UMEÅ UNIVERSITY MEDICAL DISSERTATIONS New series No 842 ISBN 91-7305-447-X ISSN – 0346-6612 From the Department of Clinical Microbiology, Clinical Bacteriology and Infectious Diseases, Umeå University, Umeå, Sweden

Expansion of circulatory Vγ9Vδ2 T cells in tularemia and Pontiac fever, two intracellular bacterial diseases with widely different clinical expression

Michal Kroca

Umeå 2003

i

Copyright © 2003 Michal Kroca

ISBN 91-7305-447-X

Printed by:

Solfjädern Offset AB

Umeå, Sweden 2003

ii

CONTENTS

CONTENTS ...... iii ABSTRACT...... iv SAMMANFATTNING PÅ SVENSKA...... v PAPERS ...... vii ABBREVIATIONS...... viii INTRODUCTION...... 1 The T-cell receptor ...... 1 αβ T-cell recognition of peptides ...... 4 CD1 molecules and CD1-dependent T cells in the microbial host defence ...... 5 γδT cells and their ligands...... 8 Vγ9Vδ2 T cells...... 10 Phosphoantigens...... 11 The functions of γδ T cells ...... 12 γδ T cells in human diseases ...... 13 Tularemia ...... 14 Two entities of legionellosis – Legionnaires’ disease and Pontiac fever ...... 16 AIMS...... 18 MATERIAL AND METHODS ...... 19 RESULTS...... 26 F. tularensis-induced expansion of γδ T cells and phosphoantigen nature of stimulating antigen (I) ...... 26 Stress proteins of F. tularensis and human γδ T cells (II)...... 28 Time course of expansion and cytokine response of Vγ9Vδ2 T cells in tularemia (III)...... 30 Expansion of Vγ9Vδ2 T cells in Pontiac fever, an intracellular bacterial disease with a clinical expression different from that of tularemia (IV) ...... 31 Response of Vγ9Vδ2 T cells from individuals undergoing legionellosis or tularemia to phosphoantigen from francisella or legionella (V)...... 34 DISCUSSION ...... 36 Vγ9Vδ2 T-cell expansion as a feature of intracellular bacterial diseases ...... 36 The heat shock protein homologs Cpn60 or DnaK of F. tularensis are not recognized by Vγ9Vδ2 T cells...... 42 Implications from data on time course of expansion of Vγ9Vδ2 T cells in tularemia and legionellosis...... 43 No increase of Vγ9Vδ2 T-cell levels in response to LVS vaccination ...... 48 Role of Vγ9Vδ2 T cells in intracellular bacterial diseases...... 48 Implications on the pathophysiology of Pontiac fever ...... 51 Implications of a cross reactivity in vitro by Vγ9Vδ2 T cells from subjects primed by tularemia or Pontiac fever, respectively ...... 52 CONCLUSIONS...... 54 ACKNOWLEDGEMENTS ...... 55 REFERENCES...... 58

iii

ABSTRACT

Although well established that human Vγ9Vδ2 T cells may expand in circulation during intracellular bacterial infections, most underlying studies included only a few cases and only some diseases had been studied so far. In tularemia, a severe invasive disease, only one patient had been described. Legionellosis, including the mild flue-like Pontiac disease caused by Legionella micdadei, had not been studied at all. The aim of the present thesis was to study the circulatory Vγ9Vδ2-T cell response in these two intracellular bacterial diseases. The number of cases included was large enough to draw general conclusions. At various intervals, Vγ9Vδ2-T-cell counts and the capability of the cells to produce proinflammatory cytokines were assayed. Finally, the nature of the stimulating antigens was determined.

In the acute phase of tularemia, we showed a marked increase of circulatory Vγ9Vδ2 T cells. When 181 samples from 108 patients with ulceroglandular tularemia were assayed, the percentage of Vγ9Vδ2 T cells was found to increase from ~5 to > 20% after the first week of disease. During the ensuing 24 months, levels were normalized. Vaccination with the live attenuated vaccine strain Francisella tularensis LVS, on the other hand, did not cause an increase in numbers of Vγ9Vδ2 T cells.

Within an outbreak of Pontiac fever, 14 cases were well defined with regard to incubation time and onset of disease. In samples obtained 4 to 6 days after onset of disease, the mean percentage of Vγ9Vδ2 T cells was ~ 1%, i.e., 20% of normal values. Thereafter, a pronounced increase occurred and at 2 to 7 weeks after onset of disease, values were ~ 15%. Later, values slowly decreased. In both tularemia and Pontiac fever, the capacity of Vγ9Vδ2 T cells to produce TNF-α in response to phorbol myristate acetate in vitro was transiently decreased, in tularemia up to 6 weeks after onset of disease and in Pontiac fever in samples obtained 5-7 weeks after onset of disease.

Nonpeptidic pyrophosphorylated molecules, referred to as phosphoantigens, are powerful stimuli for Vγ9Vδ2 T cells. Various strains of F. tularensis, including LVS, and a strain of L. micdadei were shown to produce Vγ9Vδ2 T-cell stimulating phosphoantigen. Notably, stimulation with an extract from each agent caused a similar degree of expansion of cells from subjects infected with the homologous and heterologous agent and also of cells from healthy subjects. Thus no immunospecific memory was detected in the Vγ9Vδ2-T cell response.

Since it had been suggested that homologs of the conserved heat shock protein, chaperon-60, may be recognized by human Vγ9Vδ2 T cells, we determined the subpopulation of T cells responding to this protein as well as to DnaK, another heat shock protein. Under in vitro conditions allowing a vigorous expansion of Vγ9Vδ2 T in response to a phosphoantigen, no expansion of γδ T cells in response to Cpn60 or DnaK of F. tularensis occurred. αβ T cells of tularemia-primed subjects, on the other hand, responded vigorously to the heat-shock proteins.

In conclusion, two intracellular bacterial diseases with widely varying clinical expression were both associated with expansion of circulating Vγ9Vδ2 T cells. The expansion was prominent, long-lasting, and consistent within large numbers of individuals tested. In Pontiac fever, the expansion of Vγ9Vδ2 T cells was preceded by a depletion of the cells in circulation, implicating a possible extravasal migration into an infected site before the occurrence of rapid expansion and reentrance to blood. Both in tularemia and Pontiac fever, a modulation of the cytokine expression of Vγ9Vδ2 T cells was demonstrated in vitro, suggesting the presence of modulation of the inflammatory response. In extracts from in vitro culture of F. tularensis and L. micdadei, Vγ9Vδ2 T-cell stimulating phosphoantigens were identified and according to cross stimulation experiments, they induced expansion in vitro of Vγ9Vδ2 T cells without regard to immunospecific memory.

iv

SAMMANFATTNING PÅ SVENSKA

Vid intracellulära bakteriella sjukdomar har en ökning av antalet Vγ9Vδ2 T-celler påvisats i perifert blod. När detta avhandlingsarbete påbörjades, hade endast vissa sjukdomstillstånd studerats och studierna omfattade i regel endast ett litet antal patienter. Vid tularemi (harpest) fanns en patient och vid legionellos ingen enda patient beskriven med avseende på Vγ9Vδ2 T-celler. Syftet med avhandlingsarbetet var att studera Vγ9Vδ2 T-cellsvaret vid denna typ av infektion. Studierna kom att inkludera ett stort antal fall av tularemi, som orsakas av Francisella tularensis, liksom ett utbrott av Pontiac-feber, orsakad av Legionella micdadei. Tidsförloppet i ökningen av Vγ9Vδ2 T-celler studerades och cellernas förmåga att producera cytokiner som svar på stimuli mättes.

I den akuta fasen av tularemi fann vi en markant ökning av cirkulatoriska Vγ9Vδ2 T-celler. När 181 prover från 108 patienter med ulceroglandulär tularemi undersöktes, befanns Vγ9Vδ2 T-cellerna öka från ~ 5 till > 20% efter den första sjukdomsveckan. Under de följande 24 månaderna normaliserades nivåerna successivt. Vaccination med den levande attenuerade vaccinstammen F. tularensis LVS följdes inte av någon ökning av Vγ9Vδ2 T-cellerna.

I samband med ett utbrott av Pontiac-feber undersökte vi 14 fall, vilka alla var väldefinierade med avseende på inkubationstid och sjukdomsdebut. I prover erhållna 4-6 dagar efter sjukdomsdebut uppmättes i blodet i medeltal ~ 1% Vγ9Vδ2 T-celler bland totala antalet T-celler, det vill säga en femtedel av den normala nivån. Därefter skedde en uttalad stegring, efter 2-7 veckor till ~ 15%. Därpå sjönk nivåerna långsamt. Vid både tularemi och Pontiac-feber var förmågan hos Vγ9Vδ2 T-cellerna att in vitro producera TNF-α som svar på forbolmyristatacetat övergående sänkt, vid tularemi upp till 6 veckor efter insjuknandet och vid Pontiac-feber i prover tagna 5-7 veckor efter sjukdomsdebut.

Små pyrofosforylerade molekyler av icke-peptidnatur, så kallade fosfoantigener, är kraftfulla stimuli för Vγ9Vδ2 T-celler. Olika stammar av F. tularensis, inklusive LVS, och en stam av L. micdadei befanns producera Vγ9Vδ2 T-cellstimulerande fosfoantigen. I experiment med blodprover från patienter med tularemi eller Pontiac-feber fann vi att Vγ9Vδ2 T-cellerna svarade lika bra oavsett om de stimulerades med fosfoantigen från den homologa eller heterologa bakterien. Således påvisades inget immunspecifikt minne bland Vγ9Vδ2 T-cellerna.

Då det av andra forskare hade föreslagits, att homologer av det konserverade heat-shock-proteinet chaperon-60 (cpn-60) känns igen av humana Vγ9Vδ2 T-celler, ville vi bestämma de subpopulationer av T-celler som svarar på detta protein liksom på DnaK isolerade från F. tularensis. Under sådana in vitro-förhållanden, som tillät en kraftig expansion av Vγ9Vδ2 T-celler som svar på fosfoantigen, kunde ingen expansion av γδ T-celler påvisas som svar på stimulering med cpn60 eller DnaK från F. tularensis. αβ T-celler, däremot, svarade starkt på dessa heat-shock-proteiner.

Sammanfattningsvis befanns två intracellulära bakteriella sjukdomar med olika kliniskt uttryck båda vara förbundna med en expansion av Vγ9Vδ2 T-celler i cirkulationen. Förhöjningen var uttalad, långvarig och regelbundet förekommande bland ett stort antal patienter. Vid Pontiac-feber föregicks expansionen av en minskning av antalet Vγ9Vδ2 T-celler i blodet, vilket kan tyda på att cellerna migrerar ut från blodbanan till ett infekterat område, innan de återkommer och expanderar i blodbanan. Vid både tularemi och Pontiac-feber påvisades en modulering av cytokinuttrycket i Vγ9Vδ2 T-celler in vitro, vilket kan tyda på en nedreglering av det inflammatoriska svaret. I extrakt från in vitro kulturer av F. tularensis och L. micdadei identifierades Vγ9Vδ2 T-cellstimulerande fosfoantigener. I korsvisa stimuleringsförsök in vitro svarade Vγ9Vδ2 T-cellerna utan tecken på immunspecificitet.

v

vi

PAPERS

The thesis is based on the following papers, which will be referred to in the text by their Roman numerals.

I. Poquet Y, Kroca M, Halary F, Stenmark S, Peyrat M-A, Bonneville M, Fournie JJ, Sjöstedt A. Expansion of Vγ9Vδ2 T cells is triggered by Francisella tularensis-derived phosphoantigens in tularemia but not after tularemia vaccination. Infect Immun 1998;66:2107-2114. (Poquet and Kroca contributed equally)

II. Ericsson M, Kroca M, Johansson T, Sjöstedt A, Tärnvik. A. Long-lasting recall response of CD4+ and CD8+ αβ T cells, but not γδ T cells, to heat shock proteins of Francisella tularensis. Scand J Infect Dis 2001;33:145- 152.

III. Kroca M, Tärnvik A, Sjöstedt A. The proportion of circulating γδ T cells increases after the first week of onset of tularaemia and remains elevated for more than a year. Clin Exp Immunol 2000;120:280-284.

IV. Kroca M, Johansson A, Sjöstedt A, Tärnvik A. Vγ9Vδ2 T cells in human legionellosis. Clin Diagn Laboratory Immunol 2001;8:949-954.

V. Kroca M, Johansson A, Sjöstedt A, Tärnvik A. Vγ9Vδ2 T cells from individuals undergoing Pontiac fever or tularemia respond at a similar extent to phosphoantigen from Francisella tularensis or Legionella micdadei. Manuscript.

vii

ABBREVIATIONS

AIDS acquired immunodeficiency syndrome BCG bacillus Calmette-Guerin BCYE buffered charcoal yeast extract CCR CC chemokine receptor CD cluster of differentiation CMV cytomegalovirus Cpn chaperon dTMP deoxythymidin monophosphate ELISA enzyme-linked immunosorbent assay FACS fluorescence-activated cell sorting Fas fibroblast-associated FE Francisella extract FITC fluorescein isothiocyanate FSC Francisella strain collection HLA human histocompatibility leucocyte antigen HSP heat-shock protein IFN interferon IL interleukin IPP isopentenyl pyrophosphate KIR killer inhibitory receptor LE Legionella extract LVS live vaccine strain MHC major histocompatibility complex MICA/B MHC class I restricted chains A/B NK natural killer NKG2 natural killer receptor family member G2 PBL peripheral blood lymphocytes PBMC peripheral blood mononuclear cells PE phycoerythrin PerCP peridin-chlorophyll protein PMA phorbol myristate acetate PPD purified protein derivative RANTES regulated on activation, normal T cell expressed and secreted RT room temperature TB tuberculosis TCR T-cell receptor Th T-helper TNF tumor necrosis factor TUBag tuberculosis antigen UMP uridin monophosphate

viii Introduction

INTRODUCTION

The T-cell receptor In intracellular infections, T cells are orchestrating the immune response and establishing protective immunity. They specifically recognize foreign antigen and respond by proliferation and differentiation into effector cells as well as by mediating information to other cells of the immune system.

The T-cell receptor (TCR) for recognition of antigen is a clonally distributed heterodimer protein, located on the cell surface and noncovalently linked to the CD3 signaling complex. Based on discrete receptor chains, T cells can be divided in two major subsets, αβ and γδT cells. Both types of TCRs consist of domain structures similar to those of the immunoglobulins, and they are classified as members of the immunoglobulin superfamily. Each of the two chains of the TCR has two domains – one variable (V) and one constant (C) – with three complementarity-determining regions in the V domain. The chains of the TCR are connected by disulfide-linking and the TCR is anchored in the plasma membrane by a transmembrane region (Fig.1).

The functional TCR and TCR genes are assembled by genetic recombination from separate V, D, J, and C region segments (in the case of TCRα and TCRγ genes V, J, and C regions only). The δ-chain gene segments are located on the chromosome between the Vα and Jα segments. This location of the δ-chain gene family is significant because productive rearrangement of the α-chain gene segments deletes Cδ, so that the γδ TCR cannot be co-expressed with the αβ receptor by a given T cell. The process of rearrangement is similar to that observed in the immunoglobulin loci during B-cell development. V, D, and J (or V and J) gene segments rearrange to form a unique TCR gene sequence encoding a functional variable domain, which is joined to the C region sequence by RNA splicing following transcription.

1 Introduction

Figure 1. Schematic organization of γδ T-cell receptor

TCR Antigen-binding site

CHO Vγ Vδ CHO Variable region (V) CD3 CD3

CHO Cγ Cδ CHO Constant region (C) εεδ γ Disulfide bridge

Cell membrane

Transmembrane region

Cytoplasmic tail

ζ ζ

CD3

The great diversity among αβ and γδ TCRs is created by mutually similar genetic mechanisms. For the αβ TCR, a large number of alternate α or β V gene segments and V, D, and J segments are utilized. To further increase the diversity of the TCR, additional nucleotides are inserted in the junctions between the V, D, and J gene segments of the TCR β-chain coding sequence and between the V

2 Introduction and J gene segments of the α-gene sequence. The rearrangement is primarily a random process, although restricted to allow recognition only of antigen presented in association with self MHC molecules and to avoid reactivity to self antigen. These two prerequisites are fulfilled by positive selection of cells binding to self MHC I or MHC II molecules and negative selection, i.e., deletion of cells reactive with self peptides bound to the MHC molecules. Theoretically, the diversity of the αβ TCR is estimated to be in the range of 1015 [1].

As a result of all these events, the αβ TCR is capable to recognize both a very large number of different processed antigens and a relatively small number of self-MHC molecules. It has been suggested that the limited number of germ- line V gene segments carried by T cells may generate the diversity needed for MHC recognition, whereas the enormous diversity generated at the junctional regions facilitates the recognition of antigen [2, 3].

