A Dissertation Entitled Approaches to Increase the Immunogenicity Of

A Dissertation

entitled

Approaches to Increase the Immunogenicity of Carbohydrate Antigens Using PS A1 and Subsequent Immunotherapies by

Kevin R. Trabbic

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the

Doctor of Philosophy Degree in

Chemistry

______Dr. Peter R. Andreana, Committee Chair

______Dr. Ronald E. Viola, Committee Member

______Dr. Katherine A. Wall, Committee Member

______Dr. Amanda C. Bryant-Friedrich, Committee Member

______Dr. Amanda C. Bryant-Friedrich, Dean College of Graduate Studies

The University of Toledo

August 2016

This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of

Approaches to Increase the Immunogenicity of Carbohydrate Antigens Using PS A1 and Subsequent Immunotherapies

Kevin R. Trabbic

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in Chemistry

The University of Toledo

August 2016

Zwitterionic polysaccharides (ZPS) are emerging as a viable alternative to protein carriers for vaccines and immunotherapeutics. PS A1 and PS B from Bacteroides fragilis

(ATCC® 25285™/NTTC® 9343™) are natural, zwitterionic carbohydrate-based polymers that can generate a CD4+ T cell mediated immune response and have recently been investigated as T cell carriers for tumor associated carbohydrate antigens (TACAs).

TACAs represent suitable targets for cancer immunotherapies because, they are expressed on virtually all cancers and are known to be weakly immunogenic. The immune response to TACAs can be increased by conjugation to immunogenic materials such as proteins or lipids. Therefore, we hypothesize using ZPS as immunogenic carriers for TACAs, can augment the immune response by generating entirely carbohydrate specific antibodies. The rationale behind this carbohydrate-based construct was to fine- tune the immune response to target carbohydrate specific lectins and generate antibodies that exclusively recognize carbohydrates without the background binding to proteins or peptides, a long outstanding problem in increasing immunogenicity. To take advantage

iii of the unique immune response to ZPS, we aimed to generate immunotherapies to target tumor glycosides.

In previous work emanating from our group, the Thomsen-Nouveau (Tn antigen,

α-D-GalNAc) was conjugated to PS A1, creating an entirely carbohydrate vaccine or immunotherapeutic (Tn-PS A1) and illustrated to have a robust immune response.

Adapting the same approach to investigating the TF antigen (Thomsen Friedenreich antigen, α-D-Gal-(1,3)-β-D-GalNAc), TF was conjugated to another ZPS PS B. The novel TF-PS B conjugate was immunized in Jax C57BL/6 mice to produce both IgG and

IgM antibody responses specific for the TF antigen. The study was concluded by showing enhanced binding to the TF-containing MCF-7 breast cancer cell line by fluorescence activated cell sorting (FACS). Additionally, TF-PS A1 elicited similar augmented immune responses to the TF antigen, which enabled in vitro cytotoxicity of tumor cells. In comparison to Tn-PS A1, both the TF-PS B and TF-PS A1 immunogens generated substantial decreased IgG antibody production, which is a main component of the mechanism for tumor elimination. However, an innovative strategy was used to increase the IgG immune responses to the TF antigen through the design and synthesis of a novel bivalent PS A1 construct design capitalizing on the knowledge gained through experimentation with first generation constructs.

The importance of cancer vaccine design and development was demonstrated through an immunological investigation of monovalent Tn- and TF-PS A1 constructs leading to a novel, unimolecular Tn-TF-PS A1 bivalent immunogen which significantly increased immunogenicity of the TF antigen (recall: TF-PS A1 did not render a high antibody titer response in mice). This additive “Tn adjuvanting effect” was also

iv demonstrated to generate enhanced IgG binding to tumor cell lines MCF-7 and OVCAR-

5 in FACS analysis and in a complement dependent cytotoxicity (CDC) assay monitoring lactate dehydrogenase (LDH) release from noted tumor cells. The results from the CDC assay demonstrated increased tumor cell lysis from Tn-TF-PS A1 sera compared to sera from monovalent vaccines Tn-PS A1 and TF-PS A1. Furthermore, a macrophage galactose lectin 2 (MGL2) assay was used, in conjunction with designed biotinylated probes, to study binding interactions of Tn and TF conjugated to PS A1 vaccine constructs. Our observations concluded that, in the case of the TF antigen, when a unimolecular bivalent Tn-TF-PS A1 immunogen was used, immunogenicity of the TF antigen was increased 50 times over a monovalent TF-PS A1 construct and resulted in a more potent and selective immune response. This work not only validated a MGL2 targeted vaccine design but the premise of which would influence other peptide, protein, or lipid vaccine designs by incorporating Tn antigen. To prove the utility of unimolecular bivalent immunogens, this model was adapted to Globo H-PS A1 construct consisting of Globo H and Tn. Similar to the biological results of Tn-TF-PS A1, the Tn-

Globo H-PS A1 immunogen produced a robust IgG immune response with cytotoxicity towards both MCF-7 and HCT-116 cancer cells.

In expanding the scope of our work, antibodies have emerged as promising cancer immunotherapies by binding specifically to tumor cells. The generation of the Tn-TF-PS

A1, TF-PS A1, and TF-PS B constructs represents entirely carbohydrate moieties that can assist in both tumor binding and killing. Therefore, the generation of monoclonal antibodies (mAb) from these constructs can provide entirely carbohydrate recognition without the influence from peptides or proteins. To this end and to validate our

v hypothesis, a IgM mAb was produced that demonstrated selective binding of the Tn antigen and showed potent in vitro and in vivo activity against the MCF-7 tumor cell line.

IgM antibodies have often demonstrated recognition of carbohydrate antigens greater than their IgG counterparts through higher avidity. Since, TACAs are present on almost all cancers, having an immunotherapy that can recognize specific glycosides may prove to be an efficient strategy against cancer. More importantly, IgM antibodies have been shown to be effective at mediating complement directed killing of tumor cells. This approach potentially offers a clinical therapeutic benefit.

I would like to dedicate this dissertation to my wife Ashley, may our lives be filled with happiness and love. I am grateful for all of your support and patience during the course of my academic desires.

Acknowledgements

I would like to thank Dr. Peter Andreana for all of his support and guidance over the course of my Ph.D. I also owe an immense amount of gratitude to my brother, Dr.

Chris Trabbic for inspiring and encouraging me to pursue a career in science. To my lab mates, thank you for your assistance and hours of intense thought provoking debate.

Table of Contents

Abstract ...... iii

Acknowledgements ...... v

Table of Contents ...... vi

List of Tables ...... xi

List of Figures ...... xii

List of Schemes ...... xv

List of Abbreviations ...... xvi

1 Introduction to Carbohydrate-Based Immunology ...... 1

1.1. Historical Perspective of Immunity and Vaccines ...... 1

1.2. Vaccine Immunology: Innate and Adaptive Immunity ...... 5

1.3. Biological Significance of Zwitterionic Polysaccharides ...... 12

1.4. Tumor Associated Carbohydrate Antigens (TACAs) ...... 20

1.5. Carbohydrate-Based Cancer Vaccines ...... 25

1.6. Structural Insights of Tn-PS A1 and Biological activity of Tn-PS A1 ...... 32

1.6.1. Conclusions ...... 39

1.6.2. Experimental ...... 39

1.6.2.1. Bacterial Growth and Isolation ...... 39

1.6.2.2. Purification of PS A1 ...... 41

1.6.2.3. Circular Dichroism...... 43 vi

1.7. References ...... 44

2 Immunological evaluation of the entirely carbohydrate-based ......

Thomsen-Friedenreich PS B conjugate ...... 55

2.1. Introduction ...... 55

2.2. Results and Discussion ...... 60

2.3. Conclusions ...... 70

2.4. Future Work ...... 72

2.5. Experimental ...... 73

2.5.1. Culturing B. fragilis and purification of PS B (2)...... 73

2.5.2. Synthesis of anomeric aminooxy TF (3)...... 74

2.5.3. Synthesis of TF-PS B (4) ...... 74

2.5.4. TF-BSA (5) ...... 74

2.5.5. NMR and MS analysis for compound (7) ...... 74

2.5.6. Analysis of compound (5) ...... 75

2.5.7. Immunizations...... 75

2.5.8. PS B poly-L-lysine (PS B-PLL) and TF-PS B poly-L-lysine ......

(TF-PSB-PLL)...... 75

2.5.9. ELISA ...... 76

2.5.10. Flow Cytometry ...... 77

2.5.11. Synthesis of STn-PS B (9) ...... 77

2.5.12. Periodate-Rescorinol Assay for Sialic Acid ...... 77

2.5.13. Percent loading of STn-PS B ...... 78

2.5.14. Alexa Fluor®488 percent loading ...... 78

vii

2.6. References ...... 79

3 Increasing the immunogenicity of the TF antigen by Targeting MGL2 Receptors ...

using a Bivalent Tn-TF-PS A1 conjugate ...... 85

3.1. Introduction ...... 85

3.2. Results and Discussion ...... 89

3.3. Conclusions ...... 100

3.4. Future Work ...... 103

3.5. Experimental ...... 104

3.5.1. Synthesis ...... 104

3.5.2. Synthesis of Tn-TF-PS A1 (4c) ...... 104

3.5.3. Biotinylated PS A1/Conjugate Probes (5a-d) ...... 104

3.5.4. 3-oxopropyl ethanethioate (mercaptoaldehyde) (8) ...... 105

3.5.5. General Procedure for TACA linkers 9 (Tn) and 10 (TF)...... 105

3.5.6. BSA-Maleimide (11) ...... 106

3.5.7. Tn-BSA (12) ...... 107

3.5.8. TF-BSA (13)...... 107

3.5.9. Immunizations...... 108

3.5.10. Enzyme Linked Immunosorbant Assay (ELISA)...... 109

3.5.11. MGL2 binding assay...... 110

3.5.12. Flow Cytometry ...... 110

3.5.13. Complement Dependent Cytotoxicity Assay ...... 111

3.5.14. Cytokine Profile...... 112

3.6. References ...... 114

viii

4 Immunological evaluation of Globo H-PS A1 conjugates ...... 119

4.1. Introduction ...... 119

4.2. Results and Discussion ...... 121

4.3. Conclusions ...... 130

4.4. Experimental ...... 131

4.4.1. GH-PS A1 (1a) ...... 131

4.4.2. Bivalent Tn-GH-PS A1 (1b) ...... 132

4.4.3. GB3-PS A1 (1c) ...... 132

4.4.4. Immunizations...... 132

4.4.5. Enzyme Linked Immunosorbent Assay (ELISA) ...... 133

4.4.6. Synthesis of GH-Thio linker ...... 133

4.4.7. Globo H-BSA...... 133

4.4.8. GB3 Thiol linker ...... 134

4.4.9. Flow Cytometry...... 134

4.5. References ...... 136

5 Murine IgM monoclonal antibody generated from Tn-PS A1 with in vivo and in ....

vitro activity ...... 138

5.1. Introduction ...... 138

5.2. Results and Discussion ...... 142

5.3. Conclusions ...... 149

5.4. Experimental ...... 152

5.4.1. Immunizations...... 152

5.4.2. Hybridoma Fusion Protocol ...... 152

5.4.3. IgM Purification ...... 153

5.4.4. Complement Dependent Cytotoxicity...... 153

5.4.6. SCID Mice tumor implantation and adoptive transfer of ......

immunotherapeutic...... 154

5.5. References ...... 155

A Supplemental Information to Chapter 1 ...... 160

A2 Supplemental Information to Chapter 2 ...... 162

A3 Supplemental Information to Chapter 3 ...... 170

A4 Supplemental Information to Chapter 4 ...... 187

List of Tables

1.1 Bacterial Polysaccharides and Conjugate Vaccines...... 3

1.2 Examples of CLRs and carbohydrate specificities...... 9

1.3 Examples of TACAs, their expression in tumors and interaction with CLRs ...... 23

1.4 Examples of TACAs vaccines, the adjuvants used, and current clinical trials ...... 28

1.5 Percent loading of AF488 on PS A1 with varying equivalents of NaIO4 ...... 38

2.1 Evaluating PS B (2) and TF-PS B (4) constructs through immunizations in Jax .....

C57BL/6J mice...... 64

2.2 Reaction of TF-ONH2 with Maleic Anhydride (MA) coated ELISA plates to ......

observe IgG immune response from TF-BSA and TF-PS B as a comparison. ....65

3.1 Cytokine kits used from Peprotech ...... 113

List of Figures

1.1 Illustration of antigen uptake and presentation to T cells by APC ...... 5

1.2 Interaction of calcium active site of CLR binding to carbohydrate domains...... 8

1.3 Cytokine production influences T cell production and proliferation ...... 12

1.4 Zwitterionic polysaccharides produced from bacteria ...... 13

1.5 Demonstration of PS A1 but not Sp 1 binds to DC-SIGN ...... 15

1.6 Confocal Microscopy Images of PS A1 interacting with MHC II ...... 16

1.7 Examples of adjuvants used to increase the immunogenicity of ......

carbohydrate-based antigens ...... 29

1.8 CD data of PS A1 and Tn-PS A1 ...... 35

1.9 Overview of PS A1 isolation and purification ...... 40

2.1 Structures of ZPS PS A1 and PS B from B. fragilis...... 58

2.2 Sialic acid determination of STn-PS B using periodate-rescorinol assay ...... 61

2.3 General description of ELISA...... 63

2.4 General description of Fluorescence Activated Cell Sorting (FACS)...... 67

2.5 IgG and IgM tumor cell binding (A) MCF-7 and (B) HCT-116 ...... 68

2.6 Cytotoxicity of MCF-7 using TF-PS B ...... 69

3.1 Synthetic modifications to PS A1 and 1H NMR overlay of PS A1 conjugates .....90

3.2 ELISA antibody specificity of TACA conjugates ...... 92

3.3 MGL2 binding assay and inhibition using probes 5a-d ...... 95 xii

3.4 Flow cytometry with anti-serum from 1 and 4a-c with secondary Alexa Fluor® ....

488 anti-IgG using human tumor cell lines...... 96

3.5 Antibody mediated CDC with anti-serum from 1 and 4a-4c ...... 97

3.6 Cytokine profile for SAS and TMG immunizations of PS A1, Tn-PS A1, ......

TF-PS A1, and Tn-TF-PS A1 ...... 99

3.7 Immune modulation by carbohydrate addition relative to PS A1 ...... 102

4.1 Synthesis of GH-PS A1 Conjugates 1A-C. A) Globo H-PS A1 B) Bivalent ......

Tn-GH-PS A1 and C) GB3-PS A1 ...... 122

4.2 The IgG and IgM immune response from Globo H conjugates...... 124

4.3 Cross reactivity of IgG and IgM antibodies from GH-PS A1 conjugates to ......

GB3-BSA...... 125

4.4 The immune response generated from GB3-PS A1 and recognition of ......

GB3-BSA ...... 126

4.5 Cross reactivity of anti-serum (1:100 dilution) of GH-PS A1 constructs to......

blood group A and blood group B...... 127

4.6 Flow cytometry with anti-serum from PS A1, Globo H-PS A1, and Tn-PS A1 ......

with secondary Alexa Fluor® 488 anti-IgG using human tumor cell lines ...... 128

4.7 Antibody mediated CDC with anti-serum from PS A1, Globo H-PS A1, and ......

Tn-PS A1 plus rabbit complement...... 130

5.1 The production and screening of mAbs ...... 143

5.2 Antibody Titration of Kt-IgM-8 on ELISA for effective concentration ...... 143

5.3 Carbohydrate Specificity for Kt-IgM-8 ...... 144

5.4 Flow Cytometry of Kt-8-IgM binding to A) MCF-7 and B) HCT-116 ...... 145

xiii

5.5 CDC activity of KT-IgM-8 on MCF-7 cells ...... 146

5.6 Kt-IgM-8 displays tumor volume (mm3) reduction of MCF-7 Tumors in SCID .....

mice for 39 days ...... 148

xiv

List of Schemes

1.1 Synthesis of Tn-PS A1 by the oxidation and conjugation to PS A1...... 33

1.2 Conjugation of AF488-hydrazide to PS A1 ...... 38

2.1 Semi-synthetic TF-PS B (4) immunogen ...... 59

2.2 Conjugation of TF-linker (7) to BSA-Maleimide...... 65

3.1 General synthesis of Tn-BSA and TF-BSA ...... 91

3.2 Syntheses of biotinylated TACA-PS A1 (5a-c) from TACA-conjugates (4a-c) as ...

MGL2 assay probes ...... 94

List of Abbreviations

ADCC ...... Antibody Dependent Cellular Cytotoxicity APC ...... Antigen Presenting Cells

BSA ...... Bovine Serum Albumin

CD ...... Circular Dichroism CD4 ...... Cluster of Differentiation 4 CD8 ...... Cluster of Differentiation 8 CDC ...... Complement Dependent Cytotoxicity CLR ...... C-type Lectin Receptors CMP ...... Cytidine 5’-Monophosphate CRM197 ...... Non-toxic mutant of DT CSC ...... Cancer Stem Cells

DC ...... Dendritic Cell DC-SIGN ...... Dendritic Cell-Specific Intercellular adhesion molecule-3- Grabbing Non-integrin DMEM ...... Dulbecco's Modified Eagle's medium DT ...... Diphtheria Toxin

EDTA ...... Ethylenediaminetetraacetic acid ELISA ...... Enzyme Linked Immunosorbent Assay

FACS...... Fluorescence-activated cell sorting FBS ...... Fetal Bovine Serum FDA...... Federal Drug Administration

GalNAc ...... N-acetylgalactosamine GH ...... Globo H

I.P ...... Intraperitoneal

xvi

IFN-γ ...... Interferon gamma IL ...... Interleukin

KLH ...... Keyhole Limpet Hemocyanin

LDH ...... Lactate Dehydrogenase LPS ...... Lipid Polysaccharides

MA ...... Maleic Anhydride mAb...... Monoclonal Antibody MAC ...... Membrane Attack Complex MAL ...... Maleimide MALDI-TOF...... Matrix Assisted Laser Desorption/Ionization-Time of Flight MGL ...... Macrophage Galactose Lectin MHC ...... Major histocompatibility complex MPLA ...... Monophosphoryl Lipid A MR ...... Mannose Receptor

NADH/NAD ...... Nicotinamide Adenine Dinucleotide Neu5Ac ...... N-acetylneuraminic acid or sialic acid NHS...... N-hydroxysuccinimide NK cells ...... Natural Killer Cells NO ...... Nitric Oxide

OD ...... Optical Density OMPC ...... Outer Membrane Protein Complex OVA ...... Ovalbumin

PAMPS ...... Pathogen Associated Molecular Patterns PBS ...... Phosphate Buffered Saline PEG ...... Polyethylene Glycol PLL ...... Poly-L-Lysine PS A1 ...... Polysaccharide A1 PS B ...... Polysaccharide A1

RNS ...... Reactive Nitrogen Species ROS ...... Reactive Oxygen Species

SAS ...... Sigma Aldrich Adjuvant® SCID ...... Severe Combined Immunodeficiency

xvii

Siglec...... Sialic Acid Binding Ig-like Lectins T reg ...... T regulatory Cell TACA ...... Tumor Associated Carbohydrate Antigens TCR ...... T Cell Receptor TF ...... Thomsen-Friedenreich antigen Th ...... T helper Cells TLR ...... Toll-like Receptors TMG ...... TiterMax Gold® Tn ...... Thomsen-nouveau antigen TT ...... Tetanus Toxoid

UDP...... Uridine Diphosphate

VSSP ...... Very Small Size Proteoliposomes

ZPS ...... Zwitterionic Polysaccharides

xviii

Chapter 1

Introduction to Carbohydrate-Based Immunology

1.1 Historical Perspective of Immunity and Vaccines

The first scientifically documented case of vaccine investigation began with

Edward Jenner in 1796,1 in which he administered protection against small pox by exposing a child to the cow pox virus followed by the small pox virus in subsequent days.

Unbeknownst at the time, his scientific observations for conferring immunity would radically change the approach that medicine uses to protect against virulent agents. Over the course of the next century, many scientists such as Louis Pasteur, the developer of a vaccine for chicken cholera2, had modified the scientific approaches Jenner conceived for prophylactic vaccines by the attenuation of virulent agents. This approach has set the golden standard for the induction of immunity by vaccination against pathogenic materials as many vaccines have been approved such as small pox, measles, and polio.

However, the attenuation of pathogenic material did not always produce desired immunity, and a new approach was discovered by the isolation of bacterial polysaccharides.

While investigating a strain of Pneumococcus, Dochez and Avery observed a soluble substance shed into bacterial growth media, which was purified until no remaining nitrogen was present.3 This substance reacted with antisera from infected patients, causing a precipitate.4 This substance would eventually be identified as a polysaccharide-based material. Ultimately, this polysaccharide-based material was shown by Francis and Tillet in 1930 to induce immunity in patients.5 These scientific observations gave relevancy to the biological importance of carbohydrates and polysaccharides.

One of the first examples of bioconjugate chemistry involving carbohydrate materials was documented when Goebel and Avery conjugated the Pneumococcus polysaccharide to a purified globulin from horse sera through use of a diazonium derivative of the polysaccharide.6 This polysaccharide-protein conjugate demonstrated a strong immune response in rabbits, where the polysaccharide was previously noted to be non-immunogenic.7 The pioneering work by Avery and Goebel in the development of polysaccharide-based bacterial vaccines eventually declined for the more favorable approach of using small molecule antibiotics to treat bacterial infections.8 The concepts of vaccines and immunity were not entirely understood at the time, and antibiotics appeared to be the antidote for many infections. With time, the use of antibiotics led to a large amount of antibiotic-resistant bacterial strains, which has become a major public health concern of the 21st century. Therefore, the recurring nature of ideas has come full circle and the employment and advancement of vaccines is at the forefront of contemporary medicine. Currently only bacterial- and viral-based prophylactic vaccines have been approved by the Federal Drug Administration (FDA). Table 1.1 depicts the 2

Table 1.1. Bacterial Polysaccharide and Conjugate Vaccines

Bacterial Polysaccharide Vaccines Bacteria Vaccine Polysaccharides 1,2,3,4,5,6B,7F,8,9N,9V, 10A, Streptococcus Pneumovax 23 (Merck) 11A, 12F, 14, 15B, 17F, 18C, 19F, pneumoniae 19A, 20, 22F, 23F, 33Fa Neiseria meningitidis AC Vax (GSK) A and C Menomune (Aventis) A,C,W-135, Yb

Salmonella typhi Typhim Vi (Aventis) Vic

Bacterial Polysaccharide Conjugate Bacteria Vaccine Protein Conjugate Haemophilus HibTITER (Wyeth influenzae type b Hib-CRM197d Vaccines) (Hib) PedvaxHIB (Merck) Hib-OMPCd

ActHB (Sanofi Pasteur) Hib-TTd

Streptococcus (1,3,4,5,6A,6B,7F,9V, Prevnar 13 (Wyeth) pneumoniae 14,18C,19A,19F, 23F)-CRM197e Neiseria meningitidis Menveo (Novartis) (A,C,Y, W-135)-CRM197

Menactra (Sanofi (A,C,Y, W-135)-DT Pasteur)

a. Capsular polysaccharides isolated from 23 serotypes of S. pneumoniae. These polysaccharides compose a 23 valent vaccine and are unconjugated. b. Menomune is composed of 4 capsular polysaccharides isolated from serogroups A, C, W-135, and Y from Neiseria meningitides. This vaccine helps prevent bacterial meningitis. c. Typhim Vi is composed of a surface polysaccharide Vi from S. typhi Ty2 strain. It is a vaccine for Typhoid fever. d. Vaccines HibTITER, PedvaxHIB, ActHB are polysaccharide-protein conjugate vaccines to prevent bacterial meningitis. CRM-197 is a detoxified carrier protein of diphtheria toxin. TT is tentanus toxid, it is used to increase the immune response of carbohydrate antigens. OMPC is an outer membrane protein complex used to increase the immune response of carbohydrate antigens.

major components of approved bacterial polysaccharides and polysaccharide-protein

conjugate vaccines.9 It is important to note that the approaches used for bacterial-based 3

vaccines do not necessarily translate to cancer vaccines due to the immunosuppression of self-antigens.

Cancer immunotherapy relies on biological material (T cells, B cells, and antibodies) in order to assist in the removal of cancer cells, however the approach has not been as successful as it has been for bacterial-based vaccines.10 In 1891, William Coley published the first successful clinical treatment of cancer with a 10% success rate through use of heat-killed Streptococci and Serratia marcescens in sarcoma of the soft tissue.11

The mechanism of this approach is not entirely understood but the inactivated bacteria could have generated cross reactive antibodies or acted as an adjuvant (lipids and polysaccharides) creating a proinflammatory environment to stimulate an immune response towards tumors. This event is representative of the current direction of cancer immunotherapies through using biological material to assist the immune system in elimination of cancer.

Paul Ehrlich was one of the first pioneers to believe that the immune system could be harnessed to suppress and recognize many carcinomas.12 These ideas led Sir

MacFarlane Burnet and Lewis Thomas to introduce and develop the theory of immune surveillance. Burnet believed that tumor-specific antigens could provoke an anti-tumor immune response in order to eliminate developing tumors. However, Thomas’ views were more evolutionary in nature due to the conclusion that complex organisms must have developed protective mechanisms in order to combat abnormal cell growth. These ideas eventually stagnated, but technology led to an expansion of what Ehrlich, Thomas, and Burnet hypothesized to include the innate and adaptive immune responses. It was discovered that the immune system promotes survival of tumor cells that have a reduced 4

expression of immunogenic markers, which can ultimately evade the immune response.

In a type of natural selection, the tumor cells with reduced cancer antigens can escape

immune recognition to continue uncontrolled proliferation. These ideas led to the concept

of cancer immunoediting, which is the process by which the body is able to control tumor

cell proliferation that if left unchecked would lead to cancer.

1.2. Vaccine Immunology: Innate and Adaptive Immunity

In order to further understand and develop the fundamental concepts of vaccine

immunology and the chemical design of vaccines, the innate and adaptive immune

responses (figure 1.1) need to be adequately described in relation to carbohydrate-based

antigens. The innate immune response is the first line of defense towards invading

pathogens and oncogenesis. One of the first responses of innate immunity is

inflammation caused by white blood cells. Its main purpose is to limit the spread of the

Figure 1.1. Illustration of antigen uptake and presentation to T cells by APC.

infection. The first cells that are present at the first sign of infection are phagocytes antigen presenting cells (APCs), which are composed of macrophages and dendritic cells.

Macrophages are derived from monocytes and are essential in the capture of pathogens; the pathogens are phagocytized by APCs and subsequently degraded by proteases, reactive oxygen species (ROS) and reactive nitrogen species (RNS). Immunogenic components of processed pathogens are able to bind to either major histocompatibility complex I or II, MHC I and MHC II respectively. The antigenic fragment complexed with MHC is presented on the surface of APC to T cells and recognized by the T cell receptor (TCR). In addition to an MHC interaction with TCR, co-stimulation between

CD28 on T cells and either CD80 or CD86 will facilitate T cell proliferation. The antigenic presentation and subsequent T cell activation signifies a key interaction between the innate and adaptive immune responses.

There has been extensive research oriented around the evidence for carbohydrate

T cell immunity (Chapter 1.3). This largely involves bacterial-based zwitterionic polysaccharides (ZPS) that form carbohydrate-protein complexes with MHC II.

However, ZPS or polysaccharides have not been shown to interact with MHC I but there is an MHC-I-like family of CD1 glycoproteins that can bind glycolipids. These interactions predominantly occur between hydrophobic glycolipids such as self- GM3- ceramide and GB3-ceramide or foreign glycolipids such phosphatidylinositol mannosides

(mycobacterium tuberculosis) or αGal-ceramide (marine sponges). Then the hydrophobic lipid component binds in the hydrophobic CD1 binding groove, which exposes the carbohydrate portion to bind to CD1 restricted T cells or a polyclonal T cell receptor on invariant natural killer T cells (iNKT cells).13 Transitioning from the 6

adaptive immune response to the innate immune response, natural killer cells (NK cells)

(not to be confused with iNKT cells) are activated independently of MHC proteins. NK cells are important for the immunosurveillance of cancer cells, inpart due to surface expressed lectin proteins Siglec 7 and Siglec 9 which recognize sialic acid.14 The role and importance of carbohydrate based immunity depends on the innate immune response.

The innate immune receptors recognize pathogen associated molecular patterns

(PAMPS) by pattern recognition receptors (PPR). PPRs consist of Toll-like receptors

(TLRs), C-type lectins receptors (CLR), and NOD-like receptors (NLR). TLRs and CLRs are membrane-bound receptors, which contribute to cell signaling and antigen internalization. These surface receptors are not clonally distributed meaning the genes do not undergo somatic recombination to alter protein expression in response pathogens.15

TLRs recognize a broad array of pathogen membrane components such as lipid polysaccharides (LPS), capsular polysaccharides, bacterial DNA, RNA (double stranded and single stranded), and glycoproteins. There are five major TLRs that are present on the plasma membrane of innate immune cells, TLR 1, 2, 4, 5, and 6, and three that are present in intracellular compartments of endosomes, TLR 3, 7, and 9.

The TLRs on the surface of cells are critical for recognizing cell wall components of pathogens which are primarily composed of bacterial polysaccharides and LPS, which are often expressed in response to pathogens. The endosomal TLRs are essential for recognizing viral and foreign DNA. Upon pathogen interaction with TLRs, the receptors initiate a signaling transduction cascade, which stimulates the transcription factor NF-κB.

This leads to the production of pro-inflammatory and anti-inflammatory cytokines and chemokines such as TNF-α, IL-1, and IL-6. These cytokines and chemokines then recruit 7

and stimulate immune cells.16 Since TLRs play a large role in the initiation of immune

responses, they have been investigated as immune adjuvants in both cancer and pathogen

related vaccines. For example, monophosphoryl lipid A (MPLA) commonly found in

gram negative bacteria, is a receptor agonist for TLR4.17 MPLA has been conjugated to

tumor associated carbohydrate antigens (TACA) to form a self-adjuvanting vaccine,

which influenced TLR-4 stimulated cytokine production such as IL-4 and IL-12.18

Figure 1.2. Interaction of calcium active site of CLR binding to carbohydrate domains.21-22 Adapted and modified with permission of Oxford University Press and Frontiers.

Another immune receptor group that was once believed to be a scavenger receptor, but

has been shown to be potent immune stimulators, is CLRs.19 This is due to the fact that

not only can they increase antigen uptake, but also stimulate the production of pro-

inflammatory cytokines.

CLRs have a calcium (Ca2+) active site, which can bind carbohydrate antigens by

having either 6, 7, or 8 coordination sites to the proteins.20 The calcium ion is stabilized

by the protein side chains (asparagine and glutamine), carbonyls, and water. When a

carbohydrate binds to the lectins, calcium active sites, two hydroxyls coordinate and stabilize the interaction, and subsequently facilitate antigen endocytosis. Figure 1.221 depicts (A) glucose binding to bovine minicle through the 3’ and 4’ hydroxyls and (B) molecular modeling of macrophage galactose lectin (MGL) binding to a monosaccharide component of N-acetyl galactosamine.22 There are many CLRs with known carbohydrate specificity and pathogen preference (Table 1.2). These receptors include dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-

SIGN), langerin, macrophage galactose type c-type lectin (MGL), dectin-1, dectin-2, minicle, and mannose receptor (MR). These receptors have been extensively studied and have demonstrated known carbohydrate specificity, which results in broad recognition of carbohydrate containing pathogens including viral, fungal, parasite, and bacterial-based antigens (Table 1.2).19 In addition to antigen binding to and endocytosis stimulated by

Table 1.2. Examples of CLRs and carbohydrate specificities

CLRs Carbohydrate Specificity Pathogen Interaction HIV, M. tuberculosis, S. DC-SIGN High Mannose/Fucose Mansoni MGL Terminal GalNAc/Galactose S. mansoni, Filoviruses

Dectin-1 β-1,3 glucans S.cerevisiae, C. albicans C. albicans , S.cerevisiae, A. Dectin-2 α-Mannose/Fucose/GlcNAc fumigatus C. albicans , S.cerevisiae, A. Minicle α-Mannose/Fucose/GlcNAc fumigatus High Mannose/Fucose / Sulphated C. albicans, M. tuberculosis, T. MR Sugars cruzi High Mannose/Fucose/GlcNAc/β- C. albicans , S.cerevisiae, M. Langerin 1,3 glucans furfur Dec-205 Unknown Dead Cells, Y. petis

CLRs, they can also initiate a signal transduction cascades, which can influence cytokine production.

These cytokines can help influence the differentiation of T helper cells. For example, interaction with dectin-1 can influence the production of Th1 and Th17 cells by production of pro-inflammatory cytokines IL-1β, IL-6, IL-12, IL-23, and reduction of IL-

10 through a dectin-1 specific agonist curdlan (β-1,3 glycan).23 Contrarily, spontaneous interactions of zymosan (β-1,3 glycan) with both dectin-1 and TLR 2 augments the production of IL-10, which leads to the production of T regulatory cells (T regs) and reduces the Th1 and Th17 activation observed with curdlan.23 In conclusion, either innate receptors alone or with co-stimulation from TLRs are prime pathways to be exploited to augment the immune response against carbohydrate-based antigens.

The innate and adaptive immune systems converge to provide immunity against invading pathogens. The adaptive immune response primarily involves T and B cells.

