MECHANISM OF MYELOID-DERIVED SUPPRESSOR CELL
ACCUMULATION IN CANCER AND SUSCEPTIBILITY TO REVERSAL
BY SUNITINIB
by
JENNIFER SUSAN KO, M.D.
Submitted in partial fulfillment of the requirements
For the degree of Doctor of Philosophy
Thesis Adviser: Dr. James H. Finke
Department of Pathology
CASE WESTERN RESERVE UNIVERSITY
January, 2010
CASE WESTERN RESERVE UNIVERSITY
SCHOOL OF GRADUATE STUDIES
We hereby approve the thesis of
Jennifer Susan Ko
candidate for the Doctor of Philosophy degree*.
(signed) Alan Levine Ph.D.
David Kaplan M.D., Ph.D.
Clark Distelhorst M.D.
James Finke Ph.D.
Charles Tannenbaum Ph.D.
(date) October 12th, 2009
*We also certify that written approval has been obtained for any proprietary material
contained therein.
2 TABLE OF CONTENTS
Title Page 1
Signature Sheet 2
Table of Contents 3
List of Tables 6
List of Figures 7
Acknowledgements 9
List of Abbreviations 10
Abstract 14
Chapter 1: Introduction 16
Overview: Myeloid-derived suppressor cells in cancer: a novel therapeutic target. 16
Immunotherapy in cancer 16
Myeloid-derived suppressor cells limit immunotherapy 22
Myeloid-derived suppressor cells limit anti-angiogenic therapy 28
Multiple factors are implicated in MDSC formation 30
Vascular Endothelial Growth Factor 30
Stem Cell Factor 32
Granulocyte- and Granulocyte/Monocyte Colony Stimulating Factors 33
S100A9 and Inflammation 34
Intracellular signaling implicated in MDSC programming 36
3 Chapter 2: Sunitinib Mediates Reversal of Myeloid-Derived Suppressor Cell Accumulation in Renal Cell Carcinoma Patients 44
Statement of Clinical Relevance 44
Abstract 45
Introduction 46
Materials and Methods 48
Results 53
Discussion 71
Chapter 3: Direct and Differential Suppression of Myeloid-derived Suppressor Cell Subsets by Sunitinib is Compartmentally Constrained 75
Abstract 75
Introduction 76
Materials and Methods 77
Results 81
Discussion 97
Chapter 4: Tumor-Derived Products Dynamically Activate CD15+ Myeloid-Derived Suppressor Cells and are Partially Modulated by Sunitinib 102
Abstract 102
Introduction 103
Materials and Methods 106
Results 111
Discussion 129
Chapter 5: Dissertation Discussion and Future Directions 134
Rationale 134
4 Summary of Critical Findings 135
Future Directions 144
Sunitinib’s Impact on MDSC Expansion 144
The Role of GM-CSF and Gangliosides in MDSC Activation and the Impact of Sunitinib on MDSC Activation 146
Sunitinib’s Impact on MDSC Viability 150
Further Defining the RTK Targets of Sunitinib in MDSC Inhibition 156
Regarding Intratumoral Resistance to Sunitinib 159
Bibliography 163
5 LIST OF TABLES
TABLE PAGE
1-1 MDSC Phenotype and Subsets 25
1-2 MDSC Suppressive Mechanisms 27
2-1 Sunitinib Study Patient Characteristics 54
2-2 Sunitinib-induced Alterations in White Blood Cell Parameters 59
3-1 Characteristics of RCC Patients for Tumor Analysis 95
6 LIST OF FIGURES
FIGURE PAGE
2-1 Elevated MDSC in mRCC patients decline in response to sunitinib 57
2-2 Sunitinib mediated normalization of MDSC is associated with sunitinib-mediated enhancement in T cell production of IFNγ in mRCC patients 61
2-3 In vitro depletion of MDSCs restores patient T cell production of IFN-γ 63
2-4 Effect of sunitinib on MDSC suppressive function in vitro 65
2-5 Effect of sunitinib on MDSC viability and differentiation in vitro 68
2-6 Patient MDSCs correlate with Tregs in response to sunitinib treatment 70
3-1 Sunitinib reverses immune suppression in tumor-bearing hosts 82
3-2 Anti-proliferative and pro-apoptotic effects of sunitinib on MDSC subsets in the spleen 85
3-3 MDSC in 4T1 tumor bed are relatively protected from sunitinib- Mediated downregulation 88
3-4 GM-CSF is unique in its ability to protect MDSC in the presence of sunitinib, possibly via Stat3 repression 91
3-5 Relatively high levels of GM-CSF are unique to tumor microenvironment 93
7 3-6 Tumor microenvironment limits local anti-MDSC effect of sunitinib in RCC patients 96
4-1 N-MDSC comprise the predominant MDSC subset within RCC tissue 113
4-2 Tumor conditioned media (TCM) from RCC lines induce N-MDSC from whole blood 116
4-3 N-MDSC induced from whole blood by TCM express markers of activated neutrophils 119
4-4 Tumor conditioned media prolongs survival of cultured N-MDSC 121
4-5 GM-CSF reproduces effect of TCM on N-MDSC activation and viability 123
4-6 Sunitinib treatment of tumor cells causes a partial reduction in the ability of TCM to induce N-MDSC 125
4-7 RCC derived gangliosides induce N-MDSC and sunitinib reduced expression of gangliosides 128
5-1 Proposed model of MDSC accumulation and sunitinib-mediated MDSC reversal 143
8 ACKNOWLEDGEMENTS
The PhD process has perhaps been one of the most difficult in my life up until this point, but its success was highly attributable to several individuals beyond myself. For their patience, understanding, and unconditional love, I thank my husband, Tim, my daughter,
Eva, and my parents and brothers. Life could have been easier, but thank you for understanding why I took the road less traveled by medical doctors. For their encouragement, inspiration, and teaching, I thank Dr. Finke, Dr. Borden, and Dr. Cohen.
Thank you for seeing the potential in me, and my project, especially at times when I leant toward self doubt. I especially thank Dr. Finke for the intellectual freedom he was brave enough to bestow me with from day one. For their support, friendship, tolerance, and technical guidance I thank the members of the Finke lab – Pat, Joanna, Kaushik,
Soumika, Cyndi, and Li, and other past and rotating members. Your patience and good humor helped make all the learning fun. I especially thank Pat and Joanna, two people who know no technical or organizational obstacle, even when it comes to patient studies with clinical samples. Thank you to Drs. Levine, Kaplan, Tannenbaum, Distelhorst, and
Hamlin for your intellectual discussions and support of this project. I especially thank
Dr. Levine, for always being available to me as a teacher and as a chairman. Finally, I thank God, who created all this beauty and wonder which sustains me with the excitement of a child, even as I continue to age.
9 LIST OF ABBREVIATIONS
7AAD – 7 aminoactinomycin D
ACT – adoptive cellular therapy
AECCM – activated endothelial cell conditioned medium
Ag – antigen
AMN – age-matched normals
APC – antigen presenting cell
ARG1 – arginase 1
ATRA – all-trans retinoic acid
ATT – adoptive T cell therapy
Bcl-2 – B-cell lymphoma 2
Bcl-xl – B-cell lymphoma – extra large
BCR/abl–breakpoint cluster region/Abelson murine leukemia viral oncogene homolog1
BM – bone marrow
CD – cellular differentiation
CD40L – CD40 ligand
COX2 – cyclooxygenase 2
CSF-1R – colony stimulating factor-1 receptor
CT – computed tomography
DC – dendritic cell
EGOG – Eastern Cooperative Oncology Group
FACS – fluorescence-activated cell sorting
G-CSF – granulocyte colony stimulating factor
10 GFP – green fluorescent protein
GM-CSF – granulocyte/monocytes colony stimulating factor
HLA – human leukocyte antigen
HPC – hematopoietic progenitor cell
HPLC – high-performance liquid chromotography
IFNα – interferon-alpha
IFNγ – interferon-gamma
Ig - immunoglobulin
IL – interleukin
IRB – Institutional Review Board
JAK – Janus kinase
KO – knock out
M-CSFR – macrophage colony stimulating factor receptor
MDSC – myeloid-derived suppressor cells
MHC – major histocompatibility complex m-MDSC – monocytic myeloid-derived suppressor cells mMM – metastatic malignant melanoma
MMP – matrix metalloproteinase mRCC – metastatic renal cell carcinoma
MRP – myeloid related protein
MSKCC – Memorial Sloan-Kettering Cancer Center
NAC – N-acetylcysteine
NADPH – nicotinamide adenine dinucleotide phosphate
11 NK – natural killer n-MDSC – neutrophilic myeloid-derived suppressor cells
NO – nitric oxide
NOS2 – nitric oxide synthase 2
PBMC – peripheral blood mononuclear cells
PDGFR – platelet-derived growth factor receptor
PGE2 – prostaglandin E2
RAG- recombination activating gene
RCC – renal cell carcinoma
RECIST – response evaluation criteria in solid tumors
ROS – reactive oxygen species
RTK – receptor tyrosine kinase
RTKI – receptor tyrosine kinase inhibitor
SCF – stem cell factor
SOCS – suppressor of cytokine signaling
STAT – signal transducer and activator of transcription
TAA – tumor-associated antigen
TCR – T cell receptor
TCM – tumor-conditioned media
TGFβ – transforming growth factor- beta
TKI – tyrosine kinase inhibitor
TLR – toll-like receptor
TNFα- tumor necrosis factor-alpha
12 Treg – regulatory T cells
VEGF – vascular endothelial growth factor
VEGFR – vascular endothelial growth factor receptor
VHL – Von Hippel-Lindau
13 Mechanism of Myeloid-Derived Suppressor Cell Accumulation in Cancer and
Susceptibility to Reversal by Sunitinib
ABSTRACT
by
Jennifer Susan Ko, M.D.
Tumor-driven accumulation of myeloid-derived suppressor cells (MDSC) facilitates tumor immune evasion via T-cell inhibition, therefore limiting therapeutic approaches. MDSC accumulate in tumor-bearing hosts via several factors, and suppress type-1 T-cell function via multiple mechanisms. Elevated MDSC in the blood of renal cell carcinoma patients, were first shown to significantly decline following treatment with sunitinib (inhibits VEGFr, ckit, flt3, PDGFr) treatment. This decline was correlated with a recovery in patients’ T-cell function, an effect which could be reproduced with in vitro
MDSC depletion. Sunitinib induced MDSC apoptosis in vitro. Sunitinib-mediated declines in MDSC occurred even in non-responder patients and were not correlated with tumor shrinkage.
Studies in several murine tumor models confirmed sunitinib’s ability to universally inhibit peripheral MDSC accumulation and restore normal splenic T-cell function, even when sunitinib had negligible impacts on tumor progression. Sunitinib directly inhibited pathologic expansion of mononuclear m-MDSC, and directly reduced the viability of neutrophilic n-MDSC- which was pathologically extended in the tumor- bearing state. MDSC in bone marrow and tumor proved resistant to sunitinib however, and GM-CSF availability predicted sunitinib resistance. Recombinant GM-CSF could confer sunitinib resistance to MDSC in vitro. MDSC conditioning with GM-CSF
14 inhibited STAT3 and promoted STAT5 activation, whereas other hematopoietic support
factors preferentially preserved STAT3 activation and sunitinib susceptibility.
In vitro activation of MDSC from healthy human whole blood with tumor-
conditioned media (TCM), further implicated GM-CSF, as well as gangliosides, in
MDSC activation and survival. Sunitinib could indirectly inhibit MDSC activation, via
drug-mediated reductions in ganglioside production by tumor cells. The phenotypic and
functional resemblance of TCM-induced MDSC to mRCC patient MDSC indicated that this model of n-MDSC activation from a subset of normal neutrophils may be a contributing mechanism to MDSC accumulation in tumor-bearing hosts.
Sunitinib’s anti-MDSC activities at each of three steps were thus described.
Increased MDSC expansion, activation, and viability all proved to contribute to MDSC
accumulation. The drug’s ability to act at each of these three steps was limited in the
presence of GM-CSF. Ancillary strategies to build upon sunitinib’s potency both as an
immunomodulator and as a cancer therapy will be further investigated in future studies.
15 CHAPTER 1
INTRODUCTION
Overview: Myeloid-derived suppressor cells in cancer: a novel therapeutic target.
An extensive amount of research worldwide is focused on finding new
approaches to the treatment of solid tumors. Diseases such as metastatic renal cell
carcinoma (mRCC) and metastatic malignant melanoma (mMM) are known to be
resistant to traditional chemotherapy and radiotherapy (1-3). A novel class of drugs,
which effectively function to inhibit the angiogenesis required for tumor growth have
become first-line treatment for mRCC, and are being tested in mMM (4-6). Despite the
initial and significant therapeutic activity of the anti-angiogenic tyrosine kinase inhibitors
(TKI) (ie sunitinib), all treated patients eventually develop progressive disease. Indeed,
historic observations support the idea that the immune system may present the most
viable option for curative therapy, due to the potential for durable responses in this
setting (7-9). However, recent insights in the field of tumor immunology show that while
the immune system serves to keep early oncogenic processes at bay, it is co-opted by
advanced tumors to promote tumor growth and spread (10-12). Tumor-recruited immune
cells function both to suppress the anti-tumor function of cytotoxic effector T cells and to
promote tumor-dependent angiogenesis as well as tumor invasion and metastasis (13-17).
Bone marrow-derived myeloid cells are crucial in both of these tumor-promoting
pathways, and therefore should be regarded as an important therapeutic target.
Immunotherapy in cancer
Accumulating evidence shows that the immune system recognizes and can
eliminate transformed tumor cells (11, 18). In addition to preventing clinically apparent
16 cancer, the immune system may also be able to eradicate established tumors. This
concept is best supported by the emergence of adoptive cell therapy (ACT) with
melanoma tumor-infiltrating lymphocytes as the most effective treatment for patients
with mMM (19). Furthermore, Interleukin-2 (IL-2), a T cell dependent growth factor,
has historically been given to patients with metastatic renal tumors and melanomas, and
confers a durable, curative response in 5-10% of patients (20, 21). Interferon-alpha is
another immune cell-stimulating protein that is an FDA-approved drug for treatment of
RCC and MM, where it is standard of care for use in Stage III melanoma, and up until
recently, was a first-line treatment for patients with RCC. These agents are thought to act
in part to support the outgrowth and function of tumor-antigen specific T cells which
target and destroy microscopic or macroscopic tumor; however their precise mechanism
of action is not yet fully characterized. Several excellent reviews can provide a more in-
depth discussion of adoptive cellular therapy and cytokine therapy (22-26). It will be important in the future to determine what distinguishes patients who respond to immunotherapy from those who do not respond, and to improve upon current regimens with treatments used in combination.
Tumor-induced immune suppression
Tumor-immune cell interactions
In 1909 Ehrlich first postulated that the immune system provides its host with protection against cancer. Burnet and Thomas modified this concept in the 1950’s, and proposed that the immune system performed a “surveillance” function to eliminate precancerous or cancerous cells before they could become clinically apparent tumors.
While this concept was challenged in the 1980’s, it has since regained favor and been
17 expanded based on additional findings (27). Evidence supports that what is now called
“cancer immunoediting” occurs in three phases- elimination, equilibrium, and escape
(11). The initial phase of editing, elimination, involves the identification and elimination of transformed cells by T and NK cells primarily, and is supported by the fact that mice and humans with deficient or suppressed adaptive immune responses (RAG-2 KO mice,
IFNg-deficient mice, or immunosuppressed patients) display a greater incidence of tumor development (11, 12, 18, 27-29). During “equilibrium” tumor cells and immune cells constantly interact to induce reciprocal modifications in each other, and immune cells maintain tumors in a somewhat dormant state. Indeed, wild type mice that clinically resisted tumor challenge with low-dose methylcholanthrene after 200 days, rapidly developed sarcomas if depleted of CD4+ and CD8+ T cells. When these sarcomas were transplanted into naïve, WT mice, they displayed more aggressive features, due to the editing they’d undertaken during their “dormant” state under selective immune pressure
.(18). The final phase of immune editing is the “escape” phase, where tumors employ several mechanisms that culminate in immune evasion, and the ultimate outgrowth of clinically apparent cancer (11, 30). These mechanisms will be further discussed below.
The processes underlying the early phases of immune editing perhaps need be considered most when developing tumor vaccines or treatments to be used adjuctively to surgery in patients without evidence of residual disease. However, active immune therapy for clinically apparent, unresectable tumors requires a more in depth understanding of those mechanisms which provide tumors with continued protection from immune destruction.
In addition to more potent T cell stimulation, immunotherapy will need to include approaches which reverse the immunosuppressive networks put in play by tumor cells.
18 T cell dysfunction in tumor bearing hosts
Tumor growth occurs in the face of expanded numbers of anti-tumor-Ag-specific T
cells, thus a clear understanding of the tumor-driven immunosuppressive networks which
limit the anti-tumor response is essential (31, 32). Defects exist at several of the steps
required to mount an effective adaptive immune response. The binding of antigen, either
tumor-associated or otherwise, to an antigen-presenting cell (APC) and the presentation
of antigen to a cognate T cell receptor (TCR) is a crucial point in the initiation of an
immune response. It results in T cell activation and clonal expansion, and the nature of
this interaction determines the amount and type of T cell response that ensues (33). The
primary determinant of T cell responsiveness to tumor-associated antigen (TAA) is the
context of antigen presentation by bone marrow-derived APC- the most potent of which
are dendritic cells (DC), rather than the intrinsic antigen-presenting capacity of tumor
cells themselves (34). CD4 helper T cells are essential to ensure that TAA is presented properly by DC to cytotoxic T cells. Through cognate interactions with DC that lead to increased DC maturity and longevity, surface stabilization of peptide-MHC, increased costimulatory expression, and production of IL-12, CD4 T cells (primarily via CD40L)
have the capacity to determine whether the responding CD8 cytotoxic T cells will
undergo induction of tolerance or activation. This CD4 T cell help is especially
important in the context of cancer, where TLR stimulation may be lacking (35-37). That
said, defective antigen processing and presentation by dendritic cells have been observed
in mice and patients bearing several types of tumors, and this is described further in
reviews by several authors (38-46). This defect does not relate to a lack of tumor-
associated antigen expression by tumor cells, as this notion to explain the absence of a
19 robust anti-tumor immune response has been dispelled, and over 300 TAAs have been identified, along with their HLA class I and II restrictions (47), (48). An updated peptide database for T-cell-defined tumor antigens is also available at http://www.cancerimmunity/peptidedatabase/Tcell epitopes.htm.
Type-1 T cells secrete pro-inflammatory cytokines such as interferon-gamma
(IFNγ) that activate surrounding APC and tumor cells to present antigen more efficiently, and that lead to the formation of cytotoxic T cells which target tumor cells for destruction
(10, 28, 49). In addition, interferons can be cytotoxic or cytostatic to tumor cell growth, and have inhibitory effects on tumor-dependent angiogenesis (28, 49). It has been shown numerous times using various experimental approaches, that a type-1 cytotoxic T cell response, rather than a type-2 humoral T cell response, is necessary for immune-mediated tumor rejection (11, 28). Yet, for various reasons revolving around tumor cell adaptation through immune-editing, as described above, most clinically apparent tumors subvert the immune response such that T cells are dysfunctional, and selectively deficient in several of the activities necessary to generate a type-1 response- such as CD3ζ signaling, proliferation, and/or the production of IFNγ (50-53). Hence, even if deficiencies in dendritic cell function can be circumvented through clinical delivery of vaccines, TLR ligands, or stimulated dendritic cells themselves, activated T cells are likely to encounter several other obstacles hampering tumor cell destruction. These obstacles are further addressed below.
Mechanisms of Immune Suppression
Several tumor produced molecules are implicated in T cell suppression, including tumor-derived gangliosides (54, 55), transforming growth factor beta (TGFβ) (56-60),
20 and cyclooxygenase enzyme 2 (COX2) (61) and its downstream production of prostaglanding E2 (PGE2) (62-65). In addition, B7-H1 and B7-H4 molecules on tumor cells may interact directly with T cells to produce deleterious T cell effects (66-71).
Tumor-derived gangliosides
Gangliosides can be defined as sialic acid containing glycosphingolipids. There
are multiple species of gangliosides which are classified based on the number and
location of constituent sialic acid subgroups (72, 73). Physiological ganglioside species of primarily GM1 and GM3 are ubiquitous in normal animal cells, including immune cells. These gangliosides are located in the plasma membrane of cells where they localize specifically to lipid raft portions of membrane. Here they play a role in cell development, growth, and communication via the regulation of various signaling events
(74). Renal tumor cells produce excessive amounts of gangliosides (GM2, GM1 and
GD1a) which are likely shed into the tumor microenvironment (55, 75, 76). In cancer patients with melanoma and neuroblastoma, elevated levels of serum gangliosides have been shown to correlate with tumor growth (77, 78). No such data has yet been reported
for RCC though such correlative studies are underway.
There have been several immunosuppressive functions attributed to gangliosides.
Both tumor-derived and commercially available bovine brain-derived gangliosides are
reported to induce T cell apoptosis (79-81). Additionally, lower concentrations of
gangliosides are capable of suppressing T cell effector function. They have been shown to inhibit T cell proliferation in response to various mitogens through a mechanism that likely involves interference in IL-2 signaling (82). T cell production of IFN-γ has also
recently been shown to be selectively suppressed by gangliosides (55, 83). Furthermore,
21 in vitro maturation of dendritic cells is inhibited in the presence of gangliosides,
suggesting that select gangliosides can impair antigen presentation (84). The suppressive
activity of gangliosides is enabled by their ability to insert themselves into the
membranes of cells they come in contact with. Once integrated into target cell
membranes, they are proposed to alter or interfere with various signaling events
emanating from the cell surface, and to produce intracellular reactive oxygen species
which also modulate signaling events (85).
Myeloid-derived suppressor cells limit immunotherapy
Importantly, tumors also promote the accumulation of immune cells which are
suppressive to T effector cells. The most common of these include regulatory T cells
(Treg) and myeloid-derived suppressor cells (MDSC). Elevated levels of Treg are
detected in the peripheral blood of patients bearing tumors of several types. Their
elimination is thought to be crucial in breaking immune tolerance to self-antigens, and
thus clinical strategies aimed at their depletion are currently under investigation in cancer patients (86, 87). The molecules listed above, as well as regulatory T cells, are discussed thoroughly in several other excellent reviews (52, 87, 88). Elevated levels of MDSC have been reported in the blood of cancer patients bearing several types of tumors. Their removal from tumor-bearing mice leads to tumor regression (89), and so clinical methods which effectively deplete MDSC are desirable to improve the anti-tumor immune response and prolong patient survival. Based on work published by several others in addition to that presented in the chapters to follow, MDSC seem to represent a major contributor to immune suppression in cancer patients and tumor-bearing mice (15, 51, 89-
108). Thus a significant portion of the remaining introduction is devoted to describing
22 the phenotype and function of MDSC as well as the cytokines/growth factors important
for their accumulation in the tumor-bearing host.
Phenotype and subsets of MDSC
The accumulation of myeloid cells that suppress T cell activation in tumor-
bearing mice was first reported in the late 1980’s. In a series of studies it was found that
colony-stimulating factor activity produced by tumor cells increased hematopoiesis, and
resulted in the formation of suppressor cells in the bone marrow, followed by their appearance in the blood and spleen (109, 110). These cells resembled immature bone marrow-derived myeloid cells which could also appear during bone marrow recovery following cyclophosphamide treatment (111). Importantly, the suppressive activity of these cells had varying degrees of sensitivity to indomethacin (inhibits prostaglandins) and catalase (inhibits reactive oxygen species), yet those suppressive myeloid cells generated from mice with increasingly metastatic tumors became increasingly insensitive to indomethacin and catalase (109-111).
As shown in Table 1-1, accumulating evidence to date has shown MDSC to be a largely heterogeneous group of immature myeloid cells of macrocytic/monocytic, granulocytic, or dendritic cell lineage. In the mouse model, they are broadly defined as being CD11b+Gr1+ and are capable of suppressing antigen-specific or nonspecific T cell activation. The recent identification of MDSC subpopulations in the mouse model has shown that granulocytic/neutrophilic MDSC are CD11b+Gr1hiLy6G+Ly6Clo cells with
high side scatter that suppress T cell function via arginase activity and reactive oxygen
species production. Importantly, these cells could not be differentiated from normal
neutrophils phenotypically, except by virtue of their reduced density. Monocytic MDSC
23 were identified as CD11b+Gr1loLy6G-Ly6Chi cells which suppress T cells via arginase
activity and nitric oxide production. Their phenotypic appearance resembled that of
inflammatory monocytes (112, 113). A selective role for prostaglandins in the
suppression mediated by either of these subsets has not been defined. MDSC detected in
the peripheral blood of cancer patients have been quantified using three criteria, all under
the umbrella of being CD33+HLADR-, which likely identify MDSC subpopulations that
overlap. MDSC were first defined in patients with squamous cell carcinoma of the
head/neck/lung, or adenocarcinoma of the breast/lung, to be Lineage(CD3,14,19,56)-
CD33+HLADR-CD15- and to inhibit T cell function using an unclear mechanism (114).
These cells likely represent those in the mouse model that exist at early stages of
differentiation as precursor cells (Ly6G-Ly6C+). A CD11b+CD14-CD15+ MDSC subset has been shown to suppress T cell function in patients with kidney cancer, among others, through an arginase and/or reactive oxygen species-dependent mechanism (108, 115).
These cells likely represent those that are granulocytic/neutrophilic (Ly6G+Ly6Clo) in the mouse model. Finally, a newer subset of MDSC has been defined in patients with melanoma and hepatocellular carcinoma (HCC), and they are CD14+HLADR-/dim and suppress T cell function via TGFβ and/or arginase production, as well as the induction of
Treg formation (116, 117). These may parallel MDSC in the mouse model that are monocytic/macrocytic (Ly6G-F4/80+CD115+). Our own experience with MDSC subsets will be detailed in chapters that follow.
24
Table 1-1 MDSC – Phenotype and Subsets (Murine) / (Human)
Total MDSC N-MDSC M-MDSC
Origin* Immature or alternatively Immature or Immature or activated myeloid cells alternatively activated alternatively activated Neutrophils Monos/Macs
Markers CD11b+Gr1+ CD11b+Gr1+hi CD11b+Gr1+lo Ly6G- CD33+HLADR- Ly6G+F4/80- F4/80+/- CD15+CD14- CD15-CD14+DR-dim
Morphology* heterogeneous early or completed med w/ round nuclei lg polymorphonuclear w/ nephric shape
Prevalence/ Ratio Naïve BM- ~25-40% TB BM- 35-80% See Left N:M Naïve Spleen- ~2-3% TB Spleen- 8-50% Naïve LN- <.1% TB LN- .1-4% Naïve Spleen- 1:1 TB Spleen Ratio- 4:1 Naive Blood- <.5% TB Bld- 1.2-10% TB Bld Ratio- varies
* Denotes same in mice and humans
Suppressive pathways used by MDSC to impair T cell effector function
It has become apparent that MDSC inhibit T cells via multiple mechanisms which
may be somewhat dependent on tumor type, tumor burden, and the anatomical
compartment from which MDSC are derived. These mechanisms are outlined in Table 1-
2. Existing knowledge gaps regarding such differences make it difficult to set strict rules
about MDSC phenotype, however their suppressive mechanisms can largely be grouped
into L-arginine dependent and independent ones. MDSC rapidly metabolize L-arginine
via two main pathways- arginase 1 (ARG1) and/or nitric oxide synthase 2 (NOS2).
ARG1 activation occurs in response to Type 2 cytokines (IL-4, IL-13), TGFβ, GM-CSF, and PGE2 among others, via STAT6 dependent and independent pathways (118-123).
25 ARG1 upregulation leads to the depletion of environmental L-arginine, and consequently compromises CD3ζ chain expression and TCR-mediated T cell activation, proliferation,
and cytokine production (122-124). Zea et. al. confirmed the clinical relevance of this
mechanism of MDSC-mediated T cell suppression when they looked at peripheral blood
lymphocytes of 123 renal cell carcinoma patients. They found that patients had elevated
arginase levels which correlated with elevated ornithine levels and decreased expression
of CD3ζ, and decreased T cell function. Arginase activity was limited to CD11b+CD14-
CD15+ MDSC. Removal of this cell population restored T cell expression of CD3ζ and
T cell function (108).
Arginine metabolism by upregulated levels of NOS2 in MDSC occurs in response to prostaglandin E2, and inflammatory cytokines such as IFNγ, TNFα, or IL-1α. NOS2 activation is thought to interfere with downstream signaling molecules such as
JAK/STAT proteins, which are important in T cell function (99, 125, 126). When L- arginine is depleted by arginase, NOS2 is thought to produce superoxide and NO, which rapidly combine to form highly reactive peroxynitrites. Peroxynitrites can modify and inactivate proteins via tyrosine nitrosylation and can also induce cell death via mitochondrial damage (127). In this regard, ARG1 and NOS2 pathways can actually operate in parallel, and also synergize to inhibit T cell function. Other suppressive mechanisms used by MDSC, which are L-arginine independent, include L-tryptophan metabolism by indoleamine-2,3-dioxygenase, as well as the production of reactive oxygen species (ROS) described earlier. ROS production likely occurs via NADPH oxidase machinery possessed by all phagocytic cells (128).
26 Table 1-2 MDSC– Suppressive Mechanisms
Employing MDSC Immune Effects Molecular Mechanisms
ARG1 N- and M-MDSC • T cell proliferation • L-arg depletion (arginase 1) • CD3ζ, cyclinD3, and cdk4
ROS N- MDSC - - • T cell proliferation •O2 >> ONOO (see (reactive oxygen below) species) • T cell IFNγ prdxn. • CD3ζ iNOS M-MDSC - • T cell proliferation •NO >> ONOO (inducible nitric oxide production synthase) • T cell IFNγ prdxn. •Tyrosine nitrosylation • pMHC recog
Cytokines N- and M-MDSC •DC inhibition •NFkB inhibition, (VEGF, IL-10, TGFβ) STAT3 activation •T2 polarization •Treg formation
27 Myeloid-derived suppressor cells limit anti-angiogenic therapy
It is now widely accepted that tumor angiogenesis is a crucial step in tumor development (13). This has led to the approval of rationally designed, angiogenesis inhibitors for the treatment of several tumor types. Bevacizumab (Avastin,
Genentech/Roche), a vascular endothelial growth factor (VEGF) trapping antibody, has been FDA approved for the treatment of patients with late-stage colon cancer, non-small-
cell lung cancer and breast cancer, all in combination with chemotherapy. Sorafenib
(Nexavar, Bayer) and sunitinib (Sutent, Pfizer) are two receptor tyrosine kinase inhibitors
which target the VEGFR family of pro-angiogenic signaling pathways, and are approved
for treating renal cell carcinoma as well as hepatocellular carcinomas (sorafenib) and
gastrointestinal stromal tumor (sunitinib). There are currently numerous clinical trials
testing the use of these drugs in other tumor types, and many other angiogenesis
inhibitors are being clinically evaluated.
While bevacizumab, sunitinib, and sorafenib have achieved relative success in the
setting of aggressive metastatic tumors such as RCC, they have failed to cure patients,
and the clinical responses produced can be short-lived (129). This is because they
typically stabilize tumor growth in most patients and shrink tumors in some, however all
patients eventually progress after a time period usually measured in months (130). This
is thought to reflect an adaptive response to drug by tumors, which allows for “evasive
resistance” to angiogenesis inhibitors (131). Likewise, patients who never experience a
tangible benefit in response to drug are thought to already have existing “intrinsic
resistance”. Both of these resistance patterns are thought to occur via four primary
mechanisms which all result in the employment of alternative ways to sustain tumor
28 growth in the face of inhibition of the specific therapeutic target. Experimental evidence in both human patients and mice show that these alternatives include: the upregulation of substitiute pro-angiogenic signaling pathways, the recruitment of bone-marrow derived pro-angiogenic cells which negate the need for VEGF signaling, the increase in pericyte coverage and protection of vasculature, and finally, the activation of invasive and metastatic activities which provide access to normal tissue vasculature (13).
MDSC are one of the myeloid cell subtypes recruited by tumors that regulate tumor growth and angiogenesis. There is an extensive amount written about the role that neutrophils and macrophages play in angiogenesis, and some of these functions may overlap with that of MDSC (15, 132). However, several new studies have directly implicated MDSC in angiogenesis. When Yang et. al. co-injected MDSC along with tumor cells subcutaneously into mice, those tumors grew faster and had an increased blood vessel density (133). MDSC were later shown to differentiate into CD31+ cells which incorporate into newly forming endothelium, and the prevention of MDSC recruitment to tumors led to reduced tumor angiogenesis (134). MDSC secretion of matrix metalloproteinases (MMP), especially MMP9, is thought to increase the bioavailability of several angiogenic molecules, including VEGF. Along these lines, tumor refractoriness to anti-VEGF treatment was shown by Shojaei et. al. to be dependent on the recruitment of CD11b+Gr1+ myeloid cells to tumors. These cells apparently produce Bv8, or prokineticin-2, a homologue of endocrine-gland-derived
VEGF (EG-VEGF), which circumvents the need for VEGF in tumor-dependent angiogenesis (135). Finally, VEGFR1+ bone marrow-derived myeloid cells dictate tumor metastasis via the formation of distant permissive niches for incoming tumor cells
29 (14). Tangential evidence suggests that these cells may also be a subtype of MDSC (95).
