Intratumoral STING Activation Sensitizes Poorly Immunogenic Cancers to Checkpoint Blockade Immunotherapy

The Texas Medical Center Library DigitalCommons@TMC

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

12-2018

Intratumoral STING Activation Sensitizes Poorly Immunogenic Cancers to Checkpoint Blockade Immunotherapy

Casey Ager

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

Part of the Medical Immunology Commons, and the Translational Medical Research Commons

Recommended Citation Ager, Casey, "Intratumoral STING Activation Sensitizes Poorly Immunogenic Cancers to Checkpoint Blockade Immunotherapy" (2018). The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences Dissertations and Theses (Open Access). 908. https://digitalcommons.library.tmc.edu/utgsbs_dissertations/908

This Dissertation (PhD) is brought to you for free and open access by the The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences at DigitalCommons@TMC. It has been accepted for inclusion in The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences Dissertations and Theses (Open Access) by an authorized administrator of DigitalCommons@TMC. For more information, please contact [email protected]. Approval Page INTRATUMORAL STING ACTIVATION SENSITIZES POORLY IMMUNOGENIC CANCERS

TO CHECKPOINT BLOCKADE IMMUNOTHERAPY

Casey Ager, B.S.

APPROVED:

______Michael A. Curran, Ph.D. Advisory Professor

______James P. Allison, Ph.D.

______Joya Chandra, Ph.D.

______Kimberly S. Schluns, Ph.D.

______Stephanie S. Watowich, Ph.D.

APPROVED:

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

Title Page

INTRATUMORAL STING ACTIVATION SENSITIZES POORLY IMMUNOGENIC

CANCERS TO CHECKPOINT BLOCKADE IMMUNOTHERAPY

DISSERTATION

Presented to the Faculty of

The University of Texas

MD Anderson Cancer Center UTHealth

Graduate School of Biomedical Sciences

in Partial Fulfillment

of the Requirements

for the Degree of

DOCTOR OF PHILOSOPHY

Casey Roy Ager, B.S.

Houston, Texas

December, 2018

Dedication

The work of this dissertation is dedicated to my father, Jeffrey Ager, whose example and

memory will forever be a guiding light in my life.

iii

Acknowledgments

The work in this dissertation could not have been completed without the guidance, encouragement, and support of an amazing collection of people whom I am incredibly fortunate to have in my life. First and foremost, the mentorship I have received during my time in the Curran lab has been instrumental in the success of this dissertation. Mike, thank you for your endless hours of hard work and the time you’ve invested in teaching me how to be an independent scientist, and for fostering a lab environment that is motivating, exciting, and always so fun. You’ve been an amazing mentor and role model, and you have provided me with many opportunities that have shaped me and accelerated my career as a scientist. Thank you as well to the rest of the Curran lab; members past and present. From 5am (or earlier!) experiments to great conversation and so much laughter in my little corner, it has been an honor, privilege, and joy to work with all of you. There’s not a better lab around to work in and not a nicer group of people to learn from. I would also like to thank the faculty and staff at the

GSBS for creating an environment in which I could grow as a student, scientist, and person.

The Deans Mike and Shelley have shown me an incredible level of personal support and encouragement since day one, and I am indebted to them for that. To past and current members of my thesis and candidacy committees, thank you for taking time out of your busy schedules to engage with my work and to guide my path through these projects. Special thanks to Dr. Sastry, who has written more recommendation letters for me than I thought could be humanly possible. To my sweet MB, thank you for being there through the twists and turns, ups and downs, and the long hours associated with this thesis. Your support means the world to me! And finally, thank you to my family. Mom, Bryan, Katie, and now little Addie; you have watched me grow from a young little nerd to an older taller nerd, and through all that life has brought us I would not be where I am or who I am without your love and support.

Abstract

INTRATUMORAL STING ACTIVATION SENSITIZES POORLY IMMUNOGENIC

CANCERS TO CHECKPOINT BLOCKADE IMMUNOTHERAPY

Casey Ager, B.S.

Advisory Professor: Michael A. Curran, Ph.D.

Unprecedented durable tumor regression can be achieved in patients with checkpoint blockade immunotherapy, yet these responses are rare and are not observed in all tumor types. Castration resistant prostate cancer (CRPC) and pancreatic adenocarcinoma (PDAC) are two common lethal cancers that remain refractory to checkpoint blockade, requiring development of new therapeutic approaches to promote responsivity in patients with these diseases. Both CRPC and PDAC are characterized by an immunosuppressive tumor milieu rich in cells of the myeloid lineage, which regulate T cell activity in many ways independent of the coinhibitory receptor axis. We investigated whether activation of acute inflammatory signaling through the cytosolic DNA sensing pathway component stimulator of interferon genes (STING) could reprogram tumor associated macrophages (TAMs) and myeloid derived suppressor cells (MDSCs) to support rather than suppress T cell activity, converting CRPC and PDAC to settings in which checkpoint blockade can be effective. To this end, we established therapeutic synergy between cyclic dinucleotide (CDN) STING agonists and checkpoint modulating agents in the multi-focal TRAMP-C2 murine model of CRPC, and described how the balance between systemic efficacy and local pathology can be balanced by modifying CDN dosing frequency. We characterized a novel STING agonist IACS-8803, which possesses potent stimulatory activity and therapeutic efficacy in vitro and in vivo relative to known natural and synthetic clinical CDN compounds. We described the phenotypic and functional effects of CDNs on suppressive MDSCs and TAMs of murine and human origin in the in vitro setting, and utilized high parameter flow cytometry to validate these changes in the tumor microenvironment upon local CDN injection. Utilizing a novel transplantable model of orthotopic PDAC, we confirm that intra-pancreatic delivery of 8803 can synergize with checkpoint blockade to significantly prolong survival in a highly refractory setting. Together, these studies lay the groundwork for translation of CDNs in combination with checkpoint blockade for patients with CRPC and PDAC.

Table of Contents

Approval Page ...... i

Title Page ...... ii

Dedication ...... iii

Acknowledgments ...... iv

Abstract ...... v

Table of Contents ...... vi

List of Illustrations ...... ix

List of Tables ...... xi

Chapter 1: Introduction ...... 1

1.1: Tumor Immunogenicity: Beyond “Hot” and “Cold” ...... 2

1.1.1: Defining Characteristics of “Hot” and “Cold” Tumors ...... 3

1.1.2: Castration Resistant Prostate Cancer (CRPC); a Quintessential “Cold” Tumor ..14

1.1.3: Pancreatic Ductal Adenocarcinoma (PDAC); a Uniquely “Cold” Tumor ...... 19

1.1.4: Summary ...... 26

1.2: Myeloid Immunosuppression: A Critical Barrier to Antitumor Immunity ...... 27

1.2.1: Ontogeny of Myeloid Derived Suppressor Cells and Tumor Associated

Macrophages ...... 28

1.2.2: Mechanisms of T Cell Immunosuppression by MDSCs and TAMs ...... 34

1.2.3: Therapeutic Targeting of MDSCs and TAMs in Cancer ...... 39

1.2.4: Summary ...... 44

1.3: STING: A Powerful Activator of Innate Immunity in the Tumor Microenvironment .....46

1.3.1: Discovery and Characterization of the cGAS-STING Pathway ...... 47

1.3.2: Role of the STING Pathway in the Antitumor Immune Response ...... 54

1.3.3: Current Knowledge of the Therapeutic Utility of STING Agonists in Cancer ...... 60

1.3.4: Summary ...... 67

Chapter 2: Methodology ...... 68

Chapter 3: Intratumoral STING Activation with T-Cell Checkpoint Modulation Generates

Systemic Antitumor Immunity ...... 74

3.1: Introduction ...... 74

3.2: Results...... 77

3.2.1: Intratumoral CDG potentiates systemic checkpoint modulation in TRAMP-C2 ...77

3.2.2: Intratumoral Flt3L at each lesion potentiates systemic checkpoint modulation in

TRAMP-C2 ...... 80

3.2.3: Intratumoral checkpoint modulation and CDG mobilizes abscopal immunity ...... 82

3.2.4: Intratumoral combination therapy with CDG elicits skin pathology ...... 85

3.2.5: Flt3L with checkpoint modulation failed to generate abscopal immunity ...... 87

3.2.6: High parameter analysis of TRAMP-C2 immune contexture in response to local

therapy ...... 92

3.2.6: Summary ...... 99

Chapter 4: Intratumoral Delivery of a Novel STING Agonist Repolarizes Suppressive Myeloid

Cells and Sensitizes Pancreatic Ductal Adenocarcinoma to Checkpoint Blockade ...... 101 vii

4.1: Introduction ...... 101

4.2: Results...... 104

4.2.1: In vitro characterization of IACS-8803, a novel highly potent STING agonist .... 104

4.2.2: In vivo characterization of 8803 relative to known CDNs ...... 108

4.2.3: Profiling effects of CDNs on murine bone marrow-derived MDSCs ...... 111

4.2.4: Profiling effects of CDNs on human M2c-polarized macrophages ...... 115

4.2.5: Combination therapy with 8803 and checkpoint blockade cures multi-focal mT4-LS

PDAC...... 118

4.2.6: Analysis of orthotopic and subcutaneous mT4 tumors following 8803 + checkpoint

therapy by flow cytometry ...... 122

4.2.7: 8803 potentiates checkpoint blockade against orthotopic mT4-LA and does not

require chemotherapy...... 125

4.2.8: Summary ...... 132

Chapter 5: Discussion ...... 133

Bibliography ...... 142

Vita ...... 217

viii

List of Illustrations

Figure 1: The “Hot” vs “Cold” tumor paradigm ...... 10

Figure 2: Models of macrophage polarization and plasticity...... 33

Figure 3: The canonical STING signaling pathway ...... 53

Figure 4: Local CDG potentiates IP checkpoint modulation against bilateral TRAMP-C2 ....78

Supplemental Figure S1: Individual tumor growth curves; IP antibody + IT agonists ...... 79

Figure 5: Local Flt3L potentiates IP checkpoint modulation against bilateral TRAMP-C2 ....81

Figure 6: Intratumoral combination CDG and checkpoint modulation mobilizes abscopal

responses against distal TRAMP-C2 ...... 83

Supplemental Figure S2: Individual tumor growth curves; IT antibody + IT agonists ...... 84

Figure 7: Intratumoral CDG with checkpoint modulation invokes dose-dependent and tumor

size-dependent ulcerative pathology at the tumor site ...... 86

Figure 8: Intratumoral combination Flt3L and checkpoint modulation does not mobilize

abscopal responses against distal TRAMP-C2 ...... 89

Supplemental Figure S3: Tumor relapse rates ...... 90

Supplemental Figure S4: Asymmetric TRAMP-C2 tumor growth kinetics ...... 91

Figure 9: Analysis of the TRAMP-C2 immune infiltrate at local and distal tumors following

intratumoral therapy at a single lesion ...... 96

Supplemental Figure S5: TIL flow analysis experimental setup and gating strategy ...... 97

Supplemental Figure S6: TIL Flow Population Quantifications ...... 98

Figure 10: In vitro characterization of IACS-8803 ...... 106

Supplemental Figure S7: Extended in vitro characterization of IACS-8803 ...... 107

Figure 11: In vivo characterization of 8803 in subcutaneous KPC-derived PDAC ...... 110

Figure 12: STING activation reduces the suppressive capacity of bone marrow-derived MDSC

...... 113

Supplemental Figure S8: MDSC subsets in unstimulated BM-MDSC cultures...... 114

Figure 13: STING activation phenotypically repolarizes human M2c macrophages ...... 117

Figure 14: Intra-pancreatic 8803 potentiates checkpoint blockade to cure mice with multi-focal mT4-LS ...... 120

Supplemental Figure S9: Mutational analysis and tumor growth kinetics of mT4 PDAC model

...... 121

Figure 15: Analysis of orthotopic and subcutaneous mT4 tumors following combination 8803

+ checkpoint blockade therapy ...... 124

Figure 16: Intra-pancreatic 8803 synergizes with dual checkpoint blockade to prolong survival in mT4-LA independent of chemotherapy ...... 129

Supplemental Figure S10: Extended data on combination 8803 and checkpoint blockade in mT4-LA PDAC ...... 131

List of Tables

There are no tables in this dissertation.

Chapter 1: Introduction

The theory of cancer immunosurveillance was first proposed by Thomas and Burnet in 1957,

and posits that the host’s immune system is capable of specifically sensing and disposing of

malignant cells bearing “non-self” genetic alterations. Despite historical records of the use of

heat-attenuated bacteria as “immunotherapy” by Coley dating back to the late 19th century,

this theory was largely rejected until two major intellectual advances in the late 20th century.

Schreiber and colleagues definitively showed that cancer could be surveilled by the immune system, and that the interplay between immune cells and tumors could endow cancer cells with immune-evasive characteristics (1). Additionally, Allison and colleagues identified a key mechanism by which tumors counteract productive immunosurveillance, and produced both preclinical and clinical evidence that modulation of this immune regulatory pathway could liberate the immune system to regress advanced cancer (2, 3). These advances ushered in a new era in oncology, with immunotherapy being elevated amongst the three traditional pillars of cancer therapy; surgery, chemotherapy, and radiation. Over the past decade, immune checkpoint blockade therapies have been evaluated and FDA approved to treat an array of cancers, and unprecedented durable responses can now be achieved for patients that were recently without any hope of a cure. However, many patients still fail to respond to current immunotherapies, and some malignancies including prostate and pancreatic cancer remain almost entirely refractory. In the following chapter, I aim to (i) provide an overview of the current understanding of why tumors respond or fail to respond to immunotherapy, with a

focus on prostate and pancreatic cancers, (ii) discuss the role of myeloid cells as mediators

of immune suppression in the tumor microenvironment, and (iii) review the stimulator of

interferon genes (STING) pathway and discuss its potential role in modulating tumor myeloid

cells and sensitizing refractory cancers to immunotherapy.

1.1: Tumor Immunogenicity: Beyond “Hot” and “Cold”

The ultimate goal of clinical oncology is to identify and prescribe for each patient the most efficacious therapeutic modality available based upon objective, observable characteristics of a patient’s disease. Cancer is not one disease but many; tumors differ not only across tissues, but tumors of the same histology can differ drastically in nature. Thus, a major goal of cancer research is to discover which molecular and cellular traits of a tumor are most informative for guiding clinical decisions and predicting how patients will respond to therapy. Traditionally, tumors are grouped according to histology, as malignancies that arise from a common tissue tend to share common traits. Within a given histology, tumors are classified according to their size and the extent to which they have metastasized. Classical

TNM staging involves characterizing a patient’s disease by the size of the primary tumor (T), the number and location of lymph nodes that contain cancer (N), and the degree to which disease has metastasized to distal regions of the body (M). Algorithms exist to guide clinicians to chemotherapeutic, surgical, or radiation interventions based upon specific TNM status.

While TNM staging is practical, accessible, and built upon decades of clinical research, our increasing understanding of the genetic basis of cancer and its interaction with the host is illuminating novel avenues for classifying tumors in specific but non-traditional ways. For example, a number of targeted therapies have been designed that attack specific oncogenic pathways found in many tumors. Thus, mutational analysis of a patient’s tumor can reveal a specific mutation such as BRAF V600E that can be directly targeted regardless of tumor histology with much greater therapeutic potential than standard TNM-guided therapy (4). The acceptance of immunotherapy at the clinical level has added even greater complexity, as a patient’s response to immune-based interventions can be related to both tumor genetics as well as the dynamic composition of the tumor immune microenvironment. Patients with the optimal combination of immune and genetic factors can achieve unprecedented durable tumor

regressions never before seen in previously-terminal cancers, however these patients remain relatively rare. Thus, there is great interest in understanding what characteristics define a “hot” tumor that is primed to respond to immunotherapy in contrast to “cold” tumors that do not respond, so that therapeutic approaches can be designed to render more tumors sensitive to therapy.

This section aims to provide a discussion of the key genetic, molecular, and cellular factors that are thought to contribute to “hot” and “cold” tumors in general, and provide a summary of two clinically “cold” tumors studied in this dissertation: castration resistant prostate cancer (CRPC) and pancreatic ductal adenocarcinoma (PDAC).

1.1.1: Defining Characteristics of “Hot” and “Cold” Tumors

Tumor Antigenicity. The human immune system is fundamentally fashioned to distinguish

“self” from “non-self.” Specifically, both innate and adaptive immune cells are trained to ignore or protect any and all self tissues, yet express receptors that together are theoretically able to recognize any possible protein epitope not found natively within the host. For example, viruses contain genetic material encoding proteins that differ in sequence and structure compared to human proteins; thus these non-self protein epitopes can serve as targets for host immune cells. Though cancer is indeed derived from native host tissue, it becomes malignant through the obligatory accumulation of genetic mutations that ultimately alter cellular function. As these mutated gene products by definition differ in protein sequence compared to their unmutated healthy forms, they can theoretically be recognized by the immune system as

“altered-self” targets, making cancer an immunologically visible disease. Therefore, the number and nature of mutations carried by a given tumor dictates its antigenicity, which in turn impacts the likelihood of an immune response being generated against it.

The idea that tumors by nature possess antigens that are targets of immune surveillance was first proposed in the mid 20th century by Burnet and Thomas (5), however not until the turn of the century were there proper experimental tools to decisively confirm the theory. Experimental confirmation of the concept of immunosurveillance (1) garnered interest in defining what antigens the immune system specifically recognizes in cancer. Early work defined many “shared” tumor antigens, which represent non-mutated molecules aberrantly overexpressed in cancer relative to normal tissue. These antigens have been studied as targets for vaccination, however endogenous immunity against them is limited due to mechanisms of central and peripheral immune tolerance that protect against autoimmune responses to self tissues. More recently, the advent of next-generation genetic sequencing together with immunogenic peptide epitope prediction algorithms has facilitated the identification of bona fide tumor-specific mutated antigens, termed neoantigens. In contrast to shared tumor antigens, neoantigens are found only in tumor cells and not in healthy tissues, thus T cell responses can be generated against them without the barrier of overcoming central tolerance. It has been clearly demonstrated in both mice and humans that neoantigens can be targeted by T cells and that checkpoint blockade immunotherapies can enhance the function of these neoantigen-specific T cells (6, 7). Therefore, there is currently significant interest in understanding how tumor mutational burden influences neoantigen quantity and quality, in order to develop clinically informative associations between tumor antigenicity and response to immunotherapy.

In theory, the more genetic mutations a tumor possesses, the greater the probability it will contain targetable neoantigens available for immune surveillance. In support of this theory, there is a compelling clinical observation that tumors which commonly arise due to carcinogen exposure tend to exhibit relatively high response rates to immunotherapy. For instance, melanoma and lung cancer – associated with exposure to UV light and cigarette smoke,

respectively – on average possess very high somatic mutation rates of around 10 mutations per megabase of DNA (8). Indeed, these cancers were the first to receive FDA approval for checkpoint blockade immunotherapies (9, 10). Interestingly, in a later study correlating tumor

mutational burden with response to immunotherapy, Yarchoan and colleagues found that the

median value of 10 coding somatic mutations per megabase generally separates tumors with

high response rate (>25%) from tumors with lower response rates (<20%)(11). While

carcinogen exposure is a common harbinger of high mutational burden in human cancer,

defects in DNA damage repair machinery can additionally contribute to “hypermutated” tumor

genomes. A small percentage of patients across all cancer types possess defects in the DNA

mismatch repair (MMR) pathway, which senses and repairs single nucleotide mutants or

minor insertions and deletions that occur during DNA replication (12). In colorectal cancer,

nearly 15% of patients carry tumors with MMR defects, resulting in a hypermutation phenotype

called microsatellite instability (MSI) (13). While colorectal cancer is historically refractory to

immunotherapy, patients with MSI exhibit response rates analogous to those seen in patients

with metastatic melanoma. In a landmark phase II study of pembrolizumab in metastatic

colorectal cancer, MSI patients exhibited a response rate of 40%, while none of the 18

microsatellite stable patients responded (14). These results were rapidly generalized to any

patients with MMR deficiency, and pembrolizumab was shown to be similarly effective

regardless of histology (15). These findings precipitated FDA approval of pembrolizumab for

any patient with MMR defects; the first ever tissue-agnostic approval of a cancer therapy

based on MMR biomarker testing. Together, these clinical observations suggest that tumors

with high mutational burden are more likely to be immunologically “hot” and possess higher

response rates to immunotherapy.

While the correlation between tumor mutational burden and response to

immunotherapy is strong, it is not perfect. For example, after melanoma and lung cancer,

kidney cancer was the third to receive FDA approval for checkpoint blockade targeting PD-1, however it is characterized by a relatively low median mutation rate of ~3 mutations per megabase (8). A growing body of evidence suggests that not all mutations equally engender relevant neoantigens; rather, the specific nature of the somatic genetic alteration may influence its relative immunogenicity. An early observation in support of this was provided by

Snyder et al, who performed whole exome sequencing on malignant and healthy tissue from

64 melanoma patients that received ipilimumab or tremelimumab. In this study, tumor mutational load was not sufficient to predict response to immunotherapy; however, a specific neoantigen “landscape” was found uniquely in responders, and constituted a strong predictive marker of response (7). The authors mapped this signature to short 4 amino acid tetrapeptide motifs found within immunogenic neoantigen peptides, which may by sequence and structure have a greater tendency to form high-affinity interactions with host T cell receptors. While the feasibility of identifying a similar neoantigen signature has proven controversial in other datasets (16), these results provide a mechanistic hypothesis to explain why neoantigens may differ in immunogenicity. A recent study by Turajlic and colleagues comprehensively probed whole-exome sequencing data from The Cancer Genome Atlas to characterize neoantigens arising from single nucleotide variants as well as small insertions and deletions (indels). The authors found that frameshifts resulting from indel alterations are nine times more likely to generate neoantigen peptides that are strong predicted binders to host MHC compared to single nucleotide variants (17). Furthermore, they determined that kidney cancer possesses the highest frequency of indel mutations, which may explain its sensitivity to checkpoint immunotherapy despite its low average tumor mutational load (8). These data further suggest that, with regard to antigenic epitopes, there may exist a spectrum of “non-selfness,” with some genetic alterations engendering more immunogenic epitopes than others. A precise mechanistic explanation for this has not yet been described in the field, however algorithms

to predict high-quality non-self neoantigens have been reported. Using a unique cohort of pancreatic cancer patients with extremely long-term survival, Balachandran and colleagues described a neoantigen quality model that predicts patient survival based upon homology between their expressed neoantigens and infectious disease-related epitopes, and demonstrated that this model predicts survival with greater accuracy than a model that associates immunogenicity with the number of tumor neoantigens (18). Based upon these key observations, it is becoming clear that the mere quantity of neoantigens is not sufficient to fully predict whether a tumor will be immunologically “hot”; rather, the quality of those antigens must also be taken into account to more accurately predict a tumor’s immunogenicity.

Tumor Microenvironment. While tumor antigens provide the necessary means of detection for an immune response against cancer, the eradication of a tumor ultimately relies on the capacity for immune cells to exert potent and specific cytolytic effector functions at the tumor site. This is no small task, as the human immune system is endowed with a wealth of regulatory mechanisms to suppress or prevent untoward damage of self tissue. In the local battle for survival that occurs upon immune recognition of cancer, it is evident that immune killing creates an evolutionary pressure that selects for outgrowth of tumor clones that summon these mechanisms to locally disable the immune response (19). As early as 1997, it was known that cancer cells bearing a targetable antigen could grow progressively while non- malignant skin grafts bearing the same antigen could be rejected in the same host (20). These observations suggested that it is tumor-mediated locoregional control of immune response, rather than systemic immune dysfunction, that allows for outgrowth of antigenic tumors.

Therefore, it follows that an understanding of the relevant components of the tumor immune microenvironment (TME) and how they interact can aid in predicting whether a tumor will

respond or is actively responding to immunotherapy.

In most cases, cytotoxic CD8 T cells are the indispensable mediators of immune-

mediated tumor rejection, thus the prevalence of CD8 T cells in a tumor ought to have

prognostic or predictive value in cancer. Early studies quantifying T cell numbers in colorectal

tumor biopsies found that a greater density of infiltrating CD8 T cells correlates with both

earlier stage disease and the absence of metastatic invasion (21, 22). Incredibly, a simple

histopathological algorithm termed the Immunoscore which enumerates the total number of

CD3+ and CD8+ T cells in the center and periphery of a tumor has high prognostic value in

these patients, and outperforms both TNM staging and MSI status in predicting recurrence as

well as survival in colorectal cancer (23, 24). While the Immunoscore remains to be validated

in other histologies, independent reports have associated CD8 T cell infiltration with tumor

stage, recurrence rate, or survival in melanoma (25, 26), ovarian cancer (27, 28), head and

neck cancer (29), liver cancer (30), bladder cancer (31), and gastrointestinal stromal tumors

(GIST) (32). Therefore, a high density of CD8 T cells within the tumor core is a generalizable

characteristic of immunologically “hot” tumors across many human cancers.

Given the critical import of CD8 T cells in controlling cancer progression and lethality,

it is important to understand how and why tumors become T cell-infiltrated. Interestingly, a

recent report by Spranger et al suggests that the prevalence of CD8 T cells in melanoma is

not related to the number of nonsynonymous somatic mutations in a given tumor, and these

findings could be generalized across solid tumors in the TCGA database (33). These data

suggest that the majority of solid tumors likely possess sufficient neoantigen burden to support

a T cell response, thus the capacity or failure to achieve CD8 T cell infiltration is chiefly

determined by molecular and cellular characteristics of the tumor microenvironment. Gene expression analysis of human tumors bearing differing levels of T cell infiltration have hinted that a foundational determinant underlying the presence of CD8 T cells is the host type I interferon response (34). Historically associated with antiviral immunity, type I interferons in

cancer are critical for facilitating myeloid activation in the sterile tumor microenvironment.

Specifically, IFN-β released in the microenvironment acts upon Batf3-lineage dendritic cells

(DCs), and signals through the IFN-α/βR and Stat1 to induce a transcriptional program associated with antiviral antigen presentation (35). These DCs, characterized by expression of the surface markers CD8α or CD103 in mice and CD141 in humans (36), are capable of

tumor antigen cross presentation, which allows for priming of CD8 T cells against tumor

antigens exogenously acquired by tumor DCs. Moreover, type I interferons trigger release of chemokines by myeloid cells including CCL5 and CXCL10 that are required to recruit CD8 T cells to the tumor bed (37, 38). Mice deficient in type I interferon signaling or Batf3-lineage

DCs thus exhibit deficiencies in T cell infiltration and spontaneous immunity against otherwise immunogenic tumors (34, 38-40). Therefore, the presence of type I interferon signaling and

Batf3-lineage DCs capable of tumor antigen cross presentation are critical determinants of immunologically “hot” tumors bearing CD8 T cell infiltration.

If priming and recruitment of tumor antigen-specific CD8 T cells are necessary

characteristics of immunogenic “hot” tumors, then it follows that any factors that interfere with

T cell priming, recruitment, or cytolytic function at the microenvironment can contribute to

poorly immunogenic “cold” tumors. Again, given the evolutionary emphasis on immune

regulation to avoid autoimmunity, such mechanisms are numerous and diverse in the human

immune system. Broadly, these immunoregulatory influences can be conceptually grouped

into three categories; immune suppression, immune exclusion, and immune ignorance (Fig

1).

Suppression of CD8 T cell cytolytic function can be achieved by a number of cell types

often present in the tumor microenvironment, and occurs through modulation of T cell

signaling, polarization, and metabolic fitness. After activation, T cells will begin to express an

array of coinhibitory receptors including CTLA-4, PD-1, LAG-3, Tim-3, and TIGIT, among 9

Figure 1: The “Hot” vs “Cold” tumor paradigm. Tumor immunogenicity is broadly determined by tumor antigen burden and the composition and quality of the tumor immune microenvironment. Represented are human solid tumors ranked in rough order of relative immunogenicity, as well as the multiple mechanisms by which tumors can be immunologically “cold.” It should be emphasized that multiple mechanisms can be present and relevant at the level of a given tumor type as well as within an individual patient’s tumor.

others. Ligation of these receptors generally suppresses signaling downstream of the T cell

receptor (TCR) and costimulatory receptors, limiting T cell function and longevity (41).

Specifically, PD-1 is thought to be a dominant coinhibitory factor for T cells within the tumor

microenvironment (42). PD-1 signals through SHP1 and SHP2 phosphatases to dephosphorylate and inactivate ITAM motifs on CD28, and additionally removes phosphate groups that are required for the function of downstream TCR signaling proteins such as ZAP70 and PI3K (43-45). Expression of the ligands for PD-1 (PD-L1 and PD-L2) can be found on

both myeloid cells infiltrating tumors (46), and on tumor cells themselves, and can be further

induced on these cells in response to T cell-derived IFN-γ (47, 48). These observations

provide rationale for the blockade of T cell coinhibitory molecules in cancer, an approach

which has revolutionized clinical management of cancer over the past decade. However, a

number of other T cell suppressive mechanisms exist in the TME that occur independently of

coinhibitory checkpoint receptors. Many of these are the product of immune cells that possess

an anti-inflammatory influence in cancer, including FoxP3+ regulatory T cells (Tregs), myeloid derived suppressor cells (MDSCs), and tumor associated macrophages (TAMs). Three critical immunosuppressive mechanisms include the presence of suppressive cytokines that limit inflammatory T cell function and longevity, the production of reactive oxygen and nitrogen species that have toxic effects on cytotoxic T cells, and the enzymatic catalysis of T cell- essential nutrients. As suppressive myeloid cells are key contributors to these methods of T cell inhibition in the TME, the above mechanisms will be reviewed in detail in the next chapter.

Overall, an increased density of coinhibitory ligands in the TME together with the coordinated actions of Tregs, MDSCs, and TAMs can render the TME a hostile environment for the survival and functional fitness of CD8 T cells.

Immune exclusion involves the construction of physical or metabolic barriers that preclude access of CD8 T cells to some or all of the tumor bed. For circulating T cells primed

in secondary lymphoid organs against tumor antigens, neovascular tumor endothelium

presents the first barrier to T cell entry. Whereas healthy endothelium facilitates T cell

extravasation into tissues through clustered expression of adhesion molecules such as ICAM1

and VCAM1, aberrant tumor neovasculature often displays reduced expression of these

proteins (49) and aberrant expression of inhibitory ligands such as PD-L1 and PD-L2 (50, 51), anti-adhesive molecules such as endothelin B receptor (52), or apoptosis-inducing ligands

such as TRAIL and FasL (53). Thus, the tumor vasculature can suppress T cell infiltration into

tumors by promoting anergy or outright apoptosis in circulating T cells. If extravasation through the neovasculature is permitted, T cells often encounter stromal barriers at the tumor site.

Mesenchymal fibroblasts have been shown to be immunosuppressive in many tumors (54), and possess the capacity to lay down thick webs of extracellular matrix composed of collagen, fibrinogen, laminin, and other glycoproteins. The dense patterns of ECM present in many tumors can either block T cell entry or interfere with chemotaxis, thus inhibiting T cell entry into the tumor bed (55, 56). Blockade of chemokine signals that recruit cancer-associated fibroblasts (CAFs) can boost T cell access and sensitize tumors with dense stroma to immune- based therapies (57), however outright depletion of CAFs has had mixed results in mouse and man, with some reports suggesting stromal depletion can accelerate tumor progression and metastasis (58). Absent stromal barriers to T cell chemotaxis, many tumors possess regions which are metabolically incompatible with T cell survival. Segments of a tumor that are spatially separated from sufficient vasculature can be deeply hypoxic, and limited access to oxygen promotes immunosuppressive phenotypes in both T cells and myeloid cells through activation of the hypoxia-responsive signaling protein Hif1α (59-61). Elimination of intratumoral hypoxia through hypoxia-activated pro-drug therapy or respiratory hyperoxemia can restore T cell infiltration and sensitize tumors to immunotherapy (62, 63). Therefore, exclusion of T cells from the tumor via non-permissive neovasculature, dense ECM

deposition, or intratumoral hypoxia can contribute significantly to a poorly-immunogenic, “cold”

tumor phenotype.

Immune ignorance chiefly involves a failure of the innate immune system to sense the

presence of a tumor, as innate antigen presenting cells such as DCs are required to prime a

T cell response against cancer. In this context, tumors with exceptionally low mutational

burden may possess an insufficient number of quality neoantigens to contain an epitope

capable of being targeted by host T cells. While this has been postulated to contribute to the

poorly immunogenic nature of genetically engineered murine models (GEMMs) (64), it is

unclear whether this is a relevant mechanism for human tumors that naturally develop over

long periods of time, where chromosomal instability makes the accumulation of some number

of passenger mutations essentially inevitable. An alternative explanation is that host DCs may

be shielded from tumors, denying them access to tumor antigens to utilize for T cell priming.

An interesting mechanism for this was provided by Spranger and colleagues, who found that

human melanoma tumors that lack T cell infiltration exhibit activation of the WNT/β-catenin

signaling pathway. Analysis of the effects of β-catenin signaling revealed an inhibition of the

cytokine CCL4, which is utilized to recruit Batf3-lineage DCs to the tumor microenvironment

(65). Thus, oncogenic signaling through the WNT/β-catenin pathway can further promote

tumor progression by precluding recruitment of DC populations required to mount antitumor

T cell immunity. As this pathway is active in other cancers beyond melanoma, it may constitute

a generalizable mechanism of immune ignorance across histologies (66).

In summary, the success or failure of an antitumor immune response is governed in

large part by the dynamic composition and character of the tumor microenvironment.

Immunogenic “hot” tumors tend to be characterized by robust infiltration of CD8 T cells which

have been primed by Batf3-lineage DCs activated by type I interferon to mediate tumor antigen cross-presentation. In contrast, poorly immunogenic “cold” tumors are either devoid 13

of T cells through immune ignorance or immune exclusion, or possess functionally inept T

cells through the concerted actions of immunosuppressive cells and coinhibitory ligands within the tumor microenvironment. In the future, coordinate analysis of both microenvironmental immune factors together with tumor antigen burden may be a unique method to comprehensively approximate tumor immunogenicity and provide powerful prognostic and predictive value (67).

1.1.2: Castration Resistant Prostate Cancer (CRPC); a Quintessential “Cold” Tumor

Biology and Clinical Management of CRPC. Prostate cancer is the most commonly

diagnosed cancer in men, and is one of the leading causes of male death in western countries

(68). Of the 164,690 men predicted to be diagnosed with prostate cancer this year, nearly

80% will present with localized disease (68), which is relatively indolent and associated with

good prognosis. Active surveillance through PSA screenings and MRI screening – in place of

traditional interventions including radical prostatectomy, radiotherapy, or brachytherapy – is

sufficient to allow for a 1% risk of death over 10 years for these low risk patients (69, 70).

However, around 35% of prostate cancer patients either treated or monitored for localized

disease will eventually exhibit increases in serum PSA levels associated with disease

progression, and require further therapy (71). Unfortunately, serum PSA is the only

standardized method for monitoring prostate tumor progression, as imaging techniques of the

organ are insufficient to provide definitive diagnoses. The first-line intervention for advanced

prostate cancer patients is surgical or chemical castration through use of antiandrogens such

as bicalutamide. While temporary control of PSA is common with this therapy, eventual

progression to metastatic castration-resistant disease is inevitable given sufficient time. In

contrast to localized disease, metastatic castration resistant prostate cancer (mCRPC) is an

incurable disease with a 5-year survival rate of 29% (68). Most commonly, these patients see metastatic spread to bone or distal lymph nodes, though involvement of visceral organs such as the liver can occur and is associated with more aggressive disease (72, 73). Traditionally, chemotherapy with docetaxel is the standard of care for these patients (74, 75), however in the last decade a number of novel therapies have been successfully evaluated in large randomized studies including an autologous vaccine sipuleucel-T (76), hormonal agents abiraterone (77, 78) and enzalutamide (79-81), the novel taxane cabazitaxel (82, 83), and a bone metastasis targeting radionuclide radium-223 (84, 85).

Genetic and molecular classification of mCRPC has proven a difficult task, due to the rare but complex genetic alterations associated with the disease. The mCRPC genome generally exhibits a low mutation rate of 2 mutations per megabase, with common mutations targeting tumor suppressor genes TP53, PTEN, and SPOP, as well as components of the androgen receptor (AR) signaling pathway (86, 87). Copy number alterations are also relatively infrequent, but target common pathways such as PTEN, MYC, and AR (88). A characteristic quality of advanced prostate cancer is the presence of complex coordinated chromosomal translocations, termed “chromoplexy” to represent a rewiring of the genome

(89). Specifically, mCRPC contains frequent gene fusions involving members of the ETS transcription factor family, which often fuse with components of the AR signaling pathway such as TMPRSS2 (90, 91). This has led to an effort to classify tumors based upon the presence of ETS fusions, as common mutations in genes such as PTEN, TP53, SPOP, or SPINK1 tend to be enriched in either ETS-positive or ETS-negative tumors (86, 92-94). Unfortunately, these common alterations are not amenable to available targeted therapeutic approaches, rendering personalized therapy using clinical genomics currently untenable for patients with mCRPC.

Immunotherapy for CRPC. Inflammation is linked to prostate carcinogenesis, suggesting the

host immune system interacts with and plays a role in the initiation and progression of CRPC

(95, 96). To date, the most commonly evaluated and only successful immune-based

interventions in CRPC are therapeutic vaccines. The first therapeutic vaccine created for

prostate cancer was GVAX, which initially consisted of autologous irradiated prostate cancer

cells retrovirally engineered to produce GM-CSF (97), though for large randomized studies

allogeneic cell lines LNCaP and PC3 were utilized in place of autologous tissue (98, 99).

Unfortunately, two phase III studies of GVAX for CRPC were halted early due to lack of

efficacy in the treatment arms (100), preventing clinical approval of GVAX as a single agent

therapy. An alternative vaccine, Sipuleucel-T, is a live cell-based vaccine consisting of

autologous monocytic DCs differentiated with a fusion peptide containing GM-CSF and a

prostate antigen prostatic acid phosphatase (PAP). In a phase III study of men with mCRPC,

Sipuleucel-T extended survival 4.1 months compared to placebo, despite having no significant

effect on PSA or time to disease progression (76). Based on these results, Sipuleucel-T

received FDA approval for mCRPC in 2010. However, some controversy clouds these data,

as patients age 65 or older in the placebo arm of the trial appear to have exhibited

uncharacteristically short survival relative to historical controls, which may have skewed the

survival data in favor of the Sipuleucel-T arm (101). Additionally, the cellular product infused

into patients in the placebo arm was kept without supportive cytokine at 2-8oC during the culture period in which GM-CSF-PAP was added to the cultured monocytic DCs in the treatment arm, thus the quality (or lack thereof) of the “placebo” infusion product has been questioned (101). Despite these allegations, Sipuleucel-T remains the only FDA-approved immunotherapy for prostate cancer patients. A number of other vaccines have reached late- phase clinical investigation, including a poxvirus-based vaccine named ProstVac-VF (102). A large phase III study of this vaccine was recently halted however due to futility upon interim analysis (103). Together, vaccine approaches have been rigorously evaluated clinically in

prostate cancer but to date have demonstrated little or marginal benefit in comparison to

chemotherapy or androgen modulating therapies.

Given the unprecedented clinical success of monoclonal antibodies targeting T cell

coinhibitory checkpoints CTLA-4 and PD-1 in metastatic melanoma, lung cancer, and numerous other tumor types, there has been significant interest in evaluating checkpoint blockade immunotherapy in CRPC. To date, two phase III studies have been completed

evaluating ipilimumab in men with mCRPC. The first trial evaluated ipilimumab in combination

with low-dose radiation versus radiation and placebo (104), while the second trial tested ipilimumab or placebo in earlier-stage chemotherapy-naïve patients (105). Neither study

demonstrated a significant enhancement in survival in the treatment arms, thus ipilimumab

has not received FDA approval for CRPC. However, close analysis of the data illuminates

trending advantages in either overall survival or progression-free survival in the respective

studies, suggesting ipilimumab as a single agent may possesses marginal, non-negligible

efficacy in prostate cancer. In light of this, ipilimumab is currently being evaluated in combinatorial regimens together with vaccines such as Sipuleucel-T (106) or ProstVac-VF

(NCT02506114), or with androgen deprivation therapies including leuprolide, goserelin, or degarelix (NCT01377389). Blockade of PD-1 is additionally being evaluated in CRPC, based in part upon the observation that approximately 32% of mCRPC tumors test positive for PD-

L1 expression (107). Preliminary data from the first large-scale evaluation of pembrolizumab in patients with docetaxel-refractory CRPC show a meager response rate of 9% in patients with RECIST-measurable disease (108). However, higher response rates are observed in patients with mutations in DNA repair genes such as BRCA1/2 or ATM, indicating biomarker evaluation for mutations in DNA repair genes, MMR genes, or MSI status can be beneficial in finding rare patients capable of responding to PD-1, as seen in other tumor types. Fortunately, a relatively high proportion of patients (~12%) are believed to bear this hypermutated

phenotype in CRPC (109). Additionally, PD-1 blocking agents are being utilized in the combination setting with vaccines (110), androgen deprivation therapy (111), or together with

CTLA-4 blockade (112). It remains to be seen whether these combinatorial regimens will prove successful in a setting where single agent immunotherapy has largely failed, or whether altogether novel approaches are required to benefit patients with CRPC. One such novel approach may be targeting of prostate tumor myeloid cells such as TAMs or MDSCs. A critical relationship has recently been established between MDSCs and CRPC cells, wherein MDSCs secrete factors such as IL-23 to promote AR-signaling and tumor survival even in the presence of antiandrogen therapies (113), while additionally facilitating suppression of cytotoxic T cell responses. As such, the use of myeloid-targeting therapies appears to sensitize prostate tumors to checkpoint blockade (114, 115). A critical question for the field going forward – addressed in more detail later in this dissertation – is how to most effectively modulate tumor myeloid cells to optimally synergize with checkpoint blockade, and whether these approaches will translate into effective clinical management of human CRPC.

Together, prostate cancer is a largely manageable disease at early stages, however progression to its metastatic, castration-resistant form results in incurable disease. Thus far, greater understanding of the complex genomic landscape in CRPC has not translated to effective targeted therapy. Moreover, the first generation of clinical immune-based interventions have provided either modest or negligible benefit, resulting in the classification of CRPC as an immunologically cold tumor. Future success in this disease requires better prognostic and predictive biomarkers to select patients likely to respond to available treatments, more rationally-designed combination approaches, and development of novel therapeutic regimens aimed at targeting tumor myeloid cells to convert the prostate tumor microenvironment from “cold” to “hot.”

1.1.3: Pancreatic Ductal Adenocarcinoma (PDAC); a Uniquely “Cold” Tumor

Biology and Clinical Management of PDAC. Pancreatic ductal adenocarcinoma (PDAC) is perhaps the most lethal of all common cancers, with a number of defining characteristics that act in concert to produce a highly metastatic, largely uncontrollable disease. Symptoms associated with early-stage localized pancreatic tumors – such as stomach pain, appetite loss, or back pain – are diffuse and non-specific, thus early detection of PDAC is rare. Even for high-risk patients with hereditary predisposition to PDAC, there remains no accepted screening method due to a lack of sensitivity and specificity in traditional techniques including imaging or quantification of serum tumor markers such as CA19-9 (116, 117). As a result, only

10% of PDAC patients are diagnosed with localized disease, whereas the overwhelming majority present with metastatic tumor burden (68). Surgery is the only therapy with curative potential in PDAC, however it is reserved for the minority of patients with resectable or borderline resectable disease, and even among these early-stage surgical candidates the rate of relapse is nearly 90% and 5-year overall survival is low (~20%) (118). For patients with unresectable PDAC, chemotherapy is the current standard of care. Gemcitabine has been the primary chemotherapeutic for PDAC since the late 1990s, when it demonstrated superiority over fluorouracil in extending median overall survival from 4.4 to 5.6 months (119). More recently, two unique combinatorial regimens have been approved for use based upon superiority over gemcitabine in randomized phase III studies; the first a multi-drug cocktail consisting of fluorouracil, oxaliplatin, leucovorin, and irinotecan (FOLFIRINOX) (120), and the second a two-drug combination consisting of gemcitabine and albumin-bound paclitaxel (nab- paclitaxel) (121). While these two regimens have not been evaluated in a head-to-head clinical study, it appears that FOLFIRNOX may be marginally more effective, as Conroy et al report a median overall survival of 11.1 months for patients receiving FOLFIRINOX, versus 8.5 months in patients receiving gemcitabine and nab-paclitaxel in the study reported by Von Hoff 19

et al. However, FOLFIRINOX is substantially more toxic, thus clinically it is generally reserved

for patients under 76 years of age with excellent performance status (122). Despite these

recent advances, the long-term outlook for patients with PDAC remains quite poor, as the

overall median survival rate remains 8 months. This grim reality necessitates the design of

novel interventions based upon the unique genetic and microenvironmental characteristics of

the disease.

One of the central defining characteristics of PDAC is the presence of activating

mutations in the KRAS oncogene. Analysis of the earliest detectable PDAC precursor lesions,

termed pancreatic intraepithelial neoplasias (PanINs), show that KRAS is mutated in more

than 90% of these lesions (123), suggesting oncogenic KRAS is a critical initiator of PDAC

tumorigenesis. Additional common mutations found in 50-80% of patients occur in TP53,

CDKN2A, or SMAD4 tumor suppressors. These mutations are often found in later-stage

PanINs, suggesting they occur after initial KRAS activation and are necessary for progression

into PDAC (124-126). This concept has been corroborated with genetically engineered mouse

models bearing pancreas-specific expression of oncogenic KRAS in the presence or absence

of mutated or null alleles of TP53, Ink4a/CDKN2A, or SMAD4. Mice expressing a constitutively

active KRASG12D allele develop PanIN lesions, but only rarely progress to invasive metastatic

PDAC (127). Concurrent expression of KRASG12D and inactivated TP53, Ink4a/CDKN2A, or

SMAD4 results in rapid formation of invasive PDAC with complete penetrance (124, 128, 129).

In contrast, enforced expression of mutated TP53, Ink4a/CDKN2A, or SMAD4 in the absence

of KRASG12D does not lead to observable disease (125, 130). Thus, oncogenic KRAS is

required for initiation of PDAC precursor lesions, which then acquire additional loss of TP53,

Ink4a/CDKN2A, or SMAD4 to progress to invasive PDAC in mice. Given the critical role of these four genes in PDAC tumorigenesis and their presence in essentially all PDAC patients, approaches to target them specifically would be ideal. Unfortunately, no therapies currently

exist to restore function of mutated TP53, CDKN2A, or SMAD4, and oncogenic KRAS retains

the fateful designation “undruggable” (131). Thus, targeted therapy is not a tenable option for

PDAC patients at this time.

Extensive desmoplasia is an additional defining histological characteristic of PDAC

tumors that promotes tumor progression and drug resistance. It has been estimated that

malignant cells compose only 10% of the PDAC tumor microenvironment, the rest consisting

of both cellular and non-cellular constituents. The non-cellular extracellular matrix found

densely deposited within PDAC lesions contains proteins such as collagen, fibronectin, and

laminin, in addition to numerous and diverse glycosylated proteins. It is thought that a

population of pancreas-resident myofibroblasts termed pancreatic stellate cells (PSCs) are largely responsible for the generation of this proteinaceous milieu (122, 132). Malignant PDAC cells participate in crosstalk with PSCs through production of cytokines such as TGF-β, FGF,

HGF, EGF, and PDGF (133, 134), which stimulate stromal matrix deposition by PSCs (135).

This in turn promotes tumor progression, creating a positive feedback loop accelerating matrix deposition and tumor invasion (136). In addition, dense desmoplasia restricts the tumor vasculature, resulting in a hypovascular state in which diseased pancreas in mouse and man possesses a mere 25% of the average vessel density found in normal tissue (137). This functionally limits drug perfusion into the tumor bed (137), and further results in the formation of deeply hypoxic regions within the tumor tissue (138). Not only does hypoxia itself feedback to stimulate a stronger desmoplastic reaction (139), it also promotes tumor migration and metastasis (140). Oncogenic KRAS supports metabolic adaptations by PDAC tumor cells to survive within these hypoxic zones, promoting Warburg-like anaerobic glycolysis (141), increased macropinocytotic uptake of environmental nutrients (142), autophagy (143), and upregulation of Nrf2 to maintain cellular redox states (144). In contrast, hypoxia blunts the antitumor immune response, creating sites of immune privilege inaccessible to cytotoxic T

cells that are metabolically sensitive to oxygen deprivation (63), while fortifying

immunosuppressive phenotypes in tumor myeloid populations (60, 145). Therefore,

adaptation to hypoxia in highly desmoplastic, hypovascular tumor regions affords PDAC cells

a safe haven from cytotoxic therapies and immune cells, resulting in drug resistance and rapid

progression to metastatic disease.

Finally, hematopoietic cells are an additional critical component of the PDAC tumor

stroma that largely contribute to – rather than contradict – its designation as an immunologically “cold” tumor. Characterization of the tumor-infiltrating immune milieu in GEM

models of PDAC such as the Kras+/G12DTP53+/R172HPdx1-Cre (KPC) mouse has profoundly

advanced our knowledge of the dynamic interactions between the host immune system and

PDAC from early stages through to metastatic disease. A landmark study by Clark and

colleagues described the tendency of PDAC tumors to recruit CD45+ immune cells even at pre-malignant PanIN stages, which further increases upon progression to invasive PDAC

(146). This immune infiltrate is largely composed of classically immunosuppressive cell types

including Tregs, TAMs, and MDSCs. Interestingly, PanIN lesions efficiently recruit

macrophages, while MDSCs accumulate in later stage disease. The authors describe a

striking mutually exclusive relationship between CD11b+Gr-1+ MDSC and CD3+CD8+ T cells in PDAC tumors, suggesting suppressive MDSCs may contribute to immune exclusion in

PDAC (146). Recruitment of immunosuppressive MDSC can be attributed largely to tumor-

intrinsic oncogenic KRAS signaling, which causes PDAC cells to secrete cytokines and growth

factors that mobilize and polarize MDSC including GM-CSF (147, 148) and IL-6 (149). In

support of this, blockade of tumor GM-CSF production significantly reduces MDSC

accumulation, restoring CD8 infiltration into the TME (147, 148). Macrophages also exhibit a

mutually exclusive relationship with CD8 T cells in PDAC (146), and have in some instances

been shown to exclude T cell infiltration in a way resembling the activity of cancer associated

fibroblasts (150). Macrophages are thought to arise from two distinct sources in PDAC: circulating monocytes that are recruited to the TME and differentiate into TAMs in situ, or from a pool of embryonically-derived tissue resident macrophages that possess self-renewing capacity (151). While macrophages are known to possess both immunosuppressive and pro- tumorigenic roles in PDAC (152), a recent report by Zhu et al suggests that these dual roles are mediated by individual TAM compartments, rather than by monocyte-derived TAMs alone as previously thought (153, 154). Specifically, monocyte-derived TAMs in PDAC exhibit a gene expression profile in line with immunosuppression, while embryonically-derived resident macrophages predominantly contribute matrix remodeling factors to facilitate local desmoplasia (151). While further studies are required to thoroughly unravel the complex biology of macrophage plasticity and function in PDAC, it is clear that TAMs and MDSCs together contribute significantly to PDAC tumor progression, local immunosuppression, and therapeutic resistance. Therefore, the PDAC myeloid compartment is a major target for development of novel therapeutics.

