The Immunomodulatory Effects of Arsenic Trioxide in Autoimmunity and Alloreactivity Yishan Ye

To cite this version:

Yishan Ye. The Immunomodulatory Effects of Arsenic Trioxide in Autoimmunity and Alloreactivity. Immunology. Sorbonne Université, 2019. English. ￿NNT : 2019SORUS426￿. ￿tel-03233559￿

HAL Id: tel-03233559 https://tel.archives-ouvertes.fr/tel-03233559 Submitted on 25 May 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. Sorbonne Université Ecole doctorale 394: Physiologie, Physiopathologie et Thérapeutique

Centre de recherche Saint Antoine / Equipe Mohty

The Immunomodulatory Effects of Arsenic Trioxide in Autoimmunity and Alloreactivity

Par Yishan YE

Thèsededoctoratdebiologie

Spécialité : Immunologie

Dirigée par Mohamad Mohty

Présentée et soutenue publiquement le 23 Mai 2019

Devant un jury composé de : AUCOUTURIER Pierre, PU-PH, Président du Jury BAZARBACHI Ali, Professeur, Rapporteur SAAS Philippe, Professeur, Rapporteur HERMINE Olivier, PU-PH, Examinateur MALARD Florent, MCU-PH, Co-directeur de thèse MOHTY Mohamad, PU-PH, Directeur de thèse

1 Acknowledgments

First I would like to thank the jury members, Professor Pierre Aucouturier, Professor Ali Bazarbachi, Professor Philippe Saas, Professor Olivier Hermine for having accepted to evaluate my work.

I can still remember the day in 2014 when I asked my distinguished guest from France, the president of the European Society for Blood and Marrow Transplantation, to write me a recommendation letter for a future PhD. The answer was ‘Perhaps you can come to my lab’, which opened the door to a long and fascinating journey.

To Professor Mohty, despite tons of work that you are dealing with on a daily basis, you are always the first to respond to my requirements, and the first to congratulate me when something good happens. Your global vision, leadership nature, passion towards work, great care and patience for me make you a role model that I will always follow. Thank you Professor Mohty, with my greatest appreciation.

To Dr. Florent Malard, you are really the most outstanding young physician scientist that I have ever had the privilege to meet. You are in charge of all the detailsofmy project, and always help me make correct choice when problem happens. It is my honor to have you, a man with sharp mind, accurate choice, impartial judgement as my co-director. Thank you Florent.

To Dr. Béatrice Gaugler, you were the first person who gave me greetings during my first visit to the hospital, the moment that I will always remember. As the scientific leader of the lab, you offered me the first idea for my project, and later became the main body of this thesis. I can’t finish this thesis without the ideas, and the continuous and unselfish intellectual supports from you. Thank you Béatrice.

To Dr. Laure Ricard and Dr. Nicolas Stocker, my greatest colleagues, for fighting together, for creating a really welcoming and harmonious atmosphere, for helping and encouraging me during the hard time of research. Thank you! To Prof. Arsène

2 Mekinian, for great support on the clinical samples and suggestions offered during my PhD study. Thank you! To Dr. Lama Siblany, an excellent start and I am sure that you will do great academic work in the near future. Good luck! Nevertherless, toDr. Charlotte Laurent, Mr. Maxime Tenon, Mr. Christophe de Vassoigne, and all the other people that I have had the chance to meet in the lab, thank you so much.

To Dr. Baptiste Lamarthée, Mr. Frédéric de Vassoigne, and Dr. Ruoping Tang, you were the first who offered me, a nervous foreigner without any lab techniques, the warmest welcome and the orientation courses with great patience. Thank youso much.

It is impossible to list all the people who have helped and supported me during the three years. France has become so warm and lovely because of you.

To Prof. He Huang and all the colleagues in Hangzhou, for offering me the best clinical training, and continuous care during my PhD study. Thank you.

To China Scholarship Council for the financial support.

Finally, I would like to extend my most sincere thanks to my parents, my girlfriend, my friends, and all the other people who cared and supported me continuously, far away in China. Love you forever, and see you soon.

3 List of abbreviations

ABT-199: chemical name of venetoclax G-CSF: granulocyte-colony stimulating aGVHD: acute graft-versus-host disease factor AHR: airway hyperresponsiveness GM-CSF: granulocyte-macrophage AP3: adapter protein-3 colony-stimulating factor APCs: antigen presenting cells GRFS: GVHD-free/relapse-free survival APL: acute promyelocytic leukemia GVHD: graft-versus-host disease As(III): trivalent arsenicals GVL: graft-versus-leukemia

As2O3: arsenic trioxide H2O2: hydrogen peroxide ATG: antithymocyte globulin HEVs: high endothelial venules ATL: adult T-cell leukemia/lymphoma HLA: human leukocyte antigen ATRA: all-trans retinoic acid HMGB1: high mobility group box 1 BALF: broncho-alveolar lavage fluid HSC: hematopoietic stem cell BST2: bone marrow stromal antigen 2 HCT: hematopoietic cell transplantation BU: busulfan iAs: inorganic arsenic compounds cDC: conventional dendritic cell IBD: inflammatory bowel disease CDP: common DC progenitor ID2: Inhibitor Of DNA Binding 2 cGAMP: cyclic guanosine IDO: indoleamine 2,3-dioxygenase monophosphate-adenosine IFNAR: IFN-I receptor monophosphate IFN-I: type-I interferons CLP: common lymphoid precursor IKK: IĸB kinase CMP: common myeloid precursor IRF7: interferon regulatory factor 7 CNI: calcineurin inhibitor LAK: lymphokine activated killer CpG-ODN: CpG LAP: LC3-associated phagocytosis oligodeoxyribonucleotides LC3: microtubule-associated protein CsA: cyclosporine 1A/1B-light chain 3 G-CSF: granulocyte-colony stimulating LFA-1: Lymphocyte function-associated factor antigen 1 M-CSFR: macrophage Lin: lineage markers colony-stimulating factor receptor LMPP: lymphoid-primed multi-potent CX3CR1: CX3C chemokine receptor 1 progenitor CY: cyclophosphomide LP: lymphoid precursor DAMPs: damage-associated molecular LPS: lipopolysaccharide patterns MA: myeloablative DC: dendritic cell mAb: monoclonal antibodies EC50: half maximal effective MAPKs: mitogen-activated protein concentration kinases ER: endoplasmic reticulum MCMV: mouse cytomegalovirus FKBP12: FK506-binding protein 12 MDDC: monocyte derived dendritic cell Flt3L: Fms-like tyrosine kinase 3 ligand MDP: macrophage and DC precursor GC: glucocorticoid MHC: major histocompatibility complex GI: gastrointestinal tract miHA: minor histocompatibility antigen

4 miRNA: microRNA STAT3: signal transducer and activator MM: multiple myeloma of transcription 3 MMF: mycophenolate mofetil Syk: spleen tyrosine kinase MP: myeloid precursor TBI: total body irradiation mRNA: messenger RNA TCF4: transcription factor 4 mTOR: mammalian target of rapamycin Th: T helper cell MTX: methotrexate TLR: toll-like receptor MYD88: myeloid differentiation TNF: tumor necrosis factor primary response protein 88 Treg: regulatory T cell

NaAsO2: sodium arsenite UPR: unfolded protein response NADPH: nicotinamide adenine dinucleotide phosphate NBs: nuclear bodies NETs: neutrophil extracellular traps Nf-κB: nuclear factor-κB NK cell: natural killer cell NOD: non-obese diabetic Nrf2: nuclear factor erythroid 2-related factor 2 NRM: non-relapse mortality PAMPs: pathogen-associated molecular patterns PBMC: peripheral blood mononuclear cell pDC: plasmacytoid dendritic cell PML: promyelocytic leukemia PTCy: posttransplantation cyclophosphomide RA: rheumatoid arthritis RARE: retinoic acid response elements RARα: retinoic acid receptor-α RIC: reduced-intensity conditioning RNS: reactive nitrogen species ROR: retinoic-related orphan receptor ROS: reactive oxygen species RXR: retinoid X receptor Sca-1: stem cells antigen-1 Siglec-H: sialic acid-binding immunoglobulin-like lectin H SLE: systemic lupus erythematosus SPF: specific pathogen-free SSc: systemic sclerosis

5 List of Figures

Figure 1 . Reaction of vicinal sulfhydryl groups in a protein structure with trivalent arsenic (Emadi et al., Blood Rev, 2010)...... 10

Figure 2 . Dual targeting of PML-RARα and PML by As2O3 lead to APL cell differentiation and loss of self-renewal (de Thé., Nat Rev Cancer, 2018) ...... 11 Figure 3 . Diverse functions of As(III) on immune cells...... 15 Figure 4 . Mechanisms of action: As2O3 effects on CD4+ T cell...... 20 Figure 5 . The proposed spectrum between pDC and cDC (Reizis, Immunity, 2019) 30 Figure 6 . Possible pathways of plasmacytoid dendritic cell development (modified from Shortman et al, Adv Immunol, 2013)...... 34 Figure 7 . Factors involved in pDC trafficking (modified from Swiecki et al,NatRev Immunol, 2015)...... 36 Figure 8 . TLR9 signaling (Swiecki et al, Nat Rev Immunol, 2015)...... 40 Figure 9 . Cell-intrinsic and cooperative sensing of pDCs (modified from Reizis, Immunity, 2019)...... 41 Figure 10 . aGVHD pathophysiology (Ghimire et al, Front Immunol, 2017)...... 51 Figure 11 . Initiation phase of aGVHD (Ghimire et al, Front Immunol, 2017)...... 52 Figure 12 . Chemotherapy and HCT protocol ...... 97 Figure 13 . GVHD Clinical manifestations...... 98 Figure 14 . Donor engraftment and lymphoid subpopulations...... 100 Figure 15 . Effects of different drugs on aGVHD...... 101

6 List of Tables

Table 1 . Trivalent arsenic induces immune cell apoptosis: sensitivity and mechanisms of action...... 13 Table 2 . Use of trivalent arsenic in mouse models of immune-mediated diseases.....24 Table 3 . Trivalent arsenic facilitates tumor immunotherapy...... 27 Table 4 . Revised Glucksberg aGVHD grading system (Harris et al., Biol Blood Marrow Transplant, 2016)...... 50 Table 5 . Overview of mouse models for aGVHD based on TBI (Boieri et al., Front Immunol, 2016)...... 58 Table 6 . Mouse GVHD scoring (modified from Cooke et al., Blood, 1996)...... 104

7 Table of Contents

Acknowledgments...... 2 List of abbreviations...... 4 List of Figures...... 6 List of Tables...... 7 1. Introduction...... 9 1.1. Immunomodulatory properties of trivalent arsenic...... 10 1.1.1. Pharmaceutical mechanisms of action...... 10 1.1.2. The multifaceted effects of trivalent arsenic on immune cells...... 15 1.1.3. Trivalent arsenic in mouse models of immune-mediate diseases...... 22 1.1.4. Trivalent arsenic and tumor immunotherapy...... 26 1.2. Plasmacytoid dendritic cells in autoimmunity/alloreactivity...... 29 1.2.1. Definition of pDCs...... 29 1.2.2. Development of pDCs...... 31 1.2.3. Functions of pDCs...... 35 1.2.4. pDCs in autoimmunity...... 43 1.2.5. pDCs in alloreactivity...... 45 1.3. Mouse models of acute GVHD...... 48 1.3.1. Acute GVHD (aGVHD)...... 48 1.3.2. Challenges in aGVHD prophylaxis/treatment...... 54 1.3.3. Translational value of existing aGVHD mouse models...... 57 2. Results...... 62 2.1. Part 1. Immunomodulatory effects of arsenic trioxide on plasmacytoid dendritic cells and study of mechanism...... 62 2.1.1. Article 1: Arsenic trioxide induces regulatory functions of plasmacytoid dendritic cells through interferon-alpha inhibition...... 63 2.2. Part 2. Effects of arsenic trioxide and other immunomodulatory drugs in a novel mouse models of acute graft-versus-host disease...... 96 2.2.1. Results...... 97 2.2.2. Materials and Methods...... 102 3. Discussions...... 105

3.1. As2O3, a promising drug for diseases with IFN-I signature...... 105 3.2. A novel aGVHD model with chemotherapy-based conditioning and G-CSF mobilized graft: advances and limitations...... 108 4. Conclusion...... 111 5. References...... 112 6. Annex...... 149 Article 2...... 149 Article 3...... 162

8 1.Introduction

Inorganic arsenic compounds (iAs) have been used in traditional Chinese and Western medicine for over 2400 years (Emadi et al., 2010). Arsenic trioxide (As2O3)was rediscovered in the 1970s in the treatment of acute promyelocytic leukemia(APL) with striking efficacy and good safety profile (Wang et al., 2008; Shen et al., 1997).

So far, As2O3, together with all-trans retinoic acid (ATRA), has revolutionized the treatment of APL (Lo-Coco et al., 2013; Cicconi et al., 2016).

In the recent decades, the successful treatment of As2O3 in mouse models of several autoimmune and inflammatory diseases has shed light on this old drug as a ‘novel’ immunomodulator (Bobe et al., 2006; Kavian et al., 2012a; 2012b). Consistently,

As2O3 has also shown efficacy in a mouse model of graft-versus-host disease (GVHD), which is a major inflammatory complication after allogeneic hematopoietic cell transplantation (HCT). Alloreactivity is the cause of GVHD, which happens when immunocompetent T cells in the donated tissue (the graft) recognize the recipient (the host) as foreign, and attack the target organs. However, despite these exciting findings, the mechanisms underneath are largely unknown. Herein, we focused on plasmacytoid dendritic cells (pDCs), which is a unique subset of dendritic cells specialized in secreting high levels of type-I interferons (IFN-I), and have been reported to be implicated in the pathogenesis of autoimmune diseases such as systemic sclerosis (SSc) (Ah Kioon et al., 2018). Moreover, pDCs are also demonstrated to play a important role in GVHD pathophysiology (Bossard et al., 2012; Malard et al., 2013; Waller et al., 2014).

Given the therapeutic potential of As2O3, and pathogenetic role of pDCs in both autoimmunity and alloreactivity, respectively, the first part of this thesis aimed to explore the in vitro effects of As2O3 on pDC from both healthy donors and from SSc. Following the in vitro discoveries, in the second part we constructed a novel clinical-relevant mouse model of acute GVHD (aGVHD), and tested efficacy of

As2O3 and other potential anti-aGVHD drugs in this model.

9 1.1.Immunomodulatory properties of trivalent arsenic

1.1.1.Pharmaceutical mechanisms of action

1.1.1.1.As (III) biochemistry

The immune-modulation of iAs has been studied and put in clinical use with trivalent arsenicals (As(III)) including As2O3 and sodium arsenite (NaAsO2). Both As2O3 and sodium arsenite (NaAsO2) transform to arsenous acid in aquarious solution. After being up-taken by cells through the aquaglyceroporin 9 (AQP9) transmembrane protein, the intracellular As(III) exerts the subsequent biochemical effects (Leung et al., 2007). Interaction with the thiol (or sulfhydryl) groups (-SH) of proteins with a high cysteine content constitutes the basic biochemical reaction of As(III) (Emadi et al., 2010), as shown in Figure 1, which alters the conformation, resulting in loss of their function, and affect their recruitment and interaction with other proteins and DNA (Shen et al., 2013).

Figure 1. Reaction of vicinal sulfhydryl groups in a protein structure with trivalent arsenic (Emadi et al., Blood Rev, 2010)

Apart from the direct arsenic-protein binding, recent studies reveal that many key proteins are modulated through more complicated post-translational stepwise regulations (de Thé et al., 2018).

1.1.1.2.PML and PML nuclear body regulation

It was initially found that the APL promotor promyelocytic leukemia-retinoic acid receptor-α (PML/RARα) was especially sensitive to As2O3-induced degradation, which was the critical step of disease eradication (Chen et al., 1996). The following

10 comprehensive studies by de Thé et al reveal that As2O3 specifically targets the PML moiety (Lallemand-Breitenbach et al., 2018). PML nuclear bodies (NBs), nucleared by the PML protein, could recruit and sumoylate dozens of partner proteins, leading to a variety of biological processes (Lallemand-Breitenbach et al., 2001; Sahin et al.,

2014). Notably, As2O3-induced oxidative stress could enhance formation of PML NBs, leading to p53 activation in vivo in normal mice (Niwa-Kawakita et al., 2017). In APL, through both reactive oxygen species (ROS) production and direct binding,

As2O3 exerts its dual-targeting effects (Jeanne et al., 2010). On one hand, As2O3 induces PML/RARα sumoylation, proteasomal degradation, and APL cell differentiation (Lallemand-Breitenbach et al., 2018; Fasci et al., 2015). On the other hand, As2O3 targets the wild-type PML proteins, leading to re-formation of PML NBs, subsequent p53 activation, and ultimately APL clearance (Niwa-Kawakita et al., 2017;

Ablain et al., 2014). The mechanisms of action for As2O3-induced PML/RARα and PML NB regulation in the context of APL are shown in Figure 2.

Figure 2. Dual targeting of PML-RARα and PML by As2O3 lead to APL cell differentiation and loss of self-renewal (de Thé., Nat Rev Cancer, 2018) AS, arsenic trioxide; RXR, retinoid X receptor; RARE, retinoic acid response elements;

11 It is also found that in adult T-cell leukemia/lymphoma (ATL), As2O3, together with IFN-α, reaches disease complete remission both in mice and in human through degradation of the disease driver oncoprotein Tax (Kchour et al., 2009; El-Sabban et al., 2000; El Hajj et al., 2010). Interestingly, this process was also mediated through a

As2O3 enforcement of PML NBs formation, Tax sumoylation and proteasomal degradation (Dassouki et al., 2015). In addition, it is also discovered that the

As2O3/ATRA combination significantly decreases NPM1-mutant AML leukemia blast in patients though oxidative stress generation, p53 activation, and ultimately mutant NPM1 degradation (El Hajj et al., 2015; Martelli et al., 2015). Apart from tumor suppression, PML and PML NBs are recently found to play key role in mediating innate immune responses (Lallemand-Breitenbach et al., 2001; Hsu et al., 2018; Scherer et al., 2016). Studies have identified PML as a direct, positive regulator of IFN-I signaling (Kim et al., 2015), and is implicated in the regulation of and extended spectrum of cytokines such as the pro-inflammatory IL-1 β and IL-6 (Scherer et al., 2016; Lo et al, 2013). It is thus with great interest to investigate if PML and PML NBs modulations explain As2O3’ efficacy in autoimmune/inflammatory diseases with IFN-I signature and abnormal cytokine profile.

1.1.1.3.Apoptosis induction

As(III) can induce immune cell apoptosis through both the mitochondrial-mediated and the receptor-mediated pathways (Gupta et al., 2003; Yu et al., 2002). The pro-apoptotic mechanisms include generation of oxidative stress, caspase activation, alteration of the Bcl-2 family proteins, and up/down-regulation of several survival-related signaling pathways such as nuclear factor- κ B (Nf- κ B), Rho-kinase/p38-kinase, and TNF-R1 apoptotic signalings (Gupta et al., 2003; Yu et al., 2002; Lemarie et al., 2006a; Lemarie et al., 2006b). Within the therapeutic range, As(III) is able to induce apoptosis of specific types of sensitive immune cells (Bobe et al., 2006; Thomas-Schoemann et al., 2012). In MRL/lpr mice which develop a lupus like disease mimicking autoimmune lymphoproliferative syndrome, the pathogenetic

12 abnormally activated T cells can be selectively eliminated by As2O3, leading to disease regression and prolonged survival (Bobe et al., 2006). Moreover, in a mouse model of colon cancer, regulatory T cells are preferentially depleted, leading to immunomodulatory effects (Thomas-Schoemann et al., 2012). The in vitro EC50 of different immune subsets and the possible mechanisms of apoptosis induction are summarized in Table 1.

Table 1. Trivalent arsenic induces immune cell apoptosis: sensitivity and mechanisms of action Cell subtype Species Type of EC50 (culture Possible mechanism References As(III) time) PBMC Human NaAsO2 5  mol/L (48h N/A Yu et al, and 72h) 2002; Neutrophil Human As2O3 <5  mol/L 1. H2O2 generation Binetetal, (22h) 2. Activation of caspases 2006; 3. Denovoprotein synthesis Antoine et al, 4. Syk kinase activation 2010 Monocyte Monocyte Human As2O3 ~1mol/L (6d) 1. Inhibition of NF-κB Lemarie et al, (during related survival pathways 2006b macrophagic differentiation) Promonocytic Human As2O3 >4mol/L (4d) U937 cells Dendritic cell Monocyte-deri Human NaAsO2 >2mol/L (6d) N/A Macoch et al, ved dendritic 2013 cells Tcell Primary CD4+ Human As2O3 >5  mol/L 1. ROS generation Gupta et al, Tcell (48h) 2. Regulation of the Bcl-2 2003 Primary CD8+ Human As2O3 >5  mol/L family proteins Tcell (48h) CD4+ T cell Human NaAsO2 N/A 1. TNF-R1 apoptotic Yu et al, 2002 within PBMCs signaling CD4+CD25+ Human As2O3 ~1  mol/L 1. ROS and RNS Thomas-Scho Treg (38h) accumulation emann et al, 2. High sensitivity to 2012 oxydative stress CD4+CD25- Human As2O3 ~2.5  mol/L N/A Thomas-Scho effector T cell (38h) emann et al, 2012

13 1.1.1.4.Reactive oxygen species

ROS accumulation leads to oxidative stress, causing damage to nucleic acids, proteins, and lipids, which can lead to cell signaling change, and, ultimately, apoptosis (Schieber et al., 2014). As(III) induces immune cell intracellular ROS accumulation via inhibition of antioxidants such as glutathione, glutathione peroxidase and glutathione reductase (Gupta et al., 2003; Singh et al., 2010) and activation of ROS generation enzymes such as NADPH oxidase (Chou et al., 2004; Lemarie et al., 2008). As a consequence, expressions of redox-sensitive genes such as HMOX1, NQO1and GCLM are up-regulated (Bourdonnay et al, 2009a). Especially, nuclear factor erythroid 2-related factor 2 (Nrf2), a stress-activated transcription factor responsible for inducing a battery of cytoprotective genes, is found to be activated by As(III) within mouse splenocytes, human T cells, human primary macrophages, and human monocyte derived dendritic cells (MDDCs) (Bourdonnay et al, 2009a; Duan etal., 2017; Morzadec et al., 2014; Macoch et al., 2015). The Nrf2 modulation is probably due to As(III)-induced Nrf2 sumoylation, recruitment to the PML NBs and degradation by RNF4-mediated proteolysis (Malloy et al., 2013).

1.1.1.5.Signaling pathway regulation

As(III) has the potential to selectively target specific signaling pathways. In macrophages, for example, non-toxic concentrations of As2O3 induce up-regulation of 32 and depression of 91 genes, indicating a global multi-directional change of the cell-signaling (Bourdonnay et al., 2009b). Moreover, As(III) may inactivate up to 200 enzymes, and many are crucial regulators of important signaling pathways (Shen et al., 2013). As(III) inhibits IĸB kinase (IKK), whose integrity is key to the activation of the NFĸB pathway (Lemarie et al., 2006b). Other pathways affected include the Rho-kinase/p38-kinase pathway, c-jun NH2-terminal mitogen-activated protein kinases (MAPK), etc, which will be described in details in the next section (Lemarie et al., 2008; Binet et al., 2008a).

14 1.1.2.The multifaceted effects of trivalent arsenic on immune cells

1.1.2.1.Systemic immunomodulation

Chronic arsenic exposure leads to systemic immunosuppression (Danglebenetal.,

2013). However, clinical use of As2O3 has shown good safety-profile, with limited side effects (Shen et al., 1997). No long-term treatment-related immune-mediated disease or tumorigenesis is reported. During the As2O3 single agent treatment for APL, there are inhibitory effects on hematopoietic progenitor cells, and it takes 145 and 265 days for circulating T and B cells, and 655 days for natural killer cells (NKs) to achieve the median normal levels (Alex et al., 2018). Comprehensive studies have revealed the multifaceted effects of As(III) on each immune subset, which will be introduced in the next section (summarized in Figure 3).

Figure 3. Diverse functions of As(III) on immune cells Trivalent arsenicals (As(III)) can drive both promoting (top of figure in green) and suppressive (bottom of figure in blue) immune responses in different immune cell subsets.

15 1.1.2.2.Trivalent arsenic and granulocytes

High concentrations of As2O3 induce apoptosis of human neutrophils via generation of

H2O2, de novo protein synthesis, caspase activation, and spleen tyrosine kinase (Syk) activation (Binet et al., 2006; Antoine et al., 2010). The As2O3-induced de novo synthesized proteins include annexin-1 and heat shock proteins (Binet et al., 2008b).

Moreover, As2O3 induces endoplasmic reticulum (ER) stress within human neutrophils, which elicits either self-protective mechanisms via the activation of unfolded protein response (UPR), or cell apoptosis independent of caspase-4 activation (Binet et al., 2010a; 2010b). Importantly, As2O3 recruits the MAPKs, activates p38 and c-jun NH2-terminal, and ultimately enhances the major neutrophil functions including adhesion, migration, degranulation, and phagocytosis of opsonized sheep red blood cells (Binet et al., 2008a). Syk activation also helps in this agonistic process (Antoine et al., 2010). In addition, NaAsO2 promotes the formation of neutrophil extracellular traps (NETs), which play an important role in defense against pathogen infection (Wei et al., 2018). Moreover, it was shown in an acute

As2O3-exposed mouse model that the neutrophil cell numbers are increased in the broncho-alveolar lavage fluid (BALF) (Li et al., 2017). Collectively, neutrophils are recruited and functionally enhanced during As(III) treatment, leading toa pro-inflammatory response. For eosinophil, in a mouse model of asthma, eosinophil recruitment and a reduction of chemotaxis level in BALF were detected after As2O3 treatment (Zhou et al., 2006; Chu et al., 2010). Therefore, eosinophil functions may be impaired during As(III) exposure. For basophils, the direct effects of As(III) on basophils have not yet been investigated. However, the activation of basophils may be impaired in vivo during As(III) administration, because a decreased level of IgE, an important driver of basophil activation, was discovered in the BALF of As2O3-treated mice (Zhou et al.,

2006; Chu et al., 2010). For mast cells, it is reported that NaAsO2 inhibits anti-IgE stimulated degranulation via suppression of early tyrosine phosphorylation

16 (Hutchinson et al., 2011). Therefore, As(III) may impair mast cell activation, especially during the process of allergy (Shim et al., 2016).

1.1.2.3.Trivalent arsenic and monocytes/macrophages

Recent studies have shown that monocytes are not simply precursors of macrophages as previously thought, but in fact give rise to functionally distinct monocyte-derived cells during inflammation (Guilliams et al., 2018). As2O3 induces human monocyte apoptosis during macrophage differentiation through down-regulation of the Nf- κ B related pathway (Lemarie et al., 2006b). Moreover, As2O3 increases lipopolysaccharide (LPS) dependent expression of the inflammatory IL-8 gene, by stimulating a redox-sensitive pathway that strengthens p38-kinase activation (Bourdonnay et al., 2011).

For macrophages, high concentration of As2O3 induces cell apoptosis through a mitochondrial-dependent pathway (Sengupta et al., 2002; Srivastava et al., 2016). It changes human monocyte-derived macrophages’ morphology, reduces their adhesive capacity, decreases macrophagic surface marker expressions, and impairs phagocytosis of E.coli in vitro (Lemarie et al., 2006a). Interestingly, As2O3 potentiates macrophages’ ability to secrete inflammatory TNF-α, IL-1α, CCL18, and to induce allogeneic or autologous T cell responses (Lemarie et al., 2008; Sakurai et al., 2005). These in vitro observations are confirmed by ex vivo data (Banerjee et al., 2009; Bishayi et al., 2003). The functional changes are probably due to superoxide generation, activation of the Rho-kinase/p38-kinase pathway (Lemarie et al., 2006a), and modulation of the UPR signaling (Srivastava et al., 2013). Especially, activating transcription factor 4 (TCF4, or E2-2) protein, a UPR transcription factor, plays a key role in As2O3-mediated regulation of macrophage functions (Srivastava et al., 2016).

In addition, As2O3 also globally regulates redox-sensitive gene expression in human macrophages (Bourdonnay et al., 2009a, b). Therefore, As(III) exerts double-sided effects on macropahges by imparing their clearance capacity while enhancing the pro-inflammatory functions.

17 1.1.2.4.Trivalent arsenic and dendritic cells

In vivo we distinguished two types of dendritic cells (DCs), namely conventional dendritic cells (cDCs) and plasmacytoid dendritic cells (pDCs). MDDCs arebroadly used for research in vitro for availability reason. NaAsO2 inhibits dendritic differentiation of monocytes in vitro (Bahari et al., 2017). It also decreases MDDCs viability, maturation, phagocytic capacity, as well as their ability to secrete the pro-inflammatory cytokines IL-12 and IL-23, and to stimulate T helper cell(Th)to secrete IFN- γ (Macoch et al., 2013). The inhibition of IL-12 is mediated by the induced expression of Nrf2 (Macoch et al, 2015).

1.1.2.5.Trivalent arsenic and T cells

T cells play a central role in cell-mediated immunity. High concentrations of As(III) induce apoptosis of T helper cells (CD4+ T cell), cytotoxic T cells (CD8+ T cell), and regulatory T cells (Treg) (Gupta et al., 2003; Thomas-Schoemann et al., 2012; Tenorio et al., 2005). Both mitochondrial-mediated and the receptor-mediated pathways are involved in T cell apoptosis. Gupta et al. revealed that As2O3 induces apoptosis of CD4+ and CD8+ T cells via the mitochondrial pathway by enhancing the generation of oxidative stress and by regulating the expression of Bcl-2 family proteins (Gupta et al., 2003). When CD4+ T cells were investigated within human peripheral blood mononuclear cells (PBMC), tumor necrosis factor α released from

+ other mononuclear cells after NaAsO2 exposure induced CD4 T cell apoptosis through TNF-R1 apoptotic signaling (Yu et al., 2002). There is different sensitivity to As(III) toxicity among the T cell subsets. CD4+ are

+ more sensitive than CD8 T cells to the pro-apoptotic toxicity of NaAsO2,as discovered both in vitro and in vivo in human and mouse model (Yu et al., 2002; Duan et al., 2017; Vega et al., 2004). This difference leads to a reduced CD4+/CD8+ ratio, which is associated with immunodysfunction (Lu et al., 2015). Tregsare

+ important CD4 T cells mediating immune tolerance. In vitro tests show that As2O3, at the same concentration, preferentially induces apoptosis of human purified

18 CD4+CD25+ Tregs than CD4+CD25- effector T cells, and decreases the Treg frequency in APL patients’ peripheral blood (Xu et al., 2018). As2O3-mediated selective Treg depletion is also discovered in vivo in mouse model

(Thomas-Schoemann et al., 2012). As2O3 induces Treg apoptosis through ROS and reactive nitrogen species (RNS) generation, and the differential effect of As2O3 on Treg versus other CD4+ cells may be related to differences in the cells’ redox status (Thomas-Schoemann et al., 2012). However, some other studies report opposite results (Zhao et al., 2018a; Tohyama et al., 2013). After exposure to NaAsO2,the mRNA level of the Treg specific transcription factor forkhead box P3 (Foxp3) is upregulated in the spleen and thymus of the experimental mice (Duan et al., 2017). Interestingly, Tohyama et al report that in mitogen activated human PBMC, 5μM

As2O3 decreases the Treg frequency after treatment for 48h and, conversely, increases its frequency after 96h of culture (Tohyama et al., 2013). These results indicate that short-term exposure to As(III) may deplete Treg, in contrast to long-term exposure leading to Treg percentage increase. Upon activation, naïve CD4+ T cells proliferate and differentiate into fully functional effector T cells, which lead to either inflammatory responses or immune suppression (Wan et al., 2009). It is reported that As(III) exposure inhibits the proliferation of both human and murine activated T cells, by affecting the initial activation step of T cell receptor signaling, and by inhibiting IL-2 expression, at both the proteinandmRNA levels (Soto-Pena et al., 2008; Morzadec et al., 2012a; Vega et al., 1999; Conde et al.,

2007). In addition, NaAsO2 blocks T cells in the G1 phase and, thus, retards the entry into cell cycle (Galicia et al., 2003). More importantly, As(III) significantly alters Th cell differentiation. As2O3 inhibits IFN-γ, the characteristic Th1 cytokine, expression at both the mRNA and protein levels during the activation with anti-CD3/anti-CD28 of both human and murine T cells (Duan et al., 2017; VanDenBerg et al., 2017). However, As(III) do not change, and sometimes increase the secretion of IL-4 and IL-13, the characteristic Th2 cytokines, from activated T cells (Soto-Pena et al., 2008; Galicia et al., 2003). These observations indicate that As(III) probably alters the Th1/Th2 balance towards a Th2 response. However, In ATL patients treated with

19 arsenic/IFN/zidovudine, this combination therapy revert the Treg/Th2 cytokine profile at diagnosis towards a Th1 profile, with a diminution of IL-10 (Kchour et al., 2013). This effect is probably due to the depletion of ATL leukemia cells that disrupt the immune system initially. Given that many malignancies go along with immunosuppression, successful As(III) treatment will not augment immune restrain, but rather induce ‘immune-reactivation’ during cancer depletion. Moreover, As(III) is found to be an especially potent inhibitor of IL-17. Ata concentration which is not able to affect IFN-γ secretion by Th1 cells, NaAsO2 almost totally blocks the IL-17 secretion by human Th17 cells, via mRNA reduction of the retinoic-related orphan receptor (ROR)C gene which encodes RORγt, the key transcription factor of Th17 (Morzadec et al., 2012b). Since the Th17/Tregbalanceis critical in keeping immune homeostasis, As(III) treatment may lead to a regulatory response due to the potent inhibition on Th17. The effects of As2O3 on apoptosis and functions of CD4+ T cell are summarized in Figure 4.

+ Figure 4. Mechanisms of action: As2O3 effects on CD4 T cell

20 1.1.2.6.Trivalent arsenic and B cells

Clinical relevant concentrations of As2O3 blocks the mitogen-mediated B-cell differentiation towards plasmacytes, and their IgM secretion (Rousselot et al., 1999).

In addition, As2O3, together with leflunomide, reduces the deposition and expression of IgG and IgM in both the xenograft and recipient sera in a heart xenotransplant model (Jiao et al., 2016). Collectively, As(III) treatment probably leads to a global suppression of the humoral immune system.

1.1.2.7.Trivalent arsenic and natural killer cells

As2O3 reduces hematopoietic stem cell differentiation to NK cells, leading to a delayed NK cell reconstitution when treating APL (Alex et al., 2018). As2O3 also facilitates NK cell mediated cytotoxicity towards several cancer cell lines, through modulation of both NK cell receptors and malignant cell ligand profile (Alex et al,

2018; Kim et al., 2008a). Collectively, As2O3 may enhance the cytotoxicity of NK cells.

1.1.2.8.Trivalent arsenic and inflammasomes

In the human monocyte cell line THP-1 and mouse bone marrow derived macrophages, As2O3 inhibits NLRP3 inflammasome and subsequent IL-1β and IL-18 secretion by targeting PML (Lo et al., 2013; Ahn et al., 2018). Moreover, As2O3 inhibits NLRP1, NAIP5/NLRC4 inflammasomes in the same subset of cells (Maier et al., 2014). However, in the human keratinocyte cell line and mouse skin tissue,

NaAsO2 promotes IL-1β and IL-18 secretion via AIM2 inflammasome activation (Zhang et al., 2016). In addition, the inflammasome NALP2 polymorphism is associated with arsenic-induced skin lesions in human (Bhattacharjee et al., 2013). Therefore, As(III) exerts bi-directional regulation on inflammasomes, which is probably tissue dependent.

21 1.1.3.Trivalent arsenic in mouse models of immune-mediate diseases

1.1.3.1.Autoimmune and inflammatory diseases

Recently, As(III) was found to have therapeutic efficacy in several mouse models of autoimmune and inflammatory diseases (summarized in Table 2). As2O3 achieves quasi-total lesion regression in MRL/lpr mice, a mouse model of systemic lupus erythematosus (SLE) and lymphoproliferative syndrome, by elimination ofthe unusually activated T lymphocytes, and reduction of the related autoantibodies (Bobe et al., 2006). In the 2,4,6-trinitrobenzene sulfonic acid-induced murine model of inflammatory bowel disease, As2O3 reduces the induced colitis via Nf- κ B down-regulation and caspase-3 activation (Singer et al., 2011). In a murine model of systemic sclerosis constructed by intradermal injections of hypochlorous acid, As2O3 improves skin and lung fibrosis through ROS-mediated killing of activated fibroblasts

(Kavian et al., 2012b). Moreover, As2O3 improves the pathologic changes in a rat model of rheumatoid arthritis (RA) through a pro-apoptotic effect on RA fibroblast-like synoviocytes (Mei et al., 2011). In addition, NaAsO2 decreases the disease incidence and delays its onset in non-obese diabetic (NOD) mouse, a mouse model of autoimmune type 1 diabetes, which is due to the reduced proliferation and activation of T cells (Lee et al., 2015). Asthma is a T-helper type 2 (Th2) lymphocyte-mediated chronic inflammatory disorder characterized by airway eosinophilia and airway hyperresponsiveness (AHR).

As2O3 is observed to ameliorate the allergen-driven AHR in a mouse model of asthma, through modulation of the NF- κ B pathway, airway eosinophil recruitment and functions (Zhou et al., 2006; Chu et al., 2010). Overall, As(III) has therapeutic potential on several autoimmune and inflammatory diseases through a pro-apoptotic effect and effector-cell functional modulation.

22 1.1.3.2.Allogeneic organ/stem cell transplantation

During organ or allogeneic stem cell transplantation, human leukocyte antigen (HLA) mismatches lead to severe T- and B-cell-mediated alloreactivity, which may eventually lead to severe inflammatory transplantation complications of allograft rejection or GVHD, respectively. As2O3 is shown to prolong the allograft survival in immunocompetent mouse heart transplant models, with reduction of the proportions of CD4+ and CD8+ memory T cells, and increase of Tregs in recipient spleen and lymph nodes (Yan et al., 2009; Xu et al., 2010; Lin et al., 2011). Meanwhile, the IFN-γ expression is reduced and TGF-β expression is increased in both the recipient serum and the graft (Xu et al., 2010; Lin et al., 2011). Further studies show that As2O3 inhibits accelerated allograft rejection mediated by alloreactive CD8+ and/or CD4+ memory T cells, and prolongs allograft survival (Yan et al., 2013; Li et al., 2015a).

Moreover, in two models (one allo- and one xeno-) of islet transplantation,As2O3 is shown to prolong the graft survival by inhibiting inflammatory reactions and T cell responses (Gao et al., 2015; Zhao et al., 2018b). GVHD occurs when immunocompetent T cells in the graft recognize the recipient as foreign, and attacks the target organs. A study reveals that As2O3 prevents disease occurrence in a murine sclerodermatous GVHD model mediated by overproduction of H2O2 which kills activated CD4+ T cells and pDCs (Kavian et al., 2012a). To summarize, As(III) could be a promising drug for allo-reactivity through effector-cell modulations.

1.1.3.3.Clinical trials of trivalent arsenic on immune-mediated diseases

Given the discoveries in basic research and pre-clinical models, interests grow on As(III) as a clinical therapeutic agent in immune-mediated diseases. An on-going randomized clinical trial (NCT02966301) is currently testing As2O3 as first-line treatment of chronic GVHD. Results of a phase 2a randomized clinical trial

(NCT01738360) aiming to evaluate the therapeutic efficacy of As2O3 in SLE are expected soon.

23 Table 2. Use of trivalent arsenic in mouse models of immune-mediated diseases

Underlying disease Mouse strain Optimal therapeutic regimen Clinical efficacy Possible mechanism

SLE/lymphoprolifer MRL/lpr mice As2O3 (i.p) 5g/g/d for 2 1.Prolong survival 1. Anti-DNA autoantibody↓ ative syndrome months starting from 2 2.Quasi-total RF↓IL-18↓ IFN-γ↓nitric oxide (Bobe et al, 2006) months of age regression metabolite↓TNF-α↓ Fas ligand↓IL-10↓ in serum 2. Reduction of immune-complex deposits in glomeruli

Inflammatory TNBS-induced As2O3 (i.p) 1.Prolong survival 1. Nf-κB inhibition bowel diseases colitis mouse Prevention: 5g/g/d from 4d 2.Reduced the 2. TNF-α↓ IL-1β↓ (Singer et al, 2011) model before TNBS adminitration induced colitis IL-12↓IL-17↓IL-18↓IL-23↓ (BALB/c) Treatment: 5g/g/d when the expressions in colonic extracts disease was noted 3. Elimination of inflamed cells through apoptosis

Systemic sclerosis SSc induced by As2O3 (i.p) 5g/g/d for 6 1. Reduction in skin 1. ROS generation that (Kavian et al, 2012) intradermal weeks, simultaneously with and lung fibrosis selectively kills activated injections of HOCl injection 2. prevention of fibroblasts HOCl (BALB/c) endothelial injuries 2. Autoantibody↓ IL-4↓ and IL-13↓ from activated T cells

Rheumatoid Collagen-induce As2O3 (N/A) Inhibition of 1. Apoptosis induction of RA arthritis d arthritis rat synovial fibroblast-like synoviocytes (Mei et al, 2011) model hyperplasia and through NF-ĸB signaling inflammation pathway and caspase cascade

Type I diabetes Non-obese NaAsO2 (oral) 5g/g/d for 8 1. Risk of disease 1. Reduction of infiltration of (Lee et al, 2015) diabetic mice weeks from 8 weeks of age and onset delay immun cells in islets 2. Inhibition of T cell proliferation/activation Airway chronic inflammatory disorder

Asthma Chicken As2O3 (i.p) 4g/g/d for 7 1.Amelioration of 1. Attenuation of airway (Zhou et al, 2006; OVA-challenge days, the airway eosinophils chemotaxin and Chu et al, 2010) d murine model 30 min before each challenge hyperresponse recruitment in bronchoalveolar of asthma lavage fluid (BALF) (BALB/c) 2. IkBα expression increase and NF-ĸB decrease

24 Table 2. Continued

Underlying disease Mouse strain Optimal therapeutic regimen Clinical efficacy Possible mechanism

Cardiac allograft Model 1.C3H→ Model 1. As2O3 (i.p) (1 1. Prolong allograft 1. Inhibition of graft lymphocyte rejection C57BL/6 g/g/d, days -3- 10)+CsA survival infiltration

(Yan et al, 2009; Model Model 2. As2O3 (i.p) (5 2. IFN-γ↓ IL-2↓ TGF-β↑ in Xu et al., 2010; 2.BALB/c→ g/g/d, days 0- 10) + both recipient serum and the Lin et al., 2011) C57BL/6 (anti-CD154 and graft Model 3. anti-LFA-1） 3. CD4+ Tm(%)↓ CD8+ Tm(%)

BALB/c→ Model 3. As2O3 (i.p) ↓ but Treg(%)↑ in the spleen of C57BL/6 (3 g/g/d, days 0- 10) + recipients (allo-primed T (anti-CD154 and anti-LFA-1 cells pre-transferred)

Cardiac allograft Model 1. Model 1&2. As2O3 (i.p) 1. Prolong 1. CD4+ Tm(%)↓and CD8+ rejection C57BL/6 → (3 g/g/d, days 0-10) accelerated allograft Tm(%)↓ in recipient spleen and (Yan et al, 2013; nude mouse rejection lymph nodes Li et al, 2015) (CD4+ Tm 2. IFN-γ↓ IL-2↓ and TGF-β↑ pre-transferred) IL-10↑ in both recipient serum Model and allograft 2.C57BL/6 → nude mouse (allo-primed CD8+ Tm pre-transferred)

Islet allgraft Model Model 1. As2O3 (i.p) (2 1. Prolong islet 1. CD4+ T(%)↓ CD8+ T(%)↓ rejection 1.BALB/c→ g/g/d,days 0-9) allograft and and Foxp3+ Treg (%)↑ in

(Gao et al., 2015; C57BL/6 Model 2. As2O3 (i.p) (10 xenograft survival recipient spleen and lymph nodes Zhao et al., 2018b Model 2. Lewis g/g/d, days 0-9) 2. IFN-γ↓ IL-2↓ and TGF-β↑ in rats→C57BL/6 recipient serum and allograft

Chronic GVHD B10.D2→ As2O3 (i.p) (5 g/g/d, 3 1.Reduction of the 1. Activated CD4+ T cells(%)↓ (Kavian et al, 2012) BALB/c weeks from day 7 post fibrotic changes and and pDCs(%)↓ in splenocytes transplantation) prevention of the through apoptosis clinical symptoms 2. Splenic IL-4↓, IL-17↓ and serum Anti-DNA-topoisomerase 1 autoantibody↓

Abbreviations: SLE, systemic lupus erythematosus; As2O3, arsenic trioxide; RF, rheumatoid factor; TNBS, 2,4,6-trinitrobenzene sulfonic acid; HOCl, hypochlorous acid; SSc, systemic sclerosis; ROS, reactive oxygen species; RA, rheumatoid arthritis;

NaAsO2: sodium arsenite; NF-κB, nuclear factor-kappa B; OVA, ovalbumin; IkBα, nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha; Tm, memory T cell; pDC, plasmacytoid dendritic cell

25 1.1.4.Trivalent arsenic and tumor immunotherapy

1.1.4.1.Trivalent arsenic facilitates cellular immunotherapy

The direct effects of As(III) on tumor cells are not introduced here, but arereviewed by Emadi et al (Emadi et al., 2010). The past decade has witnessed breakthroughs which bring immunotherapy to the forefront of cancer therapy. As(III) acts in cellular immunotherapy in two ways (summarized in Table 3). On the one hand, it selectively depletes tumor promoting cells. The ratio of effector to regulatory T cells (Teff/Treg) is of great importance in immune modulation, and targeting Tregs to enhance anti-tumor immune responses has become an important strategy in cancer immunology (Roychoudhuri et al., 2015). It was observed that As2O3 reduces the increased Treg numbers and Foxp3 mRNA levels in ex vivo malignant ascites (Hu et al., 2018). In two murine models of colon and liver cancer, As2O3 showed its anti-tumor effect in both in-situ carcinoma and colon cancer lung metastasis through selective depletion of the infiltrated Tregs (Thomas-Schoemann et al., 2012; Wang et al., 2016; Wang et al., 2017). On the other hand, As(III) enhances cytotoxicity of tumor-killing cells. It has been shown that As2O3 increases the cytotoxic activity of cytokine-induced killer cells and the IFN-γ secretion in vitro,whichmayhelpin tumor control (Wang et al., 2016; Wang et al., 2017). Moreover, exposure of myeloma cell lines to As2O3 increases lymphokine activated killer (LAK)-mediated killing by up-modulation of CD38 and CD54 on the myeloma cells, and increased expression of CD31 (CD38 ligand) and CD11a (CD54 ligand) on LAKs, suggesting that the improved killing is mediated by increased adhesion (Deaglio et al., 2001). In addition, exposure of NK and leukemic cells to low doses of As2O3 modulates NK cell receptors and malignant cell ligand profile in a direction that enhances NK cell-mediated cytolytic activity (Alex et al., 2018; Kim et al., 2008a). This effect was proved in a mouse model of APL, where the As2O3+NK treatment promoted longer survival as compared with As2O3 alone (Alex et al., 2018).

26 1.1.4.2.Trivalent arsenic synergize with other strategies of immunotherapy

B7-H3 (CD276) is an important immune checkpoint member of the B7 and CD28

families, which have been shown to induce antitumor immunity. As2O3 synergized with B7H3-mediated immunotherapy to eradicate hepatocellular carcinomaina mouse model, along with the generation of potent systemic antitumor immunity

mediated by CD8+ T and NK cells (Luo et al., 2006). Moreover, As2O3 was also shown in a mouse model of bladder cancer to facilitate intravesical bacillus Calmette-Guerin immunotherapy, by targeting the IER3/Nrf2 pathway (Mao et al., 2018).

Table 3. Trivalent arsenic facilitates tumor immunotherapy

Malignancy Ex vivo/mouse model Optimal therapeutic regimen Clinical Efficacy Possible mechanism Cellular immunotherapy

Malignant TILs derived from gastric As2O3 (in vitro, 5 or 10M N/A 1. CD4+ T cell (%)↓CD8+ T ascite cancer ascites for 48h) cell (%)↑Treg (%)↓with (Hu et al, Foxp3 expression↓in TILs 2018) 2. IL-10↓TGF-β↓IFN-γ↑TIL cytotoxicity↑

Colon cancer Model 1. Colon-cancer Model 1. As2O3 (i.p) 1g/g 1.Delay tumor 1. Selective Treg depletion (Thomas-Scho bearing mouse (BALB/c, once at day 10 of tumor growth with Foxp3 expression↓in emann et al, injection s.c. CT-26 colon injection 2.Inhibition of spleen and tumor tissue, due

2012; Wang et cancer cell) Model 2. As2O3 (i.v) 6g/g lung metastasis to ROS and RNS al, 2016) Model 2. Lung metastasis every 2 days for 2 weeks and prolong mouse accumulation in Tregs. model of colon cancer from day 3 of tumor injection survival 2. Cytokine induced killer (BALB/c, injection i.v. cells (in vitro): CT-26) cytotoxicity↑IFN-γ↑

Hepatic cancer Hepatic cancer bearing As2O3 (i.v) 6g/g every 2 1. Prolong mouse 1. Selective Treg depletion (Wang et al, mouse (KM strain, H22 days for 2 weeks right after survival and delay 2. Serum 2017) hepatic cancer cell liver liver tumor implantation tumor growth IFN-γ↑IL-10↓TGF-β↓ implantation) 3. Abdominal 3. IOD of CD3+T adhesion and cell↑Foxp3+ cell↓in tumor ascites↓

Multiple Co-culture of LAK and As2O3 (in vitro,0.5Mfor N/A 1. Increased lysis via CD54 myeloma myeloma cell 72h) /CD38 up-regulation on (Deaglio et al, myeloma cells and 2001) CD31(CD38 ligand)/CD11a (CD54 ligand) on LAKs

27 Malignancy Ex vivo /mouse model Optimal therapeutic regimen Clinical Efficacy Possible mechanism

APL APL bearing mouse As2O3 (i.p) (5 µg/g/d, day 8 1. Prolong survival 1. Enhancement of NK cell (Alex et al, model (FVB/N, APL to day 35 of tumor injection) comparing cytolytic activity

2018; Kim et blast cells from AND syngeneic NK cells As2O3+NK group 2. Alteration of NK cell al, 2008) MRP8-PML-RARa (i.v) (5*10^5 cells/dose, day and As2O3 group receptor and ligand Profile on transgenic FVB/N mouse) 8, 18, 28 of tumor injection) tumor cell Other strategies of immunotherapy Hepatic cancer Hepatic cancer bearing Intratumoral injection of 1.Transplanted 1. Generation of anti-tumor (Luo et al, mouse (BALB/c, H22 100µg B7H3 plasmid (on day hepatoma immunity relies largely on 2006) hepatic cancer cell) 14 of tumor injection) AND eradication and CD8+ T cells and NK cells

5µg As2O3 (every 2 days) regression of 2. Serum IFN-γ↑and CTL until tumor disappear metastasis activity↑ Bladder cancer Orthotopic bladder cancer Intravesical instillation of 0.2 1. Reduction of 1. Apoptosis of tumor (Mao et al, model (C3H /HeN, ml BCG (600 mg/L) AND tumor weight and cells↑via IER3/Nrf2 pathway↓

2018) human bladder cancer 0.2 ml of As2O3 (30 µmol/L) volume 2. DC(%) in cancer cells 5637) for 4 weeks (once a week) tissue↑Expression of CD83/CD86↑IL-6/IL-8 in DCs↑

Abbreviations: TIL, tumor-infiltrating lymphocytes; As2O3, arsenic trioxide; ROS, reactive oxygen species; RNS, reactive nitrogen species; IOD, integrated optical density; LAK, lymphokine activated killers; APL, acute promyelocytic leukemia; BCG, bacillus Calmette-Guerin;

1.1.4.3.Summary and perspectives

As(III), when used in a clinically relevant dose, i.e., between 0.5 M and 3 M, as

found in the plasma of APL patients treated with As2O3 (Shen et al., 1997), can target specific immune cell subsets exert immunomodulatory effects. The efficacy of As(III) on mouse models of autoimmune and inflammatory diseases highlights its therapeutic potential in humans. It emerges as a promising adjuvant to cancer immunotherapy agents in the treatment of both hematological and solid tumors. Several clinical trials with As(III) in autoimmune and inflammatory diseases are on-going to provethe ‘bench to bedside’ indication of As(III). It is noteworthy that the As(III) effect is dose, cellular and tissue dependent, which requires deeper understanding of this drug in vitro, in vivo and in the clinical practice. For this purpose, comprehensive genomic, proteomic and metabolomic profiling will be critical for identifying and validating potential molecular targets of As(III) for future therapeutic use.

