Mechanisms of Extracellular Nucleotide Accumulation

MECHANISMS OF EXTRACELLULAR NUCLEOTIDE ACCUMULATION

DURING REGULATED CELL DEATH IN TUMOR CELLS

ANDREA MICHELLE BOYD TRESSLER

Submitted in partial fulfillment of the requirements

For the degree of Doctor of Philosophy

Dissertation Advisor: George R Dubyak, PhD

Department of Pharmacology

Case Western Reserve University

May 2016

CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

Andrea Michelle Boyd Tressler

candidate for the degree of Doctor of Philosophy*.

Ruth Keri (Committee Chair)

George Dubyak (Dissertation Advisor)

Vera Moiseenkova-Bell (Committee Member)

Clark Distelhorst (Committee Member)

Thomas Kelley (Committee Member)

Date of Defense

March 22, 2016

*We also certify that written approval has been obtained for any proprietary material contained therein.

Dedication

I would like to dedicate this to my parents, Nik and Debbie, who have been my biggest supporters my entire life. From ballet lessons to my PhD, they have always challenged and encouraged me, sacrificing time and money to make sure they were front and center at everything I did. I would not be the woman I am today without them. Also, to my niece, Taylor, who is my favorite person and a big reason I’ve never given up.

Table of Contents

List of Figures ………………….……………………..…………...……………...... vi

Acknowledgements …………………………………………………………………x

List of Abbreviations………………………………………………………………...xii

Abstract………………………………………………………………………………….1

Chapter 1: Introduction and Background…………………………………………3

1.1 Tumor immunotherapy………………….………………………………………….3

1.2 Programmed Cell Death………………….………………………………………..8

1.2.1 Apoptosis………………………………………………………………………….8

1.2.1.1 Extrinsic…………………………………………………….……………………8

1.2.1.2 Intrinsic…………………………………………………………………………12

1.2.2 Necroptosis………………………………………………………………………13

1.3 Chemotherapeutic Drugs…………………………………………..………….....18

1.3.1 Doxorubicin……………………………………...…………………..…………..18

1.3.2 Etoposide……………………………………………………………..………….19

1.3.3 Staurosporine…………………………………………………………..………..20

1.4 Release of Intracellular ATP…………………………………………….……….21

1.4.1 Purinergic Signaling…………………………………………………….………21

1.4.2 Mechanisms of ATP Release…………………………………………..……...22

1.4.2.1 Exocytosis …………………………………………………………………….23

1.4.2.2 Plasma Membrane Permeabilization……………………………………….23

1.4.2.3 Regulated efflux via channels……………………………………………….23

iii

1.5 Pannexin 1 Channel……………………………………………………………....26

1.6 ATP metabolism…………………………………………………………………...31

1.6.1 Ecto-Nucleotide triphosphate diphosphohydrolase………………………...32

1.6.2 Ecto-nucleotide pyrophosphatase/phosphodiesterase…………………….32

1.6.3 CD73……………………………………………………………………………..32

1.7 Statement of Purpose…………………………………………………………….33

Chapter 2: Materials and Methods………………………………………………...36

Chapter 3: Chemotherapeutic Drugs Induce ATP Release via Caspase- gated Pannexin-1 Channels and a Caspase/Pannexin-1-independent

Mechanism………………………………………………………………………….....47

3.1 Abstract…………………………………………………………………………….47

3.2 Introduction………………………………………………………………………..48

3.3 Results……………………………………………………………………………..51

3.4 Discussion………………………………………………………………………....67

3.5 Acknowledgements……………………………………………………………….77

Chapter 4: Upregulated ectonucleotidases in FADD- and RIP1-deficient

Jurkat leukemia cells counteracts extracellular ATP/AMP accumulation via

pannexin-1 channels during chemotherapeutic drug-induced apoptosis

...…………………..……………...... 101

4.1 Abstract…………………………………………………………………………...101

4.2 Introduction……………………………………………………………………....102

4.3 Results………………………………………………………………………..…..106

4.5 Discussion…………………………………………………………………..……121

4.5 Acknowledgements……………………………………………………………...130

Chapter 5: Regulation of Adenine Nucleotide release through Pannexin 1 channels in EG7 Murine Lymphoma Cells…………………………………...... 147

Chapter 6: Discussion and Future Directions……………………………..…..157

6.1 Mechanisms of Programmed Cell Death Induced Adenine Nucleotide

Release……………………………………………………………………..………...159

6.2 Role of FADD and RIP1 in ectonucleotidase activity……………..………….160

6.3 Decrease in immune cell activation………………………………..…………..163

6.4 Post-translational modification of Panx1………………………………………164

6.5 Concluding Remarks…………………………………………………………….166

Appendix……………………………………………………………………………..170

References………………………………………………………………………...…185

List of Figures

Figure 1.1: Activation of the purinergic anti-tumor immune response after treatment of chemotherapy………………………………………………………..…...5

Figure 1.2: Intrinsic and Extrinsic Activation of Apoptosis……………………...... 9

Figure 1.3: TNF signaling……………………………………………………………..15

Figure 1.4: Mechanisms of ATP Release…………………………………………...24

Figure 1.5: Pannexin 1 Channel……………………………………………………..27

Figure 3.1: Comparative time courses for accumulation of active caspase-3 and loss of viability in Jurkat leukemic T cells treated with different chemotherapeutic agents…………………………………………………………………………………..78

Figure 3.2: Chemotherapeutic drugs induce caspase-3 mediated cleavage of the pannexin-1 C-terminal autoinhibitory domain………………………………………80

Figure 3.3: Chemotherapeutic drugs induce accumulation of active pannexin-1 channels via a caspase dependent activation mechanism………………………..82

Figure 3.4: Efflux of both ATP and ATP metabolites is triggered during chemotherapeutic drug-induced apoptosis of Jurkat cells………………………...84

Figure 3.5: Caspase-activated pannexin-1 channels mediate the efflux of ATP and ATP metabolites during chemotherapeutic drug-induced apoptosis but an alternative ATP release mechanism is engaged in the context of suppressed caspase activity………………………………………………………………………..87

Figure 3.6: Carbenoxolone blocks the efflux of ATP and ATP metabolites during chemotherapeutic drug-induced apoptosis…………………………………………89

Figure 3.7: Caspase-insensitive ATP release stimulated by chemotherapeutic drugs is resistant to carbenoxolone blockade but suppressed by intracellular

Ca2+ buffering………………………………………………………………………….91

Figure 3.8: Caspase-insensitive ATP release stimulated by staurosporine does

not involve direct Ca2+mobilization by staurosporine or activation of phosphatidyl

inositol phospholipase C signaling…………………………………………………..94

Figure 3.9: Proteosome inhibition induces caspase-3-mediated cleavage of the

pannexin-1 C-terminal autoinhibitory domain and pannexin-1-mediated release of

adenine nucleotides…………………………………………………………………...96

Figure 3.10: Pannexin-1 is more highly expressed in human leukemic leukocytes

than in normal human T cells…………………………………………………………99

Figure 4.1: TNFα-induction of necroptosis or extrinsic apoptosis induces release

of adenine nucleotides from Jurkat cancer cells via mechanistically distinct

pathways………………………………………………………………………..…….131

Figure 4.2: Intrinsic apoptotic signaling and Panx1 channel cleavage in FADD-

deficient Jurkat cancer cells is uncoupled from accumulation of extracellular

adenine nucleotides………………………………………………………………….134

Figure 4.3: Caspase-3-cleaved Panx1 channels are functionally active in FADD-

deficient Jurkat cancer cells during intrinsic apoptosis………………………...... 136

Figure 4.4: CD73 ecto-nucleotidase activity is upregulated in FADD-deficient

Jurkat cancer cells and counteracts Panx1 channel-mediated efflux of ATP/AMP

during intrinsic apopotosis…………………………………………………………..138

vii

Figure 4.5: Apoptotic signaling and Panx1 channel activation in RIP1-deficient

Jurkat cancer cells is also uncoupled from accumulation of extracellular adenine

nucleotides……………………………………………………………………………140

Figure 4.6: RIP1 deficient cells have increased CD73 activity………………….143

Figure 4.7: Increased expression of CD39 ecto-ATPase in RIP1-deficient Jurkat

cells relative to FADD-deficient or wildtype Jurkat cells………………………….145

Figure 5.1: Comparative activation of apoptosis in Jurkat T cells and EG7 cells

treated with chemotherapeutic agent ……..……………………………………….151

Figure 5.2: Activation of apoptosis in EG7 cells by chemotherapies does not lead

to Panx1 mediated Adenine Nucleotide accumulation…………………………..153

Figure 5.3: Cleavage of Panx1 in EG7 cells leads to increased YO-Pro dye

influx…………………………………………………………………………………...155

Appendix 1: Full Western Blot of Panx1 in Anti-Fas-treated Jurkat T cells in the presence of absence of zVAD………………………………………………………170

Appendix 2: Full Western Blot of Panx1 in STS-treated Jurkat T cells in the presence of absence of zVAD………………………………………………………171

Appendix 3: Full Western Blot of Panx1 in Dox-treated Jurkat T cells in the presence of absence of zVAD………………………………………………………172

Appendix 4: Full Western Blot of Panx1 in Etop-treated Jurkat T cells in the presence of absence of zVAD………………………………………………………173

Appendix 5: Full Western Blot of Panx1 in TS- or Anti-Fas-treated WT Jurkat T cells…………………………………………………………………………………….174

viii

Appendix 6: Full Western Blot of Panx1 in TS- or TS+zVAD-treated FADD-def

Jurkat T cells………………………………………………………………………….175

Appendix 7: Full Western Blot of Panx1 in STS-, Etop- or Anti-Fas-treated WT

Jurkat T cells………………………………………………………………………….176

Appendix 8: Full Western Blot of Panx1 in STS-, Etop- or Anti-Fas-treated FADD- def Jurkat T cells……………………………………………………………………..177

Appendix 9: Full Western Blot of Panx1 in STS-, or Etop-treated WT and RIP1- def Jurkat T cells…………………………………………………………………….178

Appendix 10: Full Western Blot of Panx1 in STS- or Dox-treated EG7 Lymphoma cells…………………………………………………………………………………….179

Appendix 11: Copy right permission forms………………………………………..180

Acknowledgements

I would first like to thank my advisor, Dr. George Dubyak for giving me the amazing opportunity to work in his lab. When I was trying to find a lab in my first year, everyone told me to pick his because of how great a mentor he is. They were not wrong. George has helped me grow not only as a scientist, by challenging me to think critically about data and the literature, but also as a person. It has been a long, tough road, but George has been there to help guide me, especially when the data shows unexpected (and unwelcome) findings, proving that persistence and a little reimagining can get me through even the most unpleasant of situations.

Secondly, I would like to thank my lab members for making lab a fun place to be. Current members; Dr. Mausita Karmakar, Hana Russo, Kristin Lozada-

Soto, Nitish Rana, former memers; Dr. Kate Trueblood, Dr. Domenick “Tony”

Prosdocimo, Dr. Andrew Blum, Dr. Christina Antonopolous-Buzzy, Dr. Michael

Katsnelson, Dan Chopick, and Caroline El Sanadi, and honorary lab member

Brian Fort. From playing “Push It” when we got good data to being there when I was frustrated, you made getting my PhD a little easier.

I would like to thank my committee of Dr. Ruth Keri, Dr. Vera

Moiseenkova-Bell, Dr. Clark Distelhorst, and Dr. Tom Kelley for always giving me great feedback and advice on my project. Additionally, I would like to thank Dr.

Krzysztof Palczewski for genuinely caring about my success as a student and scientist, and for making sure I came to seminar.

Thank you to Dr. John Mieyal for always being available to talk about any issues I had during my matriculation. Also, to the Pharmacology department staff and the Department of Physiology and Biophysics staff for making sure I had everything in on time.

Finally, I would like to thank my family and friends. Without you this would have been impossible. To my parents, Nik and Debbie, for always believing in me and doing whatever they needed to do to make sure I had everything I needed. To my brothers Eric, Chris, and Ben for being my protectors and giving me tough skin. My niece, Taylor and nephews Mason, Caiden, Jorden, Benjamin, and Liam for being the little eyes that were always watching and who gave me the motivation to keep going. Thank you to my friends, both inside and outside of school, for making life fun and exciting. You guys made grad school bearable, and almost fun.

And remember, the data’s the data.

List of Abbreviations

Panx1: pannexin 1

Casp: caspase

Z: benzyloxycarbonyl

CBX: carbenoxolone

Dox: doxorubicin

STS: staurosporine

BafA: bafilomycin A

Etop: etoposide

PI-PLC: phosphatidylinositol phospholipase C

PARP: poly(ADP-ribose) polymerase

3-MA: 3-methyladenine

ANex: extracellular adenine nucleotide

FLAAM: Firefly luciferase ATP assay mix

FLAAB: Firefly luciferase ATP assay buffer

TCR: T cell receptor

BAPTA: 1,2-bis(2-aminophenoxy)ethane-N,N,N’,N’-tetraacetic acid

BSS: basal salt solution

DC: dendritic cell

Trov: Trovafloxacin

RIP: Receptor interacting protein

RHIM: RIP homotypic interaction motif

TNF: Tumor necrosis factor

xii

MLKL: mixed-lineage kinase like pseudo-kinase

pMLKL: phosphor-MLKL

FADD: Fas associated death domain

DISC: Death inducing signaling complex

FADD-def: FADD deficient Jurkat T cells

RIP1-def: RIP1 deficient Jurkat T cells

APCP: αβ-methylene-adenosine 5’-diphosphate

Tetra: Tetramisole

ε-AMP: 1,N6-etheno-AMP

NSA: Necrosulfonimide

LDH: lactate dehydrogenase

TS: TNF-α+BV6

ARL: ARL67156

SMAC: second mitochondrial-derived activator of caspases

MHC: major histocompatibility complex

DAMP: Danger associated molecular patters

PRR: Pattern recognition receptors

TLR: Toll-like receptor

NK: Natural killer cells

PCD: Programmed cell death

TRAIL: TNF-related apoptosis inducing ligand

FASL: Fas ligand

DD: Death domain

xiii

DED: Death effector domain cIAP: cellular inhibitor of apoptosis

TRAF: TNF receptor associated factor

Bcl-2: B cell CLL/lymphoma-2

BH: BCL-2 homology

MOMP: mitochondrial outer membrane permeabilization

OMM: mitochondrial outer membrane

CALHM1: calcium homeostasis modulator 1

Cx43: Connexin 43 ns, not significant

xiv

Mechanisms of Extracellular Nucleotide Accumulation During Regulated

Cell Death in Tumor Cells

ANDREA MICHELLE BOYD TRESSLER

Abstract

Accumulation of extracellular adenine nucleotides is a key regulator of the purinergic anti-tumor immune response that is triggered after treatment with certain chemotherapeutic agents. ATP released from dying tumor cells has been shown to engage P2 purinergic receptors on nearby leukocytes to activate this response. Conversely, accumulation of adenosine within the tumor microenvironment causes suppression of the immune response. The balance between ATP and adenosine in the tumor microenvironment can dictate the aggressiveness of the cancer and its response to chemotherapy treatment. To examine the regulated release of ATP and its metabolites, as well as their extracellular accumulation, we conducted three main studies. In the first study, we used human leukemic Jurkat T cells to characterize the role of Pannexin 1

(Panx1) channels in the release of ATP during treatment with three chemotherapeutic agents: doxorubicin (Dox), Etoposide (Etop), and staurosporine (STS). These diverse pro-apoptotic drugs were able to induce functional activation of the Panx1 channel by cleavage of the autoinhibitory C terminal domain. Activation of Panx1 led to the release of ATP, but the majority of nucleotide release was in the form of ADP and AMP. Our second study examined the how different rates of adenine nucleotide release and

ectometabolism dictate the composition of pro- or anti- inflammatory nucleotides using Jurkat T cells that included clonal variants lacking either FADD or RIP1and

treated with pro-apoptotic chemotherapeutic drugs. Despite similar apoptotic

induction, the accumulation of ATP+ADP+AMP was decreased in the clonal

variants and this decrease was linked to increased expression of the

ectonucleotidase CD73. Finally, we explored the role of Panx1-mediated ATP

release in the murine lymphoma cell line, EG7, that have been used for in vivo

studies of the purinergic anti-tumor immune response. We discover that while the

EG7 cells activate Panx1 and release adenine nucleotides, the release is

independent of Panx1. Together these studies provide further understanding of

the regulated release and accumulation of ATP, ADP, and AMP within the tumor

microenvironment after treatment with chemotherapeutic drugs.

Chapter 1: Introduction and Background

Efficacy of chemotherapies has conventionally been thought to hinge on

the direct cytotoxic effects on the tumor. However, it has been shown that during

treatment with certain chemotherapies, such as doxorubicin or oxaliplatin, part of

the efficacy is linked to the ability of these therapies to activate a purinergic anti-

tumor immune response. This immune response falls under the larger category

of tumor immunotherapy and is dependent on the ATP, released by dying tumor

cells, signaling through purinergic cascades on responding immune cells. The

direct targets of chemotherapies, cell death signaling, release and metabolism of

ATP and activation of P2 purinergic receptors all play an important role in

activating the anti-tumor immune response and elimination of the tumor cells.

1.1 Tumor immunotherapy

Tumor immunotherapy is a growing field that has led to the development of therapies designed to harness the body’s immune system in the fight against cancer. Even before the development of effective immunotherapies against cancer, it was noted that patients with skin infections or autoimmune diseases had a better prognosis than those without [1]. In a famous experiment the idea of

utilizing the patient’s immune system gained recognition in the late 1800’s when

William Coley injected bone or soft tissue sarcoma patients with a bacterial toxin,

the “Coley toxin”, and noted partial responses in some patients and complete

remission in others [2]. However, it was not until recently with the success of anti-

CTLA-4 and anti-PD1 immune checkpoint inhibitors that optimism for this treatment option took hold [3].

Currently, immunotherapy is used in combination with chemotherapy for maximal efficacy and has become the standard of care for many malignancies.

The use of chemotherapies in conjunction with the immunotherapy takes advantage of three effects of the chemotherapeutic agent. First is the reduction of tumor burden. Immunotherapies are most effective when the tumor burden is small, so chemotherapy reduces that burden in patients with large tumors.

Second, it directly affects tumor stromal cells. Third, the immune response is enhanced by the exposure of a neoantigen. Neoantigens are novel protein sequences that arise as a result of mutations. These antigens can be processed into peptide antigens that are presented by the major histocompatibility complex

(MHC) and recognized by T cells as foreign [4].

Chemotherapies are currently vital to the efficiency of eliminating cancer cells from patients, and combination with immunotherapy highlights their importance. However, treatment with certain chemotherapies, such as oxaliplatin or doxorubicin, can trigger the anti-tumor immune response without use of immunotherapies (Fig 1) [5, 6]. As mentioned above, one of the key functions of the chemotherapies is their ability to expose neoantigens to responding macrophages and dendritic cells (DCs). This process hinges on purinergic signaling pathways in resident macrophages and DCs [5].

Figure 1.1: Activation of the purinergic anti-tumor immune response after

treatment of chemotherapy

Treatment with chemotherapy drugs induces programmed cell death. This leads

to the release of ATP which binds to P2Y2 receptors on nearby macrophages

and DC’s to enhance chemotaxis and phagocytosis as well as P2X7 receptor to induce assembly of the inflammasome complex and release of mature IL-1β.

Phagocytosis of dying tumor cells leads to antigen presentation to CD8+ T cells.

This, along with IL-1β dependent polarization, activates the T cells against the neoantigen leading to direct killing of the tumor cells and recruitment of NK cells.

Figure 1.1: Activation of the purinergic anti-tumor immune response after treatment of chemotherapy

In addition to exposure of the neoantigen, treatment of the tumor cells with

the chemotherapies induces cell death and the release of danger associated

molecular patterns (DAMPs) such as ATP [7]. The released ATP acts as a “find

me” signal by binding to the G protein-coupled receptor P2Y2 on resident

macrophages and DCs and enhances the recognition and phagocytosis of the

dying tumor cells and processing of tumor antigen [8-10]. In addition, ATP will

bind to P2X7 receptor on the macrophages/DCs and stimulate the assembly of

the caspase-1 inflammasome and release of the cytokine IL-1β [5]. The

macrophages/DCs then present the tumor antigen to CD8+ T cells, and in

combination with IL-1β dependent polarization of the T cells, the T cells recognize the antigen as foreign and commence direct killing of the tumor cells as well as recruitment of natural killer cells (NKs).

Activation of this response is contingent on the release of ATP from the

dying tumor cells and its binding to the purinergic receptors on the resident

macrophages/DCs. Ghiringhelli et al. demonstrated that the anti-tumor immune

response initiated by oxaliplatin was abolished in P2X7-/- mice due to the inability

of ATP to stimulate the release of IL-1β and subsequently, the efficacy of

oxaliplatin was decreased [5]. This highlights the importance of the ATP released

from cells undergoing programmed cell death in achieving maximal efficacy of

chemotherapies that are able to activate the anti-tumor immune response.

Due to the cytotoxic nature of the chemotherapies, the process of

programmed cell death, in particular apoptosis, is vital to the release of ATP and

other DAMPs.

1.2 Programmed Cell Death

Programmed cell death (PCD) describes modes of cell death that are controlled by the expression of certain genes. In the 1960’s and 1970’s two forms of cell death were described, apoptosis was the PCD form and necrosis was used to describe accidental, non-programmed cell death [11, 12]. However, since the 70s, multiple forms of PCD have been discovered. While apoptosis remains the most well studied and vital form of PCD, necroptosis is emerging as another important pathway.

1.2.1 Apoptosis

There are two pathways that trigger apoptosis: extrinsic and intrinsic.

While both pathways converge on the activation of executioner caspases 3 and

7, the signals and pathways leading up to this differ (Fig 1.2).

1.2.1.1 Extrinsic

The extrinsic pathway is activated by the binding of cell death ligands to their respective cell death receptors located in the plasma membrane of the cell.

One very common cell death ligand/receptor pair is the FAS ligand (FASL) and the FAS receptor. The FAS receptor (also known as APO-1 or CD95) belongs to the tumor necrosis factor receptor (TNF-R) family, which includes TNF and TNF- related apoptosis inducing ligand (TRAIL) [13]. Upon binding of FASL to the FAS receptor, assembly of the death inducing signaling complex (DISC) occurs. The

DISC complex consists of FAS, the adaptor protein FAS associated death domain (FADD), and pro-caspase 8 [14]. Both FAS and FADD contain homotypic

Figure 1.2: Intrinsic and Extrinsic Activation of Apoptosis

A. Extrinsic apoptosis is triggered by binding of cell death ligands to their cell death receptor. This triggers the assembly of the DISC that includes FADD and procaspase 8. Procaspase 8 cleaves itself and acts as an initiator caspase to cleave and activate executioner caspases 3 and 7. B. Intrinsic apoptosis can be activated by DNA damage which can lead to increased transcription of pro- apoptotic Bcl-2 proteins that will inhibit the anti-apoptotic proteins as well as lead to MOMP and release of cytochrome c and assembly of the apoptosome complex of APAF-1, procaspase 9, and cytochrome c. Apoptosome assembly induces activation of procaspase 9, which is an initiator caspase that will activate caspase 3 and 7.

Figure from Long. Oncogene 2012, 31: 5045-5060. For copyright permission see appendix 11.

Figure 1.2: Intrinsic and Extrinsic Activation of Apoptosis

Long. Oncogene 2012, 31: 5045-5060

death domains (DD) that allow for interaction between the two proteins. When

activated, FADD is recruited to the intracellular domain of FAS. This promotes

the dimerization of pro-caspase 8 to FADD via the death effector domains (DED)

present on both proteins [15]. This interaction allows for the conformational

change of pro-caspase 8 needed for enzymatic activation and auto-proteolytic processing that allows for its release from the DISC complex. Once released, caspase 8 will process effector caspases 3 and 7 that lead to the systematic disassembly of the cell.

FAS induced cell death can be regulated at the level of DISC assembly. c-

FLIP is a modulator of caspase-8 that contains a DED but has no enzymatic activity [16]. Depending of the relative levels of c-FLIP, it can compete with caspase-8 for binding with FADD and reduce the amount of active caspase 8 able to activate apoptosis.

An intriguing member of the TNF-R family is the TNF receptor 1 (TNF-R1).

Activated by the binding of the cytokine TNF-α, TNF-R1 can signal through three

mechanistically distinct pathways: cell survival, apoptosis and necroptosis (Fig

1.3) [17]. Similar to FAS signaling, when TNF-α binds to TNF-R1 it facilitates the

binding of a DD containing protein, in this case the TNF-R1 associated death

domain (TRADD). In addition, the receptor interacting protein kinase 1 (RIP1) is recruited to the complex. Once RIP1 is bound, E3 ubiquitin ligases TRAF2 and 5

(TNF receptor associated factor) and the cellular inhibitors of apoptosis 1 and 2

(cIAPs) are recruited to the complex, forming complex I. The cIAPs and TRAF2 catalyze the ubiquitination of RIP1 which is necessary for signaling through the

NFκB cell survival pathway [17]. The poly-ubiquitinated RIP1 acts as an adaptor

protein and is bound by the ubiquitin-binding domains of IκB kinase (IKK) and

TAK-1 kinase complexes leading to TAK-1-mediated activation of NFκB and

upregulation of pro-survival proteins such as cFLIP. However, if there is an

upregulation of the cIAP inhibitor second mitochondrial-derived activator of

caspases (SMAC also known as DIABLO) or an upregulation of the CLY and A20 proteins that can remove the ubiquitin from RIP1, RIP1 will leave complex 1 and bind to FADD and procaspase 8 forming the DISC complex (also known as complex II). Similar to DISC assembly in FAS activated cells, complex II will initiate activation of apoptosis [17]. Necroptosis can also be activated via TNF-R1

signaling and will be discussed in detail below.

1.2.1.2 Intrinsic

The intrinsic pathway of apoptosis is also known as the mitochondrial

pathway due to the importance of the mitochondria in the execution of the cell.

The mitochondrial pathway is controlled by the B cell CLL/lymphoma-2 (BCL-2)

family of proteins that consist of both pro- and anti-apoptotic proteins. Divided

into three categories, the BCL-2 family proteins are characterized by the BCL-2

homology (BH) domain [18, 19]. The anti-apoptotic BCL-2-like proteins include

BCL-2, BCL-xL, BCL-w, Mcl-1, and A1/Bfl1 and contain four BH domains. The

pro-apoptotic effector proteins BCL-2 antagonist killer 1 (Bak) and BCL-2-

associated x protein (Bax) also contain four BH domains and are the mediators

of mitochondrial outer membrane permeabilization (MOMP). The third group is

made up of the BH3-only proteins, which are also pro-apoptotic. These proteins,

including Bid, Bim, Bad, and Puma, contain only a short BH3 motif. The BH3-only

proteins can interact with either the anti-apoptotic proteins to inhibit them or Bax

and Bak to stimulate their activity. Activation and insertion of Bax/Bak oligomers

into the mitochondrial outer membrane (OMM) can be achieved by direct

activation by BH3-only proteins Bid or Bim or by inhibition of anti-apoptotic BCL-2 proteins by Bad, or Puma [20]. Once inserted into the OMM, Bax/Bak cause

MOMP that leads to the release of cytochrome c. Cytochrome c associates with apoptotic protease-activating factor 1 (APAF1) and procaspase 9, an initiator caspase, to form the apoptosome [21]. Formation of the apoptosome is thought to be the point at which the cell commits to apoptosis due to the activation of caspase 9 and subsequent activation of casp3/7. The relative levels of the pro- and anti-apoptotic BCL-2 proteins will determine the fate of the cell.

DNA damage caused by chemotherapies is one mechanism for the activation of the intrinsic apoptotic pathway. Breaks in the DNA can lead to the activation of p53 [22]. Low levels of damage can cause p53 to signal for cell cycle arrest and repair of the damage. If the damage is extensive, p53 will cause the activation of pro-apoptotic BCL-2 proteins, leading to mitochondrial mediated apoptosis. p53 is often mutated in many malignancies, however these cells can still undergo apoptosis in response to DNA damage. The homolog to p53, p73, can be upregulated during the DNA damage response and induce apoptosis in a similar fashion to p53 [23].

1.2.2 Necroptosis

Necroptosis, or programmed necrosis, describes a form of non-apoptotic

cell death that results in plasma membrane rupture. Necroptosis has been shown

to be activated in cells that are infected with viruses that have evaded apoptosis,

such as the cytomegalovirus that encodes inhibitors of caspase 8 mediated

apoptosis [24]. The best characterized model for the activation of necroptosis is

TNF-α binding to the TNF-R1, however other receptors and intracellular sensors

have also been shown to activate it. As described above, TNF-R1 can signal via

three distinct mechanisms: cell survival, apoptosis, and necroptosis [17].

Activation of necroptosis in this system is dependent upon the kinase activity of

RIP1 (Fig 1.3). When RIP1 is associated with complex II after TNF-R1 activation

it is cleaved and inactivated by caspase 8. However, when caspase 8 is inhibited

or complex II is unable to form RIP1 associates with RIP3 inducing auto-

phosphorylation of RIP1 and transphosphorylation of RIP3, resulting in the

formation of the necrosome. Recruitment and phosphorylation of additional RIP3

to the complex is followed by recruitment of mixed-lineage kinase domain-like protein (MLKL). Once in the complex, MLKL is phosphorylated by RIP3 and the pMLKL oligomerizes and translocates to the plasma membrane where it inserts and compromises plasma membrane integrity leading to plasma membrane rupture. Inhibition of pMLKL oligomerization by necrosulfonimide (NSA) protects cells from necroptotic rupture. MLKL has also been shown to promote the generation of reactive oxygen species (ROS) which aids in the progression of

TNF-α induced necroptosis [25]. There is also evidence that the necrosome can interact with the mitochondrial protein phosphatase PGAM5 leading to

Figure 1.3: TNF signaling

Binding of TNF-α to the TNF-R leads to the formation of complex I consisting of

TRADD and RIP1. RIP1 is ubiquitinated by TRAF2, TRAF5, cIAP1 and cIAP1 which leads to TAK1/IKK-mediated NFκB activation. In the presence of IAP inhibitors complex II forms with FADD, procaspase 8, and RIP1. Auto-processing of caspase 8 leads to activation of apoptosis. Pellino3 targets RIP1 to inhibit formation of complex II. Under caspase 8 inhibition RIP1 induces necroptosis and forms the necrosome with RIP3 leading to phosphorylation of both proteins, recruitment of MLKL, and MLKL phosphorylation by RIP3. Phosphorylated MLKL oligomerizes and inserts into the plasma membrane leading to membrane rupture. RIP3 can also interact with MLKL, PYGEL, GLUL, and GLUD1 to induce mitochondrial production of ROS. PGAM5 can also interact with the mitochondrial fission protein Drp1 leading to mitochondrial fragmentation.

Figure from Humphries. Cell Death and Differentiation, 2015. 22: 225-236.

For copyright permission see appendix 11.

Figure 1.3: TNF signaling

Humphries. Cell Death and Differentiation, 2015. 22: 225-236

16 mitochondrial fragmentation by the mitochondrial fission protein Drp1 in HT-29 cells [26]. However, recent studies have called into question the role of PGAM5 in TNF-α induced necroptosis. Knockdown of PGAM5 in murine fibroblasts showed no increase in cell viability after induction of necroptosis [27]. These two studies may highlight the cell type selectivity for effector mechanisms of the necroptotic pathways.

RIP1 and RIP3 are key components to the activation of necroptosis. They belong to the RIP kinase family, a seven member family that share a homologous

N terminal kinase domain (KD). RIP1, in addition to the KD, contains a C terminal

DD and an intermediate domain that includes the RIP homotypic interaction motif

(RHIM). RIP3 contains the N terminal KD and a RHIM domain at the C terminus.

RIP1 and RIP3 interact at the RHIM domains present on both proteins [17].

Structural analyses of this interaction indicate that the proteins form heterodimeric filamentous structures that are typical of beta-amyloids, and these amyloids form the active necrosome. The importance of the kinase activity of both RIP1 and RIP3 have been shown in numerous studies where loss of either

RIP1 or RIP3 results in cells resistant to TNF-α induced necroptosis. Use of necrostatin-1, an inhibitor of RIP1 kinase activity, has been useful in determining the importance of RIP1. Interestingly, knocking out RIP3 was able to rescue the embryonic lethality of caspase 8 knockout mice, and RIP1 knockout was able to rescue the embryonic lethality due to cardiac failure in FADD knockout embryos

[28, 29]. This indicates that not only are RIP1 and RIP3 vital to execution of

necroptosis, but also confirms a regulatory role of caspase 8 and FADD on the activation of this cell death pathway.

In addition to regulation by caspase 8, RIP3 has also been shown to be

regulated by RIP1. In the absence of receptor signals, RIP1 was shown to bind to

RIP3 thereby preventing the spontaneous oligomerization and activation of RIP3

mediated necroptosis [30]. Under these conditions, RIP1 kinase activity is not

necessary, but its role in inhibiting RIP3 mediated necroptosis may partially

explain the perinatal lethality of RIP1 knockout mice.

