The microenvironment differentially impairs passive and active immunotherapy in Chronic lymphocytic - Potential therapeutic synergism of CXCR4 antagonists Maike Buchner, Philipp Brantner, Gabriele Prinz, Meike Burger, Constance Baer, Christine Dierks, Dietmar Pfeifer, Roland Mertelsmann, John G Gribben, Hendrik Veelken, et al.

To cite this version:

Maike Buchner, Philipp Brantner, Gabriele Prinz, Meike Burger, Constance Baer, et al.. The microen- vironment differentially impairs passive and active immunotherapy in Chronic lymphocytic leukemia - Potential therapeutic synergism of CXCR4 antagonists. British Journal of Haematology, Wiley, 2010, 151 (2), pp.167. ￿10.1111/j.1365-2141.2010.08316.x￿. ￿hal-00569407￿

HAL Id: hal-00569407 https://hal.archives-ouvertes.fr/hal-00569407 Submitted on 25 Feb 2011

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. British Journal of Haematology

The microenvironment differentially impairs passive and active immunotherapy in Chronic lymphocytic leukemia - PotentialFor therapeutic Peer synergism Review of CXCR4 antagonists

Journal: British Journal of Haematology

Manuscript ID: BJH-2010-00258.R1

Manuscript Type: Ordinary Papers

Date Submitted by the 26-May-2010 Author:

Complete List of Authors: Buchner, Maike; University Hospital Freiburg, Medical Department I Brantner, Philipp; University Hospital Freiburg, Medical Department I Prinz, Gabriele; University Hospital Freiburg, Medical Department I Burger, Meike; University Hospital Freiburg, Medical Department I Baer, Constance; University Hospital Freiburg, Medical Department I Dierks, Christine; University Hospital Freiburg, Medical Department I Pfeifer, Dietmar; University Hospital Freiburg, Medical Department I Mertelsmann, Roland; University Hospital Freiburg, Medical Department I Gribben, John; Barts and The London School of Medicine, Institute of Cancer Veelken, Hendrik; University Hospital Freiburg, Medical Department I Zirlik, Katja; University Hospital Freiburg, Medical Department I

CHRONIC LYMPHOCYTIC LEUKAEMIA, STROMAL CELLS, Key Words: IMMUNOTHERAPY, DRUG RESISTANCE, CHEMOKINES

Page 1 of 34 British Journal of Haematology

1 2 3 The microenvironment differentially impairs passive and active immunotherapy in 4 5 Chronic Lymphocytic Leukemia - 6 7 8 CXCR4 antagonists as potential adjuvants for monoclonal antibodies 9 10 11 12 Running title: Microenvironmental impact on immunotherapy in CLL 13 14 15 16 Maike Buchner 1, Philipp Brantner 1, Natalie Stickel 1, Gabriele Prinz 1, Meike Burger 1, 17 18 Constance Bär 1, Christine Dierks 1, Dietmar Pfeifer 1, Ariane Ott 1, Roland Mertelsmann 1, John 19 20 2 For1 Peer1 Review 21 G. Gribben , Hendrik Veelken , Katja Zirlik 22 23 24 25 1Department of Hematology and , University Medical Center Freiburg, Freiburg, 26 27 Germany 28 29 2Barts and The London School of Medicine, Institute of Cancer, London, United Kingdom 30 31

32 33 34 Corresponding Author: Katja Zirlik, MD 35 36 Department of Hematology/Oncology 37 38 University Medical Center Freiburg 39 40 Hugstetter Strasse 55 41 42 79106 Freiburg, Germany 43 44 Phone: +49-761-270-7181 45 46 47 Fax: +49-761-270-7177 48 49 e-mail: [email protected] 50 51 52 53 Number of figures: 5 54 55 Supporting information figures: 7, Supporting information table: 1 56 57 Number of references: 61 58 59 60 Word count: 3991, Abstract word count: 191

British Journal of Haematology Page 2 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 Summary 4 5 Direct contact with stromal cells protects chronic lymphocytic leukemia (CLL) B cells from 6 7 8 -induced apoptosis in vitro . Blockade of CXCR4 signaling antagonizes stroma- 9 10 mediated interactions and restores CLL chemosensitivity. In vivo , administration of CXCR4 11 12 antagonists effectively mobilizes hematopoetic progenitor cells . Therefore, 13 14 combinations of CXCR4 blockade with cytoreductive treatment with selective activity on CLL 15 16 cells may avoid potential hematotoxicity. Hence, we tested CXCR4 antagonists in the context 17 18 of passive and active immunotherapeutic approaches. We evaluated how efficiently 19 20 For Peer Review 21 , alemtuzumab, and cytotoxic T cells killed CLL cells co-cultured with stromal cells 22 23 in the presence and absence of a CXCR4 antagonist. Stromal cell contact attenuated 24 25 rituximab- and alemtuzumab-induced complement-dependent cytotoxicity (CDC) of CLL 26 27 cells. Addition of CXCR4 antagonists abrogated the protective effect of stroma. In contrast, 28 29 stromal cells did not impair antibody-dependent cell-mediated cytotoxicity (ADCC) and 30 31 cytotoxicity induced by activated T cells. Destruction of microtubules in CLL target cells 32 33 34 restored the protective effect of stroma coculture for CLL cells during NK cell attack by 35 36 preventing mitochondrial relocalization towards the immunological synapse. Our data identify 37 38 the combination of CXCR4 antagonists with passive - but not active - immunotherapy as a 39 40 promising potential treatment concept in CLL. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2 Page 3 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 Introduction 4 5 B cell chronic lymphocytic leukemia (B-CLL) represents one of the most commonly 6 7 8 diagnosed lymphoid malignancies in Western countries, and is characterized by continuous 9 + 10 accumulation of mature, CD5 B cells. CLL remains incurable with cytoreductive therapy. The 11 12 microenvironment in bone marrow and secondary lymphoid tissues favours survival and 13 14 chemoresistance of CLL cells by upregulation of antiapoptotic molecules such as Mcl-1 and 15 16 Bcl-xL (Caligaris-Cappio 2003, Chiorazzi , et al 2005). In contrast to the prolonged survival in 17 18 vivo , CLL cells undergo rapid apoptosis in the absence of stroma in vitro . Mimicking the in 19 20 For Peer Review 21 vivo microenvironment by stromal cell coculture protects CLL cells from spontaneous and 22 23 chemotherapy-induced apoptosis in vitro (Burger , et al 2005, Kurtova , et al 2009). This cell 24 25 contact-mediated chemoresistance is described for other malignancies as well and is termed 26 27 "cell adhesion-mediated drug resistance" (CAM-DR) (Damiano , et al 1999). Residual 28 29 leukemic cells protected by stromal cells in the bone marrow or lymphatic tissue are thought 30 31 to contribute to disease relapse after chemotherapy. 32 33 34 The homing of CLL cells towards the protective niches is mediated by the crosstalk between 35 36 the chemokine CXCL12 and its receptor CXCR4. CXCR4 antagonists resensitize CLL cells 37 38 to spontaneous and chemotherapy-induced apoptosis during stromal coculture in vitro 39 40 (Burger , et al 1999, Burger , et al 2005). Combination therapy with CXCR4 antagonists is 41 42 therefore predicted to enhance efficacy of conventional cytoreductive treatment in CLL 43 44 (Burger and Kipps 2006). On the other hand, the CXCL12-CXCR4 interaction provides a 45 46 47 physiological homing signal to the bone marrow for hematopoietic stem cells (HSCs). Thus, 48 49 administration of CXCR4 antagonists mobilizes HSCs to the periphery and may render 50 51 them sensitive to concomitant chemotherapy. To avoid this risk of stem cell toxicity, 52 53 combination of CXCR4 antagonists with antineoplastic agents targeted to CLL cells would be 54 55 conceptually advantageous. 56 57 Due to their exclusive action on cells of leukocyte lineages without decreasing HSC 58 59 60 viability (Lim , et al 2008), passive and active immunotherapy represent promising treatment strategies in CLL (Faderl , et al 2005, Kater , et al 2007, Stanglmaier , et al 2004). The

3 British Journal of Haematology Page 4 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 monoclonal anti-CD52 and anti-CD20 antibodies (mAb) alemtuzumab (ALT, Campath-1H) 4 5 and rituximab (RIT) have been approved for passive immunotherapy of relapsed or refractory 6 7 8 CLL. Their precise mode of action is not fully understood. Monoclonal antibodies are 9 10 capable to activate the classic complement pathway which results in pore formation 11 12 and target cell lysis (Zipfel and Skerka 2009), known as complement dependent 13 14 cytotoxicity (CDC). Furthermore, antibodies may bind to Fc γγγ receptors on various 15 16 effector cells, such as natural killer (NK) cells, fixed macrophages, dendritic cells 17 18 19 (DCs), neutrophils, and eosinophils (Nimmerjahn and Ravetch 2006). Efficient lysis of 20 For Peer Review 21 the obsoned cells by these effectors has been termed antibody-dependent cell- 22 23 mediated cytotoxicity (ADCC). Additional direct apoptotic effects of, e.g. RIT are under 24 25 controversial discussion (Lim , et al 2010). 26 27 The graft-versus-leukemia effect observed after allogeneic hematopoietic stem cell 28 29 transplantation is evidence for anti-leukemic activity of T cells against CLL (Giannopoulos 30 31 32 and Schmitt 2006, Khouri , et al 1998). A potential CLL-associated T cell antigen is survivin, a 33

