THE ABO POLYMORPHISM AND PLASMODIUM FALCIPARUM MALARIA

Similar documents
Malaria Life Cycle Life Cycle

Immunohematology (Introduction) References: -Blood Groups and Red Cell Antigens (Laura Dean) -Cellular and molecular immunology, 8 th edition

Blood. Biol 105 Lecture 14 Chapter 11

Ultrastructural Studies on Plasmodium vivax

Functions of Blood. Transport. Transport. Defense. Regulation. Unit 6 Cardiovascular System: Blood

Immunohematology (Introduction)

4/5/17. Blood. Blood. Outline. Blood: An Overview. Functions of Blood

Antigen Presentation to T lymphocytes

Chapter 19: The Cardiovascular System: The Blood. Copyright 2009, John Wiley & Sons, Inc.

Malaria. Population at Risk. Infectious Disease epidemiology BMTRY 713 (Lecture 23) Epidemiology of Malaria. April 6, Selassie AW (DPHS) 1

Traits of Simple Inheritance. Polymorphisms in Human Blood

M.Sc. III Semester Biotechnology End Semester Examination, 2013 Model Answer LBTM: 302 Advanced Immunology

Topics of this lecture : RBC. Structural characteristics Hemoglobin Erythropoiesis Erythrocytes destruction

Principles of Adaptive Immunity

The Innate Immune Response

Immunohematology. Done by : Zaid Al-Ghnaneem

Zeina Al-Assaf. Mustafa Khader. Nayef Karadsheh

Overview: Blood Composition and Function

Agenda. Components of blood. Blood is Fluid Connective Tissue. Blood: General functions

Serology: ABO and Rh (D) blood group system

Lecture 2: Virology. I. Background

Chapter 22: The Lymphatic System and Immunity

Immunity. Acquired immunity differs from innate immunity in specificity & memory from 1 st exposure

Chapter 19 Cardiovascular System Blood: Functions. Plasma

Medical Virology Immunology. Dr. Sameer Naji, MB, BCh, PhD (UK) Head of Basic Medical Sciences Dept. Faculty of Medicine The Hashemite University

ACTIVATION OF T LYMPHOCYTES AND CELL MEDIATED IMMUNITY

Unit Seven Blood and Immunity

Overview of the Lymphoid System

I. Lines of Defense Pathogen: Table 1: Types of Immune Mechanisms. Table 2: Innate Immunity: First Lines of Defense

The recruitment of leukocytes and plasma proteins from the blood to sites of infection and tissue injury is called inflammation

The Adaptive Immune Responses

The Distribution of Human Differences. If all this genetic variation is so recent and continuous, why do we think of it in categorical terms?

Third line of Defense

Adaptive Immunity: Humoral Immune Responses

A. Blood is considered connective tissue. RBC. A. Blood volume and composition 1. Volume varies - average adult has 5 liters

BCM 317 LECTURE OJEMEKELE O.


Immune System AP SBI4UP

How the Innate Immune System Profiles Pathogens

Functions of Blood. 1. Transportation 2. Protection against invasion 3. Blood clotting 4. Regulation. Blood: Functions Blood clotting

Chapter 14. Blood. Blood Volume. Blood Composition. Blood

The hemoglobin (Hb) can bind a maximum of 220 ml O2 per liter.

The Distribution of Human Differences. If all this genetic variation is so recent and continuous, why do we think of it in categorical terms?

Erythrocytes. Dr. MOHAMED SAAD DAOUD BCH 471 1

Blood ---Ch 14. Whole blood consists of a plasma and cellular component.

Other Blood group systems

Blood consists of red and white blood cells suspended in plasma Blood is about 55% plasma and 45% cellular elements Plasma 90% water 10% dissolved

!!!!!Blood ---Ch 14. Whole blood consists of a plasma and cellular component.

The Immune System. These are classified as the Innate and Adaptive Immune Responses. Innate Immunity

Blood: Functions. Liquid connective tissue 3 general functions 1. Transportation. 2. Regulation. 3. Protection

Fluid movement in capillaries. Not all fluid is reclaimed at the venous end of the capillaries; that is the job of the lymphatic system

CHAPTER 10 BLOOD GROUPS: ABO AND Rh

I. Critical Vocabulary

Blood and Defense. Chapter 11

Hemolytic anemias (2 of 2)

Cellular Pathology of immunological disorders

HLA and antigen presentation. Department of Immunology Charles University, 2nd Medical School University Hospital Motol

The ABO and Rh system. Dr U. La Rocca 03 th Novembre 2017

Chapter 19. Openstax: Chapter 18. Blood

Physiology Unit 3. ADAPTIVE IMMUNITY The Specific Immune Response

Major Histocompatibility Complex (MHC) and T Cell Receptors

Chapter 19. Blood Types

CH 11 Blood OUTLINE: Functions of Blood Composition of Blood Blood Cell Disorders Blood Types Blood Clotting Functions of Blood Transportation

Haemoglobin BY: MUHAMMAD RADWAN WISSAM MUHAMMAD

Blood. The only fluid tissue in the human body Classified as a connective tissue. Living cells = formed elements Non-living matrix = plasma

Average adult = 8-10 pints of blood. Functions:

Ph.D. Thesis: Protective immune response in P.falciparum malaria 2011 CHAPTER I: Introduction. S.D. Lourembam 16

Hemoglobin and anemia BCH 471

BLOOD AND THE IMMUNE SYSTEM. Chapter 11

Third line of Defense. Topic 8 Specific Immunity (adaptive) (18) 3 rd Line = Prophylaxis via Immunization!

temperature 38 o C (100.4 o F) The temperature of blood is slightly higher than the normal body temperature of 37 o C (98.6 o F).

DONE BY : RaSHA RAKAN & Bushra Saleem

ANATOMY OF THE IMMUNE SYSTEM

Campbell's Biology: Concepts and Connections, 7e (Reece et al.) Chapter 24 The Immune System Multiple-Choice Questions

Hematology. The Study of blood

Invest in the future, defeat malaria

T-cell activation T cells migrate to secondary lymphoid tissues where they interact with antigen, antigen-presenting cells, and other lymphocytes:

T-cell activation T cells migrate to secondary lymphoid tissues where they interact with antigen, antigen-presenting cells, and other lymphocytes:

CELL BIOLOGY - CLUTCH CH THE IMMUNE SYSTEM.

APPENDIX -1 LIST OF TABLES. 01 The blood group systems Blood group collections

The Adaptive Immune Response. B-cells

Biology 2C03: Genetics What is a Gene?

There are 2 major lines of defense: Non-specific (Innate Immunity) and. Specific. (Adaptive Immunity) Photo of macrophage cell

Parasitic Protozoa, Helminths, and Arthropod Vectors

Principles of Anatomy and Physiology

What are the functions of blood?

Introduction and II. Blood Cells A. Introduction

Blood ESSENTIALS OF HUMAN ANATOMY & PHYSIOLOGY ELAINE N. MARIEB EIGHTH EDITION

Significance of the MHC

The Immune System is the Third Line of Defense Against Infection. Components of Human Immune System

Antibody Structure and Function*

All animals have innate immunity, a defense active immediately upon infection Vertebrates also have adaptive immunity

Chapter 19: The Cardiovascular System: The Blood. Copyright 2009, John Wiley & Sons, Inc.

Chapter 3, Part A (Pages 37-45): Leukocyte Migration into Tissues

VI. ABO and H Blood Groups

Hematopoiesis. Hematopoiesis. Hematopoiesis

Anatomy and Physiology

RBCs Disorders 2. Dr. Nabila Hamdi MD, PhD

IMMUNITY AND ANTIBODIES

Topic (6): The Complement System

Transcription:

THE ABO POLYMORPHISM AND PLASMODIUM FALCIPARUM MALARIA by Kayla Wolofsky A thesis submitted in conformity with the requirements for the degree of Master of Science Graduate Department of Institute of Medical Sciences, Department of Medicine University of Toronto Copyright by Kayla Wolofsky 2009

The ABO polymorphism and Plasmodium falciparum malaria Master s of Science, 2009 Kayla Wolofsky Institute of Medical Science University of Toronto Abstract Malaria has exerted a major selective pressure for red blood cell (RBC) polymorphisms that confer protection to severe disease. There is a predominance of blood type O in malaria endemic regions, and several lines of evidence suggest that the outcome of Plasmodium falciparum infection may be influenced by ABO blood type antigens. Based on observations that enhanced phagocytosis of infected polymorphic RBCs is associated with protection to malaria in other RBC disorders, we hypothesized that infected type O RBCs may be more efficiently cleared by the innate immune system than infected type A and B RBCs. The present work demonstrates human macrophages in vitro and murine monocytes in vivo phagocytosed P. falciparum infected O RBCs more avidly than infected A and B RBCs independent of macrophage donor blood type. This difference in clearance may confer relative resistance to severe malaria in individuals with blood type O. ii

Acknowledgments I would like to thank and extend my sincere gratitude and appreciation to my supervisor, Dr. Kevin Kain. He has been exceptionally supportive and his guidance, knowledge, and mentorship made this project not only possible, but also propelled it into new and exciting directions. As a result of the experience and opportunity to work with Dr. Kain and other gifted researchers, I was able to be included and contribute to the scientific community in ways I had not thought were possible for a Master s student. Thank you for allowing me these opportunities. I would like to thank members of my committee, Dr. Conrad Liles, Dr. Christine Cserti-Gazdewich, and Dr. Don Branch for their advice, expertise, and encouragement. You have all taught me that the most valuable education is not one learned solely by text books, but through collaboration, experience, and application. Thank you for donating your time and going above and beyond what was expected of you. I would also like to thank all of my fellow lab members for being exceptionally supportive and sharing their knowledge. A special thank you to Dr. Ayi Kodjo for being my mentor in the lab, continually teaching me techniques and encouraging me to believe in myself and be independent. You have taught me lifelong skills and your guidance helped me manage and learn from the frustrations and enjoy the successes. Finally, I would like to extend a sincerely deep gratitude to my parents, Ewa and Stan, and to my sister, Samara, for all of their support in pursuing my education and believing I will always succeed. Thank you for always being there to share the great moments. Thank you all for your encouragement, without it, I would not be where I am today. iii

TABLE OF CONTENTS PART I: LITERATURE REVIEW...1 SECTION 1: MALARIA BACKGROUND...2 1.1.1 The evolution and global distribution of malaria...3 1.1.2 Life cycle of Plasmodium falciparum malaria...6 1.1.3 Pathophysiology of Plasmodium falciparum malaria...9 1.1.4 Innate immunity to Plasmodium falciparum malaria...14 SECTION 2: THE RED BLOOD CELL...17 1.2.1 Erythropoiesis and physiology of the RBC...18 1.2.2 Essential components of the RBC...19 1.2.2.1 Hemoglobin...19 1.2.2.2 RBC enzymes...20 1.2.2.3 RBC membrane and aging (senescence)...20 SECTION 3: ABO BLOOD TYPE...24 1.3.1 History of ABO...25 1.3.2 Genetics and biochemistry of ABO antigens...25 1.3.3 ABO subtypes...29 1.3.4 ABO antibodies...29 1.3.5 ABO and infectious diseases...30 1.3.6 Geographic distribution of ABO blood types...32 SECTION 4: MALARIA, RED CELL POLYMORPHISMS AND NATURAL SELECTION...35 1.4.1 RBC polymorphisms...36 1.4.1.1 Hemoglobin mutations...37 1.4.1.2 RBC enzymes...39 iv

1.4.1.3 RBC membrane disorders...39 1.4.2 Inherited disease specific mechanisms of protection...40 1.4.2.1 Invasion and growth...41 1.4.2.2 Cytoadherence and rosetting...42 1.4.2.3 Clearance of infected polymorphic RBCs...42 1.4.3 ABO polymorphism and Plasmodium falciparum malaria...44 1.4.4 Potential mechanisms of protection afforded by blood type O...48 1.4.4.1 Invasion and maturation...48 1.4.4.2 Rosetting and sequestration...49 1.4.4.3 Additional mechanisms of protection...51 SECTION 5- AIMS AND HYPOTHESIS...53 PART II :MATERIALS AND METHODS...54 2.1 Reagents...55 2.2 Methods...56 PART III: RESULTS...62 PART IV: DISCUSSION...74 CONCLUSIONS AND FUTURE DIRECTIONS...90 v

LIST OF TABLES SECTION 1: MALARIA BACKGROUND Table 1. Prevalence of ABO frequencies in malaria-endemic regions....6 Table 2. Factors contributing to the clinical outcome of Plasmodium falciparum infection...10 SECTION 3: ABO BLOOD TYPE Table 3. The ABO blood type: Genotype, phenotype and antibodies...26 Table 4. ABO antigen sites on the red blood cell..28 Table 5. ABO frequencies in human ethnic populations...34 vi