1.3 Chemotherapeutic Drugs

Chemotherapeutic drugs like doxorubicin (Dox) and etoposide (Etop) are

often used as the first line defense for treatment of malignant neoplasias. Often

targeted to the DNA, the goal of chemotherapies is to induce apoptosis or

senescence of the tumor cells. Commonly used in combination with other types

of chemotherapies, targeted therapies, or immunotherapies, chemotherapies are

the standard for treatment of many tumors. Additionally, both Dox and Etop have

been shown to trigger the purinergic anti-tumor immune response.

1.3.1 Doxorubicin

Doxorubicin (Dox), brand name Adriamycin, belongs to the anthracycline

antibiotic class of antitumor agents. Derived from the fungus Streptococcus

peucetius var. caesius, Dox and its family members contain a tetracyclic ring

structure attached to a sugar, daunosamine. They contain quinone and

hydroquinone moieties that are located on adjacent rings that allow for the gain

18 and loss of electrons. These drugs have the ability to intercalate into the DNA, but a more important mechanism is their ability to form a tripartite complex with topoisomerase II and DNA. As a topoisomerase II inhibitor, Dox inhibits the re- ligation of broken DNA strands leading to apoptosis. Dox is commonly used for malignant lymphomas and breast cancers [31].

Dox is administered intravenously at a recommended dose of 50 to 75 mg/m2 via a rapid single dose, repeated after 21 days. It is cleared by hepatic metabolism and biliary excretion with a multiphasic half-life of 3 hours and approximately 30 hours. It is metabolically converted to various aglycones and then eliminated [31].

The presence of the quinone moiety causes production of reactive oxygen species (ROS), especially in the presence of iron. The generation of ROS is thought to lead to the most serious clinical toxicity of both acute and prolonged use of Dox: cardiomyopathy. A single dose can produce an acute reversible cardiomyopathy whereas chronic use can lead to congestive heart failure that is unresponsive to digitalis. Other major dose-limiting complications include myelosuppression, thrombocytopenia, anemia, stomatitis, GI disturbances, and alopecia [31].

1.3.2 Etoposide

Etoposide (Etop), brand name Vepesid, is another topoisomerase II inhibitor whose mechanism of action is similar to that of the anthracyclines.

Derived from an extract from the mandrake plant, Etop is a member of the

Epipodophyllotoxins. It is used primarily for the treatment of testicular cancer but

is also effective in treating small cell carcinoma of the lung, non-Hodgkin’s

lymphomas, acute nonlymphocytic leukemia, and Kaposi’s sarcoma [31].

Etop is commonly administered intravenously at 50-100mg/m2 for 5 days

for testicular cancer, 50-120 mg/m2 intravenously for 3 days or 50 mg per day

orally for 21 days for the treatment of small cell carcinoma of the lung.

Intravenous administration should be done over a 30-60 minute period.

The dose-limiting toxicity of Etop is leukopenia. Other side effects include thrombocytopenia, nausea, vomiting, stomatitis, diarrhea, alopecia, fever, dermatitis, and hepatic toxicity. One major complication of Etop treatment was seen in patients with childhood acute lymphoblastic leukemia. In a long-term

follow-up these patients developed an unusual form of acute nonlymphoblastic

leukemia. This leukemia appeared 1-3 years after the end of treatment for the

acute lymphoblastic leukemia. Patients who received weekly or bi-weekly doses, with cumulative doses above 2000 mg/m2 were at a higher risk of developing the

leukemia[31].

1.3.3 Staurosporine

Staurosporine (STS) is an indolocarbazole alkaloid isolated from the

bacterium Streptomyces staurosporine. It is a broad protein kinase inhibitor, with potent effects against PKC activity and a strong affinity for Ser/Thr protein kinases as well as many tyrosine kinases. To date, STS has been shown to bind to 40 protein kinase domains. STS is able to activate intrinsic apoptosis in a

20 mitochondrial independent fashion by activating initiator caspase 9[32]. Due to its broad range of targets, STS proved to be too toxic for use as a therapeutic but because of its potent apoptotic induction it is widely used to study cell death signaling in a number of cell models [33].

Several analogs of STS have been developed and are currently undergoing clinical trials. Among them are Lestaurtinib, which is currently in phase III clinical trials for infants with newly diagnosed acute lymphoblastic leukemia, and 7-hydroxystaurosporine which is in phase II clinical trials for pancreatic cancer, breast cancer and lymphoma [33].

1.4 Release of Intracellular ATP

1.4.1 Purinergic Signaling

ATP is a vital energy source within the cell that drives effectively all cellular functions. However, in the 1970’s Geoffrey Burnstock proposed that extracellular ATP and its metabolites could act as neurotransmitters. He identified two types of extracellular receptors, P1 and P2, which recognize adenosine and ATP respectively [34, 35]. P2 receptors are further divided into ionotropic P2X receptors (P2XR) and metabotropic P2Y receptors (P2YR). The ability of ATP to bind to the P2 receptors is necessary for a multitude of cellular responses, including vasodilation and chemotaxis.

P2XR monomers have two membrane-spanning domains with a long extracellular binding loop and intracellular N and C termini. Trimers of P2XR monomers form the channel which allow for the passage of monovalent and

divalent cations such as K+, Ca2+, and Na+ upon binding of extracellular ATP.

There are seven known P2XRs (1-7) [36]. Binding of three ATP to the

extracellular loop promotes rearrangement of the P2XR subunits allowing for the

opening of the channel and the influx of Na+ and Ca2+ and the efflux of K+ leading

to depolarization of the cell as well as induction of Ca2+ signaling cascades. One well characterized P2XR, P2X7, leads to activation of the NOD-like-receptor- mediated assembly and activation of the inflammasome that ultimately induces activation and release of key inflammatory cytokines IL-1β and IL-18.

P2YRs are seven transmembrane G-protein coupled receptors (GPCR) with an extracellular amino terminus and an intracellular carboxy terminus. There are eight mammalian P2YRs that signal through Gq/11 to induce phospholipase C

mediated calcium release or alternatively via Gi/0 to inhibit adenylate cyclase and

modulate ion flow [36]. P2YRs induce a wide variety of cellular responses. P2Y2

has been shown to play a vital role in recruitment of leukocytes to the site of

infection and/or cell death. Under these circumstances, ATP acts as a ‘find-me’

signal, inducing chemotaxis of leukocytes to the site of damage, as well as

promoting phagocytosis of dying cells or pathogens.

1.4.2 Mechanisms of ATP Release

One of the earliest explorations into purinergic signaling was an

examination into ATP’s role as a neurotransmitter. This hypothesis was proposed

in 1972 and was one of the catalysts for the development of the study of

purinergic signaling [35]. ATP was identified as a neurotransmitter using non-

adrenergic, non-cholinergic inhibitory nerves and subsequently identified as a co- transmitter in the sympathetic and parasympathetic nervous system, as well as

both an individual and co-transmitter in the central and peripheral nervous

system. Since then, ATP has been shown to be released due to a multitude of

stimuli, including hypoxia, hypo-osmotic stress, mechanical stress and apoptosis.

While the initiation signals for ATP release may be diverse, there are three main release mechanisms: exocytosis, cell lysis, and channel mediated (Figure 1.4).

1.4.2.1 Exocytosis

ATP can be packaged into vesicles via anionic transporter proteins,

specifically SLC17A9 that transports ATP down a proton gradient. These vesicles

can then be released via calcium mediated exocytosis [37, 38]. Exocytotic

release of ATP is vital in neurons during neuro transmission as well as during

platelet aggregation.

1.4.2.2 Plasma Membrane Permeabilization

Opening of large pores within the plasma membrane or decreases in its

integrity may result in plasma membrane rupture. When a cell lyses all of the

intracellular contents, including ATP, are now extracellular. Stimuli such as

necrosis or osmotic swelling leading to lysis can trigger this mode of adenine

nucleotide release. Permeabilization of the plasma membrane is associated with

increased inflammation due to release of multiple DAMPs in addition to ATP [39].

1.4.2.3 Regulated efflux via channels

Figure 1.4: Mechanisms of ATP Release

ATP can be packaged into secretory lysosomes or auto-phagolysosomes and released via exocytosis. ATP can also be released through plasma membrane rupture. Finally, it can be released either by transient gating or proteolytic gating of membrane channels such as Panx1.

Figure 1.4: Mechanisms of ATP Release

Cell

ATP

Another mechanism of ATP release is via channels located in the plasma

membrane. Multiple channels have been implicated in the release of ATP due to

multiple stimuli including connexin 43, CALHM1 (calcium homeostasis modulator

1), and pannexin 1 (panx1).

Connexin 43 (Cx43) is a member of the connexin family of channel

proteins which form gap junctions that can transport ATP between cells. Cx43, however, is able to form a hemi-channel and has been proposed to be the conduit of ATP release during neutrophil activation [40]. Other connexin family members involved in ATP release include connexin 26 and connexin 31[41].

Release of ATP via CALHM1 is crucial for neuro-transmission of taste signals.

Upon stimulation of the taste receptors on type II taste cells, G-protein coupled receptors, are activated and lead to release of Ca2+ which triggers TRPM5 mediated depolarization and gating of the CALHM1 channel [42]. Panx1 has been identified as the main ATP release channel during multiple stimuli including hypo-osmotic stress, mechanical stress, and apoptotic cell death [43].

1.5 Pannexin 1 Channel

Panx1 is a four transmembrane domain spanning protein consisting of 426 amino acids that contains two extracellular loops and has both the amino and carboxy termini in the cytosol (Fig 1.5). It belongs to the pannexin family of proteins that include Panx1, Panx2, and Panx3 and is the most widely expressed. Six panx1 subunits combine to form the functional panx1 channel, termed the pannexon, which is synthesized in the ER/Golgi network. Trafficking

Figure 1.5: Pannexin 1 Channel

One panx1 subunit is a 4 transmembrane spanning protein, with 6 subunits forming the channel. Panx1 can be gated into an open channel and mediate passage of ATP.

Figure 1.5- Pannexin 1 Channel

of Panx1 from the ER/Golgi to the plasma membrane is regulated by post- translational modifications, specifically glycosylation. Panx1 is a glycoprotein consisting of one glycan chain with a consensus N-linked glycosylation site located on the second extracellular loop at N254. Panx1 exists in three major glycosylation states, Gly0, Gly1 and Gly2. Gly0 is the core un-glycosylated protein, Gly1 in the mono-glycosylated protein and is located mainly within the

ER. Gly2, the highly glycosylated form of the protein, is the most abundant form of the protein at the plasma membrane and is modified in the Golgi apparatus prior to trafficking to the plasma membrane. Glycosylation to the Gly2 state has been shown to be critical to the trafficking of Panx1 from the Golgi to the plasma membrane [44-46]. Gehi et al. demonstrated that not only is glycosylation necessary for trafficking but the C terminus of panx1 is needed for glycosylation of the protein from the Gly1 form to the Gly2 form [46]. Absence of either

glycosylation or the C terminus showed decreased trafficking of Panx1 to the

plasma membrane. However, while glycosylation to Gly2 yields maximum

delivery to the plasma membrane, the Gly0 and Gly1 forms of panx1 are also

able to translocate to the plasma membrane.

Once at the plasma membrane, the Panx1 channel, under basal

conditions, has a low conductance due to the flexible C terminus localizing inside

of the pore forming a functional plug. Upon stimulation due to mechanical stress,

hypo-osmotic stress or depolarization, the C terminus domain of Panx1 can be

transiently removed to gate the channel into an open conformation allowing the

flow of ions into and out of the cell. In addition to transient gating of the channel a

29 consensus caspase cleavage site located at 376Asp-Val-Val-Asp in human

Panx1 and 376Asp-Ile-Ile-Asp in murine Panx1 allowed for the proteolytic cleavage of the C terminus by the executioner caspases 3 and 7 [47, 48].

Described as a “ball-and-chain” mechanism of inhibition, the C terminus inhibition of the channel is relieved when cleavage at the caspase cleavage site releases the C terminal from the rest of the protein and diffuses away, leaving the channel in an open confirmation. To further the notion that the C terminus is acting as a plug for the channel, Sandilos, et al. created disulfide bonds between the C terminus and the trans-membrane 1 domain that would allow the C terminal to remain inside the pore even after cleavage [47]. Upon cleavage, the researchers found that with the C terminal still present in the pore, the channel activity remained low indicating the importance of removal of the C terminus from the pore. In addition, they injected large amounts of the free form of the C terminus to cells and found that the excess C termini were able to inhibit the channel confirming that the C terminus acts as a plug to inhibit Panx1 channel activity.

Panx1 channels are nonselective ion channels that have a single-channel conductance of ~500 picosiemens and allow for the passage of molecules up to

1kDa in size. Electrophysiological studies have indicated that Panx1 can be gated in response to plasma membrane potentials that are greater than 10mV, and direct mechanical stress at hyperpolarized (-50mV) potential [43]. One of the common ways to study Panx1 channel activity is to use influx of organic cation dyes such as YO-PRO2+ or Propidum Iodide (PI).

In addition to influx of organic cations, Panx1 has also been shown to release anions such as ATP. Chekeni et al. demonstrated that, during apoptosis, activation of executioner caspases lead to the cleavage of Panx1 and the release of ATP [49]. Knock-down of Panx1 along with pharmacologic inhibition of Panx1 showed decreased accumulation of ATP outside of the cell. The identification of

Panx1 as the ATP release channel was a key step in helping further the understanding of the interplay between damaged/dying cells and responding immune cells.

1.6 ATP metabolism

While extracellular ATP is an important molecule for activation of a pro- inflammatory response that is necessary to remove dying cells and bacteria, too much inflammation can be detrimental to the host. To reduce the duration of ATP mediated inflammation it is rapidly hydrolyzed to ADP, AMP and finally adenosine. Adenosine is an anti-inflammatory molecule. Adenosine binds to the

P1 purinergic receptors. There are four adenosine receptors, A1, A2A, A2B, and A3

[50]. These are G Protein coupled receptors that have unique tissue distributions.

Most notably, A2A and A2B receptors are Gs-coupled receptors that increase the production of cAMP and induce the downregulation of key pro-inflammatory cytokines such as TNF-α and IL-12 [51]. In addition, A2 adenosine receptors have been shown to shift the balance of type 1 T-helper (Th1) cells to type 2 T- helper (Th2) cells by decreasing the ability of dendritic cells to present antigen and initiate a Th1 response. Loss of A2 adenosine receptors in murine models leads to increased tissue damage due to increased inflammation highlighting the

importance of adenosine in regulating pro-inflammatory signaling. To reduce to

pro-inflammatory impact of ATP, ectonucleotidases convert ATP to adenosine

and are often upregulated in cancer.

1.6.1 Ecto-Nucleotide triphosphate diphosphohydrolase

Ecto-nucleotide triphosphate diphosphohydrolases (NTPDase) are a family of enzymes which sequentially hydrolyze ATP to ADP and ADP to AMP.

There are 4 ecto-NTPDase family members (NTPDase1, 2, 3 and 8) with

NTPDase1, also known as CD39, abundantly found in B lymphocytes, vascular endothelium, and smooth muscle cells [52]. Expression of CD39 on the vascular endothelium has been shown to decrease platelet aggregation. In neutrophils, monocytes, and specific T and B cell subsets, it serves as a key nucleotide- inactivating enzyme [53].

1.6.2 Ecto-nucleotide pyrophosphatase/phosphodiesterase

Ecto-nucleotide pyrophosphatase/phosphodiesterases (E-NPP) are a

family of seven ectoenzymes (1-7) capable of hydrolyzing pyrophosphate and

phosphodiester bonds on nucleic acids, nucleotides, phospholipids, and

nucleotide sugars. Only the first three members are able to target nucleotides. E-

NPP1 is expressed in cartilage, heart, kidney, liver, and testis but is most notably

involved in bone formation by generating pyrophosphate necessary for bone

mineralization [54].

1.6.3 CD73

CD73 is an ectoenzyme that belongs to the ecto-5’-nucleotidase family of proteins. CD73 is characterized by its ability to convert AMP to adenosine with a

Km of 3-50μM. It is an approximately 60-70 kDa protein that consists of two glycoproteins linked by noncovalent bonds and is bound to the cell surface by glycosyl-phosphatidylinositol (GPI) anchoring [55].

CD73 is expressed on a variety of tissues with high expression located in the kidney, brain, heart, colon, lung, and liver. While its expression on lymphocytes is restricted to selected T and B cell populations and correlates with maturity, it is often co-expressed with CD39. The combination of CD39 and CD73 leads to the breakdown of pro-inflammatory ATP to adenosine and leads to decreased T cell proliferation and inhibition of inflammation. Expression of CD73 can be induced by a number of molecules such as hypoxia-inducible factor-1α, interferon ϒ and HMGB-1 [53, 55].

CD73 is often upregulated in cancer and is a predictor of poor prognosis.

Increasing evidence also links an upregulation of CD73 in multi-drug resistant cancers leading to poor patient outcome [56]. While the mechanism for the upregulation of CD73 during chemotherapy-induced drug resistance is unknown, the increase points to an additional way cancer cells are able to evade cell death by decreasing immune responses.

1.7 Statement of Purpose

The subsequent studies aim to address the role of programmed cell death pathways in the release of ATP, ADP, and AMP as well as the generation of

33 adenosine within the tumor microenvironment by answering three main questions: 1) What is the mechanism of ATP release during chemotherapy treatment and are ADP and AMP released in addition to ATP, 2) what is the mechanism of necroptotic adenine nucleotide release, and 3) once released, what is the fate of the adenine nucleotides in cells that lack key cell death proteins. Taken together, these studies provide insight into the role of cell death signaling in the regulated release of nucleotides, as well as describe how perturbation in expression of key cell death proteins can modulate the accumulation of extracellular ATP and adenosine, and therefore the immune response.

First, we addressed the mechanism underlying ATP release after tumor cells are treated with chemotherapies. We demonstrate that chemotherapeutic treatment of cancer cells leads to the gating of the Panx1 channel via cleavage of the auto-inhibitory C terminus by the executioner caspases 3 and 7. Once gated, not only is ATP released via the channel, but ADP and AMP as well, and these two nucleotides comprise the majority of the nucleotide released.

Second, we examine the mechanism of nucleotide release during necroptosis. We demonstrate that during necroptosis the adenine nucleotide release is dependent on the insertion of pMLKL into the plasma membrane, and occurs simultaneously with cell lysis.

Finally, we analyze the fate of the nucleotides once released into the extra cellular space. We show that in FADD-deficient and RIP1-deficient Jurkat cells

there is increased ecto-nucleotidase activity causing an increased breakdown of

AMP. We propose that the process of mutation during chemotherapy treatment

can give rise to increases in adenosine production that ultimately leads to

decreased efficacy of the drug due to decreased immune response.

Together, these studies explore and highlight the release of ATP and its

breakdown during the course of chemotherapy treatment. This has implications

for the study of the purinergic anti-tumor immune response after chemotherapy treatment by underlining the importance of examining the ATP vs. adenosine balance and potential mechanisms of decreased drug efficacy via immune cell

evasion.

Chapter 2: Materials and Methods

Cell Models and Reagents—The human Jurkat T cell acutelymphocytic leukemia,

FADD deficient human Jurkat T cell acute leukemia cell line (I 2.1) and human

THP1 promonocytic leukemia lines were obtained from ATCC. The RIP1

deficient human Jurkat T cell acute leukemia cell line [57, 58] was generously

provided by Zheng-gang Liu (National Cancer Institute/ NIH). The human CEM T

lymphoblastic leukemia line was generously provided by Dr. Aaron Weinberg

(Case Western Reserve University). Normal human T lymphoblasts, expanded

from human peripheral blood leukocytes cultured with phytohemagglutinin and

IL-2 as described previously [59], were generously provided by Dr. Alan Levine

(Case Western Reserve University). Key reagents were obtained from the

following sources. Anti-human Fas (CH11 clone) was from Millipore; anti-human

CD3 (OKT3 clone) was from BioLegend; doxorubicin was from (LC Laboratories

or Sigma); staurosporine was from LC Laboratories; ATP, phosphoenolpyruvate,

etoposide, carbenoxolone, trovafloxacin, probenecid, and 3-methyladenine were

from Sigma; Necrosulfonamide (NSA), U73122 and ARL67156 were from Tocris;

Z-VAD-fmk was from Tocris or APExBio; MG132 and bafilomycin were from

APExBIO; lyophilized Firefly luciferase ATP assay mix (FLAAM), Firefly

luciferase ATP assay buffer (FLAAB), pyruvate kinase (P-1506), myokinase (M-

3003), 1,N6-etheno-AMP, DAPI stain, were from Sigma. EnzChek Caspase 3 assay kit, AlamarBlue® cell viability reagent, YO-PRO, LysoTracker Red-DND 99,

Fluo4-AM dyes, Calcein-AM, MitoTracker® Red CM-H2XRos were from

Invitrogen. TNF-α was from PeproTech. Cell Titer-Glo luminescent viability assay

kit was from Promega. The cytotoxicity detection kit (LDH release) was from

Roche Applied Science. αβ-methylene-adenosine 5’-diphosphate (APCP) was

from Jena Bioscience. Anti-PARP (9542) was from Cell Signaling. Anti-β actin

(sc-1615) and all horseradish peroxidase (HRP)-coupled secondary antibodies

were from Santa Cruz Biotechnology. Pierce® ECL Western blotting substrate

was from ThermoScientific. The noncommercial rabbit anti-human and murine

pannexin-1 antisera have been described previously [49, 60].

Cell Culture and Induction of Apoptosis— Wildtype, FADD-def, and RIP1-def

Jurkat cells were maintained in RPMI 1640 medium supplemented with 10%

newborn bovine calf serum (Hyclone), 100 units/ml penicillin, 100 μg/ml

streptomycin (Invitrogen), and 2 mM L-glutamine (Lonza) at 37 °C in 5% CO2.

EG7 murine cells were maintained in RPMI 1640 medium supplemented with

10% newborn bovine calf serum (Hyclone), 100 units/ml penicillin, 100 μg/ml

streptomycin (Invitrogen), 2 mM L-glutamine (Lonza) and 0.4 mg/mL G-418 at 37

°C in 5% CO2. For induction of apoptosis and analysis of released adenine

nucleotides, the cells were resuspended at 2x106 cells/ml (2 ml/well; 6-well plates) in RPMI 1640 medium as above but with 10% bovine calf serum treated for 2 h at 65°C to inactivate serum nucleotidases. In some experiments, cells were resuspended as above but in serum-free RPMI 1640 medium. The resuspended cells were preincubated for 1 h at 37 °C in 5% CO2 prior to induction of apoptosis by addition of 250 ng/ml anti-Fas, 3 μM staurosporine

(STS), 25 μM doxorubicin (Dox), 20 μM etoposide (Etop), or 3 μM MG132.

Where indicated, 100 or 500 μM carbenoxolone (CBX), 50 or 100μM Z-VAD, 50

nM bafilomycinA (BafA), 5mM 3-methyladenine, 10 μM trovafloxacin, 100 μM

ARL67156, or 10μM U73122 were added 30–60 min before the pro-apoptotic stimuli. For TNF-α induced apoptosis, the resuspended cells were pre-incubated with 3μM BV6 for 2 hours at 37°C in 5% CO2 prior to addition of 20ng/mL TNF-α.

Apoptosis in all Jurkat cell types was induced by 3μM STS, 20μM etoposide, or

250ng/mL anti-Fas. For induction of necroptosis, FADD deficient Jurkat cells were resuspended at 2x106 cells/mL and pre-incubated with 3μM BV6 for 2 hours for 1 hour at 37°C in 5% CO2 prior to addition of 20ng/mL TNF-α. Where indicated, 20μM, 50μM, or 100μM zVAD, 1μM NSA, or 30μM Trovafloxacin were added to the cell cultures 1 hour prior to addition of TNF-α, anti-Fas, STS, or

Etop. Where indicated, 50μM APCP was added 10 min prior the addition of TNF-

α, anti-Fas, STS, or Etop.

Collection of Conditioned Medium and Measurement of Released Adenine

Nucleotides—Samples of the conditioned medium supernatants were taken at

the indicated times (routinely 0, 2, 4, 8, 12, and 18 h) after addition of pro-

apoptotic stimuli and centrifuged at 13,500 rpm for 15 s to pellet any cells. The

cell-free supernatants were transferred to fresh tubes for analysis of either ATP

only or total adenine nucleotides (ATP+ADP+AMP). For luciferase-based

quantification of ATP only, a 50-μl aliquot of conditioned medium supernatant

was supplemented with 46 μl of FLAAB and 4 μl of concentrated FLAAM

(reconstituted with 5 ml of sterile water per vial of lyophilized luciferase/luciferin mix) and transferred to a well of a 96-well white plate. The ATP dependent

bioluminescence was measured with a BioTek Synergy HT plate reader (1 s

38 integration of emitted light) and quantified by comparison with ATP standards assayed under identical conditions. For quantification of total adenine nucleotides, samples were subjected to a protocol modified from Hampp et al.

[61] whereby AMP and ADP were rephosphorylated to ATP (in a cycling reaction driven by excess phosphoenolpyruvate in the presence of pyruvate kinase and myokinase) prior to the luciferase analysis described above. 50 μl of conditioned medium supernatant was supplemented with 8.3 μl of rephosphorylation mixture

(25 mM K-HEPES, pH 8.0, 50 mM MgSO4, 8.3 mM phosphoenolpyruvate, 600 units/ml myokinase, 300 units/ml pyruvate kinase) and incubated for 90 min at 37

°C. 37.7 μl of FLAAB and 4 μl concentrated FLAAM were then added to each resphosphorylated sample, and ATP dependent bioluminescence was measured.

Measurement of Intracellular ATP, ADP, and AMP Content— Jurkat cells were incubated in the absence or presence of anti-Fas (4 h), STS (4 h), Dox (8 h), or

Etop (8 h). Aliquots of cell suspension were centrifuged, and the cell pellets were resuspended in assay buffer (25 mM HEPES, 15 mM KOH, pH 8.0, 50 mM

MgSO4) and lysed by heating in a boiling water bath for 1 min. The lysates were centrifuged to sediment denatured protein, and 25-μl aliquots were immediately assayed for ATP, ADP, or AMP content by adaptation of the rephosphorylation protocol described above but with a separate pyruvate kinase-only incubation to assay ADP versus combined pyruvate kinase/myokinase incubation to assay

AMP. Quantification of each nucleotide (ATP, ADP, and AMP) in the lysates was determined relative to parallel rephosphorylation reactions containing known concentrations of ATP, ADP, or AMP standards.

Caspase-3 Activity—Jurkat cell suspensions were treated with pro-apoptotic stimuli as indicated above for the adenine nucleotide release experiments. At various times post-apoptotic induction, aliquots of cell suspension were centrifuged to pellet the cells. The cell pellets were washed, resuspended in

PBS, and then mixed with EnzChek Caspase-3 kit (Invitrogen) lysis buffer.

Caspase 3 activity in the cell lysates was assayed using caspase 3 reaction reagents as described in the vendor protocol.

Measurement of Cell Viability by AlamarBlue Metabolism or Intracellular ATP

Content—Cell viability was measured using the AlamarBlue Cell Viability reagent® (Invitrogen) as described in the vendor protocol. Quantification of the fluorescent resorufin product produced by viable cells was measured with the

BioTek Synergy HT plate reader using a 540/620-nm filter set. As an alternative assay of cell viability directly correlated with intracellular ATP, we used the Cell

Titer-Glo® luminescent cell viability assay reagent (Promega) as described in the vendor protocol. This assay reagent combines a cell lysis buffer and proprietary thermostable recombinant luciferase for quantification of cell viability based on

ATP content. At various times post-apoptotic induction, 25-μl aliquots of Jurkat cell suspensions were diluted to 100 μl with culture medium and mixed with 100

μl of reconstituted Cell Titer-Glo reagent per well of a 96-well white plate, and the

ATP-dependent bioluminescence was measured with the BioTek plate reader.

Measurement of Cell Lysis by LDH Release- Cell were treated with apoptotic or necroptotic stimuli as described above. After brief centrifugation of the cell suspensions, the cell-free supernatants were assayed for LDH enzyme activity

using the cytotoxicity detection kit according to the protocol of Roche Applied

Science. The released LDH was normalized to total LDH content measured in

1% Triton X-100-permeabilized samples of untreated Jurkat T cells.

Western Blot Analysis—1-ml aliquots of Jurkat cell suspension or EG7 cell

suspension (2x106 cells) were centrifuged, and the cell pellets were washed in

PBS. Whole cell lysates were prepared by detergent based extractions prior to

standard processing by SDS-PAGE (12% polyacrylamide), transfer to PVDF

membranes, and Western blot analysis as described previously [62]. Primary

antibodies were used at the following concentrations or dilutions: anti-human

Panx1 serum (1:5000), anti-murine Panx1 serum (1:5000), anti-PARP (0.05

μg/ml), and anti-actin (1 μg/ml). HRP-conjugated secondary antibodies were

used at a final concentration of 0.13 μg/ml. Chemiluminescent images of the

blots were developed with ECL reagent, imaged, and quantified using a

FluorChemE processor and AlphaView SA imaging software (Cell Biosciences).

YO-PRO Dye Uptake by End Point Assay—500-μl aliquots of Jurkat cell

suspension (106/ml) were treated with anti-Fas (4 h), STS (4 h), Etop (8 h), Dox

(12 h), MG132 (8 h) in the absence or presence of 100 μM Z-VAD, collected by centrifugation, and washed once with PBS. The washed cell pellets were resuspended in 500μl of basal salt solution (BSS) containing 130mM NaCl, 5mM

KCl, 1mM MgCl2, 1.5mM CaCl2, 25mM NaHEPES, pH 7.5, 5 mM glucose, and

0.1% bovine serum albumin. This suspension was divided into two 250-μl

aliquots. One was supplemented with 250 μl of BSS containing 200 μM CBX

(final concentration 100 μM), and the other was supplemented with 250 μl of BSS

lacking CBX. Both aliquots were preincubated at room temperature for 15 min

prior to addition of 1 μM YOPRO dye and incubation for an additional 20 min.

The cells were pelleted by brief centrifugation, washed once in PBS, and

resuspended in 250 μl of fresh BSS. 200-μl aliquots were transferred to wells in a

96-well black wall/clear bottom plate, and the fluorescence (485 nm/540 nm) was measured on the BioTek Synergy HT plate reader. Afterward, phase contrast and epifluorescence images of the cells in each well were viewed and recorded using a Zeiss Axiovert 25 microscope equipped with a 485/540-nm filter set,

QCam1394 digital camera, and QCapturePro imaging software (QImaging).

YO-PRO Dye Uptake by On-line Kinetic Assay—500-μl aliquots of Jurkat cell

suspension or EG7 cell suspension (106/ml) were suspended in BSS containing

5 mM glucose, 0.1% BSA, and 1 μM YO-PRO and transferred to the wells of a

24-well plate. The cell suspensions in some wells were additionally supplemented with 100 μM Z-VAD or 100 μM CBX, and the 24-well plate was

equilibrated to 37 °C in the BioTek Synergy HT plate reader. Baseline

fluorescence (485/540 nm) in each well was measured at 1-min intervals for 5

min prior to the addition of 3 μM STS (or vehicle). The fluorescence in each well was measured at 1-min intervals for an additional 4 h prior to the addition of digitonin (100 μM) to permeabilize all cells. Digitonin-induced increases in fluorescence were measured at 30-s intervals for 10 min.

Calcein Dye Efflux- Jurkat cells were suspended at 2x106cells/mL in RPMI+10%

heat inactivated calf serum+100 units/ml penicillin, 100 μg/ml streptomycin, and 2

mM L-glutamine. The suspensions were supplemented with1μM calcein-AM and

250μM probenecid and incubated for 45 minutes at 37°C. Cells were washed in

PBS, resuspended in the same RPMI test medium, and then treated with 3μM

STS in the presence or absence of 100μM zVAD or 30μM Trovafloxacine for 4

hours at 37°C and 5% CO2. After brief centrifugation, the cell pellets were

washed and resuspended in BSS+5mM Glucose+0.1% BSA and the

fluorescence (485nm/528nm) was measured on the BioTek Synergy HT plate

reader.

Fluo-4 Assay of Cytosolic [Ca2+]—Washed Jurkat cells were resuspended

(106/ml) in BSS supplemented with 1 μM fluo-4 AM (premixed with pluronic F-127

at 1:1 by volume) and 2.5 mM probenecid. Cells were incubated at 37 °C for 45

min, washed in PBS, and resuspended in fresh BSS in the absence or presence

of 2.5 mM probenecid. 500-μl aliquots were transferred to the wells of a 24-well plate that was placed into the BioTek Synergy HT plate reader preheated to 37

°C. Baseline fluorescence (485/540 nm) in each well was measured at 20-or 60-s

intervals for 10 min. Cells were then stimulated 1 μg/ml anti-CD3, 10 μM

ionomycin, 3 μM STS, or 5mM ATP for 0.5 or 4 h. Where indicated, cells were

treated with 5, 10, or 20 μM U73122 for 10 min prior to addition of anti-CD3. At the end of each assay, cells were permeabilized with 1% Triton X-100 to determine maximum Ca2+-dependent fluorescence and then supplemented with

15mM EGTA and 50mM Tris to determine Ca2+-independent fluorescence.