Investigations further defining the role of MDSC in angiogenesis are needed. It will be particularly important in the future to determine whether part of MDSC-mediated angiogenesis enhancement is due to MDSC-mediated suppression of T cells or NK cells, and their production of interferon, as interferons are known to be anti-angiogenic.
Multiple factors are implicated in MDSC formation
Myeloid derived suppressor cells are thought to accumulate in tumor-bearing hosts as a result of alterations in myelopoiesis influenced by tumor-derived soluble factors (92, 99). Indeed, MDSC recruitment is directly correlated with tumor burden, and tumor resection decreases MDSC numbers (136-138). Implicated tumor-derived cytokines and growth factors can be partially shared between various tumor types, and one single factor has not been conclusively shown responsible. The most strongly implicated factors will be discussed.
VEGF
Vascular endothelial growth factor (VEGF) is a secreted growth factor produced by many tumors, and it is crucial in the formation and maintenance of blood vessels and blood cells. In addition to its well-characterized role in angiogenesis, VEGF inhibits dendritic cell differentiation, inhibits T cell formation, and most importantly, leads to the accumulation of phenotypic MDSC in naïve mice upon chronic infusion. VEGF infusion over 28 days was shown to prevent dendritic cell maturation in vivo, in favor of MDSC accumulation. In these same studies, VEGF or conditioned-medium from VEGF- activated endothelial cells (AECCM) acted similarly on “late” progenitor cells to enhance the growth of myeloid and erythroid progenitors. In contrast, the effects on “early”
30 progenitors and stem cells were different because recombinant VEGF had an inhibitory
effect on the total number of colonies, but increased the relative proportion of myeloid colonies and megakaryocytes; while, AECCM increased the total number of colonies formed, with little effect on their histological subtype. This suggests that VEGF may act directly on stem cells and early progenitor cells to direct them towards the myeloid lineage, but that other factors are likely needed, some of which are produced by endothelial cells in response to VEGF, to enhance and sustain the shear numbers of
MDSC seen to accumulate in the tumor-bearing host (139).
In a similar, follow-up study, VEGF infusion caused an increase in splenic and circulating, but not bone marrow, CD11b+Gr1+ cells (140). This accumulation, although modest compared to some mouse tumor models, was shown to be dependent upon signaling through VEGFR2 rather than VEGFR1. This second study was largely observational and failed to demonstrate which cells, endothelial and/or hematopoietic precursor cells, VEGF was acting on through VEGFR2 signaling. In addition, the question of whether the effects of VEGF were direct or indirect, via the induction of
another putative cytokine, was not addressed.
Indeed it has been shown by Larrivee et. al. that CD11b+Gr1+ cell accumulation
in response to VEGFR2 signaling is dependent on VEGFR2 induced increases in GM-
CSF. In these studies BM cells were virally transduced to express GFP and VEGFR2 that was subject to chemical activation, and infused into lethally irradiated mice. Results
showed that VEGFR2 activation induced an expansion only of transduced, GFP+
CD11b+Gr1+ cells, but not GFP-CD11b+Gr1+ cells, while it induced an expansion of
both GFP+ and GFP-, earlier-staged myeloid colony-forming progenitor cells. VEGFR2
31 signaling caused transduced cells to upregulate GM-CSF production (but not SCF, IL-6, notch ligands, flt3 ligand, or M-CSF) which affected non-transduced cells in a paracrine fashion. The authors thus concluded that even though GM-CSF can stimulate the proliferation of myeloid colony-forming cells, this cytokine alone is not sufficient to induce a marked expansion of CD11b+ and Gr-1+ cells, at least in the time frame of their experiments. By contrast, VEGFR-2 signaling by itself can induce expansion of differentiated cells, but cannot increase the number of myeloid colony-forming cells without the action of GM-CSF, so that only when VEGFR-2 signaling is combined with that of GM-CSF can a marked expansion of both progenitors and differentiated myeloid cells be observed (141).
Hence it is likely that VEGF does not act alone to stimulate myelopoeisis, but this has been difficult to determine because VEGF receptor expression on hematopoietic cells is either quite rare under steady-state conditions, or is technically difficult to detect. More probable is the possibility that VEGF may enhance and/or alter the production of colony- stimulating cytokines and growth factors, such as GM-CSF, by a limited number of receptor-positive cells present in hematopoietic organs, which directly stimulate myeloid cell formation. In addition, the vascular components of hematopoetic organs may require
VEGF support to upregulate their production of myeloid cells.
Stem cell factor
Stem cell factor (SCF) is a growth factor important for the survival, proliferation, and differentiation of hematopoietic stem cells and other hematopoietic progenitor cells.
It binds to the c-kit receptor (CD117) to induce signaling, and is thus also known as kit
32 ligand. Pan et. al. have shown that stem cell factor is highly expressed in several human
and murine tumors (102). Mice bearing SCF knockdown tumors or tumor-bearing mice
treated with anti-ckit blocking antibodies experienced a reduction in bone marrow MDSC
that were defined in this study as being restricted to cells falling in fraction II of a Percoll
density gradient and expressing the marker CD115. This blockade of stem cell factor
also led to a reduction in the suppressive function of these MDSC. Concurrent reductions
in Treg cell numbers and improvements in effector T cell function were also observed.
The authors did not, however, show whether MDSC numbers in the spleen were reduced,
and they did not show conclusively that the reductions in MDSC were directly
responsible for the improvements seen in T cell function. In addition, significant tumor-
shrinkage was reported following SCF blockade, and so it is not clear whether this confounded results with respect to immune function and MDSC accumulation. Finally,
the definition of MDSC employed in this study is only used by this particular laboratory,
and so its implication for other MDSC subsets is not known.
Stem cell factor is likely to play a role in MDSC accumulation, considering its role in hematopoetic stem cell maintenance. However, it is not likely to be produced by all tumors in sufficiently large amounts. Like VEGF, it is unlikely to act in an isolated fashion to expand MDSC, as it requires synergy with other cytokines to stimulate myeloid cell outgrowth in vitro.
G-CSF and GM-CSF
G-CSF or Granulocyte-Colony-Stimulating Factor can be viewed as an immunosuppressive cytokine because its overproduction results in a shift from a Type-1,
IFN-gamma mediated immune response to a Type-2, IL-4/IL-13 mediated immune
33 response in T cells, and can also impair T cell proliferation (142, 143). Part of this effect may be due to its ability to expand MDSC (30, 99, 103, 137, 144). GM-CSF, on the other hand, was long considered an immune adjuvant, which has lead to its widespread use in combination with tumor cell vaccines. Yet, its production by numerous types of tumor cells was shown to ultimately benefit tumors, as it was associated with enhanced tumor spread, as well as increased MDSC formation from BM, recruitment to lymphoid organs, and suppression of CD8+ T cell function (145-148). Importantly, evidence in both tumor-bearing mice and humans show that GM-CSF administration with vaccine, while efficacious at low concentrations of GM-CSF, can actually be immunosuppressive at high concentrations of GM-CSF as a result of MDSC recruitment (104, 116).
S100A9 and inflammation
Evidence also exists to suggest that MDSC accumulation is a result of chronic inflammation and the cytokines produced as a result of this inflammation (149). Indeed,
MDSC also have an important function in infection, chronic autoimmune inflammation, and the prevention of graft rejection, where they help contain host immunity (92, 150-
156). In the setting of infection, MDSC attenuate with the clearance of microbes, however in the setting of chronic infection with microbes which cannot be cleared, and in the setting of cancer which is not self-limiting by nature, MDSC persist. Studies out of
Dr. Ostrand-Rosenberg’s lab have supported this notion by showing that prostaglandin
(PGE2) and interleukin-1 beta (IL-1β) signaling can enhance MDSC accumulation and suppressive function in tumor-bearing animals (107, 157, 158).
Along these same lines, the newly characterized S100 inflammatory proteins of myeloid origin- S100A8 and S100A9 (myeloid-related proteins-MRP 8/calgranulinA and
34 MRP 14/calgranulin B respectively) have recently been implicated in MDSC accumulation in tumor bearing mice (89). Cheng et. al. have found Gr-1+ myeloid cells from tumor-bearing and naïve mice to express S100A8 and S100A9 proteins, which are at least partially regulated by the Stat3 transcription factor, and act at least partially to upregulate intracellular levels of reactive oxygen species (ROS). In these studies the ability of tumor-conditioned media (TCM) to block dendritic cell (DC) and macrophage formation in favor of MDSC formation from hematopoietic progenitor cells (HPC) was dependent on progenitor cell expression of S100A9. Likewise, embryonic stem cells overexpressing S100A9, when subjected to in vitro DC differentiation cultures, were unable to form into DC, but remained as immature myeloid progenitor cells. In vivo studies were not as clear-cut, but showed that S100A9 KO mice had normal myeloid cell development under steady-state conditions, but formed less splenic MDSC in response to tumor, although these results are likely confounded by the more important finding that 9 out of 12 S100A9 KO mice immunologically rejected EL-4 lymphoma tumors, while WT mice were susceptible to these same tumors. In S100A9 over-expressing, transgenic mice, the authors found the relative percentage of splenic macrophages and DC to be decreased among GFP-reporter positive cells. While the cell number of MDSC among
GFP+ cells was increased, relative percentages weren’t shown in this case, and the numbers suggested that overexpression of S100A9 does not alone bestow mice with the large increases in splenic MDSC seen in the tumor-bearing state. S100A9 transgenic mice did, however, have a significant accumulation of ckit+ myeloid progenitor cells among GFP+ bone marrow cells. This suggests that intracellular S100A9 may prevent stem cell maturation and enhance stem cell proliferation, but that other factors derived
35 from tumors may be important in the expansion of the relatively more mature,
CD11b+Gr1+ cell population.
Indeed, a related study done by Sinha et. al.found that MDSC express, secrete and bind S100A8 and S100A9 proteins (159). Once bound to receptors on MDSC, S100 proteins were found to induce MDSC chemotaxis through NFkB activation. The blocking of S100 A8/A9 binding with antibodies recognizing carboxylated N-glycans expressed on the receptors for S100A8/A9 proteins, reduced MDSC accumulation in the spleens and lymph nodes of metastatic tumor-bearing mice, but did not reduce the
suppressive function of MDSC on a per cell basis. This suggests that S100A8/9 proteins
also provide a positive autocrine feedback loop that ensures the maintenance of MDSC in
the inflammatory tumor environment.
Intracellular signaling implicated in MDSC programming
The vast majority of data indicate that signal transducer and activator of
transcription 3 (STAT3) signaling is important in MDSC biology (92, 103). Exposure of
bone marrow derived progenitor cells to tumor-conditioned media was shown to result in
Janus kinase 2 (JAK2) and STAT3 activation, which lead to the in vitro expansion of
MDSC. Such expansion was abrogated with the inhibition of STAT3 expression (160,
161). Phosphorylated-STAT3 was later found to be elevated in the immature myeloid
cells found in tumor-bearing mice, compared to those found in naïve mice (162); and
STAT3 ablation through the use of conditional knockout mice or selective inhibitors,
reduced MDSC expansion and improved T cell function in tumor-bearing mice (162,
163) This may relate to STAT3’s newly recognized role in S100A8 and S100A9 protein
expression, as described above. Under steady state conditions, however, STAT3 plays a
36 negative regulatory role in basal neutrophil production, via its control of SOCS3, a crucial feedback inhibitor of G-CSF signaling (164, 165). Conversely, during G-CSF- mediated emergency granulopoiesis, STAT3 KO mice have an inappropriately low ratio of immature, Gr1lo neutrophils to mature, Gr1hi neutrophils, and fail to elicit acute neutrophil mobilization from bone marrow, indicating a subtle positive role for STAT3 in emergency granulopoiesis which likely relates to its ability to maintain cells in an immature state (166).
Further justification for STAT3’s potential role in MDSC accumulation is the fact that many of the factors implicated in this phenomenon activate signal transduction through STAT3 phosphorylation. Numerous oncogenic signaling pathways frequently activated in cancer converge on STAT3 (167). Indeed, STAT3 is constitutively active in several tumor cell types, and thereby contributes not only to tumor cell survival and proliferation, but also to the production of immunosuppressive molecules including
VEGF, IL-10, and TGF-beta (168, 169). The primary targets for VEGF in adult animals are endothelial cells, which have relatively high levels of VEGFR1 and VEGFR2 and respond to low in vitro concentrations of VEGF (1-10ng/mL), and VEGF conclusively activates STAT3 in this cell type (170, 171). However, evidence for STAT3 activation in response to VEGF receptor signaling on normal hematopoietic cells is quite limited
(172). This may relate to the fact that VEGFR1 is considered to be relatively kinase dead, and yet it is the receptor form thought to be expressed by myeloid cells under certain conditions; meanwhile VEGFR2 is kinase active, yet has not conclusively been shown to be present on normal hematopoietic cells (173, 174). IL-10 is known to activate STAT3 signaling in myeloid cells (175-177); yet, IL-10 has not been implicated
37 in the accumulation of MDSC, but rather is implicated in MDSC suppressive function. In
addition, TGF-beta, is proposed to inhibit STAT3 signaling, and abrogation of TGF-beta
signaling in mammary carcinomas enhanced MDSC recruitment in tumor-bearing mice
(134, 178, 179).
Importantly, SCF is considered incapable of tyrosine-phosphorylating STAT3 as a singular agent, but can serine-phosphorylate STAT3 (Ser727) to enhance maximal
STAT3 transcriptional activity when given in combination with G-CSF. The proliferative synergy between these agents was shown to rely on complete STAT3 phosphorylation through distinct pathways partially dependent on phosphatidylinositol-3 kinase and ERK’s (180). G-CSF, rather, is known to be a potent activator of STAT3
(181-186).
The potential involvement of STAT5 signaling in MDSC expansion has, by and large, been overlooked. Loss of STAT5 function in primary bone marrow cells leads to a reduction in CFU-G colony formation in vitro (187, 188), and BM cells from mice lacking both STAT5 isoforms are unable to repopulate the neutrophilic granulocyte lineage of lethally irradiated wild-type hosts (189, 190). Furthermore, STAT5 favors the survival of myeloid progenitors by inducing the expression of the anti-apoptotic Bcl-2 family member, Bcl-xL (191). In addition, the antiapoptotic function of STAT5 in granulocytic and progenitor cells, is required for maintenance of neutrophil homeostasis, especially during an inflammatory response (192). Furthermore, GM-CSF is known to activate STAT5, and previous work has shown that GM-CSF addition to BM cultures in vitro drives STAT5-dominated cultures producing an abundance of CD11b+Gr1+ cells that were toxic to T cell cultures (193). Importantly, the experiments outlined in chapter
38 4 suggest that STAT3 signaling may be dominant in peripheral MDSC, and may sanction sunitinib-sensitivity in MDSC. However, they also suggest that STAT5 is additionally capable of serving as a powerful sponsor of MDSC expansion and survival, and one that protects MDSC in the presence of sunitinib.
STAT5 and STAT3 compared
The STAT family of proteins has known importance in mediating signals initiated by a wide variety of cytokines (194). STAT5 and STAT3 in particular appear to be crucial regulators of both T cell development and T cell differentiation (195). Knock out of both STAT5a and STAT5b, which are largely redundant, results in >99% perinatal lethality (196). However, analysis of fetal lymphoid development in these mice, along with the analysis of hypomorphic STAT5 mice with a less severe phenotype indicates
STAT5 activity is required for the normal development of all normal lymphoid lineages
(196, 197). This is most likely because all cytokines using the common gamma chain receptor (γc) signal through STAT5, including IL-7 and IL-15, such that deficient mice exhibit a severe combined immunodeficiency picture that is similar to deficiency of the
γc itself (198, 199). STAT5 also contributes to T cell subset development and deficiency in STAT5a and STAT5b leads to a loss of Treg cells and inability to induce Treg in vitro
(200). Likewise, constitutive STAT5 activation enforces Treg formation that is independent of other upstream signals (201). In contrast to STAT5’s role in lymphoid development, deletion of STAT3 is not reported to effect T or B cell development (202).
Rather, STAT3 is important only in T cell differentiation, where it is critical in the development and function of Th17 cells- helper T cells which produce IL-17, among other cytokines, which attract neutrophils and other inflammatory cells to sites of tissue
39 damage (203, 204). Thus, the balance of Treg and Th17 T cells, which sit on opposite
ends of the inflammatory spectrum, appears to be regulated by STAT5 and STAT3.
In myeloid lineage cells, the roles of STAT5 and STAT3 are perhaps not as clear cut. STAT3 total knock out mice die early in embryogenesis, but Cre-loxP cell line-
specific KO mice have allowed the study of STAT3 KO in vivo (205). STAT3 appears to function in myelopoeisis mainly via activation initiated by G-CSF receptor ligation
(206). One of STAT3’s main downstream targets is Socs3 (suppressor of cytokine 3
protein), which functions to block generation of the Stat signal from a cytokine receptor,
including the G-CSFR (207). When STAT3 was conditionally ablated in early
hematopoeitic development, excessive numbers of neutrophils accumulated in bone
marrow and blood. This phenotype was also apparent in Socs3 ablated mice (164, 165,
208). Socs3 is known to bind G-CSFR and reduce the amplitude of STAT3 signaling in
response to receptor ligation. Thus loss of Socs3, either by direct ablation or by the
ablation of STAT3, caused increased G-CSFR signaling and increased neutrophil
production. This indicates that Socs3 negatively regulates neutrophil production by
regulating G-CSFR signaling generally, and not just via specific inhibitory effects on
STAT3. In the absence of STAT3, there is likely to be elevated signaling through G-
CSFR, perhaps via excessive STAT5 or MAP kinase signaling (204). In terminal neutrophils, STAT3 functions independently of Socs3 to enhance neutrophil migration and chemotaxis (166). In mature monocytic myeloid cells, STAT3 mediates all of the anti-inflammatory effects of IL-10 (209-211). However, STAT3 is also activated in response to IL-6 (via gp130 receptor), which is largely pro-inflammatory. In the case of
IL-10, however, STAT3 signaling turns on the “anti-inflammatory response” (AIR) gene,
40 which inhibits the transcription of pro-inflammatory genes. In the case of IL-6 signaling
through gp130, however, IL-6R cannot activate the AIR gene(s) unless Socs3 is absent,
and Socs3 is induced by STAT3 in response to both IL-10 and IL6. Thus the anti-
inflammatory signal generated by IL-10R, via STAT3, is not unique to the IL-10R, but is
actively repressed from other receptors by Socs3 (212). Thus, phospho-STAT3 status
alone is insufficient to describe what is going on in a cell.
The involvement of STAT5 in myeloid cell development and function is largely
attributable to its role in GM-CSF and IL-3-induced signal transduction (213).
STAT5a/b hypomorphic mice had reduced white blood cell counts, but this was mostly
the result of decreased lymphocytes. However, a reduction in myeloid colony
proliferation and expansion in response to GM-CSF, IL-3 and G-CSF was found (214).
STAT5 is also an important mediator of the increase in proliferation and viability
conferred to mature myeloid cells in response to GM-CSF (215). GM-CSF has long been
considered the dominant pathway for dendritic cell generation, via STAT5 signaling.
However, CD34+ progenitor cell-derived DC generated in GM-CSF via STAT5
signaling, were recently shown to be inferior to DC generated in IL-6+Flt3L via STAT3
signaling, at inducing at Th1 type T cell response (193).
Targeted cancer therapy with sunitinib
Recent treatment strategies targeting tumor angiogenesis have produced frequent
therapeutic effects in RCC patients. One such agent, sunitinib, is a multitargeted receptor
tyrosine kinase inhibitor (RTKI) of VEGF and related receptors. It has produced
significant objective responses in patients with metastatic RCC and a superior progression-free survival when compared to IFN-alpha (216-218). While anti-angiogenic
41 agents such as sunitinib, bevacizumab and sorafenib produce disease stabilization in many RCC patients and shrink tumors in some, all patients eventually incur disease progression after a time period usually measured in months (130). Such disease progression is thought to reflect an evasive response to drug by the tumors (131). The optimal role of immunotherapy following or simultaneous to such anti-angiogenic therapy remains to be determined.
Sunitinib inhibits signaling not only through the vascular endothelial growth factor receptors (VEGFR), but also through platelet-derived growth factor receptor (PDGFR), stem cell factor receptor (c-Kit), Flt3, and colony stimulating factor-1 receptor (CSF-1R) tyrosine kinases (219). While it is well documented that a major effect of sunitinib is to block angiogenesis, our studies with RCC patients revealed that sunitinib was also effective in reversing cancer-associated immune suppression, though the mechanism was previously unknown (53). Recently published studies in two different murine tumor models for which sunitinib was used as monotherapy have confirmed the capacity of suntinib to enhance immune cell function (220, 221). These beneficial immunomodulatory effects were not observed in mice treated with the promiscuous
RTKI sorafenib, despite its activity in RCC. Furthermore, in contrast to sunitinib, sorafenib impaired primary T cell responses in vivo (222). Similarly, we have thus far found that no single RTK inhibitor or combination of inhibitors has been able to reproduce single agent sunitinib’s immune stimulating effects. Our observations for singular VEGFr inhibition with vatalanib are consistent with reports showing no MDSC inhibition in patients by either VEGF-trap (223) or bevacizumab (anti-VEGF mAb) alone (224). While the VEGF and c-kit receptors have each been implicated in MDSC
42 generation (102, 139), our preliminary data suggest that additional tumor-host
interactions are at least as important. We found that in both the 4T1 and CT26 tumor
models, maximally tolerated doses of either imatinib (targets ckit, PDGFR and BCR/abl),
vatalanib (selectively targets VEGFRs), or both drugs together minimally inhibited
MDSC generation. Unlike sunitinib, these RTKIs conspicuously do not block Flt3 (225).
The redundancy of factors implicated in MDSC accumulation is mirrored by the
unique redundancy of receptor tyrosine kinases targeted by the anti-angiogenic drug
sunitinib, several of which have been implicated in MDSC formation. Hence the
following chapters chronical a series of experiments designed first to test the efficacy of
sunitinib to reduce the MDSC burden in RCC patients; second to characterize sunitinib’s mechanism of action on MDSC; and, third to delineate sunitinib’s therapeutic limitations, as they relate to MDSC, and offer novel strategies for future clinical trial development. They also suggest that sunitinib’s unique capacity to block all three steps of
MDSC development (m-MDSC proliferation, ganglioside-mediated n-MDSC activation, and promotion of n-MDSC viability) may potentially depend upon a different set of RTK blockades at each step, thereby benefitting from sunitinib’s targeting promiscuity.
43 CHAPTER 2
SUNITINIB MEDIATES REVERSAL OF MYELOID-DERIVED SUPPRESSOR
CELL ACCUMULATION IN RENAL CELL CARCINOMA PATIENTS
Statement of Clinical Relevance
The following manuscript, “Sunitinib Mediates Reversal of Myeloid-Derived
Suppressor Cell Accumulation in Renal Cell Carcinoma Patients,” demonstrates that tumor-induced immune suppression in metastatic renal cell carcinoma patients, mediated
by myeloid derived suppressor cells, can be reversed by a sunitinib-induced reduction in
MDSC numbers. Because MDSC are known to inhibit T cell sensitization to tumor
antigen, their depletion may be clinically desirable prior to the initiation of
immunotherapy modalities such as adoptive T cell transfer, dendritic cell-based vaccines,
or cytokine therapy. Other modalities such as all-trans retinoic acid (ATRA) or certain
chemotherapy regimens can also deplete MDSC. However, the well known anti-
angiogenic effects of sunitinib, combined with its known tolerability and objective
benefit in the clinical setting, may render it a superior adjunct agent for immunotherapy
trials. The observed impact of sunitinib on host immune cells is likely to be independent of its anti-tumor effect, and so its potential benefit in immunotherapy may not be specific to tumor type. These data should thus be applicable to the design of future clinical trials.
44 Abstract
Purpose: Immune dysfunction reported in renal cell carcinoma (RCC) patients may
contribute to tumor progression. Myeloid derived suppressor cells (MDSC) represent one
mechanism by which tumors induce T cell suppression. Several factors pivotal to the
accumulation of MDSC are targeted by the tyrosine-kinase inhibitor, sunitinib. The
impact of sunitinib on MDSC-mediated immune suppression in RCC patients has been
investigated. Experimental Design: Patient peripheral blood levels of MDSC and Treg, and T cell production of interferon-gamma (IFNγ) were evaluated before and after sunitinib treatment. Correlations between MDSC and Treg normalization, as well as T cell production of IFNγ were examined. The in vitro impact of sunitinib on patient
MDSC was evaluated. Results: mRCC patients had elevated levels of CD33+HLADR- and CD15+CD14- MDSC, and these were partially overlapping populations. Treatment with sunitinib resulted in significant reduction in MDSC measured by several criteria.
Sunitinib-mediated reduction in MDSC was correlated with reversal of Type-1 T cell suppression, an effect which could be reproduced by the depletion of MDSC in vitro.
MDSC reduction in response to sunitinib correlated with a reversal of
CD3+CD4+CD25hiFoxp3+ Treg cell elevation. No correlation existed between a change
in tumor burden and a change in MDSC, Treg, or T cell production of IFNγ. In vitro
addition of sunitinib reduced MDSC viability and suppressive effect when used at
1.0ug/mL or greater. Sunitinib did not induce MDSC maturation in vitro. Conclusions:
Sunitinib-based therapy has the potential to modulate anti-tumor immunity by reversing
MDSC-mediated tumor-induced immunosuppression.
45 Introduction
This year approximately 36,000 Americans will be newly diagnosed with kidney cancer, resulting in 12,890 deaths (226). Poor outcome in renal cell carcinoma (RCC) is related to its late disease presentation, propensity for recurrence, and refractoriness to traditional chemotherapy or radiotherapy. Additionally, treatment with cytokine therapy has resulted in limited success, despite the known immunogenicity of RCC (50, 227-
230). The receptor tyrosine-kinase inhibitor- sunitinib is an oral medication being used with significant clinical impact in metastatic clear-cell RCC (217, 218, 231, 232).
Sunitinib inhibits signaling through the vascular endothelial growth factor receptors
(VEGFR), as well as platelet-derived growth factor receptor (PDGFR), stem cell factor receptor (c-Kit), Flt3, and colony stimulating factor-1 receptor (CSF-1R) (219).
VEGF signaling plays a prominent role in the pathogenesis of clear-cell RCC, particularly due to the common occurrence of Von Hippel-Lindau (VHL) gene inactivation in these tumors (232, 233). VEGF overproduction, resulting from unchecked transcription by hypoxia-inducible factor (HIF1-alpha) that occurs in the absences of
HIF1a’s normal inhibitor, VHL, promotes tumor-associated angiogenesis required for tumor growth and metastasis (232, 233). VEGF may also support tumor growth via negative effects on host anti-tumor adaptive immunity, as increased VEGF levels have been associated with alterations in myeloid cell differentiation, which impair competent dendritic cell formation and encourage suppressive myeloid cell formation in cancer patients (91, 234-238).
Myeloid-derived suppressor cells (MDSC) impair T cell effector function, and represent a heterogeneous population of cells that accumulate in tumor-bearing hosts as a
46 result of tumor induced alterations in myelopoiesis (98). MDSC accumulating in the
tumors and lymphoid organs of tumor-bearing mice are CD11b+Gr-1+ and mediate T cell
impairment that is reversed with tumor removal or CD11b+ or Gr-1+ cell depletion (91,
98, 239, 240). MDSC detected in the peripheral blood of patients bearing several tumor
types (108, 114-117, 234, 241), express the common myeloid marker CD33, but lack markers of mature myeloid cells such as the MHC class II molecule HLA-DR.
Expression of the granulocytic marker CD15 divides patient MDSC into at least two subsets. The CD15 positive subset has been shown to suppress T cell function in patients with kidney cancer, among others, through an arginase and/or reactive oxygen species- dependent mechanism (108, 115). A CD15 negative subset of MDSC was also shown to suppress T cell function in patients with squamous cell carcinoma of the head/neck/lung, or adenocarcinoma of the breast/lung, through an unclear mechanism (114). These subsets likely parallel those recently identified in the mouse model, where the CD15+
(human) and the Gr1hi (mouse) MDSC are granulocytic, and the CD15- (human) and the
Gr1lo (mouse) MDSC are monocytic (112). New evidence in hepatocellular carcinoma
and malignant melanoma patients suggests that a third, CD14+HLADRdim subset of
MDSC also exists (116, 117).
Because VEGF has been implicated in the generation of MDSC (139, 140), we
have evaluated the effect of sunitinib, which blocks signaling through multiple receptors,
including VEGF receptors (219), on MDSC in mRCC patients. MDSC are reported to
inhibit T cell function directly (98) as well as indirectly, via the induction of regulatory T
cell (Treg) formation (93, 117). We have therefore examined patient T cell production of
IFNγ, as well as patient Treg levels, before and after treatment with sunitinib. We
47 observed an elevation in CD15+ and CD15- MDSC in RCC patients compared to age-
matched normal donors. We show that treatment with sunitinib in mRCC patients
reduces the elevated levels of both MDSC subsets to near normal levels. A reduction in
MDSC measured by two criteria (CD15+CD14- and CD33+HLADR-) was correlated with
a subsequent recovery in patient T cell effector function, but not with tumor regression.
A reduction in CD15+CD14- MDSC was associated with a reduction in patient T regulatory cell levels. mRCC patient MDSC-mediated T cell suppression in vitro was reversible with MDSC depletion and modestly with the addition of excess L-arginine.
Sunitinib, when used in vitro, partially blocked the suppressive function of MDSC, possibly due to its effects on MDSC viability when added at 1.0ug/mL or greater, rather than its effects on MDSC maturation. Sunitinib-based combination immunotherapy may be a promising option for the future treatment of metastatic RCC.
Subjects, Materials and Methods
Patient population and treatment. Patients received sunitinib monotherapy for mRCC
(50mg by mouth daily) for 28 days followed by 14 days of rest, comprising one 6-week
cycle. Patients were excluded if they received any anti-cancer therapy concomitant with
sunitinib, if they had a diagnosis other than clear cell RCC or did not receive at least 28 days of sunitinib. Patients underwent disease assessment (CT and bone scans) at baseline, and after every 2 cycles (approximately every 12 weeks). Objective response according to the Response Evaluation Criteria In Solid Tumor (RECIST) criteria (242) and tumor burden shrinkage were determined by investigator assessment of radiographs.
Patients were treated until RECIST-defined disease progression or unacceptable toxicity occurred. Dose interruption and modification was performed according to treating
48 physician discretion. All patients and healthy volunteer blood donors signed an IRB- approved, written informed consent for collection of blood samples. Age-matched normal donors were healthy volunteers over the age of 50 years.
Reagents. Human IgG and L-arginine were from Sigma-Aldrich (St. Louis, MO).
Catalase was from Calbiochem (La Jolla, CA). 3H-thymidine was from Amersham-
Buchler (Braunschweig, Germany). Anti-human CD3, CD4, IFNγ, IL-4, CD11c, CD13,
CD14, CD15, CD33, HLA-DR, annexin V, 7AAD, and the annexin V staining kit were from BD Biosciences (San Jose, California). Anti-human Foxp3, CD11b, CD19, CD40,
CD56, CD80, and CD86 were from eBioscience (San Diego, CA). Mouse isotype control antibodies were from BD, eBioscience, or Immunotech. Anti-human CD15 and
CD33 antibody-coated magnetic microbeads and LS magnetic columns were from
Miltenyi Biotec (Auburn, CA). GM-CSF and IL-4 were from R&D systems. Sunitinib in pure powder form was from Pfizer.
Peripheral blood mononuclear cell isolation. Peripheral blood (60ml) was drawn from metastatic RCC (mRCC) patients prior to sunitinib treatment (cycle 1 day 1) and on day
28 after one cycle of treatment (cycle 1 day 28), and, in a subset of patients, on day 28 after two cycles of treatment (cycle 2 day 28), and from age-matched normal donors.
Peripheral blood was drawn in heparin containing tubes. Peripheral blood mononuclear cells (PBMC) were isolated within two hours of blood draw and either used fresh, or frozen for later use, according to the methods previously described (53). For phenotypic and functional studies where multiple time points were available, all time points for an individual patient were thawed together and used in the same experiment.
49 FACS analysis of patient PBMC. Analysis of MDSC percentages in patient PBMCs, were done on thawed samples. Cells were stained in FACS buffer (1x PBS with 2% heat inactivated FBS and 0.02% sodium azide). Non-specific antibody binding was blocked by pre-treatment of cells with 10ug/mL of human IgG for 20 minutes at room temperature. Surface stains were added to cells for 30 minutes at 4°C. Cells were stained with anti-human CD11b, CD11c, CD14, CD15, CD33, HLA-DR. In a subset of patients, cells were also stained with anti-human CD3, CD19, CD56, CD80, and CD86. Cells were washed in buffer and then fixed in 1% paraformaldehyde (PFA) and ran for FACS.
T-regulatory cells were enumerated by FACS on thawed patient PBMC rested in complete RPMI 1640 media overnight (37°C, 5% CO2) then stained with anti-CD3, CD4,
CD25 (Stemcell Technologies, Vancover BC), and Foxp3 according to the instructions included in the BD intracellular staining kit. PBMC were incubated overnight to facilitate experiments that could examine both Treg cells and T cell stimulation from the same sample. Thawed T cells stimulate better after some amount of resting period, and multiple past experiments have shown numerically equivalent results for Treg staining done immediately after thawing samples or done after an overnight culture. All data was acquired using Cellquest on a BD FACSCalibur, and analyzed using either Flow Jo (Tree
Star Incorporated, Ashland, OR) or Cellquest software. At least 300,000 live cell events were collected for each tube used in analysis.
