Dokumentvorlage für Diplomarbeiten

I

Friedrich-Alexander-Universität Erlangen-Nürnberg

Naturwissenschaftliche Fakultät

In-depth characterization of myeloid enhanced humanized mouse strains

Der Naturwissenschaftlichen Fakultät der Friedrich-Alexander-Universität

Erlangen-Nürnberg zur Erlangung des akademischen Grades

Doktor der Naturwissenschaften (Dr. rer. nat.)

vorgelegt von

Ilona-Petra Maser

aus München

Index of abbreviations

II

Als Dissertation genehmigt

von der Naturwissenschaftlichen Fakultät

der Friedrich-Alexander-Universität Erlangen-Nürnberg

Tag der mündlichen Prüfung: 26. Februar 2021

Gutachter: Prof. Dr. Falk Nimmerjahn

Prof. Dr. Alexander Steinkasserer

Vorsitzender des Promotionsorgans: Prof. Dr. Wolfgang Achtziger


III

INDEX OF ABBREVIATIONS ..................................................................................................... VI

LIST OF FIGURES ........................................................................................................................ IX

LIST OF TABLES......................................................................................................................... 10

SUMMARY ................................................................................................................................... 11

ZUSAMMENFASSUNG ............................................................................................................... 13

1. INTRODUCTION ................................................................................................................. 16

1.1 The innate and adaptive immune system .............................................................................16

1.1.2 The role of immune cells in tumor development ................................................................18

1.1.2.1 Lymphoid immune cells in tumors..................................................................................19

1.1.2.2 Myeloid immune cells in tumors ....................................................................................19

1.2 Cancer immunotherapy ........................................................................................................20

1.3 Promising in vivo models for cancer immunotherapy: humanized mice ...............................23

1.3.1 Background .......................................................................................................................23

1.3.2 Classical humanized mouse models ...................................................................................24

1.3.3 Limitations of classical humanized mice: cytokines are the key factor ................................26

2. AIM OF THE STUDY ........................................................................................................... 28

3. MATERIALS AND METHODS ........................................................................................... 29

3.1 Materials ..............................................................................................................................29

3.1.1 Antibodies .........................................................................................................................29

3.1.1.1 Flow cytometry antibodies ................................................................................................29

3.1.1.2 IHC antibodies ...................................................................................................................30

3.1.2 Buffer and media ...............................................................................................................30

3.1.3 Cell lines ............................................................................................................................31

3.1.4 Consumable material .........................................................................................................31

3.1.5 Reagents............................................................................................................................32

3.1.6 Instruments .......................................................................................................................34


IV

3.1.7 Software ............................................................................................................................36

3.2 Methods ...............................................................................................................................36

3.2.1 In vivo techniques ..............................................................................................................36

3.2.1.1 Generation of humanized mice ......................................................................................37

3.2.1.2 Blood and serum sampling .............................................................................................37

3.2.1.3 Tissue sampling..............................................................................................................37

3.2.1.4 Generation of humanized mice by depletion of mouse counterparts .............................37

3.2.1.5 Tumor growth in humanized mice..................................................................................37

3.2.2 Ex vivo / in vitro experiments .............................................................................................38

3.2.2.1 Preparation of human hematopoietic stem cells ............................................................38

3.2.2.2 Preparation of single cell suspensions for flow cytometry ..............................................38

3.2.2.3 Immunohistochemistry ..................................................................................................39

3.2.2.4 Phagocytosis Assay .......................................................................................................40

3.2.2.5 ELISA for detection of total mouse CSF-1.......................................................................40

3.2.2.6 Preparation of tumor lysate from frozen tissue .............................................................41

3.2.2.7 Bio-Plex multiplex immunoassay ...................................................................................41

3.2.2.8 Isolation of human PBMCs from human whole blood ....................................................42

3.2.2.9 Isolation of human pDCs and ex vivo stimulation ...........................................................42

3.2.2.10 Immunoassay for human-specific IgG antibodies .............................................................42

3.2.2.11 Flow cytometry ................................................................................................................43

3.2.2.12 Statistical analyses ...........................................................................................................44

4. RESULTS .............................................................................................................................. 45

4.1 Analysis of human hematopoietic stem cells ........................................................................45

4.2 Characterization of humanized NOG mice in comparison to myeloid enhanced humanized

mice ..............................................................................................................................................46

4.2.1 Generation of humanized mice and macroscopic differences .............................................46

4.2.2 Human immune cell engraftment in different organs .........................................................47

4.2.2.1 Peripheral blood ............................................................................................................47

4.2.2.2 Primary lymphoid organs ...............................................................................................52

4.2.2.3 Secondary lymphoid organs ...........................................................................................56

4.2.2.3 Other tissue ...................................................................................................................59

4.2.2.4 Summary .......................................................................................................................59


V

4.3 Depletion of murine macrophages and granulocytes leads to higher overall humanization

but decrease in survival....................................................................................................................60

4.3.1 Background .......................................................................................................................60

4.3.2 Depletion studies with huNOG-EXL mice ............................................................................61

4.3.2.1 Peripheral blood analysis ...............................................................................................61

4.3.2.2 Other Organs .................................................................................................................63

4.3.3 Summary ...........................................................................................................................66

4.4 Functionality of human macrophages from huNOG-EXL mice ..............................................66

4.4.1 Stimulation with Toll-like receptor agonist 8 in vivo ...........................................................66

4.4.2 Analysis of human macrophages ex vivo ............................................................................67

4.5 Functionality of human pDCs from huNOG-EXL mice............................................................69

4.6 Growth of human xenograft tumors in humanized mice ......................................................70

4.6.1 Tumor growth kinetics .......................................................................................................71

4.6.2 Characterization of different tumor lines or PDX in different mouse strains .......................71

4.6.3 pDC occurrence in tumors is dependent on several human cytokines and chemokines ......73

4.6.4 Functional analysis of tumor-resident pDCs ...................................................................75

5 DISCUSSION ........................................................................................................................ 80

5.1 Generation of humanized mice .............................................................................................80

5.2 Characterization of human HSCs and donor variability .........................................................80

5.3 Comparison of immune cell compositions in different organs ..............................................81

5.4 Functional characterization of human macrophages and pDCs ............................................85

5.5 Depletion with anti-mouse CSF-1R .......................................................................................86

5.6 Comparison of tumor growth and tumor immune cell composition .....................................87

5.7 Outlook .................................................................................................................................89

REFERENCES ............................................................................................................................. XCI

ACKNOWLEDGEMENT ......................................................................................................... XCIX


VI


ACT Adoptive cell therapy

ADCC Antibody dependent cellular cytotoxicity

ANOVA Analysis of variance

APCs Antigen-presenting cells

BM Bone marrow

BRG Balb/c Rag2−/− Il2rγ -/-

CAR-T Chimeric antigen receptor T cells

CCL C-C motif chemokine ligand

CD Cluster of differentiation

CDC Complement dependent cytotoxicity

cDC1 Cross-presenting CD141+ DCs

cDC2 Classical CD1c+ DCs

CLL Chronic lymphocytic leukemia

CLP Common lymphoid progenitors

CM Central memory

CMP Common myeloid progenitor

CSF-1 Colony stimulating factor 1

CSF-1R Colony stimulating factor 1 receptor

CTLA-4 Cytotoxic T lymphocyte antigen 4

CXCL Chemokine (C-X-C motif) ligand

CYTOF Cytometry by time of flight

DAB 3,3′-diaminobenzidin

DC Dendritic cells

DN Double negative

DNA Deoxyribonucleic acid

DP Double positive

e.g. Exemply gratia

EDTA Ethylenediaminetetraacetic acid

ELISA Enzyme-linked immunosorbent assay

EM Effector memory

EMEM Eagle's minimum essential medium

FcR Fc receptor

FCS Fetal calf serum

FDA Food and Drug Administration

G-CSF Granulocyte colony-stimulating factor

GM-CSF Granulocyte-macrophage colony-stimulating factor

GMP Granulocyte-macrophage progenitor

GvHD Graft versus host disease

h Hour

HIS Human immune system

HIV Human immunodeficiency viruses

HLA Human leukocyte antigen

HLH Hemophagocytic lymphohistiocytosis

HSC Hematopoietic stem cells

huNOG-EXL Humanized NOG


VII

huNOG-EXL Humanized NOG-EXL

huNSG-SGM3 Humanized NSG-SGM3

i.p. Intraperitoneal

i.v. Intravenous

IFN Interferon

Ig Immunoglobulin

IHC Immunohistochemistry

IL Interleukin

IL-15Rα Interleukin-15 receptor α

IQR Interquatile range

KIT Receptor tyrosine kinase type III

LPS Lipopolysaccharide

LT Long term

MDSCs Myeloid-derived suppressor cells

MEP Megakaryocyte/erythroid progenitor

MFI Mean fluorescence intensity

MHC-II Major histocompatibility complex class II

MIP-1 Macrophage inflammatory protein-1

mM Millimolar

MPPS Multipotent progenitors

NEAA Non-essential amino acid

Neg. Negative

NK Natural killer

NOD Non-obese diabetic

NOG NOD/Shi-scid/IL-2Rγnull mouse

NSG NOD scid gamma mouse

ON Overnight

PAMPs Pathogen-associated molecular patterns

PBMCs Peripheral blood mononuclear cells

PBS Phosphate-buffered saline

PD-1 Programmed cell death protein 1

pDCs Plasmacytoid dendritic cells

PD-L1 Programmed death ligand 1

Pos. Positive

PRR Pathogen recognition receptor

Rag Recombination activating gene

RBC Red blood cell

RNA Ribonucleic acid

RT Room temperature

SCF Stem cell factor

SCID Severe combined immunodeficiency

SEM Standard error of the mean

SIRPα Signal regulatory protein α

SPF Specific pathogen free

ST Short term

TAA Tumor-associated antigens

TAM Tumor-associated macrophages


VIII

TCBs T cell bispecific antibodies

TLR Toll-like receptor

TME Tumor microenvironment

TNFα Tumor necrosis factor α

TPO Thrombopoietin

Tregs Regulatory T cells

WBC White blood cells

wk Week

List of figures

IX

List of figures

FIGURE 1: HIERARCHY OF HUMAN HEMATOPOIETIC STEM CELLS. ....................................................................17

FIGURE 2: DIFFERENT METHODS TO GENERATE HUMANIZED MICE. .................................................................25

FIGURE 3: GATING STRATEGY FOR DC SUBPOPULATIONS .................................................................................43

FIGURE 4: GATING STRATEGY FOR LYMPHOID CELLS ........................................................................................43

FIGURE 5: GATING STRATEGY FOR MYELOID CELLS ..........................................................................................44

FIGURE 6: CHARACTERIZATION OF CD34+ HEMATOPOIETIC STEM CELLS...........................................................45

FIGURE 7: DIFFERENCES IN BODY WEIGHT AND SURVIVAL ...............................................................................46

FIGURE 8: PERIPHERAL BLOOD ANALYSIS IN HUMANIZED MICE .......................................................................48

FIGURE 9: HETEROGENEITY IN ENGRAFTMENT BY HSC DONOR ........................................................................49

FIGURE 10: ANALYSIS OF MONOCYTES IN PERIPHERAL BLOOD .........................................................................50

FIGURE 11: ANALYSIS OF HUMAN CYTOKINES/CHEMOKINES IN SERUM OF HUMANIZED MICE .........................51

FIGURE 12: HUMAN IgG LEVEL IN SERUM OF HUMANIZED MICE ......................................................................52

FIGURE 13: IMMUNE CELL ANALYSIS IN BONE MARROW OF HUMANIZED MICE ...............................................53

FIGURE 14: SPECIFIC FLOW CYTOMETRY DC MARKERS .....................................................................................54

FIGURE 15: ANALYSIS OF THYMI FROM HUMANIZED MICE ...............................................................................55

FIGURE 16: REPRESENTATIVE IMAGE OF IHC STAINING IN huNOG-EXL THYMUS ...............................................55

FIGURE 17: MACROSCOPIC AND IN-DEPTH CHARACTERIZATION OF SPLEENS ...................................................57

FIGURE 18: ANALYSIS OF AXIAL LYMPH NODES ................................................................................................58

FIGURE 19: MACROSCOPIC CHARACTERIZATION OF LIVERS ..............................................................................59

FIGURE 20: PERIPHERAL BLOOD ANALYSIS IN ΑCSF-1R TREATED huNOG-EXL mice ...........................................62

FIGURE 21: ANALYSIS OF PERIPHERAL ORGANS AFTER αCSF-1R TREAMENT .....................................................65

FIGURE 22: CYTOKINE RESPONSE OF HUMAN MACROPHAGES IN huNOG-EXL MICE .........................................67

FIGURE 23: IMMUNOFLUORESCENCE STAINING OF ENRICHED BM CELLS FROM huNOG-EXL ............................68

FIGURE 24: PHAGOCYTOSIS ASSAY WITH MACROPHAGES DIFFERENTIATED FROM huNOG-EXL BM ..................69

FIGURE 25: huNOG-EXL DEVELOP FUNCTIONAL HUMAN PDCs .........................................................................70

FIGURE 26: TUMOR GROWTH KINETICS OF DIFFERENT TUMORS IN HUMANIZED MICE ....................................71

FIGURE 27: COMPARISON OF OTAL LEUKOCYTES AND MYELOID CELLS IN PERIPHERAL BLOOD .........................72

FIGURE 28: COMPARISON OF TUMOR IMMUNE CELL INFILTRATE ....................................................................73

FIGURE 29: ANALYSIS OF PDC-ASSOCIATED CHEMOKINES AND CYTOKINES OF DIFFERENT TUMORS .................74

FIGURE 30: FUNCTIONAL ANALYSIS OF TUMOR-RESIDENT PDCs .......................................................................76

FIGURE 31: HUMAN CYTOKINES IN SERUM OF TLR TREATED HUMANIZED MICE ...............................................77

FIGURE 32: CYTOKINE PROFILE OF TUMORS IN HUMANIZED MICE AS HEAT MAP .............................................79

Summary

10

List of tables

TABLE 1: CATEGORIES OF CANCER IMMUNOTHERAPY ............................................................................... 23

TABLE 2: FACS ANTIBODIES ........................................................................................................................ 29

TABLE 3: PRIMARY ANTIBODIES ................................................................................................................. 30

TABLE 4: SECONDARY ANTIBODIES ............................................................................................................ 30

TABLE 5: MEDIA USED FOR IN VITRO EXPERIMENTS ................................................................................... 30

TABLE 6: CELL LINES ................................................................................................................................... 31

TABLE 7: MATERIAL LIST ............................................................................................................................ 31

TABLE 8: REAGENT LIST.............................................................................................................................. 32

TABLE 9: USED INSTRUMENTS ................................................................................................................... 34

TABLE 10: SOFTWARE ................................................................................................................................ 36

Summary

11

Summary

Despite all progress and developments in the last decades, cancer still is the second leading

cause of death. Depending on many factors, such as gender, age, lifestyle and income,

some cancer types have decreased in incidence and mortality, whereas other cancer types

such as breast cancer incidence and mortality is still rising (1). Unfortunately, oncology is a

particular challenging field for developing new drug candidates and heterogeneity of the

different tumor types poses an additional challenge. Moreover, anti-cancer drugs have the

lowest clinical success rate compared with other disease areas. This is partly due to the

inadequate translation of preclinical findings into clinic trials. Consequently, the improve-

ment of preclinical models is required to enhance the predictive power of preclinical studies

and to reduce the failure rate of novel oncology drugs. In recent years cancer immunother-

apy became a promising approach in the fight against cancer and thus the need for in vivo

models with a functional human immune system has increased (2-5).

With the development of classically humanized mice a big hurdle of preclinical

mouse-models was overcome by adequately reflecting human lymphoid system. However,

these classical humanized mouse models still have critical limitations such as the poor re-

constitution of human innate immune development. Simultaneously, the significance of tu-

mor-associated myeloid cells and their contribution in tumor progression became more ev-

ident. Consequently, transgenic mouse strains expressing human interleukin-3 (IL-3) and

granulocyte-macrophage colony-stimulating factor (GM-CSF), so called myeloid enhanced

NOG-EXL and NSG-SGM-3 mice, were generated based on the importance of these cyto-

kines for innate immune cell differentiation.

In the present study a comprehensive characterization and direct comparison of

these novel humanized mouse models was performed with focus on their myeloid reconsti-

tution to determine which model is most suitable for studies involving human myeloid cell

reconstitution. An in-depth characterization of humanization levels in peripheral blood as

well as primary and secondary lymphoid organs was conducted with special attention on

the myeloid cell populations. It was found that novel models NOG-EXL and NSG-SGM3

display more myeloid cell populations in peripheral blood as well as in primary and second-

ary lymphoid organs. Besides monocytes and macrophages also increased levels for hu-

man dendritic cells were found. But not only was the occurrence of these cell populations

analyzed, but also their functionality. Macrophages as well as pDCs from humanized NOG-

EXL (huNOG-EXL) were found to be functional as they are able to react to exogenous stim-

uli with TLR agonists. Thus, this work is the first to describe functional human macrophages

and pDCs in huNOG-EXL mice.

Summary

12

Additionally, adaptive immunity was analyzed. Impaired IgG class switch is a com-

mon limitation in classical humanized mice, this study also compared the herein analyzed

models macroscopically for lymph node and thymus development. In addition, antigen-spe-

cific IgG production was investigated and high levels were found in the myeloid enhanced

models.

The analyzed mouse models showed increased levels of myeloid cells in the

periphery. These cells are reported to play an essential role in a variety of solid tumors.

Therefore, further investigation concentrated on myeloid immune infiltration in tumors. This

is a crucial aspect in cancer immunotherapy, as the abundance of myeloid cells in solid

tumors often correlates with poor clinical outcome. Consequently, myeloid enhanced hu-

manized mice were transplanted with human xenograft tumors and tumor immune infiltrate

was analyzed. The tumor itself contributes to the human cytokine milieu and this raised the

question whether immune cell composition in humanized mice is more influenced by the

mouse strain used or by the chosen tumor model. Therefore, studying the relative effects

of transgenic and tumor-derived cytokines to myeloid cell reconstitution in humanized mice

was of paramount importance in this study.

In summary, huNOG-EXL mice were identified as a robust mouse strain to study

human immunity. Due to increased IgG levels and the reconstitution of important immune

subclasses in different tumor models, they offer a clinical relevant model for cancer immu-

notherapy. Additionally, the relatively high percentage of human pDCs allows studying this

rare cell population without further manipulation. Specific tumor models were identified that

are able of generating functional tumor-associated pDCs that are not based on the cytokine

expression of transgenes and that functionally responded to TLR activation. While the

choice of the mouse model was essential for differentiation of innate immune subsets in

periphery, this impact was lost when focusing on the tumor microenvironment. Tumor-de-

rived cytokines might be able to attract and differentiate innate immune cells even if their

frequency in the periphery is low. Thus, it was shown for first time that the tumor milieu but

not the mouse background defines intratumoral frequencies of innate immune cell subsets

such as pDCs.

Zusammenfassung

13

Zusammenfassung

Trotz aller Fortschritte und Entwicklungen in den letzten Jahrzehnten ist Krebs weltweit im-

mer noch die zweithäufigste Todesursache. Abhängig von vielen Faktoren wie Geschlecht,

Alter, Lebensstil und Einkommen haben einige Krebsarten an Inzidenz und Mortalität ab-

genommen, während andere Krebsarten wie Brustkrebs in Inzidenz und Mortalität immer

noch zunehmen (1). Leider ist die Onkologie ein besonders herausforderndes Gebiet für

die Entwicklung neuer Medikamentenkandidaten und die Heterogenität der verschiedenen

Tumor-Typen stellt eine zusätzliche Herausforderung dar. Darüber hinaus weisen Krebs-

medikamente im Vergleich zu anderen Krankheitsbereichen die niedrigste klinische Erfolgs-

rate auf. Dies ist teilweise auf die fehlende Translatierbarkeit präklinischer Befunde in klini-

sche Studien zurückzuführen. Folglich ist eine Verbesserung der präklinischen Modelle er-

forderlich, um die Vorhersagekraft präklinischer Studien zu verbessern und die Ausfallrate

neuartiger onkologischer Arzneimittel zu verringern. In den letzten Jahren wurde die Krebs-

immuntherapie zu einem vielversprechenden Ansatz im Kampf gegen Krebs und daher hat

der Bedarf an in vivo Modellen mit einem funktionierenden menschlichen Immunsystem

zugenommen (2-5).

Mit der Entwicklung klassisch humanisierter Mäuse, welche die Rekonstitution der

menschlichen Lymphozyten angemessen reflektieren, wurde bereits eine große Hürde in

präklinischen Maus-Modellen überwunden. Diese humanisierten Mausmodelle weisen je-

doch immer noch kritische Einschränkungen, wie beispielsweise die schlechte Rekonstitu-

tion des menschlichen angeborenen Immunsystems, auf. Gleichzeitig wurde die Bedeutung

von Tumor-assoziierten myeloischen Zellen und ihr Beitrag zur Tumorprogression deutli-

cher. Folglich wurden basierend auf der Wichtigkeit dieser Zytokine für die Differenzierung

der Zellen der angeborenen Immunität, transgene Mausstämme entwickelt, die menschli-

ches Interleukin-3 (IL-3) und Granulozyten-Makrophagen-Kolonie-stimulierenden Faktor

(GM-CSF) exprimieren, sogenannte myeloid-verbesserte NOG-EXL- und NSG-SGM-3-

Mäuse.

In dieser Arbeit wurde eine umfassende Charakterisierung und ein direkter Vergleich

dieser neuartigen Mausmodelle durchgeführt. Dabei lag der Schwerpunkt auf der myeloi-

schen Rekonstitution, um festzustellen, welches Modell für Studien mit menschlichen

myeloischen Zellpopulationen am besten geeignet ist. Es wurde eine eingehende Charak-

terisierung des Humanisierungsniveaus hinsichtlich der myeloischen Rekonstitution im pe-

ripheren Blut sowie in primären und sekundären lymphoiden Organen durchgeführt. Dabei

fand sich, dass die neuartigen Modelle NOG-EXL und NSG-SGM3 einen Anstieg an myelo-

ischen Zellpopulationen im peripheren Blut sowie in primären und sekundären lymphoiden

Organen aufweisen. Neben Monozyten und Makrophagen wurden auch erhöhte Level für

Zusammenfassung

14

humane dendritische Zellen gefunden. Es wurde jedoch nicht nur das Auftreten dieser Zell-

populationen analysiert, sondern auch deren Funktionalität. Dabei wurde festgestellt, dass

sowohl Makrophagen als auch pDCs von huNOG-EXL funktionell sind, da sie mit TLR A-

gonisten auf exogene Stimuli reagieren können. Somit ist die hier vorliegende Arbeit die

erste, die funktionelle humane Makrophagen und pDCs in huNOG-EXL-Mäusen beschreibt.

Zusätzlich wurde die adaptive Immunität analysiert. Da ein beeinträchtigter IgG

Klassenwechsel bei klassischen humanisierten Mäusen eine häufige Einschränkung dar-

stellt, wurden in dieser Studie auch die hier analysierten Modelle makroskopisch auf Lymph-

knoten- und Thymusentwicklung verglichen. Zusätzlich wurde die Antigen-spezifische IgG-

Produktion untersucht und hohe Level in den myeloischen verbesserten Modellen gefun-

den.

Zusammengefasst wiesen die analysierten Mausmodelle erhöhte Level myeloischer

Zellen in der Peripherie auf. Es wurde zudem berichtet, dass diese Zellen eine zentrale

Rolle in einer Vielzahl von soliden Tumoren spielen. Daher konzentrierten sich weitere Un-

tersuchungen auf die Infiltration der myeloischen Immunzellen in Tumoren. Dies ist ein

wichtiger Aspekt bei der Krebsimmuntherapie, da die Häufigkeit myeloischer Zellen in soli-

den Tumoren oft mit einem schlechten klinischen Ergebnis korreliert. Myeloid-verbesserte

humanisierte Mäuse wurden folglich mit humanen Xeno-Tumoren transplantiert und das

Tumorimmuninfiltrat analysiert. Da der Tumor selbst zum humanen Zytokinmilieu beiträgt,

wurde die Frage aufgeworfen, ob die Zusammensetzung der Tumorimmunzellen bei huma-

nisierten Mäusen stärker vom verwendeten Mausstamm oder vom gewählten Tumormodell

beeinflusst wird. Die Untersuchung des relativen Einflusses von transgenen Zytokinen bzw.

von Tumoren-abgeleiteten Zytokinen auf die Rekonstitution myeloischer Zellen bei huma-

nisierten Mäusen war von zentraler Bedeutung in dieser Studie.

Zusammenfassend wurden huNOG-EXL-Mäuse als robuster Mausstamm zur Un-

tersuchung der Immunität des Menschen identifiziert. Mit erhöhten IgG-Spiegeln und der

Rekonstitution wichtiger Immun-Subpopulatoinen in verschiedenen Tumormodellen bieten

sie ein klinisch relevantes Modell für die Krebsimmuntherapie. Darüber hinaus ermöglicht

der relativ hohe Prozentsatz menschlicher pDCs die Untersuchung dieser seltenen Zellpo-

pulation ohne zusätzliche Manipulation. Es wurden spezifische Tumormodelle identifiziert,

die in der Lage sind, funktionelle Tumor-assoziierte pDCs zu erzeugen, die auf die TLR-

Aktivierung reagierten und nicht auf der Zytokinexpression von Transgenen beruhen. Wäh-

rend die Wahl des Mausmodells für die Differenzierung angeborener Immun-Subpopulatio-

nen in der Peripherie wesentlich war, ging dieser Einfluss verloren wenn man sich auf die

Tumor-Mikroumgebung fokussierte. Von Tumoren-stammende Zytokine können möglicher-

weise Zellen des angeborenen Immunsystems anziehen und differenzieren, selbst wenn

Zusammenfassung

15

ihre Anzahl in der Peripherie gering ist. So wurde erstmals gezeigt, dass das Tumor-Milieu,

nicht aber der Maushintergrund die Häufigkeiten seltener innater Immunzell-Subpopulatio-

nen wie pDCs im Tumor definiert.

