TOWARDS A DETAILED UNDERSTANDING OF THE STORAGE LESION

and its consequences for in vivo survival following transfusion

ANDREAS HULT

Department of Community Medicine and Rehabilitation Department of Surgical and Perioperative Sciences Department of Integrative Medical Umeå 2015

Responsible publisher under swedish law: the Dean of the Medical Faculty This work is protected by the Swedish Copyright Legislation (Act 1960:729) ISBN: 978-91-7601-288-8 ISSN: 0346-6612 Cover illustration by MD, PhD Gustav Andersson. Copyright with artist. Elektronisk version tillgänglig på http://umu.diva-portal.org/ Printed by: Print & Media Umeå, Sweden 2015

Till Simon och Elvira

i

ii Table of Contents Table of Contents ...... iii Abstract ...... v Abbreviations ...... vii Enkel sammanfattning på svenska ...... ix List of original papers ...... xiii Introduction ...... 1 The leukocytes ...... 1 Innate ...... 2 Pattern recognition receptors (PRRs) ...... 3 Adaptive immunity ...... 4 Tolerance vs. Immunity ...... 5 The Red blood cell ...... 5 Function ...... 5 ...... 7 The life and death of RBCs ...... 8 Eryptosis ...... 9 Recognition and of RBCs ...... 9 Immune cells of the spleen ...... 11 RBC transfusion ...... 13 History ...... 13 Demographics ...... 14 Effects of RBC transfusion ...... 14 RBC storage conditions ...... 15 The storage lesion ...... 15 Risks associated with transfusion ...... 17 How to study RBCs ...... 18 Human studies ...... 18 Animal models ...... 18 Aims ...... 20 Materials and methods ...... 21 Human studies (paper 1) ...... 21 Donation and reinfusion ...... 21 Blood analysis ...... 21 Maximal aerobic capacity (VO2max) assessment ...... 22 Animal studies (paper 2 and 3) ...... 22 Labeling of RBCs ...... 22 The mouse model for RBC storage and transfusion ...... 23 In vitro characteristics of stored RBCs ...... 24 In vitro phagocytosis ...... 24 Ethics ...... 25

iii Statistical analysis ...... 25 Results ...... 26 Paper 1 ...... 26 Transfusion of cryopreserved RBCs leads to acute accelerated RBC breakdown ...... 26 Absent pro-inflammatory response in the acute phase after transfusion ...... 26 Additional results ...... 27 Paper 2 ...... 28 Stored human RBCs contain a subset of Ca2+-high cells ...... 28 Characteristics of eryptotic murine RBCs ...... 28 In vivo uptake of Ca2+-RBCs is located to the splenic marginal zone in a CD47- independent manner ...... 28 Additional results ...... 29 Paper 3 ...... 30 Characteristics of the murine model for storage and transfusion ...... 30 Phagocytosis of stored human and mouse RBCs is serum dependent and mediated by scavenger receptors ...... 30 Additional results ...... 31 Discrimination of RBC age fractions by in vivo biotinylation ...... 31 Role of the RBC age at the start of storage for their 24 h survival post transfusion ...... 32 Phthalates during storage ...... 32 CD47 on stored murine RBCs does not influence the rapid elimination that occurs during the first 24 h ...... 33 Discussion ...... 33 Frozen vs. liquid stored RBCs ...... 34 after transfusion? ...... 34 The mouse model for storage and transfusion ...... 35 The role of calcium uptake in the storage lesion ...... 36 Impact of storage on RBCs - Senescence, eryptosis, or both? ...... 37 How are stored RBCs presented to the upon transfusion? ... 38 CD47 in senescence, eryptosis and the storage lesion ...... 39 Scavenger receptor mediated phagocytosis of stored RBCs ...... 39 The role of serum in phagocytosis of stored cells ...... 40 Concluding remarks ...... 41 Conclusions ...... 41 Acknowledgements ...... 43 References ...... 46

iv Abstract

Red blood cells (RBCs) are vital for oxygen delivery to tissues and constitute the vast majority of all cells in blood. After leaving the red bone marrow as mature cells, RBCs have a lifespan of approximately 120 days before they are removed from the circulation by , mainly in the spleen and . RBC transfusion is a common therapy in modern healthcare. Major surgery, numerous cancer treatments and other, often lifesaving, interventions would be unthinkable without available blood supply. For this reason, hospitals store donated RBCs in blood banks.

The metabolic and structural changes that occur during prolonged storage of RBCs (the storage lesion) have been studied in detail in vitro and include oxidative , a reduction in glycolysis, increased membrane rigidity and shedding of microparticles from the RBC membrane. Stored RBCs share several features of senescent RBCs, but also with RBCs undergoing an apoptotic-like process called eryptosis. A consequence of the storage lesion is the fact that as much as 25% of stored RBCs could be rapidly removed from the circulation within 24 hours after transfusion. The mechanisms behind this rapid -mediated recognition and removal of stored RBCs, and its immunological consequences, remain largely unknown. Therefore, the aims of this thesis were to investigate if cryopreserved human RBCs induced an inflammatory response following autologous transfusion into healthy volunteers, and to further understand the mechanisms behind macrophage recognition of stored RBCs in vitro and in vivo.

Autologous transfusion of two units of cryopreserved RBCs into healthy human recipients was found to be associated with an increased extravascular RBC elimination already at 2 hours after transfusion. However, there were no signs of an increased production of any of the investigated pro-inflammatory , indicating that an increase in the destruction of RBCs per se did not induce an inflammatory response.

Eryptosis is a form of induced RBC death associated with an increased cytoplasmic Ca2+ uptake. We found that a subset of human RBCs increased their Ca2+ permeability during prolonged storage at +4°C. Using a murine model, to further understand how RBCs with an increased Ca2+ permeability were eliminated by phagocytic cells in the spleen, it was found that such RBCs were taken up by marginal zone macrophages and dendritic cells (DCs) in a manner distinct from that of naturally senescent RBCs. The DC population particularly efficient in this process expressed CD207 and are known for their ability to promote immunological tolerance. Eryptotic cell uptake was not regulated by the phagocytosis-inhibitory protein CD47 on the RBCs.

v To investigate how RBCs damaged during liquid storage are recognized and taken up by macrophages, a model to store and transfuse murine RBCs was developed. This storage model generated murine RBCs with several characteristics similar to that of stored human RBCs (i.e. loss of ATP, formation of RBC microparticles and rapid clearance of up to 35% of the RBCs during the first 24 h after transfusion). In vitro phagocytosis of human as well as murine stored RBCs was serum dependent and could be inhibited by blocking class A scavenger receptors using fucoidan or dextran sulphate.

In conclusion, the findings of this thesis contribute to further understanding how changes inflicted to RBCs during storage direct the fate of these cells in their interaction with cells of the immune system after transfusion. The observation of an increased Ca2+ permeability of stored RBCs, and the possible recognition of such cells by tolerance-promoting DCs, in combination with the findings that class A scavenger receptors and serum factors may mediate recognition of stored RBCs, may result in novel new directions of research within the field of transfusion medicine.

vi Abbreviations

2,3-DPG 2,3- Diphosphogylcerate Ab Antibody APC Antigen presenting cell AS Additive solution ATP Adenosine triphosphate CPD Citrate-Phosphate-Dextrose CTL Cytotoxic T- CTLA-4 Cytotoxic T-lymphocyte-associated protein 4 DC EPO Erythropoietin ERK-1/2 Extracellular signal-regulated kinase-1/2 FCS Follicular calf serum FcγR Fc gamma receptor FSC Forward scatter Hb Hemoglobin Hct Hematocrit HS Human serum Ig Immunoglobulin IL Interleukin LPS MARCO Macrophage receptor with collagenous structure MCP-1 chemoattractant protein 1 MHC Major Histocompatibility Complex MIP-1β Macrophage inflammatory protein 1 beta MP Microparticle MZ Marginal zone MZM Marginal zone macrophages NAbs Naturally occurring antibodies NK-cell NTBI Non- bound iron PAMP -associated molecular patterns P38-MAPK P38 mitogen-activated protein kinase pCO2 Partial pressure of carbon dioxide pO2 Partial pressure of oxygen PI3-Kinase phosphatidylinositide 3-kinase Poly(I:C) Polyinosinic polycytidylic acid PRR Pattern recognition receptor PS Phosphatidylserine RBC Red Blood Cell SAG-M Saline-Adenine-Glucose-Mannitol

vii SIGLEC Sialic acid-binding immunoglobulin-type SR Scavenger receptor TCR T-cell receptor TH-cell T-helper lymphocyte TLR Toll-like receptor TNF-α - alpha TReg-cell Regulatory T-lymphocyte TSP Thrombospondin VO2max Maximal aerobic capacity

viii Enkel sammanfattning på svenska

Mot en ökad förståelse för lagringspecifika förändringar hos röda blodkroppar, vilka kan påverka dess överlevnad efter transfusion

Röda blodkroppar (RBK) är celler som saknar cellkärna och många andra cellulära komponenter och vars primära syfte är att transportera syre från lungor till kroppens celler. De cirkulerar i c:a 120 dagar innan de har ådragit sig sådana åldersskador att de känns igen och bryts ner av immunförsvarets makrofager i lever och mjälte. Kunskapen om de åldersskador som leder till att cellerna tas upp och bryts ned är i dagsläget begränsad. Den normala livslängden hos RBK kan förkortas i situationer som skadar RBK, exempelvis sjukdom eller olika typer av läkemedel. Exakt hur makrofager känner igen skadade RBK är delvis oklart. Blodtransfusioner är en vanlig behandlingsform inom modern sjukvård, och direkt avgörande för att kunna utföra stora operationer, vid flertalet cancerbehandlingar samt för att rädda traumapatienter med stora blödningar. För att tillgodose behovet på blod vid dessa ingrepp tappas blodgivare på blod som separeras i olika komponenter (vanligtvis RBK, blodplättar och plasma) och lagras i +4°C vid blodcentraler i väntan på transfusion. Det är sedan länge känt att under lagringstiden, som i dagsläget är maximalt 42 dagar, ådrar sig RBK en rad skador. Detta leder till att en substantiell andel av lagrade RBK (upp till 25%) försvinner kort efter transfusion då de känns igen av mottagarens makrofager. De möjliga riskerna för mottagaren som kan vara kopplat till detta är i dagsläget omdebatterat i den vetenskapliga litteraturen. Även om många lagringsspecifika skador har identifierats hos RBK så är mekanismerna varmed makrofager känner igen och bryter ner vissa RBK efter transfusion fortfarande mycket oklara.

Syftet med denna avhandling har därför varit att öka förståelsen för de igenkänningsmekanismer som immunförsvarets celler använder vid kontakt med RBK, samt hur immunförsvaret reagerar på denna igenkänning. Initialt studeras detta i transfusionsförsök utförda på frivilliga försökspersoner. Det finns dock inneboende etiska problem med att i tillräckligt stor detalj studera dessa förlopp i människokroppen, vilket gör djurmodeller och studier i cellkulturer till en mer framkomlig väg. Vi har därför satt upp en musmodell för lagring och transfusion av RBK.

I det första delarbetet undersöktes vilken effekt transfusion av två blodpåsar med fryslagrade RBK hade på inflammation i det akuta skedet (2 timmar) samt två dygn efter transfusion. Vi fann en ökad makrofagmedierad

ix nedbrytning av RBK i akutskedet efter transfusion, men att detta inte ledde till mätbara nivåer av inflammatoriska substanser i blodet.

Skadade (även kallade eryptotiska) RBK kännetecknas av ett ökat upptag av kalcium, vilket följs av att dessa celler snabbt tas upp av makrofager. Vi observerade att en liten andel humana RBK vid lagring får förhöjda kalciumnivåer, vilket var relaterat till en lång lagringstid. Delarbete två utforskade med hjälp av en musmodell hur sådana skadade RBK känns igen av immunförsvarets celler. Det visade sig att eryptotiska RBK togs upp av marginalzonsmakrofager och dendritiska celler (dessa celler är viktiga för reglering av immunförsvaret) i mjälten. Detta upptag liknar mycket hur uttjänta kroppsegna celler (apoptotiska celler) tas upp, men skiljer sig från upptag av normalt åldrade RBK. Upptag av apoptotiska celler motverkar som regel inflammation och hjälper snarare till att förhindra immunförsvaret från att attackera sina egna celler och vävnader. Då en liten del av lagrade RBK påvisade eryptotiska drag, öppnas möjligheten att dessa skadade celler faktiskt kan ha en positiv inverkan på mottagaren av en blodtransfusion. Att eryptos kan utgöra en del av lagringsskadan stärks av vårt fynd att upptag av lagrade RBK i mjälten delar likheter med både det som ses hos normalt åldrade och eryptotiska RBK. Även om CD47 är ett cellyteprotein som kan förhindra upptag av normala värdceller, så fann vi att varken eryptotiska eller lagrade cellers upptag påverkades av mängden CD47 på dessa cellers yta.

Det tredje delarbetet visade att det finns anmärkningsvärda likheter varmed lagrade RBK från mus och människa kan tas upp av makrofager, något som stärker relevansen för upptäckter i vår musmodell. När färska RBK från mus och människa blandades med musmakrofager skedde inget upptag. Däremot så togs lagrade RBK upp oberoende om de kom från mus eller människa. Makrofagernas upptag visade sig vara serumberoende, vilket indikerar att det finns någon komponent i serum som krävs för makrofagernas igenkänning av lagrade celler. Vidare kunde vi förhindra upptaget genom att blockera en viss typ av receptorer (klass-A scavenger receptorer) på makrofagernas yta. Detta tyder på att lagrade RBK känns igen med hjälp av dessa receptorer hos makrofagerna, samt att upptaget involverar någon serumkomponent.

Sammanfattningsvis så har detta avhandlingsarbete påbörjat en kartläggning av interaktionen mellan lagrade RBK och kroppens immunceller, vilket medfört en utökad förståelse för de skador som uppkommer hos RBK under lagring och konsekvenserna av detta efter återinförsel av dessa celler i blodet. Vi har funnit att ökad nedbrytning av RBK efter transfusion i sig inte leder till inflammation. Vidare har avhandlingsarbetet kunnat koppla en del

x av lagringskadorna som uppstår till så kallad eryptos, vilket skulle kunna innebära att vissa lagringsskadade RBK kan ha en dämpande effekt på kroppens immunförsvar. Avslutningsvis tyder arbetet på att makrofagers upptag av lagrade RBK är beroende av komponenter i serum och klass-A scavengerreceptorer på makrofagernas yta.

I en förlängning hoppas vi att kunskap om vilka förändringar hos lagrade RBK som är relevanta för deras igenkänning av makrofager kan leda till framtagandet av en mer optimerad lagringsmiljö för RBK. Sådana strategier kan leda till att minska mängden RBK som behöver överföras till patienter i samband med transfusioner om fler fungerande RBK kan behållas i cirkulationen.

xi

xii List of original papers

1. Transfusion of cryopreserved human red blood cells into healthy humans is associated with rapid extravascular without a proinflammatory response Hult A., Malm C. and Oldenborg P.A. Transfusion, 2013. 53(1): p. 28-33.

2. Splenic uptake of RBCs with an elevated cytoplasmic Ca2+- concentration primarily involves marginal zone macrophages and CD207+ dendritic cells Larsson A., Hult A., Nilsson A., Olsson M. and Oldenborg, P.A. Submitted

3. Phagocytosis of liquid-stored red blood cells in vitro requires serum and macrophage scavenger receptors

Hult A., Toss F., Malm C., and Oldenborg P.A. Submitted

The published paper was reproduced by permission from the publisher John Wiley & Sons Ltd.

xiii xiv

Introduction

The human body contains, depending on body size, between 4 and 6 liters of blood [4]. The main components of blood are the cell free plasma (50-60%) and erythrocytes (red blood cells/RBCs; 40-50%), while white blood cells (leukocytes) and (thrombocytes) together accounts for only ~1% of total blood volume [5]. The cellular components of blood originate from pluripotent hematopoietic stem cells situated in the red bone marrow. Hematopoietic stem cells give rise to erythrocytes, (which produce thrombocytes important for the process), and leukocytes of lymphoid or myeloid origin [6]. The formation of blood cells, or hematopoiesis, is strictly regulated by different hormones and cytokines in order to meet the demand of the ever shifting physiological milieu of the body, i.e. to produce more leukocytes upon or more erythrocytes when tissue oxygenation is low [5].