34 member of the inhibitor of apoptosis (IAP) gene family (Schmidt , et al 2003). This in vitro 35 36 study was conducted to determine whether stromal contact protects CLL cells against 37 38 passive and active immunotherapy. We evaluated the ex vivo cytotoxic effect of RIT and ALT 39 40 and the ability of survivin-specific and allogenic CTLs generated in mixed lymphocyte 41 42 reactions to lyse primary CLL cells in presence and absence of the murine stromal cell line 43 44 45 M2-10B4 with or without the CXCR4 antagonist TN14003. 46 47 48 49 50 51 Materials and methods 52 53 54 55 Patients and healthy donors 56 57 58 Peripheral blood samples were obtained from 32 patients at University Medical Center 59 60 Freiburg (SI Table 1) satisfying criteria for B-CLL and 10 healthy volunteers (Cheson , et al

1996). Rai stage and treatment status were abstracted from clinical records. IgV H mutation

4 Page 5 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 status, ZAP70 expression, and genomic aberrations were determined as described 4 5 previously (Hamblin , et al 1999, Osterroth , et al 1999, Pfeifer , et al 2007, Rassenti , et al 6 7 8 2004). 19 Patients had not been treated, 13 patients had received low-dose chlorambucil 9 10 and/or fludarabine, which had been discontinued at least 6 months prior to phlebotomy. 11 12 Blood sampling had been approved by the local ethics committee and all patients gave 13 14 informed consent. 15 16 Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll. Most CLL samples 17 18 contained >85% CLL B cells as determined by flow cytometry with anti-CD19 (BioLegend, 19 20 For Peer Review 21 San Diego, CA), anti-CD5 (BD Pharmingen Heidelberg, Germany), and anti-CD23 (Beckman 22 23 Coulter, Fullerton, CA) mAb. If purity was <85%, B cells were isolated by negative selection 24 25 (B-cell isolation kit II; Miltenyi Biotech, Bergisch Gladbach, Germany). CD8 + T cells were 26 27 isolated from healthy donors by negative selection (CD8 Isolation Kit II; Miltenyi Biotech) to a 28 29 purity of more than 85%. 30 31

32 33 34 Reagents and Cell lines 35 36 The highly selective tetradeca-peptide CXCR4 receptor antagonist TN14003 was a kind 37 38 gift from N. Fujii, Kyoto, Japan (Tamamura , et al 2001) . The stromal cell line M2-10B4 39 40 (ATCC; Manassas, VA), maintained in RPMI 1640 containing 10% FCS was plated on 96 41 42 well plates at 2 x 10 5 cells/well, starved overnight, and incubated with CXCR4 30min prior to 43 44 coculture with CLL cells where indicated. To remove CLL cells from stromal cell 45 46 47 coculture we performed gentle but persistent pipetting. Stromal cells were 48 49 discriminated from CLL cells according to their distinct forward/sideward scatter 50 51 properties in flow cytometry (Burger , et al 2005). 52 53 Arabinosyl-2-fluoroadenine (F-ara-A) was purchased from Sigma-Aldrich (St. Louis, MO) and 54 55 used at 20µM final concentration. ALT (Schering, Weimar, Germany) and RIT (Roche, 56 57 Basel, Switzerland) were used at final concentrations of 20µg/ml and 10µg/ml, respectively. 58 59 60 For complement-mediated lysis, 10% fresh human serum (HS) was added. Z-VAD was purchased from Promega Inc and used at a final concentration of 20 µM.

5 British Journal of Haematology Page 6 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 4 5 Intracellular staining of Mcl-1 6 7 8 After 48 hour culture of CLL cells with or without stromal cell contact in the presence and 9 10 absence of CXCR4 antagonists, cells were fixed and permeabilized according to 11 12 manufacturer’s protocol (Foxp3 staining kit, eBioscience, San Diego, CA), stained with a Mcl- 13 14 1 antibody (Epitomics Inc., San Francisco, CA), and analyzed by flow cytometry. 15 16 17 18 Generation of effector CTLs 19 20 For Peer Review 21 Effector CTLs generated in allogeneic mixed lymphocyte reactions (MLR) were stimulated 22 23 similarly as described previously (Kato , et al 1998). CLL B cells were plated into 96-well 24 25 plates at 1 x 10 5 cells/well (IMDM, 10% human serum). 1 x 10 5 healthy donor CD8 + T 26 27 cells/well were added and T cells were re-stimulated weekly with 2 x 10 4 CLL cells/well. For 28 29 CD40 activation, CLL cells were stimulated for 48 h with 100 ng/ml sCD154 (Biosource, 30 31 Camarillo, CA) before stimulation of T cells. Recombinant human IL-2 (R&D Systems, 32 33 34 Minneapolis, MN) was added at 25 U/ml three days after each stimulation. After 1 to 4 35 51 36 stimulations, cytotoxicity of effector CTLs was assessed by flow cytometry or chromium ( Cr) 37 38 release assays. CLL cells were incubated with or without the CXCR4 blocking agent 39 40 TN14003 for 30 min in a final concentration of 10 µg/ml prior to cultivation with the unlabelled 41 42 stromal cell line M2-10B4 for 2 hours where indicated. Effector T cells were added to the 43 44 respective culture for 4 hours. For generation of survivin-specific T cells, purified HLA-A2 + 45

46 + 47 healthy CD8 T cells were primarily stimulated with irradiated monocyte-derived, peptide- 48 49 pulsed dendritic cells that were pulsed prior to stimulation with the native survivin-derived 50 51 peptide (Sur9: ELTLGEFLKL) for 2 hours in the presence of ß2-microglobulin (3 µg/ml, 52 53 Sigma Aldrich). T cells were restimulated weekly using CD40-activated B cells pulsed with 54 55 the corresponding peptide. After 3 stimulations, cytotoxicity of CTLs against primary CLL 56 57 cells was assessed in chromium 51 Cr release assay. 58 59 60

6 Page 7 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 4 5 Immunoblotting 6 7 8 Immunoblotting was performed as previously described (Buchner , et al 2009). Primary 9 S473 10 antibodies included pAkt , Akt (Cell Signaling Technology), Mcl-1 (Epitomics, San 11 12 Francisco, CA), or βββ-actin (Sigma-Aldrich), secondary antibodies were horseradish 13 14 peroxidase-conjugated from Cell Signaling Technology. Densitometric analysis was 15 16 performed using ImageJ software (http://rsb.info.nih.gov/ij/). 17 18 19 20 For Peer Review 21 CLL cell treatment and viability assessment 22 23 B-CLL cells from frozen stocks thawed prior to experiments were incubated for 48 24 25 hours with F-ara-A or with RIT and ALT for 4 h (ADCC) and 12h (CDC), respectively. 26 27 Cells were activated with IL-2 (500 U/ml) overnight and used at 30:1 (PBMC:CLL). To 28 29 distinguish between PBMC and CLL cells, the latter were stained with CFSE. Cell viability 30 31 32 was determined by propidium iodide (PI) (Sigma Aldrich) for ADCC. Otherwise, Annexin V- 33 34 FITC (BD Bioscience) and PI was used. CLL cells were removed from stromal cell coculture 35 36 by gentle but persistent pipetting. 37 38 39 40 Chromium ( 51 Cr) release assay 41 42 Chromium release assay was performed as previously described (Zirlik , et al 2006). In brief, 43 44 51 45 survivin-specific CTLs or effector T cells from allogeneic MLR were incubated with Cr- 46 47 labeled target cells. CTL activity was titrated at different effector-to-target ratios (30:1, 10:1, 48 49 3:1) for 4 hours and 51 Cr was measured in supernatants. 50 51 52 53 Cell polarization 54 55 For confocal analysis of cell polarization in target cells, CLL cells were stained with 56 57 58 Mitotracker Red CM-H2XROS (Invitrogen, Carlsbad, CA) and treated for 30min with 10 µM 59 60 colchicine or control. After 3 washing steps, cells were preincubated with alemtuzumab

(20 µg/ml) before magnetically isolated, IL-2 activated NK cells labeled with Celltracker Green

7 British Journal of Haematology Page 8 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 were added with a ratio of 1:1 onto adhesion slides (Marienfeld GmbH, Lauda-Königshofen, 4 5 Germany), cells were fixed with 2% formaldehyde, mounted using DAPI containing mounting 6 7 8 medium (Vector Laboratories, Burlingame, CA) and analyzed using a Leica TCS SP2 AOBS 9 10 spectral confocal microscope (Leica Microsystems, Wetzlar, Germany). 11 12 13 14 Statistical analysis 15 16 Statistical analysis was performed using student`s t-test or Wilcoxon test using the 17 18 Graphpad Prism Software. 19 20 For Peer Review 21 22 23 24 25 Results 26 27 28 29 Stromal cells protect CLL cells from chemotherapy-induced apoptosis via CXCL12 - CXCR4 30 31 interaction , Akt activation, and upregulation of Mcl-1 32 33 34 To confirm the previously reported protective effect of the microenvironment on the viability 35 36 of CLL cells in vitro, we assessed apoptosis of CLL cells in the presence and absence of the 37 38 CXCL12-secreting stromal cell line M2-10B4 (Burger , et al 2005). After 48 hours under 39 40 standard culture conditions, only 50±7% of CLL cells remained viable. Coculture of CLL cells 41 42 with stromal cells significantly enhanced viability to 81±3% (n=16, p<0.01). The CXCR4 43 44 antagonist TN14003 reversed this protective effect (Supporting Information (SI) Fig. 1A) , 45 46 47 while it did not affect CLL cell survival in the absence of stroma (shown in Fig. 1A). Similar 48 49 results were obtained for the CXCR4 antagonist AMD3100 (SI Fig 1B). 50 51 52 53 Fludarabine is the most effective single chemotherapeutic agent in CLL. In vitro, treatment of 54 55 CLL cells with 20µM F-ara-A, the dephosphorylated, active fludarabine derivative, reduced 56 57 absolute viability of CLL cells from 27±3% to 9±3% after 48hrs (SI Fig. 2). To compare pro- 58 59 60 apoptotic F-ara-A effects under various culture conditions, we used the relative viability by defining the respective untreated control (without stroma, with stroma, with stroma and