Immunotherapy for PDAC. To date, traditional immunotherapy has not proven successful in clinical management of PDAC, however a number of novel combinatorial approaches are on the horizon supported by compelling preclinical evidence of efficacy. As in prostate cancer, the first immunotherapies for PDAC consisted of cancer vaccines, including GVAX, an allogeneic irradiated whole cell vaccine expressing GM-CSF (155). GVAX has progressed to phase II testing in combination with other regimens, including chemoradiation in the adjuvant setting (156), low-dose cyclophosphamide (157), or cyclophosphamide plus CRS-207, a live- attenuated Listeria monocytogenes based vaccine (158). Though these studies report high tolerability and generation of mesothelin tumor antigen-specific CD8 T cells, median survival rates remain at 6 months for advanced patients (158). Three other vaccines have been

evaluated unsuccessfully in randomized phase III trials: peptide vaccines targeting telomerase

(159) or VEGFR2 (160), or an allogeneic irradiated whole cell vaccine expressing α1,3-

galactosyltransferase (161). While these trials additionally report generation of CD8 T cells

against vaccine antigens, no significant clinical benefit has been observed, suggesting

significant barriers remain preventing cytotoxic regression of PDAC tumors.

Checkpoint blockade has been tested in small phase I studies alone or in combination

with chemotherapy, however meager results have prevented progression to large-scale trials.

Ipilimumab was first tested in a phase II study of 27 patients with metastatic or locally

advanced PDAC, with no patients exhibiting disease response by RECIST criteria (162).

Similarly, no responses were seen in 14 PDAC patients treated with a PD-L1 blocking

antibody in a phase I study (163). Recent trials have evaluated checkpoint blockade in

combination with chemotherapy, including two phase I studies combining CTLA-4 blocking

antibodies with gemcitabine (164, 165). While temporary stable disease was achieved in a

number of patients in these studies, the observed RECIST response rates did not exceed

historical controls for gemcitabine monotherapy. Interestingly, a recent phase I study tested nivolumab in combination with gemcitabine and nab-paclitaxel, and reported partial responses or stable disease in 12 of 17 patients (166). While not definitive given the small sample size, these results may indicate efficacy by checkpoint blockade agents when utilized in potent combinatorial regimens. Additionally, rare PDAC patients with MMRd have been shown to respond favorably to PD-1 blockade (15), further suggesting checkpoint blockade can be of

benefit in this disease given optimal antigenic conditions.

A new generation of immunotherapeutic agents are currently in early-stage testing for

PDAC based on rational targeting of key elements in the PDAC tumor stroma. Many of these approaches target myeloid cells, due to their high prevalence in PDAC tumors and potent immunosuppressive abilities. As such, these myeloid-based therapies will be reviewed in the

next chapter of this dissertation. In addition to myeloid cells, however, cancer associated

fibroblasts (CAFs) and their matrix-depositing functions are being targeted in PDAC, in large

part because of their immunosuppressive and immune excluding effects. While outright

depletion of CAFs has proven controversial, with some studies reporting tumor progression

upon depletion whereas others report tumor control (58, 137, 167), Feig and colleagues found

that FAP+ CAFs produce large amounts of the chemokine CXCL12, which is associated with

T cell exclusion from the PDAC tumor bed (57). Blockade of CXCR4, the receptor for CXCL12, results in increased T cell infiltration and sensitization of KPC tumors to checkpoint blockade.

It is currently unclear why blockade of this chemokine results in paradoxical T cell recruitment, however there are reports of CXCL12 possessing a chemorepulsive effect at high concentrations (168). Regardless, an inhibitor of CXCR4 is being evaluated in two phase I studies in pancreatic, ovarian, and colorectal cancers (NCT03277209, NCT02179970). An additional approach involves targeting of focal adhesion kinase (FAK), a signaling protein that regulates the cellular response to contact with extracellular matrix constituents (169). Hong and colleagues discovered FAK expression in PDAC to correlate with immunosuppressive microenvironments as measured by low CD8 T cell density and high granulocytic infiltrates

(170). Inhibition of FAK results in delayed tumor growth kinetics in KPC mice and a decreased prevalence of Tregs, MDSCs, and TAMs, and thus sensitizes refractory KPC tumors to checkpoint blockade (170). Based in part on these data, a number of FAK inhibitors are under clinical evaluation in phase I and II studies for patients with PDAC, with a number of trials combining FAK inhibition with pembrolizumab (NCT02546531, NCT02758587), or chemotherapy (NCT02428270, NCT02651727). Though immune modulation has proven insufficient to provide survival benefit for patients with PDAC to date, the coming decade holds great promise for successful translation of rationally-designed combination regimens that may finally provide longer-term hope for patients with this dreaded disease.

1.1.4: Summary

While immunotherapy has delivered unprecedented clinical benefit to some patients with some tumor types, the majority of patients treated with immune modulating agents fail to achieve durable responses. Moreover, some tumor types remain almost universally refractory.

Increased understanding of the malignant antigens recognized by the immune system is improving our ability to educate our own immune systems against cancer, while investigation into the complex network of interactions between immune cells and non-immune cells in the tumor microenvironment is informing our ability to identify patients poised to respond or not to therapy. Advanced prostate and pancreatic cancers remain among the most challenging indications, yet novel rationally-designed therapeutic approaches are on the clinical horizon that may finally render these cancers sensitive to therapy. Therapeutic modulation of tumor myeloid cells is one such approach. In light of their complex biology and the myriad roles these cells play in cancer, the next section will review the etiology, function, and therapeutic targeting of myeloid cells in cancer.

1.2: Myeloid Immunosuppression: A Critical Barrier to Antitumor Immunity

The term “myeloid” cell refers to any cell of primitive or definitive hematopoietic origin which descends from the common myeloid progenitor cell. This common myeloid progenitor can give rise to two daughter lineages, one a precursor of megakaryocytes and erythrocytes

(MEP), and the other a precursor of monocytes and granulocytes (171). Cells descending from the granulocyte/monocyte progenitor include polymorphonuclear neutrophils, basophils, and eosinophils, and mononuclear monocytes, macrophages, and dendritic cells. Together, these cells largely compose the innate arm of the immune system. These cells are generally highly adaptive and are readily programmed by a multitude of dynamic environmental signals in their local niche, and thus have wide-ranging roles in physiology from host defense to maintenance of tissue integrity and homeostasis. As a result of this plasticity, they are also implicated in a number of disease states including heart disease, obesity, diabetes, allergy and autoimmunity, and cancer (172).

Chronic inflammation is a long-recognized hallmark of cancer that is both a result of and contributor to the phenotype and function of myeloid cells in cancer (173, 174). Systemic inflammation interferes with normal myelopoiesis in terms of both the quantity and quality of myeloid cells produced, resulting in unique circulating myeloid populations that can seed the tumor environment and modulate tumor growth and the host response to cancer. Moreover, the complex metabolic and cellular milieu of the tumor microenvironment can further program myeloid cells to participate in largely normal physiological processes – such as those reminiscent of wound repair – that paradoxically serve to promote tumor immune evasion, invasion, and metastatic dissemination. The high plasticity characteristic of myeloid cell biology is both a blessing and a burden; it provides therapeutic potential for reprogramming myeloid function towards tumor-antagonistic functions, yet results in a complex biology that confounds identification and validation of optimal therapeutic approaches. The aim of this

section is to review the ontogeny and function of two major classes of myeloid immune cells in the context of cancer, and discuss the current state of the art in targeting myeloid cells for cancer therapy.

1.2.1: Ontogeny of Myeloid Derived Suppressor Cells and Tumor Associated Macrophages

Myeloid Derived Suppressor Cells. An enigmatic population of myeloid cells in cancer capable of apparent immune suppression but lacking in definitive phenotypic marker expression was first identified in the 1970s (175). Not until the turn of the century, however, were there sufficient technological tools and understanding of hematopoiesis to accurately describe the identity and nature of these cells. An early definitive phenotype arose with identification of the Gr-1 epitope recognized by the RB6-8C5 antibody (176), which was later shown to bind both Ly6G and Ly6C markers on granulocyte/monocyte lineage cells in mice

(177). Cells co-expressing CD11b and Gr-1 were shown to be elevated in the context of cancer (178), immunosuppressive (179-181), and dependent on myeloid colony stimulating factors such as GM-CSF for accumulation in the blood or spleen (180, 182). These cells were shown to phenotypically resemble monocytes and granulocytes, but were accepted as a unique population due to their immature nature and potent immunosuppressive function (183).

Thus, myeloid derived suppressor cells (MDSCs) constitute a unique population of myeloid cells characterized in large part by immunosuppressive function in pathologic contexts such as cancer.

Fundamentally, MDSCs consist of immature myeloid precursor cells derived from the common myeloid progenitor that have not completed differentiation into mature monocytes or granulocytes, such as neutrophils. As such, MDSCs can be separated into two main cell types; monocytic MDSCs (MO-MDSC) or polymorphonuclear MDSCs (PMN-MDSC), based upon

phenotypic and morphologic resemblance to monocytes or neutrophils, respectively (184).

The generation of MDSCs in pathological conditions such as cancer is thought to require two

discrete signals; first, colony stimulating factors such as GM-CSF, G-CSF, or SCF-1 are

required to induce myelopoiesis to expand a pool of immature myeloid progenitors (180, 185,

186). Second, low-level chronic inflammatory signals drive aberrant activation of these precursor cells, which halts their differentiation and supports transcriptional programming towards immunosuppressive phenotypes (187). The concerted actions of GM-CSF and IL-6 have been described to promote MDSC differentiation and accumulation in a number of tumor models (147, 188-190), which has shed light on the signaling pathways governing the MDSC phenotype. IL-6 is known to induce phosphorylation, homodimerization, and nuclear translocation of STAT3. Thus, inhibition of STAT3 with tyrosine kinase inhibitors such as sunitinib (191) or by genetic ablation (192, 193) can reduce accumulation of suppressive

MDSCs in tumors, while ablation of SOCS3, a negative regulator of STAT3, can have the

opposite effect (194). STAT3 likely governs multiple aspects of MDSC biology from

mobilization (195) and differentiation (196) to immunosuppression through induction of Nox2

(197) and arginase (198). In addition, GM-CSF and IL-6 can induce C/EBPβ, a transcription

factor that also been implicated as a critical mediator of MDSC differentiation (199). Like

STAT3, C/EBPβ can regulate expression of arginase and Nox2, implicating this factor in

MDSC suppressive function (199). Additional pathways thought to be involved in MDSC

differentiation include IRF8 (200, 201), Notch signaling (202), RB1 (203, 204), and adenosinergic signaling (205, 206). Moreover, suppressive activity of MDSCs may further be

bolstered by signaling through STAT6 (207, 208), NF-κB (209, 210), PGE2/COX2 (211, 212),

and the ER stress response transcription factor CHOP (213). Given the diverse chronic

inflammatory stimuli capable of naturally generating MDSCs in vivo, together with the inherent

cellular diversity within this pool of pathologically-activated myeloid precursors, MDSCs are 29

fundamentally heterogeneous cells with context-specific phenotype and function in cancer.

Tumor Associated Macrophages. The earliest postulations of cellular immunity, dating back

to the work of Elie Metchnikoff in the late 19th century, described the function of phagocytic

cells as a means of host defense, rendering macrophages among the first immune cells ever

studied (214). Macrophages have been a central focus of observation and investigation ever

since, and continue to surprise and confound researchers with their remarkable range of

phenotypes and functions in both homeostatic and pathological states across almost all

tissues. Early on, it was assumed that all macrophages present in human tissue were derived from circulating monocytes produced via “definitive” hematopoiesis in the bone marrow (215).

Classically, these circulating monocytes express Ly6C and CCR2 in mice (CD14 in humans) in contrast to patrolling monocytes that express CX3CR1 (CD16 in humans) and rarely extravasate into tissues to undergo differentiation (216, 217). However, macrophages can

also develop during “primitive” hematopoiesis occurring in the yolk sac during embryonic

development, and these macrophages have recently been shown to seed many tissues and

remain as resident macrophages where they are maintained by in situ self-renewal through

adulthood (218-220). Interestingly, absent fate mapping manipulations, tissue resident

macrophages in a given tissue are indistinguishable from macrophages recently differentiated

from recruited circulating monocytes; a testament to the remarkable plasticity of this cell

lineage in responding to loco-regional cues within the tissue (221). Upon inflammatory insult,

macrophages will expand in tissues either through in situ proliferation of tissue resident

macrophages (222, 223) or through recruitment and differentiation of circulating monocytes

(216), and the relative contribution of these two cellular sources is likely governed by the tissue

itself and the nature of the inflammatory signal. In cancer, tumor associated macrophages

(TAM) are generally thought to undergo continual replacement by circulating monocytes or

MO-MDSCs, however recent studies suggest there may be unique temporal or functional contributions by TAMs of definitive or primitive origin. For example, Franklin and colleagues report in the MMTV-PyMT breast cancer model that early expansion of macrophages in the

TME is the result of proliferation by tissue resident macrophages, however long-term TAM accumulation requires contribution from the monocyte pool (153). Moreover, Zhu et al report that TAMs in the KPC model of PDAC are derived from both primitive and definitive hematopoiesis, and that monocyte-derived TAMs may primarily participate in immunosuppression, while tissue resident TAMs promote local desmoplasia (224). Future studies are needed to illuminate with greater clarity the relative contributions of monocyte- derived and tissue resident macrophages across cancers, as tissue-specific programming is likely to play a role in dictating this balance.

A plethora of distinct factors contribute to macrophage recruitment, differentiation, and polarization during homeostasis and in the context of cancer. Fundamentally, signaling through the CSF-1 receptor is the critical factor supporting macrophage differentiation and survival. This receptor binds both CSF-1 (225) and IL-34 (226), with both cytokines likely contributing to the generation and stability of the macrophage pool. Loss of CSF-1R signaling through genetic ablation or pharmacologic inhibition results in a drastic reduction in TAMs, suggesting CSF-1R signaling is critical for macrophage recruitment and longevity in the TME

(227-229). Moreover, the chemokine CCL2 provides a critical recruitment signal for circulating monocytes to enter the tumor bed and differentiate into TAMs (216, 230), although additional factors such as VEGF can contribute to this recruitment (231). Differentiation of circulating monocytes into TAMs appears to involve Notch pathway signaling, specifically through the transcription factor RBPJ (153). Additionally, MO-MDSCs can serve as an additional source of TAMs (60). Kumar and colleagues report that under hypoxia MO-MDSCs engage the hypoxia-inducible factor Hif-1α to induce CD45 tyrosine phosphatase activity, which

downregulates STAT3 and results in differentiation of MO-MDSCs into TAMs within hypoxic zones (232). Upon differentiation, macrophages integrate numerous environmental signals to generate a context-dependent phenotype and function (233). The “M1/M2” paradigm has been created in an attempt to define these phenotypes, and, while useful, this paradigm is widely accepted as an oversimplification of macrophage diversity in vivo (234). The

“classically activated” M1 macrophage phenotype is induced by IFN-γ and LPS, while the

“alternatively activated” M2 phenotype is induced by IL-4 and IL-13, mirroring the Th1/Th2 cytokine paradigm in T cell biology (235). These factors drive distinct transcriptional programs characterized by unique transcription factor usage (STAT1 and NF-κB versus STAT6 and

Myc), cytokine output (IL-12 versus IL-10), metabolic enzyme production (iNOS versus arginase), and surface molecule expression (CD80 and CD86 versus CD163 and CD206)

(174, 236, 237). These phenotypes have been considered extremes along a spectrum of macrophage diversity, however characterization of multiple “flavors” of alternatively activated macrophages has challenged the linear spectrum model of macrophage identity (238). Thus, the M1/M2 phenotypes may be considered landmark hues within a color wheel of interrelated identities (Fig 2). This plasticity presents a challenge in understanding the critical factors that polarize TAMs in cancer, as unique anatomical cues and tumor-derived factors in different models can dynamically modulate TAM phenotype and function (154). Therefore, high- dimensional analysis of tumor myeloid cell transcriptional and phenotypic characteristics at the single cell level will be required to fully understand the complex nature and function of

TAMs in any given cancer tissue (239-241).

Figure 2: Models of macrophage polarization and plasticity. Classical M1/M2 nomenclature utilizes a linear spectrum model, with LPS/IFNγ stimulated M1 macrophages and IL-4/IL-13 stimulated macrophages as extremes along the spectrum of possible phenotypes. A more complex view of macrophage plasticity is represented by the color wheel model, where multiple landmark phenotypes exist in a non-linear, interrelated fashion. This model is likely more accurate in describing the broad range of macrophage phenotypes present in tumors across tissues.

1.2.2: Mechanisms of T Cell Immunosuppression by MDSCs and TAMs

Myeloid infiltration in cancer is increasingly appreciated as a negative prognostic marker across many histologies and is thought to underlie resistance to immunotherapy.

Immunosuppressive activities by MDSCs and TAMs are diverse and context-specific, in line with the general theme of myeloid biology. Though unique in ontogeny and phenotype, TAMs and MDSCs exhibit many overlapping functional properties in the TME, thus this section will be organized according to the individual mechanisms of immune suppression by myeloid cells, and specific contributions from individual subsets of TAMs or MDSCs will be highlighted therein.

Nutrient Deprivation. Catabolism or sequestration of macromolecular nutrients can have immunomodulatory effects, especially on cytotoxic T cells (242). Interestingly, despite the array of environmental metabolites involved in sustaining T cell effector function, only two main amino acid catabolism pathways are common among myeloid cells in cancer, targeting arginine and tryptophan. As a corollary, cysteine sequestration has been observed in MDSCs, however the generalizability of this observation has not been demonstrated (243). Arginine catabolism by myeloid cells is largely controlled by the enzymes arginase-1 and iNOS, which create distinct byproducts and are generally expressed in different cell types. Arginase hydrolyzes arginine to form ornithine and urea, compounds that are thought to be relatively indolent in regard to T cell function, though a potential suppressive role of ornithine has been reported (244, 245). Arginase is often expressed within tumor-infiltrating MO-MDSCs, PMN-

MDSCs, and TAMs, and is thought to function primarily by denying T cells access to this critical environmental metabolite. T cells deprived of arginine exhibit a profound deficiency in proliferative capacity, concomitant with downregulation of the CD3ζ chain of the T cell receptor

(246-248). Furthermore, in the absence of arginine T cells fail to induce key cell cycle

regulators and thus stall in the G0-G1 phase (249, 250). Interestingly, arginase activity may

promote MDSC accumulation in some circumstances (251), suggesting a positive feedback

loop between arginine metabolism and MDSC recruitment may exist. Arginase expression is

mainly controlled by Th2-type cytokines including IL-4 and IL-13, which engage STAT6 to drive arginase gene transcription in association with C/EBPβ in macrophages (252, 253).

Additionally, activation of Hif1-α in hypoxic conditions results in arginase upregulation in

MDSC (60). In contrast, iNOS catabolizes arginine to form citrulline, NADPH, and nitric oxide

(NO). Unlike arginase, iNOS is induced by Th1-type cytokines such as IFN-γ, leading to the

classification of iNOS as a marker of M1-like macrophages (254). The value of iNOS

expression in the tumor myeloid compartment is likely context dependent, as it is associated

with beneficial inflammatory macrophage polarization, yet can contribute to T cell suppression

through release of toxic NO species by MDSCs, as discussed below (60).

Tryptophan catabolism in mammals is primarily mediated by the enzymes indoleamine

2,3-dioxygenase-1 (IDO1), IDO2, and TDO, with IDO1 the predominant enzyme expressed in

cancer (255). These enzymes break down tryptophan to form a metabolite N-formyl-

kynurenine, which itself constitutes a molecule with immunomodulatory properties (256).

Thus, both tryptophan deprivation and downstream catabolite formation contribute to the

immunosuppressive function of IDO in the TME. IDO was discovered as an interferon- inducible protein (257), so in cancer IDO often acts as a counter-regulatory mechanism induced in response to cytotoxic T cell function (48). Both MDSCs and TAMs can express

IDO (258, 259), and tumor-intrinsic IDO expression can enhance MDSC accumulation in tumors (260). Similar to the effects of arginine depletion, IDO1-mediated depletion of tryptophan causes growth arrest in T cells, which was shown to involve the amino acid sensing pathway GCN2 (261, 262). In addition, kynurenine products activate the aryl hydrocarbon 35

receptor (AHR) (256, 263), which can downregulate the TCR CD3ζ chain or induce outright

apoptosis in T cells (264, 265). In light of these observations, therapeutic targeting of both

IDO1 and arginase are being attempted in preclinical and clinical studies at this time, and will

be reviewed later in this chapter.

Free Radical Species. Free radicals consist of an atom or group of atoms that possess

unpaired electrons. These molecules arise as normal byproducts of biochemical reactions,

and when not properly reduced are by nature extremely reactive species, stealing electrons

from surrounding molecules to initiate a chain reaction of free radical generation that can

damage or decay macromolecular structures within or around a cell. The most common free radicals produced by myeloid cells are reactive oxygen species (ROS) including superoxide

•- (O2 ) or reactive nitrogen species (RNS) such as nitric oxide (NO) (266). Myeloid ROS

production is mediated by NADPH oxidase (Nox1) and arginase-1 (Arg1), while RNS are

primarily produced by iNOS (266). In general, it is thought that MO-MDSCs and TAMs are the

predominant producers of RNS, while PMN-MDSCs primarily produce ROS (267, 268). When

Nox1/Arg1 and iNOS are active in the same location, superoxide and NO species will react

together spontaneously to form peroxynitrite (ONOO-), one of the strongest oxidizing agents

found in vivo (269, 270). In normal physiology, ROS/RNS are utilized by both macrophages

and neutrophils to kill bacteria or helminths, however in cancer these pathways can interfere

with T cell recruitment, antigen recognition, and T cell viability.

Peroxynitrite species react irreversibly with four amino acids; methionine, cysteine,

tryptophan, and tyrosine in a process called nitration (266, 271). Nitration of proteins in the

TME can be detected with antibodies that detect nitrotyrosine, which was shown to correlate

with MDSC numbers in human tumors (272). Chemokines such as CCL2 can be nitrated,

interfering with T cell recruitment to both murine and human prostate tumors (273).

Interestingly, given the role of CCL2 in chemotaxis of both T cell and myeloid populations,

nitro-CCL2 seems to selectively interfere with CD8 T cell recruitment, while myeloid cells and

Tregs remain able to infiltrate either through retained binding to nitro-CCL2 or through other

redundant chemotactic pathways (273). Additionally, nitration of tumor MHC molecules can

interfere with peptide binding, preventing presentation of tumor antigen for recognition by CD8

T cells (274). In this study, Lu and colleagues report that MDSC are the major contributors of peroxynitrite species compared to macrophages or tumor cells across a number of human and murine tumor types (274). On the other side of the equation, nitration of the T cell receptor or the CD8 surface molecule can also occur (275). These modifications further interfere with

TCR binding to peptide-MHC complexes on the surface of tumor cells, resulting in T cell hyposensitivity in the TME (276). Thus, MHC nitration by MDSCs can prevent T cell cytotoxicity independent of anergy induction or suppression of T cell effector capacity.

Nevertheless, excessive exposure to peroxynitrite in the TME can render T cells dysfunctional regardless of antigen detection efficiency, through suppression of IL-2 signaling or outright toxicity (277, 278). Therefore, the concerted actions of MDSCs and TAMs to induce ROS/RNS in the TME can lead to local nitration of critical surface markers and chemokines to selectively interfere with CD8 T cell recruitment, longevity, and tumor-specific cytotoxicity.

Immunosuppressive Ligands. Both MDSCs and TAMs express a wide variety of surface molecules and secreted factors that can suppress local T cell responses through both direct and indirect mechanisms (174). Indirect mechanisms involve recruitment or polarization of suppressive cell types, such as Tregs. For example, MDSCs and TAMs can recruit Tregs through release of chemokines that act on CCR5 and CCR6 (279, 280), and expand Tregs locally through direct antigen presentation (281, 282) or release of TGF-β, a critical mediator of Treg differentiation (283). Moreover, MDSCs and TAMs can participate in mutual crosstalk,

reinforcing suppressive phenotypes through complementary cytokine signals. For example,

IL-10 production by MDSCs can promote TAM polarization towards an M2-like phenotype associated with NO production and reduced TNF-α release (284). Moreover, through cell- contact dependent interactions MDSCs can reduce IL-12 expression in TAMs, while TAMs further promote IL-10 production by MDSCs (285). Together, these complex interactions can fortify the immunosuppressive stroma, creating a local extrinsic barrier to antitumor T cell function.

In contrast, a number of cytokines and surface ligands expressed by TAMs and

MDSCs act directly on T cells to suppress their effector functions. Cytokines such as TGF-β and IL-10 derived from MDSCs or TAMs – in addition to their indirect effects described above

– can directly suppress CD8 T cell cytotoxicity. TGF-β exposure induces downregulation of key effector molecules in CD8 T cells including perforin, granzyme B, IFN-γ, and Fas ligand

(286), reduces IL-2 signaling and cell cycle regulators needed for proliferation (287), and is associated with T cell exclusion from the tumor bed (288). IL-10 also contributes to the suppression of T cell proliferation by regulating costimulatory signaling through CD28 (289).

Thus, TGF-β and IL-10 are together considered powerful mediators of peripheral T cell tolerance (290, 291). Furthermore, ligation of PD-1 on T cells by the coinhibitory ligands PD-

L1 or PD-L2 is a common mechanism of direct T cell suppression in cancer. MDSCs are known to express PD-L1 in both murine and human tumors (292, 293), and expression may be more pronounced on PMN-MDSCs relative to MO-MDSCs (294). This ligand is functional, as antibody blockade of PD-L1 can reduce contact-dependent T cell suppression by PMN-

MDSCs in ex vivo assays (294). Moreover, TAMs are similarly observed to express PD-L1 in murine models (295) as well as a number of human tumor types (296-298). Expression of PD-

L1 can occur at a greater magnitude on TAMs compared to tumor cells, and thus can contribute to escape of immunogenic tumors from T cell-mediated killing (299). Therefore, in 38

addition to metabolic or redox-related T cell suppression through depletion of environmental nutrients and release of ROS/RNS species, TAMs and MDSCs can express a number of cell surface or secreted ligands which are capable of suppressing antitumor T cell responses through indirect or direct mechanisms.

1.2.3: Therapeutic Targeting of MDSCs and TAMs in Cancer

Myeloid Derived Suppressor Cells. Given the potent immunosuppressive capacities of

MDSCs and their negative association with clinical outcome or response to therapy in patients with a range of different tumor types (300), MDSCs present an attractive target for therapeutic intervention. Such approaches to date have targeted MDSCs in three ways; blockade of

MDSC recruitment, interference with MDSC differentiation and survival, or inhibition of immunosuppressive functions. The canonical chemokine receptors governing monocyte and neutrophil chemotaxis are CCR2 and CXCR2, respectively. Blockade of CXCR2 has been tested preclinically, with efficacy as single agent or in combination with chemotherapy or checkpoint blockade in PDAC (301), ovarian (302), and colorectal cancers (303). Based on these data, a small molecule inhibitor of CXCR2 is being evaluated in patients with mCRPC

(NCT03177187). Inhibition of CCR2 has been evaluated clinically in combination with

FOLFIRINOX in a phase Ib trial of patients with borderline resectable or locally advanced

PDAC, with 49% of patients showing objective response by RECIST to the combination therapy (304). Interestingly, individual depletion of MO-MDSCs and TAMs by CCR2 blockade or PMN-MDSCs by CXCR2 inhibition can lead to compensatory myelopoiesis, in which the untargeted myeloid population will be recruited to the tumor in greater numbers (305). Thus, blockade of both CCR2 and CXCR2 may be necessary for optimal clinical benefit. Additionally,

CCR5 expression has been described on both MO-MDSCs and PMN-MDSCs, and may describe a highly suppressive subset of MDSCs (306). This observation illuminates the high

level of redundancy in the immune chemokine receptor network on myeloid cells, which

presents a challenge even to combination approaches. Thus, it remains to be seen whether

chemokine receptor modulation will be a tenable approach to significantly ameliorate myeloid

suppression in human cancers.

In addition to inhibiting recruitment, MDSC numbers can be reduced in cancer through

interference with MDSC differentiation, direct induction of apoptosis, or maturation into DCs

or other mature myeloid populations. One such approach with both preclinical and clinical

testing involves all-trans retinoic acid (ATRA), a derivative of vitamin A. ATRA promotes

maturation of myeloid precursors into DCs through upregulation of glutathione synthase (307,

308), and when used therapeutically effectively reduces the number of MDSCs in tumor-

bearing hosts (309). This effect has been seen in a small study evaluating ATRA with IL-2 in

18 patients with metastatic renal cell carcinoma, where ATRA improved the MDSC:DC ratio

(310). Recently, ATRA was evaluated in combination with ipilimumab in a phase II study of

patients with melanoma, where it reduced the number and suppressive function of MDSCs,

while increasing the frequency of circulating mature myeloid populations. However, the overall

clinical benefit of this approach was not fully reported (311), thus it remains to be seen whether

ATRA will synergize with immunotherapy in the clinical setting. A similar approach consists of inhibition of STAT3, which plays a major role in MDSC differentiation and suppressive programming. A number of STAT3-targeting therapies are currently under design and preclinical evaluation, and it appears that this approach can promote maturation of MDSCs into inflammatory DC populations (312). In contrast, direct depletion of MDSCs by apoptosis can be achieved in a number of ways. In response to ER stress, MDSCs upregulate TRAIL receptors relative to normal monocyte and neutrophil populations. Targeting of these receptors can thus induce apoptosis somewhat selectively in the MDSC compartment (313).

Furthermore, targeting of S100A8 and S100A9 with engineered peptides can induce MDSC

depletion with greater efficiency than Gr-1 targeting antibodies in mice (314). Interestingly, many chemotherapies when given at low doses can cause preferential apoptosis in the MDSC compartment, including gemcitabine (315), 5-fluorouracil (316), docetaxel (317), and doxorubicin (318). Clinical evaluation of these approaches alone and together with immunotherapies is needed to reveal which will most effectively counter MDSC suppression and synergize with immunotherapy in patients.

Finally, the immunosuppressive programs utilized by MDSCs can be targeted to reduce their suppressive influence in the tumor microenvironment. For example, ROS/RNS production by MDSCs can be indirectly inhibited by a synthetic triterpenoid CDDO-Me, which

also induces expression of Nrf2 to promote survival under redox insult (319). Similarly,

nitroaspirin can reduce NO production by MDSCs (320). Arginase expression in MDSCs can

be reduced by inhibitors of COX-2 (211) or PDE-5, providing an opportunity for highly-utilized clinical compounds tadalafil or sildenafil to be used therapeutically in cancer patients (321,

322). As such, tadalafil has shown efficacy in patients with head and neck cancer and multiple myeloma (323, 324). Another clinical compound with an unexpected role in MDSC biology is entinostat, an epigenetic modulator that inhibits class I histone deacetylase (325). This approach may have additional effects on TAMs as well, as is common among therapies targeting general mechanisms of myeloid-mediated immunosuppression (326). Together, a number of approaches are currently under preclinical and clinical investigation to target MDSC localization, function, and viability in cancer. Preclinical and clinical studies will be required to unravel which approaches are optimal across histologies, especially in the combination setting to enhance T cell-targeted immunotherapies.

Tumor Associated Macrophages. Like MDSCs, TAMs possess potent immunosuppressive capabilities and are often associated with poor clinical outcome or response to therapy in

patients with many tumor types (327). The predominant means of targeting TAMs to date

involves modulation of the CSF-1:CSF-1R signaling axis, given the dominant and exclusive role of this pathway in macrophage differentiation, stability, and alternative polarization. In a number of murine models and in some human tissues, tumor cells directly secrete large

amounts of CSF-1 (228, 328-330), presumably to establish local immune privilege and to

receive protumorgenic signals. Blockade of CSF-1R signaling through antibody blockade or

small molecule inhibition of CSF-1R-associated tyrosine kinase activity leads to both

reductions in the frequency of intratumoral TAMs and repolarization of residual macrophages

towards inflammatory phenotypes (152, 330). This inflammatory reprogramming of the tumor

microenvironment induces expression of T cell checkpoint ligands in the tumor

microenvironment, thus combination of CSF-1R inhibition with CTLA-4 and PD-1 engenders

strong therapeutic effects in mice with orthotopically implanted PDAC (330). In addition, CSF-

1R inhibitors have shown preclinical synergy with other conventional therapeutic modalities

including androgen depletion in prostate cancer (331) or chemotherapy in PDAC (332) and

breast cancer (333). Interestingly, patients with rare diffuse-type tenosynovial giant-cell

tumors (dt-GCT) show nearly universal overexpression of CSF-1R due to characteristic

chromosomal amplifications at the CSF-1 locus (229). Thus, in a phase I study CSF-1R

blockade induced objective responses in 86% of patients, and 2 of 28 patients achieved

complete responses (334). Currently, there are over 20 clinical trials ranging from phase I to

phase III evaluating CSF-1R inhibition as single agent or in combination with chemotherapy,

radiation, or immunotherapy across most solid tumors and some hematologic malignancies

(335). However, outside of dt-GCT, response rates have been quite modest, though the

efficacy of CSF-1R in the combination setting remains to be seen (336). While CSF-1R

inhibition allows for macrophage-specific targeting, a number of other pathways play important

roles in TAM differentiation and polarization, providing other promising – if less specific –

therapeutic avenues. The γ isoform of PI3K was recently described to be a critical molecular

switch in controlling immunosuppressive macrophage phenotypes in PDAC and other cancers

(337). Like CSF1R inhibition, pharmacologic blockade of PI3Kγ reprograms TAMs to support

CD8 T cell function while inhibiting stromal desmoplasia and associated tumor metastasis

(338), and in other tumor models PI3Kγ inhibition can potentiate checkpoint blockade (339).

Based on these data, an inhibitor of PI3Kγ is currently in phase I evaluation across advanced solid tumors (NCT02637531). In addition, RORγ was recently shown to regulate myelopoiesis, contributing to both TAM differentiation and M2-like polarization, as well as MDSC accumulation in human and murine tumors (340). Greater prioritization of TAM-targeting therapies in clinical studies should reveal in the coming years which myeloid signaling pathways are the most promising therapeutic targets in cancer.

Macrophage plasticity is generally disadvantageous in cancer, as tumors efficiently polarize macrophages to support tumor growth and immune evasion. However, a growing body of work supports the idea that macrophage plasticity also presents a unique therapeutic opportunity. TAMs are often the most prevalent immune population within tumors; if repolarized from tumor-supporting to tumor-antagonistic, they could present a significant

“trojan horse” approach to turn a tumor’s defenses against itself. The most striking example

of this concept to date involves an agonistic antibody targeting the myeloid stimulatory

receptor CD40. In a study by Beatty et al, the authors report CD40 agonism leading to T cell-

independent, macrophage-mediated tumor cytotoxicity in PDAC tumor-bearing KPC mice,

presumably through release of reactive oxygen or nitrogen species and tumor cell

phagocytosis (341). PDAC is known to possess a large and heavily immunosuppressive

myeloid infiltrate, thus inflammatory repolarization by agents such as αCD40 could in theory

be quite powerful by turning the vast reservoir of TAMs against the tumor. An agonistic

antibody targeting human CD40 has progressed through phase I testing (341, 342) and is 43

currently being evaluated in a phase II study in PDAC in combination with gemcitabine and

nab-paclitaxel in the adjuvant or neoadjuvant setting (NCT02588443). As with any immune

agonist, CD40-targeting antibodies can exhibit toxicities due to systemic elaboration of inflammatory cytokines, thus the balance between efficacy and toxicity will be critical in determining the clinical feasibility and success of this approach. An alternative approach to

CD40 activation involves activation of innate pattern recognition receptors such as Toll-like receptors (TLRs). The most common TLRs targeted for TAM repolarization in cancer include nucleic acid detectors TLR3 (polyI:C; synthetic dsRNA analog), TLR7/8 (imidazoquinoline; synthetic ssRNA analog), and TLR9 (ODN1826; synthetic CpG-rich dsDNA). These agents generally facilitate NF-κB activation through classical MyD88 or TRIF signaling downstream of the respective TLR, and upon local delivery can induce expression of proinflammatory cytokines such as TNFα, IL-12, IL-6, and to a lesser extent IFN-β (343-346). Clinical translation of these agonists is underway, with over 50 trials established for phase I/II testing in many solid tumors (347). Interestingly, a more recently characterized pattern recognition receptor STING – which is critical for innate detection of cytosolic DNA species – was shown to be essential for innate recognition and response to immunogenic tumors, while TLR signaling was dispensable (348). Whether agonists of STING can promote TAM repolarization and exhibit therapeutic activity in myeloid-rich tumors such as CRPC and PDAC is a main focus of investigation in this dissertation, and will be reviewed in detail below.

1.2.4: Summary

Tumor associated macrophages and myeloid derived suppressor cells are heterogeneous cell populations with a vast array of context-dependent functions in cancer. This complex biology has presented a challenge in describing the nature of these cells across tumor types in mouse

and man, and has delayed development of therapies aimed at reducing or modulating their functions. However, recent investigation has revealed potent immunosuppressive and tumor- promoting roles for these cells, and has elevated them as key barriers precluding efficacy of both conventional cancer therapy as well as exciting new immunotherapies. Prioritizing myeloid-targeting trials in the coming years should illuminate optimal methods for myeloid modulation in cancer, and, when utilized in combinatorial regimens, these agents bring hope of converting refractory, non-responding patients to responders in tumors currently considered terminal.

1.3: STING: A Powerful Activator of Innate Immunity in the Tumor Microenvironment

In contrast to adaptive immune cells, which undergo genetic rearrangement and hypermutation to express individualized receptors with exquisite specificity for a given antigen, innate immune cells recognize generalized molecular patterns unique to non-self pathogens or states of cellular distress through expression of germline-encoded pattern recognition receptors (PRRs). Conserved families of PRRs have been described, including the membrane-bound Toll-like Receptors (TLR) and C-type Lectin Receptors (CLR), and the endosomal NOD-like Receptors (NLR) and RIG-like Receptors (RLR). These families encode numerous unique receptors capable of sensing an array of molecular structures that are essential to the viability or pathogenesis of non-self pathogens, or are associated with the host stress response to infection or tissue damage.

Particular attention is devoted to nucleic acid sensing, as nucleic acids are indispensable for essentially all viruses and bacteria, and genetic material from these pathogens can be readily discriminated from mammalian DNA and RNA. For example, TLR9 senses unmethylated cytidine-phosphate-guanosine (CpG) rich DNA sequences, as bacteria lack the epigenetic machinery needed to undergo CpG methylation that is normally found in the mammalian genome. In addition to unique physical characteristics, aberrant cellular compartmentalization of nucleic acids can also constitute a sign of infection or stress. The presence of double-stranded DNA in the cytoplasm has long been known to induce innate immune activation, for example (349). However, the specific PRRs required for the detection of and response to cytoplasmic DNA in mammalian cells eluded discovery until recent years.

Work by Barber and Chen described the cyclic GMP-AMP synthase (cGAS) – stimulator of interferon genes (STING) pathway as the major innate detection system for cytosolic DNA, and these studies revealed STING to be a potent activator of Type I interferon production in both hematopoietic and non-hematopoietic cells. In light of studies by Gajewski and others

associating a Type I interferon gene signature with an immunologically “hot,” or T cell-inflamed

tumor microenvironment, efforts are underway to investigate the potential for therapeutic

STING activation to induce Type I interferons within immunologically “cold” non-T cell-

inflamed tumors.

This section aims to describe the discovery of the cGAS-STING pathway, summarize

current knowledge regarding the role of STING in antitumor immunity, and discuss the use of

STING agonists as therapies in both mouse and man.

1.3.1: Discovery and Characterization of the cGAS-STING Pathway

Discovery of STING. A fundamental component of the host response to viral infection is

induction of type I interferons including IFN-α and IFN-β. By the early 2000s, a number of

PRRs capable of inducing type I interferons in response to viral RNA and DNA had been

identified, including membrane-bound TLR3 and TLR9, which detect the presence of

endosomal double-stranded viral RNA and unmethylated CpG DNA, respectively (350, 351).

Additionally, a number of cytoplasmic RNA helicases including RIG-I and MDA5 were

described to detect and induce interferon responses to viral RNA present in the cytosol (352,

353). However, the host response to viral double-stranded non-CpG DNA was known to occur

independent of TLR9 signaling (349, 354), suggesting the presence of an undescribed DNA

sensor capable of sensing viral DNA in the cytoplasm.

Contemporaneous reports by Ishikawa, Zhong, and Jin in 2008 identified the

TMEM173 gene product STING (also known as MITA, ERIS, MPYS) as a novel component

of innate immune signaling. Interestingly, the first report of STING by Jin et al found STING in

association with MHC-II as a mediator of apoptotic signaling in lymphoma, which remains a

cryptic function of STING that has not been further substantiated (355). In both studies by

Ishikawa and Zhong, however, STING was identified in expression cloning screens for a robust capacity to mediate release of IFN-β upon overexpression in 293T cells (356, 357).

These studies implicated IRF3, IRF7, TBK1, and NF-κB in signaling downstream of STING, which were known to be involved in the interferon response to non-CpG DNA. This, together with the observation that STING-deficient mouse embryonic fibroblasts (MEFs) exhibit increased susceptibility to infection by the DNA virus herpes simplex virus 1 (HSV1) established STING as a putative DNA sensor. At the time, a number of cytosolic DNA sensors including DAI, AIM2, IFI16, and DDX41 had been identified, however none were uniquely required for the type I interferon response to non-CpG DNA. In a subsequent study, Ishikawa provided conclusive evidence that STING is essential for induction of IFN-β following exposure to cytosolic DNA, using STING-deficient MEFs, macrophages, and dendritic cells to demonstrate a lack of interferon release following transfection with viral DNA (HSV1, HSV2,

Ad5), bacterial DNA (E. coli, Listeria monocytogenes), and synthetic DNA species (non-CpG immunostimulatory DNA, poly-dAT:dAT, poly-dGC:dGC) (358). Of note, STING-deficient cells retained TLR9-mediated responses to CpG-containing DNA as well as TLR3-mediated responses to poly(I:C) RNA analogues, suggesting STING-mediated nucleic acid signaling and TLR-mediated nucleic acid signaling are discrete and independent entities (358).

Together, these reports identified STING as a critical component of the mammalian type I interferon response to cytosolic DNA.

Given this critical role for STING in cytoplasmic DNA sensing, it was initially thought that STING was a direct sensor of non-CpG DNA. However, early reports did not demonstrate an obvious association between STING and DNA. Bearing four transmembrane domains,

STING was found to localize to the endoplasmic reticulum (ER) as well as the ER-associated mitochondrial membrane, and upon activation translocate through the golgi apparatus to perinuclear endosomal structures, which serve as foci of STING signaling (358, 359). While 48

the long C-terminal domain of STING is oriented towards the cytosol where it can theoretically sense extra-nuclear DNA, this restricted localization profile is uncommon among known DNA sensors, which are generally non-membrane-bound cytosolic proteins. In support of this,

Sauer and Burdette identified cyclic dinucleotides – rather than DNA – as direct ligands of

STING (360, 361). Using a forward genetic approach in mice mutagenized with N-ethyl-N- nitrosourea, Sauer identified a point mutation in STING which abolished responsivity of mice to L. monocytogenes infection due to a lack of response to the bacterial second messenger nucleotide cyclic di-GMP (CDG) (361). Burdette utilized this null STING allele (termed goldenticket; STINGgt) to definitively describe STING as a sensor of cyclic dinucleotides (360).

Interestingly, in this study Burdette identified an additional R231H point mutant of murine

STING which was not stimulated by CDG, but could induce IFN-β in response to DNA transfection. This suggested the presence of an additional DNA sensor capable of generating

a novel and unique agonist of STING in response to cytoplasmic DNA.

Discovery of cGAS. To identify the elusive molecule capable of activating STING in response

to cytosolic DNA, Wu et al developed a cell-free assay to evaluate lysates from DNA-

transfected L929 fibroblasts for ligands capable of inducing STING-mediated IRF3 homodimerization. Through both chromatography and STING-based affinity purification, they identified by mass spectrometry a dinucleotide cyclic GMP-AMP (cGAMP) which could bind and activate STING (362), including the R231 isoform identified previously which did not respond to CDG. This was the first description of a cyclic dinucleotide produced endogenously within a metazoan cell. Using a complex fractionation strategy of DNA-transfected L929 extracts, the same group identified a highly conserved but uncharacterized enzyme with sequence and structural homology to the human oligoadenylate synthase (OAS1), a member of the nucleotidyltransferase family of enzymes. This novel enzyme, termed cyclic GMP-AMP

synthase (cGAS), was found to bind DNA and catalyze the synthesis of cGAMP, which induced IRF3 dimerization and IFN-β production in multiple cell types in a manner dependent on STING expression (363). This study was closely followed by three others which confirmed cGAS as a DNA sensor capable of synthesizing cGAMP (364-366). Further, these reports refined the description of cGAS enzymatic activity, as initial mass spectrometry analysis of cGAMP failed to distinguish the non-canonical G(2’5’)pA(3’5’) “mixed” phosphodiester linkage present in cGAS-synthesized cGAMP, in contrast to the G(3’5’)pA(3’5’) canonical linkage in cGAMP produced by bacterial cGAMP synthases. Together, these reports conclusively identified cGAS as a novel DNA sensor capable of inducing type 1 interferon responses to cytoplasmic, non-CpG DNA through synthesis of a novel CDN bearing mixed phosphodiester linkages which potently activates STING.

The cGAS-STING Signaling Pathway. The cGAS-STING pathway is an essential, non- redundant means of cellular nucleic acid sensing with critical roles in anti-viral immunity, autoimmunity, and tumor immunity, as described herein. Both STING and cGAS are highly conserved proteins with ancestral homologs described in anemone, suggesting this pathway evolved as an early means of host defense over 500 million years ago (367, 368). In contrast to other nucleic acid sensors such as TLR9, which is expressed predominantly in rare plasmacytoid dendritic cells and B cells, STING and cGAS exhibit a broad expression pattern and are found across hematopoietic cells in addition to non-immune endothelium, epithelial tissues, neurons, and fibroblasts.

Signaling through the cGAS-STING pathway is precipitated by the presence of cytosolic double-stranded DNA (dsDNA) species, which serve as cellular danger signals by providing scaffolds for cGAS oligomerization and resulting catalytic activity. Under normal conditions, dsDNA is compartmentalized within the nucleus or mitochondria, thus the

presence of cytosolic dsDNA can be indicative of infection by intracellular bacteria or viruses,

endogenous genotoxic stress or chromosomal instability (369-372), or necrotic cell death by neighboring cells that release unprocessed DNA (348, 373-375). Binding of cGAS to dsDNA occurs along the phosphate-sugar backbone of the minor groove of B-form DNA, allowing binding to occur in a sequence-independent manner (376, 377). Once bound, cGAS covers approximately 20 DNA base pairs (bp), allowing cGAS to recognize short dsDNA fragments.

However, at low DNA concentrations approximating physiological conditions during viral infection, cGAS activity is correlated with dsDNA fragment length, with a minimum size of 40

bp required for activation (378, 379). Monomeric binding of cGAS to dsDNA does not result

in catalytic activity; rather, dimerization of two cGAS:dsDNA units induces cGAS activation

(380). While this cGAS2-DNA2 complex formation is a thermodynamically unfavorable process, it facilitates “ladder-like” binding of additional cGAS dimers, resulting in cooperative formation of cGAS2n-DNA2 complexes (378). This cooperative binding acts to amplify the

downstream 2’3’-cGAMP signal, and likely explains the dsDNA length-dependent activity of

cGAS under physiological conditions. A number of stress-related nucleic acid binding proteins

such as TFAM and HMGB are known to bind and induce “U-turn-like” structures in DNA, which

can nucleate cGAS2n-DNA2 complexes and bypass the unfavorable cGAS2-DNA2 structure

under stress conditions (378). Of note, the DNA length-dependency of cGAS is more

pronounced with the human form of the enzyme relative to that found in other mammalian

cells, owing to two amino acid substitutions on the DNA-binding surface of human cGAS that

permit preferential binding to longer DNA sequences (381). This demonstrates a potential evolutionary emphasis on selective activation in humans, as opposed to highly sensitive activation as seen in lower species (381). Recently, Du et al found that high-order cGAS-DNA

structures undergo phase separation to form liquid-like “droplets” within the cytosol, which

concentrate and amalgamate available cGAS-DNA complexes to optimize cGAMP production

efficiency upon reaching a threshold concentration of cytosolic DNA (382). Thus, catalytic

activation of cGAS upon binding cytosolic DNA is a complex process with unique regulatory

mechanisms that promote selective activation under conditions of cellular stress or infection.

Upon activation by dsDNA, cGAS undergoes a conformational change that exposes

its catalytic domain to undergo a two-step reaction by which free ATP and GTP nucleotides

are converted to 2’3’-cGAMP, a high-affinity ligand for STING. At the basal state, STING exists

as a monomer localized to the endoplasmic reticulum (ER) membrane due to RYR and RIR

signal sequences surrounding its four N-terminal transmembrane domains (359). Two STING

monomers are required to bind a single cyclic dinucleotide such as 2’3’-cGAMP, and this

binding-induced dimerization facilitates activation of downstream signaling. Specifically, seating of 2’3’-cGAMP into the binding pocket of the cytoplasmic portion of STING dimers causes a closing of the two “wing-like” C-terminal arms and the formation of a β-sheet lid that covers the dinucleotide. This conformational change is associated with a novel trafficking pattern, whereby active STING dimers translocate from the ER through the golgi apparatus and form perinuclear endosomal structures in a process that resembles autophagosome formation (358, 383). These endosomal punctae serve as foci of STING signaling, which involves recruitment of the inhibitor of κB kinase (IKK) family member TANK-binding kinase 1

(TBK1), which associates with the S366 residue of STING, resulting in autophosphorylation on S172 of TBK1 (384, 385). Upon autophosphorylation, TBK1 is activated to phosphorylate the C-terminal portion of STING at S353, S358, and S379 residues (385). This phosphorylation event is thought to promote recruitment of interferon regulatory factor 3

(IRF3), which then is phosphorylated by TBK1 at S385/S386 and T404/S405 to mediate rapid

IRF3 dimerization. Additionally, in association with tumor necrosis factor receptor-associated factor 6 (TRAF6), TBK1 induces NF-κBp65/RelA activation through the IKKαβ activation loop

Type I Interferon Production

Figure 3: The canonical STING signaling pathway. STING is activated by cyclic dinucleotide molecules, which are either derived from bacteria (c-di-GMP/c-di-AMP) or from the cytoplasmic DNA sensor cGAS (cGAMP) following detection of viral- or host- derived DNA in the cytosol. Upon binding cyclic dinucleotides, activated STING trafficks through the golgi to perinuclear endosomes, where it recruits TBK1 to activate downstream signaling components IRF3 and NF-κB (not shown). This signaling promotes robust type I interferon production, in addition to acute anti-viral inflammation-associated gene transcription.