Introduction

16

1. Introduction

1.1 The innate and adaptive immune system

The human immune system is a highly complex network and serves not only as protection

against bacteria, viruses and other pathogen but it is also able to recognize mutated endog-

enous cells such as tumor cells. In humans and many other species the immune system

can be sub classified in two main categories, innate or non-specific and adaptive or acquired

immunity. The innate system contains primary barriers that are of physical or chemical na-

ture, like skin or mucosal membranes, which act as a protective shield by separating the

external surroundings from the internal (6). When pathogens have overcome those barriers

the immune system needs to recognize them as non-self and dangerous. They distinguish

between non-self and self by Pathogen-associated molecular patterns (PAMPs), which are

recognized by the pattern recognition receptors (PRRs) of the innate immune cells. There-

fore, the innate immune system contains additionally soluble factors as well as membrane

bound receptors to recognize and bind invading pathogens. This response is fast and rather

unspecific. It identifies pathogens quickly, destroys them and is therefore called the first line

of defense. Additionally, some cells like macrophages are capable to produce cytokines,

which can further promote or suppress the immune activity (6). In comparison, the adaptive

immune system acts relatively slow, but it is specifically directed against antigens of the

pathogens (6). So called T and B lymphocytes express different antigen receptors on their

cell surface. The high number of lymphocytes guarantees that almost any antigen is even-

tually recognized by one of them. Additionally, the adaptive immune system is able to pro-

duce antibodies against the recognized antigens and it further contains a memory function

so that a second infection by the same pathogen is eliminated even faster. Although, these

two immune cell responses differ from another in cell types, responsiveness and timelines,

they also interact closely together and depend on each other (6,7). Both types of the im-

mune system contain many subsets of immune cells or so called leukocytes, which originate

in the bone marrow, as descendants from pluripotent hematopoietic stem cells (HSCs) (Fig.

1) (8). During hematopoiesis, HSC mature partly or completely within the bone marrow and

leave into the blood stream or specific tissues. There are two main lineages of cells, which

also define the innate and adaptive compartments. The common myeloid progenitor (CMP)

cells further differentiate in macrophages as well as granulocytes (neutrophils, eosinophils

and basophils), which belong to the innate immunity and are equipped with PRRs. Moreo-

ver, the CMPs give rise to erythrocytes, megakaryocytes and platelets. T, B, NK and NK-T

cells on the other side originate from the common lymphoid progenitors (CLP) and are

mainly associated with the adaptive immunity excluding NK cells. Human dendritic cells

(DC) develop from either CMPs or CLPs depending on their subtype. Human DCs can be

Introduction

17

divided into three subclasses: Classical CD1c+ DCs (cDC2), cross-presenting CD141+ DCs

(cDC1) and plasmacytoid CD303+ DCs (pDCs) (9-11). While cDC2 DCs are central in the

stimulation of Th2 and Th17 reactions in the fight against extracellular pathogens, cDC1s

can digest dead cells via CLEC9A and are primarily known to cross-present peptides via

MHC class I to CD8+ T cells (9). Both subsets leave the bone marrow as precursors and

mature in the periphery. Plasmacytoid dendritic cells (pDCs) on the other hand leave the

BM as mature cells and are therefore among the cells that are an important link between

the innate and adaptive immune system. Moreover, pDCs are capable to produce high lev-

els of all type I interferons in reaction to stimulation with toll-like receptors (TLR) 7 and 9

and hence contribute to antiviral immunity (12).

Macrophages and neutrophilic granulocytes are activated very early in an immune re-

sponse. Exogenous antigens are taken up by phagocytosis of these innate cells and are

further processed by proteases. Subsequently, they are presented on the surface of these

so called antigen-presenting cells (APCs) by a peptide called major histocompatibility com-

plex class II (MHC-II) (6). The APCs in particular classical DCs with their cross-presenting

features are then moving to the second lymphatic organs (spleen, lymph nodes) and pre-

sent their antigen to T lymphocytes. Thus and by the expression of co-stimulatory mole-

cules, T cells are subsequently activated, differentiate and an antigen specific immune re-

sponse follows.

FIGURE 1: HIERARCHY OF HUMAN HEMATOPOIETIC STEM CELLS. Adapted from Larsson et al. (8). HSC = hematopoietic stem cells, LT = long term repopulating HSC, ST = short term repopulating HSC, CMP = common myeloid progenitor, CLP = common lymphoid progenitor, MEP = megakaryocyte/erythroid progenitor, GMP = granulocyte-macrophage progenitor

Introduction

18

T cell precursors derive from the bone marrow, but ultimately develop within the thymus.

That is why they are also called thymocytes. Here they go through a series of maturation

steps, which can be identified on the basis of the expression of various cell surface markers.

The early thymocytes lack CD4 and CD8 expression and are designated as double negative

(DN) cells. In short, by rearranging the T cell receptor chains, they upregulate these expres-

sion markers and become double positive (DP) cells. DP cells need to be successfully se-

lected positive and negative before they can leave the thymus. A positive selection occurs

when DP T cells bind MHC class I or class II cells and thereby receive a survival signal.

Cells that fail to bind die during this process. On the other hand, a negative selection takes

place if DP T cells bind too strongly to APCs which express MHC class I or class II self-

antigens. They receive an apoptosis signal to avoid self-reactive T cells that are able to

induce autoimmunity. After this selection, down-regulation of one of the two co-receptors

produces either naive CD4 or CD8 single positive cells that leave the thymus and circulate

in the periphery (6,13). In the periphery, T cells can be divided into CD8+ cytotoxic, CD4+

helper and regulatory T cells, which all fulfill different tasks. T helper cells assist other im-

mune cells e.g. B cells in their activation or maturation. Cytotoxic T cells destroy previously

recognized abnormal cells and regulatory T cells (Tregs) are responsible to down-regulate

the T cell mediated immunity after an immune response. Therefore, they are also called

suppressive T cells (6). Furthermore, free antigens may also come across B lymphocytes

and thereby will activate them. The B cell will differentiate into a plasma cell and start pro-

ducing antibodies with the same antigen specificity. B cells are also able to form memory

cells.

As all these pathways and interactions are complicated and many cells and factor are in-

volved, the immune system is tightly regulated. Distributed over the different immunological

signal pathways there are so called checkpoints that regulate the immune responses in

order to allow fast elimination of pathogens but avoid an overreaction of the immune system

(6).

1.1.2 The role of immune cells in tumor development

The hypothesis of tumor immune surveillance and escape was described already 100 years

ago by Paul Ehrlich who provided first evidence that tumor proliferation is related to host

resistance. He already spoke about tumor immunity and suggested that the immune system

might play a role in tumor proliferation (14). In the following decades researchers tried to

understand the role of the immune system in cancer development, but with little success.

In the 1970s, Thomas and Burnet hypothesized that immune surveillance existed, but un-

fortunately could not find any clear evidence (15). Almost a century after Ehrlich's theory,

Introduction

19

new findings revolutionized cancer research and fueled the debate as to whether the im-

mune system is involved in tumor growth or not. In 2002 Dunn et al. showed that there is

indeed significant evidence for immunosurveillance and even for immunoediting, meaning

that the immune system does not remain inactive while the tumor grows, but the two com-

municate. They classified immunoediting in three phases: elimination, equilibrium and es-

cape. In the elimination phase immune cells identify and abolish mutant tumor cells. If the

first phase fails, the equilibrium phase follows. In this second phase the immune cells are

able to regulate the tumor growth and tumor cells may become latent for an indefinite period.

Some cells might not endure this phase, while others may be promoted and so the compo-

sition of the tumor cells is formed by the immune system. Tumor cells, which survive the

equilibrium, enter the third phase and have already developed one of various escape strat-

egies (15-18). This hypothesis gained broad acceptance and in 2011, Hanahan and Wein-

berg added “evading immune destruction” to their extended version of hallmarks of cancer

(19).

These pioneering findings led to a stronger focus on immune cell characterization in

human tumors and it was found that all immune cell types may be present in a tumor. Their

location, combination and functionality was said to be critical in host reaction to the tumor

and defined as ‘immune contexture’ (20-22). Moreover, the immune infiltrate was shown to

be heterogeneous not only between different tumor types but also between patients with

the same tumor type. Additionally, these immune cells are accompanied by chemokines

and cytokines, which further impact the tumor immune contexture.

1.1.2.1 Lymphoid immune cells in tumors

At first, studies focused mainly on lymphoid cells, in particular T cells, as they were detected

to be involved in antitumor immune response within various solid tumors (21,22). A clear

correlation was found between T cell infiltration into tumors and clinical outcome in many

cancers. Whereas high numbers of cytotoxic CD8+ T cells were associated with a good

prognosis, Treg infiltration was associated with poor clinical outcome as was shown first for

ovarian cancer patients and later for many cancer types (20,23).

1.1.2.2 Myeloid immune cells in tumors

Emerging evidence confirmed that myeloid cells were among the most abundant

tumor infiltrating immune cells playing a critical role in tumor development (24). To date,

tumor-associated macrophages (TAM) are the best analyzed myeloid population within tu-

mors, probably because they have been confirmed to boost tumor regression as well as

tumor growth, depending on their polarization (25,26). Several environmental factors, such

as the local cytokine milieu, affect the phenotype and function of macrophages. Tumors

Introduction

20

have been reported to shape their own tumor microenvironment (TME) also by manipulation

of TAMs leading to suppression cytotoxic T cells responses (26). Macrophages are typically

described as classically activated M1 macrophages, often associated with tumor regression

and alternatively activated M2 macrophages, linked to tumor progression (27,28). However,

it became clear that these phenotypes are only two of many possible activation states of

macrophages and that their characterization is far more complex as they are able to react

towards many different stimuli (27,29). Still, their contribution to tumor development is cru-

cial and in many cases associated with poor clinical outcome (24,27,30,31)

Another cell type that was found to correlate with survival of cancer patients in clini-

cal trials are plasmacytoid dendritic cells pDCs (32-34). In general pDCs are among the

cells that are capable of linking the innate to the adaptive immune response and are the

major type I interferon producers via (TLRs) 7 and 9 (32,35). They also play a pathogenic

role in autoimmune diseases, such as systemic lupus erythematosus (36-38) and in chronic

viral infections as Hepatitis B (38-40). In particular they were shown to contribute to tumor

progression by immune suppression in ovarian and breast cancer as well as melanoma

(10,41,42).

Besides TAMs and pDCs, the majority of myeloid immunosuppressive cells that ac-

cumulated in tumors are characterized to be immature. As it cannot be confirmed that all

these cells are of normal myeloid precursor origin they were termed myeloid-derived sup-

pressor cells (MDSCs) (24,43). Although, these MDCSs share some characteristics with

common myeloid cells, they stand as a unique population of tumor infiltrating innate immune

cells (24). In summary, tumors and the TME are capable to drive myeloid cells to become

immunosuppressive and pro-tumorigenic through various factors.

1.2 Cancer immunotherapy

The understanding of immunoediting mechanisms and the more detailed characterization

of tumor-infiltrating immune cells by new technological features such as multi-color flow

cytometry, CYTOF or RNA sequencing gave rise to new hope of using the immune system

in the fight against cancer. Today there are many strategies how the host’s immune system

can be redirected to kill the tumor. Based on the immune status of the patient’s tumor, can-

cer immunotherapy can be broadly grouped in three main approaches: Active, passive and

the combination of both (44-46).

Patients with an intact immune system are potential candidates for the active strat-

egy in which host immune cells are directed to bind to tumor-associated antigens (TAAs)

such as proteins or carbohydrates. Cancer vaccines are a promising approach in the active

cancer immunotherapy category. They can be used either in a preventive or therapeutic

setting and are based on tumor peptides as well as DCs or allogeneic whole cell vaccines.

Introduction

21

Cancer vaccines are composed of tumor cell material or TAAs to stimulate an immune re-

sponse. Vaccine specificity as well as tumor heterogeneity are two major points that need

to be considered when planning to use cancer vaccines in patients (44-48).

Another active approach is the use of so called checkpoint inhibitors. Identification

of immune checkpoints and their important role in immune suppression in cancers revived

the enthusiasm for the field of cancer oncoimmunology (47). Cytotoxic T lymphocyte antigen

4 (CTLA-4) was the first checkpoint that was correlated with tumor immune escape. It is

mainly expressed on activated T cells, binds on the same ligands as CD28 but acts contra-

rily. Thus, CTLA-4 is a negative checkpoint that plays an important role in the regulation of

activated T cells. In 1996 Allison et al. proposed that the annulment of CTLA-4 induced

inactivation of T cells may result in antitumor immune response (48,49). Today’s most im-

portant checkpoint inhibitor pathway involves the surface antigen PD-1 and its ligand PD-

L1. Whereas PD-1 is mainly expressed on the surface of activated, tumor reactive T cells,

PD-L1 was shown to be highly expressed by tumor cells and suppressive immune cells in

the TME. As binding of PD-L1 to its receptor renders T cells inactive, tumors can efficiently

escape immune responses by constitutively or transiently upregulating its expression. The

clinical efficacy of therapeutic antibodies blocking the PD-1/PD-L1 axis finally confirmed the

essential role of immune checkpoints in tumor immune surveillance (44-47). Driven by the

success of these two checkpoints various other receptors were identified and are still under

investigation in preclinical and clinical trials. To date, the CTLA-4 inhibitor Ipilimumab, the

PD-1 blocking antibodies Nivolumab or Pembrolizumab and the PD-L1 blocking antibody

Atezolizumab are the only approved immune checkpoint inhibitors for cancer treatment

(45,48).

The last method in active cancer immunotherapy is the application of oncolytic vi-

ruses. These engineered viruses selectively replicate in tumor cells and kill them by lysis

without damaging normal cells. Additionally, they induce a systemic tumor-specific immunity

by releasing more virus and TAAs after successful elimination of tumor cells (44,48,50). A

major advantage of this approach is that these viruses can be engineered to express spe-

cific cytokines that promote immune cell recruitment and activation, or to produce T cell co-

stimulating molecules on infected tumor cells. As a result, the anti-tumor effect is even fur-

ther enhanced. However, there are also many limitations such as that oncolytic viruses

might be harmful to immunocompromised or late stage cancer patients. Additionally, meta-

static cancer patients must also be excluded due to the local administration of the virus.

And if the virus is to be administrated systemically, toxicity can be a problem. However, a

modified herpes simplex virus named Talimogene laherparepvec (T-VEC) was approved by

the FDA for treatment of melanoma and shows in particular promising results in combination

with other therapies (44,50).

Introduction

22

As compared to the described examples of active immunotherapy, the concept of passive

immunotherapy aims to increase the hosts own immune response. This approach is often

used when patients have a weak or low responsive immune system and might be achieved

by using monoclonal antibodies, cytokines or donor T cells (44,48). Monoclonal antibodies

are developed and used for decades and are a major pillar in the treatment of many cancers

(48). They are targeted against tumor-associated proteins and cause an anti-tumor re-

sponse by either antibody dependent cellular cytotoxicity (ADCC) or complement depend-

ent cytotoxicity (CDC). Moreover, they can also induce direct cell death or function by block-

ing prosurvival signaling or angiogenesis. However, they show large differences in the pa-

tient's response rate, since their success depends on the expression of the tumor-associ-

ated protein. Their unique specificity also makes it challenging because they only recognize

specific epitopes and are unable to bind to e.g. mutated forms (44,48,51). The most promi-

nent example is Rituximab, an antibody against CD20, which is currently approved for treat-

ment of chronic lymphocytic leukemia (CLL) and non-Hodgkin lymphoma (44,46,48,51).

The essential role of cytokines in cancer immunotherapy became evident early on,

as cytokines are secreted by immune cells and can directly influence one another. By simply

administrating cytokines to patients an anti-tumor immune response can be induced. At the

moment IL-2 and Interferon-α (IFN- α) are approved by the FDA for melanoma and renal

cell carcinoma. However, their use constitutes the permanent risk of a strong systemic in-

flammation and based on the relatively low response rates their use is restricted or they are

combined with other treatments (45,48,51).

Adoptive cell therapy (ACT) is a promising approach in which patients are reinfused

with their own immune cells upon massive ex vivo expansion. Tumor infiltrating immune

cells derived directly from e.g. a metastatic lesion have the advantage that they can be

cultured outside the tumor's immunosuppressive environment before reinjection into the

patient. Additionally, improvements in T cell engineering has broaden the opportunities for

ACT in the last years. The development of chimeric antigen receptor T cells (CAR-T) with

increased specificity and antitumor mechanisms has revolutionized the field. They act HLA

independent and are not in need of an operable tumor in the beginning. However, precon-

ditioning of the patients with chemotherapy or whole body irradiation to enhance the effec-

tiveness of the transferred T cells also has severe side effects and should be evaluated

carefully. Additionally, toxicity of CAR-T cell approaches is a big risk as high potency of

engineered T cells could result in massive cytokine release limiting therapeutic windows

(44-47).

A more elegant way that does not need a harsh precondition may be the therapeutic

use of bispecific antibodies that are able to interact with the T cell on the one site and a

Introduction

23

TAAs simultaneously. Further progress in T cell engineering recently led to improved T cell

bispecific antibodies (TCBs) with lower side effects that are currently under clinical investi-

gation (44,51).

TABLE 1: CATEGORIES OF CANCER IMMUNOTHERAPY

Passive Active

Tumor-specific monoclonal antibodies Vaccines (peptides, DCs or allogeneic whole

cell)

Cytokines Checkpoint inhibitors

Adoptive cell transfer Oncolytic viruses

As outlined before, there are many different approaches currently being evaluated

in the field of cancer immunotherapy. As cancer is a heterogeneous disease in which tumor

type, progress of the disease, patient’s age and immune status are very much different from

patient to patient, combinations of different drugs and concepts will be required. Given the

complexity of tumor and immune system interaction, additional investigations but also more

appropriate pre-clinical models are needed (45). Up to now, primarily syngeneic mouse

models were used for preclinical tests of immune checkpoint inhibitors. Nevertheless, there

are significant differences between the human and murine immune systems, and therefore

not all interactions and therapeutic strategies can be examined in these mouse models

(52,53). To fully understand the interaction between a human tumor and the human immune

system a preclinical model with both features is required.

1.3 Promising in vivo models for cancer immunotherapy: humanized mice

1.3.1 Background

Immunodeficient mice, engrafted with human hematopoietic stem cells, so called “human

immune system” (HIS) or humanized mice have become an indispensable tool in the inves-

tigation of many human diseases such as autoimmunity, transplantation medicine and in-

fections, particularly in HIV research (54,55).

The humanization of mice was not possible due to the lack of appropriate recipient mouse

strains until a couple of decades ago. In the early 1980s new findings led to the development

of the first HIS mouse: The mutation in the protein kinase, DNA-activated, catalytic poly-

peptide (prkdc) gene resulted in a severe common immunodeficiency and development of

the SCID mouse. These mice lack both T and B cells and thus allow the successful engraft-

ment of human cells (52,53,56,57). Simultaneously, non-obese diabetic (NOD) mice, which

are known for their impaired innate immunity, were described. By crossing NOD and SCID

Introduction

24

mice an even more immunodeficient strain was generated deficient in both innate as well

as adaptive immunity (52). Through the introduction of a mutation in the interleukin-2-re-

ceptor γ chain (IL-2rγ) gene, NK cell development was additionally compromised (52,53,56).

Almost at the same time the BRG mouse was developed on the background of BALB/c-

Rag2 mice. These mice have a germline mutation resulting in a deletion in the coding region

of the recombination activating gene (Rag)2, which encodes for the eponymous protein that

is important in the V(D)J recombination during B and T cell development. Consequently,

these mice fail to produce murine mature B or T lymphocytes, thus resulting in a SCID-like

phenotype (58). The development of these immunodeficient mice finally allowed the gener-

ation of humanized mice by injecting human donor cells or tissue (59).

Besides infections and autoimmunity, humanized mice offer a great opportunity for

cancer research, as oncology is a particular challenging field for developing new drugs (4,5).

This is partly due to the inadequate translation of preclinical findings into clinic trials (60).

With the rise of cancer immunotherapy approaches in the last decade, preclinical cancer

research is deeply in need of in vivo models with a functional immune system that is com-

binable with tumors (61). There are many differences in the human and the murine immune

system, so humanized mice offer a great possibility to overcome this challenge and to in-

crease predictability and translation from pre-clinic to clinic. One of the major advantages is

the use of human specific therapeutics instead of murine surrogates, which are often not

available (61,62). Additionally, it is possible to inoculate human tumor xenografts and thus

enabling cancer research to investigate the interaction between human tumors and the hu-

man immune system in a mouse model.

1.3.2 Classical humanized mouse models

The most commonly used mouse strains to develop humanized mice are BRG, NOG and

NSG mice. Upon myeloablation and transplantation of human hematopoetic stem cells

these strains closely reflect human lymphopoesis as seen by thymic selection of T cells and

B cell maturation in the bone marrow (54,59). One important difference between the NOD

and the BALB/c background is a polymorphism in the mouse signal regulatory protein α

(SIRPα) (63). In general, mouse SIRPα, a protein expressed mainly on monocytes and

macrophages, is able to recognize human CD47, a marker of “self” (the so called “don’t eat

me” signal), which is also expressed on HSCs. Binding of mouse SIRPα to human CD47

inhibits murine macrophage phagocytosis and human HSCs are not destroyed. Due to the

polymorphism murine SIRPα is not capable to recognize human CD47. Consequently, en-

graftment of human HSCs is better in mice where a functional SIRPα to CD47 binding takes

place (64). Mice transgenic for human SIRPα or mice with enhanced murine SIRPα function

were developed and showed improved engraftment rates (65). But not only the choice of

Introduction

25

the mouse strain results in various differences. The establishment of HIS mice differs sig-

nificantly between the individual research groups. There are many experimental parameters

which result in different engraftment levels: The use of newborn or adult mice, HSCs or

peripheral blood mononuclear cells (PBMCs) as donor cells, different injections methods,

several ways of pretreatment, etc. (53,56,57,59,63). Figure 2 gives an overview of the dif-

ferent methods, which are used to generate humanized mice.

FIGURE 2: DIFFERENT METHODS TO GENERATE HUMANIZED MICE. Adapted from Theocharides et al.(59), NOG = NOD-Scid Il2rγ -/-, NSG = NOD-Scid Il2rγ -/-, BRG = Balb/c Rag2−/− Il2rγ -/-, PBMCs = peripheral blood mononuclear cells

Most research groups use umbilical cord blood HSCs for the reconstitution as they are com-

mercially available. But also fetal liver cells are frequently used, despite the ethical limita-

tions that are mandatory to consider. Many groups acknowledge the fact that the humani-

zation of newborn mice results in better reconstitution than adult mice but that can be logis-

tically difficult or if previous genotyping is required to select animals for the experiment.

Pretreatment to deplete mouse HSCs from the bone marrow niches can further improve

engraftment levels and is achieved either by irradiation or application of the chemothera-

peutic agent Busulfan (59,63).

Introduction

26

1.3.3 Limitations of classical humanized mice: cytokines are the key factor

Despite all the improvements in humanized mouse research in the last years, these human-

ized mouse models still have limitations. Some cell populations, such as erythrocytes

(56,66) and human NK cells but in particular myeloid cells are not differentiated adequately

(63,66), as these progenitor cells need external stimulatory factors to differentiate to their

target population. The main problem in humanized mice is the missing cross-reactivity of

murine and human cells and cytokines, which leads to an incompetent differentiation and

maturation of the majority of myeloid-derived cells (66). The introduction of human cytokines

in humanized mice was a key effort in the development of these models for the last decade

and led to the generation of various novel mouse models (53,56,59,63,66). Different meth-

ods from hydrodynamic plasmid injection (67) over transgenic mice (68,69) to knock-in mice

(66) were established. A few examples are described below.

It was found that missing cross-reactivity between human cells and murine IL-15 is the rea-

son why human NK cells do not differentiate in humanized mice (63). Consequently, admin-

istration of human IL-15 resulted in novel humanized mouse models, which support NK cell

differentiation. Already in 2009 Huntington et al. showed that the administration of combined

human IL-15 and IL-15 receptor α (IL-15R α) in BRG mice significantly increased human

NK cells (70). These findings were confirmed and expanded (71) and resulted in the devel-

opment of the transgenic NSG-Tg(Hu-IL15) mice (72) as well as the SRG-15 knock in mice

(73), which demonstrate increased development of functional human NK cells.

For limitations of erythrocyte development the NSGW41 mouse model was generated, sup-

porting human erythropoiesis and platelet formation by introduction of a loss-of function

mutation of KIT receptors in NSG mice. These mice showed significantly more human eryth-

rocytes and contributed to the knowledge that human macrophages are involved in the de-

mise of human erythrocytes in humanized mice. Remarkably, these mice also demonstrated

increased generation of myeloid cells (74,75).

NSG mice with a HLA-A2.1 transgene were originally developed during research concen-

trating on Epstein-Barr virus to improve protective T cell responses against human viral

infections (76). However, these mice also showed robust development of functional CD141+

and CD1c+ dendritic cells (11).

The administration of the human cytokines GM-CSF, CSF-1, IL-3 and SCF led to an im-

proved development of myeloid cells (63,66,67,77-79). Again, different groups developed

different models with transgenes or knock-in combinations of the cytokines mentioned. The

focus lied particular on GM-CSF and IL-3, as these have been shown to support the differ-

entiation and function of myeloid cells. Moreover, both cytokines are not cross-reactive be-

tween humans and mice (78). Therefore, both the NOG-EXL and the NSG-SGM3 mouse

Introduction

27

models were developed, which are to be compared in this work (68,69). Both mouse strains

share transgenes for human GM-CSF and IL-3 and demonstrated improved myeloid cell

differentiation (68,69).

One of the most sophisticated models is the MISTRG mouse of the Flavell group. They

performed a knock-in of the following human genes: CSF-1 (M-CSF), IL-3/GM-CSF, SIRPα

(transgene), Thrombopoietin (TPO) in Rag2-/- IL-2rγ-/- background and called it consequently

MISTRG. With this model, they were able to show improved myeloid cell differentiation in

blood and bone marrow but also myeloid immune infiltrate in transplanted human tumors

(66). However, these mice showed a reduced lifespan as well as anemia, indicating HSC

exhaustion (66,74).

Taken together, many of these novel models demonstrate that cytokines are a key factor to

improved differentiation of various immune cell populations in humanized mice. However,

their impact on the functionality is mostly unknown and many myeloid subpopulations are

not reconstituted at physiological levels. Therefore, further investigations and improvements

are needed to create humanized mouse models which more closely mimic the immune cell

composition of humans. To date, myeloid enhanced mouse models are already used in HIV

research as well as allergy studies or autoimmunity research. However, their use in cancer

research is still uncommon. This work investigates the characteristics of myeloid enhanced

humanized mice in the tumor context and provides new insights for their use in cancer re-

search.

Aim of the study

28

2. Aim of the study

Cancer immunotherapy is nowadays one of the best options left for many cancer patients.