Unlike the innate immune response, the adaptive immune response has clonal selection of receptors associated with T and B cells where receptors can adapt to increase affinity towards an antigen.24 The main mechanism for the diversity in clonal selection is the random rearrangement of the variable (V), diversity (D), and joining (J) gene segments, resulting in specific antigen binding receptors for immunoglobulins and T cell receptor loci.24-25 Antibodies (IgG, IgM, IgA, IgE, and IgD) generated from B cells which can bind to the pathogen, results in removal by mechanisms including neutralization, phagocytosis/opsonization of pathogen-antibody complex, and antibody-induced cytotoxicity. Upon binding to a pathogen and/or tumor cells, an antibody can induce cytotoxicity by antibody dependent cellular cytotoxicity (ADCC) or complement 10

dependent cytotoxicity (CDC). For ADCC, an IgG antibody can bind to its target antigen and an NK cell binds the Fc part of IgG antibody through FcγRIII. This initiates the release of cytotoxic endocytic vesicles that consist of granzymes and perforin, which initiates cell death.26 The CDC pathway takes advantage of the complement system, which is a vital component of both the innate and adaptive immune systems. Once IgG and IgM antibodies bind to the pathogen, complement protein C1q is able to bind to the

Fc region of the antibody, which signals a cascade event recruiting other complement proteins to eventually initiate a membrane attack complex (MAC). The MAC is able to form a pore in the plasma membrane resulting in cytotoxic activity.27 Both ADCC and

CDC are important functional mechanisms by which the innate immune response can interact with the adaptive immune response using NK cells and the complement system.

The other arm of the adaptive immune response is cellular-based and involves T cells. When a naïve T cell interacts with an antigen-presenting cell (B cell, macrophage, or dendritic cell), it can either interact with MHC I or MHC II, which correspond to either

CD8+ or CD4+ T cells, respectively. Upon this stimulation, either the cytotoxic CD8+ or helper CD4+ will proliferate and assist in memory for the immune response in order to help removal of the pathogen. CD8+ T cells are cytotoxic and can recognize MHC I with and antigen in the peptide binding site to stimulate the release of granzymes and perforin in order to initiate cytotoxicity. Similarly, the CD4 + helper cell interacts with MHC II plus peptide and initiates the release of cytokines, which can be seen in Figure 1.3. There are four subsets of T helper cells (Th): Th1, Th2, Th17, and T reg, and these helper cells influence B cells to produce antibodies and aid in the development of CD8 responses and the clearance of pathogens. Collectively, the immune response involves both the innate 11

and adaptive immune response. This helps the recognition and removal of pathogens, which is important for immunity.

Figure 1.3. Cytokine production influences T cell production and proliferation

1.3. Biological Significance of Zwitterionic Polysaccharides

Large polysaccharides are known to be both T cell independent and T cell dependent since they produce an antibody response with the help of B-cells and without the influence of MHC II (T cell independent)28, or because they require T cell help for isotype switching (T cell dependent).29 There are two different types of T cell independent responses i) T cell independent 1 (TI-1) and ii) T cell independent 2 (TI-2).

TI-1 is considered to be mitogenic towards B cells and stimulates polyclonal B cell 12

activation through Toll-like receptors, such has lipid polysaccharides (LPS).30 However,

TI-2 responses involve a large repeating biological polymer such as bacterial polysaccharides and elicit a large B cell response. The T cell independent paradigm changes when zwitterionic polysaccharides (ZPS) are examined for their immunological responses.

ZPS have been shown to initiate a CD4+ T cell response, which was first shown by the laboratory of Dennis Kasper from Harvard Medical School. ZPSs appear to be an extremely privileged polymeric scaffold that only appears in four strains of pathogenic bacteria including: Bacteroides fragilis (PS A1 and PS B), Streptococcus pneumoniae

(SP1), Staphylococcus aureus (CP5 and CP8 both have partial deacetylation of NHAc on

L-FucNAc), and Morganella morganii (O-chain antigen) (Figure 1.4).

PS A1, (Figure 1.4) which comes from Bacteroides fragilis (ATCC 25285/NCTC

Figure 1.4. Zwitterionic polysaccharides produced from bacteria.

9343), is only one of six zwitterionic polysaccharides known to induce a specific T cell dependent and highly prized immune response.31 PS A1 contains a repeating zwitterionic tetrasaccharide unit that contains a [3- 2,4-dideoxy-4-amino-D-N-acetylfucose (1-4), D-

N- acetylgalactosamine(1-3), D-galactopyranose(1-3), D-galactofuranose with a 4,6- pyruvate acetal (1-].32 This carbohydrate scaffold draws comparisons to proteins in two distinct ways: a) size and b) immune recognition. First, PS A1 has been shown to adapt an alpha helical character, which is a common characteristic of proteins and can be determined by circular dichroism.33 Second, it can be recognized by the immune system and processed via MHC II, which was once thought to exclusively bind peptide fragments.34 The biological significance of B. fragilis is intraperitoneal implantation, which involves abscess formation in which CD4+ T cells, macrophages, and neutrophils are thought to be recruited to assist; a classical T cell response.31,34,35 In order to prove T cell activation in early studies, mice were first immunized with PS A1 and then T cells were adoptively transferred to confer protection against B. fragilis infections. These initial results spurred the investigation of how a polysaccharide could initiate a T cell mediated response. To add a further layer of support for T cell action, similar to other

ZPSs, chemical modifications that neutralized the charges on PS A1 abrogated T cell activation.

For further explanation of its biological mechanism, PS A1 was recently identified to first interact with CLR DC-SIGN and TLR-2 on DCs/or APCs.36 The authors proposed that the galactofuranose portion of PS A1 is the most available to interact with DC-SIGN.36 To support this assumption, ZPS SP 1 was shown to have minimal binding compared to PS A1 (Figure 1.5) in an ELISA-based assay. In order to 14

confirm their hypothesis with a cell-based assay, a Raji cell line was transfected with DC-

SIGN, and was incubated with PS A1 and EGTA to chelate calcium (Figure 1.5). PS A1

Figure 1.5. Demonstration of PS A1 but not Sp 1 binds to DC-SIGN. Reproduced with permission of The Royal Society of Chemistry. showed binding DC-SIGN-Raji cells, but no interaction with the controls; additional by,

DC-SIGN binding was abrogated with EGTA.

Once PS A1 is endocytosed by DC-SIGN, it is subjected to reactive nitrogen and oxygen species in order for oxidative cleavage to occur and for it to render short carbohydrate sequences (5-15 kDa). Sequences below 5 kDa were shown to not stimulate a T cell dependent response. These short fragments bind to human leukocyte antigen-DM (HLA-DM), which transfers the processed polysaccharide to the MHC II molecule receptor where it is transported and expressed on the surface of the cell for αβ-

TCR-mediated recognition.34,37 To support PS A1 interaction with MHC II and presentation on the surface of APCs, Figure 1.6A shows confocal microscopy images of

macrophages incubated and stained with fluorescently labeled PS A1 (red)

Figure 1.6. Confocal Microscopy Images of PS A1. interacting with (A) MHC II, (B) Charge neutralized NHAc-PS A1, carbodimide reduced PS A1, and fully neutralized PS A1. (C) PS A1 (yellow) interacting with T cell receptors (green) and MHC II (red). Adapted and modified with permission of Oxford University Press. interacting with MHC II stained (green) and presented to the surface of the cell

(yellow).38 Additionally, to support T cell engagement with MHC II, Figure 1.6C shows confocal microscopy images of interaction of APCs (red) and T cells (green) and when

PS A1 is incubated in these cell cultures, an immunological synapse can be seen in yellow.34

The process of oxidative cleavage is dependent on the APC having an inducible nitric oxide synthase (iNOS) gene, which is a distinct processing method from enzymatic proteolytic cleavage.39 The radically induced depolymerization pathway proceeds 16

through deamination cleavage by nitric oxide (NO) radicals. In iNOS-/- deficient mice,

PS A1-NO (synthetically prepared) was able to activate CD4+ T cells, whereas native PS

A1 was unable to produce a T cell response.39

Furthermore, structural modifications to PS A1 were investigated by neutralization of either the free amino sugar by N-acetylation or by a carbodiimide reduction of the carboxyl group.40 In both cases, neutralization of the charges caused an abrogation of T cell mediated responses, which confirmed that the zwitterionic character of PS A1 was solely responsible for its unique biological activity.40 To confirm the effects of abrogated T cell activity, Figure 1.6B shows the confocal microscopy images of the effects of charge neutralization of PS A1. In all three images, N-acetylated-PS A1, carbodiimide reduced PS A1, and complete charge neutralized of PS A1 derivatives showed antigen endocytosis but no presentation on to the surface of the APC by MHC II, as seen in Figure 1.6A. Charge neutralization of other ZPS such as PS B also causes abrogation of T cell responses. The authors showed N-acetylation of the amine of 2- aminoethyl phosphate on PS B, and not reduction of the carboxyl group (due to charges still present from phosphate groups ), caused abrogation of T cell mediated responses.41

Interestingly, de-N-acetylation of PS B N-acetylated sugars, β-D-GlcNAc, β-D-QuiNac, and α-L-QuiNAc also abrogated the T cell activation presumably by increasing the density of amine groups.41 However, the notion of increased density of amine groups may not be a conclusive point due to harsh acidic conditions used for de-N-acetylation, which will additionally cause depolymerization of the polysaccharide.

The other zwitterionic polysaccharides, such as SP 1, have been noted to induce

CD4+42 and CD8+CD28- T cells42 in C57BL/6 mice. CP5 and CP8 both induce intra- 17

abdominal abscesses, which signifies a potent T cell response.43 For the CP8 ZPS induced potent CD4+ T cell responses,43 it was determined that CP5 showed a higher degree of N-acetylation than CP844, thus giving merit to the structural requirements of PS

A1 for T cell activation. The greater zwitterionic character leads to T cell activation. To further support this concept, chemically induced modification also revealed a loss of intra-abdominal abscess formation.43 Finally, the ZPS from M. morganii has also been shown to interact with MHC II, and stimulate T cell activation.45 An interesting structural requirement for MHC II binding was resolved when it was determined that the zwitterionic nature of the phosphocholine retains the ability to bind to MHC II after acetylation of the ZPS.45 In summary, in order for a ZPS to bind to MHC II, the polymer must contain alternating charges in order to provide binding stability for molecules.

A Group B Streptococcal type III polysaccharide-protein conjugate has been shown to be recognized by CD4+ T cell clones,46 which indicate that non-zwitterionic polysaccharides have the ability to be recognized by T cells. The charge sequence of the zwitterionic motif of the polysaccharide allows it to bind the HLA-DR family in humans,37 as opposed to other non-charged capsular polysaccharide such as mengicoccal group B polysaccharide (MGBP) that is weakly immunogenic and is most likely a T cell independent response.47 Therefore the zwitterionic motif is vital in initiating the CD4+ T cell mediated responses, due to the fact that studies have revealed that anionic or cationic derivatives are relegated to the humoral immune arm exclusively.

When the processed PS A1 (5-15 kDa) fragment is presented on MHC II, it acts as a ligand for the CD4+ T cells to bind and form an MHC II-carbohydrate-T-cell interaction instead of the conventional MHC II-peptide-T-cell interaction. The naïve cell 18

will develop effector functions through cell differentiation, and will develop into one of the four T cell helper subsets: Th1, Th2, Th17, and T regulatory (T reg) helper (Figure

1.3).48,49 The environmental cytokines are responsible for eliciting the effector functions in T cells. The genes in the APC encoding some cytokines are up-regulated when the T cell binds to the MHC complex on the cell surface of the APC. An additional signal is required by the co-stimulation receptors, B7 from the APC and CD28 from the T cell.

Once co-stimulation has been achieved, the APC can release cytokines that direct the development of helper T cells and lead to immunoglobulin isotype switching.

The release of cytokines by T cells is a hallmark of T cell activation and can be monitored by quantifying cytokine release from respective T cell populations. The activation and proliferation of T cells are initiated by cytokines produced by each individual subset (Th1, Th2, Th17, and T reg) (Figure 1.3). The differentiation of Th1 requires cytokines IL-12 and IFN-γ produced from the APCs and the T cell secretes IL-2 and IFN-γ (Figure 1.3), which are usually induced by pathogenic microbes.50 IL-2 signals the T cell for continual proliferation and increases the number of effector cells.

Similarly, the APC releases IL-4 in the activation of the Th2 helper cell, and the T cell consequently releases IL-4 and IL-5.50 These cytokines are required for B-cell differentiation and the recruitment of other macrophages. However, a PS A1 response initiates IL-2, IL-6, and IL-17 when it is administered alone.35,51 T regs and Th17 share the common cytokine TGF-β, which could result in their differentiation, but also inhibits the induction of Th1 and Th2 (Figure 1.3). An activation of Th17 would require the presence of IL-6 and TGF-β, IL-6 has been shown to impede T reg expansion.52 53 Both

T regs and Th17 can induce each other’s’ differentiation, which is known as 19

plasticity.54,55 The Th17 subset has a different effector function from the Th1/Th2 activation. The activation of Th17 will produce IL-17, 21, and 22, which could result in recruiting neutrophils to the site of infection or the induction of isotype switching in B- cells.56 Additionally, IL-23 promotes the activity and survival of Th17 cells, which produce IL-17.57,58 It is clear that the path of cytokine activation is an essential player in

T cell activation. Identifying the T cell signature is a clear indication if a specific conjugate vaccine involves T cell polarization; a critical element in evaluating our current vaccine constructs and those proposed for the future.

1.4. Tumor Associated Carbohydrate Antigens (TACAs)

Cell surface glycosides are important biological molecules, which are used for cell surface adhesion, membrane organization, and are involved in directing proper protein folding. The synthesis of carbohydrate chains through the glycosylation machinery is critical to biological function. Glycosyltransferases play an essential role by elongating glycan chains in the endoplasmic reticulum (ER), which are ultimately transported to the surface of cells. There are three main types of glycosylated biomolecules, which are O-linked glycoproteins, N-linked glycoproteins, and glycolipids.

The common monosaccharide building blocks, which are predominately α-linked to proteins or β-linked to ceramides, are N-acetylgalactosamine (GalNAc), N- acetylglucosamine (GlcNAc), galactose (Gal), mannose (Man), fucose (Fuc), glucose

(Glc), and N-acetylneuraminic acid (Neu5Ac) or sialic acid. The general mechanism for oligosaccharide elongation begins with the transfer of a monosaccharide from an activated nucleotide donor to an acceptor substrate forming a glyosidic linkage. There are two main types of linkages found to form glycoproteins. The first is O-linked to 20

serine or threonine, and the other is N-linked through asparagine, which is linked through an amide forming an N-linked oligosaccharide. Further modification occurs in the Golgi apparatus to the glycolipid/glycoprotein of both O-linked and N-linked carbohydrates.

For example, on normal cells a GalNAc-containing glycoprotein will be glycosylated with galactose from UDP-GAL via T-synthase (C1β3GalT) and a sialic acid cap from

CMP-sialic acid from ST3Gal I.

The biosynthesis of glycolipids is similar to glycoproteins, but the ceramide is synthesized in the ER and is transported to the Golgi apparatus. The glucose is then attached to the ceramide by UDP-Glc-ceramide glucosyltransferase (GlcT). The glycolipid is further elongated to incorporate galactose from UDP-Gal to form lactose ceramide (LacCer).59 LacCer is a main component of ganglioside synthesis, which can be further elongated in order to form more complex glycolipids by sialyltransferases (α-

2,3, α-2,6, and α-2,8). This forms GD2, GD3, and GM3 among other gangliosides. Once finished in the Golgi, the newly synthesized gangliosides are transported to the plasma membrane. The Lewis histoblood group antigens are found in epithelial tissues and can either be O/N-linked to protein or lipids. The basis for the Lewis histoblood group antigens is formed through the elongation of two structural isomer cores consisting of either type 1 (Galβ1-3GlcNAc) or type 2 (Galβ1-4GlcNAc). Elongation of the disaccharide structural isomer by fucose (fucosyltransferases, FucT I-VI) and sialic acid

(sialyltransferases, ST3Gal III-VI) acid leads to the type 1 core producing Lea /Leb/SLea, and type 2 leads to the production of Lex/Ley/SLex. As an indication of tumorigenesis, improper folding or increased expression of glycosyltransferases (Table 1.3) occurs, which leads to aberrant glycosylation. 21

A common observation in cancer is the aberrant glycosylation of proteins or lipids forming O-linked mucins, Lewis histoblood group antigens, globosides, and gangliosides

(Table 1.3). The underlying cause of aberrant glycosylation is the improper folding/upregulation of glycotransferases. In the case of both the Tn (Thomsen-nouveau) and TF (Thomsen-Friedenreich) antigens, the chaperone protein Cosmc (core 1 β3GalT specific molecular chaperone) incorrectly folds the β3-galactosyltransferase or T- synthase leading to TF, whereas a mutation in Cosmc leads to increased expression of

Tn.62 Additionally, STn can be observed in mutations of Cosmc leading to production of

α 1-6 sugar by ST6GalNAc-II. The Cosmc mutation occurs due to point mutations in the open reading frame, hypermethylation in the promoters, and gene mutations.63 These

TACAs are often referred as the mucin-related antigens because they are O-linked to either serine or threonine, and are expressed primarily on breast, ovarian, and colon cancers. Other TACAs, such as the Lewis histoblood group antigens, are predominately expressed in colon and lung cancers. Some of the Lewis antigens have normal cell expression, but are regulated by increased expression of FucT and STGal transferases, which leads to increased expression of Lewisx, Lewisy, Lewisa, Lewisb and their sialyated derivatives. The globosides are commonly found on many types of tumors such as breast, ovarian, and prostate. The increased expression of Globo H and its sialylic acid substituent SSEA4 results from the increased expression of B3GALNT1, FucTI/FucTII, and ST3Gal. Furthermore, the ganglioside family is β-linked to ceramide, which are

GM3, GD3, and GM1, and are the result of increased sialytransferases (ST8Sia I and

ST3GalV) and B4GalNT1. When TACAs are observed on the surface of tumor cells, their presence 22

Table 1.3. Examples of TACAs, their expression in tumors and interaction with CLRs.

Mechanism of CLR/ Lectin TACA60 Major types of Cancers Expression Interactions O-linked Mucins Tn Breast, Colon, Gastric, Prostate , Cosmc MGL Pancreas , Ovarian mutation, ↓ C1β3GalT TF Breast, Colon, Bladder, Prostate ↓ β6-GlcNAc-T, Galectin 3 ↓sialyation STn Gastric, Colon, Breast , prostate, ↑ST6Gal1, Siglecs, Galectins pancreatic, Ovarian Cosmc mutation Lewis Antigens60,61 Ley Colon, Lung ↑FucT II DC-SIGN, Langerin Lex Colon, Lung ↑FucT VI DC-SIGN SLex Breast, Colon, Kidney, Ovarian ↑FucT VI and MR FucT VII Lea Colon, Lung ↑FucT III DC-SIGN, MBL SLea Breast ↑FucT III DC-SIGN Leb Colon, Pancreatic ↑FucT I DC-SIGN, Langerin, MBL Globosides61 Globo H Breast, Ovarian, Prostate ↑B3GALNT1 MR, DC-SIGN, and FucT I/FucT Dectin-1, Langerin II SSEA4 Breast, Brain, Pancreatic, ↑ST3Gal II N/A Prostate, Ovarian Gangliosides GD2 Melanoma, Neuroblastoma, ↑B4GALNT1 Siglec Glioma GD3 Melanoma,Neuroblastoma ↑ST8Sia I Siglec GM3 Melanoma, Breast, Renal ↑ST3Gal V Siglec

alone can help evade the immune response.

Tumor glycosides can lead to immunosuppression and are considered autoantigens because they are self-antigens, which is one of the explanations for how tumor cells can evade the immune response. A reasonable explanation to describe how some TACAs evade immune responses is through the innate immune response by lectins and CLRs. The precise mechanism to CLR immune evasion and tumor eradication is often convoluted due to the promiscuity of CLRs and co-stimulation by TLRs, which leads to pro-inflammatory and anti-inflammatory cytokines. However the O-linked mucins, especially the Tn antigen, have the ability to interact with MGL and can promote tumor invasiveness.64 The cell signaling based on interactions with MGL depends on the nature of a ligand. For example, a Tn-MUC 13TR (peptide) was able to be internalized and interact with MHC II ligands. However a Tn-MUC 116TR (protein) was internalized, but did not co-localize with MHC II.65 Furthermore, DC-SIGN can interact with Lea,66

Leb,66 Lex,67 Ley68, and a partial Globo H trisaccharide-containing fucose.61 These interactions have been shown to augment anti-tumor immunity and suppress immune responses by the production of IL-10, thus promoting an immunosuppressive environment. Langerin, which is found on Langerhan cells, has been shown to bind to

Leb and Ley69, but further elucidation of the signaling pathway needs to be performed.

Despite extensive analysis of the gangliosides, TACAs binding to CLR still need to be determined, but siglecs70 which bind sialic acid have been shown to promote survival of tumor cells by NK cell inhibition. In conclusion, the presence of sialic acid generally leads to immune tolerance and decreased immunogenicity.

Revisiting the concept of cancer immunoediting, NK cells play a major role in tumor elimination. However, the presence of sialic acid can stop the immune response and cause immunosuppression by the generation of T reg cells and decreased function of

NK cells. To prove this effect, Van Kooyk et al. showed that B16SLC35A1 sialic acid knockdown mice become more susceptible to attack from effector cells compared to the regulatory responses generated from sialic acid positive B16 cells. The down regulation of sialyltransferases causes a reduction in cell migration and increased adhesion in breast cancer cell lines, which signifies reduced tumor metastases.71 Support of the idea that carbohydrates have a regulatory role is evidenced with antibodies as noted in the observation that when an antibody is sialyated, it can induce tolerance, but in the absence of sialic acid the antibody is proinflammatory.72 The expression of sialic acid on antibodies is usually generated in response to a self-antigen, which leads to immune tolerance stimulating conditions. Modulating signals through CLRs have showed how tumors and other pathogens can escape immune recognition. Lea/Leb antigens on colorectal tumors have been shown to interact with DC-SIGN and also stimulate production of anti-inflammatory cytokines.66 The TACAs initiate a T cell independent response when they are administered alone, however, when formulated in conjugated vaccines or immunotherapies they can overcome immune tolerance.

1.5. Carbohydrate-Based Cancer Vaccines

Whole tumor cells were first used in the 1970’s as vaccines to induce protective anti-tumor responses.73 This approach is similar to immunizing with attenuated pathogens to provide protection but incorporates inactivated cancer cells instead. The use of whole tumor cells is a straightforward method which can produce a broad immune 25

response to tumor-associated antigens that are composed of both glycolipids and glycoproteins. There are two types of whole tumor cell vaccinations, autologous and allogenic. With autologous immunizations, a patient’s own tumor cells are used. The primary advantage of this approach is that the cancer cell will contain several different tumor antigens on the surface of the cell, and therefore produce an immune response to multiple antigens.74 Since tumors are composed of many different cell variants, it may be difficult to focus on specific antigens due to the overall low concentration of antigens.

With allogenic immunizations, an established cell line presents quantifiable concentrations of antigens on the cell. Collectively, immunizations with both autologous and allogenic tumor cells require an adjuvant therapy to perpetuate the immune response, which has shown clinically relevant results. With the aforementioned issues concerning antigen characterization, more chemically defined immunotherapies that can elicit a strong and specific immune response are often preferred.

TACAs are broadly expressed on a variety of tumor cells, which consequently allows for them to serve as a legitimate subject with which an immune response can be provoked. The increase of the immunogenicity of carbohydrate antigens is focused around two parameters: 1) conjugation of carbohydrate antigens to immunogenic proteins or other immune stimulating molecules as carriers, and 2) the use of select adjuvants to augment the immune response to TACA conjugates through the prolongation of the duration and magnitude of antibody responses in order to promote protective immune responses.

Immunogenic proteins, polysaccharides, and lipids have been used as synthetic and semi-synthetic carriers to increase the immunogenicity of carbohydrate antigens 26

(Table 1.4). Proteins have received a majority of the attention due in part to their robust immunogenicity and FDA approval of many bacterial-based conjugate vaccines containing DT, TT, CRM197, and OMPC (Table 1.1). However, KLH (keyhole limpet hemocyanin), a carrier protein derived from the hemolymph of the Giant Keyhole

Limpet, has been the focus of numerous clinical trials. The most notable example of a

KLH conjugate is THERATOPE® (STn-KLH), which was designed by Biomira Inc. for breast cancer and subsequently failed in phase III clinical trials.75 Collectively, these proteins have been conjugated to a myriad of TACAs to direct the immune response against carbohydrate antigens towards a T cell dependent mechanism (Table 1.4).76

The use of adjuvants (Figure 1.7) utilizes immune stimulating agents in order to enhance the adaptive immune response by targeting innate immune receptors. Adjuvants can vary in composition from inactivated Mycobacterium tuberculosis (Freund’s complete adjuvant) to inorganic salts such as aluminum salts, lipid components from bacteria, or oil in water emulsions. The benefits of using adjuvants include decreasing the amount of vaccine required while increasing specific antibody responses, and polarizing

T cell responses compared to non-adjuvant vaccines.97 Freund’s complete adjuvant contains extracellular components, and lipid and carbohydrate-based components, to stimulate the innate immune response to result in a Th1 and Th17 immune response.

Alum (potassium aluminum sulfate) is an FDA approved adjuvant whose mechanism of action has been elusive, but augments the immune response thus increasing leukocyte recruitment, and cytokine98 (IL-1β, IL-1, IL-5, IL-6) and chemokine production99 (CCL2 and CCL4),

Table 1.4. Examples of TACAs vaccines, the adjuvants used, and current clinical trials.

Monomeric O-linked TACA constructs Adjuvant Clinical Trials Tn Tn-KLH77 QS-21 Phase I/II

Tn-PAM QS-21 N/A Tn-PS A178 TMG N/A TF TF-KLH79 Detox/QS-21 Phase I STn STn-KLH (Theratope®)80 Detox Phase III Clustered STn-KLH81 QS-21 Phase I Monomeric Lewis Groups constructs Lewis Y Lewis Y-KLH82 QS-21 Phase I Lewis X Lewis Y-Lewis X-KLH QS-21 N/A (heterodimer) 83 Monomeric Globosides Constructs Globo H84 Globo H- (TT/DT/BSA/KLH) C34 Phase I Globo H-KLH QS-21 Phase III SSEA4 SSEA4-CRM197 C34 N/A Monomeric Gangliosides Constructs GD2 GD2-KLH85 MPLA Phase I GD2-L-KLH86 QS-21 Phase I GD3 GD3-KLH87 QS-21 N/A GM3 GM3-OMPC CFA/IFA N/A NGc-GM3-VSSP88 N/A Phase I/II Polyvalent Vaccines (Pooled monomers) MUC 1-G5, Tn, STn, TF, KLH QS-21 Phase I Lewis Y, GM389 Lewis Y, GM3, MUC 1, KLH QS-21 N/A MUC 290 Unimolecular Vaccine Globo H, GM2, STn, TF, KLH QS-21 Phase I Tn91 Multi-Component STn STn-Linker-MPLA92 TMG N/A STn-Linker-MPLA N/A N/A Globo H Globo H-linker-MPLA18 N/A N/A Tn Tn-YAF peptide-L-rhamnose93 TMG or SAS N/A PAM3Cys-aminobutyl-di-Tn94 N/A N/A Lewis Y PAM3Cys-peptide-tri-Lewis Y95 QS-21 N/A Tn/TF/STF PAM3CysSK4-ethylene glycol- N/A N/A MUC1 Tn Rha-TLRL-MUC1-Tn-liposome96 SAS N/A

increases antigen uptake,100 and promotes Th2 responses.101 QS-21 is a naturally isolated saponin from Quillaja saponaria. It promotes Th1/Th2 responses, and although the mechanism of action is not entirely known, structural based studies have revealed the

Figure 1.7. Examples of adjuvants used to increase the immunogenicity of carbohydrate-based antigens. aldehyde present is essential for function.102 It is thought that the aldehyde binds to the T Figure 1.6. Examples of adjuvants used to increase the immunogenicity of carbohydrate- cellbased receptor antigens. via a Schiff base (imine linkage), and QS-21 does not interact with either

TLR-2 and TLR-4.102 Furthermore, oil in water emulsions such as TiterMax Gold, which is composed of conjugated polymers of hydrophilic chains (polyoxyethylene), and hydrophobic chains (polyoxpropylene), provide slow release of antigen while recruiting leukocytes to the site of administration. For example, Globo H was conjugated to KLH,

BSA, DT, CRM-197, and TT, and the immune response was adjuvanted with a series of

α-galactosylceramide derivatives but the C34 (Figure 1.7) derivative showed increased

IgG1 antibodies to all of the protein constructs compared to vaccines using QS-21.103

However DT and TT are main components of diphtheria-pertussis-tetanus vaccines, which may result in a decrease in clinical relevance by carbohydrate epitope suppression

due to previous immunizations.104

The use of polyvalent vaccines (mixtures of pooled monomeric vaccines) have shown similar immune responses of both IgG and IgM antibodies compared to monomeric (individual) components without significantly reducing the immune response.90 Moreover, use of the polyvalent vaccines consumes significant quantities of antigen without significant benefit. For example, the polyvalent vaccines of LeY, GM3,

MUC 1, and MUC 2-KLH used 20 µg per injection, whereas monomeric components used 5 µg. In order to streamline the manufacturing and scalability of multi-antigenic vaccine constructs, Livingston and co-workers engineered several unimolecular polyvalent vaccines,105,106 which contained several TACAs through a single linker on

KLH. One of the multivalent unimolecular constructs incorporated five individual glycosylamino acids (Globo H, GM2, STn, TF, and Tn) that were to be assembled to form a pentavalent glycosylated peptide backbone, which was conjugated to KLH via a thiol-maleimide linkage.105 In humans, the pentavalent unimolecular KLH was able to generate IgG and IgM antibodies to at least three of the five antigens in phase I clinical trials.91

Alternatively, lipids such as MPLA have been conjugated to TACAs to create a fully synthetic self-adjuvanting cancer vaccine that utilizes TLR4 as an agonist to influence cytokine production and B cell proliferation.18 A T cell dependent mechanism has also been described for these self-adjuvanting hybrid carbohydrate lipid molecules by the release of cytokines IL-4, IL-12, and IFN-γ.18 Other fully synthetic approaches use

MAG, RAFTs, and dendrimers to increase the immunogenicity of TACAs. Essentially, the design of these constructs is to saturate the carrier molecules with a high density of 30

TACAs. This usually results in a T cell mediated response with the help of peptides, which are known MHC II binding motifs. These creative approaches, such as a dendrimeric (MAG:PV-Tn) construct, uses a tripeptide lysine scaffold conjugated to four

T cell epitopes with twelve Tn antigens. This construct was tested in mice, which resulted in protection against Ta3/Ha tumors.107 Furthermore, fully synthetic regioselectively addressable functionalized template (RAFT) contain a cyclodecapeptide scaffold which contains two β turns by proline-glycine motifs, which incorporated Tn-

108 ONH2 through an in situ aldehyde generation from serine forming an oxime linkage.

Additionally, the RAFT scaffold contains a T cell epitope, which elicits a T cell mediated response; a common theme for increasing the immunogenicity of carbohydrate antigens by using T cell epitopes.

One of the more fascinating approaches for vaccine development involves targeting innate immune receptors on dendritic cells. These innate immune receptors, or

CLRs, can act as a bridge between the innate and adaptive immune response. These classes of molecules can be divided into two different types based on mechanism of action: direct and indirect innate immune stimulators. A direct innate immune stimulation can be described as an agonist that directly interacts with the immune receptor such as a CLR. Binding to the mannose receptor demonstrates an increased antigen uptake and higher proinflammatory cytokine production. Although no TACA- based vaccine has been utilized with this approach, the results remain promising with

Cutler and Bundle’s demonstration of a construct design that can facilitate protection against Candida albicans by appending mannose to the protein.55 However, the heterodimer Ley-Lex-KLH vaccine showed a ten-fold decrease in immune response 31

compared to monomeric Ley-KLH, and did not generate a sufficient immune response to

Lex.83 In a separate investigation, an Lex-OVA conjugate has been shown to bind to DC-

SIGN, but it did not produce a regulatory immune response denoted by cytokine production.109 However, in conjunction with a TLR-4 LPS adjuvant, the Lex-OVA conjugate increased production of IL-10, which denoted a regulatory immune response.109 Demonstrating the full versatility of targeting innate immune cells, a cancer vaccine was produced from the fusion of human chorionic gonadotropin beta chain

(hCGβ) to mAb B11 (specificity towards mannose receptor) in conjugation with a TLR agonist, which resulted in Th1 immunity.110 In regards to the innate immune response, bacterial-based antigens appear to have a clear advantage over other pathogens due to the densely populated bacterial polysaccharides on their surface. This is a fundamental reason as to why ZPS can be used as a vaccine construct.

1.6. Structural Insights and Biological activity of Tn-PS A1.

The Andreana research group designed the use of an entirely carbohydrate vaccine that incorporated the Tn antigen conjugated to ZPS PS A1 in order to increase the immunogenicity of that particular TACA.78,111 In order to conjugate Tn to PS A1, the

Tn-aminooxy sugar was synthesized to form an oxime linkage with oxidized PS A1

(Scheme 1.1). The oxime linkage was favored over hydrazone and imine linkages due to the hydrolytic stability of the oxime linkage at physiological pH.112 Hydrolysis of the hydrazone is favored over the oxime due to the lower electronegativity of the nitrogen, which is more readily protonated compared to the oxygen of the oxime.112 In accordance to vaccine design, the oxime linkage provides stability even in acidic environments (pH

3-4) which TACA-PS A1 encounters in the lysosomes after antigen uptake.

Alternatively, if PS A1 were conjugated with TACAs containing either hydrazine or hydrazides functional groups, the hydrazone linkage would be more susceptible to hydrolysis and would decrease TACA density on PS A1, which would decrease the immune response to the TACA hapten.

