Chronic Lymphocytic Leukemia-Associated Macrophages: Not Just Nurse Cells

CLL cells interact with many accessory cells in an environment mimicking that of normal mature B cells. Role of antigen, cytokines, adhesion pathways are critical for many aspects in the disease course (proliferation/survival, migration or homing, drug resistance, and presumably relapse). Nurse-like cells (NLC) belong to a monocytic-derived, bystander population among CLL lymph node and spleen stromal cells.

To investigate the nature, functions, and location of NLC within CLL microenvironment.

Gene expression profiles (GEP) from in vitro expanded NLC from patients (n=10) were produced and compared to those from normal CD14⁺ monocytes, M1-polarized macrophages, M2-polarized macrophages and tumor-associated macrophages (produced in the lab or downloaded from GEO datasets). Principal Component Analysis was used to categorize these five populations of cells and in-house-built GSEA software was used for functional interpretation of their relevant gene lists. Protein expression patterns were validated with multi-analyte ELISArray kits, proteome profiler arrays, flow cytometry (FC) or immunohistochemistry (IHC).

New insights into the physiopathological role of NLC in CLL are suggested from five lines of evidence:

1/a Òmonocytic gene signatureÓ (i.e. a set of 549 genes) is shared by the NLC and the monocyte subtypes. The genes over-represented in NLC vs normal monocytes pinpointed positive modulation of apoptotic cell clearance (scavenger, mannose and complement receptors, LXRalpha), lipid metabolism (Apolipoprotein E, PPAR signaling), extracellular matrix-receptor interactions (integrins, SPARC, Matrix MetalloProteinases) and actin cytoskeleton remodeling.

2/unsupervised clustering show that NLC represent an M2-skewed, TAM-like cell population. They down-regulate mRNA and proteins for classic M1 inflammatory markers (e.g. IL-1, IL-6, IL-12, COX2) while increase secretion of TGFbeta, IL-10, CCL17 and CCL22 soluble factors.

3/these and previously published observations suggest that B-CLL-to-NLC interactions may orchestrate immunosuppression in this disease. PBMCs from Òwatch and waitÓ CLL patients (all stage A/Rai 0, mutated IgVH, low risk cytogenetics profile) or healthy donors were stimulated with anti-CD3/CD28 beads + IL-2, either in standard RPMI+10% FCS or in conditioned medium (CM, after 14d CLL-NLC co-culture in vitro) and their proliferation/phenotype were compared after 2 weeks. Significant expansion of T cells with Treg (CD4+CD25+FoxP3+) phenotype was observed only from CLL PBMCs grown in conditioned medium (mean % Treg: 2.85 vs 3.05 in CM for normal PBMCs, and 1.54 vs 15.9 in CM for CLL PBMCs, P< 0.05).

4/although NLC make immune synapses with live B-CLL, they do not phagocytose them. Over-expression of CD47 (ÒdonÕt eat meÓ signal) by B-CLL cells (mfi= 3490 vs 2581 on normal cells, P< 0.05, n=18) may provide them with a protective signal against NLC.

5/from our GEP, flow cytometric and IHC analyses, we propose CD163 (classic M2 marker) as a reliable tool to identify NLC in vivo. Although in vitro, CLL cells can pervert healthy donor monocytes into NLC, only CLL-derived NLC are truly CD14+ CD163+. In vivo, CD163 staining reveals putative NLC in CLL lymph nodes(LN)/spleen sections but not in bone marrow. In LN from all patients, NLC reside in the subcapsular areas and line vessel structures, suggesting a role in CLL cells trafficking. Most interestingly, NLC infiltrate pseudofollicles structures only in a subset of cases. We will present updated IHC and clinical presentation correlation studies.

Our results suggest that the role of NLC in CLL might be broader than initially thought. Beside of nursing and conferring drug resistance, NLC may also be crucial in the setting of immunosuppression, of CLL cells recruitment, and should thus be considered as therapeutic targets.

Off Label Use: GA101 is not currently approved for CLL treatment.