Disponible à / Available at permalink :

- - -

- - -

Dépôt Institutionnel de l’Université libre de Bruxelles / Université libre de Bruxelles Institutional Repository

Thèse de doctorat/ PhD Thesis Citation APA:

Gu-Trantien, C. (2012). Gene expression profiling of CD4+ T cells infiltrating human breast carcinomas identified CXCL13-producing T follicular helper cells associated with tertiary lymphoid structures and better patient outcome (Unpublished doctoral dissertation). Université libre de Bruxelles, Faculté de Médecine – Sciences biomédicales, Bruxelles.

Disponible à / Available at permalink : https://dipot.ulb.ac.be/dspace/bitstream/2013/209474/5/6afcf6ab-5462-4bad-b1e4-ee5169b84e31.txt

(English version below)

Cette thèse de doctorat a été numérisée par l’Université libre de Bruxelles. L’auteur qui s’opposerait à sa mise en ligne dans DI-fusion est invité à prendre contact avec l’Université ([email protected]).

Dans le cas où une version électronique native de la thèse existe, l’Université ne peut garantir que la présente version numérisée soit identique à la version électronique native, ni qu’elle soit la version officielle définitive de la thèse.

DI-fusion, le Dépôt Institutionnel de l’Université libre de Bruxelles, recueille la production scientifique de l’Université, mise à disposition en libre accès autant que possible. Les œuvres accessibles dans DI-fusion sont protégées par la législation belge relative aux droits d'auteur et aux droits voisins. Toute personne peut, sans avoir à demander l’autorisation de l’auteur ou de l’ayant-droit, à des fins d’usage privé ou à des fins d’illustration de l’enseignement ou de recherche scientifique, dans la mesure justifiée par le but non lucratif poursuivi, lire, télécharger ou reproduire sur papier ou sur tout autre support, les articles ou des fragments d’autres œuvres, disponibles dans DI-fusion, pour autant que :

Le nom des auteurs, le titre et la référence bibliographique complète soient cités;

L’identifiant unique attribué aux métadonnées dans DI-fusion (permalink) soit indiqué;

Le contenu ne soit pas modifié.

L’œuvre ne peut être stockée dans une autre base de données dans le but d’y donner accès ; l’identifiant unique (permalink) indiqué ci-dessus doit toujours être utilisé pour donner accès à l’œuvre. Toute autre utilisation non mentionnée ci-dessus nécessite l’autorisation de l’auteur de l’œuvre ou de l’ayant droit.

--- English Version ---

This Ph.D. thesis has been digitized by Université libre de Bruxelles. The author who would disagree on its online availability in DI-fusion is invited to contact the University ([email protected]).

If a native electronic version of the thesis exists, the University can guarantee neither that the present digitized version is identical to the native electronic version, nor that it is the definitive official version of the thesis.

DI-fusion is the Institutional Repository of Université libre de Bruxelles; it collects the research output of the University, available on open access as much as possible. The works included in DI-fusion are protected by the Belgian legislation relating to authors’ rights and neighbouring rights.

Any user may, without prior permission from the authors or copyright owners, for private usage or for educational or scientific research purposes, to the extent justified by the non-profit activity, read, download or reproduce on paper or on any other media, the articles or fragments of other works, available in DI-fusion, provided:

The authors, title and full bibliographic details are credited in any copy;

The unique identifier (permalink) for the original metadata page in DI-fusion is indicated;

The content is not changed in any way.

It is not permitted to store the work in another database in order to provide access to it; the unique identifier (permalink) indicated above must always be used to provide access to the work. Any other use not mentioned above requires the authors’ or copyright owners’ permission.

UNIVERSITE Libre De Bruxelles

IL olôC'i

F^aculté D^e M^edecine Institut Jules Bordet

Laboratoire Immunologie Moléculaire

ULB-campus Erasme

Bibliothèque des Sciences , Route deLenmck, 808lBat.c,

B-1070 Bruxelles Tél.; 02/55&

Gene expression profiling of CD4^ T cells infiltrating human breast carcinomas identified

CXCL13-producing T follicular helper cells associated with Tertiary Lymphoid Structures and

better patient outcome

Université L bre de Bruxelles

Thèse présentée en vue de l'obtention du titre académique de Docteur en Sciences Biomédicales et Pharmaceutiques par

Chunyan GU

Promotrice: Dr. Dominique Bron )

C^o-^promotrice: D^r. K^aren W^illard-G^allo

2012

UNIVERSITE Libre De Bruxelles

Faculté De Medecine

Institut Jules Bordet

Laboratoire Immunologie Moléculaire

Bibliot^

:;rsdelaSanté-CP.07 Lenniclc,808lBât.E)

Gene expression profiling of CD4^ T cells infiltrating human breast carcinomas identified

CXCL13-producing T follicular helper cells associated with Tertiary Lymphoid Structures and

better patient outcome

Thèse présentée en vue de l'obtention du titre académique de Docteur en Sciences Biomédicales et Pharmaceutiques par

Chunyan GU

Promotrice: Dr. Dominique Bron

Co-promotrice: Dr. Karen Willard-Gallo

2012

^ #■

'•,1 - '

'o'i

■i. '- ^ iffll

'\ t'

V

\ •' (ëi. \

■(^]p] 0-

/

• #'Li- , _ ■ lia V

if- i«i^' i».;;;--,•

.w *m ^/:'

y

• »

»

'IT'

~ * • * -

»* • • ,

T^-

«l.* * >■

# # t

« k

m'--

t

»

* ^r

ï:

^ F'

r- ..

«ÏÀ- y

■*l '*■-

■»'■ - 'É

I r.

!»y » ^■\

,- faqînîiTTi • ^ *-

’>• •'

l-r^.-^J

c\.

•' »' \mm :•'■ ■•'■*'

*'-) iS

»*. «'f#):%

Summary

Over the past decade, studies using murine models hâve led to the démonstration that CD4^ T helper (Th) cells play a critical rôle in the control of cancer progression. Additional support for their importance cornes from the growing body of recent clinical/translational research data demonstrating the importance of tumor-infiltrating T and B lymphocytes in long-term patient survival for various types of cancer, including breast cancer (BC). As the key population coordinating adaptive immune responses, the role(s) played by individual Th subsets in cancer immunity remains largely controversial. The Thl subset has uniquely been shown to hâve a clear anti-tumor effect, guiding CD8^ cytotoxic T cells-mediated direct tumor cell lysis through IFN-y sécrétion. Although the négative regulatory rôle played by Treg cells has been extensively studied in cancer, its prognostic value along with that of Th2 and Thl7 cells hâve not been clearly demonstrated in patients. T follicular helper (Tfh) cells, a recently characterized Th subset that plays a primary rôle in the génération of B cell memory in secondary lymphoid organs, hâve not been previousiy described infiltrating solid tumors. The principal objective of this thesis was to perform an in-depth characterization of tumor-infiltrating CD4'^ T cells (TIL) and Th subsets in human BC, where very little is currently known.

Using whole genome microarrays, we analyzed the gene expression profiles of TIL relative to their counterparts from the axillary lymph nodes and peripheral blood.

Applying a novel approach, we compared TIL profiles with public microarray data for Th subsets, demonstrating: 1) the presence of ail major Th subsets (Thl, Th2, Thl7, Treg as well as Tfh) in the TIL, 2) the TIL are effector memory rather than central memory cells, 3) the TIL are concomitantly activated and suppressed and 4) TIL from tumors with extensive lymphoid infiltrâtes are more activated/less suppressed in the TCR/CD3 signaling pathway, producing higher levels and a wider panel of Th cytokines than TIL from minimally-infiltrated tumors.

