University of Southern California Dissertations and Theses

Human myeloid-derived suppressor cells in cancer: Induction, functional characterization, and therapy

(USC Thesis Other)

Human myeloid-derived suppressor cells in cancer: Induction, functional characterization, and therapy

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content HUMAN MYELOID-DERIVED SUPPRESSOR CELLS IN CANCER:
INDUCTION, FUNCTIONAL CHARACTERIZATION, AND THERAPY
by
Melissa Genevieve Lechner
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(SYSTEMS BIOLOGY AND DISEASE)
May 2011
Copyright 2011 Melissa Genevieve Lechner
ii
EPIGRAPH
The joy of discovery is certainly the liveliest that the mind of man can ever feel.
- Claude Bernard (1813-78) French physiologist
iii
DEDICATION
To Trevor, for all the reasons.
To my parents, for their unfailing support and encouragement,
which gives me the confidence to seek answers to these questions.
iv

ACKNOWLEDGEMENTS

Thesis Committee Members:
Clive R. Taylor, MD, PhD (Chair)
Alan L. Epstein, MD, PhD (Mentor)
Harvey Kaslow, PhD
Minnie McMillan, PhD
David Horwitz, MD, PhD
Alicia McDonough, PhD

This work would not have been possible without the constant encouragement and
guidance of my mentor Alan Epstein. During my time in the laboratory he has invested
significant time and resources into the development of my technical and intellectual skills
as a scientist. At every turn he has given me opportunities to participate in project design
and analysis and to present my research findings, in person and in print, at venues
ranging from Keystone Symposia to CNN. I thank him especially for his patience and
kindness throughout my PhD training.

I would like to acknowledge Carolina Megiel, who in her capacity as a researcher in the
Epstein laboratory during my tenure has spent the past three years working with me to
develop an in vitro model of human myeloid suppressor cells and uncover new therapies
to target these cells in cancer patients and experimental tumor models. She has been
v

instrumental in helping me to execute pilot studies of a myeloid-suppressor cell clinical
assay in head and neck cancer patients and has contributed greatly to our understanding
of human myeloid-suppressor cells.

I would also like to acknowledge Sarah M. Russell and Daniel J. Liebertz, who as
medical students in the Dean's Research Scholar program worked with me in the
laboratory to develop and characterize head and neck cancer cell lines. In conjunction
with immune profiling studies in head and neck cancer, these individuals assisted with
concurrent myeloid suppressor cell studies in models of this disease.

I would like to particularly acknowledge the expert work of Lillian Young in performing
the immunohistochemistry studies. I would like to thank also Dixon Gray for flow
cytometry training, James Pang for assistance with animal studies, and Mandy Han for
assistance with laboratory techniques. I wish to acknowledge the expert work of Victoria
Bedell and the City of Hope Cytogenetic Core Facility in performing the cytogenetic and
HPV in situ hybridization studies.

vi

The Epstein laboratory trains many medical, graduate (PhD, Masters), undergraduate, and
high school students and during my tenure I would like to acknowledge the following
individuals with whom I have had the pleasure of working:
Brigid (Quigley) Bingham, Nicholas Arger, Tammy Woo, Lucy Gong, Laura Andrews,
Samantha Lasarow, Connor Church, Kieumai Vo, Nicholas Landsman, Nathan Feng,
Nicole Feng, Corey Kelsom, Cynthia Neben, Julie Jang

I would like to thank and acknowledge our clinical collaborators in the Department of
Otolaryngology - Head and Neck Surgery (Uttam Sinha, Dennis Maceri, Niels Kokot,
Dale Rice, and Rizwan Masood), Department of Pathology (Adrian Correa), and Division
of Diabetes and Endocrinology, Department of Medicine (Jonanthan LoPresti) at USC for
facilitating translation of our work from benchtop to bedside.

I would like to acknowledge the many friends and family who supported me through this
endeavor, whether it was through early morning runs, a glass of wine, or doing my
laundry.

I would particularly like to thank my parents for providing me with the seed of scientific
curiosity at an early age and the opportunities to foster that interest. I thank my brother
Andrew and Trevor Angell for providing enthusiasm during moments of discovery and
empathy during moments of frustration.

vii

Financial support for these studies was provided by the following organizations:
National Institutes of Health Cellular, Biochemical and Molecular Sciences Training
Grant at the USC Keck School of Medicine to support graduate thesis research on
myeloid-derive suppressor cells in cancer.

University of Southern California Steven's Institute for Innovation Ideas Empowered
Program Award to support the development of a novel clinical assay for the detection of
myeloid suppressor cells in cancer patients.

Philanthropic Educational Organization Scholars Award to support women pursuing
graduate research awarded in research funds to support doctoral dissertation research in
tumor immunology.