Different tumor infiltrating myeloid cell populations were identified to play a role in different

cancer indications. Presence of tumor associated macrophages and myeloid derived sup-

pressor cells has been confirmed to correlate with poor clinical outcome. Consequently,

targeting of these tumor infiltrating myeloid cells is a promising approach. Nevertheless,

there are hardly any robust preclinical in vivo models for testing myeloid-targeting mole-

cules.

Due to the impaired differentiation of human myeloid cells in classical humanized

mouse, mouse strains expressing human cytokines IL-3 and GM-CSF were developed. The

aim of this study was to identify suitable preclinical models by comprehensive characteriza-

tion and direct comparison of two humanized, myeloid enhanced mouse models called hu-

NOG-EXL and huNSG-SGM3. The focus laid on their myeloid reconstitution to determine,

which model is most suitable for studies with human myeloid cell populations. Moreover,

further myeloid enhancement in huNOG-EXL mice was also investigated by using the CSF-

1 pathway through manipulation of the murine CSF-1 receptor.

Besides in-depth characterization of humanization levels in peripheral blood as well

as primary and secondary lymphoid organs, tumor immune infiltrating cells was prioritized.

Xenografted human tumors contribute to the human cytokine milieu in humanized mice

which shapes functionality and abundance of tumor-associated myeloid cells. Conse-

quently, the question was raised, if the tumor immune cell composition in humanized mice

is more influenced by the mouse strain used or by the chosen tumor model. Therefore, the

investigation of the relative contributions of transgenic and tumor-derived cytokines to my-

eloid cell reconstitution in humanized mice was of main attention in this study.

These findings bring new insights for the functionality of human myeloid cells and in partic-

ular human pDCs in huNOG-EXL in the periphery as well as in tumor context.

Materials and Methods

29

3. Materials and Methods

3.1 Materials

3.1.1 Antibodies

3.1.1.1 Flow cytometry antibodies

TABLE 2: FACS ANTIBODIES

Marker Clone

CCR7 (CD197) 3D12

CD11b M1/70

CD11c 3.9

CD123 7G3

CD14 HCD14

CD14 M5E2

CD141 1A4

CD16 3G8

CD163 Mac2-158

CD19 HIB19

CD1c F10/21A3

CD20 2H7

CD204 U23-56

CD206 19.2

CD25 2A3

CD3 OKT3

CD3 SK7

CD303 201A

CD33 HIM3-4

CD4 SK3

CD40 5C3

CD45 (human) HI30

CD45 (mouse) 30-F11

CD45 RA HI100

CD56 HCD56

CD62L DREG-56

CD66b G10F5

CD68 Y1/82A

CD69 FN50


30

CD8 RPA-T8

CD83 HB15e

CD86 BU63

CD8a SK1

CLEC9A 8F9

F4/80 BM8

FOXP3 206D

GZB QA16A02

HLA-DR G46-6

NKp46 (CD335) 9 E2

PD-1 EH12.1

TCR a/b IP26

TCR g/d B1

3.1.1.2 IHC antibodies

TABLE 3: PRIMARY ANTIBODIES

Antigen Reactivity Clone Manufacturer

CD3 Human 2GV6 Ventana

CD4 Human SP35 Ventana

HLA-DR Human CR3/43 DAKO

CD68 Human monoclonal Sigma

TABLE 4: SECONDARY ANTIBODIES

Antibody Manufacturer

Anti-rabbit Alexa647 Invitrogen by Thermo Fisher

Anti-mouse Alexa488 Invitrogen by Thermo Fisher

ImPRESS® Anti-mouse IgG HRP Vector Laboratories

ImPRESS® Anti-rabbit IgG HRP Vector Laboratories

3.1.2 Buffer and media

TABLE 5: MEDIA USED FOR IN VITRO EXPERIMENTS

Used for Basis Supplements

OVCAR-5 cell culture RPMI1640 10% FCS, 2mM L-glutamine

SK-OV-3 cell culture McCoy’s 5a 10% FCS, 2mM L-glutamine


31

SkBr3 cell culture EMEM 10%FCS, 2mM L-glutamine, 1mM sodium

pyruvate, 0.1mM NEAA

pDC cultivation RPMI1640 10% FCS, 2mM L-glutamine, 1x Penicil-

lin/streptomycin, 0.1mM 2-Mercapthoeth-

anol, 1mM Sodium pyruvate, 10mM

Hepes, 2ml Vitamins, 1x NEAA

MACS buffer 1 x PBS 0.5% BSA, 2mM EDTA

3.1.3 Cell lines

TABLE 6: CELL LINES

Cell line Supplier

OVCAR-5 (human ovarian carcinoma) Roche Diagnostics GmbH, Penzberg

SK-OV-3 (human ovarian carcinoma) Roche Diagnostics GmbH, Penzberg

SkBr3 (human breast carcinoma) Roche Diagnostics GmbH, Penzberg

3.1.4 Consumable material

TABLE 7: MATERIAL LIST

Material Manufacturer

1.5/ 2 ml Eppendorf tubes Eppendorf AG, Hamburg, Germany

10/100/200/1000 µl pipet tips Eppendorf AG, Hamburg, Germany

15/50/250 ml conical tubes Eppendorf AG, Hamburg, Germany

2/5/10/25/50 ml pipet boy tips Falcon, Corning, New York, USA

5 ml FACS tubes Falcon, Corning, New York, USA

6/12/24/96 well cell culture plates VWR International GmbH, Ismaning, Germany

Blood lancets B.Braun, Melsungen, Germany

BSA blocked NUNC streptavidin 384 well

plates Thermo Scientific, Schwerte, Germany

Cell culture bottles (T175=250ml) Greiner bio-one, Frickenhausen, Germany

Compensation beads anti mouse/rabbit/

hamster Invitrogen by Thermo Fisher

Cryo tubes 1ml VWR International GmbH, Ismaning, Germany

DAKO pen DAKO, Hamburg, Germany

Gentle MACS C tubes Miltenyi Biotec, Bergisch Gladbach, Germany

Injection needle (18G,26G,30G) B.Braun, Melsungen, Germany

Leucosep tubes PAN Biotech GmbH, Aidenbach, Germany

LS/LD magnetic separation columns Miltenyi Biotec, Bergisch Gladbach, Germany

MACS Smart strainers (30,70,100µm) Miltenyi Biotec, Bergisch Gladbach, Germany


32

Micro-FineTM insulin syringes (30G, 29G) BD bioscience, New Jersey, USA

Microtome blades Feather, Osaka, Japan

Microvette® 200 + 500 Z EDTA coated Sarstedt AG & CO, Rommelsdorf, Germany

Microvette® 500 Z-Gel Sarstedt AG & CO, Rommelsdorf, Germany

Precellys® Keramik-Kit 1.4 mm VWR International GmbH, Darmstadt, Germany

Scalpels B. Braun Aesculap AG, Tuttlingen, Germany

Superfrost slides VWR International GmbH, Ismaning, Germany

Syringes (0.5ml, 1ml, 2ml, 5ml) B. Braun Melsungen AG, Melsungen, Germany

96-well plates v-bottom Eppendorf AG, Hamburg, Germany

96-well plates u-bottom Thermo Scientific, Schwerte, Germany

3.1.5 Reagents

TABLE 8: REAGENT LIST

Reagent Manufacturer

1 x Phosphate buffered saline (DPBS) PAN Biotech GmbH, Aidenbach, Ger-

many

10x PBS Hoffmann-La Roche GmbH, Penzberg,

Germany

AMPUWA Fresenius Kabi, Bad Homburg, Ger-

many

Antibody Diluent Ventana Medical systems, Tucson,

USA

Antigen retrieval buffers pH 6 and pH9 DAKO, Hamburg, Germany

Bovine serum albumin (BSA) PAN Biotech GmbH, Aidenbach, Ger-

many

DAB+ chromogen + substrate buffer DAKO, Hamburg, Germany

DAPI solution Hoffmann-La Roche GmbH, Penzberg,

Germany

DMSO Sigma-Aldrich GmbH, Seelze, Germany

Dulbecco's modified eagle medium (DMEM) PAN Biotech GmbH, Aidenbach, Ger-

many

EDTA PAN Biotech GmbH, Aidenbach, Ger-

many

Ethanol VWR International GmbH, Ismaning,

Germany

Eukitt ® quick-hardening mounting medium O-Kindler, Freiburg, Germany


33

FACS buffer (MACSQuant Running Buffer) Miltenyi Biotec, Bergisch Gladbach,

Germany

FCS (fetal calf serum) PAN Biotech GmbH, Aidenbach, Ger-

many

Formaldehyde 4% VWR International GmbH, Ismaning,

Germany

Hämalaun Mayer Roth, Karlsruhe, Germany

Hydrogen peroxide 30% Merck Millipore, Darmstadt, Germany

Hyrogen peroxide solution 30% Sigma-Aldrich GmbH, Seelze, Germany

Isoflurane CP-Pharma, Burgdorf, Germany

L-Glutamine PAN Biotech GmbH, Aidenbach, Ger-

many

mCSF-1 Rat-5A1-IgG Hoffmann-La Roche GmbH, Penzberg,

Germany

MicroCoat, 384-Well MTP, clear Hoffmann-La Roche GmbH, Penzberg,

Germany

NEAA (100x) PAN Biotech GmbH, Aidenbach, Ger-

many

mCSF-1 Goat-IgG R&D Systems, Inc., Minneapolis, USA

rat anti-Goat-IgG-HRP conjugate R&D Systems, Inc., Minneapolis, USA

PenStrep (500 x) Hoffmann-La Roche GmbH, Penzberg,

Germany

Propidium iodide solution Miltenyi Biotec, Bergisch Gladbach,

Germany

Protein block serum free DAKO, Hamburg, Germany

RBC Lysis Buffer (eBioscience) ThermoFisher Scientifc, Braunschweig,

Germany

RPMI-1640 medium PAN Biotech GmbH, Aidenbach, Ger-

many

Sirius Red Stain Kit (Picro) Abcam, Cambridge, UK

Sodium pyruvate 100mM PAN Biotech GmbH, Aidenbach, Ger-

many

Sodium chloride B.Braun, Melsungen, Germany

StartingBlock T20 ThermoFisher Scientifc, Braunschweig,

Germany

Stemline 2 (HSC medium) Sigma-Aldrich GmbH, Seelze, Germany


34

Streptavidin-HRP Conjugate Hoffmann-La Roche GmbH, Penzberg,

Germany

TMB: BM Blue POD Substrate, soluble Hoffmann-La Roche GmbH, Penzberg,

Germany

Trypan blue Sigma-Aldrich GmbH, Seelze, Germany

Trypsin/EDTA PAN Biotech GmbH, Aidenbach, Ger-

many

Wash buffer 10x DAKO, Hamburg, Germany

Xylol Th. Geyer GmBH & Co KG

β-Mercaptoethanol PAN Biotech GmbH, Aidenbach, Ger-

many

Discovery Purple kit Ventana Medical systems, Tucson,

USA

DAB kit Ventana Medical systems, Tucson,

USA

Hematoxylin II Ventana Medical systems, Tucson,

USA

Bluing reagent Ventana Medical systems, Tucson,

USA

human Fc block Miltenyi Biotec, Bergisch Gladbach,

Germany

Mouse Fc Block CD16/CD32 BD bioscience, New Jersey, USA

Fixation/Permeabilization kit Invitrogen by Thermo Fisher

3.1.6 Instruments

TABLE 9: USED INSTRUMENTS

Instrument Manufacturer

Autotechnicon Leica Leica instruments GmBH, Nussloch

BD LSR Fortessa BD bioscience, New Jersey, USA

Bio-Plex Pro™ Wash Station Biorad, Munich, Germany

Bio-Plex® 200 Systems Biorad, Munich, Germany

Caliper Mitutoyo Europe GmbH, Neuss, Ger-

many

Coverslipper Leica instruments GmBH, Nussloch, Ger-

many


35

Cryolys-Kuhlmodul

VWR International GmbH, Darmstadt,

Germany

Cytoperm 2 incubator Heraeus

Electric shaver Harotec GmBH, Berlin, Germany

Eppendorf Centrifuge 5418 Eppendorf AG, Hamburg, Germany

Eppendorf Centrifuge 5418 Eppendorf AG, Hamburg, Germany

Eppendorf Centrifuge 5810R Eppendorf AG, Hamburg, Germany

Eppendorf MixMate® Eppendorf AG, Hamburg, Germany

gentleMACS. Dissociator

Milteny Biotec, Bergisch Gladbach, Ger-

many

Gourmet steamer Braun GmBH, Kronberg, Germany

Heating block Eppendorf AG, Hamburg, Germany

Isofluran vaporizer Eickenmeyer Medizintechnik KG, Tut-

tlingen

Laminar flow hood BDK, Sonnenbühl

MACS quant analyzer Milteny Biotec, Bergisch Gladbach, Ger-

many

MACS® cell separator Milteny Biotec, Bergisch Gladbach, Ger-

many

MACSmix. Tube Rotator


many

MACSQuantR Analyzer 10


many

Megafuge 1.0R Heraeus

Microscope Axiovert 10 Zeiss?

Necropsy instruments B. Braun AG, Melsungen, Germany

Operetta High-Content Imaging System PerkinElmer, Rodgau, Germany

Panoramic 250 Flash III digital slide scanner 3D Histech, Budapest, Hungary

Paraffin embedding machine Vogel GmbH, Gießen, Germany

Pipette (10-100 µl, 50-200 µl, 100-1000 µl) Eppendorf, Hamburg, Germany

PrecellysR 24/24-Dual . Homogenisator

VWR International GmbH, Darmstadt,

Germany

Rotary microtome, HM355S

Thermo Fisher Scientific, Braunschweig,

Germany

Surgical instruments

B. Braun Melsungen, Melsungen. Ger-

many


36

TECAN infinite 2000 Nano Quant plate rea-

der TECAN, Maennedorf, Schweiz

Thermometer Oregon Scientific, Oregon, USA

VENTANA Discovery Ventana Medical Systems, Tucson, USA

Vi-CELL cell viability analyzer XR Beckman Coulter, Krefeld, Germany

Vortex Genie 2 Scientific industries

Weighing scale BL 150 S Sartorius, Gottingen, Germany

Zeiss Axio Scan. Z1 Carl Zeiss AG, Oberkochen, Germany

3.1.7 Software

TABLE 10: SOFTWARE

Software Company

CorelDRAW 2019 Corel Corporation

Excel, Powerpoint, Word 2016 Microsoft Corporation

EndNote X8 Clarivate Analytics

FlowJo V10 FlowJo

Graph Pad Prism V7.04 GraphPad Software

Spotfire 7.11.1 Tibco

ZEN 2.3 lite Carl Zeiss AG

3.2 Methods

3.2.1 In vivo techniques

All animal experiments were performed according to Roche Diagnostics GmbH institutional

guidelines as well as the official national requirements for animal research. Female

NOD.Cg-Prkdcscid Il2rgtm1SugTg(SV40/HTLV-IL3,CSF2)10-7Jic/JicTac transgenic mice

(NOG-EXL) and non-transgenic NOD.Cg-Prkdcscid Il2rgtm1Sug/JicTac (NOG) mice were

purchased from Taconic bioscience (4-5 weeks old). Female NOD.Cg-Prkdcscid

Il2rgtm1Wjl Tg(CMV-IL3,CSF2,KITLG)1Eav/MloySzJ (NSG-SGM3) were acquired from the

Jackson laboratory at the same age. All mouse strains were held according to the applicable

animal protection law in a specific pathogen free (SPF) area. Mice were closely monitored

for body weight and general conditions.


37

3.2.1.1 Generation of humanized mice

Humanized mice were generated by precondition with a chemotherapeutic agent. Busulfan

(Busilvex®,Pierre Faber) was diluted with 0.9% sodium chloride (Braun) and 15 mg/kg was

intraperitoneally (i.p.) injected into mice. 24 hours later 1×105 human CD34+ umbilical cord

blood-derived hematopoietic stem cells (HSCs) (Stem Cell Technologies/Allcells) in 100 µl

phosphate-buffered saline (PBS) were transplanted into the mice via intra-venous (i.v.) tail

vein injection.

3.2.1.2 Blood and serum sampling

Blood sampling for flow cytometry and serum extraction was performed at different time

points after reconstitution and at termination using a lancet and puncturing the Vena facialis.

Blood for flow cytometry was collected in EDTA coated tubes to prevent aggregation. Sep-

aration of serum was carried out by centrifugation of whole blood in a gel microvette at

13000 rpm for 8 minutes at RT. Serum samples were stored at -20°C or at -80°C for further

analysis.

3.2.1.3 Tissue sampling

Mice were sacrificed by cervical dislocation of the neck before tissue collection. Depending

on the subsequent analysis, organs were stored in sterile PBS for flow cytometry, trans-

ferred to 4% formalin solution for immunohistochemistry (IHC) or directly frozen with cryo

tubes and liquid nitrogen.

3.2.1.4 Generation of humanized mice by depletion of mouse counterparts

Depletion of mouse macrophages was performed by weekly intraperitoneal injections of the

monoclonal anti-mouse CSF-1R antibody clone 2G2 into human immune system reconsti-

tuted mice. Control mice received a mouse IgG1 isotype antibody (MOPC-21) instead of

2G2. Different dosses and treatment schedules were analyzed as described in the respec-

tive figures.

3.2.1.5 Tumor growth in humanized mice

The human ovarian carcinoma cell line OVCAR-5 was purchased from the National Cancer

Institute (Cat. Nr. 0507676). The cell line SK-OV-3 is an internal cell line, which was tested

and authenticated to be SK-OV-3 after a full match with the reference databases of ATCC,

JCRB, RIKEN, KCLB and DSMZ. All cell lines were confirmed to be free of murine patho-

gens and murine viruses (Biomedical diagnostics, Hannover, Germany). To generate hu-

man xenograft tumors 3x106 OVCAR-5 or respectively 5x106 SK-OV-3 tumor cells in 100µl

PBS were injected in the right flank of isoflurane anesthetized humanized mice. The human


38

breast cancer patient-derived xenograft BC_038, a triple negative breast cancer was ob-

tained from Oncotest and transplanted in non-humanized NOG mice for three rounds before

used in this study. The tumor fragments were digested with Collagenase D and DNase I

(Roche), counted and injected into the mammary fat pad of humanized mice. Tumor growth

was monitored twice a week by perpendicular caliper measurement and tumor volume was

calculated using the following formula: volume = 0.5 × length2 × width. Animals were re-

moved from studies as soon as one of the following termination criteria was reached: Tumor

size > 2000 mm3, tumor ulceration, physical impairment or body weight loss > 20%. At day

of necropsy tumors were carefully removed and fixed in 4% formalin for immunohistochem-

istry or processed for flow cytometry.

3.2.2 Ex vivo / in vitro experiments

3.2.2.1 Preparation of human hematopoietic stem cells

Human CD34+ cord blood derived HSCs were purchased from Allcells or Stemcell technol-

ogies respectively. On the day of injection, vials were removed from liquid nitrogen and

quickly thawed using a 37°C water bath. Cells were carefully transferred using a 1000 µl

pipette in a 10 ml conical tube containing 5 ml cold HSC medium. Cryo vials were rinsed

with 1 ml HSC medium and suspension was added slowly to the cells. Cold HSC medium

was added until 10ml were reached and cells were centrifuged in a pre-cooled (4°C) centri-

fuge at 1300 rpm for 9 minutes. Supernatant was discarded carefully and pellet was resus-

pended in 37°C warm HSC medium. Cell viability was determined by trypan blue using a

hemacytometer and the formula: % viable cells = [1.00 – (number of blue cells ÷ number of

total cells)] × 100. The formula used to finally calculate the number of viable cells per ml

suspension was: Number of viable cells × 104 × 1.1 = cells/mL culture.

3.2.2.2 Preparation of single cell suspensions for flow cytometry

Blood

Equal volume of whole blood from each mouse was transferred from EDTA tubes to a 15

ml Falcon tube. For analysis of white blood cells, erythrocytes were lysed by adding 8 ml 1x

RBC lysis buffer for 10 minutes at room temperature. In general, blood was then centrifuged

at 1450 rpm for 5 minutes at 4°C. Cells were washed with 10 ml FACS buffer and centri-

fuged again at 1450 rpm for 5 minutes at 4°C. Cells used for flow cytometric analysis were

subsequently blocked with human and mouse Fc blocking agents and incubated with anti-

bodies. Optional RBC lysis was repeated if desired outcome was not achieved. Heparinized

human whole blood was collected from healthy volunteers at the Roche Medical Service in

Penzberg. Single cell suspensions from human blood were prepared like mouse samples

but without adding mouse FcR block.


39

Spleen

Spleens were removed from mice. Spleen fragments were first rinsed with FACS buffer

using a 26G needle. Remaining tissue fragments were pressed through a 100 µm cell

strainer into a 50 ml Falcon tube using the plug of a syringe. Filter was washed with FACS

buffer and cell suspension was centrifuged at 1450 rpm for 5 minutes. RBC lysis was per-

formed by resuspending the pellet in 5 ml 1x RBC lysis buffer and incubation at room tem-

perature for 5 minutes. RBC lysis was stopped by adding 10 ml PBS. Cell suspension was

filtered through a 70/30 µm filter mesh combination and centrifuged at 1450 rpm for 5

minutes. Pellets were resuspended in 10 ml PBS and cell number per ml was determined

using the MACSQuant flow cytometer. Cells were blocked with human and mouse Fc block-

ing agents and incubated with antibodies.

Bone marrow

Femur and tibia were removed from mice and joints were detached with sharp scissors.

Bones were placed with the bigger opening downward in a 0.5 ml Eppendorf tube containing

a small hole at the bottom. The 0.5 ml Eppendorf tube was closed and placed into a 1.5 ml

Eppendorf tube. Tubes were centrifuged at 10000 g for 20 seconds. The collected bone

marrow was resuspended in 200 µl PBS and placed on ice until further processing as de-

scribed in the section above (spleen).

Thymus and lymph nodes

Thymus and lymph node samples were processed like spleen and bone marrow but without

the initial rinsing step at the beginning. Lymph nodes were pooled.

Tumor

Tumor was cut to small pieces using two scalpels. Additionally, tumor mass was simultane-

ously homogenized in C-tubes with gentleMACS Dissociator from Miltenyi and incubated

with Collagenase D and DNase I (Roche). Tumor mass, bone marrow and spleen was

mashed through a combination of filters (100 µm, 70 µm and 30 µm; Miltenyi) and lysed

before samples were blocked for unspecific binding with mouse and human Fc Blocks (Bi-

olegend/Miltenyi). For intracellular staining cells were fixed for 20 minutes and permea-

bilized (Invitrogen) prior to adding the intracellular antibodies for another 20 minute staining.

3.2.2.3 Immunohistochemistry

For histological analysis a part of the organs from humanized mice was fixed in 4% para-

formaldehyde and embedded in paraffin. Paraffin sections (2µm thick) of tumors were cut

using a rotary microtome and dehydrated overnight at 37°C. The next day the sections were

deparaffinized in 100% xylene (3x5 min) and rehydrated in a descending ethanol dilution


40

series (2x 100%, 1x 90%, 1x 80% and 2x 70% ethanol, each for 2 min). Antigen retrieval

was completed by boiling the slides in citrate buffer pH 6.0 (DAKO antigen retrieval buffer)

at 97°C for 20 minutes. After 20 minutes of cooling, the slides were washed with water as

well as PBS and blocked for 10 minutes with a PBS + 0.3% H2O2 mixture. Unspecific binding

was then blocked by serum free block medium from DAKO for 20 minutes at room temper-

ature. The subsequent incubation of the primary and afterwards the corresponding second-

ary antibodies was conducted in a humidified chamber. For long term storage, sections

were covered with mounting medium. Therefore Eukitt was used for chromogenic staining

and Fluoro-Gel II with DAPI for fluorescence staining. For visualization of nuclei sections

were counterstained with Hematoxylin. Subsequently slides were washed with tap water,

dried and mounted with a cover slipper. Slides were scanned using a digital slide scanner

and analyzed with the ZEN software.

3.2.2.4 Phagocytosis Assay

Human macrophages from humanized mouse bone marrow were enriched by bead sepa-

ration from Miltenyi. Mouse CD45 positive cells were separated using anti mCD45 beads

from Miltenyi and MS Columns. 10 µl of bead suspension was used for 1x107 bone marrow

cells. 30 000 cells per Well in 200 µl media were seeded in 96 Well Plates. For differentiation

the media was supplemented with 0.02 µg/ml hCSF-1 and cultivated for 5 days. For stimu-

lation 10 ng/ml hIL-10 (R&D Systems) was added for 24h.

SkBr3 tumor cells were cultivated in media without cytokines. For optical detection the tumor

cells are stained adherent with 1 µM CMFDA in full media at 37°C for 45 minutes. Subse-

quently, supernatant was changed with fresh media. Stained tumor cells were harvested

and added to the macrophage culture – (5000 tumor cells per Well). Fc mediated phagocy-

tosis was initiated by antibody treatment ON. As positive control Herceptin (Roche) and for

isotype control human normal immune globulin (Privigen) was used. Additionally, IL-10 was

readded freshly during the phagocytosis assay.

For imaging the cells were stained with 1 µg/ml DAPI and 1 µg/ml PI one hour prior meas-

urement with the Operetta system. For distinction of mouse and human myeloid cell popu-

lation hCD163 (Alexa488) and mF4/80 (APC) staining was performed using 1 µg/ml anti-

body each.

3.2.2.5 ELISA for detection of total mouse CSF-1

To detect free mouse CSF-1 in sera of humanized mice 384-well microplates were coated

by adding 2 µg/ml anti-mouse CSF-1 coating antibody clone 5A1 in 30 µl PBS per well

followed by overnight incubation at 4°C on a plate shaker. Coated plates were blocked for

30 minutes at RT with 100 µl StartingBlock per well and washed trice in 100µl PBS-T. Seven


41

standards were prepared by serial 1:2 dilution of a mouse CSF-1 stock solution starting

from 0.16ng/ml. Serum samples of animals were diluted in StartingBlock and 30 µl of each

sample, standard and blank (StartingBlock) were transferred in triplicates to coated plates

and incubated for 20 minutes at RT. Plates were washed trice in 100 µl PBST and incubated

with 30µl biotinylated detection antibody prediluted in StartingBlock for 60 minutes at room

temperature. After another washing step in TBST, 30 µl streptavidin-HRP in 3% BSA/PBS-

T were added onto the plates and incubated for 45 minutes at RT. Plates were washed trice

in 100 µl PBST before 30 µl of BM blue substrate was added. Reaction was stopped by

adding 25 µl 1M H2SO4 and absorbance was measured at 450 nm using a plate reader.

