Inhibitor Lurbinectedin and Oncolytic Peptide LTX-401 trigger and Synergize With Immune Checkpoint Blockade Wei Xie

To cite this version:

Wei Xie. Transcription Inhibitor Lurbinectedin and Oncolytic Peptide LTX-401 trigger Immunogenic Cell Death and Synergize With Immune Checkpoint Blockade. . Université Paris-Saclay, 2020. English. ￿NNT : 2020UPASL072￿. ￿tel-03186843￿

HAL Id: tel-03186843 https://tel.archives-ouvertes.fr/tel-03186843 Submitted on 31 Mar 2021

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés.

INDEX

ABSTRACT ...... 1 RESUME...... 2 ORIGINAL PAPERS ...... 3 ABBREVIATION ...... 4 INTRODUCTION ...... 7 Cancer and anticancer immunity ...... 7 Cancer epidemiology ...... 7 Cancer immunology ...... 8 Immunogenic cell death in cancer therapy ...... 13 Major DAMPs of ICD ...... 16 Methods to detect ICD ...... 25 ICD inducers ...... 27 Induction of ICD sensitize to immune checkpoint blockade ...... 31 Transcription inhibitor Lurbinectedin ...... 33 Oncolytic peptide LTX-401 ...... 36 AIMS OF THE THESIS ...... 38 RESULTS ...... 39 Lurbinectedin synergizes with immune checkpoint blockade to generate anticancer immunity ...... 39 Tumor lysis with LTX-401 creates anticancer immunity ...... 41 CONCLUDING REMARKS ...... 42 ACKNOWLEDGEMENT ...... 45 REFERENCES ...... 46 ANNEX 1: SCIENTIFIC PUBLICATIONS ...... 66 ANNEX 2: PAPERS NOT INCLUDED IN THIS THESIS (INSERTED IN PDF) ...... 69

ABSTRACT

Cancer is the second leading cause of death worldwide, despite the existence of standard treatment, innovative therapeutic strategies and drugs are still in urgent demand. The combination of immunogenic cell death (ICD) inducing drugs and immune checkpoint blockade (ICB) seems to be a promising modality. In this thesis, we demonstrated Lurbinectedin, a transcription inhibitor newly approved for relapsed lung cancer treatment, triggers hallmarks of ICD in four different human and murine cell line in vitro. Vaccinated with Lurbinectedin-treated fibrosarcoma cell protects immunocompetent mice from rechallenge of syngeneic tumours. Lurbinectedin restrains transplanted fibrosarcoma growth in an immune dependent manner. Both transplanted MCA205 cancer and hormone/carcinogen induced breast cancer were sensitized by Lurbinectedin to PD-1 and CTLA- 4 double ICBs. Of note, long-term immunological memory was generated in cured mice. Further, we evaluated the anticancer capacity of LTX-401, an oncolytic peptide designed for local immunotherapy. Sequential intratumoral injections of LTX-401 dramatically retards subcutaneous MCA205 and TC-1 tumour growth in immunocompetent host, yet shows limited therapeutic effect on abscopal syngeneic tumours. Single LTX-401 administration boosts the efficacy of anti-CTLA-4 or anti-PD-1/anti-CTLA-4 ICBs. Moreover, sequential LTX-401 treatment with double ICBs exhibits systemic antitumor immunity to both treated and abscopal tumour. In conclusion, lurbinectedin and LTX-401 induce cancer cell immunogenic cell death and enhance the anticancer effects of immune checkpoint blockade. These results lay the experimental foundation of combination regiments and may facilitate the clinical trial designs.

Key words: cancer, immunogenic cell death, immune checkpoint blockade, transcription, oncolytic peptide

1

RESUME

Le cancer est la deuxième cause de mortalité dans le monde. Malgré l’existence des traitements standards, le développement et la recherche de stratégies thérapeutiques innovantes et de médicaments est toujours nécessaire. La combinaison des médicaments, induisant la mort cellulaire immunogène (ICD) et l’inhibition des points de contrôle immunitaire (ICB), semble être un protocole prometteur. Dans cette thèse, nous avons démontré que la Lurbinectédine, un inhibiteur de la transcription nouvellement approuvé pour le traitement du cancer du poumon récidivant, déclenche les caractéristiques de l’ICD dans quatre différentes lignées cellulaires humaines et murines in vitro. Vaccinée par des cellules de fibrosarcome traitées par la lurbinectine, les souris immunocompétentes sont protégées lors du rechallenge des tumeurs syngéniques. La lurbinectédine limite la croissance du fibrosarcome transplanté d'une manière immunodépendante. Dans les souris, le fibrosarcome murin (MCA205) transplanté et le cancer du sein, induit par des hormones en combinaison avec des cancérigènes, ont été sensibilisés par la lurbinectédine aux deux ICB : PD-1 et CTLA-4. Il convient de noter que la mémoire immunologique à long terme a été générée chez des souris guéries. En outre, nous avons évalué la capacité anticancéreuse de LTX-401, un peptide oncolytique conçu pour l'immunothérapie locale. Les injections intratumorales séquentielles de LTX-401 retardent considérablement la croissance des tumeurs sous-cutanées MCA205 et TC-1 chez un hôte immunocompétent, mais montrent un effet thérapeutique limité sur les tumeurs syngéniques abscopales. Une seule administration de LTX-401 augmente l'efficacité des ICB anti-CTLA-4 ou anti-PD-1 + anti-CTLA-4. De plus, le traitement séquentiel avec LTX-401 et les deux ICB présente une immunité antitumorale systémique à la fois contre la tumeur traitée et la tumeur abscopale. En conclusion, la lurbinectédine et le LTX-401 induisent la mort cellulaire immunogène des cellules cancéreuses et renforcent les effets anticancéreux des inhibiteurs de points de contrôle immunitaires. Ces résultats jettent les bases expérimentales de traitements combinés et peuvent faciliter les conceptions d'essais cliniques.

Mots clés: cancer, mort cellulaire immunogène, inhibiteur de point de contrôle immunitaire, transcription, peptide oncolytique

2

ORIGINAL PAPERS

This thesis is based on the following papers:

Paper I. Lurbinectedin synergizes with immune checkpoint blockade to generate anticancer immunity. Xie W, Forveille S, Iribarren K, Sauvat A, Senovilla L, Wang Y, Humeau J, Perez-Lanzon M, Zhou H, Martínez-Leal JF, Kroemer G, Kepp O. Published in Oncoimmunology, 2019, VOL. 8, NO. 11, e1656502 (9 pages).

Paper II. Tumor lysis with LTX-401 creates anticancer immunity. Xie W, Mondragón L, Mauseth B, Wang Y, Pol J, Lévesque S, Zhou H, Yamazaki T, Eksteen JJ, Zitvogel L, Sveinbjørnsson B, Rekdal Ø, Kepp O, Kroemer G. Published in Oncoimmunology, 2019, VOL. 8, NO. 7, e1594555 (8 pages).

(In agreement with the doctoral school, Paper I&II are inserted to replace chapter ‘Materials and methods’ and ‘Results’.)

3

ABBREVIATION

AIDS acquired immune deficiency syndrome ATF activating transcription factor AML acute myeloid leukaemia ATP adenosine triphosphate ANXA1 annexin A1 APCs antigen-presenting cells AMPs antimicrobial peptides LfcinB bovine lactoferricin BFA brefeldin A CALR calreticulin CRC colorectal cancer CXCL10 CXC-chemokine ligand 10 CTLA-4 cytotoxic T-lymphocyte-associated 4 DAMPs damage-associated molecular patterns DC dendritic cell DMBA dimethylbenzantracene DOXO ER endoplasmic reticulum ERAD ER-associated degradation eIF2α eukaryotic translation initiation factor 2 FDA federal drug administration FPR1 formyl peptide receptor 1 GCN2 general control nonderepressible 2 Hsp70 heat shock protein 70 HCC hepatocellular carcinoma HMGB1 high-mobility group box 1 IFNGR1 IFN-γ receptor ICB immune checkpoint blockade ICD immunogenic cell death

4

IRE1α inositol-requiring enzyme 1α IFN interferon IFNAR1 interferon α/β-receptor subunit 1 IFN-γ interferon γ IL interleukin RIDD IRE1α-dependent decay KS Kaposi's sarcoma LPS lipopolysaccharides LRP1 low density lipoprotein receptor-related protein 1 MPA medroxyprogesterone acetate MTX mitoxantrone MYD88 myeloid differentiation primary response 88 NK natural killer NSCLC non-small cell lung cancers NF-κB nuclear factor-κB ORR overall response rate OS overall survival PERK PKR-like ER-resident kinase PD-1 programmed cell death protein 1 PD-L1 programmed death-ligand 1 PKR protein kinase R P2RX7 purinergic receptor P2X7 P2RY2 purinergic receptor P2Y2 ROS reactive oxygen species RAGE receptor for advanced glycation end-products RAG recombination Activating ROCK1 Rho-associated, coiled-coil containing protein kinase 1 SCID severe combined immunodeficiency STAT signal transducer and activator of transcription SCLC small-cell lung cancer TLR Toll-like receptor

5

UPR unfolded protein response

6

INTRODUCTION

Cancer and anticancer immunity

Cancer epidemiology

Cancer, which was first recorded by Edwin Smith in 3000 B.C. and described in a scientific manner by Hippocrates in 400 B.C., refers to a series of diseases due to uncontrollably replicating cells. According to WORLD HEALTH STATISTICS:2018, cancer has become a serious global public health problem, approximately 9 million deaths were caused by cancer in 2016, which accounts for 15.8% of total 57 million deaths all over the world . As a fatal disease, cancer ranks as the 1st leading cause of premature death (30 to 69 years old) in 55 countries, including USA, japan and the majority of European countries. In other 79 countries, cancer has been the second most important killer among premature populations. Specifically, more than 1.8 million Americans are estimated to be diagnosed as new invasive cancer cases in 2020, and nearly 600 000 deaths will occur in USA in the same year (Cancer statistics, 2020). In 2018, the top 3 common cancers in men are prostate cancer, lung cancer and colorectal cancer, which contributed almost 40% of total cancer incidence. Breast cancer (30%) is the most common cancer in women, followed by lung cancer (13%) and colorectal cancer (7%) (Cancer Facts & Figures 2018) (Figure 1).

Figure 1. Leading Sites of New Cancer Cases and Deaths – 2018 Estimated (Cancer Facts & Figures 2018)

7

Cancer immunology

The has been defined as a host defence system which could recognise non- self-compartment and protect organism from a lot of diseases, especially for infectious disease. The immune system can be divided into two parts: innate immunity and adaptive immunity. More and more researches demonstrated the pivotal function of innate and adaptive immunity in anti- cancer immune response (Coca et al., 1997; DeNardo et al., 2011; Donadon et al., 2017; Fridman et al., 2012; Hildner et al., 2008; Ishigami et al., 2000; Keizman et al., 2012; Pylayeva-Gupta et al., 2016). From cancer immunosurveillance to cancer immunoediting In the early 1900s, Paul Ehrlich firstly hypothesized that immune system could recognises cancer cells as extrinsic components and eliminates them effectively (Burnet, 1970; Ehrlich, 1909). After around half centenary, the theory of ‘immunosurveillance’ was formally proposed by Macfarlane Burnet and Lewis Thomas based on the availability of inbred strain of mice and accumulated evidence of ‘tumour-specific antigen’, they speculated that the non-stop emerging transformed cells could provoke immune response and be recognised and eliminated by lymphocytes before it becomes clinically detectable (Burnet, 1957; Klein, 1966; Old and Boyse, 1964). In the next dozens of years, the concept of immunosurveillance was in great debate. One of the representative studies which challenged this theory was conducted by Osias Stutman. Stutman observed the phenomenon that compared to genetically matched wild-type mice, carcinogen injection neither induced more tumours nor shorten tumour latency period in CBA/H strain nude mice (Stutman, 1974). Moreover, similar results obtained by different groups conclusively destroyed the bases of immunosurveillance theory and led to the abandonment of this hypothesis (Outzen et al., 1975; Rygaard and Povlsen, 1974a, b; Stutman, 1979; Thomas, 1982). Due to the solid data aforementioned, the concept of cancer immunosurveillance faded away since then. The resuscitate of cancer immunosurveillance started from 1990s, due to the application of immunodeficiency mice on pure genetic backgrounds. Between 1990 and 2000, the function of two fundamental immunological components, interferon γ (IFN-γ) and perforin, in controlling primary tumour formation were revealed. In the first finding, researchers observed that transplanted fibrosarcoma grows faster when host

8

endogenous IFN-γ was neutralized by blocking antibodies (Dighe et al., 1994). Also, sarcoma cells which express dominant negative IFN-γ receptors (IFNGR1) showed increased tumorigenicity when implanted into syngeneic immunocompetent mice (Dighe et al., 1994). Both IFN-γ receptor deficient mice and STAT1-/-mice in 129/SvEv background are much more sensitive to MCA carcinogen-induced fibrosarcoma (Kaplan et al., 1998). Another study which employed different genetic background’s IFN-γ-/- mice, reinforced the protective function of endogenous IFN-γ in restraining spontaneous tumour formation and transplanted tumour growth (Street et al., 2001). The next demonstrated key factor is perforin. Perforin is a glycoprotein which is mainly produced by natural killer (NK) cells and CD8+ T cells, upon interacting with target cell’s membrane, channels and pores formed subsequently (Osinska et al., 2014; Russell and Ley, 2002). Similar to IFN-γ-/- mice, perforin-/- mice developed more tumours and showed shorter tumour latency time post MCA incubation (Smyth et al., 2000; Street et al., 2001; van den Broek et al., 1996). In conclusion, these findings indicated the indispensable role of immunological components in subverting cancer development. In 2001, a seminal study conducted by R.D. Schreiber’s group in Washington University renovated the cancer immunosurveillance concept (Shankaran et al., 2001). In this 5-pages short assay, firstly they showed inbred mice lacking either lymphocytes (RAG2–/–) or IFN-γ response (IFNGR1–/– /STAT1–/–) were prone to grow MCA-induced tumour. Second, all of the RAG2–/– or RAG2–/– /STAT1–/– mice developed spontaneous tumours in the age of 15-21 months, supported the theory of immunosurveillance. Next, they observed an interesting phenomenon that when transplanted MCA induced tumours into wild type naïve mice, all of the tumours (17/17) derived from immunocompetent mice grew normally. On the contrary, 40% of tumours (8/20) derived from RAG2–/– immunodeficient mice were rejected post transplantation. The tumour rejection effect was abrogated when RAG2–/– mice were used as recipients. This ingeniously designed experiment proved the qualitative differences between tumours from immunocompetent mice and immunodeficient mice, strongly supports the existence of tumour immunogenicity alteration during cancer development in immunocompetent hosts. In the last part, the authors proved that tumour rejection was T cell dependent. Consistent results were obtained by other group by using nude mice, SCID mice or TCR Jα281–/– mice (Engel et al., 1997; Smyth et al., 2000; Svane et al., 1996).

9

Given the fact that cancer occurs in the presence of immunosurveillance, and cancer cells with lower immunogenicity are more likely to survive compared to highly immunogenic ones, R.D. Schreiber and his colleagues proposed a more comprehensive immune theory named Immunoediting. This term was further elucidated as three different phases ‘elimination,’ ‘equilibrium,’ and ‘escape’ (Kim et al., 2007; Schreiber et al., 2011) (Figure 2). The elimination phase could be regarded as updated immunosurveillance, in this phase transformed cells undergo stress and release danger signals such as reactive oxygen species and inflammatory cytokines (Kim et al., 2007; Matzinger, 1994). These compartments could be sensed by different types of innate immune cells like dendritic cell, and NK cell, subsequently promote their activation. The activated innate immune cells could either directly kill transformed cells or trigger adaptive immune response to achieve clearance. Of note, both innate immune system and adaptive immune system are indispensable in eliminating transformed cell. If early tumours are eliminated, this phase is the end of the whole process. However, if a small fraction of transformed cells survived from elimination phase, the subsequent equilibrium phase starts. In this phase, tumour cells are in a special state named ‘functional dormancy’ and under consistent immune pressure. In this circumstance, cancer cells keep evolving to gain resistance to immune recognition or induce an immunosuppressive tumour microenvironment, resulting in the generation of less immunogenic tumour clones. Importantly, cytokine IL-12 and IL-23 were reported to play opposing roles in this process (Teng et al., 2012). This phase is a prolonged process, which may last from months to years, even to the end of host’s life. Anti-tumour immunity and tumour promoting factors are in dynamic balance. As a result, tumour growth is strictly controlled by adaptive immune system and tumour immunogenicity is sculpted in the meantime. Particularly, innate immune response seems inessential in this period (Kim et al., 2007; Mittal et al., 2014). In the escape phase, evolved tumour variants evade immunological control and grow progressively, leading to visible tumours. The escaping mechanisms are extensively discussed but not fully elucidated, the balance is towards to tumour progression perhaps due to poor immunogenicity or immunologic cytotoxicity resistance of cancer cells, or by the gradually established immunosuppressive tumour microenvironment (Khong and Restifo, 2002; Radoja et al., 2000). The vast majority of clinically diagnosed cancers are in escape phase and immune

10

system can no longer work alone to control tumour growth, therefore appropriate therapeutic strategies should be taken to achieve favourable outcomes.

Figure 2. The cancer immunoediting concept. Cancer immunoediting is an extrinsic tumor suppressor mechanism that engages only after cellular transformation has occurred and intrinsic tumor suppressor mechanisms have failed. In its most complex form, cancer immunoediting consists of three sequential phases: elimination, equilibrium, and escape (Schreiber et al., 2011).

11

Cancer immunoediting in humans To investigate the existence of cancer immunoediting in humans, cancer incidence in immunodeficient or immunosuppressive populations are studied. Compare to general population, acquired immune deficiency syndrome (AIDS) patients are prone to several types of cancer, especially for virally induced tumour like Kaposi's sarcoma (KS) (Mitsuyasu, 2009). As a consequence of immune-suppressive treatment, patients who received transplant organs show higher risk of cancer. Birkeland analysed the cancer risk of 4178 patients who received renal transplantation between 1964 to 1986 in the Nordic countries and demonstrated an overall 3- to 5- fold increase in the combined cohort. Specifically, the incidence of lip cancer, colon and rectum cancer, non- skin cancer, melanoma, lung cancer, thyroid cancer, non-Hodgkin’s lymphoma elevated strikingly in both men and women (Birkeland et al., 1995). In another transplant recipients’ study, the malignancy incidence of 905 patients who received heart or/and lung between 1989 and 2004 were evaluated. Among which 102 patients were diagnosed with cancer, more than 7 times higher than general population. The report revealed that lymphoproliferative disorders, head and neck cancer and lung cancer rank the top three common cancer types in the cohort (Roithmaier et al., 2007). Of note, a review report from Cincinnati Transplant Tumour Registry showed two-fold higher melanoma incidence in transplant patients as compared with general population (Penn, 1996). Apart from that, other highly convincing results demonstrated the correlations between infiltrating immune cells and prognosis of patient, which strongly supported the principle of immunoediting theory (Fridman et al., 2012).

12

Immunogenic cell death in cancer therapy

Depending on the capacity of activating an adaptive immune response, regulated cell death can be categorized as either immunogenic cell death (ICD) or non-immunogenic cell death (Galluzzi et al., 2018). As a host defence approach, cells infected with virus or intracellular bacterial undergo ICD release antigens and alarmins, which stimulates immune response to eliminate pathogens. The very first demonstrated ICD inducing drug was doxorubicin, other chemotherapeutic agents like oxaliplatin, cyclophosphamide, or some physical approaches (radiotherapy, photodynamic therapy, high hydrostatic pressure) were also proved to be able to induce ICD (Apetoh et al., 2007; Casares et al., 2005; Fucikova et al., 2014; Garg et al., 2012a; Obeid et al., 2007a; Tesniere et al., 2010; Vacchelli et al., 2014). These agents could trigger the emission of damage-associated molecular patterns (DAMPs) from cancer cells, resulting in innate and adaptive immune cell activation and immunological memory generation (Kepp et al., 2014; Lotze et al., 2007). The major DAMPs and cytokines involved in ICD include calreticulin (CALR), adenosine triphosphate (ATP), high‑mobility group box 1 (HMGB1), Type I interferon (IFN), CXC ‑ chemokine ligand 10 (CXCL10), annexin A1 (ANXA1) and so on. (Galluzzi et al., 2020b)(Table1). Of note, the same pathway in different forms of ICD may exhibits opposing effects (Garg et al., 2013). In the context of ICD, CALR translocate from endoplasmic reticulum (ER) to cell surface, facilitating antigen uptake by dendritic cell (DC) (Gardai et al., 2005; Obeid et al., 2007b). ATP released by dying tumour cell could be sensed by DC through purinergic receptor P2Y2 (P2RY2) and purinergic receptor P2X7 (P2RX7), which mediate DC recruitment and activation respectively (Elliott et al., 2009; Ma et al., 2013; Michaud et al., 2011). HMGB1 emitted from cancer cell nuclear binds to Toll‑like receptor 4 (TLR4) to promotes DC maturation (Sims et al., 2010).

Tumour cell secreted Type I IFN interacts with self-interferon α/β‑receptor subunit 1 (IFNAR1) to trigger autocrine and paracrine circuitries, leading to the release of CXCL10. As a chemoattractant, CXCL10 recruits both αβ and γδ T cells to tumour site (Harding et al., 2017; Mackenzie et al., 2017; Sistigu et al., 2014; Vanpouille-Box et al., 2017). Extra extracellular ANXA1 from cancer cell engages with formyl peptide receptor 1 (FPR1) on DC, resulting in the formation of corpse/DC synapses (Baracco et al., 2019; Vacchelli et al., 2015). Together, these

13

processes convert dying tumour cells into ‘cancer vaccines’, which activates DC and subsequently mediate T cell priming and memory T cell induction (Galluzzi et al., 2017)(Figure 3).

Factor Class Effect Main receptor(s)

ANXA1 Surface proteinDirects APCs to dying cellsFPR1

Promotes the recruitment, maturation and cross- ATP Nucleotide P2RX7 P2RY2 presentation activity of APCs

Promotes the uptake of dying cells and type I IFN secretion CALR ER chaperone LRP1 by APCs

CCL2 Cytokine Promotes T cell and neutrophil recruitmentCCR2

CXCL1 Cytokine Promotes T cell and neutrophil recruitmentCXCR2

CXCL10 Cytokine Promotes T cell and neutrophil recruitmentCXCR3

Cytosolic Promotes the secretion of type I IFN and other Nucleic acid MDA5 RIG-I TLR3 RNA proinflammatory factors

Cytosolic Promotes the secretion of type I IFN and other Nucleic acid AIM2 CGAS ZBP1 DNA proinflammatory factors

ERp57 ER chaperonePromotes the uptake of dying cells by APCsLRP1 (?)

Extracellular Nucleic acidPromotes the recruitment and activation of neutrophils TLR9 DNA

F-actin Cytoskeletal componentPromotes the uptake of dying cells by APCsCLEC9A

Promotes the maturation and cross-presentation activity of HMGB1 Nuclear DNA-binding protein AGER TLR2 TLR4 APCs

HSP70 ER chaperoneFavors the uptake of dying cells b y APCsLRP1

HSP90 ER chaperoneFavors the uptake of dying cells b y APCsLRP1

TFAM Transcription factorPromotes APC maturation and recruitmentAGER

Promotes APC maturation, cross-presentation, and T cell Type I IFN Cytokine IFNARs recruitment Table 1. Major immunostimulatory DAMPs and cytokines mechanistically linked to ICD in cancer. Adapted from (Galluzzi et al., 2020b).

14

Figure 3. Mechanisms of -driven ICD. In response to inducers of immunogenic cell death (ICD), such as doxorubicin or oxaliplatin, malignant cells expose calreticulin (CALR) and other endoplasmic reticulum chaperones on their surface, secrete ATP, initiate a cell-intrinsic type I interferon (IFN) response culminating in the production of CXC-chemokine ligand 10 (CXCL10), and release high- mobility group box 1 (HMGB1) and annexin A1 (ANXA1). Upon binding to cognate receptors on the surface of myeloid or lymphoid cells, these damage-associated molecular patterns favour the uptake of cell corpses and debris thereof by antigen-presenting cells, including dendritic cells (DCs) in the context of robust immunostimulatory signals, eventually leading to the priming of an adaptive immune response involving both αβ and γδ T cells. In addition to being associated with the establishment of immunological memory, such a response has the potential to eradicate malignant cells that survive chemotherapy via an IFNγ-dependent mechanism. CXCR3, CXC-chemokine receptor 3; FPR1, formyl peptide receptor 1; IFNAR1; interferon α/β-receptor subunit 1; IL, interleukin; LRP1, LDL receptor related protein 1; P2RX7, purinergic receptor P2X7; P2RY2, purinergic receptor P2Y2; TLR4, Toll-like receptor 4 (Galluzzi et al., 2017).

15

Major DAMPs of ICD

Calreticulin exposure As one of the most abundant in ER, CALR plays fundamental role in calcium and protein folding (Michalak et al., 1999). Upon ICD treatment, CALR translocate from ER to cell surface in cooperate with other chaperone proteins (Asadzadeh et al., 2020; Fucikova et al., 2011; Garg et al., 2012b; Panaretakis et al., 2008)(Figure 4). As an ‘eat me’ signal, cells exhibiting CALR on surface can be recognised by phagocytes through low density lipoprotein receptor-related protein 1 (LRP1; also known as CD91), resulting in antigen uptake and clearance of cellular components (Bergmann et al., 1992; Garg et al., 2012b). On the contrary, this process is impaired when CD47 (a don’t eat me signal) co-expressed on cancer cells (Chao et al., 2011). Also, CALR-CD91 conjunction on antigen-presenting cells (APCs) primes Th-17 cell response in an Interleukin 6 (IL-6) and tumor necrosis factor alpha (TNF-α) dependent manner (Pawaria and Binder, 2011). Of note, CALR ablation by siRNA or neutralising antibody abolish the immune response towards dying tumour cell in vivo, while incubation with recombinant CALR protein restores the immunogenicity of the cells (Liu et al., 2019; Obeid et al., 2007b). Moreover, the immunogenicity deficiency mediated by PERK, -8 or SNAREs depletion in anthracyclines induced cell death can also be restored by absorbing recombinant CRT to the cell surface (Panaretakis et al., 2009). Altogether, these results definitely indicated that the immunogenicity of ICD is defined by cell surface CALR. Importantly, CALR exposure is correlated with clinical outcomes. In human non-small cell lung cancers (NSCLC), high expression of CALR correlates with increased antitumor immune response and favourable prognosis (Fucikova et al., 2016a). CALR level cloud also be used as an independent prognostic factor in neuroblastoma (Hsu et al., 2005). Similarly, high level of CALR is associated with more infiltrated CD45RO+ cells in colon cancer, which correlates with a higher 5-year survival rate of patients (Peng et al., 2010). In acute myeloid leukaemia (AML) patients, CALR exposure on the membrane of malignant cells positively correlated with leukaemia specific immune response, further associated with both improved relapse-free survival and superior overall survival (Fucikova et al., 2016b). Moreover, low CALR level was associated with immunogenic chemotherapeutic resistance and poor survival rate in endometrial cancer patients (Xu et al., 2018).

16

Figure 4. Mechanism of CRT exposure. Several steps are involved in the exposure of CRT in response to immunogenic cell death inducers including activation of PERK, the phosphorylation of eIF2α, caspase- 8-dependent cleavage of BAP31, and the activation of Bax and Bak. Finally, a pool of CRT that has transited the Golgi apparatus is secreted by SNARE-dependent exocytosis. Abbreviations: CRT; calreticulin, PERK; protein kinase R-like ER kinase, BAP31; B cell receptor–associated protein 31, Bax; BCL2-associated X protein, Bak; Bcl-2 homologous antagonist/killer, SNARE; soluble N-ethylmaleimide- sensitive fusion protein-attachment protein receptor (Asadzadeh et al., 2020).

ER stress ER stress is considered to initiate CALR translocation in response to ICD induction. During homeostatic conditions, protein translation goes smoothly in ER, while the three unfolded protein

17

response (UPR) mediators are bind with heat shock protein 70 (Hsp70)-type chaperone protein Bip and stay in inactive state. In response to exogenous or endogenous ER stress, Bip preferably binds with misfolded proteins thus dissociates with these UPR mediators, leading to the activation of Inositol-requiring enzyme 1α (IRE1α), PKR-like ER-resident kinase (PERK) and activating transcription factor 6α (ATF6α). Upon Bip dissociation, 1) IRE1α forms a homodimer and undergo autophosphorylation, activating its C-terminal endoribonuclease activity. IRE1α splices the XBP- 1 mRNA and facilitates XBP-1 expression, which further promotes the transcription of protein chaperones, glycosylase, and ER-associated degradation (ERAD) system (Lee et al., 2003). Additionally, IRE1α could directly modulate the abundance of ER-localized mRNAs through regulated IRE1α-dependent decay (RIDD) pathway; 2) As an ER-resident eukaryotic translation initiation factor 2 (eIF2α) protein kinase, PERK phosphorylates the translation initiation factor eIF2α under ER stress conditions, abolishing 5′-cap assembly-dependent translation (Harding et al., 2000; Harding et al., 1999). Meanwhile, the general translation inhibition enhances the activating transcription factor 4 (ATF4) production, driving the expression of downstream transcription factor CHOP (Ma et al., 2002; Vattem and Wek, 2004); 3) Post dissociation with Bip, ATF6α translocates into Golgi and activated by proteolytic cleavage, promoting the accumulation of ERAD components and chaperones (Bailey and O'Hare, 2007; Haze et al., 1999; Shen et al., 2002). Although these three sensors could coordinate to combat ER stress, the individual UPR signalling pathways varies in different circumstances (Bettigole and Glimcher, 2015; Shoulders et al., 2013). Interestingly, ICD inducers like anthracyclines and oxaliplatin specifically induce PERK- dependent eIF2α phosphorylation, yet neither activating downstream ATF4 nor other two ER stress branches (Bezu et al., 2018a)(Figure 5). But the precise mechanism of ICD induced split ER stress response remains to be investigated. It’s demonstrated that PERK mediated eIF2α phosphorylation is mandatory for CALR exposure in response to ICD induction in several mouse and human cell lines (Bezu et al., 2018a; Panaretakis et al., 2009). However, it is protein kinase R (PKR) and general control nonderepressible 2(GCN2), rather than PERK, trigger eIF2α phosphorylation and CALR exposure in mitoxantrone (MTX) or doxorubicin (DOXO) treated melanoma cells (Giglio et al., 2018). Moreover, eIF2α phosphorylation and caspase-8 activation are dispensable in Hyp-PDT mediated CALR exposure, indicating that different ICD inducers may functions distinctively (Garg et al., 2012b). Of note, tumoral p-eIF2α and CALR expression

18

indicate positive prognostic value in NSCLC patients (Fucikova et al., 2016a). Altogether, p-eIF2α could be employed as a hallmark of ICD and promising biomarker in clinical prognosis prediction although exceptions exist.

Figure 5. eIF2α phosphorylation: A hallmark of immunogenic cell death. (A) The unfolded protein response of the endoplasmic reticulum in normal circumstances. Accumulation of misfolded proteins in the endoplasmic reticulum (ER) provokes the release of heat shock protein family A (Hsp70) member 5 (HSPA5, best known as binding immunoglobulin protein, BiP) and activates simultaneously the three arms of the unfolded protein response (UPR) ((i) EIF2AK3, (best known as PERK)-mediated phosphorylation of eukaryotic initiation factor α (eIF2α), which in turn halts general protein translation but favors the expression of activating transcription factor 4 (ATF4), (ii) the translocation of the ER transmembrane protein activating transcription factor 6 (ATF6) to the Golgi, proteolytic cleavage and the nuclear translocation of its cytosolic fraction (iii) the endoplasmic reticulum to nucleus signaling 1 (ERN, best known as IRE1α)-mediated splicing of X-box binding protein 1 (XBP1) to the XBP1s isoform. The signal transduction issued by the activation of the transcription factors ATF4, ATF6 and XBP1s aim at reestablishing ER homeostasis and cell survival (or the induction of cell death in conditions of enduring ER stress). (B) The unfolded protein response in the context of immunogenic cell death (ICD). First, the non- canonical activation of PERK and possibly other eIF2α kinases triggers the phosphorylation of eIF2α without BiP occupation by misfolded proteins and/or second, ICD-inducers inhibit UPR downstream signals such as ATF4 translation, ATF6 cleavage and the IRE1α−mediated XBP1 splicing (Bezu et al., 2018b).

19

ATP release As a multi-functional molecular expelled by dying cells in ICD event, the extracellular ATP plays an important role in immune response. Upon secretion, ATP binds to P2X7 and P2Y2 on APCs, promoting their maturation and facilitates chemotactic effect respectively (Elliott et al., 2009; Ghiringhelli et al., 2009; Michaud et al., 2011). In DCs, ATP triggers the aggregation of NLRP3-ASC–Casp-1 and facilitates IL-1β release, favouring the prime of antigen- specific cytotoxic T cell and IL-17 producing-γδ T cell. Thus, deficiency of any of these (Nlrp, P2rx7, Casp1) attenuates the establishment of adaptive immune response in chemotherapeutic ICD (Ghiringhelli et al., 2009; Ma et al., 2011). Besides, ATP could modulate DC activity simultaneously by stimulating the expression of costimulatory molecules include CD40, CD80, and CD86 (Idzko et al., 2007). The molecular mechanism of ATP secretion during ICD is often related to autophagy, a tightly regulated pathway mediates the recycle of cytosolic components in response to stress (Martins et al., 2014; Michaud et al., 2011). When treated with ICD inducer MTX or oxaliplatin, autophagy deficient cancer cells (by knock-down or knock-out of essential autophagy proteins or pharmacological inhibition) secrete much less ATP yet release similar amount of HMGB1 and expose equal level of CALR on cell surface . Immunization with MTX treated autophagy deficient CT26 or MCA205 cells failed to protect immunocompetent mice from re-challenging of the same cell type, while its immune stimulating capacity could be restored by co-injection of rIL-1β but not rCALR (Michaud et al., 2011). Moreover, autophagy-deficient tumour exhibits impaired tumour-infiltrating lymphocytes recruitment in response to immunogenic chemotherapy, but intra- tumoral administration of ecto-ATPase inhibitor regenerate therapeutic immune response in immunocompetent hosts (Chao et al., 2011). Accordingly, these results indicate that autophagy is required for ATP release during chemotherapy. Following research revealed a more complex machinery in regulating ATP secretion by dying cancer cell. Several relatively independent molecules include autophagy related proteins, , Rho-associated, coiled-coil containing protein kinase 1 (ROCK1), PANX1 channels and LAMP1 are involved in ICD-associated ATP release. In other words, autophagy, lysosomal exocytosis, apoptosis, membrane blebbing and plasma membrane permeabilization are orchestrated in ATP secretion (Martins et al., 2014). Apart from autophagy, the extracellular ATP level could also be regulated by CD39 and CD73 ectonucleotidase. CD39 and CD73 convert ATP to immunosuppressive adenosine, resulting in the

20

establishment of a tumour promoting microenvironment (Chalmin et al., 2012; Sun et al., 2010). Mice lacking CD73 shows resistance to MCA-induced fibrosarcoma and increased anti-tumour immunity, CD73 antibody treatment restrains prostate tumour growth and metastasis (Stagg et al., 2012)(Figure 6). Breast cancer patients bearing the loss-of-function (Glu496Ala) P2RX7 allele are prone to cancer metastasis post anthracyclines treatment, comparing to P2RX7 normal counterpart (Ghiringhelli et al., 2009). In a cohort of colorectal cancer (CRC) patients, absence of Beclin-1 expression is associated with poor prognosis (Choi et al., 2014). Similarly, another CRC study shows that the LC3 level is positively correlated with long-term survival (Wu et al., 2015).

Figure 6. Modulation of immune cell responses by extracellular ATP in the tumour microenvironment (Di Virgilio et al., 2018). 21

HMGB1 exodus HMGB1 is the first identified HMGB family protein which locates in the nuclear as non- histone chromatin-binding protein in steady state (Palumbo et al., 2004). The function of HMGB1 mainly depends on its sublocation, namely nuclear, cytoplasm or extracellular milieu (Garg et al., 2010). Within nuclear, HMGB1 binds to chromatin to stabilize the structure and modulates transcription (Garg et al., 2010; Scaffidi et al., 2002). In cytoplasm, HMGB1 can interacts with Beclin-1 to affect autophagy (Huang et al., 2012; Tang et al., 2010). It’s reported that HMGB1 regulates the expression of various genes including p53 and nuclear factor-κB (NF-κB) (Krysko et al., 2012). Extracellular HMGB1 could directly interact with different surface receptors like RAGE (receptor for advanced glycation end-products), TLR2 and TLR4, to induce significant immune response in a cytokine-like manner (Sims et al., 2010). HMGB1 stimulates the production of a series of proinflammatory cytokines in human monocytes, but not lymphocytes. Accordingly, intraperitoneal administration of rHMGB1 increase TNF concentration in mice serum (Zhang et al., 2016). In the context of ICD, TLR4 seems to be the main HMGB1 receptor which contributes to tumour eradiation. HMGB1 inhibition and TLR4 deletion compromise the protective efficiency of dying tumour in vaccination experiment. Besides, the immune response to dying tumour cell is largely depends on TLR4 expression on DC. Moreover, both myeloid differentiation primary response 88 (MYD88) −/− mice and TLR4−/− mice failed to response to chemotherapy or radiotherapy, demonstrating the contribution of HMGB1/TLR4/MyD88 axis in anticancer immunity (Apetoh et al., 2007). It should be noticed that the function of HMGB1 is determined by its redox modification, HMGB1 acts as chemoattractant DAMP, pro-inflammatory cytokine- inducing DAMP or inactivated DAMP depending on actual reduced to oxidized forms (Venereau et al., 2012; Yang et al., 2012)(Figure 7). HMGB1 can be released in two different fashions, named active secretion or passive liberation. and monocytes secret HMGB1 through a cytokine secreting manner post IL-1β, TNF or lipopolysaccharides (LPS) stimulation (Garg et al., 2010; Muller et al., 2004). Primary necrosis cells could passively release large amount of HMGB1 to extracellular space (Scaffidi et al., 2002). Also, Cells undergo secondary necrosis discharge HMGB1. In particular, Z-VAD-fmk (a broad caspase inhibitor) treatment reduces the release of HMGB1 from ICD (Apetoh et al., 2007; Bell et al., 2006).