LysoTracker Red Accumulation Assay—1-ml aliquots of Jurkat cell suspension

(2x106/ml) in standard RPMI 1640 culture medium were transferred to the wells

of a 24-well plate and supplemented with or without 250 nM bafilomycin A or

100μM Z-VAD. After 60 min at 37 °C, the cells were treated with or without 3 μM

STS and incubated for an additional 4 h. The cell suspensions were collected for centrifugation, and the cell pellets were washed once in PBS, resuspended in

500 μl of BSS containing 100 nM LysoTracker Red, and transferred to the wells of a 24-well plate. LysoTracker Red fluorescence (540 nm/620 nm) in each well was immediately measured in the BioTek plate reader.

Imaging of Mitochondria- 1x106 Jurkat cells/mL were incubated for 4 h with or

without 3 μM STS. The cells were pelleted, washed with PBS, and resuspended

in pre-warmed (37°C) staining solution consisting of 500nM MitoTracker® Red

CM-H2XRos in BSS+5mM Glucose+0.1%BSA and incubated for 30 minutes at

37°C in the dark. Cell were pelleted, washed in PBS, resuspended in BSS+5mM

glucose+0.1% BSA supplemented with a 1:100 dilution of DAPI (0.2ug/mL

working stock), and incubated in the dark at room temperature for 5-10 minutes.

Cells were pelleted, washed in PBS, fixed in 200μL of 4% Paraformaldehyde in

PBS for 15 minutes at 37°C. The fixed cells were repelleted, washed, and

resuspended in 50μL 0.5% paraformaldehyde and stored at 4°C. Images of the

cells were acquired using an Olympus FluoView™ FV1000 laser scanning

confocal microscope coupled with an IX-81 inverted microscope equipped with a

60×/1.42 NA oil immersion objective and a computer running FluoView™ FV10-

ASW Ver. 3.1b software (Olympus America Inc., Center Valley, PA).

Extracellular ATP and AMP hydrolysis assays- Wild-type, FADD-def, or RIP1-def

cells were resuspended at 2x106 cells/mL in the RPMI assay medium for assay

of extracellular AMP hydrolysis or in BSS+5mM glucose+0.1% BSA for assay of

extracellular ATP hydrolysis. After addition of 1μM AMP or ATP, the cell

suspensions were incubated for 30 min at 37°C, and then centrifuged at 13,500

RPM for 15 sec to pellet cells. The cell-free supernatants were collected and

assayed for ATP or AMP by the luciferase-based methods described above.

Where indicated, 50μM APCP or 100μM ARL67156 was added to the cell suspension 10min prior to addition of AMP or ATP.

HPLC assay of extracellular etheno-adenosine accumulation- Wild-type, FADD-

def, or RIP1-def cells were suspended at 2x106 cells/mL in BSS+5mM

Glucose+0.1%BSA and pre-incubated for 15min at 37°C. The cell suspensions

were supplemented with 10μM 1, N6-etheno-AMP (ε-AMP), incubated for 30 min

at 37°C, and then centrifuged at 13,500 RPM for 15 sec. The cell-free

supernatants were transferred to new tube, boiled for 5 min at 100°C, and then

placed on ice. Extracellular etheno-adenosine (ε-ADO) and ε-AMP were resolved and quantified by ion exchange HPLC using a Bio-Rad Gradient Module HPLC connected to a Linear™ Fluor LC305 fluorescence detector. Samples were diluted 1:10 in water and 50 μL injected onto a Hamilton PRP-X100 anion exchange column which was eluted at 1.6 mL/min with a gradient of NH4HCO3 in

30% methanol. The baseline mobile phase was 0.25 M NH4HCO3 (pH 8.5) in

30% methanol. After sample injection, the mobile phase was developed linearly

from 0.25 to 0.275 M NH4HCO3 for 8 min, remained isocratic at 0.275 M

NH4HCO3 for the next 4 min, washed for 3 min at 0.425 M NH4HCO3 and then re- equilibrated back to the baseline mobile phase for 15 min. The elution times (in

45 minutes) were: ε-ADO, 4.9; ε-AMP, 8.5. Chromatogram peaks were obtained with

Data Ally v2.06 software.

Data Analysis—All experiments were repeated 2–4 times with separate Jurkat cell culture preparations. Figures illustrating western blot results are from representative experiments. Where indicated, figures illustrating quantification of extracellular adenine nucleotide accumulation, cell viability, intracellular adenine nucleotide content, caspase-3 activity, or YO-PRO dye uptake represent either the mean (-S.E.) of data from several identical experiments or the mean (S.E.) of triplicate independent cell suspension analyses from a single representative experiment. Quantitative results were analyzed by one-way analysis of variance with Tukey post-test comparison using Prism 3.0 software.

Chapter 3: Chemotherapeutic Drugs Induce ATP Release via Caspase-

gated Pannexin-1 Channels and a Caspase/Pannexin-1-independent

Mechanism†

3.1 Abstract:

Anti-tumor immune responses have been linked to the regulated release of ATP

from apoptotic cancer cells to engage P2 purinergic receptor signaling cascades

in nearby leukocytes. We used the Jurkat T cell acute lymphocytic leukemia

model to characterize the role of pannexin-1 (Panx1) channels in the release of

nucleotides during chemotherapeutic drug-induced apoptosis. Diverse pro-

apoptotic drugs, including topoisomerase II inhibitors, kinase inhibitors, and

proteosome inhibitors, induced functional activation of Panx1 channels via

caspase-3-mediated cleavage of the Panx1 autoinhibitory C-terminal domain.

The caspase-activated Panx1 channels mediated efflux of ATP, but also ADP and AMP, with the latter two comprising >90% of the released adenine nucleotide pool as cells transitioned from the early to late stages of apoptosis.

Chemotherapeutic drugs also activated an alternative caspase- and Panx1- independent pathway for ATP release from Jurkat cells in the presence of benzyloxycarbonyl-VAD, a pan-caspase inhibitor. Comparison of Panx1 levels indicated much higher expression in leukemic T lymphocytes than in normal, untransformed T lymphoblasts. This suggests that signaling roles for Panx1 may be amplified in leukemic leukocytes. Together, these results identify

† A version of this chapter was published in the August 11, 2014 edition of The Journal of Biological Chemistry. For copyright permission see appendix 11.

chemotherapy-activated pannexin-1 channels and ATP release as possible

mediators of paracrine interaction between dying tumor cells and the effector

leukocytes that mediate immunogenic anti-tumor responses.

3.2 Introduction:

Immune responses to cancer play significant roles in suppressing or

eliminating tumor cells that survive direct killing by chemotherapeutic agents or

radiation. Anti-tumor immune responses have been linked in part to the regulated

release of ATP from apoptotic tumor cells with consequent engagement of P2

purinergic receptor signaling cascades in tumor-infiltrating leukocytes [5, 6, 63].

Tumor cell-derived extracellular ATP stimulates G protein-coupled P2Y2 receptors on dendritic cells and macrophages to amplify the recognition and phagocytosis of apoptotic cancer cells [8, 64, 65]. ATP within the tumor microenvironment additionally stimulates ionotropic P2X7 receptors on dendritic cells/macrophages to elicit caspase-1 inflammasome assembly and IL-1β

secretion [5]. The ATP-dependent release of IL-1β from dendritic cells during

presentation of tumor antigen to CD8+ T cells enhances differentiation of the

latter into tumor-reactive effectors that can directly kill cancer cells or facilitate

recruitment of natural killer cells. ATP can be released from normal or tumor cells

via exocytosis of ATP-containing vesicles or granules, selective efflux via plasma

membrane channels, or nonselective disruption of plasma membrane integrity by

mechanical stress or cytolytic mediators [66, 67]. Two mechanistically distinct

pathways of ATP release, one channel-mediated and the other involving exocytosis, have been characterized in different tumor cell models during apoptosis induced by death receptors, radiation, or chemotherapeutic drugs.

Ravichandran and co-workers [10, 49] identified a central role for plasma membrane pannexin-1 channels in facilitating ATP export from apoptotic Jurkat human leukemic T cells or normal murine thymocytes. Pannexin-1 (Panx1) exhibits similar topology, but not sequence homology, to the larger family of connexin membrane proteins that form gap junction channels [45]. Each Panx1 protein subunit contains 426 amino acids organized in four transmembrane segments, intracellular N and C termini and two extracellular loops. Six Panx1 subunits co-assemble into stable hexameric channels during synthesis in the endoplasmic reticulum/Golgi network and trafficking to the plasma membrane. In the basal state, Panx1 channels are defined by very low open probability and conductance [47, 49, 68-71]. Increased Panx1 channel activity in the plasma membrane of nonapoptotic cells can be rapidly (within seconds to minutes) induced by mechanical stresses, stimulation of some G protein-coupled

receptors, or strong depolarization of the membrane potential to >0 mV [48].

Notably, the intracellular C-terminal tail of Panx1 contains a consensus site

(376Asp-Val- Val-Asp in human Panx1 and 376Asp-Ile-Ile-Asp in murine Panx1)

for proteolytic cleavage by caspase-3 [45, 47, 49, 69]. Cleavage at Asp-379

removes a 46-amino acid C-terminal segment that otherwise acts as an

autoinhibitory blocker of the conducting pore formed by the Panx1 hexameric

complex. Thus, accumulation of active caspase-3 during apoptosis can result in

49 essentially irreversible open-gating of Panx1 channels. These C-terminally truncated Panx1 channels are characterized by increased permeability to both organic anions, such as ATP, and organic cations [47, 49, 69], such as propidium family dyes. Studies with the Jurkat lymphocytes and mouse thymocytes demonstrated caspase-3-dependent activation of Panx1 channels and consequent ATP release in response to Fas-triggered induction of the extrinsic apoptotic cascade or UV irradiation-induced activation of the intrinsic apoptosis program [49, 72, 73].

A different pathway for ATP release from apoptotic tumor cells was elucidated by Zitvogel and co-workers [5] and Kroemer and co-workers [7] based on studies with multiple murine tumor models, including EL4/EG7 thymoma,

CT26 colon carcinoma, and MCA205 fibrosarcoma, as well as some human cancer cells such as U2OS osteosarcoma. Significant release of ATP from these tumor cells was induced by a broad range of chemotherapeutic drugs or pro- apoptotic agents, including anthracyclines, oxaliplatin, etoposide, and staurosporine. The extracellular accumulation of ATP was temporally correlated with a reduction in quinacrine-labeled intracellular puncta in the apoptotic tumor cells; quinacrine acts to fluorescently label ATP-containing acidophilic organelles or granules [7]. Subsequent studies revealed that chemotherapeutic drug induced ATP release was suppressed by pharmacological or genetic suppression of the autophagy regulators Atg5 and Atg7 and correlated with exocytosis of lysosomes [65]. The autophagy proteins acted to increase ATP compartmentalization within a subset of secretory lysosomes. Parallel signals in

the apoptotic cells promoted exocytotic fusion of these ATP-containing

lysosomes with the plasma membrane. Inhibition of caspase-3 or knockdown of

Panx1 attenuated lysosome exocytosis and secretion of ATP [62, 74]. However,

Panx1 in these tumor cell models was predominantly expressed as an

intracellular protein prior to apoptosis but translocated to the plasma membrane

concomitantly with apoptotic induction of lysosome exocytosis. In this model,

Panx1 channels act as regulators of the exocytosis of ATP-containing lysosomes rather than conduits for the direct efflux of cytosolic ATP [74].

In this study, we used the Jurkat T cell acute lymphocytic leukemia models to address unresolved mechanistic questions regarding the role of Panx1 channels in the release of ATP and other metabolites from tumor cells during chemotherapeutic drug-induced apoptosis.

3.3 Results:

Chemotherapeutic Drugs Induce Caspase-3-mediated Excision of the

Panx-1 Autoinhibitory Domain—Previous studies have defined the kinetics of

caspase-3-mediated gating of Panx1 channels during Jurkat cell apoptosis

induced by Fas receptor activation but not by chemotherapeutic drugs [49, 72,

73]. Using Fas activation as a positive control stimulus, we compared the kinetics

of caspase-3 activation and loss of cell viability in Jurkat cells treated with staurosporine (STS), a broad spectrum kinase inhibitor widely used as a model pro-apoptotic inducer, or two clinically relevant topoisomerase II inhibitors, Etop

and Dox. The three drugs were tested at concentrations known to be maximally

efficacious in inducing apoptotic death in the Jurkat model. Fig. 3.1A shows that

3 μM STS mimicked the ability of anti-Fas to trigger a rapid accumulation of

active caspase-3 within 2 h that preceded measurable loss of cell viability as

indicated by the capacity to efficiently metabolize Alamar Blue dye (Fig. 3.1B). In

contrast, significant caspase-3 activation required a 4-h incubation with 20 μM

Etop or an 8-h exposure to 25 μM Dox. As with the anti-Fas or STS stimuli, the

Etop- and Dox-induced accumulation of active caspase-3 preceded the loss of

cell viability. Notably, the decreases in Jurkat cell viability induced by STS, Etop,

Dox, or anti-Fas were markedly delayed and attenuated by the pan-caspase inhibitor, Z-VAD-fmk (Fig. 3.1C).

Asparagine 254 in the second extracellular loop of human Panx1 is N-

glycosylated during synthesis, and this glycosylation enhances trafficking of

mature Panx1 channels to the plasma membrane [44, 45, 60]. Three different

glycosylation states of Panx1 have been described: the Gly-0 core

unglycosylated protein; the Gly-1 high mannose species that predominates in the endoplasmic reticulum pool; and the Gly-2 complex glycoprotein that is the most abundant form in control Jurkat cells (Fig. 3.2A) as detected using a previously

characterized polyclonal rabbit antibody. This antibody recognizes residues 412–

426 within the autoinhibitory C-terminal domain downstream of the caspase-3/7-

cleavage site at Asp-379; two nonspecific bands in Jurkat lysates above and

below the three specific Panx1 bands were also labeled by this antiserum. We

compared the expression of Panx1 in whole cell lysates from control versus

apoptotic (20 μM Etop, 8 h) Jurkat cells (Fig. 3.2A). The C-terminal autoinhibitory

segment was efficiently cleaved from all three species of Panx1 in the apoptotic

cells. The cleaved C-terminal fragment was not detectable on PVDF membranes

after transfer from the 12% gels, presumably due to its small size (~5 kDa) and possible intracellular instability. Using Fas activation as a positive control stimulus (Fig. 3.2B), we compared the kinetics of this Panx1 proteolytic processing in cells treated with STS (Fig. 3.2C), Etop (Fig. 3.2D), or Dox (Fig.

3.2E) in either the absence or presence of Z-VAD. STS mimicked the ability of

anti-Fas to rapidly trigger significant proteolytic processing of Panx1 within 2 h

and complete removal of the autoinhibitory domain by 4 h. PARP1 is a canonical

substrate for the apoptotic executioner caspases-3/7, and the rapid Panx1

processing induced by STS (or anti-Fas) was temporally well correlated with

conversion of the 115-kDa full-length PARP1 to its 89-kDa cleavage product. The kinetics of Panx1 proteolytic processing in the Etop- and Dox-treated cells were also well correlated with PARP1 cleavage, but it was delayed relative to the actions of STS or anti-Fas (Fig. 3.2, D–F) and consistent with the slower accumulation of active caspase-3 elicited by these drugs (Fig. 3.1B). The efficacy of Z-VAD in inhibiting caspase-3-dependent processing of its cellular substrates was verified by the complete suppression of PARP1 cleavage in cells incubated with any of the three chemotherapeutic drugs or anti-Fas in the presence of Z-

VAD. Notably, Z-VAD was also able to completely suppress the proteolytic processing of Panx1 at each incubation time point in the anti-Fas-treated Jurkat cells (Fig. 3.2B). In contrast, the loss of full-length Panx1 glycoprotein species

induced by STS, Etop, or Dox was markedly delayed in the presence of Z-VAD

(Fig. 3.2,C–E with band density quantification in Fig.3. 2F). However, with

prolonged exposure (~12 h) to these agents, the presence of Z-VAD had little

effect on the final decrease in the levels of intact Panx1. These observations

might indicate that the C-terminal autoinhibitory domain can be removed by a

noncaspase family protease activated by chemotherapeutic agents but not death

receptors in Jurkat cells under conditions of arrested apoptotic signaling.

Alternatively, nonprocessed Panx1 channel proteins may be internalized and

routed for degradation by other mechanisms when cells are treated with these

drugs in the presence of caspase inhibitors. These alternatives are addressed

below.

Chemotherapeutic Drugs Induce Caspase-dependent Activation of

Pannexin-1 Channels—Using Fas activation as a positive control stimulus (Fig.

3.3, C and G), we compared the levels of plasma membrane Panx1 channel

activity in Jurkat tumor cells and G), or Dox (Fig. 3.3, F and G) in either the

absence or presence of Z-VAD. As illustrated in Fig. 3A, we assayed the influx of

YO-PRO dye (over a 20-min test period) as the readout of open-gated Panx1

channel function after incubation times that produced maximal accumulation of

active caspase-3 (Fig. 3.1A) and proteolytic excision of the Panx1 autoinhibitory

domain (Fig. 3.2, B—E) in response to each pro-apoptotic agent (4 h for STS and anti-Fas, 8 h for Etop, and 12 h for Dox). YO-PRO2+ is a 375-Da divalent cation and DNA-intercalating dye (629 Da as the diiodide salt) that is normally membrane-impermeant but readily conducted by open-gated Panx1 channels;

54 the green fluorescence of YO-PRO2+ markedly increases upon binding to intracellular nucleic acids. CBX blocks ion fluxes through open-gated Panx1 channels but does not inhibit the caspase 3-mediated cleavage of the autoinhibitory domain. Previous studies have validated the use of propidium dye influx and CBX blockade for the functional analysis of Panx1 channels [47, 49,

69, 75]. Consistent with previous studies and the kinetics of Panx1 proteolytic processing (Fig. 3.2B), Jurkat cells treated with anti-Fas for 4 h were characterized by robust increases in YO-PRO accumulation as indicated by qualitative fluorescence microscopy (Fig. 3.3C) or fluorescence plate reader- based quantification (Fig. 3.3G). Dye uptake was completely suppressed when

Z-VAD was present during the 4-h incubation with anti-Fas and was markedly attenuated when CBX was present during the subsequent 20-min YO-PRO influx period. Only a minor fraction of control Jurkat cells were YO-PRO-positive under identical assay conditions (Fig. 3, B and G). Jurkat cells treated with STS for 4 h

(Fig. 3.3, D and G), with Etop for 8 h (Fig. 3.3, E and G), or with Dox for 12 h

(Fig. 3.3, F and G) were characterized by YO-PRO accumulation responses that were qualitatively and quantitatively similar to those observed with anti-Fas stimulation. Similar magnitudes of YoPro uptake were observed in cells treated with Etop for 8 h (Fig. 3.3G) or 12 h (data not shown) consistent with the near- maximal Panx1 cleavage induced by Etop at 8 h. Co-incubation with Z-VAD suppressed the ability of each drug to induce increased YOPRO uptake, although inclusion of CBX during the dye influx assay decreased net accumulation and the numbers of YO-PRO bright cells. It is important to note that

intracellular accumulation of the intensely red doxorubicin quenches the emitted

green fluorescence of the YO-PRO·DNA complexes and thus decreases the

absolute fluorescence measured in the plate reader assay (Fig. 3.3G, last data

set on right). However, similar numbers of YO-PRO-positive cells, albeit with dimmer intensity per cell, were observed by fluorescence microscopy of the Dox- treated samples (Fig. 3.3F). To determine the onset of gating of Panx1 channels with significant permeability to YO-PRO, a real time/on-line YO-PRO uptake assay was performed using STS and anti-Fas as rapid apoptotic inducers.

Significant YOPRO influx in suspensions of Jurkat cells treated with STS (Fig.

3.3H) or anti-Fas (data not shown) occurred after a 90-min lag time, and this influx was completely suppressed in the presence of Z-VAD and markedly blocked by CBX. These data taken together indicate that pro-apoptotic chemotherapeutic drugs induce accumulation of functionally active Panx1 channels in hematopoietic tumor cells via a caspase-3-dependent activation mechanism.

Caspase-gated Panx1 Channels Mediate Robust Efflux of AMP and ADP in Addition to ATP—When Jurkat T cells were treated with anti-Fas, STS, Etop, or Dox, intracellular ATP levels decreased (Fig. 3.4A) with time courses and delay phases similar to those describing the proteolytic processing of Panx1 induced by those agents. Z-VAD markedly delayed the decrease in intracellular

ATP in STS-treated cells and completely suppressed it in cells incubated with anti-Fas, Etop, or Dox (Fig. 3.4B). The decreases in intracellular ATP were also correlated with decreases in the ratios of intracellular [ATP]:[ADP]:[AMP] during

apoptotic progression induced by anti-Fas or the chemotherapeutic drugs (Fig.

3.4C). The decrease in intracellular ATP could reflect ATP efflux via the activated

Panx1 channels, its intracellular metabolism to ADP and AMP, or a combination

of the two processes. We hypothesized that active Panx1 channels will facilitate

ATP efflux during the initial phases of apoptotic signaling but will increasingly mediate release of the metabolites ADP and AMP with ongoing apoptotic progression (Fig. 3.4D). Given the rapid and robust decrease in intracellular ATP induced by the 4-h exposure to STS, we measured the corresponding extracellular accumulation of ATP only (Fig. 3.4E) versus the sum of extracellular

ATP, ADP, and AMP (ANex) (Fig. 3.4F) in medium conditioned for 4 h by STS- treated Jurkat cells. ANex accumulation was quantified by processing extracellular samples with a myokinase and pyruvate kinase-based

rephosphorylation protocol prior to assaying total ATP via luciferin-luciferase. We

additionally tested how the relative magnitudes of STS-induced ATP versus

ANex accumulation might be modulated in the presence of ARL67156, an

inhibitor of the CD39 family ecto-nucleotidases that serially metabolize extracellular ATP to ADP and ADP to AMP [76]. Only 5 nM extracellular ATP

accumulated in medium conditioned for 4 h by control Jurkat cells and ARL67156

induced a 2-fold increase in this basal ATP accumulation. STS treatment

increased extracellular ATP accumulation by 12-fold to 60 nM, and this was

modestly potentiated to 80 nM in the presence of the ecto- ATPase inhibitor.

Notably, in both the absence and presence of ARL67156, the summed

concentrations of the extracellular ATP + ADP + AMP species at this 4-h time

57 point after STS stimulation increased to the 1200–1500 nM range that was 20- fold higher than the levels of ATP only. These data indicate that apoptotic Jurkat cells directly release large amounts of ADP and AMP in addition to ATP per se, rather than generating most of this extracellular ADP/AMP as a secondary consequence of ecto-nucleotidase action on the released ATP.

We next compared the time courses of extracellular ATP versus ANex accumulation during apoptotic induction by anti-Fas, STS, Etop, and Dox.

Consistent with previous reports, anti-Fas induced a rapid 10-fold increase in extracellular ATP within 4 h to a 60nM plateau (Fig. 3.5A); after 8 h, the extracellular [ATP] gradually decreased due to the modest Jurkat cell ectonucleotidase activity described in Fig. 3.4E. The summed concentration of the three extracellular adenine nucleotide species within 4 h after Fas stimulation increased to 1200 nM, a 20-fold higher level than the ATP only; this further increased by ~2-fold over the next 14 h (Fig. 3.5E). Similar to anti-Fas, STS also triggered a rapid 10-fold increase in extracellular ATP within 4 h that dissipated over the next 14 h of apoptotic progression (Fig. 3.5B). Again, the magnitude of total ANex accumulation was much greater (1800 nM) than that of ATP only during the early phase of STS-induced apoptosis and continued to increase over the next 18 h (Fig. 3.5F). Corresponding with their slower rates of caspase-3 activation and proteolytic processing of Panx1, Jurkat cells incubated with Etop or Dox were characterized by 4–6-h lag phases before releasing significant amounts of ATP (Fig. 3.5, C and D) or total ANex (Fig. 3.5, G and H) into the extracellular compartment. However, following these lag phases, the rates and

magnitudes of ATP release and total ANex accumulation were similar to those

observed with anti-Fas or STS stimulation. The onset of ATP accumulation was

co-temporal with summed ANex accumulation in the Etop- and Dox-treated cells, but the peak amounts of released ATP were 20 –30-fold lower than the total

ANex. These data suggest that a changing mixture of ATP, ADP, and AMP is released from the Jurkat tumor cells during apoptotic progression with ADP and

AMP comprising the predominant species at later times.

Chemotherapeutic Drugs Induce Caspase-gated ATP Release via Panx1 and an Alternative ATP Release Pathway in the Presence of Caspase

Inhibition—To characterize the role of caspase-3-gated Panx1 channels in the accumulation of extracellular ATP and total ANex, Jurkat cells were treated with anti-Fas or the three pro-apoptotic drugs in the presence of Z-VAD. The accumulation of both ATP and summed ANex in anti-Fas treated cell cultures was completely suppressed by Z-VAD at all times during the 18-h incubation

(Fig. 3.5, A and E). Likewise, Z-VAD completely blocked the release of ATP and accumulation of ANex during the early phase (≤4 h) of STS treatment (Fig. 3.5, B

and F). However, with sustained (>4 h) exposure to STS plus Z-VAD, Jurkat cells

were characterized by a prolonged phase of ATP release and ANex

accumulation over the next 14 h. Although the absolute magnitude of STS-

induced ANex accumulation was markedly reduced in the presence of Z-VAD,

the ratios of extracellular ATP/ANex in the STS + Z-VAD-treated cell cultures

were invariably greater than those measured in the matched STS only-treated

cell cultures. For example, at 12 h, extracellular ATP/ANex was 0.19 in the

STS+Z-VAD-treated cells versus 0.008 in the STS-stimulated cells. In the

presence of Z-VAD, Etop treatment also elicited a sustained release of ATP that,

after 8 h, exceeded that measured in the absence of Z-VAD (Fig. 3.5C). As with

STS, Z-VAD markedly reduced but did not eliminate ANex accumulation in Etop treated cultures (Fig. 3.5G); the ratio of extracellular ATP/ANex in Etop-treated

Jurkat cells was increased by Z-VAD at each time point after 8 h. In Dox-treated cells, co-incubation with Z-VAD almost completely suppressed the extracellular accumulation of ATP only (Fig. 3.5D) and the summed ANex (Fig. 3.5H).

However, a modest Z-VAD-resistant ATP release was observed at the 18-h time point after addition of Dox (Fig. 3.5D) suggesting delayed activation of the caspase-insensitive ATP secretion response. These data suggest that chemotherapeutic agents can activate an alternative caspase-insensitive mechanism for secretion of ATP only (or predominantly) and that this response that is engaged in the context of caspase inhibition prevents gating of Panx1 channels that facilitate efflux of AMP and ADP in addition to ATP.

To verify that the Z-VAD-sensitive nucleotide release was mediated by activated Panx1 channels, we compared extracellular accumulation of ATP or summed ANex in Jurkat cultures incubated with the four pro-apoptotic stimuli in the absence or presence of CBX (Fig. 3.6). For each agent, we assayed the CBX sensitivity at time points where Z-VAD-sensitive ATP release was maximal (from

Fig. 3.5: 4 h for anti-Fas and STS, 8 h for Etop, and 12 h for Dox). Because CBX by itself induced modest increases in extracellular ATP or ANex in control (Con) cells, we quantified the CBX-sensitive responses in the apoptotic stimulus (Stim)-

60 treated cells using the equation: [Stim plus CBX] - [Con plus CBX]/[Stim - CBX] -

[Con - CBX]. In anti-Fas-treated cells, CBX reduced ANex by 81% (Fig. 3.6A) and ATP release by 68% (Fig. 3.6E). In STS-treated cells, CBX reduced ANex by

95% (Fig. 3.6B) and ATP release by 65% (Fig. 3.6F). In Etop-treated cells, CBX reduced ANex by 90% (Fig. 3.6C) and ATP release by 95% (Fig. 3.6G). Finally, in Dox-treated cells, CBX reduced ANex by 86% (Fig. 3.6D) and ATP release by

59% (Fig. 3.6H). Thus, CBX-sensitive Panx1 channels include the predominant mechanism for release of ATP, ADP, and AMP from anti-Fas- or chemotherapeutic drug-treated Jurkat cells with intact apoptotic caspase signaling cascades.

Caspase-insensitive ATP Release Is Resistant to Carbenoxolone

Blockade but Suppressed by Intracellular Ca2+ Buffering— We next considered possible mechanisms (Fig. 3.7A) for the alternative caspase-insensitive ATP release observed in Jurkat cells treated with STS (Fig. 3.5B) or Etop (Fig. 3.5C) in the presence of Z-VAD. In addition to excision of the autoinhibitory domain,

Panx1 channels can be reversibly activated by other stimuli, including G protein- coupled receptor signaling or mechanical perturbation of the subplasma membrane cytoskeleton [48, 77, 78]. We tested whether noncleaved Panx1 channels might be gated by signals that progressively accumulate in Jurkat cells treated with chemotherapeutic drugs in the context of suppressed caspase signaling. These experiments used STS as the model stimulus given its more rapidly induced actions and tested for Panx1 channel involvement based on the known ability of CBX to block YO-PRO or ATP fluxes through activated but

noncleaved Panx1 channels. Fig. 3.7B illustrates the typical several fold increase

in ATP release (at 8 h) induced by combined STS + Z-VAD treatment versus

STS alone. The additional presence of CBX did not prevent this enhanced ATP release but further increased it by 2.5-fold (CBX also modestly enhanced STS-

induced ATP even in the absence of Z-VAD). This potentiating action of CBX on

ATP accumulation was also observed in cells co-treated with STS plus Z-VAD for

12 h (data not shown). In the same conditioned samples, we observed only

inhibitory effects of CBX, Z-VAD, or combined Z-VAD+CBX on the STS-

stimulated extracellular accumulation of the summed adenine nucleotide species

(Fig. 3.7C). We also tested the effects of trovafloxacin that has recently been

identified as another pharmacological inhibitor of Panx1 channels (19). Similar to

the CBX experiments, the presence of 30 μM trovafloxacin (trova) did not

antagonize the STS + Z-VAD-induced ATP release at 8 h (STS only, 14.5 ± 3.1

nM; STS ± Z-VAD, 115.6 ± 6.6 nM; STS + trova, 13.5 ± 0.9 nM; STS + Z-VAD +

trova, 109.4 ± 6.5 nM, n = 6 from two experiments). As expected, trovafloxacin

markedly inhibited the caspase-dependent release of summed adenine

nucleotides induced by an 8-h exposure to STS (STS only, 1589 ± 197 nM; STS

+ Z-VAD, 372±18 nM; STS + trova, 616±56 nM; STS+Z-VAD+ trova, 380 ± 24

nM, n= 6). These data indicate the caspase-independent ATP release pathway is

also independent of Panx1 channel activity.

Recent studies have indicated that several cell types, including astrocytes

(32), Jurkat T cells (33), and THP1 monocytes (34), have a subpool of ATP-

containing lysosomes that can be mobilized for rapid exocytosis in response to

increases in cytosolic [Ca2+] (Fig. 3.7A). We tested the effects of either deleting

the normal 1.5 mM CaCl2 from the extracellular saline or loading Jurkat cells with

BAPTA, a high affinity Ca2+ chelator that buffers and blunts increases in cytosolic

[Ca2+]. Although the absence of extracellular Ca2+ did not attenuate the caspase independent ATP release (Fig. 3.7D), BAPTA loading completely suppressed this

response (Fig. 3.7E). BAPTA also partially attenuated the extracellular accumulation of the summed adenine nucleotide species triggered by STS in the absence of Z-VAD (Fig. 3.7F), suggesting possible modulation of the caspase-3/

Panx1 signaling axis by Ca2+. The potent inhibitory effect of BAPTA loading

implicates a critical role for increased cytosolic [Ca2+] in supporting the caspase-

independent ATP release.

We tested whether this Ca2+-dependent response involved exocytotic

mobilization of secretory lysosomes or other secretory granules. ATP is

concentrated within lysosomes or specialized secretory granules by the

SLC17A9 anion transport protein [38]. The ability of SLC17A9 to actively

transport nucleotides is energized by the organellar membrane potential (Δψ;

inside positive) and ΔpH established by the bafilomycin A (BafA)-sensitive

vacuolar proton ATPase. Thus, we compared STS+Z-VAD-induced ATP secretion in Jurkat cells incubated in the absence or presence of BafA. Initial experiments verified the efficacy of BafA to decrease lysosomal/vesicular acid loading in Jurkat cells by measuring its ability to reduce intracellular accumulation of the acidophilic LysoTracker Red dye under the experimental conditions used to stimulate ATP release (Fig. 3.7G). Interestingly, apoptotic induction by STS

also reduced LysoTracker Red accumulation, and this effect was reversed in the presence of Z-VAD. Despite the ability of BafA to collapse the ΔpH across

lysosomal/vesicular membranes, it did not suppress the caspase-independent

ATP release response (Fig. 3.7H).

Exocytosis of ATP-containing autophagolysosomes has been also been

linked to chemotherapeutic drug-induced ATP secretion [74, 79].

Autophagosome biogenesis requires the activity of class III phosphatidylinositol

3-kinases that are pharmacologically inhibited by 3-methyladenine [80, 81].

