-BP1_pT37_T46 EIF4EBP1 CST 9459 7 53BP1 53BP1 TP53BP1 CST 4937 8 Acetyl CoA Carboxylase (phospho S79) ACC_pS79 ACACA, ACACB CST 3661 9 Acetyl CoA Carboxylase 1 ACC1 ACACA Abcam ab45174 10 ACVRL1 ACVRL1 ACVRL1 Abcam ab108207 11 ADAR1 ADAR1 ADAR Abcam ab88574 12 Akt Akt AKT1,2,3 CST 4691 13 Akt (phospho S473) Akt_pS473 AKT1,2,3 CST 9271 14 Akt (phospho T308) Akt_pT308 AKT1,2,3 CST 2965 15 AMPK alpha AMPKa PRKAA1 CST 2532 16 AMPK alpha (phospho T172) AMPKa_pT172 PRKAA1 CST 2535 17 Androgen Receptor AR AR Abcam ab52615 18 Annexin I Annexin-I ANXA1 BD Biosciences 610066 19 Annexin VII Annexin-VII ANXA7 BD Biosciences 610668 20 A-Raf A-Raf ARAF CST 4432 21 ARHI ARHI DIRAS3 MDACC Laboratory Bast Lab 22 ARID1A ARID1A ARID1A Sigma-Aldrich HPA005456 23 Atg3 Atg3 ATG3 CST 3415 24 Atg7 Atg7 ATG7 CST 8558 25 ATM ATM ATM CST 2873 23 26 ATM (phospho S1981) ATM_pS1981 ATM CST 5883 27 ATP5A ATP5A ATP5A Abcam ab14748 28 ATR (Phospho S428) ATR_pS428 ATR Abcam ab178407 29 Aurora B/AIM1 Aurora-B AIM1 CST 3094 30 Axl Axl AXL CST 8661 31 B7-H3 B7-H3 CD276 CST 14058 32 B7-H4 B7-H4 VTCN1 CST 14572 33 Bad (phospho S112) Bad_pS112 BAD CST 9291 34 Bak Bak BAK1 Abcam ab32371 35 BAP1 BAP1 BAP1 Santa Cruz sc-28383 36 Bax Bax BAX CST 2772 37 Bcl2 Bcl2 BCL2 Dako M0887 38 Bcl2A1 Bcl2A1 BCL2A1 Abnova PAB8528 39 Bcl-xL Bcl-xL BCL2L1 CST 2762 40 Beclin Beclin BECN1 Santa Cruz sc-10086 41 beta Actin b-Actin ACTB CST 4970 42 beta Catenin b-Catenin CTNNB1 CST 9562 43 beta Catenin (phospho T41/S45) b-Catenin_pT41_S45 CTNNB1 CST 9565 44 Bid Bid BID Abcam ab32060 45 Bim Bim BCL2L11 Abcam ab32158 46 B-Raf B-Raf BRAF Abcam ab33899 47 B-Raf (phospho S445) B-Raf_pS445 BRAF CST 2696 48 BRD4 BRD4 BRD4 CST 13440 49 c-Abl c-Abl ABL CST 2862 50 Caspase-3 active Caspase-3 CASP3 Abcam ab32042 51 Caspase-7 (cleaved D198) Caspase-7-cleaved CASP7 CST 9491 52 Caspase-8 Caspase-8 CASP8 CST 9746 53 Caveolin-1 Caveolin-1 CAV1 CST 3238 54 CD171 (L1) CD171 L1CAM BioLegend 826701 55 CD26 CD26 CD26 Abcam ab28340 24 56 CD29 CD29 ITGB1 BD Biosciences 610467 57 CD31 CD31 PECAM1 Dako M0823 58 CD44 CD44 CD44 CST 3570 59 CD49b CD49b ITGA2 BD Biosciences 611016 60 cdc25C cdc25C CDC25C CST 4688 61 CDK1 CDK1 CDK1 Abcam ab32384 62 CDKN2A/p16INK4a p16INK4a CDKN2A Abcam ab81278 63 Chk1 Chk1 CHEK1 CST 2360 64 Chk1 (phospho S296) Chk1_pS296 CHEK1 Abcam ab79758 65 Chk1 (phospho S345) Chk1_pS345 CHEK1 CST 2348 66 Chk2 Chk2 CHEK2 CST 3440 67 Chk2 (phospho T68) Chk2_pT68 CHEK2 CST 2197 68 c-Jun ( phospho S73) c-Jun_pS73 JUN CST 9164 69 c-Kit c-Kit KIT Abcam ab32363 70 Claudin 7 Claudin-7 CLDN7 Novus Biologicals NB100-91714 71 c-Met c-Met MET CST 3127 72 c-Met (phospho Y1234/Y1235) c-Met_pY1234_Y1235 MET CST 3129 73 c-Myc c-Myc MYC Santa Cruz sc-764 74 COG3 COG3 COG3 ProteinTech 11130-1-AP 75 COL6A1 Collagen-VI COL6A1 Santa Cruz sc-20649 76 Complex II Subunit Complex-II-Subunit SDHB Invitrogen 459230 77 Connexin 43 Connexin-43 CNST43 CST 3512 78 Coup-TFII Coup-TFII NR2F2 CST 6434 79 Cox IV Cox-IV COX4I1 Abcam ab14744 80 Cox2 Cox2 PTGS2 CST 4842 81 C-Raf (phospho S338) C-Raf_pS338 RAF1 CST 9427 82 C-Raf/Raf-1 C-Raf RAF1 Millipore 04-739 83 CXCR4 CXCR4 CXCR4 Abcam ab2074 84 Cyclin B1 Cyclin-B1 CCNB1 Epitomics 1495-1 85 Cyclin D1 Cyclin-D1 CCND1 Santa Cruz sc-718 25 86 Cyclin E1 Cyclin-E1 CCNE1 Santa Cruz sc-247 87 Cyclophilin F Cyclophilin-F PPIF Abcam ab110324 88 Cytokeratin 19 Cytokeratin-19 KRT19 Dako M0888 89 Detyrosinated alpha-Tubulin D-a-Tubulin TUBA1A Abcam ab48389 90 Dimethyl-Histone H3 (Lys4) DM-Histone-H3 HISTH3 Millipore 07-030 91 Dimethyl-K9 Histone H3 DM-K9-Histone-H3 H3K9ME2 Abcam ab32521 92 DUSP4/MKP2 DUSP4 DUSP4 CST 5149 93 Dvl3 Dvl3 DVL3 CST 3218 94 E2F-1 E2F1 E2F1 Santa Cruz sc-251 95 E-Cadherin E-Cadherin CDH1 CST 3195 96 eEF2 eEF2 EEF2 CST 2332 97 eEF2K eEF2K EEF2K CST 3692 98 EGFR EGFR EGFR CST 2232 99 EGFR (phospho Y1068) EGFR_pY1068 EGFR CST 2234

100 EGFR (phospho Y1173) EGFR_pY1173 EGFR Abcam ab32578 101 eIF4E eIF4E EIF4E CST 9742 102 eIF4G eIF4G EIF4G1 CST 2498 103 Elk1 (phospho S383) Elk1_pS383 ELK1 CST 9181 104 Epithelial Membrane Antigen EMA EMA Dako M061329-2 105 ErbB2/HER2 HER2 ERBB2 Lab Vision MS-325-P1 106 ErbB2/HER2 (phospho Y1248) HER2_pY1248 ERBB2 R&D Systems AF1768

107 ErbB3/HER3 HER3 ERBB3 Santa Cruz sc-285 108 ErbB3/HER3 (phospho Y1289) HER3_pY1289 ERBB3 CST 4791 109 ERCC1 ERCC1 ERCC1 Santa Cruz sc-17809 110 ERCC5 ERCC5 ERCC5 ProteinTech 11331-1-AP 111 ERRFI1/MIG6 MIG6 ERRFI1 Sigma-Aldrich WH0054206M1 112 Estrogen Receptor ER ESR1 Lab Vision RM-9101 113 Ets-1 Ets-1 ETS1 Bethyl A303-501A 26 114 FAK FAK PTK2 Abcam ab40794 115 FAK (phospho Y397) FAK_pY397 PTK2 CST 3283 116 Fatty Acid Synthase FASN FASN CST 3180 117 Fibronectin Fibronectin FN1 Epitomics 1574-1 118 FoxM1 FoxM1 FOXM1 CST 5436 119 FoxO3a FoxO3a FOXO3 CST 2497 120 FoxO3a (phospho S318/S321) FoxO3a_pS318_S321 FOXO3 CST 9465 121 FRA-1 FRA-1 FRA1 Santa Cruz sc-605 122 G6PD G6PD G6PD Santa Cruz sc-373887 123 Gab2 Gab2 GAB2 CST 3239 124 GAPDH GAPDH GAPDH Life Technologies AM4300 125 GATA3 GATA3 GATA3 BD Biosciences 558686 126 GCN5L2 GCN5L2 KAT2A CST 3305 127 Glutamate Dehydrogenase1/2 Glutamate-D1-2 GLUD CST 12793 128 Glutaminase Glutaminase GLS Abcam ab156876 129 Glycogen Synthase Gys GYS1 CST 3886 130 Glycogen Synthase (phospho S641) Gys_pS641 GYS1 CST 3891 131 GPBB GPBB PYGB Novus Biologicals NBP1-32799 132 GSK-3alpha/beta GSK-3a-b GSK3A, GSK3B Santa Cruz sc-7291 133 GSK-3alpha/beta (phospho S21/S9) GSK-3a-b_pS21_S9 GSK3A, GSK3B CST 9331 134 H2AX (phospho S140) H2AX_pS140 H2AX Pierce Biotechnology MA1-2022 135 Heregulin Heregulin NRG1 CST 2573 136 HES1 HES1 HES1 CST 11988 137 Hexokinase II Hexokinase-II HK2 CST 2867 138 HIAP HIAP BIRC2 Millipore 07-759 139 Hif-1 alpha Hif-1-alpha HIF1A BD Biosciences 610958 140 Histone H3 Histone-H3 H3F3A, H3F3B Abcam ab1791 141 HSP27 HSP27 HSP27 CST 2402 142 HSP27 (phospho S82) HSP27_pS82 HSBP1 CST 2401 143 HSP70 HSP70 HSP70 CST 4872 27 144 IGF1R (phospho Y1135/Y1136) IGF1R_pY1135_Y1136 IGF1R CST 3024 145 IGFBP2 IGFBP2 IGFBP2 CST 3922 146 IGFBP5 IGFBP5 IGFBP5 Santa Cruz sc-6006 147 IGFRb IGFRb INSR CST 3027 148 INPP4b INPP4b INPP4B CST 4039 149 IRF-1 IRF-1 IRF1 Santa Cruz sc-497 150 IRS1 IRS1 IRS1 Millipore 06-248 151 JAB1 JAB1 COPS5 Santa Cruz sc-13157 152 Jagged1 Jagged1 JAG1 Abcam ab109536 153 Jak2 Jak2 JAK2 CST 3230 154 JNK/SAPK (phospho T183/Y185) JNK_pT183_Y185 MAPK8 CST 4668 155 JNK2 JNK2 MAPK9 CST 4672 156 LC3A/B LC3A-B LC3AB CST 4108 157 Lck Lck LCK CST 2752 158 LDHA LDHA LDHA CST 3582 159 MAPK (phospho T202/Y204) MAPK_pT202_Y204 MAPK1, MAPK3 CST 4377 160 Mcl 1 Mcl-1 MCL1 CST 5453 161 MDM2 (phospho S166) MDM2_pS166 MDM2 CST 3521 162 MEK1 MEK1 MAP2K1 Epitomics 1235-1 163 MEK1 (phospho S217/S221) MEK1_pS217_S221 MAP2K1 CST 9154 MAP2K2 164 MEK2 MEK2 MAP2K2 CST 9125 165 Merlin/NF2 Merlin NF2 Novus Biologicals 22710002 166 MIF MIF MIF Santa Cruz sc-20121 167 Mitochondria Mitochondria MTCO2 Abcam ab3298 168 MMP2 MMP2 MMP2 CST 4022 169 Mnk1 Mnk1 MKNK1 CST 2195 170 Monocarboxylic Acid Transporter 4 MCT4 SLC16A4 Millipore AB3314P 171 MSH2 MSH2 MSH2 CST 2850 172 MSH6 MSH6 MSH6 Novus Biologicals 22030002 28 173 mTOR mTOR MTOR CST 2983 174 mTOR (phospho S2448) mTOR_pS2448 MTOR CST 2971 175 Myosin heavy chain 11 Myosin-11 MYH11 Novus Biologicals 21370002 176 Myosin IIa (phospho S1943) Myosin-IIa_pS1943 MYH9 CST 5026 177 Myt1 Myt1 MYT1 CST 4282 178 NAPSIN A NAPSIN-A NAPSA Abcam ab129189 179 N-Cadherin N-Cadherin CDH2 CST 4061 180 NDRG1 (phospho T346) NDRG1_pT346 NDRG1 CST 3217 181 NDUFB4 NDUFB4 NDUFB4 Abcam ab110243 182 NF-kappaB p65 (phospho S536) NF-kB-p65_pS536 RELA CST 3033 183 Notch1 Notch1 NOTCH1 CST 3268 184 Notch3 Notch3 NOTCH3 Santa Cruz sc-5593 185 N-Ras N-Ras NRAS Santa Cruz sc-31 186 p21 p21 CDKN1A Santa Cruz sc-397 187 p27 KIP 1 p27-Kip-1 CDKN1B Abcam ab32034 188 p27/KIP 1 (phospho T198) p27_pT198 CDKN1B Abcam ab64949 189 p27/Kip1 (phospho T157) p27_pT157 CDKN1B R&D Systems AF1555 190 p38 MAPK p38 MAPK14 CST 9212 191 p38 MAPK (phospho T180/Y182) p38_pT180_Y182 MAPK14 CST 9211 192 p44/42 MAPK p44-42-MAPK MAPK3 CST 4695 193 p53 p53 TP53 CST 9282 194 p70 S6 Kinase (phospho T389) p70-S6K_pT389 RPS6KB1 CST 9205 195 p70/S6K1 p70-S6K1 RPS6KB1 Abcam ab32529 196 p90RSK (phospho T573) p90RSK_pT573 RPS6K CST 9346 197 PAI-1 PAI-1 SERPINE1 BD Biosciences 612024 198 PAR PAR PAR Trevigen 4336-BPC-100 199 PARK7/DJ1 DJ1 PARK7 Abcam ab76008 200 PARP-1 PARP1 PARP1 Santa Cruz sc-7150 201 Paxillin Paxillin PXN Epitomics 1500-1 202 P-Cadherin P-Cadherin CDH3 CST 2130 29 203 PCNA PCNA PCNA Abcam ab29 204 Pdcd-1L1 Pdcd-1L1 CD274 Santa Cruz sc-19090 205 Pdcd4 Pdcd4 PDCD4 Rockland 600-401-965 206 PDGFR beta PDGFR-b PDGFRB CST 3169 207 PDK1 PDK1 PDPK1 CST 3062 208 PDK1 (phospho S241) PDK1_pS241 PDPK1 CST 3061 209 PD-L1 PD-L1 CD274 CST 13684 210 PEA-15 PEA-15 PEA15 CST 2780 211 PED/PEA-15 (phospho S116) PEA-15_pS116 PEA15 Invitrogen 44-836G 212 PI3 Kinase p110 alpha PI3K-p110-a PIK3CA CST 4255 213 PI3K p110 beta PI3K-p110-b PIK3CB Santa Cruz sc-376412 214 PI3K p85 PI3K-p85 PIK3R1 Millipore 06-195 215 PKA RI alpha PKA-a PRKAR1A CST 5675 216 PKC alpha PKC-a PRKCA Millipore 05-154 217 PKC alpha (phospho S657) PKC-a_pS657 PRKCA Millipore 06-822 218 PKC beta II (phospho S660) PKC-b-II_pS660 PRKCA, PRKCB CST 9371 PRKCD, PRKCE PRKCH, PRKCQ 219 PKC delta (phospho S664) PKC-delta_pS664 PRKCD Millipore 07-875 220 PKM2 PKM2 PKM2 CST 4053 221 PLC gamma2 (phospho Y759) PLC-gamma2_pY759 PLCG2 CST 3874 222 PLK1 PLK1 PLK1 CST 4513 223 PMS2 PMS2 PMS2 Novus Biologicals 22510002 224 PRAS40 PRAS40 AKT1S1 Invitrogen AHO1031 225 PRAS40 (phospho T246) PRAS40_pT246 AKT1S1 Life Technologies 441100G 226 PREX1 PREX1 PREX1 Abcam ab102739 227 Progesterone Repector PR PGR Abcam ab32085 228 PTEN PTEN PTEN CST 9552 229 Puma Puma BBC3 CST 4976 230 PYGM PYGM PYGM Novus Biologicals H00005837-M10 30 231 Rab11 Rab11 RAB11A,B CST 3539 232 Rab25 Rab25 RAB25 CST 4314 233 Rad50 Rad50 RAD50 Millipore 05-525 234 Rad51 Rad51 RAD51 CST 8875 235 Raptor Raptor RPTOR CST 2280 236 Rb Rb RB1 CST 9309 237 Rb (phospho S807/S811) Rb_pS807_S811 RB1 CST 9308 238 RBM15 RBM15 RBM15 Novus Biologicals 21390002 239 Rheb Rheb RHEB R&D Systems MAB3426 240 Rictor Rictor RICTOR CST 2114 241 Rictor (phospho T1135) Rictor_pT1135 RICTOR CST 3806 242 Rock-1 Rock-1 ROCK1 Santa Cruz sc-5560 243 RPA32 RPA32 RPA32 CST 2208 244 RPA32 (Phospho S4/S8) RPA32_pS4_S8 RPA32 Bethyl A300-245A 245 RSK RSK RPS6KA1 CST 9347 RPS6KA2 RPS6KA3 246 S6 (phospho S235/S236) S6_pS235_S236 RPS6 CST 2211 247 S6 (phospho S240/S244) S6_pS240_S244 RPS6 CST 2215 248 S6 Ribosomal Protein S6 RPS6 CST 2317 249 SCD SCD SCD Santa Cruz sc-58420 250 SDHA SDHA SDHA CST 11998 251 SF2/ASF SF2 SRSF1 Invitrogen 32-4500 252 Shc (phospho Y317) Shc_pY317 SHC1 CST 2431 253 SHP-2 (phospho Y542) SHP-2_pY542 PTPN11 CST 3751 254 SLC1A5 SLC1A5 SLC1A5 Sigma-Aldrich HPA035240 255 Smac/Diablo Smac DIABLO CST 2954 256 Smad1 Smad1 SMAD1 Abcam ab33902 257 Smad3 Smad3 SMAD3 Abcam ab40854 258 Smad4 Smad4 SMAD4 Santa Cruz sc-7966 31 259 Snail Snail SNAI1 CST 3895 260 SOD2 SOD2 SOD2 CST 13141 261 Sox2 Sox2 SOX2 CST 2748 262 Src Src SRC Millipore 05-184 263 Src (phospho Y527) Src_pY527 SRC, YES1, FYN CST 2105 FGR 264 Src Family (phospho Y416) Src_pY416 SRC, LYN, FYN CST 2101 LCK, YES1, HCK 265 Stat3 Stat3 STAT3 CST 4904 266 Stat3 (phospho Y705) Stat3_pY705 STAT3 CST 9131 267 Stat5a Stat5a STAT5A Abcam ab32043 268 Stathmin 1 Stathmin-1 STMN1 Abcam ab52630 269 Syk Syk SYK Santa Cruz sc-1240 270 Tau Tau TAU Millipore 05-348 271 TAZ TAZ TAZ CST 4883 272 TFAM TFAM TFAM CST 7495 273 TIGAR TIGAR C12ORF5 Abcam ab137573 274 Transferrin Receptor TFRC TFRC Novus Biologicals 22500002 275 Transglutaminase II Transglutaminase TGM2 Lab Vision MS-224-P1 276 TSC1/Hamartin TSC1 TSC1 CST 4906 277 TSC2/Tuberin (phospho T1462) Tuberin_pT1462 TSC2 CST 3617 278 TTF1 TTF1 NKX2-1 Abcam ab76013 279 Tuberin Tuberin TSC2 Abcam ab32554 280 Twist TWIST TWIST2 Santa Cruz sc-81417 281 Tyro3 Tyro3 TYRO3 CST 5585 282 UBAC1 UBAC1 UBAC1 Sigma-Aldrich HPA005651 283 Ubiquityl Histone H2B Ubq-Histone-H2B H2BFM Millipore 05-1312 284 UGT1A UGT1A UGT1A1 Santa Cruz sc-271268 285 VDAC1/Porin Porin VDAC1 Abcam ab14734 286 VEGF Receptor 2 VEGFR-2 KDR CST 2479 32 287 **VHL-EPPK1** VHL-EPPK1 EPPK1 BD Biosciences 556347