8 Page 9 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 TN14003 - as indicated) as 100% (Burger , et al 1999) accounting for variabilities in 4 5 , et al . 6 spontaneous apoptosis rates in different patients' samples (Kurtova 2009) 7 8 Coculture with stromal cells significantly enhanced the relative viability after F-ara-A 9 10 treatment from 35±9% to 59±8% after 48hrs (Fig. 1A, n=8, p<0.05). The protective effect of 11 12 stromal cells on chemotherapy-induced apoptosis was mainly prevented by the CXCR4 13 14 receptor antagonist TN14003 (Fig. 1 A, p<0.05), in line with previous findings (Burger , et 15 16 al 2005). The viability of stromal cells and their capacity to secrete CXCL12 was not 17 18 19 influenced by 48 h F-ara-A treatment (SI Fig. 3A and B). 20 For Peer Review 21 22 23 Mechanistically, Mcl-1 protein expression was significantly increased in primary CLL 24 25 cells after coculture with M2-10B4 cells, an effect abrogated by the CXCR4 antagonist 26 27 TN14003 (Fig. 1B). Addition of the caspase inhibitior Z-VAD did not affect Mcl-1 28 29 expression after treatment with the CXCR4 inhibitor. Therefore, the increased 30 31 32 apoptosis induction and emerging caspase activation does not account for the Mcl-1 33 34 downregulation. 35 36 In line with previous reports, we found increased activation of Akt upon stromal cell 37 38 coculture. Again, this effect could be blocked by CXCR4 inhibition (Fig. 1C). In 39 40 contrast, the expression of the other proapoptotic mediators Bid and Bim was not 41 42 regulated and Bad was not expressed at all (data not shown). These data suggest that 43 44 45 stromal cell-induced apoptosis resistance is largely mediated by Mcl-1 upregulation 46 47 (Balakrishnan , et al 2009) . 48 49 50 51 CXCR4 antagonists increase the efficacy of CDC but not ADCC of alemtuzumab and 52 53 rituximab by attenuating the protective effects of stromal cells 54 55 To investigate whether stromal cells protect CLL cells in the setting of passive 56 57 58 immunotherapy, we examined the in vitro activity of ALT and RIT in the presence of human 59 60 complement or activated natural killer (NK) cells from healthy PBMCs on CLL cells alone and

in co-culture with stromal cells. Both antibodies significantly reduced viability of CLL cells by

9 British Journal of Haematology Page 10 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 CDC and ADCC compared to IgG control as determined by %viability in flow cytometry (SI 4 5 Fig. 4A, C and D). Similar results for CDC were obtained when total viable cell count 6 7 8 was determined (SI Fig. 4B). In accordance with previously published data, ALT mediated 9 10 CDC in vitro more efficiently than RIT (Voso , et al 2002). Coculture with stromal cells 11 12 inhibited CDC of CLL cells for both antibodies. The CXCR4 antagonist TN14003 abrogated 13 14 this protective effect (Fig. 2A and B). Again, similar results were obtained when total 15 16 viable cell count was determined (SI Fig. 5A and B). The viability and the capacity to 17 18 secrete CXCL12 of stromal cells were not influenced by the treatment of ALT and RIT 19 20 For Peer Review 21 (SI Fig. 3C and D). 22 23 24 25 We next evaluated the influence of the microenvironment on ALT- and RIT-mediated ADCC 26 27 of primary CLL cell using activated NK cells from healthy PBMCs as effector cells. PBMCs 28 29 from healthy donors were used since NK from CLL patients are severely impaired 30 31 (Gorgun , et al 2005, Katrinakis , et al 1996, Kay and Zarling 1987, Kay and Zarling 1984, 32 33 34 Ramsay , et al 2008) and autologous NK cells from patients with CLL show inferior 35 36 rituximab antibody-dependent cellular cytotoxicity than allogeneic NK cells 37 38 (Weitzman , et al 2009). In contrast to CDC, we detected no increase in CLL cell viability 39 40 upon coculture with stromal cells (Fig. 2C and D). This effect was independent of various 41 42 effector-to-target ratios (Fig. 2C and D; E:T ratio 30:1, and SI Fig. 6A and B; E:T ratio 10:1). 43 44 Taken together, this suggests that the microenvironment exerts no protective effect against 45 46 47 cellular effector mechanisms. 48 49 50 51 CTL-mediated killing of allogeneic CLL cells is not impaired by stromal cell contact 52 53 To examine CTL-mediated apoptosis, allogeneic CTL were incubated with target cells in the 54 55 presence and absence of stromal cells and CXCR4 antagonists. To exclude differences in T 56 57 cell quality as cause for variable effector function, CD8 + T cells from one donor were 58 59 60 stimulated against a total of 19 different CLL samples in parallel. T cells from a healthy donor were used as CLL-derived T-cell function is impaired (Gorgun , et al 2005,

10 Page 11 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 Ramsay , et al 2008). Overall, CLL cells had inefficient allostimulatory capacity, since we 4 5 observed measurable killing of CLL cells in only 13 of 19 cases (Fig. 3A). CTL-mediated 6 7 8 apoptosis was CLL-cell specific and dose-dependent (Fig. 3B and C). At an effector-to-target 9 10 ratio of 30:1, the measurable median target cell killing was 9% (range 4-36%, n=13). No 11 12 differences in cytotoxicity were observed according to prognostic factors (data not shown). 13 14 Coculture of CLL cells with the stromal cell line or with the stromal cell line and TN14003 did 15 16 not modulate apoptosis induced by allogeneic CTLs (Fig. 4A+B). Allogeneic CTLs were 17 18 cytotoxic only for the allogeneic CLL cells against which they were originally primed, as these 19 20 For Peer Review 21 cells did not manifest any cytotoxicity of autologous PBMCs (Fig. 4B). However, even under 22 23 conditions of prestimulation with CD40L, stromal cells did not limit CTL-induced apoptosis of 24 25 CLL cells (data not shown). 26 27 28 29 To analyse possible effects of stromal contact on T cell-mediated cytotoxicity independent of 30 31 the variable and rather weak allostimulatory capacity of CLL cells, we validated our results 32 33 34 with survivin-specific, HLA-A2-restricted healthy T cells. Survivin-specific T cells effectively 35 + + - - 36 lysed HLA-A2 survivin CLL cells but not HLA-A2 survivin CLL cells in a dose-dependent 37 38 manner (Fig. 4C). In line with a previous report (Schmidt , et al 2003), survivin was expressed 39 40 in malignant B cells from a HLA-A2 positive patient with CLL as determined by quantitative 41 42 real-time PCR (data not shown). Stromal cell contact did not significantly reduce antigen- 43 44 specific cytotoxicity, nor did the addition of the CXCR4 antagonist increase cytotoxicity, 45 46 47 indicating that T cell-mediated cytotoxicity is indeed independent of stromal cell contact (Fig. 48 49 4C). 50 51 52 53 Target cell polarisation is required for efficient cytotoxicity in stroma context 54 55 To functionally explain the observed differences in stroma protection capacity in terms of cell- 56 57 mediated or complement-mediated apoptosis, we added the toxin colchicine to the 58 59 60 cytotoxicity assay. Colchicine destroys microtubule formation and therefore abrogates target cell polarisation and mitochondrial relocalization (Goping , et al 2008). Incubation with the

11 British Journal of Haematology Page 12 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 toxin did not influence the survival of CLL or stromal cells in the observed time frame (Fig. 4 5 5A, SI Fig. 3E). NK cell activity was limited during toxin treatment (data not shown) but was 6 7 8 maintained when the target cells were preincubated and repeatedly washed prior to addition 9 10 of the activated PBMCs. In line with our previous finding (Fig. 2 C+D) we observed no 11 12 protective effect of stroma in ADCC-mediated cytotoxicity, suggesting that effector cells may 13 14 overcome stromal cell mediated protective effects. However, inhibition of microtubule 15 16 formation by colchicine treatment resulted in a significant -albeit minor- protection of 17 18 CLL cells only in stroma context (Fig. 5 A, p<0.01; SI Fig.7 for absolute viability). 19 20 For Peer Review 21 Morphologically, colchicine treatment prevented mitochondrial accumulation towards the 22 23 immunological synapse as assessed by confocal microscopy (Fig. 5B). 24 25 26 27 28 29 Discussion 30 31 Interactions between CLL cells and the microenvironment provide prolonged cell survival and 32 33 34 growth advantage and confer drug resistance to CLL cells (Caligaris-Cappio 2003, Chiorazzi , 35 36 et al 2005). In vitro , the inhibition of CXCL12-CXCR4 interactions restores susceptibility of 37 38 CLL cells in stromal contact to spontaneous and drug-induced apoptosis, suggesting 39 40 synergistic effects of combined therapy of CXCR4 antagonists and chemotherapy in vivo . 41 42 Although recent studies suggest safety and efficacy of combining cytotoxic 43 44 chemotherapy with CXCR4 antagonists in acute myeloid leukemia (AML) (Uy GL 2008), 45 46 47 this combination may provoke toxicity towards mobilized healthy hematopoietic 48 49 progenitors that are normally protected in the marrow microenvironment. Therefore, 50 51 we evaluated the effect of stromal coculture in the presence and absence of CXCR4 52 53 antagonists in more specific passive and active immunotherapeutic approaches targeting 54 55 CLL cells but not HSCs. 56 57

58 59 60 Indeed, stromal cell contact reduced complement-mediated cytotoxicity of alemtuzumab and rituximab. The addition of the CXCR4 antagonist TN14003 reversed the protective effect of