The primary advantage of using PS A1 as an immunogenic carrier of carbohydrate-based antigens is that it represents an entirely carbohydrate-based vaccine

Scheme 1.1. Synthesis of Tn-PS A1 by the oxidation and conjugation to PS A1. that is able to initiate T cell dependent responses. Our group’s previous work had also shown that the Tn antigen conjugated to PS A1 elicits an IL-17 cytokine response that correlates to Th17 helper T cell activation.111 Th17 helper T cells have gained interest over the past several years because of IL-17, which has been shown to have antitumor functions by the adoptive transfer of Th17 cells. This transfer is essential in the co- stimulation of CD8+ cell-mediated killing of tumor cells.113 The Andreana group was the first to demonstrate that Tn antigen conjugated to PS A1 could elicit a selective and specific immune response in mice. This interaction released cytokines IL-10, IL-17a, IL-

4, and IL-2.111 Cytokine profiling assays can determine which cytokines are being produced from the immune response, and provide insight into a particular T cell signature. Currently, the mechanism of Tn-PS A1 processing is unclear, but it is most likely similar to the proposed mechanism of PS A1 by radically induced depolymerization, which generates 5-15 kDa fragments to interact with MHC II.

With our newly formed Tn-PS A1 vaccine construct, we sought in the current study, to understand the impact of Tn coupling on the helical content of PS A1 through circular dichroism (CD) studies; a common technique applied to monitor optical properties of the solution structure of molecules which can reveal information about secondary structure. Spectra are known to vary with protein conformation upon ligand binding or thermodynamic unfolding. CD spectroscopy uses optical rotation and changes in elipticity to determine the difference of asymmetric monosaccharides in solution, which can provide structural information such as conformations and generation of dipole moments.16-17

PS A1 is naturally substituted with specific chromophores containing defined optical activity such as the N-acetate groups, which are a common post-glycosylational marker. The conformational changes of PS A1 may affect the polarizability, static field contributions, and orientation of the chromophores, which would result in changes in optical activity. The optically active absorption bands of the substituted chromophores arise from the n–π* transitions located in the range 200-250 nm.114-116 Therefore, circular dichroism (CD) could provide a convenient method of investigating the conformational change of polysaccharide substituents. In light of our promising biological results and

what is based on the aforementioned, we elected to characterize and compare the structural integrity of PS A1 in our Tn-PS A1 vaccine construct utilizing CD.

To this end, we sought to determine whether the Tn antigen conjugated to PS A1 would affect its helical character by comparing the CD data of pure PS A1 with Tn-PS

A1 at room temperature (Figure 1.8A). CD spectra of Tn-PS A1 showed a distinct decrease in ellipticity as compared to that of pure PS A1 at wavelength 217 nm in a 0.1

A Conditions: Tn, Tn-PS A1, and PS A1 and B Conditions: Tn-PS A1, Temperature 30 C, Temperature 30 C and pH = 7 and pH = 2-7

Figure 1.8. CD data of PS A1 and Tn-PS A1. A) CD spectra of PS A1 (1), Tn (2) and Tn-PS A1 (3) at 30 0C and pH = 7.0. B) CD spectra of Tn-PS A1 at varying pH values. mg/mL concentration. That loss of helical character resulting from Tn conjugation to PS

A1 could be due to a change in conformation and/or a direct result of the oxime chromophore.

To confirm that the change in ellipticity observed with Tn-PS A1 was due to a direct effect of the Tn monosaccharide attached to oxidized PS A1, the CD of Tn was compared with PS A1 and Tn-PS A1 (Figure 1.8A). Tn alone did not give any significant

CD spectrum. Therefore, it was clear that the α-helical character of Tn-PS A1 came entirely from PS A1. 35

Within the APC, a number of events must take place in order to facilitate antigen processing, and then MHC II loading followed by presentation. First, it is critical for the pH to be reduced to fairly acidic conditions, usually in the 4.5-5.0 range by the action of an ATPase proton pump. This drop in pH triggers activation of lysosomal enzymes. It is known that blockage of acidification results in a loss of T-cell activation mediated by

ZPS.117 In vitro binding studies show that pH does not play a significant role in ZPS binding to MHC II proteins.9 Furthermore, in vitro oxidative ZPS depolymerisation cannot occur at acidic pH, which indicates that processing must occur before acidification and protease activation. According to Kasper, ZPS molecules must be internalized, processed at neutral pH through oxidation, and co-localized with the MHC II proteins in an acidic environment in order to facilitate MHC II binding and presentation to T cells.9

However, the positive MHC II mediated immune responses against our Tn-PS A1 suggests that the construct retains the ability to induce production of IgG antibodies even when it has lost the helical character due to changes in conformation. Based on our data and in support of Kasper’s investigation on charge neutralization of PS A1, we now believe that it is critical to retain the alternating charge character of Tn-PS A1 and not preserve the α-helical character. In order to further validate our data, the CD spectra of

Tn-PS A1 were collected at varying pH values. The results showed identical spectra for

Tn-PS A1 at pH 5 and 7, while results for decreasing the pH showed a gradual loss of α- helical content (Figure 1.8B).

These results indicate that the degree of α-helicity of Tn-PS A1 did not change when the pH of the solution was adjusted from 7 to 5. A better way to compare secondary structure at pH 7 and 5 is to execute the coupling reactions at the respective pH 36

and compare the CD of those with conjugates. In order for this method to have validity, the percent loading of Tn to oxidized PS A1 (Scheme 1.1) at pH 7 and 5 must be equal.

In addition to structural confirmation of Tn-PS A1, we devised simple but reliable methods in order to determine the Tn loading to PS A1.

An accurate assessment of the molar ratio of carbohydrate antigen conjugated to

PS A1 is important for the reproducibility and conducting the dose-response immunization studies. There are numerous methods to determine the number of antigens coupled to various carrier proteins. For example, the number of STn carbohydrate antigens bound to a carrier protein can be determined by either a resorcinol assay or fluorescence detection of 1,2-diamino-4,5-methylenedioxybenzyl (DMB) derivatives of sialic acid.36, 37 The approaches for determining the extent of coupling for antigen- protein conjugates can be placed into three broad classes: radioactive tracer, spectrophotometry/mass spectrometry and amino acid analysis.118 In radioassays, a small of amount carbon-14 is incorporated during the synthesis. Alternatively, radiolabeling can be performed using iodine. However, the accuracy of this method depends on the extent of labelling, and suffers from the instability of commonly used radiolabeled derivatives. Therefore the radiolabeling approach to Tn-PS A1 conjugate by reducing the

3 oxime bond by [H ]-NaBH4 was avoided. The amino acid analysis method, although robust, is unsuitable for our completely carbohydrate-based Tn-PS A1 conjugate.

Ultimately, we focused on the development of a new coupling reaction monitoring technique based on a spectrophotometric method.

AlexaFluor® 488 dye conjugates provide a powerful and simple method employing UV absorbance. Based on the success of oxime bond formation with Tn- 37

ONH2 and aldehyde generation by oxidation of D-galactofuranose from PS A1 proved to be a very reliable technique. Therefore, we decided to couple the AlexaFluor®488 dye to

Scheme 1.2. The conjugation of AF488-hydrazide to PS A1.

PS A1 through hydrazone bond formation with oxidized PS A (Scheme 1.2). After purification through dialysis, the UV absorbance of the pure conjugates was recorded at

493 nm. Percent loading was determined using Beer’s law. Table 1.5 shows the degree of Tn loading onto PS A1. Our results indicate that when 1 equivalent of NaIO4 is used for the oxidation of PS A1, greater than 100% loading of AF488® occurs. This implies

Table 1.5. Percent loading of AF488 on PS A1 with varying equivalents of NaIO4.

a Equivalents of NaIO4 Percent Loading using AF488® 1 110d 0.75 42d 0.50 37d 0.25 12d aequivalents of NaIO4 = Moles PS A1 x 120. The number 120 is derived based on the number of repeating tetrameric oligosaccharide units composing PS A1. bNumber of repeating units coupled with AF488® as a percentage.

that over-oxidation occurs with other vicinal diols in the PS A1 repeating unit such as those of D-galactofuranose (the secondary 2-OH and 3-OH). When 0.50 equivalents of

NaIO4 is used to oxidize PS A1, 37% loading with AF488 hydrazide was found. Using less than 0.50 equivalents of NaIO4 as the oxidant gave relatively low percent loading values.

1.6.1. Conclusions

We have established that the structure of Tn-PS A1 has less α-helical content then

PS A1 as determined by CD, and the percent loading of Tn on PS A1 can be determined by using spectrophotometric methods. Our results suggest that it is not the α-helical nature of the polysaccharide, but the alternating charges that contribute to MHC class II binding. Our data also shows how the α-helical content is diminished under physiological conditions using pH and temperature. This leads us to believe the Tn-PS

A1 construct binds MHC II through electrostatics and that structural aspects do not play a critical role, contrary to what has been previously proposed. The successful addition of

35 molecules of Tn per oxidized unit of PS A1 reduces the α-helical content of PS A1, but leaves the alternating charge character on adjacent repeating units. However, immune activation is evident.

1.6.2 Experimental

1.6.2.1 Bacterial Growth and Isolation

B. fragilis (ATCC 25285/NCTC 9141) was purchased from Presque Isle Cultures.

To begin the initial growth procedure, the bacteria were streaked on blood agar- containing BBE plates. The plates were prepared in an anaerobic glove bag in a CO2 environment. After the cultures were initiated, the plates were transferred to an anaerobic 39

jar with gas packs in the presence of O2 indicator strips and placed in an incubator at 37

°C. Note: Freeze-dried samples were initially purchased directly from ATCC but after several growth attempts we deemed the samples to be non-virulent.

Figure 1.9. Overview of PS A1 isolation and purification.

PYG broth was used for the growth of B. fragilis. Proteose-peptone (20 g), yeast extract (5 g), NaCl (5 g), and 0.001 g of reazurin per 1 L of nanopure H2O were autoclaved. Glucose 25% (2 mL), potassium phosphate 25% (2 mL), cysteine 5% (1 mL), 0.5% of hemin in 1N NaOH (100 µL), and 0.5% vitamin K1 in absolute ethanol (50

µL) were filtered using a 0.22 µm filter, and added to the autoclaved PYG broth.

Anaerobic conditions were achieved by degassing solutions for 30 min under an atmosphere of 80% N2, 10% CO2, 10% H2. A resazurin indicator was used to assure an

anaerobic environment. The agar plates or liquid media were ready for inoculation as soon as the media was no longer pink in color. The agar plates were cut in sections and placed into the degassed media under an inert atmosphere. For liquid media transfer, 5 mL of culture was seeded in a degassed jar by cannulation. Every 24 hr the pH of the media was tested and adjusted to 7.2. During the first 24 h of growth, the pH would drop to 5, and 5M NaOH was used to adjust the pH in 1 mL portions until pH 7.2 was noted.

A total of 20 L of bacteria fermentation was accomplished.

1.6.2.2 Purification of PS A1

The growth media was centrifuged at 4,000 X g for 20 min at 4 °C in 500 mL bottles. The supernatant was poured off and the cells were resuspended and washed in

500 mL of 0.15 M NaCl. Then 500 mL of 75 % phenol was stirred with the washed cells at 70 °C for 30 min. The phenol layer was separated by centrifuging at 5,000 X g for 30 min at 4 °C. The aqueous layer was then extracted three times with ether. After extraction the aqueous layer was concentrated under reduced pressure at 60 °C, and redissovled in a minimal amount of water and subjected to dialysis for 7 days and lyophilized. The crude material was then subjected to 5.0 mg/mL of RNase (Promega) and 5.0 mg/mL DNase (Promega) in 0.1 M acetate buffer followed by 10 mg/mL of

Protease (Sigma-Aldrich) to degrade any RNA, DNA, and protein. The material was then purified on two size exclusion columns and an anion exchange column. The first size exclusion column was packed with Sephacryl S-400 (GE Lifesciences) using 0.5 % sodium deoxycholate, 50 mM glycine, and 10 mM EDTA (pH 9.8). Crude bacterial lysate was then loaded onto the column and 2 mL fractions were collected and analyzed by UV absorbance measuring at 220, 260, and 280, and TLC charring with anisaldehyde. 41

Fractions containing more than 0.1 ABS at 260 and 280 nm were pooled for further purification. The fractions that showed absorbance at 220 nm were pooled and dialyzed.

The polysaccharide obtained was further purified using Sephacryl S-300 (GE

Lifesciences) to remove excess buffer and further separate lipid capsular polysaccharides.

UV absorbance and TLC charring again analysed fractions. Finally, the last step in the purification was use of anion exchange chromatography. The crude PS A1 was treated with 5 % acetic acid for 1 h at 100 °C, loaded onto the column, and eluted with 50 mM

Tris-HCl, pH 7.3 and an increasing NaCl concentration from 0 M to 2 M. Nuclear magnetic resonance (NMR) was used to determine purity, and gel electrophoresis was used to determine size and was stained with a carbohydrate staining kit.

The AlexaFluor488® labelled PS A1 was purified by size exclusion column

(Sephacryl S-300 HR). Beer’s law (equation 1) was used to calculate percent loading using the UV wavelength of AlexaFluor488® dye.

A = Ecl (eq.1) A is UV absorbance at wave length 495 nm, E is the molar absorptivity

-1 -1, with units of L mol cm ; for AlexaFluor488® dye at 495 nm 71,000, l = part length = 1 cm, c = concentration in mol L-1

From equation 1, the concentration of AlexaFluor488® attached to PS A1 (CAF488)

-1 -1 = A495/(71,000 L mol cm *1 cm). By measuring absorbance of the AlexaFluor488® labelled PS A1 at 495 nm, the concentration was determined and the number of moles of the dye attached to PS A1 (PS A1 does not absorb at wave length 495 nm) was then calculated. The number of moles for PS A1 added for the reaction is known, therefore the molar ratio between AlexaFluor488® dye and PS A1 can be calculated (eq. 2).Molar ratio = AlexaFluor488® dye moles/ AlexaFluor488®-PS A1 dye moles (eq. 2). Under the 42

assumption that PS A1 has 120 repeating units, the percent loading is given by the following equation (equation 3): Percent Loading = AlexaFluor488® dye moles x 100/(

AlexaFluor488®-PS A1 dye moles x 120) (eq. 3).

1.6.2.3. Circular Dichroism

All CD measurements were carried out on a Chirascan CD Spectrometer (Applied

Photophysics, Kingston Rd, Leatherhead, UK) equipped with a temperature controlled cuvette holder. All samples were analysed in PBS. The concentration of all the samples were 2.5 mg/mL and a quartz cuvette with 0.2 cm path length was used to collect the data.

1.7 References

(1) Riedel, S. Proc. (Bayl Univ Med Cent) 2005, 18, 21.

(2) Pasteur, M. Br. Med. J. 1881, 2, 283.

(3) Dochez, A. R.; Avery, O. T. J. Exp. Med. 1917, 26, 477.

(4) Heidelberger, M.; Goebel, W. F.; Avery, O. T. J. Exp. Med. 1925, 42, 727.

(5) Tillett, W. S.; Francis, T. J. Exp. Med. 1930, 52, 561.

(6) Goebel, W. F.; Avery, O. T. J. Exp. Med. 1931, 54, 431.

(7) Avery, O. T.; Goebel, W. F. J. Exp. Med. 1931, 54, 437.

(8) Weintraub, A. Carbohydr. Res. 2003, 338, 2539.

(9) Finn, A. Br. Med. Bull 2004, 70, 1.

(10) Ichim, C. V. J. Transl. Med. 2005.

(11) Coley, W. B. Am. J. Med. Sci. 1893.

(12) Dunn, G. P.; Old, L. J.; Schreiber, R. D. Immunity 2004, 21, 137.

(13) Tashiro, T.; Sekine-Kondo, E.; Shigeura, T.; Nakagawa, R.; Inoue, S.;

Omori-Miyake, M.; Chiba, T.; Hongo, N.; Fujii, S.-I.; Shimizu, K.;

Yoshiga, Y.; Sumida, T.; Mori, K.; Watarai, H.; Taniguchi, M. Int.

Immunol. 2010, 22, 319.

(14) Jandus, C.; Boligan, K. F.; Chijioke, O.; Liu, H.; Dahlhaus, M.;

Démoulins, T.; Schneider, C.; Wehrli, M.; Hunger, R. E.; Baerlocher, G.

M.; Simon, H.-U.; Romero, P.; Münz, C.; von Gunten, S. J. Clin. Invest.

2014, 124, 1810.

(15) Akira, S.; Uematsu, S.; Takeuchi, O. Cell 2006, 124, 783.

(16) Schaefer, T. M.; Desouza, K.; Fahey, J. V.; Beagley, K. W. Immunology

2004.

(17) Raetz, C. R. H.; Reynolds, C. M.; Trent, M. S.; Bishop, R. E. Annu. Rev.

Biochem. 2007, 76, 295.

(18) Zhou, Z.; Liao, G.; Mandal, S. S.; Suryawanshi, S.; Guo, Z. Chem. Sci.

2015, 6, 7112.

(19) Geijtenbeek, T. B. H.; Gringhuis, S. I. Nat. Rev. Immunol. 2009, 9, 465.

(20) Katz, A. K.; Glusker, J. P.; Beebe, S. A. J. Am. Chem. Soc. 1996, 118,

5752.

(21) Richardson, M. B.; Williams, S. J. Front. Immunol. 2014, 5, 288.

(22) Jégouzo, S. A. F.; Quintero-Martínez, A.; Ouyang, X.; dos Santos, Á.;

Taylor, M. E.; Drickamer, K. Glycobiology 2013, 23, 853.

(23) Agrawal, S.; Gupta, S.; Agrawal, A. PLoS ONE 2010, 5, e13418.

(24) Borghesi, L.; Milcarek, C. Cancer Res. 2007, 67, 3989.

(25) Litman, G. W.; Rast, J. P.; Fugmann, S. D. Nat. Rev. Immunol. 2010, 10,

543.

(26) Topham, N. J.; Hewitt, E. W. Immunology 2009, 128, 7.

(27) Dunkelberger, J. R.; Song, W.-C. Cell Res. 2010, 20, 34.

(28) Mond, J. J.; Lees, A.; Snapper, C. M. Annu. Rev. Immunol. 1995, 13, 655.

(29) Vos, Q.; Lees, A.; Wu, Z. Q.; Snapper, C. M.; Mond, J. J. Immunol. Rev.

2000, 176, 154.

(30) Obukhanych, T. V.; Nussenzweig, M. C. J. Exp. Med. 2006, 203, 305.

(31) Tzianabos, A.; Wang, J. Y.; Kasper, D. L. Carbohydr. Res. 2003, 338,

2531.

(32) Baumann, H.; Tzianabos, A. O.; Brisson, J. R.; Kasper, D. L.; Jennings, H.

J. Biochemistry 1992, 31, 4081.

(33) Kreisman, L. S. C.; Friedman, J. H.; Neaga, A.; Cobb, B. A. Glycobiology

2007, 17, 46.

(34) Cobb, B. A.; Wang, Q.; Tzianabos, A. O.; Kasper, D. L. Cell 2004, 117,

677.

(35) Chung, D. R.; Kasper, D. L.; Panzo, R. J.; Chitnis, T.; Grusby, M. J.;

Sayegh, M. H.; Tzianabos, A. O.; Chtinis, T. J. Immunol. 2003, 170, 1958.

(36) Bloem, K.; García-Vallejo, J. J.; Vuist, I. M.; Cobb, B. A.; van Vliet, S. J.;

van Kooyk, Y. Front Immunol 2013, 4, 1.

(37) Cobb, B. A.; Kasper, D. L. Glycobiology 2008, 18, 707.

(38) Cobb, B. A.; Kasper, D. L. Glycobiology 2008, 18, 707.

(39) Duan, J.; Avci, F. Y.; Kasper, D. L. Proc. Natl. Acad. Sci. U.S.A. 2008,

105, 5183.

(40) Tzianabos, A. O.; Onderdonk, A. B.; Rosner, B.; Cisneros, R. L.; Kasper,

D. L. Science 1993, 262, 416.

(41) Tzianabos, A. O.; Onderdonk, A. B.; Smith, R. S.; Kasper, D. L. Infect.

Immun. 1994, 62, 3590. 46

(42) Groneck, L.; Schrama, D.; Fabri, M.; Stephen, T. L.; Harms, F.;

Meemboor, S.; Hafke, H.; Bessler, M.; Becker, J. C.; Kalka-Moll, W. M.

Infect. Immun. 2009, 77, 3705.

(43) Tzianabos, A. O.; Wang, J. Y.; Lee, J. C. Proc. Natl. Acad. Sci. USA 2001,

98, 9365.

(44) Watts, A.; Ke, D.; Wang, Q.; Pillay, A.; Nicholson-Weller, A.; Lee, J. C.

Infect. Immun. 2005, 73, 3502.

(45) Young, N. M.; Kreisman, L. S. C.; Stupak, J.; MacLean, L. L.; Cobb, B.

A.; Richards, J. C. Glycobiology 2011, 21, 1266.

(46) Avci, F. Y.; Li, X.; Tsuji, M.; Kasper, D. L. Nat. Protoc. 2012, 7, 2180.

(47) Frosch, M.; Görgen, I.; Boulnois, G. J.; Timmis, K. N.; Bitter-Suermann,

D. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 1194.

(48) Korn, T.; Bettelli, E.; Oukka, M.; Kuchroo, V. K. Annu. Rev. Immunol.

2009, 27, 485.

(49) Liu, Z.; Fan, H.; Jiang, S. Immunol. Rev. 2013, 252, 183.

(50) Romagnani, S. Ann. Allergy Asthma Immunol. 2000, 85, 9.

(51) Mazmanian, S. K.; Round, J. L.; Kasper, D. L. Nature 2008, 453, 620.

(52) Gnerlich, J. L.; Mitchem, J. B.; Weir, J. S.; Sankpal, N. V.; Kashiwagi, H.;

Belt, B. A.; Porembka, M. R.; Herndon, J. M.; Eberlein, T. J.;

Goedegebuure, P.; Linehan, D. C. J. Immunol. 2010, 185, 4063.

(53) Kimura, A.; Kishimoto, T. Eur. J. Immunol. 2010, 40, 1830.

(54) Gaffen, S. L. Current Opinion in Immunology 2011, 23, 613.

(55) Yang, X. O.; Nurieva, R.; Martinez, G. J.; Kang, H. S.; Chung, Y.; Pappu,

B. P.; Shah, B.; Chang, S. H.; Schluns, K. S.; Watowich, S. S.; Feng, X.-

H.; Jetten, A. M.; Dong, C. Immunity 2008, 29, 44.

(56) Mitsdoerffer, M.; Lee, Y.; Jäger, A.; Kim, H.-J.; Korn, T.; Kolls, J. K.;

Cantor, H.; Bettelli, E.; Kuchroo, V. K. Proc. Natl. Acad. Sci. U.S.A.

2010, 107, 14292.

(57) Murugaiyan, G.; Saha, B. J. Immunol. 2009, 183, 4169.

(58) Muranski, P.; Boni, A.; Antony, P. A.; Cassard, L.; Irvine, K. R.; Kaiser,

A.; Paulos, C. M.; Palmer, D. C.; Touloukian, C. E.; Ptak, K.; Gattinoni,

L.; Wrzesinski, C.; Hinrichs, C. S.; Kerstann, K. W.; Feigenbaum, L.;

Chan, C. C.; Restifo, N. P. In Blood; American Society of Hematology:

2008; Vol. 112, p 362.

(59) Daniotti, J. L.; Vilcaes, A. A.; Torres Demichelis, V.; Ruggiero, F. M.;

Rodriguez-Walker, M. Front. Oncol. 2013, 3, 306.

(60) Cazet, A.; Julien, S.; Bobowski, M.; Burchell, J.; Delannoy, P. Breast

Cancer Res. 2010, 12, 204.

(61) van Liempt, E.; Bank, C. M. C.; Mehta, P.; García-Vallejo, J. J.; Kawar, Z.

S.; Geyer, R.; Alvarez, R. A.; Cummings, R. D.; Kooyk, Y. v.; van Die, I.

FEBS Lett. 2006, 580, 6123.

(62) Ju, T.; Wang, Y.; Aryal, R. P.; Lehoux, S. D.; Ding, X.; Kudelka, M. R.;

Cutler, C.; Zeng, J.; Wang, J.; Sun, X.; Heimburg-Molinaro, J.; Smith, D.

F.; Cummings, R. D. Proteomics Clin. Appl. 2013, 7, 618.

(63) Ju, T.; Cummings, R. D. Proc. Natl. Acad. Sci. USA 2002, 99, 16613. 48

(64) Zizzari, I. G.; Napoletano, C.; Battisti, F.; Rahimi, H.; Caponnetto, S.;

Pierelli, L.; Nuti, M.; Rughetti, A. J. Immunol. Res. 2015, 2015, 450695.

(65) Napoletano, C.; Rughetti, A.; Agervig Tarp, M. P.; Coleman, J.; Bennett,

E. P.; Picco, G.; Sale, P.; Denda-Nagai, K.; Irimura, T.; Mandel, U.;

Clausen, H.; Frati, L.; Taylor-Papadimitriou, J.; Burchell, J.; Nuti, M.

Cancer Res. 2007, 67, 8358.

(66) Nonaka, M.; Ma, B. Y.; Murai, R.; Nakamura, N.; Baba, M.; Kawasaki,

N.; Hodohara, K.; Asano, S.; Kawasaki, T. J. Immunol. 2008, 180, 3347.

(67) Chen, H.; Yuan, B.; Zheng, Z.; Liu, Z.; Wang, S. Cell. Immunol. 2011,

269, 144.

(68) García-Vallejo, J. J.; van Liempt, E.; da Costa Martins, P.; Beckers, C.;

van het Hof, B.; Gringhuis, S. I.; Zwaginga, J.-J.; van Dijk, W.;

Geijtenbeek, T. B. H.; van Kooyk, Y.; van Die, I. Mol. Immunol. 2008, 45,

2359.

(69) de Jong, M. A. W. P.; Vriend, L. E. M.; Theelen, B.; Taylor, M. E.;

Fluitsma, D.; Boekhout, T.; Geijtenbeek, T. B. H. Mol. Immunol. 2010,

47, 1216.

(70) Rapoport, E.; Mikhalyov, I.; Zhang, J.; Crocker, P.; Bovin, N. Bioorg.

Med. Chem. Lett. 2003, 13, 675.

(71) Hsu, C.-C.; Lin, T.-W.; Chang, W.-W.; Wu, C.-Y.; Lo, W.-H.; Wang, P.-

H.; Tsai, Y.-C. Gynecol. Oncol. 2005, 96, 415.

(72) Collin, M.; Ehlers, M. Exp. Dermatol. 2013, 22, 511.

(73) Hanna, M. G.; Peters, L. C. Cancer Res. 1978, 38, 204. 49

(74) Srivatsan, S.; Patel, J. M.; Bozeman, E. N.; Imasuen, I. E.; He, S.; Daniels,

D.; Selvaraj, P. Hum. Vaccin. Immunother. 2014, 10, 52.

(75) Zeichner, S. B. J. Am. Osteopath Assoc. 2012, 112, 482.

(76) Jocham, D.; Richter, A.; Hoffmann, L.; Iwig, K. Lancet 2004.

(77) Slovin, S. F.; Ragupathi, G.; Musselli, C.; Olkiewicz, K.; Verbel, D.;

Kuduk, S. D.; Schwarz, J. B.; Sames, D.; Danishefsky, S.; Livingston, P.

O.; Scher, H. I. J. Clin. Oncol. 2003, 21, 4292.

(78) De Silva, R. A.; Wang, Q.; Chidley, T.; Appulage, D. K.; Andreana, P. R.

J. Am. Chem. Soc. 2009, 131, 9622.

(79) Adluri, S.; Helling, F.; Ogata, S.; Zhang, S.; Itzkowitz, S. H.; Lloyd, K.

O.; Livingston, P. O. Cancer Immunol. Immunother. 1995, 41, 185.

(80) Miles, D. W.; Towlson, K. E.; Graham, R.; Reddish, M.; Longenecker, B.

M.; Taylor-Papadimitriou, J.; Rubens, R. D. Br. J. Cancer

1996, 74, 1292.

(81) Gilewski, T. A.; Ragupathi, G.; Dickler, M.; Powell, S.; Bhuta, S.;

Panageas, K.; Koganty, R. R.; Chin-Eng, J.; Hudis, C.; Norton, L.;

Houghton, A. N.; Livingston, P. O. In Clinical Cancer Research 2007;

Vol. 13, p 2977.

(82) Sabbatini, P. J.; Kudryashov, V.; Ragupathi, G.; Danishefsky, S. J.;

Livingston, P. O.; Bornmann, W.; Spassova, M.; Zatorski, A.; Spriggs, D.;

Aghajanian, C.; Soignet, S.; Peyton, M.;Flaherty, C.; Curtin, J.; Lloyd, K.

O. Int. J. Cancer 2000, 87, 79.

(83) Buskas, T.; Li, Y.; Boons, G.-J. Chem. Eur. J. 2005, 11, 5457. 50

(84) Slovin, S. F.; Ragupathi, G.; Adluri, S.; Ungers, G.; Terry, K.; Kim, S.;

Spassova, M.; Bornmann, W. G.; Fazzari, M.; Dantis, L.; Olkiewicz, K.;

Lloyd, K. O.; Livingston, P. O.; Danishefsky, S. J.; Scher, H. I. Proc. Natl.

Acad. Sci. USA 1999, 96, 5710.

(85) Becker, R.; Eichler, M. K.; Jennemann, R.; Bertalanffy, H. In Br J

Neurosurg 2002; Vol. 16, p 269.

(86) Ragupathi, G.; Livingston, P. O.; Hood, C.; Gathuru, J.; Krown, S. E.;

Chapman, P. B.; Wolchok, J. D.; Williams, L. J.; Oldfield, R. C.; Hwu,

W.-J. Clin. Cancer. Res. 2003, 9, 5214.

(87) Ragupathi, G.; Meyers, M.; Adluri, S.; Howard, L.; Musselli, C.;

Livingston, P. O. Int. J. Cancer 2000, 85, 659.

(88) Osorio, M.; Gracia, E.; Reigosa, E.; Hernandez, J.; de la Torre, A.; Saurez,

G.; Perez, K.; Viada, C.; Cepeda, M.; Carr, A.; Avila, Y.; Rodríguez, M.;

Fernández, L. E. Cancer Manag. Res. 2012, 4, 341.

(89) Ragupathi, G.; Koide, F.; Sathyan, N.; Kagan, E.; Spassova, M.;

Bornmann, W.; Gregor, P.; Reis, C. A.; Clausen, H.; Danishefsky, S. J.;

Livingston, P. O. Cancer. Immunol. Immunother. 2003, 52, 608.

(90) Ragupathi, G.; Cappello, S.; Yi, S. S.; Canter, D.; Spassova, M.;

Bornmann, W. G.; Danishefsky, S. J.; Livingston, P. O. Vaccine 2002, 20,

1030.

(91) O'Cearbhaill, R. E.; Ragupathi, G.; Zhu, J.; Wan, Q.; Mironov, S.; Yang,

G.; Spassova, M. K.; Iasonos, A.; Kravetz, S.; Ouerfelli, O.; Spriggs, D.

R.; Danishefsky, S. J.; Sabbatini, P. J. Cancers (Basel) 2016, 8. 51

(92) Zhou, Z.; Mondal, M.; Liao, G.; Guo, Z. Org. Biomol. Chem. 2014, 12,

3238.

(93) Sarkar, S.; Lombardo, S. A.; Herner, D. N.; Talan, R. S.; Wall, K. A.;

Sucheck, S. J. J. Am. Chem. Soc. 2010, 132, 17236.

(94) Toyokuni, T.; Dean, B.; Cai, S.; Boivin, D.; Hakomori, S.; Singhal, A. K.

J. Am. Chem. Soc. 1994, 116, 395.

(95) Kudryashov, V.; Glunz, P. W.; Williams, L. J.; Hintermann, S.;

Danishefsky, S. J.; Lloyd, K. O. Proc. Natl. Acad. Sci. USA 2001, 98,

3264.

(96) Sarkar, S.; Salyer, A. C. D.; Wall, K. A.; Sucheck, S. J., Bioconjugate

Chem. 2013, 24 (3), 363-375.

(97) Coffman, R. L.; Sher, A.; Seder, R. A. Immunity 2010, 33, 492.

(98) McKee, A. S.; Munks, M. W.; MacLeod, M. K. L.; Fleenor, C. J.; Van

Rooijen, N.; Kappler, J. W.; Marrack, P. J. Immunol. 2009, 183, 4403.

(99) Lu, F.; HogenEsch, H. Vaccine 2013, 31, 3979.

(100) Ghimire, T. R.; Benson, R. A.; Garside, P.; Brewer, J. M. Immunol. Lett.

2012, 147, 55.

(101) Brewer, J. M.; Conacher, M.; Hunter, C. A.; Mohrs, M.; Brombacher, F.;

Alexander, J. J. Immunol. 1999, 163, 6448.

(102) Fernández-Tejada, A.; Chea, E. K.; George, C.; Pillarsetty, N.; Gardner, J.

R.; Livingston, P. O.; Ragupathi, G.; Lewis, J. S.; Tan, D. S.; Gin, D. Y.

Nature Chem. 2014, 6, 635.

(103) Huang, Y.-L.; Hung, J.-T.; Cheung, S. K. C.; Lee, H.-Y.; Chu, K.-C.; Li,

S.-T.; Lin, Y.-C.; Ren, C.-T.; Cheng, T.-J. R.; Hsu, T.-L.; Yu, A. L.; Wu,

C.-Y.; Wong, C.-H. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 2517.

(104) Peeters, C. C.; Tenbergen-Meekes, A. M.; Poolman, J. T.; Beurret, M.;

Zegers, B. J.; Rijkers, G. T. Infect. Immun. 1991, 59, 3504.