Cancer Therapeutics Laboratories, Inc., Los Angeles, California
vi
TABLE OF CONTENTS
Epigraph ii
Dedication iii
Acknowledgements iv
List of Tables viii
List of Figures ix
Abstract xii
Chapter 1. Introduction 1
Chapter 2. Induction of human myeloid-derived suppressor cells by a
diverse set of human tumor cell lines
10
Chapter 3. Functional characterization of CD33
+
and CD11b
+
human
myeloid-derived suppressor cells induced by human solid
tumor cell lines
39
Chapter 4. Functional characterization of human myeloid-derived
suppressor cells induced by cytokines from normal
peripheral blood mononuclear cells
64
Chapter 5. USC-HN2, a new model cell line for recurrent oral cavity
squamous cell carcinoma with immunosuppressive
characteristics
91
Chapter 6. Conclusions and future directions 120
Bibliography 144
Appendix 156
viii
LIST OF TABLES
Table 1: Forty-five of 100 human solid tumor cell lines induce functionally
suppressive CD33
+
myeloid suppressor cells from volunteer
normal human PBMC after one-week co-culture in vitro
21
Table 2: Analysis of USC-HN2 surface markers by flow cytometry 111
Table 3: Selected up-regulated genes identified in USC-HN2 and SCCL-
MT1 cell lines also present in HNSCC tumor biopsies
114
Table 4: GSEA identified up-regulated genes in both USC-HN2 and SCCL-
MT1 cells compared with USC-HN1 cells
115
Table 5: Summary of current therapies for MDSC 134
ix
LIST OF FIGURES
Figure 1: Tumor immune escape is associated with decreased cell-mediated
immunity and accumulation of immune suppressor cells
5
Figure 2: The development of myeloid-derived-suppressor cells 8
Figure 3: Schematic of Co-culture and MDSC Suppression Assays for the in
vitro generation of tumor associated myeloid suppressor cells
19
Figure 4: HNSCC-induced MDSC inhibit autologous T cell proliferation and
IFNγ production
24
Figure 5: Expression of putative MDSC-inducing factors by human solid
tumor cell lines and PBMC
26
Figure 6: MDSC-inducing cell lines produce increased IL-1 β, IL-6, TNF α,
and VEGF
29
Figure 7: Removal of GM-CSF, IL-6, or IL-1β from co-culture impairs
CD33
+
MDSC induction by tumor cell lines
30
Figure 8: Cytokine-induced CD33
+
MDSC demonstrate potent suppressive
function
33
Figure 9: Up-regulation of GM-CSF by PBMC in tumor co-culture
accompanies induction
33
Figure 10: Induction of a second CD11b
+
MDSC subset by breast, lung, and
brain cancer cell lines
35
Figure 11: Morphology of human MDSC 48
Figure 12: Human CD33
+
and CD11b
+
MDSC are distinct subsets with a
common HLA-DR
low
Lineage
-
phenotype
50
Figure 13: Human MDSC mediate suppression through up-regulation of
ARG-1, NOX2, iNOS, VEGF, and TGF β
53
Figure 14: Transcription factors driving human MDSC suppressive function 57
Figure 15: Cytokine-induced CD33
+
MDSC demonstrate potent suppressive
function
75
x
Figure 16: Morphology of CD33
+
suppressor cells resembles tumor-induced
MDSC
78
Figure 17: Phenotype of cytokine-induced CD33
+
MDSC 80
Figure 18: The expression of suppressive genes by cytokne- and tumor-
induced CD33
+
MDSC varies with the inducing cytokine milieu
82
Figure 19: CD33
+
MDSC-mediated suppression of autologous T cells is
contact dependent
83
Figure 20: Cellular context for cytokine induction of CD33
+
MDSC from
normal donor PBMC
85
Figure 21: Histology of the original tumor and heterotransplants 103
Figure 22: Morphologic and cytogenetic analysis of USC-HN2 cell line 106
Figure 23: Characterization of the original tumor biopsy, USC-HN2 cell line,
and heterotransplanted tumor
109
Figure 24: USC-HN2 is highly immunomodulatory and induces suppressor
cells
112
Figure 25: In situ Hybridization for the presence of Human Papillomavirus
Type 16 and 18
115
Figure 26: Schematic showing a novel, minimally-invasive clinical assay for
cancer detection and monitoring
124
Figure 27: Preliminary data demonstrating MDSC in the peripheral blood of
cancer patients using a recently identified phenotype: CD33
+
HLA-
DR
low
HIF1a
+
127
Figure 28: Immunotherapy of Colon 26 tumor model shows cure when
immune stimulation is paired with regulatory T cell depletion
130
Figure 29: Immunohistochemistry studies of tumor-infiltrating leukocytes in
4T1 breast carcinoma murine tumor model
132
Figure 30: Immunotherapy of 4T1 breast carcinoma in BALB/c mice with
simulataneous Treg and MDSC depletion
137
xi
Figure 31: Schematic for the induction of human MDSC in cancer. 142
Figure 32: Figure 32. Reversal of CD33
+
MDSC suppressive function by
celecoxib and analogs through a non-COX2 dependent mechanism.
156
xii
ABSTRACT
Tumor immune tolerance can derive from the recruitment of suppressor cell populations,
including myeloid-derived suppressor cells (MDSC). MDSCs inhibit anti-tumor T cell
responses through a variety of mechanisms including nutrient depletion, production of
reactive oxygen and nitrogen species, VEGF expression, and regulatory T cell expansion.
In cancer patients and murine tumor models, MDSC accumulation correlates with
increased stage and tumor burden, but the frequency and mechanisms of MDSC
induction in human cancer remain poorly understood. This study examined the ability of
a diverse set of human solid tumor cell lines to induce MDSC using in vitro tumor co-
culture methods. Newly induced suppressor cells were characterized for morphology,
phenotype, and gene expression. Of over 100 solid tumor cell lines examined, 45
generated canonical CD33
+
HLA-DR
low
Lineage
-
MDSC, with high frequency of induction
by cervical, ovarian, colorectal, renal cell, and head and neck squamous cell carcinoma
cell lines. CD33
+
MDSC could be induced by some cancer cell lines of all tumor types
examined with the notable exception of breast cancer cell lines (0/9, inclusive of models
with different hormone and HER2 mutation status). Upon further examination, breast
cancer cell lines and other tumor types with infrequent CD33
+
MDSC induction
preferentially induced an undiscovered CD11b
+
CD33
low
HLA-DR
low
Lineage
-
MDSC
subset. Gene and protein expression, neutralization, and cytokine induction experiments
determined that the ability of tumor cell lines to induce CD33
+
MDSC depended upon
over-expression of IL-1 β, IL-6, TNF α, VEGF, and GM-CSF, while CD11b
+
MDSC
xiii
induction correlated with over-expression of FLT3L and TGF β. Both CD33
+
and
CD11b
+
MDSC subsets appeared as immature myeloid cells by Wright-Giemsa staining
and had significantly up-regulated expression of inducible nitric oxide synthase, TGF β,
NADPH oxidase, VEGF, and arginase-1 genes compared with normal myeloid cells.
Furthermore, increased expression of transcription factors HIF1 , STAT3, and C/EBP 
distinguished suppressive from non-suppressive tumor cell line-educated myeloid cells.
Interestingly, CD33
+
and CD11b
+
subsets showed differential expression of these
transcription factors and therapeutic reversal of suppressive function coincided with
decreasing STAT3 and HIF1 α in CD33
+
cells but with decreasing C/EBP β in CD11b
+
MDSC. These studies demonstrate the universal nature of MDSC induction by human
cancers and characterize two distinct populations of MDSC recognized by CD33
+
HLA-
DR
low
HIF1 α
+
and CD11b
+
HLA-DR
low
C/EBP β
+
biomarkers. The unraveling of these
different subsets in human cancers should enable the development of novel diagnostic
and therapeutic reagents for cancer immunotherapy.
1
CHAPTER 1.
INTRODUCTION
Cancers, or malignant tumors, are populations of abnormal cells that proliferate
uncontrollably, invade surrounding tissues, and eventually spread to distant sites in the
body and kill the host [1]. Cancerous lesions have been described by medical
practitioners and healers for centuries, and the disease continues to cause significant
human morbidity and mortality. The American Cancer Society estimated 1,529,560 new
cases diagnosed in 2010 and a lifetime risk of cancer of one in two or one in three for
men and women, respectively [2]. The past decades have seen great advances in the
diagnosis, management, and treatment of cancer, but, with the exception of prophylactic
vaccinations for Human Papilloma Virus
+
cervical carcinoma and Hepatitis B Virus
+
hepatocellular carcinoma, a cure has remained elusive. Although surgery, radiation
therapy, and chemotherapy remain mainstays of current treatment for solid tumors, these
modalities are largely ineffective against metastatic disease and the elimination of
minimal residual disease. These conventional therapies also have serious limitations
including lack of tumor specificity, significant toxicities and damage to normal tissues.
The promise of immunotherapy
It is now understood that the immune system is capable of recognizing and eliminating
cancer cells, but that tumors evade and suppress host immune responses and thus persist
and spread [3,4]. Immunotherapy seeks to overcome tumor-mediated immune
dysfunction and activate a cell-mediated immune response against cancer cells [5]. Such
2
an approach holds great promise for reducing damage to collateral tissue by taking
advantage of the inherent specificity of the human immune system. Systemic trafficking
and monitoring by immune cells also provides for superior treatment of metastatic and
inoperable lesions compared with external beam irradiation and surgical therapies.
Perhaps most importantly, the generation of immunologic memory following a robust
anti-tumor immune response prevents the recurrence of tumors.
Tumor immune escape
Despite active surveillance and elimination of cancer cells by the immune system, many
tumors persist and grow to reach clinically relevant size. These tumors escape immune
destruction by evasion or suppress anti-tumor immunity and overwhelm host defenses
(Figure 1). Neoplastic cells derive from normal cells but have lost physiologic control of
cell cycle, differentiation, and adhesion. Therefore they often closely resemble normal
tissue at the molecular level. The majority of tumor-associated antigens (TAA) are self
antigens to which the body is tolerized at birth (thymic deletion of autoreactive T cells) or
later in the periphery (e.g. induced regulatory T cells during chronic inflammation) [4].
In addition, many tumors down-regulate expression of antigen display receptors on their
cell surfaces [e.g. major histocompatibility complexes (MHC)] to decrease display of
normal and mutated cell proteins [4,6]. During tumor progression, immunoediting leads
to the selection of tumor subclones that lack expression of immunogenic TAA and are
resistant to natural killer (NK) cells that kill cells with low MHC I expression [4,6].
These adaptations challenge the ability of the immune system to identify and eliminate
3
abnormal cells, and may be the main mechanism by which poorly immunogenic and
slowly growing cancers escape. In the case of aggressively dividing and invasive cancer
cells, additional mechanisms act to achieve tumor immune tolerance. With rapid cellular
division these tumors quickly outgrow the nutrient and oxygen supply of their
vasculature, producing a hypoxic environment and an inner core of necrotic cells [7].
The expanding tumor mass displaces and invades nearby tissues, causing cell death in
the tumor microenvironment and the release of pro-inflammatory signals. The
confluence of necrosis and inflammation attracts innate immune cells to the tumor site,
and antigen presenting cells may become activated and migrate to draining lymphoid
tissues to recruit adaptive immune cells [7]. As this nascent anti-tumor immune response
is generated, successful tumors adapt to suppress the host immune system. Infiltrating
effector cells become anergic or apoptotic upon entering the tumor microenvironment
and fail to kill their target tumor cells [3,4]. Tumors show diverse adaptations to inhibit
effector immune cells, including surface expression of cytotoxic and tolerizing ligands
(e.g. B7-H4, programmed death ligands (PDL)-1 and 2, Fas Ligand) and the release of
soluble immune modulatory factors in the tumor microenvironment (Th2 cytokines: IL-4,
IL-5, transforming growth factor ) [3-7]. In addition to direct suppression of immune
cells, tumors generate tolerance indirectly through the recruitment and activation of
suppressor cell populations, namely regulatory T cells (Treg), myeloid-derived
suppressor cells (MDSC), tumor-associated macrophages (TAM), plasmacytoid dendritic
cells (pDC), and NK regulatory cells [4,6].
4
These tumor-mediated changes are often accompanied by a shift toward humoral or Th2
immunity, and away from cell-mediated Th1 type immunity [3-5]. Like viral infection,
cancer is a disease state in which host cells become abnormal and their elimination is
needed to control or eradicate the disease. Recognition and elimination of damaged or
infected cells occurs when innate immune cells act cooperatively with Th1 helper T cells
and cytotoxic T lymphocytes (CTL). Thus Th2 shifted immunity results in impaired
ability to kill abnormal self cells [3]. In summary, failed anti-tumor immunity occurs by
the accumulation of immune suppressor cells and the concurrent absence of activated
cell-mediated effector cells (CTL, NK cells) and cytokines (Th1) (Figure 1). In
developing an immunotherapy for cancer, reversing tolerance and stimulating antigen-
specific immune activation are the two main strategies most likely to yield successful
therapy.
Immune suppressor cells as a barrier to effective immunotherapy
Previously, Epstein and colleagues [5] showed that while different tumors may elicit
tolerance and immune dysfunction via distinct mechanisms, a common result is tumor
suppression of activated cell-mediated anti-tumor immunity. An effective
immunotherapy for cancer will require targeted and antigen specific immune activation
concurrent with the reversal of tumor-driven immune suppression. Immune activation
may be achieved through tumor lysate or DC vaccination, adoptive transfer of transgenic
T cells, and infusion of stimulatory immunoligands and cytokines. Many currently
approved immunotherapies for cancer rely upon such immune stimulation, but have
5
failed to reverse tumor immune tolerance, the second half of any effective
immunotherapy, and consequently have produced only minor clinical benefits [6].
Figure 1. Tumor immune escape is associated with decreased cell-mediated immunity and
accumulation of immune suppressor cells. Tumor immune surveillance mechanisms, including antigen
presenting cells (dendritic cells (DC), macrophages (M φ)) and effector cells (T helper (Th) cells, cytotoxic
T lymphocytes (CTL), natural killer (NK) cells), recognize and attack neoplastic cells in the normal host to
control tumor growth. When tumor immune escape occurs, immune suppressor cells recruited to the tumor
microenvironment and secondary lymphoid tissues inhibit cell-mediated immunity to facilitate tumor
progression. Immune suppressor cells that accumulate in the tumor setting include derivatives of innate
(myeloid-derived suppressor cells (MDSC), plasmacytoid DC (pDC), tumor-associated M φ (TAM)) and
adaptive (regulatory T cells (Treg)). In addition to the naturally occurring Treg populations that inhibit
reactions to self antigens at steady state, induced Treg cells arise in the periphery in the setting of chronic
inflammation and cancer. Lastly, a major shift in immune cytokines from cell-mediated promoting (Th1)
cytokines (interferon γ, tumor necrosis factor α, interleukin (IL)-2) in effective tumor immune surveillance
to humoral immunity promoting Th2 cytokines (IL-4, IL-5) in the cancer setting.
6
Immune suppressor cells are now recognized as a key component of tumor immune
tolerance and the major impediment to successful immunotherapy in cancer patients and
experimental solid tumor models. These regulatory cell populations, including Treg and
MDSC, have become increasingly the focus of immunotherapy studies and great strides
have been made in understanding their biology.
Regulatory T cells (Treg)
It is now well established that Tregs are a subset of CD4
+
T cells present in normal mice
and humans, and are essential for both homeostasis and the maintenance of tolerance to
tissue-specific antigens [8,9]. Naturally occurring T regulatory (nTreg) cells arise in the
thymus, constitutively display CD4, CD25, CTLA-4, GITR, OX40, and FoxP3
High
and
classically suppress the proliferation of responder T cells in vitro through a cell contact-
dependent, cytokine-independent mechanism [10]. The in vivo mechanisms of nTreg
suppression are incompletely characterized, but include expression of surface molecules
(e.g. CTLA-4), membrane bound cytokines (e.g. TGF- β), and pericellular generation of
adenosine [11]. Adaptive or induced regulatory T cells (iTreg) arise in the periphery
after TCR stimulation in the context of IL-10 or TGF- β in the case of T regulatory type 1
(Tr1) or T helper type 3 (Th3) subsets, respectively [11]. In contrast to nTreg, iTregs
exert suppression primarily through soluble factors (IL-10 and/or TGF- β) and their
suppressive function is not strictly associated with high FoxP3 expression. Tumor
tolerance develops when immune suppression is favored so that Tregs detrimentally
suppress an anti-tumor immune response. Tregs in the tumor microenvironment arise
7
from proliferation of pre-existing Tregs as well as conversion of peripheral naïve
CD4
+
CD25
-
T cells into newly derived Tregs [12]. Tumor-derived or MDSC-derived
TGF- β and indoleamine 2,3-dioxygenase (IDO) activity, as well as suboptimal antigen
presentation by MDSC, may play a role in Treg expansion and functional activation [12].
Potential therapeutic approaches to relieving Treg-mediated tumor tolerance include:
blocking generation of new Tregs; inactivating existing Tregs; and inhibiting the
suppression by Tregs of effector T cells [11]. Treg-mediated tumor tolerance is but one
mechanism of tumor immune evasion, and a failure to address other mechanisms and
their interaction with Tregs may limit the ability to rescue anti-tumor immune responses.
Myeloid-derived suppressor cells
Myeloid-derived suppressor cells (MDSC) have recently been recognized as a subset of
innate immune cells that can alter adaptive immunity and produce immunosuppression
[4]. In mice, MDSC are identified by CD11b
+
, IL-4R α
+
, and GR-1
low/int
expression, with
recognized granulocytic and monocytic subsets [13-17]. Human MDSC are less
understood and comprise a heterogeneous population of immature myeloid (CD33
+
) cells
consisting of dendritic cell, macrophage, and granulocyte progenitors that lack lineage
maturation markers (Figure 2) [13,16]. MDSC inhibit T cell effector functions through a
range of mechanisms, including: arginase 1 (ARG-1)-mediated depletion of L-arginine
[18]; inducible nitric oxide synthase (iNOS) and NADPH oxidase (NOX2) production of
reactive nitrogen and oxygen species [19,20]; vascular endothelial growth factor over-
expression [21]; cysteine depletion [22]; and the expansion of T-regulatory (Treg) cell
8
populations [23,24]. While rare or absent in healthy individuals, MDSC accumulate in
the settings of trauma, severe infection or sepsis, and cancer (Figure 2) [17], possibly as a
result of the hypoxic environment and hypoxia-inducible factor (HIF)-1α expression
[25,26].
Figure 2. The development of myeloid-derived-suppressor cells. Myeloid-derived suppressor cells arise
from common myeloid progenitors in the setting of cancer, trauma, or severe infection. MDSCs accumulate
at the site of pathology as well as in peripheral lymphoid organs in response to tumor-derived factors or
systemic cytokines. (Adapted from Gabrilovich DI and S Nagaraj. 2009. Nature Reviews, 9: 162.
Understanding the mechanisms by which tumors recruit and activate MDSC and other
suppressor cell populations is central to combating tumor-driven immune tolerance.
Identification of the major transcription factors that drive MDSC suppressive function
and/or survival could allow inhibition of all MDSC subsets by a common master switch.
Given the multitude of immune modulatory factors produced by tumors, different subsets
of MDSC may be generated in the setting of neoplasia, depending on the unique profile
of factors secreted by the tumor [27-30]. Development of robust preclinical models of
9
tumor-induced MDSC could significantly aid the study of induction and suppressive
function for the development of translational MDSC therapies in humans. In this study,
human tumor cell line-induced MDSC were used to characterize the morphology,
phenotype, gene expression and function of human MDSC, yielding several targets for
inhibition of this suppressor cell population in cancer.
10
CHAPTER 2.
INDUCTION OF HUMAN MYELOID-DERIVED SUPPRESSOR CELLS
BY A DIVERSE SET OF HUMAN TUMOR CELL LINES
Immune suppressor cells, such as myeloid-derived suppressor cells (MDSC) are
increasingly recognized as contributing to tumor immune tolerance and the failure of
immunotherapy regimens in cancer patients and experimental tumor models [4,6]. As a
recently discovered population of immune regulatory cells, MDSC remain poorly
understood, particularly in humans [13]. Because MDSC are absent or rare in healthy
hosts but common in patients with significant illness (cancer, trauma, sepsis), obtaining
sufficient numbers of human MDSC for study is challenging and has limited our
understanding of these cells in malignant disease. To elucidate better the contribution of
human MDSC to tumor immune tolerance and identify potential therapeutic targets for
inhibition, we need to learn more about their prevalence, phenotype, and functions. As in
mice, human MDSC comprise a heterogeneous population of immature myeloid (CD33
+
)
cells consisting of dendritic cell (DC), macrophage, and granulocyte progenitors that lack
lineage maturation markers and inhibit effector T cell functions [16]. MDSC have been
identified in patients with colon carcinoma [31], melanoma [29,30], hepatocellular
carcinoma [32,33], head and neck squamous cell carcinoma [20], non-small cell lung
carcinoma [34], renal cell carcinoma [35,36], pancreatic adenocarcinoma [37], ovarian
[38], and breast carcinoma [39], and multiple myeloma [40]. In each case, their
accumulation correlates with more advanced disease and poorer prognosis [39].
However, many clinical reports have relied upon enumerating cells with poorly
11
understood phenotypes, rather than upon functional studies demonstrating a role of
suppressive myeloid populations, leaving the true meaning of MDSC induction unclear.
Increasing evidence suggests that MDSC are recruited by tumor-derived factors including
IL-1 β, IL-6, vascular endothelial growth factor (VEGF), granulocyte/macrophage colony
stimulating factor (GM-CSF), macrophage colony stimulating factor (M-CSF), stem cell
factor (c-kitL), fms-related tyrosine kinase 3 ligand (FLT3L), and cyclo-oxygenase 2
(COX2)-derived prostaglandins [27,28,41-43]. Importantly, some of these are also
associated with tissue hypoxia and inflammation [26,44]. Activation of MDSC also is
still poorly understood, but evidence supports a role for interferon gamma (IFN-γ), IL-4,
IL-13, transforming growth factor beta (TGF- β), and the toll-like receptor ligand,
lipopolysaccharide endotoxin [16,42,45]. Given the multitude of immune modulatory
factors produced by tumors, different subsets of MDSC may be generated depending on
the unique profile of factors secreted by the tumor [27-29]. Preclinical models of human
tumor-induced MDSC will significantly advance knowledge of their induction and
function in the tumor setting.
We report here the development of a novel in vitro method to induce human MDSC from
healthy donor peripheral blood mononuclear cells (PBMC) by co-culturing these PBMC
with human solid tumor cell lines and subsequently measuring their suppressive ability.
Using this model system, we have determined the frequency of MDSC induction in
human cancers of varied histiologic types, and have elucidated key tumor-derived factors
12
that drive MDSC induction. Our methods generated highly purified human MDSC in
quantities sufficient for morphology, phenotype, gene expression, and functional studies
and pre-clinical testing of MDSC inhibitors, as described in subsequent chapters
1
.
MATERIALS AND METHODS
Cell Lines and Cell Culture
Tumor cell lines were obtained from the American Type Culture Collection (ATCC) or
gifted to the Epstein laboratory. Notable gifts include the SW cell lines from the Scott
and White Clinic (Temple, TX) and pancreatic cell lines from Dr. Liz Jaffe (Johns
Hopkins Medical Center, Baltimore, MD). Tumor cell line authenticity was performed
by cytogenetics and surface marker analysis performed at ATCC or in our laboratory. All
cell lines were maintained at 37
o
C in complete medium [(RPMI-1640 with 10% fetal calf
serum (characterized FCS, Hyclone, Inc., Logan, UT), 2mM L-Glutamine, 100U/mL
Penicillin, and 100µg/mL Streptomycin)] in tissue culture flasks in humidified, 5% CO
2
incubators and passaged 2-3 times per week by light trypsinization.
Tumor-Associated MDSC Generation Protocol
i. Induction by tumor cell lines
1
Portions of this chapter were previously published in the Journal of Immunology (Lechner et al. Characterization of
cytokine-induced myeloid-derived suppressor cells from normal human peripheral blood mononuclear cells. Journal of
Immunology. 2010;185(4): 2273-2284) or are undergoing revision for publication (Lechner et al. Functional
characterization of human myeloid-derived suppressor cells induced from peripheral blood mononuclear cells cultured
with a diverse set of human tumor cell lines. Journal of Immunology, under review)
13
Human PBMC were isolated from healthy volunteer donors by venipuncture (60mL total
volume) followed by differential density gradient centrifugation (Ficoll Hypaque, Sigma,
St. Louis, MO). PBMC were cultured in complete medium (6x10
5
cells/mL,
supplemented with rhGM-CSF (10ng/mL, R&D Systems, Minneapolis, MN) to support
viability) in T-25 culture flasks with human tumor cell lines for one week. Tumor cells
were seeded to achieve confluence by day 7 (approximately 1:100 ratio with PBMC), and
samples in which tumor cells overgrew were excluded from analysis and were repeated
with adjusted ratios. Alternatively, irradiated tumor cells (3500 rad) were initially seeded
at 1:10 ratio in co-cultures to examine whether induction was dependent upon actively
dividing tumor cells. PBMC cultured in medium alone were run in parallel as an
induction negative control for each donor to control for any effects of FCS. For
neutralization experiments, PBMC-tumor cell line co-cultures were repeated in the
presence or absence of neutralizing monoclonal antibodies for a subset of HNSCC cell
lines: anti-VEGF (Avastin, Genetech, San Francisco, CA), anti-TNF  (Humira, Abbott,
Abbott Park, IL), anti-IL-1 (clone AB-206-NA, Abcam, Cambridge, MA), anti-IL-6
(clone AB-201-NA, Abcam), or anti-GM-CSF (clone BVD2). CD33
+
or CD11b
+
cells
were subsequently isolated and tested for suppressive function as described below. For
these studies 39 male and 22 female donors ages 23 to 62 were used under USC
Institutional Review Board-approved protocol HS-06-00579. Data were derived from at
least two individuals and no inter-donor differences in MDSC induction or function were
observed.
14
ii. Induction by cytokines
Human PBMC were isolated from healthy volunteer donors by venipuncture, followed by
differential density gradient separation (Ficoll Hypaque, Sigma, St. Louis, MO). PBMC
were cultured in T-25 flasks at 5x10
5
cells/mL in complete medium for seven days,
supplemented with cytokines as indicated. Recombinant human cytokines used for
induction include: IL-1β (10ng/mL, Sigma), IL-6 (10ng/mL, Sigma), PGE
2
(1µg/mL,
Sigma), TGF β1 (2ng/mL, R&D Systems), TNF α (10ng/mL, R&D), VEGF (10ng/mL,
Sigma), and GM-CSF (20ng/mL, R&D). For combination-cytokine induction
experiments, the cytokines used are indicated in the Results section. PBMC cultured in
medium alone were run in parallel as a control for each donor. Cultures were run in
duplicate, and medium and cytokines were refreshed every two-three days. For all
studies, USC Institutional Review Board approval was obtained and a total of 61 unique
donors (39 male, 22 female) ages 23-62 were used under protocol HS-06-00579. Data
were derived from at least two individuals and no inter-donor differences in MDSC
induction or function were observed.
iii. MDSC Isolation
After one week, all cells were collected from tumor-PBMC co-cultures. Adherent cells
were removed using the non-protease cell detachment solution Detachin (GenLantis, San
Diego, CA). Myeloid (CD33
+
) cells were then isolated from the co-cultures using anti-
15
CD33 magnetic microbeads and LS column separation (Miltenyi Biotec, Germany) as per
manufacturer’s instructions. Purity of isolated cell populations was found to be greater
than 90% by flow cytometry and morphological examination. Viability of isolated cells
was confirmed using trypan blue dye exclusion and samples with viability less than 80%
were excluded from analysis and repeated in subsequent experiments. Yields of CD33
+
cells from co-cultures averaged 8.75% of PBMC (range 6-12%), while yields from
medium only cultures averaged 10%. For isolation of CD11b
+
suppressor cells from
breast carcinoma co-cultures, anti-CD11b microbeads were used instead of anti-CD33
microbeads.
iv. Suppression Assay
The suppressive function of tumor-educated CD33
+
or CD11b
+
cells was measured by
their ability to inhibit the proliferation of autologous T cells in the following Suppression
Assay: CD8
+
T cells isolated from 30mL of PBMC from returning healthy donors were
CFSE-labeled (3 μM, Sigma) and seeded in 96-well plates at 2x10
5
cells/well. CD33
+
cells isolated previously (ii. MDSC isolation, above) were added to the 96-well plates at a
ratio of 1:4 relative to the T cells. T cell stimulation was provided by anti-CD3/CD28
stimulation beads (Invitrogen, Carlsbad, CA). Suppression Assay wells were analyzed by
flow cytometry for T cell proliferation after three days and supernatants were analyzed
for IFNγ levels (below). Controls included a positive T cell proliferation control (T cells
alone) and an induction negative control of CD33
+
or CD11b
+
cells isolated from PBMC
cultured in medium only. Samples were run in duplicate and data were collected as
16
percent proliferation for 15,000 cells. Samples were run on a FACSCalibur flow
cytometer (BD Biosciences, San Jose, CA) and data acquisition and analysis were
performed using CellQuestPro software (BD) at the USC Flow Cytometry core facility.
v. IFNγ cytokine bead array
Interferon gamma (IFNγ) production by T cells in the MDSC Suppression Assay was
measured as a correlate of T cell activation. Supernatant was collected at the conclusion
of the Suppression Assay from each sample and stored at -80°C until analysis. Protein
levels of IFNγ were measured using BD Cytometric Bead Array human IFNγ flex set and
human soluble protein master buffer kit (BD), per manufacturer’s instructions. Samples
were run on a FACSCalibur flow cytometer (BD) and data acquisition and analysis were
performed using CellQuestPro and FCAP Array software (BD).
Real-time RT-PCR for gene expression of tumor cell lines
For gene expression studies, tumor cell lines were collected from culture flasks by
trypsinization and RNA was isolated from tumor cells by Qiagen’s RNeasy mini kit
(Valencia, CA) followed by DNase (TurboDNase, Applied Biosystems, Foster City, CA)
treatment. For real-time RT-PCR, 100ng of DNase-treated RNA was amplified with
gene specific primers using one-step Power SYBR green RNA-to-Ct kit (Applied
Biosystems) and run in an MX3000P Strategene thermocycler (La Jolla, CA). Data were
acquired and analyzed using MxPro software (Stratagene). Gene expression was
normalized to housekeeping gene GAPDH and fold change determined relative to
17
Stratagene QPCR Human Reference Total RNA (DNase treated RNA composed of total
RNA from 10 human cell lines of varied histologic origin). Primer sequences were
obtained from the NIH qRT-PCR database (http://primerdepot.nci.nih.gov) and were
synthesized by the USC Core Facility [46].
Measurement of tumor-derived factors by ELISA
Supernatants were collected from cell line cultures, passed through a 0.2µm syringe filter
unit to remove cell debris, and stored in aliquots at -20˚C. Levels of IL-1 β, IL-6, TNF α,
VEGF and GM-CSF in supernatant samples were measured using ELISA DuoSet kits
(R&D) per manufacturer's instructions. Plate absorbance was read on an ELX-800 plate
reader (Bio-Tek, Winooski, VT) and analyzed using KC Junior software (Bio-Tek).
Statistical analysis
Changes in mean T cell proliferation and mean IFNγ production in the presence or
absence of tumor-educated or cytokine-induced MDSC were tested for statistical
significance by one-way ANOVAs followed by Dunnett test for pairwise comparisons of
experimental samples to T cells alone. Mean gene expression of 15 tumor-derived
factors between HNSCC cell lines with and without CD33
+
MDSC induction capacity
was compared by ANOVA followed by Tukey’s test for pairwise comparisons. For those
factors with statistically significant different mean expression between suppressor cell
inducing and non-inducing cell line groups, a linear regression analysis was performed to
evaluate for a linear correlation between strength of suppressor cell induction and gene
18
expression levels. Changes in mean T cell proliferation stimulated in the presence of
suppressive CD33
+
induced by HNSCC cell lines for neutralization experiments were
evaluated by ANOVA followed by Tukey’s test for pairwise comparisons between all
groups. Statistical tests were performed using GraphPad Prism software (La Jolla, CA)
with a significance level of 0.05. Graphs and figures were produced using GraphPad
Prism, Microsoft Excel, and Adobe Illustrator and Photoshop software (San Jose, CA).
RESULTS
Induction of tumor cell line-associated human monocytic MDSC
A protocol for the generation of tumor cell line educated-CD33
+
human MDSC from
normal donor PBMC was developed, as outlined schematically in Figure 3. Briefly,
PBMC-tumor cell line co-cultures were established in tissue culture flasks for one week.
Tumor-educated CD33
+
cells were then isolated, checked for viability, and tested for
suppressive function by co-culture with fresh, autologous T cells in the presence of T cell
stimuli. Use of irradiated tumor cells in co-cultures yielded comparable suppressor cell
induction, suggesting that tumor cells need not be actively dividing to mediate the
observed induction of suppressive function (Table 1). Unfractionated PBMC
preparations were used in evaluating the ability of human solid tumor cell lines to
generate myeloid suppressor cells to best approximate an in vivo setting, but CD33
+
suppressor cells were also generated successfully from T cell-depleted PBMC by co-
culture with 4-998 osteogenic sarcoma or SCCL-MT1 head and neck squamous cell
carcinoma (HNSCC) cells (Table 1).
19
Figure 3. Schematic of Co-culture and MDSC Suppression Assays for the in vitro generation of
tumor-associated myeloid suppressor cells. Induction: Normal donor PBMC are co-cultured with human
solid tumor cell lines for one week. MDSC Isolation: CD33
+
cells are isolated from PBMC-tumor co-
cultures by anti-CD33 microbead labeling and magnetic column separation. Suppression Assay: Tumor-
educated CD33
+
cells are subsequently co-cultured with fresh, autologous CFSE-labeled T cells at a 1:4
ratio in the presence of T cell stimuli. After 3 days, T cell proliferation is measured as CFSE-dilution using
flow cytometry. Suppressive function is evaluated as the ability of CD33
+
cells to inhibit autologous CD8
+
T cell proliferation.
20
Strong CD33
+
MDSC induction capability by a subset of human tumor cell lines
One-hundred-one human solid tumor cell lines were tested for their ability to induce
MDSC using the tumor co-culture assay using 61 unique healthy, volunteer donors (39
male, 22 female) ranging in age from 23-62. CD33
+
MDSC could be generated by at
least one cell line of every human tumor type examined (cervical/endometrial, ovarian,
pancreatic, lung, head and neck, renal cell, liver, colorectal, prostate, thyroid, gastric,
bladder, sarcoma, and glioblastoma), with the exception of breast carcinoma (Table 1).
CD33
+
cells from tumor co-cultures were categorized as strong MDSC (ability to
suppress autologous T cell proliferation by greater than 50%), weak MDSC (50-20%,
meeting statistical significance), or non-suppressive myeloid cells (n  2 biologic
samples). Of 101 tumor cell lines examined, 45 consistently generated MDSC with 30 of
those generating strongly suppressive MDSC. A range of suppressor cell ability
appeared to exist within histologic types for the majority of tumor cell lines examined,
suggesting that subclones within a tumor may drive MDSC induction. Notably, myeloid
cells from PBMC cultured in medium alone or co-cultured with fibroblast cell lines were
not suppressive (Table 1).
21
Table 1. Forty-five of 100 human solid tumor cell lines induce functionally suppressive CD33
+
myeloid suppressor cells from volunteer normal human PBMC after one-week co-culture in vitro.
Tumor cell lines inducing CD33
+
with statistically significant suppressive function are indicated in light
pink, and those with strong MDSC inducing capacity (mean T cell suppression by CD33
+
cells ≥ 50%) are
indicated by **/dark pink. CD33
+
cells from PBMC cultured in complete medium alone (non-suppressive
control), co-cultured with fibroblast cell lines (induction negative control), and cytokine-induced MDSC
(GM-CSF + IL-6, suppressive control) were run in parallel for comparison.
22
Table 1, continued.
23
Suppressive CD33
+
myeloid cells inhibit T cell proliferation and IFNγ production
To characterize further CD33
+
suppressor cells generated by tumor cell line co-culture,
IFN  production was evaluated in addition to T cell proliferation in MDSC suppression
assays from co-culture with seven different human head and neck squamous cell
carcinoma (HNSCC) cell lines: SCCL-MT1, SCC-4, CAL-27, FaDu, RPMI 2650, and
SW 2224 (Figure 4A and 4B). The suppressive capability of HNSCC induced MDSC was
compared with that of a positive T cell proliferation control (T cells alone) and an
induction negative control of CD33
+
cells isolated from PBMC cultured in medium only,
and an induction positive control of CD33
+
cells isolated from PBMC cultured with GM-
CSF and IL-6. In comparison with T cells alone, CD33
+
cells induced by SCCL-MT1,
SCC-4, CAL-27, SW 451 and the induction positive control demonstrated statistically
significant strong inhibition of autologous T cell proliferation, while FaDu induced
weakly suppressive MDSC (p<0.05).
Consistent with strong suppression of T cell proliferation (Table 1, Figure 4A), CD33
+
cells induced by SCCL-MT1, SCC-4, and the induction positive control significantly
inhibited IFN  production (Figure 4B) (p<0.05). Although CD33
+
suppressor cells
induced by FaDu did not strongly inhibit T cell proliferation, these co-cultures did
demonstrate significantly decreased IFN  production (p<0.05). Conversely, Cal-27 and
SW 451 induced MDSC that strongly suppressed T cell proliferation (Table 1, Figure
4A), but only weakly inhibited IFN  (Figure 4B). CD33
+
cells induced by RPMI 2650
and SW 2224 did not demonstrate suppression of T cell proliferation or inhibition of
24
IFN . These findings suggest that MDSC may impede T cell responses through multiple
avenues, including inhibition of proliferation and IFN  production.
Figure 4. HNSCC-induced MDSC inhibit autologous T cell proliferation and IFNγ production. A
subset of HNSCC cell lines induces a CD33
+
population with suppressive function characteristic of MDSC,
including inhibition of autologous CD8
+
T cell proliferation (A) and IFNγ secretion (B). Tumor cell lines
are grouped by strength of MDSC induction: strong (black), weak (gray), and non-inducing (white). For
both A and B, mean shown (n 2 donors) +SEM. * indicates statistical significance by ANOVA followed by
Dunnett post-test for comparison to T cells alone, p<0.05.
Human solid tumor cell lines over-express multiple putative MDSC-promoting factors
In this study, cancer cell line induction of human CD33
+
MDSC was used as a model for
tumor induction of human MDSC. MDSC-inducing and non-inducing human tumor cell
lines were used to examine the mechanisms by which tumors generate MDSC. Gene
expression levels of putative MDSC inducing immune-modulatory factors were
compared for both groups of cancer cell lines using quantitative RT-PCR techniques
25
(Figure 5). Gene expression by tumor cell lines was examined for 15 immune
modulatory factors: ARG-1, IL-1β, IL-4, IL-6, IL-10, iNOS, c-kit L, COX2, FLT3L,
GM-CSF, IDO, M-CSF, TGF β, TNF α, and VEGF.
All tumor cell lines examined were found to have a statistically significant increase in
expression of at least one of these factors relative to a reference human RNA standard;
however no single factor was shown to be critical for MDSC induction (Figure 5).
MDSC-inducing tumor cell lines showed increased expression of COX2, IL-1β, IL-6,
TNF , M-CSF, IL-10, and IDO (Figure 5). FLT3L and c-Kit L expression appeared to
be increased in both groups of tumor cell lines, suggesting that these are not singular
distinguishing factors for MDSC induction. TGF β, GM-CSF, and ARG-1 showed
consistent down-regulation amongst MDSC-inducing tumor cell lines. These results
suggest multiple pathways of MDSC induction for the tested human cancer cell lines.
26
Figure 5. Expression of putative MDSC-inducing factors by human solid tumor cell lines and PBMC.
The fold change in gene expression of putative MDSC-inducing immune modulatory factors was
determined by quantitative RT-PCR techniques for MDSC-inducing and non-inducing human solid tumor
cell lines. In addition, the expression of these factors by freshly isolated PBMC from healthy volunteer
donors was analyzed in parallel. Due to the large variation in gene expression levels amongst tumor cell
lines, the log
10
fold change is shown (e.g. fold increases of 10 and 100 are represented as a log
10
changes of
2.0 and 3.0, respectively). Genes whose expression was higher in the tumor cell line than the reference
sample are shown in red (90th percentile); those whose expression was lower than the reference sample are
shown in green (10th percentile). Note, the differential expression of factors COX2, IL-1 β, and IL-6 in
MDSC-inducing and non-inducing cell lines and the high expression of GM-CSF in normal donor PBMC
compared to tumor cell lines.
27
Figure 5, continued.
28
A focus on HNSCC Cell Lines: MDSC-induction capacity correlates with tumor cell
line expression of IL-1β, IL-6, TNF α, VEGF, and GM-CSF
A subset of head and neck squamous cell carcinoma cell lines consisting of both MDSC-
inducing and non-inducing models was further studied for expression of these putative
inducing factors to reduce background differences in gene expression related to tissue-
specific expression patterns. HNSCC tumor cell lines showed a high frequency of CD33
+
MDSC induction (Table 1) and thus were good models for further studies of induction.
Expression of immune modulatory factors (c-kitL, COX2, FLT3L, GM-CSF, IL-1β, IL-4,
IL-6, IL-10, IDO, iNOS, M-CSF, TGF β, TNF α, VEGF) was measured in eight HNSCC
cell lines: CAL-27, FaDu, RPMI 2650, SCC-4, SCCL-MT1, SW 2224, SW 451, USC-
HN2, by using quantitative RT-PCR techniques. MDSC-induction capacity correlated
directly with tumor cell line expression of IL-1 β, IL-6, TNF α, VEGF, and GM-CSF
(Figure 6A) (p<0.05 for ANOVA followed by Dunnett-test for pairwise comparisons
between inducing and non-inducing cell lines for each factor, and p<0.05 for linear
regression analysis of suppressive induction capacity with level of cytokine production).
Differential gene expression of IL-6, TNF α, VEGF, and GM-CSF was confirmed at the
protein level by ELISA techniques (Figure 6B); IL-1β levels were below the sensitivity of
the assay. These studies demonstrated the presence of these five cytokines correlated
significantly with the ability of human tumor cell lines to induce CD33
+
MDSC.
29
Figure 6. MDSC-inducing cell lines produce increased IL-1 β, IL-6, TNF α, and VEGF. A, Expression
of ten putative MDSC-inducing factors was measured in MDSC-inducing and non-inducing HNSCC cell
lines by qRT-PCR. Increased MDSC-induction capacity was associated with greater expression of IL-1 β,
IL-6, TNF α, and VEGF (p<0.05). Mean fold change (n=2) relative to human reference RNA (red =
increased, green = decreased expression), p value shown is for linear regression analysis for factors having
significantly higher gene expression in MDSC-inducing compared with non-inducing human HNSCC cell
lines by one-way ANOVA followed by Tukey’s post-test. B, Protein secretion of these cytokines by
HNSCC cell lines was measured in supernatants using ELISA techniques to confirm gene expression
findings. Mean protein levels shown (two independent experiments each run in triplicate), +SEM.
B
A
30
Neutralizing antibodies to cytokines GM-CSF, IL-1 , IL-6, VEGF, or TNF  were tested
in PBMC-tumor cell line co-cultures to determine which factor(s) was most important for
induction (Figure 7). Neutralization of GM-CSF, IL-6, or IL-1 in tumor cell line-PBMC
co-cultures abrogated significant induction of CD33
+
suppressor cell function (p<0.05)
and restored T cell proliferation to levels comparable to controls (p=NS). COX2
expression was also elevated in many of the MDSC-inducing cell lines, particularly
ovarian and cervical cancer cell lines, and PGE
2
in combination with GM-CSF induced
weak suppressive function in CD33
+
cells. However, addition of COX2 inhibitors to
ovarian and cervical tumor cell line-PBMC co-cultures did not significantly decrease
MDSC induction (data not shown).
Figure 7. Removal of GM-CSF, IL-6, or IL-1 β from co-culture impairs CD33
+
MDSC induction by
tumor cell lines. T cell proliferation when co-cultured with CD33
+
MDSC from tumor cell line (SCCL-
MT1 or USC-HN2) co-cultures with neutralizing antibodies to GM-CSF, IL-6, IL-1 , TNF , or VEGF.
Mean shown (n=5, four independent experiments), +SEM. * indicates statistical significance, p<0.05.
31
Human CD33
+
MDSC are generated in vitro by soluble immune modulatory factors
Using MDSC-induction by cancer cell lines as a model, cytokine mixtures were designed