12 Page 13 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 stromal cells resulting in decreased viability of CLL cells identifying CXCR4 antagonists as 4 5 potential adjuvants for selective therapeutic intervention. Promising results in preclinical 6 7 8 tumor models indicate that CXCR4 antagonists may have anti-tumor activity in patients with 9 10 various malignancies without serious side effects (Khan , et al 2007, Stone , et al 2007, Voso , 11 12 et al 2002, Wendt , et al 2008, Yoon , et al 2007). 13 14 15 16 Molecular mechanisms of the antiapoptotic function of stromal cell contact remain largely 17 18 unknown. Although up-regulated in CLL, Bcl-2 does not seem to be critical for protection of 19 20 For Peer Review 21 CLL cells in stroma contact (Balakrishnan , et al 2009). We and others observed upregulation 22 23 of the anti-apoptotic Bcl-2 family member Mcl-1 upon stromal cell contact (Balakrishnan , et al 24 25 2009). The potential role of Mcl-1 as an oncogene in lymphoid malignancies was confirmed 26 27 in transgenic mouse models in which animals with deregulated expression of Mcl-1 28 29 eventually developed widely disseminated B-cell (Zhou , et al 1998). 30 31 Interestingly, in CLL, stromal cell contact also upregulates Akt activation (Edelmann, 32 33 34 et al 2008), an effect abrogated by CXCR4 antagonists (Fig. 1C). Since active Akt also 35 36 increases Mcl-1 half-live (Maurer , et al 2006), the Akt/Mcl-1 pathway might indeed mediate 37 38 antiapoptotic signals during stromal cell support. Accordingly, the inhibition of this 39 40 pathway may explain the efficacy of CXCR4 antagonists in this setting. In line with our 41 42 observations, Mcl-1 has been reported to mediate apoptosis resistance to fludarabine and 43 44 rituximab in vivo (Awan, et al 2009, Johnston, et al 2004, Pepper, et al 2008, Zhou, et al 45 46 47 1998) , and downregulation of Mcl-1 by siRNA enhances RIT-mediated apoptosis in vitro 48 49 (Hussain , et al 2007) . 50 51 52 53 A strong correlation of high-affinity F γR polymorphisms with clinical response to rituximab in 54 C 55 56 follicular lymphoma (FL) (Cartron , et al 2002) provides evidence for an important role of 57 58 ADCC in clinical lymphoma responses . In CLL, however, this correlation does not exist 59 60 which argues against a dominant role of ADCC in CLL. Although mainly natural killer

(NK) cells are established ADCC effector cells, other cells such as monocytes and

13 British Journal of Haematology Page 14 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 macrophages may contribute to ADCC (Dalle , et al 2008, Lefebvre , et al 2006). These 4 5 cell types were not analyzed in our study. 6 7 8 9 10 With respect to active immunotherapy, B-CLL cells per se are inefficient antigen-presenting 11 12 cells (APCs), largely due to an inadequate co-stimulatory capacity (Dazzi , et al 1995, Wolos 13 14 and Davey 1979) and the induction of an anergic phenotype in T cells by CLL cells. The 15 16 latter phenomenon is not only induced in autologous but also in allogeneic T cells, resulting 17 18 in down-modulation of genes important for proliferation and effector function (Gorgun , et al 19 20 For Peer Review 21 2005). Recently, T cells of CLL patients and also healthy T cells after direct CLL cell contact 22 23 were shown to have impaired ability to form immunological synapses and subsequent 24 25 polarisation events (Ramsay , et al 2008). In agreement with these findings we observed 26 27 efficient lysis of CLL cells in only 13 of 19 cases and overall poor cytotoxic T cell responses. 28 29 Effective cellular immune therapy, such as vaccines (Hus , et al 2008) or genetically 30 31 engineered T cell transfer (Hollyman , et al 2009) will have to overcome factors responsible 32 33 34 for the disease-related immune deficiency of patients with CLL and the capacity of leukemia 35 36 cells to effect cellular immune tolerance. Recent work has focused on strategies to overcome 37 38 these limitations and encouraging biologic effects and clinical responses have been 39 40 observed (Biagi , et al 2005, Hus , et al 2005, Kater , et al 2004a, Kater , et al 2004b, Mayr , et 41 42 al 2005, Zirlik , et al 2006). 43 44

45 46 47 Unlike in the setting of chemotherapy- and CDC-induced apoptosis, CLL-stromal cell 48 49 interactions did not impair T cell-mediated cytotoxicity suggesting that the 50 51 microenvironment does not limit the efficacy of active immunotherapy in B-CLL. This 52 53 conclusion could be drawn from the unspecific setting in MLRs as well as from the specific 54 55 setting using survivin-specific T cells. Of note, no change in stroma cell viability was 56 57 observed after addition of any of the components, including activated NK and 58 59 60 allospecific T cells (data not shown). In view of our data showing no influence on the susceptibility of B-CLL cells to ADCC and allogeneic CTL attack by stroma, it might be

14 Page 15 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 speculated that the mode of action to induce apoptosis is critical. Allo-responsive CTLs 4 5 are able to lyse resting and CD40-stimulated B-CLL cells with equal efficiency, although 6 7 8 CD40 stimulation results in upregulation of anti-apoptotic molecules, including Mcl-1 (Chu , et 9 10 al 2000, Kater , et al 2004a, Kitada , et al 1999). In line with our findings, Mcl-1 downregulation 11 12 has been shown to sensitize CLL cells against CDC but not ADCC mediated cytotoxicity of 13 14 rituximab (Hussain , et al 2007). Directed exocytosis of lytic molecules such as granzymes 15 16 toward the target cell is one of the major mechanisms used by CTLs an NK cells. By 17 18 mitochondrial relocalization towards the site of granzyme B entry during CTL attack, 19 20 For Peer Review 21 activated T cells are able to overcome upregulation of Bcl-2, usually blocking Granzyme B 22 23 (Goping , et al 2008). This suggests that cell polarization by immunological synapse formation 24 25 may overcome antiapoptotic events mediated by stroma. For CLL, disruption of the 26 27 microtubulus network restored to some extent the stroma protective effect, although 28 29 the biological significance remains to be determined. A direct effect of microtubule 30 31 disruption is unlikely as it would also affect ADCC efficacy without stroma. A potential 32 33 34 explanation for this effect may be the missing mitochondria relocation towards the 35 36 immunological synapse observed after colchicine treatment. 37 38 39 40 However, the fact that cell-mediated immunotherapy, such as active vaccinations strategies, 41 42 T cell transfer or "graft vs. leukemia"-effect after allogeneic hematopoietic stem cell 43 44 transplantation, is not limited by contact to stromal cells supports the application of this 45 46 47 therapeutic strategy. Therefore, we propose this approach to be further investigated in 48 49 vivo . 50 51 52 53 In conclusion, this study demonstrates that adhesion-mediated drug resistance protects CLL 54 55 cells against monoclonal antibody treatment, an effect abrogated by CXCR4 antagonists. 56 57 Therefore, combined therapy with CXCR4 antagonists may represent a promising strategy to 58 59 60 enhance the efficacy of ALT and RIT therapy in CLL. In addition, we show that CLL-stromal cell interactions do not impair T cell-mediated cytotoxicity on CLL cells.

15 British Journal of Haematology Page 16 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 Acknowledgements 4 5 We thank Max Warncke, Paul Fisch and Markus Simon for helpful discussions. This work 6 7 8 was supported by the Deutsche Krebshilfe grant 107275 (to M. Burger, H. Veelken, and K. 9 10 Zirlik). 11 12 13 14 15 16 17 18 19 20 For Peer Review 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