We aiso performed in vitro experiments to study tumor microenvironment effects on TIL by treating normal CD4^ T cells from healthy donor blood with primary tumor supernatants (SN). Tumor SN largely reproduces the TIL profile in normal Th cells, totally suppressing their activation and inhibiting their cytokine production.

Intriguingly, the highiy restricted number of cytokines induced by tumor SN included several tumor-promoting factors, such as IL-8 and TNF. SN from an extensively-infiltrated tumor was found to be less immune-suppressive than SN from minimally-infiltrated tumors. In line with this, TIL from minimally-infiltrated tumors

are doser to SN-treated (suppressed) activated donor cells whereas TIL from extensively-infiltrated tumors are more similar to activated cells without SN treatment.

These results led us to further investigate the observed différences between TIL from extensive and minimally-infiltrated tumors. Genes characterizing Thl and Tfh cells were enriched in the extensively-infiltrated tumors. PD-1*’'CD200*’' Tfh cells were specifically detected in extensively-infiltrated tumors by flow cytometry and these cells were determined to be the major source of the chemokine CXCL13.

Immunohistochemical analysis demonstrated highiy-organized tertiary lymphoid structures (TLS) within the tumor, containing a CD4VCD8^ T cell zone and a B cell zone with reactive germinal centers where Tfh cells and follicular dendritic cells (FDC) are résident. Their presence suggests the origin of an effective memory anti-tumor immune response.

Finally, we generated Tfh- and Thl-specific gene signatures reflecting différences between extensive and minimal TIL and tested their prognostic value in large-patient-scale public data sets. Our Tfh signature predicts better 10-year disease-free survival for ail BC subtypes, outperforming the Thl signature, suggesting that Tfh cells play a more central rôle than Thl cells in anti-tumor immunity. CXCL13 is the déterminant gene of our Tfh signature, showing particularly strong prognostic power for the HER2^ subtype. Additionally, these signatures aiso predict a better response to neoadjuvant chemotherapy.

This thesis research has demonstrated that a previousiy undetected Th subset, Tfh cells, infiltrâtes solid tumors and shown that their presence signais enhanced anti-tumor immunity.

Résumé

Durant cette dernière décennie, des travaux menés dans des modèles murins ont permis de mettre en évidence le rôle crucial joué par les lymphocytes T auxiliaires CD4^ (Th) dans le contrôle de la progression des cancers. De plus, de nombreuses études cliniques et/ou translationnelles récentes corroborent ces observations en montrant une corrélation entre l'importance de l'infiltration intra-tumorale par les lymphocytes T et B et la survie à long terme des patients atteints de différents types de cancer, dont le cancer du sein. En tant que chefs d'orchestre de la réponse immune adaptative, les rôles spécifiques des sous-populations des cellules Th restent controversés. Les Thl sont la seule population exerçant une claire réponse anti-tumorale, qui est liée à la sécrétion d'IFN-y, une cytokine primordiale à l'action des lymphocytes T cytotoxiques CDS"^.

Bien que le rôle néfaste des T régulateurs (Treg) a été largement étudié dans le cancer, leur implication pronostique ainsi que celle des Th2 et Thl7 n'ont pas encore été clairement démontrées. La présence d'une sous-population de CD4, les T auxiliaires folliculaires (Tfh), cellules clés dans la différenciation des lymphocytes B mémoires au sein des organes lymphoïdes secondaires, n'a jamais été décrite dans les cancers solides. Le but principal de ce travail est de caractériser les sous-populations des lymphocytes T CD4'^ infiltrant la tumeur (TIL) en prenant comme modèle le cancer du sein humain. A l'heure actuelle, il existe très peu de données sur les TIL CD4 dans ce type de cancer.

Nous avons d'abord établi le profil génique des TIL en les comparant avec ceux provenant des ganglions axillaires ou du sang périphérique. En appliquant une nouvelle approche, nous avons comparé les profils des TIL avec les données publiques de sous-populations de Th et démontré que : 1) toutes les sous-populations de cellules Th (Thl, Th2, Thl7, Treg et Tfh) infiltrent la tumeur, 2) les TIL ont un phénotype plus proche de celui des cellules mémoires effectrices que des cellules mémoires centrales, 3) les TIL sont simultanément activés et réprimés et 4) les TIL provenant des tumeurs massivement infiltrées («extensives») par des lymphocytes sont mieux activés et moins réprimés que les TIL des tumeurs peu infiltrées («minimales») dans la voie de signalisation TCR et produisent des cytokines d'une quantité plus élevée et d'une répertoire plus large.

Nous avons également effectué des expériences in vitro pour étudier l'effet de l'environnement tumoral sur les TIL en traitant des CD4 normaux (provenant des donneuses saines) par le surnageant (SN) extrait des tumeurs fraiches. Le SN tumoral

5

induit un profil génique proche de celui des TIL en inhibant l'activation et la production de cytokines de ces cellules stimulées. Curieusement, parmi le peu de cytokines induites par le SN tumoral, des facteurs pro-tumoraux comme IL-8 et TNF sont détectés. Le surnageant provenant d'une tumeur «extensive» est moins immunosuppresseur que ceux des tumeurs «minimales». Conformément, les TIL provenant des tumeurs «minimales» ont un profil génique proche des cellules normales activées et traitées (réprimées) par le SN tumoral tandis que les TIL des tumeurs «extensives» ressemblent aux cellules activées non traitées.

Ces résultats nous avaient guidés à investiguer plus profondément les différences observées entre les TIL des tumeurs «extensives» et «minimales». Les gènes caractéristiques des Thl et Tfh sont enrichis dans les tumeurs «extensives».

Les cellules Tfh PD1^'CD200^' sont spécifiquement détectées par cytométrie de flux dans les tumeurs «extensives» et sont identifiées comme les producteurs principaux de la chimiokine CXCL13. L'examen par immunohistochimie a permis de détecter des structures lymphoïdes tertiaires (TLS) dans la tumeur, composées d'une zone T (CD4 et CD8) et d'une zone B au sein de laquelle se trouve parfois un centre germinatif actif contenant des Tfh et des cellules dendritiques folliculaires (FDC). La présence de ces structures suggère l'origine d'une réponse immune mémoire anti-tumorale.

Finalement, nous avons établi des signatures géniques spécifiques aux Tfh et Thl et recherché leur impact pronostique dans deux bases de données publiques à grande échelle. Notre signature Tfh est positivement corrélée avec la survie à 10 ans des patientes de tous les sous-types de cancer du sein, et est plus performante que la signature Thl. Ceci suggère que les Tfh pourraient jouer un rôle plus crucial que les Thl dans la réponse immune anti-tumorale. CXCL13 est le gène déterminant de notre signature Tfh et son expression est fortement associée à une meilleure survie chez les patientes du sous-type HERZ"^. De plus, ces signatures prévoient également une meilleure réponse à la chimiothérapie néoadjuvante (préopératoire).

Cette étude a démontré qu'une nouvelle sous-population de CD4, les Tfh, infiltre la tumeur solide et leur présence indique l'existence d'une immunité anti-tumorale renforcée.