LIST OF FIGURES PART I: LITERATURE REVIEW SECTION 1: MALARIA BACKGROUND Figure 1. Geographic distribution of Plasmodium falciparum malaria... 4 Figure 2. Life cycle of Plasmodium falciparum malaria......9 Figure 3. The PfEMP-1 molecule and associated host receptors..14 SECTION 2: THE RED BLOOD CELL Figure 4. Red cell senescence and band 3 aggregation.....23 SECTION 3: ABO BLOOD TYPES Figure 5. Structure of ABO antigens.....26 Figure 6. Biosynthesis of the ABO antigens.....28 PART III: RESULTS Figure 7. Similar invasion and growth of P. falciparum in A, B or O blood type red blood cells....64 Figure 8. Phagocytosis of A, B and O ring-infected RBCs by human monocyte derived macrophages.. 66 Figure 9. Increased phagocytosis of schizont infected O red blood cells compared to A and B infected red blood cells by human monocyte derived macrophages.....67 Figure 10. Schizont infected O RBCS are preferentially phagocytosed independent of macrophage donor blood type....69 Figure 11. Peritoneal monocytes of C57/B6 mice clear O-infected red blood cells more efficiently than A or B infected red blood cell...71 vii

Figure 12. Increasing phagocytosis of RBCs bearing decreasing A surface antigen....73 viii

ABBREVIATIONS AMA APC ATP CFU Apical membrane antigen Antigen presenting cells Adenosine triphosphate Colony forming unit CR1 Complement receptor 1 DARC EBA EtBr FcR FITC GPI G6PD GM-CSF Hb HbAS HbS HE HO HS Duffy antigen receptor for chemokines Erythrocyte binding protein Ethidium bromide Fc receptor Fluorescein isothiocyanate Glycosylphosphatidylinositol Glucose-6-phosphate-dehydrogenase Granulocyte-macrophage colony stimulating factor Hemoglobin Sickle cell trait Sickle cell anemia Hereditary elliptocytosis Hereditary ovalocytosis Heparan sulfate ix

ICAM-1 IFN MHC MSP NO PBMC Intercellular adhesion molecule-1 Interferon Major histocompatibility complex Merozoite surface protein Nitric oxide Peripheral blood mononuclear cell P.falciparum Plasmodium falciparum PfEMP PFK PfRBP PfRh PK PS Plasmodium falciparum erythrocyte membrane protein Phosphofructokinase Plasmodium falciparum reticulocyte binding protein Plasmodium falciparum reticulocyte homologue Pyruvate kinase Phosphatidylserine P. vivax Plasmodium vivax RBCs TNF Red blood cells Tumor necrosis factor VCAM-1 Vascular cell adhesion molecule 1 VSA WHO Variant surface antigen World Health Organization x

1 PART I LITERATURE REVIEW

2 SECTION 1: MALARIA BACKGROUND Malaria is estimated to account for 1-3 million deaths per year with the major burden of disease occurring in resource poor areas of the world. 1 Plasmodium falciparum malaria, accounts for the highest morbidity and mortality, and has a complicated pathogenesis that is influenced by geographic factors, parasite virulence factors, and host genetic determinants. Although a highly effective vaccine is not yet available, an increased understanding of the interaction between P. falciparum and the above determinants will contribute to interventions to achieve near or complete elimination of this disease.

3 1.1.1 The evolution and global distribution of malaria Malaria afflicts 300-500 million people each year and remains a leading cause of death worldwide, killing between 1-3 million people annually. 1,2 Malaria is one of the strongest known forces for evolutionary selection in the recent history of the human genome. 1,3,4 Understanding the historical and global relationship between malaria and human genetic diversity provides powerful insights into the evolution of the human genome and the pathogenesis of malaria. There are five Plasmodium species that infect humans; Plasmodium falciparum, Plasmodium vivax, Plasmodium ovale, Plasmodium malariae, and Plasmodium knowlesi. 5 These species differ in their morphology, immunology, and geographic distribution. 6 Among the five species that cause malaria in humans, Plasmodium falciparum (P. falciparum) is the most virulent resulting in the greatest number of complications and the great majority of malaria-related deaths in children under the age of five. 7,8 By killing children before they reach reproductive age, P. falciparum has, in essence, naturally selected for gene variants capable of conferring a survival advantage. Research which examines the key protective factors against P. falciparum malaria, could contribute to control of this major global health threat and further our understanding of human genetics and natural selection. The evolutionary history and geographical distribution of P. falciparum reflects a three-way interaction between the parasite, the host, and the Anopheles sp. mosquito (the vector for transmission). Circa 1900, prior to the widespread use of anti-malarials, the distribution of malaria reached the geographic latitudes of 64º north and 32º south. 9 Efforts during the 20 th century to control malaria restricted the global expansion of disease, however it remains endemic to climatic regions which facilitate continuous transmission and in the last two decades has recurred in several regions that had previously eradicated transmission (Figure 1). 2 P. falciparum

4 and its Anopheles vector, are normally confined to tropical, subtropical and warm temperate regions. 10 These regions are predominantly resource poor, highly populated areas with limited access to adequate malaria prevention and treatment programs. However, genetic history and the co-evolution of P. falciparum with humans suggest this has not always been the geographic model. The closest relative to the modern day P. falciparum is the chimpanzee malaria parasite, Plasmodium reichenowi. 3,10,11 It has been argued that P. falciparum is of African origin because P. reichenowi is a parasite that infects African chimpanzees. 10 Despite some controversy, it is generally accepted that the divergence of these two species of malaria occurred approximately 9-10 million years ago, prior to the divergence of humans from non-human primate relatives such as the chimpanzees. 3,10-12 It is believed that the major spread of P. falciparum in Africa occurred during the Agrarian Revolution (4000-5000 years ago) when small nomadic groups began to establish larger settled communities; this lifestyle change provided ideal conditions for sustained P. falciparum transmission. 10 Around 15 A.D. 10, malaria arrived in the Americas carried in the blood of European colonists and their slaves, and thereafter, became indigenous to the tropical regions of Central and South America, and some southern parts of North America. 13 Childhood infection prevalence Hyper-endemic: >50% Meso-endemic: 11-50% Hypo-endemic: <10% Unclassified areas: <6% Figure from Snow, R.et al. Nature.2005;434:214-217 Figure 1. Geographic distribution of Plasmodium falciparum malaria. 2

5 P. falciparum has been referred to as the the strongest known force for evolutionary selection in the recent history of the human genome. 3,4 P. falciparum malaria and humans have co-evolved and have adapted to one another, thereby selecting for survival genes simultaneously in both species. It has been noted that in African populations where P. falciparum is highly prevalent, certain polymorphic traits of the red blood cell, such as glucose-6-phosphate dehydrogenase deficiency (G6PD), sickle cell trait (HbAS), and α-thalassaemia, have conferred a survival advantage against severe malaria and death. 14-17 Blood type O, also a polymorphic trait, is present worldwide, which suggests that this trait must have been present when humans migrated out of Africa. 3,18,19 Interestingly, however, there is also a higher than expected prevalence of blood type O (along with the previously noted specific RBC polymorphisms) in Africa, especially in areas where P. falciparum is endemic (Table 1). 3,20-25 This suggests that like other RBC polymorphisms, blood type O may confer protection against severe malarial disease. To understand how the ABO blood types may be associated with malaria, the life cycle, and pathophysiology of malaria will be discussed. In addition, the innate immune system of the host, the RBC, and other known protective polymorphisms will be reviewed.

6 Table 1. Prevalence of ABO frequencies in malaria-endemic regions. 3,20-25 Region A B O Sickle cell trait/ thalassaemia Malaria Norway 50% 8% 38% No No Portugal 53% 8% 35% No No USA 42% 10% 44% No No Canada 44% 9% 36% No No Nigeria 21.3% 23.3% 51.5% Yes Yes Ghana 23% - 47% Yes Yes Kenya 19% 20% 60% Yes Yes Papua New Guinea 27% 26% 41% Yes Yes 1.1.2 Life cycle of Plasmodium falciparum malaria The P. falciparum infection begins when a human host is bitten by an infected female Anopheles mosquito, and the mosquito injects sporozoites into the subcutaneous tissue of the human host (Figure 2). 55,26 The sporozoites find their way into the blood stream where, within one hour of entering the human host, they travel to the liver and infect hepatocytes. 55,26 The duration of the asymptomatic liver (exo-erythrocytic cycle) stage of the infection is approximately one-two weeks. During this stage, each sporozoite may yield thousands of merozoites. 27

7 Invasion The hepatocytes rupture releasing the merozoites into the blood stream (the beginning of clinical disease) where they are able to enter into RBCs by a complex invasion process comprised of four phases: (a) initial recognition and reversible attachment of the merozoite to the RBC membrane, (b) reorientation, (c) invagination of the RBC membrane around the merozoite, and (d) resealing of the RBC membrane after completion of merozoite invasion. 5,6 A number of interactions and organelles have been identified between the RBCs and the merozoite. There are three organelles on the invasion (apical) end of the parasite, the rhoptries, micronemes and dense granules all which define the phylum Apicomplexa. 5 These three organelles contain the receptors that mediate invasion of the merozoite. RBC invasion is a rapid process that is governed by molecular interactions between the merozoites and the host cell surface. 28 Primary contact is initiated by a surface coat of proteins that is largely comprised of glycosylphosphatidylinositol (GPI)-anchored membrane proteins. There are at least nine recognized GPI anchored proteins that are predicted to be potential RBC ligands. 29 Merozoite surface protein-1 (MSP-1) is the dominant antigen and is essential for parasite survival as MSP-1 is involved in the initial recognition of the RBC via sialic acid residues found on the RBC membrane. Other important proteins are MSP-2, -3 and -4. 30 P. falciparum apical membrane antigen-1 (PfAMA-1) is also essential for successful invasion as it is translocated to the merozoites surface before invasion of the RBCs, and is also present on the sporozoite for invasion into hepatocytes. Other interactions between the merozoites and the RBCs include the erythrocyte binding antigen-175 (EBA-175), EBA- 140 and EBA-181 found on the merozoites which bind to glycophorin A, glycophorin C, and band 4.1 (respectively) on the RBC. 6,30,31 EBA-175 binds via sialic acid residues that are in an α-2,3 conformation, and binds specifically to glycophorin A. 32 P. falciparum is not able to invade RBCs that are missing glycophorin A. 33,34 Although some P. falciparum strains are reliant

8 on sialic acid to invade RBCs several strains have demonstrated the ability to shift to sialic acidindependent pathways (3D7). 35 This pathway utilizes a different family of ligands called P. falciparum reticulocyte binding proteins (PfRBP), or P. falciparum reticulocyte protein homologues (PfRh). 36 PfRh2b and PfRh4 are important in the sialic acid-independent invasion pathway, however the receptors responsible for bindings these ligands is unknown. 36,37 Redundancy in invasion pathways may provide an advantage to the parasite in case it encounters polymorphisms in host receptors. Once merozoites have successfully bound and invaded the RBC, the asexual stage of development of P. falciparum begins. Maturation Initially, the merozoites develop into an early trophozoite stage known as the ring form. The ring form persists for 24 hours and matures inside the RBC through a highly active metabolic state. The P. falciparum ring feeds from the host cytoplasm, importing glucose and breaking down hemoglobin into constituent amino acids. 6 Following the ring stage, P. falciparum matures and develops to a late stage trophozite. The mature trophozoite stage parasite replicates by nuclear division resulting in schizont stage parasites. Each schizont is comprised of 20-24 merozoites, which are released upon rupture of the infected RBC. 6 When the infected RBCs rupture, merozoites and parasite metabolic waste products such as hemozoin, degradation of hemoglobin, and parasite toxins are released. The majority of the merozoites will invade other RBCs continuing the asexual cycle; however, some parasites will form sexual stage forms called gametocytes which are then transmitted to new hosts by the Anopheles vector. 5,26 The asexual stages are solely responsible for the pathology associated with malaria which may manifest in a diverse array of pathological conditions.

9 Figure from Miller, L.H et al. Nature.2002;415:673-679 Figure 2. Life cycle of Plasmodium falciparum malaria. 5 1.1.3 Pathophysiology of Plasmodium falciparum malaria Infection with P. falciparum results in considerable morbidity and without treatment may be fatal. The clinical outcome of malaria depends on many contributing factors including the parasite s virulence, the host s response, geographical, and socio-economic factors (Table 2). 5 The combination of these factors result in a range of possible outcomes for the host, including asymptomatic infection, uncomplicated malaria infection, severe infection (severe malaria anemia and cerebral malaria) and death.

10 Table 2. Factors contributing to the clinical outcome of P. falciparum infection. Parasite Factors Host Factors Geographic and Social factors -Drug Resistance -Multiplication rate -Invasion Pathways -Cytoadherence -Rosetting -Malaria toxins (hemozoin) -Antigenic Variation (PfEMP1) -Immunity -Genetics : Sickle cell, thalassaemia, ABO blood type (?) etc. -Age -Pregnancy -Pro-inflammatory cytokines -Transmission intensity -Culture and economic factors -Access to treatment Asymptomatic Clinical Outcome Death Adapted from Miller, LH et al. Nature. 2002; 415:673-679 Clinical stages of malaria pathogenesis There are three defined clinical stages of malaria pathogenesis: uncomplicated malaria, severe malaria, and cerebral malaria. Uncomplicated malaria initially presents with fever and chills, nausea and headache, sometimes associated with diarrhea and vomiting. 7 Unfortunately, because of the similarity in symptoms, malarial infection is often mistaken for many other infections including influenza or gastro-intestinal infection and is therefore not properly treated. 7 The liver stage of the infection does not produce symptoms, but once the merozoites are released from the hepatocytes, symptoms usually occur approximately two weeks after the primary infection. The fever that individuals experience may become cyclical during the course of illness because of synchronized release of the merozoites from the RBCs. If left untreated, patients may progress to severe malaria.