28 1.2.Plasmacytoid dendritic cells in autoimmunity/alloreactivity

1.2.1.Definition of pDCs

Human plasmacytoid dendritic cells (pDCs) were initially described 20 years ago by the Liu and the Colonna groups (Siegal et al., 1999; Cella et al., 1999). pDCsmanifest a secretory plasmacytoid morphology, and produce massive amount of IFN-I following viral recognition (Swiecki et al., 2015). pDCs develop from bonemarrow hematopoietic cells and constitute 0.1% to 0.5% of PBMCs (Reizis, 2019). Human pDCs do not express the lineage-associated markers (Lin) CD3, CD19, CD14, CD16 and CD11c, but selectively express the C-type lectin CD303 (BDCA2), CD304 (BDCA4), and immunoglobulin-like transcript 7 (Dzionek et al., 2001; Cao et al., 2006). They also express CD4, CD68, ILT3, and CD123 (IL-3 receptor α-subunit), which confers IL-3 supportive role for pDC survival in vitro (Cella et al., 1999). Mouse pDCs were described in 2001, and are now identified with CD11c, B220, LY6C, bone marrow stromal antigen 2 (BST2; also known as tetherin14) and sialic acid-binding immunoglobulin-like lectin H (Siglec-H) (Nakano et al., 2001; Asselin-Paturel et al., 2001). Recenly, single-cell analysis has revealed the heterogeneity of pDCs, and identified a novel DC subset which were regardedaspDC. These cells, although expressing the pDC markers CD123, CD303 and CD304, do not secrete IFN-I. Meanwhile, upon activation, they secrete IL-12 and prime potently T and B cell responses (Reizis, 2019). Human ‘non-canonical’ pDCs express typical marker Axl, and some other distinct markers including Siglec1 and Siglec6.These cells are either named as CD123+CD11c-/lo AS DCs (Axl+ Siglec6+CD123+CD11c–/lo) (Villani et al., 2017) or pre-DCs (CD33+ CX3CR1+ CD2+ CD5+ Siglec6+)(Seeetal., 2017) in different studies. Along with the previously reported heterogeneity of pDCs (Zhang et al., 2017a; Matsui et al., 2009), non-canonical pDCs are distinct cell subsets which manifest functions between those of canonical pDCs and canonical cDCs. Moreover, non-canonical Axl+ pDCs are also identified in mouse, which lack IFN-α secretion, but have increased ability to induce T cell proliferation (Dekker et al.,

29 2018). Given the observations above, it was hypothesized that pDC are bona fide interferon-α producing cells, but their reported interleukin-12 (IL-12) production and T/B cell allostimulatory capacity are attributed to “contaminating” non-canonical pDCs. However, the capacity of pDCs to differentiate into cDC-like cells discovered in both human and mouse reveal a intrinstic plasticity of pDC, which can’t be explained by this “contamination” theory (Grouard et al., 1997; Schlitzer et al., 2011). Despite similarities in functions between non-canonical pDC and canonical cDC, comparison of gene expression profiles also reveal that non-canonical pDCs co-express signature transcription factors of both canonical pDCs (TCF4)and canonical cDCs (ID2, inhibitor of DNA Binding 2), respectively (Reizis, 2019). Collectively, non-canonical pDCs may evolutionarily stand between canonical pDC and canonical cDCs, and a spectrum between pDC and cDC is proposed in Figure 5.

Figure 5. The proposed spectrum between pDC and cDC (Reizis, Immunity, 2019) The IFN-I secretion, antigen presentation capacity, as well as the expression of key transcription factors, markers on canonical pDC, non-canonical pDC, non-canonical cDC, and canonical cDC are shown.

30 1.2.2.Development of pDCs pDCs are continuously generated from hematopoietic stem cells (HSCs) in the bone marrow (BM) via intermediate progenitors. It was initially found that the Fms-like tyrosine kinase 3 ligand (Flt3L), in the absence of other signal, drove both common myeloid precursors (CMP) and common lymphoid precursors (CLP) differentiation towards pDCs, as confirmed by in vivo reconstitution assays of both murine (D’Amico et al., 2003; Shigematsu et al., 2004; Sathe et al., 2012) and humanCMPs and CLPs (Ishikawa et al., 2007). Within the myeloid pathway, the CMPs develop firstly into earlier precursors named myeloid precursors (MP), which is identified in mouse as lineage (Lin) markers (CD11b, CD3, CD19, NK1.1, Iab, CD11c, B220, TER-119, and Gr1) negative, stem cells antigen-1 (Sca-1) negative, c-Kit (or CD117) high, and CX3C chemokine receptor 1 (CX3CR1) negative. MPs subsequently differentiate into macrophage and DC precursors (MDP), which are identified in mouse as Lin-CX3CR1+CD11b-c-KithiFlt3+, and macrophage colony-stimulating factor receptor (M-CSFR or CD115) positive. Finally, MDPs give rise to monocytes and common DC progenitors (CDP) (Fogg et al., 2006; Liu et al., 2009; Auffray et al., 2009). CDPs were identified in the bone marrow of mice as Lin-c-KitintFlt3+M-CSFR+IL-7R-,and differentiate exclusively into pDCs and pre-DCs (progenitors of cDCs) (Naik et al., 2007; Onai et al., 2007). Later it was found that only a subset of CDPs express M-CSFR. A group of pDC-biased M-CSFR negative (M-CSFR-) progenitors (M-CSFR- CDP) were identified in mouse bone marrow, which express Lin-c-Kitint/loFlt3+IL-7R-M-CSFR-. These cells express high-level of TCF-4 (E2-2), the essential and specificpDC transcription factor, and have prominent pDC differentiation potential (Onaietal, 2013). These cells could be derived from either M-CSFR+ CDPs or lymphoid-primed multi-potent progenitors (LMPPs), with the activation of cytokines that upregulate TCF4 expression, i.e M-CSF or thrombopoietin, in vitro (Onai et al, 2013). The plasticity of M-CSFR+ CDPs towards M-CSFR- CDPs give rise to a doubt whether

31 M-CSFR+ CDPs give rise to pDCs exclusively via these M-CSFR- down-stream progenitor, or these two cells exist in parallel. In retrospect, the pro-DCor“CDP” population previously isolated from Flt3L-stimulated BM cultures, regardless of M-CSFR expression, had given a good yield of pDC, suggesting that it could include the M-CSFR- progenitors (Naik et al., 2007). Collectively, the ‘refined’ CDPs were proposed, which express Lin-c-Kitint/loFlt3+IL-7R- and include M-CSFR+ (TCF4lo)and M-CSFR- (TCF4hi) cells which preferentially give rise to cDCs and pDCs, respectively (Onai et al, 2013; Shortman et al., 2013). Afterwards, the M-CSFR- DC progenitor differentiates into a pDC progenitor, which shares most properties with mature pDC, but does not express CCR9, and express low class II major histocompatibility complex (MHC II). CCR9- pDC progenitors account for around 20% of bone marrow pDC, and are also identified in circulation. They could migrate into peripheral organs, and undergo tissue-specific differentiation into either terminal CCR9+ pDCs or cDC-like cells (Schlitzer et al., 2012). The plasticity of this CCR9- pDC progenitor indicates that the conversion of pDC to cDC could happen close to terminal differentiation. pDCs of the lymphoid origin are investigated in detail more recently. A pDC progenitor was recently identified within the IL-7R+ lymphoid precursors (LP). SiglecH+Ly6D+ double positive LP gave rise exclusively to pDCs when cultured in the presence of Flt3L (Rodrigues et al., 2018). Furthermore, the Ly6D+SiglecH–single positive LPs were able to differentiate into pDCs and B cells, depending on the expression of IRF8 or EBF1, respectively (Rodrigues et al., 2018). This common progenitor population of pDCs and B cells were also identified following single-cell RNA-seq analysis (Herman et al., 2018). Importantly, the non-canonical Axl+ mouse pDCs are derived from lymphoid progenitors shared with B cells, which offers an explanation for their distinct function, and indicates that their Axl+ human counterparts are probably also within the lymphoid pathway (Dekker et al., 2018). When comparing the proportion of myeloid and lymphoid originated pDCs, it seems that the majority of pDCs are derived through the myeloid pathway, with the lymphoid originated pDCs as minority. Lineage tracing using the CDP (myeloid

32 origin) marker Csf1r showed that the majority (~80%) of pDCs became labeled (Loschko et al., 2016). Furthermore, progenitors with transcriptomic features of pDCs emerge before lymphoid progenitors (Upadhaya et al., 2018) and pDCs develop from stem cells in vivo with the same kinetics as myeloid cells including cDCs (Sawai et al., 2016). The development of pDCs are schematically shown in Figure 6. Flt3 and its ligand Flt3L are crucial for pDC, as well as the lymphoid tissue-resident cDC development (Gilliet et al., 2002; O’Keeffe et al.,2002; Schmid et al., 2010). Mice genetically deficient in Flt3L, or Flt3 are poor producers of pDC and cDC in vivo (McKenna et al., 2000; Waskow et al., 2008). During the pDC development, Flt3L activates the STAT3 pathway, and the subsequent expression of TCF4, the master transcription factor of pDC development (Laouar et al., 2003; Li et al., 2012a). Moreover, a synergistic effect of IFN-I and Flt3L was identified during the differentiation of CLPs towards pDCs, where Flt3L induced IFN-I in CLPs, resulting in up-regulation of Flt3 that facilitated survival, proliferation, and differentiation of CLPs (Chen et al., 2013). Another important cytokine promoting pDC development is M-CSF (encoded by csf-1), which is able to drive pDCs and cDCs from BM precursor cells in vitro and in vivo (Fancke et al., 2008). This observation is consistent with the fact that part of CDPs express M-CSFR. Notably, the csf-1 deficient mice produce less pDC and cDC in vivo (MacDonald et al., 2005). Meanwhile, the granulocyte-macrophage colony-stimulating factor (GM-CSF) blocks the Flt3L-induced pDC development through activation of the STAT5 pathway, and subsequent suppression of the transcription factor IRF8 (Gilliet et al., 2002). The pDC transcription program seems to initiate from progenitors expressing IRF8 (Upadhaya et al., 2018). The specific development of pDCs requires the E protein transcription factor TCF4 (Cisse et al., 2008; Ghosh et al., 2010). Indeed,TCF4is responsible for pDC development through all pathways because no pDCs are generated without TCF4 (Shortman et al., 2013). As master regulator, TCF4 acts with its transcription co-factors including SPIB, IRF8, RUNX2 etc, which are involved in the development, homeostasis and function of pDCs (Swiecki et al., 2015).

33 Figure 6. Possible pathways of plasmacytoid dendritic cell development (modified from Shortman et al, Adv Immunol, 2013) Major (heavy arrows) and minor (light arrows) hematopoietic pathways found to have a potential to produce plasmacytoid dendritic cells (pDC) or conventional dendritic cells (cDC) are outlined. The intermediate precursors are: HSC, hematopoietic stem cells; LMPP, lymphocyte primed multipotent precursors; CMP, common myeloid precursors; CLP, common lymphoid precursors; CDP, common dendritic precursors; pre-cDC, precursors of cDC; LP, lymphoid progenitor; p-pDC: precursors ofpDC;. The relative importance of different pathways and the pDC to cDC ratio will vary with environmental conditions. Especially, the CCR9- pDC progenitor could differentiate to terminal CCR9+ pDCs, or CD11b+MHCIIhigh cDC-like cells, which is tissue dependent. There seems to be a direct M-CSFR- CDP differentiation from LMPP, with mechanisms yet to be understood.

34 1.2.3.Functions of pDCs

1.2.3.1.Trafficking of pDCs

Circulating pDCs migrate from the blood compartment into lymph nodes majorly through high endothelial venules (HEVs) but not afferent lymphatic vessels, the route for cDC trafficking (Sozzani et al., 2010; Russo et al., 2016). In addition to secondary lymphoid organs, pDCs also migrate from blood into peripheral tissues. pDCs constitutively express CXCR4, and the CXCR4-CXCL12 signaling is crucial for the early development of pDCs within the bone marrow stromal cell niches, and their migration towards splenic white pulp (Kohara et al., 2007; Umemoto et al., 2012). In addition, CXCR4 also mediates the recruitment of pDCs towards tumors expressing CXCL12 (Zou et al., 2001). CCR2, which is expressed in only a fraction of pDCs, is responsible for the migration of pDCs towards skin under topical imiquimod treatment (Sawai et al., 2013; Drobits et al., 2012). More recently, CCR2 is found to be important for pDC migration in steady-state and during inflammation, and that microbiota supports homeostatic trafficking of pDCs by eliciting constitutive levels of the chemokine ligand of CCR2, CCL2 (Swiecki et al., 2017). CCR5 is responsible for the migration of pDC from bone marrow to peripheral (Sawai et al., 2013). CCR6 and CCR10 express on a group of tonsil pDCs, and facilitate the homing of these cells towards inflamed epithelia expressing their respective ligands CCL20 andCCL27 (Sisirak et al., 2011). CCR7 contributes to lymph node homing of pDCs in both steady-state and inflammation (Seth et al., 2011). CCR9 mediates thymic entry of pDCs and the subsequent deletion of antigen-specific thymocytes (Hadeibaetal., 2012). Meanwhile, CCR9 is also the key homing receptor for pDC to the small intestine and colon (Wendland et al., 2007; Wurbel et al., 2011). During inflammation, additional molecules are involved in pDCs homing to lymph nodes, such as PSGL-1, the ligand for E-selectin, β1andβ2 integrins and the chemokine receptors CCR5 and CXCR3 (Krug et al., 2002; Diacovo et al., 2005)

35 The mucosal addressin cell-adhesion molecule-1 is found to be important for the β7 integrin-dependent intestinal localization of pDCs in homeostasis (Clahsen et al., 2015). In addition, the ChemR23 expressed on pDCs facilitates their recruitment towards chemerin expression tissue including HEVs in secondary lymphoid organs and in inflamed skin lesions (Parolini et al.,2017; Vermi et al., 2005; Albanesi et al., 2009). IFN-β also mediates the homing of pDCs towards lymph nodes in steady-state and during inflammation (Gao et al., 2009). Despite receptors expressed on the surface, several intracellular signaling molecular were identified to play decisive role in pDC migration. The pDC-specific CD2-associated protein is correlated with pDCs’ lymph node migration under conditions of inflammation, which is correlated with defects in actin dynamics (Marafioti et al., 2008; Srivatsan et al., 2013). Moreover, DOCK2, a hematopoietic cell-specific CDM family protein, is found to be indispensable for migration of plasmacytoid DCs (Gotoh et al., 2008). The proteins involved in pDC trafficking are summarized in Figure 7.

CCR2, IFN-β IFN-β

IFN-β

Figure 7. Factors involved in pDC trafficking (modified from Swiecki et al,Nat Rev Immunol, 2015)

36 1.2.3.2.Mechanisms of type I interferon production by pDCs

The type I interferon secretion by pDCs is mainly (albeit not exclusively) mediated through the activation of the endosomal toll-like receptors (TLRs) TLR7 and TLR9, and the subsequent myeloid differentiation primary response protein 88 (MYD88)-interferon regulatory factor 7 (IRF7) pathway (Honda et al., 2005a). Another important signaling is the MYD88-Nf- κ B pathway, which leads to the secretion of pro-inflammatory cytokines (e.g TNF-α) and chemokines, as well as the expression of co-stimulatory molecules (Zhang et al., 2017b). TLR7 sensesRNA viruses, endogenous RNA and synthetic oligoribonucleotides, whereas TLR9 detects prokaryotes containing unmethylated CpG-rich DNA sequences, endogenous DNA and synthetic CpG oligodeoxyribonucleotides (Swiecki et al., 2015). It is recently found that the endosomal TLR sensing is not exclusive, and multiple alternative sensing systems initiated by cytosolic receptors are recently revealed. pDCs sense Polysaccharide A through the non-endosomal TLR2, and stimulate CD4+T cell to secrete protective cytokine IL-10 (Dasgupta et al., 2014). Importantly, the cytosolic DNA sensor, cyclic guanosine monophosphate-adenosine monophosphate (cGAMP) synthase, is activated after binding to generic DNA, and generates the second messenger cGAMP, which binds to and activates stimulator of interferon genes, thereby triggering a IRF3-mediated type I IFN interferon production (Sun et al., 2013; Li et al., 2013). Moreover, the cytosolic RNA sensor, retinoic acid-inducible gene I sense replicate viral RNA, recruits the mitochondrial antiviralsignaling protein adaptor protein, and finally leads to type I IFN interferon also via IRF3 (Kumagai et al., 2009; Bruni et al., 2015). Other cytosolic sensors include (DExD/H)-box helicases DHX36 and DHX9, with the former selectively binding to CpG-A and activating the IRF7 pathway, and the latter selectively binding CpG-B, leading to subsequent activation of the Nf-ĸB pathway (Kim et al., 2010). Collectively, these cytosolic receptor initiating pathways may play important supplementary role in pDC immunity.

37 Despite their low frequency, pDCs produce most of the type I IFN that is detectable in the blood following viral infection. Meanwhile, upon in vivo CpG activation in mouse, the IFN-I response is mediated exclusively by pDCs (Asselin-Paturel et al., 2003; Blasius et al., 2004; Cervantes-Barragan et al., 2012). Given that TLR-7 and TLR-9 are also expressed on B cells and several myeloid cell types, a important question is raised: why and how pDCs, but not other cell types, activate this signaling pathway for IFN induction? So-far, it seems that the this phenomenon can not be explained by a single mechanism or molecule, but instead a combination of cellular processes contribute to it. pDCs consitutively express higher level of IRF7 than other cell types (Izaguirre et al., 2003), and secrete IFN-I rapidly and independently of the IFN-I receptor IFNAR-based feedback signaling, which is required for most cell types for IFN-I secretion (Barchet et al., 2002). In addition, CpG-A is retained for long periods in the early endosome of pDCs, together with the MYD88-IRF-7 complex, whereas in cDCs CpG-A is quickly transferred to lysosomal vesicles (Honda et al., 2005b; Guiducci et al, 2006). Moreover, protein kinase C and casein kinase substrate in neurons (PACSIN) 1 is specifically expressed on pDCs, and is involved in the type I IFN, but not the proinflammatory cytokine secretion in response to TLR9 ligand (Esashi et al., 2012). Given that both IRF7 and the Nf-κB pathways depend on MYD88 and UNC93B, why and how pDCs ‘select’ the IRF7 pathway to secrete IFN-I is under intensive investigation. The compartment in which TLRs encounter their ligands seems to be the decisive factor (Guiducci et al, 2006). For TLR-9, activation of TLR9 bythe multimeric CpG-A occurs in the early endosomes and leads to activation of the MYD88-IRF7 pathway, whereas monomeric CpG-B rapidly translocates to the Nf-ĸB endosome, and induce the MYD88-Nf-ĸB pathway (Honda et al., 2005b; Guiducci et al, 2006). CpG-C, or the similar but more potent CpG-P, transfers to the two endosomes and induce activation of both pathways (Vollmer et al., 2004). Another important factor mediating the preferential secretion of IFN-I, but not proinflammatory cytokines, is the adapter protein-3 (AP3) (Sasai et al., 2010). The AP3 adaptor complex and the AP-3-interacting cation transporter Slc15a4 are

38 responsible for the trafficking of TLR9 from the early endosome to a specialized lysosome-related organelle (IRF7 endosome), where TLR9 can engage molecules such as TRAF3 and IRF7, activate the MYD88 signaling complex, and finally leads to IFN-I secretion (Blasius et al., 2010). In addition, a non-canonical recognition process was identified when pDCs were exposed to large DNA containing immune complexes. This process is named microtubule-associated protein 1A/1B-light chain 3 (LC3)-associated phagocytosis (LAP), which requires the convergence of the phagocytic and autophagic pathways (Henault et al., 2012). It is recently found that LAP is also involved in CpG induced TLR-9 sensing, where LC3 directly binds to IKKα, and the LC3-IKKα complex is further recruited to endosomes containing TLR9, and induce the subsequent IFN-I secretion through IRF7 activation (Hoshino et al., 2006; Hayashi et al., 2018). Moreoover, the recognition of certain single-stranded RNA (ssRNA) viruses by TLR7 also requires autophagy process, which transports cytosolic viral replication intermediates into the lysosome, and induce the subsequent IFN-α secretion (Lee et al., 2007). Other key components involved in the MYD88-IRF7 signaling are osteopontin (Shinohara et al., 2006), and the mammalian target of rapamycin (mTOR) signaling (Cao et al., 2008). IRF5, on the other hand, is essential for MYD88-NF-ĸB signaling (Takaoka et al., 2005). As mentioned above, pDCs could produce IFN-I independent of IFNAR-based feedback signaling. Relatively, pDCs response to vesicular stomatitis virus (VSV) and mouse cytomegalovirus (MCMV) in-vivo independent of IFNAR (Barchet et al., 2002; Tomasello et al., 2018). Unexpectedly, IFNAR knock-out pDCs are not impaired for IFN-I production in response to MCMV, despite reduction in their IRF7 expression (Tomasello et al., 2018). By contrary, IFNAR appears necessary for the fullIFN-I response to TLR ligands in vivo (Asselin-Paturel et al., 2005; Blasius et al., 2010), and IFNAR is found to be essential for the optimal function of CD34+ hematopoietic stem and progenitor cells (HSPC) derived pDC percursor in vitro (Laustsen et al., 2018). Collectively, IFNAR signaling and high level of IRF7 may be dispensable for normal IFN-I response by pDCs towards specific viruses, however is needed for the entire pDC function.

39 pDCs produce high amounts of IFN-I during MCMV infection in-vivo through the TLR9-to-MyD88-to-IRF7 signaling pathway. Surprisingly, this process is dependent on neither AP3-driven endosomal routing nor the autophagy-related 5 (Atg5)-dependent LAP, indicating a potentially unknown mechanism involved in TLR sensing (Tomasello et al., 2018). The TLR9 sensing and signaling, as a typically example, is shown in Figure 8.

Figure 8. TLR9 signaling (Swiecki et al, Nat Rev Immunol, 2015) DNA-IC, DNA immune complexes; CpG-ODN, CpG oligodeoxyribonucleotides; AP3, adaptor protein 3; LC3, microtubule-associated protein 1A/1B-light chain 3;

Apart from the described cell-intrinsic mechanism for type I interferon production, recently studies indicate the involvement of co-operative mechanism. It was previous discovered that in vivo pDC activation induced their tight clustering, with unknown mechanism (Asselin-Paturel et al., 2005). It was also observed that pDCs produce higher IFN-I when cultured in vitro with high cell density, and was proved in a single-cell activation assay (Kim et al., 2014; Wimmers et al., 2018). Apart from the important role of autocrine/paracrine mechanisms, cell-cell contacts between

40 clustered pDCs were also found to prime the IFN-I secretion (Hagberg et al, 2011; Saitoh et al., 2017; Tomasello et al., 2018). Lymphocyte function-associated antigen 1 (LFA-1) belongs to the integrin superfamily of adhesion molecules, which is found to be responsible for cell-cell contact of pDCs. LFA-1 promotes IFN-I production by human (Hagberg et al, 2011) and mouse (Saitoh et al, 2017) pDCs in vitro. Moreover, the optimal activation of IFN-I production by pDC in vivo during MCMV infection or TLR9 ligand activation also requires LFA-1 expression (Tomasello et al., 2018). Therefore, the optimal IFN-I responses of pDCs in vivo and in vitro require cell-cell interaction. The cell-intrinsic and cooperative sensing of pDCs are schematically shown in Figure 9.

Figure 9. Cell-intrinsic and cooperative sensing of pDCs (modified from Reizis, Immunity, 2019) (A) Activation by TLR ligands or viruses that do not infect pDCs. (B) Activation by viruses that infect and replicate in pDCs. An infected cell is highlighted in dark gray.

41 1.2.3.3.pDCs as antigen-presenting cells pDCs express MHC class II molecules and co-stimulatory markers, which work together to cross-prime CD8+ T cells and present antigen to CD4+ T cells (Swiecki et al., 2015). Antigen presentation by pDCs can lead to CD4+ T cell activation or tolerance induction, depending on the context. Upon activated by influenza virus, pDCs drive a potent Th1 polarization (Cella et al., 2000). Meanwhile, CD40L activated pDCs induce strong Th2 response (Rissoan et al., 1999). Nevertheless, upon TLR7 activation or TGF-β exposure, pDCs favor Th17 commitment (Yu et al., 2010; Bonnefoy et al., 2011). When pDCs are either unstimulated or alternatively activated, they express the context-dependent expression of indoleamine 2,3-dioxygenase (IDO) (Munn et al., 2004; Fallarino et al., 2004; Boasso et al., 2007; Pallotta et al., 2011), inducible costimulator ligand (Ito et al., 2007), OX40 ligand (CD252) (Diana et al, 2009), PD-L1 (Diana et al., 2011) and Granzyme B (Jahrsdörfer et al., 2010),and induce regulatory T cell responses during viral infection, tumor, and autoimmune disorders. It is reported that pDCs transported antigens to the thymus and deleted antigen-specific thymocytes in a CCR9 dependent manner, contributing to immune tolerance (Hadeiba et al., 2012). In addition, CCR9+ pDCs inhibit acute GVHD in a mouse model induced by allogeneic CD4+ T cells (Hadeiba et al., 2008). Moreover, in mouse models of lung inflammation and asthma, adoptive transfer of CD8α+ pDCs induces Foxp3+ regulatory T cells and prevents the airway hyper-reactivity, but not CD8α- pDCs (Lombardi et al., 2012; Maazi et al., 2013). Indeed, CD8α+ pDC is not a certain subset of pDC as CD8α is expressed in variable manner on pDC and can be upregulated upon activation (Asselin-Paturel et al., 2001; Dalod et al., 2002). To identify the antigen-presenting role of specific surface molecules expressed on pDCs, monoclonal antibodies (mAb) were used. By using a mouse model expressing human CD303 specifically in pDCs together with a anti-CD303 mAb, it was confirmed that antigen delivery to pDCs through CD303 decreased effector CD4+ T cells and preserved Foxp3+ Tregs, thus promoting tolerance (Chappell et al., 2014).

42 Using similar methods, it was found that Siglec-H-mediated antigen delivery induced a hyporesponsive state of T cells via reducing expansion of CD4+ T cells and inhibiting Th1/Th17 cell polarization but not conversion to Foxp3+ Tregs (Loschko et al., 2011a; Meyer-Wentrup et al., 2008). On the other hand, antigen delivered to murine pDCs via BST2 in combination with TLR agonists as adjuvants is specifically presented by pDCs in vivo and elicits strong cellular and humoral immune responses (Loschko et al., 2011b). These results indicate that antigens targeted to pDCs can either lead to pro-inflammatory or regulatory immune responses, which is probably dependent on the type of antigen, the surface molecule, as well as the mode ofpDC activation. Apart from antigen-presenting to CD4+ T cells, pDCs are also able to perform cross-presentation to CD8+ T cells. Seminal study showed that human pDCs cross-presented vaccinal lipopeptides and HIV-1 antigens from apoptotic cells to specific CD8+ T cells, with similar potency as cDCs (Hoeffel et al., 2007). Later, mouse splenic pDCs were found to cross-present and generate effective CD8+ Tcells in vivo after activation, but not at steady-state (Mouriès et al., 2008). Mechanism for CD8+ T cross-priming is due to the recycling endosomes within pDCs, which offer sites for loading peptide onto MHC class I, and subsequent cross-presentation to CD8+ T cells (Di Pucchio et al., 2008). Therefore, pDCs may contribute to viral infection through specific stimulation of antiviral cytotoxic CD8+ Tcells.

1.2.4.pDCs in autoimmunity

Autoimmune disease arises from an abnormal immune response of the body against normal tissues. The role for pDCs in the pathogenesis of autoimmune diseasewas proposed soon after the definitive description of pDCs (Rönnblom and Alm., 2001). So-far, pDCs have been revealed to be implicated in almost all autoimmune and inflammatory diseases (Panda et al., 2017). SSc is a complex autoimmune disease interconnecting vasculopathy, autoimmunity, and fibrosis features. Abnormally activated pDCs are infiltrated in the target organs

43 such as skin, lung and bronchoalveolar lavage, and secrete IFN-α and CXCL4,which are both hallmarks of SSc (Ah Kioon et al., 2018; Kafaja et al, 2018; van Bon etal., 2014). It is reported that pDCs are responsible for most of the IFN-α secretioninSSc patients, and plays a critical role during the process of fibrosis (Kim et al., 2008b). Moreover, abnormal T and B cell responses are both key factors in the pathogenesis of SSc (de Bourcy et al., 2017; Ricard et al., 2019; Liu et al., 2016). Since pDCs can induce T/B cell activation, an interaction probably exists between pDC hyperactivation and abnormal T and B cell responses in SSc. Moreover, in a SSc mouse model with bleomycin-induced fibrosis, depletion of pDCs could not only prevent the disease initiation, but ameliorate the established fibrosis (Ah Kioon et al., 2018; Kafaja et al, 2018). pDCs are also found to play pathogenetic role in SLE, the chronic systemic autoimmune disease characterized by autoantibody production, complement activation, and immune complex deposition (Panda et al., 2017). Raised serum levels of IFN-α and constitutive upregulation of IFN-α–inducible genes have been observed in SLE patients, and are correlated with both disease activity and severity (Bengtsson et al., 2000, Baechler et al., 2003). Moreover, pDCs are decreased in peripheral blood, activated and accumulated in the tissue lesions of SLE patients, indicating a migration process (Blomberg et al., 2001; Tucci et al., 2008). pDCs sense the immune complexes formed by antinuclear antibodies and endogenous nucleic acids through the Fc receptor CD32, and produce IFN-α subsequently through TLR7 and/or TLR9 signaling (Båve et al., 2003; Barrat et al., 2005; Means et al., 2005). This process could be enhanced by high mobility group box 1 (HMGB1), a nuclear DNA-binding protein released from necrotic cells (Tian et al., 2007). In addition, activation by TLRs for SLE pDCs is responsible for their glucocorticoid (GC) resistance (Guiducci et al., 2010; Lepelletier et al., 2010). The pathogenic role of pDCs in SLE have been revealed more clearly recently with the development of genetic models. Diphtheria toxin receptor based transient depletion of pDC in lupus-prone mice before disease onset resulted in amelioration of disease (Rowland et al., 2014; Davison and Jørgensen, 2015). In addition, constitutive

44 impairment of pDCs by monoallelic deletion of tcf4 (gene encoding TCF4) strongly reduced autoantibody production and all disease manifestations in two different spontaneous models of SLE (Sisirak et al., 2014). Moreover, in healthy donors, pDCs drive the differentiation of regulatory B cells (Bregs), which conversely restrain IFN- α production by pDCs via IL-10 release. However in SLE, pDCs promote plasmablast differentiation but fails to induce Breg cells (Menon et al., 2016). In type I diabetes, pDCs are proportional expanded in patients at disease onset (Allen et al., 2009). Indeed, pDCs are recruited and activated in the pancreas of nonobese diabetic (NOD) mice (Diana et al., 2013), and that tcf4 knockout in NOD mice ameliorated insulitis and reduced diabetes incidence (Hansen et al., 2015). pDCs are not always disease-promoting. In certain diseases, the role of pDCs may be protective. pDC deficiency in a genetic mouse model led to exacerbation of joint inflammation. Moreover, topical application of the TLR-7 agonist imiquimod recruit activated pDCs to arthritic joints and ameliorated arthritis (Nehmar et al., 2017). Depletion of pDCs prior to psoriasis induction ameliorate psoriasis in a genetic mouse model (Glitzner et al., 2014) whereas not in a chemical-based model (Wohn etal., 2013). Even though pDCs infiltrate intestinal mucosa of inflammatory bowel disease (IBD) patients, Monoallelic tcf4 deletion in two genetic models of IBD had no effect on the development of colitis (Sawai et al., 2018). In addition, studies show conflicting role of pDC in atherosclerosis, as constitutively or transient depletion of pDCs prevented (Sage et al., 2014) or aggravated (Yun et al., 2016) atherosclerosis in genetic mouse models, respectively.

1.2.5.pDCs in alloreactivity

Alloreactivity is identified when immunocompetent T cells in the donated tissue (the graft) recognize the recipient (the host) as foreign, and attacks the target organs in the immune compromised hosts (Ferrara et al., 2009). In clinical condition, alloreactivity happens during GVHD, a major inflammatory complication for patients that underwent allogeneic hematopoietic cell transplantation (allo-HCT). Host DCs are

45 critical antigen presenting cells (APCs) which initiate aGvHD via direct host antigen presentation, whereas donor APCs present host antigen to donor T cells via indirect host antigen presentation (Reviewed by Yu et al., 2019). Pioneering study showed that MHC-expressing host pDCs alone was sufficient to prime alloreactive Tcellsand cause GVHD, and pDC maturation was mediated by the inflammatory environment created by irradiation (Koyama et al., 2009). However, in vivo depletion of host pDC, alone or together cDC depletion, did not ameliorate GVHD in mouse models (Lietal., 2012b). This is in consistent with studies revealing that in allo-HCT, many other cells, including donor APCs and recipient nonhematopoietic APCs are sufficient to induce GVHD with enough potency (Koyama et al., 2011). In clinical practice, cytokines with different mechanisms are used for hematopoietic stem cell mobilization. Given their roles in mobilizing DCs and T cells and in modulating their expression and function, they influence the GVHD occurrence and severity. Granulocyte-colony stimulating factor (G-CSF) treated experimental animals or donors preferentially induce T cells secreting regulatory cytokines including IL-10, partly due to mobilization of pDCs (Arpinati et al., 2000; Klangsinsirikul et al., 2002; Morris et al., 2004). Plerixafor, the antagonist of CXCR4, enhance HSC mobilization alone or combined with G-CSF. Plerixafor was shown to induce higher proportion of Tregs (Kean et al., 2011). However, addition of plerixafor to G-CSF did not ameliorate, but even exacerbate mouse GVHD (Arbez et al., 2015). The effects of pDC on the major organ of aGVHD, the gastrointestinal tract (GI), were under intensive investigation. Seminal study showed that CCR9+ pDCs were recruited to the intestines and attenuated aGVHD in mouse model induced by allogenetic CD4+ T cells, probably via induction of Tregs (Hadeiba et al., 2008). On the other hand, the pro-inflammatory Th17 cells, together with pDCs, were upregulated in the intestinal mucosa of patients with aGVHD, as compared with patients without aGVHD (Bossard et al., 2012). Moreover, this co-upregulation of pDC and Th17 were also shown in the skin of aGVHD patients, as compared with healthy individuals (Malard et al., 2013). However, these studies did not reveal a direct link between pDC and Th17 in the context of aGVHD. Apart from Hadeiba et

46 al, more studies have identified the tolerogenic effect of pDCs on inducing donor T cells against host tissues. Importantly, BM, but not G-CSF mobilized graft contained the precursor pDCs, which were shown to attenuate GVHD in mouse models (Banovic et al., 2009). This effect was probably due to the process where IFN-γ synthesized by donor T cells induced IDO synthesis by donor precursor pDCs,which induced subsequent Treg generation (Lu et al., 2011). Consisting with these pre-clinical data, unrelated BM allograft with higher content of pDC led to improved survival in GVHD patients (Waller et al., 2014). Post-transplantation reconstitution of pDCs is predictive for subsequent GVHD risk. Patients developing aGVHD after myeloablative (MA) allo-HCT were shown tohave significantly lower numbers of both circulating cDCs and pDCs compared with non-GVHD patients, and that low DC counts were associated with severe aGVHD (Vakkila et al., 2005; Horvath et al., 2009;). Similar to MA allo-HCT, low pDC count in patients receiving reduced-intensity conditioning (RIC) allo-HCT were also correlated with severe grades II to IV aGVHD (Mohty et al., 2005). Moreover, steroid treatment rapidly decreased pDC counts at all time points after transplantation (Fagnoni et al., 2004; Arpinati et al., 2004). Nevertheless, recent studies in mouse models show that not only the quantity, but the quality of DCs are altered during GVHD. On one hand, GVHD impairs the DC ability to prime the virus-specific T cells (Wikstrom et al., 2015). On the other hand, antigen-presentation through MHC II is also impaired during aGVHD, leading to Treg deficiency and consequent chronic GVHD (cGVHD) (Leveque-El mouttie et al., 2016).

47 1.3.Mouse models of acute GVHD

1.3.1.Acute GVHD (aGVHD)

1.3.1.1.Definition of aGVHD

GVHD is a major inflammatory complication for patients that underwent allogeneic HCT. GVHD occurs when immunocompetent donor T cells recognize the recipient host with intensive immune suppression as foreign and mount an immune response to allogeneic antigen-bearing cells with subsequent destruction of host tissues (Ferrara et al., 2009). GVHD can be defined as acute aGVHD and cGVHD, depending on the time of onset and clinical manifestations. The most affected organs at the onset of aGVHD are skin (81%), gastrointestinal tract (54%), and liver (50%) (Ghimire et al., 2017). Previously, it was believed that aGVHD occurs within day 100 after transplantation and cGVHD occurs beyond day 100. However, we now know that aGVHD can occur after day 100 as late-onset aGVHD, with an incidence of 11% at2 years after HCT (Holtan et al., 2016). Late-onset aGVHD can be de novo aGVHD occurring beyond 100 days after transplantation, or aGVHD recurrence during donor lymphocyte infusion or after taper/withdrawal of immune suppression (Jagasia et al., 2015). Moreover, there can be overlap of late-onset aGVHD and cGVHD. Despite decades of effort on prevention and treatment of aGVHD, it remains amajor cause of morbidity and mortality, and treatment of established GVHD can be difficult, with only about 40% of patients having a durable response to corticosteroid therapy (Hamilton, 2018a). In recent years, regimens containing immunomodulatory drugs, such as posttransplantation cyclophosphomide (PTCy), have emerged as promising strategies for aGVHD prevention (Kanakry et al., 2013). However, more progress are expected and new drugs are under investigation to improve aGVHD prophylaxis and treatment.

48 1.3.1.2.Risk factors of aGVHD

Numerous studies have identified the following risk factors for the development of aGVHD (Gale et al., 1987; Hahn et al., 2008; Flowers et al., 2011). (1) Degreeof HLA disparity (HLA mismatch or unrelated donor) (2) Donor and recipient gender disparity (female donor leads to higher risk of aGVHD) (3) Intensity of the transplant conditioning regimen (4) aGVHD prophylactic regimen used (5) Source of graft (peripheral blood or bone marrow greater than umbilical cord blood). A study conducted by the Center for International Blood and Marrow Transplant Research (CIBMTR) included a cohort of 5561 patients who underwent allo-HCT from HLA-identical sibling donors (SDs; n = 3191) or unrelated donors (URDs; n = 2370). They divided the transplants into 6 groups according to the conditioning regimen and stem cell source, and found that patients receiving SD transplants with MA + non-total body irradiation (TBI) + bone marrow and RIC + peripheral blood stem cells had significantly lower risks of severe aGVHD than patients in other treatment categories (Jagasia et al., 2012). Therefore, there is combined effects of conditioning regimen and graft source on aGVHD risk. Other less well-established risk factors include increasing age of the host (Hahn et al., 2008), the cytomegalovirusstatusof the donor and host (Cantoni et al., 2010), donor Epstein-Barr virus seropositivity (Styczynski et al., 2016), etc. The incidence and severity of aGVHD also appears to increase with pre-transplant comorbidities. In one study of 2,985 patients who underwent MA or RIC followed by allo-HCST for myeloid or lymphoid malignancies, the incidence and severity of aGVHD increased with increasing hematopoietic cell transplantation-specific comorbidity index (Sorror et al., 2014).

49 1.3.1.3.Risk stratification of aGVHD aGVHD was staged on a scale of 0 to IV based on a composite score derived from each of the three target organs (skin, liver, GI tract). An updated international consensus for aGVHD grading is shown in Table 4.

Table 4. Revised Glucksberg aGVHD grading system (Harris et al., Biol Blood Marrow Transplant, 2016)

BSA, body surface area;

In addition, many plasma biomarkers have emerged to have prognostic value for aGVHD. The MAGIC biomarkers including regenerating islet-derived 3-α (Reg3-α) and suppression of tumorigenicity 2 (ST2) predict long term outcomes in steroid-resistant aGVHD (Major-Monfried et al., 2018). Moreover, Plasma levels of miRNAs have also been discovered to be have predictive and prognostic valuein aGVHD (Xiao et al., 2013).

50 1.3.1.4.aGVHD pathophysiology aGVHD has been attributed to three stages. Initially, there is tissue damage due to conditioning that in turn activates the host APCs. Secondly, APCs activate donor T cells, also known as an afferent phase. Finally, in efferent phase, cellular and inflammatory factors work together to damage the target organs. The pathophysiology of aGVHD is schematically shown in Figure 10.

Figure 10. aGVHD pathophysiology (Ghimire et al, Front Immunol, 2017) LPS, lipopolysaccharide; Mф, macrophage; CTL, cytotoxic T lymphocyte; APC, antigen presenting cell; Treg, regulatory T cell;

Conditioning is crucial in allo-HCT in order to eradicate the underlying disease and to ensure the successful engraftment of donor cells (Jethava et al., 2017). Additionally, for many patient, the underlying disease, as well as the previous treatments have largely damaged the tissue (Ferrara et al., 2009). Consequently, danger signals expressed on damaged tissues (damage-associated molecular patterns (DAMPs))

51 and/or pathogens (pathogen-associated molecular patterns (PAMPs)) activate the host APCs (Blazar et al., 2012). The important DAMPs include uric acid and adenosine triphosphate (ATP) (Apostolova et al., 2016), and PAMPs include multiple microbiota derived signals which induce differential activation of TLRs and NOD like receptors (NLRs) (Penack et al., 2010). Indeed, loss of homeostasis of microbiota and intestinal epithelial cells are related with GVHD outcome (Mathewson et al., 2016; Peled et al., 2016). Pathophysiology of the initiation phase of aGVHD is summarized in Figure 11.

Figure 11. Initiation phase of aGVHD (Ghimire et al, Front Immunol, 2017) DAMPs, damage-associated molecular patterns; PAMPs, pathogen-associated molecular patterns; ATP, adenosine triphosphate; TLR, Toll-like receptor; NLR, NOD like receptors; P2XR, P2X receptor;

52 As introduced in chapter 1.2.5, Both donor and recipient APCs are able to present antigen to donor T cells, and these APCs are not restricted to DCs. CD4+ T cells are activated after recognition of MHC II disparity while CD8+ T cells respond to MHC I variations (Sprent et al., 1988). Transplantation from HLA-identical donor also develop GVHD, which is due to minor histocompatibility antigens (miHA) mismatch (Goulmy et al., 1996). miHAs can be presented on MHC I or MHC II, and so far already >100 miHAs have been identified and sequenced (Roy et al., 2017). Recent study based on genome-wide array indicated that number of miHA mismatches doubled in unrelated vs sibling HLA-matched transplants, and an increase in miHA antigen disparity increased risk of severe (grade III-IV) and stage 2-4 gut GVHD (Martin et al., 2017). Broadly expressed miHAs can cause both GVHD and graft-versus-leukemia (GVL) effects, while hematopoietic system-restricted miHAs are more prone to induce GVL responses (Spierings, 2014). In the efferent phase, both innate and adaptive immune cells work synergistically to exacerbate the T cell-induced inflammation (Holtan et al., 2014). Cytotoxic T cells induce the apoptosis of target cells through lytic Fas and perforin pathways (Lowin et al., 1994). Other immune effector cells have inflammatory roles in the pathophysiology of aGVHD, including neutrophils, NK cells, B cells, and mononuclear phagocytes (Holtan et al., 2014). Moreover, microbiota components such as LPS leak through damaged intestinal mucosa, which then recruits myeloid cells (monocytes/macrophages) to further produce pro-inflammatory cytokines and thus enhance the cytokine storm (Ferrara et al., 2009). By contrary, several suppressor populations such as Tregs (Edinger et al., 2003) and myeloid-derived suppressor cells (Highfill et al., 2010) have been shown to prevent aGVHD in mouse models and in patients. The final components of the effector phase of aGVHD are cytokines. They can amplify or attenuate GVH reactions, and the balance between proinflammatory and anti-inflammatory cytokines shapes the overall milieu and GVHD response (reviewed by Henden et al., 2015)

53 1.3.2.Challenges in aGVHD prophylaxis/treatment

1.3.2.1.aGVHD prophylaxis

A calcineurin inhibitor (CNI) in combination with a short course of methotrexate (MTX) have been the backbone for aGVHD prevention in most allo-HCTs (Hamilton, 2018a). MTX, a folate antagonist and antimetabolite, attenuates T-cell activation at low noncytotoxic doses, and have been used for prevention of aGVHD for long time. The first CNI was cyclosporine (CsA), and the combination of MTX and CsA is better than either drug alone in aGVHD prophylaxis (Storb et al., 1986). Another calcineurin inhibotor Tacrolimus was subsequently developed, and was found to prevent aGVHD together with MTX (Storb et al., 1993). CsA binds to cyclophilin, and Tacrolimus binds to FKBP12 (FK506-binding protein 12), but both CsA-cyclophilin and Tacrolimus-FKBP12 complexes block the phosphatase activity of calcineurin, thus preventing the dephosphorylation of nuclear factor of activated T-cells, which moves to the nucleus and increases the activity of genes coding for IL-2 and related cytokines (Liu et al., 1991). Two phase 3 multicenter prospective clinical trials showed superiority of Tacrolimus/MTX over CsA/MTX in the prevention of aGVHD with no difference in disease-free or overall survival(OS). Therefore both regimens are standard prophylaxis for aGVHD (Ratanatharathorn et al., 1998; Nash et al., 2000). Mycophenolate mofetil (MMF) is a prodrug of mycophenolic acid (MPA), which blocks de novo biosynthesis of purine nucleotides via inosine monophosphate dehydrogenase inhibition. De novo synthesized purines are critically needed by T and B cells for their proliferation. Therefore MMF exerts cytostatic effects on lymphocytes (Ransom, 1995). MMF in combination with a CNI is also a widely used prophylaxis regimen for aGVHD (Sabry et al., 2009) and this combination was related to faster hematopoietic engraftment (Bolwell et al., 2004; Neumann et al., 2005) and less risk of mucositis (Bolwell et al., 2004; Ram et al., 2004) as compared with CNI/MTX. However, it remains controversial when comparing the survival and rate

54 of GVHD between CNI/MTX and CNI/MMF. In MA allo-HCTs, some studies showed similar risk of acute/chronic GVHD, as well as comparable survival (Bolwell et al., 2004; Neumann et al., 2005; Gupta et al., 2016) while some others reported a higher risk of grades II-IV aGVHD and worse OS (Hamilton et al., 2018). In RIC condition, conflicting data also exist, as Piñana et al previously showed similar risks of acute/chronic GVHD and no difference in survival (Piñana et al, 2010) while Eapen et al reported higher risk of severe (grades II-IV and grades III-IV) aGVHD, higher non-relapse mortality (NRM) and inferior OS (Eapen et al., 2015) for CNI/MMF. Sirolimus, also known as rapamycin, binds to FKBP12 in a manner similar to tacrolimus. However, unlike tacrolimus-FKBP12 complex which inhibits calcineurin, the sirolimus-FKBP12 complex inhibits the mTOR pathway, and further blocks activation of T and B cells though inhibition of IL-2 and other cytokines (Lietal., 2014). Moreover, sirolimus-based graft-versus-host disease prophylaxis promotes the in vivo expansion of Tregs (Akimova et al., 2012; Peccatori et al., 2015). A phase 3, multicenter, randomized trial comparing Tacrolimus/Sirolimus and Tacrolimus/MTX as prophylaxis regimen for aGVHD showed no significant difference in grades II-IV aGVHD, relapse-free survival, and OS between Tacrolimus/Sirolimus groupand Tacrolimus/MTX group. However, the Tacrolimus/Sirolimus group showed less incidence of oropharyngeal mucositis, making this combination a good alternative for patients prone to have mucositis (Cutler et al., 2014). PTCy was developed by the Johns Hopkins Hospital, based on the pre-clinical observation that it could selectively remove alloreactive T cells without compromising engraftment (Luznik et al., 2001). Moreover, the Treg resistance to Cy may contribute to the clinical activity of PTCy in preventing GVHD (Kanakryetal., 2013). This protocol has allowed successful transplantation of HLA-haploidentical grafts, thus expanding the donor selection (Robinson et al., 2016). Single-agent PTCy has been shown to be an efficacious method of aGVHD prophylaxis in bone marrow transplantation (Kanakry et al., 2014). However, in peripheral blood stemcell transplantation, several studies have reported high rates of severe aGVHDwith

55 single-agent PTCy (Holtick et al., 2016; Bradstock et al., 2015). In addition, regimen combining PTCy and CNI or other immunomodulatory drugs are under clinical trials, and the combination of PTCy+Tacrolimus+MMF is shown to reach a good GVHD-free, relapse-free survival in a prospective multicentre phase 2 trial (Bolaños-Meade et al., 2019), but there is report indicating a possible higher rate of late-onset aGVHD (Shah et al., 2019). Another important reagent for aGVHD prevention is antithymocyte globulin (ATG), which exerts its anti-GVHD effects majorly through in-vivo T cell depletion (Mohty et al., 2007). ATG in combination with standard GVHD prophylaxis (CNI+MTX)is widely used in Europe as a standard aGVHD prophylaxis for allo-HCTs from HLA-identical related or unrelated donors (Mohty et al., 2017). Several phase III randomized trials have definitely shown that ATG reduces the risk of grades II-IV aGVHD and cGVHD without an increase in NRM (Finke et al., 2009; Kröger et al., 2016; Bonifazi et al., 2019), as compared with standard prophylaxis alone. Meanwhile, most studies report no difference in OS between ATG+ and ATG- standard GVHD prophylaxis (Finke et al., 2009; Kumar et al., 2018). Many other strategies for aGVHD prophylaxis including ex vivo T-cell depletion, Treg infusion, IL-6 neutralization, abatacept, statins, vorinostat etc have shown promising results in either pre-clinical mouse models or some single-center studies, however fail or remain to be validated in large multicenter clinical trials (reviewed by Hamilton, 2018a)

1.3.2.2.aGVHD treatment

Glucocorticoids remain the only standard initial treatment of aGVHD, despite response rates of only 40% to 60% (Martin et al., 2012). Patients with aGVHD requiring systemic therapy (grades II-IV) are typically started on 2 mg/kg/d of methylprednisolone or a prednisone equivalent (Pavletic et al., 2012). Studies have shown that initial treatment of grade II aGVHD with as low as prednisone at 1 mg/kg/day does not adversely affected survival as comparedwith2

56 mg/kg/day. Moreover, high-dose such as 2.5 mg/kg/day does not improve the response rate nor prevent the evolution to grade III-IV aGVHD (Van Lint et al., 1998; Mielcarek et al., 2015). Many studies have tried to combine other immunomodulatory drugs such as ATG (Cragg et al., 2000), infliximab (Couriel et al., 2009), etanercept, MMF (Alousi et al., 2009), etc with corticosteroids as first-line treatment for aGVHD but non is shown to be superior to corticosteroids alone. Steroid-refractory aGVHD is defined as progressive disease by day-3 of therapy or lack of improvement after1to2 weeks, depending on severity of symptoms. Steroid-refractory aGVHD or poor-tolerance for high-dose steroid are even more challenging as no effective second-line treatment is identified as standard therapy. Several novel approaches for aGVHD treatment are under investigation (reviewed by Hamilton, 2018a)

1.3.3.Translational value of existing aGVHD mouse models

Experimental models of GVHD are essential for advancing our fundamental understanding of this disorder, and findings generated from them have led to the current gold standards for GVHD prophylaxis and therapy (Zeiser et al., 2016). However, the translational value, or in other word, how predictive the results got from mouse models are for clinical practice, is always limited. Indeed, profound differences between mouse and human, as well as the differences in HCT process (e.g conditioning) may work together and lead to failure of human translation from many pre-clinical observations.

1.3.3.1.aGVHD mouse models based on TBI conditioning

Numerous models of aGVHD have been developed, with most of them (reviewed by Schroeder et al., 2011 and Boieri et al., 2016 and summarized in Table 5) involving the transplantation of T-cell-depleted bone marrow supplemented with varying numbers and phenotypic classes of donor lymphocytes (either splenocytes or lymph node T cells) into lethally irradiated recipients. The bone marrow provides donor stem cells that allow hematopoietic reconstitution after transplant; T-cell depletion is

57 carried out to avoid the effects of T cells contained within the bone marrow containing on GVHD. Either major MHC mismatch, haploidentical and miHA mismatch models are used. TBI remains the dominant conditioning regimen even though some groups have developed models based on chemotherapy conditioning.

Table 5. Overview of mouse models for aGVHD based on TBI (Boieri et al., Front Immunol, 2016) Model MHC haplotype Conditioning MHC mismatch C57BL/6→BALB/c H2b→H2d TBI Complete C3H/HeJ→C57BL/6 H2k→H2d C57BL/6→B10.BR H2b→H2k C57BL/6→B6C3F1 H2b→H2k/b TBI Haploidentical C57BL/6→B6D2F1 H2b→H2b/d C57BL/6→B6AF1 H2b→H2b/a C57BL/6→B6.C-H2bm1 H2b→H2bm1 TBI or none MHC-I C57BL/6→B6.C-H2bm12 H2b→H2bm12 MHC-II B10.D2→DBA2 H2d→H2d TBI miHA B10.D2→BALB/c H2d→H2d B10→BALB.b H2b→H2b C57BL/6→BALB.b H2b→H2b DBA2→B10.D2 H2d→H2d miHA, minor histocompatibility antigen

1.3.3.2.aGVHD mouse models based on chemotherapy conditioning

Given that the majority of HCTs conducted in clinic are based on chemotherapy conditioning (chemo-based), some chemo-based aGVHD mouse models were constructed. Sadeghi et al constructed the first MHC major mismatch [C57BL/6 (H-2 Kb) to BALB/c (H-2Kd)] model based on busulfan/cyclophosphomide (BU-CY) conditioning and allogeneic bone marrow plus splenocytes as graft. Splenocytes were added to increase alloreactivity. In this model, Allogeneic transplantedmice developed lethal aGVHD starting from day +7 with both histological and clinical signs. Donor T cells accumulated in recipient skin and intestine with GVHD progression (Sadeghi et al 2008). He et al developed a similar model with thesame

58 donor-recipient, but was based on BU/Fludarabine. Mice in this model had less severe aGVHD and longer survival as compared with those treated with BU-CY (He et al., 2016). In addition, a chemo-based haploidentical model was also developed, where C57BL/6 (H-2Kb) and CB6F1 (H-2Kd/b) mice were used as donor and recipient, and the conditioning regimen was BU/Fludarabine. The mice transplanted with bone marrow plus splenocytes developed lethal aGVHD and all died within 30 days, with histological features in the target organs (Li et al., 2015). Moreover, Riesner et al developed a MHC matched, miHA mismatched GVHD model [LP/J (H2kb) to C57BL/6 (H2kb)] based on BU-CY. The transplanted mice manifested typical clinical and histological features of aGVHD, and the mice that survived aGVHD developed scleroderma-like aGVHD. Notably, the aGVHD is less severe and the mice survival was longer in this model as compared with MHC major mismatched models introduced above, which is consistent with the fact that miHA mismatchs lead to less risk and severity of aGVHD (Riesner et al., 2016). In spite of conditioning regimen, many factors may affect the translational value of mouse models (reviewed by Stolfi et al., 2016).