22

A study showed both TLR4 Asp299Gly and Thr399Ile polymorphisms were positively associated with cancer development and progression in CRC patients, and the expression of these mutations impact patients’ survival (Messaritakis et al., 2018). Also, Thr399Ile polymorphism was significantly associated with an elevated gastrointestinal cancer risk (Jing et al., 2012). Breast cancer patients bearing mutated TLR4 alleles relapse faster post anthracycline-based chemotherapy and local irradiation (Apetoh et al., 2007). However, controversial results showed no correlations between CRC risk and Asp299Gly or Thr399Ile alleles (Davoodi and Seow, 2011).

23

Figure 7. HMGB1 release and extracellular function. HMGB1 can be actively secreted by immune inflammatory cells or passively released by dead, dying or injured cells into the extracellular milieu by several different mechanisms as indicated. Extracellular HMGB1 acts as a DAMP molecule and plays a vital role in several pathophysiological processes. The activity of extracellular HMGB1 depends on redox state, receptors and their interactive partners (Chen et al., 2013).

24

Methods to detect ICD

As aforementioned, ICD inducers cause cancer cell death in parallel with the emission of DAMPs, which includes split ER stress induced CALR translocation, autophagy dependent ATP release, HMGB1 exodus and Type I IFN secretion. These DAMPs define its adjuvanticity and triggers adaptive immune response in company with antigenicity. One thing should be emphasized here is the precise mechanisms of ICD are not fully addressed, the immunogenicity of agents should be tested both in vitro and in vivo. A drug discovery pipeline was constructed to identify ICD inducers from large libraries (Bezu et al., 2018a; Kepp et al., 2019)(Figure 8). This pipeline consists four procedures, 1) use artificial intelligence to pre-select drug candidates; 2) high content screening to find hits; 3) in vitro validation and 4) in vivo validation.

Figure 8. Discovery pipeline. (Kepp et al., 2019)

In vivo vaccination assay is employed as the ‘Gold Standard’ to evaluate immunogenicity. In brief, tumour cells were treated with cytotoxic ICD inducer candidates in vitro, then immunocompetent naïve mice were injected with these dying tumour cells without adjuvant in the flank subcutaneously, after one to two weeks these vaccinated mice were rechallenged by the syngeneic living tumour cell in the contralateral flank, tumour emergence and growth were monitored. The ‘real’ ICD inducer treated cells could protect mice from growing tumours, or at least delay tumour growth and extends lifespan compare to ‘PBS’ group and ‘non-ICD’ group (Humeau et al., 2019)(Figure 9). Given this in vivo vaccination experiment is only applicable for

25

mouse cancer cell line, the generation of humanized mice would be particularly useful in predicting ICD in human cancers.

Figure 9. Assessing the ability of a cytotoxicant to stimulate cancer ICD in vivo. (Humeau et al., 2019)

26

ICD inducers

Doxorubicin was the first identified ICD inducer in 2005, as tumour cells treated by DOXO could effectively elicit anticancer immunity in the absence of adjuvant (Casares et al., 2005). Since then, numerous ICD inducers were discovered by different research groups. ICD inducers could be categorized into four categories: chemotherapy, physical approach, immunotherapy and target therapy. As a well-studied ICD inducer cluster, chemotherapeutic drugs such as anthracyclines, platinum derivatives, alkylating agents and proteasome inhibitors are proved to trigger ICD potently (Casares et al., 2005; Panaretakis et al., 2009; Schiavoni et al., 2011; Spisek et al., 2007). Radio therapy and Photodynamic therapy are also effective approaches to induce ICD (Demaria et al., 2004; Formenti et al., 2018; Korbelik and Dougherty, 1999; Korbelik et al., 2011; Krosl et al., 1995; Paz-Ares et al., 2020). Some oncolytic virus and oncolytic peptides induce traits of ICD (Koks et al., 2015; Yamazaki et al., 2016; Zamarin et al., 2014; Zhou et al., 2016a; Zhou et al., 2018). Only specific target therapy drugs like Crizotinib and Cetuximab are capable to generate ICD (Liu et al., 2019; Pozzi et al., 2016). Confirmed ICD inducers are listed in Table 2 (Galluzzi et al., 2020a).

27

Agent P-eIF2α DAMPsMain cytokines Immune infiltrateAntagonized by Co mbination with ICIsa

7A7Yes CALR ↑IFNγ↑DCs, CD8 + CTLs, CD4+ T cells CD8+ T cell depletion ND

↑Type I IFNs, IL-6, ↑DCs, monocytes,CD8+ CTLs, NK CD8+, CD4+ or NK cell 8-MOP PCT ND CALR, ATP, HMGB1 ND CCL2 cells depletion nu/nu genotype, ↑IL-1β, IL-6, IL-12, IL- CD8+ T cell depletion, CALR, ATP, HMGB1, ↑DCs, CD8+ CTLs, γδ TH17 cells; Anthracyclines Yes 17, CXCL10, type I IFNγ or IL-17 Anti-PD-1, anti-CTLA4 ANXA1 ↓Treg cells, MDSCs IFNs, IFNγ neutralization, IL-17 receptor inhibition CD8+ T cell depletion, BleomycinYes CALR, ATP, HMGB1 ↑IFNγ, TGFβ ↑CD8+ CTLs;↑Treg cells ND IFNγ neutralization nu/nu or Rag -/- Bortezomib Yes CALR, HSP70 ↑IFNγ↑DCs, CD8 + CTLs genotype, CD8+ T cell ND depletion ↑CD8+ CTLs, CD4+ T cells, γδ TH17 Cardiac glycosides ND CALR, ATP, HMGB1 ↑IFNγ nu/nu genotype ND cells

Cetuximab ND CALR, HMGB1 ND ↑DCs, CD8+ CTLs NDND

Clostridium difficile No CALR, HMGB1, ATP NDNDNDND toxin B ↑DCs, CD8+ CTLs, NKT cells; CrizotinibYes CALR, ATP, HMGB1 ↑Type I IFNs, IL-17 nu/nu genotype Anti-PD-1, anti-PD-L1 ↓Treg cells ↑Type I IFNs, IFNγ, IL- ↑DCs, CD8+ CTLs, NK cellsb, NKT Cyclophosphamide ND CALR, HMGB1, ATP ND Anti-PD-1, anti-CTLA4 17 cells; ↓MDSCs, Treg cells nu/nu genotype, ↑Type I IFNs, IFNγ, IL- Dactinomycin Yes CALR, ATP, HMGB1 ↑DCs, CD8+ CTLs, NK cells; ↓Treg cells CD8+ and CD4+ T cell Anti-PD-1 17 depletion

28

Dinaciclib ND CALR, HMGB1, ATP ↑IFNγ↑DCs, CD8 + CTLs, CD4+ T cells Rag-/- genotype Anti-PD-1

CD8+ and CD4+ T cell DTT-205 ND CALR, HMGB1 ↑Type I IFNs ND ND depletion CD8+ and CD4+ T cell DTT-304 ND CALR, HMGB1 ↑Type I IFNs ND ND depletion

Electrical nanopulses ND CALR, ATP, HMGB1 ND ↑CD8+ CTLs NDND

Hypericin-based PDT No CALR, ATP, HMGB1 ↑IL-1β↑DCs NDND

JQ1Yes CALR, ATP, HMGB1 ND ↑DCs, CD8+ CTLs; ↓MDSCs nu/nu genotype ND

↑IL-1β, IL-6, type I ↑DCs, macrophages, CD8+ CTLs, TH1 LTX-315 No CALR, ATP, HMGB1 ND Anti-PD-1, anti-CTLA4 IFNs CD4+ cells; ↓MDSCs, Treg cells

LTX-401 No CALR, ATP, HMGB1 ↑Type I IFNs ↑ CD3+ T cells ND Anti-PD-1, anti-CTLA4

CD8+ and CD4+ T cell Lurbinectedin Yes CALR, ATP, HMGB1 ↑Type I IFNs ND Anti-PD-1, anti-CTLA4 depletion Microwave thermal ND CALR, ATP, HMGB1 ↑IFNγ, TNF ↑CD8+ CTLsCD8 + T cell depletion ND ablation Newcastle disease ↑DCs, ↑CD8+ CTLs, CD4+ T cells, NK Rag2-/- genotype, ND CALR, HMGB1 ↑IFNγ Anti-CTLA4 virus and NKT cells;↓MDSCs CD8+ T cell depletion ↑ CXCL10, IFNγ, TNF; Non-thermal plasma ND CALR, ATP, HMGB1 ↑DCs ND Anti-PD-L1 ↓TGFβ nu/nu genotype, CD8+ T cell depletion, CALR, ATP, HMGB1, ↑IL-1β, type I IFNs, Oxaliplatin Yes ↑DCs, CD8+ CTLs, γδ TH17 cells IFNγ or IL-17 Anti-PD-1, anti-CTLA4 HSP70 IFNγ neutralization, IL-17R blockage

29

↑Monocytes, neutrophils, CD8+ CTLs, Photofrin-based PDT ND CALR, HMGB1 ND CD8+ T cell depletion ND NK cells ↑DCsb, CD8+ CTLs; ↓macrophageb, Anti-PD-1, anti-PD-L1, PT-112 ND CALR, ATP, HMGB1 ND ND Treg cellsb anti-CTLA4 nu/nu genotype, ↑IL-17, type I IFNs, Anti-PD-1, anti-PD-L1, Radiotherapy ND CALR, ATP HMGB1 ↑DCs, CD8+ CTLs CD8+ T cell depletion, IFNγ, TGFβ anti-CTLA4 IFNAR1 blockage

Redaporfin-based PDT YesCALR, ATP, HMGB1 NDNDNDND

Septacidin ND CALR, ATP, HMGB1 NDND nu/nu genotype ND

Talimogene ↑IL-1β, type I IFNs, ND CALR, ATP, HMGB1 ↑CD8+ CTLs ND Anti-PD-1 laherparepvec TNF

Teniposide ND CALR, HMGB1 ↑IL-2, type I IFNs, IFNγ ↑DCs, CD8+ CTLs ND Anti-PD-1

UVC light ND CALRb, HMGB1b ↑IFNγ↑DCs, CD8 + CTLs NDND

CALR, ATP, HMGB1, WogoninYes ND ↑DCs, CD3+ T cells NDND ANXA1

Table 2. Confirmed inducers of immunogenic cell death. All agents confer enduring immunological memory, excluding dactinomycin, electrical nanopulses, non-thermal plasma and redaporfin-based photodynamic therapy (PDT), for which evidence suggests only partial immunological memory. Cardiac glycosides and crizotinib confer immunological memory only in combination with cytotoxic agents. 8-MOP, 8- methoxypsolaren; ANXA1, annexin A1; CALR, calreticulin; CTLA4, cytotoxic T lymphocyte-associated protein 4; CTL, cytotoxic T lymphocyte; DAMP, damage-associated molecular pattern; DC, dendritic cell; HMGB1, high mobility group box 1; HSP70, heat shock protein family A; ICI, immune checkpoint inhibitor; IFN, interferon; MDSC, myeloid-derived suppressor cell; ND, not determined; NK, natural killer; NKT, natural killer

a b T; P-eIF2α, eIF2α phosphorylation; PCT, photochemotherapy; TH, T helper; Treg, regulatory T; UVC, ultraviolet C. In mouse models. Observed in some but not all models investigated. Adapted from (Galluzzi et al., 2020a).

30

Induction of ICD sensitize cancers to immune checkpoint blockade

Immunotherapy, particularly immune checkpoint blockade (ICB), is now largely applied to treat various cancers. But due to the immunosuppressive nature of tumour microenvironment, the objective response rates (ORRs) of standard ICB treatment in solid tumours are relatively low (Lim et al., 2016; Rosenberg et al., 2016; Topalian et al., 2012). ICD induction in tumour sites enables to recruit antigen presenting cells and CD8 T cells to the tumour bed, thus convert immunologically ‘cold’ tumours into ‘hot’ lesions. Given that ICB’s therapeutic function dedicated to the presence of CD8 T cells, pre-treatment with ICD inducers sensitize tumours to ICB in several studies (Liu et al., 2019; Pfirschke et al., 2016; Yamazaki et al., 2020). In a cohort of NSCLC patients, chemoradiotherapy plus programmed death-ligand 1 (PD-L1) blockade dramatically prolonged the overall survival of combination treatment group against placebo group (Antonia et al., 2018). In another study involves metastatic triple-negative breast cancer patients, doxorubicin treatment upregulates the expression of key genes which related to programmed cell death protein 1 (PD-1)/PD-L1 and T cell cytotoxicity pathways, thus evaluated the objective response rate to PD-1 blockade (Voorwerk et al., 2019). In addition, bunch of clinical trials that involves the combination of immunogenic chemotherapy and ICBs are still ongoing (Galluzzi et al., 2020a)(Figure 10). Altogether, both preclinical and clinical studies reveal that ICD-inducing chemotherapy may bring benefits to cancer patients in some immunotherapy included settings. But the optimal therapeutic combination regiments remain to be critically evaluated.

31

Figure 10. Overview of ongoing clinical trials involving one or more immunogenic cell death inducers in combination with immune checkpoint inhibitors in patients with cancer. As of 1 March, 2020, ClinicalTrials.gov listed 55 ongoing (defined as those that are not completed, withdrawn, suspended, terminated or have unknown status) clinical trials in which patients are receiving chemotherapy plus agents with known immunogenic cell death (ICD)-inducing potential in combination with one or more FDA-approved immune checkpoint inhibitors in patients with cancer that were initiated after 1 January, 2019. CTLA4, cytotoxic T lymphocyte-associated protein (Galluzzi et al., 2020a).

32

Transcription inhibitor Lurbinectedin

Lurbinectedin (PM01183) is a novel synthetic marine-derived anticancer agent developed by PharmaMar Company. It was recently approved by Federal Drug Administration (FDA) for the salvage treatment of small-cell lung cancer (SCLC). As an alkaloid, lurbinectedin could binds to the DNA minor groove, induces double-strand DNA breaks, selectively blocks RNA polymerase II activity and promotes its degradation, disrupts mitosis and eventually induces cell death (Leal et al., 2010; Santamaria Nunez et al., 2016; Vidal et al., 2012). As an analogue of , the structure of lurbinectedin is similar to trabectedin except the modification in C subunit (Leal et al., 2010). Thus, lurbinectedin gains different pharmacokinetic and pharmacodynamics properties from trabectedin (Bueren-Calabuig et al., 2011; Leal et al., 2010; Soares et al., 2011). It’s reported that lurbinectedin has potent antiproliferative activity in numerous cancer cell lines in vitro, including cisplatin-sensitive and resistant cells (Leal et al., 2010; Santamaria Nunez et al., 2016; Soares et al., 2011). Single lurbinectedin treatment exhibits antitumor/antimetastatic activity in MNMCA1, M5076 murine reticulosarcoma, HOC8 and MNBPTX human ovarian carcinoma xenograft in vivo (Romano et al., 2013). In preclinical uterine cervical cancer models, lurbinectedin inhibits tumour growth by eliminating both cancer stem cell and non-cancer stem cell (Yokoi et al., 2019). Lurbinectedin inactivates oncoprotein EWS-FLI1 and synergized with irinotecan to retard Ewing sarcoma xenograft growth (Harlow et al., 2016). Also, lurbinectedin therapy overcomes cisplatin resistance in in epithelial ovarian cancer (Vidal et al., 2012). In particular, lurbinectedin depletes tumour-associated macrophages and modulates inflammatory microenvironment in both mouse fibrosarcoma tumour model and patient-derived (AVATAR) PDA xenografts (Belgiovine et al., 2017; Cespedes et al., 2016; Farago et al., 2019)(Figure 11).

33

Figure11. Chemical structure and schematized actions of lurbinectedin. (A) Chemical structure of lurbinectedin. (B) Schematic representation of lurbinectedin mechanism of action for binding DNA, inhibiting active transcription and inducing DNA breaks. (C) Schematic representation of lurbinectedin interactions with the tumor microenvironment. Lurb: Lurbinectedin (Farago et al., 2019).

The approval of lurbinectedin for second line SCLC treatment was based on several clinical trials. In a Phase I study enrolled 27 relapsed SCLC patients, the combination regiment of lurbinectedin and doxorubicin showed remarkable anticancer activity. As second-line treatment, 11 out of 12 patients with platinum-sensitive disease and 3 out of 9 patients with platinum-resistant disease were responded to the combination treatment (Calvo et al., 2017). In a single-arm, open- label phase II basket trial, 105 patients with relapsed SCLC were treated with lurbinectedin alone. The median overall survival (OS) was 9.3 months, 11.9 months, 5.0 months in the whole cohort, patients with chemotherapy sensitive or resistant disease respectively. Among all patients the

34

overall response rate (ORR) was 35.2%, whereas 45.0% in chemotherapy sensitive subgroup and 22.2% in chemotherapy resistant subgroup. Further, lurbinectedin was of good safety, no patient dead due to treatment related advise effects (Trigo et al., 2020). It’s notable that the combination therapy of lurbinectedin and atezolizumab (anti-PD-L1 antibody) in patients with progressed SCLC is under evaluation (ClinicalTrials.gov, NCT04253145). Apart from SCLC, lurbinectedin is in therapeutic trials for multiple cancer types, including breast cancer, ovarian cancer, sarcoma, acute leukaemia and different advanced solid tumours (Markham, 2020)(Table 3).

Table 3. Key clinical trials of lurbinectedin. mCRC metastatic colorectal cancer, NSCLC non-small cell lung cancer, SCLC small cell lung cancer (Markham, 2020).

35

Oncolytic peptide LTX-401

Antimicrobial peptides (AMPs) are a series of short cationic amphipathic peptides that protect hosts from various (Fjell et al., 2011; Hancock and Sahl, 2006). More than 2000 kinds of natural AMPs were discovered currently, but no conserved motif or sequence has been identified yet (Wang et al., 2016). Apart from anti- activity, some AMPs exert anticancer property by directly lysing cancer cells (Al-Benna et al., 2011; Gaspar et al., 2013). Bovine lactoferricin (LfcinB) is the best studied natural AMP which is derived from milk protein. It’s reported that LfcinB could induce cancer cell necrosis and apoptosis by destabilise membranes (Eliassen et al., 2006; Eliassen et al., 2002; Mader et al., 2005; Yoo et al., 1997). Based on the structure of LfcinB, a large number of oncolytic peptides were designed and synthetized to treat cancer. One of these oncolytic peptides is LTX-401, an amphiphatic β (2,2)-amino acid derivate that is able to lysis different type of cells (Eike et al., 2016) (Figure 12).

Figure 12. Chemical structure of the small amphipathic β(2,2)-amino acid-derived antitumor molecule LTX-401 (MWnet = 367.53) (Eike et al., 2016).

Previous data demonstrated that LTX-401 preferentially enriches in Golgi apparatus and destroys its structure, subsequently inducing cell death. The presence of brefeldin A (BFA) reduces the cytotoxicity of LTX-401 not only due to the prophylactic dissociation of Golgi, but also through decreasing mitochondrial permeabilization. BAX/BAK deficient or mitochondrial depleted cells show partial resistance to LTX-401 mediated killing, while pan-caspase inhibitor Z- VAD-fmk doesn’t exhibit any protective effect (Zhou et al., 2016b). In addition, LTX-401 treatment impairs lysosomal integrity and induces the release of reactive oxygen species (ROS).

36

Massive DAMPs are released following LTX-401 treatment, including ATP, Type I IFN, HMGB1, cytochrome c and CALR exposure. Intra-tumour injection of LTX-401 causes focal necrosis and T lymphocytes infiltration (Eike et al., 2016; Mauseth et al., 2019; Zhou et al., 2016b). The antitumor capacity of LTX-401 were verified in several cancer models. Local administration of LTX-401 in 3 consecutive days cured more than 80% of B16F1 melanoma (Eike et al., 2016). In subcutaneous JM1 tumour model, 7 out of 10 tumours were completed regressed post LTX-401 treatment. All cured animals were resistant to both subcutaneous and intrahepatic rechallenge of the same tumour cell. More convincing date was obtained from rat orthotopic JM1 liver cancer model, a preclinical model of human hepatocellular carcinoma (HCC). 5 out of 9 mice were cured from orthotopic JM1 liver tumour, by intrahepatic injection of LTX-401 on day 6 and day 8 post tumour incubation. Again, cured mice generated long-term antitumor immunity (Mauseth et al., 2019).

37

AIMS OF THE THESIS

Cancer cells undergo immunogenic cell death effectively provoke anticancer immune response in immunocompetent host. Therefore, ICD inducers could somehow convert dying tumour cells to therapeutic ‘vaccines’, attracting leukocytes to tumour beds and thus switching from the immunosuppressive tumour microenvironment to a proinflammatory context. In addition, both preclinical study and clinical trials support the hypothesis that pre-treatment with a panel of ICD inducers sensitize tumours to immune checkpoint blockade. Based on this rationale, it’s of great value to evaluate the capacity of drugs/compounds in inducing ICD and potential benefits in combination with ICBs. Lurbinectedin is a RNA polymerase II inhibitor which was newly approved by FDA for cancer treatment. The majority of previous studies investigated its anticancer ability in immunodeficient host, thus the immunogenicity of lurbinectedin induced cell death is still unknown. As an oncolytic peptide, LTX-401 is designed for local administration. Past research shows that LTX-401 could selectively concentrate in and destroy Golgi apparatus, inducing all the hallmarks of ICD in vitro. But the capacity of LTX-401 to stimulate anticancer immune responses remains to be elucidated.

The specific aims of this thesis are:

To evaluate the capacity of lurbinectedin in stimulating the hallmarks of ICD.

To investigate the anticancer immune responses of lurbinectedin and LTX-401 in different experimental cancer models.

To explore the combination therapeutic regiments in cancer treatment.

38

RESULTS

Lurbinectedin synergizes with immune checkpoint blockade to generate anticancer immunity

Accumulating results indicate that chemotherapeutic drugs which efficiently inhibit transcription and promote eIF2α phosphorylation are potent ICD inducers, indicating the potential of lurbinectedin in inducing ICD. Both preclinical data and clinical trials demonstrate that pre- treatment with certain ICD inducers sensitizes to immune checkpoint blockade, it’s worthwhile to investigate the role of lurbinectedin in the era of immunotherapy. To evaluate the immunogenicity of lurbinectedin, different parameters of ICD hallmarks were measured by robotized epifluorescence microscopy followed by automated image analysis. Lurbinectedin triggered CALR exposure, ATP release, HMGB1 exodus, Type I IFN release and cell death in a dose- and time-dependent manner in 4 human or mouse cancer cell lines. Also, lurbinectedin activated two ICD pathways named split ER stress and transcription inhibition, featured by elevated eIF2α phosphorylation level and decreased co-localization of nucleolin and fibrillarin. Thus, Lurbinectedin induced all traits of ICD in vitro. To further confirm its immunogenicity in vivo, we employed the vaccination experiment in immunocompetent mice. Lurbinectedin-treated cells significantly reduced tumour growth and led to increased overall survival. Next, we investigated the therapeutic efficacy of lurbinectedin in immunocompetent and immunodeficient mice. Lurbinectedin significantly retarded MCA205 tumour growth and extended survival span in immunocompetent host but failed to do so in immunodeficient counterpart. Implying that the anticancer efficiency of lurbinectedin in MCA205 cancer depends on immune system. Moreover, lurbinectedin, αCTLA-4 and αPD-1 triple combination regiment dramatically inhibited tumour progression and induced complete regression in 3 out of 8 animals. Cured mice exhibited systemic antitumor immunity to the same type of cancer cell. In particular, T lymphocyte depletion abolished the effect of lurbinectedin with αCTLA-4/αPD-1 dual checkpoint blockade.

39

Finally, we tested the therapeutic effect of lurbinectedin in MPA/DMBA-induced spontaneous breast cancer model. Both lurbinectedin alone and in combination with double ICBs significantly reduced tumour growth and increased overall survival.

40

ONCOIMMUNOLOGY 2019, VOL. 8, NO. 11, e1656502 (9 pages) https://doi.org/10.1080/2162402X.2019.1656502

ORIGINAL RESEARCH Lurbinectedin synergizes with immune checkpoint blockade to generate anticancer immunity Wei Xiea,b,c,d*, Sabrina Forveillea,b,c,d*, Kristina Iribarren a,b,c,d*, Allan Sauvat a,b,c,d, Laura Senovillaa,b,c,d, Yan Wanga,b,c,d, Juliette Humeaua,b,c,d, Maria Perez-Lanzon a,b,c,d, Heng Zhoua,b,c,d†, Juan F. Martínez-Leale, Guido Kroemer a,b,c,d,f,g,h, and Oliver Kepp a,b,c,d aMetabolomics and Cell Biology Platforms, Gustave Roussy Comprehensive Cancer Institute, Villejuif, France; bEquipe 11 labellisée Ligue contre le Cancer, Centre de Recherche des Cordeliers, Paris, France; cUniversité Paris Descartes, Sorbonne Paris Cité, Paris, France; dUniversité Pierre et Marie Curie, Paris, France; ePharmaMar, Madrid, Spain; fSuzhou Institute for Systems Medicine, Chinese Academy of Medical Sciences, Suzhou, China; gDepartment of Women’s and Children’s Health, Karolinska Institutet, Stockholm, Sweden; hPôle de Biologie, Hôpital Européen Georges Pompidou, AP-HP, Paris, France

ABSTRACT ARTICLE HISTORY Systemic treatment with the active transcription inhibitor lurbinectedin aims at inducing tumor cell death in Received 20 May 2019 hyperproliferative . Here we show that cell death induced by lurbinectedin reinstates and Revised 29 July 2019 enhances systemic anticancer immune responses. Lurbinectedin treatment showed traits of immunogenic Accepted 13 August 2019 cell death, including the exposure of calreticulin, the release of ATP, the exodus of high mobility group box 1 KEYWORDS (HMGB1) and type 1 interferon responses in vitro. Lurbinectedin treated cells induced antitumor immunity Anticancer immunity; when injected into immunocompetent animals and treatment of transplanted fibrosarcomas reduced tumor immunogenic cell death; growth in immunocompetent yet not in immunodeficient hosts. Anticancer effects resulting from lurbinec- checkpoint blockade; tumor tedin treatment were boosted in combination with PD-1 and CTLA-4 double immune checkpoint blockade clearance (ICB), and lurbinectedin combined with double ICB exhibited strong antineoplastic effects. Cured animals exhibited long term immune memory effects that rendered them resistant to rechallenge with syngeneic tumors underlining the potency of combination therapy with lurbinectedin.

Introduction stressed and dying cells,10,11 the cell death-associated exodus of nuclear high mobility group box 1 (HMGB1)12,13 and the Primary or transplantable tumors react to anthracycline-based stimulation of an autocrine or paracrine type-1 interferon chemotherapy with durable response in syngeneic immunocom- 14 – response. CALR serves as a de novo uptake signal and petent mice yet fail to do so in immunodeficient hosts.1 3 stimulates the engulfment of dying cancer cells by dendritic Consistently, retrospective clinical studies in patients with solid cells (DCs).3 HMGB1 binds to toll-like receptor-4 (TLR4) tumors subjected to chemotherapy showed that severe lympho- entities on DC, eliciting MYD88-dependent signaling that penia negatively affects prognosis,4,5 which points to the fact that facilitates tumor antigen processing.3,15 ATP ligates purinergic chemotherapy-elicited anticancer immunity plays a critical role receptors of the P2X type and thus activates the NLRP3 for the outcome of anticancer therapy.6,7 Based on these – inflammasome to stimulate the production of interleukin-1β findings,1 3 we introduced the hypothesis that some chemother- (IL-1β) by DC and eventually interferon-γ (IFNγ) by CD8+ apeutic agents can induce immunogenic cell death (ICD) in cytotoxic T lymphocytes (CTL).10,16 tumors and convert them into a therapeutic vaccine, hence The sum of danger associated molecular patterns (DAMP) stimulating an immune response that can control residual cancer emitted during ICD is necessary to generate anticancer immu- cells. − − − − nity, thus tumors growing in Tlr4 /-, P2rx7 /-, Myd88 /-, Nlrp3 /- Selected chemotherapeutics such as anthracyclines and −/- −/- −/- −/- + – , Il1r , Ifnγ , Ifnγr , Fpr1 , athymic or CD8 Tcell-depleted oxaliplatin are able to induce ICD1 3 while many other anti- mice fail to respond to immunogenic chemotherapeutic regi- neoplastic agents including cisplatin and mitomycin C fail to mens. Loss-of-function mutations of FPR1, P2RX7 or TLR4 in do so. Cancer cells undergoing ICD can evoke anticancer breast cancer are negatively correlated with clinical response to immunity and protect against a subsequent challenge with 3,10,13,14,17–19 – adjuvant chemotherapy with anthracyclines. These living cells exhibiting the same antigenic profile in mice1 3 results imply the obligate contribution of anticancer immune or elicit anticancer immune responses during chemotherapy responses to the success of ICD-inducing . in patients.8 Distinctive properties of immunogenic cell death Lurbinectedin is a selective inhibitor of active transcription of include the exposure of calreticulin (CALR) at the cytoplasmic protein-coding genes20 that is currently undergoing clinical surface,3,8,9 the autophagy-dependent liberation of ATP from investigation and has recently gained orphan drug approval for

Guido Kroemer [email protected]; CONTACT Oliver Kepp [email protected] Metabolomics and Cell Biology Platforms, Gustave Roussy Comprehensive Cancer Institute, Villejuif, France*contributed equally †current address: Institute of Modern Physics, Chinese Academy of Sciences, Lanzhou, China © 2019 Taylor & Francis Group, LLC e1656502-2 W. XIE ET AL. the treatment of small cell lung cancer (SCLC). Here, we inves- the biosensor cells. Following the type 1 IFN response was tigated the capacity of lurbinectedin to stimulate the emission of monitored by means an increase in GFP fluorescence inten- immunogenic DAMPs and tested anticancer immune responses sity. A significant increase in de novo GFP signal intensity was in three experimental in vivo models. Our results support the detected for both lurbinectedin and MTX throughout all time contention that lurbinectedin causes immunogenic cell death in points (Figure 1(a)). Similar results were obtained when the tumors and creates anticancer immunity. approach was repeated in human breast cancer HCC70 cells (Figure 1(b)), human colon carcinoma HT29 (Figure 1(c)) and mouse fibrosarcoma MCA205 cells (Figure 1(d)). Next, Results and discussion we investigated the capacity of lurbinectedin to activate two additional characteristics of common ICD inducers, the phos- Emission of immunogenic signals by lurbinectedin phorylation of the eukaryotic translation initiation factor 2 The known parameters determining ICD are the translocation alpha (eIF2α) and the inhibition of general transcription. of CALR to the surface of the plasma membrane, the autop- Indeed, lurbinectedin led to a dose-dependent phosphoryla- hagy-dependent liberation of ATP and the release of the non- tion of eIF2α monitored by fluorescence microscopy upon histone binding protein HMGB1, which occur before, during immunostaining with a phosphoneoepitope-specific antibody and after apoptosis, respectively. The production of type (Figure 2(a,b)). Lurbinectedin also inhibited mRNA transcrip- I interferons (IFNs) has been added to the list of ICD hall- tion at a level comparable to a known transcription-inhibitor, marks as it controls autocrine or paracrine circuits that as assessed by visualizing the dissociation of nucleolin and underlie cancer immunosurveillance. fibrillarin by microscopy (Figure 2(b,c)), an accepted proxy of In a systematic screening campaign, the capacity of lurbi- suppressed transcription.21 Lurbinectedin holds many of the nectedin to induce immunogenic cell death in cancer cells was described in vitro parameters of ICD, thus qualifying for assessed in human osteosarcoma U2OS cells stably expressing further in vivo investigations in immunocompetent animals, fluorescent biosensors for the detection of CALR-relocation which remains the gold standard assay for the determination (as a surrogate marker for CALR surface exposure), HMGB1 of ICD-mediated anticancer immunity. release and Type I IFN responses together with U2OS WT cells stained with the ATP-sensitive dye quinacrine. ICD- related parameters were measured at 4, 8, 16 and 32 hours Anticancer immunity induced by lurbinectedin post exposure to lurbinectedin from 1 nM to 1 µM by robot- ized epifluorescence microscopy followed by automated image In order to assess the capacity of lurbinectedin to stimulate analysis (Figure 1). The induction of cell death was evaluated anticancer immunity in a monotherapeutic approach and to based on changes in the nuclear morphology visualized by convert tumor cells into a therapeutic vaccine we exposed means of the DNA intercalating dye Hoechst 33342. murine fibrosarcoma cells to the drug in vitro (in conditions Lurbinectedin caused a dose- and time-dependent drop in previously established to induce a sufficient amplitude of cell cellular viability comparable to mitoxantrone (MTX) that death) and then injected the dying cancer cells into syngeneic was used at 1 and 3 µM as a positive control throughout all immunocompetent mice. One week later, the animals were re- experiments. The translocation of a CALR-GFP (green fluor- challenged injecting live tumor cells of the same kind into the escent protein) fusion protein from the perinuclear ER to the opposite flank, (Figure 3(a)). In this setting, a decrease of cellular periphery was measured by assessing cytoplasmic tumor growth can be interpreted as sign of a productive antic- “granularity” (see Materials and Methods) as an indicator ancer immune response. Indeed, lurbinectedin-treated cells for the formation of CALR-containing vesicles and as significantly reduced tumor growth (p = .0094) (Figure 3(b)) a surrogate marker for CALR exposure. Lurbinectedin, similar and led to an increase in overall survival (Figure 3(c)). As – to MTX, induced a time- and dose-dependent increase in compared to know ICD inducers1 3 the vaccination effects CALR-granularity as compared to untreated controls. The observed here were rather limited yet statistically significant. reduction of intracellular ATP (as an indicator for ATP Next we evaluated the effect of lurbinectedin on established release) was assessed by measuring the decrease in the cyto- cancers growing on immunocompetent or immunodeficient plasmic granularity of ATP containing vesicles stained with mice. MCA 205 tumors were implanted subcutaneously on the fluorescent probe quinacrine. As compared to untreated immunocompetent C57BL/6 as well as in athymic nu/nu controls a significant decrease in ATP signal similar to MTX mice. When the tumors became palpable, the animals were was detectable for lurbinectedin. The effect was dose- treated with three consecutive intravenous injections of dependent and decreased over time in line with the fragile 0.18 mg/kg lurbinectedin on day 1, 7 and 14. (Figure 4(a)). nature of the metabolite. HMGB1 release was detected as The treatment with lurbinectedin had significant therapeutic a loss in the nuclear fluorescence of an HMGB1-GFP chimera. benefit in immunocompetent animals. The tumor growth was A significant decrease in nuclear GFP signal was detected for significantly reduced as compared to control animals MTX and lurbinectedin at medium to late time points. Type (p < .0001) (Figure 4(b)) and overall survival was increased I interferon (IFN) production was measured using U2OS (Figure 4(c)). This effect was exclusively observed when tumors biosensor cells stably expressing a GFP under the control of grew on immunocompetent mice, yet was lost when the tumors the MX1 (a Type I IFN response ) promoter. To this aim proliferated on athymic (nu/nu)mice(Figure 4(d,e)). These the supernatant of U2OS cells following treatment and addi- results underscore the obligate contribution of the immune tional 48 hours incubation with fresh media was used to treat system to the chemotherapeutic activity of lurbinectedin. ONCOIMMUNOLOGY e1656502-3

a 4 h 8 hU2OS 16 h 32 h Ctrl ** *** *** *** *** *** * *** *** *** *** ** *** *** *** ****** *** *** *** *** ****** ****** *** *** *** *** *** * ** *** *** ***** ** *** *** *** *** ** * **

* *** *** *** *** * ** *** ** ** Lurbi MTX ** * *** *** *** *** **** *** *** * ** ATP ATP ATP ATP MX1 MX1 MX1 MX1 VIAB VIAB VIAB VIAB CALR CALR CALR CALR HMGB1 HMGB1 HMGB1 HMGB1

b 4h 8 hHCC70 16 h 32 h Ctrl ** *** ** ** ** **** ***** * *** ** *** ** *** *** **** *** ********* *** ** * *** *** * * ** *** *** ***** *** ** *** *** **

*** *** ** *** *** *** ***** *** * Lurbi MTX *** *** *** ** *** *** * *** *** ** * ATP ATP ATP ATP MX1 MX1 MX1 MX1 VIAB VIAB VIAB VIAB CALR CALR CALR CALR HMGB1 HMGB1 HMGB1 HMGB1

c 4 h 8 hHT29 16 h 32 h Ctrl **** *** *** *** *** ** *** *** * ** ** *** **** *** *** *** *** *** * *** *** *** ** * * ** ** ** ** *** *** ** *** *** *** *

*** * ***** *** ***** *** *** *** *** *** *** Lurbi MTX *** ** ** * *** ** *** *** ** *** *** *** ** ATP ATP ATP ATP MX1 MX1 MX1 MX1 VIAB VIAB VIAB VIAB CALR CALR CALR CALR HMGB1 HMGB1 HMGB1 HMGB1

d 4 h 8 hMCA 16 h 32 h Ctrl ********* *** *** *** ** *** *** ** * *** ** *** *** ** *** *** *** *** *** ** * * *** ** *** * ** *** * *** *** *** ***

**** ** * *** *** *** *** ** *** Lurbi MTX ** ** ** *** ** *** * *** *** *** * ATP ATP ATP ATP MX1 MX1 MX1 MX1 VIAB VIAB VIAB VIAB CALR CALR CALR CALR HMGB1 HMGB1 HMGB1 HMGB1

Figure 1. Immunogenic cell death assessment in solid tumors. (a) Human osteosarcoma U2OS (a), human breast cancer HCC70 (b) human colon cancer HT29 cells (c) and murine methylcholantrene-induced fibrosarcoma MCA205 cells (d) were treated with lurbinectedin (Lurbi, 1 nM, 10 nM, 100 nM and 1 µM) for the indicated times. Subsequently, the cells were stained with 1 µM Hoechst 33342 and 1 µM propidium iodide and assessed for the loss of viability by automated image acquisition. Images were segmented, cellular debris was excluded and the number of cells with normal nuclear morphology was enumerated. Cells stably expressing CALR-GFP were treated as above. Following the cells were fixed with 3.7% of PFA, stained with 1 µM Hoechst 33342 and assessed by automated image acquisition. Images were segmented, cellular debris was excluded and CALR-GFP granularity (a surrogate marker of CALR exposure) was evaluated in the cytoplasmic region of cells with normal nuclear morphology. Wild type cells were treated as above and then assessed for cytoplasmic quinacrine granularity (after staining with the ATP-sensitive dye quinacrine together with Hoechst 33342) by automated image acquisition, segmentation and analysis. Cells stably expressing HMGB1-GFP were treated as above and then assessed for nuclear HMGB1-GFP fluorescence intensity. The cells were fixed and stained with Hoechst 33342 and images were acquired, segmented and analyzed. WT cells were treated as above and following the media was changed and the cells were incubated for 48 hours before the supernatant was used to treat MX1-GFP biosensor cells for additional 48 hours. The cells were fixed and stained with Hoechst 33342 before type 1 IFN responses were monitored by means of automated microscopy as an increase in GFP fluorescence intensity. Mitoxantrone (MTX, 1 and 3 µM) was used as a positive control. The means of quadruplicate assessments and p-values are depicted as heat maps. (*p < .01; **p < .005; ***p < .001, two-tailed Student’s t test).