For the γδ TCR, the number of Vγ and Vδ gene segments is limited. Human γδ T cells express one of six Vγ gene segments (Vγ1, Vγ3, Vγ4, Vγ5, Vγ8, and Vγ9) and three major functional Vδ genes (Vδ1, Vδ2, and Vδ3). Despite this limited number of gene segments, the theoretical diversity of the γδ TCR is estimated to be several orders of magnitude higher than that of the αβ TCR [4]. This is due to a more diverse junctional region complexity, including use of multiple gene segments and addition and removal of germ-line encoded nucleotides. It is unknown why the mammalian host may need such an enormous diversity of the γδ T cells. It may be relevant in the context, that TCR genes do not undergo somatic mutation and do not increase the avidity in their ligand binding [5].

3 Introduction

αβ T-cell recognition of peptides αβ T cells consist of two subsets, easily distinguishable by their alternate expression of the CD4 or the CD8 molecule. The differential expression is determined during T-cell development in the thymus. At a premature stage, thymocytes express both CD4 and CD8 molecules. While undergoing positive selection, CD4+CD8+ thymocytes mature in so called single positive T cells, CD4+CD8- T cells being formed by interaction with MHC class II molecules and CD4-CD8+ T cells by interaction with MHC I.

T cells contribute to host defence in situations where antibodies are of little help. CD4 T cells are crucial to the host defence against intracellular bacteria and protozoans, agents that primarily infect mononuclear phagocytes, whereas CD8 T cells are instrumental in the host defence against viral agents. The T cells seem to be especially designed to combat these infections, because they detect antigen present inside cells. CD4 T cells recognize antigens of bacteria or protozoans endocytosed by mononuclear phagocytes and a few other types of cells. In the phagocytic endosomes, microbial peptides are released and bound to MHC class II molecules. The peptide-MHC complex is subsequently presented on the cell surface to specifically sensitized CD4 T cells. The CD4 T cells proliferate and differentiate into cytokine producing effector cells. The exclusive presence of class II molecules on mononuclear phagocytes and a few other types of cells implies a restriction of inflammation associated with cytokine production to sites of infection of these cells.

CD8 T cells contribute to the defence against viral infections. They recognize peptide antigens processed in the cytoplasm of infected cells. Peptide- MHC class I molecule complexes are formed in the endoplasmatic reticulum and presented to CD8 T cells on the cell surface. Most nucleated mammalian cells express class I molecules, thus being able to participate in the CD8 T-cell response. A CD8 T-cell response is evoked also in intracellular bacterial

4 Introduction infections, either due to an ability of the pathogen to localize in the host cell cytoplasm or to an association of bacterial antigen with MHC class I by some less well defined route. The CD8 T cells serve mainly by affording cytotoxicity against infected cells. Moreover, they produce cytokines. For their proliferation, CD8 T cells depend on CD4 T cell help [6], a mechanism whereby the CD8 T- cell response may be restricted to avoid unnecessary cytotoxicity and cytokine- mediated inflammation.

Work performed on murine experimental models during the 1980s, and confirmed during the 1990s on human cells, clarified that CD4 T cells differentiate upon stimulation by peptide-MHC II complexes into one of two alternative phenotypes, Th1 or T1 cells and Th2 or T2 cells [7]. Th1 cells produce IFN-γ, a cytokine of crucial importance to activate the antimicrobial effects of mononuclear phagocytes. Th2 cells produce IL-4, IL-5, and IL-10, and afford help in B-cell differentiation. The differentiation pathways of CD4 T cells are not yet fully elucidated, and are presently extended to comprise phenotypes other than Th1 and Th2 [8, 9].

CD1 molecules and CD1-dependent T cells in the microbial host defence Until the end of the 1980s, the concept describing T-cell recognition of antigen remained rather simple. In contrast to B cells, which can produce antibodies against a wide variety of molecular species, the recognition of antigen by T cells was believed to be restricted to proteins. During the 1990s, however, T-cell subsets predestined to recognise molecular species other than peptides and restricted by elements other than MHC molecules were successively disclosed. The existence of these recognition events is irrefutable, but the functions of the T cells recognizing non-peptidic ligands in host-parasite interaction are still virtually unknown. Remarkably, some of the recognition events are present only

5 Introduction in humans but not in the mouse [10], either due to a lack of restriction element or a lack of subset of T cells recognizing the pattern of molecules.

The first evidence implicating CD1 proteins as antigen-presenting molecules was obtained, when a human CD4-CD8- αβ+ T-cell line was found to present mycobacterial antigen in a CD1-restricted manner [11]. Similar to class I MHC molecules, CD1 molecules are heterodimers containing a β2-microglobulin chain. They are, however, otherwise unrelated to MHC proteins and nonpolymorphic. They seem to have evolved to present hydrophobic antigens to T cells. Analysis of the crystal structure of a CD1 molecule showed a deep hydrophobic ligand-binding groove, compatible with binding of fatty acid tails of lipid molecules [12].

CD1 comprises two groups of genes. Group 1, which comprises CD1a, CD1b, CD1c, and CD1e, is present in humans but not in mice. Group 2, which comprises CD1d, seems to be generally conserved in mammalian species [13]. Work aiming to define microbial components presented by CD1 has focused on mycobacteria. Discrete mycobacterial cell wall lipids and glycolipids are known to be presented by members of group 1 CD1 molecules [14]. Of special interest in relation to host defence is the demonstration of a CD1c-restricted in vitro proliferative response to mannosyl phosphodolichols. These are semisynthetic analogs of mycobacterial hexose-1-phosphoisoprenoids found in blood lymphocytes of patients recently infected with Mycobacterium tuberculosis [15]. Initially, CD1-restricted recognition of lipid/glycolipid antigens was demonstrated in CD4-CD8- αβ T cells, a phenotype which comprises a minute part of T cells in blood [16]. By ensuing investigations, mainly on T cell lines, CD1-restricted T cells have been found also among CD4+ and CD8+ αβ T cells [17, 18]. Moreover, a subset of human γδ T cells expressing the Vδ chain are CD1c restricted [19].

6 Introduction

In contrast to group 1 CD1 molecules, presentation of microbial ligands by the group 2, CD1d molecules has not yet been demonstrated. One special subset of T cells associated with CD1d is the natural killer T cell. In humans, this subset includes cells with a TCR consisting of an invariant, germline-encoded, TCR α-chain, Vα24-JαQ, associated with a Vβ11 chain. These cells, as well as an analogous subset in mice, recognize an unusual α-galactosylated sphingolipid (α-GalCer), extracted from a marine sponge [20]. Although this molecule bears resemblance to mammalian lipids [20], a corresponding mammalian ligand has not been identified. CD1d molecules might hypothetically be involved in recognition of molecules closely related to self and have a role in the anti-tumor response or in autoimmune disease, whereas members of the group 1 CD1 family might have a role in the host defence to intracellular infectious agents [14].

At present, there is no direct evidence available to show a role in host protection to infection for CD1-restriced T cells. However, indirect evidence from in vitro studies indicates that CD1-presented antigens trigger CD8-CD4- and CD8+CD4- T-cells to proliferate, secrete IFN-γ, and exert cytotoxicity on infected macrophages [16, 17]. Cytotoxicity is effectuated by perforin and granulysin secretion by CD8+CD4- cells [21, 22] and induction of apoptosis in infected macrophages via the Fas pathway by CD8-CD4- cells [23]. There are also reports indicating that CD1-antigen complexes may be recognized by γδ T cells, and that both perforin and Fas-mediated cytotoxicity is induced [19].

A role of CD1 molecules in the host response to intracellular agents is suggested by analysis of their distribution in intracellular compartments [24]. The data suggests that different CD1 molecules might sample lipid/glycolipid antigens in different compartments. This would enable successive sampling of antigens released from an intracellular pathogen, such as mycobacteria, along the intracellular pathway from early endosomes to lysosomes.

7 Introduction

γδT cells and their ligands

The role of peptide-MHC-recognizing αβ CD4 and CD8 T cells in the host defence to various types of infectious agents is well characterized. By their receptor diversity, readiness to expand in response to specific antigens, and capability to form a useful memory, these cells constitute a major part of the adaptive immune system. Although their alternative differentiation pathways remain to be completely delineated, peptide-MHC recognizing αβ Τ cells are conceptually simple to understand. In contrast, the general biologic reasons for the enormous diversity and the role of γδ T cells are not well understood.

γδ T cells are not easily classified as belonging to the innate or the adaptive immune system. Despite an enormous diversity in the complementarity- determining regions of their TCR structure, conforming with cells involved in adaptive immunity, more than one per cent of all γδ T cells of a nonprimed subject may recognize one and the same antigen molecule. Similar to cells of the innate immune system, γδ T cells seem to recognize molecular patterns rather than discrete antigenic epitopes. Nonetheless, γδ T cells expand in response to antigen stimulation and form a memory-like phenotype, as do cells involved in adaptive immunity. Attempts to name this functionally hybrid lineage of cells have resulted in the hybrid term “natural memory cells”. Such a name would probably not have been coined if these cells had been detected before the terms “innate” and “adaptive” immunity were firmly established. The question is not only semantic, because extrapolations may be erroneously performed from knowledge on innate or adaptive immunity when trying to delineate mechanisms of antigen recognition and role in the immune response of the γδ T-cell lineage.

To understand the nature of various subsets of γδ T cells, it is important to distinguish among cells with regard to Vγ and Vδ chain usage. This might be as fundamental as to distinguish CD4 from CD8 or Th1 from Th2 cells within the

8 Introduction

αβ T-cell lineage. In mice, γδ T cells expressing a given Vγ segment are exported from the thymus at a specific time period of development and migrate to a given epithelial-rich organ. For example, Vγ5Vδ1 T cells migrate to the skin and Vγ6Vδ1 Τ cells to the uterine epithelia (for references, see [4]). Presentation of ligands to γδ T cells by association to restricting elements does not seem to be required.

Here I will divide human γδ T cells in two major parts, due to VγVδ chain usage. One is the Vγ9Vδ2 subset, the topic of the present thesis. It predominates among γδ T cells in peripheral blood and lymph nodes. Before going into details, the second major part will first be described. These cells express the Vδ1 chain combined with different Vγ chains. They occur in epithelial-rich tissues, such as the skin and intestine, and may constitute up to 50% of all T cells. Their pattern of recognition differs decisively from that of Vγ9Vδ2 and is yet less well defined.

Two significant pieces of data on ligands for Vδ1 cells will be described.

One demonstrated the recognition by human Vγ1Vδ1 γδ T cells of the β2- independent MHC class Ib molecules MICA and MICB, cell surface proteins induced by stress in gut epithelial cells and upregulated in carcinoma cells [25, 26]. The observations are in line with a sentinel function of γδ T cells, as part of a first line defence. An important detail was a failure by the authors [25] to demonstrate antigen bound to MICA/MICB, indicating that these stress-induced molecules would by themseleves be targets of recognition and not part of a complex such as classical MHC-peptide or CD1-lipid. Questions remain to be resolved, however, because it was later claimed by the same group that recognition of MICA requires interaction with NKG2D, the NK cell activation receptor, which is present on γδ T cells [27].

9 Introduction

A second piece of data on human γδ T cells expressing the Vδ1 chains derives from studies in viral disease. In HIV infection, an increase of γδ T cells expressing the Vδ1 TCR chain has been reported [28]. When renal allograft recipients acquired CMV infection, they developed expansion in circulation of γδ T cells. Within 3 months of transplantation, a peak level of ~15% of T cells (i.e., CD3+ cells) were γδ T cells, expressing predominantly Vδ1 and to some extent Vδ3 but not at all the Vδ2 TCR chain. Vδ1 and Vδ3 T cells from CMV- infected kidney recipients showed a proliferative response to lysates from CMV- infected fibroblast or free CMV [29]. The results afford strong evidence in favor of a role for γδ T cells other than Vδ2-expressing γδ T cells in this kind of viral disease. This is supported by data from a murine model on herpes simplex, another DNA virus, showing that γδ T cells mediate control of disease [30].

Vγ9Vδ2 T cells

When turning to the Vγ9Vδ2 subset, the contrast between γδ T cells using the Vδ2 gene segment and those using the Vδ1 and Vδ3 gene segments becomes obvious. This relates to their localization and pattern of antigen recognition. Differences in effector functions are less well studied so far.

Vγ9Vδ2 T cells predominate among γδ T cells in blood and lymph nodes. At birth, only 1-2% of CD3+ cells are γδ T cells [3]. At this stage, various subsets of γδ T cells are well distributed. At the age of 3-10, the Vγ9Vδ2 (Vγ2Vδ2 in alternative nomenclature) T cell subset expands. In adolescents and adults, ~5% of CD3+ cells are γδ T cells. In these age groups, Vγ9Vδ2 cells predominate among γδ T cells in peripheral blood and can account for up to 95% of the γδ T cells [31, 32]. Environmental factors like a childhood infection or symptomless exposure to some microorganisms may further increase the proportion of Vγ9Vδ2 T cells in periphery. In blood, most of the remaining γδ T cells express

10 Introduction

Vδ1 in association with different Vγ elements. These cells usually account for 10-30% of peripheral blood γδ T cells.

Phosphoantigens

The first evidence indicating that Vγ9Vδ2 T cells might recognize nonpeptide antigen derives from work on mycobacteria. Proteinase-resistant low-molecular mass fractions of cell lysates induced proliferative responses in human peripheral blood γδ T cells [33]. Later on, some of these molecules were partially characterized and found to be specific for the Vγ9Vδ2 T cells [34]. Also certain synthetic compounds were found to have a similar activity [35].

Vγ9Vδ2 T-cell stimulating phosphoantigens are widely distributed in procaryotic and eucaryotic cells. One essential prerequisite for activity by these ligands is the presence of a terminal phosphate group [36]. Other general features are their low molecular weight (150–300 Da), the lack of MHC restriction, involvement of the TCR in the process of activation, and a broad cross reactivity. With regard to their biological activity, phosphoantigens of bacterial origin were found to be several logs more potent stimulators of human Vγ9Vδ2 T cells than their synthetic or eucaryotic analogs [37].

In their recognition of phosphoantigens, Vγ9Vδ2 T cells have features associated with cells of the innate as well as the adaptive immune systems. Although the heterogeneity of their TCR structure is even more extensive than that of αβ T cells, they are nonetheless broadly reactive with phosphoantigens similar in structure. Their broad recognition of phosphoantigens conforms with cells of the innate system. However, they also clonally expand upon antigen recognition, a trait defining adaptive immunity.

11 Introduction

Proportion of γδ T cells in Major γδ T-cell subset in Organ mice human mice human Peripheral blood < 5 % ~ 5 % Vγ4/1-Vδ1 Vγ9-Vδ2 Lymph nodes 5 – 10 % ~ 2 % Vγ4/1-Vδ1 Epithelia majority ~ 20 % Vγ4/6/7-Vδ1 Vγ1-Vδ1/3 Thymus < 1 % Vγ2 Liver < 5 % few Skin > 20 % rare Vγ3/5-Vδ1 Spleen < 5 % Vγ4/1-Vδ1

The functions of γδ T cells

From studies on murine models, a contribution by γδ T cells to the innate immunity against intracellular bacterial pathogens like M. tuberculosis [38-41] or Listeria monocytogenes [42, 43] was demonstrated. Early activation of γδ T cells and direct cytokine production [44-46] or regulatory effects on cytokine production by other cell types [47] were also observed in experimentally infected animals.

A possible role of Vγ9Vδ2 T cells to host control of infection cannot be studied in murine models, because of an apparent lack of any analog of the subset in mice. In a macaque model, Vγ9Vδ2 T cells expanded during BCG infection and a recall expansion occurred in response to reinfection [48]. In macaques, Vγ9Vδ2 T-cell expansion coincided with a clearance of BCG bacteriemia, both during primary infection and reinfection.

Effector functions of γδ T cells include production of cytokines, and TNF-α and IFN-γ in particular [49-51]. Specific cytotoxicity towards infected cells or tumour cells is yet another feature of γδ T cells. Selective lysis of autologous tumour cells is generally associated with the Vδ1 phenotype of γδ T cells [52, 53], but also Vγ9Vδ2 expressing cells were found to have such an ability [49, 54]. The lysis of bacteria or virus-infected cells by γδ T cells is, on contrary,

12 Introduction restricted to the Vγ9Vδ2 subpopulation [55-58]. Many reports from work on animal models document the decisive role of γδ T cells in maintenance of homeostasis in the immune system. Their regulation of TNF-α production by macrophages [47], mitigation of excessive tissue damage and promotion of granuloma formation after intracellular infections [43], [39], and maintenance of epithelial integrity [59] are such functions.

