CCL3 and CCL4, the Major Chemokines Produced by CD38+ Chronic Lymphocytic Leukemia Cells, Facilitate Microenvironmental Interactions of Neoplastic Cells Via the CD49d/VCAM Pair.

CD38, a negative prognostic marker for patients with CLL, has been demonstrated to be a key molecule in the interactions occurring in the context of tumor microenvironment, mediating both survival and migratory signals for CLL cells. By taking advantage of gene expression profiling studies (GEP) comparing 11 CD38pos (CD38>30%) and 15 CD38neg (CD38<10%) CLLs, we identified as over-expressed in CD38pos CLL cells: i) genes for the two C-C chemokines CCL3 and CCL4 (median-log difference, MLD-CCL3= 3.5; MLD-CCL4=4.4); real-time quantitative PCR (RTQ-PCR) of selected cases confirmed GEP results; ii) the gene for CD49d (MLD=4.4); a high correlation between CD38 and CD49d protein expression, also characterizing the CLL series of the present study, has been reported previously. In vitro experiments, performed on purified tumor cells from additional 11 CD38pos CLL cases cultured for 14 (t14) and 24 (t24) hours in the presence of either the agonist anti-CD38 monoclonal antibody (mAb) IB4 or the non-agonistic anti-CD38 mAb IB6 as control, demonstrated upregulation of CCL3/CCL4 transcripts at t14 (CCL3: mean fold increase=18, p=0.041; CCL4: mean fold increase=13.8, p=0.005), as assessed by RTQ-PCR, and an increased release of CCL3/CCL4 proteins at t24 (CCL3: mean =0.9 ng/mL, mean fold increase=14, p=0.003; CCL4: mean =1.7 ng/mL, mean fold increase=49, p=0.01), as assessed by ELISA. Consistently, immunohistochemistry (IHC) analysis performed in bone marrow biopsies (BMB) from 20 CLL patients (10 CD38pos and 10 CD38neg cases) showed detectable levels of CCL3 in 8 cases, all but one belonging to the CD38pos group (p=0.02). Expression of the CCL3/CCL4 specific receptors CCR1 and CCR5 was examined by flow cytometry in peripheral blood cell subpopulations from 30 CLL (12 CD38pos and 18 CD38neg). Irrespectively of CD38 expression by CLL cells, monocytes showed the highest expression levels for CCR1 and, although at a lesser extent, CCR5. Consistently, CCL3 was able to attract CLL-derived monocytes by in-vitro chemotaxis experiments, and a higher number of infiltrating CD68pos macrophages were found in BMB of CD38pos compared to CD38neg CLLs (p=0.016). In parallel experiments, conditioned media (CM) from CCL3-stimulated macrophages were collected; these CM were able to induce expression of the CD49d-ligand VCAM in human umbilical vein endothelial cells (HUVEC) and human microvascular endothelial cells (ADMEC). As shown by ELISA, TNFalpha was among the cytokines contained in macrophage-CM. This citokine was likely responsible for VCAM up-regulation by HUVEC and ADMEC, as suggested by TNFalpha neutralization experiments leading to a suppression of VCAM-1 induction in endothelial cell models. Again, IHC analysis of CLL BMB showed a meshwork of VCAM-1-positive cells more prominent in the context of lymphoid infiltrates of CD38pos, as compared to CD38neg cases (p=0.002). To verify whether CD49d engagement through VCAM-1 could enhance the protection against spontaneous apoptosis of CLL cells in vitro, we cultured purified CD38pos/CD49dpos CLL cells from 5 cases onto VCAM-1-transfected L cells or mock-transfected L cells. Results demonstrated a substantial improvement in cell viability after CD49d engagement: as high as 70%±25 cells were viable after 10 days of culture on L-VCAM cells compared to 50%±25 in control conditions (p=0.009). Altogether, these results identify molecules involved in a functional cross-talk between CD38/CD49d-expressing CLL and cells of the tumor microenvironment. This interplay may eventually affect survival and recirculation of tumor cells via the CD49d/VCAM pair.