(386). As a result, activated IRF3 and NF-κB/p65 undergo rapid translocation into the nucleus, where they act in concert to drive anti-viral gene transcription, predominantly characterized by induction of type I interferons (Fig 3). While STING-mediated activation of NF-κB/p65 promotes “canonical” NF-κB signaling, activated STING is also capable of mediating non- canonical NF-κB signaling in a TBK1-independent manner, which involves interactions with

TRAF3 to promote IKKα-mediated activation of NF-κBp52 (386). Furthermore, activated

STING was shown to physically associate with STAT6, which can be phosphorylated on S407 by TBK1, leading to homodimerization and nuclear translocation to induce a specific repertoire of chemokines including CCL2, CCL20, and CCL26 (387). Therefore, while canonical STING signaling results in potent type I interferon release due to the coordinated activation of IRF3 and NF-κBp65, STING does engage other non-canonical signaling pathways to achieve a robust acute pro-inflammatory program upon cellular infection or therapeutic activation.

1.3.2: Role of the STING Pathway in the Antitumor Immune Response

STING in Spontaneous Tumor Immunity. Immune control of cancer generally requires the generation of a cytotoxic CD8 T cell response, as these cells are in most circumstances the central mediators of tumor cell cytolysis. Naïve CD8 T cells capable of specifically sensing tumor antigens can only gain cytolytic effector function after priming by professional antigen presenting cells (APCs), however, as these cells alone are capable of delivering the costimulatory signals required to unleash effector programs in CD8 T cells following recognition of cognate tumor antigen through MHC-peptide-TCR engagement. Thus, proper

APC function is a critical rate-limiting step in the process of generating antitumor T cell responses. In 2005, a study by Dunn et al demonstrated a critical role for type I interferons

(IFN-α/β) in establishing tumor immune surveillance, and the cellular target of IFNα/β was

mapped to host hematopoietic cells (39). Contemporaneous reports by Diamond and Fuertes

et al in 2011 identified a unique lineage of dendritic cells (DCs) driven by the Batf3

transcription factor as the specific cellular target of IFN-α/β, as the action of IFN-α/β on these

cells was required to facilitate CD8 T cell priming and accumulation within tumors (34, 388).

These DCs are characterized by expression of cell surface markers CD8α or CD103, and are

thought to play a critical role in tumor antigen cross-presentation, a unique process by which

exogenous tumor antigens are internalized and shunted away from the MHC class II peptide

processing pathway and loaded on to MHC class I, allowing professional APCs to induce CD8

priming with exogenous, rather than endogenous, antigens. However, at the time it was

unclear what stimulus induced IFNα/β and resulting cross-presentation activity by Batf3-

lineage DCs, as the sterile tumor microenvironment was thought to be bereft of traditional

pathogen associated molecular patterns (PAMPs) only present during viral or bacterial

infection.

To identify the innate signaling pathway capable of mediating sterile activation of

Batf3-lineage DCs in the context of cancer, Woo et al implanted immunogenic B16-SIY tumors

into mice lacking key components of TLR signaling (MyD88, TRIF, TLR4, or TLR9), antiviral

RNA signaling (MAVS), purinogenic signaling (P2x7R), or STING signaling (STING, or IRF3).

Of all pathways tested, only host STING and IRF3 were required for generation of a tumor

specific CD8 T cell response and resulting regression of implanted tumors (348). Tumor DNA,

rather than any other component of the B16 tumor extract, was sufficient to induce IFN-β in

DCs in vitro, and using imaging cytometry, DCs carrying tumor-derived DNA were identified

in the tumor microenvironment. These same cells exhibited increased levels of IFN-β

transcripts relative to DCs from lymph nodes or spleen. Thus, DNA from dying tumor cells

constitutes a key danger associated molecular pattern (DAMP) capable of inducing STING-

mediated type I interferon responses within the sterile tumor microenvironment (348). A 55

following study by Klarquist et al confirmed a role for STING in mediating both antitumor and autoimmune responses following uptake of DNA from dying cells (373). In inducible models of colorectal cancer and glioma, STING deficiency was associated with enhanced tumor growth relative to wild type animals, further suggesting STING signaling can counteract tumor progression (389, 390). Interestingly, epigenetic silencing of both STING and cGAS promoters has also been described in human colorectal carcinoma and melanoma (391, 392), and HPV viral oncoproteins E7 and E1A have been shown to bind STING and block its activation (393).

This apparent evolutionary pressure to limit STING expression or function further supports a protective role for STING-mediated IFNα-β responses in limiting tumorigenesis. Therefore,

STING activation in host DCs following sensing of tumor-derived DNA and the associated

IFNα/β response is a critical rate-limiting step in the induction of cytolytic CD8 T cell responses

against cancer.

STING in Radiation Therapy. Localized tumor irradiation has long been utilized in cancer

therapy because of its ability to induce genotoxic stress in chromosomally-unstable malignant

cells, leading to tumor cell death. Various immune cell populations are also susceptible to radiation therapy (RT), which initially led many to believe that RT efficacy is unrelated to host

immune function (394). However, recent studies suggest the efficacy of RT is in fact

dependent on the immune system, as ablative radiation is largely ineffective in

immunodeficient mice (395). Further, RT can induce adaptive immune responses and act

synergistically with checkpoint blockade (395, 396). Delineation of the cGAS-STING pathway

and its role in cellular DNA sensing led to the hypothesis that radiation-induced DNA damage

may constitute a DAMP capable of promoting innate sensing of irradiated tumors and resulting

induction of adaptive antitumor immunity. Deng et al provided evidence for this hypothesis,

using mice with genetic deficiencies in innate immune signaling pathways to demonstrate that

STING and cGAS are uniquely required for the antitumor efficacy of RT (397). Exposing

tumors to ablative RT results in cGAS-STING mediated production of IFN-β primarily by tumor

CD11c+ DCs, presumably due to sensing of tumor DNA released upon radiation-induced

necrotic cell death. This has been confirmed by the observation that induced expression of

the DNA exonuclease three prime repair exonuclease 1 (TREX1) in irradiated tumors

degrades cytosolic DNA and attenuates RT-induced activation of the cGAS-STING pathway

(398). Interestingly, TREX1 induction is highly dependent on radiation dosing, with high doses applied in single fractions between 12 and 18 Gy constituting a threshold level capable of inducing physiologically-relevant levels of the exonuclease. However, repeated administration of low-dose radiation does not cause TREX1 upregulation, and appears optimal for inducing

cGAS-STING mediated IFN-β induction, Batf3-lineage DC cross-presentation, and resulting

priming of adaptive immune responses (398). It is intriguing to note that a collection of clinical

case studies reporting systemic regression of metastatic cancer following local irradiation of

a single primary lesion – termed the “abscopal” effect – utilized this fractionated low-dose

approach (399-402). A number of these reports involved the use of fractionated RT together

with checkpoint blockade immunotherapy, suggesting the vaccine-like immune potentiating

effects of cGAS-STING activation in tumor DCs can prime adaptive immune responses that

can be bolstered by other immune modulators. Therefore, the cGAS-STING pathway plays a

critical role in the immune-potentiating effects of RT, and fine-tuning clinical RT protocols to

optimize STING activation may result in superior clinical activity especially when combined

with checkpoint blockade immunotherapy.

STING in Non-Myeloid Tumor Stroma. Most early studies that have validated STING as a

critical component of the antitumor immune response describe its function in promoting

IFNα/β responses in Batf3-lineage DCs, as this promotes tumor-specific CD8 T cell priming

(348, 373). However, cGAS and STING are widely expressed among cells of the hematopoietic lineage as well as non-hematopoietic tissues, and engagement of this pathway can have unique effects depending on the cell type in which it is acting. For example, while

STING activation fortifies DC survival and antigen presentation function, a number of reports have suggested an inhibitory role for STING activation within T cells. Even before the cGAS-

STING pathway was described, it was known that bacterial cyclic dinucleotides such as CDG could inhibit proliferation of Jurkat T cells (403). More recently, a collection of studies provided evidence of STING signaling triggering T cell-intrinsic IRF3-mediated IFNα/β production, which was associated with decreased proliferative capacity, increased cellular stress, and proapoptotic gene expression (404-406). Specifically, these studies together report unique engagement of proapoptotic genes such as BAX (406), PUMA, and BIM (405), genes associated with the unfolded protein response including Bip/HSPA5 and GADD34 (406), and genes associated with the TP53 pathway (405). However, these studies utilized non-canonical forms of STING activation including the highly cell-permeable small molecules 5,6-

Dimethylxanthenone-4-acetic acid (DMXAA) or 10-carboxymethyl-9-acridanone (CMA), or enforced expression of a constitutively active form of STING bearing the V155M point mutation. Therefore, despite these insights, it remains to be determined whether physiologic

STING activation within the confines of the tumor microenvironment following cGAS-mediated

DNA sensing or therapeutically-administered cyclic dinucleotide agonists is sufficient to significantly attenuate T cell proliferation or survival.

DNA damage is known to upregulate expression of stress-related ligands such as

NKG2D on tumor cells in a STING-dependent fashion, which serve as activation signals for cytotoxic natural killer (NK) cells (407, 408). While type I interferons are known to be essential for NK cell homeostasis and antitumor effector functions (409), few studies have deeply probed the effects of STING activation on NK cell cytotoxicity in tumors. Data from Takashima

et al suggest that STING signaling in both B16D8 melanoma tumor cells and host

hematopoietic cells contribute to NK cell recruitment, possibly through chemokines CXCL10

and CCL5 (410). While NK cell recruitment correlated with suppression of tumor growth in this

model, it was not determined whether NK cell function was absolutely required for the

response, or clearly dependent on STING signaling. In an in vitro system evaluating

cetuximab-mediated NK-DC crosstalk against HPV+ head and neck squamous cell carcinoma

(HNSCC) cell lines, Lu et al found cGAMP able to synergize with cetuximab to enhance human NK cell expression of HLA-DR and IFN-γ, markers of NK cell activation (411). Further,

Kim et al reported a role for NK cell function in the efficacy of STINGVAX, a GM-CSF-secreting cellular vaccine formulated with CDN. These data also suggested CDN are capable of elaborating IFN-γ production by NK cells (412). Together, studies suggest STING activation can promote NK cell cytotoxicity against cancer, however it remains to be definitively shown to what extent the efficacy of CDN is dependent on NK function, and in which tumor indications

STING-mediated NK cell activation is required.

In addition to infiltrating lymphocytes, STING is also highly expressed in tumor endothelial cells (413). In an early study evaluating the efficacy of intratumoral cGAMP in melanoma, Demaria and colleagues found that within 5 hours of injection, tumor endothelial cells are the highest producers of IFN-β, in both murine and human tumor extracts (414).

Additionally, using endothelial cell lines and purified DC populations in vitro, they demonstrated a greater responsivity of endothelial cells to tumor DNA and cGAMP as measured by IFN-β mRNA induction. These data are seemingly in conflict with the notion that activation of Batf3-lineage DCs is central for the antitumor efficacy of STING agonists, however this study does not demonstrate a sufficiency for endothelial cells in promoting cGAMP-mediated immunity in mice deficient for Batf3. In addition, these experiments utilized lipofectamine-encapsulated cGAMP for both in vitro and in vivo experiments, rather than 59

naked CDNs as is used in the majority of preclinical studies as well as in early clinical trials

for STING agonists. Lipid formulation may lead to preferential uptake of CDNs by endothelial

cells which would not occur using naked CDN. Thus, this study does not disprove a role for

STING-mediated DC activation in antitumor immunity elicited by CDNs, but identifies endothelial cells as an additional unexpected cellular source of STING-induced IFN-β in the tumor microenvironment.

1.3.3: Current Knowledge of the Therapeutic Utility of STING Agonists in Cancer

Classes of STING Agonists. There is currently significant interest in the development of agonistic ligands for STING for therapeutic use in cancer. While STING is traditionally known to bind and undergo activation in response to cyclic dinucleotide molecules, the first STING agonist tested as an anti-cancer agent was a small molecule compound 5,6-

Dimethylxanthenone-4-acetic acid (DMXAA), which underwent preclinical and clinical evaluation before the STING pathway was identified. DMXAA was initially developed as an analog of flavone-8-acetic acid (FAA), which in 1989 demonstrated potent antitumor effects against multiple murine tumor models (415). This activity was initially ascribed to effects on the tumor vasculature, however later studies showed additional requirements for immune activation by DMXAA, including induction of IFN-β in tumor macrophages and DCs and recruitment of CD8 T cells (416-419). At the time, it was known that FAA – the parent compound of DMXAA – failed in clinical trials due to a lack of activity against human cells

(420), yet DMXAA was rapidly translated to clinical evaluation due to its significantly enhanced potency in murine models relative to FAA (421, 422). DMXAA progressed through phase I and II trials (423) before failing in definitive randomized phase III studies (424). These disappointing results were explained following identification of the cGAS-STING pathway, as

it was determined that DMXAA binds and potently activates murine STING, but does not bind

or stimulate human STING due to the presence of a single amino acid substitution in the

human protein (365, 425, 426). DMXAA remains useful as a tool for inducing STING activation in murine models, however it is better utilized as a cautionary tale to properly study the mechanisms of action of potential therapies on cells of human origin before translation into

large-scale clinical trials.

Despite the checkered past of DMXAA, great interest remains in the study of cyclic

dinucleotides as therapeutic STING agonists in mouse and man. The most basic CDNs first

evaluated as preclinical agents for cancer are cyclic di-GMP (CDG) and cyclic di-AMP (CDA),

which consist of two nitrogenous purine nucleobases cyclized by canonical 3’-5’

phosphodiester linkages. These molecules are utilized uniquely as second messenger

molecules in bacteria, and thus serve as immune PAMPs indicative of bacterial infection.

Bacterial CDG exhibits a Kd of 1.21µM for STING homodimer binding, and this relatively weak binding is an exothermic process (366). CDG and CDA are readily hydrolyzed by phosphodiesterase enzymes, potentially limiting their potential bioavailability as therapeutic molecules (427, 428), although the specific hydrolases expressed by humans able to degrade bacterial CDNs have not been rigorously described. The central weakness of bacterial CDNs as therapeutic compounds in humans relates to their inability to activate unique isoforms of the STING gene that are commonly expressed in the human population. Specifically, an allele bearing arginine to histidine substitution at residue 232 (R232H) of STING found in approximately 14% of humans exhibits reduced binding and activation in response to bacterial

CDNs (429, 430). Thus, bacterial CDNs are relatively weak agonists of STING and are unlikely to be translated as clinical compounds due to an inability to broadly activate all naturally- occurring isoforms of STING found in the human population.

In contrast, the metazoan CDN 2’3’-cyclic GMP-AMP (2’3’-cGAMP) is a structurally

unique molecule allowing for high-affinity binding to all human STING alleles. Unlike CDG and

CDA, 2’3’-cGAMP is cyclized by a non-canonical G(2’5’)pA(3’5’) phosphate bridge, commonly

described as a “mixed” phosphodiester linkage (364-366). This mixed linkage structure is

unique to metazoan CDNs, and endows the molecule with a flexibility not found in bacterial

CDNs. This allows 2’3’cGAMP to bind more deeply within the CDN binding groove of

dimerized STING and mediate conformational changes in the protein associated with

enhanced activation. These include a narrowing of the C-terminal wing-like arms of STING,

which draw together nearly 20 Å closer than occurs after CDG binding, and the formation of

a β-sheet lid that covers the CDN binding pocket (366). Additionally, the free 3’OH present on

the GMP residue of 2’3’-GAMP interacts specifically with the serine 162 residue of STING,

which enhances the affinity of this interaction (366). As a result, 2’3’-cGAMP exhibits a Kd of approximately 4 nM with dimerized STING, nearly an order of magnitude greater binding

affinity than that of CDG (366). The mixed linkage of 2’3’-cGAMP additionally confers binding

to the R232H allele of human STING and permits downstream signaling, unlike bacterial

CDNs (429, 430). However, the mixed linkage structure of 2’3’-cGAMP does not render the

CDN resistant to phosphodiesterase cleavage, as Li et al identified an enzyme ecto-nucleotide pyrophosphatase/phosphodiesterase (ENPP1) capable of specifically hydrolyzing 2’3’- cGAMP (431). ENPP1 deficiency in murine lung fibroblasts resulted in enhanced IFN-β production in response to 2’3’-cGAMP and HSV infection, demonstrating ENPP1 is able to exert regulatory control over 2’3’-cGAMP responses in vivo. Therefore, 2’3’-cGAMP is a high affinity ligand for all naturally-occurring STING isoforms in mouse and man, yet its in vivo potency can be limited due to sensitivity to enzymatic hydrolysis.

Synthetic analogs of naturally-occurring CDNs represent a novel class of STING agonists designed for high stability, affinity, and specificity for STING isoforms in the therapeutic setting. A critical modification that enhances CDN stability is the inclusion of sulfur

residues within the phosphate bridge, as such phosphorothioate diester linkages inhibit

hydrolysis by phosphodiesterases (432, 433). Inclusion of bisphosphorothioate linkages within

2’3’-cGAMP inhibits cleavage by ENPP1, and does not negatively affect STING binding or activation (431). Furthermore, unique bisphosphorothioate diastereomers possess subtle differences in potency, as Corrales et al report that RP-RP diastereomers elicit greater IFN-β release by THP-1 monocytes than RP-SP forms (429). Based upon these insights, the synthetic CDN ML-RR-S2-CDA is currently being investigated as the first-in-class STING targeting therapeutic in clinical trials (NCT02675439, NCT03172936). Together, these studies identify mixed linkage CDNs bearing bisphosphorothioate bridges of RP-RP diastereometry as

ideal therapeutic CDNs with high stability due to phosphodiesterase resistance, robust per-

molecule potency due to high STING affinity, and broad applicability to stimulate all common

isoforms of STING found in mouse and man.

Preclinical Evaluation of STING Agonists in Cancer. In the decade since the cGAS-STING

pathway was identified and characterized, a wealth of data has been generated evaluating

each class of STING agonist as a cancer therapy across a broad array of common preclinical

tumor models. The earliest reports of CDNs in cancer utilized the bacterial compound CDG

as an adjuvant to potentiate the therapeutic effect of cancer vaccines. Chandra et al

incorporated CDG as an adjuvant for an attenuated Listeria monocytogenes-based vaccine

expressing the MAGE-b antigen for use in the 4T1 murine breast cancer model, and found

CDG capable of suppressing metastasis, boosting CD8 T cell responses, and stimulating IL-

12 production by MDSCs when combined with vaccine (434). Interestingly, this early report is

one of the only studies to date identifying specific effects of CDNs on MDSCs. The adjuvant-

like effects of CDG in boosting tumor-specific CD8 T cells was confirmed in a similar study

evaluating CDG in combination with a “TriVax” strategy of poly-IC, CD40 agonist antibody,

and a modified gp100 peptide in the B16-F10 model of murine melanoma (435). While these early studies utilized CDG in combination with other potential innate stimulators, Hanson et al demonstrated that nanoparticle encapsulation of CDG alone can result in robust adjuvant properties by suppressing systemic diffusion and promoting accumulation of the compound within lymph nodes, allowing for more profound expansion of viral specific T cells and antibody titers relative to naked CDN or other common adjuvants (436). These studies established an early framework for the use of CDNs in cancer, demonstrating potent adjuvant-like properties of CDG including activation of myeloid antigen presenting cells, induction of antigen specific

CD8 T cells, the possibility of unique formulations for enhanced efficacy, and a broad applicability to diverse histologies including breast cancer and melanoma.

While CDNs were rapidly accepted as potent adjuvants in both viral and cancer vaccine models, the next conceptual advancement in the field involved the idea of in situ vaccination; that intratumoral delivery of a CDN can directly facilitate antigen-specific T cell priming in the absence of an a priori designed vaccine. Support for this idea was provided in contemporaneous reports by Woo and Klarquist et al who discovered a role for the cGAS-

STING pathway in controlling the innate immune response to tumor-derived DNA DAMPs, suggesting natural STING activation in this context is necessary and sufficient to induce antigen specific antitumor T cell-mediated immunity, as the tumor and its associated antigen repertoire act as an intrinsic vaccine (348, 373). In light of this, the first report of in situ vaccination through intratumoral CDN delivery was provided by Ohkuri and colleagues, who locally administered CDG into orthotopically implanted GL261 gliomas. Despite the technical challenge associated with this model, they found CDG capable of eliciting CCL5 and CXCL10 cytokines within the intracranial space, recruiting both CD8 and CD4 T cells that were capable of fully regressing of gliomas in half of mice tested (389). Perhaps the seminal report of in situ vaccination through intratumoral CDN injection came from Corrales et al, who generated novel

synthetic CDNs of high potency relative to CDG and evaluated the efficacy of local CDN

administration in murine models of B16 melanoma, 4T1 breast cancer, and CT26 colon

carcinoma (429). This study identified ML-RR-S2-CDA and provided the first evidence of superior therapeutic effects of this bisphosphorothioate modified CDN as monotherapy in treating established B16-F10 tumors relative to CDG, 2’3’-cGAMP, and ODN1668, a DNA- based agonist of TLR9. Furthermore, Corrales et al demonstrated local delivery of ML-RR-

S2-CDA could mediate regression not only of locally-injected tumors, but also was sufficient to achieve control of synchronously-implanted distal uninjected 4T1 or CT26 lesions. Similar effects have been observed in the field of radiation oncology, as in rare cases local irradiation of a primary lesion can be followed by spontaneous regression of systemic disease. Termed the “abscopal effect,” this is largely thought to be an immune-mediated phenomenon, and likely involves STING activation in response to radiation-induced tumor DNA release (397).

Thus, in situ vaccination by local CDN administration can induce curative T cell responses

that are not just localized to the environment of the injected lesion, but are capable of

circulating throughout the host and exerting cytolytic immune pressure on uninjected lesions such as distal metastases.

Building upon these insights, there has been great interest in incorporating CDNs into rationally-designed combinatorial approaches based upon complementary mechanisms of action between CDNs and other immunotherapies. For example, an existing antitumor T cell repertoire is thought to be a prerequisite for response to checkpoint blockade immunotherapy, thus agents like CDNs that are capable of boosting tumor-specific CD8 T cell frequencies are likely to act synergistically with checkpoint blockade. In addition, the interferon response characteristically associated with STING activation is known to induce expression of coinhibitory receptor ligands including PD-L1 (48), providing rationale to coordinately block the PD-1:PD-L1 axis upon intratumoral CDN therapy. Fu and colleagues were the first to

report the combination of PD-1 blockade with a STING-based therapeutic. Formulating ML-

RR-S2-CDA into a GM-CSF-secreting irradiated cellular vaccine (STINGVAX) reduced growth

kinetics of various tumor models including B16 melanoma, TRAMP-C2 prostate cancer, and

CT26 colon carcinoma, however only the combination of STINGVAX with αPD-1 was sufficient to induce outright tumor regression (437). This study was followed by another demonstrating synergy between local ML-RR-S2-CDG with PD-L1 blockade in the MOC1 model of head and neck squamous cell carcinoma (438). However, ML-RR-S2-CDG was ineffective in controlling another model of the same histology, MOC2, which was characterized as a non-T-cell-inflamed tumor. While the basis for this lack of efficacy was not explicitly determined, a similar report in the tolerized neu/N model of HER-2+ breast cancer suggested

that rationally combining ML-RR-S2-CDA with additional T cell coinhibitory and costimulatory

molecules including αPD-L1 and αOX40 could convert non-responding tumors to responders,

highlighting the capacity for CDNs to potentiate checkpoint modulation to promote cytolytic T

cell immunity in the face of pre-existing T cell tolerance (439). CDNs have additionally been

utilized in combination with agonists of 4-1BB (114) or GITR (440), innate agonists such as

TLR9 (441) or TLR7/8 (440), and radiation therapy (442). Thus, rational combinatorial

approaches incorporating CDNs with other immunotherapies can display therapeutic synergy

and are attractive for use even in immunologically “cold” tumors.

Despite the therapeutic benefits associated with STING-mediated acute inflammation

in cancer, excessive or prolonged engagement of this pathway can also lead to inflammatory

pathologies. A number of groups have observed robust local ulceration in tumors exposed to

CDNs, regardless of the tumor model or mouse strain under investigation (114, 442, 443). In

line with the initial description of DMXAA as a vascular remodeling agent, local CDN

administration appears to promote leakage or outright collapse of neovascular endothelium

within hours of injection (442). This is associated with a potent early elaboration of

inflammatory cytokines including IFN-β, TNFα, IL-6, and CCL2, and has been described as

hemorrhagic necrosis (442). Interestingly, this necrosis occurs in mice which lack adaptive

immune cells (RAG-/-) or interferon signaling (IFNαR-/-), but is reduced in mice deficient in

TNFα (443). Blockade of TNFα at the time of CDN injection also suppresses this pathology, suggesting STING-induced TNFα is central to local hemorrhagic necrosis (442). While even

excessive necrotic responses generally result in formation of an eschar that resolves over time, such pathology can be a source of infection for patients with superficial disease, or could have more complex and lasting effects if present at an injected visceral tumor. Data presented in this dissertation suggest that reduced dosing of CDNs can limit observed necrosis without significantly blunting therapeutic efficacy, and that the extent to which ulceration is observed can be related to both CDN dose and tumor size at the time of injection (114). Thus, balancing the immune promoting effects of CDNs with potential inflammatory necrotic pathology will be critical for successfully translating CDNs into clinical compounds.

1.3.4: Summary

The cGAS-STING pathway is central to innate immune responses in the context of bacterial infection, viral invasion, and cancer. STING activation liberates potent type I interferon release through the TBK1-IFR3 pathway, and also engages NF-κB signaling to produce a broad array of inflammatory cytokines, chemokines, and surface receptors across cells of both hematopoietic and non-hematopoietic origin. Cyclic dinucleotide agonists of STING are thus attractive candidates for cancer therapy, owing to their capacity to mediate in situ vaccine-like responses upon local injection by activating local myeloid cells, leading to enhanced antitumor

T cell immunity even in immunologically “cold” tumors.

Chapter 2: Methodology

Mice

Male C57BL/6 and B6(Cg)-Tyrc-2J/J albino mice were purchased from The Jackson

Laboratory, and STINGGt/Gt mice were a kind gift from the Dr. Kimberly Schluns. Mice were approximately 5 to 8 weeks old at the time of use. All mice were housed in accordance with the Association for Assessment and Accreditation of Laboratory Animal Care and NIH standards. All experiments were conducted according to protocols approved by the University of Texas MD Anderson Cancer Center Institutional Animal Care and Use Committee.

Cell Lines

The TRAMP-C2 tumor cell line was provided by Dr. Norman Greenberg and all

experiments in this manuscript were performed using passage 21-22 tumor cells. TRAMP-C2

cells were maintained as previously described (444). The mT4-2D cell line was a kind gift from

Dr. David Tuveson (445), and cells were maintained in complete DMEM containing 10% fetal

bovine serum (Gibco), 100 U/mL penicillin, 100 mg/mL streptomycin sulfate, and 2mM L-

glutamine. THP-1 Dual and J774-Dual reporter cells were obtained from InvivoGen and

cultured according to manufacturer’s instructions. HEK-Blue reporter cells were a kind gift from Dr. Simon Yu at the University of Texas MD Anderson Institute for Applied Cancer

Science (IACS) and cultured in cDMEM supplemented with zeocin (100µg/mL) and blasticidin

(30µg/mL) for antibiotic selection.

Cyclic Dinucleotides

Cyclic di-GMP (CDG) and 2’3’ cyclic GMP-AMP (cGAMP) were purchased from

InvivoGen and reconstituted in water according to manufacturer’s instructions. ML-RR-CDA was synthesized by Wuxi AppTec and reconstituted in water. IACS-8803 was designed in

collaboration with the University of Texas MD Anderson Cancer Center and synthesized by

Wuxi AppTec, and was reconstituted in water or PBS for in vitro or in vivo use. 68

Antibodies and Chemotherapy

Antibodies used in in vivo studies were purchased from BioXCell or Leinco

Technologies and were administered via intraperitoneal (IP) injection or intratumoral (IT)

injection as indicated at the following concentrations: anti-CTLA-4 (9H10; 100μg IP, 50μg IT),

anti-PD-1 (RMP1-14; 250μg IP, 50μg IT), anti-4-1BB (3H3; 200μg IP, 50μg IT). Flt3L-Ig (Flt3L;

Bio X Cell) was administered IT at 50μg per mouse, or 200μg per mouse where described.

Chemotherapy agents gemcitabine (Gemzar®; Eli Lily) and nab-paclitaxel (Abraxane®;

Celgene) were obtained from the University of Texas MD Anderson pharmacy and were

reconstituted according to manufacturer’s instructions. Both agents were administered IP at

120mg/kg. Antibodies used for flow cytometry were obtained from BD Biosciences,

BioLegend, or Roche.

STING Reporter Cell Assays

All STING reporter cell assays were conducted according to manufacturer’s protocols.

In short, CDN or control agonists were added at desired concentrations in 20µl PBS to desired

wells of flat-bottom 96-well plates. HEK-Blue, J774-Dual, THP-1 Dual, or THP-1 parental cell

lines were harvested, washed, counted, and resuspended for addition of 5x104 – 1x105 cells per well in 180µl recommended media. Cells were incubated with CDN for 20-24 hours, when supernatants were harvested for luciferase or SEAP assays using QuantiLuc® or

QuantiBlue® solutions according to manufacturer’s instructions. For ELISA, THP-1 parental cells were incubated with CDN for 72 hours, then supernatants were harvested and IFNβ levels were quantified using the PBL Human IFNβ ELISA kit according to manufacturer’s protocols.

MDSC Generation and T Cell Suppression Assays

Bone marrow was isolated from naïve male C57BL/6 mice through aspiration of the femur and/or tibia as described. Red blood cells were lysed using RBC lysis buffer (Sigma), 69

and 4x106 cells were plated in 10mL in non-TC treated 10cm petri dishes (Falcon) in RPMI containing 10% FBS, 100 U/mL penicillin, 100mg/mL streptomycin sulfate, recombinant mouse GM-CSF (40ng/mL; BioLegend) and IL-6 (40ng/mL; BioLegend), and 2-

Mercaptoethanol (β-ME; 55µM) to induce MDSC differentiation. After 4 days in culture, BM-

MDSC suspension cells were harvested, washed with cRPMI to remove cytokine, and re-

plated in 10mL cRPMI with no cytokine at 4x105 cells/mL in 10cm petri dishes. CDN were added at 1-10µg/mL as indicated for 24-48 hour repolarization.

For suppression assays, naïve splenocytes were obtained from male C57BL/6 mice, and CD8 T cells were isolated by negative selection with the Miltenyi Mouse CD8 T Cell

Isolation Kit. CD8 T cells were labeled with CellTrace CFSE (ThermoFisher) as described

(446). 1x105 CFSE-labeled CD8 T cells were plated at indicated ratios with unstimulated or

repolarized BM-MDSC in 200µl cDMEM containing 10% FBS, 50U/mL IL-2, 27.5µM β-ME,

and 2µg/mL αCD28 (clone 37.51; BioLegend) in round bottom 96-well plates coated with

αCD3 (BioXCell clone 145-2C11; coated overnight at 10µg/mL in PBS at 4oC). After 72-96

hours, CD8 T cells were analyzed for CFSE dilution by flow cytometry.

Human Macrophage Generation

Healthy donor buffy coats were obtained from the University of Texas MD Anderson

Cancer Center Blood Bank. Peripheral blood mononuclear cells were isolated by ficoll

gradient centrifugation (Histopaque 1077; Sigma), and CD14+ monocytes were isolated by negative selection using the Miltenyi Classical Monocyte Isolation Kit. Monocytes were differentiated into M2c macrophages as described (447). In short, monocytes were plated in

T25 flasks (1x106 cells/mL) in RPMI containing 20% FBS and 100ng/mL recombinant human

M-CSF for 6 days, with a media refresh on day 4. On day 6, media was replaced, with the

addition of recombinant human TGF-β (10ng/mL) and IL-10 (10ng/mL). On day 8, adherent cells were washed with PBS and cultured with cRPMI + 100ng/mL M-CSF + 10µg/mL 70

indicated CDN for repolarization over 72 hours. Adherent macrophages were harvested with

Detachin cell detachment solution (Genlantis) for downstream applications. All recombinant

human cytokines were purchased from PeproTech and reconstituted according to

manufacturer’s instructions.

Luminex and RPPA Sample Preparation

Repolarized human M2c macrophages or murine BM-MDSC were generated as

described above. Supernatants from repolarized cultures were isolated, spun to remove

contaminating cells, then frozen at -80o. All harvested supernatants were thawed on ice and were analyzed with the Cytokine & Chemokine 36-Plex Mouse ProcartaPlex or

Cytokine/Chemokine/Growth Factor 45-Plex Human ProcartaPlex Panels

(Invitrogen/ThermoFisher) using a Luminex MAGPIX® machine according to manufacturer’s instructions. For RPPA, a minimum of 1x106 cells were washed twice with cold PBS, and dry pellets were stored at -80o. Cell lysis and analysis was conducted at the University of Texas

MD Anderson Functional Proteomics RPPA Core Facility.

Subcutaneous Tumor Implantation and Treatment

Mice were implanted at day 0 on one flank with 2.5x105 mT4-LA or mT4-LS cells, or on both flanks with 1 x 106 TRAMP-C2 cells. Refer to indicated figures for exact treatment

schedules. Tumor volumes were measured with calipers to define the largest length (a), width

(b), and height (c), and volumes were calculated as (a x b x c).

Orthotopic Pancreatic Tumor Implantation and IVIS Imaging

Subconfluent cultures of mT4-2D, mT4-LA, or mT4-LS were harvested with 0.05%

trypsin, washed with PBS, and resuspended at desired concentration in ice cold PBS

containing 30% Matrigel (Corning) by volume for implantation. Each mouse was anesthetized

through isoflurane inhalation and was administered 3µg buprenorphine hydrochloride analgesic (Sigma) by intraperitoneal injection. Following sterilization of the surgical field with

70% isopropyl alcohol, a small (~1-2cm) incision was made on the left flank skin as well as the underlying peritoneal lining. Blunt-end forceps were used to access the spleen and accompanying pancreas. 50µl of cell suspension or CDN solution was administered into the head of the pancreas using a U-40 insulin syringe with 29.5 gauge needle. The peritoneal wound and surrounding skin were cleaned with betadine antiseptic swabs, and the wound was closed using 4-0 absorbable sutures (Oasis) for the peritoneal incision and the Autoclip

Wound Closing System for the skin (Braintree Scientific). Mice were allowed to recover under a heat lamp until displaying normal ambulance, and were monitored daily for signs of pain or discomfort. Wound clips were removed 10-14 days following application.

Mice were imaged weekly to monitor tumor growth using an IVIS® Spectrum in vivo imaging system (PerkinElmer). Mice were anesthetized with isoflurane and received 200µl D-

Luciferin (15mg/mL in PBS; GoldBio) by IP injection. Bioluminescent images were acquired

11-12 minutes following administration of D-Luciferin. All imaging was conducted in the

University of Texas MD Anderson Small Animal Imaging Facility.

Flow Cytometry Analysis of Tumor Immune Infiltrates

1 x 106 TRAMP-C2 cells were mixed with 30% collagen matrix (Matrigel, BD) and implanted at day 0 on both flanks. Mice received treatment when tumors reached ~200mm3 volume, on days 31, 35, and 39, and were sacrificed on day 42. Following euthanasia, primary tumors were harvested, massed, and finely diced into 70µm filters within 6cm petri dishes.

Diced tumors were enzymatically digested in a 37o incubator for 30 minutes in digestion media consisting of X-Vivo15 media (Lonza) supplemented with collagenase H (1mg/mL; Sigma) and DNAse (160µg/mL; Roche), then tumors were physically mashed through 70µm filters to create single cell suspensions. Total cells were counted, then live immune cells were purified by ficoll gradient centrifugation (Histopaque 1119; Sigma). Samples were fixed overnight and permeabilized using the FoxP3/Transcription Factor Staining Buffer Set 72

(eBioscience/ThermoFisher) for staining with fluorescently labeled antibodies in 96-well U-

bottom plates. Stained samples were analyzed using a BD LSRII or X-30 prototype flow

cytometer.

mT4-2D or mT4-LA cells were harvested with 0.05% trypsin, washed with PBS,

counted, and resuspended in ice cold PBS containing 30% Matrigel for implantation of 2.5x105 cells in 100µl subcutaneously or 3.5x104 cells in 50µl orthotopically as described. Mice were treated with intratumoral CDN, checkpoint blockade antibodies, or chemotherapy according to treatment schedules included in associated figures. Preparation of tumor samples for flow cytometry was completed as described above.

Chapter 3: Intratumoral STING Activation with T-Cell Checkpoint Modulation

Generates Systemic Antitumor Immunity

This work is based upon “Intratumoral STING Activation with T-cell Checkpoint Modulation

Generates Systemic Antitumor Immunity” by Ager et al. July 25 2017 Cancer Immunology

Research DOI: 10.1158/2326-6066.CIR-17-0049. It is presented with permission from

Cancer Immunology Research and the American Association for Cancer Research (AACR).

3.1: Introduction

Therapeutic blockade of T-cell coinhibitory receptors CTL-associated antigen 4

(CTLA-4) and programmed death 1 (PD-1) is sufficient to promote durable immune-mediated tumor regression across multiple cancers in mouse and man. Neither ipilimumab (anti-CTLA-

4) nor nivolumab (anti-PD-1), however, provide clinical benefit in the setting of castration-

resistant prostate cancer (CRPC) (448, 449). Prostate cancer is the second leading cause of

cancer-related death in men, and patients with CRPC have a median overall survival of 14

months, highlighting the urgent need for alternative immunotherapeutic approaches for this

disease (450). Preclinical models of prostate cancer, including the transgenic adenocarcinoma of the mouse prostate (TRAMP) mouse and its derivative TRAMP-C2 implantable cell line (444), reveal two major barriers to antitumor immunity in CRPC: (i) poor

CD8+ T-cell infiltration into the prostate tumor microenvironment (TME), and (ii) robust infiltration of immunosuppressive myeloid populations including tumor associated macrophages (TAM) and myeloid derived suppressor cells (MDSC). Given the well-described means by which TAM and MDSC impede T-cell infiltration and effector function independent of the CTLA-4/PD-1 axes (451), therapies which deplete or biologically reprogram TAM and

MDSC may render the prostate TME amenable to T-cell activity. Absent these barriers to T- 74

cell infiltration and function, we hypothesize that CRPC patients would be likely to experience durable benefit from T-cell checkpoint blockade.

In this study, we pursued a rational approach to design and evaluate multi-modal therapies for CRPC consisting of antagonistic antibodies targeting T-cell coinhibitory receptors CTLA-4 and PD-1, agonistic antibodies targeting the T-cell costimulatory receptor

4-1BB, and agonistic molecules targeting myeloid activation pathways including stimulator of interferon genes (STING) or FMS-like tyrosine kinase 3 (Flt3). Our lab previously established synergistic relationships between anti-CTLA-4 and anti-PD-1 antibodies, owing to their non- redundant roles in regulating central T-cell priming and peripheral effector function (452).

These data preceded successful phase III trials and subsequent FDA approval of this combination in patients with metastatic melanoma (453, 454). Additionally, we described complementary effects of CTLA-4 blockade with concomitant ligation of the stimulatory receptor 4-1BB, which promotes CD8+ T-cell survival, proliferation, and TH1 cytokine production (455). In each setting, combination therapies were maximally effective in the context of an irradiated cellular vaccine secreting Flt3-ligand (Flt3L). Known for its classical role in promoting survival, proliferation, and differentiation of hematopoietic progenitors, Flt3L can induce dendritic cell (DC) chemotaxis and conversion to a CD8α+CD103+ phenotype associated with cross-presentation and robust T-cell priming potential (456, 457). In contrast,

Flt3L also promotes differentiation of plasmacytoid dendritic cells, which can exert tolerogenic effects in cancer through induction of regulatory T cells (Treg) and limit the efficacy of Flt3L as a cancer immunotherapy (458).

The cytosolic nucleic acid sensor Stimulator of Interferon Induced Genes (STING) acts to promote tumor-specific T-cell responses by activating interferon secretion and costimulatory ligand expression by myeloid cells after uptake of tumor cell-derived DNA, and is indispensable for spontaneous rejection of immunogenic cancers (348). Cyclic dinucleotide

agonists of STING activate macrophages and DC in vitro and have demonstrated therapeutic

utility alone and in combination with checkpoint blockade across multiple tumor models (429,

437).

Using mice bearing bilaterally-implanted TRAMP-C2 tumors, we found that

intratumoral STING activation with the dinucleotide cyclic di-GMP (CDG) mediates regression

of injected tumors, but fails to elicit systemic “abscopal” immunity targeting distal lesions.

Triple checkpoint modulation with anti-CTLA-4, anti-PD-1, and anti-4-1BB, however, converts

intratumoral CDG into a potentiator of systemic immunity capable of rejecting distal, uninjected

lesions. In contrast, intratumoral Flt3L augments the benefit of triple checkpoint modulation,

but only when administered locally at each lesion. Given the potential toxicity of anti-CTLA-4,

anti-PD-1, and anti-4-1BB combination therapy, we investigated whether intratumoral administration of this combination at low doses could serve as a safer alternative to systemic antibody therapy. These studies demonstrate that curative abscopal immunity can be generated against multi-focal TRAMP-C2 when local triple checkpoint modulation is used in combination with intratumoral CDG. This therapeutic synergy is associated with infiltration of tumor-specific T cells and enhanced ratios of CD8+ T cells to Treg, MDSC, and TAM, as well

as pronounced downregulation of the M2 macrophage marker CD206 on TAM. Alternatively,

Flt3L failed to mobilize abscopal immunity in combination with triple checkpoint modulation,

likely due its expansion of CD4+ Teff and Treg populations with similar potency to CD8+ T cells at local and distal lesions. We observed dose-dependent skin pathology in mice receiving intratumoral CDG in combination with triple checkpoint blockade, although the therapeutic benefit of this combination appears modestly (but not significantly) dose-dependent. These data support the development of combinatorial strategies targeting adaptive and innate immune populations in prostate cancer, and inform the design of dosing strategies for STING agonists and checkpoint modulators to maximize clinical benefit while minimizing undue

toxicity.

3.2: Results

3.2.1: Intratumoral CDG potentiates systemic checkpoint modulation in TRAMP-C2

In both clinical and preclinical settings, prostate cancer remains refractory to

checkpoint blockade therapy targeting CTLA-4 or PD-1. To identify combinatorial strategies

capable of sensitizing prostate tumors to immunotherapy, we augmented our checkpoint

blockade cocktail of antagonistic anti-CTLA-4 and anti-PD-1 antibodies with an agonistic

antibody targeting 4-1BB, a costimulatory receptor of the Tumor Necrosis Factor Receptor

(TNFR) superfamily which promotes CD8+ T-cell survival and cytotoxic effector function (459).

We next investigated whether intratumoral STING activation using the STING agonist c-di-

GMP (CDG) can potentiate checkpoint modulation by promoting in situ vaccine-like

augmentation of T-cell priming and pro-inflammatory reprogramming of the murine prostate

tumor myeloid compartment. Mice were challenged with 1x106 TRAMP-C2 cells on both flanks, and received three doses of therapy every four days beginning when tumors were palpable on day 14. Tumor growth and survival were monitored until tumors reached

1000mm3 (Fig. 4A). We found that as a single agent, intratumoral CDG mediated rejection of

essentially all locally injected tumors; however, CDG failed to elicit a systemic immune

response capable of clearing distal, uninjected lesions (Fig. 4B-C; Supplementary Fig. S1).

Although systemic triple checkpoint modulation with anti-CTLA-4, anti-PD-1, and anti-4-1BB

demonstrated efficacy in this model, curing both bilateral tumors in 40% of mice, we found

that concurrent administration of CDG and immunotherapeutic antibodies, even at a single

lesion, induced complete bilateral tumor rejection in ~75% of mice (Fig. 4D). This revealed a

therapeutic synergy between CDG and our antibody cocktail, and demonstrated the abscopal

potential of CDG when combined with systemic checkpoint modulating therapies.

Figure 4: Local CDG potentiates IP checkpoint modulation against bilateral TRAMP-

C2. (A) Five to eight-week-old C57BL/6 mice were challenged with 1x106 TRAMP-C2 cells on both flanks, and were treated on days 14, 18, and 22 post-implantation with intraperitoneal (IP) checkpoint modulators and/or intratumoral (IT) CDG at either the right (R) or both right and left flank tumors (R&L). (B) Percent of mice tumor free at each flank, (C) tumor growth kinetics, and (D) overall survival over the treatment period are shown. Mice were deemed moribund when tumor volume reached 1000 mm3. Data is cumulative of n = 3 experiments with 5 mice per group. Statistical significance for survival was calculated using the Gehan-Breslow-Wilcoxon test. ns = not significant, * = P < 0.05, ** = P < 0.01, *** = P < 0.001, **** = P < 0.0001.

Supplemental Figure S1: Individual tumor growth curves; IP antibody + IT agonists. Individual tumor growth kinetics at the injected (right) and contralateral (left) flanks of mice treated with systemic (IP) antibody cocktail and/or intratumoral (IT) myeloid agonists.

3.2.2: Intratumoral Flt3L at each lesion potentiates systemic checkpoint modulation in

TRAMP-C2

Given our previous work combining checkpoint blockade with irradiated cellular

vaccines secreting Flt3L (460), and the potential for Flt3L to act as a chemotactic and

differentiation factor for DCs, we also chose to evaluate the capacity of intratumoral Flt3L to

potentiate triple checkpoint modulation in our bilateral TRAMP-C2 model (Fig. 5A). In contrast

to CDG, intratumoral Flt3L exhibited modest local efficacy as a single agent, yet achieved

curative responses in ~15% of mice (Fig. 5B-D). Similar to CDG, Flt3L synergized with triple checkpoint modulation to cure 75% of mice in this model; however, this maximal response required intratumoral administration of Flt3L at each tumor (Fig. 5D). Therefore, in optimal conditions, Flt3L and CDG were capable of potentiating triple checkpoint modulation with similar efficacy, curing up to 75% of mice with multi-focal, poorly immunogenic TRAMP-C2 tumors. When combined with systemic checkpoint modulators, however, only the STING agonist mobilized abscopal immunity.

Figure 5: Local Flt3L potentiates IP checkpoint modulation against bilateral TRAMP- C2. (A) Five to eight-week-old C57BL/6 mice were challenged with 1x106 TRAMP-C2 cells on both flanks, and were treated on days 14, 18, and 22 post-implantation with intraperitoneal (IP) checkpoint modulators and/or intratumoral (IT) Flt3L at either the right (R) or both right and left flank tumors (R&L). (B) Percent of mice tumor free at each flank, (C) tumor growth kinetics, and (D) overall survival over the treatment period are shown. Mice were deemed moribund when tumor volume reached 1000 mm3. Data is cumulative of n = 3 experiments with 5 mice per group. Statistical significance for survival was calculated using the Gehan-Breslow-Wilcoxon test. ns = not significant, * = P < 0.05, ** = P < 0.01, *** = P < 0.001, **** = P < 0.0001.

3.2.3: Intratumoral checkpoint modulation and CDG mobilizes abscopal immunity

In the clinic, individual checkpoint modulators can trigger severe, dose-limiting

immune-related adverse events in many patients (461). Colitis, pruritus, fatigue, and rash are

toxicities associated with anti-CTLA-4 and anti-PD-1. Agonistic antibodies targeting 4-1BB are

associated with rare instances of fatal hepatotoxicity (462-464). Therefore, combinatorial

strategies to target CTLA-4, PD-1, and 4-1BB must be sub-optimally dosed or administered

in a manner that limits systemic toxicity while preserving therapeutic potency. Based on a

number of studies investigating intratumoral administration of checkpoint modulators including

anti-CTLA-4 (465, 466), we asked whether intratumoral administration of low-dose combination checkpoint modulation with anti-CTLA-4, anti-PD-1, and anti-4-1BB could prime

local and systemic immunity against bilateral TRAMP-C2, and whether concomitant local

STING activation with CDG could further potentiate these effects.

Mice bearing 14-day established bilateral TRAMP-C2 tumors received three

intratumoral injections of therapy only at the right flank tumor four days apart (Fig. 6A). In this

setting, intratumoral low-dose anti-CTLA-4, anti-PD-1, and anti-4-1BB mediated regression of

~50% of injected tumors, but failed to engender systemic immunity against distal lesions (Fig.

6B-D; Supplementary Fig. S2). Concurrent administration of CDG augmented the abscopal

potential of local checkpoint modulation, however, promoting abscopal regression of distal

lesions and bilateral tumor rejection in up to 52% of mice. Although not statistically significant

with the number of mice tested, we observed a trend toward a dose-dependent increase in

efficacy with additional CDG doses in potentiating local checkpoint modulation (Fig. 6D).

Thus, we conclude that CDG can enhance the curative abscopal potential of intratumoral anti-

CTLA-4, anti-PD-1, and anti-4-1BB against bilateral TRAMP-C2 in a dose-dependent manner.

Figure 6: Intratumoral combination CDG and checkpoint modulation mobilizes abscopal responses against distal TRAMP-C2. (A) Five to eight-week-old C57BL/6 mice were challenged with 1x106 TRAMP-C2 cells on both flanks, and were treated on days 14, 18, and 22 post-implantation with intratumoral CDG and/or checkpoint modulators at the right flank tumor only. (B) Percent of mice tumor free at each flank, (C) tumor growth kinetics, and (D) overall survival over the treatment period are shown. Mice were deemed moribund when tumor volume reached 1000 mm3. Data is cumulative of n = 3-4 experiments with 5-10 mice per group. Statistical significance for survival was calculated using the Gehan-Breslow-Wilcoxon test. ns = not significant, * = P < 0.05, ** = P < 0.01, *** = P < 0.001, **** = P < 0.0001.

Supplemental Figure S2: Individual tumor growth curves; IT antibody + IT agonists. Individual tumor growth kinetics at the injected (right) and contralateral (left) flanks of mice treated with intratumoral (IT) antibody cocktail and/or IT myeloid agonists.