Remerciements

Mes remerciement vont en premier lieu à ma promotrice Karen Willard-Gallo qui m'a conduite dans ce domaine fascinant et qui, avec sa rigueur scientifique, sa grande patience et son enthousiasme communicatif, m'a soutenue tout au long d'un vrai chemin de recherche.

J'exprime également ma gratitude à ma promotrice Dominique Bron pour m'avoir acceuillie et m'avoir fort encouragée dans mon travail.

Ma reconnaissance à toutes les personnes qui ont contribué directement à ce travail et plus particulièrement Benjamin pour son initiation dans les analyses des données de microarray, Myriam qui m'a tant appris en cytométrie de flux, Roland (Alexandre) qui m'a appris la pathologie tumorale, Hélène qui m'a guidée dans mes premières expériences du labo, Catherine et Soizic pour leur précieux critiques, Sherene, Christine et Christos pour leur discussions fructueuses ainsi que Carole, Germain, Hugues, Marie, Françoise, Roberto, Sandeep et Stefan pour leurs aides techniques ou statistiques.

Merci à tous les membres des laboratories de recherche et du service d'anatomie pathologique à l'Institut Jules Bordet. J'apprécie en particulier les aides techniques, les conseils et les encouragements venant de Samira Majjaj, Laurence Buisseret, Debora Fumagalli, Yan Ma, Redouane Rouas, Ghiziane Rouas, Françoise Lallemand, Sandy Haussy, loanna Liaos et Jeanne Richard.

Je remercie les membres de mon jury de thèse d'avoir accepté de juger ce travail : Christophe Caux et Pierre van der Bruggen, qui ont endossé le rôle d'expert étranger, Alain Le Moine, Pierre Heimann, Raphaël Maréchal et Françoise Mascart.

Enfin, je tiens à remercier mes amis et ma famille pour leur soutient et encouragements.

Merci à Milan, mon bébé porte-bonheur et son papa, souvent un peu têtu mais qui me donne toujours du bon conseil aux moments les plus difficiles durant ma thèse. Egalement à mon futur petit Marin...

Financial Support

The \A/ork presented in this thesis was supported by the Belgian Funds for Scientific Research (FNRS) through a Télévie grant.

This thesis has been written under the supervision of Prof. Karen Willard-Gallo and Prof.

Dominique Bron.

The members of the Jury are:

• Prof. Dominique BRON (Université Libre de Bruxelles, Belgium)

• Prof. Karen WILLARD-GALLO (Université Libre de Bruxelles, Belgium)

• Prof. Alain LE MOINE (Université Libre de Bruxelles, Belgium)

• Prof. Pierre HEIMANN (Université Libre de Bruxelles, Belgium)

• Prof. Raphaël MARECHAL (Université Libre de Bruxelles, Belgium)

• Prof. Françoise MASCART (Université Libre de Bruxelles, Belgium)

• Prof. Pierre VAN DER BRUGGEN (Université Catholique de Louvain, Belgium)

• Prof. Christophe CAUX (Centre Léon Bérard, Lyon, France)

CONTENTS

List of figures and tables 12

List of abbreviations 13

INTRODUCTION...15

Part I. Breast cancer biology...16

1.1. Définition and incidence 16

1.2. Clinico-pathological classification 18

1.3. Molecular subtypes 21

1.4. Prognostic gene signatures 23

Part II. CD4"^ T helper cells in the immune response... 25

2.1. CD4^ T cell activation 25

2.1.1. TCR triggering 26

2.1.2. Costimulatory signais 27

2.1.3. Inhibitory signais 28

2.2. CD4^ T cell memory 29

2.3. CD4"^ T cell subsets 32

2.3.1. Thi cells 32

2.3.2. Th2 cells 35

2.3.3. Thi7 cells 37

2.3.4. Regulatory T (Treg) cells 39

2.3.5. T follicular helper (Tfh) cells 41

2.3.6. Tfh cell memory and plasticity 45

48 2.4. Tfh cells in secondary lymphoid organs (SLO)

2.4.1. Lymph node (LN) structure and germinal centers (GC) 48 2.4.2. Cell-to-cell interactions and migration 49

• Migration from the periphery to the lymph node

• T cell zone; from naive CD4* T cells to pre-Tfh cells

• Primary follicle: B cell encounter with pre-Tfh cells

• Tfh-B-FDC interactions in the germinal center

• Th cytokine influence on antibody responses

Part III. CD4"^ T cells in cancer immunity 56

3.1. Immune infiltrate in human cancer 56

3.2. CD4^ T cells in cancer 57

3.2.1. The emerging rôle of CD4"^ T effector/helper cells in cancer immunity 57

3.2.2. Th1 anti-tumor immunity 59

3.2.3. Th2 cells in anti- / pro-tumor immunity 60 3.2.4. Thi 7 cells in anti- / pro-tumor immunity 62 3.2.5. Treg cells in anti-/ pro-tumor immunity 63

3.2.6. Tfh cells in cancer 65

3.3. Other immune cells in cancer:

CDS"" T cells, NK cells, macrophages, MDSCs, DCs, invariant NKT cells

and y5 T cells 66

3.4. Immune suppressive factors in the tumor microenvironment:

TGF-p, IL-10and IDO

69

3.5. Tertiary Lymphoid Structures (TLS) 71

3.5.1. Ectopic tertiary lymphoid tissue in inflammatory conditions 71

3.5.2. TLS in cancer 72

3.6. Chemokines with an anti-tumor effect: their rôle in leukocyte

trafficking and cancer progression 73

3.6.1. The IFN-y-induced chemokines (CXCL9, CXCL10, CXCL11) 75

3.6.2. CXCL13 77

RESULTS ⁸⁰

CXCL13-producing follicular helper CD4^ T cells infiltrating tumors signal an organized immune response and predict improved patient outcome in breast cancer...81

DISCUSSION... 82

1. Novel approaches for studying CD4^ TIL subset 84

2. TIL dysfunction and the effect of tumor SN 86

3. TLS in cancer 87

4. The phenotype of tumor-infiltrating Tfh cells 90

♦ PD-1 and CXCR5 expression on Tfh cells 91

♦ PD-1 expression on exhausted TIL 92

♦ CD200 as an additional Tfh cell marker 93

♦ CXCL13 as a Tfh cell cytokine (chemokine) 95

5. Prognostic immune gene signatures in breast cancer 96

CONCLUSIONS AND PERSPECTIVES... 98

REFERENCES... 101

ANNEXE...136

Molecular characterization of CD3~ CD4^ T cells from patients with the

lymphocytic variant of hyperéosinophilie syndrome...137

List of Figures and Tables

Figure 1. Structure of the female breast 16

Figure 2. Annual age-adjusted cancer diagnosis and death rates among

females for selected cancers, United States 17

Figure 3. Estimated age-standardized breast cancer incidence and mortality rates and lifetime risk for a female aged between 15 and 79 years old to hâve

breast cancer in worldwide régions 18

Figure 4. Five-year survival rates of breast cancer subtypes according to ER,

PR and HER2 status 21

Figure 5. TCR/CD3 complex and interaction between TCR on CDA"" T cells

and pMHC on APCs 26

Figure 6. Models for memory T cell sélection 31

Figure 7. Lymph node structure 49

Figure 8. Stromal cell populations in the T cell zone and B cell follicle of a

murine lymph node 50

Figure 9. Migration of Tfh cells and their interactions with B cells in SLOs 51 Figure 10. GC B cell maturation and sélection 53 Figure 11. MRC and FDC networks in a primary follicle and after maturation

into a GC-containing secondary follicle 54

Figure 12. Immune cell populations in the tumor microenvironment and their

currently accepted rôles in tumorigenesis. 57

Figure 13. Chemokines and their receptors 74

Table 1. Cytokines and transcription factors in Th différentiation 33

Table 2. Effects of gene deficiency on Tfh cell formation 43

List of abbreviations

AITL angioimmunoblastic T-cell lymphoma AP-1 activating protein-1

APC antigen presenting cell

(i)BALT (inducible) bronchus-associated lymphoid tissue Bcl6 B-cell lymphoma 6

BCR B cell receptor

CAF cancer-associated fibroblaste

CCL CC chemokine ligand

CCR CC chemokine receptor CD cluster of différentiation CTL CD8^ cytotoxic T lymphocyte CXCL CXC chemokine ligand CXCR CXC chemokine receptor