1.3.3.3.Differences between human and mouse as HCT recipients

Imbred mice are genetically homozygous, age- and sex-matched, fed a consistent diet, and housed in tightly regulated specific pathogen-free (SPF) conditions, providing the researchers a reproducible platform to investigate a certain scientific question in a univariant way. However, patient populations exhibit enormous genetic and MHC diversity, range in age and health status. Moreover, patients have experienced great amount of immunomodulatory challenges throughout life and especially during the previous intensive treatments for the underlying disease. In humans, HLA typing occurs via high-resolution DNA typing, and most transplants are matched at the allele level with respect to MHC loci, although they usually differ at the level of miHAs that lie outside of the MHC locus. By contrast, most aGVHD mouse models used now are major MHC mismatched (Boieri et al., 2016). Moreover,

59 there exists other major immunological differences between human and mice, which include involution of the thymus in humans but not mice, and expression of MHC II by human but not mouse T cells (Schroeder et al., 2011). Finally, even the same strain of mice from different vendors may can carry genetic polymorphisms outsideofthe MHC locus that can affect the clinical phenotype of aGVHD (Stolfi et al., 2016) Aging is accompanied with immunosenescence, which will definitely affectthe GVHD response (Nikolich-Žugich et al., 2018). Indeed, age has been found tobe associated with GVHD incidence and severity in both mouse (Ordemann et al., 2002) and human (Flowers et al., 2011). Importantly, age is reported in a study to be strongly correlated with the GVHD-free/relapse-free survival (GRFS) of HCT recipients, as adults aged 21+ had 2-fold worse GRFS vs children (Holtan et al., 2015). In addition, aging impairs immune reconstitution post-HCT (Castermans et al., 2011). The prevalence of overweight and obesity has increased substantially in the recent decades. Many studies have investigated the relations between overweight and GVHD, and have identified that obesity is probably related with higher risk of GVHD (both acute and chronic), and higher NRM (Fuji et al., 2009; Gleimer et al., 2015).However, mice used to study GVHD are typically 8- to 14-weeks old (and have a 2-year lifespan) with normal body weight, corresponding to young, healthy teenagers in human. As is shown in chapter 1.3.1.3 that the link between intestinal microbiota and aGVHD is crucial. In mice, changes in the environment, which in turn affect the composition of gut flora, have been shown to affect disease severity (Gorski et al., 2007). There are profound differences between human and mouse in pre-transplant baseline gut microbiota. Humans have an extreme complexity of gut flora due to age, diet or other environmental factors, while mice heve relatively simple, and even over-simplified gut microbiota due to hygienic control. Moreover, another crucial factor affecting the gut microbiota is the conditioning regimen, and the method of conditioningisalways different between mouse and human HCT (introduced in the next section).

60 1.3.3.4.Differences in source of donor cells and conditioning regimen

G-CSF mobilized peripheral blood stem cells are now most commonly used for human HCT, which contains donor immune cells from the circulation (Ferraraetal., 2009). The number, origin and type of circulating immune cells can be different depending on the mobilization regimen used. By contrast, grafts for mouse models are usually isolated from bone marrow, spleen or lymph nodes. Indeed, immune cells derived from different sources might have different trafficking capacities and composition (i.e. different proportions of T-cell subsets, dendritic cells and early progenitors), and therefore exert different influence on the GVHD phenotype. Furthermore, strain-specific disparities in the proportion of lymphocyte subsets (such as CD4+,CD8+ and Treg cells) influence the phenotype between different mouse models of GVHD. Conditioning regimen is another decisive factor for HCT, which will affect the donor engraftment as well as incidence of GVHD. In HCT for human, chemotherapy-based conditioning is used for almost all cases, however in most mouse models the conditioning remains irradiation-based. Profound differences of these two types of conditioning regimens may lead to distinct tissue damages and proinflammatory reactions in mice and humans and therefore influence the GVHD phenotype (Schroeder et al., 2011). In addition, conditioning regimens in the clinicvaryin intensity from high-dose MA regimens, to RIC regimens based on the patient’s age and disease-related factors (Santoro et al., 2019). In mice, however, the vast majority of experiments use MA regimens comprising one or two doses, which can not reflect the complexity of conditionings in clinic. Moreover, mice do not receive pharmacologic GVHD prevention when investigating GVHD, which allows us tosee the uncompromised GVHD effects. Observations derived from these models may not be translated to clinic where patients routinely receive pre-transplant GVHD prophylaxis.

61 2.Results

2.1.Part 1. Immunomodulatory effects of arsenic trioxide on

plasmacytoid dendritic cells and study of mechanism

Systemic sclerosis (SSc) is an immune-mediated rheumatic disease that is characterized by fibrosis of the skin and internal organs and vasculopathy, and owns the highest mortality among all the rheumatic diseases (Denton et al., 2017). Treatments exist for many aspects of the disease but are not curative (Denton, 2014). Immuno-modulation drugs have become the mainstream for treatment howeveritis still crutial to balance the risk and benefits (Laurent et al., 2018). New drugs with good safety profile are needed for SSc.

As2O3 was found to have therapeutic potential in a mouse model of SSc (Kavian et al., 2012b), however lack of knowledge for its therapeutic mechanism has prevented its further translation towards human. IFN-I signature (increased expression and activation of IFN-regulated genes) has long been discovered in the patients with SSc (Brkic et al., 2016). The interaction between pDCs and IFN-I signature in SSc was recently identified. In a SSc mouse modelwith bleomycin-induced fibrosis, depletion of pDCs not only prevented the disease initiation, but ameliorated the established fibrosis (Ah Kioon et al., 2018; Kafaja et al, 2018). Furthermore, in human, abnormally activated pDCs are infiltrated in the target organs such as skin, lung and bronchoalveolar lavage, and secrete IFN-α andCXCL4, which are both hallmarks of SSc (Kafaja et al, 2018; van Bon et al., 2014).

Given the pathogenetic role and therapeutic potential of pDCs and As2O3 on SSc, it is interesting to test the targeting effects of As2O3 on pDCs. The aims of this part of thesis were to (1) observe the effects of different concentrations of As2O3 on the viability and functions of pDCs, and investigate the underlying mechanism. (2) Test

As2O3 on pDCs derived from SSc patients, and compare the result with healthy pDCs.

62 2.1.1.Article 1: Arsenic trioxide induces regulatory functions of plasmacytoid dendritic cells through interferon-alpha inhibition

Article 1

Arsenic trioxide induces regulatory functions of plasmacytoid

dendritic cells through interferon-alpha inhibition

Yishan Ye, Laure Ricard, Nicolas Stocker, Frédéric De Vassoigne, Eolia Brissot， Baptiste Lamarthée, Arsène Mekinian, Mohamad Mohty, Béatrice Gaugler, and Florent Malard

Presented as Oral presentation in the 60th ASH Annual Meeting and Exposition, 2018 Under revision in Journal of Investigative Dermatology

63 Arsenic trioxide induces regulatory functions of plasmacytoid dendritic cells through interferon-alpha inhibition

Short title: As2O3 effect on plasmacytoid dendritic cells

Yishan Ye, M.D.1,2, Laure Ricard, M.D.1, Nicolas Stocker, M.D.1, Frédéric De Vassoigne,3 Eolia Brissot M.D., Ph.D.,1,3 Baptiste Lamarthée Ph.D.,1,Arsène Mekinian M.D., Ph.D.,1,4 Mohamad Mohty M.D., Ph.D.1,3, Béatrice Gaugler, Ph.D.1, Florent Malard M.D., Ph.D.1,3

1 Sorbonne Université, INSERM, Centre de Recherche Saint-Antoine (CRSA),

F-75012 Paris, France

2 Bone Marrow Transplantation Center, The First Affiliated Hospital, School of

Medicine, Zhejiang University, Hangzhou, 310003, China

3 AP-HP, Hôpital Saint-Antoine, Service d’Hématologie Clinique et Thérapie

Cellulaire, F-75012, Paris, France

4AP-HP, Hôpital Saint-Antoine, Service de Médecine Interne et de l'Inflammation-(DHU i2B), F-75012, Paris, France ORCID Yishan Ye https://orcid.org/0000-0003-3706-4959 Laure Ricard https://orcid.org/0000-0003-4416-3285 Nicolas Stocker https://orcid.org/0000-0001-7316-5832 Frédéric De Vassoigne https://orcid.org/0000-0003-4133-4456 Eolia Brissot https://orcid.org/0000-0003-4471-418X Baptiste Lamarthée https://orcid.org/0000-0002-8417-1661 Arsène Mekinian https://orcid.org/0000-0003-2849-3049 Mohamad Mohty https://orcid.org/0000-0002-7264-808X Béatrice Gaugler https://orcid.org/0000-0001-9434-7176 Florent Malard https://orcid.org/0000-0002-3474-0002

64 Abbreviations

APL, acute promyelocytic leukemia; As2O3, arsenic trioxide; CpG ODN, CpG oligodeoxyribonucleotides; CpG-A, CpG ODN 2216; CpG-B, CpG ODN 2006; CpG-P, CpG ODN 21798; FVD, Fixable Viability Dye; IFN-I, type I interferon;IRF7, interferon regulatory factor 7; MFI, mean florescent intensity; NAC, N-acetylcysteine; pDC, plasmacytoid dendritic cell; RFI, relative fluorescence intensity;ROS, reactive oxygen species; SSc, systemic sclerosis;

Correspondence and preprint requests: Florent Malard, M.D., Ph.D.; Service d’Hématologie Clinique et de Thérapie Cellulaire, Hôpital Saint Antoine, APHP, 184 rue du Faubourg Saint-Antoine, 75012, Paris, France. Phone : +33 149282629 ; Fax : +33 149283375 Email : [email protected]

65 Abstract

Arsenic trioxide (As2O3) is recently found to have therapeutic potential in SSc, a life-threatening multi-system fibrosing autoimmune disease with type I interferon (IFN-I) signature. Chronically activated plasmacytoid dendritic cells (pDCs) are responsible for IFN-I secretion and are closely related with fibrosis establishment in

SSc. In this study, we showed that high concentrations of As2O3 induced apoptosis of pDCs via mitochondrial pathway with increased Bax/Bcl-2 ratio, while independent of reactive oxygen species generation. Notably, at clinical relevant concentrations,

As2O3 preferentially inhibited IFN-α secretion as compared to other cytokines such as TNF-α, probably due to potent down-regulation of the total protein and mRNA expression, as well as phosphorylation of the interferon regulatory factor 7 (IRF7). In addition, As2O3 induced a suppressive phenotype, and in combination with cytokine inhibition, it down-regulated pDCs’ capacity to induce CD4+ T cell proliferation, Th1/Th22 polarization, and B cell differentiation towards plasmablasts. Moreover, chronically activated pDCs from SSc patients were not resistant to the selective IFN-α inhibition, and regulatory phenotype induced by As2O3. Collectively, our data suggest pDC as a promising target for As2O3 treatment efficacy in SSc, and more skin-related autoimmune disorders with IFN-I signature.

66 Introduction

Arsenic trioxide (As2O3), an old drug well-known for its rediscovery in the treatment of acute promyelocytic leukemia (APL) (Lo-Coco et al., 2013), was shown to have therapeutic potential in a number of mouse models of skin-related autoimmune disorders including systemic sclerosis (SSc) (Bobe et al., 2006; Kavian et al., 2012), with largely unknown mechanism. Apoptosis induction and Nf-ĸB pathway inhibition are two mechanisms generally known for As2O3 efficacy, but are not disease-specific (Emadi et al., 2010; Yu et al., 2002). Plasmacytoid dendritic cells (pDCs) are a unique subset of dendritic cells specialized in secreting high levels of IFN-I, and are lately identified to play crucial pathogenetic role in SSc (Laurent et al., 2018). Thus, in a SSc mouse model with bleomycin-induced fibrosis, depletion of pDCs not only prevented the disease initiation, but ameliorated the established fibrosis (Ah Kioon et al., 2018; Kafaja et al, 2018). Furthermore, in human, abnormally activated pDCs are infiltrated in the target organs such as skin, lung and bronchoalveolar lavage, and secrete IFN-α andCXCL4, which are both hallmarks of SSc (Kafaja et al, 2018; van Bon et al., 2014). Abnormal T and B cell responses are both key factors in the pathogenesis of SSc (Liu et al., 2016; de Bourcy et al., 2017; Ricard et al., 2019). Upon activation bydifferent signals, pDCs mature and present antigen to CD4+ T cells, leading to Th1, Th2, Th17 or Treg responses (Cella et al., 2000; Ito et al., 2007; Rissoan et al., 1999). On the other hand, IFN-α and IL-6 secreted by pDCs, as well as cell-to-cell contacts mediate the differentiation of B cells into plasmablasts and immunoglobulin-secreting plasma cells (Jego et al., 2003, Ding et al., 2009; Shaw et al., 2010). Overall, pDCs appear to be a promising therapeutic target in SSc treatment.

We conducted experiments to look into how As2O3 affects pDCs from both healthy donors and SSc patients, and to investigate the underlying mechanisms.

67 Results

High concentrations of As2O3 induces pDC apoptosis via mitochondrial pathway with Bax/Bcl-2 ratio increase

In the treatment of APL, patients’ plasma As2O3 reaches the peak level of around 6  M about 3 to 4 hours after administration and quickly decreases to a stable concentration between 0.5M and 3M (Shen et al., 1997). Therefore, concentrations up to 2  M were considered as clinical relevant. We cultured pDCs purified from healthy donors in the presence of different doses of As2O3 for 6h or 24h, respectively.

After 24h, there was a As2O3 dose-dependent decrease of viable pDC percentages, with concentrations equal or over 1  M significantly decreased viability of both non-activated and CpG-A activated pDCs. The CpG-P activated pDCs were more resistant to As2O3, whose viability significantly decreased with equal or over 2 M (Figure 1a).

After 6h of treatment, the viability of pDCs remained over 95% even with 5MAs2O3

(Figure 1b). Given that As2O3 induced apoptosis of multiple tumor and non-malignant cells (Chen et al., 1996; Gupta et al., 2003), we consequently investigatedthe

As2O3-induced apoptosis in pDCs. After 24h of As2O3 treatment, the apoptotic population was not observed (data not shown). After 6h, clinical relevant concentrations of As2O3 did not induce apoptosis of both non-activated and CpG-A activated pDCs (Figure 1c,d). Given that serum concentration of As2O3 could transiently reach 6  M, we subsequently tested 5  MAs2O3, and observed that it induced significantly apoptosis of non-activated and CpG-A activated pDCs after 6h. Meanwhile, there was a significantly lower degree of apoptosis in the latter group, revealing a protective nature of CpG-A activation against apoptosis induction (Figure 1). Bax and Bcl-2 are key members of the Bcl-2 family proteins within the mitochondrial apoptotic pathway, and the Bax to Bcl-2 ratio determines cell survival/death following an apoptotic stimulus (Raisova et al., 2001). We observed that Bax was greatly upregulated, while Bcl-2 was slightly decreased in pDCs treated with 5MofAs2O3

68 for 6h (Figure 1f,g), leading to a significantly increased Bax/Bcl-2 ratio (Figure 1h).

As2O3 is a well-known inducer of oxidative stress, an initiator of apoptosis. However, we observed that although 5MofAs2O3 increased the intracellular reactive oxygen species (ROS) in non-activated pDCs (Figure 1i,j), the anti-oxidant N-acetylcysteine (NAC) pre-treatment, which significantly reduced the intracellular ROS level (Figure S1a, b) without affecting pDCs viability (Figure S1c, d), did not, or very slightly reversed the As2O3-induced pDC apoptosis (Figure S1e).

Clinical-relevant As2O3 inhibits cytokine secretion of pDCs, with special potency on IFN-α

PBMCs were incubated with CpG-A and up to 2 MAs2O3 for 6h, which induced neither death nor apoptosis, and IFN-α/TNF-α secretions were analyzed by intracellular staining on gated living pDCs. Surprisingly, we observed that As2O3 significantly inhibited the production of IFN-α dose-dependently, but not TNF-α (Figure 2a). However, ELISA analysis of supernatants of purified pDCs after 24h culture with non-lethal As2O3 showed that As2O3 inhibited not only IFN-α, TNF-α, but also IL-6 and CXCL10 production (Figure 2b).

As2O3 blocks IFN-α secretion from pDCs via IRF7 inhibition Decrease of TNF-α, IL-6 and CXCL10 secretion from pDCs were probably due to

As2O3 inhibition of the Nf-ĸB pathway (Emadi et al., 2010; Swiecki et al., 2015). However, the reason for the quick and potent inhibition of IFN-α remained tobe elucidated. We focused on IRF7, a crucial and specific regulator of both the induction and maintenance of IFN-α secretion by pDCs (Honda et al., 2006). Six hours incubation with 5MofAs2O3 induced a dramatic decrease of both the percentage of IRF7+ pDCs and the mean florescent intensity (MFI) of IRF7 from both non-activated and CpG-A activated pDCs, gated on living pDCs (Figure 2c-e). Further RT-PCR experiments demonstrated the inhibition on a mRNA level, especially when the pDCs were activated with CpG-A (Figure 2f). Moreover, the MFI

69 of phospho-IRF7 was also significantly decreased when the CpG-A activatedpDCs were treated with 5MofAs2O3 (Figure 2g,h).

As2O3 induces regulatory phenotype of pDCs

It has been shown that As2O3 induced a suppressive phenotype on immature DCs (Macoch et al., 2013). We addressed whether pDC maturation could also be influenced by As2O3 treatment. For this purpose, isolated pDCs were incubated with non-lethal doses of As2O3 for 24h, simultaneously activated with CpG-A, and checked for the expression of maturation markers (Figure 3a). The results showed that

0.5MofAs2O3 significantly decreased the relative fluorescence intensity (RFI) ratio of CD80, CD86, and HLA-DR, as well as the percentages of CD86+ and CCR7+ activated pDCs. We also observed an up-regulated expression of programmedcell death-ligand 1 (PD-L1) (Figure 3b,c).

As2O3 impairs pDCs ’ capacity to induce CD4+ T cell proliferation and Th1/Th22 polarization

We used a pDC/CD4+ T cell co-culture system to investigate As2O3’ effects on pDCs’ capacity to induce T cell proliferation and polarization. Flow cytometry analysis showed that when pDCs were pretreated with 0.5MofAs2O3 for 24h, together with CpG-B activation, they induce significantly lower percentages of proliferating CD4+ T cells after 5 days of co-culture (Figure 4a,b). We then investigated the effect of

As2O3 treatment on the pDC capacity to polarize allogeneic naïve CD4+ T cells. Flow cytometry analysis showed that when naïve CD4+ T were co-cultured with activated pDCs pre-treated with As2O3, there was a significant decrease in the percentages of IFN-γ (Figure 4c,d) and IL-22 (Figure 4e,f) positive proliferating CD4+ T cells after 7 days. However, As2O3 did not alter the percentages of IL-10 or TNF-α positive proliferating CD4+ T cells significantly (Figure S2).

70 As2O3 inhibits pDCs’ capacity to induce plasmablast differentiation of B cells

We used a pDC/B cell co-culture model to investigate As2O3’ effects pDCs’ capacity to induce B cell differentiation towards plasmablasts (Menon et al., 2016). The flow cytometry results showed that pDCs, together with CpG-P, induced CD27hiCD38hi plasmablast differentiation of syngeneic B cells. When pDCs were pretreated with

0.5  MofAs2O3 for 24h, significantly lower percentages of plasmablasts were induced (Figure 5a,b). Given that IFN-α/IL-6 secretion and cell-to-cell contacts mediate pDC-induced B cell differentiation, we subsequently observed that 24h culture of as low as 0.25  MofAs2O3 decreased greatly the IFN-α secretion from CpG-P activated pDCs, with IL-6 inhibition shown at a higher concentration (Figure

5c). Meanwhile, As2O3 significantly decreased CD86+ pDCs, and increased PD-L1 expression on CpG-P activated pDCs (Figure 5d), with the CD80, HLA-DR, and CCR7 expressions not significantly altered (Figure S3).

SSc pDCs are sensitive to As2O3-induced selective IFN-α inhibition, and regulatory phenotype

In order to further investigate As2O3 as a potential therapeutic agent for SSc, we first tested As2O3 on PBMCs from 12 untreated SSc patients. After incubation with CpG-A and clinical relevant doses of As2O3 for 6h, IFN-α production was inhibited dose-dependently, but not TNF-α from pDCs of SSc patients (Figure 6a). We then checked how As2O3 affected the viability and phenotype of pDCs purified from 6 additional SSc patients (for all patients information, see Table S1). After 24h of culture, 0.5  MofAs2O3, which is non-lethal for pDCs from healthy donors, significantly decreased, albeit not hugely, the viability of both non-activated and CpG-A activated SSc pDCs (Figure 6b). An increased Bax/Bcl-2 ratio was also observed when these cells were cultured for 6h with high dose of 5MAs2O3 (Figure

6c-e). For phenotype, 24h culture with 0.5MofAs2O3 decreased significantly both the MFI and the percentages of CD80+ and CD86+ SSc pDCs, while increased both the MFI and the percentage of PD-L1+ SSc pDCs, with CCR7 and HLA-DR expressions unchanged (Figure 6f,g).

71 Discussion

With this study, we concluded that in clinical conditions, As2O3 may induce pDC apoptosis during the first hours of drug administration. Afterwards, clinical relevant concentrations of As2O3 do not alter viability, but induce mostly functional alterations of pDCs. The survival of the resting state pDCs depend predominantly on the mitochondrial Bcl-2 pathway, while the survival of activated pDCs is regulated by several pathways (Lee et al., 2015; Zhan et al., 2016). The pro-apoptotic protein Bax and the anti-apoptotic protein Bcl-2 are important players in the mitochondrial apoptotic pathway (Kale et al., 2018), and the Bax/Bcl-2 ratio determines survival or death following an apoptotic stimulus (Raisova et al., 2001). We found thatAs2O3 induced a significantly increased Bax/Bcl-2 ratio in pDCs. Moreover, antioxidant

NAC did not reverse As2O3-induced pDC apoptosis. Collectively, As2O3 induced pDC apoptosis via the mitochondrial pathway with increased Bax/Bcl-2 ratio, and independent of ROS generation.

We observed that As2O3 inhibited pDC secretion of IFN-α, which may consequently impair the pDCs’ capacity to promote effector CD8+ and Th1 cell responses, to drive B cell activation and plasma cell generation (Reizis, 2019). Meanwhile, the observed inhibition of TNF-α, IL-6 and CXCL10 could reduce capacity of pDCs to upregulate inflammatory reactions and to attract immune cells to sites of infection or inflammation (Swiecki et al., 2015). A ‘cross-regulation’ effect between IFN-I and TNF-α was previously described in pDCs where TNF-α blockade decreased pDC maturation and promoted their ability to produce IFN-I, leading to possible novel autoimmune side-effects (Cantaert et al.,

2010; Conrad et al., 2018). In our study, As2O3 inhibited both IFN-α and TNF-α secretion, as well as maturation of pDCs. These effects were probably due to

‘double-target’ effects of As2O3 on both IRF7 and the Nf-ĸB pathways (Emadi et al., 2010; Swiecki et al., 2015). Moreover, the IRF7 pathway seems to be much more sensitive to As2O3, as compared to the Nf-ĸB pathway. We observed that clinical relevant concentrations of As2O3, which induced neither pDC death nor apoptosis, potently inhibited IFN-α, while leaving TNF-α unchanged at 6h of culture.

72 Meanwhile, IRF7 expression and phosphorylation were potently inhibited (Kim et al., 2014, Reizis, 2019). We speculate that the IRF7 protein may contain a special domain, offering itself high affinity for soluble trivalent arsenic (Shen et al., 2013). Upon TLR7 or TLR9 mediated activation, pDCs mature and express MHC class I (MHCI) and class II (MHCII) molecules and co-stimulatory markers, which operate together to cross-prime CD8+ T cells and present antigen to CD4+ T cells (Swiecki et al., 2015). Mature pDCs also express co-inhibitory molecules such as PD-L1(Diana et al.,2011; Wolfle et al.,2011), and induce regulatory T cell responses (Reizis, 2019).

In this study, As2O3 inhibited expression of co-stimulatory molecules and chemokine receptor, indicating that it impaired pDCs’ trafficking and antigen-presenting capacity.

In line with these observations, we showed that As2O3 treatment significantly impaired activated pDCs to promote CD4+ T cell proliferation and Th1/Th22 polarizations. The preferential inhibition of type-I IFN secreted by pDCsafterAs2O3 treatment probably contributed to the deficiency of Th1 pro-inflammatory response (Swain et al., 2012). Since abnormal T cell proliferation plays an important role in the pathogenesis of SSc (Bobe et al.,2006; Liu et al., 2016), and both Th1 and Th22 immune responses are involved in the development of SSc (Antonelli et al., 2008;

Fard et al., 2016), these observations highlight an important role of As2O3 in modulating T cell responses in SSc. Nevertheless, pDCs regulate B cell growth and differentiation via both cytokine secretion and cell-to-cell contact (Jego et al., 2003; Ding et al., 2009; Shaw et al., 2010). Altered B cell homeostasis characterized by hyperactivity of plasmablasts and autoantibodies production are reported in patients with SSc (Sato et al., 2004). We observed in this study that As2O3 potently impaired the pDC ability to induce B cell differentiation towards plasmablasts, revealing another important role of As2O3 in B cell regulation in SSc.

Regarding effects of As2O3 on other immune subsets, previous studies have shown that T cells’ viability were not significantly affected by clinical relevant concentrations of As2O3 (Gupta et al., 2003), while it induce significant apoptosis in monocytes (Lemarie et al., 2006). More interestingly, clinical relevant dose of As2O3

73 have been reported to repress the monocyte-derived dendritic cells’ capacity to induce Th1 and Th17 responses (Macoch et al., 2013). Overall, both conventional DCs and pDCs probably contribute to As2O3 induced immunomodulation in vivo. Chronically activated pDCs are responsible for most of the IFN-α secretion in SSc patients, and play a critical role during the process of fibrosis (Ah Kioon et al., 2018;

Kim et al., 2008). We observed that similar to healthy pDCs, As2O3 induced preferential inhibition of IFN-α secretion, pro-apoptotic effects, and regulatory phenotype in SSc pDCs , indicating that As2O3 effects on pDCs from SSc patients were uncompromised.

Overall, our observations on immunomodulatory properties of As2O3 offer an important theoretical explanation for the efficacy of As2O3 on SSc, and may pave the way to As2O3 utilization in more skin-related autoimmune diseases with type-I IFN signature.

Materials & Methods

Media and reagents Complete medium was RPMI-1640 supplemented with 1mM sodium pyruvate, 2mM L-glutamine, MEM vitamin solutions, 100U/mL penicillin, 100 g/mL streptomycin and 10% heat-inactivated fetal bovine serum (all from Thermo Fisher Scientific, Villebon-sur-Yvette, France). 1.5M of Class A CpG ODN 2216 (CpG-A), 1.5Mof Class B CpG ODN 2006 (CpG-B), or 1M of Class P CpG ODN 21798 (CpG-P) (all from Miltenyi Biotec, Paris, France) were used. CpG oligodeoxyribonucleotides (ODNs) are TLR9 agonists. CpG-A is a strong inducer of type I IFNs, whereas CpG-B is a potent stimulator of maturation and the production of cytokines and chemokines. CpG-P exhibits properties of both CpG-A and CpG-B. As2O3 (Sigma-Aldrich, Saint-Quentin Fallavier, France) was used at indicated concentrations. pDC isolation and culture

74 PBMCs were obtained from buffy coats by Ficoll density centrifugation, from healthy donors (Etablissement Français du Sang, Paris Saint-Antoine-Crozatier) or untreated SSc patients (Hôpital Saint-Antoine, Paris, France) after informed consent. The study was approved by the local institutional review board and the Comité de Protection des Personnes Ile-de France VII (CPP Ouest-1, reference 2017-A03380-53). pDCs were negatively selected with EasySepTM Human Plasmacytoid DC Enrichment Kit (Stem cell, Grenoble, France). The purity of isolated pDCs, verified by flow cytometry using PE-Vio770-BDCA2 (Miltenyi Biotec), ECD-CD123 (Beckman Coulter, Villepinte, France), was >90%. pDC viability and apoptosis Isolated pDCs were cultured in the presence of 10 ng/mL IL-3 (Miltenyi Biotec), or activated with CpG-A/CpG-P in the presence of the indicated doses of As2O3. Viability and apoptosis were checked with Fixable Viability Dye eFluorTM 506 (FVD) (Thermo Fisher Scientific) and FITC Annexin V (Biolegend, Ozyme, Saint-Quentin-Fallavier, France). FVD- cells were regarded as viable, FVD+ cells as dead, and FVD-Annexin V+ cells as apoptotic. For NAC (Sigma-Aldrich) treatment, pDCs were pre-treated with 1mM of NAC for 1h, washed and placed in culture for6h. For Bcl-2 and Bax staining, isolated pDCs were cultured for 6h in complete culture medium in the presence of 10 ng/mL IL-3, or activated with CpG-A in the presence of

5MAs2O3. Afterwards cells were stained with PE-Bcl-2 (BD Biosciences, Le Pont de Claix, France) and Alexa Fluor® 488-Bax (Biolegend, Ozyme) or the corresponding isotype controls, using the Foxp3/Transcription Factor Staining Buffer Set (Thermo Fisher Scientific)

Reactive oxygen species Isolated pDCs were cultured for 30 min in complete medium in the presence of 10 ng/mL IL-3, with or without activation with CpG-A and 5  MAs2O3. Instead of

As2O3, pDCs were cultured simultaneously with 200 M of tert-butyl hydroperoxide as positive control. For negative control, pDCs were pre-treated with 1mM of

75 antioxidant NAC for 1h. ROS level was detected with CellROX® Green Flow Cytometry Assay Kits (Thermo Fisher Scientific).

Cytokine secretion analysis

PBMCs were incubated with increasing doses of As2O3, activated with CpG-A for 6h, and 1l/mL Golgi plug (BD Biosciences) was added in the last 3 h. The cells were firstly stained with FVD eFluorTM 506, then with PE-Vio770-BDCA2 (Miltenyi Biotec) and ECD-CD123 (Beckman Coulter). Cells were then stained with FITC-IFN-α and APC-Vio770-TNF-α (both from Miltenyi Biotec) with Cytofix/Cytoperm Buffer (BD Biosciences).

Cell signaling staining Isolated pDCs were cultured for 6h in the absence of IL-3, with or without CpG-A

TM and 5MofAs2O3. Afterwards, the cells were fixed with Cytofix Fixation Buffer (BD Biosciences), permeabilized with Phosflow™ Perm Buffer III (BD Biosciences), and then stained with APC-IRF7 and PE-IRF7 pS477/pS479 (both from Miltenyi Biotec).

Gene expression analysis RNA was extracted using RNAeasy Mini kit (QIAGEN, Les Ulis, France). RNA was subjected to reverse transcription (High Capacity RNA-to-cDNA Master Mix, ThermoFisher Scientific) and quantified by real-time quantitative PCR using commercially available primer/probes sets (Assay-On-Demand, ThermoFisher Scientific): GAPDH (Hs99999905_m1), IRF7 (Hs01014809_g1). Real-Time PCR were performed on a 7500 Fast Dx Real-Time PCR Instrument (ThermoFisher Scientific). Relative expression for the mRNA transcripts were calculated using the ∆ ∆Ct method and GAPDH mRNA transcript as reference.

76 Phenotype evaluation After 24h of culture, pDCs were harvested and stained with FVD eFluorTM 506, treated with human Fc block (Miltenyi Biotec) and stained with the following antibodies: AA750-CD80, PC5.5-CD86, FITC-HLA-DR (Beckman Coulter), PE-CCR7 (Thermo Fisher Scientific), APC-PD-L1 (BD Biosciences) or the corresponding isotype controls. The RFI ratios were calculated by normalizing the

RFI of the indicated to the condition of non-activated pDCs without As2O3 treatment.

Mixed lymphocyte reaction (MLR) For the pDC/CD4+ T cell coculture system, pDCs were cultured for 24h with 0.25 to

0.5 μM As2O3 in the presence of 10 ng/mL IL-3, simultaneously activated with CpG-A or CpG-B, and washed twice before co-culture. Allogeneic naïve CD4+ T cells were isolated from PBMC using MagniSortTM Human CD4 Naïve T cell Enrichment Kit (Thermo Fisher Scientific). After isolation, CD4+ T cells were labeled with Cell Proliferation Dye eFluor® 450 (Thermo Fisher Scientific) and co-cultured with pDCs at a 2:1 ratio for 7 days. T cell proliferation was assessed at day-5 of culture by flow cytometry. For intracellular cytokine detection,cellswere harvested at day-7 of culture and stimulated for 5h with 25ng/mL phorbol-12-myristate-13-acetate (PMA) (Sigma-Aldrich) and 1  g/mL ionomycin (Sigma-Aldrich), and 1L/mL Golgi plug. Afterwards, cells were stained with FVD 575V (BD Biosciences), followed by staining with PE/Dazzle 594-CD3 (Biolegend) and PC7-CD4 (BD Biosciences). Finally, cells were stained with PE-IFN-γ, APC-Vio770-TNF-α, eFluor660-IL-22 (Thermo Fisher Scientific) and Vio-515-IL-10 (Miltenyi) with Cytofix/Cytoperm Buffer (BD Biosciences). For the pDC/B cell co-culture system, pDCs were pre-treated for 24h with 0.5μM

As2O3 in the presence of 10 ng/mL IL-3, and washed twice before co-culture. Syngeneic CD19+ B cells were isolated from PBMC using MagniSort® Human CD19 Positive Selection Kit (Thermo Fisher Scientific), and co-cultured with pDCs at a 3:1 ratio, in the presence of 1μM CpG-P for 3 days. Cells were then stained with

77 FITC-CD19, ECD-CD24, PC5.5-CD38 (all from Beckman Coulter), and BV421-CD27 (Biolegend).

ELISA ELISA kits of IFN-α (Thermo Fisher Scientific), TNF-α, IL-6 (PeproTech, Neuilly-sur-Seine, France) and CXCL10 (Biolegend) were used to detect these cytokine/chemokine concentrations in supernatants of pDC cultures.

Flow cytometry Analyses were performed with CytoFLEX Flow Cytometer (Beckman Coulter) and Kaluza Flow Cytometry Analysis Software version 1.5a (Beckman Coulter).

Statistical analysis The Student’s t-test was used for comparison between conditions. All data were analysed using GraphPad Prism 5.0 (GraphPad Software, San Diego, CA). A p<0.05 was considered to be significant.

Conflict of Interest The authors state no conflict of interest.

Acknowledgments The authors acknowledge the Association for Training, Education and Research in Hematology, Immunology and Transplantation for the generous and continuous support to the research work. Y.Y thanks to China Scholarship Council for financial support (CSC No. 201606320257). M.M. thanks Prof. J.V. Melo for critical reading of the manuscript.

78 References Ah Kioon MD, Tripodo C, Fernandez D, Kirou KA, Spiera RF, Crow MK, et al. Plasmacytoid dendritic cells promote systemic sclerosis with a key role for TLR8. Sci Transl Med. Jan 10. 2018;10(423) Antonelli A, Ferri C, Fallahi P, Colaci M, Giuggioli D, Ferrari SM, et al. Th1 and Th2 Chemokine Serum Levels in Systemic Sclerosis in the Presence or Absence of Autoimmune Thyroiditis. J Rheumatol. 2008;35(9):1809–11 Bobe P, Bonardelle D, Benihoud K, Opolon P, Chelbi-Alix MK. Arsenic trioxide: A promising novel therapeutic agent for lymphoproliferative and autoimmune syndromes in MRL/lpr mice. Blood. 2006;108(13):3967–75 Cantaert T, Baeten D, Tak PP, van Baarsen LGM. Type I IFN and TNFα cross-regulation in immune-mediated inflammatory disease: Basic concepts and clinical relevance. Arthritis Res. Ther. 2010;12(5):1–10 Cella M, Facchetti F, Lanzavecchia A, Colonna M. Plasmacytoid dendritic cells activated by influenza virus and CD40L drive a potent TH1 polarization. Nat Immunol. 2000;1(4):305–10 Chen GQ, Zhu J, Shi XG, Ni JH, Zhong HJ, Si GY, et al. In vitro studies on cellular and molecular mechanisms of arsenic trioxide (As2O3) in the treatment of acute promyelocytic leukemia As2O3 induces NB4 cell apoptosis with downregulation of Bcl-2 expression and modulation of PML-RAR alpha/PML proteins. Blood. 1996;88(3):1052–61 Conrad C, Di Domizio J, Mylonas A, Belkhodja C, Demaria O, Navarini AA, et al. TNF blockade induces a dysregulated type i interferon response without autoimmunity in paradoxical psoriasis. Nat. Commun. Springer US; 2018;9(1):1–11 de Bourcy CFA, Dekker CL, Davis MM, Nicolls MR, Quake SR. Dynamics of the human antibody repertoire after B cell depletion in systemic sclerosis. Sci. Immunol. 2017;2(15):1–9 Diana J, Brezar V, Beaudoin L, Dalod M, Mellor A, Tafuri A, et al. Viral infection prevents diabetes by inducing regulatory T cells through NKT cell-plasmacytoid dendritic cell interplay. J Exp Med. 2011;208(4):729–45

79 Ding C, Cai Y, Marroquin J, Ildstad S, Yan J. Plasmacytoid dendritic cells regulate autoreactive B cell activation via soluble factors and in a cell-to-cell contact manner. J Immunol. 2009;183(11):7140–9 Emadi A, Gore SD. Arsenic trioxide - An old drug rediscovered. Blood Rev. 2010;24(4–5):191–9 A Fard NA, Azizi G, Mirshafiey A. The Potential Role of T Helper Cell 22 and IL-22 in Immunopathogenesis of Multiple Sclerosis. Innov Clin Neurosci. 2016;13(7–8):30–6 Guiducci C, Gong M, Xu Z, Gill M, Chaussabel D, Meeker T, et al. TLR recognition of self nucleic acids hampers glucocorticoid activity in lupus. Nature. Nature Publishing Group; 2010;465(7300):937–41 Gupta S, Yel L, Kim D, Kim C, Chiplunkar S, Gollapudi S. Arsenic trioxide induces apoptosis in peripheral blood T lymphocyte subsets by inducing oxidative stress: a role of Bcl-2. Mol Cancer Ther. 2003;2(8):711–9 Honda K, Taniguchi T. IRFs: master regulators of signalling by Toll-like receptors and cytosolic pattern-recognition receptors. Nat Rev Immunol. 2006;6(9):644–58 Ito T, Yang M, Wang YH, Lande R, Gregorio J, Perng OA, et al. Plasmacytoid dendritic cells prime IL-10-producing T regulatory cells by inducible costimulator ligand. J Exp Med. 2007;204(1):105–15 Jego G, Palucka AK, Blanck JP, Chalouni C, Pascual V, Banchereau J. Plasmacytoid Dendritic Cells Induce Plasma Cell Differentiation through Type I Interferon and Interleukin 6. Immunity. 2003;19(2):225–34 Kafaja S, Valera I, Divekar AA, Saggar R, Abtin F, Furst DE, et al. pDCs in lungand skin fibrosis in a bleomycin-induced model and patients with systemic sclerosis. JCI Insight. 2018;3(9) Kavian N, Marut W, Servettaz A, Nicco C, Chereau C, Lemarechal H, et al. Reactive oxygen species-mediated killing of activated fibroblasts by arsenic trioxide ameliorates fibrosis in a murine model of systemic sclerosis. Arthritis Rheum. 2012;64(10):3430–40 Kim D, Peck A, Santer D, Patole P, Schwartz SM, Molitor JA, et al. Induction of interferon-alpha by scleroderma sera containing autoantibodies to topoisomerase I:

80 association of higher interferon-alpha activity with lung fibrosis. Arthritis Rheum. 2008;58(7):2163–73 Kim S, Kaiser V, Beier E, Bechheim M, Guenthner-Biller M, Ablasser A, et al. Self-priming determines high type I IFN production by plasmacytoid dendritic cells. Eur J Immunol. 2014;44(3):807–18 Laurent P, Sisirak V, Lazaro E, Richez C, Duffau P, Blanco P, et al. Innate immunity in systemic sclerosis fibrosis: Recent advances. Front. Immunol. 2018;9(JUL) Lee EF, Fairlie WD, Ottina E, Tarlinton DM, Strasser A, Vikstrom IB, et al. Prosurvival Bcl-2 family members reveal a distinct apoptotic identity between conventional and plasmacytoid dendritic cells. Proc. Natl. Acad. Sci. 2015;112(13):4044–9 Lemarie A, Morzadec C, Merino D, Micheau O, Fardel O, Vernhet L. Arsenic trioxide induces apoptosis of human monocytes during macrophagic differentiation through nuclear factor-kappaB-related survival pathway down-regulation. J Pharmacol Exp Ther. 2006;316(1):304–314b Lepelletier Y, Zollinger R, Ghirelli C, Raynaud F, Hadj-Slimane R, CappuccioA,et al. Toll-like receptor control of glucocorticoid-induced apoptosis in human plasmacytoid predendritic cells (pDCs). Blood. 2010;116(18):3389–97 Liu M, Wu W, Sun X, Yang J, Xu J, Fu W, et al. New insights into CD4+T cell abnormalities in systemic sclerosis. Cytokine Growth Factor Rev. Elsevier Ltd; 2016;28:31–6 Lo-Coco F, Avvisati G, Vignetti M, Thiede C, Orlando SM, Iacobelli S, et al. Retinoic acid and arsenic trioxide for acute promyelocytic leukemia. N Engl J Med. 2013;369(2):111–21 Macoch M, Morzadec C, Fardel O, Vernhet L. Inorganic arsenic impairs differentiation and functions of human dendritic cells. Toxicol Appl Pharmacol. 2013;266(2):204–13 Menon M, Blair PA, Isenberg DA, Mauri C. A Regulatory Feedback between Plasmacytoid Dendritic Cells and Regulatory B Cells Is Aberrant in Systemic Lupus Erythematosus. Immunity. 2016;44(3):683–97

81 Raisova M, Hossini A, Eberle J, Riebeling C, Wieder T, Sturm I, et al. The Bax/Bcl-2 Ratio Determines the Susceptibility of Human Melanoma Cells to CD95/Fas-Mediated Apoptosis. J. Invest. Dermatol. 2001;117(2):333–40 Reizis B. Plasmacytoid Dendritic Cells: Development, Regulation, and Function. Immunity. 2019;50(1):37–50 Ricard L, Jachiet V, Malard F, Ye Y, Stocker N, Riviere S, et al. Circulating follicular helper T cells are increased in systemic sclerosis and promote plasmablast differentiation through the IL-21 pathway which can be inhibited by ruxolitinib. Ann Rheum Dis. 2019. Rissoan MC, Soumelis V, Kadowaki N, Grouard G, Briere F, R. de WM, et al. Reciprocal control of T helper cell and dendritic cell differentiation. Science (80-. ). 1999;283(5405):1183–6 Sato S, Fujimoto M, Hasegawa M, Takehara K. Altered blood B lymphocyte homeostasis in systemic sclerosis: expanded naive B cells and diminished but activated memory B cells. Arthritis Rheum. 2004;50(6):1918–27 Shaw J, Wang YH, Ito T, Arima K, Liu YJ. Plasmacytoid dendritic cells regulate B-cell growth and differentiation via CD70. Blood. 2010;115(15):3051–7 Shen S, Li XF, Cullen WR, Weinfeld M, Le XC. Arsenic binding to proteins. Chem Rev. 2013;113(10):7769–92 Shen ZX, Chen GQ, Ni JH, Li XS, Xiong SM, Qiu QY, et al. Use of Arsenic Trioxide (As2O3) in the Treatment of Acute Promyelocytic Leukemia (APL) II. Clinical Efficacy and Pharmacokinetics in Relapsed Patients. Blood. 1997;89(9):3354–60 Swain SL, McKinstry KK, Strutt TM. Expanding roles for CD4(+) T cells in immunity to viruses. Nat Rev Immunol. 2012;12(2):136–48 Swiecki M, Colonna M. The multifaceted biology of plasmacytoid dendritic cells. Nat Rev Immunol. 2015;15(8):471–85 van Bon L, Affandi AJ, Broen J, Christmann RB, Marijnissen RJ, Stawski L, et al. Proteome-wide analysis and CXCL4 as a biomarker in systemic sclerosis. N Engl J Med. 2014;370(5):433–43

82 Wolfle SJ, Strebovsky J, Bartz H, Sahr A, Arnold C, Kaiser C, et al. PD-L1 expression on tolerogenic APCs is controlled by STAT-3. Eur J Immunol. 2011;41(2):413–24 Yu H-S, Chen G-S, Liao W-T, Chang K-L, Yu C-L. Arsenic Induces Tumor Necrosis Factor α Release and Tumor Necrosis Factor Receptor 1 Signaling in T Helper Cell Apoptosis. J. Invest. Dermatol. 2002;119(4):812–9 Kale J, Osterlund EJ, Andrews DW. BCL-2 family proteins: Changing partnersinthe dance towards death. Cell Death Differ. Nature Publishing Group; 2018;25(1):65–80 Zhan Y, Chow K V., Soo P, Xu Z, Brady JL, Lawlor KE, et al. Plasmacytoid dendritic cells are short-lived: Reappraising the influence of migration, genetic factors and activation on estimation of lifespan. Sci. Rep. Nature Publishing Group; 2016;6(April):1–12

83 Figure Legends

Figure 1 As2O3 induces pDC apoptosis via mitochondrial pathway with Bax/Bcl-2 ratio increase.

Purified pDCs were cultured with indicated doses of As2O3 before tests. % of viable cells after (a) 24h or (b) 6h culture (n=4 independent healthy donors (HD)). (c)

Representative graph of pDC apoptosis after 6h of culture with 1 MAs2O3 (n=4 HD).

% of apoptotic cells cultured with (d) 0to2M (n=4 HD), or (e) 5MAs2O3 (n=4 HD) for 6h. (f) % of Bax+ cells (n=4 HD), MFI of Bax (n=4 HD), (g) MFI of Bcl-2 (n=4

HD), and (h) Bax/Bcl-2 MFI ratio (n=4 HD) after 6h culture with 5  MAs2O3. (i)

Representative graph of intracellular ROS with (black) or without (grey) 5MAs2O3 for 30 min (n=5 HD). (j) MFI of the ROS (n=5 HD). Data are represented as mean± SEM. *p<0.05 ,**p<0.01 and *** p<0.001 by t-test.

Figure 2 As2O3 preferentially blocks IFN-α production from pDCs via IRF7 inhibition. (a) % of IFN-α/TNF-α positive living pDCs after PBMC incubation with indicated doses of As2O3 for6h(n=4HD).(b) Concentrations of indicated cytokines in supernatants of purified pDCs for 24h with As2O3 (n=5 HD). For Figures (c) to (h), purified pDCs were cultured with As2O3 for 6h before tests, gated on living cells. (c)

Representative graph of IRF7, after incubation with (black) or without 5MAs2O3 (grey) (n=4 HD). (d) % of IRF7+ cells (n=4 HD). (e) MFI of IRF7 (n=4 HD). (f) IRF7 mRNA expression (n=4 HD). (g) Representative graph of phospho-IRF7, with (black line, not shaded), or without (grey line) CpG-A activation, or with CpG-A activation and As2O3 (black line, dark grey shaded) (n=4 HD). (h) MFI of phospho-IRF7 (n=4 HD). Data are represented as mean±SEM. *p<0.05 and **p<0.01 by t-test.

84 Figure 3 As2O3 induces regulatory phenotype of pDCs Isolated pDCs were incubated with/without CpG-A activation, with indicated doses of

As2O3 for 24h before flow cytometry analysis. (a) Representative graph showing cells positive for the indicated surface molecule in the presence of 0.5MAs2O3 (black), control (dark grey) and the isotype control (light grey) (n=4 HD). (b) Relative fluorescence intensity (RFI) ratios of indicated surface molecules (n=4 HD). (c) Percentages of positive cells for indicated surface molecule (n=4 HD). Data are represented as mean±SEM. *p<0.05 and **p<0.01 by t-test.

Figure 4 As2O3 impairs pDCs’ capacity to induce CD4+ T cell proliferation and Th1/Th22 polarization. A pDC/CD4+ cell co-culture system was used. T cell proliferation and polarization was detected on day-5 and day-7 of co-culture, respectively. (a) Representative graph of cell proliferation (n=4). (b) CD4+ T cells negative for Cell Proliferation Dye (n=4). (c) Representative graph of IFN-γ+ and TNF-α+ proliferating T cells gated on Cell Proliferation Dye-negative cells (n=4). (d) % of IFN-γ+ proliferating T cells (n=4). (e) Representative graph of IFN-γ+ and IL-22+ proliferating T cells gated on Cell Proliferation Dye-negative cells (n=4). (f) % of IL-22+ proliferating T cells (n=4). Data are represented as mean±SEM. *p<0.05 and **p<0.01 by t-test.

Figure 5 As2O3 reduces pDCs’ ability to induce plasmablast differentiation of B cells. A pDC/B cell co-culture model was used. CD38hiCD27hi plasmablast differentiation was analyzed on day-3 of co-culture. (a) Representative graph of CD38hiCD27hi plasmablasts, gated on CD19+ cells (n=4). (b) %ofCD38hiCD27hi plasmablasts among all gated B cells (n=4). For Figures (c) to (d), isolated pDCs were incubated with/without CpG-P activation, with indicated doses of As2O3 for 24h before analysis. (c) IFN-α and IL-6 concentrations in supernatants of purified pDCs (n=4 HD). (d) RFI ratios and percentages of positive cells for CD86 and PD-L1 (n=4 HD). Data are represented as mean±SEM. *p<0.05 and **p<0.01 by t-test.

85 Figure 6 As2O3-induced pro-apoptotic effects, selective IFN-α inhibition, and regulatory phenotype are not resistant by SSc pDCs (a) % of IFN-α and TNF-α positive living pDCs after PBMCs from SSc patients were incubated with indicated doses of As2O3 for 6h (n=12 SSc patients (SP)). For (b) to

(g), purified SSc pDCs were cultured with indicated doses of As2O3 before tests. (b) % of viable SSc pDCs after 24h (n=5 SP). (c) % of Bax+ cells, and MFI of Bax expression in all pDCs, after 6h (n=3 SP). (d) MFI of Bcl-2 in all pDCs after 6h (n=3 SP). (e) Bax/Bcl-2 MFI ratio after 6h (n=3 SP). (f) RFI ratios and (g) % of positive cells for indicated surface molecules, after 24h (n=5 SP). Data are represented as mean±SEM. *p<0.05 ,**p<0.01 and *** p<0.001 by t-test.