3.2.4: Intratumoral combination therapy with CDG elicits skin pathology

In the course of the experiments described above, we observed an inflammatory reaction provoked at the injection site in mice receiving intratumoral administration of CDG together with anti-CTLA-4, anti-PD-1, and anti-4-1BB. Within two days of the final therapeutic injection, we observed accumulation of dark, bloody fluid in and around the tumor site that

yielded to ulceration, scabbing, and eventual resolution after 2-3 weeks. This pathology required concomitant local administration of CDG and checkpoint modulators, as no ulceration occurred following CDG or checkpoint modulation therapy alone. This effect was dose-

dependent with respect to the number of CDG doses, with mild inflammation and no ulceration

observed in mice that received one or two CDG doses (Fig. 7A). Tumor size at the time of therapy may also play a role in determining the extent of ulceration, as injection of larger, 28- day established TRAMP-C2 tumors resulted in reduced ulceration generally localized to the

tumor core and not evident in the surrounding skin (Fig. 7B). Given that three concurrent doses of CDG with intratumoral anti-CTLA-4, anti-PD-1, and anti-4-1BB was associated with ulcerative pathology at the injection site yet delivered little survival benefit over one or two concurrent CDG doses (Fig. 6D), it may be prudent to limit STING agonist dosing in the combination setting in order to maximize therapeutic benefit while minimizing adverse event frequency and severity.

Figure 7: Intratumoral CDG with checkpoint modulation invokes dose-dependent and tumor size-dependent ulcerative pathology at the tumor site. (A) Photographs of representative mice treated as described in Fig. 3 were taken six days following final intratumoral injection (Day 28), when ulceration is most intense, and two weeks later during the resolution phase (Day 42). (B) Mice bearing either 14-day or 28-day established TRAMP-C2 tumors were given three intratumoral injections of CDG with anti-CTLA-4, anti- PD-1, and anti-4-1BB every four days for a total of three injections, according to normal protocol. Photographs of ulceration were taken 2 (Day 24) and 6 (Day 28) days following final intratumoral injection.

3.2.5: Flt3L with checkpoint modulation failed to generate abscopal immunity

Using the treatment schedule described in Fig. 8A, we evaluated whether local Flt3L could mobilize abscopal immunity against TRAMP-C2 when combined with intratumoral triple checkpoint modulation. In contrast to CDG, we found that Flt3L failed to synergize with local checkpoint modulation in mediating systemic tumor immunity, delivering no survival benefit over anti-CTLA-4, anti-PD-1, and anti4-1BB alone (Fig. 8B-D). Given the possibility that the therapeutic potency of our human Flt3L-Ig protein could be limited by sub-optimal receptor binding affinity and generation of neutralizing antibodies, we tested a saturating dose of 200μg per injection (Fig. 8B-D; Flt3L 3x “HIGH”), and found that, although local efficacy against the injected tumor was increased, abscopal immunity failed to improve. Of all mice treated we noted that the incidence of relapse was higher in groups that received intratumoral Flt3L compared to groups treated with CDG, although the results were not statistically significant, and time to relapse was not disparate between these groups (Supplementary Fig. S3).

Although these data suggest that in situ vaccination with Flt3L was inferior to CDG in the context of checkpoint modulation, we wished to further confirm that this discrepancy was a generalizable phenomenon and not dependent on tumor size at the time of injection. To address this, we asymmetrically implanted mice with TRAMP-C2 to obtain 28-day established tumors (~200mm3) on the right flank and 14-day established tumors (~20mm3) on the left flank before beginning local treatment (Supplementary Fig. S4A). Though the long engraftment period led to variability in both tumor size at the right flank and engraftment efficiency of distal tumors by the time of treatment, we were able to observe more effective local control of large

TRAMP-C2 tumors following in situ delivery of combination checkpoint modulation and CDG in comparison to local checkpoint antibodies and Flt3L. Despite the inferior regression of the injected lesion, Flt3L did show an equivalent capacity to 3xCDG to promote abscopal responses targeting the smaller, uninjected tumor on the opposite flank (Supplementary Fig. 87

S4B). Together, we conclude that CDG is superior to Flt3L in potentiating local checkpoint modulation to induce both local control and to mobilize abscopal effects against TRAMP-C2 tumors.

Figure 8: Intratumoral combination Flt3L and checkpoint modulation does not mobilize abscopal responses against distal TRAMP-C2. (A) Five to eight-week-old C57BL/6 mice were challenged with 1x106 TRAMP-C2 cells on both flanks, and were treated on days 14, 18, and 22 post-implantation with intratumoral Flt3L and/or checkpoint modulators in the right flank only. (B) Percent of mice tumor free at each flank, (C) tumor growth kinetics, and (D) overall survival over the treatment period are shown. Mice were deemed moribund when tumor volume reached 1000 mm3. Data is cumulative of n = 3-4 experiments with 5 mice per group. Statistical significance for survival was calculated using the Gehan-Breslow-Wilcoxon test. ns = not significant, * = P < 0.05, ** = P < 0.01, *** = P < 0.001, **** = P < 0.0001.

Supplemental Figure S3: Tumor relapse rates. Overall comparison of tumor relapse rates between mice exposed to CDG vs Flt3L treatments. Relapse was defined as any tumor that initially grew to at least 25mm3 in volume that regressed below 25mm3, then relapsed to grow again above 25mm3. (A) Number of relapses across all mice treated with CDG vs Flt3L. (B) Comparison of relapse rate between CDG-treated and Flt3L-treated mice. Relapse rate is defined elapsed time between the day of regression (when initial tumor shrunk below 25mm3) and the day at which the relapsed tumor mass passed the indicated volume. (C) Raw data used to calculate relapse rates. Statistical significance was calculated using Student’s T test.

Supplemental Figure S4: Asymmetric TRAMP-C2 tumor growth kinetics. (A) To obtain large (~200mm3) right flank tumors for IT injection, mice were implanted on day -14 with 1x106 TRAMP-C2 on the right flank, were implanted on day 0 with 1x106 TRAMP-C2 on the left flank, and treatment occurred on days 14, 18, and 22. (B) Tumor growth kinetics on the injected right flank and uninjected left flank are shown. Due to some variation in tumor size at the right flank during the long 28-day engraftment period, below is shown normalized tumor growth as a percentage of tumor size at day 14.

3.2.6: High parameter analysis of TRAMP-C2 immune contexture in response to local therapy

To further understand the biological mechanisms driving response to intratumoral therapy at both injected and distal TRAMP-C2 tumors, as well as the differential capacity of

CDG and Flt3L to potentiate local checkpoint modulation, we utilized 18-color flow cytometry to phenotype and functionally characterize the immune composition of bilateral TRAMP-C2 tumors following intratumoral therapy. As local CDG injection of 14-day established TRAMP-

C2 mediates rapid, complete regression leaving minimal tumor tissue for analysis, we allowed tumors to establish for 31 days before initiating therapy, performing intratumoral injections on right flank tumors on days 31, 35, and 39, and harvesting all tumors on day 42 (Supplementary

Fig. S3).

At baseline, TRAMP-C2 tumors are dominated by a dense myeloid infiltrate consisting of CD11b+F4/80+Gr-1- tumor associated macrophages (TAM) and CD11b+Gr-1+ myeloid

derived suppressor cells (MDSC; Fig. 9B). T cells constitute only 5% of all CD45+ cells, and are composed of roughly even proportions of CD8+ cytolytic T lymphocytes (CTL),

CD4+FoxP3- effector T lymphocytes (Teff) and CD4+FoxP3+ regulatory T cells (Treg; Fig. 9C).

Intratumoral CDG administration enhanced CD8+ T-cell infiltration of injected tumors, and

caused local accumulation of CD11b+Gr-1+F4/80neg granulocytes, likely indicating neutrophil

infiltration (Fig. 9D; Supplementary Fig. 6). In contrast, intratumoral Flt3L elicits increased

CD11c+CD11b- dendritic cell (DC) accumulation, yet triggered expansion of Treg in both injected and uninjected tumors (Fig. 9B,C). Intratumoral checkpoint modulation enhanced the proportion of T cells infiltrating TRAMP-C2 tumors, and diminished the Treg pool, likely through antibody-dependent cellular cytotoxic (ADCC) depletion of αCTLA-4 coated Treg

(467). Concurrent administration of CDG and checkpoint modulating antibodies further enhanced CD8+ T-cell accumulation at each tumor, resulting in favorable CD8:Treg ratios that

were superior to those seen in response to combination therapy with Flt3L (Fig. 9E). Using 92

tetramer staining to track the known immunodominant SPAS-1 neoepitope expressed in

TRAMP-C2 (468), we found that combination therapy with CDG enhanced the density of

SPAS-reactive CD8+ T-cells at the injected tumor and exhibited a trending capacity to mobilize

these cells to the distal tumor, which was not observed in mice receiving Flt3L therapy (Fig.

9F). Local CDG synergized with checkpoint modulation to reduce the percentage of SPAS- specific T cells within the total pool of infiltrating CD8+ T-cells, suggesting that this therapeutic

approach may diversify the T-cell receptor repertoire to more effectively target subdominant

epitopes in addition to SPAS-1 compared to checkpoint modulation alone or combination

therapy with Flt3L (Fig. 9G). These data indicate a mechanistic basis underlying the discrepancy in efficacy between local combination therapy incorporating CDG and Flt3L wherein CDG elicits numerous and diverse CD8+ T-cell responses against injected and

uninjected tumors, whereas the efficacy of Flt3L is limited by the tendency to promote CD4+

T-cell responses.

The dense myeloid infiltrate characteristic of TRAMP-C2 tumors poses a barrier to the

function of infiltrating cytolytic T cells. Even in this context, we found that local priming with

3XCDG promoted improved CD8:TAM ratios relative to the untreated animals (Fig. 9H). This

was also evident for 1xCDG vs Flt3L for CD8:MDSC ratios; however, this data for the 3xCDG

group was confounded by the large neutrophil infiltration in response to the skin inflammation

shown in Fig. 7, and our inability to distinguish these neutrophils from PMN-MDSC using the

antibodies in our panel (Supplementary Fig. S5A). Overall though, few, if any, tumors reached

ratios at which CD8+ T-cells are expected to fully escape myeloid suppression. Therefore, we

hypothesize that intratumoral delivery of myeloid agonists can reprogram tumor myeloid

populations to reduce their immunosuppressive activity, creating a microenvironment that is

amenable to T-cell effector function despite unfavorable relative cellular ratios. Supporting this

hypothesis, we observed downregulation of the M2 macrophage marker CD206 on TAM in

tumors receiving CDG concurrently with all three doses of anti-CTLA-4, anti-PD-1, and anti-

4-1BB, potentially indicating repolarization of TAM in response to CDG combined with checkpoint modulation (Fig. 9I). Although combination therapy with Flt3L promoteed DC expansion and a reduction in MDSC populations, Flt3L failed to induce CD206 downregulation on TAM, suggesting this effect is specific to CDG. We conclude, therefore, that relative to

Flt3L, CDG better potentiates intratumoral anti-CTLA-4, anti-PD-1, and anti-4-1BB activity by increasing infiltration and diversification of tumor-specific CD8+ T cells at local and distal

tumors, enhancing CD8:Treg ratios, and by reprogramming suppressive TAM to a pro-

inflammatory phenotype.

Figure 9: Analysis of the TRAMP-C2 immune infiltrate at local and distal tumors following intratumoral therapy at a single lesion. (A) To obtain cell numbers sufficient for flow analysis, TRAMP-C2 tumors were established for 31 days, exposed to three doses of intratumoral therapy at the right flank, then were harvested for multi-parameter flow cytometry as described in Methods. See Supplementary materials for antibodies used and gating strategy. (B) Relative proportions of unique cell subsets within the CD45+ immune infiltrate, normalized to the total number of gated cells. (C) Relative proportions of T-cell subsets within the CD3+ lymphocyte infiltrate, normalized to the total number of gated T cells. (D) Relative proportions of MDSC subsets within the CD11b+Gr-1+ immune infiltrate. (E) Calculated ratios of CD8+ T-cells to Treg, (F) density of SPAS-1 tetramer+ CD8+ T-cells, (G) percentage of SPAS-1 reactive T cells within all infiltrating CD8 T-cells, and ratio of CD8+ T-cells to (H) macrophages are shown. (I) Percentage of macrophages expressing the M2 macrophage marker CD206. Data is cumulative of 2 independent experiments with 5 mice per group. Statistical significance was calculated using Student’s T test. ns = not significant, * = P < 0.05, ** = P < 0.01, *** = P < 0.001, **** = P < 0.0001.

Supplemental Figure S5: TIL flow analysis experimental setup and gating strategy. As described in the treatment schedule, 31-day established subcutaneous TRAMP-C2 were exposed to 3 IT injections 4 days apart, and tumors were harvested for flow analysis

3 days following the final therapy. Sample gating strategy displays all analyzed cell populations from a fully-stained naïve spleen sample. 18-color compensation was performed using single stained spleen controls and the compensation matrix from TreeStar FlowJo ® version 7.6.5, and all gating was performed in FlowJo ® version 10.1.

Supplemental Figure S6: TIL Flow Population Quantifications. Cell populations harvested from TRAMP-C2 tumors as described in Supplemental Fig S6 were enumerated using tumor suspension cell counts and relative percentages defined by flow cytometry, and densities were calculated by dividing population cell count by tumor volume. (A) Ratio of CD8 T cells to CD11b+Gr-1+ MDSC. Individual cell populations shown are (B) CD8 cytotoxic T cells (CD45+ CD3+ CD8+), (C) CD4 effector T cells (CD45+ CD3+ CD4+ FoxP3- ), (D) CD4 regulatory T cells (CD45+ CD3+ CD4+ FoxP3+), (E) myeloid derived suppressor cells (CD45+ CD3- CD11b+ Gr-1+), (F) dendritic cells (CD45+ CD3- CD11c+ CD11b-), and (G) tumor associated macrophages (CD45+ CD3- CD11b+ F4/80+ Gr-1-). Data is cumulative of 2 independent experiments with 5 mice per group. Statistical significance was calculated using Student’s T test. ns = not significant, * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

3.2.6: Summary

Simultaneous manipulation of multiple, non-redundant immune regulatory systems

can activate potent antitumor immune responses in mouse and man. Translation of this

approach, however, has been limited by cumulative inflammatory side effects which limit the

tolerability of high-order combination therapies in patients. At the time of this writing, the only

combination immunotherapy with FDA approval consists of ipilimumab (Yervoy®; anti-CTLA-

4) and nivolumab (Opdivo®; anti-PD-1) for patients with Braf V600 wild-type unresectable or metastatic melanoma. Though effective and tolerable in this indication, a number of solid tumors, including castration-resistant prostate cancer, remain resistant to these therapies. In this chapter, I report that an optimized cocktail of immunomodulators engaging both innate and adaptive immunity mediates efficient rejection of poorly immunogenic murine prostate cancer. Although local stimulation of the STING or Flt3 pathways can complement checkpoint modulating antibodies targeting CTLA-4, PD-1, and 4-1BB to delay tumor growth and enhance survival of mice with multi-focal TRAMP-C2, I find that the STING agonist CDG mobilizes abscopal immunity more efficiently than Flt3L. In an effort to mitigate the toxicity of these high- order combinations, I report that intratumoral injection of low-dose checkpoint modulators

together with CDG, but not Flt3L, is sufficient to induce curative abscopal immunity in ~50%

of mice. I identify a potential mechanistic basis for the discrepancy in synergy between checkpoint modulators and CDG vs Flt3L, as CDG promotes CD8 skewing, infiltration of tumor specific T cells, and enhanced intratumoral ratios of CD8+ T cells to Treg, TAM, and MDSC, while Flt3L facilitates expansion of CD4 T cells, including Treg, in both local and distal tumors.

Combination therapy with CDG may reduce the suppressive capacity of TAM in TRAMP-C2 tumors, as I observe downregulation of the M2 macrophage marker CD206 in response to

this therapy. These data suggest STING-mediated myeloid programming can potentiate

intratumoral T-cell checkpoint modulation to render multi-focal prostate cancer sensitive to 99

immunotherapy.

100

Chapter 4: Intratumoral Delivery of a Novel STING Agonist Repolarizes

Suppressive Myeloid Cells and Sensitizes Pancreatic Ductal Adenocarcinoma

to Checkpoint Blockade

4.1: Introduction

Pancreatic ductal adenocarcinoma (PDAC) is a lethal malignancy whose defining characteristics render it resistant to available therapies. Over 80% of PDAC patients present with metastatic disease precluding surgical resection, and relapse occurs in the majority of early-stage surgical candidates within one year (68, 118). First-line chemotherapy regimens

consisting of gemcitabine and albumin-bound paclitaxel (Abraxane®) or a cocktail of

leucovorin, fluorouracil, irinotecan, and oxaliplatin (FOLFIRINOX) extend median overall

survival up to 11 months, yet 5-year survival rates remain dismal (~7%) and treatment related

toxicities can be severe (120, 121). The PDAC genome bears few actionable targets, as the

characteristic genetic alterations common to nearly all PDAC patients (KRAS, TP53,

CDKN2A, and SMAD4) are considered “undruggable” by current targeted therapies. A

uniquely desmoplastic tumor microenvironment also contributes to the refractory nature of

PDAC tumors. High interstitial pressures from densely-deposited extracellular matrix limits drug perfusion into the tumor bed, and restriction of the sparse tumor vasculature creates regions of deep hypoxia (137, 138). While oncogenic Kras drives metabolic adaptions to sustain tumor growth under these conditions (141-144), many anti-tumor components of the

PDAC stroma, chiefly cytotoxic CD8 T cells, are excluded or rendered locally dysfunctional by this hostile environment (146). As a result, PDAC tumors are generally considered immunologically “cold”, and immunotherapies targeting T cell checkpoints CTLA-4 or PD-1 have failed in clinical evaluation in this indication (162, 164, 165, 469).

However, contrary to its perception as an immunologically silent tumor, PDAC relies on numerous intimate interactions with the host immune system from its earliest discernable 101

pre-invasive stages through metastatic disease (146). Early activation of oncogenic Kras

drives tumor-intrinsic production of colony stimulating factors including GM-CSF and CSF-1,

which stimulates myelopoiesis and recruits monocytes that undergo polarization to an “M2-

like” macrophage phenotype in situ (151, 330). Further release of Kras-driven cytokines

including IL-6 can induce aberrant activation of immature myeloid precursors, which

accumulate as myeloid derived suppressor cells (MDSC) within tumors and lymphoid organs

(147-149). Once recruited, these myeloid cells can participate in crosstalk with cancer-

associated fibroblasts (CAF) that construct fibrotic barriers that interfere with CD8 T cell

infiltration (57, 151). Moreover, any CD8 T cells which gain entry to the tumor bed are subject

to the myriad mechanisms of immunosuppression by both tumor associated macrophages

(TAM) and MDSC, which include local nutrient deprivation, secretion of suppressive

cytokines, release of reactive oxygen and nitrogen species, engagement of inhibitory T-cell

checkpoint receptors, and promotion of FoxP3-driven CD4 regulatory T cell (Treg)

phenoconversion (470). Thus, an early proclivity towards myeloid recruitment and polarization

is an intrinsic property of PDAC which cloaks developing and established neoplasms from

immune sensing and rejection (146). As a result, a number of approaches are currently under

preclinical and clinical evaluation to target myeloid cells in PDAC, including interference with

myeloid recruitment and signaling through blockade of chemokine receptors CCR2 and

CXCR2 (301, 305) or key macrophage signaling molecules CSF-1R or PI3Kγ (330, 338, 339),

as well as activation of tumor macrophages through agonism of the TNF superfamily receptor

CD40 (341, 342). However, it remains to be determined which method of myeloid modulation engenders optimal conditions to sensitize PDAC tumors to immunotherapy.

Myeloid cells are canonically activated through a repertoire of pattern recognition receptors (PRR), which sense molecular patterns associated with infection or cellular stress and induce transcriptional programs associated with acute inflammation and antigen

102

presentation. Recently, stimulator of interferon genes (STING) was identified as a novel sensor of cytosolic nucleic acids that is capable of liberating potent type I interferon responses to cytosolic DNA (358). Signaling through STING, rather than any other innate PRR, was shown to be required for spontaneous immunity against immunogenic tumors, suggesting the

STING pathway may uniquely engender tumor-antagonistic phenotypes and function in tumor

myeloid populations (348). STING is specifically activated by cyclic dinucleotide (CDN)

compounds, initiating a signaling cascade through IRF3 and NFκB pathways to elaborate

inflammatory cytokines including IFN-β and TNFα, chemokines including CXCL10 and CCL5,

and surface molecules associated with antigen presentation including MHC-II, CD86, and

CD40 (360, 373, 429). Intratumoral delivery of CDN has shown preclinical efficacy across an

array of ectopically-implanted tumor models as monotherapy and in combination with

checkpoint blockade (114, 389, 429, 434, 437, 439, 440, 442), and synthetic CDN with enhanced stability and potency relative to the mammalian CDN 2’3’ cyclic guanine monophosphate-adenosine monosphosphate (cGAMP) are currently under early-phase clinical evaluation in solid tumors with superficial involvement (NCT02675439, NCT03172936,

NCT03010176). Unknown, however, remains the feasibility and efficacy of local delivery of

CDN to visceral tumors such as PDAC, as well as a comprehensive analysis of the effects of

CDN on suppressive myeloid cell phenotypes at the transcriptional, translational, and

functional levels. In the following study, we describe and characterize a novel, highly potent

STING agonist IACS-8803, and perform in depth profiling of its effects relative to other known

CDN on in vitro polarized populations of MDSC and M2-like macrophages of murine and

human origin. Furthermore, we demonstrate that CDN can be locally delivered within

orthotopically-implanted pancreatic lesions, and provide preclinical evidence that local STING

activation can potentiate the effects of checkpoint blockade immunotherapy in a highly

aggressive, refractory model of PDAC.

103

4.2: Results

4.2.1: In vitro characterization of IACS-8803, a novel highly potent STING agonist

STING activation via cyclic dinucleotide (CDN) compounds is one of the most efficient means of generating a type I interferon response in both hematopoietic and non- hematopoietic cells, making CDNs an attractive therapeutic modality for tumors which lack a

T cell inflamed signature. A number of CDNs are produced naturally, including cyclic di- guanine monophosphate (CDG) and cyclic di-adenosine monophosphate (CDA), which are broadly conserved second messenger molecules in bacteria (360). Mammalian cells utilize the cytoplasmic DNA sensor cGAS to synthesize cyclic guanine-monophosphate adenosine monophosphate (cGAMP), which possesses higher affinity for STING due to a non-canonical

3’5’-2’5’ phosphodiester “mixed” linkage that results in deeper binding within the STING CDN- binding pocket and more potent induction of downstream signaling relative to bacterial CDN

(364-366). Recently, synthetic CDNs have been designed bearing 2’3’ mixed linkages in addition to bisphosphothionate substitutions that inhibit CDN enzymatic degradation by phosphodiesterases such as ENPP1 (431). The molecule ML-RR-S2-CDA (ML-RR) bears these modifications and is the first CDN under clinical evaluation in cancer (ADU-S100;

Novartis) (429, 437). We designed and synthesized a number of novel CDNs containing additional proprietary alterations to ML-RR, and evaluated these CDNs for the capacity to activate STING in human THP-1 Dual reporter monocytes and murine J774 Dual reporter macrophages (Invivogen). We identified a novel CDN IACS-8803 that induces downstream

IRF3 and NFκB signaling with enhanced potency relative to ML-RR, with a specific advantage at reduced concentrations below 1µg/mL (Figure 10A-B, Supplementary S7A). To confirm these data, we quantified STING-mediated release of Type I IFN by parental THP-1 monocytes and murine splenocytes by ELISA, and found that stimulation with 8803 induced

104

higher IFN-β release relative to ML-RR or cGAMP in both conditions (Fig 10C, Supplemental

Fig S7B). STING-deficient splenocytes from STINGgt/gt mice failed to induce IFN-β in response to 8803, suggesting the enhanced potency of 8803 is not due to promiscuous activation of non-STING targets (Supplemental Fig S7B). Structurally, 8803 is a 2’3’-2’5’ mixed linkage

CDA molecule with identical bisphosphothioate substitutions of R,R chirality as found in ML-

RR-S2-CDA. However, IACS-8803 contains an additional fluorine substitution in place of the

2’ hydroxyl group on the 3’ linked ribose residue (Fig 10D). These data demonstrate 8803 is a highly potent novel CDN capable of efficiently activating STING in both murine and human myeloid cell lines in vitro.

A number of unique germline variants of the human STING gene have been described, with the HAQ and R232H alleles each being found in 10-20% of the human population (430).

As these isoforms may possess unique CDN binding characteristics compared to wild-type

STING, we next evaluated the capacity of 8803 to activate the most common human STING variants. We utilized STING-deficient HEK-Blue reporter cells stably transduced with WT,

HAQ, R232H, or G230A STING alleles, and measured induction of STING signaling in response to 8803. We found 8803 sufficient to robustly activate each isoform (Fig 10E). While

ML-RR can activate the HAQ and G230A alleles as previously described, we found ML-RR incapable of activating the R232H allele at the concentrations evaluated, suggesting 8803 may be more broadly effective than ML-RR in 13.7% of the human population known to express the R232H isoform of STING.

105

Figure 10: In vitro characterization of IACS-8803. (A) 1x105 THP-1 Dual and 5x104 J774 Dual reporter cells (Invivogen) were stimulated for 20 hours with 1μg/mL of indicated CDN. IRF3 activity was measured by luciferase assay according to manufacturer’s instructions. (B) CDN titration assay using the THP-1 Dual reporter cell line as in (A). (C) 1x105 THP-1 parental cells were plated in flat-bottom 96 well plates and were incubated with 10μg/mL indicated CDN. After 72 hours, IFN-β levels in the supernatant were measured by ELISA (PBL Assay Science). (D) Chemical structures of 2’3’ cGAMP, ML-RR-S2-CDA, and IACS- 8803. (E) 5x104 HEK-Blue reporter cells expressing the indicated human STING allele were incubated with 100μg/mL of the corresponding CDN for 24 hours. IRF3/9 activity was measured by SEAP assay according to manufacturer’s instructions. Statistical significance was calculated using Student’s T test. ns = not significant, * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

106

Supplemental Figure S7: Extended in vitro characterization of IACS-8803. (A) 1x105 THP-1 Dual reporter cells (Invivogen) were stimulated for 20 hours with a range of indicated CDN. NF-κB activity was measured by SEAP assay according to manufacturer’s instructions. (B) Splenocytes were harvested from WT C57BL/6 or STINGGt/Gt mice and 2x105 cells were plated in flat bottom 96-well plates in the presence of 10μg/mL CDN. After 72 hours, IFN-β levels in the supernatants were measured by ELISA (PBL Assay Science). Statistical significance was calculated using Student’s T test. ns = not significant, * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

107

4.2.2: In vivo characterization of 8803 relative to known CDNs

To determine whether the superior potency of 8803 in in vitro cell lines translates to enhanced in vivo anti-tumor efficacy compared to other known CDNs, we performed

intratumoral injection of CDNs into subcutaneously-implanted PDAC tumors using a

Kras+/G12DTP53+/R172HPdx1-Cre (KPC) derived cell line mT4-2D (445). Tumors were harvested

following three successive injections with 5µg of indicated CDN for mass measurement and

analysis by multi-parameter flow cytometry, probing for compositional and functional

modulation of the tumor immune microenvironment (Fig 11A). We find that relative to PBS

vehicle injection, CDG and cGAMP at this dose have minor effects on tumor mass. However,

ML-RR and 8803 induce regression of mT4-2D tumors, with 8803 delivering superior efficacy

over ML-RR (Fig 11B). Local delivery of 8803 triggers expansion of the CD45+ immune

infiltrate (Fig 11C), which is dominated by recruitment of CD11b+Ly6G+Ly6CmidF4/80- granulocytes that likely consist largely of inflammatory neutrophils (Fig 11D). This granulomatous response occurs at similar or reduced levels in tumors exposed to ML-RR or

CDG, but interestingly is not as prominent in cGAMP-treated mT4-2D tumors. Nearly all other analyzed immune cell populations exist at higher densities following injection with 8803 relative to vehicle (Fig 11E), however among lymphocytes, particular expansion occurs within the CD8 T cell compartment, with these cells exhibiting increased expression of functional markers including Ki67 and Ly6C (Fig 11E-G). In the myeloid compartment, expansion of

CD11b+Ly6C+Ly6G-F4/80mid/- monocytes occurs in response to 8803, with resulting monocytes bearing reduced expression of the functional M2-like marker Arginase (Fig 11E,I).

Tumor associated macrophages (TAM) as defined by CD11b+Ly6C-Ly6G-F4/80+ marker

expression do not significantly expand in response to 8803, yet undergo phenotypic

conversion through downregulation of CD206 (Fig 1H). Together, these data describe in vitro

108

stimulatory capacity and in vivo therapeutic effects of a novel STING agonist IACS-8803, and detail its high relative potency compared to natural STING ligands CDG and cGAMP and the clinical compound ML-RR-S2-CDA.

109

Figure 11: In vivo characterization of 8803 in subcutaneous KPC-derived PDAC. (A) Mice received subcutaneous injection of 2.5x105 mT4-2D cells in 30% matrigel, then were injected intratumorally with 5μg CDN on days 15, 18, and 21 before tumor harvest on day 23. Tumors were massed, counted, and processed for flow cytometry analysis as described in methods. Data shown represent (B) tumor mass, (C) CD45+ infiltration, (D) overall composition of analyzed CD45+ cells, (E) cellular densities, (F) Ki67 and (G) Ly6C expression on tumor CD8 T cells, (H) normalized CD206 expression on macrophages, and (I) arginase expression in MO-MDSC. Data are cumulative of 2 independent experiments with 5 mice per group. Statistical significance was calculated using Student’s T test. ns = not significant, * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

110

4.2.3: Profiling effects of CDNs on murine bone marrow-derived MDSCs

STING agonists are known to act on a number of different cell types within the tumor microenvironment, including dendritic cells, macrophages, endothelial cells, and radioresistant stromal cells (414, 429, 443), and induction of CD8 T cell priming by activated professional antigen presenting cells (APCs) is thought to be central to the efficacy of CDNs.

Phenotypic “M2 to M1” conversion of macrophages following STING activation has been documented (442, 471), suggesting repolarization of suppressive myeloid populations may contribute to the efficacy of CDNs in tumors with high myeloid infiltrates. However, it is currently unknown whether CDNs modulate the phenotype or function of myeloid derived suppressor cells (MDSC), which are numerous and highly immunosuppressive in many tumor types. Therefore, we sought to describe the effects of CDNs on MDSC signaling and T cell suppressive function.

We chose to study in vitro differentiated populations of bone marrow-derived MDSCs

(BM-MDSCs) due to the sensitivity of tumor MDSCs to ex vivo culture and the lack of definitive phenotypic markers to differentiate bona fide MDSCs from mature neutrophils or monocytes recruited following intratumoral CDN injection. Culture of murine bone marrow aspirates with recombinant GM-CSF and IL-6 for 4 days yields consistent populations of both polymorphonuclear (PMN-) and monocytic (MO-) MDSCs as defined by expression of CD11b,

Ly6G, and/or Ly6C (Fig 12A, Supplemental Fig S8) (199). In these cultures, PMN-MDSCs predominate numerically over MO-MDSC at a ~2:1 ratio, which reflects the relative abundance of these populations observed in many tumor models (Supplemental Fig S8). We found this mixed population of MDSCs capable of efficiently suppressing proliferation of naïve CFSE- labeled splenic CD8 T cells in the presence of polyclonal αCD3/αCD28 stimulation, indicating they are functionally valid MDSC as has been previously published (Fig 12B). To evaluate

111

whether CDN are capable of modulating the suppressive function of MDSCs, we harvested

BM-MDSCs following 4-day in vitro differentiation and stimulated them with increasing

concentrations of 8803 for 1-2 days, then evaluated their suppressive function by T cell suppression assay (Fig 12A). In comparison to unstimulated BM-MDSCs, 8803-treated BM-

MDSCs exhibit reduced suppressive function in a dose-dependent manner, with maximum repolarization occurring at 10µg/mL 8803 (Fig 12B). These data suggest STING-activating

CDNs can reduce the suppressive function of BM-MDSCs in a dose-dependent manner. To evaluate the relative capacity of different STING agonists to modulate MDSC function, we stimulated BM-MDSCs with 2.5µg/mL of indicated CDN, then evaluated suppressive function by T cell suppression assay. Under these conditions, CDG has no observable effects on BM-

MDSC function, while cGAMP stimulation yields BM-MDSCs with a trending decrease in suppressive capacity at a 1:1 MDSC:T cell ratio. In contrast, both ML-RR and 8803 mediate significant reversal of MDSC suppression, with a majority of T cells escaping suppression by treated BM-MDSC at the equivalent 1:1 ratio (Fig 12C). Compared to ML-RR, 8803-treated

BM-MDSCs exhibit a trending, but not significant reduction in suppressive function. Together, these data indicate the CDN can reverse the T-cell suppressive activity of in vitro differentiated

MDSC populations, and synthetic dinucleotides 8803 and ML-RR-CDA are superior to natural

STING ligands at mediating these effects.

112

Figure 12: STING activation reduces the suppressive capacity of bone marrow- derived MDSC. (A) Generation of BM-MDSC and repolarization protocol as described in detail in methods. (B) Evaluati on of MDSC suppression by CD8 T cell suppression assay following MDSC repolarization using indicated concentration of 8803. (C) Evaluation of MDSC suppression by CD8 T cell suppression assay following repolarization with 2.5μg/mL indicated CDN. Data are representative of at least two independent experiments. Statistical significance was calculated using Student’s T test. ns = not significant, * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

113

Supplemental Figure S8: MDSC subsets in unstimulated BM-MDSC cultures. Bone marrow harvested from WT C57BL/6 mice was cultured in the presence of GM-CSF and IL-6 for 4 days to induce differentiation of suppressive BM-MDSC as described in methods. Flow cytometric analysis of untreated BM-MDSC revel ~95% expression of CD11b, and a 2:1 ratio of Ly6G+Ly6Cmid PMN-MDSC (49.4%) to Ly6C+Ly6G- MO-MDSC (24.0%).

114

4.2.4: Profiling effects of CDNs on human M2c-polarized macrophages

A number of studies indicate a capacity for STING-activating CDN to facilitate “M2 to

M1-like” phenotypic conversion of in vitro-differentiated and tumor associated macrophages.

However, in depth profiling of CDN-mediated macrophage repolarization in controlled in vitro

settings has not been reported. Additionally, analysis of myeloid polarization by CDNs has

predominately focused on murine macrophage populations, leaving little known about the

potential for STING agonists to modulate macrophages of human origin. We generated

human macrophages through magnetic bead isolation of healthy donor peripheral blood

mononuclear cell (PBMC) CD14+ monocytes, which were cultured in the presence of M-CSF for 6 days to induce macrophage differentiation. Polarization to an “M2c-like” phenotype was induced through addition of IL-10 and TGF-β cytokines for 2 days as previously described

(447), yielding macrophages with high expression of CD163, CD206, PD-L2, and IRF4.

Following differentiation and polarization, resulting M2c-like macrophages were exposed to

CDNs at 10 µg/mL concentration for 3 days in the presence of supportive M-CSF (Fig 13A).

We found unstimulated M2c macrophages to exhibit classical M2-like morphology

characterized by firmly adherent, elongated cellular architecture (Fig 13 B). In contrast,

addition of CDNs results in a shift in morphology towards a less-adherent, “fried egg” M1-like

shape (343, 472). Interestingly, the degree to which this morphological shift occurs is

correlated with CDN potency, which can be measured roughly by side scatter characteristics

upon analysis by flow cytometry (Fig 13C). Exposure to ML-RR or 8803 induce potent

upregulation of M1-like markers CD86, HLA-DR, and IL-6 relative to unstimulated, CDG-

treated, or cGAMP-treated M2c macrophages, concomitant with downregulation of classical

M2-like markers CD206 and CD163. Furthermore, 8803 is superior to ML-RR in inducing

expression of CD80 and PD-L1, which are generally unresponsive to natural CDNs at the 115

dose examined (Fig 13C). These markers are expressed at lower levels than CD86 or IL-6, thus we speculate that their expression may be a result of secondary effects from CDN- induced primary secretion of IFNβ, IFNγ, or TNFα cytokines. In addition, we observe a profound proliferative block in M2c macrophages responding to ML-RR or 8803 as measured by Ki67 downregulation. Interestingly, proinflammatory triggers such as LPS were recently shown to block proliferation controlled by Myc, which also governs M2-like polarization in macrophages (473-475). Thus, it is tempting to speculate that potent STING activation by ML-

RR or 8803 counteracts Myc-dependent proliferation and suppressive polarization. Together, these data indicate that CDNs are capable of mediating direct inflammatory repolarization of human M2c-like macrophages at morphological and phenotypic levels.

116

Figure 13: STING activation phenotypically repolarizes human M2c macrophages. (A) Protocol for generation of M2c polarized macrophages from CD14+ PBMC-derived monocytes and subsequent CDN stimulation, as detailed in methods. (B) Representative images of morphological changes in M2c macrophages following 3 day exposure to indicated CDNs. (C) Analysis of macrophage phenotypic polarization following 3 day CDN stimulation by flow cytometry. Data are cumulative of two independent experiments with 3 unique donors each. Statistical significance was calculated using Student’s T test. ns = not significant, * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

117

4.2.5: Combination therapy with 8803 and checkpoint blockade cures multi-focal mT4-LS

PDAC.

Given the capacity of 8803 to induce phenotypic and functional repolarization of

suppressive myeloid cells – which pose a major barrier to the efficacy of immunotherapy in

PDAC – we hypothesized that intratumoral delivery of 8803 could remodel the tumor

microenvironment to permit CD8 T cell cytotoxicity, creating an opportunity for checkpoint

blockade to be of benefit in PDAC. Furthermore, we tested whether local STING activation in

the context of checkpoint blockade could promote systemic antitumor immunity against distal

uninjected PDAC lesions, as we previously demonstrated in a bilateral model of TRAMP-C2

prostate cancer. For this, we utilized mT4-2D, a C57BL/6 syngeneic PDAC cell line derived

from Kras+/G12DTP53+/R172HPdx1-Cre (KPC) tumor organoid cultures. We found mT4-2D to

possess an extremely low mutational burden, as is common of KRAS-driven spontaneous

murine tumor models (Supplemental Fig S9A-B). Retroviral transduction of an optimized firefly luciferase gene facilitated longitudinal monitoring of orthotopically-implanted mT4-2D-

Luciferase (termed mT4-LA) using IVIS bioluminescent imaging. We found low doses of

<5x104 mT4-LA cells capable of aggressive local invasion and metastatic spread to the liver, peritoneum, and mesentery, with mice succumbing to disease within three weeks on average.

Additionally, we isolated a single cell clone from the mT4-LA population that exhibits reduced growth kinetics compared to mT4-LA in vivo following both subcutaneous and orthotopic implantation (Termed mT4-LS for its characteristically “slow” growth pattern; Supplemental

Fig S9C). We leveraged the larger therapeutic window afforded by mT4-LS to screen potential

combination approaches, and performed validation of effective therapies in the highly

aggressive orthotopic mT4-LA model.

To evaluate the efficacy of intratumoral 8803 alone and in combination with checkpoint

blockade against multi-focal PDAC, we simultaneously implanted mice with both orthotopic 118

and subcutaneous mT4-LS tumors to model a primary and metastatic lesion, respectively.

Mice received localized 8803 or PBS vehicle within the orthotopic tumor by survival surgery

on days 10 and 22 post-implantation, and received checkpoint blockade antibodies targeting

CTLA-4 or PD-1 by intraperitoneal (IP) injection on days 10, 14, and 18 (Fig 14A). We found monotherapy with αCTLA-4, αPD-1, or intra-pancreatic 8803 capable of curing 50% of mice of both lesions in this model, with combination αCTLA-4/αPD-1 not demonstrating significant benefit over individual checkpoints. However, the combination of intratumoral 8803 with concurrent checkpoint blockade cures all mice tested of both lesions, regardless of the checkpoint being targeted (Fig 14B-C). IVIS imaging reveals a rapid kinetic of tumor regression following combination therapy, with tumors falling below the level of detection within six days of initiating treatment. These data demonstrate that local STING activation by

8803 within orthotopically-implanted PDAC lesions can potentiate checkpoint blockade and induce robust curative immunity against both injected and distal uninjected mT4-LS tumors.

119

Figure 14: Intra-pancreatic 8803 potentiates checkpoint blockade to cure mice with multi-focal mT4-LS. (A) Implantation and treatment schedule. Intra-pancreatic delivery of 8803 involves survival surgery to inject 8803 directly into primary mT4-LS lesions. (B) Representative longitudinal IVIS imaging of mice treated as in (A). (C) Survival data indicating therapeutic additivity between 8803 and CTLA-4 or PD-1, or the combination of CTLA-4 and PD-1. Data are cumulative of two independent experiments with 5-10 mice per group. Statistical significance was calculated using the Log-rank Mantel-Cox test. ns = not significant, * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

120

Supplemental Figure S9: Mutational analysis and tumor growth kinetics of mT4 PDAC model. (A) Whole genome sequencing was performed on parental mT4-2D in collaboration with Colm Duffy of the Allison lab. Identified exonic nonsynonymous mutations were analyzed for predicted neoantigen epitopes using the NetMHC 4.0 algorithm. (B) Comparison of nonsynonymous mutational burden of mT4-2D relative to PancO2 (data adapted from Bhadury et al 2013) and B16-F10 (data adapted from Castle et al 2012). (C) Subcutaneous tumor growth kinetics and survival following orthotopic implantation of mT4-LA and mT4-LS cells. For subcutaneous implantation, 2.5x105 cells were implanted in 3 mice each. For orthotopic implantation, 0.35x105 cells were implanted in 30% matrigel v/v in the head of the pancreas by survival surgery in 10 mice each. Survival endpoint indicates death from tumor burden or observed overt morbidity.

121

4.2.6: Analysis of orthotopic and subcutaneous mT4 tumors following 8803 + checkpoint

therapy by flow cytometry

To understand how this combination approach achieves such rapid and robust

clearance of multi-focal PDAC, we performed multi-parameter flow cytometry on both

pancreatic and subcutaneous tumors four days following delivery of intra-pancreatic 8803. For

this early timepoint, we modified the checkpoint antibody schedule as detailed in Fig 15A, and

utilized the parental mT4-2D cell line to obtain sufficient tumor tissue for analysis. We found

local 8803 sufficient to increase CD8 T cell density within orthotopic tumors, which can be

further potentiated by αCTLA-4 or the combination of αCTLA-4/αPD-1 (Fig 15B). This

correlates with a decreased proportion of Foxp3+ Tregs in tumors exposed to 8803 and

αCTLA-4, resulting in improved CD8:Treg ratios within the pancreatic microenvironment (Fig

15C). This is expected, given the well-documented capacity for the 9H10 clone of αCTLA-4 to facilitate Treg depletion in murine tumors. In contrast, αPD-1 does not appear to drive CD8

T cells into PDAC tumors as single agent or in combination with 8803, or productively modulate tumor T cell composition to numerically favor CD8 T cells over Tregs (Fig 15C).

However, Tregs within all tumors locally-exposed to 8803 exhibit reduced expression of

CD103, a marker that has been associated with Treg effector function (Fig 15D). Thus, direct effects of 8803 may reduce the suppressive capacity of Tregs and promote CD8 T cell cytotoxicity even in αPD-1 treated mice which retain poor CD8:Treg ratios. This is substantiated by the observation that phenotypic changes within CD8 T cells associated with enhanced cytotoxicity – including expression of the cytotoxic granule Granzyme B and proliferation marker Ki67 – are observed in response to 8803 and αPD-1 (Fig 15E-F). While similar effects are noted in response to 8803 and αCTLA-4 or the combination of αCTLA-

4/αPD-1, we do observe a unique capacity for 8803 and αPD-1 to most efficiently upregulate

Ly6C expression on CD8 T cells within pancreatic tumors (Fig 15H). Interestingly, we also 122

observe analogous induction of Granzyme B in CD8 T cells within distal subcutaneous tumors in mice treated with intra-pancreatic 8803 and systemic checkpoint blockade, which indicates local 8803 can exert systemic effects on T cell function (Fig 15G). However, it is unclear whether this is bona fide “abscopal” immunity, or due to diffusion of small CDN molecules into the circulation that can induce residual STING activation within the uninjected lesion.

Downregulation of CD206 is observed on TAM from orthotopic lesions exposed to 8803 alone or together with checkpoint blockade, however the degree of downregulation is minor relative to that seen in the subcutaneous model (Fig 15I). This may be due to the transient polarizing effects of CDN in vivo following a single – rather than multiple – injections. Infiltrating DCs exhibited increased proliferation through Ki67 staining in response to 8803 and checkpoint blockade, possibly indicating expansion of the DC compartment by 8803 (Fig 15J). Together, these data describe early modulation of the PDAC TME by intra-pancreatic 8803 alone and together with checkpoint blockade, and suggest distinct mechanisms by which 8803 may potentiate checkpoint blockade by αCTLA-4 vs αPD-1 to modulate the quantity and quality of

CD8 T cells within PDAC tumors following therapy.

123

Figure 15: Analysis of orthotopic and subcutaneous mT4 tumors following combination 8803 + checkpoint blockade therapy. (A) Implantation and treatment schedule. Both orthotopic and subcutaneous tumors were harvested following local 8803 into the orthotopic lesion only. Data are from pancreatic lesions unless otherwise indicated, and represent (B) density of CD8 T cells, (C) CD8:Treg ratios, (D) CD103 expression on Tregs, (E) functional markers Ki67 and (F) granzyme B on CD8 T cells, (G) granzyme B expression on CD8 T cells within the uninjected subcutaneous tumor, (H) Ly7C on CD8 T cells, (I) CD206 expresssion on F4/80+ macrophages, and (J) Ki67 expression on CD11c+ dendritic cells.

124

4.2.7: 8803 potentiates checkpoint blockade against orthotopic mT4-LA and does not require

chemotherapy.

Given the therapeutic additivity observed between intratumoral 8803 and systemic

checkpoint blockade in the responsive mT4-LS model, we sought to test whether synergy

would be observed in the mT4-LA model which more accurately reflects the aggressive and

refractory nature of late-stage PDAC. We implanted mice simultaneously with orthotopic and

subcutaneous mT4-LA to model primary and metastatic lesions, and performed intratumoral

injection of 8803 or PBS vehicle within the orthotopic tumor by survival surgery on days 10

and 22 post-implantation. Mice received αCTLA-4 and/or αPD-1 IP on days 10, 14, 18, 22,

and 26 (Fig 16A). In this model, local 8803 induces transient regressions that moderately

extend survival compared to PBS-treated mice, however no mice are cured. Monotherapy

checkpoint blockade with αCTLA-4 or αPD-1 do not deliver significant survival benefit over

PBS-treated mice, and the combination of αCTLA-4 or αPD-1 with 8803 does not significantly extend survival compared to monotherapies (Fig 16B, Supplemental Fig S10A). Therefore, in

contrast to the mT4-LS model, 8803 does not synergize with individual checkpoint blockade

against aggressive mT4-LA tumors. However, dual blockade of CTLA-4 and PD-1 extends

survival compared to PBS-treated mice, and the combination of local 8803 with αCTLA-

4/αPD-1 shows significant survival benefit over 8803 or αCTLA-4/αPD-1 alone (Fig 16B,

Supplemental Fig S10A). Additionally, this combination alone leads to a trend towards

reduced growth kinetics at the uninjected subcutaneous tumor (Supplemental Fig S10B).

These data demonstrate that 8803 can be delivered into late-stage, aggressive tumors and

synergize with dual αCTLA-4/αPD-1 checkpoint blockade to mediate local and systemic

control of mT4-LA PDAC tumors.

Longitudinal IVIS imagining of orthotopically implanted tumors indicates the

125

combination of local 8803 and dual checkpoint blockade can achieve significant, but

incomplete regression of mT4-LA tumors. We hypothesized that concurrent administration of

first-line chemotherapy agents gemcitabine (Gem) and Abraxane (nab-paclitaxel; nP) could

further enhance tumor control by (i) curtailing tumor growth kinetics and metastasis at initiation

of immunotherapy, (ii) mediating immunogenic tumor cell death, increasing tumor antigen

availability to APCs being activated by local CDN, and (iii) promoting depletion or further M1-

like polarization of tumor myeloid cells, which has been reported in response to nab-Paclitaxel

(476). While gemcitabine has controversial effects on macrophage polarization (477, 478), it has been successfully combined in preclinical models with myeloid-targeting agents including

CD40 agonists, CCR2 and CXCR2 inhibitors, and CSF-1R inhibitors (305, 330, 341).

Therefore, we implanted mice orthotopically with mT4-LA and evaluated whether Gem/nP administered IP (120mg/kg) on days 10 and 22 could enhance the efficacy of intratumoral

8803 and αCTLA-4/αPD-1 dual checkpoint blockade. In this experiment, mice were not inoculated with subcutaneous “pseudo-metastatic” lesions due to the highly metastatic nature of orthotopic mT4-LA and the tendency of subcutaneous tumors to ulcerate before reaching

1000mm3 in volume, requiring early euthanasia of long-term responders (Supplemental Fig

S10C). Chemotherapy with Gem/nP alone delivers no observable benefit by IVIS imaging or

survival, and does not significantly boost the efficacy of 8803 or αCTLA-4/αPD-1

(Supplemental Fig 10D). When combined with both 8803 and αCTLA-4/αPD-1, Gem/nP promotes early survival, with no mice succumbing to disease until day 41 post-implantation.

However, this early benefit does not translate to long-term survival advantage, as the median overall survival rate of this cohort is identical to mice treated with 8803 and αCTLA-4/αPD-1 without chemotherapy (Fig 16C). Additionally, we observe signs of increased treatment- related toxicity in mice exposed to Gem/nP, with 31.5% of mice failing to recover from the final therapeutic surgery (Supplemental Fig S10E). A smaller proportion of mice receiving Gem/nP 126

with 8803 or αCTLA-4/αPD-1 exhibit similar morbidity, while all mice receiving combination

8803 and αCTLA-4/αPD-1 recover without incident. Therefore, we conclude that standard of care chemotherapy is not required to achieve maximum therapeutic benefit when combining local 8803 with systemic dual checkpoint blockade with αCTLA-4 and αPD-1 against orthotopic mT4-LA PDAC tumors.

127

128

Figure 16: Intra-pancreatic 8803 synergizes with dual checkpoint blockade to prolong survival in mT4-LA independent of chemotherapy. (A) Implantation and treatment schedule. (B) Survival of mice bearing orthotopic and subcutaneous mT4-LA treated as described in (A), with n=5-10 mice per group. (C) Mice bearing only orthotopic mT4-LA (0.35x105 cells) were treated as described in (A) with additional standard of care chemotherapy consisting of gemcitabine (120mg/kg) and nab-paclitaxel (120 mg/kg) (Gem/nP) administered on days 10 and 22 in indicated groups. Longitudinal tumor growth by IVIS imaging and overall survival are shown. Data are cumulative of two independent experiments with 5-10 mice per group. Statistical significance was calculated using the Log-rank Mantel-Cox test. ns = not significant, * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

129

130

Supplemental Figure S10: Extended data on combination 8803 and checkpoint blockade in mT4-LA PDAC. (A) Longitudinal IVIS imaging for mice treated as described in Fig 16A. (B) Growth kinetics of subcutaneous mT4-LA tumors in mice shown in (A) and treated as described in Fig 16A. (C) Enumeration of mice deemed moribund due to ulceration >5mm in diameter at the uninjected subcutaneous tumor in the experiment shown in (A). (D) Survival of mice shown in Fig 16C in control groups consisting of Gem/nP and 8803 or Gem/nP and CTLA-4/PD-1 compared to relevant controls. (E) Percent of mice that failed to recover from final survival surgery on day 22 in experiments shown in Fig 16C. Statistical significance was calculated using the Log-rank Mantel-Cox test. ns = not significant, * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001.