288 Vimentin Vimentin VIM Dako M0725 289 Wee1 Wee1 WEE1 CST 4936 290 XBP1 XBP1 XBP1 Santa Cruz sc-32136 291 XIAP XIAP XIAP CST 2042 292 XPA XPA XPA Santa Cruz sc-56813 293 XPF XPF XPF Abcam ab3299 294 XRCC1 XRCC1 XRCC1 CST 2735 295 YAP YAP YAP1 Santa Cruz sc-15407 296 YAP (phospho S127) YAP_pS127 YAP1 CST 4911 297 YB1 YB1 YBX1 Novus Biologicals 17250002 298 YB1 (phospho S102) YB1_pS102 YBX1 CST 2900

33 Western Blot

Protein was isolated from NK cells treated with a non-targeting siRNA pool or AMPKα1/2 siRNA 24 hours post-electroporation. To analyze protein expression, cells were lysed in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl, pH 7.4; 1% NP-40; 0.25% sodium deoxycholate; 150 mM NaCl; 1 mM EDTA) supplemented with a complete protease inhibitor cocktail (Roche, Basel, Switzerland). The same amount of proteins was separated by

12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes using the Transfer Blot Turbo (Biorad, Hercules, California). The primary antibodies used were monoclonal anti-AMPKα and anti–glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (both Cell Signaling, Danvers, MA. The membranes were blotted with appropriate primary antibodies and horseradish peroxidase (HRP)–conjugated secondary antibody to detect the proteins of interest. The anti-mouse and anti-rabbit horseradish peroxidase

(HRP)–conjugated secondary antibodies were from Santa Cruz Biotechnology (Dallas, TX).

NMR

The global metabolic profile of licensed and unlicensed NK cells was examined using NMR.

Ten million NK cells were lysed with 0.5 mL of polymer vortex beads in a 3 mL solution of methanol and water (2:1). The suspended cells were vortexed, flash-frozen in liquid nitrogen, and allowed to thaw in ice. After three vortex–freeze–thaw cycles, the samples were centrifuged for 10 minutes, and the supernatant, containing water-soluble metabolites, was extracted. The supernatant was placed in a lyophilizer overnight where the samples were freeze-dried to purify

2 the metabolites. The metabolites were then dissolved in 600 µL of deuterium oxide ( H2O), 36

µL of PBS, and 4 µL of 80 mM DSS standard (4,4-dimethyl-4-silapentane-1-sulfonic acid).

Deuterium oxide was used for field-locking, PBS was used to control the pH of the solution, and

34 DSS was used as an internal standard with a final concentration of 0.5 mM in 640 µL of solution. NMR sample preparation was completed by placing the sample solutions in NMR tubes.

NMR spectra were obtained using a Bruker AVANCE III HD 500 MHz NMR scanner (Bruker

BioSpin Corporation, The Woodlands, TX). The spectrometer was equipped with a cryogenic temperature probe to improve the sensitivity of the acquisition. A pre-saturation technique was implemented for water suppression. The spectra were obtained using a 90° flip angle, 10,240 Hz bandwidth, 1.09 s acquisition time (16,000 complex points), and 6.0 s scan delay to allow for complete T1 relaxation of the metabolites. A total of 256 scans were collected and averaged for each spectrum. The time domain signal was apodized using an exponential function.

The acquired NMR spectra were phase-corrected to produce Lorentzian peaks, baseline- corrected to reduce noise and flatten the spectra, and reference-corrected to set the DSS reference peak to 0.00 ppm. Peak identification was performed using Chenomx NMR Suite 8.1 software (Chenomx Inc., Edmonton, Canada). The identified peaks were then integrated using

MestReNova software (Mestrelab Research, A Coruña, Spain), and the intensities were converted to metabolite concentrations.

Statistics

Prism was used to calculate p values. The statistical tests used are indicated in the figure legends. For pairwise analysis of licensed and unlicensed cells from the same donor, we used paired t tests. Otherwise, unpaired t tests were used.

Metabolite concentrations were averaged separately for the licensed and unlicensed groups and compared using an unpaired, two-sample, two-tailed t-test with equal variance and correction 35 for multiple comparisons using the Holm-Sidak method. RPPA differences were identified by calculating multiple unpaired t-tests.*p£0.05, **p£0.01, ***p£0.001.

36 Chapter 3: Specific Aims

Specific Aim One: Understanding intrinsic mechanistic differences between licensed and unlicensed NK cells

Unlicensed NK cells are hypo responsive against targets such as cancer cells (Kim et al.,

2005). I hypothesized that licensed NK cells are functional due to the extrinsic interaction of inhibitory KIR with self-HLA leading to an intrinsic signal that occurs during education rendering licensed NK cells functional, able to proliferate, and survive. To establish preliminary data on mechanisms and models of NK cell licensing, I determined the ability of the two subsets cells to expand ex vivo, measured their telomere lengths using mass cytometry coupled with telofish, and identified signaling pathway differences by reverse phase protein array.

Specific Aim Two: Understanding the role of metabolism in licensed and unlicensed NK cell cytotoxicity

Observations from aim one suggested that expanded licensed NK cells might be more metabolically active than unlicensed NK cells. p38 and AMPK signaling after KIR cross-linking was increased in the licensed NK cell subsets. First I determined the engagement of glycolysis and mitochondrial respiration between the two subsets. Next I determined the necessity of mitochondrial respiration and glycolysis pathways for NK cell cytolysis. I then assessed the role of AMPK signaling in NK cell expansion and NK cell cytolysis. Finally, I determined the role of AMPK and p38 signaling in regulating glycolysis in expanded licensed NK cells.

37 Chapter 4: Understanding intrinsic mechanistic differences between licensed and unlicensed NK cells

Rationale

Licensed and unlicensed NK cell subsets respond differently to targets, I hypothesized that the two NK cell subsets might also proliferate in vitro differently as well as engage different signaling pathways that contribute to the heightened sensitivity of licensed NK cells to leukemia targets (Kim et al., 2005). I determined the telomere lengths of licensed and unlicensed NK cells to try and differentiate whether unlicensed NK cells are hypo responsive because they are exhausted due to proliferating without inhibition (“disarming” model) or whether unlicensed

NK cells are hypo responsive and don’t proliferate as quickly because they have never received their educational signal through self-HLA interaction (“arming” model). Understanding the pathways and mechanisms that exist between licensed and unlicensed NK cells could provide a method for exploiting these pathways to license all NK cells, thereby producing more functionally mature NK cells to fight cancer and infections. Additionally, there may be opportunities to turn off NK cell activation is the setting of autoimmune diseases.

Because of the limited number of licensed and unlicensed NK cells that can be isolated from peripheral blood, we employed a novel expansion platform developed in our lab to elucidate the pathway differences responsible for the enhanced cytolytic functions of licensed

NK cells.

38 4.1 Licensed and unlicensed NK cell subset distribution before and after expansion

There are several findings to explain whether KIR acquisition is random or whether KIR expression is skewed by HLA expressed in the individual, however studies have been conflicting (Beziat et al., 2010; Sleiman et al., 2014). If KIR acquisition were random, we would expect to find equal populations of licensed and unlicensed single KIR positive NK cells in the peripheral blood of normal healthy donors. However, one can speculate that there may be other selective forces that determine which NK cell populations persist in vivo. Using flow cytometry, I determined the distribution or percentages of single-KIR-positive licensed or unlicensed NK cells fresh from the peripheral blood and after expansion ex vivo (Figure 1A).

Fresh unexpanded licensed NK cells were more abundant in the peripheral blood than unlicensed NK cells. The percentage of licensed NK cells in the peripheral blood had a mean of

18.45% compared to 10.27% of unlicensed NK cells (Figure 1B). Next, purified NK cells were expanded for 14 days in the presence of irradiated 721.221 leukemia feeder cells (Figure1A).

Upon expansion for 14 days, the single-KIR-positive licensed NK cells remained at the same percentages, whereas unlicensed NK cell percentages decreased to a mean of 5.3% (Figure 1B).

39 Figure One: Distribution of licensed and unlicensed NK cell subsets before and after expansion.

40 Figure One: Distribution of licensed and unlicensed NK cell subsets before and after expansion. Freshly purified NK cells from the peripheral blood of normal donors were expanded for 14 days in the presence of 721.221 feeder cells. Percentages of licensed and unlicensed NK cells were determined by gating single KIR (licensed or unlicensed) positive NK cells by flow cytometry. Mean percent positive of licensed NK cells at day 0, 18.45 and 10.27, day 7, 18.6 and 16.9, and day 14, 20.03 and 5.36 for licensed and unlicensed NK cells respectively. Day 0 unlicensed compared to day 7(p=.032) or day 14(p=.017), respectively.

Licensed single KIR positive cells compared to unlicensed KIR positive cells day 0 (p =.049), day 7(p=.0014) and day 14(p=.00069), respectively. Paired t-tests were used to obtain p-values.

*p£0.05, **p£0.01, ***p£0.001.

41 4.2 Determining the proliferative capacity between licensed and unlicensed NK cells

Licensed and unlicensed NK cell populations have unequal percentage distributions in the peripheral blood and after expansion with 721.221 feeder cells (Figure 1B). This finding suggested that the licensed NK cells are able to proliferate more or faster than unlicensed NK cells both ex vivo and in vivo, but could also be explained by heightened cell death in unlicensed

NK cells. To assess the ability of the licensed and unlicensed NK cells to expand ex vivo, single-

KIR-positive licensed and unlicensed NK cells, along with KIR-negative NK cells were sorted by FACS (Figure 2). NK cells were first purified from buffy coat or whole blood PBMCs. NK cells were then gated and sorted according to the flow chart in Figure 2. KIR-negative NK cells were also included because this population is an immature NK cell population, KIR acquisition occurs later in development in NK cells (Beziat et al., 2010; Bjorkstrom et al., 2010). Sorted

NK cell populations were expanded for 21 days in the presence of the feeder cells K562

Clone9.mbIL21. We determined that licensed NK cells were able to expand to a mean fold expansion of 47,920 whereas unlicensed NK cells were only able to expand to a mean fold expansion of 896.5 after 3 weeks ex vivo p value=.0329 (Figure 3B). Immature KIR-negative

NK cells expanded to a mean fold of 21,200 after 3 weeks of expansion ex vivo (Figure 3B).

Figure Two: Gating strategy for FACS of single KIR+ and KIR- NK cells.

43 Figure Two: Gating strategy for FACS of single KIR+ and KIR- NK cells. One representative donor of cell sorting at day zero before expansion of fresh NK cells for signaling analysis experiments. NK cells were purified from whole blood or buffy coat PBMCs. NK cells were analyzed by flow cytometry for FACS. PBMCs cells were first gated for lymphocytes.

Next cells were gated for CD56+ CD3- NK cells. Next cells were gated for

KIR2DL1+KIR2DL2/3-, these cells were then gated for KIR2DL1 single positive (SP) NK cells.

The KIR2DL1-KIR2DL2/3- NK cells were further sorted into KIR- and KIR3DL1 SP populations. The KIR2DL1-KIR2DL2/3+ NK cells were further sorted into KIR2DL2/3 SP NK cells.

44 Figure Three: Licensed NK cells expand to greater numbers than unlicensed NK cells ex vivo.

45 Figure Three: Licensed NK cells expand to greater numbers than unlicensed NK cells ex vivo. A) Schematic of licensed and unlicensed NK cell sorting and expansion. B) Licensed and unlicensed NK cells were sorted for single KIR positive NK cells either licensed or unlicensed as described in Figure 2. Sorted NK cell populations were expanded for 3 weeks using the feeder cells K562 Clone9.mbIL21. P-value=.0329. Multiple t tests correcting for multiple comparisons were used to obtain p-values. Error bars depict mean +/- SD.

46 Figure Four: Licensed NK cells remain better killers than unlicensed NK cells after expansion for 3 weeks.

Figure Four: Licensed NK cells remain better killers than unlicensed NK cells after expansion for 3 weeks. Cytotoxicity of K562 Clone9.mbIL21 expanded licensed single-KIR- positive NK cells (black box) and expanded unlicensed single-KIR-positive NK cells (grey box) against 721.221 target cells. Effector target ratio is 1.25:1 Data is representative of 3 donors, 5 licensed data points and 4 unlicensed data points. (P value=.0052). Unpaired t tests were used to obtain p values. Error bars are min to max.

47 4.3 Determining the telomere length between licensed and unlicensed NK cells

To attempt to understand the possible mechanism of licensing, telomere lengths were measured in fresh PBMCs from typed healthy normal donors using a novel method that combines telo-fish and cytometry by time-of-flight (CyTOF). The two models of NK cell licensing that we evaluated NK cells for were the arming model and the disarming model. The

NK cell expansion data of the licensed and unlicensed NK cell subsets showed that licensed NK cells proliferated to greater numbers ex vivo than unlicensed NK cells. This evidence suggested that licensed NK cells might also proliferate more in vivo. We hypothesized that the licensed

NK cells, which are more abundant in the peripheral blood and also expand better ex vivo would have shorter telomere lengths than the unlicensed NK cells. Telomeres protect the chromosomes from deterioration. After each cycle of replication and division, the telomeres become shorter.