Combinatorial effects of lurbinectedin and αPD-1/αCTLA-4 with immune checkpoint blockers targeting CTLA-4 or PD-1. double immune checkpoint blockade For this, established MCA205 fibrosarcomas were treated with Given the capacity of lurbinectedin to induce immune- Lurbinectedin as before and subjected to immunotherapy dependent anticancer effects on established tumors, we inves- with antibodies specific for CTLA-4, PD-1 or a combination tigated whether this agent could sensitize cancers to therapy of both on day 6, 9 and 12, when the anticancer immune e1656502-4 W. XIE ET AL.

*** ab+ 80

Ctrl Lurbi Thaps 60 * *** 40

20

% of cells p-eIF2a 0 Hoechst 33342 α-p-eIF2a Ctrl Lurbi Thaps 0.8 cd* 0.6 Ctrl Lurbi ActD *** 0.4 *** *** 0.2 SOC NCL FBL SOC NCL 0 a-NCL α-FBL Ctrl

Lurbi ActD

Figure 2. Traits of immunogenic cell death. Human osteosarcoma U2OS cells were treated with 10, 50 or 100 nM lurbinectedin (Lurbi) for 6 hours. Thapsigargin (Thaps, 3 µM) was used as a positive control. The cells were fixed with 3.7% PFA and DNA was stained with 1 µM Hoechst 33342. Following the phosphorylation of the eukaryotic translation initiation factor 2 alpha (eIF2a) was assessed with phosphoneoepitope-specific antibody and was monitored by means of automated microscopy as an increase in cytoplasmic fluorescence intensity. (a,b) The level of transcription was measured in U2OS cell treated as above with Lurbi. The transcription inhibitor actinomycin D (ActD) was used as a control. The cells were fixed as above and following the colocalization of nucleolin and fibrillarin was assessed as an indicator for transcriptional activity (c,d) Scale bar equals 10 µm and bar charts depict depict mean values ± SD of quadruplicate assessments (*p < .01; ***p < .001, two-tailed Student’s t test).

a One week Day0 of latency Day0 Tumor monitoring sc sc Vaccination Rechallenge Lurbi treated MCA205 MCA205 b c C57BL/6 200 PBS (n=6) 100 ) 2 Lurbi (n=6)

100 50

%survival PBS (n=8) * Lurbi (n=8) Tumor size (mm Tumor 0 0 0102030 0210 030 Days post treatment Days post treatment

Figure 3. Anticancer vaccination efficacy of lurbinectedin-treated cells. MCA205 cells treated for 20 h with 1 µM lurbinectedin were inoculated subcutaneously (s.c.) into immunocompetent C57BL/6 mice, which were rechallenged 7 days later s.c. with living cells of the same type. The tumor growth was measured until endpoints were reached and overall survival was evaluated regularly for the following 30 days (n = 6). (*p < .01, two-tailed Student’s t test, compared to all other groups). Data were analyzed with TumGrowth. response in the tumor peaks (Figure 5(a)). Tumor monitoring upon rechallenge with the same cancer cell type from which led to the deduction that the most efficient therapeutic regi- they had been cured (MCA205), yet developed cancers when men was a combination of all three anticancer agents (lurbi- rechallenged with TC1 tumor cells (Figure 5(f,g)). Thus, mice nectedin, αCTLA-4 and αPD-1) in contrast to single-ICB that had been cured by a combination of systemic lurbinecte- therapies that appeared to be relatively inefficient in this din-based chemotherapy and immunotherapy had established setting (Figure 5(b–e)). The combination of lurbinectedin a specific anticancer immune response that generated immu- with αCTLA-4/αPD-1 dual checkpoint blockade in tumor- nological memory. bearing animals significantly extended life expectancy and, moreover, led to tumor clearance in 3/8 mice in the time Lurbinectedin retards the growth of carcinogen-induced frame of the experiment (Figure 5(e)). The effect of lurbinec- and spontaneous breast cancer tedin with αCTLA-4/αPD-1 dual checkpoint blockade was abrogated in conditions in which CD4+ and CD8+ cytotoxic To explore the potential lurbinectedin for the therapy of breast T lymphocytes (CTLs) were depleted. Mice that had been cancer, we took advantage of a hormone/carcinogen induced rendered tumor-free for more than 50 days rejected tumors breast cancer model activated by the continuous stimulation of ONCOIMMUNOLOGY e1656502-5

a One week of latency Day06917 214 sc iv iv iv MCA205 Lurbi Lurbi Lurbi Bc 200 200 ) ) C57BL/6 nu/nu 2 2

100 100

PBS (n=8) PBS (n=8) Lurbi (n=8) Lurbi (n=8) Tumor size (mm Tumor Tumor size (mm Tumor 0 0 0102030 0102030 Days post treatment Days post treatment d 100 C57BL/6 e 100 nu/nu

50 rvival 50 su %survival PBS (n=8) % PBS (n=8) Lurbi (n=8) ** Lurbi (n=8) 0 0 02100 30 02100 30 Days post treatment Days post treatment

Figure 4. Therapeutic efficacy of lurbinectedin in immunocompetent and immunodeficient hosts. Live MCA205 cells were injected subcutaneously (s.c.) into immunocompetent C57BL/6 mice or immunodeficient nu/nu mice as depicted in the scheme in (a) When tumors became palpable, mice were intravenously (i.v.) injected with 0.14 mg/Kg lurbinectedin (on day 1,7 and 14). Tumor growth was assessed regularly for the following 30 days. Data is depicted as tumor growth curves (b,d) and overall survival plots (c,e). Data were analyzed with TumGrowth. progesterone receptors by medroxyprogesterone acetate (MPA) combined with clinically established immune checkpoint block- and the repeated exposure to the DNA-damaging agent ade regimens. dimethylbenzantracene (DMBA). This induced model of breast cancer is known to be modulated by the immune system.22 We treated mice with palpable MPA/DMBA-induced tumors by Materials & methods systemic injection with lurbinectedin alone or in combination Cell culture and chemicals with double immune checkpoint blockade neutralizing CTLA-4 and PD-1 (Figure 6(a)). Both interventions significantly reduced All media and cell culture supplements were from Thermo Fisher tumor growth and increased overall survival. However, only the Scientific (Carlsbad, CA, US). Lurbinectedin was provided by combination with αCTLA-4/αPD-1 yielded tumor clearance in PharmaMar (Madrid, Spain). Cell culture plastics and consum- the time frame of the experiment (Figure 6(b–d)). ables were purchased from Greiner Bio-One (Kremsmünster, Austria). Human osteosarcoma U2OS cells previously genetically altered as described earlier,23 murine methylcholanthrene- induced fibrosarcoma MCA-205 cells and murine lung cancer Concluding remarks TC-1 cells were cultured in Glutamax®-containing DMEM med- The results of this study suggest that lurbinectedin efficiently ium supplemented with 10% fetal bovine serum (FBS), and 10 mM induces cell death in a broad panel of solid tumors. This proce- HEPES. Cells were cultured in a temperature-controlled environ- dure likely does not only cause the cells to succumb to disinte- ment at 37°C with a humidified atmosphere containing 5% CO2. gration but rather triggers traits of immunogenic cell death, including the phosphorylation of eIF2a and the release of danger Automated image acquisition and analysis associated molecular patterns (DAMPs). Irrespective of the exact molecular mechanisms accounting for these effects, there are One day before the experiment 5 × 103 cells were seeded in 96- a number of evidences advocating for lurbinectedin-triggered well µClear imaging plates (Greiner BioOne) and let adhere cancer-specific immunogenicity. Thus, animals that had been under standard culture conditions. The following day cells were cured by lurbinectedin from established cancers became resis- treated with lurbinectedin at 0.001, 0.01, 0.1 and 1 µM for 4, 8, 16 tance to rechallenge with the same cancer type. The therapeutic or 32 hours. Then cells were fixed with 3.7% formaldehyde effect of lurbinectedin was neutralized in conditions in which supplemented with 1 μg/ml Hoechst 33342 for 30 min at RT. either the host was immunocompromised or T-cell had been The fixative was changed to PBS and the plates were analyzed by depleted. Furthermore, the recapitulation in a heterogeneous automated microscopy. For the detection of ATP enriched vesi- spontaneous tumor model of effects that were previously cles, the cells were labeled after 4, 8, 16 or 32 hours of incubation observed in homogenous transplanted tumors indicates that with the fluorescent dye quinacrine (as described before23). In the results presented here hold a high translational value. short, cells were incubated with 5 µm quinacrine and 1 µg/ml Altogether, these results convincingly demonstrate that lurbi- Hoechst 33342 in Krebs-Ringer solution (125 mM NaCl, 5 mM nectedin mediated immunochemotherapy may be advantageously KCl, 1 mM MgSO4, 0.7 mM KH2PO4,2mmCaCl2,6mM e1656502-6 W. XIE ET AL.

a One week of latency Day06917 214 sc ip ip ip iv iv iv MCA205 ICBICB ICB Lurbi Lurbi Lurbi b c PBS (n=7) αPD-1/αCTLA4 (n=7) Lurbi + αCTLA4 (n=8) Lurbi + αPD-1/αCTLA4+ αCD4/CD8 (n=8) Lurbi + αPD1 (n=8) Lurbi+ αPD-1/αCTLA4 (n=8) 200 200 ) ) 2 2

100 100 Tumor clearance n=3 Tumor size (mm Tumor Tumor size (mm Tumor 0 0 0102030 0102030 Days post treatment Days post treatment d e PBS PBS (n=7) αPD-1/αCTLA4 Lurbi + αCTLA4 (n=8) Lurbi + αPD-1/αCTLA4 + αCD4/CD8 Lurbi + αPD1 (n=8) Lurbi + αPD-1/αCTLA4 100 100

# ival

50 rv 50 %survival %su

0 0 02100 30 02100 30 Days post treatment Days post treatment

f Naive (n=3) g Naive (n=3) 200 200 ) ) Lurbi + αPD-1/αCTLA4 (n=2) Lurbi + αPD-1/αCTLA4 (n=2) 2 2 MCA205 TC-1

100 100 Tumor rejection n=2 Tumor size (mm Tumor Tumor size (mm Tumor 0 0 01020 01020 Days post rechallenge Days post rechallenge

Figure 5. Sequential lurbinectedin treatment with double immune checkpoint blockade exhibits systemic antitumor immunity C57BL/6 mice were inoculated subcutaneously (s.c.) with murine fibrosarcoma MCA205. Palpable tumors were treated with sequential intravenous (i.v.) injections of 0.14 mg/Kg lurbinectedin (Lurbi) as indicated in (a). Single- or double-immune checkpoint blockade was mounted by sequential intraperitoneal (i.p.) injections of monoclonal antibodies targeting CTLA-4 or PD-1 at day 6, 9 and 12 post treatment and tumor growth (b,c) and overall survival (d,e) were assessed regularly for the following 30 days. (f,g) The generation of immunological memory was assessed in cured animals by rechallenged with MCA205 and TC-1. Naïve animals were used as controls. Individual tumor growth curves are depicted. Data were analyzed with TumGrowth. glucose and 25 mM Hepes, pH 7.4) for 30 minutes at 37°C. nuclear fluorescence intensity. Cellular debris was excluded Thereafter, cells were rinsed with Krebs-Ringer and viable cells from the analysis and secondary cytoplasmic ROIs were used were microscopically examined. For automated fluorescence for the quantification of CALR-GFP or quinacrine containing microscopy a robot-assisted Molecular Devices IXM XL vesicles. For the latter, the images were segmented and analyzed BioImager (Molecular Devices, Sunnyvale, CA, USA) equipped for GFP granularity by comparing the standard deviation of the with SpectraX light source (Lumencor, Beaverton, OR, USA), mean fluorescence intensity of groups of adjacent pixels (coeffi- adequate excitation and emission filters (Semrock, Rochester, cient of variation) within the cytoplasm of each cell to the mean NY, USA) and a 16-bit monochromes sCMOS PCO.edge 5.5 fluorescence intensity in the same ROI using the MetaXpress camera (PCO Kelheim, Germany) and a 20 X PlanAPO objective software (Molecular Devices). (Nikon, Tokyo, Japan) was used to acquire a minimum of 9 view fields, followed by automated image processing with the custom module editor within the MetaXpress software (Molecular In vivo experimentation Devices). Depending on the utilized biosensor cell line the pri- mary region of interest (ROI) was defined by a polygon mask Six- to eight-week-old female wild-type C57BL/6 and nu/nu mice around the nucleus allowing for the enumeration of cells, the were obtained from Envigo France (Huntingdon, UK) and were detection of morphological alterations of the nucleus and kept in the animal facility at the Gustave Roussy Campus Cancer ONCOIMMUNOLOGY e1656502-7

a Several month of latency Day 067 91214 ip ip ip iv iv iv DMBA/MPA ICBICB ICB Lurbi Lurbi Lurbi b c PBS (n=4) αPD-1/αCTLA-4 (n=5) Lurbi (n=5) Lurbi + αPD-1/αCTLA-4 (n=4) 150 150 ) ) 2 2

100 100

50 50 Tumor clearance n=1

Tumor size (mm Tumor 0 size (mm Tumor 0 0102030 0102030 Days post treatment Days post treatment d PBS Lurbi αPD-1/αCTLA-4 Lurbi + αPD-1/αCTLA-4 100

** 50 ** %survival

0 02100 30 Days post treatment

Figure 6. Lurbinectedin retards the growth of spontaneous tumors. Medroxyprogesterone acetate (MPA) pellets (50 mg, 90-day release) were implanted subcutaneously into the interscapular area of immunocompetent C57BL/6 mice. Then the animals received 1 mg dimethylbenzantracene (DMBA) administered by oral gavage 6 x during 7 weeks. When spontaneous tumors became palpable mice were randomly assigned to receive 0,14 mg/Kg lurbinectedin (Lurbi) alone or in combination with double immune checkpoint blockade with monoclonal antibodies targeting CTLA-4 and PD-1 at day 6, 9 and 12 post treatment (a). The tumor area and overall survival were measured regularly until ethical endpoints were reached (b,c,d). Data were analyzed with TumGrowth (https://github.com/kroemerlab). in a specific pathogen–free and temperature-controlled environ- p < .01 and ***p < .001). Murine fibrosarcoma MCA205 cells were ment with 12 h day, 12 h night cycles and received food and incubated with 1 µM lurbinectedin for 24 h, resulting in approxi- water ad libitum. Animal experiments were conducted in com- mately 70% cell death. For vaccination experiments, 3 × 105 dying pliance with the EU Directive 63/2010 and protocols 2013_094A MCA205 cells were inoculated s.c. into the left flanks of six-week- and were approved by the Ethical Committee of the Gustave old female C57BL/6 mice. Seven to ten days later, animals were re- Roussy Campus Cancer (CEEA IRCIV/IGR no. 26, registered at challenged in the opposite flank with 3 × 105 living MCA205 cells, the French Ministry of Research). As described,24,25 MCA205 and tumor growth and incidence were monitored. Six-week-old tumors were established in C57BL/6 mice by subcutaneously (s. female C57BL/6 mice (n = 12 per group) underwent surgical c.) injection of 5 × 105 cells. When tumors became palpable, implantation of slow-release medroxyprogesterone acetate 0.18 mg/Kg lurbinectedin was injected sequentially once a week (MPA) pellets (50 mg, 90-day release; Innovative Research of intravenously into the tail vein and animal well-being and tumor America, Sarasota, Fl, US) s.c. Two-hundred μLof5mg/mL growth were monitored. A total of 0.5 mg of anti-CD8 (clone 2.43 dimethylbenzantracene(DMBA,SigmaAldrich,St.Louis,MO, BioXCell BE0061) and anti-CD4 (clone GK1.5 BioXCell BE0003- US) dissolved in corn oil was administered by oral gavage once 1) intraperitoneal (i.p.) injections were repeated every 7 days to per week for 7 weeks. assure the complete depletion of both T cell populations during thewholeexperiment.Miceweresacrificedwhentumorsize reached end-point or signs of obvious discomfort associated to Immune checkpoint blockade the treatment were observed following the EU Directive 63/2010 and our Ethical Committee advice. Tumor-free animals were kept Double or single immune checkpoint blockade was applied by for more than 30 days before testing the generation of immuno- repeated intraperitoneal injections of monoclonal antibody logical memory by s.c.rechallengewith5×105 TC-1 in one flank specific to PD-1 (200 µg, Clone 29F.1A12, BioXcell, West and 5 × 105 MCA205 cells injected in the contralateral flank. Lebanon, NH, USA) or CTLA-4 (200 µg, Clone 9D9, Animals were monitored and tumor growth documented regu- BioXcell) at day 6, 9 and 12 upon initiation of the treatment larly until end-points were reached. Statistical analysis was per- with lurbinectedin. Animals were monitored regularly and the formed by applying 2-way ANOVA analysis followed by tumor growth was documented until ethical end-points were Bonferroni’s test comparing to control conditions (* p <.05,** reached. Statistical analysis was performed employing 2-way e1656502-8 W. XIE ET AL.

ANOVA analysis followed by Bonferroni’s test comparing to 3. Obeid M, Tesniere A, Ghiringhelli F, et al. Calreticulin exposure control conditions (* p < .05, ** p < .01 and ***p < .001). dictates the immunogenicity of cancer cell death. Nat Med. 2007;13(1):54–61. doi:10.1038/nm1523. 4. Ray-Coquard I, Cropet C, Van Glabbeke M, et al. Lymphopenia as Statistical procedures a prognostic factor for overall survival in advanced carcinomas, sarcomas, and lymphomas. Cancer Res. 2009;69(13):5383–5391. Unless otherwise specified, experiments were performed in doi:10.1158/0008-5472.CAN-08-3845. quadruplicate instances. Data were analyzed with the freely 5. Fridman WH, Zitvogel L, Sautes-Fridman C, Kroemer G. The immune contexture in cancer prognosis and treatment. Nat Rev available software R (https://www.r-project.org). Significances Clin Oncol. 2017;14(12):717–734. doi:10.1038/nrclinonc.2017.101. were calculated using a student t-test with Welch correction. 6. Lesterhuis WJ, Haanen JB, Punt CJ. Cancer immunotherapy– Thresholds for each assay were applied based on the Gaussian revisited. Nat Rev Drug Discovery. 2011;10(8):591–600. distribution of positive and negative controls. In vivo tumor doi:10.1038/nrd3500. growth was analyzed with the help of the TumGrowth software 7. Zitvogel L, Kepp O, Kroemer G. Immune parameters affecting the 26 efficacy of chemotherapeutic regimens. Nat Rev Clin Oncol. package freely available at https://github.com/kroemerlab. 2011;8(3):151–160. doi:10.1038/nrclinonc.2010.223. 8. Zappasodi R, Pupa SM, Ghedini GC, et al. Improved clinical outcome in indolent B-cell lymphoma patients vaccinated with Abbreviations autologous tumor cells experiencing immunogenic death. Cancer Res. 2010;70(22):9062–9072. doi:10.1158/0008-5472. DAMP danger-associated molecular pattern CAN-10-1825. ICD immunogenic cell death 9. Fucikova J, Kralikova P, Fialova A, et al. Human tumor cells killed by anthracyclines induce a tumor-specific immune response. Cancer Res. 2011;71(14):4821–4833. doi:10.1158/ Acknowledgments 0008-5472.CAN-11-0950. 10. Ghiringhelli F, Apetoh L, Tesniere A, et al. Activation of the WX and YW were supported by the China Scholarship Council, JH is the NLRP3 inflammasome in dendritic cells induces IL-1beta- recipient of a PhD fellowship from the Foundation Philanthropia; GK is dependent adaptive immunity against tumors. Nat Med. 2009;15 supported by the Ligue contre le Cancer (équipe labellisée); Agence (10):1170–1178. doi:10.1038/nm.2028. National de la Recherche (ANR) – Projets blancs; ANR under the 11. Michaud M, Martins I, Sukkurwala AQ, et al. Autophagy- frame of E-Rare-2, the ERA-Net for Research on Rare Diseases; dependent anticancer immune responses induced by chemother- Association pour la recherche sur le cancer (ARC); Cancéropôle Ile-de- apeutic agents in mice. Science. 2011;334(6062):1573–1577. France; Chancelerie des universités de Paris (Legs Poix), Fondation pour doi:10.1126/science.1208347. la Recherche Médicale (FRM); a donation by Elior; European Research 12. Apetoh L, Ghiringhelli F, Tesniere A, et al. The interaction Area Network on Cardiovascular Diseases (ERA-CVD, MINOTAUR); between HMGB1 and TLR4 dictates the outcome of anticancer Gustave Roussy Odyssea, the European Union Horizon 2020 Project chemotherapy and radiotherapy. Immunol Rev. 2007;220:47–59. Oncobiome; Fondation Carrefour; High-end Foreign Expert Program doi:10.1111/j.1600-065X.2007.00573.x. in China (GDW20171100085 and GDW20181100051), Institut National 13. Tesniere A, Schlemmer F, Boige V, et al. Immunogenic death of du Cancer (INCa); Inserm (HTE); Institut Universitaire de France; colon cancer cells treated with oxaliplatin. Oncogene. 2010;29 LeDucq Foundation; the LabEx Immuno-Oncology; the RHU Torino (4):482–491. doi:10.1038/onc.2009.356. Lumière; the Seerave Foundation; the SIRIC Stratified Oncology Cell 14. Sistigu A, Yamazaki T, Vacchelli E, et al. Cancer cell-autonomous DNA Repair and Tumor Immune Elimination (SOCRATE); and the contribution of type I interferon signaling to the efficacy of SIRIC Cancer Research and Personalized Medicine (CARPEM). chemotherapy. Nat Med. 2014;20(11):1301–1309. doi:10.1038/ nm.3708. 15. Apetoh L, Ghiringhelli F, Tesniere A, et al. Toll-like receptor Conflict of interest disclosure 4-dependent contribution of the immune system to anticancer che- – This study was supported by Pharmamar. JFM-L is a full-time employee motherapy and radiotherapy. Nat Med. 2007;13(9):1050 1059. of Pharmamar. doi:10.1038/nm1622. 16. Ma Y, Adjemian S, Mattarollo SR, et al. Anticancer chemotherapy-induced intratumoral recruitment and differen- ORCID tiation of antigen-presenting cells. Immunity. 2013;38 (4):729–741. doi:10.1016/j.immuni.2013.03.003. Kristina Iribarren http://orcid.org/0000-0002-2930-8573 17. Vacchelli E, Ma Y, Baracco EE, et al. Chemotherapy-induced Allan Sauvat http://orcid.org/0000-0001-7076-8638 antitumor immunity requires formyl peptide receptor 1. Science. Maria Perez-Lanzon http://orcid.org/0000-0002-6306-6520 2015;350(6263):972–978. doi:10.1126/science.aad0779. Guido Kroemer http://orcid.org/0000-0002-9334-4405 18. Kroemer G, Senovilla L, Galluzzi L, Andre F, Zitvogel L. Natural Oliver Kepp http://orcid.org/0000-0002-6081-9558 and therapy-induced immunosurveillance in breast cancer. Nat Med. 2015;21(10):1128–1138. doi:10.1038/nm.3944. 19. Vacchelli E, Enot DP, Pietrocola F, Zitvogel L, Kroemer G. References Impact of pattern recognition receptors on the prognosis of breast cancer patients undergoing adjuvant chemotherapy. 1. Casares N, Pequignot MO, Tesniere A, et al. Caspase-dependent Cancer Res. 2016;76(11):3122–3126. doi:10.1158/0008-5472. immunogenicity of doxorubicin-induced tumor cell death. J Exp CAN-16-0294. Med. 2005;202(12):1691–1701. doi:10.1084/jem.20050915. 20. Santamaría Nuñez G, Robles CM, Giraudon C, et al. 2. Mattarollo SR, Loi S, Duret H, Ma Y, Zitvogel L, Smyth MJ. Lurbinectedin specifically triggers the degradation of phosphory- Pivotal role of innate and adaptive immunity in anthracycline lated RNA polymerase II and the formation of DNA breaks in chemotherapy of established tumors. Cancer Res. 2011;71 cancer cells. Mol Cancer Ther. 2016;15(10):2399–2412. (14):4809–4820. doi:10.1158/0008-5472.CAN-11-0753. doi:10.1158/1535-7163.MCT-16-0172. ONCOIMMUNOLOGY e1656502-9

21. Peltonen K, Colis L, Liu H, et al. A targeting modality for destruc- 24. Iribarren K, Buque A, Mondragon L, et al. Anticancer effects of tion of RNA polymerase I that possesses anticancer activity. anti-CD47 immunotherapy in vivo. Oncoimmunology. 2019;8 Cancer Cell. 2014;25(1):77–90. doi:10.1016/j.ccr.2013.12.009. (3):1550619. doi:10.1080/2162402X.2018.1550619. 22. Pietrocola F, Pol J, Vacchelli E, et al. Caloric restriction mimetics 25. Zhou H, Mondragon L, Xie W, et al. Oncolysis with DTT-205 and enhance anticancer immunosurveillance. Cancer Cell. 2016;30 DTT-304 generates immunological memory in cured animals. (1):147–160. doi:10.1016/j.ccell.2016.05.016. Cell Death Dis. 2018;9(11):1086. doi:10.1038/s41419-018-1127-3. 23. Martins I, Kepp O, Schlemmer F, et al. Restoration of the immu- 26. EnotDP,VacchelliE,JacquelotN,ZitvogelL,TumGrowth:KG.An nogenicity of cisplatin-induced cancer cell death by endoplasmic open-access web tool for the statistical analysis of tumor growth curves. reticulum stress. Oncogene. 2011;30(10):1147–1158. doi:10.1038/ Oncoimmunology. 2018;7(9):e1462431. doi:10.1080/2162402X.20 onc.2010.500. 18.1490854.

Tumor lysis with LTX-401 creates anticancer immunity

Previous results demonstrated that LTX-401 could induce ICD hallmarks in vitro, and intra- tumour injection of LTX-401 showed potent antitumor capacity in few cancer models (Eike et al., 2016; Mauseth et al., 2019; Zhou et al., 2016b). The capacity of LTX-401 to stimulate anticancer immune response and combination effect with ICBs remains to be elucidated. Since LTX-401 is designed for future clinical usage, we first tested its safety profile at therapeutically effective dose. LTX-401, similar to LTX-315 that is under clinical evaluation in patients with transdermal accessible tumours, did not depict any signs of adverse toxicity in vivo. We then investigated the therapeutic effect of LTX-401 in subcutaneous MCA205 fibrosarcoma and TC-1 lung cancer. Intratumoral delivery of LTX-401 reduced tumour growth and increased survival rate in both models. 3 out of 12 and 8 out of 12 mice were cured in MCA205 and TC-1 settings, respectively. All these cured mice were resistant to the rechallenge of same cell type. To study the abscopal effects of LTX-401-mediated tumour lysis, mice were implanted with MCA205/TC-1 subcutaneously on one flank and four days later on the contralateral flank. Only primary tumour was treated by LTX-401 and the growth of contralateral tumour was monitored. The local treatment had a significant tumour growth reducing effect on primary tumours yet only limited effects on secondary contralateral tumours. Given that LTX-401 could induce anticancer immune response and facilitate T cell infiltration in tumour bed, we combined LTX-401 and ICBs to explore their potential synergic effects. Pre- treatment with LTX-401 greatly improved the therapeutic efficiency of αCTLA-4, αPD-1 or αCTLA-4 plus αPD-1. The combination of LTX-401 and dual checkpoint blockade (αCTLA-4 plus αPD-1) or the combination of LTX-401 and αCTLA-4 (but not αPD-1) alone were able to significantly extend life expectancy. Again, cured tumour-free mice generated tumour specific immunological memory. Particularly, sequential LTX-401 treatment with double ICBs exhibits superior systemic antitumor immunity. 100% of treated tumours and 6 out of 9 contralateral tumours were completely regressed.

41

ONCOIMMUNOLOGY 2019, VOL. 8, NO. 7, e1594555 (8 pages) https://doi.org/10.1080/2162402X.2019.1594555

BRIEF REPORT Tumor lysis with LTX-401 creates anticancer immunity Wei Xiea,b,c,d, Laura Mondragóna,b,c,d*, Brynjar Mausethe,f,g*, Yan Wanga,b,c,d, Jonathan Pola,b,c,d, Sarah Lévesquea,b,c,d, Heng Zhoua,b,c,d,h, Takahiro Yamazaki i,j,k,l,m, Johannes J. Eksteenn, Laurence Zitvogel i,j,k,l, Baldur Sveinbjørnssone,o,p, Øystein Rekdal e,o,p, Oliver Kepp a,b,c,d, and Guido Kroemer a,b,c,d,p,q aMetabolomics and Cell Biology Platforms, Gustave Roussy Comprehensive Cancer Institute, Villejuif, France; bEquipe 11 labellisée Ligue contre le Cancer, Centre de Recherche des Cordeliers, INSERM, Paris, France; cUniversité Paris Descartes, Sorbonne Paris Cité, Paris, France; dUniversité Pierre et Marie Curie, Paris, France; eLytix Biopharma, Oslo, Norway; fDivision of Cancer, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo, Norway; gInstitute of Clinical Medicine, University of Oslo, Oslo, Norway; hInstitute of Modern Physics, Chinese Academy of Sciences, Lanzhou, China; iGustave Roussy Comprehensive Cancer Center, Villejuif, France; jUniversity of Paris Sud XI, Kremlin Bicêtre, France; kInstitut National de la Santé et de la Recherche Medicale (INSERM), Villejuif, France; lCenter of Clinical Investigations in Biotherapies of Cancer (CICBT), Villejuif, France; mWeill Cornell Medical College, New York, NY, USA; nNorut Northern Research Institute, SIVA Innovation Centre, Tromsø, Norway; oInstitute of Medical Biology, University of Tromsø, Tromsø, Norway; pKarolinska Institutet, Department of Women’s and Children’s Health, Stockholm, Sweden; qPôle de Biologie, Hôpital Européen Georges Pompidou, AP-HP, Paris, France

ABSTRACT ARTICLE HISTORY Local immunotherapies such as the intratumoral injection of oncolytic compounds aim at reinstating Received 22 January 2019 and enhancing systemic anticancer immune responses. LTX-315 is a first-in-class, clinically evaluated Revised 2 March 2019 oncolytic peptide-based local immunotherapy that meets these criteria. Here, we show that LTX-401, yet Accepted 9 March 2019 another oncolytic compound designed for local immunotherapy, depicts a similar safety profile and that KEYWORDS sequential local inoculation of LTX-401 was able to cure immunocompetent host from subcutaneous Oncolysis; anticancer MCA205 and TC-1 cancers. Cured animals exhibited long-term immune memory effects that rendered therapy; immunogenic cell them resistant to rechallenge with syngeneic tumors. Nevertheless, the local treatment with LTX-401 death; checkpoint blockade alone had only limited abscopal effects on secondary contralateral lesions. Anticancer effects resulting from single as well as sequential injections of LTX-401 were boosted in combination with PD-1 and CTLA-4 immune checkpoint blockade (ICB), and sequential LTX-401 treatment combined with double ICB exhibited strong abscopal antineoplastic effects on contralateral tumors underlining the potency of this combination therapy.

– Introduction injected into rodents.3,7,9 13 Moreover, LTX-315 treatments sensitize to subsequent immune checkpoint blockade with The word ‘oncolysis’ usually evokes oncolytic viruses, i.e. αCTLA-4 antibodies,11 while tumors that are successfully viruses that selectively infect and kill cancer cells.1,2 cured with DTT-205 or DTT-304 established a long-term However, oncolysis can also be achieved by applying physical, immune response rendering them resistant against later chemical or pharmacological agents to tumor in a way that rechallenge with the cancer cells that had been eliminated the cancer is locally destroyed. For example, cationic ampho- from them.8 lytic peptides or peptide mimetics can be used to target Here, we compared the safety profiles of different lytic cellular membranes, causing their disruption by detergent- peptides to discover that LTX-401 (similar to LTX-315) was like effects. We have been characterizing the mode of action far less toxic than DTT-205 and DTT-304. We also investi- of several oncolytic reagents of this class such as LTX-315, gated the capacity of LTX-401 to stimulate anticancer LTX-401, DTT-205 and DTT-304. LTX-315 turned out to immune responses in three different experimental setups. preferentially target mitochondrial membranes;3,4 while – Our results support the hypothesis that oncolysis by LTX- LTX-401 destroys Golgi apparatus-associated membranes,5 7 401 is highly immunogenic. thus setting off a different cell-destroying cascade.7 DTT-205 and DTT-304, instead, have a tropism for lysosomal mem- branes, thus activating yet another cell death subroutine.8 Results and discussion Despite their diversity, these oncolytic peptide and peptide Differential toxicity of oncolytic agents derivatives all appear to be able to stimulate anticancer immune responses, at least in specific circumstances. Thus, In a retrospective study, the safety profiles of oncolytic pep- cancer cells treated with LTX-315 or LTX-401 in vitro can tides and peptidomimetics at their therapeutically effective stimulate a protective antineoplastic immune response when doses were compared. The onset of manifest discomfort or

CONTACT Oliver Kepp [email protected]; Guido Kroemer [email protected] Metabolomics and Cell Biology Platforms, Gustave Roussy Comprehensive Cancer Institute, Villejuif, France *Contributed equally © 2019 Taylor & Francis Group, LLC e1594555-2 W. XIE ET AL. death that was defined as an endpoint for in vivo experimen- supporting the idea that they had developed a long-term immu- tation was evaluated for the sequential treatment of tumors nological memory. Very similar results were obtained when established on C57Bl/6 mice with therapeutically effective MCA205 fibrosarcomas were replaced by non-small cell lung doses of LTX-315, LTX-401, DTT-205 and DTT-304 (Figure adenocarcinoma TC-1 cells (that have been engineered to express 1A). Data from in vivo experimentation utilizing oncolytic the human papilloma virus antigen E6). Again, local therapy with compounds in C57BL/6 mice bearing palpable subcutaneous LTX-401injectionswasabletogreatlyreducethegrowthof solid tumors were evaluated by enumerating global survival primary TC-1 carcinomas (Figure 2E), yielding a 60% survival (Figure 1B). Whereas both experimental compounds DTT- rate at 60 days (Figure 2F). Again, mice that were cured became 205 and DTT-304 led to death in a certain number of animals, resistant to rechallenge with TC-1 (but not MCA205) cells (Figure LTX-401, similar to LTX-315, that is under clinical evaluation 2G), confirming the contention that a tumor-specific memory in patients with transdermally accessible tumors, did not immune response had been established. depict any signs of adverse toxicity in vivo.

Limited abscopal effects of LTX-401-mediated tumor lysis Stimulation of anticancer immunity by LTX-401 In the next step, we interrogated the experimental system for the To identify the capacity of LTX-401 to stimulate anticancer induction of abscopal effects, i.e. a presumably immune-mediated immune responses, we treated established MCA205 fibrosarco- long-distance effect of the injected LTX-401 peptide that would mas growing subcutaneously (s.c.) in C57BL/6 mice by daily reduce the growth of a non-injected tumor. For this, we inocu- intratumoral injection (up to 4 times) of the compound (or, as lated mice with MCA205 fibrosarcomas into the right flank to a control, its vehicle PBS) until macroscopically detectable tumors create a primary tumor (that became palpable 8 days later) and disappeared or persisted (Figure 2A). This procedure greatly shortly later (4 days after the first injection) a similar amount of reduced tumor growth (Figure 2B) and extended the longevity MCA205 cells into the left flank to create a secondary tumor. of the mice, yielding a survival rate of ~22% at 60 days post- Only the primary tumor was injected for up to 4 days until tumor diagnosis (Figure 2C). At this point, tumor-free mice were rechal- disappearance with LTX-401 (or PBS as a vehicle control) and the lenged by s.c. inoculation of the same tumor type (MCA205, growth of both the primary (treated) and secondary (untreated) injected into the opposite flank) or antigenically unrelated mam- tumors were monitored. The local treatment had a significant mary carcinoma AT3 cells (injected into the same flank, though at tumor growth reducing effect on primary tumors (Figure 3A,B, a site distant from the primary MCA205 fibrosarcoma). D) yet only limited effects on secondary contralateral tumors Importantly, all mice that were cured by LTX-401 treatment (Figure 3C,E). Hence, LTX-401-mediated oncolysis alone is not from their primary MCA205 tumor became resistant to rechal- sufficient to induce a therapeutically relevant abscopal effect. lenge with MCA205 (but not AT3) cancer cells (Figure 2D),

a Evaluation of adverse effects Day -8 Day 0 Day 3

Tumor cell Intratumoral injection Further experimentation inocculation until tumor disappearance and analysis

b n=36 n=111 n=52 100 n=63 80 n=68 60 40 %survival 20 0 Vehicle LTX- LTX- DTT- DTT- control 315 401 205 304

Figure 1. Absence of adverse toxicity in LTX-401-treated mice. In a retrospective study, the adverse toxicity of sequential intratumoral injections of oncolytic peptides and peptidomimetics at they are respective therapeutically effective dose such as LTX-315 (0.3 mg/injection), LTX-401 (0.25 mg/injection), DTT-205 and DTT-304 (both at 1.5 mg/injection) was evaluated. Note that these doses have been optimized to obtain an optimal antitumor effect. Data from in vivo experimentation utilizing oncolytic compounds in C57BL/6 mice bearing palpable subcutaneous solid tumors were evaluated and global survival was enumerated (A). The percentage of animals that survived without severe discomfort requiring their euthanasia is indicated (B). ONCOIMMUNOLOGY e1594555-3

a Day -8 Day 0 Day 3 Day 80

sc it sc MCA205 or TC1 LTX-401 (until tumor disappearance) Rechallenge

b e 300 MCA205 PBS (n=12) 300 TC1 PBS (n=5) ) 2 )

LTX-401 (n=12) 2 LTX-401 (n=12) 200 200

100 100 Tumor Tumor

Tumor size (mm Tumor clearance n=3 0 clearance n=8

Tumor size (mm Tumor 0 0 20406080 0204060 Days post treatment Days post treatment c MCA205 f TC1 100 PBS (n=12) 100 LTX-401 (n=12)

*** 50 50 % survival ** % survival PBS (n=5) LTX-401 (n=12) 0 0 0 204060 0204060 Days post treatment Days post treatment d Rechallenge after cure from MCA205 (n=4) g Rechallenge after cure from TC1 (n=5) 200 MCA205 ) 200 TC-1 ) 2 2 AT3 MCA-205 Tumor Tumor immunity n=4 immunity n=5 100 100 Tumor size (mm Tumor Tumor size (mm Tumor 0 0 0204060 0 1020304050 Days post treatment Days post treatment

Figure 2. In vivo activity of LTX-401 on subcutaneous MCA205 fibrosarcoma in immunocompetent animals. Mouse fibrosarcoma MCA205 cells were injected subcutaneously in syngeneic C57BL/6 mice and palpable tumors arising thereof were treated with sequential intratumoral injections of 0.25 mg LTX-401 as indicated in (A). LTX-401 induced efficient oncolysis an effect that is reflected in reduced tumor growth (B), and increased overall survival (C) (Chi2 test, **p < 0.01, n = 12). Rechallenge of animals cured from MCA205 fibrosarcoma (pooled from several experiments) with MCA205 several weeks after the initial therapy on the contralateral flank and challenged with syngeneic mouse AT3 breast cancer cells on the ipsilateral side resulted in efficient rejection of MCA205 but aggressive tumor growth of AT3 (D). Thus, LTX-401 caused the generation of immunological memory that sufficed in rejecting isogenic tumors. Similar effects were obtained when mouse lung cancer TC-1 cells were inoculated subcutaneously in syngeneic C57BL/6 mice and tumors and were treated when palpable with repeated injections of 0.25 mg LTX- 401 intratumorally as shown in (A). LTX-401 induced efficient oncolysis and tumor control in some animals reflected in reduced tumor growth and cure of some animals (E), and increased overall survival (F) (Chi2 test, ***p < 0.001). Some of the animals cured from subcutaneous TC-1 lung cancers were rechallenged with TC-1 several weeks after the initial therapy on the contralateral and challenged with syngeneic mouse MCA205 fibrosarcoma cells on the ipsilateral side. This maneuver resulted in efficient rejection of TC-1 but aggressive tumor growth of MCA205 (G). Thus, also in this model system LTX-401 caused the generation of immunological memory in cured animals that sufficed in rejecting isogenic tumors.