86 Figure 1 ab

** *** 100 100 ** * ** * ) ** 80 80 * 60 60

s s after 24h (% 40 C 40

20 (%) pDCs Viable 20 Viable pD

0 0 0 0.5 1 2 5 0 0.5 1 2 5 0 5 1 2 5 0 1 2 5 0 1 2 5 .25 0. 25 0.5 0.5 0 0. 0.25 CpG-A ------+ + + + + + CpG-A ------+ + + + + + ------CpG-P ------+ + + + + + As2O3 (M) As2O3 (M) c d IL-3 IL-3+CpG-A 100

0.4 0.4 80 ns

60

As2O3 40 ns

0μM 20

0 Apoptotic pDCs after 6h (%) 0 0.5 1 2 0 0.5 1 2 94.5 5.0 93.8 5.6 CpG-A - - - - + + + +

As2O3 (M)

e ** 0.6 0.4 100 *** 80

As2O3 60 * 1μM 40 20

0 91.3 8.1 92.7 6.7 Apoptotic pDCs after 6h (%) 0 5 0 5 CpG-A - - + +

Annexin V As2O3 (M) f g h 80 100000 8000 20 ** * * * ** ** 80000 60 6000 15 60000 40 4000 10 40000 Bax MFI Bax Bcl-2 Bcl-2 MFI

Bax+ (%)cells 20 2000 5 20000 Bax/Bcl-2 MFI ratio Bax/Bcl-2 MFI

0 0 0 0 0 5 0 5 0 5 0 5 0 5 0 5 0 5 0 5 CpG-A - - + + CpG-A - - + + CpG-A - - + + CpG-A - - + +

As2O3 (M) As2O3 (M) As2O3 (M) As2O3 (M)

** i IL-3 IL-3+CpG-A j 120000 ** * 100000 0.07 Green

 80000

60000

40000

20000 MFI of CellROX 0 0 5 0 5 ® CellROX Green CpG-A - - + + As O (M) CTRL As2O3 5 μM 2 3

87 Figure 2 a b * 80000 * 40 3000 0.06 ** 60000 30 2000 40000 20 (pg/mL)  1000 IL-6 (pg/mL) IL-6 IFN- 20000

producing cells (%) cells producing 10 

IFN- 0 0 0 0 0.5 1 2 0 0.5 1 2 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 CpG-A - - - + + + - - - CpG-A - - - - + + + + CpG-A - - - + + + - - - CpG-B ------+ + + CpG-B ------+ + + As2O3 (M) As2O3 (M) As2O3 (M)

* 20000 60 ns 8000 * 6000 15000 40 10000

(pg/mL) 4000  20

producing cells(%) 5000

TNF- 2000 CXCL10 (pg/mL) CXCL10 

TNF- 0 0 0 0 0.5 1 2 0 0.5 1 2 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 CpG-A - - - - + + + + CpG-A - - - + + + - - - CpG-A - - - + + + - - - CpG-B ------+ + + CpG-B ------+ + + As2O3 (M) As2O3 (M) As2O3 (M) c d e pDC pDC+CpG-A 100 25000 ** ** ** 80 20000

60 15000

40 10000 IRF7 MFI

IRF7+ pDCs (%) pDCs IRF7+ 20 5000

0 0 IRF7 0 5 0 5 0 5 0 5 CpG-A - - + + CpG-A - - + +

CTRL As2O3 5 μM As2O3 (M) As2O3 (M) fhg

* 150000 ** 3 *

100000 2

50000 1 Phospho-IRF7 MFI Phospho-IRF7

IRF7 IRF7 relative expression 0 0 0 5 0 5 0 5 0 5 CpG-A - - + + CpG-A - - + + Phospho-IRF7 As2O3 (M) As2O3 (M) CTRL +CpG-A +CpG-A+As2O3

88 Figure 3 a b c IL-3 IL-3+CpG-A * 100 10 80 8 60 6 40 4 CD80 RFI ratio 2 (%) pDCs CD80+ 20

0 0 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 CD80 CpG-A - - - + + + CpG-A - - - + + +

As2O3 (M) As2O3 (M)

10 ** 80 *

8 60 6 40 4

CD86 RFI ratio 20

2 CD86+ pDCs (%)

0 0 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 CD86 CpG-A - - - + + + CpG-A - - - + + +

As2O3 (M) As2O3 (M)

1.5 * 100

1.0

50 0.5 HLA-DR RFI ratio RFI HLA-DR HLA-DR+ pDCs (%)pDCs HLA-DR+

0.0 0 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 HLA-DR CpG-A - - - + + + CpG-A - - - + + +

As2O3 (M) As2O3 (M)

0.05 * 8 60

6 40

4

20 2 CCR7 RFI ratio CCR7 RFI CCR7+ pDCs (%) pDCs CCR7+

0 0 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 CCR7 CpG-A - - - + + + CpG-A - - - + + + As2O3 (M) As2O3 (M)

0.09 15 100 ** * 80

10 60

40 5

PD-L1 RFI PD-L1ratio RFI 20 PD-L1+ pDCs (%) pDCs PD-L1+

0 0 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 PD-L1 CpG-A - - - + + + CpG-A - - - + + + As2O3 (M) As2O3 (M)

ISOTYPE CTRL As2O3 0.5 μM

89 Figure 4 a b

** 50

40

30

20

10 Proliferating T (%)cells 0 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 CpG-A - - - + + + - - - CpG-B ------+ + +

As2O3 (M)

c d

50 **

40

30

20

10 + proliferating T cells (%) γ

IFN 0 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 CpG-A - - - + + + - - - CpG-B ------+ + +

As2O3 (M) e f

* 15

10

5

IL-22+ proliferatingIL-22+ cells(%) T 0 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 CpG-A - - - + + + - - - CpG-B ------+ + +

As2O3 (M)

90 Figure 5 a b 10 **

8

6

4

Plasmablasts (%) 2

0

P G- p B cellsC B+ C+CpG-P D p + B c B+pre-treated pDC+CpG-P

50000 ** 6000 * ** 40000 ** 4800

30000 3600 (pg/mL)

 20000 2400 IL-6 (pg/mL) IFN- 10000 1200

0 0 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 CpG-P - - - - + + + + CpG-P - - - - + + + +

As2O3 (M) As2O3 (M) d

40 15 * * 30 * 10

20

5

CD86 CD86 RFI ratio 10 PD-L1 RFI ratio RFI PD-L1

0 0 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 CpG-P - - - - + + + + CpG-P - - - - + + + +

As2O3 (M) As2O3 (M) ** 100 100 * ** 80 * 80 60 60

40 40

CD86+ pDCs (%) 20 20 PD-L1+ pDCs (%)

0 0 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 CpG-P - - - - + + + + CpG-P - - - - + + + +

As O (M) As2O3 (M) 2 3

91 Figure 6

a b *** ns 80 ** * 60 ** 100 ** 60 80 40 60 40 40 20

(%)producing cells 20 producing cells (%) cells producing  20  TNF- IFN- 0 0 0 0 0.5 1 2 0 0.5 1 2 0 0.5 1 2 0 0.5 1 2 (%) 24h after pDCs SSc Viable 0 0.5 0 0.5 CpG-A - - - - + + + + CpG-A - - - - + + + + CpG-A - - + +

As2O3 (M) As2O3 (M) As2O3 (M)

cde 80 0.07 80000 ns ns * * 6000 15 * *

60 60000 4000 10 40 40000

2000 5 20 20000 Bax/Bcl-2 ratioMFI Bax MFI of SSc pDCs SSc of MFI Bax Bcl-2 MFI of SSc pDCs

Bax positive SSc pDCs(%) 0 0 0 0 0 5 0 5 0 5 0 5 0 5 0 5 0 5 0 5 CpG-A - - + + CpG-A - - + + CpG-A - - + + CpG-A - - + +

As2O3 (M) As2O3 (M) As2O3 (M) As2O3 (M) f

* * ns ns * 60 15 6 8 20

6 15 40 10 4

4 10

20 5 2 2 5 SSc pDC CD80 RFI ratio pDC RFI SSc CD80 ratio pDC RFI SSc CD86 SSc pDC PD-L1 RFI pDC ratio PD-L1 SSc RFI SSc ratioSSc pDC RFI CCR7

0 0 ratioRFI SSc pDC HLA-DR 0 0 0 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 CpG-A - - + + CpG-A - - + + CpG-A - - + + CpG-A - - + + CpG-A - - + +

As2O3 (M) As2O3 (M) As2O3 (M) As2O3 (M) As2O3 (M) g

** ** 100 ns * 100 50 50 60 80 80 40 40 60 40 60 30 30

40 20 40 20 20 20 10 20 10 CD80+ SSc pDCs (%) CD86+ SSc pDCs (%) PD-L1+ SSc pDCs (%) CCR7+ SSc pDCs (%) HLA-DR+ SSc pDCs (%) pDCs SSc HLA-DR+ 0 0 0 0 0 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 0 0.5 CpG-A - - + + CpG-A - - + + CpG-A - - + + CpG-A - - + + CpG-A - - + +

As2O3 (M) As2O3 (M) As2O3 (M) As2O3 (M) As2O3 (M)

92 Supplementary materials

Figure S1 NAC pre-treatment significantly decreases ROS generation but does not restore the apoptosis of pDCs

a b c 100000 * 100

80000 80 Green  60000 60

40000 40

20000 Viable pDCs (%) 20 MFI of CellROX 0 0 CellROX® Green 0 NAC 0 NAC CTRL NAC 1mM

d 100 e 100

80 80 *

60 60

40 40

Viable pDCs (%)Viable 20 20 Apoptotic pDCs Apoptotic (%) 0 0 0 NAC 5 5+NAC 0 NAC 5 5+NAC 0 NAC 5 5+NAC 0 NAC 5 5+NAC CpG-A - - - - + + + + CpG-A - - - - + + + +

As2O3 (M) As2O3 (M)

(a) Representative graph of one from 5 independent experiments for intracellular ROS levels without (grey) or with (black) 1mM NAC pre-treatment for 1h, 1mM NAC were used in Figures (a) to (e). (b) Mean fluorescence intensity (MFI) of the intracellular ROS levels with/without NAC pre-treatment for 1h (n=5 HD). (c) %of viable pDCs with/without NAC pre-treatment for 1h (n=5 HD). (d) %ofviablepDCs with/without NAC pre-treatment for 1h, followed by incubation with/without 5 μM

As2O3 for 6h (n=5 HD). (e) % of apoptotic (Annexin V+ FVD-) cells with/without

NAC pre-treatment for 1h, followed by incubation with/without 5 μM As2O3 for 6h (n=5 HD). Data are represented as mean±SEM. *p<0.05 by t-test.

93 Figure S2 As2O3 does not alter pDC ability to induce IL-10 or TNF-α positive proliferating CD4+ T cells

ab

15 100

80 10 60

40 5

+ proliferating T cells (%) 20  IL-10+ proliferatedIL-10+ cells T (%)

0 TNF- 0 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 0.5 CpG-A - - - + + + - - - CpG-A - - - + + + - - - CpG-B ------+ + + CpG-B ------+ + +

As2O3 (uM) As2O3 (uM)

Isolated pDCs were incubated with non-lethal doses of As2O3 for 24h, simultaneously activated with CpG-A/CpG-B, washed and co-cultured with isolated cell proliferation dye-labeled naïve CD4+ T cells for 7 days. (a) % of IL-10 positive CD4+ T cells (n=4 independent experiments). (b) % of TNF-α positive CD4+ T cells (n=4). Data are represented as mean±SEM.

94 Figure S3 As2O3 effects on the expression of CD80, HLA-DR, and CCR7 on CpG-P activated pDCs

a

30 2.0 15

1.5 20 10

1.0

10 5

CD80 RFI ratio RFI CD80 0.5 CCR7 RFI ratioCCR7 RFI HLA-DR RFI ratio RFI HLA-DR

0 0.0 0 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 CpG-P - - - - + + + + CpG-P - - - - + + + + CpG-P - - - - + + + +

As2O3 (M) As2O3 (M) As2O3 (M) b

100 100 100

80 80 80

60 60 60

40 40 40 CD80+ pDCs (%)CD80+ pDCs 20 20 CCR7+ (%)pDCs 20 HLA-DR+ pDCs (%)pDCs HLA-DR+

0 0 0 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 0 0.25 0.5 1 CpG-P - - - - + + + + CpG-P - - - - + + + + CpG-P - - - - + + + +

As2O3 (M) As2O3 (M) As2O3 (M)

Isolated pDCs were incubated with non-lethal doses of As2O3 for 24h, simultaneously activated with CpG-P, before flow cytometry analysis. (a) RFI ratios of indicated surface molecules (n=4 HD). (b) Percentages of positive cells for indicated surface molecule (n=4 HD). Data are represented as mean±SEM.

Table S1 Disease characteristics of 18 patients with systemic sclerosis.

Limited/extensive cutaneous systemic sclerosis, no (%) 16 (89%) / 2 (11%) Female sex, no (%) 15 (83%) Median age (range), years 67.5 (31-84) Median disease duration (range), years 19 (0-30) Median modified Rodnan skin-thickness score 5 (0-35) Pulmonary arterial hypertension, no (%) 3 (17%) Lung fibrosis, no (%) 6 (33%) abbreviation: no, number.

95 2.2.Part 2. Effects of arsenic trioxide and other immunomodulatory

drugs in a novel mouse models of acute graft-versus-host disease

Acute graft-versus-host disease (aGVHD) is a major inflammatory complication for patients that underwent allogeneic hematopoietic cell transplantation (HCT). Despite decades of effort on prevention and treatment of aGVHD, it remains amajor cause of morbidity and mortality (Ferrara et al., 2009; Holtan et al., 2014). Many aGVHD mouse models have been constructed, however poor translational value remains their limitations. Most allo-HCTs performed now in clinic are based on chemotherapy conditioning, however the majority of aGVHD mouse models are still based on radiation conditioning (Schroeder et al., 2011). Meanwhile, grafts for most aGVHD mouse models are unmanipulated BM cells, in contract with that most centres use G-CSF mobilized graft for human allo-HCTs. Some aGVHD mouse models based on chemotherapy conditioning have been developed. Sadeghi etal constructed the first MHC major mismatch [C57BL/6 (H-2 Kb) to BALB/c (H-2Kd)] model based on busulfan/cyclophosphomide (BU-CY) conditioning (Sadeghietal., 2008). Subsequently, some other chemo-based aGVHD mouse models were developed to investigate the different contexts of haploidentical (Li et al., 2015) and miHA mismatched (Riesner et al., 2016) allo-HCTs. Moreover, an aGVHD model based on TBI and G-CSF mobilized graft was also described (Arbez et al., 2015). However, none of them have combined these two clinical factors in one model. Our group has demonstrated an important role of pDC in the development of aGVHD

(Malard et al., 2013). Given the specific effects of As2O3 on pDCs revealed in the results of Part 1, we sought to conduct in-vivo experiments and test As2O3,aswellas other new drugs in the context of aGVHD in this new model. The aims of this part of the thesis were to: (1) construct a more clinical-relevant aGVHD mouse model based on chemotherapy conditioning and G-CSF mobilized graft and (2) to evaluate As2O3, as well as several other immunomodulatory drugs in this novel clinical relevant mouse model.

96 2.2.1.Results

HCT and clinical manifestations of aGVHD We developed a MHC major mismatched aGVHD model using C57BL/6 (H-2 Kb, female) as donor and BALB/c (H-2Kd, female) as recipient, based on BU-CY conditioning and G-CSF mobilized graft. Dose optimization revealed the suitable concentrations of BU (80mg/kg) and CY (200mg/kg). All mice conditioned with increased BU (100mg/kg) died within 4 days after transplant, with drastic weight loss (data not shown). Protocol for transplantation and G-CSF mobilization is shown in Figure 12.

Figure 12. Chemotherapy and HCT protocol Female BALB/c received IP doses of BU-CY (80mg/kg-200mg/kg) per mouse beginning at day -7 before HCT. For transplantation, 10*10^6 splenic cells from G-CSF mobilized C57BL/6 mice were given by tail vein injection on day 0.

Appropriate conditioned and syngeneic transplanted mice (Syn group) survived over +60 day of transplant, with no signs of aGVHD. Mice who received conditioning but no transplantation (Chemo group) started to die from day +2 and more than 70%of mice died within 15 days. Appropriate conditioned and allogeneic transplanted mice

97 (Allo group) started to die from day +4, with 65% of mice died within 20 days, and all died within 40 days (median survival=12 days) (Figure 13a). Mice in the Allo group exhibited features of aGVHD including hunched posture (Figure 13b), hair loss (Figures 13c, d), ruffled fur (Figure 13d) and most importantly weight loss (Figure 13e). Notably, diarrhea was found significant in mice of Allo group but not in Syn group, which happened simultaneously with weight loss. a b 100 Chemo (n=17) 80 Syn (n=17) Allo (n=17) 60 *** 40 Survival (%)Survival 20

0 0 20 40 60 Days after transplant c d

e f 110 6

100

4 90 Chemo (n=17) *** Syn (n=17) Chemo (n=17) 80 *** Allo (n=17) 2 Syn (n=17) GVHD score 70 Allo (n=17) Weight change(%)

60 0 -5 0 5 10 15 20 25 30 35 40 45 50 55 60 0 5 10 15 20 25 30 35 40 45 50 55 60 Days before and after transplant Days after transplant

Figure 13. GVHD Clinical manifestations Female BALB/c mice were conditioned with BU-CY (80mg/kg-200mg/kg) and transplanted with 10*10^6 G-CSF mobilized splenocytes from female BALB/c(Syn group) or C57BL/6 mice (Allo group). In Chemo group recipient mice did not undergo transplant. (a) Survival after HCT of 17 mice per group from 3 independent experiments, ***p<0.001 by Mantel–Cox log-rank test. Mice in Allo group exhibited features of aGVHD including hunched posture (b), hair loss (c and d), and ruffled fur (d). (e) and (f) showed mean ± SEM of 17 mice per group from 3 independent

98 experiments (e) Weight loss in aGVHD phase. Between d+5 and +25 after HCT, weights of mice from Allo group were significant lower than those in Syn group. (f) GVHD score after HCT. Mice were scored from day +1 for five clinical parameters (weight loss, posture, activity, fur and skin) on a scale from 0 to 2. Clinical GVHD score was generated by summation of these five parameters. ***p<0.001 by t-test.

Weight loss is a good indicator of GVHD in mice (Sadeghi et al., 2008). In all groups, a weight loss was seen between day -7 and day +5, largely due to the conditioning toxicity. In Syn group, recipients regained their weight after day +5 and gradually reached their original weight around day +20. In Allo group, the weight losswenton to reach the nadir on day +7 and remained at a low level, until the second phaseof weight loss on day +30. GVHD score represented the development of aGVHD (Figure 13f).

Engraftment We observed a stable donor chimerism of over 95% in BM, and an increasing donor cell chimerism in spleen, from around 80% in day +7 to full chimerism on day +14 (Figure 14a). For T cells and granulocytes, both showed a gradual engraftment in spleen, together with the development of aGVHD (Figure 14b, c). Moreover, a reversed CD4/CD8 ratio was seen in the bone marrow (Figure 14d) and spleen (Figure 14e) of Allo recipients, from day +7 to day +28, but not in Syn group, indicating an aGVHD process mainly mediated by CD8+ T cells. a b c

BM and spleen chimerism T cell/granulocyte chimerism in BM T cell/granulocyte chimerism in SP 100 100 100 BM T cells T cells 80 SP 80 GNs 80 GNs

60 60 60

40 40 40

20 20 20 Donor chimerism Donor(%) chimerism (%)Donor chimerism (%)Donor chimerism 0 0 0 7 14 21 7 14 21 7 14 21 Days after HSCT Days after HSCT Days after HSCT d e

99 Figure 14. Donor engraftment and lymphoid subpopulations Female BALB/c mice were conditioned with BU-CY and transplanted with 10*10^6 G-CSF mobilized splenocytes from female BALB/c (Syn group) or C57BL/6 mice (Allo group). In Chemo group recipient mice did not undergo transplant. On day+7, +14, +21, and +28 at least 3 mice from each group were sacrificed and BM and spleen cells were collected for FACS analysis. (a) Donor cell engraftment in BM and SP of recipients in Allo group. (b) T cell and granulocytes (GNs) engraftment in BM of recipients in Allo group. (c) T cell and GNs engraftment in SP of recipients in Allo group. (d) CD4/CD8 ratio in BM of recipients in three groups (e) CD4/CD8 ratio in SP of recipients in three groups. All values were mean±SEM for at least 3 animals in each group per time point. **p<0.01 by two sided unpaired t-test.

Effects of As2O3 and some other potential drugs on aGVHD In order to investigate the anti-aGVHD effects of several potential drugs, we firstly tried to use CsA, the well documented drug in preventing human aGVHD. CsA was administrated 0.5mg/mouse/day through IP from day 0 to d +4. Surprisingly,CsA treatment did not prevent aGVHD in this model. Moreover, two out of six mice in the Syn+CsA group died early after transplantation, indicating a probable toxicity of CsA in this model (Figure 15a). Subsequently, we tested As2O3 (0.1mg/mouse/day, IP, day +3 to day +12) in our model, and no superior survival was observed in the

Allo+As2O3 group as compared with Allo group. In addition, one out of six mice in the Syn+As2O3 group died early after transplantation (Figure 15b). Finally, we tested venetoclax, a selective inhibitor of Bcl-2 (Souers et al., 2013) (2mg/mouse/day, gavage, day 0 to day +14) in this model. Unfortunately, venetoclax also did not prevent aGVHD in this model. Nevertheless, all the mice in the Syn+venetoclax group died quickly after transplantation, indicating a combined toxicity of venetoclax and chemothrapy conditioning (Figure 15c).

Gut decontamination delayed aGVHD onset Given the potential combined toxicity of chemotherapy conditioning and the drugs investigated above, we sought to test some ‘preventive’ regimen, instead of potentially cytotoxic drugs in this model. Gut decontamination was found to prevent mouse aGVHD in some seminal studies (van Bekkum et al., 1977), we used an

100 antibiotic cocktail of 1g/L of kanamycin sulfate, amoxicillin, vancomycin hydrochloride and metronidazole (drinking water, day -7 until the end of experiment) to construct a germ-free recipient, and observed that the antibody cocktail delayed the onset of aGVHD, and significantly prolonged the survival of allogeneic-transplanted mice (Figure 15d). Moreover, we tried to use anti-fungal drugs Flucinazole (0.5 mg/ml) or Amphoterecin B (0.1mg/mL), respectively to prevent aGVHD (drinking water, day -7 until the end of experiment), however non of them prevented aGVHD occurrence in this model (Figure not shown). a b

100 100 Chemo (n=6) Chemo (n=6) 80 Allo (n=6) 80 Allo (n=6) Syn (n=6) Syn (n=6) 60 60 Allo+CsA (n=6) Allo+As2O3 (n=6) Syn+CsA (n=6) Syn+As2O3 (n=6) 40 40 Survival (%) Survival (%) 20 20

0 0 0 20 40 60 0 20 40 60 Days after transplant Days after transplant

c d

100 Chemo (n=6) 80 Allo (n=6) Syn (n=6) 60 Allo+antibiotics (n=6) Syn+antibiotics (n=6) 40 Survival (%) 20 Allo+antibiotics vs Allo p value<0.001 0 0 20 40 60 Days after transplant

Figure 15. Effects of different drugs on aGVHD Survival curves of one experiment with 6 mice in each group. Female BALB/c mice were conditioned with optimal dose of BU-CY and transplanted with G-CSF mobilized splenocytes from female BALB/c (Syn group) or C57BL/6 mice (Allo group). In Chemo group recipients did not undergo transplant. (a) In Allo+CsA and Syn+CsA group mice received through IP cyclosporine A (CsA) (0.5mg/mouse/day) from day +3 to day +12. (b) In Allo+As2O3 and Syn+As2O3 group mice received through IP As2O3 (0.1mg/mouse/day) from day +3 to day +12. (c) In Allo+venetoclax and Syn+venetoclax group mice received through gavage venetoclax (2mg/mouse/day) from day 0 to day +14. (d) In Allo+antibiotics and Syn+antibiotics group mice received via drinking water an antibiotics cocktail of 1g/L kanamycin sulfate, amoxicillin, vancomycin hydrochloride and metronidazole from day -7 until the end of experiment. Survival was analyzed using the log-rank test. ***p<0.001

101 2.2.2.Materials and Methods

Mice Female C57BL/6j mice (H2kb) and female BALB/c (H2kd) mice were purchased from Janvier (Genest-St-Isle, France). All mice were used at 11-13 weeks of age.Mice were maintained under pathogen free controlled conditions and 12hr light/darkness. Animals had access to food and water ad libitum. All protocols were performed according to approval of the “Service Vétérinaires de la Santé et de la Protection Animal” delivered by the Ministry of Agriculture of France.

HCT procedures BU (Pierre Fabre Médicament, Idron, UK) and CY (Baxter, Saint-Quentin-en-Yvelines, France) were used for conditioning. BU was diluted with PBS to reach 4mg/ml, and CY (20mg/ml) was injected without dilution. Female BALB/c mice received IP doses of BU (20mg/kg/d) for 4 days (day -7 to day -4), followed by CY (100mg/kg/d) for 2 days (day -3 to day -2). After 1 day rest, the recipient mice were injected IV with 10*10^6 splenic cells from female C57BL/6 mice by tail vein injection. PBS was administrated as a vehicle control. Forgraft preparation, donors were previously treated by subcutaneous injection of10μgper animal of recombinant human G-CSF (Hospira, Hurley, UK) once daily from day-5 to day -1. On the day of transplant, the donors were killed by cervical dislocation. Donor spleens were disrupted in RPMI and erythrocytes were then lysed with RBC Lysis Buffer (Multi-species) (Thermo Fisher Scientific, Villebon-sur-Yvette, France), washed twice with RPMI before passing through a 70 mm strainer. Cell were then respended in PBS for injection. Cell viability (>90%) was confirmed with trypan blue.

Grouping and Treatments Each experiment included 5 to 6 mice each group. Mice from Chemo group received conditioning only. Mice from Syn group received syngenic graft and Allo group received allogeneic graft. Mice from Syn+treatment group or Allo+treatment group

102 received syngenic graft/allogeneic graft plus treatment. Cyclosporine A(CsA) (5mg/mL) (Novartis Pharma, Rueil-Malmaison, France) was diluted in PBS and injected IP as 0.5mg/mouse/day from day 0 to day +4. Arsenic trioxide (1mg/mL) (Medsenic, Strasbourg, France) was diluted in PBS and injected IP as 0.1mg/mouse/day from day +3 to day +12, and control mice were injected IP with the same volume of PBS. Venetoclax (ABT-199) (10mg/mL) (AbbVie, Rungis, France) was dissolved in 5% DMSO+50% PEG 300+5% Tween 80+ddH2O, and was given through gavage as 2mg/mouse/day from day 0 to day +14. Control mice received through gavage the same volume of diluent. Gavage with same volume of diluent did not show any toxicity in BALB/c mice after 14 days in a pre-experiment (data not shown). The Antibiotic cocktail containing 1g/L of kanamycin sulfate (VWR, Fontenay-sous-Bois, France), 1g/L of amoxicillin (Panpahrma, Boulogne-Billancourt, France), 1g/L of vancomycin hydrochloride (Mylan, Saint-Priest, France) and 1g/L of metronidazole (B.Braun, Saint-Cloud, France) or the anti-fungal drugs 0.5 mg/mL of flucinazole (Fresenius Kabi, Sèvres, France) or 0.1mg/mL of AmphoterecinB (Bristol-Myers Squibb, Rueil-Malmaison Cedex, France) were dissolved in drinking water and supplied from day -7 till the end of experiment. Control mice has access to water ad libitum.

Assessment of GVHD Mice were monitored for survival and individually scored every two days forfive clinical parameters (posture, activity, fur, skin and weight loss) on a scale from 0 to 2 (Table 6). Clinical GVHD score was assessed by summation of these parameters. Animals with severe aGVHD (scores>6) were killed according to ethical guidelines. For histological analysis, the skin,colon and liver were removed immediately after sacrifice on day +7, +14 and +21, and then fixed in 4% formalin.

103 Table 6. Mouse GVHD scoring (modified from Cooke et al., Blood, 1996)

FACS analysis On day +7, +14 and +21, splenocytes and BM cells were collected, and erythrocytes were lysed with RBC Lysis Buffer. Cells were then passed through a 40μm cell strainer, washed twice and stained for 20 min at 4 °C in PBS/0.5mM EDTA/0.5% BSA with the following mouse mAbs: FITC-CD8a (SONY, Weybridge, UK), PE-CD4 (SONY), PerCPCy5.5-CD11c (SONY), PC7-H2Db (Biolegend, Ozyme, Saint-Quentin-Fallavier, France), A647-H2Dd (Biolegend), APCCY7-CD3 (SONY) and PACIFIC BLUE-CD45 (Biolegend). Analyses were performed with CytoFLEX Flow Cytometer (Beckman Coulter) and Kaluza Flow Cytometry Analysis Software version 1.5a (Beckman Coulter).

Statistical analysis Survival data were analyzed using the Kaplan-Meier method and Mantel-Cox log-rank test. For all other data, the two-sided unpaired t-test was used. Normality tests and F test confirmed Gaussian distribution and equality of variance between different groups. Values were presented as mean ± SEM. A value of P < 0.05 was considered statistically significant in all experiments. Data was computed using GraphPad Prism 5.0 (GraphPad Software).

104 3.Discussions

3.1.As2O3, a promising drug for diseases with IFN-I signature

In Part 1 of the thesis, we comprehensively investigated the effects of As2O3 on pDCs viability and functions. We found that high doses of As2O3 induced pDC apoptosis induction through mitochondiral pathway. Clinical relevant doses of As2O3 preferentially inhibited IFN-α secretion and induced regulatory phenotype of both healthy and SSc pDCs. The selective IFN-α inhibition was probably due to potent down-regulation of IRF-7. Moreover, As2O3 inhibited the maturation of pDCs, and impaired their capacity to induce CD4+ T cell proliferation, Th1/Th22 polarization, and B cell differentiation towards plasmablasts. Moreover, pDCs from SSc patients were not resistant to the selective IFN-α inhibition, and regulatory phenotype induced by As2O3.

In the treatment of APL, patients’ plasma As2O3 reaches the peak level of around 6  M and quickly decreases afterwards, with the mean half-life of 0.89 hours. The stable drug concentrations were between 0.5  M and 3  M (Shen et al., 1997). By comparing the phamacokinetic curve and our results, we concluded that in clinical conditions, As2O3 may induce pDC apoptosis during the first hours, and afterwards

As2O3 does not alter viability, but induce mostly functional alterations of pDCs. We found in both healthy pDCs and pDCs from SSc patients that Bcl-2 expression was not largely altered, but Bax was significantly increased by As2O3. Meanwhile,

ROS generation is not responsible for As2O3-indcued pDC apoptosis. Given that the interactions between the Bcl-2 family proteins under pro-apoptotic stimuli trigger apoptosis by forming pores within the mitochondrial outer membrane, which is due to Bax/Bak homo-oligomerization and pore formation (Kale et al., 2018). Therefore,

As2O3 induces pDC apoptosis via the mitochondrial pathway, however not through modulation of the Bcl-2 protein, nor through an upstream ROS generation. Moreover, In line with the discovery that pDCs rely heavily on Bcl-2 for survival, it was proved in vivo that Bcl-2 inhibitor selectively killed pDCs from lupus-prone

105 mouse (Zhan et al., 2015). We could suspect that As2O3 and commercially available Bcl-2 inhibitors could have synergic effects on pDCs apoptosis by acting on different players of the Bcl-2 apoptosis pathway. A ‘cross-regulation’ effects between IFN-I and TNF-α was previous described in pDCs (Kadowaki et al., 2002; Bennett et al., 2005), where TNF-α blockade decreases pDC maturation and extends their ability to produce IFN-I, which constitutes an obstacle for TNF-α blockade treatment in autoimmune diseases because IFN-I elevation could induce novel autoimmune side-effects (Cantaert et al., 2010; Conrad et al., 2018). In our study both IFN-α and TNF-α secretion were reduced by As2O3, which was probably due to an ‘double-target’ effect, where the IFN-α inhibition was caused by down-regulation of the IRF7 pathway, and the inhibition of pDC maturation and decrease of TNF-α production were probably due to inhibition of the

Nf- κ B pathway, a well-known target of As2O3 (Emadi et al., 2010). Moreover, we observed that clinical relevant concentrations of As2O3, which induced neither pDC death nor apoptosis, potently inhibited IFN-α, while leaving TNF-α unchanged at 6h of culture, indicating an especially high sensitivity of IRF7 pathway to As2O3. Activated pDCs express MHC molecules and co-stimulatory markers which work together to cross-prime CD8+ T cells and present antigen to CD4+ T cells (Swiecki et al., 2015). Mature pDCs also express higher levels of co-inhibitory molecules such as PD-L1 (Diana et al.,2011; Wolfle et al.,2011). In addition, expression of chemokine receptors facilitates pDCs to migrate from the blood to peripheral tissues. Inhibition of As2O3 on expression of these molecules indicate that As2O3 impairs mature pDCs’ antigen presenting and trafficking capacity.

In line with these observations above, we discovered that As2O3 treatment significantly impaired activated pDCs to promote CD4+ T cell proliferation. Since abnormal T cell proliferation plays an important role in the pathogenesis of SSc, this result highlights the therapeutic potential of As2O3 by revealing its effects on T cell proliferation through modulation of pDC functions (Bobe et al.,2006; Liu et al., 2016).

In addition, we reported that As2O3 impairs pDCs’ ability to drive Th1/Th22 polarization of CD4+ T cells. The decrease of type-I IFN secretion by pDCs after

106 As2O3 treatment probably contributes to the deficiency of Th1 pro-inflammatory response (Swain et al., 2012). Moreover, both Th1 and Th22 immune response were reported to be involved in development of SSc (Antonelli et al., 2008; Fard et al.,

2016), indicating that As2O3 may exert therapeutic effects on SSc through T cell response regulations. pDCs have the ability to regulate B cell growth and differentiation via both cytokine secretion and cell-to-cell contact (Jego et al., 2003; Shaw et al., 2010). Altered B cell homeostasis characterized by hyperactivity of plasmablasts and memory B cells and autoantibodies production are reported in patients with SSc (Sato et al., 2004). We observed in this study that As2O3 potently impaired the pDC ability to induce B cell differentiation towards plasmablasts, revealing another important roleofAs2O3 in B cell regulation in SSc. Chronically activated pDCs are responsible for most of the IFN-α secretion in SSc patients, and plays a critical role during the process of fibrosis (Ah Kioon et al., 2018; Kim et al., 2008). We observed a potent and preferential inhibition of IFN-αsecretion from the pDCs of untreated SSc patients, indicating that As2O3 is able to control IFN-α secretion from these dysfunctional pDCs. Moreover, we initially hypothesized a possible resistence of chronically activated SSc pDCs towards As2O3-induced apoptosis, which was discovered in CpG-A activated pDCs and that pDCs in SLE patients are reported to be resistant to glucocorticoid due to hyperactivation (Guiducci et al., 2010; Lepelletier et al., 2010). However, resistence was not discovered, indicating a ‘low level’ of activation for SSc pDCs. Furthermore, the regulatory induction were also similarly discovered in SSc pDCs. Collectively, theseresults indicate that As2O3 effects on pDCs from SSc patients are uncompromised.

Overall, our observations on immunomodulatory properties of As2O3 offer an important theoretical explanation for the efficacy of As2O3 on SSc, and may pave the waytoAs2O3 utilization in more autoimmune diseases with type-I IFN signature.

107 3.2.A novel aGVHD model with chemotherapy-based conditioning

and G-CSF mobilized graft: advances and limitations

In Part 2 of this thesis, we introduce a brand new mouse model of aGVHD based on both chemotherapy conditioning and G-CSF mobilized graft. Briefly, it wasamajor MHC mismatched model [C57BL/6 (H-2 Kb) to BALB/c (H-2Kd)] based on BU-CY conditioning and G-CSF mobilized graft. This model showed good donor engraftment and typical clinical signs of aGVHD, which may be advantageous in reflecting the clinical situation of HCT and aGVHD. So-far the great majority of aGVHD mouse models are based on conditioning with TBI (reviewed by Schroeder et al., 2011). However, the use of lethal TBI without chemotherapy in mouse models contrasts the standard procedure in clinical allo-HCT, and profound differences of these two types of conditioning regimens may lead to distinct GVHD phenotype (Stolfi et al., 2016). Another crucial factor affecting aGVHD incidence and outcome is the source of donor cells. G-CSF mobilized peripheral blood stem cells are now most commonly used for human HCT (Ferrara et al., 2009). By contrast, grafts for mouse models are usually unmanipulated bone marrow. Mouse models based on chemotherapy conditioning (Sadeghi et al.,2008; Li et al., 2015; He et al., 2016; Riesner et al., 2016) or G-CSF mobilized graft (Arbez et al., 2015) have been previously described, but none of them have combined these two factors in one model. Sadeghi et al established conditioning with busulfan(BU (80mg/kg)) and cyclophosphamide (CY (200mg/kg)) in a MHC mismatched model [C57BL/6 (H2kb) to BALB/c (H2kd)] using bone marrow graft. We observed in our model the same optimized dose of conditioning. Therefore, the requirementfor successful engraftment of bone marrow graft and G-CSF mobilized graft is comparable in this MHC mismatched model. Weight loss was the most well documented manifestation in GVHD mouse model.In the model of Sadeghi et al, a bi-phasic pattern was observed. We observed a similar

108 model of weight loss. In both Allo and Syn group, the weight loss was seen between day -7 and day +5, with the Allo group went on to reach the nadir on day +7 and remained at a low level until the second phase of weight loss on day +30. In theSyn group, mice started to regain their weight from day +5, and gradually returned to the original weight at day +20. However, it remained a question whether the first phase of weight loss in the Allo group was solely due to chemotherapy toxicity or there was a involvement of aGVHD. Indeed, at day +7, we observed more than 90% donor engraftment in the bone marrow and around 60% donor chimerism in the spleen (target organ). At day +21, there was total (~100%) donor engraftment in both bone marrow and spleen. T cell engraftment was slower in the spleen as compared with granulocytes, which is consistent with kinetic of immune reconstitution after Allo-HCT in both human and mouse (Ogonek et al., 2016; Sadeghi et al.,2008). Moreover, reversed CD4/CD8 ratio was observed in the Allo group bone marrowand spleen from day +7 up until day +28. Collectively, aGVHD in this model initiated before day +7, and it induced the first phase of weight-loss together with conditioning toxicity. The second phase of weight loss was solely due to aGVHD. Apart from weight loss, posture change, loss of activity, fur ruffling, and skin damage were all observed in our model. Severe diarrhea was observed in the Allo group, representing GI aGVHD, though histology was not yet done in this model. In our study, the germ-free environment constructed by oral administration of the antibiotic cocktail significantly delayed the onset of aGVHD. This is consistent with the pioneering study showing that germ-free mice developed less severe aGVHD following allo-HCT (van Bekkum et al., 1977). However, the concept that gut decontamination prevents aGVHD following allo-HCT in human is controversial given that clinical trials have failed to demonstrate consistent benefits (Whangbo et al., 2017). Our hypothesis is that in human, it is very difficult to achieve complete gut contamination, and that partial decontamination may lead to dysbiosis and therefore impaired outcome.For As2O3, the only report by Kavian et al showed that it prevented cGVHD in a mouse model. However the major effects of As2O3 was amelioration of systemic sclerosis, and survival was not evaluated (Kavian et al., 2012b). In our

109 mouse model the aGVHD manifested most evidently as weight loss due to severe diarrhea, which was the major cause of death. As2O3 did not show aGVHD preventive effect in our model, showing that As2O3 may not be preventive for GI aGVHD.

Moreover, one out of six mice in the Syn+As2O3 group died at day +5 of transplantation, indicating a potential combined toxicity of As2O3 and the conditioning regimen. In contract to intensive investigation of using of CsA for aGVHD prevention in clinic, the reports of CsA in mouse aGVHD prophylaxis were limited (Satake et al., 2014). In addition, it was also shown that high dose CsA had toxicity on mice (Boland et al., 1984). In our model, two out of six mice in the Syn+CsA group died earlyafter transplantation, alerting the combined toxicity of pre-existing conditioning and CsA. We also tested venetoclax in this model, and discovered no effect on aGVHD. There is study indicating that activated T cells are more resistant to Venetoclax as compared with naïve T cells, indicating that venetoclax may not be able to control the efferent phase of aGVHD (Mathew et al., 2018). Furthermore, the combined toxicity of conditioning and venetoclax is obvious in this study. Despite the advantages, limitations of this model are also obvious, which is due to the shortcomings of chemotherapy conditioning in the context of mouse HCT. Firstly, the one-week long conditioning is more complicated to practice as compared with the ‘single shot’ TBI. Secondly, inevitable differences of the drug concentrations, drug temperature, time of injection between each days due to human and system error will affect the homogeneity of conditioning and subsequently reduce experiment reproductivity. Thirdly, given that chemotherapy conditioning induces obvious weight loss during the first 5 days post-transplantation, it might be tricky to usedrugwhich has potential cytotoxicity early after transplantation, as observed in our study. Nevertheless, this is a MHC major mismatched model, which does not mimic the clinical conditions where major MHC matched/minor MHC mismatched and haploidentical HCTs are more routinely practiced. Therefore, in our lab weplanto construct a haploidentical model [C57BL/6 (H2kb) to CB6F1 (H2kb/d)] and a MHC minor mismatched model [129S2/SvPasCrl (H2kb) to C57BL/6 (H2kb)] based on chemotherapy conditioning and G-CSF mobilized graft in the future.

110 4.Conclusion

In Part 1 of the thesis, we showed that clinical relevant concentrations of As2O3 preferentially inhibited IFN-α secretion as compared to other cytokines such as TNF-α, probably due to potent down-regulation of the IRF7 pathway. In addition,

As2O3 induced a suppressive phenotype, and in combination with cytokine inhibition, it down-regulated pDCs’ capacity to induce T and B cell responses. Given that aGVHD is an inflammatory disorder driven by allo-antigen presentation by DCs including pDCs, and the abnormal T cell activation, we sought to test As2O3 in this context. For this purpose, in Part 2 of thesis, a novel aGVHD mouse model based on chemotherapy-based conditioning and G-CSF mobilized graft was constructed. This model showed good donor engraftment and typical clinical signs of aGVHD.

Furthermore, even though As2O3 did not show aGVHD preventive effects, gut decontamination delayed the onset of aGVHD in this model. Overall, our results indicate As2O3 as a promising immunomodulatory drug, especially in autoimmune diseases with IFN-I signature.

111 5.References

Ablain J, Rice K, Soilihi H, de Reynies A, Minucci S, de The H. Activation of a promyelocytic leukemia-tumor protein 53 axis underlies acute promyelocytic leukemia cure. Nat Med. 2014;20(2):167–74

Ah Kioon MD, Tripodo C, Fernandez D, Kirou KA, Spiera RF, Crow MK, et al. Plasmacytoid dendritic cells promote systemic sclerosis with a key role for TLR8. Sci Transl Med. Jan 10. 2018;10(423)

Ahn H, Kim J, Kang SG, Yoon SI, Ko HJ, Kim PH, et al. Mercury and arsenic attenuate canonical and non-canonical NLRP3 inflammasome activation. Sci Rep. 2018;8(1):13659

Akimova T, Kamath B, Goebel J, Meyers K, Rand E, Hawkins A, et al. Differing Effects of Rapamycin or Calcineurin Inhibitor on T-Regulatory Cells in Pediatric Liver and Kidney Transplant Recipients. Am. J. Transplant. 2012;12(12):3449–61

Albanesi C, Scarponi C, Pallotta S, Daniele R, Bosisio D, Madonna S, et al. Chemerin expression marks early psoriatic skin lesions and correlates with plasmacytoid dendritic cell recruitment. J. Exp. Med. 2009;206(1):249–58

Alex AA, Ganesan S, Palani HK, Balasundaram N, David S, Lakshmi KM, et al. Arsenic Trioxide Enhances the NK Cell Cytotoxicity Against Acute Promyelocytic Leukemia While Simultaneously Inhibiting Its Bio-Genesis. Front Immunol. 2018;9:1357

Allen JS, Pang K, Skowera A, Ellis R, Rackham C, Lozanoska-Ochser B, et al. Plasmacytoid dendritic cells are proportionally expanded at diagnosis of type 1 diabetes and enhance islet autoantigen presentation to T-cells through immune complex capture. Diabetes. 2009;58(1):138–45

Alousi A, Weisdorf D, Logan B, Bolaños-Meade J, Carter S, Difronzo N, et al. Etanercept, mycophenolate, denileukin, or pentostatin plus corticosteroids for acute graft-versus-host disease: a randomized phase 2 trial from the Blood and Marrow Transplant Clinical Trials Network. Blood. 2009;114(3):511–7

Antoine F, Ennaciri J, Girard D. Syk is a novel target of arsenic trioxide (ATO) and is involved in the toxic effect of ATO in human neutrophils. Toxicol Vitr. 2010;24(3):936–41

112 Antonelli A, Ferri C, Fallahi P, Colaci M, Giuggioli D, Ferrari SM, et al. Th1 and Th2 Chemokine Serum Levels in Systemic Sclerosis in the Presence or Absence of Autoimmune Thyroiditis. J Rheumatol. 2008;35(9):1809–11

Apostolova P, Zeiser R. The role of danger signals and ectonucleotidases in acute graft-versus-host disease. Hum. Immunol.2016;77(11):1037–47

Arbez J, Saas P, Lamarthée B, Malard F, Couturier M, Mohty M, et al. Impact of donor hematopoietic cells mobilized with G-CSF and plerixafor on murine acute graft-versus-host-disease. Cytotherapy. 2015;17(7):948–55

Arpinati M, Green CL, Heimfeld S, Heuser JE, Anasetti C. Granulocyte-colony stimulating factor mobilizes T helper 2-inducing dendritic cells. Blood. 2000;95(8):2484–90

Arpinati M, Chirumbolo G, Urbini B, Bonifazi F, Bandini G, Saunthararajah Y, et al. Acute graft-versus-host disease and steroid treatment impair CD11c+ and CD123+ dendritic cell reconstitution after allogeneic peripheral blood stem cell transplantation. Biol. Blood Marrow Transplant. 2004;10(2):106–15

Asselin-Paturel C, Boonstra A, Dalod M, Durand I, Yessaad N, Dezutter-Dambuyant C, et al. Mouse type I IFN-producing cells are immature APCs with plasmacytoid morphology. Nat Immunol. 2001;2(12):1144–50

Asselin-Paturel C, Brizard G, Chemin K, Boonstra A, O ’ Garra A, Vicari A, et al. Type I interferon dependence of plasmacytoid dendritic cell activation and migration. J. Exp. Med. 2005;201(7):1157–67

Asselin-Paturel C, Brizard G, Pin J, Brière F, Trinchieri G. Mouse Strain Differences in Plasmacytoid Dendritic Cell Frequency and Function Revealed by a Novel Monoclonal Antibody. J. Immunol. 2003;171(12):6466–77

Auffray C, Fogg DK, Narni-Mancinelli E, Senechal B, Trouillet C, Saederup N, et al. CX 3 CR1 + CD115 + CD135 + common macrophage/DC precursors and the role of CX 3 CR1 in their response to inflammation. J. Exp. Med. 2009;206(3):595–606

Baechler E, Batliwalla F, Karypis G, Gaffney P, Ortmann W, Espe K, et al. Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupus. Proc. Natl. Acad. Sci. 2003;100(5):2610–5

Bahari A, Salmani V. Environmentally relevant dose of arsenic interferes in functions of human monocytes derived dendritic cells. Toxicol Lett. 2017;275:118–22

113 Banerjee N, Banerjee S, Sen R, Bandyopadhyay A, Sarma N, Majumder P, et al. Chronic arsenic exposure impairs macrophage functions in the exposed individuals. J Clin Immunol. 2009;29(5):582–94

Banovic T, Markey KA, Kuns RD, Stuart D, Raffelt NC, Don AL, et al. Graft-versus-host disease prevents the maturation of plasmacytoid dendritic cells. J Immunol. 2009;182(2):912–20

Barchet W, Cella M, Odermatt B, Asselin-Paturel C, Colonna M, Kalinke U. Virus-induced Interferon α Production by a Dendritic Cell Subset in the Absence of Feedback Signaling In Vivo. J. Exp. Med. 2002;195(4):507–16

Barrat F, Meeker T, Gregorio J, Chan J, Uematsu S, Akira S, et al. Nucleic acids of mammalian origin can act as endogenous ligands for Toll-like receptors andmay promote systemic lupus erythematosus. J. Exp. Med. 2005;202(8):1131–9

Båve U, Magnusson M, Eloranta M, Perers A, Alm G, Rönnblom L. Fc RIIa Is Expressed on Natural IFN- -Producing Cells (Plasmacytoid Dendritic Cells) and Is Required for the IFN- Production Induced by Apoptotic Cells Combined with Lupus IgG. J. Immunol. 2003;171(6):3296–302

Bengtsson A, Sturfelt G, Truedsson L, Blomberg J, Alm G, Vallin H, et al. Activation of type I interferon system in systemic lupus erythematosus correlates with disease activity but not with antiretroviral antibodies. Lupus. 2000;9(9):664–71

Bennett L, Blanck J-P, Pascual V, Banchereau J, Palucka AK. Cross-regulation of TNF and IFN- in autoimmune diseases. Proc. Natl. Acad. Sci. 2005;102(9):3372–7

Bhattacharjee P, Das N, Chatterjee D, Banerjee A, Das JK, Basu S, et al. Association of NALP2 polymorphism with arsenic induced skin lesions and other health effects. Mutat Res. 2013;755(1):1–5

Binet F, Cavalli H, Moisan E, Girard D. Arsenic trioxide (AT) is a novel human neutrophil pro-apoptotic agent: effects of catalase on AT-induced apoptosis, degradation of cytoskeletal proteins and de novo protein synthesis. Br J Haematol. 2006;132(3):349–58

Binet F, Chiasson S, Girard D. Arsenic trioxide induces endoplasmic reticulum stress-related events in neutrophils. Int Immunopharmacol. 10(4):508–12

Binet F, Chiasson S, Girard D. Evidence that endoplasmic reticulum (ER) stress and caspase-4 activation occur in human neutrophils. Biochem Biophys Res Commun. 2010;391(1):18–23

114 Binet F, Girard D. Novel human neutrophil agonistic properties of arsenic trioxide: involvement of p38 mitogen-activated protein kinase and/or c-jun NH2-terminal MAPK but not extracellular signal-regulated kinases-1/2. J Leukoc Biol. 2008a;84(6):1613–1622

Binet F, Chiasson S, Girard D. Arsenic trioxide induces de novo protein synthesis of annexin-1 in neutrophils: association with a heat shock-like response and not apoptosis. Br J Haematol. 2008b;140(4):454–463

Bishayi B, Sengupta M. Intracellular survival of Staphylococcus aureus due to alteration of cellular activity in arsenic and lead intoxicated mature Swiss albino mice. Toxicology. 2003;184(1):31–9

Blasius A, Arnold C, Georgel P, Rutschmann S, Xia Y, Lin P, et al. Slc15a4, AP-3, and Hermansky-Pudlak syndrome proteins are required for Toll-like receptor signaling in plasmacytoid dendritic cells. Proc. Natl. Acad. Sci. 2010;107(46):19973– 8

Blasius A, Vermi W, Krug A, Facchetti F, Cella M, Colonna M. A cell-surface molecule selectively expressed on murine natural interferon-producing cells that blocks secretion of interferon-alpha. Blood. 2004;103(11):4201–6

Blazar BR, Murphy WJ, Abedi M. Advances in graft-versus-host disease biology and therapy. Nat. Rev. Immunol. 2012;12(6):443–58

Blomberg S, Eloranta M, Cederblad B, Nordlin K, Alm G, Rönnblom L. Presence of cutaneous interferon-a producing cells in patients with systemic lupus erythematosus. Lupus. 2001;10(7):484–90

Boasso A, Herbeuval J, Hardy AW, Anderson SA, Dolan MJ, Fuchs D, et al. HIV inhibits CD4+ T-cell proliferation by inducing indoleamine 2,3-dioxygenase in plasmacytoid dendritic cells. Blood. 2007;109(8):3351–9

Bobe P, Bonardelle D, Benihoud K, Opolon P, Chelbi-Alix MK. Arsenic trioxide: A promising novel therapeutic agent for lymphoproliferative and autoimmune syndromes in MRL/lpr mice. Blood. 2006;108(13):3967–75

Boieri M, Shah P, Dressel R, Inngjerdingen M. The role of animal models in the study of hematopoietic stem cell transplantation and GvHD: A historical overview. Front. Immunol. 2016;7(AUG)

Boland J, Atkinson K, Britton K, Darveniza P, Johnson S, Biggs J. Tissue distribution and toxicity of cyclosporin a in the mouse. Pathology. 1984;16(2):117–23

115 Bolaños-Meade J, Reshef R, Fraser R, Fei M, Abhyankar S, Al-Kadhimi Z, et al. Three prophylaxis regimens (tacrolimus, mycophenolate mofetil, and cyclophosphamide; tacrolimus, methotrexate, and bortezomib; or tacrolimus, methotrexate, and maraviroc) versus tacrolimus and methotrexate for prevention of graft-versus-host disease with . Lancet Haematol. 2019;6(3):e132–43

Bolwell B, Sobecks R, Pohlman B, Andresen S, Rybicki L, Kuczkowski E, et al. A prospective randomized trial comparing cyclosporine and short course methotrexate with cyclosporine and mycophenolate mofetil for GVHD prophylaxis in myeloablative allogeneic bone marrow transplantation. Bone Marrow Transplant. 2004;34(7):621–5

Bonifazi F, Solano C, Wolschke C, Sessa M, Patriarca F, Zallio F, et al. Acute GVHD prophylaxis plus ATLG after myeloablative allogeneic haemopoietic peripheral blood stem-cell transplantation from HLA-identical siblings in patients with acute myeloid leukaemia in remission: final results of quality of life and long-term outco. Lancet Haematol. 2019;6(2):e89–99

Bonnefoy F, Couturier M, Clauzon A, Remy-Martin J-P, Gaugler B, TiberghienP,et al. TGF-bata-Exposed Plasmacytoid Dendritic Cells Participate in Th17 Commitment. J. Immunol. 2011;186(11):6157–64

Bossard C, Malard F, Arbez J, Chevallier P, Guillaume T, Delaunay J, et al. Plasmacytoid dendritic cells and Th17 immune response contribution in gastrointestinal acute graft-versus-host disease. Leukemia. 2012;26(7):1471–4

Bourdonnay E, Morzadec C, Fardel O, Vernhet L. Arsenic increases lipopolysaccharide-dependent expression of interleukin-8 gene by stimulating a redox-sensitive pathway that strengthens p38-kinase activation. Mol Immunol. 2011;48(15–16):2069–78

Bourdonnay E, Morzadec C, Fardel O, Vernhet L. Redox-sensitive regulationofgene expression in human primary macrophages exposed to inorganic arsenic. J Cell Biochem. 2009a;107(3):537–547

Bourdonnay E, Morzadec C, Sparfel L, Galibert MD, Jouneau S, Martin-ChoulyC,et al. Global effects of inorganic arsenic on gene expression profile in human macrophages. Mol Immunol. 2009b;46(4):649–656

Bradstock KF, Bilmon I, Kwan J, Micklethwaite K, Blyth E, Deren S, et al. Single-agent high-dose cyclophosphamide for graft-versus-host disease prophylaxis in human leukocyte antigen-matched reduced-intensity peripheral blood stem cell transplantation results in an unacceptably high rate of severe acute graft-versus-host diseas. Biol. Blood Marrow Transplant.2015;21(5):941–4

116 Brkic Z, Van Bon L, Cossu M, Van Helden-Meeuwsen CG, Vonk MC, Knaapen H, et al. The interferon type i signature is present in systemic sclerosis before overt fibrosis and might contribute to its pathogenesis through high BAFF gene expression and high collagen synthesis. Ann. Rheum. Dis. 2016;75(8):1567–73

Bruni D, Chazal M, Sinigaglia L, Chauveau L, Schwartz O, Desprès P, et al. Viral entry route determines how human plasmacytoid dendritic cells produce type I interferons. Sci. Signal. 2015;8(366):ra25-ra25

Cantaert T, Baeten D, Tak PP, van Baarsen LGM. Type I IFN and TNF α cross-regulation in immune-mediated inflammatory disease: Basic concepts and clinical relevance. Arthritis Res. Ther. 2010;12(5):1–10

Cantoni N, Hirsch HH, Khanna N, Gerull S, Buser A, Bucher C, et al. Evidence for a bidirectional relationship between cytomegalovirus replication and acute graft-versus-host disease. Biol. Blood Marrow Transplant. 2010;16(9):1309–14

Cao W, Rosen DB, Ito T, Bover L, Bao M, Watanabe G, et al. Plasmacytoid dendritic cell-specific receptor ILT7-Fc epsilonRI gamma inhibits Toll-like receptor-induced interferon production. J Exp Med. 2006;203(6):1399–405

Cao W, Manicassamy S, Tang H, Kasturi SP, Pirani A, Murthy N, et al. Toll-like receptor-mediated induction of type I interferon in plasmacytoid dendritic cells requires the rapamycin-sensitive PI(3)K-mTOR-p70S6K pathway. Nat. Immunol. 2008;9(10):1157–64

Castermans E, Hannon M, Dutrieux J, Humblet-Baron S, Seidel L, Cheynier R, et al. Thymic recovery after allogeneic hematopoietic cell transplantation with non-myeloablative conditioning is limited to patients younger than 60 years of age. Haematologica. 2011;96(2):298–306

Cella M, Jarrossay D, Facchetti F, Alebardi O, Nakajima H, Lanzavecchia A, et al. Plasmacytoid monocytes migrate to inflamed lymph nodes and produce large amounts of type I interferon. Nat Med. 1999;5(8):919–23

Cervantes-Barragan L, Lewis K, Firner S, Thiel V, Hugues S, Reith W, et al. Plasmacytoid dendritic cells control T-cell response to chronic viral infection. Proc. Natl. Acad. Sci. 2012;109(8):3012–7

Chappell C, Giltiay N, Draves K, Chen C, Hayden-Ledbetter M, Shlomchik M, etal. Targeting Antigens through Blood Dendritic Cell Antigen 2 on Plasmacytoid Dendritic Cells Promotes Immunologic Tolerance. J. Immunol. 2014;192(12):5789– 801