3.2.2.6 Preparation of tumor lysate from frozen tissue

Fresh frozen tumor fragments were weighed and placed in Precellys tubes with ceramic

beads. Ice-cold lysis buffer was added and tumor fragments were homogenized with Pre-

cellys 24 Homogenizer (at 6°C, 2 times 30 seconds with 6000 rpm). Lysates were frozen in

liquid nitrogen and thawed on ice. Tubes were centrifuged at 4°C for 5 minutes at 10000

rpm. Supernatant was collected, protein concentration was determined by BCA and ad-

justed to 2000 µg/ml.

3.2.2.7 Bio-Plex multiplex immunoassay

To measure levels of different human cytokines in sera and tumor lysate of TLR agonist

treated or non-treated humanized mice Bio-Plex Multiplex immunoassay was performed.

Firstly, serum samples of humanized mice were diluted 1:4 and tumor lysates were diluted

1:2.5 (800 µg/ml; equates to 40 µg/well) in BioPlex dilution buffer and following steps were

performed according to the manufacturer’s protocol. Briefly, 50 µl of beads were added to

wells. After two washing steps, 50 µl standards, samples and blanks were added to the

wells and incubated for 30 minutes to 1 hour at RT using a shaker at 850 rpm. Plates were

washed three times in wash buffer before adding 25 µl of detection antibody. Subsequently,

the plates were incubated for 30 minutes at room temperature and washed again three

times. Then, 50 µl streptavidin-PE was added and incubated on a shaker at 850 rpm for 10

minutes at RT. After the last washing step 125 µl assay buffer was added and the plates

were incubated on a shaker for 30 seconds at 850 rpm and analyzed on the BioPlex plate

reader system. Cytokine concentrations were quantified on the basis of a unique standard

curve for each analyte. Cytokine levels in serum and tumor lysates from humanized mice

were measured using the Bio-Plex human cytokine 48 or 17-Plex kit as well as single anal-

ysis for IFN-α2 according to manufacturer’s protocol (Bio-Rad). Chemokine levels in serum

and tumor lysates from humanized mice were measured using the Bio-Plex system with

human chemokine 40-Plex kit (Bio-Rad). Human and murine FLT3-L were measured by

specific ELISA assays (R&D systems) (80).


42

3.2.2.8 Isolation of human PBMCs from human whole blood

Heparinized blood was collected from healthy volunteers at the Roche Medical Service in

Penzberg. Human peripheral blood mononuclear cells (PBMCs) were isolated from whole

blood by density gradient centrifugation by ready-to-use Leucosep tubes pre-filled with 15ml

PANCOLL human lymphocyte separation medium, according to the manufacturer’s instruc-

tions. Briefly, whole blood was diluted in equal volume of 1x RPMI 1640 medium and over-

laid carefully onto the porous barrier of the Leucosep tubes. Tubes were centrifuged for 20

minutes at 800 g with breaks off. After centrifugation the phase with the enriched PBMC

fraction was carefully removed and transferred into a 50ml falcon tube. Cells were washed

two times with 10 ml PBS and centrifuged for 10 minutes at 300 g.

3.2.2.9 Isolation of human pDCs and ex vivo stimulation

Bone-marrow derived cells from up to ten humanized NOG-EXL mice were used for pDC

isolation with a customized isolation kit from Stemcell Technologies after manufactures in-

structions. Briefly, cells were incubated with normal rat serum, pDC isolation cocktail, biotin

and rapidshperes. Subsequently, cells were placed in a magnet from Stemcell technologies

and the flow-through was collected. Enriched pDCs were than checked for purity and

counted. Human CD303+ CD123+ cells were used for ex vivo stimulation. Cells were stim-

ulated with 0.25 µg/µl of TLR agonists for 3 hours and supernatants were collected for cy-

tokine analysis

3.2.2.10 Immunoassay for human-specific IgG antibodies

To measure levels of IgG antibodies in sera of humanized mice an immunoassay was per-

formed, as formerly described by Kay Stubenrauch (81,82). In brief, anti-human Fcγ-pan

R10Z8E9 was digoxigenylated and subsequently biotinylated MAB anti-human Fcγ-pan

R10Z8E9 was bound to streptavidin-coated microtiter plates at a concentration of 0.5

μg/mL. After incubation for 1 hour the unbound antibody was removed by washing. Samples

and standards were pre-incubated with 0.05 μg/mL of digoxigenylated MAB anti-human

Fcγ-pan R10Z8E9 for 1 h. Thereafter, the mixture was added to wells of microtiter plates

coated with the biotinylated anti-human IgG antibodies and incubated for 1 hour. After wash-

ing, a polyclonal anti-digoxigenin-horseradish peroxidase conjugate (Roche, cat. no.

11633716001) was used to detect the bound digoxigenylated MAB anti-human Fcγ-pan

R10Z8E9. After incubation for 1 hour the HRP of the antibody–enzyme conjugates cata-

lyzed the color reaction of ABTS substrate. The signal was measured by plate reader at

405 nm wavelength (reference wavelength: 490 nm). Absorbance values of each serum

sample were determined.


43

3.2.2.11 Flow cytometry

Blood leukocytes were tested for human CD45, CD3, CD33, CD14, CD16, CD56 and CD19

(REAfinity/Miltenyi). Live/Dead staining was performed with propidium iodide (Miltenyi).

Eight-color flow cytometry analyses were performed on Miltenyi MACS Quant 10.

Splenocytes, BM-derived cells and tumor cells were tested in different panels for human

CD1c, CD3, CD4, CD8, CD11b, CD11c, CD14, CD16, CD19, CD25, CD33, CD40, CD45,

CD45RA, CD56, CD62L, CD66b, CD68, CD69, CD83, CD86, CD123, CD141, CD163,

CD204, CD206, CD279, CD303, CLEC9A, FOXP3, HLA-DR and Nkp46 in 18-color combi-

nations. Live/dead staining was performed with fixable ZOMBIE™ UV dye (BioLegend).

Antibodies and isotype controls were obtained from BD bioscience or BioLegend. 18-color

flow cytometry analyses were performed on BD Fortessa. FCS files were analyzed with

FlowJo (Version 10).

FIGURE 3: GATING STRATEGY FOR DC SUBPOPULATIONS

FIGURE 4: GATING STRATEGY FOR LYMPHOID CELLS


44

FIGURE 5: GATING STRATEGY FOR MYELOID CELLS Representative plots and gating strategy for myeloid cells, first five gates (morphology to exclude lineage) are the same as in Fig. 3.

3.2.2.12 Statistical analyses

Statistical analysis was performed using Prism 7.0 software (GraphPad Software). Normal

distributed data was expressed as mean ± standard error of the means (SEM) unless noted

otherwise. Nonparametric data was expressed as median ± interquatile range (IQR) and is

clearly indicated in the figure legend as such. The student t test was used to compare nor-

mally distributed two-group data. Ordinary one-way ANOVA with Tukey Multiple comparison

or Kruskal-Wallis ANOVA with Dunn’s multiple comparison was performed if more than two

groups were analyzed. Survival curve analysis was achieved using the log-rank (Mantel-

Cox) and Gehan-Breslow-Wilcoxon tests. Individual data points are shown if n < 5. Each

dot represents one individual mouse or analyzed sample. Statistical tests used are indicated

in the figure legends (80).

Results

45

4. Results

4.1 Analysis of human hematopoietic stem cells

To initially confirm that purchased human hematopoietic stem cells are capable of long term

engraftment and multi-lineage differentiation, HSCs from different donors were examined

by flow cytometric analysis. Only HSCs with a viability of more than 90% were chosen for

humanization of mice. Flow cytometry confirmed manufacturer specifications and revealed

that >90% of HSCs are expressing CD34+ and CD3- surface markers. An immature

CD90+CD45RA- HSC population as well as a CD90-CD45RA- differentiated multipotent pro-

genitor cell (MPP) population was found. Figure 6 shows a representative flow cytometry

density dot plot as example.

FIGURE 6: CHARACTERIZATION OF CD34+ HEMATOPOIETIC STEM CELLS. Neg = negative, pos = postitive, MPPs= multipotent progenitor cells. The CD34+CD38- population could be divided into three subpopulations: 1) CD90+CD45RA-, 2) CD90-CD45RA-, 3) CD90-

CD45RA+.

Results

46

4.2 Characterization of humanized NOG mice in comparison to myeloid enhanced hu-

manized mice

4.2.1 Generation of humanized mice and macroscopic differences

To compare the human immune system in conventional and myeloid enhanced humanized

mice, immunodeficient NOG mice as well as mice transgenic for hIL-3, hGM-CSF (NOG-

EXL) and additionally hSCF (NSG-SGM3) were humanized using a Roche internal protocol.

Briefly, for precondition, mice (5-6 weeks) were pretreated with Busulfan and injected intra-

venously with human CD34+ hematopoietic stem cells (HSCs). Body weight was monitored

and peripheral blood was screened biweekly to ensure successful human immune cell en-

graftment (Fig. 7A). Even though all mouse strains were humanized with the same protocol

(Fig. 7A), differences in body weight and survival were observed (Fig. 7B + C). Although

humanized NSG-SGM3 (huNSG-SGM3) showed the highest body weight from beginning

until end of analysis, they had a decrease in survival compared to other humanized mouse

strains (80).

FIGURE 7: DIFFERENCES IN BODY WEIGHT AND SURVIVAL A) Reconstitution protocol, graphic view: NOG, NOG-EXL and NSG-SGM3 (5-6 weeks old) were injected with Busulfan 24 h prior to intravenous injection of CD34+ human HSCs. Body weight was monitored and peripheral blood was screened from week 8 to 16 to analyze human hematopoiesis. At termination (week 16), different peripheral organs were analyzed for human leukocytes. B) Body weight kinetic over time: NOG, NOG-EXL and NSG-SGM3 were monitored from day 0 until termina-tion, n= 18 to 341, ordinary one-way ANOVA, Tukey’s multiple comparison test. ****, p < 0.0001. C) Survival curve comparison between individual humanized mouse strains, log-rank (Mantel-Cox) or

Results

47

Gehan-Breslow-Wilcoxon test yielded in the same statistical results, error bars indicate mean ± SD, n= 20/ strain, *, p < 0.05.

4.2.2 Human immune cell engraftment in different organs

4.2.2.1 Peripheral blood

The frequency of human CD45 positive cells in peripheral blood is known as reliable readout

to determine the success of humanization in immune system reconstituted humanized mice

(7,66,83). Here, flow cytometry was used to determine the humanization status of human-

ized NOG, NOG-EXL and NSG-SGM3 mice at different time points after humanization (Fig.

8A). The gating strategy is shown in Figure 7A. Detailed phenotyping of the human white

blood cells (WBCs) revealed the majority of human CD45+ cells in blood of huNOG mice

were B cells, whereas huNOG-EXL and huNSG-SGM3 presented higher myeloid cell fre-

quencies and numbers. When comparing both transgenic models differences in WBC com-

position were detected despite similar genetic backgrounds. HuNSG-SGM3 present a 2-

fold higher CD3+ T cell frequency in peripheral blood than huNOG-EXL mice (Fig. 8B). In

general, a large heterogeneity in frequencies and counts of myeloid cells was detected in

both transgenic models (Fig. 8C + D). These findings are consistent with previously pub-

lished studies on these two transgenic mouse strains (68,69).

Analysis over time revealed that huNOG-EXL and huNSG-SGM3 continuously exhibit

significantly higher humanization levels than conventional huNOG mice. Interestingly, in all

strains frequencies of some populations are expanding over time, e.g. CD3+ T cells,

whereas others are decreasing, such as CD19+ B cells. NK cells (CD56+) are detectable in

all three strains with only minor changes over the course of analysis. However, in both NOG

strains frequencies of NK cells are increasing with huNOG-EXL having 2-fold higher levels

in week 14 compared to huNOG, whereas the levels fall in huNSG-SGM3 over time. Myeloid

cells (CD33+/ CD14+) are present at higher frequency in myeloid enhanced humanized mice

throughout the entire time of observation (Fig. 8E).

Results

48

FIGURE 8: PERIPHERAL BLOOD ANALYSIS IN HUMANIZED MICE A) Flow cytometry gating strategy for detection of all relevant lineage markers, n= 28 to 40. B) Com-parison of human myeloid cells (CD33+) in human leucocytes (CD45+) in the same mice as in A as percentages, ordinary one-way ANOVA, Tukey’s multiple comparison test. *, p < 0.05; ***, p < 0.0005; ****, p < 0.0001, n= 10 to 39. C) Level of myeloid cells in week 12 and 16 as percentage. D) Total counts/µl of immune cells (CD45+) and myeloid cells (CD33+) in week 16. E, Different immune cell populations (CD45+ in single viable; CD3+, CD19+, CD56+, CD33+ and CD14+ in CD45+) in hu-NOG, huNOG-EXL and huNSG-SGM3 over time (week 8 to 14), 2-way ANOVA, Tukey’s multiple comparison test. *, p < 0.05; **, p < 0.001; ***, p < 0.0005; ****, p < 0.0001, n=28 to 121.

As mentioned before, high variation of humanization levels were observed among

individual animals (Fig. 9). To analyze if the observed discrepancy was HSC donor depend-

ent, humanization levels and WBC composition was compared between different batches

of humanized mice. A huge difference was detected in immune cell engraftment between

animals reconstitute from various donors in frequencies as well as in counts/µl. The median

range in overall humanization showed a 3-fold difference from 20% to 60% between differ-

ent donors. In mice generated of donors (D) with a low overall engraftment rate (hCD45%)

Results

49

also less CD3+ T cells were detected. Inconsistently, they expressed no reduction in CD33+

myeloid cells (e.g. D2, D5).

Next to inter-donor variablity a high heterogeneity was also observed in individual mice re-

constituted from the same HSC donor. As all mice were injected with the identical amount

of cells, one reason for the shown discrepancy in engraftment rate could lie in the clearing

of the bone marrow niches with Busulfan. Taken together, these huge variances between

donors but also in-between animals from the same donor, led to decision to randomize

animals for all experiments, using the donor origin as one criteria next to humanization level

and tumor size where applicable.

FIGURE 9: HETEROGENEITY IN ENGRAFTMENT BY HSC DONOR Comparison of total leukocytes (human and murine CD45+), T cells (CD3+) and myeloid cells (CD33+) in mice humanized with ten different donors (D) 1 to 10, n= 5- 13, week 16 after humanization.

Human myeloid cells in peripheral blood were further characterized in more detail.

Human monocytes can be subdivided in three subpopulations based on their surface

marker expression CD14 and CD16 (classical CD14+/CD16-, non-classical CD14-/CD16+,

intermediate CD14dim/CD16+) (84). Remarkably, humanized transgenic mice not only

showed more myeloid cells in total, but in particular huNOG-EXL showed higher frequencies

of CD14-/CD16+ and CD14/CD16 double positive monocytes, thus mimicking human mon-

ocyte composition more realistically. However, there is still a clear difference when com-

pared to human PBMCs (Fig. 10).

Results

50

FIGURE 10: ANALYSIS OF MONOCYTES IN PERIPHERAL BLOOD A) Human monocytes composition in total CD33+ in percentage of peripheral blood 16 weeks after reconstitution or in human PBMCs, n= 5 to 40. B) Representative flow cytometry density plots for human monocyte populations (CD16+/-/CD14+/- in CD45+) in peripheral blood of humanized mice in week 16 or human PBMCs.

Next to cellular components, the human cytokine milieu was assessed in serum of

humanized mouse models using Multiplex technology. Many cytokines and chemokines

(e.g. CCL-8, CCL-13, CCL21, CCL-23, CXCL9, CXCL10, CXCL12, IL-4, IL-5, IL-7, hFLT-3-

L, mFLT3-L) were detectable but at rather low levels or did not differ significantly (data not

shown). However, some cytokines were significantly elevated in myeloid enhanced mice

(Fig.11). The most significant increase compared to huNOG mice was detected for GM-

CSF in serum of huNSG-SGM3. This finding is in line with GM-SCF transgenic mice and

the breeders strain description. Of note, inflammatory cytokines such as IL-6, IL-8, MCP-1

and MIP-1β and accompanying contrary-acting suppressive cytokines such as IL-10 were

found to be elevated in transgenic mice but particularly in huNSG-SGM3 (Fig. 11), as de-

scribed (85,86).

Results

51

FIGURE 11: ANALYSIS OF HUMAN CYTOKINES/CHEMOKINES IN SERUM OF HU-MANIZED MICE Human cytokine/chemokine analysis by Multiplex in serum at termination (week 16), Two-tailed un-paired t-test, *, p < 0.05; **, p < 0.005; ***, p < 0.0005; ****, p < 0.0001; n= 4 to 16.

A major limitation of conventional humanized mouse models is the lack of antibody class

switch. Remarkably, as compared to huNOG mice, myeloid enhanced humanized mice and

in particular huNOG-EXL mice showed consistently high levels of human IgG 16 weeks

after humanization (Fig.12). Furthermore IgG molecules were also detectable as early as

week 10 albeit at lower concentration (Fig. 12).

Results

52

FIGURE 12: HUMAN IgG LEVEL IN SERUM OF HUMANIZED MICE Total human IgG and IgG1 antibody concentration in serum of humanized mice in week 10 and 16, n= 10 / strain, Kruskal-Wallis ANOVA with Dunn’s multiple comparison test, error bars indicate me-dian ± IQR, *, p < 0.05; **, p < 0.05.

In summary, higher overall engraftment and an increase in myeloid cells in periph-

eral blood of both transgenic mouse strains was confirmed in the present study. Further

characterization over time in peripheral blood additionally gave a clear overview, which cell

populations develop at a specific time point after humanization. huNSG-SGM3 present un-

usual high T cell levels and serum cytokines revealed increased levels of inflammatory cy-

tokines, indicating an inflammatory immune status. HuNOG-EXL show enhanced myeloid

cell differentiation, thus mimicking the situation in human blood more realistically. Addition-

ally, both transgenic strains presented physiological levels of human IgG antibodies in se-

rum, indicating that B cells undergo terminal maturation leading to improved functionality.

4.2.2.2 Primary lymphoid organs

Bone marrow

To characterize human immune cell composition in bone marrow of humanized mice, bone

marrow was isolated from humanized mice and single cell suspension was analyzed by flow

cytometry. The overall humanization in bone marrow showed significantly less human

CD45+ cells for NSG-SGM3 mice compared to huNOG and huNOG-EXL (Fig. 13A). An

increase in myeloid cells was detectable in both transgenic mice. However, in huNSG-

SGM3 this was also accompanied by a 3-fold increase of CD3+ T cells compared to huNOG-

EXL or huNOG mice (Fig. 13B). Deeper characterization of T cell subsets showed no dif-

ference in CD4/ CD8 ratio between humanized mouse strains. However, myeloid enhanced

humanized mice presented a decrease in effector memory (EM) cells in favor of an increase

of central memory (CM) cells. Additionally, BM of huNSG-SGM3 contained significantly

more Tregs compared to huNOG, a finding already described (Fig. 13C) (68).

Although transgenic mice were expected to share patterns of myeloid differentiation,

significant differences were detected. While in huNOG and huNOG-EXL mice macrophages

Results

53

(CD68+) are most abundant, huNSG-SGM3 show significantly higher levels for

CD66b+granulocytes (Fig. 13D+E). These findings are consist with previous publications,

reporting, that huNSG-SGM3 also reconstitute higher frequencies of human granulocytes

(87). Elevated levels for G-CSF in huNSG-SGM3 in combination with high levels of GM-

CSF might be a possible explanation (Fig. 11). Surprisingly, a high number of double posi-

tive (CD68+/CD66b+) cells was detected in bone marrow of both transgenic mice. It was

described that CD68+ is not an exclusive marker for monocytes and macrophages, as pre-

viously assumed, but also a marker for neutrophil granulocytes (88). Another explanation

might be that CMPs in bone marrow possible express both markers.

Remarkably, the three dendritic cell (DC) subpopulations were detectable in all hu-

manized mouse strains albeit at different levels. HuNSG-SGM3 mice present less DCs com-

pared to huNOG and in particular huNOG-EXL (Fig.13F). Representative flow cytometry

plots are showing differences in the size of pDC population between the strains (Fig. 13G).

FIGURE 13: IMMUNE CELL ANALYSIS IN BONE MARROW OF HUMANIZED MICE A) Human CD45+ comparison in bone marrow; ordinary one-way ANOVA, Tukey’s multiple compar-ison test, *, p < 0.05, n= 14 to 23. B) Composition of immune cells in CD45+. C) T cell characterization in BM at week 16, Ratio of CD4 and CD8 T cells in CD3+, Ratio of EM (Effector memory) and CM (Central memory) cells in CD3+, Frequency of Tregs in CD4+ T cells. D) Composition of myeloid cells in CD33+ in bone marrow of humanized mice. E) Representative flow cytometry plots for human

Results

54

macrophages (CD68) and neutrophils (CD66b), gated on single, viable, lineage negative, CD33+. F, Composition of DC subpopulations in HLA-DR+ in bone marrow of humanized mice. G, Representa-tive flow cytometry plots for human pDCs (CD303+/CD123+) gated on single, viable, lineage negative, HLA-DR+ in C, bone marrow and F, in spleens of huNOG, huNOG-EXL or huNSG-SGM3 mice.

For detection of these rare cell populations, specific cell surface markers, such as

CLEC9A, CD1c, and CD303 (BDCA-2), which are exclusively expressed on cDC1, cDC2,

pDCs were used (Fig. 14). Other markers such as CD123, CD141 and CD11c are not as

specific (89). CD123 is also found on CD11b positive cells and CD141 on CD303 CD123

double positive cells. CD11c is a common DC marker and consequently found on many

populations but also on CD11b+ myeloid cells (Fig. 14).

FIGURE 14: SPECIFIC FLOW CYTOMETRY DC MARKERS Representative histograms from huNOG-EXL and huNOG mice of DCs in spleens, CD303 is ex-pressed on pDCs (CD123+/ CD303+), CLEC9A on cDC1s (CD141+/CLEC9A+) and CD1c on cDC2s (CD11c+/ CD1c+).

Thymus

The existence of differentiated human T cell subsets in peripheral blood of humanized mice

indicates that human T cell development is functional in these mice. Therefore, thymus

samples of myeloid enhanced humanized mice as well as control huNOG mice were ana-

lyzed. Thymus size was comparable between all strains and no further macroscopic differ-

ences were detectable (Fig. 15A). Flow cytometry analysis revealed similar CD45+ levels in

thymus of all strains (Fig. 15B). Yet again, transgenic humanized mice presented a different

T / B cell ratio with a clear shift towards T cells. However, thymi of huNOG mice comprised

of mainly B cells (Fig. 15B). Characterization of T cell subpopulations showed comparable

levels of CD4+ cells, but a higher frequency for double positive cells in huNOG and more

CD8+ T cells in the transgenic strains (Fig. 15B). Immunohistochemistry (IHC) of thymi from

huNOG-EXL confirmed cell contact between human CD4+ T cells and human MHC-II (yel-

low areas), indicating direct interaction between T helper cells and APCs in thymi occurring

(Fig.16).

Results

55

FIGURE 15: ANALYSIS OF THYMI FROM HUMANIZED MICE A) Representative images of thymi from humanized mice, B) Human leukocyte composition in thy-mus as percentage, n= 6 to 10, Ordinary one-way ANOVA, Tukey’s multiple comparison test, ****, p < 0.0001, Immune cell composition in CD45+ and T cell subsets.

FIGURE 16: REPRESENTATIVE IMAGE OF IHC STAINING IN HUNOG-EXL THYMUS Blue= DAPI, Red= human CD4 (Alexa647), green = human MHC-II (Alexa488)

Results

56

4.2.2.3 Secondary lymphoid organs

Spleen

Both myeloid enhanced mouse strains showed significantly increased spleen size com-

pared to huNOG (Fig. 17A). Spleen weights of huNOG were comparable to the spleen

weights of immunocompetent mice such as BALB/cJ and C57BL/6J as indicated by the

breeder (Fig. 17B). These results confirm published data, which already described spleno-

megaly in huNSG-SGM3 as well as huNOG-EXL mice (69,85).

In general, spleens of huNOG-EXL mice have a higher human immune cell fre-

quency and humanized transgenic mice demonstrate higher levels of myeloid cells and less

B cells as compared to huNOG mice (Fig. 17C+D). In huNOG-EXL mice T cell levels were

elevated as compared to huNOG. Further T cell characterization showed a slight increase

in CD4+ cells in myeloid enhanced mice compared to huNOG (Fig. 17E). Further, T cell

subset analysis of spleens from huNOG mice was not possible because the total cell

amount was too low. In myeloid enhanced mice no significant difference was detected in

the ratio of EM to CM cells. However, huNSG-SGM3 mice presented higher Tregs, as pre-

viously described (61,67).

In line with the bone marrow data, huNOG and huNOG-EXL splenocytes also have

higher frequencies of macrophages (CD68+) and huNSG-SGM3 show significantly higher

levels for granulocytes (CD66b+), which are barely present in splenocytes of huNOG (Fig.

17F + G). Moreover, pDCs were also the largest group of DCs in huNOG and huNOG-EXL

spleens (Fig. 17H), indicating that these cells are able to exit the bone marrow and migrate

to lymphoid organs. In contrast, human DCs were barely detectable in spleen samples of

huNSG-SGM3 mice as also indicated in representative flow cytometry (Fig 17I).

Results

57

FIGURE 17: MACROSCOPIC AND IN-DEPTH CHARACTERIZATION OF SPLEENS A) Representative images of spleen from humanized mice. B) Spleen weight in mg (week 16), Ordi-nary one-way ANOVA, Tukey’s multiple comparison test, n= 17 to 23. *, p < 0.05; ***, p < 0.001; ****, p < 0.0001. C) Human CD45+ comparison in spleens, Ordinary one-way ANOVA, Tukey’s multiple comparison test, *, p < 0.05, n= 20 to 32. D) Immune cell composition in CD45+ from splenocytes. E) T cell characterization in spleens at week 16. EM (Effector memory), CM (Central memory). F, Com-position of myeloid cells in CD33+ from splenocytes. G, Representative dot plots of human macro-phages (CD68) and neutrophils (CD66b) gated on single, viable, lineage negative, CD33+ in huNOG, huNOG-EXL and huNSG-SGM3. H, DC subpopulations in HLA-DR+ from splenocytes. I, Repre-sentative dot plots of human pDCs (CD303+/CD123+) gated on single, viable, lineage negative, HLA-DR+ in huNOG, huNOG-EXL and huNSG-SGM3.