Murine γδ T cells share features with NK cells [60, 61], including a high level of spontaneous or induced cytotoxicity and a bias toward IFN-γ secretion. Since such effector functions might be harmful to the host, a down regulation of γδ T cells may be required. Regulated expression in the γδ T cells of killer-cell inhibitory receptors (KIRs) specific for HLA class I molecules seems to be such an event [62].

γδ T cells in human diseases

In humans, expansion of γδ T cells is observed in disease caused by various intracellular pathogens, including salmonellosis [63], toxoplasmosis [56], schistosomiasis [64], malaria [65], brucellosis [66] and listeriosis [67] and in some viral infections including HIV infection [68-71].

In human tuberculosis, data on levels of circulating Vγ9Vδ2 T cells are controversial. Although both increased and decreased levels have been reported, most studies have shown the absolute numbers and relative percentages of γδ T cells to be unchanged. A reactivity of the Vγ9Vδ2 T cells to in vitro stimulation with mycobacterial antigens seems to be preserved during the course of disease [72]. A reported increase of γδ T-cell numbers among healthy individuals being in close contact with tuberculosis patients is remarkable [73].

When the involvement of γδ T cells in the response to infection is studied, a certain number of determining factors have to be kept in mind. For example, the

13 Introduction blood level of γδ T cells in the population may vary. Because of a high endemicity of malaria in some areas, increased numbers of γδ T cells are often observed in the normal population [74].

Tularemia Tularemia is a zoonotic disease caused by the intracellular bacterial pathogen Francisella tularensis, a gram-negative rod-shaped or coccoid bacterium with unknown reservoir in nature. Two major subspecies, differing in virulence in humans, are of clinical significance. The highly virulent subspecies F. tularensis tularensis (Jellison type A) is the causative agent in North America, while the less virulent subspecies F. tularensis holarctica (Jellison type B) is spread widely in the Northern hemisphere. The bacteria are spread to humans by direct contact with infected animals (mostly rodents or lagomorphs), arthropod bites, ingestion of contaminated food or water, or by inhalation of infected particles.

Depending on route of infection, the clinical expression of tularemia varies significantly. Glandular forms of tularemia may be manifested as ulceroglandular, caused mostly by vector-borne transmission, oculoglandular, when the eye is the point of entry, and oropharyngeal, acquired by ingestion of contaminated food or water. Typical features are the formation of a primary ulcer at the point of entry and enlargement of draining lymph nodes. Inhalation of aerosolised bacteria results in the respiratory form of tularemia and may result in primary pneumonia. Enteral tularemia is caused by ingestion of contaminated food or water, while typhoidal tularemia refers to cases with no indicated route of entrance.

The incubation period of tularemia is mostly short (3-5 days), with a sudden onset of disease, including high fever, chills, fatigue, general body aches, headache, nausea, abdominal pain, vomiting, and diarrhoea, with a wide

14 Introduction individual variation in severity. Type A tularemia caused by subspecies tularensis has a mortality rate of 5-10% without effective treatment, while type B disease caused by subspecies holarctica is only rarely lethal in humans even without treatment. Nonetheless, also type B tularemia is invasive. The acute phase is ensued by a considerable period of convalescence and frequently by suppurative complications.

Tularemia occurs endemically from North America across Japan to the northern and central parts of Eurasia. In Europe, the disease is reported in most countries, although not on the British Islands. Reported outbreaks of tularemia in Europe are often associated with the use of contaminated water, or with exposure in farming or hunting. The most prevalent form is ulceroglandular tularemia.

In the immune defence to F. tularensis, like that to other intracellular bacteria, cell-mediated immunity plays a major role. Multiplication of bacteria occurs in macrophages and these cells, when appropriately activated, are also the major effector population, eliminating invading bacteria [75]. To become activated, macrophages require an external source of proinflammatory cytokines, among which the IFN-γ is of major importance [76]. Successful control of intracellular infections, including tularemia, is associated with involvement of T cells. Both CD4 and CD8 T cells are active in the defense to F. tularensis [77]. In human tularemia, the T-cell response is targeted to a vast array of bacterial proteins without any clearly immunodominant antigen [78]. The resolution of F. tularensis infection is followed by long-lasting protective immunity and a T-cell memory for as long as 25 years after infection [79]. When the present work was initiated, virtually all data on the T-cell response to F. tularensis referred to αβ T cells. In one single case, blood γδ T-cell levels had been recorded and the Vγ9Vδ2 T-cell percentage found to be increased [80].

15 Introduction

Two entities of legionellosis – Legionnaires’ disease and Pontiac fever Legionellosis is another human disease caused by gram-negative intracellular bacteria. It occurs after inhalation of aerosol particles originating from water sources contaminated with bacteria of the Legionella genus. Frequent sources are air conditioning systems, cooling towers, showers, and whirlpool spas.

There are two clinical entities of legionellosis, which differ strikingly in severity. Legionnaires’ disease is a severe multisystem illness involving pneumonia. Features include fever, non-productive cough, headache, myalgias, rigors, dyspnea, diarrhea, and delirium. Time from exposure to onset of disease varies from 2 to 10 days and the attack rate is relatively low (< 10%). Risk factors include increasing age, smoking, male sex, chronic lung disease, immunosuppression (cancer, AIDS, immunosuppressive treatment), and diabetes mellitus. As the disease develops, the patients may become very sick and sometimes need intensive care including ventilation. The mortality rate of Legionnaires’ disease varies from 5 to 30%. Legionella pneumophila causes approximately 90% of reported cases of the entity.

Pontiac fever, the second entity, is usually a non-pneumonic infection caused by Legionellae. It is a self-limited, flu-like, febrile illness. After an incubation period of about 36 hours (few hours – two days), the onset is usually acute. The attack rate is high, reaching 95 – 100% of exposed individuals. Symptoms include myalgias, chills, fatigue, fever, headache, chest pain, muscle aches, cough, and nausea. Pontiac fever occurs in otherwise healthy individuals and the recovery usually ensues within 2 to 5 days of disease, even without treatment. Due to its mild expression and the difficulty to cultivate invading microbes, it has even been questioned whether the exposure to dead bacteria or bacterial toxins or other products may be the cause of symptoms.

Members of the Legionella genus multiply intracellularly in protozoans or mammalian phagocytic cells. Therefore, the pathology in humans is similar to

16 Introduction that of other intracellular infections, like tuberculosis, listeriosis, brucellosis, tularemia, and Q fever. To control and resolve the infection, a proper host T-cell response is critical.

The role of T cells in the immune response to L. pneumophila is well documented [81]. However, the involvement of γδ T cells in the immune response to legionellae seems not to have been at all studied.

17 Aims

AIMS

The general aim of the present work was to explore the expansion of Vγ9Vδ2 T cells in peripheral blood as a feature of the human response to intracellular bacteria. The work was enabled by our encounter during the study period of outbreaks of tularemia and Pontiac fever, two intracellular bacterial diseases with widely different clinical expression. Specific aims were to

- characterize the course of Vγ9Vδ2 T-cell expansion in tularemia, after vaccination of healthy subjects with the live tularemia vaccine, and in Pontiac fever

- determine the phosphoantigen nature of Vγ9Vδ2 T-cell stimulating antigen(s) from Francisella tularensis and Legionella micdadei

- determine whether Vγ9Vδ2 T cells from subjects undergoing tularemia or Pontiac fever respond in a cross-reactive manner to phosphoantigen from Legionella micdadei and Francisella tularensis

- determine whether Vγ9Vδ2 T cells respond to the GroEL and DnaK heat shock protein homologs of Francisella tularensis

- assay the capability of γδ Τ cells to produce TNF-α and IFN-γ in response to phorbol myristate acetate, a polyclonal stimulant, during the course of infection

18 Material and Methods

MATERIAL AND METHODS

Bacterial strains. F. tularensis LVS was obtained from the US Army Medical Research Institute of Infectious Diseases, Fort Detrick, Frederick, Md. and cultivated on McLeod agar plates, supplemented with IsoVitaleX (Becton Dickinson, Cockeysville, MD). For preparation of bacterial extracts, bacteria were grown in Chamberlain medium to an optical density at 540 nm of 0.7 and harvested by centrifugation. Strain FSC171 was isolated from a wound sample of a patient with typical ulceroglandular tularemia and classified as F. tularensis subsp. holarctica. The Schu strain is a prototypic representative of the highly virulent F. tularensis subsp. tularensis. Both strains were cultivated under the same conditions as used for LVS.

L. micdadei strain CCUG 31229 (Culture Collection, University of Göteborg, Sweden) was grown on selective legionella BCYE agar plates and 24 h cultures were used for preparation of extract.

Patients, vaccinees and control subjects. Through general practitioners, blood samples for the study were obtained from patients with ulceroglandular tularemia at various intervals after onset of disease during outbreaks of tularemia in 1996 – 2000 in northern Finland and Sweden. Only patients with serologically confirmed tularemia were included in the study. Most individuals were administered doxycycline or ciprofloxacin at their first medical visit, usually 5-10 days after onset of disease. According to medical records, treatment was generally effective.

All vaccinees included in the study were healthy individuals without a known previous history of Francisella or Legionella infections. Vaccination was performed with live tularemia vaccine (Salk Institute, Swiftwater, Pa.) by

19 Material and Methods scarification, following the manufacturer´s instruction. Vaccinees donated blood for our studies up to one year after vaccination.

Blood samples were obtained on several occasions from 14 patients undergoing Legionella micdadei-caused Pontiac fever during an outbreak in April 1999. This group of six men and eight women (22 to 57 years old, mean age 46 years) was sampled also 12 months after onset of disease.

As control individuals, healthy subjects without previous known exposure to intracellular pathogens were selected, with sex and age distribution close to experimental groups. Their blood γδ T-cell levels were quite similar to those previously reported for a normal adult Swedish population [82].

An informed consent was obtained from all individuals included in the study, using a protocol approved by the Committee on Research Ethics at the Medical Faculty, Umeå University, Umeå, Sweden.

Monoclonal antibodies. For surface phenotyping of monitored cells and intracellular cytokine detection following anti-human monoclonal antibodies were used:

Specificity Clone Conjugation Manufacturer CD3 SK7 PerCP BectonDickinson, Sunnyvale, Calif. UCH-T1 PE Serotec, Oxford, United Kingdom HIT 3a CyChrome Pharmingen, San Diego, Calif. CD4 Leu-3a FITC BectonDickinson, Sunnyvale, Calif. CD8 Leu-2a PE BectonDickinson, Sunnyvale, Calif. CD19 4G7 FITC BectonDickinson, Sunnyvale, Calif. FMC63 FITC Serotec, Oxford, United Kingdom CD25 FITC BectonDickinson, Sunnyvale, Calif. CD56 MY31 PE BectonDickinson, Sunnyvale, Calif. B-A19 FITC Serotec, Oxford, United Kingdom

20 Material and Methods

CD94 HP-3D9 FITC Pharmingen, San Diego, Calif. TCR-αβ WT31 FITC BectonDickinson, Sunnyvale, Calif. BMA 031 FITC Serotec, Oxford, United Kingdom TCR-γδ 11F2 FITC, PE BectonDickinson, Sunnyvale, Calif. 5.A6.E9 FITC Serotec, Oxford, United Kingdom TCR-Vγ9 B3 FITC Pharmingen, San Diego, Calif. 7A5 FITC Serotec, Oxford, United Kingdom TCR-Vδ2 B6 PE Pharmingen, San Diego, Calif. 15D FITC Serotec, Oxford, United Kingdom IL-4 8604B312 PE Biosource, Fleurus, Belgium IFN-γ 4S.B3 PE Pharmingen, San Diego, Calif. TNF-α Mab11 PE Pharmingen, San Diego, Calif.

Surface staining of PBMC. Two- or three-color analysis was carried out for surface phenotyping of cell suspensions, following the manufacturer´s protocol (Becton Dickinson, Sunnyvale, Calif.) and using directly conjugated monoclonal antibodies. Sample aliquots were incubated with 10 µl of normal mouse serum (DAKO, Glostrup, Denmark) at a dilution 1:500 for 15 min at room temperature before the appropriate MAbs were added, followed by incubation for 25 min at room temperature in dark. After incubation, cells were washed twice with Cell- Wash solution (Becton Dickinson). Finally, cells were resuspended in 500 µl of FACS-Flow solution (Becton Dickinson) and analyzed immediately on flow cytometer.

Intracellular staining of stimulated cells. For determination of cytokine- producing cells, a combination of surface staining for cell markers and intracellular staining for cytokines was applied after “in vitro” cell stimulation. A sample of whole blood was diluted (1:4) with RPMI 1640 medium and cells

21 Material and Methods were stimulated with 1 ng/ml of PMA and 1 µM of ionomycin (both from Sigma). Extracellular transport of produced cytokines was blocked by addition o of monensin (3 µM, Sigma) and cells were incubated for 4 h at 37 C in 5% CO2 – conditions previously found to be optimal [83]. After stimulation, cells were stained for surface phenotype markers, followed by permeabilization and intracellular staining according to the Pharmingen protocol for intracellular detection of cytokines. Briefly, surface-stained cells were fixed in 4% paraformaldehyde solution for 15 min on ice, washed twice and resuspended in permeabilization solution (0.1% saponin from Sigma in Cell-Wash). Permeabilization was carried out for 15 min, before the cells were pelleted by centrifugation and stained with monoclonal antibodies for cytokines. Stained cells were washed twice in permeabilization solution and resuspended in FACS- Flow solution before analysis on flow cytometer.

Flow cytometry analysis. Evaluation of surface lymphocyte staining was performed with a FACSort instrument (Becton Dickinson) by the use of CellQuest software. Lymphocytes were gated according to their morphological parameters by means of forward scatter and side scatter analysis or by expression of CD3. Data for 10,000 gated events per sample were recorded. Thresholds for marker positivity were determined by staining of samples with isotype-control antibodies. Results were expressed as percentage of cells positive for a given marker of a subpopulation.

Phosphoantigen preparation and treatment. For preparation of crude water extract, bacterial cultures were harvested by centrifugation. The cell pellet was treated overnight with 50 ml of chloroform-methanol (1:1, vol/vol) to kill bacteria. The material was vacuum-dried and resuspended in chloroform-water. The water fractions from five repeated extractions were collected and

22 Material and Methods concentrated. Delipidation of the material was performed by flash chromatography on a Sep-Pak C18 cartridge (Waters-Millipore, Boston, Mass.) [84]. Fractions containing water, methanol, and a mixture of methanol and acetronitrile (Scharlau, Barcelona, Spain) were subsequently eluted and tested for the presence of phosphate by use of colorimetric reaction with aminoheptamolybdenate in acidic environment. Fractions containing > 0.2 mg of P-/mg of dry weight were pooled, resuspended in water at a concentration of 50 mg/ml, and kept frozen. These fractions, referred to as phosphoantigen extract of Francisella (FE) or Listeria (LE), were used for further analysis.

Enzymatic treatment of extracts was performed in order to determine the nature of active substance stimulating human γδ T cells. Ten microliters of extracts (FE, and or LE) were added to 3 µl (2U) of calf intestinal alkaline phosphatase (Boehringer, Meslan, France), 3 µl (0.2U) of Crotalus adamanteus venom nucleotide pyrophosphatase (Sigma), and 1 µl of dephosphorylation buffer (Boehringer) and incubated for 30 min at 37oC before the bioactivity test. Although DC protein assay, performed according to the manufacturer’s instructions (Bio-Rad Laboratories), did not detect any protein in the extracts, the extracts were subjected to proteinase treatment to degrade any remaining rest. Ten microliter aliquots of FE or LE were mixed with 1 µg of proteinase K (Sigma) in acidic environment and incubated for 30 min at 37oC before the bioactivity test.

In vitro polyclonal γδ T-cell expansion. To analyze the ability of bacterial extracts to induce proliferation of human γδ T cells in vitro, PBMC were prepared from blood samples obtained in heparinized tubes by centrifugation on a Ficoll-Metrizoate gradient (Lymphoprep; NYCOMED AS, Oslo, Norway). o Purified cells were cultivated at 37 C, 5% CO2 in RPMI 1640 medium supplemented with 10% pooled human serum, 100 mg/l of gentamicin, in the

23 Material and Methods presence or absence of IL-2 (10 U/ml; Bionative, Umeå, Sweden) and with addition of bacterial extracts. As a positive control for expansion of human γδ T cells, isopentenyl pyrophosphate (2 µg/ml; Sigma) was used as described previously [Wesch, 1997]. The proportion of γδ T cells and their subpopulations was assayed in cultures after eight days of cultivation by flow cytometry.