DC dendritic cell

EAE experimental autoimmune encephalomyelitis

ER estrogen receptor

FoxP3 forkhead box P3

FRC fibroblastic reticular cell

GC germinal center

HER2 human epidermal growth factor receptor 2 HEV high endothélial venule

IDO Indoleamine 2,3 dioxygenase

IFN interferon

Ig immunoglobulin

IHC immunohistochemistry

IL interleukin

ITAM immunoreceptor tyrosine-based activation motif

LCMV lymphocytic choriomeningitis virus

LN lymph node

LTi lymphoid tissue-inducer LT-a (P) lymphotoxin a 0)

MDSC myeloid-derived suppressor cells

MHCI/II major histocompatibility complex class l/ll

MRC B cell follicle marginal zone fibroblastic reticular cell N FAT nuclear factor of activated T cells

NF

B nuclear factor kappa-B

OSCC oral squamous cell carcinoma

pCR pathological complété response (to neoadjuvant therapy) PD-1 program death receptor 1

PD-L1/2 program death receptor ligand 1/2 pMHC peptide-MHCII complex

PNAd peripheral node addressin PRR pathogen récognition receptor SLO secondary lymphoid organ

STAT signal transducer and activator of transcription TAP1/2 transporter associated with antigen processing 1/2 TCR T cell receptor

Tfh CD4'^ T follicular helper cells TGF transforming growth factor TIL tumor-infiltrating lymphocyte TLS tertiary lymphoid structure TNBC “triple-negative” breast cancer TNF tumor necrosis factor

TRC T cell zone fibroblastic reticular cell

(n/i)Treg (natural/induced) CD4'^ regulatory T cells

VEGF vascular endothélial growth factor

Introduction

Part I. Breast cancer biology

1.1 .Définition and incidence

The female breast has 6-10 interweaving duct Systems which successively branch into numerous smaller ducts, eventually giving rise to the terminal duct-lobular unit, each comprising a single lobule (Figure 1). The entire female ductal-lobular System is lined by an inner epithelium and outer myoepithelium.

Based upon tumor histology, breast neoplasms are classified as breast sarcomas (which arise from the interlobular stroma), fibroadenomas (which arise from the intralobular stroma) and épithélial tumors (=carcinomas; which arise anywhere along the ductal-lobular System). Most malignant épithélial neoplasms arise from the inner epithelium of the terminal duct-lobular unit and are sub-divided into ductal carcinoma and lobular carcinoma, the two principle types of breast carcinomas [Yoder et al., 2007].

Lymph nodes Lymph

vessels

Figure 1. Structure of the female breast.

From the web site http://www.news-medical.net/health/What-is-Breast-Cancer.aspx

Breast cancer (BC) is the most frequently diagnosed cancer and the leading cause of cancer deaths in females worldwide (Figure 2: an example of US data), accounting for 23% of total new cancer cases and 14% of total cancer deaths in 2008 [Jemal, 2011]. Each year, more than 1.3 million women are diagnosed with breast cancer. Up to one in eight (~12.3%) women in North American and some European

100-1 140 n

<0 '

EO

U. 60-

o

§

8'

i' O

« =

20-

Diagnosis (year)

Lung and Bronchus

Il I I I II I I i M I I M 1 I I M I I I M I I I i 19S0 1985 1990 1995 2000 2005 ' I I i I I I I I I I I I I I I M I I I I I I I I I I I I I I I I I I i I I I I I I I I I I I

1930 1935 1940 1945 1950 1955 1960 1965 1970 1975

Death (year)

Figure 2. Annual age-adjusted cancer diagnosis and death rates among females for selected cancers, United States. Modified from Siegel et al., CA Cancer J Clin. 2011 Jul-Aug;61(4):212-36.

countries develop BC during their life time (Figure 3). Between 1980 and 2010, the global BC incidence and death increased annually 3.1% and 1.8%, respectively [Forouzanfar, 2011].

Incidence increases with âge, varying from <0.5% in women under 40 years old to -3.5% in women over 60 years old [Maxmen, 2012]. Rising population numbers in women of at-risk âges and the availability of early détection services are the main reasons for the higher incidence rates observed in developed régions compared to developing countries (Figure 3). In contrast, mortality rates are much lower in developed régions because of the more favorable BC survival. About 60% of deaths are estimated to occur in the developing régions [Jemal, 2011]. BC death rates hâve been decreasing in North America and several European countries over the past 25 years, largely as a resuit of early détection through mammography screening and improved thérapies. In the US, the five-year survival rate increased from about 40%

in 1950 to 85.8% in 2004 [Maxmen, 2012].

H

H

15 ^{ri 00}

Mean

Figure 3. Estimated age-standardized breast cancer incidence and mortality rates (per 100,000 females) and lifetime risk (%) for a female aged between 15 and 79 years old to bave breast cancer in worldwide régions. Modified from GLOBOCAN, 2008. Web site: http://globocan.iarc.fr/

1.2.Clinico-pathological classification

For a given patient, many factors influence BC survival rate, including how early the tumor is detected (tumor stage, grade, and metastatic status) and tumor subtype.

Women with localized (confined to the primary site), régional [spread to axillary lymph nodes (LN)] or distant (metastasized to distant organs) tumors at diagnosis hâve a five-year survival rate of 98.6%, 83.8% or 23.3%, respectively [Maxmen, 2012].

HistoloRic classification

Ductal and lobular carcinomas are the most common types, accounting for 75-80% and 10-15%, respectively, of ail mammary carcinomas, although other rarer forms aiso exist. Both types can be in situ and/or invasive (infiltrating), depending on whether the outer myoepithelium is disrupted by neoplastic cells that invade the surrounding intra- or interlobular stroma (Figure 1). Invasive ductal carcinoma has the worst prognosis among ail histologie types of breast carcinoma followed by invasive lobular carcinoma, where the prognosis is a little better [Yoder et al., 2007].

However, it is important to further sub-classify invasive ductal carcinomas because the prognosis varies considerably between subtypes; although 80% are unclassifiable (histologically) and thus designated as the unspecified type. The unspecified type has a better prognosis than some subtypes of ductal carcinoma (including micropapillary and metaplastic carcinomas) and a worse prognosis than other subtypes (including tubular, papillary and medullary carcinomas). Medullary carcinoma, often high grade, is characterized by a frequent dense lymphoid infiltrate and has a relatively favorable prognosis among the different types or subtypes of breast carcinomas.

