Leukemia (2011) 25, 1268–1277 & 2011 Macmillan Publishers Limited All rights reserved 0887-6924/11 www.nature.com/leu ORIGINAL ARTICLE A novel role of the CX3CR1/CX3CL1 system in the cross-talk between chronic lymphocytic leukemia cells and tumor microenvironment E Ferretti1, M Bertolotto2, S Deaglio3, C Tripodo4, D Ribatti5, V Audrito3, F Blengio6, S Matis7, S Zupo8, D Rossi9, L Ottonello2, G Gaidano9, F Malavasi10, V Pistoia1,11 and A Corcione1,11 1Laboratory of Oncology, IRCCS G. Gaslini, Genova, Italy; 2Laboratory of Phagocyte Physiopathology and Inflammation, Department of Internal Medicine, University of Genova, Genova, Italy; 3Department of Genetics, Biology & Biochemistry, Human Genetics Foundation (HuGeF), University of Torino, Torino, Italy; 4Tumor Immunology Unit, Department of Health Science, Human Pathology Section, University of Palermo, Palermo, Italy; 5Department of Human Anatomy and Histology, University of Bari, Bari, Italy; 6Laboratory of Molecular Biology, Gaslini Institute, Genova, Italy; 7Medical Oncology C, National Cancer Research Institute, Genova, Italy; 8Laboratory of Diagnostics of Lymphoproliferative Disorders, National Cancer Research Institute, Genova, Italy; 9Division of Hematology, Department of Clinical and Experimental Medicine, Amedeo Avogadro, University of Eastern Piedmont, Novara, Italy and 10Department of Genetics, Biology & Biochemistry, Research Center for Experimental Medicine (CeRMS), University of Torino, Torino, Italy Several chemokines/chemokine receptors such as CCR7, Additional prognostic markers are surface CD38 and intra- CXCR4 and CXCR5 attract chronic lymphocytic leukemia cellular ZAP-70 whose expression in leukemic cells correlates (CLL) cells to specific microenvironments. Here we have with unfavorable clinical outcome.2,4,5 Finally, different chro- investigated whether the CX3CR1/CX3CL1 axis is involved in the interaction of CLL with their microenvironment. CLL cells mosomal abnormalities detected in CLL such as 13q deletion, from 52 patients expressed surface CX3CR1 and CX3CL1 and 17deletion, 11q deletion and 12q trisomy are prognostically 6 released constitutively soluble CX3CL1. One third of these were relevant. attracted in vitro by soluble CX3CL1. CX3CL1-induced phos- Chemokines are low molecular weight cytokines specialized phorylation of PI3K, Erk1/2, p38, Akt and Src was involved in in leukocyte recruitment to inflammatory sites and in the correct induction of CLL chemotaxis. Leukemic B cells upregulated positioning of lymphocytes in secondary lymphoid organs.7 CXCR4 upon incubation with CX3CL1 and this was paralleled by increased chemotaxis to CXCL12. Akt phosphorylation was Chemokines are also involved in different pathological pro- involved in CX3CL1-induced upregulation of CXCR4 on CLL. In cesses including growth and dissemination of solid tumors and proliferation centers from CLL lymph node and bone marrow, hematological malignancies.8,9 CX3CL1 was expressed by CLL cells whereas CX3CR1 was Several chemokines/chemokine receptors control the migra- detected in CLL and stromal cells. Nurselike cells (NLCs) tion of CLL cells to lymph nodes, secondary lymphoid organs generated from CLL patient blood co-expressed surface and bone marrow. Extravasation and migration across endo- CX CR1 and CX CL1, but did not secrete soluble CX CL1. Only 3 3 3 thelium are mediated by CCR7, expressed on CLL cells, and its half of NLC cell fractions were attracted in vitro by CX3CL1. In conclusion, the CX3CR1/CX3CL1 system may contribute to ligands CCL21 and CCL19 expressed on high endothelial interactions between CLL cells and tumor microenvironment by venules of secondary lymphoid organs.10 CLL cells overexpress increasing CXCL12-mediated attraction of leukemic cells to CXCR5 that interacts with its ligand CXCL13 produced by NLC and promoting directly adhesion of CLL cells to NLC. CD68 þ macrophages in lymphoid tissues, thus contributing to Leukemia (2011) 25, 1268–1277; doi:10.1038/leu.2011.88; leukemic cell migration to the latter sites.11 Finally, migration of published online 6 May 2011 Keywords: chemokines; chronic lymphocytic leukemia (CLL); CLL cells to bone marrow is mediated by CXCR4, highly nurselike cells (NLCs); tumor microenvironment expressed in leukemic cells, and its ligand CXCL12, produced by stromal cells.12 CX3CL1/fractalkine, a chemokine constitutively expressed by many hematopoietic and non-hematopoietic tissues,13,14 is synthesized as membrane-bound protein but can also be Introduction released by proteolytic cleavage. Membrane-bound CX3CL1 functions as an adhesion molecule whereas the secreted form Chronic lymphocytic leukemia (CLL) is characterized by the þ triggers chemotaxis of lymphocytes and monocytes to inflam- monoclonal expansion of