for the in vitro generation of human MDSC from healthy donor PBMC. Briefly, PBMC
were cultured for one week in the presence or absence of immune modulatory factors
GM-CSF and IL-1β, IL-6, VEGF, TGF β, TNF α, or PGE
2
(as a proxy for COX2 over-
expression [47]), and then CD33
+
cells were isolated and tested for suppressive function
by MDSC Suppression Assays. The ability of cytokine-induced CD33
+
to inhibit
autologous T cell proliferation was evaluated at cell ratios of 1:2 and 1:4 (Figure 8).
Results from these studies show the generation of potent CD33
+
MDSC after incubation
of non-suppressive PBMC with select cytokine combinations. Notably, GM-CSF + IL-6
generated MDSC with the ability to suppress autologous T cell proliferation by 80.6
percent (mean, n=4), at a 1:4 ratio (Figure 8). Furthermore, GM-CSF alone, GM-CSF +
IL-1 β, GM-CSF + TNF α, and GM-CSF + VEGF conditions yielded MDSC with
significant suppressive function (range 43-57% suppression at 1:4 ratio). CD33
+
cells
treated with GM-CSF + PGE
2
exhibited weak suppressive ability. While these data
suggest a positive role for IL-6, VEGF, IL-1β, TNF α, and GM-CSF in MDSC generation,
TGF β does not appear to have a major role in the promotion of suppressive function in
CD33
+
cells since TGF β treatment consistently decreased the potency of other cytokines
in the induction of suppressive function (Figure 8 and Chapter 4). GM-CSF was found to
be critical to maintain myeloid cell viability, as reported previously [35], and was
included in all cytokine mixtures, though at lower concentrations than used by Ko et al.
(20ng/mL versus 50ng/mL) [35]. However, even this low level of GM-CSF alone
32
appeared to induce some suppressive function in CD33
+
cells, as shown in Figure 8, and
warrants further study as an inducing cytokine for human MDSC. The observation that
GM-CSF alone generated CD33
+
cells with some suppressive function may be due to the
promotion of immature dendritic cells (DC) that in the absence of IL-4 do not complete
maturation to functional antigen presenting cells and poorly stimulate autologous T cells
[48]. These results, in conjunction with neutralization experiments, provide strong
evidence for IL-1β and IL-6 with GM-CSF as critical factors for the induction of human
CD33
+
MDSC.
GM-CSF expression by PBMC is up-regulated by co-culture with tumor cell lines
Cytokine induction studies showed that GM-CSF alone or in combination with IL-6, IL-
1β, and VEGF generates potent CD33
+
suppressor cells. While this cytokine appears to
be critical in human myeloid suppressor cell induction, many MDSC-inducing tumor cell
lines were found to have very low GM-CSF expression (Figure 5). PBMC, however, do
have strong expression of GM-CSF (Figure 5) and thus may provide a source of GM-CSF
for MDSC induction in the tumor micro-environment. GM-CSF expression by PBMC in
the presence or absence of inducing tumor cell line was measured by quantitative RT-
PCR (Figure 9A). These data showed strong up-regulation of GM-CSF production by
normal donor PBMC following direct co-culture with MDSC-inducing cell lines 4-998
(osteogenic sarcoma) or SCCL-MT1 (HNSCC), but not culture in medium alone.
Analysis of the CD33
+
cell fraction showed that the increase in GM-CSF production was
not localized to this population (Figure 9B).
33
Figure 8. Cytokine-induced CD33
+
MDSC demonstrated potent suppressive function. PBMC from
normal donors were cultured for one week in the presence of different cytokine mixtures. CD33
+
cells
were then isolated and tested for their ability to suppress the proliferation of autologous CD3/CD28-
stimulated T cells at ratios of 1:2 and 1:4. CD33
+
cells from cultures treated with GM-CSF and IL-6, Il-1 β,
VEGF, PGE
2
, or TNF α demonstrated suppressive function. Mean (n=3) is shown with SEM. Conditions
with statistically significant decreases in mean T cell proliferation compared to stimulated T cells alone are
indicated by an asterisk.
Figure 9. Up-regulation of GM-CSF by PBMC in tumor co-culture accompanies induction. (A and B)
To shed light on possible sources of GM-CSF for the induction of suppressor cells in tumor-PBMC co-
cultures, gene expression of GM-CSF in PBMC following direct co-culture with inducing tumor cell lines
was measured by qRT-PCR techniques and compared to cells cultured in medium alone (mean shown, +
SD). A, GM-CSF expression by PBMC was strongly up-regulated following tumor cell line co-culture. B,
34
Analysis of GM-CSF expression in the CD33
+
cell fraction suggests that the increase in GM-CSF observed
is not localized to this subpopulation.
Preferential induction of CD11b
+
MDSC by human breast cancer cell lines
Interestingly, no human breast cancer cell lines (0/9) tested generated suppressive CD33
+
cells. This finding led us to investigate the induction of other MDSC phenotypes by
these models. Human MDSC have been reported to express a wide range of surface
markers, as discussed in more detail in Chapter 3, and likely consist of several subtypes.
In addition to the common myeloid antigen CD33, CD11b is another marker reported to
be expressed on some human MDSC [14,16,49]. Breast carcinoma cell lines
preferentially induced CD11b
+
MDSC (Figure 10), suggesting that this component of the
MAC-1 phagocytic complex may be a more specific marker for the subset of MDSC
induced by these tumor models. Lung carcinoma and glioma cell lines, which had a low
frequency of CD33
+
MDSC induction, also were found to induce with moderate
frequency this CD11b
+
MDSC subset (Figure 10). Taken collectively with our survey of
CD33
+
MDSC induction, these data suggest that the induction of MDSC is a universal
feature of human cancers with some variation in the phenotype of induced MDSC subsets
observed. These data further emphasize the importance of functionally defining this
heterogeneous population of suppressor cells until specific activation-associated markers
are identified.
35
Figure 10. Induction of a second CD11b
+
MDSC subset by breast, lung, and brain cancer cell lines.
A, CD11b
+
cells from breast cancer, lung cancer, or glioma cell line-PBMC co-cultures were evaluated for
suppressive function against CD3/CD28 stimulated autologous T cells. Mean (n=2) T cell proliferation +
SEM or T cell proliferation (n=1) is shown from Suppression Assays of CD33
+
or CD11b
+
cells with
autologous T cells, respectively. * indicates statistically significant suppression of T cells by CD11b
+
cells
from co-culture (p<0.05, ANOVA followed by Dunnett’s test for comparison to T cells alone); significance
for suppression by CD33
+
cells is found in Table I. Note that some tumor cell lines induce both subsets,
while others induce only one subset or neither. CD33
+
and CD11b
+
cells from medium only cultures were
not suppressive.
DISCUSSION
The recruitment, expansion, and activation of immune suppressor cells by tumors
contribute significantly to disease progression, and as such pose significant barriers to
successful cancer immunotherapy. In this study, we describe a method for generating
functional human monocytic MDSC using tumor cell line co-culture methods, and
quantify their induction by an extensive panel of human tumor cell lines with diverse
histologic origins. Of the 101 human solid tumor cell lines we examined by tandem in
vitro co-culture and suppression assays, 45 of these induced human CD33
+
myeloid
36
suppressor cells; of those 45, a subset of 30 consistently induced strongly suppressive
MDSC (Table 1). This comprehensive survey of human solid tumor models
demonstrates the overarching importance of CD33
+
MDSC-mediated immune
suppression in human cancer, and identifies specific tumor models for future studies of
human monocytic MDSC. Of considerable clinical importance, we found that tumor
types with strong inductive potential of CD33
+
MDSC included HNSCC, cervical and
ovarian, and colorectal carcinomas. In addition, the preferential induction of CD11b
+
suppressor cells over CD33
+
MDSC that we observed by breast cancer cell lines clearly
warrants further investigation. The breast tumor cell lines evaluated here included
hormone responsive and non-responsive tumors as well as Her2/erb mutation positive
and negative tumors. This array decreased the likelihood that the observed lack of CD33
+
myeloid suppressor cells was due to these particular factors per se.
Furthermore, it is well known that solid tumors are internally heterogeneous, with
subclones demonstrating variable growth, invasion, and immunomodulatory properties
[50]. Cell lines are derived from individual subclones within a solid tumor and
collectively may recapitulate the composite tumor. Our use of multiple human tumor cell
lines facilitated identification of canonical MDSC-inducing factors by directly comparing
strongly vs. weakly inductive clones that shared a common background of typical
HNSCC features. This finding suggests that only a subpopulation of cells within a solid
tumor may drive MDSC induction, and conversely should not be interpreted as an
inability to induce MDSC by these types of cancers. Indeed, data from experimental
37
tumor models and cancer patients suggest that MDSC accumulation correlates highly
with tumor burden and spread [21,35]. In fact, these data suggest the nearly universal
induction of MDSC by human cancers, albeit with some subset variance that emphasizes
the importance of further studies to characterize these cells.
Using the model system described here we were able to probe MDSC induction pathways
to identify therapeutic targets for MDSC inhibition in cancer patients. Analysis of tumor-
derived factors from MDSC-inducing and non-inducing HNSCC cell lines identified IL-6
and IL-1 as well as GM-CSF as critical tumor-derived factors (TDF) for human CD33
+
MDSC induction, in agreement with previous reports [51,52] and our own work [28]. IL-
6 and IL-1 are the predominant cytokines found to be elevated in the sera of patients
with cancer cachexia and sepsis, two common scenarios in which MDSC accumulation
has been reported [26,44]. Particularly in trauma and sepsis, systemic inflammation and
immune activation may lead to significant immune-mediated damage of normal host
tissues and may increase the danger of autoimmune reactions [44]. In this setting,
regulatory cells may be induced as a physiologic countermeasure to control widespread
autoimmunity and non-specific, collateral immune attack on self tissues. Remarkably, the
cytokine profile created by tumor-derived factors appears to mimic this situation of
systemic immune activation in its high levels of IL-6 and IL-1. In this setting, however,
the immune tolerance has additional pathologic consequences that promote tumor
immune tolerance through the generation of MDSC. The absence of GM-CSF expression
by inducing tumor cell lines appeared at first paradoxical, in light of the demonstrated
38
ability of GM-CSF alone or in combination with other cytokines to induce CD33
+
suppressor cells. GM-CSF, in conjunction with IL-4, has long been recognized to
promote DC development in vitro and as reported previously by others [35], we noted
that low levels of GM-CSF greatly enhanced myeloid cell viability in culture and
expanded CD33
+
cells. GM-CSF may be expressed by stromal cells or PBMCs that
traffic to the tumor site in response to tumor cell contact. This hypothesis is supported by
the marked up-regulation of GM-CSF expression by PBMCs that we observed following
direct co-culture with either suppressive or non-suppressive tumor cell lines (Figure 9 and
data not shown). While the focus of this study was on the induction of the canonical
CD33
+
MDSC subset, it may also shed light on the induction of CD11b
+
MDSC. This
subset was preferentially induced over CD33
+
MDSC by breast carcinoma cell lines, and
our gene expression studies of these tumor models showed up-regulation of FLT3L and
TGF  (Figure 5). Additional studies will be needed to determine the induction pathways
for CD11b
+
MDSC and other MDSC subsets (e.g. CD66b
+
).
In targeting MDSC induction in the cancer setting, a therapeutic approach must
necessarily reverse a highly evolved natural mechanism of immune tempering and
protection of self. This suggests that attenuating or blocking MDSC induction must
proceed with vigilance against generating widespread autoimmune disease by targeting
its inhibition in the tumor environment. Indeed, reversal of MDSC-mediated immune
tolerance may be achieved with less risk by inhibition of suppressive mechanisms at the
cellular level.
39
CHAPTER 3.
FUNCTIONAL CHARACTERIZATION OF HUMAN CD33
+
AND CD11b
+
MYELOID-DERIVED SUPPRESSOR CELLS INDUCED
BY HUMAN CANCER CELL LINES
Myeloid-derived suppressor cells (MDSC) have been described in patients with diverse
types of cancer [20,29-40], and our survey of over 100 human solid tumor cell lines
suggests that the capacity to induce MDSC is a nearly universal feature of human cancers
(Chapter 2). As such, MDSC pose a primary barrier to restoring anti-tumor immune
responses in cancer patients. Unfortunately, studies to understand human MDSC lag
greatly behind work done with murine MDSC [6,13,14]. In mice, rapid advances have
been aided by a clearly defined phenotype associated with active suppressor functions
(CD11b+ Gr-1
int/low+
) and the limitless supply of MDSC for study from tumor-bearing
experimental animals [13,14]. Human MDSC comprise a phenotypically and perhaps
functionally heterogeneous population of immature myeloid (CD33
+
) cells derived from
dendritic cell (DC), macrophage, and granulocyte progenitors that lack lineage
maturation markers and inhibit effector T cell functions [16]. Apart from such functional
definitions, there is no specific marker or phenotype yet identified that reproducibly
distinguishes human MDSC while encompassing their diversity. The canonical
phenotype for human MDSC is lineage negative (CD14
-
CD3
-
CD19
-
CD56
-
), myeloid
(CD33
+
), and with low expression of the antigen presentation marker HLA-DR [13,16].
Diaz-Montero et al. [39] showed a direct correlation between the accumulation of
monocytic CD33
+
HLA-DR
low
MDSC in cancer patients and disease burden and spread,
suggesting that this population in particular may be a key player in tumor immune
40
tolerance. However, human MDSC have also been reported as CD15
+
Arg1
+
granulocytes
by Zea et al. [36] in renal cell carcinoma, or as CD14
+
HLA-DR
-
cells in melanoma
[29,30], ovarian carcinoma [38], and hepatocellular carcinoma [33]. In addition, work
by Ochoa and colleagues identified granulocytic CD66b
+
MDSC in patients with renal
cell carcinoma [53]. Given the diverse tumor mileu that can lead to suppressor cell
induction (Chapter 2) and the importance of regulatory mechanisms in controlling
autoimmune responses, it is likely that several subsets of human MDSC exist and may
show functional redundancy. In mice, monocytic and granulocytic subsets of MDSC
with subtle phenotypic and functional differences arise in the tumor microenvironment
compared with secondary lymphoid tissues. Our poor understanding of human MDSC
phenotype and function has constrained development of effective therapies for their
inhibition in cancer immunotherapy. A major difficulty in studying human MDSC is their
evident accumulation in individuals with significant illness (cancer, trauma, sepsis [17])
but relative absence in healthy individuals. This limits the availability of MDSC for study
and makes identifying the factors governing a single population difficult given cancer’s
complicated and highly dysfunctional clinical picture. The development of a novel in
vitro system to generate several subsets of human MDSC (Chapter 2) facilitated the
isolation of highly enriched and homogeneous MDSC subsets for characterization and
allowed manipulation of induction conditions to identify the key factors for induction of
two different MDSC subsets selected by CD33
+
and CD11b
+
. In the present study we
describe the morphology, phenotype, and function of these two MDSC subpopulations.
From these data we have identified common features among human MDSC, but also
41
transcriptional differences that influence the susceptibility of each subset to therapeutic
inhibition.
1
MATERIALS AND METHODS
Cell Lines and Cell Culture
Tumor cell lines were obtained from the American Type Culture Collection (ATCC) or
gifted to the Epstein laboratory. Notable gifts include the SW cell lines from the Scott
and White Clinic (Temple, TX) and pancreatic cell lines from Dr. Liz Jaffe (Johns
Hopkins Medical Center, Baltimore, MD). Tumor cell line authenticity was performed
by cytogenetics and surface marker analysis performed at ATCC or in our laboratory. All
cell lines were maintained at 37
o
C in complete medium [(RPMI-1640 with 10% fetal calf
serum (characterized FCS, Hyclone, Inc., Logan, UT), 2mM L-Glutamine, 100U/mL
Penicillin, and 100µg/mL Streptomycin)] in tissue culture flasks in humidified, 5% CO
2
incubators and passaged 2-3 times per week by light trypsinization.
Tumor-Associated MDSC Generation Protocol
i. Induction
Human peripheral blood mononuclear cells (PBMC) were isolated from healthy volunteer
donors by venipuncture (60mL total volume) followed by differential density gradient
centrifugation (Ficoll Hypaque, Sigma, St. Louis, MO). PBMC were cultured in complete
1
Portions of this chapter are undergoing revision for publication (Lechner et al. Functional characterization of
human myeloid-derived suppressor cells induced from peripheral blood mononuclear cells cultured with a
diverse set of human tumor cell lines. Journal of Immunology, under review)
42
medium (6x10
5
cells/mL, supplemented with rhGM-CSF (10ng/mL, R&D Systems,
Minneapolis, MN) to support viability) in T-25 culture flasks with human tumor cell lines
for one week. Tumor cells were seeded to achieve confluence by day 7 (approximately
1:100 ratio with PBMC), and samples in which tumor cells overgrew were excluded from
analysis and were repeated with adjusted ratios. PBMC cultured in medium alone were
run in parallel as an induction negative control for each donor. For neutralization
experiments, PBMC-tumor cell line co-cultures were repeated in the presence or absence
of for a subset of HNSCC cell lines. CD33
+
or CD11b
+
cells were subsequently isolated
and tested for suppressive function as described below. For these studies 39 male and 22
female healthy, volunteer donors ages 23 to 62 were used under USC Institutional
Review Board-approved protocol HS-06-00579. Data were derived from at least two
individuals and no inter-donor differences in MDSC induction or function were observed.
ii. MDSC Isolation
After one week, all cells were collected from tumor-PBMC co-cultures. Adherent cells
were removed using the non-protease cell detachment solution Detachin (GenLantis, San
Diego, CA). Myeloid cells were then isolated from the co-cultures using anti-CD33 or
anti-CD11b magnetic microbeads and LS column separation (Miltenyi Biotec, Germany)
as per manufacturer’s instructions. Purity of isolated cell populations was found to be
greater than 90% by flow cytometry and morphological examination and viability of
isolated cells was confirmed using trypan blue dye exclusion.
43
iii. Suppression Assay
The suppressive function of tumor-educated CD33
+
or CD11b
+
cells was measured by
their ability to inhibit the proliferation of autologous T cells in the following Suppression
Assay: CD3
+
T cells isolated from 30mL of PBMC from returning healthy donors were
CFSE-labeled (3 μM, Sigma) and seeded in 96-well plates at 2x10
5
cells/well. CD33
+
or
CD11b
+
cells isolated previously (ii. MDSC isolation, above) were added to the 96-well
plates at a ratio of 1:4 relative to the T cells. T cell stimulation was provided by anti-
CD3/CD28 stimulation beads (Invitrogen, Carlsbad, CA).
Suppression Assay wells were analyzed by flow cytometry for T cell proliferation after
three days and supernatants were analyzed for IFNγ levels by ELISA (R&D Systems).
Controls included a positive T cell proliferation control (T cells alone) and an induction
negative control of CD33
+
or CD11b
+
cells isolated from PBMC cultured in medium
only. As previously reported by our laboratory, MDSC induced by granulocyte
macrophage colony stimulating factor (GM-CSF) and interleukin (IL)-6 served as a
positive suppressor cell control [28]. Where indicated specific inhibitors of MDSC were
added to suppression assays: all-trans retinoic acid (ATRA, 100nM, Sigma, St. Louis,
MO), sunitinib (0.1 g/mL, ChemieTek, Indiannapolis, IN), celecoxib (15 M, Pfizer,
New York, NY), nor-NOHA (500 M, CalBiochem, San Diego, Ca), L-NMMA (500 M,
Calbiochem), apocynin (0.1mM, Sigma), 1D11 antibody (10 g/mL), SB431542 (5 M,
Tocris, Ellisville, MO), or Avastin (10 g/mL, Genetech, San Francisco, CA). Samples
were run in duplicate and data were collected as percent proliferation for 15,000 cells.
44
Samples were run on a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA)
and data acquisition and analysis were performed using CellQuestPro software (BD) at
the USC Flow Cytometry core facility.
Characterization of myeloid suppressor cells
i. Morphology of MDSC
Wright-Giemsa staining (Protocol Hema 3, Fisher, Kalamazoo, MI) of CD33
+
or CD11b
+
cell cytospin preparations was performed to assess the morphology of tumor-educated
myeloid cells. Freshly isolated PBMC and CD33
+
cultured in medium only or induced
by cytokines GM-CSF + IL-6 were prepared in parallel for comparison. Observation,
evaluation, and image acquisition were performed using a Leica DM2500 microscope
(Leica Microsystems, www.leica-microsystems.com) connected to an automated, digital
SPOT RTke camera and SPOT Advanced Software (SPOT Diagnostic Instrument Inc.,
www.diaginc.com). Images were resized, brightened, and adjusted for color for
publication using Adobe Photoshop software (Adobe, www.adobe.com).
ii. Flow cytometry analyses of cell phenotypes
The phenotype of in vitro-generated MDSC was examined for expression of myeloid,
antigen-presenting, and suppressor cell markers and compared to non-suppressive tumor-
educated CD33
+
or CD11b
+
cells. For staining, cells were collected from flasks using
Detachin to minimize cell surface protein digestion, and washed twice with FACS buffer
(2% FCS in PBS) before resuspending 10
6
cells in 100µl FACS buffer. Cells were
45
stained for 1hr on ice with cocktails of fluorescently-conjugated monoclonal antibodies
or isotype-matched controls, washed twice with FACS buffer, and resuspended in FACS
buffer for analysis. For intracellular staining, cells were fixed and permeabilized using
Fixation/Permeabilization Kit (eBioscience, San Diego, CA) after surface staining.
Antibodies used were purchased either from BD Biosciences: CD11c (B-ly6), CD33
(HIM3-4), HLA-DR (L243), CD11b (ICRF44), CD66b (G10F5), CD14 (M5E2), CD68
(Y1/82A), (41BBL (C65-485), OX40L (Ik-1); or eBioscience: CD30 (Ber-H2), CD103
(B-Ly7), GITRL (eBioAITR-L), CD56 (MEM-188). Samples were run on a BD
FACSCalibur flow cytometer and data acquisition and analysis were performed as above.
Data are from three unique donors and expressed as a fraction of labeled cells within a
live-cell gate set for 15,000 events. CD33
+
or CD11b
+
cells from PBMC cultured in
medium alone were run in parallel for comparison.
iii. Real-time RT-PCR for gene expression of myeloid suppressor cells and tumor
cell lines
For gene expression studies, tumor-educated CD33
+
or CD11b
+
cells were isolated from
tumor-PBMC co-cultures by fluorescence activated cell sorting after Induction (i.
Induction, above) and RNA was isolated from MDSC and DNase-treated using Qiagen's
RNeasy micro kit. For real-time RT-PCR, 100ng of DNase-treated RNA was amplified
with gene specific primers using one-step Power SYBR green RNA-to-Ct kit (Applied
Biosystems) and run in an MX3000P Strategene thermocycler (La Jolla, CA). Data were
acquired and analyzed using MxPro software (Stratagene). Gene expression was
46
normalized to housekeeping gene GAPDH and fold change determined relative to
expression levels in medium only-cultured cells. Primer sequences were obtained from
the NIH qRT-PCR database (http://primerdepot.nci.nih.gov) and were synthesized by the
USC Core Facility [46].
Immunohistochemistry
Immunohistochemistry studies were performed by Lillian Young in the Department of
Pathology, USC (Los Angeles, CA) on cytospin preparations of suppressive and non-
suppressive myeloid cells using antibodies against p-STAT3 (clone 6D779, dilution
1:400) and C/EBP  (clone H-7, dilution 1:100) (Santa Cruz Biotech).
Statistical analysis
Changes in mean T cell proliferation and mean IFNγ production in the presence or
absence of tumor-educated MDSC were tested for statistical significance by one-way
ANOVAs followed by Dunnett test for pairwise comparisons of experimental samples to
T cells alone. Changes in mean T cell proliferation in suppression assays in the presence
or absence of single inhibitors of suppressive mechanisms were evaluated by ANOVA
followed by Tukey’s test for pairwise comparisons between all groups. Differences in
mean expression of phenotypic markers between pooled groups of suppressive and non-
suppressive CD33
+
or CD11b
+
cells were tested for significance by Student’s t test for
independent samples. Differences in mean transcription factor or suppressive gene
expression between CD11b
+
and CD33
+
MDSC were tested for significance by Student’s
47
t test. Statistical tests were performed using GraphPad Prism software (La Jolla, CA)
with a significance level of 0.05. Graphs and figures were produced using GraphPad
Prism, Microsoft Excel, and Adobe Illustrator and Photoshop software (San Jose, CA).
RESULTS
Morphology of human CD33
+
and CD11b
+
suppressor cells induced by tumor cell lines
Previously, two subsets of human myeloid cells (CD11b
+
or CD33
+
cells) induced from
normal donor PBMC by tumor-cell line co-culture were found to have suppressive
capacity. To characterize better these suppressor cell populations, morphology,
phenotype, and gene expression studies were performed. The morphology of suppressive
tumor-co-cultured CD33
+
and CD11b
+
populations was compared to that of freshly
isolated PBMC and myeloid cells cultured in medium only by Wright-Giemsa staining
(Figure 11 and data not shown). Healthy donor PBMC showed occasional mononuclear
cells with pale and scant cytoplasm, scattered amongst predominant lymphocytes (data
not shown). CD33
+
cells from PBMC cultured in medium alone (with rhGM-CSF for
growth support) for one week were predominantly large, mononuclear cells having
abundant basophilic cytoplasm with occasional granulocytes and other myeloid lineage
cells (e.g. eosinophils) (left panel). In contrast to the mature lineages seen in medium
only myeloid cells, CD33
+
and CD11b
+
suppressor cells isolated from PBMC after tumor
co-culture showed an abundance of immature cells, including metamyelocytes or band
cells and blast-like cells (middle and right panels). Significantly, no obvious morphologic
48
differences were observed between CD33
+
and CD11b
+
MDSC suggesting similarity
between these subsets.
Figure 11. Morphology of human MDSC. Morphology of human CD33
+
and CD11b
+
MDSC subsets
isolated after tumor cell line co-culture and normal myeloid counterparts from medium only cultures
(Wright-Giemsa staining, x400, original magnification). CD33
+
MDSC appear slightly more differentiated
than CD11b
+
MDSC after induction. Images are representative of data from more than five donors and
three independent experiments using SCCL-MT1 or USC-HN2 for CD33
+
MDSC induction and MCF7 or
60 for CD11b
+
MDSC induction.
Phenotype of MDSC shows CD33
+
and CD11b
+
subsets, both HLA-DR
low
and Lineage
-
Further characterization of CD33
+
and CD11b
+
MDSC subsets examined their expression
of a wide range of proposed MDSC and mature innate immune cell markers (CD33,
CD11b, CD66b, CD14, CD11c, HLA-DR, GITRL, OX40L, 41BBL (CD137L), CD56).
Human MDSC were isolated by magnetic bead column separation after one-week co-
culture with SCCL-MT1 or USC-HN2 HNSCC cell lines (CD33
+
) or MCF-7 breast
cancer cell line (CD11b
+
) and non-suppressive CD33
+
or CD11b
+
control cells were
isolated from medium only PBMC cultures. The purity for column isolated populations
was found to be >90% by flow cytometry. Positivity for MDSC and mature myeloid
lineage markers was measured by flow cytometry for each population and compared
between CD33
+
and CD11b
+
MDSC subsets and between suppressive and non-
49
suppressive populations(Figure 12). Interestingly, CD11b expression levels were
inversely correlated with suppressive function in CD33
+
cells in these studies, and
similarly CD33 positivity was inversely correlated with suppressive function in CD11b
+
cells, suggesting a divergence in the two populations during induction (Figure 12A).
Notably, both CD33
+
and CD11b
+
suppressive populations showed decreased expression
of activation marker HLA-DR and mature DC marker CD11c compared with non-
suppressive populations of CD11b
+
and CD33
+
cells. These data are consistent with an
accumulation of immature myeloid lineage cells coincident with the induction of
suppressive function in either CD11b
+
or CD33
+
cells. Differentiated DC markers and T
cell co-stimulatory ligands were further examined on the CD33
+
subset of MDSC and
found to be expressed at similarly low levels between suppressive and non-suppressive
CD33
+
cells isolated from tumor co-cultures (p=NS) (Figure 12B), suggesting that the
maturation and antigen presenting defects of MDSC are not primary in T cell
suppression. This is consistent with therapeutic studies we have performed in our
laboratory in which the addition of T cell co-stimulatory ligands (Fc-GITRL, Fc-
CD137L, Fc-B7.1) or agonist antibodies (anti-CD137, anti-GITR, anti-CD28) to
suppression assays failed to significantly reverse inhibition of T cell proliferation (p=NS)
(data not shown). Two surface markers, CD30 and CD103, found on other immune
suppressor cell populations (32,33) were examined in this study as potential unique
markers of active MDSC but were not found to correlate with their suppressive function
(p=NS) (Figure 12). Macrophage marker CD68 and granulocyte marker CD66b
expression were low or absent and not differentially expressed by suppressive and non-
50
Figure 12. Human CD33
+
and CD11b
+
MDSC are distinct subsets with a common HLA-DR
low
Lineage
-
phenotype. A, Phenotype of HNSCC cell line-induced CD33
+
and breast cancer cell line-induced
CD11b
+
MDSC compared with medium only, non-suppressive CD33
+
and CD11b
+
cells as measured by
flow cytometry. Mean percent positive cells (n ≥ 2) + SD shown, data from three unique donors.
Differences in percent positive cells analyzed by ANOVA then Bonferroni’s multiple comparison test for
selected pairs (* indicates statistically significant difference in mean percent positive between MDSC and
medium control for each subset, p<0.05). B, Expression of antigen presenting cell (left) and suppressor cell
(right) markers on strongly suppressive (induced by HNSCC cell lines SCCL-MT1, SCC-4, CAL-27)
versus non-suppressive (induced by SW 2224, RPMI 2650, or medium only) CD33
+
myeloid cells as
measured by flow cytometry. Mean fluorescence above isotype control (data from 3 unique donors; mean
shown for all three induction conditions (n = 9) +SEM). * indicates statistical significance, p<0.05, †
indicates p=0.59 for comparisons between suppressive cell and non-suppressive cells mean. NS = not
significant.
A
B
51
suppressive CD33
+
or CD11b
+
cells in this study, emphasizing that that these phenotypes
likely do not encompass tumor-associated macrophages [4] and the granulocytic MDSC
subsets described elsewhere [53].
CD33
+
MDSC up-regulate iNOS, TGF β, NOX2, VEGF, and ARG-1
Further characterization of tumor-induced CD33
+
suppressor cells by analysis of the
expression of putative MDSC suppression genes was performed. For these studies,
expression of putative suppression genes was measured in tumor cell line-induced
MDSC, including those induced by SCCL-MT1 HNSCC cell line, 4-998 osteogenic
sarcoma, DU 145 prostate carcinoma, CAKI-1 renal cell carcinoma, SK-OV-3 ovarian
carcinoma, and SW 608 and SW 732 colorectal adenocarcinoma cell lines (Figure 13A),
and compared to expression levels in normal, non-suppressive CD33
+
myeloid cells from
medium only cultures. These MDSC consistently showed statistically significant up-
regulation of arginase (ARG-1), inducible nitric oxide synthase (iNOS), NADPH oxidase
(NOX2), vascular endothelial growth factor (VEGF), and/or transforming growth factor
(TGF) β compared with control CD33
+
cells from medium-only cultures (Figure 13A).
These results are consistent with the gene expression data reported for MDSC isolated
from cancer patients and murine tumor models [13,16]. Subtle variations were observed
in the gene expression patterns of these tumor-induced MDSC, which is consistent with
the hypothesis that different MDSC subsets are generated by different tumors dependent
upon the specific profile of immune factors produced by each tumor. To determine the
dominant mechanism of T cell suppression by this canonical CD33
+
MDSC subset,
52
suppression assays were repeated in the presence or absence of specific inhibitors of
ARG-1 (nor-NOHA), iNOS (L-NMMA), NOX2 (apocynin), VEGF (neutralizing
antibody Avastin), or TGF β1 (SB431542 or neutralizing antibody 1D11). In these
studies no one inhibitor was found to completely reverse suppression (Figure 13B),
consistent with the pleotropic actions of monocytic MDSC, but inhibitors of ARG-1 and
NOX2 did produce statistically significant decreases in suppression by CD33
+
MDSC.
These results were confirmed by siRNA knockdown of individual suppression genes:
ARG-1, iNOS, NCF1 (NOX2 component), TGF β1, or VEGFA (data not shown). To
further explore the differences in CD33
+
and CD11b
+
subsets of human MDSC, as
preferential induction was observed for breast carcinoma cell lines (Chapter 2), gene
expression of suppressive mechanisms was measured by qRT-PCR techniques. A
comparison of ARG-1, iNOS, and NOX2-component NCF1 gene expression in CD33
+
and CD11b
+
human MDSC induced by HNSCC or breast carcinoma cell lines revealed
similar levels of expression between these subsets with a trend toward increased ARG-1
and NOX2 expression in CD33
+
MDSC (Figure 13C). While these findings need
confirmation at the protein level (studies in progress), they suggest partial or complete
functional overlap of these MDSC subsets. Furthermore, these data suggest that effective
abrogation of human MDSC activities by any single inhibitor or subset depletion is
unlikely to yield significant therapeutic benefit in cancer patients.
53
Figure 13. Human MDSC mediate suppression through up-regulation of ARG-1, NOX2, iNOS,
VEGF, and TGF β. A, Expression of putative suppressive genes ARG-1, iNOS, NOX2-component NCF1,
VEGF, and TGF  in a subset of tumor cell line-induced CD33
+
MDSC. Mean fold change (n 2 donors per
tumor cell line) +SEM, relative to CD33
+
cells cultured in medium alone, are shown. * indicates statistical
significance, p<0.05, by ANOVA followed by Dunnett test for pairwise comparisons to medium only
CD33
+
controls. B, Tumor cell line-induced CD33
+
MDSC inhibit proliferation of autologous, CD3/CD28-
stimulated T cells. Specific inhibitors of MDSC suppressive mechanisms ARG-1 and NOX2 mediate
partial but incomplete reversal of suppression. * indicates statistical significant difference in mean T cell
proliferation (mean shown + SEM, n 7 for each inhibitor, data from 2 independent experiments with
similar results), p<0.05, by ANOVA followed by Tukey test for pairwise comparisons. C, Comparison of
ARG-1, iNOS, and NOX2-component NCF1 gene expression in CD33
+
and CD11b
+
human MDSC
revealed similar levels of expression between these subsets. Mean fold change shown relative to medium
only controls (n = 3 unique donors for MDSC from co-cultures with each of three inducing tumor models)
+ SEM. No statistically significant difference between means as determined by Student’s t test for each
gene.
54
C
Figure 13, continued.
Higher Hif1 α, STAT3, and C/EBP  gene expression distinguish tumor cell line-
induced human MDSC from normal myeloid cells and delineate subsets
It is apparent that human-myeloid derived suppressor cells can be induced by multiple
factors present in the tumor setting (Chapter 2, [28]). Furthermore, as a consequence of
these multiple different induction routes at least two distinct phenotypes of human MDSC
emerge that can both mediate suppression of T cell responses (Figure 12). Interestingly,
these CD33
+
and CD11b
+
MDSC subsets showed some phenotypic (HLA-DR
low
and
lineage
-
) and functional convergence (Figure 13C) despite preferential induction by
different tumor models and predominant expression of either CD33 or CD11b. We
wondered whether a common transcription factor was activated by these multiple
pathways and might be act as a "master switch" to control both of these human MDSC.
Several transcription factors have been proposed for control of MDSC, primarily in mice,
including CCAAT-enhancer-binding proteins (C/EBP) β [51], hypoxia inducible factor
(HIF) 1 α [25], and Signal transducer and activator of transcription (STAT) 3 [35,54],
STAT5 [54], and STAT6 [13]. Previously identified as transcriptional regulators in some
55
murine tumor-derived MDSC subsets, we show that these transcription factors are
elevated in human MDSC and, importantly, are differentially expressed in CD33
+
versus
CD11b
+
MDSC subsets. We examined the expression of HIF1 α, STAT3, and C/EBP  in
tumor cell line (SCCL-MT1 or USC-HN2)-induced CD33
+
or (MCF7 breast carcinoma)
CD11b
+
human suppressor cells compared with medium only controls by qRT-PCR
techniques (data from six unique donors, two independent experiments) (Figure 14A) and
immunohistochemistry (Figure 14B). Both CD33
+
and CD11b
+
functionally active
human MDSC showed significant up-regulation of transcription factors STAT3, C/EBP ,
and HIF1  compared with non-suppressive myeloid cells from medium only cultures.
However, CD33
+
and CD11b
+
MDSC subsets showed differences in transcriptional
changes for these factors that were suggestive of different induction or activation
pathways. As shown previously, CD33
+
or CD11b
+
MDSC may be induced under a
variety of different tumor conditions and following incubation with several distinct
cytokine mixtures (Chapter 2, [28]). CD33
+
MDSC showed stronger up-regulation of
STAT3 and HIF1  while CD11b
+
MDSC showed comparably greater up-regulation of
C/EBP  (Figure 14A). Differences in pSTAT3 and C/EBP  were confirmed by
immunohistochemistry studies, but the rapid degradation of HIF1  inhibited its detection
by these methods. Current studies are ongoing to confirm transcriptional differences in
these MDSC by immnoblotting techniques. Treatment of either CD33
+
or CD11b
+
tumor-cell line-induced MDSC with lipopolysaccharide (LPS), a known activator of
MDSC function [52], caused further up-regulation of STAT3, C/EBP , and HIF1 
concurrent with increased expression of ARG-1, iNOS, and NOX2-component NCF1
56
(Figure 14C). These results further support a role for these transcription factors in
promoting human MDSC suppressive function. While suppressive abilities in both
CD11b
+
and CD33
+
subsets correlated with increased expression of STAT3, C/EBP ,
and HIF1 , the dominant transcriptional pathway may be different. Indeed, therapeutic
reversal of CD11b
+
or CD33
+
MDSC-mediated suppression corresponded with different
transcriptional changes.
Tyrosine kinase inhibitor Sunitinib and all-trans retinoic acid (ATRA) have previously
been shown to inhibit MDSC [35,49]. Studies in our lab have also identified Celecoxib
and analogs dimethyl celecoxib (DMC) and unmethylated celecoxib (UMC) as inhibitors
of suppressive function in CD33
+
, but not CD11b
+
, MDSC in vitro (Appendix A). Of
note, the reversal of MDSC effects by CXB and analogs DMX and UMC does not appear
to rely upon cyclo-oxygenase (COX)2 enzyme inactivation, as demonstrated by the
persistence of therapeutic effects in the presence of prostaglandin E
2
rescue, efficacy of
analog DMC with low to absent COX inhibitory action, and the absence of effect seen
with the structurally-unrelated COX2-selective inhibitor naproxen. Gene expression
patterns in ATRA, Sunitinb, or CXB-treated CD33
+
or CD11b
+
human MDSC were used
to understand better factors promoting suppressive function in these cells (Figure 14D).
Functional inhibition of human CD33
+
MDSC by ATRA, Sunitinib, and Celecoxib
correlated with decreased STAT3 and HIF1  transcription. In comparison, functional
inhibition of human CD11b
+
MDSC by ATRA and Sunitinib correlated with decreased
C/EBP  levels, but no change in STAT3 and HIF1  mRNA levels. Celecoxib was not
57
found to have inhibitory actions on CD11b
+
MDSC and it was not observed to decrease
C/EBP  levels in this population. While preliminary, these data suggest that HIF1 α,
STAT3, and C/EBP  may be key transcription factors related to suppressive function in
tumor cell line-induced human MDSC, as was recently demonstrated for murine MDSC,
and warrant further studies at the protein level.
Figure 14. Transcription factors promoting human MDSC suppressive function. A, HIF1 α and STAT3
expression in tumor cell line-induced CD33
+
or CD11b
+
MDSC compared with medium only controls as
measured by qRT-PCR. Mean shown (data from six unique donors, two independent experiments) +SEM;
* indicates statistical significance, p<0.05,  indicates p=0.06. B, Immunohistochemisty of C/EBP β, p-
STAT3, and HIF1 α in CD33
+
(left) and CD11b
+
(right) MDSC and CD33
+
medium controls (middle).
Representative images shown from multiple samples stained (400x, original magnification) with arrows
showing positive staining areas for p-STAT3 and HIF1 α. C, Activation of human MDSC subsets by toll-
like receptor agonist lipopolysaccharide produces up-regulation of HIF1 α, C/EBP β, and STAT3,
concurrent with increased expression of suppressive mediators ARG-1, iNOS, and NOX2. Mean shown
(data from three unique donors) + SEM. D, Transcriptional changes in MDSC subsets associated with
inactivation of suppressive function. Reversal of CD33
+
MDSC suppressive function by ATRA, sunitinib,
and CXB correlated with decreased STAT3 and HIF1  expression. Functional inhibition of human CD11b
+
MDSC by ATRA and Sunitinib correlated with decreased C/EBP  levels, but no change in STAT3 and
HIF1  mRNA levels. CXB was not found to have inhibitory actions on CD11b
+
MDSC and it was not
observed to decrease C/EBP  levels in this population. Mean shown (data from three unique donors) +
SEM, * indicates statistically significant decrease (p<0.05) in transcript level in drug-treated MDSC
compared with untreated MDSC (ANOVA with Dunnett post-test).
A
58
Figure 14, continued.
59
Figure 14, continued.
DISCUSSION
Human MDSC comprise a diverse and complex group of suppressive cells that have been
poorly characterized to date.. In the present study, we used human MDSC induced in
vitro by co-culture with head and neck cancer and breast cancer cell lines to characterize
differences in the morphology, phenotype, function, and cytokine gene expression
between CD33
+
and CD11b
+
MDSC subsets and identify those features most clearly
associated with suppressive function. These studies confirmed the heterogeneous nature
of MDSC, showing a full spectrum of myeloid progenitor morphologies. Phenotypic
studies also supported an immature population of myeloid lineage cells (CD14
-
CD68
-
CD11c
-
) with both CD33
+
and CD11b
+
suppressor cells distinguished from non-
60
suppressive counterparts by low HLA-DR expression. Interestingly, this induction of
suppressive function in myeloid cells appeared to diverge at the level of CD33
+
and
CD11b
+
expression, which may correlate with monocytic and granulocytic subsets as
reported for murine MDSC [13-17]. The heterogeneity exhibited in the phenotype of
these MDSC is likely to reflect discrete pathways for induction, subtleties in functional
repertoires, and even their ultimate fates within the immune system.
MDSC have multiple mechanisms by which they can suppress T cell effector responses,
and both CD33
+
and CD11b
+
subsets of MDSC showed up-regulation of canonical
suppressive mechanisms (ARG-1, iNOS, NOX2). Previously, we demonstrated that
subtle variations emerged in the patterns of suppressive genes that were up-regulated in
human myeloid suppressor cells by different cytokine mixtures associated with active
suppressive function [28]. Similarly, human MDSC induced by a range of human solid
tumor cell lines exhibited small differences in the up-regulation of suppressive genes that
likely result from subsets within the broadly defined CD33
+
myeloid suppressor cell
population. Stratification into CD11b
+
and CD33
+
subsets showed a trend toward greater
ARG-1 and NOX2 expression in CD33
+
MDSC.. These results likely reflect the
complexity of myeloid suppressor cells, and will require finer dissection in future studies.
Another conclusion that can be made from the characterization of suppressive function in
CD33
+
MDSC is that great redundancy exists in the pleotropic mechanisms of
suppression, such that selective inhibition of ARG-1, iNOS, NOX2, TGF β, or VEGF
alone is unable to produce significant reversal of tolerance.
61
A better therapeutic approach, then, is likely to evolve from inhibition of the transcription
factors promoting the suppressive phenotype. Here we showed that HIF1α and STAT3
are critical transcription factors in CD33
+
human MDSC and C/EBP β in CD11b
+
MDSC,
respectively, and that effective inhibition of these subsets is accompanied by selective
down-regulation of these transcription factors. Transcriptional up-regulation and cellular
accumulation of HIF1 α, which dimerizes and translocates to the nucleus with
constitutively expressed HIF1β protein, is induced by hypoxia, a common feature of the
tumor microenvironment [26]. Furthermore, under normoxic pro-inflammatory
conditions, HIF1 α may be induced by inflammatory cytokines and/or protected from
degradation by NO-mediated nitrosylation [55-57]. As such, HIF1 α is a logical candidate
to mediate MDSC induction in the tumor microenvironment. Indeed Corzo et al. [25]