However, we observed that inclusion of 5 mM 3-MA during stimulation of Jurkat cells with STS + Z-VAD for 8 h did not reduce ATP release but rather increased this response by 3-fold (Fig. 3.7I). Similar effects of 3-MA on the STS + Z-VAD- induced ATP accumulation were observed when the incubation was extended to

12 h (data not shown). Taken together, the findings illustrated in Fig. 3.7, G–I, argue against an obvious role for exocytosis of secretory lysosomes or autophagolysosomes in supporting the caspase-and Panx1-independent ATP release elicited by pro-apoptotic drugs in the presence of Z-VAD.

The suppressive effect of BAPTA-loading on the STS + Z-VAD-induced

ATP release suggested that STS may directly or indirectly elicit mobilization of intracellular Ca2+ stores. We tested whether STS might elicit global changes in

cytosolic [Ca2+] using Jurkat cells loaded with the fluo-4 indicator dye (Fig. 3.8B).

In initial experiments, we observed rapid (within 60 min) and massive efflux of pre-loaded cytosolic fluo-4, which resulted in increased fluorescence upon binding extracellular Ca2+ (Fig. 3.8A). This constitutive fluo-4 efflux was mediated

by organic anion transporter activity because it was suppressed by millimolar

probenecid (which is also a Panx1 channel blocker). Thus, the measurement of

cytosolic [Ca2+] in Jurkat cells over prolonged (>30 min) incubations requires the

presence of probenecid. Under these assay conditions, no significant elevation of

cytosolic [Ca2+] was observed in Jurkat cells during a >3-h incubation with STS

(Fig. 3.8B). This contrasted with the sustained increase in [Ca2+] induced by

ionomycin and the robust, but transient, Ca2+ mobilization triggered by T cell

receptor (TCR) activation with anti-CD3 OKT antibody. As illustrated in Fig. 3.5C, these Jurkat cells exhibited a modest increase in cytosolic [Ca2+] in response to 5

mM ATP, which will activate the low ATP affinity P2X7 purinergic receptors

expressed in these cells, but no response to 50μM ATP. Although the Fig. 3.8B

data argue against direct Ca2+-mobilizing actions of STS, it is possible that the longer (4–8 h) STS treatment times associated with the caspase-independent

ATP release cause elevated Ca2+ by indirect mechanisms. These could involve

the STS-induced release of autocrine mediators (e.g. bioactive lipids) that target

Gq- or Gi-coupled receptors with consequent activation of phosphatidylinositol

phospholipase C (PI-PLC) effector enzymes and 1,4,5-inositol trisphosphate-

sensitive Ca2+ stores. This was investigated using U73122, an inhibitor of both G protein-regulated PI-PLCβ enzymes and the tyrosine kinase regulated PI-PLCγ

effectors in most cell types, including Jurkat T cells [82, 83]. Fig. 3.8D shows that

5–20 μM U73122 markedly inhibited the PI-PLCγ-mediated Ca2+ mobilization

response to TCR activation. However, 10 μM U73122 produced no obvious

attenuation of the caspase-independent ATP release induced by co-treatment

with STS + Z-VAD (Fig. 3.8E). Additionally, this PLC inhibitor had no effect on the

caspase-sensitive accumulation of summed ANex observed in STS-treated

Jurkat cells (Fig. 3.8F).

Proteosome Inhibitors Induce Caspase-3 Gating of Panx1 Channels and

ATP Release—As indicated in Fig. 3.2, chemotherapeutic drugs induced a slow

decrease in Panx1 levels in the presence of Z-VAD, which prevents caspase-

mediated excision of the C-terminal inhibitory segment. This suggested that

Panx1 channels may be degraded by other mechanisms in cells treated with

chemotherapeutic drugs. We investigated possible involvement of the

proteasomal pathway by treating Jurkat cells for 12 h with STS, Etop, or Dox in

the absence or presence of various combinations of Z-VAD and/or MG132, a

proteasome inhibitor. MG132 per se is a potent activator of apoptosis in various

tumor cell types, including Jurkat cells [84], and proteasome inhibitors are used

clinically as cancer chemotherapeutics [85, 86]. MG132 alone induced cleavage of the Panx1 autoinhibitory domain, which was co-temporal with proteolytic

processing of PARP1 (Fig. 3.9, A and B). Although Z-VAD markedly attenuated

the MG132-induced decrease in full-length (i.e. with the retained autoinhibitory

segment) Panx1 levels, the overall expression of intact Panx1 protein remained

lower than in control Jurkat cells or cells treated with only Z-VAD. Likewise, the additional presence of MG132 did not prevent the down-regulation of full-length

Panx1 protein content was observed in cells treated with STS, Etop, or Dox plus

Z-VAD (Fig. 9, A and B). This argues against a role for proteasome-mediated degradation of Panx1 channels.

Consistent with the ability of MG132 to induce caspase-3-mediated cleavage of the Panx1 autoinhibitory domain, MG132-treated Jurkat cells were also characterized by time-dependent and Z-VAD-inhibitable increases in Panx1 channel activity as indicated by release of adenine nucleotides (Fig. 3.9, C and

D) and accumulation of YO-PRO dye (Fig. 3.9, E and F). Both indices of MG132- stimulated Panx1 activity were also blocked in the presence of CBX.

Panx1 Is More Highly Expressed in Human Leukemic Leukocytes than in

Normal Human T Cells—We compared the expression levels of Panx1 protein in

Jurkat cells, two other human leukemia cell lines (CEM T lymphoblastic leukemia and THP1 promonocytic leukemia), and normal human blood-derived T lymphoblasts (Fig. 3.10). Although the Panx1 content of the CEM and THP1 cells was similar to that in Jurkat cells, the normal T cells expressed an ~20-fold lower level of Panx1 relative to the leukemia cells. Treatment of the T lymphoblasts with anti-Fas or STS further decreased the immunoreactive Panx1 bands indicating cleavage of the autoinhibitory domain.

3.4 Discussion:

This study provides several new mechanistic insights regarding the role of

Panx1 channels in the release of adenine nucleotides from apoptotic tumor cells.

First, our experiments demonstrate that diverse pro-apoptotic chemotherapeutic agents, including topoisomerase II inhibitors, kinase inhibitors, and proteosome inhibitors, induce functional activation of Panx1 channels via caspase-3-mediated cleavage of the Panx1 autoinhibitory C-terminal domain. This observation

extends the seminal demonstration by Chekeni et al. [49] regarding the caspase

mediated gating of Panx1 channels during apoptosis initiated by death receptors

or UV irradiation. Second, we found that the profile of adenine nucleotide species

transported through these caspase-activated channels includes ATP but also its

immediate ADP and AMP breakdown products, with these metabolites

comprising >90% of the released adenine nucleotide pool that progressively

accumulates as cells transition from the early to late stages of apoptosis. Third,

other experiments revealed that chemotherapeutic drugs, but not the Fas death

receptor, activate an alternative caspase- and Panx1-independent pathway for

ATP release from Jurkat leukemic T cells in the context of suppressed caspase activity. Thus, although caspase inhibition represses the ATP export normally mediated by the proteolytically gated Panx1 channels, it facilitates other

chemotherapeutic drug-induced signaling pathways that entrain the alternative

ATP release mechanism. Finally, comparison of Panx1 levels in different human

leukocyte types indicated much higher expression in leukemic cells (lymphoid or

myeloid) than in normal, untransformed T lymphoblasts. This suggests that

various signaling functions of Panx1 may be amplified in leukemic leukocytes.

Caspase-mediated Gating of Panx1 Channels in Response to Diverse

Chemotherapeutic Agents—Anti-Fas and the four different chemotherapeutic

drugs (STS, Etop, Dox, and MG132) used in this study initiate and drive

apoptotic progression in Jurkat cells with differing and distinctive lag times and

rates of activation. This allowed us to establish that the kinetics of Panx1

proteolytic modification and accumulation of functionally active Panx1 channels

68 were precisely correlated with, and limited by, the rates of active caspase-3 accumulation in response to each pro-apoptotic drug. For example, topoisomerase inhibitors, such as Etop or Dox, initially trigger DNA damage, which then drives transcriptional cascades that alter the expression of the various pro- and anti-apoptotic Bcl2 family proteins, which ultimately control the rate at which cytochrome c is released from mitochondria to induce caspase-9 apoptosome assembly [32, 87]. Accumulation of apoptosome complexes will be rate-limited by the upstream transcriptional cascades for Bcl2 family protein expression and, in turn, will rate-limit activation of the downstream executioner caspases-3/7. Integration of the lag times between these multiple steps upstream of caspase-3 activation per se likely underlies the 4-h delay before significant cleavage of the Panx1 autoinhibitory domain is observed in Etop- or Dox-treated

Jurkat cells. In contrast, proteolytic excision of the Panx1 autoinhibitory domains with consequent accumulation of active Panx 1 channels was near-maximal within 90 min after addition of STS. This was similar to the kinetics of Panx1 activation by the Fas death receptor that rapidly engages the FADD/caspase-8 cascade to drive accumulation of active caspase-3. This very rapid activation of

Panx1 in response to STS is consistent with recent studies in Jurkat cells [32] and mouse embryo fibroblasts [88], which found that STS induced activation of caspase-9 independently of Bcl2 proteins or apoptotosome assembly. This underlies the unusually rapid induction of active caspase-3 and apoptotic progression widely observed in many STS-treated cell types.

We also observed that the chemotherapeutic agents, but not anti-Fas, induced a slow down-regulation of Panx1 protein levels in the presence of Z-

VAD. Previous studies have indicated that cell surface Panx1 in healthy cells is a

very stable protein characterized by only slow rates of internalization and

endosomal trafficking [46, 60, 89-91]. However, point mutations, changes in

glycosylation status, or disruption of the actin cytoskeleton alter Panx1 trafficking

to result in increased intracellular accumulation and lysosomal degradation [46,

60, 91]. Given that mutated Panx1 protein is targeted for proteosomal degradation [46], we tested whether the combination of chemotherapeutic drugs plus caspase inhibition might also promote proteosomal clearance of Panx1.

However, the additional presence of MG132 did not rescue Panx1 expression levels.

Dynamic Changes in the Profile of Adenine Nucleotide Species Exported via Active Panx1 Channels during Apoptotic Progression— Release of ATP from multiple cell types due to diverse stimuli, such as mechanical stress or osmotic swelling, has been extensively studied [67]. Less understood is the release of other major adenine nucleotide species, ADP and AMP, which are maintained at much lower steady-state concentrations in the cytosol of viable cells. Previous studies of Panx1 channels in apoptotic cells have focused on their ability to facilitate the efflux of cytosolic ATP into extracellular compartments and predominantly assayed accumulation of extracellular ATP as the readout of this activity [49, 72]. However, apoptosis is an energy-consuming process that will

increase the rate of intracellular ATP utilization in cells with a decreased capacity

70 for mitochondrial ATP synthesis due to the changes in mitochondrial function that define the apoptotic phenotype. Even in tumor cells wherein glycolysis, rather than oxidative phosphorylation, may include the major pathway for ATP production, disruption of mitochondrial integrity will perturb the NAD/NADH cycling required for optimal glycolytic metabolism. Thus, apoptotic induction will perturb the steady-state bioenergetic networks that normally maintain high intracellular phosphate potentials ([ATP]/[ADP][phosphate] and

[ATP]/[AMP][pyrophosphate]). This perturbation of intracellular adenine nucleotide homeostasis will be exacerbated with apoptotic activation of ATP- permeable Panx1 channels. Previous studies of intracellular ATP homeostasis during apoptosis indicated that ATP levels are maintained at normal levels, or even increase, in some cell types [92-94] but rapidly decrease in other cell models [95, 96]. It is interesting to speculate that these cell-specific differences in intracellular ATP levels during apoptotic progress reflect differences in Panx1 expression levels or rates of proteolytic activation. Normal or cancer cell types with high Panx1 expression, such as Jurkat cells, would be characterized by marked decreases in cytosolic ATP defined by the rate of accumulation of active caspase-3. Consistent with this scenario, each chemotherapeutic drug used in our experiments induced a kinetically distinct time course of decreasing intracellular ATP content that was rate-limited by the accumulation of active caspase-3 and markedly delayed by the Z-VAD pancaspase inhibitor. The decrease in total intracellular ATP was matched by changes in the ratio of intracellular ATP/ADP/AMP during apoptotic progression and correlated with

71 robust extracellular accumulation of the ATP metabolites, ADP and AMP, in addition of ATP per se. Indeed, our data indicate ADP and AMP increasingly constitute a major fraction of the adenine nucleotide pool released into the extracellular compartment of the Jurkat cells as they transition from the early to the later phases of apoptosis.

Although active Panx1 channels have been best characterized as conduits for the efflux of ATP (and UTP), they likely function as nonselective channels for the release of other small cytosolic metabolites with molecular masses (< 600–

700 Da) similar to that of ATP [45, 97]. In addition to direct release of intracellularly generated ADP and AMP, a fraction of the extracellular ADP and

AMP pool will accumulate secondary to the hydrolysis of the directly released

ATP pool by cell surface or serum ecto-nucleotidases. Indeed, we observed that the rapid accumulation of ATP in the extracellular medium of Jurkat cells treated with anti-Fas or STS transiently peaked at 4 h and then progressively declined. In contrast to many hematopoietic or nonhematopoietic tumor cell types, Jurkat cells are characterized by an atypically low expression of cell surface CD39 ecto-

ATPase/ADPase and no detectable expression of the glycosylphosphatidylinositol-anchored CD73 ecto-AMPase/ecto-5’-nucleotidase

[98]. Likewise, our experiments utilizing the ARL67156 ecto-ATPase/ADPase inhibitor suggested that only a minor fraction of the extracellular ADP/AMP measured in the conditioned medium of STS-treated Jurkat cells was produced by the extracellular hydrolysis of released ATP. Notably, Yamaguchi et al. [99]

recently reported that AMP is the major adenine nucleotide released via Panx1

channels from murine WR-19L lymphoma cells during Fas-induced apoptosis.

Alternative ATP Release Pathway Induced by Chemotherapeutic Drugs in

the Context of Suppressed Caspase Activity— We unexpectedly observed that

prolonged treatment of Jurkat T cells with staurosporine or etoposide, but not

anti-Fas, in the presence of Z-VAD revealed an alternative CBX-insensitive

mechanism of ATP release that was induced more slowly than the CBX-sensitive

Panx1 channel pathway. Recent studies in Jurkat cells and normal human T cells

have indicated that the T cell receptor (TCR) activation triggers a very rapid but

transient (within minutes) release of ATP via both Panx1 channel-mediated and

exocytotic pathways [37, 100, 101]. Both ATP release mechanisms were

dependent on the increase in cytosolic [Ca2+] elicited by TCR activation. Notably,

the alternative Panx1-independent ATP release response elicited by STS

treatment was completely suppressed in BAPTA-loaded Jurkat cells implicating

regulation by Ca2+. However in contrast to the TCR-stimulated ATP release, which was partially inhibited in bafilomycin-treated Jurkat cells, we observed no suppressive effect of this V-ATPase inhibitor of the STS+Z-VAD-induced ATP release. This argues against exocytotic mobilization of ATP-containing secretory lysosomes as the mechanism for this response to STS. Likewise, the ability of

STS + Z-VAD to trigger robust ATP release in the presence of 3-MA indicated no obvious role for autophagy-based mechanisms. However, experiments using

Jurkat cells with knockdown of Atg5 or Atg7 would provide a more rigorous test for involvement of autophagy signaling.

In addition to exocytosis, the Ca2+-dependent, but Panx1-independent,

ATP release response to STS + Z-VAD may involve another ATP-permeable

channel. For example, the CALHM1 gene product encodes the subunits for an

ATP-permeable, but CBX-insensitive, channel with structural features very

similar to those of Panx1 [102, 103]. Although CALHM1 is mainly expressed in taste-sensing cells [42] and the nervous system [104], five other human CALHM homologues have been reported [105]. Additionally, the so-called maxi-anion channels have been reported as a functionally characterized but molecularly undefined efflux pathway in multiple cell types [106-108], including Jurkat cells

[101]. Future studies should test the possible involvement of other ATP- permeable channels that might be activated in response to chemotherapeutic drugs. Finally, we observed that STS did not directly trigger an acute increase in cytosolic [Ca2+] and that STS + Z-VAD-induced ATP release was not attenuated by inhibition of PI-PLC signaling. These findings argue against feed-forward loops involving release of autocrine mediators, such as bioactive lipids or nucleotides per se, that can activate G protein-coupled receptors and 1,4,5- inositol trisphosphate-mediated Ca2+ mobilization. How chemotherapeutic drugs,

such as STS and Etop, may elicit slowly developing increases in cytosolic [Ca2+]

in the context of suppressed caspase signaling is another relevant question for

investigation.

Elevated Expression of Panx1 Channels in Leukemic Versus Normal

Leukocytes—Studies in Jurkat cells and primary mouse thymocytes (from wild

type and Panx1-/- mice) have provided most of the current understanding

regarding natively expressed Panx1 channels [49, 72]. Although apoptosis

increased ATP efflux via Panx1 channels both in Jurkat cells and thymocytes, the

magnitude of the peak extracellular ATP accumulation response was 5–8-fold greater in the Jurkat cells [10, 49]. Although this suggested that leukemic T cells may express higher levels of Panx1 per cell than normal T cells, it is difficult to interpret comparisons between cell types from different species. However, our studies, which directly compared Panx1 protein content in Jurkat and CEM human leukemic T lymphocytes with normal human T lymphoblasts, support this possibility. We have also found that murine EL4/EG7 thymic lymphoma cells express higher levels of Panx1 protein than normal mouse thymocytes (data not shown). It will be relevant to extend such comparisons to a broader range of normal leukocytes (e.g. monocytes and B cells) and their corresponding leukemia/lymphoma variants. Previous studies have suggested that increased

Panx1 expression in different hematopoietic tumor models can induce either pro- tumorigenic or anti-tumorigenic effects [109-113]. Besides caspase-mediated proteolytic activation, Panx1 channels can be rapidly and reversibly gated by hypoxia, altered tyrosine phosphorylation, and signals driven by several G protein-coupled receptors [45, 48, 114]. Thus, release of ATP, ATP metabolites, or arachidonic acid metabolites from Panx1-overexpressing tumor cells into different tumor microenvironments may entrain autocrine and paracrine signaling networks that differentially enhance or suppress tumor growth [115]. Subsequent chemotherapy-induced apoptosis of tumor cells would skew these local networks

by driving proteolytic gating of Panx1 channels and markedly enhance

extracellular accumulation of ATP and other cell-derived metabolites.

Significance of Panx1-mediated Efflux of ATP and ATP Metabolites from

Chemotherapeutic Drug-treated Tumor Cells—Extracellular ATP and ADP, but not AMP, function as agonists for several P2Y and P2X receptor subtypes that activate immune or inflammatory responses [116-118]. In contrast, extracellular

AMP is rapidly hydrolyzed by CD73 expressed on nearby stromal or recruited immune effector cells to generate adenosine that acts as an agonist for G protein-coupled adenosine receptor subtypes. Activation of A2a or A2b adenosine receptors on effector leukocytes typically (but not always) drives anti- inflammatory or immunosuppressive responses [119-121]. Thus, the relative concentrations of ATP versus ADP versus AMP that dynamically accumulate in the extracellular microenvironment of apoptotic tumor cells in vivo can variously skew local immune signaling networks between pro-inflammatory versus anti- inflammatory or immunogenic versus tolerogenic settings with significant consequences on the development of immunogenic anti-tumor responses [122-

126]. Moreover, most types of tumor cells per se express different cassettes of

P2 purinergic or P1 adenosine receptors that can modulate cell growth, survival, and sensitivity to pro-apoptotic mediators. Indeed, the leukemic Jurkat cells used in this study express several ionotropic P2 receptor subtypes, including P2X1,

P2X4, and P2X7, and these receptors act as autocrine modulators of TCR signaling pathways that regulate growth and cytokine production [100, 101]. It is interesting to speculate whether autocrine activation of these receptors during

chemotherapeutic drug-induced Panx1 activation may also modulate apoptotic

progression in Jurkat or other leukemic cells. Results from this study highlight the

ability of chemotherapeutic drugs to activate cancer cell Panx1 channels as efflux

conduits for both pro-inflammatory ATP and the anti-inflammatory precursor

AMP. Understanding how different chemotherapies initiate efficacious anti-tumor immune responses may be improved by further analyses of the tumor-specific

differences in Panx1 expression levels, drug-induced changes in intracellular

adenine nucleotide homeostasis, and the dynamics of extracellular adenine

nucleotide metabolism.

3.5 Acknowledgement:

Acknowledgments—We thank Dr. Alan Levine for providing normal human T

lymphoblasts, Dr. Aaron Weinberg for providing the CEM cell line, and Christina

Antonopoulos for performing the Western blot analyses of Panx1 expression in

MG132-treated Jurkat cells. This work was supported, in whole or in part, by

National Institutes of Health Grant R01-GM36387 (to G. R. D.). This work was

also supported by Canadian Institutes of Health Research MOP130530 (to S. P.

and D. W. L.). Supported in part by National Institutes of Health Training Grant

T32-GM008803.

Figure 3.1: Comparative time courses for accumulation of active caspase-3

and loss of viability in Jurkat leukemic T cells treated with different

chemotherapeutic agents.

(A) Jurkat T cells were treated with anti-Fas (250 ng/mL), STS (3 μM), Etop (20

μM), or Dox (25 μM). Cell lysates were prepared at the indicated times and then

assayed for caspase 3 activity. Experiments with each agent were repeated 2-5

times with data indicating mean ± SE for n=4 (anti-Fas), n=5 (STS and Dox); n=2

(Etop). (B and C) Jurkat T cells were treated with anti-Fas (250 ng/mL), STS (3

μM), Etop (20 μM), or Dox (25 μM) in the absence or presence of zVAD (50 μM for anti-Fas-stimulated cells or 100 μM for cells stimulated with STS, Etop, or

Dox). At the indicated times, aliquots were taken and immediately assayed for cell viability by measuring metabolism of AlamarBlue® dye to its fluorescent resorufin product. Data indicates mean ±SE for n=3 experiments.

Figure 3.1: Comparative time courses for accumulation of active caspase-3 and loss of viability in Jurkat leukemic T cells treated with different chemotherapeutic agents

Figure 3.2: Chemotherapeutic drugs induce caspase-3 mediated cleavage

of the pannexin-1 C-terminal autoinhibitory domain.

(A) Membrane topography of Panx1 protein subunit indicating the relative

positions of the caspase-3 cleavage site and epitope recognized by the anti-

Panx1 antibody used in the western blot of cell lysates from control Jurkat cells

or apoptotic Jurkat cells (20 μM Etop, 12 h). The indicated bands show the three

different glycosylation states of Panx1 (Gly 0, core, Gly 1, high mannose species,

Gyl 2, complex glycoprotein); * indicates non-specific immunoreactive bands. (B-

E)* Jurkat T cells were treated with 250 ng/mL anti-Fas (B), 3 μM STS (C), 20 μM

Etop (D), or 25 μM Dox (E) in the absence or presence of zVAD (50 μM for anti-

Fas or 100 μM for drugs). At the indicated times, aliquots were taken for western blot analysis of Panx1, PARP1, and actin. Data are representative of 2-3 experiments with each pro-apoptotic stimulus. (F). Densitometric quantification of

Panx1 bands in western blots from B-E; bands were normalized to the densities of the t=0 h samples for each experiment. *For uncropped Panx1 western blot see Appendix 1-4.

Figure 3.2: Chemotherapeutic drugs induce caspase-3 mediated cleavage of the pannexin-1 C-terminal autoinhibitory domain

Figure 3.3: Chemotherapeutic drugs induce accumulation of active pannexin-1 channels via a caspase dependent activation mechanism.

(A) Schematic of the YoPro dye uptake endpoint assay. Jurkat T cells were

incubated for with no stimulus for 4 h (B), with 250 ng/mL anti-Fas, 4 h (C), 3 μM

STS, 4 h (D), 20 μM Etop, 8 h (E), or 25 μM Dox (F) in the absence or presence

of 100 μM zVAD. The treated cells were then washed, resuspended in basal

saline supplemented with 1 μM YoPro ± 100 μM CBX, and incubated for 20 min

prior to plate reader quantification of accumulated YoPro fluorescence per well

(G) or phase contrast and epifluorescence imaging (B-F). (G) Data indicates mean ± SE for n=3 experiments. *P<0.05; **P<0.01; ***P<0.001. (H) Schematic of the YoPro dye uptake kinetic assay. Jurkat cells were suspended in basal saline supplemented 1 μM YoPro ± 100 μM CBX ± 100 μM zVAD transferred to the wells of a 24-well plate. Fluorescence (485 nm/540 nm) was measured at 1 min intervals for 15 min prior to addition of 3 μM STS (or vehicle) and then at 1 min intervals for an additional 4 h prior to the addition of digitonin (Dig) to permeabilize the cells. Data are representative of 3 experiments.

Figure 3.3: Chemotherapeutic drugs induce accumulation of active pannexin-1 channels via a caspase dependent activation mechanism

Figure 3.4: Efflux of both ATP and ATP metabolites is triggered during

chemotherapeutic drug-induced apoptosis of Jurkat cells.

Jurkat T cells were treated with 250 ng/mL anti-Fas, 3 μM STS, 20 μM Etop, or

25 μM Dox in the absence or presence of zVAD (50 μM for anti-Fas or 100 μM for drugs) as indicated. (A-B) Cell lysates were prepared at the indicated times

and processed for measurement of intracellular ATP content as described in

Experimental Procedures. ATP content in cells treated with pro-apoptotic agents was normalized to the ATP content in control cells. Experiments with each agent were repeated 2-5 times with data indicating mean ± SE for n=4 (anti-Fas), n=5

(STS and Dox); n=2 (Etop). (C) Jurkat cells were treated with anti-Fas or STS for

4 h or with Etop or Dox for 8 h. Heat-denatured cell extracts were prepared and assayed for intracellular ATP, ADP, and AMP content as described in

Experimental Procedures. Data (mean±SE) is from one experiment performed in triplicate. (D) Schematic of Panx1 activation by caspase-3 cleavage secondary to chemotherapeutic drug-induction of intrinsic apoptosis via release of mitochondrial cytochrome c. Caspase-gated Panx1 channels will mediate efflux of cytosolic ATP and ADP (and AMP, not shown), the levels of which will vary with progressive mitochondrial dysfunction. Following release, extracellular ATP and ADP can also hydrolyzed by ectonucleotidases. (E-F) Jurkat cells were suspended in serum-free RPMI1640 medium and treated with 3 μM STS for 4 h in the absence or presence of 100 μM ARL67156 ecto-nucleotidase inhibitor.

Samples of the extracellular medium were processed for analysis of ATP only (E)

84 or summed ATP+ADP+AMP (F). Data indicates mean ± SE for n=6 wells from 2 experiments.

Figure 3.4: Efflux of both ATP and ATP metabolites is triggered during chemotherapeutic drug-induced apoptosis of Jurkat cells.

Figure 3.5: Caspase-activated pannexin-1 channels mediate the efflux of

ATP and ATP metabolites during chemotherapeutic drug-induced

apoptosis but an alternative ATP release mechanism is engaged in the

context of suppressed caspase activity.

Jurkat T cells were treated with 250 ng/mL anti-Fas (panels A and E), 3 μM STS

(panels B and F), 20 μM Etop (panels C and G), or 25 μM Dox (panels D and H) in the absence or presence of zVAD (50 μM for anti-Fas or 100 μM for drugs) as indicated. Samples of the extracellular medium were processed for analysis of

ATP only (A, B, C, D) or summed ATP+ADP+AMP (E, F, G, H). Data indicates mean ± SE for n=4-6 wells from 2-3 experiments (n=4 for panels A, B, D, G, H;

n=6 for panels E, F).

Figure 3.5: Caspase-activated pannexin-1 channels mediate the efflux of

ATP and ATP metabolites during chemotherapeutic drug-induced apoptosis but an alternative ATP release mechanism is engaged in the context of suppressed caspase activity

Figure 3.6: Carbenoxolone blocks the efflux of ATP and ATP metabolites

during chemotherapeutic drug-induced apoptosis.

Jurkat T cells were incubated with no stimulus, with 250 ng/mL anti-Fas, 4 h (A,

E), 3 μM STS, 4 h (B, F), 20 μM Etop, 8 h (C, G), or 25 μM Dox, 12 h (D, H) in the absence or presence of 100 μM CBX (A, B, C, D) or 500 μM CBX (E, F, G,

H). Samples of the extracellular medium were processed for analysis of summed

ATP+ADP+AMP (A, B, C, D) or ATP only (E, F, G, H). Data indicates mean ± SE for n=2-3 experiments each performed in triplicate. *P<0.05; **P<0.01;

***P<0.001.

Figure 3.6: Carbenoxolone blocks the efflux of ATP and ATP metabolites during chemotherapeutic drug-induced apoptosis

Figure 3.7: Caspase-insensitive ATP release stimulated by

chemotherapeutic drugs is resistant to carbenoxolone blockade but

suppressed by intracellular Ca2+ buffering.

(A) Schematic of alternative ATP release pathways. When caspases are

inhibited, ATP may be secreted via activation of non-cleaved Panx1 channels or

exocytosis of secretory lysosomes or autophagolysosomes. (B, C) Jurkat T cells

were incubated with no stimulus or with 3 μM STS for 8 h in the absence or

presence of 100 μM zVAD, 100 μM CBX, or both inhibitors. Samples of the

extracellular medium were processed for analysis of ATP only (B) or summed

ATP+ADP+AMP (C). Data indicates mean ± SE for n=6 wells from 2

experiments; ***P<0.001. (D) Jurkat T cells were incubated for 1 h at 37°C in

either standard BSS (1.5 mM CaCl2) or Ca-free BSS (0 mM CaCl2) in the absence or presence 25 μM BAPTA-AM and then resuspended in fresh 1.5 Ca-

BSS or 0 Ca BSS. The BAPTA-loaded or mock-loaded cells were then incubated with no stimulus or with 3 μM STS f or 8 h in the absence or presence of 100 μM zVAD. Samples of the extracellular medium were processed for analysis of ATP.

Data indicates mean ± SE for 1 experiment performed in triplicate; ns: P>0.05;

***P<0.001. (E, F) Jurkat T cells were incubated for 1 h at 37°C in standard BSS

(1.5 mM CaCl2) in the absence or presence of 25 μM BAPTA-AM and then resuspended in fresh BSS. The BAPTA-loaded or mock-loaded cells were then incubated with no stimulus or with 3 μM STS for 8 h in the absence or presence of 100 μM zVAD. Samples of the extracellular medium were processed for analysis of ATP only (E) or summed ATP+ADP+AMP (F). Data indicates mean ±

SE for n=3 experiments; *P<0.05; **P<0.01; ***P<0.001. (G) Jurkat cell were

preincubated for 1 h in the absence or presence of 250 nM bafilomycin A (BafA),

100 μM zVAD, or both inhibitors. The cells were incubated for an additional 4 h

with or without 3 μM STS, and then assayed for LysoTracker Red accumulation.

Data indicates average ± range from 1 experiment performed in duplicate. (H)

Jurkat T cells were incubated with no stimulus or with 3 μM STS for 8 h in the absence or presence of 100 μM zVAD, 250 nM BafA, or both inhibitors. Samples

of the extracellular medium were processed for analysis of ATP. Data indicates

mean ± SE for n=3 experiments each performed in duplicate; ns: P>0.05;

*P<0.05. (I) Jurkat T cells were incubated with no stimulus or with 3 μM STS for 8

h in the absence or presence of 100 μM zVAD, 5 mM 3-methyladenine (3MA), or both inhibitors. Samples of the extracellular medium were processed for analysis of ATP. Data indicates mean ± SE for n=6 wells from 2 experiments; ns: P>0.05;

*P<0.05; ***P<0.001.

Figure 3.7: Caspase-insensitive ATP release stimulated by chemotherapeutic drugs is resistant to carbenoxolone blockade but suppressed by intracellular Ca2+ buffering.

Figure 3.8: Caspase-insensitive ATP release stimulated by staurosporine

does not involve direct Ca2+mobilization by staurosporine or activation of phosphatidyl inositol phospholipase C signaling.

(A, B, C, D) Jurkat cells loaded with fluo-4 Ca2+ indicator dye were assayed for

changes in cytosolic [Ca2+] as described in Experimental Procedures. Each trace

is representative of 2-5 similar test recordings from 3 separate batches of fluo-4

loaded cells. Where indicated in panel A and in all traces of panels B, C and D,

the fluo-4 fluorescence signals (RFU) were assayed in the presence of 2.5 mM probenecid. In panels A and B, the Ca2+ measurements were terminated by addition of Triton-X100 to release dye from all cells followed by addition of excess EGTA to indicate background Ca2+-independent fluorescence. Positive

control stimuli for increasing cytosolic [Ca2+] included 10 μM ionomycin (panels A

and B) and 1 μg/ml anti-CD3 TCR activating antibody (panels B, C and D). In

panel D, the indicated wells of Jurkat cell suspension were supplemented with 5,

10, or 20 μM U73122 PI-PLC inhibitor 15 min before stimulation with anti-CD3.

(E, F) Jurkat T cells were incubated with no stimulus or with 3 μM STS for 8 h in

the absence or presence of 100 μM zVAD, 10 μM U73122, or both inhibitors.