Protein purification. Protein antigens were prepared from F. tularensis LVS cultures as previously described [85]. Bacteria were grown in liquid Chamberlain medium [86] to an optical density of 0.7 at 450 nm. To enhance the amount of stress proteins, cultures were incubated for 1 hour in medium with 5 mM hydrogen peroxide. Pelleted cells were mixed with lysing buffer containing 9 M urea, 2% [v/v] Ampholine pH 3.5-10.0 [Bio-Rad Laboratories, Richmond, CA], 2% [v/v] Triton X-100 and 8 mM phenylmethylsulphonyl fluoride and stored at –80oC. Thawed material was diluted 1/6 with sample buffer (8 M urea, 2% [v/v] β-mercaptoethanol, 2% [v/v] Ampholine pH 3.5-10.0, 0.5% [v/v] Triton X-100) and subjected to two-dimensional gel electrophoresis. In the first dimension, separation according to pI was performed using immobilized pH gradient strips (immobiline DryPlate, pH 4.0-7.0; Pharmacia LKB Biotechnology, Uppsala, Sweden), following the manufacturer’s instructions. Separation in a second dimension according to molecular weight was performed on uniform 14% acrylamide gel. Proteins were visualized by staining with CuCl2 for 30 min. Relevant proteins were destained, eluted from gel pieces, and dialyzed. The protein concentration in concentrated samples was determined by use of DC protein assay (Bio-Rad Laboratories).

Lymphocyte stimulation. PBMC for in vitro stimulation with protein antigens were prepared as described above. For assay of proliferative response, 200-µl

24 Material and Methods cultures in 96-well tissue culture plates were incubated for 6 days at 37oC and

5% CO2. Cultures contained RPMI 1640 medium supplemented with 15% pooled human serum and 100 mg/l of gentamicin. Purified F. tularensis LVS proteins were used for stimulation at a concentration of 1 µg/ml, as estimated by DC protein assay. As control antigens, heat-killed F. tularensis LVS (106/ml) and purified protein derivative (PPD) from M. tuberculosis (10 µg/ml; Statens Seruminstitut, Copenhagen, Denmark) were used. Twelve repeats of each combination of stimulants were established. On day 5, a 100-µl volume of supernatant was collected from each well and stored at –70oC for cytokine determination by an ELISA, and wells were restored with 100 µl of culture media. Sixteen hours prior to termination of cultures, 0.5 µCi of [3H]thymidine was added to half of wells for liquid scintillation analysis of cell proliferation. The rest of the wells were used to perform flow cytometry.

In vitro cytokine release. Amounts of cytokines produced into supernatants of cell cultures were measured by a sandwich enzyme-linked immunosorbent assay (DuoSet; Genzyme Diagnostics, Cambridge, MA) according to the manufacturer´s instructions.

25 Results

RESULTS

F. tularensis-induced expansion of γδ T cells and phosphoantigen nature of stimulating antigen (I)

In one single case of tularemia, the proportion of blood γδ T cells had been previously determined [80]. In a sample obtained from this patient before treatment, T cells bearing the γδ TCR constituted as much as 32.7% of all blood lymphocytes. If such an expansion could be shown to occur more generally in the disease, it should be pathologically important and possibly a denominator of host control of tularemia. Our motive to undertake the study was also related to data previously presented on mycobacterial infection [50, 87], indicating the presence of γδ T-cell stimulating antigens in bacterial extracts [34-36]. Tularemia and tuberculosis are both caused by facultative intracellular bacteria and show similar histopathological features. If we were able to find a similar antigen activity in extracts from F. tularensis, a γδ T-cell expansion in the disease would be better understood.

In August 1996, ulceroglandular tularemia occurred in Northern Sweden and in Finland. Through general practitioners, we obtained a blood sample from each of 12 patients, all later serologically verified, during their second or third week of disease. Most patients were administered doxycycline at their first medical visit, usually 7 to 12 days after onset of symptoms and responded well to treatment. We also repeatedly sampled a patient hospitalized in Umeå with ulceroglandular tularemia. Analysis of the first 13 samples, collected 7 to 21 days after the onset of disease, revealed a significant increase of the proportion of γδ T cells among CD3+ T cells, the mean value being 30.5% as compared to 6.5% in samples from healthy age-matched control subjects. In samples from the patient hospitalized in Umeå, obtained on day 8, 12, and 27, the values were 26, 48, and 42%. The increase was consistently attributed to cells expressing the Vγ9Vδ2 pair of TCR chains. Although the level of γδ T cells seemed to peek

26 Results within the first two weeks of onset, it should be recalled that only one sample was tested outside this interval.

At the time when these studies were undertaken, antigens recognized by Vγ9Vδ2 T cells had been identified in extracts from M. tuberculosis and Plasmodium falciparum and characterized as nonpeptidic alkyl pyrophosphate esters, either alone or as a 5´-phosphate substituent of UMP or dTMP [84, 88]. Four of them, all identified in mycobacteria, were called TUBag 1 – 4. Chemically synthesized alkyl pyrophosphates were known to be active as well. One of them, a general cellular constituent, is isopentenyl pyrophosphate (IPP). These antigens, collectively called phosphoantigens, had been shown to stimulate Vγ9Vδ2 T cells in a TCR-dependent way, although without regard to known previous sensitization of the host to the homologous microbial agent. Starting from a cell pellet of F. tularensis FSC171, an isolate from a wound sample of a patient with ulceroglandular tularemia, we prepared phosphoantigen as previously performed for mycobacteria. The procedure included chlorophorm-methanol extraction and flash chromatography on a Sep-Pak C18 cartridge. The water extract was shown to induce expansion in vitro of Vγ9Vδ2 T cells of PBL from healthy donors as efficiently as did TUBag3 or IPP.

The phosphoantigen nature of the F. tularensis cell extract was confirmed by phosphatase treatment. The untreated extract induced a strong TNF-α release and autocytotoxicity in the Vγ9Vδ2 T-cell clone G12, activities which were completely abrogated by phosphatase treatment.

In order to compare the phosphoantigenic loads of F. tularensis strains of different virulence, aqueous extracts were prepared from two virulent isolates (FSC171 and Schu S4) and the live vaccine strain (LVS), using Mycobacterium fortuitum-derived extract as a reference. In both assays used, TNF-α release by the G12 clone and in vitro expansion of Vγ9Vδ2 T cells of PBL cultures from healthy donors, stimulation of Vγ9Vδ2 T cells was demonstrable at comparable

27 Results titers for the four different extracts. Thus, no quantitative correlation between virulence of the strain of F. tularensis and phosphoantigen load was demonstrated.

With this result in mind, we investigated the proportion of Vγ9Vδ2 T cells in blood from subjects undergoing vaccination with the LVS strain of F. tularensis. When 7 healthy vaccinees were followed for up to 4 months after vaccination, no significant expansion was demonstrated. The peak values observed at day 8 after vaccination were only slightly higher than the initial values. Thus, the prominent increase of Vγ9Vδ2 T cells in peripheral blood seen in tularemia patients was not observed after vaccination with the LVS strain.

In conclusion, we found γδ T-cell expansion in blood to be a general event in the course of tularemia. The patients underwent the ulceroglandular form of the disease and were successfully treated, indicating that the response occurred in uncomplicated disease. We also demonstrated the presence of phosphoantigen in F. tularensis. Despite a similarity in phosphoantigen activity between the live vaccine strain and a virulent strain derived from the region where the patients lived, vaccinees showed only a minor increase or no increase at all of γδ T cells in circulation.

Stress proteins of F. tularensis and human γδ T cells (II)

The Vγ9Vδ2 T-cell response in vitro to phosphoantigen produced by F. tularensis was a finding of putative relevance to the expansion of the subset seen in subjects undergoing tularemia. Before accepting that proteins are not involved, previous studies on heat shock proteins were taken in consideration. It had been claimed that human γδ T cells can recognize heat shock proteins [89- 91], implicating that they may have a surveillance role in bacterial disease as well as in the response to tumor development. Most pertinent, Daudi Burkitt´s

28 Results lymphoma cells, which express the GroEL-related heat shock protein chaperon- 60 (Cpn-60) on their surface, were found to induce expansion and activation of human Vγ9Vδ2 T cells, events which were inhibited by mAb raised against human or mycobacterial GroEL homolog [92-93]. Since in our laboratory, the GroEL- and DnaK-homologs of F. tularensis had been isolated [85, 94], we were ready to try them as antigens for Vγ9Vδ2 T cells.

Purified francisella-analogs to heat shock proteins (Ft-DnaK, the GroEL homolog Ft-cpn60, and the GroES homolog cpn10), a 17kDa membrane protein, heat-killed F. tularensis microbes, and IPP were used for in vitro stimulation of PBL. Compared with reference subjects lacking a history of tularemia or tularemia vaccination, individuals undergoing tularemia 10 or 30 years ago showed a significantly (p = .01) higher proliferative T cell response to all three heat shock proteins. By flow cytometry, blast cells were shown to express the αβ TCR. Under conditions that allowed vigorous expansion of Vγ9Vδ2 T cells in response to IPP, no expansion of Vγ9Vδ2 T cells occurred in response to DnaK or Cpn60 of F. tularensis. Neither did heat-killed F. tularensis induce expansion of Vγ9Vδ2 T cells. Similar to the response to various membrane proteins of F. tularensis, recall responses to the heat shock proteins were restricted to αβ T cells, with a predominant response in the CD4 T-cell fraction and a minor response in the CD8 T-cell fraction.

Thus, we found no Vγ9Vδ2 T-cell response to heat shock proteins of F. tularensis. Although this contradicted a previously suggested role of the subset in survey of intruders based on these conserved proteins, our results were in line with another study on the mycobacterial GroEL homolog, indicating that αβ T cells, but not γδ T cells, recognize the proteins [95].

29 Results

Time course of expansion and cytokine response of Vγ9Vδ2 T cells in tularemia (III)

To better characterize the course of the γδ T-cell elevation after onset of tularemia, 108 patients in the range of 2 to 83 years old were assayed, giving in total 181 blood samples. By collaboration with general practitioners, samples were obtained as soon as the first day after the onset of disease and the period of sampling was extended up to 24 months after onset.

In samples collected during the first week of the disease, the proportion of γδ T cells of CD3+ cells did not differ significantly from that observed in a group of healthy reference subjects. Of 48 samples, only 2 showed >10% γδ T cells. These samples were from two 2-year old children.

After the first week the level increased rapidly, a peak being reached at the interval 18 – 40 days. By analysis of samples taken at 3 – 4, 12, 18, and 24 months after onset of disease, a slow decrease in γδ T-cell proportions were found, but even after 18 months the level was still significantly higher than that of samples from control subjects. When data from 10 repeatedly sampled patients were analysed, the time curve was similar to that obtained by connecting the mean values of all samples, supporting the generally observed trend. In 42/45 subjects sampled within 3 months of onset, a Vγ9Vδ2 T-cell percentage of > 10% was found.

The increase of γδ T-cell percentage in PBL of the tularemia patients might have been due to reduced absolute numbers of αβ T cells in periphery. Therefore, absolute numbers of αβ T cells were calculated for eight patients sampled repeatedly during the first week and also at the interval of 8 – 30 days after the onset of disease. No significant differences in absolute numbers of αβ T cells were found over time, confirming the presence of γδ T-cell expansion.

30 Results

γδ T cells are effective producers of TNF-α and IFN-γ, cytokines of crucial importance to the activation of mononuclear phagocytes and control of intracellular bacterial infections and also potent mediators of inflammation. In the present study, the percentage of cells with intracellulary accumulated cytokines was determined after a 4-hour period of in vitro exposure of PBL to the polyclonal stimulant, phorbol myristate acetate (PMA). Patients were assayed at various intervals after onset of tularemia and compared to control subjects. The proportion of γδ T cells capable to accumulate TNF-α was significantly reduced during the first week after onset of tularemia, i.e., even before Vγ9Vδ2 T cells started to expand. This reduction was even more pronounced at the 8-17 day interval, i.e., when expansion of Vγ9Vδ2 T cells had been found to occur. A restoration to normal capability to accumulate the cytokine was demonstrated at 3 months after onset. As for IFN-γ production by γδ T cells, a slight although nonsignificant decrease of IFN-γ stained cells was observed after the first week of onset of disease. Irrespective of interval tested, no accumulation of IL-4 was induced by PMA in the γδ T cells.

In summary, the results of paper III showed a consistent increase of circulating Vγ9Vδ2 T cells in tularemia. An initial delay of one week and a prolonged course of the elevation were demonstrated. The percentage of γδ Τ cells expressing TNF-α in response to PMA was decreased during the first month of onset, possibly reflecting a modulation of an inflammatory response.

Expansion of Vγ9Vδ2 T cells in Pontiac fever, an intracellular bacterial disease with a clinical expression different from that of tularemia (IV)

The accumulated data on Vγ9Vδ2 T cells in human diseases is fragmentary. Only a handful number of bacterial diseases has been studied and in some of them, only a few patients were included. When faced with an outbreak of

31 Results suspect legionellosis, we therefore immediately asked for ethical evaluation to investigate the Vγ9Vδ2 T-cell response of the patients involved. Our wish was founded on two premises. One was related to the possibility to study a disease with a clinical expression different from intracellular infections previously studied. The epidemiological circumstances and clinical signs and symptoms suggested that this was an outbreak of Pontiac fever, a flue-like mild and transient disease different from other conditions, tularemia included. A Vγ9Vδ2 T-cell expansion in Pontiac fever would support the assumption that intracellular bacterial diseases share an involvement of the T-cell subset.

A second premise was a possibility to sample a group of patients, showing a well-defined date of exposure and onset of disease. Preliminary data indicated that the patients had been exposed during one single weekend, by visiting a whirlpool, and they all developed disease within 1 to 2 days. By repeated sampling, the Vγ9Vδ2 T-cell response was individually monitored. The source of the outbreak was retrospectively determined to be the whirlpool [96] and the causative agent serologically confirmed to be Legionella micdadei, an agent previously associated with whirlpool-associated outbreaks of Pontiac fever [97, 98].

Repeated blood samples were obtained on five occasions from 14 patients. Upon first sampling performed 4 to 6 days after the onset of disease, γδ T-cell levels were in the range of normal values. Thereafter, the proportion of γδ T cells of CD3+ cells rose significantly. After a peak at 2 to 7 weeks of onset, the proportion of γδ T cells decreased slowly for 6 months, although without reaching normal values. Results were similar when expressed as absolute numbers of γδ T cells per milliliter of blood, indicating that the percentage of γδ T cells increased because of expansion and not as a result of concomitant decrease in number of αβ T cells.

32 Results

The expansion of γδ T cells was attributed to Vγ9Vδ2 T cells. When we separately assayed Vγ9Vδ2 T cells, an interesting observation was done. The proportion of Vγ9Vδ2 T cells was significantly reduced at the time of first sampling, 4 to 6 days after onset of disease. At this early stage, the mean percentage of Vγ9Vδ2 T cells among CD3+ cells was only 1.0%, as compared to 5% in referent subjects. This phenomenon had not been observed among patients undergoing ulceroglandular tularemia, where Vγ9Vδ2 T cells predominated among γδ T cells in all investigated time periods (Paper III).

Similarly to our work on γδ T-cell function in tularemia (Paper III), functional capabilities of the cells and their possible regulation were tested in the course of L. micdadei infection. The capacity of γδ T cells in peripheral blood of L. micdadei-infected individuals to produce TNF-α and IFN-γ cytokines early after expansion was significantly higher than that of control subjects. Later during expansion, at 5 to 7 weeks, a striking reduction in cytokine-producing ability was observed. At 6 months after onset, the capacity of γδ T cells to produce cytokines upon in vitro stimulation was restored and in the case of IFN- γ, it was even higher than in control subjects.

Because of lack of MHC restriction, other mechanisms must ensure regulation of γδ T cells. One of them might be the expression of regulatory receptor molecules typical for NK cells [99]. In the course of legionellosis, the proportion of CD94 expressing cells among Vγ9Vδ2 T cell subset was monitored. A significant increase (p = 0.01) of surface CD94 expression was observed along with expansion of γδ T cells at 5 to 7 weeks after onset of disease, followed by normalization at the 6-month interval.

Together with previous studies on a range of intracellular bacterial diseases, the results of paper IV suggested that Vγ9Vδ2 T-cell expansion is a common feature irrespective of a varying clinical expression of the conditions. A novel

33 Results finding was an initial dramatic decrease in Vγ9Vδ2 Τ cells in circulation, before start of the expansion, suggesting an extravasation of the cells in response to the infective agent.

Response of Vγ9Vδ2 T cells from individuals undergoing legionellosis or tularemia to phosphoantigen from Francisella or Legionella (V) In the present outbreak of Pontiac fever, all bacteriological culture attempts had been unsuccessful, and the disease was confirmed only by serological investigation to be caused by L. micdadei. To identify phosphoantigen in L. micdadei, we analyzed bacteria of a reference strain. By the same procedure as used for F. tularensis (Paper I), we prepared a water-methanol extract from pelleted L. micdadei and tested the extract on PBL from blood donors. Phosphatase and protease treatment of the active fraction confirmed the phosphoantigenic nature of the active compound.