identified a role for HIF1 α in promoting ARG-1 and iNOS expression in tumor-
associated murine monocytic MDSC. Here we show that HIF1 α level correlates with
suppressive function in the CD33
+
subset of human MDSC in conjunction with STAT3
up-regulation. Furthermore, we propose that HIF1 α would be an ideal tumor-specific
target for MDSC inactivation due to the concentration of low oxygen tension and tumor-
derived inflammatory mediators (e.g. IL-6, IL-1 β) in the tumor microenvironment.
C/EBP β appears to be a key regulator of the human CD11b
+
MDSC, in agreement with a
previous study showing inhibition of tumor-derived murine and cytokine-induced human
CD11b
+
MDSC by transcript silencing of C/EBP β [51]. C/EBP  has been shown to
work cooperatively with nuclear factor kappa B (NF- B) in promoting the transcription
62
of inflammatory cytokines in myeloid cells [58], suggesting a role for NF B signaling in
the promotion of MDSC function as well. This is consistent with our findings that
inflammatory mediators IL-6, IL-1, tumor necrosis factor (TNF)- , and GM-CSF can
induce suppressive function in myeloid cells (Chapter 2) and are over-expressed
proportionally to MDSC induction capacity by human tumor cell lines. Furthermore, NF-
B and HIF1  signaling pathways have been shown to cause transactivation in some
immune cell populations, particularly upon LPS stimulation [59]. In this study, LPS
stimulation of human MDSC, regardless of subset, resulted in increased expression of
HIF1  and C/EBP , which is consistent with transactivation of these signaling pathways
in human MDSC. The studies presented here highlight the functional diversity within
human myeloid suppressor cells and provide a foundation for necessary future
investigations of these transcription factors in controlling MDSC fate. This
characterization of human CD33
+
and CD11b
+
MDSC induced by tumor cell lines
supports the concept of functionally convergent cell populations that nevertheless retain
divergent induction programs. These data suggest that therapies seeking to inhibit human
MDSC at the level of conversion from normal myeloid cells will need to target multiple
paths of induction occurring through STAT3 and HIF1 α or C/EBP β. Inflammatory
pathways appear to be major drivers of the suppressive functions in human MDSC
induced by tumors and should be investigated as means of MDSC generation in sepsis
and trauma patients where elevations of IL-6, IL-1 , and TNF-  are common [26,44].
These studies also highlight a potential means of high-throughput screening for MDSC-
targeted therapies using the down-regulation of STAT3/HIF1 α or C/EBP β as correlates
63
of inhibited suppressor function. Lastly these studies suggest that CD33
+
HLA-
DR
low
HIF1 α
+
and CD11b
+
HLA-DR
low
C/EBP β
+
are highly specific phenotypes that may
be used to isolate and study MDSC in cancer patients.
64
CHAPTER 4.
FUNCTIONAL CHARACTERIZATION OF HUMAN MYELOID-
DERIVED SUPPRESSOR CELLS INDUCED BY CYTOKINES FROM
NORMAL PERIPHERAL BLOOD MONONUCLEAR CELLS
Our work on human myeloid-derived suppressor cells (MDSC) was approached from the
perspective of cancer immunotherapy, in which suppressor cells are a barrier to effective
anti-tumor immune responses and a primary goal becomes inhibiting such suppression.
However, immune tolerance has both homeostatic and pathologic functions, so that
suppression of immune responses against self tissues is a key component of normal
immune function. Consequently, activation of robust immunity is almost immediately
followed by an accompanying increase in suppressive mechanisms like CD28 competitor
CTLA-4, induction of peripheral regulatory T cells (Treg) during chronic inflammation,
activation induced T cell death that temper and terminate immune responses that have run
their course [61]. It is not surprising then that MDSC are also noted to accumulate in the
settings of acute severe infection, sepsis, and chronic inflammation [16,62]. In this
capacity, the peripheral tolerance mediated by MDSC is protective and may help to
prevent unintended immune attack of self tissues or excessive activation of immune
signaling and mediators that overwhelms the host homeostatic mechanisms (i.e. sepsis).
It is possible, then, that MDSC could be used therapeutically in autoimmune disease or
organ transplantation to provide immune tolerance. We identified interleukin (IL)-1 and
IL-6 with granulocyte macrophage colony stimulating factor (GM-CSF) as potent
inducers of functionally active human MDSC from normal, non-suppressive myeloid
cells (Chapter 2, [28]). In the present study, we have characterized these cytokine-
65
induced MDSC as to their morphology, phenotype, gene expression, and suppressive
function, with the aim of identifying an immunosuppressive population of human cells
that can be expanded in vivo or adoptively transferred for the treatment of autoimmune
reactions in humans.
1
Autoimmunity often arises in situations of failed or insufficient immune tolerance, and
accordingly, many therapeutic approaches to the treatment of autoimmune disease aim to
temper immune responses and induce tolerance [61]. While reversal of immune
tolerance is a goal of cancer therapy, it is important to remember that such regulatory
cells in the body serve an essential physiologic function of protecting self tissues during
periods of immune activation. Myeloid-derived suppressor cells and regulatory T cells
that are barriers to successful therapy in cancer may in fact be simultaneously the
therapeutic solution for autoimmune disease and organ transplantation. In the same way
that adoptive transfer of activated and antigen-specific T cells and dendritic cells (DC)
could boost anti-tumor immunity in a tolerized cancer patient, adoptive transfer of MDSC
and Treg that are generated ex vivo could augment tolerance in a patient with
autoimmune reactions. The ability to generate suppressive cells ex vivo from an
individual’s normal or auto-reactive immune cells holds great possibility for
understanding and treating autoimmune disease.
1
Parts of this chapter were previously published in the Journal of Immunology (Lechner et al. Characterization
of cytokine-induced myeloid-derived suppressor cells from normal human peripheral blood mononuclear cells.
Journal of Immunology. 2010;185(4): 2273-2284).
66
Studies in murine models of autoimmune disease have demonstrated the therapeutic
potential of adoptive transfer of tolerogenic cells to attenuate or abrogate inappropriate
immune responses. Two collaborations, Clare-Salzler et al. [63] and Zhang et al. [64],
showed prevention of insulinitis and development of diabetes in murine models of insulin
dependent diabetes mellitus by transfer of ex vivo induced tolerogenic immature dendritic
cells (iDC). In the latter study, iDC were generated from bone marrow by culture with
GM-CSF and subsequently were injected intraperitoneally prior to streptozoicin-induced
β-islet deletion. Adoptively transferred, non-antigen-primed DC successfully inhibited
the onset of insulinitis via the expansion of Treg cells. In vitro generated tolerogenic
MDSC or iDC have also been shown to mediate potent immune suppression in models of
experimental autoimmune encephalomyelitis, collagen-induced arthritis, and
endotoxemia [65]. In murine transplantation studies, transfer of tolerogenic DC or
MDSC reduced the incidence of allograft rejection and graft-versus host disease [65-67].
Adoptive transfer of tolerogenic antigen-presenting cells holds great potential for the
immunotherapy of autoimmune disease and tissue transplantation, as demonstrated by
these reports. However, much of this foundational work has been performed in a murine
system with derivation of tolerogenic cells from embryonic stem cell, bone marrow, or
splenic populations [64-68]. Recently, we identified GM-CSF and IL-6 as the most
potent cytokines for the induction and expansion? of human MDSC from normal
peripheral blood mononuclear cells (PBMC) [28]. Concurrent with our report, Marigo et
al. [51] demonstrated that GM-CSF and IL-6-treated myeloid cells from naïve mouse
splenocytes had suppressive function and could be adoptively transferred to inhibit the
67
onset of autoimmune diabetes in a mouse model. To promote the translation of adoptive
tolerogenic DC therapy to the clinical setting, human-specific induction protocols are
needed that rely upon practically obtainable source cells. Further characterization of ex
vivo-generated MDSC is important to understand biology of these cells and to facilitate
their translation into clinical use. In this report, we characterize the morphology,
phenotype, gene expression pattern, and suppressive function of cytokine-induced human
CD33
+
suppressor cells.
MATERIALS AND METHODS
In vitro generation of human MDSC
i. Induction by cytokines
Human PBMC were isolated from healthy volunteer donors by venipuncture, followed by
differential density gradient separation (Ficoll Hypaque, Sigma, St. Louis, MO). PBMC
were cultured in T-25 flasks at 5x10
5
cells/mL in complete medium for seven days,
supplemented with cytokines as indicated. Recombinant human cytokines used for
induction include: IL-1β (10ng/mL, Sigma), IL-6 (10ng/mL, Sigma), PGE
2
(1µg/mL,
Sigma), TGF β1 (2ng/mL, R&D Systems, Minneapolis, MN), TNF α (10ng/mL, R&D),
VEGF (10ng/mL, Sigma), and GM-CSF (20ng/mL, R&D). For combination-cytokine
induction experiments, the cytokines used are indicated in the Results section. GM-CSF
was added to the mixtures of cytokines to support cell viability, as described previously
[35]. PBMC cultured in medium alone were run in parallel as a control for each donor.
Cultures were run in duplicate, and medium and cytokines were refreshed every two-
68
three days. For all studies, USC Institutional Review Board approval was obtained and a
total of 18 male and 7 female donors ages 23-62 were used under protocol HS-06-00579.
Data were derived from at least two individuals and no inter-donor differences in MDSC
induction or function were observed.
ii. MDSC Isolation
After one week, all cells were collected from cytokine-induced PBMC cultures. Adherent
cells were removed using the non-protease cell detachment solution Detachin (GenLantis,
San Diego, CA). Myeloid (CD33
+
) cells were then isolated from the co-cultures using
anti-CD33 magnetic microbeads and LS column separation (Miltenyi Biotec, Germany)
as per manufacturer’s instructions. Purity of isolated cell populations was found to be
greater than 90% by flow cytometry and morphological examination. Viability of
isolated cells was confirmed using trypan blue dye exclusion and samples with viability
less than 80% were excluded from analysis and repeated in subsequent experiments.
iii. Suppression Assay
The suppressive function of tumor-educated CD33
+
cells was measured by their ability to
inhibit the proliferation of autologous T cells in the following Suppression Assay: CD8
+
T cells isolated from 30mL of PBMC from returning healthy donors were CFSE-labeled
(3 μM, Sigma) and seeded in 96-well plates at 2x10
5
cells/well. CD33
+
cells isolated
previously (ii. MDSC isolation, above) were added to the 96-well plates at a ratio of 1:4
or 1:2 relative to the T cells. T cell stimulation was provided by anti-CD3/CD28
69
stimulation beads (Invitrogen, Carlsbad, CA). Suppression Assay wells were analyzed by
flow cytometry for T cell proliferation after three days. Controls included a positive T
cell proliferation control (T cells alone) and an induction negative control of CD33
+
cells
isolated from PBMC cultured in medium only. Samples were run in duplicate and data
were collected as percent proliferation for 15,000 cells. Samples were run on a
FACSCalibur flow cytometer (BD Biosciences, San Jose, CA) and data acquisition and
analysis were performed using CellQuestPro software (BD) at the USC Flow Cytometry
core facility.
Morphology of cytokine-induced MDSC
Wright-Giemsa staining (Protocol Hema 3, Fisher, Kalamazoo, MI) of CD33
+
cell
cytospin preparations was performed to assess the morphology of cytokine-induced
CD33
+
cells. Freshly isolated PBMC and CD33
+
cells from healthy donors were
prepared in parallel for comparison. Observation, evaluation, and image acquisition were
performed using Leica DM2500 microscope (Leica Microsystems, www.leica-
microsystems.com) connected to an automated, digital SPOT RTke camera and SPOT
Advanced Software (SPOT Diagnostic Instrument Inc., www.diaginc.com). Images were
resized and brightened for publication using Adobe Photoshop software (Adobe,
www.adobe.com).
70
Flow cytometry analyses of cell phenotypes
The phenotype of in vitro-generated MDSC was evaluated for expression of CD33, HLA-
DR, CD11b, CD11c, CD66b, CD68, CD14, and IL-13R α2 and compared to whole
PBMC and non-induced CD33
+
cells. Changes in PBMC subpopulations during cytokine
induction and the expansion of Treg cells in CD33
+
-T cell co-cultures also were
measured by flow cytometry. For staining, cells were collected from flasks using
Detachin (Genlantis) to minimize cell surface protein digestion, washed twice with FACS
buffer (2% FCS in PBS), and 10
6
cells were resuspended in 100ul FACS buffer. Cells
were stained for 1hr on ice with cocktails of fluorescently-conjugated monoclonal
antibodies or isotype-matched controls, then washed twice with FACS buffer, and
resuspended in FACS buffer for analysis. For intracellular staining, cell surface staining
was performed first, followed by buffer fixation/permeabilization (eBioscience, San
Diego, CA) and intracellular staining. Antibodies used were from BD: CD4 (RPA-T4),
CD8 (RPA-T8), CD11b (ICRF44), CD11c (B-ly6), CD14 (M5E2), CD25 (M-A251),
CD33 (HIM3-4, WM53), CD66b (G10F5), FoxP3 (259D/C7), HLA-DR (G46-6); Santa
Cruz Biotech: IL-13R α2 (B-D13); Miltenyi Biotec: CD25 (4E3); and eBioscience: CD1a
(HI149), CD3 (OKT3), CD20 (2H7), CD56 (MEM188), CD68 (Y1/82A), and isotype
controls. Samples were run on a FACSCalibur flow cytometer (BD) and data acquisition
and analysis were performed as above. Samples were run in duplicate and PBMC
cultured in medium alone were run in parallel for comparison. Data were acquired as the
fraction of labeled cells within a live-cell gate set for 15,000 events.
71
Quantitative RT-PCR for gene expression of cytokine-induced MDSC
For gene expression studies, cytokine-induced CD33
+
cells isolated from whole PBMC
cultures by fluorescence activated cell sorting after Cytokine Induction described above.
RNA was isolated from cells and DNase-treated using Qiagen RNeasy Micro kit. For
quantitative RT-PCR, 100ng of DNase-treated RNA was amplified with gene specific
primers using one-step Power SYBR green RNA-to-Ct kit (Applied Biosciences) and run
in an Mx300P Strategene thermocycler. Data were acquired and analyzed using MxPro
software (Stratagene). Gene expression was normalized to a housekeeping gene
(GAPDH), and fold change determined relative to human standard RNA (Stratagene) and
medium only control CD33
+
cells. Primer sequences were obtained from the NIH qRT-
PCR database (http://primerdepot.nci.nih.gov) and were synthesized by the USC Core
Facility [46].
Transwell assays
The MDSC Suppression Assay was modified to test contact dependency of MDSC-
mediated T cell suppression. For these studies, cytokine induced CD33
+
cells isolated
from Cytokine Induction cultures and fresh autologous CFSE-labeled T cells were co-
cultured at a 1:4 ratio in the presence of T cell stimuli for 3 days, as described above. In
deviation from the methods above, transwell plates (0.4 μm pores, Griener Bio-One,
Wemmel, Belgium) were used for these studies, with T cells cultured in plate wells and
CD33
+
cells cultured in transwell inserts to inhibit direct T cell – CD33
+
cell contact.
72
Cytokine induction of CD33
+
MDSC from T cell-depleted PBMC
To determine the role, if any, of T cells, the Cytokine Induction Assay was repeated for
select cytokine mixtures using T cell-depleted PBMC from healthy volunteers, rather
than whole PBMC. For these studies, PBMC obtained after density gradient separation
were depleted of T cells using anti-CD3 magnetic microbeads and LD column separation
(Miltenyi Biotec), per manufacturer’s instructions. Subsequent culture, cell isolation, and
function evaluation of cytokine induced CD33
+
cells from these cultures was performed
as described in the MDSC Isolation and Suppression Assays above.
Statistical analysis
For MDSC Suppression Assays, CFSE fluorescence was measured by flow cytometry to
determine the percent proliferation for each sample relative to T cell controls. Where
possible, mean relative proliferation and SD were calculated and graphed using Microsoft
Excel software (Microsoft, Redmond, WA). A change in mean T cell proliferation in the
presence or absence of tumor-educated MDSC was tested for statistical significance
using the student t test for independent samples. For all analyses in which multiple
experimental samples were compared against one another (e.g. qRT-PCR), a one-way
ANOVA was performed followed by Dunnett post-test analysis using GraphPad Prism
(La Jolla, CA). For all statistical tests, significance level α = 0.05.
73
RESULTS
Human CD33
+
MDSC are generated in vitro by soluble immune modulatory factors
Studies of human tumor cell line induction of MDSC identified IL-1β and IL-6 with GM-
CSF as key factors for the generation of active suppressor cells from normal healthy
donor PBMC. Briefly, PBMC were cultured for one week in the presence or absence of
immune modulatory factors GM-CSF and IL-1β, IL-6, VEGF, TGF β, TNF α, or PGE
2
(as
a proxy for COX2 over-expression [47]), and then CD33
+
cells were isolated and tested
for suppressive function by MDSC Suppression Assays. In these studies cytokine-
educated were shown to be sufficient to inhibit the proliferation of CD3/CD28 stimulated
autologous T cells at ratios of 1:4 and 1:2 (Chapter 2 and Figure 15A). Of the cytokine
mixtures evaluated for induction, IL-6 and GM-CSF showed the greatest efficacy and
GM-CSF alone or with IL-1, VEGF, or TNF  also generated suppressive CD33
+
myeloid cells. In this study, we expand upon these data by evaluating the ability of
cytokine-educated MDSC to inhibit IFNγ production in addition to T cell proliferation.
Multi-cytokine combinations were also tested for the induction of CD33
+
MDSC. The
ability of cytokine-induced CD33
+
to inhibit autologous T cell proliferation and IFNγ
production was evaluated at cell ratios of 1:2 and 1:4 (Figure 15A and 15B).
As shown previously (Chapter 2), potent CD33
+
suppressor cells were induced with IL-6
and GM-CSF, as well as GM-CSF alone or with IL-1, VEGF, and TNF  to a lesser
extent (Figure 15A). Combinations of these cytokines did not significantly increase the
suppressive capacity of induced MDSC, suggesting a final common pathway for these
74
cytokines leading to a common functional change in myeloid cells. Of note, TGF β does
not appear to have a major role in the promotion of suppressive function in CD33
+
cells
since TGF β treatment consistently decreased the potency of other cytokines in the
induction of suppressive function (Figure 15A and 15B).
Consistent with strong inhibition of autologous T cell proliferation, CD33
+
cells induced
by GM-CSF + IL-6 significantly inhibited IFNγ production in autologous T cells at 1:2
and 1:4 ratios (Figure 15B). CD33
+
suppressor cells induced by other cytokine mixtures
also inhibited IFNγ expression by autologous T cells, though none to the same extent as
GM-CSF + IL-6-generated CD33
+
cells. Mean IFNγ levels in responder T cells were
statistically lower, though to a lesser extent, in co-cultures with CD33
+
cells induced by
cytokine mixtures GM-CSF + IL-1β, GM-CSF + PGE
2
, GM-CSF + VEGF, and GM-CSF
alone at ratios of 1:2 and 1:4 (Figure 15B). Induction of MDSC by combination of
multiple cytokines identified, GM-CSF + IL-6 + VEGF and GM-CSF + IL-1β + PGE
2
combinations as capable of inducing moderately strong CD33
+
suppressor cells (Figure
15A and 15B).
75
Figure 15. Cytokine-induced CD33
+
MDSC demonstrate potent suppressive function. PBMC from
normal donors were cultured for one week in the presence of different cytokine mixtures. CD33
+
cells
were then isolated and tested for their ability to suppress the (A) proliferation and (B) INFγ production by
autologous T cells at ratios of 1:2 and 1:4. CD33
+
cells from cultures treated with GM-CSF and IL-6, Il-
1 β, VEGF, PGE
2
, or TNF α demonstrated suppressive function. For both graphs, mean is shown with SEM.
Conditions with statistically significant decreases in mean T cell proliferation compared to stimulated T
cells alone are indicated by an asterisk. C, Treg expansion by cytokine-induced MDSC in Suppression
Assays: the fraction of CD25
+
FoxP3
+
T cells (CFSE-labeled) at the conclusion of a three day Suppression
Assay with cytokine-induced CD33
+
cells and fresh autologous T cells was analyzed by flow cytometry.
Co-cultures with CD33
+
cells induced by GM-CSF + IL-6 or GM-CSF + TNF α showed increases in
CD25
+
FoxP3
+
T cells relative to stimulated T cells cultured alone or with CD33
+
cells from medium-only
cultures (n=1).
76
Figure 15, continued.
Although, TGF β-cultured CD33
+
cells did not inhibit T cell proliferation, these co-
cultures did demonstrate decreased IFNγ production. This finding suggested that
expansion of Treg cells may be occurring from fresh CD3
+
cells in the Suppression Assay
co-cultures [23,24]. To evaluate this possibility, responder T cells from Suppression
Assay co-cultures were stained for Treg markers after three days, and the fraction of
CD25
+
FoxP3
+
T cells (CFSE-labeled) was measured by flow cytometry (Figure 15C).
The fraction of CD25
+
FoxP3
+
from Suppression Assays of CD33
+
treated with GM-CSF
+ IL-6 or GM-CSF + TNF α was slightly increased (7.04% and 9.29%, respectively)
compared to stimulated T cells alone (6.07%) and T cells co-cultured with CD33
+
cells
from medium-only culture (5.82%) (Figure 15C). Suppression Assays with CD33
+
cells
induced by some other cytokine mixtures showed similar increases in the fraction of
CD25
+
FoxP3
+
T cells: GM-CSF + VEGF (8.49%), GM-CSF + IL-1β (7.55%), GM-CSF
+ TGF β (6.93%), GM-CSF (9.20%) (data not shown). These results suggest that Treg
expansion may contribute to MDSC-mediated T cell suppression in some settings, as
77
reported previously [23,24], though conclusions are limited by the low sample number
for these studies.
Morphology of cytokine induced CD33
+
MDSC
The morphology of cytokine-induced CD33
+
cell populations was examined by Wright-
Giemsa staining of cytospin cell preparations (Figure 16). The CD33
+
cells isolated from
cytokine-treated PBMC cultures showed remarkably homogeneous morphology and were
distinctly different from starting PBMC and freshly isolated CD33
+
cells. The
morphology of starting PBMC following differential density gradient separation was
typical of healthy human donors, showing numerous lymphocytes and mixed monocyte
populations with rare granulocytes and eosinophils. Non-induced CD33
+
cells isolated
from PBMC demonstrated a lymphocyte-depleted monocyte population of small cells, as
expected. In contrast, cytokine-induced CD33
+
cells were consistently large
mononuclear cells with basophilic, granular-appearing cytoplasm (Figure 16). As
reported previously, bi-nucleate cells were commonly observed in CD33
+
cells cultured
under numerous conditions [68]. With cytokine exposure, cells appeared to change from
blast-like cells with scant cytoplasm (Figure 16, panels 1-2) to cytoplasm predominant
cells (Figure 16, panels 3-8). However, no discernable difference was observed in
morphology between suppressive and non-suppressive cytokine-treated CD33
+
cells by
Wright-Giemsa staining.
Figure 16. Morphology of CD33+ suppressor cells resembles tumor
cytokine-induced CD33
+
populations compared to PBMC by Wright
magnification). Starting PBMC population (1) shows small granulocytic and monocytic cells scattered
amongst lymphocytes. CD33
following cytokine induction
suppressive cytokine-induced CD33
CD33
+
cells, though the cytoplasm of the latter frequently appear more basophilic
Phenotype of cytokine induced CD33
The surface phenotype of cytokine
and compared to the phenotypes of whole PBMC and CD33
alone (Figure 17). Flow cytometry analyses of cell forward and side scatter demonstrated
a primarily granulocytic CD33
. Morphology of CD33+ suppressor cells resembles tumor-induced MDSC.
populations compared to PBMC by Wright-Giemsa staining (x400, original
magnification). Starting PBMC population (1) shows small granulocytic and monocytic cells scattered
amongst lymphocytes. CD33
+
cells isolated after one week culture with complete medium alone (2) or
following cytokine induction (3-8) appear as large mononuclear cells with abundant cytoplasm. Non
induced CD33
+
cells (3,4) are similar in morphology to suppressive cytokine
cells, though the cytoplasm of the latter frequently appear more basophilic.
Phenotype of cytokine induced CD33
+
MDSC resembles tumor-induced MDSC
The surface phenotype of cytokine-treated CD33
+
cells was analyzed by flow cytometry
and compared to the phenotypes of whole PBMC and CD33
+
cells cultured in medium
). Flow cytometry analyses of cell forward and side scatter demonstrated
a primarily granulocytic CD33
+
population induced by cytokines, as compared to the
78
induced MDSC. Morphology of
staining (x400, original
magnification). Starting PBMC population (1) shows small granulocytic and monocytic cells scattered
cells isolated after one week culture with complete medium alone (2) or
8) appear as large mononuclear cells with abundant cytoplasm. Non-
cells (3,4) are similar in morphology to suppressive cytokine-induced
induced MDSC
cells was analyzed by flow cytometry
cells cultured in medium
). Flow cytometry analyses of cell forward and side scatter demonstrated
population induced by cytokines, as compared to the
79
largely monocytic morphology seen in PBMC and medium-only CD33
+
cell populations.
Staining for CD33 confirmed a high purity of target cells following anti-CD33 magnetic
bead labeling and column separation. The reported phenotype of human MDSC is
CD33
+
HLA-DR
low
CD11b
+
CD66b
+
, with low expression of differentiated macrophage
and dendritic cell markers (4,5). CD33
+
cells from all cytokine-treated cultures showed
similar increases in CD11b and CD66b expression relative to isotype controls, with low
to intermediate expression of monocyte/macrophage-associated markers CD68 and
CD14. Low expression of antigen presentation protein HLA-DR appeared to distinguish
suppressive from non-suppressive cytokine-treated CD33
+
cells. A small increase in IL-
13R α2 expression was also observed in suppressive, but not non-suppressive or control,
CD33
+
cells. For GM-CSF + PGE
2
induced CD33
+
cells, two discrete populations of
cells were noted. For this cytokine mixture, only the granulocytic population (shown
gated separately in Figure 17) expressed a phenotype consistent with human MDSC.
Overall, suppressive cytokine-induced CD33
+
cells generated by in vitro culture
demonstrated a phenotype consistent with that previously reported for human MDSC
[16,69].
80
Figure 17. Phenotype of cytokine-induced CD33
+
MDSC. Phenotypes of cytokine-induced CD33
+
cells
compared to whole PBMC and CD33
+
cultured in medium alone were determined by flow cytometry. The
expression of putative MDSC markers (CD33, CD11b, IL-13R α2, CD66b) and markers of mature antigen
presenting cells (HLA-DR, CD11c, CD14, CD68) was evaluated for each sample (black line) relative to
isotype controls (gray). In addition, forward and side scatter analyses of cells were performed to compare
size and granularity of cytokine-induced CD33
+
cells to controls. For GM-CSF + PGE
2
induced CD33
+
cells, two discrete populations of cells were noted. For this cytokine mixture, only the granulocytic
population (shown gated) expressed a phenotype consistent with human MDSC.
81
Cytokine-induced CD33
+
MDSC have up-regulated iNOS, TGF β, VEGF, and NOX2
MDSC-mediated suppression of effector T cell responses has been shown to correlate
with increased expression of ARG-1, iNOS, NOX2, VEGF, and TGF β by suppressor
cells [18-23]. To better characterize the nature of suppressive cytokine-induced CD33
+
cells, gene expression of these putative mechanisms of MDSC suppression were
evaluated by quantitative RT-PCR techniques (Figure 18). Cytokine-induced CD33
+
suppressor cells demonstrate significant up-regulation of iNOS, NOX2, VEGF, and/or
TGF β compared to freshly isolated, non-induced CD33
+
cells. iNOS, VEGF, and TGF β are consistently up-regulated by various inducing cytokine mixtures, while up-regulation
of NOX2 relates most closely with TNF α induction. GM-CSF + IL-6-induced CD33
+
cells appeared to mediate their potent suppressive function through ARG-1, iNOS,
VEGF, and TGF β, with minimal contribution from NOX2 (Figure 18A). Also, different
cytokine-induced myeloid suppressor cell groups demonstrated various patterns of gene
expression (Figure 18A). Previous reports have shown increased expression of inhibitory
B7-homologues on murine MDSC derived from some experimental tumor models,
though the functional significance of these markers is debated [14,70]. Gene expression
levels of PDL1 (B7H1), PDL2 (B7H2), and B7H4 were compared between suppressive
and non-suppressive cytokine-induced CD33
+
cells by quantitative RT-PCR studies
(Figure 18B). PDL1 expression was generally observed to be up-regulated in all groups
of cytokine-induced CD33
+
cells, though the increase was statistically significant only for
GM-CSF-induced cells (p<0.05). Cytokine-induced CD33
+
cells showed consistent
down-regulation of PDL2 and B7H4 genes relative CD33
+
from medium-only cultures,
82
with no differences between suppressive and non-suppressive cells (Figure 18B). These
results are consistent with reports from experimental tumor models in mice [14,70], and
do not support a role for PDL1, PDL2, or B7H4 in suppression mediated by these human
cytokine-induced CD33
+
cells.
Figure 18. The expression of suppressive genes by cytokine- and tumor-induced CD33
+
MDSC varies
with the inducing cytokine milieu. A, Gene expression of reported suppressive mechanisms in MDSC
(ARG-1, iNOS, NOX2, VEGF, TGF β) in different subsets of cytokine-induced CD33
+
MDSC as
determined by quantitative RT-PCR techniques. Mean fold change relative to CD33
+
cells cultured in
medium alone shown, +SEM. B, Gene expression levels of PDL1 (B7H1), PDL2 (B7H2), and B7H4 were
compared between suppressive and non-suppressive cytokine-induced CD33
+
cells by quantitative RT-PCR
studies. Mean fold change in expression relative to CD33
+
from medium-only cultures shown, with SEM.
Statistically significant values are indicated by an asterisk.
83
Suppression by human CD33
+
MDSC is contact dependent
A variety of suppressive mechanisms have been reported for human and murine MDSC,
though it is unclear whether or not they require cell-to-cell contact with target T cells. In
our assays, separation of suppressive cytokine-induced CD33
+
cells from responder T
cells by a transwell inserts (0.4 μm pores) was found to abrogate the suppressive effects of
all cytokine-induced MDSC (Figure 19). These data suggest that many, if not all,
suppressive mechanisms employed by these cells are contact dependent.
Figure 19. CD33
+
MDSC-mediated suppression of autologous T cells is contact dependent.
Suppressive cytokine-induced CD33
+
cells were co-cultured with fresh T cells isolated from the same
donor in a single well or separated by a transwell insert at a 1:4 ratio in a modified MDSC Suppression
Assay. Mean T cell proliferation is shown, with SEM. For all transwell samples, mean T cell proliferation
in the presence of cytokine-induced CD33
+
cells was not statistically significantly different from stimulated
T cells alone.
Cytokine induction expands CD56
+
, CD33
+
, and CD14
+
cell populations
Suppressive CD33
+
cells with MDSC-like morphology, phenotype, and gene expression
were successfully generated from whole PBMC derived from healthy volunteer donors.
To shed light on the cellular context of CD33
+
suppressor cell induction, changes in cell
types and numbers during cytokine induction were measured by flow cytometry analyses
84
(Figure 20A). All cytokine mixtures, with the exception of GM-CSF + VEGF, showed
expansion of CD56
+
(natural killer cells (NK)), CD33
+
(myeloid), CD14
+
(monocytes),
and to a lesser extent CD1a
+
(DC) cell populations relative to PBMC cultured in medium
alone. This effect occurs with GM-CSF alone, suggesting that this cytokine promotes
proliferation of cells in the myeloid and granulocyte compartment, consistent with its
known function in normal hematopoeisis [48]. Expansion of CD66b
+
cells above the
levels observed with GM-CSF alone occurred with the addition of IL-1β or PGE
2
to
cytokine mixtures. T cell numbers and CD4/CD8 ratio appeared to be largely unaffected
by cytokine induction. Cytokine mixtures also did not appear to affect the frequency of B
cells (CD20
+
) or macrophages (CD68
+
). Cytokine mixtures inducing CD33
+
suppressor
cells did not produce changes in cell types and numbers distinct from those observed for
cytokine mixtures that did not induce MDSC.
T cells are not needed for cytokine induction of CD33
+
MDSC from normal donor cells
T cells comprise a large fraction of PBMC and depletion of this population may increase
the yields of CD33
+
cells from cytokine-induced cultures. To determine whether T cells
were necessary for the induction of suppressive CD33
+
cells by the cytokine mixtures
used above, cytokine induction experiments were repeated using T cell-depleted PBMC.
The suppressive function of CD33
+
cells from whole PBMC or from T-cell depleted
PBMC cultures treated with cytokine mixes was then compared, as shown in Figure 20B.
CD33
+
MDSC generated in the absence of T cells demonstrated a comparable
suppressive capacity to those generated from whole PBMC for most cytokine mixtures
85
Figure 20. Cellular context for cytokine induction of CD33
+
MDSC from normal donor PBMC. A,
Changes in cell types and frequencies during cytokine induction were measured by flow cytometry.
Significant expansion of CD56
+
(NK), CD33
+
(myeloid), and CD14
+
(monocyte) cell populations was
observed in PBMC treated with all cytokine mixtures, with the exception of GM-CSF + VEGF. The
addition of IL-1 β or PGE
2
appeared to increase the frequency of CD66b
+
cells beyond that observed with
GM-CSF treatment alone. The number and CD4/CD8 ratio of T cells did not appear to be affected by the
cytokine combinations examined here, nor did the frequency of B cells (CD20
+
) or macrophages (CD68
+
).
Cytokine mixtures inducing CD33
+
suppressor cells did not produce changes in cell types and numbers
distinct from those observed for cytokine mixtures that did not induce MDSC. B, The suppressive function
of CD33
+
cells from whole PBMC or from T-cell depleted PBMC cultures treated with cytokine mixtures
was compared to determine the requirement of T cells in the induction of MDSC. CD33
+
MDSC generated
in the absence of T cells demonstrated a comparable suppressive capacity to those generated from whole
PBMC for most cytokine mixtures examined. However, for cytokine mixtures GM-CSF alone, GM-CSF +
TNF α, and GM-CSF + VEGF, CD33
+
generated in the absence of T cells were more suppressive than those
generated from whole PBMC (p<0.05).
86
examined. Interestingly, induction studies using GM-CSF alone, GM-CSF + TNF α, and
GM-CSF + VEGF showed that CD33
+
generated in the absence of T cells were more
suppressive than those generated from whole PBMC (p<0.05).
DISCUSSION
Studies of MDSC induction by human cancer cell lines identified key tumor-derived
factors associated with MDSC generation from non-suppressive, normal donor PBMC in
the cancer setting. These data were then used to identify specific cytokine combinations
necessary and sufficient to induce functionally suppressive myeloid (CD33
+
) cells from
healthy donor PBMC as an in vitro model of tumor-MDSC induction. In the rpesent
study, these cytokine-induced myeloid suppressor cells were characterized for
morphology, phenotype, suppressive capacity, and gene expression and compared with
human tumor cell line-induced MDSC.
Six cytokine combinations were identified for the in vitro induction of potent CD33
+
MDSC-like suppressor cells from normal donor PBMC: GM-CSF, GM-CSF + IL-1β,
GM-CSF + IL-6, GM-CSF + PGE
2
, GM-CSF + TNF α, and GM-CSF + VEGF. Of these
cytokine mixtures, IL-6 and GM-CSF consistently generated the most suppressive CD33
+
cells, as measured by the ability to suppress proliferation of, and IFNγ production by,
autologous T cells at different ratios (Figure 15). This study also demonstrated that the
induction of CD33
+
MDSC from immune competent PBMC by select cytokine
combinations occurs in a context of CD33
+
/CD14
+
/CD56
+
cell expansion and does not
87
require T cells. Wright-Giemsa staining and flow cytometry analyses of cytokine-
induced CD33
+
cells and freshly isolated CD33
+
cells demonstrate a shift from small,
monocytic cells to a homogeneous population of large mononuclear cells with abundant
basophilic, granular cytoplasm (Figure 16). As reported previously for tolerogenic DC
[68], bi-nucleated cells were commonly observed in CD33
+
cells cultured under different
conditions. Importantly, these studies demonstrate that cell morphology alone does not
distinguish between CD33
+
MDSC and non-suppressive CD33
+
cells. In characterizing
the surface phenotype of cytokine-induced MDSC, this study examined expression of
putative MDSC-associated surface markers as well as lineage specific markers for DC
(CD11c), granulocytes (CD66b), and macrophages (CD68) (Figure 17). These results
corroborate CD66b and CD11b in conjunction with CD33 positivity as markers of human
myeloid cells with suppressive function [53,62,69]. Consistent with our phenotype of
tumor cell-line educated MDSC (Chapter 3) and reports from MDSC in cancer patients,
cytokine-induced CD33
+
cells with suppressive function showed lower expression of
HLA-DR.
CD33
+
MDSC induced by GM-CSF and IL-6 mediated potent suppression of autologous
T cell proliferation and IFNγ production. Suppressive myeloid cells were also generated
by cytokine mixtures of IL-1β, VEGF, TNF α, and PGE
2
, with GM-CSF added to support
cell viability, but the resulting suppressor cells were noticeably less potent (Figure 15).
In vitro-generated suppressive MDSC exhibited increased expression of iNOS, NOX2,
TGF β, and VEGF (Figure 18). Interestingly, slight variations in the gene expression
88
profile of, or Treg expansion by, suppressive CD33
+
cells occurred with different
inducing cytokines. IL-6 and GM-CSF induced MDSC appeared to exert strong
suppression through up-regulation of ARG-1, iNOS, VEGF, and TGF β, with minimal
contribution from NOX2 expression. This diversity reflects the variation in suppressive
mechanisms reported for MDSC from cancer patients and murine tumor models [18-23]
and further suggests the existence of MDSC subpopulations. As noted in Chapter 2,
MDSC-inducing tumor cell lines were observed to up-regulate expression of different
factors that may be responsible for MDSC generation (Figure 3), which may explain the
presence of different MDSC populations amongst tumors. In this study, the generation of
suppressive CD33
+
MDSC using multiple distinct cytokine mixtures also supports the
existence of diverse MDSC subpopulations (Figure 15).
The MDSC Suppression Assay described in this study tested the ability of CD33
+
cells to
inhibit T cell proliferation and IFNγ production in the presence CD3/CD28 stimulation.
This method is similar to suppression assays frequently used to evaluate the function of
regulatory T cells [12]. It is important to note that many investigators use antigen
presenting cell populations as feeder cells in Treg suppression assays. Just as Treg may
be expanded from responder CD3
+
T cell populations and contribute to suppression in
MDSC Suppression Assays (Figure 15), Treg induction of MDSC from antigen
presenting cells present in Treg Suppression Assays may contribute to inhibition of T
cells responses. As demonstrated in this study, CD33
+
cells may be induced from normal
89
donor cells by select cytokines and mediate potent suppression of autologous T cells
(Figure 15).
Previously, we reported the in vitro generation of human CD33
+
MDSC from PBMC by
direct co-culture with select human solid tumor cell lines (Chapter 2). In this study,
cytokines IL-6 and GM-CSF, and secondarily GM-CSF in combination with IL-1β,
PGE
2
, TNF α, or VEGF, are shown to be sufficient for the induction of functionally
suppressive CD33
+
cells from healthy donor PBMC. MDSC are not typically present in
healthy individuals [17], and the ability to generate these suppressor cells in vitro from
non-suppressive populations represents an important discovery. Furthermore, in
developing ex vivo MDSC induction protocols for autoimmunity treatment, a tumor cell-
free system with defined induction factors is highly preferable. Suppressor cells induced
by IL-6 and GM-CSF may enable the generation of tolerogenic myeloid cells for the
adoptive immunotherapy of autoimmune disease. The cytokine-induction combinations
identified here allow derivation of human MDSC from a starting population of common,
healthy PBMC. This is in contrast to reports of in vitro generation of murine MDSC
populations from embryonic, splenic, or bone marrow derived cells [51,67] – cell
fractions that are less readily available from human donors. While many similarities exist
between murine and human immune systems, studies in Treg highlight the need to be
cognizant of interspecies differences in surface markers and biology of immune
suppressor cells [12,71]. Thus, the discovery of human-specific MDSC induction
methods is a valuable tool for the further study of these suppressor cells. Based upon
90
these results, future investigations are warranted to study the potential of IL-6 and GM-
CSF as therapeutic targets for the inhibition of MDSC induction in cancer patients or as
means of induction for expansion of these cells in autoimmunity.
91
CHAPTER 5.
USC-HN2, A NEW MODEL CELL LINE FOR RECURRENT ORAL CAVITY
SQUAMOUS CELL CARCINOMA WITH IMMUNOSUPPRESSIVE
CHARACTERISTICS
During our screen of human tumor cell lines for the induction of myeloid-derived
suppressor cells (MDSC), we noticed that head and neck cancers appeared to be frequent
inducers of this population. At the time our laboratory had an ongoing collaboration with
the American Tissue Culture Collection (ATCC) and the USC Head and Neck Surgery
Group to develop an expanded collection of head and neck cancer cell lines to be made
publically available to facilitate studies in this disease. Two cell lines generated early in
this project, USC-HN1 and USC-HN2, displayed strikingly different immunomodulatory
behavior. Head and neck squamous cell carcinoma (HNSCC) is known for its strong
modulation of the immune system and is a frequent model for the development of new
immunotherapies. These new cell lines were characterized for morphology, phenotype,
tumor suppressor and oncogene expression, cytokine production, heterotransplantation
into nude mice, and cytogenetic abonormalities. From the results of these studies we
formulated a hypothesis describing two major mechanisms of tumor immune escape
taken by tumors, or by subsets of cells within them, as modeled by cell lines USC-HN1
and USC-HN2. USC-HN1 cell line modeled predominantly immune evasion as a method
of immune escape, showing down-regulation of select HLA receptors and antigen
presentation machinery concomitant with low expression of immune modulatory
cytokines and surface ligands. By contrast, USC-HN2 aggressively produced
immunosuppressive cytokines, induced MDSC and regulatory T cells, and up-regulated
92
antigen presentation. Ongoing studies in the lab are evaluating this hypothesis in head
and neck tumor biopsy specimens from patients and should shed light on the nature of
immune escape mechanisms in head and neck cancer. USC-HN1 and USC-HN2 have
been made publically available through ATCC and should facilitate pre-clinical studies of
immunotherapy for human HNSCC.
This project was done jointly with Sarah Russell
1
, with assistance from Lucy Gong
1
, and
collaboration with Rizwan Masood
2
, Adrian Correa
1
, and Uttam Sinha
2
for the collection
of patient biopsy specimens and analysis of pathology. Microarray experiments were
performed and analyzed by Jing Han
3
and Raj K Puri
3
. The cell lines USC-HN1 and
USC-HN2 were established by Alan Epstein.
1
Department of Pathology, USC Keck School of Medicine, Los Angeles, California
2
Department of Otolaryngology, USC Keck School of Medicine, Los Angeles, California
3
Tumor Vaccines and Biotechnology Branch, Division of Cellular and Gene Therapies, Center for
Biologics Evaluation and Research, Food and Drug Administration, Bethesda, Maryland, United States
93
Head and neck cancer is the sixth most common solid tumor malignancy worldwide, and
despite available surgical and adjuvant therapies, continues to account for significant
morbidity and mortality in the United States with 36,000 incident cases and 8,000 deaths
in 2010 [73,74]. Significantly, the five-year survival rate for patients with head and neck
squamous cell carcinoma (HNSCC) is poor (30-40%) and has seen only marginal
improvement in the past four decades [75,76]. These cancers can arise from the
epithelium of the sinonasal tract, oral cavity, pharynx, or larynx, and are predominantly
(>90%) squamous cell in origin [73]. A history of tobacco smoking, excessive alcohol
consumption, and, more recently, human papillomavirus (HPV) infection have been
identified as risk factors for HNSCC and each appear to promote tumorigenesis through
different mechanisms [73,75,77]. HPV-positive tumors are now considered to be a
distinct biomodel from HPV-negative HNSCC, whereby cellular transformation occurs
secondary to virally-mediated degradation of tumor suppressor proteins p53 and Rb
rather than mutation inactivation of genes p53 and p16
INK4
(Rb pathway downstream
component), as is seen in tumors from patients with significant smoking or alcohol use