PBMCs isolated from peripheral blood were first labeled with antibodies using standard mass- cytometry methods. After cell staining, PBMCs undergo DNA denaturation followed by hybridization of the peptide nucleic acid (PNA) probe conjugated to a Cy-5 fluorophore that is specific for the TelC region of the DNA. Lastly, a Cy-5 specific heavy metal conjugated antibody is used to label the telomeres with a CyTOF antibody. A representative donor using spanning-tree progression analysis of density-normalized events (SPADE) analysis was conducted to visualize telomere length from PBMC subsets and is depicted in Figure 5. SPADE utilizes unsupervised clustering and allows a visualization of a branch tree structure of multiple cell types by selecting markers for clustering (Qiu et al., 2011).

The telo-fish mass cytometry assay easily depicts telomere lengths within different

PBMC subsets. In the SPADE tree, red is associated with higher expression and is equated to 48 longer telomere lengths; blue is representative of lower expression and is equated with shorter telomere lengths. The size of the circle is related to the number of cell. For example, the larger the circle, the more cells are represented. To validate the assay, telomere lengths within PBMCs subsets from normal healthy donors were determined. Reasonably, naïve CD4 and CD8 T cell subsets have longer telomeres than their memory counterparts (Figure 5, 6) CD56dim (mature)

NK cells have shorter telomere lengths in comparison to CD56bright (immature) NK cells (Figure

5,6).

Licensed NK cells had a trend of shorter telomere lengths in comparison to unlicensed

NK cell populations isolated from PBMCs from normal healthy donors (Figure 7). What we believe to be more interesting is the fact that unlicensed NK cells did not have shorter telomere lengths. We speculated that if licensed NK cells are more abundant in the peripheral blood because they proliferate more in vivo then the licensed NK cells would have shorter telomeres.

This data is in favor of the “arming” model of NK cell education. However, more replicates will need to be performed to determine whether statistical significance can be achieved.

49 Figure Five: Spade analysis of Telomere Length Comparison between PBMC subsets.

Figure Five: Spade Analysis of Telomere Length Comparison between PBMC subsets.

Representative donor of TelC MMI spade tree of PBMC gated subsets. Central Memory (CM),

Effector Memory (EM). 50 Figure Six: Quantitative Analysis of Telomere Length Comparison between PBMC subsets.

Figure Six: Quantitative Analysis of Telomere Length Comparison between PBMC subsets. Quantified TelC mean metal intensity (MMI) compared between PBMC subsets. TelC

MMI was normalized to B cell to account for donor-to-donor variability. Data is representative of 5 donors. Central Memory (CM), Effector Memory (EM). Error bars depict mean +/- SD.

51 Figure Seven: Licensed NK cells have shorter telomere lengths than unlicensed NK cells.

Figure Seven: Licensed NK cells have shorter telomere lengths than unlicensed NK cells.

Telomere length of licensed and unlicensed single-KIR-positive fresh NK cells from peripheral blood were determined. Data is representative of 3 donors, 5 licensed data points and 4 unlicensed data points. Data is not significant. Error bars depict mean +/- SD.

52 4.4 Determining differences in signaling pathways between licensed and unlicensed

NK cells

The differences in expansion could be attributed to pro-survival pathways being turned on in the licensed and KIR-negative NK cell populations. Or the differences in expansion could be attributed to pro-apoptosis, pro-death pathways that are engaged in the unlicensed NK cell population. To determine what signaling differences could be playing a role in the variation of licensed and unlicensed NK cell populations during expansion and also variations between the two cell subsets and their ability to elicit a cytotoxic response against a tumor target, we performed signaling analysis between the three expanded cell subsets, licensed, unlicensed and

KIR-negative NK cells.

Licensed NK cells have heightened cytolytic responses against tumor targets missing

HLA, compared to unlicensed NK cells. I have shown that licensed NK cells exhibit increased proliferation upon ex vivo expansion with K562 Clone9.mbIL21 feeder cells compared to unlicensed NK cells. Both of these characteristics of licensed NK cells suggest that licensed and unlicensed NK cells engage different signaling pathways that are responsible for differences in ex vivo expansion and NK cell mediated lysis of targets. I hypothesized that licensed NK cells differentially engage signaling pathways compared to unlicensed NK cells at baseline, upon inhibitory stimulation or upon activation. Understanding differences in baseline expression between licensed and unlicensed NK cells, and pathways engaged during inhibitory receptor signaling of licensed NK cells, may provide insight into what pathways to turn on in unlicensed

NK cells, thus releasing the break on NK cell hypo responsiveness. Due to the low number of licensed and unlicensed NK cells that can be isolated from the peripheral blood, licensed, unlicensed and KIR-negative NK cell populations were sorted at day 0 and expanded for 21 53 days as described in 4.2. After expansion, NK cell populations were assessed by flow cytometry for the purity of the populations (Figure 8). All single-KIR-positive or KIR-negative NK cell populations had a purity of greater than 90% to ensure any signaling differences observed between the different subsets could be observed.

Expanded licensed, unlicensed and KIR- NK cell populations were then further divided up into three subgroups each. Signaling at baseline, after activation, or after inhibition were the three subgroups. Signaling through receptor cross-linking was achieved by using antibodies specific for the receptor to first coat the cell receptors, followed by the addition of a secondary antibody to achieve receptor cross-linking to induce signaling. To assess baseline signaling, the baseline subgroups were cross-linked with isotype control antibodies. To assess inhibitory pathways that are engaged between the licensed and unlicensed populations, KIR specific anti-

KIR antibodies were used to induce cross-linking. To assess activation signaling pathways that are engaged between the different populations, the subgroups were cross-linked with anti-

NKp46 antibodies to induce signaling. Protein was then isolated and sent to the reverse phase protein array (RPPA) core at MD Anderson Cancer Center.

Baseline Expression Differences Between Expanded Licensed and Unlicensed NK cells

Many proteins were found to be differentially expressed between the expanded licensed and unlicensed NK cells at baseline. At baseline, expression of proteins involved in cell death or senescence, such as phospho-Bad, Bax, Pdcd4, CDK2 (p16INK4a) and PUMA (Gu et al., 1992;

Jeffers et al., 2003; Howells et al., 2011; Westphal et al., 2011; Zhen et al., 2016), were significantly increased in the expanded unlicensed NK cells compared to the licensed NK cells

(Figure 10). Phospho-PI3K(p85) and annexin VII expression was also increased in the

54 unlicensed NK cells at baseline. By contrast, expression of proteins involved in cell metabolism, such as MIF, MTCO2, SDHB, and PKM2 (King et al., 2006; Gao et al., 2007; Sakuragi et al.,

2007; Gupta and Bamezai, 2010; Yoshikawa et al., 2011), were greater at baseline in the licensed expanded NK cells than in the unlicensed subset (Figure 10). Interestingly, proteins involved in cell cycle arrest and DNA damage such as CDK1, Chk1, phopho-Chk2(pT68) and

FoxM1 (Matsuoka et al., 1998; Geng et al., 2007; Millour et al., 2011; Wang et al., 2013), all had increased expression in the licensed NK cells at baseline (Figure 10). Other proteins with increased expression in the licensed NK cells at baseline were PRAS40, D-a-Tublin, MSH6, and phospho-NDRG1(pT346) (Figure 10). We also determined proteins that were upregulated in

KIR-negative NK cells to have an immature NK cell population to compare the licensed and unlicensed NK cells to. The KIR-negative NK cells likely have NKG2A expression to contribute to their responsiveness. The KIR-negative NK cells protein expression analysis revealed that KIR-negative NK cells had overlapping protein expression changes with both licensed and unlicensed NK cells. However, the majority of protein expression in KIR-negative

NK cells modeled protein expression similar to licensed NK cells (Figure 10).

Protein Expression Differences Between Expanded Licensed and Unlicensed NK Cells

Upon NKp46 Cross-Linking

Following Nkp46 cross-linking to induce an activating signal, the adhesion proteins

CD31 and CD171 were upregulated in both licensed and unlicensed expanded NK cells.

Interestingly, after NKp46 cross-linking, phospho-ATM expression was lower in licensed than unlicensed expanded NK cells (Figure 11). Other proteins that were upregulated in both licensed and unlicensed NK cells after activation through NKp46 were EMA (MUC1), XPF, and HER2.

55 Protein Expression Differences Between Expanded Licensed and Unlicensed NK Cells

Upon KIR Cross-Linking

Two proteins known to affect cellular metabolism, phospho-AMPKα and phospho-p38, was upregulated upon KIR cross-linking in expanded licensed NK cells (Figure 12). Other protein expression that was found to be only significantly increased in licensed NK cells was c-

MET, JAB1, TWIST, Porin, BAP1 and phospho-NF-kB (p65_pS536). Interesting, phospho-NF- kB expression was decreased in the unlicensed NK cells after KIR cross-linking (Figure 12).

Protein Expression Changes That Occur in Both Expanded Licensed and Unlicensed NK

Cells Upon KIR Cross-Linking

Upon KIR cross-linking, numerous proteins were upregulated or downregulated in both licensed and unlicensed expanded NK cells (Figure 13A, B). Proteins involved in cell adhesion that had increased expression after KIR cross-linking in both licensed and unlicensed expanded cell subsets included proteins such as, CD44, CD49b and CD171. Other protein expression that was increased after KIR cross-linking in both licensed and unlicensed NK cells was N-Ras,

Transglutaminase, Smad4, Snail, MSH2, MTOC2, HSP27, UGT1A, PRAS40, E2F1, 14-3-3- epsilon, Src and SCD. Proteins that had decreased expression after KIR cross-linking in both licensed and unlicensed NK cells were YWHAZ, phospho-AKTS1, VTCN1 (B7H4) and

TFAM.

Validation of proteins involved in cellular metabolism

Several proteins involved in cellular metabolism were increased in licensed NK cells either at baseline, PKM2, or after KIR cross-linking, phospho-AMPKα and phospho-p38.

Pyruvate kinase isoform 2 (PKM2) catalyzes the last step of glycolysis. AMP activated protein

56 kinase (AMPK) signaling and p38 mitogen activated protein kinase (p38 MAPK) signaling has been implicated in promoting glycolysis. Before we investigated the role of these proteins in licensed and unlicensed NK cell cytolysis or expansion, we verified the expression differences between the licensed and unlicensed NK cells. PKM2 expression was validated by flow cytometry. Both mean fluorescence intensity (MFI) and percent positive cells were increased in the licensed NK cells compared to the unlicensed NK cells (Figure 14A, B). Additionally, phospho-AMPKα and phospho-p38 expression levels upon KIR cross-linking were validated by phospho-CyTOF. Both phospho-AMPKα and phospho-p38 mean metal intensity (MMI) were increased in the licensed NK cells compared to the unlicensed NK cells.

I hypothesize that cellular metabolism is responsible for some or all of the cytolytic differences observed between licensed and unlicensed NK cell subsets.

57 Figure Eight: NK cell single-KIR-positive populations after sorting on day 0 and expanding for 21 days.

58 Figure Eight: NK cell single-KIR-positive populations after sorting on day 0 and expanding for 21 days. NK cells were sorted by FACS for single-KIR-positive populations at day 0 and expanded for 21 days using the feeder cells K562 Clone9.mbIL21. Data is from one representative donor. After expansion, purity of each single KIR positive population was greater than 90%.

59 Figure Nine: Schematic of experimental workflow to understand signaling pathways utilized by licensed and unlicensed NK cell subsets at baseline (isotype), inhibitory (KIR) and activating (NKp46) induced signaling.

Figure Nine: Schematic of experimental workflow to understand signaling pathways utilized by licensed and unlicensed NK cell subsets at baseline (isotype), inhibitory (KIR) and activating (NKp46) induced signaling. First, freshly isolated NK cells from the peripheral blood were sorted by FACS for KIR-negative, and single-KIR-positive licensed or unlicensed

NK cells. The sorted populations were expanded for 21 days in the presence of the feeder cells

K562 Clone9.mbIL21 to obtain enough cells to perform signaling analysis. To induce signaling, the NK cell subsets were cross-linked with isotype control (baseline), KIR (inhibitory) or

NKp46 (activating) antibodies. Protein was then isolated for reverse phase protein array analysis.

60 Figure Ten: Baseline (isotype) signaling in expanded licensed and unlicensed NK cells.

Figure Ten: Baseline (isotype) signaling in expanded licensed and unlicensed NK cells.

Protein expression of expanded licensed or unlicensed single-KIR-positive NK cells, and expanded KIR-negative NK cells at baseline

(isotype cross-linking). P-values comparing licensed and unlicensed NK cells are as follows p= Annexin-VII=0.012, Bad_pS112=0.019, 61 Bax=0.010, Bcl2=0.004, p16INK4a=0.006, Pdcd4=0.0007, PDK1=0.013, PI3K-p85=0.023, Puma=0.023, CDK1=0.030, Chk1=0.038,

Chk2_pT68=0.049, SDHB=0.007, FoxM1=0.024, MIF=0.0026, MTCO2=0.03, MSH6=0.00005, NDRG1_pT346=0.0012, PKM2=0.007,

PRAS40=0.027. Unpaired t-tests were used to determine the p-values. Error bars are min to max.

62 Figure Eleven: Protein expression in expanded licensed and unlicensed NK cells after

NKp46 cross-linking.

Figure Eleven: Protein expression in expanded licensed and unlicensed NK cells after

NKp46 cross-linking. P-values= CD31=0.003, ATM pS1987=0.0027, EMA=0.0053,

CD44=0.024, XPF=0.010, HER2=0.013, CD171=0.018. Unpaired t-tests were used to determine the p-values. Error bars are min to max.

Figure Twelve: Upregulated protein expression in licensed NK cells upon KIR cross- linking.

Figure Twelve: Upregulated protein expression in licensed NK cells upon KIR cross- linking. P-values= p38_pT180_Y182 .0092, c-MET .0014, JAB1 .017, TWIST .023, Porin

64 .025, BAP1 .032, NFKB1 .04, AMPKα .041. Unpaired t-tests were used to determine the p- values. Error bars are min to max.

65 Figure Thirteen: Upregulated (A) or down regulated (B) protein expression in NK cells upon KIR cross-linking regardless of

licensing.

66 Figure Thirteen: Upregulated (A) or down regulated (B) protein expression in NK cells upon KIR cross-linking regardless of licensing. Unpaired t-tests were used to determine the p-values. All p-values were <0.05. Error bars are min to max.

67 Figure Fourteen: Validation of RRPA by flow cytometry and CyTOF.

68 Figure Fourteen: Validation of RRPA by flow cytometry and CyTOF. Flow cytometry analysis of PKM2 expression. A) Percent positive of PKM2 expression (p=.016) and B) Mean fluorescence intensity (MFI) of PKM2 (p=.0083) expression in expanded licensed single-KIR- positive or expanded unlicensed single-KIR-positive NK cells. Unpaired t tests were used to determine p values. Mass cytometry analysis of phospho-AMPKα and phospho-p38 after KIR- cross-linking in expanded licensed or unlicensed NK cells. C) Mean Metal Intensity (MMI) of phospho-AMPKα (p=.05), D) phospho-p38 (p=.05) Unpaired t test was used to determine p values. Error bars depict mean +/- SD.

69 Chapter 4 Discussion

In mice, skewing of the NK cell receptor repertoire is dependent on MHC-I level expression (Brodin et al., 2012). After expanding NK cells ex vivo, we observed that licensed

NK cells are more abundant before and after expansion suggesting either that the mechanism that determines increased proliferative potential is stable throughout the expansion process or that both trans and cis licensing contribute to human NK cell repertoire skewing. The percentages of licensed NK cells freshly isolated from the peripheral blood is greater than unlicensed NK cells, suggesting that the HLA environment in the individual skews the KIR repertoire and selects for NK cell populations expressing licensed KIR. If KIR expression was stochastic and proliferation was equal, we would expect KIR distribution frequency to be the same between licensed and unlicensed NK cells. Additionally, sorted licensed NK cells expanded to greater numbers ex vivo compared to unlicensed NK cells and increased expression of many pro-apoptotic proteins were observed in the unlicensed NK cell subsets after expansion.

Studies by Brodin et al., observing enhanced survival in educated NK cell subsets in mice also agree with our findings. This suggests that the licensed NK cells might also proliferate in vivo more than unlicensed NK cells, perhaps in the setting of viral infections or tumor occurrence when HLA is down regulated, like in the setting of 721.221 and K562 Clone9.mbIL21 ex vivo expansions.

Sorted licensed NK cells expanded to notably greater numbers than unlicensed NK cells, even though the telomeres of the licensed NK cells were slightly shorter than unlicensed NK cells, but this difference was not statistically significant. Suggesting, differences in ex vivo expansion could not be attributed to NK cell exhaustion of unlicensed cells in vivo resulting in 70 shorter telomeres. Although licensed NK cells are inhibited or disarmed via KIR, we propose that NK cells require an arming signal (arming model) to function, rather than a disarming signal, because they proliferate more and have shorter telomeres than unlicensed cells. We would expect to see unlicensed NK cells with shorter telomere lengths due to over stimulation in vivo if the disarming model is true. KIR-negative NK cells were also able to expand considerably ex vivo, likely due to NKG2A expression, because NKG2A expression precedes

KIR expression and is also thought to contribute to NK cell education (Braud et al., 1998;

Yawata et al., 2008; Beziat et al., 2010; Bjorkstrom et al., 2010; Lisovsky et al., 2015). We did not verify NKG2A expression as this was not the main focus of the study, our interest in using

KIR-negative NK cells stemmed from using an immature NK cell population to compare with the licensed and unlicensed NK cell expansion.