By use of blood samples from subjects recovering from tularemia and Pontiac fever, we approached the question of the specificity of the Vγ9Vδ2 T- cell response. γδ T cells show traits referred to cells of both the innate and the adaptive immune system. Although Vγ9Vδ2 T cells vary extensively in fine structure of their TCR, they recognize in a cross-reactive manner a restricted type of molecules, preferentially small nonpeptidic alkyl pyrophosphates and alkylamines. Such a pattern of recognition might indicate a capability of Vγ9Vδ2 T cells to recognize a wide variety of microbial agents and therefore, priming due to infection with one agent might be of value when later encountering a new agent.

More specifically, we asked the question of whether subjects recovering from tularemia would react more strongly to a crude phosphoantigen extract prepared from F. tularensis than to an extract from L. micdadei and vice versa. PBL from patients undergoing legionellosis or tularemia about 12 months

34 Results previously, tularemia vaccinees, and healthy donors were incubated in vitro for 8 days in the presence of Francisella or Legionella extract. In all groups of subjects, expansion of Vγ9Vδ2 T cells occurred and the rate of expansion was quite similar irrespective of use of Francisella or Legionella extract as antigen. A great majority of Vγ9Vδ2 T cells of individuals undergoing Pontiac fever were found to express CD45RO, a marker for the memory phenotype. Thus, a Vγ9Vδ2 T cell memory was demonstrated, however no memory specific to phosphoantigen from the causative agent.

In essence, our results were in line with the concept of a broad-reactive set of memory Vγ9Vδ2 T cells built up by exposure to various causative agents. Exposure to one facultative intracellular bacterium seems to endow the host with an increased capability to react to an unrelated microbial intruder within the group of facultative intracellular bacteria.

35 Discussion

DISCUSSION

Vγ9Vδ2 T-cell expansion as a feature of intracellular bacterial diseases

With regard to host-parasite interaction, extracellular and intracellular bacteria differ completely. Extracellular bacteria are killed by phagocytosis, whereas intracellular bacteria need the intracellular environment for their replication. The host defence is adapted to these discrete principles of attack, including a mandatory role of humoral immunity in extracellular and of cell-mediated immunity in intracellular infection. When data now accumulate indicating that a select T-cell population, the Vγ9Vδ2 T-cell subset, is expanded in the circulation of patients acquiring intracellular bacterial diseases, the generalisation of the phenomenon seems important to substantiate.

So far, Vγ9Vδ2 T cells have been studied only in some of all intracellular bacterial diseases and except for tuberculosis, the data generated is based on relatively small clinical materials (Table 2). In general, expansion of γδ cells was restricted to the Vγ9Vδ2 subset, while Vδ1 T-cell expansion was detected only sporadically. In brucellosis, expansion of Vγ9Vδ2 T cells was documented in blood from six patients, reaching nearly 50% of T cells [66]. Similarly, in one report including eleven samples from four patients with Q-fever, an expansion of Vγ9Vδ2 T cells was shown [100]. In listeriosis, another intracellular bacterial infection, two studies each including eight patients showed increased percentages of γδ T cells in blood [67, 101]. In one of these reports, expansion of both δ1 and δ2 subsets was documented [101]. When patients with Salmonella infection were assayed, increased average values of Vγ9Vδ2 T cells were recorded among seventeen samples from six cases with systemic infection, whereas a lower elevation was found also in fifteen samples taken from seven patients with gastroenteritis [63]. Out of the range of other observations were those on ehrlichiosis described by Caldwell [102]. Mean numbers of Vγ9Vδ2 T

36 Discussion cells among 18 assayed patients exceeded 50 % of T cells (with maximal value reaching 97 %), including one sample from each patient.

Except for the present studies, one single case of tularemia has been recorded, and in legionellosis no case at all. The studies of paper I, III, and IV added 195 samples from 121 patients with tularemia and 53 samples from 14 patients with legionellosis. The results did indeed consolidate the assumption that Vγ9Vδ2 T cells are involved in the pathophysiology of intracellular bacterial diseases. Moreover, the increase of Vγ9Vδ2 T-cell levels in tularemia and legionellosis was indeed a consistent finding. It should be remarked that the blood samples were obtained from tularemia patients during several years and from several geographic regions in Sweden and Finland. Based on our demonstration of a consistent increase of Vγ9Vδ2 T-cell levels in tularemia and legionellosis, it seems relevant to ask the question of whether a deficiency of the subset may occur and what might be the consequences of a deficiency for the pathophysiology of these infections. So far, there are no data on such human cases reported in the literature. According to studies of host defence in animal models, αβ T cells may compensate more or less for the γδ T cells. However, the relevance of the comparison is uncertain, especially since a subset corresponding to the Vγ9Vδ2 T cells cannot be found in rodents. A future discovery of a Vγ9Vδ2 T-cell deficiency would be of great informative value and screening of T-cell populations in cases of intracellular bacterial infections with severe manifestations therefore seems warranted.

In the context, mycobacterial disease needs to be separately discussed. A number of studies determined the involvement of γδ T cells during in vivo infection with M. tuberculosis. In studies involving adult TB patients, however, results on levels of peripheral γδ T cells were inconsistent. In some reports, increased numbers were documented in patients [87, 103], or among hospital workers in contact with TB patients [73]. Other reports showed no changes in γδ

37 Discussion

T-cell frequency [50, 104]. These studies, since they were performed mainly on adult patients, may reflect a situation separate from primary infection. In this regard, studies involving children might be more relevant [72, 105]. In such a study, the ability of peripheral blood γδ T cells to proliferate in vitro in response to mycobacteria-derived antigens was demonstrated only in PPD+ children (both patients and controls) and increased levels of Vγ9Vδ2 T cells were detected in the cerebrospinal fluid of children undergoing tuberculous meningitis. The in vitro proliferative activity as well as the in vivo increase of Vγ9Vδ2 T-cell levels was significantly reduced by chemotherapy. A failure of Vγ9Vδ2 T cells to respond to M. tuberculosis-derived antigens was found in patients with active pulmonary tuberculosis [106], a phenomenon later suggested to be explained by activation-induced cell death [107]. Taken together, an activation of Vγ9Vδ2 T cells seems to be associated with a successful immune response to infection with M. tuberculosis, while chronic infection may possibly be linked to a deficiency in Vγ9Vδ2 T-cell functions.

Little data is available on Vγ9Vδ2 T cells in extracellular bacterial diseases. In one report [108], expansion of γδ T cells was described. However, enumeration of these cells was based on negative ex-vivo selection, instead of direct labeling of peripheral blood cells with anti-TCR antibodies, as in other reports. Moreover, these cells were CD4+ and although their γδ TCR-chain usage was not investigated, they did obviously not belong to the Vγ9Vδ2 T-cell subset. The nature of the subset recorded by the authors remains unclear. In a study of pertussis, an extracellular infection associated with tissue damage, an increase of CD4+ and CD8+ T cells and B cells and a decrease in the γδ T-cell level in blood were recorded [109]. In two studies on intracellular bacterial infections, referent groups were included comprising 13 and 4 patients with extracellular invasive bacterial diseases, showing normal γδ T-cell values [63, 100].

38 Discussion

Altogether, those sparse data that exist on extracellular bacterial diseases show a normal proportion of Vγ9Vδ2 T cells in peripheral blood, affording odds for the assumption that the cell subset is indeed involved more specifically in the pathophysiology of intracellular bacterial infections. Provided it can be confirmed by larger studies on extracellular bacterial diseases, that the Vγ9Vδ2 T-cell expansion is an event more closely related to intracellular infections, the Vγ9Vδ2 T-cell expansion may become useful as a diagnostic marker, and perhaps, when its relation to pathophysiology is better understood, may be helpful in determination of disease severity.

Similar to intracellular bacterial diseases, an expansion of circulatory Vγ9Vδ2 T cells has been found also in infections caused by protozoans. Among them, toxoplasmosis and malaria have been more carefully studied. In toxoplasmosis, the increase in Vγ9Vδ2 T cells was found to be only moderate, average values varying around 10 % of all T cells. For malaria, data from natural [74, 110-112] and experimental infections [113] have demonstrated a γδ T-cell expansion at levels similar to those of intracellular bacterial diseases. Moreover, Vγ9Vδ2 T cells are stimulated in vitro in the presence of live, rather than killed malaria parasites [114], and respond by production of cytolytic and proinflammatory molecules including cytokines of importance to the host defence [115]. Thus, the role of γδ T cells is not only limited to bacterial diseases, but may be a general feature of the host response to protozoan and intracellular bacterial infections.

39

TABLE 2. Human γδ T-cell expansion in vivo in various infectious diseases Mean (max) Expansion of No. of patients Agents Disease Persistence Reference γδ T-cell % γ9δ2 δ1 (samples)

Bacteria M. tuberculosis tuberculosis 8 (26) ? ? 14 (29) < 4 weeks [103] pulmonary tb 6.4 (16.5) ? ? 20 ? [87] tb contacts 10.4 (18) ? ? 17 ? [73] M.t. or M.a-i. and HIV+ 19.7 + + 9 ? [116] F. tularensis tularemia 32.7 + - 1 > 1 month [80] tularemia 30.5 (47.8) + - 13 (14) > 3 weeks Paper I tularemia 25.2 (47.8) + - 108 (181) > 18 months Paper III Legionella micdadei Pontiac fever 18 (41) + - 14 (53) > 6 months Paper IV Brucella melitensis brucellosis 29 (48) + - 6 > 4 m (3/6) [66] Coxiella burnetii Q-fever 16 (30) + - 4 (11) > 14 m (2) [100] Listeria monocytogenes listeriosis 11.7 (35.3) + + 8 (23) > 12 m (1) [101] listeriosis 13 + - 8 > 40 d (2) [67] Salmonella sp. systemic 28.9 (47.7) + - 6 (17) > 17 m (1) [63] gastroenteritis 10.5 (23.8) + (1) - (1) 7 (15) < 5 m (2) Ehrlichia chaffeensis ehrlichiosis 57 (97) + - 18 4 weeks [102]

Haemophilus influenzae 27 (37) ? ? 12 ? Streptococcus pneumoniae meningitis 25 (42) ? ? 8 ? [108] Neisseria meningitidis 35 (46) ? ? 8 ?

40

Protozoa Leishmania donovani visceral leishmaniasis 44 ? ? 8 ? [64] Toxoplasma gondii toxoplasmosis 9.4 (14.9) + - 8 ? [57] toxoplasmosis 10.3 + - 12 ? [56] acute congenital tox. 12 (17.6) + - 3 ? [117] Plasmodium falciparum malaria (adult endemic) 16 (26) ? ? 4 (12) > 28 days (1) [74] malaria (untreated) 15.5 ? ? 9 > 28 days [110] malaria (non-endemic) 17.9 + + 22 (66) > 28 days [65] malaria (non-endemic) 19 + - 29 > 4 months [111]

Plasmodium vivax malarial paroxysm (non-endemic) 11 (28) ? ? 25 ? [112]

Eucaryota Onchocerca volvulus onchocerciasis 6.7 (17.8) - + 8 (24) > 14 months [118]

41 Discussion

The heat shock protein homologs Cpn60 or DnaK of F. tularensis are not recognized by Vγ9Vδ2 T cells

During the last decade it has become clear that αβ T cells constitute a more heterogenous subset of microbe-recognizing cells than has been previously recognized. Besides CD4+ and CD8+ αβ T cells, CD1-restricted cells have been found to recognize lipids/glycolipids from mycobacteria. Most likely, a variety of bacteria residing in the macrophage vacuole may be targeted by these cells. The event is novel and has not yet been extrapolated to agents other than mycobacteria.

The sentinel role of γδ T cells in infections is similarly unexplored. Based on this background, it seemed important to consider results presented more than a decade ago [93], indicating that γδ T cells, and Vγ9Vδ2 T cells in particular, recognize the conserved heat shock protein homologs Cpn-60 and DnaK. In some early reports, reactivity of several γδ T-cell lines towards the mycobacterial HSP65 (homologous to Cpn-60) [119, 120] was demonstrated while in others, the reactivity of γδ T cells towards recombinant HSP65 was found to be poor [95].

Although it is clearly established that mycobacteria induce Vγ9Vδ2 T-cell stimulation in vitro [121, 122], the role of heat shock protein as antigen for the stimulation seems doubtful. Small protease-resistant phosphate molecules have been shown to induce expansion of Vγ9Vδ2 T cells in vitro [33, 34, 36], including antigens found in mycobacteria [35, 123]. Not only the HSP65, but also HSP10 (Cpn-10) protein of M. tuberculosis has been implicated as a γδ T- cell stimulant. However, the stimulatory activity was found to be associated with carried phosphoantigen and was not due to the protein part of the preparation used [124]. It should be remarked that bacterial heat shock proteins have been found to be recognized by αβ T cells [63, 100].

42 Discussion

In our first experiments (Paper I), we found the stimulatory moiety in F. tularensis to be a low-molecular weight non-peptidic phosphate-containing molecule, similar to those found in M. tuberculosis. To find out whether the F. tularensis-derived stress proteins DnaK, Cpn60 or Cpn10 can participate in the stimulation of Vγ9Vδ2 T cells, the experiments described in Paper II were performed. These stress proteins were able to stimulate αβ T cells from previously infected individuals, while γδ T cells from the same individuals did not at all respond to such stimulation. Nevertheless, the proliferative response of Vγ9Vδ2 T cells from these individuals was retained, as confirmed by their expansion upon stimulation with IPP. We concluded that in the case of F. tularensis infection, the in vivo observed expansion of Vγ9Vδ2 T cells is most probably a result of phosphoantigenic stimulation, rather than a response to protein-derived antigens.

Implications from data on time course of expansion of Vγ9Vδ2 T cells in tularemia and legionellosis The present investigations (Paper I, III, and IV) afforded comprehensive information on the time course of the Vγ9Vδ2 T-cell elevation in both tularemia and Pontiac fever. Tularemia is an invasive disease affecting several organ systems, whereas Pontiac fever is a transient flu-like disease. In spite of such a wide difference in clinical expression between the diseases, the course of the Vγ9Vδ2 T-cell expansion was quite similar.

In tularemia patients, the level of γδ T cells in peripheral blood increased after the first week of disease, reached a maximum within the third and fourth weeks, and slowly decreased during 18 months (Paper III). A similar course was observed in legionellosis (Paper IV). Values were maximal at two to six weeks after onset and declined over a period of six months, although not to normal levels. So far, data on γδ T-cell kinetics in other diseases are only sporadic. In

43 Discussion brucellosis, elevated numbers of Vγ9Vδ2 T cells were reported in three patients at 4 months after onset [66]. Similarly, in two patients with Q-fever increased levels of Vγ9Vδ2 T cells were observed for as long as 14 months [100]. In listeriosis, one report described a persisting increase of Vγ9Vδ2 T cells in one patient for 12 months [101], and a second report in two patients for more than 40 days [67]. Sporadic information is also available for Salmonella infection. In one case of systemic infection, an elevated Vγ9Vδ2 T-cell number was observed 17 months after onset, while in two cases of gastroenteritis, values were normalized within 5 months [63].

More comprehensive are data from malaria patients. Elevated values in individual cases were reported on day 28 after onset [65, 74, 110]. In a more extensive study [111], increased levels of Vγ9Vδ2 T cells were observed among 29 patients for more than 4 months. Together with these reports, the data of paper I, III, and IV allow a generalisation on the course of elevation of the Vγ9Vδ2 T cells in intracellular infections, namely a week´s delay, a peak 2-4 weeks after onset, and a slow continuous decrease over several months or even a year.

A week’s delay before start of replication of γδ Τ cells might be expected, because they are believed to be dependent on cytokine support from CD4 T cells [125, 126]. In line with this, a recall response of CD4 cells to protein antigens of F. tularensis is usually first demonstrated during the second week of tularemia [127].

Only two of all blood specimens from tularemia patients showed values out of the range, both obtained from 2-year-old children (Paper III). In these cases, significantly elevated γδ T-cell numbers were found as early as on the first day after onset of the disease. The reason for this aberration is unknown. It should be emphasized, however, that the exact day of onset of tularemia in small children may be less exactly recorded and the most simple explanation may be that the

44 Discussion disease developed insidiously. If, on the other hand, the day of onset were indeed appropriately observed, the values would suggest that an increased level of Vγ9Vδ2 T cells may appear earlier after a primary exposure than after secondary exposure to phosphoantigens. In conflict with such a possibility, however, vaccination with BCG enhanced the response of γδ Τ cells to mycobacterial antigens [128].

The slow decline of values of Vγ9Vδ2 T cells in tularemia and Pontiac fever over time is most likely not due to a persisting presence of microbes, as latency is not documented to occur in these diseases. On the contrary, bacteria are eradicated efficiently. Moreover, the sporadic occurrence of F. tularensis in the environment argues against a possible reexposure.