Lymph nodes

Immunodeficient mice commonly lack lymph node structures as a consequence of the func-

tional knockouts that were introduced to achieve the immunodefiency. Even in conventional

humanized mice with lymphoid reconstitution, lymph node development is impaired. Hence,

it was remarkable to see clearly detectable lymph nodes in myeloid enhanced humanized

mice. While mesenteric lymph nodes could be found in all three humanized mouse strains,

Results

58

although at much different sizes, only myeloid enhanced mice exhibited axial lymph node

development. As compared to huNSG-SGM3 mice, lymph nodes found in huNOG-EXL gen-

erally appeared to be larger (Fig. 18A). Flow cytometry analysis revealed only a small pop-

ulation of human CD45+ cells in LNs of huNOG compared to huNOG-EXL and huNSG-

SGM3, consistent with the number and size of LNs (Fig. 18B + C). Analysis of the immune

cell composition in LNs revealed a large population of T cells in the transgenic mice whereas

in huNOG B cells were the most abundant lymphocyte population with more than 25% (Fig.

18C). In all mouse models evaluated low numbers of human CD33+ myeloid cells were

detected in lymph nodes, indicating that human myeloid cells might not efficiently differen-

tiate in myeloid enhanced humanized mice. Deeper characterization of the T cell subsets

revealed more than 50% being CD4 positive T cells. A minor double positive subset was

detected in the transgenic strains as well as an increased population of CD8+ T cells. Lymph

node development and class switching of antibodies provide evidence for improved func-

tionality of adaptive immunity huNOG-EXL mice.

FIGURE 18: ANALYSIS OF AXIAL LYMPH NODES A) Representative images of axial LN, B) Axial LN amount from all humanized mouse strains ana-lyzed, n=10/strain. Ordinary one-way ANOVA, Tukey’s multiple comparison test. *, p < 0.05. C) Hu-man leukocyte composition (CD45+) in lymph nodes (LNs) as percentage, n= 6 to 10, Ordinary one-way ANOVA, Tukey’s multiple comparison test, ****, p < 0.0001; Immune cell composition in CD45+ and T cells in CD3+.

Results

59

4.2.2.3 Other tissue

Liver

Although the liver is neither a primary nor a secondary lymphoid organ it is a relevant organ

for inflammation and other pathologies. Therefore, livers from humanized mice were ana-

lyzed with focus on macroscopic anomalies and liver weight. Interestingly, livers from

huNSG-SGM3 mice were bigger (Fig. 19A+B) and showed an increase in weight compared

to huNOG and huNOG-EXL mice (Fig. 19B). These findings are in accordance with Yoshi-

ara et al. (86), who also showed hepatomegaly in this humanized mouse strain as a poten-

tial consequence of ongoing inflammation. No further flow cytometry analysis was per-

formed on livers.

FIGURE 19: MACROSCOPIC CHARACTERIZATION OF LIVERS A, Representative images of livers. B) Liver weight at termination (week 16), black = huNOG, light grey= huNOG-EXL, darker grey= huNSG-SGM3. Ordinary one-way ANOVA, Tukey’s multiple com-parison test, each dot represents one individual mouse, error bars indicate mean ± SEM, n= 17 to 23. *, p < 0.05; ***, p < 0.001; ****, p < 0.0001.

4.2.2.4 Summary

Taken together, this comprehensive characterization of the immune cell composition high-

lights the superiority in terms of myeloid cell differentiation of the myeloid enhanced trans-

genic mice compared to non-transgenic NOG mice. Moreover, the direct head-to-head com-

parison helps to understand essential differences between the models and serves as prac-

tical assistance for the choice of a suitable mouse model. Higher overall engraftment and

an increase in myeloid cells in both transgenic mouse strains, as described before (68,69),

were confirmed. Different immune subsets in humanized mice follow different engraftment

kinetics, in-depth characterization in peripheral blood provides clear guidance on when to

start specific investigation. Although huNOG-EXL and huNSG-SGM3 mice share two of the

transgenic cytokines with huNSG-SGM3, they exhibit quite different phenotypes. The com-

bination of high levels of human G-CSF and supraphysiological levels of human GM-CSF

very likely explains the shift towards granulocyte development in huNSG-SGM3 mice. The

additional expression of human SCF might account for several other differences observed

in huNSG-SGM3 mice such as increased frequency of regulatory T cells.

Results

60

HuNSG-SGM3 have been reported to have a reduced life span and to show signs

of anemia (85,86). The presented results from this study emphasize these findings. In par-

ticular, hepatosplenomegaly, elevated levels of inflammatory cytokines in serum and in-

creased mortality are indicators for a chronic inflammation. Due to the described limitations

of huNSG-SGM3, and in particular the reduced life span which impairs long term

investigations such as tumor studies, only huNOG-EXL mice were subjected to further

analysis and compared to huNOG as non-transgenic control.

HuNOG-EXL mice offer a promising and robust mouse model to study phenotype

and functionality of human myeloid cells in vivo. HuNOG-EXL mice show improved myeloid

cell reconstitution with an appropriate proportion of macrophages and DCs as well as in-

creased IgG levels and thus, are a more translatable in vivo model. Although the high pro-

portion of human pDCs in this mouse model is not physiological when compared to human

BM, it finally enables to study this rare cell population in more detail (90).

4.3 Depletion of murine macrophages and granulocytes leads to higher overall hu-

manization but decrease in survival

4.3.1 Background

As described in the introduction, cytokines are the key factor to improved differentiation of

various immune cell populations. For human myeloid cell differentiation there is another

important cytokine besides IL-3 and GM-CSF. It is well known that human CSF-1 is crucial

for the development and differentiation of human myeloid cells in humanized mice (67).

Moreover, it was reported that the knock-in of human CSF-1 led to improved differentiation

and functions of human myeloid cells in humanized BRG mice (65,66), as the example of

the MISTRG mouse demonstrates. These findings are supported by data of a previous PhD

Thesis by Dr. J. Eckmann (former Engelhardt) (91). He demonstrated that elevated murine

CSF-1 levels induced by anti-CSF-1R blocking and depletion of murine macrophages leads

to an increase of human macrophages in BRG mice. In order to reduce the existing chimer-

ism and to support the ratio in favor of human macrophages over mouse macrophages the

question was raised, if it is possible to use increased murine CSF-1 levels in huNOG-EXL

mice as well. Besides, the expectation that this would lead to enhanced overall humaniza-

tion and higher myeloid cell frequencies, blocking murine CSF-1R and deleting murine mac-

rophages would also help to understand possible effects in tumor microenvironment in sub-

sequent studies. However, safety of this antibody treatment had to be evaluated, because

mice with markedly enhanced myeloid reconstitution have a reduced life span such as MIS-

TRG or NSG-SGM3 (66,85).

Results

61

4.3.2 Depletion studies with huNOG-EXL mice

4.3.2.1 Peripheral blood analysis

Compared to conventional humanized mice, such as BRG mice, the expression of human

GM-CSF and human IL-3 in huNOG-EXL mice has already improved myeloid differentiation

and migration significantly. Thus it was hypothesized that lower doses of the mouse mac-

rophage depleting antibody 2G2 (anti-mouse CSF-1R, here αCSF-1R) would trigger human

macrophage development and terminal differentiation. Therefore, huNOG-EXL mice were

treated with 2.5, 5 or 10mg/kg of αCSF-1R or the corresponding mouse IgG1 isotype control

antibody (MOPC-21) respectively. Humanization levels were analyzed in peripheral blood.

Remarkably, already after the first treatment with different doses of αCSF-1R ratio of human

to mouse immune cells was slightly enhanced in dose-dependent manners. This effect was

further enhanced by a second treatment with αCSF-1R antibody (Fig. 20A).

Results

62

FIGURE 20: PERIPHERAL BLOOD ANALYSIS IN αCSF-1R TREATED huNOG-EXL MICE A) Comparison of the effect of different treatment doses at baseline (=before 2G2 treatment) or after first (1st) or second (2nd) treatment with the blocking αCSF-1R antibody on immune cells (CD45+), T cells (CD3+) and myeloid cells (CD33+), Vehicle = isotype antibody control, Ordinary two-way ANOVA, *, p < 0.05; ***, p < 0.001; ****, p < 0.0001, n = 5 to 9. B) Representative dot plots of human monocytes in healthy humans and huNOG-EXL mice without αCSF-1R treatment or with different doses respectively. C) Murine CSF-1 levels measured in serum of huNOG-EXL mice with αCSF-1R treatment. D) Survival curve comparison between mice treated with different αCSF-1R doses, log-rank (Mantel-Cox) or Gehan-Breslow-Wilcoxon test yielded in the same statistical results, n= 5 to 9, *, p < 0.05. E) Human cytokine levels measured in serum of huNOG-EXL mice treated with different αCSF-1R doses or isotype control antibody respectively. No αCSF-1R = no treatment at all, Vehicle = isotype.

Results

63

The increased humanization over time was accompanied by increased frequencies of

human CD33+ myeloid as well as CD3+ T cells. Further characterization of the myeloid sub-

sets in peripheral blood revealed a dose-dependent increase in intermediate

CD14dim/CD16+ and nonclassical CD14+/CD16+ monocytes in αCSF-1R treated mice versus

control mice (Fig. 20B). In addition, serum samples were analyzed for mouse CSF-1 levels.

Whereas control mice did not show significant changes over time, CSF-1 was elevated in

mice treated with a dose of 5 mg/kg αCSF-1R or higher. Additionally it was found, that 2.5

mg/kg of αCSF-1R had a clear correlation between increase in mouse CSF-1 levels and

human myeloid cell frequencies (Fig. 20C). Furthermore, serum was analyzed to check for

human cytokines levels. In αCSF-1R treated mice cytokine concentrations were increased,

especially in the higher dosed group with 10mg/kg. Of note, not only myeloid-associated

cytokines such as CCL-2 were increased, but also inflammatory cytokines for instance IL-6

and IL-8 (Fig. 20E). The downside of increased humanization, enhanced myeloid and T cell

reconstitution was a lower survival rate of αCSF-1R treated huNOG-EXL mice compared to

control mice (Fig. 20D). Again, a dose-dependent effect was detectable.

4.3.2.2 Other Organs

To test whether improved reconstitution of monocytes was accompanied by terminal differ-

entiation of human tissue human macrophages, lymphoid as well as non-lymphoid organs

were analyzed by flow cytometry. Bone marrow was extracted from femurs of αCSF-1R

treated huNOG-EXL or control mice, respectively. It was detected that higher doses of the

anti-mouse CSF-1R blocking antibody led to higher overall engraftment in mice. In the high-

est concentration group more than 80% of the cells were human origin, consisting with find-

ings in the necropsy where these mice often displayed pale bones. However, the higher

levels of immune cells did not result in higher frequencies of CD33+ myeloid cells or CD68+

macrophages in bone marrow. On the contrary, a trend in the other direction could be iden-

tified. In the 10mg/kg group myeloid cells and macrophages are slightly reduced in bone

marrow, indicating that these cells migrated to the periphery. Remaining macrophages how-

ever, were prone to differentiate into CD163+ macrophages, often classified as M2 macro-

phages (Fig. 21A).

In spleens it was shown, that αCSF-1R treatment led to increased levels of myeloid

cells and in particular CD163+ macrophages in a dose-dependent manner (Fig. 21B). IHC

analysis of the same organs confirmed these findings (data not shown).

Livers of αCSF-1R treated mice often looked conspicuous during necropsy, IHC

analysis of this organ was additionally performed. Livers appeared to have issues with tis-

sue integrity, as they were soft and instable. Therefore, livers were stained for human mac-

rophages (CD68), T cells (CD3), murine macrophages (F4/80) and connective tissue (Sirius

Results

64

red). It was found that human macrophages were localized in clusters often around vessels.

Moreover, more human macrophages were detected in higher dosed groups, whereas al-

most no murine F4/80+ macrophages were found in the 5mg/kg group, indicating that this

might be the cut off to achieve depletion of murine macrophages accompanied by an in-

creased murine CSF-1 level. Remarkably, collagen was also increased with elevated αCSF-

1R levels as confirmed by Sirius red staining, thus indicating starting or ongoing liver fibrosis

(Fig. 21C).

Results

65

FIGURE 21: ANALYSIS OF PERIPHERAL ORGANS AFTER αCSF-1R TREAMENT A) Flow cytometry analysis of bone marrow for immune cells (CD45+), myeloid cells (CD33+), mac-rophages in general (CD68+) and M2-like macrophages (CD163+). B) Flow cytometry analysis of spleens from the same mice as in A. C) IHC staining of liver with or without αCSF-1R treatment for human macrophages & T cells, murine macrophages and collagen.

Results

66

4.3.3 Summary

In general, frequencies of total human immune cells as well as human monocytes in human

CD45+ cells increased in peripheral blood and lymphoid organs upon treatment with αCSF-

1R. However, a clear dose threshold was detectable, as 2.5mg/kg of the αCSF-1R antibody

was not sufficient to enhance mouse CSF-1 levels and consequently increase human im-

mune cells and in particular human macrophages. In summary, it was demonstrated that

repetitive injection with anti-mouse CSF-1R significantly improves reconstitution of human

peripheral blood monocytes as well as human tissue infiltrating macrophages. Unfortunately

it was not possible to establish a dose that facilitates to increase murine CSF-1 levels and

subsequently human macrophages without affecting the overall survival of the mice. On the

contrary, these results are indicating that once the system is activated by exceeding the

critical value of the murine CSF-1 levels, the side effects cannot be stopped. Consequently,

no further experiments using huNOG-EXL mice in combination with αCSF-1R treatment

were performed.

4.4 Functionality of human macrophages from huNOG-EXL mice

4.4.1 Stimulation with Toll-like receptor agonist 8 in vivo

Human macrophages were found in huNOG-EXL mice at higher frequencies compared to

huNOG mice. However, their functionality in vivo still remains to be confirmed. Hence, the

cytokine response of human macrophages in huNOG-EXL mice was tested in vivo by i.p.

injection of the TLR8-agonist TL8-506. Four hours after injection serum of mice was taken

and analyzed by multiplex. Pro-inflammatory cytokines associated with monocytes/macro-

phages, such as IL-6, IL-8, MCP-1, MIP-1β and TNF-α were elevated in mice treated with

TLR8-agonist compared to vehicle treated mice (Fig. 22). Serum concentrations of human

TNF- α and IL-6 were elevated as well, though the difference was not significant. In general,

cytokine response was heterogeneous among individual mice most likely due to different

frequencies of human macrophages in humanized mice. In summary, release of human

pro-inflammatory cytokines in response to TLR8 stimulation provided first evidence for func-

tionality of human macrophages in huNOG-EXL mice in vivo.

Results

67

FIGURE 22: CYTOKINE RESPONSE OF HUMAN MACROPHAGES IN huNOG-EXL MICE Cytokine analysis measured by multiplex in serum of huNOG-EXL mice treated with 35 µg TLR-8/mouse or serum-free water as Vehicle control. Two-tailed unpaired t-test, *, p < 0.05; **, p < 0.005; ***, p < 0.0005; ****, p< 0.0001; n= 8 – 10.

4.4.2 Analysis of human macrophages ex vivo

As part of the innate immune response, one key function of macrophages is their ability to

antibody-dependent cellular phagocytosis (ADCP) for example of pathogens but also of hu-

man tumor cells. To check if macrophages from huNOG-EXL mice are functional in terms

of ADCP, bone marrow single cell suspensions were depleted of mouse immune cells and

differentiated into macrophages. After five days, CMFDA labeled, HER-2 positive SkBr3

breast cancer cells were added to the cell culture (92). To induce ADCP, Herceptin, a mon-

oclonal antibody targeting HER-2, was added to the macrophage-tumor cell co-culture. As

isotype control a IgG1 antibody (Privigen®) was used. To confirm the presence of human,

but not mouse macrophages, co-cultures were stained with fluorescently labeled antibodies

against F4/80 and CD163. Hoechst dye 33343 was used as counterstain (Fig. 23).

Results

68

FIGURE 23: IMMUNOFLUORESCENCE STAINING OF ENRICHED BM CELLS FROM huNOG-EXL First column shows staining with mouse macrophage marker F4/80 (red). Second column displays staining of human macrophages with the marker CD163 (green). Nuclei were stained with Hoechst dye 33342 (blue). m1= mouse 1, m2= mouse 2, m3=mouse 3

For phagocytosis assay wells were prepared with an equal amount of BM cells (~20000)

and untreated tumor cells (~3000) (Fig. 24A). In wells were Herceptin was added phagocy-

tosis was significantly increased compared to IgG control. In turn, number of SkBr3 tumor

cell count was significantly reduced (Fig. 24B - D). These results present that human mac-

rophages differentiated from huNOG-EXL bone marrow are capable of FcγR-mediated

phagocytosis and are functional.

Results

69

FIGURE 24: PHAGOCYTOSIS ASSAY WITH MACROPHAGES DIFFERENTIATED FROM huNOG-EXL BM A) Total cell count of untreated BM cells with equal amounts as untreated control. B) Comparison of phagocytosis level between wells cultured with the IgG antibody Privigen (control) and wells treated with Herceptin as percentage. C) Comparison of total cell count of SkBr3 tumor cells between un-treated cells, cells cultured with Privigen and cells cultured with Herceptin. D) Images of fluorescent SkBr3 tumor cells (green) in wells cultured with Herceptin 5µg/ml or IgG1 Isotype 5g/ml (Privigen®) respectively. n=3/mouse/condition. Loss of CMFDA florescence indicates phagocytosis of tumor cells after 24h. m1= mouse 1, m2= mouse 2, m3=mouse 3.

4.5 Functionality of human pDCs from huNOG-EXL mice

Beside human macrophages, the next aim was to check functionality of another important

but rare myeloid subset, the pDCs. In BM of huNOG-EXL mice a robust amount of human

pDCs in huNOG-EXL bone marrow was identified (Fig. 13F). Therefore, huNOG-EXL bone

marrow was pooled and pDCs were enriched using a customized pDC isolation kit. To com-

pare functionality between myeloid enhanced and conventional humanized mice, bone mar-

row from huNOG mice was subjected to the same procedure, however cell counts were too

low to perform a thourough comparison. Since TLR activation-induced cytokine release was

validated as reliable readout for pDCs function, enriched pDCs from huNOG-EXL mice were

cultivated for 3h or overnight (ON) either with or without the addition of TLR agonists. Next,

supernatant was taken for cytokine analysis, while cells were stained for flow cytometry.

Upon ex vivo stimulation with TLR agonists it was expected that pDCs upregulate co-stim-

ulatory molecules. Indeed, flow cytometry showed that CD83, CD69 and CD86 mean fluo-

rescence intensity (MFI) were elevated on pDCs that were stimulated for 3h or overnight

Results

70

(ON) compared to non-stimulated control (Fig. 25A). pDCs are known to be the major pro-

ducers of type I interferon, therefore IFN-α2 production was analyzed. Treatment with TLR

agonists resulted in a significantly increased IFN-α2 production already after 3h and even

higher concentration after stimulation ON (Fig. 25B). The early activation marker CD69

shifted in expression already after 3h while CD83 and CD86 expression levels were signif-

icantly changed only after ON stimulation (Fig. 25C).

FIGURE 25: HuNOG-EXL DEVELOP FUNCTIONAL HUMAN PDCs A) Activation marker MFIs (CD83, CD69, CD86), gated on pDCs (CD123+/CD303+) enriched from huNOG-EXL BM with TLR7/8 or TLR9, measured after 3h or overnight (ON) stimulation, Ordinary one-way ANOVA, Tukey’s multiple comparison test. *, p < 0.05; **, p < 0.001; ***, p < 0.0001, n= 3-5/condition. B) IFN-α2 levels in supernatant of TLR agonist treated or non-treated pDCs enriched from huNOG-EXL BM. C) Representative activation marker histograms with (TLR7/8 or TLR9) or without stimulation (media control = med. ctl.) ON, gated on lin-, HLA-DR+, pDC+

4.6 Growth of human xenograft tumors in humanized mice

Emerging evidence confirms that myeloid tumor infiltration plays a critical role in tumor de-

velopment and progression. Tumor infiltrating myeloid cells can efficiently suppress anti-

tumor immune responses and thus are attractive therapeutic targets in order to prevent

tumor progression and relapse (93,94). Well-known examples are macrophages, which

contribute depending on their various phenotypes and surface markers to survival of pa-

tients (25). Another cell type that was found to be correlated with the survival of patients in

clinical trials are pDCs (10,33,34,95). Especially in ovarian and breast cancer as well as

melanoma, pDC infiltration has been reported to correlate with bad prognosis for patients

(10,41,42). However, the complex role of innate tumor infiltrating immune cells is still not

fully understood also because predictive preclinical models to study human tumor and im-

mune cell interaction have not yet established. In general, humanized mice provide a unique

possibility to investigate the growth of human tumor cells in the presence of an existing

Results

71

human immune system and thus offer an advanced preclinical model to study mechanisms

of cancer immunotherapy in vivo.

4.6.1 Tumor growth kinetics

In a first step tumor growth of different cell lines and one patient derived breast cancer

xenograft were compared in myeloid enhanced versus classical humanized NOG mice. All

three tumor models were able to engraft in huNOG as well as in huNOG-EXL as confirmed

by increased tumor volume over time. For the ovarian cancer cell line SK-OV-3 no differ-

ences in tumor growth was detectable between mouse strains. Conversely, OVCAR-5,

which is also an ovarian cancer cell line showed a considerably reduced tumor growth in

huNOG-EXL mice compared to huNOG mice (Fig. 26). The breast cancer PDX BC_038

had a much slower tumor growth onset compared to the two cell line based xenografts and

growth kinetics were slightly but statistically significant different between the strains.

FIGURE 26: TUMOR GROWTH KINETICS OF DIFFERENT TUMORS IN HUMANIZED MICE Tumor volume [mm3±SEM] in humanized mice injected with different tumor lines or PDX, respec-tively. 2-way ANOVA, Tukey’s multiple comparison test. ****, p < 0.0001, error bars indicate median ± 95% CI, n=27 to 67.

4.6.2 Characterization of different tumor lines or PDX in different mouse strains

To clarify whether xenografted tumors are able to influence the engrafted human immune

system, immune cells in blood were analyzed during tumor growth. Human immune cell

(CD45+) and myeloid cell (CD33+) frequencies were preserved in peripheral blood inde-

pendent of tumor growth (Fig. 27). Other cell populations such as T, B and NK cells were

also analyzed and did not differ between mice inoculated with different tumors (data not

shown).

17 24 31 37 42 490

100

200

300

400

SK-OV-3

17 22 28 36 42 480

250

500

750

1000

OVCAR-5

****

33 42 46 50 54 610

200

400

600

800

1000

BC_038

*

NOG

days after tumor cell inoculation

NOG-EXL

Tu

mo

r vo

lum

e [

mm

3]

Results

72

FIGURE 27: COMPARISON OF OTAL LEUKOCYTES AND MYELOID CELLS IN PE-RIPHERAL BLOOD Comparison of human CD45+ in peripheral blood of mice inoculated with either SK-OV-3, OVCAR-5 or patient-derived xenograft BC_038, week 20. 2-way ANOVA, Tukey’s multiple comparison test, error bars indicate median ± IQR. n= 22 to 26, ns= not significant.

Although there were no obvious differences detected in the periphery, the local tu-

mor cytokine milieu can be expected to influence immune contexture. Therefore, character-

ization of the tumor immune cell infiltrate was performed by flow cytometry. Low numbers

of immune cells were found in OVCAR-5 tumors regardless of the mouse strain. In compar-

ison, higher frequencies of human CD45 positive cells were detectable in SK-OV-3 and

BC_038 tumors (Fig. 28A). In general there was a clear trend for increased baseline infil-

tration of immune cells in tumors engrafted into huNOG-EXL as compared to huNOG mice

(Fig.28A). Strikingly, an in-depth characterization of immune subsets in tumor samples re-

vealed a dominant effect of the tumor model over the humanized mouse strain. SK-OV-3

had a different immune cell arrangement compared to OVCAR-5 and BC_038, which were

quite comparable. In detail, SK-OV-3 had lower T cell levels but instead remarkably high

frequencies of NK cells independent of the mouse strain (Fig. 28B + C). Interestingly, hu-

NOG-EXL mice did not show more myeloid cells in any tumor model tested (Fig. 28C).

Results

73

In a next step further characerization of tumor associated myeloid cells in the different tu-

mors models was performed. Across all samples analyzed the majority of myeloid cells were

identified as CD68+CD66b- tumor associated macrophages (TAMs). TAMs were highly

abundant in SK-OV-3 tumors whereas reasonable numbers of CD68-CD66b+ granulocytes

were only detected in BC_038 and OVCAR-5 tumors. Interestingly, in all xenografts a

CD33+ but CD66b- CD68- double negative population was detected, which might indicate

the presence of myeloid derived suppressor cells (Fig. 28D). Last but not least, human DCs

were detected in all tumors albeit at different frequencies (Fig. 28D). In two out of three

models, BC_038 and OVCAR-5, all three major subsets of DCs were identified.

FIGURE 28: COMPARISON OF TUMOR IMMUNE CELL INFILTRATE A) Total immune cell infiltrate in different tumor models from huNOG-EXL and huNOG, 2-way ANOVA, Tukey’s multiple comparison test, n= 3 to 6. B) Human leukocytes composition (CD45+) in different tumor models in huNOG and huNOG-EXL mice, n=4 to 12. C) Pie charts of immune subsets in SK-OV-3, OVCAR-5 or BC_038 tumors. Myeloid cells (CD68/CD66b) in CD33+ and DCs (CD303+/CD123+, CD1c+/CD11c+, CD141+/CLEC9A+) in HLA-DR+.

4.6.3 pDC occurrence in tumors is dependent on several human cytokines and

chemokines

Since pDCs were detected in OVCAR-5 and BC_038 but not in SK-OV-3 tumors, further

investigation of factors that are associated with pDCs were performed. hFlt3-L, hGM-CSF

and hIL-3 are long known to be correlated with pDC survival and differentiation (96-98).

However, lately various new migratory cues have been identified to direct pDCs from BM to

blood or into diseased tissues including tumor-derived CXCL12 (SDF-1), CXCR3 ligand

CXCL9/10 and CCR7 ligands CCL19 and CCL21 (12,99-105). Hence, expression levels of

Results

74

these cytokines and chemokines in tumor lysates of humanized mice inoculated with the

characterized tumors SK-OV3, OVCAR-5 and BC_038 were investigated (80). The strong-

est pDC recruiting factor human Flt3-L showed significantly higher expression levels in the

breast cancer PDX BC_038 compared to the cell lines SK-OV-3 and OVCAR-5 (Fig. 29).