Samples of the extracellular medium were processed for analysis of ATP only

(panel E) or summed ATP+ADP+AMP (panel F). Data indicates mean ± SE for

n=6 wells from 2 experiments. Data indicates mean ± SE for n=6 wells from 2

experiments; ns: P>0.05; *P<0.05; ***P<0.001.

Figure 3.8: Caspase-insensitive ATP release stimulated by staurosporine

does not involve direct Ca2+mobilization by staurosporine or activation of phosphatidyl inositol phospholipase C signaling.

Figure 3.9: Proteosome inhibition induces caspase-3-mediated cleavage of the pannexin-1 C-terminal autoinhibitory domain and pannexin-1-mediated release of adenine nucleotides.

(A) Jurkat T cells were treated with no stimulus, 3 μM MG132, 3 μM STS, 20 μM

Etop, or 25 μM Dox in the absence or presence of 100 μM zVAD or combined

100 μM zVAD plus 3 μM MG132 for 12 h. Aliquots were taken for western blot analysis of Panx1, PARP1, and actin. Panx1 blots are from 2 separate experiments; PARP and actin blots are representative both of experiments; * indicates non-specific immunoreactive band. (B). Densitometric quantification of

Panx1 bands in western blots from A; bands were normalized to the densities of the control cell samples for each experiment. (C) Jurkat cells were incubated with no stimulus or 3 μM MG132 in the absence or presence of 100 μM zVAD. At the indicated times, samples of the extracellular medium were processed for analysis of summed ATP+ADP+AMP. Data indicates mean ± SE from 1 experiment performed in triplicate. (D) Jurkat T cells were incubated with no stimulus or with

3 μM MG132 for 8 h in the absence or presence of 100 μM zVAD or 100 μM

CBX. Samples of the extracellular medium were processed for analysis of summed ATP+ADP+AMP. Data indicates mean ± SE for n=3-9 wells from 3 experiments; ***P<0.001. (E, F) Jurkat T cells were incubated with no stimulus or with 3 μM MG132 for 8 h in the absence or presence of 100 μM zVAD. Samples of the extracellular medium were processed for analysis of YoPro accumulation

(in the absence or presence or 100 μM CBX) as described in Figure 3. (E)

Representative phase-contrast and epifluorescence microscopy images. (F) Data indicates mean ± SE for n=6 wells from 2 experiments; ***P<0.001.

Figure 3.9: Proteosome inhibition induces caspase-3-mediated cleavage of the pannexin-1 C-terminal autoinhibitory domain and pannexin-1-mediated release of adenine nucleotides.

Figure 3.10: Pannexin-1 is more highly expressed in human leukemic

leukocytes than in normal human T cells.

Left Panel: Whole cell lysates from Jurkat T cell lymphocytic leukemia, THP1

promonocytic leukemia, or CEM T cell lymphoblastic leukemia were processed

for western blot analysis of Panx1 and actin; * indicates non-specific

immunoreactive band. Right Panel: Normal human T lymphoblasts were

incubated for 4 h with or without 250 ng/mL anti-Fas or 3 μM STS the absence or presence of 100 μM zVAD. The cells were processed for western blot analysis of

Panx1 and actin; aliquots of cell lysates from an equivalent number (106) of

Jurkat cells were run on the same gel as positive controls.

Figure 3.10: Pannexin-1 is more highly expressed in human leukemic leukocytes than in normal human T cells.

100

Chapter 4: Upregulated ectonucleotidases in FADD- and RIP1-deficient

Jurkat leukemia cells counteract extracellular ATP/AMP accumulation via pannexin-1 channels during chemotherapeutic drug-induced apoptosis‡

4.1 Abstract

Pannexin-1 (Panx1) channels mediate the regulated efflux of ATP, ADP, and

AMP from cancer cells in response to induction of extrinsic apoptosis by death

receptors or intrinsic apoptosis by chemotherapeutic agents. We characterized

how different pathways of nucleotide efflux and ectometabolism determine the

composition of extracellular adenine nucleotide pools during chemotherapy-

induced apoptosis in Jurkat human leukemia cells that included mutated variants

lacking either the FADD or RIP1 adapter proteins. TNF-α induced extrinsic

apoptosis in control Jurkat cells but necroptosis in the FADD-deficient line;

treatment of either line with chemotherapeutic drugs elicited similar intrinsic

apoptosis. Jurkat cells lacking RIP1 were deficient in necroptotic signaling but

retained their proapoptotic response to chemotherapeutic agents. Although

apoptotic activation triggered robust proteolytic gating of Panx1 channels in all

three Jurkat cell lines, extracellular ATP/AMP accumulation was markedly

reduced in the FADD-deficient cells and completely suppressed in the RIP1-

deficient line. These differences were due to upregulated expression of

ectonucleotidases that included the CD73 ecto-AMPase in both mutant Jurkat

lines, as well as the CD39 ecto-ATPase and another yet-to-be-defined AMP-

degrading ectoenzyme in the RIP1-deficient cells. However, robust ATP/AMP

‡ A version of this chapter was submitted for review to the journal Molecular Pharmacology February, 2016

101

accumulation was observed in the FADD-deficient during necroptotic activation of

MLKL pores. Thus, extracellular adenine nucleotide accumulation during regulated tumor cell death requires ATP/AMP efflux pathways that exceed the rate of nucleotide clearance by cell-autonomous ecto-nucleotidases. Differential expression of various ecto-nucleotidases in tumor cell variants will determine whether chemotherapy-induced activation of the Panx1 channels drives accumulation of immunostimulatory ATP versus immunosuppressive adenosine within the tumor microenvironment.

4.2 Introduction

The efficacy of cancer chemotherapeutic regimens is markedly enhanced

by induction of a sustainable anti-tumor immune response. This is determined in

part by the accumulation of immunogenic mediators within the tumor micro-

environment. These include bioactive peptides, lipids, and nucleotides that are

either constitutively released from viable cancer cells or inducibly released via

chemotherapy- or receptor-induced cell death signaling cascades [5, 6, 127-131].

Given its high intracellular concentration in all cell types, ATP comprises a

ubiquitous tumor cell-derived immunogenic mediator. Extracellular ATP acts as

a “find-me” signal for G protein-coupled P2Y-family receptors expressed on nearby macrophages and dendritic cells (DCs) [8, 10, 49]. This amplifies chemotaxis of these phagocytic leukocytes to the vicinity of dying cells and thereby facilitates the phagocytosis of tumor cells required for processing and presentation of tumor antigens to T lymphocytes.

102

Pannexin-1 (Panx1) channels comprise a major pathway for the regulated

release of ATP to extracellular compartments in response to induction of either

extrinsic apoptosis by death receptors or intrinsic apoptosis by chemotherapeutic

agents [49, 132]. Apoptotic induction of active caspase-3 results in proteolytic

excision of an autoinhibitory C-terminal segment to drive open-gating of the large-pore Panx1 channels and consequent efflux of intracellular metabolites such as ATP [48, 49, 132]. Although the Panx1/ATP efflux pathway contributes to the efficient immunogenicity of several anti-tumor therapies, important questions regarding this model remain. These include: 1) how alternative mechanisms for extracellular ATP accumulation may be utilized during nonapoptotic modes of regulated tumor cell death; and 2) whether altered activity of such nucleotide accumulation pathways underlies the acquisition of resistance to anti-tumor therapies that can characterize clonal variants of particular tumor models.

Several therapeutic agents, including agonists for TNF/TRAIL-family death receptors and various chemotherapeutic drugs, can trigger regulated death of certain cancer cell types either by engagement of canonical apoptotic signaling or by necroptosis [17, 20]. The latter involves formation of necrosome complexes between RIP1 and RIP3 (receptor interacting protein kinases 1 and 3) to drive

RIP3-mediated phosphorylation of the mixed-lineage kinase like (MLKL) pseudokinase. Phosphorylated MLKL (pMLKL) oligomerizes and inserts into the plasma membrane as multimeric pores with significant permeability to both inorganic ions and organic metabolites such as ATP [17, 133-136]. However, the relative

103

kinetics and magnitudes of extracellular ATP accumulation as mediated by RIP3-

activated MLKL pores versus caspase-3-activated Panx1 channels in a given tumor cell model have not been quantified and compared.

Although ATP directly released from apoptotic or necroptotic tumor cells

supports immunogenic anti-tumor responses by stimulating P2Y receptor

signaling, it can also drive the generation of adenosine which activates

immunosuppressive A2A/A2B receptors expressed on both myeloid and

lymphoid immune cells [50, 51]. The balance between local ATP and adenosine

accumulation within the tumor microenvironment will determine the net efficacy of

the immunogenic response to programmed death of cancer cells. Multiple studies

have linked the decreased efficacy of cancer chemotherapies to increased levels

of ectonucleotidases that metabolize released adenine nucleotides to adenosine

in extracellular compartments [56, 126, 137]. CD39-family ectonucleotidases

break down ATP and ADP to AMP while the CD73 ectonucleotidase hydrolyzes

AMP to adenosine [52, 123, 138, 139]. While both ectonucleotidases can be

upregulated in certain types of cancers, CD73 in particular has been linked to

chemotherapy-resistant cancers and high CD73 levels are an indicator of poor

prognosis [123, 124, 126, 140-142].

In this study, we characterized how alternative pathways of nucleotide

efflux and ectometabolism determine the composition of extracellular adenine

nucleotide pools during death receptor-induced apoptosis or necroptosis versus

chemotherapeutic drug-induced apoptosis in Jurkat human leukemia cell models.

These included Jurkat variants lacking the FADD (Fas associated death domain)

104

adapter protein or RIP1 kinase. TNF-α receptor activation induces death via

extrinsic apoptosis in control Jurkat cells but necroptosis in the FADD-deficient line; treatment of either Jurkat line with chemotherapeutic drugs elicits similar intrinsic apoptotic signaling [143]. Jurkat cells that lack RIP1 are deficient in necroptotic signaling to MLKL but retain apoptotic signaling to Panx1. Although apoptotic activation triggered similarly robust proteolytic gating of Panx1

channels in all three Jurkat cell lines, extracellular adenine nucleotide

accumulation was markedly reduced in the FADD-deficient cells and completely suppressed in the RIP1-deficient line. The quantitative differences in extracellular nucleotide accumulation correlated with the relative expression of CD73 ecto-

AMPase activity in the three lines (RIP1-deficient>>FADD-deficient>control).

RIP1-deficient cells were characterized by additional upregulation of the CD39 ecto-ATPase and another as-yet-unidentified AMP-degrading ectoenzyme. In contrast to the marked inhibition of extracellular nucleotide accumulation in

FADD-deficient cells with caspase-3-gated Panx1 channels, robust nucleotide accumulation was observed in the FADD-deficient Jurkat cells during necroptotic activation of MLKL pores. These results indicate that net extracellular adenine nucleotide accumulation during various modes of tumor cell death signaling requires a nucleotide efflux pathway that exceeds the rate of clearance by cell- autonomous ecto-nucleotidases. Differential expression of CD73 ecto-AMPase and other ecto-nucleotidases in clonal tumor variants may determine whether chemotherapeutic drug-induced activation of the Panx1 channel pathway

105 predominantly drives accumulation of immunostimulatory ATP versus immunosuppressive adenosine.

4.3 Results

TNF α-induction of necroptosis or extrinsic apoptosis induces release of adenine nucleotides from Jurkat cancer cells via mechanistically distinct pathways- Apoptosis and necroptosis comprise two distinct forms of programmed cell death (Fig. 4.1A). In wildtype (WT) Jurkat human leukemic T cells, TNF-α (T) binding to type 1 TNF receptors (TNFR1) induces formation of “complex 1” signaling platforms containing TRADD, RIP1, and cIAP1/2. This elicits NF-κB signaling and cell survival. However, when mitochondria release SMAC/Diablo or the cells are treated with SMAC mimetic drugs (S), the cIAPs are downregulated to facilitate the assembly of “complex 2” platforms consisting of FADD, procaspase-8 (casp8), and RIP1. Under these conditions, casp8 is activated to both initiate apoptosis and cleave and inactivate RIP1. Conversely, when Jurkat cells deficient in FADD (FADD-def) are treated with TNF-α and SMAC mimetic, complex 2 cannot assemble leaving RIP1 free to form necrosome platforms with

RIP3; this drives MLKL phosphorylation, oligomerization, and pore insertion into the plasma membrane and thereby triggers necroptosis. Fas receptor-induced apoptosis of WT Jurkat cells has been previously shown to elicit ATP efflux via caspase-3 (casp3) mediated cleavage of the C-terminus of Panx1 channels (Fig.

4.1A). While the kinetics and magnitude of this Fas-induced ATP release from

Jurkat cells has been defined, it is not known whether similar Panx1 cleavage and ATP efflux parameters characterize TNFR1-induced apoptosis or how MLKL

106

pores may facilitate alternative ATP efflux parameters during TNFR1-induced

necroptosis.

Using WT and FADD-def Jurkat cells identically treated with TS, we compared the kinetics of lytic plasma membrane disruption (as indicated by release of the cytosolic macromolecule lactate dehydrogenase, LDH) during apoptotic versus necroptotic progression. While WT cells treated with TS for up to 12 h did not release LDH (Fig. 4.1B), FADD-def cells began to release LDH

within 4 h and this progressively increased to ~60% cell lysis by 12 h (Fig. 4.1C).

The TS-induced lysis of FADD-def Jurkat cells was completely suppressed by necrosulfonamide (NSA), an inhibitor of the phosphorylated MLKL oligomerization necessary for insertion of MLKL pores into the plasma membrane. In contrast, there was no attenuation of LDH release from FADD-def cells treated with TS in the presence of the pan-caspase inhibitor zVAD (Fig.

4.1C). These observations confirm that TNF-α induced necroptotic signaling is defined by rapid plasma membrane disruption independently of apoptotic caspases and that TNF-α induced apoptotic signaling over a similar 12 h test period does not result in secondary necrosis.

Despite these marked differences in lytic plasma membrane disruption,

TS induced similar extracellular adenine nucleotide accumulation (1-1.5 μM at 8-

12 h) by the WT and FADD-def Jurkat cells cultured at 2x106/ml. We

characterized the kinetics and underlying apoptotic mechanism of TS-induced

adenine nucleotide release from the WT cells by using Fas receptor activation as

a positive control and measuring the summed accumulation of extracellular ATP,

107

ADP and AMP (ANex). Our previous study, described in chapter 3, determined

that analysis of ANex, rather than ATP alone, provides a better readout of net

Panx1 channel-mediated adenine nucleotide efflux during apoptotic progression

because the intracellular nucleotide pool shifts from initially high [ATP] and

low[ADP]+[AMP] to lower [ATP] and higher [ADP]+[AMP] (Fig 3.4) as a

consequence of the increased ATP utilization and decreased mitochondrial ATP

synthesis that characterizes apoptotic signaling [132]. TS-induced apoptosis resulted in a steady and significant increase in ANex after a ~ 4 h lag time and

this was suppressed in the presence of zVAD but not NSA (Fig 4.1E). These

ANex responses were similar to those observed in αFas-treated WT Jurkat cells

but with modestly slower kinetics (Fig 4.1D). In both TS- and αFas-treated WT

Jurkat cells, the time course of the ANex accumulation correlated with the

activation of caspase-3 as indicated by cleavage of the canonical casp3

substrate PARP (Figs. 4.1G and H). In addition, we compared the proteolytic

processing of Panx1 channels during the TS- and αFas-treatments. As indicated in Fig 3.2A, Panx1 exists in three different glycosylation states: un-glycosylated

Gly-0, monoglycosylated Gly-1, and Gly-2 which is the fully glycosylated and most abundant form that accumulates in the plasma membrane as a 48-50 kDa protein. All three forms are detected using a previously characterized polyclonal antibody that recognizes the extreme intracellular C-terminus of Panx1 downstream of the casp3 cleavage site. Casp3-mediated cleavage of Panx1 during apoptotic progression results in a decreased anti-Panx1 48-50 kDa band on western blots of whole cell lystates due to loss of the antibody binding site.

108

Panx1 cleavage was evident within 4 h in αFas-treated WT cells and within 8 h in the TS-treated cells (Fig 4.1G and H). These data, together with the observed lack of LDH release, indicate that the ANex accumulation triggered during TNF-

receptor mediated apoptosis is mediated by efflux through caspase-3-activated

Panx1 channels prior to secondary necrosis. This confirms that engagement of

multiple death receptors in Jurkat leukemia activates similar extrinsic apoptotic

signaling programs leading to proteolytic gating of Panx1 channels that mediate

efflux of cytosolic ATP, ADP and AMP pools.

In contrast to the WT cells, FADD-def Jurkat cells responded to TS with robust ANex accumulation that started at 2 h and peaked at 8 h. This response was completely suppressed by NSA but insensitive to zVAD (Fig 4.1F) indicating

that ANex accumulation was dependent on insertion of non-selective MLKL pores

into the plasma membrane. Also consistent with this mechanism, no obvious

cleavage of PARP or Panx1 was evident in TS-stimulated FADD-def Jurkat cells at the 4- and 8 h-time points coinciding with significant nucleotide release (Fig.

4.1I). It is important to note that the decrease in the Panx1 (and uncleaved

PARP1) western blot signals occurring after 8h is due to loss of viable cell mass as indicated by the decreased actin signal. Taken together, these data from

FADD-def Jurkat cells indicate that although TNFR1 signaling is redirected from extrinsic apoptosis to necroptosis, this regulated cell death results in extracellular accumulation of ATP, ADP and AMP released via MLKL pores concurrent with cell lysis.

109

Intrinsic apoptotic signaling and Panx1 channel cleavage in FADD-

deficient Jurkat cancer cells is uncoupled from accumulation of extracellular

adenine nucleotides- Many tumor cells lose the ability to undergo extrinsic

apoptosis in response to the various death receptors such as Fas, TNFR1, and

TRAIL, due to up-regulation of pro-survival proteins or down- regulation of pro-

apoptotic proteins [144]. Additionally, down-regulation of some signaling proteins

classically associated with death receptor-induced extrinsic apoptosis can be associated with resistance to chemotherapeutic drug-induced intrinsic apoptosis

in certain cancer cell models [56]. Although FADD is best characterized for its

critical roles in mediating extrinsic apoptosis by Fas and TNFR1, the assembly of

RIP1/FADD/caspase-8 ripoptosomes can amplify death signaling by

chemotherapeutic drugs that induce intrinsic apoptosis due to release of

mitochondrial SMAC/Diablo and down-regulation of IAPs [145]. We tested whether the absence of FADD modulated various intrinsic apoptotic signaling responses, including Panx1-mediated release of adenine nucleotides to extracellular compartments, in Jurkat cells stimulated with chemotherapeutic drugs. As described in chapter 3, multiple chemotherapeutic or pro-apoptotic drugs, including staurosporine (STS) and etoposide (Etop), trigger caspase-3- mediated activation of Panx1 channels and consequent adenine nucleotide efflux in WT Jurkat T cells [132]. Using these responses in WT cells as positive controls, we similarly treated FADD-def Jurkat cells with STS or Etop for up to12 h and assayed Panx1 and PARP cleavage (Figs. 4.2A, B), secondary necrosis

(Figs. 4.2C, D), and ANex accumulation (Figs. 4.2E, F). Both STS and Etop

110

induced robust intrinsic apoptosis in the FADD-def cells as indicated by cleavage

of PARP and Panx1 albeit with slightly delayed kinetics compared to the

responses in WT cells. We verified that the FADD-def Jurkat cells were unable

to undergo extrinsic apoptosis as indicated by the absence of PARP and Panx1

cleavage in response to αFas (Fig 4.2B). STS induced near-maximal apoptotic

activation within 4-6 h in both Jurkat lines while significant responses to Etop

required a 6-8h exposure. Secondary necrosis was observed when either cell

line was treated with STS, but not Etop, for >8h (Fig. 4.2C,D); this lytic release of

LDH was completely suppressed by zVAD.

Consistent with our previous findings, STS and Etop induced robust ANex

accumulation (Fig. 4.2E) in WT Jurkat cells that was: 1) temporally correlated

with the kinetics of Panx1 proteolytic processing (Fig. 4.2A); 2) insensitive to

NSA; and 3) markedly suppressed by zVAD. We previously demonstrated that

the zVAD-insensitive component of nucleotide release in STS-treated Jurkat cells

is insensitive to blockade of Panx1 channels but suppressed by intracellular Ca2+

buffering [132]. To our surprise, very little extracellular ANex accumulation was

observed in the FADD-def Jurkat cells treated with STS or Etop (Fig 4.2G and H)

despite the robust proteolytic processing of Panx1 channels. These data indicate

that FADD deficiency uncouples accumulation of ANex from casp3-mediated cleavage of Panx1 channels.

Caspase-3-cleaved Panx1 channels are functionally active in FADD-

deficient Jurkat cancer cells during intrinsic apoptosis- The decreased ANex

111

accumulation by apoptotic FADD-def Jurkat cells could be due to four possible mechanisms: 1) lack of functional Panx1 channels; 2) altered ion selectivity or rectification properties of active Panx 1 channels; 3) altered localization of ATP-

producing mitochondria within the local membrane environment of active Panx1

channels; or 4) increased ecto-nucleotidase activity. As illustrated in Fig. 4.3A,

caspase-3-mediated excision of the C-terminal autoinhibitory domains of Panx1

channels gates their permeability as ATP efflux conduits, but also as influx/efflux

pathways for normally impermeant organic dyes. To test whether caspase-3- processed Panx1 channels are functionally blocked in apoptotic FADD-def Jurkat cells, we adapted our previously described assay (Fig 3.3) for the influx of YO-

PRO2+, a 375-Da divalent cationic dye that intercalates with DNA to produce

green fluorescence. WT or FADD-def cells (control and apoptotic) were pulsed with extracellular YO-PRO2+ at times corresponding to maximal Panx1 cleavage by STS (4 h) or Etop (12 h), and dye uptake was quantified by fluorescence plate reader analysis (Fig. 4.3B) and visualized by fluorescence microscopy (Fig.

4.3D). Significant increases in YO-PRO2+ dye uptake were observed in both WT

and FADD-def cells treated with STS or Etop relative to untreated controls for

each cell line (Figs. 4.3B and D). Although the magnitude of YO-PRO2+

accumulation was ~2-fold lower in the apoptotic FADD-def cells relative to

apoptotic WT cells, these data indicate that intrinsic apoptosis does induce

accumulation of functionally active Panx1 channels in the FADD-def Jurkat cells.

However, this assay of Panx1 function measures the influx of an organic cation

while ATP is an organic anion that effluxes from the cell. To test whether FADD

112

deficiency might alter the ion selectivity or rectification properties of Panx1

channels, we loaded WT or FADD-def Jurkat cells with the anionic calcein4- dye

prior to the 4 h treatment with STS. Similar amounts of calcein4- were released

from the apoptotic WT and FADD-def cells (Fig 4.3C). We confirmed that this

calcein4- efflux was mediated by casp3-activated Panx1 channels by performing

parallel measurements in the presence of zVAD or the Panx1 channel blocker

trovafloxacin (Trov). Both inhibitors caused complete retention of calcein2- within the WT and FADD-def cells during STS treatment. These results indicate that changes in Panx1 channel ion selectivity, conductance, or rectification are unlikely mechanisms for the marked suppression of ANex accumulation in FADD-

def Jurkat cells.

FADD deficiency does not alter mitochondrial localization in control or

apoptotic Jurkat cancer cells- Recent studies have reported that, during T cell

activation, mitochondria translocate to the plasma membrane subdomains of the

immune synapse resulting in highly localized increases in ATP concentration

near to Panx1 channels and T cell receptor (TCR) signaling complexes [146-

148]. Junger and colleagues demonstrated that these local pools of ATP are

necessary for optimal ATP release via Panx1 channels in T cells and that

inhibition of mitochondrial translocation results in decreased ATP efflux. Given

the T cell lineage of the Jurkat leukemia cells, we considered the possibility that

mitochondria might also localize at the plasma membrane during apoptosis to

support Panx1-mediated ATP efflux and that FADD deficiency might disrupt this

113

localization to decrease release of adenine nucleotides. We used MitoTracker®

Red to visualize mitochondrial localization in WT and FADD-def Jurkat cells before and after STS treatment for 4 hours (Fig. 4.3E). DAPI staining verified that the nucleus occupies much of the cytoplasmic volume in both cell lines prior to apoptotic induction and that mitochondria are localized to the thin shell of cytosol between the nucleus and plasma membrane. STS induced equivalent nuclear fragmentation in the WT and FADD-def cells to result in similar diffusion of mitochondria into the expanded cytosolic volume. While these data neither rule out nor demonstrate a role for subcellular mitochondrial localization in supporting

Panx1-mediated ATP efflux from apoptotic Jurkat cells, there is no indication that

FADD deficiency significantly alters mitochondrial distribution.

CD73 ecto-nucleotidase activity is upregulated in FADD-deficient Jurkat

cancer cells and counteracts Panx1 channel-mediated efflux of ATP/AMP during intrinsic apopotosis - The major cell surface nucleotidases include: 1) the CD39-

family enzymes which serially hydrolyze ATP to ADP and ADP to AMP [140]; and

2) the GPI-anchored CD73 enzyme which breaks down AMP to adenosine (Fig.

4.4C). Previous studies have shown that WT Jurkat T cells express very low

CD39 (see Fig 3.4) and CD73 activity. However, expression of both

ectoenzymes is highly regulated (both positively and negatively) during the

differentiation, polarization, and activation of normal T lymphocytes [149].

Moreover, multidrug resistant clonal variants of Jurkat cells selected for

suppressed Fas expression have increased CD73 expression [56]. Thus, we

114

tested whether the decreased ANex accumulation by apoptotic FADD-def Jurkat cells was due to increased extracellular breakdown of the adenine nucleotides released via active Panx 1 channels. As noted previously, we routinely measure the summed accumulation of extracellular ATP, ADP, and AMP by apoptotic cells because the progressive decline in the intracellular ATP/AMP ratio results in greatly increased efflux of AMP, rather than ATP, via activated Panx1 channels as cells transition from early to later phases of the apoptotic cascade. Indeed, comparison of the magnitudes of the summed extracellular [ATP+ADP+AMP]

(Fig 4.4A; replotted data from Figs. 4.2E and F) versus [ATP] alone (Fig. 4.4B)

during STS treatment of WT Jurkat cells demonstrates ATP constitutes only a

few percent of total adenine nucleotide release. However, similar accumulation

of extracellular ATP was observed during the first 4 h of STS in both WT and

FADD-def cells which contrasted with the ~10-fold difference in ANex at these

time points. These data suggest that Panx1-mediated ATP and AMP efflux are

similar in both cell lines but that the AMP is rapidly cleared from the extracellular

compartment. Similar comparisons of ANex (Fig. 4.4A, middle panel) versus ATP

(Fig. 4.4B, middle panel) accumulation in WT and FADD-def cells during the

slower apoptotic progression induced by Etop revealed a different pattern of

extracellular nucleotide dynamics. Etop induced similar time courses of ATP and

ANex accumulation in the WT cells (near-linear increases with time after an initial

4 h lag period), but almost no accumulation above basal levels of either ATP or

ANex in the FADD-def cells. This suggests both ATP and AMP can be efficiently

cleared from the extracellular compartment of the FADD-def cells due to the

115

slower rates of Panx1-mediated ATP and AMP efflux triggered by Etop- versus

STS-induced apoptotic execution. Finally, comparisons of ANex (Fig. 4.4A, right panel) versus ATP (Fig. 4.4B, right panel) accumulation during TNFα (+Smac mimetic)-induced apoptosis in WT cells and TNFα (+Smac mimetic)-induced necroptosis in FADD-def cells yielded yet another extracellular nucleotide profile.

The kinetics and magnitudes of ATP and ANex accumulation during TS-induced

extrinsic apoptosis in WT cells were very similar (4 h lag phase followed by

steady increase over the next 8 h) to those measured during Etop-induced

intrinsic apoptosis. In contrast, TS-induced necroptosis in the FADD-def cells

rapidly triggered a 15-fold in ATP accumulation that peaked at 4 h and then

declined, while the ANex accumulation continued to increase until 8h. This robust accumulation of both ATP and ANex during TS-induced necroptosis of the

FADD-def cells contrasted with the near-complete absence of ATP/ANex accumulation during Etop-induced apoptosis in same FADD-def background.

Thus, markedly different profiles of extracellular nucleotide accumulation can occur in a given cancer cell lineage depending on the modes of regulated cell death.

Given that AMP is the predominant nucleotide released via Panx1 channels in apoptotic Jurkat cells, we hypothesized that FADD-deficiency resulted in an increased CD73 activity that efficiently hydrolyzes the released

AMP to adenosine. To test this, we compared the magnitudes of ANex

accumulation by WT and FADD-def Jurkat cells during 4 h STS treatment in the

absence or presence of the CD73 inhibitor α,β-methylene-adenosine 5’-

116

diphosphate (APCP). The presence of 50μM APCP markedly increased (by 6-

fold) ANex accumulation in the STS-stimulated FADD-def Jurkat cells; this ANex

level was similar to that in STS-treated WT cells in the absence of APCP (Fig.

4.4D). Similar potentiation of ANex accumulation by APCP was observed in

FADD-def Jurkat cells treated with Etop for 12 h (data not shown). The

increased ANex accumulation observed with APCP was not observed when the

FADD-def cells were cotreated with STS and zVAD consistent with AMP being

released via casp3-gated Panx1 channels. Notably, the presence of APCP

further increased the ANex accumulated during TS-induced necroptotic activation

of FADD-def Jurkat cells (Fig. 4.4E).

To confirm that the APCP-dependent increase in ANex accumulation was due to inhibition of AMP breakdown, we compared the abilities of FADD-def versus WT Jurkat cells to hydrolyze a 1 μM pulse of exogenous AMP over a 30 min test period in the absence or presence of APCP (Fig 4.4F). The WT cells hydrolyzed <20% of the added AMP while FADD-def cells cleared 50% of the

AMP; this accelerated breakdown was suppressed by APCP.

Apoptotic signaling and Panx1 channel activation in RIP1-deficient Jurkat cancer cells is also uncoupled from accumulation of extracellular adenine nucleotides- RIP1 plays a central role in directing TNF/TRAIL-family receptor signaling along the cell survival, extrinsic apoptotic, or necroptotic pathways [17]

(Fig. 4.1A). However, as previously noted, RIP1/FADD/caspase-8 ripoptosome

complexes can also assemble during intrinsic apoptosis triggered by

117

chemotherapeutic drugs [145]. Given the striking phenotypic effects of FADD-

deletion on modulating ANex accumulation driven by pro-apoptotic

chemotherapeutic drugs, we tested whether deficiency in RIP1, as another

ripoptosome component, results in a similar phenotype. We induced extrinsic or

intrinsic apoptosis in RIP1-def Jurkat cells (and WT cells as matched controls)

with anti-Fas for 4 h, STS for 4 h or Etop for 12 h, respectively. RIP1-def Jurkat cells were characterized by the complete absence of detectable ANex

accumulation in response to apoptotic induction by STS (Fig. 4.5A), Etop (Fig.

4.5B), and anti-Fas (Fig 4.5C). Western blot analysis indicated identical casp3- mediated processing of both Panx1 and PARP in the WT and RIP1-deficient cells

treated with STS or Etop. Similar to the FADD-def cells (Fig. 4.3E), there was no obvious alteration in mitochondrial localization in RIP1-def Jurkat cells before or after apoptotic induction with STS (Fig. 4.5E). Likewise, analysis of YO-PRO2+

dye influx (Fig. 4.5F and 4.5G) or calcein4- efflux (Fig. 4.5H) revealed no differences in Panx1 channel functional activity between the WT and RIP1-def cells during apoptotic progression.

Increased expression of CD73 ecto-AMPase in RIP1-deficient Jurkat cells relative to FADD-deficient or wildtype Jurkat cells- Given the similar phenotypes of the RIP1-def and FADD-def Jurkat cells, we tested whether the APCP inhibitor of CD73 would also rescue the ability of STS to induce ANex accumulation in the

RIP1-def cells. To our surprise, inclusion of 50 μM APCP during the STS treatment did not facilitate increased ANex accumulation in these Jurkat cells in

118

contrast to its effects in the FADD-def line (Fig. 4.6A). We then used the same

ecto-AMPase assay described for Fig.4F to compare the abilities of the WT,

FADD-def, and RIP1-def cells – at the same cytocrit of 2 x 106 cells/ ml used in

the ANex measurements – to metabolize an exogenous pulse of 1μM AMP in the

absence or presence of 50 μM APCP (Fig. 4.6B). The RIP1-def cells completely

hydrolyzed this exogenous AMP within 30 min even in the presence of APCP;

this contrasted with the respective 20% and 50% AMP clearance activities

(inhibitable by APCP) of the WT and FADD-def lines. We considered whether upregulated ecto-alkaline phosphatase (eAP), another ecto-AMPase, might

contribute to the greater AMP clearance by the RIP1-def cells. However,

inclusion of the eAP inhibitor tetramazole (Tetra) did not prevent the AMP

clearance (Fig. 4.6B). We then assayed ecto-AMPase activity at progressively

lower cytocrits of the FADD-def Jurkat cells (Fig. 4.6C). At 2 x 104/ml, the RIP1-

def cells cleared ~40-50% of the 1 μM exogenous AMP and this hydrolysis was

suppressed by APCP but not tetramazole.

