Chronic Lymphocytic Leukemia Cells With Trisomy 12 Home To The Bone Marrow In a CXCR4-Independent Manner and Are Prone To Proliferate In Vitro

Trisomy 12 (tri12) is a frequent chromosomal aberration in chronic lymphocytic leukemia (CLL) associated with atypical cell morphology, high in vivo tumor proliferation activity and a predisposition to Richter’s transformation. Tri12 harboring CLL cells express increased levels of the negative prognostic marker CD49d, the α4 subunit of the integrin very late antigen 4 (VLA-4), which we previously identified as a key regulator of CLL cell homing to bone marrow (BM). During this process, inside-out activation of VLA-4 upon CXCR4 binding to endothelially displayed CXCL12 is thought to upregulate the adhesive properties of VLA-4 and augment the arrest of CLL cells on the VCAM-1 presenting vessels.

Here, we investigated the functional interplay of VLA-4 and CXCR4 in CLL carrying tri12. We first found that the upregulation of CD49d expression in this subset (MFIR CD49d 9.8±5.3 (n=22) vs. 2.7±3.9 (n=126), p<0.0001) was paralleled by their reduced CXCR4 expression (MFIR CXCR4 11.8±7.2 (n=22) vs. 22.7±14.2 (n=126), p=0.0003). Using short term adoptive transfers, we compared the ability of tri12 and no tri12 CLL cells to home to the BM of NOD/SCID mice. 5-10x10⁶ CLL cells were injected into tail vein and homing was evaluated after 3 hours. Based on their more frequent CD49d high phenotype, we observed increased homing rates (homed human CLL cells per 10⁶ injected cells per 10⁶ acquired murine cells) of tri12 compared to no tri12 CLL (225±160 (n=7) vs. 90±117 (n=20), p=0.025). However, when comparing CD49d+ tri12 and CD49d+ no tri12 subsets, we did not observe any significant differences in their homing capacity. To further study CXCL12/CXCR4 function in BM homing, we pretreated mice with either the novel CXCL12 antagonist NOX-A12 or the CXCR4 inhibitor AMD3100 prior to CLL cell injection. While homing of no tri12 CLL cells (n=3, in duplicates) was reduced by both pretreatments (homing rates 137 vs 38 vs 30), the homing capacity of tri12 CLL cells (n=3, in duplicates) was not affected.

We next tested whether VLA-4 expressed on these cells was able to undergo CXCL12-induced activation and support cell arrest under shear conditions. To this end, we perfused CLL cells over VCAM-1 or VCAM-1/CXCL12 substrates and analyzed rates and categories of cell tethering at a single cell level by videomicroscopy. CXCL12 induced the arrests of no tri12 CLL cells (n=3) on VCAM-1 under shear flow in a CXCR4 and VLA-4 dependent manner. In contrast, tri12 CLL cells (n=3) robustly tethered to VCAM-1 in the absence of the chemokine, and interactions could not be further enhanced by additional CXCL12 nor could they be abrogated by use of AMD3100. This failure of CXCR4-induced adhesion was not based on a general defect in CXCR4 functionality as in vitro chemotaxis of tri12 CLL cells (n=5) towards CXCL12 was fully maintained. To detect potential differences in VLA-4 affinity regulation, we used a conformationally sensitive antibody that recognizes epitopes induced by VLA-4 ligation, and an LDV-containing VLA-4 specific ligand to probe resting integrin affinity. Also, we used a small fluorescent ligand to study rapid VLA-4 affinity changes during inside-out chemokine induced activation. On resting tri12 CLL, VLA-4 exhibited an affinity state similar to that observed on circulating lymphocytes, and tri12 CLL cells failed to undergo the rapid affinity up-regulation triggered by CXCL12 pretreatment, in keeping with tethering experiments.

Next, we investigated whether the tumor microenvironment has a different influence on the behavior of the tri12 subset. Therefore we subjected the cells to in vitro co-cultures mimicking the lymphoid proliferation centers. Basal levels of the early activation marker CD69 were similar in tri12 CLL compared to no tri12 cases. Tri12 CLL, however, underwent stronger activation when cultured in presence of accessory cells (%CD69+ cells 60.0±18.5 (n=4) vs. 17.7±20.1 (n=19), p=0.008). Moreover, in several setups, proliferation rates of these cells were increased, irrespective of the proliferative stimulus and detection method used.

In summary, our results provide a mechanistical basis at least in part explaining the peculiar and clinical features of the tri12 CLL subset. In light of the specific migratory and proliferative properties of tri12 cells and novel agents targeting particularly these functions, our findings may also imply therapeutical consequences.