16 Page 17 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 References 4 5 Awan, F.T., Kay, N.E., Davis, M.E., Wu, W., Geyer, S.M., Leung, N., Jelinek, D.F., 6 7 Tschumper, R.C., Secreto, C.R., Lin, T.S., Grever, M.R., Shanafelt, T.D., Zent, C.S., 8 Call, T.G., Heerema, N.A., Lozanski, G., Byrd, J.C. & Lucas, D.M. (2009) Mcl-1 9 expression predicts progression-free survival in chronic lymphocytic leukemia patients 10 treated with pentostatin, cyclophosphamide, and rituximab. Blood, 113, 535-537. 11 Balakrishnan, K., Burger, J.A., Wierda, W.G. & Gandhi, V. (2009) AT-101 induces apoptosis 12 in CLL B cells and overcomes stromal cell-mediated Mcl-1 induction and drug 13 resistance. Blood, 113, 149-153. 14 Biagi, E., Rousseau, R., Yvon, E., Schwartz, M., Dotti, G., Foster, A., Havlik-Cooper, D., 15 Grilley, B., Gee, A., Baker, K., Carrum, G., Rice, L., Andreeff, M., Popat, U. & 16 Brenner, M. (2005) Responses to human CD40 ligand/human interleukin-2 17 autologous cell vaccine in patients with B-cell chronic lymphocytic leukemia. Clin 18 Cancer Res, 11, 6916-6923. 19 Buchner, M., Fuchs, S., Prinz, G., Pfeifer, D., Bartholome, K., Burger, M., Chevalier, N., 20 For Peer Review 21 Vallat, L., Timmer, J., Gribben, J.G., Jumaa, H., Veelken, H., Dierks, C. & Zirlik, K. 22 (2009) Spleen tyrosine kinase is overexpressed and represents a potential 23 therapeutic target in chronic lymphocytic leukemia. Cancer Res, 69, 5424-5432. 24 Burger, J.A., Burger, M. & Kipps, T.J. (1999) Chronic lymphocytic leukemia B cells express 25 functional CXCR4 chemokine receptors that mediate spontaneous migration beneath 26 bone marrow stromal cells. Blood, 94, 3658-3667. 27 Burger, J.A. & Kipps, T.J. (2006) CXCR4: a key receptor in the crosstalk between tumor cells 28 and their microenvironment. Blood, 107, 1761-1767. 29 Burger, M., Hartmann, T., Krome, M., Rawluk, J., Tamamura, H., Fujii, N., Kipps, T.J. & 30 Burger, J.A. (2005) Small peptide inhibitors of the CXCR4 chemokine receptor 31 (CD184) antagonize the activation, migration, and antiapoptotic responses of 32 33 CXCL12 in chronic lymphocytic leukemia B cells. Blood, 106, 1824-1830. 34 Caligaris-Cappio, F. (2003) Role of the microenvironment in chronic lymphocytic leukaemia. 35 Br J Haematol, 123, 380-388. 36 Cartron, G., Dacheux, L., Salles, G., Solal-Celigny, P., Bardos, P., Colombat, P. & Watier, H. 37 (2002) Therapeutic activity of humanized anti-CD20 monoclonal antibody and 38 polymorphism in IgG Fc receptor FcgammaRIIIa gene. Blood, 99, 754-758. 39 Cheson, B.D., Bennett, J.M., Grever, M., Kay, N., Keating, M.J., O'Brien, S. & Rai, K.R. 40 (1996) National Cancer Institute-sponsored Working Group guidelines for chronic 41 lymphocytic leukemia: revised guidelines for diagnosis and treatment. Blood, 87, 42 4990-4997. 43 Chiorazzi, N., Rai, K.R. & Ferrarini, M. (2005) Chronic lymphocytic leukemia. N Engl J Med, 44 352, 804-815. 45 46 Chu, P., Wierda, W.G. & Kipps, T.J. (2000) CD40 activation does not protect chronic 47 lymphocytic leukemia B cells from apoptosis induced by cytotoxic T lymphocytes. 48 Blood, 95, 3853-3858. 49 Dalle, S., Thieblemont, C., Thomas, L. & Dumontet, C. (2008) Monoclonal antibodies in 50 clinical oncology. Anticancer Agents Med Chem, 8, 523-532. 51 Damiano, J.S., Cress, A.E., Hazlehurst, L.A., Shtil, A.A. & Dalton, W.S. (1999) Cell adhesion 52 mediated drug resistance (CAM-DR): role of integrins and resistance to apoptosis in 53 human myeloma cell lines. Blood, 93, 1658-1667. 54 Dazzi, F., D'Andrea, E., Biasi, G., De Silvestro, G., Gaidano, G., Schena, M., Tison, T., 55 Vianello, F., Girolami, A. & Caligaris-Cappio, F. (1995) Failure of B cells of chronic 56 lymphocytic leukemia in presenting soluble and alloantigens. Clin Immunol 57 Immunopathol, 75, 26-32. 58 59 Edelmann, J., Klein-Hitpass, L., Carpinteiro, A., Fuhrer, A., Sellmann, L., Stilgenbauer, S., 60 Duhrsen, U. & Durig, J. (2008) Bone marrow fibroblasts induce expression of PI3K/NF-kappaB pathway genes and a pro-angiogenic phenotype in CLL cells. Leuk Res, 32, 1565-1572.

17 British Journal of Haematology Page 18 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 Faderl, S., Coutre, S., Byrd, J.C., Dearden, C., Denes, A., Dyer, M.J., Gregory, S.A., 4 Gribben, J.G., Hillmen, P., Keating, M., Rosen, S., Venugopal, P. & Rai, K. (2005) 5 The evolving role of alemtuzumab in management of patients with CLL. Leukemia, 6 7 19, 2147-2152. 8 Giannopoulos, K. & Schmitt, M. (2006) Targets and strategies for T-cell based vaccines in 9 patients with B-cell chronic lymphocytic leukemia. Leuk Lymphoma, 47, 2028-2036. 10 Goping, I.S., Sawchuk, T., Rieger, A., Shostak, I. & Bleackley, R.C. (2008) Cytotoxic T 11 lymphocytes overcome Bcl-2 inhibition: target cells contribute to their own demise. 12 Blood, 111, 2142-2151. 13 Gorgun, G., Holderried, T.A., Zahrieh, D., Neuberg, D. & Gribben, J.G. (2005) Chronic 14 lymphocytic leukemia cells induce changes in gene expression of CD4 and CD8 T 15 cells. J Clin Invest, 115, 1797-1805. 16 Hamblin, T.J., Davis, Z., Gardiner, A., Oscier, D.G. & Stevenson, F.K. (1999) Unmutated Ig 17 V(H) genes are associated with a more aggressive form of chronic lymphocytic 18 leukemia. Blood, 94, 1848-1854. 19 Hollyman, D., Stefanski, J., Przybylowski, M., Bartido, S., Borquez-Ojeda, O., Taylor, C., 20 For Peer Review 21 Yeh, R., Capacio, V., Olszewska, M., Hosey, J., Sadelain, M., Brentjens, R.J. & 22 Riviere, I. (2009) Manufacturing validation of biologically functional T cells targeted to 23 CD19 antigen for autologous adoptive cell therapy. J Immunother, 32, 169-180. 24 Hus, I., Rolinski, J., Tabarkiewicz, J., Wojas, K., Bojarska-Junak, A., Greiner, J., 25 Giannopoulos, K., Dmoszynska, A. & Schmitt, M. (2005) Allogeneic dendritic cells 26 pulsed with tumor lysates or apoptotic bodies as immunotherapy for patients with 27 early-stage B-cell chronic lymphocytic leukemia. Leukemia, 19, 1621-1627. 28 Hus, I., Schmitt, M., Tabarkiewicz, J., Radej, S., Wojas, K., Bojarska-Junak, A., Schmitt, A., 29 Giannopoulos, K., Dmoszynska, A. & Rolinski, J. (2008) Vaccination of B-CLL 30 patients with autologous dendritic cells can change the frequency of leukemia 31 antigen-specific CD8+ T cells as well as CD4+CD25+FoxP3+ regulatory T cells 32 33 toward an antileukemia response. Leukemia, 22, 1007-1017. 34 Hussain, S.R., Cheney, C.M., Johnson, A.J., Lin, T.S., Grever, M.R., Caligiuri, M.A., Lucas, 35 D.M. & Byrd, J.C. (2007) Mcl-1 is a relevant therapeutic target in acute and chronic 36 lymphoid malignancies: down-regulation enhances rituximab-mediated apoptosis and 37 complement-dependent cytotoxicity. Clin Cancer Res, 13, 2144-2150. 38 Johnston, J.B., Paul, J.T., Neufeld, N.J., Haney, N., Kropp, D.M., Hu, X., Cheang, M. & 39 Gibson, S.B. (2004) Role of myeloid cell factor-1 (Mcl-1) in chronic lymphocytic 40 leukemia. Leuk Lymphoma, 45, 2017-2027. 41 Kater, A.P., Evers, L.M., Remmerswaal, E.B., Jaspers, A., Oosterwijk, M.F., van Lier, R.A., 42 van Oers, M.H. & Eldering, E. (2004a) CD40 stimulation of B-cell chronic lymphocytic 43 leukaemia cells enhances the anti-apoptotic profile, but also Bid expression and cells 44 remain susceptible to autologous cytotoxic T-lymphocyte attack. Br J Haematol, 127, 45 46 404-415. 47 Kater, A.P., Remmerswaal, E.B., Nolte, M.A., Eldering, E., van Oers, M.H. & van Lier, R.A. 48 (2004b) Autologous cytomegalovirus-specific T cells as effector cells in 49 immunotherapy of B cell chronic lymphocytic leukaemia. Br J Haematol, 126, 512- 50 516. 51 Kater, A.P., van Oers, M.H. & Kipps, T.J. (2007) Cellular immune therapy for chronic 52 lymphocytic leukemia. Blood, 110, 2811-2818. 53 Kato, K., Cantwell, M.J., Sharma, S. & Kipps, T.J. (1998) Gene transfer of CD40-ligand 54 induces autologous immune recognition of chronic lymphocytic leukemia B cells. J 55 Clin Invest, 101, 1133-1141. 56 Katrinakis, G., Kyriakou, D., Papadaki, H., Kalokyri, I., Markidou, F. & Eliopoulos, G.D. 57 (1996) Defective natural killer cell activity in B-cell chronic lymphocytic leukaemia is 58 59 associated with impaired release of natural killer cytotoxic factor(s) but not of tumour 60 necrosis factor-alpha. Acta Haematol, 96, 16-23. Kay, N.E. & Zarling, J. (1987) Restoration of impaired natural killer cell activity of B-chronic lymphocytic leukemia patients by recombinant interleukin-2. Am J Hematol, 24, 161- 167.