The leukocytes

The immune system is built up by a variety of lymphoid tissues, leukocytes and molecules, which are designed to protect us from infectious agents such as , , fungi and parasites. In addition to protecting us from , the immune system also helps in maintaining body homeostasis by removing and restoring damaged endogenous cells and tissues. After differentiation and maturation, predominantly in the bone marrow, leukocytes migrate to the blood, peripheral tissues and the lymphatic system. In all these tissues, leukocytes carry out their various immunologic functions. The immune system can be divided into two branches; the innate and the adaptive system. The innate immune system is a rapidly responding system that recognizes and eliminates a wide range of pathogen-associated molecular patterns (PAMPs) found on various infectious agents. PAMPs are identified by their binding to different pattern recognition receptors (PRRs) expressed by the immune cells [6, 7]. The is initially much slower, but instead more efficient in clearing the body from infectious agents. Furthermore, the adaptive system has a memory, and the ability to generate long lasting immunity against different pathogens. Via this memory function, pathogens will be swiftly recognized and eliminated if invading a second time [6].

1

Innate immunity

The cellular components of the innate immune system derive mainly from myeloid progenitors and include mast cells, , macrophages, dendritic cells (DCs), and , and . Despite their lymphoid origin, natural killer cells (NK-cells) are also considered as innate immune cells which identify and kill -infected or cancerous endogenous cells. Mast cells, macrophages and DCs are mainly found in various tissues, while monocytes and granulocytes are mainly found in the blood of a healthy individual. The neutrophil is the most abundant leukocyte in blood, constituting 50-70% of all leukocytes. It is readily recruited to infected areas by inflammatory and cytokines and play a key role in the initial defense against pathogens. Blood monocytes eventually leave the circulation and migrate to organs and peripheral tissues where they mature into macrophages. As they mature, macrophages will display a unique receptor repertoire depending on their microenvironment, making for instance a liver macrophage somewhat phenotypically and functionally different from a marginal zone macrophage in the spleen [8]. Upon recognition of infectious agents, the innate immune system launches an inflammatory response by the release of , pro- inflammatory cytokines and chemokines in order to recruit and activate other immune cells to the infected area for the participation in the fight of the infection [6]. The innate immune system also includes the ability of , whereby macrophages or DCs scavenge foreign or endogenous antigens in peripheral tissues by phagocytosis or macropinocytosis for later presentation to the adaptive immune system.

In addition to fighting infectious agents, the innate immune system (DCs and macrophages in particular) is also involved in maintaining normal homeostasis in its microenvironment by engulfment of senescent or apoptotic endogenous cells and cell fragments, a process mostly resulting in anti-inflammatory responses [9]. The discrimination between (modified) self and non-self is dependent on which receptor(s) that recognize(s) the cell or substance destined for phagocytosis (or endocytosis) by the phagocytic cells and ultimately determines if the cell will inhibit or induce an [7]. Aside from the cellular components of the innate immune system, there is also the which consists of several plasma proteins with the ability to bind to various pathogens and either initiate of the pathogen (mainly bacteria) or elicit an inflammatory response by the release of pro-inflammatory molecules and the formation of opsonins on the surface of the pathogen. Opsonins target pathogens for elimination by phagocytic cells that carry complement receptors (CRs) [10].

2

Pattern recognition receptors (PRRs)

The cells of the innate immune system carry a panel of receptors, PRRs, that recognize evolutionary conserved patterns on pathogens - PAMPs as well as molecules from damaged endogenous cells and tissues (danger-associated molecular patterns - DAMPs) [11, 12]. These PRRs include the Toll-like receptors (TLRs), which recognize bacterial patterns like lipopolysaccharide (LPS) as well as viral RNA. Another member of the PRRs is the C-type receptors (CLRs). These receptors recognize carbohydrate structures on pathogens. Endogenous carbohydrate structures, capable of binding to CLRs, are often hidden by terminal sialic acid, thus hindering auto- recognition [6]. Binding to TLRs and CLRs can elicit production of the pro- inflammatory cytokines interleukin-1β (IL-1β), tumor necrosis factor-α (TNF-α) and IL-6 [11] , (Figure 1).

Scavenger receptors (SRs) also belong to the PRRs and can be involved in recognition of PAMPs on pathogens as well as ligands on apoptotic cells [13]. The discrimination between self and non-self will ultimately determine if the

Pathogen

PRRs Phagocytic cell Degradation of pathogen PAMP Recognition Release of Nucleus pro-inflammatory cytokines

Cytokine IL-1β, IL-6, TNF-α production

Figure 1. PRR mediated pathogen recognition by a phagocytic cell. PAMP recognition by PRRs results in; 1, Phagocytosis and subsequent degradation of the pathogen. 2, Transcription, production and secretion of pro-inflammatory cytokines e.g. IL-1β, TNF-α and IL-6. The release of pro-inflammatory cytokines will initiate an inflammatory response that will recruit and activate other leukocytes to the site of infection.

3

phagocytic cell will generate a pro-, or anti-inflammatory response. SRs are a heterogeneous family of receptors with an extremely diverse ligand repertoire, where most ligands are polyanionic and include proteins, polyribonucleotides, polysaccharides and lipids [13, 14]. SRs are subdivided into classes (class A-I, each with several members) and are known to partner with various co-receptors for their function [13]. Macrophages express several SR classes, including class A-SRs. SR-A1 and MARCO are two receptors that belong to the class A-SRs found in the spleen. SR-A1 is expressed by most macrophages while the expression of MARCO is restricted to marginal zone macrophages (MZM) [13].

Adaptive immunity

The cells of the adaptive immune system derive from a common lymphoid progenitor cell and can be divided into B- (B-cells) and T- lymphocytes (T-cells). Lymphocytes are, in contrast to leukocytes of the innate immune system, highly specific in their ability to recognize antigen in that it will only bind to one specific motif. During development, the genes that are coding for the variable region of the B- or T-cell receptor undergoes rearrangements and hypermutations, resulting in a wide range of receptor variability and subsequent pathogen recognition within the lymphocyte population. Each individual lymphocyte, however, is still highly specific in its antigen recognition. Lymphocytes that recognize self-antigens undergo () during development in order to avoid autoimmune reactions. Those that survive this selection are allowed to migrate from the bone marrow (or thymus in the case of T-cells) into the blood and lymphatic system, including the spleen and lymph nodes [6].

When a B-cell has bound an antigen to its receptor (called B-cell receptor/BCR), the antigen is internalized and later presented on MHC class II to T-cells for T-cell dependent activation. Upon activation, the B-cell starts to proliferate and changes phenotype into either a memory cell or a plasma cell. The plasma cell starts to produce immunoglobulins (antibodies or ab:s) which are then distributed via the lymph system out into the body, with the ability to bind a specific antigen. Binding of ab:s to their antigens induces opsonization of the pathogen which can then be recognized by phagocytic cells having Fcγ-receptors (FcγR) that recognizes the constant region of the ab:s. The binding of an ab to its antigen can, in addition, initiate activation of the complement system which in turn increases the opsonization and leads to further promotion of phagocytosis [10]. The memory B-cells can swiftly be activated if presented with the antigen upon a second infection and hence establishes a long lasting immunity against its specific antigen [6].

4

T-cells recognize its antigen via the T-cell Receptor (TCR). Antigens are presented to T-cells by macrophages or DCs. When T-cells recognizes its antigen, it can be activated and will then differentiate into one of the following subsets; 1) Cytotoxic T-lymphocytes (CTL) whose main function is to kill infected host cells, 2) T-helper cells (TH-cells) who are involved in co- activation of B-cells as well as other immune cells and, 3) Regulatory T-cells (TReg-cells) which can promote suppression of lymphocytes. [6] This is of course a simplified description of the extremely complex functions of the immune system in general and the adaptive immune system in particular.

Tolerance vs. Immunity

DCs are particularly efficient antigen presenting cells (APCs) that control the T-cell mediated immune response via paracrine and contact-dependent signaling mechanisms. When a DC has taken up and processed an antigen, it will migrate to the T-cell areas of secondary lymphoid organs and present the MHC class II-bound antigen to naïve T-cells. If the engulfed antigen was recognized as foreign by the DC, it will also express co-stimulatory receptors (e.g. CD80 and CD86) and secrete pro-inflammatory cytokines (i.e. interleukin 12 (IL-12) and IL-18) to promote activation of TH-cells and CTLs [15] (Figure 2A). On the other hand, if the antigen was recognized as a self- antigen, the DC will display the antigen to T-cells together with lower levels of CD80 and CD86, which will not be accompanied by pro- production. Instead, the anti-inflammatory cytokine IL-10 will be released. Presentation of self-antigens may result in either T-cell anergy or the formation of TReg-cells and promotion of tolerance against the self- antigen (Figure 2B) [16]. The activation of T-cells is hence a two-step process in which binding of the T-cell receptor to the MHC-bound antigen on DCs is the first step. The second step is either the ligation of CD80/CD86 on the DC to CD28 on the T-cell to promote an immune response, or binding of CD80/CD86 to CTLA-4 on the T-cell to promote tolerance.

The Red blood cell

Function

Erythrocytes (from now on referred to as RBCs), is the most abundant cell type in the human body, and play a key role in oxygen delivery from the to tissues throughout the body. The transport of oxygen is mediated by the iron-containing metalloprotein hemoglobin (Hb), which is highly abundant in RBCs. The tertiary structure of adult Hb consists of four polypeptide chains (two α-globin and two β-globin) which all carry an iron- containing heme group to which oxygen is reversibly bound, forming

5

Pathogen A PRRs Antigen presentation on MHC class II

PAMP Naїve Recognition Degradation of pathogen T-cell TCR Nucleus CD28

APC Pro-inflammatory Co-receptors

Pro-inflammatory cytokines IL-1β, IL-6, IL-12, IL-18, TNF-α

Apoptotic cell B PRRs Antigen presentation on MHC class II

Recognition of Naїve modified self Degradation of cell T-cell TCR Nucleus CTLR-4

APC Anti-inflammatory Co-receptors

Anti-inflammatory cytokines IL-10

Figure 2. Recognition patterns on ingested material determine how antigens are co-presented to T-cells. Ingestion of antigen by an APC is degraded and later presented on MHC class II on the cell surface. A, Recognition of foreign (e.g. pathogens) by APCs induce a pro- inflammatory response and subsequent T-cell mediated immunity. B, Recognition of modified self (e.g. apoptotic cells) by APCs induces T-cell mediated tolerance in an anti-inflammatory manner. oxyhemoglobin (hemoglobin that has not bound oxygen is denoted

6

deoxyhemoglobin). The affinity of oxygen to hemoglobin is affected by the partial pressure of oxygen (pO2) and carbon dioxide (pCO2), temperature, pH and the intracellular concentration of 2,3-diphosphoglycerate (2,3-DPG). Thus, the affinity regulation of oxygen to Hb optimizes oxygen uptake in the lungs, and oxygen delivery to metabolically active tissues, dependent on the physiological milieu at the respective site [17, 18].

RBCs also facilitate CO2 transport in two different forms; a small fraction of CO2 is bound to the globin part of Hb, while the bulk is converted by the carbonic anhydrase into carbonic acid (H2CO3) which dissociates + - - into H and bicarbonate (HCO3 ) in the cytosol of RBCs. The HCO3 is then exchanged with chloride (Cl-) across the plasma membrane by the anion transporter membrane protein Band 3 (an abundant membrane protein in RBCs) to further be transported in the to the lungs where the reaction is reversed, allowing CO2 to diffuse into the alveoli of the lungs and finally be exhaled. [5]

The delivery of O2 to, and the removal of CO2 from, the tissues enables adenosine triphosphate (ATP) production from oxidative phosphorylation to occur in the mitochondria of the target cells and thus permitting an efficient energy utilization throughout the body. Because mature RBCs do not contain any organelles, including mitochondria, it will not consume any of the oxygen it carries and are hence solely dependent on anaerobic glycolysis for ATP production, leaving lactate as an end product. Lactate is finally expelled from the RBC via symport of hydrogen ions in a pH dependent manner [5, 19].

Erythropoiesis

The maturation of RBCs from hematopoietic stem cells involves several intermediates as the precursor cells divide and proliferate, starting to produce Hb and extrudes organelles (including their nucleus), to finally reach the state of . Still carrying ribosomes and mRNA for protein synthesis, the reticulocytes are then released from the red bone marrow into the blood stream. During the first couple of days in circulation, the reticulocytes maturate into RBCs which have lost their capability to produce proteins. At any given time, 1-2 % of all circulating RBCs are reticulocytes. The concentration can hence be used to assess the rate of the RBC production, since a higher percentage of reticulocytes is indicative of increased erythropoiesis [20].

The regulation of erythropoiesis under normal conditions is highly dependent on the peptide hormone erythropoietin (EPO), which is mainly

7

synthesized in the kidneys. EPO promotes committed erythroblasts in the red bone marrow to proliferate and differentiate to RBCs [21]. The EPO production is in turn regulated by the oxygen availability in the EPO producing cells, thus creating an efficient feedback loop in where more EPO, and consequently more RBCs, are produced when tissue oxygenation is low [20, 22]. However, in cases of increased hemolysis and chronic anemia, EPO regulation of erythropoiesis is not sufficient to compensate for the vast reduction of RBCs. Instead, and stem cell factor have been suggested to facilitate the recovery from severe anemia by stimulating extended proliferation of early erythroid progenitor cells [23].

The life and death of RBCs

The mature human RBC has a lifespan of approximately 120 days before it is targeted for destruction by macrophages, mainly in the spleen and liver. Although not fully understood, the senescence of RBCs is likely to involve several structural changes on the cell surface, which ultimately leads to macrophage recognition and phagocytosis [24]. Following ingestion of RBCs, the macrophages recycle iron from Hb which is then transported back to the bone marrow by the transporter protein transferrin in order to facilitate continued erythropoiesis. In situations with increased RBC breakdown, transferrin can become saturated. The highly reactive iron ions, capable of forming and tissue damage in its unbound state, is then found as non-transferrin bound iron (NTBI) in the plasma [25]. The remaining heme-group of Hb (without iron) is converted to bilirubin in macrophages and later secreted in the blood plasma [26]. Changes in the plasma bilirubin concentration and transferrin saturation can hence be used as indicators to assess the rate of RBC phagocytosis.

Senescence markers of RBCs include the clustering and/or degradation of the Band 3 membrane protein which has been shown to be formed after various oxidation processes. This may be attributed to the binding of oxidized Hb to the cytoplasmic domain of Band 3 [24, 27, 28]. Alterations in Band 3 can also be induced trough the activation of calcium- dependent proteases like calpain, at least in the elderly [29]. RBCs shrink in size and increase their density as they age, which can be seen by lower forward scatter (FSC) as assessed by flow cytometry and by density centrifugation fractionation [30]. The decrease in cell size is contributed to by the shedding of microparticles (MPs) throughout the lifespan of the RBC [31]. MPs contain various amounts of Hb and their formation seems to be, at least partly, facilitated by the spleen [31, 32]. The MPs display the phospholipid phosphatidylserine (PS) on their outer membrane leaflet and are quickly removed from the circulation by macrophages [33]. If human RBCs

8

themselves externalize PS as a natural part of their senescence is questionable [30], but cannot be completely ruled out due to the fast recognition of PS+ cells by macrophages, leaving them little time to be detected in circulation. Reductions in terminal sialic acids [34, 35] and in the expression of the cell surface glycoprotein CD47 [36] have also been observed in the membrane of senescent RBCs. Increased non-enzymatic glycation (forming advanced glycation end products (AGEs)) of Hb as well as membrane proteins is also found to occur as RBCs age [37, 38]. Finally, the metabolic rate of RBCs decrease over time as seen by reduced activity of glycolytic and a consequent reduction in ATP and lactate production [39].

Eryptosis

Besides normal senescence, RBCs can also enter a programed cell death (apoptotic)-like process called eryptosis, which can be induced by an number of endogenous and xenobiotic substances, as well as by oxidative stress, osmotic chock and energy depletion [40]. Eryptosis is not a pure analogue of apoptosis in nucleated cells, as RBCs lack mitochondria and nucleus. Eryptosis and apoptosis yet share many common features like cell shrinkage, MP formation and PS externalization, which in RBCs is induced by increased cytosolic Ca2+ levels [41, 42]. The actions of increased cytosolic Ca2+ involves the activation of K+ ion channels and subsequent reduction in cell size, cytoskeleton degradation (and membrane blebbing) by activated calpain, and an increased scrambling activity of the phospholipids in the lipid bilayer resulting in PS-externalization [42]. Eryptosis has also been reported to occur in a number of diseases, including type-2 diabetes [43, 44], [45] and heart failure [46], often related to oxidative events.