117 Chen GQ, Zhu J, Shi XG, Ni JH, Zhong HJ, Si GY, et al. In vitro studies on cellular and molecular mechanisms of arsenic trioxide (As2O3) in the treatment of acute promyelocytic leukemia As2O3 induces NB4 cell apoptosis with downregulation of Bcl-2 expression and modulation of PML-RAR alpha/PML proteins. Blood. 1996;88(3):1052–61

Chen Y-L, Chen T-T, Pai L-M, Wesoly J, Bluyssen HAR, Lee C-K. A type I IFN– Flt3 ligand axis augments plasmacytoid dendritic cell development from common lymphoid progenitors. J. Exp. Med. 2013;210(12):2515–22

Chou WC, Jie C, Kenedy AA, Jones RJ, Trush MA, Dang C V. Role of NADPH oxidase in arsenic-induced reactive oxygen species formation and cytotoxicity in myeloid leukemia cells. Proc Natl Acad Sci U S A. 2004;101(13):4578–83

Chu KH, Lee CC, Hsin SC, Cai BC, Wang JH, Chiang BL. Arsenic trioxide alleviates airway hyperresponsiveness and eosinophilia in a murine model of asthma. Cell Mol Immunol. 2010;7(5):375–80

Cicconi L, Lo-Coco F. Current management of newly diagnosed acute promyelocytic leukemia. Ann Oncol. 2016;27(8):1474–81

Cisse B, Caton ML, Lehner M, Maeda T, Scheu S, Locksley R, et al. Transcription Factor E2-2 Is an Essential and Specific Regulator of Plasmacytoid Dendritic Cell Development. Cell. 2008;135(1):37–48

Clahsen T, Pabst O, Tenbrock K, Schippers A, Wagner N. Localization of dendritic cells in the gut epithelium requires MAdCAM-1. Clin. Immunol. 2015;156(1):74–84

Conde P, Acosta-Saavedra LC, Goytia-Acevedo RC, Calderon-Aranda ES. Sodium arsenite-induced inhibition of cell proliferation is related to inhibition of IL-2 mRNA expression in mouse activated T cells. Arch Toxicol. 2007;81(4):251–9

Conrad C, Di Domizio J, Mylonas A, Belkhodja C, Demaria O, Navarini AA, et al. TNF blockade induces a dysregulated type i interferon response without autoimmunity in paradoxical psoriasis. Nat. Commun. 2018;9(1):1–11

Cooke K, Kobzik L, Martin T, Brewer J, Delmonte JJ, Crawford J, et al. An experimental model of idiopathic pneumonia syndrome after bone marrow transplantation: I. The roles of minor H antigens and endotoxin. Blood. 1996;88(8):3230–9

118 Couriel D, Saliba R, de Lima M, Giralt S, Andersson B, Khouri I, et al. A Phase III Study of Infliximab and Corticosteroids for the Initial Treatment of Acute Graft-versus-Host Disease. Biol. Blood Marrow Transplant. 2009;15(12):1555–62

Cragg L, Blazar BR, Defor T, Kolatker N, Miller W, Kersey J, et al. A Randomized Trial Comparing Prednisone with Antithymocyte Globulin/Prednisone as anInitial Systemic Therapy for Moderately Severe Acute Graft-Versus-Host Disease. Biol. Blood Marrow Transplant. 2000;6(4 A):441–7

Cutler C, Logan B, Nakamura R, Johnston L, Choi S, Porter D, et al. Tacrolimus/sirolimus vs tacrolimus/methotrexate as GVHD prophylaxis after matched, related donor allogeneic HCT. Blood. 2014;124(8):1372–7

Dalod M, Salazar-Mather T, Malmgaard L, Lewis C, Asselin-Paturel C, BrièreF,etal. Interferon α / β and Interleukin 12 Responses to Viral Infections. J. Exp. Med. 2002;195(4):517–28

D’Amico A, Wu L. The Early Progenitors of Mouse Dendritic Cells and Plasmacytoid Predendritic Cells Are within the Bone Marrow Hemopoietic Precursors Expressing Flt3. J. Exp. Med. 2003;198(2):293 – 303 Dangleben NL, Skibola CF, Smith MT. Arsenic immunotoxicity: a review. Env. Heal. 2013;12(1):73

Dasgupta S, Erturk-Hasdemir D, Ochoa-Reparaz J, Reinecker HC, Kasper DL. Plasmacytoid dendritic cells mediate anti-inflammatory responses to a gut commensal molecule via both innate and adaptive mechanisms. Cell Host Microbe. 2014;15(4):413–23

Dassouki Z, Sahin U, El Hajj H, Jollivet F, Kfoury Y, Lallemand-BreitenbachV,etal. ATL response to arsenic/interferon therapy is triggered by SUMO/PML/RNF4-dependent Tax degradation. Blood. 2015;125(3):474–82

Davison LM, Jørgensen TN. Sialic acid-binding immunoglobulin-type lectin H-positive plasmacytoid dendritic cells drive spontaneous lupus-like disease development in B6.Nba2 mice. Arthritis Rheumatol. 2015;67(4):1012–22 de Bourcy CFA, Dekker CL, Davis MM, Nicolls MR, Quake SR. Dynamics of the human antibody repertoire after B cell depletion in systemic sclerosis. Sci. Immunol. 2017;2(15):1–9 de The H. Differentiation therapy revisited. Nat Rev Cancer. 2018;18(2):117–27

Deaglio S, Canella D, Baj G, Arnulfo A, Waxman S, Malavasi F. Evidence of an immunologic mechanism behind the therapeutical effects of arsenic trioxide (As2O3) on myeloma cells. Leuk Res. 2001;25(3):227–35

119 Dekker JD, Rhee C, Hu Z, Lee B-K, Lee J, Iyer VR, et al. Lymphoid origin of a lineage of intrinsically activated plasmacytoid dendritic cell in mice and humans. bioRxiv. 2018;(512):310680

Denton C. Advances in pathogenesis and treatment of systemic sclerosis. Clin Med. 2014;18(99):195–201

Denton C, Khanna D. Systemic sclerosis. Lancet. 2017;390(10103):1685–99

Di Pucchio T, Chatterjee B, Smed-Sörensen A, Clayton S, Palazzo A, Montes M,etal. Direct proteasome-independent cross-presentation of viral antigen by plasmacytoid dendritic cells on major histocompatibility complex class I. Nat. Immunol. 2008;9(5):551–7

Diacovo TG, Blasius AL, Mak TW, Cella M, Colonna M. Adhesive mechanisms governing interferon-producing cell recruitment into lymph nodes. J. Exp. Med. 2005;202(5):687–96

Diana J, Brezar V, Beaudoin L, Dalod M, Mellor A, Tafuri A, et al. Viral infection prevents diabetes by inducing regulatory T cells through NKT cell – plasmacytoid dendritic cell interplay. J. Exp. Med. 2011;208(4):729–45

Diana J, Griseri T, Lagaye S, Beaudoin L, Autrusseau E, Gautron A, et al. NKT Cell-Plasmacytoid Dendritic Cell Cooperation via OX40 Controls Viral Infectionina Tissue-Specific Manner. Immunity. 2009;30(2):289–99

Diana J, Simoni Y, Furio L, Beaudoin L, Agerberth B, Barrat F, et al. Crosstalk between neutrophils, B-1a cells and plasmacytoid dendritic cells initiates autoimmune diabetes. Nat. Med. 2013;19(1):65–73

Drobits B, Holcmann M, Amberg N, Swiecki M, Grundtner R, Hammer M, et al. Imiquimod clears tumors in mice independent of adaptive immunity by converting pDCs into tumor-killing effector cells. J. Clin. Invest. 2012;122(2):575–85

Duan X, Gao S, Li J, Wu L, Zhang Y, Li W, et al. Acute arsenic exposure induces inflammatory responses and CD4(+) T cell subpopulations differentiation in spleen and thymus with the involvement of MAPK, NF- κ B, and Nrf2. Mol Immunol. 2017;81:160–72

Dzionek A, Sohma Y, Nagafune J, Cella M, Colonna M, Facchetti F, et al. BDCA-2, a Novel Plasmacytoid Dendritic Cell – specific Type II C-type Lectin, Mediates Antigen Capture and Is a Potent Inhibitor of Interferon α/β Induction. J ExpMed. 2001;194(12):1823–34

120 Eapen M, Logan B, Horowitz M, Zhong X, Perales M, Lee S, et al. Bone Marrow or Peripheral Blood for Reduced-Intensity Conditioning Unrelated Donor Transplantation. J. Clin. Oncol. 2015;33(4):364–9

Edinger M, Hoffmann P, Ermann J, Drago K, Garrison Fathman C, Strober S, et al. CD4+CD25+ regulatory T cells preserve graft-versus-tumor activity while inhibiting graft-versus-host disease after bone marrow transplantation. Nat. Med. 2003;9(9):1144–50

El Hajj H, Dassouki Z, Berthier C, Raffoux E, Ades L, Legrand O, et al. Retinoic acid and arsenic trioxide trigger degradation of mutated NPM1, resulting in apoptosis of AML cells. Blood. 2015;125(22):3447–754

El Hajj H, El-Sabban M, Hasegawa H, Zaatari G, Ablain J, Saab ST, et al. Therapy-induced selective loss of leukemia-initiating activity in murine adult T cell leukemia. J Exp Med. 2010;207(13):2785–92

El-Sabban ME, Nasr R, Dbaibo G, Hermine O, Abboushi N, Quignon F, et al. Arsenic-interferon-α – triggered apoptosis in HTLV-I transformed cells is associated with Tax down-regulation and reversal of NF-κB activation. Blood. 2000;96(8):2849– 55

Emadi A, Gore SD. Arsenic trioxide- an old drug rediscovered. Blood Rev. 2010;24(4–5):191–9

EsashiE,BaoM,WangYH,CaoW,LiuYJ.PACSIN1regulatesthe TLR7/9-mediated type I interferon response in plasmacytoid dendritic cells. Eur. J. Immunol. 2012;42(3):573–9

Fagnoni F, Oliviero B, Giorgiani G, De Stefano P, Dehò A, Zibera C, et al. Reconstitution dynamics of plasmacytoid and myeloid dendritic cell precursors after allogeneic myeloablative hematopoietic stem cell transplantation. Blood. 2004;104(1):281–9

Fallarino F, Asselin-Paturel C, Vacca C, Bianchi R, Gizzi S, Fioretti M, et al. Murine Plasmacytoid Dendritic Cells Initiate the Immunosuppressive Pathway of Tryptophan Catabolism in Response to CD200 Receptor Engagement. J. Immunol. 2004;173(6):3748–54

Fancke B, Suter M, Hochrein H, O ’ Keeffe M. M-CSF: A novel plasmacytoid and conventional dendritic cell poietin. Blood. 2008;111(1):150–9

121 Fard NA, Azizi G, Mirshafiey A. The Potential Role of T Helper Cell 22 and IL-22 in Immunopathogenesis of Multiple Sclerosis. Innov Clin Neurosci. 2016;13(7–8):30–6

Fasci D, Anania VG, Lill JR, Salvesen GS. SUMO deconjugation is required for arsenic-triggered ubiquitylation of PML. Sci Signal. 2015;8(380):ra56

Ferrara JL, Levine JE, Reddy P, Holler E. Graft-versus-host disease. Lancet. 2009;373(9674):1550–61

Finke J, Bethge WA, Schmoor C, Ottinger HD, Stelljes M, Zander AR, et al. Standard graft-versus-host disease prophylaxis with or without anti-T-cell globulin in haematopoietic cell transplantation from matched unrelated donors: a randomised, open-label, multicentre phase 3 trial. Lancet Oncol. 2009;10(9):855–64

Flowers MED, Inamoto Y, Carpenter PA, Lee SJ, Kiem H-P, Petersdorf E, et al. Comparative Analysis of Risk Factors for Acute GvHD and for Chronic GvHD disease according to NIH consensus Criteria. Blood. 2011;117(11):3214–9

Fogg D, Sibon C, Miled C, Jung S,Aucouturier, Dan R. Littman, Ana Cumano FG. A clonogenic bone marrow progenitor specific for macrophages and dendritic cells. Science (80-. ). 2006;311(5757):83–7

Fuji S, Kim SW, Yoshimura K ichi, Akiyama H, Okamoto S ichiro, Sao H, et al. Possible Association between Obesity and Posttransplantation Complications Including Infectious Diseases and Acute Graft-versus-Host Disease. Biol. Blood Marrow Transplant. American Society for Blood and Marrow Transplantation; 2009;15(1):73–82

Gale RP, Bortin MM, Van Bekkum DW, Biggs JC, Dicke KA, Gluckman E, et al. Risk factors for acute graft-versus-host disease. Br. J. Haematol. 1987;67(4):397–406

Galicia G, Leyva R, Tenorio EP, Ostrosky-Wegman P, Saavedra R. Sodium arsenite retards proliferation of PHA-activated T cells by delaying the productionand secretion of IL-2. Int Immunopharmacol. 2003;3(5):671–82

Gao C, Jiang J, Ma P, Cheng P, Lian Y, Zhao B, et al. Arsenic Trioxide Induces T Cell Apoptosis and Prolongs Islet Allograft Survival in Mice. Transplantation. 2015;99(9):1796–806

Gao Y, Majchrzak-kita B, Fish EN, Gommerman JL. Dynamic accumulation of plasmacytoid dendritic cells in lymph nodes is regulated by IFN-beta Dynamic Accumulation of Plasmacytoid Dendritic Cells in Lymph Nodes is regulated by IFN-beta Toronto General Research Institute , University Health Network , Toro. Blood. 2009;114(13):2623–32

122 Ghimire S, Weber D, Mavin E, Wang XN, Dickinson AM, Holler E. Pathophysiology of GvHD and other HSCT-related major complications. Front. Immunol. 2017;8(MAR)

Ghosh HS, Cisse B, Bunin A, Lewis KL, Reizis B. Continuous Expression of the Transcription Factor E2-2 Maintains the Cell Fate of Mature Plasmacytoid Dendritic Cells. Immunity. 2010;33(6):905–16

Gilliet M, Boonstra A, Paturel C, Antonenko S, Xu X-L, Trinchieri G, et al. The Development of Murine Plasmacytoid Dendritic Cell Precursors Is Differentially Regulated by FLT3-ligand and Granulocyte/Macrophage Colony-Stimulating Factor. J. Exp. Med. 2002;195(7):953–8

Gleimer M, Li Y, Chang L, Paczesny S, Hanauer D, Frame D, et al. Baseline body mass index among children and adults undergoing allogeneic hematopoietic cell transplantation: clinical characteristics and outcomes. Bone Marrow Transplant. 2015;50(3):402–10

Glitzner E, Korosec A, Brunner P, Drobits B, Amberg N, Schonthaler H, et al. Specific roles for dendritic cell subsets during initiation and progression of psoriasis. EMBO Mol. Med. 2014;6(10):1312–27

Gorski J, Chen X, Gendelman M, Yassai M, Krueger A, Tivol E, et al. Homeostatic expansion and repertoire regeneration of donor T cells during graft versus host disease is constrained by the host environment. Blood. 2007;109(12):5502–10

Gotoh K, Tanaka Y, Nishikimi A, Inayoshi A, Enjoji M, Takayanagi R, et al. Differential requirement for D0CK2 in migration of plasmacytoid dendritic cells versus myeloid dendritic cells. Blood. 2008;111(6):2973–6

Goulmy E, Schipper R, Pool J, Blokland E, Falkenburg J, Vossen J, et al. Mismatches of Minor Histocompatibility Antigens between HLA-Identical Donors and Recipients and the Development of Graft-Versus-Host Disease after Bone Marrow Transplantation. N. Engl. J. Med. 1996;334(5):281–5

Grouard G, Rissoan MC, Filgueira L, Durand I, Banchereau J, Liu YJ. The enigmatic plasmacytoid T cells develop into dendritic cells with interleukin (IL)-3and CD40-ligand. J. Exp. Med. 1997;185(6):1101–11

Guiducci C, Gong M, Xu Z, Gill M, Chaussabel D, Meeker T, et al. TLR recognition of self nucleic acids hampers glucocorticoid activity in lupus. 2010;465(7300):937–41

123 Guiducci C, Ott G, Chan J, Damon E, Calacsan C, Matray T, et al. Properties regulating the nature of the plasmacytoid dendritic cell response to Toll-like receptor 9 activation. J. Exp. Med. 2006;203(8):1999–2008

Guilliams M, Mildner A, Yona S. Developmental and Functional Heterogeneity of Monocytes. Immunity. 2018;49(4):595–613

Gupta A, Punatar S, Mathew L, Kannan S, Khattry N. Cyclosporine Plus Methotrexate or Cyclosporine Plus Mycophenolate Mofetil as Graft Versus Host Disease Prophylaxis in Acute Leukemia Transplant: Comparison of Toxicity, Engraftment Kinetics and Transplant Outcome. Indian J. Hematol. Blood Transfus. 2016;32(3):248–56

Gupta S, Yel L, Kim D, Kim C, Chiplunkar S, Gollapudi S. Arsenic trioxide induces apoptosis in peripheral blood T lymphocyte subsets by inducing oxidative stress: a role of Bcl-2. Mol Cancer Ther. 2003;2(8):711–9

Hadeiba H, Lahl K, Edalati A, Oderup C, Habtezion A, Pachynski R, et al. Plasmacytoid Dendritic Cells Transport Peripheral Antigens to the Thymusto Promote Central Tolerance. Immunity. 2012;36(3):438–50

Hadeiba H, Sato T, Habtezion A, Oderup C, Pan J, Butcher EC. CCR9 expression defines tolerogenic plasmacytoid dendritic cells able to suppress acute graft-versus-host disease. Nat. Immunol. 2008;9(11):1253–60

Hagberg N, Berggren O, Leonard D, Weber G, Bryceson Y, Alm G, et al. IFN- α Production by Plasmacytoid Dendritic Cells Stimulated with RNA-Containing Immune Complexes Is Promoted by NK Cells via MIP-1 and LFA-1. J. Immunol. 2011;186(9):5085–94

Hahn T, McCarthy PL, Zhang MJ, Wang D, Arora M, Frangoul H, et al. Risk factors for acute graft-versus-host disease after human leukocyte antigen-identical sibling transplants for adults with leukemia. J. Clin. Oncol. 2008;26(35):5728–34

Hamilton BK. Current approaches to prevent and treat GVHD after allogeneicstem cell transplantation. Hematol. Am. Soc. Hematol. Educ. Progr. 2018a;2018(1):228 – 235

Hamilton B, Liu Y, Hemmer M, Wang T, Chhabra S, Costa LJ, et al. Cyclosporine in Combination with Mycophenolate Mofetil Leads to Increased Incidence of Graft-Versus-Host Disease and Inferior Outcomes after Myeloablative Allogeneic Hematopoietic Cell Transplantation: A CIBMTR Analysis. Biol. Blood Marrow Transplant. 2018b;24(3):S185

124 Hansen L, Schmidt-Christensen A, Gupta S, Fransén-Pettersson N, Hannibal T, Reizis B, et al. E2-2 Dependent Plasmacytoid Dendritic Cells Control Autoimmune Diabetes. PLoS One. 2015;10(12):e0144090

Harris AC, Young R, Devine S, Hogan WJ, Ayuk F, Bunworasate U, et al. International, Multicenter Standardization of Acute Graft-versus-HostDisease Clinical Data Collection: A Report from the Mount Sinai Acute GVHD International Consortium. Biol. Blood Marrow Transplant. 2016;22(1):4–10

Hayashi K, Taura M, Iwasaki A. The interaction between IKKα and LC3 promotes type i interferon production through the TLR9-containing LAPosome. Sci. Signal. 2018;11(528):1–11

He X, Ye Y, Xu X, Wang J, Huang Y, Weng G, et al. Conditioning with Fludarabine-Busulfan versus Busulfan-Cyclophosphamide Is Associated with Lower aGVHD and Higher Survival but More Extensive and Long Standing Bone Marrow Damage. Biomed Res. Int. 2016;2016(3071214)

Henault J, Martinez J, Riggs JM, Tian J, Mehta P, Sasai M, et al. Noncanonical Autophagy Is Required for Type I Interferon Secretion in Response to DNA-Immune Complexes. Immunity. 2012;37(6):986–97

Henden AS, Hill GR. Cytokines in Graft-versus-Host Disease. J. Immunol. 2015;194(10):4604–12

HermanJS,GrünS,GrünD.FateIDinfers cell fate bias in multipotent progenitors from single-cell RNA-seq data. Nat. Methods. 2018;15(5):379–86

Highfill S, Rodriguez P, Zhou Q, Goetz C, Koehn B, Veenstra R, et al. Bone marrow myeloid-derived suppressor cells (MDSCs) inhibit graft-versus-host disease (GVHD) via an arginase-1-dependent mechanism that is up-regulated by interleukin-13. Blood. 2010;116(25):5738–47

Hoeffel G, Ripoche A, Matheoud D, Nascimbeni M, Escriou N, Lebon P, et al. Antigen Crosspresentation by Human Plasmacytoid Dendritic Cells. Immunity. 2007;27(3):481–92

Holtan SG, DeFor TE, Lazaryan A, Bejanyan N, Arora M, Brunstein CG, et al. Composite end point of graft-versus-host disease-free, relapse-free survival after allogeneic hematopoietic cell transplantation. Blood. 2015;125(8):1333–8

Holtan SG, Khera N, Levine JE, Chai X, Storer B, Liu HD, et al. Late acute graft-versus-host disease: A prospective analysis of clinical outcomes and circulating angiogenic factors. Blood. 2016;128(19):2350–8

125 Holtan SG, Marcelo P, Weisdorf DJ. Acute graft-versus-host disease: A bench-to-bedside update. Blood. 2014;124(3):363–73

Holtick U, Chemnitz JM, Shimabukuro-Vornhagen A, Theurich S, Chakupurakal G, Krause A, et al. OCTET-CY: A phase II study to investigate the efficacy of post-transplant cyclophosphamide as sole graft-versus-host prophylaxis after allogeneic peripheral blood stem cell transplantation. Eur. J. Haematol. 2016;96(1):27–35

Honda K, Yanai H, Negishi H, Asagiri M, Sato M, Mizutani T, et al. IRF-7 is the master regulator of type-I interferon-dependent immune responses. Nature. 2005;434(7034):772–777a

Honda K, Ohba Y, Yanai H, Negishi H, Mizutani T, Takaoka A, et al. Spatiotemporal regulation of MyD88–IRF-7 signalling for robust type-I interferon induction. Nature. 2005;434(7036):1035–1040b

Horváth R, Budinský V, Kayserová J, Kalina T, Formánková R, Starý J, et al. Kinetics of dendritic cells reconstitution and costimulatory molecules expression after myeloablative allogeneic haematopoetic stem cell transplantation: Implications for the development of acute graft-versus host disease. Clin. Immunol. 2009;131(1):60–9

Hoshino K, Sugiyama T, Matsumoto M, Tanaka T, Saito M, Hemmi H, et al. I κ B kinase-α is critical for interferon-α production induced by Toll-like receptors 7 and 9. Nature. 2006;440(7086):949–53

Hsu KS, Kao HY. PML: Regulation and multifaceted function beyond tumor suppression. Cell Biosci. 2018;8:5

Hu Z, Hu S, Wu Y, Li S, He C, Xing X, et al. Accumulation and suppressive function of regulatory T cells in malignant ascites: Reducing their suppressive function using arsenic trioxide in vitro. Oncol Lett. 2018;15(4):5384–90

Hutchinson LM, Trinh BM, Palmer RK, Preziosi CA, Pelletier JH, Nelson HM, etal. Inorganic arsenite inhibits IgE receptor-mediated degranulation of mast cells. J Appl Toxicol. 2011;31(3):231–41

Ishikawa F, Niiro H, Iino T, Yoshida S, Saito N, Onohara S, et al. The developmental program of human dendritic cells is operated independently of conventional myeloid and lymphoid pathways. Blood. 2007;110(10):3591–660

126 Ito T, Yang M, Wang Y, Lande R, Gregorio J, Perng O, et al. Plasmacytoid dendritic cells prime IL-10 – producing T regulatory cells by inducible costimulator ligand. J. Exp. Med. 2007;204(1):105–15

Izaguirre A, Barnes B, Amrute S, Yeow W, Megjugorac N, Dai J, et al. Comparative analysis of IRF and IFN-alpha expression in human plasmacytoid and monocyte-derived dendritic cells. J. Leukoc. Biol. 2003;74(6):1125–38

Jagasia M, Arora M, Flowers M, Chao N, McCarthy P, Cutler C, et al. Risk factors for acute GVHD and survival after hematopoietic cell transplantation. Blood. 2012;119(1):296–307

Jagasia MH, Greinix HT, Arora M, Williams KM, Wolff D, Cowen EW, et al. National Institutes of Health Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-versus-Host Disease: I. The 2014 Diagnosis and Staging Working Group Report. Biol. Blood Marrow Transplant. 2015;21(3):389–401.e1

Jahrsdörfer B, Vollmer A, Blackwell S, Maier J, Sontheimer K, Beyer T, et al. Granzyme B produced by human plasmacytoid dendritic cells suppresses T-cell expansion. Blood. 2010;115(6):1156–65

Jeanne M, Lallemand-Breitenbach V, Ferhi O, Koken M, Le Bras M, Duffort S, etal. PML/RARA Oxidation and Arsenic Binding Initiate the Antileukemia Response of As2O3. Cancer Cell. 2010;18(1):88–98

Jego G, Palucka AK, Blanck JP, Chalouni C, Pascual V, Banchereau J. Plasmacytoid Dendritic Cells Induce Plasma Cell Differentiation through Type I Interferon and Interleukin 6. Immunity. 2003;19(2):225–34

Jethava YS, Sica S, Savani B, Socola F, Jagasia M, Mohty M, et al. Conditioning regimens for allogeneic hematopoietic stem cell transplants in acute myeloid leukemia. Bone Marrow Transplant. 2017;52(11):1504–11

JiaoZX,LengY,XiaJJ,WuHQ,JinN,FuJZ,etal.As2O3combinedwith leflunomide prolongs heart xenograft survival via suppressing the response of Th1, Th2, and B cells in a rat model. Xenotransplantation. 2016;23(3):237–48

Kadowaki N, Antonenko S, Lau JY-N, Liu Y-J. Natural Interferon α/β–Producing Cells Link Innate and Adaptive Immunity. J. Exp. Med. 2002;192(2):219–26

Kafaja S, Valera I, Divekar AA, Saggar R, Abtin F, Furst DE, et al. pDCs in lungand skin fibrosis in a bleomycin-induced model and patients with systemic sclerosis. JCI Insight. 2018;3(9)

127 Kale J, Osterlund EJ, Andrews DW. BCL-2 family proteins: Changing partnersinthe dance towards death. Cell Death Differ. 2018;25(1):65–80

Kanakry CG, Ganguly S, Zahurak M, Bolanos-Meade J, Thoburn C, Perkins B, et al. Aldehyde Dehydrogenase Expression Drives Human Regulatory T Cell Resistance to Posttransplantation Cyclophosphamide. Sci. Transl. Med. 2013;5(211):211ra157-211ra157

Kanakry C, Tsai H, Bolaños-Meade J, Smith B, Gojo I, Kanakry J, et al. Single-agent GVHD prophylaxis with posttransplantation cyclophosphamide after myeloablative, HLA-matched BMT for AML, ALL, and MDS. Blood. 2014;124(25):3817–27

Kavian N, Marut W, Servettaz A, Laude H, Nicco C, Chereau C, et al. Arsenic trioxide prevents murine sclerodermatous graft-versus-host disease. J Immunol. 2012a;188(10):5142–5149

Kavian N, Marut W, Servettaz A, Nicco C, Chereau C, Lemarechal H, et al. Reactive oxygen species-mediated killing of activated fibroblasts by arsenic trioxide ameliorates fibrosis in a murine model of systemic sclerosis. Arthritis Rheum. 2012b;64(10):3430–3440

Kchour G, Tarhini M, Kooshyar MM, El Hajj H, Wattel E, Mahmoudi M, et al. Phase 2 study of the efficacy and safety of the combination of arsenic trioxide, interferon alpha, and zidovudine in newly diagnosed chronic adult T-cell leukemia/lymphoma (ATL). Blood. 2009;113(26):6528–32

Kchour G, Rezaee SAR, Farid R, Ghantous A, Rafatpanah H, Tarhini M, et al. The combination of arsenic, interferon-alpha, and zidovudine restores an “ immunocompetent-like” cytokine expression profile in patients with adult T-cell leukemia lymphoma. Retrovirology. 2013;10(1):1–12

Kean L, Sen S, Onabajo O, Singh K, Robertson J, Stempora L, et al. Significant mobilization of both conventional and regulatory T cells with AMD3100. Blood. 2011;118(25):6580–90

Kim D, Peck A, Santer D, Patole P, Schwartz SM, Molitor JA, et al. Induction of interferon-alpha by scleroderma sera containing autoantibodies to topoisomerase I: association of higher interferon-alpha activity with lung fibrosis. Arthritis Rheum. 2008;58(7):2163–73

Kim JY, Bae JH, Lee SH, Lee EY, Chung BS, Kim SH, et al. Induction of NKG2D ligands and subsequent enhancement of NK cell-mediated lysis of cancer cells by arsenic trioxide. J Immunother. 2008;31(5):475–486a

128 Kim S, Kaiser V, Beier E, Bechheim M, Guenthner-Biller M, Ablasser A, et al. Self-priming determines high type I IFN production by plasmacytoid dendritic cells. Eur. J. Immunol. 2014;44(3):807–18

Kim T, Pazhoor S, Bao M, Zhang Z, Hanabuchi S, Facchinetti V, et al. Aspartate-glutamate-alanine-histidine box motif (DEAH)/RNA helicase A helicases sense microbial DNA in human plasmacytoid dendritic cells. Proc. Natl. Acad. Sci. 2010;107(34):15181–6

Kim YE, Ahn JH. Positive Role of Promyelocytic Leukemia Protein in Type I Interferon Response and Its Regulation by Human Cytomegalovirus. PLoS Pathog. 2015;11(3):1–24

Klangsinsirikul P, Russell NH. Peripheral blood stem cell harvests from G-CSF-stimulated donors contain a skewed Th2 CD4 phenotype and a predominance of type 2 dendritic cells. Exp. Hematol. 2002;30(5):495–501

Kohara H, Omatsu Y, Sugiyama T, Noda M, Fujii N, Nagasawa T. Development of plasmacytoid dendritic cells in bone marrow stromal cell niches requires CXCL12-CXCR4 chemokine signaling. Blood. 2007;110(13):4153–60

Koyama M, Hashimoto D, Aoyama K, Matsuoka K, Karube K, Niiro H, et al. Plasmacytoid dendritic cells prime alloreactive T cells to mediate graft-versus-host disease as antigen-presenting cells. Blood. 2009;113(9):2088–95

Koyama M, Kuns R, Olver S, Raffelt N, Wilson Y, Don A, et al. Recipient nonhematopoietic antigen-presenting cells are sufficient to induce lethal acute graft-versus-host disease. Nat. Med. 2011;18(1):135–42

Kröger N, Solano C, Wolschke C, Bandini, G, Patriarca F, Pini M, et al. Antilymphocyte Globulin for Prevention of Chronic Graft-versus-Host Disease. N. Engl. J. Med. 2016;374(1):43–53

Krug A, Uppaluri R, Facchetti F, Dorner BG, Sheehan KCF, Schreiber RD, et al. Cutting Edge: IFN-Producing Cells Respond to CXCR3 Ligands in the Presenceof CXCL12 and Secrete Inflammatory Chemokines upon Activation. J. Immunol. 2002;169(11):6079–83

Kumagai Y, Kumar H, Koyama S, Kawai T, Takeuchi O, Akira S. Cutting Edge: TLR-Dependent Viral Recognition Along with Type I IFN Positive Feedback Signaling Masks the Requirement of Viral Replication for IFN- Production in Plasmacytoid Dendritic Cells. J. Immunol. 2009;182(7):3960–4

129 Kumar A, Reljic T, Hamadani M, Mohty M, Kharfan-Dabaja MA. Antithymocyte globulin for graft-versus-host disease prophylaxis: an updated systematic review and meta-analysis. Bone Marrow Transplant. 2018; [Epub ahead of print]

Lallemand-Breitenbach V, de Thé H. PML nuclear bodies from architecture to function. Curr Opin Cell Biol. 2018;52:154–61

Lallemand-Breitenbach V, Zhu J, Puvion F, Koken M, Honoré N, Doubeikovsky A, et al. Role of Promyelocytic Leukemia (Pml) Sumolation in Nuclear Body Formation, 11s Proteasome Recruitment, and as 2 O 3 -Induced Pml or Pml/Retinoic Acid Receptor α Degradation. J. Exp. Med. 2001;193(12):1361–72

Laouar Y, Welte T, Fu XY, Flavell RA. STAT3 Is Required for Flt3L-Dependent Dendritic Cell Differentiation. Immunity. 2003;19(6):903–12

Laurent P, Sisirak V, Lazaro E, Richez C, Duffau P, Blanco P, et al. Innate immunity in systemic sclerosis fibrosis: Recent advances. Front. Immunol. 2018;9(JUL)

Laustsen A, Bak R, Krapp C, Kjær L, Egedahl J, Petersen C, et al. Interferon priming is essential for human CD34+ cell-derived plasmacytoid dendritic cell maturation and function. Nat. Commun. 2018;9(1)

Lee HK, Lund JM, Ramanathan B, Mizushima N, Iwasaki A. Autophagy-dependent viral recognition by plasmacytoid dendritic cells. Science (80-. ). 2007;315(5817):1398–401

Lee YS, Kim D, Lee EK, Kim S, Choi CS, Jun HS. Sodium meta-arsenite prevents the development of autoimmune diabetes in NOD mice. Toxicol Appl Pharmacol. 2015;284(2):254–61

Lemarie A, Morzadec C, Bourdonnay E, Fardel O, Vernhet L. Human Macrophages Constitute Targets for Immunotoxic Inorganic Arsenic. J. Immunol. 2006a;177(5):3019–3027

Lemarie A, Morzadec C, Merino D, Micheau O, Fardel O, Vernhet L. Arsenic trioxide induces apoptosis of human monocytes during macrophagic differentiation through nuclear factor-kappaB-related survival pathway down-regulation. J Pharmacol Exp Ther. 2006b;316(1):304–314

Lemarie A, Bourdonnay E, Morzadec C, Fardel O, Vernhet L. Inorganic Arsenic Activates Reduced NADPH Oxidase in Human Primary Macrophages through a Rho Kinase/p38 Kinase Pathway. J. Immunol. 2008;180(9):6010–7

130 Lepelletier Y, Zollinger R, Ghirelli C, Raynaud F, Hadj-Slimane R, CappuccioA,et al. Toll-like receptor control of glucocorticoid-induced apoptosis in human plasmacytoid predendritic cells (pDCs). Blood. 2010;116(18):3389–97

Leung J, Pang A, Yuen WH, Kwong YL, Tse EW. Relationship of expression of aquaglyceroporin 9 with arsenic uptake and sensitivity in leukemia cells. Blood. 2007;109(2):740–6

Leveque-El Mouttie L, Koyama M, Le Texier L, Markey K, Cheong M, Kuns R, et al. Corruption of dendritic cell antigen presentation during acute GVHD leadsto regulatory T-cell failure and chronic GVHD. Blood. 2016;128(6):794–804

LiHS,YangCY,NallaparajuKC,ZhangH,LiuYJ,GoldrathAW,etal.Thesignal transducers STAT5 and STAT3 control expression of Id2 and E2-2 during dendritic cell development. Blood. 2012a;120(22):4363–4373

Li H, Demetris A, McNiff J, Matte-Martone C, Tan H, Rothstein D, et al. Profound Depletion of Host Conventional Dendritic Cells, Plasmacytoid Dendritic Cells, and B Cells Does Not Prevent Graft-versus-Host Disease Induction. J. Immunol. 2012b;188(8):3804–3811

Li C, Guan T, Gao C, Lin Y, Yan G, Zhu M, et al. Arsenic trioxide inhibits accelerated allograft rejection mediated by alloreactive CD8(+) memory T cells and prolongs allograft survival time. Transpl Immunol. 2015a;33(1):30–36

[Establishment of a graft-versus-host disease mouse model for haplo-identical bone marrow transplantation with busulfan and fludarabine conditioning regimen]. Zhonghua Xue Ye Xue Za Zhi. 2015b;36(3):238–240

Li J, Zhao L, Zhang Y, Li W, Duan X, Chen J, et al. Imbalanced immune responses involving inflammatory molecules and immune-related pathways in the lung of acute and subchronic arsenic-exposed mice. Env. Res. 2017;159:381–93

Li J, Kim SG, Blenis J. Rapamycin: One drug, many effects. Cell Metab. Elsevier Inc.; 2014;19(3):373–9

Li N, Hu W, Li X, Yang W, Chen P, Chen Y. Li X, Wu J, Gao D, Wang H, Sun L, Chen Z. Pivotal roles of cGAS-cGAMP signaling in antiviral defense and immune adjuvant effects. Science (80-. ). 2013;341(6152):1390–4

Lin Y, Dai H, Su J, Yan G, Xi Y, Ekberg H, et al. Arsenic trioxide is a novel agentfor combination therapy to prolong heart allograft survival in allo-primed T cells transferred mice. Transpl Immunol. 2011;25(4):194–201

131 Liu J, Farmer JD, Lane WS, Friedman J, Weissman I, Schreiber SL. Calcineurinisa common target of cyclophilin-cyclosporin A and FKBP-FK506 complexes. Cell. 1991;66(4):807–15

Liu K, Victora G, Schwickert T, Guermonprez P, Meredith M, Yao K, et al. In Vivo Analysis of Dendritic Cell. Science. 2009;324(April):392–7

Liu M, Wu W, Sun X, Yang J, Xu J, Fu W, et al. New insights into CD4+T cell abnormalities in systemic sclerosis. Cytokine Growth Factor Rev. 2016;28:31–6

Lo YH, Huang YW, Wu YH, Tsai CS, Lin YC, Mo ST, et al. Selective inhibition of the NLRP3 inflammasome by targeting to promyelocytic leukemia protein in mouse and human. Blood. 2013;121(16):3185–94

Lo-Coco F, Avvisati G, Vignetti M, Thiede C, Orlando SM, Iacobelli S, et al. Retinoic acid and arsenic trioxide for acute promyelocytic leukemia. N Engl J Med. 2013;369(2):111–21

Lombardi V, Speak AO, Kerzerho J, Szely N, Akbari O. CD8α +Β - and CD8α +Β - plasmacytoid dendritic cells induce Foxp3 + regulatory T cells and prevent the induction of airway hyper-reactivity. Mucosal Immunol. 2012;5(4):432–43

Loschko J, Heink S, Krug AB, Hackl D, Reindl W, Korn T, et al. Antigen Targeting to Plasmacytoid Dendritic Cells via Siglec-H Inhibits Th Cell-Dependent Autoimmunity. J. Immunol. 2011a;187(12):6346–6356

Loschko J, Schlitzer A, Dudziak D, Drexler I, Sandholzer N, Bourquin C, et al. Antigen Delivery to Plasmacytoid Dendritic Cells via BST2 Induces Protective T Cell-Mediated Immunity. J. Immunol. 2011b;186(12):6718–6725

Loschko J, Rieke GJ, Schreiber HA, Meredith MM, Yao KH, Guermonprez P, et al. Inducible targeting of cDCs and their subsets in vivo. J. Immunol. Methods. 2016;434:32–8

Lowin B, Hahne M, Mattmann C, Tschopp J. Cytolytic T-cell cytotoxicity is mediated through perforin and Fas lytic pathways. Nature. 1994;370(6491):650–2

Lu W, Mehraj V, Vyboh K, Cao W, Li T, Routy JP. CD4:CD8 ratio as a frontier marker for clinical outcome, immune dysfunction and viral reservoir size in virologically suppressed HIV-positive patients. J Int AIDS Soc. 2015;18:20052

Lu Y, Giver C, Sharma A, Li J, Darlak K, Owens L, et al. IFN- and indoleamine 2,3-dioxygenase signaling between donor dendritic cells and T cells regulates graft versus host and graft versus leukemia activity. Blood. 2011;119(4):1075–85

132 Luo L, Qiao H, Meng F, Dong X, Zhou B, Jiang H, et al. Arsenic trioxide synergizes with B7H3-mediated immunotherapy to eradicate hepatocellular carcinomas. Int J Cancer. 2006;118(7):1823–30

Luznik L, Jalla S, Engstrom LW, Iannone R, Fuchs EJ. Durable engraftment of major histocompatibility complex-incompatible cells after nonmyeloablative conditioning with fludarabine, low-dose total body irradiation, and posttransplantation cyclophosphamide. Blood. 2001;98(12):3456–64

Maazi H, Lam J, Lombardi V, Akbari O. Role of plasmacytoid dendritic cell subsets in allergic asthma. Allergy. 2013; 68(6): 695-701

MacDonald KPA, Rowe V, Hume DA, Sasmono T, Bofinger HM, Thomas R, et al. The Colony-Stimulating Factor 1 Receptor Is Expressed on Dendritic Cells during Differentiation and Regulates Their Expansion. J. Immunol. 2005;175(3):1399–405

Macoch M, Morzadec C, Fardel O, Vernhet L. Inorganic arsenic impairs differentiation and functions of human dendritic cells. Toxicol Appl Pharmacol. 2013;266(2):204–13

Macoch M, Morzadec C, Genard R, Pallardy M, Kerdine-Romer S, Fardel O, et al. Nrf2-dependent repression of interleukin-12 expression in human dendritic cells exposed to inorganic arsenic. Free Radic Biol Med. 2015;88(Pt B):381–90

Maier NK, Crown D, Liu J, Leppla SH, Moayeri M. Arsenic trioxide and other arsenical compounds inhibit the NLRP1, NLRP3, and NAIP5/NLRC4 inflammasomes. J Immunol. 2014;192(2):763–70

Major-monfried H, Renteria AS, Pawarode A, Reddy P, Ayuk F, Holler E, et al. MAGIC biomarkers predict long-term outcomes for steroid-resistant acute GVHD. Blood. 2018;131(25):2846–55

Malard F, Bossard C, Brissot E, Chevallier P, Guillaume T, Delaunay J, et al. Increased plasmacytoid dendritic cells and ROR γ t-expressing immune effectors in cutaneous acute graft-versus-host disease. J. Leukoc. Biol. 2013;94(6):1337–43

Malard F, Labopin M, Cho C, Blaise D, Papadopoulos EB, Passweg J, et al. Ex vivo and in vivo T cell-depleted allogeneic stem cell transplantation in patients with acute myeloid leukemia in first complete remission resulted in similar overall survival: On behalf of the ALWP of the EBMT and the MSKCC. J. Hematol. Oncol. 2018;11(1):1–11

133 Malloy MT, McIntosh DJ, Walters TS, Flores A, Goodwin JS, Arinze IJ. Trafficking of the transcription factor Nrf2 to promyelocytic leukemia-nuclear bodies: Implications for degradation of nrf2 in the nucleus. J. Biol. Chem. 2013;288(20):14569–83

Mao MH, Huang HB, Zhang XL, Li K, Liu YL, Wang P. Additive antitumor effect of arsenic trioxide combined with intravesical bacillus Calmette-Guerin immunotherapy against bladder cancer through blockade of the IER3/Nrf2 pathway. Biomed Pharmacother. 2018;107:1093–103

Marafioti T, Paterson JC, Ballabio E, Reichard KK, Tedoldi S, Hollowood K, et al. Novel markers of normal and neoplastic human plasmacytoid dendritic cells. Blood. 2008;111(7):3778–92

Martelli MP, Gionfriddo I, Mezzasoma F, Milano F, Pierangeli S, Mulas F, et al. Arsenic trioxide and all-trans retinoic acid target NPM1 mutant oncoprotein levels and induce apoptosis in NPM1-mutated AML cells. Blood. 2015;125(22):3455–65

Martin P, Levine D, Storer B, Warren E, Zheng X, Nelson S, et al. Genome-wide minor histocompatibility matching as related to the risk of graft-versus-host disease. Blood. 2017;129(6):791–8

Martin P, Rizzo J, Wingard J, Ballen K, Curtin P, Cutler C, et al. First- and Second-Line Systemic Treatment of Acute Graft-versus-Host Disease: Recommendations of the American Society of Blood and Marrow Transplantation. Biol. Blood Marrow Transplant. 2012;18(8):1150–63

Mathew R, Haribhai D, Kohlhapp F, Duggan R, Ellis P, Riehm J, et al. The BCL-2-Selective Inhibitor Venetoclax Spares Activated T-Cells during Anti-Tumor Immunity. Blood. 2018;132(Suppl 1):3704

Mathewson N, Jenq R, Mathew A, Koenigsknecht M, Hanash A, Toubai T, et al. Gut microbiome – derived metabolites modulate intestinal epithelial cell damage and mitigate graft-versus-host disease. Nat. Immunol. 2016;17(5):505–13

Matsui T, Connolly JE, Michnevitz M, Chaussabel D, Yu C-I, Glaser C, et al. CD2 Distinguishes Two Subsets of Human Plasmacytoid Dendritic Cells with Distinct Phenotype and Functions. J. Immunol. 2009;182(11):6815–23

McKenna H, Stocking K, Miller R, Brasel K, De Smedt T, Maraskovsky E, et al. Mice lacking flt3 ligand have deficient hematopoiesis affecting hematopoietic progenitor cells, dendritic cells, and natural killer cells. Blood. 2000;95(11):3489–97

134 Means TK, Latz E, Hayashi F, Murali MR, Golenbock DT, Luster AD. Human lupus autoantibody-DNA complexes activate DCs through cooperation of CD32 and TLR9. J. Clin. Invest. 2005;115(2):407–17

Mei Y, Zheng Y, Wang H, Gao J, Liu D, Zhao Y, et al. Arsenic trioxide induces apoptosis of fibroblast-like synoviocytes and represents antiarthris effect in experimental model of rheumatoid arthritis. J Rheumatol. 2011;38(1):36–43

Menon M, Blair PA, Isenberg DA, Mauri C. A Regulatory Feedback between Plasmacytoid Dendritic Cells and Regulatory B Cells Is Aberrant in Systemic Lupus Erythematosus. Immunity. 2016;44(3):683–97

Meyer-wentrup F, Benitez-ribas D, Tacken PJ, Punt CJA, Figdor CG, Vries IJMDe, et al. Targeting DCIR on human plasmacytoid dendritic cells results in antigen presentation and inhibits IFN- α production. Blood. 2008;111(8):4245–53

Mielcarek M, Furlong T, Storer B, Green M, McDonald G, Carpenter P, et al. Effectiveness and safety of lower dose prednisone for initial treatment of acute graft-versus-host disease: a randomized controlled trial. Haematologica. 2015;100(6):842–8

Mohty M. Mechanisms of action of antithymocyte globulin: T-cell depletionand beyond. Leukemia. 2007;21(7):1387–94

Mohty M, Gastaut J-A, Viens P, Olive D, Blaise D, Bardou V-J, et al. Impact of plasmacytoid dendritic cells on outcome after reduced-intensity conditioning allogeneic stem cell transplantation. Leukemia. 2005;19(1):1–6

Mohty M, Malard F. Antithymocyte Globulin for Graft-Versus-Host Disease Prophylaxis After Allogeneic Hematopoietic Stem-Cell Transplantation. J. Clin. Oncol. 2017;35(36):3993–5

Morris ES, MacDonald KPA, Rowe V, Johnson DH, Banovic T, Clouston AD, et al. Donor treatment with pegylated G-CSF augments the generation of IL-10-producing regulatory T cells and promotes transplantation tolerance. Blood. 2004;103(9):3573– 81

Morzadec C, Bouezzedine F, Macoch M, Fardel O, Vernhet L. Inorganic arsenic impairs proliferation and cytokine expression in human primary T lymphocytes. Toxicology. 2012a;300(1–2):46–56

Morzadec C, Macoch M, Robineau M, Sparfel L, Fardel O, Vernhet L. Inorganic arsenic represses interleukin-17A expression in human activated Th17 lymphocytes. Toxicol Appl Pharmacol. 2012b;262(3):217–222

135 Morzadec C, Macoch M, Sparfel L, Kerdine-Romer S, Fardel O, Vernhet L. Nrf2 expression and activity in human T lymphocytes: stimulation by T cell receptor activation and priming by inorganic arsenic and tert-butylhydroquinone.FreeRadic Biol Med. 2014;71:133–45

Mouriès J, Moron G, Schlecht G, Escriou N, Dadaglio G, Leclerc C. Plasmacytoid dendritic cells efficiently cross-prime naive T cells in vivo after TLR activation. Blood. 2008;112(9):3713–22

Munn DH, Sharma MD, Hou D, Baban B, Lee JR, Antonia SJ, et al. Expression of indoleamine 2,3-dioxygenase by plasmacytoid dendritic cells in tumor-draining lymph nodes. J. Clin. Invest. 2004;114(2):280–90

Naik SH, Sathe P, Park H-Y, Metcalf D, Proietto AI, Dakic A, et al. Development of plasmacytoid and conventional dendritic cell subtypes from single precursor cells derived in vitro and in vivo. Nat. Immunol. 2007;8(11):1217–26

Nakano H, Yanagita M, Gunn MD. Cd11c+B220+Gr-1+Cells in Mouse Lymph Nodes and Spleen Display Characteristics of Plasmacytoid Dendritic Cells. J Exp Med. 2001;194(8):1171–8

Nash RA, Antin JH, Karanes C, Fay JW, Avalos BR, Yeager AM, et al. Phase 3 study comparing methotrexate and tacrolimus with methotrexate and cyclosporine for prophylaxis of acute graft-versus-host disease after marrow transplantation from unrelated donors. Blood. 2000;96(6):2062–8

Nehmar R, Alsaleh G, Voisin B, Flacher V, Mariotte A, Saferding V, et al. Therapeutic Modulation of Plasmacytoid Dendritic Cells in Experimental Arthritis. Arthritis Rheumatol. 2017;69(11):2124–35

Neumann F, Graef T, Tapprich C, Vaupel M, Steidl U, Germing U, et al. Cyclosporine A and Mycophenolate Mofetil vs Cyclosporine A and Methotrexate for graft-versus-host disease prophylaxis after stem cell transplantation from HLA-identical siblings. Bone Marrow Transplant. 2005;35(11):1089–93

Nikolich-Žugich J. The twilight of immunity: Emerging concepts in aging ofthe immune system. Nat. Immunol. 2018;19(1):10–9

Niwa-Kawakita M, Ferhi O, Soilihi H, Le Bras M, Lallemand-Breitenbach V, deThe H. PML is a ROS sensor activating p53 upon oxidative stress. J Exp Med. 2017;214(11):3197–206

136 Ogonek J, Kralj JM, Ghimire S, Varanasi P, Holler E, Greinix H, et al. Immune Reconstitution after Allogeneic Hematopoietic Stem Cell Transplantation. Front Immunol. 2016;(7):507

O’Keeffe M, Hochrein H, Vremec D, Pooley J, Evans R, Woulfe S, et al. Effects of administration of progenipoietin 1, Flt-3 ligand, granulocyte colony-stimulating factor, and pegylated granulocyte-macrophage colony-stimulating factor on dendritic cell subsets in mice. Blood. 2002;99(6):2122–30

Onai N, Kurabayashi K, Hosoi-Amaike M, Toyama-Sorimachi N, Matsushima K, Inaba K, et al. A Clonogenic Progenitor with Prominent Plasmacytoid Dendritic Cell Developmental Potential. Immunity. 2013;38(5):943–57

Onai N, Obata-Onai A, Schmid MA, Ohteki T, Jarrossay D, Manz MG. Identification of clonogenic common Flt3+M-CSFR+ plasmacytoid and conventional dendritic cell progenitors in mouse bone marrow. Nat. Immunol. 2007;8(11):1207–16

Ordemann R, Hutchinson R, Friedman J, Burakoff S, Reddy P, Duffner U, et al. Enhanced allostimulatory activity of host antigen-presenting cells in old mice intensifies acute graft-versus-host disease. J. Clin. Invest. 2002;109(9):1249–56

Pallotta MT, Orabona C, Volpi C, Vacca C, Belladonna ML, Bianchi R, et al. Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells. Nat. Immunol. 2011;12(9):870–8

Panda SK, Kolbeck R, Sanjuan MA. Plasmacytoid dendritic cells in autoimmunity. Curr. Opin. Immunol. Elsevier Ltd; 2017;44:20–5 Parolini S, Santoro A, Marcenaro E, Luini W, Massardi L, Facchetti F, et al. The role of chemerin in the colocalization of NK and dendritic cell subsets into inflamed tissues. Blood. 2007;109(9):3625–32

Pavletic SZ, Fowler DH. Are we making progress in GVHD prophylaxis and treatment? Hematol. / Educ. Progr. Am. Soc. Hematol. 2012;2012:251–64

Peccatori J, Forcina A, Clerici D, Crocchiolo R, Vago L, Stanghellini M, et al. Sirolimus-based graft-versus-host disease prophylaxis promotes the in vivo expansion of regulatory T cells and permits peripheral blood stem cell transplantation from haploidentical donors. Leukemia. 2015;29(2):396–405

Peled JU, Jenq RR, Holler E, van den Brink MRM. Role of gut flora after bone marrow transplantation. Nat. Microbiol. 2016;1(4):16036

Penack O, Holler E, Van Den Brink MRM. Graft-versus-host disease: Regulation by microbe-associated molecules and innate immune receptors. Blood. 2010;115(10):1865–72

137 Piñana JL, Valcárcel D, Fernández-Avilés F, Martino R, Rovira M, Barba P, etal. MTX or mycophenolate mofetil with CsA as GVHD prophylaxis after reduced-intensity conditioning PBSCT from HLA-identical siblings. Bone Marrow Transplant. 2010;45(9):1449–56

Ram R, Yeshurun M, Vidal L, Shpilberg O, Gafter-Gvili A. Mycophenolate mofetil vs. methotrexate for the prevention of graft-versus-host-disease - Systematic review and meta-analysis. Leuk. Res. 2014;38(3):352–60

Ransom J. Mechanism of action of mycophenolate mofetil. Ther Drug Monit. 1995;17(6):681–4

Ratanatharathorn V, Nash RA, Przepiorka D, Devine SM, Klein JL, Weisdorf D,etal. Phase III study comparing methotrexate and tacrolimus (prograf, FK506) with methotrexate and cyclosporine for graft-versus-host disease prophylaxis after HLA-identical sibling bone marrow transplantation. Blood. 1998;92(7):2303–14