Combinatorial effects of LTX-401 and anti-CTLA4- we investigated the possibility that this agent would sensitize mediated immune checkpoint blockade cancers to therapy with immune checkpoint blockers target- ing CTLA-4 or PD-1. For this, established MCA205 fibrosar- Given the capacity of local LTX-401 therapy to induce mem- comas were injected only once intratumorally and then ory immune responses and abscopal effects on distant lesions, e1594555-4 W. XIE ET AL.

a MCA205 or TC1 (contralateral) sc Day -8 Day -4 Day 0 Day 3

sc it MCA205 or TC1 LTX-401 (until tumor disappearance) Treated MCA205 tumors Contralateral MCA205 tumors bcPBS (n=12) PBS (n=12) 300 ) ) 300 2 LTX-401 (n=8) 2 LTX-401 (n=8)

200 200

100 100 Tumor

Tumor size (mm Tumor clearance n=5 size (mm Tumor 0 0 0 1020304050 0 1020304050 Days post treatment Days post treatment deTreated TC1 tumors Contralateral TC1 tumors 300 PBS (n=6) 300 PBS (n=6) ) ) 2 LTX-401 (n=10) 2 LTX-401 (n=10) 200 200

100 100 Tumor Tumor Tumor size (mm Tumor clearance n=8 size (mm Tumor clearance n=3 0 0 0204060 0204060 Days post treatment Days post treatment

Figure 3. LTX-401 treatment of subcutaneous MCA205 fibrosarcoma induces limited abscopal effects. C57BL/6 mice were inoculated with murine fibrosarcoma MCA205 subcutaneously on one side and four days later on the contralateral flank. Palpable primary tumors arising from the first injection were treated with sequential intratumoral injections of 0.25 mg LTX-401 as indicated in (A). LTX-401 induced efficient oncolysis in primary tumors which is reflected in reduced tumor growth and tumor clearance in some animals (B). In parallel tumors arising from untreated secondary lesions were monitored and abscopal effects were documented as tumor growth in separate curves (C). Lung cancer TC-1 cells were injected subcutaneously into C57BL/6 mice on one flank and four days later on the contralateral side. Palpable primary tumors arising from the first injection were treated with sequential injections of 0.25 mg LTX-401 intratumorally as indicated in (A). LTX-401 induced efficient oncolysis in primary tumors which is reflected in reduced tumor growth and tumor clearance in some animals (D). In parallel tumors arising from untreated secondary lesions were monitored and abscopal effects were documented as tumor growth curves (E).

subjected to immunotherapy with antibodies specific for immunotherapy had established a specific anticancer immune CTLA-4, PD-1 or a combination of both (Figure 4A). response. Sequential treatments of LTX-401 in combination Continuous tumor monitoring led to the conclusion that the with dual checkpoint blockade increased this effect and was most efficient therapeutic regimen consisted in a combination able to achieve 100% clearance of treated tumors (Figure 5A, of all three anticancer agents (LTX-401, αCTLA-4 and αPD- B,D). Of note in six out of nine animals the combination of 1). In contrast, single-agent therapies appeared to be relatively LTX-401 with dual checkpoint blockade not only cleared the inefficient in this setting (Figure 4B–F). When the survival of treated tumor but also exhibited abscopal neoplastic effects on tumor-bearing animals was monitored, the combination of distant tumors (Figure 5C,E). LTX-401 and dual checkpoint blockade (αCTLA-4 plus αPD-1) or the combination of LTX-401 and αCTLA-4 (but Concluding remarks not αPD-1) alone were able to significantly extend life expec- tancy (Figure 4G–H). Again, mice that had been rendered The results of this study suggest that LTX-401 can provide tumor-free for more than 50 days resisted rechallenge with oncolysis with a relatively favorable safety profile when the the same cancer cell type from which they had been cured agent is injected into established tumors. Indeed, LTX-401 (MCA205), yet readily developed TC-1 cancers (Figure 4I). could be injected daily on 4 subsequent days with zero fatalities. Thus, as to be expected, mice that had been cured by Oncolysis by LTX-401 yielded a rapid macroscopic response and a combination of local oncolysis and systemic allowed most of the tumors to become undetectable following ONCOIMMUNOLOGY e1594555-5

a Day -8 Day 0 Day 6 Day 9 Day 12

sc ip MCA205 LTX-401 Immune checkpoint blockade

b 300 PBS (n=7) c 300 α-PD-1 (n=7) ) ) 2 LTX-401 (n=7) 2 LTX-401+α-PD-1 (n=6) 200 200

100 100 Tumor Tumor

Tumor size (mm Tumor clearance n=1 size (mm Tumor clearance n=1 0 0 0 1020304050 0 1020304050 deDays post treatment Days post treatment 300 αCTLA-4 (n=7) 300 αPD-1/CTLA-4 (n=7) ) ) 2 LTX-401+αCTLA-4 (n=7) 2 LTX-401+αPD-1/CTLA-4 (n=7) 200 200

100 100 Tumor Tumor clearance n=2 clearance n=4 Tumor size (mm Tumor Tumor size (mm Tumor 0 0 0 1020304050 0 1020304050 Days post treatment Days post treatment f g 100 PBS 100 αPD-1/CTLA-4 LTX-401 LTX-401+ αPD-1 αPD-1/CTLA-4 LTX-401+ αPD-1 50 50 ** % survival %survival

0 0 0210030 40 50 0210030 40 50 Days post treatment Days post treatment hiRechallenge after cure from MCA205 100 α 300 MCA205 (n=7)

CTLA-4 ) LTX-401+ 2 TC-1 (n=7) αCTLA-4 200 Tumor immunity n=7 50 ** 100 % survival

0 size (mm Tumor 0 0210030 40 50 0 10203040 Days post treatment Days post treatment

Figure 4. Combination of single LTX-401 treatment with immune checkpoint blockade increases its efficacy. C57BL/6 mice subcutaneously implanted with MCA205 fibrosarcoma were injected intratumorally when tumors became palpable with a single injection of 0.25 mg LTX-401 as indicated in (A). Immune checkpoint blockade was mounted by sequential injections of monoclonal antibodies targeting CTLA-4 or PD-1 alone or in combination at day 6, 9 and 12 post treatment. LTX-401 induced efficient oncolysis which is reflected in reduced tumor growth and tumor clearance in some animals. This effect was increased in combination with CTLA-4 and PD-1cotreatmentas depicted in separate tumor growth curves (B-E) and overall survival plots (Chi2 test, **p < 0.01) (F-H). Rechallenge of animals cured from MCA205 fibrosarcoma with MCA205 several weeks after the initial therapy on the contralateral and challenged with syngeneic mouse TC-1 lung cancer cells on the ipsilateral side resulted in efficient rejection of MCA205 but aggressive tumor growth of TC-1 (I). Thus, LTX-401 caused the generation of immunological memory that led to the rejection of isogenic tumors. several local injections. This procedure likely does not simply induction of immunogenic cell death or by the liberation of cause necrotic (and passive) lysis of neoplastic and stromal cells. tumor-associated antigens from their tumor towards the draining Rather, it triggered some immunogenic events, be it the lymph nodes.14 Irrespective of the exact molecular mechanisms e1594555-6 W. XIE ET AL.

MCA205 (contralateral) a Day -8 Day -4 Day 0 Day 3 Day 6 Day 9 Day 12 sc it ip MCA205 LTX-401 Immune checkpoint blockade

Treated MCA205 tumors Contralateral MCA205 tumors bcPBS (n=8) PBS (n=8) 300 αPD-1/CTLA-4 (n=9) 300 αPD-1/CTLA-4 (n=9) ) ) 2 2

200 200

100 100 Tumor

Tumor size (mm Tumor size (mm Tumor clearance n=1 0 0 0 1020304050 0 1020304050 Days post treatment Days post treatment

PBS (n=8) PBS (n=8) deα α 300 LTX-401+ PD-1/CTLA-4 (n=9) 300 LTX-401+ PD-1/CTLA-4 (n=9) ) ) 2 2

200 200

100 100 Tumor Tumor Tumor size (mm Tumor Tumor size (mm Tumor clearance n=9 clearance n=6 0 0 0 1020304050 0 1020304050 Days post treatment Days post treatment

Figure 5. Sequential LTX-401 treatment with double immune checkpoint blockade exhibits systemic antitumor immunity C57BL/6 mice were inoculated with murine fibrosarcoma MCA205 subcutaneously on one side and four days later on the contralateral flank. Palpable primary tumors arising from the first injection were treated with sequential injections of 0.25 mg LTX-401 as indicated in (A). Immune checkpoint blockade was mounted by sequential injections of double immune checkpoint blockade with monoclonal antibodies targeting CTLA-4 or PD-1 at day 6, 9 and 12 post treatment. Immune checkpoint inhibition alone did not control tumor growth of LTX-401-treated primary (B) and distant (C) tumors. In contrast, combination of sequential LTX-401 treatment with double immune checkpoint blockade cleared 100% of LTX-401-treated primary tumors (D) and depicted abscopal effects on distant tumors (E). accounting for these effects, there are at least three lines of Materials & methods evidence pleading in favor of LTX-401-triggered cancer-specific Chemicals, cell culture media and supplements immunogenicity. First, animalsthathadbeencuredfromtheir established cancers by LTX-401 became resistance to rechallenge All supplements and cell culture media were purchased from with the same cancer type. Second, one single intratumoral Thermo Fisher Scientific (Carlsbad, CA, US). LTX-401 was pro- injection of LTX-401 (which turned out to be inefficient on its vided by Lytix Biopharma (Oslo, Norway). Cell culture support own in reducing tumor volume) sensitized the cancers to sub- and consumables were obtained from Greiner Bio-One (Monroe, sequent immunotherapy with antibodies blocking CTLA-4 or, CA, US). Mouse fibrosarcoma cells MCA-205 and murine lung more so, a combination of antibodies targeting CTLA-4 and PD- cancer TC-1 cells were cultured in Glutamax®-containing DMEM 1. Third, in mice bearing two neoplastic lesions, one which was medium supplemented with 10% fetal bovine serum (FBS), and treated and another one that was left untreated, the combination 10 mM HEPES. Cells were maintained in a humidified incubator of LTX-401 with double immune checkpoint blockade did not at 37°C with an atmosphere containing 5% CO2. only reduce the growth of the locally injected tumor but also mediated long-distance (abscopal) effects resulting in the control of most of the non-injected cancers. In vivo experimentation Altogether, these results convincingly demonstrate that Female wild-type C57BL/6 mice at the age of 6–8 weeks were LTX-401 mediated oncolysis may advantageously combine obtained from Harlan France (Gannat, France) and kept with established immunotherapeutic regimens. under controlled conditions in the animal facility at the ONCOIMMUNOLOGY e1594555-7

Gustave Roussy Campus Cancer in specific pathogen-free and combination at day 6, 9 and 12 post treatment. Animals temperature-controlled environment with 12 h day, 12 h night were monitored and tumor growth documented regularly cycles and received food and water ad libitum. Animal experi- until end-points were reached or signs of obvious discomfort ments were conducted in compliance with the EU Directive were observed. Statistical analysis was performed employing 63/2010, and protocols 2013_094A and were approved by the two-way ANOVA analysis followed by Bonferroni’s test com- Ethical Committee of the Gustave Roussy Campus Cancer paring to control conditions (* p < 0.05, ** p < 0.01 and ***p < (CEEA IRCIV/IGR no. 26, registered at the French Ministry 0.001). of Research). MCA205 or TC-1 tumors were established in C57BL/6 hosts by subcutaneously inoculating 5 × 105 cells. Disclosure of potential conflicts of interest When tumors became palpable, either 0.25 mg (for each injection of sequential treatment) or 0.4 mg (for single injec- No potential conflicts of interest were disclosed. tion) of LTX-401 was injected intratumorally. Remaining tumor tissue was treated on subsequent days accordingly Funding and animal well-being and tumor growth were monitored. Mice were sacrificed when tumor size reached end-point or WX,HZ and YW were supported by the China Scholarship Council, GK signs of obvious discomfort associated with the treatment and LZ are supported by the Ligue contre le Cancer (équipes labelisées); Agence National de la Recherche (ANR) – Projets blancs; ANR under the were observed always following the EU Directive 63/2010 frame of E-Rare-2, the ERA-Net for Research on Rare Diseases; and our Ethical Committee advice. Surviving and tumor-free Association pour la recherche sur le cancer (ARC); Cancéropôle Ile-de- animals were analyzed and kept for more than 30 days before France; Institut National du Cancer (INCa); Fondation Bettencourt- subcutaneous rechallenge with 5 × 105 live TC-1 in one flank Schueller; Fondation de France; Fondation pour la Recherche Médicale and 5 × 105 live MCA205 cells injected in the contralateral (FRM); the European Commission (ArtForce); the European Research Council (ERC); the LabEx Immuno-Oncology; the SIRIC Stratified flank. In the case of previously injected TC-1 tumor-free Oncology Cell DNA Repair and Tumor Immune Elimination animals, the location of the injected cells was inverted. (SOCRATE); the SIRIC Cancer Research and Personalized Medicine Animals were monitored and tumor growth documented reg- (CARPEM); the Seerave foundation; the Swiss Bridge Foundation, ularly until end-points were reached or signs of obvious dis- ISREC and the Paris Alliance of Cancer Research Institutes (PACRI). comfort were observed. Statistical analysis was performed This project was supported by the Norwegian Research Council (254800, 257967). employing two-way ANOVA analysis followed by Bonferroni’s test comparing to control conditions (* p < 0.05, ** p < 0.01 and ***p < 0.001). Abscopal effects were Abbreviations analyzed by inoculating C57BL/6 hosts subcutaneously with 5 5 ×10 MCA205 or TC-1 cells. Four days later the same DAMP danger-associated molecular pattern number of cells was inoculated on the contralateral side to ICD immunogenic cell death establish a second syngeneic tumor. Following exclusively primary tumors were treated with the oncolytic LTX-401 as ORCID described above and tumor sizes of both primary and second- ary tumor were monitored and documented regularly until Takahiro Yamazaki http://orcid.org/0000-0002-7420-4394 end-points were reached or signs of obvious discomfort were Laurence Zitvogel http://orcid.org/0000-0003-1596-0998 Øystein Rekdal http://orcid.org/0000-0001-5563-6709 observed. Oliver Kepp http://orcid.org/0000-0002-6081-9558 Guido Kroemer http://orcid.org/0000-0002-9334-4405 Retrospective study The adverse toxicity of sequential intratumoral injections of References oncolytic peptides and peptidomimetics at their respective 1. Fend L, Yamazaki T, Remy C, Fahrner C, Gantzer M, Nourtier V, therapeutically effective dose such as LTX-315 (0.3 mg/injec- Preville X, Quemeneur E, Kepp O, Adam J, et al. Immune check- tion), LTX-401 (0.25 mg/injection), DTT-205 and DTT-304 point blockade, immunogenic chemotherapy or IFN-alpha block- (both at 1.5 mg/injection) was evaluated. Data from in vivo ade boost the local and abscopal effects of oncolytic virotherapy. Cancer Res. 2017;77(15):4146–4157. doi:10.1158/0008-5472.CAN- experimentation utilizing oncolytic compounds in C57BL/6 16-2165. mice bearing palpable subcutaneous solid tumors were col- 2. Bommareddy PK, Shettigar M, Kaufman HL. Integrating oncoly- lected from these two studies,11,8 and unpublished data. tic viruses in combination cancer immunotherapy. Nat Rev Animals that depicted adverse toxic effects were enumerated Immunol. 2018;18(8):498–513. doi:10.1038/s41577-018-0014-6. and the percentage of mice that depicted an onset of discom- 3. Zhou H, Forveille S, Sauvat A, Yamazaki T, Senovilla L, Ma Y, Liu P, Yang H, Bezu L, Muller K, et al. The oncolytic peptide fort upon treatment was plotted. LTX-315 triggers immunogenic cell death. Cell Death Dis. 2016;7: e2134. doi:10.1038/cddis.2016.47. 4. Zhou H, Forveille S, Sauvat A, Sica V, Izzo V, Durand S, Muller K, Immune checkpoint blockade Liu P, Zitvogel L, Rekdal O, et al. The oncolytic peptide LTX-315 Immune checkpoint blockade was mounted by sequential kills cancer cells through Bax/Bak-regulated mitochondrial mem- brane permeabilization. Oncotarget. 2015;6(29):26599–26614. intraperitoneal injections of 200 µg monoclonal antibodies doi:10.18632/oncotarget.5613. targeting PD-1 (Clone 29F.1A12, BioXcell, West Lebanon, 5. Gomes-da-Silva LC, Jimenez AJ, Sauvat A, Xie W, Souquere S, NH, USA) or CTLA-4 (Clone 9D9, BioXcell) alone or in Divoux S, Storch M, Sveinbjornsson B, Rekdal O, Arnaut LG, e1594555-8 W. XIE ET AL.

et al. Recruitment of LC3 to damaged Golgi apparatus. Cell Death 10. Sveinbjornsson B, Camilio KA, Haug BE, Rekdal O. LTX-315: a Differ. 2018. doi:10.1038/s41418-018-0221-5. first-in-class oncolytic peptide that reprograms the tumor 6. Gomes-da-Silva LC, Zhao L, Bezu L, Zhou H, Sauvat A, Liu P, microenvironment. Future Med Chem. 2017;9(12):1339–1344. Durand S, Leduc M, Souquere S, Loos F, et al. Photodynamic doi:10.4155/fmc-2017-0088. therapy with redaporfin targets the endoplasmic reticulum and 11. Yamazaki T, Pitt JM, Vetizou M, Marabelle A, Flores C, Rekdal O, Golgi apparatus. Embo J. 2018;37(13). doi:10.15252/ Kroemer G, Zitvogel L. The oncolytic peptide LTX-315 overcomes embj.201798354. resistance of cancers to immunotherapy with CTLA4 checkpoint 7. Zhou H, Sauvat A, Gomes-da-Silva LC, Durand S, Forveille S, blockade. Cell Death Differ. 2016;23(6):1004–1015. doi:10.1038/ Iribarren K, Yamazaki T, Souquere S, Bezu L, Muller K, et al. The cdd.2016.35. oncolytic compound LTX-401 targets the Golgi apparatus. Cell 12. Camilio KA, Rekdal O, Sveinbjornsson B. LTX-315 (Oncopore): Death Differ. 2016;23(12):2031–2041. doi:10.1038/cdd.2016.86. a short synthetic anticancer peptide and novel immunotherapeu- 8. Zhou H, Mondragon L, Xie W, Mauseth B, Leduc M, Sauvat A, tic agent. Oncoimmunology. 2014;3:e29181. doi:10.4161/ Gomes-da-Silva LC, Forveille S, Iribarren K, Souquere S, et al. onci.29181. Oncolysis with DTT-205 and DTT-304 generates immunological 13. Eike LM, Mauseth B, Camilio KA, Rekdal O, Sveinbjornsson B. memory in cured animals. Cell Death Dis. 2018;9(11):1086. The cytolytic amphipathic beta(2,2)-Amino acid LTX-401 Induces doi:10.1038/s41419-018-1127-3. DAMP release in melanoma cells and causes complete regression 9. Nestvold J, Wang MY, Camilio KA, Zinocker S, Tjelle TE, of B16 Melanoma. PLoS One. 2016;11(2):e0148980. doi:10.1371/ Lindberg A, Haug BE, Kvalheim G, Sveinbjornsson B, Rekdal O. journal.pone.0148980. Oncolytic peptide LTX-315 induces an immune-mediated absco- 14. Galluzzi L, Buque A, Kepp O, Zitvogel L, Kroemer G. pal effect in a rat sarcoma model. Oncoimmunology. 2017;6(8): Immunogenic cell death in cancer and infectious disease. Nat e1338236. doi:10.1080/2162402X.2017.1338236. Rev Immunol. 2017;17(2):97–111. doi:10.1038/nri.2016.107.

CONCLUDING REMARKS

The main conclusions are:

I. Lurbinectedin triggers immunogenic cell death and induce anticancer immunity

II. Pretreatment with Lurbinectedin sensitize cancer to subsequent αPD-1/αCTLA-4 double immune checkpoint blockade

III. Local injection of LTX-401 generates anticancer immunity

IV. Combination of sequential LTX-401 treatment and αPD-1/αCTLA-4 double immune checkpoint blockade exhibits striking antineoplastic effects on both local and abscopal tumours

Perspectives In the first part of this thesis, we investigated the immunogenicity of Lurbinectedin and defined it as a bona fide ICD inducer. In vivo results indicate Lurbinectedin may synergizes clinically approved immunotherapy regimens in cancer treatment. In order to maximize the beneficial effects of lurbinectedin, following issues need to be addressed:

I. Why the vaccination effect of lurbinectedin-killed MCA205 cancer cell is limited compared to other classical ICD inducers? Does this phenomenon exist in other cancer types? To optimise lurbinectedin mediated protecting effects, different drug concentrations and time points should be tested to find a proper condition. Given that the sensitivity of cancer cells to lurbinectedin may varies depends on genetic background, tumour heterogeneity or other factors, it’s also an important issue to identify drug sensitive cancer types before clinical usage.

42

II. Does all of transcription inhibitors induce ICD? If negative, what determines the immunogenicity of transcription inhibitors? A newly published paper shows that transcription inhibition may serve as an initial event for ICD induction (Humeau et al., 2020), but whether this stress pathway in sufficient to trigger ICD is not defined yet. To answer this question, the immunogenicity of transcription inhibitors needs to be systemically evaluated. By taking advantage of the ICD predication system (in silicon), fluorescent biosensor-based platform (in vitro) (Kepp et al., 2019), high-throughput screening could be launched to identify ICD candidates in this cluster. Further validated by vaccination assay (in vivo). Conclusions can be made only by this means.

III. Why lurbinectedin sensitize MCA205 cancer to ICBs but the synergistic effect is not significant in spontaneous breast cancer? This could be further addressed by tumour infiltrated lymphocytes analysis. Maybe the immunostimulating effect of lurbinectedin induced cancer cell death was unable to neutralise the immunosuppressive microenvironment in this circumstance.

IV. Perspective on clinical application of lurbinectedin-immunotherapy combination in cancer therapy. As a newly approved drug in relapsed SCLC treatment, the potential benefits of lurbinectedin- immunotherapy combination regiments are under evaluation in 3 different clinical trials. Lurbinectedin (PM01183) will combine with Atezolizumab (anti-PD-L1), or Pembrolizumab (anti-PD-1) or Nivolumab (anti-PD-1) plus Ipilimumab (anti-CTLA-4) to treat SCLC in NCT04253145, NCT04358237 and NCT04610658 respectively. Except SCLC, the application of lurbinectedin in breast cancer, sarcoma, and other advanced solid tumours treatment are still in clinical trials. In our ongoing study, we demonstrated that lurbinectedin not only facilitates cytotoxic T cell infiltration, but also relives the immunosuppressive nature of tumour microenvironment by depleting macrophage. As CD8 T cell is regarded as a positive prognostic factor while macrophage exhibits opposite function in multiple solid cancers, lurbenectedin shows great therapeutic potential in these cancer types. Besides, an induction of PD-1 and CTLA-4 expression was observed in tumour infiltrate CD8 T cell, which theoretically supported the

43

hypothesis that immune checkpoint blockade synergize with lurbinectedin in different types of cancer (unpublished data).

In the second topic, we demonstrated that oncolysis by LTX-401 is highly immunogenic and elicits systemically anticancer response in combination with ICBs. Also, some questions need be answered to extend its future application. I. Why LTX-401 preferably synergize αCTLA-4 rather than αPD-1 in this tumour model? It’s reported that within oncolytic peptide LTX-315 treated tumour, CTLA-4 molecule was unregulated while PD-1 expression was decreased on CD8+ TILs. This may explain the different outcomes in these two different combinatorial regimens. Maybe LTX-401 shares the same property with LTX-315 in modulating tumour microenvironment, TIL analysis may give some clues in this question. II. How to achieve suitable and sustainable concentration of LTX-401 intratumorally? This is a major obstacle for clinical application of oncolytic compounds, given its unspecific oncolysis property, local administration seems to be the only delivery technique. The best solution for this problem is to identify new oncolytic drugs or coating biomaterials which minimize the injection times but reach the same therapeutic effects. This improvement may benefit beyond superficial cancer patients.

44

ACKNOWLEDGEMENT

I would like to give biggest thanks to all of you who helps me, supports me and companies me in the past years. I cannot enjoy a so inspiring and fantastic doctoral journey without your presence.

To my supervisor Guido, thanks for your kind supervision during my Ph.D. study, your insightful perspectives and passions on science always inspire me to think big scientific questions.

To my co-supervisor Oliver, thanks for welcoming to the lab. You are always reliable and available whenever I need guidance or help, your supports and encouragements mean a lot to me.

To Dr. Éric Tartour and Dr. François Ghiringhelli, thanks for joining my jury as rapporteurs.

To Dr. Donnadieu Emmanuel, thanks for serving as the president of the jury and examine my study.

To Dr. Catherine Brenner and Dr. Sébastien Apcher, thanks for accepting my invitation to examine this work.

To Guo, Yan and Shuai, thanks for helping me in finalizing experiments and your useful advices.

To Peng, Liwei and Chen, thanks for the happiness you brought to me, I do enjoy the time we spent together both in and out of the lab.

To Giulia and Juliette H, thanks for your patience and kindness, I cannot adapted to the French life smoothly without your help.

To all the members in our lab, I do appreciate your generous help and advices in my Ph.D projects and future careers.

To Chinese Scholarship Council, thanks for financial support during my study.

To my beloved parents and little sister, I cannot appreciate more for your consistent love and support, you give me the courage to explore, to extend my limitation without worry.

To my girlfriend Yanli, thanks for your unswerving love, your company, your tolerance, your comfort and inspiration. You lightened my life and make me better and better. Love you so much.

45

REFERENCES

Al-Benna, S., Shai, Y., Jacobsen, F., and Steinstraesser, L. (2011). Oncolytic activities of host defense peptides. International journal of molecular sciences 12, 8027-8051.

Antonia, S.J., Villegas, A., Daniel, D., Vicente, D., Murakami, S., Hui, R., Kurata, T., Chiappori, A., Lee, K.H., de Wit, M., et al. (2018). Overall Survival with Durvalumab after Chemoradiotherapy in Stage III NSCLC. The New England journal of medicine 379, 2342-2350.

Apetoh, L., Ghiringhelli, F., Tesniere, A., Obeid, M., Ortiz, C., Criollo, A., Mignot, G., Maiuri, M.C., Ullrich, E., Saulnier, P., et al. (2007). Toll-like receptor 4-dependent contribution of the immune system to anticancer chemotherapy and radiotherapy. Nature medicine 13, 1050-1059.

Asadzadeh, Z., Safarzadeh, E., Safaei, S., Baradaran, A., Mohammadi, A., Hajiasgharzadeh, K., Derakhshani, A., Argentiero, A., Silvestris, N., and Baradaran, B. (2020). Current Approaches for Combination Therapy of Cancer: The Role of Immunogenic Cell Death. Cancers 12.

Bailey, D., and O'Hare, P. (2007). Transmembrane bZIP transcription factors in ER stress signaling and the unfolded protein response. Antioxidants & redox signaling 9, 2305-2321.

Baracco, E.E., Stoll, G., Van Endert, P., Zitvogel, L., Vacchelli, E., and Kroemer, G. (2019). Contribution of annexin A1 to anticancer immunosurveillance. Oncoimmunology 8, e1647760.

Belgiovine, C., Bello, E., Liguori, M., Craparotta, I., Mannarino, L., Paracchini, L., Beltrame, L., Marchini, S., Galmarini, C.M., Mantovani, A., et al. (2017). Lurbinectedin reduces tumour- associated macrophages and the inflammatory tumour microenvironment in preclinical models. British journal of cancer 117, 628-638.

Bell, C.W., Jiang, W., Reich, C.F., 3rd, and Pisetsky, D.S. (2006). The extracellular release of HMGB1 during apoptotic cell death. American journal of physiology Cell physiology 291, C1318- 1325.

46

Bergmann, J.F., Chassany, O., Geneve, J., Abiteboul, M., Caulin, C., and Segrestaa, J.M. (1992). Endoscopic evaluation of the effect of ketoprofen, ibuprofen and aspirin on the gastroduodenal mucosa. European journal of clinical pharmacology 42, 685-687.

Bettigole, S.E., and Glimcher, L.H. (2015). Endoplasmic reticulum stress in immunity. Annual review of immunology 33, 107-138.

Bezu, L., Sauvat, A., Humeau, J., Gomes-da-Silva, L.C., Iribarren, K., Forveille, S., Garcia, P., Zhao, L., Liu, P., Zitvogel, L., et al. (2018a). eIF2alpha phosphorylation is pathognomonic for immunogenic cell death. Cell death and differentiation 25, 1375-1393.

Bezu, L., Sauvat, A., Humeau, J., Leduc, M., Kepp, O., and Kroemer, G. (2018b). eIF2alpha phosphorylation: A hallmark of immunogenic cell death. Oncoimmunology 7, e1431089.

Birkeland, S.A., Storm, H.H., Lamm, L.U., Barlow, L., Blohme, I., Forsberg, B., Eklund, B., Fjeldborg, O., Friedberg, M., Frodin, L., et al. (1995). Cancer risk after renal transplantation in the Nordic countries, 1964-1986. International journal of cancer 60, 183-189.

Bueren-Calabuig, J.A., Giraudon, C., Galmarini, C.M., Egly, J.M., and Gago, F. (2011). Temperature-induced melting of double-stranded DNA in the absence and presence of covalently bonded antitumour drugs: insight from molecular dynamics simulations. Nucleic acids research 39, 8248-8257.

Burnet, F.M. (1970). The concept of immunological surveillance. Progress in experimental tumor research 13, 1-27.

Burnet, M. (1957). Cancer: a biological approach. III. Viruses associated with neoplastic conditions. IV. Practical applications. British medical journal 1, 841-847.

Calvo, E., Moreno, V., Flynn, M., Holgado, E., Olmedo, M.E., Lopez Criado, M.P., Kahatt, C., Lopez-Vilarino, J.A., Siguero, M., Fernandez-Teruel, C., et al. (2017). Antitumor activity of lurbinectedin (PM01183) and doxorubicin in relapsed small-cell lung cancer: results from a phase I study. Annals of oncology : official journal of the European Society for Medical Oncology 28, 2559-2566.

47

Casares, N., Pequignot, M.O., Tesniere, A., Ghiringhelli, F., Roux, S., Chaput, N., Schmitt, E., Hamai, A., Hervas-Stubbs, S., Obeid, M., et al. (2005). Caspase-dependent immunogenicity of doxorubicin-induced tumor cell death. The Journal of experimental medicine 202, 1691-1701.

Cespedes, M.V., Guillen, M.J., Lopez-Casas, P.P., Sarno, F., Gallardo, A., Alamo, P., Cuevas, C., Hidalgo, M., Galmarini, C.M., Allavena, P., et al. (2016). Lurbinectedin induces depletion of tumor-associated macrophages, an essential component of its in vivo synergism with gemcitabine, in pancreatic adenocarcinoma mouse models. Disease models & mechanisms 9, 1461-1471.

Chalmin, F., Mignot, G., Bruchard, M., Chevriaux, A., Vegran, F., Hichami, A., Ladoire, S., Derangere, V., Vincent, J., Masson, D., et al. (2012). Stat3 and Gfi-1 transcription factors control Th17 cell immunosuppressive activity via the regulation of ectonucleotidase expression. Immunity 36, 362-373.

Chao, M.P., Majeti, R., and Weissman, I.L. (2011). Programmed cell removal: a new obstacle in the road to developing cancer. Nature reviews Cancer 12, 58-67.

Chen, R., Hou, W., Zhang, Q., Kang, R., Fan, X.G., and Tang, D. (2013). Emerging role of high- mobility group box 1 (HMGB1) in liver diseases. Molecular medicine 19, 357-366.

Choi, J.H., Cho, Y.S., Ko, Y.H., Hong, S.U., Park, J.H., and Lee, M.A. (2014). Absence of autophagy-related proteins expression is associated with poor prognosis in patients with colorectal adenocarcinoma. Gastroenterology research and practice 2014, 179586.

Coca, S., Perez-Piqueras, J., Martinez, D., Colmenarejo, A., Saez, M.A., Vallejo, C., Martos, J.A., and Moreno, M. (1997). The prognostic significance of intratumoral natural killer cells in patients with colorectal carcinoma. Cancer 79, 2320-2328.

Davoodi, H., and Seow, H.F. (2011). Variant Toll-like receptor4 (Asp299Gly and Thr399Ile alleles) and Toll-like receptor2 (Arg753Gln and Arg677Trp alleles) in colorectal cancer. Iranian journal of allergy, asthma, and immunology 10, 91-99.

Demaria, S., Ng, B., Devitt, M.L., Babb, J.S., Kawashima, N., Liebes, L., and Formenti, S.C. (2004). Ionizing radiation inhibition of distant untreated tumors (abscopal effect) is immune mediated. International journal of radiation oncology, biology, physics 58, 862-870. 48

DeNardo, D.G., Brennan, D.J., Rexhepaj, E., Ruffell, B., Shiao, S.L., Madden, S.F., Gallagher, W.M., Wadhwani, N., Keil, S.D., Junaid, S.A., et al. (2011). Leukocyte complexity predicts breast cancer survival and functionally regulates response to chemotherapy. Cancer discovery 1, 54-67.

Di Virgilio, F., Sarti, A.C., Falzoni, S., De Marchi, E., and Adinolfi, E. (2018). Extracellular ATP and P2 purinergic signalling in the tumour microenvironment. Nature reviews Cancer 18, 601-618.

Dighe, A.S., Richards, E., Old, L.J., and Schreiber, R.D. (1994). Enhanced in vivo growth and resistance to rejection of tumor cells expressing dominant negative IFN gamma receptors. Immunity 1, 447-456.

Donadon, M., Hudspeth, K., Cimino, M., Di Tommaso, L., Preti, M., Tentorio, P., Roncalli, M., Mavilio, D., and Torzilli, G. (2017). Increased Infiltration of Natural Killer and T Cells in Colorectal Liver Metastases Improves Patient Overall Survival. Journal of gastrointestinal surgery : official journal of the Society for Surgery of the Alimentary Tract 21, 1226-1236.

Ehrlich, P. (1909). Ueber des jetzigen Stand der Karzinomforschung.

Eike, L.M., Mauseth, B., Camilio, K.A., Rekdal, O., and Sveinbjornsson, B. (2016). The Cytolytic Amphipathic beta(2,2)-Amino Acid LTX-401 Induces DAMP Release in Melanoma Cells and Causes Complete Regression of B16 Melanoma. PloS one 11, e0148980.

Eliassen, L.T., Berge, G., Leknessund, A., Wikman, M., Lindin, I., Lokke, C., Ponthan, F., Johnsen, J.I., Sveinbjornsson, B., Kogner, P., et al. (2006). The antimicrobial peptide, lactoferricin B, is cytotoxic to neuroblastoma cells in vitro and inhibits xenograft growth in vivo. International journal of cancer 119, 493-500.

Eliassen, L.T., Berge, G., Sveinbjornsson, B., Svendsen, J.S., Vorland, L.H., and Rekdal, O. (2002). Evidence for a direct antitumor mechanism of action of bovine lactoferricin. Anticancer research 22, 2703-2710.

Elliott, M.R., Chekeni, F.B., Trampont, P.C., Lazarowski, E.R., Kadl, A., Walk, S.F., Park, D., Woodson, R.I., Ostankovich, M., Sharma, P., et al. (2009). Nucleotides released by apoptotic cells act as a find-me signal to promote phagocytic clearance. Nature 461, 282-286.

49

Engel, A.M., Svane, I.M., Rygaard, J., and Werdelin, O. (1997). MCA sarcomas induced in scid mice are more immunogenic than MCA sarcomas induced in congenic, immunocompetent mice. Scandinavian journal of immunology 45, 463-470.

Farago, A.F., Drapkin, B.J., Lopez-Vilarino de Ramos, J.A., Galmarini, C.M., Nunez, R., Kahatt, C., and Paz-Ares, L. (2019). ATLANTIS: a Phase III study of lurbinectedin/doxorubicin versus topotecan or cyclophosphamide/doxorubicin/vincristine in patients with small-cell lung cancer who have failed one prior platinum-containing line. Future oncology 15, 231-239.

Fjell, C.D., Hiss, J.A., Hancock, R.E., and Schneider, G. (2011). Designing antimicrobial peptides: form follows function. Nature reviews Drug discovery 11, 37-51.