We propose that the arming signal through the inhibitory KIR endows the licensed NK cells with proliferative potential, allowing them to expand to greater numbers than unlicensed

NK cells. Further studies are needed to clarify whether licensed NK cells and NKG2A-positive,

KIR-negative NK cells engage the same metabolic signaling pathways.

At baseline, many proteins involved in cell death and senescence were upregulated in the unlicensed KIR-positive cells. This increase in cell death and senescence related proteins could explain the marked decrease in expansion capabilities of the unlicensed NK cells. This provides further evidence that an arming signal is needed for an NK cell to proliferate as well as assume full cytolytic effector functions. Interestingly, proteins involved in cell cycle arrest have increased expression at baseline in the expanded licensed NK cells, which is surprising as this subset is able to proliferate well. This could be explained by the decrease in phospho-ATM

71 expression in expanded licensed NK cells upon NKp46 cross-linking. ATM interacts with

CHK1/2, CDK1, alpha-tubulin and FoxM1 (proteins that had increased expression in licensed

NK cells at baseline) (Matsuoka et al., 1998; Geng et al., 2007; Millour et al., 2011; Wang et al., 2013), and a decrease in ATM signaling could prevent cell death.

We also characterized KIR-negative NK cell protein expression at baseline. KIR- negative NK cells are immature developmentally compared to KIR-positive NK cells. KIR- negative NK cells likely have expression of NKG2A. NKG2A can also contribute to licensing and NK cell responsiveness (Ivarsson et al., 2013; Lisovsky et al., 2015). My project did not specifically look at the role of NKG2A in NK cell education, however the KIR-negative NK cells are a good comparison to use to compare to licensed and unlicensed NK cells because they do not express KIR and they are at an earlier developmental stage and are not exhausted. There were some overlapping similarities in protein expression from the KIR-negative NK cells with both licensed and unlicensed NK cells. At baseline protein expression levels that were similar between unlicensed NK cells and the KIR-negative NK cells were Annexin-VII, Pdcd4, PDK1, phospho-PI3K and Puma. However, at baseline, there was an overwhelming similarity between the licensed and KIR-negative NK cell populations. The proteins that had similar expression level patterns between the licensed and KIR-negative NK cells were phospho-Bad, Bax, p16INK4a, CDK1, Chk1, SDHB, D-α-tubulin, FoxM1, Mif, MTCO2, MSH6, phospho-

NDRG1, PKM2 and PRAS40. The great parallels between licensed and KIR-negative NK cells suggests that the inhibitory signal through KIR and NKG2A provides an important function that allows both KIR-negative and licensed NK cells to proliferate more ex vivo.

72 A greater number of proteins were differentially expressed in the expanded licensed and unlicensed NK cells after KIR cross-linking than after NKp46 cross-linking. Previously, KIR signaling could only be attributed to signaling through SHP1, 2. However, after evaluating protein expression changes after KIR cross-linking, it can be appreciated that KIR signaling contributes to many pathways outside of SHP-1, 2. Phospho-AMPK-α and phospho-p38 were upregulated in licensed NK cells after KIR cross-linking, and both of these pathways are known to play a role in stress and metabolic responses. We proposed that the changes in protein expression after KIR cross-linking might aid in the “arming” of licensed NK cells to achieve a heightened cytolytic response to targets that lack HLA. The increase in phospho-AMPK-α and phospho-p38 led us to delve deeper into the metabolic pathways engaged by the licensed and unlicensed NK cell subsets.

Some proteins had similar expression patterns after both KIR and NKp46 cross-linking such as CD44 and CD171. Both of these proteins are involved in cell-to-cell interactions and cell adhesion. This suggests that NK receptor engagement alone can increase expression of proteins involved in adhesion that will aid the NK cell in forming a tight synapse with the target cell and achieve NK cell mediated cytolysis. Protein expression levels that increased in both licensed and unlicensed NK cells after NKp46 cross-linking included EMA (MUC1), XPF and

HER2, and the role of these proteins have not been explored in NK cells. Many of the pathway differences that we have described were in expanded NK cells, which are in a heightened activation state compared to NK cells freshly isolated from peripheral blood. We believe this model system is applicable to fresh NK cells because, after expansion, differences in cytotoxicity were still observed between expanded licensed and unlicensed NK cells, as has been shown in freshly isolated NK cells. Our findings open up many new opportunities for

73 understanding NK cell activation and inhibition signaling pathways that could provide targetable treatment options to improve NK cell immunotherapy.

74 Chapter 5: Understanding the role of metabolism in licensed and unlicensed NK cell cytotoxicity

Rationale

Preliminary results from aim one revealed that p-AMPKα and p-p38 levels were increased in licensed NK cells compared to unlicensed NK cells upon KIR cross-linking (Figure

12, 14 C, D). AMPK and p38 signaling are both involved in sensing stress and have been show to increase glycolysis. At baseline, PKM2 levels were increased in licensed NK cells (Figure 10,

14 A, B). PKM2 catalyzes the last rate-limiting step to convert phosphoenolpyruvate into pyruvate. All of these proteins are involved in the glycolysis pathway, suggesting that glycolysis may be different in the two subsets. AMPK is activated when AMP levels are high indicating a need for increased energy production. Based on these data, we hypothesized that KIR signaling during development leads to increased AMPK, which enhances glycolysis and thereby enables increased lytic function in licensed cells. Licensing can be measured by performing cytotoxicity or degranulation assays against 721.221 leukemia target cells as a surrogate for cytotoxicity.

Licensed NK cells have greater cytotoxicity and degranulation compared to unlicensed NK cells. Studies observing T cell metabolism have provided evidence that metabolism can be responsible for differences in effector function. We further investigated the role of metabolism in the licensed and unlicensed NK cells. We also determined the role AMPK and p38 signaling play in regulating glycolysis in licensed NK cells.

75 5.1 Determining the effect AMPK signaling has on NK cell cytotoxicity

The increase in phospho-AMPKα levels after KIR cross-linking in licensed NK cells suggested that AMPK signaling may be playing a role in the heightened responsiveness licensed

NK cells have against their targets. Using the pharmacological AMPK agonist, Metformin, and

AMPK inhibitor Compound C, I determined whether AMPK activation or inhibition altered licensed or unlicensed NK cell cytotoxicity against 721.221 targets. I hypothesized that AMPK activation in the unlicensed NK cell subsets will enhance their cytolytic ability against 721.221 targets. However, activation of AMPK with Metformin did not lead to a significant increase in

721.221 cell lysis in either licensed or unlicensed NK cells (Figure 15). Inhibition of AMPK with Compound C prevented all cytolytic ability in both licensed and unlicensed NK cells

(Figure 15).

76 Figure Fifteen: The effect of AMPK activation or inhibition on licensed and unlicensed NK cell cytotoxicity.

Short-Term AMPK Inhibition/Activation Cytotoxicity Against 721.221 Targets 100 Complete Media 80 Metformin 2mM Compound C 10uM 60

40 % Lysis

Licensed Unlicensed

Figure Fifteen: The effect of AMPK activation or inhibition on licensed and unlicensed NK cell cytotoxicity. Licensed and unlicensed NK cells were incubated with 2mM Metformin or

10uM Compound C for 3 hours prior to starting the cytotoxicity assay. NK cells were washed and co-cultured with 721.221 targets at a 10:1 effector target ratio for 4 hours. Percent lysis of

721.221 targets by licensed or unlicensed NK cells is shown. Error bars represent mean +/- min to max.

77 5.2 Determining the effect AMPK signaling has on NK cell expansion

AMPK signaling in licensed and unlicensed NK cells was investigated for its role in NK cell expansion. As shown above in Figure 3, the licensed NK cells expanded to greater numbers than unlicensed NK cells. To determine whether the increased expansion abilities of the licensed

NK cells was attributed to AMPK signaling, licensed and unlicensed NK cells were sorted and then expanded as described in Figure 3A, with the addition of Metformin or Compound C, to determine the effect AMPK activation or inhibition, respectively, has on NK cell expansion.

Opposing to my hypothesis, activating AMPK signaling with Metformin resulted in decreased expansion of both licensed and unlicensed NK cell subsets. This suggested that AMPK plays an important role in NK cell proliferation, because inhibiting AMPK signaling killed NK cells, however prolonged activation of AMPK was also detrimental to NK cell expansion. It is understandable that prolonged activation of AMPK prevents NK cell expansion because AMPK is involved in stress responses. Stressing the NK cells for a long period of time was harmful to

NK cell proliferation.

78 Figure Sixteen: The effect of AMPK activation or inhibition on licensed and unlicensed

NK cell expansion.

Figure Sixteen: The effect of AMPK activation or inhibition on licensed and unlicensed

NK cell expansion. Licensed and unlicensed NK cells were sorted by FACS at day zero. Sorted

NK cells were expanded for 21 days with the feeder cells K562 Clone9.mbIL21 in the presence of 2mM Metformin or 10uM Compound C. NK cell fold expansion with no drug, Metformin, or

Compound C. Licensed NK cells are shown with a solid line, unlicensed NK cells are shown with a dashed line. No drug, circle; Metformin, square; Compound C, triangle.

79 5.3 Evaluation of licensed and unlicensed NK cell subsets utilization of glycolysis and mitochondrial respiration

The upregulation of PKM2, phospho-AMPKα and phospho-p38 in expanded licensed

NK cells at baseline led us to investigate metabolic pathways in the two subsets. Using a

Seahorse XFe96 analyzer (Agilent Technologies) that measures extracellular acidification rate

(ECAR) and oxygen consumption rate (OCR), we determined the extent to which licensed and unlicensed NK cells depend on glycolysis and mitochondrial respiration. We performed the mito stress test and the glycolysis stress tests to observe licensed and unlicensed NK cell’s ability to engage mitochondrial respiration and glycolysis, respectively, under stress conditions. ECAR is largely determined by the rate of glycolysis and OCR is an indicator of mitochondrial respiration. First we assessed mitochondrial respiration and observed that there was no significant difference in OCR between expanded licensed and unlicensed NK cells (Figure 17).

Next we measured ECAR to calculate glycolysis, glycolytic capacity and glycolytic reserve.

Glycolysis in the seahorse assay is a measurement of glycolysis in a cell after a saturating amount of glucose is added. Glycolytic capacity refers to the maximum glycolytic rate after mitochondrial respiration is inhibited. Glycolytic reserve is representative of the maximum glycolytic function of a cell and is calculated by glycolytic capacity-glycolysis. Expanded licensed NK cells exhibited greater glycolysis, higher glycolytic capacity and greater glycolytic reserve than unlicensed expanded NK cells (Figure 18).

Since our metabolism studies were conducted in expanded licensed and unlicensed NK cells, we wanted to determine whether differences in glycolysis could also be observed in fresh

NK cells isolated from the peripheral blood of normal healthy donors. We assessed glycolysis in fresh licensed and unlicensed NK cells. First we determined glycolytic capacity and glycolytic 80 reserve in unstimulated fresh licensed and unlicensed NK cells. We observed a trend toward lower glycolysis in unlicensed NK cells compared to licensed NK cells, albeit not significant

(Figure 19). We also stimulated the licensed and unlicensed NK cells during the glycolysis stress test assay with particles made from K562 Clone9.mbIL21 by our collaborators in Florida

(Oyer et al., 2016). Stimulation with the particles increased the rate of glycolysis, however there was not a significant difference in the rate of glycolysis between the licensed and unlicensed NK cell subsets (Figure 20). For this experiment, only two donors were used due to donor availability, further experiments will need to include more donors in order to determine whether the observed increase in glycolysis in licensed NK cells is true in fresh human NK cells.

81 Figure Seventeen: Expanded licensed versus unlicensed NK cell assessment of mitochondrial respiration.

Figure Seventeen: Expanded licensed versus unlicensed NK cell assessment of mitochondrial respiration.

A) Mitochondrial Respiration, one representative donor. B) Basal Respiration and spare respiratory capacity calculated for 3 donors (5 licensed populations, 4 unlicensed populations). Error bars measure standard deviation.

82 Figure Eighteen: Expanded licensed versus unlicensed NK cell assessment of glycolysis.

Figure Eighteen: Expanded licensed versus unlicensed NK cell assessment of glycolysis. A) Glycolytic function, one representative donor. B) Glycolysis (p=.026), glycolytic capacity (p=.0019), and glycolytic reserve (p=.036) calculated for 3 donors (5 licensed populations, 4 unlicensed populations). Multiple t tests (unpaired) were used to determine p values. Errors bars measure min to max.

83 Figure Nineteen: Assessment of glycolysis in fresh NK cells.

A Fresh NK Glycolytic Capacity B Fresh NK Glycolytic Reserve

25 20

20 15 15 10 10

5 ECAR (mpH/min) ECAR 5 (mpH/min) ECAR

0 0 Licensed Unlicensed Licensed Unlicensed

Figure Nineteen: Assessment of glycolysis in fresh NK cells.

NK cells were purified from peripheral blood from typed KIR/HLA normal donors. NK cells were sorted by FACS for single KIR positive licensed or unlicensed NK cell populations A)

Fresh NK cell glycolytic capacity. B) Fresh NK cell glycolytic reserve. Error bars measure min to max.

84 Figure Twenty: Assessment of glycolysis in fresh NK cells after stimulation with PM21.

Figure Twenty: Assessment of glycolysis in fresh NK cells after stimulation with PM21.

Glycolytic capacity and glycolytic reserve were determined after Stimulating the freshly isolated licensed and unlicensed NK cell populations with PM21. Data is representative of two donors.

Error bars measure min to max.

85 5.4 Determining the role of glycolysis and OXPHOS in licensed and unlicensed NK cell cytolytic effector function

Licensed NK cells achieved greater cytolysis of 721.221 leukemia targets (Figure 4).

Metabolism has been shown to play a role in T cell effector functions, so we next investigated whether the increase in cytolysis of licensed NK cell targets could be attributed to the increase in glycolysis. Using metabolic pathway inhibitors, we investigated whether licensed and unlicensed NK cells depend on glycolysis or mitochondrial respiration for cytotoxic effector functions. In order to achieve inhibition of the different metabolic pathways in the NK cells without also inhibiting the metabolic pathways in the tumor target, which might change the sensitivity of the targets to NK cell mediated cytolysis, NK cells were pre-incubated in the presence inhibitors specific to glycolysis or OXPHOS pathways. The drug was then removed during the assay to prevent any off-target effects on the tumor cells.

Pre-incubating expanded licensed or unlicensed NK cells in the presence of the glycolysis inhibitor 2-DG did not affect NK cell cytotoxicity against 721.221 targets (Figure

21). 2-DG is a glucose analog that inhibits the enzyme hexokinase, the enzyme that catalyzes the first step during glycolysis. After expansion, licensed NK cells were pre-incubated in the presence of oligomycin, an oxidative phosphorylation inhibitor; licensed NK cells had a modest reduction (mean, 31%) in cytotoxicity (Figure 21). However, when unlicensed expanded NK cells were pre-incubated with oligomycin, a striking reduction in unlicensed NK cell cytotoxicity was observed (mean, 87.8%) (Figure 21). Oligomycin is an ATP synthase inhibitor that blocks oxidative phosphorylation by blocking ATP synthase’s proton channel. 2-DG and oligomycin pretreatment in combination led to mean reductions in cytotoxicity of 31.3% and

95% in expanded licensed and unlicensed NK cells, respectively (Figure 21). Only when 2-DG 86 was combined with glucose-free medium and oligomycin to pretreat the licensed NK cells did we see a large reduction (mean, 88.1%) in licensed NK cell cytotoxicity to 721.221 target cells

(Figure 21). Pre-treating licensed and unlicensed NK cells with an AMPK agonist (AICAR) did not affect NK cell cytotoxicity against 721.221 targets (Figure 21). These findings suggest that licensed NK cells utilize both mitochondrial respiration and glycolysis to elicit a cytolytic response against leukemia targets, however unlicensed NK cells are dependent upon mitochondrial respiration to perform cytolysis. Additionally, activating AMPK signaling for a short amount of time did not improve unlicensed NK cell cytolysis. The differences in the observed cytolysis between licensed and unlicensed NK cells were partially if not completely attributed to differences in aerobic glycolysis.

87 Figure Twenty-One: Inhibition of metabolic pathways in NK cells to determine metabolic

pathways necessary for NK cell mediated lysis.

88 Figure Twenty-One: Inhibition of metabolic pathways in NK cells to determine metabolic pathways necessary for NK cell mediated lysis. Single-KIR-positive expanded NK cells were pre-treated for 3 hours with the following metabolic inhibitors: Vehicle, 2mM 2DG, 40nM

Oligomycin, 2mM+40nM Oligomycin, RPMI Glucose Free Media+ 2mM 2DG+ 40nM

Oligomycin, 500uM AICAR. NK cells were then incubated with 721.221 target cells at a 2.5:1 effector target ratio for 4 hours. Circles represent licensed single-KIR-positive expanded NK cells, squares represent unlicensed single-KIR-positive expanded NK cells. Unpaired t tests were used to determine p values of licensed vehicle control compared to GF+2DG+Oligo p- value=.0008, Oligomycin treated licensed and unlicensed comparison p=0.00017,

2DG+Oligomycin treated licensed NK cells compared to unlicensed NK cells p=0.000012.