Reizis B. Plasmacytoid Dendritic Cells: Development, Regulation, and Function. Immunity. 2019;50(1):37–50

Ricard L, Jachiet V, Malard F, Ye Y, Stocker N, Riviere S, et al. Circulating follicular helper T cells are increased in systemic sclerosis and promote plasmablast differentiation through the IL-21 pathway which can be inhibited by ruxolitinib. Ann Rheum Dis. 2019;78(4):539-550

Riesner K, Kalupa M, Shi Y, Elezkurtaj S, Penack O. A preclinical acute GVHD mouse model based on chemotherapy conditioning and MHC-matched transplantation. Bone Marrow Transplant. Nature Publishing Group; 2016;51(3):410–7

Rissoan MC, Soumelis V, Kadowaki N, Grouard G, Briere F, R. de WM, et al. Reciprocal control of T helper cell and dendritic cell differentiation. Science (80-. ). 1999;283(5405):1183–6

Robinson T, O’Donnell P, Fuchs E, Luznik L. Haploidentical bone marrow and stem cell transplantation: experience with post-transplantation cyclophosphamide. Semin Hematol. 2016;53(2):90–7

Rodrigues PF, Alberti-Servera L, Eremin A, Grajales-Reyes GE, Ivanek R, Tussiwand R. Distinct progenitor lineages contribute to the heterogeneity of plasmacytoid dendritic cells. Nat. Immunol. 2018;19(7):711–22

138 Rönnblom L, Alm G V. A Pivotal Role for the Natural Interferon α–producing Cells (Plasmacytoid Dendritic Cells) in the Pathogenesis of Lupus. J. Exp. Med. 2001;194(12):F59–64

Rousselot P, Labaume S, Marolleau JP, Larghero J, Noguera MH, Brouet JC, et al. Arsenic trioxide and melarsoprol induce apoptosis in plasma cell lines andinplasma cells from myeloma patients. Cancer Res. 1999;59(5):1041–8

RowlandS,RiggsJ,GilfillanS,Bugatti M, Vermi W, Kolbeck R, et al. Early, transient depletion of plasmacytoid dendritic cells ameliorates autoimmunity in a lupus model. J. Exp. Med. 2014;211(10):1977–91

Roy DC, Perreault C. Major vs minor histocompatibility antigens. Blood. 2017;129(6):664–6

Roychoudhuri R, Eil RL, Restifo NP. The interplay of effector and regulatory T cells in cancer. Curr Opin Immunol. 2015;33:101–11

Russo E, Teijeira A, Vaahtomeri K, Willrodt A-H, Bloch JS, Nitschké M, et al. Intralymphatic CCL21 Promotes Tissue Egress of Dendritic Cells through Afferent Lymphatic Vessels. Cell Rep. 2016;14(7):1723–34

Sabry W, Le Blanc R, Labbé AC, Sauvageau G, Couban S, Kiss T, et al. Graft-versus-Host Disease Prophylaxis with Tacrolimus and Mycophenolate Mofetil in HLA-Matched Nonmyeloablative Transplant Recipients Is Associated with Very Low Incidence of GVHD and Nonrelapse Mortality. Biol. Blood Marrow Transplant. 2009;15(8):919–29

Sadeghi B, Aghdami N, Hassan Z, Forouzanfar M, Rozell B, Abedi-Valugerdi M,et al. GVHD after chemotherapy conditioning in allogeneic transplanted mice. Bone Marrow Transplant. 2008;42(12):807–18

Sage AP, Murphy D, Maffia P, Masters LM, Sabir SR, Baker LL, et al. MHC Class II-restricted antigen presentation by plasmacytoid dendritic cells drives proatherogenic T cell immunity. Circulation. 2014;130(16):1363–73

Sahin U, Ferhi O, Jeanne M, Benhenda S, Berthier C, Jollivet F, et al. Oxidative stress-induced assembly of PML nuclear bodies controls sumoylation of partner proteins. J Cell Biol. 2014;204(6):931–45

Saitoh SI, Abe F, Kanno A, Tanimura N, Mori Saitoh Y, Fukui R, et al. TLR7 mediated viral recognition results in focal type i interferon secretion by dendritic cells. Nat. Commun. 2017;8(1):1592

139 Sakurai T, Ohta T, Fujiwara K. Inorganic arsenite alters macrophage generation from human peripheral blood monocytes. Toxicol Appl Pharmacol. 2005;203(2):145–53

Santoro N, Labopin M, Ciceri F, Van Lint MT, Nasso D, Blaise D, et al. Impact of conditioning intensity on outcomes of haploidentical stem cell transplantation for patients with acute myeloid leukemia over 45 years of age. Cancer. 2019; 125(9):1499-1506

Sasai M, Linehan M, Iwasaki A. Bifurcation of Toll-like receptor 9 signaling by adaptor protein 3. Science (80-. ). 2010;329(5998):1530–4

Satake A, Schmidt AM, Nomura S, Kambayashi T. Inhibition of calcineurin abrogates while inhibition of mTOR promotes regulatory T cell expansion and graft-versus-host disease protection by IL-2 in allogeneic bone marrow transplantation. PLoS One. 2014;9(3):1–8

Sathe P, Vremec D, Wu L, Corcoran L, Shortman K. Convergent differentiation: myeloid and lymphoid pathways to murine plasmacytoid dendritic cells. Blood. 2012. p. 11–9

Sato S, Fujimoto M, Hasegawa M, Takehara K. Altered blood B lymphocyte homeostasis in systemic sclerosis: expanded naive B cells and diminished but activated memory B cells. Arthritis Rheum. 2004;50(6):1918–27

Sawai CM, Babovic S, Upadhaya S, Knapp DJHF, Lavin Y, Lau CM, et al. Hematopoietic Stem Cells Are the Major Source of Multilineage Hematopoiesis in Adult Animals. Immunity. 2016;45(3):597–609

Sawai CM, Sisirak V, Ghosh HS, Hou EZ, Ceribelli M, Staudt LM, et al. Transcription factor Runx2 controls the development and migration of plasmacytoid dendritic cells. J. Exp. Med. 2013;210(11):2151–9

Sawai C, Serpas L, Neto A, Jang G, Rashidfarrokhi A, Kolbeck R, et al. Plasmacytoid Dendritic Cells Are Largely Dispensable for the Pathogenesis of Experimental Inflammatory Bowel Disease. Front. Immunol. 2018;9(October):1–16

Scherer M, Stamminger T. Emerging Role of PML Nuclear Bodies in Innate Immune Signaling. J Virol. 2016;90(13):5850– 4 Schieber M, Chandel NS. ROS function in redox signaling and oxidative stress. Curr Biol. 2014;24(10):R453-62

Schlitzer A, Heiseke AF, Einwachter H, Reindl W, Schiemann M, Manta C-P, et al. Tissue-specific differentiation of a circulating CCR9- pDC-like common dendritic cell precursor. Blood. 2012;119(25):6063–71

140 Schlitzer A, Loschko J, Mair K, Vogelmann R, Henkel L, Einwachter H, et al. Identification of CCR9- murine plasmacytoid DC precursors with plasticity to differentiate into conventional DCs. Blood. 2011;117(24):6562–70

Schmid M, Kingston D, Boddupalli S, Manz M. Instructive cytokine signals in dendritic cell lineage commitment. Immunol. Rev. 2010;234(1):32–44

Schroeder MA, DiPersio JF. Mouse models of graft-versus-host disease: advances and limitations. Dis. Model. Mech. 2011;4(3):318–33

See P, Dutertre CA, Chen J, Gunther P, McGovern N, Irac SE, et al. Mapping the human DC lineage through the integration of high-dimensional techniques. Science (80-. ). 2017;356(6342)

Sengupta M, Bishayi B. Effect of lead and arsenic on murine macrophage response. Drug Chem Toxicol. 2002;25(4):459–72

Seth S, Oberdorfer L, Hyde R, Hoff K, Thies V, Worbs T, et al. CCR7 Essentially Contributes to the Homing of Plasmacytoid Dendritic Cells to Lymph Nodes under Steady-State As Well As Inflammatory Conditions. J. Immunol. 2011;186(6):3364– 72

Shah MV, Saliba RM, Rondon G, Chen J, Soebbing D, Rus I, et al. Pilot study using post-transplant cyclophosphamide (PTCy), tacrolimus and mycophenolateGVHD prophylaxis for older patients receiving 10/10 HLA-matched unrelated donor hematopoietic stem cell transplantation. Bone Marrow Transplant. Springer US; 2019; 54(4):601-606

Shaw J, Wang YH, Ito T, Arima K, Liu YJ. Plasmacytoid dendritic cells regulate B-cell growth and differentiation via CD70. Blood. 2010;115(15):3051–7

Shen S, Li XF, Cullen WR, Weinfeld M, Le XC. Arsenic binding to proteins. Chem Rev. 2013;113(10):7769–92

Shen ZX, Chen GQ, Ni JH, Li XS, Xiong SM, Qiu QY, et al. Use of Arsenic Trioxide (As2O3) in the Treatment of Acute Promyelocytic Leukemia (APL) II. Clinical Efficacy and Pharmacokinetics in Relapsed Patients. Blood. 1997;89(9):3354–60

Shigematsu H, Reizis B, Iwasaki H, Mizuno SI, Hu D, Traver D, et al. Plasmacytoid dendritic cells activate lymphoid-specific genetic programs irrespective of their cellular origin. Immunity. 2004;21(1):43–53

141 Shim J, Kennedy RH, Weatherly LM, Hutchinson LM, Pelletier JH, Hashmi HN, etal. Arsenic inhibits mast cell degranulation via suppression of early tyrosine phosphorylation events. J Appl Toxicol. 2016;36(11):1446–59

Shinohara M, Lu L, Bu J, Werneck M, Kobayashi K, Glimcher L, et al. Osteopontin expression is essential for interferon- α production by plasmacytoid dendritic cells. Nat. Immunol. 2006;7(5):498–506

Shortman K, Sathe P, Vremec D, Naik S, O’Keeffe M. Plasmacytoid Dendritic Cell Development. 1st ed. Adv. Immunol. 2013; 120:105-26

Siegal FP, Kadowaki N, Shodell M, Fitzgerald-bocarsly PA, Shah K, Ho S. The Nature of the Principal Type 1 Interferon – Producing Cells in Human Blood. Science (80-. ). 1999;284(June):1835–8

Singer M, Trugnan G, Chelbi-Alix MK. Arsenic trioxide reduces 2,4,6-trinitrobenzene sulfonic acid-induced murine colitis via nuclear factor-kappaB down-regulation and caspase-3 activation. Innate Immun. 2011;17(4):365–74

Singh N, Kumar D, Lal K, Raisuddin S, Sahu AP. Adverse health effects due to arsenic exposure: modification by dietary supplementation of jaggery in mice. Toxicol Appl Pharmacol. 2010;242(3):247–55

Sisirak V, Vey N, Vanbervliet B, Duhen T, Puisieux I, Lira SA, et al. CCR6/CCR10-mediated plasmacytoid dendritic cell recruitment to inflamed epithelia after instruction in lymphoid tissues. Blood. 2011;118(19):5130–40

Sisirak V, Ganguly D, Lewis K, Couillault C, Tanaka L, Bolland S, et al. Genetic evidence for the role of plasmacytoid dendritic cells in systemic lupus erythematosus. J. Exp. Med. 2014;211(10):1969–76

Sorror ML, Martin PJ, Storb RF, Bhatia S, Maziarz RT, Pulsipher MA, et al. Pretransplant comorbidities predict severity of acute graft-versus-host disease and subsequent mortality. Blood. 2014;124(2):287–95

Soto-Pena GA, Vega L. Arsenic interferes with the signaling transduction pathway of T cell receptor activation by increasing basal and induced phosphorylation of Lck and Fyn in spleen cells. Toxicol Appl Pharmacol. 2008;230(2):216–26

Souers A, Leverson J, Boghaert E, Ackler S, Catron N, Chen J, et al. ABT-199, a potent and selective BCL-2 inhibitor, achieves antitumor activity while sparing platelets. Nat. Med. 2013;19(2):202–8

142 Sozzani S, Vermi W, Del Prete A, Facchetti F. Trafficking properties of plasmacytoid dendritic cells in health and disease. Trends Immunol. 2010;31(7):270–7

Spierings E. Minor histocompatibility antigens: Past, present, and future. Tissue Antigens. 2014;84(4):374–360

Sprent J, Schaefer M, Gao E, Korngold R. Role of T cell subsets in lethal graft-versus-host disease (GVHD) directed to class I versus class II H-2 differences. I. L3T4+ cells can either augment or retard GVHD elicited by Lyt-2+ cells in class I different hosts. J. Exp. Med. 1988;167(2):556–69

Srivastava RK, Li C, Chaudhary SC, Ballestas ME, Elmets CA, Robbins DJ, et al. Unfolded protein response (UPR) signaling regulates arsenic trioxide-mediated macrophage innate immune function disruption. Toxicol Appl Pharmacol. 2013;272(3):879–87

Srivastava RK, Li C, Wang Y, Weng Z, Elmets CA, Harrod KS, et al. Activating transcription factor 4 underlies the pathogenesis of arsenic trioxide-mediated impairment of macrophage innate immune functions. Toxicol Appl Pharmacol. 2016;308:46–58

Srivatsan S, Swiecki M, Otero K, Cella M, Shaw AS. CD2-Associated Protein Regulates Plasmacytoid Dendritic Cell Migration, but Is Dispensable for Their Development and Cytokine Production. J. Immunol. 2013;191(12):5933–40

Stolfi JL, Pai CCS, Murphy WJ. Preclinical modeling of hematopoietic stem cell transplantation - Advantages and limitations. FEBS J. 2016;283(9):1595–606

Storb R, Raff R, Appelbaum F, Deeg H, Fitzsimmons W, Graham T, et al. FK-506 and methotrexate prevent graft-versus-host disease in dogs given 9.2 Gy total body irradiation and marrow grafts from unrelated dog leukocyte antigen-nonidentical donors. Transplantation. 1993;56(4):800–7

Storb R, Deeg HJ, Whitehead J, Appelbaum F, Beatty P, Bensinger W, et al. Methotrexate and Cyclosporine Compared with Cyclosporine Alone for Prophylaxis of Acute Graft versus Host Disease after Marrow Transplantation for Leukemia. N. Engl. J. Med. 1986;314(12):729–35

Styczynski J, Tridello G, Gil L, Ljungman P, Hoek J, Iacobelli S, et al. Impact of Donor Epstein-Barr Virus Serostatus on the Incidence of Graft-Versus-Host Disease in Patients With Acute Leukemia After Hematopoietic Stem-Cell Transplantation: A Study From the Acute Leukemia and Infectious Diseases Working Parties of the Europ. J. Clin. Oncol. 2016;34(19):2212–20

143 Sun L, Wu J, Du F, Chen X, J CZ. Cyclic GMP-AMP synthase is a cytosolic DNA sensor that activates the type I interferon pathway. Science (80-. ). 2013;339(6121):786–91

Swain SL, McKinstry KK, Strutt TM. Expanding roles for CD4(+) T cells in immunity to viruses. Nat Rev Immunol. 2012;12(2):136–48

Swiecki M, Miller HL, Sesti-Costa R, Cella M, Gilfillan S, Colonna M. Microbiota induces tonic CCL2 systemic levels that control pDC trafficking in steady state. Mucosal Immunol. 2017;10(4):936–45

Swiecki M, Colonna M. The multifaceted biology of plasmacytoid dendritic cells. Nat. Rev. Immunol. 2015;15(8):471–85

Takaoka A, Yanai H, Kondo S, Duncan G, Negishi H, Mizutani T, et al. Integral role of IRF-5 in the gene induction programme activated by Toll-like receptors.Nature. 2005;434(7030):243–9

Tenorio EP, Saavedra R. Differential effect of sodium arsenite during the activation of human CD4+ and CD8+ T lymphocytes. Int Immunopharmacol. 2005;5(13 – 14):1853–69

Thomas-Schoemann A, Batteux F, Mongaret C, Nicco C, Chereau C, Annereau M, et al. Arsenic trioxide exerts antitumor activity through regulatory T cell depletion mediated by oxidative stress in a murine model of colon cancer. J Immunol. 2012;189(11):5171–7

Tohyama N, Tanaka S, Onda K, Sugiyama K, Hirano T. Influence of anticancer agents on cell survival, proliferation, and CD4+CD25+Foxp3+ regulatory T cell-frequency in human peripheral-blood mononuclear cells activated byT cell-mitogen. Int Immunopharmacol. 2013;15(1):160–6

Tomasello E, Naciri K, Chelbi R, Bessou G, Fries A, Gressier E, et al. Molecular dissection of plasmacytoid dendritic cell activation in vivo during a viral infection. EMBO J. 2018;37(19):e98836

Tucci M, Quatraro C, Lombardi L, Pellegrino C, Dammacco F, Silvestris F. Glomerular accumulation of plasmacytoid dendritic cells in active lupus nephritis: Role of interleukin-18. Arthritis Rheum. 2008;58(1):251–62

Umemoto E, Otani K, Ikeno T, Verjan Garcia N, Hayasaka H, Bai Z, et al. Constitutive Plasmacytoid Dendritic Cell Migration to the Splenic White Pulp Is Cooperatively Regulated by CCR7- and CXCR4-Mediated Signaling. J. Immunol. 2012;189(1):191–9

144 Upadhaya S, Sawai C, Papalexi E, Rashidfarrokhi A, Jang G, Chattopadhyay P,etal. Kinetics of adult hematopoietic stem cell differentiation in vivo. J. Exp.Med. 2018;215(11):2815–32

Vakkila J, Thomson AW, Hovi L, Vettenranta K, Saarinen-Pihkala UM. Circulating dendritic cell subset levels after allogeneic stem cell transplantation in children correlate with time post transplant and severity of acute graft-versus-host disease. Bone Marrow Transplant. 2005;35(5):501–7 van Bekkum D, Knaan S. Role of bacterial microflora in development of intestinal lesions from graft-versus-host reaction. J Nat Cancer Inst. 1977;58(3):787–90 van Bon L, Affandi AJ, Broen J, Christmann RB, Marijnissen RJ, Stawski L, et al. Proteome-wide analysis and CXCL4 as a biomarker in systemic sclerosis. N Engl J Med. 2014;370(5):433–43

Van Lint MT, Uderzo C, Locasciulli A, Majolino I, Scimé R, Locatelli F, et al.Early treatment of acute graft-versus-host disease with high- or low-dose 6-methylprednisolone: a multicenter randomized trial from the Italian Group for Bone Marrow Transplantation. Blood. 1998;92(7):2288–93

VanDenBerg KR, Freeborn RA, Liu S, Kennedy RC, Zagorski JW, Rockwell CE. Inhibition of early T cell cytokine production by arsenic trioxide occurs independently of Nrf2. PLoS One. 2017;12(10):e0185579

Vega L, Montes de Oca P, Saavedra R, Ostrosky-Wegman P. Helper T cell subpopulations from women are more susceptible to the toxic effect of sodium arsenite in vitro. Toxicology. 2004;199(2–3):121–8

Vega L, Ostrosky-Wegman P, Fortoul TI, Diaz C, Madrid V, Saavedra R. Sodium arsenite reduces proliferation of human activated T-cells by inhibition of the secretion of interleukin-2. Immunopharmacol Immunotoxicol. 1999;21(2):203–20

Vermi W, Riboldi E, Wittamer V, Gentili F, Luini W, Marrelli S, et al. Role of ChemR23 in directing the migration of myeloid and plasmacytoid dendritic cells to lymphoid organs and inflamed skin. J. Exp. Med. 2005;201(4):509–15

Villani AC, Satija R, Reynolds G, Sarkizova S, Shekhar K, Fletcher J, et al. Single-cell RNA-seq reveals new types of human blood dendritic cells, monocytes, and progenitors. Science (80-. ). 2017;356(6335)

145 Vollmer J, Weeratna R, Payette P, Jurk M, Schetter C, Laucht M, et al. Characterization of three CpG oligodeoxynucleotide classes with distinct immunostimulatory activities. Eur. J. Immunol. 2004;34(1):251–62

WallerE,LoganB,HarrisW,DevineS,PorterD,MineishiS,etal.Improved Survival After Transplantation of More Donor Plasmacytoid Dendritic or Naïve T Cells From Unrelated-Donor Marrow Grafts: Results From BMTCTN 0201. J. Clin. Oncol. 2014;32(22):2365–72

Wan YY, Flavell RA. How diverse--CD4 effector T cells and their functions. JMol Cell Biol. 2009;1(1):20 – 36 Wang L, Hu X, Xu Y, Liu Z. Arsenic trioxide inhibits lung metastasis of mouse colon cancer via reducing the infiltration of regulatory T cells. Tumour Biol. 2016;37(11):15165–73

Wang L, Wang R, Fan L, Liang W, Liang K, Xu Y, et al. Arsenic trioxide is an immune adjuvant in liver cancer treatment. Mol Immunol. 2017;81:118–26

Wang ZY, Chen Z. Acute promyelocytic leukemia from highly fatal to highly curable. Blood. 2008;111(5):2505–15

Waskow C, Liu K, Darrasse-Jèze G, Guermonprez P, Ginhoux F, Merad M, et al. The receptor tyrosine kinase Flt3 is required for dendritic cell development in peripheral lymphoid tissues. Nat. Immunol. 2008;9(6):676–83

Wei Z, Zhang X, Wang J, Wang Y, Yang Z, Fu Y. The formation of canine neutrophil extracellular traps induced by sodium arsenic in polymorphonuclear neutrophils. Chemosphere. 2018;196:297–302

Wendland M, Czeloth N, Mach N, Malissen B, Kremmer E, Pabst O, et al. CCR9 is a homing receptor for plasmacytoid dendritic cells to the small intestine. Proc. Natl. Acad. Sci. 2007;104(15):6347–52

Whangbo J, Ritz J, Bhatt A. Antibiotic-mediated modification of the intestinal microbiome in allogeneic hematopoietic stem cell transplantation. Bone Marrow Transplant. 2017;52(2):183–90

Wikstrom ME, Fleming P, Kuns RD, Schuster IS, Voigt V, Miller G, et al. Acute GVHD results in a severe DC defect that prevents T-cell priming and leads to fulminant cytomegalovirus disease in mice. Blood. 2015;126(12):1503–14

Wimmers F, Subedi N, van Buuringen N, Heister D, Vivié J, Beeren-Reinieren I, et al. Single-cell analysis reveals that stochasticity and paracrine signaling control interferon-alpha production by plasmacytoid dendritic cells. Nat. Commun. 2018;9(1):1–12

146 Wohn C, Ober-Blöbaum J, Haak S, Pantelyushin S, Cheong C, Zahner S, et al. Langerinneg conventional dendritic cells produce IL-23 to drive psoriatic plaque formation in mice. Proc. Natl. Acad. Sci. 2013;110(26):10723–8

Wolfle SJ, Strebovsky J, Bartz H, Sahr A, Arnold C, Kaiser C, et al. PD-L1 expression on tolerogenic APCs is controlled by STAT-3. Eur J Immunol. 2011;41(2):413–24

Wurbel MA, McIntire MG, Dwyer P, Fiebiger E. CCL25/CCR9 interactions regulate large intestinal inflammation in a murine model of acute colitis. PLoS One. 2011;6(1)

Xiao B, Wang Y, Li W, Baker M, Guo J, Corbet K, et al. Plasma microRNA signature as a noninvasive biomarker for acute graft-versus-host disease. Blood. 2013;122(19):3365–75

Xu S, Chen J, Wang F, Kang X, Lan T, Wang F, et al. Arsenic trioxide combined with co-stimulatory molecule blockade prolongs survival of cardiac allografts in alloantigen-primed mice. Transpl Immunol. 2010;24(1):57–63

Xu W, Li X, Quan L, Yao J, Mu G, Guo J, et al. Arsenic trioxide decreases the amount and inhibits the function of regulatory T cells, which may contribute to its efficacy in the treatment of acute promyelocytic leukemia. Leuk Lymphoma. 2018;59(3):650–9

Yan G, Xi Y, Xu S, Lin Y, Chen J, Dai H, et al. Inhibition of accelerated rejection mediated by alloreactive CD4(+) memory T cells and prolonged allograft survival by arsenic trioxide. Immunol Invest. 2013;42(5):438–54

Yan S, Zhang QY, Zhou B, Xue L, Chen H, Wang Y, et al. Arsenic trioxide attenuated the rejection of major histocompatibility complex fully-mismatched cardiac allografts in mice. Transpl. Proc. 2009;41(5):1855–8

Yu C-F, Peng W-M, Oldenburg J, Hoch J, Bieber T, Limmer A, et al. Human Plasmacytoid Dendritic Cells Support Th17 Cell Effector Function in Response to TLR7 Ligation. J. Immunol. 2010;184(3):1159–67

Yu H, Tian Y, Wang Y, Mineishi S, Zhang Y. Dendritic Cell Regulation of Graft-Vs.-Host Disease: Immunostimulation and Tolerance. Front. Immunol. 2019;10(February):14–7

Yu H-S, Chen G-S, Liao W-T, Chang K-L, Yu C-L. Arsenic Induces Tumor Necrosis Factor α Release and Tumor Necrosis Factor Receptor 1 Signaling in T Helper Cell Apoptosis. J. Invest. Dermatol. Elsevier Masson SAS; 2002;119(4):812–9

147 Yun TJ, Lee JS, Machmach K, Shim D, Choi J, Wi YJ, et al. Indoleamine 2,3-Dioxygenase-Expressing Aortic Plasmacytoid Dendritic Cells Protect against Atherosclerosis by Induction of Regulatory T Cells. Cell Metab. 2016;23(5):852–66

Zeiser R, Blazar BR. Preclinical models of acute and chronic graft-versus-host disease: How predictive are they for a successful clinical translation? Blood. 2016;127(25):3117–26

Zhan Y, Carrington EM, Ko HJ, Vikstrom IB, Oon S, Zhang JG, et al. Bcl-2 antagonists kill plasmacytoid dendritic cells from lupus-prone mice and dampen interferon-α production. Arthritis Rheumatol. 2015;67(3):797–808

Zhang M, Qi Y, Li H, Cui J, Dai L, Frank JA, et al. AIM2 inflammasome mediates Arsenic-induced secretion of IL-1 beta and IL-18. Oncoimmunology. 2016;5(6):e1160182

Zhang H, Gregorio JD, Iwahori T, Zhang X, Choi O, Tolentino LL, et al. A distinct subset of plasmacytoid dendritic cells induces activation and differentiation of B and T lymphocytes. Proc. Natl. Acad. Sci. 2017a;114(8):1988–1993

Zhang Q, Lenardo MJ, Baltimore D. 30 Years of NF-κB: A Blossoming of Relevance to Human Pathobiology. Cell. 2017b;168(1–2):37–57

Zhao JJ, Yang SW, Zhang F, Yuan XL, ZM. Z, Fang BJ, et al. As2O3 Up-Regulates the Proportion of CD4+CD25+CD127low Tregs in Peripheral Blood of Patients with Severe Aplastic Anemia. Zhongguo Shi Yan Xue Ye Xue Za Zhi. 2018a;26(3):854– 858

Zhao B, Xia JJ, Wang LM, Gao C, Li JL, Liu JY, et al. Immunosuppressive effect of arsenic trioxide on islet xenotransplantation prolongs xenograft survival in mice. Cell Death Dis. 2018b;9(3):408

Zhou LF, Zhu Y, Cui XF, Xie WP, Hu AH, Yin KS. Arsenic trioxide, a potent inhibitor of NF-kappaB, abrogates allergen-induced airway hyperresponsiveness and inflammation. Respir Res. 2006;7:146

Zou W, Machelon V, Coulomb-L ’ Hermin A, Borvakz J, Nome F, Isaeva T, et al. Stromal-derived factor-1 in human tumors recruits and alters the functionof plasmacytoid precursor dendritic cells. Nat. Med. 2001;7(12):1339–46

148 6.Annex

Article 2

Circulating follicular helper T cellsareincreasedinsystemic

sclerosis and promote plasmablast differentiation through the IL-21

pathway which can be inhibited by ruxolitinib

Laure Ricard, Vincent Jachiet, Florent Malard, Yishan Ye, Nicolas Stocker, Sébastien Rivière, Patricia Senet, Jean-Benoit Monfort, Olivier Fain, Mohamad Mohty, Béatrice Gaugler, Arsène Mekinian

Published on Annals of the Rheumatic Diseases 2019 Apr;78(4):539-550

149 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from Epidemiological science Circulating follicular helper T cells are increased in systemic sclerosis and promote plasmablast differentiation through the IL-21 pathway which can be inhibited by ruxolitinib Laure Ricard,1,2 Vincent Jachiet,1,2 Florent Malard,1,3 Yishan Ye,1 Nicolas Stocker,1 Sébastien Rivière,2 Patricia Senet,4 Jean-Benoit Monfort,4 Olivier Fain,2 Mohamad Mohty,1,3 Béatrice Gaugler,1,3 Arsène Mekinian1,2

Handling editor Josef S Abstract Key messages Smolen Objectives systemic sclerosis (SSc) is an autoimmune 1 disease characterised by widespread fibrosis, Centre de Recherche Saint- What is already known on this topic? Antoine (CRSA), Sorbonne microangiopathy and autoantibodies. Follicular helper + + + + ►► CD4 Tcells with a Tfh phenotype can infiltrate Université, INSERM U938, Paris, T (Tfh) cells CD4 CXCR5 PD-1 cooperate with B the skin of SSc patients, and induce in vitro France lymphocytes to induce the differentiation of plasmocytes 2 myofibroblast differentiation. Service de Médecine Interne secreting immunoglobulins (Ig). Circulating Tfh (cTfh) et de l’Inflammation D( HU i2B), cells are increased in several autoimmune diseases. AP-HP, Hôpital Saint-Antoine, What this study adds? Paris, France However, there are no data about cTfh cells and their 3 ► Circulating Tfh cells are increased in SSc and Service d’Hématologie Clinique interaction with B cells in SSc. The aim of this study was ► correlate with SSc severity. et Thérapie Cellulaire, AP-HP, to perform a quantitative and functional analysis of cTfh Hôpital Saint-Antoine, Paris, ► Tfh cells from patients with SSc induce B-cell cells in SSc. ► France differentiation into plasmablasts secreting Ig 4 Methods Using flow cytometry, we analysed cTfh cells Service de Dermatologie, AP- via IL-21 secretion with increased capacity than HP, Hôpital Tenon, Paris, France from 50 patients with SSc and 32 healthy controls (HC). healthy controls. In vitro coculture experiments of sorted cTfh and B cells ► IL-21 or JAK1/2 blockade by ruxolitinib could be Correspondence to were performed for functional analysis. IgG and IgM ► a promising strategy in the treatment of SSc. Dr Arsène Mekinian, AP-HP, production were measured by ELISA. Hôpital Saint-Antoine, Service Results We observed that cTfh cell numbers are de Médecine Interne et de How might this impact on clinical practice or increased in patients with SSc compared with HC. l’Inflammation-(DHU i2B), future developments ? Université Paris, Paris F-75012, Furthermore, the increase in cTfh cells was more potent ► Clinical data with anti-JAK inhibition could be France; in patients with severe forms of SSc such as diffuse SSc ► interessting in SSc patients arsene.​ ​mekinian@aphp.​ ​fr and in the presence of arterial pulmonary hypertension. BéaG and AènM contributed cTfh cells from patients with SSc present an activated

equally. Tfh phenotype, with high expression of BCL-6, increased http://ard.bmj.com/ capacity to produce IL-21 in comparison with healthy diffuse SSc (dSSc) with skin involvement proximal Received 4 September 2018 controls. In vitro, cTfh cells from patients with SSc Revised 18 December 2018 to elbows and knees. dSSc is a rapidly progressing had higher capacity to stimulate the differentiation of Accepted 19 December 2018 + + hi disorder with more frequent visceral involvement Published Online First CD19 CD27 CD38 B cells and their secretion of IgG and increased mortality rates.2 The exact patho- 13 February 2019 and IgM through the IL-21 pathway than Tfh cells from physiology of the disease is largely unknown and healthy controls. Blocking IL-21R or using the JAK1/2 includes endothelial cell impairment, fibrosis and inhibitor ruxolitinib reduced the Tfh cells’ capacity on 12 April 2019 by guest. Protected copyright. immune dysfunction. The CD4+ T cells are a major to stimulate the plasmablasts and decreased the Ig component of the infiltrate in the skin in the acute production. 3 Conclusions circulating Tfh cells are increased in SSc inflammatory stage of the disease. Peripheral T and correlate with SSc severity. The IL-21 pathway or lymphocytes have increased expression of activa- JAK1/2 blockade by ruxolitinib could be a promising tion markers and secrete proinflammatory cyto- 4 strategy in the treatment of SSc. kines that activate fibroblasts and plasma cells. B lymphocytes are also reported to play a predom- inant role in SSc: autoantibodies are usually detected in the sera of patients with SSc5 and B-cell © Author(s) (or their homeostasis is disturbed in SSc. Despite increased employer(s)) 2019. No Introduction Systemic sclerosis (SSc) is an autoimmune disease expression of activation markers (CD80, CD86 and commercial re-use. See rights 6 and permissions. Published characterised by fibrosis of the skin and other CD95), SSc B lymphocytes secrete higher levels 7 by BMJ. organs, vascular impairment and deficient immune of the profibrotic cytokines IL-6 and TGF-β. The 1 To cite: Ricard L, responses. Patients with SSc are classified into two levels of the cytokine B-cell activating factor and Jachiet V, Malard F, groups depending on the extent of the skin involve- a proliferation-inducing ligand implicated in B-cell et al. Ann Rheum Dis ment: the limited cutaneous SSc (lSSc) with skin survival are also increased in SSc serum and appear 2019;78:539–550. involvement limited to face and hands, and the to be correlated with disease severity.8–10

Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 539 150 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from

Follicular helper T (Tfh) cells are a subset of CD4+ T cells Table 1 Patients’ characteristics able to provide help to B cells to undergo proliferation, isotype switch and somatic hypermutation, resulting in long-lasting Patients with Patients with Patients with antibody responses.11 12 They are localised next to the germinal SSc dSSc lSSc Characteristics (n=52) (n=15) (n=37) centre (GC) in secondary lymphoid organs and are characterised by their surface expression of C-X-C chemokine receptor type Age (years) 61 (32–81) 59.5 (32–75) 61 (32–81) 5 (CXCR5), inducible costimulatory (ICOS) and programmed Age at first non-Raynaud 48 (26–72) 42 (27–67) 49 (26–72) symptom (years) death cell protein 1 (PD-1), by the expression of the tran- Female sex n (%) 42 (81) 12 (73) 31 (84) scription factor B-cell lymphoma 6 (BCL-6) and the secretion of IL-21 and IL-4 cytokines.13–15 A circulating Tfh population Disease duration (years) 12 (0–27) 12 (0–24) 10 (0–27) (cTfh) has also been identified, which expresses CXCR5, PD-1 European ethnicity n (%) 31 (64) 8 (53) 23 (69) and ICOS and can help B-cell differentiation into plasma cells Skin involvement via IL-21 secretion.16 Increased frequencies of cTfh have been Diffuse systemic sclerosis 15 (29) – – n (%) reported in several autoimmune diseases such as rheumatoid arthritis, juvenile dermatomyositis, Sjögren’s syndrome and Limited systemic sclerosis 37 (71) – – n (%) systemic lupus erythematosus.16–19 Several subtypes of cTfh have Rodnan score 10 (0–36) 15 (10–36) 8.5 (0–27) been described whose frequencies could be disturbed in autoim- Active digital ulcers n (%) 8 (16) 4 (27) 4 (11) mune diseases.16 Recently, a pathologically expanded population Pulmonary involvement of CXCR5−PD-1+CD4+ T cells called T peripheral helper cells has been identified in the synovium of patients with rheumatoid Interstitial lung disease n (%) 21 (40) 11 (73) 10 (27) arthritis which could also promote plasma cell differentiation.20 Pulmonary arterial 5 (10) 3 (20) 2 (5) hypertension n (%) Furthermore, CD4+ T cells with a Tfh phenotype can infiltrate FVC (% of the predicted 103 (37–154) 75 (37–111) 106 (38–154) the skin of patients with SSc and induce in vitro myofibroblast value) differentiation.21 Thus, Tfh may play a predominant role in the DLCO (% of the predicted 61 (14–98) 55 (23–82) 62 (14–98) pathophysiology of SSc disease and promote B-cell stimulation value) and activation favouring increased persistent inflammation. We Other organ involvement performed a quantitative and functional analysis of the cTfh Joint involvement n (%) 6 (12%) 1 (7) 5 (14) in patients with SSc in comparison with healthy controls. We Kidney sclerosis crisis n (%) 1 (2%) 1 (7) 0 (0) demonstrate that cTfh are increased in SSc and correlated with Heart impairment n (%) 1 (2%) 1 (7) 0 (0) disease severity. Moreover, cTfh from patients with SSc are more Laboratory data efficient than healthy controls to induce B-cell differentiation BNP (mg/L) 34 (10–300) 39 (18–134) 34 (10–300) into plasma cells secreting IgG via the secretion of IL-21 cyto- C reactive protein (mg/L) 1 (1–50) 8.9 (1–50) 1 (1–16) kine. Ruxolitinib, a JAK1/2 inhibitor, significantly reduced in Autoantibodies vitro cTfh capacity to induce plasma cells and their IL-21 secre- Anti-centromeres n (%) 26 (50) 0 (0) 25 (71) tion, suggesting that the JAK/STAT pathway blockade could be a potential therapeutic approach in SSc. Anti-Scl70 n (%) 11 (22) 5 (33) 7 (20) Anti-RNAPol3 n (%) 5 (10) 5 (33) 0 Values are medians with ranges and frequencies with percentages. Patients and methods Anti-RNAPol3, anti-RNA polymerase 3 antibody; Anti-Scl70, anti-topoisomerase http://ard.bmj.com/ Fifty-two consecutive patients with SSc (median age 61 years 1 antibody; BNP, Brain natriuretic peptide; DLCO, Diffusing capacity for carbon monoxide; FVC, Forced vital capacity. (range, 32–81), 80% women) and 38 healthy controls (median age 53 years (range, 23–64), 46% women) were prospectively included in the study. All patients with SSc enrolled in this Biological samples study were followed at Saint-Antoine Hospital. All patients Blood samples were collected in EDTA tubes (BD Biosciences, fulfilled the 2013 American College of Rheumatology/Euro- Le Pont de Claix, France). Healthy controls (HC) were obtained 22 pean League Against Rheumatism criteria for SSc. Patients from healthy donors (‘Etablissement Français du Sang, Paris on 12 April 2019 by guest. Protected copyright. with an associated autoimmune systemic disease, concomitant Saint-Antoine-Crozatier’). Peripheral blood mononuclear cells infectious disease or active neoplasm were excluded. None of (PBMCs) were isolated with a standard gradient centrifugation the patients with SSc were receiving treatment with steroids or procedure on a lymphocyte separation medium (Lymphosep other immunosuppressive therapy at the time of the analysis. separation media; Dutscher, Issy-les-Moulineaux, France). For each patient, the following data were analysed: age, gender, disease duration (from the date of the first non-Raynaud’s Phenotype analysis by flow cytometry phenomenon), type of SSc (dSSc or lSSc), presence of active After isolation, PBMCs were stained with the following fluoro- digital ulcers, presence of joint, heart, gastrointestinal and lung chrome-conjugated antibodies: CXCR3 (CD183) FITC, CCR6 involvement, modified Rodnan skin score (mRss) and presence PE-Vio770 (Miltenyi Biotec, Paris, France); CXCR5 (CD185) of pulmonary arterial hypertension. The lung involvement was PE (eBioscience, ThermoFisher, Villebon, France); CD45RA defined as the presence of interstitial lung disease on high-res- ECD, PD-1 (CD279) PECY5.5, ICOS (CD278) APC, CD3 olution CT scan. Laboratory data included C reactive protein AA750 (Beckman Coulter, Villepinte, France); CD45 BV510, level, total lymphocyte count, plasma creatinine, urea, creati- CD4 BV650, HLA-DR BV786 (BD Biosciences, Rungis, France). nine phosphokinase enzymes and antinuclear autoantibodies For B-lymphocyte analysis, the following fluorochrome-conju- (anti-centromeres, anti-topoisomerase I, anti-PM-Scl, anti-RNA gated antibodies were used: IgD FITC, CD21 BV450, CD45 polymerase III autoantibodies). Characteristics of patients with BV510, IL-21R APC (BD Biosciences); CD27 PE, CD24 ECD, SSc are depicted in table 1. CD38 PECY5.5, CD19 PC7 (Beckman Coulter); CD5 AA700

540 Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 151 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from

(Biolegend, Ozyme, Montigny-le-Bretonneux). T-lymphocyte (Hs00757930_m1), BLIMP-1 (Hs00757930_m1). Real-Time and B-lymphocyte absolute numbers were determined using PCR were performed on a 7500 Fast Dx Real-Time PCR Instru- 50 µL of whole blood from patients and HC using Trucount ment (ThermoFisher Scientific). Relative expression for the tubes (BD Biosciences). Samples were stained with the following mRNA transcripts was calculated using the ∆∆Ct method and antibodies: CD19 FITC, CD45 ECD, CD3 AA750 (Beckman GAPDH mRNA transcript as reference. Coulter) for 15 min and then incubated for 15 min with 450 µL BD FACS Lysis Solution and subsequently analysed by flow Statistical analysis cytometry. Compensation beads were used for compensation Data are expressed as means±SD, medians with ranges and setting (VersaComp; Beckman Coulter). Cells were analysed on numbers with their frequencies. The χ2 test or Fisher test was a Cytoflex flow cytometer (Beckman Coulter). Data were anal- used to compare qualitative values according to distribution and ysed using Kaluza V.5.1 software (Beckman Coulter). the Student t-test, Mann-Whitney U test or Wilcoxon test for continuous quantitative variables. The Pearson test was used to Cell sorting determine the correlation between variables. All analyses were Total CD4+ T cells were isolated from PBMCs using a Magnisort performed using GraphPad Prism V.5.0 (GraphPad Software, Human CD4 T-cell negative selection kit (Invitrogen, Ther- San Diego, California, USA). A p value <0.05 was considered as moFisher). T-cell purity was always more than 94%. CD4+ T statistically significant. cells were then stained with the following fluorochrome-con- jugated antibodies: CXCR5 (CD185) PE (eBioscience), CD4 Results Pacific Blue, PD-1 APC (Beckman Coulter). Tfh cells were Circulating Tfh cells are increased in SSc and express high defined as CD4+CXCR5+PD-1+ cells. Tfh cells, CD4+CX- levels of PD-1 and activation markers CR5−PD-1+ and CD4+CXCR5−PD-1− were sorted using a flow We analysed circulating Tfh (cTfh) cells in the blood of 50 cytometer (FACS Aria III; BD Biosciences). B lymphocytes were patients with SSc and 32 healthy controls by flow cytometry. The isolated from PBMCs by a magnetic beads CD19+ positive selec- cTfh cells were defined as CD4+CXCR5+PD-1+ cells among tion kit (eBioscience). All sorted cell populations exhibited high CD4+ T cells. The gating strategy used to identify cTfh cells is purity (>90%). represented in figure 1A. The percentage and absolute numbers of cTfh cells among CD4+ T cells were significantly increased Stimulation of Tfh cells in patients with SSc as compared with HC (mean 7.2±3.6% vs + Sorted Tfh cells or non-Tfh CD4 T cells from nine patients 4.4±1.5% (p<0.0001) and 69.1±38.9 cells/µL vs 33.2±19.2 with SSC were cultured at 50 000 cells per well in 200 µL of cells/µL (p<0.0001), respectively) (figure 1B,C). RPMI 1640 medium (Eurobio, Courtaboeuf, France) supple- We also evaluated the CXCR5−PD-1+CD4+ T cells and we mented with 10% fetal bovine serum (Gibco, ThermoFisher) observed that their number was also significantly increased in and with phorbol 12-myristate 13-acetate (PMA, 50 ng/mL) and patients with SSc in comparison with HC: 128±79.4 cells/µL ionomycin (1 µg/mL) (Sigma-Aldrich, St. Louis, Missouri, USA). versus 77±60 cells/µL; p=0.001 (figure 1D). We then analysed Supernatants were then collected after 36 hours of culture and the three major subsets of Tfh cells in human blood as previously IL-21 levels were measured by ELISA (IL-21 Human uncoated described by Morita et al,16 namely Tfh1, Tfh2 and Tfh17 subsets ELISA kit; ThermoFisher). which are characterised by their different expression of CXCR3 and CCR6 (figure 1A). As shown in figure 1E, the percentages Coculture of T and B cells of Tfh1 cells CXCR3+CCR6−, Tfh2 cells CXCR3−CCR6− and + + − +

Sorted Tfh cells, non-Tfh CD4 T cells and purified CD19 Tfh17 CXCR3 CCR6 cells did not differ in patients with SSc http://ard.bmj.com/ B cells were cocultured at 50 000 cells/well in 96-well plates in comparison with healthy controls. (Starstedt) at a 1:1 ratio in 200 µL of TexMACS medium PD-1 signalling on Tfh cells promotes the selection and the (Miltenyi Biotec) supplemented with gentamycine (Gibco) and survival of GC B cells, and regulates formation of long-lived 0.1 µg/mL SEB (Toxin Technology, Sarasota, Florida, USA) from plasma cells via interaction with the ligand PD-L1 on B cells. 13 patients with SSc (3 dSSc and 10 lSSc) and 6 HC. Super- We evaluated the level of expression of PD-1 on Tfh cells from natants were collected after 6 days of culture. For blocking patients with SSc. High expression of PD-1 (PD-1high) or inter- int experiments, 10 µg/mL recombinant Human IL-21 Fc Chimaera mediate expression (PD-1 ) by Tfh cells were defined according on 12 April 2019 by guest. Protected copyright. Protein (IL-21RFc) (R&D Systems, Lille, France), or 10 µg/mL to the PD-1 intensity as shown in figure 1F. We observed that anti-IL6R-IgG1 (Roactemra, tociluzimab; Roche), or 0.1 µg/mL the frequencies of CD4+CXCR5+PD-1high cells were signifi- ruxolitinib (InvivoGen, San Diego, California, USA) were added cantly increased in patients with SSc as compared with healthy in the culture. The total IgG and IgM levels were measured in controls (1.6±0.95% vs 1.1±0.6%; p=0.01) and for CD4+CX- coculture supernatants by ELISA (IgG Total Human uncoated CR5+PD-1int cells (4.9±2.31% vs 3.2±0.96%; p=0.003) ELISA kit and IgM human uncoated ELISA kit; ThermoFisher) (figure 1G). The CD4+CXCR5+PD-1high Tfh cells have an according to the manufacturers’ instructions. increased expression (MFI) of ICOS (mean 4984±2716 vs 7845±3187; p<0.0001) and of HLA-DR (mean 10 590±5525 vs 14 560±8466; p=0.023) as compared with CD4+CX- RT-PCR analyses + int RNA extractions from eight patients with SSc (n=3 dSSC and CR5 PD-1 cells (figure 1H,I). Thus, cTfh cells are increased n=5 lSSc) and six HC were performed using RNeasy Mini kit in patients with SSc and express higher levels of PD-1 and other (Qiagen, Cergy Pontoise, France). RNA was subjected to reverse activation markers, such as ICOS and HLA-DR. transcription (High Capacity RNA-to-cDNA Master Mix; Ther- moFisher) and quantified by real-time quantitative PCR using Circulating Tfh are increased in severe subtypes of SSc commercially available primer/probes sets (Assay-On-Demand; We then analysed the proportions of cTfh according to SSc ThermoFisher Scientific): GAPDH (Hs99999905_m1), IL-21 characteristics and subtypes. As expected, median Rodnan scale (Hs00222327_m1), BCL-6 (Hs00153368_m1), CXCL13 was significantly different in patients with dSS compared with

Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 541 152 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from http://ard.bmj.com/

Figure 1 Expansion of cTfh cells in the blood of patients with systemic sclerosis (SSc). (A) Gating strategy to identify cTfh and Tfh subsets by flow on 12 April 2019 by guest. Protected copyright. cytometry. (B) Frequencies of cTfh within CD4+ T cells in patients with SSc (n=50) and in healthy controls (HC) (n=32). (C) Absolute numbers of cTfh cells in patients with SSc (n=50) and in HC (n=24). (D) Absolute numbers of CXCR5−PD-1+CD4+ cells in patients with SSc and in HC. (E) Frequencies of cTfh subsets (Tfh1, Tfh2, Tfh17) within Tfh cells in patients with SSc and in HC. (F) Identification of PD-1hi Tfh by flow cytometry. (G) Frequencies of PD-1hi Tfh cells (left panel) and CXCR5+PD-1int CD4+ cells (right panel) among CD4+ T in SSc and HC. (H) Expression of inducible costimulatory (ICOS) mean fluorescence intensity (MFI arithmetic mean) by PD-1hi and PD-1int cells in SSc. (I) Expression of HLA-DR MFI arithmetic mean by PD-1hi and PD- 1int cells in SSc. Data represent the mean with SD; *p<0.05, **p<0.01, ***p<0.001 by Mann-Whitney U test. lSSc (p=0.0002) (table 1). We observed that cTfh cells were (p=0.053) or interstitial lung disease (figure 2D,E). Interest- significantly expanded in the severe dSSc as compared with ingly, cTfh frequencies correlated positively with mRSS scale lSSc (9.3±4.9% vs 6.3±2.7%; p=0.04) (figure 2A). Only cTfh (r=+0.33; p=0.023), but no correlation of Tfh cells and mRSS CXCR5+PD-1+ but not CXCR5−PD-1+CD4+ T cells were was observed when patients with dSSc and lSSc were analysed significantly increased in patients with dSSc (figure 2B). Circu- separately, probably because of the higher mRSS values in dSSc lating Tfh cell frequencies were also increased in patients with (figure 2F). The cTfh cell frequency did not correlate with C reac- pulmonary arterial hypertension (13.4±5.8% vs 6.2±2.2%; tive protein levels, brain natriuretic peptide, diffusing capacity p=0.003) (figure 2C). However, we did not observe differences for carbon monoxide, forced vital capacitylevels, the presence of in the proportions of cTfh in patients with active digital ulcers scleroderma renal crisis and other organ involvements (data not

542 Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 153 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from http://ard.bmj.com/

Figure 2 Circulating Tfh cells are increased according to disease severity. (A) Frequencies of cTfh among CD4+ T cells in limited cutaneous systemic − + + + sclerosis (lSSc) and in diffuse SSc (dSSc). (B) Frequencies of CXCR5 PD-1 CD4 cells among CD4 T cells in lSSc and in dSSc. (C) Frequencies of on 12 April 2019 by guest. Protected copyright. cTfh among CD4+ T cells in patients with SSc with or without pulmonary arterial hypertension (PAH). (D) Frequencies of cTfh among CD4+ T cells in patients with SSc with or without active digital ulcers. (E) Frequencies of cTfh among CD4+ T cells in patients with SSc with or without interstitial lung disease (ILD). (F) Correlation between frequencies of cTfh among CD4+ T cells and modified Rodnan skin score (mRSS) scale (A–E). Data represent the mean with SD; *p<0.05, **p<0.01, ***p<0.001 by Mann-Whitney U test. (F) Spearman test for patients with SSc (left panel), dSSc and lSSc (middle and right panel, respectively). ns, not significant.

shown). These data suggest that cTfh frequencies are increased numbers of CD19+CD27+38hiIgD− plasma B cells were not in the severe dSSc subtype. statistically different in patients with SSc and in HC (figure 3B). However, cTfh cell frequencies and numbers positively correlated Circulating Tfh cells express IL-21 and induce plasma cell with frequencies and numbers of CD19+CD27+38hiIgD− plasma differentiation through IL-21 B cells (r=+0.38; p=0.006) in patients with SSc, whereas no The cTfh cells are able to induce differentiation of naive B cells into significant correlation was found in HC (figure 3C). Circulating plasma cells through IL-21 secretion.16 20 As cTfh are increased in CXCR5−PD-1+CD4+ T-cell numbers (r=+0.37; p=0.025) and patients with SSc, we evaluated the frequency of plasma cells in frequencies (r=+0.514; p=0.0005) also positively correlated these patients, which were defined as CD19+CD27+CD38hiIgD− with CD19+CD27+CD38hiIgD− plasma B cells in patients with B cells in SSc and HC (figure 3A). The frequencies and absolute SSc, whereas no correlation was found for HC (figure 3D).

Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 543 154 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from

Figure 3 Circulating Tfh cells frequencies correlates with plasmablasts (PB) in systemic sclerosis (SSc). (A) Gating strategy to identify circulating B-cell subsets. PB were defined as CD19+CD27+CD38hiIgD− cells by flow cytometry. (B) Absolute numbers and frequencies of PB in SSc (n=50) and in healthy controls (HC) (n=24). (C) Correlation of PB and cTfh frequencies in SSc and in HC. (D) Correlation of PB and CXCR5−PD-1+CD4+ T-cell http://ard.bmj.com/ frequencies in SSc and in HC. (B) Data represent the mean with SD, *p<0.05, **p<0.01, ***p<0.001 by Mann-Whitney U test. (C–D) Spearman test. ns, not significant.