Recognition and phagocytosis of RBCs

The mechanisms behind macrophage removal of normally senescent RBCs, reaching the 120 day limit of their life-span, are still not completely understood [2]. This lack of detailed knowledge is one important reason why there is an even greater deficit in the understanding of the mechanisms behind the rapid post-transfusion clearance of stored RBCs, which will be discussed further in a later section. In addition to the biochemical changes that occur during RBC senescence, clearance of senescent RBCs requires molecular changes to the cell surface which can then be recognized by macrophages (Figure 3). Most likely, senescent RBCs gradually accumulate these cell surface changes, suggested to include loss of terminal sialic acids and the subsequent exposure of underlying carbohydrate structures [34], a reduction of the anti-phagocytic protein CD47 on the cell surface [36],

9

binding of naturally occurring antibodies (NAbs) [30, 47] and oxidative damage to Hb and phospholipids [48]. Many of these changes may well be recognized by NAbs, but the most convincing target for such antibodies so far is the conformational changes that take place in the Band 3 protein as a result of oxidative damage to hemoglobin [27, 28, 49]. Although NAbs are of low affinity and at low concentrations in circulation, binding of NAbs could result in FcγR-mediated phagocytosis of senescent RBCs but also complement deposition and CR participation [50]. Finally, while being questioned [30], senescent RBCs may also, like nucleated apoptotic cells, eventually expose PS on the surface to promote their phagocytosis by macrophages [51]. PS can be recognized by several receptors on phagocytic cells, either directly by TIM1, TIM4 or stabilin-2, or via bridging molecules like Gas-6, thrombospondin (TSP) and lactadherin, following interaction with integrins on the macrophage [2].

Normal RBC Senescent RBC Sialic acid PS Band 3 SIGLEC C3b CD47 Tim-4 SIRP-α TSP NAbs CR Integrin FcγR Anti-phagocytic Pro-phagocytic receptors receptors

Macrophage

Figure 3. Pro-, and anti-phagocytic receptors on macrophages. Abbreviations: SIGLEC - Sialic acid binding immunoglobulin lectins, SIRP-α - Signal-regulatory protein-α, PS – Phosphatidylserine, TIM-4 - T-cell immuno- globulin mucin protein 4, TSP - Thrombospondin, NAbs - Naturally occurring antibodies, FcγR – Fc gamma receptor, CR – , C3b – complement protein 3b. Modified from de Back et al. [2].

10

In addition to the pro-phagocytic receptors mentioned above, macrophages also carry anti-phagocytic receptors (figure 3). Signal-regulatory protein α (SIRPα) is an inhibitory receptor expressed on phagocytic cells that recognizes the widely expressed “self-antigen” CD47 [52]. Mouse RBCs lacking CD47 was shown to be rapidly phagocytized by red pulp macrophages in the spleen, indicating the anti-phagocytic properties of the CD47-SIRPα interaction. Another family of receptors that mediates recognition of self and inhibits phagocytosis is the sialic acid binding immunoglobulin lectins (SIGLECs) that recognizes terminal sialic acids on host cells [53]. A reduction in sialic acid on RBCs (by neuraminidase treatment) leads to the rapid clearance of these cells by macrophages [54]. A reduction in CD47 and terminal sialic acid on aging RBCs have been observed [34-36], which may well contribute to macrophage recognition of senescent RBCs.

Immune cells of the spleen

Principally, the spleen has two separate compartments; the immunological and lymphocyte-rich white pulp, and the RBC-rich red pulp having more of a filtering function [1, 55, 56]. The splenic white pulp is divided into separate T-cell and B-cell areas. Macrophages can be found in both the red and white pulp, as well as in the area separating these two domains – the marginal zone (MZ). Separate DC subsets are also distributed within the red pulp, MZ and white pulp (T-cell) areas. In the murine spleen, three major macrophage populations can be identified; the F4/80+ red pulp macrophages, the MARCO+ MZ macrophages and the MOMA-1+ marginal metallophilic macrophages (MMM) [1, 3] (Figure 4). F4/80+ macrophages are mainly involved in the clearance of cells trapped in the red pulp, whereas the other two macrophage populations are more specialized in trapping microbes and PS+ apoptotic cells. Due to their close localization to the splenic white pulp, MARCO+ and MOMA-1+ macrophages (together with DCs) are more actively interacting with lymphocytes to regulate the adaptive immune system. When investigating macrophage phagocytosis of RBCs, it is important to remember that macrophages in vivo are very heterogeneous as a group, with many specific functions depending on their phenotype and anatomical location [8]. Adding on to this complexity, activated macrophages may also promote as well as dampen inflammation, making analysis of specific macrophage subsets even more important [8].

11

Figure 4. Anatomical structure of the spleen. Large arteriole (lower part to the left) is branching and delivers blood to the marginal zone (MZ) and sinusoids of the spleen. Blood is then filtered through the MZ or arteriole sinusoids in to the red pulp and finally collected in sinusoids connected to venules. Upper part represent a magnification of the MZ. Marginal metallophilic macrophages (MMM) is lining the white pulp side of the MZ. Adjacent to MMM are dendritic cells (DCs) and MZ macrophages (MZM). Blood is filtered through the MZ an in to the red pulp where red pulp macrophages are situated. (Modified from Junqueira et al., [1] and Cyster [3]).

12

RBC transfusion

History

The first successful blood transfusion in humans was carried out by obstetrician Dr. James Blundell in 1818 when he transfused approximately 400 ml of whole blood, which was collected from multiple donors and administrated ten times over a 40 min period of time [57]. The patient, suffering from gastric carcinoma (described at the time as “scirrhosity of the pylorus”), showed clear symptoms of anemia and was barely conscious at the time of transfusion, but displayed marked improvement in the following 24 h post transfusion. His symptoms then reemerged and were followed by death a little over two days after the transfusion. Dr. Blundell continued his pioneering work in blood transfusion with a total of 10 documented transfusions of which 5 were successful [58].

The next big step towards modern transfusion medicine was taken by Karl Landsteiner in the early 20th century. He discovered that sera from different donors sometimes, but not always, agglutinated blood cells of other donors. He proposed the ABO blood group system (although Dr. Landsteiner at the time classified blood into groups A, B and C) [59] as a method to conduct compatible blood transfusions, a discovery that ultimately rendered him the Nobel prize in physiology or medicine in 1930. The identification of the Rhesus blood group system (the second most important blood group system after the ABO system) in the late 1930’s [60] is also to be considered an important step on the way to safer and more compatible blood transfusions.

Today, more than 300 antigens belonging to over 30 blood group systems have been described[61]. Blood group antigens are proteins or carbohydrate structures present on the of RBCs. Alloimmunization against blood group antigens can occur after RBC transfusion, owing to blood group incompatibilities between donor and recipient [61]. Alloimmunization results in ab-mediated immunity against the foreign antigen. In most cases, the production of ab:s against blood group antigens is an induced response which follows after contact with the respective antigen. In contrast, ab:s against A and B antigens of the ABO blood group system starts to be produced during the first year after birth [62], irrespective of contact with the antigens. Enzymatic removal of the B-antigen from RBCs of blood group B has been described, making them transfusable to donors of all blood groups [63]. The combined removal of both the A and B antigens have later been reported [64], opening the door for universal RBCs that can be transfused to everyone, regardless of the ABO-blood group of the donor or recipient.

13

The use of citrate as an anticoagulant in donated blood made transfusions easier to manage as the risk for coagulation was vastly reduced [65]. The addition of glucose (or dextrose) in combination with citrate enabled prolonged storage of blood possible [66]. The introduction of first Acid- Citrate-Dextrose (ACD) and later Citrate-Phosphate-Dextrose (CPD) as a combined anticoagulant and energy source to the collected blood (stored in plastic bags) enabled donation and transfusion to be separated in time by as much as 28 days when stored at +4°C [67], thereby making storage of blood in blood banks a routine in many hospitals.

Demographics

According to the World Health Organization (WHO), 108 million units of blood were donated worldwide in 2012, with approximately half of these donations occurring in high-income countries where only 18% of the world’s population lives. In low-income countries, a majority of the collected blood is stored and later on transfused as whole blood, as compared to high income countries where 95% of the donated blood is further processed into RBC-, plasma-, and thrombocyte concentrates [68].

Effects of RBC transfusion

Blood transfusions are used both in the setting of acute and chronic bleeding, as well as in patients with bone marrow suppression, e.g. cytostatic treatment, and in patients with hemoglobinopathies (e.g. thalassemia and ) [69, 70]. Symptoms of anemia (i.e. fatigue, tachycardia and dyspnea) usually only arises at Hb concentrations below 90-100 g/L due to compensatory mechanisms that increase cardiac output and replenish blood plasma volume [71]. However, symptom onset is largely dependent on patients’ related factors such as age, cardiorespiratory fitness, rate of anemia development and functional requirements. The recommendations to transfuse hemodynamically stable patients is generally recommended at Hb 70-80 g/L [71]. The restoration of blood volume and viscosity after profound hemorrhagic shock and hemodilution is also important in order to restore a functional microcirculation by increasing blood perfusion [72]. Moreover, the oxygen carrying capacity following transfusion has been documented to increase in non-anemic men as assessed by increased arterial oxygen content and greater resistance to hypoxic conditions during exercise [73], an observation that has opened up for the misuse of blood transfusions as a doping agent in endurance sports [74].

14

RBC storage conditions

To allow for a more efficient use of donated blood, blood donation and transfusion have to be optimally separated in place and time. This can be achieved by avoiding coagulation and assuring a sufficient ex vivo storage milieu with respect to nutrients and buffered pH conditions [66]. The first step involves collecting donated blood in CPD. As mentioned above, donated blood is today often separated into its different components. In this process, RBCs are isolated from the other cellular components of the blood, as well as from the majority of the plasma, by centrifugation and filtration techniques carried out inside a closed system containing the necessary components. The last step in this process involves the addition of an additive solution (AS) to the packed RBCs in order to optimize storage conditions. In Europe, a solution of Saline-Adenine-Glucose-Mannitol (SAG-M) is widely used, while closely related solutions (abbreviated AS-1, AS-3 and AS-5) are used in America. Blood stored at +4°C in these solutions can be kept for up to 42 days and typically display an in vivo recovery of 78-82% 24 h after transfusion into healthy individuals [75]. This may, however, not reflect the 24 h recovery in a clinical setting, since there are reports involving transfusions to patients with various underlying diseases showing 24 h recoveries after transfusion of around 75% already after 25-35 days of storage [76, 77]. Recently, a new additive solution (AS-7) has been approved for prolonged storage of RBCs for up to 8 weeks [78]. The AS-7 differs from the previous storage solutions by containing NaHCO3 and NaPO4 but no saline, having a higher pH and lower osmolality and displaying an in vivo 24-h recovery of 88% at day 42 and 82% at day 56 of storage [78].

The storage lesion

RBCs in liquid storage undergo time-dependent metabolic and structural changes, collectively referred to as the storage lesion. The storage lesion has been well characterized, based on in vitro measurements on RBC concentrates over time. Metabolic alterations are manifested by the reduction in ATP and 2,3-diphosphoglycerate (2,3-DPG) content. As RBCs are purely glycolytic, a time-dependent increase in levels of lactate and a corresponding decrease in pH is also noticed in the storage media. Structural changes, partly due to oxidative damage, include conformational changes in the Band 3 anion transport protein, IgG binding as well as increased rigidity and osmotic fragility. Shape shifting and a reduction in size is also noted in a subset of stored RBCs, attributed to the shedding of MPs and by reduced levels of cytoplasmic K+ [76, 79-82]. The amount of CD47 may also be reduced during storage [83, 84], although as to which extent is not clear. Thus, stored RBCs show several similarities to what has been described for

15

normally senescent RBCs in circulation (Figure 5). ATP and 2,3-DPG levels have been shown to recover shortly after transfusion, indicating that at least part of the storage lesion is reversible in vivo following transfusion [85]. However, it is presently unclear if the storage lesion simply reflects an accelerated aging of the RBCs, if it is merely a stress-induced eryptosis, a combination of both, or something else. Perhaps more important, the physiological consequences in vivo (after transfusion) remains largely unknown [86]. It is also unclear if the storage lesion affects all age-fractions of the RBCs in a random way, or if it primarily affects the older cells within the RBC population. This distinction may be of importance to understand the storage lesion and in the long run to develop strategies to prevent it.

In addition to liquid storage, RBCs can be frozen (cryopreserved) for extended periods of time, using glycerol as a cryoprotectant. Cryopreserved RBCs are more fragile and display increased hemolysis upon thawing and washing, as compared to liquid stored RBCs, but instead maintain their ATP and 2,3-DPG content better [87, 88]. In addition, RBC cryopreservation does not result in MP formation, PS exposure or reduced levels of CD47 [89]. The 24 h in vivo recovery of cryopreserved RBCs has been reported to be ~85%, making cryopreservation of RBCs a valuable alternative to liquid storage, especially for patients with rare blood groups or when there is a shortage of liquid stored RBCs [88].

Senescent Stored Eryptotic RBC RBC RBC

Microparticle formation + + + Glycolysis ↓ ↓ - ATP ↓ ↓ ? 2,3-DPG ? ↓ ? Lactate - + - Oxidative damage + + - Ca2+ influx ? ? + Band 3 aggregation/degradation + + + Phosphatidylserine exposure ? ? + IgG binding + + ? Sialic acid ↓ ? - CD47 ↓ ↓(?) ?

Figure 5. A comparison between senescent, stored or eryptotic RBCs. Presented is a selection of metabolic and structural changes reported in the literature.

16

Risks associated with transfusion

Transfusion-related acute injury (TRALI) is the leading cause of fatalities directly related to transfusion, followed by hemolytic transfusion reactions due to blood group incompatibility [71]. In a systemic review of 45 cohort studies, evaluating the efficacy of RBC transfusions in high-risk patients, the risks associated with transfusion were found to outweigh the benefits in all but three studies [90]. Firm conclusions of the relation of risks and benefits of transfusions are however difficult to draw because of the inherent bias in cohort studies as sicker patients tend to receive more transfusions [91]. There has also been an ongoing debate as to whether RBCs that have been stored for longer periods of time will have adverse effects following transfusion in various clinical settings. These effects include increased risks of infection, multiple organ failure, length of hospital stay and mortality [92-94]. A compilation of 32 studies on the topic by Aubron et al. [95] found that 18 studies reported clinically harmful effects and 14 did not. The discrepancy is likely due to various confounders not taken in to account, as well as problems with underpowered studies. It is also possible that transfusion of older blood can be particularly harmful to certain patient groups [92].

Recently, a large randomized controlled trial (Age of blood evaluation, ABLE) was published in which over 2400 ICU-patients had been enrolled and randomly selected to receive either fresh RBCs (stored for 6.1±4.9 days) or standard issued RBCs (stored for 22.0±8.4 days) [96]. The authors found neither an increased risk of 90 day mortality, nor any other risks that could be associated with transfusion of stored RBCs. Similar findings have been reported from randomized controlled trials conducted on cardiac surgery patients (the RECESS trial) [97] and in premature infants (the ARIPI trial) [98], indicating that there are no major risks associated with transfusion of older blood, at least not for the patient groups that have been investigated. Still, because the maximum storage of RBCs in AS ranges from 42 to 56 days, it could be argued that the average storage times of 22 days (ABLE), 28 days (RECESS) or 14.6 days (ARIPI) may not be relevant to asses “older” RBCs. This argument becomes particularly important in case the possible adverse effects related to prolonged RBC storage would be manifested closer to the end of the maximum storage time [99]. In addition, despite the relatively large number of participants, the above mentioned studies have been challenged as underpowered, thus failing to detect possible small but clinically relevant differences [99, 100].

Several animal studies have been conducted in the search for the mechanistic explanation to the supposed risks with transfusion of stored

17

RBCs. Hod et al. reported that mice receiving the equivalent of 2 units of stored (but not fresh) RBCs manifested with high levels of pro-inflammatory cytokines and NTBI in the plasma 2 h after transfusion [101]. In addition, they found that the pro-inflammatory response to lipopolysaccharide (LPS) stimulation was enhanced by the transfusion of stored, but not fresh RBCs [101]. Comparable inflammatory responses were obtained in a similar study in canines [102]. “The iron hypothesis” have been suggested to explain the inflammatory response seen in animal studies. This states that inflammation is mediated by an increased erythrophagocytosis and iron overload in the macrophages, resulting in release of reactive iron ions in a quantity that exceeds transferrin binding capacity [103]. However, in a human study where 1 unit of RBCs stored for 42 days was transfused, there were indications of NTBI in the plasma but no signs of a pro-inflammatory response [104].