The effect of the two transgenes hGM-CSF and hIL-3 in NOG-EXL mice showed no effect

on the inoculated tumors, as levels are comparable between the two mouse strains. How-

ever, a difference is detectable between the tumors. SK-OV-3 tumors express significantly

higher levels of hGM-CSF and hIL-3. Yet, IL-3 levels are generally very low in all tumor

models. CXCR3 ligand CXCL9 and CXCL10 are increased in OVCAR-5 and BC_038 as is

CCL19 and CXCL12. CCL21 showed higher levels in BC_038 compared to the two ovarian

tumors (Fig. 29). Other chemokines showed only minor or no differences between the tumor

models. These results indicate that not one single cytokine/chemokine is responsible for

pDC occurrence and migration in tumors but the combination of several (80).

FIGURE 29: ANALYSIS OF PDC-ASSOCIATED CHEMOKINES AND CYTOKINES OF DIFFERENT TUMORS Cytokines hFLT-3L, hGM-CSF, hIL-3 and chemokines associated with pDC migration hCXCL9,

hCXCL10. hCXCL12, hCCL19, and hCCL21analyzed by Bio-Plex from tumor lysates of humanized

mice inoculated with SK-OV-3, OVCAR-5 or BC_038, Kruskal-Wallis ANOVA, error bars indicate

median ± IQR, *p < 0.05; **p < 0.01, ***p < 0.0005, ****p < 0.0001, n = 3–19.

Results

75

4.6.4 Functional analysis of tumor-resident pDCs

The finding of this rather huge and functional population of pDCs in huNOG-EXL periphery

as well as in the two tumor models OVCAR-5 and BC_038 led to the question if these cells

are also able to react to TLR agonists. The presence of pDCs is known to be a prognostic

factor in ovarian and breast cancer (32-34), therefore further functional analysis of pDCs

was performed by intra-tumoral injection of the TLR7/8 agonist R848. Mice were sacrificed

4h later and pDC activation marker expression as well as intra-tumoral IFN-α2 levels were

measured. Activation markers on tumor-derived pDCs were found to be upregulated upon

TLR stimulation (Fig. 30A + B). Additionally, whole tumor lysates from TLR7/8 treated mice

confirmed an increase not only of IFN-α2 but also several other inflammatory cytokines such

as IL-6, IL-8, MIP-1β and TNF-α upon TLR engagement (Fig. 30C). pDCs are known to not

only produce IFN-α but also other cytokines such as the above mentioned (35,106,107).

However, it cannot completely be excluded that other cells also react to TLR agonists. To

investigate if tumor cells themselves are among the cytokine producers, OVCAR-5 was

cultivated in vitro, cells were treated with a wide dose range of TLR7/8 agonists and super-

natant of the tumor cells was analyzed. All IFN-α2 levels were below detection level (data

not shown). A possible reason can be found in the TLR signaling cascade. TLR cleavage

is essential for functioning TLR signaling and impaired cleavage capacity might conse-

quently lead to unavailable IFN-a2 production (108,109). As BC_038 is a primary xenograft

with no existing cell line, the experiment could not be performed with this tumor model.

Results

76

FIGURE 30: FUNCTIONAL ANALYSIS OF TUMOR-RESIDENT PDCs A) Activation markers (CD69, CD83, CD86) on pDCs in BC_038 tumor upon stimulation. B) Repre-sentative histograms of activation markers with stimulation (TLR7/8) or without stimulation (vehicle), gated on lin-, HLA-DR+, pDC+ in BC_038 tumors. C) Human cytokines analyzed by Multiplex from tumor lysates of humanized mice treated i.t. with TLR7/8 or Vehicle for 4h; 2-way ANOVA, uncor-rected Fisher's Least Significant Difference (LSD); *, p < 0.05; **, p < 0.001; ***, p < 0.0005; ****, p < 0.0001; n= 4 to 12

In theory, a major advantage of intra-tumoral administration of TLR agonists is that

local cytokine release prevents systemic toxicity caused by exuberant cytokine release in

consequence of immune stimulation. Consequently, serum of TLR stimulated tumor bearing

mice was analyzed to check for side effects in the periphery. Although a systemic increase

of INF-α2 as well as other cytokines was detected after 4h (Fig. 31A), serum levels were

significantly lower than in humanized, non-tumor bearing mice in which systemic cytokine

release was induced by i.p. injection of TLR7/8 agonist.

Results

77

These results clearly indicate that cytokine responses induced by intra-tumoral TLR

stimulation are mostly retained within the tumor tissue (Fig. 31B).

FIGURE 31: HUMAN CYTOKINES IN SERUM OF TLR TREATED HUMANIZED MICE A) Human IFN-a2 levels analyzed by Multiplex from serum of tumor- and non-tumor-bearing human-ized mice treated i.t or i.p respectively for 4h. B) Human cytokines analyzed by Multiplex from serum of humanized mice treated i.t. with TLR7/8 or Vehicle for 4h, 2-way ANOVA, uncorrected Fisher's Least Significant Difference (LSD); *, p < 0.05; **, p < 0.001; ***, p < 0.0005; ****, p < 0.0001; n= 4 to 12.

In summary, two ovarian cancer cell lines and one breast cancer PDX were analyzed for

deeper characterization of human immune cell composition in the myeloid enhanced hu-

NOG-EXL strain compared to classical huNOG. Comparison of tumor immune contexture

revealed that the tumor model itself had a more pronounced impact than the humanized

Results

78

mouse strain used. Striking differences in WBC composition were detected between con-

ventional and myeloid enhanced humanized mice, therefore increased overall infiltration

and myeloid reconstitution in tumors of huNOG-EXL mice could have been expected. In-

deed, a higher frequency of immune cells were present in the tumors of huNOG-EXL mice,

except for OVCAR-5. However, the effect of the mouse strain on the immune cell infiltrate

in the tumors was only minor compared to the effect of the tumor itself. This might be ex-

plained by the cytokine milieu of the tumors. Cytokines are not solitary produced by immune

cells but also by tumor cells. Hence, tumor cells are for example able to attract tumor-favor-

able promoting immune cells (110-112). BC_038 and OVCAR-5 share a similar cytokine

profile as well as a comparable immune cell composition even though one is a PDX and the

other an established cancer cell line from a different tumor entity (Fig. 28 + 32). SK-OV-3

on the other hand has a different tumor cytokine milieu. In conclusion, the choice of the right

tumor model is often underestimated when designing in vivo studies with myeloid targets.

Results

79

FIGURE 32: CYTOKINE PROFILE OF TUMORS IN HUMANIZED MICE AS HEAT MAP Heat Map showing cytokine levels at baseline of tumors from SK-OV3-3, OVCAR-5 and BC_038. Row dendrogram: Clustering method: complete linkage, distance measure: Correlation, ordering weight: Average value, Normalization: None, Scale: Logarithmic. Column dendrogram: Clustering method: Complete linkage, Distance measure: City Block, Ordering weight: Average value, Normal-ization: None

Discussion

80

5 Discussion

5.1 Generation of humanized mice

Humanized mice have been demonstrated to be a useful and important tool for cancer immu-

notherapy studies. Different mouse strains are generated in various ways with diverse immune

cell reconstitution levels (59). Classical humanized mice, such as BRG, NOG and NSG pri-

marily reconstitute T and B cells (59,67,87) and offer a strong model to study tumor interaction

with these lymphoid populations e.g. for checkpoint inhibitor studies (113,114). However, these

models lack proper antigen dependent immune responses (115,116) and show low myeloid

cell development due to impaired cross-reactivity between human and murine cytokines

(59,87). In consequence, several humanized mouse models expressing human cytokines have

been developed in the past applying different methods such as hydrodynamic plasmid injection

(67), random insertion of transgenes (68,69) or targeted genetic knock-ins (66). As a result,

myeloid enhanced humanized mouse models were generated and provided somewhat im-

proved reconstitution with human myeloid immune subsets (66-69). Humanized mice can be

generated in various ways and differ significantly between the individual research groups.

There are many experimental parameters which result in different engraftment levels: The use

of newborn or adult mice, the source of the human donor cells, different injections methods as

well as several ways of preconditioning (59,63). Different laboratories vary in their animal care-

taking conditions with housing as well as diet playing an essential role in the development of

the immune system. These are all factors that should be taken into account, when comparing

individual humanized mice data from different research groups. Yet, a phenotypic and func-

tional head to head comparison of novel myeloid enhanced mouse strains NSG-SGM3 and

NOG-EXL using identical humanization protocols was still missing. Therefore, a comprehen-

sive characterization was performed to validate the superiority of these models over classical

humanized NOG mice.

5.2 Characterization of human HSCs and donor variability

On one hand, HSCs are defined as multipotent, which means the ability to form all differenti-

ated blood cells, and on the other hand, they can renew themselves in the long term, which

indicates that cell division can produce descendants that are identical to the parents. In gen-

eral, HSCs are characterized as lineage negative, CD34+ and CD38-. The engraftment and

development of mature human T cells in immunodeficient mice could lead to xeno GvHD (117).

Therefore, it is important to ensure that cord blood HSCs are depleted of human T cells. This

was confirmed by flow cytometry, which showed that only a minor population was still CD3

positive (Fig. 6). Further, there are hematopoietic cells that are still multipotent, but no longer

capable of long-term self-renewal, as they have already further differentiated. These CD90-

Discussion

81

CD45RA− cells are termed multipotent progenitors (MPP) in contrast to the more immature

CD90+ CD45RA- HSCs (118,119). Both populations have been shown to successfully develop

into stable human hematopoiesis in NOG mice, albeit to different engraftment levels (118). In

this work, HSCs were not further sorted or specific subpopulation isolated. Instead the mixture

of HSCs and MPPs were injected and long-term as well as stable human cell engraftment was

achieved in all mouse strains. The purchased umbilical cord blood derived HSCs used in this

work for humanization showed a homogeneous distribution of these HSCs and MPP popula-

tions across all donors. However, in individual cases high variability in humanization levels

from different human hematopoietic stem cell donors could be observed. This resulted in dif-

ferences of overall engraftment and in particular in the development of T cells, which was

already described by others (73,120). It is most likely that these variations result from variable

frequencies of MPPs and HSCs in the purchased cord-blood derived HSCs. To overcome this

hurdle, mice were randomized using the donor origin as one additional criterion besides hu-

manization level and tumor size. Thus, donor heterogeneity was bypassed.

5.3 Comparison of immune cell compositions in different organs

All mice arrived with the same age and were held one week in acclimatization before humani-

zation. On arrival, NSG-SGM3 mice showed a higher body weight than the two NOG strains

and were able to maintain this throughout the entire study. Nevertheless, huNSG-SGM3

started to drop in survival around day 100 (85). Taken together, other findings such as anemia,

increased inflammatory cytokines and high T cell levels are more likely to lead to this decline

in survival, indicating that body weight alone is not sufficient to predict an overall poor health

condition.

In a first step, higher overall engraftment and an increase in myeloid cells in peripheral

blood of both transgenic mouse strains was confirmed. Further characterization over time in

peripheral blood additionally gave a clear overview, which cell populations develop at a specific

week. This might be a helpful tool to decide when to start an experiment. While myeloid en-

hanced humanized mice are better mimicking human immune cell composition, all mice still

show discrepancies when compared to human PBMCs. Especially huNOG mice mainly ex-

press B cells. Transgenic mice are better reflecting the composition, however huNSG-SGM3

mice are prone to increased and un-physiological T cell development. Although, myeloid en-

hanced mice displayed all monocyte subsets, they did not reflect human white blood compo-

sition, which showed in particular more classical CD14+/CD16- monocytes but also more non-

classical and intermediate subsets. Another missing cytokine, namely CSF-1 might be respon-

sible, as it was previously shown that CSF-1 is able to increase human monocyte reconstitution

in humanized mice (67).

Discussion

82

Baseline serum analysis of humanized mice also showed huge differences between

the strains. Serum cytokine characterization revealed increased levels of inflammatory cyto-

kines in transgenic mice, especially huNSG-SGM3 compared to the huNOG non-transgenic

strain, as already reported (85,86). Inflammatory cytokines such as IL-6, IL-8 and TNF-α, which

are most likely produced by activated macrophages, were significantly elevated compared to

the other strains. But also IFN-γ from activated T cells was increased. Higher levels of IL-10

likely represents a counteracting response to this ongoing inflammation (121). These findings

are in accordance with previously published data and might indicate that these mice experi-

ence a so called hemophagocytic lymphohistiocytosis (HLH), which is caused by activated T

cells and macrophages and their cytokines (86). The unexpectedly high T cell levels in huNSG-

SGM3 in blood and in the periphery might correlate and contribute to these increased cytokine

levels. Control sera from non-humanized mice served to control the specificity of the capture

antibodies for human cytokines in the multiplex assays used (data not shown).

Another often-criticized limitation of classical humanized mice is the impaired ability of

human B cells to perform class switching and consequently minimal antigen-specific IgG pro-

duction (115,116). This disadvantage of humanized mouse models is a major problem, partic-

ularly in vaccine and infectious disease research, where the humoral response is of crucial

importance. Although some models react to antigen-specific antibodies with generation of IgM,

IgG levels remain rather low (122,123). Many factors are considered as reasons for this defi-

ciency, such as compromised T and B cell development, deficiency in lymph node structure as

well as deprived differentiation of myeloid cells (122,124). Consequently many research

groups have focused on this topic and novel humanized mouse models developed, e.g.

mCD47/BALB-HIS mice, IL-6 knock-in mice or the BRGST mice, which showed that B cell

function can be improved (64,116,125). Macroscopic comparison of the herein analyzed mod-

els demonstrated more and bigger LNs especially in huNOG-EXL compared to huNOG. Unlike

the above mentioned mouse models that were specifically produced for LN development, the

LNs in this study presented mostly T cells instead of B cells. However, serum analysis pre-

sented physiological levels of human IgG antibodies, indicating that in myeloid enhanced but

especially in huNOG-EXL mice B cells undergo terminal maturation leading to improved func-

tionality. This is in accordance with former studies, demonstrating that expression of human

cytokines SCF, GM-CSF and IL-3 in NSG BLT mice not only led to improved human myeloid

cell reconstitution but also improved B cell development and antigen-specific antibodies

(122,126,127). Concluding, the herein presented results show that the transgenic expression

of human cytokines are an important factor for the development of the adaptive immunity in

humanized mice. This makes these mouse strains even more promising models for infection

and vaccine research. However, further experiments, e.g. immunization assays are needed for

deeper B cell characterization.

Discussion

83

Further characterization of the immune cell composition in bone marrow presented that

huNSG-SGM3 have higher levels of T cells and granulocytes, as was already seen in periph-

eral blood. Deeper investigation of the myeloid compartment in BM led to the finding of all three

dendritic cell subpopulations in bone marrow of humanized mice. Remarkably, is the high fre-

quency of pDCs, particularly in huNOG-EXL mice. Compared to previously published data in

humanized NSG mice even the herein analyzed huNOG mice show high pDC frequencies in

bone marrow. However, huNSG-SGM3 present a much lower frequency compared to huNOG

and huNOG-EXL. Outstandingly, pDC levels in peripheral blood and lymphoid organs of hu-

NOG-EXL mice are greatly increased compared to humans (90). This fact enables functional

studies of this rare pDC population without additional manipulation such as ex vivo expansion.

The existence of differentiated human T cell subsets in peripheral blood of humanized

mice indicates that human T cell development is functional in these mice. Flow cytometry dis-

played similar CD45+ levels in thymi of all strains with transgenic humanized mice showing a

clear shift towards T cells. Thymi of huNOG mice were composed mainly of B cells. In humans

thymic B cells are a unique population of B lymphocytes that reside mainly at the cortico-me-

dullary junction of the thymus. These B cells are distinct from peripheral B cells in terms of

their origin as well as their phenotype (88,128-130). For example in thymi, immature B cells

proliferated more than mature subsets (131). Furthermore, it is described for huNOG mice that

human B cells were only partially differentiated and that B cell precursors accumulated in the

spleens of these mice (123). It might be possible that B cells also accumulate in the thymus.

This also correlates with the poor production of IgG, which was confirmed in this study. B cells

are seldom investigated in thymi of humanized mice, so there is no available literature to com-

pare this population in these models. It is assumed, that huNOG mice present higher levels of

thymic B cells compared to the myeloid enhanced strains due to overall poor differentiation of

B cells in this strain and that these cells accumulate partly in the thymus. However, to charac-

terize thymic B cells in detail in humanized mice further studies are needed. Especially matu-

ration markers need to be evaluated as well as the localization by histology. Further charac-

terization of T cell subpopulations in all strains showed comparable levels of CD4+ cells, but a

higher frequency for CD4+CD8+ double positive cells in huNOG and more CD8+ T cells in the

transgenic strains. Immunohistochemistry (IHC) of thymi from huNOG-EXL confirmed cell con-

tact between human CD4+ T cells and human MHC-II+ areas (yellow areas). As outlined in

detail in the introduction the interaction between MHC molecules and T cells is an important

part of thymic T cell selection. T cells which are not able to bind, die during the selection pro-

cess. Thus, these findings prove that reconstituted human CD4+ T cells are able to interact

with human MHC-II. This indicates that human T cells are primed in a mouse thymus by human

MHC molecules and not only by murine cells.

Discussion

84

Spleens and livers of myeloid enhanced mice, and here in particular of huNSG-SGM3

mice, were enlarged and presented higher organ weights compared to organs of huNOG mice.

For hematological diseases the phenomenon of enlarged spleens and extramedullary hema-

topoiesis is well known. Immunodeficient mouse models transplanted with leukemic cells

(AML/CML) showed significantly increased spleens (55,132). As possible reason the unusual

proliferation of myeloid precursors and myeloid cells are stated (132). However, this is also

observed and reported for huNSG-SGM3 mice. Wunderlich et al. showed evidence of extrame-

dullary hematopoiesis for this strain, accompanied by significantly enlarged spleens after hu-

manization with human HSCs (85). They further showed that treatment with a CD33 targeting

chemotherapeutic was able to reverse spleen sizes, suggesting a macrophage activation syn-

drome a subtype of the before mentioned HLH which destroys BM cellularity. These findings

were further confirmed by Yoshihara et al. who showed that this strains develops hepatosple-

nomegaly after humanization (86). Taken together, the findings presented in this work and

previously published data suggest that infiltration of lymphocytes but especially myeloid cells

such as macrophages in spleen and liver are the reason for the increase in size and weight of

these organs.

A key finding during the characterization of human myeloid subsets was that despite a

similar genetic background huNSG-SGM3 tend to develop more granulocytes in bone marrow

and spleens whereas in huNOG-EXL myeloid precursors tend to differentiate into macro-

phages. High cytokine levels for G-CSF and mainly for GM-CSF might be responsible for the

tendency to granulocyte development in huNSG-SGM3 mice. In general, the high cytokine

levels in huNSG-SGM3 indicate that these mice are prone to inflammation, most likely due to

their unusually high T cell levels and in consequence their activation and production of these

cytokines. Together with the findings of hepatosplenomegaly and a reduced life span their

application is limited (85,86). huNOG-EXL on the other hand, offer a promising and robust

mouse model to study human myeloid cells. With a higher myeloid cell reconstitution, func-

tional macrophages as well as pDCs, physiological IgG levels and the ability to engraft different

tumor models they provide many advantages and can be applied in safety experiments or

molecule target identification.

As described above, there are other humanized mouse models which showed en-

hanced myeloid cell levels in peripheral blood, spleen and bone marrow (66,68,133). In partic-

ular, the MISTRG mouse of the Flavell group demonstrated remarkable myeloid cell levels in

humanized mice but still not comparable to fresh human blood (66). The additional knock-ins

of CSF-1 and TPO in MISTRG mice may be possible reasons for the improved differentiation

of myeloid cells. Both are reported to increase the engraftment of human hematopoietic cells

in mice and to induce myeloid cell differentiation (66,77,134). It should further be considered

that the MISTRG mouse contains a knock-in of the relevant genes, whereas the strains used

Discussion

85

in this work as are transgenic for human GM-CSF and IL-3 (and SCF). A gene knock-in means

that the gene of interest is inserted in a specific locus in the mouse genome. Usually that is the

locus where the mouse pendant is located. Through this targeted insertion the knocked-in gene

is exposed to biological expression patterns and levels (135). A transgenic approach including

a constitutively active promoter on the other side means a random insertion of the gene of

interest in the mouse genome. Therefore, this method may not faithfully display the right posi-

tion on the locus and may lead to different results (135). Nonetheless, both mouse strains

demonstrated improved myeloid cell differentiation, albeit with different distribution of immune

cell subsets. The herein performed direct comparison helps to understand essential differ-

ences between the models and serves as a practical assistance for the choice of a suitable

mouse model.

5.4 Functional characterization of human macrophages and pDCs

Human macrophages were found in huNOG-EXL mice at higher frequencies compared to hu-

NOG mice. However, their functionality was still unknown. Therefore, an ex vivo phagocytosis

assay was performed. Macrophages differentiated from huNOG-EXL bone marrow was cul-

tured for 5 days and showed FcγR-mediated phagocytosis of fluorescent labeled tumor cells.

This provided first evidence that macrophages isolated from huNOG-EXL are able to differen-

tiate in culture to macrophages that are able to phagocyte. However, the actual in vivo func-

tionality remained to be confirmed. Thus, release of human pro-inflammatory cytokines in re-

sponse to TLR8 stimulation was investigated and provided first proof for functionality of human

macrophages in huNOG-EXL mice in vivo. Ex vivo BM differentiation does not truly mimic the

situation in vivo, therefore further studies are necessary to complete this dataset. Ideally, an in

vivo phagocytosis assay based on fluorescence-bioparticles should be performed as it was

previously described (136).

Former studies in humanized mice often focused on the conventional dendritic cell pop-

ulations CD1c+ and the cross-presenting DCs (11). However, these studies were performed

in different mouse strains (137) or they used an additional enhancement with Flt3-ligand injec-

tions (138,139). Only few data concerning pDCs in any humanized mouse strain are available.

For huNOG-EXL a pDC population was earlier described albeit with rather unspecific markers

CD11c and CD123 (140). As shown herein and previously by others, several other myeloid

cells as well as CD34+ cells also express one or both markers (141,142). Using CD303, a

specific marker for pDCs, it was found that this rare subset is found mainly in the bone marrow

of huNOG-EXL and to lesser amounts also in huNOG mice. pDCs are known to be the major

type I interferon producers in reaction to viral contamination. Signaling is induced via TLR 7

and 9 (35,95). According to former studies, which used TLR agonists to stimulate pDCs, func-

tional characterization of pDCs from huNOG-EXL bone marrow was performed (11,143,144).

Discussion

86

In summary, pDCs were able to react to exogenous delivered TLR agonists. In response to

this treatment pDCs upregulated their surface activation markers and produced high amounts

of IFNα. Thus, this work is the first to describe functional human pDCs in huNOG-EXL mice.

Although, huNSG-SGM3 mice express the same human cytokines as huNOG-EXL, pDCs

were not detected in sufficient amounts in these mice. This is in accordance, with previously

published data demonstrating that GM-CSF inhibits pDC production in Flt3 bone-marrow cul-

tures (145). Increased GM-CSF levels are associated with various pathologies in particular of

hematological origin, therefore GM-CSF levels should be more closely considered in the con-

text of pDC differentiation. It is possible, that GM-CSF levels above a certain critical threshold

may impair pDC development. Indeed, Zhan et al. titrated in vitro GM-CSF concentrations in

the presence of a constant amount of Flt3-L and found, that increasing murine GM-CSF levels

reduced pDC subpopulations with 10 ng/mL of GM-CSF reaching complete pDC inhibition

(146). Hence, it can be hypothesized that NSG-SGM3 mice with systemic huGM-CSF levels

of ~ 2ng/ml in serum show much less pDCs compared to NOG-EXL mice with 5 -10 pg/ml of

huGM-CSF serum levels. Another important factor is that GM-CSF may have different roles

depending on the site of production. Although, it might be possible that GM-CSF in high levels

is able to inhibit pDC development from hematopoietic precursors, it was also shown in culture

experiments with isolated human pDCs by Ghirelli et al., that GM-CSF together with IL-3 are

essential for human pDC viability and differentiation (147). As those two cytokines are system-

ically expressed in both transgenic mice, but not in huNOG mice, the positive role on peripheral

pDC differentiation by GM-CSF should also be considered. To finally clarify the role of GM-

CSF in pDC development in humanized mice further studies are needed.

5.5 Depletion with anti-mouse CSF-1R

The question if it is possible to additionally improve myeloid differentiation in huNOG-EXL and

at the same time reduce chimerism led to experiments using an anti-CSF-1R antibody. Previ-

ously generated in-house data suggested, that murine CSF-1 is indeed able to bind to human

CSF-1 receptors albeit with a much lower affinity. Therefore, the CSF-1R blocking antibody

was used to deplete mouse macrophages and in turn increase murine CSF-1 levels. Peripheral

blood as well as primary and secondary lymphoid organs showed an increase in human mye-

loid cells (91). Macrophages in BM and spleen in higher treated groups, were prone to differ-

entiate into CD163+ macrophages, often classified as M2 macrophages, clearly indicating the

effect of CSF-1. It was previously shown, that CSF-1 is associated with M2 macrophage po-

larization, whereas GM-CSF on the other side is associated with M1 macrophages (148,149).

Unfortunately, huNOG-EXL mice developed a rapid onset of signs of anemia, bad over-

all condition with hunchback posture and body weight loss. These graft versus host disease

(GvHD)-like symptoms but especially the perturbed erythropoiesis point towards pathogenic

Discussion

87

changes in the bone marrow environment induced by αCSF-1R treatment. One possible rea-

son why macrophages are not tolerant against the mouse host lies in the incompatibility of

human SIRP-α and mouse CD47. But the integrity of the BM tissue is also a problem if too

many activated monocytes and macrophages are formed. Because the BM is the niche for the

transplanted HSCs, it is also the first organ to be infiltrated with differentiated human macro-

phages derived from myeloid precursors. This myeloid-driven xeno GvHD was also described

for huNSG-SGM3 mice, where it was shown that treatment with an anti-CD33 is able to im-

prove overall condition and survival of these mice (85). Moreover this is in accordance with

recent results, which were able to correlate the CSF-1 pathway with the onset of chronic GvHD.

They showed that CSF-1 differentiated macrophages are a crucial factor in chronic GvHD and

that the depletion of macrophages in this setting improved long term survival of mice by reduc-

ing chronic GvHD (150).