Formenti, S.C., Rudqvist, N.P., Golden, E., Cooper, B., Wennerberg, E., Lhuillier, C., Vanpouille- Box, C., Friedman, K., Ferrari de Andrade, L., Wucherpfennig, K.W., et al. (2018). Radiotherapy induces responses of lung cancer to CTLA-4 blockade. Nature medicine 24, 1845-1851.

Fridman, W.H., Pages, F., Sautes-Fridman, C., and Galon, J. (2012). The immune contexture in human tumours: impact on clinical outcome. Nature reviews Cancer 12, 298-306.

Fucikova, J., Becht, E., Iribarren, K., Goc, J., Remark, R., Damotte, D., Alifano, M., Devi, P., Biton, J., Germain, C., et al. (2016a). Calreticulin Expression in Human Non-Small Cell Lung Cancers Correlates with Increased Accumulation of Antitumor Immune Cells and Favorable Prognosis. Cancer research 76, 1746-1756.

Fucikova, J., Kralikova, P., Fialova, A., Brtnicky, T., Rob, L., Bartunkova, J., and Spisek, R. (2011). Human tumor cells killed by anthracyclines induce a tumor-specific immune response. Cancer research 71, 4821-4833.

Fucikova, J., Moserova, I., Truxova, I., Hermanova, I., Vancurova, I., Partlova, S., Fialova, A., Sojka, L., Cartron, P.F., Houska, M., et al. (2014). High hydrostatic pressure induces immunogenic cell death in human tumor cells. International journal of cancer 135, 1165-1177.

50

Fucikova, J., Truxova, I., Hensler, M., Becht, E., Kasikova, L., Moserova, I., Vosahlikova, S., Klouckova, J., Church, S.E., Cremer, I., et al. (2016b). Calreticulin exposure by malignant blasts correlates with robust anticancer immunity and improved clinical outcome in AML patients. Blood 128, 3113-3124.

Galluzzi, L., Buque, A., Kepp, O., Zitvogel, L., and Kroemer, G. (2017). Immunogenic cell death in cancer and infectious disease. Nature reviews Immunology 17, 97-111.

Galluzzi, L., Humeau, J., Buque, A., Zitvogel, L., and Kroemer, G. (2020a). Immunostimulation with chemotherapy in the era of immune checkpoint inhibitors. Nature reviews Clinical oncology.

Galluzzi, L., Vitale, I., Aaronson, S.A., Abrams, J.M., Adam, D., Agostinis, P., Alnemri, E.S., Altucci, L., Amelio, I., Andrews, D.W., et al. (2018). Molecular mechanisms of cell death: recommendations of the Nomenclature Committee on Cell Death 2018. Cell death and differentiation 25, 486-541.

Galluzzi, L., Vitale, I., Warren, S., Adjemian, S., Agostinis, P., Martinez, A.B., Chan, T.A., Coukos, G., Demaria, S., Deutsch, E., et al. (2020b). Consensus guidelines for the definition, detection and interpretation of immunogenic cell death. Journal for immunotherapy of cancer 8.

Gardai, S.J., McPhillips, K.A., Frasch, S.C., Janssen, W.J., Starefeldt, A., Murphy-Ullrich, J.E., Bratton, D.L., Oldenborg, P.A., Michalak, M., and Henson, P.M. (2005). Cell-surface calreticulin initiates clearance of viable or apoptotic cells through trans-activation of LRP on the phagocyte. Cell 123, 321-334.

Garg, A.D., Dudek, A.M., Ferreira, G.B., Verfaillie, T., Vandenabeele, P., Krysko, D.V., Mathieu, C., and Agostinis, P. (2013). ROS-induced autophagy in cancer cells assists in evasion from determinants of immunogenic cell death. Autophagy 9, 1292-1307.

Garg, A.D., Krysko, D.V., Vandenabeele, P., and Agostinis, P. (2012a). Hypericin-based photodynamic therapy induces surface exposure of damage-associated molecular patterns like HSP70 and calreticulin. Cancer immunology, immunotherapy : CII 61, 215-221.

51

Garg, A.D., Krysko, D.V., Verfaillie, T., Kaczmarek, A., Ferreira, G.B., Marysael, T., Rubio, N., Firczuk, M., Mathieu, C., Roebroek, A.J., et al. (2012b). A novel pathway combining calreticulin exposure and ATP secretion in immunogenic cancer cell death. The EMBO journal 31, 1062-1079.

Garg, A.D., Nowis, D., Golab, J., Vandenabeele, P., Krysko, D.V., and Agostinis, P. (2010). Immunogenic cell death, DAMPs and anticancer therapeutics: an emerging amalgamation. Biochimica et biophysica acta 1805, 53-71.

Gaspar, D., Veiga, A.S., and Castanho, M.A. (2013). From antimicrobial to anticancer peptides. A review. Frontiers in microbiology 4, 294.

Ghiringhelli, F., Apetoh, L., Tesniere, A., Aymeric, L., Ma, Y., Ortiz, C., Vermaelen, K., Panaretakis, T., Mignot, G., Ullrich, E., et al. (2009). Activation of the NLRP3 inflammasome in dendritic cells induces IL-1beta-dependent adaptive immunity against tumors. Nature medicine 15, 1170-1178.

Giglio, P., Gagliardi, M., Tumino, N., Antunes, F., Smaili, S., Cotella, D., Santoro, C., Bernardini, R., Mattei, M., Piacentini, M., et al. (2018). PKR and GCN2 stress kinases promote an ER stress- independent eIF2alpha phosphorylation responsible for calreticulin exposure in melanoma cells. Oncoimmunology 7, e1466765.

Hancock, R.E., and Sahl, H.G. (2006). Antimicrobial and host-defense peptides as new anti- infective therapeutic strategies. Nature biotechnology 24, 1551-1557.

Harding, H.P., Zhang, Y., Bertolotti, A., Zeng, H., and Ron, D. (2000). Perk is essential for translational regulation and cell survival during the unfolded protein response. Molecular cell 5, 897-904.

Harding, H.P., Zhang, Y., and Ron, D. (1999). Protein translation and folding are coupled by an endoplasmic-reticulum-resident kinase. Nature 397, 271-274.

Harding, S.M., Benci, J.L., Irianto, J., Discher, D.E., Minn, A.J., and Greenberg, R.A. (2017). Mitotic progression following DNA damage enables pattern recognition within micronuclei. Nature 548, 466-470.

52

Harlow, M.L., Maloney, N., Roland, J., Guillen Navarro, M.J., Easton, M.K., Kitchen-Goosen, S.M., Boguslawski, E.A., Madaj, Z.B., Johnson, B.K., Bowman, M.J., et al. (2016). Lurbinectedin Inactivates the Ewing Sarcoma Oncoprotein EWS-FLI1 by Redistributing It within the Nucleus. Cancer research 76, 6657-6668.

Haze, K., Yoshida, H., Yanagi, H., Yura, T., and Mori, K. (1999). Mammalian transcription factor ATF6 is synthesized as a transmembrane protein and activated by proteolysis in response to endoplasmic reticulum stress. Molecular biology of the cell 10, 3787-3799.

Hildner, K., Edelson, B.T., Purtha, W.E., Diamond, M., Matsushita, H., Kohyama, M., Calderon, B., Schraml, B.U., Unanue, E.R., Diamond, M.S., et al. (2008). Batf3 deficiency reveals a critical role for CD8alpha+ dendritic cells in cytotoxic T cell immunity. Science 322, 1097-1100.

Hsu, W.M., Hsieh, F.J., Jeng, Y.M., Kuo, M.L., Chen, C.N., Lai, D.M., Hsieh, L.J., Wang, B.T., Tsao, P.N., Lee, H., et al. (2005). Calreticulin expression in neuroblastoma--a novel independent prognostic factor. Annals of oncology : official journal of the European Society for Medical Oncology 16, 314-321.

Huang, J., Liu, K., Yu, Y., Xie, M., Kang, R., Vernon, P., Cao, L., Tang, D., and Ni, J. (2012). Targeting HMGB1-mediated autophagy as a novel therapeutic strategy for osteosarcoma. Autophagy 8, 275-277.

Humeau, J., Levesque, S., Kroemer, G., and Pol, J.G. (2019). Gold Standard Assessment of Immunogenic Cell Death in Oncological Mouse Models. Methods in molecular biology 1884, 297- 315.

Humeau, J., Sauvat, A., Cerrato, G., Xie, W., Loos, F., Iannantuoni, F., Bezu, L., Levesque, S., Paillet, J., Pol, J., et al. (2020). Inhibition of transcription by dactinomycin reveals a new characteristic of immunogenic cell stress. EMBO molecular medicine 12, e11622.

Idzko, M., Hammad, H., van Nimwegen, M., Kool, M., Willart, M.A., Muskens, F., Hoogsteden, H.C., Luttmann, W., Ferrari, D., Di Virgilio, F., et al. (2007). Extracellular ATP triggers and maintains asthmatic airway inflammation by activating dendritic cells. Nature medicine 13, 913- 919.

53

Ishigami, S., Natsugoe, S., Tokuda, K., Nakajo, A., Che, X., Iwashige, H., Aridome, K., Hokita, S., and Aikou, T. (2000). Prognostic value of intratumoral natural killer cells in gastric carcinoma. In Cancer, pp. 577-583.

Jing, J.J., Li, M., and Yuan, Y. (2012). Toll-like receptor 4 Asp299Gly and Thr399Ile polymorphisms in cancer: a meta-analysis. Gene 499, 237-242.

Kaplan, D.H., Shankaran, V., Dighe, A.S., Stockert, E., Aguet, M., Old, L.J., and Schreiber, R.D. (1998). Demonstration of an interferon gamma-dependent tumor surveillance system in immunocompetent mice. Proceedings of the National Academy of Sciences of the United States of America 95, 7556-7561.

Keizman, D., Ish-Shalom, M., Huang, P., Eisenberger, M.A., Pili, R., Hammers, H., and Carducci, M.A. (2012). The association of pre-treatment neutrophil to lymphocyte ratio with response rate, progression free survival and overall survival of patients treated with sunitinib for metastatic renal cell carcinoma. European journal of cancer 48, 202-208.

Kepp, O., Sauvat, A., Leduc, M., Forveille, S., Liu, P., Zhao, L., Bezu, L., Xie, W., Zitvogel, L., and Kroemer, G. (2019). A fluorescent biosensor-based platform for the discovery of immunogenic cancer cell death inducers. Oncoimmunology 8, 1606665.

Kepp, O., Senovilla, L., Vitale, I., Vacchelli, E., Adjemian, S., Agostinis, P., Apetoh, L., Aranda, F., Barnaba, V., Bloy, N., et al. (2014). Consensus guidelines for the detection of immunogenic cell death. Oncoimmunology 3, e955691.

Khong, H.T., and Restifo, N.P. (2002). Natural selection of tumor variants in the generation of "tumor escape" phenotypes. Nature immunology 3, 999-1005.

Kim, R., Emi, M., and Tanabe, K. (2007). Cancer immunoediting from immune surveillance to immune escape. Immunology 121, 1-14.

Klein, G. (1966). Tumor antigens. Annual review of microbiology 20, 223-252.

54

Koks, C.A., Garg, A.D., Ehrhardt, M., Riva, M., Vandenberk, L., Boon, L., De Vleeschouwer, S., Agostinis, P., Graf, N., and Van Gool, S.W. (2015). Newcastle disease virotherapy induces long- term survival and tumor-specific immune memory in orthotopic glioma through the induction of immunogenic cell death. International journal of cancer 136, E313-325.

Korbelik, M., and Dougherty, G.J. (1999). Photodynamic therapy-mediated immune response against subcutaneous mouse tumors. Cancer research 59, 1941-1946.

Korbelik, M., Zhang, W., and Merchant, S. (2011). Involvement of damage-associated molecular patterns in tumor response to photodynamic therapy: surface expression of calreticulin and high- mobility group box-1 release. Cancer immunology, immunotherapy : CII 60, 1431-1437.

Krosl, G., Korbelik, M., and Dougherty, G.J. (1995). Induction of immune cell infiltration into murine SCCVII tumour by photofrin-based photodynamic therapy. British journal of cancer 71, 549-555.

Krysko, D.V., Garg, A.D., Kaczmarek, A., Krysko, O., Agostinis, P., and Vandenabeele, P. (2012). Immunogenic cell death and DAMPs in cancer therapy. Nature reviews Cancer 12, 860-875.

Leal, J.F., Martinez-Diez, M., Garcia-Hernandez, V., Moneo, V., Domingo, A., Bueren-Calabuig, J.A., Negri, A., Gago, F., Guillen-Navarro, M.J., Aviles, P., et al. (2010). PM01183, a new DNA minor groove covalent binder with potent in vitro and in vivo anti-tumour activity. British journal of pharmacology 161, 1099-1110.

Lee, A.H., Iwakoshi, N.N., and Glimcher, L.H. (2003). XBP-1 regulates a subset of endoplasmic reticulum resident chaperone genes in the unfolded protein response. Molecular and cellular biology 23, 7448-7459.

Lim, S.H., Sun, J.M., Lee, S.H., Ahn, J.S., Park, K., and Ahn, M.J. (2016). Pembrolizumab for the treatment of non-small cell lung cancer. Expert opinion on biological therapy 16, 397-406.

Liu, P., Zhao, L., Pol, J., Levesque, S., Petrazzuolo, A., Pfirschke, C., Engblom, C., Rickelt, S., Yamazaki, T., Iribarren, K., et al. (2019). Crizotinib-induced immunogenic cell death in non-small cell lung cancer. Nature communications 10, 1486.

55

Lotze, M.T., Deisseroth, A., and Rubartelli, A. (2007). Damage associated molecular pattern molecules. Clinical immunology 124, 1-4.

Ma, Y., Adjemian, S., Mattarollo, S.R., Yamazaki, T., Aymeric, L., Yang, H., Portela Catani, J.P., Hannani, D., Duret, H., Steegh, K., et al. (2013). Anticancer chemotherapy-induced intratumoral recruitment and differentiation of antigen-presenting cells. Immunity 38, 729-741.

Ma, Y., Aymeric, L., Locher, C., Mattarollo, S.R., Delahaye, N.F., Pereira, P., Boucontet, L., Apetoh, L., Ghiringhelli, F., Casares, N., et al. (2011). Contribution of IL-17-producing gamma delta T cells to the efficacy of anticancer chemotherapy. The Journal of experimental medicine 208, 491-503.

Ma, Y., Brewer, J.W., Diehl, J.A., and Hendershot, L.M. (2002). Two distinct stress signaling pathways converge upon the CHOP promoter during the mammalian unfolded protein response. Journal of molecular biology 318, 1351-1365.

Mackenzie, K.J., Carroll, P., Martin, C.A., Murina, O., Fluteau, A., Simpson, D.J., Olova, N., Sutcliffe, H., Rainger, J.K., Leitch, A., et al. (2017). cGAS surveillance of micronuclei links genome instability to innate immunity. Nature 548, 461-465.

Mader, J.S., Salsman, J., Conrad, D.M., and Hoskin, D.W. (2005). Bovine lactoferricin selectively induces apoptosis in human leukemia and carcinoma cell lines. Molecular cancer therapeutics 4, 612-624.

Markham, A. (2020). Lurbinectedin: First Approval. Drugs 80, 1345-1353.

Martins, I., Wang, Y., Michaud, M., Ma, Y., Sukkurwala, A.Q., Shen, S., Kepp, O., Metivier, D., Galluzzi, L., Perfettini, J.L., et al. (2014). Molecular mechanisms of ATP secretion during immunogenic cell death. Cell death and differentiation 21, 79-91.

Matzinger, P. (1994). Tolerance, danger, and the extended family. Annual review of immunology 12, 991-1045.

56

Mauseth, B., Camilio, K.A., Shi, J., Hammarstrom, C.L., Rekdal, O., Sveinbjornsson, B., and Line, P.D. (2019). The Novel Oncolytic Compound LTX-401 Induces Antitumor Immune Responses in Experimental Hepatocellular Carcinoma. Molecular therapy oncolytics 14, 139-148.

Messaritakis, I., Stogiannitsi, M., Koulouridi, A., Sfakianaki, M., Voutsina, A., Sotiriou, A., Athanasakis, E., Xynos, E., Mavroudis, D., Tzardi, M., et al. (2018). Evaluation of the detection of Toll-like receptors (TLRs) in cancer development and progression in patients with colorectal cancer. PloS one 13, e0197327.

Michalak, M., Corbett, E.F., Mesaeli, N., Nakamura, K., and Opas, M. (1999). Calreticulin: one protein, one gene, many functions. The Biochemical journal 344 Pt 2, 281-292.

Michaud, M., Martins, I., Sukkurwala, A.Q., Adjemian, S., Ma, Y., Pellegatti, P., Shen, S., Kepp, O., Scoazec, M., Mignot, G., et al. (2011). Autophagy-dependent anticancer immune responses induced by chemotherapeutic agents in mice. Science 334, 1573-1577.

Mitsuyasu, R.T. (2009). AIDS-related malignancies.

Mittal, D., Gubin, M.M., Schreiber, R.D., and Smyth, M.J. (2014). New insights into cancer immunoediting and its three component phases--elimination, equilibrium and escape. Current opinion in immunology 27, 16-25.

Muller, S., Ronfani, L., and Bianchi, M.E. (2004). Regulated expression and subcellular localization of HMGB1, a chromatin protein with a cytokine function. Journal of internal medicine 255, 332-343.

Obeid, M., Panaretakis, T., Joza, N., Tufi, R., Tesniere, A., van Endert, P., Zitvogel, L., and Kroemer, G. (2007a). Calreticulin exposure is required for the immunogenicity of gamma- irradiation and UVC light-induced apoptosis. Cell death and differentiation 14, 1848-1850.

Obeid, M., Tesniere, A., Ghiringhelli, F., Fimia, G.M., Apetoh, L., Perfettini, J.L., Castedo, M., Mignot, G., Panaretakis, T., Casares, N., et al. (2007b). Calreticulin exposure dictates the immunogenicity of cancer cell death. Nature medicine 13, 54-61.

57

Old, L.J., and Boyse, E.A. (1964). Immunology of Experimental Tumors. Annual review of medicine 15, 167-186.

Osinska, I., Popko, K., and Demkow, U. (2014). Perforin: an important player in immune response. Central-European journal of immunology 39, 109-115.

Outzen, H.C., Custer, R.P., Eaton, G.J., and Prehn, R.T. (1975). Spontaneous and induced tumor incidence in germfree "nude" mice. Journal of the Reticuloendothelial Society 17, 1-9.

Palumbo, R., Sampaolesi, M., De Marchis, F., Tonlorenzi, R., Colombetti, S., Mondino, A., Cossu, G., and Bianchi, M.E. (2004). Extracellular HMGB1, a signal of tissue damage, induces mesoangioblast migration and proliferation. The Journal of cell biology 164, 441-449.

Panaretakis, T., Joza, N., Modjtahedi, N., Tesniere, A., Vitale, I., Durchschlag, M., Fimia, G.M., Kepp, O., Piacentini, M., Froehlich, K.U., et al. (2008). The co-translocation of ERp57 and calreticulin determines the immunogenicity of cell death. Cell death and differentiation 15, 1499- 1509.

Panaretakis, T., Kepp, O., Brockmeier, U., Tesniere, A., Bjorklund, A.C., Chapman, D.C., Durchschlag, M., Joza, N., Pierron, G., van Endert, P., et al. (2009). Mechanisms of pre-apoptotic calreticulin exposure in immunogenic cell death. The EMBO journal 28, 578-590.

Pawaria, S., and Binder, R.J. (2011). CD91-dependent programming of T-helper cell responses following heat shock protein immunization. Nature communications 2, 521.

Paz-Ares, L., Spira, A., Raben, D., Planchard, D., Cho, B.C., Ozguroglu, M., Daniel, D., Villegas, A., Vicente, D., Hui, R., et al. (2020). Outcomes with durvalumab by tumour PD-L1 expression in unresectable, stage III non-small-cell lung cancer in the PACIFIC trial. Annals of oncology : official journal of the European Society for Medical Oncology.

Peng, R.Q., Chen, Y.B., Ding, Y., Zhang, R., Zhang, X., Yu, X.J., Zhou, Z.W., Zeng, Y.X., and Zhang, X.S. (2010). Expression of calreticulin is associated with infiltration of T-cells in stage IIIB colon cancer. World journal of gastroenterology 16, 2428-2434.

Penn, I. (1996). Malignant melanoma in organ allograft recipients. Transplantation 61, 274-278.

58

Pfirschke, C., Engblom, C., Rickelt, S., Cortez-Retamozo, V., Garris, C., Pucci, F., Yamazaki, T., Poirier-Colame, V., Newton, A., Redouane, Y., et al. (2016). Immunogenic Chemotherapy Sensitizes Tumors to Checkpoint Blockade Therapy. Immunity 44, 343-354.

Pozzi, C., Cuomo, A., Spadoni, I., Magni, E., Silvola, A., Conte, A., Sigismund, S., Ravenda, P.S., Bonaldi, T., Zampino, M.G., et al. (2016). The EGFR-specific antibody cetuximab combined with chemotherapy triggers immunogenic cell death. Nature medicine 22, 624-631.

Pylayeva-Gupta, Y., Das, S., Handler, J.S., Hajdu, C.H., Coffre, M., Koralov, S.B., and Bar-Sagi, D. (2016). IL35-Producing B Cells Promote the Development of Pancreatic Neoplasia. Cancer discovery 6, 247-255.

Radoja, S., Rao, T.D., Hillman, D., and Frey, A.B. (2000). Mice bearing late-stage tumors have normal functional systemic T cell responses in vitro and in vivo. Journal of immunology 164, 2619-2628.

Roithmaier, S., Haydon, A.M., Loi, S., Esmore, D., Griffiths, A., Bergin, P., Williams, T.J., and Schwarz, M.A. (2007). Incidence of malignancies in heart and/or lung transplant recipients: a single-institution experience. The Journal of heart and lung transplantation : the official publication of the International Society for Heart Transplantation 26, 845-849.

Romano, M., Frapolli, R., Zangarini, M., Bello, E., Porcu, L., Galmarini, C.M., Garcia-Fernandez, L.F., Cuevas, C., Allavena, P., Erba, E., et al. (2013). Comparison of in vitro and in vivo biological effects of trabectedin, lurbinectedin (PM01183) and Zalypsis(R) (PM00104). International journal of cancer 133, 2024-2033.

Rosenberg, J.E., Hoffman-Censits, J., Powles, T., van der Heijden, M.S., Balar, A.V., Necchi, A., Dawson, N., O'Donnell, P.H., Balmanoukian, A., Loriot, Y., et al. (2016). Atezolizumab in patients with locally advanced and metastatic urothelial carcinoma who have progressed following treatment with platinum-based chemotherapy: a single-arm, multicentre, phase 2 trial. Lancet 387, 1909-1920.

Russell, J.H., and Ley, T.J. (2002). Lymphocyte-mediated cytotoxicity. Annual review of immunology 20, 323-370.

59

Rygaard, J., and Povlsen, C.O. (1974a). Is immunological surveillance not a cell-mediated immune function? Transplantation 17, 135-136.

Rygaard, J., and Povlsen, C.O. (1974b). The mouse mutant nude does not develop spontaneous tumours. An argument against immunological surveillance. Acta pathologica et microbiologica Scandinavica Section B: Microbiology and immunology 82, 99-106.

Santamaria Nunez, G., Robles, C.M., Giraudon, C., Martinez-Leal, J.F., Compe, E., Coin, F., Aviles, P., Galmarini, C.M., and Egly, J.M. (2016). Lurbinectedin Specifically Triggers the Degradation of Phosphorylated RNA Polymerase II and the Formation of DNA Breaks in Cancer Cells. Molecular cancer therapeutics 15, 2399-2412.

Scaffidi, P., Misteli, T., and Bianchi, M.E. (2002). Release of chromatin protein HMGB1 by necrotic cells triggers inflammation. Nature 418, 191-195.

Schiavoni, G., Sistigu, A., Valentini, M., Mattei, F., Sestili, P., Spadaro, F., Sanchez, M., Lorenzi, S., D'Urso, M.T., Belardelli, F., et al. (2011). Cyclophosphamide synergizes with type I interferons through systemic dendritic cell reactivation and induction of immunogenic tumor apoptosis. Cancer research 71, 768-778.

Schreiber, R.D., Old, L.J., and Smyth, M.J. (2011). Cancer immunoediting: integrating immunity's roles in cancer suppression and promotion. Science 331, 1565-1570.

Shankaran, V., Ikeda, H., Bruce, A.T., White, J.M., Swanson, P.E., Old, L.J., and Schreiber, R.D. (2001). IFNgamma and lymphocytes prevent primary tumour development and shape tumour immunogenicity. Nature 410, 1107-1111.

Shen, J., Chen, X., Hendershot, L., and Prywes, R. (2002). ER stress regulation of ATF6 localization by dissociation of BiP/GRP78 binding and unmasking of Golgi localization signals. Developmental cell 3, 99-111.

Shoulders, M.D., Ryno, L.M., Genereux, J.C., Moresco, J.J., Tu, P.G., Wu, C., Yates, J.R., 3rd, Su, A.I., Kelly, J.W., and Wiseman, R.L. (2013). Stress-independent activation of XBP1s and/or ATF6 reveals three functionally diverse ER proteostasis environments. Cell reports 3, 1279-1292.

60

Sims, G.P., Rowe, D.C., Rietdijk, S.T., Herbst, R., and Coyle, A.J. (2010). HMGB1 and RAGE in inflammation and cancer. Annual review of immunology 28, 367-388.

Sistigu, A., Yamazaki, T., Vacchelli, E., Chaba, K., Enot, D.P., Adam, J., Vitale, I., Goubar, A., Baracco, E.E., Remedios, C., et al. (2014). Cancer cell-autonomous contribution of type I interferon signaling to the efficacy of chemotherapy. Nature medicine 20, 1301-1309.

Smyth, M.J., Thia, K.Y., Street, S.E., Cretney, E., Trapani, J.A., Taniguchi, M., Kawano, T., Pelikan, S.B., Crowe, N.Y., and Godfrey, D.I. (2000). Differential tumor surveillance by natural killer (NK) and NKT cells. The Journal of experimental medicine 191, 661-668.

Soares, D.G., Machado, M.S., Rocca, C.J., Poindessous, V., Ouaret, D., Sarasin, A., Galmarini, C.M., Henriques, J.A., Escargueil, A.E., and Larsen, A.K. (2011). Trabectedin and its C subunit modified analogue PM01183 attenuate nucleotide excision repair and show activity toward platinum-resistant cells. Molecular cancer therapeutics 10, 1481-1489.

Spisek, R., Charalambous, A., Mazumder, A., Vesole, D.H., Jagannath, S., and Dhodapkar, M.V. (2007). Bortezomib enhances dendritic cell (DC)-mediated induction of immunity to human myeloma via exposure of cell surface heat shock protein 90 on dying tumor cells: therapeutic implications. Blood 109, 4839-4845.

Stagg, J., Beavis, P.A., Divisekera, U., Liu, M.C., Moller, A., Darcy, P.K., and Smyth, M.J. (2012). CD73-deficient mice are resistant to carcinogenesis. Cancer research 72, 2190-2196.

Street, S.E., Cretney, E., and Smyth, M.J. (2001). Perforin and interferon-gamma activities independently control tumor initiation, growth, and metastasis. Blood 97, 192-197.

Stutman, O. (1974). Tumor development after 3-methylcholanthrene in immunologically deficient athymic-nude mice. Science 183, 534-536.

Stutman, O. (1979). Chemical carcinogenesis in nude mice: comparison between nude mice from homozygous matings and heterozygous matings and effect of age and carcinogen dose. Journal of the National Cancer Institute 62, 353-358.

61

Sun, X., Wu, Y., Gao, W., Enjyoji, K., Csizmadia, E., Muller, C.E., Murakami, T., and Robson, S.C. (2010). CD39/ENTPD1 expression by CD4+Foxp3+ regulatory T cells promotes hepatic metastatic tumor growth in mice. Gastroenterology 139, 1030-1040.

Svane, I.M., Engel, A.M., Nielsen, M.B., Ljunggren, H.G., Rygaard, J., and Werdelin, O. (1996). Chemically induced sarcomas from nude mice are more immunogenic than similar sarcomas from congenic normal mice. European journal of immunology 26, 1844-1850.

Tang, D., Kang, R., Livesey, K.M., Cheh, C.W., Farkas, A., Loughran, P., Hoppe, G., Bianchi, M.E., Tracey, K.J., Zeh, H.J., 3rd, et al. (2010). Endogenous HMGB1 regulates autophagy. The Journal of cell biology 190, 881-892.

Teng, M.W., Vesely, M.D., Duret, H., McLaughlin, N., Towne, J.E., Schreiber, R.D., and Smyth, M.J. (2012). Opposing roles for IL-23 and IL-12 in maintaining occult cancer in an equilibrium state. Cancer research 72, 3987-3996.

Tesniere, A., Schlemmer, F., Boige, V., Kepp, O., Martins, I., Ghiringhelli, F., Aymeric, L., Michaud, M., Apetoh, L., Barault, L., et al. (2010). Immunogenic death of colon cancer cells treated with oxaliplatin. Oncogene 29, 482-491.

Thomas, L. (1982). On immunosurveillance in human cancer. The Yale journal of biology and medicine 55, 329-333.

Topalian, S.L., Hodi, F.S., Brahmer, J.R., Gettinger, S.N., Smith, D.C., McDermott, D.F., Powderly, J.D., Carvajal, R.D., Sosman, J.A., Atkins, M.B., et al. (2012). Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. The New England journal of medicine 366, 2443-2454.

Trigo, J., Subbiah, V., Besse, B., Moreno, V., Lopez, R., Sala, M.A., Peters, S., Ponce, S., Fernandez, C., Alfaro, V., et al. (2020). Lurbinectedin as second-line treatment for patients with small-cell lung cancer: a single-arm, open-label, phase 2 basket trial. The Lancet Oncology 21, 645-654.

62

Vacchelli, E., Aranda, F., Eggermont, A., Galon, J., Sautes-Fridman, C., Cremer, I., Zitvogel, L., Kroemer, G., and Galluzzi, L. (2014). Trial Watch: Chemotherapy with immunogenic cell death inducers. Oncoimmunology 3, e27878.

Vacchelli, E., Ma, Y., Baracco, E.E., Sistigu, A., Enot, D.P., Pietrocola, F., Yang, H., Adjemian, S., Chaba, K., Semeraro, M., et al. (2015). Chemotherapy-induced antitumor immunity requires formyl peptide receptor 1. Science 350, 972-978. van den Broek, M.E., Kagi, D., Ossendorp, F., Toes, R., Vamvakas, S., Lutz, W.K., Melief, C.J., Zinkernagel, R.M., and Hengartner, H. (1996). Decreased tumor surveillance in perforin-deficient mice. The Journal of experimental medicine 184, 1781-1790.

Vanpouille-Box, C., Alard, A., Aryankalayil, M.J., Sarfraz, Y., Diamond, J.M., Schneider, R.J., Inghirami, G., Coleman, C.N., Formenti, S.C., and Demaria, S. (2017). DNA exonuclease Trex1 regulates radiotherapy-induced tumour immunogenicity. Nature communications 8, 15618.

Vattem, K.M., and Wek, R.C. (2004). Reinitiation involving upstream ORFs regulates ATF4 mRNA translation in mammalian cells. Proceedings of the National Academy of Sciences of the United States of America 101, 11269-11274.

Venereau, E., Casalgrandi, M., Schiraldi, M., Antoine, D.J., Cattaneo, A., De Marchis, F., Liu, J., Antonelli, A., Preti, A., Raeli, L., et al. (2012). Mutually exclusive redox forms of HMGB1 promote cell recruitment or proinflammatory cytokine release. The Journal of experimental medicine 209, 1519-1528.

Vidal, A., Munoz, C., Guillen, M.J., Moreto, J., Puertas, S., Martinez-Iniesta, M., Figueras, A., Padulles, L., Garcia-Rodriguez, F.J., Berdiel-Acer, M., et al. (2012). Lurbinectedin (PM01183), a new DNA minor groove binder, inhibits growth of orthotopic primary graft of cisplatin-resistant epithelial ovarian cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 18, 5399-5411.

63

Voorwerk, L., Slagter, M., Horlings, H.M., Sikorska, K., van de Vijver, K.K., de Maaker, M., Nederlof, I., Kluin, R.J.C., Warren, S., Ong, S., et al. (2019). Immune induction strategies in metastatic triple-negative breast cancer to enhance the sensitivity to PD-1 blockade: the TONIC trial. Nature medicine 25, 920-928.

Wang, G., Li, X., and Wang, Z. (2016). APD3: the antimicrobial peptide database as a tool for research and education. Nucleic acids research 44, D1087-1093.

Wu, S., Sun, C., Tian, D., Li, Y., Gao, X., He, S., and Li, T. (2015). Expression and clinical significances of Beclin1, LC3 and mTOR in colorectal cancer. International journal of clinical and experimental pathology 8, 3882-3891.

Xu, Q., Chen, C., Chen, G., Chen, W., Zhou, D., and Xie, Y. (2018). Significance of calreticulin as a prognostic factor in endometrial cancer. Oncology letters 15, 8999-9008.

Yamazaki, T., Buque, A., Ames, T.D., and Galluzzi, L. (2020). PT-112 induces immunogenic cell death and synergizes with immune checkpoint blockers in mouse tumor models. Oncoimmunology 9, 1721810.

Yamazaki, T., Pitt, J.M., Vetizou, M., Marabelle, A., Flores, C., Rekdal, O., Kroemer, G., and Zitvogel, L. (2016). The oncolytic peptide LTX-315 overcomes resistance of cancers to immunotherapy with CTLA4 checkpoint blockade. Cell death and differentiation 23, 1004-1015.

Yang, H., Lundback, P., Ottosson, L., Erlandsson-Harris, H., Venereau, E., Bianchi, M.E., Al- Abed, Y., Andersson, U., Tracey, K.J., and Antoine, D.J. (2012). Redox modification of cysteine residues regulates the cytokine activity of high mobility group box-1 (HMGB1). Molecular medicine 18, 250-259.

Yokoi, E., Mabuchi, S., Shimura, K., Komura, N., Kozasa, K., Kuroda, H., Takahashi, R., Sasano, T., Kawano, M., Matsumoto, Y., et al. (2019). Lurbinectedin (PM01183), a selective inhibitor of active transcription, effectively eliminates both cancer cells and cancer stem cells in preclinical models of uterine cervical cancer. Investigational new drugs 37, 818-827.

64

Yoo, Y.C., Watanabe, R., Koike, Y., Mitobe, M., Shimazaki, K., Watanabe, S., and Azuma, I. (1997). Apoptosis in human leukemic cells induced by lactoferricin, a bovine milk protein-derived peptide: involvement of reactive oxygen species. Biochemical and biophysical research communications 237, 624-628.

Zamarin, D., Holmgaard, R.B., Subudhi, S.K., Park, J.S., Mansour, M., Palese, P., Merghoub, T., Wolchok, J.D., and Allison, J.P. (2014). Localized oncolytic virotherapy overcomes systemic tumor resistance to immune checkpoint blockade immunotherapy. Science translational medicine 6, 226ra232.

Zhang, H., Liu, Z., and Liu, S. (2016). HMGB1 induced inflammatory effect is blocked by CRISPLD2 via MiR155 in hepatic fibrogenesis. Molecular immunology 69, 1-6.

Zhou, H., Forveille, S., Sauvat, A., Yamazaki, T., Senovilla, L., Ma, Y., Liu, P., Yang, H., Bezu, L., Muller, K., et al. (2016a). The oncolytic peptide LTX-315 triggers immunogenic cell death. Cell death & disease 7, e2134.

Zhou, H., Mondragon, L., Xie, W., Mauseth, B., Leduc, M., Sauvat, A., Gomes-da-Silva, L.C., Forveille, S., Iribarren, K., Souquere, S., et al. (2018). Oncolysis with DTT-205 and DTT-304 generates immunological memory in cured animals. Cell death & disease 9, 1086.

Zhou, H., Sauvat, A., Gomes-da-Silva, L.C., Durand, S., Forveille, S., Iribarren, K., Yamazaki, T., Souquere, S., Bezu, L., Muller, K., et al. (2016b). The oncolytic compound LTX-401 targets the Golgi apparatus. Cell death and differentiation 23, 2031-2041.

65

ANNEX 1: SCIENTIFIC PUBLICATIONS

1. Xie W, Forveille S, Iribarren K, Sauvat A, Senovilla L, Wang Y, Humeau J, Perez-Lanzon M, Zhou H, Martínez-Leal JF, Kroemer G, Kepp O. Lurbinectedin synergizes with immune checkpoint blockade to generate anticancer immunity. Oncoimmunology. 2019 Sep 5;8(11):e1656502. doi: 10.1080/2162402X.2019.1656502.

2. Xie W, Mondragón L, Mauseth B, Wang Y, Pol J, Lévesque S, Zhou H, Yamazaki T, Eksteen JJ, Zitvogel L, Sveinbjørnsson B, Rekdal Ø, Kepp O, Kroemer G. Tumor lysis with LTX-401 creates anticancer immunity. Oncoimmunology. 2019 Apr 13;8(7):1594555. doi: 10.1080/2162402X.2019.1594555.

3. Wang Y, Xie W*, Humeau J, Chen G, Liu P, Pol J, Zhang Z, Kepp O, Kroemer G. Autophagy induction by thiostrepton improves the efficacy of immunogenic chemotherapy. J Immunother Cancer. 2020 Mar;8(1):e000462. doi: 10.1136/jitc-2019-000462. (*co-first author)

4. Zhou H, Mondragón L, Xie W*, Mauseth B, Leduc M, Sauvat A, Gomes-da-Silva LC, Forveille S, Iribarren K, Souquere S, Bezu L, Liu P, Zhao L, Zitvogel L, Sveinbjørnsson B, Eksteen JJ, Rekdal Ø, Kepp O, Kroemer G. Oncolysis with DTT-205 and DTT-304 generates immunological memory in cured animals. Cell Death Dis. 2018 Oct 23;9(11):1086. doi: 10.1038/s41419-018-1127-3. (*co-first author)

5. Loos F, Xie W*, Sica V, Bravo-San Pedro JM, Souquère S, Pierron G, Lachkar S, Sauvat A, Petrazzuolo A, Jimenez AJ, Perez F, Maiuri MC, Kepp O, Kroemer G. Artificial tethering of LC3 or p62 to organelles is not sufficient to trigger autophagy. Cell Death Dis. 2019 Oct 10;10(10):771. doi: 10.1038/s41419-019-2011-5. (*co-first author)

6. Zhang S, Wang Y, Xie W*, Howe ENW, Busschaert N, Sauvat A, Leduc M, Gomes-da-Silva LC, Chen G, Martins I, Deng X, Maiuri L, Kepp O, Soussi T, Gale PA, Zamzami N, Kroemer G. Squaramide-based synthetic chloride transporters activate TFEB but block

66

autophagic flux. Cell Death Dis. 2019 Mar 11;10(3):242. doi: 10.1038/s41419-019-1474-8. Erratum in: Cell Death Dis. 2019 Apr 3;10(4):301. (*co-first author)

7. Chen G, Xie W, Nah J, Sauvat A, Liu P, Pietrocola F, Sica V, Carmona-Gutierrez D, Zimmermann A, Pendl T, Tadic J, Bergmann M, Hofer SJ, Domuz L, Lachkar S, Markaki M, Tavernarakis N, Sadoshima J, Madeo F, Kepp O, Kroemer G. 3,4-Dimethoxychalcone induces autophagy through activation of the transcription factors TFE3 and TFEB. EMBO Mol Med. 2019 Nov 7;11(11):e10469. doi: 10.15252/emmm.201910469. Epub 2019 Oct 14.