(105) Ragupathi, G.; Koide, F.; Livingston, P. O.; Cho, Y. S.; Endo, A.; Wan,

Q.; Spassova, M. K.; Keding, S. J.; Allen, J.; Ouerfelli, O.; Wilson, R. M.;

Danishefsky, S. J. J. Am. Chem. Soc. 2006, 128, 2715.

(106) Sabbatini, P. J.; Ragupathi, G.; Hood, C.; Aghajanian, C. A.; Juretzka, M.;

Iasonos, A.; Hensley, M. L.; Spassova, M. K.; Ouerfelli, O.; Spriggs, D.

R.; Tew, W. P.; Konner, J.; Clausen, H.; Abu Rustum, N.; Dansihefsky, S.

J.; Livingston, P. O. Clin. Cancer. Res. 2007, 13, 4170.

(107) Lo-Man, R.; Bay, S.; Vichier-Guerre, S.; Deriaud, E.; Cantacuzène, D.;

Leclerc, C. Cancer Res. 1999, 59, 1520.

(108) Grigalevicius, S.; Chierici, S.; Renaudet, O.; Lo-Man, R.; Dériaud, E.;

Leclerc, C.; Dumy, P. In Bioconjugate Chem. 2005; Vol. 16, p 1149.

(109) Wang, J.; Zhang, Y.; Wei, J.; Zhang, X.; Zhang, B.; Zhu, Z.; Zou, W.;

Wang, Y.; Mou, Z.; Ni, B.; Wu, Y. Immunology 2007, 121, 174.

(110) He, L.-Z.; Weidlick, J.; Sisson, C.; Marsh, H. C.; Keler, T. In Cell. Mol.

Immunol. 2015; Vol. 12, p 719.

(111) De Silva, R. A.; Appulage, D. K.; Pietraszkiewicz, H.; Bobbitt, K. R.;

Media, J.; Shaw, J.; Valeriote, F. A.; Andreana, P. R. Cancer Immunol.

Immunother. 2012, 61, 581. 53

(112) Kalia, J.; Raines, R. T. Angew. Chem. Int. Ed. Engl. 2008, 47, 7523.

(113) Martin-Orozco, N.; Muranski, P.; Chung, Y.; Yang, X. O.; Yamazaki, T.;

Lu, S.; Hwu, P.; Restifo, N. P.; Overwijk, W. W.; Dong, C. Immunity

2009, 31, 787.

(114) Buffington, L. A.; Pysh, E. S.; Chakrabarti, B.; Balazs, E. A. J Am Chem

Soc 1977, 99, 1730.

(115) Listowsky, I.; Avigad, G.; Englard, S. The Journal of organic chemistry

1970, 35, 1080.

(116) Southwick, J. G.; McDonnell, M. E.; Jamieson, A. M.; Blackwell, J.

Macromolecules 1979, 12, 305.

(117) Stephen, T. L.; Niemeyer, M.; Tzianabos, A. O.; Kroenke, M.; Kasper, D.

L.; Kalka-Moll, W. M. Infect. Immun. 2005, 73, 2184.

(118) Erlanger, B. F. Methods in Enzymology 1980, 70, 85.

Chapter 2

Immunological evaluation of the entirely carbohydrate- based Thomsen-Friedenreich PS B conjugate.

[Trabbic, K. R.; Bourgault, J. P.; Shi, M.; Clark, M.; Andreana, P. R. Org. Biomol. Chem.

2016, 14 (13), 3350.] - Reproduced by permission of The Royal Society of Chemistry

2.1. Introduction

Economic projections forecast the global market for cancer immunotherapies will increase by 20% annually and expand to 7.1 billion dollars by 2018.2 This is directly related to the Food and Drug Administration (FDA) and the European Commission (EC) approval of personalized DC-based prostate cancer vaccine sipuleucel-T, commonly known as Provenge.3 Provenge® mechanism of action involves in vitro priming of autologous dendritic cells with a chimeric protein containing a cancer antigen, prostatic acid phosphatase (PAP) fused to cytokine GM-CSF. The activated DC is reintroduced to the patient to stimulate an immune response against prostate cancer. Major disadvantages of this immunotherapy are cost ($90,000) and a marginal increase in life expectancy (4-6 months). Remarkably, there are no protein conjugate vaccines on the market designed for the treatment of cancer; despite the numerous commercially available protein conjugate bacterial vaccines including PedvaxHIB®, HIBERIX®, 55

Menjugate®, and Prevenar 13®.4 There have also been clinical trials using carrier protein conjugates with surface tumor associated carbohydrate antigens (TACAs) (e.g.

GM2, TF, STn (Theratope®)), which have fallen short of any measurable therapeutic benefit.5-10 As important cell surface tumor markers, targeting TACAs requires innovative strategies that utilize immunogens in creative ways that allow for facile synthetic accessibility with a concomitant increased therapeutic value.11

TACAs are a consequence of mutated chaperone proteins that cause aberrant glycosylations on the surface of tumor cells.12, 13 These surface antigens are considered to be a focus of prophylactic and therapeutic vaccination efforts arising from their synthetic accessibility.14 It is well documented that TACAs themselves do not elicit a robust anti-carbohydrate immune response which is a direct result of the T-cell independent nature of sugars and our immune preference towards peptide/protein/amino acids.15 In order to overcome the weakly immunogenic nature of carbohydrate antigens,

TACAs have been conjugated to highly immunogenic protein carriers such as keyhole limpet hemocyanin (KLH) and tetanus toxoid (TT) amongst others including recent bioconjugate studies with CRM 197.16-19 However, the inefficiency of these glycoprotein conjugate immunizations can be attributed to several deficiencies associated with the protein carriers themselves such as a) carbohydrate epitope suppression,7 b) protein aggregation/stability,5 and c) large immunogenic linkers that induce unwanted immune responses on their own.20, 21 One approach to circumvent the complexities of protein carriers would be in accessing alternative “carriers” such as entirely carbohydrate constructs that we are currently studying to negate the aforementioned issues.22-24 The discovery and mechanistic insights of zwitterionic polysaccharides (ZPSs), especially PS 56

A1, illustrates a validation of how polysaccharides can invoke both an innate and adaptive immune response mimicking essential characteristics of heavily employed carrier proteins.6, 25

ZPS vaccine development/immunotherapy aims to increase the immunogenicity and specificity to carbohydrate antigens exclusively (all donor and acceptor interactions are carbohydrate-based). The design of ZPS-conjugates takes advantage of antigen uptake via C-type lectin receptors (CLRs) and interaction with toll-like-receptor (TLRs) which leads to antigen presentation to CD4+ T-cells.26 Interactions with TLR2, have produced an adjuvant effect by increasing the immune response towards a chemically derived ZPS-protein conjugate compared to the native polysaccharide analogue.27 ZPS- conjugates are distinct from the more general concept of bacterial-based vaccines which, in some instances, use isolated capsular polysaccharides from bacteria. These bacterial- based vaccines are thought to generate a T-cell independent response via cross-linking with B-cell receptors and invoke short-lived IgM immune responses.15 In the clinical setting there are several licensed non-protein conjugated bacterial-based vaccines targeting Streptococcus pneumoniae (Pneomovax 23)28, 29, Neisseria meningitidis

(Menomune)30 and Salmonella typhi (Typhim Vi)31, 32. Herein, the focus will center on investigating ZPS PS B as an antigen “carrier” for the TACA Thomsen-Friedenreich (TF) disaccharide.

Both ZPSs PS A1 (1) and PS B (2) isolated from B. fragilis (Figure 2.1) have been shown to be potential alternatives to traditional immunogenic carrier proteins in immunization studies.22, 23, 33 Initial investigations of ZPS revealed adoptively transferred splenic T-cells from ZPS immunizations demonstrated protection in other animals against 57

intra-abdominal abscesses from a challenge with B. fragilis or other pathogenic bacteria, which signifies definitive T-cell activation.34-37,38 The importance of ZPSs has been further illustrated by their ability to induce an immune response involving lymphocytes, neutrophils, and macrophages to the intraperitoneal cavity.39 This effect is facilitated through increased expression of intercellular adhesion molecule 1 (ICAM-1) to attract polymorphonuclear leukocytes (PMNLs) mediated by ZPS.39 PS B, itself, is a high molecular weight ZPS with repeating sugars: β-D-QuiNac (1→4), α-D→Gal (1-4), α-L-

QuiNAc (1→3), and branched from 3’-galactose is β-D-GlcNAc (1→3), α-D-

GalA(1→3), and α-L-Fucp(1→2). In addition, a bacterial exclusive 2-aminoethyl

Figure 2.1. Structures of ZPS PS A1 (1) and PS B (2) from B. fragilis. phosphonate moiety is substituted on the 4’ position of β-D-GlcNAc.

In our previous studies, we reported on a novel Tn-PS A1 construct that was able to elicit specific IgG3 antibodies to Tn and bind to Tn-containing tumor cell lines.22, 23

IgG3 antibodies are carbohydrate specific antibodies40 that result from a T cell mediated antibody isotype switching event41 and is an effective inducer of complement dependent cytotoxicity.42 This semi-synthetic approach was utilized to take advantage of

Scheme 2.1. Semi-synthetic TF-PS B (4) immunogen. regio-specific PS A1 oxidation and site-selective conjugation using the synthetic anomeric aminooxy GalNAc (Tn = TACA) forming a highly stable oxime link23 to avoid all non-natural hydrocarbon linkers which have been shown to induce unwanted immune responses.20 Subsequent to this work, we reported on a similar approach for aminooxy

Thompsen-Friedenreich (TF or T) antigen.33 Additionally, using circular dichroism in our studies confirmed the importance of a structure-activity relationship, which noted the importance of an alternating charge character on adjacent monosaccharides as being essential for biological activity.24 Based on the success of our TACA-PS A1 conjugates, our focus has also turned to utilizing other zwitterionic polysaccharides that can elicit specific immune responses to TACAs in the absence of carrier proteins and/or peptide epitopes. To this end, the TF antigen (Galβ1→3GalNAcα1→Ser/Thr) represents an

excellent TACA target due to its high level of expression in breast and prostate tumors as well as playing a significant role in cancer metastasis. TF is thought to be involved in cancer metastasis through galectin-3 which is a galactoside binding lectin.43 Therefore, synthesizing a construct that could specifically focus on the TF disaccharide

(Galβ1→3GalNAcα1) portion and would also allow for a novel immunogen which could mount a directed immune response was paramount. The concepts of our specific approach were designed to examine carbohydrate selective immune processing with regards to antibody production without the “help” of proteins/peptides or lipids and to avoid heterogeneous mixtures of carbohydrate-peptide, carbohydrate-protein and carbohydrate-lipid combinations.

2.2. Results and Discussion

In order to accomplish our aims, we first elected to oxidize PS B (2) using the

Malaprade reaction44 with 10 mM sodium periodate in NaOAc buffer (pH 5) to reveal aldehydes that would specifically react with aminooxy-TF antigen (3) for conjugation to produce TF-PS B (4) (Scheme 2.1).33 Our strategy to utilize an oxime bond was chosen for purposes that include stability and efficiency.45 Other orthogonal conjugation chemistries, such as click chemistry forming triazoles were noted, however, in many documented cases recognition of the immunological antigen was impaired and therefore was avoided in our approach.46, 47

Once the conjugation was complete, as noted by 1H NMR following the formation of oxime characterization, we set out to determine the percent loading of the TF antigen on PS B. Determining percent loading without the capability of mass spectroscopic techniques (polysaccharides do not ionize well) can be challenging, however, to 60

overcome this limitation we employed two indirect methods known for quantitative analysis pertaining to percent loading. The first method we employed was a periodate- rescorinol sialic acid assay using STn-PS B and the second method was an Alexa Fluor®

488-hydrazide fluorophore conjugation protocol.48 In the first method, the periodate- rescorinol strategy was preferred over a phenol-sulfuric acid method as the latter is non- specific towards carbohydrates and PS B posed some interference in the development of a calibration curve. Sialic acid based conjugates are optimal in this scenario and are preferred over the TF-antigen (3) due to the vicinal diols of sialic acid requiring low

Sialic Acid Periodate-Rescorinol Assay

1.0 Sialic Acid

STn-PSB

M n

PS B

0 8

5 GalNac

Galactose Amine

a V

0.5

D O

Interpolated value of sialic acid for PS B = 2.586 mg/ 50 mg of Construct

0.0 0 10 20 30 40 50

Sugar (mg)

Figure 2.2. Sialic acid determination using periodate-rescorinol assay. sodium periodate concentrations for oxidation (1mM), where (2 and 3) require higher concentrations (10mM) of sodium periodate, potentially leading to undesired fluorophore 61

generation. In short, a standard curve of sialic acid was used to interpolate the concentration of sialic acid on STn-ONH2 conjugated to PS B from the in situ fluorophore generated using the periodate-rescorinol method between aldehydes and rescorinol which gave ~10% loading (Figure 2.2). PS B, GalNAc, and Galactose Amine was used as controls to demonstrate the selectivity of this assay towards sialic acid.

Whereas, the Alexa Fluor® 488-hydrazide labeling method gave ~6% by conjugating the reactive fluorophore via a hydrazone linkage directly to oxidized PS B to be quantified.

The difference in loading levels can be attributed to steric hindrance affiliated with the bulky Alexa Fluor® 488-hydrazide binding in close proximity to available oxidized PS B carbonyl aldehydes and electrostatic repulsion from the nature of Alex Fluor® 488 itself.

In order to begin determining the effectiveness of our novel TF-PS B construct

(4), Jax C57BL/6J mice were immunized, blood sera were collected and anti-TF immune responses examined. Three different immunogens were administered to the mice: 1) PS

B (2), 2) TF-PS B (4), and 3) TF-BSA (5) (Scheme 2.2) with and without TiterMax®

Gold adjuvant to determine, amongst a host of assays, antibody binding and specificity towards the TF antigen. To identify the specificity and selectively of an antibody immune response, the ELISA (Enzyme Linked Immunosorbent Assay) (Figure 2.3) represents the first way to detect and quantify an antibody response. As an essential assay, data obtained from these assays can provide in vitro insights into the specificity and selectivity to the TF hapten generated from vaccine immunizations. The general procedure for ELISA begins coating a 96-well plate with TACA-protein conjugates to measure antibody binding to a hapten (TF without PS B). The primary anti-serum from immunizations is serially diluted and incubated on the well plate containing TACA-BSA 62

conjugates. The plate is washed to remove primary antibodies and enzyme-linked secondary antibodies are incubated with the purpose to detect bound primary antibodies.

The plates are again washed to remove unbound secondary antibodies and a substrate such as 4-nitro phenylphosphate is added to the enzyme (alkaline phosphatase) linked secondary antibody to cleave the phosphate to produce ρ-nitrophenol chromophore which can be monitored at 405 nm.

Figure 2.3. General description of ELISA.

The ELISA results, as noted in (Table 2.1; entry A), indicate mice immunized with PS B alone produced titers with the respective isotypes of IgM, IgG1 and IgG2b which is indicative of a Th1-type immune response.49 We determined this convention using an ELISA plate coating construct of PS B-poly-L-lys (PSB-PLL). Interestingly, when PS B was administered with TiterMax® Gold (TMG) adjuvant, there was an observed decreased titer of Kappa antibodies but an increase in IgG1 titers (Table 2.1; entry B). This is likely attributed to the adjuvant emulsion of TiterMax® Gold permitting slow release of PS B while promoting antigen presenting cells to the site of delivery.50 Gratifyingly, TF-PS B (Table 2.1; entries C-F) constructs produced similar antibody isotype profiles in comparison to PS B with the caveat of specific antibodies

recognizing the TF antigen. Data for entries E-I in Table 2.1 were obtained using a TF-

BSA (5) ELISA coating construct to screen for selectivity.

To validate that antibodies generated from TF-PS B construct, with or without TMG,

Table 2.1. Evaluating PS B (2) and TF-PS B (4) constructs through immunizations in Jax C57BL/6J mice.

entry antigen plate coating Kappa[c] IgM[c] IgG[c] IgG1[c] IgG2b[c]

A PS B PS B-PLL[a] 39344 31252 7840 N.O.[b] 2273 PS B B PS B-PLL[a] 26250 28174 7609 190 9450 (TMG) TF-PS B TF-PS B- C 22336 12682 7659 64 2392 (TMG) PLL[a] TF-PS B- D TF-PS B 33613 23415 22686 N.O.[b] 645 PLL[a] TF-PS B E TF-BSA[a] 3782 2676 5431 372 93 (TMG) N.O.[b] F TF-PS B TF-BSA[a] 5250 683* 1608* N.O.[b]* * PS B N.O.[b] G TF-BSA[a] 319 18*# 2*# N.O.[b]* (TMG) * TF- [b] [a] # # N.O. [b] H ONH2 TF-BSA 641 62* 3* N.O. * * (TMG) N.O.[b] N.O.[b] N.O.[b] I PBS TF-BSA[a] 21 N.O.[b]* *# *# *

[a]Negative control: Anti-sera from PBS mice with negligible binding (O.D. ≤ 0.1) [b] N.O.: not observed (O.D. ≤ 0.12) [c] Experiments were performed in triplicate, [* or #] Statistically Significant (P < .005) compared with E* or F#

could recognize the TF antigen alone, sera from PS B TMG immunizations (Table 2.1;

entry G) was screened using a TF-BSA (Scheme 2.2) coating construct. We observed

significant IgG/IgM binding difference (P < 0.005) to TF-BSA with entries E and F

compared to control entries H and I. Therefore, the antibodies generated from TF-PS B construct (Table 2.1; entries E and F) were noted as being selective and specific in regards to the TF antigen (3).

Table 2.2. Reaction of TF-ONH2 with Maleic Anhydride (MA) coated ELISA plates to observe IgG immune response from TF-BSA and TF-PS B as a comparison. The plates were blocked with 2% casein to avoid reactivity with anti-BSA sera.

Since polyclonal antibodies from anti-TF-PS B (TMG) immunized mice (Table

2.1; entry E) recognize the TF-antigen, we felt it important to compare the efficacy of the anti-TF antibodies generated from a common protein construct in order to parse out the

PS B binding contribution of the TF-PS B immunogen. In order to answer this question,

Scheme 2.2. Conjugation of TF-linker (7) to BSA-maleimide. we prepared a TF-BSA conjugate (Scheme 2.2), where TF (3) was reacted with mercaptoaldehyde (6) to yield the TF linker (7). Zemplen conditions were used to 65

deacetylate the thioacetate to compound 8, which was used to react with BSA-maleimide to afford semi-synthetic TF-BSA. There was a loading of 34 molecules of TF per unit of

BSA determined by MALDI-TOF and was consequently immunized in Jax C57BL/6 mice.

In this case, when we specifically examined the anti-IgG isotypes from TF-BSA immunization, an observable larger titer response was generated towards TF-PS B (with

TiterMax® Gold) (Table 2.2) when a TF-maleimide coating was used (Pierce™ Maleic

Anhydride Activated Plates). ELISA with TF-maleic anhydride (TF-MA) coated plates were used to determine the titer of TF-PS B (4) and TF-BSA (5). TF-maleic anhydride plates were used because TF-ONH2 (3) could be conjugated to the maleic anhydride

(MA) coated plates without the need for protein conjugates or linkers, therefore allowing us to obtain true recognition of the TF-antigen. Another method for screening TF was used (TF-KLH) and it was constructed using similar conditions to TF-BSA but it was concluded that the maleimide linker augmented the binding data and subsequently the titers. Therefore, maleimide free ELISA coated plates were required to determine the specificity and selectivity of TF-BSA immunizations. From (Table 2.2), the TF-MA plates were used as a common platform to compare the titer data from TF-BSA and TF-

PS B. The data from TF-PS B was similar to what was seen in (Table 2.1; entry E) but anti-TF-BSA antibodies had minimal IgG recognition to TF. A possible explanation for this observation is TF-PS B may be able to act as a bridge between the innate and adaptive immune responses, producing specific anti-TF antibodies. Whereas TF-BSA contains a (4-maleimidmethyl)cyclohexane-1-carboxylate linker, which has been noted in literature20 to elicit strong immune responses against the linker and suppressing the 66

immune response against the carbohydrate based TF antigen.

Figure 2.4. General description of Fluorescence Activated Cell Sorting (FACS).

Based on our ELISA results, which concluded that a TF-PS B construct could elicit selective TF binding polyclonal antibodies generated in mice, we decided to examine the binding preference of those antibodies towards a human MCF-7 (breast) tumor cell line. It is known that MCF-7 cells express the TF antigen.51-53 To achieve this aim we elected to employ fluorescent binding technique and use flow cytometry (Figure

2.4) to determine binding efficiency. The principles of FACS (Fluorescent Activated Cell

Sorting) is similar to the ELISA technique but primary antibodies and secondary fluorescently labeled cells can be sorted based on fluorescent intensity. The TF-PS B

(with TMG) anti-serum (blue line) produced higher fluorescent IgM/IgG binding events to MCF-7 (Figure 2.5A and 2.5C) than did PS B (with TiterMax® Gold) serum alone

(orange line). However, PS B had more fluorescent IgG events than TF-PS B on the

HCT116 (colon) cells. Both TF-PS B and PS B anti-IgM (Figure 2.5D) responses did not

sufficiently recognize HCT-116. A rational explanation for the larger IgG recognition of anti-PS B over anti-TF-PS B (Figure 2.5B) might be because HCT-116 expresses lower quantities of the TF-antigen (CD176) compared to MCF-7 and is a possible reason or the

Figure 2.5. IgG tumor cell binding (A) MCF-7 (blue line) and (B) HCT-116 (blue line). IgM tumor cell binding (C) MCF-7 and (D) HCT-116. N.B. Serum IgG antibodies were detected using commercially available 2o Alexa Fluor488® anti IgG antibody. Serum IgM antibodies were detected using commercially available 2o Alexa Fluor647® anti IgM antibody. disparity in antibody recognition for different carcinomas.46, 54-58

Antibody dependent cellular cytotoxicity (ADCC) is an in vivo and in vitro technique that can be used to determine the potency of antibody responses (Figure 2.6A).

Once an IgG antibody binds to target cells, the Fc portion of the antibody can recruit the

Fc receptor on NK cells (either CD16 or FcRγIII) which triggers the release of granzymes

to lyse the target cell. To further support the potency of the TF-PS B (Figure 2.6B),

ADCC was used to assess the activity of the anti-TF-PS B serum to initiate cell mediated killing. The anti-TF-PS B serum was able to produce 26 % cytotoxicity which was statistically significant compared to PS B, TF-BSA, and the control serum (both anti-TF and PBS). This result noticeably demonstrated the effectiveness of comparing cytotoxicity of TF-PS B to both PS B and TF-BSA. Another method to evaluate antibody responses is complement dependent cytotoxicity (CDC) (Figure 2.6C). Similar to

ADCC, once an antibody is bound to a target cell complement binds to the Fc portion of the antibody which initiates a membrane attack complex to lyse the cell. Interestingly, in

A B

C D

Figure 2.6. Cytotoxicity of MCF-7 using TF-PS B. A) Schematic representation of ADCC. B) MCF-7 ADCC with TF-PS B. C) Schematic representation of CDC. D) MCF-7 CDC with TF-PS B.

(Figure 2.6d) the anti-TF-PS B serum did not produce any complement mediated toxicity.

There are two explanations for the lack of complement mediated cytotoxicity: 1) the IgG antibodies out competed IgM antibodies for binding to MCF-7 and 2) some classes of

IgG antibodies are not effective at fixing complement compared to IgM antibodies.

2.3. Conclusions

The basis for using ZPS-PS B as an immunogen is to support the notion that zwitterionic polysaccharides can be a viable alternative to protein carriers in cancer vaccine development. In thinking about the aforementioned issues associated with carrier proteins and linkers, we have been emphasizing entirely carbohydrate based immune constructs for specific anti-carbohydrate immune responses as opposed to heterogeneous protein constructs consisting of peptide(s)/protein(s) and sugars combined. One key feature of our approach is that the zwitterionic charges on polysaccharides 1 and 2, which are essential components for immune activation, are most likely due to the electrostatic similarities of peptides and specific uptake through C-type lectins. Therefore, using ZPSs as immunogens in cancer vaccine development can be supported through the innate and adaptive immune responses for ZPSs.

The immune response(s) generated from TF-PS B resulted in antibodies specific for the TF disaccharide, void of amino acids, chemical linkers or proteins. The majority of antibody isotypes obtained were IgM; their pentavalent nature allows for increased binding due to higher avidity towards glycans which can result in complement mediated killing.59, 60 The generation of IgG1 and IgG2b isotypes indicates the activation of Th2 and Th1 mediated immunity61, which could be useful in antibody directed cellular cytotoxicity. This contrasts the immune response generated by TACA-PS A1 which we 70

previously reported induces a Th1/Th17 immunity.22 However, IgM/IgG antibodies generated by TF-PS B showed greater fluorescent binding events in flow cytometry than anti-PS B immunoglobulins (Figure 2.5A and 2.5C) by binding to TF expressing MCF-7 cells. Additionally, anti-TF-PS B antibodies showed a preference towards MCF-7 over

HCT-116; it is known that MCF-7 cells have a higher expression level of TF (CD176) than do HCT-116 carcinomas.55-57 Collectively, the anti-TF PS B immune response was able to recognize the TF antigen in both flow cytometry and ELISA, which demonstrated

ZPS-based tumor antigen conjugates can be a viable protein alternative for TACA based cancer vaccines.

In order for us to determine the efficacy of the ZPS-based tumor immune responses we elected to compare data with a TF-BSA protein conjugate (Table 2.2). The results indicated the protein construct was not as equally sufficient in generating higher immunological titers towards TF than the TF-PS B equivalent. These results support a notion that using ZPSs as immunogens increases the immunogenicity of carbohydrate antigens by exploiting innate and adaptive immune responses. One advantage in using

ZPS conjugates is bacterial polysaccharides can cross-link surface receptors on dendritic cells to promote efficient antigen uptake through large carbohydrate oligomers. PS B generates a distinct immune response differing from PS A1 as noted by the absence of expressed IgG3 antibodies that are correlated to a Th17 immune response.62 A plausible explanation for this differentiation could be based on varying interactions with CLRs where PS A1 interacts with DC-SIGN and although currently not entirely understood, PS

B could have interactions with other lectins that have a preference for N-acetylated sugars or even D-fucose.63-65 The importance of using TF-PS B as an immunogen, therefore, is 71

to facilitate uptake on APCs and generate antibodies that can mitigate metastasis of TF containing carcinomas to promote tumor cell killing. The utility of anti-TF antibodies from TF-PS B could also assist in halting metastasis by preventing galectin-3 recognition and by promoting antibody directed cytotoxicity towards cancer cells. The comparison between TF-PS B and Tn-PS A1 is not a valid comparison due to the differences in carbohydrate antigens. A valid comparison in Chapter 3 will investigate an adjuvant effect the Tn antigen has on overall immunogenicity. This will also permit a comparison between investigating TF on different ZPS on both PS A1 and PS B. We are committed to understanding these subtle differences in the immune responses generated from ZPS by future studies which will be reported in due time.

2.4. Future Work

PS B is an interesting ZPS and further investigation is warranted due to both IgG and IgM antibodies specific towards human tumor cell lines. Additionally, the ADCC activity of PS B is a key factor in further studies because the anti-TF antibodies generated were able to produce activity. One could envision using other TACA-PS B conjugates as a mechanism to increase the immune response of entirely carbohydrate based vaccines.

Additionally, further mechanistic studies are needed to be performed such as varying concentration of TACA-PS B conjugates and determining cytokine profiles to assess the efficacy of adjuvant effects.

2.5. Experimental

2.5.1. Culturing B. fragilis and purification of PS B (2). The growth of B. fragilis has been reported previously33 (Chapter 1.5). 20 L of B. fragilis was harvested after 48 h of growth and centrifuged at 4,000 X g for 20 min at 4 °C in 500 mL centrifuge 72

bottles. Cell supernatant was removed and the pellet was re-suspended in 500 mL of 0.15

M NaCl. Liquefied phenol (EMD Millipore) (500 mL) was added to the re-suspended cell pellet and stirred at 70 °C for 30 min. The aqueous layer was removed from the liquefied phenol using a separatory funnel. The aqueous layer was back extracted three times with diethyl ether and dialyzed with SnakeSkin™ dialysis tubing (10K MWCO).

Crude bacterial lysate was treated with RNase (Sigma) and DNase (Sigma) in 0.1 M sodium acetate buffer (pH 4.5), followed by Pronase® (Roche) treatment (pH 7.0) and finally dialysis. The crude mixture was purified by size exclusion chromatography

(Sephacryl S-300 HR) with elution buffer (0.5% sodium deoxycholate, 50 mM glycine, and 10 mM EDTA (pH 9.8)). Fractions were collected and analyzed using UV- spectroscopy; fractions were pooled if there was no absorbance at 260 and 280 nm. The elution buffer was removed by dialysis and crude samples were analyzed by 1H NMR.

The final step in purification required anion-exchange chromatography (DEAE-

Sepharose) to separate the zwitterionic polysaccharides using Tris-HCl (pH 7.3) and a salt gradient from 0 M – 2 M NaCl for elution of the polysaccharides was used. Purity of

PS B was assessed by 1H and 31P NMR.

2.5.2. Synthesis of anomeric aminooxy TF (3). Synthesis of TF-ONH2 has previously been reported.32

2.5.3. Synthesis of TF-PS B (4). Random oxidization of 1.0 mg of PS B using 10 mM of sodium periodate in 0.5 mL 0.1 M sodium acetate buffer pH 5.0 was accomplished by allowing the reaction to stir for 90 min in the dark, followed by quenching with 1 M KCl. TF-ONH2 (3) 2.0 mg was then added to the solution of oxidized PS B and the reaction was allowed to stir overnight. TF-PS B was dialyzed and 73

lyophilized. Conjugation was observed by oxime formation (7.4 - 8.0 ppm) using H1

NMR (see below for spectral data).

2.5.4. TF-BSA (5). Aminooxy TF (3) 5.0 mg was reacted with mercaptoaldehyde

(6)1 for 18 h in sodium acetate buffer (pH 5.5) at room temperature and purified using

Sephadex G-10 and deionized/distilled H2O as the eluent. Fractions containing the TF- linker were lyophilized. 2.5 mg of (7) was deacetylated using Zemplen’s method with

NaOMe in methanol followed by base neutralization with DOWEX 50W x 8-100 ion exchange resin. The solution was then filtered and concentrated under reduced pressure.

1 2.5.5. NMR and MS analysis for Compound (7). H NMR (600 MHz, D2O): (E and Z isomers): δ 7.4 (dd, J1 = 15.1 Hz, J2 = 8.6 Hz, 1HE), 5.3 (dd, J1 = 13.9 Hz, J2 = 4.0

Hz), 4.4-4.3 (m, 4 H), 4.2 (dd, J1 = 12 Hz, J2 = 2.3 Hz, 1 H), 4.0 (d, J1 = 5.7 Hz , 1 H),

3.9-3.8 (m, 2 H), 3.8 (d, J = 3.1 Hz, 2 H), 3.6-3.5 (m, 9 H), 3.5-3.5 (m, 4 H), 3.4-3.4 (m, 2

H), 3.1 (q, 3H), 3.0-2.9 (m, 3 H), 2.3-2.2 (m, 4 H), 1.9 (d, 7 H), 1.5-1.4 (m, 5 H), 1.2 (t, 6

13 H), 0.8-0.8 (m, 6 H). C NMR (150 MHz, D2O): (E and Z isomers): δ 200.9, 174.6,

158.3, 104.7, 98.6, 76.9, 75.0, 72.5, 71.0, 68.6, 60.8, 52.1, 47.6, 46.6, 42.4, 42.0, 30.8,

30.0, 24.7, 21.9, 10.9. LRMS:ESI [M+(Na)+] calcd for 563.19 found 563.1.

2.5.6. Analysis of Compound (5). 2.0 mg of BSA-maleimide (Pierce

Biotechnology) was dissolved in 0.3 mL of reaction buffer (1 x PBS buffer with 0.1 M

EDTA (pH 7.2). Compound 8 was then dissolved in 0.2 mL of reaction buffer and added to a solution containing BSA-maleimide. The reaction proceeded for 24 hr at 4 °C and was extensively dialyzed at 4 °C. Conjugation was analyzed by MALDI-TOF (M/Z

90080.640). Mass loading was calculated using the following equation: (MW of TF-BSA

– MW of BSA Mal)/ (MW of TF-linker). We determined there were 34 molecules of TF- linker conjugated per BSA-maleimide.

2.5.7. Immunizations. Jax C57BL/6 male mice, 6 weeks, were obtained from

Jackson Laboratories and maintained by the Department of Laboratory Animal Resources

(DLAR). All animal protocols were performed in compliance with the relevant laws and institutional guidelines and has been approved by the Institutional Animal Care and Use

Committee (IACUC) of the University of Toledo. Mice were immunized by intraperitoneal injections (i.p.) with 10 μg of TF-PS B, PS B and TF-BSA with and without TiterMax® Gold. Injections were performed on Day 0, 7, 14, and 28. Blood was collected and pooled in a BD Vacutainer® SST™ on Day 32 using a cardiac puncture technique to draw blood. Blood was allowed to clot and serum was separated in

BD Vacutainer® SST™ using a manufacture protocol.

2.5.8. PS B poly-L-lysine (PS B-PLL) and TF-PS B poly-L-lysine (TF-PSB-

PLL).

100 μg of PS B or TF-PS B was added to a test tube containing 0.5 mL of 0.01 M NaOH

(0.001% phenolphthalein indicator) and 0.5 mg of cyanuric chloride. The mixture was vortexed for 1 min and 0.1 mL of 0.1 % poly-L-lysine (PLL) was then added to the mixture, vortexed for 1 min and allowed to react for 3 h at 4 °C on a shaker. The conjugate was diluted to 30 mL with 0.1 M carbonate buffer (pH 9.2).