18 Page 19 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 Kay, N.E. & Zarling, J.M. (1984) Impaired natural killer activity in patients with chronic 4 lymphocytic leukemia is associated with a deficiency of azurophilic cytoplasmic 5 granules in putative NK cells. Blood, 63, 305-309. 6 7 Khan, A., Greenman, J. & Archibald, S.J. (2007) Small molecule CXCR4 chemokine receptor 8 antagonists: developing drug candidates. Curr Med Chem, 14, 2257-2277. 9 Khouri, I.F., Keating, M., Korbling, M., Przepiorka, D., Anderlini, P., O'Brien, S., Giralt, S., 10 Ippoliti, C., von Wolff, B., Gajewski, J., Donato, M., Claxton, D., Ueno, N., Andersson, 11 B., Gee, A. & Champlin, R. (1998) Transplant-lite: induction of graft-versus- 12 malignancy using fludarabine-based nonablative chemotherapy and allogeneic blood 13 progenitor-cell transplantation as treatment for lymphoid malignancies. J Clin Oncol, 14 16, 2817-2824. 15 Kitada, S., Zapata, J.M., Andreeff, M. & Reed, J.C. (1999) Bryostatin and CD40-ligand 16 enhance apoptosis resistance and induce expression of cell survival genes in B-cell 17 chronic lymphocytic leukaemia. Br J Haematol, 106, 995-1004. 18 Kurtova, A.V., Balakrishnan, K., Chen, R., Ding, W., Schnabl, S., Quiroga, M.P., Sivina, M., 19 Wierda, W.G., Estrov, Z., Keating, M.J., Shehata, M., Jager, U., Gandhi, V., Kay, 20 For Peer Review 21 N.E., Plunkett, W. & Burger, J.A. (2009) Diverse marrow stromal cells protect CLL 22 cells from spontaneous and drug-induced apoptosis: development of a reliable and 23 reproducible system to assess stromal cell adhesion-mediated drug resistance. 24 Blood, 114, 4441-4450. 25 Lefebvre, M.L., Krause, S.W., Salcedo, M. & Nardin, A. (2006) Ex vivo-activated human 26 macrophages kill chronic lymphocytic leukemia cells in the presence of rituximab: 27 mechanism of antibody-dependent cellular cytotoxicity and impact of human serum. J 28 Immunother, 29, 388-397. 29 Lim, C.K., Sun, L., Feng, Q., Law, P., Chua, W.T., Lim, S.N. & Hwang, W.Y. (2008) Effect of 30 anti-CD52 antibody alemtuzumab on ex-vivo culture of umbilical cord blood stem 31 cells. J Hematol Oncol, 1, 19. 32 33 Lim, S.H., Beers, S.A., French, R.R., Johnson, P.W., Glennie, M.J. & Cragg, M.S. (2010) 34 Anti-CD20 monoclonal antibodies: historical and future perspectives. Haematologica, 35 95, 135-143. 36 Maurer, U., Charvet, C., Wagman, A.S., Dejardin, E. & Green, D.R. (2006) Glycogen 37 synthase kinase-3 regulates mitochondrial outer membrane permeabilization and 38 apoptosis by destabilization of MCL-1. Mol Cell, 21, 749-760. 39 Mayr, C., Kofler, D.M., Buning, H., Bund, D., Hallek, M. & Wendtner, C.M. (2005) 40 Transduction of CLL cells by CD40 ligand enhances an antigen-specific immune 41 recognition by autologous T cells. Blood, 106, 3223-3226. 42 Nimmerjahn, F. & Ravetch, J.V. (2006) Fcgamma receptors: old friends and new family 43 members. Immunity, 24, 19-28. 44 Osterroth, F., Alkan, O., Mackensen, A., Lindemann, A., Fisch, P., Skerra, A. & Veelken, H. 45 46 (1999) Rapid expression cloning of human immunoglobulin Fab fragments for the 47 analysis of antigen specificity of B cell and anti-idiotype lymphoma 48 vaccination. J Immunol Methods, 229, 141-153. 49 Pepper, C., Lin, T.T., Pratt, G., Hewamana, S., Brennan, P., Hiller, L., Hills, R., Ward, R., 50 Starczynski, J., Austen, B., Hooper, L., Stankovic, T. & Fegan, C. (2008) Mcl-1 51 expression has in vitro and in vivo significance in chronic lymphocytic leukemia and is 52 associated with other poor prognostic markers. Blood, 112, 3807-3817. 53 Pfeifer, D., Pantic, M., Skatulla, I., Rawluk, J., Kreutz, C., Martens, U.M., Fisch, P., Timmer, 54 J. & Veelken, H. (2007) Genome-wide analysis of DNA copy number changes and 55 LOH in CLL using high-density SNP arrays. Blood, 109, 1202-1210. 56 Ramsay, A.G., Johnson, A.J., Lee, A.M., Gorgun, G., Le Dieu, R., Blum, W., Byrd, J.C. & 57 Gribben, J.G. (2008) Chronic lymphocytic leukemia T cells show impaired 58 59 immunological synapse formation that can be reversed with an immunomodulating 60 drug. J Clin Invest, 118, 2427-2437. Rassenti, L.Z., Huynh, L., Toy, T.L., Chen, L., Keating, M.J., Gribben, J.G., Neuberg, D.S., Flinn, I.W., Rai, K.R., Byrd, J.C., Kay, N.E., Greaves, A., Weiss, A. & Kipps, T.J. (2004) ZAP-70 compared with immunoglobulin heavy-chain gene mutation status as

19 British Journal of Haematology Page 20 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 a predictor of disease progression in chronic lymphocytic leukemia. N Engl J Med, 4 351, 893-901. 5 Schmidt, S.M., Schag, K., Muller, M.R., Weck, M.M., Appel, S., Kanz, L., Grunebach, F. & 6 7 Brossart, P. (2003) Survivin is a shared tumor-associated antigen expressed in a 8 broad variety of malignancies and recognized by specific cytotoxic T cells. Blood, 9 102, 571-576. 10 Stanglmaier, M., Reis, S. & Hallek, M. (2004) Rituximab and alemtuzumab induce a 11 nonclassic, caspase-independent apoptotic pathway in B-lymphoid cell lines and in 12 chronic lymphocytic leukemia cells. Ann Hematol, 83, 634-645. 13 Stone, N.D., Dunaway, S.B., Flexner, C., Tierney, C., Calandra, G.B., Becker, S., Cao, Y.J., 14 Wiggins, I.P., Conley, J., MacFarland, R.T., Park, J.G., Lalama, C., Snyder, S., 15 Kallungal, B., Klingman, K.L. & Hendrix, C.W. (2007) Multiple-dose escalation study 16 of the safety, pharmacokinetics, and biologic activity of oral AMD070, a selective 17 CXCR4 receptor inhibitor, in human subjects. Antimicrob Agents Chemother, 51, 18 2351-2358. 19 Tamamura, H., Omagari, A., Hiramatsu, K., Gotoh, K., Kanamoto, T., Xu, Y., Kodama, E., 20 For Peer Review 21 Matsuoka, M., Hattori, T., Yamamoto, N., Nakashima, H., Otaka, A. & Fujii, N. (2001) 22 Development of specific CXCR4 inhibitors possessing high selectivity indexes as well 23 as complete stability in serum based on an anti-HIV peptide T140. Bioorg Med Chem 24 Lett, 11, 1897-1902. 25 Uy GL, R.M., McFarland KM, et al. (2008) Mobilization and chemosensitization of AML with 26 the CXCR4 antagonist Plerixafor (AMD3100): a phase I/II study of AMD3100+MEC in 27 patients with relapsed or refractory disease. . Blood, 112, 678a. 28 Voso, M.T., Pantel, G., Rutella, S., Weis, M., D'Alo, F., Urbano, R., Leone, G., Haas, R. & 29 Hohaus, S. (2002) Rituximab reduces the number of peripheral blood B-cells in vitro 30 mainly by effector cell-mediated mechanisms. Haematologica, 87, 918-925. 31 Weitzman, J., Betancur, M., Boissel, L., Rabinowitz, A.P., Klein, A. & Klingemann, H. (2009) 32 33 Variable contribution of monoclonal antibodies to ADCC in patients with chronic 34 lymphocytic leukemia. Leuk Lymphoma, 50, 1361-1368. 35 Wendt, M.K., Cooper, A.N. & Dwinell, M.B. (2008) Epigenetic silencing of CXCL12 increases 36 the metastatic potential of mammary carcinoma cells. Oncogene, 27, 1461-1471. 37 Wolos, J.A. & Davey, F.R. (1979) Depressed stimulation in the MLR by B lymphocytes on 38 chronic lymphocytic leukemia: failure to demonstrate a suppressor cell. Clin Immunol 39 Immunopathol, 14, 77-85. 40 Yoon, Y., Liang, Z., Zhang, X., Choe, M., Zhu, A., Cho, H.T., Shin, D.M., Goodman, M.M., 41 Chen, Z.G. & Shim, H. (2007) CXC chemokine receptor-4 antagonist blocks both 42 growth of primary tumor and metastasis of head and neck cancer in xenograft mouse 43 models. Cancer Res, 67, 7518-7524. 44 Zhou, P., Qian, L., Bieszczad, C.K., Noelle, R., Binder, M., Levy, N.B. & Craig, R.W. (1998) 45 46 Mcl-1 in transgenic mice promotes survival in a spectrum of hematopoietic cell types 47 and immortalization in the myeloid lineage. Blood, 92, 3226-3239. 48 Zipfel, P.F. & Skerka, C. (2009) Complement regulators and inhibitory proteins. Nat Rev 49 Immunol, 9, 729-740. 50 Zirlik, K.M., Zahrieh, D., Neuberg, D. & Gribben, J.G. (2006) Cytotoxic T cells generated 51 against heteroclitic peptides kill primary tumor cells independent of the binding affinity 52 of the native tumor antigen peptide. Blood, 108, 3865-3870. 53 54 55 56 57 58 59 60

20 Page 21 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 Figure legends 4 5 6 7 Figure 1: Stromal cells protect CLL cells from chemotherapy-induced apoptosis via 8 9 CXCR4 - CXCL12 interactions and pAkt-dependent Mcl-1 upregulation. (A) Viability of 10 11 CLL cells cultured with or without stromal cells and treated with 20µM F-ara-A or the 12 13 combination of F-ara-A and 10µg/ml TN14003 was determined by annexin V/PI staining. 14 15 Relative CLL cell viability was determined by defining the respective control as 100% (as 16 17 indicated with the control bar) after 48hrs treatment. Data represent means±SEM (n=8). (B) 18 19 Means (± SEM ) of mean fluorescent intensity (MFI by flow cytometry) of Mcl-1 expression in 20 For Peer Review 21 22 CLL cells cultured with or without stroma in the presence or absence of the CXCR4 23 24 antagonist TN14003 are displayed on the left (n=12). A representative experiment shows 25 26 MFI of Mcl-1 expression in primary CLL cells cultured with or without stromal cells on the 27 28 right. Below, a representative Mcl-1 immunoblot of CLL cells is shown in the presence and 29 30 absence of stroma, with and without addition of TN14003 and/or the caspase inhibitor Z-VAD 31 32 33 (20 µM). (C) Immunoblot shows pAkt and Akt of a representative CLL sample after 15min 34 35 with or without coculture of stroma in the presence and absence of the CXCR4 antagonist 36 37 TN14003. 38 39 40 41 Figure 2: CXCR4 antagonists increase the efficacy of ALT and RIT complement 42 43 induced cytotoxicity by attenuating the protective effects of stromal cells. (A+B) CLL 44 45 46 cells (n=7) cultured with or without stromal cells in the presence or absence of the CXCR4 47 48 antagonist TN14003 treated with 20µg/ml ALT (A) or 10µg/ml RIT (B) for 12hrs. Relative 49 50 viabilities were determined by staining with Annexin V and PI and presented as means±SEM. 51 52 (C+D) Relative viability of CLL cells (n=12) cultured with or without stromal cells in the 53 54 presence or absence of the CXCR4 antagonist TN14003 treated with 20µg/ml ALT (C) or 55 56 10µg/ml RIT (D) for 4hrs in the presence of IL-2-activated PBMC at a ratio 30:1. IgG control 57 58 59 of the respective culture setting (w/o stroma, w/stroma, or stroma+TN14003) was defined as 60 100% relative viability. Using annexin V and PI staining, CLL relative cell viability was

determined and presented as means±SEM.