8. Humeau J, Sauvat A, Cerrato G, Xie W, Loos F, Iannantuoni F, Bezu L, Lévesque S, Paillet J, Pol J, Leduc M, Zitvogel L, de Thé H, Kepp O, Kroemer G. Inhibition of transcription by dactinomycin reveals a new characteristic of immunogenic cell stress. EMBO Mol Med. 2020 May 8;12(5):e11622. doi: 10.15252/emmm.201911622. Epub 2020 Apr 23.

9. Liu P, Zhao L, Loos F, Marty C, Xie W, Martins I, Lachkar S, Qu B, Waeckel-Énée E, Plo I, Vainchenker W, Perez F, Rodriguez D, López-Otin C, van Endert P, Zitvogel L, Kepp O, Kroemer G. Immunosuppression by Mutated Calreticulin Released from Malignant Cells. Mol Cell. 2020 Feb 20;77(4):748-760.e9. doi: 10.1016/j.molcel.2019.11.004. Epub 2019 Nov 27.

10. Iribarren K, Buque A, Mondragon L, Xie W, Lévesque S, Pol J, Zitvogel L, Kepp O, Kroemer G. Anticancer effects of anti-CD47 immunotherapy in vivo. Oncoimmunology. 2018 Dec 11;8(3):1550619. doi: 10.1080/2162402X.2018.1550619.

11. Gomes-da-Silva LC, Jimenez AJ, Sauvat A, Xie W, Souquere S, Divoux S, Storch M, Sveinbjørnsson B, Rekdal Ø, Arnaut LG, Kepp O, Kroemer G, Perez F. Recruitment of LC3 to damaged Golgi apparatus. Cell Death Differ. 2019 Aug;26(8):1467-1484. doi: 10.1038/s41418-018-0221-5. Epub 2018 Oct 22.

67

12. Kepp O, Sauvat A, Leduc M, Forveille S, Liu P, Zhao L, Bezu L, Xie W, Zitvogel L, Kroemer G. A fluorescent biosensor-based platform for the discovery of immunogenic cancer cell death inducers. Oncoimmunology. 2019 Apr 26;8(8):1606665. doi: 10.1080/2162402X.2019.1606665.

68

ANNEX 2: PAPERS NOT INCLUDED IN THIS THESIS (INSERTED IN PDF)

69 Open access Original research J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from Autophagy induction by thiostrepton improves the efficacy of immunogenic chemotherapy

1,2,3,4,5,6,7,8,9 1,2,3,4,5,6,7 1,2,3,4,5,6,7 Yan Wang, Wei Xie , Juliette Humeau, Guo Chen,1,2,3,4,5,6,7,10 Peng Liu,1,2,3,4,5,6,7 Jonathan Pol,1,2,3,4,5,6,7 Zhen Zhang,8,9 1,2,3,4,5,6,7 1,2,3,4,5,6,7,11,12,13 Oliver Kepp , Guido Kroemer

To cite: Wang Y, Xie W, AbstrACt Only a fraction of cytotoxicants are able to Humeau J, et al. Autophagy background Immunogenic cell death (ICD) is a peculiar stimulate ICD, which requires the premortem induction by thiostrepton modality of cellular demise that elicits adaptive immune improves the efficacy of induction of specific stress pathways, in responses and triggers T cell- dependent immunity. immunogenic chemotherapy. particular, autophagy and partial endo- Journal for ImmunoTherapy Methods Fluorescent biosensors were employed for an plasmic reticulum (ER) stress response with of Cancer 2020;8:e000462. unbiased drug screen approach aiming at the identification the phosphorylation of eukaryotic initiation doi:10.1136/jitc-2019-000462 of ICD enhancers. factor 2α (eIF2α).8–10 Autophagy is required results Here, we discovered thiostrepton as an for the lysosomal secretion of ATP, a chemo- enhancer of ICD able to boost chemotherapy- induced ► Additional material is tactic factor that is released from stressed/ published online only. To view ATP release, calreticulin exposure and high- mobility Medicale. Protected by copyright. please visit the journal online group box 1 exodus. Moreover, thiostrepton enhanced dying cancer cells and attracts dendritic cell (http://dx. doi. org/ 10. 1136/ jitc- anticancer immune responses of oxaliplatin (OXA) in precursors into the tumor bed via the action 2019-000462). vivo in immunocompetent mice, yet failed to do so in on purinergic receptors and may contribute to local inflammasome activation as well.11 12 YW and WX contributed equally. immunodeficient animals. Consistently, thiostrepton combined with OXA altered the ratio of cytotoxic T ER stress facilitates the surface exposure of YW and WX are joint first lymphocytes to regulatory T cells, thus overcoming the most abundant ER luminal protein, calre- authors. immunosuppression and reinstating anticancer ticulin (CALR).9 10 13 14 Once present at the ZZ, OK and GK are joint senior immunosurveillance. surface of the plasma membrane of stressed/ authors. Conclusion Altogether, these results indicate that dying cancer cells, CALR serves as a potent thiostrepton can be advantageously combined with ‘eat- me’ signal to facilitate the uptake of Accepted 11 March 2020 http://jitc.bmj.com/ chemotherapy to enhance anticancer immunogenicity. tumor antigens by immature dendritic cells.15 High-mobility group box 1 (HMGB1), which is the most abundant non-histone chromatin- IntroduCtIon binding protein, is released from dying/dead Although it was thought for more than 50 cells to reach toll-like receptor 4 (TLR4) on dendritic cells, stimulating their maturation

years that anticancer cytotoxicants would kill on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque tumor cells, thus yielding a direct antineo- for the cross-presentation of tumor antigens 16 17 plastic effect,1 it has become clear over the to CTL. Hence, this cascade (ATP, CALR, 2 18 past 15 years that some particularly successful HMGB1) contributes to the stepwise initi- chemotherapeutics do not act in such a direct ation of the anticancer immune response © Author(s) (or their within the tumor bed and composes the triad employer(s)) 2020. Re- use fashion but rather stress and kill a fraction of permitted under CC BY- NC. No the tumor cells in a way that they elicit an of ICD. commercial re- use. See rights immune response against tumor- associated and permissions. Published by 2 BMJ. antigens. Thus, immunogenic cell death 3 For numbered affiliations see (ICD) causes tumor growth control via the MAterIAls And Methods end of article. induction of antitumor immunity requiring Cell culture and chemicals the implication of dendritic cells and inter- Culture media and supplements for cell Correspondence to feron-γ producing cytotoxic T lymphocytes Dr Oliver Kepp; culture were obtained from Life Technologies captain. olsen@ gmail. com (CTLs). This implies that some chemother- (Carlsbad, California, USA) and plasticware apeutics can be advantageously combined came from Greiner BioOne (Kremsmün- Professor Zhen Zhang; with immune checkpoint blockers targeting ster, Austria). PC12 cells stably expressing zhen_ zhang@ fudan. edu. cn the programmed cell death death 1/ doxycycline-inducible Q74-GFP were cultured Guido Kroemer; programmed death ligand 1 (PD-1/PD-L1) in Roswell Park Memorial Institute (RPMI)- kroemer@ orange. fr interaction.4–7 1640 containing 5% fetal bovine serum and

Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 1 Open access J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from

10% horse serum.19 Human osteosarcoma U2OS TFE-/- LC3, wild- type C57BL/6 mice were treated with thiost- cells carrying a knockout for TFE3 and TFEB,20 MCA205 repton, liver were thereafter collected and processed. murine fibrosarcoma stably expressing shRNAs inter- For the GFP- LC3 assessment, 8- week- old GFP- LC3 trans- fering with the expression of TFE3 and TFEB,20 and all genic (Tg-GFP- LC3) mice, containing a rat LC3- EGFP the other cells were maintained in Dulbecco’s modified fusion protein under the control of the chicken b-actin, Eagle’s medium, supplemented with 10% fetal bovine were treated with thiostrepton with or without 40 mg/kg serum and both containing 100 units/mL penicillin G of the autophagic flux inhibitor leupeptin 2 hours before sodium and 100 mg/mL streptomycin sulfate. All cells the organs were collected and processed. For tumor were cultured at 37°C under 5% CO2. The Prestwick growth experiments, MCA205 or TC-1 tumors were estab- Chemical Library was obtained from Prestwick Chemical lished in C57BL/6 hosts by subcutaneously inoculating (Illkirch, France), flavonoids were purchased from Extra- 3×105 cells. Thiostrepton was dissolved in 3.34% dimeth- synthese (Genay, France), fatty acids came from Larodan ylsulfoxide (DMSO), 10.02 % PG400 (Sigma-Aldrich), (Malmö, Sweden); known calorie restriction mimetics, 10.02% TWEEN80 (Sigma- Aldrich) and 76.6% corn oil cisplatin, dactinomycin, cycloheximide, staurosporine, (Sigma- Aldrich). thapsigargin, tunicamycin, bafilomycin A1, and cyclo- When tumors became palpable, mice were treated with heximide were from Sigma-Aldrich (St. Louis, Missouri, thiostrepton in 200 mg/kg 1 day before, on, and 1 day USA). after the chemotherapy (10 mg/kg oxaliplatin, Sigma- Aldrich) by intraperitoneal injection. On the following high-content microscopy days, thiostrepton were given 40 mg/kg alone three times Human osteosarcoma U2OS cells wild type or stably per week, and mice well-being and tumor growth were expressing GFP- LC3; GFP- TFEB; mRFP- GFP- LC3; monitored and documented. Animals were sacrificed mCherry- GFP- p62, HMGB1- GFP and CALR- RFP, or when tumor size reached endpoint or signs of obvious human glioma H4 cells stably expressing GFP- LC3 or discomfort were observed following the EU Directive

PC12 cells expressing GFP-Q74 were seeded in 384-well 63/2010 and our Ethical Committee advice. Medicale. Protected by copyright. black microplates. After treatment, cells were fixed with 4% paraformaldehyde (PFA, w/v in phosphate buff- detection of protein deacetylation ered saline (PBS)) overnight at 4°C and stained with U2OS- GFP- LC3 stable cells were seeded in 384-well micro- 10 µg/mL Hoechst 33 342 in PBS. Images were acquired plates for 24 hours. After experimental treatments, cells using an ImageXpress micro XL automated microscope were fixed with 4% PFA for 20 min at room temperature. (Molecular Devices, San Jose, California, USA) with a ×20 Thereafter, cells were incubated with an antibody specific PlanApo objective (Nikon, Tokyo, Japan). A minimum of for acetyl-alpha- tubulin (#5335, 1:500, Cell Signaling four view fields was acquired per well, and experiments Technology) in 5% bovine serum albumin (BSA, w/v in involved at least triplicate assessments. Quantitation PBS) for 1 hour to block non- specific binding sites and was done on 1000–2400 cells per condition. Following acetylated tubulins, followed by overnight incubation at http://jitc.bmj.com/ images were analyzed using the Custom Module Editor 4°C with specific antibody to detect acetylated proteins of the MetaXpress software (Molecular Devices). Briefly, at lysines (#623402, 1:400, BioLegend, San Diego, Cali- cells were segmented and divided into nuclear and cyto- fornia, USA). After washing several times with PBS, cells plasmic regions based on the nuclear Hoechst staining were incubated with Alexa Fluor conjugates (Life Tech- and cytoplasmic GFP or RFP signal. GFP-LC3 dots were nologies) against the primary antibody for 2 hours at measured in the cytoplasmic compartment. For quina- room temperature. Nuclei were stained by incubating the on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque crine assays, cells were incubated with Krebs- Ringer solu- cells with 10 µg/mL Hoechst 33 342 in PBS. Fluorescent tion containing 5 µM quinacrine and 20 µg/mL Hoechst images were acquired and analyzed as described before. 33 342 at 37°C for 30 min, then cells were washed and imaged alive. Quinacrine dots were measured in the cyto- determination of extracellular hMGb1 concentration plasmic compartment. GFP-TFEB or HMGB1 fluorescence Cell culture supernatants were collected and centrifuge intensities were measured in the nuclear compartment. to remove debris. The concentration of HMGB1 was CALR-RFP aggregation was measured perimembranous detected by ELISA (HMGB1 ELISA kit II; Shino Test while excluding perinuclear signal. Data processing and Corporation, Tokyo, Japan) according to the manufactur- statistical analyses were performed using the R software er's instructions. The positive control was from the same (http://www. r- project. org/). kit. Absorbance was analyzed by means of an Paradigm I3 multimode plate reader (Molecular Devices). In vivo experimentation All mice were kept in an environmental-controlled and Immunoblotting pathogen-free facility with 12 hours light/dark cycles and After treatment, cells were collected and lysed in cold food and water ad libitum. Female wild- type C57BL/6 RIPA lysis and extraction buffer (Thermo Fisher Scientific, mice and athymic female nude mice (nu/nu) at the age Carlsbad, California, USA) containing Pierce Protease and of 6–8 weeks were obtained from Envigo, France (Envigo, Phosphatase Inhibitor Mini Tablet (Thermo Fisher Scien- Huntingdon, UK). For immunoblotting assessment of tific) on ice for 40 min. After centrifugation at 12 000g

2 Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 Open access J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from Medicale. Protected by copyright. http://jitc.bmj.com/ on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque

Figure 1 Identification of thiostrepton as autophagy inducer and immunogenic cell death enhancer. (A,B) Human neuroglioma H4 stably expressing GFP- LC3 were treated for 6 hours at two to five concentrations with 1428 chemical compounds originating from a selection of the Prestwick library (at 0.5, 5 and 50 µM), a flavonoids library (at 10, 30, and 90 µM) and a fatty acid library (at 31.25, 62.5, 125, 250 and 500 µM), together with their respective control vehicles (complete medium, DMSO or ethanol). This resulted in 4147 tested conditions in total. Torin 1 at 300 nM was used as positive control for autophagy induction. The number of GFP- LC3 positive puncta was measured as a proxy for autophagy and the number of cells that confer a regular nuclear phenotype was assessed as an indicator for viability. Data were normalized to controls and are depicted as ranked percentage of the normalized GFP- LC3 dots. A total of 175 compounds with increased GFP- LC3 dots count (and number of healthy cells >0.8) were selected for further experiments. Normalized data are shown as a dot plot in (A) and representative images are depicted in (B). (C, D) Human osteosarcoma U2OS cells stably expressing a tandem mRFP-GFP- LC3 biosensor were treated with the selected compounds to assess autophagic flux. Bafilomycin A1 (BAFA1) at 100 nM was used as a prototype flux inhibitor. The number of GFP positive and RFP positive dots was assessed and normalized to untreated controls. Cellular viability was assessed and normalized as described above. Data are depicted as ranked percentage of the ratio of RFP/GFP dots. A total of 34 autophagic flux inducing compounds (with numbers of healthy cells >0.8) were selected for further experiments. Normalized data are shown as a dot plot in (C) and representative images are depicted in (D). (E,F) U2OS wild-type cells were tested with the selected hit compounds at 50 µM in the presence of low- dose oxaliplatin (OXA at 60 µM) for 6 hours. High- dose OXA (400 µM) was used as positive control for ATP release. Quinacrine positive dots were assessed by microscopy and then values were normalized to controls. Normalized data are shown as a dot plot in (E) and representative images are depicted in (F). Scale bar equals 10 nm.

Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 3 Open access J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from Medicale. Protected by copyright.

Figure 2 Thiostrepton induces autophagic flux. (A–D) Human osteosarcoma U2OS cells stably expressing GFP- LC3 were treated with thiostrepton (THIOS; 1, 3, 9 µM), spermidine (SPDN; 200 µM), or vehicle (CTRL) for 6 hours. The cells were fixed and http://jitc.bmj.com/ acetylated tubulin was blocked by means of a specific antibody. Thereafter, cytoplasmic fluorescence intensity of acetylated lysine residues was assessed by appropriate Alexa Fluor 568- conjugated secondary antibodies (A). All data were normalized to controls (CTRL). Viable cells with regular nuclear phenotype are shown in (B). Protein acetylation are shown in (C), and GFP-LC3 dots were counted as a proxy for autophagic activity in (D). Representative images of acetylation and GFP-LC3 are shown in (A). Scale bar equals 10 µm. Data are means±SD of four replicates (**p<0.01, ***p<0.001; Student’s t-test). (E,F) Human glioma H4 cells stably expressing GFP- LC3 were treated with thiostrepton (1.9, 5.6, 16.7 and 50 µM), torin 1 (TORIN, 300 nM), or vehicle (CTRL), with or without bafilomyicn A1 (BAFA1) for 6 hours. GFP- LC3 dots were counted to assess autophagic activity. on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque Representative images of GFP- LC3 are shown in (E). Data are means±SD of four replicates (***p<0.001; Student’s t-test). (G,H) H4 wild- type cells were treated with thiostrepton (5 µM) or vehicle (CTRL), with or without BafA1 for 6 hours. Then cells were processed to measure LC3 and p62 protein levels by SDS–PAGE and immunoblot. GAPDH was used as a loading control. The experiments were repeated at least three times. Band intensities of p62, GAPDH, and LC3-II were measured, and ratios of p62 or LC3- II versus GAPDH (LC3- II/GAPDH, p62/GAPDH) were calculated in (H). Data are means±SEM of three independent experiments (*p<0.05 compared with untreated CTRL; #p<0.05, ##p<0.01, compared with no BAFA1; paired Student’s t-test). (H–I) U2OS cells stably expressing RFP- GFP- p62 tandem reporter were treated with thiostrepton (15, 30, 60 µM), vehicle (CTRL), or torin 1 (TORIN, 300 nM). GFP and RFP intensity were measured. The ratio of GFP/RFP intensity indicated p62 degradation (I). Representative images of RFP-GFP- p62 are shown in (I). Data are means±SD of four replicates (*p<0.05, ***p<0.001; Student’s t- test) (J). (K,L) Rat adrenal gland PC12 cells expressing inducible variant of Q74-GFP cells were treated with doxycycline (1 µg/mL) for 8 hours for the induction of Q74 expression. Then the medium was changed and thiostrepton (0.25, 0.5, 1, 2 µM), torin 1 (TORIN, 300 nM), or vehicle (CTRL) were added for 24 hours. The cells were fixed and GFP- Q74 dots were counted to assess Q74 degradation (L). Representative images of GFP- Q74 are shown in (K) (*p<0.05, ***p<0.001; Student’s t- test). (M,N) C57BL/6 mice were injected intraperitoneally with thiostrepton (200 mg/kg) or vehicle for 24 hours every day for 3 days, and tissues were collected 6 hours after the last injection. Then the tissue was subjected to SDS–PAGE and immunoblot (M). GAPDH was measured as a loading control. Band intensities of GAPDH and LC3- II were measured, and the ratio of LC3- II versus GAPDH (LC3- II/GAPDH) was calculated in (N). Data are means±SEM of 6 mice per group (***p<0.001; Student’s t-test). (O–P) GFP- LC3- expressing mice were i.p. injected with thiostrepton for three times (same as the first three times in tumor growth experiments). Leupeptin (Leu) was used to test autophagic flux in vivo, and GFP- LC3 dots were measured in liver tissue. Data are means±SEM of at least three mice (*p<0.05 vs CTRL without Leu; #p<0.05 vs CTRL with Leu; Student’s t- test).

4 Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 Open access J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from

Figure 3 Thiostrepton- induced autophagy is TFE- dependent. (A–D) Human osteosarcoma U2OS cells wild type, TFEB- TFE3 double knockout (B) or ATG5 knockout (C) cells expressing GFP- LC3 were treated with thiostrepton (2 µM) for 6 hour. GFP- LC3 dots were counted to measure autophagy activity (D). Representative images of GFP-LC3 are shown in (A). Data

are means±SD of four replicates (*p<0.05; Student’s t- test). (E–F) U2OS cells stably expressing TFEB- GFP fusion protein Medicale. Protected by copyright. were treated with indicated concentrations of thiostrepton (THIOS; 2, 5, 8, 10 µM) or torin 1 (TORIN, 300 nM) for 6 hours. Immunofluorescence staining was conducted with antibodies against TFE3 and appropriate Alexa Fluor 647- conjugated secondary antibody before the assessment of nuclear fluorescence intensity. GFP- TFEB nuclear intensities and Alexa Fluor 647(cy5)- TFE3 nuclear intensities were assessed (F). Data are means±SD of four replicates (**p<0.01;***p<0.001; Student’s t- test). Representative images are shown in (E). Scale bar equals 10 µm. (G,H) U2OS wild- type cells were treated with thiostrepton (0.1, 0.5, 1, 5, 10 µM), thapsgargin (3 µM), or vehicle for 6 hours. Immunofluorescence staining was conducted with antibodies against phospho- eIF2 alpha (Ser51) and appropriate Alexa Fluor 647- conjugated secondary antibody before the assessment of cytoplasmic fluorescence intensity (H). Representative images of phospho- eIF2 alpha (Ser51) are shown in (G). Data are means±SD of four replicates (*p<0.05; Student’s t- test). http://jitc.bmj.com/ for 15 min, supernatants were heated in sample buffer permeabilized with 0.1% Triton X100 on ice, and blocked (Thermo Fisher Scientific) at 100°C for 10 min. Protein with 5% bovine serum albumin (BSA, w/v in PBS) for samples were separated on precast 4%–12% polyacryl- 1 hour. For phospho-eIF2 alpha (Ser51) staining, fixed amide NuPAGE Bis–Tris gels (Thermo Fisher Scientific) cells were directly blocked with 5% BSA for 1 hour. Then and electrotransferred to Polyvinylidene difluoride (PVDF) cells were incubated with antibodies specific to TFE3 membranes (Millipore Corporation, Burlington, Massa- (#PA5-54909, 1:500, Invitrogen) or phospho-eIF2 alpha chusetts, USA). Membranes were probed overnight at 4°C (Ser51) (#ab32157, 1:500, Abcam) overnight (ON) incu- on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque with primary antibodies specific to LC3 (#2775, 1:1000, bation at 4°C. Thereafter, Alexa Fluor 647 nm conjugates Cell Signaling Technology), p62 (#ab56416, 1:10 000, (Thermo Fisher Scientific) against the primary antibody Abcam, Cambridge, UK), Atg5 (#A2859, 1:1000, Sigma- were applied for 2 hours at RT. Cells were then washed Aldrich, St. Louis, Missouri, USA), GAPDH (#ab8245, 1:10 and imaged by high-content microscopy as described 000, Abcam), TFEB (#4240, 1:1000, Cell Signaling Tech- above. The nuclear intensity of TFE3 and cytoplasmic nology, Danvers, Massachusetts, USA), TFE3 (ab93808, intensity of phospho-eIF2 alpha (Ser51) were measured 1:2000, Abcam), and β-actin (#ab49900, 1:10 000, Abcam) and normalized on Ctrl. followed by incubation with appropriate horseradish peroxidase–conjugated secondary antibodies (Southern Autophagy measurement on tissue sections Biotech, 1:5000, Birmingham, Alabama, USA). Immuno- Liver tissues were fixed with 10% neutral buffered reactive bands were visualized with ECL prime western formalin at overnight at 4°C and then transferred into blotting detection reagent (Sigma-Aldrich). Equal loading 30% sucrose diluted in PBS for 24 hours at 4°C. There- was controlled for all blots and one representative loading after, the organs were embedded in optimal cutting control is depicted, as indicated in the figure legends. temperature (OCT) solution, and consecutive tissue sections were cut by cryostat. Samples were stained with Immunofluorescence 4′,6-diamidino-2- phenylindole (DAPI) to detect nuclei, Cells were incubated with 3.7% PFA for 20 min at room and images were acquired by confocal microscopy. GFP- temperature. For TFE3 staining, fixed cells were then LC3 dot area was measured by ImageJ.

Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 5 Open access J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from

Figure 4 Validation of thiostrepton facilitated ATP release with mitoxantrone (MTX) and oxaliplatin (OXA). (A,B) Thiostrepton (THIOS; 0.5, 1, 2, 3, 4, 6 µM) decreased intracellular ATP content in human osteosarcoma U2OS wild-type cells within 6 hours in the presence of low doses of the ICD inducers mitoxantrone (MTX, 1 µM) and oxaliplatin (OXA; 200 µM) detected by quinacrine. High- dose MTX (5 µM) and OXA (400 µM) were used as positive controls. Representative images are shown in (A). Scale bar equals 10 µm. Data are means±SD of four replicates (compared with vehicle control (CTRL), *p<0.05, **p<0.01, ***p<0.001 compared with CTRL; ##p<0.01, ###p<0.001 compared with cells treated with low- dose MTX; &p<0.05, &&&p<0.01 compared with cells treated with low- dose OXA; unpaired Student’s t- test). (C–E) Thiostrepton increased extracellular ATP levels in murine fibrosarcoma MCA205 wild- type cells in the presence of low doses of the ICD inducers MTX (0.5 µM) and OXA (150 µM), but Medicale. Protected by copyright. much less efficient in MCA205- TFE knockdown or ATG5 knockdown cells. Cells were treated with THIOS (1.56, 3.12, 6.25, 12.5 µM) alone or in combination with low- dose MTX or OXA for 6 hours as described above. Data are means±SD of at least three replicates (*p<0.05, **p<0.01, ***p<0.001 compared with thiostrepton single treatment; #p<0.05 compared with cells treated with low-dose MTX alone, &&&p<0.01 compared with cells treated with low- dose OXA alone; unpaired Student’s t- test). extracellular AtP level 30 min at room temperature and images were acquired Cellular supernatants were collected and processed with and analyzed by automated microscopy. the ENLITEN ATP Assay System Bioluminescence Detec- tion Kit (Promega, Madison, Michigan, USA; #FF2000) ex vivo–phenotyping of the tumor immune infiltrate http://jitc.bmj.com/ following the manufacturer’s methods. Fluorescence was Tumors were harvested, weighed and transferred on detected by means of a Paradigm I3 multimode plate ice into gentleMACS C tubes (Miltenyi Biotec, Bergisch reader (Molecular Devices). Gladbach, Germany) containing 1 mL of RPMI medium. Tumors were dissociated first mechanically with scis- Flow cytometry analysis of CAlr surface exposure sors, then enzymatically using Miltenyi Biotec mouse

Cells were detached and collected and sequentially on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque tumor dissociation kit and a GentleMACS Octo Dissoci- incubated with Zombie UV (#423114; Biolegend), anti- ator following the manufacturer’s instructions. Tumor CALR rabbit monoclonal antibody (#ab2907; Abcam, homogenates were filtered through 70 µm MACS Smart- Cambridge, UK) for 30 min at 4°C, followed by Alexa Strainer (Miltenyi Biotec) and washed twice with PBS. Fluor 488, 546 or 647 goat anti-rabbit IgGs (#A11034, Finally, bulk tumor cells were homogenized in PBS at Thermo Fisher Scientific) for 30 min at 4°C, then fixed a concentration corresponding to 250 mg of the initial with 4% PFA in PBS for flow cytometric assessment. tumor weight per millilitre. Bulk tumor cell homoge- Flow cytometric analysis of annexin V and dAPI nates, each corresponding to 50 mg of the initial tumor Cells were detached and collected and incubated with sample, were stained with LIVE/DEAD Fixable Yellow annexin V (Biolegende #640943) and DAPI according to dye (Thermo Fisher Scientific). Fc receptors were the manufacturer’s advice. Following cells were assessed blocked with anti- mouse CD16/CD32 (clone 2.4G2, by flow cytometry. Mouse BD Fc Block, BD Pharmingen). Surface staining of murine immune cell populations infiltrating the reactive oxygen species detection tumor was performed with the following fluorochrome- Cells were treated as indicated and menadione was used conjugated antibodies: (1) ‘Neutrophils’ panel: anti- CD45 for 1 hour at 100 µM as a positive control. CellROX deep APC- Fire750 (clone 30 F-11, BioLegend), anti- Ly-6G PE red (#C10422, Thermo Fisher Scientific) was added for (clone 1A8, BD Pharmingen), anti-Ly- 6C FITC (clone 30 min at a final concentration of 5 µM. Then cells were AL-21, BD Pharmingen), anti-CD11b V450 (clone fixed and stained with 4% PFA containing Hoechst for M1/70, BD Pharmingen); (2) ‘T- cell activation/exhaustion’

6 Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 Open access J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from Medicale. Protected by copyright. Figure 5 Thiostrepton facilitates calreticulin (CALR) exposure and high-mobility group box 1 (HMGB1) release in the presence of immunogenic cell death (ICD) inducer. (A,B) Thiostrepton increased CALR exposure in human osteosarcoma U2OS cells stably expressing CALR- RFP. Thiostrepton (THIOS: 0.5, 1, 2, 3, 4, 6 µM) was employed in the presence of low- dose mitoxantrone (MTX, 1 µM) and oxaliplatin (OXA, 200 µM). High- dose MTX (5 µM) and OXA (400 µM) were used as positive controls for 18 hours. Representative images are shown in (A). Scale bar equals 10 µm. Data are means±SD of four replicates (*p<0.05, **p<0.01, ***p<0.001; Student’s t- test; compared with vehicle control (CTRL)). (C,D) Thiostrepton increased CALR exposure in MCA205 wild- type cells detected by flow cytometry. Cells were treated with THIOS (12.5 µM) alone or in combination with OXA (150 µM) cells for 6 hours. Representative histograms are shown in (L). Data are means±SEM of at least three replicates (***p<0.001, compared with CTRL; ###p<0.001, compared with OXA: Student’s unpaired t-test). (E) THIOS alone or in combination (COMB) at the same concentration as in (A–B) increased extracellular HMGB1 after 48 hours in MCA205 wild- type cells monitored by ELISA. Data are means±SEM of at least three replicates ((*p<0.05, **p<0.01, ***p<0.001, compared http://jitc.bmj.com/ with CTRL; ##p<0.01, compared with OXA: unpaired Student’s t- test). (F–G) Thiostrepton increased extracellular HMGB1 in U2OS cells and decreased HMGB1 content in U2OS HMGB1- GFP biosensor cells. Cells were treated as in (A,B) for 72 hours. Representative images are shown in (F). Scale bar equals 10 µm. Data are means±SD of four replicates (*p<0.05, **p<0.01, ***p<0.001, compared with CTRL; Student’s t- test).

panel: anti- CD3 APC (clone 145–2 C11, BioLegend), for cytometry. (Representative gating strategies are on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque anti-CD8 PE (clone 53–6.7, BD Pharmingen), anti- depicted in online supplementary figure S5 and S6.) CD4 PerCP- Cy5.5 (clone RM4-5, Thermo Fisher Scien- tific), anti- CD25 PE-Cy7 (clone PC61.5, Thermo Fisher statistical analysis Scientific). Then, cells were fixed and permeabilized Data are reported as means±SD of triplicate determi- in eBioscience Foxp3/Transcription Factor Staining nations and experiments were repeated at least twice, Buffer (Thermo Fisher Scientific). To complete the ‘T- yielding similar results if not otherwise reported. Statistical cell activation/exhaustion’ panel, an intranuclear staining significance was assessed by Student’s t- test. TumGrowth was performed with anti- FoxP3 FITC (clone FJK- 16s, and GraphPad were used to analyze in vivo data arising 21 Thermo Fisher Scientific). Finally, stained samples were from murine models. run through a BD LSR II flow cytometer. Data were acquired using BD FACSDiva software (BD Biosciences) and analyzed using FlowJo software (TreeStar). Abso- results And dIsCussIon lute counts of leukocytes and tumor cells were normal- A pharmacologic screen identifies thiostrepton as an ized considering the following parameters: weight of autophagy inducer the harvested tumor and total volume of the dissociated To identify bona fide autophagy inducers, we screened tumor cell suspension (cell concentration typically set a total of 1560 compounds contained in the Prestwick to 250 mg/mL in PBS), proportion of the whole cell library, as well as home- made libraries containing flavo- suspension and proportion of the cell suspension used noids18 20 and a range of saturated and unsaturated fatty

Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 7 Open access J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from Medicale. Protected by copyright. http://jitc.bmj.com/

Figure 6 Thiostrepton improves the efficacy of anticancer chemotherapy in vivo. (A,B) Effects of thiostrepton (THIOS) on tumor growth of TC-1 murine lung cancers (A) and MCA205 mouse fibrosarcomas cancers (B) in syngeneic immunocompetent C57Bl/6 mice. From left to right: (i) growth curves (mean±SEM); (ii) tumor size distributions at day 21 for TC1 tumor or day 25 for murine MCA205 fibrosarcoma of the data shown in (i); (iii) individual tumor growth curves of mice treated with oxaliplatin (OXA) on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque alone or in combination with THIOS of the data shown in (i); (iv) Kaplan-Meier curves with tumor size exceeding 250 mm2 as an endpoint. (B) n=8 for mice treated with OXA and n=6 without OXA), (C) n=10 for single treatments or untreated animals, n=13 for the group receiving combination treatment). (C,D) Effects of THIOS on the growth of MCA205 fibrosarcomas in immunodeficient nu/nu mice (C) and MCA205- TFE knockdown cells in immunocompetent C57Bl/6 mice (D). From left to right: (i) growth curves (mean±SEM); (ii) tumor size distributions at day 19 on nude mice or day 37 on C57Bl/6 mice of data shown in (i); (iii) individual tumor growth curves of mice treated with OXA alone or in combination with THIOS of data shown in (i); (iv) Kaplan-Meier curves tumor size exceeding 300 mm2 (nude mice) or 250 mm2 (C57Bl/6 mice) as an endpoint. (C): n=7 for untreated controls and the groups receiving single treatments, n=10 for the group receiving the combination treatment); (D) n=6 for each group. acids22 for their capacity to induce the formation of red fluorescent protein (mRFP)- GFP-LC3 tandem autophagic puncta in human H4 neuroblastoma cells reporter23 for their capacity to induce autophagic flux that express a green fluorescent protein (GFP) fused (meaning that they increase the ratio of RFP- positive to microtubule- associated proteins 1A/1B light chain over GFP- positive puncta) (figure 1C,D; online supple- 3B (best known as LC3) (figure 1A,B; online supple- mentary table S2). The 34 most effective candidate mentary table S1). The 175 compounds that were most compounds (online supplementary table 1) were finally efficient at one of the chosen concentrations were investigated for their capacity to reduce intracellular re-tested on another cell line, namely, human U2OS ATP concentrations measured with the fluorochrome osteosarcoma cells stably expressing the monomeric quinacrine in combination with the prototypic ICD

8 Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 Open access J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from

Figure 7 Cytofluorometric characterization of tumor- infiltrating lymphocytes (TIL). The number of CD3+ CD8+ cytotoxic T lymphocytes (A), CD4+ FOXP3+ CD25+ regulatory T cells (Treg) (B), the ratio of CD3+ CD8+ T lymphocytes over Treg (C), (compared with CTRL: ***p<0.001; compared with OXA: #p<0.05, NS, not statistically significant; Student’s t- test) (n=10 for the untreated controls and the groups receiving single treatments, n=13 for the group receiving the combination treatment). inducers mitoxantrone, using 3,4-dimethoxychalcone 20 translation in human cells cultured in vitro (see online as a positive control (figure 1E,F; online supplemen- supplementary figure S2A–D), indicating that it can tary table S3). This approach led to the identification interfere with protein synthesis in eukaryotes as it does in of thiostrepton, an antibiotic produced by Streptomyces bacteria.30 Thiostrepton did not lead to an overproduc- azureus and Streptomyces laurentii,24–26 as a potential ICD tion of reactive oxygen species (see online supplementary enhancer. figure S2E,F), yet had (limited) cytotoxic effects when cells were exposed to the drug for an extended amount of Medicale. Protected by copyright. Validation of autophagy induction by thiostrepton time (24 hours) (see online supplementary figure S2G–J). When used at micromolar concentrations for 6 hours, This effect was further increased when thiostrepton was thiostrepton had little or no cytotoxic activity on cultured combined with low-dose oxaliplatin. Altogether, these cells and reduced cytoplasmic protein acetylation, as results confirm that thiostrepton is a potent autophagy- measured by immunofluorescence staining, as it induced inducing and cell stress- inducing agent. cytoplasmic GFP-LC3 puncta in human and murine biosensor cells (figure 2A–D, online supplementary figure thiostrepton amplifies ICd surrogate markers S1A,B). In the presence of bafilomycin A1, thiostrepton Mitoxantrone (MTX) and oxaliplatin (OXA) are potent continued to increase the number of cytoplasmic GFP- ICD inducers.31 As such, they elicit the ICD hallmarks in LC3 puncta per cell (figure 2E,F, online supplementary vitro, in cultured cancer cells, including the autophagy- http://jitc.bmj.com/ figure S1A,B). Moreover, the autophagy-associated lipida- dependent ATP release into the supernatant, calreticulin tion of LC3 was enhanced by thiostrepton, more so in the (CALR) exposure on the cell surface, as well as nuclear presence of bafilomycin A1 (figure 2G,H, online supple- exodus of HMGB1.32–34 We determined the capacity mentary figure S1C,D), indicating that thiostrepton truly of thiostrepton to enhance these stigmata of ICD in a induced autophagy. Accordingly, thiostrepton acted on series of in vitro assays. The importance of thiostrepton- another biosensor that measures autophagic flux, namely, induced ATP release needs to be further validated in vivo. on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque a mCherry- GFP- sequestosome-1 (SQSTM1, best known as Nevertheless, it appears clear that thiostrepton reduced p62) to preferentially reduce the GFP-dependent fluores- the abundance of quinacrine-detectable ATP in the cyto- cence, indicating an increase in autophagy (figure 2I,J).27 plasm of cultured cells, more so if combined with low Moreover, thiostrepton stimulated the degradation concentrations of MTX or OXA, suggesting additive of a Q74-huntingtin- GFP fusion protein indicative of interactions (figure 4A,B). Accordingly, ATP concentra- an increase of autophagic flux28 (figure 2K,L). When tions in the supernatant increased with the thiostrepton/ systemically injected into mice, thiostrepton also caused MTX or thiostrepton/OXA concentration, and this ATP increased LC3 lipidation in the liver, suggesting that it is release was inhibited in cells in which ATG5 or both active in vivo (figure 2M–P). Of note, thiostrepton lost TFEB and TFE3 had been knocked down (figure 4C–I) its capacity to induce GFP- LC3 puncta in U2OS cells even after extended incubation for 24 hours (see online that lack the essential autophagy gene ATG5 or the two supplementary figure S3A,B). Thiostrepton was able to proautophagic transcription factors TFEB and TFE3 cause the relocation of an RFP- CALR fusion protein to (figure 3A–D). Indeed, thiostrepton induced the trans- the periphery of the cells either alone or in the pres- location of TFEB and TFE3 from the cytoplasm to the ence of MTX or OXA (figure 5A,B), and this effect did nucleus, as did the positive control torin (figure 3E,F). not require TFEB plus TFE3 (see online supplementary Moreover, thiostrepton induced the phosphorylation of figure S4A). The capacity of thiostrepton to cause CALR eIF2α, an event that is often associated with autophagy.29 exposure was validated by surface immunofluorescence However, thiostrepton did inhibit transcription and staining (figure 5C,D). Finally, protracted incubation

Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 9 Open access J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from

(48–72 hours) with thiostrepton could cause HMGB1 inducers such as MTX and OXA. Moreover, thiostrepton release, more so when it was combined with low- dose enhanced the anticancer activity of OXA in vivo. This MTX or OXA (figure 5E–G), which was reduced when adjuvant effect required the expression of proautophagic TFEB plus TFE3 was knocked down (see online supple- transcription factors by the cancer cells and was fully lost mentary figure S4B). Hence, thiostrepton can amplify in the absence of T lymphocytes. Indeed, thiostrepton ICD surrogate markers in the context of low-dose MTX combined with OXA stimulated a major change in the or OXA. tumor immune infiltrate, enhancing the density of CTL, but reducing that of Tregs and G-MDSC. Thiostrepton thiostrepton enhances the efficacy of immunogenic is an antibiotic,24 and the intestinal microbiota has been chemotherapy shown to contribute to anticancer immune responses Although thiostrepton alone had no tumor growth elicited by some immunogenic chemotherapies37 38 or reducing effect against TC1 non- small cell lung cancers immunotherapy, meaning that broad-spectrum antibi- or MCA205 fibrosarcomas established in immuno- otics causing sterilization of the intestinal tract reduce the competent C57/BL6 mice, it did increase the anti- efficacy of anticancer treatments.39 40 For this reason, it cancer activity of the ICD inducer OXA (but not appears unlikely that the immune-dependent anticancer cis-diamminedichloridoplatinum(II), abbreviated CDDP, effects of thiostrepton would be related to its antibiotic best known as cisplatin, which is unable to trigger ICD activity. Thus, thiostrepton could be repurposed as an see online supplementary figure S5),7 further dimin- enhancer of immunogenic chemotherapies, indepen- ishing tumor size and extending the survival of the dent of its antibiotic activity. animals (figure 6A,B). This thiostrepton/OXA combina- tion effect depended on T lymphocytes because tumor Author affiliations 1 developing in athymic nu/nu mice failed to reduce Gustave Roussy Cancer Campus, Villejuif, France 2INSERM, UMR1138, Centre de Recherche des Cordeliers, Paris, France their growth in response to the treatment (figure 6C). 3Equipe 11 labellisée par la Ligue Nationale contre le Cancer, Centre de Recherche Moreover, tumors formed by cancer cells lacking TFEB des Cordeliers, Paris, France Medicale. Protected by copyright. and TFE3 expression that were implanted in immuno- 4Université de Paris, Paris, France competent C57/BL6 mice, failed to exhibit the combi- 5Metabolomics and Cell Biology Platforms, Gustave Roussy Cancer Campus, nation effect as well (although they responded to OXA) Villejuif, France 6Sorbonne Université, Paris, France (figure 6D), supporting the idea that TFEB/TFE3- 7Faculté de Médecine, Université Paris- Saclay, Kremlin-Bicêtre, France dependent autophagy induction is critical for the anti- 8Department of Radiation Oncology, Fudan University Shanghai Cancer Center, cancer action of thiostrepton. Analysis of the leukocyte Shanghai, China infiltrate by immunophenotyping and cytofluorimetry 9Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, (see online supplementary figure S6) revealed that thio- China 10College of Life Sciences, Nankai University, Tianjin, China strepton had no significant effect on the abundance of 11Pôle de Biologie, Paris, France, Hôpital Européen Georges Pompidou, AP-HP, Paris, + + http://jitc.bmj.com/ CD8 CTLs (figure 7A) but reduced the number of CD4 France + CD25 Foxp3 regulatory T cells (Tregs) if combined with 12Suzhou Institute for Systems Medicine, Chinese Academy of Sciences, Suzhou, OXA (figure 7B), thus improving the CTL/Treg ratio China 13 (figure 7C). Moreover, the combination treatment was Department of Women’s and Children’s Health, Karolinska University Hospital, Karolinska Institutet, Stockholm, Sweden particularly efficient in increasing the number of tumor infiltrating neutrophils (see online supplementary figure Acknowledgements We thank David Rubinstein for the PC12- Q74-GFP cell line. S7 and S8). In conclusion, thiostrepton can synergize with on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque Contributors YW, WX and JH conducted experiments and analyzed data. GC, PL OXA to improve tumor control in an immune- dependent and JP assisted with experimental design. ZZ, OK and GK conceived the study and fashion, correlating with reprogramming of the tumor wrote the manuscript. microenvironment. Funding YW and WX are supported by the China Scholarship Council. YW and ZZ were supported by the National Nature Science Foundation of China (81502049 Concluding remarks for YW, 81572955 and 81773357 for ZZ). GK is supported by the Ligue contre Thiostrepton is a fully synthetic oligothiopeptide antibi- le Cancer (équipe labellisée); Agence National de la Recherche (ANR) – Projets blancs; ANR under the frame of E-Rare-2, the ERA- Net for Research on Rare otic that has been approved for veterinary but not human Diseases; AMMICa US23/CNRS UMS3655; Association pour la recherche sur le use, for the treatment of superficial infections including cancer (ARC); Association “Le Cancer du Sein, Parlons- en!”; Cancéropôle Ile- dermatitis and mastitis.35 Preclinical experiment suggest de- France; Chancelerie des universités de Paris (Legs Poix), Fondation pour la that it may inhibit human breast cancer growth and Recherche Médicale (FRM); a donation by Elior; European Research Area Network 36 on Cardiovascular Diseases (ERA- CVD, MINOTAUR); Gustave Roussy Odyssea, the overcome cisplatin resistance. Here, we identified thio- European Union Horizon 2020 Project Oncobiome; Fondation Carrefour; High-end strepton in an unbiased screen designed to identify auto- Foreign Expert Program in China (GDW20171100085), Institut National du Cancer phagy inducers, and subsequently discovered that this (INCa); Inserm (HTE); Institut Universitaire de France; LeDucq Foundation; the LabEx agent can enhance the stigmata of ICD in cultured cells, Immuno- Oncology (ANR-18- IDEX-0001); the RHU Torino Lumière; the Seerave Foundation; the SIRIC Stratified Oncology Cell DNA Repair and Tumor Immune thus enhancing ATP release, CALR exposure and HMGB1 Elimination (SOCRATE); and the SIRIC Cancer Research and Personalized Medicine exodus. Although thiostrepton had some activities on (CARPEM). these ICD hallmarks when used alone, it was particularly Competing interests OK and GK are scientific co- founders of Samsara active when combined with low doses of established ICD Therapeutics.

10 Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 Open access J Immunother Cancer: first published as 10.1136/jitc-2019-000462 on 26 March 2020. Downloaded from

Patient consent for publication Not required. innate and adaptive immunity, and tissue repair. Immunol Rev 2017;280:74–82. ethics approval Animal experiments were conducted in compliance with the EU 17 Apetoh L, Ghiringhelli F, Tesniere A, et al. Toll- like receptor Directive 63/2010). 4-dependent contribution of the immune system to anticancer Provenance and peer review Not commissioned; externally peer reviewed. chemotherapy and radiotherapy. Nat Med 2007;13:1050–9. 18 Zimmermann A, Kainz K, Hofer SJ, et al. 4,4'Dimethoxychalcone: a data availability statement Data are available on reasonable request. All data natural flavonoid that promotes health through autophagy- dependent relevant to the study are included in the article or uploaded as supplementary and -independent effects. Autophagy 2019;15:1662–4. information. 19 Ravikumar B, Duden R, Rubinsztein DC. Aggregate-prone proteins with polyglutamine and polyalanine expansions are degraded by open access This is an open access article distributed in accordance with the autophagy. Hum Mol Genet 2002;11:1107–17. Creative Commons Attribution Non Commercial (CC BY- NC 4.0) license, which 20 Chen G, Xie W, Nah J, et al. 3,4-Dimethoxychalcone induces permits others to distribute, remix, adapt, build upon this work non-commercially, autophagy through activation of the transcription factors TFE3 and and license their derivative works on different terms, provided the original work is TFEB. EMBO Mol Med 2019;11:e10469. properly cited, appropriate credit is given, any changes made indicated, and the use 21 Enot DP, Vacchelli E, Jacquelot N, et al. TumGrowth: an open- is non-commercial. See http://creativecommons. org/ licenses/ by- nc/ 4. 0/. access web tool for the statistical analysis of tumor growth curves. Oncoimmunology 2018;7:e1462431. orCId ids 22 Sauvat A, Chen G, Müller K, et al. Trans-Fats inhibit autophagy induced by saturated fatty acids. EBioMedicine 2018;30:261–72. Wei Xie http://orcid. org/ 0000- 0001- 5554- 2075 23 Kimura S, Noda T, Yoshimori T. Dissection of the autophagosome Oliver Kepp http://orcid. org/ 0000- 0002- 6081- 9558 maturation process by a novel reporter protein, tandem fluorescent- tagged LC3. Autophagy 2007;3:452–60. 24 Donovick R, Pagano JF, Stout HA, et al. Thiostrepton, a new antibiotic. I. In vitro studies. Antibiot Annu 1955;3:554–9. reFerenCes 25 Dutcher JD, Vandeputte J. Thiostrepton, a new antibiotic. II. Isolation 1 Lowe SW, Lin AW. Apoptosis in cancer. Carcinogenesis and chemical characterization. Antibiot Annu 1955;3:560–1. 2000;21:485–95. 26 Jambor WP, Steinberg BA, Suydam LO. Thiostrepton, a new 2 Galluzzi L, Buqué A, Kepp O, et al. Immunogenic cell death in cancer antibiotic. III. In vivo studies. Antibiot Annu 1955;3:562–5. and infectious disease. Nat Rev Immunol 2017;17:97–111. 27 Pankiv S, Clausen TH, Lamark T, et al. p62/SQSTM1 binds directly 3 Garg AD, Dudek-Peric AM, Romano E, et al. Immunogenic cell death. to Atg8/LC3 to facilitate degradation of ubiquitinated protein Int J Dev Biol 2015;59:131–40. aggregates by autophagy. J Biol Chem 2007;282:24131–45. 4 Pfirschke C, Engblom C, Rickelt S,et al . Immunogenic chemotherapy 28 Narain Y, Wyttenbach A, Rankin J, et al. A molecular investigation of true dominance in Huntington's disease. J Med Genet

sensitizes tumors to checkpoint blockade therapy. Immunity Medicale. Protected by copyright. 2016;44:343–54. 1999;36:739–46. 5 Kepp O, Zitvogel L, Kroemer G. Clinical evidence that immunogenic 29 Tallóczy Z, Jiang W, Virgin HW, et al. Regulation of starvation- and cell death sensitizes to PD-1/PD- L1 blockade. Oncoimmunology virus-induced autophagy by the eIF2alpha kinase signaling pathway. 2019;8:e1637188. Proc Natl Acad Sci U S A 2002;99:190–5. 6 Burlion A, Ramos RN, Kc P, et al. A novel combination of 30 Naaktgeboren N, Roobol K, Gubbens J, et al. The mode of action chemotherapy and immunotherapy controls tumor growth in mice of thiostrepton in the initiation of protein synthesis. Eur J Biochem with a human immune system. Oncoimmunology 2019;8:1596005. 1976;70:39–47. 7 Liu P, Zhao L, Pol J, et al. Crizotinib- induced immunogenic cell death 31 Garg AD, More S, Rufo N, et al. Trial watch: immunogenic cell death in non- small cell lung cancer. Nat Commun 2019;10:1486. induction by anticancer chemotherapeutics. Oncoimmunology 8 Kepp O, Semeraro M, Bravo- San Pedro JM, et al. eIF2α 2017;6:e1386829. phosphorylation as a biomarker of immunogenic cell death. Semin 32 Obeid M, Tesniere A, Ghiringhelli F, et al. Calreticulin exposure Cancer Biol 2015;33:86–92. dictates the immunogenicity of cancer cell death. Nat Med 9 Bezu L, Sauvat A, Humeau J, et al. eIF2α phosphorylation is 2007;13:54–61.

pathognomonic for immunogenic cell death. Cell Death Differ 33 Tesniere A, Schlemmer F, Boige V, et al. Immunogenic death of colon http://jitc.bmj.com/ 2018;25:1375-1393. cancer cells treated with oxaliplatin. Oncogene 2010;29:482–91. 10 Giglio P, Gagliardi M, Tumino N, et al. PKR and GCN2 stress 34 Menger L, Vacchelli E, Adjemian S, et al. Cardiac glycosides exert kinases promote an ER stress- independent eIF2α phosphorylation anticancer effects by inducing immunogenic cell death. Sci Transl responsible for calreticulin exposure in melanoma cells. Med 2012;4:ra199. Oncoimmunology 2018;7:e1466765. 35 Nesbitt GH, Fox PR. Clinical evaluation of Panolog cream used 11 Nayagom B, Amara I, Habiballah M, et al. Immunogenic cell death to treat canine and feline dermatoses. Vet Med Small Anim Clin in a combined synergic gene- and immune- therapy against cancer. 1981;76:535–8. Oncoimmunology 2019;8:e1667743. 36 Kwok JM-M, Peck B, Monteiro LJ, et al. FOXM1 confers acquired 12 Ghiringhelli F, Apetoh L, Tesniere A, et al. Activation of the NLRP3 cisplatin resistance in breast cancer cells. Mol Cancer Res on October 19, 2020 at Institut Gustave-Roussy-Bibliotheque inflammasome in dendritic cells induces IL- 1beta- dependent 2010;8:24–34. adaptive immunity against tumors. Nat Med 2009;15:1170–8. 37 Viaud S, Saccheri F, Mignot G, et al. The intestinal microbiota 13 Giglio P, Gagliardi M, Bernardini R, et al. Ecto- Calreticulin is essential modulates the anticancer immune effects of cyclophosphamide. for an efficient immunogenic cell death stimulation in mouse Science 2013;342:971–6. melanoma. Genes Immun 2019;20:509–13. 38 Iida N, Dzutsev A, Stewart CA, et al. Commensal bacteria 14 Li W, Yang J, Luo L, et al. Targeting photodynamic and photothermal control cancer response to therapy by modulating the tumor therapy to the endoplasmic reticulum enhances immunogenic cancer microenvironment. Science 2013;342:967–70. cell death. Nat Commun 2019;10:3349. 39 Zitvogel L, Ma Y, Raoult D, et al. The microbiome in cancer 15 Gardai SJ, McPhillips KA, Frasch SC, et al. Cell-surface calreticulin immunotherapy: diagnostic tools and therapeutic strategies. Science initiates clearance of viable or apoptotic cells through trans- 2018;359:1366–70. activation of LRP on the phagocyte. Cell 2005;123:321–34. 40 Routy B, Le Chatelier E, Derosa L, et al. Gut microbiome influences 16 Bianchi ME, Crippa MP, Manfredi AA, et al. High-mobility group box efficacy of PD-1- based immunotherapy against epithelial tumors. 1 protein orchestrates responses to tissue damage via inflammation, Science 2018;359:91–7.

Wang Y, et al. J Immunother Cancer 2020;8:e000462. doi:10.1136/jitc-2019-000462 11 Zhou et al. Cell Death and Disease (2018) 9:1086 DOI 10.1038/s41419-018-1127-3 Cell Death & Disease

ARTICLE Open Access Oncolysis with DTT-205 and DTT-304 generates immunological memory in cured animals Heng Zhou1,2,3,4,18,LauraMondragón1,2,3,4, Wei Xie1,2,3,4,BrynjarMauseth5,6,7,MarionLeduc1,2,3,4, Allan Sauvat1,2,3,4, Lígia C. Gomes-da-Silva1,2,3,4,8, Sabrina Forveille1,2,3,4, Kristina Iribarren1,2,3,4, Sylvie Souquere9,10, Lucillia Bezu1,2,3,4, Peng Liu 1,2,3,4,LiweiZhao1,2,3,4, Laurence Zitvogel9,11,12,13, Baldur Sveinbjørnsson5,14,15, J. Johannes Eksteen16, Øystein Rekdal5,14, Oliver Kepp 1,2,3,4 and Guido Kroemer1,2,3,4,15,17

Abstract Oncolytic peptides and peptidomimetics are being optimized for the treatment of cancer by selecting agents with high cytotoxic potential to kill a maximum of tumor cells as well as the capacity to trigger anticancer immune responses and hence to achieve long-term effects beyond therapeutic discontinuation. Here, we report on the characterization of two novel oncolytic peptides, DTT-205 and DTT-304 that both selectively enrich in the lysosomal compartment of cancer cells yet differ to some extent in their cytotoxic mode of action. While DTT-304 can trigger the aggregation of RIP3 in ripoptosomes, coupled to the phosphorylation of MLKL by RIP3, DTT-205 fails to activate RIP3. Accordingly, knockout of either RIP3 or MLKL caused partial resistance against cell killing by DTT-304 but not DTT-205. In contrast, both agents shared common features in other aspects of pro-death signaling in the sense that their cytotoxic effects were strongly inhibited by both serum and antioxidants, partially reduced by lysosomal inhibition

1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; with bafilomycin A1 or double knockout of Bax and Bak, yet totally refractory to caspase inhibition. Both DTT-304 and DTT-205 caused the exposure of calreticulin at the cell surface, as well as the release of HMGB1 from the cells. Mice bearing established subcutaneous cancers could be cured by local injection of DTT-205 or DTT-304, and this effect depended on T lymphocytes, as it led to the establishment of a long-term memory response against tumor-associated antigens. Thus, mice that had been cured from cancer by the administration of DTT compounds were refractory against rechallenge with the same cancer type several months after the disappearance of the primary lesion. In summary, DTT-205 and DTT-304 both have the capacity to induce immunotherapeutic oncolysis.

Introduction Peptides can be synthetically generated and potentially provide pharmacological leads or final agents for multiple purposes. In the field of cancer research, so-called onco- Correspondence: Oliver Kepp ([email protected])or lytic peptides have been conceived with the objective of Guido Kroemer ([email protected]) selectively killing tumor cells. This may be achieved by 1 Metabolomics and Cell Biology Platforms, Gustave Roussy Comrehensive fusing targeting sequences (that interact with proteins Cancer Institute, Villejuif, France 2Equipe 11 labellisée Ligue contre le Cancer, Centre de Recherche des specifically expressed on the surface of malignant cells or Cordeliers, INSERM U, 1138 Paris, France tumor vasculature) with effector sequences (that cause the – Full list of author information is available at the end of the article. lysis of the targeted cell type)1 3, or alternatively by local These author contributed equally: Heng Zhou, Laura Mondragón, Wei Xie, Brynjar Mauseth administration of the oncolytic peptide into the neoplastic Edited by H.-U Simon

© The Author(s) 2018 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a linktotheCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 2 of 14

– lesion, notably by direct injection4 8. Lytic peptides One mechanism that appears to be particularly impor- usually combine two physicochemical features, namely tant for triggering antineoplastic immune responses is the lipophilicity and cationic charge, meaning that they con- induction of immunogenic cell death (ICD), tied to the tain hydrophobic and positively charged amino acids cell surface exposure of the DAMP calreticulin, an “eat- (mostly arginine and lysine residues) that may be inter- me” signal for dendritic cell (DC) precursors, as well as spersed in a way to create an amphipathic structure9.Itis the release of several DAMPs including ATP and HMGB1 thought that this design facilitates the enrichment of the that attract DCs into the proximity of cancer cells and peptides within the cell and, in particular, the mito- activate them for optimal tumor antigen cross- – chondrial matrix as a result of their electrophoretic dis- presentation to cytotoxic T lymphocytes17 21. Of note, tribution following the Nernst equation10, hence allowing both LTX-315 and LTX-401 are able to stimulate ICD, them to mediate local membrane-permeabilizing effects meaning that cancer cells killed by these agents in vitro that compromise organellar and cellular integrity11. can elicit protective antitumor immune responses upon Although the overall molecular design of oncolytic their inoculation in vivo and that local administration of peptides follows the rules exposed above, there may be either LTX-315 or LTX-401 triggers signs of an immune – major, hitherto unexplained differences in the subcellular response in established tumors4 8,13,22,23. For this indi- distribution of such agents. For example, the oncolytic cation, LTX-315 is currently undergoing clinical trials peptide LTX-315 follows a classical pattern of mito- (NCT01986426; clinicaltrials.gov). chondrial distribution causing early permeabilization of Based on the aforementioned premises, the develop- this organelle with the dissipation of the mitochondrial ment of oncolytic peptides is currently expanding. transmembrane potential and the release of inter- Recently, a new class of oncolytic peptides (DTT peptides) membrane proteins including cytochrome c and DIABLO with antilymphoma activity were reported24. Several through the outer membrane11,12. In sharp contrast, new analogues were designed and tested against hepato- oncolytic, LTX-401, an amphipathic β(2,2)-amino acid cellular and colorectal carcinoma cell lines (Table 1). In derivative, tends to enrich in the Golgi apparatus and this report, we analyzed the cellular and immunological dismantles the organelle before mitochondrial integrity is mode of action of two novel oncolytic DTT-analogues, compromised4,8,13. This mitochondrial step of the cell DTT-205 and DTT-304. Here, we report that both death cascade appeared to be important for cell death agents selectively target lysosomes and share some but induction by both LTX-315 and LTX-401, because not all of their cytotoxic mechanisms of action. Impor- knockout of the proapoptotic multidomain BCL2 family tantly, both DTT-205 and DTT-304 can cause the total proteins BAX and BAK attenuated cell killing by both eradication of established mouse cancers as they induce LTX-315 and LTX-4018,11. These differences and simila- a potent, therapeutically relevant antitumor immune rities illustrate the complexity of pro-death signaling response. mediated by agents that apparently share comparable physicochemical properties. Results and discussion Over the past few years, it has become increasingly Morphological effects of DTT-205 and DTT-304 clear that anticancer drugs should not only be optimized When added to human cancer cells, DTT-205 or DTT- with respect to their capacity to kill a significant (and 304 stimulated a necrosis-like disruption of cellular ideally close-to-total) fraction of malignant cells and morphology. In cells in which the plasma membrane was hence to “debulk” the primary tumor and its metastases. still intact and hence retaining the cytoplasm, lipid dro- Rather, antineoplastics should also be able to stimulate plets in the cytoplasm were a prominent morphological anticancer immune responses, an effect that can be feature (Fig. 1a) that was induced in a dose-dependent achieved by a variety of mechanisms, namely, (i) killing fashion (Fig. 1b, c), as determined by transmission elec- of cancer cells in a way that leads to the release or tron microscopy. Staining with the red fluorescent lipo- exposure of danger-associated molecular patterns philic dye Nile red confirmed the formation of lipid (DAMPs) that will alert the innate and later the acquired droplets in the cytoplasm of cells treated with DTT-205 or immune system to recognize tumor-associated antigens, DTT-304 that occurred in a time and dose-dependent (ii) direct stimulation of immune effectors or subversion fashion (Fig. 1d–f). Of note, these effects were only – of immunosuppressive mechanisms14 16.Indeed,itis obtained when cellular metabolism and membrane traf- difficult to conceive that any kind of cancer treatment ficking were active at 37 °C, yet not at lower temperatures would be able to eliminate the very last malignant cell such as 14 or 22 °C (Fig. 1g, h). As for other oncolytic (among 1010 to 1012 neoplastic cells) and hence to peptides such as LTX-31525 and LTX-4018, DTT-205 and achieve definitive cure, a scenario that would apply if DTT-304 lost their cytotoxic potential in the presence of anticancer immunosurveillance was not involved in serum (Supplementary Figure S1). However, DTT-205 long-term treatment outcomes. and DTT-304 differed from LTX-315 and LTX-401 in the

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 3 of 14

Table 1 Peptide design and IC50 values of DTT peptides against various cell lines

Compound JM1 BEL-7402 HEPG2 HepaRG CT-26 HT-29 CC531 MRC-5 HUVEC hRBCs

DTT-106 17.5 ± 4.2 27.1 ± 0.5 27.05 ± 0.6 24.1 9.7 ± 0.4 25.8 ± 1.2 26.7 ± 4.3 28.6 ± 2.4 29.3 ± 0.2 ND DTT-201 26 ± 3.0 > 31 > 31 ND 10.8 ± 1.3 > 31 > 31 > 31 > 31 ND DTT-202 15.9 ± 2.3 > 31 > 31 23.3 ± 0.9 7.4 ± 1.6 23.4 ± 0.5 16.5 ± 1.4 23.3 ± 2.2 29.9 ± 1.2 ND DTT-203 13.4 ± 5.9 19.0 ± 1.3 23.4 ± 3.1 22.6 8.1 ± 0.6 26.5 ± 0.9 > 31 30.5 ± 1.7 29.5 ± 0.8 ND DTT-204 16.4 ± 1.0 22.9 ± 0.8 > 31 ND 8.1 ± 1.2 16.5 ± 3.4 25.2 ± 0.4 24.0 ± 3.0 27.5 ± 2.5 ND DTT-205 6.9 ± 1.7 10.3 ± 0.3 23.6 ± 0.7 9.0 ± 0.3 8.2 ± 0.5 9.8 ± 0.9 13.0 ± 0.3 11.2 ± 0.1 10.6 ± 2.6 526–614 DTT-122 > 31 > 31 > 31 ND ND > 31 ND ND ND ND DTT-301 26.0 ± 2.9 > 31 > 31 ND 6.7 ± 0.6 > 31 > 31 > 31 > 31 ND DTT-302 16.3 ± 1.2 > 31 24.5 ± 3.4 21.1 10.1 ± 1.5 > 31 > 31 > 31 > 31 ND DTT-303 13.0 ± 0.8 19.5 ±2.4 20.2 ± 3.6 28.5 6.0 ± 0.5 > 31 25.8 ± 5.8 26.7 ± 3.4 > 31 ND DTT-304 13.0 ± 1.3 16.4 ± 0.2 21.4 ± 0.3 15.0 ± 2.1 8.6 ± 2.6 21.1 ± 1.7 24.3 ± 6.7 29.0 ± 2.0 27.9 ± 2.1 > 928 DTT-305 12.4 ± 2.0 15.8 ± 0.05 20.6 ± 2.4 14.8 ± 0.4 7.8 ± 2.0 9.0 ± 0.8 15.7 ± 0.3 12.4 ± 0.2 15.0 ± 0.3 ND DTT-306 8.7 ± 0.7 11.6 ± 0.5 20.1 ± 2.8 12.5 7.5 ± 1.7 4.7 ± 0.9 13.1 ± 2.1 17.5 ± 2.3 16.4 ± 1.5 ND hRBCs human red blood cells, ND not determined Data represents two or more independent experiments conducted in triplicates (IC50 μM ± SD). Standard concentration gradient 1–100 μg/ml, equaling ∼0.3–31 μM

sense that only the DTT peptides, but not the LTX For this, we used BAFA1 to avoid their lysosomal compounds, induced lipid droplets (Supplementary accumulation (Fig. 2b), finding that this maneuver par- Figure S2). tially reduced cell killing by DTT-205 or DTT-304 (Fig. 2h, i). Lysosomal tropism of DTT-205 and DTT-304 Intrigued by the peculiar morphology of DTT-treated Apoptosis-related signaling induced by DTT-205 and cells, we investigated the subcellular distribution of DTT-304 these agents, taking advantage of DTT-205 and DTT- Caspase-3 activation is one of the biochemical hall- 304 that had been modified to attach a blue-fluorescent marks of apoptosis27.Whileasignificant fraction of cells moiety to the molecules. Cells that express fluorescent treated with the positive control, the pan-tyrosine kinase biosensors in the nucleus (histone H2B fused to red inhibitor staurosporine stained positively with an anti- fluorescent protein, RFP), endoplasmic reticulum (cal- body recognizing the proteolyticallymaturefragmentof reticulin, CALR fused to green fluorescent protein, caspase-3, only a minor fraction of cells exhibited cas- GFP), Golgi apparatus (GALT1 fused to GFP), mito- pase activation after treatment with DTT-205 or DTT- chondria (DIABLO fused to GFP), or lysosomes 304 (Fig. 3a–d). Both DTT-205 and DTT-304 caused (LAMP1 fused to GFP) were incubated with the fluor- nuclear shrinkage, though without the formation of escent derivatives of DTT-205 and DTT-304. Confocal apoptotic bodies, and a decrease in the number of fluorescence microscopy revealed that both DTT-205 analyzable cells (Fig. 3a–c, e, f; Supplementary Fig. 3). and DTT-304 co-localized with the lysosomal The pan-caspase inhibitor z-VAD-fmk failed to prevent marker LAMP1-GFP but not with any other organellar cell killing by DTT-205 or DTT-304 (Fig. 3g, h), as did probe (Fig. 2a–d). Of note, this colocalization was the ferroptosis inhibitor ferrostatin-1. In contrast, two abrogated upon preincubation of the cells with the antioxidants, namely N-acetylcysteine and glutathione vacuolar ATPase inhibitor bafilomycin A1 (BAFA1), strongly reduced the cytotoxic activity of DTT-205 and which is known to abolish lysosomal acidification DTT-304 (Fig. 3g–j, Supplementary Figure 4). Apopto- (Fig. 2b–d). As other lysosomotropic agents26, DTT-205 tic signaling also involves the mitochondrial membrane and (less so) DTT-304-affected lysosomal stability and permeabilization that often depends on the expression let to a reduction in measurable organellar surface of proapoptotic multidomain members of the BCL2 area as detected by LysoTrackerTM (Fig. 2e–g). Next, family such as BAX and BAK. Indeed knockout of BAX, we determined whether the lysosomotropism of DTT- alone or together with BAK, reduced killing by DTT-205 205 and DTT-304 might explain their cytotoxic activity. or DTT-304 to a variable extent (Fig. 3k, l). In synthesis,

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 4 of 14

A B s

Ctr DTT-205 DTT-304 tel 2 DTT-205 pordd 1.5 **

i **

pil 1 f o#naeM 0.5

0 5 Ctr 2.5

C 1.25 s te 2 DTT-304

L lpor L **

1.5 ddi

L * p

i 1 L l L fo#naeM L 0.5 L 0 5 Ctr 2.5 D EF1.25

DTT-205 Nile-Red Hoechst aera aera 3.5 DTT-205 3.5 DTT-304

3 6 h 24 h ** 3 6 h 24 h

+ + deReliN.mroN deReliN.mroN 2.5 2.5 * * * * 2 * 2 * * * * Ctr 6 h 24 h DTT-304 Nile-Red Hoechst 1.5 1.5 1 1 0.5 0.5 0 0 5 5 5 5 10 10 10 10 Ctr Ctr 2.5 2.5 2.5 2.5 1.25 0.65 0.65 1.25 0.65 0.65 1.25 Ctr 6 h 24 h 1.25 G H

DTT-205 14°C 22°C 37°C DTT-304 14°C 22°C 37°C aera a ** era 2 2 **

* * +

+ * deReliN.mro deReliN 1.5 1.5

1 1

. 0.5 0.5 mroN

0 N 0 Ctr 0.651.252.5 5 10 Ctr 0.651.252.5 5 10

Fig. 1 DTT peptides induce the formation of lipid droplets. Human osteosarcoma U2OS cells were treated with 2.5 µM of either DTT-205 or DTT- 304 for 6 h. Cells were fixed and subjected to electron microscopy. The overview micrographs in the upper panel depict necrotic morphologies (N) of some of the treated cells and the high magnification micrographs in the lower panel show the formation of lipid droplets (L) in intact cells. Size bars equals 10 µm (upper panel) and 1 µm (lower panel). Representative images (a) and quantifications (b, c) are depicted (mean ± SD of a minimum of five view fields). The formation of lipid droplets in response to increasing doses from 0.65 to 10 µM DTT peptides was quantified by means of the lipophilic dye Nile Red at 6 h and 24 h post treatment in epifluorescence microscopy. Representative images (d) and quantifications (e, f) are shown (mean ± SD of triplicate assessments, Student’s t test, p < 0.5, **p < 0.01). Temperature dependency of the lipid droplet formation was assessed by keeping the cell cultures upon treatment with 0.65 to 10 µM DTT-205 (g) and DTT-304 (h) at the indicated temperature for 6 h before Nile Red staining. Increased number of lipid droplets at physiological temperature is indicative for an underlying active biochemical reaction (mean ± SD of triplicate assessments, Student’s t test, *p < 0.5, **p < 0.01)

it appears that pro-oxidant and BAX/BAK-dependent Necroptotic signaling induced by DTT-205 and DTT-304 processes, but not caspase activation, participate to cell To investigate the potential role of necroptotic signaling killing by DTT-205 or DTT-304. in cell death induction by DTT compounds, we took

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 5 of 14

A E DTT-205 Organelle Merge DTT-304 Organelle Merge Hoechst LysoTracker DTT-205 CALR DTT-304 Ctr0.65 1.25

F G aera aera DTT-205 DTT-304

4 4

+ +

rek r

3 ekcarT 3

c arT

DIABLO GALT1 2 2 o

osyL.le * syL.leR 1 1 LAMP1 H2B R 0 0 Ctr Ctr 0.65 1.25 0.65 1.25

B H DTT-205 DTT-205 100 +BAFA1 BAFA1 BAFA1 80

DTT-205 Organelle Merge DTT-304 Organelle Merge ytilibaiv% 60 * * 40 * ** LAMP1 20 *** *** 0 Ctr 0.651.252.5 5 10 C D I DTT-205 DTT-205 DTT-304 DTT-304 DTT-304 DTT-304

0.5 +BAFA1 0.5 +BAFA1 100 +BAFA1 ec * ec

0.4 0.4 80

nerrucooc n

y tiliba errucooc.ler * 0.3 0.3 60 *

i *

0.2 0.2 v% 40 .ler 0.1 0.1 20 *** *** *** 0 0 0 *** Ctr 0.651.252.5 5 10 ER ER GA GA Nuc Nuc Mito Mito Lyso Lyso Lyso Lyso Fig. 2 Organellar targeting of DTT-205 and DTT-304. Human osteosarcoma U2OS cells stably expressing the nuclear marker histone H2B together with red fluorescent protein (RFP), the ER marker calreticulin (CALR) labeled with green fluorescent protein (GFP), galactose-1-phosphate uridylyltransferase GALT1, a marker of the Golgi apparatus fused to GFP, DIABLO co-expressing GFP as an indicator for mitochondria and LAMP1-GFP as a lysosomal marker were treated with 1.25 µM Pacific blue-labeled DTT peptides in the presence or absence of lysosomal acidification that was blocked or not with bafilomycin A1 (BAFA1). Both DTT-205 and DTT-304 accumulated in lysosomal structures, an effect that was inhibited with BAFA1. Representative images of confocal assessment (a, b; size bar equals 5 µm) and relative cooccurrence of Pacific Blue label with organellar markers was assessed (c, d; mean ± SEM of a minimum of five view fields). Wild-type U2OS cells were stained with LysoTracker green and the decrease in lysosomal content was assessed upon treatment 0.65 or 1.25 µM DTT peptides for 6 h by epifluorescence microscopy. Representative images (e) and quantifications (f, g) are depicted (size bar equals 10 µm; mean ± SD of triplicate assessments, Student’s t test, *p < 0.5). Viability was measured in living cells by means of the exclusion dye propidium iodide 6 h post treatment with 0.65 to 10 µM DTT-205 (h) and DTT-304 (i) in the presence or absence of BAFA1 by live cell microscopy (mean ± SD of triplicate assessments, Student’s t test, *p < 0.5, **p < 0.01, ***p < 0.001). BAFA1 partially decreased the cytotoxic effect of the DTT peptides

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 6 of 14

A B

DTT-205 Hoechst CASP3active DTT-304 Hoechst CASP3active 6 h 6 h 24 h 24 h

5 10 5 10 5 10 5C10 trSTS

C D G

+ + ev evitca DTT-205 6 h 24 h DTT-304 6 h 24 h DTT-205 25 itca 25 100

*** *** 3PSACslle

3PSACsllec 20 20 80

y

t

i l 15 15 iba 60 **

10 10 iv% 40 *** *** cfo%

fo% 5 5 20 ND ND 0 ND ND 0 ND 0 5 5 5 5

10 10 10 10

rtC rt Ctr Ctr

2.5 2.5 2.5 2.5

1-REF

DAV-Z HSG C

STS STS STS STS

0.65 1.25 0.65 1.25 0.65 0.65 1.25 1.25

C AN

E F H DTT-205 6 h 24 h DTT-304 6 h 24 h DTT-304

250 250 )

) 100

2 2

mµ(aeraraelcuN mµ(aeraraelcuN 200 ** 200 ** 80

* ** ** y t

** ilibaiv% 150 150 *** *** 60 ** *** *** ** 100 100 40 *** *** 50 50 20 0 0 0 5 5 5 5

10 10 10 10

rtC rtC 1-RE Ctr Ctr

2.5 2.5 2.5 2.5

HSG D CAN

STS STS STS STS

0.65 0.65 0.65 1.25 0.65 1.25 1.25 1.25

A

V

-Z F

I K DTT-205 WT Bax-/- Bax/Bak-/- DTT-205 Ctrl NAC 100 100 80 *

ytil * 80

y

ti i

60 l * i

ba 60 * bai i ** ** v 40 *** ** v% 40

% ** 20 *** *** 20 *** *** 0 0 Ctr 0.651.252.5 5 10 Ctr 0.651.252.5 5 10

J L DTT-304 WT Bax-/- Bax/Bak-/- DTT-304 Ctrl NAC 100 100 80 *

ytilib 80 ** ytilibai

60 60 * aiv% 40 ***

v ** *** 40 20 % 20 0 0 Ctr 0.651.252.5 5 10 Ctr 0.651.252.5 5 10

Fig. 3 (See legend on next page.)