11 In 1990, the World Health Organization (WHO) established criteria for the diagnosis of severe malaria. The major criteria include neurological involvement (cerebral malaria), pulmonary edema, acute renal failure, and severe anemia. 8 Severe anemia is the second most common symptom of P. falciparum infection and is caused by the destruction of RBCs and overall decreased erythropoiesis. 38,39 Acidosis and hypoglycemia are the most common metabolic complications. 8 Cerebral malaria is the most common cause of death in adults and children with severe malaria. 40 According to the WHO, the strict definition of cerebral malaria requires the presence of P. falciparum parasitemia and unarousable coma with a Glasgow Coma score of 9 or less; all other causes of coma, such as hypoglycemia, bacterial meningitis and viral encephalitis, need to be excluded. 40,41 Typical neurological symptoms include coma, seizures, edema, and brainstem damage. 40,41 Engorgement of cerebral capillaries and venules filled with infected RBCs and noninfected RBCs are typical histopathological findings in cerebral malaria. 41 As the infection progresses, the increasingly detrimental pathogenesis of P. falciparum malaria is believed to be caused by two main factors: 1) an imbalance of cytokine production; and 2) the sequestration of infected RBCs in the microvasculature of vital organs. Inflammatory response P. falciparum infection results in an increase of both pro- inflammatory cytokines and antiinflammatory cytokines. 42-44 However, in cerebral malaria, there is an unbalanced and excessive production of the pro-inflammatory response. 45-47 This phenomenon has been studied extensively but is beyond the scope of this thesis and will only be briefly reviewed.

12 Blood concentrations of pro-inflammatory cytokines, especially tumor necrosis factor (TNF), interferon gamma (IFN-γ), interleukin-1ϐ (IL-1ϐ), and IL-6, have been shown to be raised in cerebral malaria. 41,46,47 TNF may contribute to malaria pathogenesis including cerebral malaria. TNF up regulates endothelial cytoadherence receptors such as intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1), and E-selectin. TNF may cause hypoglycemia and dyserthryopoesis, and has been shown to induce the release of nitric oxide (NO) which interferes with synaptic transmission. 38,41,48 Parasite sequestration P. falciparum has a unique ability to adhere to host microvasculature endothelium, a process known as sequestration. Sequestration causes microvascular obstruction and compromises the blood flow through tissues such as the liver, spleen, lung, and brain. 38 The effects of sequestration include mechanical obstruction (which can lead to hypoxia), metabolic disturbances and is a central point where parasite toxins and inflammatory mediators concentrate. 38,49 Increased expression of cytoadherence receptors enhances infected RBC sequestration to the endothelium via parasite derived proteins (expressed on the surface of the infected RBC ), such as PfEMP-1 (Figure 3). 40 It is broadly accepted that ~12 hours after a merozoite invades a RBC, the principal parasite surface protein and sequestration ligand known as P. falciparum erythrocyte membrane protein 1 (PfEMP-1), encoded by var genes, is expressed. 50 It is predominantly mature stage parasites (trophozoites and schizonts) that adhere to the microvasculature. The PfEMP-1 molecule has a pivotal role in the pathogenesis of P. falciparum as a number of host receptors are recognized by the various extracellular binding domains of PfEMP-1 (Figure 3), thus permitting the infected RBCs to adhere to host endothelium. 51 In addition, PfEMP-1 binds to a number of different host receptors on both

13 monocytes and other RBCs. 44 In the case of cerebral malaria, PfEMP-1 may mediate adhesion to several adhesion molecules, in particular ICAM-1 which is unregulated on the cerebral vascular endothelium. 40 Post-mortem studies have shown that sequestration is greater in the brain than the other organs and correlates with ICAM-1 expression in cerebral vessels. 43,52 However, in some post mortem studies, not all individuals who die from cerebral malaria present with sequestered parasites. 38 In addition, the schizont infected RBC, to avoid clearance from the spleen, may bind to other schizont infected RBCs (agglutination), to other non-infected RBCs (rosetting), or to platelets bound to other infected RBCs or the endothelium. 53 Binding is mediated by a number of variant surface antigens expressed on the RBC, such as PfEMP-1, and any abnormality on either the receptor or ligands may prevent adhesion. Important receptors found on host tissues involved in sequestration and rosetting are heparan sulphate (HS), complement receptor 1 (CR1), thrombospondin, A and B blood type antigens, and CD36 (Figure 3). 38,49,54,55 HS is a proteoglycan found on the RBC, and along with CR1, acts as a receptor in the formation of rosettes. A and B blood type antigens have also been shown to act as co-receptors in the formation of rosettes, and depending on the blood type of the person can produce different rosetting rates and sizes (see section 1.4.4.2). 56 The role of CD36 in the pathogenesis of malaria is controversial and there is little evidence it contributes to cerebral malaria since there is little expression of CD36 in cerebral vessels. 38 CD36 is also found on platelets, monocytes and dendritic cells. 57 Parasite interaction with CD36 on monocytes has been shown to be an important interaction in non-opsonic phagocytosis and innate clearance of infected RBCs. 58-60

14 In summary, high parasite burdens leading to sequestration and dysregulated immune responses are thought to make important contributions to the pathogenesis of cerebral malaria and severe disease. Adapted from Cserti, CM et al. Blood. 2007; 110:2250-2258 Figure 3. The PfEMP-1 molecule and associated host receptors. 1.1.4 Innate immunity to Plasmodium falciparum malaria The innate immune response is crucial to the outcome during a P.falciparum infection. Innate immune responses take effect immediately and provide an early defense until the adaptive immune response is engaged. In some cases, an infection by P. falciparum may be controlled by the innate immune system. 61 Parasite burdens observed in non-immune individuals with acute P. falciparum malaria are lower than expected based on parasite replication rates observed in vitro, suggesting that the innate immune system can contribute to effective control of acute parasite replication before the adaptive immune response develops. 58,59,62 The innate immune system functions to limit the maximum parasite density, but gradually acquired adaptive mechanisms complete parasite elimination. The innate immune system is essential for most inflammatory responses that are triggered by monocytic cells, other leukocytes and mast cells through their innate sensing receptors. 63 Studies have shown that macrophages are important in innate immunity as they are able to clear parasitized RBCs in the absence of opsonizing malariaspecific antibodies. 59 It is hypothesized that there are two methods of infected RBC uptake by macrophages. The predominant method of uptake involves the binding of non-specific IgG and

15 complement to the surface of infected RBCs, and increased exposure of senescent RBC markers such as exposure of phosphatidylserine (PS). This method induces the release of pro-inflammatory cytokines. The second method of uptake is CD36 mediated, which involves the binding of CD36 on the macrophage to PfEMP-1 on the infected RBCs. This method does not involve the release of pro-inflammatory cytokines. 64,58 There are three main biochemical pathways that result in activation of the complement system: the classical complement binding pathway; the mannose-binding lectin pathway; and the alternative pathway. All three lead to the formation of C3 and C5 convertase which results in the cleavage of C3 and C5 into C3a, C3b, C5a and C5b, respectively. 65 RBCs opsonized by IgG and complement (C3b) are recognized by the Fc receptor (FcR) and CR1 (respectively), and phagocytosed by macrophages. 66,67 This method of clearance is effective in senescent and damaged RBCs, and also in P.falciparum infected RBCs. 67 Ayi et al compared the uptake of ring infected RBCs and mature infected RBCs and found an overall higher affinity for the uptake of mature stage parasites over ring stage. 68 This uptake may be due to the structural changes within the RBC, resembling RBC senescence. Specifically, in mature infected RBCs there is an increase in hemichrome deposition and band 3 aggregation as well as increased expression of PfEMP-1 on the surface of the RBC. 68,69 The exact mechanism of phagocytosis of infected RBCs by monocytes and macrophages is unknown; however the primary mechanism is thought to be through complement and IgG binding, and even possibly via PS dependent pathways. 66,70-72 Any trait of the RBC that enhances clearance of infected RBCs may confer a survival advantage. This is especially true if there is enhanced phagocytosis of schizont stage parasites, as it will

16 reduce the number of infected RBCs available to bind within the micro-vascular beds of vital organs.

17 SECTION 2: THE RED BLOOD CELL At first, RBCs seem to be one of the simplest cells in the human body, with no organelles, no nucleus and only two major functions, to deliver oxygen and remove carbon dioxide to and from the tissues. 73 However, blood, more specifically, the RBC is often referred to as the essence of life. The RBC is rich in nutrients, continually renewing, and is a shelter not recognized by the immune response, thereby rendering it the perfect target for a hemotropic pathogen, such as P. falciparum.

18 1.2.1 Erythropoiesis and physiology of the RBC RBCs are anucleate cells devoid of typical organelles. 74 They are found in the blood stream and their primary function is to deliver oxygen and remove carbon dioxide from the tissues. Erythropoiesis is the process by which RBCs are produced in the bone marrow. 73 The process of RBC maturation involves a series of differentiation steps which are tightly regulated. RBCs develop from the multi-potential progenitor cell (CFU-GEMM) under the influence of erythropoietin (EPO), granulocyte-macrophage colony stimulating factors (GM-CSF), IL-3, and IL-4. 74 This gives rise to the erythropoietin sensitive cell CFU-E which when stimulated by erythropoietin, develops into a RBC. 74 At this stage of RBC development, the cell is released from the bone marrow and is called a reticulocyte. In healthy, non-infected blood, 1-2% of the total RBC counts are reticulocytes. Reticulocytes have stage-specific surface antigens specific to this stage of development, making them targets for invasion by Plasmodium vivax (P. vivax). 75 After a day of circulating in the blood stream, they develop into mature RBCs. 74 The mature RBC has a biconcave structure that is deformable. It only consists of cytoplasm and a cell membrane, which allows the RBC to pass through small blood vessels and narrow capillaries. The disruption of erythropoiesis is important to the development of severe malaria anemia (SMA). In cases of SMA, low reticulocytosis is observed, suggesting insufficient erythropoiesis as a major factor for anemia. In one study, researchers found reduced proliferation and terminal differentiation of erythroblasts to become mature hemoglobin-producing cells. 76 The suppression of initial proliferation, differentiation and maturation, and inadequate reticulocytosis has been proposed as the basis of insufficient erythropoiesis during malaria.

19 1.2.2 Essential components of the RBC As mature RBCs do not differ phenotypically, do not contain internal mechanisms of synthesis or traffic proteins, or express class I or II MHC molecules on their surface, they are the ideal vehicle for the parasite to evade the immune system of the host. 77 The RBC is central to the life cycle of P. falciparum, and alterations to the RBC may impair parasite growth and replication and provide a survival advantage to the host. There are three major components to the RBC: hemoglobin, metabolic enzymes, and the RBC membrane. All three have a pivotal role in parasite development. 1.2.2.1 Hemoglobin Hemoglobin is an assembly of 4 globin protein chains (2 -α globins and 2-ϐ globins chains) arranged into a set of α-helix structures and 4 heme groups. 73,77,78 The hemoglobin within the RBC transports oxygen from the lungs to the tissues and trades it for carbon dioxide to take back to the lungs. The binding of the oxygen to the iron molecule in hemoglobin causes the hemoglobin to undergo a conformational change. There are two main states of hemoglobin - the deoxyhemoglobin (taut state, no bound oxygen) and the oxyhemoglobin (the relaxed state, bound oxygen). 79 Over time this process causes oxidative stress on the RBC. 80 Hemoglobin is central to malarial infection as the parasite internalizes and degrades massive amounts of hemoglobin from the host RBC during the blood stage of infection. 81 Hemoglobin is hydrolyzed to free amino acids, which are subsequently incorporated into parasites proteins. 82 The released heme is stored as a polymerized byproduct called hemozoin. In vitro studies have shown that macrophages that have engulfed hemozoin, are unable to digest it and are unable to repeat phagocytosis, or kill bacteria. 83 Mutations in the genes for the hemoglobin proteins results

20 in hemoglobin variants, resulting in a group of hereditary diseases termed hemoglobinopathies (discussed in section 1.4.1.1), which undoubtedly have an impact on P. falciparum malaria. 1.2.2.2 RBC enzymes Malaria parasites spend virtually all of their asexual life cycle inside the RBCs of their host. They must continually adapt themselves to the host s RBC environment, and need a very specific environment to grow. 84 Malaria-infected RBCs are consistently under oxidative stress, and have high levels of H 2 O 2 and OH - radials, which are primarily produced by the digestion of the hemoglobin by the parasite. 85 There are two important groups of enzymes involved in the metabolic pathway in the RBC: enzymes involved in energy metabolism, and enzymes to prevent/reverse oxidative damage. The mature RBC does not have mitochondria or storage capacity, and therefore utilizes anaerobic glycolysis as a source of energy. 79 The enzyme pyruvate kinase (PK) is a member of the anaerobic pathway and produces a molecule of ATP. 78 Enzyme deficiency in the anaerobic pathway leads to reduced viability of the RBC. The RBC is continually susceptible to oxidative damage and at high concentrations of radicals or insufficient protection, oxidative damage leads to a loss of the ability of the RBCs to transfer O 2 and CO 2 and eventually the cells hemolyze. 86 Glucose-6-phosphate dehydrogenase (G6PD), an enzyme in the aerobic pathway produces NADPH, which is essential for the production of pentoses needed for the anaerobic pathway and is a major reducing agent in the RBC. 78,87 Two additional enzymes that prevent and reverse oxidative denaturation of the hemoglobin are methemoglobin reductase and superoxide dismutase. 86,88 1.2.2.3 RBC membrane and aging (senescence) Several properties of the RBC membrane are essential to the invasion and pathogenesis of P. falciparum malaria. The membrane contains the structures required for invasion and gives the