21 British Journal of Haematology Page 22 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 4 5 Figure 3: CTL-mediated killing of allogeneic CLL cells. (A) Killing of CLL cells of 19 CLL 6 7 8 patients induced by effector T cells generated in allogeneic MLR with spontaneous levels of 9 10 apoptosis of B-CLL cells subtracted as determined by flow cytometry (CFSE/PI staining) at 11 12 an effector-to-target ratio of 30:1 (M=mutated IgV H; UM=unmutated IgV H). (B) Flow cytometry 13 14 analysis of cytolytic activity of effector T-cells from allogeneic MLR against CLL cells by 15 16 staining with CFSE/PI presented as effector-to-target ratios 3:1 and 30:1. (C) Dose- 17 18 dependent killing of CLL cells induced by effector T cells with spontaneous levels of 19 20 For Peer Review 21 apoptosis of B-CLL cells subtracted as determined by flow cytometry (CFSE/PI staining) 22 23 represent means±SEM of triplicates. 24 25 26 27 Figure 4: CLL - stromal cell interactions do not impair T cell mediated cytotoxicity. (A) 28 29 Viability of CLL cells (n=11) in the presence or absence of stromal coculture after incubation 30 31 with T cells generated in allogeneic mixed lymphocyte reaction as assessed by flow 32 33 + - 34 cytometry (CFSE /PI ). Untreated control without addition of T cells of the respective culture 35 36 setting was defined as 100% viability. Data represent means±SEM. (B) Cytolytic activity of 37 38 effector CTLs generated in allogeneic MLR as assessed by flow cytometry against CLL cells 39 40 in presence of the stromal cell line M210B4 with or without the CXCR4 inhibitor TN14003 41 42 and against autologous PMBCs at three different effector-to-target ratios. Displayed are 43 44 means±SEM of PI +-values of samples from 3 representative patients. (C) Killing by 45 46 47 allogeneic CTLs generated against a native (Sur9: ELTLGEFLKL) survivin-derived peptide of 48 49 HLA-A2-positive CLL cells cultured with or without stromal cells in the presence or absence 50 51 of the CXCR4 antagonist TN14003 and HLA-A2-negative CLL cells at 3 different effector- 52 53 target ratios. 54 55 56 57 Figure 5: Prevention of mitochondrial relocalization in target cells restores protective 58

59 + 60 stromal effect. (A) CFSE -CLL cells were pretreated with 10 µM colchicine after incubation with and without stroma before addition of ALT or IgG control and IL-2-activated PBMCs.

22 Page 23 of 34 British Journal of Haematology

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 After 2hrs, viabililty was measured by %CFSE +/PI - and the relative viability calculated by 4 5 defining the respective control culture as 100 % relative viability. Data are presented as 6 7 8 means±SEM (n=7). (B) Confocal images of colchicine- or control-pretreated CLL cells 9 10 stained for mitochondria (red) in direct contact to IL-2-activated NK cells (green) during 11 12 alemtuzumab treatment. 13 14 15 16 17 18 Supporting information Figure Legends: 19 20 For Peer Review 21 SI Fig. 1: (A) Viability of CLL cells cultured with or without stromal cells in the presence or 22 23 absence of the CXCR4 antagonist TN14003 determined by annexin V/PI staining. Displayed 24 25 are the mean percentage viability values (± SEM) of leukemic cells assessed by staining with 26 27 Annexin V and PI (n=16). (B) Direct comparison of the CXCR4 antagonist TN14003 and 28 29 AMD1300 (10 µM, Sigma Aldrich) revealed equal effects. Both antagonists significantly 30 31 reduced stromal cell induced survival (p<0.05, n=5, respectively). 32 33 34 35 - - 36 SI Fig. 2: Absolute viability of CLL cells defined as Annexin /PI , cultivated with or without 37 38 stromal cells and treated with 20 µM F-ara-A or the combination of F-ara-A and/or 10 µg/ml 39 40 TN14003. Data are presented as mean ± SEM values (n=8). 41 42 43 44 SI Fig. 3: Standard MTT assay and CXCL12 ELISA (BD Bioscience) were performed after 45 46 47 incubation of stromal cells with 20 µM F-ara-A (A+B), IgG, rituximab, or alemtuzumab (C+D) 48 49 for the same time frames as during CLL treatment. (E) Incubation with 10 µM colchicine for 50 51 30 min did not alter stromal cell viability as determined by MTT assay. 52 53 54 55 SI Fig. 4: (A) Relative viability of CLL cells cultured with ALT (20 µg/ml) or RIT (10 µg/ml) for 56 57 12 hours in the presence of 10% fresh human serum (n=7) for complement mediated 58 59 60 cytotoxicity determined by Annexin V/PI staining. (B) Viable cell count of CLL cells treated with IgG (20 µg/ml), ALT (20 µg/ml) or RIT (10 µg/ml) for 12 hours in the presence of 10%

23 British Journal of Haematology Page 24 of 34

Buchner et al. 1 Microenvironmental impact on immunotherapy in CLL 2 3 fresh human serum (n=7) for complement mediated cytotoxicity determined by Guava 4 5 viacount. (C+D) Relative viability of CLL cells cultured with ALT (20 µg/ml) or RIT (10 µg/ml) 6 7 8 for 4 h with IL-2 activated PBMC (n=8) for ADCC at a effector-to-target ratio of 30:1 (C) and 9 10 10:1 (D), IgG control was defined as 100% viability. Data are presented as mean ± SEM 11 12 values. 13 14 15 16 SI Fig. 5: Viable cell count of CLL cells treated with rituximab (A) or alemtuzumab (B) in the 17 18 presence of fresh human stroma, measured using the Guava viacount technology for the 19 20 For Peer Review 21 indicated culture conditions (n=7). Data are presented as mean ± SEM values. 22 23 24 25 SI Fig. 6: (A+B) Relative viability of CLL cells (n=8) cultured with or without stromal cells in 26 27 the presence or absence of the CXCR4 antagonist TN14003 treated with 20 µg/ml ALT (A) or 28 29 10 µg/ml RIT (B) for 4 hours in the presence of IL-2 activated PBMC at a ratio 10:1. IgG 30 31 control of the respective culture setting (w/o stroma, w/ stroma, or stroma+TN14003) was 32 33 34 defined as 100% relative viability. Data are presented as mean ± SEM values. 35 36 37 38 SI Fig. 7: CFSE + CLL cells were pretreated with 10 µM colchicine after incubation with and 39 40 without stroma before IL-2 activated PBMCs were added. After 2 hours, viabililty was 41 42 + - measured by %CFSE /PI . Data are presented as mean ± SEM values. 43 44

45 46 47 Supporting information Table 1: Clinical characteristics of CLL patients. 48 49 50 51 52 53 54 55 56 57 58 59 60

24 Page 25 of 34 British Journal of Haematology Figure 1 w/o stroma w/ stroma 1 A 2 3 4 5 100 6 7 8 9 10 50 11

12 viability %

13 (control=100%) 14 15 16 0 17 CLL CLL CLL 18 CLL For Peer Review 19 20 21 TN14003 22 TN14003 CLL + + CLL F-ara-A 23 + CLL F-ara-A CLL + + CLL TN14003 CLL + + CLL TN14003 CLL + + CLL TN14003 CLL + + CLL TN14003 CLL + + CLL F-ara-A + 24 + CLL F-ara-A + 25 p<0.01 p<0.05 26 B 27 100 28 250 29 80 30 200 31 60 32 150 33 40 34 % Max of % 20 35 Mcl-1 MFI 100 36 ∆ ∆ ∆ ∆ 0 37 50 38 Mcl-1 FITC 39 40 0 isotype control 41 42 Stroma - + + w/o stroma 43 TN14003 - - + 44 w/ stroma 45 46 noStroma stroma stroma+TN14003 47 48 49 Mcl-1 50 β actin 51 52 Caspase inhibition - + - + - + 53 54 densitometric quotient 0.4 0.7 1.0 1.0 0.8 0.3 55 56 57 C pAkt 58 59 Akt 60 Stroma - - + + TN14003 - + - + British Journal of Haematology Page 26 of 34 Figure 2

1 2 3 4 A CDC Alemtuzumab CDC Rituximab 5 B 6 7 p<0.01 p<0.05 p<0.05 p<0.05 8 100 9 10 80 11 100 12 60 13 14 15 40 16 50