How to study RBCs

Human studies

As mentioned earlier, the current (poor) understanding of the storage lesion and its effects upon transfusion is largely based on in vitro studies of stored RBC concentrates, which may or may not be relevant to explain the swift macrophage recognition of these cells after transfusion. In vivo studies are usually limited to non-invasive monitoring of 24 h survival assessments using 51Cr or biotin in vitro-labeled RBCs [105, 106]. As for the efficacy of transfusion, there are indications that stored RBCs actually decrease tissue oxygenation upon transfusion [107, 108]. This is possibly attributed to a reduced activity as a consequence of a decreased ATP content and nitric oxide (NO) availability [109]. However, there are limitations in the studies conducted to measure tissue oxygenation in humans, which complicate the interpretation of these findings [110].

Animal models

Human physiology or is, of course, best studied directly in humans. However, that is not always practical or possible to do, in part due to ethical considerations. Therefore, the use of the mouse as a model system is common in order to enable studies not easily performed in humans. The mouse offers advantages by its small size, short generation time and both homologous and modifiable genetic backgrounds. Although mice are not a man, they are still a close relative when comparing the genomes with ~99% of the human genes having mouse homologues [111]. Even though the immune system of humans is quite similar to that of mice, it is nonetheless

18

important to mention that there are differences between species which needs to be taken in to account before making generalizations of findings in mice and its potential impact for humans. For instance; the main leukocyte cell type in human blood is neutrophil granulocytes which constitute 50-70% of total leukocytes. In mice, only account for 10-25% and lymphocytes are instead the most abundant among blood leukocytes. Immunological differences between species also includes the occurrence and distribution of various kinds of Toll-like receptors (TLRs), CD4 and CD33, Fc receptors and immunoglobulin subclasses, as well as differences in cytokines secreted by the cells of the immune system. This can be seen as a reflection of different strategies used by the two species to cope with the environmental demands that they have been subjected to during the 65 to 75 million years of evolution that separates us [112].

Mouse RBCs have a shorter lifespan than human RBCs, typically about 50 days [113]. Senescent mouse RBCs are recognized by macrophages in spleen and liver [114]. The senescence mechanisms seem to be similar in mice and humans with respect to rheological properties and cell shrinkage [115]. Although at an accelerated rate, human and mouse RBCs also display similar storage characteristics [116]. Taken together, this indicates that a mouse model of storage and transfusion of RBCs could prove useful in getting a deeper understanding of implications of the human red blood cell storage lesion after transfusion.

19

Aims

Conflicting results have been reported regarding inflammation as a potential side effect of transfusion of stored RBCs.

• The aim was to investigate if cryopreserved human RBCs induce an increased erythrophagocytosis and subsequent inflammatory response following autologous transfusion into healthy volunteers.

The storage lesion of RBCs may resemble senescence, but could potentially also include eryptosis.

• We aimed at explore the role of eryptosis in the storage lesion and then investigate the splenic uptake of eryptotic cells after transfusion.

Although the biochemical changes described in stored RBCs may have an impact on their survival after transfusion, it seems logical that there is another level of the problem that needs to be understood in order to explain the rapid loss of RBCs after transfusion, namely; what molecular changes to the cell surface of the stored RBCs are recognized by macrophages and eventually results in RBC phagocytosis?

• The final aim was to study the receptor-ligand interaction between liquid stored RBCs and macrophages in order to better understand the consequences of the storage lesion after transfusion.

20

Materials and methods

Human studies (paper 1)

Ten healthy male participants, age 22 to 44 years, with Hb in the range of 140-155 g/L, were included in the study. The subjects, most of them committed to recreational training, were not allowed to compete in any events associated with the Swedish sports confederation during the entire time period of the study, including 4 weeks after reinfusion, due to the fact that they would potentially have a competitive advantage over any opponent related to the increase in oxygen carrying capacity after reinfusion. In addition, 7 control participants, matching the inclusion criteria, were also included for the evaluation of maximal aerobic capacity (VO2max). Control participants, which did not donate or reinfuse RBCs, conducted the VO2max tests at the same time points as the transfusion group.

Donation and reinfusion

Study participants donated one unit of blood (450 ml) on two occasions separated by one week. Blood was collected in CPD, followed by leukocyte and thrombocyte reduction, after which the majority of the plasma was removed by centrifugation. One hundred ml of SAG-M was then added to the remaining cells after which they were stored for 48 to 72 h before cryopreservation. Glycerolization of RBCs was carried out using the ACP 215 automated system (Hemonetics, Inc., Tamarac, FL) by the addition of a 57% glycerolyte solution (Fenwal, Lake Zurich, IL), after which the cells were kept frozen (-80°C) for 15 to 16 weeks. RBCs were then thawed, deglycerolized and washed, resulting in a post thaw recovery of ~80%. Reinfusion of RBC units was carried out by transfusing each unit under approximately 30 min. Venous whole blood samples from study participants were collected in EDTA, gel or Li- containing tubes (BD Vacutainer, Franklin Lake, NJ) at 2 h before, and at 2 and 48 h post transfusion for whole blood, serum and plasma analysis, respectively.

Blood analysis

Whole blood samples were used to analyze blood cell counts and Hb concentrations (Sysmex 2100, Sysmex, Kobe, Japan), while serum samples were used in the analysis of bilirubin, transferrin and serum iron (Vitros 5.1, Ortho Clinical Diagnostics, Inc., Raritan, NJ). Results from the two latter were combined to calculate transferrin saturation [117]. Serum samples were also used for the analysis of pro-, and anti-inflammatory cytokines, using a

21

cytokine bead array, custom designed to detect and quantify IL-1β, IL-6, IL- 8, IL-10, monocyte chemoattractant protein 1 (MCP-1), macrophage inflammatory protein 1β (MIP-1β) and TNF-α (7-Plex kit, Bio-Rad, Hercules, CA). Analyses were made according to the manufactures instructions, with the addition of an extra concentration for the generation of the standard curve in order to increase detection of cytokines of minute concentrations. Plasma samples were used for the quantification of total haptoglobin by an immunoturbidimetric assay (Cobas 6000, Roche Diagnostics Scandinavia AB, Bromma, Sweden).

Maximal aerobic capacity (VO2max) assessment

Study participants conducted treadmill running to exhaustion on a constant speed (individually set to be maintainable for 10 km) with a fixed stepwise incline increase (0.5-1°) each minute. Oxygen consumption (VO2) was measured on a Jaeger Oxycon Pro (Erich Jaeger GmbH, Hoechberg, Germany). VO2max tests were performed before and after blood donation, as well as before and (48 h) after reinfusion.

Animal studies (paper 2 and 3)

In the subsequent experiments, we used adult (8-25 weeks of age) male and female CD47-/-, CD47+/-, or wild type (wt) Balb/c mice, or wild type C57BL/6J mice. Mice were kept under low pathogen conditions and bred in our local facility with free access to water and food in accordance with local guidelines. All the experiments were performed in compliance with relevant Swedish and institutional laws and guidelines and approved by the Umeå research animal ethics committee (A14-12).

Labeling of RBCs

Labeling of RBCs before transfusion enabled the tracing of these cells after transfusion. We have used two different methods to label mouse RBCs. First, in vitro labeling with the fluorescent lipophilic dye PKH26 (according to the manufacturer’s instructions, Sigma Aldrich, St. Louis, MO). The PKH26 dye readily incorporates in the lipids of the RBC cell membrane [118]. Second, in vivo biotinylation with biotin (N-hydroxysuccinimido-biotin, EZ-Link NHS- Biotin, Thermo Scientific, Rockford, IL) which forms irreversible amine bonds to membrane proteins with exposed amine containing side chains (commonly lysine) and the N-terminal of proteins and polypeptides [106].

PKH26 labeling of RBCs was preferably used to identify splenic uptake of transfused cells. Tissue sections of spleens from transfused animals were

22

incubated with fluorescently labeled ab:s against surface markers of phagocytic cells in order to visualize specific cell populations, i.e. F4/80 on red pulp macrophages, MARCO on MZM, MOMA-1 on MMM and CD11c on dendritic cells. Sections were then analyzed by laser scanning confocal microscopy (Leica TSP-2, Heidelberg, Germany).

We used biotin-labeled cells in survival experiments, where blood samples could be drawn from mice at different time points after transfusion. Blood samples were then incubated with fluorochrome-conjugated streptavidin, which forms a strong bond to biotin. In this way, the fraction of the labeled cells, i.e. the transfused cells, could be monitored over time by flow cytometry (FACS Calibur, BD Biosciences, San Jose, CA). In vivo biotinylation was carried out by tail vein injection of 0.4 to 1.5 mg of biotin, dissolved in 150 µl of sterile PBS. This efficiently labeled all circulating cells. The mice were sacrificed the following day, and blood (with >98% biotin- labeled cells) collected by heart puncture. This blood was used for storage and transfusion. The double in vivo biotinylation technique enabled discrimination between different age fractions within the RBC population [119]. This was made possible by the injection of 1.1 mg of biotin per day for three days and then by waiting until half of the RBC population had been regenerated (~25 days), after which a lower dose of biotin (0.4 mg) was injected. This resulted in half of the RBCs binding higher levels of biotin (biotinhi; the older fraction) and the other half binding intermediate levels of biotin (biotinint; the young fraction). The two fractions of biotinylated cells can easily be discriminated individually from biotin negative RBCs of the recipient mice after transfusion.

We also labeled human RBCs with FLUO-3 (Molecular Probes) in order to quantify intracellular Ca2+ uptake. This was done by incubating RBCs with FLUO-3 in Ca2+-free media for 1 h in 37°C. Cells were then washed and later suspended in Ca2+-containing media, followed by incubation for 1 h in 37°C. Fluorescence intensity was then determined by flow-cytometric analysis (FACSCalibur flow cytometer, BD Biosciences).

The mouse model for RBC storage and transfusion

Whole blood was collected in CPD (blood to CPD-ratio - 7:1) by heart puncture of euthanized mice and spun at +4°C and 500 x g for 5 min. Plasma was then transferred to a separate tube, while the buffy coat and the top layer of the RBCs were discarded. Plasma was spun at +4°C and 1300 x g for 10 min, in order to remove any residual cells, after which the supernatant was transferred into a new tube. The RBCs were washed 3 times in PBS and finally mixed with SAG-M and plasma (ratio 3:1) and either used for direct

23

transfusion or stored for 10 days at +4°C in 0.5 ml test tubes. For Phthalate experiments, test tubes were coated in plastic film containing DEPH- phthalates.

In eryptosis experiments, the RBCs were incubated with the calcium ionophore ionomycin. For this, RBCs were suspended in Krebs-Ringer Hepes (KRH) medium at a hematocrit of 0.5% with 1µM ionomycin for 1 h at 37°C, followed by 3 washes in PBS. Ionophore-treated RBCs were finally resuspended in sterile PBS at a hematocrit 0.5% for phagocytosis experiments, or at 20% hematocrit for transfusion experiments.

Mice anesthetized with isofluorane (Baxter, Deerfield, IL) were transfused with 150 µl of fresh or 10 day stored RBCs, administered by tail vein injection. Blood samples from the opposite tail vein were taken at multiple time points after transfusion and analyzed by flow cytometry for the presence of transfused cells as described above.

In vitro characteristics of stored RBCs

To elucidate if the mouse model for RBC storage shared characteristics with that of human stored RBCs, in vitro measurements on intracellular ATP and the formation of MPs were conducted. For ATP measurements, 5 x 107 murine RBCs, stored for 0, 4, 7, or 10 days, were lysed in hypotonic buffer. An ATP standard curve was established by diluting ATP samples of known concentrations. Samples were mixed with ATP-reagent, containing luciferin and luciferase, followed by analysis of peak luminescence using a Triathler luminometer (Hidex, Turku, Finland). For MP quantification, fresh and 10 day stored murine RBC concentrates were labeled with fluorescein isothiocyanate (FITC)-conjugated anti-mouse Ter-119 (ImmunoTools, Freisoythe, Germany) and analyzed for MPs by logarithmic scaling for FSC and SSC on a FACSCalibur (BD Biosciences) flow cytometer.

In vitro phagocytosis

Mouse peritoneal macrophages were used for phagocytosis experiments, basically as described elsewhere [120]. Fresh or stored human and mouse RBCs, or eryptotic mouse RBCs, were then incubated together with adherent macrophages for 1 h at 37°C, 5% CO2. Incubation was carried out with or without the addition of 10% FCS or human serum (HS). The effect of the class A-SR-inhibitors fucoidan and dextran sulphate, or the control substance chondroitin sulphate, was also evaluated. For signaling inhibitor experiments, macrophages were pre-incubated for 30 min in the presence or absence of the phosphatidylinositide 3-kinases (PI3-kinase)–inhibitor

24

LY294002 (50µM), the P38 mitogen-activated protein kinase (p38 MAPK)- inhibitor SB203580 (30 µM), or the extracellular signal-regulated kinase-1/2 (ERK1/2)-inhibitor PD98059 (30 µM). When phagocytosis was terminated, non-ingested RBCs were lysed with dH2O for 30 sec, followed by washing, fixation and staining, using May-Grunewald/Giemsa. In every experiment, at least two technical replications per condition were made. Macrophages with ingested RBCs were then counted to generate a phagocytosis index by dividing the number of phagocytized RBCs with the total number of macrophages.

Ethics

The human transfusion study was conducted in accordance with the declaration of Helsinki and approved by the Swedish central ethics committee (Dnr Ö 1-2009). For in vitro experiments, human RBCs were acquired from blood bags (containing saline-adenine-glucose-mannitol (SAG-M), kept at a constant temperature of +4˚C and stored in the normal deposit at the Blood Bank at Norrlands University Hospital (Umeå, Sweden). A part of a plastic hose, filled with ~1 ml RBCs, was welded off from blood bags stored for up to six weeks. The acquired samples were anonymized and carried only information of blood group and the date of donation. Since the results obtained in the present study could not be traced back to a specific blood donor, ethical approval was not needed according to Swedish legislation (41, SFS 2003:460). All animal experiments were approved by Umeå research animal ethics committee (A14-12).

Statistical analysis

Two tailed student’s paired and unpaired t-test was used for comparisons between samples and groups respectively, when normal distribution could be assumed. Wilcoxon´s related samples signed rank test were used of paired comparisons when data were not normally distributed (IBM SPSS statistics Base 19, IBM, Armonk, NY, USA and JMP 11.0, SAS Campus Drive, Cary, NC).

25

Results

Paper 1

Transfusion of liquid-stored RBCs has been associated with worsened outcomes after transfusion [92, 121]. In addition, an animal study has shown that transfusion of the equivalent of 2 units of liquid-stored RBCs resulted in a profound proinflammatory response, suggested to be the result of an increased RBC uptake by macrophages shortly after transfusion [101]. To investigate the response to transfusion of cryopreserved RBCs in humans, we transfused 2 units of cryopreserved RBCs into 10 healthy human volunteers. Indicators of erythrophagocytosis, inflammation and the number of leukocytes in blood were assessed at 2 h and 48 h after transfusion and compared with that at 2 h before transfusion.

Transfusion of cryopreserved RBCs leads to acute accelerated RBC breakdown As expected, these RBC transfusions resulted in a significant increase in blood Hb. The number of circulating neutrophil granulocytes was significantly higher at 2 h and 48 h after transfusion. This was accompanied by an increase in circulating monocytes at 48 h after transfusion, thus indicating a transfusion-mediated effect on the immune system. We also found evidence of an acute clearance of transfused RBCs as the serum iron concentration, the transferrin saturation and the serum bilirubin concentration were all significantly increased at 2 h post transfusion, but had then returned toward baseline at 48 h. We further noticed a decrease in serum haptoglobin levels at 2 h after transfusion, which again returned towards baseline levels at 48 h, indicating that part of the hemolysis was in fact intravascular.

Absent pro-inflammatory response in the acute phase after transfusion After concluding that an acute hemolysis occurred after transfusion, we next investigated if there were any signs of inflammation. For this, we screened for known pro-inflammatory cytokines; IL-1β, IL-6, IL-8, MCP-1, MIP-1β, TNF-α, and the anti-inflammatory cytokine IL-10. None of the investigated cytokines increased in serum samples, neither at 2 h nor at 48 h after transfusion.