Error bars represent min to max.

89 5.5 Determining the role of AMPKα and p38 signaling in NK cell glucose metabolism

AMPK signaling is known to play a role in promoting glycolysis, and was increased in licensed NK cells upon KIR cross-linking. I hypothesized that AMPK signaling would lead to an increase in glycolysis and that this signaling could be responsible for the increase in cytotoxicity of licensed expanded NK cells (Hardie, 2011; Wu and Wei, 2012). We explored the contribution of AMPKα and p38 signaling to the increased glycolysis in expanded licensed NK cells. AMPK protein expression was knocked down in expanded licensed NK cells using

AMPKα1/2 siRNA (Figure 22). Knockdown of AMPKα1/2 did not affect expanded licensed

NK cell cytotoxicity. Knockdown of AMPKα1/2 resulted in a decrease in glycolytic capacity and glycolytic reserve in expanded licensed NK cells, but this decrease was not statistically significant (Figure 22B). p38 protein expression was knocked down in licensed NK cells, using the same method as AMPK knockdown, with p38 siRNA (Figure 23). By contrast, whereas knockdown of p38 again did not change cytotoxicity of licensed NK cells, it did increase glycolytic capacity (Figure 23B). With these results, it is unclear if AMPK signaling upon KIR cross-linking is upstream of regulating glycolysis. However, p38 signaling downstream of KIR signaling negatively regulates glycolysis. This suggests that KIR signaling is important for tightly regulating glycolysis in licensed NK cells.

Figure Twenty-Two: Knockdown of AMPKα1/2 in licensed NK cells to determine the role

AMPKα plays in NK cell glycolysis.

Figure Twenty-Two: Knockdown of AMPKα1/2 in licensed NK cells to determine the role

AMPKα plays in NK cell glycolysis. A) Knockdown of AMPK in expanded licensed NK cells using AMPK siRNA after 24 hours. AMPKα protein expression was determined by Western

Blot. B) Glycolytic function of expanded licensed NK cells after AMPK knockdown. Circles represent control (non-targeting siRNA) and squares represent AMPKα1/2 siRNA.

91 Figure Twenty-Three: Knockdown of p38 in expanded licensed NK cells to determine the role p38 plays in NK cell glycolysis.

Figure Twenty-Three: Knockdown of p38 in expanded licensed NK cells to determine the role p38 plays in NK cell glycolysis. A) Knockdown of p38 in expanded licensed NK cells after 24 hours. p38 expression was determined by flow cytometry. B) Glycolytic function of expanded licensed NK cells after p38 knockdown, circles are control no siRNA, squares are

2uM p38siRNA, unpaired t test was used to generate a p=.037.

92 5.6 Comparison of intracellular metabolites in expanded licensed and unlicensed

NK cells

We hypothesized that intracellular metabolites would provide us with further insight into which metabolic pathways were engaged by licensed and unlicensed NK cells. Therefor we explored intracellular metabolites that were upregulated in licensed and unlicensed expanded

NK cells using nuclear magnetic resonance (NMR) in collaboration with Dr. Pratip

Bhattacharya’s lab (MD Anderson Cancer Center).

Using NMR, we identified a trend toward higher (not statistically significantly) acetate and lactate concentrations in expanded licensed NK cells than in the unlicensed cells (Figure

24). Conversely, higher intracellular concentrations of glutamate, aspartate, and taurine were found in expanded unlicensed NK cells than in the licensed subset (Figure 24). The differences observed in intracellular metabolite concentrations by NMR suggests that unlicensed NK cells may utilize different metabolic pathways that may hinder or alter their cytolysis.

93 Figure Twenty-Four: Comparison of intracellular metabolites found in expanded licensed and unlicensed single-KIR-positive NK cells.

94 Figure Twenty-Four: Comparison of intracellular metabolites found in expanded licensed and unlicensed single-KIR-positive NK cells. Mean concentration of various metabolites quantified by NMR. P-values were determined by calculating standard error (SE) of licensed and unlicensed NK cells. 3 Donors were used in these experiments and three replicates of each donor were performed using NMR. P-values are as follows, glutamate p=.012, aspartate p=

.047, taurine p= .041. NMR experiments were performed with the help of Travis Salzillo, in the laboratory of Dr. Pratip Bhattacharya.

95 Chapter 5 Discussion

Mitochondrial respiration is considered an efficient, yet slow, way to produce energy.

Aerobic glycolysis is a quick and inefficient, method of energy production but also generates biosynthetic substrates needed during proliferation. Mitochondrial respiration was similar between the expanded licensed and unlicensed NK cells; however, glycolysis, glycolytic capacity and glycolytic reserve were significantly greater in the expanded licensed NK cell population than in unlicensed cells. These findings suggest that the metabolic pathway responsible for the heighted cytotoxic response of licensed NK cells is not due to mitochondrial respiration, as both subsets utilized this pathway similarly. However, mitochondrial respiration was still important for both licensed and unlicensed NK cell cytolysis. This was shown by the decreased cytolysis of leukemia targets by licensed NK cells and the even more apparent large decrease of the unlicensed NK cell cytolysis after OXPHOS was inhibited with oligomycin.

One explanation of the increase in glycolysis between the licensed and unlicensed NK cells could be the increase in PKM2 expression observed at baseline in the licensed NK cells.

PKM2 catalyzes the last rate-limiting reaction in glycolysis to produce pyruvate. This increase was interesting because others have reported that resting NK cells have low levels of glycolysis and mitochondrial respiration, but that upon activation, glycolysis and mitochondrial respiration increase (Donnelly et al., 2014; Marcais et al., 2014; Keppel et al., 2015). Our results indicated that unlicensed NK cells, even after activation through expansion ex vivo, couldn’t increase glycolysis to the high levels observed by licensed NK cells. Examining glycolysis and other metabolic pathways in specific cell subsets is an important distinction, as bulk analysis of NK cell metabolism may therefore miss what is occurring within specific NK cell subsets. Until

96 now, notable intrinsic differences in licensed and unlicensed NK cells, other than their function, have not been described.

The heightened glycolytic state of the expanded licensed NK cells compared with the unlicensed NK cells suggested that increased aerobic glycolysis might explain, or at least contribute to, the greater sensitivity of licensed NK cells to tumor targets lacking HLA. Upon inhibition of glycolysis with 2-DG, we did not see a reduction in cytotoxicity against parental cells to either the licensed or unlicensed expanded NK cells. The finding that licensed and unlicensed NK cells were highly cytolytic even when glycolysis was inhibited for a short period of time was encouraging because the tumor microenvironment will likely have less glucose available, as many tumors are highly glycolytic (Chang et al., 2015; Kouidhi et al., 2016). This has important implications for NK cell adoptive therapy for solid tumors because NK cells could be given in multiple doses to overcome any inhibition that prolonged exposure to low glucose in the tumor microenvironment. However, upon inhibition of mitochondrial respiration, we observed a modest reduction in the cytotoxicity of licensed NK cells and an almost complete inhibition of cytotoxicity in the unlicensed NK cell subset against parental cells. Inhibition of both mitochondrial respiration and glycolysis did not affect licensed expanded NK cell cytolytic function, but it did inhibit the remaining cytotoxicity of unlicensed NK cells after oxidative phosphorylation inhibition alone. Only after inhibiting glycolysis by pre-treating licensed NK cells for 3 hours in the presence of two glycolytic pathway inhibitors and an oxidative phosphorylation inhibitor did we observe an abolition of the expanded licensed NK cell cytotoxicity against 721.221 cells. Having to use two glycolytic pathway inhibitors in combination to abolish licensed NK cell cytotoxicity suggested that the expanded licensed NK cells are in a highly glycolytic state. High levels of glycolysis can be used to generate lots of

97 ATP very quickly and can be used for generating building blocks for the cell that might be necessary for NK cell cytolytic function. We propose that increased glycolysis was responsible for the increase in licensed NK cell cytotoxicity compared to unlicensed NK cells. Inhibition of mitochondrial respiration eliminated the ability of the unlicensed NK cells to kill their targets, we propose that the increase in glycolysis in the licensed NK cells is a redundant energy producing pathway that equips licensed NK cells to have heightened cytolysis.

The signaling subunit of AMPK, AMPKα, did not appear to play a role in NK cell glycolysis. After activating AMPK signaling with Metformin or AICAR, licensed and unlicensed NK cells did not increase their cytolytic abilities against 721.221 cells. However,

AMPK signaling does appear to be important for NK cell cytolytic ability, because inhibition of

AMPK signaling with Compound C resulted in complete inhibition of both licensed and unlicensed NK cells to lyse their targets. AMPK signaling also appears to be important for NK cell expansion, as inhibition of AMPK signaling with Compound C resulted in complete inhibition of licensed and unlicensed NK cell expansion. Interestingly, constitutive activation of

AMPK signaling with Metformin led to a decrease in both licensed and unlicensed NK cell expansion. This was likely due to the cells staying in a stressed state for too long, AMPK is involved in stress signaling and prolonged activation was detrimental to NK cell expansion. It is possible that AMPK could still play a role in glycolysis and licensed NK cell cytotoxicity, but that the effects may not be seen after knockdown for just 24 hours. This is a limitation of this model system as knockdown can only be achieved temporarily.

The function of p38 in regulating glycolysis known to vary between cell types (Jaswal et al., 2007; Kalender et al., 2011; Bolanos, 2013; Liu et al., 2015). Most commonly, p38

98 functions to promote glycolysis. In contrast to its common role in other cell types, we demonstrated that p38 is a negative regulator of glycolysis in NK cells. Signaling through KIR increased p38 signaling, which to our surprise inhibited glycolysis. Future studies are needed to further explore how p38 is regulating glycolysis in licensed NK cells. Licensed NK cells readily upregulate proteins that increase glycolysis, such as PKM2 and phospho-AMPK, but also increase phospho-p38 expression, a negative regulator of glycolysis. This expression pattern suggests that glycolysis is tightly regulated in licensed NK cells, adding to the importance of the glycolytic pathway in these cells. Future directions will include investigating the consequence of too much glycolysis in licensed NK cells to better understand why glycolysis is tightly regulated as it pertains to NK cell effector functions and proliferation.

Using NMR, increased acetate concentrations, although not significant, were observed in expanded licensed single KIR positive NK cells compared to the unlicensed NK cells (Figure

24). Increased concentrations of acetate has been linked to improved CD8+ T cell memory responses by increasing glycolytic activity in CD8+ memory T cells (Balmer et al., 2016). The increased concentration of acetate further supports the highly glycolytic nature of the expanded licensed single KIR positive NK cells. To definitively determine the role of acetate in our system further studies will be necessary. Increased concentrations (not significant) of lactate were found in the licensed expanded NK cells (Figure 24). The increased levels of lactate were in agreement with the increased glycolysis that we found in the licensed NK cells compared to the expanded unlicensed NK cells.

Higher glutamate concentrations were found in the expanded unlicensed single KIR positive NK cells compared to licensed NK cells. Increased glutamate concentrations have been

99 associated with inhibiting T cell PHA or CD3/CD28 induced proliferation (Lombardi et al.,

2004; Pacheco et al., 2004). Glutamate may also be playing the same inhibitory role in the unlicensed NK cell proliferation as the unlicensed NK cells had substantially lower proliferation than licensed NK cell expansion ex vivo. Higher concentrations of glutamate and aspartate in expanded unlicensed single-KIR-positive NK cells could suggest that unlicensed NK cells depend upon the glutaminolysis pathway for proliferation and cytolytic function (Figure 24).

Expanded unlicensed single-KIR-positive NK cells had higher intracellular concentrations of the metabolite taurine compared to licensed NK cells. Taurine has been implicated in regulating T cell size and effector responses (Bachmann, 2012; Kaesler et al.,

2012). Additionally, taurine is an antioxidant and might be increased in the unlicensed NK cells, which rely heavily on mitochondrial respiration (Schaffer et al., 2014). Mitochondrial respiration is known to be a source of reactive oxygen species. Further studies are needed to examine the role of taurine in NK cells, we can speculate that taurine acts as an antioxidant in

NK cells, however validation studies need to be performed. Further experiments will need to be performed to validate the NMR findings; we did not perform validations because we were focusing mainly on glycolysis and mitochondrial respiration between licensed and unlicensed

NK cells. Gluconate levels were higher (not significant) in the licensed NK cells than the unlicensed NK cells. Gluconate metabolism is well characterized in prokaryotes, however the role of metabolized gluconate in humans has not been determined (Rohatgi et al., 2014).

Gluconate is a derivative of glucose and has been demonstrated to be an antioxidant (Rohatgi et al., 2014). Future studies looking at the role of gluconate and NK cell function are necessary to determine if gluconate metabolism plays a role in NK cell education.

100 We identified novel findings between licensed and unlicensed NK cells that have not previously been described, however we acknowledge that there are limitations of the study. For instance, determining the role of AMPK and p38 signaling by performing transient knockdowns may not best reflect the actual role of these signaling molecules since the knockdown could only be achieved temporarily and the functions of these molecules might only be observed after a more stable knockdown. Further investigations into the effector function of licensed and unlicensed NK cell subsets in vivo are required, particularly in the cancer microenvironment where metabolite availability and pH levels will be quite different to those in the in vitro setting.

101 Summary and Future Directions

We found that licensed and unlicensed NK cells are different in several ways that have not been previously appreciated. We observed that the licensed NK cells expand to greater numbers than the unlicensed NK cells ex vivo. Additionally, licensed NK cells have slightly shorter telomeres than unlicensed NK cells. This suggests an intrinsic ability in licensed NK cells of greater proliferation in vivo in agreement with increased proliferation during our ex vivo expansion. The underlying mechanisms that are responsible for heightened effector function in licensed NK have remained unknown until now. One reason for this is the limited number of licensed and unlicensed NK cells that can be obtained from the peripheral blood. By expanding licensed and unlicensed NK cell populations ex vivo, I determined that licensed and unlicensed

NK cells differentially engage a variety of signaling pathways. Of importance, unlicensed NK cells had increased expression of proteins involved in pro-apoptotic and pro-cell death signaling pathways at baseline. Conversely, expanded licensed NK cells had increased expression of proteins involved in cellular metabolism at baseline, suggesting that the licensed NK cells are more metabolically active. I investigated the cellular metabolism pathways engaged by licensed and unlicensed NK cells. I found that licensed NK cells had higher levels of glycolysis than expanded unlicensed NK cells. This is contradictory to what was previously shown, that upon activation NK cells upregulate glycolytic machinery and increase glycolysis (Donnelly et al.,

2014; Keppel et al., 2015). In my findings, only the licensed NK cells could increase the rate of aerobic glycolysis. Mitochondrial respiration was found to be the same in both expanded licensed and unlicensed NK cell subsets.

102 Next, the role of mitochondrial respiration and glycolysis in NK cell mediated lysis of

721.221 leukemia targets was determined by using inhibitors of the two pathways. Inhibition of mitochondrial respiration with oligomycin resulted in loss of the expanded unlicensed NK cells to lyse their targets, however there was only a minor suppression of expanded licensed NK cell mediated lysis of the 721.221 targets. Only after inhibiting the licensed NK cells with both a mitochondrial respiration inhibitor and two separate glycolysis inhibitors (2-DG with glucose free media) was a drastic decrease in NK cell mediated lysis inhibition observed. The difficulty inhibiting the licensed NK cells cytolysis was encouraging as using licensed NK cells in solid tumors might be able to withstand low nutrient conditions in the tumor microenvironment.

My data raises several new questions that will need to be investigated in the field of NK cell education in regards to metabolism. The first question that needs to be answered is, what are the downstream signaling proteins after KIR signaling that leads to an increase in glycolysis in the licensed NK cells. It needs to be investigated whether KIR signaling keeps licensed NK cell inhibited allowing licensed NK cells to increase their glycolytic machinery or whether KIR signaling leads to an upregulation of glycolytic machinery thereby increasing glycolysis.

Another important question that was outside the scope of my project is to determine the importance of aerobic glycolysis for NK cell killing. Does the aerobic glycolysis lead to a quick increase in energy production that allows the licensed NK cells to elicit a quick and powerful cytolytic response or is it that during glycolysis biosynthetic precursors, such as amino acids and nucleotides, are being produced allowing for the production of perforin and granzyme B.

Another question that needs to be addressed is whether licensed or unlicensed expanded NK cells will be inhibited by the tumor microenvironment. Tumors are known to deplete exogenous levels of glucose, I assessed whether short term glycolysis inhibition had an effect on licensed

103 and unlicensed NK cell cytotoxicity, however determining how longterm decreased glycolysis would affect licensed and unlicensed NK cells is important when using licensed NK cells clinically for the treatment of solid tumors. Additionally, since tumors are highly glycolytic, the concentration of lactic acid in the tumor microenvironment will likely be high. Lactate concentrations within the tumor microenvironment can reach up to 40mM (Walenta et al.,

2000). Furthermore, one study found that treatment of NK cells with lactate impaired NK cells cytotoxicity (Husain et al., 2013). Another study demonstrated that tumors with lactate dehydrogenase (Ldha) knocked down was associated with improved cytolytic function in NK cells (Husain et al., 2013). Further studies are necessary to determine the effect of lactic acid on licensed and unlicensed NK cell effector functions.