The present results may have some relevance when trying to understand the difference in peripheral blood Vγ9Vδ2 T-cell levels in various geographical regions. When normal values of γδ T cells in groups of healthy individuals from different areas of the world were analysed [82], significant differences were found. The mean value for a group of 24 Swedish donors was found to be 4.2%, ranging from 0.5 to 9.4% - an observation in good agreement with results obtained in our studies (Paper I, III and IV). Significantly higher values were documented for a group of 26 Turkish citizens (9.3%) and also 14 non-Japanese Asian donors (9.2%). The mean value of a group consisting of 35 Japanese donors was 4.5%, although in this group 6 individuals expressed values > 10%. Even though subpopulations of γδ T cells were not analyzed, it seems probable that a major fraction had the Vγ9Vδ2 phenotype. These cells are known to form only a minor portion of γδ T cells upon birth and are later undergoing extrathymic expansion in response to environmental factors like intracellular infections. A high incidence of tuberculosis and other intracellular diseases in the areas, where the donors with high values came from, suggests that environmental factors rather than a genetic background may explain the

45 Discussion differences. The results obtained in our studies lend further support to this assumption, as the expansion of Vγ9Vδ2 T cells was long-term persistent.

Our study on Pontiac fever (Paper IV) disclosed novel information on the very early stage of infection. In this study, the exact time point of exposure and onset of disease could be determined, the first blood samples being taken on day 4-6 of exposure. At this point, the total γδ T-cell values did not differ significantly from values of a control group, whereas Vγ9Vδ2 T-cell numbers in peripheral blood were significantly lower (P<0.001). Such a phenomenon had not been observed previously. To my knowledge, the only data showing such a tendency originated from a report on experimental infection of two volunteers with P. falciparum [113]. In this experiment, decreased numbers of total γδ T cells in peripheral blood were reported, together with a decrease of total numbers of all lymphocytes up to day 12 after infection.

It seems reasonable to assume that the early decrease of the Vγ9Vδ2 T-cell levels in Pontiac fever reflects extravasal migration of the cells to the infected tissue. This possibility is in agreement with recent data on expression of the RANTES/macrophage inflammatory protein-1α/-1β receptor CCR5 by Vγ9Vδ2 T cells [129]. If extravasation can be confirmed to occur, the implication of the Vγ9Vδ2 T cells in the host response would be strengthened. The reason why we did not observe the phenomenon in tularemia, if at all present, is subjected to several possible explanations. First, the exact day of infection is less readily notified in tularemia than in Pontiac fever and our samples may not have been obtained early enough. Moreover, the development of tularemia, especially the ulceroglandular form, is slower than that of Pontiac fever. Second, the area of infection may be larger and a recruitment of Vγ9Vδ2 T cells more extensive in a respiratory than a cutaneous infection.

Extravasation from blood as part of a first line defense is an intriguing possibility. There is some data from both animal experiments and humans to

46 Discussion indicate that γδ T cells may have a role in the local infectious process. In mouse models, γδ T cells were found to promote granuloma formation and regulation of neutrophil activity in tissues infected with M. tuberculosis [130] or L. monocytogenes [43]. In leprosy patients, an accumulation of Vγ9Vδ2 T cells was reported in granulomatous skin reactions [131], thus supporting their role in the local immune defence.

An early depletion of Vγ9Vδ2 T cells from blood into infected tissue does not necessarily implicate that the cells do immediately start to proliferate. To proliferate, γδ T cells require the presence of exogenous T-cell growth factor. γδ T cells are themselves poor producers of IL-2 [58]. Beside activated CD4+ T cells, which are efficient producers of IL-2, also infected mononuclear phagocytes might be the source of IL-15 – another cytokine stimulating expansion of γδ T cells. Other cytokines, like IL-12, IL-7, IL-4 and IL-1 may also serve as growth factors in such a situation, although with lower efficiency [132-134]. Even before proliferation is commenced, a capability of Vγ9Vδ2 T cells to produce IFN-γ and TNF-α may give them an important role at an early stage of tissue infections [135-138]. More recently, a direct antimicrobial effect of Vγ9Vδ2 T cells mediated by cytotoxicity towards infected macrophages has been shown [139].

Little is known about differences in reactivity of primed and unprimed Vγ9Vδ2 T cells. A relatively high expression of CD45RO marker on the surface of Vγ9Vδ2 T cells, even increasing in the course of infection (Paper V; [113]), suggest a memory phenotype. Thanks to this, the Vγ9Vδ2 T cells may respond by activation early after attack by an invading pathogen. Such a role may be meaningful, as a first line of defence. Thus, an early migration of Vγ9Vδ2 T cells may be an important part of the defence.

47 Discussion

No increase of Vγ9Vδ2 T-cell levels in response to LVS vaccination

In spite of the presence of phosphoantigen in the vaccine strain F. tularensis LVS, at a titer similar to that of a wild strain of the species, no Vγ9Vδ2 T-cell expansion in peripheral blood was found after vaccination (paper I). The live tularemia vaccine was developed by attenuation of F. tularensis subsp. holarctica, which thereby lost much of its virulence for humans and also for rodents [140]. The loss in virulence was associated with a slower multiplication of bacteria in vivo, probably leading to a lower production of antigenic metabolic components. Obviously, this may be the reason for a less prominent activation of Vγ9Vδ2 T cells. Vaccination might be associated with a less efficient stimulation of CD4 T cells, a source of IL-2, required for Vγ9Vδ2 T- cell proliferation. Nonetheless, the local natural immunity in response to vaccination seems to be powerful enough to restrict the infectious process without the development of symptoms of disease. Obviously, more data on the immune response after vaccination as compared to natural infection are needed to explain the difference in the Vγ9Vδ2 T-cell response.

In a single vaccinee (not included in the study because of reluctance to give blood samples throughout the study), symptoms of disease appeared. When sampled at the time of symptoms, an elevated level of Vγ9Vδ2 T cells in the blood was demonstrated, exceeding 20% of all T cells (Kroca, unpublished data). Such a high value has not been reported in healthy control subjects from the region.

Role of Vγ9Vδ2 T cells in intracellular bacterial diseases

The Vγ9Vδ2 T-cell subset seems to be present only in primates. No subset with a pattern of recognition similar to that of the subset is found in mice. The lack of experimental models (except macaques) is a main obstacle when trying to

48 Discussion elucidate functions of the Vγ9Vδ2 T-cell subset. It cannot be taken for certain, moreover, that reliable conclusions may be drawn from murine models regarding γδ T cells in total. In murine listeriosis, several roles for γδ T cells have been suggested. Arguments for an importance in the host defence are their early appearance in the peritoneal cavity after intraperitoneal infection and the presence of higher bacterial loads at early stages of infection in γδ T-cell depleted animals than in littermate mice [42]. Beside IFN-γ production, production of TNF-α by γδ T cells has been demonstrated in murine listeriosis, a cytokine which promotes the production of IFN-γ by NK cells [141]. IFN-γ production is of special importance for the activation of infected macrophages at an early stage of infection.

Histopathological studies in mice have shown a regulatory role of γδ T cells. In γδ T-cell depleted animals, extensive liver lesions were observed together with high levels of neutrophils as compared to a granulomatous response in normal mice [142]. A down-regulating effect of γδ T cells on inflammation was further supported, when increased levels of IL-10 producing γδ T cells were observed in the phase of infection, when IFN-γ production was maximal [143]. Moreover, increased levels of the proinflammatory cytokines IL-6, IL-12 and IFN-γ have been found in γδ T-cell deficient mice [144]. Another possible role is their cytotoxic activity towards low-density activated macrophages, inducing cell death and down-modulating inflammation during terminal stages of infection [145].

Altogether, the role of γδ T cells seems to be complex. They may have different effects, including production of the proinflammatory cytokines and down-regulation of inflammation, during various phases of the process of intracellular infections.

49 Discussion

Our data on the cytokine response of γδ T cells from subjects undergoing tularemia or Pontiac fever (paper III and IV) may be in line with the presence of a dual role of the subset. Peripheral blood lymphocytes were incubated for 4 h in vitro in the presence of a T-cell stimulant (phorbol myristate acetate) and after monensin-induced blockage of cytokine secretion, the intracellular accumulation of cytokines by γδ T cells was assayed by flow cytometry. In tularemia, the ability to produce the proinflammatory cytokine TNF-α was significantly decreased among γδ T cells in blood samples obtained as early as during the first week of infection. This down-modulation was preserved over the period of maximal expansion (up to 40 days) and restored nearly to normal levels at 3 months. The decrease in IFN-γ production was not significant and appeared slightly later – at a time when a massive expansion of γδ T cells had started. In Pontiac fever, a significant increase in levels of both IFN-γ and TNF-α producing γδ T cells was observed at 14-16 days after infection followed by a substantial decrease at 5-7 weeks. At 6 months, the level of TNF-α producing γδ T cells was restored, while the ability of γδ T cells to produce IFN-γ even exceeded control values. Together, these data suggest that Vγ9Vδ2 T cells might have a dual role in intracellular bacterial diseases, first serving as a source of IFN-γ and TNF-α, important as early mediators of host protection, thereafter being subjected to down-regulation, in order to avoid exaggeration of inflammation and tissue necrosis.

Vγ9Vδ2 T cells are known to express several types of regulatory molecules, including the CD94/NKG2 complex [62]. Suggested roles for this complex in Vγ9Vδ2 T cells are inhibition of cytotoxic activity [146], down-modulation of activation after stimulation with phosphoantigens [147], and inhibition of proliferation and cytokine secretion by these cells [148]. Upon stimulation with phosphoantigens, a translocation of CD94 to the cell surface has been observed [149], which may increase the threshold for activation of Vγ9Vδ2 T cells by

50 Discussion antigenic stimulation. In our study (Paper IV), an increased expression of CD94 on Vγ9Vδ2 T cells was observed during the course of Pontiac fever, temporarily correlating with the level of these cells in peripheral blood and a down- regulation of their cytokine production. Our observations support the presence of a role of Vγ9Vδ2 T cells in the host response to intracellular infections, including a regulatory function during the course of infection.

In a recently published study, a transient functional deficiency of Vγ9Vδ2 T cells in children with tuberculosis was observed [150]. During acute phase of disease, these cells retained their proliferative capability in response to in vitro stimulation with phosphoantigens, while their IFN-γ production and granulysin expression were abolished. After about 12 months, restoration of these parameters was observed in response to successful chemotherapy. These observations may be in accordance with our results obtained in tularemia and Pontiac fever, taking into account that recovery in tuberculosis is a much slower process.

Implications on the pathophysiology of Pontiac fever Except for the study presented in paper IV, legionellosis seems not to have been investigated with regard to γδ T cells. Actually, the study was a test of the assumption that facultative intracellular bacterial diseases are more generally associated with Vγ9Vδ2 T-cell expansion. Due to the transient nature of Pontiac fever, it has been questioned whether the disease is due to infection or to the exposure to dead bacteria or toxins inhaled from a contaminated whirlpool or other sources. The results presented in paper IV support the former alternative. By inference, the long-lasting increase of a T-cell subset in Pontiac fever is more probably an equivalent to what happens in several other intracellular infections rather than a phenomenon that occurs only as a result of exposure to some yet

51 Discussion not defined bacterial toxin. The Vγ9Vδ2 T-cell stimulating phosphoantigens may, in the case of bacterial infection, be metabolic products of isoprenoid biosynthesis [123, 151, 152] and a metabolic activity by growing bacteria would be more likely than the presence of dead bacteria or a bacterial toxin as a prerequisite for the expansion of Vγ9Vδ2 T cells.

Implications of a cross reactivity in vitro by Vγ9Vδ2 T cells from subjects primed by tularemia or Pontiac fever, respectively When incubated in vitro, both L. micdadei, implicated as a causative agent of the present outbreak of Pontiac fever, and F. tularensis, were shown to produce phosphoantigen (paper I and V). Aiming to compare the in vitro response of subjects previously undergoing tularemia or Pontiac fever to antigen from the homologous vs the heterologous agent, we prepared a crude extract of phosphoantigen from each organism. Both categories of patients had undergone their disease one year previously and they still had an increased proportion of Vγ9Vδ2 T cells. They responded at a quite similar degree to the heterologous and homologous antigens. The responses were not the result of stimulation with an exceedingly high amount of material.

Antigens of different origin have been reported to stimulate Vγ9Vδ2 T cells, including the synthetic molecule monoethylphosphate [35], four mycobacterial molecules [34], isopentenyl pyrophosphate and related prenyl pyrophosphate derivatives [153], as well as additional natural and synthetic molecules like diphosphoglyceric acid, xylose- and ribose-1-phosphate, glycerol-3-phosphate and dimethylallylpyrophosphate [154]. More recently, alkylamines from microbes, edible plants and tea have been shown to be Vγ9Vδ2 T-cell antigens [155]. Despite their structural differences, these antigens seem to be recognized by the same cells. It has not been possible to isolate clones of Vγ9Vδ2 T cells specific to discrete antigen molecules [154]. The broad reactivity of these cells

52 Discussion stands in contrast to their capability of discriminating structurally related substances. Actually, minor changes in phosphoantigen structure may result in loss of stimulatory properties [156]. Recognition by Vγ9Vδ2 T cells seems to depend on antigen pattern rather than discrete chemical moieties [157, 158]. Dephosphorylation of TCR-bound phosphoantigen is necessary for proper stimulation and molecules resistant to pyrophosphate hydrolysis may specifically but reversibly inhibit activation of Vγ9Vδ2 T cells [158]. Altogether, it is not surprising that Vγ9Vδ2 T cells from patients infected with unrelated agents may respond to stimulation with phosphoantigens from these two agents in cross-reactive manner (Paper V). The increased expression of the CD45RO marker on these cells suggested a memory state. The ability to recognize invading pathogens at an early stage of infection before acquired immunity is established may be of vital importance for the host.

In the context, an early extravasation of 80% of Vγ9Vδ2 T cells from the circulation during the first days of onset of Pontiac fever may be of special significance. A broad reactivity of Vγ9Vδ2 T cells to evolutionary distinct intracellular pathogens seems to be a feature of the human host immune response. These cells will probably be ready for immediate in situ production of IFN-γ and TNF-α. For proliferation, however, they might need additional stimulus, like IL-2 support by CD4+ T cells or another T-cell growth factor.

53 Conclusions

CONCLUSIONS

1. A significant expansion of peripheral blood Vγ9Vδ2 T cells occurs consistently in patients undergoing tularemia or Pontiac fever. After a week´s delay, levels increase rapidly, peak levels being reached within a month. Thereafter, values decrease slowly during several months. Vaccination with the live tularemia vaccine, on the other hand, is not associated with an increase of blood levels of Vγ9Vδ T cells.

2. Francisella tularensis and Legionella micdadei both contain Vγ9Vδ2 T-cell stimulating non-peptidic phosphorylated products, so called phosphoantigens.

3. Vγ9Vδ2 T cells from patients undergoing tularemia or Pontiac fever expand in vitro to a similar extent irrespective of whether they are stimulated with the homologous or heterologous phosphoantigen. Such a broad reactivity of memory cells may endow the host with a capability of surveying a wide range of microbial intruders.

4. Under in vitro conditions allowing a vigorous expansion of Vγ9Vδ2 T cells, no expansion of γδ Τ cells occurs in response to the Cpn60 or DnaK homologs of Francisella tularensis.

5. During the course of both tularemia and Pontiac fever, the capacity of Vγ9Vδ2 T cells to produce TNF-α and IFN-γ is transiently decreased, suggesting the presence of modulation of the inflammatory response.

Based on our observations in tularemia and Pontiac fever and also from literature data on other intracellular infections, the expansion of Vγ9Vδ2 T cells seems to be a feature shared by various human intracellular bacterial diseases.

54 Acknowledgements

ACKNOWLEDGEMENTS

I would like to thank all the people, who contributed to this work in any style and format. Over those years there were many of them I have met and work with. If some of you are not listed here, it is caused by my weakening memory and not because I am angry on anyone.

Especially I am grateful to the following people;

My supervisor, Arne Tärnvik for initiating all of this work (without his conviction and reliance I would never started with it) and for sharing his endless knowledge on infections and tularemia in particular. Also his contribution to scientific writing cannot be appreciated enough.

Anders Sjöstedt for always being a few steps ahead, always having new ideas and suggestions, forcing me to continue further.

Ales Macela for always pushing me forward, in gentle style but with really irresistible power.

Gunnar Sandström and Karin Hjalmarsson for supporting my stay at FOI and allowing me to perform most of my experimental work there.

Thorsten Johansson for introducing me to the mysterious world of flow cytometry and for being such a good friend.

Åke Forsberg, Anders Norqvist and Mats Forsman for helpful scientific discussions at FOI and their suggestive ideas.