Determination of patient T cell IFNγ response. Patient PBMC samples were stimulated with anti-CD3/28 bound beads (Dynal, Carlsbad, CA) and IL-2 (Chiron,
Emeryville, CA) for 72hrs. Golgi plug was added to cells for the last 6 hours, and harvested cells were stained with anti-human CD3, CD4, IFNγ and IL-4 according to the
50 protocol provided using the BD intracellular staining kit. Non-stimulated cells from each
donor served as a negative control. Additionally, specificity of cytokine staining was
confirmed in each sample via subtraction of any non-specific staining occurring in
samples pretreated with unlabeled anti-cytokine antibodies prior to the addition of
fluorochrome-labeled antibodies.
MDSC depletion. One-half of each patient sample was treated with anti-CD15
antibody-coated magnetic microbeads. Cells were incubated at 4°C for 20 minutes, then
washed and resuspended in PBE (Phosphate Buffered Saline with BSA and EDTA).
Cells were run over an LS magnetic column for the depletion of bead-labeled cells, as per
the manufacturer’s instructions. FACS analysis was done on a small aliquot of cells to
assure that MDSC had been effectively depleted. Cells were then resuspended in
complete RPMI media, and then activated and stained for intracellular cytokines (IFNγ
and IL-4) as described above.
In vitro sunitinib culture assays. Fresh patient PBMCs were incubated with a mixture
of anti-human CD33 and CD15 antibody conjugated magnetic beads, CD33+ and CD15+ myeloid cells were positively selected with an LS column. Cells which flowed through the columns were at least 80% CD3+ T cells, while positively selected myeloid cells were
all CD11b+ and on average 40% were negative for HLADR, and thus represented MDSC.
For studies evaluating the in vitro effect of sunitinib on MDSC, a minimal
concentration of 0.1ug/mL was used, as this is equivalent to the levels detected in human
plasma (243), and sunitinib was tittered up to 10 to 50-fold from there. To measure the
impact of sunitinib on MDSC viability, myeloid cells containing MDSC were incubated
in complete RPMI media with 20% SK-RC26B (gift from Dr. Neil Bander, Cornell
51 Medical Center) RCC cell line-conditioned media (TCM) and 50ng/mL of GM-CSF to support cell viability over 48 hours (observed and published findings, 22). Sunitinib, suspended in plain RPMI media was added at 0.1ug/mL, 1.0ug/mL, and 5.0ug/mL to three groups of cells. After 48 hours, cells were harvested, and surface stained for
HLADR and CD33, for 15 min at 4°C in FACS buffer, then stained for annexin V and
7AAD per BD provided protocol for 15 minutes, and then ran for FACS.
For studies examining the impact of sunitinib on MDSC differentiation, patient myeloid cells were isolated and incubated in GM-CSF and TCM-containing cultures as above, with the addition of 50ng/mL of IL-4, with or without sunitinib at either 0.1ug/mL or 1.0ug/mL. Half the media was replaced after 3 days, and after 6 days, the cells were harvested and analyzed for the expression of HLADR, CD40, CD80, and CD86 by
FACS. Remaining cells were irradiated at 3000Rad and used as stimulators in mixed lymphocyte reactions (MLR) with normal donor, allogeneic, fresh T cells. T cell proliferation was determined after 6 days by the incorporation of tritiated thymidine.
Finally, for studies examining the mechanism of MDSC-mediated T cell suppression, and the impact of sunitinib on this, patient T cells and MDSC-containing myeloid cells isolated as before were co-cultured 1:1 in the presence of either catalase
200U/mL, L-arginine (2mM), or sunitinib (0.1 or 1.0ug/mL), and T cells were stimulated with anti-CD3/28 coated beads for 72h with intracellular cytokine production being evaluated as described above.
Statistical analysis. The Wilcoxon rank sum test was used to compare mRCC patients and healthy donors with respect to MDSCs, intracellular γ-interferon, and Tregs, and to compare these parameters in patients who achieved a partial response by RECIST criteria
52 versus patients whose best response was stable disease or progression. Spearman rank
correlations were used to assess associations between immune parameters and associations with changes in tumor burden. The t-test was used to compare results of in
vitro experiments. All statistical tests were 2-sided and all analyses were conducted
using SAS (version 9.1; SAS Institute Inc., Cary, NC).
Results
Patient characteristics and clinical response to sunitinib.
Data from twenty-three mRCC patients treated with sunitinib monotherapy between 8/05 and 8/07 were available for analysis. Patient characteristics are summarized in Supplementary Table 1. Overall, seventy-six percent of patients were male, median age was 57 (range 41-80) and most patients (86%) had ECOG performance status 0 or 1. Eighty-one percent of patients had prior nephrectomy, 43% had received prior systemic therapy- primarily sorafenib, thalidomide, interferon-α, and/or interleukin-
2, and one patient had received prior radiotherapy. Using the MSKCC criteria for previously untreated patients (34), 29% of patients were considered to have a favorable risk profile, 62% were considered intermediate, and 10% were considered poor risk.
Forty-three percent of patients achieved a partial response by RECIST criteria; the median change in tumor burden was a 22.5% decrease (range 60% decrease to 50% increase). Seven patients have progressed and 5 patients have died. Median follow-up for the patients still being followed is 4.4 months (range 3.0-16.6).
53 Supplementary Table 2-1- Patient Characteristics1
Factor N (%)
Gender Male 16 (76%) Female 5 (24%)
Age at Start of Treatment Mean + s.d. 58.5 + 2.0 Median 57 Range 41 – 80
ECOG Performance Status 0 7 (33%) 1 11 (52%) 2 3 (14%)
Prior Treatment Nephrectomy 17 (81%) Systemic Therapy 9 (43%) Radiation 1 ( 5%)
MSKCC Risk Group Favorable 6 (29%) Intermediate 13 (62%) Unfavorable 2 (10%)
Best Response Partial 9 (43%) Stable Disease 10 (48%) Progression 2 (10%)
% Tumor Reduction Mean + s.e. 15.9 + 6.9% Median 22.5% Range -50 – 60%
1 Data unavailable for two patients
54 Elevated MDSC in mRCC patients decline in response to sunitinib.
mRCC patient PBMCs were analyzed before the start of treatment (Cycle 1 Day
1), and after one or two cycles of treatment with sunitinib (Cycle 1 Day 28 and Cycle 2
Day 28 respectively) for MDSC, and their levels were compared to those in age-matched
normal donors (AMN). MDSC previously described in RCC patients-CD14-CD15+
MDSC as well as CD33+HLADR- MDSC (108, 241) were quantified as shown (Figure
1A). These cells were also confirmed to be positive for myeloid markers CD11b, CD11c,
and CD13 in a subset of patients (data not shown). When each of the MDSC populations
were calculated as a percentage of total PBMC, a highly significant increase in the
number of both circulating CD14-CD15+ MDSC and CD33+HLADR- MDSC were seen
in mRCC patients (mean=5.49% and 5.42% respectively) when compared to healthy age-
matched normal donors (mean=.23% and .76%, p<.001 and p=.002 respectively, Figure
1B). MDSC by both criteria significantly declined after one cycle of treatment
(mean=2.21% and 2.28%, p=0.005 and p=.007 respectively, Figure 1B). In the subset of patients available who were treated with two cycles of sunitinib, MDSC continued to decline with an additional cycle of therapy (mean= 0.75% and 1.29%, p<.001 and p=.02 respectively Figure 1B).
In order to confirm our suspicion that some degree of overlap existed between the two populations of MDSC, as well as to assure that CD15- cells were also a target of sunitinib, a subset of 15 patients was analyzed by 4-color FACS whereby CD11c, CD33,
HLADR, and CD15 were all in the same acquisition tube. A representative dot plot and histogram detail the analysis undertaken (Figure 1C), which allowed for the quantification of both CD33+HLADR-CD15- immature myeloid cells which are likely to
55 be more monocytic in nature (114), as well as CD33+HLADR-CD15+ immature myeloid
cells which are likely to be more granulocytic in nature (108, 115). Both populations of
MDSC, which are likely similar to those recently characterized in the mouse tumor model
(112), declined in response to treatment with sunitinib (CD15+ p=.02, CD15- p=.005,
Figure 1C). MDSC declining in response to sunitinib were confirmed to be lineage negative in a subset of patients (p<.006, Figure 1D).
56
57
Sunitinib suppresses bone marrow production of myeloid cells but enhances
lymphoid cell production.
Because sunitinib induced marked changes in MDSC, and because some of the
receptors targeted by sunitinib influence hematopoiesis, we analyzed twenty patient complete blood counts (CBC) with white blood cell differentials (w/ diff) that had been reported at the appropriate time relative to sunitnib treatment (Supplementary Table 2 ).
Total WBC counts declined with treatment from a median pretreatment amount of 7.7
K/µL, to 4.1 K/µL (pretreatment to cycle 1 day 28, p<.001), but stayed within the range of normal for most patients. The percentage of neutrophils significantly decreased, but stayed within the range of normal for most patients, (71% to 63%, p<.001). In contrast, a decline in monocyte percentage was slight (8% to 7%), while the decline in absolute monocyte counts was significant (0.689 to 0.281, p<.001). Meanwhile, the percentage of lymphocytes significantly increased (16% to 27%, p<.001) into the normal range after treatment in most patients. Thus, sunitinib may have a myelospecific effect on bone marrow function.
58
Supplementary Table 2- Sunitinib alters several white blood cell parameters1
Pretreatment Cycle 1 Day 28
(n=20) (n=20)
WBC (K/µL) 7.7 4.1*
%Neutrophils 71% 63%*
%Lymphocytes 16% 27%*
%Monocytes 8% 7%†
1 Complete blood counts with differentials were often times sent at the same time as study bloods were drawn on mRCC patients receiving sunitinib, and whole bloods were processed by the CCF automated laboratory services. (*p<.001, †p=.04)
59 Declines in MDSC are associated with increases in IFNγ-producing T cells after
sunitinib therapy.
In a related study, we have evaluated CD3+ T cell production of IFNγ in mRCC
patients before and after treatment with sunitinib, and found that mRCC patients have
significantly reduced amounts of IFNγ production, and that this type 1 response increases after sunitinib treatment (p=<.001, n=38, (53)). We observed similar results in this study, which included a smaller subset of patients. Pretreatment patient T cell production of
IFNγ was reduced (median=7.71%), when compared to AMN donors (median=16.43%, p=.008). Treatment with sunitinib increased the amount of IFNγ-producing T cells in mRCC patients after 1 cycle (median=13.81%) and 2 cycles of therapy (median=15.93%) although this did not reach statistical significance in this subset (Figure 2A). Type I T cell IFNγ response normalization following sunitinib treatment was seen to coincide with normalizing numbers of MDSC, suggesting that MDSC levels may need to be reduced below a certain threshold before adaptive T cell immunity can be recovered. Indeed, as seen in Figure 2B, reductions in MDSC after two cycles of therapy were directly correlated with an overall increase in patient T cell IFNγ production from baseline (r=-
.66, p=.03). Additionally, mRCC patients with relatively larger numbers of persisting
MDSC after sunitinib treatment had relatively lower amounts of plasma IFNγ (r=-0.81, p=0.02, n=8, data not shown).
60
61 In vitro depletion of patient MDSC partially restores patient T cell production of
IFNγ Because of the negative association between mRCC patient MDSC and T cell
IFNγ production, we sought to determine whether removal of MDSC in vitro could
render patient T cells capable of a type-1 response. Selected PBMC samples taken from
patients with high levels of MDSC prior to the initiation of therapy were chosen and half
of each sample was depleted of MDSC with anti-CD15 magnetic beads (Figure 3A), and
then both conditions were stimulated with anti-CD3/CD28 for 72 hours as previously
described. CD3+ cells were analyzed for IFNγ and IL-4 production (Figure 3A), and the
mean levels of these cytokines in normal donor T cells or patient T cells with or without
the removal of MDSC was calculated. MDSC depletion improved the ability of mRCC
patient T cells to produce IFNγ (p<.05, Figure 3B). In all groups, there were low levels
of IL-4 production seen at baseline that did not change with MDSC depletion.
62
63 In vitro impact of sunitinib on MDSC-mediated T cell suppression.
MDSC characterized previously in RCC patients have been shown to inhibit T cell function in an arginase-dependent manner (108). In addition, CD15+ and CD33+
MDSC were also suggested to inhibit T cells via the production of reactive oxygen species (115, 128). We therefore wanted to compare sunitinib’s ability to reverse patient
MDSC-mediated T cell suppression in vitro to the ability of L-arginine (2mM, (108)) and catalase (200U/mL, (115)) to do the same. In order to ensure that both MDSC subtypes would be included in our experiments, we positively selected for RCC patient myeloid cells with a combination of anti-CD33 and anti-CD15 magnetic beads. Positively selected MDSC were added to patient T cell cultures in the presence or absence of the various potential MDSC inhibitors, and T cell production of IFNγ at 72 hours following polyclonal stimulation was compared to cultures where no MDSC were added. MDSC isolated from patients were highly suppressive of patient T cell function, and the addition of L-arginine to cultures resulted in significant, although modest, reversal of T cell suppression (Figure 4). Sunitinib used at 0.1ug/mL, a level equivalent to that detected in patient plasma (244), induced a trend toward normalization of T cell function which was equivalent to that seen with the addition of catalase, although it did not reach significance. Finally, in the presence of increased concentrations of sunitinib
(1.0ug/mL), there appeared to be significant, but modest, reversal of MDSC-mediated T cell suppression (Figure 4).
64
65 In vitro impact of sunitinib on MDSC viability and differentiation.
We next asked whether the sunitinib-induced depression of MDSC accumulation
observed in patients could be related to sunitinib-mediated MDSC apoptosis or
maturation. It was necessary to add patient MDSC to cultures containing 20% SK-
RC26B cell-line tumor conditioned media (TCM) and/or 50ng/mL GM-CSF to support
long-term MDSC viability (unpublished results). To assess the impact of sunitinib on
MDSC apoptosis, patient myeloid cells were treated, or not, with sunitinib at 0.1ug/mL,
1.0ug/mL, or 5.0ug/mL for 48 hours, and then analyzed by FACS for annexin V staining
in both the CD33+HLADR- MDSC, as well as the CD33+HLADR+ monocytes/DC.
Patient T cells were also separately cultured for 48 hours in complete RPMI with or without the same amounts of sunitinib, and assessed for viability. We found that, relative
to lymphocytes, patient myeloid cells displayed an increased sensitivity to sunitinib-
induced cell death (Figure 5A). Patient MDSC seemed also to be somewhat more
sensitive to sunitinib in vitro when compared to patient monocytes. To assess
patient MDSC differentiation in response to sunitinib, isolated myeloid cells were
cultured with 50ng/mL of GM-CSF and IL-4 to stimulate DC maturation as well as 20%
TCM to prevent indiscriminant maturation. Sunitinib was added, or not, to cultures at
0.1ug/mL or 1.0ug/mL. Figure 5B shows that patient CD33+HLADR- myeloid cells
added to cultures were negative for the expression of CD80 and CD86 as compared to
CD33+HLADR+ cells. After 6 days in culture, cells were harvested and their expression
of DC maturity markers compared by single-color FACS. Remaining cells were used in
MLRs to stimulate allogeneic T cell proliferation. Comparison of the expression of
HLADR, CD40, CD80, CD86, as well as the ability to stimulate T cell proliferation in an
66 MLR, shows that sunitinib treated and untreated cells became essentially equivalent DC
(Figure 5C). Sunitinib did not increase MHC class II and costimulatory marker expression on patient myeloid cells, nor did it increase the T cell stimulating capabilities of these cells, suggesting that it is not functioning to induce MDSC differentiation.
67
68 Changes in patient MDSC and T regulatory cells in response to sunitinib are
directly associated.
Elevated Treg have been observed in the blood of cancer patients and are believed
to suppress the development of anti-tumor immunity (245, 246). It has been shown that
MDSC in tumor bearing mice and hepatocellular carcinoma patients can induce
CD4+CD25+hiFoxp3+ Treg formation (93, 117), and several clinical strategies aimed at
Treg depletion are under investigation (247). We thus investigated modulations in
mRCC patient Treg levels in response to sunitinib. Elevated CD3+CD4+CD25hiFoxp3+
Tregs, (median=2.59% of CD3+4+ cells in mRCC versus 1.41% in AMN, p=.002, (53)),
quantified as shown (Figure 6A), and confirmed to be suppressive in multiple in vitro
experiments (53), also declined after treatment with sunitinib (although this decline did
not reach statistical significance). In the cohort of patients included in this study, we
detected a positive correlation between the numbers of Treg and MDSC remaining after
two cycles of therapy (r=.75, p=.008, data not shown). In addition, a change in MDSC
levels between cycles of treatment, was positively correlated with a change in Treg levels
over two treatments (Figure 6B, r=.93, p<.001). This warrants further investigation into
the possibility that sunitinib may interfere with Treg formation, possibly via MDSC, or
possibly via a shared sunitinib target which impacts both MDSC and Treg.
69
70 Discussion
It is now widely accepted that there are several tumor-mediated
immunosuppressive networks operational in kidney cancer that impede the success of
immune-based therapies (50, 229, 248, 249) One of these networks involves the tumor-
induced accumulation of myeloid-derived suppressor cells (103, 120). Elevated levels of
peripheral blood MDSC in mRCC patients have been shown to decline after treatment
with all-trans-retinoic acid (ATRA), presumably due to ATRA induced maturation of
these cells (241). Here we report that sunitinib monotherapy, a frontline treatment
for patients with mRCC, induces a significant reduction in circulating MDSC in mRCC
patients. In addition, this reduction was associated with an improvement in effector T
cell function, and direct correlations were observed between the drug-mediated reduction
in MDSC numbers and an improvement in T cell IFNγ production, as well as a decline in
Treg numbers. Interestingly, we did not see a correlation between changes in any of the
immune parameters tested and changes in tumor burden, response to treatment, or
survival.
The mechanism by which sunitinib improves Type 1 T cell function in mRCC
patients is currently under investigation. It is likely to be in part due to the reduction of
MDSC which inhibit effector T cell function directly. Results from in vitro experiments implicate that a portion of MDSC-mediated T cell suppression in patients was mediated by L-arginine depletion and perhaps ROS production, however further investigation is currently underway to identify additional mechanisms that may be relatively active in all
MDSC or MDSC subsets. Importantly, patient T cells can be rendered functional when
MDSC are depleted. Indeed, removal of MDSC from the PBMC of selected patients
71 prior to the stimulation of their T cells, resulted in a significant improvement in the
ability of those T cells to produce IFNγ.
Our in vitro results support the notion that patient T cells are viable and functional even in the presence of concentrations of sunitinib that depress MDSC viability and function. Exposure to 1ug/mL sunitinib induced 30% of patient MDSC to undergo apoptosis, and over 60% of MDSC were killed with 5.0ug/mL sunitinib over 48 hours,
although no effect on patient T cell viability was seen. Thus it is likely that the recovery
of IFNγ production by patient T cells that were cocultured with their own MDSC in the
presence of 1.0ug/mL sunitinib was a result of sunitinib-induced MDSC apoptosis.
Modest recovery in T cell IFNγ production in cocultures exposed to 0.1ug/mL sunitinib
may relate to sunitinib-mediated inhibition of MDSC suppressive function. Sunitinib-
mediated improvements in patient T cell function likely occur via effects on MDSC that
are exerted by multiple mechanisms including an effect on MDSC viability and function.
Because it is difficult to directly recalculate the in vitro and in vivo doses of sunitinib, we
have begun parallel experiments in several mouse tumor models where we have also
observed a dramatic suppression of MDSC in response to sunitinib (Ko JS et al,
manuscript under review). We plan to evaluate the effects of sunitinib on MDSC
function and viability in tumor bearing mice receiving sunitinib, and also to determine the
effects of drug on MDSC formation in the bone marrow and tumor.
We have previously shown that sunitinib induces a reduction in Treg levels in
mRCC patients, and that Treg reductions are associated with increases in patient T cell
function as measured via IFNγ production (53). In the current study, an increase in T cell
function was also directly correlated with a reduction in MDSC numbers, suggesting that
72 both MDSC and Treg are contributing to immune dysfunction in mRCC patients.
Sunitinib may be unique because it not only reduces the numbers of MDSC in mRCC patients, but it also reduces the number of Tregs. Similar to what we have reported in patients receiving sunitinib, we have seen that tumor-bearing mice receiving sunitinib experience a decline in Treg levels (unpublished observations). This is in agreement with what was recently published by Hipp et. al. where it was observed that sunitinib, but not another tyrosine-kinase inhibitor- sorafenib, reduced Treg levels (222). In the current study, a highly significant correlation between the decline in MDSC and the decline in
Treg in response to sunitinib was seen. Further investigation is therefore warranted to determine whether this is due to the influence of MDSC on Treg formation, or rather due to a common target of sunitinib which is shared by MDSC and Treg.
It is still possible that the immune effects of sunitinib observed in patients could be direct – via its interactions with receptors on hematopoietic cells, or indirect – via its effects on tumor cells. No statistically significant associations between any of the immunological parameters investigated and either objective response or tumor shrinkage were found (p>0.12 in all cases, data not shown). Similarly, even those patients whose tumors progressed during the course of treatment saw a decline in MDSC with sunitinib treatment. In parallel, our ongoing mouse experiments have shown a reduction in MDSC in tumor models such as 4T1 and CT26 where virtually no direct anti-tumor effects were evident (Ko JS et al. manuscript under review). These findings are consistent with the possibility that sunitinib directly suppresses MDSC independently of its anti-tumor impact. However, the possibility that sunitinib induces specific, functional changes in
73 tumor makeup which in turn affect MDSC, without inducing an apparent change in tumor
size, is currently under investigation.
It’s been shown that several tumor-produced growth factors targeted by sunitinib
are implicated in MDSC accumulation. Continuous VEGF infusion in naïve mice induced MDSC formation via VEGFR2 signaling (140) in one report, while MDSC
accumulation in mice bearing colon tumors has been attributed to stem cell factor (SCF)
(102). Indeed, experiments ongoing in our lab, as well as those published have identified
a portion of MDSC which express VEGFR1 and VEGFR2 (S.George1 (95). Sunitinib-
targeted receptors include all of those implicated thus far (VEGFR and c-kit), among
others such as PDGF, Flt-3, and CSF-1 receptor whose mechanistic importance cannot be
ruled out (219).
The accumulation of MDSC and Treg, as well as the suppression of T cell IFNγ in
metastatic RCC patients is consistent with previous reports (108, 241, 247). Here we
show that MDSC elevation as well as IFNγ suppression can be reversed in response to sunitinib, a drug with unparalleled activity in mRCC. The generation of an effective anti- tumor adaptive immune response requires the elimination of MDSC that likely initiate several T cell deficits (52). The inclusion of clinical treatments aimed at MDSC removal may be an important part of future immunotherapeutic protocols. These data further the rationale for sunitinib-based combination therapy with immunomodulators to enhance the anti-tumor effect and the effect on patient survival.
1 George S, Rayman P, Biswas S, Smith-Williams T, Ko J, Wood L, Elson P, Rini B, Bukowski R, Finke J. “Expression of Flt-3 and VEGFR1 on myeloid derived suppressor cells in renal cell carcinoma.” ASCO, Chicago, IL 2008.
74 CHAPTER 3
DIRECT AND DIFFERENTIAL SUPPRESSION OF MYELOID-DERIVED
SUPPRESSOR CELL SUBSETS BY SUNITINIB IS COMPARTMENTALLY
CONSTRAINED
Abstract
The anti-angiogenic drug sunitinib is a receptor tyrosine-kinase (RTK) inhibitor with significant, yet not curative, therapeutic impacts in metastatic renal cell carcinoma
(mRCC). Sunitinib is also an immunomodulator, potently reversing myeloid-derived suppressor cell (MDSC) accumulation and T-cell inhibition in the peripheral blood even of non-responder RCC patients. We observed that sunitinib similarly prevented massive
MDSC accumulation and restored normal T-cell function in the spleens of tumor-bearing
mice, even when sunitinib had negligible impacts upon tumor progression itself. Both
monocytic and neutrophilic splenic MDSC were highly inhibitable by sunitinib. In
contrast, MDSC in bone marrow and tumor proved highly resistant to sunitinib, and
ambient T-cell function remained suppressed in these compartments. Proteomic analyses comparing tumor to peripheral compartments demonstrated that GM-CSF predicted sunitinib resistance, and recombinant GM-CSF could confer sunitinib resistance to
MDSC in vitro. MDSC conditioning with GM-CSF inhibited STAT3 and promoted
STAT5 activation, whereas other hematopoietic support factors preferentially preserved both STAT3 activation and sunitinib susceptibility. We conclude that compartmental- dependent GM-CSF exposure may account for sunitinib’s regionalized impact upon host
MDSC modulation, and hypothesize that ancillary strategies to decrease such
75 regionalization will enhance sunitinib’s potency both as an immunomodulator and as a
cancer therapy.
Introduction
Angiogenesis is a crucial step in tumor progression; hence the rational development of several angiogenesis inhibitors. One such drug, sunitinib (Sutent, Pfizer), which inhibits signaling through RTKs including VEGFR1-3, PDGFR, ckit, flt3, and m-CSF receptors, is typically first-line therapy for metastatic renal cell carcinoma (RCC) (217-220). Yet all patients eventually progress on sunitinib, and this progression is thought to reflect an adaptive tumor resistance to angiogenesis inhibition (13, 250, 251). While the basis of such resistance is incompletely understood, accumulating evidence suggests that host myeloid-derived suppressor cells (MDSC) are recruited by tumors to mediate resistance to anti-angiogenic drugs (17, 135). Indeed, MDSC themselves can promote angiogenesis
(15, 16, 133, 134), in addition to serving as pivotal agents of tumor-induced T2-type biasing and the escape from cell-mediated immunity (106, 149, 252).
We recently demonstrated that sunitinib therapy significantly reversed RCC-induced
MDSC accumulation in patients’ peripheral blood, correlating with improvements in peripheral T-cell function (51). Sunitinib inhibited peripheral accumulation of all currently identified CD33+ human-MDSC subsets, including immature (lineage
negative), neutrophilic (CD15+CD14-), and monocytic (CD14+HLADR-/dim) MDSC (51,
108, 114-117, 224, 253). Such pan-inhibition was unlikely to be mediated solely through
an indirect anti-tumor effect, since sunitinib therapy inhibited peripheral MDSC
76 accumulation even when RCC patients displayed tumor progression (i.e., sunitinib
treatment failure) (51). We therefore investigated whether sunitinib’s remarkable impact
on peripheral MDSC was also evident when animal tumors were treated with sunitinib,
and whether sunitinib’s seemingly decisive negative modulation of MDSC extended into the tumor compartment.
We here report that sunitinib is uniformly effective for suppressing peripheral
(splenic) CD11b+Gr-1+ MDSC accumulation in all tested tumor models, despite widely
ranging anti-tumor effects among these models. Sunitinib inhibited proliferation of Ly6G- monocytic (m-) MDSC and also impaired survival of Ly6G+ neutrophilic (n-) MDSC.
We also observed, however, that sunitinib’s dramatic abrogation of splenic MDSC and normalization of bystander T cell function did not extend to either tumor or bone marrow
(BM) compartments. MDSC resistance to sunitinib corresponded to compartmental availability of GM-CSF, and recombinant GM-CSF itself conferred sunitinib resistance.
Paralleling our prior observations for dendritic cell differentiation, (193) superimposed exposure to GM-CSF dominantly preempted STAT3-activation in MDSC with steady- state STAT5 activation, providing a likely escape mechanism from STAT3-dependent sunitinib toxicity (221). Such regional disparities may explain how residual intratumoral
MDSC contribute to anti-angiogenic resistance despite pronounced drug-mediated declines in peripheral MDSC.
Materials and Methods
Mice, tumors, and treatment. Experiments performed under institutionally approved animal research committee protocols adhering to USDA guidelines. Female BALB/c mice from NCI, Frederick, MD were maintained pathogen-free (USDA) and studied at 8-
77 12wks; with 4T1-mammary, CT26-colonic, and RENCA-renal carcinomas syngeneic to
BALB/c mice maintained and injected into mice as previously described (193). Sunitinib
40mg/kg/day i.p. was initiated for 9 days total, except where indicated otherwise, after tumors reached 7mm diameter. Intraperitoneal treatment yielded the same MDSC reductions as oral treatment. Spleens, BM and tumors were processed as previously described (193).
Reagents. Culture medium consisted of RPMI-1640+10%FCS and conventional additives (51, 254). Pan-T-cell isolation kit and magnetic beads conjugated to anti-
CD11b, -FITC, or -APC were from Miltenyi Biotec (Auburn, CA). Functional grade anti- mouse CD3e and anti-CD28 were from BD Pharmingen (San Jose, CA). rmflt3L, rmSCF, rmIL-6, rmG-CSF, rmGM-CSF were from Peprotech (Rocky Hill, NJ). CFSE was from
Invitrogen (San Diego, CA). Luminex Bioplex Mouse Cytokine Array was from Biorad
(Hercules, CA). Proteome Profile Array was from RnD Systems (Minneapolis, MN).
Fluorochrome-coupled antibodies were all purchased from either ebiosciences (San
Diego, CA) or BD Biosciences (San Jose, CA). FACS data was collected on
FACSCalibur (BD) and analyzed using Cellquest software (BD) or FlowJo software
(Tree Star, Ashland, OR).
Determination of T-cell response. IFN-gamma production was assayed following polyclonal stimulation as described previously (51). Proliferation was assayed with either
CFSE dilutions or tritiated thymidine incorporation as previously described (193, 255).
In vivo proliferation and viability assays. One hour following sunitinib treatment, mice were injected i.p. with 1mg of sterile BrdU/mouse. Four hours following BrdU injection, cells obtained from spleen, BM and in some cases tumor and blood were stained
78 according to the BrdU-APC kit (BD). AnnexinV staining was done as previously
described (51).
MDSC subset isolation and staining. Splenic cell suspensions from 4T1+ mice were
depleted of T-cells, B cells, macrophages, and dendritic cells using APC-labeled antibodies to CD3, CD4, CD8, and MHCII and anti-APC magnetic beads. Remaining
cells contained mostly MDSC, and n-MDSC were isolated using anti-Ly6G-FITC and
anti-FITC beads. m-MDSC were isolated from remaining cells using anti-CD11b
magnetic beads. LS magnetic columns from Miltenyi were used. Cells were then stained
for FACS analysis or were adhered to glass slides using a cytospin at 60 rpm for 5
minutes. Adherent cells were then air dried and subsequently stained using a Fisher
Diagnostics Hema 3 Manual Staining Kit which is comparable to the Wright-Giemsa
method.
Cytokine analysis using proteome profile array. Blood was obtained via cardiac
puncture and equal volumes were added to ETDA to prevent clotting. Cell-free plasma
was obtained following centrifugation at 4°C at 13,000 for 10 minutes and frozen at -
80°C. 580uL of pooled plasma was added to each membrane in the array. SuperSignal
West Femto Maximum Sensitivity Substrate from ThermoScientific was used for
development. Additionally, lysates from frozen tumor tissue was obtained using the lysis
buffer recipe provided in the instruction manual with protease (ThermoScientific) and
phosphatase inhibitors (Pierce) added. Protein was assayed using the Biorad DC Protein
Assay kit according to the manual. 150ug of protein was added to each membrane.
Cytokine analysis using Bioplex array. Plasma was obtained as above and pooled.
Spleen and tumor tissue was immediately frozen in eppendorf tubes on dry ice. Lysates
79 were obtained from thawed, sonicated tissue in the same lysis buffer as above. Protein
was quantified as above. Plasma was assayed according to the Bio-Plex Pro custom assay
(Biorad). Tissue lysates were brought to 1mg/mL using 1xPBS with 0.5% BSA, and
50uL added to each well. A separate standard curve was made using lysis buffer as
diluent for the analysis of cell lysates. Data was acquired and analyzed on a Luminex
device using the Biorad Bio-Plex System and Bio-Plex Manager 3.0 software.
pSTAT signaling analysis in MDSC. BM-derived MDSC were cultured in stem cell
factor+flt3L +/- either IL-6, G-CSF, or GM-CSF and +/- sunitinib. pStat staining was
done according to the BD Phosflow Fix Buffer I and Perm Buffer III products. Briefly,
culture plates were spun down and supernatants removed. Fix buffer I was added to wells
for 10min at 37°. Washed cells were surface stained (CD11b and Gr1),
fixed/permeabilized with ice cold Perm Buffer III for 30 minutes on ice, then Fc blocked
with 10% heat-inactivated rat serum, rat IgG, and anti-CD16/32 antibodies in FACS
buffer prior to staining. Anti-pStat antibodies and isotypes were Alexa-Fluor-647 conjugated.
Analysis of human RCC tumor tissue. Freshly explanted clear-cell RCC tumors were digested. A portion of cells were stained for MDSC. The other portion was cultured at 1 million lymphocytes/1mL in complete RPMI media and stimulated for IFNγ production as above. Some tumors (RCC- 0885, 0803, 0958, and 3081427) were cultured short-term
(1-2 passages) and GM-CSF was quantified in the resultant cell-free supernatants, or in supernatants derived from the long-term clear-cell RCC line, SK-RC26b, (previously acquired from Dr. Neil Bander at Cornell Medical Center (256)) in the presence or absence of sunitinib.