HuNOG-EXL mice have relatively high numbers of myeloid cells at baseline, so our

concern was that even a low dose treatment with the CSF-1R blocking antibody would be

sufficient to push the differentiation of monocytes into macrophages and subsequently to in-

duce myeloid-driven GvHD. Indeed, it was impossible to establish a dose that was able to

increase murine CSF-1 levels and does not affect survival of the animals. The presented re-

sults confirm that it is a huge challenge to fine-tune myeloid differentiation in humanized mice

using the CSF-1 axis. HuNOG-EXL mice showed no survival issues without αCSF-1R treat-

ment, so the question came up if myeloid cells from huNOG-EXL are perhaps not functional

without further CSF-1 differentiation. To answer this question ex vivo as well as in vivo exper-

iments were performed and it could be shown that macrophages isolated from huNOG-EXL

mice were indeed able to phagocyte tumor cells via ADCP. Additionally, it was shown that

human macrophages are able to upregulate cytokine production in response to human specific

TLR-8 stimulation in vivo, indicating that macrophages from huNOG-EXL are functional.

5.6 Comparison of tumor growth and tumor immune cell composition

Recent studies showed that myeloid tumor infiltration is a critical part in tumor fate. Many my-

eloid cells contribute to either tumor progression or regression (93,94). Well known examples

are macrophages, which contribute depending on their various phenotypes and surface mark-

ers to cancer outcome. Another cell type that was found to be correlated with survival of pa-

tients in clinical trials are pDCs (10,33,34,95). Especially in ovarian and breast cancer pDCs

were shown to contribute to tumor progression by immune suppression (10). Consequently,

two ovarian cancer cell lines and one breast cancer PDX were chosen for deeper characteri-

zation of human immune cell composition in the myeloid enhanced huNOG-EXL. The three

tumor models were able to engraft in huNOG as well as in huNOG-EXL. For the ovarian cancer

Discussion

88

cell line SK-OV-3 no differences in tumor growth was detectable between mouse strains. Con-

versely, OVCAR-5, which is also an ovarian cancer cell line showed a considerably reduced

tumor growth in huNOG-EXL mice compared to huNOG mice (Fig. 4A, Suppl. 4B). The patient-

derived breast cancer model BC_038 had a much slower tumor growth kinetic and presented

only minor differences in tumor growth.

The differences in blood and in periphery between the mouse strains were distinctive,

therefore it was also expected to find more immune cells and increased myeloid cells in tumors

of huNOG-EXL mice especially as differences in tumor growth were observed. Consequently,

the immune cell composition from myeloid enhanced NOG-EXL mice with the control huNOG

strain was compared. Depending on the tumor models a higher frequency of immune cells was

present in the tumors of huNOG-EXL mice, except for OVCAR-5. However, the effect of the

mouse strain on the immune cell infiltrate in the tumors was only minor compared to the effect

of the tumor model itself. This might be explained by the cytokine milieu of the tumor models.

Cytokines are not solitary produced by immune cells but also by tumor cells. Hence, tumor

cells are for example able to attract tumor-favorable promoting immune cells (110-112).

BC_038 and OVCAR-5 share a similar cytokine profile as well as a comparable immune cell

composition even though one is a PDX and the other an established cancer cell line from a

different tumor entity. SK-OV-3 on the other site showed a completely different tumor cytokine

composition. For example SK-OV-3 tumors not only display a rather huge population of NK

cells but also express high levels of IL-15, IL-2 as well as SDF-1. The cytokine IL-15 is known

to be essential for the differentiation, proliferation and survival of NK cells in vivo. IL-2 is also

an important growth factor for NK cell precursors and helps in the maturation of NK cells. In

addition, SDF-1a (CXCL12) was identified to recruit NK cells (151). Further, SK-OV-3 tumors

were found to be highly infiltrated by macrophages and indeed they also expressed higher

levels of macrophage migratory cytokines, such as CSF-1, GM-CSF and MIP-1β (148,149).

Simultaneously they did show no pDC content, which might be due to high GM-CSF levels as

explained above. BC_038 tumor on the other side had a higher T cell infiltration and associated

with that showed higher levels for RANTES, MIP-1α and IL-6, which are found in T cell migra-

tion pathways (152,153). OVCAR-5 is in general not a very highly infiltrated tumor which is

also reflected in the low cytokine levels. In addition, the transgenic cytokines GM-CSF as well

as IL-3 were also shown to have direct effects on tumor growth (154,155). The influence of

these two transgenes on the existing tumor cytokine milieu might further impact differences in

tumor growth and tumor immune cell composition. Therefore, it is concluded that the choice of

the right tumor model is often underestimated when designing in vivo studies with myeloid

targets.

Nevertheless, human pDCs in OVCAR-5 and BC_038 tumors were found, albeit to

much lower frequency as in the periphery. Further characterization of these tumor-associated

Discussion

89

pDCs showed upregulation of activation markers and increased IFN-α2 levels upon TLR ago-

nist treatment. This indicates that tumor-derived pDCs are as well functional. However, it was

not feasible to isolate the rare population of pDCs from the tumors due to low cell numbers.

Therefore, an in vitro assay with cell lines to show that the IFN-α2 is not produced by the tumor

cells themselves was performed. Additionally, literature concerning this topic supports the hy-

pothesis that IFN- α2 is produced by tumor-resident pDCs. pDCs are also named type I Inter-

feron-producing cells and are the major producers of IFN-α as was described many times for

many tissues. However, cellular sources of IFN type I can vary with the type of viral infection.

In infections with e.g. influenza virus, epithelial cells and alveolar macrophages in the airways

are known to provide the primary source of IFN type I. pDC on the other hand secrete IFN-α

when the virus bypasses the local barrier and becomes systemic. As the TLR agonist was

injected intratumoraly IFN-α producing effects of epithelial cells as well as alveolar macro-

phages can be excluded. In addition, other cell types such as B cells which express TLR7 and

TLR9 are unable to produce IFN-α upon activation (156), moreover they are prone to produce

IFN-β (157). A study by Bender et al. confirmed that pDCs are solitary producers of IFN-α after

4 h stimulation of human PBMCs with TLR7 or TLR9 agonists. They showed, that lymphocytes

did not respond significantly to treatment with either TLR agonist. However, monocytes were

highly stimulated by TLR8 activation and produced TNF-a, IL-6 and IL-1β but not IFN-α. They

further showed that pDCs produced TNF- α and IL-1β (158), confirming previous studies, which

independently showed that pDCs yield not only IFN-α but also TNF-α, IL-6, IL-8 and MIP-1β

upon stimulation (35,159). Given the intratumoral TLR agonist application and the early

readout 4 hours after the agonist injection, it is very likely that the observed increase in IFN-α

in tumor lysates is due to pDC. To clearly show that pDC are the IFN-α source, pDC deplet-

ing/blocking studies would be required. Further investigations of tumor immune cell infiltration

in human tumor models and especially the role of pDCs are certainly needed. IFN-α is de-

scribed to be a promising target in cancer immunotherapy as it was shown that it is able to

promote tumor regression (160).Therefore, an in vivo experiment with an IFN-α blocking anti-

body and the analysis of tumor growth after TLR treatment over time should be performed.

5.7 Outlook

In conclusion, it must be said that despite the efforts made to improve preclinical models there

are still some limitations. Further investigation is needed to develop mouse models that truth-

fully mimic human tumor environment at best in interaction with a human immune system. The

combination with a human immune system allowed the investigation of the interaction between

tumor and immune system and enables testing of immunomodulatory drugs. In terms of per-

sonalized therapy, humanized mice may be an adequate test system. A future-oriented next

step would be an autologous setting, meaning the transplantation of tumor samples and human

Discussion

90

hematopoietic stem cells from the same patients. However, such an approach is logistically

very challenging and until improved preclinical models are ready to be adopted in translational

research it will take some time.

As described herein, pDCs play a crucial role in antitumor immunity through their ability

to activate adaptive immune responses, in particular T cells. Infiltration of tumors by pDCs has

been found in many types of cancers especially in ovarian and breast cancer, however asso-

ciated with poor prognosis. In melanoma, pDC activation by TLR agonist treatment induced

functional immune responses in these tumors and inhibited tumor growth (41,42). As these

studies in melanoma proved it is important to have a model to study human tumors in the

context of human pDCs. This work presents such a tumor-bearing humanized mouse model

and lays the foundation for further experiments on pDC activation in ovarian as well as breast

cancer. Further, it is described how important the choice of the right tumor model is for inves-

tigating such complex interactions. Not only is the mouse strain an unstable variable but espe-

cially the tumor model. The conclusion from the presented study might help other groups in

planning in vivo studies for analyzing tumor immune infiltrate. Instead of buying expensive

mice or evolving time-consuming new mouse models, selection of the right tumor cell line or

PDX model can result in the right humanized mouse model with the desired characteristics. In

this work pDCs were found in two tumor models, one a cell line and one a PDX and further it

was shown, that pDCs are able to react to TLR agonist treatment in tumor. The next step would

be to assess the long-term effects of pDC infiltration and IFN-α production on tumor growth

and other immune cells in the tumor. This model can be used to further optimize dosing and

timing of TLR agonist treatment or to test TLR agonists in combination with other immunother-

apies. Furthermore, it can be used to study the interaction of tumor cytokine milieu and pDC

migration pathways by e.g. pDC depletion studies. For example, pDC depletion with the mon-

oclonal anti-BDCA2 antibody 15B could be performed, as it was shown that 15B is able to

deplete pDCs in blood and tissue, resulting consequently in impaired IFN-α release and thus

proving that pDCs are the primary source of type I IFN (161-163).

References

XCI

References

1. Ritchie MRH. Cancer. Our world in data 2020. 2. Bray F, Jemal A, Grey N, Ferlay J, Forman D. Global cancer transitions according to the Human

Development Index (2008–2030): a population-based study. The Lancet Oncology 2012;13(8):790-801 doi 10.1016/s1470-2045(12)70211-5.

3. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer 2015;136(5):E359-86 doi 10.1002/ijc.29210.

4. Kola I, Landis J. Can the pharmaceutical industry reduce attrition rates? Nature Reviews Drug Discovery 2004;3(8):711-6.

5. Hay M, Thomas DW, Craighead JL, Economides C, Rosenthal J. Clinical development success rates for investigational drugs. Nat Biotechnol 2014;32(1):40-51 doi 10.1038/nbt.2786.

6. Murphy K, Weaver C. Janeway's immunobiology. New York, NY: Garland Science/Taylor & Francis Group, LLC; 2016. xx, 904 pages p.

7. Chaplin DD. Overview of the immune response. J Allergy Clin Immunol 2010;125(2 Suppl 2):S3-23 doi 10.1016/j.jaci.2009.12.980.

8. Larsson J, Karlsson S. The role of Smad signaling in hematopoiesis. Oncogene 2005;24(37):5676-92 doi 10.1038/sj.onc.1208920.

9. Collin M, McGovern N, Haniffa M. Human dendritic cell subsets. Immunology 2013;140(1):22-30 doi 10.1111/imm.12117.

10. Koucky V, Boucek J, Fialova A. Immunology of Plasmacytoid Dendritic Cells in Solid Tumors: A Brief Review. Cancers (Basel) 2019;11(4) doi 10.3390/cancers11040470.

11. Minoda Y, Virshup I, Leal Rojas I, Haigh O, Wong Y, Miles JJ, et al. Human CD141(+) Dendritic Cell and CD1c(+) Dendritic Cell Undergo Concordant Early Genetic Programming after Activation in Humanized Mice In Vivo. Front Immunol 2017;8:1419 doi 10.3389/fimmu.2017.01419.

12. Swiecki M, Colonna M. The multifaceted biology of plasmacytoid dendritic cells. Nat Rev Immunol 2015;15(8):471-85 doi 10.1038/nri3865.

13. Klein L, Kyewski B, Allen PM, Hogquist KA. Positive and negative selection of the T cell repertoire: what thymocytes see (and don't see). Nature Reviews Immunology 2014;14(6):377-91 doi 10.1038/nri3667.

14. Ehrlich P. Über den jetztigen Stand der karzinomforschung. Ned Tijdschr Geneeskd 1908;5(273).

15. Dunn GP, Old LJ, Schreiber RD. The immunobiology of cancer immunosurveillance and immunoediting. Immunity 2004;21(2):137-48 doi 10.1016/j.immuni.2004.07.017.

16. Gavin P. Dunn ATB, Hiroaki Ikeda, Lloyd J. Old and Robert D. Schreiber. Cancer immunoediting from immunosurveillance to tumor escape. nature immunology 2002;3:991-8.

17. Muenst S, Laubli H, Soysal SD, Zippelius A, Tzankov A, Hoeller S. The immune system and cancer evasion strategies: therapeutic concepts. J Intern Med 2016;279(6):541-62 doi 10.1111/joim.12470.

18. Schreiber RD, Old LJ, Smyth MJ. Cancer immunoediting: integrating immunity's roles in cancer suppression and promotion. Science 2011;331(6024):1565-70 doi 10.1126/science.1203486.

19. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011;144(5):646-74 doi 10.1016/j.cell.2011.02.013.

20. Barnes TA, Amir E. HYPE or HOPE: the prognostic value of infiltrating immune cells in cancer. Br J Cancer 2017;117(4):451-60 doi 10.1038/bjc.2017.220.

21. Fridman WH, Pages F, Sautes-Fridman C, Galon J. The immune contexture in human tumours: impact on clinical outcome. Nat Rev Cancer 2012;12(4):298-306 doi 10.1038/nrc3245.

22. Galon J, Fridman WH, Pages F. The adaptive immunologic microenvironment in colorectal cancer: a novel perspective. Cancer Res 2007;67(5):1883-6 doi 10.1158/0008-5472.CAN-06-4806.

23. Curiel TJ, Coukos G, Zou LH, Alvarez X, Cheng P, Mottram P, et al. Specific recruitment of regulatory T cells in ovarian carcinoma fosters immune privilege and predicts reduced survival. Nat Med 2004;10(9):942-9 doi 10.1038/nm1093.

24. Gabrilovich DI, Ostrand-Rosenberg S, Bronte V. Coordinated regulation of myeloid cells by tumours. Nat Rev Immunol 2012;12(4):253-68 doi 10.1038/nri3175.

25. Mantovani A, Sica A. Macrophages, innate immunity and cancer: balance, tolerance, and diversity. Curr Opin Immunol 2010;22(2):231-7 doi 10.1016/j.coi.2010.01.009.

26. Pollard JW. Tumour-educated macrophages promote tumour progression and metastasis. Nat Rev Cancer 2004;4(1):71-8 doi 10.1038/nrc1256.

27. Murray PJ, Allen JE, Biswas SK, Fisher EA, Gilroy DW, Goerdt S, et al. Macrophage activation and polarization: nomenclature and experimental guidelines. Immunity 2014;41(1):14-20 doi 10.1016/j.immuni.2014.06.008.

28. Majety M, Runza V, Lehmann C, Hoves S, Ries CH. A drug development perspective on targeting tumor-associated myeloid cells. FEBS J 2018;285(4):763-76 doi 10.1111/febs.14277.

29. Martinez FO, Gordon S. The M1 and M2 paradigm of macrophage activation: time for reassessment. F1000Prime Rep 2014;6:13 doi 10.12703/P6-13.

30. Ries CH, Cannarile MA, Hoves S, Benz J, Wartha K, Runza V, et al. Targeting Tumor-Associated Macrophages with Anti-CSF-1R Antibody Reveals a Strategy for Cancer Therapy. Cancer Cell 2014;25(6):846-59 doi 10.1016/j.ccr.2014.05.016.

31. Biswas SK, Mantovani A. Macrophage plasticity and interaction with lymphocyte subsets: cancer as a paradigm. Nature Immunology 2010;11(10):889-96 doi 10.1038/ni.1937.

32. Labidi-Galy SI, Sisirak V, Meeus P, Gobert M, Treilleux I, Bajard A, et al. Quantitative and functional alterations of plasmacytoid dendritic cells contribute to immune tolerance in ovarian cancer. Cancer Res 2011;71(16):5423-34 doi 10.1158/0008-5472.CAN-11-0367.

33. Conrad C, Gregorio J, Wang YH, Ito T, Meller S, Hanabuchi S, et al. Plasmacytoid dendritic cells promote immunosuppression in ovarian cancer via ICOS costimulation of Foxp3(+) T-regulatory cells. Cancer Res 2012;72(20):5240-9 doi 10.1158/0008-5472.CAN-12-2271.

34. Sisirak V, Faget J, Gobert M, Goutagny N, Vey N, Treilleux I, et al. Impaired IFN-alpha production by plasmacytoid dendritic cells favors regulatory T-cell expansion that may contribute to breast cancer progression. Cancer Res 2012;72(20):5188-97 doi 10.1158/0008-5472.CAN-11-3468.

35. Decalf J, Fernandes S, Longman R, Ahloulay M, Audat F, Lefrerre F, et al. Plasmacytoid dendritic cells initiate a complex chemokine and cytokine network and are a viable drug target in chronic HCV patients. J Exp Med 2007;204(10):2423-37 doi 10.1084/jem.20070814.

36. Smith N, Rodero MP, Bekaddour N, Bondet V, Ruiz-Blanco YB, Harms M, et al. Control of TLR7-mediated type I IFN signaling in pDCs through CXCR4 engagement-A new target for lupus treatment. Sci Adv 2019;5(7) doi ARTN eaav901910.1126/sciadv.aav9019.

37. Onuora S. Connective tissue diseases: depleting plasmacytoid dendritic cells: a new therapeutic approach in SLE? Nat Rev Rheumatol 2014;10(10):573 doi 10.1038/nrrheum.2014.160.

38. Seto WK, Yuen MF. New pharmacological approaches to a functional cure of hepatitis B. Clin Liver Dis (Hoboken) 2016;8(4):83-8 doi 10.1002/cld.577.

39. Barrat FJ, Su L. A pathogenic role of plasmacytoid dendritic cells in autoimmunity and chronic viral infection. J Exp Med 2019;216(9):1974-85 doi 10.1084/jem.20181359.

40. Hirsch I, Caux C, Hasan U, Bendriss-Vermare N, Olive D. Impaired Toll-like receptor 7 and 9 signaling: from chronic viral infections to cancer. Trends Immunol 2010;31(10):391-7 doi 10.1016/j.it.2010.07.004.

41. Aspord C, Leccia MT, Charles J, Plumas J. Plasmacytoid Dendritic Cells Support Melanoma Progression by Promoting Th2 and Regulatory Immunity through OX40L and ICOSL. Cancer Immunology Research 2013;1(6):402-15 doi 10.1158/2326-6066.cir-13-0114-t.

42. Aspord C, Tramcourt L, Leloup C, Molens JP, Leccia MT, Charles J, et al. Imiquimod Inhibits Melanoma Development by Promoting pDC Cytotoxic Functions and Impeding Tumor Vascularization. J Invest Dermatol 2014;134(10):2551-61 doi 10.1038/jid.2014.194.

43. Gabrilovich DI, Bronte V, Chen SH, Colombo MP, Ochoa A, Ostrand-Rosenberg S, et al. The terminology issue for myeloid-derived suppressor cells. Cancer Res 2007;67(1):425; author reply 6 doi 10.1158/0008-5472.CAN-06-3037.

44. Papaioannou NE, Beniata OV, Vitsos P, Tsitsilonis O, Samara P. Harnessing the immune system to improve cancer therapy. Ann Transl Med 2016;4(14):261 doi 10.21037/atm.2016.04.01.

45. Sambi M, Bagheri L, Szewczuk MR. Current Challenges in Cancer Immunotherapy: Multimodal Approaches to Improve Efficacy and Patient Response Rates. J Oncol 2019;2019:4508794 doi 10.1155/2019/4508794.

46. Galluzzi L, Vacchelli E, Bravo-San Pedro JM, Buque A, Senovilla L, Baracco EE, et al. Classification of current anticancer immunotherapies. Oncotarget 2014;5(24):12472-508 doi 10.18632/oncotarget.2998.

47. Ventola CL. Cancer Immunotherapy, Part 1: Current Strategies and Agents. P T 2017;42(6):375-83.

48. Mellman I, Coukos G, Dranoff G. Cancer immunotherapy comes of age. Nature 2011;480(7378):480-9 doi 10.1038/nature10673.

49. Leach DR, Krummel MF, Allison JP. Enhancement of antitumor immunity by CTLA-4 blockade. Science 1996;271(5256):1734-6 doi 10.1126/science.271.5256.1734.

50. Farkona S, Diamandis EP, Blasutig IM. Cancer immunotherapy: the beginning of the end of cancer? BMC Med 2016;14:73 doi 10.1186/s12916-016-0623-5.

51. Klener P, Jr., Otahal P, Lateckova L, Klener P. Immunotherapy Approaches in Cancer Treatment. Curr Pharm Biotechnol 2015;16(9):771-81 doi 10.2174/1389201016666150619114554.

52. Sanmamed MF, Chester C, Melero I, Kohrt H. Defining the optimal murine models to investigate immune checkpoint blockers and their combination with other immunotherapies. Ann Oncol 2016;27(7):1190-8 doi 10.1093/annonc/mdw041.

53. Rongvaux A, Takizawa H, Strowig T, Willinger T, Eynon EE, Flavell RA, et al. Human hemato-lymphoid system mice: current use and future potential for medicine. Annu Rev Immunol 2013;31:635-74 doi 10.1146/annurev-immunol-032712-095921.

54. Walsh NC, Kenney LL, Jangalwe S, Aryee KE, Greiner DL, Brehm MA, et al. Humanized Mouse Models of Clinical Disease. Annu Rev Pathol 2017;12:187-215 doi 10.1146/annurev-pathol-052016-100332.

55. Yong KSM, Her ZS, Chen QF. Humanized Mice as Unique Tools for Human-Specific Studies. Arch Immunol Ther Ex 2018;66(4):245-66 doi 10.1007/s00005-018-0506-x.

56. Ito R, Takahashi T, Katano I, Ito M. Current advances in humanized mouse models. Cell Mol Immunol 2012;9(3):208-14 doi 10.1038/cmi.2012.2.

57. Victor Garcia J. Humanized mice for HIV and AIDS research. Curr Opin Virol 2016;19:56-64 doi 10.1016/j.coviro.2016.06.010.

58. Shinkai Y, Rathbun G, Lam KP, Oltz EM, Stewart V, Mendelsohn M, et al. RAG-2-deficient mice lack mature lymphocytes owing to inability to initiate V(D)J rearrangement. Cell 1992;68(5):855-67 doi 10.1016/0092-8674(92)90029-c.

59. Theocharides AP, Rongvaux A, Fritsch K, Flavell RA, Manz MG. Humanized hemato-lymphoid system mice. Haematologica 2016;101(1):5-19 doi 10.3324/haematol.2014.115212.

60. Kamb A. What’s wrong with our cancer models? NATURE REVIEWS | DRUG DISCOVERY 2005;4:161 - 5.

61. Shultz LD, Brehm MA, Garcia-Martinez JV, Greiner DL. Humanized mice for immune system investigation: progress, promise and challenges. Nat Rev Immunol 2012;12(11):786-98.

62. Gonzalez L, Strbo N, Podack ER. Humanized mice: novel model for studying mechanisms of human immune-based therapies. Immunol Res 2013;57(1-3):326-34.

63. Drake AC, Chen Q, Chen J. Engineering humanized mice for improved hematopoietic reconstitution. Cell Mol Immunol 2012;9(3):215-24 doi 10.1038/cmi.2012.6.

64. Legrand N, Huntington ND, Nagasawa M, Bakker AQ, Schotte R, Strick-Marchand H, et al. Functional CD47/signal regulatory protein alpha (SIRP(alpha)) interaction is required for optimal human T- and natural killer- (NK) cell homeostasis in vivo. Proc Natl Acad Sci U S A 2011;108(32):13224-9 doi 10.1073/pnas.1101398108.

65. Strowig T, Rongvaux A, Rathinam C, Takizawa H, Borsotti C, Philbrick W, et al. Transgenic expression of human signal regulatory protein alpha in Rag2-/-gamma(c)-/- mice improves engraftment of human hematopoietic cells in humanized mice. Proc Natl Acad Sci U S A 2011;108(32):13218-23 doi 10.1073/pnas.1109769108.

66. Rongvaux A, Willinger T, Martinek J, Strowig T, Gearty SV, Teichmann LL, et al. Development and function of human innate immune cells in a humanized mouse model. Nat Biotechnol 2014;32(4):364-72.

67. Qingfeng Chen MKaJ. Expression of human cytokines dramatically improves reconstitution of specific human-blood lineage cells in humanized mice. PNAS 2009;106(51):21783-8.

68. Billerbeck E, Barry WT, Mu K, Dorner M, Rice CM, Ploss A. Development of human CD4+FoxP3+ regulatory T cells in human stem cell factor-, granulocyte-macrophage colony-stimulating factor-, and interleukin-3-expressing NOD-SCID IL2Rgamma(null) humanized mice. Blood 2011;117(11):3076-86 doi 10.1182/blood-2010-08-301507.

69. Ito R, Takahashi T, Katano I, Kawai K, Kamisako T, Ogura T, et al. Establishment of a human allergy model using human IL-3/GM-CSF-transgenic NOG mice. J Immunol 2013;191(6):2890-9.

70. Huntington ND, Legrand N, Alves NL, Jaron B, Weijer K, Plet A, et al. IL-15 trans-presentation promotes human NK cell development and differentiation in vivo. J Exp Med 2009;206(1):25-34 doi 10.1084/jem.20082013.

71. Pek EA, Chan T, Reid S, Ashkar AA. Characterization and IL-15 dependence of NK cells in humanized mice. Immunobiology 2011;216(1-2):218-24 doi 10.1016/j.imbio.2010.04.008.

72. Brehm MA, Aryee K-E, Bruzenksi L, Greiner DL, Shultz LD, Keck J. Transgenic expression of human IL15 in NOD-scid IL2rgnull (NSG) mice enhances the development and survival of functional human NK cells. The Journal of Immunology 2018;200(1 Supplement):103.20-.20.

73. Herndler-Brandstetter D, Shan L, Yao Y, Stecher C, Plajer V, Lietzenmayer M, et al. Humanized mouse model supports development, function, and tissue residency of human natural killer cells. Proc Natl Acad Sci U S A 2017;114(45):E9626-E34 doi 10.1073/pnas.1705301114.

74. Cosgun KN, Rahmig S, Mende N, Reinke S, Hauber I, Schafer C, et al. Kit regulates HSC engraftment across the human-mouse species barrier. Cell Stem Cell 2014;15(2):227-38 doi 10.1016/j.stem.2014.06.001.