104 Figure Twenty-Five: Summary of licensed and unlicensed NK cell characteristics.

105 Figure Twenty-Five: Summary of licensed and unlicensed NK cell characteristics. Licensed NK cells have increased proliferation ex vivo compared to unlicensed NK cells. Licensed NK cells have increased expression of phospho-AMPKa and phospho- p38 upon KIR cross-linking. P38 signaling negatively regulates glycolysis signaling. At baseline, PKM2 expression is increased in expanded licensed NK cells. Expanded licensed NK cells have increased glycolysis compared to unlicensed NK cells. Mitochondrial respiration is similar between licensed and unlicensed expanded NK cells.

106 Additionally, further evaluation of fresh licensed and unlicensed NK cell glycolysis engagement will need to be performed to definitively determine whether or not fresh licensed

NK cells also exhibit heightened glycolysis. Differences will likely only be seen after activation of the licensed and unlicensed NK cells in vitro. Unstimulated freshly isolated NK cells, along with other lymphocytes have low levels of metabolism and differences between subsets have been difficult to observe without additional stimulation. In T cell subsets, differences in metabolism were only observed after activating the T cells in vitro (Guppy et al., 1993;

Frauwirth et al., 2002; Jacobs et al., 2008; Buck et al., 2015). One major obstacle is overcoming limited numbers of licensed and unlicensed NK cells obtained from the donor for characterization. One solution could be using apheresis products; however, it will be difficult to justify such drastic measures for experimental purposes because individuals will still need to be

KIR and HLA typed to determine which NK cell cells are licensed. Additionally, it is important to determine whether or not human unlicensed NK cells can achieve greater cytolytic properties by turning on glycolysis in the unlicensed NK cells. Evidence from haploidentical HSCT studies and mouse models studying NK cell education suggest that licensing is dynamic and can be altered. Future studies will need to determine whether over expression of PKM2, or inducing factors that regulate increased expression of PKM2 could rescue unlicensed NK cell cytolytic deficiency. Or perhaps, other pathways that have yet to be discovered play a greater role in NK cell mediated cytotoxicity.

In conclusion, we have identified differences between licensed and unlicensed NK cells in expansion, telomere lengths, and signaling pathways, including large differences in glycolysis. We believe this is the first study to clearly demonstrate the mechanistic differences responsible for the different cytolytic functions of licensed and unlicensed human NK cells.

107 These important findings will pave the way for further investigations into NK cell education and artificially increasing the responsiveness of unlicensed NK cells. Our results have broad and significant implications, both for basic NK cell biology and for the clinical use of NK cells in adoptive therapy. We propose these metabolic findings explain the differences in cytolytic function between licensed and unlicensed NK cells. Our findings are highly relevant to the clinic, as clinical trials of expanded NK cell infusion for the treatment of various tumor types are ongoing. Our finding that licensed NK cells are highly glycolytic is important, because solid tumors are also highly glycolytic and the two cell types may compete for nutrients.

Further human studies regarding selecting donors with more certain combinations of licensed and unlicensed NK cells and considering KIR mismatch when selecting NK cell donors will be important when therapeutics are used that may interfere with tumor and or immune cell metabolism. Additionally, our findings provide many potential paths to further explore differences in apoptosis, senescence, DNA damage and, potentially, alternative metabolic pathways that may also contribute to licensed and unlicensed NK cell function. Using apoptosis inhibitors during NK cell expansion with unlicensed NK cells may provide evidence of why unlicensed NK cells are unable to proliferate as well. Additionally, determining what factors regulate the pro-apoptotic proteins in unlicensed NK cells could provide insight into why unlicensed NK cells cannot expand ex vivo as well as licensed NK cells.

Differential intracellular metabolites such as glutamate, aspartate and taurine in unlicensed NK cells need to be investigated further for their role in dampening cytotoxicity in unlicensed NK cells. Unlicensed NK cells appeared to depend solely on mitochondrial respiration for their cytotoxicity; however, licensed NK cells use both mitochondrial respiration and glycolysis for their cytotoxicity. Other pathways such as the tricarboxylic acid cycle (TCA)

108 were not investigated in this present study. The TCA cycle could also be mediating metabolism differences between the two NK cell subsets that contributes to differences in their effector functions.

Overall we have described novel and targetable differences between licensed and unlicensed NK cell metabolism that affects NK cell cytotoxicity.

109 Bibliography

Anfossi, N., P. Andre, S. Guia, C.S. Falk, S. Roetynck, C.A. Stewart, V. Breso, C. Frassati, D. Reviron, D. Middleton, F. Romagne, S. Ugolini and E. Vivier, 2006. Human nk cell education by inhibitory receptors for mhc class i. Immunity, 25(2): 331-342. Available from http://www.ncbi.nlm.nih.gov/pubmed/16901727. DOI 10.1016/j.immuni.2006.06.013. Aquino-Lopez, A., V.V. Senyukov, Z. Vlasic, E.S. Kleinerman and D.A. Lee, 2017. Interferon gamma induces changes in natural killer (nk) cell ligand expression and alters nk cell-mediated lysis of pediatric cancer cell lines. Frontiers in immunology, 8: 391. Available from http://www.ncbi.nlm.nih.gov/pubmed/28428785. DOI 10.3389/fimmu.2017.00391. Bachmann, M.F., 2012. Taurine: Energy drink for t cells. European journal of immunology, 42(4): 819-821. Available from http://www.ncbi.nlm.nih.gov/pubmed/22531908. DOI 10.1002/eji.201242450. Balmer, M.L., E.H. Ma, G.R. Bantug, J. Grahlert, S. Pfister, T. Glatter, A. Jauch, S. Dimeloe, E. Slack, P. Dehio, M.A. Krzyzaniak, C.G. King, A.V. Burgener, M. Fischer, L. Develioglu, R. Belle, M. Recher, W.V. Bonilla, A.J. Macpherson, S. Hapfelmeier, R.G. Jones and C. Hess, 2016. Memory cd8(+) t cells require increased concentrations of acetate induced by stress for optimal function. Immunity, 44(6): 1312-1324. Available from http://www.ncbi.nlm.nih.gov/pubmed/27212436. DOI 10.1016/j.immuni.2016.03.016. Barnes, K., J.C. Ingram, O.H. Porras, L.F. Barros, E.R. Hudson, L.G. Fryer, F. Foufelle, D. Carling, D.G. Hardie and S.A. Baldwin, 2002. Activation of glut1 by metabolic and osmotic stress: Potential involvement of amp-activated protein kinase (ampk). Journal of cell science, 115(Pt 11): 2433-2442. Available from http://www.ncbi.nlm.nih.gov/pubmed/12006627. Bendall, S.C., E.F. Simonds, P. Qiu, A.D. Amir el, P.O. Krutzik, R. Finck, R.V. Bruggner, R. Melamed, A. Trejo, O.I. Ornatsky, R.S. Balderas, S.K. Plevritis, K. Sachs, D. Pe'er, S.D. Tanner and G.P. Nolan, 2011. Single-cell mass cytometry of differential immune and drug responses across a human hematopoietic continuum. Science, 332(6030): 687-696. Available from http://www.ncbi.nlm.nih.gov/pubmed/21551058. DOI 10.1126/science.1198704. Bessoles, S., G.S. Angelov, J. Back, G. Leclercq, E. Vivier and W. Held, 2013. Education of murine nk cells requires both cis and trans recognition of mhc class i molecules. Journal of immunology, 191(10): 5044-5051. Available from http://www.ncbi.nlm.nih.gov/pubmed/24098052. DOI 10.4049/jimmunol.1301971. Beziat, V., B. Descours, C. Parizot, P. Debre and V. Vieillard, 2010. Nk cell terminal differentiation: Correlated stepwise decrease of nkg2a and acquisition of kirs. PloS one, 5(8): e11966. Available from http://www.ncbi.nlm.nih.gov/pubmed/20700504. DOI 10.1371/journal.pone.0011966. Beziat, V., J.A. Traherne, L.L. Liu, J. Jayaraman, M. Enqvist, S. Larsson, J. Trowsdale and K.J. Malmberg, 2013. Influence of kir gene copy number on natural killer cell education. Blood, 121(23): 4703-4707. Available from http://www.ncbi.nlm.nih.gov/pubmed/23637128. DOI 10.1182/blood-2012-10- 461442.

110 Bjorkstrom, N.K., P. Riese, F. Heuts, S. Andersson, C. Fauriat, M.A. Ivarsson, A.T. Bjorklund, M. Flodstrom-Tullberg, J. Michaelsson, M.E. Rottenberg, C.A. Guzman, H.G. Ljunggren and K.J. Malmberg, 2010. Expression patterns of nkg2a, kir, and cd57 define a process of cd56dim nk-cell differentiation uncoupled from nk-cell education. Blood, 116(19): 3853-3864. Available from http://www.ncbi.nlm.nih.gov/pubmed/20696944. DOI 10.1182/blood-2010-04- 281675. Bolanos, J.P., 2013. Adapting glycolysis to cancer cell proliferation: The mapk pathway focuses on pfkfb3. The Biochemical journal, 452(3): e7-9. Available from http://www.ncbi.nlm.nih.gov/pubmed/23725459. DOI 10.1042/BJ20130560. Braud, V.M., D.S. Allan, C.A. O'Callaghan, K. Soderstrom, A. D'Andrea, G.S. Ogg, S. Lazetic, N.T. Young, J.I. Bell, J.H. Phillips, L.L. Lanier and A.J. McMichael, 1998. Hla-e binds to natural killer cell receptors cd94/nkg2a, b and c. Nature, 391(6669): 795-799. Available from http://www.ncbi.nlm.nih.gov/pubmed/9486650. DOI 10.1038/35869. Brodin, P., K. Karre and P. Hoglund, 2009. Nk cell education: Not an on-off switch but a tunable rheostat. Trends in immunology, 30(4): 143-149. Available from http://www.ncbi.nlm.nih.gov/pubmed/19282243. DOI 10.1016/j.it.2009.01.006. Brodin, P., T. Lakshmikanth, S. Johansson, K. Karre and P. Hoglund, 2009. The strength of inhibitory input during education quantitatively tunes the functional responsiveness of individual natural killer cells. Blood, 113(11): 2434-2441. Available from http://www.ncbi.nlm.nih.gov/pubmed/18974374. DOI 10.1182/blood- 2008-05-156836. Brodin, P., T. Lakshmikanth, K. Karre and P. Hoglund, 2012. Skewing of the nk cell repertoire by mhc class i via quantitatively controlled enrichment and contraction of specific ly49 subsets. Journal of immunology, 188(5): 2218-2226. Available from http://www.ncbi.nlm.nih.gov/pubmed/22287714. DOI 10.4049/jimmunol.1102801. Brodin, P., T. Lakshmikanth, R. Mehr, M.H. Johansson, A.D. Duru, A. Achour, M. Salmon- Divon, K. Karre, P. Hoglund and S. Johansson, 2010. Natural killer cell tolerance persists despite significant reduction of self mhc class i on normal target cells in mice. PloS one, 5(10). Available from http://www.ncbi.nlm.nih.gov/pubmed/20957233. DOI 10.1371/journal.pone.0013174. Buck, M.D., D. O'Sullivan and E.L. Pearce, 2015. T cell metabolism drives immunity. The Journal of experimental medicine, 212(9): 1345-1360. Available from http://www.ncbi.nlm.nih.gov/pubmed/26261266. DOI 10.1084/jem.20151159. Burns, L.J., D.J. Weisdorf, T.E. DeFor, D.H. Vesole, T.L. Repka, B.R. Blazar, S.R. Burger, A. Panoskaltsis-Mortari, C.A. Keever-Taylor, M.J. Zhang and J.S. Miller, 2003. Il-2-based immunotherapy after autologous transplantation for lymphoma and breast cancer induces immune activation and cytokine release: A phase i/ii trial. Bone marrow transplantation, 32(2): 177-186. Available from http://www.ncbi.nlm.nih.gov/pubmed/12838283. DOI 10.1038/sj.bmt.1704086. Chang, C.H., J. Qiu, D. O'Sullivan, M.D. Buck, T. Noguchi, J.D. Curtis, Q. Chen, M. Gindin, M.M. Gubin, G.J. van der Windt, E. Tonc, R.D. Schreiber, E.J. Pearce and E.L. Pearce, 2015. Metabolic competition in the tumor microenvironment is a driver of cancer progression. Cell, 162(6): 1229-1241. Available from http://www.ncbi.nlm.nih.gov/pubmed/26321679. DOI 10.1016/j.cell.2015.08.016. Charoudeh, H.N., L. Schmied, A. Gonzalez, G. Terszowski, K. Czaja, K. Schmitter, L. Infanti, A. Buser and M. Stern, 2012. Quantity of hla-c surface expression and licensing of 111 kir2dl+ natural killer cells. Immunogenetics, 64(10): 739-745. Available from http://www.ncbi.nlm.nih.gov/pubmed/22772778. DOI 10.1007/s00251-012-0633-1. Cheent, K. and S.I. Khakoo, 2009. Natural killer cells: Integrating diversity with function. Immunology, 126(4): 449-457. Available from http://www.ncbi.nlm.nih.gov/pubmed/19278418. DOI 10.1111/j.1365- 2567.2009.03045.x. Coudert, J.D., L. Scarpellino, F. Gros, E. Vivier and W. Held, 2008. Sustained nkg2d engagement induces cross-tolerance of multiple distinct nk cell activation pathways. Blood, 111(7): 3571-3578. Available from http://www.ncbi.nlm.nih.gov/pubmed/18198346. DOI 10.1182/blood-2007-07- 100057. Demanet, C., A. Mulder, V. Deneys, M.J. Worsham, P. Maes, F.H. Claas and S. Ferrone, 2004. Down-regulation of hla-a and hla-bw6, but not hla-bw4, allospecificities in leukemic cells: An escape mechanism from ctl and nk attack? Blood, 103(8): 3122-3130. Available from http://www.ncbi.nlm.nih.gov/pubmed/15070694. DOI 10.1182/blood- 2003-07-2500. Denman, C.J., V.V. Senyukov, S.S. Somanchi, P.V. Phatarpekar, L.M. Kopp, J.L. Johnson, H. Singh, L. Hurton, S.N. Maiti, M.H. Huls, R.E. Champlin, L.J. Cooper and D.A. Lee, 2012. Membrane-bound il-21 promotes sustained ex vivo proliferation of human natural killer cells. PloS one, 7(1): e30264. Available from http://www.ncbi.nlm.nih.gov/pubmed/22279576. DOI 10.1371/journal.pone.0030264. Domenech, E., C. Maestre, L. Esteban-Martinez, D. Partida, R. Pascual, G. Fernandez- Miranda, E. Seco, R. Campos-Olivas, M. Perez, D. Megias, K. Allen, M. Lopez, A.K. Saha, G. Velasco, E. Rial, R. Mendez, P. Boya, M. Salazar-Roa and M. Malumbres, 2015. Ampk and pfkfb3 mediate glycolysis and survival in response to mitophagy during mitotic arrest. Nature cell biology, 17(10): 1304-1316. Available from http://www.ncbi.nlm.nih.gov/pubmed/26322680. DOI 10.1038/ncb3231. Donnelly, R.P., R.M. Loftus, S.E. Keating, K.T. Liou, C.A. Biron, C.M. Gardiner and D.K. Finlay, 2014. Mtorc1-dependent metabolic reprogramming is a prerequisite for nk cell effector function. Journal of immunology, 193(9): 4477-4484. Available from http://www.ncbi.nlm.nih.gov/pubmed/25261477. DOI 10.4049/jimmunol.1401558. Ebihara, T., A.H. Jonsson and W.M. Yokoyama, 2013. Natural killer cell licensing in mice with inducible expression of mhc class i. Proceedings of the National Academy of Sciences of the United States of America, 110(45): E4232-4237. Available from http://www.ncbi.nlm.nih.gov/pubmed/24145414. DOI 10.1073/pnas.1318255110. Elliott, J.M., J.A. Wahle and W.M. Yokoyama, 2010. Mhc class i-deficient natural killer cells acquire a licensed phenotype after transfer into an mhc class i-sufficient environment. J Exp Med, 207(10): 2073-2079. Available from https://www.ncbi.nlm.nih.gov/pubmed/20819924. DOI 10.1084/jem.20100986. Fauriat, C., M.A. Ivarsson, H.G. Ljunggren, K.J. Malmberg and J. Michaelsson, 2010. Education of human natural killer cells by activating killer cell immunoglobulin-like receptors. Blood, 115(6): 1166-1174. Available from http://www.ncbi.nlm.nih.gov/pubmed/19903900. DOI 10.1182/blood-2009-09- 245746. Fernandez, N.C., E. Treiner, R.E. Vance, A.M. Jamieson, S. Lemieux and D.H. Raulet, 2005. A subset of natural killer cells achieves self-tolerance without expressing inhibitory receptors specific for self-mhc molecules. Blood, 105(11): 4416-4423. Available 112 from http://www.ncbi.nlm.nih.gov/pubmed/15728129. DOI 10.1182/blood-2004-08- 3156. Frauwirth, K.A., J.L. Riley, M.H. Harris, R.V. Parry, J.C. Rathmell, D.R. Plas, R.L. Elstrom, C.H. June and C.B. Thompson, 2002. The cd28 signaling pathway regulates glucose metabolism. Immunity, 16(6): 769-777. Available from http://www.ncbi.nlm.nih.gov/pubmed/12121659. Gao, L., C. Flores, S. Fan-Ma, E.J. Miller, J. Moitra, L. Moreno, R. Wadgaonkar, B. Simon, R. Brower, J. Sevransky, R.M. Tuder, J.P. Maloney, M. Moss, C. Shanholtz, C.R. Yates, G.U. Meduri, S.Q. Ye, K.C. Barnes and J.G. Garcia, 2007. Macrophage migration inhibitory factor in acute lung injury: Expression, biomarker, and associations. Translational research : the journal of laboratory and clinical medicine, 150(1): 18-29. Available from http://www.ncbi.nlm.nih.gov/pubmed/17585860. DOI 10.1016/j.trsl.2007.02.007. Geng, L., X. Zhang, S. Zheng and R.J. Legerski, 2007. Artemis links atm to g2/m checkpoint recovery via regulation of cdk1-cyclin b. Molecular and cellular biology, 27(7): 2625-2635. Available from http://www.ncbi.nlm.nih.gov/pubmed/17242184. DOI 10.1128/MCB.02072-06. Gerriets, V.A. and J.C. Rathmell, 2012. Metabolic pathways in t cell fate and function. Trends in immunology, 33(4): 168-173. Available from http://www.ncbi.nlm.nih.gov/pubmed/22342741. DOI 10.1016/j.it.2012.01.010. Gu, Y., J. Rosenblatt and D.O. Morgan, 1992. Cell cycle regulation of cdk2 activity by phosphorylation of thr160 and tyr15. The EMBO journal, 11(11): 3995-4005. Available from http://www.ncbi.nlm.nih.gov/pubmed/1396589. Guppy, M., E. Greiner and K. Brand, 1993. The role of the crabtree effect and an endogenous fuel in the energy metabolism of resting and proliferating thymocytes. European journal of biochemistry, 212(1): 95-99. Available from http://www.ncbi.nlm.nih.gov/pubmed/8444168. Gupta, V. and R.N. Bamezai, 2010. Human pyruvate kinase m2: A multifunctional protein. Protein science : a publication of the Protein Society, 19(11): 2031-2044. Available from http://www.ncbi.nlm.nih.gov/pubmed/20857498. DOI 10.1002/pro.505. Hardie, D.G., 2011. Amp-activated protein kinase: A cellular energy sensor with a key role in metabolic disorders and in cancer. Biochemical Society transactions, 39(1): 1-13. Available from http://www.ncbi.nlm.nih.gov/pubmed/21265739. DOI 10.1042/BST0390001. Hayakawa, Y., J.M. Kelly, J.A. Westwood, P.K. Darcy, A. Diefenbach, D. Raulet and M.J. Smyth, 2002. Cutting edge: Tumor rejection mediated by nkg2d receptor-ligand interaction is dependent upon perforin. Journal of immunology, 169(10): 5377-5381. Available from https://www.ncbi.nlm.nih.gov/pubmed/12421908. Hayakawa, Y. and M.J. Smyth, 2006. Innate immune recognition and suppression of tumors. Adv Cancer Res, 95: 293-322. Available from https://www.ncbi.nlm.nih.gov/pubmed/16860661. DOI 10.1016/S0065- 230X(06)95008-8. He, Y. and Z. Tian, 2017. Nk cell education via nonclassical mhc and non-mhc ligands. Cellular & molecular immunology, 14(4): 321-330. Available from http://www.ncbi.nlm.nih.gov/pubmed/27264685. DOI 10.1038/cmi.2016.26.