21 parasite a medium to express surface antigens such as PfEMP-1, which are ultimately important for its survival. The primary function of the RBC membrane is to form a phospholipid bilayer to protect and contain hemoglobin allowing the RBC to carry oxygen. Membrane The RBC membrane is an exceptionally complex membrane with over 50 transmembrane proteins identified. Key structures on the membrane include the Rh complex, Duffy glycoprotein, ABO antigen, and band 3. 89 The Rh complex has been implicated in RBC structure and is hypothesized to have a role in ammonium and CO 2 transport. 90-92 The Rh complex is important in blood transfusion, and has a primary role in hemolytic disease of the newborn. 90-92 The Duffy glycoprotein is present on the RBC and corresponds to the chemokine receptor DARC (Duffy Antigen Receptor for Chemokines) which is a general receptor for several chemokines and removes excessive chemokines from the blood and tissue. 93 The ABO antigen is a carbohydrate based structure, and will be further discussed in section 3. Band 3 is a transport protein that is associated with RBC senescence as it aggregates and signals for RBC removal. 94,95 Senescence RBC aging is thought to be accelerated by the intracellular presence of P. falciparum. 96 This has been attributed to decreased levels of antioxidants and ATP, coupled with enhanced flux of ions, especially calcium. P. falciparum induces biochemical modifications in the membranes of infected RBCs, mimicking the physiologic aging process of the cell, most likely produced through P. falciparum-induced oxidative stress. 97 There are also a number of key structures on the membrane that are associated with RBC aging and senescence such as band 3. Band 3 comprises 25% of the total membrane protein and while it has a structural and transporter role it

22 also supports a number of RBC antigens (ABO, Diego and RH). Band 3 is important to maintain the flexibility of RBCs and also has a role in anion exchange as it exchanges Cl - for HCO - 3, thus removing CO 2 from tissues. 95 Consequently, it is believed to have an essential role in the removal of aged or defective cells. 98 The average life span of RBCs is 120 days after which most are phagocytosed by macrophages in the spleen and liver. 74 The trigger mechanism for senescence is not completely understood; however there are RBC modifications that occur both internally and externally that may contribute to senescence. Senescent RBCs are smaller, more rigid, and their surface is partially desialylated. 96 Many of the primary enzymes have reduced activity, thereby increasing reactive oxygen species (ROS). This leads to accelerated oxidative injury and the denaturation of hemoglobin to form hemichromes which bind to the cytoplasmic domain of band 3 (Figure 4). 94,98,99 The binding of hemichromes to band 3 causes cross-linking of the cytoplasmic domains resulting in clustering. Clustering is detected by autologous antibodies to the altered band 3 molecules. Fc and complement receptors on the macrophage recognize these signals and the RBC is cleared by the macrophages 94,98,99. In addition, it is thought there may be neo-antigens exposed during band 3 degradation, including increased exposure of PS on the RBC surface, which collectively transforms the RBC from self to non-self. 60,66,100

23 Figure from Pantaleo, A et al.autoimmunity Reviews.2008;7:457-462 Figure 4. RBC senescence and band 3 aggregation. 1. Oxidative denaturation of hemoglobin leading to hemichrome formation 2. Hemichrome binds to cytoplasmic domain of band 3 and causes oxidative cross-linking 3. Band 3 dissociates from cytoskeletal proteins and 4. Formation of band 3/hemichrome clusters and opsonization by anti-band 3 antibodies and C3b

24 SECTION 3: ABO BLOOD TYPE The ABO polymorphism is the most recognized and the most clinically important antigen classification system to date. Its recognition is central to the practice of transfusion medicine, because of the immediate recognition and rejection of major incompatible non-self cells. In the past century since the ABO discovery, scientists have been able to identify an association between the ABO blood type and a number of infectious diseases, some of which exert genetic selection. However, the cause of the molecular and geographic diversity, along with the evolutionary basis for the origin of the ABO blood type, remains a mystery.

25 1.3.1 History of ABO The ABO blood type was first recognized in 1900 by Karl Landsteiner. 101,102 The ABO blood type is the most well known and the most clinically important of the 29 blood type systems due to its significant role in successful blood transfusion and organ transplantation. In this blood type system, there are three blood type antigens, A, B and O. These antigenic determinants are oligosaccarchides located on glycol proteins and glycolipids expressed on RBCs, and are predominantly found on band 3. 103,104 ABO antigens are found on other cell tissues, epithelia, various body fluids, secretions, lymphocytes, and platelets, however they are expressed most densely on RBCs. 104,105 At first glance, this system seems to be relatively simple; three antigens (A, B, O), six genotypes (AA, AO, BB, BO, OO, AB), and 4 phenotypes (A, B, O, and AB). 102 However, within 30 years of discovery of the ABO blood type, subtypes of each blood type were discovered (section 1.3.3). Molecular studies on the ABO blood types show that the O gene appears to be a mutation of the A gene, suggesting that type O appeared later than the other ABO blood types, and therefore, blood type O is considered to be the polymorphic type. 18,19,106 1.3.2 Genetics and biochemistry of ABO antigens The ABO blood type antigens are subject to strict genetic control and are a perfect example of Mendelian genetics. Genes A and B each express themselves dominantly with respect to the O silent gene, but A and B are co-dominant and in a heterozygote, both A and B antigens are expressed on the RBC (Table 3). 107

26 Table 3. The ABO blood type: Genotype, phenotype and antibodies. 108 Genotype Phenotype Antibodies present in serum OO O Anti-A,B AO AA A Anti-B BO BB B Anti-A AB AB None The ABO gene itself does not encode the carbohydrate ABO antigen but rather a glycosyltransferase, which transfers either a α-1,3 linked N-acetylgalactosamine by α -1,3 N- acetyl galactosaminyltransferase or a α-1,3 linked galactose by α-1,3 galactosyltransferase to form A or B antigens, respectively, on the precursor molecule (H) (Figure 5). 107,109,110 O (H) B A 2 A 1 Figure from Sheffield, WP et al.transfus Med Rev.2005;19:295-307 Figure 5. Structure of ABO antigens. The precursor structure (H) to which A and B trisaccharides are attached is genetically independent of the ABO system. It represents the precursor substance in the biosynthetic pathway leading to the A and B determinant structures. The O gene represents an amorph or

27 silent gene as there is no antigenic expression or glycosyltransferase enzyme formed. 108 Based on this close chemical relationship, the H system is now included in the discussion of the ABO system, and is referred to as the ABO (H) system. 111 The blood type O character is found on RBCs and in secretions of blood type O individuals, and is considered the product of the O (H) gene. O( H) molecules are also present in non-o individuals, but to a lesser degree, as the remainders of chains are left unmodified by the glycosyltransferase activity. 112,113 The ABO locus, located on chromosome 9, position 9q34,1-q34.2 has seven exons ranging from 28-688 base pairs, and six introns with 554 to 12983 base pairs. 110,114 The last two exons, 6 and 7, which encode 823 of the 1062 base pairs of the transcribed mrna encode for the catalytic domain of the ABO glycosyltransferase. 114 Most variants of the A and B antigens are encoded by missense mutations in exon 7. 105 The blood type O allele has a single nucleotide deletion, found in G261 in exon 6. This deleted nucleotide is not found in A and B sequences. This deletion causes a frame shift that alters the protein sequence after amino acid 88 and introduces a stop codon after nucleotide 352 in the consensus sequence. Enzyme activity is never achieved after the pre-catalytic translation stop signal at amino acid 117. 110 The O (H) locus is localized on the long arm of chromosome 19 at position q13.3. 111 As earlier stated, the ABO antigens are not confined to RBCs, but are also found in other organs, various secretions, and tissues. The ability to secret ABO antigens is closely linked to the H locus and is found on chromosome 19q13.3. This gene is known as the secretor gene (Se/se). Individuals who have the homozygous or heterozygous Se gene, are able to secrete ABO active material according to their genotype. In contrast, individuals who are homozygous for the ``se`` gene are not able to secret antigens. 103 The formation of the ABO antigens is a multistep process that involves a number of different enzymes and pathways depending on an individual s ABO blood type ( Figure 6). 115

28 Figure 6. Biosynthesis of the ABO antigens. Adapted from Hosoi, EJ Med Invest.2008;55:174-183 Although the values somewhat differ, the average number of ABO antigen sites per RBC are in the range of 1.5-2 million. 111 All RBCs, independent of the ABO blood type have O(H) antigen present and the subtype determines the number of ABO determinant sites per RBC (Table 4). 103,111,116 Table 4. ABO antigen sites on the RBC. 104,111,116 Blood Type A sites B sites O(H) sites A1 910,000-1,300,000-70,000-170,000 A2 160,000-290,000-1,080,000-1,210,000 B - 610,000-830,000 540,000-760,000 O - - 1,590,000-1,740,000

29 1.3.3 ABO subtypes In addition to the 4 major types A 1, B, A 1 B and O there are additional subtypes. Subtypes of ABO are distinguished by decreased amounts of A, B, or O(H) antigens on the RBCs and they are classified based on the differences in the strength of agglutination of RBCs with anti-a, anti- B, and anti-a,b reagents. 117 Blood type A has the most variation and is classified by the quantity of A-antigen present. A 1 is considered the normal A blood type with the highest number of A antigens present. 117 In decreasing order by the quantity of A-antigen, other subtypes include: A 2, A 3, A x, A end, A m, and A el. 111,115 In general 80% of people who are blood type A are A 1 and the remaining 20% are A 2. 115 A 1 is said to be more branched (Figure 5) and has a greater number of A antigenic sites (Table 4). 117 The difference between A 1 and A 2 can be observed by the agglutination of A 1 RBCs with anti-a 1 lectin extract of Dolichos biflorus seeds. To date, studies have shown that the A transferase of A 1 and A 2 are both N-acetyl-galactosaminyltransferases, however their kinetic properties differ slightly. The α-1,3-n-acetylgalactosaminyltransferase activity is 5-10 times higher in A 1 than that of A 111 2. The A 2 enzyme adds only a single N- acetylgalactosamine group on the H substance and modifies fewer O (H) antigens. 118 Studies have shown that people of A 2 subtype can have antibodies against A 103,119 1. Similarly, subtypes of blood type B are classified by the quantity of B antigen, and the amount of B antigen decreases in the order B, B 3, B x, B m, and B el. 108,111 1.3.4 ABO antibodies Anti-A and anti-b antibodies of IgM and IgG origin were long thought to be naturally occurring and produced at birth. However, the discovery of the presence of A and B antigens on plants, bacteria, and parasites led to the current widely accepted position that initial formation of anti-a and anti-b antibodies are likely stimulated in newborns by exposure to environmental antigens in the aerodigestive tract which are similar to or identical in structure to the A and B antigens found

30 on these living organisms. 108,120,121 This was demonstrated in a classical experiment performed on White Leghorn chickens raised in a germ free environment. Initially, the chickens had no anti-a or anti-b titers. However, when exposed to Escherichia coli (E.coli), which has been shown to express B antigen, chickens formed anti-b antibodies. 108 This model demonstrates how antibodies directed against RBC antigens may result from natural exposure to carbohydrates that mimic some blood type antigens. RBC antibodies cause breakdown of the RBCs by either C3b-mediated cell phagocytosis or through direct lysis due to activation of the terminal complement components (C5b-9). C3b mediated RBC phagocytosis is referred to as extravascular immune hemolysis and is the most common. 122 This is the removal and destruction of RBCs by macrophages of the spleen and liver. During the normal aging of RBCs in circulation, RBCs are destroyed and degraded by macrophages through the extravascular pathway. This pathway is primarily mediated by IgG1 and IgG3 as splenic macrophages have receptors specific for Fc fragments. 123 1.3.5 ABO and infectious diseases Many studies have attempted to establish a relationship between the ABO blood type system and various infectious diseases, and recent studies have focused on establishing a scientific rational for these statistical relationships. 10,124-128 The most convincing statistical associations have been found to be between ABO and diseases such as cancer, peptic ulcers, coagulation and infection from various viruses and pathogens. 124,129,130 In 1954, Aird et al reported that type O individuals were 50% more likely to have duodenal ulcers than type A, B and AB individuals. 130 The rationale for this observation was that the blood type Le b, which is closely associated with the ABO system, is lower in individuals of blood type A. Boren et al, found that Le b is a receptor for Helicobacter pylori. 18,131 Another group found that individuals of blood type O were more likely