Ingela Göransson for the help in taking blood samples from almost everyone around.

55 Acknowledgements

Kerstin Kuoppa, Ulla Eriksson and Anna Macellaro for their kind help in laboratory work every time I needed some.

Pär Larsson for his efforts when helping me with searches in the Francisella genome.

Maiken Karlsson for excellent secretary help during my work at FOI.

Igor Golovliov for sharing his experience from experimental work on Francisella and from foreigners’ live in Sweden.

My colleagues from the Department of Clinical Microbiology, Mats Ericsson, Anders Johansson and Stephan Stenmark for their contributions to this work and their encouragement.

Gunborg Eriksson for her great administrative support and intensive care any time it was needed and for the book revision and preparation.

All these nameless physicians in Sweden and Finland who provided us with blood samples from tularemia and legionellosis patients, and all patients themselves.

All the other people I met in Sweden throughout all these years and who made this country to be nearly my second home.

Our collaborators in France, Jean-Jacques Fournié and members of his group in Toulouse, Yannick Poquet, Christian Belmant and Eric Champagne for the help with preparation and characterization of phosphoantigens from Francisella strains, and also Marc Bonneville and Marie-Alix Peyrat from Nantes for the experiments on the gamma/delta T-cell clone.

All colleagues in my home institute in Czech Republic, who had to take over my part of work in the times I was working on my thesis abroad.

56 Acknowledgements

And last but not least I would like to express my special thanks to my family, in particular to my wife Zuzana and my daughters Alena and Eliska for their deep understanding and great support they provided to me for all this time. Without this help it would never be possible to do all of the work being 2000 km away from home.

These studies were supported by grants from the Medical Faculty of Umeå University, Västerbottens läns landsting, the Swedish National Defense Research Agency, and the Swedish Medical Research Council.

57 References

REFERENCES

1. Kuby, J., Immunology. Second edition ed. 1994, New York: W.H.Freeman and Company. 2. Davis, M.M. and P.J. Bjorkman, T-cell antigen receptor genes and T-cell recognition. Nature, 1988. 334(6181): p. 395-402. 3. McVay, L.D. and S.R. Carding, Generation of human gammadelta T-cell repertoires. Crit Rev Immunol, 1999. 19(5-6): p. 431-60. 4. Carding, S.R. and P.J. Egan, Gammadelta T cells: functional plasticity and heterogeneity. Nat Rev Immunol, 2002. 2(5): p. 336-45. 5. Steele, C.R., D.E. Oppenheim, and A.C. Hayday, Gamma(delta) T cells: non-classical ligands for non-classical cells. Curr Biol, 2000. 10(7): p. R282-5. 6. Husmann, L.A. and M.J. Bevan, Cooperation between helper T cells and cytotoxic T lymphocyte precursors. Ann N Y Acad Sci, 1988. 532: p. 158-69. 7. Mosmann, T.R. and R.L. Coffman, TH1 and TH2 cells: different patterns of lymphokine secretion lead to different functional properties. Annu Rev Immunol, 1989. 7: p. 145-73. 8. Mosmann, T.R. and S. Sad, The expanding universe of T-cell subsets: Th1, Th2 and more. Immunol Today, 1996. 17(3): p. 138-46. 9. Groux, H., An overview of regulatory T cells. Microbes Infect, 2001. 3(11): p. 883-9. 10. Ohno, S., MHC evolution and development of a recognition system. Ann N Y Acad Sci, 1994. 712: p. 13-9. 11. Beckman, E.M., et al., Recognition of a lipid antigen by CD1-restricted alpha beta+ T cells. Nature, 1994. 372(6507): p. 691-4. 12. Zeng, Z., et al., Crystal structure of mouse CD1: An MHC-like fold with a large hydrophobic binding groove. Science, 1997. 277(5324): p. 339-45. 13. Calabi, F., et al., Two classes of CD1 genes. Eur J Immunol, 1989. 19(2): p. 285-92. 14. Gumperz, J.E. and M.B. Brenner, CD1-specific T cells in microbial immunity. Curr Opin Immunol, 2001. 13(4): p. 471-8. 15. Moody, D.B., et al., CD1c-mediated T-cell recognition of isoprenoid glycolipids in Mycobacterium tuberculosis infection. Nature, 2000. 404(6780): p. 884-8. 16. Porcelli, S., et al., Recognition of cluster of differentiation 1 antigens by human CD4- CD8-cytolytic T lymphocytes. Nature, 1989. 341(6241): p. 447-50. 17. Rosat, J.P., et al., CD1-restricted microbial lipid antigen-specific recognition found in the CD8+ alpha beta T cell pool. J Immunol, 1999. 162(1): p. 366-71. 18. Sieling, P.A., et al., Evidence for human CD4+ T cells in the CD1-restricted repertoire: derivation of mycobacteria-reactive T cells from leprosy lesions. J Immunol, 2000. 164(9): p. 4790-6. 19. Spada, F.M., et al., Self-recognition of CD1 by gamma/delta T cells: implications for innate immunity. J Exp Med, 2000. 191(6): p. 937-48. 20. Kawano, T., et al., CD1d-restricted and TCR-mediated activation of valpha14 NKT cells by glycosylceramides. Science, 1997. 278(5343): p. 1626-9.

58 References

21. Ochoa, M.T., et al., T-cell release of granulysin contributes to host defense in leprosy. Nat Med, 2001. 7(2): p. 174-9. 22. Smith, S.M. and H.M. Dockrell, Role of CD8 T cells in mycobacterial infections. Immunol Cell Biol, 2000. 78(4): p. 325-33. 23. Stenger, S., et al., Differential effects of cytolytic T cell subsets on intracellular infection. Science, 1997. 276(5319): p. 1684-7. 24. Jayawardena-Wolf, J. and A. Bendelac, CD1 and lipid antigens: intracellular pathways for antigen presentation. Curr Opin Immunol, 2001. 13(1): p. 109-13. 25. Groh, V., et al., Broad tumor-associated expression and recognition by tumor-derived gamma delta T cells of MICA and MICB. Proc Natl Acad Sci U S A, 1999. 96(12): p. 6879-84. 26. Groh, V., et al., Recognition of stress-induced MHC molecules by intestinal epithelial gammadelta T cells. Science, 1998. 279(5357): p. 1737-40. 27. Das, H., et al., MICA engagement by human Vgamma2Vdelta2 T cells enhances their antigen-dependent effector function. Immunity, 2001. 15(1): p. 83-93. 28. Boullier, S., et al., V delta 1 T cells expanded in the blood throughout HIV infection display a cytotoxic activity and are primed for TNF-alpha and IFN-gamma production but are not selected in lymph nodes. J Immunol, 1997. 159(7): p. 3629-37. 29. Dechanet, J., et al., Implication of gammadelta T cells in the human immune response to cytomegalovirus. J Clin Invest, 1999. 103(10): p. 1437-49. 30. Johnson, R.M., et al., A murine CD4-, CD8- T cell receptor-gamma delta T lymphocyte clone specific for herpes simplex virus glycoprotein I. J Immunol, 1992. 148(4): p. 983- 8. 31. Parker, C.M., et al., Evidence for extrathymic changes in the T cell receptor gamma/delta repertoire. J Exp Med, 1990. 171(5): p. 1597-612. 32. Hinz, T., et al., Identification of the complete expressed human TCR V gamma repertoire by flow cytometry. Int Immunol, 1997. 9(8): p. 1065-72. 33. Pfeffer, K., et al., Primary responses of human T cells to mycobacteria: a frequent set of gamma/delta T cells are stimulated by protease-resistant ligands. Eur J Immunol, 1990. 20(5): p. 1175-9. 34. Constant, P., et al., Stimulation of human gamma delta T cells by nonpeptidic mycobacterial ligands. Science, 1994. 264(5156): p. 267-70. 35. Tanaka, Y., et al., Nonpeptide ligands for human gamma delta T cells. Proc Natl Acad Sci U S A, 1994. 91(17): p. 8175-9. 36. Schoel, B., S. Sprenger, and S.H. Kaufmann, Phosphate is essential for stimulation of V gamma 9V delta 2 T lymphocytes by mycobacterial low molecular weight ligand. Eur J Immunol, 1994. 24(8): p. 1886-92. 37. Fournie, J.J. and M. Bonneville, Stimulation of gamma delta T cells by phosphoantigens. Res Immunol, 1996. 147(5): p. 338-47. 38. Inoue, T., et al., Early appearing gamma/delta-bearing T cells during infection with Calmette Guerin bacillus. J Immunol, 1991. 146(8): p. 2754-62.

59 References

39. Kaufmann, S.H. and C.H. Ladel, Role of T cell subsets in immunity against intracellular bacteria: experimental infections of knock-out mice with Listeria monocytogenes and Mycobacterium bovis BCG. Immunobiology, 1994. 191(4-5): p. 509-19. 40. Ladel, C.H., et al., Contribution of alpha/beta and gamma/delta T lymphocytes to immunity against Mycobacterium bovis bacillus Calmette Guerin: studies with T cell receptor-deficient mutant mice. Eur J Immunol, 1995. 25(3): p. 838-46. 41. Nabeshima, S., et al., Infection of Mycobacterium bovis bacillus Calmette-Guerin in antibody-mediated gamma delta T-cell-depleted mice. Immunology, 1995. 84(2): p. 317- 21. 42. Hiromatsu, K., et al., A protective role of gamma/delta T cells in primary infection with Listeria monocytogenes in mice. J Exp Med, 1992. 175(1): p. 49-56. 43. Mombaerts, P., et al., Different roles of alpha beta and gamma delta T cells in immunity against an intracellular bacterial pathogen. Nature, 1993. 365(6441): p. 53-6. 44. Yamamoto, S., et al., Listeria monocytogenes-induced gamma interferon secretion by intestinal intraepithelial gamma/delta T lymphocytes. Infect Immun, 1993. 61(5): p. 2154-61. 45. Ferrick, D.A., et al., Differential production of interferon-gamma and interleukin-4 in response to Th1- and Th2-stimulating pathogens by gamma delta T cells in vivo. Nature, 1995. 373(6511): p. 255-7. 46. Berguer, R. and D.A. Ferrick, Differential production of intracellular gamma interferon in alpha beta and gamma delta T-cell subpopulations in response to peritonitis. Infect Immun, 1995. 63(12): p. 4957-8. 47. Nishimura, H., et al., The role of gamma delta T cells in priming macrophages to produce tumor necrosis factor-alpha. Eur J Immunol, 1995. 25(5): p. 1465-8. 48. Shen, Y., et al., Adaptive immune response of Vgamma2Vdelta2+ T cells during mycobacterial infections. Science, 2002. 295(5563): p. 2255-8. 49. Flanagan, B.F., et al., T cell receptor junctional regions of V gamma 9+/V delta 2+ T cell clones in relation to non-MHC restricted cytotoxic activity. Mol Immunol, 1993. 30(7): p. 659-67. 50. Barnes, P.F., et al., Gamma delta T lymphocytes in human tuberculosis. J Infect Dis, 1992. 165(3): p. 506-12. 51. Barnes, P.F., et al., Patterns of cytokine production by mycobacterium-reactive human T-cell clones. Infect Immun, 1993. 61(1): p. 197-203. 52. Zocchi, M.R., M. Ferrarini, and C. Rugarli, Selective lysis of the autologous tumor by delta TCS1+ gamma/delta+ tumor-infiltrating lymphocytes from human lung carcinomas. Eur J Immunol, 1990. 20(12): p. 2685-9. 53. Mavaddat, N., et al., An analysis of the relationship between gamma delta T cell receptor V gene usage and non-major histocompatibility complex-restricted cytotoxicity. Immunol Cell Biol, 1993. 71(Pt 1): p. 27-37. 54. Bank, I., et al., V delta 2+ gamma delta T lymphocytes are cytotoxic to the MCF 7 breast carcinoma cell line and can be detected among the T cells that infiltrate breast tumors. Clin Immunol Immunopathol, 1993. 67(1): p. 17-24.

60 References

55. Bukowski, J.F., C.T. Morita, and M.B. Brenner, Recognition and destruction of virus- infected cells by human gamma delta CTL. J Immunol, 1994. 153(11): p. 5133-40. 56. De Paoli, P., et al., Phenotypic profile and functional characteristics of human gamma and delta T cells during acute toxoplasmosis. J Clin Microbiol, 1992. 30(3): p. 729-31. 57. Scalise, F., et al., Lymphocytes bearing the gamma delta T-cell receptor in acute toxoplasmosis. Immunology, 1992. 76(4): p. 668-70. 58. Tsukaguchi, K., K.N. Balaji, and W.H. Boom, CD4+ alpha beta T cell and gamma delta T cell responses to Mycobacterium tuberculosis. Similarities and differences in Ag recognition, cytotoxic effector function, and cytokine production. J Immunol, 1995. 154(4): p. 1780-96. 59. Komano, H., et al., Homeostatic regulation of intestinal epithelia by intraepithelial gamma delta T cells. Proc Natl Acad Sci U S A, 1995. 92(13): p. 6147-51. 60. Giacomelli, R., et al., Increase of circulating gamma/delta T lymphocytes in the peripheral blood of patients affected by active inflammatory bowel disease. Clin Exp Immunol, 1994. 98(1): p. 83-8. 61. Mingari, M.C., et al., Cytolytic T lymphocytes displaying natural killer (NK)-like activity: expression of NK-related functional receptors for HLA class I molecules (p58 and CD94) and inhibitory effect on the TCR-mediated target cell lysis or lymphokine production. Int Immunol, 1995. 7(4): p. 697-703. 62. Battistini, L., et al., Phenotypic and cytokine analysis of human peripheral blood gamma delta T cells expressing NK cell receptors. J Immunol, 1997. 159(8): p. 3723-30. 63. Hara, T., et al., Predominant activation and expansion of V gamma 9-bearing gamma delta T cells in vivo as well as in vitro in Salmonella infection. J Clin Invest, 1992. 90(1): p. 204-10. 64. Raziuddin, S., et al., Gamma delta T cells and the immune response in visceral leishmaniasis. Eur J Immunol, 1992. 22(5): p. 1143-8. 65. Ho, M., et al., Polyclonal expansion of peripheral gamma delta T cells in human Plasmodium falciparum malaria. Infect Immun, 1994. 62(3): p. 855-62. 66. Bertotto, A., et al., Lymphocytes bearing the gamma delta T cell receptor in acute Brucella melitensis infection. Eur J Immunol, 1993. 23(5): p. 1177-80. 67. Bertotto, A., et al., Peripheral blood gamma delta T cells in human listeriosis. Acta Paediatr, 1995. 84(12): p. 1434-5. 68. De Paoli, P., et al., A subset of gamma delta lymphocytes is increased during HIV-1 infection. Clin Exp Immunol, 1991. 83(2): p. 187-91. 69. De Paoli, P., et al., Gamma delta T cell receptor-bearing lymphocytes during Epstein- Barr virus infection. J Infect Dis, 1990. 161(5): p. 1013-6. 70. Kozbor, D., et al., Analysis of gamma delta+ T cells in peripheral blood of children with perinatal human immunodeficiency virus (HIV) infection. J Clin Immunol, 1993. 13(3): p. 193-203. 71. De Maria, A., et al., Selective increase of a subset of T cell receptor gamma delta T lymphocytes in the peripheral blood of patients with human immunodeficiency virus type 1 infection [published erratum appears in J Infect Dis 1992 Oct;166(4):950]. J Infect Dis, 1992. 165(5): p. 917-9.