131

4.2.8: Summary

Targeting the suppressive myeloid stroma of pancreatic ductal adenocarcinoma

(PDAC) is an attractive approach to sensitize these tumors to checkpoint blockade immunotherapy. Given the capacity for STING to control innate immune activation in both pathogenic and malignant settings, we investigated the effects of cyclic dinucleotide (CDN) agonists of STING on suppressive myeloid populations in vitro and in in vivo models of PDAC.

We characterized a novel CDN IACS-8803, which exhibits equivalent – if not superior – activity in all settings relative to the first-in-class clinical CDN ML-RR-S2-CDA (ML-RR). As a single agent, local delivery of 8803 mediates efficient regression of subcutaneous PDAC tumors compared to ML-RR or natural CDNs cGAMP or CDG, inducing a profound inflammatory remodeling of the PDAC tumor microenvironment. In isolated in vitro settings, 8803 reduces the T cell suppressive capacity of murine bone marrow-derived MDSCs, and mediates morphological and phenotypic M1-like repolarization of human M2c-like macrophages. Using a novel transplantable model of KPC-derived PDAC, we find that intratumoral administration of 8803 into visceral orthotopically-implanted tumors potentiates the effects of checkpoint blockade immunotherapy, leading to the rapid eradication of an early-stage responsive disease model and extending survival in the context of a late-stage aggressive disease model.

We find that standard of care chemotherapy is not required to achieve maximal survival benefit with this approach. Therefore, intratumoral delivery of potent STING agonists is an effective approach to sensitize aggressive PDAC tumors to checkpoint blockade in murine preclinical models, suggesting this combination may be of benefit in a clinical setting in urgent need of effective treatment options.

132

Chapter 5: Discussion

Castration resistant prostate cancer (CRPC) and pancreatic ductal adenocarcinoma

(PDAC) are terminal diseases that constitute a large proportion of all cancer-related deaths

each year (68). Though immense progress has been made over the last decade identifying immune-based treatments that can be curative for patients with otherwise lethal malignancies,

CRPC and PDAC remain refractory to these modalities. Therefore, new approaches must be

developed based upon a deep understanding of the underlying barriers to immunological

detection and rejection in these settings. A growing consensus in the world of immuno- oncology attributes poor tumor immunogenicity in large part to the wayward activities of innate immune cells within the tumor microenvironment, prompting the discovery of therapeutic regimens capable of depleting or reprogramming these cells to support – rather than suppress

– T cell-mediated antitumor immunity. In this dissertation, I studied the general hypothesis that activation of the STING pathway in suppressive tumor myeloid cells could mediate their functional repolarization, creating a context in which checkpoint blockade immunotherapy could liberate significant antitumor T cell responses in the context of poorly immunogenic cancers. In the course of these studies, I learned how intratumoral delivery of cyclic dinucleotide (CDN) STING agonists in combination with checkpoint modulators affects their ability to generate systemic immunity against multi-focal tumor burden, and how modulating

CDN dosing frequency can alter the balance between efficacy and inflammatory pathology. I had the opportunity to characterize a novel STING agonist, IACS-8803, which consistently outperformed the first-in-class clinical CDN ML-RR-S2-CDA in both in vitro and in vivo testing.

I further established a novel transplantable model of orthotopic PDAC in mice, and used this model to evaluate whether surgical delivery of 8803 into malignant pancreatic tissue could sensitize PDAC to checkpoint blockade or synergize with standard of care chemotherapy. In this chapter, I will discuss how my findings support, contradict, or advance the knowledge of

133

the field, and speculate about the potential future implications and applications of this work.

The strong preclinical foundation for targeting STING in cancer has spurred interest in translating CDNs as clinical compounds. To this end, a number of synthetic CDNs such as

ML-RR-S2-CDA have been created for potent activation of both murine and human STING and reduced sensitivity to enzymatic degradation (429, 431). This effort was made with the assumption that greater CDN potency in vitro directly correlates with optimal therapeutic benefit in vivo, a theory that has been partially substantiated by dose titrations experiments

(429) but has not been rigorously tested in the literature through comparison of different CDNs at equivalent concentrations. Some groups have argued that excessive STING activation can be a negative influence in cancer by upregulating counter-regulatory suppressive molecules such as IDO and PD-L1 (479) or through inhibition of T cell proliferation and survival (404,

405). Moreover, I found that cyclic di-GMP (CDG), one of the weakest agonists of STING utilized in the literature, possesses significant efficacy in the TRAMP-C2 model of prostate cancer, and that reduced dosing of CDG suppresses local pathologies without significantly blunting efficacy. These results would argue that the “more is better” approach may not universally apply to STING agonists. Our identification of a novel CDN with even greater potency than ML-RR-S2-CDA (ML-RR) gave me the opportunity to ask with a panel of unique agonists whether greater in vitro CDN potency does in fact correlate with ultimate therapeutic benefit in PDAC tumors. In the context of a non-saturating dose of CDN, I found that CDN potency does correlate, albeit imperfectly, with tumor regression following intratumoral injection. Synthetic CDNs ML-RR and 8803 are clearly superior to natural CDNs CDG and cGAMP in the subcutaneous mT4 model. However, cGAMP does not outperform CDG, and in fact may trend towards inferiority at the dosage tested. An intriguing observation with these data is that cGAMP, unlike CDG or synthetic CDNs, fails to induce a substantial granulocytic infiltrate upon local injection. Though the nature of this granulomatous response has not been

134

fully described in the literature, it has been suggested that these cells may contribute to

hemorrhagic necrosis at the injection site, possibly through TNFα release (442, 443). Whether recruitment of inflammatory neutrophils is a critical component of the CDN-mediated antitumor

response remains an open question for further investigation. Nevertheless, in in vitro myeloid

polarization experiments, I found that phenotypic and functional myeloid reprogramming

correlates directly with CDN potency. Thus, with the corollary that CDG and cGAMP don’t fit perfectly within the paradigm, my results argue that translating CDNs with high in vitro potency is more likely to lead to more effective tumor control in vivo.

Knowing this, an interesting area of investigation moving forward would be to determine exactly why synthetic CDNs possess such greater potency than cGAMP. It is not currently clear that these molecules bind with higher affinity to STING. The modifications that differentiate natural and synthetic CDNs tend to be associated with phosphodiesterase resistance, indicating that enhanced stability may underlie their greater efficacy (431).

However, CDNs are very small molecules that can be rapidly cleared from the circulation, so

it is not established that enzymatic degradation is rate-limiting in this context. An alternative

explanation could be that synthetic CDNs exhibit differential rates of cellular uptake relative

to natural CDNs, or are preferentially internalized by different cell types. Little is known about

how unformulated CDNs ultimately gain access to the cytoplasm, thus greater understanding

of the mechanisms controlling CDN uptake would be fruitful for the field. In addition, the

enhanced potency of synthetic CDNs may also be explained by a capacity to engage

qualitatively disparate signaling pathways relative to natural CDNs. One may imagine that

CDNs bearing unnatural thiol substitutions or fluorinated sugar residues induce unique

conformations in STING that affect recruitment of downstream signaling adaptors. Or, the

rapid activation of STING by synthetic CDNs may elaborate primary cytokine release with

faster kinetics compared to natural CDNs, leading to autocrine or paracrine signaling and

135

activation of secondary pathways that diversify the cytokine output. Experiments in this vein

are an area of current investigation in follow up of my dissertation studies.

In comparing the efficacy of known CDNs in vivo and in in vitro functional assays, my

results have both corroborated and deepened our understanding of the effects of CDNs on

suppressive myeloid populations. Polarization of macrophages along the M1/M2 spectrum

has been studied in response to DMXAA and RR-CDG, focusing on the level of RNA

transcription of selected genes such as arginase and iNOS (471) or release of cytokines

including IL-10 or TNFα (442). My studies are the first to monitor in detail the broad phenotypic

changes in human M2c polarized macrophages at the protein level by a panel of common

CDNs. These experiments illuminated the tendency of CDNs to potently induce CD86 on

suppressive macrophages, while CD80 can only be upregulated to a physiologically relevant

level by 8803 at the dose examined. Moreover, these studies were the first to describe

downregulation of M2-like markers CD163 and CD206 by human macrophages, coincident

with a clear morphological shift away from M2-like morphology (343, 472). Importantly, my

studies were also the first to describe detailed effects on MDSC function in response to CDN

exposure. By establishing an assay system to quantitatively evaluate T cell suppression by

homogenous populations of in vitro-differentiated bone marrow-derived MDSCs following

CDN repolarization, I was able to demonstrate that STING activation reduces the suppressive

capacity of MDSCs, with the most pronounced effects achieved by the more potent CDNs.

Further studies into the nature of this reversal of function are warranted, as it is unclear

whether observed functional changes are due to a shift in the relative contributions of PMN-

MDSCs versus MO-MDSCs in the culture, preferential cytotoxic effects on the most

suppressive cells in the culture, or outright downregulation of factors responsible for

ROS/RNS production, metabolic deprivation, or cytokine-mediated suppression of T cells.

Detailed analysis of signaling programs engaged by CDNs in human M2c macrophages and

136

murine MDSCs through RNA sequencing and RPPA protein analysis are currently underway,

which should shed light on these questions. At present, however, it appears that CDNs can

achieve phenotypic and functional reprogramming of suppressive myeloid populations, which

is likely to bolster the therapeutic effects of these molecules when delivered locally into

myeloid-rich suppressive tumor microenvironments.

The requirement for local delivery of CDNs presents a number of interesting questions

and obstacles for successful translation into humans. Chief among these are the issues of

dosing and applicability to visceral tumor settings. There is currently no consensus

concentration at which CDNs should be used in preclinical settings, with published reports

ranging from 5 µg to 100 µg per injection (429, 440). It can be argued that traditional “mg/kg” dosing is irrelevant in the context of intratumoral therapy, where therapies are intended to act at the site of injection rather than throughout the body. Rather, dosing based upon the size of the injected tumor may be more relevant, in light of the relationship I found between tumor size and the extent of inflammation at the site on injection with a flat dose of CDG. There is currently one FDA approved therapy administered by intratumoral injection in cancer; the oncolytic virus talimogene laherparepvec (T-VEC) (480). This drug is administered with a hybrid flat dosing scheme, in which patients receive either 106 pfu/mL or 108 pfu/mL T-VEC, but receive different injection volumes according to tumor size; starting at 0.1 mL for lesions under 0.5cm in diameter, ranging up to 4.0 mL for lesions greater than 5.0 cm in diameter

(481). Experience with T-VEC may be useful in guiding clinical dosing strategies for CDNs.

Additionally, few studies have evaluated different dosing frequencies, in terms of the number

of CDN injections or days between injections, if multiple are used. My studies were the first to

directly report on the effects of reducing CDN dosing frequency, as I found that a single CDN

injection invokes less toxicity without significantly sacrificing systemic efficacy compared to

three injections when combined with local checkpoint modulators in the bilateral TRAMP-C2

137

model. Francica et al found that within 24 hours following CDN injection, tumors exhibit widespread cleaved caspase 3 expression, indicative of a rapid apoptotic response in the

TME (443). As STING activation is known to exert toxicity in the T cell compartment (404,

405), one can imagine that use of frequent successive CDN injections may prove counter- productive if this results in apoptotic death of newly primed and recruited T cells. In this case, frequent high-dose CDNs may deliver favorable cytoreduction of large bulky tumors, but this may occur at the expense of a maintained memory T cell response as a result of T cell toxicity.

Thus, the relationship between CDN dosing, tumor reduction, and generation of memory T cell response is an important area of investigation for the field moving forward.

Irrespective of this biology, the use of multiple CDN injections may be practically untenable in the clinic for the majority of patients that would require delivery of these agonists to sites of visceral disease. In theory, any organ in the body can be accessed through endoscopic, laparoscopic, or interventional radiologic means, and these are regularly utilized to biopsy human tumors (482). However, these procedures are associated with significant cost and risk, especially if multiple injections are preferred. My studies indicate that a single or two CDN injections are sufficient to elicit significant therapeutic benefit in both subcutaneous TRAMP-C2 or orthotopically-implanted mT4 PDAC, suggesting that local administration at the time of biopsy may be feasible and effective for patients even with late stage disease. Though rare in PDAC, the adjuvant setting presents an intriguing place to test

the effects of local CDNs in patients already undergoing surgery for resectable or borderline

resectable disease, especially given the high rate of relapse in these patients (68). Two

innovative approaches have been proposed to prolong or optimize the benefit associated with

local innate agonists in this setting through slow-release formulations. Smith and colleagues

developed a biopolymer film that could be loaded with CDNs and CAR T cells and placed

within the peritoneal cavity in proximity to growing PDAC tumors, allowing for extended

138

release of both constituents after a single procedure (483). In this study, the biopolymer approach outperformed local injection of naked CDN, and was sufficient to cure a portion of mice with aggressive orthotopically implanted KPC-derived tumors. While striking, the efficacy of this therapy was heavily dependent on CAR T cells, which have not been successfully utilized in the context of solid cancers and thus have many barriers preventing their use in human PDAC, for example. An alternative approach involves formulating CDNs or other innate agonists into a biodegradable hydrogel matrix, which can be locally implanted and allow for slow release of CDNs into tumors over one to two weeks. To date, two studies have incorporated CDNs into hydrogels for use in cancer therapy, and have demonstrated immune potentiating effects and therapeutic benefit in 4T1 breast cancer and MOC2 head and neck cancer models (484, 485). This approach, combined with checkpoint blockade therapy, may enhance the efficacy of my local injection method without requiring use of CAR T cells as used in the study by Smith et al. This is thus an area of active investigation with exciting translational potential.

Ultimately, the holy grail of any intratumoral immunotherapy is generation of abscopal immunity. A number of studies have reported that local delivery of STING agonists in the single agent setting is sufficient to induce systemic immunity against distal uninjected tumors

(429, 439). In contrast, a number of other reports have found the opposite to be true; that

STING agonists alone are insufficient to generate abscopal immunity on their own, but only affect outgrowth of distal tumors when used in combination with other systemic immunotherapies (438, 440, 442). My results have indicated that either can be true. In all likelihood, the ability of CDNs alone to generate abscopal immunity may be dependent on tumor antigenicity as well as the size and composition of the tumor at the time of therapy. For example, using the less aggressive mT4-LS model of PDAC, I found that single agent 8803 delivered into orthotopic lesions can elicit curative responses against the injected tumor as

139

well as the uninjected subcutaneous tumor in 50% of mice tested. However, in the aggressive

mT4-LA model, no curative responses are observed using the equivalent treatment schedule.

Thus, in two models which are theoretically equivalent in terms of antigen burden but differ in

in vivo growth kinetics, abscopal immunity was only generated in the context of lesser tumor

burden. Interestingly, this too may be context-dependent, as no significant abscopal immunity

could be generated using CDG against small ~25mm3 TRAMP-C2 tumors, which are known to exhibit relatively slow growth kinetics in vivo. In this context, CDN potency and the immune composition of distal TRAMP-C2 tumors may be critical in dictating responsivity of the uninjected tumor to abscopal T cell-mediated immunity. A final factor to be considered in this discussion is that the small size and highly-diffusive nature of CDNs makes systemic distribution likely even following local intratumoral injection. Without rigorous evaluation of the levels of CDN reaching the uninjected tumor that may have direct immune-potentiating effects,

it is difficult to make firm conclusions about abscopal immunity, especially when large doses

(~50-100 µg) are administered one or more times (429, 443). Therefore, while CDNs are

theoretically capable of promoting priming of systemic T cell responses, it is difficult to predict

under which circumstances these responses will be seen, and detailed validation needs to be

provided to determine whether any such response is the result of bona fide abscopal

immunity.

Together, my results help lay the groundwork to test CDNs in combination with

checkpoint blockade therapy in patients with CRPC or PDAC, two settings with urgent need

for effective treatment modalities. These studies have contributed to the field at large by

providing insight into CDN dosing strategies and local pathologies when combined with

checkpoint modulating therapies, by identifying a novel potent CDN and demonstrating that

CDN potency does correlate with efficacy in both in vitro functional assays and in vivo tumor

models, by enriching our understanding of the effects of CDNs on M2c polarized macrophages

140

and bone marrow-derived MDSCs of human and murine origin, and by providing proof of

concept that CDNs can be delivered into visceral PDAC tumors and synergize with checkpoint

blockade immunotherapy in the setting of highly aggressive and metastatic disease. The

coming years should see rapid clinical translation of therapies targeting myeloid cells in many

cancers including CRPC and PDAC, where the myeloid infiltrate is numerous and functionally

corrupt. My hope is that, in some part due to the work presented in this thesis, these powerful

STING-targeting therapies may help lead the way in this endeavor, and that patients in need of hope may find it as a result.

141

Bibliography

1. Shankaran, V., H. Ikeda, A. T. Bruce, J. M. White, P. E. Swanson, L. J. Old, and R.

D. Schreiber. 2001. IFNgamma and lymphocytes prevent primary tumour development and

shape tumour immunogenicity. Nature 410: 1107-1111.

2. Krummel, M. F., and J. P. Allison. 1995. CD28 and CTLA-4 have opposing effects on

the response of T cells to stimulation. J Exp Med 182: 459-465.

3. Leach, D. R., M. F. Krummel, and J. P. Allison. 1996. Enhancement of antitumor

immunity by CTLA-4 blockade. Science 271: 1734-1736.

4. Chapman, P. B., A. Hauschild, C. Robert, J. B. Haanen, P. Ascierto, J. Larkin, R.

Dummer, C. Garbe, A. Testori, M. Maio, D. Hogg, P. Lorigan, C. Lebbe, T. Jouary, D.

Schadendorf, A. Ribas, S. J. O'Day, J. A. Sosman, J. M. Kirkwood, A. M. M. Eggermont, B.

Dreno, K. Nolop, J. Li, B. Nelson, J. Hou, R. J. Lee, K. T. Flaherty, and G. A. McArthur.

2011. Improved Survival with Vemurafenib in Melanoma with BRAF V600E Mutation. New

England Journal of Medicine 364: 2507-2516.

5. Burnet, M. 1957. Cancer; a biological approach. I. The processes of control. Br Med

J 1: 779-786.

6. Gubin, M. M., X. Zhang, H. Schuster, E. Caron, J. P. Ward, T. Noguchi, Y. Ivanova,

J. Hundal, C. D. Arthur, W. J. Krebber, G. E. Mulder, M. Toebes, M. D. Vesely, S. S. Lam, A.

J. Korman, J. P. Allison, G. J. Freeman, A. H. Sharpe, E. L. Pearce, T. N. Schumacher, R.

Aebersold, H. G. Rammensee, C. J. Melief, E. R. Mardis, W. E. Gillanders, M. N. Artyomov,

and R. D. Schreiber. 2014. Checkpoint blockade cancer immunotherapy targets tumour-

specific mutant antigens. Nature 515: 577-581.

7. Snyder, A., V. Makarov, T. Merghoub, J. Yuan, J. M. Zaretsky, A. Desrichard, L. A.

Walsh, M. A. Postow, P. Wong, T. S. Ho, T. J. Hollmann, C. Bruggeman, K. Kannan, Y. Li,

C. Elipenahli, C. Liu, C. T. Harbison, L. Wang, A. Ribas, J. D. Wolchok, and T. A. Chan.

142

2014. Genetic basis for clinical response to CTLA-4 blockade in melanoma. The New

England journal of medicine 371: 2189-2199.

8. Martincorena, I., and P. J. Campbell. 2015. Somatic mutation in cancer and normal

cells. Science 349: 1483-1489.

9. Borghaei, H., L. Paz-Ares, L. Horn, D. R. Spigel, M. Steins, N. E. Ready, L. Q. Chow,

E. E. Vokes, E. Felip, E. Holgado, F. Barlesi, M. Kohlhäufl, O. Arrieta, M. A. Burgio, J.

Fayette, H. Lena, E. Poddubskaya, D. E. Gerber, S. N. Gettinger, C. M. Rudin, N. Rizvi, L.

Crinò, G. R. Blumenschein, S. J. Antonia, C. Dorange, C. T. Harbison, F. Graf Finckenstein,

and J. R. Brahmer. 2015. Nivolumab versus Docetaxel in Advanced Nonsquamous Non–

Small-Cell Lung Cancer. New England Journal of Medicine 373: 1627-1639.

10. Hodi, F. S., S. J. O'Day, D. F. McDermott, R. W. Weber, J. A. Sosman, J. B. Haanen,

R. Gonzalez, C. Robert, D. Schadendorf, J. C. Hassel, W. Akerley, A. J. van den Eertwegh,

J. Lutzky, P. Lorigan, J. M. Vaubel, G. P. Linette, D. Hogg, C. H. Ottensmeier, C. Lebbe, C.

Peschel, I. Quirt, J. I. Clark, J. D. Wolchok, J. S. Weber, J. Tian, M. J. Yellin, G. M. Nichol,

A. Hoos, and W. J. Urba. 2010. Improved survival with ipilimumab in patients with metastatic melanoma. The New England journal of medicine 363: 711-723.

11. Yarchoan, M., A. Hopkins, and E. M. Jaffee. 2017. Tumor Mutational Burden and

Response Rate to PD-1 Inhibition. New England Journal of Medicine 377: 2500-2501.

12. Kunkel, T. A., and D. A. Erie. 2015. Eukaryotic Mismatch Repair in Relation to DNA

Replication. Annual review of genetics 49: 291-313.

13. Boland, C. R., and A. Goel. 2010. Microsatellite instability in colorectal cancer.

Gastroenterology 138: 2073-2087.e2073.

14. Le, D. T., J. N. Uram, H. Wang, B. R. Bartlett, H. Kemberling, A. D. Eyring, A. D.

Skora, B. S. Luber, N. S. Azad, D. Laheru, B. Biedrzycki, R. C. Donehower, A. Zaheer, G. A.

Fisher, T. S. Crocenzi, J. J. Lee, S. M. Duffy, R. M. Goldberg, A. de la Chapelle, M. Koshiji,

143

F. Bhaijee, T. Huebner, R. H. Hruban, L. D. Wood, N. Cuka, D. M. Pardoll, N.

Papadopoulos, K. W. Kinzler, S. Zhou, T. C. Cornish, J. M. Taube, R. A. Anders, J. R.

Eshleman, B. Vogelstein, and L. A. Diaz. 2015. PD-1 Blockade in Tumors with Mismatch-

Repair Deficiency. New England Journal of Medicine 372: 2509-2520.

15. Le, D. T., J. N. Durham, K. N. Smith, H. Wang, B. R. Bartlett, L. K. Aulakh, S. Lu, H.

Kemberling, C. Wilt, B. S. Luber, F. Wong, N. S. Azad, A. A. Rucki, D. Laheru, R.

Donehower, A. Zaheer, G. A. Fisher, T. S. Crocenzi, J. J. Lee, T. F. Greten, A. G. Duffy, K.

K. Ciombor, A. D. Eyring, B. H. Lam, A. Joe, S. P. Kang, M. Holdhoff, L. Danilova, L. Cope,

C. Meyer, S. Zhou, R. M. Goldberg, D. K. Armstrong, K. M. Bever, A. N. Fader, J. Taube, F.

Housseau, D. Spetzler, N. Xiao, D. M. Pardoll, N. Papadopoulos, K. W. Kinzler, J. R.

Eshleman, B. Vogelstein, R. A. Anders, and L. A. Diaz, Jr. 2017. Mismatch repair deficiency

predicts response of solid tumors to PD-1 blockade. Science 357: 409-413.

16. Van Allen, E. M., D. Miao, B. Schilling, S. A. Shukla, C. Blank, L. Zimmer, A. Sucker,

U. Hillen, M. H. G. Foppen, S. M. Goldinger, J. Utikal, J. C. Hassel, B. Weide, K. C. Kaehler,

C. Loquai, P. Mohr, R. Gutzmer, R. Dummer, S. Gabriel, C. J. Wu, D. Schadendorf, and L.

A. Garraway. 2015. Genomic correlates of response to CTLA-4 blockade in metastatic

melanoma. Science 350: 207-211.

17. Turajlic, S., K. Litchfield, H. Xu, R. Rosenthal, N. McGranahan, J. L. Reading, Y. N.

S. Wong, A. Rowan, N. Kanu, M. Al Bakir, T. Chambers, R. Salgado, P. Savas, S. Loi, N. J.

Birkbak, L. Sansregret, M. Gore, J. Larkin, S. A. Quezada, and C. Swanton. 2017. Insertion- and-deletion-derived tumour-specific neoantigens and the immunogenic phenotype: a pan- cancer analysis. Lancet Oncol 18: 1009-1021.

18. Balachandran, V. P., M. Luksza, J. N. Zhao, V. Makarov, J. A. Moral, R. Remark, B.

Herbst, G. Askan, U. Bhanot, Y. Senbabaoglu, D. K. Wells, C. I. O. Cary, O. Grbovic-Huezo,

M. Attiyeh, B. Medina, J. Zhang, J. Loo, J. Saglimbeni, M. Abu-Akeel, R. Zappasodi, N.

144

Riaz, M. Smoragiewicz, Z. L. Kelley, O. Basturk, I. Australian Pancreatic Cancer Genome,

R. Garvan Institute of Medical, H. Prince of Wales, H. Royal North Shore, G. University of,

H. St Vincent's, Q. B. M. R. Institute, C. f. C. R. University of Melbourne, I. f. M. B. University

of Queensland, H. Bankstown, H. Liverpool, C. O. B. L. Royal Prince Alfred Hospital, H.

Westmead, H. Fremantle, H. St John of God, H. Royal Adelaide, C. Flinders Medical, P.

Envoi, H. Princess Alexandria, H. Austin, I. Johns Hopkins Medical, A. R.-N. C. f. A. R. o.

Cancer, M. Gonen, A. J. Levine, P. J. Allen, D. T. Fearon, M. Merad, S. Gnjatic, C. A.

Iacobuzio-Donahue, J. D. Wolchok, R. P. DeMatteo, T. A. Chan, B. D. Greenbaum, T.

Merghoub, and S. D. Leach. 2017. Identification of unique neoantigen qualities in long-term survivors of pancreatic cancer. Nature 551: 512-516.

19. Dunn, G. P., A. T. Bruce, H. Ikeda, L. J. Old, and R. D. Schreiber. 2002. Cancer immunoediting: from immunosurveillance to tumor escape. Nature immunology 3: 991-998.

20. Wick, M., P. Dubey, H. Koeppen, C. T. Siegel, P. E. Fields, L. Chen, J. A. Bluestone, and H. Schreiber. 1997. Antigenic Cancer Cells Grow Progressively in Immune Hosts without Evidence for T Cell Exhaustion or Systemic Anergy. The Journal of Experimental

Medicine 186: 229-238.

21. Galon, J., A. Costes, F. Sanchez-Cabo, A. Kirilovsky, B. Mlecnik, C. Lagorce-Pages,

M. Tosolini, M. Camus, A. Berger, P. Wind, F. Zinzindohoue, P. Bruneval, P. H. Cugnenc, Z.

Trajanoski, W. H. Fridman, and F. Pages. 2006. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science 313: 1960-1964.

22. Pagès, F., A. Berger, M. Camus, F. Sanchez-Cabo, A. Costes, R. Molidor, B.

Mlecnik, A. Kirilovsky, M. Nilsson, D. Damotte, T. Meatchi, P. Bruneval, P.-H. Cugnenc, Z.

Trajanoski, W.-H. Fridman, and J. Galon. 2005. Effector Memory T Cells, Early Metastasis, and Survival in Colorectal Cancer. New England Journal of Medicine 353: 2654-2666.

23. Mlecnik, B., G. Bindea, H. K. Angell, P. Maby, M. Angelova, D. Tougeron, S. E.

145

Church, L. Lafontaine, M. Fischer, T. Fredriksen, M. Sasso, A. M. Bilocq, A. Kirilovsky, A. C.

Obenauf, M. Hamieh, A. Berger, P. Bruneval, J. J. Tuech, J. C. Sabourin, F. Le Pessot, J.

Mauillon, A. Rafii, P. Laurent-Puig, M. R. Speicher, Z. Trajanoski, P. Michel, R. Sesboue, T.

Frebourg, F. Pages, V. Valge-Archer, J. B. Latouche, and J. Galon. 2016. Integrative

Analyses of Colorectal Cancer Show Immunoscore Is a Stronger Predictor of Patient

Survival Than Microsatellite Instability. Immunity 44: 698-711.

24. Mlecnik, B., M. Tosolini, A. Kirilovsky, A. Berger, G. Bindea, T. Meatchi, P. Bruneval,

Z. Trajanoski, W. H. Fridman, F. Pages, and J. Galon. 2011. Histopathologic-based prognostic factors of colorectal cancers are associated with the state of the local immune reaction. J Clin Oncol 29: 610-618.

25. Azimi, F., R. A. Scolyer, P. Rumcheva, M. Moncrieff, R. Murali, S. W. McCarthy, R.

P. Saw, and J. F. Thompson. 2012. Tumor-infiltrating lymphocyte grade is an independent predictor of sentinel lymph node status and survival in patients with cutaneous melanoma. J

Clin Oncol 30: 2678-2683.

26. Thomas, N. E., K. J. Busam, L. From, A. Kricker, B. K. Armstrong, H. Anton-Culver,

S. B. Gruber, R. P. Gallagher, R. Zanetti, S. Rosso, T. Dwyer, A. Venn, P. A. Kanetsky, P.

A. Groben, H. Hao, I. Orlow, A. S. Reiner, L. Luo, S. Paine, D. W. Ollila, H. Wilcox, C. B.

Begg, and M. Berwick. 2013. Tumor-infiltrating lymphocyte grade in primary melanomas is independently associated with melanoma-specific survival in the population-based genes, environment and melanoma study. J Clin Oncol 31: 4252-4259.

27. Zhang, L., J. R. Conejo-Garcia, D. Katsaros, P. A. Gimotty, M. Massobrio, G.

Regnani, A. Makrigiannakis, H. Gray, K. Schlienger, M. N. Liebman, S. C. Rubin, and G.

Coukos. 2003. Intratumoral T cells, recurrence, and survival in epithelial ovarian cancer. The

New England journal of medicine 348: 203-213.

28. Hwang, W. T., S. F. Adams, E. Tahirovic, I. S. Hagemann, and G. Coukos. 2012.

146

Prognostic significance of tumor-infiltrating T cells in ovarian cancer: a meta-analysis.

Gynecol Oncol 124: 192-198.

29. Balermpas, P., F. Rodel, R. Liberz, J. Oppermann, J. Wagenblast, S. Ghanaati, P. N.

Harter, M. Mittelbronn, C. Weiss, C. Rodel, and E. Fokas. 2014. Head and neck cancer relapse after chemoradiotherapy correlates with CD163+ macrophages in primary tumour and CD11b+ myeloid cells in recurrences. Br J Cancer 111: 1509-1518.

30. Gabrielson, A., Y. Wu, H. Wang, J. Jiang, B. Kallakury, Z. Gatalica, S. Reddy, D.

Kleiner, T. Fishbein, L. Johnson, E. Island, R. Satoskar, F. Banovac, R. Jha, J. Kachhela, P.

Feng, T. Zhang, A. Tesfaye, P. Prins, C. Loffredo, J. Marshall, L. Weiner, M. Atkins, and A.

R. He. 2016. Intratumoral CD3 and CD8 T-cell Densities Associated with Relapse-Free

Survival in HCC. Cancer Immunol Res 4: 419-430.

31. Sharma, P., Y. Shen, S. Wen, S. Yamada, A. A. Jungbluth, S. Gnjatic, D. F. Bajorin,

V. E. Reuter, H. Herr, L. J. Old, and E. Sato. 2007. CD8 tumor-infiltrating lymphocytes are predictive of survival in muscle-invasive urothelial carcinoma. Proc Natl Acad Sci U S A 104:

3967-3972.

32. Rusakiewicz, S., M. Semeraro, M. Sarabi, M. Desbois, C. Locher, R. Mendez, N.

Vimond, A. Concha, F. Garrido, N. Isambert, L. Chaigneau, V. Le Brun-Ly, P. Dubreuil, I.

Cremer, A. Caignard, V. Poirier-Colame, K. Chaba, C. Flament, N. Halama, D. Jager, A.

Eggermont, S. Bonvalot, F. Commo, P. Terrier, P. Opolon, J. F. Emile, J. M. Coindre, G.

Kroemer, N. Chaput, A. Le Cesne, J. Y. Blay, and L. Zitvogel. 2013. Immune infiltrates are prognostic factors in localized gastrointestinal stromal tumors. Cancer Res 73: 3499-3510.

33. Spranger, S., J. J. Luke, R. Bao, Y. Zha, K. M. Hernandez, Y. Li, A. P. Gajewski, J.

Andrade, and T. F. Gajewski. 2016. Density of immunogenic antigens does not explain the presence or absence of the T-cell–inflamed tumor microenvironment in melanoma.

Proceedings of the National Academy of Sciences 113: E7759-E7768.

147

34. Fuertes, M. B., A. K. Kacha, J. Kline, S. R. Woo, D. M. Kranz, K. M. Murphy, and T.

F. Gajewski. 2011. Host type I IFN signals are required for antitumor CD8+ T cell responses

through CD8{alpha}+ dendritic cells. J Exp Med 208: 2005-2016.

35. Le Bon, A., N. Etchart, C. Rossmann, M. Ashton, S. Hou, D. Gewert, P. Borrow, and

D. F. Tough. 2003. Cross-priming of CD8+ T cells stimulated by virus-induced type I interferon. Nature immunology 4: 1009-1015.

36. Jongbloed, S. L., A. J. Kassianos, K. J. McDonald, G. J. Clark, X. Ju, C. E. Angel, C.

J. Chen, P. R. Dunbar, R. B. Wadley, V. Jeet, A. J. Vulink, D. N. Hart, and K. J. Radford.

2010. Human CD141+ (BDCA-3)+ dendritic cells (DCs) represent a unique myeloid DC subset that cross-presents necrotic cell antigens. J Exp Med 207: 1247-1260.

37. Harlin, H., Y. Meng, A. C. Peterson, Y. Zha, M. Tretiakova, C. Slingluff, M. McKee, and T. F. Gajewski. 2009. Chemokine expression in melanoma metastases associated with

CD8+ T-cell recruitment. Cancer Res 69: 3077-3085.

38. Spranger, S., D. Dai, B. Horton, and T. F. Gajewski. 2017. Tumor-Residing Batf3

Dendritic Cells Are Required for Effector T Cell Trafficking and Adoptive T Cell Therapy.

Cancer Cell 31: 711-723 e714.

39. Dunn, G. P., A. T. Bruce, K. C. Sheehan, V. Shankaran, R. Uppaluri, J. D. Bui, M. S.

Diamond, C. M. Koebel, C. Arthur, J. M. White, and R. D. Schreiber. 2005. A critical function for type I interferons in cancer immunoediting. Nature immunology 6: 722-729.

40. Hildner, K., B. T. Edelson, W. E. Purtha, M. Diamond, H. Matsushita, M. Kohyama,

B. Calderon, B. U. Schraml, E. R. Unanue, M. S. Diamond, R. D. Schreiber, T. L. Murphy, and K. M. Murphy. 2008. Batf3 deficiency reveals a critical role for CD8alpha+ dendritic cells in cytotoxic T cell immunity. Science 322: 1097-1100.

41. Chen, L., and D. B. Flies. 2013. Molecular mechanisms of T cell co-stimulation and co-inhibition. Nature reviews. Immunology 13: 227-242.

148

42. Juneja, V. R., K. A. McGuire, R. T. Manguso, M. W. LaFleur, N. Collins, W. N.

Haining, G. J. Freeman, and A. H. Sharpe. 2017. PD-L1 on tumor cells is sufficient for

immune evasion in immunogenic tumors and inhibits CD8 T cell cytotoxicity. J Exp Med 214:

895-904.

43. Chemnitz, J. M., R. V. Parry, K. E. Nichols, C. H. June, and J. L. Riley. 2004. SHP-1 and SHP-2 associate with immunoreceptor tyrosine-based switch motif of programmed death 1 upon primary human T cell stimulation, but only receptor ligation prevents T cell activation. J Immunol 173: 945-954.

44. Sheppard, K. A., L. J. Fitz, J. M. Lee, C. Benander, J. A. George, J. Wooters, Y. Qiu,

J. M. Jussif, L. L. Carter, C. R. Wood, and D. Chaudhary. 2004. PD-1 inhibits T-cell receptor induced phosphorylation of the ZAP70/CD3zeta signalosome and downstream signaling to

PKCtheta. FEBS letters 574: 37-41.

45. Hui, E., J. Cheung, J. Zhu, X. Su, M. J. Taylor, H. A. Wallweber, D. K. Sasmal, J.

Huang, J. M. Kim, I. Mellman, and R. D. Vale. 2017. T cell costimulatory receptor CD28 is a primary target for PD-1-mediated inhibition. Science 355: 1428-1433.

46. Herbst, R. S., J. C. Soria, M. Kowanetz, G. D. Fine, O. Hamid, M. S. Gordon, J. A.

Sosman, D. F. McDermott, J. D. Powderly, S. N. Gettinger, H. E. Kohrt, L. Horn, D. P.

Lawrence, S. Rost, M. Leabman, Y. Xiao, A. Mokatrin, H. Koeppen, P. S. Hegde, I. Mellman,

D. S. Chen, and F. S. Hodi. 2014. Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients. Nature 515: 563-567.

47. Loke, P., and J. P. Allison. 2003. PD-L1 and PD-L2 are differentially regulated by

Th1 and Th2 cells. Proc Natl Acad Sci U S A 100: 5336-5341.

48. Spranger, S., R. M. Spaapen, Y. Zha, J. Williams, Y. Meng, T. T. Ha, and T. F.

Gajewski. 2013. Up-regulation of PD-L1, IDO, and T(regs) in the melanoma tumor microenvironment is driven by CD8(+) T cells. Sci Transl Med 5: 200ra116.

149

49. Clark, R. A., S. J. Huang, G. F. Murphy, I. G. Mollet, D. Hijnen, M. Muthukuru, C. F.

Schanbacher, V. Edwards, D. M. Miller, J. E. Kim, J. Lambert, and T. S. Kupper. 2008.

Human squamous cell carcinomas evade the immune response by down-regulation of vascular E-selectin and recruitment of regulatory T cells. The Journal of Experimental

Medicine 205: 2221-2234.

50. Mazanet, M. M., and C. C. Hughes. 2002. B7-H1 is expressed by human endothelial cells and suppresses T cell cytokine synthesis. J Immunol 169: 3581-3588.

51. Rodig, N., T. Ryan, J. A. Allen, H. Pang, N. Grabie, T. Chernova, E. A. Greenfield, S.

C. Liang, A. H. Sharpe, A. H. Lichtman, and G. J. Freeman. 2003. Endothelial expression of

PD-L1 and PD-L2 down-regulates CD8+ T cell activation and cytolysis. Eur J Immunol 33:

3117-3126.

52. Buckanovich, R. J., A. Facciabene, S. Kim, F. Benencia, D. Sasaroli, K. Balint, D.

Katsaros, A. O'Brien-Jenkins, P. A. Gimotty, and G. Coukos. 2008. Endothelin B receptor mediates the endothelial barrier to T cell homing to tumors and disables immune therapy.

Nat Med 14: 28-36.

53. Motz, G. T., S. P. Santoro, L. P. Wang, T. Garrabrant, R. R. Lastra, I. S. Hagemann,

P. Lal, M. D. Feldman, F. Benencia, and G. Coukos. 2014. Tumor endothelium FasL establishes a selective immune barrier promoting tolerance in tumors. Nat Med 20: 607-615.

54. Kraman, M., P. J. Bambrough, J. N. Arnold, E. W. Roberts, L. Magiera, J. O. Jones,

A. Gopinathan, D. A. Tuveson, and D. T. Fearon. 2010. Suppression of antitumor immunity by stromal cells expressing fibroblast activation protein-alpha. Science 330: 827-830.

55. Hartmann, N., N. A. Giese, T. Giese, I. Poschke, R. Offringa, J. Werner, and E.

Ryschich. 2014. Prevailing role of contact guidance in intrastromal T-cell trapping in human pancreatic cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 20: 3422-3433.

150

56. Salmon, H., K. Franciszkiewicz, D. Damotte, M.-C. Dieu-Nosjean, P. Validire, A.

Trautmann, F. Mami-Chouaib, and E. Donnadieu. 2012. Matrix architecture defines the preferential localization and migration of T cells into the stroma of human lung tumors. The

Journal of Clinical Investigation 122: 899-910.

57. Feig, C., J. O. Jones, M. Kraman, R. J. Wells, A. Deonarine, D. S. Chan, C. M.

Connell, E. W. Roberts, Q. Zhao, O. L. Caballero, S. A. Teichmann, T. Janowitz, D. I.

Jodrell, D. A. Tuveson, and D. T. Fearon. 2013. Targeting CXCL12 from FAP-expressing carcinoma-associated fibroblasts synergizes with anti-PD-L1 immunotherapy in pancreatic cancer. Proc Natl Acad Sci U S A 110: 20212-20217.

58. Ozdemir, B. C., T. Pentcheva-Hoang, J. L. Carstens, X. Zheng, C. C. Wu, T. R.

Simpson, H. Laklai, H. Sugimoto, C. Kahlert, S. V. Novitskiy, A. De Jesus-Acosta, P.

Sharma, P. Heidari, U. Mahmood, L. Chin, H. L. Moses, V. M. Weaver, A. Maitra, J. P.

Allison, V. S. LeBleu, and R. Kalluri. 2014. Depletion of carcinoma-associated fibroblasts and fibrosis induces immunosuppression and accelerates pancreas cancer with reduced survival. Cancer Cell 25: 719-734.

59. Ben-Shoshan, J., S. Maysel-Auslender, A. Mor, G. Keren, and J. George. 2008.

Hypoxia controls CD4+CD25+ regulatory T-cell homeostasis via hypoxia-inducible factor-

1alpha. Eur J Immunol 38: 2412-2418.

60. Corzo, C. A., T. Condamine, L. Lu, M. J. Cotter, J. I. Youn, P. Cheng, H. I. Cho, E.

Celis, D. G. Quiceno, T. Padhya, T. V. McCaffrey, J. C. McCaffrey, and D. I. Gabrilovich.

2010. HIF-1alpha regulates function and differentiation of myeloid-derived suppressor cells in the tumor microenvironment. J Exp Med 207: 2439-2453.

61. Dang, E. V., J. Barbi, H. Y. Yang, D. Jinasena, H. Yu, Y. Zheng, Z. Bordman, J. Fu,

Y. Kim, H. R. Yen, W. Luo, K. Zeller, L. Shimoda, S. L. Topalian, G. L. Semenza, C. V.

Dang, D. M. Pardoll, and F. Pan. 2011. Control of T(H)17/T(reg) balance by hypoxia-

151

inducible factor 1. Cell 146: 772-784.

62. Hatfield, S. M., J. Kjaergaard, D. Lukashev, T. H. Schreiber, B. Belikoff, R. Abbott, S.

Sethumadhavan, P. Philbrook, K. Ko, R. Cannici, M. Thayer, S. Rodig, J. L. Kutok, E. K.

Jackson, B. Karger, E. R. Podack, A. Ohta, and M. V. Sitkovsky. 2015. Immunological

mechanisms of the antitumor effects of supplemental oxygenation. Sci Transl Med 7:

277ra230.

63. Jayaprakash, P., M. Ai, A. Liu, P. Budhani, T. Bartkowiak, J. Sheng, C. Ager, C.

Nicholas, A. R. Jaiswal, Y. Sun, K. Shah, S. Balasubramanyam, N. Li, G. Wang, J. Ning, A.

Zal, T. Zal, and M. A. Curran. 2018. Targeted hypoxia reduction restores T cell infiltration and sensitizes prostate cancer to immunotherapy. J Clin Invest 128: 5137-5149.

64. Evans, R. A., M. S. Diamond, A. J. Rech, T. Chao, M. W. Richardson, J. H. Lin, D. L.

Bajor, K. T. Byrne, B. Z. Stanger, J. L. Riley, N. Markosyan, R. Winograd, and R. H.

Vonderheide. 2016. Lack of immunoediting in murine pancreatic cancer reversed with neoantigen. JCI Insight 1.

65. Spranger, S., R. Bao, and T. F. Gajewski. 2015. Melanoma-intrinsic β-catenin signalling prevents anti-tumour immunity. Nature 523: 231.

66. Sweis, R. F., M. W. Drazer, and M. J. Ratain. 2016. Analysis of Impact of Post-

Treatment Biopsies in Phase I Clinical Trials. J Clin Oncol 34: 369-374.

67. Cristescu, R., R. Mogg, M. Ayers, A. Albright, E. Murphy, J. Yearley, X. Sher, X. Q.

Liu, H. Lu, M. Nebozhyn, C. Zhang, J. K. Lunceford, A. Joe, J. Cheng, A. L. Webber, N.

Ibrahim, E. R. Plimack, P. A. Ott, T. Y. Seiwert, A. Ribas, T. K. McClanahan, J. E.

Tomassini, A. Loboda, and D. Kaufman. 2018. Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy. Science 362.

68. Siegel, R. L., K. D. Miller, and A. Jemal. 2018. Cancer statistics, 2018. CA Cancer J

Clin 68: 7-30.

152

69. Klotz, L., L. Zhang, A. Lam, R. Nam, A. Mamedov, and A. Loblaw. 2010. Clinical results of long-term follow-up of a large, active surveillance cohort with localized prostate cancer. J Clin Oncol 28: 126-131.

70. Selvadurai, E. D., M. Singhera, K. Thomas, K. Mohammed, R. Woode-Amissah, A.

Horwich, R. A. Huddart, D. P. Dearnaley, and C. C. Parker. 2013. Medium-term outcomes of active surveillance for localised prostate cancer. Eur Urol 64: 981-987.

71. Paller, C. J., E. S. Antonarakis, M. A. Eisenberger, and M. A. Carducci. 2013.

Management of patients with biochemical recurrence after local therapy for prostate cancer.

Hematol Oncol Clin North Am 27: 1205-1219, viii.

72. Gandaglia, G., F. Abdollah, J. Schiffmann, V. Trudeau, S. F. Shariat, S. P. Kim, P.

Perrotte, F. Montorsi, A. Briganti, Q. D. Trinh, P. I. Karakiewicz, and M. Sun. 2014.

Distribution of metastatic sites in patients with prostate cancer: A population-based analysis.

Prostate 74: 210-216.

73. Pond, G. R., G. Sonpavde, R. de Wit, M. A. Eisenberger, I. F. Tannock, and A. J.

Armstrong. 2014. The prognostic importance of metastatic site in men with metastatic castration-resistant prostate cancer. Eur Urol 65: 3-6.

74. Petrylak, D. P., C. M. Tangen, M. H. Hussain, P. N. Lara, Jr., J. A. Jones, M. E.

Taplin, P. A. Burch, D. Berry, C. Moinpour, M. Kohli, M. C. Benson, E. J. Small, D.

Raghavan, and E. D. Crawford. 2004. Docetaxel and estramustine compared with mitoxantrone and prednisone for advanced refractory prostate cancer. The New England journal of medicine 351: 1513-1520.

75. Tannock, I. F., R. de Wit, W. R. Berry, J. Horti, A. Pluzanska, K. N. Chi, S. Oudard,

C. Theodore, N. D. James, I. Turesson, M. A. Rosenthal, M. A. Eisenberger, and T. A. X.

Investigators. 2004. Docetaxel plus prednisone or mitoxantrone plus prednisone for advanced prostate cancer. The New England journal of medicine 351: 1502-1512.

153

76. Kantoff, P. W., C. S. Higano, N. D. Shore, E. R. Berger, E. J. Small, D. F. Penson, C.

H. Redfern, A. C. Ferrari, R. Dreicer, R. B. Sims, Y. Xu, M. W. Frohlich, P. F. Schellhammer, and I. S. Investigators. 2010. Sipuleucel-T immunotherapy for castration-resistant prostate cancer. The New England journal of medicine 363: 411-422.

77. de Bono, J. S., C. J. Logothetis, A. Molina, K. Fizazi, S. North, L. Chu, K. N. Chi, R.

J. Jones, O. B. Goodman, Jr., F. Saad, J. N. Staffurth, P. Mainwaring, S. Harland, T. W.

Flaig, T. E. Hutson, T. Cheng, H. Patterson, J. D. Hainsworth, C. J. Ryan, C. N. Sternberg,

S. L. Ellard, A. Flechon, M. Saleh, M. Scholz, E. Efstathiou, A. Zivi, D. Bianchini, Y. Loriot,

N. Chieffo, T. Kheoh, C. M. Haqq, H. I. Scher, and C.-A.-. Investigators. 2011. Abiraterone and increased survival in metastatic prostate cancer. The New England journal of medicine

364: 1995-2005.

78. Ryan, C. J., A. Molina, and T. Griffin. 2013. Abiraterone in metastatic prostate cancer. The New England journal of medicine 368: 1458-1459.

79. Beer, T. M., A. J. Armstrong, D. E. Rathkopf, Y. Loriot, C. N. Sternberg, C. S.

Higano, P. Iversen, S. Bhattacharya, J. Carles, S. Chowdhury, I. D. Davis, J. S. de Bono, C.

P. Evans, K. Fizazi, A. M. Joshua, C.-S. Kim, G. Kimura, P. Mainwaring, H. Mansbach, K.

Miller, S. B. Noonberg, F. Perabo, D. Phung, F. Saad, H. I. Scher, M.-E. Taplin, P. M.

Venner, and B. Tombal. 2014. Enzalutamide in Metastatic Prostate Cancer before

Chemotherapy. New England Journal of Medicine 371: 424-433.

80. Scher, H. I., K. Fizazi, F. Saad, M.-E. Taplin, C. N. Sternberg, K. Miller, R. de Wit, P.

Mulders, K. N. Chi, N. D. Shore, A. J. Armstrong, T. W. Flaig, A. Fléchon, P. Mainwaring, M.

Fleming, J. D. Hainsworth, M. Hirmand, B. Selby, L. Seely, and J. S. de Bono. 2012.

Increased Survival with Enzalutamide in Prostate Cancer after Chemotherapy. New England

Journal of Medicine 367: 1187-1197.

81. Tran, C., S. Ouk, N. J. Clegg, Y. Chen, P. A. Watson, V. Arora, J. Wongvipat, P. M.

154

Smith-Jones, D. Yoo, A. Kwon, T. Wasielewska, D. Welsbie, C. D. Chen, C. S. Higano, T.