80 Biostatistics. Unless indicated otherwise, all experiments were completed at least 3 times, with at least 5 mice per group. Results were pooled and the mean +/- standard errors were expressed graphically as columns. Treatment groups were compared using t- test for two samples assuming equal variances. A 2-tailed p-value less than 0.05 was deemed significant.
Results
Sunitinib reverses MDSC-mediated immune suppression in mice bearing renal and non-renal tumors. We first studied sunitinib’s ability to act as a broad immunopotentiator in several mouse tumor models. Treatment of either Renca-kidney,
CT26-colon or 4T1-breast tumor-bearing mice, or even naïve mice, with sunitinib
(40mg/kg) daily for 6-9 days significantly reduced both the percentage and total numbers of CD11b+Gr1+ MDSC detected in the spleen (Figure 1A). Such reductions were associated with significant disinhibition of T-cells which were otherwise suppressed in the tumor-bearing state. T-cells from 4T1 tumor-bearing mice were less able to produce
IFNγ in response to polyclonal stimulation with anti-CD3/28 when compared to naïve, non-tumor bearing mice. Such T-cell suppression was reversible with either in vivo
MDSC depletion using sunitinib, or in vitro MDSC depletion using anti-Gr-1 magnetic beads; finally, bead-isolated MDSC could be introduced to suppress T-cells from naïve mice as well (Figure 1B). Similar findings occurred in Renca and CT26 tumor-bearing mice (data not shown). Deficits in T-cell proliferation were also attributable to the presence of MDSC (Figure 1C), and CD4+ and CD8+ T-cells regained function when
MDSC were depleted with sunitinib or mechanically (Figure 1D). These findings suggest
81 that MDSC are major mediators of T-cell suppression in tumor-bearing hosts, and that suppressed T-cells can be rendered functional if activated in the absence of MDSC.
82 Sunitinib inhibits intrasplenic proliferation of m-MDSC. We next asked whether
sunitinib functions to inhibit MDSC expansion in vivo. The proliferative rate of MDSC
was kinetically quantified by the percentage of BrdU+ cells. Within the MDSC population, cells brightest for Gr1 staining (Gr1hi) were relatively non-proliferative in
naïve (pink squares, <10%BrdU+) and tumor-bearing mice (green squares and triangles,
10-20%BrdU+). This was echoed in BM (BM) Gr1hi-MDSC (not shown). In contrast,
MDSC which were dim for Gr1 (Gr1lo) were very proliferative in the spleens (and BM) of tumor-bearing mice (blue squares and triangles, ~50%BrdU+ (BM not shown))
compared to naïve mice, where proliferation was only present in the BM (not shown),
and not in spleen (red circles, <10%BrdU+) (p<.0002, Figure 2A), indicating pronounced
proliferative pathology of splenic Gr1lo MDSC in the tumor-bearing state. Sunitinib was able to reverse this pathology by significantly inhibiting the expansion of Gr1lo MDSC on
day 6 following treatment (blue triangles, p<.02, Figure 2A).
These findings suggested that MDSC could be functionally divided into two groups
based on their proliferative potential in vivo. Upon isolation, as previously reported (112,
113) Ly6G+ n-MDSC were indeed Gr1hi, Ly6G+, and F4/80-, and displayed early or
completed polymorphonuclear features on cytospins (Figure 2B), consistent with their lower rate of proliferation. Ly6G- m-MDSC were Gr1lo, Ly6G- and evenly distributed in
regards to F4/80 expression and monocytic versus immature morphology on cytospins
(Figure 2B), consistent which their higher rate of proliferation. Thus, splenic m-MDSC
were highly proliferative compared to n-MDSC, and sunitinib inhibited this proliferation,
significantly contributing to the overall reduction in splenic MDSC accumulation.
83 Sunitinib impairs the viability of splenic n-MDSC. We found over half the Gr1hi n-
MDSC in the spleens of naïve mice were undergoing apoptosis, consistent with their being normal neutrophils with a rapid rate of turnover in naïve animals (Figure 2C). The rate of n-MDSC apoptosis in tumor-bearing mice was significantly reduced compared to naïve, indicating that they have a prolonged lifespan in vivo. Sunitinib significantly reduced the viability of splenic n-MDSC in vivo on day 6 following treatment (p<.00005,
Figure 2C). In contrast, m-MDSC generally had lower rates of apoptosis and these were only modestly affected by sunitinib in vivo (naive= 25%, 4T1=15%, 4T1+Sunitinib=32% on average, data not shown).
84
85 Compared to spleen, intratumoral MDSC are sunitinib-resistant. In contrast to the
global effect of sunitinib to improve splenic T-cell function via reductions in MDSC, the
drug’s effect on tumor growth varied considerably. Renca kidney tumors were
disproportionately sensitive to sunitinib, and tumors enzymatically digested at the end of
the treatment period yielded no viable cells; nonetheless, tumors eventually progressed
following discontinuation of drug, indicating that the mice had not been cured of disease
(Figure 3A-top). In contrast, 4T1 mammary tumors continued to progress, albeit more slowly, during sunitinib treatment, and tumors digested at the end of treatment contained viable tissue. Furthermore, 4T1 immediately resumed normal growth kinetics with
sunitinib discontinuation (Figure 3A-bottom). Finally, CT26 displayed negligible growth
inhibition during sunitinib treatment (not shown), despite the drug’s ability to abrogate
splenic MDSC accumulation even in this model.
We examined sunitinib’s impact upon cellular tumor constituents which remained
viable during treatment, focusing predominantly on 4T1. Compared to splenic MDSC,
there was a much more modest decline in tumor-associated MDSC (Figure 3B) in
response to drug. Tumor-infiltrating T-cells from sunitinib-treated mice did not function significantly better than those from untreated mice, in marked contrast to the improved splenic T-cell function obtainable from the same mice (Figure 3C). Indeed, intratumoral
MDSC from sunitinib-treated mice retained T-cell suppressive capabilities equivalent to those from untreated mice (Figure 3D). Our findings suggest that even when sunitinib broadly reverses peripheral MDSC accumulation (51), intratumoral MDSC are significantly less impacted.
86 Similarly to tumor MDSC, MDSC in BM resist sunitinib. The BM is the originating repository of myeloid, stem and progenitor cells; and 30-40% of BM cells are
CD11b+Gr1+, even in naïve mice. This frequency rises in all tumor models tested thus far
(4T1, Renca, and CT26), and reaches ~80% in the 4T1 model, so sunitinib’s primary site of MDSC inhibition might logically be expected to be BM. Nevertheless, we observed no sunitinib-induced decrease in the % of CD11b+Gr1+ MDSC in the BM of any treatment group tested (naïve, 4T1+, Renca, and CT26, Figure 2D and data not shown). Consistent with this, we did not see a sustained anti-proliferative effect of sunitinib on MDSC residing in the BM compartment (data not shown). No BrdU+ cells appeared in the blood of untreated or drug-treated mice after a 4 hour BrdU pulse, ruling out the possibility that spleen and BM results were confounded by variations in the rate of MDSC egress (data not shown).
87
88 GM-CSF uniquely protects MDSC in the presence of sunitinib in vitro, possibly via
preemption of STAT3. Despite their disparate susceptibilities to in vivo sunitinib
exposure, purified splenic-, tumor- and BM-derived MDSC proved equally sensitive to
sunitinib-mediated apoptosis in culture (not shown), prompting us to investigate
regionally produced factors which could account for in vivo disparities. Because several
host- and tumor-produced factors such as stem cell factor (SCF), IL-6, G-CSF, and GM-
CSF have been implicated in MDSC accumulation (92), we cultured fresh BM containing
MDSC in the presence of SCF and flt3 ligand to provide basal support, with the further
addition of either IL-6, G-CSF, or GM-CSF +/- sunitinib. Sunitinib significantly impaired
the viability of CD11b+Gr1+ MDSC, but GM-CSF, and to a lesser extent G-CSF, could
partially restore MDSC viability when added to cultures (Figures 4A-B). Indeed,
10ng/mL of GM-CSF protected MDSC significantly better than G-CSF added at
100ng/mL. In addition, G-CSF and GM-CSF, when added to SCF and Flt3L, induced the
strongest proliferative response in these cultures; and, while sunitinib totally prevented
MDSC expansion in the presence of G-CSF, it had a relatively modest effect in the
presence of GM-CSF (Figure 4C).
To test whether a particular signaling profile could distinguish sunitinib-sensitive
from sunitinib-insensitive cultures, we performed intracellular staining for steady state
levels of pY705STAT3 and pY694STAT5 expression in MDSC cultured in non-
protective (SCF,Flt3L, +/- IL-6 or G-CSF) or protective (SCF,Flt3L, + GM-CSF) conditions. Consistent with our prior studies (193), MDSC maintained in the absence of
GM-CSF displayed a STAT3-driven signaling signature (Figure 4D), which was repressible with sunitinib in vitro. In contrast, also consistent with our prior studies,
89 addition of GM-CSF dominantly suppressed pY705STAT3 and activated pSTAT5
(Figure 4D). Cultures which were STAT3-driven showed sensitivity to sunitinib, and this resulted in downregulation of pSTAT3. Meanwhile, cultures which were not STAT3- driven were insensitive to sunitinib, and ultimately unchanged in their signaling pattern.
90
91 GM-CSF is selectively expressed in the tumor microenvironment in vivo. Because GM-
CSF was unique in its protection of MDSC in vitro, we tested its relative levels in peripheral versus tumor compartments. Mouse cytokine proteome profile arrays and
Luminex assays of untreated and sunitinib-treated 4T1+ mice showed that GM-CSF was not present at detectible levels in the plasma or spleens of these mice, but was readily detected in the tumor bed. In contrast, G-CSF, among others, was very readily detectible peripherally (Figures 5A-B). The preferential localization of GM-CSF to the tumor bed was consistent with its unique ability to protect MDSC from the effects of sunitinib in vitro, and with the persistence of MDSC in 4T1 tumors during sunitinib treatment.
92
93 Tumor microenvironment limits sunitinib’s anti-MDSC effect in RCC patients. We
next examined RCC patients’ tumor specimens (described in Table 3-1). Three subsets of
CD33+HLADR- MDSC were identifiable: CD33hi patient MDSC were CD15- and
mononuclear, and included equal numbers of CD14+ and lineage negative subsets (Figure
6A). CD33lo patient MDSC were CD15+ and had polymorphonuclear features (Figure
6A, cytospins not shown). This neutrophilic MDSC subset, as in the mouse tumor model, was the most prevalent type observed. Figure 6B shows the average percentage of each
MDSC subset detected in tumors from untreated patients, as well as the amount detected in two tumors from sunitinib-treated patients. Unlike the pronounced decline in peripheral blood MDSC observed in RCC patients treated with sunitinib (51), tumors obtained from sunitinib-treated patients have not demonstrated declines in MDSC. The continuing presence of MDSC in the tumors of sunitinib-treated patients was associated with continuing T-cell suppression, as measured by IFNγ production, compared to normal donor T-cells and T-cells from untreated tumors (Figure 6C). In addition, several short-term tumor cell lines derived from Cleveland Clinic surgical patients’ RCC tumors
(1-2 passages) produced abundant amounts of GM-CSF in vitro, and this production was not significantly reduced in the presence of sunitinib (Figure 6D). This suggests that compartmental constraints limit sunitinib’s ability to act as an immunopotentiator, and that locally produced GM-CSF may promote local sunitinib resistance.
94
Supplementary Table 3-1- Patient Characteristics (All Tumors2)
Total Tumors3 37 (100%)
Gender Male 24 (65%) Female 13 (35%)
Age at Tumor Removal Mean + s.d. 60.1 + 10.6 Median 61 Range 32 – 78
Tumor Spread Localized 25 (68%) Metastatic4 12 (32%)
Tumor Size (cm) Mean + s.d. 7.4 + 3.6 Median 6 Range 2.3 – 15
Fuhrman Nuclear Grade 2/4 15 (43%) 3/4 11 (30%) 4/4 9 (27%)
2 All donors signed an IRB-approved, written informed consent. All Tumors were Clear Cell RCC excepting 1 Papillary and 1 Onchocytoma. 3 21 tumors used in Figure 6B and 24 tumors used in Figure 6C with several tumors measured for both MDSC and IFNγ and included in both graphs. 4 All tissues received were from kidney excepting 1 femur and 1 lymph node (LN) for which a Grade is not included. Sites of metastasis at time of surgery included LN-5, adrenal-3, kidney-2, bone-2, and liver, stomach, and lung – 1 each. No previous treatment had been given excepting 4 patients who received prior sunitinib as indicated in the related Figure 6. Two patients had a previous nephrectomy – one of which also received prior sunitinib.
95
96 Discussion
While the role of immunosurveillance to inhibit tumor progression is well documented, there are many obstacles which prevent the destruction of advanced tumors by immune cells, one of which is MDSC accumulation. Sunitinib targets several RTKs and has experienced relative clinical success, especially in the setting of mRCC (218).
We recently reported sunitinib’s ability to reverse immune suppression in mRCC patients via MDSC inhibition (51), yet sunitinib’s broad potential to modulate immune function independently of its anti-tumor effect was unknown. We here report that sunitinib can inhibit MDSC accumulation, and thereby restore normal T-cell function, in the spleens of mice bearing both sunitinib-sensitive and sunitinib-insensitive tumors. This suggests that sunitinib’s immunomodulatory activities occur independently of its anti-tumor potency.
As such, sunitinib may prove a useful adjunct agent in immunotherapy trials (220).
The ability of sunitinib to normalize splenic MDSC even in the 4T1 model, where an enormous degree of MDSC accumulation occurs, confirms sunitinib’s pronounced anti-
MDSC activities. In the present report we demonstrate two mechanisms by which sunitinib inhibits MDSC accumulation. Pilot studies identified the spleen and BM as sites of MDSC proliferation, with BrdU-positive cells appearing after only a 1hour BrdU pulse. Using in vivo BrdU administration, Gr1lo m-MDSC were found to proliferate at a rapid rate in the BM of naïve and tumor-bearing mice, and in the spleens of tumor- bearing mice, suggesting that the extramedullary proliferation is the most pathological, because it does not normally occur in naïve mice under steady-state conditions. Sunitinib had an anti-proliferative effect on splenic m-MDSC in vivo, and this effect could be duplicated in vitro.
97 The anti-proliferative effect of sunitinib is unlikely to account solely for the dramatic
declines seen in MDSC in response to drug. Because previous studies had shown
sunitinib to have a toxic effect on RCC patient MDSC in vitro (51), its effect on Gr1hi, neutrophilic MDSC (n-MDSC) viability was also studied in vivo. Our findings show that
Gr1hi n-MDSC were relatively non-proliferative, and suggest that the accumulation of
this MDSC subset in tumor-bearing mice is more related to an abnormally prolonged
lifespan. Indeed, n-MDSC from tumor-bearing mice were much less likely to be
undergoing apoptosis at any given time, compared to naïve mice. Sunitinib seemed to
normalize this in vivo by increasing the frequency of n-MDSC apoptosis, as measured by
AnnexinV staining, an effect which was reproducible in vitro. The data thus show that sunitinib acts to inhibit both the abnormal expansion of m-MDSC and the abnormally extended survival of n-MDSC. An alternative interpretation is that sunitinib prevents the expansion of mononuclear MDSC which may, in part, represent n-MDSC precursors
(257). Such activity alone could lead to the perceived increase in n-MDSC apoptosis, because terminal n-MDSC would eventually undergo spontaneous apoptosis and fail to be replaced. This possibility is currently being tested.
In addition to being MDSC subset-specific, our data strongly suggest that sunitinib’s
inhibition of MDSC is direct. Proteome profile and luminex arrays did not identify a
decline in any of the factors implicated in MDSC expansion or activation. In fact, several cytokines such as G-CSF and IFNγ were present at increased amounts in sunitinib treated mice, consistent with Kerbel et. al’s parallel observations in naïve mice (258).
The receptors involved in sunitinib’s anti-MDSC activity are currently under investigation. Putative RTK’s inhibited by sunitinib include targets of VEGF, SCF (ckit),
98 Flt3L (flt3), m-CSF and PDGF (218, 219, 259). SCF has recently been implicated in
MDSC accumulation (102), and sunitinib inhibits its receptor, ckit. However, culture with RTK ligands such as SCF and/or flt3L produced only limited MDSC expansion and viability unless combined with proliferatively synergizing cytokines such as IL-6, G-CSF or GM-CSF (193). This suggests that while the sunitinib RTK targets ckit and/or flt3 may provide permissive signals for MDSC expansion, tumor-promoted excessive MDSC accumulation requires synergy from additional cytokines.
G-CSF and GM-CSF have both previously been implicated in MDSC expansion (92,
99, 104, 116), and vaccines which produce higher concentrations of GM-CSF have been reported to be less effective as a consequence of promoting MDSC (104, 116).
Additionally, however, our studies show the divergent abilities of G-CSF and GM-CSF to confer resistance to sunitinib in vitro, paralleling the compartmental disparities of these cytokines in vivo. Because the Renca model is paradoxically far more vulnerable to sunitinib treatment than human RCC, we believe other models such as 4T1 are potentially more predictive of escape mechanisms relevant to human cancer. Notably, both mouse
4T1 and human RCCs displayed disparately greater GM-CSF production than sunitinib- sensitive peripheral compartments (plasma and spleen), whereas all compartments were rich in G-CSF. Although BM could not be promptly frozen and lysed in parallel with the other compartments analyzed, we hypothesize that a local abundance of GM-CSF accounts for the sunitinib resistance of BM MDSC.
We hypothesize that the striking ability of GM-CSF but not other pro-proliferative cytokines to promote sunitinib resistance is due to its dominant capacity to preempt
STAT3 programming in favor of STAT5 programming. Our previous studies showed
99 that treatment of CD34+ hematopoieitic progenitors with GM-CSF promoted STAT5
activation while dominantly blocking both Flt3L- and IL6-induced STAT3 activation
(24). Despite preventing STAT3 activation, GM-CSF continued to display proliferative
synergy with Flt3L and IL-6 which was undiminished even for STAT3 knockout BM
(193). We here demonstrate a similar pSTAT5-dominated signaling pathway for MDSC exposed to GM-CSF. Conversely, in the absence of GM-CSF, combinations of SCF,
Flt3L, IL-6 and/or G-CSF all promoted STAT3-activated MDSC which were highly
susceptible to inhibition by sunitinib. While previous studies have highlighted the
importance of STAT3 signaling in MDSC accumulation and function (89, 260), as well
as in emergency granulopoiesis (166), the present report is the first to demonstrate
generation of MDSC in a steady state of STAT5 rather than STAT3 activation, the
consequence of GM-CSF exposure. Because sunitinib’s capacity to inhibit MDSC was
recently shown to depend upon STAT3 blockade (221), we hypothesize that GM-CSF
treatment dominantly reprograms MDSC to function independently of STAT3, rendering
them resistant to sunitinib.
MDSC-mediated diminutions in T-cell-mediated immunity may contribute to the
currently limited effectiveness of immunotherapy in RCC and other tumors (31, 50-53).
Our previous studies in RCC patients showed that MDSC declines and improvements in
T-cell function were not contingent upon tumor shrinkage in response to sunitinib, and
even sunitinib-induced tumor cytoreduction is not associated with cure (51, 53). Our data
thus far obtained in human RCC patients supports the model established in 4T1 tumor-
bearing mice, where peripheral compartment reductions in MDSC ubiquitously occur in
response to sunitinib, but where the drug’s anti-MDSC activity is much less pronounced
100 intratumorally as a result of a relative abundance of non-sunitinib-targeted growth factors, such as GM-CSF, which provide alternative sunitinib-resistant survival signals.
We are investigating ancillary strategies to abrogate such regional resistance, in order to enhance sunitinib’s potency both as an immunomodulator and as a cancer therapy.
101 CHAPTER 4
Tumor-Derived Products Dynamically Activate CD15+ Myeloid-Derived
Suppressor Cells and are Partially Modulated by Sunitinib
Abstract
In human renal cell carcinoma (RCC) the increased accumulation of myeloid- derived suppressor cells (MDSC) is significantly reduced by treatment with sunitinib resulting in an improved type-1 T cell response. Here we show that of the different
MDSC subsets, neutrophilic MDSC (N-MDSC) are the most predominant in the blood and tumors of RCC patients. Soluble products present in supernatants derived from RCC lines induce N-MDSC formation/activation from healthy donor blood. The induced N-
MDSC which co-purify with mononuclear cells, are suppressive to T cell proliferation and interferon-gamma (IFNγ) production. They resemble “primed” neutrophils which have upregulated expression of CD11b and CD66b, and down regulated CD62L, and display a relatively lower density compared to normal neutrophils. The phenotype of induced N-MDSC was similar to the phenotype of CD15+ MDSC present in the tumor and blood of RCC patients. The potency of RCC supernatants to induce N-MDSC was reduced by approximately 50% when tumor cell lines were treated with a non-toxic dose of sunitinib. Additional studies show that GM-CSF may account for the induction of N-
MDSC that is not affected by sunitinib while gangliosides may account for the induction activity that is sensitive to sunitinib. These data explain a mechanism by which tumor- derived products induce the appearance of N-MDSC in patient blood and tumor. They also characterize the indirect (via actions on tumor cells) actions of sunitinib on N-MDSC
102 which occur via the modulation of gangliosides, and which are yet limited, via the
continued production of N-MDSC activating cytokines such as GM-CSF.
Introduction
Renal cell carcinoma (RCC) is a cancer distinguished by both poor outcome as
well as immunogenicity. It is known to be resistant to traditional chemotherapy and radiotherapy and to have a high propensity for recurrence (1, 261). Immunotherapy for the treatment of metastatic RCC has demonstrated a 15-20% response rate, while in a small minority of patients, treatment with interleukin-2 induced long-term survival (262).
The relatively weak response to immunotherapy by the majority of patients is likely related to tumor-induced suppression of T cell immunity (50, 120, 229, 263). Indeed, peripheral blood T cells from RCC patients display a diminished capacity to generate a type-1 T cell IFNγ response, which is considered critical for the development of an antitumor immune response (50, 51, 108). Tumor-induced immune suppression is a challenge which need be reversed before any immune-based treatment can succeed (31,
52). Recently the small molecule receptor tyrosine kinase inhibitor, sunitinib, which targets several receptors in the vascular endothelial growth factor receptor (VEGFR) family has become first-line treatment for patients with metatstatic RCC (mRCC). This drug has increased the response rate over that observed with cytokine therapy and has been relatively successful at extending patient survival, although it is not curative (216).
We have shown that treatment of RCC patients with sunitinib reverses T cell suppression and reduces the number of myeloid-derived suppressor cells (MDSC) (51, 53). The role of tumor-induced MDSC in such T cell suppression has been well documented (92, 101),
103 and these cells appear to play a particularly important role in the setting of RCC (51, 108,
120).
As currently understood, MDSC are a heterogeneous group of immature myeloid cells that suppress T cell function and that accumulate in tumor-bearing hosts as a result of tumor-induced myelopoietic expansion (96, 98, 239). In kidney cancer, two main groups of MDSC have been described and both subsets are positive for the myeloid markers CD11b, and CD33, and negative for the maturation marker MHCII or HLA-DR.
Within this umbrella, neutrophil-like CD15+CD14- MDSC from patients inhibit T cells
through the production of arginase and reactive oxygen species, and most likely
correspond to CD11b+Gr1+hiLy6G+Ly6Cint MDSC in the mouse tumor model (51, 108,
112, 113). Mononuclear lineage negative (Lin-) MDSC from patients inhibit T cells
through reactive intermediates, and likely correspond to CD11b+Gr1+loLy6GloLy6C+hi
MDSC in the mouse tumor model (114). In addition, CD14+ MDSC from melanoma and
hepatocellular carcinoma patients, which are also CD11b+CD33+HLADR- have been
shown to produce T cell suppressive cytokines and induce regulatory T cell formation;
and these cells likely correspond to CD11b+Gr1+loLy6GloF4/80+ and/or CD115+ MDSC described in the mouse tumor model (116, 117).
The critical components involved in MDSC accumulation in RCC patients have not been well addressed. Several tumor and host-produced factors are thought to play a role in the myelopoietic expansion of MDSC. Vascular endothelial growth factor
(VEGF) has been linked to MDSC accumulation in cancer patients as well as mice, and
RCC tumors and patient plasma are known to have an abundance of VEGF (140, 234,
264-266). Stem cell factor (SCF) was shown to be present in several human and mouse
104 tumors, and blockade of this factor reduced MDSC-mediated immune suppression in the mouse tumor model (102). The importance of these factors in MDSC accumulation in
RCC patients has been indirectly validated because the drug sunitinib, which inhibits
signaling through both VEGF and SCF receptors (VEGFR1/2 and ckit respectively)
among others, was shown to significantly reduce MDSC accumulation after 1 and 2
cycles of treatment (51). Because sunitinib targets many receptors (VEGFR, ckit, flt3,
M-CSFR, PDGFR), several of which are implicated in hematopoiesis, it may predictably
act to limit MDSC expansion. Yet, myeloid cell expansion is not likely to be the only
contributing factor to MDSC accumulation.
Here we have sought to understand whether tumor-produced factors can directly
activate MDSC and whether sunitinib has a role in inhibiting this activation which
mechanistically contributes to MDSC accumulation. The focus of this study was on
neutrophilic, N-MDSC (CD11b+CD14-CD15+) which have recently been identified as the
most prevalent MDSC subset in both patients (51, 108, 224, 253) as well as tumor-
bearing mice (113). Moreover, total neutrophil levels are significantly elevated in RCC
patients and are associated with poor prognosis (267-269). Here we show the novel
activation of neutrophilic MDSC (N-MDSC) from healthy donor whole blood (WB) by
RCC cell line-derived products present in tumor-conditioned media. Tumor-induced N-
MDSC were phenotypically activated when compared to normal neutrophils because they
were less dense, had upregulated CD66b and CD11b, and downregulated CD62L. This
activated phenotype was additionally mirrored in RCC patient peripheral blood and
tumor-derived N-MDSC. Similarly to patient-derived MDSC, in vitro activated N-
MDSC suppressed the proliferation and the interferon-gamma production of T cells.
105 Rather than undergoing rapid apoptosis following activation, viable N-MDSC could be maintained in vitro in the presence of tumor-conditioned media. GM-CSF was found to be a particular cytokine, present in RCC cell line supernatants, which could both activate
and maintain the viability of N-MDSC in vitro. While the pretreatment of whole blood
with sunitinib was unable to prevent the induction of MDSC by TCM, TCM derived from
sunitinib (0.1ug/mL)-treated tumor cells was less able to activate N-MDSC, yet provided
equal support for long-term viability of previously activated N-MDSC. This was
consistent with the fact that at the non-toxic dose of 0.1ug/mL, sunitinib did not modulate
tumor cell production of GM-CSF, but rather decreased tumor cell production of gangliosides, which also had the capability to potently activate N-MDSC, yet could not maintain their viability in vitro. The findings herein, describe a novel mechanism by which RCC induce N-MDSC and defines one of the mechanisms behind sunitinib- mediated MDSC reductions.
Materials and Methods
Patient Populations
Peripheral blood was collected from patients with metasatic RCC, all clear cell histology (n=7), and from 15 age-matched normal healthy donors. Tumor infiltrating
MDSC were obtained from 21 patients with localized clear cell RCC undergoing nephrectomy or partial nepherctory at the Cleveland Clinic. All patients and healthy volunteer blood donors signed an IRB-approved, written informed consent for collection of blood and/or tumor samples.
Reagents
106 RPMI media, Hanks Balanced Salt Solution without Calcium or Magnesium,
ammonium chloride, human IgG for blocking non-specific antibody binding, and
dimethyl sulfoxide (DMSO) were obtained from Sigma-Aldrich (St. Louis, MO). Ficoll-
Hypaque was from GE Healthcare (Upsala, Sweden). Fetal Bovine Serum was purchased
from Hyclone (Logan, UT). Anti-CD3 (OKT3,) and anti-CD28 antibodies obtained from
OrthoBiotech (Bridgewater, NJ) and BD Bioscience (San Jose, CA) respectively.
Unconjugated anti-human IFNγ, anti-human IFNγ-FITC or -APC, and anti-human- CD3,
CD4, CD14, CD15, CD33, HLA-DR, CD66b, or CD62L, as well as annexinV, 7AAD, and the annexinV staining kit were all from BD Biosciences (San Jose, California). Anti- human CD11b was from eBioscience (San Diego, CA). Golgi Plug, Fix Perm and Perm
Wash were part of an Intracellular Cytokine Staining Kit from BD Biosciences. Mouse
IgM kappa-APC, mouse IgG1 kappa-PECy5, mouse IgG2a kappa-APC, mouse IgG2a-
PerCP, Anti-rat IgG2a FITC isotype and mouse IgM kappa-PE were from BD Biosience.
Mouse IgM kappa-FITC and mouse IgG1 kappa-PE were from eBioscience. 3H-
Thymidine was purchased from PerkinElmer, Boston, MA). Human GM-CSF, IL-6, IL-8
and G-CSF was obtained from RnD Systems (Minneapolis, MN). Anti-GM-CSF
neutralizing antibody mIgG1 was from Abcam and mIgG1 isotype control was from BD.
Peripheral blood mononuclear and T cell isolation
Peripheral blood was centrifuged over a Ficoll-Hypaque density gradient
(Amersham Pharmacia Biotech AB, Uppsala, Sweden) to obtain total leukocytes. For the suppression assay, T cells were purified by negative magnetic selection using microbeads coated with antibodies to CD14 (monocytes/macrophages), CD16 (NK cells), CD19 (B cells), CD56 (NK cells) and glycophorin A (RBCs) (Stem Cell Technologies, Vancouver,
107 Canada). The T-cell isolation procedure yielded cells that were more than 95% positive
for CD3 as defined by immunocytometry.
Neutrophil isolation
Peripheral blood was centrifuged over a Ficoll-Hypaque density gradient in 50mL conical
tubes. The top 1-2mL of several RBC pellets were collected and combined to a total
volume of 10mL and added to 12mL of a 6% dextran in 0.9%NaCl solution. The tube
was filled to 50mL with Hanks BSS without calcium or magnesium to give a final
concentration of 1.5% dextran. The tube was inverted several times to mix and left for 1
hour protected from light. The supernatant was collected, and pelleted, and remaining
RBCs lysed with hypotonic NH4Cl.
Ex vivo processing of tumors and staining of MDSC
Redundant tissue from RCC tumors was digested with Collagenase, DNAse and
Hylaronidase(all from Sigma), for 1hr and mashed through sieve to obtain a single cell
suspension. Cells were then stained with anti-human- HLA-DR FITC, CD15 PE, CD14
PerCP and CD33 APC in FACS buffer containing 2% heat-inactivated FCS following Fc
block with human IgG. All samples were acquired on a BD FACSCalibur and analyzed
using CellQuest software.
Preparation of tumor conditioned media
Established RCC cell lines, SK-RC-7 and SK-RC-26b were obtained from Dr.
Neil Bander (The New York Hospital, Cornell University Medical College, NY, NY) while line 0827LM was generated at the Cleveland Clinic. Additionally, short term
cultures (1-2 passages) were generated from freshly resected RCC tissue. Tumor cells
were seeded and maintained in complete RPMI 1640 (10% fetal bovine serum) medium
108 at 37ºC with 5% CO2 and were allowed to reach confluence in 6 well plates. Tumor condition media was obtained by collecting the supernatants from confluent tumor cultures treated or not with 0.1ug/ml Sunitinib for 6 days. All TCM were centrifuged at
2000rpm for 10 minutes to remove any cells. The supernatants were stored in aliquots at -
20oC. In some experiments, TCM was treated for 30 minutes with either neutralizing antibody to GM-CSF or the isotype control antibody ( both at 1ug/mL) prior to the addition of each TCM to whole blood for N-MDSC activation.
Induction of N-MDSC
Cell free, tumor-conditioned media (TCM) or complete RMPI media (as a control) was added at 10%v/v to healthy-donor whole blood for one hour at room temperature (RT). Blood was then Ficolled for PBMC isolation as above. Induced N-
MDSC were identified in the PBMC layer. In some cases, induced N-MDSC were isolated using anti-human CD15 antibody and either FACS cell sorting, or Miltenyi magnetic beads and columns.
T cell suppression assays
Suppression of T cell proliferation was performed using mixed-lymphocyte- reactions (MLR). T cells were isolated from the same donor blood as the induced MDSC for the T-effectors. WB was incubated with TCM for 1 hr, ficolled, stained with anti- human- CD33 APC and HLA-Dr FITC and then sorted on a BD FACSAria. The
CD33+HLA-DR- cells were incubated with Teffectors at a 1:1 and 1:10 ratio along with
1 hr adherent allogeneic PBMC which had been irradiated at 5000 Rad prior to usage as stimulator cells. After 6 days, 1uCi 3H-thymidine was added to each well and then harvested after 18 hr.
109 The suppression of intracellular levels of IFNγ by MDSC was determined by stimulation PBMC isolated from WB incubated with tumor supernatant or not, with anti-
CD3/anti-CD28 antibodies for 3 days. Golgi plug was added to cells for 6 hours, and then cells were harvested and stained for FACS analysis. Non-stimulated cells from each
donor served as a negative control. Additionally, specificity of cytokine staining was
confirmed in each sample via subtraction of any non-specific staining occurring in
samples pretreated with unlabeled anti-cytokine antibodies prior to the addition of
fluorochrome-labeled antibodies.