61 References

72. Dieli, F., et al., Ligand-specific alphabeta and gammadelta T cell responses in childhood tuberculosis. J Infect Dis, 2000. 181(1): p. 294-301. 73. Ueta, C., et al., Increase of gamma/delta T cells in hospital workers who are in close contact with tuberculosis patients. Infect Immun, 1994. 62(12): p. 5434-41. 74. Hviid, L., et al., The gamma/delta T-cell response to Plasmodium falciparum malaria in a population in which malaria is endemic. Infect Immun, 1996. 64(10): p. 4359-62. 75. Anthony, L.D., R.D. Burke, and F.E. Nano, Growth of Francisella spp. in rodent macrophages. Infect Immun, 1991. 59(9): p. 3291-6. 76. Fortier, A.H., et al., Activation of macrophages for destruction of Francisella tularensis: identification of cytokines, effector cells, and effector molecules. Infect Immun, 1992. 60(3): p. 817-25. 77. Sjöstedt, A., et al., Various membrane proteins of Francisella tularensis induce interferon-gamma production in both CD4+ and CD8+ T cells of primed humans. Immunology, 1992. 76: p. 584-592. 78. Surcel, H.M., Diversity of Francisella tularensis antigens recognized by human T lymphocytes. Infect Immun, 1990. 58(8): p. 2664-8. 79. Ericsson, M., et al., Persistence of cell-mediated immunity and decline of humoral immunity to the intracellular bacterium Francisella tularensis 25 years after natural infection. J Infect Dis, 1994. 170(1): p. 110-4. 80. Sumida, T., et al., Predominant expansion of V gamma 9/V delta 2 T cells in a tularemia patient. Infect Immun, 1992. 60(6): p. 2554-8. 81. Horwitz, M.A., Cell-mediated immunity in Legionnaires' disease. J Clin Invest, 1983. 71(6): p. 1686-97. 82. Esin, S., et al., Different percentages of peripheral blood gamma delta + T cells in healthy individuals from different areas of the world. Scand J Immunol, 1996. 43(5): p. 593-6. 83. Mascher, B., P. Schlenke, and M. Seyfarth, Expression and kinetics of cytokines determined by intracellular staining using flow cytometry. J Immunol Methods, 1999. 223(1): p. 115-21. 84. Behr, C., et al., Plasmodium falciparum stimuli for human gammadelta T cells are related to phosphorylated antigens of mycobacteria. Infect Immun, 1996. 64(8): p. 2892-6. 85. Ericsson, M., et al., Characterization of the nucleotide sequence of the groE operon encoding heat shock proteins chaperone-60 and -10 of Francisella tularensis and determination of the T-cell response to the proteins in individuals vaccinated with F. tularensis. Infect Immun, 1997. 65(5): p. 1824-9. 86. Chamberlain, R.E., Evaluation of live tularemia vaccine prepared in a chemicaly defined medium. Appl Microbiol, 1965. 13: p. 232-5. 87. Ito, M., et al., Increased proportions of peripheral blood gamma delta T cells in patients with pulmonary tuberculosis. Chest, 1992. 102(1): p. 195-7. 88. Poquet, Y., et al., A novel nucleotide-containing antigen for human blood gamma delta T lymphocytes. Eur J Immunol, 1996. 26(10): p. 2344-9.

62 References

89. Havlir, D.V., et al., Selective expansion of human gamma delta T cells by monocytes infected with live Mycobacterium tuberculosis. J Clin Invest, 1991. 87(2): p. 729-33. 90. O'Brien, R., et al., Stimulation of a major subset of lymphocytes expressing T cell receptor gamma delta by an antigen derived from Mycobacterium tuberculosis. Cell, 1989. 57(4): p. 667-74. 91. O'Brien, R. and W. Born, Heat shock proteins as antigens for gamma delta T cells. Semin Immunol, 1991. 3(2): p. 81-7. 92. Fisch, P., et al., Recognition by human V gamma 9/V delta 2 T cells of a GroEL homolog on Daudi Burkitt's lymphoma cells. Science, 1990. 250(4985): p. 1269-73. 93. Kaur, I., et al., Human peripheral gamma delta T cells recognize hsp60 molecules on Daudi Burkitt's lymphoma cells. J Immunol, 1993. 150(5): p. 2046-55. 94. Ericsson, M., et al., Increased synthesis of DnaK, GroEL, and GroES homologs by Francisella tularensis LVS in response to heat and hydrogen peroxide. Infect Immun, 1994. 62(1): p. 178-83. 95. Kabelitz, D., et al., A large fraction of human peripheral blood gamma/delta + T cells is activated by Mycobacterium tuberculosis but not by its 65-kD heat shock protein. J Exp Med, 1990. 171(3): p. 667-79. 96. Gotz, H.M., et al., A whirlpool associated outbreak of Pontiac fever at a hotel in Northern Sweden. Epidemiol Infect, 2001. 126(2): p. 241-7. 97. Luttichau, H.R., et al., An outbreak of Pontiac fever among children following use of a whirlpool. Clin Infect Dis, 1998. 26(6): p. 1374-8. 98. Fallon, R.J. and T.J. Rowbotham, Microbiological investigations into an outbreak of Pontiac fever due to Legionella micdadei associated with use of a whirlpool. J Clin Pathol, 1990. 43(6): p. 479-83. 99. Ferrini, S., et al., T cell clones expressing the natural killer cell-related p58 receptor molecule display heterogeneity in phenotypic properties and p58 function. Eur J Immunol, 1994. 24(10): p. 2294-8. 100. Schneider, T., et al., The number and proportion of Vgamma9 Vdelta2 T cells rise significantly in the peripheral blood of patients after the onset of acute Coxiella burnetii infection. Clin Infect Dis, 1997. 24(2): p. 261-4. 101. Jouen-Beades, F., et al., In vivo and in vitro activation and expansion of gammadelta T cells during Listeria monocytogenes infection in humans. Infect Immun, 1997. 65(10): p. 4267-72. 102. Caldwell, C.W., et al., Lymphocytosis of gamma/delta T cells in human ehrlichiosis [see comments]. Am J Clin Pathol, 1995. 103(6): p. 761-6. 103. Behr-Perst, S.I., et al., Phenotypically activated gammadelta T lymphocytes in the peripheral blood of patients with tuberculosis. J Infect Dis, 1999. 180(1): p. 141-9. 104. Fajac, I., et al., Lymphocytes infiltrating normal human lung and lung carcinomas rarely express gamma delta T cell antigen receptors. Clin Exp Immunol, 1992. 87(1): p. 127- 31. 105. Poccia, F., et al., In vivo gammadelta T cell priming to mycobacterial antigens by primary Mycobacterium tuberculosis infection and exposure to nonpeptidic ligands. Mol Med, 1999. 5(7): p. 471-6.

63 References

106. Li, B., et al., Disease-specific changes in gammadelta T cell repertoire and function in patients with pulmonary tuberculosis. J Immunol, 1996. 157(9): p. 4222-9. 107. Li, B., et al., Involvement of the Fas/Fas ligand pathway in activation-induced cell death of mycobacteria-reactive human gamma delta T cells: a mechanism for the loss of gamma delta T cells in patients with pulmonary tuberculosis. J Immunol, 1998. 161(3): p. 1558-67. 108. Raziuddin, S., et al., Gamma delta T lymphocytes and proinflammatory cytokines in bacterial meningitis. J Allergy Clin Immunol, 1994. 93(4): p. 793-8. 109. Bertotto, A., et al., Gamma delta T cells are decreased in the blood of children with Bordetella pertussis infection. Acta Paediatr, 1997. 86(1): p. 114-5. 110. Ho, M., et al., Increased gamma delta T cells in acute Plasmodium falciparum malaria. Immunol Lett, 1990. 25(1-3): p. 139-41. 111. Roussilhon, C., et al., Human TcR gamma delta+ lymphocyte response on primary exposure to Plasmodium falciparum. Clin Exp Immunol, 1994. 95(1): p. 91-7. 112. Perera, M.K., et al., Transient increase in circulating gamma/delta T cells during Plasmodium vivax malarial paroxysms. J Exp Med, 1994. 179(1): p. 311-5. 113. Rzepczyk, C.M., et al., Experimental human Plasmodium falciparum infections: longitudinal analysis of lymphocyte responses with particular reference to gamma delta T cells. Scand J Immunol, 1996. 43(2): p. 219-27. 114. Waterfall, M., A. Black, and E. Riley, Gammadelta+ T cells preferentially respond to live rather than killed malaria parasites. Infect Immun, 1998. 66(5): p. 2393-8. 115. Hensmann, M. and D. Kwiatkowski, Cellular basis of early cytokine response to Plasmodium falciparum. Infect Immun, 2001. 69(4): p. 2364-71. 116. Jason, J., et al., Natural T, gammadelta, and NK cells in mycobacterial, Salmonella, and human immunodeficiency virus infections. J Infect Dis, 2000. 182(2): p. 474-81. 117. Hara, T., et al., Human V delta 2+ gamma delta T-cell tolerance to foreign antigens of Toxoplasma gondii. Proc Natl Acad Sci U S A, 1996. 93(10): p. 5136-40. 118. Munk, M.E., et al., Onchocerca volvulus provides ligands for the stimulation of human gamma/delta T lymphocytes expressing V delta 1 chains. J Infect Dis, 1993. 168(5): p. 1241-7. 119. Haregewoin, A., et al., Human gamma delta+ T cells respond to mycobacterial heat- shock protein. Nature, 1989. 340(6231): p. 309-12. 120. O'Brien, R.L., et al., Stimulation of a major subset of lymphocytes expressing T cell receptor gamma delta by an antigen derived from Mycobacterium tuberculosis. Cell, 1989. 57(4): p. 667-74. 121. Kabelitz, D., et al., The primary response of human gamma/delta + T cells to Mycobacterium tuberculosis is restricted to V gamma 9-bearing cells. J Exp Med, 1991. 173(6): p. 1331-8. 122. Panchamoorthy, G., et al., A predominance of the T cell receptor V gamma 2/V delta 2 subset in human mycobacteria-responsive T cells suggests germline gene encoded recognition. J Immunol, 1991. 147(10): p. 3360-9. 123. Belmant, C., et al., 3-Formyl-1-butyl pyrophosphate A novel mycobacterial metabolite- activating human gammadelta T cells. J Biol Chem, 1999. 274(45): p. 32079-84.

64 References

124. Boom, W.H., et al., Characterization of a 10- to 14-kilodalton protease-sensitive Mycobacterium tuberculosis H37Ra antigen that stimulates human gamma delta T cells. Infect Immun, 1994. 62(12): p. 5511-8. 125. Pechhold, K., et al., Primary activation of V gamma 9-expressing gamma delta T cells by Mycobacterium tuberculosis. Requirement for Th1-type CD4 T cell help and inhibition by IL-10. J Immunol, 1994. 152(10): p. 4984-92. 126. Elloso, M.M., et al., Human gamma delta T cell subset-proliferative response to malarial antigen in vitro depends on CD4+ T cells or cytokines that signal through components of the IL-2R. J Immunol, 1996. 157(5): p. 2096-102. 127. Tärnvik, A., G. Sandström, and S. Löfgren, Time of lymphocyte response after onset of tularemia and after tularemia vaccination. J Clin Microbiol, 1979. 10(6): p. 854-60. 128. Hoft, D.F., R.M. Brown, and S.T. Roodman, Bacille Calmette-Guerin vaccination enhances human gamma delta T cell responsiveness to mycobacteria suggestive of a memory-like phenotype. J Immunol, 1998. 161(2): p. 1045-54. 129. Glatzel, A., et al., Patterns of chemokine receptor expression on peripheral blood gamma delta T lymphocytes: strong expression of CCR5 is a selective feature of V delta 2/V gamma 9 gamma delta T cells. J Immunol, 2002. 168(10): p. 4920-9. 130. D'Souza, C.D., et al., An anti-inflammatory role for gamma delta T lymphocytes in acquired immunity to Mycobacterium tuberculosis. J Immunol, 1997. 158(3): p. 1217- 21. 131. Modlin, R.L., et al., Lymphocytes bearing antigen-specific gamma delta T-cell receptors accumulate in human infectious disease lesions. Nature, 1989. 339(6225): p. 544-8. 132. Skeen, M.J. and H.K. Ziegler, Intercellular interactions and cytokine responsiveness of peritoneal alpha/beta and gamma/delta T cells from Listeria-infected mice: synergistic effects of interleukin 1 and 7 on gamma/delta T cells. J Exp Med, 1993. 178(3): p. 985- 96. 133. Skeen, M.J. and H.K. Ziegler, Induction of murine peritoneal gamma/delta T cells and their role in resistance to bacterial infection. J Exp Med, 1993. 178(3): p. 971-84. 134. Garcia, V.E., et al., IL-15 enhances the response of human gamma delta T cells to nonpeptide [correction of nonpetide] microbial antigens. J Immunol, 1998. 160(9): p. 4322-9. 135. Goodier, M.R., et al., Cytokine profiles for human V gamma 9+ T cells stimulated by Plasmodium falciparum. Parasite Immunol, 1995. 17(8): p. 413-23. 136. Sireci, G., et al., Patterns of phosphoantigen stimulation of human Vgamma9/Vdelta2 T cell clones include Th0 cytokines. Hum Immunol, 1997. 58(2): p. 70-82. 137. Garcia, V.E., et al., Single-cell cytokine analysis of gamma delta T cell responses to nonpeptide mycobacterial antigens. J Immunol, 1997. 159(3): p. 1328-35. 138. Troye-Blomberg, M., et al., Human gamma delta T cells that inhibit the in vitro growth of the asexual blood stages of the Plasmodium falciparum parasite express cytolytic and proinflammatory molecules. Scand J Immunol, 1999. 50(6): p. 642-50. 139. Dieli, F., et al., Vgamma9/Vdelta2 T lymphocytes reduce the viability of intracellular Mycobacterium tuberculosis. Eur J Immunol, 2000. 30(5): p. 1512-9.

65 References

140. Fortier, A.H., et al., Live vaccine strain of Francisella tularensis: infection and immunity in mice. Infect Immun, 1991. 59(9): p. 2922-8. 141. Ladel, C.H., C. Blum, and S.H. Kaufmann, Control of natural killer cell-mediated innate resistance against the intracellular pathogen Listeria monocytogenes by gamma/delta T lymphocytes. Infect Immun, 1996. 64(5): p. 1744-9. 142. Fu, Y.X., et al., Immune protection and control of inflammatory tissue necrosis by gamma delta T cells. J Immunol, 1994. 153(7): p. 3101-15. 143. Hsieh, B., et al., In vivo cytokine production in murine listeriosis. Evidence for immunoregulation by gamma delta+ T cells. J Immunol, 1996. 156(1): p. 232-7. 144. Skeen, M.J., et al., Exaggerated proinflammatory and Th1 responses in the absence of gamma/delta T cells after infection with Listeria monocytogenes. Infect Immun, 2001. 69(12): p. 7213-23. 145. Egan, P.J. and S.R. Carding, Downmodulation of the inflammatory response to bacterial infection by gammadelta T cells cytotoxic for activated macrophages. J Exp Med, 2000. 191(12): p. 2145-58. 146. Halary, F., et al., Control of self-reactive cytotoxic T lymphocytes expressing gamma delta T cell receptors by natural killer inhibitory receptors. Eur J Immunol, 1997. 27(11): p. 2812-21. 147. Carena, I., et al., Major histocompatibility complex class I molecules modulate activation threshold and early signaling of T cell antigen receptor-gamma/delta stimulated by nonpeptidic ligands. J Exp Med, 1997. 186(10): p. 1769-74. 148. Poccia, F., et al., CD94/NKG2 inhibitory receptor complex modulates both anti-viral and anti-tumoral responses of polyclonal phosphoantigen-reactive V gamma 9V delta 2 T lymphocytes. J Immunol, 1997. 159(12): p. 6009-17. 149. Boullier, S., et al., Phosphoantigen activation induces surface translocation of intracellular CD94/NKG2A class I receptor on CD94- peripheral Vgamma9 Vdelta2 T cells but not on CD94- thymic or mature gammadelta T cell clones. Eur J Immunol, 1998. 28(11): p. 3399-410. 150. Dieli, F., et al., Selective depression of interferon-gamma and granulysin production with increase of proliferative response by Vgamma9/Vdelta2 T cells in children with tuberculosis. J Infect Dis, 2002. 186(12): p. 1835-9. 151. Putra, S.R., et al., Distribution of mevalonate and glyceraldehyde 3-phosphate/pyruvate routes for isoprenoid biosynthesis in some gram-negative bacteria and mycobacteria. FEMS Microbiol Lett, 1998. 164(1): p. 169-75. 152. Rohmer, M., The discovery of a mevalonate-independent pathway for isoprenoid biosynthesis in bacteria, algae and higher plants. Nat Prod Rep, 1999. 16(5): p. 565-74. 153. Tanaka, Y., et al., Natural and synthetic non-peptide antigens recognized by human gamma delta T cells. Nature, 1995. 375(6527): p. 155-8. 154. Burk, M.R., L. Mori, and G. De Libero, Human V gamma 9-V delta 2 cells are stimulated in a cross-reactive fashion by a variety of phosphorylated metabolites. Eur J Immunol, 1995. 25(7): p. 2052-8. 155. Bukowski, J.F., C.T. Morita, and M.B. Brenner, Human gammadelta T cells recognize alkylamines derived from microbes, edible plants, and tea: implications for innate immunity. Immunity, 1999. 11(1): p. 57-65.

66 References

156. Espinosa, E., et al., Chemical synthesis and biological activity of bromohydrin pyrophosphate, a potent stimulator of human gamma delta T cells. J Biol Chem, 2001. 276(21): p. 18337-44. 157. Belmant, C., et al., Conventional and non-conventional recognition of non-peptide antigens by T lymphocytes. C R Acad Sci III, 1999. 322(11): p. 919-24. 158. Belmant, C., et al., A chemical basis for selective recognition of nonpeptide antigens by human delta T cells. Faseb J, 2000. 14(12): p. 1669-70.

67