17 control) (% of 20 control) (% of 18 viability relative For Peer Review viability relative 19 20 0 0 21 Stroma - + + Stroma - + + 22 TN14003 - - + TN14003 - - + 23 24 25 26 27 28 C D ADCC Rituximab 29 ADCC Alemtuzumab 30 100 100 31 32 80 80 33 34 35 60 60 36 37 40 40 38 39 control) (% of relative viability relative 40 20 control) (% of 20 41 viability relative 42 0 0 43 44 Stroma - + + Stroma - + + 45 TN14003 - - + TN14003 - - + 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 27 of 34 British Journal of Haematology Figure 3

1 40 2 A 3 4 30 5 6 7 20

8 cells (%)

9 +

10 PI 10 11 12 13 0 14 15 16 17 18 For Peer Review 19 # 143 (UM) # 143 # 161 (M) # 161 # 85 (M) # 85 (UM) # 88 # 28 (UM) # 28 (UM) # 90 (M) # 14 (M) # 56 (M) # 176 # 95 (M) # 95 (M) # 101 (M) # 166 (M) # 202 (UM) # 209 (M) # 108 (M) # 62 # 210 (M) # 210 (UM) # 59 20 (M/UM) # 43 21 22 23 B 24 25 8 3:1 26 32 PI 27 95 28 Cells # 29 Lin: SS SS 30 31 32 FS Lin: FS CFSE CFSE 33 34 35 36 26 30:1 37 PI 38 39 40 41 42 CFSE 43 44 45 46 30 47 C

48 ) # 210 49 % # 43 50 ( 20 51 s # 56

l 52 l 53 e

c

54 + 55 I 10 56 P 57 58 59 0 60 3:1 10:1 30:1 Effector-to-Target Ratio British Journal of Haematology Page 28 of 34 Figure 4

1 2 3 4 5 A B 6 100 30 Stroma + CLL 7 Stroma + CLL + TN14003 8 80 )

% autologous PBMC

9 ( 10 20

s

60 l

11 l

12 e 13 40 c

+

14 I

(% of control) (% of 10

15 P relative viability relative 16 20 17 18 0 For Peer Review0 19 20 Stroma - + + 3:1 10:1 30:1 21 22 TN14003 - - + Effector-to-Target Ratio 23 24 25 26 27 28 C 29 ) 40

%

30 ( 31

s 32 i 30 33 s TC survivin - CLL HLA A2 pos. y

34 l TC survivin - CLL + stroma

35 c

i 20 36 f TC survivin - CLL + stroma + TN14003

i

37 c TC survivin - CLL HLA A2 neg. 38 e

39 p 10 40 S 41 42 0 43 3:1 10:1 30:1 44 45 Effector-to-Target Ratio 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 29 of 34 British Journal of Haematology Figure 5

1 2 A 3 w/o stroma w/ stroma 4 5 n.s. 6 100 7 n.s. ** 8 80 9 10 11 60 12 13 40 14 % relative viability % 15 20 16 17 18 For0 Peer Review 19 20 ALT - - + + - - + + 21 Colchicine - + - + - + - + 22 23 24 25 B control Colchicine 26 27 28 29 30 31 32 CLL #225 33 34 35 36 37 38 39 40 41 42 CLL #109 43 44 45 46 47 48 49 50 51 CLL #54 52 53 54 55 56 57 58 59 60 British Journal of Haematology Page 30 of 34

1 SI Figure 1 2 p<0.01 p<0.05 3 100 4 A 5 6 75 7 8 9 50 10 11 12 13 25 14 15 cellCLL viability (%) 16 0 17 Stroma - + + 18 For Peer Review 19 TN14003 - - + 20 21 p<0.05 22 B 23 p<0.05 24 100 25 26 27 80 28 29 60 30 31 32 40 33 34 20 35 relative viability % 36 37 0 38 ma o ma ma 39 o o Str 40 Str Str 41 +TN14003 +AMD3100 42 43 44 SI Figure 2 45 w/o stroma w/ stroma 46 47 80 48 49 50 60 51 52 53 40 54 55 56 20 57 absolute% viability 58 59 0 60 F-Ara-A - - + + - - + + TN14003 - + - + - + - + Page 31 of 34 British Journal of Haematology SI Figure 3

1 2 n.s. 3 n.s. 4 5 A 1.5 B 4 6 7 8 3 9 1.0 10 11 2 12 13 0.5 14 1 OD (MTT assay)

15 [ng/ml] CXCL12 16 17 0.0 0 18 For Peer Reviewuntreated F-ara-A 19 Untreated F-Ara-A 20 21 22 n.s. n.s. 23 24 C D 25 1.5 4 26 27 28 3 29 1.0 30 31 2 32 33 0.5 34 1 OD (MTT assay) 35 [ng/ml] CXCL12 36 37 0.0 0 38 39 40 IgG RIT ALT IgG ALT RIT 41 42 43 44 E 45 n.s. 46 1.0 47 48 0.8 49 50 51 0.6 52 53 0.4 54 55 56 OD (MTT assay) 0.2 57 58 0.0 59 60 Untreated Colchicine British Journal of Haematology Page 32 of 34

SI Figure 4 1 A CDC B 2 CDC 3 p<0.01 4 p<0.0001 p<0.01 5 100 p<0.0001 6 /ml) 7 5 10 80 8 9 10 60 11 12 40 5 13 14 (% of control) (% of 15 20 relative viability relative 16 17 0 cell viable count (x10 0 18 For Peer Review 19 IgG RIT ALT RIT 20 IgG ALT 21 ADCC 22 C D ADCC 23 p<0.0001 24 p<0.0001 p<0.05 p<0.0001 25 100 26 100 10:1 27 30:1 80 28 80 29 30 60 31 60 32 40 33 40 34 (% of control) (% of

35 20 control) (% of

relative viability relative 20

36 viability relative 37 0 38 0 39 IgG RIT ALT 40 IgG RIT ALT 41 SI Figure 5 42 43 A B 44 p<0.05 p<0.05 p<0.01 p<0.05 45 10

46 /ml) /ml) 5 47 8 5 10 48 49 50 6 51 52 4 5 53 54 55 2 56 viable cell count (x10 57 0 viable cell count (x10 0 58 59 4 60 R noStr Rit Str ALT Str Rit ALT noStr Str CXCR4 Rit Str CXC ALT Page 33 of 34 British Journal of Haematology

1 2 3 4 5 6 7 SI Figure 6 8 9 10 A ADCC Alemtuzumab B ADCC Rituximab 11 100 100 12 13 14 80 80 15 16 60 60 17 18 For Peer Review 19 40 40 20 (% of control) (% of

21 viability relative 20 control) (% of 20 22 relative viability relative 23 24 0 0 25 Stroma - + + Stroma - + + 26 27 TN14003 - - + TN14003 - - + 28 29 30 31 32 33 34 SI Figure 7 35 36 37 38 w/o stroma w/ stroma 39 40 100 41 n.s. 42 ** 43 90 44 45 46 80 47 48

49 % viability 50 70 51 52 53 60 54 ALT - - + + - - + + 55 Colchicine - + - + - + - + 56 57 58 59 60 SI Table 1 British Journal of Haematology Page 34 of 34

CLL Age Sex Rai Binet IgV H % ZAP70 Gen. Ab. Prior therapy 1 mut. (at least 6 month # ago) 2 3 95 69 M 2 B M 2.1 positive Del13q14 Chlorambucil 4 5 210 68 F 1 A M 6.1 negative n.d. None 6 101 67 M 4 C M 3.1 negative Del13q14 Chlorambucil 7 8 166 85 F 0 A M 6.8 negative Del13q14 None 9 202 10 69 M 4 C n.d. negative Del13q14 Fludarabine 11 209 66 M 0 A UM 0 positive n.d. None 12 13 57 M 2 B M / 7.2 / negative Trisomy 12 None 14 43 UM 0.7 15 59 56 M 4 C UM 0 positive Del17p13 Fludarabine 16 17 108 68 M 0 A n.d. negative n.d. None 18 For Peer Review 19 28 66 F 1 A UM 0 positive None Chlorambucil 20 90 74 M 3 B UM 0 n.d. Del13q14 Chlorambucil 21 22 14 71 M 0 A M 4.1 negative Del13q14 None 23 homozygot 24 25 56 67 F 1 A M 7.5 negative Del13q14 None 26 176 71 F 0 A n.d. negative Del13q14 None 27 28 85 73 M 2 A M 10.8 negative Del13q14 None 29 30 143 69 F 1 B UM 0 n.d. Trisomy 12 Chlorambucil 31 88 66 M 1 B UM 0 n.d. Trisomy 12 Chlorambucil 32 33 62 68 F 2 A M 3.8 negative Trisomy 3 None 34 35 161 71 M 0 A M 6.8 n.d. Del13q14 none 36 biallelic 37 140 55 M 2 A M 4,8 n.d. Del13q14 none 38 39 146 69 M 0 A M 8,9 negativ Del13q14 none 40 41 77 58 M 1 A n.d. n.d. none 42 43 23 67 M 2 B UM 0 n.d. Del13q14, none 44 Del17p13 45 69, 78 M 3 C UM 1,3 positiv none none 46 183 47 48 91 45 M 0 A M 3,4 n.d. Del13q14 none 49 50 107 74 F 2 B M 3,1 negativ keine Chlorambucil 51 100 73 F 2 A M 8,2 negativ Del13q14 Chlorambucil, 52 Fludarabin 53 54 45 59 M 2 A M 6,9 negativ Del13q14 none 55 56 191 69 F 0 A M 9,9 n.d. n.d. none 57 58 109 70 M 2 A UM 0 n.d. Del13q14 Fludarabin, 59 Bendamustin 60 225 60 F 4 C M n.d. Del 13q14 Chlorambucil

54 48 M 2 B UM 0 pos Del18p19 Fludarabin