31 to develop hemorrhage rather than deep venous thrombosis, the latter more common in blood type A individuals. This phenomenon may be attributed to the fact that type A individuals have higher levels of factor VIII (anti-hemophilic globulin/von Willebrand factor levels), and thus a higher than average level of coagulation capacity. 18,132 More relevant to the hypothesis of this thesis is the relationship between the ABO blood types and infectious diseases. It has been shown that many gram-negative organisms, such as E. coli, have chemical moieties resembling either the A or B antigen. E. coli itself is said to have A and B -like antigen present on its surface, thereby stimulating anti-a and anti-b antibodies. A study done in 1978 by Marsh et al demonstrated how intestinal E.coli stimulated the production of anti-b antibody in a 20 day old child. 133 There are a number of very interesting associations and while some are relatively minor (e.g. E. coli and blood type A and B), some relationships between ABO blood types and infectious diseases have been speculated to account for the various worldwide distribution of the ABO blood type. It has been suggested that the distribution of ABO blood types in various parts of the world may have been influenced by the occurrence of endemic and epidemic diseases such as small pox and plague. 124,134 Association studies have shown there is an increased incidence of small pox in people of blood type A. Small pox has been shown to express A-like antigens, and during viremia the virus is thought to be a cognate target of anti-a antibody. This suggests that individuals with blood type O or B who contract small pox may have an immunological advantage over those who are blood type A. 124,126,135 Conversely, the pathogenesis of the plague and cholera has been said to be influenced by blood type O. It has been reported that an H-like antigen is expressed on the plague bacilli and Vibrio cholera. Therefore, individuals with blood type O may tolerate these bacterial infections better than individuals with blood type A or B. 18

32 P. falciparum malaria is an infectious disease that may express A and B-like antigens. It has been shown that patients with P. falciparum malaria have higher anti-a and anti-b agglutination titer when compared with those who were test negative for malaria. 125,128,136 Gonzales and colleagues were the first to compare the anti-a and anti-b titers of sera from blood type O patients infected with malaria and from individuals blood type O negative for malaria. They found the anti-a titers were ~250 x higher for individuals infected with P. falciparum, and anti-b titers were 32 x higher. 125 These datas alone suggest that P. falciparum may express A- and B- like antigens and may be recognized by the immune system as non-self sooner in individuals with blood type O. This may be due to an additional antibody (anti-a and anti-b antibodies) immune response alongside the traditional immune response to P. falciparum infected RBCs. 125 1.3.6 Geographic distribution of ABO blood types There are striking differences with respect to the frequency and distribution of the ABO blood types between populations in varied geographic locations. These differences are related to a number of factors including natural selection, genetic drift, founder effect, and mutations. Although the frequency of the ABO blood type distribution worldwide is not exactly known, it is commonly accepted that blood type O is the most common (45%), followed by type A (40%), B (11%) and AB (4%). 137 However some blood types, like O in Africa, are more prevalent in certain geographical locations and ethnic groups.this leads to two questions: 1) if blood type O is determined by the recessive gene, what occurred to make the O phenotype so common?; and 2) if A and B are dominant alleles, why are these two phenotypes not predominant in any given population? To answer these questions, one must look at the distribution of ABO across the world and look at the factors that may influence positive selection for the O phenotype. It is important to approach this question from three perspectives: 1) how do the frequencies of ABO differ in identified ethnic populations?; 2) where do these ethnic populations predominate

33 geographically?; and 3) what factors (including historical health epidemics) may have occurred to influence the outcome? In temperate regions above the tropic of Cancer and below the tropic of Capricorn, blood type A is more prevalent. In Asia, blood type B is the most common blood type and around the equator, especially sub-saharan Africa, blood type O is the most prevalent. When comparing ABO frequencies in certain parts of the world to the presence or absence of malaria, it appears as though blood type O is prevalent in areas of endemic malaria where certain RBC polymorphic traits are also elevated (Table 1, Section 1.1.1). 3,20-25 If one is to consider ethnicity alone, amongst Caucasians, the distribution of those with either blood type A and O is equal (44% each), while those with blood type B is 9% ( Table 4). 108 This implies that neither the A nor O blood type is favored or advantageous. However, there is a significant difference when comparing the A and O blood types in those of African descent. The frequency of blood type O is approximately 2 times greater than that of blood type A and 3 times greater than B (O=49%,A=27%, B=19%) (Table 5). 108 This suggests that O is being selected for within this population. These data are even more interesting given that blood type O is a recessive gene. Clearly, it is important to further examine the relationship between people of African descent, ABO, and geographic factors (and the infectious diseases that reside there).

34 Table 5. ABO frequencies in human ethnic populations. Blood Type Caucasian (%) Black African (%) A 1 A 2 B AB O 34 10 9 4 44 19 8 19 4 49 Adapted from Gibbs, FL et al. Blood group systems, ABH and Lewis. American Association of Blood Banks.1986.

35 SECTION 4: MALARIA, RED CELL POLYMORPHISMS AND NATURAL SELECTION Malaria is one of the leading causes of deaths worldwide from a single infectious agent, killing ~1 million people per year, 85% of which are children under the age of five. 1 P. falciparum malaria has made a larger footprint on the human genome than any other pathogen. It has coevolved with the human population and selected for survival genes by causing fatalities in individuals before the age of reproduction. 3 The current form of P. falciparum has existed for over 8 million years and caused massive mortality among modern humans and our hominid ancestors. Any gene conferring protection from fatal P. falciparum infection has undeniably been subjected to the true process of natural selection.

36 1.4.1 RBC polymorphisms The RBC is a complicated cell, with a complex array of intrinsic and extrinsic proteins, carbohydrates, receptors, transporters, and enzymes, all of which have distinct functions. P. falciparum spends most of its growth cycle in the intra-erythrocytic stage which is the stage when disease pathogenesis occurs. The relationship between the RBC and the parasite is complex and integrated; the parasite is RBC centric. Any defect directly affecting a membrane component can modify the shape of the RBC and reduce its survival capacity, and ultimately create an inhospitable environment for the malaria parasite to invade and grow within. 77 The same effect may result from molecular lesions of cytoplasmic proteins or enzymes, all of which could contribute to, or even cause, cell death. Over time, the morbidity and mortality of malaria has produced great selection pressures in affected populations for mutations that reduce the impact of the disease. There is strong evidence supporting an association between severity of malaria, RBC polymorphisms, the geographic distribution of malaria transmission, and the increased frequency of various polymorphic traits. This is also known as the malaria hypothesis as originally proposed by Haldane, which states common abnormalities in RBCs have been selected because of the fitness advantage they confer against malaria. 51 Inherited RBC disorders with altered membrane and functions can be broadly divided into three classes. The first class encompasses altered functions due to mutations in the various membrane or cytoskeletal proteins, including hereditary disorders such as elliptocytosis (HE), ovalocytosis (HO), and stomatocytosis. 138 The second major class of disorders include altered function due to secondary effects on the membrane resulting from mutations in globin gene such as sickle cell disease (HbS) and α-and ϐ-thalaseemia. 139,140 A consistent feature of these two classes is

37 decreased cell deformability. The structural integrity of the membrane allows the RBC to perform its biological role of oxygen transport and to deform reversibly during flow. 77 The final class of disorders is whereby a key enzyme involved in the glycolysis or the pentose pathway has an altered structure, thereby impairing the ability of the RBC to withstand oxidative stress conditions. 77 Examples of this condition include G6PD deficiency and pyruvate kinase (PK) deficiency. 141 Most RBC diseases are detrimental if a person is homozygous for the defect, however, heterozygosity for disease-causing alleles of several RBC disorders have been shown to confer protection against severe malaria. 1.4.1.1 Hemoglobin mutations Sickle cell anemia Sickle cell anemia (HbS) is an inherited disorder of hemoglobin and is due to a single nucleotide substitution of ϐ6 valine for glutamic acid in the ϐ-globin gene (known as ϐ s -globin). 15 In homozygous sickle cell anemia (HbS), both alleles of the ϐ s -globin are inherited resulting in a mutant hemoglobin. 139 The life span of the RBC in HbS is significantly shortened as sickle cells are usually cleared after about 10-20 days. 142 During the deoxygenation stage, the HbS polymerizes causing the RBCs to become abnormally rigid and non deformable leading to an irregular sickled shape. 143 Sickled Hb has the ability to undergo auto-oxidation in the presence of oxygen to produce superoxide and hydroxyl radicals in RBCs. In addition to the increase of reactive oxygen species, there is also a decrease in antioxidant enzymes and oxygen radical scavengers (e.g., glutathione peroxidase and superoxide dismutase) which further leads to abnormalities in the RBC membrane. 144-148 HbS exhibits accelerated auto oxidation at a rate of

38 1.7 times that of normal HbA. 149,150 In people who are heterozygotic and only have one sickle gene, this trait is referred to as HbAS or sickle trait. Sickle cell trait it usually regarded as a benign condition and rarely has complications. This is due to the fact that 55-75% of the hemoglobin in circulation is normal, and hemoglobin concentrations and other indices including RBC life span are normal. 16,151 HbS was one of the first of the structural hemoglobin variants to be associated with malaria protection. It has been estimated that 300 million people worldwide carry the sickle cell trait, with the highest concentration in the African region where P. falciparum is present. 152 In sub- Saharan Africa, there is an allelic frequency of the HbS gene of 0.15, suggesting that 30% of people carry the sickle cell trait. 153 This apparent natural selection for the sickle gene is referred to as balanced polymorphism in that the polymorphic trait is relatively protected from P. falciparum. It has been shown that heterozygotes (HbAS) are at least 90% protected from developing severe and lethal malaria, but are not protected from asymptomatic parasitemia. 154,155 Thalassemia Thalassaemia genes are widely distributed in the world but are found most often among people in the Mediterranean, the Middle East, and Southern Asia. Thalassaemia is an inherited autosomal recessive blood disorder where the genetic defect results in the reduction of one of the globin chains that make hemoglobin. There are two primary types of thalassaemia: α- and ϐ- thalassaemia. Both are caused by more than 150 different mutations, most of which are single nucleotide substitutions. 13 Both of these disorders involve the loss of either the α- or ϐ-chain of the molecule. 140 Some mutations completely eliminate the α-or ϐ-chain, whereas others dampen, but do not abolish globin synthesis. J.B.S Haldane was the first to propose that the high gene frequency of thalassaemia in certain populations reflected protection against severe disease by

39 heterozygous individuals. 13 It is believed that the potentially lethal thalassaemia gene is retained in the population because it provides some protection from malaria in the heterozygous state. Case controlled studies in Papua, New Guinea and Africa have shown that the homozygous and heterozygous states for ϐ-thalassaemia both protect against severe disease, where as α- thalassaemias offer approximately 50% reduction against severe malaria anemia. 17,156 1.4.1.2 RBC enzymes G6PD deficiency G6PD deficiency is the most common human RBC enzyme deficiency. G6PD provides RBCs with crucial protection against oxidant damage. G6PD converts glucose-6-phosphate (the initial product of glycolysis) into 6-phosphogluconate (6PD), while converting NADP to NADPH which is essential for the NADPH-dependent enzyme methemoglobin reductase. 13 Although G6PD deficiency is caused by point mutations that reduce enzymatic activity there are a few mutants where enzymatic activity is normal or even enhanced. There is a striking overlap between the areas where G6PD deficiency is common and where P. falciparum malaria is endemic, providing circumstantial evidence that G6PD deficiency confers protection against severe malaria. 157 There have been a number of case-control studies which examined the relationship between G6PD-deficient individuals and protection from malaria. 8,14,157 Amongst these studies is the study conducted by Ruwende et al in 1998, who demonstrated that there is a level of protection of 50% against severe malaria in heterozygous females and hemizygous males. 158 1.4.1.3 RBC membrane disorders RBC Duffy Negative

40 The Duffy negative phenotype is caused by a single nucleotide substitution polymorphism in the promoter region of the gene for the Duffy antigen. The Duffy antigen has been identified as a chemokine receptor as well as an essential receptor for P. vivax merozoites. 159,160 Individuals that do not express the Duffy antigen are protected from P. vivax. Barnwell et al showed that P. vivax merozoites are incapable of invading Duffy negative RBCs. 55 South Asian Ovalocytosis A deletion in the structural protein band 3 results in a condition known as South Asian ovalocytosis (SAO). This condition is only present in a heterozygous form, as no infants homozygous for this band 3 gene deletion survive intrauterine development. 161 This polymorphism is highly prevalent in Melanesian populations where P. falciparum is prevalent. 162 It has been shown that carriers of this trait are at a reduced risk of infection with P. falciparum 163, 19 due to impaired invasion and multiplication. 1.4.2 Inherited disease specific mechanisms of protection Based on studies of blood disorders such as those noted above, it is evident that genetic factors of the host contribute to the variability and disease complexity of P. falciparum and that there is also a relationship between polymorphic RBCs and protection from disease severity of P. falciparum malaria. Carriers of HbAS, α- and ϐ- thalassemia, G6PD deficiency, Duffy negative and ovalocytosis are all protected from developing severe malaria and are at a survival advantage. However, there is no clear consensus regarding how hemoglobinopathies and defective RBC enzymes and membranes are able to confer protection from malaria. Each RBC polymorphism comes with its own pathology; HbS causes auto oxidation of the hemoglobin, ϐ-thalassemia