Inflammatory carcinoma, a rare form of breast carcinoma that can be of any histologie type, has the worst prognosis among ail forms of breast carcinoma, with much lower survival rates compared to other locally advanced BC [Bastawisy et al., 2012]. Overall, BC is a highiy heterogeneous disease, and even with the best current treatments, the 10-year disease-free survival rates vary from 98% to 10%.

Furthermore, within stage IV disease, there are still wide variations in survival from 11% for inflammatory to 34% for mucinous or papillary carcinoma [Gloeckler Ries and Eisner, 2004].

TMN staging

Staging is the process used to evaluate the extent or severity of a patient's cancer. Cancer stages range from Stage 0 (a very early form of cancer) to Stage IV (advanced, metastatic cancer). The staging System is called TNM which stands for

"Tumor, Node, Metastasis". For BC, TNM staging takes into account tumor size and extent of its spread to the skin or chest wall under the breast (T; from 0 to 4), whether the cancer has spread to nearby LNs and, if so, how many LNs are affected (N; from 0 to 3), and whether it has spread to distant organs (M; from 0 to 1).

TNM stage is strongly associated with patient outcome. For patients of ail âges, those diagnosed in stages 0 and I had a 100% 5-year relative survival rate while those diagnosed with stage IV had a rate of 21% [Gloeckler Ries and Eisner, 2004].

Histological grading

Histological grading, from 1 to 3 for "low", "intermediate" and "high" grade, refers to the degree of de-differentiation and prolifération rate of the malignant cells.

The Bloom-Richardson Scoring System is used to define tumor grades. The main éléments of this method are the extent of tumor tubule formation (for ductal carcinoma), the level of mitotic activity or rate of cell division, and the degree of nuclear atypia ("pleomorphism") of the tumor cells.

Histologie grade is aiso a predictor of outcome independent of clinical stage.

Except for stages 0-1, histologie grade plays an important prognostie rôle with patient survival rate deereasing 15-20% from grade 1 to grade 3 within eaeh satge (ll-IV) [Gloeckler Ries and Eisner, 2004].

Lymph node involvement

Axillary LN status (positive or négative, indieating with or without LN metastasis) is of primary importanee as a predietor for reeurrenee and survival [Donegan, 1992].

Generally, régional LN involvement inereases with tumor size, with more than 60% of women with >5 em tumors having positive nodes at the time of diagnosis [Gloeckler Ries and Eisner, 2004]. Examination of the sentinel LN is a reliable teehnique for evaluating axillary LN involvement, whieh spares some patients from unneeessary LN removal.

Hormone reeeptor and growth faetor reeeptor status

Another important diagnostie praetiee is to measure the expression levels for the two hormone receptors, the estrogen reeeptor (ER) and the progestérone reeeptor (PR), and the expression level or gene amplifieation of the human epidermal growth faetor reeeptor 2 (HER2, aIso known as HER2/neu, ErbB-2 or ERBB2). The status of these three reeeptors is used to define the subtypes of BC and is of primordial prognostie and therapeutie importanee (Figure 4).

Approximately 75% of breast tumors express ER and most of them aiso express PR. HER2 overexpression or amplifieation is found in 20-30% of BC patients. HER2'^

and ER“ tumors tend to be of high grade (poorly differentiated and highly-proliferating), hâve more frequent p53 mutations [Rossner et al., 2009;

Melhem-Bertrandt étal., 2012] and a worse prognosis eompared with ER"^ disease. A subtype of tumors négative for hormone reeeptors and HER2 amplifieation, ealled

"triple-negative" breast eaneer (TNBC), whieh aeeounts for 10-20% of breast

carcinomas, is particularly aggressive, with a five-year survival rate of -77%

comparée! with 93% for other types of breast carcinomas [Powell, 2012], Triple négative patients aiso tend to be younger and more frequently carry an inherited mutation in the tumor suppressor gene BRCAl or BRCA2 [Hudis & Gianni, 2011].

About 75% of TNBC are classified as the "basal-like" molecular subtype (detailed in Section 1.3).

0-5 year relative cumulative sur\'ival

100 90 80

53.2% .... 70

13.1% fî?

60 11.4%

(O>

.... 50 {/)3

9.9% 40

(U>»

7.2% 30

U.>

3.5%

1.7% (other)

20

■ ER+/PR+/HER2- | ER+/PR-/HER2+

■ ER+/PR-/HER2- ■ ER-/PR-/HER2- ... 10

■ ER+/PR+/HER2+ ■ ER-/PR-/HER2+

--- 0

Figure 4. Five-year survival rates of breast cancer subtypes according to ER, PR and HER2 status. From Maxmen, Nature 2012. May 30;485(7400):S50-1.

1.3. Molecular subtypes

Overthe past decade, large scale gene expression profiling of tumor tissue using microarray technologies hâve permitted the reclassification of BC based on différences in the expression of hundreds to thousands of genes within groups of patients.

In 2000, Pérou and colleagues defined the first molecular BC subtypes naturally displaying distinct gene expression patterns using unsupervised hierarchical clustering analysis [Pérou et al., 2000]. The resulting clusters included a large

proliferation-related gene cluster with smaller clusters containing genes reflecting spécifie activities such as the interferon pathway. These expression patterns aiso distinguish genes characteristic of multiple cell types, including the two major épithélial cell types (basal/myoepithelial and luminal épithélial cells, indicating the tumor origin), as well as endothélial cells, stromal cells, adipose-enriched/normal breast cells, B lymphocytes, T lymphocytes and macrophages, which reflect the composition of the tumor microenvironment. The four major subtypes (luminal, basal, ERBB2'^ and normal breast-like) were shown to cluster separately. Compatible with the traditional classification of BC, this molecular classification related ER'^

patients to luminal épithélial cell-like, TNBC patients to basal épithélial cell-like, HER2^ patients to overexpression of ERBB2 locus genes and additionally defined a normal breast cell-like subtype.

In a subséquent study, Sorlie, Pérou and colleagues further separated ER^

luminal patients into A, B and C subtypes reflecting low to high tumor prolifération rates, and for the first time correlated BC molecular subtypes with patient survival [Sorlie et al., 2001]. Luminal A patients hâve excellent survival rates while basal, ERBB2'^ patients hâve significantly reduced survival rates with the normal breast-like subtype found to be intermediate. Distinct from the luminal A subtype, luminal B and C patients (now considered together to be the luminal B subtype) aIso hâve reduced survival rates and a high frequency of p53 mutations, to an extent comparable with the aggressive basal and ERBB2'^ subtypes. The identification of a more aggressive subtype of ER'^ patients previousiy considered to be in the good prognostic group constitutes a major finding from these first studies using gene expression profiling of BC.

Independent from tumor size or nodal status, the molecular subtypes were repeatedly detected in independent gene expression data sets with similar survival rate stratification, which was especially clear for the luminal A, luminal B, basal and HER2 subtypes [Sorlie et al., 2003; Sotiriou et al., 2003; Hu Z et al., 2006; Wirapati et al., 2008; Mackay et al., 2011]. Although these molecular subtypes are significantly associated with patient outcome, the classification of an individual

patient is not practicable using this clustering method due to the nature of the analysis (based on assessment of many patients to generate natural clusters) and its inhérent instability (the individuel branches in each cluster can change considerably between slightiy different patient sets), thus remaining a diagnostic challenge.