We further explored the functional capacity of cTfh from (figure 4D). In patients with SSc and HC, cTfh cells expressed patients with SSc. We first characterised the expression of the significantly higher transcriptional levels of CXCL-13 than transcription factor BCL-6 required for Tfh differentiation and CXCR5−PD-1+CD4+ T cells (p=0.008 and p=0.02 for patients on 12 April 2019 by guest. Protected copyright. B-cell help23 24 and Blimp-1 which antagonises BCL-6.25 We with SSc and HC, respectively) (figure 4C), whereas we did not analysed by RT-PCR the expression of BCL-6 and Blimp-1 in detect any CXCL-13 expression by CXCR5−PD-1−CD4+ T cells sorted CXCR5+PD-1+CD4+ and CXCR5−PD-1−CD4+ T cells (data not shown). cTfh cells from patients with SSc expressed from patients with SSc and HC. As shown in figure 4A, for significantly higher IL-21 than cTfh from HC (260±232 vs patients with SSc, BCL-6 expression was significantly higher in 100±0; p=0.02) (figure 4E). cTfh cells CXCR5+PD-1+CD4+ than in CXCR5−PD-1−CD4+ In order to determine the capacity of cTfh cells from patients T cells (p=0.016), whereas Blimp-1 expression was not signifi- with SSc to induce plasmablast differentiation, sorted cTfh were cantly different (p=0.55) (figure 4B). In addition, cTfh cells cocultured with autologous CD19+ B cells. The proportion of from patients with SSc also expressed higher BCL-6 than CD19+CD27+CD38hi plasmablasts was determined by flow CXCR5−PD-1+CD4+ T cells (figure 4A). In contrast, cTfh cells cytometry after 6 days of coculture with cTfh cells (figure 5A). from HC did not express differently BCL-6 nor BLIMP1 than In the cocultures of cells from patients with SSc, we observed a CXCR5−PD-1−CD4+ T cells (figure 4A and B). Then, we anal- significant increase of plasmablasts from B cells cocultured with ysed the expression of IL-21 and CXCL-13 cytokines by cTfh cTfh cells as compared with B cells cocultured with non-Tfh and CXCR5−PD-1− CD4+ T cells. Sorted cTfh from patients CXCR5−PD-1−CD4+ T cells (41.9±16.0% vs 7.1±5.6%; with SSc expressed higher transcriptional levels of IL-21 than p<0.0001) (figure 5B) or B cells alone. The levels of total IgG CXCR5−PD-1−CD4+ T cells (p=0.03) (figure 4D), and this and IgM were measured by ELISA in the supernatants of CD19+ increase was not observed for cTfh cells from HC (p=0.06) B cells cocultured with cTfh and non-Tfh cells. IgG and IgM

544 Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 155 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from http://ard.bmj.com/ on 12 April 2019 by guest. Protected copyright.

Figure 4 BCL-6 and IL-21 are increased in cTfh from patients with systemic sclerosis (SSc). (A–B) Quantification of the expression of the transcription factor Bcl-6 (A) and Blimp-1 (B) in blood CD4+ T cells by reverse transcription PCR (RT-PCR) (n=8 patients with SSc, n=6 healthy controls (HC)). (C–D) RT-PCR analysis of chemokine CXCL13 (C) and cytokine IL-21 (D) and in blood CD4+ T cells from patients with SSc (n=7 for IL-21 experiment and n=8 for CXCL13) or HC (n=6). (E) RT-PCR analysis of cytokine IL-21 in cTfh or CXCR5−PD-1+ CD4+ T cells from patients with SSc (n=6) or HC (n=6). (A–E) Data represent mean with SD, *p<0.05, **p<0.01, ***p<0.001 by Wilcoxon test. ns, not significant. concentrations were significantly increased in cTfh and B-cell promoting IgM secretion (p=0.01) (figure 5D). In the cocultures cocultures as compared with non-Tfh CXCR5−PD-1−CD4+ of cells from HC, the same results were observed for plasmab- T-cell and B-cell cocultures (2384±2510 ng/mL vs 190±287 ng/ last frequencies and Ig production (figure 5B–D). Interestingly, mL; p<0.0001) and (4592±3191 ng/mL vs 1315±1436 ng/mL; cTfh cells from patients with SSc were more efficient to induce p=0.005), respectively (figure 5C,D). Also, CXCR5−PD-1+CD4+ plasmablasts and IgG or IgM production in comparison with T cells induced higher plasmablast frequencies and IgG produc- healthy controls (figure 5B–D). tion than CXCR5−PD-1−CD4+ T cells (p=0.01 and p=0.0004, Altogether, these data suggest that cTfh from patients with SSc respectively) (figure 5B–C). Finally, CXCR5+PD-1+CD4+ T Tfh are fully functional in promoting plasmablast activation and Ig cells were more efficient than CXCR5−PD-1+CD4+ T cells in production.

Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 545 156 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from http://ard.bmj.com/ on 12 April 2019 by guest. Protected copyright.

Figure 5 Plasmablast (PB) cell frequencies and Ig secretion are increased in Tfh coculture with B cells. (A) Identification of PB by flow cytometry after 6 days of Tfh and B-cell cocultures, one representative example of patient. (B) Frequencies of PB after Tfh and B-cell cocultures or B-cell culture alone by flow cytometry (n=13 patients with systemic sclerosis (SSc), n=6 healthy controls (HC)). IgG (C) and IgM (D) in supernatants of coculture by ELISA (n=13 patients with SSc, n=6 HC). (B–D) Black bars represent patients with SSc, white bars represent HC. Data represent mean with SD, Mann- Whitney U test. ns, not significant.

IL-21 blockade and JAK1/2 inhibition suppress the cTfh cells’ After 6 days of cTfh and B-cell coculture with IL-21RFc ability to induce plasma B-cell proliferation and Ig secretion from patients with SSc, plasmablast cell frequency significantly Since IL-21 support the expansion of plasmablasts induced by decreased (38.4±19.6% without IL-21RFc vs 13.2±13.8% cTfh, we performed blocking experiments of the IL-21 signal- with IL-21RFc; p=0.002) (figure 6A). In addition, a signif- ling pathway. Thus, cTfh sorted from the blood of patients with icant decrease of IgG (2821±3051 ng/mL vs 508±616 ng/ SSc were cocultured with autologous B cells in the presence mL; p=0.004) (figure 6B) and IgM secretion (4592±3191 ng/ of recombinant IL-21RFc or in the presence of ruxolitinib, a mL vs 962±978 ng/mL; p=0.0003) (figure 6C) was observed JAK1/2 inhibitor which can affect CD4+ T-cell proliferation and with IL-21RFc in cTfh and B-cell cocultures from patients with inflammatory cytokine production.26 SSc. IL-21 blocking also reduced plasmablast cell frequencies

546 Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 157 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from http://ard.bmj.com/

Figure 6 cTfh cells promote plasmablast (PB) cell differentiation and Ig secretion through IL-21 that can be inhibited by ruxolitinib in patients with on 12 April 2019 by guest. Protected copyright. systemic sclerosis (SSc). (A) PB cell frequencies in Tfh-cell and B-cell cocultures in the presence of IL-21RFc (n=12 patients with SSc, n=6 healthy controls (HC)), or with anti-IL-6R (tocilizumab) (n=8 patients with SSc, n=6 HC). (B) IgG concentration in supernatants of Tfh and B-cell cocultures with or without IL-21RFc (n=11 patients with SSc, n=5 HC). (C) IgM concentration in supernatants of Tfh and B-cell cocultures with or without IL- 21RFc (n=11 patients with SSc, n=5 HC) or anti-IL-6R (n=7 patients with SSc, n=5 HC). (D–F) PB frequencies, IgG and IgM levels in supernatants after Tfh-cell and B-cell cocultures with or without ruxolitinib (n=6 patients with SSc, n=6 HC). (G) Cytokine IL-21 production by CD4+ T cells after phorbol 12-myristate 13-acetate and ionomycin stimulation with or without ruxolitinib (n=9 patients with SSc). (A–G) Data represent mean with SD, *p<0.05, **p<0.01***p<0.001 by Mann-Whitney U test. ns, not significant. and IgM production in cocultures of Tfh and B cells from HC The addition of ruxolitinib, a JAK1/2 inhibitor, to CD19+ (figure 6A–C). We also assessed the role of IL-6 in plasmablast B cells and cTfh cocultures from patients with SSc resulted in expansion induced by cTfh. IL-6 has been previously reported to a significant decrease of plasmablast expansion (44.4±7.8% be highly secreted by B cells in SSc7 and able to stimulate IL-21 vs 4.7±2.9%; p=0.0007), of IgG (1568±937 ng/mL vs production by CD4+ T cells.27 The addition of tocilizumab, an 56.8.6±75.7 ng/mL; p=0.02) and IgM secretions (4575±2447 anti-IL-6 receptor monoclonal antibody, to CD19+ B and Tfh ng/mL vs 105±132 ng/mL; p=0.002) (figure 6D–F). Similarly, cell cocultures did not significantly reduce the frequency of in cTfh and B-cell cocultures from HC, the addition of ruxoli- plasmablasts (figure 6A) or their secretion of IgM (figure 6C) in tinib resulted in a significant reduction of plasmablast frequen- patients with SSc and in HC. cies, IgG and IgM productions (figure 6D–F). The addition of

Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 547 158 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from

IL-21RFc or ruxolitinib to CXCR5−PD-1+CD4+ T-cell and been shown to infiltrate the skin of patients with SSc.21 The CD19+ B-cell cocultures from patients with SSc also reduced the increased frequencies of cTfh cells that we observed in patients plasmablast cell differentiation and IgG secretion (online supple- with SSc could be related to overall increased Tfh cells in blood mentary figure 1). and tissues, or represent Tfh cell homeostasis disturbance with Thus, these results indicated that ruxolitinib could affect increased levels of circulating cells. Few studies have analysed plasmablast differentiation and their capacity to secrete immu- circulating or tissue-infiltrating Tfh cell frequencies and homeo- noglobulins. In order to determine if ruxolitinib can directly stasis in SSc. Two studies described increased ICOS+ T cell in impact the IL-21 production by cTfh, we stimulated in vitro the skin of patients with SSc30 31 and one recent study showed cTfh from patients with SSc by PMA and ionomycin in the that CD4+ICOS+PD-1+CXCR5+ Tfh cells infiltrate the skin of presence of ruxolitinib. We observed that IL-21 levels in the patients with SSc. Moreover, the presence of Tfh cells in the supernatants of activated cTfh were significantly reduced skin was correlated with clinical disease and dermal fibrosis.21 by ruxolitinib (3033±1419 pg/mL vs 1455±1101 pg/mL; These CD4+ICOS+PD-1+CXCR5+ Tfh cells from SSc were able p=0.013). Conversely, IL-21 secretion by CXCR5−PD-1+CD4+ to induce dermal fibroblast differentiation into myofibroblasts was not affected by the presence of ruxolitinib (figure 6G). Taken in vitro.21 In a murine model of graft-versus-host disease, SSc together, these results demonstrate that cTfh from patients with ICOS+ Tfh cells were increased and neutralisation of IL-21 or SSc can drive the expansion of plasma cells and immunoglobulin the administration of an anti-ICOS antibody prevented clin- production, which can be successfully inhibited by IL-21 neutral- ical dermal fibrosis.21 In our study, we observed that cTfh cells isation with ruxolitinib. expressing high levels of PD-1 also expressed high levels of ICOS. We showed that cTfh cells from patients with SSc had an Discussion increased expression of the transcriptional factors BCL-6, and of Autoimmune diseases are usually characterised by increased CXCL13 and IL-21 cytokine. These factors are associated with autoantibody production which can result in organ dysfunction B-cell interaction, suggesting that these cTfh are able to coop- through various mechanisms, such as increased apoptosis, recruit- erate with B cells. BCL-6 is a the master regulator of Tfh cells.14 ment of inflammatory cells or direct cytotoxicity. This increased This observation is in line with previous studies, showing that production of autoantibodies can result from a tolerance break- only Tfh cells from germinal centres (GC Tfh) expressed BCL-6, 28 down during the B-cell development or be the consequence whereas BCL-6 expression is not detected in blood cTfh.16 32 29 of somatic mutation and aberrant selection into the GCs. Thus, our observation of the increased expression of BCL-6 in Tfh cells are crucial immune regulators of B-cell activation and cTfh from patients with SSc together with the higher expression differentiation and recently cTfh have been implicated in several of IL-21 suggest that cTFh from patients with SSc may have a autoimmune diseases such as lupus erythematosus, Sjogren’s phenotype of effector cells, in contrast to the resting phenotype syndrome and rheumatoid arthritis. Moreover, cTfh frequencies of cTfh from HC which do not express BCL-6. have been shown to correlate with disease severity and levels Using cocultures with B cells, we show that Tfh from patients 17–19 of autoantibodies. In SSc, B cells can secrete various auto- with SSc were fully functional and more efficient to induce antibodies that correlate with disease subtypes, disease severity plasmablast differentiation and their IgG and IgM secretion than and prognosis. Thus, several studies showed that B-cell homeo- cTfh from HC. This higher capacity could rely on the higher stasis was altered in SSc; however, the precise mechanisms of IL-21 expression. The IL-21 cytokine is known to be a potent B-cell excessive maturation remain understudied. In this study, plasma cell inducer33–35 and IL-21R is highly expressed in the we performed the functional analysis of Tfh cells from patients skin of patients with SSc.36 37 In a murine model of bleomycin-in- + with SSc and analysed their interaction with B cells. First, we duced fibrosis, increased IL-21 production from CD4 T cells http://ard.bmj.com/ demonstrate that cTfh are increased in SSc and are correlated was correlated with dermal fibrosis and lung fibrosis.38 More- with disease severity. Indeed, the cTfh frequencies were signifi- over, IL-21 could induce fibrosis through activation of CD8+ T cantly increased in patients with SSc with dSSc subtypes, which cells.39 The CD4+ T cells secreting IL-21 sorted from synovial are characterised by more aggressive disease outcome and fluid of patients with rheumatoid arthritis induce the produc- increased mortality. cTFH cells were also increased in patients tion of MMP-1 and MMP-3 by a fibroblast-like synoviocyte with pulmonary arterial hypertension which represents a severe that promotes joint inflammation, further supporting a role for and life-threating complication of SSc. The skin progression, IL-21 beyond the regulation of B-cell response.40 We demon- on 12 April 2019 by guest. Protected copyright. as represented by the mRSS scale, is one of best predictors of strated that IL-21 is highly secreted by cTfh from patients with disease activity in SSc, and the Tfh cell frequency is significantly SSc and blocking Il-21R significantly reduced Tfh cells’ capacity correlated with the mRSS in our study. Further studies would be to induce plasmablast cells. needed to analyse the cTfh cell frequencies and their functional Blocking proinflammatory cytokines such as IL-17, TNFα capacity in patients under treatment who respond in particular or IL-6 with monoclonal antibodies has been successfully used with skin improvement. We did not analyse cTfh correlation to treat several immune diseases such as psoriasis,41 rheuma- with autoantibody levels in the serum of patients with SSc as toid arthritis or ankylosing arthritis.42 However, the cytokine ELISA tests used to detect autoantibodies provided only qualita- neutralisation strategy presents some limitations: some patients tive data, and more than 80% of patients with SSc were positive did not respond sufficiently or presented secondary treatment for autoantibodies in the serum. Another subset of CD4+ T cells, failure because of drug immunogenicity.43 The targeting of the which can also stimulate plasma B-cell differentiation, has been Janus Kinase pathway has been recently experimented because recently described in rheumatoid arthritis and characterised as the JAK signalling pathway is associated with various cytokine CXCR5−PD-1+CD4+ peripheral helper T cells. The T follicular receptors.22 Ruxolitinib is a JAK1 and JAK2 antagonist, which helper-like cell frequencies were similar to HC, and these cells inhibits intracellular signalling of multiple proinflammatory did not correlate with SSc severity. The CXCR5−PD-1+ CD4+ cytokines including IL-21.44 IL-21 expression has been signifi- peripheral helper T cells were expanded in the synovium of cantly reduced by ruxolitinib in a rodent model of rheuma- patients with rheumatoid arthritis20 and tissular Tfh cells have toid arthritis.45 In graft-versus-host disease, ruxolitinib showed

548 Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 159 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from promising results in patients with steroid-refractory disease46 6 Sato S, Fujimoto M, Hasegawa M, et al. Altered blood B lymphocyte homeostasis in and is currently being tested in prospective studies in rheuma- systemic sclerosis: expanded naive B cells and diminished but activated memory B cells. Arthritis Rheum 2004;50:1918–27. toid arthritis, psoriasis, myositis and graft-versus-host disease. 7 Dumoitier N, Chaigne B, Régent A, et al. Scleroderma peripheral B lymphocytes Treatment with ruxolitinib reduced skin and pulmonary fibrosis secrete interleukin-6 and transforming growth factor β and activate fibroblasts. in bleomycin-induced pulmonary fibrosis and in adTBR-in- Arthritis Rheumatol 2017;69:1078–89. duced dermal fibrosis.47 We report that ruxolitinib inhibits the 8 Matsushita T, Hasegawa M, Yanaba K, et al. Elevated serum BAFF levels in patients cTfh capacity to secrete IL-21 and decrease the proportions of with systemic sclerosis: enhanced BAFF signaling in systemic sclerosis B lymphocytes. Arthritis Rheum 2006;54:192–201. plasmablasts induced by cTfh cells in vitro, similarly to IL-21R 9 Wutte N, Kovacs G, Berghold A, et al. CXCL13 and B-cell activating factor as putative blockade. biomarkers in systemic sclerosis. Br J Dermatol 2013;169:723–5. The molecular events leading to the expansion of cTfh in 10 Bielecki M, Kowal K, Lapinska A, et al. Increased production of a proliferation- patients with SSc have not been explored. Human cTfh cell inducing ligand (APRIL) by peripheral blood mononuclear cells is associated with antitopoisomerase I antibody and more severe disease in systemic sclerosis. J differentiation and expansion pathways are not actually well Rheumatol 2010;37:2286–9. established. IL-12 produced by activated dendritic cells could 11 King C, Tangye SG, Mackay CR. T follicular helper (Tfh) cells in normal and promote human Tfh cell differentiation.48 Other cytokines such dysregulated immune responses. Annu Rev Immunol 2008;26:741–66. as IL-6 and IL-21 could contribute to Tfh cell differentiation, 12 Fazilleau N, Mark L, McHeyzer-Williams LJ, et al. Follicular helper T cells: lineage and + 15 location. Immunity 2009;30:324–35. inducing BCL-6 expression by CD4 T cells. Recently, plas- 13 Breitfeld D, Ohl L, Kremmer E, et al. Follicular B helper T cells express CXC chemokine macytoid dendritic cells have been involved in SSc pathogenesis receptor 5, localize to B cell follicles, and support immunoglobulin production. J Exp via the aberrant expression of TLR-8. One can speculate that Med 2000;192:1545–52. plasmacytoid dendritic cells could contribute to Tfh cell expan- 14 Crotty S. Follicular helper CD4 T cells (Tfh). Annu Rev Immunol 2011;29:621–63. sion by increasing proinflammatory cytokine production in the 15 Nurieva RI, Chung Y, Martinez GJ, et al. BCL6 mediates the development of T follicular 49 helper cells. Science 2009;325:1001–5. context of SSc. STAT-3 is the major transducer of IL6-R and 16 Morita R, Schmitt N, Bentebibel SE, et al. Human blood CXCR5(+)CD4(+) T cells are 50 51 IL-21R on T and B cells. In addition, IL-21 activates B cells counterparts of T follicular cells and contain specific subsets that differentially support via the JAK–STAT signalling pathway.52 53 STAT3 regulates IL-21 antibody secretion. Immunity 2011;34:108–21. expression by CD4+ T cells and is important for Tfh differen- 17 Ma J, Zhu C, Ma B, et al. Increased frequency of circulating follicular helper T cells in 17 54 patients with rheumatoid arthritis. Clin Dev Immunol 2012;2012:1–7. tiation. Furthermore, an increased expression of phosphor- 18 Li XY, Wu ZB, Ding J, et al. Role of the frequency of blood CD4(+) CXCR5(+) CCR6(+) ylated STAT3 has been reported in skin biopsies from patients T cells in autoimmunity in patients with Sjögren’s syndrome. Biochem Biophys Res with SSc.55 The mechanism of Tfh cell inhibition by ruxolitinib Commun 2012;422:238–44. could be the inhibition of STAT-3 phosphorylation,45 which will 19 Le Coz C, Joublin A, Pasquali JL, et al. Circulating Tfh subset distribution is strongly affected in lupus patients with an active disease. PLoS One 2013;8:e75319. modulate the proinflammatory cytokine receptor expression on 20 Rao DA, Gurish MF, Marshall JL, et al. Pathologically expanded peripheral T helper cell T and B cells. subset drives B cells in rheumatoid arthritis. Nature 2017;542:110–4. In conclusion, circulating Tfh cells are increased in SSc and 21 Taylor DK, Mittereder N, Kuta E, et al. T follicular helper-like cells contribute to skin correlate with severe SSc, notably in dSS subtypes. cTfh from fibrosis. Sci Transl Med 2018;10:eaaf5307. patients with SSc present an activated phenotype and induce 22 van den Hoogen F, Khanna D, Fransen J, et al. 2013 classification criteria for systemic sclerosis: an American College of Rheumatology/European League Against B-cell differentiation into plasmablasts secreting Ig via IL-21 Rheumatism collaborative initiative. Arthritis Rheum 2013;65:2737–47. secretion. IL-21 or JAK1/2 blockade by ruxolitinib could be a 23 Yu D, Rao S, Tsai LM, et al. The transcriptional repressor BCL-6 directs T follicular promising strategy in the treatment of SSc. helper cell lineage commitment. Immunity 2009;31:457–68. 24 Hatzi K, Nance JP, Kroenke MA, et al. BCL6 orchestrates Tfh cell differentiation via Acknowledgements We would like to thank Dr Hélène Fohrer-Ting for providing multiple distinct mechanisms. J Exp Med 2015;212:539–53. help in cTfh sorting experiments (Centre d’Histologie, d’Imagerie et de Cytométrie 25 Johnston RJ, Poholek AC, DiToro D, et al. BCL6 and Blimp-1 are reciprocal (CHIC), Centre de Recherche des Cordeliers, 75006 Paris, France). We also thank and antagonistic regulators of T follicular helper cell differentiation. Science Frederic Devassoigne and Christophe Devassoigne for their help in collecting 2009;325:1006–10. http://ard.bmj.com/ patients’ blood samples (Tumorothèque Saint-Antoine, APHP, Hôpital Saint-Antoine, 26 Keohane C, Kordasti S, Seidl T, et al. JAK inhibition induces silencing of T helper 75012 Paris, France). cytokine secretion and a profound reduction in T regulatory cells. Br J Haematol Contributors All authors listed on the manuscript have substantially contributed 2015;171:60–73. to this work. BG, LR and AM designed the experimental research, performed 27 Suto A, Kashiwakuma D, Kagami S, et al. Development and characterization of IL-21- experiments, interpreted data and wrote the manuscript. FM, MM, VJ, SR, NS, producing CD4+ T cells. J Exp Med 2008;205:1369–79. YY, PS, J-BM and OF interpreted the data and participated in the writing of the 28 Yurasov S, Nussenzweig MC. Regulation of autoreactive antibodies. Curr Opin manuscript. Rheumatol 2007;19:421–6. 29 Vinuesa CG, Sanz I, Cook MC. Dysregulation of germinal centres in autoimmune on 12 April 2019 by guest. Protected copyright. Funding This work was supported by the Aterhit Foundation and received grants disease. Nat Rev Immunol 2009;9:845–57. from ’Groupe Francophone de Recherche sur la Sclérodermie’ (GFRS). 30 Yanaba K, Asano Y, Noda S, et al. Increased production of soluble inducible Competing interests None declared. costimulator in patients with diffuse cutaneous systemic sclerosis. Arch Dermatol Res 2013;305:17–23. Patient consent for publication Obtained. 31 Hasegawa M, Fujimoto M, Matsushita T, et al. Augmented ICOS expression Ethics approval This study was approved by the Ethic committee of ’Kremlin in patients with early diffuse cutaneous systemic sclerosis. Rheumatology Bicêtre University’ (no. ID-RCB 2017-AO3380-53). 2013;52:242–51. 32 Chevalier N, Jarrossay D, Ho E, et al. CXCR5 expressing human central memory Provenance and peer review Not commissioned; externally peer reviewed. CD4 T cells and their relevance for humoral immune responses. J Immunol 2011;186:5556–68. References 33 Kuchen S, Robbins R, Sims GP, et al. Essential role of IL-21 in B cell activation, D1 umoitier N, Lofek S, Mouthon L. Pathophysiology of systemic sclerosis: state of the expansion, and plasma cell generation during CD4+ T cell-B cell collaboration. J art in 2014. Presse Med 2014;43(10 Pt 2):e267–78. Immunol 2007;179:5886–96. 2 Gabrielli A, Avvedimento EV, Krieg T. Scleroderma. N Engl J Med 34 Ettinger R, Sims GP, Fairhurst AM, et al. IL-21 induces differentiation of human 2009;360:1989–2003. naive and memory B cells into antibody-secreting plasma cells. J Immunol 3 Kalogerou A, Gelou E, Mountantonakis S, et al. Early T cell activation in the skin from 2005;175:7867–79. patients with systemic sclerosis. Ann Rheum Dis 2005;64:1233–5. 35 Bryant VL, Ma CS, Avery DT, et al. Cytokine-mediated regulation of human B cell 4 O’Reilly S, Hügle T, van Laar JM. T cells in systemic sclerosis: a reappraisal. differentiation into Ig-secreting cells: predominant role of IL-21 produced by CXCR5+ Rheumatology 2012;51:1540–9. T follicular helper cells. J Immunol 2007;179:8180–90. 5 Kraaij MD, van Laar JM. The role of B cells in systemic sclerosis. Biologics 36 Distler JH, Jüngel A, Kowal-Bielecka O, et al. Expression of interleukin-21 receptor in 2008;2:389–95. epidermis from patients with systemic sclerosis. Arthritis Rheum 2005;52:856–64.

Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 549 160 Systemic sclerosis Ann Rheum Dis: first published as 10.1136/annrheumdis-2018-214382 on 13 February 2019. Downloaded from

37 pesce J, Kaviratne M, Ramalingam TR, et al. The IL-21 receptor augments Th2 effector 47 Zhang Y, Liang R, Chen CW, et al. Jak1-dependent transphosphorylation of JAK2 limits function and alternative macrophage activation. J Clin Invest 2006;116:2044–55. the antifibrotic effects of selective ak2J inhibitors on long-term treatment. Ann Rheum 38 Lei L, He ZY, Zhao C, et al. Elevated frequencies of CD4(+) IL-21(+) T, CD4(+) IL- Dis 2017;76:1467–75. 21R(+) T and IL-21(+) Th17 cells, and increased levels of IL-21 in bleomycin-induced 48 Schmitt N, Morita R, Bourdery L, et al. Human dendritic cells induce the differentiation mice may be associated with dermal and pulmonary inflammation and fibrosis. Int J of interleukin-21-producing T follicular helper-like cells through interleukin-12. Rheum Dis 2016;19:392–404. Immunity 2009;31:158–69. 39 Brodeur TY, Robidoux TE, Weinstein JS, et al. IL-21 promotes pulmonary fibrosis 49 Ah Kioon MD, Tripodo C, Fernandez D, et al. Plasmacytoid dendritic cells promote through the induction of profibrotic CD8+ T cells. J Immunol 2015;195:5251–60. + systemic sclerosis with a key role for TLR8. Sci Transl Med 2018;10:eaam8458. 40 Lebre MC, Vieira PL, Tang MW, et al. Synovial IL-21/TNF-producing CD4 T cells 50 Spolski R, Leonard WJ. Interleukin-21: basic biology and implications for cancer and induce joint destruction in rheumatoid arthritis by inducing matrix metalloproteinase autoimmunity. Annu Rev Immunol 2008;26:57–79. production by fibroblast-like synoviocytes. J Leukoc Biol 2017;101:775–83. 51 Rayanade RJ, Ndubuisi MI, Etlinger JD, et al. Regulation of IL-6 signaling by p53: 41 Leonardi C, Matheson R, Zachariae C, et al. Anti-interleukin-17 monoclonal antibody STAT3- and STAT5-masking in p53-Val135-containing human hepatoma Hep3B cell ixekizumab in chronic plaque psoriasis. N Engl J Med 2012;366:1190–9. lines. J Immunol 1998;161:325–34. 42 McInnes IB, Schett G. The pathogenesis of rheumatoid arthritis. N Engl J Med 52 Berglund LJ, Avery DT, Ma CS, et al. IL-21 signalling via STAT3 primes human naive 2011;365:2205–19. 43 Garcês S, Demengeot J. The immunogenicity of biologic therapies. Curr Probl Dermatol B cells to respond to IL-2 to enhance their differentiation into plasmablasts. Blood 2018;53:37–48. 2013;122:3940–50. 44 Ghoreschi K, Gadina M. Jakpot! New small molecules in autoimmune and 53 Deenick EK, Avery DT, Chan A, et al. Naive and memory human B cells have distinct inflammatory diseases. Exp Dermatol 2014;23:7–11. requirements for STAT3 activation to differentiate into antibody-secreting plasma 45 Fridman JS, Scherle PA, Collins R, et al. Selective inhibition of JAK1 and Jak2 is cells. J Exp Med 2013;210:2739–53. efficacious in rodent models of arthritis: preclinical characterization ofINC B028050. J 54 Wei L, Laurence A, Elias KM, et al. IL-21 is produced by Th17 cells and drives IL-17 Immunol 2010;184:5298–307. production in a STAT3-dependent manner. J Biol Chem 2007;282:34605–10. 46 Zeiser R, Burchert A, Lengerke C, et al. Ruxolitinib in corticosteroid-refractory graft- 55 Chakraborty D, Šumová B, Mallano T, et al. Activation of STAT3 integrates common versus-host disease after allogeneic stem cell transplantation: a multicenter survey. profibrotic pathways to promote fibroblast activation and tissue fibrosis. Nat Commun Leukemia 2015;29:2062–8. 2017;8:1130. http://ard.bmj.com/ on 12 April 2019 by guest. Protected copyright.

550 Ricard L, et al. Ann Rheum Dis 2019;78:539–550. doi:10.1136/annrheumdis-2018-214382 161 Article 3

Old dog, New trick: Trivalent Arsenic as an Immunomodulatory

Drug

Yishan Ye, Béatrice Gaugler, Mohamad Mohty, Florent Malard

Under review in Frontiers in Immunology

162 Immunomodulation by trivalent arsenic

1 Old Dog, New Trick: Trivalent Arsenic as an Immunomodulatory 2 Drug

3

4 Yishan Ye, M.D.1,2, Béatrice Gaugler, Ph.D.1, Mohamad Mohty M.D., Ph.D.1,3, Florent Malard 5 M.D., Ph.D.1,3*

6 1Sorbonne Université, INSERM, Centre de Recherche Saint-Antoine (CRSA), F-75012 Paris, France 7 2Bone Marrow Transplantation Center, The First Affiliated Hospital, School of Medicine, Zhejiang 8 University, Hangzhou, 310003, China 9 3AP-HP, Hôpital Saint-Antoine, Service d’Hématologie Clinique et Thérapie Cellulaire, Sorbonne 10 Université, F-75012, Paris, France

11 * Correspondence: 12 Florent Malard, M.D., Ph.D., Service d’Hématologie Clinique et de Thérapie Cellulaire, Hôpital 13 Saint Antoine, APHP, Sorbonne Université and INSERM, UMRs 938, 184 rue du Faubourg Saint- 14 Antoine, 75012, Paris, France. Phone : +33 149282629 ; Fax : +33 149283375

15 E-mail: [email protected]

16 Keywords: Trivalent arsenic1, Immunomodulation2, Mechanisms of action3, Immune subset4, 17 Mouse model5, Autoimmune disease6, Cancer immunotherapy7

18

19 Abstract

20 Trivalent arsenic (As(III)), known for its rediscovery in the treatment of acute promyelocytic 21 leukemia, is recently found to be an immunomodulatory agent. As(III) have therapeutic potential in 22 mouse models of several autoimmune and inflammatory diseases. In vitro, it selectively induces 23 apoptosis of immune cells due to different sensitivity. At a non-toxic level, As(III) shows its 24 multifaceted nature by inducing either pro- or anti-inflammatory functions of immune subsets due to 25 signaling pathway modulation. These effects are exerted by either As(III)-protein interactions or as a 26 consequence of As(III) induced homeostasis imbalance. The immunomodulatory properties also 27 confer As(III) synergistic effect with cancer immunotherapy. In this review, we summarize the 28 immunomodulatory effects of As(III) , focusing on the effects As(III) on immune subsets in vitro, on 29 mouse models of immune-related diseases, and the role of As(III) in cancer immunotherapy. The 30 updates in mechanisms of action, and the pioneer clinical trials are also described.

31

32

33

163 Immunomodulation by trivalent arsenic

34 1 Introduction

35 Inorganic arsenic compounds (iAs) have been used in traditional Chinese and Western medicine for 36 over 2400 years (1). Arsenic trioxide (As2O3) was rediscovered in the 1970s in the treatment of acute 37 promyelocytic leukemia (APL) with striking efficacy and good safety profile (2,3). So far, As2O3, 38 together with all-trans retinoic acid (ATRA), has revolutionized the treatment of APL: the outcome 39 changed from 35% to 45% long-term overall survival (OS) by chemotherapy to a complete remission 40 (CR) rate of over 95% and a long-term survival rate up to 90% with a chemotherapy free approach 41 (4,5). Furthermore, As2O3 has also shown promising results in many other malignancies such as adult 42 T-cell leukemia/lymphoma (ATL), and NPM1 mutant acute myeloid leukemia (AML) (6,7).

43 The immune-modulation of iAs has been studied and put in clinical use with trivalent arsenicals 44 (As(III)) including As2O3 and sodium arsenite (NaAsO2). As(III) has high affinity for sulfhydryl 45 groups and can bind to cysteines in peptides and proteins, leading to protein conformation change 46 and malfunction (8). As(III) is also a potent reactive oxygen species(ROS) inducer (9). Moreover, 47 many key proteins are indirectly regulated by As(III) through post-translational regulations such as 48 sumoylation (10) (mechanisms reviewed by Shen et al) (8).

49 iAs are regarded as having general immunotoxity to the immune system (11). However, in vitro 50 studies have provided a more detailed picture, revealing the multifaceted effects of As(III) on 51 different immune subsets (12,13). Recent reports on animal models reveal the efficacy of As(III) on 52 several autoimmune and inflammatory diseases (14,15). In addition, in the era of cancer 53 immunotherapy, emerging evidence indicates As(III) as a promising immunoadjuvant for 54 hematological malignancies and solid tumors (16,17).

55 Here, in an aim to promote the better use of this double-edged sword in immunotherapy, we 56 summarize the mechanisms of action of As(III), the modulation on each immune cell subset, the use 57 of As(III) on pre-clinical mouse models of immune-mediated diseases, as well as the pioneer studies 58 of As(III) as an adjuvant for tumor immunotherapy.

59

60 2 Pharmaceutical mechanisms of action

61 2.1 As(III) biochemistry

62 Both As2O3 and NaAsO2 transform to As(OH)3 in aquarious solution. After being up-taken by cells 63 through the aquaglyceroporin 9 (AQP9) transmembrane protein, the intracellular As(III) exerts the 64 subsequent biochemical effects (18). Interaction with the thiol (or sulfhydryl) groups (-SH) of 65 proteins with a high cysteine content constitutes the basic biochemical reaction of As(III) (1), which 66 alters the conformation, resulting in loss of their function, and affect their recruitment and interaction 67 with other proteins and DNA (8). Apart from the direct arsenic-protein binding, recent studies reveal 68 that many key proteins are modulated through more complicated post-translational stepwise 69 regulations (19).

70 2.2 PML and PML nuclear body regulation

71 It was initially found that the APL promotor promyelocytic leukemia-retinoic acid receptor-α 72 (PML/RARα) was especially sensitive to As2O3-induced degradation, which was the critical step of 73 disease eradication (20). The following comprehensive studies by de Thé etalrevealthatAs2O3

164 Immunomodulation by trivalent arsenic

74 specifically targets the PML moiety (21). PML nuclear bodies (NBs), nucleared by the PML protein, 75 could recruit and sumoylate dozens of partner proteins, leading to a variety of biological processes 76 (10,22). Notably, As2O3-induced oxidative stress could enhance formation of PML NBs, leading to 77 p53 activation in vivo in normal mice (23).

78 In APL, through both reactive oxygen species (ROS) production and direct binding, As2O3 exerts its 79 dual-targeting effects (24). On one hand, As2O3 induces PML/RARα sumoylation, proteasomal 80 degradation, and APL cell differentiation (21,25). On the other hand, As2O3 targets the wild-type 81 PML proteins, leading to re-formation of PML NBs, subsequent p53 activation, and ultimately APL 82 clearance (23,26). It is also found that in ATL, As2O3, together with interferon-α, reaches disease 83 complete remission both in mice and in human through degradation of the disease driver oncoprotein 84 Tax (6,27,28). Interestingly, this process was also mediated through a As2O3 enforcement of PML 85 NBs formation, Tax sumoylation and proteasomal degradation (29). In addition, it is also discovered 86 that the As2O3/ATRA combination significantly decreases NPM1-mutant AML leukemia blast in 87 patients though oxidative stress generation, p53 activation, and ultimately mutant NPM1 degradation 88 (7,30).

89 Apart from tumor suppression, PML and PML NBs are recently found to play key role in mediating 90 innate immune response (10,31,32). Studies have identified PML as a direct, positive regulator of 91 IFN I signaling (33), and is implicated in the regulation of and extended spectrum of cytokines such 92 as the pro-inflammatory IL-1β and IL-6 (32,34). It is thus with great interest to investigate if PML 93 and PML NBs modulations explain As2O3’ efficacy in autoimmune/inflammatory diseases with type- 94 I IFN signature and abnormal cytokine profile.

95 2.3 Apoptosis induction

96 As(III) can induce immune cell apoptosis through both the mitochondrial-mediated and the receptor- 97 mediated pathways (12,35). The pro-apoptotic mechanisms include generation of oxidative stress, 98 caspase activation, alteration of the Bcl-2 family proteins, and up/down-regulation of several 99 survival-related signaling pathways such as nuclear factor-kB (Nf-kB), Rho-kinase/p38-kinase, and 100 TNF-R1 apoptotic signalings (12,35-37). Within the therapeutic range, As(III) is able to induce 101 apoptosis of specific types of sensitive immune cells (14,16). In a mouse model of colon cancer, 102 regulatory T cells are preferentially depleted, leading to immunomodulatory effects (16). The in vitro 103 EC50 of different immune subsets and the possible mechanisms of apoptosis induction are 104 summarized in Table 1.

105 2.4 Reactive oxygen species

106 ROS accumulation leads to oxidative stress, causing damage to nucleic acids, proteins, and lipids, 107 which can lead to cell signaling change, and, ultimately, apoptosis (38). As(III) induces immune cell 108 intracellular ROS accumulation via inhibition of antioxidants such as glutathione (GSH), glutathione 109 peroxidase (Gpx) and glutathione reductase (GR) (12,39), and activation of ROS generation enzymes 110 such as NADPH oxidase (40). As a consequence, expressions of redox-sensitive genes such as 111 HMOX1, NQO1 and GCLM are up-regulated (41). Especially, nuclear factor erythroid 2-related 112 factor 2 (Nrf2), a stress-activated transcription factor responsible for inducing a battery of 113 cytoprotective genes, is found to be activated by As(III) within mouse splenocytes, human T cells, 114 human primary macrophages, and human monocyte derived dendritic cells (MDDCs) (41-44). The 115 Nrf2 modulation is probably due to As(III)-induced Nrf2 sumoylation, recruitment to the PML NBs 116 and degradation by RNF4-mediated proteolysis (45).

165 Immunomodulation by trivalent arsenic

117 2.5 Signaling pathway regulation

118 Due to the different affinity of different proteins to As(III), and the potential post-translational 119 modulation by other structures such as PML NBs, As(III) has the potential to selectively target 120 specific signaling pathways. In macrophages, for example, non-toxic concentrations of As2O3 induce 121 up-regulation of 32 and depression of 91 genes, indicating a global multi-directional change of the 122 cell-signaling (46). Moreover, As(III) may inactivate up to 200 enzymes, and many are crucial 123 regulators of important signaling pathways (8). As(III) inhibits IĸB kinase (IKK), whose integrity is 124 key to the activation of the NFĸB pathway (36). Other pathways affected include the Rho- 125 kinase/p38-kinase pathway, c-jun NH2-terminal MAPK, etc, which will be described in details in the 126 next section (40,47).

127

128 3 The multifaceted effects of trivalent arsenic on immunocytes

129 3.1 Systemic immunomodulation

130 Chronic arsenic exposure leads to systemic immunosuppression (11). However, clinical use of As2O3 131 has shown good safety-profile, with limited side effects (3). No long-term treatment-related immune- 132 mediated disease or tumorigenesis is reported. During the As2O3 single agent treatment for APL, 133 there are inhibitory effects on hematopoietic progenitor cells, and it takes 145 and 265 days for 134 circulating T and B cells, and 655 days for natural killer (NK) cells to achieve the median normal 135 levels (48). The detailed effects of As(III) on each immune subset are described in the following 136 sections.

137 3.2 Trivalent arsenic and granulocytes

138 The multifaceted effects of As(III) on different subsets of immnuocytes aresummarizedinFigure1. 139 High concentrations of As2O3 induce apoptosis of human neutrophils via generation of H2O2, de novo 140 protein synthesis, caspase activation, and Syk kinase activation (49,50). The As2O3-induced de novo 141 synthesized proteins include annexin-1 and heat shock proteins (51). Moreover, As2O3 induces 142 endoplasmic reticulum (ER) stress within human neutrophils, which elicits either self-protective 143 mechanisms via the activation of unfolded protein response (UPR), or cell apoptosis independent of 144 caspase-4 activation (52,53). Importantly, As2O3 recruits the mitogen-activated protein kinases 145 (MAPKs), activates p38 and c-jun NH2-terminal (JNK), and ultimately enhances the major 146 neutrophil functions including adhesion, migration, phagocytosis and degranulation (47). Syk 147 activation also helps in this agonistic process (50). In addition, NaAsO2 promotes the formation of 148 neutrophil extracellular traps (NETs), which play an important role in pathogen infection, 149 cardiovascular disease and several cancers (54). Moreover, it was shown inanacuteAs2O3-exposed 150 mouse model that the neutrophil cell numbers are increased in the broncho-alveolar lavage fluid 151 (BALF) (55). Collectively, neutrophils are recruited and functionally enhanced during As(III) 152 treatment, leading to a pro-inflammatory response.

153 For eosinophil, in a mouse model of asthma, eosinophil recruitment and a reduction of chemotaxis 154 level in BALF were detected after As2O3 treatment (56,57). Therefore, eosinophil functions may be 155 impaired during As(III) exposure. For basophils, the direct effects of As(III) on basophils have not 156 yet been investigated. However, the activation of basophils may be impaired in vivo during As(III) 157 administration, because a decreased level of IgE, an important driver of basophil activation, was 158 discovered in the BALF of As2O3-treated mice (56,57). For mast cells, it is reported that NaAsO2

166 Immunomodulation by trivalent arsenic

159 inhibits anti-IgE stimulated degranulation via suppression of early tyrosine phosphorylation (58). 160 Therefore, As(III) may impair mast cell activation, especially during the process of allergy (59).

161 3.3 Trivalent arsenic and monocytes/macrophages

162 Recent studies have shown that monocytes are not simply precursors of macrophages as previously 163 thought, but in fact give rise to functionally distinct monocyte-derived cells during inflammation (60). 164 As2O3 induces human monocyte apoptosis during macrophage differentiation through down- 165 regulation of the Nf-kB related pathway (36). Moreover, As2O3 increases lipopolysaccharide (LPS) 166 dependent expression of the inflammatory IL-8 gene, by stimulating a redox-sensitive pathway that 167 strengthens p38-kinase activation (61).

168 For macrophages, high concentration of As2O3 induces cell apoptosis through a mitochondrial- 169 dependent pathway (62,63). It changes their morphology, reduces their phagocytic and adhesive 170 capacity, and decreases macrophagic surface marker expressions on human monocyte-derived 171 macrophages in vitro (37). Interestingly, As2O3 potentiates macrophages’ ability to secrete 172 inflammatory TNF-α, IL-1α, CCL18, and to induce allogeneic or autologous T cell responses (40,64). 173 These in vitro observations are confirmed by ex vivo data (65,66). The functional changes are 174 probably due to superoxide generation, activation of the Rho-kinase/p38-kinase pathway (37), and 175 modulation of the UPR signaling (67). Especially, activating transcription factor 4 protein, a UPR 176 transcription factor, plays a key role in As2O3-mediated regulation of macrophage functions (63). In 177 addition, As2O3 also globally regulates redox-sensitive gene expression in human macrophages 178 (41,46). Therefore, As(III) exerts double-sided effects on macropahges by imparing their clearance 179 capacity while enhancing the pro-inflammatory functions.

180 3.4 Trivalent arsenic and dendritic cells

181 Two types of dendritic cells (DCs), namely conventional dendritic cells (cDC) and plasmacytoid 182 dendritic cells (pDCs), exist in vivo, though MDDCs are broadly used for research in vitro for 183 availability reason. NaAsO2 inhibits dendritic differentiation of monocytes in vitro (68). It also 184 decreases MDDCs viability, maturation, phagocytic capacity, as well as their ability to secrete the 185 pro-inflammatory cytokines IL-12 and IL-23, and to stimulate T helper (Th) cells to secrete IFN-γ 186 (69). The inhibition of IL-12 is mediated by the induced expression of Nrf2 (44). Moreover, As2O3 187 inhibits IFN-α secretion by murine pDCs (15).

188 3.5 Trivalent arsenic and T cells

189 T cells play a central role in cell-mediated immunity. As a typical example, the effects of As2O3 on 190 apoptosis and functions of CD4+ T cell are summarized in Figure 2. High concentrations of As(III) 191 induce apoptosis of T helper cells (CD4+ T cell), cytotoxic T cells (CD8+ T cell), and regulatory T 192 cells (Treg) (12,16,70). Both mitochondrial-mediated and the receptor-mediated pathways are + + 193 involved in T cell apoptosis. Gupta et al. revealed that As2O3 induces apoptosis of CD4 and CD8 T 194 cells via the mitochondrial pathway by enhancing the generation of oxidative stress and by regulating 195 the expression of Bcl-2 family proteins (12). When CD4+ T cells were investigated within human 196 peripheral blood mononuclear cells (PBMC), tumor necrosis factor α released from other + 197 mononuclear cells after NaAsO2 exposure induced CD4 T cell apoptosis through TNF-R1 apoptotic 198 signaling (35).

199 There is different sensitivity to As(III) toxicity among the T cell subsets. CD4+ are more sensitive + 200 than CD8 T cells to the pro-apoptotic toxicity of NaAsO2, as discovered both in vitro and in vivo in

167 Immunomodulation by trivalent arsenic

201 human and mouse model (35,42,71). This difference leads to a reduced CD4+/CD8+ ratio, which is 202 associated with immunodysfunction (72). Tregs are important CD4+ T cells mediating immune 203 tolerance. In vitro tests show that As2O3, at the same concentration, preferentially induces apoptosis 204 of human purified CD4+CD25+ Tregs than CD4+CD25- effector T cells, and decreases the Treg 205 frequency in APL patients’ peripheral blood (73). As2O3-mediated selective Treg depletion is also 206 discovered in vivo in mouse model (16). As2O3 induces Treg apoptosis through ROS and reactive + 207 nitrogen species (RNS) generation, and the differential effect of As2O3 on Treg versus other CD4 208 cells may be related to differences in the cells’ redox status (16). However, some other studies report 209 opposite results (74,75). After exposure to NaAsO2, the mRNA level of the Treg specific 210 transcription factor Foxp3 is upregulated in the spleen and thymus of the experimental mice (42). 211 Interestingly, Tohyama et al report that in mitogen activated human PBMC, 5μM As2O3 decreases 212 the Treg frequency after treatment for 48h and, conversely, increases its frequency after 96h of 213 culture. These results indicate that short-term exposure to As(III) may deplete Treg, in contrast to 214 long-term exposure leading to Treg percentage increase.

215 Upon activation, naïve CD4+ T cells proliferate and differentiate into fully functional effector T cells, 216 which lead to either inflammatory responses or immune suppression (76). It is reported that As(III) 217 exposure inhibits the proliferation of both human and murine activated T cells, by affecting the initial 218 activation step of T cell receptor signaling, and by inhibiting IL-2 expression, at both the protein and 219 mRNA levels (77-80). In addition, NaAsO2 blocks T cells in the G1 phase and, thus, retards the entry 220 into cell cycle (81). More importantly, As(III) significantly alters Th cell differentiation. As2O3 221 inhibits IFN-γ, the characteristic Th1 cytokine, expression at both the mRNA and protein levels 222 during the activation with anti-CD3/anti-CD28 of both human and murine T cells (42,82). However, 223 As(III) do not change, and sometimes increase the secretion of IL-4 and IL-13, the characteristic Th2 224 cytokines, from activated T cells (77,81). These observations indicate that As(III) probably alters the 225 Th1/Th2 balance towards a Th2 response. However, In ATL patients treated with 226 arsenic/IFN/zidovudine, this combination therapy revert the Treg/Th2 cytokine profile at diagnosis 227 towards a Th1 profile, with a diminution of IL-10 (83). This effect is probably due to the depletion of 228 ATL leukemia cells that disrupt the immune system initially. Given that many malignancies go along 229 with immunosuppression, successful As(III) treatment will not augment immune restrain, but rather 230 induce ‘immune-reactivation’ during cancer depletion.

231 Moreover, As(III) is found to be an especially potent inhibitor of IL-17. At a concentration which is 232 not able to affect IFN-γ secretion by Th1 cells, NaAsO2 almost totally blocks the IL-17 secretion by 233 human Th17 cells, via mRNA reduction of the retinoic-related orphan receptor (ROR)C gene which 234 encodes RORγt, the key transcription factor of Th17 (84). Since the Th17/Treg balance is critical in 235 keeping immune homeostasis, As(III) treatment may lead to a regulatory response due to the potent 236 inhibition on Th17.

237 3.6 Trivalent arsenic and B cells

238 Baysan A et al. reveal that As2O3 induces apoptosis of the Bcl-2 negative human B-cell line Ramos 239 via the mitochondrial pathway, by upregulating the expression of pro-apoptotic proteins Bax and Bim 240 (13). Moreover, it also blocks the mitogen-mediated B-cell differentiation towards plasmacytes, and 241 their IgM secretion (85). In addition, As2O3, together with leflunomide, reduces the deposition and 242 expression of IgG and IgM in both the xenograft and recipient sera in a heart xenotransplant model 243 (86). Collectively, As(III) treatment probably leads to a global suppression of the humoral immune 244 system.

168 Immunomodulation by trivalent arsenic

245 3.7 Trivalent arsenic and natural killer cells

246 As2O3 reduces hematopoietic stem cell differentiation to NK cells, leading to a delayed NK cell 247 reconstitution when treating APL (46). As2O3 also facilitates NK cell mediated cytotoxicity towards 248 several cancer cell lines, through modulation of both NK cell receptors and malignant cell ligand 249 profile (46,87). Collectively, As2O3 may enhance the cytotoxicity of NK cells.

250 3.8 Trivalent arsenic and inflammasomes

251 In the human monocyte cell line THP-1 and mouse bone marrow derived macrophages (BMDMs), 252 As2O3 inhibits NLRP3 inflammasome and subsequent IL-1β and IL-18 secretion by targeting PML 253 (34,88). Moreover, As2O3 inhibits NLRP1, NAIP5/NLRC4 inflammasomes in the same subset of 254 cells (89). However, in the human keratinocyte cell line and mouse skin tissue, NaAsO2 promotes IL- 255 1β and IL-18 secretion via AIM2 inflammasome activation (90). In addition, the inflammasome 256 NALP2 polymorphism is associated with arsenic-induced skin lesions in humans (91). Therefore, 257 As(III) exerts bi-directional regulation on inflammasomes, which is probably tissue dependent.

258

259 4 Trivalent arsenic in mouse models of immune-mediate diseases

260 4.1 Autoimmune diseases and asthma

261 Recently, As(III) was found to have therapeutic efficacy in several mouse models of autoimmune and 262 inflammatory diseases (summarized in Table 2). As2O3 achieves quasi-total lesion regression in 263 MRL/lpr mice, a mouse model of systemic lupus erythematosus (SLE) and lymphoproliferative 264 syndrome, by elimination of the unusually activated T lymphocytes, and reduction of the related 265 autoantibodies (14). In the 2,4,6-trinitrobenzene sulfonic acid-induced murine model of inflammatory 266 bowel disease, As2O3 reduces the induced colitis via Nf-kB down-regulation and caspase-3 activation 267 (92). In a murine model of systemic sclerosis constructed by intradermal injections of HOCl, As2O3 268 improves skin and lung fibrosis through ROS-mediated killing of activated fibroblasts (93). 269 Moreover, As2O3 improves the pathologic changes in a rat model of rheumatoid arthritis (RA) 270 through a pro-apoptotic effect on RA fibroblast-like synoviocytes (94). In addition, NaAsO2 271 decreases the disease incidence and delays its onset in non-obese diabetic (NOD) mouse, a mouse 272 model of autoimmune type 1 diabetes, which is due to the reduced proliferation and activation of T 273 cells (95).

274 Asthma is a T-helper type 2 (Th2) lymphocyte-mediated chronic inflammatory disorder characterized 275 by airway eosinophilia and airway hyperresponsiveness (AHR). As2O3 is observed to ameliorate the 276 allergen-driven AHR in a mouse model of asthma, through modulation of the NF-kB pathway, 277 airway eosinophil recruitment and functions (56,57).

278 Overall, As(III) has therapeutic potential on several autoimmune and inflammatory diseases through 279 a pro-apoptotic effect and effector-cell functional modulation.

280 4.2 Allogeneic organ/stem cell transplantation

281 During organ or allogeneic stem cell transplantation, human leukocyte antigen mismatches lead to 282 severe T- and B-cell-mediated alloreactivity, which may eventually lead to severe inflammatory 283 transplantation complications of allograft rejection or graft-versus-host disease (GVHD), respectively. 284 As2O3 is shown to prolong the allograft survival in immunocompetent mouse heart transplant models,

169 Immunomodulation by trivalent arsenic

285 with reduction of the proportions of CD4+ and CD8+ memory T cells, and increase of Tregs in 286 recipient spleen and lymph nodes (96-98). Meanwhile, the IFN-γ expression is reduced and TGF-β 287 expression is increased in both the recipient serum and the graft (97,98). Further studies show that + + 288 As2O3 inhibits accelerated allograft rejection mediated by alloreactive CD8 and/or CD4 memory T 289 cells, and prolongs allograft survival (99,100). Moreover, in two models (one allo- and one xeno-) of 290 islet transplantation, As2O3 is shown to prolong the graft survival by inhibiting inflammatory 291 reactions and T cell responses (101,102). GVHD occurs when immunocompetent T cells in the graft 292 recognize the recipient as foreign, and attacks the target organs. A study reveals that As2O3 prevents 293 disease occurrence in a murine sclerodermatous GVHD model mediated by overproduction of H2O2 294 which kills activated CD4+ T cells and pDCs (15). To summarize, As(III) could be a promising drug 295 for allo-reactivity through effector-cell modulations.