and no HPV infection [73,75,78]. Histologically, HNSCC demonstrates features typical
of squamous cell carcinomas with varying degrees of keratinization; large, pleomorphic
nuclei with multiple, prominent nucleoli; abundant cytoplasm with vacuolation;
intercellular bridging; and frequent mitotic figures [73]. Surface and intracellular
markers also aid classification of HNSCC, with specific positivity seen for fatty acid
binding protein 5 (FABP5), epidermal growth factor receptor (EGFR), E-cadherin, CD74,
94
CD24, IL13R 2, and tumor stem cell markers CD44v6 and CD133 [79-81]. Cytogenetic
analysis of HNSCC typically reveals an aberrant chromosome set with various deletions,
translocations, and double chromosomes, including isochromosome formation and
breakpoints at or near the centromeres [81].
Until recently, HNSCC was considered to be a relatively uniform clinical disease with a
standard treatment regimen including surgical resection with adjuvant chemotherapy and
radiation for advanced disease [73]. However, recent epidemiologic trends, as well as
molecular, genetic and immunologic studies, have demonstrated that despite their
homogenous morphology HNSCC tumors in fact exhibit a diverse range of clinical
behaviors [73-75, 79,80,82]. The driving factor(s) behind indolent and locally invasive
compared with highly aggressive, widely metastatic neoplasms is the focus of much
ongoing research and will significantly impact future therapies. The immunologic
properties of HNSCC are of particular interest in this new era of cancer immunotherapy
where a growing understanding of the complex relationship between tumors and the host
immune system has allowed for the development of more individualized and targeted
treatments for many cancers [83]. It is now recognized that the immune system is
capable of recognizing and eliminating cancer cells in the body, termed immune
surveillance, but that tumors adapt to evade and escape immune attack [4,6]. Because
HNSCC arises from normal oropharyngeal and laryngeal mucosa, cancer cells can largely
resemble normal tissues to which the immune system is tolerized [84]. Furthermore,
immune evasion can be achieved through down-regulation of genes related to antigen
95
presentation, including select human leukocyte antigen (HLA) class I molecules, which
up to 50% of HNSCC tumors in patients have been found to exhibit [84,85].
Additionally, numerous groups have provided evidence of the immunomodulatory effects
of HNSCC, including the local and regional suppression of the immune system by
interleukins, vascular endothelial growth factor (VEGF), cyclo-oxygenase 2 (COX2), and
matrix metalloproteinases [84,86-89], suggesting that not all HNSCC tumors are able to
rely solely on evasive mechanisms for survival. These cytokines facilitate tumor
progression through their actions to promote angiogenesis, basement membrane invasion,
and the induction of immune tolerance [84,86-89]. Specifically, individuals with
aggressive HNSCC tumors are observed to have a Th2–weighted immune response, with
increased IL-6 and IL-10 cytokine production and decreased Th1 effector cells (natural
killer cells and cytotoxic T lymphocytes) and cytokines (IL-12, interferon gamma and
tumor necrosis factor alpha) [84,90,91]. It is generally recognized that a Th1 response
promotes the cell-mediated immunity that is needed for effective tumor cell killing and
disease regression, while a Th2-weighted response inhibits such activity [90,91].
HNSCC appear to exhibit a range of immune escape mechanisms that likely influence
disease progression and response to therapy. Moreover, the influence of HPV infection
on HNSCC tumors' ability to escape immune detection and killing remains unclear;
HPV-positive HNSCC tumors have significantly better prognoses than HPV-negative
tumors, which may be related to differences in the tumor-host immune response and
tumor-derived immune tolerance mechanisms [75,92]. These unanswered questions and
96
the alarmingly poor prognosis for HNSCC patients has prompted a greater need for pre-
clinical models with which to investigate the immunological properties of HNSCC.
Cancer cell lines are important models for pre-clinical studies of the molecular
pathogenesis and immunologic properties of disease, as well as the development of new
therapies. With regards to HNSCC in particular, while a variety of cell lines have been
generated, few have been made publicly available (12 HNSCC cell lines currently
available through the American Tissue-type Cell Collection (ATCC),
http://www.atcc.org), and many lack complete characterization. In this report, we
describe the establishment and characterization of a unique HNSCC cell line, designated
USC-HN2, derived from a tumor biopsy specimen from a patient with invasive, recurrent
buccal squamous cell carcinoma. USC-HN2 recapitulated well the phenotype of the
original tumor biopsy, and heterotransplantation and cytogenetic studies confirmed its
oncogenic derivation and monoclonality, respectively. By comparing it to previously
established HNSCC cell lines USC-HN1 [93] and SCCL-MT1, the latter having never
before been characterized in the literature, we demonstrate two divergent mechanisms of
immune escape among these models of HNSCC. As a heterogeneous group of tumors,
HNSCC likely requires more individualized treatment targeting these divergent
mechanisms of immune escape, the lack of which may account for the high rates of
primary-site recurrence and metastases to loco-regional lymph nodes leading to the
dismal long-term prognosis for HNSCC [77].
97
MATERIALS AND METHODS
Cell lines and tissues
Tumor cell lines were obtained from ATCC or gifted to the Epstein laboratory and
authenticity was verified by cytogenetics and surface marker analysis. Cell lines were
maintained in complete medium (RPMI-1640 with 10% FCS, 2mM L-Glutamine, 100
U/ml Penicillin, and 100ug/ml Streptomycin) in a humidified 5% CO
2
, 37°C incubator.
Informed consent and HNSCC tumor biopsies were obtained as described previously
[93]. IRB approval from the USC Keck School of Medicine (HS-09-00048) was
obtained for the collection and use of HNSCC tumor biopsies.
Establishment of cell line USC-HN2
Tumor explants were used to develop the USC-HN2 cell line, as described previously
[89]. After establishment of the cell line, interval screening was performed using
MycoAlert Mycoplasma Detection Kit (Lonza, Rockland, ME). Cell doubling time was
determined for USC-HN2 by cell count measurements at 24 hour intervals for one week.
Heterotransplantation in Nude mice
Twelve-week-old female Nude mice (n=3, Simonsen Laboratory, Gilroy, CA) were
injected with cultured USC-HN2 cells for heterotopic (s.c. flank, 7.5x10
6
cells) or
orthotopic (base of the tongue, 3x10
6
cells) heterotransplantation studies. Tumor
measurements were made twice weekly and animals were sacrificed two (oral cavity) or
98
four (flank) weeks after implantation. Institutional Animal Care and Use Committee-
approved protocols were followed.
Immunohistochemistry (IHC)
Cytospin preparations of USC-HN2 cells from culture and tissue sections of patient
biopsy and heterotransplanted tumors were used for IHC studies, as described previously
[93,94]. Wright-Giemsa staining (Protocol Hema 3, Fisher, Kalamazoo, MI) of USC-
HN2, USC-HN1, and SCCL-MT1 cytospin preparations was performed to assess and
compare morphology, as described previously [93,94]. Both USC-HN2 cytospin and
paraffin tissue slides were stained for specific antigens with monoclonal antibodies
including: CD44 (DF1485; Dako Corp., Carpinteria, CA), E-cadherin (4A2C7;
Invitrogen, Carlsbad, CA), epidermal growth factor receptor (EGFR) (E30; Biogenex,
San Ramon, CA), keratin (AE1/AE-3; Covance, Berkeley, CA), p53 (1801; CalBiochem,
San Diego, CA), Rb (RbG3-245; BD Biosciences, San Diego, CA), p16 (INK4) and
FABP5 (311215) (R&D Systems, Minneapolis, MN). Observation, evaluation, and
image acquisition were made as described previously [93,94].
Analysis of surface markers by flow cytometry
Single cell suspensions (10
6
cells in 100µl) in 2% FCS in PBS were stained with
fluorescence-conjugated antibodies, as described previously [93,94]. For intracellular
stains, buffer fixation/permeabilization (eBioscience, San Diego, CA) was performed
prior to staining. Antibodies were from BD Biosciences: CD24 (ML5), CD74 (M-B741),
99
E-cadherin (36/Ecadherin), EGF receptor (EGFR1), Nanog (N31-355), Oct 3/4 (40/Oct-
3), SOX2 (245610), isotype controls; Santa Cruz Biotechnology (Santa Cruz, CA): IL-
13R α2 (B-D13), c-kit (104D2); Abcam (Cambridge, MA): CD44v6 (VFF-7); and
eBioscience: CD133 (TMP4), isotype controls.
Cytogenetics and in situ hybridization
Karyotype analysis using Giemsa staining and in situ hybridization for HPV DNA
sequences were performed by the Division of Anatomic Pathology, City of Hope Medical
Center (Duarte, CA) using early passages of USC-HN2, USC-HN1, and SCCL-MT1. To
screen for HPV infectivity, single color FISH was preformed using Enzo Life Sciences
HPV16/18 probe (ENZO-32886, Plymouth Meeting, PA) followed by tyramide signal
amplification (TSA kit#21, Invitrogen).
Microarray gene expression profiling
Total RNA was isolated from USC-HN2, USC-HN1, and SCCL-MT1 using RNeasy
Mini Kit (Qiagen, Valencia, CA) and analyzed by microarray, as previously described
[93]. Human universal RNA (huRNA; Stratagene, Santa Clara, CA) was used as a
common reference for all experiments. For data analysis, data files were uploaded into
mAdb database and analyzed by the software tools provided by the Center for
Information Technology (CIT), NIH. SAM (Significance Analysis of Microarray) and t-
test analyses were performed to identify differentially expressed genes. In addition,
100
GSEA (Gene Set Enrichment Analysis) provided in mAdb was also performed to
distinguish groups of differentially expressed genes in these cell lines.
TP53 mutation analysis
Genomic DNA isolated as above, was amplified using primers for exons 5-9 of TP53, as
described by Dai et al. [78]. Purified PCR products were sequenced by the USC DNA
core facility using ABI 3730 DNA Analyzer (Applied Biosystems) and screened for
mutations using BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi).
Cytokine and oncogene analysis by quantitative(q)RT-PCR
Gene expression analyses by qRT-PCR were performed on USC-HN2, USC-HN1, and
SCCL-MT1 cell lines as described previously [28,93].
Measurement of tumor-derived factors by ELISA
Three-day supernatants were collected from cell line cultures at 90% confluence (all cell
lines have comparable growth rates), passed through a 0.2µm filter unit to remove cell
debris, and analyzed for protein levels of IL-1β, IL-6, IL-8, TNF α, VEGF and GM-CSF
using ELISA DuoSet kits (R&D) per manufacturer's instructions. Plate absorbance was
read on an ELX-800 plate reader (Bio-Tek, Winooski, VT) and analyzed using KC Junior
software (Bio-Tek).
101
Induction of regulatory T cells and myeloid-derived suppressor cells
USC-HN2, USC-HN1 and SCCL-MT1 cell lines were assessed for their ability to induce
regulatory T cells (Treg) and myeloid-derived suppressor cells (MDSC), as described
previously [28,95]. Briefly, PBMCs obtained from healthy volunteers were co-cultured
in complete medium (6x10
5
cells/mL) with tumor cells for one week. Tumor cells were
seeded to achieve confluence by day seven. After one week in co-culture, CD33
+
MDSC
or CD4
+
CD25
high
Treg cells were isolated using magnetic bead separation techniques.
The suppressive function of tumor-educated MDSC or Treg cells was assessed by their
ability to inhibit the proliferation of fresh, autologous CFSE-labeled (3 μM) T cells
stimulated with CD3/CD28 beads in vitro. T cell proliferation was measured by flow
cytometry after three days.
Statistical analysis
To identify statistically significant differences in gene and protein expression by HNSCC
cell lines and suppression of T cell proliferation by co-cultured CD33
+
or CD4
+
CD25
high
cells, one-way ANOVA followed by Dunnett post-test was applied. Statistical analyses
for microarray experiments were performed as described above. Statistical tests were
performed using GraphPad Prism software (La Jolla, CA) at a significance level of 0.05.
Graphs and figures were produced using GraphPad Prism, Microsoft Excel, and Adobe
Illustrator and Photoshop software.
102
RESULTS
Case report of patient with recurrent invasive left buccal squamous cell carcinoma
The patient is an 81-year-old female with a past medical history of recurrent left sided
oral cancer and a 50-pack year history of tobacco smoking and occasional alcohol
consumption. The patient initially presented to USC University Hospital in March, 2002
with complaints of a left buccal ulcerated, non-healing lesion. A 3.5cm left cheek mass
was resected in April, 2002. Surgical pathology of the resected mass showed moderate-
to-poorly differentiated squamous cell carcinoma (SCC) of the oral cavity. After a
second surgical resection for recurrence in August, 2002, the patient remained disease
free until July, 2009, when she presented with a left buccal mass with contiguity to the
left mandible. The patient underwent a third surgical resection for suspected recurrence
in August, 2009 which revealed a 4cm moderately differentiated SCC of the buccal
mucosa with bone and perineural invasion, but no evidence of vascular invasion or tumor
metastasis to submental, submandibular, maxillary, oral cavity, or floor of mouth lymph
nodes (Stage IV, T4N0M0; Figure 21). The patient did not receive any radiation or
chemotherapy treatment. The patient is currently tumor-free and continues to have
routine follow-up at the USC University Hospital.
103
Figure 21. Histology of the
original tumor and hetero-
transplants. (Upper and lower
left panels) H&E stained sections
of the original tumor show
groups of cells infiltrating stroma
with a desmoplastic and dense
lympho-plasmacytic reaction,
and occasional keratin pearl
formation (arrow). The cells
feature an increased nuclear to
cytoplasmic ratio with
occasionally prominent nucleoli
and scattered mitotic figures
(H&E x200 and x400 original
magnification). (Upper and
lower right panels)
Subcutaneous heterotransplanta-
tion of USC-HN2 cell line
demonstrates a keratinizing
tumor (arrow) recapitulating the
original tumor histology (H&E
x200 and x400 original magnification).
Establishment of USC-HN2 cell line
The USC-HN2 cell line was derived from the patient’s recurrent buccal mucosal SCC
resected in August, 2009 using explant fragments that adhered to the surface of a plastic
flask. After 2-3 weeks, tumor cells were removed by trypsinization and placed in petri
dishes to enable cloning procedures required to isolate a cell line from normal stromal
cells. Tumor cells isolated from several fragments were pooled and used to establish the
USC-HN2 cell line. These cells were found to have a rapid doubling time of 22 hours,
which is comparable to the previously reported growth rates of other HNSCC cell lines
(26.5 hours) [76]. Once a morphologically uniform population of cells was established,
several freezings were performed to obtain early passages of USC-HN2. From one of
these early freezings, vials of frozen cells were sent to ATCC for distribution to other
investigators.
104
Heterotransplantation in Nude mice
Viable USC-HN2 cells from culture were injected in the oral cavity or subcutaneously in
athymic Nude mice (n=3) and tumors were excised after two (tongue) or four
(subcutaneous) weeks (Figure 21). Subcutaneous tumors grew to between 110mm
3
and
150mm
3
and oral cavity tumors were excised once visible tumors had grown (3mm
3
; data
not shown). H&E stained sections of the heterotransplants showed a moderately to
poorly differentiated, keratinizing SCC. Surrounding the invasive tumor, a mild to
moderate chronic and acute inflammatory infiltrate was present. These findings
demonstrate that USC-HN2 is transplantable in xenograft models and that
heterotransplanted tumors closely resemble the original tumor.
Morphology of USC-HN2 cell line is typical of oral cavity squamous cell carcinoma
Phase-contrast photomicrographs of cultured cells and Wright-Giemsa stained cytospins
were used to assess the morphology of USC-HN2 cell line as compared to established
HNSCC cell lines USC-HN1 and SCCL-MT1 (Figure 22A). All three cell lines
demonstrated characteristic features of oral cavity squamous cell carcinoma. Cytology of
the USC-HN2 cell line showed nuclear pleomorphism with prominent nucleoli, frequent
mitotic activity, and an abundance of vacuolated cytoplasm. USC-HN1, in contrast,
demonstrated a more blast-like cytology when compared with USC-HN2 and SCCL-MT1
and grew as a more tightly packed monolayer.
105
Cytogenetics
Cytogenetic analysis of USC-HN2 and SCCL-MT1 cells was performed in order to
confirm the unique identify of each cell line. All mitotic cells collected for GTG-band
analysis from USC-HN2 cell culture were clonally abnormal. The hyperdiploid cell line
was characterized by unbalanced translocation suspected to occur between the short arm
of chromosome 2 and the distal long arm of chromosome 18, trisomy 5 and 9, partially
trisomy for distal 2p, and tetrasomy for 8q with a modal number of 51 chromosomes
(Figure 22B and 22C). The karyotype of USC-HN2 contains characteristic features of
HNSCC, including isochromosome formation with resultant loss/deletion of the short
arm of chromosome 8, and breakpoints at or near the centromeres. Aneusomy is also a
recurrent finding in cytogenetically abnormal head and neck tumor specimens [73].
106
Figure 22. Morphologic and cytogenetic analysis of USC-HN2 cell line. A, Phase-contrast
photomicrographs (top) and Wright-Giemsa stained cytospins (bottom) of USC-HN2, USC-HN1, and
SCCL-MT1 cells (cytospin, Wright-Giemsa stain, x100 original magnification). All cell lines demonstrate
squamous cell morphology with varied numbers of mitotic cells (rounded, luminescent cells). B, The
karyotype of USC-HN2 contains characteristic features of HNSCC, including (C) isochromosome
formation with resultant loss/deletion of the short arm of chromosome 8. A modal number of 50
chromosomes, including 2 marker chromosomes are seen in this sample karyotype.
107
Figure 22, continued.
Phenotype of USC-HN2 cell line and heterotransplants closely resemble the original
tumor biopsy
Immunophenotypic characterization of USC-HN2 cells in culture and from tumors grown
in Nude mice demonstrated similarity to the original tumor and confirmed a keratinizing
squamous cell carcinoma (Figure 23). Neither the original tumor nor USC-HN2 cell line
expressed CD45, S100, or vimentin, consistent with its epithelial origin. USC-HN2 cells
demonstrate positive expression of keratin, FABP5, E-cadherin, and CD44, as well as
108
strong nuclear Rb and p53 expression in situ, consistent with HNSCC and the original
tumor biopsy [73,76,79-81]. Epidermal growth factor receptor (EGFR) and CD44
staining was increased in the cytospin and heterotransplant samples in comparison with
the original tumor biopsy. Flow cytometry studies were completed to characterize the
phenotype of USC-HN2 compared with USC-HN1 and SCCL-MT1 (Table 2).
Compared to isotype controls, all cell lines displayed positive staining for HNSCC
biomarkers EGFR, CD24, and E-cadherin. USC-HN2 and SCCL-MT1 also showed
increased staining for CD44v6, whereas staining for CD74, CD133 and IL-13R α2 was
negative in all three cell lines [76,77,88,89]. Expression of stem cell-associated
transcription factors c-Kit, Nanog, OCT3/4, and SOX2 was measured, and with the
exception of positive staining for c-Kit in SCCL-MT1, these factors were not detected
(data not shown) [96,97].
109
Figure 23. Characterization of the original tumor biopsy, USC-HN2 cell line, and heterotransplanted
tumor. Photomicrograph of immunoperoxidase staining of original tumor biopsy (left panels), USC-HN2
cells from culture in a cytospin preparation (middle panels), and formalin-fixed paraffin-embedded tissue
sections of USC-HN2 Nude mouse subcutaneous heterotransplant (right panels) for CD45, S100,
Vimentin, p53, Rb, EGFR, FABP5, E-cadherin, CD44, and Keratin (x400 original magnification).
110
USC-HN2 has increased expression of immune modulatory cytokines
The expression of pertinent oncogenes and cytokines was examined for USC-HN2, USC-
HN1, and SCCL-MT1 using qRT-PCR techniques. USC-HN2 showed a statistically
significant increase in mean expression of immune modulatory cytokines IL-1β, IL-6,
and IL-8 as compared to universal human reference RNA (Figure 24A, p<0.0005), which
was confirmed at the protein level by ELISA techniques (Figure 24B, p<0.05). IL-8
mRNA expression was found to be significantly increased in USC-HN2 and SCCL-MT1,
but not USC-HN1, while protein secretion of IL-8 was elevated in all three cell lines. All
three cell lines demonstrated significant protein secretion of GM-CSF and VEGF, though
mRNA expression was not significantly increased for these genes. USC-HN2 also had
increased TNF α protein levels compared with USC-HN1 and SCCL-MT1. The overall
expression profile of USC-HN2 is highly immune modulatory and more closely
resembles that of SCCL-MT1 than of USC-HN1.
To elucidate further the functional implications of this highly immune modulatory
cytokine profile, the USC-HN2 cell line was assessed for its ability to induce Treg and
MDSC suppressor cell populations from healthy volunteer peripheral blood mononuclear
cells after one-week co-culture. Suppressive function of tumor-educated CD33
+
MDSC
or CD4
+
CD25
high
Treg cells was assessed by their ability to inhibit the proliferation of
fresh, autologous T cells stimulated with CD3/CD28 beads in vitro. USC-HN2 and
SCCL-MT1 induced strongly suppressive MDSC (Figure 24C) and weakly suppressive
111
Treg cells (data not shown), consistent with previous reports that demonstrate HNSCC to
be highly immune modulatory in patients [28,81,82,95].
% Positive MFI
Target
Isotype
Control Antibody
Isotype
Control Antibody
USC-HN2
CD24 0.90 76.11 56.76 609.77**
E-cadherin 0.90 35.81 56.76 303.60**
EGFR 0.72 92.84 21.38 479.34**
CD44v6 0.90 7.75 56.76 152.86*
CD74 0.90 0.49 56.76 41.59
CD133 0.79 0.61 32.68 26.84
IL-13R α2 0.38 0.24 19.23 12.15
USC-HN1
CD24 0.92 29.34 32.75 92.92**
E-cadherin 0.92 8.01 32.75 109.56*
EGFR 0.89 94.74 8.64 341.99*
CD44v6 0.92 3.50 32.75 49.37
CD74 0.92 0.73 32.75 26.64
CD133 0.99 11.33 24.64 56.04
IL-13R α2 1.61 1.07 58.62 6.43
SCCL-MT1
CD24 1.37 24.7 65.13 203.06**
E-cadherin 1.37 8.87 65.13 215.69**
EGFR 0.34 98.34 16.20 1392.73**
CD44v6 1.37 6.03 65.13 133.36*
CD74 1.37 0.61 65.13 49.12
CD133 1.32 0.98 31.02 27.16
IL-13R α2 1.04 0.27 24.13 13.44
* MFI 50-100 above isotype control
** MFI >100 above isotype control
Table 2. Analysis of USC-HN2 surface markers by flow cytometry. Flow cytometry studies of USC-
HN2, USC-HN1, and SCCL-MT1 demonstrate surface markers characteristic of HNSCC cell lines.
Percent of positive staining cells (middle columns) and mean fluorescence intensity (MFI, right columns)
are shown for each antibody target and isotype control. Positive findings are shown in bold.
112
Figure 24. USC-HN2 is
highly immunomodulatory
and induces suppressor
cells. A, qRT-PCR analysis
of cytokine mRNA levels in
USC-HN2 compared with
established HNSCC cell
lines. Cytokine levels are
displayed as the fold change
in gene expression of each
cell line as compared with
human reference RNA levels.
USC-HN2 and SCCL-MT1
both showed increased
expression of IL-1 β, IL-6, IL-
8, and COX2, while USC-
HN1 had decreased cytokine
expression. B, Secreted
protein levels measured by
ELISA confirmed similar,
highly immunomodulatory
cytokine profiles for USC-
HN2 and SCCL-MT1. C,
USC-HN2 and SCCL-MT1
induced strongly suppressive
MDSC after one-week co-
culture with healthy donor
PBMC. For all samples mean
(n ≥2) data shown +SD;
*indicates p<0.05.
113
Microarray gene expression analysis
Microarray gene expression analyses results from USC-HN2, SCCL-MT1, and USC-
HN1 cell lines were compared with the data obtained from HNSCC tumor biopsy
samples previously reported [79]. A total of 243 genes were significantly differentially
expressed in both USC-HN2 and SCCL-MT1 cell lines. A comparison of gene
expression among USC-HN2, SCCL-MT1, and USC-HN1 cell lines identified many
common differentially expressed genes, consistent with their shared HNSCC origin.
Many of the up-regulated genes identified were also present in HNSCC tumor biopsies,
suggesting that USC-HN2 has an expression profile typical of HNSCC (Table 3). GSEA
(gene set enrichment analysis, [98]) was used to elucidate the differences between cell
lines. These analyses showed that USC-HN1 had significant down-regulation of genes
related to antigen processing and presentation pathways as compared with USC-HN2 and
SCCL-MT1, which demonstrate up-regulation of these genes (Table 4).
114
GeneBank Access Gene Symbol (Annotation) Log
2
Ratio
Immune Response
NM_002117 HLA-C (major histocompatibility complex, class I C) 2.6
NM_004048 B2M (beta-2 microglobulin) 2.1
NM_005514 HLA-B (major histocompatibility complex, class I B) 1.8
NM_002116 HLA-A (major histocompatibility complex, class I A) 1.7
NM_013230 CD24 (CD24 antigen) 1.3
Cell Growth, Maintenance/Cell cycle Regulation
NM_000424 KRT5, keratin 5 2.9
NM_000526 KRT14, keratin 14 2.0
NM_033666 ITGB1, integrin, beta 1 2.0
NM_002273 KRT8, keratin 8 1.5
NM_006088 TUBB2C, tubulin beta 2C 1.5
NM_006082 TUBA1B, tubulin alpha 1b 1.4
NM_005507 CFL1, cofilin 1 1.3
NM_002628 PFN2, profilin 2 1.3
NM_005022 PFN1, profilin 1 1.0
NM_004360 CDH1, E-cadherin 1.0
Translation and Protein Synthesis
NM_000971 RPL7, ribosomal protein L7 1.7
NM_006013 RPL10, ribosomal protein L10 1.4
NM_000979 RPL18, ribosomal protein L18 1.2
NM_001042559 EIF4G2, translation initiation factor 4 gamma 2 1.2
NM_001006 RPS 3A, ribosomal protein S3A 1.2
Metabolism
NM_001135700 YWHAZ 2.5
NM_002808 PSMD2, proteasome 26S subunit 1.8
NM_002794 PSMB2, proteasome subunit beta 2 1.6
NM_021130 PPIA, peptidylprolyl isomerase A (cyclophilin A) 1.5
NM_005561 LAMP1, lysosomeal-associated membrane protein 1 1.4
NM_001165415 LDHA, lactate dehydrogenase A 1.4
NM_005348 HSP90AA1, heat shock 90kDa alpha class A member 1 1.4
NM_001689 ATP5G3, ATP synthase H+ transporting subunit 1.0
NM_002715 PPP2CA, protein phosphatase 2 catalytic subunit 1.0
Others
NM_005978 S100A2, S100 calcium binding protein A2 2.8
NM_005953 MT2A, metallothionein 2A 2.6
NM_003329 TXN, thioredoxin 2.3
NM_006096 NDRG1, N-myc downstream regulated 1 2.2
NM_021103 TMSB10, thymosin, beta 10 1.9
NM_021009 UBC, ubiquitin C 1.7
NM_199185 NPM1, nucleophosmin 1.6
NM_001428 ENO1, enolase 1 1.2
Table 3. Selected up-regulated genes identified in USC-HN2 and SCCL-MT1 cell lines also present in
HNSCC tumor biopsies. Log
2
ratio of 1 signifies a 2-fold difference in the mean gene expression of the
cell line versus human reference RNA (p<0.05).
115
RefSeq Gene
USC-HN2
Log
2
Ratio
USC-HN1
Log
2
Ratio
SCCL-MT1
Log
2
Ratio
NM_002116 HLA-A 2.55226 1.2207 0.93383
NM_005514 HLA-B 2.12008 -0.51712 1.57735
NM_002117 HLA-C 3.19975 -0.67547 2.09203
NM_002127 HLA-G 2.12753 -0.746 1.25325
NM_001434 HLA-H 2.88822 -0.41665 2.05369
NM_002127 HLA-J 2.35369 -0.04189 1.32591
NR_027822 HLA-L 2.73574 -0.22241 1.64753
Table 4. GSEA identified up-regulated genes in both USC-HN2 and SCCL-MT1 cells compared with
USC-HN1 cells. Log
2
ratio of 1 signifies a 2-fold difference in the mean gene expression of the cell line
versus human reference RNA (p<0.05).
Figure 25. In situ Hybridization for the presence of Human Papillomavirus Type 16 and 18. Single
color FISH using an HPV16/18 probe demonstrates the HPV
-
status of USC-HN2, USC-HN1, and SCCL-
MT1 cell lines as compared with the HPV
+
control cell line HeLa.
116
Viral Screen and TP53 mutation analysis
All three cell lines, as well as the original tumor tissue from USC-HN1 and USC-HN2
(SCCL-MT1 original tumor not available) were screened for HPV by in situ
hybridization (Figure 25). Consistent with the oral cavity origin of all these cell lines, no
evidence of HPV 16 or 18 was found [75,78]. DNA from the each of the cell lines was
also screened for TP53 mutations, which are found in approximately half of all HNSCC
tumors and are typically absent in HPV
+
samples [73,78]. TP53 mutations were
identified in both USC-HN1 and SCCL-MT1; however no mutation was found USC-
HN2 (data not shown).
DISCUSSION
In this report, we describe the establishment and characterization of USC-HN2, a novel
cell line derived from a patient with recurrent, invasive HPV
-
buccal SCC with a past
medical history significant for a 50-pack-year history of tobacco smoking and no pre-
operative chemotherapy or radiation therapy. USC-HN2 cultured cells and
heterotransplanted tumors closely resembled the original tumor biopsy specimen with
respect to HNSCC-associated markers (keratin, E-cadherin, FABP5), HPV infection, and
cytogenetic abnormalities. One difference noted was the outgrowth of a highly
proliferative, EGFR
+
subclone from a largely EGFR
-
original tumor during establishment
of the cell line to yield an EGFR
+
cell line. Overall, the USC-HN2 cell line appeared
typical of HNSCC and showed similar morphology, growth rate, phenotype, and tumor
suppressor and oncogene expression to previously established HNSCC cell lines USC-
117
HN1 and SCCL-MT1. Despite these similarities, striking differences in the immune
modulatory capabilities of USC-HN2, SCCL-MT1, and USC-HN1 models were
discovered.
Immune evasion and suppression are two mechanisms by which tumors escape immune
destruction and evidence exists for the employment of both by HNSCC tumors [4,6,84].
The results of this study revealed USC-HN1 to be a poorly antigenic tumor model with
little to no capacity to induce immune tolerance while USC-HN2 and SCCL-MT1
appeared to be highly immunogenic tumor models with strong immune suppression
capacity. USC-HN1 cells showed significant down-regulation of antigen processing and
presentation pathways compared with human reference RNA and USC-HN2 and SCCL-
MT1 cell lines. Additionally, the USC-HN2 cultured cells and heterotransplants, as well
as the SCCL-MT1 cells, showed strong positivity for the head and neck cancer stem cell
marker CD44v6, while USC-HN1 demonstrated lower levels of expression. Cancer stem
cell populations within tumors are reported to have greater expression of immunogenic
tumor-associated antigens [99,100], a hypothesis that was supported here by microarray
data demonstrating significant up-regulation of antigen-presentation-related genes in
USC-HN2 and SCCL-MT1 but not USC-HN1. In order for immunogenic tumor cells to
persist in the face of infiltrating host immune cells, they must adapt to acquire
immunosuppressive capabilities, as demonstrated by USC-HN2 and SCCL-MT1 cell
lines. Tumors can also elicit immune suppression either directly, through the release of
soluble factors, or indirectly through the induction of suppressor cells, a mechanism that
118
has not been well characterized in HNSCC [84]. USC-HN2 and SCCL-MT1 were found
to have elevated expression of inflammatory and Th2 cytokines IL-1β, IL-6, IL-8, GM-
CSF and VEGF, while USC-HN1 demonstrated increased expression of only IL-8 and
VEGF. Previously, we have identified IL-1β, IL-6, and GM-CSF as key factors for the
induction of myeloid-derived suppressor cells, a population of innate immune suppressor
cells that mediate direct suppression of effector T cells and expand regulatory T cell
populations [28]. Indeed, co-culture of USC-HN2 and SCCL-MT1 but not USC-HN1
with normal healthy donor peripheral blood cells generated functionally suppressive
MDSC in vitro. Of note, these three cell lines were all found to be HPV
-
, which does not
suggest HPV infection as a major causative factor for the observed differences, but
additional studies in HNSCC cancer patients are needed before HPV can be excluded as
an influence on immune escape mechanisms.
Previously, we showed in breast and colorectal cancer patients that different tumors may
elicit tolerance and immune dysfunction via distinct mechanisms [5]. We now extend
this hypothesis to HNSCC cancer and suggest that two divergent mechanisms for
immune escape are evident in HNSCC, as modeled by the recently characterized cell
lines USC-HN2, USC-HN1, and SCCL-MT1. USC-HN2 and SCCL-MT1 appear to
model highly immunogenic cancers with robust cytokine production and strong induction
of suppressor cell populations, in contrast to USC-HN1, which demonstrates a more
benign cytokine profile, lack of suppressor cell induction, and down-regulation of MHC
class I molecules. Immunotherapy seeks to overcome tumor-mediated immune
119
dysfunction and activate a cell-mediated immune response against cancer cells. Such an
approach holds great promise for reducing damage to collateral tissue by taking
advantage of the inherent specificity of the human immune system. Systemic trafficking
and monitoring by immune cells also provides for superior treatment of metastatic and
inoperable lesions compared with external beam irradiation and surgical therapies.
Perhaps most importantly, the generation of immunologic memory following a robust
anti-tumor immune response prevents the recurrence of tumors. Immune evasion and
suppression are two mechanisms by which tumors escape immune destruction and
evidence exists for the employment of both by HNSCC tumors [84]. Whether these
different paths are taken by different tumors or different subclones within a single tumor
is unknown, but effective therapy for HNSCC will necessarily need to address both
mechanisms. Collectively these well-characterized cell lines provide a valuable spectrum
of HNSCC models and may aid in the development of new immunotherapies for these
difficult to treat tumors. Based upon our results, USC-HN1 could facilitate the
development of cancer vaccine and immune activation therapies that seek to aid tumor
cell recognition and targeting by the immune system while USC-HN2 and SCCL-MT1
provide excellent models for the development of suppressor cell-targeted therapies.
120
CHAPTER 6.
CONCLUSIONS AND FUTURE DIRECTIONS
Our investigation of human myeloid-derived suppressor cells was the natural culmination
of many past studies to characterize, understand, and ultimately develop effective
therapies to treat the immune dysfunctions seen in cancer patients. While our results
have at times left us perplexed, I have gained great admiration for the complexity and
nuance of the human immune system. We have learned a great deal about this newly
discovered population and its role in mediating immune tolerance, and I summarize the
most salient points in the following pages.
Earlier work in the laboratory hinted at the diverse ways by which tumors achieve
immune escape [5]. This finding helped to explain the variable therapeutic results
reported in clinical trials and experimental tumor models for a given protocol.
Importantly, some features of tumor immune escape emphasized the recurrent theme,
albeit reached along different paths among all cancers, of deficient cell-mediated
immunity to tumor antigens and the presence of immune suppression. Furthermore, the
immune suppression present in tumors appeared nearly always to include the recruitment
and activation of immune suppressor cells. This fact then makes characterization of these
cells extremely important as we seek to develop more effective immunotherapy for
cancer patients.
121
While other researchers sought to unravel MDSC in the murine system, we focused on
understanding human MDSC.This was a decision governed by clinical trials history that
emphasized mice were not small humans (e.g. murine Tregs are now known to have
significant differences from human Tregs). Indeed, one of two defining markers for
murine MDSC, namely Gr-1, has no known analog in humans. The decision to study
human MDSC was driven also by the desire to advance a field being avoided by many
due to the challenges of obtaining cells for study and determining the function of a cell
population with no definitive phenotype.
An initial question we sought to answer was whether human MDSC were even induced
as a regular feature of human cancers. The literature contained sporadic reports of
MDSC accumulation in a variety of cancer patients, but these were based upon poorly
defined surface markers rather than functional studies of suppression. During the course
of this project such clinical reports have become more frequent. We now understand
from our own screen of over 100 human cancer cell lines that MDSC induction appears
to be a universal feature of human cancers. Thus we can now state that MDSC are a
unique suppressor cell population in that they consistently accumulate in many types of
cancer but are rare or absent in healthy individuals. This feature provided the idea to
translate the presence of MDSC in patients into a clinical assay for cancer detection.
122
MDSC as a universal test for cancer detection
With improvements in living conditions and medical care, life expectancies are
increasing worldwide. This has translated to an aging population in which more
individuals are being diagnosed and living with cancer,often now as a chronic disease.
Collectively these trends create an increased need for tumor surveillance and long-term
cancer management. Given the significant economic impact of this disease, investment in
more effective clinical detection and therapies is warranted.
Through partnership with the USC Stevens Institute for Innovation, we are developing a
novel, minimally-invasive and universal clinical test for cancer using the presence of
MDSC in the peripheral blood of cancer patients. While MDSC are present in the tumors
as well as peripheral blood, and secondary lymphoid tissues of cancer patients, they are
rare or absent in the blood and tissues of healthy individuals [13,17]. As demonstrated by
our screen of more than 100 different human cancer cell lines of varied histologic types
(Chapter 2), the induction of CD33
+
or CD11b
+
MDSC appears to be a universal and
mechanistically meaningful feature of human cancers. Furthermore the extent of MSC
accumulation has been shown to correlate directly with tumor stage [21,39]. We propose
that the presence of MDSC in the peripheral blood of cancer patients can be used as a
highly specific clinical test for the presence and even the severity of neoplastic diseases
(Figure 26).
123
Previously, using a constellation of seven markers [Lineage
-
(CD3
-
, CD14
-
, CD19
-
,
CD57
-
), HLA-DR
-,
CD11b
+
, and CD33
+
] and functional studies to detect MDSC in the
blood of cancer patients and healthy controls, Diaz-Montero et al. [39] established that
MDSC accumulate in the peripheral blood of cancer patients. They also demonstrated
that there is a statistically significant difference among early stage (1.96% of blood
leukocytes, 124/ μL of whole blood) and late stage (3.77%, 260 cells/ μL) patients, and
that patients with highly metastatic disease had the greatest accumulation of these cells
(4.37%, 326 cells/ μL) [39]. Additionally, increased MDSC numbers were detected in
patients found to have radiographic evidence of disease progression and decreased
numbers were observed with radiographic evidence of disease response to therapy and
regression [39]. While the use of seven markers and functional studies is not feasible in
the routine clinical laboratory setting, this study established that in the peripheral blood of
cancer patients, MDSC are present in quantifiable amounts and changes in MDSC levels
are detectable using routine flow cytometry.
To use MDSC numbers as a marker for tumor burden in patients, a simplified phenotype
of MDSC was needed that distinguished these pathologic cells from normal myeloid cells
present in blood, and that correlated with active suppressive function. In developing an
in vitro model to study the induction of human MDSC from normal donor PBMC using
cancer cell lines, we generated a highly enriched population of human MDSC in
sufficient numbers to perform characterization studies. Using tumor cell-line generated
124
Figure 26. Schematic showing a novel, minimally-invasive clinical assay for cancer detection and
monitoring. Patient peripheral blood cells are analyzed by routine flow cytometry for the presence of
myeloid suppressor cells (MSC) as a marker for tumor presence. Active MSC are distinguished from
normal blood cells by a unique 3-marker phenotype that correlates directly with suppressive function.
Accumulation of active MSC correlates directly with disease stage and tumor burden, allowing physicians
to track disease stage, tumor response to therapy, and tumor recurrence or progression by a simple blood
test.
125
human MDSC, we described the morphology, phenotype, gene expression, and
suppressive function of this newly characterized cell population, including a unique set of
biomarkers expressed by this suppressive human MDSC phenotype [CD33
+
HLA-DR
low
HIF1 
+
or CD11b
+
C/EBP 
+
(preferentially induced by breast carcinoma models)]. We
developed monoclonal antibodies against these antigen targets to create a clinical assay
that detects and quantifies MDSC in patient blood samples by flow cytometry as a
diagnostic and treatment monitoring tool. Preliminary results from a pilot study of this
MDSC Clinical Assay in patients with head and neck squamous cell carcinoma (HNSCC)
clearly documented the greater accumulation of CD33
+
HLA-DR
low
HIF1 α
+
MDSC in the
peripheral blood of cancer patients compared with healthy volunteer donors (Figure 27).
Using these ongoing pilot studies in head and neck cancer patients, assay sensitivity and
specificity are being optimized to produce a robust, minimally invasive, and cost-
effective test that can be used routinely in hospitals. As our data collection proceeds, it
appears that HIF1 α and CD33 may be sufficient as markers for these cells in cancer
patient PBMC samples. Ongoing studies are evaluating this clinical assay as a method
for physicians and surgeons to measure tumor response to therapy and to detect disease
recurrence or metastasis during long term follow-up and monitoring of cancer patients.
Planned future clinical studies will evaluate the same assay as a means to distinguish
benign from malignant nodules, as in those patients with thyroid or breast nodules for
whom routine imaging or fine needle aspiration biopsy results are inconclusive.
126
Existing paradigms for cancer detection and monitoring tend to envision the disease as a
diverse collection of neoplasms that arise from varied tissues in the body. Those
approaches therefore use different biomarkers and assays for each tumor type and often
require that tumor cells be significantly mutated to be distinguishable from their normal
tissue of origin. In contrast to such thinking, we propose viewing cancer as a single
disease in which neoplastic cells progressively grow and invade normal tissue through
their induction of immune tolerance, at least locally and perhaps systemically. In our
paradigm, immune evasion and tolerization are common characteristics of malignant
tumors that can be quantified as a direct marker of tumor presence, mass, and perhaps
aggressiveness. Unlike existing tests to screen for cancer presence, such as
mammograms, colonoscopies, or whole body PET scans, the proposed assay is minimally
invasive, utilizes routine clinical laboratory techniques, and is applicable to cancers of
any histologic origin. As a universal test for cancer identification, this assay could aid
physicians in the early detection of cancer, distinguish benign from malignant tumors,
monitor cancer response to therapy, detect the presence of occult metastatic disease, and
identify early recurrences.
127
Figure 27. Preliminary data demonstrating MDSC in the peripheral blood of cancer patients using a
recently identified phenotype: CD33
+
HLA-DR
low
HIF1 
+
. Twenty milliliters of peripheral blood was
collected from normal, healthy volunteers or HNSCC cancer patients and PBMC were isolated by density
gradient separation. PBMC were stained for CD33
+
and HLA-DR
+
using fluorescence-labeled monoclonal
antibodies, then cells were fixed and permeabilized for intracellular staining of HIF1  by a third antibody.
Stained sample PBMC and isotype controls were analyzed on a FACSCalibur flow cytometry using
CellQuestPro software and collecting 50,000 live leukocyte events. CD33
+
HLA-DR
low
HIF1 
+
cells were
found to be 15.78-16.23% of myeloid cells in cancer patients (A) compared with 0.12-1.99% in healthy
controls (B).
128
Figure 27, continued.
129
MDSC-targeted immunotherapy
The apparently universal induction of immune suppressor cells by cancer is a double-
edged sword. The presence of MDSC in cancer patients is an early warning sign for
disease, but is also a major impediment to generating a strong anti-tumor immune
response for therapy. Surgery, radiation therapy, and chemotherapy for solid tumors are
largely ineffective against metastatic disease and the elimination of minimal residual
disease. Like other investigators, we think that the human immune system holds
unrealized potential as a resource for cancer therapy and has the added value of built-in
memory required for sustained cure.
An effective immunotherapy of cancer must achieve antigen-specific immune activation
accompanied by the reversal of tumor-driven immune suppression. Immune activation
may be achieved through a variety of approaches such as tumor lysate or DC vaccination,
adoptive transfer of transgenic T cells, and infusion of stimulatory immunoligands and
cytokines. Many currently approved immunotherapies for cancer rely upon such immune
stimulation, but their clinical successes have been limited by lagging methods to reverse
tumor immune tolerance, the second half of any effective immunotherapy [6]. Immune
suppressor cells are now recognized as a key component of tumor immune tolerance and
also as a major impediment to successful immunotherapy in cancer patients and
experimental solid tumor models [3,4,6]. Regulatory T cells (Treg) were quickly
recognized as an important population mediating tolerance in the tumor
microenvironment. Several methods for the elimination of Treg in cancer patients have
been described, including selective depletion by cyc
antibodies against CD25 (PC61/Zenapax)
excitement when first introduced
immune stimulation, when paired with Treg depletion, would yield tumor regression and