113 Hoglund, P. and P. Brodin, 2010. Current perspectives of natural killer cell education by mhc class i molecules. Nature reviews. Immunology, 10(10): 724-734. Available from http://www.ncbi.nlm.nih.gov/pubmed/20818413. DOI 10.1038/nri2835. Howells, C.C., W.T. Baumann, D.C. Samuels and C.V. Finkielstein, 2011. The bcl-2-associated death promoter (bad) lowers the threshold at which the bcl-2-interacting domain death agonist (bid) triggers mitochondria disintegration. Journal of theoretical biology, 271(1): 114-123. Available from http://www.ncbi.nlm.nih.gov/pubmed/21130780. DOI 10.1016/j.jtbi.2010.11.040. Husain, Z., Y. Huang, P. Seth and V.P. Sukhatme, 2013. Tumor-derived lactate modifies antitumor immune response: Effect on myeloid-derived suppressor cells and nk cells. Journal of immunology, 191(3): 1486-1495. Available from http://www.ncbi.nlm.nih.gov/pubmed/23817426. DOI 10.4049/jimmunol.1202702. Ivarsson, M.A., L. Loh, N. Marquardt, E. Kekalainen, L. Berglin, N.K. Bjorkstrom, M. Westgren, D.F. Nixon and J. Michaelsson, 2013. Differentiation and functional regulation of human fetal nk cells. J Clin Invest, 123(9): 3889-3901. Available from https://www.ncbi.nlm.nih.gov/pubmed/23945237. DOI 10.1172/JCI68989. Jacobs, S.R., C.E. Herman, N.J. Maciver, J.A. Wofford, H.L. Wieman, J.J. Hammen and J.C. Rathmell, 2008. Glucose uptake is limiting in t cell activation and requires cd28- mediated akt-dependent and independent pathways. Journal of immunology, 180(7): 4476-4486. Available from http://www.ncbi.nlm.nih.gov/pubmed/18354169. Jaswal, J.S., M. Gandhi, B.A. Finegan, J.R. Dyck and A.S. Clanachan, 2007. P38 mitogen- activated protein kinase mediates adenosine-induced alterations in myocardial glucose utilization via 5'-amp-activated protein kinase. American journal of physiology. Heart and circulatory physiology, 292(4): H1978-1985. Available from http://www.ncbi.nlm.nih.gov/pubmed/17172269. DOI 10.1152/ajpheart.01121.2006. Jeffers, J.R., E. Parganas, Y. Lee, C. Yang, J. Wang, J. Brennan, K.H. MacLean, J. Han, T. Chittenden, J.N. Ihle, P.J. McKinnon, J.L. Cleveland and G.P. Zambetti, 2003. Puma is an essential mediator of p53-dependent and -independent apoptotic pathways. Cancer cell, 4(4): 321-328. Available from http://www.ncbi.nlm.nih.gov/pubmed/14585359. Joncker, N.T., N.C. Fernandez, E. Treiner, E. Vivier and D.H. Raulet, 2009. Nk cell responsiveness is tuned commensurate with the number of inhibitory receptors for self-mhc class i: The rheostat model. Journal of immunology, 182(8): 4572-4580. Available from http://www.ncbi.nlm.nih.gov/pubmed/19342631. DOI 10.4049/jimmunol.0803900. Jonsson, A.H., L. Yang, S. Kim, S.M. Taffner and W.M. Yokoyama, 2010. Effects of mhc class i alleles on licensing of ly49a+ nk cells. Journal of immunology, 184(7): 3424-3432. Available from http://www.ncbi.nlm.nih.gov/pubmed/20194719. DOI 10.4049/jimmunol.0904057. Kadri, N., A.K. Wagner, S. Ganesan, K. Karre, S. Wickstrom, M.H. Johansson and P. Hoglund, 2016. Dynamic regulation of nk cell responsiveness. Current topics in microbiology and immunology, 395: 95-114. Available from http://www.ncbi.nlm.nih.gov/pubmed/26658943. DOI 10.1007/82_2015_485. Kaesler, S., M. Sobiesiak, M. Kneilling, T. Volz, W.E. Kempf, P.A. Lang, K.S. Lang, T. Wieder, B. Heller-Stilb, U. Warskulat, D. Haussinger, F. Lang and T. Biedermann, 2012. Effective t-cell recall responses require the taurine transporter taut. European

114 journal of immunology, 42(4): 831-841. Available from http://www.ncbi.nlm.nih.gov/pubmed/22531910. DOI 10.1002/eji.201141690. Kalender, A., A. Selvaraj and G. Thomas, 2011. A matter of energy stress: P38beta meets mtorc1. Cell research, 21(6): 859-861. Available from http://www.ncbi.nlm.nih.gov/pubmed/21483449. DOI 10.1038/cr.2011.65. Keating, S.E., V. Zaiatz-Bittencourt, R.M. Loftus, C. Keane, K. Brennan, D.K. Finlay and C.M. Gardiner, 2016. Metabolic reprogramming supports ifn-gamma production by cd56bright nk cells. Journal of immunology, 196(6): 2552-2560. Available from http://www.ncbi.nlm.nih.gov/pubmed/26873994. DOI 10.4049/jimmunol.1501783. Keppel, M.P., N. Saucier, A.Y. Mah, T.P. Vogel and M.A. Cooper, 2015. Activation-specific metabolic requirements for nk cell ifn-gamma production. Journal of immunology, 194(4): 1954-1962. Available from http://www.ncbi.nlm.nih.gov/pubmed/25595780. DOI 10.4049/jimmunol.1402099. Kim, S., K. Iizuka, H.L. Aguila, I.L. Weissman and W.M. Yokoyama, 2000. In vivo natural killer cell activities revealed by natural killer cell-deficient mice. Proceedings of the National Academy of Sciences of the United States of America, 97(6): 2731-2736. Available from https://www.ncbi.nlm.nih.gov/pubmed/10694580. DOI 10.1073/pnas.050588297. Kim, S., J. Poursine-Laurent, S.M. Truscott, L. Lybarger, Y.J. Song, L. Yang, A.R. French, J.B. Sunwoo, S. Lemieux, T.H. Hansen and W.M. Yokoyama, 2005. Licensing of natural killer cells by host major histocompatibility complex class i molecules. Nature, 436(7051): 709-713. Available from http://www.ncbi.nlm.nih.gov/pubmed/16079848. DOI 10.1038/nature03847. King, A., M.A. Selak and E. Gottlieb, 2006. Succinate dehydrogenase and fumarate hydratase: Linking mitochondrial dysfunction and cancer. Oncogene, 25(34): 4675- 4682. Available from http://www.ncbi.nlm.nih.gov/pubmed/16892081. DOI 10.1038/sj.onc.1209594. Kishton, R.J., C.E. Barnes, A.G. Nichols, S. Cohen, V.A. Gerriets, P.J. Siska, A.N. Macintyre, P. Goraksha-Hicks, A.A. de Cubas, T. Liu, M.O. Warmoes, E.D. Abel, A.E. Yeoh, T.R. Gershon, W.K. Rathmell, K.L. Richards, J.W. Locasale and J.C. Rathmell, 2016. Ampk is essential to balance glycolysis and mitochondrial metabolism to control t-all cell stress and survival. Cell metabolism, 23(4): 649-662. Available from http://www.ncbi.nlm.nih.gov/pubmed/27076078. DOI 10.1016/j.cmet.2016.03.008. Kouidhi, S., M.Z. Noman, C. Kieda, A.B. Elgaaied and S. Chouaib, 2016. Intrinsic and tumor microenvironment-induced metabolism adaptations of t cells and impact on their differentiation and function. Frontiers in immunology, 7: 114. Available from http://www.ncbi.nlm.nih.gov/pubmed/27066006. DOI 10.3389/fimmu.2016.00114. Lanier, L.L., 2008. Up on the tightrope: Natural killer cell activation and inhibition. Nature immunology, 9(5): 495-502. Available from http://www.ncbi.nlm.nih.gov/pubmed/18425106. DOI 10.1038/ni1581. Lisovsky, I., G. Isitman, R. Song, S. DaFonseca, A. Tremblay-McLean, B. Lebouche, J.P. Routy, J. Bruneau and N.F. Bernard, 2015. A higher frequency of nkg2a+ than of nkg2a- nk cells responds to autologous hiv-infected cd4 cells irrespective of whether or not they coexpress kir3dl1. Journal of virology, 89(19): 9909-9919. Available from http://www.ncbi.nlm.nih.gov/pubmed/26202228. DOI 10.1128/JVI.01546-15. Liu, J., D. Wen, X. Fang, X. Wang, T. Liu and J. Zhu, 2015. P38mapk signaling enhances glycolysis through the up-regulation of the glucose transporter glut-4 in gastric 115 cancer cells. Cellular physiology and biochemistry : international journal of experimental cellular physiology, biochemistry, and pharmacology, 36(1): 155-165. Available from http://www.ncbi.nlm.nih.gov/pubmed/25925563. DOI 10.1159/000374060. Ljunggren, H.G. and K. Karre, 1985. Host resistance directed selectively against h-2- deficient lymphoma variants. Analysis of the mechanism. The Journal of experimental medicine, 162(6): 1745-1759. Available from http://www.ncbi.nlm.nih.gov/pubmed/3877776. Lombardi, G., G. Miglio, C. Dianzani, R. Mesturini, F. Varsaldi, A. Chiocchetti, U. Dianzani and R. Fantozzi, 2004. Glutamate modulation of human lymphocyte growth: In vitro studies. Biochemical and biophysical research communications, 318(2): 496-502. Available from http://www.ncbi.nlm.nih.gov/pubmed/15120628. DOI 10.1016/j.bbrc.2004.04.053. Marcais, A., J. Cherfils-Vicini, C. Viant, S. Degouve, S. Viel, A. Fenis, J. Rabilloud, K. Mayol, A. Tavares, J. Bienvenu, Y.G. Gangloff, E. Gilson, E. Vivier and T. Walzer, 2014. The metabolic checkpoint kinase mtor is essential for il-15 signaling during the development and activation of nk cells. Nature immunology, 15(8): 749-757. Available from http://www.ncbi.nlm.nih.gov/pubmed/24973821. DOI 10.1038/ni.2936. Marcais, A. and T. Walzer, 2014. Mtor: A gate to nk cell maturation and activation. Cell cycle, 13(21): 3315-3316. Available from http://www.ncbi.nlm.nih.gov/pubmed/25485573. DOI 10.4161/15384101.2014.972919. Marin, R., F. Ruiz-Cabello, S. Pedrinaci, R. Mendez, P. Jimenez, D.E. Geraghty and F. Garrido, 2003. Analysis of hla-e expression in human tumors. Immunogenetics, 54(11): 767- 775. Available from http://www.ncbi.nlm.nih.gov/pubmed/12618909. DOI 10.1007/s00251-002-0526-9. Marra, J., J. Greene, J. Hwang, J. Du, L. Damon, T. Martin and J.M. Venstrom, 2015. Kir and hla genotypes predictive of low-affinity interactions are associated with lower relapse in autologous hematopoietic cell transplantation for acute myeloid leukemia. Journal of immunology, 194(9): 4222-4230. Available from http://www.ncbi.nlm.nih.gov/pubmed/25810393. DOI 10.4049/jimmunol.1402124. Martin, M.P., Y. Qi, X. Gao, E. Yamada, J.N. Martin, F. Pereyra, S. Colombo, E.E. Brown, W.L. Shupert, J. Phair, J.J. Goedert, S. Buchbinder, G.D. Kirk, A. Telenti, M. Connors, S.J. O'Brien, B.D. Walker, P. Parham, S.G. Deeks, D.W. McVicar and M. Carrington, 2007. Innate partnership of hla-b and kir3dl1 subtypes against hiv-1. Nature genetics, 39(6): 733-740. Available from http://www.ncbi.nlm.nih.gov/pubmed/17496894. DOI 10.1038/ng2035. Matsuoka, S., M. Huang and S.J. Elledge, 1998. Linkage of atm to cell cycle regulation by the chk2 protein kinase. Science, 282(5395): 1893-1897. Available from http://www.ncbi.nlm.nih.gov/pubmed/9836640. Middleton, D., A. Meenagh, J. Moscoso and A. Arnaiz-Villena, 2008. Killer immunoglobulin receptor gene and allele frequencies in caucasoid, oriental and black populations from different continents. Tissue antigens, 71(2): 105-113. Available from http://www.ncbi.nlm.nih.gov/pubmed/18069936. DOI 10.1111/j.1399- 0039.2007.00973.x. Miller, J.S., Y. Soignier, A. Panoskaltsis-Mortari, S.A. McNearney, G.H. Yun, S.K. Fautsch, D. McKenna, C. Le, T.E. Defor, L.J. Burns, P.J. Orchard, B.R. Blazar, J.E. Wagner, A. 116 Slungaard, D.J. Weisdorf, I.J. Okazaki and P.B. McGlave, 2005. Successful adoptive transfer and in vivo expansion of human haploidentical nk cells in patients with cancer. Blood, 105(8): 3051-3057. Available from http://www.ncbi.nlm.nih.gov/pubmed/15632206. DOI 10.1182/blood-2004-07-2974. Millour, J., N. de Olano, Y. Horimoto, L.J. Monteiro, J.K. Langer, R. Aligue, N. Hajji and E.W. Lam, 2011. Atm and p53 regulate foxm1 expression via e2f in breast cancer epirubicin treatment and resistance. Molecular cancer therapeutics, 10(6): 1046- 1058. Available from http://www.ncbi.nlm.nih.gov/pubmed/21518729. DOI 10.1158/1535-7163.MCT-11-0024. Miyamoto, S., A.N. Murphy and J.H. Brown, 2008. Akt mediates mitochondrial protection in cardiomyocytes through phosphorylation of mitochondrial hexokinase-ii. Cell death and differentiation, 15(3): 521-529. Available from http://www.ncbi.nlm.nih.gov/pubmed/18064042. DOI 10.1038/sj.cdd.4402285. Moesta, A.K., P.J. Norman, M. Yawata, N. Yawata, M. Gleimer and P. Parham, 2008. Synergistic polymorphism at two positions distal to the ligand-binding site makes kir2dl2 a stronger receptor for hla-c than kir2dl3. Journal of immunology, 180(6): 3969-3979. Available from http://www.ncbi.nlm.nih.gov/pubmed/18322206. Newsholme, E.A., B. Crabtree and M.S. Ardawi, 1985. Glutamine metabolism in lymphocytes: Its biochemical, physiological and clinical importance. Quarterly journal of experimental physiology, 70(4): 473-489. Available from http://www.ncbi.nlm.nih.gov/pubmed/3909197. Newsholme, E.A., B. Crabtree and M.S. Ardawi, 1985. The role of high rates of glycolysis and glutamine utilization in rapidly dividing cells. Bioscience reports, 5(5): 393- 400. Available from http://www.ncbi.nlm.nih.gov/pubmed/3896338. Novellasdemunt, L., L. Bultot, A. Manzano, F. Ventura, J.L. Rosa, D. Vertommen, M.H. Rider, A. Navarro-Sabate and R. Bartrons, 2013. Pfkfb3 activation in cancer cells by the p38/mk2 pathway in response to stress stimuli. The Biochemical journal, 452(3): 531-543. Available from http://www.ncbi.nlm.nih.gov/pubmed/23548149. DOI 10.1042/BJ20121886. Nsiah-Sefaa, A. and M. McKenzie, 2016. Combined defects in oxidative phosphorylation and fatty acid beta-oxidation in mitochondrial disease. Bioscience reports, 36(2). Available from http://www.ncbi.nlm.nih.gov/pubmed/26839416. DOI 10.1042/BSR20150295. Oppenheim, D.E., S.J. Roberts, S.L. Clarke, R. Filler, J.M. Lewis, R.E. Tigelaar, M. Girardi and A.C. Hayday, 2005. Sustained localized expression of ligand for the activating nkg2d receptor impairs natural cytotoxicity in vivo and reduces tumor immunosurveillance. Nature immunology, 6(9): 928-937. Available from http://www.ncbi.nlm.nih.gov/pubmed/16116470. DOI 10.1038/ni1239. Orange, J.S., 2013. Natural killer cell deficiency. The Journal of allergy and clinical immunology, 132(3): 515-525; quiz 526. Available from http://www.ncbi.nlm.nih.gov/pubmed/23993353. DOI 10.1016/j.jaci.2013.07.020. Oyer, J.L., V. Pandey, R.Y. Igarashi, S.S. Somanchi, A. Zakari, M. Solh, D.A. Lee, D.A. Altomare and A.J. Copik, 2016. Natural killer cells stimulated with pm21 particles expand and biodistribute in vivo: Clinical implications for cancer treatment. Cytotherapy, 18(5): 653-663. Available from http://www.ncbi.nlm.nih.gov/pubmed/27059202. DOI 10.1016/j.jcyt.2016.02.006.