2.5.9. ELISA. Immulon™ 4 HBX 96 well plates (coated with either PS B/TF-PS

B-PLL or TF-BSA) and maleic anhydride activated 96 well plates (coated with TF-

ONH2) (Thermo Scientific) were used to determine titers from immunized PS B, TF-PS 75

B, and TF-BSA mice. The Immulon™ 4 HBX plates were coated with TF-BSA or TF-

PSB/ or PS B-PLL (3 μg/mL in 0.1 M carbonate buffer (pH 9.2). Maleic anhydride plates were coated with TF-ONH2 (3) as per manufacture instructions. Plates were left at

37 °C for 1 h with shaking and then continued overnight at 4 °C. The plates were then washed three times with washing buffer (1 x TBS, 0.05% Tween 20, pH 7.3) and blocked with blocking buffer (2% BSA, 1 x TBS, pH 7.3) and incubated for 1 h, followed by washing three more times with washing buffer. Anti-sera were initially diluted 1:300 for total antibody titers and 1:100 for IgG isotypes, then serially diluted in half-log10 dilutions and incubated for 2 hr at 37 °C followed by washing three times with washing buffer. Alkaline phosphatase secondary antibodies anti- (kappa, IgG) diluted (1:2000) and (IgM, IgG1, IgG2a, IgG2b, and IgG3) were purchased from (Southern Biotech) diluted (1:1000) and incubated for 1 h, followed by washing three times with washing buffer. PNPP tablets (Pierce) were dissolved in diethanolamine substrate buffer (pH 9.8) and then 100 μL was added to each well for 30 min for sufficient color to develop to detect secondary antibodies. The reaction was quenched with 2 M NaOH. Optical density measurements were obtained using a UV plate reader (Bio-Tek PowerWave HT) and the 96 well plates were read at 405 nm using Gen5 2.0 plate reading software. All assays were performed in triplicate. Titers were determined by regression analysis with half-log10 dilutions plotted against absorbance. The titer cutoff value was set at 0.2 for titer determination. Statistically analysis from ELISAs for experimental groups were compared with the controls using paired t test using GraphPad Prism 6.

2.5.10. Flow Cytometry. MCF-7 and HCT-116 cells lines were provided by (Dr.

Frederick Valeriote, Henry Ford Health Systems). Anti-sera were diluted to 1:200 with 76

FACS buffer (1X PBS, 2% FBS, and 0.001% azide) and incubated with the cell lines (1 x

106 cells) for 30 min on ice. Cells were washed with FACS buffer three times and incubated with secondary antibodies using either AlexaFluor® 488/647 and washed three times. Cells were analyzed by flow cytometry using BD FACSCalibur™ and data analysis obtained using FlowJo software.

2.5.11. Synthesis of STn-PS B (9). Random oxidization of 1.0 mg of PS B using

10 mM of sodium periodate in 0.5 mL 0.1 M sodium acetate buffer pH 5.0 was accomplished by allowing the reaction to stir for 90 min in the dark, followed by quenching with 1 M KCl. 2.0 mg of STn-ONH2 was then added to the solution of oxidized PS B and the reaction was allowed to stir overnight. TF-PS B was dialyzed and lyophilized. Conjugation was observed by oxime formation (7.4 - 8.0 ppm) using H1

NMR (see below for spectral data).

2.5.12. Periodate-Rescorcinol Assay for Sialic Acid. A linear gradient of sialic acid, N-acetyl galactose, and galactose amine was generated from 40, 35, 30, 25, 20, 15,

10, 7.5, 5, 2.5, 1, and 0.5 µg. STn-PS B (9) and PS B (2) were added in triplicate in separate wells at 50 µg per well. 40 µL was placed in triplicate for each concentration in a 96 well plate. 10 µL of 5 mM NaIO4 was placed in each well and incubated for 35 min at 4 °C making a final concentration of 1 mM. 100 µL of rescorinol solution (0.6 g of resorcinol in 100 mL of 17% HCl solution and 0.0025 mM of CuSO4) was added to the well-plate and incubated for 60 min at 90 °C. The unknowns were determined from the sialic acid concentration at 580 nM.

2.5.13. Percent loading of STn-PS B. Sialic acid by weight was determined from the periodate rescorinol assay and STn percent loading was calculated by the following equation.

(Amount of sialic acid (µg) from assay) / (weight of glyconjugate) x (molecular weight STn/molecular weight of sialic acid) x 100%.

2.5.14. Alexa Fluor® 488 percent loading. 100 µg of oxidized PS B was reacted with

100 µg of Alex Fluor® 488-hydrazide (Molecular Probes) for 24 h in PBS buffer pH 7.4 followed by dialysis. The solution was lyophilized and re-dissolved with 100 µl of PBS.

Optical density measurements were obtained using a UV plate reader (Bio-Tek

PowerWave HT) and then the 96 well plates were read at 495 nm using Gen5 2.0 plate reading software. The percent loading of Alexa Fluor® 488 was determined using the manufacturer protocol(s) (Invitrogen Alex Fluor® 488 Protein labeling Kit). The

-1 -1 following equation was used: Moles of dye per mol of PS B = A494 / (71000 cm M X

PS B concentration)

2.6 References

(1) Silverman, E. Biotechnol. Healthc. 2012, 9, 13.

(2) Britten, C. M.; Singh-Jasuja, H.; Flamion, B.; Hoos, A.; Huber, C.; Kallen,

K.-J.; Khleif, S. N.; Kreiter, S.; Nielsen, M.; Rammensee, H.-G.; Sahin,

U.; Hinz, T.; Kalinke, U. Nat. Biotechnol. 2013, 31, 880.

(3) Finn, A. Brit. Med. Bull. 2004, 70, 1.

(4) Chen, D.; Kristensen, D. Expert Rev. Vaccines 2009, 8, 547.

(5) Dasgupta, S.; Erturk-Hasdemir, D.; Ochoa-Reparaz, J.; Reinecker, H.-C.;

Kasper, D. L. Cell Host Microbe 2014, 15, 413.

(6) Leclerc, C.; Schutze, M. P.; Deriaud, E.; Przewlocki, G. J. Immunol. 1990,

145, 1343.

(7) Miles, D.; Papazisis, K. Clin. Breast Cancer 2003, 3, 134.

(8) Slovin, S. F.; Ragupathi, G.; Musselli, C.; Fernandez, C.; Diani, M.;

Verbel, D.; Danishefsky, S.; Livingston, P.; Scher, H. I. Cancer Immunol.

Immunother. 2005, 54, 694.

(9) Livingston, P. O.; Adluri, S.; Helling, F.; Yao, T. J.; Kensil, C. R.;

Newman, M. J.; Marciani, D. Vaccine 1994, 12, 1275.

(10) Xu, Y.; Sette, A.; Sidney, J.; Gendler, S. J.; Franco, A. Immunol. Cell Biol.

2005, 83, 440.

(11) Ju, T.; Aryal, R. P.; Stowell, C. J.; Cummings, R. D. J. Cell Biol. 2008,

182, 531.

(12) Ju, T.; Lanneau, G. S.; Gautam, T.; Wang, Y.; Xia, B.; Stowell, S. R.;

Willard, M. T.; Wang, W.; Xia, J. Y.; Zuna, R. E.; Laszik, Z.; Benbrook,

D. M.; Hanigan, M. H.; Cummings, R. D. Cancer Res. 2008, 68, 1636.

(13) Cazet, A.; Julien, S.; Bobowski, M.; Burchell, J.; Delannoy, P. Breast

Cancer Res. 2010, 12, 204.

(14) Wilson, R. M.; Danishefsky, S. J. J. Am. Chem. Soc. 2013, 135, 14462.

(15) Kirkwood, J. M.; Ibrahim, J. G.; Sosman, J. A.; Sondak, V. K.; Agarwala,

S. S.; Ernstoff, M. S.; Rao, U. J. Clin. Oncol. 2001, 19, 2370.

(16) Miles, D.; Roché, H.; Martin, M.; Perren, T. J.; Cameron, D. A.; Glaspy,

J.; Dodwell, D.; Parker, J.; Mayordomo, J.; Tres, A.; Murray, J. L.;

Ibrahim, N. K.; Group, T. S. Oncologist 2011, 16, 1092.

(17) Ibrahim, N. K.; Murray, J. L. Clin. Breast Cancer 2003, 3, 139.

(18) Danishefsky, S. J.; Shue, Y.-K.; Chang, M. N.; Wong, C.-H. Acc. Chem.

Res. 2015, 48, 643.

(19) Buskas, T.; Li, Y.; Boons, G.-J. Chem. Eur. J. 2004, 10, 3517.

(20) Kudryashov, V.; Glunz, P. W.; Williams, L. J. Proc. Natl. Acad. Sci.

U.S.A. 2001, 96, 3264.

(21) De Silva, R. A.; Appulage, D. K.; Pietraszkiewicz, H.; Bobbitt, K. R.;

Media, J.; Shaw, J.; Valeriote, F. A.; Andreana, P. R. Cancer Immunol.

Immunother. 2012, 61, 581.

(22) De Silva, R. A.; Wang, Q.; Chidley, T.; Appulage, D. K.; Andreana, P. R.

J. Am. Chem. Soc. 2009, 131, 9622.

(23) Trabbic, K. R.; De Silva, R. A.; Andreana, P. R. Med. Chem. Commun.

2014, 5, 1143.

(24) Wang, Q.; McLoughlin, R. M.; Cobb, B. A.; Charrel-Dennis, M.; Zaleski,

K. J.; Golenbock, D.; Tzianabos, A. O.; Kasper, D. L. J. Exp. Med. 2006,

203, 2853.

(25) Monzavi-Karbassi, B.; Pashov, A.; Kieber-Emmons, T. Vaccines 2013, 1,

174.

(26) Gallorini, S.; Berti, F.; Mancuso, G.; Cozzi, R.; Tortoli, M.; Volpini, G.;

Telford, J. L.; Beninati, C.; Maione, D.; Wack, A. Proc. Natl. Acad. Sci.

U.S.A. 2009, 106, 17481.

(27) CDC Ann. Intern. Med. 1984, 101, 348.

(28) Butler, J. C.; Hofmann, J.; Cetron, M. S.; Elliott, J. A.; Facklam, R. R.;

Breiman, R. F. J. Infect. Dis. 1996, 174, 986.

(29) Armand, J.; Arminjon, F.; Mynard, M. C.; Lafaix, C. Journal of Biological

Standardization 1982, 61, 317.

(30) Keitel, W. A.; Bond, N. L.; Zahradnik, J. M.; Cramton, T. A.; Robbins, J.

B. Vaccine 1994, 12, 195.

(31) Klugman, K. P.; Gilbertson, I. T.; Koornhof, H. J.; Robbins, J. B.;

Schneerson, R.; Schulz, D.; Cadoz, M.; Armand, J. Lancet 1987, 2, 1165.

(32) Bourgault, J. P.; Trabbic, K. R.; Shi, M.; Andreana, P. R. Org. Biomol.

Chem. 2014, 12, 1699. 81

(33) Tzianabos, A. O.; Onderdonk, A. B.; Zaleznik, D. F.; Smith, R. S.;

Kasper, D. L. Infect. Immun. 1994, 62, 4881.

(34) Chung, D. R.; Kasper, D. L.; Panzo, R. J.; Chitnis, T.; Grusby, M. J.;

Sayegh, M. H.; Tzianabos, A. O.; Chtinis, T. J. Immunol. 2003, 170, 1958.

(35) Surana, N. K.; Kasper, D. L. Immunol. Rev. 2012, 245, 13.

(36) Tzianabos, A. O.; Finberg, R. W.; Wang, Y.; Chan, M.; Onderdonk, A. B.;

Jennings, H. J.; Kasper, D. L. J. Biol. Chem. 2000, 275, 6733.

(37) Tzianabos, A. O.; Onderdonk, A. B.; Rosner, B.; Cisneros, R. L.; Kasper,

D. L. Science 1993, 262, 416.

(38) Gibson, F. C.; Onderdonk, A. B.; Kasper, D. L.; Tzianabos, A. O. J.

Immunol. 1998, 160, 5000.

(39) Perlmutter, R. M.; Hansburg, D.; Briles, D. E.; Nicolotti, R. A.; Davie, J.

M. J. Immunol. 1978, 121, 566.

(40) Markham, R. B.; Pier, G. B.; Schreiber, J. R. J. Immunol. 1991, 146, 316.

(41) Mitsdoerffer, M.; Lee, Y.; Jäger, A.; Kim, H.-J.; Korn, T.; Kolls, J. K.;

Cantor, H.; Bettelli, E.; Kuchroo, V. K. Proc. Natl. Acad. Sci. U.S.A.

2010, 107, 14292.

(42) Yu, L.-G.; Andrews, N.; Zhao, Q.; McKean, D.; Williams, J. F.; Connor,

L. J.; Gerasimenko, O. V.; Hilkens, J.; Hirabayashi, J.; Kasai, K.; Rhodes,

J. M. J. Biol. Chem. 2007, 282, 773.

(43) Malaprade, L. Bull. Soc. Chim. Franc. 1934, 1, 683.

(44) Kalia, J.; Raines, R. T. Angew. Chem. Int. Ed. 2008, 47, 7523.

(45) Buskas, T.; Thompson, P.; Boons, G.-J. Chem. Commun. 2009, 5335. 82

(46) Yin, Z.; Chowdhury, S.; McKay, C.; Baniel, C.; Wright, W. S.; Bentley,

P.; Kaczanowska, K.; Gildersleeve, J. C.; Finn, M. G.; BenMohamed, L.;

Huang, X. ACS Chem. Biol. 2015, 10, 2364.

(47) Svennerholm, L. Biochim. Biophys. Acta 1957, 24, 604.

(48) Gracie, J. A.; Bradley, J. A. Eur. J. Immunol. 1996, 26, 1217.

(49) Harold F Stils, J. ILAR J. 2005, 46, 280.

(50) Hanisch, F.-G.; Stadie, T.; Bosslet, K. Cancer Res. 1995, 55, 4036.

(51) Li, Q.; Li, Z.; Duan, X.; Yi, W. J. Am. Chem. Soc. 2014, 136, 12536.

(52) Zhu, J.; Wan, Q.; Lee, D.; Yang, G.; Spassova, M. K.; Ouerfelli, O.;

Ragupathi, G.; Damani, P.; Livingston, P. O.; Danishefsky, S. J. J. Am.

Chem. Soc. 2009, 131, 9298.

(53) Dippold, W.; Steinborn, A.; Meyer zum Büschenfelde, K. H. Environ.

Health Perspect. 1990, 88, 255.

(54) Ferguson, K.; Yadav, A.; Morey, S.; Abdullah, J.; Hrysenko, G.; Eng, J.

Y.; Sajjad, M.; Koury, S.; Rittenhouse-Olson, K. Future Oncol. 2014, 10,

385.

(58) Lei, X.; Liu, C.; Azadzoi, K.; Li, C.; Lu, F.; Xiang, A.; Sun, J.; Guo, Y.;

Zhao, Q.; Yan, Z.; Yang, J. Sci. Rep. 2015, 5, 7824.

(59) Navin, T. R.; Krug, E. C.; Pearson, R. D. Infect. Immun. 1989, 57, 1343.

(60) Lefeber, D. J.; Benaissa-Trouw, B.; Vliegenthart, J. F. G.; Kamerling, J.

P.; Jansen, W. T. M.; Kraaijeveld, K.; Snippe, H. Infect. Immun. 2003, 71,

6915.

(61) Mitsdoerffer, M.; Lee, Y.; Jäger, A.; Kim, H.-J.; Korn, T.; Kolls, J. K.;

Cantor, H.; Bettelli, E.; Kuchroo, V. K. Proc. Natl. Acad. Sci. U.S.A. 2

010, 107, 14292.

(62) Bloem, K.; García-Vallejo, J. J.; Vuist, I. M.; Cobb, B. A.; van Vliet, S. J.;

van Kooyk, Y. Front. Immunol. 2013, 4, 1.

(63) Håkansson, K.; Reid, K. B. Protein Sci. 2000, 9, 1607.

(64) van Vliet, S. J.; den Dunnen, J.; Gringhuis, S. I.; Geijtenbeek, T. B.; van

Kooyk, Y. Curr. Opin. Immunol. 2007, 19, 435.

Chapter 3

Increasing Immunogenicity of the TF-Antigen by Targeting MGL2 Receptors Using a Bivalent Tn-TF- PS A1 Conjugate

3.1. Introduction

PS A1, a zwitterionic capsular polysaccharide isolated from the commensal bacteria, Bacteroides fragilis ATCC 25285/NCTC 9343, initiates CD4+ T cell responses.1 The current understanding of zwitterionic polysaccharides as immune stimulants has been suggested to rival the protein paradigm for T cell activation2 bridging the innate and adaptive immune gap.3 Innate immunological mechanistic studies of PS

A1 clearly show interactions with toll-like receptor 2 (TLR-2)/CD282 and DC-SIGN, which are essential for efficient uptake by antigen-presenting cells (APCs) or dendritic cells (DCs).4,5 While PS A1 has shown promise as an immunogenic stimulant and a possible alternative to protein-based cancer vaccines,6,7 it can additionally contribute to the production of Th17 immunity.7 The production of Th17 cells has been shown to be essential in protection against Staphylococcus aureus,8 Mycobacterium tuberculosis,9 and in particular cancers.10,11 The activation of Th17 immunity is bifunctional in that the production of TGF-β can influence the valuable Th17 immune response but can also

influence the production of T regulatory cells (T regs).12 Therefore, investigating ZPSs as “carriers” in vaccine development requires innovative design strategies to decrease the regulatory immune responses through selective modifications that allow for interactions with other innate immune receptors.

The design and development of carbohydrate-based vaccines remains critical for targeting disease states where unique sugar antigens are normally the first-line in immune surveillance. Carbohydrates, however, have long been known to be weakly immunogenic in eliciting valued T cell responses. One strategy for improving immunogenicity involves tailoring the immunogen to target specific innate immune receptors found on

APCs or DCs.13 These immune response receptors have evolved to differentiate self- versus pathogen-associated molecular patterns (PAMPs). PAMPs are often composed of diverse glycan structures that can be broadly recognized by C-type lectin receptors

(CLRs) and TLRs.14 CLRs are best characterized as calcium-dependent proteins expressed on myeloid cells to promote efficient antigen uptake which ultimately leads to presentation to CD4+ T cells.15 Fortunately, CLRs are selective for conserved carbohydrate recognition domains concomitantly leading to pathogen clearance.16-18

A specific CLR, macrophage galactose binding lectin 2 (MGL2) serves as a valuable surface receptor for vaccine development due to selectivity towards N- acetylgalactosamine (GalNAc), the sugar component of the Thomsen nouveau (Tn) cancer antigen.19 The Tn antigen is an important tumor associated carbohydrate antigen

(TACA) involved in the onset and progression of tumors.15,20,21 Since carbohydrates/TACAs are known for being weakly immunogenic, targeting CLRs with a simple, covalently linked sugar may prove useful for enhanced phagocytosis and 86

increased immunogenicity. Therefore, attaching N-acetylgalactosamine (GalNAc) as a small molecule activator for innate and adaptive immune responses holds promise for an increased uptake of TACA-based vaccine constructs in which a myriad of carriers are currently being examined. 6,22-26

While the Tn antigen has been extensively studied in cancer vaccine related studies, it has only recently been investigated for its ability to bind MGL2.27 Leclerc and co-workers demonstrated a correlation between increased Tn density on MUC-6 (15 amino acid peptide fragment) and enhanced antigen uptake by APCs in comparison to a peptide fragment alone.28 This result was believed to be the consequence of improved binding to MGL2.28 Furthermore, Danishefsky and co-workers engineered a comprehensive polyvalent vaccine mix composed of seven monomeric TACA conjugates: 1) a MUC 1-G5 peptide containing 8 conjugated Tns, 2) a Tn cluster, 3) an

STn cluster, 4) a TF cluster, 5) one consisting of Globo-H, 6) one consisting of GM3, and

7) a Lewis Y immunogen.29 The polyvalent mixture proved to be effective in the recognition of respective TACAs but there was no discernible immunological titer difference or adjuvant effect between the monovalent (single TACA) to heptavalent

(mixture of TACAs) immunizations with keyhole limpet hemocyanin (KLH) conjugates.

However, there was a noted two-fold increase in IgG titer values when peptide MUC1-

G5 was used in monovalent to polyvalent immunizations.29 The likely rationale for Tn not contributing to an adjuvant effect to other TACAs is the notion of super cross-linking

CLRs which can decrease antigen uptake and presentation and impair proper immune recognition.30 This effect is also observed in the natural setting with microorganisms and endogenous glycosylated MUC-1 clearance leading to tumor evasion. Collectively, two 87

independent studies showed the propensity of the Tn antigen to either increase antigen uptake or increase specific antibody generation. Therefore, with the aforementioned issues in over stimulating CLRs with regards to vaccine design, incorporating Tn on a unimolecular construct with another TACA could help promote efficient antigen uptake through MGL2 and subsequently increase immunogenicity.

In addition to targeting CLRs such as MGL2, a balanced immune response is critical for overall efficacy in which Th1, Th2, and Th17 can assist in antibody production. Along those lines, C-type lectins have the ability to influence cytokine production and can be considered targets for self-adjuvanting vaccine constructs.31,32

Cytokines are often associated with an induction of suppressive T reg responses initiated through interactions with TLR-2.33 The presence of proinflammatory markers such as

IL-6, IL-4, and IFN-γ can negate side effects of T regs. One reagent that leads to the production of IL-6 is adjuvant monophosphoryl lipid A (MPLA), which can overcome a decrease in suppressive immune responses (T regs).34,35 Although the mechanistic function of MGL2 remains elusive, a similar signalling event belongs to the family of the well-studied asialoglycoprotein receptors (ASGP-R), which initiate pro-inflammatory cytokines.21 As a charge for this current work, and based on our previous results,7 the

Tn-PS A1 cytokine data showed reduced IL-10 and increased IL-4/IL-17 expression, which was found to be distinct from PS A1 alone. Furthermore, interaction with MGL2 is known to produce IL-4.36 It is possible that this switch in cytokine profiles could be explained as Tn interacting with MGL2 thereby providing access to an alternate processing pathway as opposed to one for PS A1 alone.7 Since MGL2 has been demonstrated to skew the immune response to Th2, targeting this receptor with the Tn 88

antigen is a viable strategy to increase immunogenicity of various TACA-PS A1 constructs.

Therefore, using the aforementioned knowledge, we sought to create a novel immunogen with dual functionality through: 1) an ability to increase the immune response towards TACA Thomsen Friedenreich (TF = D-Galβ1,3-D-GalNAc) by conjugating both Tn and TF to PS A1 and 2) the use of Tn to target the MGL2 receptor.

Our approach has the advantage of using ZPS PS A1 4a-c (Figure 3.1) which can augment multivalency effects37 leading to higher degrees of interactions by Tn on the surface of APCs and thus antigen internalization. Collectively, incorporating Tn on 1 will induce an adjuvant effect by involving key components of innate immune receptors such as MGL2 and subsequently activate adaptive immune responses with T and B cells.

Herein, we present our data demonstrating this effect through increased antibody recognition of TACAs observed with anti-serum from a bivalent vaccine construct containing both Tn and TF antigens as compared to their monovalent counterparts. Our working hypothesis is that a Tn-TF-PS A1 (4c) bivalent conjugate will interact with

MGL2 and increase APC uptake thus increasing immunogenicity towards the TF antigen as opposed to monovalent TACA conjugate TF-PS A1 (4b) (Figure 3.1).

3.2. Results and Discussion

The synthesis of Tn-TF-PS A1 (4c) (Figure 3.1A) was achieved using sodium periodate oxidation of 16 followed by conjugation of 26 and 338 in a 1:1 molar ratio. This led to an overall loading of 29% (16.5% TF and 12.5% Tn). The loading of 4c was determined by NMR integration of the N-acetyl groups from compound 1. Figure 3.1B illustrates the 1H NMR overlays of 1 and 4a-c to denote the chemical transformation 89

Figure 3.1. (A) Synthetic modification of PS A1 (1) to incorporate Tn and TF and (B) 1H NMR overlay of PS A1 conjugates 1 and 4a-c at 60 oC.

characterized by the oxime link (see Appendix C for NMR details of 4c).

Scheme 3.1. General synthesis of Tn-BSA and TF-BSA.

To evaluate the immunological potency of compounds 4a-c, Jax C57BL/6 mice were vaccinated using two separate adjuvants: 1) TiterMax Gold® (TMG) and 2) Sigma adjuvant system® (SAS), and antibody responses were determined by ELISA.

Additionally, binding specificity of polyclonal antibodies towards either Tn-BSA or TF-

BSA (see Scheme 3.1) were used to parse out individual contributions of constructs 4a-c.

Firstly, 4a (Figure 3.2A-B) demonstrates strong IgG/IgM specificity towards Tn-BSA when TMG is used as the adjuvant, however, the overall titer value is increased when

SAS is employed. Furthermore, 4a has minimal cross-binding with TF-BSA (Figure

3.2B-C) suggesting that the immune response that is generated in mice favors the Tn antigen. However, 4b has minimal IgG binding to TF-BSA (Figure 3.2C) and Tn-BSA

(Figure 2A) when both TMG and SAS were employed. Important to note is that the flexibility of the α-glycosidic linkage in the TF antigen decreases its immunogenicity.39

We confirmed that there were statistically significant differences in IgG production 91

between 4b and 1 as imunogens (see Appendix 3 , Figure A3.1). This result verified that there was an IgG specific response generated towards TF but that the PS A1 was required for ultimate IgG recognition of the TF hapten.

It was initially thought that using SAS adjuvant could help boost the IgG immune response similar to the effect noticed with 4a, however, no discernible differences were observed. The immune response with 4b remained exclusively an IgM isotype with high

Figure 3.2. ELISA specificity of TACA-conjugates (4a-c). (A) IgG specificity towards Tn-BSA. (B) IgM specificity towards Tn-BSA. (C) IgG specificity towards TF-BSA. (D) IgM specificity towards TF-BSA. Both PS A1 and PBS control mice sera showed no cross-reactivity to either Tn-BSA or TF-BSA. Data are illustrated as mean ± s.e.m. * P < 0.05; two tailed Student’s t-test. titer values observed towards TF-BSA (Figure 3.2D) and moderate cross-reactivity to Tn-

BSA (Figure 3.2B). An interesting caveat occurred with 4b and SAS as this combination decreased IgM production compared to 4b and TMG (Figure 3.2D). Since TMG and

SAS have been proven to have minimal effects in producing specific IgG antibodies towards the TF antigen in 4b, a new construct had to be designed to incorporate the immune stimulating properties of 4a while a continuing focus on TF remained. For this matter, we turned to the simultaneous conjugation of both 2 and 3 to oxidized PS A1 giving 4c (Figure 3.1A). Prior to immunological evaluation, we were under the impression that the addition of Tn would interact with CLR MGL2 to promote a bivalent- targeted immunogen for increased antigen uptake and presentation of both Tn and TF antigens. As noted from Figure 3.2, when monovalent constructs 4a and 4b were administered to Jax C57BL/6 mice, there was minimal cross-reactivity to either immunogen. However, incorporating both Tn and TF onto PS A1 led to enhanced

IgG/IgM responses when coating constructs Tn-BSA and TF-BSA were used in the

ELISA. Both TMG and SAS (Figure 3.2A-2D) were used as adjuvants in separate immunization studies. While the IgG/IgM specificity towards Tn-BSA, when 4c was used as the immunizing construct, remained similar to monovalent counterpart 4a, there was a drastic change in TF-BSA IgG specificity from 4b to 4c. Moreover, 4c with SAS led to an ~2.5 fold change from 4c when TMG was used. This indicates that in addition to Tn interacting with MGL2, MPLA (the active component in SAS) augmented the immunogenicity of 4c.

In order to validate MGL2 interaction with 4a and 4c, four biotinylated PS A1 conjugate probes (Scheme 3.2; 5a-5d) were synthesized. The probes were constructed using 1 or 4a-c with sulfo-NHS-biotin. Constructs 5a-d were used in a colorimetric assay where MGL2 coated ELISA plate wells and streptavidin-alkaline phosphatase detected binding interactions. Compounds 5a-d were evaluated for their ability to bind to

MGL2 (Figure 3.3A). It is important to note that 5d was used as a negative control

Scheme 3.2. Syntheses of biotinylated TACA-PS A1 (5a-c) from TACA-conjugates (4a-c) as MGL2 assay probes because it is known not to interact with MGL2 and would account for a biotinylated linear probe similar to constructs 5a-c. Only compounds 5a and 5c showed binding to MGL2 due to the presence of Tn. However, 5b showed binding that was most likely augmented by

multivalent interactions with MGL2. Constructs 5a-c (10 µg/mL) (Figure 3.3B) were

Figure 3.3. (A) MGL2 binding assay and inhibition using probes 5a-d (Scheme 1). (B) Percent inhibition by Tn-BSA (10µg/mL) with 5a-c (10 µg/mL). Data are illustrated as mean ± s.e.m. * denotes % inhibition by Tn-BSA. shown to be competitively inhibited by 10 µg/mL of Tn-BSA giving 44% inhibition for

5a, 64% for 5b, and 53% for 5c. Compound 5b was inhibited the most by Tn-BSA due to MGL2 binding preference of Tn over TF. However, 5a was favored over 5c due to the presence of TF which most likely interfered in the binding event.

To further support the notion of increased antibody generation through bivalent construct Tn-TF-PS A1, flow cytometry was used to determine polyclonal antibody 95

binding to human tumor cells MCF-7 (Figure 3.4A) and OVCAR-5 (Figure 3.4B).

Validation of the anti-serum of 4c showed a 97% gated-shift in fluorescently labeled cell

Figure 3.4. Flow cytometry with anti-serum from 1 and 4a-c with secondary Alexa Fluor® 488 anti-IgG using human tumor cell lines. (A) MCF-7 human breast tumor cell line. (B) OVCAR-5 human ovarian tumor cell line. populations compared to MCF-7 cells alone. For comparison, the shift in fluorescent cells for control serum was 8%, 1 gave 10%, 4a gave 23%, and 4b gave 41%. Similar binding events were seen using human ovarian tumor cell line OVCAR-5 with 4c giving a 98% shift in fluorescently sorted cell population compared to the PBS control at 7%, 1 at 5%, 4a at 23%, and 4b giving 49%. The increased fluorescent cell populations of anti-

4b serum to MCF-7 and OVCAR-5 comes as a surprise due to the low IgG binding to

TF-BSA on ELISA (Figure 2C).

After evaluating antibody binding by flow cytometry, our final step was to assess

antibody function using complement dependent cytotoxicity (CDC) (Figure 3.5A-B). An

LDH assay was used to measure the amount of LDH released from either MCF-7 (Figure

3.5A) or OVCAR-5 cells (Figure 3.5B) by lysis of cancer cells with antibodies generated from (1 and 4a-4c) and rabbit complement. LDH is an oxidoreductase enzyme which catalyzes the conversion of lactate to pyruvate coupled with the reduction of NAD+ to

NADH. Subsequently, Diaphorase uses NADH to reduce iodonitrotetrazolium to formazan which can be analyzed at 490 nm (Figure 3.5C)

Figure 3.5. Antibody mediated CDC with anti-serum from 1 and 4a-4c plus rabbit complement. (A) MCF-7 human breast tumor cell line. (B) OVCAR-5 human ovarian tumor cell line. Data are illustrated as mean ± s.e.m. * P < 0.05, ** P < 0.005, *** P < 0.0005; two tailed Student’s t-test. (C) Overview of LDH assay.

In Figure 3.5A, 4c had 59% cytotoxicity towards MCF-7 with statistically significant values (P-value < 0.05) in comparison with 4a 52% and 4b 50%.

Additionally, 4c had 53% cytotoxicity towards OVCAR-5 which again produced statistically significant values (P-value < 0.005) over 4a 39% and 4b 43%. Collectively,

4c gave greater cytotoxicity over monovalent equivalents 4a and 4b which is an additional advantage of an increased immune response.

The induction of Th17 immune responses by IL-23 has been shown to induce B cell proliferation and antibody production by the production of IL-17 and IL-21.40 Both of these cytokines are involved with isotype switching that causing the production of

IgG1, IgG2a, IgG2b, and IgG3 antibodies.

The addition of Tn to PS A1 initiated a profound promotion of cytokines associated with the initiation and proliferation of the Th17 immune response (IL-17, IL-

22, and IL-23) (Figure 3.6). The Th1 associated cytokine IFN-γ also was increased based on cells immunized with adjuvants TMG or SAS. Additionally, the addition of sugars such as the TF antigen showed increased levels of IL-10 which has been shown to dampen an immune response. This is a potential reason why TF-PS A1 IgG antibody production is decreased in Figure 3.2C. Considering the augmented immune response of

Tn-TF-PS A1, the increased immune response could be directly related to the decreased

IL-10 and increased IL-17 production observed with Tn-TF-PS A1 (Figure 3.6), ultimately increasing the immune response.

The effects of adjuvant between TMG and SAS, shows SAS increases

Figure 3.6. Cytokine profile for SAS and TMG immunizations of PS A1, Tn-PS A1, TF-PS A1, and Tn-TF-PS A1 immune response for PS A1 and TACA-PS A1 conjugates by increasing cytokine expression of Th17 associated cytokines (IL-17 and IL-22) and Th1 cytokines of IFN-γ.

The increase of these cytokines while using SAS is correlated exclusively to IgG2b. 99

3.3. Conclusions

The development of a bivalent Tn-TF-PS A1 construct, using a semi-synthetic approach, has led to the increased immunogenicity of the TF antigen. This increased immune response can be attributed to a targeted MGL2 strategy leading to an increased uptake of TACAs. This stands in contrast to other multivalent approaches that have been engineered in which there was no major effect on the individual TACAs alone.29 The success of 4c is distinct from other polyvalent immunogens (globular protein conjugates) mostly likely due to the linear and repetitive nature of 1 and 4c leading to higher surface area contact to DCs and multivalent interactions. We note that this is the first time in which a bivalent Tn-TF conjugate has had an enhanced immune response in increasing the binding events to TF.