M. Beer, D. T. Hung, H. I. Scher, M. E. Jung, and C. L. Sawyers. 2009. Development of a

second-generation antiandrogen for treatment of advanced prostate cancer. Science 324:

787-790.

82. de Bono, J. S., S. Oudard, M. Ozguroglu, S. Hansen, J. P. Machiels, I. Kocak, G.

Gravis, I. Bodrogi, M. J. Mackenzie, L. Shen, M. Roessner, S. Gupta, A. O. Sartor, and T.

Investigators. 2010. Prednisone plus cabazitaxel or mitoxantrone for metastatic castration- resistant prostate cancer progressing after docetaxel treatment: a randomised open-label trial. Lancet 376: 1147-1154.

83. Eisenberger, M., A. C. Hardy-Bessard, C. S. Kim, L. Geczi, D. Ford, L. Mourey, J.

Carles, P. Parente, A. Font, G. Kacso, M. Chadjaa, W. Zhang, J. Bernard, and J. de Bono.

2017. Phase III Study Comparing a Reduced Dose of Cabazitaxel (20 mg/m(2)) and the

Currently Approved Dose (25 mg/m(2)) in Postdocetaxel Patients With Metastatic

Castration-Resistant Prostate Cancer-PROSELICA. J Clin Oncol 35: 3198-3206.

84. Henriksen, G., K. Breistol, O. S. Bruland, O. Fodstad, and R. H. Larsen. 2002.

Significant antitumor effect from bone-seeking, alpha-particle-emitting (223)Ra demonstrated in an experimental skeletal metastases model. Cancer Res 62: 3120-3125.

85. Parker, C., and O. Sartor. 2013. Radium-223 in prostate cancer. The New England journal of medicine 369: 1659-1660.

86. Barbieri, C. E., S. C. Baca, M. S. Lawrence, F. Demichelis, M. Blattner, J. P.

Theurillat, T. A. White, P. Stojanov, E. Van Allen, N. Stransky, E. Nickerson, S. S. Chae, G.

Boysen, D. Auclair, R. C. Onofrio, K. Park, N. Kitabayashi, T. Y. MacDonald, K. Sheikh, T.

Vuong, C. Guiducci, K. Cibulskis, A. Sivachenko, S. L. Carter, G. Saksena, D. Voet, W. M.

Hussain, A. H. Ramos, W. Winckler, M. C. Redman, K. Ardlie, A. K. Tewari, J. M. Mosquera,

N. Rupp, P. J. Wild, H. Moch, C. Morrissey, P. S. Nelson, P. W. Kantoff, S. B. Gabriel, T. R.

155

Golub, M. Meyerson, E. S. Lander, G. Getz, M. A. Rubin, and L. A. Garraway. 2012. Exome

sequencing identifies recurrent SPOP, FOXA1 and MED12 mutations in prostate cancer.

Nat Genet 44: 685-689.

87. Grasso, C. S., Y. M. Wu, D. R. Robinson, X. Cao, S. M. Dhanasekaran, A. P. Khan,

M. J. Quist, X. Jing, R. J. Lonigro, J. C. Brenner, I. A. Asangani, B. Ateeq, S. Y. Chun, J.

Siddiqui, L. Sam, M. Anstett, R. Mehra, J. R. Prensner, N. Palanisamy, G. A. Ryslik, F.

Vandin, B. J. Raphael, L. P. Kunju, D. R. Rhodes, K. J. Pienta, A. M. Chinnaiyan, and S. A.

Tomlins. 2012. The mutational landscape of lethal castration-resistant prostate cancer.

Nature 487: 239-243.

88. Robinson, D., E. M. Van Allen, Y. M. Wu, N. Schultz, R. J. Lonigro, J. M. Mosquera,

B. Montgomery, M. E. Taplin, C. C. Pritchard, G. Attard, H. Beltran, W. Abida, R. K. Bradley,

J. Vinson, X. Cao, P. Vats, L. P. Kunju, M. Hussain, F. Y. Feng, S. A. Tomlins, K. A.

Cooney, D. C. Smith, C. Brennan, J. Siddiqui, R. Mehra, Y. Chen, D. E. Rathkopf, M. J.

Morris, S. B. Solomon, J. C. Durack, V. E. Reuter, A. Gopalan, J. Gao, M. Loda, R. T. Lis,

M. Bowden, S. P. Balk, G. Gaviola, C. Sougnez, M. Gupta, E. Y. Yu, E. A. Mostaghel, H. H.

Cheng, H. Mulcahy, L. D. True, S. R. Plymate, H. Dvinge, R. Ferraldeschi, P. Flohr, S.

Miranda, Z. Zafeiriou, N. Tunariu, J. Mateo, R. Perez-Lopez, F. Demichelis, B. D. Robinson,

M. Schiffman, D. M. Nanus, S. T. Tagawa, A. Sigaras, K. W. Eng, O. Elemento, A. Sboner,

E. I. Heath, H. I. Scher, K. J. Pienta, P. Kantoff, J. S. de Bono, M. A. Rubin, P. S. Nelson, L.

A. Garraway, C. L. Sawyers, and A. M. Chinnaiyan. 2015. Integrative clinical genomics of

advanced prostate cancer. Cell 161: 1215-1228.

89. Baca, S. C., D. Prandi, M. S. Lawrence, J. M. Mosquera, A. Romanel, Y. Drier, K.

Park, N. Kitabayashi, T. Y. MacDonald, M. Ghandi, E. Van Allen, G. V. Kryukov, A. Sboner,

J. P. Theurillat, T. D. Soong, E. Nickerson, D. Auclair, A. Tewari, H. Beltran, R. C. Onofrio,

G. Boysen, C. Guiducci, C. E. Barbieri, K. Cibulskis, A. Sivachenko, S. L. Carter, G.

156

Saksena, D. Voet, A. H. Ramos, W. Winckler, M. Cipicchio, K. Ardlie, P. W. Kantoff, M. F.

Berger, S. B. Gabriel, T. R. Golub, M. Meyerson, E. S. Lander, O. Elemento, G. Getz, F.

Demichelis, M. A. Rubin, and L. A. Garraway. 2013. Punctuated evolution of prostate cancer

genomes. Cell 153: 666-677.

90. Berger, M. F., M. S. Lawrence, F. Demichelis, Y. Drier, K. Cibulskis, A. Y.

Sivachenko, A. Sboner, R. Esgueva, D. Pflueger, C. Sougnez, R. Onofrio, S. L. Carter, K.

Park, L. Habegger, L. Ambrogio, T. Fennell, M. Parkin, G. Saksena, D. Voet, A. H. Ramos,

T. J. Pugh, J. Wilkinson, S. Fisher, W. Winckler, S. Mahan, K. Ardlie, J. Baldwin, J. W.

Simons, N. Kitabayashi, T. Y. MacDonald, P. W. Kantoff, L. Chin, S. B. Gabriel, M. B.

Gerstein, T. R. Golub, M. Meyerson, A. Tewari, E. S. Lander, G. Getz, M. A. Rubin, and L.

A. Garraway. 2011. The genomic complexity of primary human prostate cancer. Nature 470:

214-220.

91. Tomlins, S. A., D. R. Rhodes, S. Perner, S. M. Dhanasekaran, R. Mehra, X. W. Sun,

S. Varambally, X. Cao, J. Tchinda, R. Kuefer, C. Lee, J. E. Montie, R. B. Shah, K. J. Pienta,

M. A. Rubin, and A. M. Chinnaiyan. 2005. Recurrent fusion of TMPRSS2 and ETS transcription factor genes in prostate cancer. Science 310: 644-648.

92. Lin, C., L. Yang, B. Tanasa, K. Hutt, B. G. Ju, K. Ohgi, J. Zhang, D. W. Rose, X. D.

Fu, C. K. Glass, and M. G. Rosenfeld. 2009. Nuclear receptor-induced chromosomal proximity and DNA breaks underlie specific translocations in cancer. Cell 139: 1069-1083.

93. Rubin, M. A., C. A. Maher, and A. M. Chinnaiyan. 2011. Common gene rearrangements in prostate cancer. J Clin Oncol 29: 3659-3668.

94. Young, A., N. Palanisamy, J. Siddiqui, D. P. Wood, J. T. Wei, A. M. Chinnaiyan, L. P.

Kunju, and S. A. Tomlins. 2012. Correlation of urine TMPRSS2:ERG and PCA3 to ERG+ and total prostate cancer burden. Am J Clin Pathol 138: 685-696.

95. Dennis, L. K., C. F. Lynch, and J. C. Torner. 2002. Epidemiologic association

157

between prostatitis and prostate cancer. Urology 60: 78-83.

96. Roberts, R. O., E. J. Bergstralh, S. E. Bass, M. M. Lieber, and S. J. Jacobsen. 2004.

Prostatitis as a risk factor for prostate cancer. Epidemiology 15: 93-99.

97. Simons, J. W., B. Mikhak, J. F. Chang, A. M. DeMarzo, M. A. Carducci, M. Lim, C. E.

Weber, A. A. Baccala, M. A. Goemann, S. M. Clift, D. G. Ando, H. I. Levitsky, L. K. Cohen,

M. G. Sanda, R. C. Mulligan, A. W. Partin, H. B. Carter, S. Piantadosi, F. F. Marshall, and

W. G. Nelson. 1999. Induction of immunity to prostate cancer antigens: results of a clinical

trial of vaccination with irradiated autologous prostate tumor cells engineered to secrete

granulocyte-macrophage colony-stimulating factor using ex vivo gene transfer. Cancer Res

59: 5160-5168.

98. Simons, J. W., M. A. Carducci, B. Mikhak, M. Lim, B. Biedrzycki, F. Borellini, S. M.

Clift, K. M. Hege, D. G. Ando, S. Piantadosi, R. Mulligan, and W. G. Nelson. 2006. Phase I/II trial of an allogeneic cellular immunotherapy in hormone-naive prostate cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 12:

3394-3401.

99. Simons, J. W., and N. Sacks. 2006. Granulocyte-macrophage colony-stimulating factor-transduced allogeneic cancer cellular immunotherapy: the GVAX vaccine for prostate cancer. Urol Oncol 24: 419-424.

100. Drake, C. G., and E. S. Antonarakis. 2010. Update: immunological strategies for prostate cancer. Curr Urol Rep 11: 202-207.

101. Huber, M. L., L. Haynes, C. Parker, and P. Iversen. 2012. Interdisciplinary critique of sipuleucel-T as immunotherapy in castration-resistant prostate cancer. J Natl Cancer Inst

104: 273-279.

102. Kantoff, P. W., T. J. Schuetz, B. A. Blumenstein, L. M. Glode, D. L. Bilhartz, M.

Wyand, K. Manson, D. L. Panicali, R. Laus, J. Schlom, W. L. Dahut, P. M. Arlen, J. L.

158

Gulley, and W. R. Godfrey. 2010. Overall survival analysis of a phase II randomized

controlled trial of a Poxviral-based PSA-targeted immunotherapy in metastatic castration- resistant prostate cancer. J Clin Oncol 28: 1099-1105.

103. Gulley, J. L., M. Borre, N. J. Vogelzang, S. Ng, N. Agarwal, C. C. Parker, D. W.

Pook, P. Rathenborg, T. W. Flaig, J. Carles, N. D. Shore, L. Chen, C. R. Heery, W. R.

Gerritsen, F. Priou, and P. W. Kantoff. 2018. Results of PROSPECT: A randomized phase 3 trial of PROSTVAC-V/F (PRO) in men with asymptomatic or minimally symptomatic metastatic, castration-resistant prostate cancer. Journal of Clinical Oncology 36: 5006-5006.

104. Kwon, E. D., C. G. Drake, H. I. Scher, K. Fizazi, A. Bossi, A. J. van den Eertwegh, M.

Krainer, N. Houede, R. Santos, H. Mahammedi, S. Ng, M. Maio, F. A. Franke, S. Sundar, N.

Agarwal, A. M. Bergman, T. E. Ciuleanu, E. Korbenfeld, L. Sengelov, S. Hansen, C.

Logothetis, T. M. Beer, M. B. McHenry, P. Gagnier, D. Liu, and W. R. Gerritsen. 2014.

Ipilimumab versus placebo after radiotherapy in patients with metastatic castration-resistant prostate cancer that had progressed after docetaxel chemotherapy (CA184-043): a multicentre, randomised, double-blind, phase 3 trial. Lancet Oncol 15: 700-712.

105. Beer, T. M., E. D. Kwon, C. G. Drake, K. Fizazi, C. Logothetis, G. Gravis, V. Ganju,

J. Polikoff, F. Saad, P. Humanski, J. M. Piulats, P. Gonzalez Mella, S. S. Ng, D. Jaeger, F.

X. Parnis, F. A. Franke, J. Puente, R. Carvajal, L. Sengelov, M. B. McHenry, A. Varma, A. J. van den Eertwegh, and W. Gerritsen. 2017. Randomized, Double-Blind, Phase III Trial of

Ipilimumab Versus Placebo in Asymptomatic or Minimally Symptomatic Patients With

Metastatic Chemotherapy-Naive Castration-Resistant Prostate Cancer. J Clin Oncol 35: 40-

47.

106. Scholz, M., S. Yep, M. Chancey, C. Kelly, K. Chau, J. Turner, R. Lam, and C. G.

Drake. 2017. Phase I clinical trial of sipuleucel-T combined with escalating doses of ipilimumab in progressive metastatic castrate-resistant prostate cancer. Immunotargets Ther

159

6: 11-16.

107. Haffner, M. C., G. Guner, D. Taheri, G. J. Netto, D. N. Palsgrove, Q. Zheng, L. B.

Guedes, K. Kim, H. Tsai, D. M. Esopi, T. L. Lotan, R. Sharma, A. K. Meeker, A. M.

Chinnaiyan, W. G. Nelson, S. Yegnasubramanian, J. Luo, R. Mehra, E. S. Antonarakis, C.

G. Drake, and A. M. De Marzo. 2018. Comprehensive Evaluation of Programmed Death-

Ligand 1 Expression in Primary and Metastatic Prostate Cancer. Am J Pathol 188: 1478-

1485.

108. Bono, J. S. D., J. C. Goh, K. Ojamaa, J. M. P. Rodriguez, C. G. Drake, C. J. Hoimes,

H. Wu, C. H. Poehlein, and E. S. Antonarakis. 2018. KEYNOTE-199: Pembrolizumab

(pembro) for docetaxel-refractory metastatic castration-resistant prostate cancer (mCRPC).

Journal of Clinical Oncology 36: 5007-5007.

109. Pritchard, C. C., C. Morrissey, A. Kumar, X. Zhang, C. Smith, I. Coleman, S. J.

Salipante, J. Milbank, M. Yu, W. M. Grady, J. F. Tait, E. Corey, R. L. Vessella, T. Walsh, J.

Shendure, and P. S. Nelson. 2014. Complex MSH2 and MSH6 mutations in hypermutated

microsatellite unstable advanced prostate cancer. Nature Communications 5: 4988.

110. Stein, M. N., L. Fong, R. F. Tutrone, A. E. Mega, M. Lobo, Q. Hong, and N. B. Haas.

2018. KEYNOTE-046: ADXS-PSA plus pembrolizumab (pembro) in metastatic castration-

resistant prostate cancer (mCRPC). Journal of Clinical Oncology 36: 5019-5019.

111. Graff, J. N., J. J. Alumkal, R. F. Thompson, A. Moran, G. V. Thomas, M. A. Wood, C.

G. Drake, R. Slottke, and T. M. Beer. 2018. Pembrolizumab (Pembro) plus enzalutamide

(Enz) in metastatic castration resistant prostate cancer (mCRPC): Extended follow up.

Journal of Clinical Oncology 36: 5047-5047.

112. Boudadi, K., D. L. Suzman, V. Anagnostou, W. Fu, B. Luber, H. Wang, N. Niknafs, J.

R. White, J. L. Silberstein, R. Sullivan, D. Dowling, R. Harb, T. R. Nirschl, B. A. Veeneman,

S. A. Tomlins, Y. Wang, A. Jendrisak, R. P. Graf, R. Dittamore, M. A. Carducci, M. A.

160

Eisenberger, M. C. Haffner, A. K. Meeker, J. R. Eshleman, J. Luo, V. E. Velculescu, C. G.

Drake, and E. S. Antonarakis. 2018. Ipilimumab plus nivolumab and DNA-repair defects in

AR-V7-expressing metastatic prostate cancer. Oncotarget 9: 28561-28571.

113. Calcinotto, A., C. Spataro, E. Zagato, D. Di Mitri, V. Gil, M. Crespo, G. De Bernardis,

M. Losa, M. Mirenda, E. Pasquini, A. Rinaldi, S. Sumanasuriya, M. B. Lambros, A. Neeb, R.

Luciano, C. A. Bravi, D. Nava-Rodrigues, D. Dolling, T. Prayer-Galetti, A. Ferreira, A.

Briganti, A. Esposito, S. Barry, W. Yuan, A. Sharp, J. de Bono, and A. Alimonti. 2018. IL-23 secreted by myeloid cells drives castration-resistant prostate cancer. Nature 559: 363-369.

114. Ager, C. R., M. J. Reilley, C. Nicholas, T. Bartkowiak, A. R. Jaiswal, and M. A.

Curran. 2017. Intratumoral STING Activation with T-cell Checkpoint Modulation Generates

Systemic Antitumor Immunity. Cancer Immunol Res 5: 676-684.

115. Lu, X., J. W. Horner, E. Paul, X. Shang, P. Troncoso, P. Deng, S. Jiang, Q. Chang,

D. J. Spring, P. Sharma, J. A. Zebala, D. Y. Maeda, Y. A. Wang, and R. A. DePinho. 2017.

Effective combinatorial immunotherapy for castration-resistant prostate cancer. Nature 543:

728-732.

116. Canto, M. I., F. Harinck, R. H. Hruban, G. J. Offerhaus, J. W. Poley, I. Kamel, Y. Nio,

R. S. Schulick, C. Bassi, I. Kluijt, M. J. Levy, A. Chak, P. Fockens, M. Goggins, M. Bruno, and C. International Cancer of Pancreas Screening. 2013. International Cancer of the

Pancreas Screening (CAPS) Consortium summit on the management of patients with increased risk for familial pancreatic cancer. Gut 62: 339-347.

117. Kenner, B. J., V. L. W. Go, S. T. Chari, A. E. Goldberg, and L. J. Rothschild. 2017.

Early Detection of Pancreatic Cancer: The Role of Industry in the Development of

Biomarkers. Pancreas 46: 1238-1241.

118. Oettle, H., P. Neuhaus, A. Hochhaus, J. T. Hartmann, K. Gellert, K. Ridwelski, M.

Niedergethmann, C. Zulke, J. Fahlke, M. B. Arning, M. Sinn, A. Hinke, and H. Riess. 2013.

161

Adjuvant chemotherapy with gemcitabine and long-term outcomes among patients with resected pancreatic cancer: the CONKO-001 randomized trial. JAMA 310: 1473-1481.

119. Burris, H. A., 3rd, M. J. Moore, J. Andersen, M. R. Green, M. L. Rothenberg, M. R.

Modiano, M. C. Cripps, R. K. Portenoy, A. M. Storniolo, P. Tarassoff, R. Nelson, F. A. Dorr,

C. D. Stephens, and D. D. Von Hoff. 1997. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 15: 2403-2413.

120. Conroy, T., F. Desseigne, M. Ychou, O. Bouche, R. Guimbaud, Y. Becouarn, A.

Adenis, J. L. Raoul, S. Gourgou-Bourgade, C. de la Fouchardiere, J. Bennouna, J. B.

Bachet, F. Khemissa-Akouz, D. Pere-Verge, C. Delbaldo, E. Assenat, B. Chauffert, P.

Michel, C. Montoto-Grillot, M. Ducreux, U. Groupe Tumeurs Digestives of, and P. Intergroup.

2011. FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. The New England journal of medicine 364: 1817-1825.

121. Von Hoff, D. D., T. Ervin, F. P. Arena, E. G. Chiorean, J. Infante, M. Moore, T. Seay,

S. A. Tjulandin, W. W. Ma, M. N. Saleh, M. Harris, M. Reni, S. Dowden, D. Laheru, N.

Bahary, R. K. Ramanathan, J. Tabernero, M. Hidalgo, D. Goldstein, E. Van Cutsem, X. Wei,

J. Iglesias, and M. F. Renschler. 2013. Increased Survival in Pancreatic Cancer with nab-

Paclitaxel plus Gemcitabine. New England Journal of Medicine 369: 1691-1703.

122. Kleeff, J., M. Korc, M. Apte, C. La Vecchia, C. D. Johnson, A. V. Biankin, R. E.

Neale, M. Tempero, D. A. Tuveson, R. H. Hruban, and J. P. Neoptolemos. 2016. Pancreatic cancer. Nat Rev Dis Primers 2: 16022.

123. Kanda, M., H. Matthaei, J. Wu, S. M. Hong, J. Yu, M. Borges, R. H. Hruban, A.

Maitra, K. Kinzler, B. Vogelstein, and M. Goggins. 2012. Presence of somatic mutations in most early-stage pancreatic intraepithelial neoplasia. Gastroenterology 142: 730-733.e739.

124. Aguirre, A. J., N. Bardeesy, M. Sinha, L. Lopez, D. A. Tuveson, J. Horner, M. S.

162

Redston, and R. A. DePinho. 2003. Activated Kras and Ink4a/Arf deficiency cooperate to

produce metastatic pancreatic ductal adenocarcinoma. Genes Dev 17: 3112-3126.

125. Bardeesy, N., A. J. Aguirre, G. C. Chu, K.-h. Cheng, L. V. Lopez, A. F. Hezel, B.

Feng, C. Brennan, R. Weissleder, U. Mahmood, D. Hanahan, M. S. Redston, L. Chin, and

R. A. DePinho. 2006. Both p16^Ink4a and the p19^Arf-p53 pathway

constrain progression of pancreatic adenocarcinoma in the mouse. Proceedings of the

National Academy of Sciences 103: 5947-5952.

126. Hustinx, S. R., L. M. Leoni, C. J. Yeo, P. N. Brown, M. Goggins, S. E. Kern, R. H.

Hruban, and A. Maitra. 2005. Concordant loss of MTAP and p16/CDKN2A expression in pancreatic intraepithelial neoplasia: evidence of homozygous deletion in a noninvasive

precursor lesion. Modern pathology : an official journal of the United States and Canadian

Academy of Pathology, Inc 18: 959-963.

127. Hingorani, S. R., E. F. Petricoin, A. Maitra, V. Rajapakse, C. King, M. A. Jacobetz, S.

Ross, T. P. Conrads, T. D. Veenstra, B. A. Hitt, Y. Kawaguchi, D. Johann, L. A. Liotta, H. C.

Crawford, M. E. Putt, T. Jacks, C. V. Wright, R. H. Hruban, A. M. Lowy, and D. A. Tuveson.

2003. Preinvasive and invasive ductal pancreatic cancer and its early detection in the

mouse. Cancer Cell 4: 437-450.

128. Hingorani, S. R., L. Wang, A. S. Multani, C. Combs, T. B. Deramaudt, R. H. Hruban,

A. K. Rustgi, S. Chang, and D. A. Tuveson. 2005. Trp53R172H and KrasG12D cooperate to

promote chromosomal instability and widely metastatic pancreatic ductal adenocarcinoma in

mice. Cancer Cell 7: 469-483.

129. Izeradjene, K., C. Combs, M. Best, A. Gopinathan, A. Wagner, W. M. Grady, C. X.

Deng, R. H. Hruban, N. V. Adsay, D. A. Tuveson, and S. R. Hingorani. 2007. Kras(G12D)

and Smad4/Dpc4 haploinsufficiency cooperate to induce mucinous cystic neoplasms and

invasive adenocarcinoma of the pancreas. Cancer Cell 11: 229-243.

163

130. Bardeesy, N., K. H. Cheng, J. H. Berger, G. C. Chu, J. Pahler, P. Olson, A. F. Hezel,

J. Horner, G. Y. Lauwers, D. Hanahan, and R. A. DePinho. 2006. Smad4 is dispensable for

normal pancreas development yet critical in progression and tumor biology of pancreas

cancer. Genes Dev 20: 3130-3146.

131. Ostrem, J. M., and K. M. Shokat. 2016. Direct small-molecule inhibitors of KRAS:

from structural insights to mechanism-based design. Nat Rev Drug Discov 15: 771-785.

132. Erkan, M., J. Kleeff, A. Gorbachevski, C. Reiser, T. Mitkus, I. Esposito, T. Giese, M.

W. Buchler, N. A. Giese, and H. Friess. 2007. Periostin creates a tumor-supportive

microenvironment in the pancreas by sustaining fibrogenic stellate cell activity.

Gastroenterology 132: 1447-1464.

133. Lohr, M., C. Schmidt, J. Ringel, M. Kluth, P. Muller, H. Nizze, and R. Jesnowski.

2001. Transforming growth factor-beta1 induces desmoplasia in an experimental model of

human pancreatic carcinoma. Cancer Res 61: 550-555.

134. Ohuchida, K., K. Mizumoto, M. Murakami, L. W. Qian, N. Sato, E. Nagai, K.

Matsumoto, T. Nakamura, and M. Tanaka. 2004. Radiation to stromal fibroblasts increases

invasiveness of pancreatic cancer cells through tumor-stromal interactions. Cancer Res 64:

3215-3222.

135. Bachem, M. G., M. Schunemann, M. Ramadani, M. Siech, H. Beger, A. Buck, S.

Zhou, A. Schmid-Kotsas, and G. Adler. 2005. Pancreatic carcinoma cells induce fibrosis by

stimulating proliferation and matrix synthesis of stellate cells. Gastroenterology 128: 907-

921.

136. Hwang, R. F., T. Moore, T. Arumugam, V. Ramachandran, K. D. Amos, A. Rivera, B.

Ji, D. B. Evans, and C. D. Logsdon. 2008. Cancer-associated stromal fibroblasts promote

pancreatic tumor progression. Cancer Res 68: 918-926.

137. Olive, K. P., M. A. Jacobetz, C. J. Davidson, A. Gopinathan, D. McIntyre, D. Honess,

164

B. Madhu, M. A. Goldgraben, M. E. Caldwell, D. Allard, K. K. Frese, G. Denicola, C. Feig, C.

Combs, S. P. Winter, H. Ireland-Zecchini, S. Reichelt, W. J. Howat, A. Chang, M. Dhara, L.

Wang, F. Ruckert, R. Grutzmann, C. Pilarsky, K. Izeradjene, S. R. Hingorani, P. Huang, S.

E. Davies, W. Plunkett, M. Egorin, R. H. Hruban, N. Whitebread, K. McGovern, J. Adams, C.

Iacobuzio-Donahue, J. Griffiths, and D. A. Tuveson. 2009. Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science 324:

1457-1461.

138. Koong, A. C., V. K. Mehta, Q. T. Le, G. A. Fisher, D. J. Terris, J. M. Brown, A. J.

Bastidas, and M. Vierra. 2000. Pancreatic tumors show high levels of hypoxia. Int J Radiat

Oncol Biol Phys 48: 919-922.

139. Masamune, A., K. Kikuta, T. Watanabe, K. Satoh, M. Hirota, and T. Shimosegawa.

2008. Hypoxia stimulates pancreatic stellate cells to induce fibrosis and angiogenesis in pancreatic cancer. Am J Physiol Gastrointest Liver Physiol 295: G709-717.

140. Brizel, D. M., S. P. Scully, J. M. Harrelson, L. J. Layfield, J. M. Bean, L. R. Prosnitz, and M. W. Dewhirst. 1996. Tumor oxygenation predicts for the likelihood of distant metastases in human soft tissue sarcoma. Cancer Res 56: 941-943.

141. Ying, H., A. C. Kimmelman, C. A. Lyssiotis, S. Hua, G. C. Chu, E. Fletcher-

Sananikone, J. W. Locasale, J. Son, H. Zhang, J. L. Coloff, H. Yan, W. Wang, S. Chen, A.

Viale, H. Zheng, J. H. Paik, C. Lim, A. R. Guimaraes, E. S. Martin, J. Chang, A. F. Hezel, S.

R. Perry, J. Hu, B. Gan, Y. Xiao, J. M. Asara, R. Weissleder, Y. A. Wang, L. Chin, L. C.

Cantley, and R. A. DePinho. 2012. Oncogenic Kras maintains pancreatic tumors through regulation of anabolic glucose metabolism. Cell 149: 656-670.

142. Commisso, C., S. M. Davidson, R. G. Soydaner-Azeloglu, S. J. Parker, J. J.

Kamphorst, S. Hackett, E. Grabocka, M. Nofal, J. A. Drebin, C. B. Thompson, J. D.

Rabinowitz, C. M. Metallo, M. G. Vander Heiden, and D. Bar-Sagi. 2013. Macropinocytosis

165

of protein is an amino acid supply route in Ras-transformed cells. Nature 497: 633-637.

143. Yang, S., X. Wang, G. Contino, M. Liesa, E. Sahin, H. Ying, A. Bause, Y. Li, J. M.

Stommel, G. Dell'Antonio, J. Mautner, G. Tonon, M. Haigis, O. S. Shirihai, C. Doglioni, N.

Bardeesy, and A. C. Kimmelman. 2011. Pancreatic cancers require autophagy for tumor

growth. Genes & Development 25: 717-729.

144. DeNicola, G. M., F. A. Karreth, T. J. Humpton, A. Gopinathan, C. Wei, K. Frese, D.

Mangal, K. H. Yu, C. J. Yeo, E. S. Calhoun, F. Scrimieri, J. M. Winter, R. H. Hruban, C.

Iacobuzio-Donahue, S. E. Kern, I. A. Blair, and D. A. Tuveson. 2011. Oncogene-induced

Nrf2 transcription promotes ROS detoxification and tumorigenesis. Nature 475: 106-109.

145. Leblond, M. M., A. N. Gerault, A. Corroyer-Dulmont, E. T. MacKenzie, E. Petit, M.

Bernaudin, and S. Valable. 2016. Hypoxia induces macrophage polarization and re- education toward an M2 phenotype in U87 and U251 glioblastoma models.

Oncoimmunology 5: e1056442.

146. Clark, C. E., S. R. Hingorani, R. Mick, C. Combs, D. A. Tuveson, and R. H.

Vonderheide. 2007. Dynamics of the immune reaction to pancreatic cancer from inception to invasion. Cancer Res 67: 9518-9527.

147. Bayne, L. J., G. L. Beatty, N. Jhala, C. E. Clark, A. D. Rhim, B. Z. Stanger, and R. H.

Vonderheide. 2012. Tumor-derived granulocyte-macrophage colony-stimulating factor

regulates myeloid inflammation and T cell immunity in pancreatic cancer. Cancer Cell 21:

822-835.

148. Pylayeva-Gupta, Y., K. E. Lee, C. H. Hajdu, G. Miller, and D. Bar-Sagi. 2012.

Oncogenic Kras-induced GM-CSF production promotes the development of pancreatic

neoplasia. Cancer Cell 21: 836-847.

149. Ancrile, B., K. H. Lim, and C. M. Counter. 2007. Oncogenic Ras-induced secretion of

IL6 is required for tumorigenesis. Genes Dev 21: 1714-1719.

166

150. Beatty, G. L., R. Winograd, R. A. Evans, K. B. Long, S. L. Luque, J. W. Lee, C.

Clendenin, W. L. Gladney, D. M. Knoblock, P. D. Guirnalda, and R. H. Vonderheide. 2015.

Exclusion of T Cells From Pancreatic Carcinomas in Mice Is Regulated by Ly6C(low)

F4/80(+) Extratumoral Macrophages. Gastroenterology 149: 201-210.

151. Zhu, Y., J. M. Herndon, D. K. Sojka, K. W. Kim, B. L. Knolhoff, C. Zuo, D. R.

Cullinan, J. Luo, A. R. Bearden, K. J. Lavine, W. M. Yokoyama, W. G. Hawkins, R. C. Fields,

G. J. Randolph, and D. G. DeNardo. 2017. Tissue-Resident Macrophages in Pancreatic

Ductal Adenocarcinoma Originate from Embryonic Hematopoiesis and Promote Tumor

Progression. Immunity 47: 323-338 e326.

152. Candido, J. B., J. P. Morton, P. Bailey, A. D. Campbell, S. A. Karim, T. Jamieson, L.

Lapienyte, A. Gopinathan, W. Clark, E. J. McGhee, J. Wang, M. Escorcio-Correia, R.

Zollinger, R. Roshani, L. Drew, L. Rishi, R. Arkell, T. R. J. Evans, C. Nixon, D. I. Jodrell, R.

W. Wilkinson, A. V. Biankin, S. T. Barry, F. R. Balkwill, and O. J. Sansom. 2018. CSF1R(+)

Macrophages Sustain Pancreatic Tumor Growth through T Cell Suppression and

Maintenance of Key Gene Programs that Define the Squamous Subtype. Cell Rep 23: 1448-

1460.

153. Franklin, R. A., W. Liao, A. Sarkar, M. V. Kim, M. R. Bivona, K. Liu, E. G. Pamer, and

M. O. Li. 2014. The cellular and molecular origin of tumor-associated macrophages. Science

344: 921-925.

154. Movahedi, K., D. Laoui, C. Gysemans, M. Baeten, G. Stange, J. Van den Bossche,

M. Mack, D. Pipeleers, P. In't Veld, P. De Baetselier, and J. A. Van Ginderachter. 2010.

Different tumor microenvironments contain functionally distinct subsets of macrophages derived from Ly6C(high) monocytes. Cancer Res 70: 5728-5739.

155. Jaffee, E. M., R. H. Hruban, B. Biedrzycki, D. Laheru, K. Schepers, P. R. Sauter, M.

Goemann, J. Coleman, L. Grochow, R. C. Donehower, K. D. Lillemoe, S. O'Reilly, R. A.

167

Abrams, D. M. Pardoll, J. L. Cameron, and C. J. Yeo. 2001. Novel allogeneic granulocyte- macrophage colony-stimulating factor-secreting tumor vaccine for pancreatic cancer: a

phase I trial of safety and immune activation. J Clin Oncol 19: 145-156.

156. Lutz, E., C. J. Yeo, K. D. Lillemoe, B. Biedrzycki, B. Kobrin, J. Herman, E. Sugar, S.

Piantadosi, J. L. Cameron, S. Solt, B. Onners, I. Tartakovsky, M. Choi, R. Sharma, P. B.

Illei, R. H. Hruban, R. A. Abrams, D. Le, E. Jaffee, and D. Laheru. 2011. A lethally irradiated

allogeneic granulocyte-macrophage colony stimulating factor-secreting tumor vaccine for

pancreatic adenocarcinoma. A Phase II trial of safety, efficacy, and immune activation. Ann

Surg 253: 328-335.

157. Laheru, D., E. Lutz, J. Burke, B. Biedrzycki, S. Solt, B. Onners, I. Tartakovsky, J.

Nemunaitis, D. Le, E. Sugar, K. Hege, and E. Jaffee. 2008. Allogeneic granulocyte

macrophage colony-stimulating factor-secreting tumor immunotherapy alone or in sequence

with cyclophosphamide for metastatic pancreatic cancer: a pilot study of safety, feasibility,

and immune activation. Clinical cancer research : an official journal of the American

Association for Cancer Research 14: 1455-1463.

158. Le, D. T., A. Wang-Gillam, V. Picozzi, T. F. Greten, T. Crocenzi, G. Springett, M.

Morse, H. Zeh, D. Cohen, R. L. Fine, B. Onners, J. N. Uram, D. A. Laheru, E. R. Lutz, S.

Solt, A. L. Murphy, J. Skoble, E. Lemmens, J. Grous, T. Dubensky, Jr., D. G. Brockstedt,

and E. M. Jaffee. 2015. Safety and survival with GVAX pancreas prime and Listeria

Monocytogenes-expressing mesothelin (CRS-207) boost vaccines for metastatic pancreatic

cancer. J Clin Oncol 33: 1325-1333.

159. Middleton, G., P. Silcocks, T. Cox, J. Valle, J. Wadsley, D. Propper, F. Coxon, P.

Ross, S. Madhusudan, T. Roques, D. Cunningham, S. Falk, N. Wadd, M. Harrison, P.

Corrie, T. Iveson, A. Robinson, K. McAdam, M. Eatock, J. Evans, C. Archer, T. Hickish, A.

Garcia-Alonso, M. Nicolson, W. Steward, A. Anthoney, W. Greenhalf, V. Shaw, E. Costello,

168

D. Naisbitt, C. Rawcliffe, G. Nanson, and J. Neoptolemos. 2014. Gemcitabine and capecitabine with or without telomerase peptide vaccine GV1001 in patients with locally advanced or metastatic pancreatic cancer (TeloVac): an open-label, randomised, phase 3 trial. Lancet Oncol 15: 829-840.

160. Yamaue, H., T. Tsunoda, M. Tani, M. Miyazawa, K. Yamao, N. Mizuno, T. Okusaka,

H. Ueno, N. Boku, A. Fukutomi, H. Ishii, S. Ohkawa, M. Furukawa, H. Maguchi, M. Ikeda, Y.

Togashi, K. Nishio, and Y. Ohashi. 2015. Randomized phase II/III clinical trial of elpamotide for patients with advanced pancreatic cancer: PEGASUS-PC Study. Cancer science 106:

883-890.

161. Hardacre, J. M., M. Mulcahy, W. Small, M. Talamonti, J. Obel, S. Krishnamurthi, C.

S. Rocha-Lima, H. Safran, H. J. Lenz, and E. G. Chiorean. 2013. Addition of algenpantucel-

L immunotherapy to standard adjuvant therapy for pancreatic cancer: a phase 2 study.

Journal of gastrointestinal surgery : official journal of the Society for Surgery of the

Alimentary Tract 17: 94-100; discussion p. 100-101.

162. Royal, R. E., C. Levy, K. Turner, A. Mathur, M. Hughes, U. S. Kammula, R. M.

Sherry, S. L. Topalian, J. C. Yang, I. Lowy, and S. A. Rosenberg. 2010. Phase 2 trial of single agent Ipilimumab (anti-CTLA-4) for locally advanced or metastatic pancreatic adenocarcinoma. J Immunother 33: 828-833.

163. Brahmer, J. R., S. S. Tykodi, L. Q. Chow, W. J. Hwu, S. L. Topalian, P. Hwu, C. G.

Drake, L. H. Camacho, J. Kauh, K. Odunsi, H. C. Pitot, O. Hamid, S. Bhatia, R. Martins, K.

Eaton, S. Chen, T. M. Salay, S. Alaparthy, J. F. Grosso, A. J. Korman, S. M. Parker, S.

Agrawal, S. M. Goldberg, D. M. Pardoll, A. Gupta, and J. M. Wigginton. 2012. Safety and activity of anti-PD-L1 antibody in patients with advanced cancer. The New England journal of medicine 366: 2455-2465.

164. Aglietta, M., C. Barone, M. B. Sawyer, M. J. Moore, W. H. Miller, Jr., C. Bagala, F.

169

Colombi, C. Cagnazzo, L. Gioeni, E. Wang, B. Huang, K. D. Fly, and F. Leone. 2014. A phase I dose escalation trial of tremelimumab (CP-675,206) in combination with gemcitabine in chemotherapy-naive patients with metastatic pancreatic cancer. Ann Oncol 25: 1750-

1755.

165. Kalyan, A., S. M. Kircher, N. A. Mohindra, H. S. Nimeiri, V. Maurer, A. Rademaker,

A. B. Benson, and M. F. Mulcahy. 2016. Ipilimumab and gemcitabine for advanced pancreas cancer: A phase Ib study. Journal of Clinical Oncology 34: e15747-e15747.

166. Wainberg, Z. A., H. S. Hochster, B. George, M. Gutierrez, M. E. Johns, E. G.

Chiorean, E. L. Kwak, A. Kalyan, V. Manax, M. Ye, T. Chen, N. Trunova, and P. J. O'Dwyer.

2017. Phase I study of nivolumab (nivo) + nab-paclitaxel (nab-P) ± gemcitabine (Gem) in solid tumors: Interim results from the pancreatic cancer (PC) cohorts. Journal of Clinical

Oncology 35: 412-412.

167. Provenzano, P. P., C. Cuevas, A. E. Chang, V. K. Goel, D. D. Von Hoff, and S. R.

Hingorani. 2012. Enzymatic targeting of the stroma ablates physical barriers to treatment of pancreatic ductal adenocarcinoma. Cancer Cell 21: 418-429.

168. Poznansky, M. C., I. T. Olszak, R. Foxall, R. H. Evans, A. D. Luster, and D. T.

Scadden. 2000. Active movement of T cells away from a chemokine. Nat Med 6: 543-548.

169. Zhao, X. K., Y. Cheng, M. Liang Cheng, L. Yu, M. Mu, H. Li, Y. Liu, B. Zhang, Y.

Yao, H. Guo, R. Wang, and Q. Zhang. 2016. Focal Adhesion Kinase Regulates Fibroblast

Migration via Integrin beta-1 and Plays a Central Role in Fibrosis. Sci Rep 6: 19276.

170. Jiang, H., S. Hegde, B. L. Knolhoff, Y. Zhu, J. M. Herndon, M. A. Meyer, T. M.

Nywening, W. G. Hawkins, I. M. Shapiro, D. T. Weaver, J. A. Pachter, A. Wang-Gillam, and

D. G. DeNardo. 2016. Targeting focal adhesion kinase renders pancreatic cancers responsive to checkpoint immunotherapy. Nature Medicine 22: 851.

171. Akashi, K., D. Traver, T. Miyamoto, and I. L. Weissman. 2000. A clonogenic common

170

myeloid progenitor that gives rise to all myeloid lineages. Nature 404: 193-197.

172. Schultze, J. L., A. Schmieder, and S. Goerdt. 2015. Macrophage activation in human diseases. Semin Immunol 27: 249-256.

173. Hanahan, D., and R. A. Weinberg. 2011. Hallmarks of cancer: the next generation.

Cell 144: 646-674.

174. Ugel, S., F. De Sanctis, S. Mandruzzato, and V. Bronte. 2015. Tumor-induced myeloid deviation: when myeloid-derived suppressor cells meet tumor-associated macrophages. J Clin Invest 125: 3365-3376.

175. Bennett, J. A., V. S. Rao, and M. S. Mitchell. 1978. Systemic bacillus Calmette-

Guerin (BCG) activates natural suppressor cells. Proc Natl Acad Sci U S A 75: 5142-5144.

176. Hestdal, K., F. W. Ruscetti, J. N. Ihle, S. E. Jacobsen, C. M. Dubois, W. C. Kopp, D.

L. Longo, and J. R. Keller. 1991. Characterization and regulation of RB6-8C5 antigen expression on murine bone marrow cells. J Immunol 147: 22-28.

177. Fleming, T. J., M. L. Fleming, and T. R. Malek. 1993. Selective expression of Ly-6G on myeloid lineage cells in mouse bone marrow. RB6-8C5 mAb to granulocyte- differentiation antigen (Gr-1) detects members of the Ly-6 family. J Immunol 151: 2399-

2408.

178. Pak, A. S., M. A. Wright, J. P. Matthews, S. L. Collins, G. J. Petruzzelli, and M. R.

Young. 1995. Mechanisms of immune suppression in patients with head and neck cancer: presence of CD34(+) cells which suppress immune functions within cancers that secrete granulocyte-macrophage colony-stimulating factor. Clinical cancer research : an official journal of the American Association for Cancer Research 1: 95-103.

179. Bronte, V., E. Apolloni, A. Cabrelle, R. Ronca, P. Serafini, P. Zamboni, N. P. Restifo, and P. Zanovello. 2000. Identification of a CD11b(+)/Gr-1(+)/CD31(+) myeloid progenitor capable of activating or suppressing CD8(+) T cells. Blood 96: 3838-3846.

171

180. Bronte, V., D. B. Chappell, E. Apolloni, A. Cabrelle, M. Wang, P. Hwu, and N. P.

Restifo. 1999. Unopposed production of granulocyte-macrophage colony-stimulating factor by tumors inhibits CD8+ T cell responses by dysregulating antigen-presenting cell maturation. J Immunol 162: 5728-5737.

181. Kusmartsev, S. A., Y. Li, and S. H. Chen. 2000. Gr-1+ myeloid cells derived from tumor-bearing mice inhibit primary T cell activation induced through CD3/CD28 costimulation. J Immunol 165: 779-785.

182. Serafini, P., R. Carbley, K. A. Noonan, G. Tan, V. Bronte, and I. Borrello. 2004. High- dose granulocyte-macrophage colony-stimulating factor-producing vaccines impair the immune response through the recruitment of myeloid suppressor cells. Cancer Res 64:

6337-6343.

183. Gabrilovich, D. I., V. Bronte, S. H. Chen, M. P. Colombo, A. Ochoa, S. Ostrand-

Rosenberg, and H. Schreiber. 2007. The terminology issue for myeloid-derived suppressor cells. Cancer Res 67: 425; author reply 426.

184. Youn, J. I., S. Nagaraj, M. Collazo, and D. I. Gabrilovich. 2008. Subsets of myeloid- derived suppressor cells in tumor-bearing mice. J Immunol 181: 5791-5802.

185. Kowanetz, M., X. Wu, J. Lee, M. Tan, T. Hagenbeek, X. Qu, L. Yu, J. Ross, N.

Korsisaari, T. Cao, H. Bou-Reslan, D. Kallop, R. Weimer, M. J. Ludlam, J. S. Kaminker, Z.

Modrusan, N. van Bruggen, F. V. Peale, R. Carano, Y. G. Meng, and N. Ferrara. 2010.

Granulocyte-colony stimulating factor promotes lung metastasis through mobilization of

Ly6G+Ly6C+ granulocytes. Proc Natl Acad Sci U S A 107: 21248-21255.

186. Menetrier-Caux, C., G. Montmain, M. C. Dieu, C. Bain, M. C. Favrot, C. Caux, and J.

Y. Blay. 1998. Inhibition of the differentiation of dendritic cells from CD34(+) progenitors by tumor cells: role of interleukin-6 and macrophage colony-stimulating factor. Blood 92: 4778-

4791.

172

187. Condamine, T., and D. I. Gabrilovich. 2011. Molecular mechanisms regulating

myeloid-derived suppressor cell differentiation and function. Trends Immunol 32: 19-25.

188. Lechner, M. G., D. J. Liebertz, and A. L. Epstein. 2010. Characterization of cytokine-

induced myeloid-derived suppressor cells from normal human peripheral blood mononuclear

cells. J Immunol 185: 2273-2284.

189. Oh, K., O. Y. Lee, S. Y. Shon, O. Nam, P. M. Ryu, M. W. Seo, and D. S. Lee. 2013.

A mutual activation loop between breast cancer cells and myeloid-derived suppressor cells facilitates spontaneous metastasis through IL-6 trans-signaling in a murine model. Breast

Cancer Res 15: R79.

190. Wu, C. T., C. C. Hsieh, C. C. Lin, W. C. Chen, J. H. Hong, and M. F. Chen. 2012.

Significance of IL-6 in the transition of hormone-resistant prostate cancer and the induction of myeloid-derived suppressor cells. J Mol Med (Berl) 90: 1343-1355.

191. Xin, H., C. Zhang, A. Herrmann, Y. Du, R. Figlin, and H. Yu. 2009. Sunitinib inhibition

of Stat3 induces renal cell carcinoma tumor cell apoptosis and reduces immunosuppressive

cells. Cancer Res 69: 2506-2513.

192. Abad, C., H. Nobuta, J. Li, A. Kasai, W. H. Yong, and J. A. Waschek. 2014. Targeted

STAT3 disruption in myeloid cells alters immunosuppressor cell abundance in a murine

model of spontaneous medulloblastoma. J Leukoc Biol 95: 357-367.

193. Zhou, J., Z. Qu, F. Sun, L. Han, L. Li, S. Yan, L. P. Stabile, L. F. Chen, J. M.

Siegfried, and G. Xiao. 2017. Myeloid STAT3 Promotes Lung Tumorigenesis by

Transforming Tumor Immunosurveillance into Tumor-Promoting Inflammation. Cancer

Immunol Res 5: 257-268.

194. Yu, H., Y. Liu, B. C. McFarland, J. S. Deshane, D. R. Hurst, S. Ponnazhagan, E. N.

Benveniste, and H. Qin. 2015. SOCS3 Deficiency in Myeloid Cells Promotes Tumor

Development: Involvement of STAT3 Activation and Myeloid-Derived Suppressor Cells.

173

Cancer Immunol Res 3: 727-740.

195. Nefedova, Y., M. Huang, S. Kusmartsev, R. Bhattacharya, P. Cheng, R. Salup, R.

Jove, and D. Gabrilovich. 2004. Hyperactivation of STAT3 is involved in abnormal

differentiation of dendritic cells in cancer. J Immunol 172: 464-474.

196. Cheng, P., C. A. Corzo, N. Luetteke, B. Yu, S. Nagaraj, M. M. Bui, M. Ortiz, W.

Nacken, C. Sorg, T. Vogl, J. Roth, and D. I. Gabrilovich. 2008. Inhibition of dendritic cell

differentiation and accumulation of myeloid-derived suppressor cells in cancer is regulated

by S100A9 protein. J Exp Med 205: 2235-2249.

197. Corzo, C. A., M. J. Cotter, P. Cheng, F. Cheng, S. Kusmartsev, E. Sotomayor, T.

Padhya, T. V. McCaffrey, J. C. McCaffrey, and D. I. Gabrilovich. 2009. Mechanism

regulating reactive oxygen species in tumor-induced myeloid-derived suppressor cells. J

Immunol 182: 5693-5701.

198. Vasquez-Dunddel, D., F. Pan, Q. Zeng, M. Gorbounov, E. Albesiano, J. Fu, R. L.

Blosser, A. J. Tam, T. Bruno, H. Zhang, D. Pardoll, and Y. Kim. 2013. STAT3 regulates

arginase-I in myeloid-derived suppressor cells from cancer patients. The Journal of Clinical

Investigation 123: 1580-1589.

199. Marigo, I., E. Bosio, S. Solito, C. Mesa, A. Fernandez, L. Dolcetti, S. Ugel, N. Sonda,

S. Bicciato, E. Falisi, F. Calabrese, G. Basso, P. Zanovello, E. Cozzi, S. Mandruzzato, and

V. Bronte. 2010. Tumor-induced tolerance and immune suppression depend on the

C/EBPbeta transcription factor. Immunity 32: 790-802.

200. Stewart, T. J., K. M. Greeneltch, J. E. Reid, D. J. Liewehr, S. M. Steinberg, K. Liu, and S. I. Abrams. 2009. Interferon regulatory factor-8 modulates the development of tumour- induced CD11b+Gr-1+ myeloid cells. J Cell Mol Med 13: 3939-3950.

201. Waight, J. D., C. Netherby, M. L. Hensen, A. Miller, Q. Hu, S. Liu, P. N. Bogner, M.

R. Farren, K. P. Lee, K. Liu, and S. I. Abrams. 2013. Myeloid-derived suppressor cell

174

development is regulated by a STAT/IRF-8 axis. J Clin Invest 123: 4464-4478.