perhaps even tumor cure. In our own laboratory, such therapy indeed did lead to cure of
Colon 26 tumors in mice (Figure 28
Figure 28. Immunotherapy of Colon 26 tumor model shows cure when immune stimulation is paired
with regulatory T cell depletion.
rejected in mice treated with LEC/chTNT
cells into the tumor microenvironment, and Treg depletion. Tumors in mice that rece
LEC/chTNT-3 or Treg depletion did not show significant inhibition of growth or regression compared with
tumors in untreated mice.
been described, including selective depletion by cyclophosphamide and monoclonal
antibodies against CD25 (PC61/Zenapax) [11]. These drugs generated
first introduced and it was expected that cancer vaccines or other
immune stimulation, when paired with Treg depletion, would yield tumor regression and
perhaps even tumor cure. In our own laboratory, such therapy indeed did lead to cure of
26 tumors in mice (Figure 28) [101].
. Immunotherapy of Colon 26 tumor model shows cure when immune stimulation is paired
with regulatory T cell depletion. Established Colon 26 tumors grown in BALB/c mice were successfully
rejected in mice treated with LEC/chTNT-3, a tumor-targeted chemokine stimulating infiltation of immune
cells into the tumor microenvironment, and Treg depletion. Tumors in mice that rece
3 or Treg depletion did not show significant inhibition of growth or regression compared with
130
ophosphamide and monoclonal
. These drugs generated considerable
and it was expected that cancer vaccines or other
immune stimulation, when paired with Treg depletion, would yield tumor regression and
perhaps even tumor cure. In our own laboratory, such therapy indeed did lead to cure of
. Immunotherapy of Colon 26 tumor model shows cure when immune stimulation is paired
Established Colon 26 tumors grown in BALB/c mice were successfully
targeted chemokine stimulating infiltation of immune
cells into the tumor microenvironment, and Treg depletion. Tumors in mice that received only
3 or Treg depletion did not show significant inhibition of growth or regression compared with
131
Despite such success, in many other tumors even the combination immune stimulation
and regulatory T cell depletion/inhibition failed to produce tumor regression. At this
juncture, proponents of immunotherapy began to look for additional tumor-driven
immune problems, and soon thereafter myeloid-derived suppressor cells (MDSC)
emerged as possible culprits. In recent years, we and others have studied these cells in
hopes of finding the remaining key suppressor population in cancer. We were
unprepared for the diversity and complexity governing these cells, which appear to have a
key role in trauma, sepsis, and autoimmune disease in addition to their role in many
forms of cancer.
Murine breast carcinoma 4T1 was one such model that did not respond to immune
stimulation and Treg depletion, and so we used it to investigate the possibility of an
MDSC-driven component to tumor immune tolerance. Immune profile studies of 4T1
revealed tumor infiltrating T cells and innate effector cells that were unable to effectively
kill the tumor cells (Figure 29). These results suggested an immunogenic tumor with
strong induction potential of suppressor cells and other immunosuppressive mechanisms.
Indeed, immunohistochemistry and flow cytometry studies of 4T1 tumor-bearing mice
showed frequent MDSC (Figure 29) in addition to Treg, providing for the first time a
plausible explanation for the failure of immune stimulation with Treg-inhibition therapy
to halt tumor progression.
132
Figure 29. Immunohistochem-
istry studies of tumor-
infiltrating leukocytes in 4T1
breast carcinoma murine
tumor model. Expression of
MDSC-associated molecules
CD11b and ARG-1 was frequent
in this tumor model, as were
markers of activated T cells
(CD4
+
CD25
+
) and Treg cells
(CD4
+
CD25
+
FoxP3
+
) (FoxP3
data not shown). This immune
profile is suggestive of an
immunogenic tumor model in
which activated effector immune
cells infiltrate into the tumor
microenvironment but are
inhibited from acting by immune
suppressor cells (MDSC, Treg)
also present.
Once MDSC were recognized as contributors to tumor immune tolerance, a next step was
developing MDSC-targeted immunotherapy reagents. Compared with Treg-inhibition,
effective methods for the inhibition of MDSC, which have only recently been recognized
as major mediators of tumor tolerance and progression, are less certain. Previously
described MDSC-targeted therapies, summarized in Table 5, acted through selective
depletion of MDSC (5-fluorouracil, gemcitabine, docetaxel, sunitinib), inhibition of
signaling pathways (sunitinib, GW2580, amiloride), or inhibition of suppressive
mechanisms (celecoxib, sildenafil, ATRA) [23,35,49,102-111]. Unfortunately, with the
exception of sunitinib, none of these reagents has been tested or shown efficacy in
humans. Ko et al. [35] first demonstrated decreased MDSC accumulation in renal cell
carcinoma patients after treatment with sunitinib, a tyrosine kinase inhibitor with
selective action on the JAK/STAT3 signaling pathway. Subsequent research has shown
133
STAT3 signaling to be a key mediator of suppressor cell function, and a more recent
study demonstrated partial tumor regression when dendritic cell vaccination was
combined with sunitinib therapy in mice [110]. Sunitinib has become a boon to cancer
therapy, but has significant side effects that limit its use clinically, including left
ventricular myopathy and hypertension, mucositis/stomatitis, and myelosuppression. In
addition, Ko et al. [35] found no tumor regression in renal cell carcinoma patients treated
with sunitinib, despite decreased accumulation and selective depletion of MDSC.
Identifying better immunotherapy protocols to inhibit human MDSC expansion should be
a focus of future research. Our characterization studies of human MDSC also highlighted
the canonical transcription factors, hypoxia inducible factor (HIF)1  and C/EBP , as
potential therapeutic targets in human MDSC in addition to STAT3. These and other
therapies from murine MDSC research should be further investigated using in vivo
experimental tumor models and cancer patients.
134
Drug Tumor Model Therapeutic effect
All-trans retinoic acid
(ATRA)
Vitamin A derivative
[49]
CT-26 colon carcinoma, EL4
thymoma, and MC38 colon
carinoma murine tumor
models; Renal cell carcinoma
patients
Differentiation of MDSC into mature
myeloid cells via neutralization of ROS by
GSH
Amiloride
K
+
-sparing diuretic [102]
EL4 thymoma, CT-26 colon
carcinoma, and TS/A
mammary carcinoma murine
tumor models; H23 lung
adenocarinoma human cancer
cell line
Inhibition of tumor-derived exosome-
associated Hsp72 triggered STAT3
activation in MDSC by inhibition of
exosome formation
Celecoxib
COX2 inhibitor [103]
AB1 mesothelioma murine
tumor model
Decreased accumulation of MDSC in
spleens of tumor-bearing mice and
decreased ROS production by granulocytic
MDSC
CpG ODNs
Toll-like receptor 9
agonist [104]
CT-26 colon carcinoma and
CEA242-Tag murine tumor
models
Decreased suppressive function and
maturation of Ly6G
high
MDSC
Docetaxel
Antimicrotubule
chemotherapeutic
[105]
4T1-Neu mammary
carcinoma tumor model
Polarization of MDSC toward a type 1
macrophage (M1) phenotype, selective
depletion of type 2 (mannose receptor
+
)
MDSC over M1 cells, and inhibition of
STAT3
5-Fluorouracil
Pyrimidine analog
chemotherapeutic [106]
EL4 thymoma murine tumor
model
Selective cytotoxic depletion of MDSC in
the tumor and secondary lymphoid organs
of tumor bearing mice
Gemcitabine
Pyrimidine analog
chemotherapeutic [107]
4T1 mammary carcinoma
murine tumor model
Selective cytotoxic depletion of MDSC in
the tumor and secondary lymphoid organs
of tumor bearing mice
GW2580
Inhibitor of CSF receptor
signaling [108]
3LL lung carcinoma, RM-1
prostate carcinoma, and
B16F1 melanoma tumor
models
Inhibition of CSFR1 signaling decreases
recruitment of TAM and monocytic MDSC
Sildenafil
Phosphodiesterase type 5
inhibitor [23,109]
CT-26 colon, 4T1 mammary,
and A20 B-cell lymphoma
murine tumor models
Down regulation of MDSC suppressive
marker IL-4R α
Sunitinib
Tyrosine kinase small
molecule inhibitor
[35,54,110]
Renal cell carcinoma patients
B16.Ova melanoma murine
tumor model
Decreased MDSC accumulation in cancer
patients; decreased viability and
suppressive function in vitro
Decreased MDSC accumulation in the
tumor microenvironment and improved
cancer vaccine efficacy
Triterpenoid CDDO-
Me
Antioxidant [111]
Renal cell carcinoma
patients; MC38 colon, LLC
lung, and EL-4 thymoma
murine tumor models
Decreased suppressive function (without
change in accumulation) by decreasing
reactive oxygen species
Table 5. Summary of current therapies for MDSC.
135
Depletion or inhibition of single suppressor cell populations (i.e. Treg or MDSC) appears
to have limited effects in eliciting anti-tumor immune responses in vivo due to redundant
tolerizing mechanisms and a lack of immune activation [12,35]. Optimal immunotherapy
is more likely to result from a decrease in suppressor cell accumulation and suppressive
function that coincides with augmentation of cell-mediated immunity. To test this
hypothesis, DC vaccination was evaluated in combination with dual MDSC and Treg
inhibition for the treatment of 4T1 breast carcinoma tumors in BALB/c mice. We
hypothesized that DC vaccination alone would be ineffective for therapy due to the
tolerizing effects of immune suppressor cells. In addition, due to the frequency of both
MDSC and Treg observed in the tumor microenvironment and draining lymphoid tissues
in 4T1 tumor-bearing mice (Figure 29), we hypothesized that inhibition of both
suppressor populations would be needed for complete reversal of tumor-driven immune
tolerance. For these studies MDSC were selectively depleted by the chemotherapeutic
drug 5-fluorouracil (Table 5) as described previously. In agreement with our hypothesis,
simultaneous depletion of MDSC (by 5-FU) and Treg (by cyclophosphamide) with tumor
DC vaccination halted the growth of established 4T1 tumors in a recent pilot study
(Figure 30). Despite these promising results, chemotherapy has significant collateral
toxicity to normal tissues and host defense systems, particularly neutrophils and other
rapidly dividing immune cells. Therefore, an immunotherapeutic approach that inhibits
MDSC function and possible matures them to antigen presenting cells would be
preferable to non-specific elimination of myeloid cells by chemotherapy. HIF1  was
identified as a mediator of suppressive function in murine and human MDSC [25;
136
Chapter 3]. There we hypothesized that HIF1  inhibition might be an alternative
mechanism to inactivate MDSC in 4T1 tumors. Indeed, preliminary results showed that
replacing 5-FU dosing with the HIF1  inhibitor CAY10585 (Cayman Chemical, Ann
Arbor, MI) in the treatment of 4T1 tumor-bearing mice produced a comparable halt in
progression when administered with DC vaccination and Treg depletion (Figure 31B).
Current immune monitoring studies are underway to expand our mechanistic
understanding of the observed treatment effects and determine whether HIF1  inhibition
is an effective means of MDSC inhibition in cancer.
137
Figure 30. Immunotherapy of 4T1 breast carcinoma in BALB/c mice with simulataneous Treg and
MDSC depletion. A, Dendritic cell vaccination with or without MDSC or Treg inhibition provided some
therapeutic benefit compared with untreated controls. However, DC vaccination concurrent with inhibition
of both MDSC and Treg populations was found to be superior and completely controlled tumor growth. B,
Similar effects were observed using alternative MDSC-targeted therapy HIF1 α inihibtor CAY10585. Mean
tumor volume shown (n=5) + SEM.
138
A working model for understanding human MDSC in the cancer setting
From our own data and other reports in the literature, we can propose a model by which
hypoxia and inflammatory signals within the tumor microenvironment induce
suppressive function in an expanding population of human MDSC (Figure 31A and B).
IL-1 and IL-6 in combination with GM-CSF appear to be necessary and sufficient for the
induction of functionally active human CD33
+
MDSC (Chapter 2, [28]). Other
inflammatory mediators also probably contribute to MDSC generation, given the
diversity of human cancers and the heterogeneous phenotypes reported for human
MDSC. We showed a role for TNF-  and VEGF in the induction of human CD33
+
MDSC in addition to IL-1, IL-6, and GM-CSF (Chapter 2, [28]). Gene expression
studies of MDSC-inducing human tumor cell lines suggested that IL-1  and COX-2-
derived prostaglandin E2 are inducers of suppressive myeloid cells. Analysis of changes
in immune cell populations caused by IL-1 or PGE
2
treatment of PBMC showed
expansion of CD66b
+
cells, suggesting that these cytokines may promote induction and
expansion of CD66b
+
MDSC, a subset reported by Ochoa and colleagues [36]. A
CD11b
+
subset of human MDSC was preferentially induced by breast carcinoma cell
lines. These tumor cell lines showed increased expression of putative MDSC-inducing
factors FLT3L and TGF  (Chapter 2) and additional studies are needed to more fully
describe the induction of this subset. It is now clear that human cancers have multiple
ways to induce and activate MDSC and that many of these pathways involve
inflammatory mediators.
139
Inflammation, necrosis, and hypoxia are common features of the neoplastic
microenvironment, particularly for tumors with aggressive growth and strong leukocyte
infiltration [7,26]. Inflammation is also a hallmark of systemic infection and trauma –
two other settings in which MDSC are known to accumulate [44]. Cellular regulators of
MDSC induction and function thus are likely to involve inflammatory (cytokine and toll-
like receptor) and oxygen-sensitive RE-DOX signaling pathways. Indeed, the up-
regulation and phosphorylation of signal tranducer and activator of transcription (STAT)s
1, 3, 5, and 6, common downstream mediators of cytokine receptor binding, have been
described in MDSC [14,35,54,62]. Nuclear factor  B (NF- B) signaling is also
upregulated by many inflammatory signals, most notably IL-1 , TNF- , and toll-like
receptor (TLR)4 agonist lipopolysaccharide endotoxin (LPS) derived from gram-negative
bacteria. C/EBP  is a transcription factor that works coordinately with NF- B to
regulate gene expression in inflammation [59] and recently has been implicated in
maintaining CD11b
+
murine MDSC [51]. Regulation of hypoxia signaling is complex in
immune cells and can be stimulated through oxygen-dependent or independent pathways
[57]. In both pathways, activation of HIF1-dependent signaling requires upregulation of
the hypoxia inducible factor (HIF)1  gene and cytoplasmic accumulation of HIF1 
protein. During normoxia, regular ubiquitination by the Von-Hippel Lindau tumor
suppressor protein and proteasomal degradation normally maintains low cytoplasmic
levels of HIF1  and prevents its significant dimerization with the constitutively
expressed binding partner HIF1  that is necessary for nuclear translocation. Interaction
with the VHL protein for degradation requires prolyl hydroxylation of HIF1 , and under
140
hypoxic conditions when oxygen is less available as a substrate for this reaction,
cytoplasmic HIF1  accumulates [57]. In addition, in specific cell types activation of
HIF1  signaling under normoxic conditions can be stimulated by cytokine and growth
factor signals through phosphatidylinositol 3-kinase or mitogen-activated protein kinase
pathways, as well as stabilization of HIF1  protein by NO-mediated S-nitrosylation
[57,58,112]. Recently, Corzo et al. [35] showed that HIF1 α accumulation in murine
myeloid cells promotes expression of MDSC suppressive gene products iNOS and ARG-
1. Doedens et al. [113] found that murine tumor-associated macrophage suppression of T
cells also was HIF1 -dependent. These studies suggest that future investigation of the
transcriptional regulators of human MDSC will be illuminating.
In this report we show strong upregulation of STAT3, HIF1 , and C/EBP  in human
MDSC, with some variability between CD33
+
and CD11b
+
subsets (Chapter 3). In both
subsets, expression of these transcription factors increased following LPS-induced
activation in vitro concurrent with upregulation of putative suppressive genes ARG-1,
iNOS, and the NADPH oxidase (NOX2) component, NCF1. These data are consistent
with STAT3, HIF1 , and C/EBP  as promoters of suppressive function in human
MDSC. Furthermore, therapeutic inhibition of MDSC with ATRA, Sunitinib, or
Celecoxib was accompanied by significant down-regulation of these transcription factors.
Interestingly, inhibition of suppressive function corresponded with down-regulation of
STAT3 and HIF1  in CD33
+
MDSC and C/EBP  in CD11b
+
MDSC suggesting that the
predominant drivers of suppressive function may be different in these subsets.
141
We propose that the inflammatory cytokines identified here, namely tumor-derived IL-
1β, IL-6, GM-CSF, TNF- α, and VEGF, in concert with tumor microenvironment hypoxia
lead to increased STAT3, NF-κB-C/EBP , and HIF1 α-dependent gene transcription to
produce the suppressive myeloid phenotype (Figure 31). While STAT3 and HIF1 
appear to be primary in CD33
+
MDSC, and C/EBP  in CD11b
+
MDSC, there is also
ample evidence for transactivation among these pathways in immune cells. Such
transactivation could amplify their separate induction effects and account for overlapping
functions of human MDSC subsets. Studies in myeloid cells suggest that HIF1 α and
NF B-C/EBP  signaling induce autocrine production of inflammatory mediators IL-1β,
IL-6, and TNF- α [26,55,56,60,112,114], which could serve to maintain existing MDSC
and induce new MDSC from myeloid progenitors. While this model remains
preliminary, it unites the germinal findings of this investigation with other published data
to provide a working hypothesis of the MDSC induction and activation steps that are
required for effective immunosuppression in the host.
142
Figure 31. Schematic for the induction of human CD33
+
and CD11b
+
MDSC in cancer. A, Hypoxia
and tumor-derived cytokines IL-1 β, IL-6, TNF α, VEGF, FLT3L, and TGF  in the tumor microenvironment
promote signaling through STAT3, NFκB/C/EBP, SMAD2/4, and HIF1α pathways in myeloid cells. In
addition to oxygen-dependent HIF1  regulation, inflammatory cytokines up-regulate HIF1  transcription
(via PI3K or MAPK) and NO stabilizes HIF1  protein (via S-nitrosylation). Other factors influencing
MDSC function include PBMC and tumor-derived GM-CSF, which supports expansion of myeloid
progenitors and survival of MDSC, and IFN , which contributes to MDSC activation. Transactivation (*)
between JAK/STAT, HIF1 , and NF B signaling pathways amplifies the induction effects of tumor-
derived cytokines and hypoxia in MDSC. B, Activated transcription factors translocate to the nucleus
where they up-regulate expression of suppressive genes (iNOS, NOX2, ARG-1, VEGF) and autocrine
production of putative MDSC inducers (e.g. IL-6, IL-1 , TNF , and VEGF). The transcription factors
driving suppressive function (and by extension potential therapeutic targets) in human MDSC appear to
vary by subset, with a dominant role for STAT3 and HIF1  in CD33
+
MDSC (purple) and a dominant role
for NF B-C/EBP  in CD11b
+
MDSC (pink).
143
Figure 31, continued.
144
BIBLIOGRAPHY
[61] Abbas AK and Lichtman AH. Basic Immunology Updated Edition: Functions
and Disorders of the Immune System, 3
rd
ed. China: Saunders, 2011.
[44] Adib-Conquy M and Cavaillon JM. Stress molecules in sepsis and systemic
inflammatory response syndrome. FEBS Letters. 2007, 581(19): 3723-3733.
[83] Albers AE, Strauss L, Liao T, Hoffmann TK, and Kaufmann AM. T cell-tumor
interaction directs the development of immunotherapies in head and neck cancer.
Clin Dev Immunol. 2010; 2010: 236378.
[2] American Cancer Society, Surveillance and Health Policy Research, 2010,
Cancer Facts and Figures 2010. Atlanta, GA: vAmerican Cancer Society, Inc,
2010.
[18] Bak SP, Alonso A, Turk MJ, and Berwin B. Murine ovarian cancer vascular
leukocytes require arginase-1 activity for T cell suppression. Mol Immunol. 2008,
46: 258-268.
[86] Bergmann C, Strauss L, Wang Y, Szczepanski MJ, Lang S, Johnson JT,
Whiteside TL. T regulatory type 1 cells in squamous cell carcinoma of the head and
neck: mechanisms of suppression and expansion in advanced disease Clinical
Cancer Research. 2008, 14: 3706-3715.
[10] Bluestone JA and Abbas AK. Natural versus adaptive regulatory T cells.
Nature Rev. Immunol. 2003, 3: 253–257.
[110] Bose A, Taylor JL, Alber S, Watkins SC, Garcia JA, Rini BI, Ko JS, Cohen
PA, Finke JH, and Storkus WJ. Sunitinib facilitates the activation and recruitment
of therapeutic anti-tumor immunity in concert with specific vaccination.
International Journal of Cancer. 2010 Dec 17. [Epub ahead of print] DOI:
10.1002/ijc.25863
[40] Brimnes MK, Vangsted AJ, Knudsen LM, Gimsing P, Gang AO, Johnsen HE,
and Svane IM. Increased level of both CD4+FOXP3+ regulatory T cells and
CD14+HLA-DR ⁻/low myeloid-derived suppressor cells and decreased level of
dendritic cells in patients with multiple myeloma. Scand J Immunol. 2010, 72(6):
540-547.
[52] Bunt SK, Clements VK, Hanson EM, Sinha P, and Ostrand-Rosenberg S.
Inflammation enhances myeloid-derived suppressor cell cross-talk by signaling
through Toll-like receptor 4. Journal of Leukocyte Biology. 2009, 85(6): 996-1004.
145
[102] Chalmin F, Ladoire S, Mignot G, Vincent J, Bruchard M, Remy-Martin JP,
Boireau W, Rouleau A, Simon B, Lanneau D, De Thonel A, Multhoff G, Hamman
A, Martin F, Chauffert B, Solary E, Zitvogel L, Garrido C, Ryffel B, Borg C,
Apetoh L, Rébé C, and Ghiringhelli F. Membrane-associated Hsp72 from tumor-
derived exosomes mediates STAT3-dependent immunosuppressive function of
mouse and human myeloid-derived suppressor cells. Journal of Clinical
Investigation. 2010, 120: 457-471.
[100] Chikamatsu K, Takahashi G, Sakakura K, Ferrone S, and Masuyama K.
Immunoregulatory properties of CD44+ cancer stem-like cells in squamous cell
carcinoma of the head and neck. Head Neck. 2010, 33(2): 208-215.
[97] Chiou SH, Yu CC, Huang CY, Lin SC, Liu CJ, Tsai TH, Chou SH, Chien CS,
Ku HH, Lo JF. Positive correlations of Oct-4 and Nanog in oral cancer stem-like
cells and high-grade oral squamous cell carcinoma. Clinical Cancer Research.
2008, 14: 4085-4095.
[63] Clare-Salzler MJ, Brooks J, Chai A, Van Herle K, and Anderson C. Prevention
of diabetes in nonobese diabetic mice by dendritic cell transfer. Journal of Clinical
Investigation. 1992, 90(3): 741-748.
[12] Colombo MP and Piconese S. Regulatory-T-cell inhibition versus depletion:
the right choice in cancer immunotherapy. Nat Rev Cancer. 2007, 7(11): 880-887.
[16] Condamine T and Gabrilovich DI. Molecular mechanisms regulating myeloid-
derived suppressor cell differentiation and function. Trends in Immunology. 2011,
32(1): 19-25.
[25] Corzo CA, Condamine T, Lu L, Cotter MJ, Youn JI, Cheng P, Cho HI, Celis E,
Quiceno DG, Padhya T, McCaffrey TV, McCaffrey JC, and Gabrilovich DI. HIF-
1α regulates function and differentiation of myeloid-derived suppressor cells in the
tumor microenvironment. Journal of Experimental Medicine. 2010, 207: 2439-
2453.
[20] Corzo CA, Cotter MJ, Cheng P, Cheng F, Kusmartsev S, Sotomayor E, Padhya
T, McCaffrey TV, McCaffrey JC, and Gabrilovich DI. Mechanism regulating
reactive oxygen species in tumor-induced myeloid-derived suppressor cells. Journal
of Immunology. 2009,182: 5693-5701.
[46] Cui W, Taub DD, and Gardner K. qPrimerDepot: a primer database for
quantitative real time PCR. Nucleic Acids Res. 2007, 35(Database issue): D805-
809.
146
[78] Dai M, Clifford GM, le Calvez F, Castellsagué X, Snijders PJ, Pawlita M,
Herrero R, Hainaut P, and Franceschi S. IARC Multicenter Oral Cancer Study
Group. Human papillomavirus type 16 and TP53 mutation in oral cancer: matched
analysis of the IARC multicenter study. Cancer Research. 2004, 64(2): 468-471.
[39] Diaz-Montero CM, Salem ML, Nishimura MI, Garrett-Mayer E, Cole DJ, and
Montero AJ. Increased circulating myeloid-derived suppressor cells correlate with
clinical cancer stage, metastatic tumor burden, and doxorubicin-cyclophosphamide
chemotherapy. Cancer Immunol Immunother. 2009, 58: 49-59.
[113] Doedens AL, Stockmann C, Rubinstein MP, Liao D, Zhang N, DeNardo DG,
Coussens LM, Karin M, Goldrath AW, and Johnson RS. Macrophage expression of
hypoxia-inducible factor-1 alpha suppresses T-cell function and promotes tumor
progression. Cancer Research. 2010, 70(19): 7465-7475.
[69] Dong RS Adams S, Bouma G , Eddouda A , Chana P, Moulding D, Duncan A,
and J Anderson. Newly identified multinuclear cells in cultured human dendritic
cells possess the phenotype of regulatory DC. In Molecular and Cellular Biology
and Immune Escape in Cancer, February 7–12, 2010. Keystone Symposia on
Molecular and Cellular Biology. Silverthorne, CO. p. 171.
[21] Donkor MK, Lahue E, Hoke TA, Shafer LR, Coskun U, Solheim JC, Gulen D,
Bishay J, and Talmadge JE. Mammary tumor heterogeneity in the expansion of
myeloid-derived suppressor cells. Int Immunopharmacol. 2009, 9: 937-948.
[24] Dumitriu IE, Dunbar DR, Howie SE, Sethi T, and Gregory CD. Human
dendritic cells produce TGF-beta 1 under the influence of lung carcinoma cells and
prime the differentiation of CD4+CD25+Foxp3+ regulatory T cells. Journal of
Immunology. 2009, 182: 2795-2807.
[26] Eltzschig HK and Carmeliet P. Hypoxia and inflammation. New England
Journal of Medicine. 2011, 364(7): 656-665.
[88] Erdem NF, Carlson ER, and Gerard DA. Characterization of gene expression
profiles of 3 different human oral squamous cell carcinoma cell lines with different
invasion and metastatic capacities. J Oral Maxillofac Surg. 2008, 66: 918-927.
[30] Filipazzi, P., R. Valenti, V. Huber, L. Pilla, P. Canese, M. Iero, C. Castelli, L.
Mariani, G. Parmiani, and L. Rivoltini. 2007. 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.
147
[17] Gabrilovich DI and Nagaraj S. Myeloid-derived suppressor cells as regulators
of the immune system. Nat Rev Immunol. 2009, 9: 162-174.
[75] Goon PK, Stanley MA, Ebmeyer J, Steinsträsser L, Upile T, Jerjes W, Bernal-
Sprekelsen M, Görner M, Sudhoff HH. HPV & head and neck cancer: a descriptive
update. Head Neck Oncol. 2009, 1: 36-43.
[38] Gordon IO and Freedman RS. Defective antitumor function of monocyte-
derived macrophages from epithelial ovarian cancer patients. Clinical Cancer
Research. 2006, 12: 1515-1524.
[79] Han J, Kioi M, Chu WS, Kasperbauer JL, Strome SE, and Puri RK.
Identification of potential therapeutic targets in human head & neck squamous cell
carcinoma. Head Neck Oncol. 2009, 1:27.
[82] Heo DS, Snyderman C, Gollin SM, Pan S, Walker E, Deka R, Barnes EL,
Johnson JT, Herberman RB, Whiteside TL. Biology, cytogenetics, and sensitivity to
immunological effector cells of new head and neck squamous cell carcinoma lines.
Cancer Research. 1989, 49: 5167-5178.
[50] Heppner GH. Tumor heterogeneity. Cancer Research. 1984, 44: 2259-2265.
[33] Hoechst B, Ormandy LA, Ballmaier M, Lehner F, Krüger C, Manns MP,
Greten TF, and Korangy F. A new population of myeloid-derived suppressor cells
in hepatocellular carcinoma patients induces CD4(+)CD25(+)Foxp3(+) T cells.
Gastroenterology. 2008, 135: 234-243.
[32] Hoechst B, Voigtlaender T, Ormandy L, Gamrekelashvili J, Zhao F,
Wedemeyer H, Lehner F, Manns MP, Greten TF, and Korangy F. Myeloid derived
suppressor cells inhibit natural killer cells in patients with hepatocellular carcinoma
via the NKp30 receptor. Hepatology. 2009, 50: 799-807.
[114] Huang WL, Yeh HH, Lin CC, Lai WW, Chang JY, Chang WT, Su WC.
Signal transducer and activator of transcription 3 activation up-regulates
interleukin-6 autocrine production: a biochemical and genetic study of established
cancer cell lines and clinical isolated human cancer cells. Mol Cancer. 2010, 9: 309.
[87] Issa A, Le TX, Shoushtari AN, Shields JD, and Swartz MA. Vascular
endothelial growth factor-C and C-C chemokine receptor 7 in tumor cell-lymphatic
cross-talk promote invasive phenotype. Cancer Research. 2009, 69: 349-357.
[74] Jemal A, Siegel R, Xu J, and Ward E. Cancer Statistics, 2010. CA Cancer J
Clin. 2010, 60: 277-300.
148
[55] Kaluz S and Van Meir EG. At the crossroads of cancer and inflammation: Ras
rewires an HIF-driven IL-1 autocrine loop. J Mol Med. 2011, 89(2): 91-94.
[77] Kaur J and Ralhan R. Establishment and characterization of a cell line from
smokeless tobacco associated oral squamous cell carcinoma. Oral Oncology. 2003,
39: 806-820.
[54] Ko JS, Rayman P, Ireland J, Swaidani S, Li G, Bunting KD, Rini B, Finke JH,
and Cohen PA. Direct and differential suppression of myeloid-derived suppressor
cell subsets by sunitinib is compartmentally constrained. Cancer Research. 2010,
70(9): 3526-36.
[35] Ko JS, Zea AH, Rini BI, Ireland JL, Elson P, Cohen P, Golshayan A, Rayman
PA, Wood L, Garcia J, Dreicer R, Bukowski R, Finke JH. Sunitinib mediates
reversal of myeloid-derived suppressor cell accumulation in renal cell carcinoma
patients. Clinical Cancer Research. 2009, 15: 2148-2157.
[105] Kodumudi KN, Woan K, Gilvary DL, Sahakian E, Wei S, and Djeu JY. A
novel chemoimmunomodulating property of docetaxel: suppression of myeloid-
derived suppressor cells in tumor bearers. Clinical Cancer Research. 2010, 16:
4583-4594.
[27] Kusmartsev S and Gabrilovich DI. Effect of tumor-derived cytokines and
growth factors on differentiation and immune suppressive features of myeloid cells
in cancer. Cancer Metastasis Rev. 2006, 25: 323-331.
[90] Lathers DM, Achille NJ, and Young MR. Incomplete Th2 skewing of
cytokines in plasma of patients with squamous cell carcinoma of the head and neck.
Hum Immunol. 2003, 64: 1160-1166.
[107] Le HK, Graham L, Cha E, Morales JK, Manjili MH, and Bear HD.
Gemcitabine directly inhibits myeloid derived suppressor cells in BALB/c mice
bearing 4T1 mammary carcinoma and augments expansion of T cells from tumor-
bearing mice. International Immunopharmacology. 2009, 9: 900-909.
[45] Lechner MG and Epstein AL. A New Mechanism for Blocking Myeloid-
Derived Suppressor Cells by CpG. Clinical Cancer Research. 2011 Feb 2.[Epub
ahead of print]
[94] Lechner MG, Lade S, Liebertz DJ, Prince HM, Brody GS, Webster HR, and
Epstein AL. Breast implant-associated, ALK-negative, T-cell, anaplastic, large-cell
lymphoma: Establishment and characterization of a model cell line (TLBR-1) for
this newly emerging clinical entity. Cancer. 2010 Nov 8. [Epub ahead of print]
PMID: 21061340
149
[28] Lechner MG, Liebertz DJ, and Epstein AL. Characterization of cytokine-
induced myeloid-derived suppressor cells from normal human peripheral blood
mononuclear cells. Journal of Immunology. 2010, 185: 2273-2284.
[95] Lechner MG, Megiel C, Liebertz DJ, and Epstein AL. Myeloid- derived
suppressor cell induction and mechanisms of suppression in 100 human solid tumor
cell lines. In Molecular and Cellular Biology and Immune Escape in Cancer,
February 7–12, 2010. Keystone Symposia on Molecular and Cellular Biology.
Silverthorne, CO. p.182, abstract 228.
[72] Leslie M. Biomedical research: immunology uncaged. Science. 2010, 327:
1573.
[58] Li F, Sonveaux P, Rabbani ZN, Liu S, Yan B, Huang Q, Vujaskovic Z,
Dewhirst MW, and Li CY. Regulation of HIF-1 α Stability through S-nitrosylation.
Mol Cell. 2007, 26(1): 63–74.
[101] Li J, Hu P, Khawli LA, and Epstein AL. Complete regression of experimental
solid tumors by LEC/chTNT-3 immunotherapy and CD25(+) T-cell depletion.
Cancer Research. 2003, 63(23): 8384-8392.
[93] Liebertz DJ, Lechner MG, Masood R, Sinha UK, Han J, Puri RK, Correa AJ,
and Epstein AL. Establishment and characterization of a novel head and neck
squamous cell carcinoma cell line USC-HN1. Head Neck Oncol. 2010, 2:5.
[92] Liebertz DJ, Lechner MG, Sinha UT, and Epstein AL. Mechanisms of immune
escape in head and neck squamous cell carcinoma depend upon human papilloma
virus infectivity. In Molecular and Cellular Biology and Immune Escape in Cancer,
February 7–12, 2010. Keystone Symposia on Molecular and Cellular Biology.
Silverthorne, CO. p.183, abstract 232.
[76] Lin CJ, Grandis JR, Carey TE, Gollin SM, Whiteside TL, Koch WM, Ferris
RL, Lai SY. Head and neck squamous cell carcinoma cell lines: established models
and rationale for selection. Head Neck. 2007, 29: 163-188.
[34] Liu CY, Wang YM, Wang CL, Feng PH, Ko HW, Liu YH, Wu YC, Chu Y,
Chung FT, Kuo CH, Lee KY, Lin SM, Lin HC, Wang CH, Yu CT, and Kuo HP.
Population alterations of L: -arginase- and inducible nitric oxide synthase-expressed
CD11b(+)/CD14 (-)/CD15 (+)/CD33 (+) myeloid-derived suppressor cells and CD8
(+) T lymphocytes in patients with advanced-stage non-small cell lung cancer. J
Cancer Res Clin Oncol. 2009, 136: 35-45.
150
[71] Liu Y, Zeng B, Zhang Z, Zhang Y, and Yang R. B7-H1 on myeloid- derived
suppressor cells in immune suppression by a mouse model of ovarian cancer. Clin.
Immunol. 2008, 129: 471–481.
[3] Lizée G, Cantu MA, and Hwu P. Less yin, more yang: confronting the barriers
to cancer immunotherapy. Clinical Cancer Research. 2007,13(18 Pt 1): 5250-5255.
[31] Mandruzzato S, Solito S, Falisi E, Francescato S, Chiarion-Sileni V, Mocellin
S, Zanon A, Rossi CR, Nitti D, Bronte V, and Zanovello P. IL4Ralpha+ myeloid-
derived suppressor cell expansion in cancer patients. Journal of Immunology. 2009,
182: 6562-6568.
[51] Marigo I, Bosio E, Solito S, Mesa C, Fernandez A, Dolcetti L, Ugel S, Sonda
N, Bicciato S, Falisi E, Calabrese F, Basso G, Zanovello P, Cozzi E, Mandruzzato
S, and Bronte V. Tumor-induced tolerance and immune suppression depend on the
C/EBPbeta transcription factor. Immunity. 2010, 32(6): 790-802.
[59] Matsusaka T, Fujikawa K, Nishio Y, Mukaida N, Matsushima K, Kishimoto T,
and Akira S. Transcription factors NF-IL6 and NF-kappa B synergistically activate
transcription of the inflammatory cytokines, interleukin 6 and interleukin 8. Proc
Natl Acad Sci U S A. 1993, 90(21): 10193-10197.
[85] Meissner M, Reichert TE, Kunkel M, Gooding W, Whiteside TL, Ferrone S,
and Seliger B. Defects in the human leukocyte antigen class I antigen processing
machinery in head and neck squamous cell carcinoma: association with clinical
outcome. Clinical Cancer Research. 2005, 11(7): 2552-2560.
[1] Moossa AR, Schimpff SC, and Robson MC. Comprehensive Textbook of
Oncology, Vol. 1. 2nd ed. San Francisco: Williams & Wilkins, 1991. Chapter 22,
p.159-160.
[66] Morelli AE and Thomson AW. Tolerogenic dendritic cells and the quest for
transplant tolerance, Nat. Rev. Immunol. 2007, 7: 610–621.
[37] Morse MA, Hall JR, and Plate JM. Countering tumor-induced
immunosuppression during immunotherapy for pancreatic cancer. Expert Opin Biol
Ther. 2009, 9: 331-339.
[48] Morse MA, Zhou LJ, Tedder TF, Lyerly HK, and Smith C. Generation of
dendritic cells in vitro from peripheral blood mononuclear cells with granulocyte-
macrophage-colony-stimulating factor, interleukin-4, and tumor necrosis factor-
alpha for use in cancer immunotherapy. Annals of Surgery. 1997, 226: 6–16.
151
[15] Movahedi K, Guilliams M, Van den Bossche J, Van den Bergh R, Gysemans
C, Beschin A, De Baetselier P, and Van Ginderachter JA. Identification of discrete
tumor-induced myeloid-derived suppressor cell subpopulations with distinct T cell-
suppressive activity. Blood. 2008, 111(8): 4233-4244.
[62] Munera V, Popovic PJ, Bryk J, Pribis J, Caba D, Matta BM, Zenati M, and
Ochoa JB. Stat 6-dependent induction of myeloid derived suppressor cells after
physical injury regulates nitric oxide response to endotoxin. Annals of Surgery.
2010, 251: 120-126.
[19] Nagaraj S and Gabrilovich DI. Tumor escape mechanism governed by
myeloid-derived suppressor cells. Cancer Research. 2008, 68: 2561-2563.
[111] Nagaraj S, Youn JI, Weber H, Iclozan C, Lu L, Cotter MJ, Meyer C, Becerra
CR, Fishman M, Antonia S, Sporn MB, Liby KT, Rawal B, Lee JH,and
Gabrilovich DI. Anti-inflammatory triterpenoid blocks immune suppressive
function of MDSCs and improves immune response in cancer. Clinical Cancer
Research. 2010, 16(6): 1812-1823.
[67] Natarajan S and Thomson AW. Tolerogenic dendritic cells and myeloid-
derived suppressor cells: potential for regulation and therapy of liver auto- and
alloimmunity. Immunobiology. 2010, 215(9-10): 698-703.
[49] Nefedova Y, Fishman M, Sherman S, Wang X, Beg AA, and Gabrilovich DI.
Mechanism of all-trans retinoic acid effect on tumor-associated myeloid-derived
suppressor cells. Cancer Research. 2007, 67(22): 11021-11028.
[11] Nizar S, Meyer B, Galustian C, Kumar D, and Dalgleish A. T regulatory cells,
the evolution of targeted immunotherapy. Biochim Biophys Acta. 2010, 1806(1): 7-
17.
[96] Okamoto A, Chikamatsu K, Sakakura K, Hatsushika K, Takahashi G, and
Masuyama K. Expansion and characterization of cancer stem-like cells in squamous
cell carcinoma of the head and neck. Oral Oncol. 2009;45:633-639.
[42] Ostrand-Rosenberg S and Sinha P. Myeloid-derived suppressor cells: linking
inflammation and cancer. Journal of Immunology. 2009, 182: 4499-4506.
[73] Pai SI and Westra WH. Molecular pathology of head and neck cancer:
implications for diagnosis, prognosis, and treatment. Annu Rev Pathol. 2009, 4: 49-
70.
152
[13] Peranzoni E, Zilio S, Marigo I, Dolcetti L, Zanovello P, Mandruzzato S,
Bronte V. Myeloid-derived suppressor cell heterogeneity and subset definition.
Curr Opin Immunol. 2010, 22: 238-244.
[56] Peyssonnaux C, Cejudo-Martin P, Doedens A, Zinkernagel AS, Johnson RS,
and Nizet V. Cutting edge: Essential role of hypoxia inducible factor-1alpha in
development of lipopolysaccharide-induced sepsis. Journal of Immunology. 2007,
178: 7516-7519.
[29] Poschke I, Mougiakakos D, Hansson J, Masucci GV, and Kiessling R.
Immature immunosuppressive CD14+HLA-DR-/low cells in melanoma patients are
Stat3hi and overexpress CD80, CD83, and DC-sign. Cancer Research. 2010, 70:
4335-4345.
[108] Priceman SJ, Sung JL, Shaposhnik Z, Burton JB, Torres-Collado AX,
Moughon DL, Johnson M, Lusis AJ, Cohen DA, Iruela-Arispe ML, Wu L.
Targeting distinct tumor-infiltrating myeloid cells by inhibiting CSF-1 receptor:
combating tumor evasion of antiangiogenic therapy. Blood. 2010, 115: 1461-1471.
[81] Prince ME and Ailles LE. Cancer stem cells in head and neck squamous cell
carcinoma. J Clin Oncol. 2008, 26: 2871-2875.
[60] Rius J, Guma M, Schachtrup C, Akassoglou K, Zinkernagel AS, Nizet V,
Johnson RS, Haddad GG, and Karin M. NF-kappaB links innate immunity to the
hypoxic response through transcriptional regulation of HIF-1alpha. Nature. 2008,
453: 807-811.
[53] Rodriguez PC, Ernstoff MS, Hernandez C, Atkins M, Zabaleta J, Sierra R, and
Ochoa AC. Arginase I-producing myeloid-derived suppressor cells in renal cell
carcinoma are a subpopulation of activated granulocytes. Cancer Research. 2009,
69(4): 1553-1560.
[5] Sadun RE, Sachsman SM, Chen X, Christenson KW, Morris WZ, Hu P, and
Epstein AL. Immune signatures of murine and human cancers reveal unique
mechanisms of tumor escape and new targets for cancer immunotherapy. Clinical
Cancer Research. 2007,13: 4016–4025.
[9] Sakaguchi S. Naturally arising Foxp3-expressing CD25+CD4+ regulatory T
cells in immunological tolerance to self and non-self. Nature Immunology. 2005, 6:
345–352.
[57] Semenza GL. Targeting HIF-1 for cancer therapy. Nat Rev Cancer. 2003,
3(10): 721-732.
153
[41] Serafini P, Borrello I, and Bronte V. Myeloid suppressor cells in cancer:
recruitment, phenotype, properties, and mechanisms of immune suppression. Semin
Cancer Biol. 2006, 16: 53-65.
[109] Serafini P, Meckel K, Kelso M, Noonan K, Califano J, Koch W, Dolcetti L,
Bronte V, and Borrello I. Phosphodiesterase-5 inhibition augments endogenous
antitumor immunity by reducing myeloid-derived suppressor cell function. Journal
of Experimental Medicine. 2006;203: 2691-2702.37.
[23] Serafini P, Mgebroff S, Noonan K, and Borrello I. Myeloid-derived suppressor
cells promote cross-tolerance in B cell lymphoma by expanding regulatory T cells.
Cancer Research. 2008, 68: 5439-5449.
[47] Sinha P, Clements VK, Fulton AM, and Ostrand-Rosenberg S. Prostaglandin
E2 promotes tumor progression by inducing myeloid-derived suppressor cells.
Cancer Research. 2007, 67: 4507–4513.
[91] Sparano A, Lathers DM, Achille N, Petruzzelli GJ, and Young MR.
Modulation of Th1 and Th2 cytokine profiles and their association with advanced
head and neck squamous cell carcinoma. Otolaryngol Head Neck Surg. 2004,
131(5): 573-576.
[22] Srivastava MK, Sinha P, Clements VK, Rodriguez P, and Ostrand-Rosenberg
S. Myeloid-derived suppressor cells inhibit T-cell activation by depleting cystine
and cysteine. Cancer Research. 2010, 70: 68-77.
[80] Stadler ME, Patel MR, Couch ME, and Hayes DN. Molecular biology of head
and neck cancer: risks and pathways. Hematol Oncol Clin N Am. 2008, 22: 1099-
1124.
[4] Stewart TJ and Abrams SI. How tumours escape mass destruction. Oncogene.
2008, 27: 5894–5903.
[6] Stewart TJ and Smyth MJ. Improving cancer immunotherapy by targeting
tumor-induced immune suppression. Cancer Metastasis Reviews. 2011, 30(1): 125-
140.
[98] Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA,
Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. Gene set enrichment
analysis: a knowledge-based approach for interpreting genome-wide expression
profiles. Proc Natl Acad Sci. 2005, 102: 15545-15550.
154
[8] Takahashi T, Kuniyasu Y, Toda M, Sakaguchi N, Itoh M, Iwata M, Shimizu J,
and Sakaguchi S. Immunologic self-tolerance maintained by CD25+CD4+ naturally
anergic and suppressive T cells: induction of autoimmune disease by breaking their
anergic/suppressive state. Int. Immunol.1998, 10: 1969–1980.
[65] Thomson AW and Robbins PD. Tolerogenic dendritic cells for autoimmune
disease and transplantation, Ann. Rheum. Dis. 2008, 67 (Suppl. 3): iii90–iii96.
[70] Ugel S, Delpozzo F, Desantis G, Papalini F, Simonato F, Sonda N, Zilio S, and
Bronte V. Therapeutic targeting of myeloid-derived suppressor cells. Curr. Opin.
Pharmacol. 2009, 9: 470–481.
[99] van Staveren WC, Solís DY, Hébrant A, Detours V, Dumont JE, and Maenhaut
C. Human cancer cell lines: Experimental models for cancer cells in situ? For
cancer stem cells? Biochim Biophys Acta. 2009, 2: 92-103.
[103] Veltman JD, Lambers ME, van Nimwegen M, Hendriks RW, Hoogsteden
HC, Aerts JG, and Hegmans JP. COX-2 inhibition improves immunotherapy and is
associated with decreased numbers of myeloid-derived suppressor cells in
mesothelioma. Celecoxib influences MDSC function. BMC Cancer. 2010, 10: 464.
[106] Vincent J, Mignot G, Chalmin F, Ladoire S, Bruchard M, Chevriaux A,
Martin F, Apetoh L, Rébé C, Ghiringhelli F. 5-Fluorouracil selectively kills tumor-
associated myeloid-derived suppressor cells resulting in enhanced T cell-dependent
antitumor immunity. Cancer Research. 2010, 70: 3052-3061.
[112] Walmsley SR, Chilvers ER, and Whyte MKB. Hypoxia, hypoxia inducible
factor and myeloid cell function. Arthritis Research and Therapy. 2009, 11: 219-
225.
[89] Walsh JE, Lathers DM, Chi AC, Gillespie MB, Day TA, and Young MR.
Mechanisms of tumor growth and metastasis in head and neck squamous cell
carcinoma. Curr Treat Options Oncol. 2007, 8: 227-228.
[7] Whiteside TL. The tumor microenvironment and its role in promoting tumor
growth. Oncogene. 2008, 27, 5904–5912
[43] Xiang X, Poliakov A, Liu C, Liu Y, Deng ZB, Wang J, Cheng Z, Shah SV,
Wang GJ, Zhang L, Grizzle WE, Mobley J, and Zhang HG. Induction of myeloid-
derived suppressor cells by tumor exosomes. International Journal of Cancer.
2009, 124: 2621-2633.
155
[14] Youn JI, Nagaraj S, Collazo M, and Gabrilovich DI. Subsets of myeloid-
derived suppressor cells in tumor-bearing mice. Journal of Immunology. 2008,
181(8): 5791-8502.
[84] Young MR. Protective mechanisms of head and neck squamous cell
carcinomas from immune assault. Head Neck. 2006, 28: 462-470.
[36] Zea AH, Rodriguez PC, Atkins MB, Hernandez C, Signoretti S, Zabaleta J,
McDermott D, Quiceno D, Youmans A, O'Neill A, Mier J, and Ochoa AC.
Arginase-producing myeloid suppressor cells in renal cell carcinoma patients: a
mechanism of tumor evasion. Cancer Research. 2005, 65: 3044-3048.
[64] Zhang C, Zou X, Leluo G, Xu J, Xiang M. Prevention of type 1 diabetes by
immature dendritic cells treated with an ethanol extract of Paecilomyces hepiali
Chen mycelium. Methods Find Exp Clin Pharmacol. 2008, 30(6): 421-429.
[68] Zhou Z, French DL, Ma G, Eisenstein S, Chen Y, Divino CM, Keller G, Chen
SH, Pan PY. Development and function of myeloid-derived suppressor cells
generated from mouse embryonic and hematopoietic stem cells. Stem Cells. 2010,
28: 620–632.
[104] Zoglmeier C, Bauer H, Noerenberg D, Wedekind G, Bittner P, Sandholzer N,
Rapp M, Anz D, Endres Prof S, and Bourquin C. CpG blocks immune suppression
by myeloid- derived suppressor cells in tumor-bearing mice. Clinical Cancer
Research. 2011 Jan 13. [Epub ahead of print] doi:10.1158/1078-0432.
156
APPENDIX
Studies in our lab have identified Celecoxib and analogs dimethyl celecoxib (DMC) and
unmethylated celecoxib (UMC) as inhibitors of suppressive function in CD33
+
MDSC in
vitro. Of note, the reversal of MDSC effects by CXB and analogs DMX and UMC does
not appear to rely upon cyclo-oxygenase (COX)2 enzyme inactivation, as demonstrated
by the persistence of therapeutic effects in the presence of prostaglandin E
2
rescue,
efficacy of analog DMC with low to absent COX inhibitory action, and the absence of
effect seen with the structurally-unrelated COX2-seletcive inhibitor naproxen. Lin et al.
1
have reported potent Akt-pathway inhibition as a mechanism of action for CXB and
analogs DMC and UMC. Akt signaling is involved in NF B-C/EBP  transactivation of
JAK/STAT signaling and stimulation of autocrine IL-6 production, providing a possible
mechanism of action for inhibition of human MDSC that warrants further study (Chapter
3). For these studies, human CD33
+
MDSC induced by cancer cell lines were co-cultured
with fresh, autologous CFSE-labeled T cells at a ratio of 1:4 in the presence of absence of
drugs and vaqdThtaglandin E2 as indicated. T cell stimulation was provided by anti
CD3/CD28 microbeads. After three days in culture, T cell proliferation was measured as
CFSE dilution by flow cytometry. Mean T cell proliferation shown + SEM, two
independent experiments.
Figure 32. Reversal of CD33
+
MDSC suppressive function by celecoxib and analogs through a non-COX2
dependent mechanism.
1
Lin HP. Celecoxib: Its non-COX-2 targets and its anti-cancer effects (Thesis Dissertation). Ohio: Ohio
State University, 2005.