Construct 4a was consistent in mounting an IgG specific immune response to the

Tn antigen when TMG or SAS were used as external adjuvants. However, proving the same strategy to accommodate the TF antigen was more challenging in 4c. The results indicated that the use of different adjuvants had relatively no effect on IgG titer values.

When PS A1 was bivalently conjugated with both Tn and TF (4c), there was a profound difference in anti-TF IgGs compared to 4b. A comparable result was also observed in

Figure 3.4 and 3.5 where the anti-serum from 4c was able to bind and contribute to the cytotoxicity of human tumor cell lines MCF-7 and OVCAR-5 greater than the monovalent equivalents. The addition of Tn signifies the importance of binding MGL2 which corresponds to higher immunological activity. To differentiate between multivalent polyclonal antibodies from 4c, either Tn- or TF-BSA ELISA coatings were screened to assess antibody specificity. To determine if MGL2 contributed to an increase 100

in TF immune response towards 4c, four biotinylated probes 5a-d were evaluated for binding to MGL2. Both 5a and 5c had similar binding profiles to MGL2, which signifies the addition of the Tn antigen promoted efficient uptake of the immunogen. However,

TF has been documented to have lower affinity towards MGL2, which was confirmed by examining data from Figure 4.19,27 When 10 µg/mL Tn-BSA was used to compete with the binding of PS A1-biotin derivatives, Tn-PS A1 biotin was shown to be inhibited at

44% whereas Tn-TF-PS A1 was inhibited at 53% when an equivalent concentration of 10

µg/mL was used. The inhibition of Tn-TF-PS A1 appears to be affected by the conjugation of TF because TF has less affinity for MGL2 and therefore more susceptible to inhibition by Tn-BSA. The negative control in the experiment was PS A1 due to the fact that it has no binding to MGL2 (Figure 3.4).

The rationale behind using Sigma Aldrich Adjuvant (SAS) was the incorporation of an

MPLA-based adjuvant to overcome potential suppressive T reg responses.34 This adjuvant is distinct from TMG, which is a potent oil and water emulsion that allows slow release of the immunogen. When SAS was administered with 4c, there was increased immunogenicity to the TF antigen but also enhanced immunogenicity was observed in formulations with TMG. It is important to note the overall titers from Figure 3.3 suggest the combined effects of MGL2 and SAS decrease the suppressive effects of IL-10 from the added cytokine production from MPLA based adjuvant by creating a proinflammatory environment and increasing IgG responses. Contrastingly, Van kooyk and co-workers have shown that targeting TLR-2 and MGL2 separately augmented IL-10 values;41 an effect that was not seen with Tn-PS A1.7 Additionally, targeting the MGL2 receptor alone showed a decrease in production of IL-10.42 Therefore, the semi-synthetic 101

modifications to PS A1 in the forms of 4a and 4c could impair TLR-2 interactions and decrease endogenous IL-10 thus promoting a pro-inflammatory immune response. The change in zwitterionic character did not previously shown a significant change in antigen uptake by APCs. Therefore the terminal sugar whether it comes from naturally occurring

PS A1 through Galactofuranose (Galf), or conjugated sugars GalNAc (Tn) and Gal (TF) plays a significant role in modulating the immune response noted by cytokine responses

(Figure 3.7).

↑IFN-γ, ↑IL-17, ↑IL-22, ↑IL-23 ↓IL-22, ↓IL-23, ↓IL17 ↓IL-10 ↑IL-10

Figure 3.7. Immune modulation by carbohydrate addition relative to PS A1

The incorporation of the Tn antigen to TF-PS A1 has had a profound influence on the respective immunological activity corresponding to an increase of the following parameters: a) IgG antibodies specific towards TF, b) binding to tumor cells and c) complement dependant cytotoxicity. The mechanism behind this activity is increased

MGL2 binding by the Tn antigen which reveals a novel targeted vaccine method for enhanced antigen uptake and greater immunological activity. Since PS A1 has been

102

noted to bind to DC-SIGN,4 it is a possibility that other lectins could be involved in the initiation of this immune response. This model has the potential of being adapted to multiple vaccines formats including peptides, proteins, nanoparticles, and lipids to increase the therapeutic potential of carbohydrate-based vaccines.

3.4. Future Work

The foundation to unimoleular bivalent Tn-TF-PS A1 construct has been established with data strongly suggesting the importance of the addition of Tn to PS A1.

Furthermore, adaptations of this work include varying the concentration ratio of Tn and

TF on PS A1, prophylactic, and therapeutic tumor challenges. To validate the increased tumor cell binding and cytotoxicity, the monovalent and bivalent PS A1 constructs need to be examined to assess therapeutic efficiency.

Since PS A1 has been originally demonstrated to interact with DC-SIGN and Tn-

TF-PS A1 has been shown to interact with MGL2, the change in terminal sugars could affect CLR interactions. Therefore a potential study to determine if binding to DC-SIGN is abrogated is to use biotinylated probes 5a-d to determine if binding is still present in the TACA-PS A1 conjugates. The predicted outcome is reduced binding to DC-SIGN but function will still be present due to the total percent conjugation which is approximately 30-35 %. This indicates that galactofuranose (proposed to interact exclusively with DC-SIGN) remains unmodified and available to interact with DC-SIGN.

Additionally, increasing the immune response to the TF antigen using a bivalent approach incorporating the Tn antigen is a major discovery in the field of glyco-based immunology. These results present the opportunity to adapt this approach to protein

103

based vaccines to replace the failed phase III clinical trial of the vaccine Theratope (STn-

KLH).

3.5. Experimental

3.5.1. Synthesis

Growth of Bacteroides fragilis ATCC 25285/NCTC 9343 and isolation of PS A1 (1) has been previously reported Chapter 1.5.

1 Tn-ONH2 (2) has been previously reported.

2 TF-ONH2 (3) has been previously reported.

Tn-PS A1 (4a) has been previously reported.1

TF-PS A1 (4b) has been previously reported.38

3.5.2. Synthesis of Tn-TF PS A1 (4c). A 2 mM solution of NaIO4 was used to oxidize 1 mg of PS A1 in 0.5 mL of NaOAc buffer pH 5.2 for 90 min. KCl was used to quench excess NaIO4. A 1:1 molar ratio of Tn-ONH2 (2) to TF-ONH2 (3) (1.0 mg and

1.7 mg respectfully) were allowed to react with oxidized PS A1 for 24 hours followed by a long stint of dialysis using 10 kDa MWCO Snakeskin™ tubing. Percent loading was calculated from the following formula: For TF-ONH2 (% loading = MW TF-ONH2/MW

TF-PS A1 hexasaccharide conjugate x mol fraction). The mol fraction was obtained from

NMR integration of the respective two NHAc methyl protons from PS A1 and the NHAc from the TF antigen. Percent Tn-ONH2 loading was determined by using the formula:

For Tn-ONH2 (% loading = MW Tn-ONH2/MW Tn-PS A1 pentasaccharide conjugate x mol fraction). The mol fraction was obtained from NMR integration of the respective two NHAc methyl protons from PS A1 and the NHAc from Tn antigen.

3.5.3. Biotinylated PS A1/Conjugate Probes (5a-d). 1.0 mg of either PS A1 (1), 104

Tn-PS A1 (4a), TF-PS A1 (4b), or Tn-TF-PS A1 (4c) was reacted with 0.5 mg of sulfo-

NHS-biotin (100X equivalents) (ProteoChem) in 0.5 mL of 1XPBS buffer pH 7.4 for 24 h at room temperature. The PS A1 probes were dialyzed, lyophilized, and reconstituted in 1X DPBS buffer (with CaCl2/MgCl2) pH 7.2 at a concentration of 1 mg/mL. Activity of the probes were evaluated in streptavidin based assays, as described below, in the

MGL2 binding assay.

3.5.4. 3-oxopropyl ethanethioate (mercaptoaldehyde) (8) A catalytic amount of of piperidine (5.0 µL) was added to 0.5 mL acrolein (6) at 0 oC. Then 0.52 mL of thioacetic acid (7) was added dropwise over a period of 30 minutes. The reaction was carried out for 12 hours and the reaction mixture was then concentrated under vacuum and purified by column chromatography using 30% EtOAc/70% DCM as the eluent to give mercaptoaldehyde (8) in 95% yield. 1H NMR (CHLOROFORM-d ,600MHz): δ 9.67

(d, J=1.0 Hz, 1 H), 3.03 (t, J=1.0 Hz, 2 H), 2.73 (t, J=1.0 Hz, 2 H), 2.25 ppm (d, J=1.0

13 Hz, 3 H); C NMR (150 MHz, D2O): δ 200.1, 195.6, 43.8, 30.7, 21.7. LRMS:ESI

[M+(Na)+] calcd for 155.01 found 155.0.

3.5.5. General Procedure for TACA linkers 9 (Tn) and 10 (TF). Aminooxy Tn

(2) (5.0 mg) was reacted with 2.8 mg of mercaptoaldehyde (8) for 18 h in sodium acetate buffer (pH 5.5) at room temperature and purified using Sephadex G-10 and deionized/distilled H2O as the eluent. Fractions containing the Tn-linker (9) were

1 lyophilized. H NMR (D2O,600MHz): δ 7.51 - 7.54 (m, 1 H), 6.89 - 6.92 (m, 1 H), 5.37

(d, J=3.7 Hz, 1 H), 5.28 - 5.30 (m, 1 H), 4.15 - 4.21 (m, 1 H), 3.90 - 3.95 (m, 1 H), 3.80 -

3.88 (m, 2 H), 3.59 - 3.68 (m, 2 H), 2.92 - 3.03 (m, 2 H), 2.58 - 2.71 (m, 1 H), 2.43 (ddd,

J=9.5, 6.2, 2.9 Hz, 1 H), 2.24 - 2.29 (m, 2 H), 1.91 - 1.97 (m, 3 H), 1.80 ppm (br. s., 1 H); 105

13 C NMR (150 MHz, D2O): δ 200.9, 174.5, 154.6, 104.7, 98.7, 76.8, 74.9, 72.4, 70.5,

68.5, 60.8, 47.6, 30.0, 29.3, 25.8, 25.4, 25.0, 23.2, 21.9. LRMS:ESI [M+(Na)+] calcd for

373.10 found 373.1

Aminooxy TF (3) (5.0 mg) was reacted with 1.7 mg of mercaptoaldehyde (8) for

18 h in sodium acetate buffer (pH 5.5) at room temperature and purified using Sephadex

G-10 and deionized/distilled H2O as the eluent. Fractions containing the TF-linker (10)

1 were lyophilized. H NMR (D2O,600MHz): δ 7.53 (t, J=6.2 Hz, 1 H), 6.91 (s, 1 H), 5.37

(d, J=4.0 Hz, 1 H), 5.28 (d, J=4.0 Hz, 1 H), 4.40 (d, J=8.1 Hz, 1 H), 4.32 - 4.39 (m, 1 H),

4.19 (d, J=2.9 Hz, 1 H), 4.16 (d, J=2.9 Hz, 1 H), 3.95 (dt, J=11.2, 2.7 Hz, 1 H), 3.84 -

3.89 (m, 1 H), 3.80 (d, J=2.9 Hz, 1 H), 3.59 - 3.69 (m, 2 H), 3.49 - 3.58 (m, 1 H), 3.42

(dt, J=9.9, 7.7 Hz, 1 H), 2.92 - 3.05 (m, 1 H), 2.57 - 2.75 (m, 1 H), 2.44 (q, J=6.6 Hz, 1

H), 2.33 (s, 1 H), 2.24 - 2.28 (m, 1 H), 1.91 (s, 1 H), 1.80 ppm (s, 1 H); 13C NMR (150

MHz, D2O): δ 200.9, 174.5, 154.6, 104.7, 98.7, 76.8, 74.9, 72.4, 70.5, 68.5, 60.8, 47.6,

30.0, 29.3, 25.8, 25.4, 25.0, 23.2, 21.9. LRMS:ESI [M+(Na)+] calcd for 535.15 found

535.1.

3.5.6. BSA-Maleimide (11). 1.0 mg of BSA was dissolved in 300 µL of 1XPBS buffer with 1 mM EDTA (pH 7.2) and reacted for 30 min with 100 µL of a 2 mM of 3-

(maleimido)propionic acid N-hydroxysuccinimide solution in 1 mL of DMF. Excess 3- maleimidopropionate was removed by centrifugal ultrafiltration (Vivaspin® 6 MWCO 10 kDa) and washed three times with 5 mL of 1XPBS buffer containing 1 mM EDTA (pH

7.2). Conjugation was analyzed by MALDI-TOF and determined to be M/Z 71686.967.

Mass loading was calculated using the following equation: (MW of BSA-maleimide –

MW of BSA (664303)/(MW of maleimide linker). Based on the molecular weight, we 106

were able to determine that there were 34 molecules of maleimide linked to BSA.

3.5.7. Tn-BSA (12). 2.5 mg of Tn-linker (9) was deacetylated using Zemplen’s method consisting of NaOMe in methanol followed by neutralization with DOWEX 50W x 8-100 ion exchange resin. The solution was then filtered and concentrated under reduced pressure. The deacetylated Tn-linker was dissolved in 0.1 mL of 1XPBS buffer with 1 mM EDTA (pH 7.2) and added to a 1.0 mg solution of (11) in 0.2 mL 1XPBS buffer. Conjugation was analyzed by MALDI-TOF (M/Z 78273.845). Mass loading was calculated using the following equation: (MW of Tn-BSA – MW of BSA- maleimide)/(MW of Tn-linker). From this method, we determined that there were 14 molecules of Tn-linker conjugated per BSA-maleimide.

3.5.8. TF-BSA (13). 2.5 mg of TF-linker (10) was deacetylated using Zemplen’s method consisting of NaOMe in methanol followed by neutralization with DOWEX 50W x 8-100 ion exchange resin. The solution was then filtered and concentrated under reduced pressure. The deacetylated TF-linker was dissolved in 0.1 mL of 1XPBS buffer with 1 mM EDTA (pH 7.2) and added to a 1.0 mg solution of (11) in 0.2 mL 1XPBS buffer. Conjugation was analyzed by MALDI-TOF (M/Z 78273.845). Mass loading was calculated using the following equation: (MW of TF-BSA – MW of BSA- maleimide)/(MW of TF-linker). We determined there were 14 molecules of TF-linker conjugated per BSA-maleimide.

1 H NMR (D2O,600MHz): δ 7.53 (t, J=6.2 Hz, 1 H), 6.91 (s, 1 H), 5.37 (d, J=4.0 Hz, 1

H), 5.28 (d, J=4.0 Hz, 1 H), 4.40 (d, J=8.1 Hz, 1 H), 4.32 - 4.39 (m, 1 H), 4.19 (d, J=2.9

Hz, 1 H), 4.16 (d, J=2.9 Hz, 1 H), 3.95 (dt, J=11.2, 2.7 Hz, 1 H), 3.84 - 3.89 (m, 1 H),

3.80 (d, J=2.9 Hz, 1 H), 3.59 - 3.69 (m, 2 H), 3.49 - 3.58 (m, 1 H), 3.42 (dt, J=9.9, 7.7 107

Hz, 1 H), 2.92 - 3.05 (m, 1 H), 2.57 - 2.75 (m, 1 H), 2.44 (q, J=6.6 Hz, 1 H), 2.33 (s, 1

13 H), 2.24 - 2.28 (m, 1 H), 1.91 (s, 1 H), 1.80 ppm (s, 1 H); C NMR (150 MHz, D2O): δ

200.9, 174.5, 154.6, 104.7, 98.7, 76.8, 74.9, 72.4, 70.5, 68.5, 60.8, 47.6, 30.0, 29.3, 25.8,

25.4, 25.0, 23.2, 21.9. LRMS:ESI [M+(Na)+] calcd for 535.15 found 535.1.

3.5.9. Immunizations. Jax C57BL/6 male mice (6 weeks) were obtained from

Jackson Laboratories and maintained by the Department of Laboratory Animal Resources

(DLAR) at the University of Toledo. All animal protocols were performed in compliance with the relevant laws and institutional guidelines set forth by the Institutional Animal

Care and Use Committee (IACUC) of the University of Toledo.

Sample sizes (n=5) were chosen based on desired amount of blood sera (1 mL/mouse). Mice were distributed randomly without bias. Criterion for inclusion of mice depended on the health status of the mouse. If mice were shown to have ascites or signs of distress the mouse was euthanized. However, no abnormalities occurred throughout the duration of the experiment.

Vaccinations with Titermax Gold Individual Tn-, TF- and Tn-TF-PS A1 constructs (20 µg) were mixed in a 1:1 ratio of 50 uL of TiterMax® Gold and injected into 7 wk old C57BL/6 mice (Jackson Laboratory) (each construct was administered individually – not mixed). Mice groups (n=5) were immunized by intraperitoneal injections (i.p.) on day 0, 14, 28, 42. Blood sera were obtained using a cardiac puncture technique on day 52.

Vaccinations with Sigma Adjuvant System. Individual Tn-, TF- and Tn-TF-PS A1 constructs (20 µg) were mixed in a 1:1 ratio of 100 µL of Sigma Adjuvant System

(Sigma-Aldrich) and injected into 7 wk old C57BL/6 mice (Jackson Laboratory) (each 108

construct was administered individually – not mixed). Mice groups (n=5) were immunized by intraperitoneal injections (i.p.) on day 0, 21, 42, per manufacture’s instructions. Blood sera were obtained using a cardiac puncture technique on day 52.

3.5.10. Enzyme Linked Immunosorbant Assay (ELISA). Either Tn- or TF-

BSA was coated on Immulon® Microtiter™ 4 HBX 96 well plates using 3µg/mL in carbonate buffer (pH 9.2) and then the plates were incubated for 18 h at 4 °C. Plates were washed three times with 200 µL of washing buffer (1XPBS buffer with 0.05%

Tween® 20) and blocked with 200 µL of 3% BSA for 1 h. Serum from mice was initially diluted at 1:100 and then serially half-log10 diluted, put into wells and incubated for 2 h at 37 °C for a final volume of 100 µL in each well. After incubation for 2 h, the plates were washed three times with 200 µL of washing buffer. Alkaline phosphatase linked secondary antibodies (Anti-IgM and Anti-IgG) was used to detect primary antibodies bound to either Tn- or TF-BSA. The procedure for the use of secondary anti-

IgM (Southern Biotech) were diluted (1:1000) and 100 µL were placed in wells corresponding for IgM detection and incubated for 1 h at 37°C. The procedure for the use of secondary anti-IgG antibodies (light chain, Jackson Immunoresearch) were diluted

(1:5000) and 100 µL were placed into wells corresponding to light chain IgG detection and incubated for 1 h at 37°C. The plates were washed three times with 200 µL of washing buffer and p-Nitrophenyl Phosphate (PNPP) (1 mg/mL) in diethanolamine buffer (pH 9.8) was added at a 100 µL per well and incubated for 30 min and optical density was read at 405 nm using BioTek PowerWave HT Microplate

Spectrophotometer. All assays were performed in triplicate. Titers were determined by regression analysis with dilutions plotted against absorbance. The titer cutoff value was 109

set at 0.2 for titer determination, which is two times the value from control mice.

Statistical analysis from ELISAs for experimental groups were compared with the controls using paired t test using GraphPad Prism 6.

3.5.11. MGL2 binding assay. Mouse recombinant MGL2 (R&D systems) 2.5

µg/mL was used to coat Immulon® Microtiter™ 4 HBX 96 well plates in 1XDPBS buffer (with CaCl2/MgCl2 ) pH 7.2 for 18 h at 4 ºC. The plates were then washed with

200 µL of 1XDPBS washing buffer (with CaCl2/MgCl2 and 0.05% Tween 20) three times. PS A1-biotin and respective biotinylated conjugates (4a-4c) were serially diluted from 40-0.625 µg/mL and incubated for 2 h at 37 ºC in 1XDPBS with CaCl2/MgCl2.

Plates were then washed with 200 µL of 1XDPBS buffer three times. A strepavidin- alkaline phosphatase (Sigma Aldrich) was diluted (1:1000) and 100 µL was added to each well and incubated for 1 h at 37 ºC. The plates were washed three times with 200

µL of 1X DPBS washing buffer and PNPP (1 mg/mL) in diethanolamine buffer (pH 9.8) was added at a 100 µL per well and incubated for 30 min and optical density was read at

405 nm. Experiments were performed in triplicate and data are illustrated as mean ± s.e.m.

Percent inhibition by Tn-BSA following the same procedure noted above was then conducted, however, 10 µg/mL was co-incubated with 4a-4c before binding competition to MGL2 was attempted. Percent inhibition was calculated using equation:

[(O.D of 4a-4c binding to MGL2) - (O.D. of co-incubation of 4a-c with Tn-BSA)/(O.D. of 4a-4c binding to MGL2)]x100.

3.5.12. Flow Cytometry. MCF-7 and OVCAR-5 (obtained from Henry Fold

Health Systems mycoplasma free) were cultured in 10% FBS RPMI 1640. 1.0 x 106 cells 110

of each cell line was incubated at 4 ºC for 1 h in the dark with 1:50 dilution of the following separate anti-sera: 1X PBS control, 1, 4a-4c. The cells were washed three times in 250 µL of FACs buffer (2% FBS in 1XPBS, 0.001 % sodium azide) by centrifuging at 1000 rpm. 100 µL Anti-IgG Alexa Fluor® 488 (1:50 dilution) was added to the cells and incubated at 4 oC in the dark for 1 h followed by three washes with 250

µL of FACS staining buffer. The cells were fixed with freshly prepared 1% paraformaldehyde and analyzed using BD Biosciences FACSCalibur at the University of

Toledo Core Flow Cytometry Facility. FlowJo analysis software was used to process flow cytometry data.

3.5.13. Complement Dependent Cytotoxicity Assay

MCF-7 cells (1.0x104) and OVCAR-5 cells (1.0x104) were seeded in 96 wells plates and incubated overnight in a 5% CO2 incubator at 37ºC. The plates were washed with 2%

BSA in DPBS and 100 µL of experimental anti-serum in 1:20 dilution of (1, 4a-4c, and

PBS control) was incubated for 1h. The experimental wells were washed and incubated with 10% rabbit complement (Pel-Freez) for 1 h at 37ºC. The control values of the LDH assay kit (Roche) was determined from spontaneous LDH release (low control) and 1%

Triton X-100 (high control) and incubated for 1 h at 37ºC, simultaneously with the experimental values. 50 µL of cell supernatant was transferred to a new 96 well plate containing 50 µL of DPBS. According to manufactures protocols 100 µL of the colorimetric LDH detection reagent was added to each well and the O.D was read at 490 nm. The percentage cellular cytotoxicity was calculated by the following equation: Cell cytotoxicity % = (experimental values – low control values) / (high control values – low control values) x 100. 111

3.5.14. Cytokine Profile. For the primary method of in vitro splenic cytokine release assay, the immunized spleens were removed, cultured, and stimulated for 72 h in the presence of PS A1, Tn-PS A1, TF-PS A1, and Tn-TF-PS A constructs. Mice were injected on days 0, 21, 35 and spleens were removed 10 days later. The spleens were harvested 10 days later to promote higher Th1 and Th17 immune response because these cells require longer to mature. The spleens were homogenized into single cell suspensions and strained over a 70 µm cell strainer into a 50 mL tube. The cells were centrifuged at 1000 rpm for 10 minutes. The cell pellet was treated with red blood cell lysis buffer (3 mL) for 3-4 min then diluted with 35 mL of cell culture media and centrifuged at 1000 rpm for 10 min to pellet the cells and buffer removal. The cells were strained over a 70 µm strainer again to remove dead clumps of cells. The spleen cells were counted and diluted to 1.0x106/ml and were seeded into 96 well plates. For stimulation of the splenocytes (mixture T cells, B cells, and APCs), 20 µg of TACA-PS

A1 conjugates and 20 µg of PS A1 were added to individual sections of the well plates.

The details of the ELISA experiments to detect cytokines were very similar to what was described previously above but a sandwich ELISA was used. This method used respective rabbit anti-mouse cytokine capture antibodies (IL-2, IL-4, IL-6, IL-10, IL-12,

IL-17, IL-22, INF-γ) (Peprotech) and were diluted to a concentration of 1.0 µg/mL and

100 µL was added to each ELISA well and to incubate overnight. The plates were washed four times in 300 µL of washing buffer (PBS with 0.1% Tween 20) and blocked with 300 µL of blocking buffer (PBS with 1% BSA), followed by washing. Then 100 µL experimental cell culture supernatant (TACA-PS A1, PS A1, and unstimulated splenic cells for baseline cytokine expression) were incubated for 2 h at 37 °C and washed. 100 112

µL of biotin conjugated detection antibodies (Table 3.1) were added to each well and incubated at room temperature for 2 h, followed by washing. Again, enzyme

(Strepavidin-HRP or Avidin-HRP) linked secondary antibodies specific for respective cytokines were used at 100 µL per well and incubated for 30 min, followed by washing.

Colorimetric detection of Strepavidin-HRP secondary antibodies used 100 µL of TMB substrate solution for 20 min and optical density were read at 450 nm with a wavelength correction at 640 nm.

Two other cytokines were investigated, (IL-23 and TGF-β (EBIOSCIENCE).

The capture antibodies anti-mouse IL-23 (250X) and anti-mouse TGF-β (250X) were diluted in 1X PBS and coated on well plates overnight at 4 °C. The sandwich ELISA assays described above was followed for these cytokine detection assays.

Table 3.1. Cytokine kits used from Peprotech

Cytokines IL-2 IL-4 IL-6 IL-10 IL-12 IL-17a IL-22 INF-γ (Peprotech) Capture mAb 1.0µg/m 1.0µg/ 1.0µg/ 1.0µg/ 1.0µg/ 1.0µg/ 1.0µg/ 1.0µg/ Concentration L mL mL mL mL mL mL mL

Detection mAb 0.1 0.125µ 0.1 0.125 0.25 0.25 1.0 0.25 concentration µg/mL g/mL µg/mL µg/mL µg/mL µg/mL µg/mL µg/mL

Standard 4000 3000 4.0 2000 700 1.0 3.0 1.0 Diluents (serially pg/mL pg/mL pg/mL pg/mL pg/mL ng/mL ng/mL ng/mL diluted)

113

3.6. References

(1) Kalka-Moll, W. M.; Tzianabos, A. O.; Bryant, P. W.; Niemeyer, M.;

Ploegh, H. L.; Kasper, D. L. J. Immunol. 2002, 169, 6149.

(2) Cobb, B. A.; Wang, Q.; Tzianabos, A. O.; Kasper, D. L. Cell 2004, 117,

677.

(3) Troy, E. B.; Kasper, D. L. Front. Biosci. 2010, 15, 25.

(4) Bloem, K.; García-Vallejo, J. J.; Vuist, I. M.; Cobb, B. A.; van Vliet, S. J.;

van Kooyk, Y. Front. Immunol. 2013, 4, 1.

(5) Wang, Q.; McLoughlin, R. M.; Cobb, B. A.; Charrel-Dennis, M.; Zaleski,

K. J.; Golenbock, D.; Tzianabos, A. O.; Kasper, D. L. J. Exp. Med. 2006,

203, 2853.

(6) De Silva, R. A.; Wang, Q.; Chidley, T.; Appulage, D. K.; Andreana, P. R.

J. Am. Chem. Soc. 2009, 131, 9622.

(7) De Silva, R. A.; Appulage, D. K.; Pietraszkiewicz, H.; Bobbitt, K. R.;

Media, J.; Shaw, J.; Valeriote, F. A.; Andreana, P. R. Cancer Immunol.

Immunother. 2012, 61, 581.

(8) Cho, J. S.; Pietras, E. M.; Garcia, N. C.; Ramos, R. I.; Farzam, D. M.;

Monroe, H. R.; Magorien, J. E.; Blauvelt, A.; Kolls, J. K.; Cheung, A. L.;

Cheng, G.; Modlin, R. L.; Miller, L. S. J. Clin. Invest. 2010, 120, 1762.

114

(9) Monin, L.; Griffiths, K. L.; Slight, S.; Lin, Y.; Rangel-Moreno, J.; Khader,

S. A. Mucosal Immunol. 2015, 8, 1099.

(10) Martin-Orozco, N.; Muranski, P.; Chung, Y.; Yang, X. O.; Yamazaki, T.;

Lu, S.; Hwu, P.; Restifo, N. P.; Overwijk, W. W.; Dong, C. Immunity

2009, 31, 787.

(11) Hamaï, A.; Pignon, P.; Raimbaud, I.; Duperrier-Amouriaux, K.; Senellart,

H.; Hiret, S.; Douillard, J.-Y.; Bennouna, J.; Ayyoub, M.; Valmori, D.

Cancer Res. 2012, 72, 1059.

(12) Wan, Y. Y.; Flavell, R. A. Immunol. Rev. 2007, 220, 199.

(13) Yang, X. O.; Nurieva, R.; Martinez, G. J.; Kang, H. S.; Chung, Y.; Pappu,

B. P.; Shah, B.; Chang, S. H.; Schluns, K. S.; Watowich, S. S.; Feng, X.-

H.; Jetten, A. M.; Dong, C. Immunity 2008, 29, 44.

(14) Geijtenbeek, T. B. H.; Gringhuis, S. I. Nat. Rev. Immunol. 2009, 9, 465.

(15) Denda-Nagai, K.; Aida, S.; Saba, K.; Suzuki, K.; Moriyama, S.; Oo-

puthinan, S.; Tsuiji, M.; Morikawa, A.; Kumamoto, Y.; Sugiura, D.;

Kudo, A.; Akimoto, Y.; Kawakami, H.; Bovin, N. V.; Irimura, T. J. Biol.

Chem. 2010, 285, 19193.

(16) Kang, Y.-S.; Kim, J. Y.; Bruening, S. A.; Pack, M.; Charalambous, A.;

Pritsker, A.; Moran, T. M.; Loeffler, J. M.; Steinman, R. M.; Park, C. G.

Proc. Natl. Acad. Sci. USA 2004, 101, 215.

(17) Glinsky, V. V.; Glinsky, G. V.; Rittenhouse-Olson, K.; Huflejt, M. E.;

Glinskii, O. V.; Deutscher, S. L.; Quinn, T. P. Cancer Res. 2001, 61, 4851.

115

(18) Streng-Ouwehand, I.; Unger, W. W. J.; van Kooyk, Y. Cancers (Basel)

2011, 3, 3169.

(19) Tsuiji, M.; Fujimori, M.; Ohashi, Y.; Higashi, N.; Onami, T. M.; Hedrick,

S. M.; Irimura, T. J. Biol. Chem. 2002, 277, 28892.

(20) Singh, S. K.; Streng-Ouwehand, I.; Litjens, M.; Kalay, H.; Saeland, E.;

van Kooyk, Y. Int. J. Cancer 2011, 128, 1371.

(21) van Vliet, S. J.; Saeland, E.; van Kooyk, Y. Trends Immunol. 2008, 29, 83.

(22) Slovin, S. F.; Ragupathi, G.; Musselli, C.; Olkiewicz, K.; Verbel, D.;

Kuduk, S. D.; Schwarz, J. B.; Sames, D.; Danishefsky, S.; Livingston, P.

O.; Scher, H. I. J. Clin. Oncol. 2003, 21, 4292.

(23) Dziadek, S.; Kowalczyk, D.; Kunz, H. Angew. Chem. Int. Ed. 2005, 44,

7624.

(24) Lakshminarayanan, V.; Thompson, P.; Wolfert, M. A.; Buskas, T.;

Bradley, J. M.; Pathangey, L. B.; Madsen, C. S.; Cohen, P. A.; Gendler, S.

J.; Boons, G. J. Proc. Natl. Acad. Sci. USA 2012, 109, 261.

(25) Yin, Z.; Chowdhury, S.; McKay, C.; Baniel, C.; Wright, W. S.; Bentley,

P.; Kaczanowska, K.; Gildersleeve, J. C.; Finn, M. G.; BenMohamed, L.;

Huang, X. ACS Chem. Biol. 2015, 10, 2364.

(26) Sarkar, S.; Lombardo, S. A.; Herner, D. N.; Talan, R. S.; Wall, K. A.;

Sucheck, S. J. J. Am. Chem. Soc. 2010, 132, 17236.

(27) Singh, S. K.; Streng-Ouwehand, I.; Litjens, M.; Weelij, D. R.; García-

Vallejo, J. J.; van Vliet, S. J.; Saeland, E.; van Kooyk, Y. Mol. Immunol.

2009, 46, 1240. 116

(28) Freire, T.; Lo-Man, R.; Bay, S.; Leclerc, C. J. Biol. Chem. 2011, 286,

7797.

(29) Ragupathi, G.; Koide, F.; Sathyan, N.; Kagan, E.; Spassova, M.;

Bornmann, W.; Gregor, P.; Reis, C. A.; Clausen, H.; Danishefsky, S. J.;

Livingston, P. O. Cancer. Immunol. Immunother. 2003, 52, 608.

(30) Engering, A.; Geijtenbeek, T.; van Kooyk, Y. Trends Immunol. 2002, 23,

480.

(31) Lang, R.; Schoenen, H.; Desel, C. Immunobiology 2011, 216, 1184.

(32) Miyake, Y.; Toyonaga, K.; Mori, D.; Kakuta, S.; Hoshino, Y.; Oyamada,

A.; Yamada, H.; Ono, K.-i.; Suyama, M.; Iwakura, Y.; Yoshikai, Y.;

Yamasaki, S. Immunity 2013, 38, 1050.