26

Additional results Paper 1 is based on data that originates from a larger study aimed at evaluating the enhanced aerobic capacity as a result of autologous transfusion and to identify biological markers of blood doping. Maximal aerobic capacity (VO2max) was evaluated before and after blood donation as well as reinfusion. Donation of 900 ml of blood led to a significant reduction in VO2max of 8±3%. Surprisingly, this reduction was not compensated for during the 15 weeks of recovery after donation even though Hb concentration was restored (Figure 6). In contrast, reinfusion of the cryopreserved RBCs led to a 13±8% increase in VO2max, close to the starting value.

) 4 Maximal oxygen uptake (VO ) % 2max ( 2 e u l 0 a v

l -2 a i

t -4 i n i -6

m *

o -8 r f -10 * e

g D R

n -12 a

h -14 C 0 2 4 6 8 10 12 14 16

) 10

% Blood Hb concentration ( e u

l 5 a v l a

i 0 t i n i

m -5 o r f e

g -10

n D Control R

a * Transfusion h C -15 0 2 4 6 8 10 12 14 16 Week Figure 6. Restoration of [Hb] 15 weeks after donation does not result in restored maximal aerobic capacity (Vo2MAX).

VO2MAX and blood Hb concentration was measured before and after blood donation as well as reinfusion. Values for control group (dotted line, n=7) and transfusion group (continuous line, n=9), are given as means±SEM. Unpaired student’s t-test was used for group comparisons. Indicated is the time of blood donation (D) and time of RBC reinfusion (R).

27

Paper 2

Eryptosis, a Ca2+ dependent stress-induced cell death in RBCs, results in a shorter than normal lifespan and could theoretically be a part of the RBC storage lesion. However, the understanding on how eryptotic RBCs are taken up by phagocytic cells in the spleen is virtually unknown. Thus, this study aimed at first investigating if there were any signs of eryptosis in liquid store RBCs, and second to investigate what phagocytic populations of the spleen that mediate uptake of eryptotic cells.

Stored human RBCs contain a subset of Ca2+-high cells It is currently unknown if eryptosis plays a role in the storage lesion. Because calcium uptake is a hallmark of eryptosis, we tested if liquid stored human RBCs display an increased Ca2+ permeability. RBCs, stored in SAG-M for 0, 2, 4 and 6 weeks were labelled with the FLUO-3 calcium indicator followed by incubation in Ca2+ containing medium for 1 h at 37°C. At the beginning of storage, human RBCs stored in SAG-M contained a low fraction of cells with an increased cytoplasmic Ca2+-level (0.7±0.3%). However, we found a time- dependent increase in the fraction of cells with increased Ca2+, reaching 3.6±0.8% after 6 weeks of storage. This suggests that part of the storage lesion could be associated with some RBCs having an eryptotic phenotype.

Characteristics of eryptotic murine RBCs A mouse model was used to investigate the clearance of transfused eryptotic cells, as this is difficult to assess in humans. It was found that RBCs treated with 1 µM ionomycin increased their intracellular Ca2+ concentration and that around 50% of the cells were PS+ as accessed with Annexin V binding. This was accompanied by a reduction in cell size as seen by lower FSC in flow cytometric analysis.

In vivo uptake of Ca2+-RBCs is located to the splenic marginal zone in a CD47-independent manner Transfusion of untreated RBCs, labeled with PKH26 or biotin, resulted in a rather linear clearance over the entire expected lifespan of 50 days, where the splenic uptake of these cells was located to F4/80+ macrophages in the red pulp area . The fact that the clearance kinetics of PKH26 or biotin labeled cells was virtually identical indicated that the labeling per se had no effect on the removal of the transfused cells. In contrast to that for freshly isolated RBCs, transfusion of PKH26-labeled Ca2+-RBCs (ionophore treated) resulted in a swift clearance of the majority of the transfused cells within 30 minutes. The splenic uptake of Ca2+-RBCs was located to the MZ, and appeared to be preferentially mediated by MARCO+ MZ macrophages (MZM). Uptake of Ca2+-RBCs was also seen in CD11c+ DCs in the MZ at 1 h after transfusion. At

28

24 h after transfusion, we also found CD11c+ DCs that were positive for PKH26 in the T-cell area of the splenic white pulp. Flow cytometric analysis at 24 h after transfusion revealed that more than 20% of the total CD11chi DCs, and about half of the CD11chi/CD207hi DC subset, were positive for PKH26+ Ca2+-RBCs. The uptake of Ca2+-RBCs by macrophages or DCs was not affected by CD47 on the RBCs since CD47-/- Ca2+-RBCs were equally ingested by peritoneal macrophages in vitro, or by DCs in vivo, as compared with that for CD47+/+ Ca2+-RBCs.

Additional results Murine RBCs, labeled with PKH26 and stored for 10 days were injected in to recipient mice. Spleens were harvested after 24 h and cryosections were labeled with antibodies against of MARCO+ MZM and F4/80+ red pulp macrophages. The splenic uptake of stored RBCs was found to co-localize with both F4/80+, as well as MARCO+, splenic macrophages (Figure 7.)

Figure 7. Splenic uptake of liquid stored murine erythrocytes is located both in the marginal zone and the red pulp. PKH26+ stored erythrocytes were shown to be taken up by both F4/80+ red pulp macrophages (top row), and MARCO+ MZM (bottom row). Co-location of erythrocytes and macrophages is seen as yellow on merged pictures to the right.

29

Paper 3

Much is still unknown regarding the storage lesion and the mechanisms mediating rapid RBC elimination after transfusion. The lack of detailed knowledge can in part be explained by ethical limitations in human studies. To circumvent these limitations, we have adopted a murine model of RBC storage and transfusion. This model was designed to mimic that of the human setting as closely as possible, e.g. by using CPD and SAG-M in the collection and storage of the RBCs. In addition to the in vivo model of RBC transfusions in mice, we also investigated if scavenger receptors were involved in the phagocytosis of stored murine or human RBCs, using an in vitro mouse macrophage phagocytosis assay.

Characteristics of the murine model for storage and transfusion Murine RBCs have a shorter lifespan and are generally considered as more fragile than human RBCs [116], which was also reflected in the shorter storage time needed to induce the storage lesion (10 days for mouse RBCs and 42 days for human RBCs). At this time point, stored murine RBCs displayed similar characteristics as previously described for stored human RBCs, including about 50% reduction in ATP content and formation of MP. In addition, about 35% of the stored murine RBCs were eliminated from the circulation within 24 h, while the RBCs surviving this time period showed normal long-term clearance kinetics after transfusion.

Phagocytosis of stored human and mouse RBCs is serum dependent and mediated by scavenger receptors The initial rapid clearance kinetics of stored RBCs likely represents molecular changes caused to a fraction of the RBCs during storage, which are efficiently recognized by macrophages leading to phagocytosis after transfusion. We hypothesized that these changes may not be species- specific, but rather that the recognition mechanisms could be part of the evolutionary conserved systems that identify damaged cells. To test this hypothesis, both human and murine RBCs (fresh or stored) were incubated with murine peritoneal macrophages in an in vitro phagocytosis assay. When macrophages were incubated with fresh RBCs, neither murine nor human RBCs were phagocytized. In contrast, stored murine and human RBCs were readily phagocytized by macrophages in the presence of heat inactivated FCS (and heat inactivated HS for human RBCs). Interestingly, phagocytosis of stored murine or human RBCs was markedly reduced by 70-75% under serum-free conditions. To investigate if phagocytosis of stored RBCs was dependent on class A-SR, we studied phagocytosis in the presence of the class A-SR ligands fucoidan and dextran sulphate. The results showed that both fucoidan and dextran sulphate inhibited phagocytosis of stored murine

30

or human RBCs by more than 90%. Phagocytosis of stored RBCs was also significantly reduced in the presence of the PI3 kinase-inhibitor LY294002. In addition, the ERK1/2-inhibitor PD98059 and the p38MAPK-inhibitor SB203580 each reduced phagocytosis of stored RBCs by 30% and 36%, respectively.

Additional results

Discrimination of RBC age fractions by in vivo biotinylation Double in vivo biotinylation was carried out with 25 days in between biotin injections. This enabled the RBCs to be separated into age fractions. Biotinhi cells represented RBCs >25 days old and biotinint cells represented the remaining RBCs which were <25 days old (figure 8A). On day 26, the mice were bled and RBCs were isolated before transfusion into recipient mice. Blood samples from transfused mice showed a close to linear clearance of >25 days old biotinhi cells during the first 25 days after transfusion. The <25 days old biotinint RBCs were, as expected, initially cleared slower and could be detected up to 55 days after transfusion (figure 8B).

A. B.120 >25 days old <25 days old 100 Total >25 days old 80 PE - <25 days old 60

40 RBC recovery RBC recovery (%) Streptavidin 20

0 0 20 40 60 FSC Time (days) Figure 8. The use of biotin to label circulating erythrocytes in donor mice and subsequent analysis of their clearance following transfusion into recipient mice. A, Donor mice were injected i.v. with 1.1 mg biotin at day 0, 1 and 2. At day 25, another biotin injection of 0.4 mg was made. On day 26, mice were bled, erythrocytes were stained with streptavidin-PE and analyzed by flow cytometry. A representative dot-plot shows that >25 days old biotinhi erythrocytes can be separated from <25 days old biotinint erythrocytes. B, Recovery of freshly isolated >25 days old (closed symbols) or <25 days old RBCs (open symbols) after transfusion into syngeneic recipient mice. Grey symbols represents recovery of all transfused RBCs. Data are mean±SD of three separate recipient mice.

31

Role of the RBC age at the start of storage for their 24 h survival post transfusion RBCs from double in vivo biotinylated mice, freshly isolated or stored for 10 days, were transfused into recipient mice in order to investigate if RBCs >25 days old at the start of storage would be more severely affected by storage and therefore eliminated to a larger extent during the first 24 h after transfusion. The 24 h recovery of freshly transfused RBCs was high among both younger and older fractions (figure 9 A and B). Interestingly, when 10 days stored RBCs were transfused, we found that virtually all cells disappearing rapidly within the first 24 h belonged to the fraction of RBCs >25 days old at the start of storage (figure 9 B). When comparing the survival kinetics of the fraction of >25 days old freshly transfused or stored RBCs that remained in circulation at 24 h after the transfusion, we found that old fresh and stored RBCs had virtually similar recovery and expected life span (figure 9 C).

A Young RBC B Old RBC C RBCs remaining after 24 h 120 Fresh 120 Fresh 120 Stored Stored Old Fresh 100 100 100 Old stored

80 80 80

60 60 60

40 40 40 RBC recovery RBC recovery (%) RBC recovery RBC recovery (%) RBC recovery RBC recovery (%) 20 20 20

0 0 0 0 10 20 30 40 Time (days)

Figure 9. Virtually all of the stored murine RBCs cleared within 24 h belong to the fraction of older RBCs. Murine RBCs were labeled twice with biotin in vivo to enable separation of RBCs with an age of <25 days (young) or >25 days (old). Isolated RBCs were either freshly transfused (Fresh) or stored in SAG-M and autologous plasma at +4°C for 10 days before transfusion. (A) Recovery of fresh or stored younger RBCs, or (B) fresh or stored older RBCs, at 24 h after transfusion. (C) Recovery kinetics of older fresh or older stored RBCs, where the amount of each RBC fraction remaining in circulation at 24 h after transfusion was set to 100 per cent. Data are means±SD of 4 separate recipient mice.

Phthalates during storage Plastic softeners or phthalates are known to be released from blood bags in to the membranes of the stored RBCs [122-124]. Trying to investigate a potential effect of phthalates in our murine storage model, RBCs were stored in test tubes coated with or without phthalate containing PVC-plastic film.

32

Transfusion of phthalate stored RBCs demonstrated near identical 24 h survival characteristics as RBCs stored without phthalates (Figure 10A.) This finding could indicate that plastic softeners may not affect murine RBCs during storage with respect to their chances of surviving after transfusion.

CD47 on stored murine RBCs does not influence the rapid elimination that occurs during the first 24 h The cell surface protein CD47 has been shown to serve as an anti-phagocytic ligand, protecting RBCs from being phagocytized by macrophages [52]. Furthermore, the amount of CD47 on the cell surface has been implicated to decrease during storage of human RBCs [83]. To test if a reduction in CD47 on stored RBCs could affect the rate of rapid 24 h elimination after transfusion, stored RBCs from Wt (CD47+/+) or CD47+/- mice (in which the CD47 expression is 50 % of that in CD47 Wt) were transfused into Wt mice. The 24 h survival was nearly identical when comparing Wt and CD47+/- RBCs, indicating that CD47 may not be involved in regulating the acute phagocytosis of stored erythrocytes (figure 10B).

A 120 B 120

100 100

80 80

60 60

40 40 RBC recovery RBC recovery (%) RBC recovery RBC recovery (%) 20 20 Phthalates CD47+/+ RBCs Control CD47+/- RBCs 0 0 0 4 8 12 16 20 24 0 4 8 12 16 20 24 Time (hours) Time (hours) Figure 10. Acute clearance of stored murine RBCs is not affected by phthalates of RBC CD47. (A) RBCs were stored in the presence or absence of phthalates before transfusion to syngenic mice. Data are means±SD of 2 separate recipient mice. (B) Transfusion of stored Wt (CD47+/+) or CD47+/- RBCs displays near identical RBC recovery during the first 24 h after transfusion Data are means±SD of 4 separate recipient mice.

33

Discussion

Frozen vs. liquid stored RBCs

At present, the vast majority of RBC concentrates are stored under liquid conditions [125]. Cryopreservation of RBCs are mainly restricted to donors with rare blood groups and for autologous transfusions [126]. Cryopreserved RBCs have been shown to be spared from many of the hallmarks of the liquid storage lesion including a reduction in ATP and 2,3-DPG and the formation of MPs [87-89]. Since little is known about the clinical consequences of the storage lesion, it is problematic to evaluate the two storing methods based on in vitro characteristics, but the phenotype of cryopreserved RBCs suggest that the method may have advantages over liquid storage. On the other hand, the time consuming procedure (3-4 hours) of thawing RBCs limits its use to planned procedures. Our data indicated that transfusion of two units of cryopreserved RBCs had a modest effect on the Hb concentration, which was only significantly elevated at 2 h after transfusion (from 145.4±4.9 g/L to 151.1±4.6 g/L). After 48 h, it had dropped 3 g/L and was no longer significantly different from the Hb concentration before transfusion. On the other hand, aerobic capacity assessment at 48 h after transfusion, revealed a 13% increase in VO2max as compared to values obtained the week before transfusion (Figure 6). Thus, differences in the venous Hb concentration may not be the best assessment of neither RBC volume nor aerobic capacity after RBC transfusion to non-anemic men, possibly explained by the capacity of the spleen to store RBCs [127, 128].

Inflammation after transfusion?

There is at present a controversy regarding risks with transfusion of RBCs stored for prolonged periods of time [92, 93]. Alarming results from animal studies have also emerged, in were Hod et al. [101] could show a profound pro-inflammatory and the presence of NTBI 2 h after transfusion of the equivalent of two units of stored RBCs into mice. The cytokine release was suggested to originate from the large amount of iron delivered to macrophages upon the rapid phagocytosis of a fraction of the transfused RBCs [103]. However, no clear evidence has been presented for such a response in humans, with the exception of a report in preterm infants [129]. To test if increased erythrophagocytosis alone could provoke such a response in healthy humans, we conducted experiments where human subjects received two units of cryopreserved RBCs. Although the storage conditions was not the same (frozen vs. liquid stored), it has been shown that approximately 15 % of the previously frozen, transfused RBCs are rapidly

34

cleared from circulation upon transfusion [88, 130]. Indeed, our results also showed a marked increase in serum bilirubin, iron and transferrin saturation at 2 h post transfusion, which all indicate increased erythrophagocytosis. However, the rise in transferrin saturation (from 28.1 ±1.3 % to 43.2 ±1.5 %) did not indicate that non-transferrin bound iron were present in the plasma since most of the transferrin were still not saturated with iron. It can however not be ruled out that the peak of serum iron came about after the measurement at 2 h post transfusion and that NTBI might have been present at a later time point. A human study on transfusion of one unit of liquid stored RBCs, stored for 42 days, to healthy recipients showed a larger increase in transferrin saturation than observed in our study [104]. That study also detected NTBI in plasma after transfusion, indicating that there could be a difference as to how frozen and liquid stored RBCs are received after transfusion [104]. This discrepancy is probably related to the storage lesion which primarily is restricted to liquid stored RBCs.