Abstract (if available)

Abstract Tumor immune tolerance can derive from the recruitment of suppressor cell populations, including myeloid-derived suppressor cells (MDSC). MDSCs inhibit anti-tumor T cell responses through a variety of mechanisms including nutrient depletion, production of reactive oxygen and nitrogen species, VEGF expression, and regulatory T cell expansion. In cancer patients and murine tumor models, MDSC accumulation correlates with increased stage and tumor burden, but the frequency and mechanisms of MDSC induction in human cancer remain poorly understood. This study examined the ability of a diverse set of human solid tumor cell lines to induce MDSC using in vitro tumor co-culture methods. Newly induced suppressor cells were characterized for morphology, phenotype, and gene expression. Of over 100 solid tumor cell lines examined, 45 generated canonical CD33+ HLADRlow Lineage- MDSC, with high frequency of induction by cervical, ovarian, colorectal, renal cell, and head and neck squamous cell carcinoma cell lines. CD33+ MDSC could be induced by some cancer cell lines of all tumor types examined with the notable exception of breast cancer cell lines (0/9, inclusive of models with different hormone and HER2 mutation status). Upon further examination, breast cancer cell lines and other tumor types with infrequent CD33+ MDSC induction preferentially induced an undiscovered CD11b+ CD33low HLADRlow Lineage- MDSC subset. Gene and protein expression, neutralization, and cytokine induction experiments determined that the ability of tumor cell lines to induce CD33+ MDSC depended upon over-expression of IL-1β, IL-6, TNFα, VEGF, and GM-CSF, while CD11b+ MDSCinduction correlated with over-expression of FLT3L and TGFβ. Both CD33+ and CD11b+ MDSC subsets appeared as immature myeloid cells by Wright-Giemsa staining and had significantly up-regulated expression of inducible nitric oxide synthase, TGFβ, NADPH oxidase, VEGF, and arginase-1 genes compared with normal myeloid cells.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The effect of inhibiting the NFkB pathway on myeloid derived suppressor cell cytokine production