117 Pacheco, R., F. Ciruela, V. Casado, J. Mallol, T. Gallart, C. Lluis and R. Franco, 2004. Group i metabotropic glutamate receptors mediate a dual role of glutamate in t cell activation. The Journal of biological chemistry, 279(32): 33352-33358. Available from http://www.ncbi.nlm.nih.gov/pubmed/15184389. DOI 10.1074/jbc.M401761200. Phatarpekar, P.V., D.A. Lee and S.S. Somanchi, 2016. Electroporation of sirna to silence gene expression in primary nk cells. Methods in molecular biology, 1441: 267-276. Available from http://www.ncbi.nlm.nih.gov/pubmed/27177673. DOI 10.1007/978-1- 4939-3684-7_22. Pittari, G., X.R. Liu, A. Selvakumar, Z. Zhao, E. Merino, M. Huse, J.H. Chewning, K.C. Hsu and B. Dupont, 2013. Nk cell tolerance of self-specific activating receptor kir2ds1 in individuals with cognate hla-c2 ligand. J Immunol, 190(9): 4650-4660. Available from https://www.ncbi.nlm.nih.gov/pubmed/23554313. DOI 10.4049/jimmunol.1202120. Qiu, P., E.F. Simonds, S.C. Bendall, K.D. Gibbs, Jr., R.V. Bruggner, M.D. Linderman, K. Sachs, G.P. Nolan and S.K. Plevritis, 2011. Extracting a cellular hierarchy from high- dimensional cytometry data with spade. Nature biotechnology, 29(10): 886-891. Available from http://www.ncbi.nlm.nih.gov/pubmed/21964415. DOI 10.1038/nbt.1991. Raulet, D.H., 2004. Interplay of natural killer cells and their receptors with the adaptive immune response. Nature immunology, 5(10): 996-1002. Available from http://www.ncbi.nlm.nih.gov/pubmed/15454923. DOI 10.1038/ni1114. Raulet, D.H., 2006. Missing self recognition and self tolerance of natural killer (nk) cells. Seminars in immunology, 18(3): 145-150. Available from http://www.ncbi.nlm.nih.gov/pubmed/16740393. DOI 10.1016/j.smim.2006.03.003. Rohatgi, N., T.K. Nielsen, S.P. Bjorn, I. Axelsson, G. Paglia, B.G. Voldborg, B.O. Palsson and O. Rolfsson, 2014. Biochemical characterization of human gluconokinase and the proposed metabolic impact of gluconic acid as determined by constraint based metabolic network analysis. PLoS One, 9(6): e98760. Available from https://www.ncbi.nlm.nih.gov/pubmed/24896608. DOI 10.1371/journal.pone.0098760. Romee, R., M. Rosario, M.M. Berrien-Elliott, J.A. Wagner, B.A. Jewell, T. Schappe, J.W. Leong, S. Abdel-Latif, S.E. Schneider, S. Willey, C.C. Neal, L. Yu, S.T. Oh, Y.S. Lee, A. Mulder, F. Claas, M.A. Cooper and T.A. Fehniger, 2016. Cytokine-induced memory-like natural killer cells exhibit enhanced responses against myeloid leukemia. Science translational medicine, 8(357): 357ra123. Available from http://www.ncbi.nlm.nih.gov/pubmed/27655849. DOI 10.1126/scitranslmed.aaf2341. Romero-Garcia, S., M.M. Moreno-Altamirano, H. Prado-Garcia and F.J. Sanchez-Garcia, 2016. Lactate contribution to the tumor microenvironment: Mechanisms, effects on immune cells and therapeutic relevance. Frontiers in immunology, 7: 52. Available from http://www.ncbi.nlm.nih.gov/pubmed/26909082. DOI 10.3389/fimmu.2016.00052. Sakuragi, T., X. Lin, C.N. Metz, K. Ojamaa, N. Kohn, Y. Al-Abed and E.J. Miller, 2007. Lung- derived macrophage migration inhibitory factor in sepsis induces cardio- circulatory depression. Surgical infections, 8(1): 29-40. Available from http://www.ncbi.nlm.nih.gov/pubmed/17381395. DOI 10.1089/sur.2006.031.

118 Schaffer, S.W., C.J. Jong, T. Ito and J. Azuma, 2014. Role of taurine in the pathologies of melas and merrf. Amino acids, 46(1): 47-56. Available from http://www.ncbi.nlm.nih.gov/pubmed/23179085. DOI 10.1007/s00726-012-1414-8. Schmid, I. and B.D. Jamieson, 2004. Assessment of telomere length, phenotype, and DNA content. Current protocols in cytometry, Chapter 7: Unit 7 26. Available from http://www.ncbi.nlm.nih.gov/pubmed/18770803. DOI 10.1002/0471142956.cy0726s29. Sleiman, M., N.H. Brons, T. Kaoma, F. Dogu, A. Villa-Forte, P. Lenoble, F. Hentges, K. Kotsch, S.D. Gadola, C. Vilches and J. Zimmer, 2014. Nk cell killer ig-like receptor repertoire acquisition and maturation are strongly modulated by hla class i molecules. Journal of immunology, 192(6): 2602-2610. Available from http://www.ncbi.nlm.nih.gov/pubmed/24554773. DOI 10.4049/jimmunol.1302843. Somanchi, S.S., V.V. Senyukov, C.J. Denman and D.A. Lee, 2011. Expansion, purification, and functional assessment of human peripheral blood nk cells. Journal of visualized experiments : JoVE(48). Available from http://www.ncbi.nlm.nih.gov/pubmed/21339714. DOI 10.3791/2540. Tamas, P., S.A. Hawley, R.G. Clarke, K.J. Mustard, K. Green, D.G. Hardie and D.A. Cantrell, 2006. Regulation of the energy sensor amp-activated protein kinase by antigen receptor and ca2+ in t lymphocytes. The Journal of experimental medicine, 203(7): 1665-1670. Available from http://www.ncbi.nlm.nih.gov/pubmed/16818670. DOI 10.1084/jem.20052469. Tandon, P., C.A. Gallo, S. Khatri, J.F. Barger, H. Yepiskoposyan and D.R. Plas, 2011. Requirement for ribosomal protein s6 kinase 1 to mediate glycolysis and apoptosis resistance induced by pten deficiency. Proceedings of the National Academy of Sciences of the United States of America, 108(6): 2361-2365. Available from http://www.ncbi.nlm.nih.gov/pubmed/21262837. DOI 10.1073/pnas.1013629108. Tripathy, S.K., P.A. Keyel, L. Yang, J.T. Pingel, T.P. Cheng, A. Schneeberger and W.M. Yokoyama, 2008. Continuous engagement of a self-specific activation receptor induces nk cell tolerance. The Journal of experimental medicine, 205(8): 1829- 1841. Available from http://www.ncbi.nlm.nih.gov/pubmed/18606857. DOI 10.1084/jem.20072446. Uhrberg, M., N.M. Valiante, B.P. Shum, H.G. Shilling, K. Lienert-Weidenbach, B. Corliss, D. Tyan, L.L. Lanier and P. Parham, 1997. Human diversity in killer cell inhibitory receptor genes. Immunity, 7(6): 753-763. Available from http://www.ncbi.nlm.nih.gov/pubmed/9430221. van der Windt, G.J. and E.L. Pearce, 2012. Metabolic switching and fuel choice during t-cell differentiation and memory development. Immunological reviews, 249(1): 27-42. Available from http://www.ncbi.nlm.nih.gov/pubmed/22889213. DOI 10.1111/j.1600- 065X.2012.01150.x. Vander Heiden, M.G., L.C. Cantley and C.B. Thompson, 2009. Understanding the warburg effect: The metabolic requirements of cell proliferation. Science, 324(5930): 1029- 1033. Available from http://www.ncbi.nlm.nih.gov/pubmed/19460998. DOI 10.1126/science.1160809. Viant, C., A. Fenis, G. Chicanne, B. Payrastre, S. Ugolini and E. Vivier, 2014. Shp-1-mediated inhibitory signals promote responsiveness and anti-tumour functions of natural killer cells. Nature communications, 5: 5108. Available from http://www.ncbi.nlm.nih.gov/pubmed/25355530. DOI 10.1038/ncomms6108. 119 Wahle, J.A., K.H. Paraiso, R.D. Kendig, H.R. Lawrence, L. Chen, J. Wu and W.G. Kerr, 2007. Inappropriate recruitment and activity by the src homology region 2 domain- containing phosphatase 1 (shp1) is responsible for receptor dominance in the ship- deficient nk cell. Journal of immunology, 179(12): 8009-8015. Available from http://www.ncbi.nlm.nih.gov/pubmed/18056340. Walenta, S., M. Wetterling, M. Lehrke, G. Schwickert, K. Sundfor, E.K. Rofstad and W. Mueller-Klieser, 2000. High lactate levels predict likelihood of metastases, tumor recurrence, and restricted patient survival in human cervical cancers. Cancer research, 60(4): 916-921. Available from http://www.ncbi.nlm.nih.gov/pubmed/10706105. Wang, B., Z. Li, C. Wang, M. Chen, J. Xiao, X. Wu, W. Xiao, Y. Song and X. Wang, 2013. Zygotic g2/m cell cycle arrest induced by atm/chk1 activation and DNA repair in mouse embryos fertilized with hydrogen peroxide-treated epididymal mouse sperm. PloS one, 8(9): e73987. Available from http://www.ncbi.nlm.nih.gov/pubmed/24040138. DOI 10.1371/journal.pone.0073987. Wang, R., C.P. Dillon, L.Z. Shi, S. Milasta, R. Carter, D. Finkelstein, L.L. McCormick, P. Fitzgerald, H. Chi, J. Munger and D.R. Green, 2011. The transcription factor myc controls metabolic reprogramming upon t lymphocyte activation. Immunity, 35(6): 871-882. Available from http://www.ncbi.nlm.nih.gov/pubmed/22195744. DOI 10.1016/j.immuni.2011.09.021. Warburg, O., K. Gawehn and A.W. Geissler, 1958. [metabolism of leukocytes]. Zeitschrift fur Naturforschung. Teil B, Chemie, Biochemie, Biophysik, Biologie und verwandte Gebiete, 13B(8): 515-516. Available from http://www.ncbi.nlm.nih.gov/pubmed/13593654. Warburg, O., F. Wind and E. Negelein, 1927. The metabolism of tumors in the body. The Journal of general physiology, 8(6): 519-530. Available from http://www.ncbi.nlm.nih.gov/pubmed/19872213. Wei, H., W.T. Nash, A.P. Makrigiannis and M.G. Brown, 2014. Impaired nk-cell education diminishes resistance to murine cmv infection. European journal of immunology, 44(11): 3273-3282. Available from http://www.ncbi.nlm.nih.gov/pubmed/25187217. DOI 10.1002/eji.201444800. Westphal, D., G. Dewson, P.E. Czabotar and R.M. Kluck, 2011. Molecular biology of bax and bak activation and action. Biochimica et biophysica acta, 1813(4): 521-531. Available from http://www.ncbi.nlm.nih.gov/pubmed/21195116. DOI 10.1016/j.bbamcr.2010.12.019. Wu, J. and L.L. Lanier, 2003. Natural killer cells and cancer. Adv Cancer Res, 90: 127-156. Available from https://www.ncbi.nlm.nih.gov/pubmed/14710949. Wu, S.B. and Y.H. Wei, 2012. Ampk-mediated increase of glycolysis as an adaptive response to oxidative stress in human cells: Implication of the cell survival in mitochondrial diseases. Biochimica et biophysica acta, 1822(2): 233-247. Available from http://www.ncbi.nlm.nih.gov/pubmed/22001850. DOI 10.1016/j.bbadis.2011.09.014. Yawata, M., N. Yawata, M. Draghi, F. Partheniou, A.M. Little and P. Parham, 2008. Mhc class i-specific inhibitory receptors and their ligands structure diverse human nk-cell repertoires toward a balance of missing self-response. Blood, 112(6): 2369-2380. Available from http://www.ncbi.nlm.nih.gov/pubmed/18583565. DOI 10.1182/blood- 2008-03-143727.

120 Yoshikawa, S., K. Muramoto and K. Shinzawa-Itoh, 2011. Proton-pumping mechanism of cytochrome c oxidase. Annual review of biophysics, 40: 205-223. Available from http://www.ncbi.nlm.nih.gov/pubmed/21545285. DOI 10.1146/annurev-biophys- 042910-155341. Zhen, Y., D. Li, W. Li, W. Yao, A. Wu, J. Huang, H. Gu, Y. Huang, Y. Wang, J. Wu, M. Chen, D. Wu, Q. Lyu, W. Fang and B. Wu, 2016. Reduced pdcd4 expression promotes cell growth through pi3k/akt signaling in non-small cell lung cancer. Oncology research, 23(1-2): 61-68. Available from http://www.ncbi.nlm.nih.gov/pubmed/26802652. DOI 10.3727/096504015X14478843952861.

121 Vita

Jolie Rae Schafer was born in Austin, Texas on August 20, 1990, the daughter of Suzette Graf and Tom Graf. After completing her work at McNeil High School, Austin, Texas in 2008, she entered Houston Baptist University in Houston, Texas. She received the degree of Bachelor or

Science with a double major in Biology and Biochemistry Molecular Biology from HBU in

May, 2012. For the next year, she worked as a research assistant in the Department of

Experimental Radiation Oncology at MD Anderson Cancer Center. In May of 2013 she entered

The University of Texas MD Anderson Cancer Center UTHealth Graduate School of

Biomedical Sciences.

122