Assay of tumor conditioned media for growth factors
Cell free TCM from sunitinib-treated and untreated tumor cells were frozen at -
80C and semiquantitative analysis of numerous cytokines, chemokines, and growth
factors were done according to the instructions provided in the human cytokine and human angiogenesis proteome profile arrays (RnD Systems). Analysis was done using
NIH image J software.
Ganglioside isolation and quantification
Gangliosides were isolated from tumor cells as described before (55) with minor modifications. Briefly, extraction of gangliosides was performed with chloroform- methanol (1:1) for 18h at 4ºC followed by partitioning in 10ml Diisopropyl ether/1-
Butanol/0.1% aqueous NaCl. The lyophilized, final aqueous phase was passed through a
Sephadex G-25 column to remove the salts and small molecular weight impurities.
Isolated gangliosides eluting in the void volume were lyophilized and dissolved in chloroform-methanol. Gangliosides were quantified using a lipid-bound sialic acid
(LSA) assay as previously described (55). Briefly, lipids were extracted using a 2:1 ratio
110 of ice-cold chloroform-methanol and the lipid extract was subsequently separated with cold distilled water. The LSA content of the aqueous phase was quantified using resorcinol-HCL, by comparing the absorption at 580nm with the standard curve generated using known amounts of free sialic acid (n-acetyl-neuraminic acid). (34)
Statistical analysis
Statistical comparisons between treatment groups were done using two-tailed t- tests for 2 samples assuming equal variances. Differences in values were considered significant if the p-value was <.05.
Results
Neutrophilic MDSC are the most prevalent MDSC subtype in renal cell carcinoma tumors.
We have previously shown that MDSC accumulate in the peripheral blood of
RCC patients and suppress patient T cell function. Importantly, the majority of the
MDSC detected were neutrophilic CD33+dimHLADR-CD15+CD14- MDSC (N-MDSC)
(51). Here we analyzed fresh tumor specimens from 21 RCC patients for the presence of
MDSC. Figure 1A outlines the analysis undertaken. Within the CD33+HLADR- MDSC present in the tumor, all three previously reported subpopulations of MDSC were present
– neutrophilic, monocytic, and lineage negative MDSC (108, 114, 116). The majority of the CD33+lo cells were also CD15+CD14- and were neutrophilic N-MDSC. These cells were over 6% of cells falling within the scatter-based myelo-gated events, and were 60% of tumor-associated MDSC (Figure 1B). We also observed two MDSC subsets within the CD33+hi cells, one which was CD14+ and thus monocytic (M-MDSC), as well as one which was CD14-, and therefore lineage negative and precursor type (Lin-MDSC)
111 (Figure 1A). The Lin-MDSC were the second most common MDSC subset found in the
tumor bed, and represented over 1% of cells falling within the myelogate, and over 20%
of tumor-associated MDSC (Figure 1B). Finally, the M-MDSC were the third most
common MDSC subset found in RCC tumors, and represented 1% of cells falling within
the myelogate and over 10% of tumor-associated MDSC (Figure 1B) Because
CD33+loHLADR-CD15+CD14- MDSC which are neutrophilic (N-MDSC) represented the most prevalent MDSC subset in the blood and tumors of RCC patients, we focused
further efforts to develop a model for the evaluation of tumor induction of N-MDSC.
112
113 Soluble tumor-derived products activate neutrophilic MDSC from healthy donor whole
blood.
It has been shown in the mouse tumor model that neutrophilic MDSC can be
differentiated from normal neutrophils primarily by their lower density, which allows
them to copurify with mononuclear cells (112). In cancer patients, CD15+ granulocytes were shown to abnormally co-purify with mononuclear cells and suppress T cell function, an effect that could be reproduced in healthy donor whole blood by the addition of
FMLP, a bacterial peptide which is known to activate neutrophils (224, 270). However,
it is clinically important to determine whether tumor-derived products can activate
neutrophils or neutrophils precursors to become N-MDSC, and also to identify which
tumor-derived products are responsible for this phenomenon. We have found that the
addition of 10%v/v cell-free supernatants from the RCC cell line SK-RC-26B, caused
CD15+ N-MDSC to copurify with mononuclear cells after Ficoll density centrifugation.
After 1 hour of incubation, the percentage of N-MDSC in PBMC increased to over 15% on average, while one hour of incubation in complete media alone resulted in virtually no
N-MDSC induction (average<1% of PBMC) (Figure 2A). The appearance of the induced
N-MDSC phenotypically duplicated the appearance of those seen in the PBMCs and tumors of mRCC patients - CD33+loHLADR-CD15+CD14- (Figure 2B). In addition, cell-free supernatants generated from multiple RCC cell cultures were able to activate N-
MDSC at levels which were several fold above background (1.5-30 fold, Figure 2C), suggesting that N-MDSC induction by TCM is a common feature variably shared by most RCCs. Supernatants from well-established tumor cell lines derived from RCC tumors induced N-MDSC formation; but, most importantly, soluble products present in
114 supernantants from freshly cultured RCC cells, derived from freshly explanted tumors,
could replicate what was observed with established lines. Supernatants from 7 different
freshly derived RCC cells induced N-MDSC several fold above background, and to
varying degrees (Figure 2C). Interestingly, supernatants from two RCC cell lines were
incapable of inducing N-MDSC (Figure 2C).
In order to determine whether the induced N-MDSC could functionally act like
MDSC which accumulate in tumor-bearing hosts, we tested their ability to suppress T
cell function. N-MDSC induced from the whole blood of healthy donors were isolated
by FACS sorting of CD15+ cells and added to T cells in vitro at a 1:1 or 1:10 N-MDSC
to T cell ratio. T cells were activated with irradiated allogeneic stimulator (adherent) cells at a ratio of 1 T cell to 5 stimulator cells, and T cell proliferation was quantified using tritiated-thymidine incorporation for the last 18 hours of a 7 day stimulation period.
At a ratio of 1:1 and 1:10 (N-MDSC:T cell) T cell proliferation was suppressed in the presence of induced N-MDSC (Figure 2D). In addition, T cells in PBMC derived from whole blood cultured for 1 hour with media alone, which contained <2% N-MDSC, formed healthy blasts and produced normal amounts of interferon-gamma, as determined by intracellular cytokine staining with FACS, at the end of a 72 hour stimulation with anti-CD3/28 antibody-bound beads. However, T cells in PBMC derived from whole blood which had been cultured in 10% TCM and contained approximately 15-20%
MDSC, did not form into healthy blasts and failed to produce interferon-gamma upon stimulation with anti-CD3/28 antibodies (Figure 2E).
115
116 mRCC patient derived N-MDSC and in vitro activated N-MDSC resemble “primed” neutrophils
Peripheral blood neutrophils are known to express CD11b, CD66b and CD62L.
CD11b, the alphaM-integrin and Mac-1 component, is mobilized to the cell surface of neutrophils in larger amounts upon activation, and serves as a surrogate marker of neutrophil “priming.” Primed neutrophils with high expression of CD11b were shown to produce a larger respiratory burst and therefore generate larger amounts of reactive oxygen species upon further encounters (270, 271). CD66b upregulation is also an indicator of neutrophil priming and is thought to signify neutrophil degranulation (272).
CD62L, the L-selectin adhesion molecule, is proteolytically shed from the surface of activated neutrophils. We therefore asked whether any or all of these three markers could be used, in addition to density, to discriminate N-MDSC, from normal neutrophils which are dense and purify with red blood cells over a Ficoll density gradient.
We compared the expression levels of these three markers in CD15+ gated cells purifying with PBMCs following a 1 hour incubation with TCM. We used the bacteria- derived peptide fmlp as a positive control in these experiments, because it is known to induce neutrophils activation (224, 270). Figure 3A shows that both fmlp and TCM upregulated the expression of CD11b and CD66b, and downregulated the expression of
CD62L in hypodense N-MDSC, as compared to normal neutrophils isolated from RBC
pellets using dextran. The phenotypic analysis of CD15+ cells purifying with PBMC
from mRCC patient peripheral blood, or from patient RCC tumor specimens, show that,
compared to normal neutrophils, patient N-MDSC are phenotypically “activated”
similarly to those neutrophils which become activated with either fmlp or TCM. (Figure
117 3B). Representative data looking at the overlaid histograms from CD15-gated normal
neutrophils (purple), and hypodense, fmlp-activated neutrophils (green), and hypodense
CD15+ cells from patients (pink) is shown (Figure 3C). Importantly, the very small
amount of CD15+ cells found in the PBMC layer isolated from healthy donor blood, may
be incidentally present due to imperfections in Ficoll isolation, and display the same
levels of surface markers as normal, resting neutrophils (Figure 3B). In contrast, CD15+
cells purifying with PBMCs from patients have marked upregulation of CD11b and
CD66b, even greater than that induced with fmlp, indicating that they are in an activated
state. Additionally, the dramatic increase in the expression of CD66b in patient MDSC
may indicate that these N-MDSC had likely undergone extensive degranulation (Figure
3B). Importantly, the N-MDSC which were induced from healthy donor whole blood,
using tumor conditioned media (TCM) over one hour, phenotypically resembled the N-
MDSC identified in mRCC patients, with upregulated levels of CD11b and CD66b, and
downregulated CD62L, indicating that the model of N-MDSC induction using tumor-
derived products is likely to reflect what may be contributing to N-MDSC accumulation in mRCC cancer patients.
118
119 Tumor-derived products sustain N-MDSC viability in vitro
Neutrophils are known to rapidly turnover in vivo, with a lifespan of less than 24 hours, and neutrophil activation is typically associated with subsequent cell death (224,
271). We therefore tested whether the induced N-MDSC would quickly die in vitro. As expected, CD15+ N-MDSC purified with PBMCs over Ficoll following activation in whole blood, were largely undergoing spontaneous apoptosis after 48hours, when cultured in media alone. However, when 20% RCC tumor cell-conditioned media was added to these cultures, N-MDSC reduced apoptosis after 48 hours culture in vitro. The analysis undertaken is shown in Figure 4A and graphically summarized in Figure 4B.
120
121 GM-CSF reproduces the effect of tumor-conditioned media on N-MDSC activation and viability in vitro
GM-CSF has been implicated in MDSC formation in both mouse tumor models as
well as in human patients, and RCC tumors are known to express GM-CSF (92, 104, 116,
273). In addition, GM-CSF is a factor reported to not only to activate neutrophils, but
also to maintain neutrophil viability (274, 275). We therefore tested the effects of GM-
CSF in our model of N-MDSC activation. When added to whole blood for 1 hour, GM-
CSF induced the appearance of N-MDSC in the PBMC of healthy donors, similarly to
TCM (Figure 5A). GM-CSF additionally induced phenotypic changes consistent with
cellular activation in isolated neutrophils, as compared to media alone (Figure 5B).
Importantly, GM-CSF was also quite effective at maintaining the long-term viability of
N-MDSC in vitro (Figure 5C). We therefore tested supernatants from our SK-RC26b cell
line to confirm whether GM-CSF was a product of these tumor cells. The semi-
quantitative results obtained from proteome profile array show that tumor cells were
making abundant levels of GM-CSF (Figure 5D, squared). Importantly, other cytokines
which were found in tumor-cell supernatants and which are thought to contribute to N-
MDSC accumulation, were unable to mimic the effects of GM-CSF. As seen in Figure
5E, neither G-CSF, IL-8, or IL-6 could effectively induce N-MDSC from whole blood.
Fifty or 10ng/mL of GM-CSF induced significantly more N-MDSC than 100ng/mL of G-
CSF, IL-8 or IL-6, which all had modest to no activity. Finally, the pretreatment of TCM
with neutralizing antibody to GM-CSF, but not with isotype control antibody, could significantly reduce N-MDSC induction by TCM (Figure 5F, right).
122
123 Sunitinib indirectly reduces MDSC activation by tumor-derived products
We recently showed that sunitinib reduced MDSC accumulation in mRCC
patients following one or two cycles of treatment (51). To explore the mechanism of
sunitinib’s impact on N-MDSC accumulation in our system, we first pretreated whole
blood with sunitinib at 0.1 or 1.0ug/mL for 30 minutes prior to adding tumor conditioned
media. This pretreatment was not able to inhibit the activation of N-MDSC and equal
numbers of CD15+ cells were induced in the presence and absence of drug (Figure 6A).
We next asked whether sunitinib treatment of tumor cells, would modify the resultant
TCM such that less N-MDSC would be induced. Equal numbers of confluent SK-RC-
26B cells were cultured with or without 0.1ug/mL of sunitinib for three days, and then
harvested cell-free supernatants were used for N-MDSC induction. Sunitinib had no
significant effect on SK-RC-26B tumor cell viability or growth rate at up to 5.0ug/mL
(cell counts with trypan blue, and AnnexinV/7AAD staining, Figure 6B). However, supernatants from cells cultured in 0.1ug/mL of sunitinib were less able to induce N-
MDSC formation from healthy donor whole blood (Figure 6C). In contrast, sunitinib treatment of tumor cell lines did not significantly impact the ability of the resultant supernatants to sustain N-MDSC viability in vitro (Figure 6D).
To determine whether sunitinib was reducing N-MDSC induction mechanistically by reducing the amount of activating cytokines produced by tumor cells; we tested culture supernatants from untreated and sunitinib-treated tumor cells for several cytokines and chemokines. Sunitinib treatment of tumor cells, at non-toxic doses, did not significantly alter the amount of any protein cytokine or factor produced by tumor cells
124 (Figure 6E). Importantly, sunitinib did not decrease the amount of GM-CSF produced by tumor cells, which can both induce and maintain the viability of N-MDSC.
125
Sunitinib modulates tumor cell production of N-MDSC activating gangliosides
Sunitinib did not alter tumor cell production of any of the protein cytokines tested,
but several bioactive molecules are known to be lipid-based, and thus cannot be detected
by conventional methodologies. One such product which our lab has found to be made
and shed in large amounts by tumor cells is siacylated glycosphingolipids, or
gangliosides. These products have previously been implicated in T cell and dendritic cell
suppression, and were shown to be present in tumor cell supernatants and to insert
themselves in T cell membranes5 (55). We therefore tested whether gangliosides might
be a product in TCM which is capable of inducing N-MDSC formation from healthy-
donor whole blood. Figure 7A shows that purified gangliosides from the SK-RC-26B cell line, when added to whole blood for 1 hour, led to the activation of N-MDSC which co-purified with PBMC and appeared identical to those induced by either TCM or GM-
CSF. As seen in Figure 7B, SK-RC-26B-derived ganglioside activation of N-MDSC was dose responsive, and was effective at a concentration as low as 1.0 ug/mL (6-fold induction). In contrast to TCM, and GM-CSF, however, gangliosides were unable to sustain the viability of N-MDSC in vitro (data not shown).
To determine whether sunitinib-mediated suppression of ganglioside production could be responsible for the reduced N-MDSC-activating capacity of supernatants from sunitinib-treated tumor cells, we compared ganglioside expression in untreated and sunitinib-treated cells. We have previously identified GM2 to be an RCC-specific
ganglioside which has deleterious effects on T cell function, and which can be used as a
5 Biswas S, Richmond A, Biswas K, Ko J, Ghosh S, Simmons M, Rayman P, Rini B, Gill I, Tannenbaum C et. al. Elevated levels of select gangliosides in T cells from renal cell carcinoma patients is associated with T cell dysfunction. Accepted by Journal of Immunology.
126 marker of tumor-derived ganglioside expression (55). We found that tumor-specific
GM2 expression was significantly reduced on the surface of sunitinib-treated tumor cells,
as assessed by FACS (Figure 7C). We further quantified total ganglioside levels, using
the sialic acid quantification technique previously described (55), which were shed from equal numbers of tumor cells that were either untreated, or treated with sunitinib at
0.1ug/mL. As seen in Figure 7D, equal amounts of cell-free supernatant taken from untreated tumor cells contained more total ganglioside than cell-free supernatant taken
from sunitinib-treated cells. Taken together, RCC-derived gangliosides induce N-MDSC, and tumor cell treatment with sunitinib leads to a reduction in ganglioside production, cellular expression, and therefore shedding into culture supernatants. This reduction may account for the partial reversal of N-MDSC induction by supernatants of tumor cells which are treated in vitro with sunitinib.
127
128 Discussion
The accumulation of myeloid cells which are capable of inhibiting an anti-tumor
T cell response has been recognized in tumor-bearing mice for many years. In the mouse tumor model, the large majority of these myeloid-derived suppressor cells (MDSC) have the same marker expression as murine neutrophils (CD11b+Gr1+), and yet are reduced in
density, and are found to home with lymphocytes in the spleen, unlike normal neutrophils
(112). More recently, MDSC have been detected in the peripheral blood of cancer patients; and while neutrophilic, as well as monocytic and less mature subsets seem to accumulate, neutrophilic MDSC appear to be the most prevalent MDSC subset in many cancer types, including RCC (51, 108). In parallel to mouse tumor models, these cells isolated from patients share marker expression with normal neutrophils (CD33+HLADR-
CD15+), and yet have been distinguished by their reduced density which allows them to copurify with, and inhibit lymphocytes in vitro. Amazingly, the mechanisms which differentiate normal neutrophils and N-MDSC have not been adequately examined. The data included in this report suggest that one of the mechanisms responsible for N-MDSC accumulation in tumor-bearing hosts is the dynamic activation of neutrophils, or a subset of neutrophils, which circulate during steady-state conditions in healthy hosts and which, once activated, engage to suppress T cells, and have an abnormally prolonged lifespan.
The dynamic nature of this process may explain why MDSC accumulation is quickly reversed following tumor removal, and why MDSC fail to survive when transferred into naïve hosts (97).
We have previously shown the relative prevalence of N-MDSC in the blood of
RCC patients. However, this is the first report, to our knowledge, which characterizes
129 MDSC subsets isolated from fresh RCC tumors. The majority of the cells falling within the myelogate were likely tumor cells and other stromal cells which were negative for CD markers. However, the presence of nearly 10% MDSC is likely to be physiologically
significant, considering this would allow for nearly a 1:1 ratio with tumor-infiltrating
CD3+ T cells, which represented 15% of cells in digested tumors on average (data not shown).
The in vitro activation of normal-donor neutrophils into phenotypic and functional
N-MDSC by soluble tumor cell-derived products suggests that the N-MDSC endpoint may be reached via more than one pathway. Indeed, within the population of so-called
“terminally differentiated” neutrophils, a subset may exist which can become either de- differentiated or activated, rapidly, via a mechanism not previously recognized.
Additionally, relatively mature cells may become activated to become cells which resemble immature MDSC because they become both immunosuppressive, as well as have a relatively prolonged lifespan (in the presence of tumor-derived products). This point was also recently demonstrated by Kusmarsev et. al. (95). In their report, PBMC isolated from the blood of healthy human donors were exposed to tumor-conditioned media for 24-48 hours to induce the formation of CD11b+VEGFR1+ suppressive myeloid cells. In these studies MDSC were presumably induced from so-called
“terminally differentiated” cells which circulate in healthy hosts under steady-state conditions as well. The MDSC induced in these experiments were likely to be of the monocytic subset, however, because no neutrophils would have been present in the ficolled PBMC which were subsequently exposed to TCM.
130 CD66+ neutrophilic MDSC were recently described by Rodriguez et al, and resembled fmlp-induced neutrophils (224). This is the first report, to our knowledge, which shows that products derived from tumor cells can induce the formation of N-
MDSC. In addition, this induction was marked by the upregulation of surface expression of CD66b, along with CD11b, and the downregulation of CD62L, a signature that we found strongly manifest in tumor-associated N-MDSC, which exist in close proximity to activating factors. While fmlp can induce the formation of N-MDSC-like cells, our model of tumor-derived products activating N-MDSC is more clinically relevant to what may occur in cancer patients. Importantly, N-MDSC, following activation, could be maintained in vitro for several days in the presence of tumor-conditioned media. The failure of activated neutrophils to undergo normal physiological apoptosis is likely to contribute to N-MDSC accumulation which occurs in cancer patients. Humans produce billions of neutrophils daily with the expectation that they will be short-lived; such that a failure of these cells to turnover, would quickly lead to systemic accumulations such as those seen in tumor-bearing hosts (271). The relatively high prevalence of N-MDSC, compared to other subsets of MDSC, is therefore not surprising.
Neutrophils are capable of responding to several cytokines, but GM-CSF is one particular cytokine which has been shown to cause MDSC accumulation when given in large amounts, as part of vaccines (104, 105, 116). It has furthermore been implicated in immune dysfunction in several chronic inflammatory states, likely due to its pleiotrophic abilities to stimulate myeloid cell production from stem cells, as well as to act to functionally enhance mature myeloid cells (276). GM-CSF is made by many renal tumors, and its presence in the TCM may have partially been responsible for the activities
131 of TCM to activate N-MDSC and to prolong N-MDSC viability, because it could reproduce each of these actions when added as a single agent, and neutralization of GM-
CSF with antibody partially inhibited the ability of TCM to induce N-MDSC. However, our data indicate that other factors, such as gangliosides were likely contributing to N-
MDSC activation as well. Ganglioside have been shown to inhibit dendritic cell generation and function although the responsible mechanism is not well defined (277).
Our studies suggest that ganglioside can also promote immune suppression by inducing
N-MDSC from whole blood. The mechanism by which sunitinib lowers tumor cell production of gangliosides is currently under investigation.
Previously we reported that treatment of metastatic RCC patients with the tyrosine kinase inhibitor sunitinib significantly reduced the number of circulating MDSC and resulted in an improved type-1 T cell response (51). In the current studies, the potency of RCC supernatants to induce N-MDSC was reduced by approximately 50% when tumor cell lines were treated with a non-toxic dose of sunitinib. This reduction could not be accounted for by reductions in any of the tumor-produced cytokines known to act on myeloid cells, including GM-CSF. Rather, sunitinib reduced tumor cell production of gangliosides, which could activate N-MDSC in vitro but not maintain their viability in vitro. This is consistent with the fact that TCM from untreated and sunitinib- treated tumor cells were equally able to maintain MDSC viability in vitro, and had equivalent amounts of GM-CSF. These data explain one mechanism by which tumor- derived products induce the appearance of N-MDSC in patient blood and tumor. They also characterize one of the mechanisms by which sunitinib may reduce circulating levels of N-MDSC in patients- which is via actions on tumor cells and the resultant modulation
132 of ganglioside production, and which is yet limited by the continued production of N-
MDSC activating cytokines such as GM-CSF.
133 CHAPTER 5
DISSERTATION DISCUSSION AND FUTURE DIRECTIONS
Rationale
The unique power of our immune system to identify and destroy specific targets
makes it a prime candidate for the treatment of cancer, especially when used adjunctively
with other modalities. Unfortunately, the clinical investigation of immune-based
therapeutic strategies, especially vaccine-based strategies, has yielded disappointing
results. Perhaps the most promising strategy is that which employs the adoptive transfer
of tumor-infiltrating T cells which are expanded ex vivo and then re-infused into patients
in very large quantities. This may be due in part to the fact that these T cells are grown in
the absence of immunosuppressive networks, including MDSC, that exist in the tumor-
bearing hosts, while vaccines require T cell expansion to occur in vivo, in the presence of
tumor-mediated immune suppression. Future clinical trials will require modalities that
combat such immune suppression throughout the delivery of cellular immune therapy.
Both lymphodepleting and myeloablative chemotherapies prolong the engraftment of
adoptively transferred T cells and increase the objective anti-tumor response to adoptive
T cell therapy (ATT) in the setting of metastatic melanoma, yet these regimens are not
well tolerated by all patients, and the chemotherapies used generally provide no additional anti-tumor effects. We have therefore explored the potential of the anti-
angiogenic drug, sunitinib, to be used as an adjunct agent for immunotherapy in cancer.
Based on our findings and those of others, MDSC represent a major source of T cell suppression in tumor bearing hosts; and sunitinib has the ability to restore T cell function, thus opening the way for combinational studies that provide relief from immune
134 suppression while allowing T cell activation. However, the development of effective
immunotherapy will benefit from further understanding of the processes regulating
MDSC accumulation and sunitinib’s mechanistic impacts on this process. Our studies
using human RCC cancer patient blood and tumor tissue, several mouse tumor models, as
well as a model of tumor-driven MDSC activation from whole blood; have shown very
complimentary results, which have allowed us to form a unifying model of MDSC
accumulation, comprised of many conclusions drawn from parallel observations taken
from all three manuscripts.
Summary of Critical Findings
1. MDSC accumulate in tumor-bearing hosts. This finding has been reported by
several others in mouse tumor models, as well as in patients bearing certain tumors. Our
finding of ~5% MDSC in total PBMC in RCC patients is slightly lower that what was
reported by Zea et al. in his RCC cohort, yet above what was reported by Almand et. al.
in the blood of patients bearing in several types of squamous and adenocarcinomas (108,
114, 234). Importantly, the level was very significantly elevated when compared to that
detected in the blood of numerous age-matched normal blood donors (51). In the mouse
tumor models used, Renca, CT26, and 4T1, splenic MDSC were several fold above the
amount detected in the spleens of naïve mice and this was significant in every model.
Our detected levels were consistent with those previously published (113, 220, 221).
Finally the in vitro addition of 10%v/v cell-free tumor conditioned media to healthy human donor whole blood was able to activate a subset of neutrophils to phenotypically and functionally resemble MDSC derived from RCC patients, further implicating tumor-
135 derived products in the systemic accumulation of MDSC. At least half of the MDSC-
inducing capacity of TCM could be attributed to tumor-produced GM-CSF in this study.
2. Sunitinib reduces peripheral MDSC accumulation (blood/spleen). We were the
first to report on such a phenomenon, which has also now been reported by several other
groups (94, 220, 221) (and Cohen P, Talmadge, JE and Bronte, V by personal
communication). This finding has been consistent in both tumor-bearing patients and
mice, and even in naïve, non-tumor-bearing mice, suggesting that at least a large portion
sunitinib’s anti-MDSC effects are direct, via interactions with normal host cells, rather
than indirect, via tumor cytoreduction. However, in vitro experiments using TCM to
activate MDSC from whole blood suggest that a portion of sunitinib’s anti-MDSC impact
may result from a drug-induced reduction in MDSC-activating gangliosides, and
potentially other biologically active lipids.
3. The anti-MDSC effects of sunitinib occur independently of its anti-tumor effects.
This was first suggested by the fact that nearly every RCC patient receiving sunitinib saw
declines in MDSC, regardless of whether they had stable disease, tumor shrinkage, or
even disease progression while on drug (51). We did not assay levels of circulating
cytokines in this study, but many of these patient plasmas are currently frozen, and we
are beginning to test for correlations in various cytokines and MDSC, as well as cytokine
changes in response to drug. In the mouse tumor model, we found that declines in
splenic MDSC occurred regardless of whether the tumor itself was sensitive (Renca) or
resistant to drug (4T1/CT26). Additionally, sunitinib treatment did not lead to declines in
circulating levels of any of the cytokines thought to support MDSC. Rather, and
consistent with what was published to occur even in naïve, non-tumor-bearing mice,
136 several implicated cytokines went up in response to drug, including VEGF, G-CSF, IFN-
gamma, and SDF-1-alpha (258). Overall, this suggests that sunitinib does not remove the
drive for MDSC formation, in fact it may even increase this drive. Instead, sunitinib
must block one or several of the molecular signals that are required for MDSC to
accumulate.
Consistent with this, in vitro data examining the impact of sunitinib treatment to tumor
cell cultures show that the drug (when used at physiologically relevant doses) does not
have a direct anti-proliferative or toxic effect on tumor cells. Likewise, it did not
significantly reduce the levels of any MDSC-supportive cytokines or growth factors
produced by tumor cells in vitro. Sunitinib did, however, induce a much more subtle
change in tumor cells, which could only be appreciated thus far by a reduction in
ganglioside production, which may have been responsible for the decreased ability of
resultant TCM to induce MDSC from whole blood.
4. Sunitinib-mediated declines in MDSC lead to a restoration in bystander T cell function. In addition to experiencing declines in MDSC, RCC patients on sunitinib saw
improvements in Type-1 T cell function (IFNg production). This restoration in T cell
function was directly correlated with declines in MDSC, and depletion of MDSC from untreated PBMC using magnetic beads could reproduce this effect (51). The preserved function of patient T cells when activated in the absence of MDSC is consistent with previous reports (108, 241). T cell recovery upon MDSC removal was mirrored exactly
in tumor-bearing mice, where splenic T cell function significantly improved following
sunitinib treatment or MDSC removal with magnetic beads. In vitro, PBMC rich in
TCM-induced MDSC could not properly proliferate or produce IFNg in response to
137 stimulation, when compared to PBMC stimulated in the absence of TCM-induced-
MDSC.
5. Sunitinib induces n-MDSC apoptosis in relatively unprotected environments.
Sunitinib’s mechanism of action on MDSC is difficult to completely elucidate using strictly human samples; however, our studies with RCC patient-derived MDSC indicated that sunitinib treatment in vitro had a direct, pro-apoptotic effect on MDSC. This effect was myelospecific, and appeared to be MDSC-specific at 1.0ug/mL (51). In these experiments, however, TCM and GM-CSF were added to cultures, to maintain MDSC viability, and this likely dampened the appreciable effect of sunitinib in vitro, because
GM-CSF was later shown to be protective for MDSC in the presence of drug. These in vitro studies done with patient-derived MDSC strongly paralleled findings in the 4T1+ tumor model. Here, splenic n-MDSC were shown to have a lower rate of steady-state apoptosis in tumor-bearing mice, compared to naïve mice. Additionally, this survival advantage was corrected following sunitinib treatment, and sunitinib significantly increased the rate of n-MDSC apoptosis in 4T1+ mice. Similarly, MDSC derived from tumor-bearing mice underwent apoptosis upon exposure to 1.0ug/mL sunitinib in vitro.
MDSC could be protected from the apoptotic effects of drug, however, if cultured with
GM-CSF (and less so with TCM). Parallel findings seem to occur with in vitro-induced n-MDSC in preliminary studies done thus far, because, in addition to being a superior n-
MDSC activator, GM-CSF was found to be superior among all myelo-supportive cytokines at maintaining MDSC viability in vitro.
6. Sunitinib inhibits m-MDSC expansion. Examination of MDSC proliferation in vivo, possible only in the mouse model, showed monocytic, m-MDSC, to be the most
138 proliferative, yet least abundant, MDSC subset. Additionally, a pathologic increase in the rate of m-MDSC proliferation was appreciable in tumor bearing mice, which was reversible upon sunitinib treatment to tumor-bearing mice. Furthermore, sunitinib significantly inhibited MDSC proliferation in vitro over 72 hours, at 1.0ug/mL (lower doses not yet attempted). The MDSC thus far induced in vitro using TCM have been of the neutrophilic subset, and hence relatively non-proliferative. Future studies will sort m-
MDSC from patients and confirm sunitinib’s anti-proliferative impact on these cells.
Additionally, a method of in vitro m-MDSC induction has previously been reported (95), and is currently being tested in our lab, and sunitinib’s anti-proliferative effects on such cells will be examined.
7. Sunitinib does not induce MDSC differentiation into dendritic cells. It was previously published that the drug all-trans-retinoic acid (ATRA), can induce MDSC differentiation into DC (241). This drug, however, is yet to show therapeutic impact in patients. When tested for its ability to do the same, sunitinib did not mature RCC patient derived-MDSC into DC in vitro (51). This is consistent with our in vivo finding in RCC patients, and in naïve and tumor-bearing mice, that sunitinib treatment results in reduced dendritic cell levels (data not shown). This effect is to be somewhat expected, considering the importance of flt3-flt3L interactions to steady-state DC production and sunitinib’s ability to potently inhibit signaling through the flt3 receptor. Indeed, sunitinib’s reliance on flt3 blockade for its anti-MDSC effects will also be thoroughly tested in future experiments as well.
8. Sunitinib’s anti-MDSC impact is limited in tumor bed and bone marrow, possibly due to local presence of GM-CSF. Despite dramatic reductions in peripheral
139 MDSC in patients in tumor-bearing mice, MDSC were unexpectantly found to persist in
the tumor bed. In the mouse tumor models tested, Renca tumors were extremely
sensitive to drug, in a way that is not likely replicated in human tumors. However, 4T1
tumors had moderate sensitivity and resumed normal growth kinetics promptly upon
stopping drug. This is likely to reflect the in vivo situation in human patients, and in this model, there was a disparate effect of sunitinib on peripheral versus tumor-associated
MDSC. Also associated with a persistence of MDSC in human RCC and mouse 4T1 tumors was a continued depression in tumor-infiltrating T cell function. We additionally could not detect a reduction in BM-associated MDSC in naïve mice or Renca, CT26, or
4T1 tumor-bearing mice in response to drug. The local persistence of MDSC in the tumor bed was tentatively attributed (at least in part) to local availability of GM-CSF which was present in the tumor bed, but not detectible systemically, and which could protect MDSC from the effects of sunitinib in vitro. We hypothesize that bone marrow naturally contains abundant GM-CSF as well, although technical limitations prevented its detection in this compartment. This conclusion will be more rigorously tested as outlined below.
9. GM-CSF-mediated MDSC protection may result from its reprogramming of
MDSC from a STAT3-dominated to a STAT5-dominated pathway. Consistent with previously reported results, GM-CSF exposure to bone marrow cells preempted STAT3 signaling in favor of STAT5 signaling (193). These cultures were insensitive to the effects of sunitinib in vitro. In contrast, all other tested cytokines, drove STAT3- dominated cultures, which were uniformly sensitive to sunitinib-mediated inhibition. In vivo staining for pSTAT3 and pSTAT5 in MDSC derived from spleen/blood versus
140 tumor (or bone marrow) of tumor-bearing mice and patients, is currently underway, but
will likely be technically complicated. Furthermore, in vitro experiments with STAT3
and STAT5 knockout bone marrow, or using pharmacologic inhibitors, will also be attempted in future experiments.