Although these data indicate that a CD73-type ectonucleotidase activity is

upregulated in RIP1-deficient Jurkat cells, the inability of APCP to facilitate

significant rescue of apoptotic ANex accumulation suggests that these cells

express an additional pathway for extracellular AMP clearance. Studies in

muscle tissues have identified an extracellular AMP deaminase activity within

neuromuscular junctions. AMP deaminase (AMPD), which catalyzes conversion

of AMP to IMP (and is distinct from adenosine deaminase), is best characterized

as an intracellular purine salvage enzyme [150]. To distinguish CD73-mediated

119

hydrolysis of AMP to adenosine from AMPD-mediated conversion of AMP to

IMP, we incubated intact WT, FADD-def, and RIP1-def Jurkat cells with 10 μM ε-

AMP and measured both its decrease and corresponding conversion to the ε- adenosine (ε-ADO) product by HPLC and fluorescence detection. Etheno- modification of the adenine moiety on AMP or adenosine prevents their recognition by the deaminases. Figs. 4.6D and E verify that functional CD73 ecto-AMPase activity in RIP1-def Jurkat cells is >8-fold higher than in the WT line and 4-fold higher than in the FADD-def cells.

Increased expression of CD39 ecto-ATPase in RIP1-deficient Jurkat cells relative to FADD-deficient or wildtype Jurkat cells- As noted above, the initial phases of STS-stimulated extracellular ATP accumulation (Fig. 4.4B) were similar in the WT and FADD-def Jurkat cells despite marked differences in total

ANex accumulation (Fig.4.4A). This suggested that the FADD-def cells do not

express higher CD39 ecto-ATPase activity than WT Jurkat cells. We and others

have previously reported that functional CD39 expression is very low in WT

Jurkat cells (Fig 3.3) [10, 132]. Given the higher CD73 ecto-AMPase levels in

RIP1-def cells relative to the WT and FADD-def lines, we tested if CD39 ecto-

ATPase expression was also greater in the RIP1-def cells by incubating each of

the three Jurkat cell lines with 1 μM exogenous ATP for 30 min in the absence or

presence of the CD39 inhibitor ARL67156 (ARL). Consistent with previous

reports, the WT and FADD-def Jurkat cells hydrolyzed only 5-6% of the added

ATP and this breakdown was blocked in the presence of ARL (Fig. 4.7B). In

120

contrast, the RIP1-def cells cleared ~40% of the ATP via an ARL-inhibited

pathway. The increased capacity of the RIP1-def Jurkat cells to metabolize

exogenously added ATP correlated with a markedly decreased ability to

accumulate extracellular ATP during the early phase of STS-induced apoptosis

(Fig. 4.7A).

4.4 Discussion

This study provides four new insights regarding mechanisms of adenine

nucleotide release and extracellular metabolism by cancer cells (Jurkat human

leukemic T cells) during different modes of regulated cell death. First, we

extended previous findings by others [49] and us [151] verifying the critical role of

casp3-activated Panx1 channels as an ATP efflux pathway to cells undergoing

extrinsic apoptosis in response to the TNFR1 member of the TRAIL family.

Second, we demonstrated that redirection of TNFR1 signaling from apoptosis to

necroptosis results in redirection of the ATP efflux mechanism from open-gating

of Panx1 channels to insertion of RIP3-phosphorylated MLKL pores into the

plasma membrane. Third, we observed that, due to its rapid intracellular

production during both apoptotic and necroptotic progression, AMP becomes a

major nucleotide substrate for both Panx1 channels and MLKL pores resulting in

its robust extracellular accumulation during both modes of regulated cell death.

Finally, we found that genetic ablation of FADD and RIP1, two critical adapter proteins for TNF/TRAIL death receptor signaling, uncouples casp3-activated

Panx 1 channels from extracellular adenine nucleotide accumulation due to up-

regulated expression of ectonucleotidases. These included the CD73 ecto-

121

AMPase in both the FADD-deficient and the RIP1-deficient Jurkat cells, as well

as the CD39 ecto-ATPase and another as-yet-unidentified AMP-degrading enzyme in the RIP1-deficient cells. This provides insight into the regulation of

ATP, AMP and adenosine levels during mutation of cancer cells.

TNFR1 activation induces adenine nucleotide release via pannexin-1 channels in apoptotic Jurkat cells versus MLKL pores in necroptotic Jurkat cells

Depending on cellular and metabolic context, TNFR1 can signal through three mechanistically distinct pathways in leukemic Jurkat T cells and other cancer cell types. Activation of TNFR1 with TNF-α alone induces NFκB signaling and cell survival. However, in the presence of SMAC mimetic drugs (such as

BV6 used in this study) that downregulate cIAP proteins and thereby decrease

K67-ubiquitination of RIP1, TNF-α drives assembly of pro-apoptotic

RIP1/FADD/caspase-8 ripoptosomes in WT Jurkat cells, but pro-necroptotic

RIP1/RIP3 necrosomes in FADD-deficient Jurkat cells. These model systems facilitated direct comparison of the adenine nucleotide efflux responses during

TNFR1-induced apoptosis versus TNFR1-induced necroptosis in the same tumor cell background. As expected, TNFR1 activation of BV6-treated WT Jurkat cells induced a progressive increase in extracellular ATP+ADP+AMP (ANex)

accumulation over 12 h (after an initial 2-h lag phase). This accumulation was

mediated by active Panx1 channels in the absence of overt cell lysis because it:

1) coincided with casp3-mediated proteolytic processing of both Panx1 channels

122

and PARP; 2) was suppressed by the zVAD pan-caspase inhibitor; and 3) preceded secondary necrotic lysis as indicated by accumulation of extracellular

LDH. These responses to TNFa largely recapitulated those elicited by Fas activation, but with slower kinetics. This suggests that all members of the

Fas/TNFR/TRAIL death receptor family drive extrinsic apoptototic cascades that similarly converge on Panx1 channels as the predominant pathway for nucleotide release.

Notably, nucleotide release in response to the same TNFa/BV6 stimuli was completely redirected from a Panx1 channel-based mechanism to a MLKL pore-based pathway in Jurkat cells that lacked expression of FADD. Some characteristics of this ANex accumulation in necroptotic FADD-def cells were

similar to those in apoptotic WT cells including: 1) a 2 h lag phase before

significant nucleotide release; and 2) a quantitatively equivalent magnitude (1-2

mM) of peak [ANex]. Despite these similarities, the underlying MLKL-dependent

efflux pathway and its regulation were very distinct and defined by: 1)

concurrence or near-concurrence with lytic collapse of membrane integrity; 2)

insensitivity to caspase inhibition; 3) suppression by the NSA inhibitor of pMLKL

oligomerization and insertion into the plasma membrane pore. Two recent

studies have also described the release of ATP during necroptotic progression in

human THP-1 monocytic leukemia cells [152] and a human bronchial epithelial

cell line [153]. These reports additionally noted that conditioned media from the

necroptotic cells stimulated the P2Y receptor-mediated migration of phagocytic

leukocytes in transwell mobility assays similar to the earlier findings of

123

Ravichandran and colleagues who used conditioned medium from apoptotic

Jurkat cells and murine thymocytes [10, 49]. Thus, our studies support a model whereby the regulated release of ATP and other nucleotides acts as a major

“find-me” signal for phagocytic leukocytes to similarly migrate towards either apoptotic or necroptotic tumor cells as a critical prelude to tumor cell clearance and processing of tumor antigens for presentation to T cells.

Differential release of ATP versus AMP during the initial versus later phases of both apoptotic and necroptotic progresssion

Most studies of nucleotide release from dying cancer cells have focused on extracellular accumulation of ATP given its role as both a find-me ligand for directed migration of phagocytes to the tumor microenvironment and a pro- immunogenic/ pro-inflammatory mediator [5, 7, 9, 10, 49, 65, 74]. However, it is important to consider that both Panx1 – as a so-called large pore channel [154]– and MLKL [155-157] – as a bona fide pore-forming protein – are defined by permeability to a broad range of organic/inorganic anions and cations with molecular masses in the 1000 Da range or greater. These can include nucleotides other than ATP, such as UTP, which also acts as find-me ligand for

P2Y2 receptor-expressing phagocytes [49, 158] and intracellular metabolites of

ATP, such as AMP [99, 151]. The results from this study (Figs. 4.4A and B) confirm and extend those latter findings that AMP becomes the predominant nucleotide released via Panx1 channels during the progression of apoptotic

124 execution. Similarly, AMP, in addition to ATP, was increasingly released via

MLKL pores during progression of necroptosis in TNFa/BV6-stimulated FADD- deficient Jurkat cells (Figs. 4.4A and B). Previous studies by Green and colleagues determined that the intracellular nucleotide pool of apoptotic cells shifts from an initially high [ATP] to progressively lower [ATP] as a consequence of the increased ATP utilization and decreased mitochondrial ATP synthesis that characterizes apoptotic signaling [159, 160]. This is due in part to the caspase-3- mediated cleavage and inactivation of the p75 subunit of complex I of the electron transport chain [159]. Although less characterized, it is likely that a similar decrease in the cytosolic [ATP]/[AMP] ratio occurs during necroptosis.

Using the same TNFa/BV6 stimuli in FADD-deficient Jurkat cells as in our studies, Schenk and Fulda recently reported that necroptotic progression is faciliated by a positive feed-back loop whereby activated MLKL promotes accumulation of mitochondrial reactive oxygen species (ROS) which, in turn, further stabilize formation of RIP1/RIP3 necrosomes [161]. Such diversion of mitochondrial function to ROS production will decrease ATP production, while insertion of MLKL pores into the plasma will increase ATP breakdown to ADP and AMP as a result of increased Na+, K+-ATPase and Ca2+-ATPase activity secondary to the enhanced Na+ influx [162] and Ca2+ influx [163] mediated by

MLKL pores.

An increasing accumulation of AMP, rather than ATP, in the in vivo extracellular microenvironment of apoptotic or necroptotic tumor cells may be significant due to its modulatory effects on pro-immunogenic and pro-

125

inflammatory signaling. In contrast to ATP (or ADP), AMP is not an agonistic

ligand for the immunostimulatory P2 receptors expressed on phagocytic

leukocytes or tumor antigen-reactive T cells. Rather, extracellular AMP is a potential source for the production of adenosine which activates the immunosuppressive A2a/ A2b adenosine receptors expressed on those leukocytes and T cells. However, this potential capacity of extracellular AMP to drive local accumulation of immunosuppressive adenosine will be determined by the relative expression of the CD73 or alkaline phosphatase ecto-AMPases on

the tumor cells per se, adjacent stromal cells, anr/or recruited leukocytes.

Upregulation of CD73 and other ecto-nucleotidases in FADD-deficient and RIP1-

deficient Jurkat cells

Reduced expression of, or accumulation of inactivating mutations in, pro-

apoptotic or pro-necroptotic signaling proteins often contributes to tumor cell

resistance to various chemotherapeutic agents. Decreased expression or activity

of FADD in Jurkat cells uncouples Fas/ TNFR1/ TRAIL-family death receptors from induction of the extrinsic apoptotic program [164] while decreases in RIP1 expression/activity suppress necroptotic induction, but not apoptotic induction, by the death receptors [165-167]. However, chemotherapeutic agents that predominantly activate the intrinsic apoptotic program can bypass such deficiencies in FADD or RIP1 mutations to induce tumor cell death. Consistent with this scenario, we verified efficient stimulation of intrinsic apoptosis in FADD-

126

or RIP1-deficient Jurkat cells treated with staurosporine or etoposide, two

mechanistically distinct chemotherapeutic/ pro-apoptotic drugs. Staurosporine, a

broad spectrum kinase inhibitor, induces exceptionally rapid intrinsic apoptosis in

most cell types by a mechanism that involves, in part, direct modulation of

caspase-9 independently of Apaf-1 [168]. Although not used as in vivo

chemotherapeutic agent, staurosporine is the prototype for related drugs, such

as 7-hydroxystaurosporine/ UCN-01, that target leukemia and lymphoma in

human subjects [169, 170]. Etoposide stimulates apoptosis via its inhibitory

action on type 2 toposiomerase and is a widely-used cancer chemotherapeutic

drug. Despite efficient induction of apoptosis, as indicated by casp3-mediated

PARP cleavage, and robust open-gating of Panx1 channels, as indicated by YO-

PRO2+ influx and calcein4- efflux, in FADD- or RIP1-deficient Jurkat cell treated with STS or Etop, little or no extracellular adenine nucleotide accumulation was observed under these conditions. This uncoupling of Panx1 channel activity from

ANex accumulation was linked to increased expression of CD73 ecto-AMPase in

both Jurkat lines, and additional upregulation of CD39 ecto-ATPase and another

AMP-degrading ecto-enzyme in the RIP1-deficient Jurkat cells.

Notably, only a 2-fold increase in CD73-type activity distinguished the

FADD-deficient cells from the wildtype Jurkat cells, but this was sufficient to

markedly reduce STS-induced ANex accumulation and completely suppress Etop-

stimulated accumulation. Interestingly, this modest increase in CD73 activity

also suppressed extracellular ATP accumulation in response to etoposide but not

staurosporine. This suggests that increased clearance of extracellular AMP by

127

CD73 may relieve product inhibition of the upstream CD39 ecto-ATPase and

thereby result in increased clearance of ATP if the latter is delivered to the

extracellular compartment at a relatively slow rate. This scenario would be

consistent with the slower generation of casp3-activated Panx1 channels during

Etop-induced apoptosis relative to STS-induced apoptosis. Moreover, the

increased CD73 activity of the FADD-deficient Jurkat cells was insufficient to

prevent robust extracellular accumulation of either the summed ATP+ADP+AMP

pool or ATP alone during TNFa/BV6-induced necroptosis. This implies that the

rate of efflux of ATP and AMP via the activated MLKL pores markedly exceeds

the rate of AMP clearance by the upregulated CD73 activity. Importantly,

inclusion of the APCP inhibitor of CD73 “rescued” the apoptotic ANex

accumulation responses to STS and Etop in the FADD-deficient cells and further enhanced the necroptotic ANex accumulation elicited by TNFa/BV6.

The absence of RIP1 expression in Jurkat cells resulted in an even more

dramatic suppression of the ANex and ATP accumulation responses during STS-

or Etop-induced intrinsic apoptosis, as well as Fas-induced extrinsic apoptosis.

Although the RIP1-deficient Jurkat line was characterized by an ~8-fold greater

level of CD73 activity relative to the parental control Jurkat cell line, inclusion of

the APCP inhibitor did not markedly restore the ANex accumulation responses

during apoptosis. This indicated that RIP1-deficient cells expressed higher levels

of other ectonucleotidases in addition to CD73. Although we found a 2-fold

upregulation of CD39-type ecto-ATPase activity, increased levels of this ecto-

enzyme cannot explain the APCP-resistant clearance of AMP. We also ruled out

128 the contribution of tissue non-selective alkaline phosphatase, another GPI- anchored ecto-enzyme that can efficiently hydrolyze AMP to adenosine [171,

172]. It will be important to identify this alternative AMP-clearing ecto-enzyme in future experiments. An interesting candidate is the type 3 AMP deaminase, the product of the Ampd3 gene which is the most widely expressed of the three major AMP deaminase isoforms in humans [173-180]. Although AMPDs are canonically associated with regulation of intracellular AMP turnover, one study

[181] identified an extracellular role for AMPD at the neuromuscular synapse.

This ecto-enzyme effectively competed with CD73 for extracellular AMP catabolism to produce IMP at the expense of the extracellular adenosine generation and consequent activation of A2 adenosine receptor signaling. The possible expression of an ecto-AMPD activity by tumor cells (and/or by tumor stroma and recruited leukocytes) would have significant implications for the accumulation of extracellular purines in the tumor microenvironment. Such an activity would divert extracellular AMP accumulation away from the production of immunosuppressive adenosine.

Significance of increased ecto-nucleotidase activities during regulated tumor cell death

Release of ATP from apoptotic or necroptotic tumor cells is important for the activation and modulation of the anti-tumor immune response that occurs during treatment with certain chemotherapeutic agents. In contrast, released

129

AMP is not an agonist for immunostimulatory P2 nucleotide receptors but is a potential substrate source for production of immunosuppressive adenosine.

Because wild-type Jurkat cells are defined by low rates of ATP and AMP hydrolysis, our studies indicate that even modest increases in CD73 activity can alter the extracellular balance of ATP versus adenosine and tip the scale towards decreased inflammation. The mechanisms by which deficiencies in FADD or

RIP1 expression modulate expression of CD73 and other ectonucleotidases remain unclear. Nonetheless, the results from this study indicate that net extracellular adenine nucleotide accumulation during various modes of tumor cell death signaling requires a nucleotide efflux pathway that exceeds the rate of clearance by cell-autonomous ecto-nucleotidases. Differential expression of

CD73 ecto-AMPase and other ecto-nucleotidases in clonal tumor variants may determine whether chemotherapeutic drug-induced activation of the Panx1 channel pathway predominantly drives accumulation of immunostimulatory ATP versus immunosuppressive adenosine. The suggests yet another pathway by which cancer cells can evade immune surveillance given the growing literature that increased CD73 in the tumor microenvironment is indicative of poor prognosis.

4.5 Acknowledgements

We would like to thank Dr. Silvia Penuela and Dr. Dale Laird for generously providing the panx1 anti-body. This work was supported, in whole or in part, by

National Institutes of Health Grant R01-GM36387 (to G. R. D.). Also, supported in part by National Institutes of Health Training Grant T32-GM008803.

130

Figure 4.1: TNFα-induction of necroptosis or extrinsic apoptosis induces

release of adenine nucleotides from Jurkat cancer cells via mechanistically

distinct pathways

A, A schematic of Fas and TNF receptor signaling. Fas can trigger the formation

of the FADD-Casp8 death-inducing signaling complex activating executioner

caspases-3/7 and leading to apoptosis. TNFR1 can trigger apoptosis by forming

a TRADD-FADD-Casp8 complex after TS treatment; apoptosis can also be

triggered by TS by formation of a RIP1-RIP3-FADD-procaspase 8/cFLIP

complex. TNFR1 can trigger necroptosis after treatment with TS+zVAD in wild-

type cells by forming the necrosome that contains RIP1-RIP3 and recruits MLKL,

and cell survival/inflammation by forming complex 1 that contains TRADD-

TRAF2-RIP1-cIAP1/2. B, E, H. WT Jurkat cells were treated with either

250ng/mL αFas for the indicated times or were pretreated with 3μM of the

SMAC-mimetic BV6 for 2 h followed by stimulation with 20ng/mL TNF-α for the indicated times. C, F, I. FADD-def Jurkat cells were pretreated with 3μM BV6 for

2 h followed by stimulation with 20ng/mL TNF-α for the indicated times. All treatments were done in the presence or absence of 20μM zVAD or 1μM NSA where indicated. Both apoptosis and necroptosis lead to ATP efflux. B and C,

Cell supernatants were taken at 0h, 2h, 4h, 8h, and 12h post treatment and measured for release of LDH. LDH released from TS treated cells was assayed, analyzed and normalized to untreated control cells that had been lysed as described in the “Experimental Procedures.” Experiments with each agent were repeated 3-4 times with data indicating mean ± S.E. for n=4 (WT) and n=3

131

(FADD-def). D-F, cells were treated with either αFas (D) or TS (E and F) and

supernatant was collected at 0h, 2h, 4h, 8h, and 12h post treatment and

analyzed for summed ATP+ADP+AMP. Data indicate mean ± S.E. of n=3. G-I*,

cell lysates were collected for Western Blot analysis as described in Chapter 2

and analyzed for expression of Panx1, PARP, and Actin. Data are representative

of 3 experiments for each stimulus. *For uncropped Panx1 Western Blot see

Appendix 5 and 6

132

Figure 4.1: TNFα-induction of necroptosis or extrinsic apoptosis induces release of adenine nucleotides from Jurkat cancer cells via mechanistically distinct pathways

133

Figure 4.2: Intrinsic apoptotic signaling and Panx1 channel cleavage in

FADD-deficient Jurkat cancer cells is uncoupled from accumulation of

extracellular adenine nucleotides

WT and FADD-def cells were treated with 3μM STS or 20μM Etop in the

presence of 100μM zVAD or 1μM NSA where indicated. A* and B*, cell lysates

were collected for Western Blot analysis and analyzed for Panx1, PARP, and

Actin. Data are representative of 3 experiments for each cell type. C and D,

supernatants were collected at 0h, 2h, 4h, 8h, and 12h post treatment and

assayed for release of LDH as described in Fig 1. Data indicating mean ± S.E.

n=3. E and F, Samples of the extracellular media were assayed for total

ATP+ADP+AMP. Data indicating mean ± S.E. of n=3. *For uncropped Panx1

western blot see Appendices 7 and 8.

134

Figure 4.2: Intrinsic apoptotic signaling and Panx1 channel cleavage in

FADD-deficient Jurkat cancer cells is uncoupled from accumulation of extracellular adenine nucleotides

135

Figure 4.3: Caspase-3-cleaved Panx1 channels are functionally active in

FADD-deficient Jurkat cancer cells during intrinsic apoptosis

A, Schematic illustrates caspase-3 activated Panx1-channel function as a pathway for ATP release or efflux/influx of charged organic dyes. B, WT, and

FADD-def Jurkat cells were treated with 3μM STS for 4 h, or 20μM Etop for12h, then washed and resuspended in basal salt solution supplemented with 1μM YO-

PRO2+ and incubated for 20 min prior to quantification of YO-PRO2+ fluorescence per well. Data indicating mean ± S.E. n=3; ns, p> 0.05; ***, p<

0.001.C, WT, and FADD-def Jurkat cells were loaded with 1μM calcein-AM as described in chapter 2 then treated with STS for 4h in the presence or absence of 100μM zVAD or 50μM Trov. Cells were then washed and resuspended in basal salt solution and calcein4- fluorescence was quantified. Data indicating mean ±S.E. n=4-6 of individual wells from 2-3 experiments; **, p< 0.01; ***, p<

0.001. D, Cells were treated with STS or Etop for 4 or 12 hours, respectively, then washed and resuspended in basal salt solution supplemented with 1μM YO-

PRO2+ and incubated for 20 min prior to phase-contrast and epifluorescence microscopy imaging. Representative images of 2-3 experiments for each stimulus. E, WT and FADD-def Jurkat cells were treated with STS for 4hr then stained with MitoTracker® Red and DAPI as described in chapter 2. Cells were imaged by confocal microscopy at a 60X magnification. Data are representative images of n=15-19 individual cells.

136

Figure 4.3: Caspase-3-cleaved Panx1 channels are functionally active in

FADD-deficient Jurkat cancer cells during intrinsic apoptosis

137

Figure 4.4: CD73 ecto-nucleotidase activity is upregulated in FADD-

deficient Jurkat cancer cells and counteracts Panx1 channel-mediated

efflux of ATP/AMP during intrinsic apoptosis

A and B, WT and FADD-def cells were treated with 3μM STS or 20μM Etop and

samples were taken at 0h, 2h, 4h, 8h, and 12h and assayed for summed

ATP+ADP+AMP (A) and ATP only (B). Data indicating mean ± S.E. of n=3. C,

Schematic of CD73 activity, Induction of apoptosis by extrinsic or intrinsic

pathways leads to the gating of the panx1 channel, which can be measure via

influx of YO-PRO dye or efflux of Calcein4- dye. This gating allows for release of

ATP, ADP, and AMP from the cell. In FADD-def cells an increase in CD73

activity causes increased hydrolysis of AMP to Ado, which can be inhibited by

APCP. D, WT and FADD-def cells were treated with STS, 4h, in the presence or

absence of 50μM APCP. Supernatants were collected and analyzed for total

ATP+ADP+AMP. Data indicate mean ± S.E. of n=6-9 wells of 2-3 experiments;

***, p< 0.001. C, WT and FADD-def Jurkat cells were treated with STS in the

presence of APCP and in the presence or absence of 100μM zVAD or 50μM

Trov. Data indicate mean ± S.E. of n=6-9 wells of 2-3 experiments; NS, p> 0.05;

***, p< 0.001. E, FADD-def Jurkat cells were treated with TS for 4h in the

presence or absence of APCP. Data indicate mean ± S.E. of n=3 wells; NS,

p>0.05; ***, p<0.001. F, 1μM AMP was added to untreated WT, FADD-def in the presence or absence of 50μM APCP. Supernatants were collected after 30 min and analyzed for levels of remaining AMP. Data indicate mean ±S.E. of n=6-9 wells of 2-3 experiments; NS, p> 0.05; *, p< 0.05; ***, p< 0.001.

138

Figure 4.4: CD73 ecto-nucleotidase activity is upregulated in FADD- deficient Jurkat cancer cells and counteracts Panx1 channel-mediated efflux of ATP/AMP during intrinsic apoptosis

139

Figure 4.5: Apoptotic signaling and Panx1 channel activation in RIP1-

deficient Jurkat cancer cells is also uncoupled from accumulation of

extracellular adenine nucleotides

WT and RIP1-def Jurkat cells were with 3μM STS or 20μM Etop for 4h or 12h,

respectively. A-C, Cells were treated with STS (A), Etop (B) or 250ng/mL αFas

(STS, αFas) n=7 (Etop) of individual wells from three experiments; NS, p> 0.05;

***, p< 0.001. D*, cell lysates were collected for Western Blot analysis and

analyzed for Panx1, PARP and Actin. Data are representative of 3 experiments.

E, WT and RIP1-def Jurkat cells were treated with STS for 4hr then stained with

MitoTracker® Red and DAPI. Cells were imaged by confocal microscopy at a

60X magnification. Data are representative images of n=15-19 individual cells. F,

Cells were treated with STS or Etop for 4 of 12 hours, respectively and cell

suspensions were analyzed for YO-PRO2+ accumulation as described in Fig 3.

Representative phase-contrast epifluorescence microscopy images of 3 experiments. G, WT, and RIP1-def Jurkat cells were treated with 3μM STS, 4h, or 20μM Etop, 12h,then washed and resuspended in basal salt solution supplemented with 1μM YO-PRO2+ and incubated for 20 min prior to quantification of YO-PRO fluorescence per well. Data indicating mean ± S.E. n=3; p> 0.05; ***, p< 0.001. H, WT, and RIP1-def Jurkat cells were loaded with

1μM calcein-AM then treated with STS for 4h in the presence or absence of

100μM zVAD or 50μM Trov. Cells were then washed and resuspended in basal salt solution and calcein4- fluorescence was measured. Data indicating mean

140

±S.E. n=4-6 of individual wells from 2-3 experiments; **, p< 0.01; ***, p< 0.001.

*For uncropped Panx1 western blot see Appendix 9.

141

Figure 4.5: Apoptotic signaling and Panx1 channel activation in RIP1- deficient Jurkat cancer cells is also uncoupled from accumulation of extracellular adenine nucleotides

142

Figure 4.6: Increased expression of CD73 ecto-AMPase in RIP1-deficient

Jurkat cells relative to FADD-deficient or wildtype Jurkat cells

A, WT, FADD-def, and RIP1-def Jurkat cells were treated with STS, 4h, in the

presence or absence of 50μM APCP. Supernatants were collected and analyzed

for total ATP+ADP+AMP. Data indicate mean ± S.E. of n=6-9 wells of 2-3

experiments; ***, p< 0.001. B, WT, FADD-def and RIP1-def cells were

suspended at 2x106 cells/mL. 1μM AMP was added in the presence or absence of 5mM Tetra and supernatant was collected after 30 min C, RIP1-def Jurkat cells were resuspended at cell concentrations of 2x106 cells/mL, 2x105 cells/mL,

or 2x104 cells/mL. 1μM AMP was added and supernatant was collected after 30

min. AMP levels were assayed as described above. Data indicate mean ± S.E. of

n=6 wells of 2 experiments. Insert: 1 μM AMP was added to RIP1-def cells

resuspended at 2x104 cells/mL in the presence or absence of APCP or 5mM

Tetra. Supernatants were collected after 30 min and analyzed for remaining

levels of AMP. Data indicate mean ± S.E. of n=6 wells of 2 experiments; *, p<

0.05. D, E, 10μM ε-AMP was added to untreated cells that were resuspended at

2x106 cells/mL in BSS+5mM glucose+0.1% BSA or BSS+5mM glucose+0.1%

BSA alone. Supernatants were collected after 30min and processed for separation and quantification of ε-ADO and ε-AMP by HPLC and fluorescence detection as described in chapter 2. Data are representative of 3 experiments.

143

Figure 4.6: Increased expression of CD73 ecto-AMPase in RIP1-deficient

Jurkat cells relative to FADD-deficient or wildtype Jurkat cells

144

Figure 4.7: Increased expression of CD39 ecto-ATPase in RIP1-deficient

Jurkat cells relative to FADD-deficient or wildtype Jurkat cells

A, WT, FADD-def, and RIP1-def Jurkat cells were treated with 3μM STS for 4

hours. Supernatants were collected and assayed for extracellular ATP. Data

indicate mean ± S.E. of n=3; NS, p>0.05; *, p< 0.05. B, WT, FADD-def, and

RIP1-def cells were suspended at 2x106 cells/mL. 1μM of ATP was added in the presence or absence of ARL. Supernatants were collected after 30min and analyzed for remaining levels of ATP. Data indicate mean ± S.E. of n=3-6 wells of 2 experiments; NS, p> 0.05; *, p< 0.05; **, p< 0.01.

145

Figure 4.7: Increased expression of CD39 ecto-ATPase in RIP1-deficient

Jurkat cells relative to FADD-deficient or wildtype Jurkat cells

146

Chapter 5: Regulation of Adenine Nucleotide release through Pannexin 1

channels in EG7 Murine Lymphoma Cells

Activation of the anti-tumor immune response in vivo is commonly studied

by using EG7 murine lymphoma cells that overexpress the OVA peptide, or their parental EL4 cell line. EG7 cells treated with chemotherapeutic agents have been shown to release significant amounts of ATP, which is a key component of the activation of the immune response [5, 7]. Our chapter 3 and 4 studies

indicate the human Jurkat T cells release ATP and other adenine nucleotides via

caspase 3/7 cleavage of Panx1 during chemotherapeutic drug treatment.

However, it is unknown whether this holds true in the EG7 model. In the this

study we aimed to determine if EG7 cells utilize the Panx1 ATP release pathway

during treatment with pro-apoptotic chemotherapeutic agents doxorubicin (Dox)

and staurosporine (STS).

To determine if EG7 murine lymphoma cells undergo Panx1-mediated

ATP release, we first compared the activation of apoptosis in EG7 cells to

activation in Jurkat T cells in response to STS or Dox. We hypothesized that due

to the same treatment, the EG7 cells would also exhibit similar kinetics of

caspase 3 activation and cell death. Using an enzymatic casp3 activity assay we

measured the amount of active casp3 in EG7 cells treated with STS or Dox and

compared it to Jurkat T cells (Fig 5.1A and 5.1B). In STS-treated EG7 cells there

was an increase in active casp3 starting at ~2hours post treatment with peak

activation occurring between 4 and 8 hours. In Dox-treated cells, the casp3

activation occurred more slowly due to its mechanism of action and thus took ~8

147

hours for significant activation. The activation of casp3 due to STS or Dox was

confirmed by western blot analysis of PARP1 (a casp3 substrate) cleavage (Fig

5.1C and D and 5.1 E and F). This pattern is similar to what is observed in Jurkat

T cells. Together, these data confirm that treatment with STS or Dox in different

T cell leukemia or lymphoma cell lines elicits a similar apoptotic cell death

response.

Next, we hypothesized that, due to the similar kinetics of casp3 activation,

the cleavage of Panx1 would be similar between cell lines. We performed western blot analysis of EG7 cells treated with either STS or Dox. We found that both Dox and STS treated cells induced proteolytic processing of Panx1 with

similar kinetics to that observed in Jurkat T cells (Fig 5.2A and B). These data

correspond with the casp3 activation kinetics.

We next postulated that due to the similar activation of casp3 and

cleavage of Panx1 when compared to Jurkat T cells, EG7 cells would release

significant adenine nucleotides via Panx1 following treatment with either STS or

Dox. To evaluate this hypothesis, we measured total extracellular

ATP+ADP+AMP (ANex) accumulation from EG7 cells treated with either STS or

Dox over an 18 hour time course. Surprisingly, while Dox-treated EG7 cells

showed significant ANex co-temporal with cleavage of Panx1, STS-treated EG7 cells did not release significant ANex until approximately 8 hours post treatment, 6

hours after significant Panx1 cleavage (Figure 5.2C). This suggests that adenine

nucleotide release during STS treatment of EG7 cells occurs via a Panx1-

independent mechanism. Additionally, the presence of zVAD did not attenuate

148

STS or Dox induced ANex accumulation (Figure 5.2C). These data indicate that

while Panx1 is being cleaved during both STS and Dox treatment in EG7 cells,

this is not correlated with adenine nucleotide efflux.

To determine why adenine nucleotides were not being released via Panx1 we tested whether or not the channel was functional. It has been shown that S-

nitrosylation at key cysteine residues in the protein can covalently modify the

channel and decrease its function [182]. Using YO-PRO2+ dye influx as a measure of Panx1 channel gating and function, we treated EG7 cells with STS and measured influx of the dye over a 4 hour time period (Figure 5.3).