long-lived CD5 B lymphocytes in 13,15 matory sites. the peripheral blood, bone marrow and secondary lymphoid CX C chemokine receptor 1 (CX CR1), the receptor for organs.1 The presence or absence of somatic mutations of 3 3 CX3CL1, is expressed on human NK cells, monocytes, immunoglobulin variable heavy chain (IGHV) genes in malig- 13,16 17 T lymphocytes, mast cells as well as on B cells. nant cells is the best predictor of clinical outcome of CLL Previous studies have shown that CX CR1 is upregulated in patients. In particular, unmutated IGVH genes are associated 3 different types of B cell lymphoma and in CLL. However, the with more aggressive course and shorter survival as compared 2,3 functional role of CX3CR1 in these malignancies has not been with mutated cases. 18,19 20,21 addressed. Other studies performed in breast, prostate, pancreatic adenocarcinoma22 and glioma23 models have shown Correspondence: Dr A Corcione, Laboratory of Oncology, IRCCS that CX3CR1 is involved in metastatic spreading of tumor cells to G. Gaslini, Largo G. Gaslini 5, Genova 16148, Italy. specific tissues expressing CX CL1. E-mail: [email protected] 3 11These authors contributed equally to this work Here we have investigated the functional role of the CX3CR1/ Received 21 October 2010; revised 7 February 2011; accepted 16 CX3CL1 axis in CLL and found that these molecules are involved March 2011; published online 6 May 2011 in the cross-talk between tumor cells and their microenvironment. CX3CR1/CX3CL1 system and CLL microenvironment E Ferretti et al 1269 Patients and methods with unconjugated mAb, followed by secondary reagent and analyzed. Cells were run on a FACSCalibur instrument (Becton- Patients Dickinson, BD, Franklin Lakes, NJ, USA). For each analysis Blood samples were obtained from 52 CLL patients after 10 000 events were acquired and analyzed using CellQuest informed consent in accordance with Helsinki Declaration software (BD). Results were expressed as either percentage of and study approval by the Institutional Review Boards of the positive cells or mean fluorescence intensity ratio (MFIR), University of Eastern Piedmont, Novara and the National Cancer calculated by dividing the MFI of cells stained with specific Research Institute, Genova. CLL diagnosis rested upon, blood mAb by the MFI of cells stained with isotype-control mAb. In count of at least 5 Â 109 B lymphocytes per litre, immunoglo- some experiments, cells were analyzed using a FACSort (BD bulin (Ig) light chain restriction, co-expression of CD19, CD5 Biosciences), acquiring a minimum of 1500 events. Data were and CD23, low expression of surface Ig and absent or low analyzed using the WindMDI software (freely downloadable at expression of CD79b on CLL.6 http://facs.scripps.edu/software.html) Clinical and laboratory features of CLL patients were as follows: (i) patient age: median 65, range 43–85 years; (ii) sex: 30/52 males and 22/52 females; (iii) no treatment at study entry Chemotaxis assay for at least 2 months; (iv) median B lymphocyte count PBMCs (5 Â 105) were dispensed in the upper chamber of 13.5 Â 109/l with at least 85% CD19 þ CD5 þ cells; (v) IGHV 24-transwell plates of polycarbonate membrane with 5 mm pore- mutational status (unmutated X98% homology to the germline size (Costar, Cambridge, MA, USA), and increasing concentra- 2 gene) : 28/52 mutated and 24/52 unmutated; (vi) ZAP70 tions (10, 100, 300, 600, 1000 ng/ml) of rCX3CL1 (R&D System) expression (cutoff between positive and negative samples ¼ 20% or medium alone (nil) were added to the lower chamber.17 The CD19 þ CD5 þ cells, as assessed by flow cytometry): 31/52 rCXCL12 (R&D System) was tested at 300 ng/ml. Plates were ZAP70 þ and 21/52 ZAP70À; (vii) CD38 expression (cutoff incubated for 2 h at 37 1C. Transmigrated cells were harvested, between positive and negative samples ¼ 30% CD19 þ CD5 þ counted and identified by staining with CD19/CD5 mAbs. Each cells, as assessed by flow cytometry): 30/52 CD38 þ and 22/52 experiment was performed in triplicate and results were CD38À. expressed as percentage of input calculated as ratio between rCX3CL1-induced migration and spontaneous migration. In some experiments, CLL cells were pretreated 1 h at 37 1C with 100 ng/ml Cell isolation and cultures pertussis toxin (PTX) (Sigma) and subjected to chemotaxis to Peripheral blood mononuclear cells (PBMCs) were isolated by rCX3CL1 (ref. 17). In other experiments, CLL cells were pre- Ficoll–Hypaque gradients and cryopreserved in freezing solu- incubated for 1 h at 37 1C with the following: PI3K inhibitor tion containing 50% RPMI 1640 (Sigma Chemical Company, LY294002 (20 mM; Sigma), Akt inhibitor (1 mM; Calbiochem, St Louis, MO, USA), 40% fetal bovine serum (Sigma), and 10% Gibbstown, NJ, USA), MEK inhibitor