Despite the persistence of intratumoral MDSC, the overall reduction in the systemic MDSC burden is likely to account for the synergy between sunitinib and other immunotherapies such as vaccine, which has recently been published by other groups, and which our collaborators have observed as well (220) (and Cohen, P, and Storkus, W.
manuscripts in preparation). Unfortunately, as detailed in chapter 3, and discussed further below, residual MDSC escape sunitinib-mediated toxicity, likely as a result of growth factor redundancy which exists in bone marrow and tumor in resistant tumor models. As such, future therapeutic efforts will likely benefit from the investigation of ancillary strategies to abrogate such regional resistance, in order to enhance sunitinib’s potency both as an immunomodulator and as a cancer therapy. In addition, the replacement of sunitinib-depleted dendritic cells, may also add to sunitinib’s therapeutic impact.
Regarding sunitinib’s mechanism of action
Our investigation into the role of sunitinib to reverse immune suppression in tumor- bearing hosts has revealed several factors which mechanistically contribute to MDSC accumulation in cancer, most of which are subject to modulation by this drug. A model outlining such processes is depicted in Figure 5-1. Our findings, as described above, suggest that the hyperaccumulation of MDSC occurs by at least three mechanisms, each of which is subject to inhibition by sunitinib. 1. Accumulation of (primarily) n-MDSC in
141 the tumor-bearing state occurs synchronously with a marked increased proliferation of
(primarily) Gr1lo, mononuclear, m-MDSC in the tumor-bearing state. This suggests that
an increased rate of expansion in precursor cells contributes to m-MDSC accumulation and may also contribute to n-MDSC accumulation, and that sunitinib reduces MDSC partially via its anti-proliferative effect on these cells. This observation has only been possible in the mouse tumor model thus far; however, future experiments with human
CD34+ cord-blood derived cells would likely show an increase in myeloid cell expansion
in response to TCM. In addition, sunitinib would likely have a strong anti-proliferative effect in these experiments, depending on the factors (i.e. +/- GM-CSF) added to cultures.
2. Tumor products were also shown to induce phenotypic and functional n-MDSC activation from a subset of neutrophilic constituents which circulate even under steady- state conditions in non-tumor bearing hosts. This activation likely contributes to the overall MDSC burden in tumor-bearing hosts, and sunitinib partially reversed this activation via its effects on tumor cell production of non-protein MDSC activators, such as gangliosides. This study relied on human donors, yet could likely be repeated in the mouse model. 3. Once expanded and activated, n-MDSC are bestowed with prolonged systemic survival, beyond that which is normal for neutrophils, especially activated neutrophils. Sunitinib acted to reduce this survival in vivo, and the toxic effect of sunitinib on MDSC was also shown in vitro in patient and tumor-bearing mice-derived
MDSC. A series of several experiments, outlined below, would solidify this model, however, and glean further insight into the exact molecular target(s) of sunitinib.
142 Periphery •G-CSF and STAT3 driven Bone Marrow •Susceptible to sunitinib- •G- and GM-CSF driven? mediated inhibition •STAT3 and STAT5 driven?
1. Proliferation of m-MDSC 2. n-MDSC activation m-MDSC n-MDSC
macrophage neutrophil 3. n-MDSC apoptosis
Dying dendritic cell neutrophil
=pathway increased in tumor-bearing state
Insensitive Tumors =pathway inhibited •G- and GM-CSF driven by sunitinib •STAT3 and STAT5 driven?
Figure 5-1 Model for Sunitinib-mediated Reversal of MDSC Accumulation in Tumor-bearing Hosts: 1) MDSC supply is increased by an increased rate of proliferation in m- MDSC. Sunitinib limits this via an anti-proliferative effect. 2) N-MDSC likely consist of both immature neutrophils which accumulate due to the increased rate of myeloid cell production, as well as activated neutrophils which become activated by tumor-produced products. Sunitinib limits this activation indirectly by dampening tumor cell production of gangliosides. 3) Cell death and clearance of n-MDSC is reduced, either by virtue of their cellular immaturity, or by virtue of TCM-mediated sustenance of activated n- MDSC. Sunitinib has a direct toxic effect on these cells. In the periphery, 1-3 are likely mediated by G- CSF, which exists at relatively high levels in the plasma of TB-hosts and is capable of driving expansion, activation, and viability of MDSC. G-CSF mediated effects on MDSC are STAT3-driven and reversible via sunitinib treatment. In contrast, other factors such as GM-CSF may control MDSC function in BM and tumor and render these cells less susceptible to sunitinib.
143 Future Directions
Sunitinib’s impact on MDSC expansion
Our results characterized in Chapter 3 show that m-MDSC proliferate at a much greater rate than n-MDSC, despite the fact that m-MDSC are outnumbered by n-MDSC
1:4. In addition, while around half the m-MDSC were F4/80+ and thus likely of monocyte/macrophage lineage, the other half were negative for any marker aside from
CD11b+ and dim expression of Gr1. This suggests that at least half the Gr1dim m-
MDSC may indeed be precursors to the more prevalent, Gr1hi n-MDSC (166). If proven correct, than sunitinib could be acting to limit the accumulation of BOTH MDSC subsets via its anti-proliferative effects alone. Testing whether m-MDSC are precursors to n-
MDSC is possible via FACS sorting of CD11b+Gr1loF4/80- cells from the spleens of tumor-bearing mice, followed by in vitro culture or in vivo adoptive transfer. Purified m-MDSC could be pulsed with CFSE and cultured for 3 or 6 days in the presence of SCF and Flt3L to mimic host stromal support, and variably supplemented with G-CSF, GM-
CSF, or TCM, which should enhance Gr1hi n-MDSC differentiation. Sunitinib (0.01-10 ug/ml) could be tested to confirm its ability to suppress m-MDSC proliferation as well as n-MDSC accumulation from the sorted m-MDSC. At culture’s end, cells stained with
CD11b, Gr1, Ly6G, and F4/80 would be run for FACS analyses. If m-MDSC differentiate into n-MDSC with progressive proliferative cycles, an increase in Gr1 staining fluorescence intensity and the acquisition of Ly6G staining upon successive cell divisions (apparent by successive CFSE dilutions) would be observed. In addition, if a portion of m-MDSC represents precursors to n-MDSC, then sunitinib will reduce the accumulation of both MDSC subsets by inhibiting Gr1lo m-MDSC proliferation, even
144 potentially at lower doses of sunitinib which are not directly proapoptotic to n-MDSC. In parallel in vivo experiments, sorted m-MDSC (CD11b+Gr1loF4/80-) from the spleens of
CD45.1 congenic BALB/cByJ tumor-bearing mice would be labeled with CFSE, then transferred i.v. into CD45.2 BALB/c bearing synchronous challenges of the same tumors.
3 or 6 days later FACS analysis of spleens and tumors from CD45.2 recipient mice for the presence of CD45.1+ MDSC, now with the phenotype of n-MDSC (Gr1hi, Ly6G+), would be completed. Detected CD45.1+ n-MDSC could also be isolated by FACS sort to confirm their T cell suppressive abilities.
It is additionally important to draw parallels in MDSC biology between tumor- bearing mice, and human cancer patients. Therefore CD33+CD15-CD14-, lineage- negative-MDSC in patients (the likely counterparts to Gr1loF4/80- MDSC in mice) would be tested for their ability to differentiate into n-MDSC upon consecutive proliferative cycles in vitro. In addition CD15+ n-MDSC from patients could be compared to Gr1hi n-
MDSC in mice. n-MDSC (CD15+) and lin-m-MDSC (CD33+CD15-CD14-) subsets from RCC patient blood would be sorted and labeled with CFSE then cultured in
SCF+flt3L with either G-CSF, GM-CSF, or RCC TCM and +/- sunitinib for either 3 or 6 days. Cells stained at the end of culture for CD14 and CD15 would be analyzed by
FACS, along with CFSE dilutions. If parallelism to mouse studies exists, pre-committed
CD15+ MDSC would not display CFSE dilutions and may not live out to 6 days in vitro.
Furthermore, CD15-CD14- MDSC would be the predominant proliferating subpopulation in culture, and would largely acquire CD15+ in the presence of G-CSF, TCM or GM-
CSF, as the precursors of n-MDSC.
145 Defining the role of GM-CSF and gangliosides in MDSC activation and the impact sunitinib has on MDSC activation
Results obtained in Chapter 4 suggest that GM-CSF and/or gangliosides commonly produced by RCC tumors are fundamental to RCC-mediated immunosuppression via the enhancement of n-MDSC activation. If so, GM-CSF and/or ganglioside levels would be significantly higher for RCC sources that induce n-MDSC compared to non-inducing sources. Furthermore, the selective susceptibility of gangliosides (and not GM-CSF) to sunitinib-mediated modulation, implies that ganglioside-mediated induction of n-MDSC would be significantly inhibited by sunitinib treatment (in vivo or in vitro), whereas n-
MDSC induction by GM-CSF would be unaffected. To further confirm such a model, freshly isolated RCC from two nephrectomy sources: untreated patients and patients who received neoadjuvant sunitinib pre-surgery could be compared for their n-MDSC inducing qualities. Fresh tumors could be digested enzymatically and allowed to adhere overnight prior to removing non-adherent cells. After an initial 7 day passage, adherent and replated pure tumor-cell culture supernatants could be collected at interval time periods with levels of GM-CSF determined by ELISA, and levels of shed ganglioside assessed by ganglioside isolation, followed by HPLC with on-line mass spectrometer
(LC-ESI-MS-MS). The same supernatants would also be added to whole blood (at 1:5 or
1:20) for 1 hour prior to ficoll centrifugation and subsequent staining for detection of de novo n-MDSC activation. Increased levels of GM-CSF and/or gangliosides would be expected to correlate with n-MDSC induction for each freshly explanted RCC supernatant. The activation of n-MDSC from whole blood should also be significantly
146 less, and ganglioside but not GM-CSF production significantly reduced, when the tumor
supernatants are prepared from the sunitinib-exposed nephrectomy tumor materials.
Additionally, freshly cultured tumor supernatants, both previously sunitinib-exposed
and unexposed nephrectomy explants, could also be cultured with graded doses of
sunitinib in vitro (0-10 ug/ml), and then supernatants exposed or not, to neutralizing anti-
GM-CSF or isotype ctrl Ab, and then tested for their capacity to induce n-MDSC from whole blood. Because sunitinib does not significantly reduce tumor production of GM-
CSF, n-MDSC induction should be partially reduced following either in vivo and/or in vitro exposure to sunitinib, or neutralizing anti-GM-CSF antibody; yet, only dual exposure to sunitinib and anti-GM-CSF antibody should largely prevent n-MDSC induction.
Does NAC reduce MDSC activation by inhibiting tumor expression of gangliosides?
The interactions between tumor cells and sunitinib which result in reduced ganglioside production and shedding by tumor cells are not known. The curious ability of N-acetyl cysteine (NAC), when added to tumor cells to also reduce the ability of TCM to induce n-
MDSC, suggests that NAC may also reduce tumor cell shedding of gangliosides, or tumor cell production of GM-CSF. It should therefore be determined whether NAC can, like sunitinib, reduce ganglioside production; or whether it reduces GM-CSF production by tumor cells, using the methods outlined above. If NAC, like sunitinib, also reduces ganglioside production, than redox status might be an important determinant in the production of these siacylated glycosphingolipids, and the effect of both NAC and sunitinib on the levels of ROS in tumor cells, using the ROS-sensing dye CM-H2DCFDA and FACS should be tested. The dissection of the pathway by which sunitinib alters
147 cellular redox status would then ensue.
Does n-MDSC induction by TCM/GM-CSF or Gangliosides within one hour relate to neutrophil priming with resultant degranulation and subsequent density alterations?
If GM-CSF and gangliosides are confirmed to be the major inducers of n-MDSC, mechanistic studies on the cellular changes resulting from GM-CSF and ganglioside exposure to myeloid cells would be appropriate. Neutrophils are known to exist in at least three phenotypes, the first of which is classified as “resting” or quiescent. In this state, the cells are circulating with a rounded morphology and little membrane ruffling.
In the inflammed vascular endothelium, contact with various pro-inflammatory stimuli, including TNF-alpha, IL-8 and GM-CSF, or bacterial peptides such as fmlp, or LPS, etc. induces neutrophil “priming.” Primed neutrophils acquire an adherent phenotype, partially due to upregulated surface levels of CD11b/CD18, or Mac-1, and are hyper- reactive to activation by stimuli which would not typically cause full activation. Finally,
“activated” neutrophils are fully engaged in the processes related to bacterial killing, including release of oxidative granules and phagocytosis (278, 279). Neutrophil priming is known to occur in response to GM-CSF within 15-60 minutes (280) and could potentially induce the release of secretory, tertiary, and specific granules (281-287). This is supported by the fact that both TCM and GM-CSF upregulated surface expression of
CD66b, present in specific granules, as well as CD11b, present in specific granules and secretory vesicles (288).
Future experiments would therefore solidify the working hypothesis behind the mechanism of n-MDSC induction, which occurs within only one hour’s time, to be related
to neutrophil priming with resultant degranulation and subsequent density alterations.
148 First, to confirm the origin of the induced n-MDSC as neutrophils, healthy donor blood
would be layered over ficoll and centrifuged and the RBC pellet collected, which would
include only RBC and granulocytes. The cells could then be treated with plain media,
TCM, GM-CSF, or gangliosides for one hour, and then again centrifuged over a Ficoll density gradient. FACS staining and cell counting would quantify the amount of n-
MDSC which then precipitated at the Ficoll interface where n-MDSC are found. If cells which were formerly dense granulocytes, falling with RBC over a density gradient, are induced with TCM, GM-CSF, or gangliosides to become less dense, and phenotypically resemble n-MDSC, then their suppressive nature would additionally be confirmed.
Furthermore, if n-MDSC are induced from previously pelleted granulocytes, then the requirement for platelets in this activation would also be considered less likely. It would also rule out the possibility that n-MDSC activation is indirect via TMC (etc) acting on lymphocyes/monocytes present in whole blood. Wright-Giemsa staining of cytospins would also be done to confirm a similar morphology in neutrophils and n-MDSC.
To confirm that a shift in density of resting neutrophils results from degranulation, whole blood could be pre-treated, or not, with the chemical compound, pentoxifylline, which prevents granule release (289, 290), prior to the addition of media, TCM, GM-
CSF, or gangliosides. If pentoxifylline prevents the appearance of n-MDSC in the
PBMC fraction of cells following Ficoll centrifugation, then n-MDSC activation and co- precipitation with T cells in 1 hour’s time would be attributable to a degranulation and a consequent change in cell density. Subsequent experiments would expose isolated neutrophils to media, TCM, GM-CSF, or gangliosides for one hour with or without prior pretreatment with pentoxifylline, and then stain these cells for surface versus intracellular
149 markers of azurophilic – CD63, specific – CD66b VEGF and lactoferrin, tertiary –
MMP8/9 and arginase, and secretory vescicles – CD35 and alkaline phosphatase, as well as for reactive oxygen species content using DCFDA (289-294). In addition, ELISAs could quantify the release of granule contents in the presence and absence of the various n-MDSC activators.
NADPH oxidase activation is another well reported effect of neutrophil priming, and results in elevated ROS production by activated cells (278). This, combined with release of arginase-containing granules, increases the likelihood of activated n-MDSC being more T-cell suppressive than normal neutrophils, and preliminary data indicates as much.
Such experiments are complicated, however, by the fact that even normal neutrophils may become activated in cultures with CD3/28-stimulated T cells, and even dying neutrophils are reported to release arginase into their environment. However, significant differences should still be detectible, and if so, the mechanism of n-MDSC T cell suppression can be tested via the addition of excess L-arginine or catalase (the ROS scavenger) to cultures.
The above listed strategies would also bring to light differences in the mechanism behind GM-CSF and ganglioside induced n-MDSC activation, which are highly likely.
For instance, if GM-CSF could, but gangliosides could not, induce n-MDSC appearance from RBC/granulocyte containing pellets that are subsequently re-centrifuged over a ficoll density gradient, this would point to a requirement for platelets or other plasma constituents in ganglioside-mediated induction.
Sunitinib’s impact on MDSC viability
Our studies have shown that tumor-activated n-MDSC can maintain viability for
150 several days in vitro if kept in the presence of tumor-conditioned media. Sunitinib
seemingly corrects this imbalance in neutrophil longevity in vivo and in vitro by promoting n-MDSC apoptosis, both for mice and humans respectively. Because it is
difficult to study human n-MDSC in vitro in the absence of exogenous GM-CSF or G-
GSF support, in vivo mouse experiments would be best suited to further characterize
sunitinib’s impact on n-MDSC homeostasis. Future experiments should test whether the
increase in n-MDSC longevity in tumor-bearing mice is dependent upon continuous
exposure to tumor-derived products (extrinsic factor-dependent survival) or whether it is mediated by internal changes (intrinsic survival) inherent to the MDSC created during
the tumor-bearing process. Isolated BM derived n-MDSC from naïve and tumor-bearing
mice would be labeled with CFSE, then adoptively transferred iv into both naïve and TB
mice. Cells from recipient mice spleens, BM, and tumors would be analyzed by flow
cytometry to determine the percentage of CFSE+ n-MDSC recovered over an extended
time period (24 and 72 hours). If the tumor-conditioned environment is essential to n-
MDSC survival, then CFSE+ cells would be recovered at 72 hours post-transfer only
when n-MDSC are transferred into tumor-bearers, but not into naïve mice; and this would be independent of whether cells were derived from naïve or tumor-bearing BM (extrinsic survival). Precisely this result is anticipated. To confirm that sunitinib is acting directly on peripheral n-MDSC in vivo n-MDSC taken from the BM of either untreated or sunitinib treated (6 days) tumor-bearing mice, which start with seemingly equal viability, could be tested for their ability to survive in tumor-bearing mice untreated or sunitinib- treated (6 days). The recovery of CFSE+ n-MDSC at 72 hours, only when labeled cells are injected into 4T1 untreated but not sunitinib treated mice, would confirm that
151 sunitinib directly reduces n-MDSC lifespan in vivo. The recovery of CFSE+ n-MDSC at
72 hours only when transferred cells are obtained from untreated tumor-bearing mice, but not from the BM of sunitinib-treated mice, would suggest that sunitinib permanently alters the inherent functional qualities (non-phenotypic) of MDSC during the formation process.
Defining the mechanism of sunitinib-mediated MDSC apoptosis, and its reliance on
STAT3 versus STAT5 signaling.
STAT3
Previously published work showed that sunitinib inhibited pSTAT3 activation in renal tumor cell lines via an inhibition of src tyrosine kinase activity (221). This in vitro inhibition took place within 2 hours of sunitinib treatment. Declines in pSTAT3 and p-
Src were also seen in tumor tissues after 1 day of in vivo treatment with sunitinib, and in tumor-associated myeloid cells after in vivo treatment for an unclear duration. Taken together, this indicates that sunitinib may have direct inhibitory effects on STAT3.
STAT3 activates many anti-apoptotic genes, including survivin, cyclin E, Bcl-xL and Bcl2
(221). Hence the sudden inhibition of STAT3 in STAT3-driven cultures would likely set cells up for death. IL-6 and G-CSF are primarily known to drive STAT3 activation, and this appeared to be the case in our culture system. Three questions remain - 1) Can sunitinib directly inhibit STAT3 in MDSC, 2) does STAT3 inhibition result in MDSC apoptosis in cultures reliant on STAT3 via a reduction in anti-apoptotic genes, and 3) is
STAT3 the operative survival pathway of MDSC in the periphery, and does reversal of this activation in sunitinib-treated mice predict MDSC apoptosis in vivo.
The ability of sunitinib to directly inhibit STAT3 activation in MDSC in vitro
152 should first be tested. If sunitinib has a direct effect on STAT3, this should be apparent in several early time points following stimulation with STAT3 activators. MDSC from spleen and bone marrow should be isolated and cultured in the presence and absence of the following: SCF+flt3L, SCF+flt3L+IL-6, SCF+flt3L+G-CSF, IL-6, G-CSF, with or without sunitinib pretreatment (30 minutes) for either 5min, 15min, 30min, 2 hours, or 8 hours. Cell lysates should be tested using western blots for the amount of total and phosphorylated-STAT3, and possible src family kinases, src, lyn and hck. The inhibition of STAT3 activation at early time points would point to a direct targeting of STAT3 by sunitinib. Fluorescence staining of cells treated the same could also be done to examine cells by confocal microscopy in order to determine whether STAT3 mobilization to the
G-CSFR and IL-6 receptor occur and whether there is subsequent nuclear localization in the presence of sunitinib. The identification of direct STAT3 inhibition in MDSC would be an unlikely, yet novel finding. Furthermore, STAT3 is unlikely to be activated in cultures treated with SCF+Flt3L alone, conditions where sunitinib’s pro-apoptotic effect nevertheless prevail. This would suggest that STAT3 activity does not make cells susceptible to drug in and of itself.
The levels of anti-apoptotic (bcl-xl, bcl-2, survivin, XIAP) and pro-apoptotic
(bad, bax, bak, caspase 3,8,9) genes and proteins liberated from MDSC cell lysates should be tested in each of the following conditions: 1) SCF+flt3L, 2) SCF+flt3L+IL-6,
3) SCF+flt3L+G-CSF, 4) IL-6, 5) G-CSF, 6) GM-CSF, 7) SCF+flt3L+GM-CSF, all with or without sunitinib treatment over 12, 24, and 48 hours. Apoptosis in sunitinib-sensitive cultures should be preceded by an increase in the ratio of pro- to anti-apoptotic genes.
Furthermore, if sunitinib mediated inhibition of receptor targets – ckit and flt3 are
153 important in its killing of IL-6 and G-CSF cultures, indicating a specific requirement for
parallel RTK signaling in the case of IL-6 and G-CSF, then this should be apparent by the
progressively increased ratio of proapoptotic genes as such:
6 +/- Sut = 7 +/- Sut <
2 – Sut = 3 – Sut <
2+ Sut = 3 + Sut = 4 +/- Sut = 5 +/- Sut <
1 + Sut
Finally, splenic, BM, and tumor – associated MDSC should be compared for there
STAT3 activation status in untreated and sunitinib-treated mice to test whether STAT3
downregulation precedes the initiation of apoptosis. This could be done using Gr1 and
pSTAT3 co-staining on tissue sections from each of the 3 compartments. Selective
expression of STAT3 in peripheral MDSC, with suppression following sunitinib
treatment, but before the onset of apoptosis in MDSC (3-6 days), would compartmentally
and chronologically implicate STAT3 suppression as a mechanism in sunitinib-mediated
N-MDSC apoptosis.
STAT5
In vitro experiments done in Chapter 3 with GM-CSF suggest that GM-CSF is
protective for MDSC, and this coincided with STAT5 activation and STAT3 repression.
Indeed, GM-CSF can induce the upregulation of several anti-apoptotic proteins, via
STAT5-dependent and independent pathways (257, 295). Yet three questions remain –
1) Does GM-CSF protect MDSC by virtue of STAT5 activation and the consequent
upregulation of anti-apoptotic genes, 2) does GM-CSF protect MDSC by virtue of
STAT3 downregulation, suggesting that STAT3 activation may sensitize cells to
154 sunitinib-mediated killing, 3) Is STAT5 – mediated protection of MDSC apparent in vivo.
The requirement for STAT5 in GM-CSF mediated protection from sunitinib can be addressed using bone marrow from STAT5 hypomorphic mice or STAT5 KO mice compared to that from littermate controls. If the STAT5-deficient bone marrow MDSCs remain protected from the effects of sunitinib in the presence of GM-CSF, this would argue against a protective role for STAT5. The levels of pro-versus anti-apoptotic genes present in each culture condition could be compared.
The possibility exists that STAT3 somehow sensitizes cells to sunitinib-mediated killing. If this is true, then increasing amounts of IL-6 treatment in vitro, in the presence of low, but sufficient levels of GM-CSF to mediate protection, should induce increasing amounts of MDSC apoptosis in response to sunitinib. This increasing apoptosis should coincide with increasing pSTAT3 activation.
Finally, splenic, BM, and tumor – associated MDSC should be compared for there
STAT5 activation status in untreated and sunitinib-treated mice. This could be done using Gr1 and pSTAT5 co-staining on tissue sections from each of the 3 compartments.
Selective expression of STAT5 in BM and tumor MDSC, with no suppression following sunitinib treatment, would compartmentally implicate STAT5 as a mechanism in sunitinib-mediated MDSC protection. Furthermore, GM-CSF administration to sunitinib- treated mice should render even peripheral MDSC protected from the effects of sunitinib in vivo, and this protection should coincide with the upregulation of pSTAT5.
STAT3 and STAT5 Independent Mechanism of Sunitinib-mediated Apoptosis
Another possibility is that sunitinib inhibits the expansion/proliferation of GM-
155 colony forming progenitor, ckit and or flt3+ cells at relatively low concentrations of drug.
This inhibition would prevent the replacement of dying MDSC. In the case of N-MDSC, cells most likely to have G-CSF receptors and to be sustained by G-CSF in vitro and in vivo, cells are more terminal and have little replicative potential. These cells would thus
be more dependent on replacement by newly forming cells – a process potently inhibited
by drug. Therefore, sunitinib would appear to have a larger toxic effect here, merely by
maintaining an anti-proliferative effect on receptor+ cells. In the case of M-MDSC, cells most likely to have GM-CSF receptors and to have replicative potential, as well as the potential to live several days in vitro, there is therefore likely less reliance on replacement
by newly forming cells. Thus an anti-proliferative effect of drug would not appear to resemble an apoptotic effect in vitro. The possibility that sunitinib’s main effect is anti-
proliferative and not pro-apoptotic can be tested as follows. N-MDSC can be pulsed and
fluorescently labeled with a red or far red dye prior to culture. M-MDSC and other BM
cells can be pulsed with green CFSE prior to culture. Combined mixed MDSC cultures
can then be treated with the above conditions – A-G with and without sunitinib. At the
end of 3 or 6 days, green M-MDSC cells can be examined for viability with annexinV
and for proliferation with CFSE divisions. Red N-MDSC can be examined for viability.
If sunitinib’s main action is anti-proliferative, then red cells will have similar annexinV
positivity in all cultures, and yet the annexinV positivity of red and green cells combined
will be higher in sunitinib-treated cultures {with the exception of GM-CSF treated ones)
due to a failure of green cell proliferation.
Further defining the RTKs that sunitinib targets to block MDSC accumulation
While the VEGF and c-kit receptors have each been implicated in MDSC generation,
156 our data suggest that additional tumor-host interactions are at least as important. Our preliminary experiments done thus far have found that in both the 4T1 and CT26 tumor models, maximally tolerated doses of either imatinib (targets ckit, PDGFR and BCR/abl), vatalanib (selectively targets VEGFRs), or both drugs together minimally inhibited
MDSC generation (data not shown). Unlike sunitinib, these RTKIs do not block Flt3.
Therefore, sunitinib’s unique capacity to block all three steps of MDSC development (m-
MDSC proliferation, ganglioside-mediated n-MDSC differentiation, and promotion of n-
MDSC viability) may depend upon a different set of RTK blockades at each step, thereby benefitting from sunitinib’s targeting promiscuity. The capacities of narrower spectrum
RTKIs as well as RTK-blocking mAbs to inhibit the individual steps of n-MDSC development compared to sunitinib, which inhibits all three steps, should be rigorously tested. Based on preliminary data, blocking flt3 alone (lestaurtinib or anti-CD135 mAb) or blocking flt3 along with c-kit (imatinib or anti-CD117 mAb) and/or VEGFR (vatalanib or anti-CD309 Ab) for MDSC inhibition should first be tested and compared to that with sunitinib. These experiments may implicate a particular RTK target relevant to sunitinib’s efficacy at each stage of n-MDSC development. As correlative studies, patient and healthy donor-induced MDSC, as well as spleen, bone marrow, and tumor- associated MDSC from tumor-bearing mice, should be examined for their expression of select RTKs at individual stages of differentiation. Our preliminary assays have found that only 35% of human MDSC stain positively for surface VEGFR1 and only 25% for surface flt3 (not shown), suggesting either that expression is highly heterogeneous or that staining conditions are not yet sufficiently sensitive. RTKs are known to become internalized and ubiquitinated upon interaction with ligand, and some RTKs exist at very
157 low levels on the cell surface, such that internalization could render them no longer
detectible by conventional FACS surface staining. It would thus be appropriate to
examine both surface and internalized RTK’s, and to attempt to enhance assay sensitivity
with signal amplification on MDSC either directly ex vivo or following a 4 hour exposure
to either: plain RPMI +/-serum starvation +/- PMA/I (phorbol myristate
acetate/Ionomycin) (or TCM) activation. The expression of these receptors will
furthermore be compared to that on normal neutrophils and monocytes isolated by our
standard methods, and correlated to the observed effects of inhibitors under the
experiment above.
Our findings showed that Gr1hi n-MDSC were relatively non-proliferative, and
suggested that the accumulation of this MDSC subset in tumor-bearing mice is more related to an abnormally prolonged lifespan. Sunitinib seemed to normalize this in vivo by increasing the frequency of n-MDSC apoptosis, as measured by AnnexinV staining,
an effect which was reproducible in vitro. Sunitinib also acted to inhibit the abnormal
expansion of m-MDSC, a subset of which may represent n-MDSC precursors. The
possibility exists that such anti-proliferative activity alone could lead to the perceived
increase in n-MDSC apoptosis, because terminal n-MDSC would eventually undergo
spontaneous apoptosis and fail to be replaced. This possibility could be tested in vitro by
FACS sorting bone marrow into Gr1hi n-MDSC and Gr1lo m-MDSC, then labeling the
respective groups either green or red with fluorescent, cell-permeant dye. Cells would
then be cultured together, and stimulated with G- or GM-CSF +SCF/flt3L in the presence
or absence of sunitinib. If sunitinib’s only mechanism of action is anti-proliferative, then the main difference in treated cultures would be the lack of red cell divisions. However,
158 if sunitinib acts directly on n-MDSC to induce apoptosis, then green cell death would be higher in the sunitinib-treated group, than in the untreated group.
The further possibility exists that sunitinib is able to inhibit the tyrosine phosphorylation of previously unrecognized targets. Downstream signaling from several cytokine receptors, including IL-6, G-CSF, and GM-CSF, are known to activate tyrosine kinases, and thus the tyrosine phosphorylation pattern produced in response to each of these cytokines, with and without SCF and flt3L, in the presence and absence of sunitinib should be examined. This could be done using cell lysates from the above conditions, and either a phopho-tyrosine-protein array, a luminex-based array, or phospho-tyrosine western blots. If sunitinib decreased tyrosine phosphorylation on protein targets other than the known RTKs, than further studies would describe the mechanism for this inhibition, as well as the consequences to MDSC cellular function.
Regarding intratumoral resistance to sunitinib
Treatment with sunitinib causes at least a log reduction of MDSC in the spleens of tumor bearing mice and in the peripheral blood of RCC patients. However, the absolute reduction of MDSC within the tumors and BM of 4T1+ tumor-bearing mice following sunitinib treatment was far less pronounced. Furthermore, persistent intratumoral MDSC remained T cell suppressive during sunitinib treatment in both 4T1-bearing mice and
RCC patients. This is in contrast to the RENCA tumor model wherein no viable cells
(tumor, stromal, or immune) are detectable within tumor tissue isolated during treatment.
Yet, even in the RENCA tumor model, mice cannot be considered cured of disease, because drug cessation prompts resumption in tumor growth. GM-CSF, which was selectively present in the tumor bed (and presumably bone marrow) enhanced MDSC
159 survival and expansion, and rendered MDSC resistant to the in vitro effects of sunitinib.
High concentrations of GM-CSF in vivo have putatively been associated with MDSC induction, and GM-CSF’s unique capacity to shift BM cells from STAT3- to STAT5- driven dominant programming associated with T cell suppression has been demonstrated. In addition, hypoxia-induced survival signals which revolve around HIF-
1alpha activation may also contribute to MDSC resistance to sunitinib in BM and tumor
compartments, allowing tumor-associated MDSC to function in a STAT3-independent,
HIF-1alpha-dependent manner. Such was recently suggested by Cesar and Gabrilovich in
an oral presentation at the 2009 AACR meeting. HIF-1alpha is additionally thought to be
important in myeloid cell inflammatory activation. Furthermore, BM is known to be a
relatively hypoxic environment, and sunitinib and other anti-angiogenic drugs are known
to induce further hypoxia in the tumor microenvironment. Future experiments should
delineate the roles of GM-CSF and HIF-1alpha in promoting MDSC resistance to
sunitinib within the BM and tumor environments, which currently limits the anti-cancer
potential of sunitinib as a single agent.