75. Rahmig S, Kronstein-Wiedemann R, Fohgrub J, Kronstein N, Nevmerzhitskaya A, Bornhauser M, et al. Improved Human Erythropoiesis and Platelet Formation in Humanized NSGW41 Mice. Stem Cell Reports 2016;7(4):591-601 doi 10.1016/j.stemcr.2016.08.005.

76. Strowig T, Gurer C, Ploss A, Liu YF, Arrey F, Sashihara J, et al. Priming of protective T cell responses against virus-induced tumors in mice with human immune system components. J Exp Med 2009;206(6):1423-34 doi 10.1084/jem.20081720.

77. Rathinam C, Poueymirou WT, Rojas J, Murphy AJ, Valenzuela DM, Yancopoulos GD, et al. Efficient differentiation and function of human macrophages in humanized CSF-1 mice. Blood 2011;118(11):3119-28 doi 10.1182/blood-2010-12-326926.

78. Willinger T, Rongvaux A, Strowig T, Manz MG, Flavell RA. Improving human hemato-lymphoid-system mice by cytokine knock-in gene replacement. Trends Immunol 2011;32(7):321-7 doi 10.1016/j.it.2011.04.005.

79. Willinger T, Rongvaux A, Takizawa H, Yancopoulos GD, Valenzuela DM, Murphy AJ, et al. Human IL-3/GM-CSF knock-in mice support human alveolar macrophage development and human immune responses in the lung. Proc Natl Acad Sci U S A 2011;108(6):2390-5 doi 10.1073/pnas.1019682108.

80. Maser I-P, Hoves S, Bayer C, Heidkamp G, Nimmerjahn F, Eckmann J, et al. The Tumor Milieu Promotes Functional Human Tumor-Resident Plasmacytoid Dendritic Cells in Humanized Mouse Models. Frontiers in Immunology 2020;11(2082) doi 10.3389/fimmu.2020.02082.

81. Stubenrauch K, Wessels U, Essig U, Kowalewsky F, Vogel R, Heinrich J. Characterization of murine anti-human Fab antibodies for use in an immunoassay for generic quantification of human Fab fragments in non-human serum samples including cynomolgus monkey samples. J Pharmaceut Biomed 2013;72:208-15 doi 10.1016/j.jpba.2012.08.023.

82. Stubenrauch K, Wessels U, Lenz H. Evaluation of an immunoassay for human-specific quantitation of therapeutic antibodies in serum samples from non-human primates. J Pharm Biomed Anal 2009;49(4):1003-8 doi 10.1016/j.jpba.2009.01.030.

83. Traggiai E, Chicha L, Mazzucchelli L, Bronz L, Piffaretti JC, Lanzavecchia A, et al. Development of a human adaptive immune system in cord blood cell-transplanted mice. Science 2004;304(5667):104-7 doi 10.1126/science.1093933.

84. Sampath P, Moideen K, Ranganathan UD, Bethunaickan R. Monocyte Subsets: Phenotypes and Function in Tuberculosis Infection. Front Immunol 2018;9:1726 doi 10.3389/fimmu.2018.01726.

85. Wunderlich M, Stockman C, Devarajan M, Ravishankar N, Sexton C, Kumar AR, et al. A xenograft model of macrophage activation syndrome amenable to anti-CD33 and anti-IL-6R treatment. JCI Insight 2016;1(15):e88181.

86. Yoshihara S, Li Y, Xia J, Danzl N, Sykes M, Yang YG. Posttransplant Hemophagocytic Lymphohistiocytosis Driven by Myeloid Cytokines and Vicious Cycles of T-Cell and Macrophage Activation in Humanized Mice. Front Immunol 2019;10:186.

87. Yong KSM, Her Z, Chen Q. Humanized Mice as Unique Tools for Human-Specific Studies. Arch Immunol Ther Exp (Warsz) 2018;66(4):245-66.

88. Ahmad Amanzada IAM, Martina Blaschke, Sajjad Khan, Hazir Rahman, Giuliano Ramadori, Federico Moriconi. Identification of CD68+ neutrophil granulocytes in in vitro model of acute inflammation and inflammatory bowel disease. Int J Clin Exp Pathol 2013;6(4):561-70.

89. Vermi W, Soncini M, Melocchi L, Sozzani S, Facchetti F. Plasmacytoid dendritic cells and cancer. J Leukoc Biol 2011;90(4):681-90 doi 10.1189/jlb.0411190.

90. Heidkamp GF, Sander J, Lehmann CHK, Heger L, Eissing N, Baranska A, et al. Human lymphoid organ dendritic cell identity is predominantly dictated by ontogeny, not tissue microenvironment. Sci Immunol 2016;1(6) doi 10.1126/sciimmunol.aai7677.

91. Engelhardt J. Induction of functional human macrophages in humanized mice [Dissertation]: University of Hamburg; 2016. 97 p.

92. Valabrega G, Montemurro F, Aglietta M. Trastuzumab: mechanism of action, resistance and future perspectives in HER2-overexpressing breast cancer. Ann Oncol 2007;18(6):977-84 doi 10.1093/annonc/mdl475.

93. Awad RM, De Vlaeminck Y, Maebe J, Goyvaerts C, Breckpot K. Turn Back the TIMe: Targeting Tumor Infiltrating Myeloid Cells to Revert Cancer Progression. Frontiers in Immunology 2018;9.

94. Elliott LA, Doherty GA, Sheahan K, Ryan EJ. Human Tumor-Infiltrating Myeloid Cells: Phenotypic and Functional Diversity. Front Immunol 2017;8:86 doi 10.3389/fimmu.2017.00086.

95. Labidi-Galy SI, Sisirak V, Meeus P, Gobert M, Treilleux I, Bajard A, et al. Quantitative and Functional Alterations of Plasmacytoid Dendritic Cells Contribute to Immune Tolerance in Ovarian Cancer. Cancer Research 2011;71(16):5423-34.

96. Demoulin S, Roncarati P, Delvenne P, Hubert P. Production of large numbers of plasmacytoid dendritic cells with functional activities from CD34(+) hematopoietic progenitor cells: use of interleukin-3. Experimental hematology 2012;40(4):268-78 doi 10.1016/j.exphem.2012.01.002.

97. Gilliet M, Boonstra A, Paturel C, Antonenko S, Xu XL, Trinchieri G, et al. The development of murine plasmacytoid dendritic cell precursors is differentially regulated by FLT3-ligand and granulocyte/macrophage colony-stimulating factor. J Exp Med 2002;195(7):953-8 doi 10.1084/jem.20020045.

98. Zhang H, Gregorio JD, Iwahori T, Zhang X, Choi O, Tolentino LL, et al. A distinct subset of plasmacytoid dendritic cells induces activation and differentiation of B and T lymphocytes. Proc Natl Acad Sci U S A 2017;114(8):1988-93 doi 10.1073/pnas.1610630114.

99. Bendriss-Vermare N, Chaperot L, Peoc'h M, Vanbervliet B, Jacob MC, Briere F, et al. In situ leukemic plasmacytoid dendritic cells pattern of chemokine receptors expression and in vitro migratory response. Leukemia 2004;18(9):1491-8 doi 10.1038/sj.leu.2403452.

100. Krug A, Uppaluri R, Facchetti F, Dorner BG, Sheehan KC, Schreiber RD, et al. IFN-producing cells respond to CXCR3 ligands in the presence of CXCL12 and secrete inflammatory chemokines upon activation. J Immunol 2002;169(11):6079-83.

101. Kryczek I, Wei S, Keller E, Liu R, Zou W. Stroma-derived factor (SDF-1/CXCL12) and human tumor pathogenesis. Am J Physiol Cell Physiol 2007;292(3):C987-95 doi 10.1152/ajpcell.00406.2006.

102. Penna G, Vulcano M, Sozzani S, Adorini L. Differential migration behavior and chemokine production by myeloid and plasmacytoid dendritic cells. Hum Immunol 2002;63(12):1164-71 doi 10.1016/s0198-8859(02)00755-3.

103. Shurin MR, Shurin GV, Lokshin A, Yurkovetsky ZR, Gutkin DW, Chatta G, et al. Intratumoral cytokines/chemokines/growth factors and tumor infiltrating dendritic cells: friends or enemies? Cancer Metastasis Rev 2006;25(3):333-56 doi 10.1007/s10555-006-9010-6.

104. Sozzani S, Vermi W, Del Prete A, Facchetti F. Trafficking properties of plasmacytoid dendritic cells in health and disease. Trends Immunol 2010;31(7):270-7 doi 10.1016/j.it.2010.05.004.

105. Vermi W, Bonecchi R, Facchetti F, Bianchi D, Sozzani S, Festa S, et al. Recruitment of immature plasmacytoid dendritic cells (plasmacytoid monocytes) and myeloid dendritic cells in primary cutaneous melanomas. J Pathol 2003;200(2):255-68 doi 10.1002/path.1344.

106. Hjorton K, Hagberg N, Israelsson E, Jinton L, Berggren O, Sandling JK, et al. Cytokine production by activated plasmacytoid dendritic cells and natural killer cells is suppressed by an IRAK4 inhibitor. Arthritis Res Ther 2018;20 doi ARTN 23810.1186/s13075-018-1702-0.

107. Ogata M, Ito T, Shimamoto K, Nakanishi T, Satsutani N, Miyamoto R, et al. Plasmacytoid dendritic cells have a cytokine-producing capacity to enhance ICOS ligand-mediated IL-10 production during T-cell priming. Int Immunol 2013;25(3):171-82 doi 10.1093/intimm/dxs103.

108. Kawasaki T, Kawai T. Toll-like receptor signaling pathways. Front Immunol 2014;5:461 doi 10.3389/fimmu.2014.00461.

109. Leifer CA, Medvedev AE. Molecular mechanisms of regulation of Toll-like receptor signaling. J Leukoc Biol 2016;100(5):927-41 doi 10.1189/jlb.2MR0316-117RR.

110. Binnewies M, Roberts EW, Kersten K, Chan V, Fearon DF, Merad M, et al. Understanding the tumor immune microenvironment (TIME) for effective therapy. Nat Med 2018;24(5):541-50.

111. Dranoff G. Cytokines in cancer pathogenesis and cancer therapy. Nat Rev Cancer 2004;4(1):11-22.

112. Mumm JB, Oft M. Cytokine-based transformation of immune surveillance into tumor-promoting inflammation. Oncogene 2008;27(45):5913-9.

113. Natarajan A, Mayer AT, Reeves RE, Nagamine CM, Gambhir SS. Development of Novel ImmunoPET Tracers to Image Human PD-1 Checkpoint Expression on Tumor-Infiltrating Lymphocytes in a Humanized Mouse Model. Mol Imaging Biol 2017;19(6):903-14.

114. Wang M, Yao LC, Cheng M, Cai D, Martinek J, Pan CX, et al. Humanized mice in studying efficacy and mechanisms of PD-1-targeted cancer immunotherapy. FASEB J 2018;32(3):1537-49.

115. Jones ML, Spiers H. The crowd storms the ivory tower. Nat Methods 2018;15(8):579-80. 116. Li Y, Masse-Ranson G, Garcia Z, Bruel T, Kok A, Strick-Marchand H, et al. A human immune

system mouse model with robust lymph node development. Nat Methods 2018;15(8):623-30. 117. Pearson T, Greiner DL, Shultz LD. Creation of "humanized" mice to study human immunity. Curr

Protoc Immunol 2008;Chapter 15:Unit 15 21 doi 10.1002/0471142735.im1521s81. 118. Majeti R, Park CY, Weissman IL. Identification of a hierarchy of multipotent hematopoietic

progenitors in human cord blood. Cell Stem Cell 2007;1(6):635-45 doi 10.1016/j.stem.2007.10.001.

119. Wisniewski D, Affer M, Willshire J, Clarkson B. Further phenotypic characterization of the primitive lineage− CD34+CD38−CD90+CD45RA− hematopoietic stem cell/progenitor cell sub-population isolated from cord blood, mobilized peripheral blood and patients with chronic myelogenous leukemia. Blood Cancer Journal 2011;1(9):e36-e doi 10.1038/bcj.2011.35.

120. Arrey F, Lowe D, Kuhlmann S, Kaiser P, Moura-Alves P, Krishnamoorthy G, et al. Humanized Mouse Model Mimicking Pathology of Human Tuberculosis for in vivo Evaluation of Drug Regimens. Front Immunol 2019;10:89 doi 10.3389/fimmu.2019.00089.

121. Behrens EM, Canna SW, Slade K, Rao S, Kreiger PA, Paessler M, et al. Repeated TLR9 stimulation results in macrophage activation syndrome-like disease in mice. J Clin Invest 2011;121(6):2264-77.

122. Jangalwe S, Shultz LD, Mathew A, Brehm MA. Improved B cell development in humanized NOD-scid IL2Rgamma(null) mice transgenically expressing human stem cell factor, granulocyte-macrophage colony-stimulating factor and interleukin-3. Immun Inflamm Dis 2016;4(4):427-40 doi 10.1002/iid3.124.

123. Watanabe Y, Takahashi T, Okajima A, Shiokawa M, Ishii N, Katano I, et al. The analysis of the functions of human B and T cells in humanized NOD/shi-scid/gamma c(null) (NOG) mice (hu-HSC NOG mice). Int Immunol 2009;21(7):843-58 doi 10.1093/intimm/dxp050.

124. Theocharides APA, Manz MG. Finally: development of humanized lymph nodes. Nat Methods 2018;15(8):580-2 doi 10.1038/s41592-018-0080-5.

125. Yu H, Borsotti C, Schickel JN, Zhu S, Strowig T, Eynon EE, et al. A novel humanized mouse model with significant improvement of class-switched, antigen-specific antibody production. Blood 2017;129(8):959-69 doi 10.1182/blood-2016-04-709584.

126. Covassin L, Jangalwe S, Jouvet N, Laning J, Burzenski L, Shultz LD, et al. Human immune system development and survival of non-obese diabetic (NOD)-scid IL2rgamma(null) (NSG) mice engrafted with human thymus and autologous haematopoietic stem cells. Clin Exp Immunol 2013;174(3):372-88 doi 10.1111/cei.12180.

127. Wunderlich M, Chou FS, Sexton C, Presicce P, Chougnet CA, Aliberti J, et al. Improved multilineage human hematopoietic reconstitution and function in NSGS mice. PLoS One 2018;13(12):e0209034 doi 10.1371/journal.pone.0209034.

128. Perera J, Huang H. The development and function of thymic B cells. Cell Mol Life Sci 2015;72(14):2657-63 doi 10.1007/s00018-015-1895-1.

129. Perera J, Meng L, Meng F, Huang H. Autoreactive thymic B cells are efficient antigen-presenting cells of cognate self-antigens for T cell negative selection. Proc Natl Acad Sci U S A 2013;110(42):17011-6 doi 10.1073/pnas.1313001110.

130. Xiao S, Zhang W, Manley NR. Thymic B cell development is controlled by the B potential of progenitors via both hematopoietic-intrinsic and thymic microenvironment-intrinsic regulatory mechanisms. PLoS One 2018;13(2):e0193189 doi 10.1371/journal.pone.0193189.

131. Gies V, Guffroy A, Danion F, Billaud P, Keime C, Fauny JD, et al. B cells differentiate in human thymus and express AIRE. J Allergy Clin Immunol 2017;139(3):1049-52 e12 doi 10.1016/j.jaci.2016.09.044.

132. Richter K, Pinto do OP, Hagglund AC, Wahlin A, Carlsson L. Lhx2 expression in hematopoietic progenitor/stem cells in vivo causes a chronic myeloproliferative disorder and altered globin expression. Haematologica 2003;88(12):1336-47.

133. Coughlan AM, Harmon C, Whelan S, O'Brien EC, O'Reilly VP, Crotty P, et al. Myeloid Engraftment in Humanized Mice: Impact of Granulocyte-Colony Stimulating Factor Treatment and Transgenic Mouse Strain. Stem Cells Dev 2016;25(7):530-41 doi 10.1089/scd.2015.0289.

134. Rongvaux A, Willinger T, Takizawa H, Rathinam C, Auerbach W, Murphy AJ, et al. Human thrombopoietin knockin mice efficiently support human hematopoiesis in vivo. Proc Natl Acad Sci U S A 2011;108(6):2378-83 doi 10.1073/pnas.1019524108.

135. Doyle A, McGarry MP, Lee NA, Lee JJ. The construction of transgenic and gene knockout/knockin mouse models of human disease. Transgenic Res 2012;21(2):327-49 doi 10.1007/s11248-011-9537-3.

136. Tartaro K, VanVolkenburg M, Wilkie D, Coskran TM, Kreeger JM, Kawabata TT, et al. Development of a fluorescence-based in vivo phagocytosis assay to measure mononuclear phagocyte system function in the rat. J Immunotoxicol 2015;12(3):239-46 doi 10.3109/1547691X.2014.934976.

137. Cheng L, Ma J, Li G, Su L. Humanized Mice Engrafted With Human HSC Only or HSC and Thymus Support Comparable HIV-1 Replication, Immunopathology, and Responses to ART and Immune Therapy. Front Immunol 2018;9:817.

138. Iwabuchi R, Ikeno S, Kobayashi-Ishihara M, Takeyama H, Ato M, Tsunetsugu-Yokota Y, et al. Introduction of Human Flt3-L and GM-CSF into Humanized Mice Enhances the Reconstitution and Maturation of Myeloid Dendritic Cells and the Development of Foxp3(+)CD4(+) T Cells. Front Immunol 2018;9:1042.

139. Li Y, Mention JJ, Court N, Masse-Ranson G, Toubert A, Spits H, et al. A novel Flt3-deficient HIS mouse model with selective enhancement of human DC development. Eur J Immunol 2016;46(5):1291-9.

140. Perdomo-Celis F, Medina-Moreno S, Davis H, Bryant J, Zapata JC. HIV Replication in Humanized IL-3/GM-CSF-Transgenic NOG Mice. Pathogens 2019;8(1).

141. Fromm JR. Flow cytometric analysis of CD123 is useful for immunophenotyping classical Hodgkin lymphoma. Cytometry B Clin Cytom 2011;80(2):91-9.

142. Ugo Testa EPaAF. CD123 is a membrane biomarker and a therapeutic target in hematologic malignancies. Biomarker Research 2014;2(4):1-11.

143. Audige A, Rochat MA, Li D, Ivic S, Fahrny A, Muller CKS, et al. Long-term leukocyte reconstitution in NSG mice transplanted with human cord blood hematopoietic stem and progenitor cells. BMC Immunol 2017;18(1):28.

144. Cheng L, Zhang Z, Li G, Li F, Wang L, Zhang L, et al. Human innate responses and adjuvant activity of TLR ligands in vivo in mice reconstituted with a human immune system. Vaccine 2017;35(45):6143-53.

145. Esashi E, Wang YH, Perng O, Qin XF, Liu YJ, Watowich SS. The signal transducer STAT5 inhibits plasmacytoid dendritic cell development by suppressing transcription factor IRF8. Immunity 2008;28(4):509-20 doi 10.1016/j.immuni.2008.02.013.

146. Zhan Y, Vega-Ramos J, Carrington EM, Villadangos JA, Lew AM, Xu Y. The inflammatory cytokine, GM-CSF, alters the developmental outcome of murine dendritic cells. Eur J Immunol 2012;42(11):2889-900 doi 10.1002/eji.201242477.

147. Ghirelli C, Zollinger R, Soumelis V. Systematic cytokine receptor profiling reveals GM-CSF as a novel TLR-independent activator of human plasmacytoid predendritic cells. Blood 2010;115(24):5037-40 doi 10.1182/blood-2010-01-266932.

148. Hamilton JA. Colony-stimulating factors in inflammation and autoimmunity. Nat Rev Immunol 2008;8(7):533-44 doi 10.1038/nri2356.

149. Svensson J, Jenmalm MC, Matussek A, Geffers R, Berg G, Ernerudh J. Macrophages at the fetal-maternal interface express markers of alternative activation and are induced by M-CSF and IL-10. J Immunol 2011;187(7):3671-82 doi 10.4049/jimmunol.1100130.

150. Alexander KA, Flynn R, Lineburg KE, Kuns RD, Teal BE, Olver SD, et al. CSF-1-dependant donor-derived macrophages mediate chronic graft-versus-host disease. Journal of Clinical Investigation 2014;124(10):4266-80 doi 10.1172/JCI75935.

151. Zamai L, Ponti C, Mirandola P, Gobbi G, Papa S, Galeotti L, et al. NK cells and cancer. J Immunol 2007;178(7):4011-6 doi 10.4049/jimmunol.178.7.4011.

152. Weissenbach M, Clahsen T, Weber C, Spitzer D, Wirth D, Vestweber D, et al. Interleukin-6 is a direct mediator of T cell migration. Eur J Immunol 2004;34(10):2895-906 doi 10.1002/eji.200425237.

153. Zang YC, Samanta AK, Halder JB, Hong J, Tejada-Simon MV, Rivera VM, et al. Aberrant T cell migration toward RANTES and MIP-1 alpha in patients with multiple sclerosis. Overexpression of chemokine receptor CCR5. Brain 2000;123 ( Pt 9):1874-82 doi 10.1093/brain/123.9.1874.

154. Dentelli P, Rosso A, Olgasi C, Camussi G, Brizzi MF. IL-3 is a novel target to interfere with tumor vasculature. Oncogene 2011;30(50):4930-40.

155. Hong IS. Stimulatory versus suppressive effects of GM-CSF on tumor progression in multiple cancer types. Exp Mol Med 2016;48(7):e242.

156. Guiducci C, Coffman RL, Barrat FJ. Signalling pathways leading to IFN-alpha production in human plasmacytoid dendritic cell and the possible use of agonists or antagonists of TLR7 and

TLR9 in clinical indications. J Intern Med 2009;265(1):43-57 doi 10.1111/j.1365-2796.2008.02050.x.

157. Hamilton JA, Hsu HC, Mountz JD. Role of production of type I interferons by B cells in the mechanisms and pathogenesis of systemic lupus erythematosus. Discov Med 2018;25(135):21-9.

158. Bender AT, Tzvetkov E, Pereira A, Wu Y, Kasar S, Przetak MM, et al. TLR7 and TLR8 Differentially Activate the IRF and NF-kappaB Pathways in Specific Cell Types to Promote Inflammation. Immunohorizons 2020;4(2):93-107 doi 10.4049/immunohorizons.2000002.

159. Ogata M, Ito T, Shimamoto K, Nakanishi T, Satsutani N, Miyamoto R, et al. Plasmacytoid dendritic cells have a cytokine-producing capacity to enhance ICOS ligand-mediated IL-10 production during T-cell priming. Int Immunol 2013;25(3):171-82.

160. Lee S, Margolin K. Cytokines in cancer immunotherapy. Cancers (Basel) 2011;3(4):3856-93. 161. Li G, Cheng M, Nunoya J, Cheng L, Guo H, Yu H, et al. Plasmacytoid dendritic cells suppress

HIV-1 replication but contribute to HIV-1 induced immunopathogenesis in humanized mice. PLoS Pathog 2014;10(7):e1004291 doi 10.1371/journal.ppat.1004291.

162. Li G, Zhao J, Cheng L, Jiang Q, Kan S, Qin E, et al. HIV-1 infection depletes human CD34+CD38- hematopoietic progenitor cells via pDC-dependent mechanisms. PLoS Pathog 2017;13(7):e1006505 doi 10.1371/journal.ppat.1006505.

163. Pham TNQ, Meziane O, Miah MA, Volodina O, Colas C, Beland K, et al. Flt3L-Mediated Expansion of Plasmacytoid Dendritic Cells Suppresses HIV Infection in Humanized Mice. Cell Rep 2019;29(9):2770-82 e5 doi 10.1016/j.celrep.2019.10.094.

Acknowledgement

This work has been a truly intensive experience for me and it would not have been possible

without the effort of several people to whom I am extremely grateful.

I would like to thank Roche Diagnostics GmbH for the opportunity to perform a PhD thesis in

the industry. For providing the funding which allowed me to undertake this research, but also

for giving me the opportunity to attend conferences and meetings. In particular, my gratitude

belongs to the Oncology Discovery Pharmacology department and mainly to the head of the

department Dr. Frank Herting. He created my PhD position and helped me a lot in all veterinary

affairs.

Further, I would like to thank my supervisor, Dr. Jan Eckmann, for endless scientific

discussions, for his encouragement but also for the freedom in study planning. As a result, he

made a significant contribution to the fact that my knowledge deepened and at the same time

strengthened my self-confidence.

I thank the Friedrich-Alexander University of Erlangen and my doctor father Prof. Falk

Nimmerjahn for giving me the opportunity to perform this research project.

I also thank Dr. Sabine Hoves and Dr. Gordon Heidkamp for sharing their incredible

knowledge in immunology with so much joy that inspired me genuinely. On a personal level, I

would like to thank both of them for always providing me with advice and action, not only in

immunology. Thanks belongs to Sabine also for her keen eyes and every little mistake she

found while proof reading.

I would particularly like to thank Dr. Carola H. Ries for the guidance, encouragement

and advice she has provided me even in times where she did not have to. I have been ex-

tremely fortunate to have a mentor who cared so much about my work but above all about

myself. Without her guidance and constant feedback the publication of this work would not

have been achievable.

I would also like to thank all employees of Oncology Discovery Pharmacology and CIT1

department who helped me whenever needed. In particular, I would like to thank Christa Bayer

for just everything, beginning with the possible best support during all in vivo work and ending

with the finest coffee breaks. I am also deeply thankful for the excellent technical support of

Petra Ulrich, Petra Falkner and Stefanie Lechner for in vivo assistance; Stefan Gottwald,

Monika Friedrich, Theresia Manger-Harasim and Melanie Winter for ex vivo analyses. In addi-

tion, Jakob Rosenhauer for his help with proliferation and phagocytosis assays, Afsaneh

Abdolzade-Bavil for IgG measurement, Said Aktas for his support with Spotfire and Dr. Lena

Vockentanz for the establishment and expansion of the breast cancer PDX model. It has been

a privilege to work with you all and I hope to be able to work with you again sometime.

Completing this work would have been all the more difficult were it not for the support

and the reliability of my dearest friends, who provided happy distractions to rest my mind out-

side of my research.

I would also like to say a heartfelt thank you to my mother for encouraging me to follow

my dreams and for helping me to become a strong personality.

I want to express my deepest gratitude to Michael, my husband, for his continuous sup-

port and encouragement. I was repeatedly amazed by his willingness to listen to immunobiol-

ogy topics for hours and his patience during all the ups and downs of my research. You are

my rock and always will be!

And of course, thanks to our daughter Sarah for being such a good little baby and

making it possible for me to finish what I started.

Documents

Dokumentvorlage für Diplomarbeiten