202. Cheng, P., V. Kumar, H. Liu, J. I. Youn, M. Fishman, S. Sherman, and D.

Gabrilovich. 2014. Effects of notch signaling on regulation of myeloid cell differentiation in cancer. Cancer Res 74: 141-152.

203. Casbon, A. J., D. Reynaud, C. Park, E. Khuc, D. D. Gan, K. Schepers, E. Passegue, and Z. Werb. 2015. Invasive breast cancer reprograms early myeloid differentiation in the bone marrow to generate immunosuppressive neutrophils. Proc Natl Acad Sci U S A 112:

E566-575.

204. Youn, J. I., V. Kumar, M. Collazo, Y. Nefedova, T. Condamine, P. Cheng, A. Villagra,

S. Antonia, J. C. McCaffrey, M. Fishman, A. Sarnaik, P. Horna, E. Sotomayor, and D. I.

Gabrilovich. 2013. Epigenetic silencing of retinoblastoma gene regulates pathologic differentiation of myeloid cells in cancer. Nature immunology 14: 211-220.

205. Iannone, R., L. Miele, P. Maiolino, A. Pinto, and S. Morello. 2013. Blockade of A2b adenosine receptor reduces tumor growth and immune suppression mediated by myeloid- derived suppressor cells in a mouse model of melanoma. Neoplasia 15: 1400-1409.

206. Ryzhov, S., S. V. Novitskiy, A. E. Goldstein, A. Biktasova, M. R. Blackburn, I.

Biaggioni, M. M. Dikov, and I. Feoktistov. 2011. Adenosinergic regulation of the expansion

and immunosuppressive activity of CD11b+Gr1+ cells. J Immunol 187: 6120-6129.

207. Bronte, V., P. Serafini, C. De Santo, I. Marigo, V. Tosello, A. Mazzoni, D. M. Segal,

C. Staib, M. Lowel, G. Sutter, M. P. Colombo, and P. Zanovello. 2003. IL-4-induced arginase 1 suppresses alloreactive T cells in tumor-bearing mice. J Immunol 170: 270-278.

208. Highfill, S. L., P. C. Rodriguez, Q. Zhou, C. A. Goetz, B. H. Koehn, R. Veenstra, P. A.

Taylor, A. Panoskaltsis-Mortari, J. S. Serody, D. H. Munn, J. Tolar, A. C. Ochoa, and B. R.

Blazar. 2010. Bone marrow myeloid-derived suppressor cells (MDSCs) inhibit graft-versus-

host disease (GVHD) via an arginase-1-dependent mechanism that is up-regulated by

175

interleukin-13. Blood 116: 5738-5747.

209. Bunt, S. K., V. K. Clements, E. M. Hanson, P. Sinha, and S. Ostrand-Rosenberg.

2009. Inflammation enhances myeloid-derived suppressor cell cross-talk by signaling through Toll-like receptor 4. J Leukoc Biol 85: 996-1004.

210. Zhao, X., L. Rong, X. Zhao, X. Li, X. Liu, J. Deng, H. Wu, X. Xu, U. Erben, P. Wu, U.

Syrbe, J. Sieper, and Z. Qin. 2012. TNF signaling drives myeloid-derived suppressor cell accumulation. J Clin Invest 122: 4094-4104.

211. Rodriguez, P. C., C. P. Hernandez, D. Quiceno, S. M. Dubinett, J. Zabaleta, J. B.

Ochoa, J. Gilbert, and A. C. Ochoa. 2005. Arginase I in myeloid suppressor cells is induced by COX-2 in lung carcinoma. J Exp Med 202: 931-939.

212. Sinha, P., V. K. Clements, A. M. Fulton, and S. Ostrand-Rosenberg. 2007.

Prostaglandin E2 promotes tumor progression by inducing myeloid-derived suppressor cells.

Cancer Res 67: 4507-4513.

213. Thevenot, P. T., R. A. Sierra, P. L. Raber, A. A. Al-Khami, J. Trillo-Tinoco, P. Zarreii,

A. C. Ochoa, Y. Cui, L. Del Valle, and P. C. Rodriguez. 2014. The stress-response sensor chop regulates the function and accumulation of myeloid-derived suppressor cells in tumors.

Immunity 41: 389-401.

214. E, M. 1883. Untersuchungen ueber die mesodermalen Phagocyten einiger

Wirbeltiere. Biologisches Centralblatt: 560-565.

215. van Furth, R., and Z. A. Cohn. 1968. The origin and kinetics of mononuclear phagocytes. J Exp Med 128: 415-435.

216. Geissmann, F., S. Jung, and D. R. Littman. 2003. Blood monocytes consist of two principal subsets with distinct migratory properties. Immunity 19: 71-82.

217. Hanna, R. N., C. Cekic, D. Sag, R. Tacke, G. D. Thomas, H. Nowyhed, E. Herrley, N.

Rasquinha, S. McArdle, R. Wu, E. Peluso, D. Metzger, H. Ichinose, I. Shaked, G.

176

Chodaczek, S. K. Biswas, and C. C. Hedrick. 2015. Patrolling monocytes control tumor metastasis to the lung. Science 350: 985-990.

218. Ginhoux, F., M. Greter, M. Leboeuf, S. Nandi, P. See, S. Gokhan, M. F. Mehler, S. J.

Conway, L. G. Ng, E. R. Stanley, I. M. Samokhvalov, and M. Merad. 2010. Fate mapping analysis reveals that adult microglia derive from primitive macrophages. Science 330: 841-

845.

219. Schulz, C., E. Gomez Perdiguero, L. Chorro, H. Szabo-Rogers, N. Cagnard, K.

Kierdorf, M. Prinz, B. Wu, S. E. Jacobsen, J. W. Pollard, J. Frampton, K. J. Liu, and F.

Geissmann. 2012. A lineage of myeloid cells independent of Myb and hematopoietic stem cells. Science 336: 86-90.

220. Yona, S., K. W. Kim, Y. Wolf, A. Mildner, D. Varol, M. Breker, D. Strauss-Ayali, S.

Viukov, M. Guilliams, A. Misharin, D. A. Hume, H. Perlman, B. Malissen, E. Zelzer, and S.

Jung. 2013. Fate mapping reveals origins and dynamics of monocytes and tissue macrophages under homeostasis. Immunity 38: 79-91.

221. Calderon, B., J. A. Carrero, S. T. Ferris, D. K. Sojka, L. Moore, S. Epelman, K. M.

Murphy, W. M. Yokoyama, G. J. Randolph, and E. R. Unanue. 2015. The pancreas anatomy conditions the origin and properties of resident macrophages. J Exp Med 212: 1497-1512.

222. Davies, L. C., M. Rosas, S. J. Jenkins, C. T. Liao, M. J. Scurr, F. Brombacher, D. J.

Fraser, J. E. Allen, S. A. Jones, and P. R. Taylor. 2013. Distinct bone marrow-derived and tissue-resident macrophage lineages proliferate at key stages during inflammation. Nat

Commun 4: 1886.

223. Hashimoto, D., A. Chow, C. Noizat, P. Teo, M. B. Beasley, M. Leboeuf, C. D. Becker,

P. See, J. Price, D. Lucas, M. Greter, A. Mortha, S. W. Boyer, E. C. Forsberg, M. Tanaka, N. van Rooijen, A. Garcia-Sastre, E. R. Stanley, F. Ginhoux, P. S. Frenette, and M. Merad.

2013. Tissue-resident macrophages self-maintain locally throughout adult life with minimal

177

contribution from circulating monocytes. Immunity 38: 792-804.

224. Zhu, Y., J. M. Herndon, D. K. Sojka, K. W. Kim, B. L. Knolhoff, C. Zuo, D. R.

Cullinan, J. Luo, A. R. Bearden, K. J. Lavine, W. M. Yokoyama, W. G. Hawkins, R. C. Fields,

G. J. Randolph, and D. G. DeNardo. 2017. Tissue-Resident Macrophages in Pancreatic

Ductal Adenocarcinoma Originate from Embryonic Hematopoiesis and Promote Tumor

Progression. Immunity 47: 597.

225. Kawasaki, E. S., M. B. Ladner, A. M. Wang, J. Van Arsdell, M. K. Warren, M. Y.

Coyne, V. L. Schweickart, M. T. Lee, K. J. Wilson, A. Boosman, and et al. 1985. Molecular cloning of a complementary DNA encoding human macrophage-specific colony-stimulating factor (CSF-1). Science 230: 291-296.

226. Lin, H., E. Lee, K. Hestir, C. Leo, M. Huang, E. Bosch, R. Halenbeck, G. Wu, A.

Zhou, D. Behrens, D. Hollenbaugh, T. Linnemann, M. Qin, J. Wong, K. Chu, S. K.

Doberstein, and L. T. Williams. 2008. Discovery of a cytokine and its receptor by functional screening of the extracellular proteome. Science 320: 807-811.

227. Gyori, D., E. L. Lim, F. M. Grant, D. Spensberger, R. Roychoudhuri, S. J.

Shuttleworth, K. Okkenhaug, L. R. Stephens, and P. T. Hawkins. 2018. Compensation

between CSF1R+ macrophages and Foxp3+ Treg cells drives resistance to tumor

immunotherapy. JCI Insight 3.

228. Lin, E. Y., A. V. Nguyen, R. G. Russell, and J. W. Pollard. 2001. Colony-stimulating

factor 1 promotes progression of mammary tumors to malignancy. J Exp Med 193: 727-740.

229. Ries, C. H., M. A. Cannarile, S. Hoves, J. Benz, K. Wartha, V. Runza, F. Rey-Giraud,

L. P. Pradel, F. Feuerhake, I. Klaman, T. Jones, U. Jucknischke, S. Scheiblich, K. Kaluza, I.

H. Gorr, A. Walz, K. Abiraj, P. A. Cassier, A. Sica, C. Gomez-Roca, K. E. de Visser, A.

Italiano, C. Le Tourneau, J. P. Delord, H. Levitsky, J. Y. Blay, and D. Ruttinger. 2014.

Targeting tumor-associated macrophages with anti-CSF-1R antibody reveals a strategy for

178

cancer therapy. Cancer Cell 25: 846-859.

230. Qian, B. Z., J. Li, H. Zhang, T. Kitamura, J. Zhang, L. R. Campion, E. A. Kaiser, L. A.

Snyder, and J. W. Pollard. 2011. CCL2 recruits inflammatory monocytes to facilitate breast- tumour metastasis. Nature 475: 222-225.

231. Avraham-Davidi, I., S. Yona, M. Grunewald, L. Landsman, C. Cochain, J. S.

Silvestre, H. Mizrahi, M. Faroja, D. Strauss-Ayali, M. Mack, S. Jung, and E. Keshet. 2013.

On-site education of VEGF-recruited monocytes improves their performance as angiogenic and arteriogenic accessory cells. The Journal of Experimental Medicine 210: 2611-2625.

232. Kumar, V., P. Cheng, T. Condamine, S. Mony, L. R. Languino, J. C. McCaffrey, N.

Hockstein, M. Guarino, G. Masters, E. Penman, F. Denstman, X. Xu, D. C. Altieri, H. Du, C.

Yan, and D. I. Gabrilovich. 2016. CD45 Phosphatase Inhibits STAT3 Transcription Factor

Activity in Myeloid Cells and Promotes Tumor-Associated Macrophage Differentiation.

Immunity 44: 303-315.

233. Stout, R. D., C. Jiang, B. Matta, I. Tietzel, S. K. Watkins, and J. Suttles. 2005.

Macrophages sequentially change their functional phenotype in response to changes in microenvironmental influences. J Immunol 175: 342-349.

234. Murray, P. J., J. E. Allen, S. K. Biswas, E. A. Fisher, D. W. Gilroy, S. Goerdt, S.

Gordon, J. A. Hamilton, L. B. Ivashkiv, T. Lawrence, M. Locati, A. Mantovani, F. O. Martinez,

J. L. Mege, D. M. Mosser, G. Natoli, J. P. Saeij, J. L. Schultze, K. A. Shirey, A. Sica, J.

Suttles, I. Udalova, J. A. van Ginderachter, S. N. Vogel, and T. A. Wynn. 2014. Macrophage activation and polarization: nomenclature and experimental guidelines. Immunity 41: 14-20.

235. Mills, C. D., K. Kincaid, J. M. Alt, M. J. Heilman, and A. M. Hill. 2000. M-1/M-2 macrophages and the Th1/Th2 paradigm. J Immunol 164: 6166-6173.

236. Mokoena, T., and S. Gordon. 1985. Human macrophage activation. Modulation of mannosyl, fucosyl receptor activity in vitro by lymphokines, gamma and alpha interferons,

179

and dexamethasone. J Clin Invest 75: 624-631.

237. Stein, M., S. Keshav, N. Harris, and S. Gordon. 1992. Interleukin 4 potently enhances murine macrophage mannose receptor activity: a marker of alternative immunologic macrophage activation. J Exp Med 176: 287-292.

238. Edwards, J. P., X. Zhang, K. A. Frauwirth, and D. M. Mosser. 2006. Biochemical and functional characterization of three activated macrophage populations. J Leukoc Biol 80:

1298-1307.

239. Azizi, E., A. J. Carr, G. Plitas, A. E. Cornish, C. Konopacki, S. Prabhakaran, J.

Nainys, K. Wu, V. Kiseliovas, M. Setty, K. Choi, R. M. Fromme, P. Dao, P. T. McKenney, R.

C. Wasti, K. Kadaveru, L. Mazutis, A. Y. Rudensky, and D. Pe'er. 2018. Single-Cell Map of

Diverse Immune Phenotypes in the Breast Tumor Microenvironment. Cell 174: 1293-1308 e1236.

240. Chevrier, S., J. H. Levine, V. R. T. Zanotelli, K. Silina, D. Schulz, M. Bacac, C. H.

Ries, L. Ailles, M. A. S. Jewett, H. Moch, M. van den Broek, C. Beisel, M. B. Stadler, C.

Gedye, B. Reis, D. Pe'er, and B. Bodenmiller. 2017. An Immune Atlas of Clear Cell Renal

Cell Carcinoma. Cell 169: 736-749 e718.

241. Lavin, Y., S. Kobayashi, A. Leader, E. D. Amir, N. Elefant, C. Bigenwald, R. Remark,

R. Sweeney, C. D. Becker, J. H. Levine, K. Meinhof, A. Chow, S. Kim-Shulze, A. Wolf, C.

Medaglia, H. Li, J. A. Rytlewski, R. O. Emerson, A. Solovyov, B. D. Greenbaum, C. Sanders,

M. Vignali, M. B. Beasley, R. Flores, S. Gnjatic, D. Pe'er, A. Rahman, I. Amit, and M. Merad.

2017. Innate Immune Landscape in Early Lung Adenocarcinoma by Paired Single-Cell

Analyses. Cell 169: 750-765 e717.

242. Murray, P. J. 2016. Amino acid auxotrophy as a system of immunological control nodes. Nature immunology 17: 132-139.

243. Srivastava, M. K., P. Sinha, V. K. Clements, P. Rodriguez, and S. Ostrand-

180

Rosenberg. 2010. Myeloid-derived suppressor cells inhibit T-cell activation by depleting cystine and cysteine. Cancer Res 70: 68-77.

244. Droge, W., D. Mannel, W. Falk, V. Lehmann, H. Schmidt, S. Nick, B. Hacker-Shahin, and R. Janicke. 1985. Suppression of cytotoxic T lymphocyte activation by L-ornithine. J

Immunol 134: 3379-3383.

245. Janicke, R., and W. Droge. 1985. Effect of L-ornithine on proliferative and cytotoxic

T-cell responses in allogeneic and syngeneic mixed leukocyte cultures. Cell Immunol 92:

359-365.

246. Rodriguez, P. C., A. H. Zea, K. S. Culotta, J. Zabaleta, J. B. Ochoa, and A. C.

Ochoa. 2002. Regulation of T cell receptor CD3zeta chain expression by L-arginine. J Biol

Chem 277: 21123-21129.

247. Rodriguez, P. C., A. H. Zea, J. DeSalvo, K. S. Culotta, J. Zabaleta, D. G. Quiceno, J.

B. Ochoa, and A. C. Ochoa. 2003. L-arginine consumption by macrophages modulates the expression of CD3 zeta chain in T lymphocytes. J Immunol 171: 1232-1239.

248. Zea, A. H., P. C. Rodriguez, K. S. Culotta, C. P. Hernandez, J. DeSalvo, J. B.

Ochoa, H. J. Park, J. Zabaleta, and A. C. Ochoa. 2004. L-Arginine modulates CD3zeta expression and T cell function in activated human T lymphocytes. Cell Immunol 232: 21-31.

249. Rodriguez, P. C., C. P. Hernandez, K. Morrow, R. Sierra, J. Zabaleta, D. D.

Wyczechowska, and A. C. Ochoa. 2010. L-arginine deprivation regulates cyclin D3 mRNA stability in human T cells by controlling HuR expression. J Immunol 185: 5198-5204.

250. Rodriguez, P. C., D. G. Quiceno, and A. C. Ochoa. 2007. L-arginine availability regulates T-lymphocyte cell-cycle progression. Blood 109: 1568-1573.

251. Fletcher, M., M. E. Ramirez, R. A. Sierra, P. Raber, P. Thevenot, A. A. Al-Khami, D.

Sanchez-Pino, C. Hernandez, D. D. Wyczechowska, A. C. Ochoa, and P. C. Rodriguez.

2015. l-Arginine depletion blunts antitumor T-cell responses by inducing myeloid-derived

181

suppressor cells. Cancer Res 75: 275-283.

252. Gray, M. J., M. Poljakovic, D. Kepka-Lenhart, and S. M. Morris, Jr. 2005. Induction of arginase I transcription by IL-4 requires a composite DNA response element for STAT6 and

C/EBPbeta. Gene 353: 98-106.

253. Pauleau, A. L., R. Rutschman, R. Lang, A. Pernis, S. S. Watowich, and P. J. Murray.

2004. Enhancer-mediated control of macrophage-specific arginase I expression. J Immunol

172: 7565-7573.

254. Munder, M., K. Eichmann, and M. Modolell. 1998. Alternative metabolic states in murine macrophages reflected by the nitric oxide synthase/arginase balance: competitive regulation by CD4+ T cells correlates with Th1/Th2 phenotype. J Immunol 160: 5347-5354.

255. Prendergast, G. C., W. P. Malachowski, J. B. DuHadaway, and A. J. Muller. 2017.

Discovery of IDO1 Inhibitors: From Bench to Bedside. Cancer Res 77: 6795-6811.

256. Opitz, C. A., U. M. Litzenburger, F. Sahm, M. Ott, I. Tritschler, S. Trump, T.

Schumacher, L. Jestaedt, D. Schrenk, M. Weller, M. Jugold, G. J. Guillemin, C. L. Miller, C.

Lutz, B. Radlwimmer, I. Lehmann, A. von Deimling, W. Wick, and M. Platten. 2011. An endogenous tumour-promoting ligand of the human aryl hydrocarbon receptor. Nature 478:

197-203.

257. Yoshida, R., J. Imanishi, T. Oku, T. Kishida, and O. Hayaishi. 1981. Induction of pulmonary indoleamine 2,3-dioxygenase by interferon. Proc Natl Acad Sci U S A 78: 129-

132.

258. Yu, J., W. Du, F. Yan, Y. Wang, H. Li, S. Cao, W. Yu, C. Shen, J. Liu, and X. Ren.

2013. Myeloid-derived suppressor cells suppress antitumor immune responses through IDO expression and correlate with lymph node metastasis in patients with breast cancer. J

Immunol 190: 3783-3797.

259. Zhao, Q., D. M. Kuang, Y. Wu, X. Xiao, X. F. Li, T. J. Li, and L. Zheng. 2012.

182

Activated CD69+ T cells foster immune privilege by regulating IDO expression in tumor-

associated macrophages. J Immunol 188: 1117-1124.

260. Holmgaard, R. B., D. Zamarin, Y. Li, B. Gasmi, D. H. Munn, J. P. Allison, T.

Merghoub, and J. D. Wolchok. 2015. Tumor-Expressed IDO Recruits and Activates MDSCs in a Treg-Dependent Manner. Cell Rep 13: 412-424.

261. Munn, D. H., E. Shafizadeh, J. T. Attwood, I. Bondarev, A. Pashine, and A. L. Mellor.

1999. Inhibition of T cell proliferation by macrophage tryptophan catabolism. J Exp Med 189:

1363-1372.

262. Munn, D. H., M. D. Sharma, B. Baban, H. P. Harding, Y. Zhang, D. Ron, and A. L.

Mellor. 2005. GCN2 kinase in T cells mediates proliferative arrest and anergy induction in

response to indoleamine 2,3-dioxygenase. Immunity 22: 633-642.

263. Bessede, A., M. Gargaro, M. T. Pallotta, D. Matino, G. Servillo, C. Brunacci, S.

Bicciato, E. M. Mazza, A. Macchiarulo, C. Vacca, R. Iannitti, L. Tissi, C. Volpi, M. L.

Belladonna, C. Orabona, R. Bianchi, T. V. Lanz, M. Platten, M. A. Della Fazia, D. Piobbico,

T. Zelante, H. Funakoshi, T. Nakamura, D. Gilot, M. S. Denison, G. J. Guillemin, J. B.

DuHadaway, G. C. Prendergast, R. Metz, M. Geffard, L. Boon, M. Pirro, A. Iorio, B. Veyret,

L. Romani, U. Grohmann, F. Fallarino, and P. Puccetti. 2014. Aryl hydrocarbon receptor

control of a disease tolerance defence pathway. Nature 511: 184-190.

264. Fallarino, F., U. Grohmann, C. Vacca, C. Orabona, A. Spreca, M. C. Fioretti, and P.

Puccetti. 2003. T cell apoptosis by kynurenines. Adv Exp Med Biol 527: 183-190.

265. Fallarino, F., U. Grohmann, S. You, B. C. McGrath, D. R. Cavener, C. Vacca, C.

Orabona, R. Bianchi, M. L. Belladonna, C. Volpi, P. Santamaria, M. C. Fioretti, and P.

Puccetti. 2006. The combined effects of tryptophan starvation and tryptophan catabolites

down-regulate T cell receptor zeta-chain and induce a regulatory phenotype in naive T cells.

J Immunol 176: 6752-6761.

183

266. Lu, T., and D. I. Gabrilovich. 2012. Molecular Pathways: Tumor-Infiltrating Myeloid

Cells and Reactive Oxygen Species in Regulation of Tumor Microenvironment. Clinical

Cancer Research.

267. Movahedi, K., M. Guilliams, J. Van den Bossche, R. Van den Bergh, C. Gysemans,

A. Beschin, P. De Baetselier, and J. A. Van Ginderachter. 2008. Identification of discrete tumor-induced myeloid-derived suppressor cell subpopulations with distinct T cell- suppressive activity. Blood 111: 4233-4244.

268. Nagaraj, S., J. I. Youn, and D. I. Gabrilovich. 2013. Reciprocal relationship between myeloid-derived suppressor cells and T cells. J Immunol 191: 17-23.

269. Huie, R. E., and S. Padmaja. 1993. The reaction of no with superoxide. Free radical research communications 18: 195-199.

270. Pacher, P., J. S. Beckman, and L. Liaudet. 2007. Nitric oxide and peroxynitrite in health and disease. Physiological reviews 87: 315-424.

271. De Sanctis, F., S. Sandri, G. Ferrarini, I. Pagliarello, S. Sartoris, S. Ugel, I. Marigo, B.

Molon, and V. Bronte. 2014. The Emerging Immunological Role of Post-Translational

Modifications by Reactive Nitrogen Species in Cancer Microenvironment. Frontiers in

Immunology 5.

272. Haqqani, A. S., J. F. Kelly, and H. C. Birnboim. 2002. Selective nitration of histone tyrosine residues in vivo in mutatect tumors. J Biol Chem 277: 3614-3621.

273. Molon, B., S. Ugel, F. Del Pozzo, C. Soldani, S. Zilio, D. Avella, A. De Palma, P.

Mauri, A. Monegal, M. Rescigno, B. Savino, P. Colombo, N. Jonjic, S. Pecanic, L. Lazzarato,

R. Fruttero, A. Gasco, V. Bronte, and A. Viola. 2011. Chemokine nitration prevents intratumoral infiltration of antigen-specific T cells. J Exp Med 208: 1949-1962.

274. Lu, T., R. Ramakrishnan, S. Altiok, J. I. Youn, P. Cheng, E. Celis, V. Pisarev, S.

Sherman, M. B. Sporn, and D. Gabrilovich. 2011. Tumor-infiltrating myeloid cells induce

184

tumor cell resistance to cytotoxic T cells in mice. J Clin Invest 121: 4015-4029.

275. Nagaraj, S., K. Gupta, V. Pisarev, L. Kinarsky, S. Sherman, L. Kang, D. L. Herber, J.

Schneck, and D. I. Gabrilovich. 2007. Altered recognition of antigen is a mechanism of

CD8+ T cell tolerance in cancer. Nat Med 13: 828-835.

276. Nagaraj, S., A. G. Schrum, H. I. Cho, E. Celis, and D. I. Gabrilovich. 2010.

Mechanism of T cell tolerance induced by myeloid-derived suppressor cells. J Immunol 184:

3106-3116.

277. Bronte, V., T. Kasic, G. Gri, K. Gallana, G. Borsellino, I. Marigo, L. Battistini, M.

Iafrate, T. Prayer-Galetti, F. Pagano, and A. Viola. 2005. Boosting antitumor responses of T lymphocytes infiltrating human prostate cancers. J Exp Med 201: 1257-1268.

278. Mazzoni, A., V. Bronte, A. Visintin, J. H. Spitzer, E. Apolloni, P. Serafini, P.

Zanovello, and D. M. Segal. 2002. Myeloid Suppressor Lines Inhibit T Cell Responses by an

NO-Dependent Mechanism. The Journal of Immunology 168: 689-695.

279. Liu, J., N. Zhang, Q. Li, W. Zhang, F. Ke, Q. Leng, H. Wang, J. Chen, and H. Wang.

2011. Tumor-associated macrophages recruit CCR6+ regulatory T cells and promote the development of colorectal cancer via enhancing CCL20 production in mice. PLoS One 6: e19495.

280. Schlecker, E., A. Stojanovic, C. Eisen, C. Quack, C. S. Falk, V. Umansky, and A.

Cerwenka. 2012. Tumor-infiltrating monocytic myeloid-derived suppressor cells mediate

CCR5-dependent recruitment of regulatory T cells favoring tumor growth. J Immunol 189:

5602-5611.

281. Pan, P. Y., G. Ma, K. J. Weber, J. Ozao-Choy, G. Wang, B. Yin, C. M. Divino, and S.

H. Chen. 2010. Immune stimulatory receptor CD40 is required for T-cell suppression and T

regulatory cell activation mediated by myeloid-derived suppressor cells in cancer. Cancer

Res 70: 99-108.

185

282. Serafini, P., S. Mgebroff, K. Noonan, and I. Borrello. 2008. Myeloid-derived

suppressor cells promote cross-tolerance in B-cell lymphoma by expanding regulatory T

cells. Cancer Res 68: 5439-5449.

283. Huang, B., P. Y. Pan, Q. Li, A. I. Sato, D. E. Levy, J. Bromberg, C. M. Divino, and S.

H. Chen. 2006. Gr-1+CD115+ immature myeloid suppressor cells mediate the development

of tumor-induced T regulatory cells and T-cell anergy in tumor-bearing host. Cancer Res 66:

1123-1131.

284. Beury, D. W., K. H. Parker, M. Nyandjo, P. Sinha, K. A. Carter, and S. Ostrand-

Rosenberg. 2014. Cross-talk among myeloid-derived suppressor cells, macrophages, and

tumor cells impacts the inflammatory milieu of solid tumors. J Leukoc Biol 96: 1109-1118.

285. Sinha, P., V. K. Clements, S. K. Bunt, S. M. Albelda, and S. Ostrand-Rosenberg.

2007. Cross-talk between myeloid-derived suppressor cells and macrophages subverts tumor immunity toward a type 2 response. J Immunol 179: 977-983.

286. Thomas, D. A., and J. Massague. 2005. TGF-beta directly targets cytotoxic T cell functions during tumor evasion of immune surveillance. Cancer Cell 8: 369-380.

287. Gorelik, L., and R. A. Flavell. 2001. Immune-mediated eradication of tumors through the blockade of transforming growth factor-beta signaling in T cells. Nat Med 7: 1118-1122.

288. Mariathasan, S., S. J. Turley, D. Nickles, A. Castiglioni, K. Yuen, Y. Wang, E. E.

Kadel, III, H. Koeppen, J. L. Astarita, R. Cubas, S. Jhunjhunwala, R. Banchereau, Y. Yang,

Y. Guan, C. Chalouni, J. Ziai, Y. Senbabaoglu, S. Santoro, D. Sheinson, J. Hung, J. M.

Giltnane, A. A. Pierce, K. Mesh, S. Lianoglou, J. Riegler, R. A. D. Carano, P. Eriksson, M.

Hoglund, L. Somarriba, D. L. Halligan, M. S. van der Heijden, Y. Loriot, J. E. Rosenberg, L.

Fong, I. Mellman, D. S. Chen, M. Green, C. Derleth, G. D. Fine, P. S. Hegde, R. Bourgon, and T. Powles. 2018. TGFbeta attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells. Nature 554: 544-548.

186

289. Akdis, C. A., A. Joss, M. Akdis, A. Faith, and K. Blaser. 2000. A molecular basis for T

cell suppression by IL-10: CD28-associated IL-10 receptor inhibits CD28 tyrosine

phosphorylation and phosphatidylinositol 3-kinase binding. FASEB J 14: 1666-1668.

290. Enk, A. H., J. Saloga, D. Becker, M. Mohamadzadeh, and J. Knop. 1994. Induction

of hapten-specific tolerance by interleukin 10 in vivo. J Exp Med 179: 1397-1402.

291. Oh, S. A., M. Liu, B. G. Nixon, D. Kang, A. Toure, M. Bivona, and M. O. Li. 2017.

Foxp3-independent mechanism by which TGF-β controls peripheral T cell tolerance.

Proceedings of the National Academy of Sciences 114: E7536-E7544.

292. Lu, C., P. S. Redd, J. R. Lee, N. Savage, and K. Liu. 2016. The expression profiles

and regulation of PD-L1 in tumor-induced myeloid-derived suppressor cells.

Oncoimmunology 5: e1247135.

293. Lu, C., N. Savage, N. Singh, and K. Liu. 2017. The expression profiles and regulation of PD-L1 in tumor-induced myeloid-derived suppressor cells. The Journal of Immunology

198: 124.129-124.129.

294. Ballbach, M., A. Dannert, A. Singh, D. M. Siegmund, R. Handgretinger, L. Piali, N.

Rieber, and D. Hartl. 2017. Expression of checkpoint molecules on myeloid-derived suppressor cells. Immunol Lett 192: 1-6.

295. Hartley, G. P., L. Chow, D. T. Ammons, W. H. Wheat, and S. W. Dow. 2018.

Programmed Cell Death Ligand 1 (PD-L1) Signaling Regulates Macrophage Proliferation and Activation. Cancer Immunol Res 6: 1260-1273.

296. Gottlieb, C. E., A. M. Mills, J. V. Cross, and K. L. Ring. 2017. Tumor-associated macrophage expression of PD-L1 in implants of high grade serous ovarian carcinoma: A comparison of matched primary and metastatic tumors. Gynecol Oncol 144: 607-612.

297. Tseng, Y. H., H. L. Ho, C. R. Lai, Y. H. Luo, Y. C. Tseng, J. Whang-Peng, Y. H. Lin,

T. Y. Chou, and Y. M. Chen. 2018. PD-L1 Expression of Tumor Cells, Macrophages, and

187

Immune Cells in Non-Small Cell Lung Cancer Patients with Malignant Pleural Effusion. J

Thorac Oncol 13: 447-453.

298. Petek, B., and R. L. Jones. 2014. The management of ophthalmic involvement in blue rubber bleb nevus syndrome. GMS Ophthalmol Cases 4: Doc04.

299. Noguchi, T., J. P. Ward, M. M. Gubin, C. D. Arthur, S. H. Lee, J. Hundal, M. J. Selby,

R. F. Graziano, E. R. Mardis, A. J. Korman, and R. D. Schreiber. 2017. Temporally Distinct

PD-L1 Expression by Tumor and Host Cells Contributes to Immune Escape. Cancer

Immunol Res 5: 106-117.

300. Zhang, S., X. Ma, C. Zhu, L. Liu, G. Wang, and X. Yuan. 2016. The Role of Myeloid-

Derived Suppressor Cells in Patients with Solid Tumors: A Meta-Analysis. PLOS ONE 11: e0164514.

301. Steele, C. W., S. A. Karim, J. D. G. Leach, P. Bailey, R. Upstill-Goddard, L. Rishi, M.

Foth, S. Bryson, K. McDaid, Z. Wilson, C. Eberlein, J. B. Candido, M. Clarke, C. Nixon, J.

Connelly, N. Jamieson, C. R. Carter, F. Balkwill, D. K. Chang, T. R. J. Evans, D. Strathdee,

A. V. Biankin, R. J. B. Nibbs, S. T. Barry, O. J. Sansom, and J. P. Morton. 2016. CXCR2

Inhibition Profoundly Suppresses Metastases and Augments Immunotherapy in Pancreatic

Ductal Adenocarcinoma. Cancer Cell 29: 832-845.

302. Devapatla, B., A. Sharma, and S. Woo. 2015. CXCR2 Inhibition Combined with

Sorafenib Improved Antitumor and Antiangiogenic Response in Preclinical Models of

Ovarian Cancer. PLOS ONE 10: e0139237.

303. Jamieson, T., M. Clarke, C. W. Steele, M. S. Samuel, J. Neumann, A. Jung, D.

Huels, M. F. Olson, S. Das, R. J. Nibbs, and O. J. Sansom. 2012. Inhibition of CXCR2 profoundly suppresses inflammation-driven and spontaneous tumorigenesis. J Clin Invest

122: 3127-3144.

304. Nywening, T. M., A. Wang-Gillam, D. E. Sanford, B. A. Belt, R. Z. Panni, B. M.

188

Cusworth, A. T. Toriola, R. K. Nieman, L. A. Worley, M. Yano, K. J. Fowler, A. C. Lockhart,

R. Suresh, B. R. Tan, K.-H. Lim, R. C. Fields, S. M. Strasberg, W. G. Hawkins, D. G.

DeNardo, S. P. Goedegebuure, and D. C. Linehan. 2016. Targeting tumour-associated

macrophages with CCR2 inhibition in combination with FOLFIRINOX in patients with

borderline resectable and locally advanced pancreatic cancer: a single-centre, open-label,

dose-finding, non-randomised, phase 1b trial. The Lancet. Oncology 17: 651-662.

305. Nywening, T. M., B. A. Belt, D. R. Cullinan, R. Z. Panni, B. J. Han, D. E. Sanford, R.

C. Jacobs, J. Ye, A. A. Patel, W. E. Gillanders, R. C. Fields, D. G. DeNardo, W. G. Hawkins,

P. Goedegebuure, and D. C. Linehan. 2018. Targeting both tumour-associated CXCR2(+)

neutrophils and CCR2(+) macrophages disrupts myeloid recruitment and improves

chemotherapeutic responses in pancreatic ductal adenocarcinoma. Gut 67: 1112-1123.

306. Blattner, C., V. Fleming, R. Weber, B. Himmelhan, P. Altevogt, C. Gebhardt, T. J.

Schulze, H. Razon, E. Hawila, G. Wildbaum, J. Utikal, N. Karin, and V. Umansky. 2018.

CCR5(+) Myeloid-Derived Suppressor Cells Are Enriched and Activated in Melanoma

Lesions. Cancer Res 78: 157-167.

307. Gabrilovich, D. I., M. P. Velders, E. M. Sotomayor, and W. M. Kast. 2001.

Mechanism of immune dysfunction in cancer mediated by immature Gr-1+ myeloid cells. J

Immunol 166: 5398-5406.

308. Nefedova, Y., M. Fishman, S. Sherman, X. Wang, A. A. Beg, and D. I. Gabrilovich.

2007. Mechanism of all-trans retinoic acid effect on tumor-associated myeloid-derived

suppressor cells. Cancer Res 67: 11021-11028.

309. Kusmartsev, S., F. Cheng, B. Yu, Y. Nefedova, E. Sotomayor, R. Lush, and D.

Gabrilovich. 2003. All-trans-retinoic acid eliminates immature myeloid cells from tumor- bearing mice and improves the effect of vaccination. Cancer Res 63: 4441-4449.

310. Mirza, N., M. Fishman, I. Fricke, M. Dunn, A. M. Neuger, T. J. Frost, R. M. Lush, S.

189

Antonia, and D. I. Gabrilovich. 2006. All-trans-retinoic acid improves differentiation of myeloid cells and immune response in cancer patients. Cancer Res 66: 9299-9307.

311. Tobin, R. P., K. R. Jordan, W. A. Robinson, D. Davis, V. F. Borges, R. Gonzalez, K.

D. Lewis, and M. D. McCarter. 2018. Targeting myeloid-derived suppressor cells using all- trans retinoic acid in melanoma patients treated with Ipilimumab. Int Immunopharmacol 63:

282-291.

312. Nefedova, Y., S. Nagaraj, A. Rosenbauer, C. Muro-Cacho, S. M. Sebti, and D. I.

Gabrilovich. 2005. Regulation of dendritic cell differentiation and antitumor immune response in cancer by pharmacologic-selective inhibition of the janus-activated kinase

2/signal transducers and activators of transcription 3 pathway. Cancer Res 65: 9525-9535.

313. Condamine, T., V. Kumar, I. R. Ramachandran, J. I. Youn, E. Celis, N. Finnberg, W.

S. El-Deiry, R. Winograd, R. H. Vonderheide, N. R. English, S. C. Knight, H. Yagita, J. C.

McCaffrey, S. Antonia, N. Hockstein, R. Witt, G. Masters, T. Bauer, and D. I. Gabrilovich.

2014. ER stress regulates myeloid-derived suppressor cell fate through TRAIL-R-mediated apoptosis. J Clin Invest 124: 2626-2639.

314. Qin, H., B. Lerman, I. Sakamaki, G. Wei, S. C. Cha, S. S. Rao, J. Qian, Y.

Hailemichael, R. Nurieva, K. C. Dwyer, J. Roth, Q. Yi, W. W. Overwijk, and L. W. Kwak.

2014. Generation of a new therapeutic peptide that depletes myeloid-derived suppressor cells in tumor-bearing mice. Nat Med 20: 676-681.

315. Suzuki, E., V. Kapoor, A. S. Jassar, L. R. Kaiser, and S. M. Albelda. 2005.

Gemcitabine selectively eliminates splenic Gr-1+/CD11b+ myeloid suppressor cells in tumor-bearing animals and enhances antitumor immune activity. Clinical cancer research : an official journal of the American Association for Cancer Research 11: 6713-6721.

316. Vincent, J., G. Mignot, F. Chalmin, S. Ladoire, M. Bruchard, A. Chevriaux, F. Martin,

L. Apetoh, C. Rebe, and F. Ghiringhelli. 2010. 5-Fluorouracil selectively kills tumor-

190

associated myeloid-derived suppressor cells resulting in enhanced T cell-dependent

antitumor immunity. Cancer Res 70: 3052-3061.

317. Kodumudi, K. N., K. Woan, D. L. Gilvary, E. Sahakian, S. Wei, and J. Y. Djeu. 2010.

A novel chemoimmunomodulating property of docetaxel: suppression of myeloid-derived suppressor cells in tumor bearers. Clinical cancer research : an official journal of the

American Association for Cancer Research 16: 4583-4594.

318. Alizadeh, D., M. Trad, N. T. Hanke, C. B. Larmonier, N. Janikashvili, B. Bonnotte, E.

Katsanis, and N. Larmonier. 2014. Doxorubicin eliminates myeloid-derived suppressor cells and enhances the efficacy of adoptive T-cell transfer in breast cancer. Cancer Res 74: 104-

118.

319. Nagaraj, S., J. I. Youn, H. Weber, C. Iclozan, L. Lu, M. J. Cotter, C. Meyer, C. R.

Becerra, M. Fishman, S. Antonia, M. B. Sporn, K. T. Liby, B. Rawal, J. H. Lee, and D. I.

Gabrilovich. 2010. Anti-inflammatory triterpenoid blocks immune suppressive function of

MDSCs and improves immune response in cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 16: 1812-1823.

320. De Santo, C., P. Serafini, I. Marigo, L. Dolcetti, M. Bolla, P. Del Soldato, C. Melani,

C. Guiducci, M. P. Colombo, M. Iezzi, P. Musiani, P. Zanovello, and V. Bronte. 2005.

Nitroaspirin corrects immune dysfunction in tumor-bearing hosts and promotes tumor eradication by cancer vaccination. Proceedings of the National Academy of Sciences of the

United States of America 102: 4185-4190.

321. Califano, J. A., Z. Khan, K. A. Noonan, L. Rudraraju, Z. Zhang, H. Wang, S.

Goodman, C. G. Gourin, P. K. Ha, C. Fakhry, J. Saunders, M. Levine, M. Tang, G. Neuner,

J. D. Richmon, R. Blanco, N. Agrawal, W. M. Koch, S. Marur, D. T. Weed, P. Serafini, and I.

Borrello. 2015. Tadalafil augments tumor specific immunity in patients with head and neck squamous cell carcinoma. Clinical cancer research : an official journal of the American

191

Association for Cancer Research 21: 30-38.

322. Serafini, P., K. Meckel, M. Kelso, K. Noonan, J. Califano, W. Koch, L. Dolcetti, V.

Bronte, and I. Borrello. 2006. Phosphodiesterase-5 inhibition augments endogenous

antitumor immunity by reducing myeloid-derived suppressor cell function. J Exp Med 203:

2691-2702.

323. Noonan, K. A., N. Ghosh, L. Rudraraju, M. Bui, and I. Borrello. 2014. Targeting

immune suppression with PDE5 inhibition in end-stage multiple myeloma. Cancer Immunol

Res 2: 725-731.

324. Weed, D. T., J. L. Vella, I. M. Reis, A. C. De la Fuente, C. Gomez, Z. Sargi, R.

Nazarian, J. Califano, I. Borrello, and P. Serafini. 2015. Tadalafil reduces myeloid-derived

suppressor cells and regulatory T cells and promotes tumor immunity in patients with head

and neck squamous cell carcinoma. Clinical cancer research : an official journal of the

American Association for Cancer Research 21: 39-48.

325. Orillion, A., A. Hashimoto, N. Damayanti, L. Shen, R. Adelaiye-Ogala, S. Arisa, S.

Chintala, P. Ordentlich, C. Kao, B. Elzey, D. Gabrilovich, and R. Pili. 2017. Entinostat

Neutralizes Myeloid-Derived Suppressor Cells and Enhances the Antitumor Effect of PD-1

Inhibition in Murine Models of Lung and Renal Cell Carcinoma. Clinical cancer research : an official journal of the American Association for Cancer Research 23: 5187-5201.

326. Kim, K., A. D. Skora, Z. Li, Q. Liu, A. J. Tam, R. L. Blosser, L. A. Diaz, Jr., N.

Papadopoulos, K. W. Kinzler, B. Vogelstein, and S. Zhou. 2014. Eradication of metastatic mouse cancers resistant to immune checkpoint blockade by suppression of myeloid-derived cells. Proc Natl Acad Sci U S A 111: 11774-11779.

327. Zhang, Q.-w., L. Liu, C.-y. Gong, H.-s. Shi, Y.-h. Zeng, X.-z. Wang, Y.-w. Zhao, and

Y.-q. Wei. 2012. Prognostic Significance of Tumor-Associated Macrophages in Solid Tumor:

A Meta-Analysis of the Literature. PLOS ONE 7: e50946.

192

328. Goswami, S., E. Sahai, J. B. Wyckoff, M. Cammer, D. Cox, F. J. Pixley, E. R.

Stanley, J. E. Segall, and J. S. Condeelis. 2005. Macrophages promote the invasion of breast carcinoma cells via a colony-stimulating factor-1/epidermal growth factor paracrine loop. Cancer Res 65: 5278-5283.

329. Zhu, X. D., J. B. Zhang, P. Y. Zhuang, H. G. Zhu, W. Zhang, Y. Q. Xiong, W. Z. Wu,

L. Wang, Z. Y. Tang, and H. C. Sun. 2008. High expression of macrophage colony- stimulating factor in peritumoral liver tissue is associated with poor survival after curative resection of hepatocellular carcinoma. J Clin Oncol 26: 2707-2716.

330. Zhu, Y., B. L. Knolhoff, M. A. Meyer, T. M. Nywening, B. L. West, J. Luo, A. Wang-

Gillam, S. P. Goedegebuure, D. C. Linehan, and D. G. DeNardo. 2014. CSF1/CSF1R blockade reprograms tumor-infiltrating macrophages and improves response to T-cell checkpoint immunotherapy in pancreatic cancer models. Cancer Res 74: 5057-5069.

331. Escamilla, J., S. Schokrpur, C. Liu, S. J. Priceman, D. Moughon, Z. Jiang, F. Pouliot,

C. Magyar, J. L. Sung, J. Xu, G. Deng, B. L. West, G. Bollag, Y. Fradet, L. Lacombe, M. E.

Jung, J. Huang, and L. Wu. 2015. CSF1 receptor targeting in prostate cancer reverses macrophage-mediated resistance to androgen blockade therapy. Cancer Res 75: 950-962.

332. Weizman, N., Y. Krelin, A. Shabtay-Orbach, M. Amit, Y. Binenbaum, R. J. Wong, and

Z. Gil. 2014. Macrophages mediate gemcitabine resistance of pancreatic adenocarcinoma by upregulating cytidine deaminase. Oncogene 33: 3812-3819.

333. DeNardo, D. G., D. J. Brennan, E. Rexhepaj, B. Ruffell, S. L. Shiao, S. F. Madden,

W. M. Gallagher, N. Wadhwani, S. D. Keil, S. A. Junaid, H. S. Rugo, E. S. Hwang, K.

Jirstrom, B. L. West, and L. M. Coussens. 2011. Leukocyte complexity predicts breast cancer survival and functionally regulates response to chemotherapy. Cancer Discov 1: 54-

67.

334. Cassier, P. A., A. Italiano, C. A. Gomez-Roca, C. Le Tourneau, M. Toulmonde, M. A.

193

Cannarile, C. Ries, A. Brillouet, C. Muller, A. M. Jegg, A. M. Broske, M. Dembowski, K.

Bray-French, C. Freilinger, G. Meneses-Lorente, M. Baehner, R. Harding, J. Ratnayake, K.

Abiraj, N. Gass, K. Noh, R. D. Christen, L. Ukarma, E. Bompas, J. P. Delord, J. Y. Blay, and

D. Ruttinger. 2015. CSF1R inhibition with emactuzumab in locally advanced diffuse-type tenosynovial giant cell tumours of the soft tissue: a dose-escalation and dose-expansion phase 1 study. Lancet Oncol 16: 949-956.

335. Mantovani, A., F. Marchesi, A. Malesci, L. Laghi, and P. Allavena. 2017. Tumour- associated macrophages as treatment targets in oncology. Nat Rev Clin Oncol 14: 399-416.

336. Cannarile, M. A., M. Weisser, W. Jacob, A. M. Jegg, C. H. Ries, and D. Ruttinger.

2017. Colony-stimulating factor 1 receptor (CSF1R) inhibitors in cancer therapy. J

Immunother Cancer 5: 53.

337. Kaneda, M. M., K. S. Messer, N. Ralainirina, H. Li, C. J. Leem, S. Gorjestani, G.

Woo, A. V. Nguyen, C. C. Figueiredo, P. Foubert, M. C. Schmid, M. Pink, D. G. Winkler, M.

Rausch, V. J. Palombella, J. Kutok, K. McGovern, K. A. Frazer, X. Wu, M. Karin, R. Sasik,

E. E. Cohen, and J. A. Varner. 2016. PI3Kgamma is a molecular switch that controls immune suppression. Nature 539: 437-442.

338. Kaneda, M. M., P. Cappello, A. V. Nguyen, N. Ralainirina, C. R. Hardamon, P.

Foubert, M. C. Schmid, P. Sun, E. Mose, M. Bouvet, A. M. Lowy, M. A. Valasek, R. Sasik, F.

Novelli, E. Hirsch, and J. A. Varner. 2016. Macrophage PI3Kgamma Drives Pancreatic

Ductal Adenocarcinoma Progression. Cancer Discov 6: 870-885.

339. De Henau, O., M. Rausch, D. Winkler, L. F. Campesato, C. Liu, D. H. Cymerman, S.

Budhu, A. Ghosh, M. Pink, J. Tchaicha, M. Douglas, T. Tibbitts, S. Sharma, J. Proctor, N.

Kosmider, K. White, H. Stern, J. Soglia, J. Adams, V. J. Palombella, K. McGovern, J. L.

Kutok, J. D. Wolchok, and T. Merghoub. 2016. Overcoming resistance to checkpoint blockade therapy by targeting PI3Kgamma in myeloid cells. Nature 539: 443-447.

194

340. Strauss, L., S. Sangaletti, F. M. Consonni, G. Szebeni, S. Morlacchi, M. G. Totaro, C.

Porta, A. Anselmo, S. Tartari, A. Doni, F. Zitelli, C. Tripodo, M. P. Colombo, and A. Sica.

2015. RORC1 Regulates Tumor-Promoting "Emergency" Granulo-Monocytopoiesis. Cancer

Cell 28: 253-269.

341. Beatty, G. L., E. G. Chiorean, M. P. Fishman, B. Saboury, U. R. Teitelbaum, W. Sun,

R. D. Huhn, W. Song, D. Li, L. L. Sharp, D. A. Torigian, P. J. O'Dwyer, and R. H.

Vonderheide. 2011. CD40 agonists alter tumor stroma and show efficacy against pancreatic carcinoma in mice and humans. Science 331: 1612-1616.

342. Beatty, G. L., D. A. Torigian, E. G. Chiorean, B. Saboury, A. Brothers, A. Alavi, A. B.

Troxel, W. Sun, U. R. Teitelbaum, R. H. Vonderheide, and P. J. O'Dwyer. 2013. A phase I study of an agonist CD40 monoclonal antibody (CP-870,893) in combination with gemcitabine in patients with advanced pancreatic ductal adenocarcinoma. Clinical cancer research : an official journal of the American Association for Cancer Research 19: 6286-

6295.

343. Rodell, C. B., S. P. Arlauckas, M. F. Cuccarese, C. S. Garris, R. Li, M. S. Ahmed, R.

H. Kohler, M. J. Pittet, and R. Weissleder. 2018. TLR7/8-agonist-loaded nanoparticles promote the polarization of tumour-associated macrophages to enhance cancer immunotherapy. Nat Biomed Eng 2: 578-588.

344. Singh, M., H. Khong, Z. Dai, X.-F. Huang, J. A. Wargo, Z. A. Cooper, J. P. Vasilakos,

P. Hwu, and W. W. Overwijk. 2014. Effective Innate and Adaptive Antimelanoma Immunity through Localized TLR7/8 Activation. The Journal of Immunology 193: 4722-4731.