41 causes ineffective erythropoiesis, α-thalassemia is responsible for a mild hemolytic state, G6PD deficiency causes accelerated oxidative membrane damage, and SAO causes sodium and potassium to leak out of the cell. 87,140,143,164 A number of mechanisms of protection have been proposed, including decreased invasion and growth of P. falciparum malaria in variant RBCs, decreased rosetting or cytoadherence, and increased removal of infected RBCs by the innate immune system. 1.4.2.1 Invasion and growth The process of merozoite invasion and growth involves a complex sequence of events that includes a number of specific interactions and requires a specific environment to grow. Any deviation from the normal human RBC might result in inhibition of P. falciparum to invade or grow within the RBC. Studies on the invasion and maturation of P. falciparum into RBCs carrying the sickle cell, thalassaemia, G6PD-deficiency and SAO trait have been controversial. There have been reports of decreased growth of P. falciparum malaria in HbAS, G6PDdeficiency, and ϐ-thalassemia RBCs. 165-169 Studies have suggested that oxygen radicals formed in RBCs with the sickle cell trait or are G6PD deficient, slow growth of P. falciparum as sickle trait and G6PD deficient RBCs produce higher levels of the superoxide anion and hydrogen peroxide than normal RBCs. 168,170 Studies showing inhibition of maturation of P. falciparum in RBCs with the sickle cell trait were incubated at low oxygen tension which may not represent an appropriate experimental condition since RBCs from those with the sickle trait do not sickle at normal oxygen tensions. The growth conditions in these studies also appear to have been inadequate, as these cells are sensitive to nutrient deprivation and were likely dehydrated even in the non-parasitized state. 171,172 Better designed studies found no difference in parasite growth in HbAS, G6PD deficient, or ϐ-thalassemia cells. 68,168,173 SAO RBCs have been shown to resist invasion by malarial parasites in vitro. This has been attributed to greater rigidity of ovalocytic

42 RBCs. 174,175 Studies examining the invasion of other species of Plasmodium into RBC variants have consistently reported that P. vivax malaria cannot invade into RBCs homozygous for a null Duffy allele, as Duffy negative cells are missing an essential antigen (Fy6) for invasion. 160 1.4.2.2 Cytoadherence and rosetting Decreased cytoadherence, (the process of infected RBC adhesion to the microvascular endothelium), and decreased rosetting (binding of infected RBCs to uninfected RBCs) of polymorphic RBCs have been suggested as potential mechanisms of protection from severe malaria. It has been reported that infected HbAS cells display impaired cytoadherence due to reduced PfEMP-1 expression on the RBC surface. 176 Udomsangpetch et al found that there was a significant decrease in the levels of P. falciparum antigens associated with the membrane in infected thalassemic RBCs, and as a result, cytoadherence and rosetting of thalassemic RBCs was greatly reduced. 177 1.4.2.3 Clearance of infected polymorphic RBCs Polymorphic RBCs already have increased susceptibility to RBC membrane damage, targeting them for clearance. The interaction between the parasite within the defective RBC and the host innate immune system represents a final putative mechanism of protection to malaria. Variant RBCs senesce at a faster rate than normal RBCs, and are marked for phagocytosis at a greater rate. 68 A number of studies have compared the phagocytosis of infected polymorphic RBCs to infected normal controls. 68,178-181 There is an growing body of evidence suggesting that infected variant RBCs are cleared more efficiently by the innate immune system than infected normal RBCs. It has been consistently shown that there is increased phagocytosis of malaria infected HbAS, G6PD deficient and thalassemic RBCs than infected normal controls. 66,68,178,180,182 Ayi et al demonstrated accelerated removal of ring stage infected RBCs carrying genes for HbAS,

43 G6PD deficiency and α-and ϐ-thalassaemia. 68,180 The RBC surface of non-infected variant RBCs possess abnormalities that are already recognized by pattern recognition receptors and other components of the innate immune system, and these are likely to be further modified by malarial infection, thereby enhancing the possibility of recognition and removal by macrophages. In another study, it was demonstrated that, compared to the baseline of non-infected controls, noninfected RBCs from individuals with HbAS and ϐ-thalassemia had elevated membrane bound hemichromes, aggregated band 3, bound IgG and C3b. 68 It was subsequently demonstrated that ring stage parasitized HbAS and ϐ-thalassemia RBCs displayed even higher levels of aggregated band 3, bound IgG and C3b, resulting in a dramatic increase in phagocytosis. 68 Ring parasitized HbAS and thalassemia cells are subject to multiple forms of oxidative stress, the first as a result of the developing parasite and the second generated by the RBC modification itself. However, phagocytosis of trophozoite-infected normal RBCs and trophozoite infected mutant RBCs was not significantly different nor were the levels of hemichrome deposition, band 3 aggregation, IgG or C3b deposition. It is thought that the damage inflicted by the mature parasite is so significant that it will overshadow the baseline differences in normal and mutant RBCs as the damage caused by both the polymorphism and the parasite reaches a maximium. 68 The uptake of ring stage parasites vs schizont stage parasites has been shown to be mediated by different receptors. Ring stage-infected RBCs are phagocytosed similarly to senescent or oxidative damaged RBCs, and IgG and complement signal for phagocytosis. However, schizonts are phagocytosed more intensely by IgG, but also have additional markers such as exposure of PS 66. A similar study by Cappadoro et al found that there was increased phagocytosis of infected G6PD deficient cells when compared to normal infected cells. All three of these conditions share a common trait in that they are characterized by increased generation of oxygen radicals or have a decreased ability to survive oxidative damage. 140,150,180

44 1.4.3 ABO polymorphism and Plasmodium falciparum malaria The ABO blood types were the first genetic polymorphisms found to vary significantly among human populations. In the early 1900 s, a number of studies sought to establish an association of ABO blood types with certain diseases (Section 1.3.5) One of the largest of these studies was conducted in Russia by Rubaschkin and Leisermann in 1929. 183 A χ 2 test indicated a significant association between malaria and the ABO blood types, with AB having the highest frequency of malaria patients. 183,184 However, a subsequent study by Parr et al in 1930 showed no difference between ABO blood types and the frequency of malaria. 183 The results of both studies were considered to be ambiguous, and further research to establish a relationship between ABO and P. falciparum was temporarily abandoned. 184 However, over the last four decades, the relationship between the ABO blood type system and P. falciparum disease severity has been revisited. While the results of earlier studies (1926-1998) were contradictory and heterogeneous in study design, recent studies have consistently found an association between ABO blood types and disease severity of P. falciparum malaria. The study by Rubaschkin and Leisermann, the associations noted between the ABO blood type and a variety of infectious diseases, and the discovery of the protective effect of Duffy negative phenotypes against P. vivax infection, renewed the interest of the research community in the association between ABO and malaria. 125,160,185 Renewed interest in studying the relationship between ABO and malaria emerged in the late 1970 s in Nigeria 21,186 and led to further studies between 1979-1998. These studies used parasitemia and incidence of malaria infection as endpoints. In twelve cross-sectional studies and five case-control studies, parasitemia and infection by P. falciparum were used as end point criteria. Ten cross sectional studies 21,187-195 (completed in Africa and Columbia) and four case control studies 186,196-198 (conducted in Brazil, Africa, the UK and India) did not find a significant interaction between ABO and malaria risk

45 and overall parasitemia. Only two cross sectional studies 199,200 and one case control 201 study found an interaction between parasitemia levels and infection risk of malaria and ABO blood type. All three of these studies were completed in India and these found that fewer blood type O individuals were infected with P. falciparum and had lower parasitemia. From these studies, it appeared that parasitemia and incidence of P. falciparum malaria were not associated with ABO blood types. It was argued that parasitemia and incidence of malaria were not good predictors of disease outcome in P. falciparum infection. In 1998, investigators began using clinical severity, rather than incidence or parasitemia as end points. In five well-designed studies relating ABO to the clinical severity, blood type A was found to be more prevalent in severe malaria and blood type O was underrepresented. 202-206 Severe malaria was defined according the World Health Organization criteria including: unarounsable coma (Blantyre com score of 2 with other causes of coma excluded), severe anemia (< 5g/dl), neurological impairments, repeated seizures, evidence of hepatic and/or renal failure. 8 Fischer et al compared 209 mild malaria outpatients to 280 severe malaria inpatients in Zimbabwe. 204 They focused on the relationship between severe anemia and risk of coma. Infected blood type A had overall lower hemoglobin levels than blood type O (p<0.02), and a higher risk of coma when compared to non-type A(non-A= 3%, A=9%, p=0.008). A case control study comparing children with severe malaria and uncomplicated controls was conducted in Gabon by Lell et al in 1998. Blood type, amongst other RBC polymorphisms, was determined in 100 severe malaria cases and 100 mild malaria cases. 3 They found blood type A was significantly associated with severe malaria (O.R 3.0; p<0.01). 203 Pathirana et al analyzed the ABO blood types in 243 Sri Lankan patients infected with P. falciparum malaria. 205 In total, there were 163 patients diagnosed with uncomplicated malaria, 80 with severe malaria and 65 with severe non-malarial illness. The frequencies of the occurrence of each ABO blood type

46 were then compared to a population control. Patients with severe malaria were three times more likely to be of blood type A versus O (p=0.005) and two times more likely to be of type B than O (p=0.014). 205 Another well-designed study was conducted by Rowe et al in 2007 in Mali. This study assessed the ABO blood type status of 567 blood samples of Malian children. They compared the distribution of ABO blood type frequencies in severe malaria cases to healthy controls (N=124 matched triplicates with severe, uncomplicated malaria and healthy controls). They also compared ABO frequencies in 65 matched triplicates in non-severe hyper-parasitemia cases to uncomplicated and healthy controls. The distribution of the ABO frequencies in severe malaria and the matched uncomplicated and healthy controls showed a decrease in the frequency of blood type O in severe malaria cases when compared to uncomplicated and healthy controls (SM:21, UM:44.4, HC:45.2%), and an increase of non-o blood types (SM:79.1,UM:55.6, HC:54.8% ). It was found that blood type O conferred significant protection against severe malaria when compared with the non-o blood types (O.R 0.34, p=0.0003). 206 In contrast to these results, they found the frequency of blood type O in non-severe hyperparasitemia cases did not differ from their matched controls (O.R 1.0, p=1.0), confirming from original studies that ABO is not related to overall parasitemia levels. These studies all specifically compared ABO phenotype to disease severity of P. falciparum malaria but did not discriminate between genotypes, for example AO heterozygotes vs. AA homozygotes. 206 The most recent study examining the relationship between ABO and malaria, took a novel approach by genotyping the ABO gene and then comparing ABO genotype frequencies to disease severity. 202 This was the largest of all the ABO association studies using a sample population of >9000 individuals across three African populations. When comparing ABO phenotypes alone, this study strongly supported the hypothesis that individuals with blood type O are protected from severe malaria. Blood type A was associated with increased risk of severe

47 disease (OR 1.33, p=0.0007), however they found AB individuals had the greatest risk of severe disease (OR 1.59, p=0.006). The contribution of this study to the field is that they also compared the genotype of ABO individuals to disease severity. 202 There was no significant difference in the risk estimate for severe malaria between AO and AA genotypes and the same was true for BB vs. BO. Placental malaria and ABO The association between ABO blood types and placental malaria is unclear; two studies have shown risk of active infection between blood types A and O is parity dependent, and the third showed no difference in past placental infection in women with blood type O vs. non-o. 207-209 In these studies they compared parity, active placental infection, and birth outcomes. The first study in Gambia by Loscertales et al found there was a significant difference in the birth weight of children born to type O mothers when compared to non-o mothers (O:2893±362 g, A: 2639±323 g, p=0.04). The authors found parity-related susceptibility to P. falciparum placental infection associated with O phenotype. They found increased prevalence of P. falciparum placental malaria infection in primigravid women but a reduced prevalence in multigravid mothers with blood type O. The following year, the same group replicated these findings in Malawian women. 207,209 Subsequently, another group of researchers in the Sudan, investigated the role of ABO blood types on pregnancy in mothers and placental infection. They found that blood type O was associated with past placental malaria infection in all gravities. They found a significant difference in hemoglobin levels between O vs. non-o blood types (11.8 vs. 10.9, respectively, p<0.05), but did not find a difference in birth parameters such as birth weight. 208 It remains controversial if blood type is associated with significant differences in birth weight, however both studies found a higher hemoglobin level in mothers of blood type O. 207-209

48 Collectively these studies analyzing disease severity and ABO blood types, together with those examining the association between ABO and placental malaria, have demonstrated compelling evidence of a link between ABO polymorphism and P. falciparum disease severity. It appears that blood type O protects from progression to severe disease and may offer a survival advantage. More studies are needed to affirm the link between disease severity and ABO, including studies which include mortality as the clinical end point. Examination of the direct mechanism(s) by which blood type O offers protection would validate the relationship between blood type O and disease severity of P. falciparum malaria. 1.4.4 Potential mechanisms of protection afforded by blood type O As mentioned, there is mounting evidence documenting an association between the ABO blood type and P. falciparum malaria. The evidence suggests that individuals with blood type O are as susceptible to falciparum infection as individuals with blood types A and B. However, it appears as though blood type O may be protected from progression to severe disease. The exact mechanism of protection is unclear, however it has been postulated that protection may be achieved through decreased parasite invasion and/or rosetting/cytoadherence of O RBCs compared to A and B RBCs. 1.4.4.1 Invasion and maturation Invasion of RBCs involves a number of specific receptor-ligand interactions, some of which are associated with ABO blood types,(e.g., attachment of MSP-1 and EBA-175 to band 3 and glycophorin A, both of which express ABO antigens). 210 Only one study has examined invasion of P. falciparum into A, B and O RBCs. 211 Chung et al reported an increase in the invasion of A 1 RBCs by P. falciparum when compared to invasion of A 2, O or B RBCs (p<0.05). This was shown in both sialic acid-dependent and acid-independent parasite strains. They further treated