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 7 of 14

(see figure on previous page) Fig. 3 Cell death induced by DTT-205 and DTT-304. Human osteosarcoma U2OS cells were treated with 0.65 to 10 µM DTT peptides for the indicated time and following assessed for the activation of caspase-3 (CASP3). The pan-kinase inhibitor staurosporine (STS) was used as positive control. Representative images (a, b) and quantifications of CASP3 activation (c, d) are depicted while pyknosis was assessed by measuring the decrease in Hoechst 33342-stained nuclear area (e, f) by conventional microscopy (size bar equals 10 µm; mean ± SD of triplicate assessments, Student’s t test, *p < 0.5, **p < 0.01, ***p < 0.001). In order to assess the cell death modality, U2OS cells were pretreated with the ferroptosis inhibitor ferrostatin-1 (FER-1), the pan-caspase inhibitor zVAD-fmk (zVAD) or the antioxidants N-actylcystein (NAC) and reduced glutathione (GSH) before the addition of 2.5 µM of DTT peptides (g, h). Exclusively the inhibition of reactive oxygen species (ROS) generation decrease the cytotoxic effects of both DTT-205 and DTT-304 in a dose-dependent fashion (i, j, mean ± SD of triplicate assessments, Student’s t test, *p < 0.5, **p < 0.01, ***p < 0.001). Mouse embryonic fibroblasts (MEFs) that were either wild type, single- or double knockout for the proapoptotic proteins Bax and/or Bak were treated with 0.65 to 10 µM of DTT-205 or DTT-304 and viability was assessed by means of an exclusion dye in epifluorescence microscopy (k, l; mean ± SD of triplicate assessments. Student’s t test, *p < 0.5, **p < 0.01, ***p < 0.001). Single as well as double knockouts were partially resistant to DTT-peptide- induced cell death advantage of a biosensor cell line that lacks MLKL (and oncolysis in most surviving mice. Of note, prior depletion hence cannot undergo necroptosis) yet expresses a RIP3- of T cells by injection of specific antibodies blocking CD4 GFP fusion protein that can be monitored for its aggre- and CD8 abolished these anticancer effects. These gation in the cytoplasm within so-called “necropto- immune-dependent anticancer effects of DTT-205 and somes”28,29. As a positive control, a combination of tumor DTT-304 were evident both when tumor growth kinetics necrosis factor-α (TNFα), DIABLO mimetic, and z-VAD- and overall survival of mice were monitored (Fig. 6a–g). fmk (collectively abbreviated as ‘TSZ’) induced full Rechallenge of mice that had been cured from MCA205 necroptosome activation. DTT-205 was unable to induce cancers with the same tumor type (MCA205) was this phenomenon, while DTT-304 turned out to cause incompatible with the development of cancers, although necroptosome activation in a time-dependent and antigenically distinct TC-1 lung adenocarcinoma grew in concentration-dependent fashion (Fig. 4a–d). Accord- most (eight out of nine) cases (Fig. 6h–k). Very similar ingly, cells treated with DTT-304 (but not DTT-205) results were obtained when TC-1 carcinomas instead of manifested the phosphorylation of the RIP3 substrate MCA205 sarcomas were treated with DTT-205 or DTT- MLKL (Fig. 4e, f). Cells rendered deficient for RIP3 or 304. All animals that could be monitored were cured from MLKL30 became partially resistant against cell killing by their cancers (Supplementary Figure 7A–E), and none of DTT-304 (but not DTT-205) (Fig. 4g, h), supporting the the cured animals that were rechallenged with TC-1 notion that DTT-304 (but not DTT-205) can activate a developed cancers, although MCA205 tumors readily cell death signaling pathway that involves the necroptotic developed (Supplementary Figure 7F–H). Animals bear- cascade. ing TC-1 tumors on both flanks that were treated with the oncolytic peptides only on one side, but not on the other Immunogenic cell death triggered by DTT-205 and depicted abscopal effects that led to the reduction of DTT-304 tumor growth in the distant lesion (Supplementary Fig- ICD is characterized by the cellular release/exposure of ure S8). These results indicate that immune-dependent DAMPs (such as CALR, HMGB1 and type-1 interferons) cancer cure mediated by DTT-205 or DTT-304 is coupled that make dead-cell antigens recognizable to the immune to the establishment of systemic immunity that mediates system. Both DTT-205 and DTT-304 triggered CALR long-term immune memory specific for the eradicated exposure on the cell surface (Fig. 5a, b), nuclear HMGB1 cancer type. exodus (Fig. 5c, d), and the transcription of genes coding for type-1 interferons (Fig. 5e, f). Nevertheless, the treat- Concluding remarks ment with oncolytic peptides did not lead to an increase The results of this study suggest that the oncolytic in the expression of major histocompatibility complex peptides DTT-205 and DTT-304 both target lysosomes as class I molecules (MHC-I) (Supplementary Figure 5) and indicated by their selective accumulation in this organelle. depicted a rather unspecific cytotoxicity on both cancer To this respect, DTT-205 and DTT-304 are rather dis- and immune cells (Supplementary Figure 6). Based on tinct from other oncolytic peptides that often target the these in vitro characteristics of DTT compound-induced mitochondria. Despite their common lysosomotropic cell death, we next investigated whether these agents nature, DTT-205 and DTT-304 are to a certain extent might mediate anticancer effects through the stimulation different in their mode of action. While cell death induced of an anticancer immune response. For this, established by both agents strongly depends on the production of MCA205 fibrosarcomas growing on immunocompetent reactive oxygen species exclusively DTT-304 shows pro- haploidentical C57BL/6 mice were injected with either necroptotic features, such as the formation of the DTT-205 or DTT-304, which both caused complete necroptosome and the phosphorylation of MLKL.

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 8 of 14

AB DTT-205 Hoechst RIP3-GFP

+ 80 DTT-205 6 h 24 h

rgg A

3PIRsll 60 6 h *** 40

*** ec

fo 20 24 h ND ND ND % 0 5 5 10 Ctr 0.65 1.25 2.5 5 10 TSZ 10 Ctr 2.5 2.5 0.65 1.25 0.65 1.25 TSZ TSZ

C DTT-304 Hoechst RIP3-GFP D

+ 80 DTT-304 6 h 24 h

rgg *** *** A

3P 60 6 h

I *** *** Rsllec 40 *** ** ** fo 20 * 24 h * * % 0 5 Ctr 0.65 1.25 2.5 5 10 TSZ 5 10 10 Ctr 2.5 2.5 1.25 1.25 0.65 0.65 TSZ TSZ

EFDTT-205 DTT-304 p-MLKL 54 kDa p-MLKL 54 kDa MLKL 54 kDa MLKL 54 kDa 5 5 10 10 Ctr Ctr 2.5 2.5 TSZ TSZ 0.65 1.25 0.65 1.25 G H DTT-205 -/- -/- DTT-304 -/- -/- 100 WT RIP3 MLKL 100 WT RIP3 MLKL

80 80

ytil y

t ilib

ibaiv% 60 * * * * 60 * * a 40 iv% 40 ** ** *** *** 20 *** 20 *** 0 *** *** *** 0 Ctr 0.651.252.5 5 10 TSZ Ctr 0.651.252.5 5 10 TSZ

Fig. 4 Necroptotic traits induced by DTT-304 but not by DTT-205. Human colon carcinoma HT-29 cells knockout for MLKL and stably expressing RIP3 coupled to green fluorescent protein (GFP) were treated with 0.65 to 10 µM DTT peptides for the indicated time and following assessed for the aggregation of RIP3 indicative for necroptosome formation by epifluorescence microscopy. The combination of TNFα (T), SMAC/DIABLO mimetic peptide B6 (S), and the pan-caspase inhibitor z-VAD-fmk (Z) was used as positive control for the induction of necroptosis. Representative images (a, c) and quantifications of RIP3 aggregates (b, d) are depicted (size bar equals 10 µm; mean ± SD of triplicate assessments, Student’s t test, *p < 0.5, **p < 0.01, ***p < 0.001). Downstream MLKL activation was visualized with phosphoneoepitope-specific antibody (e, f) Of note, exclusively DTT-304 but not DTT-205 depicted traits of necroptosis. Murine lung cancer TC-1 cells that were CRISPR gene edited in RIP3 and MLKL were treated with 0.65 to 10 µM of DTT-205 or DTT-304 for 6 h and viability was assessed by means of an exclusion dye (g, h; mean ± SD of triplicate assessments, Student’s t test, *p < 0.5, **p < 0.01, ***p < 0.001). Knockouts of RIP3 and MLKL were partially resistant to DTT-304 yet not to DTT-205-induced cell death

Nevertheless, analogue to similar oncolytic peptides which underlines the immunotherapeutic potential of such as LTX-315 and LTX-4014-8,12,23 both DTT-205 and oncolytic peptides in general and DTT-205 and DTT-304 DTT-304 triggered the emission of pro-immunogenic in particular. DAMPs in vitro and activated anticancer immune responses in vivo. Complete tumor lysis by DTT peptides Materials and methods required repeated intratumoral injections, which might Chemicals and cell cultures become a challenging aspect during clinical translation Cell culture media and supplements were purchased but might facilitate to render the tumor into its own from Thermo Fisher Scientific (Carlsbad, CA, USA) and vaccine31. Thus, in this study animals that were cured chemicals came from Sigma-Aldrich (St. Louis, MO, USA) from primary tumors generated long-term immunological with the exception of LTX-315, LTX-401, DTT-205, and memory that mediated the rejection of isogenic tumors, DTT-304 that were provided by Lytix Biopharma (Oslo,

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 9 of 14

A B

DTT-205 6 h 24 h DTT- 304 6 h 24 h

- -

25 IP 25

PI

+ + RLACsllecfo RLACsllecfo 20 20 15 ** * 15

10 10 ** * 5 * 5 ** ** * ND

* ND % % 0 0

rtC 5 5

56.0

52.1 5.2

5 01

52.1 56.0 5.2 01 5 10 10 Ctr 2.5 2.5 0.65 1.25 0.65 1.25 C D DTT-205 6 h 24 h DTT- 304 6 h 24 h 600 600 ** ** **

450 450 ** ** ** 300 300 * * 150 150 * * * 0 0

rtC HMGB1 release (ng/ml) 5 HMGB1 release (ng/ml) 5

56.0

52.1

5.2 5 01

52.1 56.0 5.2 5

01 10 10 Ctr 2.5 2.5 0.65 1.25 0.65 1.25 E F DTT-205 Ifna1 Ifnab1 DTT- 304 Ifna1 Ifnab1 12 *** 12

9 9

6 * 6 * * * 3 3 * * Rel. mRNA content Rel. mRNA content 0 0

rtC 5 5

56.0

52.1 5.2

5 01

56.0

52.1 5.2 5 01 10 10 Ctr 2.5 2.5 0.65 1.25 0.65 1.25 Fig. 5 DAMP release from tumor cells treated with DTT-205 and DTT-304. The exposure of calreticulin (CALR) in human osteosarcoma U2OS cells was measured by flow cytometry using polyclonal anti-CALR antibody while excluding cells that lost cytoplasmic membrane integrity and thus incorporated the exclusion dye propidium iodide (PI) (a, b). Exodus of high mobility group box 1 (HMGB1) from the cells into cell culture supernatants was monitored by HMGB1-specific enzyme-linked immunosorbent assay ELISA. Absorbance was measured and concentrations were calculated based on standards (c, d). Both CALR and HMGB1 were emitted by both DTT-205 and DTT-304 in a dose-dependent fashion. The production of type I interferon (IFN) was measured in by reverse transcription quantitative real time polymerase chain reaction qRT-qPCR from purified mRNA of cells treated with DTT compounds (e, f) (mean ± SD of triplicate assessments; Student’s t test, *p < 0.5, **p < 0.01, ***p < 0.001). In summary some ICD DAMPs were release in response to DTT-205 and DTT-304

Norway). Plasticware was obtained from Greiner Bio-One expressing GALT1-GFP, H2B-RFP, CALR-GFP, DIABLO- (Monroe, CA, USA), primary antibody (cleaved caspase-3; GFP or LAMP1-GFP, and HT-29 stably expressing RIP3- ® #9661) came from Cell Signaling (Danvers; MA; USA), and GFP28 cells were cultured in Glutamax -containing DMEM AlexaFluor-coupled secondary antibody from Thermo medium supplemented with 10% fetal calf serum (FCS), Fisher Scientific. Anti-CD4 and anti-CD8 antibodies for and 10 mM HEPES. Cells were grown in a humidified in vivo use were obtained from BioXcell (West Lebanon, incubator at 37 °C under a 5% CO2 atmosphere. NH, USA). Mouse embryonic fibroblast (MEF), mouse Murine thymocytes and splenocytes were obtained ® sarcoma cells MCA205, murine lung cancer TC-1, by dispersing organ extracts in Glutamax -containing human osteosarcoma U2OS, wild type or stably DMEM supplemented with 50 µM 2-mercaptoethanol

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 10 of 14

A MCA205 s.c. Tumor monitoring Peptides i.t. Antibodies i.p. -7 0 1530 45 60 Days of treatment

(C57BL/6, naive)

Ctr DTT-205 DTT-304 DTT-205 αCD4/CD8 DTT-304 αCD4/CD8 B C D 200 n=5 200 200 2 2 2 150 150 150

100 100 100

50 50 50 *** Tumor size mm Tumor Tumor size mm Tumor n=5 size mm Tumor n=5 0 0 0 *** 0510 20 30 40 0600510 20 30 40 0600510 20 30 40 060 E F G 200 200 100 ** 2 2 150 150 75 ** 100 100 50

50 50 % survival 25

Tumor size mm Tumor n=4 size mm Tumor n=3 0 0 0 0510 20 30 40 0600510 20 30 40 0600510 20 30 40 060 Days Days Days

MCA205 rechallenge TC-1 challenge H I K 200 DTT-205 200 DTT-304 MCA205 rechallenge 2 TC-1 challenge 2 150 150

100 100

(C57BL/6, cured from MCA205) 50 50

Tumor size mm Tumor n=6 size mm Tumor n=3 0 0 0510 20 30 40 0600510 20 30 40 060 Days Days

Fig. 6 In vivo activity of DTT-205 and DTT-304 in immunocompetent animals. Mouse fibrosarcoma cells were inoculated subcutaneously in syngenic C57BL/6 animals and arising tumors were treated when palpable with repeated injections of DTT-205 or DTT-304 in the presence of absence of intraperitoneally injected CD4/CD8 blocking antibody (a). Both DTT-205 and DTT-304-induced efficient oncolysis. Immunocompetent animals depicted long-term effects whereas tumors recurred soon after treatment in immunecompromised animals upon T-cell depletion an effect that is reflected in tumor growth (b-f) and overall survival (g) (Chi2 test, **p < 0.01, ***p < 0.001). Rechallenge of animals cured from MCA205 fibrosarcoma with MCA205 several weeks after the initial therapy on the contralateral and challenge with syngenic mouse TC-1 lung cancer cells on the ipsilateral side resulted in efficient rejection of MCA205 but aggressive tumor growth of TC-1 (h-k). DTT-205 and DTT-304 caused the generation of immunological memory that sufficed in rejection isogenic tumors and FCS and HEPES as described above. Erythrolysis was peptides. Supernatants were collected and centrifuged at conducted using ACK lysis buffer (Thermo Fisher 500 × g for 5 min in order to remove cellular debris. Scientific). Released HMGB1 protein content was quantified in supernatants by ELISA (#ST51011; IBL, Hamburg, Ger- HMGB1 release assays many) according to the manufacturer’s recommendations, Cells were seeded in 24-well plate and let adhere and and samples were measured with a Paradigm E3 multi- adapt overnight, before they were treated with the DTT label reader (Molecular Devices).

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 11 of 14

Western blot uptake of the exclusion dye in a minimum of four view Half a million cells were harvested from six-well plates fields per well by means of an ImageXpress micro XL resuspended in lysis buffer containing 150 mM sodium automated bioimager (Molecular Devices) equipped with chloride, 1.0% NP-40, 0.5% sodium deoxycholate, 0.1% a PlanApo 20 × /0.75 NA objective (Nikon, Tokyo, Japan). SDS, and protease inhibitor cocktails (Complete protease inhibitor cocktail, Roche, Basel, Switzerland) and incu- Immunostaining bated on ice for 30 min. To obtain supernatant, cell lysate In total, 5 x 103 U2OS cells were seeded into black 96- was centrifuged at 12000×g for 20 min at 4 °C to remove well µclear imaging plates (Greiner Bio-One) and allowed insoluble materials. The lysate was mixed with 4 × to adapt for 24 h. Thereafter, the cells were treated with NuPAGE® LDS sample buffer and 10 × sample reducing DTT compounds and respective controls and incubated agent, and proteins were denaturated at 100℃ for 10 min. for additional 6 or 24 h before fixation in 3.7% (w/v) NuPAGE® Novex® 4–12% Bis-Tris Protein Gels (Thermo paraformaldehyde (PFA) in PBS supplemented with 1 µM Fisher Scientific) were used for protein electrophoresis Hoechst 33342 for 20 min. Upon fixation, cells were under a 100 V constant voltage mode. Separated proteins permeabilized with 0.1% Triton in PBS for 10 min at RT. were transferred from gel to PVDF membrane (Merck- Unspecific binding was blocked with 2% BSA in PBS for Millipore, Darmstadt, Germany). After blocking with 5% 10 min at RT followed by primary antibody diluted in BSA BSA in 1 × TBS containing 0.1% Tween®−20 (1 × TBST) 2% following the manufactures recommendations over- for 1 h at room temperature, the membranes were probed night on an orbital shaker at 4 °C. The cells were rinsed with corresponding primary antibodies at 4 °C overnight: twice and stained with AlexaFluor-coupled secondary anti-mouse MLKL polyclonal antibody (AP14272b; antibodies for 1 h at RT, rinsed twice and subjected to Abgent, San Diego, CA, USA). To visualize phosphory- imaging using an ImageXpress micro XL automated lated MLKL (pMLKL), anti-mouse MLKL (phospho S345) bioimager (Molecular Devices) equipped with a PlanApo monoclonal antibody (ab196436, Abcam) was used. The 20 × /0.75 NA objective (Nikon). membranes were then washed and incubated with HRP- conjugated secondary antibodies (SouthernBiotech, Bir- High-content screening microscopy mingham, AL, USA) at room temperature for 1 h. The U2OS cells stably expressing GALT1-GFP, CALR- peroxidase activity was detected with ECL Western KDEL-GFP, H2B-RFP, LAMP1-GFP, and DIABLO-GFP Blotting Detection Reagent (GE healthcare, Chicago, IL, or HT-29 cells stably expressing RIP3-GFP were seeded in USA) and images were acquired by ImageQuant LAS 96-well black microplates for 24 h. After treatment, cells 4000 (GE healthcare). were fixed with 3.7% PFA for 20 min at room temperature and stained with 10 µg/ml Hoechst 33342 in PBS. Image RNA extraction, reverse transcription, and qRT-PCR acquisition was performed using an ImageXpress Micro Cells were collected for total RNA extraction using the XL automated microscope (Molecular Devices). A mini- RNeasy Mini kit (Quiagen, Hilden, Germany). Three mum of four view fields were captured per well. Upon micrograms of RNA were reverse-transcribed to cDNA acquisition, images were analyzed using the Custom using SuperScript® III First-Strand System (Thermo Module Editor of the MetaXpress software (Molecular Fisher Scientific). Type I IFN-related was Devices). Briefly, cells were segmented and divided into quantified with TaqMan® Gene Expression Assays using nuclear and cytoplasmic regions based on the nuclear Universal Master Mix II (with UNG) on a StepOnePlus™ Hoechst staining and GFP or RFP cytoplasmic signals. Real-Time PCR system (all from Thermo Fisher Scien- GFP-LC3 and RIP3-GFP dots were detected using auto- tific). GAPDH was used as house-keeping gene for nor- mated thresholding, and their number and surface were malization. Relative gene expression was quantified using measured in the cytoplasmic compartment. Cooccurrence the comparative Ct method and was calculated as fold of GALT1-GFP, CALR-KDEL-GFP, H2B-RFP, LAMP1- change. All experiments were conducted in triplicate GFP, and DIABLO-GFP fluorescence signals with Pacific assessment. Blue-labeled DTT compounds were systematically ana- lyzed to assess subcellular targeting. Data processing and High-throughput assessment of cell death statistical analyses were performed using the R software In total, 5 x 103 U2OS cells were seeded into black 96- (http://www.r-project.org/). well µclear imaging plates (Greiner Bio-One) and allowed to adapt for 24 h. Thereafter the cells were treated with Transmission electron microscopy the DTT compounds and respective controls and incu- For ultrastructural studies, human osteosarcoma U2OS bated for additional 6 or 24 h before either 1 µM of DAPI cells were fixed in 1.6% glutaraldehyde (v/v in 0.1 M or a mixture of 1 µM Hoechst, and 1 µM propidium phosphate buffer) for 1 h, collected by scraping, cen- iodide were added immediately before monitoring the trifuged, and the pellet was postfixed 1% osmium

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 12 of 14

tetroxide (w/v in 0.1 M phosphate buffer). Following inoculating 500.000 cells. When tumors became palpable, dehydration through a graded ethanol series, cells were 1.5 mg of DTT peptides were injected intratumorally. embedded in Epon™ 812 and ultrathin sections were Four days later, remaining tumor tissue was treated stained with standard uranyl acetate and lead citrate. accordingly and mice well-being and tumor growth were Images were taken using a Tecnai 12 electron microscope monitored. Anti-CD8 and anti-CD4 i.p. injections were (FEI, Eindhoven, the Netherlands). repeated every 7 days to assure the complete depletion of both T cell populations during the whole experiment. Data processing and statistical analyses Animals were sacrificed when tumor size reached end- Unless otherwise specified, experiments were performed point or signs of obvious discomfort associated to the in triplicate parallel instances and repeated at least once, treatment were observed following the EU Directive 63/ and data were analyzed with the R software (http://www. 2010 and our Ethical Committee advice. Surviving and r-project.org/). Microscopy images were segmented and tumor-free animals were analyzed and kept for more than analyzed by means of the MetaXpress (Molecular Devi- 30 days before rechallenge with 5 × 105 live TC-1 and ces) software and numerical data was further processed MCA205 cells injected contralateral in case of MCA205 with R. Unless otherwise specified, data are presented as previously-injected tumor-free animals. In case of means ± SD. Thresholds for the minimum number of previously-injected TC-1 tumor-free animals the location events in each analysis necessary to apply further statistics of the injected cells was inverted. Animals were mon- were calculated based on a medium effect size (according itored and tumor growth documented regularly until end- to Cohen’s conventional criteria) using the pwr package points were reached or signs of obvious discomfort were for R with a targeted value of 0.95. Samples that did not observed. Statistical analysis was performed employing match the requirements were marked ND and were two-way ANOVA analysis followed by Bonferroni’s test excluded from the analysis. comparing to Ctr conditions (*p < 0.05, **p < 0.01, and ***p < 0.001). Determination of surface-exposed CALR or MHC-I by immunofluorescence Peptide synthesis Cells were collected and rinsed twice with cold PBS. All peptides were synthesized on solid-phase with a Following the cells were incubated with an anti-CALR Prelude instrument (Protein Technologies Inc. Tucson, antibody (ab2907; Abcam, Cambridge, UK) or anti-MHC- AZ, USA) using standard Fmoc protocols and amino acid I antibody (12-5958-82, eBioscience, San Diego, CA, USA) derivatives. All synthesized peptides were prepared as C- diluted in cold blocking buffer (1% BSA in PBS) for terminal amides by using a Rink amide resin (Nova- 60 min on ice, followed by washing and incubation with biochem, Merck-Millipore, Billerica, MA, USA) as solid AlexaFluor 488-conjugates (Invitrogen) in blocking buffer support. The Fmoc-amino acids used were standard (for 30 min). Thereafter cells were washed in cold PBS, the derivatives from Novabiochem. Double couplings (2 × vital dye propidium iodide (PI) or Zombie UV was added 30 min, 5 eq to the resin) were performed. The incoming to a final concentration of 1 μg/mL, and samples were Fmoc-amino acids were activated with 5 eq (2-(6-chloro- analyzed by means of a CyAn ADP (Beckman Coulter, 1H-benzotriazole-1-yl)−1,1,3,3-tetramethylaminium hex- Brea, CA, USA) coupled to a HyperCyt autosampler afluorophosphate) (HCTU) and 10 eq diisopropylethyla- (IntelliCyt; Albuquerque, NM, USA). The analysis was mine (DIPEA) with dimethylformamide (DMF) as solvent. − limited to living (PI ) cells. Data were statistically eval- Coupling reactions were concluded with a washing (DMF, uated using R (https://www.r-project.org). 3 × 30 s) and Fmoc-removal step (20% piperidine in DMF, 5 + 10 min). For microscopy studies, the peptides were Mouse experiments fluorescently labeled at their C-termini with Pacific Blue™ Female wild-type C57BL/6 mice at the age of 6–8 weeks (Exmax = 410 nm, Emmax = 455 nm). This was achieved by were obtained from Harlan France (Gannat, France) and using a Universal NovaTag resin (Novabiochem, Merck- maintained in the animal facility at Gustave Roussy Millipore, Billerica, MA, USA) as solid support. Before Campus Cancer in specific pathogen–free conditions in a removing the final Fmoc-group, the resin was treated with temperature-controlled environment with 12 h light, 12 h 2% trifluoroacetic acid (TFA) in dichloromethane (3 × dark cycles, and received food and water ad libitum. 1 min) and washed with DMF (3 × 30 s) and 1 M DIPEA in Animal experiments were in compliance with the EU DMF (1 × 30 s). The resin was then treated with Pacific Directive 63/2010 and protocols 2013_094A and were Blue succinimidyl ester (Thermo Fisher Scientific, Wal- approved by the Ethical Committee of the Gustave Roussy tham, MA, USA) for 1 h. The resin was washed with DMF Campus Cancer (CEEA IRCIV/IGR no. 26, registered at (3 × 30 s) and the final Fmoc-group removed. Completed the French Ministry of Research). MCA205 tumors were peptides were cleaved from the resin using a cocktail established in C57BL/6 hosts by subcutaneously containing 95% TFA, 2.5% water, and 2.5%

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 13 of 14

triisopropylsilane for 3 h. TFA was removed using a isolated blood from healthy individuals who gave rotavapor (Hei-VAP Advantage rotavapor, Heidolph their signed informed consent. RBCs were resuspended to a Instruments, Schwabach, Germany) and the fully depro- 10% hematocrit solution before being incubated for 1 h at tected peptides precipitated with diethyl ether as C- 37 °C with DTT compounds dissolved in PBS at con- terminal amides. The ether was decanted and the pre- centrations ranging from 438–928 μM (1500–3000 μg/ml). cipitated crude peptide allowed to air dry before analysis RBCs with PBS and 1% Triton solution alone served as a and purification. negative and positive control, respectively. After centrifu- ging the samples at 4000 rpm for 5 min, the absorbance of Peptide purification and characterization the supernatant was measured at 405 nm on a spectro- The solvents used in the analytical and preparative photometric microliter plate reader (Thermomax, Mole- systems were MilliQ water (Solvent A) and acetonitrile cular Devices). The protocol used for blood sampling and (Solvent B), both modified with 0.1% TFA. The crude and handling has been reviewed and approved by the Norwe- purified peptides were analyzed on an ACQUITY UPLC gian Regional Ethic Committee (REK)-approved protocol H-class system with a photodiode array (PDA) detector (2016/376). (Waters, Milford, MA, USA) equipped with an ACQUITY CEH C18 UPLC column (Waters, 2.1 × 50 mm, 1.7 µm). A Acknowledgements H.Z. and P.L. were supported by the China Scholarship Council, L.Z. is gradient of 0–50% Solvent B over 30 min with a flow rate supported by the Ligue contre le Cancer, G.K. and L.Z. are supported by the of 1 mL/min was used, and detection was set at 200–500 Ligue contre le Cancer (équipes labelisées); Agence National de la Recherche – nm. The crude peptides were purified to >95% on a (ANR) Projets blancs; ANR under the frame of E-Rare-2, the ERA-Net for Research on Rare Diseases; Association pour la recherche sur le cancer (ARC); XSelect CSH C18 OBD prep column (19 × 250 mm, 5 µm; Cancéropôle Ile-de-France; Institut National du Cancer (INCa); Fondation Waters, Milford, MA, USA) installed in an AutoPur- Bettencourt-Schueller; Fondation de France; Fondation pour la Recherche fi – Médicale (FRM); the European Commission (ArtForce); the European Research i cayion System (Waters). A default gradient of 10 40% Council (ERC); the LabEx Immuno-Oncology; the SIRIC Stratified Oncology Cell Solvent B over 30 min with a flow rate of 20 mL/min was DNA Repair and Tumor Immune Elimination (SOCRATE); the SIRIC Cancer used, but it was adjusted as required. The detection was Research and Personalized Medicine (CARPEM); the Swiss Bridge Foundation, – ISREC and the Paris Alliance of Cancer Research Institutes (PACRI). This project set at 200 500 nm. The molecular weight of the peptides was supported by the Norwegian Research Council (254800, 257967). LCGdS is was confirmed on the Xevo G2 Q-TOF with ACQUITY supported by Portuguese Science Foundation (ref. SFRH/BPD/93562/2013). UPLC I-Class system (Waters). The purified peptides Author details were then freeze-dried (Labconco FreeZone 4.5 Plus, 1 − Metabolomics and Cell Biology Platforms, Gustave Roussy Comrehensive Kansas City, MO, USA) as TFA-salts and stored at 20 °C Cancer Institute, Villejuif, France. 2Equipe 11 labellisée Ligue contre le Cancer, until further use. Centre de Recherche des Cordeliers, INSERM U, 1138 Paris, France. 3Université Paris Descartes, Sorbonne Paris Cité, Paris, France. 4Université Pierre et Marie Curie, Paris, France. 5Lytix Biopharma, Oslo, Norway. 6Division of Cancer, In vitro cytotoxicity Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo, The MTT assay was adopted to determine cell viability Norway. 7Institute of Clinical Medicine, University of Oslo, Oslo, Norway. 8 9 in a panel of cancerous and non-transformed cells after Chemistry Department, University of Coimbra, Coimbra, Portugal. Gustave Roussy Comprehensive Cancer Center, Villejuif, France. 10CNRS, UMR9196, 4 h incubation with DTT peptides. Pre-cultured cells were Villejuif, France. 11University of Paris Sud XI, Kremlin Bicêtre, France. 12Institut 4 4 seeded in 96-well plates at a density of 1 × 10 –1.5 × 10 National de la Santé et de la Recherche Medicale (INSERM), U1015 Villejuif, 13 cells/well and applied for experiment as previously France. Center of Clinical Investigations in Biotherapies of Cancer (CICBT) 507, Villejuif, France. 14Institute of Medical Biology, University of Tromsø, Tromsø, described. In short, cells were washed once with serum- Norway. 15Karolinska Institutet, Department of Women’s and Children’s Health, free RPMI 1640 and incubated with increasing con- Stockholm, Sweden. 16Norut Northern Research Institute, SIVA Innovation Centre, Tromsø, Norway. 17Pôle de Biologie, Hôpital Européen Georges centrations of DTT peptides before adding 10 μl MTT 18 fi Pompidou, AP-HP, Paris, France. Present address: Institute of Modern Physics, solution to each well. Lastly, acidi ed isopropanol was Chinese Academy of Sciences, Lanzhou, China added to facilitate formazan crystal solubilization. Absorbance was measured at 570 nm on a spectro- Conflict of interest fl photometric microtiter plate reader (Thermomax Mole- The authors declare that they have no con ict of interest. cular Devices, NJ, USA). Cell survival was calculated as the A570 nm of DTT-treated cells relative to the negative Publisher’s note control (100% viable cells) using the mean of two or three Springer Nature remains neutral with regard to jurisdictional claims in fi independent experiments and expressed as a 50% inhibi- published maps and institutional af liations. tory concentration (IC50). Supplementary Information accompanies this paper at (https://doi.org/ 10.1038/s41419-018-1127-3). Hemolytic activity The cytotoxic activity of two selected DTT compounds Received: 4 July 2018 Revised: 28 September 2018 Accepted: 1 October 2018 (DTT-205andDTT-304)againsthumanredbloodcells (hRBCs) was determined by a hemolytic assay using freshly

Official journal of the Cell Death Differentiation Association Zhou et al. Cell Death and Disease (2018) 9:1086 Page 14 of 14

References 16. Ngwa, W. et al. Using immunotherapy to boost the abscopal effect. Nat. Rev. 1. Ruoslahti, E., Bhatia, S. N. & Sailor, M. J. Targeting of drugs and nanoparticles to Cancer 18,313–322 (2018). tumors. J. Cell. Biol. 188, 759–768 (2010). 17. Galluzzi, L., Buque, A., Kepp, O., Zitvogel, L. & Kroemer, G. Immunological effects 2. Iwasaki, T. et al. Selective cancer cell cytotoxicity of enantiomeric 9-mer of conventional chemotherapy and targeted anticancer agents. Cancer Cell. peptides derived from beetle defensins depends on negatively charged 28,690–714 (2015). phosphatidylserine on the cell surface. Peptides 30,660–668 (2009). 18. Zitvogel,L.,Galluzzi,L.,Kepp,O.,Smyth,M.J.&Kroemer,G.TypeIinterferonsin 3. Papo, N. & Shai, Y. Host defense peptides as new weapons in cancer treat- anticancer immunity. Nat. Rev. Immunol. 15,405–414 (2015). ment. Cell. Mol. Life Sci. 62,784–790 (2005). 19. Obeid, M. et al. Calreticulin exposure dictates the immunogenicity of cancer 4. Eike, L. M., Mauseth, B., Camilio, K. A., Rekdal, O. & Sveinbjornsson, B. The cell death. Nat. Med. 13,54–61 (2007). cytolytic amphipathic beta(2,2)-amino acid LTX-401 induces DAMP release in 20. Apetoh, L. et al. Toll-like receptor 4-dependent contribution of the immune melanoma cells and causes complete regression of B16 melanoma. PLoS. One. system to anticancer chemotherapy and radiotherapy. Nat. Med. 13, 11, e0148980 (2016). 1050–1059 (2007). 5. Sveinbjornsson, B., Camilio, K. A., Haug, B. E. & Rekdal, O. LTX-315: a first-in-class 21. Ghiringhelli, F. et al. Activation of the NLRP3 inflammasome in dendritic cells oncolytic peptide that reprograms the tumor microenvironment. Future Med. induces IL-1beta-dependent adaptive immunity against tumors. Nat. Med. 15, Chem. 9, 1339–1344 (2017). 1170–1178 (2009). 6. Yamazaki, T. et al. The oncolytic peptide LTX-315 overcomes resistance of 22. Camilio, K. A., Berge, G., Ravuri, C. S., Rekdal, O. & Sveinbjornsson, B. Complete cancers to immunotherapy with CTLA4 checkpoint blockade. Cell Death Differ. regression and systemic protective immune responses obtained in B16 mel- 23, 1004–1015 (2016). anomas after treatment with LTX-315. Cancer Immunol. Immunother. 63, 7. Zhou, H. et al. The oncolytic peptide LTX-315 triggers immunogenic cell 601–613 (2014). death. Cell Death Dis. 7, e2134 (2016). 23. Camilio, K. A., Rekdal, O. & Sveinbjornsson, B. LTX-315 (oncopore): a short 8. Zhou, H. et al. The oncolytic compound LTX-401 targets the Golgi apparatus. synthetic anticancer peptide and novel immunotherapeutic agent. Oncoim- Cell Death Differ. 23,2031–2041 (2016). munology 3, e29181 (2014). 9. Al-Benna,S.,Shai,Y.,Jacobsen,F.&Steinstraesser,L.Oncolyticactivitiesofhost 24. Eksteen, J. J. et al. Iterative design and in vivo evaluation of an oncolytic defense peptides. Int. J. Mol. Sci. 12,8027–8051 (2011). antilymphoma peptide. J. Med. Chem. 60,146–156 (2017). 10. Kroemer, G., Galluzzi, L. & Brenner, C. Mitochondrial membrane permeabili- 25. Forveille, S. et al. The oncolytic peptide LTX-315 triggers necrotic cell death. zation in cell death. Physiol. Rev. 87,99–163 (2007). Cell Cycle 14,3506–3512 (2015). 11. Zhou, H. et al. The oncolytic peptide LTX-315 kills cancer cells through Bax/ 26. Aits, S. & Jaattela, M. Lysosomal cell death at a glance. J. Cell. Sci. 126, Bak-regulated mitochondrial membrane permeabilization. Oncotarget 6, 1905–1912 (2013). 26599–26614 (2015). 27. Galluzzi, L., Lopez-Soto, A., Kumar, S. & Kroemer, G. Caspases connect cell- 12. Eike, L. M., Yang, N., Rekdal, O. & Sveinbjornsson, B. The oncolytic peptide LTX- death signaling to organismal homeostasis. Immunity 44,221–231 (2016). 315 induces cell death and DAMP release by mitochondria distortion in 28. Cho, Y. S. et al. Phosphorylation-driven assembly of the RIP1-RIP3 complex human melanoma cells. Oncotarget 6, 34910–34923 (2015). regulates programmed necrosis and virus-induced inflammation. Cell 137, 13. Sistigu, A., Manic, G. & Vitale, I. LTX-315, CAPtivating immunity with necrosis. 1112–1123 (2009). Cell Cycle 15, 1176–1177 (2016). 29. Linkermann, A. & Green,D.R.Necroptosis.N. Engl. J. Med. 370,455–465 (2014). 14. Galluzzi, L., Buque, A., Kepp, O., Zitvogel, L. & Kroemer, G. Immunogenic cell 30. Yang, H. et al. Contribution of RIP3 and MLKL to immunogenic cell death in cancer and infectious disease. Nat. Rev. Immunol. 17,97–111 (2017). death signaling in cancer chemotherapy. Oncoimmunology 5, e1149673 15. Bloy, N. et al. Immunogenic stress and death of cancer cells: contribution of (2016). antigenicity vs adjuvanticity to immunosurveillance. Immunol. Rev. 280, 31. Marabelle, A., Tselikas, L., de Baere, T. & Houot, R. Intratumoral immunotherapy: 165–174 (2017). using the tumor as the remedy. Ann. Oncol. 28, xii33–xii43 (2017).

Official journal of the Cell Death Differentiation Association