Approximately 2 hours after treatment with STS YO-PRO2+ dye uptake was

observed in the EG7 cells. This influx was suppressed in the presence of the

Panx1 inhibitor carbenoxolone (CBX). Both the kinetics and the magnitude of

influx were similar to that in Jurkat T cells treated with STS. This indicates that

Panx1 is being gated as a functional channel.

These data presented demonstrate a distinct uncoupling between Panx1

gating and ATP release. These results are similar to those obtained with FADD-

def and RIP1-def Jurkat T cells (see Chapter 4). While it is possible that EG7

cells also have increased ectonucleotidase activity that counteracts Panx1-

mediated extracellular nucleotide accumulation, the significant accumulation at

later time points in STS-treated EG7 cells and the lack of zVAD inhibition in Dox-

treated cells point to another regulatory mechanism for inhibiting Panx1-mediated

release. This, coupled with our functional YO-PRO2+ influx data, suggests

alterations in the selectivity of the channel. A potential regulating mechanism for

149 the altered selectivity is the phosphorylation of key residues located within the pore, which will be discussed in further detail in chapter 6, Discussion and Future

Directions.

150

Figure 5.1: Comparative activation of apoptosis in Jurkat T cells and EG7 cells treated with chemotherapeutic agents

Jurkat T cells and EG7 cells were treated with 3μM STS or 25 μM Dox. Samples were taken at 0hr, 2hrs, 4hrs, 8hrs, 12hrs and 18hrs. A, cells were assayed for active caspase 3. Data represent mean ± S.E. of n=3. B and C, cell samples were processed and assayed by Western Blot analysis for PARP1 and actin as outlined in chapter 2. Data is representative of n=3 experiments.

151

Figure 5.1: Comparative activation of apoptosis in Jurkat T cells and EG7 cells treated with chemotherapeutic agents

152

Figure 5.2: Activation of apoptosis in EG7 cells by chemotherapeutic

agents is not correlated with Panx1 mediated adenine nucleotide efflux

A and B*, EG7 cells were treated with STS (A) or Dox (B) for the indicated times and samples were collected and analyzed by Western Blot for Panx1 and actin.

Data is representative of n=3 experiments. C, cells were treated with STS or Dox for the indicated times and supernatants were assayed for summed

ATP+ADP+AMP as indicated in chapter 2. Data represent mean ± S.E. of n=3-6 wells of 1-2 experiments. *For uncropped Panx1 Western Blot see Appendix 10.

153

Figure 5.2: Activation of apoptosis in EG7 cells by chemotherapeutic agents is not correlated with Panx1 mediated adenine nucleotide efflux

154

Figure 5.3: Cleavage of Panx1 in EG7 cells is correlated with increased YO-

Pro dye influx

EG7 cells and Jurkat T cells were suspended in basal saline supplemented 1 μM

YoPro ± 100 μM CBX ± 100 μM zVAD transferred to the wells of a 24-well plate.

Fluorescence (485 nm/540 nm) was measured at 1 min intervals for 15 min prior

to addition of 3 μM STS (or vehicle) and then at 1 min intervals for an additional 4

h prior to the addition of digitonin (Dig) to permeabilize the cells. Data are

representative of 3 experiments.

155

Figure 5.3: Cleavage of Panx1 in EG7 cells is correlated with increased YO-

Pro dye influx

156

Chapter 6: Discussion and Future Directions

Our studies described herein provide further insight into the balance of extracellular ATP and adenosine in the tumor microenvironment. Through three distinct studies we have demonstrated: 1. the release pathway for pro- inflammatory ATP and other adenine nucleotides during both apoptosis and necroptosis, 2. the modulation of accumulation of these adenine nucleotides when key cell death pathways are mutated, and 3. other potential factors, such as phosphorylation, involved in release of ATP via Panx1. These studies highlight the importance of examining both release and breakdown of pro- inflammatory ATP when studying the anti-tumor immune response elicited by chemotherapies.

In the study presented in chapter 3, we elucidated the signaling pathway that leads to the release of ATP, ADP, and AMP from cancer cells during chemotherapy treatment. Activation of intrinsic apoptosis causes the activation of executioner caspases 3 and 7, which are able to cleave the auto-inhibitory C terminal of the Panx1 channel. Once gated, ATP, ADP, and AMP are able to traverse the channel and accumulate in the extracellular space. Furthermore, in the study presented in chapter 4, we were able to compare apoptotic adenine nucleotide release to necroptotic release. Through this we discerned the distinct mechanism of necroptotic adenine nucleotide accumulation. Upon activation of the necroptotic cell death pathway, MLKL is phosphorylated by RIP3. This phosphorylation is necessary for the oligomerization and insertion of MLKL into the plasma membrane leading to disruption of plasma membrane homeostasis,

157

cell swelling and lysis. The release of adenine nucleotides is dependent on the

oligomerization and insertion of pMLKL into the plasma membrane. Due to the

rapidity of the cell lysis once the pMLKL pores are formed, the majority of the

adenine nucleotides are released concurrently with cell lysis. Additionally, while

apoptosis and necroptosis utilize distinct pathways to release adenine

nucleotides, during both cell death processes the majority of adenine nucleotide

released is in the form of ADP and AMP. In these studies we only examined two

forms of PCD and their mechanism of adenine nucleotide release, but other

forms of PCD are able to be activated in cells to varying degrees due to other

stimuli. This raises the question as to whether all forms of PCD utilize these two

pathways or if alternative mechanisms for release exist.

In the study presented in chapter 4, we discovered that during

chemotherapeutic drug treatment in Jurkat T cells that were deficient in FADD or

RIP1 the cells were unable to accumulate adenine nucleotides to the same extent as wild-type Jurkat T cells. We determined that these cell lines had

increased activity of AMP hydrolysis by the ectonucleotidase CD73. However, it

is unknown if FADD or RIP1 directly or indirectly regulate the activity or

expression of CD73.

Finally, in the study presented in chapter 5, we examined the mechanism

of chemotherapeutic drug-induced adenine nucleotide release in a murine

lymphoma cell model, EG7 cells. Although Panx1 channels were gated by casp3

and the cells released adenine nucleotides, the release was not via Panx1. This

158

indicates other factors, potentially phosphorylation of key Ser/Thr or Tyr residues,

may be necessary for release of adenine nucleotides via Panx1.

6.1 Mechanisms of Programmed Cell Death Induced Adenine Nucleotide

Release

Apoptosis and necroptosis are only two forms of PCD. Pyroptosis is a caspase 1-mediated form of cell death that is commonly activated in macrophages and DCs after inflammasome activation. Binding of a priming signal, i.e. binding of pathogen associated molecular patterns (PAMPs) to pattern recognition receptors (PPRs), induces upregulation of the key inflammasome proteins NLRP3 and IL-1β [183]. Inflammasome activators, such as ATP binding to P2X7 receptors, induce loss of intracellular K+, and NLRP3 forms the

inflammasome complex with the adaptor protein ASC and pro-caspase-1.

Assembly of this complex allows for the auto-proteolytic activation of casp1 which leads to the formation of a pyroptotic pore or channel that disrupts membrane homeostasis and leads to cell lysis. Unpublished data from our lab has recently

shown that during pyroptosis adenine nucleotides are released and that this

release is occurring via the pyroptotic pore prior to cell lysis. Recently, two

studies were published that identified the protein gasderminD (gsdmD) as the

downstream target of casp1 [184, 185]. Similar to pMLKL, when gsdmD is

cleaved it is thought to form oligomers that are capable of insertion into the

plasma membrane. This insertion would allow for the release of ATP and its

metabolites into the extracellular space and provide another mechanistically

distinct pathway for ATP release.

159

6.2 Role of FADD and RIP1 in the regulation of ectonucleotidase activity

In chapter 4 we described how Jurkat T cells deficient in either FADD or

RIP1 had increased CD73 activity, and in the case of the RIP1-def Jurkat T cells,

increased CD39 activity as well. However, it is unknown if these two proteins regulate the activity or expression of CD73 or CD39. T cells have various expression levels of CD39 and CD73 depending on their stage of development.

Naive T cells, for example, express increased levels of CD73. This expression allows for the maintenance of memory T cell populations upon activation during

weak T cell receptor (TCR) signaling. However, when strong TCR activation

occurs and the naive T cell becomes differentiated into an effector T cell,

expression of CD73 decreases [149]. Due to the plasticity of the genes in T cells

it is highly plausible that changes in major signaling pathways can induce

upregulation of these ectonucleotidases. Future directions aim to determine if

FADD or RIP1 are either directly or indirectly involved in modulation of CD73

activity, as well as CD39 activity in the case of RIP1. To test this, we will perform

a rescue assay by reintroducing FADD or RIP1 into their respective deficient cell

lines. If, after reintroduction of the proteins, the phenotype of increased AMP

hydrolysis is reversed we would then perform a co-immunoprecipitation to

determine additional binding partners of FADD or RIP1. Additionally, RIP1 can

act as both a kinase and an adaptor protein. To determine if the role of RIP1 is

dependent on its kinase activity we will add back kinase dead RIP1. If the

phenotype is reversed in the absence of kinase activity then we know that the

160

function of RIP1 in CD73 and CD39 activity is dependent on its role as an

adaptor. Co-IPs will help determine any downstream components.

Both the FADD and the RIP1 deficient cell lines were generated through

frame shift mutations using ICR-191 and selected in the presence of activating

Fas anti-bodies for the FADD-def cell line and TNF-α for RIP1-def cells [57, 186].

Analysis of other mutations in these cell lines was limited to the cell death pathways they were directly connected to; therefore, it is possible that mutations or alterations in the transcriptional regulatory pathway for CD73 and CD39 also occurred. The ecto-5’-NT gene, which encodes CD73, contains a number of regulatory binding sites within and upstream of the promoter region. Studies have uncovered a cAMP Response Element (CRE) binding site that suppresses transcription of the gene. Mutations at its binding site in Jurkat T cells show increased CD73 mRNA expression as well as increased enzymatic activity.

Researchers were able to identify members of the ATF family, specifically ATF-1 and ATF-2, as the transcription factors that bind to the CRE binding site [187].

Due to its role in suppressing CD73 transcription, it is plausible that during mutation to generate the FADD-def and RIP1-def cell lines there was also a mutation in either the CRE binding site or of the CRE transcription factors.

Additionally, other transcriptional regulatory elements are able to promote the transcription of CD73. When the Sp1/AP2 transcriptional binding site was mutated in Jurkat T cells there was a decrease in the amount of CD73 mRNA

[187]. Mutations in the binding site that increase the affinity for the Sp1 transcription factor or an overall increase in generation of the transcription factors

161 may cause an increase in CD73 expression. Alternatively, it has been shown that during Th17 cell differentiation, both CD73 and CD39 expression is increased via

IL-6 mediated Stat3 activation and TGF-β mediated downregulation of the growth factor inhibitor-1 (Gfi-1) protein, both of which are able to bind to and regulate the transcription of the ectonucleotidases [188]. Taken together, mutations in either the binding sites of these regulatory elements or in the proteins that bind to them may be causing the upregulation of CD73 in both deficient cell lines and in CD39 in the RIP1-def cells. Future directions will be directed towards the analysis of the genomic sequence of the ecto-5’-NT gene and the ENTPD1 gene that encodes for CD39 in WT, FADD-def and RIP1-def Jurkat T cells. This will allow for determination of mutations in the regulatory binding sites in the promoter region when the deficient cell lines are compared to the WT controls. Additionally, microarray analysis of the gene expression in the mutated cell lines compared to the WT control cells can help us determine if any of the transcription factors that regulate CD73 and CD39 expression are increased or decreased. Once identified, RT-PCR will allow us to quantitatively measure the amount of the change.

Due to the use of mutagens to generate the FADD and RIP1 deficient cell lines, other methods to determine a direct relationship with ectonucleotidase expression are necessary. Since expression of these genes can increase and decrease throughout the differentiation process of T cells, it is plausible that

FADD or RIP1 could be upstream regulators of the expression of the ectonucleotidases. This can be examined through use of CRISPR/cas9 knock-

162

out of either FADD or RIP1 in WT Jurkat T cells. Additionally, while FADD knock-

out mice are embryonic lethal and RIP1 knock-out mice are perinatal lethal, use of embryonic fibroblasts from the knock-out mice would allow us to not only examine if FADD and RIP1 are modulating activity of CD73, but would also allow us to determine if the phenotype is specific for T cells.

6.3 Decreases in immune cell activation by altered ectonucleotidase activity

Increases in expression of CD39 and CD73 have been linked to poor prognosis due to the generation of anti-inflammatory adenosine [140]. While the

FADD-def and RIP1-def Jurkat cells respond to the chemotherapies in a similar fashion as WT Jurkat cells, the increase in CD73 activity would likely leads to a decreased immune response to tumors comprised of these mutant Jurkat cell lines. Binding of ATP to P2Y2 leads to increased chemotaxis toward the dying

tumor cells [49]. To determine if the increase in CD73 activity correlates to a

decrease in chemotactic capabilities of immune cells, trans-well migration assays

will be utilized. Supernatants from treated and untreated WT, FADD-def, and

RIP1-def cells are placed into 96 well plates while THP-1 monocytes are placed

on top of a 5μm-pore size Transwell and after incubation the number of cells that

migrated through the pores will be quantified using CyQuant dye. We

hypothesize there will be a decrease in migration of THP-1 monocytes with

supernatants from FADD-def and RIP1-def Jurkat cells relative to supernatants

from WT Jurkat cells. This will indicate a decreased ability of supernatants from

FADD-def and RIP1-def Jurkat cells to activate an immune response.

163

Another method to measure the ability of FADD-def and RIP1-def cells to recruit immune cells in vivo is to perform an air-pouch experiment [49]. Air

pouches are formed in mice by injection of 5mL of sterile air under the skin on

the dorsum on day 1 and maintained by injection of 3mL of sterile air on day 4.

On day 7, supernatant from treated and untreated WT, FADD-def, and RIP1-def

Jurkat cells is obtained and after filtration 1mL is injected into the air pouch. After

24 hours the air pouch will be lavaged and analyzed via flow cytometry for

infiltration of macrophages and monocytes (CD11b+/Gr-1low). Consistent with the

Transwell migration assay we predict there will be a decrease in the number of

infiltrating monocytes and macrophages in FADD-def and RIP1-def Jurkat T cell

supernatants compared to WT Jurkat T cell supernatants.

6.4 Post-translational modification of Panx1

In the chapter 5 study, we showed how Panx1 was being gated but not

utilized for adenine nucleotide release in EG7 murine thymoma cells after

treatment with STS or Dox. Previous studies have shown that due to these

chemotherapy treatments trigger ATP release from the dying EG7 cells.

However, our studies show that there is an uncoupling between the proteolytic

activation of Panx1 channels and the release of ATP and other nucleotides. This

suggests that other post-translational modifications may contribute to the gating

or conductance of the channel, such as S-nitrosylation and phosphorylation.

The role of phosphorylation of the Panx1 channel in regulating activity is

now emerging as an alternative way of modulating this channel. A study by

164

Lohman et, al. recently demonstrated that phosphorylation of Panx1 by the Src

family kinases (SRK) during TNF-α treatment of endothelial cells was necessary

for transient activation of the channel [189]. Further studies will be needed to determine if phosphorylation of the Panx1 channel can modulate the release of

ATP and adenine nucleotides during other modes of Panx1 gating. It is possible that in the EG7 cells phosphorylation of key residues necessary for full gating of the channel is unable to occur. This would explain why we see gating of Panx1 but not release of the adenine nucleotides. Future studies will address two main points: Does phosphorylation occur during chemotherapy treatment, and does this phosphorylation alter the selectivity of the channel?

To address the first question, immunoprecipitation of the Panx1 channel after treatment of chemotherapies will be done and assessed for the presence of commonly phosphorylated residues, mainly Tyr, Ser, Thr. Controls of untreated

EG7 cells as well as cells that undergo Panx1 mediated ATP release during chemotherapy treatment, i.e. Jurkat T cells, will be used to compare EG7 cells treated with the drugs. If there is higher Panx1 phosphorylation in Jurkat T cells compared to EG7 cells, we will then assess the importance of this phosphorylation on the selectivity of the Panx1 channel using patch clamp.

Identification and mutation of the phosphorylation sites will allow us to determine the role of phosphorylation during ATP release via Panx1.

While kinetic analysis of YO-Pro2+ influx during STS treatment suggested

that the Panx1 channels are functional in EG7 cells, we still cannot rule out the

role of S-nitrosylation. S-nitrosylation has been shown to modulate Panx1

165

channel activity when nitric oxide interacts with the reactive cysteine residues

causing conformational changes [182, 190]. Additionally, it has been shown that

the murine C terminus is more flexible than that of the human Panx1 and is able to exit the pore more readily [47]. It is possible that in order to reduce spontaneous gating that there is an increase in S-nitrosylation at the C terminal

within the pore and unless these bonds are reduced the C terminus will be

unable to diffuse away from the pore, even after proteolytic cleavage. Therefore, future directions will address the possibility of decreased release of ATP and adenine nucleotides via Panx1 due to increased S-nitrosylation. Measurements

of ATP and total adenine nucleotide accumulation will be conducted in the

presence of a general reducing agent, dithiothreitol (DTT). Creating point

mutations at the cysteine residues involved will allow us to further discern the

potential role of S-nitrosylation in the regulation of Panx1 mediated ATP release.

6.5 Concluding Remarks

Together, these studies and proposed future directions allow us to better

understand modulation of the regulated release of ATP, ADP and AMP and their

breakdown to adenosine by leukemic T cells. This not only has implications on

the study of the activation and suppression of the anti-tumor immune response

due to chemotherapy treatment, but also immunotherapy as a whole.

These studies revealed the release mechanism of ATP, ADP and AMP

through Panx1 after cleavage of the C terminal domain during treatment with

chemotherapies. Additionally, we showed that Panx1 is highly upregulated in

166 various leukemia cells and it has also been shown to be increased in multiple other aggressive cancers such as metastatic melanoma [113, 126]. In the study on melanoma cells, Panx1 expression was correlated with the aggressiveness of the melanoma and knockdown of Panx1 decreased cellular migration and proliferation. This study suggested that the ATP released from Panx1 overexpressing cells induces cell autonomous purinergic signaling cascades that cause increases in intracellular calcium leading to increased cell proliferation.

While this suggests that Panx1 and ATP release is pro-tumorigenic, our studies, along with others, have suggested that this response can also be anti- tumorigenic. The magnitude of cellular ATP lost via activated Panx1 channels during chemotherapeutic treatment, coupled with release of other DAMPs that aid in recruitment of immune cells, not only indicates that this process releases sufficent ATP to activate the anti-tumor response, but also aids in depletion of energy stores to futher decrease survival of the tumor cells. Panx1 may also be beneficial as a biomarker for tumors that can potentially activate a strong anti- tumor immune response when treated with select chemotherapies. While this would not apply to all Panx1 overexpressing tumor cells, it may aid in selection of proper treatments.

With regard to activation of the anti-tumor immune response, these studies also highlight the importance of monitoring the breakdown of pro- inflammatory ATP to anti-inflammatory adenosine, particularly in cases of acquired multi-drug resistance. It has been shown that during generation of multi- drug resistant cells, increases in CD73 occur and this plays a role in the

167 resistance [56]. Our studies highlight the complexity of activation vs. suppression of the anti-tumor immune response after treatment with chemotherapeutic agents. In the first study we demonstrated that the majority of the adenine nucleotide released from apoptotic cells was in the form of AMP and this was supported by a study by Yamaguchi et al [99]. This is largely due to the disruption of mitochondrial function during apoptosis by casp3 [159, 191]. Due to the inability of the mitochondria to generate ATP, increased levels of intracellular

ADP and AMP accumulate and once Panx1 channels are gated these metabolites are released. Additionally, we demonstrated differential rates of AMP hydrolysis in cell lines from the same genetic background but with mutations in key cell death signaling cascades. While AMP by itself does not directly modulate pro- or anti-immune responses, the significant amount of released AMP, coupled to its role as a substrate for adenosine generation, point to an important function in balancing these opposing processes. The increased hydrolysis of AMP in the mutated cell lines during chemotherapy treatment is indicative of an increase in extracellular adenosine generation. This can lead to decreased immune responses and thus affect the efficacy of the drug while not affecting the direct cytotoxic effects. Adenosine’s role in cancer has been well studied [50]. Binding of adenosine to A2 receptors on immune cells has been shown to promote generation of M2 macrophages that promote angiogenesis, as well as T regulatory cells that suppress the cytotoxic effects of CD8+ T cells. Adenosine is also known to suppress the action of natural killer cells by preventing adhesion to tumor cells thereby limiting their cytotoxic capacity. Additionally, the mechanism

168

of the release can determine the magnitude of release that may be able to

overwhelm the activity of the ectonucleotidases. Necroptosis in FADD-def Jurkat

cells leads to rapid release of AMP that is able to saturate CD73 allowing for

significant accumulation of AMP to occur, while the slower kinetic of Panx1

mediated nucleotide release due to Etop treatment in FADD-def Jurkat cells

allowed for sufficient clearance of AMP upon release. The differential expression

of CD73 and CD39 and the various rates of release of ATP/AMP from different

dying tumor cell models can dictate if the release of adenine nucleotides via

Panx1 is immunostimulatory or immunosuppressive. Due to increased mutations

during the course of chemotherapy treatment that may lead to increases in CD73

activity, understanding the rates at which ATP is released and subsequently

metabolized into adenosine within the tumor microenvironment can provide a

deeper understanding of drug resistant tumors, as well as the purinergic anti- tumor immune response.

169

Appendix

Appendix 1: Full Western Blot of Panx1 in Anti-Fas treated Jurkat T cells in the presence of absence of zVAD. Data shown in Figure 3.2

55 43

170

Appendix 2: Full Western Blot of Panx1 in STS treated Jurkat T cells in the presence of absence of zVAD. Data shown in Figure 3.2

55 43

171

Appendix 3: Full Western Blot of Panx1 in Dox treated Jurkat T cells in the presence of absence of zVAD. Data shown in Figure 3.2

55 43

172

Appendix 4: Full Western Blot of Panx1 in Etop treated Jurkat T cells in the presence of absence of zVAD. Cropped data shown in Figure 3.2

55 43

173

Appendix 5: Full Western Blot of Panx1 in TS or Anti-Fas treated WT Jurkat T cells. Data shown in Figure 4.1

55 43

174

Appendix 6: Full Western Blot of Panx1 in TS or TS+zVAD treated FADD-def

Jurkat T cells. Data shown in Figure 4.1

55 43

175

Appendix 7: Full Western Blot of Panx1 in STS, Etop or Anti-Fas treated WT

Jurkat T cells. Data shown in Figure 4.2

55 43

176

Appendix 8: Full Western Blot of Panx1 in STS, Etop or Anti-Fas treated FADD- def Jurkat T cells. Data shown in Figure 4.2

55 43

177

Appendix 9: Full Western Blot of Panx1 in STS, or Etop treated WT and RIP1-def

Jurkat T cells. Data shown in Figure 4.5

55 43

178

Appendix 10: Full Western Blot of Panx1 in STS or Dox treated EG7 Lymphoma cells. Data shown in Figure 5.2

55 43

179

Appendix 11: Copyright permission

NATURE PUBLISHING GROUP LICENSE TERMS AND CONDITIONS Mar 24, 2016

This is a License Agreement between Andrea Boyd Tressler ("You") and Nature Publishing Group ("Nature Publishing Group") provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the terms and conditions provided by Nature Publishing Group, and the payment terms and conditions. All payments must be made in full to CCC. For payment instructions, please see information listed at the bottom of this form. License Number 3803830760199 License date Feb 07, 2016 Licensed content publisher Nature Publishing Group Licensed content publication Oncogene New frontiers in promoting tumour cell death: Licensed content title targeting apoptosis, necroptosis and autophagy Licensed content author J S Long and K M Ryan Licensed content date Feb 6, 2012 Volume number 31 Issue number 49

Type of Use reuse in a dissertation / thesis Requestor type academic/educational Format print and electronic Portion figures/tables/illustrations Number of figures/tables/illustrations 1 High-res required no Figures 1.2.1 Author of this NPG article no Your reference number None Mechanisms of nucleotide accumulation during Title of your thesis / dissertation regulated cell death in tumor cells Expected completion date May 2016 Estimated size (number of pages) 200

Total 0.00 USD

180

NATURE PUBLISHING GROUP LICENSE TERMS AND CONDITIONS Mar 24, 2016

This is a License Agreement between Andrea Boyd Tressler ("You") and Nature Publishing Group ("Nature Publishing Group") provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the terms and conditions provided by Nature Publishing Group, and the payment terms and conditions. All payments must be made in full to CCC. For payment instructions, please see information listed at the bottom of this form. License Number 3803760903609 License date Feb 07, 2016 Licensed content publisher Nature Publishing Group Licensed content publication Cell Death and Differentiation RIP kinases: key decision makers in cell death and Licensed content title innate immunity Licensed content author F Humphries, S Yang, B Wang and P N Moynagh Licensed content date Aug 22, 2014 Volume number 22 Issue number 2

Type of Use reuse in a dissertation / thesis Requestor type academic/educational Format print and electronic Portion figures/tables/illustrations Number of figures/tables/illustrations 1 High-res required no Figures Figure 1.2.2 Author of this NPG article no Your reference number None Mechanisms of nucleotide accumulation during Title of your thesis / dissertation regulated cell death in tumor cells Expected completion date May 2016 Estimated size (number of pages) 200

Total 0.00 USD Terms and Conditions

181

Terms and Conditions for Permissions

Nature Publishing Group hereby grants you a non-exclusive license to reproduce this material for this purpose, and for no other use,subject to the conditions below:

1. NPG warrants that it has, to the best of its knowledge, the rights to license reuse of this material. However, you should ensure that the material you are requesting is original to Nature Publishing Group and does not carry the copyright of another entity (as credited in the published version). If the credit line on any part of the material you have requested indicates that it was reprinted or adapted by NPG with permission from another source, then you should also seek permission from that source to reuse the material.

2. Permission granted free of charge for material in print is also usually granted for any electronic version of that work, provided that the material is incidental to the work as a whole and that the electronic version is essentially equivalent to, or substitutes for, the print version.Where print permission has been granted for a fee, separate permission must be obtained for any additional, electronic re-use (unless, as in the case of a full paper, this has already been accounted for during your initial request in the calculation of a print run).NB: In all cases, web-based use of full-text articles must be authorized separately through the 'Use on a Web Site' option when requesting permission.

3. Permission granted for a first edition does not apply to second and subsequent editions and for editions in other languages (except for signatories to the STM Permissions Guidelines, or where the first edition permission was granted for free).

4. Nature Publishing Group's permission must be acknowledged next to the figure, table or abstract in print. In electronic form, this acknowledgement must be visible at the same time as the figure/table/abstract, and must be hyperlinked to the journal's homepage.

5. The credit line should read: Reprinted by permission from Macmillan Publishers Ltd: [JOURNAL NAME] (reference citation), copyright (year of publication) For AOP papers, the credit line should read: Reprinted by permission from Macmillan Publishers Ltd: [JOURNAL NAME], advance online publication, day month year (doi: 10.1038/sj.[JOURNAL ACRONYM].XXXXX)

Note: For republication from the British Journal of Cancer, the following credit lines apply. Reprinted by permission from Macmillan Publishers Ltd on behalf of Cancer Research UK: [JOURNAL NAME] (reference citation), copyright (year of

182

publication)For AOP papers, the credit line should read: Reprinted by permission from Macmillan Publishers Ltd on behalf of Cancer Research UK: [JOURNAL NAME], advance online publication, day month year (doi: 10.1038/sj.[JOURNAL ACRONYM].XXXXX)

6. Adaptations of single figures do not require NPG approval. However, the adaptation should be credited as follows:

Adapted by permission from Macmillan Publishers Ltd: [JOURNAL NAME] (reference citation), copyright (year of publication)

Note: For adaptation from the British Journal of Cancer, the following credit line applies. Adapted by permission from Macmillan Publishers Ltd on behalf of Cancer Research UK: [JOURNAL NAME] (reference citation), copyright (year of publication)

7. Translations of 401 words up to a whole article require NPG approval. Please visit http://www.macmillanmedicalcommunications.com for more information.Translations of up to a 400 words do not require NPG approval. The translation should be credited as follows:

Translated by permission from Macmillan Publishers Ltd: [JOURNAL NAME] (reference citation), copyright (year of publication).

Note: For translation from the British Journal of Cancer, the following credit line applies. Translated by permission from Macmillan Publishers Ltd on behalf of Cancer Research UK: [JOURNAL NAME] (reference citation), copyright (year of publication)

We are certain that all parties will benefit from this agreement and wish you the best in the use of this material. Thank you.

Special Terms: v1.1

183

184

References

1. Littman, Dan R., Releasing the Brakes on Cancer Immunotherapy. Cell, 2015. 162(6): p. 1186-1190. 2. Coley, W.B., The treatment of inoperable sarcoma with the mixed toxins of erysipelas and bacillus prodigiosus: immediate and final results in one hundred and forty cases. JAMA, 1898(31): p. 389–395

3. Sharma, P. and J.P. Allison, The future of immune checkpoint therapy. Science, 2015. 348(6230): p. 56-61.

4. Mahoney, K.M., P.D. Rennert, and G.J. Freeman, Combination cancer immunotherapy and new immunomodulatory targets. Nature Reviews Drug Discovery, 2015. 14(8): p. 561-584. 5. Ghiringhelli, F., et al., Activation of the NLRP3 inflammasome in dendritic cells induces IL- 1beta-dependent adaptive immunity against tumors. Nat Med, 2009. 15(10): p. 1170-8. 6. Kepp, O., et al., The immunogenicity of tumor cell death. Current Opinion in Oncology, 2009. 21(1): p. 71-76. 7. Martins, I., et al., Chemotherapy induces ATP release from tumor cells. Cell Cycle, 2009. 8(22): p. 3723-8. 8. Ma, Y., et al., Anticancer Chemotherapy-Induced Intratumoral Recruitment and Differentiation of Antigen-Presenting Cells. Immunity, 2013. 38(4): p. 729-741. 9. Ma, Y., et al., ATP-dependent recruitment, survival and differentiation of dendritic cell precursors in the tumor bed after anticancer chemotherapy. Oncoimmunology, 2013. 2(6): p. e24568. 10. Elliott, M.R., et al., Nucleotides released by apoptotic cells act as a find-me signal to promote phagocytic clearance. Nature, 2009. 461(7261): p. 282-6. 11. Degterev, A., M. Boyce, and J. Yuan, A decade of caspases. Oncogene, 0000. 22(53): p. 8543-8567. 12. Weinlich, R., C.P. Dillon, and D.R. Green, Ripped to death. Trends in Cell Biology. 21(11): p. 630-637. 13. Walczak, H. and P.H. Krammer, The CD95 (APO-1/Fas) and the TRAIL (APO-2L) Apoptosis Systems. Experimental Cell Research, 2000. 256(1): p. 58-66. 14. Dickens, L.S., et al., The ‘complexities’ of life and death: Death receptor signalling platforms. Experimental Cell Research, 2012. 318(11): p. 1269-1277. 15. Pop, C., et al., Role of proteolysis in caspase-8 activation and stabilization. Biochemistry, 2007. 46(14): p. 4398-4407. 16. Irmler, M., et al., Inhibition of death receptor signals by cellular FLIP. Nature, 1997. 388(6638): p. 190-195. 17. Humphries, F., et al., RIP kinases: key decision makers in cell death and innate immunity. Cell Death and Differentiation, 2015. 22(2): p. 225-236. 18. Schinzel, A., T. Kaufmann, and C. Borner, Bcl-2 family members: intracellular targeting, membrane-insertion, and changes in subcellular localization. Biochimica Et Biophysica Acta-Molecular Cell Research, 2004. 1644(2-3): p. 95-105. 19. Youle, R.J. and A. Strasser, The BCL-2 protein family: opposing activities that mediate cell death. Nature Reviews Molecular Cell Biology, 2008. 9(1): p. 47-59.