49 A 1 RBCs with N-acetyl-galactosaminidase to remove the terminal GalNAc on blood type A 1. They demonstrated a significant reduction in the invasion of A 1 RBCs but not in O treated cells. Although these findings support the hypothesis that ABO blood types and P. falciparum are associated, however this type of comparative invasion assay has only been reported once and needs to be repeated to establish its validity. 1.4.4.2 Rosetting and sequestration Rosetting and cytoadherence have been implicated in the pathogenesis of cerebral malaria. It is thought that the strong binding of infected RBCs to both uninfected RBCs and endothelial cells, may block or impair the perfusion of blood through the microvasculature. This may cause tissue ischemia and endothelial cell apoptosis, ultimately leading to severe and cerebral malaria. 212 Sequestration of infected RBCs may occur within the vascular beds of the brain and other vital organs. 42 Studies have shown a decrease of rosetting and cytoadherence in blood isolated from infected individuals with polymorphic blood disorders such as thalassaemia and sickle cell (Section 1.4.2.2) It has also been shown that the DBL-1α region on the parasitic virulence antigen PfEMP-1 demonstrates lectin like properties and can bind to A and B antigens. 44,56 Rosetting Previous studies have shown that rosetting is significantly reduced in blood type O RBCs compared to A, B and AB. 56,206,213-216 In each of these studies, blood type O RBCs were able to form rosettes, however they were smaller, had a lower rosetting rate and were more easily disrupted. The A and B trisaccarchides are thought to act as receptors on uninfected RBCs and therefore are able to bind to schizont-infected RBCs. 56 Rowe et al demonstrated that when soluble A antigen was added to culture, rosette formation was inhibited. 56 Furthermore, treatment of blood type A with α-n acetyl- galactosaminidase decreased their rosetting

50 capacity. 56 There are two important concerns with respect to the rosetting hypothesis; rosetting has yet to be shown to occur in vivo and it is strain specific as there are A and B blood type preferring strains (as reviewed in Uneke et al, 2007). 217 Overall, the evidence implicating a relationship between disease severity, ABO blood type, and rosetting is strong. One study which provides compelling support for the rosetting hypothesis was undertaken by Rowe and colleagues. They examined the relationship between the ABO blood type and P. falciparum infection in Mali. They found rosette frequencies were significantly lower in parasite isolates from patients with blood type O compared to non-o blood types, and that non-o blood types with a >5% rosette frequency were at a greater risk of severe malaria ( OR 15.23 p<0.0001). 206 It may be possible that rosetting is one mechanism that confers protection associated with blood type O. Sequestration Another putative mechanism is increased adhesion of A- infected RBCs to the endothelium (sequestration). A number of soluble adhesion molecules have been associated with ABO blood type. Circulating levels of the procoagulant von Willebrand factor (vwf) have been shown to vary between ABO blood types, with type O having lower levels of vwf than non-o types. 218 This is most likely due to the fact that vwf of type A individuals seems to be more resistant to the proteolysis by ADAMTS13 than O individuals. 219 It has also been reported that platelet CD36 also expresses A-antigen. 220 A model proposed by Cserti et al suggests that enhanced sequestration of infected and non infected blood type A to the endothelium is caused by a number of different cell types and associations. 3 As previously noted, ABO antigens are present on a number of different cell types, e.g., endothelial cells, and platelets. 221 Cserti et al explain that cytoadherence begins with the A or B uninfected RBC binding to an infected RBC through

51 A and B antigens and PfEMP-1. The infected cell further binds to the endothelium by either binding to A or B antigens present on the endothelial cells, or by binding to blood type antigens that are found on platelets or vwf. Therefore, because of a lack of lectin binding and interactions between the A and B antigens and PfEMP-1, cytoadhesion of blood type O fails. 1.4.4.3 Additional mechanisms of protection Preference of anopheline mosquitoes for specific blood types A number of studies have examined the relationship between the feeding habits of the Anopheles mosquito and a preference to feed on individuals of certain blood type. 222-224 Two different research groups have investigated the association between the attractiveness of individuals with certain ABO blood types to the biting habits of the Anopheles gambiae mosquito. These groups reported contradictory results. Wood et al found that the mosquitoes preferentially selected for hosts of blood type O (Average number of blood meals: O; 4.4 and A; 3.3, p<0.05). 223,224 A later study done by Bryan et al found that there was a preference for blood type A individuals (70% of all mosquitoes fed on blood type A, compared to the 24% that fed on blood type O, p value unknown). 222 A and B antigens on the surface of malaria It has been suggested that individuals of blood type O may have a selective advantage as they have both anti-a and anti-b antibodies in their serum. 125 This would be a potential advantage since certain micro-organisms appear to have antigenic determinants that resemble the A or B antigen. A number of studies have shown that P. falciparum malaria parasites share blood type A and (to a lesser extent) B antigens, and therefore would be better tolerated immunologically by individuals who are blood type A. 125,128,136,185,225 There is evidence that supports the presence of

52 A- and B- like determinants on P. falciparum because there is an increase in anti-a and anti-b titers in patients with chronic malaria when compared to healthy controls. Oliver-Gonzalex et al found that there was a 250-fold increase in anti-a antibodies in blood type O patients who had malaria, and an 32-fold increase in the anti-b antibodies. 185 Antibodies directed against A and B are IgM and IgG antibodies and therefore bind complement, activating the classical pathway. If P. falciparum does have some antigenic determinants that resemble A and B antigens, this would support the hypothesis that blood type O confers protection from developing severe malaria. Several studies over the past two decades have demonstrated a link between the ABO blood group system and severity of P.falciparum malaria. 186,190,192,197,198,202-206,226,227 A recent metaanalysis of data associating ABO and malaria, found an association between malaria severity and blood groups A and B while milder disease was associated with group O. 210 The mechanism of protection from malaria associated with blood group O has been the subject of much speculation. RBC polymorphisms, which are known to confer protection to clinically severe malaria, have been shown to confer protection by defective invasion, decreased rosetting, and increased clearance. Mechanisms investigated to understand why blood group O is protected, have been based on similar RBC polymorphisms mechanisms. While the mechanism underlying the protective effects of the O blood type remains unclear, the O blood type has been previously associated with decreased rosetting when compared to A and B infected RBCs and decreased invasion by P. falciparum. 56,177,206,211,215,228,229 Studies have reported increased phagocytic uptake of P. falciparum parasitized polymorphic RBCs. 68,69,178-180,182 Based upon these findings, we hypothesized that parasitized blood group O RBCs are phagocytosed more efficiently than A and B parasitized RBCs.

53 SECTION 5- AIMS AND HYPOTHESIS The purpose of this study was to examine the mechanistic basis by which blood type O may contribute to protection against severe malaria including alterations in RBC invasion and maturation as well as enhanced clearance by effector cells of the innate immune system. 1. Based on the interaction between molecules on the merozoite and the RBC (MSP-1:Band 3, EBA 175:GPA), we hypothesized that P. falciparum merozoites would preferentially invade blood type A RBCs. 2. Where enhanced phagocytosis of polymorphic infected RBCs occurred, we hypothesized that infected type O RBCs may be more efficiently cleared by the innate immune system than type A and B RBCS. The aims of this project were to identify novel mechanisms by which blood type O might confer protection from severe malaria, to confirm that mechanism in vivo, and determine whether the amount of O antigen present on the RBC correlates with the observed effect. The present research study showed that, when compared to type A and B infected RBCs, there was a significant increase in the phagocytosis of schizont-infected O RBCs. We conclude that enhanced clearance of infected O RBCs may represent an additional putative mechanism by which blood type O may contribute to protection against severe malaria.

54 PART II MATERIALS AND METHODS

55 2.1 Reagents Endotoxin free RPMI 1640 and gentamicin were purchased from Invitrogen Life Technologies (Burlington, ON, Canada). Human AB serum was purchased from Wisent Inc (St-Bruno, Quebec, Canada). Diff Quik staining kit and fetal calf serum (FCS) were both purchased from Fisher Scientific (Ottawa, ON, Canada). Both FCS and human AB serum were heat-inactivated for 30 minutes at 55 C before use to remove complement activity. Alanine was purchased from Sigma Aldrich (Oakville, Ontario, Canada). Mycoplasma removal agent was purchased from MP Biochemical (Solon, Ohio, USA). Ficoll-Paque and Percoll was purchased from GE Healthcare (Baie D Urfé, Québec, Canada). NOVACLONE blood grouping reagent was purchased from Dominion Biologicals Ltd (Dartmouth, Nova Scotia, Canada). Mice Male or female C57BL/6 mice aged 6-10 weeks (Charles River, Hollister, CA) were maintained under pathogen-free conditions, and used for experiments. All experimental procedures involving mice were conducted in accordance with the animal protocol approved by the University of Toronto Animal Use Committee or the University Health Network Animal Care Committee. Ethics Experiments involving mice were approved by the University of Toronto Animal Use Committee or the University Health Network Animal Care Committee, as appropriate, and performed in accordance with current institutional regulations.

56 P. falciparum culture P.falciparum strains ITG and 3D7 (mycoplasma-free) were maintained in continuous culture as previously described by Trager and Jensen. 230 Strains were maintained at 2% hematocrit in R- 10G media (RPMI 1640 supplemented with 10% heat-inactivated human serum, 1.8 g/l Na 2 CO 3, 6g/l HEPES, 25 mg/l gentamicin and 1.35 mg/l hypoxanthine, ph 7.4.) On a daily basis media was changed, and flasks were infused with a slow current of a gas mixture containing 7% CO 2, 5% O 2 and 88% N 2. 2.2 Methods RBC and serum isolation Whole blood was donated from healthy non-immune individuals aged 20-54 years and collected in BD Vacutainer Glass tubes containing ACD solution as anti-coagulant. RBCs were separated from whole blood by as previously described. 231 Briefly, whole blood was layered on an 80% Percoll gradient [80% (w/v) Percoll, 6% (w/v) mannitol, 10 mm glucose and 20% (v/v) PBS 10X] and spun for thirty minutes at 3000 RPM at 24 º C. The isolated RBCs were washed three times in R-0G media (RPMI 1640 medium supplemented with 10mM glucose and 10g/L gentamicin) and re-suspended in parasite growth medium R-10G media (RPMI 1640 containing 20 mm glucose, 2 mm glutamine, 6 g/l Hepes, 2 g NaHCO 3, 10 g/l gentamicin, 10% human AB serum and 1.35 mg/l hypoxanthine, ph 7.3) to approximately 20% hematocrit. Blood for serum was collected in BD Vacutainer Glass tubes with no additive. Serum was separated from RBCs by centrifugation at 1500 RPM for five minutes at 24ºC, and 200 μl aliquots were stored at -20 C for future use. To ensure complement activity was maintained, each aliquot was thawed only once and discarded after use.

57 Blood typing Blood samples were typed for ABO blood type by standard hemagglutination techniques according to manufacturer s instructions. One drop of DBL anti-a, anti-b, and anti-a,b were added to labelled test tubes. Serum free, washed RBCs were re-suspended in PBS to a final hematocrit of 10%. One drop of the RBC suspension was added to each test tube. Each tube was mixed and then centrifuged for 30 seconds at 3400 rpm. The test tubes were then gently shaken and the cells were examined for agglutination. For blood type A samples, cells were further typed using A 1 lectin (Dolichos biflorus) to further confirm presence or absence of the A1 antigen. Monocyte isolation Human monocytes were isolated and purified from the peripheral blood of healthy donors (nonimmune A or O type) and plated on glass cover slips in 24-well polystyrene plates as previously described. 83,232,233 Briefly, blood was drawn from donors using BD Vacutainers containing sodium heparin as anticoagulant. Blood was immediately mixed with warm PBS in a 1:1 ratio, and then carefully layered on Ficoll (25ml/15ml) and centrifuged at 1800 RPM for 30 minutes at 21ºC. The peripheral blood mononuclear cell (PBMC) layer was carefully removed using a Pasteur pipette and washed three times with cold PBS. The PBMCs were then re-suspended in R- 10G FCS media (RPMI 1640 medium containing L-glutamine and HEPES, supplemented with 10 % heat-inactivated FCS and 25 mg/l gentamicin). The final volume and total number of cells were adjusted to give a final concentration of 1.25x10 6 PBMCs/150μl/well. Each coverslip was plated with 150 μl of suspension and incubated for 5 days at 5% CO 2 in a humidified incubator at 37ºC.