To elucidate if the hemolysis was primarily intra- or extra-vascular (i.e. if the red blood cells burst in circulation or were phagocytized in a controlled way by macrophages), we measured the levels of haptoglobin in serum. Haptoglobin binds free Hb and a diminished plasma level of haptoglobin is an indirect marker for intravascular hemolysis, although profound extravascular hemolysis could also lead to reduced haptoglobin levels [131]. We found that haptoglobin was reduced by 0.18 g/L at 2 h after transfusion, which corresponds to the binding of approximately 0.8 g of Hb [132, 133]. Since the total amount of Hb transfused was about 80 g, intravascular hemolysis only accounted for 1% of the transfused RBCs, strongly indicating that the majority of the hemolysis was in fact extravascular. After concluding that we had an acute, extravascular hemolysis, we investigated the presence of pro-inflammatory cytokines in the plasma of transfused individuals, in search of similarities with findings in the above mentioned animal study. None of the investigated pro-inflammatory cytokines IL-1β, IL-6, IL-8, MCP- 1, MIP-1β or TNF-α increased in plasma at 2 h or 48 h after transfusion. Similar findings were also reported after transfusion of one unit of liquid stored RBCs in healthy recipients [104]. The discrepancy in the data concerning signs of inflammation between animal and human studies could be due to differences in the storage and/or administration procedure, but also due to species differences [112].

The mouse model for storage and transfusion

In order to expand our understanding on post transfusion events related to the storage lesion, we turned to an animal model. There are always drawbacks when using model systems, and one must be cautious when

35

transferring findings in animal models to human physiology [112]. Nevertheless, due to ethical considerations, an animal model can present opportunities to study mechanisms that are not easily assessed in human subjects. The use of inbred mice is an attractive alternative due to practical considerations and the genetic kinship between man and mouse [111]. Together with the increasing availability of genetic manipulations, the mouse model offers unique possibilities compared to humans investigations.

The use of a mouse model in RBC storage and transfusion has been shown to resemble human settings [134], albeit with shorter storage time of the RBCs prior to transfusion. As expected, we observed that 10 days of storage resulted in an ATP reduction, MP formation and 24 h survival kinetics resembling that found in human RBCs after prolonged storage. Human RBCs are stored in PVC-plastic bags that contain plasticizers (phthalates), known to leak into the membranes of the RBCs and thus improve their storage characteristics as they reduce the hemolysis incidence [122-124]. In an attempt to evaluate the effect of phthalates on mouse RBC storage, we coated the storage test tube with PVC-plastic film prior to the addition of the cells. RBCs stored in phthalate plastic showed similar 24 h clearance kinetics as control RBCs stored without phthalates (Figure 10A), indicating that plasticizers may not have an major influence on the 24 h survival of stored mouse RBCs after transfusion in our model.

The role of calcium uptake in the storage lesion

Increased intracellular Ca2+ is a hallmark of eryptotic cells, which leads to PS externalization and the subsequent uptake by phagocytic cells [40]. The observation that a small fraction of RBCs had an increased Ca2+ permeability after 6 weeks of storage suggests that eryptosis could be a part of the RBC storage lesion. It is not known if the fraction of Ca2+ RBCs could further increase after transfusion, but the observation that human stored RBCs starts to externalize PS after overnight incubation at 37°C suggests the possibility that transfused cells may acquire additional removing signals after transfusion [135].

The spleen consists of distinct populations of phagocytic cells with different functions and locations [56]. By studying the splenic uptake of experimentally induced eryptotic cells after transfusion, we found that the vast majority of RBCs with increased Ca2+ were taken up in the marginal zone (MZ), preferentially by MARCO+ MZM. This is in contrast to the uptake of normally senescent RBCs, which generally are phagocytized by F4/80+ red pulp macrophages. One explanation to the different micro-anatomical locations of phagocytosis could be that a substantial amount of the arterial

36

blood in the murine (but also human) spleen flows through the MZ before entering the red pulp, placing phagocytic cells of the MZ first in line to recognize the eryptotic RBCs. The uptake of stored murine RBCs was found to be located both in the marginal zone and in the red pulp (Figure 7) and as such, agrees with the hypothesis that a part of the storage lesion could in fact be explained by eryptosis. However, it also supports the view that the storage lesion is multifactorial, and that other RBC changes lead to their recognition by the red pulp macrophages.

Impact of storage on RBCs - Senescence, eryptosis, or both?

To further understand the effect of the storage lesion, we applied the double in vivo biotinylation technique [119], in which RBCs in circulation can be separated with regard to their age (Figure 8). This grants the ability to follow younger and older RBCs separately, in order to evaluate the effect of the storage lesion in separate age fractions. Interestingly, we found that almost the entire loss of transfused stored RBCs during the first 24 h involved the fraction of older RBCs (>25 days of age prior to storage), which strongly indicates that the storage lesion has a more profound effect on aged RBCs (Figure 9A, B). Markedly, the older cells that remained in circulation after 24 h showed similar clearance kinetics as freshly transfused RBCs of the same age (Figure 9C). This therefore suggests that the storage lesion could have a more profound effect on older RBCs, and that cells surviving the first 24h after transfusion seem to be rather unaffected by the storage lesion.

It is known that some of the RBCs having reduced levels of ATP and 2,3-DPG after storage may restore these levels within 24 h after transfusion [85]. Such recovery could theoretically also be connected with restoring some of the RBC cell surface changes that are involved in macrophage recognition and clearance. Furthermore, it has been suggested that the shedding of PS+ MPs from stored RBCs could reduce the amount of the externalized PS on the RBCs and thus reduce the risk of phagocytosis of these cells [135]. To investigate the possibility that some of the stored RBCs, which are normally rapidly eliminated after transfusion, may have the capacity to recover in the absence of macrophages, one could deplete recipient mice of macrophages in the spleen and liver before transfusion. The efficient depletion of macrophages in contact with the bloodstream by i.v. injection of clodronate liposomes has been demonstrated [136]. An improvement in 24 h recovery of stored RBCs in clodronate depleted macrophages have also been shown [137]. By transfusing biotin-labeled stored RBCs, we could investigate how different age fractions among the transfused cells remain in circulation over time until macrophages start to return around day 6, and elucidate if all of the rapidly cleared transfused RBCs are taken up by macrophages.

37

Transfused RBCs remaining in circulation can also be recovered by FACS- sorting for further biochemical analysis.

How are stored RBCs presented to the immune system upon transfusion?

DCs are professional APCs that trap foreign or endogenous antigens by phagocytosis or macropinocytosis. The antigens are then intracellularly processed, attached to MHC, and finally presented on the cell surface together with costimulatory molecules (Figure 2). Depending on the phenotype of the DC, and the recognition pattern of the antigen, the DC will present the antigen in either an immunogenic or tolerogenic way [15, 138]. The balance between tolerance and immunity can further be affected by the inflammatory state of the organism. Alloimmunization has been shown to be propagated after RBC transfusion of transgenic hen egg (HEL)- RBCs in mice, but only after co-stimulation with the immunostimulant poly(I:C) [139]. In contrast, the administration of LPS together with HEL- RBC transfusion not only failed to increase alloimmunization, but was also found to inhibit the same process [140]. Thus, inflammation as such may be a too broad of a label to use when assessing risks of transfusions since (mimicked) viral or bacterial inflammatory stimuli seem to have quite different effects on immunomodulation in relation to alloimmunization, at least in mice. This also highlights the importance of finding the receptor(s) involved in recognition of stored RBCs by the immune cells.

A subset of DCs which express the cell surface proteins CD103 (integrin αE) and CD207 (langerin) (CD207+ DCs) has been identified as responsible for the uptake of apoptotic blood borne cells in the marginal zone of the (mouse) spleen [141]. Their function is to maintain self-tolerance by migrating to the T-cell area of the spleen and present self-antigens to naïve T-cells. Depletion of MZ has been shown to result in loss of immunosuppression against apoptotic cells [142]. After stimulation with either poly(I:C) or LPS, the number of CD207+ DCs was found to be drastically reduced in murine spleens [141]. The finding in paper 2 of the present thesis, that stored human RBCs may contain a population of eryptotic cells and that experimentally eryptotic RBCs were recognized and taken up by CD207+ DCs in the spleen, therefore opens the door for a somewhat unexpected possibility: that part of the storage lesion may in fact have beneficial effects in terms of propagating tolerance rather than immunity. Inflammatory stimuli are likely to interfere with maintenance of this tolerance, again bringing the role of inflammation after transfusion of stored RBCs to life. In a next step, it would be intriguing to transfuse PHK26+ stored RBCs, collect spleens and (by flow cytometry) determine to which degree different macrophage and DC subsets ingest

38

stored RBCs. Different activation markers on PKH26+ phagocytic cells can then be assessed in order to understand the immunologic impact of transfusion of stored RBCs.

CD47 in senescence, eryptosis and the storage lesion

The presence or absence of CD47 on experimentally eryptotic RBCs affected neither macrophage phagocytosis in vitro nor DC uptake of these cells in vivo. There have been indications in the literature that senescent, and possibly also stored RBCs, may have reduced amounts of CD47 on their surface [36, 83]. We found that CD47+/- RBCs (expressing half of normal CD47 levels) had virtually identical clearance kinetics as normal CD47+/+ RBCs after storage and transfusion (Figure 10B). The long-term survival of CD47+/- RBCs was on the other hand slightly reduced as compared to CD47+/+ controls (data not shown). This could support the hypothesis that senescent RBCs are recognized by phagocytic cells after opsonization by NAbs, since IgG-opsonized CD47+/- RBCs were found to be cleared more rapidly after transfusion than equally IgG-opsonized CD47+/+ RBCs in vivo [143]. We have not specifically investigated the effect of CD47 in our phagocytosis assay of stored RBCs. However, since macrophage phagocytosis of stored RBCs could be efficiently blocked by addition of SR inhibitors , and SR-mediated phagocytosis of oxidatively damaged RBCs has been shown to be CD47 independent [120], there are no strong indications from our studies that CD47 (or lack thereof) is influencing macrophage recognition of stored RBCs.

Furthermore, the CD47-SIRPα interaction has been shown to be dependent on RBC membrane flexibility, which could imply that the ability of CD47 to inhibit phagocytosis could be further limited in senescent or stored RBCs [144]. In addition, an alternative role of the CD47-SIRPα interaction has been presented by Burger et al. [145], who showed that the serum protein thrombospondin-1 (TSP-1) could bind to CD47 on senescent or stored RBCs to promote their phagocytosis. A conformational change in CD47 enabled binding of TSP-1 which resulted in SIRPα-dependent phagocytosis. This indicates that the CD47-SIRPα interaction can have both pro-, and anti- phagocytic properties, thus making quantification of total CD47 per se an unsuitable indicator to predict the uptake of senescent and/or stored RBCs.

Scavenger receptor mediated phagocytosis of stored RBCs

Scavenger receptors are evolutionary conserved receptors that can mediate endocytosis and phagocytosis of molecules or cells carrying certain molecular signatures [13]. Fucoidan and dextran sulphate are ligands for

39

class A-SRs and can efficiently block binding of other ligands to these receptors [146]. Interestingly, we found that both fucoidan and dextran sulphate strongly inhibited phagocytosis of stored human as well as murine RBCs in vitro. Thus, class A-SRs may be a previously unappreciated class of receptors involved in macrophage phagocytosis of RBCs damaged by storage. It has previously been shown that fucoidan and dextran sulphate can inhibit phagocytosis of oxidatively damaged murine RBCs [120]. Inhibition of the signal transduction proteins PI3-kinase, p38-MAPK and ERK1/2 was shown to affect phagocytosis of IgG opsonized or oxidized RBCs [120]. Since we also found that the same inhibitors reduced the uptake of stored RBC, this could suggest that oxidative stress may in fact be one important contribution to the storage lesion with respect to the 24 h recovery of transfused cells. PI3- kinase is essential for phagocytosis of endogenous material as well as pathogens [147], and has been suggested to also regulate phagocytosis of stored RBCs [137]. Scavenger receptor class A-1 (SRA-1 or CD204) lacks a signaling cytoplasmic domain and may only mediate binding of damaged RBCs to the macrophage, while the intracellular signaling resulting in phagocytosis could be contributed by other receptors binding to the same RBC [148]. Ligation of fucoidan to SRA-1 has been shown to enhance TLR-4 mediated activation of nuclear factor-κβ (NF- κβ), resulting in TNF-α and IL-1β production by LPS stimulated murine peritoneal macrophages [149]. Ligation of fucoidan alone did not, however, mediate an inflammatory response. Importantly, the impact of this recognition mechanism in vivo, as well as in the macrophage subpopulations normally involved in elimination of stored RBCs in spleen or liver, has to be further investigated in both the murine and human system.

The role of serum in phagocytosis of stored cells

Both human and murine stored RBCs required heat-stable serum factors for efficient phagocytosis by macrophages in vitro, since removal of heat- inactivated FCS strongly reduced RBC uptake. Phagocytosis of stored human RBCs could also be restored by heat-inactivated human AB serum. Thus, heat-stable serum factors seem to be required for macrophage phagocytosis of stored RBCs. The fact that serum was required for phagocytosis of both human and murine stored RBCs, and that stored RBCs of both species were similarly recognized by murine macrophages, suggest that macrophages can use evolutionary conserved species-independent mechanisms for uptake of stored RBCs. Complement is not a likely candidate, since it is destroyed by heat-inactivation [150]. However, several serum proteins have been described to function as bridge molecules in macrophage recognition of damaged or apoptotic cells [2, 151]. Thus, further investigations will be

40

required to narrow down on the serum component(s) and corresponding receptor(s) that promote phagocytosis of RBCs affected by storage.

Concluding remarks The findings in paper 3 of this thesis indicate that the recognition of liquid stored RBCs is mediated by class A scavenger receptors. SR-A1 is a possible candidate which is expressed on many different macrophage populations, including murine peritoneal macrophages [149], which were used in our experiments. Inhibition of phagocytosis in our in vitro model, both at receptor and signaling transduction level, were shown to resemble that found for oxidatively damaged RBCs [120]. Together with our observation that the oldest RBCs are more prone to be rapidly removed after transfusion, this indicates that oxidative damage may very well lead to important phenotypic changes of stored RBCs that ultimately leads to their swift recognition after transfusion.

The phagocytosis of stored murine or human RBCs was found to be serum dependent. In addition, while SR-A1 lacks a signaling domain, PI3 kinase, p38MAPK and ERK1/2 were found to be important for phagocytosis. This suggests that (an)other receptor(s), possibly requiring serum proteins, could provide signaling and partner with SR-A1 for phagocytosis of stored RBCs (Figure 11). In contrast to phagocytosis of opsonized RBCs, CD47 does not seem to inhibit SR mediated RBC phagocytosis [120]. Our observation that stored CD47+/- RBCs (expressing half of normal CD47 levels) were cleared at the same rate as normal (CD47+/+) RBCs, strengthen the hypothesis for the involvement of SR in uptake of stored RBCs.

Stored RBC

Serum protein

SR (SR-A1?) Receptor x

Macrophage

PI3 kinase P38 MAPK ERK1/2

Figure 11. Working hypothesis for a two-receptor model mediating phagocytosis of stored RBCs.

41

Conclusions

This thesis is based on human and murine in vivo and in vitro studies. From human studies it is concluded:

• Increased erythrophagocytosis did not provoke a pro-inflammatory response in healthy men, thus questioning the iron hypothesis that have been suggested to occur after transfusion.

• A subset of liquid stored RBCs display a time dependent increase in eryptotic-like phenotype.

From studies conducted in the murine model system it is concluded:

• Eryptotic cells are phagocytized in the splenic marginal zone by macrophages and DCs in ways resembling uptake of apoptotic cells.

• Liquid-stored RBCs is phagocytized by macrophages in the marginal zone and red pulp of the spleen.

• The oldest fraction of liquid-stored murine RBCs is responsible for the reduced 24 h survival seen after transfusion. A reduction of CD47 on the cell surface did not affect the acute clearance of stored RBCs after transfusion.

• In vitro phagocytosis of stored human and mouse RBCs is serum dependent and can be blocked by class A scavenger receptor inhibition.