1.4.Prognostic gene signatures

To search a more suitable method for classifying individuel patients, an extensive body of work has been performed since 2001 in a direct search for genes specifically associated with known patient outcomes. The first génération prognostic gene signatures such as Gene 70 (=MammaPrint) [van 't Veer et al., 2002; van de Vijver et al., 2002] and Gene 76 [Wang Y et al., 2005], were derived from the global analysis of ail samples without taking into account the presence of molecular subtypes. Alternatively, Sotiriou and colleagues used an interesting approach to dérivé a 97 gene signature (GGI for Genomic Grade Index) by comparing histologie grade 1 and 3 ER"^ patients with the resulting signature successfully separating histologie grade 2 patients into 2 groups with distinct survival rates, comparable to either grade 1 or grade 3 patients [Sotiriou et al., 2006]. These signatures, as well as many other first génération gene signatures robustly separate ER'^ patients into high and low risk groups, corresponding to luminal B and luminal A, respectively. These gene signatures solved the problems associated with the clustering method for patient classification; however, for ER“ patients they failed to predict any outcome benefit. Examination of the genes included in these signatures led to the récognition that there was little convergence between genes from different signatures, with ail of them reflecting the single biological process of prolifération. Furthermore, their power was seriousiy questioned concerning its superiority to classical immunohistochemistry (IHC) tests that stain for prolifération marker proteins such as Ki67.

Although the initial gene expression studies detected genes associated with cellular components of the tumor microenvironment [Pérou et al., 2000], their prognostic value only began to emerge in 2007. The elevated false positivity

observed in these first microarray studies led to only the most significant genes being considered (i.e. the numerous prolifération signatures with low gene convergence among them) with ali other biologically meaningfui signais, which were generally low, initially ignored. Teschendorff and colleagues were the first group to discover the excellent prognostic value of immune response genes in ER“ patients [Teschendorff et al., 2007]. The weak signal from the immune genes compared to proliferation-associated genes was the main reason why their relevance was initially overlooked. Subsequently, several groups independently reported immune gene signatures that were linked with improved patient outcome for aggressive forms of disease, including ER" tumors (particularly basal-like and TNBC) [Desmedt et al., 2008; Rody et al., 2009; Yau et al., 2010; Lehmann et al., 2011], HER2^ [Alexe et al., 2007; Desmedt et al., 2008; Rody et al., 2009], and high proliférative tumors [Schmidt et al., 2008], These immune signatures, derived from the analysis of total tumor gene expression profiling, detect either a broad spectrum of immune response genes [Alexe et al., 2007; Lehmann et al., 2011], genes associated with spécifie pathways such as STATl [Desmedt et al., 2008] or spécifie cell types such as the T-cell metagene [Rody et al., 2009] and the B-cell metagene [Schmidt et al., 2008]. Accumulating evidence suggests that T and B lymphocytes are the major cellular sources providing these bénéficiai gene signais [Rody et al., 2011; West et al., 2011; Ascierto ML et al., 2012]. Additionally, recent studies hâve aiso demonstrated that immune signatures can efficiently predict a patient's response to chemotherapy.

Increased rates of pathological complété response (pCR; defined as the absence of residual invasive carcinoma in the breast and absence of metastatic cells in the axillary LNs after completion of neoadjuvant treatment which is known to predict an excellent long-term patient survival) after preoperative (neoadjuvant) chemotherapy were observed in patients with higher expression levels of published immune gene signatures, with the strongest association observed in the HER2'^ subtype and to a lesser extent the basal-like subtype [Ignatiadis et al., 2012].

Part II. CD4^ T helper cells in the immune response

The immune System contains complex networks of cells, each displaying distinct specificities and functions within the highiy organized and coordinated immune response. Cells of the innate immune System possess conserved pattern récognition receptors (PRR) that detect a limited range of pathogens or pathogenic host cells expressing the corresponding danger-associated molecular patterns. The innate immune cells are rapid responders that initiate the immune response. The adaptive immune System is principally composed of T and B cells, which are charged with recognizing a huge diversity of foreign antigens (including abnormally expressed host antigens) with high specificity. T and B cells aiso hâve the capability of generating immunological memory, which is needed for longtime immune surveillance and the rapid induction of a high level spécifie immune response to recall antigens.

During primary viral infections, antigen spécifie naive CDA^ T cells are activated by dendritic cells (DC) to undergo massive clonal expansion and differentiate into multifunctional effector [Thl, Th2, Thl7, follicular helper T (Tfh), etc.] and regulatory (Treg) subsets capable of producing spécifie patterns of cytokines. Help from spécifie subset(s) of 004"^ T cells is critical for promoting B cell antibody responses and generating B cell memory (help from Tfh cells) as well as the production of cytotoxic and memory CDB"^ T cells (help from Thl cells). Once the infection is resolved, the majority of effector T cells undergo apoptosis with only a small population of memory CD4^ T cells surviving long-term as the rapid and effective responders upon reinfection.

2.1 .CD4'^ T cell activation

Naive CD4'^ T cells, after their sélection and maturation in the thymus, enter the circulation and begin the search for their first antigen récognition spécifie for their T cell receptor (TCR). Generally, this encounter occurs in the T cell zone of the LN or another secondary lymphoid organ (SLO), where they can more easily interact with antigen-experienced mature DCs or another antigen presenting cell (APC). APCs

présent processed antigenic peptides, which are embedded in the major histocompatibility complex class II (MHCII) receptors on their surface, to the naive CD4'^ T cells whose TCR can recognize the spécifie peptide-MHCII complex (pMHC) and initiate activation. This process of antigen-specific activation of naive CD4^ T cells is the first step in an adaptive immune response.

2.1.1. TCR triggering

Activation of naive CD4^ T cells requires at least two separate signais. The first signal results from TCR-pMHC ligation and CD3 phosphorylation; the second signal is produced by interactions bet\A/een a number of costimulatory molécules and their paired ligands expressed on Th cells and APCs, respectively.

Responsible for spécifie antigen récognition and uniquely expressed on T cells (CD4^ or CD8^), the TCR receptor is composed of a heterodimer of TCR chains (TCRa/(3 orTCRy/6) associated with three dimers of CD3 chains (CD3y/e, CD36/e and CD3^0 [Call & Wucherpfennig, 2005; Kuhns & Davis, 2012] (Figure 5). Bach TCR heterodimer contains a variable région serving as the peptide binding site, which is produced via rearrangement of the TCR genes. The intracellular domains of the CD3 chains contain one to three immunoreceptor tyrosine-based activation motifs (ITAMs) that are critically phosphorylated leading to signal transduction.

Figure 5. TCR/CD3 complex (right) and interaction between TCR on T cells and pMHC on APCs (left). Modified from Kuhns & Davis, Front Immunol. 2012;3:159.

In naive CD4^ T cells, the inactive Src kinase Lck, associated with the intracellular domain of CD4 receptor, is held at a distance from the TCR/CD3 complex, thereby preventing ITAM phosphorylation (Figure 5). Upon TCR ligation, CD4 clusters with the TCR/CD3 complex in lipid rafts where it can then interact with the pMHC complex helping to stabilize the TCR-pMHC interaction (reviewed in [Smith-Garvin et al., 2009]). Multi-molecular assemblies of TCR/pMHC complexes are required to médiate sustained TCR signaling on CD4'^ T cells [Kuhns & Davis, 2012]. As a conséquence, Lck is brought into the proximity of the CD3 cytoplasmic domains where it then médiates ITAM phosphorylation. Recruitment of ZAP70 follows, initiating a cascade of phosphorylation events and activation of numerous signaling molécules that lead to the génération of the secondary messengers diacylglycérol (DAG) and inositol triphosphate (IP3) as well as Ras activation. DAG promotes dégradation of the NF-kB inhibitor IkB, releasing this transcription factor for entry into the nucléus. IP3 induces an increase in intracellular Ca^"^ levels, leading to NFAT déphosphorylation and nuclear translocation. Ras initiâtes the Raf-l/MEK/ERK pathway and triggers activation of the transcription factor AP-1. The combined action of the transcription factors NF-kB, NFAT and AP-1 then directs the activation process and cytokine production.

2.1.2. Costimulatory signais

Signaling solely through the TCR results in a non-responsive anergie State in naive T cells with co-ligation of other surface receptors required to provide the additional signais necessary for full T cell activation.

CD28 is the best characterized costimulatory receptor and regarded as the principle effector. CD28, constitutively expressed on >90% CD4^ T cells, binds to CD80 (B7-1) and CD86 (B7-2) expressed on APCs. Microarray gene expression studies revealed that CD28 costimulation provokes quantitative rather than qualitative changes in human T cell activation [Diehn et al., 2002]. Among the few genes induced by CD3/CD28 ligation but not by CD3 triggering alone, IL-2 is the most prominent one. The IL-2 receptor is aiso induced on recently activated CD4^ T cells

with the IL-2/IL-2 receptor autocrine loop significantly increasing activation-induced T cell prolifération. Through an IL-2-independent mechanism, CD28 costimulation induces down-regulation of the cyclin-dependent kinase inhibitor p27, driving the cells into the G1 phase of the cell cycle; however, entry into S phase is partially IL-2-dependent. In addition to their proliférative rôles, CD28 and IL-2 signaling promote T cell survival by upregulating the pro-survival protein BcI-Xl and the anti-apoptotic molécule Bcl-2, respectively.

TCR/CD3 ligation aiso induces many other costimulatory molécules, which are absent on the surface of resting CD4'^ T cells, including ICOS, 0X40 (CD134) and 4-lBB (CD137). ICOS shares several structure features with CD28 and both signal directiy through the PI3K/Akt pathway via a cytoplasmic tail associated with p85, a regulatory subunit of PI3K. In contrast to CD28, ICOS ligation is unable to induce IL-2 expression.

0X40 and 4-lBB aIso induce activation of PI3K/Akt but are not directiy associated with a protein kinase but link downstream signaling through the TRAF family of adapter proteins.

2.1.3. Inhibitory signais

The immune System has a reostatic mechanism for resetting the threshold for T cell activation, using inhibitory signais to limit the size and duration of an immune response when its function has been achieved. T cell-intrinsic inhibitory signais are mainly mediated by CD28:B7 family and TNF/TNFR family members, including cytotoxic T-lymphocyte antigen-4 (CTLA-4), programmed death-1 (PD-1) and B- and T-lymphocyte attenuator (BTLA). AH three receptors are induced by activation, with CTLA-4 aiso constitutively expressed in/on regulatory T cells.

CTLA-4 shares the two ligands CD80 and CD86 with CD28 but has higher affinity for these ligands, thus effectively competing for and limiting CD28 costimulation. The engagement of CTLA-4 results in a net decrease in the production of IL-2 and the IL-2 receptor with cell cycle arrest in Gl. Temporal and spatial control of ligand expression may aiso be important in determining the outcome of immune responses. CD86, the preferred ligand for CD28, is expressed at low levels in unactivated DCs and is rapidiy

upregulated by a variety of activating stimuli. In contrast, CD80, the preferred CTLA-4 ligand, virtually absent on unactivated DCs, is upregulated by similar stimuli, but its expression on the cell surface peaks later than CD86 expression. It seems that the requirement of the cytoplasmic tail of CTLA-4 for its inhibitory fonction dépends on the surface CTLA-4 and B7 levels [Carreno et al., 2000]. In addition, reverse signaling through B7 may be important mechanism for limitingT cell prolifération, by inducing APC production of indoleamine 2,3-dioxygenase (IDO; an enzyme to catalyze tryptophan metabolism), reducing tryptophan availability and increasing pro-apoptotic métabolite production [Fallarino et al., 2003]. CD86 can aiso be expressed on activated CD4^ T cells [Paust et al., 2004] and seems to play a rôle in negatively controlling autoreactivity through CTLA-4 on Treg.

PD-Ll (B7-H1) and PD-L2 (B7-DC) are B7 family members that serve as ligands for PD-1 and induce apoptosis in PD-1 expressing cells. PD-Ll aIso binds CD80, mediating bi-directional inhibitory signaling. In contrast to CTLA-4 which blocks PI3K downstream Akt signaling, PD-1 directiy blocks TCR/CD28-mediated activation by inhibiting PI3K activation, which results in inhibiting the expression of the survival gene Bcl-xL. Thus, PD-1 ligation is more sensitive for the induction of apoptosis than the ligation of CTLA-4 [Riley, 2009].

BTLA is another B7/CD28 family member exerting strong inhibitory signais.

However, the BTLA ligand herpesvirus entry mediator (HVEM) is a member of the TNF receptor family. BTLA is highiy expressed on naive T cells but decreases during T cell activation. HVEM aiso interacts with another B7/CD28 family inhibitory receptor, CD160 and two TNF receptor family members LIGHT and lymphotoxin a (LT-a), which can deliver positive costimulatory signais. This is a unique example of direct interactions between these two families for regulating positive and négative costimulatory signais.

2.2.CD4'^T cell memory

In general, memory CDA"^ T cells are thought to provide help for B cells and CDB"^

T cells during both primary and antigen recall responses. Recent studies in murine

models suggest they aiso hâve additional rôles, including the enhancement of inflammatory cytokine and chemokine production by the innate immune response [Strutt et al., 2010], the direction of protective responses through non-helper fonctions in the absence of CDS"^ T cells and B cells [Teijaro et al., 2010] and the médiation of perforin-dependent direct cytotoxicity [Brown et al., 2006] during influenza reinfection. Influenza virus spécifie memory 004"^ T cells hâve been shown to mobilize enhanced T cell recruitment and immune responses in the lung through an IFN-y-dependent mechanism. More recently, McKinstry and colleagues provided further detail by showing that: 1) in the absence of host B or T cells (but not both), donor memory CDA"^ T cells can protect mice from low-dose virus challenge, independent of donor cell IFN-y production, 2) in the absence of both B and T cells, memory CD4'^T cell-produced IFN-y is required fora limited protection effect, and 3) donor memory CD4^ T cells can enhance B cell responses independent of Tfh cells, translated from the no effect by CD40-L blockade or SAP deficiency in donor memory CD4^ T cells [McKinstry et al., 2012].

Vaccine-induced CD4'^ memory T-cell populations can persist for up to 75 years with a half-life of 8-15 years [Hammarlund et al., 2003]. However, very little is currently known about the general nature of fate decisions that CD4^ T cells make during the primary response. The most important factors appear to be TCR avidity and antigen dose during the primary response (reviewed in [Kim & Williams, 2010]).

The TCR répertoire of antigen-specific T cells narrows to progressively higher avidity throughout the primary response and during subséquent re-challenges. However, the avidity of CD4'^ memory T cells inversely correlates with the initial antigen dose.

Interestingly, clonal populations of CD4'^ T cells can undergo functional avidity maturation throughout the primary response even as the TCR itself remains fixed.

Additionally, different inflammatory adjuvants can select different TCR répertoires with a wider range and higher avidities than immunization with peptide alone.

Intriguingly, antigen retained in SLOs after virus clearence can continue to shape the CD4^ memory T-cell compartment for several weeks [Jelley-Gibbs et al., 2005].

An interesting model of CD4^ memory T-cell différentiation is shown in Figure 6

(from [Kim & Williams, 2010]).

(a) Low antigen stimulation

Recruitment threshold

(b) High antigen stimulation

Recruitment Effector threshold

Memory

•Maximal expansion

•Robust etfector fonction

•Long-lived memory

•Robust expansion

•Robust ef^tor fonction

•Short-lived memory

•Some expansion

•Some effector fonction

•No memory

•Poor expansion

•Poor effector fonction

•No memory

Memory

&

&

<

5

V

• Maximal expansion

• Robust effector fonction

•Long-lived memory

•Maximal expansion

• Robust effector fonction

•Long-lived memory

• Robust expansion

• Robust effector fonction

• Short-lived memory

•Some expansion

•Some effector fonction

•No memory

(c) Repeated antigen stimulation

Repeated antigen encounters Recruitment

threshold

II

;o â

üI

SCL

Ο

O O O O

<à

<à

<S)

<S) <à

Long-lived memory

Short-lived merrrory

Figure 6. Models for memory T cell sélection. From Kim & Williams, Immunology.

2010 Nov;131(3):310-7.

(a) Under conditions of low antigen availability, only high-avidity clones populate the

memory pool, (b) Under conditions of high antigen availability, intermediate and high avidity clones populate the memory pool, (c) Following repeated antigen stimulation, high-avidity clones might gain a compétitive advantage and preferentially populate the memory pools.

2.3.CD4'"T cell subsets

During the course of an immune response, CD4'^ T cells undergo complicated patterns of différentiation in the lymphoid compartment from uncommitted naive cells to highiy spécifie effector cells under the influence of spécifie cytokines. The initial discovery of CD4'^ T cell subsets was however derived from in vitro observations. In 1986, Mosmann and Goffman identified two distinct subsets of activated CD4'^ T helper/effector cells, Thl and Th2, defined by their unique cytokine profiles and fonctions [Mosmann et al., 1986]. This division fit nicely with previous démonstrations that organisms tend to mount either a cell-mediated or humoral response, but not both, in response to a given pathogen. Since then our understanding of differentiated Th cell lineages has expanded greatly. Today at least five well characterized Th subsets hâve been identified (their development is determined by the pattern of signais they reçoive during initial antigen interactions), including Thl, Th2, Thl7, Treg and the more recently characterized Tfh cells, with each subset exerting spécifie fonctions. However, the debate continues on their lineage distinction due to the flexibility of established Th subsets to adopt another phenotype under the influence of factors présent in the microenvironment, which is often referred to as their plasticity. Nevertheless, most of the studies investigating functional rôles or différentiation mechanisms for each Th subset were conducted in murine models.

2.3.1. Thl cells

Thl cells play an important rôle in cellular immunity to intracellular microorganisms and are critical in most organ-specific autoimmune diseases and anti-tumor immunity. Thl cytokines are aiso thought to be responsible for delayed

type hypersensitivity and macrophage activation.

The hallmark cytokine produced by Thl cells is IFN-y (Table 1), a cytokine whose activity can induce activated CDS"^ T cell prolifération, polarize macrophages to the type 1 effector (Ml) phenotype and increase natural killer (NK) cell lytic activity.

Table 1. Cytokines and transcription factors in Th différentiation.

Produced cytokines

Promoting cytokines

Transcription factors

Signaling molécules

Chemokine receptor

main IFN-y, IL-2, IL-12, IFN-y T-bet STATl, CXCR3

Thl TNF STAT4 CCR5

addi- IL-10, LT-a IL-18 RUNX3

tion

main IL-4, IL-5, IL-4 GATA3 STAT6 CCR3

Th2 IL-13 CCR4

addi- IL-9, lL-10, IL-2, IL-6 c-Maf STAT5 CCR8

tion IL-25,

amphiregulin

main IL-17A IL-6, TGF-P RORyt STAT3 CCR4

Thl7 _addi- IL-17F, IL-21, IL-IP, IL-21, RORa, RUNXl CCR6

tion IL-22, LT-a IL-23

main TGF-P, IL-10 TGF-P FoxP3 STAT5, CCR4

Treg ^{SMAD3 CCR5}

addi- IL-2 CCR8

tion

main IL-21, CXCL13 IL-6, IL-21 Batf, Bcl6 STAT3 CXCR5

Tfh _addi- IL-4, IFN-y IL-23, IL-27 c-Maf, IRF4 CXCR4

tion

IFN-y production is initially showed in mice to be induced through STAT4 pathway essentially by exogenous cytokine IL-12 signaling [Thierfelder et al., 1996], mainly thought to be produced by activated DCs.

In mice as well as in human, there are two IL-12 spécifie receptors, IL-12RP1 (constitutively expressed on naive CD4'^ T cells) and IL-12R32 (induced by TCR activation and maintained by IL-12 and IFN-y signaling) [Szabo et al., 1997]. IFN-y then acts through the STATl pathway to induce T-bet (TBX21), the Thl master transcription factor leading to further IFN-y production [Szabo et al., 2000] and IL-12 receptor expression [Afkarian et al., 2002]. This positive feedback loop is required to sustain Thl development while the Thl cytokine IFN-y and the Th2 cytokine IL-4 mutually inhibit the différentiation of one another. IL-18 is a potent IFN-y-inducing factor which can synergize with IL-12 to induce IFN-y production [Robinson et al., 1997] with the receptor IL-18Ra upregulated during Thl différentiation. RUNX3 is an additional Thl transcription factor induced by T-bet that is capable of binding to IFNG promoter (as T-bet) [Djuretic et al., 2007] and regulating IFN-y production through inhibition of the Th2 master transcription factor GATAS [Kohu et al., 2009].

IL-2, LT-a and TNF are aiso cytokines produced by Thl cells. IL-2 promotes the prolifération of recently activated effectorT cells but is aIso critical for Treg survival and homeostasis. LT-a has been shown to be an important mediator in autoimmune diseases such as multiple sclerosis. LT-a can aiso be expressed by human Thl7 cells [Chiang et al., 2009] and found as a cell surface heterotrimer with LT-P (LT-aip2) or as a secreted homotrimer (LT-aS) [Hansen & Caspi, 2009]. Secreted LT-aS binds to the TNF receptors I and II and shares functional similarities with TNF-a; LT-aip2 binds its own receptor LT-PR and plays a critical rôle in the formation of secondary lymphoid tissues and germinal centers (GC). Interestingly, it was revealed in mice that the IL-12p40 homodimer, which was previousiy thought to be biologically inactive, but not the classic IL-12p70 (IL-12p40:IL-12p35 heterodimer) can induce LT-a expression through IL-12RP1 but not IL-12RP2 [Jana & Pahan, 2009]. TNF is a strong inducer of chronic inflammation and anti-TNF therapy has been show to be effective for treatment of rheumatoid arthritis and other autoimmune diseases