58 Re-infection of ABO RBCs P. falciparum cultures of clones ITG and 3D7 were purified on a Percoll-mannitol gradient (3ml/6ml). Isolated schizonts were washed twice and then used to infect fresh non-immune RBCs of different blood types (A, B and O). Final hematocrit and parasitemia used varied depending on the experiment and will be discussed in further detail in each section. The newly infected RBCs were incubated with R-10G human AB serum at 37ºC. 58,230 Analysis of invasion and maturation To assess parasite invasion and maturation, purified schizonts at 0.5% parasitemia were mixed with non-immune A, B or O RBCs at a hematocrit of 2%, as previously described by Ayi et al. 68 Slides were prepared by thin blood smears from cultures at 24 h and 72 h to assess invasion and at 48 h and 96 h to assess maturation. Slides were stained with Diff-Quik, and 1000 RBCs were examined microscopically. Percent parasitemia was determined as follows: [number of parasites number of total RBCs counted] x 100 in each field of view. Parasitized RBC preparation for phagocytosis assay Ring-stage and mature-stage parasites were obtained 24 hours and 48 hours, respectively, after rupture of schizonts. The ring culture was spun and re-suspended in alanine-tris solution (300 µm alanine and 0.15 µm tris, ph 7.4 at a ratio 1:18) and incubated for 5 minutes at 37 C, causing lysis of the mature-stage parasites while leaving the ring-stage parasites intact. 234 The culture was then washed and re-suspended in R-0G media, separated using Percoll-mannitol gradient, and centrifuged at 1800 RPM for 20 minutes at room temperature. The pellet containing uninfected RBCs and ring-stage parasites was then washed and incubated with autologous non-immune serum. The final parasitemia was 15-20%. Mature parasite cultures were synchronized with alanine for 15 minutes 24 hours after re-infection of ABO RBCs,

59 washed with the R-0G media, and re-incubated with the R-10G media containing the AB serum. Mature-stage parasites were obtained 48 hours after rupture, and were spun and incubated with autologous non-immune serum. Phagocytosis in vitro Parasitized and non-parasitized cultures were incubated with autologous non-immune serum for 30 minutes at 37 C in a ratio of 1:1 (culture/serum). Cell suspensions were then washed twice with the R-0G media, and re-suspended in 1 ml of R-0G media. An aliquot of 500 ul of each sample (parasitemia 12-20% ring-stage or 5% mature-stage) were incubated in each well containing approximately 1.25 x 10 5 macrophages adhered to the glass coverslip in culture plates at ratio of 20-40:1 (parasitized RBCs: monocyte-derived macrophages). The plates were rotated gently for 90 min (ring-stages) or 120 minutes (mature-stages) at 37 C in a 5% CO 2 humidified incubator. The ring stage phagocytosis assay was performed over a shorter period (45 minutes) to prevent complete degradation of the ring stage cytoplasm (to allow accurate counting of phagocytic uptake). After incubation, all RBCs that had not been internalized were removed by hypotonic lysis with cold water to prevent counting infected RBCs not phagocytosed. Macrophages were then fixed and stained using Diff-Quik. Phagocytosis was assessed by counting the number of internalized parasites in 250 macrophages. Values were expressed in percentage as: [number internalized parasites number of total macrophages counted] x 100. RBC preparation and phagocytosis assays were preformed as previously described, 58,68,235 and all the experiments were performed in duplicate, repeated at least three times with Plasmodium falciparum strain ITG and then confirmed using the 3D7 strain.

60 Phagocytosis in vivo To assess phagocytosis of infected A,B and O RBCs in vivo, 5x10 7 mature infected RBCs and non-infected RBCs were injected into the peritoneal cavity of C57BL/6 mice as previously described by Serghides et al. 235 Briefly, 48 hours after re-infection, mature stage parasites were isolated using a Percoll-mannitol gradient. The purified mature culture and the uninfected RBCs were then incubated with autologous human serum in a 1:1 ratio (culture/serum), washed using R-0G media, and then 5x10 7 opsonized mature infected RBCs and uninfected controls were injected into the peritoneal cavity of C57BL/6 mice. Three hours after injection, peritoneal cells were collected, and washed with RPMI-1640. The cells were then separated by centrifugation and re-suspended in 500 µl of R-0G media. 150 μl aliquots of peritoneal cells from each mouse were placed on a cover slip in a 24 well plate and allowed to adhere for 30 minutes. Cells on coverslips were treated by hypotonic lysis to remove RBCs and stained with Diff-quick. In addition, 200 µl of the suspension was lysed for 45 seconds with cold water, washed, cytospun at 800 RPM for 10 minutes and stained with Diff-quick and used for images. Images were acquired with an Olympus BX41 microscope and an Infinity2 camera at 100 x magnification. Statistical analysis Statistical analysis was performed with Graphpad Prism 4 software (San Diego, CA, USA). To confirm the normal distribution of data, all data sets were assessed using the Kolmogorov- Smirnov test. Data sets that displayed normal distribution were analyzed by t-test or one way ANOVA as appropriate, and data sets that did not display normal distribution were analyzed by Mann-Whitney rank sum test or Kruskall-Wallis test, as appropriate. A general linear model was used to analyze experiments with multiple independent variables (e.g., macrophage and RBC

61 type). To test the dose-dependent effect of the A antigen, we used linear regression analysis. Conditions for all experiments were performed in duplicate, and each experiment was repeated at least three times with strain ITG. Data are either shown as box plots representing the median, inter-quartile range and range or as bar graphs representing the mean and ±SD. Differences with a p<0.05 were considered statistically significant.

62 PART III RESULTS

63 P. falciparum does not preferentially invade or mature within blood type A, B or O RBCS. Several studies examining the protective nature of heterozygous and homozygous RBC polymorphisms from P. falciparum malaria have found impaired invasion and maturation in RBCs in vitro. This finding was observed in carriers of ovalocytosis, pyruvate kinase deficiency and in thalassaemia. 68,178-180 Similarly Chung et al. reported a decrease of P. falciparum invasion into O RBCs when compared to A 1. 211 Based on these results and the interactions between merozoites and RBCs (Section 1.1.2), we hypothesized that invasion of the A RBC would be increased compared to O. To test this hypothesis, we infected A, B and O RBCs and measured the parasitemia daily for two cycles of growth. As shown in Figure 7, there was no significant difference in the invasion of P. falciparum into A, B or O RBCs (p>0.05). 211 In addition, there was no significant difference in the maturation of P.falciparum within A, B and O RBCs (p>0.05). Datas shown represent three independent experiments using P. falciparum strain ITG and five different blood type A, two blood type B and five blood type O donors. To ensure these observations were generalizable to other parasite lines, we confirmed these find ings with the P. falciparum strain 3D7. Our data indicate that the parasitic invasion and growth in O RBCs is normal, suggesting that protection against malaria in blood type O individuals is not due to decreased invasion or maturation.

64 Figure 7. Similar invasion and growth of P. falciparum in A, B or O blood type RBCs. Healthy RBCs from blood types A 1, B and O donors were infected with P.falciparum strain ITG schizonts. The initial inoculum was adjusted to 0.5% parasitemia. Parasitemia for invasion cycles were assessed by thin blood smears at 24 and 72 hours after re-infection and 48 and 96 hours for maturation cycles. Parasitemia was determined by counting the number of infected RBCs in 1000 RBCs. Data shown represent three independent experiments (A:n=7, B:n=5, O:n=7). Box plots depict the median, interquartile range and range. There was no significant difference within each cycle of invasion and maturation between blood types A, B and O RBCs, p>0.05; Kruskal- Wallis test).

65 Human monocyte-derived macrophages phagocytose infected O RBCs more efficiently than infected A and B RBCs. Mounting evidence suggests that variant RBCs infected with P. falciparum malaria are cleared more rapidly than normal RBCs. 68,69,178-180,182 This has been proposed and examined in RBCs heterozygous for G6PD, sickle cell anemia, -thalassaemia and pyruvate kinase deficiency. 236 These studies compared the phagocytic uptake of polymorphic RBCs and normal RBCs infected with ring-stage or mature schizont-stage parasites. It was consistently observed that polymorphic RBCs are phagocytosed significantly more efficiently than normal RBCs, when infected with ring-stage parasites. A similar trend has been observed in the clearance of mature stage infected polymorphic RBCs. Based on these observations, we examined the phagocytic uptake of infected A, B and O RBCs. Infected A, B and O RBCs were incubated with autologous non-immune serum and then incubated with human monocyte-derived macrophages. We hypothesized that infected O type RBCs would be preferentially phagocytosed compared to type A and B RBCs. While there was a slight increase of phagocytosis of the ring stage infected O RBCs when compared to infected A (p=0.34) and B RBCs (p=1.16), the difference did not reach statistical significance (Figure 8). Schizont-parasitized RBCs are more susceptible to phagocytic uptake as they express higher levels of parasite antigens and are more susceptible to oxidative stress. 69,94 This prompted us to examine the uptake of schizont-stage parasitized A, B and O RBCs. The uptake of schizontparasitized O RBCs was approximately 2.5-fold higher than the uptake observed with the infected A (p=0.0008) and B (p=0.004) RBCs (Figure 9). To ensure that the differences in uptake were attributable to infection by P.falciparum, we used non-infected A, B and O RBCs as controls. There was no significant difference in the uptake of non-infected A, B and O RBCs (Figure 9).

66 Figure 8. Phagocytosis of ring stage infected A, B and O RBCs by human monocyte-derived macrophages. P. falciparum ring-infected ABO and non-infected ABO RBCs were incubated with autologous non-immune serum and incubated in a 40:1 ratio with human monocyte-derived macrophages for two hours. The phagocytic index for each RBC type was calculated by counting the number of internalized parasites in 250 macrophages by microscopy and normalized to the average phagocytic index of parasitized blood type A. Data represent pooled results from three independent experiments using P. falciparum strain ITG, using at least three different donors in each blood type (A:n=7, B:n=3, O:n=7). The bar graph represents the mean ± SD. There was no significant difference in the uptake of infected O RBCs when compared to A or B infected RBCs (p=0.34, p=1.16, respectively; Mann Whitney rank sum test with Bonferroni correction for multiple comparisons).

67 *** ** Figure 9. Increased phagocytosis of schizont infected O RBCs compared to infected A and B RBCs by human monocyte-derived macrophages. P. falciparum schizont infected ABO and non-infected ABO RBCs were incubated with autologous non-immune serum and then incubated with human monocyte-derived macrophages at a 20:1 ratio for two hours. The phagocytic index was calculated by counting the number of internalized parasites in 250 macrophages by microscopy. Data was then normalized to the average phagocytic index of parasitized blood type A. Data represent three independent experiments using P. falciparum strain ITG and each blood type is represented by at least three different donors, (A:n=9, B:n=4, O:n=9). The box plots represent the median, interquartile and complete range. There was an increase in the phagocytic uptake of infected O RBCs when compared to the phagocytic uptake of both A and B infected RBCs (***, p<0.001 and **, p<0.01, respectively; Mann Whitney rank sum test with Bonferroni correction for multiple comparisons).

68 Enhanced phagocytosis of schizont-infected O RBCs is independent of macrophage donor blood type. ABO antigens are expressed in multiple cell types and not exclusively on RBCs, hence they are frequently referred to as histo blood type antigens. 68,178,180,182,237 However, their expression status on platelets, lymphocytes and macrophages is uncertain. 221,238 In order to ensure that preferential uptake of infected O RBCs was not dependent on the blood type of the macrophage donor, we performed two phagocytosis assays in parallel, one with macrophages from an A blood type donor and the other from an O donor. When comparing the uptake of infected A and O RBCs, we observed that infected O RBCs had a higher phagocytic index (p=0.0001; two-way ANOVA), independently of the blood type of the macrophage donor (Figure 10). No difference was observed in the uptake of RBCs between macrophage donors of different ABO blood types (p=0.552; two-way ANOVA).

69 *** *** Figure 10. Schizont infected O RBCS are preferentially phagocytosed independent of macrophage donor blood type. Blood was simultaneously drawn from A and O blood donors and monocytes were isolated on a Percoll-mannitol gradient. Monocytes were then plated on glass coverslips and incubated at 37 C for 5 days. Blood donors were ABO blood typed by standard hemagglutinin techniques. Schizont-infected blood type A and O cultures were incubated with autologous non-immune serum and incubated with the isolated macrophages at a 20:1 ratio. The phagocytic index was calculated by counting the number of internalized parasites within 250 macrophages. Data represent three independent experiments using P. falciparum strain ITG and each blood type is represented by at least three different donors, (A:n=8, B:n=8). The box plots represent the median, interquartile and complete range. Data was normalized to the average phagocytic uptake of infected blood type A by macrophages isolated from A donors. Using two-way analysis of variance (ANOVA), the blood type of the infected cells was found to influence the uptake of infected A and O RBCs (***, p<0.001), whereas the blood type of the macrophage donor was found not to influence the uptake of infected A and O RBCs (p>0.05).

70 Phagocytosis of infected O RBCs is increased in C57BL/6 mice. To simulate phagocytosis by peripheral blood monocytes in vivo, Serghides et al developed a method of measuring phagocytosis of P. falciparum using resident monocytes found in the peritoneal cavity of C57BL/6 mice. Fifty million purified schizont-infected A, B and O RBCs that had been previously incubated with autologous human serum were injected into the peritoneal cavity of C57BL/6 mice. Based upon our previous findings, we hypothesized that the resident monocytes would clear the infected O RBCs at a greater rate than the infected A and B RBCs. As expected infected blood type O RBCs were phagocytosed more avidly than types A and B (Figure 11a). There was a threefold increase in the uptake of infected blood type O RBCs when compared to both infected blood types A and B RBCs (O vs. A p=0.03 and B p=0.04). This is illustrated in Figure 11b showing internalization of infected non-o and O RBCs by resident monocytes from C57BL/6 mice. By comparing panels A and B to O, one can observe a substantial difference in the uptake of infected O RBCs when compared to A and B as there is a significantly greater density of internalized infected O RBCs contained within the monocyte.

71 A) * * * ** * B) Non-O O Blood Type Figure 11. Peritoneal monocytes of C57BL/6 mice clear infected O RBCs more efficiently than infected A or B RBC. The peritoneal cavity of C57BL/6 mice were injected intraperitoneally with 5.0 x 10 7 purified mature stage parasites cultivated in blood donated from A, B and O blood type donors. Three hours after injection, resident monocytes were collected, washed, and plated on glass coverslips. A) The phagocytic index was calculated by counting the number of internalized parasites within 250 monocytes and then data were normalized to the average phagocytic index of infected A