42

Acknowledgements

During my time as a PhD student, I have had the privilege to meet and work with many interesting colleagues to whom I would like to express my gratitude. First, to my supervisor, Christer Malm who finally managed to get me accepted as a PhD student. I am sincerely grateful for your support over the years and for always having my back and allowing me to go my own way. You are a warmhearted man and an inspiring entrepreneur. And thank you for trying to teach me “good enough”. It has been helpful, even though I still struggle with accepting this concept. To my co-supervisor, Per-Arne Oldenborg for taking me in and letting me try my wings as an experimental researcher. I am grateful for the many discussions that we have had and for all the knowledge you have shared with me. You have always been patient and supportive, and allowed me to try to find the right way, even if it could take some time. To my former co-supervisor, Lars Frängsmyr, who introduced me to the field of proteomics where I started this scientific journey. Thank you for the technical support over the years and for all the scientific and philosophical discussions. To my co-author Fredrik Toss at the unit for transfusion medicine, Norrlands universitetssjukhus who have be extremely helpful and aided me in getting a clinical insight regarding my research as well as helping me getting material to do research on. I would also like to acknowledge Lilian Falk and staff at Blodcentralen in Boden, Staffan Wikstrand at Sunderby sjukhus, Hans Gulliksson, Stella Larsson and Linda Larsson and staff at Blodkomponenten, Karolinska universitetssjukhuset and Karin Skoglund and staff at Blodcentralen in Umeå, for all your help in acquiring and administrating blood samples and blood bags from different corners of Sweden. To the staff of Winternet in Boden, especially Ann-Sofie Lindberg and Anna Nordin for conducting the tests for aerobic capacity so skillfully. And to Eva Edholm for taking all the blood samples. To all my colleagues at the unit for sports medicine for making my years at the unit so memorable. A special thanks to my fellow PhD-student colleagues and friends Jonas Johansson and Tommy Henriksson, it would not have been the same without you! And of course to my friend Apostolos Theos for all your encouragement. And to my roommates, André Brännström and Andreas Isaksson for all your artistic pictures and collages that you have made… And my former roommate, Niklas Boman, I

43

really miss our philosophical discussions. And of course to Gunilla Solander who have been amazingly helpful and kind during my years at IME, you are truly missed. And to Britt-Marie Eliasson for your caring way and administrative help during the more intense parts of my teaching assignments. To Fia Lindberg for your support and inspiration, you are a true friend. To Jenny Lindehall Hadrévi, who helped me in getting my masters project, which ultimately led to me getting a PhD position. To Michael B. Svensson for being such an inspiring person and to Kajsa Gilenstam for your thoughtfulness and optimism. To Robert R.N 1,0 Nilsson for your extremely convincing argumentation skills. Or, come to think of it, your way of never leaving anything up for discussion. My appreciation also go to Håkan Alfredsson, Lars-Göran Fjällborg, Tom Pietilä, Monica Sörman, Jonas Lorentzon, Lasse Stigell, Lisbeth Wikström Frissén, Jiguo Yu, Claes-Göran Sundell, Ann Sörlin, Anna Nordström, Peter Nordström, Helena Gustavsson, Nils Berginström, Gabriel Högström, Stefan Sandlund, the staff of the HAI-project; David, Jim and Magnus, and the lab personnel; Lennart, Roger and Knappen for making IME such an great place to work at. And of course to my former coworkers Per Jonsson, CJ Boraxbekk (formally known as Olsson), Peter Forsgren, Erkki Jakobsson, Kicki Olsson and Anna Oscarsson as you guys most definitely contributed to the greatness of IME. I would also like to thank my friends and colleagues at IMB. In anatomy: To Anna-Karin Olofsson, I owe you so much for all the help and support over the years. To Ludvig Backman for letting me run a few gels once in a while but more important, for valuable discussions of non-scientific importance. To Gustav Andersson for your absolutely magnificent artwork on the cover of this thesis. To Mona Lindström, for being a good listener and occasionally talking some sense in to me. To Anton Tjust for stimulating academic discussions. To Lars-Eric Thornell for lending me a pair of shorts so that we were allowed on the golf court in Georgia, USA (together with the alligators). And to Lena Carlsson, Per Stål, Björn Lindell, Sture Forsgren, Partik Danielsson, Jingxia Liu, Fatima Pedrosa Domellöf, Yafeng Song, Eva Carlsson and Johan Bagge for being great colleagues. In histology: Shrikant Kolan, for always being so kind and helpful. I wish you the best of luck on your upcoming thesis defense. To Barbro Borgström for excellent technical assistance. To Anders Larsson, for sharing your methodological skills in the lab. And to Åsa Stenberg, Sara af Bjerkén, Ana Virel, Maria Chermenina, Sanaz Hashemian, Nina Nevalainen, Janne Ylärinne, Per Linström, Anna Rehnmark, Ingrid Strömberg, Göran Dahlgren,

44

Saowaluck Sompet, Jan-Ove Sehlin, Anita Dreyer-Perkiömäki, Ann-Lena Tallander and Selamawit Kefela. To Kristina Lejon, Radha Thyagarajan, and Viqar Banday at the unit for for interesting lab meetings and journal clubs. To Elisabeth, Jenny and Sandra at the animal facility for taking good care of the mice. And to Siggan and Sofia for great company when working with the mice. To Börje Ljungberg and Sofia Wård at the dept. of surgical and perioperative sciences.

Finally, I would like to acknowledge a few persons from outside the academia that are of special importance to me. To my fantastic sister Elin, and your wonderful family. You are like a second family for me. And to my niece and goddaughter Wilma. If you still want to be a scientist when you grow up, I will help you in any way I can. To my mother and father, thank you for everything you have given me. For all the encouragement and support over the years. I know that you have always believed in me and for that I am truly grateful. This is one step closer to the blue hall. To my dear friend Erik. Thank you for all the adventures and for your support when I have needed it. We have literally been through high and low together. To Marie, for being such a good friend and listener. And to the both of you for being an important part of my children’s lives. To Johanna, finally, the three years have passed. I wish you all the best in the future. Thank you for giving me the best things in my life, namely; Simon och Elvira. Jag är så stolt och tacksam över att ni finns i mitt liv!

45

References

1. Junqueira, L.C.U.a., J. Carneiro, and R.O. Kelley, Basic histology. 7th ed. 1992, Norwalk, Conn.: Appleton & Lange. x, 518 p.

2. de Back, D.Z., et al., Of macrophages and red blood cells; a complex love story. Front Physiol, 2014. 5: p. 9.

3. Cyster, J.G., B cells on the front line. Nat Immunol, 2000. 1(1): p. 9- 10.

4. Feldschuh, J. and Y. Enson, Prediction of the normal blood volume. Relation of blood volume to body habitus. Circulation, 1977. 56(4 Pt 1): p. 605-12.

5. Alberts, B., Molecular biology of the cell. 4th ed. 2002, New York: Garland Science. xxxiv, 1548 p.

6. Murphy, K., et al., Janeway's immunobiology. 7th ed. 2008, New York: Garland Science. xxi, 887 p.

7. Medzhitov, R. and C.A. Janeway, Jr., Decoding the patterns of self and nonself by the innate immune system. Science, 2002. 296(5566): p. 298-300.

8. Gordon, S. and P.R. Taylor, Monocyte and macrophage heterogeneity. Nat Rev Immunol, 2005. 5(12): p. 953-64.

9. Wynn, T.A., A. Chawla, and J.W. Pollard, Macrophage biology in development, homeostasis and disease. Nature, 2013. 496(7446): p. 445-55.

10. Dunkelberger, J.R. and W.C. Song, Complement and its role in innate and adaptive immune responses. Cell Res, 2010. 20(1): p. 34-50.

11. Takeuchi, O. and S. Akira, Pattern recognition receptors and inflammation. Cell, 2010. 140(6): p. 805-20.

12. Areschoug, T., J. Waldemarsson, and S. Gordon, Evasion of macrophage scavenger receptor A-mediated recognition by pathogenic streptococci. Eur J Immunol, 2008. 38(11): p. 3068-79.

46

13. Canton, J., D. Neculai, and S. Grinstein, Scavenger receptors in homeostasis and immunity. Nat Rev Immunol, 2013. 13(9): p. 621- 34.

14. Peiser, L. and S. Gordon, The function of scavenger receptors expressed by macrophages and their role in the regulation of inflammation. Microbes Infect, 2001. 3(2): p. 149-59.

15. Reis e Sousa, C., Dendritic cells in a mature age. Nat Rev Immunol, 2006. 6(6): p. 476-83.

16. Rutella, S., S. Danese, and G. Leone, Tolerogenic dendritic cells: cytokine modulation comes of age. Blood, 2006. 108(5): p. 1435-40.

17. Epstein, R.M. and E.M. Papper, Respiratory factors in the uptake and excretion of anesthetics. 1965. Int Anesthesiol Clin, 1998. 36(4): p. 121-35.

18. MacDonald, R., Red cell 2,3-diphosphoglycerate and oxygen affinity. Anaesthesia, 1977. 32(6): p. 544-53.

19. Dubinsky, W.P. and E. Racker, The mechanism of lactate transport in human erythrocytes. J Membr Biol, 1978. 44(1): p. 25-36.

20. Williams, W.J., Hematology. 4th ed. 1990, New York: McGraw-Hill. xxxvii, 1882 p., 6 p. of plates.

21. Smith, B.R., Regulation of hematopoiesis. Yale J Biol Med, 1990. 63(5): p. 371-80.

22. Mulcahy, L., The erythropoietin receptor. Semin Oncol, 2001. 28(2 Suppl 8): p. 19-23.

23. Hattangadi, S.M., et al., From stem cell to red cell: regulation of erythropoiesis at multiple levels by multiple proteins, , and chromatin modifications. Blood, 2011. 118(24): p. 6258-68.

24. Bosman, G.J., F.L. Willekens, and J.M. Werre, Erythrocyte aging: a more than superficial resemblance to apoptosis? Cell Physiol Biochem, 2005. 16(1-3): p. 1-8.

25. Ganz, T. and E. Nemeth, Iron : interactions with normal and disordered erythropoiesis. Cold Spring Harb Perspect Med, 2012. 2(5): p. a011668.

47

26. Pimstone, N.R., et al., The enzymatic degradation of hemoglobin to bile pigments by macrophages. J Exp Med, 1971. 133(6): p. 1264-81.

27. Lutz, H.U. and A. Bogdanova, Mechanisms tagging senescent red blood cells for clearance in healthy humans. Front Physiol, 2013. 4: p. 387.

28. Low, P.S., et al., The role of hemoglobin denaturation and band 3 clustering in red blood cell aging. Science, 1985. 227(4686): p. 531- 3.

29. Schwarz-Ben Meir, N., T. Glaser, and N.S. Kosower, Band 3 protein degradation by calpain is enhanced in erythrocytes of old people. Biochem J, 1991. 275 ( Pt 1): p. 47-52.

30. Franco, R.S., et al., Changes in the properties of normal human red blood cells during in vivo aging. Am J Hematol, 2013. 88(1): p. 44- 51.

31. Werre, J.M., et al., The red cell revisited--matters of life and death. Cell Mol Biol (Noisy-le-grand), 2004. 50(2): p. 139-45.

32. Willekens, F.L., et al., Hemoglobin loss from erythrocytes in vivo results from spleen-facilitated vesiculation. Blood, 2003. 101(2): p. 747-51.

33. Willekens, F.L., et al., Liver Kupffer cells rapidly remove red blood cell-derived vesicles from the circulation by scavenger receptors. Blood, 2005. 105(5): p. 2141-5.

34. Vaysse, J., et al., Adhesion and erythrophagocytosis of human senescent erythrocytes by autologous monocytes and their inhibition by beta-galactosyl derivatives. Proc Natl Acad Sci U S A, 1986. 83(5): p. 1339-43.

35. Bratosin, D., et al., Cellular and molecular mechanisms of senescent erythrocyte phagocytosis by macrophages. A review. Biochimie, 1998. 80(2): p. 173-95.

36. Khandelwal, S., N. van Rooijen, and R.K. Saxena, Reduced expression of CD47 during murine red blood cell (RBC) senescence and its role in RBC clearance from the circulation. Transfusion, 2007. 47(9): p. 1725-32.

48

37. Cohen, R.M., et al., Red cell life span heterogeneity in hematologically normal people is sufficient to alter HbA1c. Blood, 2008. 112(10): p. 4284-91.

38. Vlassara, H., et al., Advanced glycosylation endproducts on erythrocyte cell surface induce receptor-mediated phagocytosis by macrophages. A model for turnover of aging cells. J Exp Med, 1987. 166(2): p. 539-49.

39. Seaman, C., S. Wyss, and S. Piomelli, The decline in energetic metabolism with aging of the erythrocyte and its relationship to cell death. Am J Hematol, 1980. 8(1): p. 31-42.

40. Lang, F., et al., Eryptosis, a window to systemic disease. Cell Physiol Biochem, 2008. 22(5-6): p. 373-80.

41. Bratosin, D., et al., Programmed cell death in mature erythrocytes: a model for investigating death effector pathways operating in the absence of mitochondria. Cell Death Differ, 2001. 8(12): p. 1143-56.

42. Lang, F., E. Lang, and M. Foller, Physiology and pathophysiology of eryptosis. Transfus Med Hemother, 2012. 39(5): p. 308-14.

43. Maellaro, E., et al., Erythrocyte -3 activation and oxidative imbalance in erythrocytes and in plasma of type 2 diabetic patients. Acta Diabetol, 2013. 50(4): p. 489-95.

44. Calderon-Salinas, J.V., et al., Eryptosis and oxidative damage in type 2 diabetic mellitus patients with chronic kidney disease. Mol Cell Biochem, 2011. 357(1-2): p. 171-9.

45. Basu, S., et al., Eryptosis in hereditary spherocytosis and thalassemia: role of glycoconjugates. Glycoconj J, 2010. 27(7-9): p. 717-22.

46. Mahmud, H., et al., Suicidal erythrocyte death, eryptosis, as a novel mechanism in heart failure-associated anaemia. Cardiovasc Res, 2013. 98(1): p. 37-46.

47. Kay, M.M., Mechanism of removal of senescent cells by human macrophages in situ. Proc Natl Acad Sci U S A, 1975. 72(9): p. 3521- 5.

48. Kiefer, C.R. and L.M. Snyder, Oxidation and erythrocyte senescence. Curr Opin Hematol, 2000. 7(2): p. 113-6.

49

49. Lutz, H.U., Naturally occurring anti-band 3 antibodies in clearance of senescent and oxidatively stressed human red blood cells. Transfus Med Hemother, 2012. 39(5): p. 321-7.

50. Turrini, F., et al., Clustering of integral membrane proteins of the human erythrocyte membrane stimulates autologous IgG binding, complement deposition, and phagocytosis. J Biol Chem, 1991. 266(35): p. 23611-7.

51. Boas, F.E., L. Forman, and E. Beutler, Phosphatidylserine exposure and red cell viability in red cell aging and in hemolytic anemia. Proc Natl Acad Sci U S A, 1998. 95(6): p. 3077-81.

52. Oldenborg, P.A., et al., Role of CD47 as a marker of self on red blood cells. Science, 2000. 288(5473): p. 2051-4.

53. Paulson, J.C., M.S. Macauley, and N. Kawasaki, Siglecs as sensors of self in innate and adaptive immune responses. Ann N Y Acad Sci, 2012. 1253: p. 37-48.

54. Durocher, J.R., R.C. Payne, and M.E. Conrad, Role of sialic acid in erythrocyte survival. Blood, 1975. 45(1): p. 11-20.

55. Kraal, G., Cells in the marginal zone of the spleen. Int Rev Cytol, 1992. 132: p. 31-74.

56. Mebius, R.E. and G. Kraal, Structure and function of the spleen. Nat Rev Immunol, 2005. 5(8): p. 606-16.

57. Blundell, J., Experiments on the Transfusion of Blood by the Syringe. Med Chir Trans, 1818. 9(Pt 1): p. 56-92.

58. Welck, M., P. Borg, and H. Ellis, James Blundell MD Edin FRCP (1790-1877): pioneer of blood transfusion. J Med Biogr, 2010. 18(4): p. 194-7.

59. Landsteiner, K., On agglutination of normal human blood. Transfusion, 1961. 1: p. 5-8.

60. Levine, P. and R.E. Stetson, Landmark article July 8, 1939. An unusual case of intra-group agglutination. By Philip Levine and Rufus E Stetson. JAMA, 1984. 251(10): p. 1316-7.