Our data, as well as that published by others, suggests that peripheral MDSC have an
RTK-dependent, STAT3-dominated activation signature which is highly susceptible to
sunitinib (221). MDSC in tumor and BM compartments, however, are likely to receive
additional STAT5- and/or HIF-1alpha dependent growth support, driven by GM-CSF
and/or hypoxia respectively, which renders them sunitinib resistant. In vivo studies
should examine the levels of pY705STAT3 and pY694STAT5 and HIF-1a in MDSC
from the spleens, BM, and tumors of naïve, treated and untreated tumor-bearing mice via
the rapid processing of tissues in sterile, cold, 1x PBS with immediate fixation of single
160 cell suspensions followed by staining for lineage markers and intracellular phospho-
STAT3, STAT5, and HIF-1alpha proteins, followed by FACS analysis. Splenic MDSC
are anticipated to have high levels of pSTAT3 and BM and tumor MDSC would likely
have high levels of pSTAT5. In addition, tumor-associated MDSC (pre-sunitinib
treatment) may have elevated HIF-1alpha protein intracellularly. pSTAT5 signaling
would be expected to persist and HIF-1alpha signaling elevate, in sunitinib-treated mice;
while pSTAT3 signaling would decline in response to sunitinib. In addition to flow-
based quantification of pSTAT3/5 and HIF-1a signaling, remaining fresh fixed/then
frozen spleens and tumors, and paraffin-embedded BM (from decalcified bone), could be
used to confirm the persistence of pSTAT5 and HIF-1a in the marrow and tumor of
treated mice.
Parallel in vitro experiments should determine whether GM-CSF and hypoxia can
protect BM cells via STAT5/HIF-1a activation pathway from sunitinib-mediated
apoptosis. MDSC viability could be tested in culture conditions containing either GM-
CSF or hypoxia, +/- sunitinib in the presence or absence of selective inhibitors of
pSTAT5 and HIF-1a respectively, with parallel measurements by FACS of the relative
amounts of pSTAT3, pSTAT5, and HIF-1a taken. To further link STAT3 or STAT5 to
sunitinib susceptibility vs resistance, respectively, cultures could also be performed to
evaluate the in vitro n-MDSC survival in the presence or absence of sunitinib, using BM
from B6-background mice with hypomorphic STAT5 or conditionally knocked out for
STAT3, versus wildtype littermate controls. GM-CSF driven, sunitinib-resistant MDSC
should be easily derived from STAT3KO mice but potentially not STAT5KO mice, whereas cytokines other than GM-CSF (flt3L, SCF, IL-6, G-CSF, etc.) would be unable
161 to promote MDSC from STAT3KO mice, and would be unchanged from STAT5KO mice.
The in vivo potential of GM-CSF and HIF-1 alpha-inducing chronic hypoxia should be tested for their ability to reverse the peripheral effects of sunitinib on MDSC. Injection of purified GM-CSF, if given at high enough concentrations, or chronic hypoxia (metered delivery of nitrogen-diluted air) would enhance the survival of splenic/peripheral MDSC even in the presence of sunitinib if these are the main agents responsible for sunitinib- resistance. Preliminary experiments could determine the dosing of GM-CSF which results in increased pSTAT5 in splenic MDCSs, and confirm that 10% FiO2 similarly
induces HIF-1alpha. The percentage of viable (annexinV-) MDSC present in the spleens,
BM, and tumors of mice untreated or treated with sunitinib would be assessed in each condition. GM-CSF and/or controlled hypoxia may sufficiently recreate BM and tumor
microenvironment-like conditions in the periphery/spleens of tumor-bearing mice, and
hence render even splenic MDSC resistant to the effects of sunitinib. This protection
should be accompanied by a loss in the recovery of T cell effector function following
sunitinib treatment, and this could be confirmed experimentally as well.
The examination of immune cells from tumors of patients previously treated with
sunitinib is very preliminary, and yet suggests that the same regionalized effect of sunitinib which occurs in 4T1+ tumor-bearing mice, may also occur in patients.
Therefore, additional tumor and blood samples from Stage IV RCC patients having had either no treatment, or sunitinib treatment, should be obtained to contrast the impacts of sunitinib exposure upon MDSC and T cells in both of these compartments. Nephrectomy and blood samples obtained from RCC patients proximally exposed to neoadjuvant sunitinib should be contrasted to matched patients who underwent nephrectomy without
162 prior sunitinib exposure. Our preliminary results indicate that MDSC and Tregs persist in
the tumors of sunitinib-treated patients, and that effector T cells remain suppressed
intratumorally; this contrasts to sunitinib’s opposite impact on MDSC, Tregs and effector
T cells in the peripheral blood of these patients. In addition, MDSC from the blood and tumors of untreated and sunitinib-treated patients should be compared for their levels of pSTAT3, pSTAT5, and HIF-1a; and plasma and tumor lysates from the same cohorts analyzed for GM-CSF and other cytokine expression.
BIBLIOGRAPHY:
1. Cohen, H. T. and McGovern, F. J. Renal-cell carcinoma. N Engl J Med, 353: 2477-2490, 2005. 2. Gogas, H. J., Kirkwood, J. M., and Sondak, V. K. Chemotherapy for metastatic melanoma: time for a change? Cancer, 109: 455-464, 2007. 3. Motzer, R. J., Bander, N. H., and Nanus, D. M. Renal-cell carcinoma. N Engl J Med, 335: 865-875, 1996. 4. Escudier, B., Lassau, N., Angevin, E., et al. Phase I trial of sorafenib in combination with IFN alpha-2a in patients with unresectable and/or metastatic renal cell carcinoma or malignant melanoma. Clin Cancer Res, 13: 1801-1809, 2007. 5. Fecher, L. A., Amaravadi, R. K., and Flaherty, K. T. The MAPK pathway in melanoma. Curr Opin Oncol, 20: 183-189, 2008. 6. Samlowski, W. E., Wong, B., and Vogelzang, N. J. Management of renal cancer in the tyrosine kinase inhibitor era: a view from 3 years on. BJU Int, 102: 162- 165, 2008. 7. Acquavella, N., Kluger, H., Rhee, J., et al. Toxicity and activity of a twice daily high-dose bolus interleukin 2 regimen in patients with metastatic melanoma and metastatic renal cell cancer. J Immunother, 31: 569-576, 2008. 8. Begley, J. and Ribas, A. Targeted therapies to improve tumor immunotherapy. Clin Cancer Res, 14: 4385-4391, 2008. 9. Moschos, S. J., Mandic, M., Kirkwood, J. M., Storkus, W. J., and Lotze, M. T. Focus on FOCIS: interleukin 2 treatment associated autoimmunity. Clin Immunol, 127: 123-129, 2008. 10. Bui, J. D. and Schreiber, R. D. Cancer immunosurveillance, immunoediting and inflammation: independent or interdependent processes? Curr Opin Immunol, 19: 203-208, 2007. 11. Dunn, G. P., Old, L. J., and Schreiber, R. D. The three Es of cancer immunoediting. Annu Rev Immunol, 22: 329-360, 2004.
163 12. Teng, M. W., Swann, J. B., Koebel, C. M., Schreiber, R. D., and Smyth, M. J. Immune-mediated dormancy: an equilibrium with cancer. J Leukoc Biol, 84: 988- 993, 2008. 13. Bergers, G. and Hanahan, D. Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer, 8: 592-603, 2008. 14. Kaplan, R. N., Riba, R. D., Zacharoulis, S., et al. VEGFR1-positive haematopoietic bone marrow progenitors initiate the pre-metastatic niche. Nature, 438: 820-827, 2005. 15. Kujawski, M., Kortylewski, M., Lee, H., et al. Stat3 mediates myeloid cell- dependent tumor angiogenesis in mice. J Clin Invest, 118: 3367-3377, 2008. 16. Murdoch, C., Muthana, M., Coffelt, S. B., and Lewis, C. E. The role of myeloid cells in the promotion of tumour angiogenesis. Nat Rev Cancer, 8: 618-631, 2008. 17. Shojaei, F., Wu, X., Malik, A. K., et al. Tumor refractoriness to anti-VEGF treatment is mediated by CD11b+Gr1+ myeloid cells. Nat Biotechnol, 25: 911- 920, 2007. 18. Koebel, C. M., Vermi, W., Swann, J. B., et al. Adaptive immunity maintains occult cancer in an equilibrium state. Nature, 450: 903-907, 2007. 19. Rosenberg, S. A., Restifo, N. P., Yang, J. C., Morgan, R. A., and Dudley, M. E. Adoptive cell transfer: a clinical path to effective cancer immunotherapy. Nat Rev Cancer, 8: 299-308, 2008. 20. Klapper, J. A., Downey, S. G., Smith, F. O., et al. High-dose interleukin-2 for the treatment of metastatic renal cell carcinoma : a retrospective analysis of response and survival in patients treated in the surgery branch at the National Cancer Institute between 1986 and 2006. Cancer, 113: 293-301, 2008. 21. Smith, F. O., Downey, S. G., Klapper, J. A., et al. Treatment of metastatic melanoma using interleukin-2 alone or in conjunction with vaccines. Clin Cancer Res, 14: 5610-5618, 2008. 22. Rosenberg, S. A. and Dudley, M. E. Adoptive cell therapy for the treatment of patients with metastatic melanoma. Curr Opin Immunol, 21: 233-240, 2009. 23. Berger, C., Turtle, C. J., Jensen, M. C., and Riddell, S. R. Adoptive transfer of virus-specific and tumor-specific T cell immunity. Curr Opin Immunol, 21: 224- 232, 2009. 24. McDermott, D. F. Immunotherapy of metastatic renal cell carcinoma. Cancer, 115: 2298-2305, 2009. 25. Melief, C. J. Cancer immunotherapy by dendritic cells. Immunity, 29: 372-383, 2008. 26. Vulink, A., Radford, K. J., Melief, C., and Hart, D. N. Dendritic cells in cancer immunotherapy. Adv Cancer Res, 99: 363-407, 2008. 27. Bhardwaj, N. Harnessing the immune system to treat cancer. J Clin Invest, 117: 1130-1136, 2007. 28. Dunn, G. P., Koebel, C. M., and Schreiber, R. D. Interferons, immunity and cancer immunoediting. Nat Rev Immunol, 6: 836-848, 2006. 29. Swann, J. B., Vesely, M. D., Silva, A., et al. Demonstration of inflammation- induced cancer and cancer immunoediting during primary tumorigenesis. Proc Natl Acad Sci U S A, 105: 652-656, 2008.
164 30. Bharadwaj, U., Li, M., Zhang, R., Chen, C., and Yao, Q. Elevated interleukin-6 and G-CSF in human pancreatic cancer cell conditioned medium suppress dendritic cell differentiation and activation. Cancer Res, 67: 5479-5488, 2007. 31. Finn, O. J. Cancer immunology. N Engl J Med, 358: 2704-2715, 2008. 32. Rosenberg, S. A., Sherry, R. M., Morton, K. E., et al. Tumor progression can occur despite the induction of very high levels of self/tumor antigen-specific CD8+ T cells in patients with melanoma. J Immunol, 175: 6169-6176, 2005. 33. Reis e Sousa, C. Dendritic cells in a mature age. Nat Rev Immunol, 6: 476-483, 2006. 34. Sotomayor, E. M., Borrello, I., Rattis, F. M., et al. Cross-presentation of tumor antigens by bone marrow-derived antigen-presenting cells is the dominant mechanism in the induction of T-cell tolerance during B-cell lymphoma progression. Blood, 98: 1070-1077, 2001. 35. Filatenkov, A. A., Jacovetty, E. L., Fischer, U. B., et al. CD4 T cell-dependent conditioning of dendritic cells to produce IL-12 results in CD8-mediated graft rejection and avoidance of tolerance. J Immunol, 174: 6909-6917, 2005. 36. Janssen, E. M., Lemmens, E. E., Wolfe, T., et al. CD4+ T cells are required for secondary expansion and memory in CD8+ T lymphocytes. Nature, 421: 852-856, 2003. 37. Smith, C. M., Wilson, N. S., Waithman, J., et al. Cognate CD4(+) T cell licensing of dendritic cells in CD8(+) T cell immunity. Nat Immunol, 5: 1143-1148, 2004. 38. Bennaceur, K., Chapman, J. A., Touraine, J. L., and Portoukalian, J. Immunosuppressive networks in the tumour environment and their effect in dendritic cells. Biochim Biophys Acta, 1795: 16-24, 2009. 39. Fricke, I. and Gabrilovich, D. I. Dendritic cells and tumor microenvironment: a dangerous liaison. Immunol Invest, 35: 459-483, 2006. 40. Gerner, M. Y., Casey, K. A., and Mescher, M. F. Defective MHC class II presentation by dendritic cells limits CD4 T cell help for antitumor CD8 T cell responses. J Immunol, 181: 155-164, 2008. 41. Pinzon-Charry, A., Ho, C. S., Laherty, R., et al. A population of HLA-DR+ immature cells accumulates in the blood dendritic cell compartment of patients with different types of cancer. Neoplasia, 7: 1112-1122, 2005. 42. Pinzon-Charry, A., Maxwell, T., and Lopez, J. A. Dendritic cell dysfunction in cancer: a mechanism for immunosuppression. Immunol Cell Biol, 83: 451-461, 2005. 43. Vicari, A. P., Caux, C., and Trinchieri, G. Tumour escape from immune surveillance through dendritic cell inactivation. Semin Cancer Biol, 12: 33-42, 2002. 44. Yang, L. and Carbone, D. P. Tumor-host immune interactions and dendritic cell dysfunction. Adv Cancer Res, 92: 13-27, 2004. 45. Zitvogel, L., Tesniere, A., and Kroemer, G. Cancer despite immunosurveillance: immunoselection and immunosubversion. Nat Rev Immunol, 6: 715-727, 2006. 46. Zou, W. Immunosuppressive networks in the tumour environment and their therapeutic relevance. Nat Rev Cancer, 5: 263-274, 2005.
165 47. van der Bruggen, P., Traversari, C., Chomez, P., et al. A gene encoding an antigen recognized by cytolytic T lymphocytes on a human melanoma. Science, 254: 1643-1647, 1991. 48. Novellino, L., Castelli, C., and Parmiani, G. A listing of human tumor antigens recognized by T cells: March 2004 update. Cancer Immunol Immunother, 54: 187-207, 2005. 49. Borden, E. C., Sen, G. C., Uze, G., et al. Interferons at age 50: past, current and future impact on biomedicine. Nat Rev Drug Discov, 6: 975-990, 2007. 50. Tatsumi, T., Kierstead, L. S., Ranieri, E., et al. Disease-associated bias in T helper type 1 (Th1)/Th2 CD4(+) T cell responses against MAGE-6 in HLA- DRB10401(+) patients with renal cell carcinoma or melanoma. J Exp Med, 196: 619-628, 2002. 51. Ko, J. S., Zea, A. H., Rini, B. I., et al. Sunitinib mediates reversal of myeloid- derived suppressor cell accumulation in renal cell carcinoma patients. Clin Cancer Res, 15: 2148-2157, 2009. 52. Rabinovich, G. A., Gabrilovich, D., and Sotomayor, E. M. Immunosuppressive strategies that are mediated by tumor cells. Annu Rev Immunol, 25: 267-296, 2007. 53. Finke, J. H., Rini, B., Ireland, J., et al. Sunitinib reverses type-1 immune suppression and decreases T-regulatory cells in renal cell carcinoma patients. Clin Cancer Res, 14: 6674-6682, 2008. 54. Sa, G., Das, T., Moon, C., et al. GD3, an overexpressed tumor-derived ganglioside, mediates the apoptosis of activated but not resting T cells. Cancer Res, 69: 3095-3104, 2009. 55. Biswas, K., Richmond, A., Rayman, P., et al. GM2 expression in renal cell carcinoma: potential role in tumor-induced T-cell dysfunction. Cancer Res, 66: 6816-6825, 2006. 56. Ahmadzadeh, M. and Rosenberg, S. A. TGF-beta 1 attenuates the acquisition and expression of effector function by tumor antigen-specific human memory CD8 T cells. J Immunol, 174: 5215-5223, 2005. 57. Li, M. O., Wan, Y. Y., Sanjabi, S., Robertson, A. K., and Flavell, R. A. Transforming growth factor-beta regulation of immune responses. Annu Rev Immunol, 24: 99-146, 2006. 58. Petrausch, U., Jensen, S. M., Twitty, C., et al. Disruption of TGF-beta signaling prevents the generation of tumor-sensitized regulatory T cells and facilitates therapeutic antitumor immunity. J Immunol, 183: 3682-3689, 2009. 59. Thomas, D. A. and Massague, J. TGF-beta directly targets cytotoxic T cell functions during tumor evasion of immune surveillance. Cancer Cell, 8: 369-380, 2005. 60. Wallace, A., Kapoor, V., Sun, J., et al. Transforming growth factor-beta receptor blockade augments the effectiveness of adoptive T-cell therapy of established solid cancers. Clin Cancer Res, 14: 3966-3974, 2008. 61. Bergmann, C., Strauss, L., Zeidler, R., Lang, S., and Whiteside, T. L. Expansion of human T regulatory type 1 cells in the microenvironment of cyclooxygenase 2 overexpressing head and neck squamous cell carcinoma. Cancer Res, 67: 8865- 8873, 2007.
166 62. Akasaki, Y., Liu, G., Chung, N. H., et al. Induction of a CD4+ T regulatory type 1 response by cyclooxygenase-2-overexpressing glioma. J Immunol, 173: 4352- 4359, 2004. 63. Baratelli, F., Lin, Y., Zhu, L., et al. Prostaglandin E2 induces FOXP3 gene expression and T regulatory cell function in human CD4+ T cells. J Immunol, 175: 1483-1490, 2005. 64. Huang, M., Stolina, M., Sharma, S., et al. Non-small cell lung cancer cyclooxygenase-2-dependent regulation of cytokine balance in lymphocytes and macrophages: up-regulation of interleukin 10 and down-regulation of interleukin 12 production. Cancer Res, 58: 1208-1216, 1998. 65. Yang, L., Yamagata, N., Yadav, R., et al. Cancer-associated immunodeficiency and dendritic cell abnormalities mediated by the prostaglandin EP2 receptor. J Clin Invest, 111: 727-735, 2003. 66. Curiel, T. J., Wei, S., Dong, H., et al. Blockade of B7-H1 improves myeloid dendritic cell-mediated antitumor immunity. Nat Med, 9: 562-567, 2003. 67. Krambeck, A. E., Thompson, R. H., Dong, H., et al. B7-H4 expression in renal cell carcinoma and tumor vasculature: associations with cancer progression and survival. Proc Natl Acad Sci U S A, 103: 10391-10396, 2006. 68. Kryczek, I., Wei, S., Zhu, G., et al. Relationship between B7-H4, regulatory T cells, and patient outcome in human ovarian carcinoma. Cancer Res, 67: 8900- 8905, 2007. 69. Kryczek, I., Wei, S., Zou, L., et al. Cutting edge: induction of B7-H4 on APCs through IL-10: novel suppressive mode for regulatory T cells. J Immunol, 177: 40-44, 2006. 70. Kryczek, I., Zou, L., Rodriguez, P., et al. B7-H4 expression identifies a novel suppressive macrophage population in human ovarian carcinoma. J Exp Med, 203: 871-881, 2006. 71. Wei, S., Shreiner, A. B., Takeshita, N., et al. Tumor-induced immune suppression of in vivo effector T-cell priming is mediated by the B7-H1/PD-1 axis and transforming growth factor beta. Cancer Res, 68: 5432-5438, 2008. 72. Ritter, G. and Livingston, P. O. Ganglioside antigens expressed by human cancer cells. Semin Cancer Biol, 2: 401-409, 1991. 73. Sorice, M., Parolini, I., Sansolini, T., et al. Evidence for the existence of ganglioside-enriched plasma membrane domains in human peripheral lymphocytes. J Lipid Res, 38: 969-980, 1997. 74. Zhang, X. and Kiechle, F. L. Review: Glycosphingolipids in health and disease. Ann Clin Lab Sci, 34: 3-13, 2004. 75. Hakomori, S. and Handa, K. Glycosphingolipid-dependent cross-talk between glycosynapses interfacing tumor cells with their host cells: essential basis to define tumor malignancy. FEBS Lett, 531: 88-92, 2002. 76. Hoon, D. S., Okun, E., Neuwirth, H., Morton, D. L., and Irie, R. F. Aberrant expression of gangliosides in human renal cell carcinomas. J Urol, 150: 2013- 2018, 1993. 77. Black, P. H. Shedding from the cell surface of normal and cancer cells. Adv Cancer Res, 32: 75-199, 1980.
167 78. Valentino, L., Moss, T., Olson, E., et al. Shed tumor gangliosides and progression of human neuroblastoma. Blood, 75: 1564-1567, 1990. 79. Chahlavi, A., Rayman, P., Richmond, A. L., et al. Glioblastomas induce T- lymphocyte death by two distinct pathways involving gangliosides and CD70. Cancer Res, 65: 5428-5438, 2005. 80. Garcia-Ruiz, C., Colell, A., Paris, R., and Fernandez-Checa, J. C. Direct interaction of GD3 ganglioside with mitochondria generates reactive oxygen species followed by mitochondrial permeability transition, cytochrome c release, and caspase activation. Faseb J, 14: 847-858, 2000. 81. Kudo, D., Rayman, P., Horton, C., et al. Gangliosides expressed by the renal cell carcinoma cell line SK-RC-45 are involved in tumor-induced apoptosis of T cells. Cancer Res, 63: 1676-1683, 2003. 82. Ladisch, S., Li, R., and Olson, E. Ceramide structure predicts tumor ganglioside immunosuppressive activity. Proc Natl Acad Sci U S A, 91: 1974-1978, 1994. 83. Irani, D. N., Lin, K. I., and Griffin, D. E. Brain-derived gangliosides regulate the cytokine production and proliferation of activated T cells. J Immunol, 157: 4333- 4340, 1996. 84. Tourkova, I. L., Shurin, G. V., Chatta, G. S., et al. Restoration by IL-15 of MHC class I antigen-processing machinery in human dendritic cells inhibited by tumor- derived gangliosides. J Immunol, 175: 3045-3052, 2005. 85. Biswas, S., Biswas, K., Richmond, A., et al. Elevated levels of select gangliosides in T cells from renal cell carcinoma patients is associated with T cell dysfunction. J Immunol, 183: 5050-5058, 2009. 86. Barnett, B. G., Ruter, J., Kryczek, I., et al. Regulatory T cells: a new frontier in cancer immunotherapy. Adv Exp Med Biol, 622: 255-260, 2008. 87. Curiel, T. J. Regulatory T cells and treatment of cancer. Curr Opin Immunol, 20: 241-246, 2008. 88. Wang, H. Y. and Wang, R. F. Regulatory T cells and cancer. Curr Opin Immunol, 19: 217-223, 2007. 89. Cheng, P., Corzo, C. A., Luetteke, N., et al. 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, 2008. 90. Ko, J. S., Bukowski, R. M., and Fincke, J. H. Myeloid-derived suppressor cells: a novel therapeutic target. Curr Oncol Rep, 11: 87-93, 2009. 91. Gabrilovich, D. Mechanisms and functional significance of tumour-induced dendritic-cell defects. Nat Rev Immunol, 4: 941-952, 2004. 92. Gabrilovich, D. I. and Nagaraj, S. Myeloid-derived suppressor cells as regulators of the immune system. Nat Rev Immunol, 9: 162-174, 2009. 93. Huang, B., Pan, P. Y., Li, Q., et al. 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, 2006. 94. Kao, J., Packer, S., Vu, H. L., et al. Phase 1 study of concurrent sunitinib and image-guided radiotherapy followed by maintenance sunitinib for patients with oligometastases: acute toxicity and preliminary response. Cancer, 115: 3571- 3580, 2009.
168 95. Kusmartsev, S., Eruslanov, E., Kubler, H., et al. Oxidative stress regulates expression of VEGFR1 in myeloid cells: link to tumor-induced immune suppression in renal cell carcinoma. J Immunol, 181: 346-353, 2008. 96. Kusmartsev, S. and Gabrilovich, D. I. Immature myeloid cells and cancer- associated immune suppression. Cancer Immunol Immunother, 51: 293-298, 2002. 97. Kusmartsev, S. and Gabrilovich, D. I. Inhibition of myeloid cell differentiation in cancer: the role of reactive oxygen species. J Leukoc Biol, 74: 186-196, 2003. 98. Kusmartsev, S. and Gabrilovich, D. I. Role of immature myeloid cells in mechanisms of immune evasion in cancer. Cancer Immunol Immunother, 55: 237-245, 2006. 99. Marigo, I., Dolcetti, L., Serafini, P., Zanovello, P., and Bronte, V. Tumor-induced tolerance and immune suppression by myeloid derived suppressor cells. Immunol Rev, 222: 162-179, 2008. 100. Nagaraj, S. and Gabrilovich, D. I. Myeloid-derived suppressor cells. Adv Exp Med Biol, 601: 213-223, 2007. 101. Nagaraj, S. and Gabrilovich, D. I. Tumor escape mechanism governed by myeloid-derived suppressor cells. Cancer Res, 68: 2561-2563, 2008. 102. Pan, P. Y., Wang, G. X., Yin, B., et al. Reversion of immune tolerance in advanced malignancy: modulation of myeloid-derived suppressor cell development by blockade of stem-cell factor function. Blood, 111: 219-228, 2008. 103. Serafini, P., Borrello, I., and Bronte, V. Myeloid suppressor cells in cancer: recruitment, phenotype, properties, and mechanisms of immune suppression. Semin Cancer Biol, 16: 53-65, 2006. 104. Serafini, P., Carbley, R., Noonan, K. A., et al. 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, 2004. 105. Serafini, P., De Santo, C., Marigo, I., et al. Derangement of immune responses by myeloid suppressor cells. Cancer Immunol Immunother, 53: 64-72, 2004. 106. Sinha, P., Clements, V. K., Bunt, S. K., Albelda, S. M., and Ostrand-Rosenberg, S. Cross-talk between myeloid-derived suppressor cells and macrophages subverts tumor immunity toward a type 2 response. J Immunol, 179: 977-983, 2007. 107. Sinha, P., Clements, V. K., Fulton, A. M., and Ostrand-Rosenberg, S. Prostaglandin E2 promotes tumor progression by inducing myeloid-derived suppressor cells. Cancer Res, 67: 4507-4513, 2007. 108. Zea, A. H., Rodriguez, P. C., Atkins, M. B., et al. Arginase-producing myeloid suppressor cells in renal cell carcinoma patients: a mechanism of tumor evasion. Cancer Res, 65: 3044-3048, 2005. 109. Young, M. R., Aquino, S., and Young, M. E. Differential induction of hematopoiesis and immune suppressor cells in the bone marrow versus in the spleen by Lewis lung carcinoma variants. J Leukoc Biol, 45: 262-273, 1989.
169 110. Young, M. R., Newby, M., and Wepsic, H. T. Hematopoiesis and suppressor bone marrow cells in mice bearing large metastatic Lewis lung carcinoma tumors. Cancer Res, 47: 100-105, 1987. 111. Nikcevich, D. A., Duffie, G. P., Young, M. R., et al. Stimulation of suppressor cells in the bone marrow and spleens of high dose cyclophosphamide-treated C57Bl/6 mice. Cell Immunol, 109: 349-359, 1987. 112. Movahedi, K., Guilliams, M., Van den Bossche, J., et al. Identification of discrete tumor-induced myeloid-derived suppressor cell subpopulations with distinct T cell-suppressive activity. Blood, 111: 4233-4244, 2008. 113. Youn, J. I., Nagaraj, S., Collazo, M., and Gabrilovich, D. I. Subsets of myeloid- derived suppressor cells in tumor-bearing mice. J Immunol, 181: 5791-5802, 2008. 114. Almand, B., Clark, J. I., Nikitina, E., et al. Increased production of immature myeloid cells in cancer patients: a mechanism of immunosuppression in cancer. J Immunol, 166: 678-689, 2001. 115. Schmielau, J. and Finn, O. J. Activated granulocytes and granulocyte-derived hydrogen peroxide are the underlying mechanism of suppression of t-cell function in advanced cancer patients. Cancer Res, 61: 4756-4760, 2001. 116. Filipazzi, P., Valenti, R., Huber, V., et al. Identification of a new subset of myeloid suppressor cells in peripheral blood of melanoma patients with modulation by a granulocyte-macrophage colony-stimulation factor-based antitumor vaccine. J Clin Oncol, 25: 2546-2553, 2007. 117. Hoechst, B., Ormandy, L. A., Ballmaier, M., et al. A new population of myeloid- derived suppressor cells in hepatocellular carcinoma patients induces CD4(+)CD25(+)Foxp3(+) T cells. Gastroenterology, 135: 234-243, 2008. 118. Bronte, V., Serafini, P., De Santo, C., et al. IL-4-induced arginase 1 suppresses alloreactive T cells in tumor-bearing mice. J Immunol, 170: 270-278, 2003. 119. Hesse, M., Modolell, M., La Flamme, A. C., et al. Differential regulation of nitric oxide synthase-2 and arginase-1 by type 1/type 2 cytokines in vivo: granulomatous pathology is shaped by the pattern of L-arginine metabolism. J Immunol, 167: 6533-6544, 2001. 120. Ochoa, A. C., Zea, A. H., Hernandez, C., and Rodriguez, P. C. Arginase, prostaglandins, and myeloid-derived suppressor cells in renal cell carcinoma. Clin Cancer Res, 13: 721s-726s, 2007. 121. Rodriguez, P. C., Hernandez, C. P., Quiceno, D., et al. Arginase I in myeloid suppressor cells is induced by COX-2 in lung carcinoma. J Exp Med, 202: 931- 939, 2005. 122. Rodriguez, P. C., Quiceno, D. G., Zabaleta, J., et al. Arginase I production in the tumor microenvironment by mature myeloid cells inhibits T-cell receptor expression and antigen-specific T-cell responses. Cancer Res, 64: 5839-5849, 2004. 123. Rodriguez, P. C., Zea, A. H., DeSalvo, J., et al. L-arginine consumption by macrophages modulates the expression of CD3 zeta chain in T lymphocytes. J Immunol, 171: 1232-1239, 2003. 124. Rodriguez, P. C., Quiceno, D. G., and Ochoa, A. C. L-arginine availability regulates T-lymphocyte cell-cycle progression. Blood, 109: 1568-1573, 2007.
170 125. Bingisser, R. M., Tilbrook, P. A., Holt, P. G., and Kees, U. R. Macrophage- derived nitric oxide regulates T cell activation via reversible disruption of the Jak3/STAT5 signaling pathway. J Immunol, 160: 5729-5734, 1998. 126. Duhe, R. J., Evans, G. A., Erwin, R. A., et al. Nitric oxide and thiol redox regulation of Janus kinase activity. Proc Natl Acad Sci U S A, 95: 126-131, 1998. 127. Talmadge, J. E. Pathways mediating the expansion and immunosuppressive activity of myeloid-derived suppressor cells and their relevance to cancer therapy. Clin Cancer Res, 13: 5243-5248, 2007. 128. Nefedova, Y., Fishman, M., Sherman, S., et al. Mechanism of all-trans retinoic acid effect on tumor-associated myeloid-derived suppressor cells. Cancer Res, 67: 11021-11028, 2007. 129. Shojaei, F. and Ferrara, N. Antiangiogenic therapy for cancer: an update. Cancer J, 13: 345-348, 2007. 130. Miller, K. D., Sweeney, C. J., and Sledge, G. W., Jr. Can tumor angiogenesis be inhibited without resistance? Exs 95-112, 2005. 131. Kerbel, R. S. Therapeutic implications of intrinsic or induced angiogenic growth factor redundancy in tumors revealed. Cancer Cell, 8: 269-271, 2005. 132. Shojaei, F., Singh, M., Thompson, J. D., and Ferrara, N. Role of Bv8 in neutrophil-dependent angiogenesis in a transgenic model of cancer progression. Proc Natl Acad Sci U S A, 105: 2640-2645, 2008. 133. Yang, L., DeBusk, L. M., Fukuda, K., et al. Expansion of myeloid immune suppressor Gr+CD11b+ cells in tumor-bearing host directly promotes tumor angiogenesis. Cancer Cell, 6: 409-421, 2004. 134. Yang, L., Huang, J., Ren, X., et al. Abrogation of TGF beta signaling in mammary carcinomas recruits Gr-1+CD11b+ myeloid cells that promote metastasis. Cancer Cell, 13: 23-35, 2008. 135. Shojaei, F., Wu, X., Qu, X., et al. G-CSF-initiated myeloid cell mobilization and angiogenesis mediate tumor refractoriness to anti-VEGF therapy in mouse models. Proc Natl Acad Sci U S A, 106: 6742-6747, 2009. 136. Abe, F., Dafferner, A. J., Donkor, M., et al. Myeloid-derived suppressor cells in mammary tumor progression in FVB Neu transgenic mice. Cancer Immunol Immunother, 2009. 137. Donkor, M. K., Lahue, E., Hoke, T. A., et al. Mammary tumor heterogeneity in the expansion of myeloid-derived suppressor cells. Int Immunopharmacol, 9: 937- 948, 2009. 138. Salvadori, S., Martinelli, G., and Zier, K. Resection of solid tumors reverses T cell defects and restores protective immunity. J Immunol, 164: 2214-2220, 2000. 139. Gabrilovich, D., Ishida, T., Oyama, T., et al. Vascular endothelial growth factor inhibits the development of dendritic cells and dramatically affects the differentiation of multiple hematopoietic lineages in vivo. Blood, 92: 4150-4166, 1998. 140. Huang, Y., Chen, X., Dikov, M. M., et al. Distinct roles of VEGFR-1 and VEGFR-2 in the aberrant hematopoiesis associated with elevated levels of VEGF. Blood, 110: 624-631, 2007. 141. Larrivee, B., Pollet, I., and Karsan, A. Activation of vascular endothelial growth factor receptor-2 in bone marrow leads to accumulation of myeloid cells: role of
171 granulocyte-macrophage colony-stimulating factor. J Immunol, 175: 3015-3024, 2005. 142. Rutella, S., Pierelli, L., Bonanno, G., et al. Role for granulocyte colony- stimulating factor in the generation of human T regulatory type 1 cells. Blood, 100: 2562-2571, 2002. 143. Volpi, I., Perruccio, K., Tosti, A., et al. Postgrafting administration of granulocyte colony-stimulating factor impairs functional immune recovery in recipients of human leukocyte antigen haplotype-mismatched hematopoietic transplants. Blood, 97: 2514-2521, 2001. 144. Sawanobori, Y., Ueha, S., Kurachi, M., et al. Chemokine-mediated rapid turnover of myeloid-derived suppressor cells in tumor-bearing mice. Blood, 111: 5457- 5466, 2008. 145. Bronte, V., Chappell, D. B., Apolloni, E., et al. 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, 1999. 146. Takeda, K., Hatakeyama, K., Tsuchiya, Y., Rikiishi, H., and Kumagai, K. A correlation between GM-CSF gene expression and metastases in murine tumors. Int J Cancer, 47: 413-420, 1991. 147. Young, M. R., Wright, M. A., and Young, M. E. Antibodies to colony-stimulating factors block Lewis lung carcinoma cell stimulation of immune-suppressive bone marrow cells. Cancer Immunol Immunother, 33: 146-152, 1991. 148. Young, M. R., Young, M. E., and Wright, M. A. Stimulation of immune- suppressive bone marrow cells by colony-stimulating factors. Exp Hematol, 18: 806-811, 1990. 149. Ostrand-Rosenberg, S. and Sinha, P. Myeloid-derived suppressor cells: linking inflammation and cancer. J Immunol, 182: 4499-4506, 2009. 150. Brem-Exner, B. G., Sattler, C., Hutchinson, J. A., et al. Macrophages driven to a novel state of activation have anti-inflammatory properties in mice. J Immunol, 180: 335-349, 2008. 151. Dietlin, T. A., Hofman, F. M., Lund, B. T., et al. Mycobacteria-induced Gr-1+ subsets from distinct myeloid lineages have opposite effects on T cell expansion. J Leukoc Biol, 81: 1205-1212, 2007. 152. Gelderman, K. A., Hultqvist, M., Pizzolla, A., et al. Macrophages suppress T cell responses and arthritis development in mice by producing reactive oxygen species. J Clin Invest, 117: 3020-3028, 2007. 153. Gomez-Garcia, L., Lopez-Marin, L. M., Saavedra, R., et al. Intact glycans from cestode antigens are involved in innate activation of myeloid suppressor cells. Parasite Immunol, 27: 395-405, 2005. 154. Goni, O., Alcaide, P., and Fresno, M. Immunosuppression during acute Trypanosoma cruzi infection: involvement of Ly6G (Gr1(+))CD11b(+ )immature myeloid suppressor cells. Int Immunol, 14: 1125-1134, 2002. 155. Heithoff, D. M., Enioutina, E. Y., Bareyan, D., Daynes, R. A., and Mahan, M. J. Conditions that diminish myeloid-derived suppressor cell activities stimulate cross-protective immunity. Infect Immun, 76: 5191-5199, 2008.