345. Wang, Y., L. Su, M. D. Morin, B. T. Jones, L. R. Whitby, M. M. R. P. Surakattula, H.

Huang, H. Shi, J. H. Choi, K.-w. Wang, E. M. Y. Moresco, M. Berger, X. Zhan, H. Zhang, D.

L. Boger, and B. Beutler. 2016. TLR4/MD-2 activation by a synthetic agonist with no similarity to LPS. Proceedings of the National Academy of Sciences 113: E884-E893.

195

346. Zhao, B. G., J. P. Vasilakos, D. Tross, D. Smirnov, and D. M. Klinman. 2014.

Combination therapy targeting toll like receptors 7, 8 and 9 eliminates large established

tumors. J Immunother Cancer 2: 12.

347. Iribarren, K., N. Bloy, A. Buque, I. Cremer, A. Eggermont, W. H. Fridman, J.

Fucikova, J. Galon, R. Spisek, L. Zitvogel, G. Kroemer, and L. Galluzzi. 2016. Trial Watch:

Immunostimulation with Toll-like receptor agonists in cancer therapy. Oncoimmunology 5: e1088631.

348. Woo, S. R., M. B. Fuertes, L. Corrales, S. Spranger, M. J. Furdyna, M. Y. Leung, R.

Duggan, Y. Wang, G. N. Barber, K. A. Fitzgerald, M. L. Alegre, and T. F. Gajewski. 2014.

STING-dependent cytosolic DNA sensing mediates innate immune recognition of immunogenic tumors. Immunity 41: 830-842.

349. Stetson, D. B., and R. Medzhitov. 2006. Recognition of cytosolic DNA activates an

IRF3-dependent innate immune response. Immunity 24: 93-103.

350. Alexopoulou, L., A. C. Holt, R. Medzhitov, and R. A. Flavell. 2001. Recognition of double-stranded RNA and activation of NF-kappaB by Toll-like receptor 3. Nature 413: 732-

738.

351. Hemmi, H., O. Takeuchi, T. Kawai, T. Kaisho, S. Sato, H. Sanjo, M. Matsumoto, K.

Hoshino, H. Wagner, K. Takeda, and S. Akira. 2000. A Toll-like receptor recognizes bacterial DNA. Nature 408: 740-745.

352. Kang, D. C., R. V. Gopalkrishnan, Q. Wu, E. Jankowsky, A. M. Pyle, and P. B.

Fisher. 2002. mda-5: An interferon-inducible putative RNA helicase with double-stranded

RNA-dependent ATPase activity and melanoma growth-suppressive properties. Proc Natl

Acad Sci U S A 99: 637-642.

353. Yoneyama, M., M. Kikuchi, T. Natsukawa, N. Shinobu, T. Imaizumi, M. Miyagishi, K.

Taira, S. Akira, and T. Fujita. 2004. The RNA helicase RIG-I has an essential function in

196

double-stranded RNA-induced innate antiviral responses. Nature immunology 5: 730-737.

354. Ishii, K. J., C. Coban, H. Kato, K. Takahashi, Y. Torii, F. Takeshita, H. Ludwig, G.

Sutter, K. Suzuki, H. Hemmi, S. Sato, M. Yamamoto, S. Uematsu, T. Kawai, O. Takeuchi,

and S. Akira. 2006. A Toll-like receptor-independent antiviral response induced by double-

stranded B-form DNA. Nature immunology 7: 40-48.

355. Jin, L., P. M. Waterman, K. R. Jonscher, C. M. Short, N. A. Reisdorph, and J. C.

Cambier. 2008. MPYS, a novel membrane tetraspanner, is associated with major

histocompatibility complex class II and mediates transduction of apoptotic signals. Mol Cell

Biol 28: 5014-5026.

356. Ishikawa, H., and G. N. Barber. 2008. STING is an endoplasmic reticulum adaptor

that facilitates innate immune signalling. Nature 455: 674-678.

357. Zhong, B., Y. Yang, S. Li, Y. Y. Wang, Y. Li, F. Diao, C. Lei, X. He, L. Zhang, P.

Tien, and H. B. Shu. 2008. The adaptor protein MITA links virus-sensing receptors to IRF3

transcription factor activation. Immunity 29: 538-550.

358. Ishikawa, H., Z. Ma, and G. N. Barber. 2009. STING regulates intracellular DNA-

mediated, type I interferon-dependent innate immunity. Nature 461: 788-792.

359. Sun, W., Y. Li, L. Chen, H. Chen, F. You, X. Zhou, Y. Zhou, Z. Zhai, D. Chen, and Z.

Jiang. 2009. ERIS, an endoplasmic reticulum IFN stimulator, activates innate immune

signaling through dimerization. Proc Natl Acad Sci U S A 106: 8653-8658.

360. Burdette, D. L., K. M. Monroe, K. Sotelo-Troha, J. S. Iwig, B. Eckert, M. Hyodo, Y.

Hayakawa, and R. E. Vance. 2011. STING is a direct innate immune sensor of cyclic di-

GMP. Nature 478: 515-518.

361. Sauer, J. D., K. Sotelo-Troha, J. von Moltke, K. M. Monroe, C. S. Rae, S. W.

Brubaker, M. Hyodo, Y. Hayakawa, J. J. Woodward, D. A. Portnoy, and R. E. Vance. 2011.

The N-ethyl-N-nitrosourea-induced Goldenticket mouse mutant reveals an essential function

197

of Sting in the in vivo interferon response to Listeria monocytogenes and cyclic dinucleotides. Infect Immun 79: 688-694.

362. Wu, J., L. Sun, X. Chen, F. Du, H. Shi, C. Chen, and Z. J. Chen. 2013. Cyclic GMP-

AMP is an endogenous second messenger in innate immune signaling by cytosolic DNA.

Science 339: 826-830.

363. Sun, L., J. Wu, F. Du, X. Chen, and Z. J. Chen. 2013. Cyclic GMP-AMP synthase is a cytosolic DNA sensor that activates the type I interferon pathway. Science 339: 786-791.

364. Ablasser, A., M. Goldeck, T. Cavlar, T. Deimling, G. Witte, I. Rohl, K. P. Hopfner, J.

Ludwig, and V. Hornung. 2013. cGAS produces a 2'-5'-linked cyclic dinucleotide second messenger that activates STING. Nature 498: 380-384.

365. Gao, P., M. Ascano, Y. Wu, W. Barchet, B. L. Gaffney, T. Zillinger, A. A. Serganov,

Y. Liu, R. A. Jones, G. Hartmann, T. Tuschl, and D. J. Patel. 2013. Cyclic [G(2',5')pA(3',5')p] is the metazoan second messenger produced by DNA-activated cyclic GMP-AMP synthase.

Cell 153: 1094-1107.

366. Zhang, X., H. Shi, J. Wu, X. Zhang, L. Sun, C. Chen, and Z. J. Chen. 2013. Cyclic

GMP-AMP containing mixed phosphodiester linkages is an endogenous high-affinity ligand for STING. Mol Cell 51: 226-235.

367. Kranzusch, P. J., S. C. Wilson, A. S. Lee, J. M. Berger, J. A. Doudna, and R. E.

Vance. 2015. Ancient Origin of cGAS-STING Reveals Mechanism of Universal 2',3' cGAMP

Signaling. Mol Cell 59: 891-903.

368. Margolis, S. R., S. C. Wilson, and R. E. Vance. 2017. Evolutionary Origins of cGAS-

STING Signaling. Trends Immunol 38: 733-743.

369. Gluck, S., B. Guey, M. F. Gulen, K. Wolter, T. W. Kang, N. A. Schmacke, A.

Bridgeman, J. Rehwinkel, L. Zender, and A. Ablasser. 2017. Innate immune sensing of cytosolic chromatin fragments through cGAS promotes senescence. Nat Cell Biol 19: 1061-

198

1070.

370. Harding, S. M., J. L. Benci, J. Irianto, D. E. Discher, A. J. Minn, and R. A. Greenberg.

2017. Mitotic progression following DNA damage enables pattern recognition within micronuclei. Nature 548: 466-470.

371. Mackenzie, K. J., P. Carroll, C. A. Martin, O. Murina, A. Fluteau, D. J. Simpson, N.

Olova, H. Sutcliffe, J. K. Rainger, A. Leitch, R. T. Osborn, A. P. Wheeler, M. Nowotny, N.

Gilbert, T. Chandra, M. A. M. Reijns, and A. P. Jackson. 2017. cGAS surveillance of micronuclei links genome instability to innate immunity. Nature 548: 461-465.

372. Yang, H., H. Wang, J. Ren, Q. Chen, and Z. J. Chen. 2017. cGAS is essential for cellular senescence. Proceedings of the National Academy of Sciences 114: E4612-E4620.

373. Klarquist, J., C. M. Hennies, M. A. Lehn, R. A. Reboulet, S. Feau, and E. M.

Janssen. 2014. STING-mediated DNA sensing promotes antitumor and autoimmune responses to dying cells. J Immunol 193: 6124-6134.

374. Rongvaux, A., R. Jackson, C. C. Harman, T. Li, A. P. West, M. R. de Zoete, Y. Wu,

B. Yordy, S. A. Lakhani, C. Y. Kuan, T. Taniguchi, G. S. Shadel, Z. J. Chen, A. Iwasaki, and

R. A. Flavell. 2014. Apoptotic caspases prevent the induction of type I interferons by mitochondrial DNA. Cell 159: 1563-1577.

375. White, M. J., K. McArthur, D. Metcalf, R. M. Lane, J. C. Cambier, M. J. Herold, M. F. van Delft, S. Bedoui, G. Lessene, M. E. Ritchie, D. C. Huang, and B. T. Kile. 2014. Apoptotic caspases suppress mtDNA-induced STING-mediated type I IFN production. Cell 159: 1549-

1562.

376. Civril, F., T. Deimling, C. C. de Oliveira Mann, A. Ablasser, M. Moldt, G. Witte, V.

Hornung, and K. P. Hopfner. 2013. Structural mechanism of cytosolic DNA sensing by cGAS. Nature 498: 332-337.

377. Tao, J., X. W. Zhang, J. Jin, X. X. Du, T. Lian, J. Yang, X. Zhou, Z. Jiang, and X. D.

199

Su. 2017. Nonspecific DNA Binding of cGAS N Terminus Promotes cGAS Activation. J

Immunol 198: 3627-3636.

378. Andreeva, L., B. Hiller, D. Kostrewa, C. Lassig, C. C. de Oliveira Mann, D. Jan

Drexler, A. Maiser, M. Gaidt, H. Leonhardt, V. Hornung, and K. P. Hopfner. 2017. cGAS senses long and HMGB/TFAM-bound U-turn DNA by forming protein-DNA ladders. Nature

549: 394-398.

379. Luecke, S., A. Holleufer, M. H. Christensen, K. L. Jonsson, G. A. Boni, L. K.

Sorensen, M. Johannsen, M. R. Jakobsen, R. Hartmann, and S. R. Paludan. 2017. cGAS is activated by DNA in a length-dependent manner. EMBO Rep 18: 1707-1715.

380. Li, X., C. Shu, G. Yi, C. T. Chaton, C. L. Shelton, J. Diao, X. Zuo, C. C. Kao, A. B.

Herr, and P. Li. 2013. Cyclic GMP-AMP synthase is activated by double-stranded DNA- induced oligomerization. Immunity 39: 1019-1031.

381. Zhou, W., A. T. Whiteley, C. C. de Oliveira Mann, B. R. Morehouse, R. P. Nowak, E.

S. Fischer, N. S. Gray, J. J. Mekalanos, and P. J. Kranzusch. 2018. Structure of the Human cGAS-DNA Complex Reveals Enhanced Control of Immune Surveillance. Cell 174: 300-311 e311.

382. Du, M., and Z. J. Chen. 2018. DNA-induced liquid phase condensation of cGAS activates innate immune signaling. Science 361: 704-709.

383. Barber, G. N. 2015. STING: infection, inflammation and cancer. Nature reviews.

Immunology 15: 760-770.

384. Liu, S., X. Cai, J. Wu, Q. Cong, X. Chen, T. Li, F. Du, J. Ren, Y. T. Wu, N. V. Grishin, and Z. J. Chen. 2015. Phosphorylation of innate immune adaptor proteins MAVS, STING, and TRIF induces IRF3 activation. Science 347: aaa2630.

385. Tanaka, Y., and Z. J. Chen. 2012. STING specifies IRF3 phosphorylation by TBK1 in the cytosolic DNA signaling pathway. Sci Signal 5: ra20.

200

386. Abe, T., and G. N. Barber. 2014. Cytosolic-DNA-mediated, STING-dependent proinflammatory gene induction necessitates canonical NF-kappaB activation through TBK1.

J Virol 88: 5328-5341.

387. Chen, H., H. Sun, F. You, W. Sun, X. Zhou, L. Chen, J. Yang, Y. Wang, H. Tang, Y.

Guan, W. Xia, J. Gu, H. Ishikawa, D. Gutman, G. Barber, Z. Qin, and Z. Jiang. 2011.

Activation of STAT6 by STING is critical for antiviral innate immunity. Cell 147: 436-446.

388. Diamond, M. S., M. Kinder, H. Matsushita, M. Mashayekhi, G. P. Dunn, J. M.

Archambault, H. Lee, C. D. Arthur, J. M. White, U. Kalinke, K. M. Murphy, and R. D.

Schreiber. 2011. Type I interferon is selectively required by dendritic cells for immune rejection of tumors. J Exp Med 208: 1989-2003.

389. Ohkuri, T., A. Ghosh, A. Kosaka, J. Zhu, M. Ikeura, M. David, S. C. Watkins, S. N.

Sarkar, and H. Okada. 2014. STING contributes to antiglioma immunity via triggering type I

IFN signals in the tumor microenvironment. Cancer Immunol Res 2: 1199-1208.

390. Zhu, Q., S. M. Man, P. Gurung, Z. Liu, P. Vogel, M. Lamkanfi, and T.-D. Kanneganti.

2014. Cutting Edge: STING Mediates Protection against Colorectal Tumorigenesis by

Governing the Magnitude of Intestinal Inflammation. The Journal of Immunology 193: 4779-

4782.

391. Xia, T., H. Konno, J. Ahn, and G. N. Barber. 2016. Deregulation of STING Signaling in Colorectal Carcinoma Constrains DNA Damage Responses and Correlates With

Tumorigenesis. Cell Rep 14: 282-297.

392. Xia, T., H. Konno, and G. N. Barber. 2016. Recurrent Loss of STING Signaling in

Melanoma Correlates with Susceptibility to Viral Oncolysis. Cancer Res 76: 6747-6759.

393. Lau, L., E. E. Gray, R. L. Brunette, and D. B. Stetson. 2015. DNA tumor virus oncogenes antagonize the cGAS-STING DNA-sensing pathway. Science 350: 568-571.

394. Wasserman, J., H. Blomgren, S. Rotstein, B. Petrini, and S. Hammarstrom. 1989.

201

Immunosuppression in irradiated breast cancer patients: in vitro effect of cyclooxygenase

inhibitors. Bull N Y Acad Med 65: 36-44.

395. Lee, Y., S. L. Auh, Y. Wang, B. Burnette, Y. Wang, Y. Meng, M. Beckett, R. Sharma,

R. Chin, T. Tu, R. R. Weichselbaum, and Y. X. Fu. 2009. Therapeutic effects of ablative

radiation on local tumor require CD8+ T cells: changing strategies for cancer treatment.

Blood 114: 589-595.

396. Deng, L., H. Liang, B. Burnette, M. Beckett, T. Darga, R. R. Weichselbaum, and Y.-

X. Fu. 2014. Irradiation and anti–PD-L1 treatment synergistically promote antitumor immunity in mice. The Journal of Clinical Investigation 124: 687-695.

397. Deng, L., H. Liang, M. Xu, X. Yang, B. Burnette, A. Arina, X. D. Li, H. Mauceri, M.

Beckett, T. Darga, X. Huang, T. F. Gajewski, Z. J. Chen, Y. X. Fu, and R. R. Weichselbaum.

2014. STING-Dependent Cytosolic DNA Sensing Promotes Radiation-Induced Type I

Interferon-Dependent Antitumor Immunity in Immunogenic Tumors. Immunity 41: 843-852.

398. Vanpouille-Box, C., A. Alard, M. J. Aryankalayil, Y. Sarfraz, J. M. Diamond, R. J.

Schneider, G. Inghirami, C. N. Coleman, S. C. Formenti, and S. Demaria. 2017. DNA exonuclease Trex1 regulates radiotherapy-induced tumour immunogenicity. Nat Commun 8:

15618.

399. Golden, E. B., A. Chhabra, A. Chachoua, S. Adams, M. Donach, M. Fenton-

Kerimian, K. Friedman, F. Ponzo, J. S. Babb, J. Goldberg, S. Demaria, and S. C. Formenti.

2015. Local radiotherapy and granulocyte-macrophage colony-stimulating factor to generate abscopal responses in patients with metastatic solid tumours: a proof-of-principle trial.

Lancet Oncol 16: 795-803.

400. Golden, E. B., S. Demaria, P. B. Schiff, A. Chachoua, and S. C. Formenti. 2013. An

abscopal response to radiation and ipilimumab in a patient with metastatic non-small cell

lung cancer. Cancer Immunol Res 1: 365-372.

202

401. Postow, M. A., M. K. Callahan, C. A. Barker, Y. Yamada, J. Yuan, S. Kitano, Z. Mu,

T. Rasalan, M. Adamow, E. Ritter, C. Sedrak, A. A. Jungbluth, R. Chua, A. S. Yang, R. A.

Roman, S. Rosner, B. Benson, J. P. Allison, A. M. Lesokhin, S. Gnjatic, and J. D. Wolchok.

2012. Immunologic correlates of the abscopal effect in a patient with melanoma. The New

England journal of medicine 366: 925-931.

402. Twyman-Saint Victor, C., A. J. Rech, A. Maity, R. Rengan, K. E. Pauken, E.

Stelekati, J. L. Benci, B. Xu, H. Dada, P. M. Odorizzi, R. S. Herati, K. D. Mansfield, D.

Patsch, R. K. Amaravadi, L. M. Schuchter, H. Ishwaran, R. Mick, D. A. Pryma, X. Xu, M. D.

Feldman, T. C. Gangadhar, S. M. Hahn, E. J. Wherry, R. H. Vonderheide, and A. J. Minn.

2015. Radiation and dual checkpoint blockade activate non-redundant immune mechanisms in cancer. Nature 520: 373-377.

403. Steinberger, O., Z. Lapidot, Z. Ben-Ishai, and D. Amikam. 1999. Elevated expression of the CD4 receptor and cell cycle arrest are induced in Jurkat cells by treatment with the novel cyclic dinucleotide 3',5'-cyclic diguanylic acid. FEBS letters 444: 125-129.

404. Cerboni, S., N. Jeremiah, M. Gentili, U. Gehrmann, C. Conrad, M. C. Stolzenberg, C.

Picard, B. Neven, A. Fischer, S. Amigorena, F. Rieux-Laucat, and N. Manel. 2017. Intrinsic antiproliferative activity of the innate sensor STING in T lymphocytes. J Exp Med 214: 1769-

1785.

405. Gulen, M. F., U. Koch, S. M. Haag, F. Schuler, L. Apetoh, A. Villunger, F. Radtke, and A. Ablasser. 2017. Signalling strength determines proapoptotic functions of STING. Nat

Commun 8: 427.

406. Larkin, B., V. Ilyukha, M. Sorokin, A. Buzdin, E. Vannier, and A. Poltorak. 2017.

Cutting Edge: Activation of STING in T Cells Induces Type I IFN Responses and Cell Death.

J Immunol 199: 397-402.

407. Gasser, S., S. Orsulic, E. J. Brown, and D. H. Raulet. 2005. The DNA damage

203

pathway regulates innate immune system ligands of the NKG2D receptor. Nature 436:

1186-1190.

408. Lam, A. R., N. L. Bert, S. S. Ho, Y. J. Shen, L. F. Tang, G. M. Xiong, J. L. Croxford,

C. X. Koo, K. J. Ishii, S. Akira, D. H. Raulet, and S. Gasser. 2014. RAE1 ligands for the

NKG2D receptor are regulated by STING-dependent DNA sensor pathways in lymphoma.

Cancer Res 74: 2193-2203.

409. Swann, J. B., Y. Hayakawa, N. Zerafa, K. C. Sheehan, B. Scott, R. D. Schreiber, P.

Hertzog, and M. J. Smyth. 2007. Type I IFN contributes to NK cell homeostasis, activation, and antitumor function. J Immunol 178: 7540-7549.

410. Takashima, K., Y. Takeda, H. Oshiumi, H. Shime, M. Okabe, M. Ikawa, M.

Matsumoto, and T. Seya. 2016. STING in tumor and host cells cooperatively work for NK cell-mediated tumor growth retardation. Biochem Biophys Res Commun 478: 1764-1771.

411. Lu, S., F. Concha-Benavente, G. Shayan, R. M. Srivastava, S. P. Gibson, L. Wang,

W. E. Gooding, and R. L. Ferris. 2018. STING activation enhances cetuximab-mediated NK cell activation and DC maturation and correlates with HPV(+) status in head and neck cancer. Oral Oncol 78: 186-193.

412. Kim, Y., D. Pardoll, J. Fu, and T. Dubensky. 2015. Abstract 2495: STING agonists formulated into cancer vaccines (STINGVAX) can cure established tumor resistant to immune checkpoint blockade by activating NK cells. Cancer Research 75: 2495-2495.

413. Baird, J. R., Z. Feng, H. D. Xiao, D. Friedman, B. Cottam, B. A. Fox, G. Kramer, R.

S. Leidner, R. B. Bell, K. H. Young, M. R. Crittenden, and M. J. Gough. 2017. STING expression and response to treatment with STING ligands in premalignant and malignant disease. PLoS One 12: e0187532.

414. Demaria, O., A. De Gassart, S. Coso, N. Gestermann, J. Di Domizio, L. Flatz, O.

Gaide, O. Michielin, P. Hwu, T. V. Petrova, F. Martinon, R. L. Modlin, D. E. Speiser, and M.

204

Gilliet. 2015. STING activation of tumor endothelial cells initiates spontaneous and

therapeutic antitumor immunity. Proc Natl Acad Sci U S A 112: 15408-15413.

415. Hill, S., K. B. Williams, and J. Denekamp. 1989. Vascular collapse after flavone

acetic acid: a possible mechanism of its anti-tumour action. Eur J Cancer Clin Oncol 25:

1419-1424.

416. Bibby, M. C., R. M. Phillips, J. A. Double, and G. Pratesi. 1991. Anti-tumour activity

of flavone acetic acid (NSC 347512) in mice--influence of immune status. Br J Cancer 63:

57-62.

417. Laws, A. L., A. M. Matthew, J. A. Double, and M. C. Bibby. 1995. Preclinical in vitro

and in vivo activity of 5,6-dimethylxanthenone-4-acetic acid. Br J Cancer 71: 1204-1209.

418. Roberts, Z. J., L. M. Ching, and S. N. Vogel. 2008. IFN-beta-dependent inhibition of

tumor growth by the vascular disrupting agent 5,6-dimethylxanthenone-4-acetic acid

(DMXAA). J Interferon Cytokine Res 28: 133-139.

419. Jassar, A. S., E. Suzuki, V. Kapoor, J. Sun, M. B. Silverberg, L. Cheung, M. D.

Burdick, R. M. Strieter, L. M. Ching, L. R. Kaiser, and S. M. Albelda. 2005. Activation of

tumor-associated macrophages by the vascular disrupting agent 5,6-dimethylxanthenone-4- acetic acid induces an effective CD8+ T-cell-mediated antitumor immune response in murine models of lung cancer and mesothelioma. Cancer Res 65: 11752-11761.

420. Kerr, D. J., T. Maughan, E. Newlands, G. Rustin, N. M. Bleehen, C. Lewis, and S. B.

Kaye. 1989. Phase II trials of flavone acetic acid in advanced malignant melanoma and colorectal carcinoma. Br J Cancer 60: 104-106.

421. Ching, L. M., W. R. Joseph, and B. C. Baguley. 1992. Antitumour responses to flavone-8-acetic acid and 5,6-dimethylxanthenone-4-acetic acid in immune deficient mice. Br

J Cancer 66: 128-130.

422. Rewcastle, G. W., G. J. Atwell, Z. A. Li, B. C. Baguley, and W. A. Denny. 1991.

205

Potential antitumor agents. 61. Structure-activity relationships for in vivo colon 38 activity

among disubstituted 9-oxo-9H-xanthene-4-acetic acids. J Med Chem 34: 217-222.

423. McKeage, M. J., M. Reck, M. B. Jameson, M. A. Rosenthal, D. Gibbs, P. N.

Mainwaring, L. Freitag, R. Sullivan, and J. Von Pawel. 2009. Phase II study of ASA404

(vadimezan, 5,6-dimethylxanthenone-4-acetic acid/DMXAA) 1800mg/m(2) combined with

carboplatin and paclitaxel in previously untreated advanced non-small cell lung cancer. Lung

Cancer 65: 192-197.

424. Lara, P. N., Jr., J. Y. Douillard, K. Nakagawa, J. von Pawel, M. J. McKeage, I. Albert,

G. Losonczy, M. Reck, D. S. Heo, X. Fan, A. Fandi, and G. Scagliotti. 2011. Randomized

phase III placebo-controlled trial of carboplatin and paclitaxel with or without the vascular

disrupting agent vadimezan (ASA404) in advanced non-small-cell lung cancer. J Clin Oncol

29: 2965-2971.

425. Conlon, J., D. L. Burdette, S. Sharma, N. Bhat, M. Thompson, Z. Jiang, V. A.

Rathinam, B. Monks, T. Jin, T. S. Xiao, S. N. Vogel, R. E. Vance, and K. A. Fitzgerald. 2013.

Mouse, but not human STING, binds and signals in response to the vascular disrupting

agent 5,6-dimethylxanthenone-4-acetic acid. J Immunol 190: 5216-5225.

426. Kim, S., L. Li, Z. Maliga, Q. Yin, H. Wu, and T. J. Mitchison. 2013. Anticancer flavonoids are mouse-selective STING agonists. ACS Chem Biol 8: 1396-1401.

427. Gao, J., J. Tao, W. Liang, M. Zhao, X. Du, S. Cui, H. Duan, B. Kan, X. Su, and Z.

Jiang. 2015. Identification and characterization of phosphodiesterases that specifically degrade 3'3'-cyclic GMP-AMP. Cell Res 25: 539-550.

428. Simm, R., M. Morr, A. Kader, M. Nimtz, and U. Romling. 2004. GGDEF and EAL domains inversely regulate cyclic di-GMP levels and transition from sessility to motility. Mol

Microbiol 53: 1123-1134.

429. Corrales, L., L. H. Glickman, S. M. McWhirter, D. B. Kanne, K. E. Sivick, G. E.

206

Katibah, S. R. Woo, E. Lemmens, T. Banda, J. J. Leong, K. Metchette, T. W. Dubensky, Jr.,

and T. F. Gajewski. 2015. Direct Activation of STING in the Tumor Microenvironment Leads

to Potent and Systemic Tumor Regression and Immunity. Cell Rep 11: 1018-1030.

430. Yi, G., V. P. Brendel, C. Shu, P. Li, S. Palanathan, and C. Cheng Kao. 2013. Single

nucleotide polymorphisms of human STING can affect innate immune response to cyclic

dinucleotides. PLoS One 8: e77846.

431. Li, L., Q. Yin, P. Kuss, Z. Maliga, J. L. Millan, H. Wu, and T. J. Mitchison. 2014.

Hydrolysis of 2'3'-cGAMP by ENPP1 and design of nonhydrolyzable analogs. Nat Chem Biol

10: 1043-1048.

432. Putney, S. D., S. J. Benkovic, and P. R. Schimmel. 1981. A DNA fragment with an alpha-phosphorothioate nucleotide at one end is asymmetrically blocked from digestion by exonuclease III and can be replicated in vivo. Proc Natl Acad Sci U S A 78: 7350-7354.

433. Yan, H., X. Wang, R. KuoLee, and W. Chen. 2008. Synthesis and immunostimulatory properties of the phosphorothioate analogues of cdiGMP. Bioorg Med

Chem Lett 18: 5631-5634.

434. Chandra, D., W. Quispe-Tintaya, A. Jahangir, D. Asafu-Adjei, I. Ramos, H. O. Sintim,

J. Zhou, Y. Hayakawa, D. K. Karaolis, and C. Gravekamp. 2014. STING ligand c-di-GMP improves cancer vaccination against metastatic breast cancer. Cancer Immunol Res 2: 901-

910.

435. Wang, Z., and E. Celis. 2015. STING activator c-di-GMP enhances the anti-tumor effects of peptide vaccines in melanoma-bearing mice. Cancer immunology, immunotherapy

: CII 64: 1057-1066.

436. Hanson, M. C., M. P. Crespo, W. Abraham, K. D. Moynihan, G. L. Szeto, S. H. Chen,

M. B. Melo, S. Mueller, and D. J. Irvine. 2015. Nanoparticulate STING agonists are potent lymph node-targeted vaccine adjuvants. J Clin Invest 125: 2532-2546.

207

437. Fu, J., D. B. Kanne, M. Leong, L. H. Glickman, S. M. McWhirter, E. Lemmens, K.

Mechette, J. J. Leong, P. Lauer, W. Liu, K. E. Sivick, Q. Zeng, K. C. Soares, L. Zheng, D. A.

Portnoy, J. J. Woodward, D. M. Pardoll, T. W. Dubensky, Jr., and Y. Kim. 2015. STING

agonist formulated cancer vaccines can cure established tumors resistant to PD-1 blockade.

Sci Transl Med 7: 283ra252.

438. Moore, E., P. E. Clavijo, R. Davis, H. Cash, C. Van Waes, Y. Kim, and C. Allen.

2016. Established T Cell-Inflamed Tumors Rejected after Adaptive Resistance Was

Reversed by Combination STING Activation and PD-1 Pathway Blockade. Cancer Immunol

Res 4: 1061-1071.

439. Foote, J. B., M. Kok, J. M. Leatherman, T. D. Armstrong, B. C. Marcinkowski, L. S.

Ojalvo, D. B. Kanne, E. M. Jaffee, T. W. Dubensky, Jr., and L. A. Emens. 2017. A STING

Agonist Given with OX40 Receptor and PD-L1 Modulators Primes Immunity and Reduces

Tumor Growth in Tolerized Mice. Cancer Immunol Res 5: 468-479.

440. Sallets, A., S. Robinson, A. Kardosh, and R. Levy. 2018. Enhancing immunotherapy of STING agonist for lymphoma in preclinical models. Blood Adv 2: 2230-2241.

441. Temizoz, B., E. Kuroda, K. Ohata, N. Jounai, K. Ozasa, K. Kobiyama, T. Aoshi, and

K. J. Ishii. 2015. TLR9 and STING agonists synergistically induce innate and adaptive type-

II IFN. Eur J Immunol 45: 1159-1169.

442. Baird, J. R., D. Friedman, B. Cottam, T. W. Dubensky, Jr., D. B. Kanne, S. Bambina,

K. Bahjat, M. R. Crittenden, and M. J. Gough. 2016. Radiotherapy Combined with Novel

STING-Targeting Oligonucleotides Results in Regression of Established Tumors. Cancer

Res 76: 50-61.

443. Francica, B. J., A. Ghasemzadeh, A. L. Desbien, D. Theodros, K. E. Sivick, G. L.

Reiner, L. Hix Glickman, A. E. Marciscano, A. B. Sharabi, M. L. Leong, S. M. McWhirter, T.

W. Dubensky, Jr., D. M. Pardoll, and C. G. Drake. 2018. TNFalpha and Radioresistant

208

Stromal Cells Are Essential for Therapeutic Efficacy of Cyclic Dinucleotide STING Agonists

in Nonimmunogenic Tumors. Cancer Immunol Res 6: 422-433.

444. Foster, B. A., J. R. Gingrich, E. D. Kwon, C. Madias, and N. M. Greenberg. 1997.

Characterization of prostatic epithelial cell lines derived from transgenic adenocarcinoma of

the mouse prostate (TRAMP) model. Cancer Res 57: 3325-3330.

445. Boj, S. F., C. I. Hwang, L. A. Baker, Chio, II, D. D. Engle, V. Corbo, M. Jager, M.

Ponz-Sarvise, H. Tiriac, M. S. Spector, A. Gracanin, T. Oni, K. H. Yu, R. van Boxtel, M.

Huch, K. D. Rivera, J. P. Wilson, M. E. Feigin, D. Ohlund, A. Handly-Santana, C. M. Ardito-

Abraham, M. Ludwig, E. Elyada, B. Alagesan, G. Biffi, G. N. Yordanov, B. Delcuze, B.

Creighton, K. Wright, Y. Park, F. H. Morsink, I. Q. Molenaar, I. H. Borel Rinkes, E. Cuppen,

Y. Hao, Y. Jin, I. J. Nijman, C. Iacobuzio-Donahue, S. D. Leach, D. J. Pappin, M. Hammell,

D. S. Klimstra, O. Basturk, R. H. Hruban, G. J. Offerhaus, R. G. Vries, H. Clevers, and D. A.

Tuveson. 2015. Organoid models of human and mouse ductal pancreatic cancer. Cell 160:

324-338.

446. Quah, B. J., and C. R. Parish. 2012. New and improved methods for measuring lymphocyte proliferation in vitro and in vivo using CFSE-like fluorescent dyes. J Immunol

Methods 379: 1-14.

447. Gabrusiewicz, K., B. Rodriguez, J. Wei, Y. Hashimoto, L. M. Healy, S. N. Maiti, G.

Thomas, S. Zhou, Q. Wang, A. Elakkad, B. D. Liebelt, N. K. Yaghi, R. Ezhilarasan, N.

Huang, J. S. Weinberg, S. S. Prabhu, G. Rao, R. Sawaya, L. A. Langford, J. M. Bruner, G.

N. Fuller, A. Bar-Or, W. Li, R. R. Colen, M. A. Curran, K. P. Bhat, J. P. Antel, L. J. Cooper,

E. P. Sulman, and A. B. Heimberger. 2016. Glioblastoma-infiltrated innate immune cells

resemble M0 macrophage phenotype. JCI Insight 1.

448. Kwon, E. D., C. G. Drake, H. I. Scher, K. Fizazi, A. Bossi, A. J. M. van den Eertwegh,

M. Krainer, N. Houede, R. Santos, H. Mahammedi, S. Ng, M. Maio, F. A. Franke, S. Sundar,

209

N. Agarwal, A. M. Bergman, T. E. Ciuleanu, E. Korbenfeld, L. Sengeløv, S. Hansen, C.

Logothetis, T. M. Beer, M. B. McHenry, P. Gagnier, D. Liu, and W. R. Gerritsen. Ipilimumab

versus placebo after radiotherapy in patients with metastatic castration-resistant prostate

cancer that had progressed after docetaxel chemotherapy (CA184-043): a multicentre, randomised, double-blind, phase 3 trial. The Lancet Oncology 15: 700-712.

449. Topalian , S. L., F. S. Hodi , J. R. Brahmer , S. N. Gettinger , D. C. Smith , D. F.

McDermott , J. D. Powderly , R. D. Carvajal , J. A. Sosman , M. B. Atkins , P. D. Leming , D.

R. Spigel , S. J. Antonia , L. Horn , C. G. Drake , D. M. Pardoll , L. Chen , W. H. Sharfman ,

R. A. Anders , J. M. Taube , T. L. McMiller , H. Xu , A. J. Korman , M. Jure-Kunkel , S.

Agrawal , D. McDonald , G. D. Kollia , A. Gupta , J. M. Wigginton , and M. Sznol 2012.

Safety, Activity, and Immune Correlates of Anti–PD-1 Antibody in Cancer. New England

Journal of Medicine 366: 2443-2454.

450. Ritch, C. R., and M. S. Cookson. 2016. Advances in the management of castration resistant prostate cancer. BMJ (Clinical research ed.) 355: i4405.

451. Gabrilovich, D. I., S. Ostrand-Rosenberg, and V. Bronte. 2012. Coordinated regulation of myeloid cells by tumours. Nature reviews. Immunology 12: 253-268.

452. Curran, M. A., W. Montalvo, H. Yagita, and J. P. Allison. 2010. PD-1 and CTLA-4 combination blockade expands infiltrating T cells and reduces regulatory T and myeloid cells within B16 melanoma tumors. Proceedings of the National Academy of Sciences of the

United States of America 107: 4275-4280.

453. Wolchok, J. D., H. Kluger, M. K. Callahan, M. A. Postow, N. A. Rizvi, A. M. Lesokhin,

N. H. Segal, C. E. Ariyan, R. A. Gordon, K. Reed, M. M. Burke, A. Caldwell, S. A.

Kronenberg, B. U. Agunwamba, X. Zhang, I. Lowy, H. D. Inzunza, W. Feely, C. E. Horak, Q.

Hong, A. J. Korman, J. M. Wigginton, A. Gupta, and M. Sznol. 2013. Nivolumab plus ipilimumab in advanced melanoma. N Engl J Med 369: 122-133.

210

454. Larkin, J., V. Chiarion-Sileni, R. Gonzalez, J. J. Grob, C. L. Cowey, C. D. Lao, D.

Schadendorf, R. Dummer, M. Smylie, P. Rutkowski, P. F. Ferrucci, A. Hill, J. Wagstaff, M. S.

Carlino, J. B. Haanen, M. Maio, I. Marquez-Rodas, G. A. McArthur, P. A. Ascierto, G. V.

Long, M. K. Callahan, M. A. Postow, K. Grossmann, M. Sznol, B. Dreno, L. Bastholt, A.

Yang, L. M. Rollin, C. Horak, F. S. Hodi, and J. D. Wolchok. 2015. Combined Nivolumab and

Ipilimumab or Monotherapy in Untreated Melanoma. N Engl J Med.

455. Curran, M. A., M. Kim, W. Montalvo, A. Al-Shamkhani, and J. P. Allison. 2011.

Combination CTLA-4 blockade and 4-1BB activation enhances tumor rejection by increasing

T-cell infiltration, proliferation, and cytokine production. PLoS One 6: e19499.

456. Maraskovsky, E., K. Brasel, M. Teepe, E. R. Roux, S. D. Lyman, K. Shortman, and

H. J. McKenna. 1996. Dramatic increase in the numbers of functionally mature dendritic cells in Flt3 ligand-treated mice: multiple dendritic cell subpopulations identified. J Exp Med

184: 1953-1962.

457. Salmon, H., J. Idoyaga, A. Rahman, M. Leboeuf, R. Remark, S. Jordan, M.

Casanova-Acebes, M. Khudoynazarova, J. Agudo, N. Tung, S. Chakarov, C. Rivera, B.

Hogstad, M. Bosenberg, D. Hashimoto, S. Gnjatic, N. Bhardwaj, A. K. Palucka, B. D. Brown,

J. Brody, F. Ginhoux, and M. Merad. 2016. Expansion and Activation of CD103(+) Dendritic

Cell Progenitors at the Tumor Site Enhances Tumor Responses to Therapeutic PD-L1 and

BRAF Inhibition. Immunity 44: 924-938.

458. Klein, O., L. M. Ebert, D. Zanker, K. Woods, B. S. Tan, J. Fucikova, A. Behren, I. D.

Davis, E. Maraskovsky, W. Chen, and J. Cebon. 2013. Flt3 ligand expands CD4+ FoxP3+ regulatory T cells in human subjects. European journal of immunology 43: 533-539.

459. Bartkowiak, T., and M. A. Curran. 2015. 4-1BB Agonists: Multi-Potent Potentiators of

Tumor Immunity. Frontiers in oncology 5: 117.

460. Curran, M. A., and J. P. Allison. 2009. Tumor Vaccines Expressing Flt3 Ligand

211

Synergize with CTLA-4 Blockade to Reject Preimplanted Tumors. Cancer Res 69: 7747-

7755.

461. Boutros, C., A. Tarhini, E. Routier, O. Lambotte, F. L. Ladurie, F. Carbonnel, H.

Izzeddine, A. Marabelle, S. Champiat, A. Berdelou, E. Lanoy, M. Texier, C. Libenciuc, A. M.

Eggermont, J. C. Soria, C. Mateus, and C. Robert. 2016. Safety profiles of anti-CTLA-4 and anti-PD-1 antibodies alone and in combination. Nature reviews. Clinical oncology 13: 473-

486.

462. Segal, N. H., T. F. Logan, F. S. Hodi, D. F. McDermott, I. Melero, O. Hamid, H.

Schmidt, C. Robert, V. Chiarion-Sileni, P. A. Ascierto, M. Maio, W. J. Urba, T. C.

Gangadhar, S. Suryawanshi, J. Neely, M. Jure-Kunkel, S. Krishnan, H. E. Kohrt, M. Sznol, and R. Levy. 2016. Results From an Integrated Safety Analysis of Urelumab, an Agonist

Anti-CD137 Monoclonal Antibody. Clin Cancer Res.

463. Hodi, F. S., S. J. O'Day, D. F. McDermott, R. W. Weber, J. A. Sosman, J. B. Haanen,

R. Gonzalez, C. Robert, D. Schadendorf, J. C. Hassel, W. Akerley, A. J. M. van den

Eertwegh, J. Lutzky, P. Lorigan, J. M. Vaubel, G. P. Linette, D. Hogg, C. H. Ottensmeier, C.

Lebbé, C. Peschel, I. Quirt, J. I. Clark, J. D. Wolchok, J. S. Weber, J. Tian, M. J. Yellin, G.

M. Nichol, A. Hoos, and W. J. Urba. 2010. Improved Survival with Ipilimumab in Patients with Metastatic Melanoma. New England Journal of Medicine 363: 711-723.

464. Topalian, S. L., F. S. Hodi, J. R. Brahmer, S. N. Gettinger, D. C. Smith, D. F.

McDermott, J. D. Powderly, R. D. Carvajal, J. A. Sosman, M. B. Atkins, P. D. Leming, D. R.

Spigel, S. J. Antonia, L. Horn, C. G. Drake, D. M. Pardoll, L. Chen, W. H. Sharfman, R. A.

Anders, J. M. Taube, T. L. McMiller, H. Xu, A. J. Korman, M. Jure-Kunkel, S. Agrawal, D.

McDonald, G. D. Kollia, A. Gupta, J. M. Wigginton, and M. Sznol. 2012. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. The New England journal of medicine

366: 2443-2454.

212

465. Fransen, M. F., T. C. van der Sluis, F. Ossendorp, R. Arens, and C. J. Melief. 2013.

Controlled local delivery of CTLA-4 blocking antibody induces CD8+ T-cell-dependent tumor eradication and decreases risk of toxic side effects. Clinical cancer research : an official journal of the American Association for Cancer Research 19: 5381-5389.

466. Marabelle, A., H. Kohrt, I. Sagiv-Barfi, B. Ajami, R. C. Axtell, G. Zhou, R. Rajapaksa,

M. R. Green, J. Torchia, J. Brody, R. Luong, M. D. Rosenblum, L. Steinman, H. I. Levitsky,

V. Tse, and R. Levy. 2013. Depleting tumor-specific Tregs at a single site eradicates disseminated tumors. J Clin Invest 123: 2447-2463.

467. Simpson, T. R., F. Li, W. Montalvo-Ortiz, M. A. Sepulveda, K. Bergerhoff, F. Arce, C.

Roddie, J. Y. Henry, H. Yagita, J. D. Wolchok, K. S. Peggs, J. V. Ravetch, J. P. Allison, and

S. A. Quezada. 2013. Fc-dependent depletion of tumor-infiltrating regulatory T cells co- defines the efficacy of anti-CTLA-4 therapy against melanoma. J Exp Med 210: 1695-1710.

468. Fasso, M., R. Waitz, Y. Hou, T. Rim, N. M. Greenberg, N. Shastri, L. Fong, and J. P.

Allison. 2008. SPAS-1 (stimulator of prostatic adenocarcinoma-specific T cells)/SH3GLB2: A prostate tumor antigen identified by CTLA-4 blockade. Proc Natl Acad Sci U S A 105: 3509-

3514.

469. Patnaik, A., S. P. Kang, D. Rasco, K. P. Papadopoulos, J. Elassaiss-Schaap, M.

Beeram, R. Drengler, C. Chen, L. Smith, G. Espino, K. Gergich, L. Delgado, A. Daud, J. A.

Lindia, X. N. Li, R. H. Pierce, J. H. Yearley, D. Wu, O. Laterza, M. Lehnert, R. Iannone, and

A. W. Tolcher. 2015. Phase I Study of Pembrolizumab (MK-3475; Anti-PD-1 Monoclonal

Antibody) in Patients with Advanced Solid Tumors. Clinical cancer research : an official journal of the American Association for Cancer Research 21: 4286-4293.

470. Gabrilovich, D. I., S. Ostrand-Rosenberg, and V. Bronte. 2012. Coordinated regulation of myeloid cells by tumours. Nature reviews. Immunology 12: 253-268.

471. Downey, C. M., M. Aghaei, R. A. Schwendener, and F. R. Jirik. 2014. DMXAA

213

causes tumor site-specific vascular disruption in murine non-small cell lung cancer, and like the endogenous non-canonical cyclic dinucleotide STING agonist, 2'3'-cGAMP, induces M2

macrophage repolarization. PloS one 9: e99988-e99988.

472. McWhorter, F. Y., T. Wang, P. Nguyen, T. Chung, and W. F. Liu. 2013. Modulation of

macrophage phenotype by cell shape. Proceedings of the National Academy of Sciences of

the United States of America 110: 17253-17258.

473. Pello, O. M., M. De Pizzol, M. Mirolo, L. Soucek, L. Zammataro, A. Amabile, A. Doni,

M. Nebuloni, L. B. Swigart, G. I. Evan, A. Mantovani, and M. Locati. 2012. Role of c-MYC in

alternative activation of human macrophages and tumor-associated macrophage biology.

Blood 119: 411-421.

474. Pello, O. M., R. Chevre, D. Laoui, A. De Juan, F. Lolo, M. J. Andres-Manzano, M.

Serrano, J. A. Van Ginderachter, and V. Andres. 2012. In vivo inhibition of c-MYC in myeloid

cells impairs tumor-associated macrophage maturation and pro-tumoral activities. PLoS One

7: e45399.

475. Liu, L., Y. Lu, J. Martinez, Y. Bi, G. Lian, T. Wang, S. Milasta, J. Wang, M. Yang, G.

Liu, D. R. Green, and R. Wang. 2016. Proinflammatory signal suppresses proliferation and

shifts macrophage metabolism from Myc-dependent to HIF1α-dependent.

Proceedings of the National Academy of Sciences 113: 1564-1569.

476. Cullis, J., D. Siolas, A. Avanzi, S. Barui, A. Maitra, and D. Bar-Sagi. 2017.

Macropinocytosis of Nab-paclitaxel Drives Macrophage Activation in Pancreatic Cancer.

Cancer Immunol Res 5: 182-190.

477. Di Caro, G., N. Cortese, G. F. Castino, F. Grizzi, F. Gavazzi, C. Ridolfi, G. Capretti,

R. Mineri, J. Todoric, A. Zerbi, P. Allavena, A. Mantovani, and F. Marchesi. 2016. Dual

prognostic significance of tumour-associated macrophages in human pancreatic

adenocarcinoma treated or untreated with chemotherapy. Gut 65: 1710-1720.

214

478. Deshmukh, S. K., N. Tyagi, M. A. Khan, S. K. Srivastava, A. Al-Ghadhban, K.

Dugger, J. E. Carter, S. Singh, and A. P. Singh. 2018. Gemcitabine treatment promotes immunosuppressive microenvironment in pancreatic tumors by supporting the infiltration, growth, and polarization of macrophages. Scientific Reports 8: 12000.

479. Lemos, H., E. Mohamed, L. Huang, R. Ou, G. Pacholczyk, A. S. Arbab, D. Munn, and A. L. Mellor. 2016. STING Promotes the Growth of Tumors Characterized by Low

Antigenicity via IDO Activation. Cancer Res 76: 2076-2081.

480. Harrington, K. J., R. H. Andtbacka, F. Collichio, G. Downey, L. Chen, Z. Szabo, and

H. L. Kaufman. 2016. Efficacy and safety of talimogene laherparepvec versus granulocyte- macrophage colony-stimulating factor in patients with stage IIIB/C and IVM1a melanoma: subanalysis of the Phase III OPTiM trial. Onco Targets Ther 9: 7081-7093.

481. Andtbacka, R. H., H. L. Kaufman, F. Collichio, T. Amatruda, N. Senzer, J. Chesney,

K. A. Delman, L. E. Spitler, I. Puzanov, S. S. Agarwala, M. Milhem, L. Cranmer, B. Curti, K.

Lewis, M. Ross, T. Guthrie, G. P. Linette, G. A. Daniels, K. Harrington, M. R. Middleton, W.

H. Miller, Jr., J. S. Zager, Y. Ye, B. Yao, A. Li, S. Doleman, A. VanderWalde, J. Gansert, and

R. S. Coffin. 2015. Talimogene Laherparepvec Improves Durable Response Rate in

Patients With Advanced Melanoma. J Clin Oncol 33: 2780-2788.

482. Marabelle, A., L. Tselikas, T. de Baere, and R. Houot. 2017. Intratumoral immunotherapy: using the tumor as the remedy. Ann Oncol 28: xii33-xii43.

483. Smith, T. T., H. F. Moffett, S. B. Stephan, C. F. Opel, A. G. Dumigan, X. Jiang, V. G.

Pillarisetty, S. P. S. Pillai, K. D. Wittrup, and M. T. Stephan. 2017. Biopolymers codelivering engineered T cells and STING agonists can eliminate heterogeneous tumors. J Clin Invest

127: 2176-2191.

484. Leach, D. G., N. Dharmaraj, S. L. Piotrowski, T. L. Lopez-Silva, Y. L. Lei, A. G.

Sikora, S. Young, and J. D. Hartgerink. 2018. STINGel: Controlled release of a cyclic

215

dinucleotide for enhanced cancer immunotherapy. Biomaterials 163: 67-75.

485. Park, C. G., C. A. Hartl, D. Schmid, E. M. Carmona, H. J. Kim, and M. S. Goldberg.

2018. Extended release of perioperative immunotherapy prevents tumor recurrence and eliminates metastases. Sci Transl Med 10.

216

Vita

Casey Roy Ager was born the son of Jeffrey and Leanne Ager. After growing up in Spokane,

Washington and graduating from Ferris High School in 2009, he received his Bachelor of

Sciences with a major in Biochemistry from the University of Washington in March, 2013.

During his undergraduate studies, Casey performed research in the David Baker lab as a member of the 2011 iGEM team at UW, and additionally worked in the Rainer Storb lab at

Fred Hutchinson Cancer Center. Casey joined The University of Texas MD Anderson Cancer

Center UTHealth Graduate School of Biomedical Sciences in August of 2013.

Permanent address:

4712 E. 40th Ct.

Spokane, WA 99223

217