185

20. Martinou, J.-C. and R.J. Youle, Mitochondria in Apoptosis: Bcl-2 Family Members and Mitochondrial Dynamics. Developmental Cell, 2011. 21(1): p. 92-101. 21. Tait, S.W.G. and D.R. Green, Mitochondria and cell death: outer membrane permeabilization and beyond. Nat Rev Mol Cell Biol, 2010. 11(9): p. 621-632. 22. Roos, W.P. and B. Kaina, DNA damage-induced cell death: From specific DNA lesions to the DNA damage response and apoptosis. Cancer Letters, 2013. 332(2): p. 237-248. 23. Roos, W.P. and B. Kaina, DNA damage-induced cell death by apoptosis. Trends in Molecular Medicine, 2006. 12(9): p. 440-450. 24. Mocarski, E.S., H. Guo, and W.J. Kaiser, Necroptosis: The Trojan horse in cell autonomous antiviral host defense. Virology, 2015. 479–480: p. 160-166. 25. Zhao, J., et al., Mixed lineage kinase domain-like is a key receptor interacting protein 3 downstream component of TNF-induced necrosis. Proceedings of the National Academy of Sciences of the United States of America, 2012. 109(14): p. 5322-5327. 26. Wang, Z., et al., The Mitochondrial Phosphatase PGAM5 Functions at the Convergence Point of Multiple Necrotic Death Pathways. Cell, 2012. 148(1-2): p. 228-243. 27. Murphy, J.M., et al., The Pseudokinase MLKL Mediates Necroptosis via a Molecular Switch Mechanism. Immunity, 2013. 39(3): p. 443-453. 28. Kaiser, W.J., et al., RIP3 mediates the embryonic lethality of caspase-8-deficient mice. Nature, 2011. 471(7338): p. 368-+. 29. Oberst, A., et al., Catalytic activity of the caspase-8-FLIPL complex inhibits RIPK3- dependent necrosis. Nature, 2011. 471(7338): p. 363-+. 30. Orozco, S., et al., RIPK1 both positively and negatively regulates RIPK3 oligomerization and necroptosis. Cell Death and Differentiation, 2014. 21(10): p. 1511-1521. 31. Goodman, L.S., Gilman, A., Brunton, L. L., Lazo, J. S., & Parker, K. L., Goodman & Gilman's the pharmacological basis of therapeutics. 12th ed. 2006: New York: McGraw-Hill. 32. Manns, J., et al., Triggering of a novel intrinsic apoptosis pathway by the kinase inhibitor staurosporine: activation of caspase-9 in the absence of Apaf-1. The FASEB Journal, 2011. 25(9): p. 3250-3261. 33. Bharate, S.B., et al., Kinase Inhibitors of Marine Origin. Chemical Reviews, 2013. 113(8): p. 6761-6815. 34. Burnstoc.G, PURINERGIC NERVES. Pharmacological Reviews, 1972. 24(3): p. 509-&. 35. Burnstock, G., Purinergic signalling: Therapeutic potential. Drug Development Research, 1998. 45(3-4): p. 86-92. 36. Burnstock, G., Purinergic signalling: from discovery to current developments. Experimental Physiology, 2014. 99(1): p. 16-34. 37. Tokunaga, A., et al., Involvement of SLC17A9-dependent Vesicular Exocytosis in the Mechanism of ATP Release during T Cell Activation. Journal of Biological Chemistry, 2010. 285(23): p. 17406-17416. 38. Sawada, K., et al., Identification of a vesicular nucleotide transporter. Proceedings of the National Academy of Sciences, 2008. 105(15): p. 5683-5686. 39. Zheng, Y., S.E. Gardner, and M.C.H. Clarke, Cell Death, Damage-Associated Molecular Patterns, and Sterile Inflammation in Cardiovascular Disease. Arteriosclerosis Thrombosis and Vascular Biology, 2011. 31(12): p. 2781-2786. 40. Eltzschig, H.K., et al., ATP release from activated neutrophils occurs via connexin 43 and modulates adenosine-dependent endothelial cell function. Circulation Research, 2006. 99(10): p. 1100-1108. 41. Li, A., et al., Mechanisms of ATP Release, the Enabling Step in Purinergic Dynamics. Cellular Physiology and Biochemistry, 2011. 28(6): p. 1135-1144.

186

42. Taruno, A., et al., CALHM1 ion channel mediates purinergic neurotransmission of sweet, bitter and umami tastes. Nature, 2013. 495(7440): p. 223-226. 43. Dubyak, G.R., Both sides now: multiple interactions of ATP with pannexin-1 hemichannels. Focus on "A permeant regulating its permeation pore: inhibition of pannexin 1 channels by ATP". American Journal of Physiology-Cell Physiology, 2009. 296(2): p. C235-C241. 44. Penuela, S., et al., Glycosylation Regulates Pannexin Intermixing and Cellular Localization. Molecular Biology of the Cell, 2009. 20(20): p. 4313-4323. 45. Penuela, S., R. Gehi, and D.W. Laird, The biochemistry and function of pannexin channels. Biochimica et Biophysica Acta (BBA) - Biomembranes, 2013. 1828(1): p. 15-22. 46. Gehi, R., Q. Shao, and D.W. Laird, Pathways Regulating the Trafficking and Turnover of Pannexin1 Protein and the Role of the C-terminal Domain. Journal of Biological Chemistry, 2011. 286(31): p. 27639-27653. 47. Sandilos, J.K., et al., Pannexin 1, an ATP release channel, is activated by caspase cleavage of its pore-associated C-terminal autoinhibitory region. J Biol Chem, 2012. 287(14): p. 11303-11. 48. Sandilos, J.K. and D.A. Bayliss, Physiological mechanisms for the modulation of pannexin 1 channel activity. The Journal of Physiology, 2012. 590(24): p. 6257-6266. 49. Chekeni, F.B., et al., Pannexin 1 channels mediate 'find-me' signal release and membrane permeability during apoptosis. Nature, 2010. 467(7317): p. 863-7. 50. Antonioli, L., et al., Immunity, inflammation and cancer: a leading role for adenosine. Nature Reviews Cancer, 2013. 13(12): p. 842-857. 51. Stagg, J. and M.J. Smyth, Extracellular adenosine triphosphate and adenosine in cancer. Oncogene, 2010. 29(39): p. 5346-5358. 52. Robson, S.C., J. Sevigny, and H. Zimmermann, The E-NTPDase family of ectonucleotidases: Structure function relationships and pathophysiological significance. Purinergic signalling, 2006. 2(2): p. 409-30. 53. Rosa Bono, M., et al., CD73 and CD39 ectonucleotidases in T cell differentiation: Beyond immunosuppression. Febs Letters, 2015. 589(22): p. 3454-3460. 54. Stefan, C., S. Jansen, and M. Bollen, NPP-type ectophosphodiesterases: unity in diversity. Trends in Biochemical Sciences, 2005. 30(10): p. 542-550. 55. Sträter, N., Ecto-5’-nucleotidase: Structure function relationships. Purinergic Signalling, 2006. 2(2): p. 343-350. 56. Mikhailov, A., et al., CD73 participates in cellular multiresistance program and protects against TRAIL-Induced apoptosis. Journal of Immunology, 2008. 181(1): p. 464-475. 57. Ting, A.T., F.X. PimentelMuinos, and B. Seed, RIP mediates tumor necrosis factor receptor 1 activation of NF-kappa B but not Fas/APO-1-initiated apoptosis. Embo Journal, 1996. 15(22): p. 6189-6196. 58. Morgan, M.J., Y.-S. Kim, and Z.-g. Liu, Membrane-Bound Fas Ligand Requires RIP1 for Efficient Activation of Caspase-8 within the Death-Inducing Signaling Complex. Journal of Immunology, 2009. 183(5): p. 3278-3284. 59. Franko, J.L. and A.D. Levine, Antigen-independent adhesion and cell spreading by inducible costimulator engagement inhibits T cell migration in a PI-3K-dependent manner. J Leukoc Biol, 2009. 85(3): p. 526-38. 60. Penuela, S., et al., Pannexin 1 and pannexin 3 are glycoproteins that exhibit many distinct characteristics from the connexin family of gap junction proteins. Journal of Cell Science, 2007. 120(21): p. 3772-3783.

187

61. Hampp, R., in Methods of Enzymatic Analysis. 1985, Bergmeyer, H. U., ed: Verlag Chemie, Weinheim. p. 370. 62. Qu, Y. and G.R. Dubyak, P2X7 receptors regulate multiple types of membrane trafficking responses and non-classical secretion pathways. Purinergic Signalling, 2009. 5(2): p. 163- 173. 63. Zitvogel, L., et al., Inflammasomes in carcinogenesis and anticancer immune responses. Nature Immunology, 2012. 13(4): p. 343-351. 64. Martins, I., et al., Premortem autophagy determines the immunogenicity of chemotherapy-induced cancer cell death. Autophagy, 2012. 8(3): p. 413-415. 65. Michaud, M., et al., Autophagy-dependent anticancer immune responses induced by chemotherapeutic agents in mice. Science, 2011. 334(6062): p. 1573-7. 66. Lazarowski, E.R., Mechanisms of Release of Nucleotides and Integration of Their Action as P2X- and P2Y-Receptor Activating Molecules. Molecular Pharmacology, 2003. 64(4): p. 785-795. 67. Lazarowski, E.R., Vesicular and conductive mechanisms of nucleotide release. Purinergic Signalling, 2012. 8(3): p. 359-373. 68. Wang, J. and G. Dahl, SCAM analysis of Panx1 suggests a peculiar pore structure. The Journal of General Physiology, 2010. 136(5): p. 515-527. 69. Dourado, M., E. Wong, and D.H. Hackos, Pannexin-1 Is Blocked by Its C-Terminus through a Delocalized Non-Specific Interaction Surface. PLoS One, 2014. 9(6): p. e99596. 70. Ma, W., et al., Pannexin 1 forms an anion-selective channel. Pflugers Arch, 2012. 463(4): p. 585-92. 71. Romanov, R.A., et al., The ATP permeability of pannexin 1 channels in a heterologous system and in mammalian taste cells is dispensable. J Cell Sci, 2012. 125(Pt 22): p. 5514- 23. 72. Qu, Y., et al., Pannexin-1 Is Required for ATP Release during Apoptosis but Not for Inflammasome Activation. The Journal of Immunology, 2011. 186(11): p. 6553-6561. 73. Poon, I.K., et al., Unexpected link between an antibiotic, pannexin channels and apoptosis. Nature, 2014. 507(7492): p. 329-34. 74. Martins, I., et al., Molecular mechanisms of ATP secretion during immunogenic cell death. Cell Death Differ, 2014. 21(1): p. 79-91. 75. Qiu, F. and G. Dahl, A permeant regulating its permeation pore: inhibition of pannexin 1 channels by ATP. AJP: Cell Physiology, 2008. 296(2): p. C250-C255. 76. Lévesque, S.A., et al., Specificity of the ecto-ATPase inhibitor ARL 67156 on human and mouse ectonucleotidases. British Journal of Pharmacology, 2007. 152(1): p. 141-150. 77. Bao, L., S. Locovei, and G. Dahl, Pannexin membrane channels are mechanosensitive conduits for ATP. FEBS Letters, 2004. 572(1-3): p. 65-68. 78. Locovei, S., L. Bao, and G. Dahl, Pannexin 1 in erythrocytes: Function without a gap. Proceedings of the National Academy of Sciences, 2006. 103(20): p. 7655-7659. 79. Wang, Y., et al., Autophagy-dependent ATP release from dying cells via lysosomal exocytosis. Autophagy, 2013. 9(10): p. 1624-5. 80. Petiot, A., et al., Distinct Classes of Phosphatidylinositol 3'-Kinases Are Involved in Signaling Pathways That Control Macroautophagy in HT-29 Cells. Journal of Biological Chemistry, 2000. 275(2): p. 992-998. 81. Wu, Y.T., et al., Dual Role of 3-Methyladenine in Modulation of Autophagy via Different Temporal Patterns of Inhibition on Class I and III Phosphoinositide 3-Kinase. Journal of Biological Chemistry, 2010. 285(14): p. 10850-10861.

188

82. Katagiri, K., Rap1-mediated Lymphocyte Function-associated Antigen-1 Activation by the T Cell Antigen Receptor Is Dependent on Phospholipase C- 1. Journal of Biological Chemistry, 2004. 279(12): p. 11875-11881. 83. An, S., Y. Zheng, and T. Bleu, Sphingosine 1-Phosphate-induced Cell Proliferation, Survival, and Related Signaling Events Mediated by G Protein-coupled Receptors Edg3 and Edg5. Journal of Biological Chemistry, 2000. 275(1): p. 288-296. 84. Park, H.S., et al., Proteasome inhibitor MG132-induced apoptosis via ER stress-mediated apoptotic pathway and its potentiation by protein tyrosine kinase p56lck in human Jurkat T cells. Biochemical Pharmacology, 2011. 82(9): p. 1110-1125. 85. Bassermann, F., R. Eichner, and M. Pagano, The ubiquitin proteasome system — Implications for cell cycle control and the targeted treatment of cancer. Biochimica et Biophysica Acta (BBA) - Molecular Cell Research, 2014. 1843(1): p. 150-162. 86. Suh, K.S., et al., The role of the ubiquitin proteasome system in lymphoma. Critical Reviews in Oncology/Hematology, 2013. 87(3): p. 306-322. 87. Ferrari, D., et al., Differential Regulation and ATP Requirement for Caspase-8 and Caspase-3 Activation during CD95- and Anticancer Drug-induced Apoptosis. Journal of Experimental Medicine, 1998. 188(5): p. 979-984. 88. Imao, T. and S. Nagata, Apaf-1- and Caspase-8-independent apoptosis. Cell Death Differ, 2012. 20(2): p. 343-352. 89. Ambrosi, C., et al., Pannexin1 and Pannexin2 Channels Show Quaternary Similarities to Connexons and Different Oligomerization Numbers from Each Other. Journal of Biological Chemistry, 2010. 285(32): p. 24420-24431. 90. Boassa, D., et al., Trafficking Dynamics of Glycosylated Pannexin1 Proteins. Cell Communication & Adhesion, 2008. 15(1-2): p. 119-132. 91. Bhalla-Gehi, R., et al., Pannexin1 and Pannexin3 Delivery, Cell Surface Dynamics, and Cytoskeletal Interactions. Journal of Biological Chemistry, 2010. 285(12): p. 9147-9160. 92. Bossy-Wetzel, E., Mitochondrial cytochrome c release in apoptosis occurs upstream of DEVD-specific caspase activation and independently of mitochondrial transmembrane depolarization. The EMBO Journal, 1998. 17(1): p. 37-49. 93. Eguchi, Y., S. Shimizu, and Y. Tsujimoto, Intracellular ATP levels determine cell death fate by apoptosis or necrosis. Cancer Research, 1997. 57(10): p. 1835-1840. 94. Zamaraeva, M.V., et al., Cells die with increased cytosolic ATP during apoptosis: a bioluminescence study with intracellular luciferase. Cell Death Differ, 2005. 12(11): p. 1390-1397. 95. Gleiss, B., et al., Fas-triggered phosphatidylserine exposure is modulated by intracellular ATP. FEBS Letters, 2002. 519(1-3): p. 153-158. 96. Vander Heiden, M.G., et al., Bcl-x(L) prevents cell death following growth factor withdrawal by facilitating mitochondrial ATP/ADP exchange. Molecular Cell, 1999. 3(2): p. 159-167. 97. Wang, N., et al., Paracrine signaling through plasma membrane hemichannels. Biochimica et Biophysica Acta (BBA) - Biomembranes, 2013. 1828(1): p. 35-50. 98. Yegutkin, G.G., et al., The evidence for two opposite, ATP-generating and ATP- consuming, extracellular pathways on endothelial and lymphoid cells. Biochem. J., 2002. 367(1): p. 121-128. 99. Yamaguchi, H., et al., Immunosuppression via adenosine receptor activation by adenosine monophosphate released from apoptotic cells. Elife, 2014. 3: p. e02172.

189

100. Woehrle, T., et al., Pannexin-1 hemichannel-mediated ATP release together with P2X1 and P2X4 receptors regulate T-cell activation at the immune synapse. Blood, 2010. 116(18): p. 3475-84. 101. Yip, L., et al., Autocrine regulation of T-cell activation by ATP release and P2X7 receptors. Faseb j, 2009. 23(6): p. 1685-93. 102. Siebert, A.P., et al., Structural and Functional Similarities of Calcium Homeostasis Modulator 1 (CALHM1) Ion Channel with Connexins, Pannexins, and Innexins. Journal of Biological Chemistry, 2013. 288(9): p. 6140-6153. 103. Ma, Z., et al., Calcium homeostasis modulator 1 (CALHM1) is the pore-forming subunit of an ion channel that mediates extracellular Ca2+ regulation of neuronal excitability. Proceedings of the National Academy of Sciences, 2012. 109(28): p. E1963-E1971. 104. Green, K.N. and F.M. LaFerla, Linking Calcium to Aβ and Alzheimer's Disease. Neuron, 2008. 59(2): p. 190-194. 105. Dreses-Werringloer, U., et al., A Polymorphism in CALHM1 Influences Ca2+ Homeostasis, Aβ Levels, and Alzheimer's Disease Risk. Cell, 2008. 133(7): p. 1149-1161. 106. Dubyak, G.R., Function without form: an ongoing search for maxi-anion channel proteins. Focus on "Maxi-anion channel and pannexin 1 hemichannel constitute separate pathways for swelling-induced ATP release in murine L929 fibrosarcoma cells". AJP: Cell Physiology, 2012. 303(9): p. C913-C915. 107. Islam, M.R., et al., Maxi-anion channel and pannexin 1 hemichannel constitute separate pathways for swelling-induced ATP release in murine L929 fibrosarcoma cells. AJP: Cell Physiology, 2012. 303(9): p. C924-C935. 108. Sabirov, R.Z. and Y. Okada, ATP release via anion channels. Purinergic Signal, 2005. 1(4): p. 311-28. 109. Bao, B.A., et al., Pannexin1 Drives Multicellular Aggregate Compaction via a Signaling Cascade That Remodels the Actin Cytoskeleton. Journal of Biological Chemistry, 2012. 287(11): p. 8407-8416. 110. Celetti, S.J., et al., Implications of pannexin 1 and pannexin 3 for keratinocyte differentiation. Journal of Cell Science, 2010. 123(8): p. 1363-1372. 111. Lai, C.P.K., et al., Tumor-Suppressive Effects of Pannexin 1 in C6 Glioma Cells. Cancer Research, 2007. 67(4): p. 1545-1554. 112. Litvin, O., et al., What is hidden in the pannexin treasure trove: the sneak peek and the guesswork. Journal of Cellular and Molecular Medicine, 2006. 10(3): p. 613-634. 113. Penuela, S., et al., Loss of Pannexin 1 Attenuates Melanoma Progression by Reversion to a Melanocytic Phenotype. Journal of Biological Chemistry, 2012. 287(34): p. 29184- 29193. 114. Chiu, Y.-H., K.S. Ravichandran, and D.A. Bayliss, Intrinsic properties and regulation of Pannexin 1 channel. Channels, 2014. 8(2): p. 103-109. 115. Adamson, S.E. and N. Leitinger, The role of pannexin1 in the induction and resolution of inflammation. FEBS Letters, 2014. 588(8): p. 1416-1422. 116. Dubyak, G.R., Purinergic signaling at immunological synapses. Journal of the Autonomic Nervous System, 2000. 81(1-3): p. 64-68. 117. Dubyak, G.R., P2X7 receptor regulation of non-classical secretion from immune effector cells. Cell Microbiol, 2012. 14(11): p. 1697-1706. 118. Junger, W.G., Immune cell regulation by autocrine purinergic signalling. Nat Rev Immunol, 2011. 11(3): p. 201-212. 119. Haskó, G., et al., A2B adenosine receptors in immunity and inflammation. Trends in Immunology, 2009. 30(6): p. 263-270.

190

120. Koroskenyi, K., et al., Involvement of Adenosine A2A Receptors in Engulfment-Dependent Apoptotic Cell Suppression of Inflammation. The Journal of Immunology, 2011. 186(12): p. 7144-7155. 121. Csoka, B., et al., Adenosine promotes alternative macrophage activation via A2A and A2B receptors. The FASEB Journal, 2011. 26(1): p. 376-386. 122. Yegutkin, G.G., et al., Altered purinergic signaling in CD73-deficient mice inhibits tumor progression. European Journal of Immunology, 2011. 41(5): p. 1231-1241. 123. Stagg, J., et al., Anti-CD73 antibody therapy inhibits breast tumor growth and metastasis. Proceedings of the National Academy of Sciences, 2010. 107(4): p. 1547-1552. 124. Stagg, J., et al., CD73-Deficient Mice Have Increased Antitumor Immunity and Are Resistant to Experimental Metastasis. Cancer Research, 2011. 71(8): p. 2892-2900. 125. Beavis, P.A., et al., Blockade of A 2A receptors potently suppresses the metastasis of CD73 + tumors. Proceedings of the National Academy of Sciences, 2013. 110(36): p. 14711-14716. 126. Loi, S., et al., CD73 promotes anthracycline resistance and poor prognosis in triple negative breast cancer. Proceedings of the National Academy of Sciences, 2013. 110(27): p. 11091-11096. 127. Gude, D.R., et al., Apoptosis induces expression of sphingosine kinase 1 to release sphingosine-1-phosphate as a "come-and-get-me" signal. Faseb Journal, 2008. 22(8): p. 2629-2638. 128. Peter, C., et al., Release of lysophospholipid 'find-me' signals during apoptosis requires the ATP-binding cassette transporter A1. Autoimmunity, 2012. 45(8): p. 568-573. 129. Peter, C., et al., Migration to apoptotic "Find-me" signals is mediated via the phagocyte receptor G2A. Journal of Biological Chemistry, 2008. 283(9): p. 5296-5305. 130. Vacchelli, E., et al., Chemotherapy-induced antitumor immunity requires formyl peptide receptor 1. Science, 2015. 350(6263): p. 972-978. 131. Zitvogel, L., et al., Immunological aspects of cancer chemotherapy. Nat Rev Immunol, 2008. 8(1): p. 59-73. 132. Boyd-Tressler, A., et al., Chemotherapeutic Drugs Induce ATP Release via Caspase-gated Pannexin-1 Channels and a Caspase/Pannexin-1-independent Mechanism. Journal of Biological Chemistry, 2014. 289(39): p. 27246-27263. 133. Su, L., et al., A Plug Release Mechanism for Membrane Permeation by MLKL. Structure, 2014. 22(10): p. 1489-1500. 134. Sun, L., et al., Mixed Lineage Kinase Domain-like Protein Mediates Necrosis Signaling Downstream of RIP3 Kinase. Cell, 2012. 148(1-2): p. 213-227. 135. Wang, H., et al., Mixed Lineage Kinase Domain-like Protein MLKL Causes Necrotic Membrane Disruption upon Phosphorylation by RIP3. Molecular Cell, 2014. 54(1): p. 133- 146. 136. Wang, Q., et al., Necroptotic cells release find-me signal and are engulfed without proinflammatory cytokine production. In Vitro Cellular & Developmental Biology-Animal, 2015. 51(10): p. 1033-1039. 137. Beavis, P.A., et al., CD73: a potent suppressor of antitumor immune responses. Trends in Immunology, 2012. 33(5): p. 231-237. 138. Picher, M., et al., Ecto 5 '-nucleotidase and nonspecific alkaline phosphatase - Two amp- hydrolyzing ectoenzymes with distinct roles in human airways. Journal of Biological Chemistry, 2003. 278(15): p. 13468-13479. 139. Resta, R., Y. Yamashita, and L.F. Thompson, Ecto-enzyme and signaling functions of lymphocyte CD73. Immunological Reviews, 1998. 161: p. 95-109.

191

140. Antonioli, L., et al., CD39 and CD73 in immunity and inflammation. Trends in Molecular Medicine, 2013. 19(6): p. 355-367. 141. Dzhandzhugazyan, K.N., et al., Ecto-ATP diphosphohydrolase/CD39 is overexpressed in differentiated human melanomas. Febs Letters, 1998. 430(3): p. 227-230. 142. Nevedomskaya, E., et al., A Systems Oncology Approach Identifies NT5E as a Key Metabolic Regulator in Tumor Cells and Modulator of Platinum Sensitivity. Journal of Proteome Research, 2016. 15(1): p. 280-290. 143. Laukens, B., et al., Smac Mimetic Bypasses Apoptosis Resistance in FADD- or Caspase-8- Deficient Cells by Priming for Tumor Necrosis Factor alpha-Induced Necroptosis. Neoplasia, 2011. 13(10): p. 971-U125. 144. Hanahan, D. and Robert A. Weinberg, Hallmarks of Cancer: The Next Generation. Cell, 2011. 144(5): p. 646-674. 145. Fulda, S. and D. Vucic, Targeting IAP proteins for therapeutic intervention in cancer. Nat Rev Drug Discov, 2012. 11(2): p. 109-124. 146. Quintana, A., et al., T cell activation requires mitochondrial translocation to the immunological synapse. Proceedings of the National Academy of Sciences of the United States of America, 2007. 104(36): p. 14418-14423. 147. Schenk, U., et al., Purinergic control of T cell activation by ATP released through pannexin-1 hemichannels. Sci Signal, 2008. 1(39): p. ra6. 148. Ledderose, C., et al., Mitochondria Are Gate-keepers of T Cell Function by Producing the ATP That Drives Purinergic Signaling. Journal of Biological Chemistry, 2014. 289(37): p. 25936-25945. 149. Knudson, K.M., et al., Cutting Edge: The Signals for the Generation of T Cell Memory Are Qualitatively Different Depending on TCR Ligand Strength. Journal of Immunology, 2013. 191(12): p. 5797-5801. 150. Magalhaes-Cardoso, M.T., et al., Ecto-AMP deaminase blunts the ATP-derived adenosine A(2A) receptor facilitation of acetylcholine release at rat motor nerve endings. Journal of Physiology-London, 2003. 549(2): p. 399-408. 151. Boyd-Tressler, A., et al., Chemotherapeutic drugs induce ATP release via caspase-gated pannexin-1 channels and a caspase/pannexin-1-independent mechanism. J Biol Chem, 2014. 289(39): p. 27246-63. 152. Wang, Q., et al., Necroptotic cells release find-me signal and are engulfed without proinflammatory cytokine production. In Vitro Cell Dev Biol Anim, 2015. 51(10): p. 1033- 9. 153. Pouwels, S.D., et al., Cigarette smoke-induced necroptosis and DAMP release trigger neutrophilic airway inflammation in mice. Am J Physiol Lung Cell Mol Physiol, 2016. 310(4): p. L377-86. 154. Chiu, Y.H., K.S. Ravichandran, and D.A. Bayliss, Intrinsic properties and regulation of Pannexin 1 channel. Channels (Austin), 2014. 8(2): p. 103-9. 155. Su, L., et al., A plug release mechanism for membrane permeation by MLKL. Structure, 2014. 22(10): p. 1489-500. 156. Galluzzi, L., O. Kepp, and G. Kroemer, MLKL regulates necrotic plasma membrane permeabilization. Cell Res, 2014. 24(2): p. 139-40. 157. Dondelinger, Y., et al., MLKL compromises plasma membrane integrity by binding to phosphatidylinositol phosphates. Cell Rep, 2014. 7(4): p. 971-81. 158. Lazarowski, E.R., Vesicular and conductive mechanisms of nucleotide release. Purinergic Signal, 2012. 8(3): p. 359-73.

192

159. Ricci, J.E., et al., Disruption of mitochondrial function during apoptosis is mediated by caspase cleavage of the p75 subunit of complex I of the electron transport chain. Cell, 2004. 117(6): p. 773-786. 160. Waterhouse, N.J., et al., Cytochrome c maintains mitochondrial transmembrane potential and ATP generation after outer mitochondrial membrane permeabilization during the apoptotic process. J Cell Biol, 2001. 153(2): p. 319-28. 161. Schenk, B. and S. Fulda, Reactive oxygen species regulate Smac mimetic/TNFalpha- induced necroptotic signaling and cell death. Oncogene, 2015. 34(47): p. 5796-806. 162. Chen, X., et al., Translocation of mixed lineage kinase domain-like protein to plasma membrane leads to necrotic cell death. Cell Res, 2014. 24(1): p. 105-21. 163. Cai, Z., et al., Plasma membrane translocation of trimerized MLKL protein is required for TNF-induced necroptosis. Nat Cell Biol, 2014. 16(1): p. 55-65. 164. Laukens, B., et al., Smac mimetic bypasses apoptosis resistance in FADD- or caspase-8- deficient cells by priming for tumor necrosis factor alpha-induced necroptosis. Neoplasia, 2011. 13(10): p. 971-9. 165. Sosna, J., et al., TNF-induced necroptosis and PARP-1-mediated necrosis represent distinct routes to programmed necrotic cell death. Cell Mol Life Sci, 2014. 71(2): p. 331- 48. 166. Linkermann, A., et al., Rip1 (receptor-interacting protein kinase 1) mediates necroptosis and contributes to renal ischemia/reperfusion injury. Kidney Int, 2012. 81(8): p. 751-61. 167. Morgan, M.J., Y.S. Kim, and Z.G. Liu, Membrane-bound Fas ligand requires RIP1 for efficient activation of caspase-8 within the death-inducing signaling complex. J Immunol, 2009. 183(5): p. 3278-84. 168. Manns, J., et al., Triggering of a novel intrinsic apoptosis pathway by the kinase inhibitor staurosporine: activation of caspase-9 in the absence of Apaf-1. FASEB J, 2011. 25(9): p. 3250-61. 169. Marti, G.E., et al., Phase I trial of 7-hydroxystaurosporine and fludararbine phosphate: in vivo evidence of 7-hydroxystaurosporine induced apoptosis in chronic lymphocytic leukemia. Leuk Lymphoma, 2011. 52(12): p. 2284-92. 170. Yamauchi, T., M.J. Keating, and W. Plunkett, UCN-01 (7-hydroxystaurosporine) inhibits DNA repair and increases cytotoxicity in normal lymphocytes and chronic lymphocytic leukemia lymphocytes. Mol Cancer Ther, 2002. 1(4): p. 287-94. 171. Yegutkin, G.G., Nucleotide- and nucleoside-converting ectoenzymes: Important modulators of purinergic signalling cascade. Biochim Biophys Acta, 2008. 1783(5): p. 673-94. 172. Millan, J.L., Alkaline Phosphatases : Structure, substrate specificity and functional relatedness to other members of a large superfamily of enzymes. Purinergic Signal, 2006. 2(2): p. 335-41. 173. Rybakowska, I., et al., Effect of AMP-deaminase 3 knock-out in mice on enzyme activity in heart and other organs. Nucleosides Nucleotides Nucleic Acids, 2014. 33(4-6): p. 319- 22. 174. Li, P., et al., AMP deaminase 3 plays a critical role in remote reperfusion lung injury. Biochem Biophys Res Commun, 2013. 434(1): p. 131-6. 175. Daniels, I.S., et al., AMP deaminase 3 deficiency enhanced 5'-AMP induction of hypometabolism. PLoS One, 2013. 8(9): p. e75418. 176. Cheng, J., et al., AMPD3-deficient mice exhibit increased erythrocyte ATP levels but anemia not improved due to PK deficiency. Genes Cells, 2012. 17(11): p. 913-22.

193

177. Lee, S., et al., AMPD3 is involved in anthrax LeTx-induced macrophage cell death. Protein Cell, 2011. 2(7): p. 564-72. 178. Mahnke, D.K. and R.L. Sabina, Calcium activates erythrocyte AMP deaminase [isoform E (AMPD3)] through a protein-protein interaction between calmodulin and the N-terminal domain of the AMPD3 polypeptide. Biochemistry, 2005. 44(14): p. 5551-9. 179. Qiu, F.H., et al., IMP and AMP deaminase in reperfusion injury down-regulates neutrophil recruitment. Proc Natl Acad Sci U S A, 2000. 97(8): p. 4267-72. 180. Mahnke-Zizelman, D.K. and R.L. Sabina, Cloning of human AMP deaminase isoform E cDNAs. Evidence for a third AMPD gene exhibiting alternatively spliced 5'-exons. J Biol Chem, 1992. 267(29): p. 20866-77. 181. Magalhaes-Cardoso, M.T., et al., Ecto-AMP deaminase blunts the ATP-derived adenosine A2A receptor facilitation of acetylcholine release at rat motor nerve endings. J Physiol, 2003. 549(Pt 2): p. 399-408. 182. Lohman, A.W., et al., S-Nitrosylation Inhibits Pannexin 1 Channel Function. Journal of Biological Chemistry, 2012. 287(47). 183. Shi, J., et al., Inflammatory caspases are innate immune receptors for intracellular LPS. Nature, 2014. 514(7521): p. 187-+. 184. Kayagaki, N., et al., Caspase-11 cleaves gasdermin D for non-canonical inflammasome signalling. Nature, 2015. 526(7575): p. 666-671. 185. Shi, J., et al., Cleavage of GSDMD by inflammatory caspases determines pyroptotic cell death. Nature, 2015. 526(7575): p. 660-665. 186. Juo, P., et al., FADD is required for multiple signaling events downstream of the receptor Fas. Cell Growth & Differentiation, 1999. 10(12): p. 797-804. 187. Spychala, J., A.G. Zimmermann, and B.S. Mitchell, Tissue-specific regulation of the ecto-5 '-nucleotidase promoter - Role of the cAMP response element site in mediating repression by the upstream regulatory region. Journal of Biological Chemistry, 1999. 274(32): p. 22705-22712. 188. Chalmin, F., et al., Stat3 and Gfi-1 Transcription Factors Control Th17 Cell Immunosuppressive Activity via the Regulation of Ectonucleotidase Expression. Immunity, 2012. 36(3): p. 362-373. 189. Lohman, A.W., et al., Pannexin 1 channels regulate leukocyte emigration through the venous endothelium during acute inflammation. Nature Communications, 2015. 6. 190. Lohman, A.W., et al., Complex interplay between Pannexin 1 cysteine residues and channel regulation. Faseb Journal, 2012. 26. 191. Ricci, J.E., R.A. Gottlieb, and D.R. Green, Caspase-mediated loss of mitochondrial function and generation of reactive oxygen species during apoptosis. Journal of Cell Biology, 2003. 160(1): p. 65-75.

194