PDF

Study of bone morphogenetic protein-2 and stromal cell derived factor-1 in prostate cancer

PDF

Suppressor cell therapy: targeting T regulatory cells in cancer

PDF

Identification and characterization of immune-escape mechanisms in solid tumors

PDF

The role of fibroblast growth factor signaling on postnatal hepatic progenitor cell expansion

PDF

Generation and characterization of fully human anti-CD19 chimeric antigen receptor T (CAR-T) cells for the treatment of hematologic malignancies

PDF

Exploration of the roles of cancer stem cells and survivin in the pathogenesis and progression of prostate cancer

PDF

Role of cancer-associated fibroblast secreted annexin A1 in generation and maintenance of prostate cancer stem cells

PDF

Novel roles for Maf1 in embryonic stem cell differentiation and adipogenesis

Asset Metadata

Creator Lechner, Melissa Genevieve (author)

Core Title Human myeloid-derived suppressor cells in cancer: Induction, functional characterization, and therapy

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Systems Biology

Publication Date 11/03/2012

Defense Date 03/11/2011

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag caat-enhancer-binding protein beta,cancer,cytokines,Fms-related tyrosine kinase 3 ligand,granulocyte macrophage-colony stimulating factor,human,hypoxia-inducible factor 1 alpha,immunology,immunotherapy,interleukin-1 beta,interleukin-6,myeloid-derived suppressor cells,OAI-PMH Harvest,signal transducer and activator of transcription 3,transforming growth factor beta,tumor cell lines,tumor immunity,tumor necrosis factor alpha,vascular endothelial growth factor

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Epstein, Alan L. (committee chair), Taylor, Clive R. (committee chair), Horwitz, David (committee member), Kaslow, Harvey R. (committee member), McDonough, Alicia A. (committee member), McMillan, Minnie (committee member)

Creator Email lechner@usc.edu,melissalechner@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3872

Unique identifier UC1334622

Identifier etd-Lechner-4434 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-463063 (legacy record id),usctheses-m3872 (legacy record id)

Legacy Identifier etd-Lechner-4434.pdf

Dmrecord 463063

Document Type Dissertation

Rights Lechner, Melissa Genevieve

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu