Année 2017
THÈSE / UNIVERSITÉ DE RENNES 1 sous le sceau de l’Université Bretagne Loire
pour le grade de DOCTEUR DE L’UNIVERSITÉ DE RENNES 1 Mention : Biologie
Ecole doctorale VAS
Présentée par Katherine Yaacoub
Préparée à l’unité de recherche INSERM U1242-COSS Chimie, Oncogenèse, Stress et Signalisation, CLCC Eugène Marquis
Thèse soutenue à BIOSIT, Rennes c-FLIP as a potent le 27 Avril 2017 anticancer target : devant le jury composé de : Enhancement of Nathalie RIOUX-LECLERCQ Professeur au CHU de Rennes / Présidente du jury cancer cell apoptosis Sylvie FOURNEL Professeur à l’Université de Strasbourg / rapporteur by compounds Bruno SEGUI Professeur à l’Université de Toulouse/ rapporteur identified through Marie-Thérèse DIMANCHE-BOITREL virtual screening DR2 à l’Université de Rennes 1/ examinateur Vincent FERRIERES Professeur à ENSCR de Rennes/ examinateur Thierry GUILLAUDEUX MCU à l’Université de Rennes 1/ directeur de thèse
Acknowledgements
First of all, I would like to express my sincere gratitude to all the jury members for being a part of my project’s evaluation and for taking the time to read this manuscript.
The success of our work relies heavily on the support and the advices of my thesis director Dr. Thierry Guillaudeux. I would like to take this opportunity to thank you for accepting me as student, for helping me to ameliorate my skills, and for encouraging me to improve my intellectual potential. You trusted me over the past three years and you taught me how a successful and independent student has to be. I am proud to have been your PhD student.
Dr. Rémy Pedeux, let me call you the “Secret Guardian” of my thesis, you have been always ready to answer my questions and always welcoming when I knocked your door ten times a day! Thank you for saving my work when I did many mistakes, and for re-establishing my experiments on the right path.
Pr. Richard Danielllou, Pr. Pascal Bonnet, Dr. Pierre Lafite, and Dr. Samia Aci-Sèche, thank you for giving me the chance to join your laboratory in Orléans and to do an important part of my thesis there. Thank you for your technical advices, constructive remarks and inspiring conversations.
I would like to thank COSS team members for these extra-professional and wonderful moments and for your conviviality and benevolence.
Special thanks for the Lebanese “Association of Specialization and Scientific Orientation”, that offered me a scholarship for higher education during these three years. Without your contribution, I would not have the chance to come to France and get this degree.
I warmly thank all my lovely friends in Rennes, who showed me their immense support. Chaza, Ramona, Nour, Bassil, Nicolas and Emna. You all supported me and listened to my complaints with interest. Thank you for sharing with me these adorable moments in Rennes. Without you I would not have this humorous and amazing stay in France.
I will end by thanking my family who supported me throughout this thesis despite the distance. Precious Mom & Dad, you are the reason of my success, thank you for helping me to strengthen my weakness. Firas & Marcel, you are the most warmhearted brothers in the world, thank you for being always by my side. Uncles, Aunties particularly Helene, and cousins, your support is much appreciated.
THANK YOU ALL…
1
Abstract
FLIP (FLICE Inhibitory Protein) is an anti-apoptotic protein which shares sequence similarity with the pro-apoptotic protein caspase-8. FLIP competes with caspase-8 for binding to the adaptor protein FADD (Fas-associated death domain), thus it inhibits caspase-8 activation, thereby blocking apoptosis. During the development of molecules interfering with anti-apoptotic proteins, searching for inhibitors of FLIP protein which is overexpressed in a very large number of cancers, has failed. This is partly due to the fact that little FLIP structural information is available at present.
TRAIL is a member of TNFα superfamily. It has been described to activate the apoptotic signaling pathways. TRAIL showed great interest in anti-cancer therapy, due to its ability to induce tumor cell death without any effect on normal cells. However, the efficacy of TRAIL is limited by several molecular mechanisms. One of these mechanisms is the overexpression of FLIP which is able to compromise the therapeutic use of TRAIL. The main goal of this project is to develop novel inhibitory molecules able to interfere with FLIP in tumor cells without any effect on the homologous protein caspase 8.
After the construction of FLIP and caspase-8 proteins on the basis of the crystallographic structure of the viral FLIP and FADD respectively, the first docking experiments using a chemical library of the National Cancer Institute NCI have been carried out. The most interesting molecules, being selective for FLIP versus caspase 8, were selected and tested on lung cancer cell lines that overexpress FLIP protein. Co-administration of FLIP inhibitors with TRAIL was performed to verify the restoration of the apoptotic pathway in cancer cells. A molecular test of "Pull down assay" was done in order to confirm the inhibition of the FLIP/FADD interaction. Finally, the evaluation of caspases activity was carried out to confirm the reactivation of the apoptotic machinery after TRAIL/FLIP-inhibitors combination.
In conclusion, the combination of TRAIL with FLIP inhibitors resulted in apoptosis restoration in resistant tumor cells. These newly identified compounds may serve later as potential elements in cancer treatment field.
2
Résumé
Plusieurs protéines anti-apoptotiques sont surexprimées dans les cellules tumorales où elles contribuent à la transformation des cellules cancéreuses et à leur résistance à la plupart des traitements. L’échappement aux mécanismes apoptotiques contribue à la carcinogénèse et à la progression tumorale, mais il participe également à la résistance aux traitements, puisque les différentes thérapies anti-cancéreuses aujourd’hui utilisées que ce soit la chimiothérapie, la radiothérapie ou l’immunothérapie agissent majoritairement en activant les voies de signalisation conduisant à la mort cellulaire et tout particulièrement à l’apoptose. Aussi, ces molécules anti-apoptotiques sont des cibles de choix dans l’élaboration de nouvelles approches thérapeutiques. Des composés ciblant les protéines anti-apoptotiques soit au niveau de leur ARNm (oligonucléotides antisens), soit au niveau protéique (petites molécules inhibitrices) ont été développés et sont actuellement en phase d’évaluation préclinique, voire clinique pour certains d’entre eux. Néanmoins dans cette course au développement de molécules interférant avec les protéines anti-apoptotiques, des inhibiteurs ciblant la protéine anti-apoptotique c-FLI font défaut. Ceci est en partie dû au fait que peu d’informations structurales de c-FLIP sont disponible à l’heure actuelle. FLIP (FLICE Inhibitory Protein) est une protéine inhibitrice qui interfère dans le recrutement des caspases initiatrices 8 et 10 de la mort cellulaire programmée (apoptose) dans la région cytoplasmique des récepteurs de mort activés. Grâce aux fortes identités de séquences partagées entre c-FLIP et les deux procaspases-8/10, c-FLIP est capable d’empêcher leur interaction avec les récepteurs de mort par l’intermédiaire du complexe supramoléculaire DISC (Death Inducing Signalling Complex), bloquant ainsi leur activation. La protéine FLIP possède deux domaines effecteurs de mort (Death Effector Domains DEDs : DED1 and DED2) positionnés en tandem qui miment le prodomaine des procaspases- 8/10. FLIP peut être recrutée avec FADD (Fas Associated Death Domain) via son DED2 au sein du DISC, empêchant ainsi l’activation des β procaspases. Trois isoformes de la protéine cytosolique FLIP ont été caractérisées à ce jour ainsi que 6 protéines homologues virales (v- FLIP) qui permettent ainsi de prolonger la survie des cellules qu’ils infectent. La structure cristallographique récente de v-FLIP a permis de révéler que les 2 domaines effecteurs de mort (DED 1/β) étaient associés l’un avec l’autre de manière très étroite principalement grâce à des interactions hydrophobes conservées. FLIP est une protéine anti-apoptotique extrêmement importante que l’on retrouve surexprimée dans un très grand nombre de tumeurs d’origines tissulaires variées, comme les
3
carcinomes colorectaux, les carcinomes gastriques, les carcinomes pancréatiques, les lymphomes de Hodgkin, les lymphomes B folliculaires, les leucémies lymphoïdes chroniques, les mélanomes, les carcinomes du sein, les carcinomes ovariens, les cancers de l’utérus, ainsi que les carcinomes de la vessie et de la prostate, et elle participe fortement au développement tumoral et à la résistance aux molécules thérapeutiques. De nombreux travaux ont permis de montrer que FLIP était un acteur déterminant dans la résistance à la mort induite par des ligands pro-apoptotiques tels que TRAIL et que la diminution de son expression sensibilisait de nombreuses cellules tumorales préalablement résistantes à la mort. A l’inverse l’expression forcée de FLIP rend les cellules résistantes au TRAIL. Ces observations démontrent bien que FLIP apparait comme une cible thérapeutique de choix, en particulier pour les différents types de tumeurs précités et pour lesquels le caractère malin agressif et la résistance aux agents thérapeutiques sont très étroitement dépendants de la surexpression de cette protéine. En outre, v-FLIP K1γ de l’herpesvirus 8 humain (HHV8, qualifié d’herpesvirus associé au sarcome de Kaposi, KSHV) joue également un rôle oncogénique en inhibant l’apoptose dépendante des récepteurs de mort. TRAIL est une cytokine de la famille du TNF qui est décrite pour activer des voies de signalisation conduisant à la mort cellulaire par apoptose. TRAIL est produit sous forme d’une protéine transmembranaire de β81 acides aminés. TRAIL possède une activité maximale sous sa forme homotrimérique dont la stabilité est due à la présence d’atome de zinc, ce qui le diffère des autres membres de la famille TNF. TRAIL est produit par les cellules du système immunitaire comme les lymphocytes T, les NK, les macrophages et il a montré un grand intérêt dans la thérapie anticancéreuse, grâce à sa capacité d’induire la mort des cellules tumorales sans aucun effet sur les cellules normales, ce qui permet de traiter les tumeurs en minimisant les effets secondaires. Cependant, l’efficacité de TRAIL est limitée par plusieurs mécanismes moléculaires qui aboutissent à la résistance des cellules cancéreuses au TRAIL. La surexpression de FLIP est un des facteurs qui peuvent compromettre l’utilisation thérapeutique de TRAIL en inhibant son action apoptotique au niveau des récepteurs de mort. Ainsi, les molécules ciblant c-FLIP et/ou v-FLIP au niveau ARN messager (oligonucléotides antisens) et protéique (petits inhibiteurs) doivent être développées et testées comme nouvelle classe de molécules anti-tumorales potentielles, mais au-delà de leur intérêt bien avéré en cancérologie, ces composés pourront également être envisagés comme nouveaux traitements adaptés à certaines affections virales. Pour cette raison, le but principal de ce projet est de développer de nouvelles molécules inhibitrices capables d’interférer avec c-FLIP dans les cellules tumorales, sans aucun effet sur la protéine
4
homologue pro-caspase 8, afin de restaurer l’activité apoptotique de TRAIL et induire la mort cellulaire dans les cellules cancéreuses. Vu que la structure cristallographique du domaine d’interaction DEDβ de c-FLIP n’a pas encore été caractérisée, tandis qu’une structure tridimensionnelle a été publiée pour une forme virale de c-FLIP :v-FLIP MC159, alors on a modélisé le domaine DED2 de c-FLIP sur la base de la structure cristallographique de DED2 de v-FLIP MC159. Pareil pour DED2 de caspase-8, on a modélisé son domaine DED2 sur la base de structure cristallographique de FADD. Après la construction des domaines DED2 de FLIP et caspase-8, premières expériences d’ancrage ou “docking” utilisant une base de données virtuelles de composés chimiques (1990 molécules connues pour posséder des propriétés anticancéreuses potentielles) du « National Cancer Institute (NCI) » aux USA ont été effectuées. Ces analyses nous ont permis de mettre en évidence 9 molécules possédant des propriétés compatibles avec les effets escomptés, c’est-à-dire une forte interaction avec le domaine DED2 de c-FLIP impliqué dans la formation du DISC, et pas d’interaction avec DEDβ de caspase-8, alors ces molécules sont nommées sélectives pour c-FLIP versus caspase-8. Les 9 molécules les plus intéressantes, étant comme sélectives pour c-FLIP et non caspase 8, ont été testées sur des lignées de cancer de poumons H1703 surexprimant de manière stable la protéine c-FLIP. La cytotoxicité des 9 composés a été testée par cytométrie en flux, et les concentrations appropriées présentant aucun effet cytotoxique étaient choisies pour faire les tests supplémentaires. Une co-administration de chacune des molécules inhibitrices de c-FLIP avec le ligand de mort TRAIL était faite pour vérifier la restauration de la voie apoptotique dans les cellules cancéreuses. Comme prévu, un avancement de mort cellulaire des lignées cancéreuses était remarqué, ce qui montre que la suppression de la fonction de c-FLIP ainsi la stimulation des récepteurs de mort par leur propre ligand induit la réactivation de l’apoptose, ce qui rend la protéine c-FLIP une importante cible thérapeutique pour le traitement de différents types du cancer. Un test moléculaire de « Pull down assay » a également montré l’effet inhibiteur de nos 9 molécules, en empêchant l’interaction entre les β protéines recombinantes c-FLIP et FADD, ce qui confirme que l’utilisation de ces composés évite le recrutement de c-FLIP au niveau du complexe DISC, et que l’inhibition de l’interaction FLIP/FADD permet la récupération de l’activité de TRAIL en test cellulaire. Finalement, pour vérifier que la mort cellulaire des cellules cancéreuses après combinaison de TRAIL avec les inhibiteurs de FLIP est faite par voie apoptotique, on a étudié l’activité
5
enzymatique des caspases-8, -γ et PARP qui est un marqueur de l’induction de l’apoptose. Alors, on a observé un clivage des caspases et du PARP après la combinaison de TRAIL avec les nouveaux composés, ce qui montre l’activation des caspases responsables du déclenchement du processus apoptotique. Par conséquent, la combinaison de TRAIL avec les inhibiteurs de FLIP aboutit à la restauration de la voie apoptotique dans des cellules cancéreuses. Ces composés nouvellement identifiés, peuvent servir ultérieurement comme des potentiels éléments des stratégies utilisées dans le domaine du traitement du cancer. Nos données nous ont permis de déposer un brevet avec «Fonds de maturation-SATT ouest valorisation» pour les nouvelles molécules que nous avons identifiées, et d'autres expériences sont effectivement en cours afin de consolider et renforcer notre invention.
6
Table of Contents
Acknowledgments……………………………………………………………………………1
Abstract……………………………………………………………………………………….2
Résumé………………………………………………………………………………………..3
Table of contents…………………………………………………………...... 7
List of abbreviations………………………………………………………………………….9
List of figures………………………………………………………………………………...1β
Bibliographic Introduction……………………………………………………………….…1γ
Chapter I: Cancer overview…………………………………………………………….…..14
What is cancer……………………………………………………………………...... 15
I) Cancerogenesis………………………………………………………………15 II) Cancer risk factors…………………………………………………………...18 III) Different types of cancer treatment………………………………………….19 IV) Treatments failure: Role of apoptosis resistance…………………………….β1
Chapter II: The apoptotic machinery…………………………………………………...…β5
I) History of apoptosis research………………………………………………...β6 II) Mechanisms of apoptosis…………………………………………………….β6 A) Different apoptotic pathways…………………………………………….β6 1. The extrinsic pathway………………………………………………..β7 2. The intrinsic pathway………………………………………………...γ0 III) Apoptosis evasion and cancer………………………………………………..γβ A) Transcriptional/translational modifications………………………………γ4 1. Expression of anti-apoptotic proteins………………………………...γ4 2. Suppressing the pro-apoptotic genes…………………………………γ6 B) Post-translational modifications………………………………………….γ8 1. Ubiquitination………………………………………………………..γ8 2. Phosphorylation………………………………………………………γ9 3. Methylation…………………………………………………………..γ9 IV) Death receptors-dependent apoptosis and DISC assembly…………………..40 A) TRAIL’s structure and role in apoptosis…………………………………40 1. Structure and expression of TRAIL………………………………….40 2. Different TRAIL’s receptors…………………………………………4β 3. Physiological roles of TRAIL………………………………………..44 a. TRAIL: a potential factor for cancer therapy…………………….45 b. TRAIL: a regulator of the immune system………………………46 c. TRAIL-mediated necroptosis……………………………………47
7
d. TRAIL-mediated non-cell death pathways………………………48 e. Different TRAIL forms for anti-cancer therapies………………..48 B) FADD: a main signal transducer for death receptors…………………….53 C) Caspases: central players in apoptosis…………………………………...55 1. Caspase-8 dual function in apoptosis and necrosis……………….....57 2. Caspase-8 deregulation in cancers…………………………………...57 D) DED chain and DISC assembly……………………………………….…58
Chapter III: c-FLIP a major inhibitor of the extrinsic apoptotic pathway and a relevant clinical target for cancer therapies…………………..………………………………….…61 I) Different isoforms and structures of c-FLIP………………………………..62 II) Different c-FLIP functions………………………………………………….64 A) Molecular function of c-FLIP in regulating apoptosis………………….64 B) Role of c-FLIP in necroptosis…………………………………………..65 C) Role of c-FLIP in inducing a survival signaling………………………..66 D) c-FLIP role in tissue homeostasis and immune system…………….…..68 III) c-FLIP: elevated level in human cancers…………………………………...69 IV) Modulation of c-FLIP expression…………………………………………..70 A) Regulation of c-FLIP on transcriptional and translational level……….70 B) Post-translational regulation and degradation of c-FLIP……………….71 V) c-FLIP: a critical target for cancer therapies………………………………..7γ A) targeting c-FLIP transcription…………………………………………..7γ B) post-transcriptionally targeting of c-FLIP………………………………74 Project’s objectives………………………………………………………………………….76 Results………………………………………………………………………………………..77 Discussion…………………………………………………………………………………….99 Annexes……………………………………………………………………………………..107
8
List of abbreviations
A AIF: Apoptosis inducing factor APAF-1: Adaptor Protein Apoptotic protease-activating factor 1 AR: Androgen Receptor B BAX: Bcl-2 Associated X Bcl-2: B-cell lymphoma 2 BIR-domains: Baculovirus-IAP-Repeat-domains
C CARP: Caspase-8/10 associated RING Proteins C-MYC: Cellular Myelocytose CRD: Cystein-rich domain CREB: cAMP response element-binding protein D DD: Death Domain DED: Death Effector Domain DIABLO: Direct IAP binding protein with low PI DR: Death Receptor E EBV: Eipsten Barr Virus EDAR: Ectodermal Dysplasia Receptor EGR1: Early Growth Response protein ERK: Extracellular signal-Regulated Kinase G G-CSF: Granulocytes colony stimulating factor
9
H HHV8: Human Herpes Virus-8
I IAPs: Inhibitors of Apoptosis Proteins IRF5: Interferon Regulatory Factor 5 M MALT: Mucosa-Associated Lymphoid Tissue MAPK: Mitogen-Activated Protein Kinase Mcl-1: Myeloid cell leukemia 1 MDR1: multidrug resistance protein 1 MDM2: mouse double minute 2 homolog MLKL: Mixed lineage kinase domain-like protein MMP-9:Matrix Metallopeptidase-9 MOMP: Mitochondrion Outer Membrane Permeabilization mTORC1: Mammalian Traget of Rapamycin complex 1 N NFAT: Nuclear Factor of activated T cells NGFR: Nerve Growth Factor Receptor NK: Natural killer O OPG: Osteoprotegerin P P-gp: P-glycoprotein PUMA: P53 Upregulated Modulator of Apoptosis PKB: Protein Kinase B
R RIP:Receptor Interacting Protein
10
S Smac: Second mitochondrial activator of caspases sp1: Specificity Protein 1 STAT3: Signal transducer and activator of transcription 3 T TNF: Tumor Necrosis Factor TRAIL: TNF-related apoptosis inducing ligand TRAF: TNFR-associated factor X XIAP: X-linked inhibitor of apoptosis
11
List of figures
Figure 1.Carcinogenesis phases...... 17
Figure 2. General principles of drug resistance...... 21
Figure 3. DED-FADD/procaspase-8 structural modeling and chain formation...... 29
Figure 4. Different receptors and Decoy Receptors...... 30
Figure 5. Extrinsic and Intrinsic Apoptosis pathways...... 32
Figure 6. Apoptosis resistance factors...... 33
Figure 7. Crystal complex of trimeric TRAIL binding to DR5 receptors.)...... 41
Figure 8. Clinical trials of TRAIL-DRs targeting therapies...... 49
Figure 9. Domain structures and classification of caspases ...... 56
Figure 10. Structures of CFLAR gene and c-FLIP protein variants...... 64
Figure 11. C-FLIP/caspase-8 dimerization in the DISC...... 68
Figure 12. FADD assembly with c-FLIP and caspase-8 through homotypic DED interactions...... 103
12
Bibliographic Introduction
The aim of this bibliographic part is to synthesize all the knowledge related to our project.
The first chapter describes cancer, cancerogenesis, cancer risk factors, the different existent treatments, as well as the majority of resistance mechanisms.
The second chapter focuses on the apoptotic machinery, particularly the extrinsic pathway of apoptosis. We describe here the function of DISC components in tumor resistance to therapies, in particular to TRAIL. We also develop TRAIL’s mechanisms of action, and its important role in cancer treatments.
Finally, the third chapter focuses on the anti-apoptotic protein c-FLIP. We describe its different molecular functions, its role in cancer development, and its importance in clinical targeted therapies.
13
Chapter I
Cancer overview
14
What is cancer?
Cancer is the uncontrolled growth of abnormal cells. It develops when gene modifications control the function of our cells, especially how they grow or divide. Cancer is a collection of more than 100 diseases in which cells start to grow out of control and crowd out normal cells. Normally, cells grow and divide in an orderly way and they die when become damaged and new cells take their place. When cancer develops, however, this process is breaking down and old or damaged cells survive and divide without stopping to form masses called “Tumors”. Cancer can start anywhere in human body: In the lungs, the breast, the pancreas, the colon, the blood…cancer cells can spread to other parts of the body through the blood or the lymph system to form “metastasis” which can cause severe damages in body’s function, thus most people die of metastatic disease.
Cancer is becoming a major cause of morbidity and mortality in the coming decades in all the world’s regions. Socioeconomic factors are associated with cancer incidence through different pathways. The human development index (HDI), a composite subscript of life expectancy, education and gross national income, is strongly correlated to overall cancer incidence which is predicted to be increasing from 12,7 million cases in 2008 to 22,2 million by 2030 (Bray et al. 2012).
I) Carcinogenesis
Carcinogenesis is a multistep process in which a sequential accumulation of mutations within cells is occurring. Despite the large number of mutations, only a small subset of cells is crucial for neoplastic development (Sjoblom et al. 2006). These mutations result in imbalance of homeostasis as the transformed cells increase their proliferation rate, decrease their death and form a survival-promoting environment. Most of cancers are characterized by a genetic instability observed by a variety of genetic alterations (Duesberg et al. 2000). Experimental evidence showed that only a small proportion of tumor cells is capable of tumor initiation because cancers arise from stem cells or cells that have acquired stem cells properties. These cancer stem cells are the principal cause of tumor progression and heterogeneity as they are extensively proliferating resulting in continuous self-renewing and tumor development (Al- Hajj et al. 2004). Tumor initiates after a process of mutation acquisition which occurs in the DNA of individual cells at any time. Thus, this process causes mistakes in DNA functions
15
which start to copy itself during cell division. This evidence is accorded to Hanahan and Weinberg paradigm who suggest that only a small number of mutations are needed for cancer initiation and these mutations are called “somatic mutations”. They are classified in three different categories:
. R-mutation represents either a gain of proto-oncogenes function leading to an overexpression of proteins controlling cell proliferation, or a loss of tumor-suppressor genes function which regulate cell cycle checkpoints and inhibit cell proliferation. . D-mutation represents the loss of genes responsible for apoptosis promotion. Thus, this mutation diminishes cell death rate and could be correlated with an extensive replication potential. . G-mutation represents the genomic instability which enhances mutations rate, and it is considered as an essential implement to acquire more mutations causing malignancy (Hanahan and Weinberg 2000).
Once cancer is initiated, changes in gene expression take place during the promoting phase with selective proliferation of initiated cells and development of pre-neoplastic cells (Oliveira et al. 2007). Undifferentiated cells and uncontrolled cellular expansion are established, and contribute to instability between growth and cell death leading to malign neoplasia appearance. Tumor masses interact with their microenvironment resulting in ulterior mutations and appearance of other signs of malignancies such as a tumor progression, tissue invasion and metastasis.
Between initiation and promotion, the identified lesions are designated as benign neoplasias. Their transformation into malign lesions in the last stage of carcinogenesis is called tumor progression, in which the neoplasic phenotype is the result of genetic and epigenetic mechanisms. Progression stage is irreversible, and characterized by genetic instability and sometimes by very fast growth. As well, angiogenesis is essential to tumor progression and contributes to malignancy. Its inhibition might be a good tool to delay tumor development (Shacter et al. 2002) (Figure 1).
One of cancer cell characteristics is the ability to propagate and move to distant sites through vascular, lymphatic or transcoelomic routes. These processes of invasion and metastasis resulted from deregulation of several molecules responsible for cell-cell adhesion. Abnormalities in these adhesion structures, including desmosomes, tight junctions (TJ) and Gap junctions, lead to cancer cell dissociation from primary tumor and metastatic spread to
16
other locations. In particular, the adhesion molecule E-cadherin plays a key role in metastasis initiation as loss or disruption of its expression reduces cell-cell adhesion and enhances metastatic dissemination of cancer cells. Metastasis is a complex process that requires increased motility, decreased adhesion and apoptosis resistance. Cellular migration is also essential for metastasis. Migration of cells requires cell surface protein expression and detection of extracellular signals. Treatments that block cell motility pathways represent a powerful tool to control metastatic dissemination. A specific tumor cell migrate to a suitable particular location, for example, breast and prostate cancers metastasize to bone environment, contrariwise, gastrointestinal cancers form lung and liver metastasis. Today, anti-angiogenic therapies showed efficient results, however, anti-metastatic options are not yet identified (Jiang et al. 2015; Wells et al. 2013).
Figure 1.Carcinogenesis phases. The initiation involves genes alterations or mutations resulting in deregulation of cell proliferation, survival and differentiation. The promotion phase is long and reversible, in which preneoplastic cells accumulate. Progression is the final stage of neoplastic transformation in which tumor volume increases rapidly and tumor cells may acquire additional mutations. Metastasis phase represents the displacement of tumor cells to other parts of the body (Siddiqui et al. 2015)
17
II) Cancer risk factors
It is difficult to know exactly why some persons develop cancers and others don’t. Studies showed that many risk factors increase the chances of developing cancer; however, not all people who are exposed to these factors develop cancer. Among risk factors, we cite chemicals and other substances; also they include uncontrolled things like age and family history which is a sign of inherited cancer syndrome. An epidemiology study is required to identify cancer risk factors. Researchers collect a large group of people and compare patients developing cancers with those who don’t. Such studies aim to find the association between an increased cancer incidence and a potential risk factor. The list below represents the most studied risk of cancer:
1. Age: age is used in all studies of cancer epidemiology and it is considered as one of the most studied risk factors. Advancing age is highly correlated with cancer, and the median age of cancer diagnosis is 66 years. The increased cancer rate is due to a prolonged exposure to carcinogens and weakening of body’s immune system. However, cancer can occur at any age. For example, bone cancer is mostly observed in patients under age 20, and 10 % of leukemia are diagnosed in adolescents and children (White et al. 2014). 2. Tobacco smoking: tobacco contains carcinogens that greatly increase the risk of different cancers including lungs, mouth, throat and many others. Smoking people or surrounded by a tobacco smoke environment have a high risk of cancer development because of tobacco contains chemicals that damage DNA (Tsugane 2013). 3. Alcohol: alcohol consumption is linked to increased risk of various cancer types, such as oral cavity, larynx, esophagus, stomach, liver and others. 3,6 % of all cancers are attributable to alcohol drinking (Boffetta et al. 2006). 4. Environmental factors: Pollutants in the air or water such as “asbestos” which is an industrial waste may cause lung cancer and mesothelioma (cancer of pleura). These two diseases develop exudates after a long latency time of asbestos fibers (Raşcu et al. 2016). Radiation: exposure to radiation of certain wavelengths, also called ionizing radiation, is a risk factor to develop cancer due to DNA damage. Extended exposure to Ultra violet radiation, primarily from sunlight or artificial tanning
18
beds are enough to induce skin damage and promote skin carcinogenesis (Sample and He 2016). High energy radiation such as X-rays, gamma rays and alpha particles can damage DNA, and are markedly correlated with increased risk of developing Thyroid tumors (Sholl, Barletta, and Hornick 2017). Exposure to the radioactive gas “Radon” which is released from soil and rock increases the risk of lung cancer (Zoliana et al. 2016). 5. Infectious agents: these agents including viruses, bacteria and parasites promote cancer development in infected persons by disrupting the normal cell growth; as well they cause chronic inflammation which leads to cancer. For example, the human papillomavirus (HPV) is a major cause of cervical cancer in women and anal cancer in men. Vaccination that prevents HPV-associated cancers is recommended for children ages 11 and 12 (Maguire et al. 2017). Some bacteria also cause cancers, including Helicobacter pylori which promotes stomach cancer and lymphomas (S. Yang et al. 2016). Some parasites like Schistosoma haematobium, found in Africa and Middle East, cause chronic inflammation of the bladder leading to cancer promotion (Bernardo et al. 2016).
Despite the progress in chronic diseases treatments, eradication of cancer remains difficult and not effective. Thus, primary prevention through lifestyle and reduction of exposure to environmental risk factors would inhibit a large proportion of cancer deaths.
III) Different types of cancer treatment
Cancer treatment is one of the most complicated aspects in medical care. Many factors are taken into consideration before starting treatment, including the probability of cure or lifetime extension, the side effects and the patient’s wishes. Once cancer is diagnosed, the main goal of treatment is removing the tumor if possible. If cure is impossible, symptoms can be relieved with treatments that improve life quality (palliative therapy). There are many categories of cancer treatment, and they depend on cancer type and how advanced it is. The main existing treatments include:
Surgery is the traditional form of cancer treatment and it is the most effective to eliminate most types of cancer before spreading to lymph nodes or other organs. If cancer is not metastasized, surgery alone is enough to cure the person. In this case, the
19
patient is under risk of cancer recurrence and need chemotherapy and/or radiotherapy after surgery to prevent a recurrence. Nevertheless, surgery is not the best treatment for all cancers because some tumors are localized in inaccessible sites and removing cancer might require removing the whole organ or impairing its function (Ercolano 2017)
Radiation therapy: it uses high-energy particles or waves, such as x-rays, gamma rays, electron beams to destroy cancer cells that divide rapidly and having difficulties in repairing their DNA. Cancer cells divide usually more quickly than normal cells; therefore, they are more likely targeted by radiation. However, some cancer cells are very resistant to radiations (Vuong, Lin, and Wei 2016). Chemotherapy: it is the use of drugs to destroy cancer cells. These drugs can work through the whole body and kill spread cancer cells (metastasized), contrary to surgery or radiation that damage cancer cells in certain area. Most drugs are not that selective, and they kill cancer cells with destroying normal cells, thus chemotherapy drugs affect normal cells and cause side effects such us nausea, vomiting, anemia, increased risk of infections…(Chabner and Roberts 2005) Immunotherapy: it is used to stimulate the body’s immune system against cancer. The immune system is able to detect and kill damaged cells and prevent cancer development. For example, antigens derived from tumors or from viral proteins( Epstein-Barr virus, hepatitis B and C virus, HPV virus), and administered as vaccines, are potentially able to boost the production of immune cells antibodies (B cells) (Tashiro and Brenner 2017). As well, monoclonal antibodies are used to target specific proteins on tumor cell surfaces. For example, Trastuzumab which interacts with HER2 receptor of cancer cells showed a significant reduction of recurrence and death from breast cancers (BAN et al. 2016). Moreover, blocking PD-L1 (Programmed Death Ligand 1) binding to its receptor PD-1, using a specific antibody enhanced immune functions and induced durable tumor regression in patients with advanced cancers including renal-cell cancer, non-small cell lung cancer, and melanoma (Brahmer et al. 2012). Targeted therapy: it is a type of treatment by which drugs can block the growth and the spread of cancer by interfering with specific molecules called “molecular targets” implicated in cancer progression. Targeted therapies are designed to block tumor cell proliferation by interacting with their specific target, whereas chemotherapy drugs are
20
not specific and are cytotoxic by killing cancer cells. One example of targeted therapies is “Vemurafenib” which directly targets and inhibits the mutant form of the BRAF protein in patients with metastatic melanoma (Peng et al. 2016).
IV) Treatments failure: Role of apoptosis resistance
Despite anti cancer drugs improvements and their use in the last decades, some tumors are refractory to drugs, and some patients who respond initially to therapies show a lesser response subsequently, resulting in tumor re-growth. The development of drug resistance results when cancer becomes tolerant to pharmaceutical drugs, either to chemotherapy or to targeted therapy. Resistance to anti cancer drugs may be inherent, due to some genetic characteristics or acquired as a cellular response to drug exposure. Drug resistance is considered as the most convincing cause for cancer therapy discontinuation. Cancers can develop drug resistance through several mechanisms including drug target alteration, altered expression of drug pumps, enhanced ability to repair DNA damage and most interesting mechanism is cell death inhibition (Figure 2).
Figure 2. General principles of drug resistance. After drug administration, the amount reached by tumor cells is limited by pharmacokinetics factors (PK) such as absorption, distribution, metabolism and elimination. In the tumor, pharmacodynamic (PD) factors mean the effects of the drug on the. The anticancer activity of a drug can be limited by poor drug influx or excessive efflux; drug inactivation or lack of activation; alterations such as changes in expression levels of the drug target; activation of adaptive prosurvival responses and a lack of cell death induction due to dysfunctional apoptosis, which is a hallmark of cancer (Holohan C et al., 2013).
21
Drug target alteration: it is possible that during the treatment, the drug target could be modified in many ways, or increased/decreased to levels where it is no longer a useful target. Anti-cancer drugs are designed to disable the activity of components that are necessary to cell survival. So when cancer cells survive the treatment, it means that target genes have been mutated and given a protein that keeps its activity and no longer binds to the drug, thus no more inhibited by it. For example, the anti-cancer drug Doxorubicin targets topoisomerase II, an enzyme that has a vital role in DNA replication. Mutations in topoisomerase II gene which lead to its down-regulation, confer resistance to cancer cells and render them insensitive to Doxorubicin (Di Nicolantonio et al. 2005). Altered expression of drug pumps: One of the most investigated mechanisms of drug resistance is the decreased accumulation of drug into the cancer cells by enhancing the efflux. This action is accomplished by a group of membrane proteins which eject drugs and keep the intracellular drug concentration below a cell-killing threshold. These proteins are members of the ATP-binding cassette (ABC) transporter superfamily, composed of 48 genes encoding ABC transporters. MDR1 gene which encodes a membrane-based pump molecule called P-gp (P-glycoprotein), was the first identified gene and has been extensively studied (Hilgendorf et al. 2007). P-gp extrudes drugs from cells at a rate that surpasses their entry, rendering cells resistant to therapies. Tumors with overexpression of MDR protein, such as hepatocellular carcinomas, leukemia and lung carcinomas show extreme intrinsic resistance (H. Yang et al. 2016; Spolitu et al. 2016; Kong et al. 2016). Moreover, anti-cancer drugs which belong to TNF/Fas ligand family play a prevalent role in apoptosis induction; however, the up-regulation of P-gp confers a resistance to Fas-induced caspase-3 activation and apoptosis (Henkart 1996). DNA damage repair: The repair of damaged DNA plays a potential role in drug resistance. Chemotherapy drugs cause DNA damage, so cancer cells develop an enhanced ability to remove DNA adducts and repair drug-induced lesions through the activation of DNA repair proteins. For example, Cisplatin cause harmful DNA crosslinks leading to apoptosis. However, a resistance to Cisplatin arises due to nucleotide excision repair accomplished by ERCC1 protein (Excision repair cross- complementing) which recognizes and reverses cisplatin-DNA damage. Its expression is found elevated in cisplatin-resistant cells of lung cancer (Bonanno, Favaretto, and Rosell 2014; Olaussen et al. 2006). 22
Apoptosis inhibition: Escape of apoptosis can eventually lead to expansion of neoplastic cells population and it helps the escape of tumor cells from the immune system surveillance. Given that the concept of chemotherapy or targeted therapy is primarily the induction of apoptosis; defects in the apoptotic pathway render cancer cells resistant to therapies. The main features in drug resistance seem to be survival signaling which prevents cell death. So, resistance to apoptosis established an important clinical problem and understanding its mechanism leads to new therapeutic approaches based on apoptosis sensitivity modulation (Hanahan and Weinberg 2000). Tumor cells acquire resistance to apoptosis by several mechanisms that interfere at different levels of the apoptotic signal, including the deregulation of the anti-apoptotic genes.
Bcl-2 is an anti-apoptotic protein which heterodimerizes with the pro-apoptotic members of the BH3 family to prevent mitochondrial pore formation and cytochrome C release, thus inhibiting of apoptotic initiation (Masood, Azmi, and Mohammad 2011). Bcl-2 is found strongly upregulated in small cell lung cancer where it decreased the apoptotic rate conferring resistance to chemotherapy (Sartorius and Krammer 2002). Mcl-1 is a member of the Bcl-2 family. It is considered as a potent anti-apoptotic protein because it is highly expressed in different cancers and its overexpression is correlated with resistance to chemotherapy drugs (Warr and Shore 2008). Mcl-1 is able to block apoptosis induced by different stimuli, including radiotherapy and chemotherapy. Downregulation of Mcl-1sensitizes leukemia cells to Etoposide and induces apoptosis (Jacquemin et al. 2012a). X-IAP is an anti-apoptotic protein which interacts with MDM2 to promote cancer cell survival through binding to p53, promoting its degradation and blocking apoptosis by precluding caspases activation. Small molecules inhibitors of MDM2/XIAP interactions decrease their expression and enhance cancer cell apoptosis by caspase-3/7 activation and p53 stabilization (Gu et al. 2016) One of the other important mechanism by which tumors resist to apoptosis is the overexpression of the anti-apoptotic protein c-FLIP which inhibits death receptors- mediated apoptosis. C-FLIP binds to FADD within the DISC and prevents caspase-8 homodimerization and self-processing, thereby its activation. c-FLIP is
23
upregulated in various cancers including Non-small cell lung cancer (NSCLC) and correlated with shorter survival (Riley et al. 2013a).
Resistance to apoptosis is multi-factorial and it is a complicated mechanism which involves different signaling pathways at different levels. Thus, a comprehensive understanding of the biological networks implicated in apoptosis resistance is required to select treatment strategies. In the last years, strategies that tumors use to acquire resistance are almost known and several mechanisms of apoptosis induction in tumor cells have been discovered. Nowadays, researchers aim to provide new insights into tumor resistance mechanisms by downregulating or inhibiting the anti-apoptotic molecules function, but a major problem is still remaining by selectively modulating tumor cells sensitivity without affecting normal cells.
24
Chapter II
The apoptotic machinery
25
One of the most widely studied field in biology for over 2 % of sciences research has been “Apoptosis”, a form of programmed cell death. Apoptosis plays a key role in development, and apoptosis mistakes are directly associated with numerous diseases including cancer, neurodegenerative disorders, tissue atrophy and auto-immune diseases. Apoptosis was defined as a collection of morphological events such as cell shrinkage, membrane ruffling (blebbing) and packaging of the dying cell into small membrane-bound vesicles called apoptotic bodies. Nuclear events were also observed including chromosome degradation and DNA break up into regular-sized fragments. Exposure of a lipid called phosphatidylserine on apoptotic cell surface is also observed. The apoptotic program occurs into every cell in our body. When cells detect DNA damage, they activate apoptosis to remove theirselves from the population. So apoptosis, a cellular suicide, is an entirely normal function of cells.
I) History of apoptosis research
Apoptosis term was used for the first time in 1972 to describe a form of cell death which is different morphologically from necrosis. However, the concept of natural dying cells is comparatively old and studied by Carl Vogt in toads in 1842. In 1885, the first description of apoptosis morphology appeared, and these morphological hallmarks were regularly reviewed in the literature for over 30 years. 1951, the first review of cell death was published and suggested that apoptosis plays an important role in embryogenesis and vertebrates development. In 1965, the Australian pathologist John F. Kerr observed an unusual form of cell death where the components of cell remain intact within small vesicles and called it “shrinkage necrosis”. In 197β, Kerr team replaced the term “shrinkage necrosis” by “apoptosis” and they defined this particular form of cell death in a paper that was extensively recognized as a master hallmark in biology field (Curtin and Cotter 2003; Kerr, Wyllie, and Currie 1972).
II) Mechanisms of apoptosis
A) Different apoptotic pathways
In the recent years, the molecular machinery of apoptosis was elucidated, revealing a number of molecules which are responsible for the characteristic changes of apoptosis phenomenon. The notion of signal transduction pathways in cell death machinery have been
26
identified, demonstrating the link between environmental stimuli and cell death/survival. To date, two main apoptotic pathways are discovered: the extrinsic or death receptor pathway, and the intrinsic or mitochondrial pathway. These two pathways are linked, and molecules of one pathway can affect the other one (Igney and Krammer 2002).
1. The extrinsic pathway:
It is now recognized as an important mechanism used by NK cells and cytotoxic T lymphocytes to kill virus-infected cells and tumor cells. In mammals, the extrinsic pathway mediates apoptosis in response to the activation of cell-surface death receptors that are members of TNF receptor gene superfamily. Their activation is initiated by TNF-ligands superfamily composed of more than 20 proteins implicated in cell death, survival, differentiation and immune regulation (Avi Ashkenazi 2002). The members of TNF- receptors superfamily are type I transmembrane proteins with an extracellular ligand-binding N-terminal domain, and a C-terminal intracellular domain. TNF receptors share similar “CRD-cysteine-rich extracellular domain” (with up to 65% sequence identity), which defines their ligand specificity. Two subsets of TNF superfamily receptors are known: “The death receptors”, and “Decoy receptors”.
Death receptors: characterized by a cytoplasmic domain of about 80 amino acids called “Death Domain (DD)” responsible for transmitting the apoptotic signal from cell surface to intracellular signaling pathways (Henning Walczak and Krammer 2000). The most broadly studied death receptors are Fas (CD95/APO-1), TNF- Receptor1 (TNF-R1/p55/CD120a), TRAIL-Receptor-1 (TRAIL-R1/DR4), and receptor 2 (TRAIL-R2/DR5/APO-2/Killer). Whereas, the role of DR3 (APO- 3/TRAMP/WSL-1/LARD), EDAR, p75-NGFR and DR6 (TR7), which are not potent inducers of apoptosis, is less known. Activation of Fas, TRAIL-R1 and TRAIL-R2 is often cell death-inducing, however, stimulation of TNF-R1 usually induces cytokine production, inflammation and cell survival (Wajant 2003). Fas, TRAIL-R1 and TRAIL-R2 are activated by their natural ligands, a group of complementary cytokines that belong to TNF protein family, including Fas-Ligand (Fas-L/CD95L/APO1-L) and TRAIL (APO2-L). These cognate ligands are type II transmembrane proteins with C- terminal extracellular domain, transmembrane region, and N-terminal intracellular domain. The extracellular C-terminal region, which is the most homologous between the TNF-superfamily ligands having 30% protein sequence identity, can be
27
proteolytically processed into a soluble form that is released and binds to their cognate receptors, knowing that the soluble form is significantly less potent to induce apoptosis compared to its corresponding membrane-bound form (Shudo et al. 2001; Wajant et al. 2001). Fas-L and TRAIL binding to their appropriate receptors results in receptor trimerization, DD clustering and the recruitment of the adaptor protein FADD. FADD is able to associate with death receptors through homotypic interactions of their DD. FADD also contains DEDs (Death Effector Domains) that mediate the recruitment of cell death effectors, namely the “initiator caspases” as inactive proforms (caspases 8 and caspases 10), as well as c-FLIP( cellular FLICE- inhibitory protein), an inhibitor of caspase-8. The generated complex (Ligands-DR- FADD-Caspases) is called DISC (Death Inducing Signaling Complex), responsible for producing an apoptotic signaling cascade initiated by activated caspases. Upon the interaction of initiator caspases with FADD and DISC formation, a conformation change in the active site of caspases is occurring, resulting in their auto-proteolytic cleavage and activation, and the release of the heterotetramic active form in the cytosol. The heterotetramic active form induces a proteolytic cascade and activates downstream effector caspases such as caspase-3 (Chao et al. 2005). For CD95 signaling pathway, there are two distinct prototypic cell types. In type I cells, the amount of active caspase-8 recruited to the DISC is sufficient to induce directly an apoptotic signaling able to activate caspase-3 within 30 min of receptor engagement, whereas in type II cells, the amount of caspase-8 is too low and not enough to activate caspase-3, so a mitochondrial amplification loop is required for a complete apoptotic signal (C. Scaffidi et al. 1998). DISC formation is an essential step to Death receptor mediated apoptosis; however, the definite mechanism underlying this assembly remains ambiguous. Several studies suggest that DISC components are formed at a ratio of 1:1, whereas a newer study by Dickens team suggested that FADD is sub-stoichiometric relative to death receptors and caspases, and showed that caspase-8 is present at 9-fold more that FADD into the DISC, where caspases-8 interact via their DED domains to form an activating chain (Dickens et al. 2012a) (Figure 3, see also Figure 5).
28
Figure 3. DED-FADD/procaspase-8 structural modeling and chain formation. The colored subunits correspond to the interaction between DED of FADD (Dark Pink) and one dimer of procaspase-8 (Blue and green), resulting in DEDs chain formation (Gray). Caspase-8 catalytic subunit dimer is shown in pink and indicates the formation of antiparallel dimers all along the DED chain. Dotted lines represent the link between procaspases-8-DED2 and p18 subunit of procaspases-8. (Dickens LS et al., 2012)
Decoy receptors DcRs are able to antagonize the apoptotic signal and fail to trigger apoptosis (Riccioni et al. 2005a). They are called “decoy” due to their competitor function to bind ligands of death receptors. Osteoprotegrin (OPG) is a soluble decoy receptor for RANKL/OPGL and probably for TRAIL (Simonet et al. 1997). DcR1 (TRAIL-R3) and DcR2 (TRAIL-R4) resemble TRAIL-R1 and TRAIL-R2 by their extracellular and transmembrane region, however, DcR1 lacks the intracellular death domain and it is Glycosylphosphatidyl-inositol (GPI)-linked plasma membrane receptor, and DcR2 has a truncated non-functional death domain. Therefore, DcR1/2 prevent TRAIL-induced apoptosis in several normal and tumor cells (Riccioni et al. 2005a). DcR3, also called TR6 and M68, is a soluble receptor of TNF superfamily that
29
inhibits CD95-mediated apoptosis and found at elevated levels in cancer cells of many tumor types (Y. Wu et al. 2003) (Figure 4).
Figure 4. Different receptors and Decoy Receptors. To date, eight death receptors DRs have been identified (TNFR1, CD95, TRAIL-R1, TRAIL-R2, DR3 (TRAMP), DR6, EDAR and NGFR), and four decoy receptors DcRs (DcR1, DcR2, DcR3 and OPG). Cystein-rich domains are represented in green (Nikolaev et al., 2009). The immunological ligand of DR6 was still unknown, that is why it was the less studied death receptor. In contrast a new study showed that Syndecan-1, a glycosylated transmembrane protein, is binding partner for DR6 and their interaction results in an autoimmune disease such as systemic lupus erythematosus (Fujikura et al. 2017)
2. The intrinsic pathway
In the intrinsic pathway, also called mitochondrial pathway, caspases activation is promoted by outer mitochondrial membrane permeabilization. Several cytotoxic signals and pro-apoptotic molecules induce mitochondria permeabilization, including Bcl-2 family proteins, mitochondrial lipids, bioenergetic metabolite regulators and components of the permeability transition pore (Green et al. 2004). Once the outer mitochondrial membrane is disrupted, a group of proteins localized between the inner and the outer membrane is released
30
into the cytosol, including cytochrome C, AIF, Smac/DIABLO, Omi/HtrA2 and endonuclease G. These apoptogenic factors promote caspases activation, thereby trigger the apoptotic performance (Saelens et al. 2004). In the cytosol, cytochrome C binds to the C-terminal domain of Apaf-1 protein, which contains an N-terminal caspase-recruitment domain “CARD” able to oligomerize with the initiator caspase-9 through CARD-CARD interactions. These three components form together a complex called “Apoptosome”, which in turn recruits the executioner caspase-3 which is activated by caspase-9. The active caspase-3 is responsible for cleaving key substrates in the cytosol to promote the apoptotic machinery (Bratton et al. 2001). The other released mitochondrial factors are Smac/DIABLO and Omi/HtrA2, which upgrade caspases activation by neutralizing the activity of the endogenous inhibitors of caspases: IAPs. The dimer Smac/DIABLO can bind by its IBM (IAP-Binding motif) one molecule of IAPs family proteins such as XIAP, cIAP1, cIAP2, surviving and Apollon, thereby it prevents IAPs binding to caspase-9 (Yihua Huang et al. 2003). As well, the dimer Omi/HtrA2 promotes the apoptotic pathway in a caspase-dependant manner by antagonizing IAPs function, and in caspase-independent manner as a protease (Wenyu Li et al. 2002) (Figure 5).
Death receptor and mitochondrial pathways of apoptosis can be interconnected at different levels. Upon death receptor activation, the active caspase-8 promotes the cleavage of the pro- apoptotic protein BID, a Bcl-2 family protein with a BH3 domain only, to form a truncated bid, t-BID, which subsequently translocates to the mitochondria and induces cytochrome C release and caspase-9 activation, thereby improving the mitochondrial amplification loop (Ghatage et al. 2012). Moreover, it has been demonstrated that cleavage and activation of caspase-6 after the proapoptotic factors release from mitochondria, may provide a positive feedback to the extrinsic pathway by enhancing caspase-8 cleavage (Cowling and Downward 2002).
31
Figure 5. Extrinsic and Intrinsic Apoptosis pathways. The extrinsic pathway involves the binding of death ligands to their cognate receptors resulting in sequential activation of caspase-8 and -3, which cleaves other substrates leading to cell death. The intrinsic pathway, initiated by several stimuli such as DNA damage or oxidative stress, leads to the permeabilization of the mitochondria membrane and the release of apoptogenic molecules and Apoptosome complex formation. The two pathways are linked by truncated BID, cleaved by caspase-8 and induces the translocation of Bax/Bak to the mitochondria, enhancing the mitochondrial apoptotic pathway. (Beesoo R et al., 2014).
III) Apoptosis evasion and cancer
Carcinogenesis is a very complex mechanism where genetic factors are a major cause. To develop cancer, several steps are occurring progressively, starting from initial genetic mutations, accumulation of a series of genetic alterations, wrongful survival, and a proliferative benefit leading finally to metastasis. During tumorigenesis, cancer cells acquire
32
several properties to overcome the apoptotic cell death that aims to remove all abnormal cells. To survive and proliferate, cancer cells have to surpass all the barriers that serve the apoptotic pathway, either by improving the antiapoptotic machinery, or mitigating the proapoptotic machinery, or maybe both (Figure 6).Thus, it is powerfully demonstrated that evading apoptosis plays a key role in cancer development. The antiapoptotic machinery performed by cancer cells for survival includes transcriptional, translational, and post-translational modifications, helping to evade apoptosis.
Figure 6. Apoptosis resistance factors. Different resistance mechanisms acquired by cancer cells are described at different levels of the extrinsic and the intrinsic pathways. An upregulation of Decoy receptors, c-FLIP, IAPs family proteins, and Bcl-2 /Bcl-XL proteins prevents the apoptotic machinery, however, a downregulation of p53, PUMA, BAX and cytochrome C factors are responsible for cell death inhibition (Picarda G et al., 2012).
33
A) Transcriptional/ Translational modifications
These abnormal modifications, which are characteristics of cancer cells, include gene amplification or deletion, gene silencing by DNA methylation and dysregulation of some transcription factors that play a role in apoptotic molecules expression. As well, miRNAs target the γ’UTR region of mRNAs and negatively control their expression, so they could decrease or increase cell death depending on their target messenger RNA (Kumar and Cakouros 2004; S. Song and Ajani 2012)
1. Expression of anti-apoptotic proteins
C-FLIP (cellular FADD-Like interleukin-1 -converting enzyme-like protease) is found at high levels in human melanomas and murine B-cell lymphomas. It blocks apoptosis induction at death receptors levels (A Krueger et al. 2001). In addition, an increased of c-FLIP/caspase- 8 ratio was correlated with inhibition of CD95-mediated apoptosis in EBV-positive Burkitt’s Lymphoma (Tepper and Seldin 1999). The viral form of the protein (v-FLIP), encoded by HHV8 viruses, is found highly expressed in Kaposi’s sarcoma (Stürzl et al. 1999). c-FLIP can be atypically expressed upon cellular stress. For example, Hyperoxia (high oxygen tension) has been reported to promote the upregulation of c-FLIP which inhibits apoptosis by suppressing both extrinsic and intrinsic pathways, the latter one via BAX inhibition. Expression of c-FLIP increased the phosphorylation of p38-MAPK, leading to BAX increased phosphorylation and inactivation, thereby protection against intrinsic cell death (X. Wang et al. 2007).
PEA-15, also called PED, is a death effector domain DED-containing protein, expressed in central nervous system particularly in astrocytes, blocks CD95-, TRAIL-, TNFα-mediated apoptosis at the extrinsic pathway by disrupting caspase-8/FADD interactions. It has been demonstrated that PEA-15 is implicated in promoting AKT-dependant chemoresistance in human breast cancer cells (Stassi et al. 2005)
A large proportion of tumors display an increased expression of IAPs proteins which are considered as cell death regulators by binding to caspases and interfering with apoptotic signaling pathways either via death receptors or mitochondrial pathway. IAPs share one to three common domains called BIR-domains allowing them to bind caspases and other proteins (Obexer and Ausserlechner 2014). XIAP is the most potent and described IAP family member because it is the only one that inhibits caspases by direct physical interaction. XIAP
34
is an inhibitor of Smac/Diablo released from mitochondria and prevents XIAP/caspases interactions, thereby activating apoptosis. XIAP contains three BIR domains (BIR1-3) in the N-terminal half. BIR-1 interacts with proteins regulating the NFκB proteins, while BIR-2 and BIR-3 are responsible for interactions with either caspase-3 and -7 (BIR-2) or caspase-9 (BIR- 3) respectively (Eckelman, Salvesen, and Scott 2006). Another IAP family member, cIAP2 is deregulated by the translocation of t(11;18) (q21;q21) and it is found at high levels in 50 % of MALT cancers (Dierlamm et al. 1999). cIAP2 binds caspase-3 and -7 but does not efficiently inhibit them by physical interaction but drives them to proteasomal degradation (Choi et al. 2009).
Cancer cells frequently overexpress anti-apoptotic proteins, notably Bcl-2 family proteins which prevent the mitochondrion outer membrane permeabilization (MOMP). Bcl-2 was identified first in B cell follicular lymphoma where a genetic translocation t (14/18) of the Bcl-2 gene led to its high expression, which is associated with poor prognosis and apoptosis resistance (Kogan et al. 2001). The tumor associated viruses such as EBV and HHV8 encode proteins homologous to human Bcl-2, BHRF1 and KSbcl-2 respectively. These two proteins contribute to apoptosis resistance in infected cells, thereby lead to tumor formation (Tarodi, Subramanian, and Chinnadurai 1994; Sarid et al. 1997).
The amplification of Mcl-1 gene, an inhibitor of the intrinsic pathway of apoptosis, has been found in various cancers including lung cancer (Beroukhim et al. 2010). The hyperactivity of PI3K/AKT pathway in tumor cells leads to transcriptional regulation and activation of CREB or STAT3 factors, which are transactivators of Mcl-1 gene, thereby promoting its expression. Moreover, it has been demonstrated that mTORC1, which is a downstream target of PI3K/AKT signaling, enhances Mcl-1 mRNA translation in a mouse lymphoma model (Mills et al. 2008).
Another mechanism by which cells acquire resistance to apoptosis is the abnormal expression of decoy receptors that represent an alternative mechanism of resistance to TRAIL- or CD95- induced cell death. Decoy receptor 3 (DcR3) is a soluble receptor that binds to CD95-L and inhibits its apoptotic action. It is amplified in several lung and colon carcinoma, the same in several Adenocarcinomas and Glioblastomas. Ectopic expression of DcR3 in a rat Glioma model contributes to a lower immune cell infiltration indicating that DcR3 is implicated in immune evasion (Roth et al. 2001). As well, DcR1 (TRAIL-R3), decoy
35
receptor for TRAIL, was reported to be overexpressed in gastric carcinomas (Sheikh et al. 1999).
2. Suppressing the pro-apoptotic genes
Genotoxic or cytotoxic stress can induce the expression of pro-apoptotic genes in normal cells. However, this mechanism is often abolished in tumor cells by mutations, silencing and downregulation of these pro-apoptotic molecules.
P53 is a considered as a critical transcription factor that regulates the expression of a group of genes known to induce apoptosis; thus, alterations of the p53 pathway affect the sensitivity of tumors to apoptosis (Ryan, Phillips, and Vousden 2001). Specific mutations of TP53 gene are correlated with resistance to Doxorubicin treatment and relapse in breast cancer patients (Hientz et al. 2015). Mutations and deficiency of Trp53 (the gene that encodes p53 in mice) showed a poor response to -Irradiations and chemotherapies in mice (Lee and Bernstein 1993). TP53 (the gene that encodes p53 in humans) is mutated in all of the major histogenetic groups, including lung, colon and breast cancers. TP53 is the most routinely mutated tumor suppressor gene in human cancers and account for more than 50 % of all cases (Feki and Irminger-Finger 2004). Several apoptotic pathways induced by p53, including BAX signaling, may be widely affected in cancer cells after p53 mutations. Thus, when p53 is lost in cancer cells, the threshold of MOMP will be elevated by limiting BAX expression and reducing the probability of its oligomerization (Mihara et al. 2003).
The pro-apoptotic Bcl-2 family member BAX is mutated in certain types of cancer. Mutations in the BH-domain that result in loss of function, also frameshift mutations that decrease its expression are common in tumor cell lines which become more resistant to apoptosis (Meijerink et al. 1998). In BAX-deficient mice, a low percentage of apoptotic cells is detected, as well as tumor growth is accelerated, indicating that BAX is essential for p53- mediated apoptosis (Yin et al. 1997).
BIM, a BH3-only protein is a member of Bcl2-family and it is induced upon growth factors retraction and other apoptotic stimuli. Different transcription factors are implicated in BIM expression, including FOXOs which have a particular interest because its activity is abolished by two major prosurvival kinases: AKT (also called PKB) and ERK. Thus, BIM expression is suppressed in cancer cells by high levels of AKT and ERK, contrariwise,
36
downregulation of AKT and ERK by TKIs results in BIM re-expression and apoptosis induction (Gilley, Coffer, and Ham 2003; Costa et al. 2007).
Another BH3-only protein is identified, PUMA, as one of the proapoptotic Bcl-2 family proteins that is extremely deleted in cancers. PUMA is a p53 target gene, but FOXO can transactivate PUMA independently of p53. A recent study showed that inhibition of PI3K/AKT pathway using TKIs, promotes FOXO expression, which in turn activates PUMA gene and triggers apoptosis. Given that BIM is also regulated by FOXOs, how PUMA and BIM are separately and independently regulated remains not well understood (Bean et al. 2013).
Among the numerous strategies to cell death resistance, an absence of surface expression of death receptors is identified. The expression of CD95 death receptor is restrained in many tumor cells, including hepatocellular carcinomas. Downregulation of CD95 gene transcription leads to CD95 expression loss, contributing to chemoresistance and immune evasion. Loss of CD95 in hepatocellular carcinomas is linked to abnormalities in p53 gene which appears to have a major role in CD95 expression in hepatocytes (Volkmann et al. 2001). Moreover, point mutations and deletions in the cytoplasmic death domain (DD) of CD95 drive to a truncated death receptor formation. This latter might prevent CD95-L mediated apoptosis and increases the risk of developing cancers (Straus et al. 2001)
Mutations in TRAIL receptors (TRAIL-R1/DR4 and TRAIL-R2/DR5) have also been detected in several cancers including metastatic breast cancer (Shin et al. 2001). An allelic loss of chromosome 8p21-22, where these TRAIL receptors are located, is frequently occurring in cancers including Non-Small Cell Lung Cancer (NSCLC) and head and neck squamous cell cancer. This genetic aberration results in two missense alterations in the ectodomain of the receptors and amino acid changes in or near the ligand-binding domain, affecting this way TRAIL ability to bind on its receptors (Fisher et al. 2001). Other TRAIL receptor mutations are localized in the death domain and induce conformational changes including reduced binding affinity of DRs to FADD, decreased exposure of FADD-DED for caspase-8 binding, thus notably decreasing the ability of cancer cells to undergo TRAIL- mediated apoptosis (W. S. Park et al. 2001). Moreover, signaling via TRAIL receptors can be impaired by downregulation of their expression as part of an adaptive stress response. An abnormal transport of TRAIL-R1/R2 from intracellular stores such as the endoplasmatic
37
reticulum, to cell surface provided colon cancer cells a resistance to TRAIL-mediated cell death (Jin et al. 2004).
The expression or the function of caspase-8 is impaired by genetic or epigenetic mechanisms in various cancers. The gene for apoptosis initiator caspase-8 is frequently inactivated by homo- or heterozygous genomic deletions in neuroblastoma with amplification of the oncogene N-Myc. Neuroblastoma cells become resistant to death receptor- and Doxorubicin-mediated apoptosis (Kidd et al. 2000). An alternative splicing of intron 8 in caspase-8 gene results in producing a novel inhibitor of itself, named caspase-8L, a catalytically inactive splice variant, whose overexpression confers neuroblastoma cells a protection against TRAIL- and not Etoposide-mediated apoptosis (Miller et al. 2006). As well, the small variant caspase-8L is found in progenitor Hematopoietic Stem Cells (HSC) and Leukemia cells, where it is recruited to the DISC after CD95 triggering, thereby preventing the activation of caspases cascade and apoptosis machinery (Mohr et al. 2005).
It is well-documented that c-FLIP is a key regulator of caspase-8 activity at DISC level (J Tschopp et al., 1998). When highly expressed, the three isoforms of c-FLIP (c-FLIP L, c- FLIP S and c-FLIP R) are recruited to DISC and inhibit death receptor-mediated apoptosis by inhibiting caspase-8 activation and processing (Andreas Krueger et al. 2001). In the absence of full processing, caspase-8 remains restricted to DISC and cannot induce apoptosis (Kavuri et al. 2011)
B) Post-translational modifications
Regulation of apoptosis is not only modulated by transcription or translation factors, but also by post-translational changes including Ubiquitination, Phosphorylation, and Methylation. Ubiquitination of a protein can modify its stability and it is forcefully regulated by ubiquitin/proteasome pathway (Vucic, Dixit, and Wertz 2011). In contrast, the other forms of post-translational modifications act as epigenetic-like codes which modulate and change specific functions of the apoptotic proteins (Dai and Gu 2010).
1. Ubiquitination
Ubiquitination (or ubiquitylation) is essential for all the intracellular processes in eukaryotes. Ubiquitin and UBLs (Ubiquitin-Like proteins) are responsible for this
38
mechanism; they bind to specific proteins, affect their function, and target them to the proteasomal degradation. The ubiquitin/proteasome pathway involves a set of biochemical events mediated by ubiquitin-activating enzyme E1, ubiquitin-conjugating enzyme E2, and ubiquitin ligase E3 (Weissman, Shabek, and Ciechanover 2011). The anti-apoptotic protein Mcl-1, described above, is characterized by its rapid turnover. To date, five E3 ligase have been identified to ubiquitinate Mcl-1 for degradation; the most important one is TRIM-17. However, impairment of Mcl-1 phosphorylation by point mutations or kinase inhibition blocks its interaction with Trim-17, thereby decreased its ubiquitination and stabilized Mcl-1 expression resulting in MOMP blockage (Magiera et al. 2013). Cancer cells may also suppress the accumulation of pro-apoptotic proteins such as BIM. BIM is characterized by a short life time; its stability is regulated by ERK phosphorylation on Serine 69 (S69). This phosphorylation upgrades BIM degradation via ubiquitin/proteasome pathway (Ley et al. 2003)
2. Phosphorylation
Phosphorylation plays a key role in Mcl-1 stability. Phosphorylation of T163 by ERK led to Mcl-1 stabilization. An upregulation of ERK activity was found in breast cancer tissues (Ding et al. 2007). Moreover, caspase-8 is considered as a new substrate for Src kinase, which phosphorylates it on Tyr380, located between the large and the small subunit of caspase-8. Tyr380 phosphorylation, frequently occurring in human colon cancer where Src is highly activated, results in downregulation of caspase-8 apoptotic function, thereby impairing Fas-inducing apoptosis (Cursi et al. 2006).
3. Methylation
DNA hypermethylation is one mechanism by which death receptors expression can be downregulated. For example, the methylation status of 28 CpG sites in the promoter region of the death receptor Fas gene correlates with suppression of Fas expression in colon carcinoma cell lines. Inhibition of this methylation upregulates Fas expression and sensitizes cells to apoptosis (Petak et al. 2003). A recent investigation has demonstrated that TRAIL resistance may be due to a hypermethylation of the DR4 gene. Indeed, a low expression of DR4 of patients with lung squamous carcinomas is associated with CpG-island hypermethylation of
39
DR4 gene promoter (W. Wang, Qi, and Wu 2015). Moreover, epigenetic silencing of caspase- 8 expression, by gene hypermethylation, is found in Small Cell Lung Cancer (SCLC) cell lines and it is responsible for acquiring resistance to TRAIL-cell death (Hopkins-Donaldson et al. 2003).
IV) Death receptor-dependant apoptosis and DISC assembly
As mentioned above, binding of the corresponding ligand to each DR (FasL for Fas, TRAIL for TRAIL-R1/R2 and TNF for TNF-R1) leads to the formation of DISC, consisting of DRs Oligomerization and recruitment of multiple molecules of FADD, procaspase-8 and/or -10, and c-FLIP. Into the DISC, interactions between its components are based on two different types of homotypic links: one is between DDs of DRs and FADD, and the other is between DEDs of FADD, procaspase-8/-10 and c-FLIP. Thus, formation of DISC is a platform for death receptor-mediated apoptosis, yet the mechanisms underlying its assembly is still unclear (J. K. Yang et al. 2005a).
A) TRAIL’s structure and role in apoptosis
1. Structure and expression of TRAIL
The gene encoding for TRAIL (Apo-2L) is localized on chromosome 3 at position 3q26, which is not close to any other member of TNF superfamily. TRAIL is composed of 281 amino acid (32 kDa), initially identified in 1995. It was cloned based on the sequence homology of its extracellular domain with CD95-L (28% identity) and TNF (23% identity) (Wiley et al. 1995). TRAIL is a type II transmembrane protein, which is anchored to the plasma membrane and presented to the cell surface. The C-terminal extracellular domain can be cleaved by cysteine protease, unlike TNF-α which is cleaved by metalloproteinases, resulting in the release of a biologically active 24 kDa soluble form comprising amino acids 114-281 (Kimberley and Screaton 2004). Like most other members of TNF superfamily, TRAIL forms homotrimers that bind three death receptors, and each receptor interacts with the crevice formed between two monomers of the trimer. Unlike other TNF members, an internal zinc atom which interacts with three cysteines residues, one from each TRAIL
40
monomer, is crucial for maintaining the stability, solubility and optimal biological activity of the trimeric TRAIL (Hymowitz et al. 2000a). (Figure 7)
Figure 7. Crystal complex of trimeric TRAIL binding to DR5 receptors. A) a side view of TRAIL/DR5 complex. Yellow, pink and turquoise represent TRAIL monomeric subunits; Red, green and blue represent three DR5 monomers. This figure shows the “ligand induced trimerization model” in which the trimeric TRAIL recruits three receptors. Receptors are positioned at the interfaces between two ligand monomers. B) The same complex shown in A but represented in a top view (Mongkolsapaya et al., 1999).
TRAIL displays a widespread expression, and it is upregulated on activated cells of the immune system such as T lymphocytes, Natural killer (NK), monocytes, macrophages, dendritic cells, and neutrophiles (Tecchio et al. 2004). It appears probably that other cytokines called “Interferons” have a key role in stimulating immune cells to produce TRAIL. For example, TRAIL expression on liver NK cells is regulated by Interferon Gamma INF- secreted from NK cells in an autocrine manner; because it has been found that a large proportion of NK cells produce both TRAIL and INF- in wild type and T cell-deficient mice (Takeda et al. 2001). Moreover, it has been demonstrated that a co-culture of Chronic Myeloid Leukemia (CML) neutrophiles and Peripheral Blood Mononuclear Cells (PBMCs) with INF-α at therapeutic doses stimulates the expression of high levels of TRAIL mRNA and the release of high amounts of a soluble active form of TRAIL (sTRAIL) from INF-α-
41
activated neutrophiles and monocytes (Tecchio et al. 2004). In contrast to other TNF family members, whose expression is only found on activated cells, TRAIL mRNA is broadly expressed in different tissues other than immune cells, including spleen, lungs and prostate (Wiley et al. 1995).
2. Different TRAIL’s Receptors
TRAIL has a potent ability to trigger apoptosis in a large variety of tumor cell lines, but not most normal cells, lighting its potential therapeutic role in cancer treatments (Seki et al. 2003a). TRAIL induces apoptosis through interacting with its receptors. So far, five distinct cognate receptors have been identified. Initially, DR4 (TRAIL-R1) was the first identified receptor, expressed at the same levels on malignant and normal cells. This equal expression does not explain TRAIL sensitivity against tumor cells (Pan et al., 1997). A second receptor named DR5 (TRAIL-R2/TRICK-2/KILLER) was then identified using the sequence of the intracellular domain of DR4 in an expressed sequence tag (EST) database search. DR5 is non- glycosylated protein, presents 58% overall homology to DR4, the greatest homology in the intracellular DD. DR5 was also widely expressed on normal and malignant cells (Screaton et al. 1997). DR5 is expressed on a wide range of tissues and it is up-regulated on activated lymphocytes. DR5 has two distinct spliced forms: DR5A (TRICK-2A, a short form) and DR5B (TRICK-2B, a long form). The difference between these two variants is the presence of a 23 amino acid extension between the transmembrane domain and the start of CRDs (the long form). The two isoforms do not have different functions. DR4 does not contain this extension and it is more similar to DR5A (Screaton et al. 1997). The main function of DR5 is mediating cell death on several different cell types. Specifically, treatment with DNA- damaging molecules enhances DR5 expression and sensitizes human acute leukemia cells to TRAIL cell death (Wen et al. 2000). A link between DR5 expression and p53 tumor suppressor gene has been found. An upregulation of DR5 expression following -Irradiation is a p53-mediated event. The link between these two factors is due to the presence of p53- response elements in DR5 promoter (Takimoto and El-Deiry 2000). DR4 has a similar functional story of DR5, and plays a similar immune surveillance role, since it has a main and crucial function in apoptosis induction. Numerous mutations of DR4 have been found in cancer cells; for example, a polymorphism in the ligand binding domain of DR4 is correlated with higher incidence of bladder cancer (Hazra et al. 2003), and a high expression of DR4 is associated with a favorable prognosis in colon cancer (Sträter et al. 2002). DR4 and DR5 share high similar structures, and they are both able to induce TRAIL-mediated apoptosis.
42
Nevertheless, different functions between DR4 and DR5 have been reported. First, DR5 has a higher affinity to bind TRAIL than DR4, however, the high affinity does not mean an enhanced DISC activation: soluble TRAIL (sTRAIL) interaction with DR5 triggers a weak DISC formation, thus it requires further cross-linking of sTRAIL, whereas, sTRAIL binding on DR4 is able to induce apoptosis independently of further cross-linking of TRAIL molecules (Truneh et al. 2000; Mühlenbeck et al. 2000). Second, DR4 appears to mediate cell death in chronic lymphocytic leukemia cells, acute myelogenous leukemia cells and pancreatic cancer cells, whereas DR5 induces apoptosis in many other epithelial-derived cancers (Sean K Kelley et al. 2004). This differential action mode between DR4 and DR5 depending on cell or cancer type might be an important therapeutic point to specifically target the corresponding receptor and induce apoptosis in a particular cancer type.
Later on, another EST database search was applied, using the extracellular domain of death receptors where TRAIL-binding region is localized, and two other receptors were identified and named Decoy receptors: DcR1 (TRID) and DcR2 (TRUNDD) (M MacFarlane et al. 1997). DcR1 does not contain a DD and it is anchored to the membrane via GPI tail. DcR1 is not widely expressed like other TRAIL receptors; its transcripts are found mostly on peripheral blood lymphocytes (PBLs). It has a low signaling capacity, and its overexpression blocks apoptosis (Horejsí et al. 1999). DcR2 shares high homology with other TRAIL receptors (58-70 %), and has a broader expression than DcR1. It has a truncated intracellular domain lacking 52 amino acids that encode the DD (H. Walczak et al. 1997). It has been demonstrated that the expression of DcRs is also enhanced in response to p53-sensed DNA damage, but this evidence does not fit with the immune surveillance role of DRs in removing damaged cells because up-regulation of DcRs can neutralize this event. So it is suggested that during p53 activation, the weak signals of DcRs are amplified via DR4/DR5 promoters due to their location on the same chromosomal region 8p21-22 (Meng et al. 2000).
Finally, the fifth receptor of TRAIL was identified, OPG (Osteoprotegerin), which is the only soluble form of TRAIL receptors. It is highly glycosylated and secreted as a disulfide linked dimer. A small proportion of OPG is found as a monomeric form (Emery et al. 1998). Unlike other TRAIL receptors, OPG contains four CRDs, and plays a role in osteoclasts activation in bone remodeling. OPG ligand, another member of TNF superfamily (also called RANK-L) binds to RANK receptor (receptor activator of NFκB) and induces osteoclastogenesis. However, OPG competes with RANK for RANK-L binding, resulting in nonlethal osteopetrosis (Simonet et al. 1997). The role of OPG as a TRAIL receptor is less
43
understood, but it has been suggested that it serves as a prosurvival factor by binding to TRAIL and blocking apoptosis (Pritzker, Scatena, and Giachelli 2004). For example, OPG is implicated in breast cancer progression. Breast tumor cells highly express OPG which via sequesterization of TRAIL, provides to tumor cells more aggressive growth and bone metastatic potential (Zauli et al. 2009).
Previously, it was supposed that TRAIL binds to all its receptors with the same affinities. But later, it was demonstrated that the rank order of affinities is robustly temperature- dependant. DR4, DR5, DcR1 and OPG have the same binding affinity to TRAIL at 4°C. However, the rank of affinities was substantially changed at 37°C, with DR5 owing the highest affinity (KD= 2 nm), and OPG the weakest one (KD=400 nm) (Truneh et al. 2000).
Expression of Decoy receptors was first found to be restricted to normal cells, while cancerous tissues preferentially express TRAIL-R1 and TRAIL-Rβ conferring TRAIL’s selectivity. So, it was assumed that the presence of DcRs on normal cells can divert away TRAIL from DRs, whereas their absence on cancer cells keeps cells susceptible to TRAIL- mediated cell death (Griffith et al. 1998). However, later studies demonstrated that many tumor tissues such as breast cancer (Ganten et al. 2009), prostate cancer (Koksal et al. 2008), and acute myeloid leukemia (Riccioni et al. 2005b) have shown a high expression of DcRs and their expression is associated with poor prognosis. In contrast, DcRs expression in tumor cells is not linked to TRAIL sensitivity (Dyer et al., 2007), and normal cells do not require DcRs to be protected from TRAIL-mediated apoptosis, suggesting that the role of DcRs is more complicated than originally thought and requires intracellular signaling regulation (van Dijk et al. 2013).
3. Physiological roles of TRAIL
TRAIL is highly expressed by NK and NK-T cells, thus it was broadly reported that TRAIL has an important role in immune surveillance against virus infected cells and cancer cells. The discovery of TRAIL’s ability to kill cancer cells while sparing normal cells represents a promising approach in the development of cancer targeted therapies.
44
3. a. TRAIL: a potential factor for cancer therapy
TRAIL has a specific role in immune-mediated tumor elimination. Unlike FasL and TNF which can trigger apoptosis in normal cells, TRAIL has a high specificity toward malignant cells. Moreover, in contrast to FasL which causes fulminant and sudden hepatic failure, and TNFα which presents an inflammatory response, administration of TRAIL is systematically safe (O Micheau et al.,2013). Since its identification in 1995, TRAIL has shown anti-cancer properties in vitro by inducing apoptosis in a wide variety of cancer cells of diverse origin (Wiley et al. 1995)
Mice express only one death-inducing receptor called mTRAIL-R (MK/mDR5) sharing 43% and 49% sequence homology with human TRAIL-R1 and –R2 respectively. In addition, mice express two decoy receptors ( mDcTRAIL-R1 and mDcTRAIL-R2) distinct from human decoy receptors (G. S. Wu et al. 1999; Schneider et al. 2003). Importantly, the use of soluble recombinant TRAIL in animal models such as mice or primates, induced a significant tumor regression without any toxicity (A Ashkenazi et al. 1999a). Besides, the potent anti-cancer function of TRAIL was demonstrated by Xenograft of B cell lymphoma in TRAIL deficient mice (Lisa M. Sedger et al. β00β), or by using specific antibodies that neutralize TRAIL (Seki et al. 2003b). These two studies demonstrated that tumor cells develop much faster and form liver metastasis in TRAIL deficient mice compared to WT BALB/c mice. Moreover, the anti- tumor activity of TRAIL was observed in mice xenografted with human multiple myeloma (MM) cells. TRAIL administration for 14 days was well tolerated and significantly reduced plasmacytomas growth. A co-administration with proteasome inhibitor PS-341 increased the pro-apoptotic activity of TRAIL against TRAIL-sensitive MM cells by sensitizing DR5 expression (Mitsiades et al. 2001). Studies also suggest that TRAIL combination with other anti-cancer drugs increases its potency against tumor cells. Treatment of human leukemic cells with Etoposide, Ara-C or Doxorubicin followed by TRAIL, induced significantly more apoptosis that treatment with each molecule alone. This phenomenon is attributed to an increase in DR5 levels which are usually upregulated in response to DNA damages (Wen et al. 2000). In 2008, Henning Walczak’s team have demonstrated that TRAIL-R deficient mice developed lymph nodes metastasis, derived from primary epithelial skin tumors, thereby suggesting that TRAIL-R is a metastasis suppressor gene (Grosse-Wilde et al. 2008). Another study has showed that TRAIL-R in mice is able to suppress inflammation and tumorigenesis. Mono- or biallelic loss of TRAIL-R decreased median lymphoma-free survival, increased metastatic potential and led to apoptosis defects (Finnberg, Klein-Szanto, and El-Deiry 2008)
45
In most human cancers, the tumor suppressor gene p53 is inactivated resulting in therapies resistance. However, TRAIL can kill these cancer cells regardless of p53 status, thus TRAIL is a useful therapeutic tool particularly in cells with a p53 downregulation (Nagane, Huang, and Cavenee 2001). During malignant transformation, TRAIL sensitivity is enhanced by a cell-autonomous manner via intracellular changes rather than cell-extrinsic factors, such as INF secretion by immune cells (Lawrence et al. 2001; van Dijk et al. 2013). The transformation from premalignant colorectal cells to colorectal carcinoma cells is associated with TRAIL sensitivity enhancement. Yet, the underlying mechanism is not fully understood (Hague et al. 2005). Substantially, malignant transformation-driven TRAIL sensitivity is a main promoter of tumor immune elimination. For example, TRAIL deficient mice were more sensitive to the carcinogen Methylcholanthrene (MCA) and more susceptible to develop fibrosarcoma (Cretney et al. 2002).
3. b. TRAIL: a regulator of the immune system
While the activity of TRAIL as an anti-tumor factor has been extensively explored, other studies focused on the role of TRAIL in immune system regulation. TRAIL function in the immune system got the attention when it was discovered as it is expressed on a variety of innate and adaptive immune cells. Its expression depends on the level of immune cells activation; for example, it is upregulated on monocytes and macrophages after INF- and LPS (lipopolysaccharide) stimulation (Ehrlich et al. 2003).
TRAIL and FAS-L are important for the regulation of T helper lymphocytes (Th1/Th2). TCR stimulation using a monoclonal antibody anti-CD3 increased Fas-L expression on Th1 and TRAIL on Th2. This expression of death ligands contributes to an auto-apoptotic process in Th1 and Th2 and it is considered as a critical step to a selective removal of differentiating T helper cells (Roberts et al. 2003)
TRAIL has also an important role in hematopoiesis regulation. TRAIL receptors are not expressed on hematopoietic progenitor cells (CD34+ cells), thus they are protected from TRAIL-induced apoptosis (Zauli et al. 2006). In addition, immature erythroblasts are more sensitive to TRAIL-induced apoptosis than the mature ones. A high expression of TRAIL was found in bone marrow which leads to immature erythroblasts death causing aplastic anaemia disease (Kakagianni et al. 2006). In contrast, a downregulation of TRAIL-R1/R2 and TRAIL
46
was found in patients with multiple myeloma where the erythropoiesis is enhanced (Grzasko et al. 2006).
Importantly, a study of TRAIL Knockout mice showed the crucial role of TRAIL in autoimmunity. An enlarged thymus was found, attributing this to a defect in thymocytes apoptosis, thereby confirming the role of TRAIL in thymic deletion. TRAIL deficient mice had a high number of immature thymocytes and failed to induce apoptosis of T cells in vivo and in vitro. They also demonstrated the role of TRAIL in mediating auto-immunity, as they have observed that TRAIL deficient mice are more susceptible to develop collagen-induced arthritis and streptozoticin-induced diabetes (Lamhamedi-Cherradi et al. 2003). Moreover, TRAIL has a crucial role in autoimmune thyroiditis. Mice treatment with the recombinant TRAIL led to a milder form of the disease with a significant decrease of mononuclear cell infiltration in the thyroid and less thyroid follicular destruction. Thus, these data supposed that TRAIL suppresses the development of the autoimmune thyroiditis by disrupting the functions of cells implicated in immune response (S. H. Wang et al. 2005).
TRAIL is also implicated in viral infections. During primary infection of influenza virus, CD8+ T cells kill virus-infected cells via TRAIL-mediated apoptosis. TRAIL deficient mice showed an increased influenza-associated morbidity and increased disease severity (Brincks et al. 2008). Human cytomegalovirus (HCMV) infection upregulated TRAIL-R1/R2 on infected fibroblasts, whereas, T and B lymphocytes, NK cells, macrophages and monocytes produce a high level on INF- which down-regulates TRAIL R1/R2 expression on uninfected fibroblasts. Therefore, TRAIL can now kill only virus-infected cells (L M Sedger et al. 1999)
3. c. TRAIL- mediated necroptosis
Not only TNF, but also TRAIL was proved to be able of necroptosis (or necrosis) induction. DRs not only trigger the apoptotic pathway, they also program cells to die in a caspase-independent pathway (Linkermann and Green 2014). Necroptosis mechanism depends on the formation of a complex called “Necrosome”, containing the kinases RIP1 and RIP3. Necrosome is mainly formed when caspase-8 is absent or inactive; it recruits and phosphorylates the pseudokinase MLKL which in turn forms oligomer via N-terminal domain, translocates to the plasma membrane and induces necrotic cell death by forming pores on the membrane (X. Chen et al. 2014)
47
3. d. TRAIL-mediated non-cell death pathway
Apart from inducing apoptosis and in some cases necroptosis, TRAIL can also trigger a variety of non-cell death signaling pathways including NFκB, MAP kinases, AKT and Src which can enhance the malignancy of cancer cells by increasing their proliferation, migration and invasion (Falschlehner et al. 2007). Induction of these pathways depends on cell type and also when apoptosis is inhibited (Tran et al. 2001). In vivo treatment of TRAIL promoted metastasis in an orthotopic Xenograft model and TRAIL-R2 expression enhanced tumor cell proliferation (Trauzold et al. 2006). Another study showed that TRAIL treatment induces proliferation and promotes migration in K-RAS mutated colorectal cancer cell lines. These findings shed the light to the pro-invasive role of endogenous TRAIL in K-RAS mutated cancer cells (Hoogwater et al. 2010). Therefore, it is very interesting to very well understand the non apoptotic signaling of TRAIL in order to anticipate its effects.
3. e. Different TRAIL forms for anticancer therapies
Targeting TNF/TNF-R and CD95/CD95-L in anticancer therapy provoked sever toxicity. However, targeting TRAIL/TRAIL-Rs system stimulates death signals in tumor cells; thereby different TRAIL forms have been developed and undergone first clinical testing. There are two categories of TRAIL pharmacological agents: Recombinant TRAIL and agonistic antibodies for TRAIL-R1 and –R2. Recombinant TRAIL which targets TRAIL-R1 and TRAIL-R2 at the same time can trigger them both simultaneously leading to a strong death signal. However, the presence of decoy receptors can bind TRAIL and tamper its activity. Thus, the use of TRAIL-Rs specific antibodies is much more recommended (Tuthill et al., 2015). But it is important to know which receptor (R1 or R2) should be targeted, depending on cancer type. For example, TRAIL-R1 mediates the apoptosis signaling in pancreatic cell lines and lymphoid malignancies (Marion MacFarlane et al. 2005; Johannes Lemke et al. 2010), whereas, it has been reported that TRAIL-R2 induces apoptosis in breast and colon cancer cells (R. F. Kelley et al. 2005) (Figure 8).
48
Figure 8. Clinical trials of TRAIL-DRs targeting therapies. Schematic representation shows the progress of TRAIL in clinical trials. Different cancer types and different forms of TRAILs are shown with the respective phases of clinical tests. For all of them, anticancer efficiency was observed in preclinical studies. However, only a minimal anticancer activity, which has not been detected in Randomized-controlled trials RTCs, was obtained (Lemke et al., 2014).
i) Recombinant form of TRAIL for therapeutic strategies
So far, the only form of recombinant TRAIL approved for clinical applications is an untagged protein, and comprises the TNF homology domain (THD) with extracellular domain of soluble TRAIL ( amino acids 114-281), called Dulanermin (Apo2L.0 or AMG-951) (A Ashkenazi et al. 1999b). This version of recombinant TRAIL is active and safe as it worked well in several xenotransplant cancer models, and was well tolerated and safely evaluated in
49
cynomolgus monkeys and chimpanzees (S K Kelley et al. 2001). Treatment of tumor-bearing mice that are lacking TRAIL-R expression, with Dulanermin results in anti-tumor effect by targeting non-cancer cells, most likely the micro-environmental cells. This study suggested that recombinant TRAIL induced apoptosis in DR5-expressing endothelial cells leading to the collapsing of blood vessels, tumor hemorrhage and reduced tumor growth (N. S. Wilson et al. 2012). Consequently, Dulanermin was tested in cancer patients of phase I clinical trials where it proved to be safe and well tolerated even when combined with Rituximab (CD-20 antibody), however, this combination did not lead to increased objective responses in patients with relapsed indolent B-cell lymphoma (Cheah et al. 2015). As well, Dulanermin is evaluated in another RCT (Randomized controlled trials) to evaluate the antitumor effect of novel pharmacological compounds. Patients with non-small-cell lung cancer were treated with chemotherapy agents alone (such as paclitaxel, carboplatin, bevacizumab) or combined with Dulanermin. Unfortunately, the addition of Dulanermin did not show any anticancer activity and did not improve the outcomes of patients (Soria et al. 2011a).
The disappointing results of Dulanermin obtained in clinical trials are due to its low pharmacokinetic profiles. Dulanermin has a short plasma half-life (3 to 5 minutes in rodents, and 23 to 31 minutes in nonhuman primates) and it has a rapid clearance from circulation (S K Kelley et al. 2001). In addition, Dulanermin can activate TRAIL-R1 by soluble and membrane-bound forms, whereas TRAIL-R2 is only activated by membrane-bound or soluble TRAIL cross-linked with monoclonal antibody (Wajant et al. 2001)
ii) Agonistics TRAIL-R specific antibodies as anticancer therapeutics
Besides Dulanermin, several agonistics TRAIL-R1/R2 antibodies are used in clinical trials, such as Mapatumumab which targets TRAIL-R1 and Conatumumab (AMG-655), Lexatumumab (HGS-ETR2), Tigatuzumab (CS-1008), Drozitumumab (Apomab) and LBY- 135) which target TRAIL-R2. They all showed an anticancer effect in preclinical tests. Therefore, agonistics antibodies were used in clinical trials and revealed safety and wide tolerability, having a long half-life in serum (hours to many days) compared to recombinant TRAIL, removing the need of repeated applications and allowing a stable concentrations into tumor tissues during the treatment. However, the activity of these specific antibodies is precluded in vivo by their requirement of external cross-linking to trigger an effective clustering of TRAIL-DRs, thereby facilitating DISC assembly and mediating apoptosis
50
(Adams et al. 2008; Dhein et al. 1992). Agonistics antibodies were used in RCTs, combined with chemotherapy of proteasome inhibitor Bortezomib, but no statistically significant anticancer activity was achieved, and results were largely disappointing, likely because an intrinsic TRAIL resistance within primary tumors or insufficient activity of TRAIL-Rs targeting drugs (Holland 2014). Thus a new strategy was developed to enhance TRAIL-Rs clustering, based on combination of TRAIL-R2 specific agonistic antibody AMG-655 (Conatumumab) with soluble TRAIL. This co-administration sensitizes cancer cells that were resistant to TRAIL, and this effect is due to a secondary cross-linking displayed by the agonistic antibody, which acted in cooperation with the normal clustering of TRAIL-R2 exerted by sTRAIL (Tuthill et al. 2015b).
In summary, during clinical studies of Dulanermin and agonistics TRAIL-Rs antibodies, a well toleration was observed, however, only a minimal anticancer activity was realized. Thus, to overcome these pharmacological downsides, new TRAIL formulations have been developed to increase its stability, efficiency and cancer-specific delivery:
iii) The tagged forms of TRAIL to increase its stability
Stability of TRAIL has a crucial role in its biological activity, and trimerization of TRAIL’s monomers is essential to trigger TRAIL-Rs clustering on cell surface and enhance the apoptotic signals. However, TRAIL monomers can also form disulfide linked dimers that derogate its apoptotic effect by up to 90-fold (Hymowitz et al. 2000b). Thus, recombinant versions of TRAIL have been developed to increase its stability, in which the amino terminus of TNF homology domain of TRAIL is fused to tags such as poly-Histidine tags (His-TRAIL) (Pitti et al. 1996) and FLAG epitope tag (FLAG-TRAIL), facilitating TRAIL purification process. It is remarkable that FLAG-TRAIL is seedy active and requires additional cross- linking by anti-FLAG antibody (M2) to oligomerize and induce apoptosis ((Wiley et al. 1995). However, his-TRAIL and FLAG-TRAIL were shown to kill freshly isolated primary human hepatocytes (PHH) in vitro, probably due to the formation of supramolecular aggregates by interactions between the tags (Lawrence et al. 2001). These results suggest that the tagged forms of TRAIL may cause hepatotoxicity when used in vivo. Thus, even Dulanermin has a lowest antitumor effect; it also has a low toxicity, that’s why it was primarily selected for clinical trials.
51
Another improved version of TRAIL that was developed to enhance the stability while maintaining the proper trimeric structure had consisted in an inclusion of a specific trimerization domain, a modified leucine zipper motif (LZ-TRAIL) or isoleucine zipper (iz- TRAIL) on the N-terminus of the extracellular domain. Interactions between these trimerization motifs form stable triple helices and provide a solid stability of TRAIL trimer. These versions of TRAIL are more active than Dulanermin, both in vitro and in vivo, with higher pharmacokinetics profiles (extended distribution and elimination half-life) and they showed no specific toxicity neither in PHH nor in mice (Rozanov et al. 2009; Ganten et al. 2006).
To improve its efficiency in vivo, TRAIL was linked covalently to molecules that have beneficial pharmacokinetics characteristics such as Human Serum Albumin (HAS) which extends TRAIL serum half-life to 15 hours after intravenous injection (Müller et al. 2010), and polyethylene glycol (PEG). PEGylated TRAIL enhanced therapeutic potentials in tumor Xenograft animal models due to its better tumor-targeting performance (Chae et al. 2010).
Among other strategies, TRAIL was fused to the Fc portion of human IgG1: Fc-TRAIL. It displays a high specific activity in vitro and a longer half-life than recombinant TRAIL. Fc- TRAIL showed an enhanced oligomerization and was more effective to inhibit tumor growth in Xenograft mice models without any remarkable hepatotoxicity (H. Wang, Davis, and Wu 2014). A hexavalent TRAIL was generated from two trimers of extracellular domain of TRAIL which were fused to Fc portion of human IgG1, creating six receptor binding sites per drug molecule. This compound, called APG 350, showed a potential apoptosis induction in cancer cell lines, primary tumor cells and mice models (Gieffers et al. 2013).
iv) Cancer-specific delivery of TRAIL
Many cancer cells are intrinsically resistant or acquire resistance to TRAIL during treatment, thus it is necessary to remove this resistance in order to kill them by TRAIL. Several studies demonstrated that co-administration of chemotherapeutic drugs can sensitize resistant cells to TRAIL-induced apoptosis (Kruyt 2008). However, chemotherapeutic drugs can damage normal cells too, thereby causing severe side effects. So it is better to target chemotherapeutic drugs, when combined with TRAIL, specifically to cancer cells. This
52
beneficial targeted delivery method can also increase TRAIL local concentration and decrease drug dilution in blood.
The first approach is TRAIL combination with nanoparticles (NPs). NPs have a diameter of 50-100 nm, composed of a large variety of compounds, and can improve pharmacokinetics, pharmacodynamics and drug stability in vivo. NPs have an enhanced permeability and retention (EPR) which means they can easily cross capillaries and lymph vessels that drain tumors (Davis et al. 2008). Several TRAIL-containing NPs werre developed so far, but TRAIL-containing liposomes have emerged as the most versatile ones and safely used in clinics. NPs containing TRAIL with other chemotherapeutic drugs are generated to enhance TRAIL-proapoptotic effect. For example, combination of TRAIL with Doxorubicin increased the therapeutic effect by enhancing the co-delivery of the drug with TRAIL, with any systemic toxicity observed in vivo (Guo et al. 2011)
In conclusion, TRAIL is a promising anticancer agent as it kills specifically tumor cells. Several clinical trials based on TRAIL therapies are developed, however the anticancer activity of TRAIL-R agonists tested in patients remains limited and disappointing.
B) FADD: a main signal transducer for Death Receptors
FADD gene (also called MORT1) is located on chromosome 11q13.3 in humans and chromosome 7 in mice (P. K. Kim et al. 1996). Locus that contains FADD gene is frequently mutated in human malignancies. For example, 11q13 region contains FGF 3 and 4 genes (Fibroblasts Growth Factors) which are co-amplified in melanoma. BCL1 is located very close to the FADD gene and it is mutated in B-cell lymphoma. However, no mutation of the FADD gene itself was reported so far (Katoh et al. 2003). Human and mouse FADD proteins share 80% similarity and 68% identity, consisting in 208 and 205 amino acids respectively (J. Zhang et al. 1996). The C-terminus Death Domain DD and the N-terminus Death Effector Domain DED of FADD are conserved between species, and they are both playing an essential role in transducing death receptors-mediated apoptotic signals (Weber et al. 2001b). FADD is considered, as caspae-8, a main regulator of death-receptor apoptosis. However, another study has reported that FADD plays non-apoptotic functions such as embryonic development and T cell proliferation (S.-M. Park et al. 2005). The dual role of FADD in apoptosis and cell survival remains controversial.
53
FADD-mediated signals of apoptosis are inhibited by c-FLIP with intervention of NFκB activation which promotes cell proliferation (Piao et al. 2012). However, an induced expression of FADD in cancer cells decreases c-FLIP protein expression and re-activates caspase-8 mediated apoptosis (Ranjan et al. 2012). Thus, a recent study explored the underlying molecular mechanism of FADD-mediated ablation of c-FLIP and has demonstrated that an ectopic expression of FADD promoted JNK-mediated activation of E3 ubiquitin ligase responsible for ubiquitination and degradation of c-FLIP. Furthermore, FADD impedes NFκB activation by ubiquitinating IKK and negatively regulates the inhibitor of apoptosis cIAP2 (Ranjan et al. 2016). In addition to its role in mediated death- receptor apoptosis, overexpression of FADD can form long filamentous aggregates called “DEFs” (Death Effector Filaments) by self-association and then initiate apoptosis independently of death receptors (Sheng Wang et al. 2017). Recent evidence demonstrated that a mutant FADD called N-FADD, truncated in C-terminus tail, can strongly increase FADD potential to self-associate, thereby enhancing apoptosis in melanoma cancer cells. Moreover, N-FADD significantly suppressed tumor growth and enhanced survival time in mice bearing melanoma tumors (Chien et al. 2016)
The first role attributed to FADD was transmitting the apoptotic signal from death receptors expressed on the cell surface, thus it was suggested that FADD is localized exclusively in the cytoplasm. However, a nuclear localization sequence (NLS) was later identified and accounts for FADD expression in the nucleus. It is certain that cytoplasmic FADD has a proapoptotic role, in contrast the role of nuclear FADD is still unclear. It is supposed that it protects cells from apoptosis, but the underlying mechanism has to be more investigated (Gómez-Angelats et al. 2003). To investigate other roles of FADD, FADD knockout mice were generated. Embryos death was observed at day 12 of development because of a remarkable abdominal hemorrhage and cardiac failure. These findings suggest that, apart from its role in cell death transduction, FADD plays an important role in the proliferation of some cell types (Yeh et al. 1998). Furthermore, FADD is implicated in the development, proliferation and function of the T cell lineage. Activation of T cells that are FADD-dominant negative, by specific antibodies or exogenous cytokines is markedly abolished, suggesting that FADD is required for T cells differentiation (Beisner et al. 2003). Mice expressing a functional FADD have a normal embryogenesis and lymphopoiesis; in contrast T cell-specific FADD deficient mice contain a normal thymocytes number but few peripheral T cells, demonstrating that FADD is not essential for thymocytes development, but
54
indispensable for peripheral T cell homeostasis (Y. Zhang et al. 2005). Moreover, FADD plays an important role in preventing chronic intestinal inflammation. Mice with Intestinal Epithelial Cells (IEC)-specific knockout of FADD developed epithelial cell necrosis, enteritis and erosive colitis (Welz et al. 2011). Thus, FADD is an important functional component for apoptotic and non-apoptotic signaling pathways.
C) Caspases: central players in apoptosis
The term “caspase” was used for the first time in 1996 to introduce the functional role of these enzymes, as “C” refers to cysteine proteases, and “aspase” to their ability to cleave at an aspartic acid residue (Alnemri et al. 1996). Two groups of caspases are identified so far, based on their functions: a. The inflammatory caspases including caspase-1, -4, -5, -11 and -12. These caspases induce the activation of “inflammasomes” responsible for inflammation process triggering, leading to an inflammatory form of cell death called “pyroptosis”. In human, only caspases-1, -4, -5 and -12 are encoded, however the mouse genome encodes caspase-1, -11 and -12. Noting that mice caspase-11 is homologous to human caspase-4 and -5 (Latz, Xiao, and Stutz 2013) b. The apoptotic caspases are responsible for initiating and executing the apoptotic machinery, and they are divided into two subgroups: the initiator caspases including caspase- 2, -8, -9 and -10, and the effector caspases including caspase-3,-6 and -7. Caspase-8 and -10 are activated into the multiprotein complex DISC; caspase-9 is activated in Apoptosome complex, and caspase-2 into PIDDosome complex. Caspase-8/10 and caspase-9 initiate the extrinsic and the intrinsic apoptosis respectively (Fuchs and Steller 2015).
Caspases are expressed in many tissues and organs, in immune cells and non-immune cells. They are initially produced as inactive procaspases that require dimerization for their activation. They are constituted of a C-terminus protease effector domain containing a large and a small subunit, and an N-terminus prodomain allowing them to have different protein- protein interactions (Taylor, Cullen, and Martin 2008). Caspases differ by their prodomain, so they can interact with different protein adaptor. For example, caspase-1,-2,-4,-5,-9,-11 and -12 have an N-terminus prodomain known as “CARD” (Caspase activation and recruitment domain), whereas caspase-8/-10 contain DEDs. However, the effector caspases -3,-6 and -7 are characterized by a short prodomain, and are inactive until cleavage and activation by the initiator caspases (Creagh 2014) (Figure 9).
55
Figure 9. Domain structures and classification of caspases. CARD refers to caspases activation and recruitment domain, DED refers to Death Effector Domain. * 12 different caspases are encoded by the human genome. Noting that caspase 12 has two isoforms: caspase 12S as an inactive truncated form, and caspase 12L as an active full-length form. However, a thirteenth caspase was identified as caspase 14, which is less easily categorized, neither an apoptotic nor an inflammatory caspase, but it has a specialized role in Keratinocytes differentiation (Man S et al., 2016).
56
When apoptosis machinery is induced, initiator caspases that exist as inactive procaspases monomers are activated, and they in turn trigger effector caspases that subsequently coordinate their activities to break down key structural proteins and activate other enzymes. Effector caspases are produced as inactive procaspases dimers and they must be cleaved by initiator caspases to their activation. This cleavage occurs between the large and the small subunits resulting in conformational changes that creates a mature functional protease (Riedl and Shi 2004). Once active, a single effector caspase can activate other executioner caspases leading to accelerated feedback loop of caspase activation.
1. Caspase 8 dual functions in apoptosis and necrosis
Initiator caspase-8 is a focal apical caspase for death receptor-mediated apoptosis signaling. It is a 55 kDa cysteine protease, consisting of 480 amino acids. It contains two DEDs and a catalytic protease domain (Fulda 2009). For its activation, an oligomerization and proteolysis cleavage are required. The first cleavage generates p43/41 and the second one generates p18/10 which is released in the cytosol (Degterev, Boyce, and Yuan 2003). Upon death receptor induction, the monomeric procaspases-8 protein is recruited by its DED to the DISC resulting in dimerization and activation of caspase-8. Cells from caspase-8 deficient mice are thus resistant to DR-mediated apoptosis (Kang et al. 2004). Caspase-8 serves a vital role in embryonic development. Deletion of caspase-8 in mice demolishes apoptosis and leads to embryonic death caused by an impaired heart muscle development and crowded accumulation of erythrocytes (Varfolomeev et al. 1998)
Apart from its important and initial role in mediating the extrinsic apoptosis, caspase-8 appears to have additional functions unrelated to apoptosis. For example, caspase-8 has a critical role in T-cell homeostasis (Salmena et al. 2003). Deletion of caspase-8 results in chronic skin inflammation in mice (Kovalenko et al. 2009). In addition, caspase-8 is able to inhibit necroptosis, an inflammatory cell death mediated by RIPK1 and RIPK3. Caspase-8 is able to cleave RIPK1 and RIPK3, thus a spontaneous inflammation resulted from caspase-8 deletion is prevented by the inhibition of necroptosis (Günther et al. 2011).
2. Caspase-8 deregulation in cancers
Squamous cell carcinoma of the oral cavity was the first tumor identified with caspase-8 mutations. This mutation alters the Stop codon and increases the size of the protein by 88
57
amino acid. Thus, the ability of the mutated protein to induce apoptosis is reduced, suggesting that this mutation affects caspase-8 activity (Mandruzzato et al. 1997). In invasive colorectal carcinomas, five somatic caspase-8 mutations were found, and markedly reduced caspase-8 activity, indicating that these mutants have a dominant negative inhibition of apoptosis (H. S. Kim et al. 2003). These caspase-8 mutations are not associated with loss of caspase-8 expression. However, a decreased caspase-8 expression was observed in Head and Neck carcinoma cells, and correlated with chemoresistance. 5’Aza-Deoxycytidine (a DNA methyltransferase inhibitor) increased caspase-8 mRNA and protein expression, suggesting that epigenetic silencing is responsible for caspase-8 downregulation (Liu et al. 2009; Eggert et al. 2000). Decreased stability and enhanced degradation of caspase-8 have been reported in TRAIL-resistant colon cancer cells. An upregulation of CARP1 and CARP2 proteins, implicated in ubiquitin-mediated proteolysis of DED containing caspases, is also found in these resistant cancer cells. Thus, restoration of caspase-8 expression re-sensitized colon cancer cells to TRAIL-induced apoptosis (McDonald and El-Deiry 2004). Several caspase-8 post-translational modifications are identified, including caspase-8 phosphorylation which is reported to inhibit its activity. The majority of caspase-8 phosphorylation studies have focused on tyrosine 380 phosphorylation (Y380) which has a critical role in cancer progression, migration and metastasis (Senft, Helfer, and Frisch 2007). A recent study has demonstrated that phosphorylation of Y380 localized in the linker region joining the large and small catalytic subunits, leads to increased resistance to CD95-mediated apoptosis and enhanced cell migration. Y380 phosphorylation inhibits caspase-8 autoproteolytic activity and release of the mature caspase-8 subunits but not its homodimerization into the DISC, thereby impeding downstream activation of caspases cascade (Powley et al. 2016)
Although caspase-8 is a key player in the most reviewed mechanism of cancer cell death, deregulation of this enzyme and its signaling pathways helped to discover an alternative non- apoptotic functions of caspase-8 such as serving in the persistence of mutated cells and promoting tumorigenesis.
D) DED chain and DISC assembly
Formation of the multiproteic Death-Inducing Signaling Complex DISC after death receptors triggering, is a critical step in mediating apoptosis. However, death receptors can also mediate cell survival signaling by the activation of nonapoptotic pathways including
58
NFκB and MAP kinases (Peter et al. 2007). As caspase-8 is the apical caspase in death receptor pathway and its activation occurs into the DISC, it is important to understand how caspase-8 is activated, what mechanisms regulate its activation and how this regulates cell fate! It has been proposed that caspase-8 activation occurs through an “induced proximity” mechanism, involving dimerization of caspase-8 (required for initial activation) and cleavage (required for an efficient apoptosis signaling) (D. W. Chang et al. 2003). These two steps are required to achieve the signaling threshold for DISC mediated apoptosis. Importantly, without the second-step cleavage event, the caspase-8 dimerization alone is insufficient to trigger apoptosis and may induce alternative pathways resulting in cell survival. Thus, conformation status of caspase-8 within the DISC determines cell fate, either death or survival (Hughes et al. 2009).
Although DISC formation is an essential step in death receptor-mediated apoptosis, mechanisms underlying its assembly remain unclear. Initially, it was proposed that within the DISC, one ligand trimer binds to one receptor trimer which in turn recruits three molecules of FADD and three caspase-8 (Weber and Vincenz 2001a). Another study suggested that DISC components are assembled in a ration 3:2:2 for DR: FADD: caspase-8 (H Berglund et al. 2000). Substantially, stoichiometry of core components and structure of native DISC have never been well defined with full length proteins. A recent work showed that the native DISC does not conform to the 1:1:1 model of DR: FADD: caspase-8; instead, they have demonstrated that FADD is substoichiomeric relative to death receptors and caspase-8. There is surprisingly up to 9-fold more caspase-8 than FADD into the DISC, suggesting that caspase-8 molecules interact sequentially by their DED domains to form a DED chain. Importantly, mutations in DED2 of caspase-8 abolishes DED chain formation and blocks activation of caspases cascade and cell death (Dickens et al. 2012b). A more recent study has demonstrated that caspase-8 activation is occurring by DED chain at the DISC, and by analyzing the molecular architecture of this chain, they found that is highly composed of caspase-8 cleavage products, in particular its prodomain (p26/p24). The relative amounts of p26/p24 in the DISC increased over time (80% of total caspase-8 amounts in 60 minutes) and the cleavage product p43/p41 is present at low levels, suggesting that rapidly processed within the DISC. They have also demonstrated that caspase-10 and c-FLIP are present in 10 times lower proportions than caspase-8 in the DED chain, noting that c-FLIP does not prevent caspase-8 recruitment neither terminates DED chain, but rather incorporates into them. DED chain termination depends on dissociation/association rates of caspase-8. The stability of the
59
longer chain decreases and caspase-8 bound at the end of a long chain has a higher dissociation probability than one bound to a short chain (Schleich et al. 2016).
In summary, the DED chain assembly model represents an essential possibility that a small amount of DISC is able to activate large amounts of caspase-8. Activation of caspase-8 has to be strictly regulated in order to prevent inappropriate cell death or activation of alternative pathways.
60
Chapter III
“C-FLIP”: a major inhibitor of the extrinsic apoptotic pathway and a relevant clinical target for cancer therapies
61
c-FLIP: Cellular FLICE (FADD-Like Interleukin-1 -Converting Enzyme)-Inhibitory Protein is a potential anti-apoptotic protein that negatively regulates death receptor- downstream signaling, and it is considered as a major resistance factor that suppresses TRAIL-mediated apoptosis, as well as apoptosis induced by chemotherapy drugs in cancer cells (Safa et al. 2011a) In this chapter, I will discuss the role of c-FLIP variants in preventing apoptosis, different regulators of c-FLIP expression, and novel c-FLIP-targeted drugs for cancer therapies which improve TRAIL efficacy and help to eliminate cancer cells.
I) Different isoforms and structures of c-FLIP
Some viruses express anti-apoptotic proteins to inhibit cell death of their host cells and to escape from the protective apoptotic machinery. One of these proteins was identified in 1997 and called Viral-FLICE Inhibitory Proteins (v-FLIPs). Viruses that contain v-FLIPs are members of the gammaherpesvirus class, including Human Herpesvirus 8 (also called Kaposi sarcoma-associated Herpesvirus), Herpesvirus saimiri and Moluscum Contagiosum Virus (MCV). V-FLIPs contain two DEDs, having the ability to interfere within the DISC, thus preventing caspase-8 activation (Margot Thome et al. 1997). The overexpression of HHV8 v- FLIP led to T cell resistance to apoptosis and enhanced tumor progression in a murine B-cell lymphoma model. V-FLIP gene deletion from Herpesvirus saimiri confirmed the antiapoptotic role of v-FLIP; however, it showed that v-FLIP is not indispensible for viral replication and pathogenicity (Glykofrydes et al. 2000).
Because of the high similarity between v-FLIP and N-terminus of human caspase-8, it was suggested that v-FLIP gene was derived from its host cell gene. Thus, researchers had later identified a highly similar gene in the human genome and called it CFLAR (Caspase-8 and FADD-Like Apoptosis Regulator) which encodes for Cellular-FLIP protein (c-FLIP, also called FLAME-1, I-FLICE, Casper, MRIT, Usurpin, and CLARP) (Jürg Tschopp et al. 1997). CFLAR gene is located on human chromosome 2q33-34 near to caspase-8 and -10 genes, suggesting that these three genes have evolved from ancient gene duplication (Rasper et al. 1998b). CFLAR gene has 13 distinct spliced variants, but at the protein level in human, three of which are expressed as proteins: the 55 kDa long form (c-FLIP L), the 26 kDa short form (c-FLIP S), and other short form protein of 24 kDa (c-FLIP R) (J Tschopp, Irmler, and Thome 1998) (Figure 10). C-FLIP L has two N-terminus DEDs (DED1 and DED2) and a C-terminus caspase-like domain (p20 and p12), thus it closely resembles to caspase-8 and -10, however,
62
c-FLIP L does not have a caspase-like proteolytic activity because of several amino acids substitution, especially the essential cysteine residue in the Gln-Ala-Cys-X-Gly motif (X refers to any amino acid) and the Histidine residue in the His-Gly motif, and they are both crucial and indispensible for caspase catalytic activity and are conserved in all caspases. C- FLIP L has the same cleavage site as caspase-8, positioned at Asp-376 which corresponds to the link between p20 and p12. Cleavage at this site generates the proteolytic fragment variant p43c-FLIP (Safa, Day, and Wu 2008a; Cohen 1997). C-FLIP S is composed of two DEDs and a short C-terminus tail of 20 amino acids that are crucial for its ubiquitination and targeting for degradation by proteasome. Structures of c-FLIP S and v-FLIP are reported to be similar (Poukkula et al. 2005a). As well, c-FLIP R contains two DEDs but it is lacking the additional C-terminus amino acids found in c-FLIPS. it is specifically expressed in some cell lines such as Raji cells and in primary human T cells. C-FLIP R and c-FLIP S share several similar characteristics, for example they have both a short half-life and a similar mode of expression after primary human T cells activation (Golks et al. 2005). Surprisingly, only a single nucleotide polymorphism called rs10190751 determines the production of either c- FLIP S or c-FLIP R, but due to the difference in protein translation rates, c-FLIP S is produced more than c-FLIP R (Nana Ueffing et al. 2009). These three isoforms of c-FLIP are all reported to interact with FADD and form heterodimers with caspase-8 via DED-DED interactions. Heterodimerization of caspase-8 with c-FLIP L and not c-FLIP S and R, results in two consecutive cleavage products of c-FLIP L, called p43-FLIP and p22-FLIP, generated from caspase-8-mediated cleavage at D376 and D196 respectively (Bagnoli, Canevari, and Mezzanzanica 2010).
Human CFLAR gene is composed of 14 exons and different isoforms are generated by alternative splicing. Inclusion of exon 7 results in c-FLIP S translation because of a stop codon; however, skipping exon 7 results in c-FLIP L translation. A stop codon at the beginning of intron 6 results in c-FLIP R translation (Djerbi et al. 2017). CFLAR gene is evolutionary conserved in vertebrates. C-FLIP S is absent in mice because of the lack of corresponding exon, however, the other short form c-FLIP R is found in mice, as well c-FLIP L. D376 of human c-FLIP L is conserved in mice as D377, in contrast D196 is only found in human c-FLIP L. Thus, the first cleavage product of c-FLIP L, p43-FLIP, is produced by cleavage in mice, but the second cleavage product p22-FLIP is only restricted to humans (N Ueffing et al. 2008; Salvesen and Walsh 2014).
63
Figure 10. Structures of CFLAR gene and c-FLIP protein variants. In gene structure, red arrow indicated CFLAR gene, and the black arrows indicate the nearby genes, mainly caspase-8/10 genes. For protein structures, the three c-FLIP isoforms contain similar DEDs in their N-terminus domain, but they differ by the length of Ct domain. The numbers indicate amino acids residues, and the small arrows indicate different cleavage sites (Tsuchiya Y et al., 2015).
II) Different c-FLIP functions
A) Molecular function of c-FLIP in regulating apoptosis
C-FLIP variants were initially described as inhibitors of DR-mediated apoptosis. The role of the short forms of c-FLIP in regulating apoptosis is well established. When c-FLIP S is recruited to the DISC via DED interactions, it blocks CD95- and TRAIL-mediated apoptosis by inhibiting caspases-8 processing and activation through heterodimerization. The same effect is reported for c-FLIP R (Jürg Tschopp et al. 1997). Moreover, it was reported that c- FLIP S efficiently inhibits oxaliplatin-induced apoptosis by increasing XIAP protein stability and AKT activation (S. Kim et al. 2008). However, due to the high homology between c-FLIP L and caspase-8, the effect of c-FLIP L on apoptosis regulation was controversially
64
investigated. It was demonstrated that c-FLIP L has an anti-apoptotic effect similar to c-FLIP S/R, when it is expressed at high levels in the cells; otherwise, when c-FLIP L is expressed at physiological relevant levels, it acts as a pro-apoptotic protein by enhancing caspase-8 activation through catalytic active caspase-8/c-FLIP L heterodimers formation (D. W. Chang et al. 2002). The pro-apoptotic effect of c-FLIP L was confirmed by a study on c-FLIP L- deficient mice which showed a heart failure and embryonic death, a phenotype similar to FADD- and caspase-8 deficient mice (Yeh et al. 2000a). Nevertheless, these data contradict with noticing that selective c-FLIP L Knockdown increased caspase-8 recruitment to DISC, processing and activation, thereby promoting DR-induced apoptosis (Sharp, Lawrence, and Ashkenazi 2005b). A recent study of Majkut et al. has demonstrated that homodimerization of two caspase-8 molecules into the DISC results in full processing of caspase-8 which effectively triggers apoptosis. However, when c-FLIP L is incorporated in a DED chain, it forms heterodimers with caspase-8 and activates it without interdomain cleavage. Thus, c- FLIP L-activated caspase-8 remains restricted to the DISC via DED-DED interactions and is able to cleave only limited nearby substrates. This finding confirms how c-FLIP L inhibits apoptosis even with its ability to activate caspase-8 (Majkut, Sgobba, Holohan, Crawford, Logan, Kerr, et al. 2014). Extensive studies involving diverse types of human cancer cells revealed that the role of c-FLIP-L is rather anti-apoptotic than pro-apoptotic. For example, in MCF-7 breast cancer cell line, c-FLIP L forms with DR5, FADD and caspase-8 a complex called AIC (Apoptotic Inhibitory Complex). C-FLIP L expression Knockdown using small interfering RNA blocks breast cancer cells proliferation and induces apoptosis by activating death receptor and mitochondrial pathways (Day, Huang, and Safa 2008). The role of c-FLIP variants in the regulation of apoptosis remains unclear. Some studies revealed that the long form acts as pro-apoptotic molecules, in contrast other studies demonstrated its antiapoptotic function. Importantly, these observations are issued from different research groups, thus the different obtained results might be dependent on cell types, c-FLIP amounts in cells, stimulus strength and specificity of death ligands.
B) Role of c-FLIP in Necroptosis
C-FLIP has been reported to play an essential role in necroptosis regulation. Similar to their different functions in apoptosis, c-FLIP L and c-FLIP S contribute differently in necroptosis pathways (Olivier Micheau 2003a). Necroptosis is an inflammatory cell death,
65
triggered by physical or chemical trauma, viral or bacterial infections and severe microenvironmental conditions. It is a caspase independent cell death, mediated by RIP1/RIP3 kinases. Necroptosis can be induced by stimulation of TNFR1, caspases inhibitors, genotoxic drugs or administration of SMAC mimetic (known as IAPs antagonists). These conditions lead to the activation of RIPK1 (Receptor Interacting Protein Kinase 1) which binds to FADD via DD-DD interactions. Subsequently, DED-containing proteins such as caspase-8 and c-FLIP (L/S) bind to FADD via DED-DED interactions to form a complex called Ripoptosome (Vandenabeele et al. 2010). When c-FLIP is absent, caspase-8 is fully processed and activated into the Ripoptosome through homodimerization, thus it cleaves and inactivates RIPK1, dissociates from Ripoptosome and executes apoptosis. In contrast, when c- FLIP L is recruited to the Ripoptosome, it forms heterodimers with caspase-8 leading to its activation. Ripoptosome-restricted active caspase-8 cleaves and inactivates RIPK1. After RIPK1 cleavage, caspase-8 disassembles from the Ripoptosome and becomes inactive. Thus, c-FLIP L incorporation into Ripoptosome blocks necroptosis as well as apoptosis, and maintains cell survival (Schilling, Geserick, and Leverkus 2014). While c-FLIP L was shown to block necroptosis, the short form c-FLIP S was reported to promote necroptosis. Procaspases-8/c-FLIP S heterodimers prevent caspase-8 processing and activation, thus fail to cleave RIPK1 which remains active. The active RIPK1 forms with RIPK3 and MLKL a necroptosis-inducing complex called “Necrosome” (Feoktistova et al. 2011). Therefore, these data suggest that c-FLIP isoforms into the Ripoptosome determine if cells undergo RIP- mediated necroptosis or caspase-dependant apoptosis.
C) Role of c-FLIP in inducing a survival signaling
Apart from its role in mediating different types of cell death, accumulating evidence has demonstrated that c-FLIP activates many cytoprotective pathways implicated in cell survival, proliferation and carcinogenesis. When overexpressed, c-FLIP L activates NF-κB and ERK signaling by binding to their adaptor proteins such as TRAF1/TRAF2/RIP and Raf-1 respectively, leading to cell proliferation (T Kataoka et al. 2000a). C-FLIP L needs to be cleaved in order to activate NF-κB. Only caspase-8 mediated cleavage product p43-FLIP is able to interact with IKK complex leading to NF-κB activation (Takao Kataoka and Tschopp 2004). The other cleavage product, p22-FLIP, generated from the three different isoforms of c-FLIP after cleavage at D196, is also reported to interact with the IKK complex and mediate
66
NF-κB activation (Golks et al. 2006). An up-regulation of c-FLIP level was found in myofibroblasts of lungs with fibrosis, resulting in the inhibition of Fas-mediated apoptosis and an enhancement of cell proliferation via TRAF and NFκB activation (Figure 11). Downregulation of c-FLIP sensitized myofibroblasts to Fas-mediated apoptosis and decreased their proliferation (Golan-Gerstl et al. 2012). The overexpression of c-FLIP L but not c-FLIP S induces the phosphorylation and activation of FAK and ERK leading to an increased expression of MMP-9 responsible for proteolytic degradation of the extracellular matrix. Thus, by activating these pathways, the overexpression of c-FLIP L promotes cancer cell adhesion and motility (D. Park et al. 2008a). The overexpression of c-FLIP L prevents the ubiquitination and the proteasomal degradation of -catenin, leading to an enhanced expression of cyclin D, colony formation and invasiveness of prostate cancer cells. By inhibiting -catenin degradation, c-FLIP L promotes canonical Wnt signaling which induces the expression of several genes such as c-myc. Thus c-FLIP L overexpression enhances tumorigenesis in addition to its role in preventing Fas-induced apoptosis in cancer cells (Naito et al. 2004). In addition, the overexpression of c-FLIP L harms the function of Ubiquitin- Proteasome System (UPS) by increasing the accumulation of various short-lived proteins including HIF1α ( Hypoxia Inducible Factor 1α) which regulates the expression of different genes implicated in cell proliferation, invasion and metastasis (Ishioka et al. 2007). Up- regulated c-FLIP L interacts with the serine-threonine kinase Akt and enhances its survival signaling by preventing its ability to bind and phosphorylate its substrates such as Gskγ . Thus, the active Gskγ induces a reduction of caspase-3 and p27 mRNA levels leading to TRAIL-mediated apoptosis reduction (Quintavalle et al. 2010). Inhibition of Akt pathway using DMNB (4,5-dimethoxy-2-nitrobenzaldehyde) led to an increased expression of DR4/DR5 mRNA and a decreased level of c-FLIP mRNA, thus sensitizing cancer cells to TRAIL-induced apoptosis (M.-J. Kim et al. 2009). C-terminus of c-FLIP L, but not c-FLIP S, interacts with Fas-binding domain of Daxx (an alternative Fas signaling adaptor and implicated in proapoptotic signaling), thus inhibiting JNK activation by preventing the normal interaction of Daxx with Fas. Therefore, inhibition of JNK activation by the overexpression of c-FLIP L renders cells resistant to Fas-induced apoptosis (Y.-Y. Kim et al. 2003).
67
Figure 11. C-FLIP/caspase-8 dimerization in the DISC. 1) When c-FLIP is absent, homodimerization of two caspase-8 results in full processing and activation of caspase-8, and apoptosis induction. 2) In the presence of c-FLIP S, caspase-8 remains uncleaved thus not functional and apoptosis is blocked. 3) In the case of c-FLIP L, it forms heterodimers with caspase-8 with partial processing of caspase-8. The active heterodimer remains restricted to the DISC and cleaves small number of substrates including RIP, as well it activates non- apoptotic pathways such as ERK and NF-κB (Shirley et al., 2013).
D) c-FLIP role in tissue homeostasis and immune system
To evaluate the physiological role of c-FLIP, the exon 1 of CFLAR gene was disrupted leading to the loss of all c-FLIP isoforms in mice embryos. Lacking of c-FLIP in the whole body prevented mice embryos from surviving past day 10.5 of embryogenesis and displayed heart development failure. Moreover, c-FLIP deficient embryonic fibroblasts were more sensitive to TNF-induced apoptosis (Yeh et al. 2000b). Mutations of c-FLIP stop codon resulted in 46-amino acid extension on its C-terminus. The resulted aberrant c-FLIP protein is polyubiquitinated by TRIM21 E3 ubiquitin ligase, thus it is more rapidly degraded by UPS leading to apoptosis of the hepatocytes during embryogenesis and mice death at around day
68
E13.5 (Shibata et al. 2015). These findings obviously demonstrated the role of c-FLIP in mammalian development; however, its implication in adult animal’s development is still unknown.
c-FLIP has also an important role in T-cell activation and development. Generation of c- FLIP deficient T-cells resulted in a severe reduction of T cell number in thymus, spleen and lymph nodes in mice, and impairment of T cells proliferation after TCR activation (Chau et al. 2005). The role of c-FLIP in B-cells was also studied and revealed that B-cell specific c-FLIP deficient mice have a decreased number of peripheral B cells that become hypersensitive to Fas-mediated apoptosis, and an impaired proliferation stimulated by TLRs (Toll-Like Receptors) and BCR (B cell receptor) (H. Zhang et al. 2009). Generation of macrophages- deficient mice by deleting c-FLIP in myeloid cells promoted the development of neutrophila, splenomegaly and delayed neutrophiles clearance, suggesting that c-FLIP deficiency contributed to macrophages failure to clear apoptotic neutrophiles (Gordy et al. 2011). Dendritic cells (DCs) are known to be crucial for immune homeostasis. Deletion of c-FLIP from DCs in mice resulted in spontaneous inflammatory arthritis development, autoreactive CD4+ T cells increase, and a reduced number of T regulatory cells (Treg) (Q.-Q. Huang et al. 2015) As well, another study showed that c-FLIP deficient DCs have an enhanced production of inflammatory cytokines (TNFα, IL-2 and G-CSF). These findings showed the unexpected anti-inflammatory activity of c-FLIP and its suppressive role against innate immunity (Y.-J. Wu et al. 2015)
III) c-FLIP: Elevated level in human cancers
An up-regulation of c-FLIP level has been found in several types of cancers, and it is correlated with cancer progression due to inhibition of the apoptotic machinery. The observation of different cell lines including ovarian carcinoma, colorectal carcinoma, and prostate carcinoma cells showed increased expression of c-FLIP. As well, primary tissues of patients with Head and neck squamous cell carcinoma (HNSCC), bladder urothelial carcinoma, gallbladder carcinoma, B-cell chronic lymphocytic leukemia, Burkitt’s lymphoma, lung adenocarcinoma, Non-Hodgkin’s lymphoma and hepatocellular carcinoma have elevated levels of c-FLIP. An overexpression of c-FLIP has been also found in primary cells of melanoma, Hodgkin’s lymphoma, non-small cell lung carcinoma, and gastric carcinoma
69
(Shirley and Micheau 2013). Eighteen out of 18 patients with primary Ewing sarcoma, including metastasis, showed an abundant expression of c-FLIP (de Hooge et al. 2007). An expression of c-FLIP variants (L/S) was detected in pancreatic ductal adenocarcinoma, however, c-FLIP expression is absent in normal pancreatic ducts (Haag et al. 2011). The evaluation of c-FLIP profile has demonstrated that c-FLIP is significantly more expressed in osteosarcoma lung metastasis than in osteosarcoma primary tumors, suggesting that c-FLIP plays an important role in the metastatic potential of osteosarcoma to the lung (Rao-Bindal et al. 2013). An up-regulation of c-FLIP transcripts was also detected in gastric adenocarcinomas and is correlated with lymph nodes metastasis and tumor progression (ZHOU et al. 2004). High grade intraepithelial lesions showed an enhanced expression of c- FLIP, contrary to normal cervix epithelium or low-grade lesions where c-FLIP is absent, suggesting that c-FLIP is a potent marker of cervical cancer progression (Ili et al. 2013). The overexpression of c-FLIP L is detected in the majority of malignancies more than c-FLIP S; otherwise some studies focused on the upregulation of c-FLIP S. An Akt-mediated upregulation of c-FLIP S promotes human gastric cancer cells survival and provides resistance to TRAIL-induced apoptosis (Nam et al. 2003a). Another study showed that an overexpression of c-FLIP S but not c-FLIP L is found in human lung adenocarcinoma with low level of E2F1, an important transcription factor during apoptosis (Salon et al. 2006). These findings prove that c-FLIP isoforms are often upregulated in tumors, and their expression is correlated with chemotherapy drugs, as well as TRAIL resistance, tumor aggressiveness, and poor clinical outcome. Thus, c-FLIP is considered as an important therapeutic target to restore the apoptotic process in cancer cells.
IV) Modulators of c-FLIP expression
A) Regulation of c-FLIP on transcriptional and translational levels
Different stimuli such as chemotherapeutic agents, TNF superfamily ligands, chemokines and interleukins, as well as growth factors can regulate the expression of c-FLIP by modulating its transcription. These stimuli activate large number of transcription factors which are reported to transcriptionally modulate c-FLIP gene. For example, it has been reported that these following transcription factors: NF-κB, EGR1, p6γ, p5γ, NFAT, AR, sp1 induce the expression of c-FLIP gene. However FOXO3a, IRF5, c-myc, c-FOS and sp3 block
70
c-FLIP transcription (Safa, Day, and Wu 2008b). The overexpression of the zinc-finger transcription factor Gli2 induces a high transcription of c-FLIP (L/S) in basal cell carcinoma BCC (Kump et al. 2008). During early infection, the influenza A virus matrix 1 protein (M1) induces the transcription of many survival genes such as c-FLIP gene through ReIB (NF-κB member) activation (Halder et al. 2013). It was demonstrated that chemotherapy-induced interleukin-8 (IL-8) signaling increased mRNA transcript levels and expression of c-FLIP L and S via NF-κB- and AR-dependant transcriptional activation, thus reducing the sensitivity of prostate cancer cell to exhibit apoptosis (C. Wilson et al. 2008). Furthermore, c-FLIP expression is transcriptionally upregulated by PI3K/AKT and MAPK pathways activation, resulting in inhibition of DR-apoptosis (Olivier Micheau 2003b). The regulation of c-FLIP R is less described in the literature. One study has demonstrated that CD40 activation results in upregulation of the short form c-FLIP R which inhibits CD95-mediated apoptosis in primary precursor B-ALL (Acute Lymphoblastic Leukemia) (Troeger et al. 2007). It is interesting to know that c-FLIP isoforms are differently regulated by the same transcription factor, and these regulation differences might be cell line-dependant. For example p63, an important transcription factor for epithelial development, enhances c-FLIP R expression but it decreases c-FLIP S transcription, while c-FLIP L remains intact in Keratinocytes (Borrelli et al. 2009).
Other studies have demonstrated the translational regulation of c-FLIP. A natural herbal compound called Rocaglamide (Roc) prevents the translation of c-FLIP S by inhibiting the activation of the translation initiation factor 4E (eIF4E), and sensitizes adult T-cell leukemia/lymphoma cells to TRAIL- and CD95L-induced cell death (Bleumink et al. 2011). Moreover, in Glioblastoma multiform (GBM) cells, translation of c-FLIP S is upregulated by the activation of mTOR-p70/S6 Kinase 1 pathway. Inhibition of mTOR suppresses c-FLIP S protein expression and removes TRAIL resistance of cancer cells (A. Panner et al. 2005).
B) Post-translational regulation and degradation of c-FLIP
Post-translational modifications can regulate c-FLIP protein concentration by promoting its degradation. Previous studies showed that c-FLIP isoforms are short-lived proteins and they are often degraded by the Ubiquitin-Proteasome System (UPS) (Fukazawa et al. 2001). A co-administration of TRAIL with γ,γ’-diindolylmethane overcomes TRAIL resistance in human cancer cells by downregulating c-FLIP protein expression through UPS (S. Zhang et al. 2005). A mathematical model in a recent study showed that c-FLIP isoforms are unstable
71
and characterized by a dynamic turnover, so that their expression is quickly altered, making them a key determinant of death receptor responses (Toivonen et al. 2011). It has been reported that c-FLIP S is more prone to ubiquitination and proteasomal degradation due to the presence of two important lysine residues, Lys192 and Lys195, in its C-terminus amino acids. These residues are not found in c-FLIP L, thus the short form has a shorter half-life (Poukkula et al. 2005b). JNK activation by TNFα contributed to phosphorylation and activation of the E3 ubiquitin ligase Itch which is responsible for c-FLIP L proteasomal degradation (L. Chang et al. 2006b). As well, c-FLIP S has been reported to be ubiquitinated by the E3 ubiquitin ligase Itch (Amith Panner et al. 2009). Other studies showed the polyubiquitylation of c-FLIP in an Itch-independent manner. Suberoylanilide hydroxamic acid (SAHA, VORINOSTAT), a histone deacetylase inhibitor sensitizes breast cancer cells to TRAIL-induced apoptosis by enhancing the proteasomal degradation of c-FLIP L/S in an Itch independent pathway (Yerbes and López-Rivas 2012). Another E3 ubiquitin ligase has been identified and called TRAF7 (TNF Receptor-Associated Factor 7). It mediates JNK activation after TNFα stimulation and promotes c-FLIP L polyubiquitylation on several lysine residues such as Lys29, Lys33, Lys48 and Lys63. Importantly, Lys29 polyubiquitylation resulted in lysosomal degradation in addition to proteasomal degradation (Scudiero et al. 2012).
Phosphorylation changes also play an important role in regulating c-FLIP isoforms on post-translational levels. Phosphorylation of serine 193 (S193) by protein Kinase C (PKC) prevents c-FLIP S/R polyubiquitylation and stabilizes their expression level, thus enhancing cell survival. However, c-FLIP L phosphorylation on S193 by PKC decreases its polyubiquitylation but never affects its stability, indicating that phosphorylation of S193 has different effects on c-FLIP isoforms (Kaunisto et al. 2009). TNF-activated p38-MAPK and c- Abl promote c-FLIP S phosphorylation, thus facilitating its interaction with c-Cbl (a E3 ubiquitin ligase) and its proteasomal degradation in Mycobacterium-infected macrophages (Kundu et al. 2009). ROS (Reactive Oxygen Species) can also regulate c-FLIP protein degradation. A recent study has demonstrated that ROS triggers the phosphorylation of c- FLIP L on Threonine 166 (Thr166) and showed that this phosphorylation is necessary for the consecutive polyubiquitylation on Lys167 and c-FLIP L degradation, thus restoring prostate cancer cells sensitivity to TRAIL (Wilkie-Grantham et al. 2013).
Heat Stress, also termed “Hyperthermia” was also reported to decrease c-FLIP L and c- FLIP S protein expression via proteasomal degradation and sensitizing primary human T lymphocytes to CD95-mediated cell death (Meinander et al. 2007). In addition, our
72
collaborator Micheau O. and his team have recently demonstrated that hyperthermia induces c-FLIP aggregation and depletion from cytosolic fraction, thus preventing its recruitment into DISC and restoring TNF ligands-mediated apoptosis (Morlé et al. 2015).
V) c-FLIP: a critical target for cancer therapies
As mentioned above, c-FLIP is found at high levels in a wide variety of cancers, and its expression is correlated with tumor aggressiveness and resistance to DR-mediated apoptosis. Thus, its downregulation by different substances can overcome apoptosis resistance in tumor cells. These evidences render c-FLIP an important therapeutic target to reactivate apoptotic signaling in tumor cells (Oztürk et al. 2012). There are many types of agents that target c- FLIP, either at the transcriptional or post-transcriptional level (Safa 2013).
A) Targeting c-FLIP transcription
DNA damaging agents have proven promising effects by downregulating c-FLIP variants levels. Pretreatment with the alkylating agent and chemotherapeutic drug Cisplatin, downregulates the expression of c-FLIP S by inhibiting its transcription, but never influences c-FLIP L expression. However, Cisplatin inhibits the phosphorylation of c-FLIP L in the resistant melanoma cells, noting that phosphorylated c-FLIP L is the only form recruited to the DISC and that inhibits caspase-8 cleavage in melanoma cells. Thus, Cisplatin restores TRAIL-induced caspase8-mediated apoptosis in melanoma cells (J. H. Song et al. 2003). Doxorubicin, a DNA intercalating agent, reduces c-FLIP transcription and expression in prostate carcinoma cell lines and prostate xenografts, resulting in more TRAIL-induced apoptotic cell death and tumor growth inhibition than a treatment with TRAIL alone (El- Zawahry, McKillop, and Voelkel-Johnson 2005a). Inhibitors of histone deacetylase (HDAC) are also reported as emerging cancer therapy by regulating c-FLIP levels. Droxinostat, which selectively inhibits HDAC3, HDAC6 and HDAC8, sensitizes prostate cancer cells to Fas apoptosis by decreasing c-FLIP L gene transcription and protein expression (Wood et al. 2010). In addition, another HDAC inhibitor called Trichostatin A sensitizes ovarian cancer cells to TRAIL-apoptosis by downregulation of c-FLIP L mRNA and protein levels through downregulation of EGFR1/2 and inhibition of its downstream targets such as Akt and ERK (S.-J. Park et al. 2009). Moreover, the HDAC inhibitor Vorinostat (SAHA), potently downregulates the expression of c-FLIP in Malignant Pleural Mesothelioma (MPM) cells,
73
thus promotes caspase-8 interaction with RIPK1 and sensitizes cells to SMAC mimetic compounds (SMCs)-mediated apoptosis (Crawford et al. 2013). Inhibition of NFκB transcription factor using Quinacrine, results in downregulation of c-FLIP gene expression, a transcriptional target of NFκB, thus promoting TRAIL-apoptosis in human colon cancers (Jani et al. 2010). Although these DNA damaging compounds showed promising results in targeting c-FLIP gene expression, their use in current therapies shows some difficulties as their effects differ from a cell type to another, and sometimes they target only one variant of c-FLIP isoforms.
B) Post-transcriptionally targeting of c-FLIP
Direct inhibition of c-FLIP mRNA translation using small interfering siRNAs, represents the most effective and specific approach to silence c-FLIP expression and sensitize cancer cells to death ligands apoptosis in vitro (Day and Safa 2009a). siRNAs –mediated silencing of c-FLIP L and S promoted spontaneous apoptosis in colorectal cancer cells in a caspase-8 and FADD dependant manner. Interestingly, intratumoral delivery of siRNAs targeting c-FLIP triggered apoptosis and prevented tumor growth in colorectal cancer xenografts in BALB/c mice (T. R. Wilson et al. 2007). Besides, c-FLIP knockdown using specific siRNAs triggered apoptosis in MCF-7 breast cancer cell lines, and eliminated neoplastic cells of breast cancer xenografts without harming the normal fibroblastic and stromal cells (Day et al. 2009). siRNA-mediated downregulation of both FLIP splice forms L and S enhanced ionizing radiation (IR)-mediated cell death of NSCLC by enhancing caspase-3, -7, -8 activity and PARP cleavage (McLaughlin et al. 2016). Despite these findings, using of siRNAs in vivo showed always many complications, and clinical trials using siRNAs to target c-FLIP have not yet started.
Treatment with the methyltransferase inhibitor DZNep enhanced sensitivity of B-cell lymphoma cells to TRAIL by accelerating c-FLIP degradation. DZNep treatment did not reduce c-FLIP mRNA levels; in contrast it affected c-FLIP mRNA stability by increasing levels of c-FLIP-targeting microRNAs such as miR-512-3p and miR-346, thus leading to an enhanced c-FLIP degradation (McLaughlin et al. 2016). Anoikis is another form of apoptosis which normally occurs upon detachment of cells from ECM (Extracellular matrix). Resistance to Anoikis resulted in malignant cells survival and metastasis formation. Treatment with a small molecule called Anisomycin sensitized prostate cancer cells to the Anoikis process and
74
prevented distal tumor formation in mouse model by inhibiting c-FLIP protein synthesis (Mawji et al. 2007).
Due to the structural similarity between c-FLIP and caspase-8, it is difficult to find small molecules that inhibit c-FLIP function, as they can block caspase-8 as well, thereby inhibiting apoptosis. Some small molecules that have a broad activity on cells have been used to downregulate c-FLIP protein expression. For example, the protease inhibitor Bortezomib effectively decreased c-FLIP L protein expression and overcame resistance to TRAIL- apoptosis in myeloma cell lines (Perez et al. 2010).Treatment of kidney carcinoma cells with CHOP-inducing drugs such as Withaferin A, Thapsigargin and Brefeldin A led to a strong reduction of c-FLIP L protein level without altering its mRNA level, associated with an increase of CHOP protein. Importantly, forced expression of CHOP facilitated c-FLIP L degradation in the ubiquitin/proteasome system (Noh et al. 2012).
Although the good results obtained by siRNAs, small molecules, and DNA damaging agents that decrease c-FLIP mRNA and protein expression levels, new small molecules that inhibit directly c-FLIP function and recruitment into DISC are needed to develop new therapeutic agents.
75
Project’s objectives
TRAIL is a mediator of cell death, and it kills preferentially cancer cells. Clinical trials have demonstrated that TRAIL is well tolerated without any remarkable cytotoxicity. However, the use of TRAIL as a single agent to treat cancer patients is not well efficient and requires an association with targeted therapy. The potential of TRAIL as a monotherapeutic agent is in part limited by the upregulation of several anti-apoptotic proteins including c- FLIP, which interferes into the DISC, binds to FADD and inhibits caspase-8-mediated apoptotic machinery. Interestingly, chemotherapy targeting c-FLIP in combination with TRAIL remains a potential strategy to overcome TRAIL-resistance in tumor cells.
As discussed above, c-FLIP is extremely regulated by different pathways, and the presence of three different isoforms which are differently modulated makes c-FLIP-targeting strategies more difficult and complicated. In addition, some inhibitors can target both isoforms of c- FLIP; however some compounds alter the expression of only one isoform (Shirley et al. 2013). Moreover, because of the high structural homology between c-FLIP and the apoptotic caspase-8 protein, the development of small molecules that block c-FLIP recruitment to DISC without affecting caspase-8 is very tedious. With the exception of siRNAs, c-FLIP inhibitors downregulate c-FLIP expression in an indirect manner. So far, there is no specific compound that targets selectively and directly c-FLIP, thus the development of new chemicals that directly target and inhibit c-FLIP proteins without inhibiting caspase-8 function remains the main goal of our project. In this context, the first aim of our work was to identify, in collaboration with specialists in modelisation and molecular docking experiments, new molecules that are able to bind selectively on c-FLIP versus caspase-8, and inhibit c- FLIP/FADD interaction within DISC. Second, we aimed to test the cytotoxicity and the efficiency of the new molecules on cancer cells that overexpress the anti-apoptotic protein c- FLIP. A combination of TRAIL with each molecule was performed to evaluate the role of new molecules in TRAIL-mediated apoptosis restoration. To confirm their inhibitory role in preventing c-FLIP/FADD interaction, a molecular test of purified recombinant proteins was assessed. Thirdly, we evaluated caspases activity and cleavage to appreciate the reactivation of the apoptotic signaling pathways after molecules combination with TRAIL. Our data led us to deposit a patent with “Fonds de maturation-SATT ouest valorisation” for the new molecules we identified, and further experiments are actually in progress especially in vivo to consolidate and strengthen our invention.
76
Results
77
Identification of new c-FLIP inhibitors to restore apoptosis in TRAIL-resistant cancer cells
Katherine Yaacoub1, 3,4, Rémy Pedeux1, 4, Pierre Lafite2, Samia ACI-Sèche2, Pascal Bonnet2, Richard Daniellou2, Thierry Guillaudeux1, 3,4 1 Université Rennes 1, 2 Rue du Thabor 35000 Rennes, France ;
2 Université Orléans, CNRS, ICOA, UMR 7311, F-45067 Orléans, France
3 UMS CNRS3480/US 018 INSERM BIOSIT, 2 Av. du Pr Léon Bernard, 35043 Rennes cedex, France
4 INSERM U1242 COSS, CLCC Eugène Marquis, Rue de la Bataille Flandres-Dunkerque, 35042 Rennes, France
c-FLIP, a catalytically inactive caspase-8-homologous protein, is a potent antiapoptotic protein highly expressed in various types of cancers. c-FLIP competes with caspase-8 for binding to the adaptor protein FADD (Fas-Associated Death Domain) following Death Receptors (DRs) activation, thereby blocking the extrinsic apoptotic machinery. Inhibition of c-FLIP activity might enhance DRs-mediated apoptosis and overcome anticancer drugs resistance.
The aim of this work was to identify, based on in silico methods, new molecules that are able to bind selectively to c-FLIP and block its anti-apoptotic activity. Using a homology 3D model of c-FLIP, in silico screening of 1880 compounds from the NCI database (National Cancer Institute) was performed. Nine molecules were selected for in vitro assays, exhibiting the strongest binding affinity on c-FLIP and the highest selectivity versus caspase-8.
Using human lung cancer cell line H1703 that overexpress c-FLIP protein, the inhibitory effect of these nine molecules was tested. Our results showed that treatment of H1703-cFLIP with TRAIL alone had no effect on cell death, because of the anti-apoptotic role of c-FLIP. However, the combination of TRAIL with selected molecules significantly enhanced TRAIL- mediated apoptosis. Moreover, the newly identified molecules were able to prevent FADD/c- FLIP interactions in a molecular pull-down assay, in addition to their role in rescuing caspases activation and cleavage in TRAIL-treated tumor cells. All together, our findings indicate that these molecules could efficiently prevent c-FLIP recruitment into the DISC complex, thus restoring the caspase-8 apoptotic cascade. These results pave the way to design new c-FLIP inhibitory compounds that may serve as anticancer agents in tumors overexpressing c-FLIP.
Key words: c-FLIP, TRAIL, apoptosis, protein-protein interaction, drug resistance, cancer treatment
78
Introduction
The development of chemotherapeutic drug resistance is one of several clinical problems causing anti-cancer drugs limited efficacy or failure. The identification of chemotherapy resistance mechanisms helps to design novel strategies(ROBERTI, LA SALA, and CINTI1 2006; Safa 2004). Over the past years, evidence has shown that abnormalities in apoptosis signaling pathways such as activation of anti-apoptotic proteins are highly associated with drug resistance(Mashima and Tsuruo 2005).
Among these proteins, c-FLIP (Cellular FLICE Inhibitory Protein) is a major anti-apoptotic and resistance protein that restrains apoptosis induced by TNF (Tumor Necrosis Factor) superfamily members, including TRAIL (TNF-Related Apoptosis Inducing Ligand), Fas-L, TNF α, et …as ell as apoptosis stimulated by chemotherapeutic drugs in cancer cells(Seol, Mihich, and Berleth 2015). c-FLIP is overexpressed in many types of human cancers such as ovarian carcinoma(Mezzanzanica et al. 2004), colorectal carcinoma(Longley et al. 2006), gastric adenocarcinoma(Nam et al. 2003b), prostate carcinoma(X. Zhang et al. 2004). An upregulated level of c-FLIP has been detected also in primary tissues from patients having lung adenocarcinoma(Brambilla, Brambilla, and Gazzeri 2006), hepatocellular carcinoma(N. S. Wilson, Dixit, and Ashkenazi 2009) and B-cell chronic lymphocytic leukemia (Marion MacFarlane et al. 2002). As well, the analysis of primary cells from patients showed an over expression of c-FLIP in melanomas(Bullani et al. 2001) a d Hodgki s l pho as(Mathas et al. 2004). c-FLIP has 13 distinct variants, only three of them are expressed as proteins in human cells. These are c-FLIP (L), c-FLIP (s) and c-FLIP(R)(Olivier Micheau 2003a). The long form c-FLIP (L) 55 kDa is similar to procaspase-8, containing N-terminal tandem DEDs (Death effector domains), and a C-terminal caspase-like domain, which lacks the catalytic cysteine residue, responsible for the proteolytic activity of caspases. The short form c-FLIP (s) (26 kDa) is composed only of two DEDs without caspase-Like domain and a short C-terminus(L. Chang et al. 2006a). Another short form protein of 27 kDa called c-FLIP (R) is particularly expressed in a number of T and B cells such as Raji cells, and human primary T cells. It also contains N- terminal DEDs but a short C-terminal composed of a stretch of residues playing a key role in the ubiquitinylation of c-FLIP proteins(Golks et al. 2005; L. Chang et al. 2006a) c-FLIP is considered as a key inhibitor of the extrinsic apoptotic pathway by preventing the homodimerization and autoactivation of procaspase-8/10, the initiator factor of apoptosis. This extrinsic pathway, also called death receptor pathway, is induced by the binding of different death ligands of the TNF superfamily (TRAIL, Fas-L, TNFα to their respective death receptors DRs (TRAIL- R1/R2, Fas, TNF receptor). This binding induces the trimerization of DRs which in turn recruit the adaptor protein FADD(J. W. Kim, Choi, and Joe 2000). Once FADD is recruited to DRs, procaspase-8/10 binds to FADD through the interaction between their DEDs, leading to DISC (Death Inducing Signaling Complex) formation, thereby the activation of downstream caspases and apoptosis events(Boldin et al. 1996). However, apoptosis machinery is attenuated by c-FLIP interference. First, it was supposed that c-FLIP impedes the recruitment of procaspase-8 to the DISC, thereby precludes its activation(Rasper et al. 1998a). Then, Scaffidi and coworkers have contradicted this theory and demonstrated that caspase-8 is always recruited to the DISC at the same time as c- FLIP(s/L) proteins(Carsten Scaffidi et al. 1999). Procaspase-8 forms a heterodimeric complex
79
with c-FLIP (L), resulting in an incomplete cleavage and limited activation of caspase-8 because of the lack of enzymatic activity of c-FLIP (L). This heterodimerization prevents further apoptotic signals transduction. However, the over expression of c-FLIP(s) can completely inhibit the processing of caspase-8 at the DISC, thus blocking the activation of the apoptotic cascade. These findings reflect different functional roles of c-FLIP(L) and c- FLIP(s) in the mechanisms of apoptosis inhibition(Carsten Scaffidi et al. 1999; Andreas Krueger et al. 2001).
TRAIL, also called APO-2L, is a member of TNF family which is mainly expressed by immune cells. It is a type II transmembrane protein with a C-terminal extracellular domain, cleaved by a cysteine protease resulting in a soluble form(Kimberley and Screaton 2004). Five distinct receptors have been identified to recognize and bind TRAIL. TRAIL-R1 (DR4) and TRAIL-R2 (DR5), are classical death receptors and are able to trigger apoptosis as they contain a functional cytoplasmic death domain (DD). TRAIL-R3 (DcR1), TRAIL-R2 (DcR2) also known as decoy receptors, as well the circulating receptor Osteoprotegerin (OPG) are not able to propagate the death signals due to death domain absence(Pan et al. 1997). TRAIL is considered as a potent anti cancer agent since it has been proven that it kills preferentially cancer cells in a wide variety of tumors, and does not have any toxicity in the majority of normal cells. However, a large number of cancers evade TRAIL-induced apoptosis and get TRAIL resistance through different mechanisms including the over expression of c-FLIP which prevents DR-induced apoptosis(Shulin Wang and El-Deiry 2003). Selective knock-down of c- FLIP(L) expression sensitizes tumor cells to TRAIL-induced cell death in human lung cancer cell lines(Sharp, Lawrence, and Ashkenazi 2005a). As well, it has been demonstrated that Withanolide E, a steroidal lactone derived from Physalis peruviana, can highly sensitize renal carcinomas cells and other human cancer cells to TRAIL-mediated apoptosis through a rapid destabilization, aggregation and proteasomal degradation of c-FLIP proteins, confirming the key role of c-FLIP in protecting cells from death ligands-mediated apoptosis(Henrich et al. 2015). So far, the inhibitors of c-FLIP that have been studied, apart from siRNAs, act indirectly on c- FLIP, such as Cisplatin which induces p53-dependent FLIP ubiquitination and degradation in ovarian cancer cells(Abedini et al. 2008), or Actinomycin D which downregulates FLIP(L) and FLIP(s) expression in B chronic lymphocytic leukemia(Olsson et al. 2001). Thus, the identification of molecules that target directly c-FLIP could be a new strategy to overcome chemotherapy resistance. c-FLIP is structurally similar to caspase-8, each DED of c-FLIP shares 25% similarity with DEDs of caspase-8, and the C-terminus 270 amino acid of c-FLIP are also 25% identical to caspase-8 C-terminus (Shu, Halpin, and Goeddel 1997). Thus, due to this homology,∼ the identification of new compounds that bind selectively on c-FLIP versus caspase-8∼ and prevent its recruitment to the DISC represents a major challenge. Based on these concepts, molecular modeling and docking experiments were set up to construct c-FLIP and caspase-8 homology models in order to find c-FLIP selective inhibitors. In vitro assays using recombinant c-FLIP(S) and FADD, coupled to in cellulo assays demonstrated the inhibitory role of these new compounds by restoring TRAIL-mediated apoptosis, by preventing specifically c-FLIP/FADD interactions. Our findings suggest that blocking c-FLIP recruitment into the DISC by specific inhibitors decreases tumor resistance to death receptors mediated apoptosis.
80
Materials and methods
Molecular modeling 1) Homology modeling of DED2 domains At the time of the first experiments, no crystallographic structure was available for the DED2 domains of c-FLIP and CASP8 target domain, so some homology models were built using MOE software (Molecular Operating Environment). First, sequences of the DED2 domains of c-FLIP and of CASP8 have been extracted from the Uniprot database and used to find homolog structures in the Protein Data Bank (PDB) using the BLAST software. Three sequences showed a reasonable percentage of identity with the target sequences (around 30%) to perform homology modeling. Three structures of the FADD DED domain (PDB.ID 1A1W, 1A1Z, 2GF5) and three structure of two v-FLIP DED2 domains (PDB.ID 2BBR, 2BBZ, 3CL3) were identified as suitable templates for modeling the DED2 domains of CASP8 and c- FLIP respectively 2) Identification of the binding site The goal of the docking study was to identify compounds able to prevent the interaction of the DED2 domain of c-FLIP with the DED domain of FADD. No precise localization of the interacting site was known, however published studies outlined the role of a conserved hydrophobic patch F-L in this interaction (Carrington et al, 2006, Dickens et al, 2012). We have thus used the SiteFinder module of the MOE software to identify the druggable pockets at the surface of our models and we only retained homology models which presents such pockets at the vicinity of the F-L hydrophobic patch. Two models of the CASP8 DED2 domain and one model of the c-FLIP DED2 were kept based on the PDB.ID 1A1W, 2GF5 and 2BBZ structure templates respectively.
3) Docking of chemical libraries The 1880 molecules of the NCI DiversitySet3 extracted from the ZINC database were virtually screened on the three models using two docking softwares AutoDock and Glide. Indeed, the ZINC database proposes ready-to-dock sets of commercially available molecules. The results of the two virtual screening were then combined using a consensus scoring method and a root mean square deviation (RMSD) filter.
4) Consensus scoring function In addition to the Glide and AutoDock scoring functions, the MOE GBVI/WS dG function was selected to rescore all poses of the docked ligands. Therefore, the binding modes of each docked ligand with Glide and AutoDock was finally assessed using four score values. The score values were normalized using the Z-score formula