172 156. Zhu, B., Bando, Y., Xiao, S., et al. CD11b+Ly-6C(hi) suppressive monocytes in experimental autoimmune encephalomyelitis. J Immunol, 179: 5228-5237, 2007. 157. Bunt, S. K., Sinha, P., Clements, V. K., Leips, J., and Ostrand-Rosenberg, S. Inflammation induces myeloid-derived suppressor cells that facilitate tumor progression. J Immunol, 176: 284-290, 2006. 158. Bunt, S. K., Yang, L., Sinha, P., et al. Reduced inflammation in the tumor microenvironment delays the accumulation of myeloid-derived suppressor cells and limits tumor progression. Cancer Res, 67: 10019-10026, 2007. 159. Sinha, P., Okoro, C., Foell, D., et al. Proinflammatory S100 proteins regulate the accumulation of myeloid-derived suppressor cells. J Immunol, 181: 4666-4675, 2008. 160. Nefedova, Y., Huang, M., Kusmartsev, S., et al. Hyperactivation of STAT3 is involved in abnormal differentiation of dendritic cells in cancer. J Immunol, 172: 464-474, 2004. 161. Kusmartsev, S., Nefedova, Y., Yoder, D., and Gabrilovich, D. I. Antigen-specific inhibition of CD8+ T cell response by immature myeloid cells in cancer is mediated by reactive oxygen species. J Immunol, 172: 989-999, 2004. 162. Nefedova, Y., Nagaraj, S., Rosenbauer, A., et al. 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, 2005. 163. Kortylewski, M., Kujawski, M., Wang, T., et al. Inhibiting Stat3 signaling in the hematopoietic system elicits multicomponent antitumor immunity. Nat Med, 11: 1314-1321, 2005. 164. Croker, B. A., Metcalf, D., Robb, L., et al. SOCS3 is a critical physiological negative regulator of G-CSF signaling and emergency granulopoiesis. Immunity, 20: 153-165, 2004. 165. Lee, C. K., Raz, R., Gimeno, R., et al. STAT3 is a negative regulator of granulopoiesis but is not required for G-CSF-dependent differentiation. Immunity, 17: 63-72, 2002. 166. Panopoulos, A. D., Zhang, L., Snow, J. W., et al. STAT3 governs distinct pathways in emergency granulopoiesis and mature neutrophils. Blood, 108: 3682- 3690, 2006. 167. Yu, H. and Jove, R. The STATs of cancer--new molecular targets come of age. Nat Rev Cancer, 4: 97-105, 2004. 168. Nabarro, S., Himoudi, N., Papanastasiou, A., et al. Coordinated oncogenic transformation and inhibition of host immune responses by the PAX3-FKHR fusion oncoprotein. J Exp Med, 202: 1399-1410, 2005. 169. Wang, T., Niu, G., Kortylewski, M., et al. Regulation of the innate and adaptive immune responses by Stat-3 signaling in tumor cells. Nat Med, 10: 48-54, 2004. 170. Chen, S. H., Murphy, D. A., Lassoued, W., et al. Activated STAT3 is a mediator and biomarker of VEGF endothelial activation. Cancer Biol Ther, 7: 1994-2003, 2008. 171. Zachary, I. and Gliki, G. Signaling transduction mechanisms mediating biological actions of the vascular endothelial growth factor family. Cardiovasc Res, 49: 568- 581, 2001.
173 172. Ye, F., Chen, H. Z., Xie, X., et al. Vascular endothelial growth factor (VEGF) and ovarian carcinoma cell supernatant activate signal transducers and activators of transcription (STATs) via VEGF receptor-2 (KDR) in human hemopoietic progenitor cells. Gynecol Oncol, 94: 125-133, 2004. 173. Tchaikovski, V., Fellbrich, G., and Waltenberger, J. The molecular basis of VEGFR-1 signal transduction pathways in primary human monocytes. Arterioscler Thromb Vasc Biol, 28: 322-328, 2008. 174. Wu, Y. and Zhu, Z. Vascular endothelial growth factor receptor 1, a therapeutic target in cancer, inflammation and other disorders. Curr Med Chem, 16: 2890- 2898, 2009. 175. Kawakami, Y., Inagaki, N., Salek-Ardakani, S., et al. Regulation of dendritic cell maturation and function by Bruton's tyrosine kinase via IL-10 and Stat3. Proc Natl Acad Sci U S A, 103: 153-158, 2006. 176. Staples, K. J., Smallie, T., Williams, L. M., et al. IL-10 induces IL-10 in primary human monocyte-derived macrophages via the transcription factor Stat3. J Immunol, 178: 4779-4785, 2007. 177. Tamassia, N., Calzetti, F., Menestrina, N., et al. Circulating neutrophils of septic patients constitutively express IL-10R1 and are promptly responsive to IL-10. Int Immunol, 20: 535-541, 2008. 178. Campbell, J. D., Cook, G., Robertson, S. E., et al. Suppression of IL-2-induced T cell proliferation and phosphorylation of STAT3 and STAT5 by tumor-derived TGF beta is reversed by IL-15. J Immunol, 167: 553-561, 2001. 179. Wierenga, A. T., Schuringa, J. J., Eggen, B. J., Kruijer, W., and Vellenga, E. Downregulation of IL-6-induced STAT3 tyrosine phosphorylation by TGF-beta1 is mediated by caspase-dependent and -independent processes. Leukemia, 16: 675-682, 2002. 180. Duarte, R. F. and Frank, D. A. SCF and G-CSF lead to the synergistic induction of proliferation and gene expression through complementary signaling pathways. Blood, 96: 3422-3430, 2000. 181. Chakraborty, A., Dyer, K. F., Cascio, M., Mietzner, T. A., and Tweardy, D. J. Identification of a novel Stat3 recruitment and activation motif within the granulocyte colony-stimulating factor receptor. Blood, 93: 15-24, 1999. 182. Chakraborty, A. and Tweardy, D. J. Stat3 and G-CSF-induced myeloid differentiation. Leuk Lymphoma, 30: 433-442, 1998. 183. Chakraborty, A. and Tweardy, D. J. Granulocyte colony-stimulating factor activates a 72-kDa isoform of STAT3 in human neutrophils. J Leukoc Biol, 64: 675-680, 1998. 184. Redell, M. S., Tsimelzon, A., Hilsenbeck, S. G., and Tweardy, D. J. Conditional overexpression of Stat3alpha in differentiating myeloid cells results in neutrophil expansion and induces a distinct, antiapoptotic and pro-oncogenic gene expression pattern. J Leukoc Biol, 82: 975-985, 2007. 185. Shao, H., Xu, X., Jing, N., and Tweardy, D. J. Unique structural determinants for Stat3 recruitment and activation by the granulocyte colony-stimulating factor receptor at phosphotyrosine ligands 704 and 744. J Immunol, 176: 2933-2941, 2006.
174 186. Tian, S. S., Lamb, P., Seidel, H. M., Stein, R. B., and Rosen, J. Rapid activation of the STAT3 transcription factor by granulocyte colony-stimulating factor. Blood, 84: 1760-1764, 1994. 187. Ilaria, R. L., Jr., Hawley, R. G., and Van Etten, R. A. Dominant negative mutants implicate STAT5 in myeloid cell proliferation and neutrophil differentiation. Blood, 93: 4154-4166, 1999. 188. Teglund, S., McKay, C., Schuetz, E., et al. Stat5a and Stat5b proteins have essential and nonessential, or redundant, roles in cytokine responses. Cell, 93: 841-850, 1998. 189. Bunting, K. D., Bradley, H. L., Hawley, T. S., et al. Reduced lymphomyeloid repopulating activity from adult bone marrow and fetal liver of mice lacking expression of STAT5. Blood, 99: 479-487, 2002. 190. Snow, J. W., Abraham, N., Ma, M. C., et al. STAT5 promotes multilineage hematolymphoid development in vivo through effects on early hematopoietic progenitor cells. Blood, 99: 95-101, 2002. 191. Kieslinger, M., Woldman, I., Moriggl, R., et al. Antiapoptotic activity of Stat5 required during terminal stages of myeloid differentiation. Genes Dev, 14: 232- 244, 2000. 192. Fievez, L., Desmet, C., Henry, E., et al. STAT5 is an ambivalent regulator of neutrophil homeostasis. PLoS One, 2: e727, 2007. 193. Cohen, P. A., Koski, G. K., Czerniecki, B. J., et al. STAT3- and STAT5- dependent pathways competitively regulate the pan-differentiation of CD34pos cells into tumor-competent dendritic cells. Blood, 112: 1832-1843, 2008. 194. Levy, D. E. and Darnell, J. E., Jr. Stats: transcriptional control and biological impact. Nat Rev Mol Cell Biol, 3: 651-662, 2002. 195. Wei, L., Laurence, A., and O'Shea, J. J. New insights into the roles of Stat5a/b and Stat3 in T cell development and differentiation. Semin Cell Dev Biol, 19: 394-400, 2008. 196. Yao, Z., Cui, Y., Watford, W. T., et al. Stat5a/b are essential for normal lymphoid development and differentiation. Proc Natl Acad Sci U S A, 103: 1000-1005, 2006. 197. Hoelbl, A., Kovacic, B., Kerenyi, M. A., et al. Clarifying the role of Stat5 in lymphoid development and Abelson-induced transformation. Blood, 107: 4898- 4906, 2006. 198. Cao, X., Shores, E. W., Hu-Li, J., et al. Defective lymphoid development in mice lacking expression of the common cytokine receptor gamma chain. Immunity, 2: 223-238, 1995. 199. Thomis, D. C., Gurniak, C. B., Tivol, E., Sharpe, A. H., and Berg, L. J. Defects in B lymphocyte maturation and T lymphocyte activation in mice lacking Jak3. Science, 270: 794-797, 1995. 200. Yao, Z., Kanno, Y., Kerenyi, M., et al. Nonredundant roles for Stat5a/b in directly regulating Foxp3. Blood, 109: 4368-4375, 2007. 201. Burchill, M. A., Yang, J., Vang, K. B., et al. Linked T cell receptor and cytokine signaling govern the development of the regulatory T cell repertoire. Immunity, 28: 112-121, 2008.
175 202. Fornek, J. L., Tygrett, L. T., Waldschmidt, T. J., et al. Critical role for Stat3 in T- dependent terminal differentiation of IgG B cells. Blood, 107: 1085-1091, 2006. 203. Korn, T., Oukka, M., Kuchroo, V., and Bettelli, E. Th17 cells: effector T cells with inflammatory properties. Semin Immunol, 19: 362-371, 2007. 204. O'Shea, J. J. and Murray, P. J. Cytokine signaling modules in inflammatory responses. Immunity, 28: 477-487, 2008. 205. Takeda, K., Noguchi, K., Shi, W., et al. Targeted disruption of the mouse Stat3 gene leads to early embryonic lethality. Proc Natl Acad Sci U S A, 94: 3801- 3804, 1997. 206. McLemore, M. L., Grewal, S., Liu, F., et al. STAT-3 activation is required for normal G-CSF-dependent proliferation and granulocytic differentiation. Immunity, 14: 193-204, 2001. 207. Hortner, M., Nielsch, U., Mayr, L. M., et al. Suppressor of cytokine signaling-3 is recruited to the activated granulocyte-colony stimulating factor receptor and modulates its signal transduction. J Immunol, 169: 1219-1227, 2002. 208. Kimura, A., Kinjyo, I., Matsumura, Y., et al. SOCS3 is a physiological negative regulator for granulopoiesis and granulocyte colony-stimulating factor receptor signaling. J Biol Chem, 279: 6905-6910, 2004. 209. El Kasmi, K. C., Holst, J., Coffre, M., et al. General nature of the STAT3- activated anti-inflammatory response. J Immunol, 177: 7880-7888, 2006. 210. Takeda, K., Clausen, B. E., Kaisho, T., et al. Enhanced Th1 activity and development of chronic enterocolitis in mice devoid of Stat3 in macrophages and neutrophils. Immunity, 10: 39-49, 1999. 211. Williams, L. M., Sarma, U., Willets, K., et al. Expression of constitutively active STAT3 can replicate the cytokine-suppressive activity of interleukin-10 in human primary macrophages. J Biol Chem, 282: 6965-6975, 2007. 212. Yoshimura, A., Naka, T., and Kubo, M. SOCS proteins, cytokine signalling and immune regulation. Nat Rev Immunol, 7: 454-465, 2007. 213. Schindler, C. and Plumlee, C. Inteferons pen the JAK-STAT pathway. Semin Cell Dev Biol, 19: 311-318, 2008. 214. Coffer, P. J., Koenderman, L., and de Groot, R. P. The role of STATs in myeloid differentiation and leukemia. Oncogene, 19: 2511-2522, 2000. 215. Feldman, G. M., Rosenthal, L. A., Liu, X., et al. STAT5A-deficient mice demonstrate a defect in granulocyte-macrophage colony-stimulating factor- induced proliferation and gene expression. Blood, 90: 1768-1776, 1997. 216. Heng, D. Y., Chi, K. N., Murray, N., et al. A population-based study evaluating the impact of sunitinib on overall survival in the treatment of patients with metastatic renal cell cancer. Cancer, 115: 776-783, 2009. 217. Motzer, R. J., Hutson, T. E., Tomczak, P., et al. Sunitinib versus interferon alfa in metastatic renal-cell carcinoma. N Engl J Med, 356: 115-124, 2007. 218. Motzer, R. J., Rini, B. I., Bukowski, R. M., et al. Sunitinib in patients with metastatic renal cell carcinoma. Jama, 295: 2516-2524, 2006. 219. Roskoski, R., Jr. Sunitinib: a VEGF and PDGF receptor protein kinase and angiogenesis inhibitor. Biochem Biophys Res Commun, 356: 323-328, 2007.
176 220. Ozao-Choy, J., Ma, G., Kao, J., et al. The novel role of tyrosine kinase inhibitor in the reversal of immune suppression and modulation of tumor microenvironment for immune-based cancer therapies. Cancer Res, 69: 2514-2522, 2009. 221. Xin, H., Zhang, C., Herrmann, A., et al. Sunitinib inhibition of Stat3 induces renal cell carcinoma tumor cell apoptosis and reduces immunosuppressive cells. Cancer Res, 69: 2506-2513, 2009. 222. Hipp, M. M., Hilf, N., Walter, S., et al. Sorafenib, but not sunitinib, affects function of dendritic cells and induction of primary immune responses. Blood, 111: 5610-5620, 2008. 223. Fricke, I., Mirza, N., Dupont, J., et al. Vascular endothelial growth factor-trap overcomes defects in dendritic cell differentiation but does not improve antigen- specific immune responses. Clin Cancer Res, 13: 4840-4848, 2007. 224. Rodriguez, P. C., Ernstoff, M. S., Hernandez, C., et al. Arginase I-producing myeloid-derived suppressor cells in renal cell carcinoma are a subpopulation of activated granulocytes. Cancer Res, 69: 1553-1560, 2009. 225. Pytel, D., Sliwinski, T., Poplawski, T., Ferriola, D., and Majsterek, I. Tyrosine kinase blockers: new hope for successful cancer therapy. Anticancer Agents Med Chem, 9: 66-76, 2009. 226. Jemal, A., Siegel, R., Ward, E., et al. Cancer statistics, 2007. CA Cancer J Clin, 57: 43-66, 2007. 227. Finke, J. H., Rayman, P., Edinger, M., et al. Characterization of a human renal cell carcinoma specific cytotoxic CD8+ T cell line. J Immunother (1991), 11: 1- 11, 1992. 228. Finke, J. H., Rayman, P., Hart, L., et al. Characterization of tumor-infiltrating lymphocyte subsets from human renal cell carcinoma: specific reactivity defined by cytotoxicity, interferon-gamma secretion, and proliferation. J Immunother Emphasis Tumor Immunol, 15: 91-104, 1994. 229. Tatsumi, T., Kierstead, L. S., Ranieri, E., et al. MAGE-6 encodes HLA- DRbeta1*0401-presented epitopes recognized by CD4+ T cells from patients with melanoma or renal cell carcinoma. Clin Cancer Res, 9: 947-954, 2003. 230. Yang, J. C. and Childs, R. Immunotherapy for renal cell cancer. J Clin Oncol, 24: 5576-5583, 2006. 231. Brugarolas, J. Renal-cell carcinoma--molecular pathways and therapies. N Engl J Med, 356: 185-187, 2007. 232. Garcia, J. A. and Rini, B. I. Recent progress in the management of advanced renal cell carcinoma. CA Cancer J Clin, 57: 112-125, 2007. 233. Kaelin, W. G., Jr. The von Hippel-Lindau tumor suppressor protein and clear cell renal carcinoma. Clin Cancer Res, 13: 680s-684s, 2007. 234. Almand, B., Resser, J. R., Lindman, B., et al. Clinical significance of defective dendritic cell differentiation in cancer. Clin Cancer Res, 6: 1755-1766, 2000. 235. Gabrilovich, D. I., Chen, H. L., Girgis, K. R., et al. Production of vascular endothelial growth factor by human tumors inhibits the functional maturation of dendritic cells. Nat Med, 2: 1096-1103, 1996. 236. Gabrilovich, D. I., Ishida, T., Nadaf, S., Ohm, J. E., and Carbone, D. P. Antibodies to vascular endothelial growth factor enhance the efficacy of cancer
177 immunotherapy by improving endogenous dendritic cell function. Clin Cancer Res, 5: 2963-2970, 1999. 237. Ohm, J. E., Shurin, M. R., Esche, C., et al. Effect of vascular endothelial growth factor and FLT3 ligand on dendritic cell generation in vivo. J Immunol, 163: 3260-3268, 1999. 238. Saito, H., Tsujitani, S., Ikeguchi, M., Maeta, M., and Kaibara, N. Relationship between the expression of vascular endothelial growth factor and the density of dendritic cells in gastric adenocarcinoma tissue. Br J Cancer, 78: 1573-1577, 1998. 239. Gabrilovich, D. I., Velders, M. P., Sotomayor, E. M., and Kast, W. M. Mechanism of immune dysfunction in cancer mediated by immature Gr-1+ myeloid cells. J Immunol, 166: 5398-5406, 2001. 240. Kusmartsev, S. A., Li, Y., and Chen, S. H. Gr-1+ myeloid cells derived from tumor-bearing mice inhibit primary T cell activation induced through CD3/CD28 costimulation. J Immunol, 165: 779-785, 2000. 241. Mirza, N., Fishman, M., Fricke, I., et al. All-trans-retinoic acid improves differentiation of myeloid cells and immune response in cancer patients. Cancer Res, 66: 9299-9307, 2006. 242. Therasse, P., Arbuck, S. G., Eisenhauer, E. A., et al. New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst, 92: 205-216, 2000. 243. Mendel, D. B., Laird, A. D., Xin, X., et al. In vivo antitumor activity of SU11248, a novel tyrosine kinase inhibitor targeting vascular endothelial growth factor and platelet-derived growth factor receptors: determination of a pharmacokinetic/pharmacodynamic relationship. Clin Cancer Res, 9: 327-337, 2003. 244. Motzer, R. J., Bacik, J., Murphy, B. A., Russo, P., and Mazumdar, M. Interferon- alfa as a comparative treatment for clinical trials of new therapies against advanced renal cell carcinoma. J Clin Oncol, 20: 289-296, 2002. 245. Cesana, G. C., DeRaffele, G., Cohen, S., et al. Characterization of CD4+CD25+ regulatory T cells in patients treated with high-dose interleukin-2 for metastatic melanoma or renal cell carcinoma. J Clin Oncol, 24: 1169-1177, 2006. 246. Curiel, T. J. Tregs and rethinking cancer immunotherapy. J Clin Invest, 117: 1167-1174, 2007. 247. Dannull, J., Su, Z., Rizzieri, D., et al. Enhancement of vaccine-mediated antitumor immunity in cancer patients after depletion of regulatory T cells. J Clin Invest, 115: 3623-3633, 2005. 248. Tatsumi, T., Herrem, C. J., Olson, W. C., et al. Disease stage variation in CD4+ and CD8+ T-cell reactivity to the receptor tyrosine kinase EphA2 in patients with renal cell carcinoma. Cancer Res, 63: 4481-4489, 2003. 249. Vieweg, J., Su, Z., Dahm, P., and Kusmartsev, S. Reversal of tumor-mediated immunosuppression. Clin Cancer Res, 13: 727s-732s, 2007. 250. Ebos, J. M., Lee, C. R., Cruz-Munoz, W., et al. Accelerated metastasis after short- term treatment with a potent inhibitor of tumor angiogenesis. Cancer Cell, 15: 232-239, 2009.
178 251. Paez-Ribes, M., Allen, E., Hudock, J., et al. Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant metastasis. Cancer Cell, 15: 220-231, 2009. 252. Rodriguez, P. C. and Ochoa, A. C. Arginine regulation by myeloid derived suppressor cells and tolerance in cancer: mechanisms and therapeutic perspectives. Immunol Rev, 222: 180-191, 2008. 253. Mandruzzato, S., Solito, S., Falisi, E., et al. IL4Ralpha+ myeloid-derived suppressor cell expansion in cancer patients. J Immunol, 182: 6562-6568, 2009. 254. Peng, L., Kjaergaard, J., Plautz, G. E., et al. Tumor-induced L-selectinhigh suppressor T cells mediate potent effector T cell blockade and cause failure of otherwise curative adoptive immunotherapy. J Immunol, 169: 4811-4821, 2002. 255. Alexander, J. P., Kudoh, S., Melsop, K. A., et al. T-cells infiltrating renal cell carcinoma display a poor proliferative response even though they can produce interleukin 2 and express interleukin 2 receptors. Cancer Res, 53: 1380-1387, 1993. 256. Ebert, T., Bander, N. H., Finstad, C. L., Ramsawak, R. D., and Old, L. J. Establishment and characterization of human renal cancer and normal kidney cell lines. Cancer Res, 50: 5531-5536, 1990. 257. Steimer, D. A., Boyd, K., Takeuchi, O., et al. Selective roles for antiapoptotic MCL-1 during granulocyte development and macrophage effector function. Blood, 113: 2805-2815, 2009. 258. Ebos, J. M., Lee, C. R., Christensen, J. G., Mutsaers, A. J., and Kerbel, R. S. Multiple circulating proangiogenic factors induced by sunitinib malate are tumor- independent and correlate with antitumor efficacy. Proc Natl Acad Sci U S A, 104: 17069-17074, 2007. 259. Rini, B. I. Sunitinib. Expert Opin Pharmacother, 8: 2359-2369, 2007. 260. Corzo, C. A., Cotter, M. J., Cheng, P., et al. Mechanism regulating reactive oxygen species in tumor-induced myeloid-derived suppressor cells. J Immunol, 182: 5693-5701, 2009. 261. Rini, B. I., Campbell, S. C., and Escudier, B. Renal cell carcinoma. Lancet, 373: 1119-1132, 2009. 262. Koneru, R. and Hotte, S. J. Role of cytokine therapy for renal cell carcinoma in the era of targeted agents. Curr Oncol, 16 Suppl 1: S40-44, 2009. 263. Campoli, M., Ferrone, S., Zea, A. H., Rodriguez, P. C., and Ochoa, A. C. Mechanisms of tumor evasion. Cancer Treat Res, 123: 61-88, 2005. 264. Ohm, J. E. and Carbone, D. P. VEGF as a mediator of tumor-associated immunodeficiency. Immunol Res, 23: 263-272, 2001. 265. Ohm, J. E., Gabrilovich, D. I., Sempowski, G. D., et al. VEGF inhibits T-cell development and may contribute to tumor-induced immune suppression. Blood, 101: 4878-4886, 2003. 266. Rini, B. I. VEGF-targeted therapy in renal cell carcinoma: active drugs and active choices. Curr Oncol Rep, 8: 85-89, 2006. 267. Choueiri, T. K., Garcia, J. A., Elson, P., et al. Clinical factors associated with outcome in patients with metastatic clear-cell renal cell carcinoma treated with vascular endothelial growth factor-targeted therapy. Cancer, 110: 543-550, 2007.
179 268. Donskov, F. and von der Maase, H. Impact of immune parameters on long-term survival in metastatic renal cell carcinoma. J Clin Oncol, 24: 1997-2005, 2006. 269. van Cruijsen, H., van der Veldt, A. A., Vroling, L., et al. Sunitinib-induced myeloid lineage redistribution in renal cell cancer patients: CD1c+ dendritic cell frequency predicts progression-free survival. Clin Cancer Res, 14: 5884-5892, 2008. 270. Goddard, D. S., Yamanaka, K., Kupper, T. S., and Jones, D. A. Activation of neutrophils in cutaneous T-cell lymphoma. Clin Cancer Res, 11: 8243-8249, 2005. 271. Theilgaard-Monch, K., Porse, B. T., and Borregaard, N. Systems biology of neutrophil differentiation and immune response. Curr Opin Immunol, 18: 54-60, 2006. 272. Borregaard, N., Sorensen, O. E., and Theilgaard-Monch, K. Neutrophil granules: a library of innate immunity proteins. Trends Immunol, 28: 340-345, 2007. 273. Gerharz, C. D., Reinecke, P., Schneider, E. M., Schmitz, M., and Gabbert, H. E. Secretion of GM-CSF and M-CSF by human renal cell carcinomas of different histologic types. Urology, 58: 821-827, 2001. 274. Stie, J. and Jesaitis, A. J. Reorganization of the human neutrophil plasma membrane is associated with functional priming: implications for neutrophil preparations. J Leukoc Biol, 81: 672-685, 2007. 275. Zhu, D., Hattori, H., Jo, H., et al. Deactivation of phosphatidylinositol 3,4,5- trisphosphate/Akt signaling mediates neutrophil spontaneous death. Proc Natl Acad Sci U S A, 103: 14836-14841, 2006. 276. Hamilton, J. A. Colony-stimulating factors in inflammation and autoimmunity. Nat Rev Immunol, 8: 533-544, 2008. 277. Shurin, G. V., Shurin, M. R., Bykovskaia, S., et al. Neuroblastoma-derived gangliosides inhibit dendritic cell generation and function. Cancer Res, 61: 363- 369, 2001. 278. Sheppard, F. R., Kelher, M. R., Moore, E. E., et al. Structural organization of the neutrophil NADPH oxidase: phosphorylation and translocation during priming and activation. J Leukoc Biol, 78: 1025-1042, 2005. 279. Silliman, C. C., Ambruso, D. R., and Boshkov, L. K. Transfusion-related acute lung injury. Blood, 105: 2266-2273, 2005. 280. Dang, P. M., Dewas, C., Gaudry, M., et al. Priming of human neutrophil respiratory burst by granulocyte/macrophage colony-stimulating factor (GM-CSF) involves partial phosphorylation of p47(phox). J Biol Chem, 274: 20704-20708, 1999. 281. Borregaard, N. Current concepts about neutrophil granule physiology. Curr Opin Hematol, 3: 11-18, 1996. 282. Borregaard, N., Kjeldsen, L., Sengelov, H., et al. Changes in subcellular localization and surface expression of L-selectin, alkaline phosphatase, and Mac-1 in human neutrophils during stimulation with inflammatory mediators. J Leukoc Biol, 56: 80-87, 1994. 283. Coffer, P. J. and Koenderman, L. Granulocyte signal transduction and priming: cause without effect? Immunol Lett, 57: 27-31, 1997.
180 284. Kuijpers, T. W., Tool, A. T., van der Schoot, C. E., et al. Membrane surface antigen expression on neutrophils: a reappraisal of the use of surface markers for neutrophil activation. Blood, 78: 1105-1111, 1991. 285. McColl, S. R., Beauseigle, D., Gilbert, C., and Naccache, P. H. Priming of the human neutrophil respiratory burst by granulocyte-macrophage colony- stimulating factor and tumor necrosis factor-alpha involves regulation at a post- cell surface receptor level. Enhancement of the effect of agents which directly activate G proteins. J Immunol, 145: 3047-3053, 1990. 286. Mege, J. L., Gomez-Cambronero, J., Molski, T. F., Becker, E. L., and Sha'afi, R. I. Effect of granulocyte-macrophage colony-stimulating factor on superoxide production in cytoplasts and intact human neutrophils: role of protein kinase and G-proteins. J Leukoc Biol, 46: 161-168, 1989. 287. Sengelov, H., Boulay, F., Kjeldsen, L., and Borregaard, N. Subcellular localization and translocation of the receptor for N-formylmethionyl-leucyl- phenylalanine in human neutrophils. Biochem J, 299 ( Pt 2): 473-479, 1994. 288. Koenderman, L., Kanters, D., Maesen, B., et al. Monitoring of neutrophil priming in whole blood by antibodies isolated from a synthetic phage antibody library. J Leukoc Biol, 68: 58-64, 2000. 289. Currie, M. S., Rao, K. M., Padmanabhan, J., et al. Stimulus-specific effects of pentoxifylline on neutrophil CR3 expression, degranulation, and superoxide production. J Leukoc Biol, 47: 244-250, 1990. 290. Elbim, C., Reglier, H., Fay, M., et al. Intracellular pool of IL-10 receptors in specific granules of human neutrophils: differential mobilization by proinflammatory mediators. J Immunol, 166: 5201-5207, 2001. 291. Condliffe, A. M., Davidson, K., Anderson, K. E., et al. Sequential activation of class IB and class IA PI3K is important for the primed respiratory burst of human but not murine neutrophils. Blood, 106: 1432-1440, 2005. 292. Gaudry, M., Bregerie, O., Andrieu, V., et al. Intracellular pool of vascular endothelial growth factor in human neutrophils. Blood, 90: 4153-4161, 1997. 293. Guthrie, L. A., McPhail, L. C., Henson, P. M., and Johnston, R. B., Jr. Priming of neutrophils for enhanced release of oxygen metabolites by bacterial lipopolysaccharide. Evidence for increased activity of the superoxide-producing enzyme. J Exp Med, 160: 1656-1671, 1984. 294. Jacobsen, L. C., Theilgaard-Monch, K., Christensen, E. I., and Borregaard, N. Arginase 1 is expressed in myelocytes/metamyelocytes and localized in gelatinase granules of human neutrophils. Blood, 109: 3084-3087, 2007. 295. Xiao, W., Hong, H., Kawakami, Y., Lowell, C. A., and Kawakami, T. Regulation of myeloproliferation and M2 macrophage programming in mice by Lyn/Hck, SHIP, and Stat5. J Clin Invest, 118: 924-934, 2008.
181