296 4.3 Clinical trials of trivalent arsenic on immune-mediated diseases

297 Given the discoveries in basic research and pre-clinical models, interests grow on As(III) as a clinical 298 therapeutic agent in immune-mediated diseases. An on-going randomized clinical trial 299 (NCT02966301) is currently testing As2O3 as first-line treatment of chronic GVHD. Results of a 300 phase 2a randomized clinical trial (NCT01738360) aiming to evaluate the therapeutic efficacy of 301 As2O3 in SLE are expected soon.

302

303 5 Trivalent arsenic and tumor immunotherapy

304 5.1 Trivalent arsenic facilitates cellular immunotherapy

305 The direct effects of As(III) on tumor cells are not discussed here, but are reviewed by Emadi et al 306 (1). The past decade has witnessed breakthroughs which bring immunotherapy to the forefront of 307 cancer therapy. As(III) acts in cellular immunotherapy in two ways (summarized in Table 3). On the 308 one hand, it selectively depletes tumor promoting cells. The ratio of effector to regulatory T cells 309 (Teff/Treg) is of great importance in immune modulation, and targeting Tregs to enhance anti-tumor 310 immune responses has become an important strategy in cancer immunology (103). It was observed 311 that As2O3 reduces the increased Treg numbers and Foxp3 mRNA levels in ex vivo malignant ascites 312 (104). In two murine models of colon and liver cancer, As2O3 showed its anti-tumor effect in both in- 313 situ carcinoma and colon cancer lung metastasis through selective depletion of the infiltrated Tregs 314 (16,105,106). On the other hand, As(III) enhances cytotoxicity of tumor-killing cells. It has been 315 shown that As2O3 increases the cytotoxic activity of CIK (cytokine-induced killer cells) and the IFN- 316 γ secretion in vitro, which may help in tumor control (105,106). Moreover, exposure of myeloma cell 317 lines to As2O3 increases lymphokine activated killer (LAK)-mediated killing by up-modulation of 318 CD38 and CD54 on the myeloma cells, and increased expression of CD31 (CD38 ligand) and CD11a 319 (CD54 ligand) on LAKs, suggesting that the improved killing is mediated by increased adhesion (17). 320 In addition, exposure of NK and leukemic cells to low doses of As2O3 modulates NK cell receptors 321 and malignant cell ligand profile in a direction that enhances NK cell-mediated cytolytic activity 322 (46,87). This effect was proved in a mouse model of APL, where the As2O3+NK treatment promoted 323 longer survival as compared with As2O3 alone (46). Overall, As2O3 could work as an effective 324 adjuvant for cellular immunotherapy in both hematological malignancies and solid tumors.

325 5.2 Synergistic effect of trivalent arsenic and other strategies of immunotherapy

170 Immunomodulation by trivalent arsenic

326 B7-H3 (CD276) is an important immune checkpoint member of the B7 and CD28 families, which 327 have been shown to induce antitumor immunity. As2O3 synergized with B7H3-mediated 328 immunotherapy to eradicate hepatocellular carcinoma in a mouse model, along with the generation of 329 potent systemic antitumor immunity mediated by CD8+ and NK cells (107). Moreover, As2O3 was 330 also shown in a mouse model of bladder cancer to facilitate intravesical bacillus Calmette-Guerin 331 (BCG) immunotherapy, by targeting the IER3/Nrf2 pathway (108).

332

333 6 Conclusions and perspectives

334 As(III), when used in a clinically relevant dose, can target specific immune cell subsets and act as an 335 immunomodulatory drug. The mechanisms of action include apoptosis induction, protein modulation, 336 and regulation of cellular functions through signaling pathway regulation. The efficacy of As(III) on 337 mouse models of autoimmune and inflammatory diseases highlights its therapeutic potential in 338 humans. It emerges as a promising adjuvant to cancer immunotherapy agents in the treatment of both 339 hematological and solid tumors. Several clinical trials with As(III) in autoimmune and inflammatory 340 diseases are on-going to prove the ‘bench to bedside’ indication of As(III).

341 It is noteworthy that the As(III) effect is dose, cellular and tissue dependent, which requires deeper 342 understanding of this drug in vitro, in vivo and in the clinical practice. For this purpose, 343 comprehensive genomic, proteomic and metabolomic profiling will be critical for identifying and 344 validating potential molecular targets of As(III) for future therapeuticuse.

345

346 7 Acknowledgments

347 The authors acknowledge the Association for Training, Education and Research in Hematology, 348 Immunology and Transplantation for the generous and continuous support to the research work. Y.Y 349 thanks to China Scholarship Council for financial support (CSC No. 201606320257).

350

351 8 Author Contributions Statement

352 YY, BG, MM and FM wrote the manuscript. MM contributed to financial support. All authors 353 contributed substantially to this work and gave final approval of manuscript.

354

355 9 Conflict of Interest Statement

356 The authors declare that the research was conducted in the absence of any commercial or financial 357 relationships that could be construed as a potential conflict of interest.

358

359

360

171 Immunomodulation by trivalent arsenic

361 10 References

362 1. Emadi A, Gore SD. Arsenic trioxide - An old drug rediscovered. Blood Rev (2010) 24:191–9. 363 doi:10.1016/j.blre.2010.04.001.

364 2. Wang ZY, Chen Z. Acute promyelocytic leukemia from highly fatal to highly curable. Blood 365 (2008) 111:2505–15. doi:10.1182/blood-200707-102798.

366 3. Shen ZX, Chen GQ, Ni JH, Li XS, Xiong SM, Qiu QY, et al. Use of Arsenic Trioxide 367 (As2O3) in the Treatment of Acute Promyelocytic Leukemia (APL) II. Clinical Efficacy and 368 Pharmacokinetics in Relapsed Patients. Blood (1997) 89:3354–60.

369 4. Lo-Coco F, Avvisati G, Vignetti M, Thiede C, Orlando SM, Iacobelli S, et al. Retinoic acid 370 and arsenic trioxide for acute promyelocytic leukemia. N Engl J Med (2013) 369:111–21. 371 doi:10.1056/NEJMoa1300874.

372 5. Cicconi L, Lo-Coco F. Current management of newly diagnosed acute promyelocytic 373 leukemia. Ann Oncol (2016) 27:1474–81. doi:10.1093/annonc/mdw171.

374 6. Kchour G, Tarhini M, Kooshyar MM, El Hajj H, Wattel E, Mahmoudi M, et al. Phase 2 study 375 of the efficacy and safety of the combination of arsenic trioxide, interferon alpha, and zidovudine in 376 newly diagnosed chronic adult T-cell leukemia/lymphoma (ATL). Blood (2009) 113:6528–32. 377 doi:10.1182/blood-2009-03-211821.

378 7. El Hajj H, Dassouki Z, Berthier C, Raffoux E, Ades L, Legrand O, et al. Retinoic acid and 379 arsenic trioxide trigger degradation of mutated NPM1, resulting in apoptosis of AML cells. Blood 380 (2015) 125:3447–754. doi:10.1182/blood-2014-11612416.

381 8. Shen S, Li XF, Cullen WR, Weinfeld M, Le XC. Arsenic binding to proteins. Chem Rev 382 (2013) 113:7769–92. doi:10.1021/cr300015c.

383 9. Chou WC, Jie C, Kenedy AA, Jones RJ, Trush MA, Dang C V. Role of NADPH oxidasein 384 arsenic-induced reactive oxygen species formation and cytotoxicity in myeloid leukemia cells. Proc 385 Natl Acad Sci U S A (2004) 101:4578–83. doi:10.1073/pnas.0306687101.

386 10. Lallemand-Breitenbach V, Zhu J, Puvion F, Koken M, Honoré N, Doubeikovsky A, et al. 387 Role of Promyelocytic Leukemia (Pml) Sumolation in Nuclear Body Formation, 11s Proteasome 388 Recruitment, and as 2 O 3 -Induced Pml or Pml/Retinoic Acid Receptor α Degradation. J Exp Med 389 (2001) 193:1361–72. doi:10.1084/jem.193.12.1361.

390 11. Dangleben NL, Skibola CF, Smith MT. Arsenic immunotoxicity: a review. Env Heal (2013) 391 12:73.

392 12. Gupta S, Yel L, Kim D, Kim C, Chiplunkar S, Gollapudi S. Arsenic trioxide induces 393 apoptosis in peripheral blood T lymphocyte subsets by inducing oxidative stress:aroleofBcl-2.Mol 394 Cancer Ther (2003) 2:711–9.

395 13. Baysan A, Yel L, Gollapudi S, Su H, Gupta S. Arsenic trioxide induces apoptosis via the 396 mitochondrial pathway by upregulating the expression of Bax and Bim in human B cells. Int J Oncol 397 (2007) 30:313–8.

172 Immunomodulation by trivalent arsenic

398 14. Bobe P, Bonardelle D, Benihoud K, Opolon P, Chelbi-Alix MK. Arsenic trioxide: A 399 promising novel therapeutic agent for lymphoproliferative and autoimmune syndromes in MRL/lpr 400 mice. Blood (2006) 108:3967–75. doi:10.1182/blood-2006-04-020610.

401 15. Kavian N, Marut W, Servettaz A, Laude H, Nicco C, Chereau C, et al. Arsenic trioxide 402 prevents murine sclerodermatous graft-versus-host disease. J Immunol (2012) 188:5142–9. 403 doi:10.4049/jimmunol.1103538.

404 16. Thomas-Schoemann A, Batteux F, Mongaret C, Nicco C, Chereau C, AnnereauM,etal. 405 Arsenic trioxide exerts antitumor activity through regulatory T cell depletion mediated by oxidative 406 stress in a murine model of colon cancer. J Immunol (2012) 189:5171–7. 407 doi:10.4049/jimmunol.1103094.

408 17. Deaglio S, Canella D, Baj G, Arnulfo A, Waxman S, Malavasi F. Evidence of an 409 immunologic mechanism behind the therapeutical effects of arsenic trioxide (As2O3) on myeloma 410 cells. Leuk Res (2001) 25:227–35.

411 18. Leung J, Pang A, Yuen WH, Kwong YL, Tse EW. Relationship of expression of 412 aquaglyceroporin 9 with arsenic uptake and sensitivity in leukemia cells. Blood (2007) 109:740–6. 413 doi:10.1182/blood-2006-04019588.

414 19. de Thé H. Differentiation therapy revisited. Nat Rev Cancer (2018) 18:117–27. 415 doi:10.1038/nrc.2017.103.

416 20. Chen GQ, Zhu J, Shi XG, Ni JH, Zhong HJ, Si GY, et al. In vitro studies on cellular and 417 molecular mechanisms of arsenic trioxide (As2O3) in the treatment of acute promyelocytic leukemia 418 As2O3 induces NB4 cell apoptosis with downregulation of Bcl-2 expression and modulation of 419 PML-RAR alpha/PML proteins. Blood (1996) 88:1052–61.

420 21. Lallemand-Breitenbach V, de Thé H. PML nuclear bodies from architecture to function. Curr 421 Opin Cell Biol (2018) 52:154–61. doi:10.1016/j.ceb.2018.03.011.

422 22. Sahin U, Ferhi O, Jeanne M, Benhenda S, Berthier C, Jollivet F, et al. Oxidative stress- 423 induced assembly of PML nuclear bodies controls sumoylation of partner proteins. J Cell Biol (2014) 424 204:931–45. doi:10.1083/jcb.201305148.

425 23. Niwa-Kawakita M, Ferhi O, Soilihi H, Le Bras M, Lallemand-Breitenbach V, de The H. PML 426 is a ROS sensor activating p53 upon oxidative stress. J Exp Med (2017) 214:3197–206. 427 doi:10.1084/jem.20160301.

428 24. Jeanne M, Lallemand-Breitenbach V, Ferhi O, Koken M, Le Bras M, Duffort S, et al. 429 PML/RARA Oxidation and Arsenic Binding Initiate the Antileukemia Response of As2O3. Cancer 430 Cell (2010) 18:88–98. doi:10.1016/j.ccr.2010.06.003.

431 25. Fasci D, Anania VG, Lill JR, Salvesen GS. SUMO deconjugation is required for arsenic- 432 triggered ubiquitylation of PML. Sci Signal (2015) 8:ra56. doi:10.1126/scisignal.aaa3929.

433 26. Ablain J, Rice K, Soilihi H, de Reynies A, Minucci S, de The H. Activation of a 434 promyelocytic leukemia-tumor protein 53 axis underlies acute promyelocytic leukemia cure. Nat 435 Med (2014) 20:167–74. doi:10.1038/nm.3441.

173 Immunomodulation by trivalent arsenic

436 27. El-Sabban ME, Nasr R, Dbaibo G, Hermine O, Abboushi N, Quignon F, et al. Arsenic- 437 interferon-α–triggered apoptosis in HTLV-I transformed cells is associated with Tax down-regulation 438 and reversal of NF-κB activation. Blood (2000) 96:2849–55.

439 28. El Hajj H, El-Sabban M, Hasegawa H, Zaatari G, Ablain J, Saab ST, et al. Therapy-induced 440 selective loss of leukemia-initiating activity in murine adult T cell leukemia. J Exp Med (2010) 441 207:2785–92. doi:10.1084/jem.20101095.

442 29. Dassouki Z, Sahin U, El Hajj H, Jollivet F, Kfoury Y, Lallemand-Breitenbach V, et al. ATL 443 response to arsenic/interferon therapy is triggered by SUMO/PML/RNF4-dependent Tax degradation. 444 Blood (2015) 125:474–82. doi:10.1182/blood-2014-04572750.

445 30. Martelli MP, Gionfriddo I, Mezzasoma F, Milano F, Pierangeli S, Mulas F, et al. Arsenic 446 trioxide and all-trans retinoic acid target NPM1 mutant oncoprotein levels and induce apoptosis in 447 NPM1-mutated AML cells. Blood (2015) 125:3455–65. doi:10.1182/blood-2014-11611459.

448 31. Hsu KS, Kao HY. PML: Regulation and multifaceted function beyond tumor suppression. 449 Cell Biosci (2018) 8:5. doi:10.1186/s13578-018-0204-8.

450 32. Scherer M, Stamminger T. Emerging Role of PML Nuclear Bodies in Innate Immune 451 Signaling. J Virol (2016) 90:5850–4. doi:10.1128/JVI.01979-15.

452 33. Kim YE, Ahn JH. Positive Role of Promyelocytic Leukemia Protein in Type I Interferon 453 Response and Its Regulation by Human Cytomegalovirus. PLoS Pathog (2015) 11:1–24. 454 doi:10.1371/journal.ppat.1004785.

455 34. Lo YH, Huang YW, Wu YH, Tsai CS, Lin YC, Mo ST, et al. Selective inhibition of the 456 NLRP3 inflammasome by targeting to promyelocytic leukemia protein in mouse and human. Blood 457 (2013) 121:3185–94. doi:10.1182/blood-2012-05432104.

458 35. Yu H-S, Chen G-S, Liao W-T, Chang K-L, Yu C-L. Arsenic Induces Tumor Necrosis Factor 459 α Release and Tumor Necrosis Factor Receptor 1 Signaling in T Helper Cell Apoptosis. J Invest 460 Dermatol (2002) 119:812–9. doi:10.1046/j.1523-1747.2002.00475.x.

461 36. Lemarie A, Morzadec C, Merino D, Micheau O, Fardel O, Vernhet L. Arsenic trioxide 462 induces apoptosis of human monocytes during macrophagic differentiation through nuclear factor- 463 kappaB-related survival pathway down-regulation. J Pharmacol Exp Ther (2006) 316:304–14. 464 doi:10.1124/jpet.105.092874.

465 37. Lemarie A, Morzadec C, Bourdonnay E, Fardel O, Vernhet L. Human Macrophages 466 Constitute Targets for Immunotoxic Inorganic Arsenic. J Immunol (2006) 177:3019–27. 467 doi:10.4049/jimmunol.177.5.3019.

468 38. Schieber M, Chandel NS. ROS function in redox signaling and oxidative stress. Curr Biol 469 (2014) 24:R453-62. doi:10.1016/j.cub.2014.03.034.

470 39. Singh N, Kumar D, Lal K, Raisuddin S, Sahu AP. Adverse health effects due to arsenic 471 exposure: modification by dietary supplementation of jaggery in mice. Toxicol Appl Pharmacol 472 (2010) 242:247–55. doi:10.1016/j.taap.2009.10.014.

174 Immunomodulation by trivalent arsenic

473 40. Lemarie A, Bourdonnay E, Morzadec C, Fardel O, Vernhet L. Inorganic Arsenic Activates 474 Reduced NADPH Oxidase in Human Primary Macrophages through a Rho Kinase/p38 Kinase 475 Pathway. J Immunol (2008) 180:6010–7. doi:10.4049/jimmunol.180.9.6010.

476 41. Bourdonnay E, Morzadec C, Fardel O, Vernhet L. Redox-sensitive regulation of gene 477 expression in human primary macrophages exposed to inorganic arsenic. J Cell Biochem (2009) 478 107:537–47. doi:10.1002/jcb.22155.

479 42. Duan X, Gao S, Li J, Wu L, Zhang Y, Li W, et al. Acute arsenic exposure induces 480 inflammatory responses and CD4(+) T cell subpopulations differentiation in spleen and thymus with 481 the involvement of MAPK, NF-kB, and Nrf2. Mol Immunol (2017) 81:160–72. 482 doi:10.1016/j.molimm.2016.12.005.

483 43. Morzadec C, Macoch M, Sparfel L, Kerdine-Romer S, Fardel O, Vernhet L. Nrf2 expression 484 and activity in human T lymphocytes: stimulation by T cell receptor activation and priming by 485 inorganic arsenic and tert-butylhydroquinone. Free Radic Biol Med (2014) 71:133–45. 486 doi:10.1016/j.freeradbiomed.2014.03.006.

487 44. Macoch M, Morzadec C, Genard R, Pallardy M, Kerdine-Romer S, Fardel O, etal.Nrf2- 488 dependent repression of interleukin-12 expression in human dendritic cells exposed to inorganic 489 arsenic. Free Radic Biol Med (2015) 88:381–90. doi:10.1016/j.freeradbiomed.2015.02.003.

490 45. Malloy MT, McIntosh DJ, Walters TS, Flores A, Goodwin JS, Arinze IJ. Trafficking of the 491 transcription factor Nrf2 to promyelocytic leukemia-nuclear bodies: Implications for degradation of 492 nrf2 in the nucleus. J Biol Chem (2013) 288:14569–83. doi:10.1074/jbc.M112.437392.

493 46. Bourdonnay E, Morzadec C, Sparfel L, Galibert MD, Jouneau S, Martin-ChoulyC,etal. 494 Global effects of inorganic arsenic on gene expression profile in human macrophages. Mol Immunol 495 (2009) 46:649–56. doi:10.1016/j.molimm.2008.08.268.

496 47. Binet F, Girard D. Novel human neutrophil agonistic properties of arsenic trioxide: 497 involvement of p38 mitogen-activated protein kinase and/or c-jun NH2-terminal MAPK but not 498 extracellular signal-regulated kinases-1/2. J Leukoc Biol (2008) 84:1613–22. 499 doi:10.1189/jlb.0708421.

500 48. Alex AA, Ganesan S, Palani HK, Balasundaram N, David S, Lakshmi KM, et al.Arsenic 501 Trioxide Enhances the NK Cell Cytotoxicity Against Acute Promyelocytic Leukemia While 502 Simultaneously Inhibiting Its Bio-Genesis. Front Immunol (2018) 9:1357. 503 doi:10.3389/fimmu.2018.01357.

504 49. Binet F, Cavalli H, Moisan E, Girard D. Arsenic trioxide (AT) is a novel human neutrophil 505 pro-apoptotic agent: effects of catalase on AT-induced apoptosis, degradation of cytoskeletal proteins 506 and de novo protein synthesis. Br J Haematol (2006) 132:349–58. doi:10.1111/j.1365- 507 2141.2005.05866.x.

508 50. Antoine F, Ennaciri J, Girard D. Syk is a novel target of arsenic trioxide (ATO) and is 509 involved in the toxic effect of ATO in human neutrophils. Toxicol Vitr (2010) 24:936–41. 510 doi:10.1016/j.tiv.2009.11.011.

175 Immunomodulation by trivalent arsenic

511 51. Binet F, Chiasson S, Girard D. Arsenic trioxide induces de novo protein synthesis of annexin- 512 1 in neutrophils: association with a heat shock-like response and not apoptosis. Br J Haematol (2008) 513 140:454–63. doi:10.1111/j.1365-2141.2007.06941.x.

514 52. Binet F, Chiasson S, Girard D. Evidence that endoplasmic reticulum (ER) stress and caspase- 515 4 activation occur in human neutrophils. Biochem Biophys Res Commun (2010) 391:18–23. 516 doi:10.1016/j.bbrc.2009.10.141.

517 53. Binet F, Chiasson S, Girard D. Arsenic trioxide induces endoplasmic reticulum stress-related 518 events in neutrophils. Int Immunopharmacol (2010) 10:508–12. doi:10.1016/j.intimp.2010.01.013.

519 54. Wei Z, Zhang X, Wang J, Wang Y, Yang Z, Fu Y. The formation of canine neutrophil 520 extracellular traps induced by sodium arsenic in polymorphonuclear neutrophils. Chemosphere (2018) 521 196:297–302. doi:10.1016/j.chemosphere.2017.12.175.

522 55. Li J, Zhao L, Zhang Y, Li W, Duan X, Chen J, et al. Imbalanced immune responses involving 523 inflammatory molecules and immune-related pathways in the lung of acute and subchronic arsenic- 524 exposed mice. Env Res (2017) 159:381–93. doi:10.1016/j.envres.2017.08.036.

525 56. Zhou LF, Zhu Y, Cui XF, Xie WP, Hu AH, Yin KS. Arsenic trioxide, a potent inhibitor of 526 NF-kappaB, abrogates allergen-induced airway hyperresponsiveness and inflammation. Respir Res 527 (2006) 7:146. doi:10.1186/1465-9921-7-146.

528 57. Chu KH, Lee CC, Hsin SC, Cai BC, Wang JH, Chiang BL. Arsenic trioxide alleviates airway 529 hyperresponsiveness and eosinophilia in a murine model of asthma. Cell Mol Immunol (2010) 530 7:375–80. doi:10.1038/cmi.2010.26.

531 58. Hutchinson LM, Trinh BM, Palmer RK, Preziosi CA, Pelletier JH, Nelson HM, et al. 532 Inorganic arsenite inhibits IgE receptor-mediated degranulation of mast cells. J Appl Toxicol (2011) 533 31:231–41. doi:10.1002/jat.1585.

534 59. Shim J, Kennedy RH, Weatherly LM, Hutchinson LM, Pelletier JH, Hashmi HN, et al. 535 Arsenic inhibits mast cell degranulation via suppression of early tyrosine phosphorylation events. J 536 Appl Toxicol (2016) 36:1446–59. doi:10.1002/jat.3300.

537 60. Guilliams M, Mildner A, Yona S. Developmental and Functional Heterogeneity of 538 Monocytes. Immunity (2018) 49:595–613. doi:10.1016/j.immuni.2018.10.005.

539 61. Bourdonnay E, Morzadec C, Fardel O, Vernhet L. Arsenic increases lipopolysaccharide- 540 dependent expression of interleukin-8 gene by stimulating a redox-sensitive pathway that strengthens 541 p38-kinase activation. Mol Immunol (2011) 48:2069–78. doi:10.1016/j.molimm.2011.06.443.

542 62. Sengupta M, Bishayi B. Effect of lead and arsenic on murine macrophage response. Drug 543 Chem Toxicol (2002) 25:459–72. doi:10.1081/DCT-120014796.

544 63. Srivastava RK, Li C, Wang Y, Weng Z, Elmets CA, Harrod KS, et al. Activating transcription 545 factor 4 underlies the pathogenesis of arsenic trioxide-mediated impairment of macrophage innate 546 immune functions. Toxicol Appl Pharmacol (2016) 308:46–58. doi:10.1016/j.taap.2016.07.015.

176 Immunomodulation by trivalent arsenic

547 64. Sakurai T, Ohta T, Fujiwara K. Inorganic arsenite alters macrophage generation from human 548 peripheral blood monocytes. Toxicol Appl Pharmacol (2005) 203:145–53. 549 doi:10.1016/j.taap.2004.08.003.

550 65. Banerjee N, Banerjee S, Sen R, Bandyopadhyay A, Sarma N, Majumder P, et al. Chronic 551 arsenic exposure impairs macrophage functions in the exposed individuals. J Clin Immunol (2009) 552 29:582–94. doi:10.1007/s10875-009-9304-x.

553 66. Bishayi B, Sengupta M. Intracellular survival of Staphylococcus aureus due to alteration of 554 cellular activity in arsenic and lead intoxicated mature Swiss albino mice. Toxicology (2003) 555 184:31–9.

556 67. Srivastava RK, Li C, Chaudhary SC, Ballestas ME, Elmets CA, Robbins DJ, et al. Unfolded 557 protein response (UPR) signaling regulates arsenic trioxide-mediated macrophage innate immune 558 function disruption. Toxicol Appl Pharmacol (2013) 272:879–87. doi:10.1016/j.taap.2013.08.004.

559 68. Bahari A, Salmani V. Environmentally relevant dose of arsenic interferes in functions of 560 human monocytes derived dendritic cells. Toxicol Lett (2017) 275:118–22. 561 doi:10.1016/j.toxlet.2017.05.005.

562 69. Macoch M, Morzadec C, Fardel O, Vernhet L. Inorganic arsenic impairs differentiation and 563 functions of human dendritic cells. Toxicol Appl Pharmacol (2013) 266:204–13. 564 doi:10.1016/j.taap.2012.11.008.

565 70. Tenorio EP, Saavedra R. Differential effect of sodium arsenite during the activation of human 566 CD4+ and CD8+ T lymphocytes. Int Immunopharmacol (2005) 5:1853–69. 567 doi:10.1016/j.intimp.2005.06.006.

568 71. Vega L, Montes de Oca P, Saavedra R, Ostrosky-Wegman P. Helper T cell subpopulations 569 from women are more susceptible to the toxic effect of sodium arsenite in vitro. Toxicology (2004) 570 199:121–8. doi:10.1016/j.tox.2004.02.012.

571 72. Lu W, Mehraj V, Vyboh K, Cao W, Li T, Routy JP. CD4:CD8 ratio as a frontier marker for 572 clinical outcome, immune dysfunction and viral reservoir size in virologically suppressed HIV- 573 positive patients. J Int AIDS Soc (2015) 18:20052. doi:10.7448/IAS.18.1.20052.

574 73. Xu W, Li X, Quan L, Yao J, Mu G, Guo J, et al. Arsenic trioxide decreases the amount and 575 inhibits the function of regulatory T cells, which may contribute to its efficacy in the treatment of 576 acute promyelocytic leukemia. Leuk Lymphoma (2018) 59:650–9. 577 doi:10.1080/10428194.2017.1346253.

578 74. ZhaoJJ,YangSW,ZhangF,YuanXL,ZM.Z,FangBJ,etal.As2O3Up-Regulates the 579 Proportion of CD4+CD25+CD127low Tregs in Peripheral Blood of Patients with Severe Aplastic 580 Anemia. Zhongguo Shi Yan Xue Ye Xue Za Zhi (2018) 26:854–8.

581 75. Tohyama N, Tanaka S, Onda K, Sugiyama K, Hirano T. Influence of anticancer agents on cell 582 survival, proliferation, and CD4+CD25+Foxp3+ regulatory T cell-frequency in human peripheral- 583 blood mononuclear cells activated by T cell-mitogen. Int Immunopharmacol (2013) 15:160–6. 584 doi:10.1016/j.intimp.2012.11.008.

177 Immunomodulation by trivalent arsenic

585 76. Wan YY, Flavell RA. How diverse--CD4 effector T cells and their functions. J Mol Cell Biol 586 (2009) 1:20–36. doi:10.1093/jmcb/mjp001.

587 77. Soto-Pena GA, Vega L. Arsenic interferes with the signaling transduction pathway of T cell 588 receptor activation by increasing basal and induced phosphorylation of Lck and Fyn in spleen cells. 589 Toxicol Appl Pharmacol (2008) 230:216–26. doi:10.1016/j.taap.2008.02.029.

590 78. Morzadec C, Bouezzedine F, Macoch M, Fardel O, Vernhet L. Inorganic arsenic impairs 591 proliferation and cytokine expression in human primary T lymphocytes. Toxicology (2012) 300:46– 592 56. doi:10.1016/j.tox.2012.05.025.

593 79. Vega L, Ostrosky-Wegman P, Fortoul TI, Diaz C, Madrid V, Saavedra R. Sodium arsenite 594 reduces proliferation of human activated T-cells by inhibition of the secretion of interleukin-2. 595 Immunopharmacol Immunotoxicol (1999) 21:203–20. doi:10.3109/08923979909052758.

596 80. Conde P, Acosta-Saavedra LC, Goytia-Acevedo RC, Calderon-Aranda ES. Sodium arsenite- 597 induced inhibition of cell proliferation is related to inhibition of IL-2 mRNA expression in mouse 598 activated T cells. Arch Toxicol (2007) 81:251–9. doi:10.1007/s00204-006-0152-7.

599 81. Galicia G, Leyva R, Tenorio EP, Ostrosky-Wegman P, Saavedra R. Sodium arsenite retards 600 proliferation of PHA-activated T cells by delaying the production and secretion of IL-2. Int 601 Immunopharmacol (2003) 3:671–82. doi:10.1016/s1567-5769(03)00049-3.

602 82. VanDenBerg KR, Freeborn RA, Liu S, Kennedy RC, Zagorski JW, Rockwell CE. Inhibition 603 of early T cell cytokine production by arsenic trioxide occurs independently of Nrf2. PLoS One 604 (2017) 12:e0185579. doi:10.1371/journal.pone.0185579.

605 83. Kchour G, Rezaee SAR, Farid R, Ghantous A, Rafatpanah H, Tarhini M, et al.The 606 combination of arsenic, interferon-alpha, and zidovudine restores an “ immunocompetent-like” 607 cytokine expression profile in patients with adult T-cell leukemia lymphoma. Retrovirology (2013) 608 10:1–12. doi:10.1186/1742-4690-10-91.

609 84. Morzadec C, Macoch M, Robineau M, Sparfel L, Fardel O, Vernhet L. Inorganic arsenic 610 represses interleukin-17A expression in human activated Th17 lymphocytes. Toxicol Appl 611 Pharmacol (2012) 262:217–22. doi:10.1016/j.taap.2012.05.004.

612 85. Rousselot P, Labaume S, Marolleau JP, Larghero J, Noguera MH, Brouet JC, et al. Arsenic 613 trioxide and melarsoprol induce apoptosis in plasma cell lines and in plasma cells from myeloma 614 patients. Cancer Res (1999) 59:1041–8.

615 86. Jiao ZX, Leng Y, Xia JJ, Wu HQ, Jin N, Fu JZ, et al. As2 O3 combined with leflunomide 616 prolongs heart xenograft survival via suppressing the response of Th1, Th2, and B cells in a rat 617 model. Xenotransplantation (2016) 23:237–48. doi:10.1111/xen.12238.

618 87. Kim JY, Bae JH, Lee SH, Lee EY, Chung BS, Kim SH, et al. Induction of NKG2D ligands 619 and subsequent enhancement of NK cell-mediated lysis of cancer cells by arsenic trioxide. J 620 Immunother (2008) 31:475–86.

178 Immunomodulation by trivalent arsenic

621 88. Ahn H, Kim J, Kang SG, Yoon SI, Ko HJ, Kim PH, et al. Mercury and arsenic attenuate 622 canonical and non-canonical NLRP3 inflammasome activation. Sci Rep (2018) 8:13659. 623 doi:10.1038/s41598-018-31717-7.

624 89. Maier NK, Crown D, Liu J, Leppla SH, Moayeri M. Arsenic trioxide and other arsenical 625 compounds inhibit the NLRP1, NLRP3, and NAIP5/NLRC4 inflammasomes. J Immunol (2014) 626 192:763–70. doi:10.4049/jimmunol.1301434.

627 90. Zhang M, Qi Y, Li H, Cui J, Dai L, Frank JA, et al. AIM2 inflammasome mediates Arsenic- 628 induced secretion of IL-1 beta and IL-18. Oncoimmunology (2016) 5:e1160182. 629 doi:10.1080/2162402X.2016.1160182.

630 91. Bhattacharjee P, Das N, Chatterjee D, Banerjee A, Das JK, Basu S, et al. Association of 631 NALP2 polymorphism with arsenic induced skin lesions and other health effects. Mutat Res (2013) 632 755:1–5. doi:10.1016/j.mrgentox.2013.04.010.

633 92. Singer M, Trugnan G, Chelbi-Alix MK. Arsenic trioxide reduces 2,4,6-trinitrobenzene 634 sulfonic acid-induced murine colitis via nuclear factor-kappaB down-regulation and caspase-3 635 activation. Innate Immun (2011) 17:365–74. doi:10.1177/1753425910371668.

636 93. Kavian N, Marut W, Servettaz A, Nicco C, Chereau C, Lemarechal H, et al. Reactive oxygen 637 species-mediated killing of activated fibroblasts by arsenic trioxide ameliorates fibrosis in a murine 638 model of systemic sclerosis. Arthritis Rheum (2012) 64:3430–40. doi:10.1002/art.34534.

639 94. Mei Y, Zheng Y, Wang H, Gao J, Liu D, Zhao Y, et al. 640 Arsenic trioxide induces apoptosis of fibroblast-like synoviocytes and represents antiarthris effect in 641 experimental model of rheumatoid arthritis. J Rheumatol (2011) 38:36–43.

642 95. Lee YS, Kim D, Lee EK, Kim S, Choi CS, Jun HS. Sodium meta-arsenite prevents the 643 development of autoimmune diabetes in NOD mice. Toxicol Appl Pharmacol (2015) 284:254–61. 644 doi:10.1016/j.taap.2014.12.016.

645 96. Yan S, Zhang QY, Zhou B, Xue L, Chen H, Wang Y, et al. Arsenic trioxide attenuated the 646 rejection of major histocompatibility complex fully-mismatched cardiac allografts in mice. Transpl 647 Proc (2009) 41:1855–8. doi:10.1016/j.transproceed.2009.02.073.

648 97. Xu S, Chen J, Wang F, Kang X, Lan T, Wang F, et al. Arsenic trioxide combinedwithco- 649 stimulatory molecule blockade prolongs survival of cardiac allografts in alloantigen-primed mice. 650 Transpl Immunol (2010) 24:57–63. doi:10.1016/j.trim.2010.07.003.

651 98. Lin Y, Dai H, Su J, Yan G, Xi Y, Ekberg H, et al. Arsenic trioxide is a novel agent for 652 combination therapy to prolong heart allograft survival in allo-primed T cells transferred mice. 653 Transpl Immunol (2011) 25:194–201. doi:10.1016/j.trim.2011.08.002.

654 99. Yan G, Xi Y, Xu S, Lin Y, Chen J, Dai H, et al. Inhibition of accelerated rejection mediated 655 by alloreactive CD4(+) memory T cells and prolonged allograft survival by arsenic trioxide. 656 Immunol Invest (2013) 42:438–54. doi:10.3109/08820139.2013.801986.

179 Immunomodulation by trivalent arsenic

657 100. Li C, Guan T, Gao C, Lin Y, Yan G, Zhu M, et al. Arsenic trioxide inhibits accelerated 658 allograft rejection mediated by alloreactive CD8(+) memory T cells and prolongs allograft survival 659 time. Transpl Immunol (2015) 33:30–6. doi:10.1016/j.trim.2015.05.004.

660 101. Gao C, Jiang J, Ma P, Cheng P, Lian Y, Zhao B, et al. Arsenic Trioxide Induces T Cell 661 Apoptosis and Prolongs Islet Allograft Survival in Mice. Transplantation (2015) 99:1796–806. 662 doi:10.1097/TP.0000000000000735.

663 102. Zhao B, Xia JJ, Wang LM, Gao C, Li JL, Liu JY, et al. Immunosuppressive effect of arsenic 664 trioxide on islet xenotransplantation prolongs xenograft survival in mice. Cell Death Dis (2018) 665 9:408. doi:10.1038/s41419-018-0446-8.

666 103. Roychoudhuri R, Eil RL, Restifo NP. The interplay of effector and regulatory T cells in 667 cancer. Curr Opin Immunol (2015) 33:101–11. doi:10.1016/j.coi.2015.02.003.

668 104. Hu Z, Hu S, Wu Y, Li S, He C, Xing X, et al. Accumulation and suppressive function of 669 regulatory T cells in malignant ascites: Reducing their suppressive function using arsenic trioxide in 670 vitro. Oncol Lett (2018) 15:5384–90. doi:10.3892/ol.2018.7974.

671 105. Wang L, Hu X, Xu Y, Liu Z. Arsenic trioxide inhibits lung metastasis of mouse colon cancer 672 via reducing the infiltration of regulatory T cells. Tumour Biol (2016) 37:15165–73. 673 doi:10.1007/s13277-016-5377-3.

674 106. Wang L, Wang R, Fan L, Liang W, Liang K, Xu Y, et al. Arsenic trioxide is an immune 675 adjuvant in liver cancer treatment. Mol Immunol (2017) 81:118–26. 676 doi:10.1016/j.molimm.2016.12.001.

677 107. Luo L, Qiao H, Meng F, Dong X, Zhou B, Jiang H, et al. Arsenic trioxide synergizes with 678 B7H3-mediated immunotherapy to eradicate hepatocellular carcinomas. Int J Cancer (2006) 679 118:1823–30. doi:10.1002/ijc.21557.

680 108. Mao MH, Huang HB, Zhang XL, Li K, Liu YL, Wang P. Additive antitumor effect of arsenic 681 trioxide combined with intravesical bacillus Calmette-Guerin immunotherapy against bladder cancer 682 through blockade of the IER3/Nrf2 pathway. Biomed Pharmacother (2018) 107:1093–103. 683 doi:10.1016/j.biopha.2018.08.057.

684

685

686

687

688

689

690

691

180 Immunomodulation by trivalent arsenic

692

693 Table 1. Trivalent arsenic induces immune-mediated cell apoptosis: sensitivity and mechanisms 694 of action

Cell subtype Species Type of EC50 (culture time) Possible mechanism Refs As(III)

PBMC Human NaAsO2 5mol/L (48h and 72h) N/A (35)

Neutrophil Human As2O3 <5mol/L (22h) 1. H2O2 generation (49,50 ) 2. Activation of caspases 3. De novo protein synthesis 4. Syk kinase activation Monocyte

Monocyte (during Human As2O3 ~1mol/L (6d) 1. Inhibition of NF-kB related (36) macrophagic differentiation survival pathways with GM-CSF or M-CSF)

Promonocytic U937 cells Human As2O3 >4mol/L (4d) Dendritic cell (DC)

Monocyte-derived DCs Human NaAsO2 >2mol/L (6d) N/A (69) Tcell

Primary CD4+ T cell Human As2O3 >5mol/L (48h) 1. ROS generation (12)

Primary CD8+ T cell Human As2O3 >5mol/L (48h) 2. Regulation of the Bcl-2 family proteins

CD4+ T cell within PBMCs Human NaAsO2 N/A 1. TNF-R1 apoptotic signaling (35)

CD4+CD25+ Treg Human As2O3 ~1mol/L (38h) 1. ROS and RNS accumulation (16) 2. High sensitivity to oxydative stress

CD4+CD25- effector T cell Human As2O3 ~2.5mol/L (38h) N/A (16) B cell

Ramos cell line Human As2O3 >10mol/L (24h) 1. ROS generation (13) 2. Upregulation of Bax and Bim expression 695

696 Abbreviations: As(III), trivalent arsenic; EC50, half maximal effective concentration; PBMC, 697 peripheral blood mononuclear cell; NaAsO2, sodium arsenite; As2O3, arsenic trioxide; H2O2, 698 hydrogen peroxide; GM-CSF, granulocyte-macrophage colony-stimulatingfactor;M-CSF, 699 macrophage colony-stimulating factor; NF-kB, nuclear factor-kappa B; ROS, reactive oxygen species; 700 RNS, reactive nitrogen species; 701

181 Immunomodulation by trivalent arsenic

Table 2. Use of trivalent arsenic in mouse models of immune-mediated diseases

Investigated disease Mouse strain Optimal therapeutic regime Clinical efficacy Possible mechanism Refs

Autoimmune diseases

SLE/lymphoproliferative MRL/lpr mice As2O3 (i.p) 5g/g/d for 2 1.Prolong survival 1. Anti-DNA autoantibody↓ RF↓IL-18↓ IFN- (14) syndrome months starting from 2 months γ↓nitric oxide metabolite↓TNF-α↓ Fas of age 2.Quasi-total regression of ligand↓IL-10↓ in serum antibody- and cell-mediated manifestations 2. Reduction of immune-complex deposits in glomeruli

Inflammatory bowel diseases TNBS-induced colitis murine As2O3 (i.p) 1.Prolong survival 1. Nf-kB inhibition (92) 182 model (BALB/c) Prevention: 5g/g/d from 4d 2.Reduced the induced colitis 2. TNF-α↓ IL-1β↓ IL-12↓IL-17↓IL-18↓IL-23↓ before TNBS adminitration as assessed by macroscopic expressions in colonic extracts and microscopic scores Treatment: 5g/g/d when the 3. Elimination of inflamed cells through disease was noted apoptosis

Systemic sclerosis SSc induced by As2O3 (i.p) 5g/g/d for 6 1. Reduction in skin and lung 1. ROS generation that selectively kills (93) weeks, simultaneously with fibrosis activated fibroblasts intradermal injections of HOCl HOCl injection (BALB/c) 2. prevention of endothelial 2. Autoantibody↓ IL-4↓ and IL-13↓ from injuries activated T cells

Rheumatoid arthritis Collagen-induced arthritis rat As2O3 (N/A) Inhibition of synovial 1. Apoptosis induction of RA fibroblast-like (94) model hyperplasia and inflammation synoviocytes through NF-ĸB signaling pathway and caspase cascade

Type I diabetes Non-obese diabetic mice NaAsO2 (oral) 5ug/g/d for 8 1. Decrease of incidence and 1. Reduction of infiltration of immunocytes in (95) weeks from 8 weeks of age disease onset delay islets Immunomodulation by trivalent arsenic

2. Inhibition of T cell proliferation/activation

Airway chronic inflammatory disorder

Asthma Chicken OVA-challenged murine As2O3 (i.p) 4g/g/d for 7 days, 1.Amelioration of the allergen- 1. Attenuation of airway eosinophils (56,5 model of asthma (BALB/c) driven airway chemotaxin and recruitment in bronchoalveolar 7) 30 min before each challenge hyperresponsiveness lavage fluid (BALF)

2. IkBα expression increase and NF-ĸB decrease

Allograft rejection AND graft-versus-host disease (GVHD)

Cardiac allograft rejection Model 1.C3H→C57BL/6 Model 1. As2O3 (i.p) (1 1. Prolong allograft survival 1. Inhibition of graft lymphocyte infiltration (96,9 g/g/d, days -3- 10)+CsA 7,98) Model 2.BALB/c→C57BL/6 2. IFN-γ↓ IL-2↓ TGF-β↑ in both Model 2. As2O3 (i.p) (5

183 Model 3. BALB/c→C57BL/6 g/g/d, days 0- 10) + (anti- recipient serum and the graft (allo-primed T cells pre- CD154 and anti-LFA-1） transferred) 3. CD4+ Tm(%)↓ CD8+ Tm(%)↓ but Model 2. As2O3 (i.p) (3 Treg(%)↑ in the spleen of recipients g/g/d, days 0- 10) + (anti- CD154 and anti-LFA-1

Cardiac allograft rejection Model 1. C57BL/6 → nude Model 1&2. As2O3 (i.p) (3 1. Prolong accelerated allograft 1. CD4+ Tm(%)↓and CD8+ Tm(%)↓in (99,1 mouse (CD4+ Tm pre- g/g/d, days 0-10) rejection mediated by CD4 and recipient spleen and lymph nodes 00) transferred) CD8 memory T cells 2. IFN-γ↓ IL-2↓ and TGF-β↑IL-10↑in both Model 2.C57BL/6 → nude mouse recipient serum and allograft (allo-primed CD8+ Tm pre- transferred)

Chronic GVHD B10.D2→BALB/c As2O3 (i.p) (5 g/g/d, 3 weeks 1.Reduction of the fibrotic 1. Activated CD4+ T cells(%)↓ and pDCs(%) (13) from day 7 post changes and prevention of the ↓ in splenocytes through apoptosis transplantation) clinical symptoms 2. Splenic IL-4↓,IL-17↓ and serum Anti- Immunomodulation by trivalent arsenic

DNA-topoisomerase 1 autoantibody↓

702

703

704 Abbreviations: SLE, systemic lupus erythematosus; As2O3, arsenic trioxide; RF, rheumatoid factor; TNBS, 2,4,6-trinitrobenzene sulfonic 705 acid; HOCl, hypochlorous acid; SSc, systemic sclerosis; ROS, reactive oxygen species; RA, rheumatoid arthritis; NaAsO2: sodium arsenite; 706 NF-ĸB, nuclear factor-kappa B; OVA, ovalbumin; IkBα,nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha; 707 Tm, memroy T cell; pDC, plasmacytoid dendritic cell;

708

709 Table 3. Trivalent arsenic facilitates tumor immunotherapy 184

Investigated disease Ex vivo sample/mouse model Optimal therapeutic regime Clinical Efficacy Possible mechanism Refs

Cellular immunotherapy

Malignant ascites TILs derived from ascites of gastric As2O3 (invitro,5or10M N/A 1. CD4+ T cell (%)↓CD8+ T cell (%)↑Treg (104) cancer patients for 48h) (%)↓with Foxp3 expression↓in TILs

2.IL-10↓TGF-β↓IFN-γ↑TIL cytotoxicity↑

Colon cancer Model 1. Colon-cancer bearing mouse Model 1. As2O3 (i.p) 1.Delay tumor growth 1.Selective Treg depletion with Foxp3 (16,10 (BALB/c, injection s.c. of CT-26 colon 1mg/kg once at day 10 of expression↓in spleen and tumor tissue, due 5) cancer cell) tumor injection 2.Inhibition of lung metastasis to ROS and RNS accumulation in Tregs. and prolong mouse survival Model 2. Lung metastasis model of Model 2. As2O3 (i.v) 2. Cytokine induced killer cells (in vitro): colon cancer (BALB/c, injection i.v. of 6mg/kg every 2 days for 2 cytotoxicity↑IFN-γ↑ CT-26) weeks from day 3 of tumor injection Immunomodulation by trivalent arsenic

Hepatic cancer Hepatic cancer bearing mouse (KM As2O3 (i.v) 1. Prolong mouse survival 1. Selective Treg depletion (106) strain, H22 hepatic cancer cell liver implantation) 6mg/kg every 2 days for 2 2. Inhibition of tumor growth 2. Serum IFN-γ↑IL-10↓TGF-β↓ weeks right after liver tumor implantation 3. Reduction of abdominal 3. IOD of CD3+T cell↑Foxp3+ cell↓in adhesion and ascites tumor

Multiple myeloma Co-culture of LAK and myeloma cell As2O3 (in vitro, 0.5Mfor N/A 1. Increase of LAK-mediated lysis via up- (17) 72h) regulation of expressions of CD54 and CD38 on myeloma cells, and CD31(CD38 ligand) and CD11a (CD54 ligand) on LAKs

Acute promyelocytic APL bearing mouse model (FVB/N, As2O3 (i.p) (5 µg/g/d, day 8 1. Prolong survival comparing 1. Enhancement of NK cell cytolytic (48,87 leukemia APL blast cells from MRP8-PML- to day 35 of tumor As2O3+NK group and As2O3 activity ) RARa transgenic FVB/N mouse) injection) AND group alone 2. Alteration of NK cell receptor and ligand Syngeneic NK cells (i.v) Profile on tumor cell (5*10^5 cells/dose, day 8,

185 18, 28 of tumor injection)

Other strategies of immunotherapy

Hepatic cancer Hepatic cancer bearing mouse Intratumoral injection of 1.Transplanted hepatoma 1. Generation of anti-tumor immunity relies (107) (BALB/c, H22 hepatic cancer cell) 100µgB7H3plasmid(on eradication and regression of largely on CD8+ T cells and NK cells day 14 of tumor injection) distant tumor nodules AND 5µg As2O3 every 2 2. Serum IFN-γ↑and CTL activity↑ days until tumor disappear

Bladder cancer Orthotopic bladder cancer model (C3H Intravesical instillation of 1. Reduction of tumor weight 1. Apoptosis of tumor cells↑via IER3/Nrf2 (108) /HeN, human bladder cancer cells 5637) 0.2 ml BCG (600 mg/L) and volume pathway↓ AND 0.2 ml of As2O3 (30 µmol/L) for 4 weeks (once a 2. DC(%) in cancer tissue↑Expression of week) CD83/CD86↑IL-6/IL-8 in DCs↑

710 Abbreviations: TIL, tumor-infiltrating lymphocytes; As2O3, arsenic trioxide; ROS, reactive oxygen species; RNS, reactive nitrogen species; 711 IOD, integrated optical density; LAK, lymphokine activated killers; APL, acute promyelocytic leukemia; BCG, bacillus Calmette-Guerin; Immunomodulation by trivalent arsenic

712 Figure legends

713 Figure 1. Diverse functions of As(III) on immune cells

714 Trivalent arsenicals (As(III)) can drive both promoting (top of figure in green) and suppressive 715 (bottom of figure in blue) immune responses in different immune cell subsets.

+ 716 Figure 2. Mechanisms of action: As2O3 effects on CD4 Tcell

717

186 Macrophage (MΦ) NK cell TNF‐α↑IL‐1α↑CCL18↑ Cytolytic activity↑ Allogeneic/autologous T cell responses Keratinocytes Neutrophil IL‐1β↑ and IL‐18 ↑ Recruitment↑ Functions (e.g degranulation)↑ NK cell receptor NETs formation↑ regulation AIM2 activation MAPKs, Syk

Immunoenhancement 187 As(III) Immunosuppression NLRP1,NAIP5/NLRC4, Eosinophil NLRP3 inhibition MΦ Chemotaxis↓ IL‐1β↓ and Recruitment↓ tyrosine Nrf2 IL‐18↓ phosphorylation TCR IFN‐γ &RORγt Mast cell signaling↓ inhibition Dendritic cell Degranula IL‐2↓ ROS generation Maturation↓ tion↓ Rho Kinase/p38 Kinase, B cell UPR signaling (e.g ATF4) Plasmablasts Phagocytosis↓ differentiation↓ IL‐12↓IL‐23↓ T cell IgG/IgM↓ Stimulation of IFN‐ Macrophage T cell Th1/Th2 ratio↓ γ secretion from T Phagocytosis↓ Activation↓ Th17 inhibition cells↓ Adhesion↓ Proliferation↓ Treg depletion IFN‐α from pDCs↓ Surface marker↓ IFN‐γ

As2O3 As2O3 AQ P9 B cl‐ 2 xL l‐ c IFNG mRNA B B ax

Cytochrome C GSH ROS H2O+O2 APAF‐1 188 Procaspase 9 ΔΨm↓

R O Caspase 9 R γ C A t as I F p Procaspase 3 Caspase 3 a s TNFR1 e

Caspase 8 3 FADD TRADD IL‐17 TNFα (secreted by other

As2O3 treated cells) Apoptosis Résumé Lors des dernières décennies, les résultats obtenus par le traitement au trioxyde de diarsenic (As2O3) de divers modèles murins de maladies auto-immunes et inflammatoires, telles que la sclérodermie systémique (SSc), ou de maladie du greffon contre l'hôte (GVHD) ont permis de considérer ce vieux médicament comme un "nouvel" immunomodulateur. Les cellules dendritiques plasmocytoïdes (pDC) jouant un rôle majeur dans l’auto-immunité et l’alloréactivité, nous avons observé dans une première étude que des doses élevées d'As2O3 induisaient l'apoptose des pDC par voie mitochondriale. Les doses, cliniquement pertinentes, d'As2O3 inhibaient préférentiellement la sécrétion d'IFN-α et induisaient un phénotype régulateur des pDC de témoins et de patients atteints de SSc. Cette inhibition sélective del'IFN-α était probablement liée à une puissante inhibition du facteur de transcription IRF-7.

De plus, l’As2O3 altérait leur capacité à induire une prolifération lymphocytaire T CD4+, la polarisation Th1/Th22 et la différenciation des lymphocytes B en plasmablastes. Dans une seconde étude, nous avons développé un nouveau modèle murin de GVHD aigue, basé à la fois sur un conditionnement par chimiothérapie et sur un greffon mobilisé par G-CSF. Ce modèle a montré une bonne prise de greffe des cellules du donneur et des signes cliniques typiques de GVHD aigue. De plus nous avons montré qu’une décontamination bactérienne du tube digestif retardait l’apparition d’une GVHD aigue.

Mots-clés : trioxyde de diarsenic, maladie auto-immune, maladie du greffon contre l'hôte, cellules dendritiques plasmocytoïdes, modèle murin Abstract

In the recent decades, the successful treatment of arsenic trioxide (As2O3) in mouse models of several autoimmune and inflammatory diseases such as systemic sclerosis (SSc) and chronic graft-versus-host disease (GVHD) has shed light on this olddrugas a ‘novel’ immunomodulator. Notably, plasmacytoid dendritic cells (pDCs)play important role in both autoimmunity and alloreactivity. In a first study, we found that high doses of As2O3 induced pDC apoptosis induction through mitochondrial pathway.

Clinically relevant doses of As2O3 preferentially inhibited IFN-α secretion and induced regulatory phenotype of both healthy and SSc pDCs. The selective IFN-α inhibition was probably due to potent down-regulation of the transcription factor

IRF-7. Moreover, As2O3 impaired the capacity of pDCs to induce CD4+ T cell proliferation, Th1/Th22 polarization, and B cell differentiation towards plasmablasts. In a second study, we successfully constructed a new acute GVHD mouse model based on both chemotherapy conditioning and G-CSF mobilized graft, which showed good donor engraftment and typical clinical signs of aGVHD. Furthermore, we showed that gut decontamination delayed the onset of aGVHD in this model.

Keywords: arsenic trioxide, autoimmune disease, graft-versus-host disease, plasmacytoid dendritic cells, mouse model

189