61. Matteocci, A. and L. Pierelli, Red blood cell alloimmunization in sickle cell disease and in thalassaemia: current status, future

50

perspectives and potential role of molecular typing. Vox Sang, 2014. 106(3): p. 197-208.

62. Mitra, R., N. Mishra, and G.P. Rath, Blood groups systems. Indian J Anaesth, 2014. 58(5): p. 524-8.

63. Goldstein, J., et al., Group B erythrocytes enzymatically converted to group O survive normally in A, B, and O individuals. Science, 1982. 215(4529): p. 168-70.

64. Liu, Q.P., et al., Bacterial glycosidases for the production of universal red blood cells. Nat Biotechnol, 2007. 25(4): p. 454-64.

65. Lewisohn, R., Twenty Years' Experience with the Citrate Method of Blood Transfusion. Ann Surg, 1937. 105(4): p. 602-9.

66. Mollison, P.L., The introduction of citrate as an anticoagulant for transfusion and of glucose as a red cell preservative. Br J Haematol, 2000. 108(1): p. 13-8.

67. Gibson, J.G., 2nd, C.B. Gregory, and L.N. Button, Citrate- phosphate-dextrose solution for preservation of human blood: a further report. Transfusion, 1961. 1: p. 280-7.

68. WHO, Blood safety and availability. World health organization, 2014. Fact sheet No279.

69. Hess, J.R., Red cell storage. J Proteomics, 2010. 73(3): p. 368-73.

70. Greenwalt, T.J. and K.R. Zelenski, Transfusion support for haemoglobinopathies. Clin Haematol, 1984. 13(1): p. 151-65.

71. Goodnough, L.T., J.H. Levy, and M.F. Murphy, Concepts of blood transfusion in adults. Lancet, 2013. 381(9880): p. 1845-54.

72. Tsai, A.G., et al., Perfusion vs. oxygen delivery in transfusion with "fresh" and "old" red blood cells: the experimental evidence. Transfus Apher Sci, 2010. 43(1): p. 69-78.

73. Pace, N., W.V. Consolazio, and E.L. Lozner, The Effect of Transfusions of Red Blood Cells on the Hypoxia Tolerance of Normal Men. Science, 1945. 102(2658): p. 589-91.

51

74. Cacic, D.L., T. Hervig, and J. Seghatchian, Blood doping: the flip side of transfusion and transfusion alternatives. Transfus Apher Sci, 2013. 49(1): p. 90-4.

75. Hess, J.R., An update on solutions for red cell storage. Vox Sang, 2006. 91(1): p. 13-9.

76. Luten, M., et al., Survival of red blood cells after transfusion: a comparison between red cells concentrates of different storage periods. Transfusion, 2008. 48(7): p. 1478-85.

77. Zeiler, T., Müller, J.T., Kretschmer, V., Flow-cytometric determination of survival time and 24-hour recovery of transfused red blood cells. Transfus Med Hemother, 2003. 30(1): p. 14-19.

78. Cancelas, J.A., et al., Additive solution-7 reduces the red blood cell cold storage lesion. Transfusion, 2015. 55(3): p. 491-8.

79. Almizraq, R., et al., Storage of red blood cells affects membrane composition, microvesiculation, and in vitro quality. Transfusion, 2013. 53(10): p. 2258-67.

80. D'Alessandro, A., et al., Time-course investigation of SAGM-stored leukocyte-filtered red bood cell concentrates: from metabolism to proteomics. Haematologica, 2012. 97(1): p. 107-15.

81. Kriebardis, A.G., et al., Storage-dependent remodeling of the red blood cell membrane is associated with increased immunoglobulin G binding, lipid raft rearrangement, and caspase activation. Transfusion, 2007. 47(7): p. 1212-20.

82. Relevy, H., et al., Blood banking-induced alteration of red blood cell flow properties. Transfusion, 2008. 48(1): p. 136-46.

83. Anniss, A.M. and R.L. Sparrow, Expression of CD47 (integrin- associated protein) decreases on red blood cells during storage. Transfus Apher Sci, 2002. 27(3): p. 233-8.

84. Stewart, A., et al., The application of a new quantitative assay for the monitoring of integrin-associated protein CD47 on red blood cells during storage and comparison with the expression of CD47 and phosphatidylserine with flow cytometry. Transfusion, 2005. 45(9): p. 1496-503.

52

85. Heaton, A., T. Keegan, and S. Holme, In vivo regeneration of red cell 2,3-diphosphoglycerate following transfusion of DPG-depleted AS-1, AS-3 and CPDA-1 red cells. Br J Haematol, 1989. 71(1): p. 131- 6.

86. Klein, H.G., D.R. Spahn, and J.L. Carson, Red blood cell transfusion in clinical practice. Lancet, 2007. 370(9585): p. 415-26.

87. Henkelman, S., et al., The effects of cryopreservation on red blood cell rheologic properties. Transfusion, 2010. 50(11): p. 2393-401.

88. Bandarenko, N., et al., Successful in vivo recovery and extended storage of additive solution (AS)-5 red blood cells after deglycerolization and resuspension in AS-3 for 15 days with an automated closed system. Transfusion, 2007. 47(4): p. 680-6.

89. Holovati, J.L., et al., The effects of cryopreservation on red blood cell microvesiculation, phosphatidylserine externalization, and CD47 expression. Transfusion, 2008. 48(8): p. 1658-68.

90. Marik, P.E. and H.L. Corwin, Efficacy of red blood cell transfusion in the critically ill: a systematic review of the literature. Crit Care Med, 2008. 36(9): p. 2667-74.

91. Carson, J.L., R.C. Reynolds, and H.G. Klein, Bad bad blood? Crit Care Med, 2008. 36(9): p. 2707-8.

92. Flegel, W.A., C. Natanson, and H.G. Klein, Does prolonged storage of red blood cells cause harm? Br J Haematol, 2014. 165(1): p. 3-16.

93. Wang, D., et al., Transfusion of older stored blood and risk of death: a meta-analysis. Transfusion, 2012. 52(6): p. 1184-95.

94. Sparrow, R.L., Red Blood Cell Storage Duration and Trauma. Transfus Med Rev, 2015. 29(2): p. 120-126.

95. Aubron, C., et al., Age of red blood cells and transfusion in critically ill patients. Ann Intensive Care, 2013. 3(1): p. 2.

96. Lacroix, J., et al., Age of Transfused Blood in Critically Ill Adults. N Engl J Med, 2015. 372(15): p. 1410-1418.

97. Steiner, M.E., et al., Effects of red-cell storage duration on patients undergoing cardiac surgery. N Engl J Med, 2015. 372(15): p. 1419- 29.

53

98. Fergusson, D.A., et al., Effect of fresh red blood cell transfusions on clinical outcomes in premature, very low-birth-weight infants: the ARIPI randomized trial. JAMA, 2012. 308(14): p. 1443-51.

99. Zimring, J.C., Established and theoretical factors to consider in assessing the red cell storage lesion. Blood, 2015. 125(14): p. 2185- 90.

100. Pereira, A., Will clinical studies elucidate the connection between the length of storage of transfused red blood cells and clinical outcomes? An analysis based on the simulation of randomized controlled trials. Transfusion, 2013. 53(1): p. 34-40.

101. Hod, E.A., et al., Transfusion of red blood cells after prolonged storage produces harmful effects that are mediated by iron and inflammation. Blood, 2010. 115(21): p. 4284-92.

102. Callan, M.B., et al., Transfusion of 28-day-old leucoreduced or non- leucoreduced stored red blood cells induces an inflammatory response in healthy dogs. Vox Sang, 2013. 105(4): p. 319-27.

103. Spitalnik, S.L., Stored red blood cell transfusions: iron, inflammation, immunity, and infection. Transfusion, 2014. 54(10): p. 2365-71.

104. Hod, E.A., et al., Transfusion of human volunteers with older, stored red blood cells produces extravascular hemolysis and circulating non-transferrin-bound iron. Blood, 2011. 18(25): p. 6675-82.

105. Mollison, P.L., Methods of determining the posttransfusion survival of stored red cells. Transfusion, 1984. 24(2): p. 93-6.

106. Mock, D.M., et al., Measurement of posttransfusion red cell survival with the biotin label. Transfus Med Rev, 2014. 28(3): p. 114-25.

107. Kiraly, L.N., et al., Transfusion of aged packed red blood cells results in decreased tissue oxygenation in critically injured trauma patients. J Trauma, 2009. 67(1): p. 29-32.

108. Marik, P.E. and W.J. Sibbald, Effect of stored-blood transfusion on oxygen delivery in patients with sepsis. JAMA, 1993. 269(23): p. 3024-9.

54

109. Roback, J.D., Vascular effects of the red blood cell storage lesion. Hematology Am Soc Hematol Educ Program, 2011. 2011: p. 475-9.

110. Doctor, A. and P. Spinella, Effect of processing and storage on red blood cell function in vivo. Semin Perinatol, 2012. 36(4): p. 248-59.

111. Taft, R.A., M. Davisson, and M.V. Wiles, Know thy mouse. Trends Genet, 2006. 22(12): p. 649-53.

112. Mestas, J. and C.C. Hughes, Of mice and not men: differences between mouse and human immunology. J Immunol, 2004. 172(5): p. 2731-8.

113. Horky, J., J. Vacha, and V. Znojil, Comparison of life span of erythrocytes in some inbred strains of mouse using 14C-labelled glycine. Physiol Bohemoslov, 1978. 27(3): p. 209-17.

114. Knutson, M. and M. Wessling-Resnick, Iron metabolism in the reticuloendothelial system. Crit Rev Biochem Mol Biol, 2003. 38(1): p. 61-88.

115. Waugh, R.E., et al., Rheologic properties of senescent erythrocytes: loss of surface area and volume with red blood cell age. Blood, 1992. 79(5): p. 1351-8.

116. Makley, A.T., et al., Murine blood banking: characterization and comparisons to human blood. Shock, 2010. 34(1): p. 40-5.

117. Vernet, M., Immunochemical assay of transferrin and iron saturation in serum. Clin Chem, 1993. 39(11 Pt 1): p. 2352-3.

118. Parish, C.R., Fluorescent dyes for lymphocyte migration and proliferation studies. Immunol Cell Biol, 1999. 77(6): p. 499-508.

119. Saxena, R.K., et al., A Double in vivo Biotinylation Technique for Objective Assessment of Aging and Clearance of Mouse Erythrocytes in Blood Circulation. Transfus Med Hemother, 2012. 39(5): p. 335-41.

120. Olsson, M. and P.A. Oldenborg, CD47 on experimentally senescent murine RBCs inhibits phagocytosis following Fcgamma receptor- mediated but not scavenger receptor-mediated recognition by macrophages. Blood, 2008. 112(10): p. 4259-67.

55

121. Goodnough, L.T. and M.F. Murphy, Do liberal blood transfusions cause more harm than good? BMJ, 2014. 349: p. g6897.

122. Hill, H.R., et al., The effects of polyvinyl chloride and polyolefin blood bags on red blood cells stored in a new additive solution. Vox Sang, 2001. 81(3): p. 161-6.

123. Dumont, L.J., et al., Exploratory in vitro study of red blood cell storage containers formulated with an alternative plasticizer. Transfusion, 2012. 52(7): p. 1439-45.

124. Haishima, Y., et al., Screening study on hemolysis suppression effect of an alternative plasticizer for the development of a novel blood container made of polyvinyl chloride. J Biomed Mater Res B Appl Biomater, 2014. 102(4): p. 721-8.

125. Wagner, C.T., et al., Biochemical stabilization enhances red blood cell recovery and stability following cryopreservation. Cryobiology, 2002. 45(2): p. 153-66.

126. Chaplin, H., Jr., The proper use of previously frozen red blood cells for transfusion. Blood, 1982. 59(6): p. 1118-20.

127. Kuwahira, I., et al., Splenic contraction-induced reversible increase in hemoglobin concentration in intermittent hypoxia. J Appl Physiol (1985), 1999. 86(1): p. 181-7.

128. Sallet, P., et al., Study of a new indirect method based on absolute norms of variation to detect autologous blood transfusion. Int J Hematol, 2008. 88(4): p. 362-8.

129. Keir, A.K., et al., Plasma cytokines and markers of endothelial activation increase after packed red blood cell transfusion in the preterm infant. Pediatr Res, 2013. 73(1): p. 75-9.

130. Dumont, L.J. and J.P. AuBuchon, Evaluation of proposed FDA criteria for the evaluation of radiolabeled red cell recovery trials. Transfusion, 2008. 48(6): p. 1053-60.

131. Beutler, E., Destruction of erythrocytes, in Williams Hematology, 8:th ed., K. Kaushasky, Lichtman, M.A., Beutler, E., Kipps, T.J., Seligsohn, U., Prchal, J.T., Editor. 2010, Mc Graw Hill: New York. p. 451-2.

56

132. Sowemimo-Coker, S.O., Red blood cell hemolysis during processing. Transfus Med Rev, 2002. 16(1): p. 46-60.

133. Jayle, M.F. and J. Moretti, Haptoglobin: Biochmical, genetic and physiopathologic aspects. Progr Hematol, 1962. 3: p. 342-59.

134. Gilson, C.R., et al., A novel mouse model of red blood cell storage and posttransfusion in vivo survival. Transfusion, 2009. 49(8): p. 1546-53.

135. Burger, P., et al., Potassium leakage primes stored erythrocytes for phosphatidylserine exposure and shedding of pro-coagulant vesicles. Br J Haematol, 2013. 160(3): p. 377-86.

136. van Rooijen, N. and E. Hendrikx, Liposomes for specific depletion of macrophages from organs and tissues. Methods Mol Biol, 2010. 605: p. 189-203.

137. Wojczyk, B.S., et al., Macrophages clear refrigerator storage- damaged red blood cells and subsequently secrete cytokines in vivo, but not in vitro, in a murine model. Transfusion, 2014. 54(12): p. 3186-97.

138. Belz, G.T., W.R. Heath, and F.R. Carbone, The role of dendritic cell subsets in selection between tolerance and immunity. Immunol Cell Biol, 2002. 80(5): p. 463-8.

139. Hendrickson, J.E., et al., Recipient inflammation affects the frequency and magnitude of immunization to transfused red blood cells. Transfusion, 2006. 46(9): p. 1526-36.

140. Hendrickson, J.E., et al., Discrete Toll-like receptor agonists have differential effects on alloimmunization to transfused red blood cells. Transfusion, 2008. 48(9): p. 1869-77.

141. Qiu, C.H., et al., Novel subset of CD8{alpha}+ dendritic cells localized in the marginal zone is responsible for tolerance to cell- associated antigens. J Immunol, 2009. 182(7): p. 4127-36.

142. McGaha, T.L., et al., Marginal zone macrophages suppress innate and adaptive immunity to apoptotic cells in the spleen. Blood, 2011. 117(20): p. 5403-12.

143. Olsson, M., A. Nilsson, and P.A. Oldenborg, Dose-dependent inhibitory effect of CD47 in macrophage uptake of IgG-opsonized

57

murine erythrocytes. Biochem Biophys Res Commun, 2007. 352(1): p. 193-7.

144. Sosale, N.G., et al., Cell rigidity and shape override CD47's "self"- signaling in phagocytosis by hyperactivating myosin-II. Blood, 2015. 125(3): p. 542-52.

145. Burger, P., et al., CD47 functions as a molecular switch for erythrocyte phagocytosis. Blood, 2012. 119(23): p. 5512-21.

146. Platt, N. and S. Gordon, Is the class A macrophage scavenger receptor (SR-A) multifunctional? - The mouse's tale. J Clin Invest, 2001. 108(5): p. 649-54.

147. Gillooly, D.J., A. Simonsen, and H. Stenmark, Phosphoinositides and phagocytosis. J Cell Biol, 2001. 155(1): p. 15-7.

148. Kelley, J.L., et al., Scavenger receptor-A (CD204): a two-edged sword in health and disease. Crit Rev Immunol, 2014. 34(3): p. 241- 61.

149. Yu, H., et al., Scavenger receptor A (SR-A) is required for LPS- induced TLR4 mediated NF-kappaB activation in macrophages. Biochim Biophys Acta, 2012. 1823(7): p. 1192-8.

150. Soltis, R.D., et al., The effect of heat inactivation of serum on aggregation of immunoglobulins. Immunology, 1979. 36(1): p. 37- 45.

151. Erwig, L.P. and P.M. Henson, Clearance of apoptotic cells by phagocytes. Cell Death Differ, 2008. 15(2): p. 243-50.

58

1