(33) Gaddis, D. E.; Maynard, C. L.; Weaver, C. T.; Michalek, S. M.; Katz, J. J.

Leukoc. Biol. 2013, 93, 21.

(34) Pasare, C.; Medzhitov, R. Science 2003, 299, 1033.

(35) Hall, J. A.; Bouladoux, N.; Sun, C. M.; Wohlfert, E. A.; Blank, R. B.; Zhu,

Q.; Grigg, M. E.; Berzofsky, J. A.; Belkaid, Y. Immunity 2008, 29, 637.

(36) Singh, S. K.; Streng-Ouwehand, I.; Litjens, M.; Kalay, H.; Saeland, E.;

van Kooyk, Y. Int. J. Cancer 2011, 128, 1371.

(37) Wall, S. T.; Saha, K.; Ashton, R. S.; Kam, K. R.; Schaffer, D. V.; Healy,

K. E. Bioconjugate Chem. 2008, 19, 806.

(38) Bourgault, J. P.; Trabbic, K. R.; Shi, M.; Andreana, P. R. Org. Biomol.

Chem. 2014, 12, 1699.

117

(39) Irazoqui, F. J.; Sendra, V. G.; Lardone, R. D.; Nores, G. A. Immunol. Cell.

Biol. 2005, 83, 405.

(40) Mitsdoerffer, M.; Lee, Y.; Jäger, A.; Kim, H.-J.; Korn, T.; Kolls, J. K.;

Cantor, H.; Bettelli, E.; Kuchroo, V. K. Proc. Natl. Acad. Sci. U.S.A.

2010, 107, 14292.

(41) van Vliet, S. J.; Bay, S.; Vuist, I. M.; Kalay, H.; Garcia-Vallejo, J. J.;

Leclerc, C.; van Kooyk, Y. J. Leuk. Biol. 2013, 94, 315.

(42) Napoletano, C.; Zizzari, I. G.; Rughetti, A.; Rahimi, H.; Irimura, T.;

Clausen, H.; Wandall, H. H.; Belleudi, F.; Bellati, F.; Pierelli, L.; Frati, L.;

Nuti, M. Eur. J. Immunol. 2012, 42, 936.

118

Chapter 4

Immunological Evaluation of Globo H-PS A1 Conjugates

4.1. Introduction

Globo H is a unique ganglioside based hexasaccharide tumor associated carbohydrate antigen (TACA) and is anchored in tumor cells through a lipid ceramide. It is overexpressed in many tumor cells such as breast, ovarian, prostate, etc., and it was first identified on the MCF-7 cell line in 1984.1 Its hexasaccharide nature is unique and has recently been focused on numerous clinical trials with Globo H conjugated to KLH or

CRM 197 but to date no TACA based vaccine has been granted approval. In clinical trials the Globo H-KLH vaccine with adjuvant QS-21 had minimal IgG production in humans; however, high quantities of IgM resulted in complement dependent cytotoxicity.2 Other studies anchored by OBI Pharma are investigating Globo H conjugated to CRM197 using glycolipid C34 adjuvant.3,4 The C34 adjuvant is a α- galactosylceramide derivative specificity designed for interaction with CD1d a protein which is classified as an MHC-I-like molecule and interacts with NKT cells to polarize toward Th1 immunity.4 Additionally, a fully synthetic Globo H-MPLA conjugate produced a T cell dependent immune response and produced high IgG antibody responses.5 119

Globo H remains an essential carbohydrate target not only because of the expression on breast cancers,6 but also its contribution to angiogenesis7 and expression on cancer stem cells (CSC) leading to tumor initiation and progression.8 Globo H has been shown to induce immunosuppression by shedding from the tumor and decreasing T and B cell populations by reducing Notch1 signaling.9 Therefore, targeting Globo H can be vital for the clearance of primary tumor cells and CSCs by halting tumor cell recurrence.

Additionally, Globo H shares a common trisaccharide core (Galα1-4Galβ1-4Glc) structure with GB3, which is also expressed on CSCs but not on normal stem cells.10 The mechanism for increased expression of gangliosides is facilitated by glycosyltransferases

A4GALT (GB3) and FUT1/FUT2 for Globo H. Therefore, not only would an effective vaccine be able to act as an angiogenesis inhibitor but also as a potent mediator of cytotoxicity by ADCC and CDC of CSC. Increasing the immunogenicity of TACAs is a common theme to clinically validate these targets, but the use of adjuvants remains essential to augment immune responses.

There are three main advantages to using PS A1 as an immune construct, 1) interactions with innate immune receptors (TLR2 and DC-SIGN), 2) mediating a T cell dependent response, and 3) polysaccharide and/or ZPS (scaffold) leading to higher degrees of interactions on the surface of antigen presenting cells, through a multivalent effect.11 These interactions include inherent binding to both TLR 2 and DC-SIGN which transforms a TACA-PS A1 construct into an exclusive self-adjuvanting system. When using PS A1 from B. fragilis, this zwitterionic polysaccharide (ZPS) can act as a bridge between the innate to adaptive immune responses. This transition first begins with cross linking to the surface of dendritic cells by TLR2 and DC-SIGN (other CLRS may be 120

involved due to promiscuity). This action stimulates both cytokine production and ZPS internalization via endocytosis. The ZPS is radically depolymerized by nitric oxide where 4-5 kDa fragments bind to MHC II and are presented on the surface to T cells.

These components on the surface of antigen presenting cells are c-type lectins, which are integral in dealing with carbohydrate-based vaccines. The loading of TACAs on a polymer can alter the immune response by incorporating known carbohydrate specificity to influence CLR binding and uptake.

CLRs are an essential part of carbohydrate based immunity (especially with ZPS) by promoting targeted carbohydrate based immunogens. The entire premise of using

ZPS is to create an entirely carbohydrate based vaccine for the generation of unique carbohydrate specific antibodies. However, focusing on certain carbohydrate antigens can modulate the immune responses by promoting proinflammatory cytokines such as IL-

6 and increased antigen uptake. These effects were noted when a unimolecular bivalent

Tn-TF-PS A1 construct was able to increase the immune response towards the TF antigen, when compared to the monovalent TF-PS A1 alone (Chapter 3). The rationale for using Globo H-PS A1 is to not only investigate the unique carbohydrate antigen but also to investigate the effects of having a terminal fucose of Globo H on the immune response. The significance of a terminal fucose; is in promoting interactions with other

CLRs (dectin-1, DC-sign, langerin, mannose receptor) to further modulate the immune response generated from PS A1. This can effectively produce a vaccine that has targeted function towards dendritic cells due to the multivalent binding effects.

4.2. Results and Discussion

PS A1 was oxidized using sodium periodate and three separate conjugates were 121

semi-synthetically prepared through an oxime link. The formation of the oxime linkage is a key component in the generation of PS A1 conjugates because they provide greater hydrolytic stability than hydrazones, hydrazides, and imines due to the electronegativity of the oxygen compared to either nitrogen or carbon.12 This added stability is critical in ensuring the TACA-ONH2 is tethered to PS A1 after being subjected to acidic lysosomes

Figure 4.1. Synthesis of GH-PS A1 conjugates 1A-C. A) Globo H-PS A1 B) Bivalent Tn-GH-PS A1 and C) GB3-PS A1

122

en route for presentation to T cells by MHC II. Globo H-PS A1 (GH-PS A1) and a unimolecular bivalent construct Tn-GH-PS A1 was injected into C57/BL6 mice and the immunological evaluation was assessed with and without Sigma Aldrich Adjuvant (SAS) or TiterMax Gold (TMG). SAS is a mixture including monophosphloryl lipid A

(MPLA), a TLR 4 agonist, and synthetic trehalose dicorynomycolate (STDCM), which binds to C-type lectins, minicle and dectin-2 which increases production of proinflammatory cytokines.13,14 TMG is a potent oil in water emulsion which provides slow release of antigens and its main component CRL-8300, is composed of conjugated copolymer of polyethylene oxide and polypropylene oxide.15 Figure 4.2 displays the selective anti-Globo-H immune response generated from a series of Globo-H based PS

A1 incorporated into constructs with different adjuvants. Examination of the GH-PS A1 constructs revealed exceptional IgG and IgM specificity towards Globo-H-BSA (see

Scheme D.1.). The GH-PS A1 (SAS) had exceptional anti-IgG and anti-IgM binding with a titer value of 22,000 and 7,300, respectively. Additionally, GH-PS A1 (TMG) showed potent anti-IgG titers with a titer value of 9,700. The difference of the administration of adjuvant between SAS compared to TMG had a significant three-fold effect on the amount of antibody towards GH-BSA.

123

When comparing both adjuvants while investigating the unimolecular bivalent construct Tn-GH-PS A1, an interesting phenomenon occurred. The Tn-GH-PS A1 with

TMG had an increased anti-IgG titer of 15,700 compared to 9,700 from GH-PS A1 with

TMG. This result suggests the presence of Tn alone can augment the selectivity and specificity of the anti-IgG immune response towards GH-BSA. However, when Tn-GH-

PS A1 was administered with SAS, there was an enormous reduction of both anti-IgG and anti-IgM. This result can potentially explained by research performed by van Kooyk

Figure 4.2. The IgG and IgM immune response from Globo H conjugates. et al., where simultaneous activation of CLRs (DC-SIGN and DCIR) showed reduced activation and presentation to T cells by APCs.16

Another interesting result that occurred through the immunological evaluation of

124

GH-PS A1 constructs, was the higher generated immune response against GB3-BSA (see

Scheme D.2) (Figure 4.3). The immune response was notably higher for all of the anti- serum generated against GH-PS A1 to GB3-BSA. Since GB3 is a part of the core structure of Globo H, it is plausible Globo H will become fragmented due to radical nitric oxide degradation. This ultimately leads to a fragmented portion of Globo H presented to

T cells to assist in antibody generation. For comparison, this resulted in close to a two- fold increase in the recognition of GB3-BSA over GH-BSA. Figure 4.4 shows the antibody response generated against GB3-PS A1 (TMG) to compare to GH-PS A1.

Figure 4.3. Cross reactivity of IgG and IgM antibodies from Globo H-PS A1 conjugates to GB3-BSA.

125

While there is a substantial titer for total antibody response (kappa), the antibody response generated as IgG and IgM antibodies is substantially decreased in comparison to anti-GH-PS A1 constructs (Figure 4.3). These results potentially validate the immune modulating properties of a terminal fucose (Globo H) compared to terminal galactose

(GB3). A comparison towards Tn-PS A1 (terminal GalNAc) and TF-PS A1 (terminal

Gal) (Chapter 3) shows the addition of galactose containing moieties seen with the TF antigen dampening the immune response by increased IL-10 values. However, further

Figure 4.4. The immune response generated from GB3-PS A1 and recognition of GB3-BSA work needs to be performed to confirm the modulation of immune responses by separate carbohydrates.

A particular concern of creating an immune response with a construct containing both fucose and N-acetyl galactosamine is there could be immense cross reactivity with blood group A and blood group B. Therefore, anti-serum from the GH-PS A1 based constructs were screened for binding to both of the blood groups in ELISA (Figure 4.5). 126

Both the GH-PS A1 (TMG and SAS) had relatively low anti-IgG and anti-IgM binding to blood group A (BGA) and blood group B (BGB) with optical density value less than 0.2.

Additionally, the Tn-GH-PS A1 (TMG and SAS) were analogous to GH-SAS where there was minimal anti-IgG and anti-IgM cross

Figure 4.5. Cross reactivity of anti-serum (1:100 dilution) of GH-PS A1 constructs to blood group A and blood group B.

127

reactivity towards BGA and BGB. This result suggests that there is not a concern with large immune responses towards Globo H and the potential of cytotoxicity of red blood cells. Flow cytometry was then used to determine the IgG response binding to human tumor cell lines MCF-7 (breast) and OVCAR-5 (ovarian) (Figure 4.6). The anti-serum from the GH-PS A1 constructs and respective adjuvant formulations were individually

Figure 4.6. Flow cytometry with anti-serum from PS A1, Globo H-PS A1, and Tn-PS A1 with secondary Alexa Fluor® 488 anti-IgG using human tumor cell lines. (A) MCF-7 human breast tumor cell line. (B) OVCAR-5 human ovarian tumor cell line.

128

used to determine binding to cancer cells. Analogous to Figure 4.2 and Figure 4.3, the anti-serum generated showed good binding to MCF-7 and OVCAR-5. When specifically examining the binding of the TMG series (GH-PS A1 and Tn-GH-PS A1), both anti sera showed exceptional binding to MCF-7 with 84 % positive shift in fluorescent intensity

(GH-PS A1 TMG) and 91 % positive shift in fluorescent intensity (Tn-GH-PS A1 TMG) compared to the controls of PBS (8%), PS A1 (10%), and auto-fluorescence of the cell line alone. Tn-GH-PS A1 TMG had the greatest fluorescent intensity when binding to

OVCAR-5, which makes an interesting discovery compared to Tn-GH-PS A1 SAS. The difference in binding may be contributed to over stimulation of CLRs with the adjuvant of SAS leading to less effective antibody binding responses.

Similar results were observed with the TMG series binding with OVCAR-5 with a 95 % positive (Tn-GH-PS A1 TMG) and 84 % positive with (GH-PS A1 TMG) compared to the controls of PBS (5 %) and PS A1 (4 %). When examining the SAS series, as expected, the Tn-GH-PS A1 SAS showed the lowest anti-IgG binding with 71

% binding to MCF-7 and 62 % OVCAR-5. However, GH-PS A1 SAS showed the highest binding to MCF-7 cell line with 94 % positive fluorescent anti-IgG binding events and 81 % binding to OVCAR-5.

For determining the LDH assay (Figure 4.7A-B), the anti-sera that demonstrated the highest binding in ELISA and flow cytometry were selected for their potential to mediate complement dependent cytotoxicity. The two that were investigated were GH-PS

A1 SAS and Tn-GH-PS A1 TMG. Tn-GH-PS A1 TMG demonstrated superior cytotoxicity towards both MCF-7 and OVCAR-5 tumor cells with 79 % and 58%,

129

Figure 4.7. Antibody mediated CDC with anti-serum from PS A1, Globo H-PS A1, and Tn-PS A1 plus rabbit complement. (A) MCF-7 human breast tumor cell line. (B) OVCAR-5 human ovarian tumor cell line. respectively. Additionally, these results are significant compared to the cytotoxicity from

PS A1 serum towards MCF-7 (40 %) and OVCAR-5 (17%).

The combined effects from having multiple TACAs on a unimolecular bivalent construct lead to greater binding to tumor cells. The Globo H-PS A1 SAS also had significant binding to MCF-7 and OVCAR-5 with 63 % and 49 %, respectively.

Collectively, both the Tn-GH-PS A1 TMG and GH-PS A1 demonstrated excellent cytotoxicity between cell lines.

4.3. Conclusions

The synthesis of Globo H-PS A1 and Tn-GH-PS A1 and subsequent immunizations has generated high immune responses towards Globo H which resulted in tumor cell binding and high cytotoxicity of both MCF-7 and OVCAR-5. The advantages of using the ZPS platform are related to the entirely carbohydrate vaccine construct with entirely carbohydrate specificity and targeted uptake by dendritic cells through CLRs. To

130

our knowledge, there are no other anti-cancer based vaccines that satisfy the aforementioned categories.

The results indicated the use of adjuvants play a major effect in the immunogenicity in both Globo H-PS A1 and Tn-Globo H-PS A1. The use of SAS had a significant impact on the anti-IgG response from GH-PS A1 with a titer of 22,000 compared to 9,700 from GH-PS A1. This suggests the proinflammatory cytokines generated from MPLA and STDCM assist in producing a larger immunological titer.

However, the same proinflammatory environment from SAS did not produce the same desired results with Tn-GH-PS A1. In fact, using SAS with Tn-GH-PS A1 dampened the

IgG antibody response nearly tenfold in comparison to the response generated from Tn-

GH-PS A1 TMG. The difference in binding may be contributed to the over stimulation of CLRs with the adjuvant of SAS leading to less effective antibody binding responses.

The decreased immunogenicity of Tn-GH-PSA1 SAS reveals that much more research needed to divulge the mechanism of entirely carbohydrate immunogens. However, an interesting phenomena has been demonstrated when a ligand interacted simultaneously with both DC-SIGN and dendritic cell immunoreceptor (DCIR), showing reduced activation and presentation to T cells.16 However, interactions with TLRs and multiple interactions with CLRs was not investigated. It can be concluded that multiple interactions from C-type lectins and multiple interactions of TLRs can ultimately affect T cell presentation and subsequent immune response. Further studies in due time should be performed to assist in the development of vaccines targeting DCs.

131

Analysis through flow cytometry of the GH-based PS A1 constructs revealed high recognition of both tumor cell lines MCF-7 and OVCAR-5 due to the cell lines expression of both Tn and Globo H.

4.4. Experimental

4.4.1. GH-PS A1 (1a). 1.0 mg of PS A1 was oxidized using 1 mM sodium periodate in sodium acetate buffer pH 5.0 for 90 min in the dark. Excess sodium periodate was quenched with KCl and desalted using centrifugal filter (10 kDa MWCO).

1.3 mg of Globo H-ONH2 reacted with oxidized PS A1 for 16 h. The resulting reaction was desalted using centrifugal filter (10 kDa MWCO). (Note: Dialysis was not used because samples formed irreversible aggregates after lyophilizing). 1H NMR was used to determine oxime formation.

4.4.2. Bivalent Tn-GH-PS A1 (1b). 1.0 mg of PS A1 was oxidized using conditions as described above. A 1:1 molar ratio of 1.1 mg of Globo H-ONH2 and 0.25 mg of Tn-ONH2 was reacted with freshly oxidized PS A1. Excess salts and by products were removed by centrifugal filtration. 1H NMR was used to determine two separate oxime formation.

4.4.3. GB3-PS A1 (1c). 1.0 mg of PS A1 was oxidized using conditions as described above. 1.2 mg of GB3-ONH2 was reacted with 1.0 mg of oxidized PS A1 for

16 h. The reaction was dialyzed and lyophilized. 1H NMR was used to determine oxime formation.

4.4.4. Immunizations. Individual GH, Tn-GH, or GB3-PS A1 constructs (20

µg) were mixed in a 1:1 ratio of 50 uL of TiterMax® Gold and injected into 7 wk old

C57BL/6 mice (Jackson Laboratory) (each construct was administered individually – not 132

mixed). Mice groups (n=5) were immunized by intraperitoneal injections (i.p.) on day 0,

14, 28, 42. Blood sera were obtained using a cardiac puncture technique on day 52.

Vaccinations with Sigma Adjuvant System. Individual GH and Tn-GH-PS A1 constructs (20 µg) were mixed in a 1:1 ratio of 100 µL of Sigma Adjuvant System

(Sigma-Aldrich) and injected into 7 wk old C57BL/6 mice (Jackson Laboratory) (each construct was administered individually not mixed). Mice groups (n=5) were immunized by intraperitoneal injections (i.p.) on day 0, 21, 42, per manufactures instructions. Blood sera were obtained using a cardiac puncture technique on day 52.

4.4.5. Enzyme Linked Immunosorbent Assay (ELISA). Either GH-BSA,GB3-

BSA, blood group A/or blood group B was coated on Immulon® Microtiter™ 4 HBX 96 well plates using 3µg/mL in carbonate buffer (pH 9.2) and then the plates were incubated for 18 h at 4 °C. ELISA procedures described previously were followed.

4.4.6. Synthesis of GH-Thio linker. 3.0 mg of Globo H-ONH2 was reacted with

1.0 mg of 3-oxopropyl ethanethioate (mercaptoaldehyde) (Chapter 3.4) for 24 h. The reaction was purified by Sephadex G-10 and deionized/distilled H2O as the eluent.

Fractions containing the GH linker were lyophilized.

1 H NMR (D2O,600MHz): δ = 7.56 (t, J=6.2 Hz, 1 H), 5.34 (d, J=3.7 Hz, 1 H), 5.12 (d,

J=4.4 Hz, 1 H), 4.75 - 4.83 (m, 8 H), 4.74 (br. s., 7 H), 4.64 - 4.70 (m, 11 H), 4.49 - 4.53

(m, 1 H), 4.40 - 4.46 (m, 3 H), 4.27 - 4.31 (m, 1 H), 4.10 - 4.15 (m, 3 H), 3.98 - 4.01 (m,

1 H), 3.90 - 3.93 (m, 2 H), 3.77 - 3.89 (m, 9 H), 3.71 - 3.76 (m, 7 H), 3.51 - 3.70 (m, 26

H), 3.49 - 3.50 (m, 1 H), 3.48 (t, J=2.0 Hz, 1 H), 2.95 - 3.03 (m, 2 H), 2.40 - 2.50 (m, 1

H), 2.24 - 2.29 (m, 3 H), 1.91 - 1.95 (m, 4 H), 1.07 - 1.13 ppm (m, 4 H).13C NMR

(D2O,151MHz): δ = 201.1, 201.0, 174.3, 104.0, 103.2, 102.0, 100.4, 99.2, 91.5, 78.2, 133

77.1, 76.3, 76.1, 75.4, 75.0, 74.6, 73.5, 72.1, 71.8, 71.7, 71.1, 70.8, 70.2, 70.1, 69.4, 69.2,

69.0, 68.4, 68.0, 67.8, 66.7, 60.9, 60.3, 59.4, 51.6, 30.0, 29.9, 29.2, 25.8, 25.4, 25.2, 22.2,

18.5, 15.3 ppm

4.4.7. Globo H-BSA. 2.0 mg of Globo H-thiol linker was deacetylated by a solution of concentrated K2CO 3 for 1.5 h. Zemplen conditions were not used because

Globo H is insoluble in MeOH. Globo H-thiol linker and reacted with freshly prepared

BSA-Malemide (procedure described previously) in PBS buffer with 1 mM EDTA pH

7.2. After 16 h at 4 C, the reaction was dialyzed 10,000 MWCO. Conjugation was confirmed with MALDI-TOF (92249.938) for a total conjugation of 15.5%.

4.4.8. GB3 Thiol linker. 2.0 mg of GB3-ONH2 reacted with 1.5 mg of mercaptoaldehyde for 18 h in sodium acetate buffer (pH 5.5) at room temperature and purified using Sephadex G-10 and deionized/distilled H2O as the eluent. Fractions containing the -GB3 linker were lyophilized. 2.5 mg of (7) was deacetylated using

Zemplen’s method with NaOMe in methanol followed by base neutralization with

DOWEX 50W x 8-100 ion exchange resin. The solution was then filtered and concentrated under reduced pressure

1 H NMR (D2O ,600MHz): δ = 7.52 (dd, J=8.5, 3.7 Hz, 1 H), 5.46 (t, J=4.2 Hz, 1 H), 4.96

(d, J=3.4 Hz, 1 H), 4.54 (dd, J=7.7, 6.5 Hz, 1 H), 4.37 (t, J=6.2 Hz, 1 H), 4.03 - 4.07 (m,

3 H), 3.90 - 3.97 (m, 4 H), 3.82 - 3.90 (m, 6 H), 3.80 (br. s., 2 H), 3.69 - 3.78 (m, 9 H),

3.57 - 3.62 (m, 1 H), 2.97 - 3.15 (m, 1 H), 2.42 (td, J=9.2, 4.6 Hz, 1 H), 2.38 (d, J=3.2

Hz, 3 H), 1.51 - 1.68 (m, 1 H), 0.90 ppm (dt, J=15.0, 7.5 Hz, 4 H). 13C NMR

(D2O,151MHz): δ = 200.9, 200.8, 181.5, 158.6, 103.1, 100.3, 99.2, 92.4, 78.1, 78.0, 77.3,

134

77.3, 75.4, 72.1, 71.6, 70.9, 70.8, 70.8, 70.3, 70.2, 69.1, 68.9, 68.5, 60.4, 60.3, 59.4, 42.2,

42.1, 30.8, 29.9, 29.9, 24.8, 24.6, 23.2, 10.8, 10.7 ppm

4.4.9. Flow Cytometry. MCF-7 and OVCAR-5 was cultured in 10% FBS RPMI

1640. 1.0 x 106 cells of each cell line was incubated at 4 C for 1 h in the dark with 1:50 dilution of the following separate anti-serums (PBS control, PS A1, Globo H-PS A1, Tn-

Globo H-PS A1). The cells were washed three times in FACs buffer (2 % FBS in PBS,

0.001 % sodium azide) by centrifuging at 1000 rpm. 100 µL Anti-IgG Alexa Fluor 488

(1:50 dilution) was added to the cells and incubated at 4 C in the dark for 1 h followed by three washes with FACS staining buffer. The cells were fixed with freshly prepared 1% paraformaldehyde and obtained using BD Biosciences FACsCaliber by the University of

Toledo Core flow cytometry facility. FlowJo FACs analysis was used to analyze the

Data.

135

4.5. References

(1) Bremer, E. G.; Levery, S. B.; Sonnino, S.; Ghidoni, R.; Canevari, S.;

Kannagi, R.; Hakomori, S. J. Biol. Chem 1984, 259, 14773.

(2) Gilewski, T.; Ragupathi, G.; Bhuta, S.; Williams, L. J.; Musselli, C.;

Zhang, X. F.; Bornmann, W. G.; Spassova, M.; Bencsath, K. P.; Panageas,

K. S.; Chin, J.; Hudis, C. A.; Norton, L.; Houghton, A. N.; Livingston, P.

O.; Danishefsky, S. J. Proc. Natl. Acad. Sci. USA 2001, 98, 3270.

(3) Huang, Y. L.; Hung, J. T.; Cheung, S. K. C.; Lee, H. Y.; Chu, K. C.; Li, S.

T.; Lin, Y. C.; Ren, C. T.; Cheng, T. J. R.; Hsu, T. L.; Yu, A. L.; Wu, C.

Y.; Wong, C. H. Proc. Natl. Acad. Sci. USA 2013, 110, 2517.

(4) Danishefsky, S. J.; Shue, Y.-K.; Chang, M. N.; Wong, C.-H. Acc. Chem.

Res. 2015, 48, 643.

(5) Zhou, Z.; Liao, G.; Mandal, S. S.; Suryawanshi, S.; Guo, Z. Chem. Sci.

2015, 6, 7112.

(6) Lou, Y.-W.; Wang, P.-Y.; Yeh, S.-C.; Chuang, P.-K.; Li, S.-T.; Wu, C.-

Y.; Khoo, K.-H.; Hsiao, M.; Hsu, T.-L.; Wong, C.-H. Proc. Natl. Acad.

Sci. U.S.A. 2014, 111, 2482.

136

(7) Cheng, J.-Y.; Wang, S.-H.; Lin, J.; Tsai, Y.-C.; Yu, J.; Wu, J.-C.; Hung,

J.-T.; Lin, J.-J.; Wu, Y.-Y.; Yeh, K.-T.; Yu, A. L. In Cancer Res. 2014;

Vol. 74, p 6856.

(8) Chen, K.; Huang, Y.-h.; Chen, J.-l. Nature 2013, 34, 732.

(9) Tsai, Y.-C. J. Cancer Sci. Ther. 2013, 05.

(10) Gupta, V.; Bhinge, K. N.; Hosain, S. B.; Xiong, K.; Gu, X.; Shi, R.; Ho,

M.-Y.; Khoo, K.-H.; Li, S.-C.; Li, Y.-T.; Ambudkar, S. V.; Jazwinski, S.

M.; Liu, Y.-Y. J. Biol. Chem. 2012, 287, 37195.

(11) Wall, S. T.; Saha, K.; Ashton, R. S.; Kam, K. R.; Schaffer, D. V.; Healy,

K. E. Bioconjugate Chem. 2008, 19, 806.

(12) Kalia, J.; Raines, R. T. Angew. Chem. Int. Ed. Engl. 2008, 47, 7523.

(13) Nishizawa, M.; Yamamoto, H.; Imagawa, H.; Barbier-Chassefière, V.;

Petit, E.; Azuma, I.; Papy-Garcia, D. J. Org. Chem. 2007, 72, 1627.

(14) Wilson, G. J.; Marakalala, M. J.; Hoving, J. C.; van Laarhoven, A.;

Drummond, R. A.; Kerscher, B.; Keeton, R.; van de Vosse, E.; Ottenhoff,

T. H. M.; Plantinga, T. S.; Alisjahbana, B.; Govender, D.; Besra, G. S.;

Netea, M. G.; Reid, D. M.; Willment, J. A.; Jacobs, M.; Yamasaki, S.; van

Crevel, R.; Brown, G. D. Cell Host Microbe 2015, 17, 252.

(15) Fox, C. B. Molecules 2009, 14, 3286.

(16) García-Vallejo, J. J.; Bloem, K.; Knippels, L. M. J.; Garssen, J.; van Vliet,

S. J.; van Kooyk, Y. Front. Immunol. 2015, 6, 87.

137

Chapter 5

Murine IgM monoclonal antibody generated from Tn- PS A1 with in vivo and in vitro activity

5.1. Introduction

Monoclonal antibodies (mAbs) are passive immunotherapeutics where B cells are isolated, fused with immortal non-antibody producing cells, and the resulting hybridoma cell lines are screened for selective antibody recognition.1 The fusion between splenocytes (diploid) and myeloma cells (tetraploid) results in a multi-nucleated cell which typically stabilizes as a hyperdiploid cell. To select favorable growth conditions for the hybridomas, myeloma cells such as Sp2/0-Ag142 which are deficient in hypoxanthine guanine phosphoribosyl transferase (HGPRT) were used for fusion. This deficiency is important for utilizing the salvage pathway for the denovo synthesis of

DNA bases of purines and pyrimidines. To remove unfused myeloma cells, aminopterin is used to block the de novo synthesis of DNA which causes myeloma death but does not affect the newly formed hybridomas because the HGPRT gene is acquired from the lymphocytes. The hybridomas are screened and selected based on antibody production and positive selections are further investigated.

138

In a cancer antigen prioritization study, 9 out of the top 75 cancer antigens were identified as being tumor-associated carbohydrate antigens (TACAs).3 Carbohydrate- based antigens have emerged as important targets for immunotherapies and TACA-based vaccines are currently being developed and evaluated in clinical trials. Recently, in a landmark decision, the FDA approved Unituxin®, as the first and only monoclonal antibody (mAb) targeting TACA GD2 (GalNAcβ1-4(Neu5AcA2-8Neu5AcA2-3) Galβ1-

4Glc)) for the treatment of high risk neuroblastoma in pediatric patients.4 Unituxin was developed from immunizations with the neuroblastoma cell line, LAN-1 which is known to contain a high density of GD2.4 However, before approval of TACA targeting

Unituxin®, only protein (non-carbohydrate)-based cancer antigens led to FDA approval of approximately 30 cancer mAbs, including Herceptin® (trastuzumab – Genentech,

Inc.), Rituxan® (rituximab – Biogen Idec Inc.), and Avastin® (bevacizumab –

Genentech, Inc.).5,6 Unlike proteins, carbohydrates suffer from T cell independent immune responses leading to weak immunogenicity. These limitations are off-set, in most cases, when TACAs are conjugated to carrier proteins, but can still have some disadvantages such as protein epitope suppression and immunogenic hydrocarbon linkers leading to non-specific antibody generation. Since there is often ambiguity in the effectiveness of TACA-conjugates, considering the failed Phase III clinical trials of

Theratope® (STn-KLH), novel immunogen strategies need to be investigated that specifically target sugar antigens in order to produce highly effective immunotherapies.

An alternative approach for overcoming the stalemate of TACA-protein conjugates is the entirely carbohydrate-based construct, that takes advantage of the

Thomsen-nouveau (Tn) antigen (D-GalpNAc) and an MHCII targeting zwitterionic 139

polysaccharide PS A1 from the capsule of commensal Bacteroides fragilis ATCC

25285/NCTC 9343.7 This Tn-PS A1 conjugate is derived from oxidized PS A1, and aminooxy Tn for spontaneous oxime formation and is stable under physiological conditions. It is now known that this unique immunogen stimulates anti-tumor responses through the induction of CD4+ T cells and through the production of cytokines; Il-10, IL-

17A, IL-4, and IL-2, encoding for a Th17 immunity.7 Furthermore, carbohydrate- specific/selective polyclonal IgG and IgM antibodies have been determined.8 This innovative design of entirely carbohydrate construct, Tn-PS A1 for augmenting the immune response towards TACAs may be a valuable tool in treating/preventing cancers.

It can also subsequently be used to produce monoclonal antibodies (mAb) due to the generation of selective and specific anti-carbohydrate immune responses.

Most of the FDA approved antibodies are IgG, but there have been ongoing clinical investigations for human IgM mAbs such as mAb216 and L612 HuMAb, which has demonstrated promising therapeutic results for melanoma and leukemia patients, respectively.9,10 IgM antibodies are gaining clinical relevance due inpart to their pentavalent nature and potent ability to initiate complement dependent cytotoxicity

(CDC).11 The pentavalent nature allows higher avidity due to increased sites of specificity. CDC activity is enhanced compared to IgG antibodies due to the initial complement protein C1q binding 1000-fold greater to IgM antibodies, which is involved in the initial C1-complex of the classical complement cascade. 12,13 Collectively through targeting TACAs by IgM antibodies, there is a therapeutic potential for binding IgM antibodies by increased avidity and complement activation.

140

There are numerous examples of mAbs (IgG and IgM) that can recognize TACAs and specifically the Tn antigen, but lack accurate specificity to glycosides. The Tn antigen is a valuable target due to its high expression on tumor cells (~80-90% of all tumors) and synthetic accessibility for biological conjugations.14,15 MAbs (IgG or IgM) as immunotherapeutics for TACAs can lack accurate specificity to glycosides due to promiscuity of the antibodies. In 2007, Gildersleeve and coworkers examined 27 carbohydrate specific mAbs to various TACAs using glycan array technology. The results of the selectivity and specificity towards TACAs were particularly concerning because over half of the TACA specific mAbs screened had cross reactivity to other carbohydrates and several of which did not bind to the target.16 The epitome of the lack of carbohydrate specificity can be observed with B1.1 (IgM), which is a commercial mAb specific for Tn that failed to bind Tn alone, but rather interacted with a cluster of Tn

(AcTn-Tn-Tn-Gly-Hex-BSA).16 For this example, the Tn cluster provided greater surface area for B1.1 to bind due to strong avidity of IgM antibodies. Therefore, devising alternative strategies to address the specific development of mAbs against TACAs is a challenging but critical aspect in ensuring carbohydrate specificity and selectivity.

An important criterion for the consideration in generating specific anti- carbohydrate mAbs is the requirement that immunogens must produce antibodies that are specific for sugar antigens without influence from peptide/hydrocarbon linkers. To avoid the cross reactivity between carbohydrate antigens, we sought to generate monoclonal antibodies from the TACA-ZPS, Tn-PS A1 to focus the immune response specifically onto Tn. Utilizing a synthetically prepared Tn-hydroxyl amine conjugated to oxidized galactofuranose, the formation of an oxime bond provides a unique entirely carbohydrate 141

immunogen without the need of bulky immunogenic linkers. The advantage of this model is to emphasize the immune response on O-linked carbohydrates by the linker-free oxime ligation and not on O-linked glycopeptides. For example, when examining mAbs towards glycopeptides, binding tends to be influenced by the original peptide sequence and is thus not glycan-specific.17,18 Traditional methods for mAb production have used naturally occurring TACAs (i.e., cancer cells and glycosylated proteins), and have led to many non-specific and commercially available mAbs such as B1.1 and Tn218 (IgM).

These two mAbs were generated from Ovine Submaxillary Mucin and screened for their ability to bind to Tn. The complication associated with glycoproteins is carrier-induced epitopic suppression that is due to the greater immunogenicity of the protein carrier, minimizing the response against TACAs. 19,20 Therefore most mAbs generated from glycopeptides/proteins/linkers will have a varying degree of sensitivity towards the peptide/linker portion, which is one of the reasons why we elected to use an entirely carbohydrate immunogen to assist in mAb development. We hypothesized that use of

Tn-PS A1 to generate mAbs would produce superior antibodies specifically for glycosides, which will lead to sufficient anti-tumor responses.

5.2 Results and Discussion

We utilized a novel technique to generate monoclonal antibodies from mice immunized with Tn-PS A1, an entirely carbohydrate immunogen. PS A1 was chosen as the immunogen because it is a zwitterionic polysaccharide that induces a T cell mediated immune response. The intended use behind this construct was to generate mAbs that

142

have exclusive carbohydrate binding donor and acceptor sites. After immunizing mice, the spleen cells were homogenized to single cell suspensions and fused with myeloma cell line Sp2/0-Ag14. The resulting hybridoma cell culture supernatants were screened to

Figure 5.1. The production and screening of mAbs. bind to Tn-BSA in order to parse out and isolate carbohydrate-specific antibodies (Figure

5.1). The hybridoma cell supernatant that demonstrated the best ability to bind to Tn-

BSA was chosen for scale up procedures for in vivo and in vitro studies. The optimal working concentration of Kt-IgM-8 (IgM) (Figure 5.2) was determined by serially

Figure 5.2. Titration of Kt-IgM-8 on ELISA to determine effective concentration.

143

diluting the antibody starting at a concentration of 60 to 0.01 µg/mL. Optimal binding in the titration of the antibody was observed at 0.3 µg/mL at an optical density of (O.D.) above 0.2. For an IgM antibody, binding at low concentrations rivals an IgG antibody but also indicates high avidity due to the pentavalent binding nature of the antibody. In order to compare the efficiency of KT-IgM-8, a commercial IgM antibody (Tn-218) was used to compare concentration for antibody recognition of the Tn antigen on ELISA

(Figure 5.3A). KT-IgM-8 and Tn-218 were screened in parallel at an initial concentration of 30 µg/mL and serially diluted to a final concentration of 0.23 µg/mL.

From Figure 3A, Tn-218 minimally recognized Tn-BSA, but Kt-IgM-8 demonstrated superior recognition, which suggests that the viability of binding Tn (GalNAc) is enhanced over what is commercially available. To expand upon the specificity of Kt-

Figure 5.3. (A) Kt-IgM-8 comparison study with Tn-218 (B). Carbohydrate specificity for Kt-IgM-8.

144

IgM-8, a small panel of TACA-related constructs were employed that displayed various

Tn antigens (α/β-Tn-Thr-BSA, α-Tn-BSA, α-TF-BSA, blood group A, and blood group

B), which were screened on ELISA (Figure 5.3A). KT-IgM-8 had no discernable preference between α or β containing-Thr-Tn glycosides and had decreased affinity for α-

TF-BSA. Additionally, KT-IgM-8 did not bind to ZPS PS A1 or BSA used to block the

ELISA plates (not shown). Fortunately, Kt-IgM-8 minimally recognized blood group A and B below the threshold value at 30 µg/mL (O.D ≤ 0.2) but did recognize them at increasing concentrations (O.D ≥ 0.2).

The next step in characterizing Kt-IgM-8 was to determine if the antibody could bind to whole cancer cells in flow cytometry. MCF-7 (Breast) and HCT-116 (Colon) were chosen due to both the presence of Tn and the fact that they represent two of the most common forms of cancers that contain Tn binding tumor cell lines. This is the first

Figure 5.4. Flow Cytometry of Kt-8-IgM binding to A) MCF-7 and B) HCT-116. step in determining how well an immunotherapeutic will stand up against in vivo models.

Anti- IgM Alexa Fluor®647 was used as the fluorescent secondary antibody to detect

IgM antibody binding to the primary antibody adhered on the cancer cell lines. Kt-IgM-8 145

shows the ability to bind to both tumor cell lines at 30 ug/mL (Figure 5.4) and showed a shift in fluorescence of 49 % in both cell lines compared to the cell lines alone.

In order to determine antibody function, a CDC assay coupled with Chromium-51 was used to detect cytotoxicity of MCF-7 cells. In Figure 5.5, Kt-IgM-8, Tn-PS A1 whole sera, (Tn-PS A1) IgG purified polyclonal sera, PS A1 sera, and control PBS sera

Figure 5.5. CDC activity of KT-IgM-8 on MCF-7 cells. Data are illustrated as mean ± s.e.m. ** P < 0.005, *** P < 0.0005; two tailed Student’s t-test. were used as a comparison to assess the potency of CDC activity. Both the Tn-PS A1 whole sera and Tn-PS A1 IgG purified polyclonal sera were used as Tn specific controls

146

that represented cytotoxicity from whole sera and purified IgG’s from the same sera. The

IgG purified polyclonal was essential in determining how effective IgGs from the immunization can be at initiating CDC without the help from IgMs. Surprisingly, Kt-

IgM-8 showed the greatest CDC activity at close to 30 % cytotoxicity, which showed more statistically significant activity than Tn-PS A1 sera (P <.005) and IgG purified Tn-

PS A1 sera P < .005) at 15 % and 8 %, respectively. Additionally, CDC activity was absent from anti-serums from PS A1 and PBS control mice. From an immunotherapeutic perspective, Kt-IgM-8 has the ability to initiate CDC greater than what can be seen from immunizations due to the overall concentration of antibody used. This leads to an important hypothesis that a Tn-specific IgM antibody can provide protection from in vivo tumor models. As a platform to examine human tumors in mice models, SCID (Severe

Combined Immunodeficient) mice provide an optimal host for xenografted human tumors. Since the SCID mice lack B and T lymphocytes, xenografted tumors are able to be implanted and grow without provoking an immune response against the tumors.

Consequently, the use of MCF-7 cells represents studying breast cancer without the need for using human models. The MCF-7 tumor growth in SCID mice was measured by tumor volume and effectiveness of the immunotherapeutic, and was assessed by the comparison of tumor volume in the control mice (PBS). Figure 5.6 represents four different treatments: PBS Control, KT-IgM-8 (Figure 5.6A), Tn-PS A1 whole sera

(Figure 5.6B), and Tn-PS A1 IgG polyclonal (Figure 5.6C). The humane endpoint of the experiment was determined when tumor volume approached 400 mm.3

147

Figure 5.6. Kt-IgM-8 displays tumor volume (mm3) reduction of MCF-7 Tumors in SCID mice for 39 days. A) KT-IgM-8 treatment of MCF-7 tumor growth in comparison to PBS control mice B) Anti-Tn-PS A1 whole serum in comparison to PBS mice over C) Anti-Tn-PS A1 polyclonal IgG in comparison to PBS mice D) Tumor volume at day 39. Data are illustrated as mean ± s.e.m. ** P < 0.005, *** P < 0.0005; two tailed Student’s t-test.

The control mice treated with PBS should have no protection against the tumors and would determine the efficiency of each antibody treatment. The Tn-PS A1 whole sera provided the greatest protection against tumor growth at 52 % difference (Figure 5.6D).

The Tn-PS A1 sera would be able to use both ADCC and CDC due to the mixture of both

IgM and IgG. Unfortunately, the IgG purified sera showed minimal protection against

148

tumor growth, which indicates that polyclonal IgGs provide minimal protection against tumor growth. However, Kt-IgM-8 demonstrated protection against tumors at 39 % difference (Figure 5.6D), which significantly defines the effectiveness of the treatment.

The data presented shows the effectiveness of IgM antibodies and their role in minimizing tumor growth.

5.3. Conclusion

The zwitterionic nature of PS A1 exploits a natural CD4+ immune response, which in theory would assist in a unique glycan-specific antibody development; a concept only seen in bacterial polysaccharides based mAbs. In order to confirm this assumption, we adapted this model to accommodate the Tn antigen. It is not surprising that the Tn antigen can make the PS A1 construct more immunogenic due to the interactions with c- type lectin receptor (MGL2), which facilitates increased antigen uptake in mice

(unpublished). The rationale behind using an entirely carbohydrate immunogen (Tn-PS

A1) was to focus the antibody recognition on glyosidic linkages in order to generate antibodies that have no specificity towards peptides/lipids. This was a challenging endeavor because antibody binding of glycans results in low Kd values and compensates by having multiple binding sites for higher avidity. This is a potential reason as to why

IgM antibodies can be favored over IgG due to the pentavalent binding compared to the bivalent bonding. Targeting the glycosides is essentially one of the most important features because Tn can be associated with different peptides and can ultimately affect antibody recognition. Therefore producing a mAb that is selective for glycosides can provide specificity for Tn by not having cross reactivity with peptides that are naturally occurring. 149

A particular concern when using an entirely carbohydrate construct is antibody cross reactivity with normally expressed carbohydrates. To examine the binding properties of the Kt-IgM-8, a small panel of Tn related antigens α/β-Tn-Thr-BSA, α-Tn-

BSA, α-TF-BSA, blood group A, and blood group B were screened by ELISA (Figure

5.3A). This panel represented different forms of the Tn-antigen, which are the primary biological expression GalNAc. However, α-Tn-Thr/Ser is distinctively exposed on the surface of tumor cells by a mutation in Cosmc, a chaperone protein responsible for the proper folding of the glycosylation machinery.21 GalNAc is also terminally expressed on normal blood group A (GalNAc(α1-3)[Fuc(α1-2)]Gal(β1-3)), but off the carbohydrate scaffold is an adjacent Fuc, which may impair antibody recognition of GalNac in this confirmation. Additionally, there are similarities between blood group A and B, where blood group B has Gal substituted for GalNAc. A challenge associated with targeting Tn or other TACAs is their ability to cross react with glycosides present on blood cells, which can promote harmful cytotoxicity. Furthermore, Kt-IgM-8 at 30 µg/mL does not preferentially differentiate between α/β-Tn-Thr-BSA, but α-TF-BSA does exhibit reduced binding due to the addition of the Gal to GalNAc. This suggests that the antibody can recognize α/β-Tn when it is exposed on the surface, but binding is reduced when Tn is masked with Gal. Kt-IgM-8 was determined to be very specific towards Tn, and thus should be able to recognize blood group A due to the terminal expression of GalNAc.

However, there was insignificant binding, which may have been impaired due to the branched structures of the blood group antigens. Therefore developing mAbs from Tn-

PS A1 produced a very specific antibody response to the Tn antigen and it clearly

150

exceeded the binding produced from other mAbs made from proteins such as Tn-218

(Figure 5.3A).

Over the past decade, IgM antibodies have proved to be effective in treating carcinomas. In order to support the shift of perception of IgM antibodies, a vaccine containing NeuGCGM3 gangliosides which form very small proteoliposomes induces

IgM protection in breast and melanoma carcinomas, and has reached phase III clinical trials.22 From our data it is reasonable to conclude that IgM antibodies, both monoclonal and polyclonal, may be more effective in killing tumor cells than IgG due to CDC activity. In Figure 5.5, Kt-IgM-8 showed a potent ability to initiate CDC compared to

Tn-PS A1 whole sera and Tn-PS A1 IgG polyclonal sera. In Figure 5.6, SCID mice with xenografted MCF-7 tumors were treated with anti-Tn-PS A1 sera, purified Tn-PS A1 IgG

(polyclonal), and Kt-IgM-8. The use of purified IgGs from Tn-PS A1 sera would mimic

IgG responses seen from vaccinations for the SCID model. The results (Figure 5.6D) showed both the anti-Tn-PS A1 sera and Kt-IgM-8 were effective in reducing the size of the MCF-7 tumors with antibody treatment alone. However, the Tn-PS A1 IgG pAbs showed no statistical difference in reducing the size of the tumors. These results are encouraging because the antibodies were naked, meaning they were not antibodies incorporated as drug conjugates and no additional cancer drugs, such as cyclophosphamide, were administered with treatment.

With the recent surge in FDA approvals of mAb as immunotherapies, the IgM antibody has been overlooked due to the superior nature of high affinity IgG’s binding peptide/protein moieties. However, IgG antibodies may not always be the preferred choice when it relates to glycosides, IgM antibodies have demonstrated potent CDC 151

responses to tumor cells. Kt-IgM-8 represents a biological tool that demonstrates in vitro complement activity and in vivo reduction of tumor growth. Additionally, only a handful of other Tn specific antibodies have been used for in vivo data, (MLS 128, GOD3-2C4, and KM3413) all of which are IgG antibodies.23-25 However, GOD3-2C4 was the only

Tn-specific antibody to be used in adoptive transfer of the antibody in SCID models, but activity was assessed by ADCC and CDC was not examined. Therefore, Kt-IgM-8 represents one of the first IgM antibodies specific for the Tn-antigen to be used in passive immunotherapies for cancer that utilizes CDC as the main source of cytotoxicity.

5.4. Experimental

5.4.1. Immunizations . Immunization of Tn-PS A1, PS A1, and PBS have been reported.26

5.4.2. Hybridoma Fusion Protocol. Mice spleens were obtained on day 60 in

DMEM media. The spleenocytes were obtained by gently homogenizing the spleens.

Cells were washed with serum free DMEM by centrifuging at 1000 rpm for 10 minutes and resuspending the final pellet in 30 ml of serum free DMEM. Simultaneously, Sp2/0-

Ag14 (ATCC CRL-1581) were cultured and washed with serum free DMEM serum free by centrifuging at 1000 rpm for 10 minutes and resuspending in 30 ml in serum free

DMEM. 2 x 107 myeloma cells and 1 x 108 viable splenocytes were added in a 50 ml centrifuge tube and were washed with serum free DMEM three times. Clona Cell-

HYPEG (1 ml) was added to the tube without stirring. Cells were stirred for a minute by gently shaking the tube. 4 ml of serum free DMEM media was added to the fusion mixture and stirred for 4 minutes. 10 ml of serum free DMEM was slowly added and incubated at 37 C for 15 minutes. 30 ml of 10% FCS-DMEM was added and washed 152

with 40 ml of DMEM and the supernatant was discarded. 10 ml of 20% FCS-DMEM resuspended the pellet and was transferred to a T-175 flask containing 20 ml of 20-

DMEM and was incubated for 24 hr in 5% CO2. Cells were centrifuged and resuspended with 10 mL of 20-DMEM and added to 90 ml of semi-solid methyl cellulose media

(ClonaCell Flex). The bottle was mixed by inverting and was aliquoted in 10 petri dishes and placed in a 5% CO2 incubator for 10-14 days. Single cell colonies were picked (5 µl) and placed in 96 well plates containing 10-DMEM in 200 µl. The cell supernatants were screened by ELISA with plates coated with Tn-BSA when sufficient antibody was produced.

5.4.3. IgM Purification. Purification of IgM antibodies followed this protocol.27

Cell culture supernatant was dialyzed against distilled water causing a precipitation of the

IgM antibody after 1 day at 4°C. The resulting precipitate was centrifuged to remove water. The precipitate was dissolved in 1x PBS buffer and was followed by ammonium sulfate precipitation by adding 17.1 g of ammonium sulfate forming a precipitate, which was concentrated and purified further with size exclusion chromatography (sephacryl S-

300). Fractions were individually checked and monitored at 280 nm. The resulting fractions containing IgM antibody were pooled, sterile filtered and stored at 4°C.

5.4.4. Complement Dependent Cytotoxicity. 2 x 104 (MCF-7) was adhered to

96 well plate overnight. 51Cr was exposed to the cells for 4 hrs and washed with cell media. 100 µL of KT-IgM-8, anti-Tn-PS A1 whole sera, anti-Tn-PS A1 IgG purified, anti-PS A1, or anti-PBS sera was added to each well in triplicate. The antibodies were incubated for 1 h at 37°C in a 5 % CO2 incubator, the cells were washed and 10% complement was added to each well. 51Cr release was measured after 18 h by liquid 153

scintillation to quantify 51Cr release and % cytotoxicity was calculated by using the following formula: (experimental – spontaneous) / (max – spontaneous) x 100.

Spontaneous wells only received media.

5.4.5. Flow Cytometry. KT-IgM-8 was diluted to 30 µg/mL and incubated with the cell lines (MCF-7 and HCT-116, 2.0 x106) for 30 min on ice and washed three times.

Cells were labeled with either Alexa Fluor® 647 and acquired using BD FACSCalibur™ and analyzed using FlowJo software.

5.4.6. SCID Mice tumor implantation and adoptive transfer of immunotherapeutic . SHOTM Mouse (Crl:SHO-PrKdcScidHrhr) (Charles River), female/4 week old, were surgically implanted with 17β-Estradiol 60 day release pellet (0.72 mg/pellet) (Innovative Research of America) behind the shoulders. Two days later 5 x

105 MCF-7 (tumors cell) were mixed with Geltrex® Matrix (1:1) at 4 °C and subcutaneously injected into mice on their flanks (2 x per mouse). The mice tumors were measured three times a week using calipers using equation (Tumor Volume mm3 =

(Length x width2) /2). Four days after tumor implantation, Tn-PS A1 whole sera, IgG purified Tn-PS A1 sera, PBS and Kt-IgM-8 were I.P injected once every week until the humane endpoint was reached. Data was analyzed using GraphPad Prism and student t- test was used for statistical significance.

154

5.5. References

(1) Reading, C. L. J. Immunol. Methods 1982, 53, 261.

(2) Köhler, G.; Milstein, C. Eur. J. Immunol. 1976, 6, 511.

(3) Cheever, M. A.; Allison, J. P.; Ferris, A. S.; Finn, O. J.; Hastings, B. M.;

Hecht, T. T.; Mellman, I.; Prindiville, S. A.; Viner, J. L.; Weiner, L. M.;

Matrisian, L. M. Clin. Cancer Res. 2009; Vol. 15, p 5323.

(4) Dhillon, S. Drugs 2015, 75, 923.

(5) Rob Aggarwal, S. Nat Biotechnol 2014; Vol. 32, p 32.

(6) Weiner, G. J. Nature 2015; Vol. 15, p 361.

(7) De Silva, R. A.; Wang, Q.; Chidley, T.; Appulage, D. K.; Andreana, P. R.

J. Am. Chem. Soc. 2009; Vol. 131, p 9622.

(8) De Silva, R. A.; Appulage, D. K.; Pietraszkiewicz, H.; Bobbitt, K. R.;

Media, J.; Shaw, J.; Valeriote, F. A.; Andreana, P. R. Cancer Immunol.

Immunother.; Springer-Verlag: 2012; Vol. 61, p 581.

(9) Irie, R. F.; Ollila, D. W.; O’Day, S.; Morton, D. L. Cancer Immunol.

Immunother. 2004, 53, 110.

(10) Liedtke, M.; Twist, C. J.; Medeiros, B. C.; Gotlib, J. R.; Berube, C.;

Bieber, M. M.; Bhat, N. M.; Teng, N. N.; Coutre, S. E. Haematologica

2012, 97, 30.

155

(11) Melis, J. P. M.; Strumane, K.; Ruuls, S. R.; Beurskens, F. J.; Schuurman,

J.; Parren, P. W. H. I. Mol. Immunol. 2015, 67, 117.

(12) Czajkowsky, D. M.; Shao, Z. Proc. Natl. Acad. Sci. U.S.A. 2009, 106,

14960.

(13) Ehrenstein, M. R.; Notley, C. A. Nature 2010, 10, 778.

(14) Desai, P. R. Transfus Med. Rev. 2000; Vol. 14, p 312.

(15) Springer, G. F. In J Mol Med 1997; Vol. 75, p 594.

(16) Manimala, J. C.; Roach, T. A.; Li, Z.; Gildersleeve, J. C. Glycobiology

2007, 17, 17.

(17) Karsten, U.; Diotel, C.; Klich, G.; Paulsen, H.; Goletz, S.; Muller, S.;

Hanisch, F. G. Cancer Res. 1998, 58, 2541.

(18) Karsten, U.; Serttas, N.; Paulsen, H.; Danielczyk, A.; Goletz, S.

Glycobiology 2004, 14, 681.

(19) Jegerlehner, A.; Wiesel, M.; Dietmeier, K.; Zabel, F.; Gatto, D.; Saudan,

P.; Bachmann, M. F. Vaccine 2010, 28, 5503.

(20) Schutze, M. P.; Leclerc, C.; Jolivet, M.; Audibert, F.; Chedid, L. J.

Immunol. 1985, 135, 2319.

(21) Wang, Y.; Ju, T.; Ding, X.; Xia, B.; Wang, W. Proc. Natl. Acad. Sci.

U.S.A 2010.

(22) Mulens, V.; de la Torre, A.; Marinello, P.; Rodríguez, R.; Cardoso, J.;

Díaz, R.; O Farrill, M.; Macias, A.; Viada, C.; Saurez, G.; Carr, A.;

Crombet, T.; Mazorra, Z.; Perez, R.; Fernández, L. E. Hum. Vaccin.

Immunother. 2014; Vol. 6, p 736. 156

(23) Welinder, C.; Baldetorp, B.; Borrebaeck, C.; Fredlund, B.-M.; Jansson, B.

Glycobiology 2011; Vol. 21, p 1097.

(24) Ando, H.; Matsushita, T.; Wakitani, M.; Sato, T.; Kodama-Nishida, S.; S

hibata, K.; Shitara, K.; Ohta, S. Biol. Pharm. Bull. 2008, 31, 1739.

(25) Nakada, H.; Inoue, M.; Numata, Y.; Tanaka, N.; Funakoshi, I.; Fukui, S.;

Mellors, A.; Yamashina, I. PNAS 1993; Vol. 90, p 2495.

(26) De Silva, R. A.; Wang, Q.; Chidley, T.; Appulage, D. K.; Andreana, P. R.

J. Am. Chem. Soc. 2009, 131, 9622.

(27) Andrew, S. M.; Titus, J. A.; Coico, R.; Amin, A. Curr. Protoc. Immunol.

2001, Chapter 2, Unit 2.9.

157

Appendix A (Chapter 1)

A.1. 1H NMR of PS A1

PS A1.esp Water 1.94

0.0030 1.96 1.36

0.0025 4.74

0.0020

3.90 4.67

0.0015 3.61

3.64

1.22

3.64

3.65

NormalizedIntensity

1.22

3.54

4.86

3.59

4.12 3.68

0.0010 3.53

3.47

4.96

3.94

3.41

3.79

4.37

4.65

3.96 5.21

0.0005 3.39

4.58 3.34

7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 Chemical Shift (ppm)

158

A.2. 1H NMR of Tn-PS A1

Tn-PS A1-3.002.esp

0.0030

0.0025

0.0020

0.0015 NormalizedIntensity 0.0010

0.0005

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0 Chemical Shift (ppm)

159

Appendix B (Chapter 2 )

B.1 1H NMR of Purified PS B (2)

160

B.3. 31P NMR of PS B (2)

22.68

1.0

0.9

0.8

0.7

0.6

0.5

0.4 -0.41

0.3

0.2

0.1

0.0

-0.1 200 150 100 50 0 -50 -100

161

B.4. 1H NMR of TF-PS B (4)

162

B.5. Expansion of 1H NMR of TF-PS B (4) from 6.8-8.0 ppm

163

B.6. MALDI-TOF of TF-BSA (5)

90080.640 Intens. [a.u.] Intens.

0 40000 50000 60000 70000 80000 90000 100000 110000 m/z

164

B.7. 1H NMR of TF-linker (7)

165

B.8. 13C NMR of TF-linker (7)

166

B.9. 1H NMR of STn-PS B (9)

167

Appendix C (Chapter 3)

C.1. 1H NMR of TF-PS A1 (4b)

JPB-TFPSA1-redialysis.001.esp

0.0015

0.0010 NormalizedIntensity

0.0005

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 Chemical Shift (ppm)

168

C.2. 1H NMR of Tn-TF-PS A1 (4c)

169

C.3. Expansion 1H NMR of Tn-TF-PS A1 (4c)

tn-tf-psa1.001.esp 0.008 PS A1-NHAc (2) 0.007 Tn-NHAc 0.006

0.005

0.004

TF-NHAc NormalizedIntensity 0.003

0.002

0.001 5.99 1.73 1.91

2.75 2.70 2.65 2.60 2.55 2.50 2.45 2.40 2.35 2.30 2.25 2.20 2.15 2.10 2.05 2.00 1.95 1.90 1.85 1.80 1.75 1.70 1.65 1.60 1.55 Chemical Shift (ppm)

170

C.4. Expansion 1H NMR of Tn-TF-PS A1 (4c)

tn-tf-psa1.001.esp

0.0040

0.0035 4.69

0.0030

0.0025

4.80

4.82

5.60

5.59 5.29

5.16

5.46

7.83 7.82

NormalizedIntensity 0.0020

5.64

4.91

5.22

5.38

5.34

7.18 5.05

0.0015

0.0010

7.5 7.0 6.5 6.0 5.5 5.0 4.5 Chemical Shift (ppm)

171

C.5. 2D H1-H1 COSY of Tn-TF-PS A1 (4c)

tn-tf-psa1.002.001.2rr.esp

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5 F1Chemical (ppm) Shift

6.0

6.5

7.0

7.5

8.0

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 F2 Chemical Shift (ppm)

172

C.6. 1D TOCSY Tn-TF-PS A1 (4c) doublet of (TF-Gal) at 4.61 ppm at room temperature with 120 ms mixing time

173

C.7. 1D TOCSY Tn-TF-PS A1 (4c) doublet of (Tn-TF-GalNac) at 4.61 ppm at room temperature with 120 ms mixing time. Note-both the GalNac, either from Tn or TF were indistinguishable due to overlap.

174

C.8. 1H NMR of (8)

kt-newlinker

2.25 1.0 0.9 0.8 0.7

0.6

0.5

3.03 2.73

NormalizedIntensity 0.4 9.67

0.3

2.74

3.04 3.02 2.72 0.2 9.67 0.1 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 Chemical Shift (ppm)

175

C.9. 13C NMR of (8)

kt-newlinkerc13

30.68 43.87

0.9 21.71 200.08 0.8

0.7

0.6

0.5 NormalizedIntensity 0.4

0.3

0.2 195.67

21.72

77.10

77.31

77.52 200.08 0.1

200 192 184 176 168 160 152 144 136 128 120 112 104 96 88 80 72 64 56 48 40 32 24 16 8 0 Chemical Shift (ppm)

176

C.10. 1H NMR of (9)

tn-new-linker.001.esp

4.71 4.70 2.25 1.93 1.93

0.035

0.030 1.93

0.025 2.27

0.020 NormalizedIntensity

0.015 3.63

3.63

3.83

3.62

2.28

4.74

1.96

4.74 3.92

0.010 3.83

5.29

4.18

2.98

2.43

2.44

7.52

2.99

5.28

3.93

2.97

4.17

2.96

2.98

7.52

1.80

1.96

5.37

4.67

3.81

2.44

4.18

4.15

7.53

2.95

7.53 3.00

0.005 4.17

2.42

4.20

6.90

7.51 4.81

3.61

2.67 3.60

7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 Chemical Shift (ppm)

177

C.11. 13C NMR of (9)

tn-new-linker.002.esp 29.89 1.0

0.9

68.29

29.26 71.21

0.8 98.54

25.44 21.85

0.7 48.89

60.85 67.57

0.6

0.5 154.63

NormalizedIntensity 0.4 71.52

0.3 174.58

0.2

0.1

180 160 140 120 100 80 60 40 20 0 Chemical Shift (ppm)

178

C.12. 1H NMR of (10)

tf-new-linker.001.esp Water

1.91 2.25

0.015 2.28

0.010

1.80

2.33

NormalizedIntensity

3.63 3.63

0.005 3.64

3.62

4.35

3.62

3.80

3.65 2.44

4.37

7.53

3.43

4.77

2.98

3.52

5.29

5.28

4.17

4.16 3.87 3.51

3.96

3.56 2.97

3.41

3.61

3.00

4.40

4.41

3.96

5.37

2.97

7.54

6.91

3.67

7.52

4.67

2.45

4.81

2.25

2.42

6.92

6.90

2.69 1.93

1.94 3.03

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 Chemical Shift (ppm)

179

C.13. 13C NMR of (10)

180

C.14. MALDI-TOF OF BSA Maleimide (11)

181

C.14. MALDI-TOF of Tn-BSA (12)

182

C.15. MALDI-TOF of TF-BSA (13)

183

Figure C.1. Anti-Serum of TF-PS A1 anti-IgG binding to TF-PS A1 coated ELISA plates in comparison to PS A1.

184

Appendix D (Chapter 4)

D.1. 1H NMR of GH-PS A1 (1a) at 60°C

0.00030 globoH-Psa1-monday.001.esp

0.00025

0.00020

0.00015 NormalizedIntensity

0.00010

0.00005

7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 Chemical Shift (ppm)

185

D.2. 1H NMR of GH-PS A1 (1a) at 22°C

gh-psa1-5-5-16.001.esp

4.78 4.74 4.71

0.0040

0.0035 4.68

2.32 4.67

0.0030

4.89

4.90 2.34

0.0025 4.06

4.05

NormalizedIntensity

4.07

4.04

1.50

4.07

1.49

4.18

4.02

4.03

4.01 3.99

0.0020 4.19

4.51

4.67

1.75

4.33

3.93

3.94

1.62

4.52

4.39

3.92

5.21

5.34

3.91

5.33

2.30

5.52 4.41

0.0015 2.36

5.51

4.65

5.40

3.86

3.88

1.45

5.81

5.63

2.23

3.62

1.39

2.99

8.03

8.04 1.13 3.58

0.0010

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 Chemical Shift (ppm)

186

D.3. 1H NMR of Tn-GH-PS A1 (1b) at 60°C

0.0020 tn-globoH-psa1-monday.001.esp

0.0015

0.0010 NormalizedIntensity

0.0005

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 Chemical Shift (ppm)

187

D.4. 1H NMR of Tn-GH-PS A1 (1b) at 22°C

tn-gh-psa1-2-22-16.001.esp Glyercol Impurity

0.00015

0.00010 NormalizedIntensity

0.00005

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 Chemical Shift (ppm)

188

D.5. 1H NMR of GB3-PS A1 (1c) at 22°C

GB3-PS A1.esp

0.0020

0.0015

0.0010 NormalizedIntensity

0.0005

8 7 6 5 4 3 2 1 0 Chemical Shift (ppm)

189

D.6. 1H NMR of GH-Thiol linker

0.0075 globoH-new-linker.001.esp 0.0070

0.0065

0.0060

0.0055

0.0050

0.0045

0.0040

0.0035

0.0030

NormalizedIntensity 0.0025

0.0020

0.0015

0.0010

0.0005

8 7 6 5 4 3 2 1 Chemical Shift (ppm)

190

D.7. 13C NMR of GH-Thiol linker

1.0 globoH-new-linker.002.esp

0.9

0.8

0.7

0.6

0.5

0.4

0.3 NormalizedIntensity 0.2

0.1

-0.1

-0.2

200 180 160 140 120 100 80 60 40 20 Chemical Shift (ppm)

191

D.8. 1H NMR of GB3-Thiol linker

gb3linker1-proton

0.015

0.010 AbsoluteIntensity

0.005

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 Chemical Shift (ppm)

192

D.9. 13C NMR of GH-Thiol linker

gb3-linker.001.esp

181.48 23.20

0.15

0.10 68.88

60.44

69.07

75.40

29.94

68.52

NormalizedIntensity

60.33

100.26

30.79 42.06

0.05 70.75

24.64 10.67

10.78

42.18

200.95

200.83

103.14

59.43

72.13

77.29

99.19

77.99

78.14

158.63 92.36

200 180 160 140 120 100 80 60 40 20 0 Chemical Shift (ppm)

193

D.15. MALDI-TOF OF BSA-Malemide for Globo H-BSA conjugate

194

D.15. MALDI-TOF OF Globo H-BSA conjugate

8000 Intens. [a.u.] Intens.

6000

92249.938 47829.055

4000

2000

50000 60000 70000 80000 90000 100000 110000 m/z

195

D.16. MALDI-TOF OF GB3-BSA conjugate

196

Scheme D1. Globo-H BSA conjugate

197

Scheme D2. GB3-BSA conjugate

198